Selenium: Its Molecular Biology and Role in Human Health

SELENIUM Its Molecular Biology and Role in Human Health, Second Edition

SELENIUM Its Molecular Biology and Role in Human Health, Second Edition

Edited by Dolph L. Hatfield National Cancer Institute, USA Maria J. Berry University of Hawaii, USA and Vadim N. Gladyshev University of Nebraska, USA

^

rineer Spri

Library of Congress Control Number: 2006924112 lSBN-10; 0-387-33826-8 ISBN-13: 978-0-387-33826-2

e-ISBN-!0: 0-387-33827-6

Printed on acid-free paper.

© 2006 Springer Science+Business Media, LLC. All rights reserved. This work may not be translated or copied in whole or in part without the written permission of the publisher (Springer Science-t-Business Media, LLC, 233 Spring Street, New York, NY 10013, USA), except for brief excerpts in connection with reviews or scholarly analysis. Use in connection with any form of information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed is forbidden. The use in this publication of trade names, trademarks, service marks and similar terms, even if they are not identified as such, is not to be taken as an expression of opinion as to whether or not they are subject to proprietary rights. Printed in the United States of America. 9 8 7 6 5 4 3 2 1 springer.com

TABLE OF CONTENTS Contributors

xi

Foreword Raymond F. Burk

xvii

Preface

xxi

Dolph L. Hatfield, Maria J. Berry and Vadim N. Gladyshev Acknowledgements

xxiii

Chapter 1 Selenium: A historical perspective James E. Oldfield

1

Part I. Biosynthesis of selenocysteine and its incorporation into protein Chapter 2 Selenium metabolism in prokaryotes August Bock, Michael Rother, Marc Leibundgut and Nenad Ban

9

Chapter 3 Mammalian and other eukaryotic selenocysteine tRNAs 29 Bradley A. Carlson, Xue-Ming Xu, Rajeev Shrimali, Aniruddha Sengupta, Min-Hyuk Yoo, Robert Irons, Nianxin Zhong, Dolph L. Hatfield, Byeong Jae Lee, Alexey V. Lobanov and Vadim N. Gladyshev Chapter 4 Evolution of selenocysteine decoding and the key role of selenophosphate synthetase in the pathway of selenium utilization 39 Gustavo Salinas, Hector Romero, Xue-Ming Xu, Bradley A. Carlson, Dolph L. Hatfield and Vadim N. Gladyshev Chapter 5 SECIS RNAs and K-turn binding proteins. A survey of evolutionary conserved RNA and protein motifs Christine Allmang and Alain Krol

51

vi

Selenium: Its molecular biology and role in human health

Chapter 6 SECIS binding proteins and eukaryotic selenoprotein synthesis Donna M. Driscoll and Paul R. Copeland Chapter 7 The importance of subcellular localization of SBP2 and EFsec for selenoprotein synthesis Peter R. Hoffmann and Maria J. Berry Chapter 8 Selenocysteine biosynthesis and incorporation may require supramolecular complexes Andrea L. Small-Howard and Maria J. Berry

63

73

83

Part II. Selenium-containing proteins Chapter 9 Selenoproteins and selenoproteomes Vadim N. Gladyshev

99

Chapter 10 Deletion of selenoprotein P gene in the mouse Raymond F. Burk, Gary E. Olsen and Kristina E. Hill

Ill

Chapter 11 Selenium and methionine sulfoxide reduction Hwa-Young Kim and Vadim N. Gladyshev

123

Chapter 12 Selenoprotein W in development and oxidative stress Chrissa Kioussi and Philip D. Whanger

135

Chapter 13 The 15-kDa selenoprotein (SeplS): functional analysis and role in cancer 141 Vyacheslav M. Labunskyy, Vadim N. Gladyshev and Dolph L. Hatfield Chapter 14 Regulation of glutathione peroxidase-1 expression Roger A. Sunde

149

Table of Contents

Chapter 15 Selenoproteins of the glutathione system Leopold Flohe and Regina Brigelius-Flohe Chapter 16 New roles of glutathione peroxidase-1 in oxidative stress and diabetes Xin Gen Lei and Wen-Hsing Cheng Chapter 17 Selenoproteins of the thioredoxin system Arne Holmgren Chapter 18 Mitochrondrial and cytosolic tliioredoxin reductase loiocliout mice Marcus Conrad, Georg W. Bornkamm and Marcus Brielmeier

vii

161

173

183

195

Chapter 19 Selenium, deiodinases and endocrine function Antonio C. Bianco and P. Reed Larsen

207

Chapter 20 Biotechnology of selenium Linda Johansson and Elias S.J. Arner

221

Part III. Selenium and human health Chapter 21 Selenium, selenoproteins and brain function Ulrich Schweizer and Lutz Schomberg

233

Chapter 22 Selenium as a cancer preventive agent Gerald F. Combs, Jr. and Junxuan Lii

249

Chapter 23 Peering down the kaleidoscope of thiol proteomics and unfolded protein response in studying the anticancer action of selenium 265 Ke Zu, Yue Wu, Young-Mee Park and Clement Ip

viii

Selenium: Its molecular biology and role in human health

Chapter 24 Genetic variation among selenoprotein genes and cancer Alan M. Diamond and Rhonda L. Brown

277

Chapter 25 Selenium and viral infections MelindaA. Beck

287

Chapter 26 Role of selenium in HIV/AIDS Marianna K. Baum and Adriana Campa

299

Chapter 27 Effects of selenium on immunity and aging Roderick C. McKenzie, Geoffrey J. Becket and John R. Arthur

311

Chapter 28 Selenium and male reproduction 323 Matilde Maiorino, Antonella Roveri, Fulvio Ursini, Regina Brigelius-Flohe and Leopold Flohe Chapter 29 Mouse models for assessing the role of selenium in health and development 333 Bradley A. Carlson, Xue-Ming Xu, Rajeev Shrimali, Aniruddha Sengupta, Min-Hyuk Yoo, Nianxin Zhong, Dolph L Hatfield, Robert Irons, Cindy D. Davis, Byeong Jae Lee, Sergey V. Novoselov and Vadim N. Gladyshev Chapter 30 Drosophila as a tool for studying selenium metabolism and role of selenoproteins Cristina Pallares, Florenci Serras and Montserrat Corominas Chapter 31 Selenoproteins in parasites Gustavo Salinas, Alexey V. Lobanov and Vadim N. Gladyshev Chapter 32 Incorporating 'omics' approaches to elucidate the role of selenium and selenoproteins in cancer prevention Cindy D. Davis and John A. Milner

343

355

367

Table of Contents

ix

Chapter 33 Selenium-induced apoptosis Ick Young Kim, Tae Soo Kim, Youn Wook Chung and Daewon Jeong

379

Chapter 34 Selenoprotein mimics Junqiu Liu and Guimin Luo

387

Chapter 35 Update of human dietary standards for selenium Orville A. Levander and Raymond F. Burk

399

Index

411

Contributors Christine Allmang

Maria J. Berry

Architecture et Reactivite de I'arN UPR 9002 du CNRS-Universite Louis Pasteur Institut de Biologie Moleculaire et Cellulaire 67084 Strasbourg, France

Department of Cell and Molecular Biology John A. Bums School of Medicine University of Hawaii at Manoa Honolulu, HI 96813, USA

Antonio C. Bianco Elias S. J. Arner Medical Nobel Institute for Biochemistry Department of Medical Biochemistry and Biophysics Karolinska Institute SE-171 77 Stockholm, Sweden

John R. Arthur Division of Vascular Health Rowett Research Institute Bucksbum, Aberdeen, Scotland AB219SB,UK

Nenad Ban Institute of Molecular Biology and Biophysics Swiss Federal Institute of Technology ETH Hfinggerberg, HPK Building CH-8093 Zurich, Switzerland

Marianna K. Baum

Thyroid Section, Division of Endocrinology Diabetes and Hypertension Department of Medicine Brigham and Women's Hospital and Harvard Medical School 77 Avenue Louis Pasteur Boston, MA 02115, USA

August Bdck Lehrstuhl flir Mikrobiologie der Universitat Munchen, D-80638 Munich, Germany

Georg W. Bornkamm Institute of Clinical Molecular Biology and Tumor Genetics GSF-Research Centre for Environment and Health 81377 Munich, Germany

Florida International University Stempel School of Public Health Department of Dietetics and Nutrition 11200 SW 8th Street Miami, FL 33199, USA

Markus Brielmeier

Melinda A. Beck

Regina Brigelius-Flohe

Department of Nutrition University of North Carolina at Chapel Hill Chapel Hill, NC 27599, USA

Department Biochemistry of Micronutrients German Institute of Human Nutrition Potsdam-Rehbruecke (DIfE) Arthur-Scheunert-Allee 114-116 D-14558 Nuthetal, Germany

Geoffrey J. Beckett Department of Clinical Biochemistry University of Edinburgh Combined Laboratories The Royal Infirmary of Edinburgh 51 Little France Cresdent Edinburgh, Scotland, EH 16 4SA, UK

Department of Comparative Medicine GSF-Research Centre for Environment and Health 85764 Neuherberg, Germany

Rhonda L. Brown Department of Human Nutrition University of Illinois at Chicago Chicago, IL 60612, USA

Xll

Selenium: Its molecular biology and role in human health

Raymond F. Burk

Paul R. Copeland

Division of Gastroenterology, Hepatology, and Nutrition Department of Medicine Vanderbilt University School of Medicine Nashville, TN 37232, USA

Department of Molecular Genetics Microbiology and Immunology UMDNJ - Robert Wood Johnson Medical School Piscataway, NJ 08854, USA

Bradley A. Carlson

Montserrat Corominas

Molecular Biology of Selenium Section Laboratory of Cancer Prevention Center for Cancer Research National Cancer Institute National Institutes of Health Bethesda, MD 20892, USA

Departament de Geneica Universitat de Barcelona, Diagonal 645 08028 Barcelona, Spain

Adriana Campa Florida International University Stempel School of Public Health Department of Dietetics and Nutrition 11200 SW 8th Street Miami, FL 33199, USA

Wen-Hsing Cheng Laboratory of Molecular Gerontology National Institute on Aging National Institutes of Health Bahimore, MD 21224, USA

Youn Wook Chung Laboratory of Cellular and Molecular Biochemistry School of Life Sciences and Biotechnology Korea University 1,5-Ka, Anam-Dong Sungbuk-Ku Seoul 136-701, Korea

Gerald F. Combs, Jr. Grand Forks Human Nutrition Research Center, USDA-ARS Grand Forks, ND 58202, USA

Marcus Conrad Institute of Clinical Molecular Biology and Tumor Genetics GSF-Research Centre for Environment and Health 81377 Munich, Germany

Cindy D. Davis Nutritional Science Research Group Division of Cancer Prevention National Cancer Institute National Institutes of Health Bethesda, MD 20892, USA

Alan M. Diamond Department of Human Nutrition University of Illinois at Chicago Chicago, IL 60612, USA

Donna M. DriscoU Department of Cell Biology Lemer Research Institute Cleveland Clinic Foundation Cleveland, OH 44195, USA

Leopold Flohe MOLISA GmbH Universitatsplatz 2 D-39106 Magdeburg, Germany

Vadim N. Gladyshev Department of Biochemistry University of Nebraska Lincoln, NE 68688, USA

Dolph L. Hatfield Molecular Biology of Selenium Section Laboratory of Cancer Prevention Center for Cancer Research National Cancer Institute National Institutes of Health Bethesda, MD 20892, USA

Contributors

xiu

Kristina E. HiU

Hwa-Young Kim

Division of Gastroenterology, Hepatology, and Nutrition Department of Medicine Vanderbilt University School of Medicine Nashville, TN 37232, USA

Department of Biochemistry University of Nebraska Lincoln, Nebraska 68588, USA

Peter R. Hoffmann John A. Bums School of Medicine Department of Cell and Molecular Biology University of Hawaii at Manoa Honolulu, HI 96813, USA

Ick Young Kim Laboratory of Cellular and Molecular Biochemistry School of Life Sciences and Biotechnology Korea University 1,5-Ka, Anam-Dong Sungbuk-Ku Seoul 136-701, Korea

Arne Holmgren Medical Nobel Institute for Biochemistry Department of Medical Biochemistry and Biophysics Karolinska Institute SE-171 77 Stockholm, Sweden

Clement Ip Department of Cancer Chemoprevention Roswell Park Cancer Institute Buffalo, NY 14263, USA

Tae Soo Kim Laboratory of Cellular and Molecular Biochemistry School of Life Sciences and Biotechnology Korea University 1,5-Ka, Anam-Dong Sungbuk-Ku Seoul 136-701, Korea

Chrissa Kioussi Robert Irons Nutritional Science Research Group Division of Cancer Prevention and Molecular Biology of Selenium Section Laboratory of Cancer Prevention Center for Cancer Research National Cancer Institute National Institutes of Health Bethesda, MD 20892, USA

Daewon Jeong BK2I HLS, Seoul National University 28 Yeonkun-Dong Chongno-Ku Seoul 110-749, Korea

Linda Johansson Medical Nobel Institute for Biochemistry Department of Medical Biochemistry and Biophysics Karolinska Institute SE-171 77 Stockholm, Sweden

Department of Biochemistry and Biophysics Oregon State University Corvallis, OR 97331, USA

Alain Krol Architecture et Reactivite de I'arN UPR 9002 du CNRS-Universite Louis Pasteur Institut de Biologic Moleculaire et Cellulaire 67084 Strasbourg, France

Vyacheslav M. Labunskyy Department of Biochemistry University of Nebraska Lincoln, NE 68688, USA

XIV

Selenium: Its molecular biology and role in human health

P. Reed Larsen

Junxuan Lii

Thyroid Section, Division of Endocrinology Diabetes and Hypertension Department of Medicine Brigham and Women's Hospital and Harvard Medical School 77 Avenue Louis Pasteur Boston, MA 02115, USA

Hormel Institute University of Minnesota Austin, MN 55912, USA

Guimin Luo Key Laboratory for Molecular Enzymology and Engineering Jilin University Changchun 130023, China

Byeong Jae Lee Laboratory of Molecular Genetics Institute of Molecular Biology and Genetics School of Biological Sciences Seoul National University Seoul 151-742, Korea

Matilde Maiorino

Xin Gen Lei

Roderick C. McKenzie

Department of Animal Science Cornell University Ithaca, NY 14853, USA

Marc Leibundgut Institute of Molecular Biology and Biophysics Swiss Federal Institute of Technology ETH Honggerberg, HPK Building CH-8093 Zflrich, Switzerland

Department of Biological Chemistry University of Padova Viale G. Colombo, 3 1-35121 Padova, Italy

Laboratory for Clinical and Molecular Virology Royal Dick Veterinary School University of Edinburgh Summerhall, Edinburgh EH9 IQH, UK

John A. Milner

Orville A. Levander

Nutritional Science Research Group Division of Cancer Prevention National Cancer Institute National Institutes of Health Bethesda, MD 20892, USA

Beltsville Human Nutrition Research Center U. S. Department of Agriculture Agricultural Research Service Beltsville, MD 20705, USA

Department of Biochemistry University of Nebraska Lincoln, NE 68688, USA

Junqiu Liu

James E. Oldfield

Key Laboratory for Supramolecular Structure and Materials Jilin University Changchun 130012, China

Alexey V. Lobanov Department of Biochemistry University of Nebraska Lincoln, NE 68688, USA

Sergey V. Novoselov

Oregon State University Corvallis, OR 97331, USA

Gary £. Olson Division of Gastroenterology, Hepatology, and Nutrition Department of Cell and Developmental Biology Vanderbilt University School of Medicine Nashville, TN 37232, USA

Contributors

XV

Cristina Pallar^s

Ulrich Schweizer

Departament de Geneica Universitat de Barcelona Diagonal 645 08028 Barcelona, Spain

Neurobiology of Selenium Neuroscience Research Center and Institute for Experimental Endocrinology Charit^-Universitatsmedizin Berlin Charite Campus Mitte D-10117 Berlin, Germany

Young-Mee Park Department of Cellular Stress Biology Roswell Park Cancer Institute Buffalo, NY 14263, USA

Hector Romero Laboratorio de Organizacion y Evoluci6n del Genoma Dpto. de Biologia Celular y Molecular Instituto de Biologia Facultad de Ciencias Igua4225 Montevideo, CP 11400, Uruguay

Michael Rother Lehrstuhl fUr Mikrobiologie der Universitat Mflnchen D-80638 Munich, Germany

Antonella Roveri Department of Biological Chemistry University of Padova Viale G. Colombo, 3 1-35121 Padova, Italy

Aniruddlia Sengupta Molecular Biology of Selenium Section Laboratory of Cancer Prevention Center for Cancer Research National Cancer Institute National Institutes of Health Bethesda, MD 20892, USA

Florenci Serras Departament de Geneica Universitat de Barcelona Diagonal 645 08028 Barcelona, Spain

Rajeev Shrimali Molecular Biology of Selenium Section Laboratory of Cancer Prevention Center for Cancer Research National Cancer Institute National Institutes of Health Bethesda, MD 20892, USA

Andrea L Small-Howard Gustavo Salinas Cdtedra de Inmunologia Facultad de Quimica-Facultad de Ciencias Universidad de la Repiiblica Instituto de Higiene Avda. A. Navarro 3051 Montevideo, CP 11600, Uruguay

Lutz Sclioinburg Institute for Experimental Endocrinology Charit^-Universitatsmedizin Berlin Charity Campus Mitte D-10117 Berlin, Germany

Department of Cell and Molecular Biology John A. Bums School of Medicine University of Hawaii at Manoa Honolulu, HI 96813, USA

Roger A. Sunde 1415 Linden Drive University of Wisconsin Madison, WI53705, USA

Fulvio Ursini Department of Biological Chemistry University of Padova Viale G. Colombo, 3 1-35121 Padova, Italy

xvi

Selenium: Its molecular biology and role in human health

Philip D. Whanger Department of Environmental and Molecular Toxicology Oregon State University Corvallis, OR 97331, USA

YueWu Department of Cancer Chemoprevention Roswell Park Cancer Institute Buffalo, NY 14263, USA

Xue-Ming Xu Molecular Biology of Selenium Section Laboratory of Cancer Prevention Center for Cancer Research National Cancer Institute National Institutes of Health Bethesda, MD 20892, USA

Min-Hyuk Yoo Molecular Biology of Selenium Section Laboratory of Cancer Prevention Center for Cancer Research National Cancer Institute National Institutes of Health Bethesda, MD 20892, USA

Nianxin Zhong Molecular Biology of Selenium Section Laboratory of Cancer Prevention Center for Cancer Research National Cancer Institute National Institutes of Health Bethesda, MD 20892, USA

KeZu Department of Cancer Chemoprevention Roswell Park Cancer Institute Buffalo, NY 14263, USA

Foreword The discovery of selenoproteins in 1973 was the starting point for today's flourishing selenium field [1,2]. It provided evidence that selenium had biochemical functions that could account for its nutritional effects [3,4]. Further, it opened the selenium field to investigation by the methods of biochemistry, which led to the identification of several more selenoproteins and showed that selenocysteine was the form of the element in animal selenoproteins and in most bacterial ones. Although noteworthy efforts were made to uncover the mechanism of selenocysteine and selenoprotein synthesis using biochemical methods, the problem yielded only when attacked with the methods of molecular biology [5,6]. The bacterial mechanism was characterized first; characterization of the animal mechanism is a work in progress. It is interesting to note that the only genes that are devoted to selenium metabolism are those that support selenoprotein synthesis and selenocysteine catabolism. Consequently, it seems likely that competition for selenium between selenoprotein synthesis and the production of selenium excretory metabolites [7] controls wholebody selenium homeostasis. The physiological functions of selenium derive fi-om the catalytic and physical properties of selenoproteins. Selenoproteins such as the glutathione peroxidases and the thioredoxin reductases have redox activities that allow them to serve in oxidant defense. The deiodinases use their redox activities to activate and inactivate thyroid hormones. From these two examples, it can be seen that selenoprotein functions are diverse while having in common a redox mechanism. Although a few of the biological functions of selenium have been identified, many have not. Application of bioinformatics techniques to genomic databases has identified 25 genes for selenoproteins in the human genome [8]. Most of the proteins represented by those genes have not been characterized to the point where their functions can be assessed. Thus, one of the major challenges in selenium research is to characterize all the selenoproteins so that their biological activities can be determined. The ultimate goal of selenium research is to improve human health. Veterinary and animal science investigators had already demonstrated that nutritional selenium deficiency occurred in animals fed plants firom areas with low soil selenium availability when, in 1979, Chinese researchers reported the existence of a selenium-responsive disease in such an area. Their study showed convincingly that the occurrence of Keshan disease, a childhood cardiomyopathy, could be prevented by selenium supplementation [9]. Although several other diseases have been postulated to be selenium deficiency conditions, studies to prove those claims have not appeared. Thus,

xviii

Selenium: Its molecular biology and role in human health

Keshan disease, which has almost disappeared from China as economic conditions have improved, remains the extreme example of pathology that can occur in selenium deficient human beings. While selenium deficiency severe enough to allow the occurrence of Keshan disease is rare, people in many areas of the world have selenium intakes that are not sufficient to allow full expression of all selenoproteins. New Zealand and many countries in Europe fall into this category. In response to learning that its selenium status was low, Finland chose in 1985 to add selenium to its fertilizer. It has thereby become a laboratory for studying the effects of supplementing a population with selenium. The selenium status of Firms rapidly became comparable to that of North Americans but without discemable effects on the incidences of major diseases [10]. This type of study without a control population would not be expected to detect subtle health effects or uncommon ones such as altered responses to drugs: so the question of whether full expression of selenoproteins is needed for optimum health must remain open. This issue needs attention from clinical investigators because of the large number of people affected and the implications it has for setting official dietary requirements for selenium. More directly related to basic selenium research, mutations and polymorphisms of selenoprotein genes and of genes involved in selenoprotein synthesis can cause human disease. An example of this is the congenital muscle disease that results from mutation of the gene for selenoprotein N, one of the selenoproteins of unknown function. Perhaps elucidation of the function of selenoprotein N will suggest a treatment for the muscle disease. Phenotypes of mice with deletion of a selenoprotein might be instructive in this respect. For example, deletion of selenoprotein P causes neurological dysfunction that can be prevented by selenium supplements above the nutritional requirement. If an analogous human condition were found, selenium supplements might be efficacious in its treatment. Examples of animal research that support understanding of human diseases stimulate basic selenium research. In addition to research on the physiological functions of selenium, considerable enthusiasm has been generated for studying the effects of pharmacological doses of the element. The results of numerous animal studies and limited human trials have suggested that administration of pharmacological doses of selenium can prevent some kinds of cancer [11]. Additional trials are underway to test this hypothesis. If such a chemopreventive effect of selenium can be proven, it would not likely be linked to the selenoproteins because the subjects in the trials were not selenium deficient before supplementation was started. This means that the selenoproteins would have been at their optimal levels initially and that selenium supplements would not have been expected to affect them. Other

Foreword

xix

metabolic effects of high selenium intake have been noted, however, and might account for its effects on cancer development. It will be important for public health reasons to determine whether selenium is an effective chemopreventive agent in human beings and, if it is, to determine the safety of pharmacological doses of selenium. Many tasks remain in the selenium field. Additional characterization of individual selenoproteins and elucidation of the mechanism of selenoprotein synthesis are needed to facilitate identification of pathological conditions involving selenium. Clinical studies are needed to determine the selenium intake needed to ensure full expression of all selenoproteins and to assess the health implications of selenium intakes that do not allow full expression of all selenoproteins. And, finally, whether selenium is efficacious as a chemopreventive agent needs to be determined. References 1. 2.

DC Turner, TC Stadtman 1973 Arch Biochem Biophys 154:366 JT Rotruck, AL Pope, HE Ganther, AB Swanson, D Hafeman WG Hoekstra 1973 Science 179:588 3. KE McCoy, PH Weswig 1969 JNutr 98:383 4. K Schwarz, CM Foltz 1957 J Amer Chem Soc 79:3292 5. A Bock, K Forchhammer, J Heider, C Baron 1991 Trends Biochem Sci 16:463 6. I Chambers, J Frampton, P Goldfarb, N Affara, W McBain, PR Harrison 1986 EMBO J 5:1221 7. Y Kobayashi, Y Ogra, K Ishiwata, H Takayama, N Aimi, KT Suzuki 2002 Proc Natl Acad Sci U S A 99: 15932 8. GV Kryukov, S Castellano, SV Novoselov, AV Lobanov, O Zehtab, R Guigo, VN Gladyshev 2003 Science 300:1439 9. Keshan Disease Research Group 1979 Chinese Medical Journal 92:471 10. M Eurola, G Alfthan, A Aro, P Ekholm, V Hietaniemi, H Rainio, R Rankanen, E-R Venalainen 2003 Agrifood Research Reports 36 Results of the Finnish selenium monitoring program MTT Agrifood Research Finland, FIN-31600 Jokioinen, Finland pp 42 11. LC Clark, GF Combs Jr, BW TumbuU, EH Slate, DK Chalker, J Chow, LS Davis, RA Glover, GF Graham, EG Gross, A Krongrad, JL Lesher, HK Park, BB Sanders, CL Smith, JR Taylor 1996 JAMA 276:1957

Raymond F. Burk

Preface Since the first edition of Selenium: Its Molecular Biology and Role in Human Health was published in 2001, many new insights into the biochemical, molecular, genetic and health aspects of this fascinating element have been elucidated. Several new human clinical trials have also been undertaken examining the role of selenium in protection against different cancers. For example, the National Cancer Institute initiated two new clinical trials involving selenium. One of these is called SELECT, Selenium and vitamin E Cancer Prevention Trial, and it involves examining the role of selenium and vitamin E in protecting against prostate cancer, with a goal of enrolling over 35,000 males in the study. The other trial involves examining the role of selenium in protection against lung cancer, a study incorporating 1960 individuals. The commitment of hundreds of millions of dollars to these trials for examining the role of selenium in protecting humans against different forms of cancer illustrates how highly important this element is regarded by the medical and scientific communities in health issues. What is of such significance to elucidating the role of selenium in health in these human clinical trials is that not only will the effect of selenium on prostate and lung cancers be assessed, but these trials will shed light on the role of many additional aspects of selenium in health such as aging, heart disease, viral inhibition and other forms of cancer including colon, liver and brain malignancies. Many exciting discoveries have occurred in the last five years which are described in the current edition. For example, the entire selenoprotein gene population, designated the selenoproteome, has been identified in humans and rodents. Furthermore, the various selenoproteins described in the last edition have been further characterized and their new features described. Numerous selenoprotein genes have been targeted for removal using standard or loxP-Cre technologies to further elucidate their functions in development and health. Selenoproteins have also been shown to be involved in different human genetic disorders. Many new and novel features have been uncovered on the biosynthesis of selenocysteine, the amino acid that contains selenium, and its incorporation into protein as the 21^' amino acid in the genetic code. Further studies on the various components involved in the biosynthesis of selenocysteine and its insertion into protein have determined that much of this vast selenoprotein machinery exists in supramolecular complexes. Finally, several mouse models that were specifically generated for examining the role of selenium and selenoproteins in health and development have been devised. The rapid expansion and many new discoveries in the selenium field in the last five years are reflected by the addition of many new chapters and a much longer current edition.

xxii

Selenium: Its molecular biology and role in human health

The purpose of the new edition book is to inform the reader of these many new discoveries and to examine our present knowledge of the molecular biology of selenium, its incorporation into proteins as selenocysteine and the role that this element and selenium-containing proteins (selenoproteins) play in health. The book's emphasis is on our understanding of selenium metabolism in mammals and the role of this element in human health. The book begins with a brief history of selenium and how its face has changed through the years from one of a toxin and possible carcinogen to one of an essential micronutrient in the diets of humans and other animals. Indeed, selenium is now touted as an important cancer chemopreventative agent, as well as for its roles in inhibiting viral expression, delaying the progression of AIDS in HIV positive patients, preventing heart disease and other cardiovascular and muscle disorders, slowing the aging process, and having roles in development, male reproduction and immune function. As more of the molecular biology of selenium is unraveled, we are understanding the manner in which this element does indeed have direct roles in each of these health issues. The present book, like the first edition, is divided into three parts with ten more chapters than the earlier edition. The chapters in Part I, which is entitled "Biosynthesis of selenocysteine and its incorporation into protein," define selenocysteine as the 21^' naturally occurring amino acid in the genetic code and describe how this amino acid is incorporated into protein. Interestingly, the inclusion of selenocysteine to the genetic code as its 21'' amino acid marks the first addition to the code since it was deciphered in the mid-1960s. Our current understanding of how selenoprotein expression is regulated and the nucleocytoplasmic shuttling of the selenocysteine biosynthesis and insertion machinery in eukaryotes is also discussed in Part I. Part II is entitled "Selenium-containing proteins" and it discusses our current understanding of selenoproteins, primarily in higher eukaryotes. Part III is entitled "Selenium and human health" and it covers our current understanding of the role of selenium in various diseases, including cancer and heart disease, in HIV infection and AIDS, in male reproduction, and as an antiviral agent. The role of small molecular weight, selenium-containing compounds (selenocompounds) in human health and the dietary selenium requirements for humans are also discussed. In summary, this book provides an up-to-date review of much of the ongoing research in the selenium field. It provides a resource for scientists working in the selenium field, as well as for physicians, other scientists and students who wish to learn more about this fascinating micronutrient.

Dolph L. Hatfield, Maria J. Berry and Vadim N. Gladyshev

Acknowledgements The support and generous help of Bradley A, Carlson throughout the preparation of this book is gratefully acknowledged. The editors also wish to thank Sergey V. Novoselov for his help with the book cover.

Chapter 1. Selenium: A historical perspective James E. Oldfield Oregon State University, Corvallis, Oregon 97331, USA

Summary: The path followed in the biochemistry of selenium has taken some sharp turns during its development. At first, feared as a poisoner of livestock and later impugned as a carcinogen, selenium has about-faced and is now recognized as an essential micronutrient with anti-carcinogenic properties. While early studies on selenium have focused on the role of this trace element in animal physiology and studies with microorganisms, the field has matured to employ molecular biology to explain and employ the protective effects of selenium against a number of human maladies, including cancer and heart disease. The emphasis of this chapter is an examination of selenium's early history as a toxin, its later recognition as an essential micronutrient in the diet of mammals and its impact in the livestock industry that provided the foundations for the vast amount of the current basic and health research on this fascinating element. Even before it had been discovered and named, there were reports of conditions occurring in animals that, in retrospect, must have been caused by an excess of selenium. The Venetian explorer, Marco Polo, wrote of problems encountered by travelers in a mountainous region of what is now Shaan-Xi province in China [1]. He noted that when horses or other beasts of burden grazed on some indigenous plants, their hooves would split and fall off In the light of present knowledge, it would seem that these plants were "selenium accumulators" that concentrate selenium fi'om the soil to levels that are toxic to grazing animals. Then, several hundred years later, Madison [2] an army surgeon stationed at Fort Randall in the Nebraska territory, described a similar condition among dragoon remount horses that had been newly introduced to the area. K.W. Franke, who was a State Chemist at South Dakota State College, headed much of the definitive work on local toxic plants [3]. Actual proof of selenium's involvement in this toxicity problem came when workers in South Dakota identified it as the toxic principle in plants causing what was locally called "alkali disease" in cattle, on range lands of the north-central United States [4].

Selenium: Its molecular biology and role in human health The discovery of selenium, as an element, was made in 1817 by a Swedish chemist, Jons Jakob Berzelius, through what was, at that time, an elegant analytical process [5]. Berzelius was investigating the cause of illnesses among workers at a sulfuric acid manufacturing plant that occurred when copper pyrites from a local mine were used as the source of sulfur. He scraped a red deposit from the walls of the lead chambers in which the pyrites were processed, anticipating that it might contain tellurium, an element he had recently discovered. Tellurium was not present but he isolated another new element which he named selenium, after Selene, the Greek goddess of the moon. Taken together, these three early indications of selenium toxicity were certainly an inauspicious beginning for what was eventually to be recognized as an essential micronutrient. At that time, if anyone thought of selenium at all, and few did, it was as a toxic element. The earliest organized research effort with selenium, then, was directed toward means of avoiding, or coping with its toxicity. It was recognized that certain areas in the United States had seleniferous soils and this, together with the identification of selenium-accumulating plants, spelled trouble for animal agriculture operations. Ranchers learned to identify and remove accumulator plants, to dilute their livestock's forage feed with nonseleniferous materials, and to move their animals around in cycles which included some time on low-selenium forage grazing areas. Then, in 1957, research by a German scientist, Klaus Schwarz, working at the U.S. National Institutes of Health in Bethesda, changed forever the way selenium was assessed by both the scientific community and the general public. Schwarz had been working in Germany on studies of brewers' yeast as a protein source, during World War II, and he continued these studies when he came to America. He found, when he fed torula yeast, rather than brewers' yeast to rats, that they developed necrotic livers and he concluded that the brewers' yeast contained some essential nutrient that the torula yeast did not. He named the unknown substance "factor 3," since two other substances that alleviated liver necrosis had already been identified: vitamin E and (mistakenly) L-cysteine, which were known as factors 1 and 2. In 1957, Schwarz and Foltz armounced that they had fractionated factor 3 and found it to contain selenium [6]. Although its toxicity remained a real and difficult problem, it was now evident that, at lower dietary levels, selenium was harmless and, indeed, was quickly recognized as a dietary essential. The response to this discovery was immediate, and surprisingly extensive, as selenium deficiency was shown to be implicated in a number of animal diseases beyond the original liver necrosis. Studies in Oregon [7] showed that it was the cause of "white muscle disease," a myopathy that affected hundreds of calves and lambs each year in the central part of the state. Then, in quick succession, selenium deficiency was linked to other diseases of domestic animals and birds, including exudative diathesis and pancreatic

Selenium: A historical perspective degeneration in poultry, hepatosis dietetica in pigs and "ill-thrift" in cattle and sheep [8]. Of these, white muscle disease is the most widespread and has the greatest economic impact - involving not only calves and lambs but also deer, goats, horses, poultry and rabbits and occurring in all the major sheepproducing countries in the world [9]. Questions naturally arose about the biochemical function of selenium: how such small amounts of it could produce such profound biological reactions. These were answered, at least in part, by research carried out simultaneously in America and Germany. At the University of Wisconsin, Rotruck and associates [10] discovered selenium's presence in the enzyme, glutathione peroxidase, while Flohe in Germany showed the precise placement of selenium in the enzyme molecule [11]. It seemed that this enzymic involvement might be one way in which selenium could perform its beneficial metabolic functions and would explain how so little selenium could accomplish so much. Farmers and ranchers are often accused of being slow in accepting and applying research results relevant to their operations but this was certainly not the case with selenium supplementation. Its benefits were so dramatic that it soon became an accepted husbandry practice in areas of selenium deficiency, worldwide. Research, too, developed a number of methods by which selenium might be made available to animals, including feed fortification, injection, and with ruminant animals, an ingenious heavy pellet that would remain in the forestomach and gradually make selenium available for periods as long as a year. Selenium was also added to fertilizer mixes used on range and pasture land to improve the selenium status of forage plants grown thereon [12]. So, early biological research with selenium was stimulated by the animal industries, which in countries like New Zealand, were major contributors to the nation's economy. At first thought, the possibility of a selenium deficiency occurring among humans seemed remote on the grounds that the great diversity of the human diet would make an overall selenium deficiency unlikely. Cases of a human selenium deficiency did emerge, however, in some rural areas in China, where the people lived almost entirely on food substances produced on their ovra (selenium-deficient) land. This led to a cardiac myopathy that was first reported in Keshan county, of Heilongjiang province in northeastern China, and was called Keshan disease [13]. It is interesting to compare these symptoms with those of white muscle disease among animals - they have much in common. One of the fascinations of selenium research has been the abrupt changes in direction that have taken place over the years as knowledge of selenium's functions developed. So it was in the latter years of the 20th century when research interest in selenium switched from animals to molecular research and an emphasis on the role of selenium in human health.. This change was

Selenium: Its molecular biology and role in human health fueled by observations that, in addition to its now accepted nutrient function, selenium could also exert beneficial effects on human health at dietary levels somewhat higher than those required for its purely-nutritional activity. In Finland, governmental agencies became concerned about the long-term effects of low-selenium diets in their country on the health of the human population. They authorized the addition of selenium, as selenate, to fertilizers applied in the production of animal and human foods and have shown that this process effectively raises the selenium content of the Finnish diet to levels consistent with good nutrition and human health. They have carefully monitored the situation since Se-fertilization began, in 1984 [14] and we can be grateful to the Finnish scientists for providing much useful information on this type of application of selenium. The Finnish experience, too, has led to studies of the selenium status of other populations where dietary levels have been decreasing, over time [15]. The health-preserving activities of selenium, at about double the dietary levels recommended by the U.S. National Research Council, have been reviewed in detail by Combs [16]. It is interesting that these studies drew on the findings of Clark and associates at the Arizona Cancer Center and this gives rise to another of selenium research's "about faces," since early research had proposed that selenium was a carcinogen [17]. The application of selenium supplementation of livestock feeds to overcome selenium deficiency was prohibited for a time by the U.S. Food and Drug Administration (FDA) because of concerns raised by studies in their own laboratories suggesting that selenium might be a carcinogen [18]. This ruling exasperated American livestock producers who pointed out that they were being denied application of research that their tax dollars had helped pay for, while their strong competitors in New Zealand and Australia were routinely applying selenium in diets of their livestock. This conflict was resolved in this country, and in fact, the use of selenium in livestock feeds has been estimated to have saved this industry hundreds of millions of dollars in preventing muscle disorders and numerous other anomalies including enhancing reproduction as discussed in detail by Combs and Combs [19]. So, to recapitulate, the trail of research with selenium has been a tortuous one, marked by sudden and sometimes dramatic changes in direction. Its discovery, by Berzelius, was serendipitous; he was expecting to find tellurium in the Swedish sulfuric acid vats, but instead, he isolated selenium. There was a corollary to this in the much later studies of its health-protecting properties. When the Oregon workers sought a cure against white muscle disease, they thought it would be vitamin E, which proved ineffective, but selenium worked. Most of the early research, done in the first half of the last century, focused on means of avoiding selenium's toxicity, but Schwarz's carefully controlled studies with yeast opened the door for investigation of its

Selenium: A historical perspective beneficial effects as a micronutrient. Commercial application of supplementary selenium in diets of farm animals was delayed for several years due to fear that it might be a carcinogen but then, in one if its most dramatic about-faces, selenium proved to be anti-carcinogenic. Interestingly, Clark's study aimed against skin cancer where selenium that proved ineffective, but it was found to have significant benefits with other types of cancer, including those of the prostate, lung and intestine/colon (see [17] and references therein). It is understandable, certainly, because of its dreaded consequences in human health that cancer should have received the major attention by investigators of this new area of selenium's activity. It is exciting, however, that it has been shown to be a useful strategy against a number of other human diseases, and the Antioxidant Vitamins newsletter published by Hoffinan La Roche company listed 50 diseases against which selenium may play a protective role [20]. These include diseases of the heart, long recognized as major killers of the world's human populations [21,22] and ADDS, which has been called the "greatest catastrophe in human history" [23]. But most importantly, these earlier studies showing the importance of selenium in the diets of laboratory animals and livestock and the finding of selenium in protein as the amino acid selenocysteine in the 1970s have provided the foundations for the remarkable transformation that this field witnessed in the last 20 years. Indeed, the basic research described in this edition specifies selenium as a preventative agent in cancer, heart disease and other cardiovascular and muscle disorders, as an inhibitor of viral expression and as a factor delaying the aging process and the progression of AIDS in HIV positive patients. Furthermore, selenium is identified as,an essential element in mammalian development, male reproduction and immune function. These many health benefits now attributed to selenium highlight the serrated road fi-om a toxin to what may now be designated as a magic bullet. References 1.

2. 3. 4. 5. 6. 7. 8.

Polo, Marco. 1967. The Travels of Marco Polo Translated by EW Marsden and revised by T Wright pp 100-101 Everymans Library, London (Cited in C Reilly 1996 Selenium in Food and Health Chapman & Hall London p 3) TC Madison 1860 Statistical Report on the Sickness and Mortality in the Army of the United States RH Cooledge ed Ex Doc 52:37 KW Franke 1934 J Nutrition 8:597-608 AL Moxon 1937 Bull. 311, S. Dakota AgExp Sta 81 pp JJ Berzelius 1818 Serie 2 7:194 (Cited in C Reilly 1996 Selenium in Food and Health Chapman & Hall London p 2) K Schwarz, CM Foltz \951JAm Chem Soc 78:3292 OH Muth, JE Oldfield, LP Remmert, JR Schubert 1958 Science 128:1090 C Reilly 1996 Selenium in Food and Health Chapman & Hall ed London 338

Selenium: Its molecular biology and role in human health

9. E Wolf, V Kollonitsch, CH Kline 1963 Agr & Food Chem 11:355 10. JT Rotruck, AL Pope, HE Ganther, AB Swanson, DG Hafeman, WG Hoekstra 1973 Science 179:588 11. L Flohe, WA Gunzler, HH Schock 1973 FEES Letters 32:132 12. JE Oldfield 1997 Biomed & Environ 10:280 13. B Gu 1993 Chinese Med J 96:25\ 14. P Koivistoinen, K Huttunen 1986 Ann Clin Res 18:13 15. MP Rayman 2000 Lancet 356:233 16. GF Combs Jr 2001 Nutrition and Cancer 40:6 17. LC Clark, GF Combs Jr, BW Tumbull, EH Slate, D Alberts, D Abele, R Allison, J Bradshaw, D Chalker, J Chow, D Curtis, J Dalen, L Davis, R Deal, M Dellasega 1996 J Am Med Assoc 216:1957 18. AA Nelson, OG Fitzhugh, HO Calvery 1943 Cancer Res 3:230 19. GF Combs Jr, SB Combs 1986 The Role of Selenium in Nutrition Academic Press Inc New York 20. 1993 Antioxidant Vitamins Newsletter Hoffinan LaRoche Co New York 7:12 21. AFM Kardinaal, FJ Kok, L Kohlmeier, M Martin-Moreno, J Ringstad, J Gomez-Aracena, VP Mazaer, M Thamm, BC Martin, P Van'tVeer, JK Huttunea 1997 Am J Epidemiology 145:373 22. JT Salonen 1985 Trace Elements in Health and Disease H Bostrom, N Ljungstedt ed Almquist and Wiksell International Stockholm 172 23. HD Foster 2002 What Really Causes AIDS? Trafford Publishing Victoria, Canada 197

Parti

Biosynthesis of selenocysteine and its incorporation into protein

Chapter 2. Selenium metabolism in prokaryotes August Bock and Michael Rother Lehrstuhlfiir Mikrobiologie der Universitat Munchen, D-80638 Munich, Germany

Marc Leibundgut and Nenad Ban Institute of Molecular Biology and Biophysics, Swiss Federal Institute of Technology, ETH Honggerberg, HPK building, CH-8093 Zurich, Switzerland

Summary: The biosynthesis and specific incorporation of selenocysteine into protein requires the function of a UGA codon determining the position of selenocysteine insertion and a secondary/tertiary structure within the mRNA, designated the SECIS element, following the UGA at its 3'side in bacteria and located in the 3 'non-translated region in archaea. Biosynthesis of selenocysteine takes place on a unique tRNA species, tRNA^**", which is charged by seryl-tRNA synthetase and serves as an adaptor for the conversion of the seryl moiety into the selenocysteyl product by selenocysteine synthase. Monoselenophosphate, provided by selenophosphate synthetase, is the selenium donor. Selenocysteyl-tRNA^**^ is bound by the special translation factor SelB, which in bacteria via its Cterminal extension interacts with the apical part of the SECIS stem-loop structure. Crystallographic and NMR structural analyses of this extension from Moorella thermoacetica SelB, either free or complexed with the SECIS element, showed that it is made up of four winged helix domains from which only the C-terminal one interacts with the RNA ligand. Structure of the entire SelB molecule from Methanococcus maripaludis in the apo- and GDP/GTP bound forms revealed that it is a chimera between elongation factor Tu and initiation factors. Comparison of the structures in the GDP and GTP forms and modelling of the interactions between selenocysteyl-tRNA and SelB provided information on how SelB may discriminate tRNA^**^ from canonical tRNAs and may differentiate between the selenocysteyl moiety and the serylresidue of the precursor. A scenario for the major steps in the decoding process is postulated and arguments are given why the interaction of SelB with the mRNA is crucial. Reasons are also presented for the necessity of a balanced ratio of the components of the selenocysteine insertion apparatus and how it is regulated in E. coli via translational repression implicating a SECIS-like element located at the ultimate 5 'end ofselAB mRNA.

Selenium: Its molecular biology and role in human health

10

Introduction When bacteria are challenged with low molecular weight selenium compounds in the medium, they can process selenium in a nonspecific or a specific manner. The nonspecific metabolism rests on the chemical similarity between selenium and its neighbor element in the periodic table, sulfijr. When present above a critical concentration in Escherichia coli, i.e., at selenite concentrations higher than 1 ^M, selenium intrudes the sulfur pathways and is metabolized along the routes of sulfur metabolism [1,2] (Figure 1). Thus, selenium in the form of selenate is taken up by the sulfate transport system and reduced to selenide via the assimilatory sulfate reduction system. When offered as selenite, reduction appears to proceed chemically by interaction with thiol compounds like glutathione (see [3] for review).

Sulfate/Selenate

4

Sulfate/Selenate '

Selenite

V

Selenite . R-SH

•f

Sulfide/Selenide O-Ac-Ser

"V

Seryl-tRNAS^":

Cysteine/Se-Cysteine^ Pool mnm^s^U S/Se-Cystationine S/Se-Cys-tRNACv^

i I

ucu uco

S/Se-Methionine

S/Se-Met-tRNA'^^t

Se-Cysteyl-tRNASe

Scicnoprotcins — » I Selenylated Proteins AUG

Figure 1. Scheme for the specific and nonspecific metabolism and incorporation of selenium into macromolecules. The specific pathway is highlighted in bold. mnm's^U is the abbreviation for 5-methylamino-methyl-2-thiouridine and mnm'se^U for 5-methylaminomethyl-2-selenouridine. 0-Ac-Ser: 0-acetylserine, [Se] designates the reactive selenium species used by the selenophosphate synthetase as a substrate for the synthesis of selenophosphate; its possible metabolic origin is indicated by dashed arrows (see Chapter 4).

Selenium metabolism in prokaryotes

11

The first organic selenium compound formed is free selenocysteine, which can be converted to selenocystathionine and eventually to selenomethionine. On the other hand, selenocysteine has been shown to be a substrate for cysteyl-tRNA synthetase, which forms selenocysteyl-tRNA'^^^ and in this way incorporates selenocysteine at cysteine positions in proteins [4-6]. The decision whether selenium is incorporated nonspecifically as either selenocysteine or selenomethionine, therefore, should be dependent on the relative catalytic efficiencies of cysteyl-tRNA synthetase and cystathionine synthetase for the substrate cysteine and its analog selenocysteine. Nonspecific incorporation into macromolecules is drastically reduced when the cysteine biosynthetic pathway is interrupted by mutations or when it is fully repressed [6]. When selenomethionine is provided in the medium, it is almost indiscriminately incorporated into protein in place of methionine. This replacement is frequently used in x-ray analysis of protein crystals by multiwavelength anomalous dispersion [7] or in NMR spectroscopy [8]. Selenomethionine as the major selenium compound has also been detected when bacteria were grown on excessive amounts of selenite [9,10]. Free selenocysteine, on the other hand, is highly toxic and therefore growth inhibitory. Its incorporation in place of cysteine requires an overexpression system like the promoter-polymerase system of phage T7 to circumvent toxicity [11,12]. The specific incorporation of selenocysteine, on the other hand, is effective at much lower concentrations of selenite in the medium. With the aid of a fdhF-lacZ fusion reporter gene, in which readthrough into lacZ is dependent on the availability of selenium (see below), saturation has already occurred by 0.1 |iM selenite [13]. Specific incorporation does not involve free, low molecular weight selenocysteine since the biosynthesis of the molecule takes place from a precursor amino acid esterified with tRNA. It should be emphasized that the capacity to synthesize selenoproteins by the specific pathway is not ubiquitous. Actually, it is absent in the majority of microorganisms [14]. In this chapter we will discuss the specific incorporation of selenocysteine by bacteria, mainly E. coli and by members of archaea. Identification of the components involved in selenocysteine biosynthesis and specific insertion rests to a considerable degree on the early work of several groups studying the anaerobic formate metabolism oiE. coli [15-20]. Genes had been analyzed which, when mutated, abolished the ability of E. coli to synthesize active isoenzymes of formate dehydrogenase known as formate dehydrogenase N and formate dehydrogenase H which couple formate oxidation to the reduction of nifrate or protons, respectively. Thus, some mechanism must have been affected in the mutants that is required for generating activity of both enzymes. The genes had been mapped on the

12

Selenium: Its molecular biology and role in human health

chromosome oiE. coli and some of them (fdhAfdhB andfdhC) turned out to be involved in selenium metabolism [21]. Merits also go to two technical developments, namely the establishment of a plate overlay technique for screening large numbers of colonies for formate dehydrogenase activity [17] and the set-up of a procedure for specific incorporation of radioactive selenium into selenopolypeptides [22]. With the aid of these techniques, it was easy to differentiate between specific and nonspecific incorporation (see Figure 1). Specific incorporation of selenocysteine by bacteria The first genes discovered to contain an in-fi-ame UGA codon directing selenocysteine insertion were gpx, coding for glutathione peroxidase fi"om mouse [23], and fdhF fi-om E. coli, coding for the selenopolypeptide of formate dehydrogenase H [24]. Whereas an amino acid sequence was available for glutathione peroxidase showing colinearity between the UGA in the mRNA and selenocysteine in the protein, this was not the case for the bacterial enzyme. Evidence was obtained, however, by leading truncations from the 3'end into the gene and showing that removal of the segment containing the UGA also abolished selenium incorporation into the truncated gene product. Definite proof for the cotranslational insertion was then provided by fusion of the /acZ reporter gene upstream and downstream of the UGA in fdhF and the demonstration that readthrough of the UGA required the presence of selenium in the medium [13]. Analysis of mutations that affected readthrough led to the identification of the genetic elements involved in selenium metabolism in E. coli [21]. After the discovery that UGA also directs selenocysteine insertion into proteins in archaea [25], and with the results of the bioinformatic analysis of whole genome sequences from several hundred organisms, it has become an accepted notion that UGA is the universally conserved codon for selenocysteine [26]. tRNA^" The key element for specific selenocysteine insertion in E. coli was identified as the product of the fdhC gene, now designated as the selC gene [27]. It codes for a tRNA with unusual sequence and structural properties (Figure 2A). With 95 nucleotides, tRNA^'' is the largest tRNA in E. coli mainly because of an aminoacyl acceptor stem of eight possible base pairs and a 22 nucleotide long extra arm. There are also a number of deviations fi-om the consensus structure characteristic of canonical elongator tRNAs, namely a G at position 8, an A at position 14, a Y-R pair at the 10-25 sites and an R-Y base pair at positions 11-24. Moreover, the R-Y Levitt pair between the positions 15-48 is missing. As expected, extensive enzymatic and chemical

Selenium metabolism in prokaryotes

13

probing of the solution structure of tRNA^^*^ from E. colt, compared with that of canonical tRNA^^', showed that these deviations, plus the fact that the D stem is closed to a six base pair helix minimizing the D loop to four nucleotides, also restrict the types of tertiary interactions within the molecule [28]. Whereas the canonical G19-C56 interaction is still present, there are new interactions between CI6 of the D loop and C59 of the T loop and the canonical A21-(U8-A14) triple pair is substituted by a G8-(A21-U14) triple interaction. The extra arm is closed by a G45-A48 pair and connected to the anticodon coaxial helix only by interaction of A44 with U26. All these unusual sequence and structural properties are conserved in the sequences of other bacterial tRNA^^*^ species [29]. In view of the still open discussion on the structure of the eukaryal (see Chapter 3) and archaeal (Figure 2B) counterparts and of the lack of an x-ray structure, the conclusions can be concentrated on three characteristic features: (i) the acceptor-T stem stacked helix is extended to 13 base pairs made up of 8 plus 5 base pairs in bacteria and 9 plus 4 in archaea and eukarya, (ii) the closure of the D stem and the deviations from the sequence in canonical positions restrict the possibilities for tertiary interactions within the molecule, and (iii) the extra arm appears to be less well fixed to the body of the molecule than in classical elongator tRNAs. k

A C C i' G G—C 6—C A—U A—U

c

G

C^G

C G G^C G—C

c—e

m u C U GM C C

c u c uG c m • M i l 1 11 • G fiS «*"uV c !0a

ll|f G A <5 G C

C - G G / C0 "^*4r « • C—c ' G_0 30—0—c m " 'A—V 0 C A U i'A U P A

E. coli

A *'

e—c 10 ™\

J-

B

9

C C

w

A

0—C C—G 5. — G lO—CTb 5b^G—C--e?a 6j> G—C _.gg 10 G—C " « " A \ U C C C C 18, \ a i i i i 50- 1 1 1 1 C G G G e- r ^ ,^ 0 0 G G T C 1 11 1 G y A G CC C G G 7-u 1 ^ G—Cg' / c aa 1 G • U A ,/ C ji^ i^ G / C 21 C—G ^ (}/ ^.^ ,j G —c » i;,/ 30—-G—C « ^ *0 A—U fi A C A U « U |- A

%

M. jannaschii

Figure 2. Cloverleaf models of the Escherichia coli (A) and Methanococcus jannaschii (B) tRNA^^ species. The modified bases are shaded. The bacterial tRNA^""^ is drawn in the 8 + 5 arrangement of the acceptor stem/T-stem helices and the archaeal one in the 9 + 4 structure.

14

Selenium: Its molecular biology and role in human health

Biosynthesis of selenocysteine Formally, the biosynthesis of selenocysteine as elucidated for E. coli takes place in the following three steps: (a) L-Serine + ATP + tRNA^' -> L-Seryl-tRNA^"= + AMP + PPj (b) L-Seryl-tRNA^" + SS ^ Dehydroalanyl-tRNA^"= + SS + H2O (c) Dehydroalanyl-tRNA^'= + SePOj^' -> Selenocysteyl-tRNA^' + P04^" tRNA^'" is charged with L-serine by seryl-tRNA synthetase (equation a) [27] which is in accordance with the presence of the serine identity elements [28]. The overall catalytic efficiency of charging, however, is only about 1% of that measured for a cognate serine acceptor which reflects the limited requirement of serine carbon flux into the minor pathway [30]. Each of the structural properties differentiating tRNA^*'' from cognate serine acceptors could be responsible for the reduced acceptor activity. The conversion of seryl-tRNA^*" into selenocysteyl-tRNA^^'^ is catalysed by selenocysteine synthase (SS), which is the selA gene product. Selenocysteine synthase is a decameric protein made up of 50 kDa subunits which contain pyridoxal-5'phosphate as prosthetic group [31]. The amino group of serine forms an aldimine linkage with the carbonyl of pyridoxal phosphate and a water molecule is eliminated yielding dehydroalanyltRNA^*'' (equation b). Chemical proof for this intermediate was brought about via reduction by potassium borohydride, which yields alanyl-tRNA^'^ [32]. Nucleophilic addition of selenide then gives rise to selenocysteyltRNA^'^ (equation c). The source of selenide is monoselenophosphate [33]. Since elevated levels of selenide can substitute for monoselenophosphate in the reaction, it has been speculated why free selenide is not used as the natural subsfrate. A possibility considered is that the phosphate serves as a specificity "handle" to discriminate selenide from the highly similar sulfide. Indeed, substituting selenophosphate by thiophosphate gave rise to cysteyltRNA^^'' [34]. Moreover, selenium in the selenophosphate molecule is in an activated state, which also may contribute to the kinetics of selenide transfer to dehydroalanyl-tRNA^' [35]. Biochemical and high-resolution electron microscopic analysis demonsfrated that two subunits of the decameric enzyme bind one seryltRNA^" molecule [31,36]. The fully loaded enzyme thus contains five molecules of charged tRNAs bound to it. Binding seems uncooperative, only depending on the stoichiometry between the protein and the subsfrate [36]. It appears that once the tRNA is charged with serine it is immediately bound to selenocysteine synthase and stays in the activated state until selenophosphate is available as the substrate molecule. The cellular numbers of tRNA^^'' molecules (about 250) [37] and selenocysteine synthase decamers (about 150, which can bind five tRNA^"' molecules simultaneously) [31,32,36]

Selenium metabolism in prokaryotes

15

favor the assumption of the enzyme functioning as a sink for capturing the charged tRNA, which may be a good way to optimize the efficiency of utiHzation of the trace element. All bacterial species capable of selenoprotein synthesis and whose genomes have been sequenced possess orthologs of selA and selD genes whose products share high sequence similarities with selenocysteine synthase and selenophosphate synthetase from E. coli, respectively. This suggests that the mechanism for the biosynthesis of selenocysteyl-tRNA^*'' from seryl-tRNA^*'' is identical in all prokaryotes [34]. On the other hand, genomes from archaea with the ability to form selenoproteins do not contain an obvious candidate of the bacterial selA gene. Although the derived amino acid sequence of the selD gene from archaea is very similar to its bacterial counterpart, the actual mechanism for the biosynthesis of selenocysteine in archaea remains to be determined [38]. Translation factor SelB To participate in the decoding process, the 20 classical aminoacyl-tRNAs each have to enter a ternary complex with elongation factor Tu and GTP. When the affinity of EF-Tu to selenocysteyl-tRNA^'^'' was determined, it was found that it is about 200-fold lower than that of the standard aminoacyltRNAs [39]. Under competitive conditions, therefore, it cannot serve as a substrate for EF-Tu. This role is taken over by the specialized translation factor SelB [40]. SelB from E. coli (encoded by the selB gene, previously fdhA) is 69 kDa in size, and in its N-terminal part, it displays significant sequence similarity to EF-Tu (Figure 3). Utilising the sequence signatures of the bacterial SelB protein as a lead, a homolog has been identified amongst the gene products of the archaeon Methanococcus (M.) jannaschii [41]. The product of this open reading frame (aSelB) was purified, shown to be a GTPase with guanosine nucleotide binding properties like bSelB, and demonstrated to bind selenocysteyltRNA^'' from the same organism, preferentially but not exclusively to seryltRNA^^". aSelB contains a C-terminal extension of only 11 kDa compared to about 27 kDa of the bacterial SelB. Knowledge of the sequence of the archaeal SelB protein immediately prompted the identification of the eukaryal counterpart [42,43]. Recently, several independent crystallographic studies of bacterial and archaeal SelB proteins have provided first insights into the strucfral basis of selenocysteyl-tRNA^" recognition and into the mode of SelB interaction with the SECIS element and the ribosome. The crystal structure of SelB from the archaeon M. maripaludis reveals a molecule with overall dimensions of 1 1 0 A x 6 6 A x 3 9 A that consists of four distinct structural domains (domain

16

Selenium: Its molecular biology and role in human health

I G1 G2 G3

bSelB

II G4

III

IV

G5

,,mDm

aSelB A1 A2

A3

A4 A5

Figure 3. Domain structures of the bacterial SelB protein and its archaeal (aSelB) homolog, in comparison to the three structural domains of elongation factor Tu (EF-Tu). The G motifs involved in binding of the guanosine nucleotides are indicated by Gl to G4. Sequence segments absent in the SelB structures relative to that of EF-Tu are indicated by gaps. Note that domains IV of bSelB and aSelB are structurally unrelated (see figure 6).

I - rV), which adopt a "molecular chalice" arrangement [44] (Figure 4A). The first three domains form the cup of the chalice, whereas its base is formed by domain FV, which is linked to the cup via two long, anti-parallel P strands. The extended linker between the core of the factor formed by domains I-III and domain IV is flexible and allows domain IV to reach more than 50 A fi-om its attachment point (Figure 4B). Interestingly, the overall domain arrangement of archaeal SelB and the topology of its domain IV resemble the structure of IF2/eIF5B [45]. Furthermore, SelB has high sequence and structural homology to the a subunit of the archaeal IF2, which is the other factor that specifically recognizes a non-canonical tRNA (initiator tRNAi'^*') [44,46,47]. Nevertheless, it is not clear if functional parallels between the initiation of protein synthesis and selenocysteine incorporation may exist. Archaeal SelB adopts an EF-Tu:GTP-like overall domain arrangement in its apo-, GDP- and GppNHp-bound forms. Upon binding of a GTP analogue, small conformational changes are observed in the Switch 2 region in the GTPase domain that, based on the comparison with the EF-Tu:tRNA ternary complex, leads to the exposure of SelB residues involved in clamping the 5' phosphate of the tRNA. This mechanism may explain how under physiological conditions SelB binds selenocysteyl-tRNA^' only in its GTP state although there are no major conformational changes at the domain level between this and the GDP-bound states. The amino acid binding pocket of SelB is highly selective, in contrast to EF-Tu, and SelB has been observed to tightly bind selenocysteyl-tRNA^"' while discriminating against its precursor, seryl-tRNA^**^ [39,40]. The combined structural and biochemical work confirmed the importance of

Selenium metabolism in prokaryotes

25

factors. The rate constant of the release of GDP firom its complex with SelB was several orders of magnitude larger than that displayed by elongation factor EF-Tu, which explains why no guanosine nucleotide release factor is required for the function of SelB. On the other hand, the rate constant for the release of GTP is two orders of magnitude lower and in the same range as that measured for elongation factor EF-Tu. When the interaction of SelB with the 17 nucleotide minihelix of the fdhF SECIS element which carried a fluorescent group was assessed, an affinity of 1 nM was observed which was even increased when selenocysteyl-tRNA^^'' was present. Binding of the charged tRNA, therefore, maximizes the affinity of SelB for the RNA ligand; dissociation of the tRNA decreases the affinity, on the other hand, which leads to dissociation of SelB fi^om the mRNA, a necessary requirement for the translation of codons downstream of UGA. In conclusion, the following scenario can be visualized for decoding UGA as selenocysteine on the basis of the information presently available (i) A quaternary complex between SelB, selenocysteyl-tRNA^'', the SECIS element of the mRNA and GTP is formed. Its formation is non-random; binding of the charged tRNA stabilizes the complex of SelB with the SECIS element, (ii) During translation, the complex is translocated towards the ribosome, the lower helical part of the SECIS element is melted, and when the UGA arrives at the A site, SelB makes contact with the ribosome, which induces GTP hydrolysis, (iii) The charged tRNA is released in the proximity of the A site; its release decreases the affinity of SelB for the mRNA and facilitates the dissociation of the SelB-SECIS complex, (iv) After translation of the SECIS sequence, the RNA can refold and serve as a target for the formation of a new quaternary complex to assist the next oncoming ribosome in decoding UGA. The model can accommodate the results of the structure/function relation of the mutant SECIS variants, like the stringent requirement of a precise distance between the UGA and the loop region. Many issues, however, remain speculative. Examples are whether GTP hydrolysis precedes or succeeds the release of charged tRNA, how domain IV communicates with the EF-Tu like part of the SelB molecule and what effect domain IV exerts on the translation process when it contacts the ribosome at the mRNA entrance cleft.

26

Selenium: Its molecular biology and role in human health

Involvement of the E. coli SECIS element in the control of gene expression In E. coli, the genes for the components of the selenocysteine insertion machinery are organised in three transcriptional units on the chromosome, namely selAB, selC and selD. Transcription of these units is constitutive, thus independent of the physiological condition [79]. At the translational level, however, the expression of the selAB unit is subject to regulation. The crucial element involved is a SECIS-like structure located in the immediate 5'-end of the non-translated region, not overlapping with the ribosomal binding site [80]. The SECIS-like structure is the target for SelB binding, whereby the affinity is about 6-fold lower compared to binding by the inframe fdhF SECIS element. It also forms a quaternary complex with GTP, selenocysteyl-tRNA^*^ and SelB, which results in repression of the selAB mRNA translation [80]. Since quaternary complex formation requires the presence of tRNA^^'' charged with selenocysteine, translational repression affords the availability of an adequate supply of selenium in the medium.

-

D °

vibF -;

selA

selB

G ^ G C U C G U A C G U A C G C —G C —G „G—C

I A —U^ A^U RBS U • G 5". U — A U C A G C C A G G U U U C C U A*U G

Figure 8. Organisation of the selAB operon (A) and structure of the SECIS-like element (B). RBS: Ribosmal binding site; the ignition codon of the selA gene is indicated by an asterisk.

What is the rational for the necessity to control selAB mRNA translation? A possible reason may exist in the requirement for a balanced ratio of the components of selenocysteine insertion machinery. Under wild-type and balanced growth conditions the supply of tRNA^^'' is limiting as judged by

Selenium metabolism in prokaryotes

27

the effect of its overexpression, which results in an about 2-fold stimulation of UGA readthrough [81]. On the other hand, overproduction of SelB is detrimental to UGA readthrough because the statistics for formation of the complex with both the SECIS and selenocysteyl-tRNA^*'' ligands is disturbed [81]. Translational repression of selAB expression thus adjusts the level of SelB to the amount required. Thus, under low mRNA levels, selenocysteine synthase and SelB formation are reduced but the readthrough of the UGA is still sufficient due to the higher affinity of its SECIS structure for SelB compared to that of the SECIS-like element. At high selenoprotein mRNA levels, such as under fermentative conditions, translational repression is relieved thus supplying the required increased amount of SelB. Acknowledgements The financial support of this work by the Deutsche Forschungsgemeinschaft and the Fonds der Chemischen Industrie to AB and of the Swiss National Science Foundation (SNSF), the NCCR Structural Biology program of the SNSF, the ETH internal research grant TH-1/01-2, and a Young Investigator grant from the Human Frontier Science Program to NB is gratefully acknowledged. References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28.

DO Cowie, GN Cohen 1957 Biochim Biophy Acta 26:252 T Tuve, HH Williams 1961 J Biol Chem 236:597 J Heider, A B6ck \99l Advances Microbial Physiol 7,5:1 \ JL Hoffmann, KP McConell, DR Carpenter 1970 Biochim Biophys .4cto 199:531 PA Young, II Kaiser 1975 Arch Biochem Biophys 171:483 S Muller, J Heider, A. B6ck 1997 Arch Microbiol 168:421 WA Hendrickson, JR Horton, DM LeMaster 1990 EMBO y 9:1665 GP Mullen, RP Dunlap, JD Odon 1986 Biochemistry 25:5625 RE Huber, RS Criddle 1967 Biochim Biophys Acta 141:587 MGN Hartmannis, TC Stadtman 1982 Proc Natl Acad Sci USA 79:4912 S Muller, H Senn, B Gsell, W Vetter, C Baron, A B6ck 1994 Biochemistry 33:3404 M-P Strub et al 2003 Structure 11: 1359 F Zinoni, A Birkmann, W Leinfelder, A B6ck 1987 Proc Natl Acad Sci USA 84:3156 GV Kryukov, VN Gladyshev 2004 EMBO Reports 5:538 A Graham et al 1980 FEMSMicrobiol Lett 7:145 B A Haddock, M-A Mandrand-Berthelot 1982 Biochem Soc Trans 10:478 M-A Mandrand-Berthelot, MYK Wee, B A Haddock 1978 FEMS Microbiol Lett 4:37 YA Begg, JN Whyte, BA Haddock 1977 FEMS Microbiol Lett 2:47 M Chippaux, F Casse, C-C Pascal 1972 JBacteriol 110:766 EL Barrett, CE Jackson, HT Fukumoto, GW Chang 1979 Mol Gen Genet 177:95 W Leinfelder et al 1988 JBacteriol 170:540 JC Cox, ES Edwards, JA DeMoss 1981 JBacteriol 145:1317 I Chambers et al 1986 EMBO J 5:\22\ F Zinoni, A Birkmann, TC Stadtman, A B6ck 1986 Proc Natl Acad Sci USA 83:4650 R Wilting, S Schoriing, BC Persson, A BOck 1997 J Mol Biol 266:637 A BOck 2005 Selenocysteine in Encyclopedia of Life Sciences John Wiley & Sons Chichester http://www.els.net W Leinfelder, E Zehelein, M-A Mandrand-Berthelot, A Bock 1988 Nature 331:723 C Baron, W Westhof, A Bock, R Giege 1993 J Mol Biol 231:274

28 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81.

Selenium: Its molecular biology and role in human health P Tommy, R Wilting, J Heider, A B6ck 1994 JBacteriol 176:1268 C Baron, J Heider, A B6ck 1990 Nud Acids Res 18:6761 K Forchhammer et al 1991 J Biol Chem 266:6318 K Forchhammer, A B6ck 1991 J Biol Chem 266:6324 Z Veres et al 1992 Proc Natl Acad Sci USA 89:2975 P Tormay et al 1998 EurJBiochem 254:655 RS Glass, TC Stadtman 1995 Methods Enzymol 252:309 H Engelhardt et al 1992 Mol Microbiol 6:3461 H Dong, L Nilsson, CG Kurland 1996 J Mol Biol 260:649 M Rother, A Resch, R. Wilting & A Bock 2001 BioFactors 14:75 C FSrster, G Ott, K Forchhammer, M Sprinzl 1990 Nucl Acids Res 18:487 K Forchhammer, W Leinfelder, A B6ck 1989 Nature 342:453 M Rother, R Wilting, S Commans, A B6ck 2000 J Mol Biol 299:351 RM Tujebajeva et al 2000 EMBO Reports 11:158 D Fagegaltier et al 2000 EMBO J17:4796 M Leibundgut, C Frick, M Thanbichler, A BOck, N Ban 2005 EMBO J 24:11 A Roll-Mecak, C Cao, TE Dever, SK Burley 2000 Cell 103:781 E Schmitt, S Blanquet, Y Mechulam 2002 EMBO 7 21:1821 CB Foster 2005 Mol Biol Evol 22: 383 C Baron, J Heider, A B8ck 1993 Proc Natl Acad Sci USA 90:4181 MJ Berry et al 1991 Nature 353:273 MJ Berry, L Banu, JW Harney, PR Larsen 1993 EMBO J 12:3315 GE Garcia, TC Stadtman 1992 J Bacteriol 174:7080 T. Gursinsky, J Jager, JR Andreesen, B Sohling 2000 Arch Microbiol 174:200 M Rother, I Matthes, F Lottspeich & A Bock 2003 J Bacteriol 185:107 M Kromayer, R Wilting, P Tormay, A B6ck \996 J Mol Biol 262:412 D Fourmy, E Guittet, S Yoshizawa 2002 J Mol Biol 324:137 S Yoshizawa et al 2005 Nat Struct Mol Biol 12:198 MSelmer,XDSu2002£A/5Oy21:4145 KS Gajiwala, SK Burley 2000 Curr Opin Struct Biol, 10:110 A Huttenhofer, A B6ck 1998 Biochemistry 37:885 M Thanbichler, A B6ck, RS Goody 2000 J Biol Chem 275:20458 J Heider, C Baron, A B6ck 1992 EMBO J11:3759 Z Liu, M Reches, I Groisman, H Engelberg-Kulka 1998 Nucl Acids Res 26:896 A Huttenhofer, E Westhof, A Bock 1996 RNA 2:354 C Allmang, P Carbon, A Krol 2002 RNA 8:1308 AM Zavacki et al 2003 Mol Cell 11:773 S Meunieretal 2000 £MSOy 19:1918 M Valle et al 2003 Nat Struct Biol 10:899 MJ Berry, AL Maia, JD Kieffer, JW Hamey, PR Larsen 1992 Endocrinology 131:1848 H Kollmus, L Floh6, JEG McCarthy 1996 Nucl Acids Res 24:1195 S Suppmann, BC Persson, A B6ck 1999 EMBO J18:2284 ESJ Amer, H Sarioglu, F Lottspeich, A Holmgren, A BOck 1999 J Mol Biol 292:1003 O Rengby et al 2004 Appl Environ Microbiol 51:59 Z Liu, M Reches, H Engelberg-Kulka \999 J Mol Biol 294:1073 JB Mansell, D Gueremont, ES Poole, WP Tate 2001 EMBO J 20:7234 ES Poole, CM Brown, WP Tate 1995 EMBO 714:151 S Mottagui-Tabar, A BjSmsson, L Isaksson EMBO J 13:249 J Parker 1989 Microbiol Rev 53:273 SJ Klug et al 1997 Proc Natl Acad Sci USA 94:6676 G Sawers, J Heider, E Zehelein, A Bock 1991 7 Bacteriol 172:4983 M Thanbichler, A B6ck 2002 EMBO y 21:6825 P Tormay, G Sawers, A Bfick 1996 Mol Microbiol 21:1253

Chapter 3. Mammalian selenocysteine tRNAs

and

other

eukaryotic

Bradley A. Carlson, Xue-Ming Xu, Rajeev Shrimali, Aniruddha Sengupta, Min-Hyuk Yoo, Robert Irons', Nianxin Zhong and Dolph L. Hatfield Molecular Biology of Selenium Section, Laboratory of Cancer Prevention, Center for Cancer Research, National Cancer Institute, National Institutes of Health, Bethesda, MD 20892, USA

Byeong Jae Lee Laboratory of Molecular Genetics, Institute of Molecular Biology and Genetics, School of Biological Sciences, Seoul National University, Seoul 151-742, Korea

Alexey V. Lobanov and Vadim N. Gladyshev Department of Biochemistry, University of Nebraska, Lincoln, NE 68688, USA

Summary: Selenocysteine (Sec) tRNA occupies a prominent position in the expression of selenoproteins as it is essential for their synthesis and it provides the means by which selenium is co-translationally inserted into protein as the amino acid. Sec. Thus, Sec tRNA is regarded as the principle constituent in selenoprotein synthesis. Many features unique to this tRNA have been characterized over the years in mammals and other eukaryotes. In the last five years, the major advances have been in an elucidation of the different roles that the two major Sec tRNA isoforms play in selenoprotein biosynthesis and in Sec biosynthesis. One isoform appears to be responsible for the synthesis of selenoproteins that have roles in housekeeping functions and are less dependent on selenium status for their expression. The second isoform, that differs by only a single methyl group at the 2'-0-hydroxylribosyl moiety at position 34 (designated Um34), appears to be responsible for the expression of selenoproteins that have roles in stress-related phenomena and are highly dependent on selenium for their expression. Several new observations regarding Sec biosynthesis, which occurs on its tRNA, have also been recently made. Other recent advances involving Sec tRNA have used this molecule as a tool for determining whether eukaryotes outside the animal kingdom contain the machinery dedicated for the insertion of Sec into protein. These recent findings are discussed in this chapter. 'RI is also affiliated with the Nutritional Science Research Group, Division of Cancer Prevention, National Cancer Institute, National Institutes of Health, Bethesda, MD 20892.

30

Selenium: Its molecular biology and role in human health

Introduction Of all the tRNAs that have been identified and whose functions have been characterized, selenocysteine (Sec) tRNA is arguably the most fascinating. For example, it is responsible for the translation of an entire class of proteins, the selenoproteins. Unlike the other 20 amino acids in protein. Sec is biosynthesized on its tRNA. Furthermore, while the other 20 aminoacyltRNAs share a common elongation factor, eEF-2, Sec-tRNA has its own dedicated elongation factor, EFsec [1,2]. In addition, there are many other novel features in its primary and secondary structures, as well as in its transcriptional properties, which are unique only to this tRNA. As these unique characteristics have been reviewed in detail recently [3], they will not be further discussed herein. Of importance to our present discussion on Sec tRNA in mammals is that it exists as two isoforms that differ from each other by a single methyl group, 2'-0-methylribose, which is present at position 34, the wobble position of the anticodon. This methyl group is designated Um34 and its synthesis is a highly specialized event in the maturation of Sec tRNA (see below). Sec tRNA has been detected in prokaryotes, eukaryotes and archaea, and therefore, is widespread in nature. However, while Sec tRNA has been characterized extensively in mammals and other animals [3], little is known about the properties, functions and distribution of this tRNA in other eukaryotic organisms. In fact. Sec tRNA, selenoproteins and/or the Sec protein insertion machinery have been reported thus far in only a handful of eukaryotes other than animals. The new data on the roles of the two mammalian isoforms in selenoprotein biosynthesis, Sec synthesis on its tRNA in mammalian cells and the occurrence of Sec tRNA in eukaryotes outside the animal kingdom are subjects of this review. For further information, the reader is referred to other, more extensive reviews on this subject [3,4]. Uin34, a highly specialized event in mammalian Sec tRNA'^*'"'^*'' Sec tRNA is designated as Sec tRNA'^"^^^*^, since it is first aminoacylated with serine and the biosynthesis of Sec occurs on the tRNA [3]. As noted above, there are two isoforms of Sec tRNA'^'^'^*'^ that differ fi-om each other by the presence of Um34. Both isoforms have a highly modified base at position 34, 5-methoxycarbonylmethyluridine (mcm^U), and they are thus designated mcm^U and mcm^Um. The synthesis of Um34, which is the last step in the maturation of Sec tRNA^^' , has been described as a highly specialized event [3] since its addition 1) is dependent on the primary and tertiary structure of Sec tRNA^^^'''^*', including the prior synthesis of all modified bases (the synthesis of other modified bases is not nearly as stringently dependent on these factors) [5], 2) is dependent upon selenium status [4], 3) alters secondary and tertiary structure [6], and 4) is responsible

Mammalian selenocysteine tRNA

31

for the synthesis of a subset of selenoproteins impHcated in stress response [7,8]. The stress-related selenoproteins include glutathione peroxidase-1 and 3, SelR, SelT and likely SelW, and the dependence of this subclass of selenoproteins on mcm^Um is discussed in greater detail in Chapter 29 (see also [7,8]). Interestingly, the expressions of stress-related selenoproteins, like the occurrence of mcm^Um, are dependent on selenium status. The lack of expression of this subclass of selenoproteins is a translational event due to the lack of Um34 on mcm^Um as the amount of mRNA encoding some of these selenoproteins is more than adequate for their expression following the reduction of mcm^Um due to selenium deficiency. However, the level of mRNA encoding some of these selenoproteins, e.g., GPxl and SelW, is substantially reduced during selenium deficiency. This dependence of selenoprotein mRNA expression on selenium status was proposed to result from the mechanism of nonsense mediated decay (NMD), wherein the UGA Sec codon in selenoprotein genes is recognized as nonsense [9,10]. Different selenoprotein mRNAs manifest varying degrees of sensitivity to NMD and the reason for this is not clearly understood. In this regard, the sensitivity of mcm^Um to selenium status is of particular interest since the overall levels of the Sec tRNA^^*'^'^'' population decrease significantly under conditions of selenium deficiency, while the amounts of mcm^U remain largely unchanged and may even be enriched (described in detail in [3]). A possible sequence of events during selenium deficiency is that Um34 synthesis is reduced which in turn prevents the expression of several stress-related selenoproteins and the corresponding mRNAs manifest varying degrees of sensitivity to NMD. The dependence of a subclass of proteins, and in this case, stress-related selenoproteins, on a methyl group for recoding a nonsense codon (UGA) to generate their expression is, to our knowledge, an unprecedented observation in translation. To elucidate the means by which Um34 is responsible for the expression of a subset of selenoproteins, we examined various parameters of selenoprotein mRNAs such as nucleotide context of the UGA codon, position of the UGA codon in the open reading floras and the class of the Sec insertion sequence element (located in the 3' untranslated region of selenoprotein mRNAs and required for the incorporation of Sec into protein [11]), but none of these features revealed any clear pattern that would explain this phenomenon (reviewed in [3]). The precise mechanism of how UGA is recoded by mcm'Um in stress-related selenoprotein mRNAs resulting in their expression is currently under investigation. Uin34 methylase Although the methylase that synthesizes Um34 on Sec tRNA'^''^" has not been identified, a protein described by Ding and Grabowski, designated SECp43 [12], may have a role in Um34 synthesis [13]. These investigators

32

Selenium: Its molecular biology and role in human health

reported that SECp43 existed in a complex with a 48 kDa protein and Sec tRNA'^"'^'' in HeLa cell extracts. A 48 kDa protein, designated soluble liver antigen (SLA), had been earlier characterized as it co-precipitated with Sec tRNA'^'^'^^'^ after being targeted by antibodies in patients with an autoimmune chronic hepatitis [14]. We recently found that the 48 kDa protein which formed a complex with SECp43 and Sec tRNA'^"^^^'' is indeed SLA [13,15]. Furthermore, we found that knockdown of SECp43 in mammalian cells using RNAi technology reduced the formation of Um34 [13] suggesting that it is involved in Um34 synthesis. Further studies are required to determine if SECp43 is associated with methylase function and how it is precisely involved in Um34 synthesis. However, we investigated additional parameters of SECp43 and SLA [13]. The targeted removal of either SECp43 or SLA affected the binding of the other to Sec tRNA'^'^'^^'", even though tRNA^^'^'^^'^ attachment was most affected when SLA was the targeted member of the complex. In addition, we observed that SECp43 is located primarily in the nucleus, while SLA occurs in the cytoplasm. Cotransfection of both proteins resulted in the nuclear translocation of SLA. This result suggested that SECp43 may also serve as a chaperone for shuttling SLA and Sec tRNA'^''^'^ between different cellular compartments. SLA likely has a role in the biosynthesis of Sec as discussed below. Sec Biosynthesis Sec is distinctive from the other 20 amino acids in protein in that Sec biosynthesis occurs on its tRNA [3,5]. Although asparagine, glutamine and cysteine biosynthesis can occur on tRNA^™, tRNA°'" [16] and tRNA^>" [17], respectively, this means of synthesizing these amino acids appears to be restricted to only a few life forms. However, Sec apparently is synthesized on its tRNA in all organisms that encode the Sec-protein insertion machinery which, as noted above, is widespread in nature. Serine is attached to Sec tRNA'^''^^" by seryl-tRNA synthetase, and thus, the identity elements for recognizing this tRNA are for serine and not Sec. The identity elements in Sec tRNAt^"''^" are located in the discriminator base and the long extra arm which are essential to its aminoacylation, even though the acceptor, Tvj/C and D stems also have roles in the identity process (reviewed in [3,5]). Following the aminoacylation of Sec tRNA^^^"^^^*^, the serine moiety serves as the backbone for the synthesis of Sec on its tRNA. The enzyme that carries out the synthesis of Sec utilizing the serine moiety on seryl-tRNA'^''^*' is designated Sec synthase and its identity and mechanism of action has been thoroughly characterized in bacteria (Chapter 2). However, in mammals, a homolog of the bacterial enzyme is absent and the overall biosynthesis of Sec on its tRNA has not been established. In 1970, a kinase activity that phosphorylated a minor species of seryl-tRNA to form phosphoseryl-tRNA was observed in rooster liver [18] and a minor

Mammalian selenocysteine tRNA

33

seryl-tRNA that decoded UGA was reported in bovine, rabbit and chicken livers [19]. The phosphoseryl-tRNA and the seryl-tRNA that decoded UGA were subsequently found to be Sec tRNA'^"'^^'' [20], but the kinase activity remained elusive until only recently. A gene for mammalian phosphoseryltRNAf^^'^''= kinase (pstk) was identified using a comparative genomics approach by searching completely sequenced archaeal genomes for a kinaselike protein with the pattern of occurrence similar to that of known components of the Sec insertion machinery [21]. A gene corresponding to a potential mouse pstk was cloned, the gene product (PSTK) expressed and characterized. PSTK specifically phosphorylated the seryl moiety on seryltRNA'^"'^' confirming that indeed the seryl-tRNA'^''^"" kinase had been correctly identified. A search of proteins with homology to mammalian PSTK revealed homologs in Drosophila, Caenorhabditis elegans, Methanopyrus kandleri and Methanococcus jannaschii. These observations suggest that the function of PSTK has been conserved in archaea and eukaryotes that synthesize selenoproteins, but this function is absent in bacteria and eukaryotes, e.g., plants and yeast, that do not synthesize selenoproteins. Since PSTK has been highly conserved in evolution, it must have an important role in selenoprotein biosynthesis and/or its regulation [21,22]. A similar reaction as that which may occur in Sec biosynthesis in mammals has been observed in archaea wherein cysteine biosynthesis on tRNA^^* takes place by the initial aminoacylation of the tRNA with phosphoserine and subsequently converting the phosphoserine moiety to cysteine by a pyridoxal phosphate-containing Cys synthase [17]. This pathway serves as an excellent model providing further evidence that phosphoseryl-tRNA'^'^'^'^^'^ is indeed an intermediate in the biosynthesis of Sec (see also [22]). A mammalian Sec synthase distinct from that described in bacteria (Chapter 2) could then act upon phosphoseryl-tRNA'^'^'^'' in removing the phosphate group and accepting the active selenium donor to make selenocysteyl-tRNA'^"'^*^. Interestingly, a candidate Sec synthase is SLA (see above for further characterization of SLA). This protein was identified several years ago [14], and has more recently been characterized as an important member of the Sec biosynthesis and protein insertion machinery [13,15] (see also Chapter 8). The active form of selenium that is donated to the intermediate in the biosynthesis of Sec has been characterized in bacteria as monoselenophosphate, which is synthesized from selenide and ATP by selenophosphate synthetase [23]. Although the active selenium donor has not been characterized in eukaryotes, two selenophosphate synthetase genes, designated Spsl and Sps2, have been detected in mammals [24-26]. SPS2, the gene product of Sps2, is a selenoprotein which suggests that it may be involved in the autoregulation of its own biosynthesis [24]. When the

34

Selenium: Its molecular biology and role in human health

activated form of selenium is donated to the intermediate, which might be a product of phosphoseryl-tRNA^^"^^" catalyzed by SLA, the biosynthesis of Sec on tRNA^^"^^^^" would be complete and the Sec moiety then poised for insertion into the nascent selenopeptide (see Chapter 8). If this hypothesis is correct, it may be anticipated that the mechanism of Sec is synthesized on the tRNA will be resolved in the near future. Occurrence and evolution of Sec tRNA'^"'^*' and its insertion machinery In 1989, the synthesis of Sec on its tRNA was reported simultaneously in bacteria [27] and mammalian cells [28]. These observations unequivocally demonstrated that Sec was the 21*' amino acid in the genetic code. Even though earlier studies had suggested that Sec was the 21*' amino acid [2931], when it was recognized that Sec biosynthesis occurred on its tRNA, the possibility that an intermediate, e.g., phosphoserine [20], might first be incorporated as the 21*' amino acid and then changed posttranslationally to Sec had to be ruled out (see discussion in [28]). Evidence was subsequently provided that the Sec tRNA'^'^^^^'' gene (trsp) was universal in the animal kingdom [32] and trsp was found in numerous species of prokaryotes [33]. The occurrence of selenoproteins and the Sec protein insertion machinery was subsequently reported in archaea (see [34] and references therein), but only recently were this machinery [35] and Sec ^j^^[Ser]Sec ^ 5.33J obscrvcd in eukaryotes outside the animal kingdom. Initially, the Sec incorporation machinery and at least 10 selenoproteins [35] were found in Chlamydomonas, a green alga and a member of the plant kingdom. Since Sec tRNA'^''^'' is a convenient marker for identifying the presence of the Sec protein insertion machinery in an organism, attention was focused on a means of demonstrating the occurrence of this nucleic acid in different eukaryotes. Initially, we devised a relatively simple procedure for partially purifying and then sequencing Sec tRNA^^"'^*'' by reverse transcription (RT)PCR and sequenced the Sec tRNA from Chlamydomonas [36]. A computational Sec tRNA genomic analysis program for detecting trsp in partially or completely sequenced genomes was also developed (A.V. Lobanov, G. V. Krjmkov and V.N. Gladyshev, unpublished data). The latter procedure in combination with isolating and sequencing the gene product by RT-PCR provided an easy and direct method for assessing whether organisms encode the machinery for incorporating Sec into protein. Sec tRNA'^''^ was identified in two model organisms, Dictyostelium discoideum and Tetrahymena thermophila [37]. The sole termination codon is T. thermophila is UGA (see [37] and references therein) and the demonstration that this organism also encodes Sec tRNA suggests that its only stop codon has a shared function. It is therefore important to note that T. thermophila does indeed utilize its Sec tRNA for making selenoproteins as

Mammalian selenocysteine tRNA

35

its genome encodes selenoprotein genes (A.V. Lobanov and V.N. Gladyshev, unpublished data), trsp was also recently identified in various species of Plasmodium [38,39] and in Toxoplasma gondii [38]. Sec tRNAs are the longest tRNAs known in either eukaryotes, archaea or prokaryotes. Prokaryotes are the longest found tRNAs thus far with some exceeding 100 bases in length [34]. All known animal Sec tRNAs^^"^^^'' are 90 bases in length as are those from Chlamydomonas [36] and Tetrahymena [37]. trsp in Toxoplasma gondii is 87 bases long [38] which indicates its gene product is 90 bases in length as the CCA terminus is added posttranslationally in all tRNAs. Dictyostelium Sec tRNA is 91 nucleotides [37] and those from various species of Plasmodium are 93 nucleotides [38,39]. The extra base in Dictyostelium Sec tRNA^^"^^'' occurs in the Dstem, while the extra three bases in Plasmodium occur in the long extra arm. The clover leaf model of mammalian Sec tRNA^^"'^'" and that of Dictyostelium are shown in Figure 1.

B

\ 3' A C C 5' G G—C C—G C —G" C—G G—C G-U "A—U"" U-U „ G—0 U U .A A CACC"""* G^GAcryac "GUGG C ^ I I I I I I k .„ ^ ^^ ^ U S^GGGG u - A c ? A - ^-^ A-U "G-C" G-C C A U i'A mcm'U g

„G • I ^ "• U ' . I'^'u <"

A

Bovine liver Sec tRNA'^^''^""

3' A C C 5' G G—C C-G Q—C" U—G G-C U-A "U —G"U-U G—C UU A A CAAC"" * G^CUACGC "GUUG,, C ^ l l l l l l A , „ U U « G^AUSCG U_A. ? A « U-A G-C "G-C" A—U C A U A U C

"c' I "• G ! *

A . u '"

*

Dictyostelium discoideum Sec tRNA'^"^'^"'=

Figure 1. Cloverleaf model of mammalian (A) and Dictyostelium (B) Sec tRNA'^"'^^". The secondary structure of the tRNA is shown in a 9/4 based paired form (i.e., 9 base pairs in the acceptor stem and 4 base pairs in the T-stem).

36

Selenium: Its molecular biology and role in human health

It is amazing that so much energy has been invested in developing a highly sophisticated system during evolution for the insertion of single amino acid, Sec, into protein. Most of the known functions of Sec in protein are involved in redox reactions and thus the driving force for evolving the Sec protein incorporation machinery most likely was to utilize the unique chemical properties of the selenium atom in Sec. Under reducing environment of the cell. Sec is ionized as its pKa (~5.5) is below physiological pH and it is a better participant in certain redox reactions than Cys. The pKa of Cys is 8.3, therefore, many cysteines are protonated under physiological conditions. The dependence of different organisms on selenoproteins for survival varies widely within the three life kingdoms, eubacteria, archaea and eukaryotes. Some organisms, such as yeasts and higher plants, do not utilize selenoproteins at all, while others, such as higher vertebrates, are dependent on selenoproteins for their survival (see [3,34] and references therein). Other organisms, such as many prokaryotes, encode the Sec protein insertion machinery, but its occurrence often is not essential to their survival, although there must be some selective advantage for them to maintain this machinery. These observations raise important unresolved questions in selenium biology. For example, why, once such a sophisticated system evolved for utilizing the unique properties of selenium, don't all organisms take advantage of selenoproteins in cellular metabolism, or alternatively, why have some organisms lost the ability to make selenoproteins? Although there is likely no single solution to this question, a better understanding of those organisms that do and do not make selenoproteins will most certainly shed light on this enigma. As Sec tRNAs provide a readily identifiable marker for the presence of the Sec protein insertion machinery, and the computational and sequencing techniques described above are so readily applicable to identifying Sec tRNAs, focusing on the occurrence of trsp and/or Sec tRNA will provide a means of assessing how widespread the use of Sec is in nature. Acknowledgements This work was supported by the Intramural Research Program of the National Institutes of Health, National Cancer Institute, Center for Cancer Research. References 1. 2. 3. 4. 5.

RM Tujebajeva, PR Copeland, XM Xu, BA Carlson, JW Harney, DM DriscoU, DL Hatfield, MJ Berry 2000 EMBO /Jep 1:158 D Fagegaltier, N Hubert, K Yamada, T Mizutani, P Carbon, A Krol 2000 EMBO J 19:4796 DL Hatfield, BA Carlson, XM Xu, H Mix, VN Gladyshev 2006 Prog Nucl Acids Res A/o/5/o/(In Press) DL Hatfield, VN Gladyshev 2002 Mol Cell Biol 22:3565 LK Kim, T Matsufuji, S Matsufuji, BA Carlson, SS Kim, DL Hatfield, BJ Lee 2000 UNA 6:1306

Mammalian selenocysteine tRNA 6. 7. 8.

9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39.

37

AM Diamond, IS Choi, PF Grain, T Hashizume, SC Pomerantz, R Cruz, CJ Steer, KE Hill, RF Burk, JA McCloskey, DL Hatfield 1993 J Biol Chem 268:14215 BA Carlson, XM Xu, VN Gladyshev, DL Hatfield 2005 J Biol Chem 280:5542 BA Carlson, XM Xu, VN Gladyshev, DL Hatfield 2005 In Grosjean (Ed.): Fine-Tuning ofRNA Functions by Modification and Editing. Topics in Current Genetics Vol. 12. pp 431-438 SL Weiss, RA Sunde 1998 RNA 4:816 PM Moriarty, CC Reddy, LE Maquat 1998 Mol Cell Biol 18:2932 SC Low, MJ Berry 1996 Trends Biochem Sci 21:203 F Ding, PJ Grabowski 1999 RNA 5:1561 XM Xu, H Mix, BA Carlson, PJ Grabowski, VN Gladyshev, MJ Berry, DL Hatfield 2005 J Biol Chem 2S0A\56S C Gelpi, EJ Sontheimer, JL Rodriguez-Sanchez 1992 Proc Natl Acad Sci USA 89:9739 A Small-Howard, N Morozova, Z Stoytcheval, EP Forryl, JB Mansell, JW Harney, BA Carlson, XM Xu, DL Hatfield, MJ Berry 2005 Mol Cell Biol (In Press) DL Tumbula, HD Becker, WZ Chang, D Soil 2000 Nature 407:106 A Sauerwald, W Zhu, TA Major, H Roy, S Palioura, D Jahn, WB Whitman, JR Yates, M Ibba, D Soil 2005 Science 307:1969 PH Maenpaa, MR Bemfield 1970 Proc Natl Acad Sci USA67M8 D Hatfield, FH Portugal 1970 Proc Natl Acad Sci USA 67:1200 D Hatfield, A Diamond, B Dudock 1982 Proc Natl Acad Sci US A79:62\5 BA Carlson, XM Xu, GV Kryukov, M Rao, MJ Berry, VN Gladyshev, DL Hatfield 2004 ProcNatlAcadSciUSA 101:12848 AM Diamond IQQA Proc Natl Acad Sci USA 101:13395 RS Glass, WP Singh, W Jung, Z Veres, TD Scholz, TC Stadtman 1993 Biochemistry 32:12555 MJ Guimaraes, D Peterson, A Vicari, BG Cocks, NG Copeland, DJ Gilbert, NA Jenkins, DA Ferrick, RA Kastelein, JF Bazan, A Zlotnik 1996 Proc Natl Acad Sci USA 93:15086 lY Kim, TC Stadtman 1995 Proc Natl Acad Sci USA 92:7710 SC Low, JW Hamey, MJ Berry 1995 J Biol Chem 270:21659 W Leinfelder, TC Stadtman, A Bock 1989 J Biol Chem 264:9720 BJ Lee, PJ Worland, JN Davis, TC Stadtman, DL Hatfield 1989 J Biol Chem 264:9724 I Chambers, J Frampton, P Goldfarb, N Affara, W McBain, PR Harrison 1986 EMBO J 5:1221 F Zinoni, A Birkmann, TC Stadtman, A Bock 1986 Proc Natl Acad Sci USA83:4650 W Leinfelder, E Zehelein, M Mandrandberthelot, A Bock 1988 Nature 331:723 BJ Lee, M Rajagopalan, YS Kim, KH You, KB Jacobson, D Hatfield 1990 Mol Cell Biol 10:1940 J Heider, A Bock 1993 Adv Microb Physiol 35:71 GV Kryukov, S Castellano, SV Novoselov, AV Lobanov, O Zehtab, R Guigo, VN Gladyshev 2003 Science 300:1439 SV Novoselov, M Rao, NV Onoshko, H Zhi, GV Kryukov, Y Xiang, DP Weeks, DL Hatfield, VN Gladyshev 2002 EMBO J 21:3681 M Rao, BA Carlson, SV Novoselov, DP Weeks, VN Gladyshev, DL Hatfield 2003 RNA 9:923 RK Shrimali, AV Lobanov, XM Xu, M Rao, BA Carlson, DC Mahadeo, CA Parent, VN Gladyshev DL Hatfield 2005 Biochem Biophys Res Commun 329:147 AV Lobanov, C Delgado, S Rahlfs, SV Novoselov, GV Kryukov, S Gromer, DL Hatfield, K Becker, VN Gladyshev 2005 Nucl Acids Res 34:496. T Mourier, A Pain, B Barrell, S Griffiths-Jones 2005 RNA 11:119

Chapter 4. Evolution of selenocysteine decoding and the key role of selenophosphate synthetase in the pathway of selenium utilization Gustavo Salinas Cdtedra de Inmunologia, Facultad de Quimica-Facultad de Ciencias, Universidad de la Republica. Instituto deHigiene, Avda. A. Navarro 3051, Montevideo, CP 11600, Uruguay

Hector Romero Laboratorio de Organizacion y Evolucion del Genoma, Dpto. de Biologia Celular y Molecular, Instituto de Biologia, Facultad de Ciencias, Igud 4225, Montevideo, CP 11400, Uruguay

Xue-Ming Xu, Bradley A. Carlson and Dolph L. Hatfield Molecular Biology of Selenium Section, Laboratory of Cancer Prevention, Center for Cancer Research, National Cancer Institute, National Institutes of Health, Bethesda, MD 20892, USA

Vadim N. Gladyshev Department of Biochemistry, University of Nebraska, Lincoln, NE 68588, USA

Summary: The complete sequencing of genomes and the development of in silico methods for identification of genes encoding selenocysteine (Sec)containing proteins have greatly contributed to shape our view on the evolution of selenium utilization in nature. Current evidence is consistent with the idea that Sec decoding is a late addition to the genetic code and it evolved once, before the separation of archaeal, bacterial and eukaryal domains. Many organisms have lost the Sec decoding trait, but recent evidence has shown that the loss is not irreversible. The distribution of organisms that use UGA as a Sec codon suggests that Sec decoding evolved as a result of speciation, differential gene loss and horizontal gene transfer. Selenium is also used in the synthesis 2-selenouridine, a modified base of unknown function located in the wobble position of certain tRNAs. It has been recently demonstrated that selenouridine and Sec-decoding traits can evolve independently of each other, but both require selenophosphate synthetase. This ATP-dependent enzyme emerged as a key feature of selenium utilization that allows separation of selenium fi-om the pathways of

40

Selenium: Its molecular biology and role in human health

sulfur utilization and non-specific use of selenium. Some animals, including mammals, evolved two selenophosphate synthetases, highlighting an unknown complexity of selenium utilization in nature. Introduction Co-translational incorporation of selenocysteine (Sec) into nascent polypeptides is neither canonical nor universal. A Sec-decoding apparatus is needed to reprogram specific UGA codons [1-3]. The Sec-decoding apparatus and selenoprotein genes are present in the three domains of life; yet, many taxa lack them. In Sec decoding species, the selenoproteome consists of a restricted number of proteins [4,5]. All these observations have raised important questions regarding the evolution of Sec utiKzation in nature. For example, how and when did the translation machinery to decode Sec evolve? If it evolved once, has it been perpetuated solely by vertical descent? Has the UGA codon evolved from nonsense to sense or vice versa? Have extant selenoproteins evolved from Cys-containing proteins or vice versa? What are the selective forces that result in maintenance, loss and acquisition of the Sec-decoding trait and selenoproteins? In a broader scenario, studies on the evolution of Sec invite more in-depth questions regarding the evolution of the genetic code and the translation machinery. Recent work allowed some of these questions to be answered providing a provisional evolutionary scenario [5-7]. At the same time, some unknowns remain. In this chapter we review the current knowledge regarding Sec and selenium utilization in nature, their evolution, and highlight a key role of selenophosphate synthetase in these processes. Sec decoding: common origin before tlie division of tlie thiree domains? Current evidence strongly suggests that the Sec decoding trait evolved once, before the division of bacterial, archaeal and eukaryal domains. For example, there are fundamental similarities in the three domains: i) Sec is decoded by UGA-matching tRNA^'*' (also known as selC) and a dedicated elongation factor (EFsec, also known as selB); ii) the translational reprogramming is fulfilled by the SECIS element present in selenoprotein mRNAs; and iii) Sec synthesis occurs on a tRNA scaffold as reviewed in [2,3,8,9]. A common origin is further supported by the recent phylogenetic analysis of the genes involved in Sec decoding [7], which indicates that the trait is monophyletic in the bacterial domain and that eukaryal and archaeal Sec-decoding genes have a common ancestor. These observations suggest that the most parsimonious and likely evolutionary scenario for the trait is a common origin between the three domains, and not independent origins. The greater similarity between archaeal and eukaryal domains may reflect the fact that the transcription and translation machinery in archaea and eukarya is thought to be of common origin. Further studies should be carried out to identify and

Evolution of Sec decoding

41

date the time of divergence of the different genes involved in Sec decoding and compare these patterns with divergence of the three domains. Syvanen [10] has proposed that the unity of the genetic code is the product of an evolutionary process that has continued since the diversification of the major domains and specifically suggested that the last common ancestor (which defines the origin of the three domains) did not use arginine and tryptophan. In this alternative scenario, horizontal gene transfer (HGT) would have played a critical role in homogenization of the code. If this proposition is correct, it is then possible that Sec might not have been the 2V^ amino acid added to the genetic code. Despite the remarkable similarities, differences do exist in the Secdecoding traits in the three domains of life, including an increased complexity of the pathway in eukaryotes [11-14]. It has been argued that the differences between bacterial and eukaryal Sec incorporation are due to a refinement of the mechanism to provide an increased efficiency in Sec incorporation [15]. It is also likely that some changes favored a greater flexibility in reprogramming. Indeed, the location of the SECIS element within the untranslated region in archaea and eukarya released the constraints imposed by the location of SECIS immediately downstream of UGA^'^ within the coding region of bacterial messages. The Sec decoding trait can be lost, but not irreversibly Recently, the distribution of the Sec-decoding trait was analyzed systematically by searching complete genomes for the presence of genes involved in Sec decoding and selenoprotein(s) and using this information to construct a provisional "Sec decoding map" within the "tree of life" [7]. This study revealed that the trait is present in most phyla, but absent in many species, and provided clues regarding the evolution of Sec. Then, the phylogenies of the Sec-decoding genes were inferred and compared to organismal phylogenies. This approach explained the spread and "holed" pattern of Sec-decoding species within the tree of life as the result of speciation, differential gene loss and horizontal gene transfer (HGT). It also revealed that the loss of the trait is a phenomenon that takes place within clades at different evolutionary levels, implying that the loss occurred not only at rather basal evolutionary levels (e.g., phylum and class), but also in recent lineages {e.g., genus and species). A stunning example of the latter is the case of the C092 and mediaevalis strains of Yersinia pestis that have lost the ability to decode Sec (possess functional tRNA^*^ and Sec synthase but an EFsec pseudogene) while the KIM strain retains this ability (unpublished). Yet, the main disclosure of the study was that it clearly demonstrated that the loss of the trait was not irreversible, indicating that the genetic code can be "rewired" by HGT, a possibility previously thought as highly unlikely [16]. This phenomenon was patently observed in the case of

42

Selenium: Its molecular biology and role in human health

Photobacterium profundum, which did not acquire its extant trait and the selenoproteins involved in the glycine reductase complex by vertical descent from a proteobacterial ancestor, but rather from a different lineage. Incongruences possibly attributable to HGT, between gene and species frees were also observed in the case of Pseudomomas spp. (y-proteobacteria), Sinorhizobium meliloti (a-proteobacterium) and Burkholderia spp (Pproteobacteria). In these species, the only selenoprotein is the a-subunit of formate dehydrogenase. In this regard, it should be noted that Pseudomonas spp, S. meliloti and Burkholderia pseudomallei are soil colonizing bacteria [17,18]; whereas S. meliloti and Pseudomonas spp. even compete for nodulation on some plants. It was speculated that a putative vector for acquisition of the frait may exist, on the basis of the observation that S. meliloti contains the genetic information for selenoproteins and the Secdecoding frait within a megaplasmid (pSymA) with a high number of transposons [15]. Selenophosphate synthetase: an essential enzyme for selenium utilization Selenophosphate synthetase (SPS) is an essential enzyme for selenium utilization: it catalyzes the synthesis of monoselenophosphate [19], a reduced and reactive form of selenium, which provides the selenium atom for synthesis of Sec and 2-selenouridine, another biologically relevant form of selenium in nature (see below and [20]). There appears to be two groups of SPS enzymes. One group contains Sec or Cys at the active site and the corresponding E. colt enzyme can catalyze, in vitro, the synthesis of selenophosphate from selenide and ATP [21] as follows: ATP + HSe" + H2O -> H2SeP03' + AMP + Pi. It should be emphasized that the Km value for selenide is 20 fjM, a concenfration that would be noxious for the cell in vivo, suggesting that selenide would not be the physiological selenium donor for SPS2/SelD. The other group (designated SPSl) is present exclusively in some eukaryotic organisms that also possess SPS2 (e.g., mammals); SPSl neither contains Sec or Cys at the predicted active site position nor appears to catalyze the in vitro reaction depicted above [22]. In vivo, human SPS2 complemented an E. coli selD mutant strain when the medium was supplemented with selenite or Sec, restoring the activity of the selenoprotein formate dehydrogenase to E. coli wild type levels. In confrast, complementation with SPSl was ineffective when selenite was used as a selenium supply, although it improved when Sec was used in the medium [22]. These results led Tamura et al [22] to propose that human SPS2 functions in the pathway of de novo synthesis of selenophosphate from selenite, after reduction of the latter, presumably, by intracellular thiols; the Sec residue of SPS2 active site would bind this reduced selenium to form an enzyme substrate complex. Alternatively, for SPSl catalysis, mammalian

Evolution of Sec decoding

43

cells would supply an atom of Se derived from a Sec salvage pathway that may recycle Sec derived from selenoproteins or from the promiscuous incorporation of selenium instead of sulfur in the Cys metabolic pathway. Sec P-lyases, enzymes that catalyze the conversion of Sec to Ala and a selenium transfer form (Se*, since the redox state of Se has not been determined), and NifS and NifS-like proteins (enzymes that provide a sulfur atom to iron-sulfur clusters by catalyzing Cys desulfuration, and also convert Sec to Ala and Se*) are candidate enzymes to participate in selenium mobilization from Sec [23,24], and have been proposed as key players for the Sec-salvage pathway. It is relevant to emphasize that the bacterial NifS-like proteins CsdB, CSD, and IscS also have an important role in selenium mobilization from Sec and selenophosphate synthesis. Indeed, Sec and bacterial NifS-like proteins can effectively replace the high level of free selenide in the in vitro SPS assay [25,26]. Furthermore, the E. coli SPS (C17S) mutant, which is inactive in the standard in vitro assay with selenide as substrate, was found to be active in the presence of Sec and NifS proteins, suggesting a selenium delivery function for these proteins [26]. Evidence that selenium is mobilized in vivo from free Sec has also been obtained by Lacourciere [27]: growth of £•. coli in the presence 0.1 |LIM ^^Se SeOs^' and increasing amounts of Sec resulted in a concomitant decrease in ^^Se incorporation in formate dehydrogenase and bulk tRNA. This led Lacourciere to propose that NifS-like proteins are components of a selenium delivery system for the biosynthesis of selenophosphate. Recently, other potential selenium-binding and delivery proteins for SPS have been characterized. Human 3-mercaptopyruvate sulfur fransferase (MST) and glyceraldehyde-3-phosphate dehydrogenase (GAPDH), bound selenium supplied as selenodiglutathione formed from SeOs^" and glutathione; the bound selenium was readily released from MST and GAPDH and available as a substrate for bacterial SPS [28]. Thus, in vitro, these selenium-binding proteins and a low level of selenodigultathione (formed from selenite and glutathione) could effectively replace the high concenfrations of selenide used in SPS assays. Despite this considerable progress, the metabolic pathway(s) of selenium assimilation and the physiological system that donates selenium to SPS remain to be completely characterized and represents one of the challenges ahead in the selenium field. Selenophosphate synthetase: one enzyme, two selenium utilization traits 2-Selenouridine is the second major biological form of Se in nature [29]. It is a modified base so far identified exclusively in the wobble position of the anticodon of tRNA^=", tRNA°'" and tRNA°'" in some bacteria [30]. 2Selenouridine is synthesized by the protein designated YbbB from

44

Selenium: Its molecular biology and role in human health

thiouridine tRNA and selenophosphate (the latter as the selenium donor). Although it was originally thought that organisms able to synthesize 2selenouridine were the same as those able to decode Sec, it has recently been shown that the set of organisms that synthesize 2-selenouridine overlaps with, yet is distinct from, the set of organisms able to decode Sec (Figure 1).

SPS n = l

Figure 1. Sec-decoding and selenouridine (SeU) synthesis traits both require selenophosphate synthetase (SPS). However, the set of species that decode Sec overlaps with, but is different from, the set of species that synthesize SeU. The representation excludes species of the eukaryal domain. Note that there is one species possessing SPS, but neither trait (see text). n= numbers of completed prokaryotic genomes. (Total number of complete genomes analyzed was 153.)

Indeed, SPS is required for both Sec and selenouridine synthesis. This study allowed us to define SPS as the gene signature of selenium utilization, YbbB as the gene signature for 2-selenouridine synthesis and EFsec and tRNA^^" as the gene signature of the Sec-decoding trait. Thus, the likely evolutionary scenario is that SPS is required for both traits, but Sec-decoding and 2selenouridine synthesis traits can evolve independently of each other. Analyses of genomic organization of Se utihzation genes in the bacterial domain revealed that SPS is more often arranged in an operon with ybbB than with EFsec (selB) and Sec synthase {selA). HGT events were also identified for the selenouridine synthesis trait; thus, the pattern of species

Evolution of Sec decoding

45

distribution for both Se traits is the result of speciation, differential gene loss and HGT. An interesting parallel that might be sketched for both traits is the restricted use of selenium: it is certainly peculiar that the Sec decoding apparatus has been maintained in some species for insertion of a single amino acid into a single protein; equally unusual is the finding that 2-selenouridine is used in only three bases in the entire transcriptomes of bacteria possessing ybbB. Although several functions have been postulated for selenouridine, its function is not known. Based on the facts that 2-selenouridine is found exclusively at the wobble position of codons ending in a purine, and that these codons pose a problem for the translation machinery, it has been postulated that 2-selenouridine would have been an adjustment of the decoding apparatus to increase translational fidelity [7]. Whether this base modification occurs outside the bacterial domain is not known; a low identity homolog to bacterial ybbB is present in Methanococcus jannaschii and Methanopyrus kandleri [7,30]. Finally, the presence of a SPS homolog with high identity in Enterococcus faecalis, a species that neither decodes Sec nor possesses ybbB is interesting. Furthermore, a Sec lyase homolog and a protein involved in sulfur reduction flank this gene in the genome. This observation suggests that there may be an additional Se utilization trait that occurs in E. faecalis and perhaps other organisms (unpublished). Looking at the phenotype: evolution of selenoproteins is a Iiighly dynamic process Evolution of selenoproteins is a highly dynamic process linked to the evolution of the Sec decoding trait itself. From the mechanistic point of view, this process involves: i) evolution of selenoproteins from Cyscontaining homologs concomitant with the acquisition of SECIS elements [5], ii) evolution of Cys-containing proteins from selenoproteins (fossil SECIS elements have been well documented even in Sec decoding species) [4,31], iii) gene duplication [32,33], and iv) conceivably "invention" of entirely new selenoprotein families for which no Cys homologs have been detected (possibly the glycine reductase component, grdA, in bacteria [5,15], and SelJ in eukarya [5,34]). This highly dynamic process is at least in part due to the following processes: i) Sec and Cys can be replaced by each other concomitant with appropriate adjustments in the protein context [35] and/or protein concentration, and ii) the SECIS element appears to evolve easily, especially in archaea and eukarya, in which the 3'-UTR location of this structure is not constrained by the coding region. Yet, the evolutionary forces that shape Sec utihzation, in particular the delicate balance between processes that maintain, acquire or lose this trait

46

Selenium: Its molecular biology and role in human health

are not clear. It has been previously postulated that the selective advantage provided by selenoenzymes over Cys-homologs (due to the better nucleophilicity and lower pKa of SeH in Sec over SH in Cys), might become a disadvantage if selenium supply becomes limiting [7,15,36]. In these situations, enzymes containing Sec as catalytic residues could have evolved into Cys-containing proteins, or alternatively, both Sec-containing and Cyscontaining forms could have been maintained allowing organisms to use Se in a facultative manner [36]. An illustrative example of a Se facultative organism is Methanococcus maripaludis, which represses the synthesis of the Cys homologs when grown in a medium that contains adequate amounts of Se, but this repression is not observed in a mutant with disrupted selB [36]. Nevertheless, the occurrence of organisms carrying only one of the selenium utilization traits indicates that Se availability might not be the sole factor involved in the loss of either trait. It is also of interest that selenoproteomes, while differing in size in composition between various organisms, always represent a very small subset of the total protein complement. One may consider Sec as a twoedged weapon - the advantage of having a highly reactive selenol group might become a disadvantage if used indiscriminately to replace Cys. This may restrict the pervasive use of Sec in the environments where Se supply may be adequate or excessive [7]. Another conspicuous feature of Sec utilization in nature is its idiosyncratic use by different taxa: different sets of selenoproteins have evolved in different lineages [6,34,37]. This finding indicates adaptations in the use of Sec, presumably to fiilfill particular needs. It has been suggested that both lineage-specific expansion (presumably of recent origin), and the presence of core selenoproteins (ancient origin) appear to contribute to extant selenoproteomes [5]. Analyses of prokaryotic selenoproteomes revealed that formate dehydrogenase is present in most Sec decoding organisms [5], suggesting that under anaerobic respiration Sec-containing formate dehydrogenase confers a specific advantage [7]. Indeed, most of selenoprotein formate dehydrogenase-containing species are obligatory anaerobes or facultative aerobes; the sole exception appears to be S. melilloti, a symbiotic nitrogenfixing obligatory aerobe that lives in the oxygen-limited environment of the nodule. On the other hand, glycine reductases, present in T. denticola, P. profundum, T. tengcongensis and several species of the genus Clostridium might have conferred a selective advantage allowing certain anaerobic bacteria to conserve energy via a soluble substrate level phosphorylation system [38]. Synthesis of Sec-tRNA^*^*^: a non-canonical mechanism A conspicuous feature of Sec synthesis is that it occurs on its tRNA, fi-om Ser-tRNA^'', which is then converted into Sec-tRNA^**^ using

Evolution of Sec decoding

47

monoselenophosphate as the selenium donor [39,40]. This latter reaction is catalyzed by Sec synthase (SelA) in the bacterial domain. The equivalent enzyme(s) in archaea and eukarya is (are) not known, although the mechanism is thought to involve a Sep-tRNA^*'' intermediate (Sep=phosphoserine) [41].

Table 1. Non-cognate charging of amino acids into tRNAs Final aa attached to tRNA

Initial aa charged to tRNA

Charging aminoacyl-tRNA synthetase

Phylogenetic Distribution

Asn

Asp, Glu

aspRS gluRS

archaea bacteria

Cys

Sep

sepRS

archaea

fMet

Met

metRS

Gin

Glu, Asp

gluRS aspRS

bacteria, organelles archaea bacteria

Pyl

Lys

lysRS1+lysRS2

archaea

Sec

Ser

serRS

archaea bacteria eukarya

Nature has evolved two different mechanisms for synthesis of aminoacyltRNAs: i) the canonical one {i.e., a specific amino acyl tRNA sjmthetase recognizes an amino acid and its cognate tRNA), and ii) the non-canonical mechanism wherein the tRNA is loaded first with a "non-cognate" amino acid that is then modified to the amino acid to be incorporated into protein. Although the amino acid biosynthesis on a tRNA scaffold is not unique to Sec (Table 1, after [42] and [43]), it is interesting to point out that Sec is the sole example in which the second mechanism appears to occur in the three domains of life, and might be the predominant or even exclusive biosynthetic mechanism. It is also important to note that Sec synthesis appears to resemble Cys synthesis on tRNA in two aspects: i) in archaea and eukarya, Ser-tRNA^^° is first converted into Sep-tRNA^", whereas Sep is the intermediate in Cys synthesis on tRNA in some archaea [44], and ii) synthesis of Cys on tRNA is the sole mechanism by which Cys is synthesized in a subset of archaea that use this strategy. It has been speculated that SepCysS (the enzyme that charges Sep directly to tRNA^^^) provided a means by which both Cys and Sec may have been originally added to the genetic code [44]. More generally, it could be speculated that in

48

Selenium: Its molecular biology and role in human health

situ synthesis of an amino acid provides a strategy by which the amino acid repertoire (and the genetic code) could have been, and conceivably can be expanded. Lessons from the genetic code Nearly four decades ago the genetic code was deciphered [45]. Yet, the signals involved in translation of information are neither universal nor completely known, and several mechanisms of reprogramming have been documented [46]. Sec decoding provides clues to how, with very few genes, a codon can be reprogrammed in specific messages, increasing the amino acid repertoire. In addition, it illustrates how flexible and dynamic is the evolution of the genetic code: the ability to decode Sec can be lost, but later reacquired by HGT of a handful of genes. The case of Sec is also interesting in that it is a "non-standard" amino acid that is co-translationally incorporated, but has not been fixed as the only (or major) function of UGA in the genetic code. In turn, this raises the question of why Sec has not been "hardwired" (i.e., reassigned a codon), and what are the differences between increasing the amino acid repertoire post-translationally or cotranslationally. Stop codons have repeatedly evolved particular meanings (Trp, Cys, Sec, Pyl) either by codon reassignment or by reprogramming, suggesting that a change of meaning could be less deleterious for a nonsense codon. In particular, whether the UGA codon originally specified Sec or stop has been a matter of debate. It was speculated [16] that UGA evolved fi"om sense to nonsense, and postulated a possible scenario where Sec was one of the earliest amino acids in evolutionary history. And once oxygen evolved in the atmosphere, the susceptibility to oxidation could have counterselected Sec resulting in the stop function of UGA and maintaining the Sec function by means of SECIS element and EF-sec. On the other hand, it was also proposed [47] that UGA evolved from nonsense to sense, and that Sec was added to the already fixed code. Although the question of the ancestral meaning of UGA is difficult to solve, the very need of reprogramming specific messages suggests that the Sec insertion has evolved on top of an existing translational machinery to acquire a new meaning for an already assigned codon. In other words, "loose" programming (for specific messages) must have been a novelty added to the "hardwired" code (i.e., universal, for all messages). Further clues to the idiosyncrasies of the genetic code and its dynamic evolution have arisen fi'om pyrrolysine (Pyl), a recently discovered "nonstandard" amino acid [48]. The mechanism of Pyl incorporation, in particular whether it requires recoding (and if it does, what are the signals) is not clear [49]. Although Sec and Pyl use dissimilar decoding strategies, there are clear parallels: Sec and Pyl are amino acids used in a small set of

Evolution of Sec decoding

49

proteins and both are expansions of the amino acid repertoire by redefining the meaning of stop codons. The diversity of mechanisms involved in translation, together with the simple machinery needed to expand the genetic code, opens the door to the possibility that 22 genetically encoded amino acids may not be the final number and that additional amino acids may exist in the genetic code. References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17.

18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32.

A Bock 2000 Biofactors 11:77 MJ Berry, RM Tujebajeva, PR Copeland, XM Xu, BA Carlson, GW Martin, 3rd, SC Low, JB Mansell, E Grundner-Culemann, JW Harney et al 2001 Biofactors 14:17 DL Hatfield, VN Gladyshev 2002 Mol Cell Biol 22:3565 GV Kryukov, S Castellano, SV Novoselov, AV Lobanov, O Zehtab, R Guigo, VN Gladyshev 2003 Science 300:1439 GV Kryukov, VN Gladyshev 2004 EMBO Rep 5:538 Y Zhang, DE Fomenko, VN Gladyshev 2005 Genome Biol 6:R37 H Romero, Y Zhang, VN Gladyshev, G Salinas 2005 Genome Biol 6:R66 M Rother, A Resch, R Wilting, A Bock 2001 Biofactors 14:75 A Lescure, D Fagegaltier, P Carbon, A Krol 2002 Curr Protein Pept Sci 3:143 M Syvanen 2002 Trends Genet 18:245 SA Kinzy, K Caban, PR Copeland 2005 Nucleic Acids Res 33:5172 XM Xu, H Mix, BA Carlson, PJ Grabowski, VN Gladyshev, MJ Berry, DL Hatfield 2005 J Biol Chem 280:41568 L Chavatte, B A Brown, DM Driscoll 2005 Nat Struct Mol Biol 12:408 AL Small-Howard, MJ Berry 2005 Biochem Soc Trans 33:1493 PR Copeland 2005 Genome Biol 6:221 A Bock, K Forchhammer, J Heider, C Baron 1991 Trends Biochem Sci 16:463 MT Holden, RW Titball, SJ Peacock, AM Cerdeno-Tarraga, T Atkins, LC Grossman, T Pitt, C Churcher, K Mungall, SD Bentley et al 2004 Proc Natl Acad Sci USA 101:14240 PM Riccillo, CI Muglia, FJ de Bruijn, AJ Roe, IR Booth, OM Aguilar 2000 J Bacterial 182:1748 W Leinfelder, K Forchhammer, B Veprek, E Zehelein, A Bock 1990 Proc Natl Acad Sci USA 87:543 TC Stadtman, JN Davis, E Zehelein, A Bock 1989 Biofactors 2:35 lY Kim, Z Veres, TC Stadtman 1992 J Biol Chem 267:19650 T Tamura, S Yamamoto, M Takahata, H Sakaguchi, H Tanaka, TC Stadtman, K Inagaki 2004 Proc Natl Acad Sci U S A \0hl6i 62 A Lescure, D Gautheret, P Carbon, A Krol 1999 J Biol Chem 274:38147 H Mihara, T Kurihara, T Watanabe, T Yoshimura, N Esaki 2000 J Biol Chem 275:6195 GM Lacourciere, TC Stadtman 1998 J Biol Chem 273:30921 GM Lacourciere, H Mihara, T Kurihara, N Esaki, TC Stadtman 2000 J Biol Chem 275:23769 GM Lacourciere 2002 JBacteriol 184:1940 Y Ogasawara, GM Lacourciere, K Ishii, TC Stadtman 2005 Proc Natl Acad Sci USA 102:1012 Z Veres, TC Stadtman 1994 Proc Natl Acad Sci USA 91:8092 MD Wolfe, F Ahmed, GM Lacourciere, CT Lauhon, TC Stadtman, TJ Larson 2003 J Biol Chem 279:lS0l A Bock, M Rother 2004 Arch Microbiol 183:148 JA Vorholt, M Vaupel, RK Thauer 1997 Mol Microbiol 23:1033

50 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49.

Selenium: Its molecular biology and role in human health S Castellano, SV Novoselov, GV Kryukov, A Lescure, E Blanco, A Krol, VN Gladyshev, R Guigo 2004 EMBORep 5:71 S Castellano, AV Lobanov, C Chappie, SV Novoselov, M Albrecht, D Hua, A Lescure, T Lengauer, A Krol, VN Gladyshev et al 2005 Proc Natl Acad 5cj [/ 5.4 102:16188 HY Kim, VN Gladyshev 2005 PLoSBiol 3x375 M Rother, I Mathes, F Lottspeich, A Bock 2003 J Bacteriol 185:107 R} Stmwell, MJBeny 2005 Proc Natl AcadSciUS A 102:16123 JR Andreesen 2004 Curr Opin Chem Biol 8:454 BJ Lee, PJ Worland, JN Davis, TC Stadtman, DL Hatfield 1989 J Biol Chem 264:9724 BJ Lee, M Rajagopalan, YS Kim, KH You, KB Jacobson, D Hatfield 1990 Mol Cell 5/0/10:1940 BA Carlson, XM Xu, GV Kryukov, M Rao, MJ Berry, VN Gladyshev, DL Hatfield 2004 Proc Natl Acad Sci U S A 101:12848 M Di Giulio 2005 Biosystems 80:175 M Ibba, D Soil 2004 Genes Dev 18:731 A Sauerwald, W Zhu, TA Major, H Roy, S Palioura, D Jahn, WB Whitman, JR Yates, 3rd, M Ibba, D Soil 2005 Science 307:1969 FH Crick 1966 Cold Spring Harbor Symp Quant Biol 3 \ A O Namy, JP Rousset, S Napthine, I Brierley 2004 Mol Cell 13:157 VN Gladyshev, GV Kryukov 2001 Biofactors 14:87 B Hao, W Gong, TK Ferguson, CM James, JA Krzycki, MK Chan 2002 Science 296:1462 Y Zhang, PV Baranov, JF Atkins, VN Gladyshev 2005 J Biol Chem 280:20740

Chapter 5. SECIS RNAs and K-turn binding proteins. A survey of evolutionary conserved RNA and protein motifs Christine Allmang and Alain Krol Architecture et Reactivite de I'arN, UPR 9002 du CNRS-Universite Louis Pasteur, Institut de Biologic Moleculaire et Cellulaire, 67084 Strasbourg, France

Summary: The SelenoCysteine Insertion Sequence (SECIS) is a stem-loop structure residing in the 3' untranslated region of all selenoprotein mRNAs. Its presence is mandatory to allow the ribosome to readthrough the UGA selenocysteine codon. The SECIS RNA possesses a well-defined secondary structure. Four consecutive non-Watson-Crick base pairs, with a central tandem of sheared G.A/A.G base pairs, constitute the functional motif of the SECIS RNA which is recognized by the SECIS binding protein SBP2. The tandem of sheared base pairs is part of a recurrent motif, the kink-turn (Ktum), occurring in a variety of different RNAs. The K-tum is a helix-internal loop-helix composed of a non-Watson-Crick stem containing the G.A base pairs and a canonical stem. The internal loop between the stems is always asymmetrical and usually contains three unpaired nucleotides on one strand and none on the other. We propose here that the SECIS RNA must represent a K-tum variant with regard to the limited structural differences that distinguish it from consensus K-tums. Work by others showed that ribosomal protein L30 also binds the SECIS RNA in a specific manner. L30 and SBP2 are members of a family of proteins sharing the same RNA-binding domain called L7A/L30. All proteins possessing this fold recognize K-tum RNAs. Three stmctures of RNA-protein complexes containing the L7A/L30 protein fold and cognate K-tum RNAs have been solved. In light of the interaction principles goveming these RNA-protein complexes, we discuss how L30 can recognize the SECIS RNA. Collectively, all the findings suggest that the L7A/L30 protein fold and the K-tum are ancient stmctural motifs that have evolved various functions, from pre-mRNA splicing to protein synthesis. Introduction The field of eukaryotic selenoprotein research is fascinating in several respects. First, the existence of taxa-specific selenoproteins altered the initial perception that mammals recapitulate the eukaryotic selenoproteome. Second, it becomes increasingly apparent that the number of molecular

52

Selenium: Its molecular biology and role in human health

partners involved in selenoprotein synthesis is larger than previously thought. Relocation of the SECIS element, from the coding frame in bacteria to the 3'-untranslated region (3'UTR) of selenoprotein mRNAs in eukaryotes, may be responsible only in part for this complexification. hideed, even selenocysteine biosynthesis itself seems to take a more sophisticated pathway in eukaryotes. The SECIS stem-loop contains a well defined structural motif composed of four consecutive non-Watson-Crick base pairs, with a central tandem of sheared G.A base pairs [1]. This motif (Figure lA) also ensures a functional role as it is essential to selenocysteine incorporation in vivo [2,3] and constitutes the binding site of SBP2, a protein binding specifically to the SECIS RNA [4,5]. The SBP2 RNA-binding domain contains a region sharing a high degree of amino acid sequence similarity to the L7A/L30 protein family containing ribosomal proteins L7Ae and L30, the 15.5kD/Snul3p spliceosomal protein and other functionally unrelated proteins [6,7]. Cocrystal structures of the L7Ae, L30 and 15.5 kD proteins in complex with their cognate RNAs revealed that the proteins fold into a highly conserved compact globular domain, the L7A/L30 domain, that binds specifically to RNAs possessing a kink-turn (K-tum) motif The canonical Ktum is a recurrent element, occurring notably in ribosomal RNAs, U4 snRNA, and box C/D regions of snoRNAs and archaeal sRNAs. It contains a tandem of sheared G.A/A.G base pairs that have an important structural role in forming and stabilizing the turn [8]. In earlier studies, we proposed a 3D model for the SECIS RNA where the phosphodiester backbone is bent at the non-Watson-Crick base pairs [1]. Combined with the presence of sheared base pairs, the proposed folding of the SECIS RNA suggests that it could be a canonical K-tum or a K-tum related RNA. From all these findings emerges the important issue of how different RNAs harboring K-tum motifs can selectively discriminate proteins sharing the same RNA-binding domain. This is a particularly burning question in light of the finding that ribosomal protein L30 is another SECIS-binding protein [9]. This chapter will describe the SECIS RNA stmcture with comparison to canonical K-tum RNAs, then highlight the similarities/differences between protein-RNA complexes formed with proteins of the L7A/L30 family and K-tum RNAs. The SECIS RNA: a K-turn variant An experimental secondary stmcture model for SECIS RNAs (Figure lA) was proposed about ten years ago based on stmcture probing in solution [1]. It was next discovered that certain SECIS RNAs can adopt a slightly different 2D stmcture at their apex [10]. Called Form 2 SECIS (Figure lA), they possess an additional helix III but a shorter apical loop, compared to Form 1 SECIS. Besides the non-Watson-Crick quartet. Form 2 shares with Form 1 the other conserved features characterizing SECIS RNAs, i.e. the run

SECIS RNAs andK-turn binding proteins

53

of As in the apical (Form 1) or internal loop II (Form 2) and the 13-15 bp long helix II. More systematic identification of a variety of novel selenoprotein mRNAs including vertebrates, invertebrates and green algae [11-20] clearly established that Form 2 SECIS are more widespread than Form 1. However swapping experiments could not establish that Form 2, although preponderant, provides a functional advantage to selenocysteine incorporation [10].

_| Apical loop 7-17nt

A/c A/c

Helix II 13-15 bp

N.N A. Q G. A U A/G

Internal loop I 3-14 nt

I Apical loop 3-1 Om

J Bulge or I Internal loop II 3-5 nt

N . N Non Watson-Crick quartet

LV

A. G G• A U.N. A/G H/G \

\ )

Helix I mm 7 bp

G

C

u

u

G-%-

U

G

C

C

U - « - U

G

C

A D—^ Q

A ® - « U u

UrG BP1 U 5,U

G
Uu„

A— A^ C C 5'

u u

. A "

A A A3,

Rat DIOI SECIS

^ I ^

GC CIS Watson-CricK AU cis walson-crick

—%—

cts Wa1$or-Cr!Ck

/ ^ ^ G G 3'

U4RNA

Q-|>- Irans

^>-

NC-stem

BP3 A 0—1> G BP2a^__Q A

A,D-|>G

BP3 A D — ^ G BP2 G 4 _ n A

G

Q ^ BP1 C ^ G 5'

P^BP4 ^ G c-stem C 3'

Consensus K-turn

Hoogsleen/Suga Edge

rails Sugar Edger'Sugar Edge

Figure 1. Structure models for the SECIS RNA. (A) Secondary structure models of Forms 1 and 2 SECIS. The conserved sequence and structural features are indicated. N, any nucleotide; A/G and A/c indicate that A is the prevalent base. (B) Secondary structure diagrams of the U4 snRNA and consensus K-tums adapted from [22,26] and the putative SECIS K-tum of the rat type I iodothyronine deiodinase (DIOI). BPl to BP5 stands for base pairs 1 to 5. Circled bases are discussed in the text. NC-stem: non-Watson-Crick stem; C-stem: Watson-Crick stem. Broken lines in U4 snRNA stand for hydrogen bonds between N6A30 and 2'OH of A44, N1A44 and 2'OH of A29 [22]; the latter interaction differs from that proposed at the homologous position in the consensus K-tum. The graphical conventions for displaying nonWatson-Crick base pairs arefrom[28].

54

Selenium: Its molecular biology and role in human health

It is striking that mRNAs encoding the same selenoprotein can harbor either a Form 1 or a Form 2 SECIS, depending on the animal species. This is well exemplified by the SelM mRNA where Form 2 occurs in mammals while zebrafish harbors Form 1 SECIS [14]. Remarkably, chemical probing experiments indicated that the conserved As are single stranded and well accessible whatever the SECIS form [21]. In a few cases, especially in the mammalian SelM and some Chlamydomonas Form 2 SECIS, Cs are found instead of As without apparently altering the SECIS function [14,15]. Thus the universal conservation of the As, which was taken for granted at the time when the number of available SECIS sequences was too little to make statistically valid comparisons, is called into question. The mechanistic role of these unpaired A/Cs is still unknown, but their functional importance has been experimentally proven in vivo by sitedirected mutagenesis. Along the same lines, the nucleotide 5' to the nonWatson-Crick quartet is A in the vast majority of SECIS RNAs. However, compilation of selenoprotein mRNA sequences in other organisms led to the conclusion that G can sometimes be found instead, an interesting example being provided by the single selenoprotein mRNA in nematodes [11,20,21]. This correlates with in vivo experiments and mobility shift assays with SBP2 and SECIS RNAs concluding that an A is preferred but not mandatory [5,21]. In conclusion, it emerges from phylogenetic studies that SECIS RNAs exhibit a remarkable conservation of the 2D structures but few invariant nucleotides. Clearly, elucidation of the function of the single stranded A/C and conserved length of helix II is a necessary step toward an in-depth understanding of the function of the SECIS RNA. The non-Watson-Crick quartet at the foot of helix II is a characteristic feature recognized by SBP2 (Figure IB). The central G.A tandem was shown by structure probing experiments and computer modeling to adopt the sheared geometry [1]. Tandem sheared base pairs were initially discovered in the crystal structure of the 5' stem-loop of U4 snRNA in complex with the 15.5 kD protein [22]. The prevalence of this RNA motif was in fact revealed by the analysis of the atomic structures of the large and small ribosomal subunits where it occurs six times in H.marismortui 23S rRNA and twice in T.thermophilus 16S rRNA [8]. Its presence was further identified in the crystal structures of three other RNA-protein complexes: the yeast ribosomal protein L30e with its pre-mRNA, and the archaeal ribosomal protein L7Ae in complex with box C/D or box H/ACA sRNAs [23-25]. A two-dimensional representation of the tertiary structure of a consensus K-tum is diagrammed in Figure IB, which was adapted from [26]. In this publication, the authors derived the consensus from examination of K-tums in crystal structures and compared them with the sequence alignments of rRNAs from the three kingdoms of life. The K-tum is a two-sfranded, helix-internal loop-helix motif comprising about 15 nucleotides. The first stem (canonical or C-stem)

SECIS RNAs and K-turn binding proteins

55

ends at the internal loop with two Watson-Crick base pairs, mostly G-C. The non-canonical stem (NC-stem) starts typically with the sheared G.A base pairs. The internal loop is always asymmetrical with three unpaired nucleotides on one strand and none on the other. Because of the cross-strand stacking of the sheared base pairs, a sharp bend of the sugar-phosphate backbone occurs between the C and NC-stems. In fact, five base pairs characterize the K-tum motif (Figure IB): the Watson-Crick C-G base pair 1, the sheared G.A base pairs 2 and 3, the triple A.C-G base pair 4, and G.A base pair 5. The adenine of base pair 4 mediates the minor groove interaction with the C-stem (A-minor motif; see reference 27) and is crucial for K-tum folding. Figure IB shows the 2D structure model of the non-Watson-Crick quartet of the rat type I iodothyronine deiodinase (DIOI) SECIS RNA [1] compared to the structure of the U4 snRNA K-tum motif adapted from [22]. Visual inspection of the SECIS 2D stmcture identified an important K-tum characteristic feature: the C-stem separated by an internal loop from the NCstem comprising the invariant sheared base pairs. Despite the similarity, a few SECIS specific stmctural features led us to ask whether they form genuine K-tums. The non-Watson-Crick U.U base pair 3' to the sheared base pairs will not be discussed further because it displays sequence variation in SECIS and other RNAs and does not participate directly in the K-tum interactions [26]. The first question concerns the U residue (circled in Figure IB) 5' to the sheared base pair 2. Chemical probing experiments detected that it forms a non-Watson-Crick U.U base pair in the naked SECIS RNA [1]. In contrast, the homologous position is unpaired in U4 snRNA and in the consensus K-tum (Figure IB; see also Figure 2B). Moreover, data from crystal stmctures of RNA-protein complexes showed that the base at this position protmdes away from the RNA chain and is tightly bound in a pocket of the protein [22-25]. However, one cannot exclude the possibility that an SBP2-promoted induced fit leads to impairing and flipping out of the U residue. It could thus be the positional analog of the protmding base in the other K-tums (Figures IB and 2B). The second question asks whether the counterparts to base pairs 4 and 5 of the consensus K-tum also exist in the SECIS RNA as chemical probing cannot detect them. Formation of base pair 4 in the SECIS RNA will only depend on the sequence of base pair 1 since A is invariant in base pair 3 (Figure IB). Base pair 1 is U-A in the SECIS RNA shown, very often C-G and G-C but rarely A-U or G.U in others [1,11-21]. Interestingly, tables of sequence variation in [26] show the prevalence of CG, G-C or U-A at base pair 1 in K-tums, indicating that formation of base pair 4 is theoretically possible in SECIS RNAs. Likewise, base pair 5 could form in SECIS RNAs as tables in [26] established that base pairing is permitted in canonical K-tums between the invariant A of base pair 2 and any nucleotide. Lastly, one could argue that the size of internal loop I of SECIS RNAs is larger than in canonical K-tums.

56

Selenium: Its molecular biology and role in human health

o N

u 1, u u G A

N' ;RU

33

G

^ ^ U

3' G ,1 .«.~A

A A^44 30 -A C — GA V !

NN-- N 5 - N -- N j

5'

SECIS RNA

G—C

^

•

•-

U

• / . , . , U-A~l" U • G, A G C-G*^

3'

U4 snRNA

» U

Ki - y 0

U3 Box B/C snoRNA

3' U G N G ^ 59^ U \ C 5 8 . ~A . c • G 53-A,

G 1^ G—u C- G 5' 3'

3' A

3' G

282-A .A-246 261 -A \ ^247 C — G ^248 G-C G—C 5' 3'

L30e pre-mRNA

L7Ae rRNA

15. A • C IS-,, e . u " • G G A - , C N G —C G- C A G A A L7Ae Box C C SRNA

hS3P2 llL5.5kD SnJl3p (*p2p yRPLBS tiRPLFA

liRP V |||i|c s V T I H E G S Q L K Q Q I Q S I l l l l l l F•

V P G3 Nk|iK DG K

lAF-qS/NSE

752 119 IZl 153 105 210

Figure 2. K-tum motifs and amino acid sequence alignments of L7A/L30 RNA-binding domains. (A) The secondary structures of the U4 snRNA, L30e pre-mRNA, L7Ae rRNA, L7Ae box C/D sRNA were taken from the crystal structures of the corresponding RNAprotein complexes, those of the SECIS RNA and U3 snoRNA result from structure probing (see text). In bold are the sheared G.A base pairs. Numbering is from the original publications, except that of the SECIS RNA which is arbitrary. The dotted line between A248 and U265 in L7Ae rRNA represents the hydrogen bond giving rise to the base triple A.U.G [30]. (B) Amino acid sequence alignment of L7A/L30 proteins. hSBP2, human SBP2; hl5.5 kD, human 15.5 kD; Snul3p, the yeast 15.5 kD ortholog; Nhp2p, the yeast core protein of box H/ACA snoRNPs; yRPL30, yeast ribosomal protein L30; hRPL7A, human ribosomal protein L7A. Identical and similar amino acids are displayed in black and gray backgrounds, respectively.

SECIS RNAs and K-turn binding proteins

57

However, the structures in [26] showed that the increased length of the 5' and 3' strands in loop I versus the consensus K-tum is not a major obstacle to Ktum formation as variable lengths do exist in the 5' strands of various Ktums. Regarding the 3' strand, examination of the K-tum crystal structures pointed to the possibility of accommodating its extra length. hi conclusion, we propose that the core of SECIS RNA is a K-tum like motif where the bend occurs at the intemal loop I, providing less structural constraint than canonical K-tums. As a consequence, the SECIS RNA must be endowed with a greater flexibihty enabling it to switch easily from an open to a closed kinked conformation, thus triggering a major conformational change of the SBP2 bound complex [9]. A phylogenetically conserved RNA-protein interface at work for selenoprotein synthesis Proteins containing the L7A/L30 RNA-binding domain include ribosomal proteins L7A (L7Ae in Archaea) and L30, human 15.5kD (Snul3p in yeast) in box H/ACA snoRNPs. Archaea contain neither 15.5 kD nor Nhp2p, L7Ae being the surrogate in box C/D and box H/ACA sRNPs. Crystal stmctures attested to the presence of a K-tum motif in the yeast L30e pre-mRNA, L7Ae rRNA and box C/D sRNA, in addition to U4 snRNA discussed above (Figure 2A). In this series, the only K-tum sequence variant is the L7Ae rRNA where U.G substitutes for the A.G (top) base pair. U.G can nevertheless form base pair 4 described in Figure IB [26,30]. U3 snoRNA contains one B/C and one C'/D box instead of the classical box C/D, both recognized by 15.5 kD/Snul3p; the B/C box stmcture shown was derived from probing experiments [31]. Figure 3A shows views of the crystal smctures of the 15.5 kD-U4 snRNA, L30e-pre-mRNA and L7Ae-box C/D sRNA complexes, adapted from [22-24]. A detailed description of the RNA-protein contacts fall beyond the scope of this review. Inspection of Figure 3A, however, reveals that the three stmctures form analogous protein-RNA interfaces despite the differential orientation of some helices (compare for example the bottom right helix in L30e with the proteins in the other two complexes). The interface is provided by the flipped-out bases U31 (U4 snRNA), U263 and U18 (rRNA and sRNA), A57 (L30e pre-mRNA) protmding into an electrostatically neutral pocket of the cognate protein, and by a few amino acids that make base-specific contacts with the guanines of the sheared G.A base pairs. Yet differences can be found. For example, binding of 15.5 kD is highly susceptible to mutations of U31 while changing A57 and U18 is less deleterious to L30e and L7Ae interaction [22-24]; the angle of the kink shows subtle variations in each complex; finally, it is worth noting that L30e and the pre-mRNA interact through a mutually induced fit [32] whereas only

58

Selenium: Its molecular biology and role in human health

the RNA component (U4 snRNA or sRNA) undergoes an induced fit upon binding to 15.5 kD and L7Ae, respectively [31, 33-37]. Earlier work localized the SBP2 RNA-binding domain in a region lying approximately between positions 500 and 750 [4,7,29, and our unpublished results]. Within this area, database searches [4,6,7] identified a subdomain homologous to the L7A/L30 RNA-binding domain (Figure 2B), SBP2 and 15.5 kD/Snul3p sharing the highest amino acid sequence similarity [7]. The RNA-binding domain of SBP2 is thus bipartite, composed of the conserved L7A/L30 module flanked by SBP2-specific sequences. A structure-guided strategy, based on the similarities between SBP2/15.5 kD and SECIS RNA/U4 snRNA, and the crystal structure of the 15.5 kD-U4 snRNA complex, predicted SBP2 amino acids that should contact the SECIS RNA [7]. Changing them to alanines led to the identification of eight amino acids critical for SECIS binding, four of them being crucial: Gly676 and Glu679 are postulated to contact the guanines of the sheared base pairs, Glu699 and Arg731 being very likely part of the pocket accommodating the SECIS RNA U2 (Figure 3B). These findings established that the recognition principles governing the 15.5 kD-U4 snRNA interaction must be similar in the SBP2SECIS RNA complex especially at the guanines of the G.A base pairs and at U2 (SECIS RNA) and U31 (U4 snRNA). Another member of the L7A/L30 family, the rat ribosomal protein L30, was recently shown to be a novel SECIS-binding protein [9]. As a follow-up, determination of the molecular basis underlying this interaction would be instructive in particular to understand how L30 can recognize the SECIS K-tum and compete with SBP2. In the absence of a structural model though, comparison of the structures of the L30e, 15.5 kD, L7Ae and SBP2 RNA-protein complexes [7, 22-25, 37] provided some clues that may explain the L30 versatility. In all of the complexes, mutations of the bases comprising the sheared G.A base pairs resulted in the complete loss of protein binding in vitro. Together with the high amino acid sequence similarity between yeast L30e and rat L30, these findings strongly suggest that rat L30 in complex with the SECIS RNA should also interact at the G.A tandem of the SECIS RNA, most likely at the guanines. An interesting difference between 15.5 kD and SBP2 on the one hand, and L7Ae and L30e on the other, occurs at the flipped-out base. In the 15.5 kD and SBP2 complexes, mutations of U31 and U2 to any nucleotide is detrimental to binding in vitro and function in vivo (for the SECIS RNA). In contrast, L7Ae can accommodate a C instead of U18 and there is little sequence preference at A57 for the binding in vitro of L30e which can tolerate G or even C [38]. Remarkably, a correlation can be made at the protein level, hi the 15.5kD-U4 snRNA and SBP2-SECIS RNA complexes, five (almost) identical amino acids contact U31 and probably U2, respectively (Figure 3B). Instead, only two L7Ae amino acids make base specific contacts with U18, L30e showing the simplest interaction scheme

SECIS RNAs and K-turn binding proteins

59

with one single contact between A57 and Asn47 (or Asn74, depending on whether the NMR or X-ray structures are considered). Given that G or C may substitute for A57, it is conceivable that U can also fit.

h15.5kD

L30e

U4 snRNA

Ala 39

L7Ae

L30e mRNA

SRNA

Gly38 Asn40

Lys 682 Lys44

h15.5kD-U4snRNA

SBP2-SECIS RNA

Figure 3. RNA-protein interfaces at various L7A/L30 protein-K-tum RNA complexes. (A) Overall crystal structures of the human hi5.5 kD-U4 snRNA, L30e-mRNA and L7Ae-box C/D sRNA complexes adapted from [22-24]. The ribbon plots of the proteins with the bound RNA fragments are shown. Figures were generated with PyMOL in an orientation expliciting structural similarities. (B) Similar interaction principles govern the 15.5 kD-U4 snRNA and SBP2-SECIS RNA complexes [7].

Taking into account that a single contact forms between L30e and A57, we propose that the SECIS RNA U2 could also hydrogen bond with L30e Asn47

60

Selenium: Its molecular biology and role in human health

or Asn74 upon repositioning of the base to offer the appropriate hydrogen bond donor and acceptor groups. As rat L30 binds the SECIS RNA, we assayed other L7A/L30 proteins for their abiUties to recognize the SECIS RNA. Snul3p and L7Ae indeed bound the SECIS RNA but the reverse did not happen since SBP2 was unable to interact with U4 snRNA or an L7Ae RNA target (A.Clery, C.A, A.K and C. Branlant, manuscript in preparation). We concluded from this experiment that the SBP2 RNA-binding domain is more complex than in the other proteins of the family, the SECIS RNA binding specificity being very likely provided by amino acids flanking the L7A/L30 subdomain. In fact, our unpublished data support this hypothesis. The analogous protein-RNA interface formed between L7A/L30 proteins and various K-tum RNAs suggests a conformational adaptability of the RNA upon binding to its cognate protein. Such a dynamic process could potentially confer the binding specificity for different K-tums, as exemplified by rat L30. This adaptability could be facilitated by the dimorphism of Ktum RNAs that are in dynamic equilibrium between a tightly kinked-tum and a more open structure [39]. A large number of proteins within the L7A/L30 family and a large number of diverse RNAs containing the K-tum motif have been identified, with the majority of the K-tums residing in the small and large ribosomal RNAs [8]. Altogether, these findings suggest that the L7A/L30 fold and the K-tum are ancient stmctural motifs that have evolved specialized roles in many different biological processes. References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17.

R Walczak, E Westhof, P Carbon, A Krol 1996 RNA 2:367 R Walczak, P Carbon, A Krol 1998 RNA 4:74 GW Martin III, JW Harney, MJ Berry 1998 RNA 4:65 PR Copeland, JE Fletcher, BA Carlson, DL Hatfield, DM Driscoll 2000 EMBO J19:306 JE Fletcher, PR Copeland, DM Driscoll, A Krol 2001 RNA 7:1442 PR Copeland, DM Driscoll 2001 Selenium: Its Molecular Biology and Role in Human Health DL Hatfield ed Kluwer Academic Publishers Boston/Dordrecht/London pp 55 C Alltnang, P Carbon, A Krol 2002 RNA 8:1308 DJ Klein, TM Schmeing, PB Moore, TA Steitz 2001 EMBO J 20:4214 L Chavatte, BA Brown, DM Driscoll 2005 Nature Struct & Mol Biol 12:408 E Grundner-Culemann, GW Martin III, JW Harney, MJ Berry 1999 RNA 5:625 C Buettner, JW Harney, MJ Berry 1999 J Biol Chem 274:21598 M Hirosawa-Takamori, H Jackie, G Vorbruggen 2000 EMBO Rep \:44\ S Castellano, N Morozova, M Morey, MJ Berry, F Serras, M Corominas, R Guigo 2001 EMBO Rep 2:691 KV Korotkov, SV Novoselov, DL Hatfield, VN Gladyshev 2002 Mol Cell Biol 22:1402 SV Novoselov, M Rao, NV Onoshko, H Zhi, GV Kryukov, Y Xiang, DP Weeks, DL Hatfield, VN Gladyshev 2002 £A/fiO J21:3681 GV Kryukov, S Castellano, SV Novoselov, AV Lobanov, O Zehtab, R Guigo, VN Gladyshev 2003 Science 300:1439 C Thisse, A Degrave, GV Kryukov, VN Gladyshev, S Obrecht-Pflumio, A Krol, B Thisse, A Lescure 2003 Gene Expression Patterns 3:525

SECIS RNAs and K-turn binding proteins 18. 19. 20. 21. 11. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39.

61

S Castellano, SV Novoselov, GV Kryukov, A Lescure, E Blanco, A Krol, VN Gladyshev, R Guigo 2004 EMBO Rep 5:71 T Mourier, A Pain, B Barrell, S Griffiths-Jones 2005 RNA 11:119 K Taskov, C Chappie, GV Kryukov, S Castellano, AV Lobanov, KV Korotkov, R Guigo, VN Gladyshev 2005 Nucleic Acids Res 33:2227 D Fagegaltier, A Lescure, R Walczak, P Carbon, A Krol 2000 Nucleic Acids Res I Vidovic, S Nottrott, K Hartmuth, R Luhrmann, R Ficner 2000 Mol Cell 6:1331 JA Chao, JR Williamson 2004 Structure 12:1165 T Moore, Y Zhang, MO Fenley, H Li 2004 Structure 12:807 T Hamma, AR Ferr6-D'Amar6 2004 Structure 12:893 A Lescoute, NB Leontis, C Massire, E Westhof 2005 Nucleic Acids Res 33:2395 P Nissen, JA Ippolito, N Ban, PB Moore, TA Steitz 2001 Proc Natl Acad Sci USA 98:4899 NB Leontis, E Westhof 2001 RNA 7:499 A Lescure, C Allmang, K Yamada, P Carbon, A Krol 2002 Gene 291:279 N Ban, P Nissen, J Hansen, PB Moore, TA Steitz 2000 Science 289:905 N Marmier-Gourrier, A Clery, V Senty-Segault, B Charpentier, F Schlotter, F Leclerc, R Foumier, C Branlant 2003 RNA 9:821 JA Chao, GS Prasad, SA White, C David Stout, JR Williamson 2003 J Mol Biol 326:999 V Cojocaru, S Nottrott, R Klement, TM Jovin 2005 RNA 11:197 B Turner, SE Melcher, TJ Wilson, DG Norman, DMJ Lilley 2005 RNA 11:1192 AK Wozniak, S Nottrott, E Kuhn-H6lsken, GF Schrdder, H Grubmuller, R Luhrmann, CAM Seidel, F Oesterhelt 2005 RNA 11:1545 C Charron, X Manival, A Cl^ry, V Senty-Segault, B Charpentier, N Marmier-Gourrier, C Branlant, A Aubry 2004 J Mol Biol 342:757 J Suryadi, EJ Tran, E Stuart Maxwell, BA Brown 2005 Biochem 44:9657 SA White, M Hoeger, JJ Schweppe, A Shillingford, V Shipilov, J Zarutskie 2004 RNA 10:369 TA Goody, SE Melcher, DG Norman, DMJ Lilley 2004 RNA 10:254

Chapter 6. SECIS binding proteins and eukaryotic selenoprotein syntliesis Donna M. Driscoll Department of Cell Biology, Lerner Research Institute, Cleveland Clinic Foundation, Cleveland. OH 44195, USA

Paul R. Copeland Department of Molecular Genetics, Microbiology and Immunology, UMDNJ - Robert Wood Johnson Medical School, Piscataway, NJ 08854, USA

Summary: In eukaryotes, the biosynthesis of selenoproteins involves a novel translational mechanism that recodes UGA as selenocysteine (Sec). The decision to incorporate Sec rather than terminate protein synthesis involves at least two RNA-binding proteins that interact specifically with a conserved motif within the 3' untranslated region (3' UTR) of the selenoprotein mRNA. The structures, functions, and interactions of these two proteins, SECIS Binding Protein 2 (SBP2) and ribosomal protein L30 (L30), are considered in detail in this chapter. Introduction Selenocysteine (Sec) is co-translationally incorporated into selenoproteins in response to a UGA codon in the mRNA. Since UGA is normally read as a stop codon, there are unique features of selenoprotein mRNA translation that distinguish it from normal protein synthesis. The recoding of UGA as Sec depends on cw-acting sequences in the selenoprotein mRNA and novel /ra«5-acting proteins to deliver the Sec-tRNA^^*''^^'' to the ribosome [1-3]. The eukaryotic pathway for Sec insertion is more complex than that utilized in bacteria since it requires a m-acting element in the 3' untranslated region (3' UTR) of the selenoprotein mRNA [4]. This Sec Insertion Sequence (SECIS) element consists of a highly conserved stem-loop structure that contains two short conserved motifs: a quartet of non-Watson-Crick base pairs comprising the SECIS core and an AAR motif found in a single stranded bulge or apical loop [5-7]. Since the discovery that the SECIS core and AAR motif are required for Sec insertion, investigators have hypothesized that the SECIS functions as a platform for recruiting RNAbinding proteins that recognize these conserved motifs. This chapter will focus on two SECIS-binding proteins, SECIS Binding Protein 2 (SBP2) and

64

Selenium: Its molecular biology and role in human health

ribosomal protein L30 (L30), both of which bind to the SECIS core and whose importance in UGA recoding has been established. SBP2 The discovery of the eukaryotic SECIS element in 1991 [4] naturally led to a search for cognate binding factors. Such a search resulted in the identification of SECIS-binding protein 2 (SBP2) derived from rat testicular extracts, a 94 kDa protein that specifically interacts with the conserved SECIS core [8,9]. SBP2 is an essential component of the UGA recoding machinery and a limiting factor for Sec incorporation in vitro and in vivo [10]. Subsequent structure/function and primary sequence analyses revealed that SBP2 is comprised of three major subdomains as illustrated in Figure 1 [11]. The N-terminal domain, which consists of amino acids (aa) 1-399 (numbering for rat SBP2), is dispensible for UGA recoding in vitro and in transfected cells. The functional domain (aa 399-517) is required for Sec incorporation in vitro but not for RNA-binding activity. Immediately downstream of this region is the RNA-binding domain (aa 517-777), which is required for interaction with the SECIS element. The C-terminal region of SBP2 (aa 777-846) has no known function. Together the functional and RNA-binding domains account for all of the known activities associated with SBP2: Sec incorporation in vitro, SECIS binding, and ribosome binding.

Putative Regulatory Domain

RNA Binding Domain

Fully functional C-terminal fragment

i Sec incorporation in vitro SECIS element binding Ribosome binding

Figure 1. Diagram of domain structure for rat SBP2. The C-terminal 447 amino acids are sufficient to carry out all known functions for SBP2 as indicated.

The N-terminal domain Recent studies have demonstrated that full-length SBP2 and the C-terminal half of the protein (aa 399-846) are indistinguishable in their abilities to

SECIS-bindingproteins: SBP2 and L30

65

incorporate Sec in vitro [12,13]. Interestingly, the N-terminal segment of the protein is found only in higher eukaryotes (primates, rodents and fish) and appears to be correlated with selenoproteome complexity. That is, the more selenoproteins a given genome encodes, the more likely it is to have the larger version of SBP2. This may suggest a selenoprotein-specific regulatory function for this domain, but there are currently no data to support this hypothesis. The N-terminal domain does possess a predicted nuclear localization signal suggesting that nuclear localization may regulate SBP2 function. The current published data, however, show no detectable nuclear localization of SBP2 in transiently transfected hepatoma cells [12], but further work in a variety of cell types and under a variety of conditions will be needed to thoroughly address this issue. Unfortunately, the sequence of the N-terminal domain of SBP2 does not offer any other clues for function because it is apparently unique to the SBP2 family. Determining the function of this domain will require long-term biochemical and/or genetic studies. The functional domain Initial structure/function analysis of SBP2 demonstrated that amino acids 399-517 are required for Sec incorporation but not for SECIS binding [11]. The sequence of this functional domain is, again, unique to the SBP2 family and contains several stretches of sequence that are conserved from protists to mammals. In addition to binding to the SECIS element, SBP2 has also been shown to associate with ribosomes. This interaction is reduced or abolished by mutations in the SECIS-binding domain of SBP2. Interestingly, truncations within the functional domain also result in significantly reduced ribosome binding activity, indicating that both the SECIS binding domain and the functional domain are required for this function. Further mutagenesis analysis has shown that one of these conserved stretches consisting of basic residues (KKGK506-510) is essential for Sec incorporation activity (P.R. Copeland, unpublished data). However, mutations within the KKGK motif do not affect the protein's ability to bind to the ribosome or to the SECIS element. Thus, the precise role of this domain in UGA recoding remains unknown. The SECIS binding domain Within the SECIS binding domain is an L7Ae RNA binding motif found in several proteins, including the ribosomal protein L30 (L30) and the translation termination factor, eukaryotic release factor 1 (eRFl). Since SBP2 was shown to bind the ribosome via 28S rRNA [11], it has been proposed that this interaction may also utilize the L7Ae motif to interact with a kink-turn motif on the ribosome (see below). Indeed, the SECIS element has recently been shown to compete with SBP2 for ribosome binding [13]. Within the L7Ae motif, only one residue is absolutely conserved in all family

66

Selenium: Its molecular biology and role in human health

members (G669 in SBP2). When this residue was changed to arginine, SBP2 lost the abihty to incorporate Sec and bind to the SECIS element, but it still retained ribosome binding activity [11]. Since truncations within the SECIS binding domain prevent ribosome binding, it was concluded that the SECIS binding domain and ribosome binding domain are overlapping but not identical. The existing structural data regarding the interaction of several L7Ae RNA binding motifs with their kink-turn target RNAs provide a fairly clear picture of how SBP2 might physically interact with both the SECIS element and the ribosome [14-16]. However, it is important to note that the RNA-binding domain of SBP2 is 250 amino acids long, which is considerably larger than the core L7Ae motif Thus additional sequences are required for the SECIS-binding activity of SBP2 and the determinants for sequence specificity remain unknown. More extensive mutagenesis within the L7Ae motif in human SBP2 has shown that not all of the conserved residues are required for SECIS binding, perhaps strengthening the argument for dual specificity (i.e., SECIS element and rRNA). This concept is certainly not unprecedented as several proteins bearing the L7Ae motif have been shown to interact with more than one type of kink-turn [16-18]. SBP2 interactions: self association. Experiments designed to study the SBP2:ribosome interaction revealed that recombinant SBP2 purified fi-om E. coli formed a salt-sensitive complex, suggesting self-association [11]. The fact that SBP2 has only one canonical RNA binding domain but is known to interact with both the SECIS element and rRNA led us to propose a model in which a head-to-head dimer configuration SBP2 would have two RNA binding domains available to simultaneously interact with both components and provide a means for communication between the ribosome and SECIS elements downstream in the 3'UTR [19]. However, recent studies employing a combination of gel filtration and pull-down assays showed that SBP2 does not self-associate [13]. In fact, the complex formation that was previously observed was shown to be a function of the N-terminal Strep-tag used in the initial work. Interestingly, it was also determined that SBP2 could not simultaneously interact with the ribosome and the SECIS element in vitro. The presence of a 2.5-fold excess of a wild type SECIS element, but not a mutant lacking the SECIS core, was able to exclude a significant fi-action of SBP2 fi-om the ribosome [13]. Taking these recent results into account, a new model for the role of SBP2 in Sec incorporation must be devised as discussed below. SBP2 interactions: eEFSec One of the predicted functions of SBP2 is to directly or indirectly link the SECIS element to eEFSec, the elongation factor that is dedicated to delivering Sec-tRNA^^"'^*^ to the ribosome. The initial characterization of

SECIS-binding proteins: SBP2 and L30

67

eEFSec showed that it could be co-immunoprecipitated with SBP2 when both factors were over-expressed in transiently transfected cells [20]. Interestingly, this complex was almost completely abrogated by treatment of the extracts with RNAse A, strongly suggesting an indirect interaction which is mediated through RNA. More recent work makes the argument that SBP2 and eEFSec directly interact via the eEFSec C-terminal extension, but this could only be shown to be direct when eEFSec was severely truncated [21]. Further experiments in transfected cells showed that SBP2/eEFSec coimmunoprecipitation was enhanced when the tRNA^^^"^^^" was co-transfected, suggesting the formation of a large macromolecular complex [21]. However, it seems clear that the full-length SBP2 and eEFSec proteins do not directly interact in vitro, and the precise nature of their association in vivo remains unknown. SBP2 interactions: tlie ribosome A potentially direct role for SBP2 in the Sec decoding process is best supported by the observation that it is stably associated with ribosomes in vitro and in vivo [11,13]. Intriguingly, both the functional domain and the SECIS binding domain are required for ribosome binding [11], suggesting that more than one binding site may be involved. One model for the action of SBP2 function puts the SBP2/ribosome interaction as dominant over the SBP2/SECIS element interaction such that the latter only occurs during UGA recoding, perhaps involving an SBP2-dependent retardation of the termination reaction. The signal for Sec incorporation in this case would be a complete or partial shift of SBP2 binding from the ribosome to the SECIS element, perhaps yielding an A-site conformational change that preferentially accommodates eEFSec/Sec-tRNA'^''^'"' at the expense of eRFl. Formal proof that ribosome binding is required for Sec incorporation awaits the identification of the appropriate mutant version of SBP2, and the inclusion of a ribosomal protein, L30, in this process (see below) adds significant complexity to this model. In addition, the potential functions of additional, as-yet unidentified factors could be critical since the known players have not yet been demonstrated to be sufficient for Sec incorporation. Ribosomal Protein L30 The discovery that SBP2 and eEFsec are SECIS-binding and tRNA-binding proteins, respectively, was a surprise since most investigators expected that these activities would be carried out by a single eukaryotic homolog of SelB. These studies significantly advanced the field but they did not provide insight into how the SECIS in the 3' UTR is recruited to the UGA/Sec codon at the upstream ribosome. Unlike the ribosome-tethered mechanism of UGA recoding in prokaryotes, the eukaryotic machinery has to solve the problem of the separation of the SECIS and the UGA/Sec codon, which can be

68

Selenium: Its molecular biology and role in human health

hundreds or even thousands of nucleotides apart [22]. This mechanistic difference between the prokaryotic and eukaryotic Sec insertion machinery suggests that a unique component might have evolved in the eukaryotic pathway to mediate interactions between the SECIS element and the ribosome. While SBP2 may play a part in this process [13], Chavatte and colleagues recently showed that ribosomal protein L30, which has no prokaryotic ortholog, is a component of the eukaryotic recoding machinery [18]. L30 was identified in UV crosslinking assays as a SECIS-binding activity, which requires the SECIS core for binding. The protein was purified by RNA affinity chromatography using a wild-type SECIS element as the ligand. Based on mass spectrometry analysis of the purified protein, the crosslinking activity was identified as L30 [18]. Previously known functions L30 has been called an RNA-binding chameleon because it binds to multiple targets. The protein auto-regulates its own expression by binding to its premRNA in the nucleus to inhibit splicing and to the mature transcript in the cytoplasm to silence translation [23-25]. L30 is also an essential component of the large ribosomal subunit in eukaryotes [26] but little is known about its function in protein synthesis. The recent cryo-electron microscopy map of the 80S wheat germ ribosome localized L30 to the interface between the large and small ribosomal subunits [27]. L30 interacts with the large subunit by binding to kink turn motifs in both helix 34 and helix 58 of the 28S rRNA [27]. How could L30 bind simultaneously to the ribosome and to the SECIS? One possibility is that L30 may have two RNA-binding interfaces, one interacting with the rRNA and the other with the SECIS element. Even if L30 has only one RNA-binding interface, one can envision a model in which ribosome-associated L30 leaves and binds to the SECIS due to differences in affinity. Alternatively, it may be that fi-ee L30, not ribosome-associated L30, is the active pool for UGA recoding. The ability of L30 to bind to multiple targets may be due to its flexibility in solution, as discussed below. Involvement in UGA recoding Multiple lines of evidence support the hypothesis that L30 is involved in UGA recoding [18]. Purified recombinant L30 binds to SECIS elements in vitro, and the endogenous protein is bound to selenoprotein mRNAs in vivo. Furthermore, overexpression of L30 in transfected cells stimulated UGA recoding activity. It is important to note that direct evidence for an essential function of L30 in Sec incorporation in vitro or in vivo is still lacking. Experiments to deplete L30 using siRNA or gene targeting strategies are likely to be difficult to perform since L30 is an essential protein in yeast and required for normal protein synthesis. As a component of the large ribosome subunit, L30 is ubiquitously expressed. The protein is also found in multiple

SECIS-binding proteins: SBP2 and L30

69

cellular compartments, including the nucleus, nucleolus, and cytoplasm [28]. To date, it is not known which cellular pool of L30 is active in UGA receding. RNA-binding properties In hindsight, the identification of L30 as a SECIS-binding protein might have been anticipated since L30 and SBP2 both belong to the L7Ae family of proteins. In fact, it was noted several years ago that the RNA-binding domain of SBP2 shares significant homology with L30 [10]. The L7Ae family, which is named after the archaeal ribosomal protein L7Ae, also includes ribosomal protein SI2, 15.5 kDa protein, and eukaryotic release factor-1 [29]. Several members of this family are known to bind to a kink-turn structure, a relatively common motif found in rRNAs, snRNAs, snoRNAs, and mRNAs [15,30]. RNAs containing a kink-turn can undergo a conformational transition from an open to kinked form [31,32]. This conformational transition can be induced by increasing magnesium concentrations in vitro or possibly by protein binding in vivo. Kink-turns differ in their sequence, secondary structure, angle of bending, and context, which may partly explain why L7Ae family members show different RNAbinding specificities. Although SBP2 and L30 bound to the SECIS element in vitro, two other L7Ae family members did not [18]. Interestingly, L7Ae itself has SECIS-binding activity but does not stimulate UGA recoding [18]. Thus, as has been previously shown for SBP2 [11], the ability of a protein to bind to the SECIS is not sufficient to enhance Sec insertion. RNA-binding domain The RNA-binding domain of L30 is highly conserved [28,33]. The structure of this protein complexed with a known target, the L30 mRNA, has been determined by NMR [17] and X-ray crystallography [33]. Both the protein and RNA undergo conformational changes upon interaction, implying a partially induced fit binding mechanism. L30, which is folded into a threelayer a/p/a sandwich, interacts with the RNA through the loops at the end of the sandwich. An important direction for future research is to resolve the structure of the L30:SECIS complex and identify the molecular determinants of this interaction. Conformation-specific SECIS-binding activities of L30 and SBP2 The fact that L30 and SBP2 have overlapping binding sites on the SECIS suggested the possibility that the two proteins may bind simultaneously to adjacent sites or competitively to the same site. However, Chavatte et al. showed that SBP2 and L30 do not bind to the same RNA molecule but instead compete for binding to the SECIS in a magnesium-dependent manner [18]. The effect of magnesium is likely due to its ability to induce a kink-turn

70

Selenium: Its molecular biology and role in human health

in the SECIS molecule. Based on in vitro binding experiments, SBP2 and L30 differ in their ability to recognize the proposed different conformers of the SECIS, with SBP2 binding preferentially to the open conformer and L30 binding to and possibly stabilizing the kinked form. An even more surprising result was obtained when competition experiments were performed with preformed complexes. In spite of having a much lower affinity for the SECIS, L30 could rapidly displace SBP2 from a preformed SBP2:SECIS complex. Likewise, an L30: SECIS complex was efficiently displaced by SBP2. Based on these results, Chavatte et al. proposed that the SECIS element acts as a molecular switch and that the dynamic exchange between L30 and SBP2 for the SECIS depends on the composition of the preformed complex [18]. These results provide a potential explanation for how SBP2 and L30 could bind sequentially to the same site during the UGA recoding process. Molecular basis for different RNA-binding specificities Since SBP2 and L30 share a similar secondary structure, why do these two proteins show differential binding to the proposed open and kinked conformers? There are several possible explanations. L30 and SBP2 have a number of amino acid differences in regions that are predicted to interact with RNA [18]. Secondly, mutagenesis studies showed that the L30-like domain in SBP2 is necessary but not sufficient for SECIS-binding, indicating that additional sequences in SBP2 are required for its binding activity [11]. Furthermore, L30 is a flexible protein in solution [17], whereas the larger RNA-binding domain of SBP2 may be more rigid. Intriguingly, replacement of the L30-like domain in SBP2 with the L30 coding sequence resulted in a protein that was completely defective in SECIS-binding (L. Chavatte and D. Driscoll, unpublished results), supporting the hypothesis that the two domains are not functionally equivalent. Mechanism of UGA recoding in eukaryotes Previous models of eukaryotic Sec insertion were based on the assumption that SBP2 remained stably bound to the SECIS through multiple rounds of translation. It was proposed that the SBP2-SECIS complex then interacted with eEFsec to recruit Sec-tRNA^^"^^''' to the ribosome. The recent findings discussed in this chapter suggest that the eukaryotic mechanism of Sec incorporation is considerably more complicated. The bulk of SBP2 in mammalian cells may be ribosome-associated rather than SECIS-bound [13]. In addition, the involvement of L30 in UGA recoding adds another layer of complexity since this protein also has ribosome-binding and SECIS-binding activities. The SECIS element appears to play a dynamic role in UGA recoding, perhaps by undergoing a conformational transition from an open to kinked form. Binding of L30 may trigger this fransition. Finally, SBP2 and

SECIS-bindingproteins: SBP2 and L30

71

L30 have the abiUty to displace each other from a pre-formed protein: SECIS complex [18]. Taken together, these results suggest that a two-step mechanism in which L30 and SBP2 (or vice versa) bind and act sequentially during UGA recoding to recruit eEFsec and deliver Sec-tRNA'^^'^^^^'' to the ribosomal A site. Although the details of the mechanism have not been elucidated, one obvious conclusion from these studies is that the multiple functions of prokaryotic SelB are carried out by at least three eukaryotic proteins. Conclusion The last few years have seen rapid advances in our understanding of the basal machinery that incorporates Sec into eukaryotic selenoproteins but many unanswered questions still remain. The exact functions of SBP2 and L30 in the recoding mechanism have not been defined. Little is known about the identity of the active cellular pools of SBP2 and L30, or the order of events during UGA recoding. When and where do SBP2 and L30 bind the SECIS element? Are either of these proteins stably associated with the SECIS or is the SECIS :protein interaction transient? What is the stoichiometry of SBP2, L30, and selenoprotein mRNAs in mammalian cells? The mechanism by which the eEFsec/Sec-tRNA'^*'^'^''' complex is recruited to the ribosome is also not understood. Although SBP2 and eEFsec have been shown to be in a complex in transfected cells, this interaction appears to be RNA-dependent rather than direct. Finally, our understanding of the UGA recoding mechanism will remain limited until the role of the highly conserved and essential AAR motif within the SECIS element is elucidated. References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16.

DM DriscoU, PR Copeland 2003 Annu Rev Nutr 23:17 DL Hatfield (ed) 2001 Selenium: Its molecular biology and role in human health Kluwer Academic Publishers, Boston DL Hatfield, VN Gladyshev 2002 Mol Cell Biol 22:3565 MJ Berry, L Banu, YY Chen, SJ Mandel, JD Kieffer, JW Harney, PR Larsen 1991 Nature 353:273 R Walczak, E Westhof, P Carbon, A Krol 1996 RNA 2:367 D Fagegaltier, A Lescure, R Walczak, P Carbon, A Krol 2000 Nucleic Acids Res 28:2679 E Grundner-Culemann, GW Martin, 3rd, JW Harney, MJ Berry 1999 RNA 5:625 A Lesoon, A Mehta, R Singh, GM Chisolm, DM Driscoll 1997 Mol Cell Biol 17:1977 PR Copeland, DM Driscoll 1999 J Biol Chem 274:25447 PR Copeland, JE Fletcher, BA Carlson, DL Hatfield, DM Driscoll 2000 EMBO J 19:306 PR Copeland, VA Stepanik, DM Driscoll 2001 Mol Cell Biol 21:1491 A Mehta, CM Rebsch, S A Kinzy, JE Fletcher, PR Copeland 2004 J Biol Chem 37852 SA Kinzy, K Caban, PR Copeland 2005 Nucleic Acids Res 33:5172 I Vidovic, S Nottrott, K Hartmuth, R Luhrmann, R Ficner 2000 Mol Cell 6:1331 DJ Klein, TM Schmeing, PB Moore, TA Steitz 2001 EMBOJ20A2\4 J Suryadi, EJ Tran, ES Maxwell, BA Brown, 2nd 2005 Biochemistry (Mosc) 44:9657

72

Selenium: Its molecular biology and role in human health

17. 18. 19.

H Mao, SA White, JR Williamson 1999 Nat Struct Biol 6:1139 L Chavatte, BA Brown, DM Driscoll 2005 Nat Struct Mol Biol 12:408 PR Copeland, DM Driscoll 2001 Selenium: Its Molecular Biology and Role in Human Health DL Hatfield (ed) Kluwer Academic Publishers, Boston pp 55 RM Tujebajeva, PR Copeland, X-M Xu, BA Carlson, JW Harney, DM Driscoll, DL Hatfield, MJ Berry 2000 EMBO Reports 1:1 AM Zavacki, JB Mansell, M Chung, B Klimovitsky, JW Harney, MJ Berry 2003 Mol Cell 11:773 GW Martin, MJ Berry 2001 Selenium: Its Molecular Biology and Role in Human Health DL Hatfield (ed) Kluwer Academic Publishers, Boston pp 45 J Vilardell, JR Warner 1997 Mol Cell Biol 17:1959 B Li, J Vilardell, JR Warner 1996 Proc Natl Acad Sci USA 93:1596 MD Dabeva, JR Warner 1993 J Biol Chem 268:19669 MD Dabeva, JR Warner \987 J Biol Chem 262:16055 M Halic, T Becker, J Frank, CM Spahn, R Beckmann 2005 Nat Struct Mol Biol 12:467 J Vilardell, SJ Yu, JR Warner 2000 Mol Cell 5:761 EV Koonin, P Bork, C Sander 1994 Nucleic Acids Res 22:2166 WC Winkler, FJ Grundy, BA Murphy, TM Henkin 2001 RNA 7:1165 TA Goody, SE Melcher, DG Norman, DM Lilley 2004 RNA 10:254 S Matsumura, Y Ikawa, T Inoue 2003 Nucleic Acids Res 31:5544 JA Chao, JR Williamson 2004 Structure (Camb) 12:1165

20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33.

Chapter 7. The importance of subcellular localization of SBP2 and EFsec for selenoprotein synthesis. Peter R. Hoffmann and Maria J. Berry Department of Cell and Molecular Biology, John A. Bums School of Medicine, University of Hawaii at Manoa, Honolulu HI 96813 USA

Summary: Receding of UGA from a stop codon to selenocysteine poses a dilemma for the protein translation machinery. The presence of a UGA codon within the coding region of selenoprotein mRNAs render them susceptible to degradation by nonsense-mediated decay (NMD) immediately after export to the cytoplasm. To avoid NMD and assure efficient translation of selenoprotein mRNA, assembly of the selenocysteine incorporation factors, including Selenocysteine Binding Protein 2 (SBP2) and selenocysteine-specific elongation factor (EFsec), onto the mRNA most likely occurs in the nucleus, at least for some selenoprotein mRNAs. The amino acid sequences of SBP2 and EFsec both contain functional nuclear localization signals (NLSs) and nuclear export signals (NESs). hi fact, several lines of evidence suggest that these proteins are capable of shuttling between the cytoplasm and nucleus. SBP2 contains multiple NLSs and NESs and may be particularly important for final export of a stable protein/mRNA complex from the nucleus to the cytoplasm for franslation. Li this context, the subcellular location of these important components of the selenocysteine incorporation machinery is as essential for proper translation and may prove to be an important regulatory step in selenoprotein synthesis. Introduction Selenium is believed to exert its biological effects through its incorporation into selenoproteins as the amino acid, selenocysteine. The human selenoprotein family contains at least 25 members that exhibit a wide range of functions, including thyroid hormone metabolism, glucose metabolism, intra- and extra-cellular antioxidation, redox regulation, selenium transport, and sperm maturation and antioxidant protection [1]. Selenoprotein mRNAs are unusual in that they contain one or more UGA codons within their coding regions. The UGA codon typically signals termination to the protein synthesis machinery. However, UGA in selenoprotein mRNA is recoded from a termination codon to a selenocysteine insertion codon. This recoding process requires assembly of dedicated factors at specific secondary

74

Selenium: Its molecular biology and role in human health

structures in the 3' untranslated regions (UTRs) of these mRNAs [2], termed SECIS elements. These factors include SBP2 [3] and EFsec [4,5] complexed with selenocysteyl-tRNA (sec-tRNA^'', reviewed in [6]). The sequence of steps in formation of the mRNA (SECIS)/sec-tRNA^7SBP2/EFsec complex has been characterized through work from several different laboratories. SBP2 is thought to first bind to the SECIS element in the 3' UTR of selenoprotein mRNAs [7]. Binding of Sec-tRNA^" by EFsec promotes interaction between the two factors, EFsec and SBP2 [8]. Binding of EFsec and SBP2 allows insertion of selenocysteine and, thus, recoding of the UGA codon. SBP2 may remain bound to the SECIS element and promote serial deliveries of Sec-tRNA^" through multiple rounds of translation, thus promoting efficient synthesis of selenoproteins. In fact, this process is most likely required in the case of SelP, in which 10-17 selenocysteines are incorporated per polypeptide. Other protein factors have been identified that may play important roles in selenoprotein synthesis, but this chapter will focus on the factors, SBP2 and EFsec. Disrupting the functions of the factors involved in selenocysteine incorporation, or limiting selenium availability results in premature termination of translation at UGA codons [3,9-13]. Under these circumstances, selenoprotein mRNAs are susceptible to degradation through NMD [14], with different mRNAs exhibiting varying sensitivities to the NMD pathway [15,16]. In order for selenoprotein mRNA to avoid degradation via NMD, proper assembly of the translational components must take place. When one considers the multiple steps that are involved in assembly of this protein/mRNA complex, it becomes clear that the subcellular location of the various factors plays a crucial role in the synthesis of selenoproteins. That is, to avoid degradation of selenoprotein mRNAs, it matters not only that the protein factors like SBP2 and EFsec be present in proper amounts, but in the proper place as well. Nonsense mediated decay and selenoproteins Also known as mRNA surveillance, NMD is a quality-control process in eukaryotes that acts to degrade mRNAs with premature termination codons [17]. Through this process, mRNAs are degraded that might otherwise encode truncated proteins that interfere with normal cellular processes. While the focus of our discussion will be on the role of NMD in preventing production of truncated protein products, it is worth mentioning that NMD may also play a speciahzed role in regulating other targets including pseudogene products, non-coding RNAs, and mRNA encoding proteins involved in amino acid metabolism [18]. Undoubtedly, NMD has evolved to deal with the premature termination codon-containing transcripts arising from routine abnormalities in gene expression. For example, approximately 55% of human pre-mRNAs are

Subcellular localization ofSBP2 and EFsec

75

targets of alternative splicing and add tremendous diversity to the human proteome [19]. However, inefficient or inaccurate intron removal during splicing may give rise to a reading frame shift that results in premature termination codons. Thus, organisms from yeast to man have evolved NMD pathways to deal with premature termination codons in a rapid, efficient manner. To this end, NMD has been postulated to occur during the "pioneer" round of translation that occurs concurrently with export of the 5' end of the mRNA through the nuclear pore [20]. Selenoprotein mRNAs undergo NMD when selenocysteine incorporation levels are reduced, such as when selenium is limiting [15,16]. Selenocysteine incorporation must occur with high efficiency during this crucial proofreading round of translation, or degradation of the corresponding mRNA would likely ensue. Interestingly, selenium-deficient conditions result in degradation of different selenoprotein mRNA at unequal rates. Several studies show that within the glutathione peroxidase (GPX) family of proteins (GPX-1 through -4) selenium deficiency leads to degradation of GPX-1 and GPX-3 mRNAs before GPX-2 and GPX-4 mRNAs [21-23]. Other studies have supported a similar notion that GPX-1 mRNA levels are decreased by selenium deficiency while other selenoprotein mRNAs are relatively unchanged [24,25]. As described above, recoding of the UGA codon to a selenocysteine insertion signal would require that the factors, SBP2 and EFsec/sec-tRNA, be assembled onto the mRNA to form the "recoding complex". To circumvent NMD and promote efficient selenocysteine incorporation, it follows that it would be advantageous if complex assembly occurred concurrent with or early after franscription of selenoprotein mRNAs. This would require that the protein factors, SBP2 and EFsec, which are synthesized in the cytoplasm, translocate independently or together into the nucleus. Once assembled, the protein-mRNA complex must then franslocate back to the cytoplasm for selenoprotein translation to proceed. The process of nucleocytoplasmic shuttling and the determinants contained within SBP2 and EFsec proteins that guide the shuttling are likely to be crucial elements of the selenoprotein synthesis process. Localization signals within SBP2 and EFsec sequences The fransport of proteins and RNA into and out of the nucleus occurs through the nuclear pore complex (NPC), a large macromolecular structure embedded in the double membrane of the nuclear envelope [26]. Cargo is transported through the nuclear pore via shuttling carrier proteins, including the importins and exportins, as well as other classes of transporters. Motifs recognized by these carrier proteins, signaling import into and export out of the nucleus, are referred to as nuclear localization signals (NLS) and nuclear export signals (NES), respectively. When amino acid sequences for SBP2 and EFsec are analyzed for the presence of NLSs and NESs, both types of

76

Selenium: Its molecular biology and role in human health

motifs are identifiable in both proteins. As examples, the sequences of murine EFsec and rat SBP2 contain putative signals as indicated in Figure 1. A putative NES is present in the amino-terminal elongation factor domain of the murine and human EFsec sequences, and is conserved in the sequences from D. melanogaster, A. gambia, C.elegans, and C. intestinalis, but not from archaea. A putative NLS is present in the carboxy-terminal region of EFsec, overlapping the region previously identified as the SBP2 interaction domain [10]. This sequence is also conserved amongst human, murine, A. gambia, and C. intestinalis sequences.

iFsec NES

Etongation factor domam

" fl=NLS 448 583 SBPl-interaction domain

| = NiS 1

S1P2 f|=NLS 399 517 777 S46 Transactiration SECIS RNA domain binding domain

Figure 1. Schematic models of EFsec and SPB2 with NES (black) and NLS (light grey) domains indicated as vertical bars. A, Mouse EFsec, with elongation factor domain in light grey, and SBP2-interaction domain in dark grey; B, Rat SBP2, with transactivation domain in light grey, and SECTS RNA binding domain in dark grey. Five putative NLSs are present in the rat SBP2 sequence (Figure 1), all of these being conserved in the murine sequence. Three putative NLSs are present in human and A. gambia sequences, and two in the D. melanogaster sequence. The rat and mouse SBP2 proteins are predicted to exhibit 74% nuclear localization, the human and D. melanogaster 65%, and the A. gambia protein 52%. hi the rat and mouse sequences, two and three NLSs, respectively, are located in the amino-terminal half of the protein previously shown to be dispensable for selenocysteine incorporation in vitro [27]. Two of these are localized in the transactivation domain (see below), and the last one lies at the junction between the transactivation and SECIS binding domains. Literestingly, the N-terminal domain is not present in the D. melanogaster and A. gambia sequences. Four NESs are predicted in the rat and mouse SBP2 sequences, three in the amino-terminal and one in the carboxy-terminal region. Of these, three are conserved in the human sequence, whereas only the C-terminal putative NES is present in the Nterminally truncated D. melanogaster and A. gambia sequences.

Subcellular localization ofSBP2 and EFsec

11

Functional analyses of NLS and NES in SBP2 and EFsec Studies have been conducted in our laboratory to determine the subcellular localization of EFsec and SBP2 in different cell lines and conditions. These studies have utilized a variety of cell lines to investigate localization of endogenous SBP2 as well as various epitope-tagged and green fluorescent protein-fused versions of EFsec and SBP2. Immunofluorescent microscopy and western blot analyses both confirm that EFsec and SBP2 are present in both the cytoplasm and nucleus (unpublished results). Furthermore, SBP2 may localize to the nucleolus, a compartment enriched in rRNA and ribosomal proteins. It is tempting to speculate that this may allow interactions to take place between SBP2 and ribosomal proteins or RNAs. Localization of EFsec and SBP2 and expression levels of SBP2 in different cell lines revealed two interesting findings. First, EFsec localization correlated with both the endogenous levels and localization of SBP2, being predominantly cytoplasmic in cells with undetectable levels of SBP2 (HEK293), but colocalizing with SBP2 when levels of the latter are significant (MST0-211H and HepG2). Second, expression levels of SBP2 in the three cell lines correlated with previously reported selenoprotein expression levels. HEK-293 cells express low levels of the glutathione peroxidases (GPXs) and thioredoxin reductases (TRs), and undetectable levels of the iodothyronine deiodinases and selenoprotein P ([28,29] and unpublished results). MSTO21IH cells, a human mesothelioma tumor-derived cell line, express high levels of the mRNA for type 2 iodothyronine deiodinase [30], and HepG2 cells endogenously express iodothyronine deiodinases, selenoprotein P, and the GPXs and TRs (our unpublished results). Taken together, these results suggest that the levels of SBP2 may be a significant determinant for both EFsec localization and selenoprotein synthesis efficiency. SBP2 is a relatively large molecule (94 kDa) that must carry out both protein-mRNA and protein-protein interactions. Several studies have focused on the functional domains within this intriguing protein [27,31]. For instance, the SECIS RNA binding domain has been mapped to amino acids 517-777 in carboxy-terminal region. The transactivation domain, required for selenocysteine incorporation in vitro, was mapped to the middle region of the protein, consisting of amino acids 399-517 (Figure 1). The segment of protein containing both regions, SBP2 399-777, constitutes the minimal functional domain. Until recently, no fianction has been identified for the Nterminal region of this protein. Due to the presence of multiple N-terminal NESs, our laboratory hypothesized that this region of SBP2 may be crucial for nuclear export of SBP2fi"omthe nucleus into the cytoplasm. In fact, our data suggest this is the case (unpublished results).

78

Selenium: Its molecular biology and role in human health

Nucleocytoplasmic shuttling and selenoprotein synthesis The implication that EFsec and SBP2 have the capability to shuttle into and out of the nucleus allow speculation how the translocation of these proteins fits into the scheme of selenoprotein synthesis. In addition to whether the proteins themselves translocate between the cytoplasm and nucleus, one must consider whether they shuttle as a complex. Because the binding domains of each protein have been mapped, one may also consider how binding to mRNA or other proteins may affect the shuttling that is observed with the over-expressed versions of the individual proteins themselves. As an example, binding of SBP2 to the SECIS in selenoprotein mRNA may mask one of the NLSs which is located in the SECIS-binding region. A schematic model that illustrates a possible series of binding and shuttling events is shown in Figure 2. Sec-tRNA^'' binds to the N-terminal portion of EFsec in the cytoplasm, possibly resulting in masking of the NES in the elongation factor domain. The complex of sec-tRNA^'^/EFsec is imported into the nucleus, likely via EFsec's C-terminal NLS. SBP2 shuttles between the cytoplasm and nucleus via its NLSs and NESs. SBP2 binds to SECIS elements of selenoprotein mRNAs in the nucleus, whereby the NLSs would be masked but also no longer needed. Shuttling of SBP2 in and out of the nucleolus may also take place either prior to or after interacting with the SECIS element and/or other factors. Interaction between SBP2 and the sectRNA^/EFsec complex could occur either in the cytoplasm or nucleus. If this occurs in the cytoplasm, the NLSs of SBP2 would still be exposed for import of the sec-tRNA^7EFsec/SBP2 complex into the nucleus and binding of selenoprotein mRNAs. Finally, either the mRNA export pathway or the NES of SBP2 would ultimately be responsible for export of the complex into the cytoplasm for binding to ribosomes and translation of selenoproteins. In this sense, selenoprotein synthesis is dependent upon not only which factors are present, but where they are located as well. Levels and localization of SBP2 as regulatory steps in selenoprotein synthesis If SBP2 levels affect the shuttling capacity of the mRNA/protein complexes involved in selenoprotein synthesis, then one must consider expression levels of SBP2 as an important point of selenoprotein synthesis. As mentioned above, numerous studies have reported on the differential sensitivity of different selenoprotein mRNAs to degradation upon selenium depletion. The decline in GPXl mRNA upon selenium depletion has been shown to exhibit all the hallmarks of NMD [32], serving to down-regulate the levels of selenoprotein mRNAs when selenium is not abundant. In another study, type 1 iodothyronine deiodinase and selenoprotein P mRNAs exhibited vastly differing levels of decline in liver compared with kidney of selenium depleted rats [33]. TR 1 mRNA is less sensitive to degradation upon

Subcellular localization ofSBP2 and EFsec

81

tempting to speculate that shuttling of SBP2 into the nucleolus is crucial for its association with other factors necessary for selenoprotein synthesis. One such factor may be the ribosomal protein L30, which has recently been implicated in selenoprotein synthesis [44]. If true, this would add emphasis to the point that it matters not only at what levels the important selenoprotein synthesis factors are present, but how they are localized within the cells as well. References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30.

GV Kryukov, S Castellano, SV Novoselov, AV Lobanov, O Zehtab, R Guigo, VN Gladyshev 2003 Science 300:1439 MJ Berry, L Banu, YY Chen, SJ Mandel, JD Kieffer, JW Harney, PR Larsen 1991 Nature 353:273 PR Copeland, JE Fletcher, BA Carlson, DL Hatfield, DM Driscoll 2000 EMBO J 19:306 RM Tujebajeva, PR Copeland, XM Xu, BA Carlson, JW Harney, DM Driscoll, DL Hatfield, MJ Berry 2000 EMBO R 2;2158 D Fagegaltier, N Hubert, K Yamada, T Mizutani, P Carbon, A Krol 2000 EMBO J 19:4796 MJ Berry 2000 Translational Control of Gene Expression N Sonenberg, JWB Hershey, MB Mathews (Eds) Cold Spring Harbor Laboratory Press Cold Spring Harbor 763 RM Tujebajeva, PR Copeland, XM Xu, BA Carlson, JW Harney, DM Driscoll, DL Hatfield, MJ Berry 2000 EMBO Rep \ A 58 AM Zavacki, JB Mansell, M Chung, B Klimovitsky, JW Harney, MJ Berry 2003 Mol Ce//11:773 MJ Berry, JW Harney, T Ohama, DL Hatfield 1994 Nucleic Acids Res 22:3753 AM Zavacki, JB Mansell, M Chung, B Klimovitsky, JW Harney, MJ Berry 2003 Mol Cell 11:773 MJ Berry, L Banu, JW Harney, PR Larsen 1993 EMBO J \2-33l5 S Himeno, HS Chittum, RF Burk \996 JBiol Chem 271:15769 S Ma, KE Hill, RM Caprioli, RF Burk 2002 JBiol Chem 111: 12749 BA Carison, XM Xu, VN Gladyshev, DL Hatfield 2005 JBiol Chem 280:5542 SL Weiss, RA Sunde 1998 mA 4:816 PM Moriarty, CC Reddy, LE Maquat 1998 Mol Cell Biol 18:2932 LE Maquat 2004 Nat Rev Mol Cell Biol 5:89 J Weischenfeldt, J Lykke-Andersen, B Porse 2005 CurrBiol 15:R559 Z Kan, EC Rouchka, WR Gish, DJ States 2001 Genome Res 11:889 N Visa, AT Alzhanova-Ericsson, X Sun, E Kiseleva, B Bjorkroth, T Wurtz, B Danehoh 1996 Ce//84:253 S Weiss Sachdev, RA Sunde 2001 Biochem J 351:851 K Wingler, C Muller, R Brigelius-Flohe 2001 Biofactors 14:43 C Muller, K Wingler, R Brigelius-Flohe 2003 Biol Chem 384:11 KE Hill, PR Lyons, RF Burk 1992 Biochem Biophys Res Commun 185:260 XG Lei, JK Evenson, KM Thompson, RA Sunde 1995 JNutr 125:1438 LF Pemberton, BM Paschal 2005 Traffic 6:187 PR Copeland, VAStepanik, DM Driscoll 2001 Mol Cell Biol 21:1491 RM Tujebajeva, JW Harney, MJ Berry 2000 JBiol Chem 275:6288 JR Gasdaska, JW Harney, PY Gasdaska, G Powis, MJ Berry 1999 J Biol Chem 274:25379 C Curcio, MM Baqui, D Salvatore, BH Rihn, S Mohr, JW Harney, PR Larsen, AC Bianco 2001 JBiol Chem 276:30183

82

Selenium: Its molecular biology and role in human health

31. 32. 33. 34. 35. 36.

PR Copeland, DM Driscoll 2002 Methods Enzymol 347:40 PP Dennis 1997 Cell 89:1007 MJ Christensen, PM Cammack, CD Wray 1995 JNutr Biochem 6:367 KB Hadley, RA Sunde 2001 JNutr Biochem 12:693 D Fagegaltier, P Carbon, A Krol 2001 Biofactors 14:5 X Sun, X Li, PM Moriarty, T Henics, JP LaDuca, LE Maquat 2001 Mol Biol Cell 12:1009 MJ Berry, PR Larsenl993 Biochem Soc Trans 21:827 L Flohe, WA Gunzler, HH Schock 1973 FEBSLett 32:132 GF Jr Combs, SB Combs 1984 Annu Rev Nutr 4:257 S Taketani, H Kohno, R Tokunaga, T Ishii, S Bannai, 1991 Biochem Int 23:625 B Gereben, A Kollar, JW Harney, PR Larsen 2002 Mol Endocrinol 16:1667 DL St Germain, RA Schwartzman, W Croteau, A Kanamori, Z Wang, DD Brown, VA Gallon 1994 Proc Natl Acad Sci USA 91:7767 AM Dumitrescu, XH Liao, MS Abdullah, J Lado-Abeal, FA Majed, LC Moeller, G Boran, L Schomburg, RE Weiss, S Refetoff 2005 Nat Genet 37:1247 L Chavatte, BA Brown, DM Driscoll 2005 Nat Struct Mol Biol 12:408 L de Jesus et al 2006 Mol Cell Biol 26:1795

37. 38. 39. 40. 41. 42. 43. 44. 45.

Chapter 8. Selenocysteine biosynthesis and incorporation may require supramolecular complexes Andrea L Small-Howard and Maria J Berry Department of Cell and Molecular Biology, John A. Burns School of Medicine, University of Hawaii at Manoa, Honolulu HI 96813, USA

Summary: Selenoproteins are a protein family in which the amino acid, selenocysteine (Sec) is incorporated cotranslationally at UGA "stop" codons. Selenoprotein synthesis requires unique RNA and protein factors in order to recode UGA. In eukaryotes, both the biosynthesis and incorporation of Sec employ nucleo-cytoplasmic shuttling of multiple protein factors, and this shuttling appears to play crucial roles in the functions of these factors. A dynamic series of supramolecular complexes may increase the efficiency of Sec incorporation, as well as the recycling of the Sec biosynthesis factors. This chapter will review the putative functions of these factors and explore the protein-nucleic acid and protein-protein interactions that define their biological roles. In addition, preliminary attempts to isolate supramolecular translation complexes from in vivo and in vitro systems will be discussed. Recent evidence suggests that both the biosynthesis of Sec-tRNA and the incorporation of Sec are physically coupled within supramolecular complexes that retain and coordinate the many factors required. tRNA channeling systems for the standard 20 amino acids rely on cotranslational complexes for spatial and temporal regulation of the multiple enzymatic factors responsible for aminoacyl-tRNA recycling. This chapter draws parallels between these functionally equivalent systems in eukaryotes. Introduction The selenoprotein family exhibits diversity of form, function, and distribution. Although many of the first described selenoproteins are enzymes that utilize the redox potential of selenium in the enzyme's active site, more recently identified selenoproteins serve a wide variety of functions, including structural, transport, signaling, or enzymatic roles [1-5]. Individual selenoproteins may be intracellular, transmembrane, or secreted proteins [1,6]. Expression of selenoproteins ranges from ubiquitous to tissuespecific [2,3,7]. Selenoproteins have been described in all three kingdoms of life, but they are not represented equally by all groups within these kingdoms [8-14]. In addition, different mechanisms for selenoprotein biosynthesis and incorporation occur in eukaryotes, prokaryotes and archaea [8,10,12,15-17].

84

Selenium: Its molecular biology and role in human health

This chapter will explore the essential role that supramolecular translation complexes may play in eukaryotic selenoprotein biosynthesis and incorporation. Significance of selenium in human health The essential micronutrient selenium is a mineral found in soil and is supplied in human diets primarily as edible plants, or secondarily in the meat of primary or secondary consumers of plant matter [18-20]. Around the world, most soils contain adequate quantities of selenium; notable exceptions include some areas of China, Russia, New Zealand, Finland and Zaire [2123]. Several diseases which have been attributed to selenium deficiency arose in the areas with selenium deficient soil [20,24,25]. These include: Keshan disease, Kashin-Beck disease, and Myxedematous Endemic Cretinism [18,21,23,25-27]. The discovery of human diseases related to selenium deficiency underscores its importance in maintaining good human health. In addition, clinical studies suggest a role for selenium supplementation in the prevention and treatment of some forms of cancer [18,21,28-32]. From selenium to selenoproteins, a review of Sec biosynthesis Selenium exerts its biological activity as Sec, the 21^' amino acid [33,34]. Whereas most amino acids are aminoacylated in mature form onto their cognate tRNAs, Sec is synthesizedfi"omserine in a multi-step reaction while attached to its unique Sec tRNA'^'^'^*''. The pathway of Sec biosynthesis has been elucidated in considerable detail for prokaryotes; researchers are working to resolve the Sec biosynthesis pathways of eukaryotes and archaea. hi prokaryotes. Sec biosynthesis occurs as follows. Sec tRNAt^''^^"' serves as the platform for the conversion of serine to Sec. Selenophosphate synthetase (SELD) converts selenide, fi-om dietary or recycled selenium, and ATP to produce the high energy selenophosphate [35]. Sec-tRNA^^"^'^" is charged with serine by the enzyme, seryl tRNA synthetase. Sec synthase (SELA), a pyridoxal-phosphate containing enzyme, converts the resultant seryltRNA^^'^^" to selenocysteyl-tRNA'^'''^" through a dehydration reaction [36]. Elimination of the water molecule creates an aminoacrylyl-tRNA-enzyme intermediate. Addition of a high energy selenophosphate across the double bond, followed by elimination of the phosphate group and release from the enzyme produces the mature Sec-tRNA'^'^'^''. Recent descriptions of the Sec biosynthesis pathways in eukaryotes and archaea demonstrate that the processes are similar, but not identical, to the pathway in prokaryotes [37]. These pathways appear to contain different enzymes at particular steps, and may also require additional steps [38-41]. hi eukaryotes, where nuclear transcription and cytoplasmic translation present an additional regulatory challenge, it is unclear whether all, or some, of the Sec biosynthesis pathway

Supramolecular complexes in selenoprotein synthesis

85

occurs in the nucleus or in the cytoplasm. Our current model of Sec biosynthesis suggests that preliminary steps may occur in the cytoplasm, but the final maturation of Sec-tRNA^^"^^^" presumably occurs in the nucleus, as demonstrated for the other aminoacylated tRNAs [42]. The presence of both nuclear localization signals and nuclear export signals in several selenoprotein biosynthesis factors suggests that eukaryotic Sec factors may shuttle between the cytoplasm and the nucleus, as necessary [43]. Sec incorporation strategies Similarly, nucleo-cytoplasmic shuttling of selenoprotein factors appears necessary to facilitate Sec incorporation into selenoproteins in eukaryotes [41]. Although selenoprotein synthesis in all three kingdoms shares involvement of the Sec-tRNA^^"^^'^ and the SECIS element, additional processes distinguish Sec incorporation in prokaryotes, eukaryotes and archaea. In prokaryotes, the SECIS element is located immediately downstream of the UGA codon, within the coding region of the selenoprotein mRNA [44]. In contrast, the SECIS element is located in the 3'-untranslated region of both eukaryote and archaea selenoprotein mRNAs. The functional implications of these differences in SECIS location are significant. In prokaryotes, the Sec cotranslational complex (i.e., SELB and Sec-tRNA'^"^^*^ interacting with the ribosomal A-site) must disassociate from the ribosome and the stem-loop of the SECIS-element must "melt" in order for translation to continue through the coding region. Because the SECIS element(s) are located in the 3'-untranslated region of selenoprotein mRNAs fi-om eukaryotes and archaea [45], the stem-loop can remain formed without interfering with the progress of the ribosome and cotranslational complexes can remain associated with the SECIS, and perhaps the ribosome. In addition, having the SECIS in the 3'-untranslated region allows for the cotranslational insertion of multiple See's within a single selenoprotein [46]. For example, human selenoprotein P (SelP) contains 10 Sec residues but only two SECIS-elements; therefore at least one of these SECIS elements must code for more than one Sec residue. Another difference between Sec incorporation in prokaryotes and in eukaryotes and archaea involves the requirement of protein factors. In prokaryotes, a single protein factor, SELB, performs multiple functions. SELB binds to the SECIS element and acts as an elongation factor by carrying the Sec-tRNA and associating with the ribosome [47]. In both eukaryotes and archaea, SelB homologs were identified that function as elongation factors, binding Sec-tRNA'^'^'^'^ [12]. In archaea the elongation factor and SECIS binding activities reside within the same protein, despite the physical separation of the UGA and SECIS (see Chapter 2). In eukaryotes, two or more protein factors are required to accomplish the elongation factor and SECIS binding functions [45]. This

86

Selenium: Its molecular biology and role in human health

chapter will explore the interactions of the many cotranslational factors involved in Sec incorporation in eukaryotes. Major question How are the factors involved in Sec biosynthesis and incorporation coordinated in eukaryotes? Our working hypothesis is that supramolecular complexes facilitate the organization of this highly complex process. First, we will describe the protein factors and their putative functions in selenium biosynthesis and incorporation. Then we will review data from interaction studies, which may provide the key to understanding Sec biosynthesis and incorporation. Preliminary data identify at least two supramolecular complexes involved in these processes. We conclude by illustrating parallels between tRNA-channeling and cotranslational selenoprotein biosynthesis/incorporation complexes in eukaryotes. Protein factors implicated in euliaryotic Sec biosyntliesis Although the Sec biosynthetic pathway in eukaryotes is still being resolved, many of the proteins involved have been identified based on the prokaryotic pathway (refer to "From selenium to selenoproteins"), hi eukaryotes, two distinct selenophosphate synthetase enzymes, SPSl and SPS2, convert selenide and ATP into the high energy selenophosphate compound. SPS2 is a selenoprotein, which may add an auto-regulatoiy element to the process, hi both prokaryotes and eukaryotes, Sec-tRNA'^'' is charged with serine by the enzjmie, seryl tRNA synthetase. The following is the part that's elusive and not fully established. By analogy with the pathway in prokaryotes, a putative Sec synthase activity would be needed to convert the resultant seryltRNA^^^'^^^' to selenocysteyl-tRNA^^^^^''. The Soluble Liver Antigen/Liver Pancreas (SLA/LP) protein, a 48 kDa protein identified in association with tRNA'^''^*^ [48], is likely to act in the Sec synthase reaction in eukaryotes. Sequence analysis of SLA/LP reveals that this protein is a member of the pyridoxal phosphate-dependent transferase superfamily, a family which also includes bacterial Sec synthase [49]. Li addition, a phosphoseryl tRNA kinase has been identified and shown to convert seryl-tRNA to phosphoseryl-tRNA, which is likely to be an intermediate in the transition to selenocysteyl-tRNA[50]. An analogous pathway in archaea shows CystRNA synthesis progressing through a phosphoseryl-tRNA intermediate with a synthase exhibiting homology to SLA/LP. The role of the phosphoseryltRNA kinase and the precise identity of the Sec synthase in eukaryotes bear further investigation. In eukaryotes, an additional methylation is necessary prior to the synthesis of some stress-related selenoproteins [40]. SECp43, a 43 kDa protein that binds Sec-tRNA'^'''^[51], is the likely candidate for the 2'-(9-methylation of the 5-methoxycarbonylmethyl uridine at the wobble base position in Sec-tRNA'^'^^" [41]. In addition, the protein binding studies

Supramolecular complexes in selenoprotein synthesis

87

described below suggest that SECp43 may have an additional role in promoting assembly of both Sec biosynthesis and incorporation complexes. Protein factors implicated in eukaryotic Sec incorporation As previously mentioned, Sec incorporation in eukaryotes requires more than one protein factor. An eukaryotic elongation factor (eEFsec), the homolog of prokaryotic SELB, is able to bind Sec-tRNA'^"^^^ but eEFsec does not bind to SECIS elements [46]. Two other protein factors bind the SECIS elements in selenoprotein mRNAs and also bind ribosomal sites. These factors are Sec Binding Protein 2 (SBP2 [52]) and ribosomal protein L30 (rpL30 [53]). SBP2 and rpL30 appear to compete for SECIS binding [53]; however, their precise roles in Sec incorporation are being investigated (see Chapter 6). SBP2 has been shown to associate with ribosomes, and rpL30 exists both free and as a component of the large ribosomal subunit [53,54]. As alluded to above, SECp43 may also play a role in promoting the association of supramolecular complexes involved in Sec incorporation. In particular, it seems to promote the association between eEFsec and SBP2. The interaction studies described below should clarify this proposed role further. Eukaryotic protein:RNA interactions in Sec biosyntliesis/incorporation Most of the protein factors involved in Sec biosynthesis and incorporation have been identified through RNA binding studies. Table 1 summarizes the proteiniRNA interactions that have been characterized to date.

Table 1. Protein:RNA interactions in Sec biosynthesis/incorporation

Protein SLA/LP SECp43 eEFsec SBP2 rpL30

RNA and reference Sec-tRNA^'^^'^J'*^ [48] Sec-tRNA^^'J'^'[51] Sec-tRNA"*''''^'= [46] SECIS mRNA [52]; rRNA [54] SECIS mRNA [52]; rRNA [53]

Three of the factors involved in Sec biosynthesis and incorporation were characterized based on their specific-binding to Sec-tRNA'^'^'^*'. The other two were identified based on interactions with the SECIS element in selenoprotein mRNAs, but were also subsequently shown to interact with ribosomes. The importance of these protein:RNA interactions will be discussed in our review of protein:protein interactions in Sec incorporation and biosynthesis.

88

Selenium: Its molecular biology and role in human health

Eukaryotic protein:proteiii interactions in Sec biosynttiesis/incorporation The involvement of protein:protein interactions in Sec biosynthesis and incorporation has been tested using several methods; however, a simple division can be made between the studies performed in vitro or in vivo. In vitro protein:protein interaction studies Expressed and purified proteins were characterized for interactions occuring without the additional factors found in cells or cell lysates. The in vitro studies assessed direct protein:protein binding between the following Sec factors, pair-wise, as summarized in Table 2. Table 2. In vitro protein:protein interactions in Sec biosynthesis/incorporation

SPSl SLA/LP SECp43 eEFsec

SBP2

eEFsec

SECp43

-

-

-

SLA/LP [551

[55]*

minus sign = no interaction; * requires Sec-tRNA'^"'^°°

In these experiments, an interaction between SPSl and SLA/LP was the only protein:protein interaction demonstrated without adding any other cellular products [55]. SPSl and SLA/LP are the enzymes purported to be responsible for the initial stages of Sec biosynthesis. More specifically, SPSl is a selenophosphate synthetase, and SLA/LP is provisionally described as the Sec synthase. When ^^Se-labeled Sec-tRNA'^''^"' was added to eEFsec, in vitro, an EMSA demonstrated a binding-shift of a Sec-tRNA'^^'^'^^VeEFsec complex [55]. Adding SECp43 super-shifted the Sec-tRNA'^'''^'7eEFsec complex [55]. This suggests an indirect interaction between SectRNA^^^'^^^7eEFsec/SECp43. One striking conclusion that can be drawn fi-om these studies is the importance of the RNAs in the protein:protein interactions demonstrated in vivo. These results suggest that the RNAs are at the core of the supramolecular complexes involved in the Sec incorporation cycle. In vivo protein:protein interaction studies In vitro studies had produced evidence for only a single protein:protein interaction (Table 2). In vivo studies were performed to reveal protein:protein interactions requiring other cellular cofactors involved in Sec biosynthesis and incorporation. In vivo protein:protein interaction studies were performed in the cellular environment, where endogenous protein:RNA interactions

89

Supramolecular complexes in selenoprotein synthesis

were, presumably, not limiting. In addition, Sec-tRNA'^^'^^^'^ was cotransfected in an effort to promote all of the possible interactions between the Sec factors. Tagged factors were cotransfected as pairs and triplets. The results of these studies are summarized in Table 3.

Table 3. In vivo protein:protein interactions in Sec biosynthesis/incorporation

SPSl SLA/LP SECp43 eEFsec

SBP2 [55]* [46, 55, 56]**

eEFsec [55]*

SECp43 [55]* [41,55]*

SLA/LP [55]®

* requires Sec-tRNA^^"^**"^ co-transfection; ®requires SECp43 co-transfection; 'enhanced by SECp43 co-transfection; minus sign = no interaction

In these studies, SECp43 enhances protein-protein interactions between two distinct sets of Sec factors [55]. Initial attempts to co-immunoprecipitate (IP) SPSl and SLA/LP in vivo produced unconvincing results despite the fact that these factors strongly interacted in the in vitro system. When SECp43 was cotransfected with SPSl and SLA/LP, the interactions between the three were all positive [55]. The interaction between eEFsec and SBP2 was used to identify eEFsec from a co-IP with SBP2, in vivo [46]. The eEFsec and SBP2 interaction is dependant on Sec-tRNA^^*'^^^"' and is also enhanced by SECp43 [55]. The in vivo interactions of Sec-tRNA'^"'^'7eEFsec/SECp43 and now with SBP2 are parallel to the in vitro reaction of these Sec factors. Taken together these results suggest that SECp43 may play a role in the formation or stability of supramolecular complexes in Sec biosynthesis and incorporation. The role of SECp43 in supramolecular complex formation will be explored in the following section on the isolation of supramolecular complexes. Isolation of both nuclear and cytoplasmic supramolecular complexes As previously mentioned, nuclear transcription and cytoplasmic translation necessitates nucleo-cytoplasmic shuttling of translational factors. In order to assess the sub-cellular locations of supramolecular complexes involved in Sec biosynthesis and incorporation, cell lysates were fractionated before isolating complexes by IP [55]. Two different methods isolated supramolecular complexes involved in Sec biosynthesis and incorporation. The first isolated complexes composed of epitope-tagged Sec factors; whereas, the second isolated complexes with endogenous Sec factors. Both methods have their own strengths and weaknesses, but taken together, they

90

Selenium: Its molecular biology and role in human health

provide a preliminary view of the supramolecular complexes involved in Sec biosynthesis and incorporation. In the first method, all five of the Sec factors listed in Table 4 have epitope-tags to facilitate the identification of individual components within the supramolecular complexes. These factors were cotransfected as a quintet (all of the factors listed plus Sec-tRNA^^^'^^^') and a quartet (the factors listed in Table 4 and Sec-tRNA'^"''^"', but without SECp43) to test the hypothesis that SECp43 affects the formation of supramolecular complexes in Sec biosynthesis and incorporation [55]. Cells were fractionated, and then aliquots of the nuclear and cytoplasmic lysates were IP'ed with each of the antibodies corresponding to the tags on the factors, independently. Western blots of each set of isolated complexes, from the different IP reactions, were then probed with each of the antibodies to the tags on the factors, in series. The results from the cotransfections with all five Sec factors are summarized in Table 4; whereas, the consequences of omitting SECp43 from the supramolecular complexes are summarized in Table 5.

Table 4. Supramolecular complexes in Sec biosynthesis/incorporation

SPSl IP Western Cyt Nuc SPSl ++ ++ SLA/LP ++ ++ SECp43 ++ ++ eEFsec SBP2 +

SLA/LP SECp43 Cyt Nuc Cyt Nuc

eEFsec Cyt Nuc

++ ++ ++

++

++

—

—

++ ++ ++

++ ++ ++

—

++ ++

— ~ —

— — ~

++ ++ ++

+

—

++ ++ —

SE P2 Cyt Nuc

"++"=strong positive reaction; "+"=weak but detectable reaction; "—"=no detectable reaction shaded boxes=the same Sec factor is both IP'ed and probed for on the western blot

In Table 4, the Sec factor selected for in the IP is listed across the top. The cytoplasmic (Cj^) and nuclear (Nuc) fractions were IP'ed separately as indicated. Any of the other Sec factors in a complex with the IP'ed Sec factor would also be removed from the respective cell lysate with the target of the IP reaction and be detected during western blot analysis with the antibody corresponding to the epitope tag on the Sec factor (listed in the far left column). The shaded boxes represent the control situation in which the same Sec factor is both IP'ed and probed for on the western blot. SPSl, SECp43, and eEFsec IP in both the cytoplasmic and nuclear fractions when the western blot is performed with the antibody corresponding to the same Sec factor as the IP. In confrast, SLA/LP and SBP2 do not IP in the nucleus under these same conditions. Presumably, the epitopes for SLA/LP and

91

Supramolecular complexes in selenoprotein synthesis

SBP2 are not accessible to the antibody during the IP reaction in nuclear lysates because transfection controls verify that these Sec factors have entered the nucleus. One noteworthy trend is the mutual exclusion of SLA/LP and eEFsec from supramolecular complexes with otherwise similar Sec factors. Taken together, the co-IP results suggest that at least two distinct supramolecular complexes participate in Sec biosynthesis and incorporation [55]. The first "biosynthesis" complex is likely composed of SPSl, SLA/LP, SECp43, and Sec-tRNA'^"^'^". The "biosynthesis" complex is likely preformed in the cytoplasm, and then enters the nucleus where SLA/LP leaves the complex and SPSl, SECp43, and Sec-tRNA^^"'^^'' are joined by eEFsec and SBP2. The new "incorporation" complex is also detected in the cytoplasm, suggesting that this complex must also shuttle. Information regarding the role of ribosomal interactions in this process might help to determine when SBP2 binds to the SECIS element within the selenoprotein mRNA. Both Sec factor complexes can be detected in cytoplasmic and nuclear fractions, this suggests that nuclear-cytoplasmic shuttling is occurring. Fluorescent microscopy studies support and extend this hypothesis [56]. The role of S£Cp43 in supramolecular complexes To test the hypothesis that SECp43 affects the formation of supramolecular complexes in Sec biosynthesis and incorporation, cells were cotransfected with the quartet of factors listed in Table 5 plus Sec-tRNAf^"^^"=, but without SECp43. Otherwise, the experiment was performed exactly as above.

Table 5. Supramolecular complexes in the absence of added SECp43

IP Western SPSl SLA/LP eEFsec SBP2

SPSl Cyt Nuc

SLA/LP Cyt Nuc

+

, +

—

~ ~ ~

~ ~ ~

— —

+

— ~ — —

e£] i'sec

SB]P2

Cyt

Nuc

Cyt

Nuc

~ ~

~ —

~ —

+ +

+ +

+ +

~ ~ ~ —

"+"=strong positive reaction; "-"=no detectable reaction; shaded boxes=the same Sec factor is both IP'ed and probed for on the western blot

These results support the hypothesis that SECp43 is necessary for the interaction of SPSl and SLA/LP. SECp43 also stabilizes the larger supramolecular complexes described in the previous section. The interactions

92

Selenium: Its molecular biology and role in human health

between cEFsec and SBP2 were detected in the absence of SECp43, even though SECp43 enhances this interaction, in vivo. Supramolecular complexes composed of endogenous Sec factors In the second approach, a single tagged Sec factor was transfected into cells with Sec-tRNA'^"^^' [55]. Nuclear and non-nuclear cell fractions were IP'ed by the tagged co-factor to capture any endogenous Sec factors associated as a complex. Electrophoresis was performed using non-reducing conditions to promote complex stability. Each of the Sec factors listed in Table 4 were tested in this manner, independently. Western blot analysis showed complexes of greater than 250 kDa in the cytoplasmic fractions of cells transfected with the following factors: SPSl, SECp43, and SBP2. A nuclear complex of approximately 150 kDa was visualized in cells transfected with eEFsec. Supramolecular complexes were not detected from cells transfected with SLA/LP. Several explanations are possible for this result. A supramolecular complex would not be captured in the assay if the epitope tag on the factor was not accessible to the antibody. Furthermore, larger complexes (i.e., attached to ribosomes) are unlikely to be resolved during electrophoresis. The next section will address the question: Which of these Sec factors interact with the ribosome? Sec factors associated with ribosomes Ribosome co-purification strategies have demonstrated that two of the Sec factors, SBP2 and rpL30, are associated with ribosomes. In addition, both SBP2 [54] and rpL30 [53] compete with each other for SECIS-binding (see Chapter 7). Additionally, cells transfected with the same five epitope-tagged Sec factors as in Table 4, or control cells, were lysed and separated on a sucrose gradient to separate the polysome-containing fractions [55]. The initial, non-ribosomal fractions contained all of the Sec factors. The fraction containing the ribosomal small sub-imits contained SPSl and Secp43. All of the whole ribosome-containing fractions contained eEFsec and SBP2. Most of the ribosome containing fractions also showed SLA/LP at a lower but detectable level. The interaction of Sec factors with ribosomes and the cycling of the Sec factors have considerable relevance to tRNA-channeling type processes defined in Sec-tRNA^^"'^^*^ biosynthesis and incorporation, as discussed below. Comparisons between Sec cycling and tRNA-channeling Large, multi-enzyme, protein complexes containing tRNA are responsible for the organization and spatial/temporal regulation of the tRNA channeling processes in eukaryotic cells [57-59]. Supramolecular complexes presumably play a similar role in increasing the efficiency of Sec biosynthesis and incorporation. Considerable evidence supports a channeled tRNA cycle in

Supramolecular complexes in selenoprotein synthesis

93

eukaryotes [57-65]. The channeled tRNA cycle functions alongside the ribosomal-based translation reactions, very much like Sec biosynthesis and incorporation processes. During the channeled tRNA cycle, aminoacyl-tRNA and tRNA are shuttled from the aminoacyl-transferase and relevant elongation factor (eEF-lA), to the ribosome for incorporation of the amino acid, and then back to the aminoacyl-transferase to recharge the tRNA [57,60,63]. This cycle occurs without significant dissociation of the factors into the cellular fluid [62,63]. Similarly, the Sec-tRNAf^'^^*^ associates with the enzymes responsible for recharging the tRNA^^''^" (SPSl and SLA/LP) and also the relevant elongation factor (eEFsec), during Sec biosynthesis and incorporation. The cotranslational, supramolecular complex, containing eEFsec/Sec-tRNAt^''^7SECp43/SPSl, is delivered to the ribosome, as has been described in tRNA channeling for the other 20 amino acids. If the analogy between Sec biosynthesis and incorporation and tRNA-channeling remains consistent, the eEFsec/Sec-tRNAf^''^'7SECp43/SPSl complex leaves the ribosome and the tRNA'^'^^*^ is recharged. The continued association of the Sec factors would increase the probability of this highly complex process recurring in eukaryotes. The intracellular tRNA cycle requires discrete changes in the composition of the supramolecular complexes, as has been described for Sec biosynthesis/incorporation (Table 4). In addition, tRNA channeling depends on interactions between eEF-lA and actin in the cytoskeleton [64], and this interaction is mediated by Rho A kinase (ROK or ROCK). Preliminary results suggest that eEFsec interacts with actin in the cytoskeleton also via ROCK [unpublished data]. Future work will explore the cytoskeletal regulation of Sec biosynthesis and incorporation. The tRNA channeling complex is shuttled through the nuclear pore complex (NPC) using several factors including vigilin [58], but less is known about the mechanism for transport of the Sec supramolecular complexes through the NPC. Several Sec factors have NLSs and NESs [43], which are specifically recognized by receptors at the NPC. Studies have shown that some of the Sec factors are able to shuttle [56]. Future work will address whether Sec-tRNA^^"'^^" and Sec factors are cycled during Sec biosynthesis and incorporation; analogous to tRNA channeling for the other twenty amino acids. Conclusion This chapter has reviewed data supporting the hypothesis that large, dynamic supramolecular complexes mediate Sec biosynthesis and incorporation. Aminoacyl-tRNA chatmeling relies on similar large supramolecular complexes to coordinate charging of tRNAs with translation at the ribosome. Sec biosynthesis and incorporation require precise spatial and temporal regulation, like the other amino acids during a tRNA channeling cycle.

94

Selenium: Its molecular biology and role in human health

Continuation of these studies should provide an explanation for the regulation of this highly complex process. References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47.

GV Kryukov, S Castellano et al. 2003 Science 300:1439 PR Hoffmann, MJ Berry 2005 nyroid 15:769 KM Brown, JR Arthur 2001 Public Health Nutr 4:593 VN Gladyshev, GV Kryukov 2001 Biofactors 14:87 S Gromer, JK Eubel etal. 2005 Cell Mol Life Sci [Epub ahead of print] F Ursini, M Maiorino et al 1995 Methods Enzymol 252:38 D Behne, A Kyriakopoulos 2001 Annu Rev Nutr 21:453 M Rother, A Resch et al 2001 Mol Microbiol 40:900 CB Foster 2005 Mol Biol Evol 22:383 M Rother, A Resch et al 2001 Biofactors 14:75 GV Kryukov, VN Gladyshev 2004 EMBO Rep 5:538 D Fagegaltier, P Carbon et al.lOOX Biofactors 14:5 S Castellano, SV Novoselov et al 2004 EMBO Rep 5:71 LH Fu, XF Wang et al 2002 J Biol Chem 277: 25983 M Leibundgut, C Frick et al. 2005 EMBO J 24:11 B Lenhard, O Orellana et al 1999 Nucleic Acids Res 27:721 PJ Keeling, NM Fast era/. 1998 JA/o/£vo/47:649 RS Bedwal, N Nair et al 1993 Med Hypotheses 41:150 JW Finley, A Sigrid-Keck et al 2005 J Nutr 135 1236 S Dodig, I Cepelak 2004 Acta Pharm 54:261 DGBarceloux 1999 J Toxicol Clin Toxicol 37:145 Y Xia, KE Hill et al 2005 Am J Clin Nutr 81:829 J Kohrle 1999 Biochimie 81:527 J Wu, GL Xu 1987 J Trace Elem Electrolytes Health Dis 1:39 AT Nesterov 1964 Arthritis Rheum 7:29 MP Burke, K Opeskin 2002 Med Sci Law 42:10 JB Vanderpas, B Contempre ef a/. 1990 Am J Clin Nutr 52:\09n OH Al-Taie, N Uceyler et al 2004 Nutr Cancer 48:6 GF Combs Jr. 1999 Med Klin Munich 94(Suppl 3): 18 V Diwadkar-Navsariwala, AM Diamond 2004 J Nutr 134:2899 HE Ganther2001 Adv Exp Med Biol 492:\\9 SV Novoselov, DV Calvisi et al 2005 Oncogene [Epub ahead of print] A Bock, K Forchhammer et al 1991 Mol Microbiol 5:515 A Bock, K Forchhammer et al. 1991 Trends Biochem Sci 16:463 A Ehrenreich, K Forchhammer et a/. 1992 Eur J Biochem 206:767 K Forchhammer, K Boesmiller et al 1991 Biochimie 73:1481 MJ Guimaraes, D Peterson e< a/. \99e, Proc Natl Acad Sci U S A 9l:\50^6 YOgasawara, GMLacourciereera/. 2005P/-ocA/a//y4carf5c/6'5^ 102:1012 RR Jameson, AM Diamond 2004 RNA 10:1142 BA Carlson, XM Xu et al 2005 J Biol Chem 280:5542 XM Xu, H Mix et al 2005 J Biol Chem. 280:41568 E Lund, JE Dahlberg 1998 Science 282:2082 AL Small-Howard, MJ Berry 2005 Biochem Soc Trans 33:1493 DM Driscoll, PR Copeland 2003 Annu Rev Nutr 23:17 A Lescure, D Fagegaltier et al. 2002 Curr Protein Pept Sci 3:143 RM Tujebajeva, PR Copeland et al 2000 EMBO Rep 1:158 K Forchhammer, W Leinfelder et al 1989 Nature 342:453

Supramolecular complexes in selenoprotein synthesis 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65.

C Gelpi, EJ Sontheimer et al. 1992 Proc Natl Acad Sci USA 89:9739 T Kemebeck, AW Lohse et al. 2001 Hepatology 34:230 BA Carlson, XMXue? a/. lOOi, Proc Natl Acad Sci USA 101:12848 F Ding, PJ Grabowski 1999 RNA 5:1561 PR Copeland, JE Fletcher et al 2000 EMBO J19:306 L Chavatte, BA Brown et al 2005 Nat Struct Mol Biol 12:408 S A Kinzy, K Caban et al 2005 Nucleic Acids Res 33:5172 AL Small-Howard, N Morozova et al. 2005 [submitted] LA de Jesus, PR Hoffinann et al 2005 Mol Cell Biol [Epub ahead of print] BS Negrutskii, MP Deutscher 1992 Proc Natl Acad Sci USA 89:3601 T Vollbrandt, D Willkomm et al. 2004 Int JBiochem Cell Biol 36:1306 H Grosshans, G Simos et al 2000 J Struct Biol 129:288 H Grosshans, E Hurt et al 2000 Genes Dev 14:830 BS Negrutskii, MP Deutscher 1991 Proc Natl Acad Sci USA 88:4991 BS Negrutskii, R Stapulionis et al 1994 Proc Natl Acad SciUSA9\ :964 R Stapulionis, MP Deutscher 1995 Proc Natl Acad Sci USA 92:7158 R Stapulionis, S Kolli et al 1997 J Biol Chem 272:24980 R Stapulionis, MP Deutscher 1998 Methods Mol Biol 77:23

95

Part II

Selenium-containing proteins

Chapter 9. Selenoproteins and selenoproteomes Vadim N. Gladyshev Department of Biochemistry, University of Nebraska, Lincoln, NE 68688, USA

Summary: In the past several years, progress in genome sequencing and development of specialized bioinformatics tools allowed efficient identification of selenocysteine-containing proteins encoded in completely sequenced genomes. Information is currently available on selenoproteomes from a variety of organisms, including humans, which contain 25 known selenoprotein genes. This review provides basic information about mammalian selenoproteins and other known selenoprotein families. Analysis of full sets of selenoproteins in organisms provides exciting avenues for examining selenoprotein evolution and dependence of organisms on the trace element selenium and allows linking selenoproteins with specific biological and biomedical effects of dietary selenium. Introduction Selenium occurs in selenoproteins in two known forms. In several bacterial selenium-containing molybdoproteins, such as nicotinic acid hydroxylase and xanthine dehydrogenase, it is present in the form of a labile cofactor that contains a Se-Mo bond that is directly involved in catalysis [1-4]. Neither the exact chemical form of selenium in this cofactor nor its biosynthetic pathway are known. However, by far the major form of selenium in proteins is selenocysteine (Sec), the 2 r ' amino acid in the genetic code. It is encoded by TGA and has been found in all three major lines of descent (bacteria, archaea and eukaryotes). Research in the past five years strengthened the idea that the essential role of selenium in biology, as well as its beneficial roles in human health, are largely due to its presence in proteins in the form of Sec. In contrast to other amino acids found in proteins. Sec is typically utilized only when it is required for protein function. Accordingly, it is normally a key functional (catalytic) group in proteins, and essentially all selenoproteins with known functions directly use Sec in redox catalysis. Bioinformatics tools for selenoprotein identification Historically, selenoproteins have been identified by following the presence of selenium in protein fractions during isolation. Sec-containing proteins can

100

Selenium: Its molecular biology and role in human health

be metabolically labeled with ^^Se, a convenient y-emitter that remains covalently bound in proteins during SDS-PAGE and can be visualized on gels and membranes using a Phosphorlmager. Using this technique, a number of proteins were identified in both prokaryotes and eukaryotes [5-8]. However, it has become apparent that the applicability of this technique is limited to the most abundant proteins, whereas less abundant selenoproteins and those characterized by unique expression patterns could not be identified by this method. Remarkable progress in genome sequencing and other large-scale sequencing projects provided an attractive and mostly untapped resource that could be utilized for selenoprotein discovery. All selenoprotein genes have two characteristic features: a Sec-encoding TGA codon and a Sec insertion sequence (SECIS) element. The TGA triplets that code for Sec do not provide sufficient information at the nucleotide sequence level to identify them computationally. However, SECIS elements are amenable to these techniques as these structures are highly specific for selenoprotein genes, have conserved segments and possess a sufficiently complex secondary structure (Chapters 5-7). Therefore, initial bioinformatics analyses focused on SECIS elements, and selenoprotein discovery followed a simple strategy: 1) finding candidate SECIS elements; 2) analyzing upstream regions to identify coding regions; and 3) testing candidate selenoproteins for the presence of selenium by metabolically labeling cells with ^^Se. The first selenoproteins identified using this technique were selenoproteins R (now known as methionine-R-sulfoxide reductase), N and T [9,10]. This strategy was initially limited to cDNA sequences, but, following improvements in the computational description of SECIS elements and selenoprotein genes [11], the searches were later carried out on entire genomes characterized by small-to-moderate sizes [12,13]. To analyze larger genomes, a pair of closely related genomes was used (e.g., human and mouse genomes), which were searched for conserved pairs of SECIS elements that belonged to selenoprotein orthologs in these organisms [14]. In parallel, bioinformatics tools were developed that screened for coding function of TGA codons by analyzing sequences downstream of TGA [13]. The searches were also greatly aided by the observation that most selenoprotein genes have homologs, in which Sec is replaced with Cys. As a result, a strategy was developed wherein large protein databases (e.g., NCBI non-redundant protein database) could be searched against large nucleotide sequence databases that are known to contain selenoprotein genes (e.g., genomes of organisms containing Sec insertion machinery genes) [14-16]. These searches were designed to identify TGA-containing nucleotide sequences which when translated align with Cys-containing protein sequences from the protein database such that the conserved Cys residues align with translated TGA codons and these pairs are flanked by conserved

Selenoproteins and selenoproteomes

101

sequences. In a further extension of this approach, computational gene predictions were carried out on pairs of genomes (at the appropriate phylogenetic distance) to identify TGA to TGA and TGA to Cys codon alignments embedded in conserved stretches of sequence. This approach has also lead to the discovery of new selenoprotein families. 100 aa

Protein name 15 kPa selenoprotein {Sep15) Fep15 lodothyronine Deiodinase (DIs) Selenoprotein H (SelH) Selenoprotein I (Sell) Selenoprotein J (SelJ) Selenoprotffln M ^SelM) SoloroprotonOiSelO) 3Hbrn[jroiHTi P I S B ' P ) Molhiofi.-ic; R Suilcxido RorJuciaso (MJ;: Selercproieiri S iSeiS) bslsrcprote-i V ;Se:Vj Seienophosphate Synthetase (SPS) Glutathione Peroxidases (G"Px) SE^k-ritjprotfjri K i^i(;:K; Solonoprolo:n M iSoiV; Snlonoprotn^n T (ScIT; Selonoprolo.n U iSolU) Selenoprotein W (SelW) Thiorscioxin RocJiiclnses ( T R P MetNo,i.ne-S-SLilto>:deHedjctase:V5rA: Praiej-i D Ej.fide Isonerase 'PO s; Peroxiredoxin (Prx) Proline reductase (PrdB) Prx-like protein Thioredoxin (Trx) Formate dehydrogenase alpha ciiain (fdhA) Glycine reductase selenoprotein A (grdA) Glycine reductase selenoprotein B (grdB) AhpD-like protein Arsenate reductase Moiyfadopterin biosynthesis protein (MoeB) Glutaredoxln (Grx) DsbA-like protein Glutathione S^transferase (GST) Thiol-disulfide isomerase'like protein CMP domain-containing protein Rhodanase* related sulfurtransferase OsmC-like protein DsrE-like protein DsbG-IIke protein HesB-like protein Formylmethanofuran dehydrogenase (FMDH) Methylviologen-reducinghydrogenase Coenzyme F42Q-r6ducing hydrogenase Heterodisulphide reductase

IZZIZ

•z

LIZ)

Figure 1. Selenoprotein families. Selenoproteins that occur in vertebrates or single celled eukaryotes are highlighted by shaded boxes, and selenoproteins found in prokaryotes are shown in bold. On the right, relative sizes of selenoproteins are shown (relative to a 100 amino acid scale) and the location of Sec within protein sequence is shown by a black line.

Thus, although searches for SECIS elements could efficiently guide computational selenoprotein predictions, the alternative methods were possible. The Sec/Cys and Sec/Sec homology approaches were completely independent of the searches for SECIS elements. Importantly, both SECIS-

102

Selenium: Its molecular biology and role in human health

based and SECIS-independent algorithms identify similar sets of selenoprotein genes in various genomes, arguing that both tools show excellent performance and that all or almost all selenoproteins could be identified by these programs in completely sequenced genomes. Selenoproteins In the past five years, numerous papers reported on novel selenoprotein genes. Figure 1 lists currently known selenoprotein families. More than half of the selenoproteins in this figure are new (e.g., have been discovered after publication of the previous edition of this book), and essentially all of them were discovered through bioinformatics approaches. Some selenoproteins were named differently by various laboratories, sometimes causing confusion in regard to the number and identity of selenoproteins. Below, we briefly describe currently knoAvn selenoproteins, with the focus on selenoproteins found in mammals. More detailed information on selected selenoproteins can be found in various chapters in this book. Mammalian selenoproteins Glutathione peroxidases. Glutathione peroxidase 1 (GPxl) is the first known animal selenoprotein [17] and also perhaps the best studied (see Chapters 1416). It is a highly efficient antioxidant enzyme that catalyzes glutathionedependent hydroperoxide reduction. Mammals contain eight glutathione peroxidase homologs, of which five are selenoproteins, including GPxl (also known as cGPx), GPx2 (also known as GI-GPx), GPx3 (also known as pGPx), GPx4 (also known as PHGPx) and GPx6. Of these, only GPx4 is known to be essential during embryogenesis in mammals [18]. In addition to its antioxidant function, this protein serves a structural role in mature sperm and was implicated in site-specific disulfide bond formation [19]. Questions remain in regard to the function of all Sec-containing glutathione peroxidases. Recently, selenoprotein GPx homologs were detected in singlecelled eukaryotes and bacteria. Thvroid hormone deiodinases. Mammals have three deiodinases (DIl, DI2 and DI3) which activate or inactivate thyroid hormones by reductive deiodination. These selenoproteins are reviewed in detail in Chapter 19. It was thought that deiodinases evolved in animals, but recently, bacterial homologs of these proteins were found [16]. Thioredoxin reductases. Mammalian thioredoxin reductase (TR) was known for decades, but only recently was it discovered that these enzymes are selenoproteins (see Chapters 17 and 18). This family of proteins was the first in which Sec was found to be located at the penultimate C-terminus [6,20], and we now know that Sec in these proteins is part of the additional active site that is itself a substrate for the N-terminal thiol-disulfide active site [21-23]. Mammals contain three thioredoxin reductases, and all three are

Selenoproteins and selenoproteomes

103

selenoproteins [24]; therefore the entire thioredoxin system in mammals is dependent on selenium. TRl (TrxRl, TxnRdl) is a cytosolic protein. Its main function is to control the reduced state of thioredoxin. However, it exhibits broad substrate specificity [25] and occurs in the form of at least six isoforms generated by alternative splicing that differ in their N-terminal sequences [26-28]. Thioredoxin/glutathione reductase (TGR, also known as TR2 and TrxR3) is a protein that, compared to other animal thioredoxin reductases, has an additional N-terminal glutaredoxin (Grx) domain [22]. TGR can catalyze many reactions specific for thioredoxin and glutathione systems. This protein was recently implicated in the formation/isomerization of disulfide bonds during sperm maturation [29]. TR3 (also known as TrxR2) is a mitochondrial protein, whose function is likely to reduce mitochondrial thioredoxin and glutaredoxin 2. TRl and TR3 are essential proteins in mammals [30,31]. Methionine-R-sulfoxide reductase 1 (MsrBl). MsrBl was initially identified using bioinformatics approaches as Selenoprotein R [9] and Selenoprotein X [10], and later was shown to catalyze stereospecific reduction of oxidized methionine residues in proteins with thioredoxin as reductant (Chapter 11) [32]. In mammals, there are two additional MsrBs (MsrB2 and MsrB3); however, these contain Cys in the active site and reside in mitochondria and/or endoplasmic reticulum [33]. MsrBl is located in the cytosol and nucleus and is the most active MsrB in mammals [34]. 15 kPa selenoprotein (Sepl5). Sep 15 is a mammalian protein, but its homologs are found in other eukaryotes (mostly in animals) (Chapter 13) [7]. It resides in the endoplasmic reticulum where it binds UDPglucose:glycoprotein glucosyltransferase, a sensor of protein folding [35]. Recent structural analyses revealed its redox function. Sep 15 is characterized by the thioredoxin-like fold and is implicated in the cancer prevention effect of dietary selenium [36,37]. Selenophosphate synthetase 2 (SPS2). By analogy to bacterial selenophosphate synthetase SelD [38], SPS2 is thought to synthesize selenophosphate, a selenium donor compound (Chapter 4) [39]. SPS2 homologs are found in all three major domains of life. Selenoprotein P (SelP). SelP is the only selenoprotein with multiple Sec residues (Chapter 10) [40]. For example, human SelP has 10 Sec and zebrafish SelPa has 17 Sec residues. However, some organisms also have smaller SelPs (e.g., zebrafish SelPb contains a single Sec) [41]. SelP is the major plasma selenoprotein, which is synthesized in the liver and delivers selenium to certain other organs and tissues [42,43]. However, brain appears to synthesize its own pool of SelP (see Chapter 21). Selenoprotein W (SelW). SelW is the smallest mammalian selenoprotein (Chapter 12) [44]. Its function is not known. It was thought to be specific to

104

Selenium: Its molecular biology and role in human health

animals, but recently SelW homologs were identified in lower eukaryotes and bacteria [16]. Selenoprotein V (SelV). SelV has a C-terminal SelW homology domain and a larger N-terminal sequence of unknown function [14]. This protein is expressed exclusively in testes. Its function is not known. Selenoprotein T (SelT). SelT is a small selenoprotein with a putative Nterminal redox motif [9]. Its function is not known. Selenoprotein M (SelM). SelM is a distant homolog of Sep 15 and also has a thioredoxin-like fold and a predicted redox motif [37,45]. SelM resides in the endoplasmic reticulum. Its function is not known. Selenoprotein H (SelH). This small selenoprotein of unknown function with a predicted redox motif was first identified as BthD in flies [12,14]. Selenoprotein O (SelO). SelO is a widely distributed protein that has homologs in animals, bacteria, yeast and plants, but the function of any of these proteins is not known [14]. Only vertebrate homologs of SelO have Sec, which is located in the C-terminal penultimate position. In SelO homologsfi-omother organisms. Sec is replaced with Cys. Selenoprotein K (SelK). SelK is unusual among mammalian selenoproteins in that it does not have a pronoimced secondary structure [14]. This small selenoprotein contains a single transmembrane helix in the N-terminal sequence that targets this protein to the plasma membrane. SelK homologs can be detected in many eukaryotes, but no information is available on the function of this protein. Selenoprotein S (SelS). Like SelK, SelS has Sec in the C-terminal sequence and a single transmembrane region at the N-terminus [14]. Recent studies revealed its role in retrotranslocation of misfolded proteins from the endoplasmic reticulum of mammalian cells to the cytosol, where these proteins are further degraded [46]. SelS binds Berlin 1, an endoplasmic reticulum protein also involved in protein retrotranslocation. In addition, SelS was implicated in inflammation and immune response [47]. Selenoprotein N fSelN). One of the first selenoprotein discovered through bioinformatics approaches [10], SelN remains a selenoprotein of unknown function. This protein was implicated in the role of selenium in muscle function [48]. Selenoprotein I (SelD. Sell is a recently evolved selenoprotein specific to vertebrates [14]. This predicted membrane selenoprotein has no known function. Other eukaryotic selenoproteins Methioninc-S-sulfoxide reductase (MsrA). MsrA is a widely distributed protein family, whose function is to repair methionine residues in proteins [49]. MsrA catalyzes a stereospecific reduction of methionine-S-sulfoxides with thioredoxin. Only one selenoprotein MsrA was described, which is

Selenoproteins and selenoproteomes

105

present in Chlamydomonas, green algae [50]. All other known MsrAs are Cys-containing proteins. MsrA should not be confused with MsrB [33]. Both MsrA and MsrB are methionine sulfoxide reductases, but they have no sequence homology and catalyze complementary reactions using different diastereomers of methionine sulfoxide. Protein disulfide isomerase (PDI). Selenoprotein PDI is narrowly distributed in eukaryotes [51], whereas Cys-containing PDI is an essential protein involved in formation of disulfide bonds in the endoplasmic reticulum of eukaryotic cells. Selenoprotein U (SelU). SelU is found in the selenoprotein form in fish and several other marine organisms, but in mammals, all three SelU homologs are Cys-containing proteins [52]. SelU function is not known. Selenoprotein J (SelJ). SelJ has been described in fish and sea urchin, with Cys homologs only found in cnidarians. This selenoprotein shows similarity to the jellyfish Jl-crystallins and together with these proteins defines a unique subfamily within the larger family of ADP-ribosylation enzymes [53]. SelJ function is not known. Fish 15 kPa selenoprotein (Fepl5). Fepl5 is homologous to mammalian Sep 15 and SelM, but could only be detected in fish [54]. Fepl5 resides in the endoplasmic reticulum and possibly in Golgi. Its function is not known. Plasmodium selenoproteins Sell. Sel2. Sel3 and Sel4. The four Plasmodium selenoproteins show no detectable homology to any other proteins [55]. Sell and Sel4 have Sec in the C-terminal regions, similar to animal SelK and SelS. Functions of these proteins are not known. Prokaryotic selenoproteins Several selenoproteins, including selenophosphate synthetase, deiodinase homologs, glutathione peroxidase and SelW, occur in both prokaryotes and eukaryotes. These selenoproteins have already been discussed above. Other prokaryotic selenoproteins are discussed below. Formate dehydrogenase (FDH). This is the most abundant and widespread prokaryotic selenoprotein [56]. Sec in this protein is coordinated to molybdenum and directly involved in the oxidation of formate to carbon dioxide [57,58]. In many bacteria, FDH is the only selenoprotein, which may be responsible for maintaining the Sec trait in these organisms [59]. Hvdrogenase. Several hydrogenases are known that contain Sec. In these proteins, Sec is bound to nickel and is directly involved in catalysis [60]. Two different hydrogenase subunits may contain Sec, including one which may have two Sec residues [61]. Formvlmethanofuran dehydrogenase (FMDH). FMDH is a distant homolog of FDH and catalyzes a similar reaction (with formylmethanofuran as the substrate) [62]. As in FDH, Sec in FMDH is coordinated to molybdenum in the enzyme active site.

106

Selenium: Its molecular biology and role in human health

Selenoprotein A (GrdA). GrdA is a selenoprotein component of a multiprotein glycine reductase complex in certain bacteria [63]. This is the only known prokaryotic selenoprotein for which no Cys homologs can be detected [38]. Selenoprotein B ("GrdB'). GrdB is a selenoprotein component of multiprotein complexes involved in the reduction of glycine, sarcosine, betaine and other substrates [64-66]. Known GrdB proteins are substratespecific and bind a single GrdA. Peroxiredoxin (Prx). Peroxiredoxins are abundant Cys-containing proteins that are present in essentially all organisms. Some bacteria contain Seccontaining Prxs [67]. However, these selenoproteins have not been functionally characterized. Thioredoxin (Trx). Trx is the major intracellular protein disulfide reductant. It occurs in all organisms and is often an essential protein. Some bacterial Trx homologs are selenoproteins [15,16]. Glutaredoxin (Grx). Like Trx, it is a well studied protein disulfide oxidoreductase. Grx function is dependent on glutathione. Some bacterial Grxs are selenoproteins [15,16]. HesB-like. This distant homolog of HesB proteins (also known as IscA) is a selenoprotein present in certain archaea and bacteria [15]. HesB/IscA proteins are involved in iron-sulfur cluster biosynthesis, but the function of their selenoprotein homolog has not been characterized. Additional prokaryotic selenoproteins. Additional selenoproteins are listed in Figure 1. Most of these proteins are homologs of thiol-dependent oxidoreductases, in which the catalytic Cys is replaced with Sec [15,16]. Selenoprotein functions From the brief description of selenoprotein functions, it is apparent that selenoproteins for which functions are known are redox proteins. In these proteins. Sec is the catalytic residue that is employed because of its strong nuclephilicity and low pKa [38,68]. Sec reversibly changes its redox state during catalysis. Functions of many selenoproteins, particularly those found in vertebrates, are not known. However, by analogy to proteins with known functions, it may be expected that the majority of these uncharacterized selenoproteins are also redox proteins. All selenoproteins may be loosely clustered into three protein groups. The most abundant selenoprotein group includes proteins containing Sec in the N-terminal regions, followed by an a-helix. Many of these selenoproteins exhibit thioredoxin or thioredoxin-like folds, but some proteins (e.g., MsrA) show different folds. In these proteins, Sec is the catalytic group, which often works in concert with a resolving Cys. In the second group. Sec is located in the C-terminal sequences. These proteins so far have been described only in eukaryotes and include

Selenoproteins and selenoproteomes

107

selenoproteins K, S, O, I and thioredoxin reductases. In thioredoxin reductases, Sec is the redox residue and a key component of the catalytic Cterminal Gly-Cys-Sec-Gly tetrapeptide. The function of Sec in other selenoproteins in this group is not known. Selenoproteins in the third group utilize Sec to coordinate redox metals (molybdenum, tungsten and nickel) in the active sites of these proteins. This protein class includes hydrogenase, formate dehydrogenase and formylmethanofuran dehydrogenase, which are found in only in prokaryotes. Catalytic advantages and disadvantages of selenocysteine Sec residues serve important catalytic functions in selenoenzymes, whereas mutants in which Sec is replaced with Cys are 100-1,000 fold less active (but interestingly, these mutants do show detectable activity) [69]. In addition, comparison of catalytic functions of selenoproteins and their natural Cyscontaining homologs revealed that selenoproteins often exhibit higher catalytic activities [38]. It was also found that, in some natural Cyscontaining proteins, replacement of catalytic Cys residues with Sec can increase the enzyme activity [70]. All these studies are consistent with the idea that Sec provides catalytic advantages over Cys in enzyme active sites. If so, why then is Sec not used universally to maximize catalytic efficiency of thiol-based redox catalysts? It was found that presence of Sec does not always lead to highly active catalysts. For example, Drosophila thioredoxin reductase is a Cys-containing protein, yet its activity is nearly equivalent to that of mammalian thioredoxin reductases, which are selenoproteins [71]. Instead, Sec in these proteins was proposed to provide a broader range of substrates and a broader range of microenvironmental conditions in which thioredoxin reductase activity is possible. More recently, this question was addressed using mammalian methionineR-sulfoxide reductases, one of which is a selenoprotein whereas the other two homologs contain Cys. It was found that Sec- and Cys-containing methionine-R-sulfoxide reductases show important differences in their catalytic mechanisms, and that having Sec has both catalytic advantages (high activity with some substrates) and disadvantages (dependence on a unique resolving Cys that assists the catalytic Sec) [34]. The use of Se should also be viewed in context of availabihty of selenium in the environment. Clearly, further studies are required to fully understand the reason why Sec is used or not in proteins. Selenoproteomes As discussed above, bioinformatics analyses allowed the identification of all or almost all selenoproteins in a variety of organisms [11]. These data about full sets of selenoproteins in organisms (selenoproteomes) are proving to be

108

Selenium: Its molecular biology and role in human health

extremely useful in addressing numerous questions relevant to the biology of selenium. This information should allow, for the first time, to fully explain biological and biomedical effects of dietary selenium. This is because it is now possible to link individual selenoproteins or selenoprotein groups (e.g., stress-related selenoproteins, housekeeping selenoproteins, etc.) to the specific effects of dietary selenium. In this respect, selenium is ahead of studies involving other trace elements (and biofactors) where new metalloproteins are still discovered biochemically and often by accident. For example, as discussed in Chapters 13, 22-24 and 32, selenoproteins are implicated in the cancer prevention effect of selenium. However, which selenoproteins are involved is a matter of continuing investigation. Of the 25 human selenoproteins [14], several proteins may likely be excluded due to their known functions or limited expression patterns, leaving only a handful of candidate selenoproteins, which individually, or in combination, may possess cancer prevention activities. Analyses of selenoproteomes also provide important insights into the requirement of selenium for various organisms as well as explaining Sec evolution. For example, formate dehydrogenase is present in essentially all prokaryotes that contain the machinery for Sec biosynthesis and insertion, suggesting that this protein is responsible for maintaining the Sec trait in these organisms [15,16,59]. Searches of nematode selenoproteomes revealed that C. elegans and C. briggsae likely have only a single UGA codon that codes for Sec in their genomes [72]. This codon is present in the thioredoxin reductase gene, and phylogenetic analyses suggested that other selenoprotein genes were lost in these nematodes during evolution. Literestingly, fiiiit flies have a Cyscontaining thioredoxin reductase, suggesting that this protein may compensate for the lack of selenoprotein thioredoxin reductase. Thus, it would not be surprising if an animal is identified in the future that lacks selenoproteins altogether. In turn, information about such animals (or other organisms that lost selenoproteins, such as yeast and higher plants) should help explain the changing requirements for selenium during evolution. Recent characterization of selenoproteomes of nematodes [72], fruit flies [12,13], mammals [14] and other vertebrates [52,53], and Apicomplexan parasites [55], as well as those of numerous bacteria and archaea [15,73], provided many clues in regard to the use of selenium in these organisms. Rapid progress in genome sequencing should allow application of previously developed bioinformatics tools to many additional genome projects. Recent studies show that large scale environmental genome surveys are also amenable to these applications [16]. And while the selenoprotein content of key model genomes is, in general, well characterized, the analysis of additional genomes, often uncovers novel selenoprotein families, suggesting that we only know afi^actionof selenoproteins in Nature.

Selenoproteins and selenoproteomes

109

Conclusions A decade ago, only several selenoproteins were known. Largely due to remarkable progress in genomic research and bioinformatics, we now have information on more than 40 selenoprotein families. In selenoproteins with known functions, Sec is a key functional group that carries out redox catalysis. Therefore, identification of each new selenoprotein provides information on the possible role of this protein in redox biology and identifies the candidate catalytic group in this protein. Importantly, this information links selenium to an ever expanding universe of biological functions dependent on this trace element. Further studies on identity and functions of selenoprotein genes should help explain known biological and biomedical effects of selenium and identify new biological processes and pathways dependent on this trace element. Acknowledgments This work is supported by NIH GM061603. References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27.

GL Dilworth 1982 Arch Biochem Biophys 219:30 VN Gladyshev, SV Khangulov, TC Stadtman \994 Proc Natl Acad Sci USA 91:232 VN Gladyshev, SV Khangulov, TC Stadtman 1996 Biochemistry 35:212 WT Self, TC Stadtman 2000 Proc Natl Acad Sci USA 97:7208 D Behne, A Kyriakopoulos, H Meinhold, J Kohrle 1990 Biochem Biophys Res Commun 173:1143 T Tamura, TC Stadtman 1996 Proc Natl Acad Sci USA 93:1006 VN Gladyshev, KT Jeang, JC Wootton, DL Hatfield 1998 J Biol Chem 273:8910 SV Novoselov, M Rao, NV Onoshko et al 2002 EMBO 721:3681 GV Kryukov, VM Kryukov, VN Gladyshev 1999 J Biol Chem 274:33888 A Lescure, D Gautheret, P Carbon, A Krol 1999 J Biol Chem 274:38147 VN Gladyshev, GV Kryukov, DE Fomenko, DL Hatfield 2004 Amu Rev Nutr 24:579 FJ Martin-Romero, GV Kryukov, AV Lobanov et al 2001 J Biol Chem 276:29798 S Castellano, N Morozova, M Morey et al 2001 EMBO Rep 2:697 GV Kryukov, S Castellano, SV Novoselov et al 2003 Science 300:1439 GV Kryukov, VN Gladyshev 2004 EMBO Rep 5:538 Y Zhang, DE Fomenko, VN Gladyshev 2005 Genome Biol 6:R37 JT Rotruck, AL Pope, HE Ganther et al 1973 Science 179:588 LJ Yant, Q Ran, L Rao et al 2003 Free Radic Biol Med 34:496 F Ursini, S Heim, M Kiess, M Maiorino, A Roveri, J Wissing, L Flohe 1999 Science 285:1393 VN Gladyshev, KT Jeang, TC Stadtman 1996 Proc Natl Acad Sci USA 93:6146 S Gromer, J Wissing, D Behne D et al 1998 Biochem J 332:591 QA Sun, L Kimarsky, S Sherman, VN Gladyshev 2001 Proc Natl Acad Sci USA 98:3673 T Sandalova, L Zheng et al 2001 Proc Natl Acad Sci USA 98:9533 QA Sun, Y Wu, F Zappacosta et al 1999 y Biol Chem 274:24522 ES Amer, A Holmgren 2000 Eur J Biochem 267:6102 QA Sun, F Zappacosta, V Factor et al 2001 J Biol Chem 276:3106 AK Rundlof, M Janard, A Miranda-Vizuete, ES Amer 2004 Free Radic Biol Med 36:641

110 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73.

Selenium: Its molecular biology and role in human health D Su, VN Gladyshev 2004 Biochemistry AZ-MXll D Su, SV Novoselov, QA Sun et al 2005 J Biol Chem 280:26491 M Conrad, C Jakupoglu, SG Moreno et al 2004 Mol Cell Biol 24:9414 C Jakupoglu, GK Przemeck, M Schneider et al 2005 Mol Cell Biol 25:1980 GV Kryukov, RA Kumar, A Koc et al 2002 Proc Natl Acad Sci USA 99:4245 HY Kim, VN Gladyshev 2004 Mol Biol Cell 15:1055 HY Kim, VN Gladyshev 2005 PLoSBiol 3:e375 KV Korotkov, E Kumaraswamy, Y Zhou et al 2001 J Biol Chem 276:15330 E Kumaraswamy, A Malykh, KV Korotkov et al 2000 J Biol Chem 275:35540 AD Ferguson, VM Labunskyy, DE Fomenko et al 2006 J Biol Chem 281:3536 TC Stadtman 1996 Annu Rev Biochem 65:83 MJ Guimaraes, D Peterson, A Vicari et al 1996 Proc Natl Acad Sci USA 93:15086 RF Burk, KE Hill 2005 Annu Rev Nutr 25:215 GV Kryukov, VN Gladyshev 2000 Genes Cells 5:1049 KE Hill, J Zhou, WJ McMahan et al 2003 J Biol Chem 278:13640 L Schomburg, U Schweizer, B Holtmann et al 2003 Biochem J 370:397 SC Vendeland, MA Beilstein, JY Yeh, W Ream, PD Whanger 1995 Proc Natl Acad Sci USA 92:8749 KV Korotkov, SV Novoselov, DL Hatfield, VN Gladyshev 2002 Mol Cell Biol 22:1402 Y Ye, Y Shibata, C Yun, D Ron, TA Rapoport 2004 Nature 429:841 JE Curran, JB Jowett, KS Elliott et al 2005 Nat Genet 37:1234 B Moghadaszadeh, N Petit, C Jaillard 2001 Nat Genet 29:17 H Weissbach, L Resnick, N Brot N 2005 Biochim Biophys .4cto 1703:203 SV Novoselov, M Rao, NV Onoshko et al 2002 EMBO y 21:3681 T Obata, Y Shiraiwa 2005 J Biol Chem 280:18462 S Castellano, SV Novoselov, GV Kryukov et al 2004 EMBO Rep 5:71 S Cd&teWaao, AM Lohmoy tt&X 2005 Proc Natl Acad Sci USA 102:16188 SV Novoselov, D Hua, AV Lobanov, VN Gladyshev 2005 Biochem J, in press AV Lobanov, CDelgado, S Rahlfs et al 2006 Nucl Acids Res 34:496 F Zinoni, A Birkmann, TC Stadtman, A Bock 1986 Proc Natl Acad Sci USA 83:4650 VN Gladyshev, SV Khangulov, MJ Axley, TC Stadtman 1994 Proc Nad Acad Sci USA 91:7708 JC Boyington, VN Gladyshev et al 1997 Science 275:1305 H Romero, Y Zhang, VN Gladyshev, G Salinas 2005 Genome Biol 6:R66 E Garcin, X Vemede, EC Hatchikian et al 1999 Structure 7:557 R Wilting, S Schorling, BC Persson, A Bock 1997 J Mol Biol 266:637 JA Vorholt, M Vaupel, RK Thauer 1997 Mol Microbiol 23:1033 GE Garcia, TC Stadtman 1991 JBacteriol 173:2093 M Wagner, D Sonntag, R Grimm et al 1999 Eur J Biochem 260:38 JR Andreesen, M Wagner, D Sonntag et al 1999 Biofactors 10:263 T Schrader, A Rienhofer, JR Andreesen 1999 Eur J Biochem 264:862 B Sohling, T Farther, KP Rucknagel, MA Wagner, JR Andreesen 2001 Biol Chem 382:979 DL Hatfield, VN Gladyshev 2002 Mol Cell Biol 22:3565 MJ Axley, A Bock, TC Stadtman 1991 Proc Nad Acad Sci USA 88:8450 S Hazebrouck, L Camoin, Z Faltin et al 2000 J Biol Chem 275:28715 S Gromer, L Johansson, H Bauer et al 2003 Proc Nad Acad Sci USA 100:12618 K Taskov, C Chappie, GV Kryukov et al 2005 Nucleic Acids Res "iZ-.llTI Y Zhang, V Gladyshev, 2005 Bioinformatics 21:25 80

Chapter 10. Deletion of selenoprotein P gene in the mouse Raymond F. Burk, Gary E. Olson and Kristina E. Hill Division of Gastroenterology, Hepatology, and Nutrition, Department of Medicine (R.F.B. and K.E.H.) and Department of Cell and Developmental Biology (G.E.O.), Vanderbilt University School of Medicine, Nashville, TN 37232, USA

Summary: Selenoprotein P is a selenium-rich extracellular protein that has a number of unusual characteristics. The recent production of mice with selenoprotein P deleted {SeppV'' mice) has fostered studies of its function. Selenoprotein P has been shown to have a role in whole-body selenium homeostasis. When fed diets containing selenium at its required concentration, Seppl'^' mice have lower tissue selenium concentrations than do Seppl^'^ mice, except in the liver. Thus, selenoprotein P appears to distribute selenium from the liver to other tissues. Also, SeppV' mice develop neurological dysfunction and have male infertility. Both brain and testis in them have very low selenium concentrations and appear to be selenium deficient. Feeding Seppl'' mice high dietary levels of selenium prevents most of the neurological dysfunction but does not raise brain selenium concentration or prevent brainstem axonal degeneration. Thus, the functions of selenoprotein P in the central nervous system appear to be complex and are not explained by a single mechanism. Spermatozoa produced by Seppl'' mice are morphologically similar to those of selenium deficient Seppl^ * mice. They lack a mitochondrial sheath covering the distal midpiece. Absence of the sheath causes axonemal disruption that leads to a hairpin bend at the midpiece-principal piece junction. The resulting spermatozoa have markedly reduced motility, presumably a cause of their ineffectiveness.These findings suggest that the primary function of selenoprotein P in the testis is to provide selenium to be used in sperm production. Introduction In 1982, the second animal selenoprotein to be recognized was given the name selenoprotein P because of its plasma location [1,2]. Subsequent work has shown that selenoprotein P is a single polypeptide that has two domains with respect to selenium content (reviewed in [3]). The iV-terminal domain is the larger and contains one selenocysteine residue in a UxxC redox motif.

112

Selenium: Its molecular biology and role in human health

compatible with it possibly having an enzymatic function [4]. A heparinbinding site, A'-linked carbohydrate, and two stretches of basic amino acids (largely histidine residues) are also present in this domain [5-7]. The Cterminal domain of the rat protein contains 9 selenocysteine residues and Olinked carbohydrate. Four isoforms of selenoprotein P are present in rat plasma [8]. All share the same amino acid sequence beginning at the A'^ terminus, but the three shorter ones terminate at the positions of internal selenocysteine residues. One isoform is the full-length protein. Another is the A'-terminal domain: it terminates at the position of the second selenocysteine residue. Two others terminate at the positions of the third and seventh selenocysteine residues, respectively. There is controversy over whether isoforms of human selenoprotein P exist [4,9,10] and no reports on isoforms of mouse selenoprotein P have appeared. As yet, no unique function has been attributed to a selenoprotein P isoform. Selenoprotein P concentration in plasma is sensitive to selenium intake in the nutritional range [11,12]. It is likely to become the preferred plasma biomarker for assessing human selenium nutritional status when an assay for it becomes available for general use. The selenoprotein P in plasma is synthesized largely in the liver [13] and suggestions have been made that it serves to transport selenium from the liver to other tissues [2,14]. Li addition, selenoprotein P mRNA is found in virtually all tissues. This suggests that the protein has a function in the interstitial space. In the previous edition of this book, we proposed, based on our research findings and those of others [15], that selenoprotein P scavenges peroxynitrite in the interstitial space [16]. Reviews that focus on various aspects of selenoprotein P have appeared in recent years and the reader is referred to them [3,17,18]. Progress in determining the function(s) of selenoprotein P has been accelerated in the last few years by the production of mice with deletion of the selenoprotein P gene (Seppl). This chapter will focus on studies of these mice. Mice with deletion of tlie selenoprotein P gene Two research groups, that of Kohrle in Germany and ours, produced selenoprotein P knockout (SeppV^) mice independently and reported on them in early 2003 [19,20]. The initial experiments indicated that selenoprotein P has a major role in selenium homeostasis. Its deletion leads to altered distribution of selenium in the animal. It has been difficult to determine whether the SeppJ'^' phenotypes that have been observed are caused by this maldistribution of selenium or by lack of a more direct function of selenoprotein P. Distinguishing between these two possibilities remains a challenge.

Deletion of selenoprotein P in the mouse

113

Each group has published follow-up studies that address selenium disposition and neurological function in these mice [21-23]. In addition, our group has examined spermatogenesis in them [24]. Effects of selenoprotein P deletion on selenium disposition The initial reports of the knockouts indicated that loss of selenoprotein P affected the selenium content and selenoprotein enzyme activity of most tissues [19,20]. There was reasonable agreement of results between the research groups even though the Kohrle group fed their mice a chow diet and our group used a selenium-deficient Torula yeast-based diet supplemented with graded amounts of selenium. The knockout of selenoprotein P depressed selenium levels in all tissues investigated except liver. Testis and brain selenium were depressed to the greatest extent (Figure 1). Kidney and heart selenium concentrations were depressed to a lesser extent (unpublished observations, Burk, R.F. and [19]). Liver selenium concentration was not affected by the knockout except when the animals were fed selenium at or below its nutritional requirement (Figure 1 and [19]). Under those conditions, liver selenium was better preserved in Seppl'^' mice than in SeppI*^* mice. This observation set liver apart from other tissues as not requiring selenoprotein P to maintain its selenium—or even as benefiting in this respect from selenoprotein P deletion. 1500

nn seppiy-)

1

rn

•

Sepp1(+I+)

-S» 1000 CO

m

500

iid liver

brain

testis

M liver

0.10

brain

testis

0.50 dietary Se (mg/kg)

Figure 1.Tissue selenium levels in SeppI'^' and Seppl^'* mice fed Torula yeast-based diets supplemented with selenite for 4 weeks beginning at weaning. Values are means ± SD, n=4-7. Asterisks signify significant differences (p<0.05) by the Student t-test.

114

Selenium: Its molecular biology and role in human health

Both groups concluded that the most direct explanation of these results was that selenoprotein P synthesized in the liver was involved in the distribution of selenium to peripheral tissues. This conclusion agreed with earlier suggestions that were based on experiments that tracked administered ^^Se [2,14]. Our experiments with SeppV^' mice included feeding diets with varying selenium content. Kidney and heart selenium levels of Seppl'^' mice normalized when dietary selenium was increased, whereas selenium levels in testis and brain were more resistant. This pointed to possible differences in mechanisms by which different tissues acquire their selenium. It suggested that brain and testis require the presence of selenoprotein P to take up selenium and that kidney and heart could readily take up another form, which becomes more abundant when selenium intake is increased, even in Seppl'^' mice. In order to determine the effect of liver-produced selenoprotein P, the Kohrle group investigated mice with liver-specific inactivation of Trsp, the gene for selenocysteine tRNA [25]. Earlier work with these mice had shown that liver selenoproteins were drastically decreased and that ^'Se labeling of plasma selenoprotein P was depressed to about 25% [13]. This means that some selenoprotein P was present in the plasma of the mice, even though liver Trsp had been inactivated. The amount of selenoprotein P in plasma was not directly assessed. Mice with liver-specific inactivation of Trsp had normal brain selenium levels but depressed kidney selenium levels [25]. Moreover, these mice did not have neurological dysfunction as had been observed in Seppl'' mice. The K6hrle group interpreted these results to indicate that preservation of brain selenium is effected by selenoprotein P synthesized and retained in the brain (presumably unaffected in these mice) and not by selenoprotein P in the plasma (depressed but not absent in these mice). This conclusion is important because it implies that receptor-mediated uptake of selenium in the form of selenoprotein P by the brain is not the mechanism that allows the brain to maintain its selenium in severe selenium deficiency. At least one other interpretation of the results obtained by the Kohrle group using mice with liver-specific inactivation of Trsp is possible: the mice were selenium deficient. Selenium deficiency might have resulted fi"om their inability to export selenium fi-om the liver in the form of selenoprotein P. Deletion of selenoprotein P, which includes loss of production of it by liver, leads to disruption of selenium homeostasis with wasting of selenium in the urine [26]. Because the liver is the organ that makes the excretory metabolites of selenium, loss of its ability to make selenoprotein P would be expected to increase liver selenium and result in an increased production of the excretory metabolites. This would result in lower whole-body selenium and the decrease in kidney selenium observed.

Deletion of selenoprotein P in the mouse

115

Further, it can be argued that use of the liver-specific deletion of the Trsp model does not allow conclusions about the mechanism of brain selenium preservation because some selenoprotein P is present in plasma in these mice, just as it is in selenium deficient mice. It appears that more work is needed before the relative roles of plasma and brain selenoprotein P can be assigned in the selenoprotein P-mediated brain retention of selenium. In summary, selenium homeostasis depends on selenoprotein P. This protein appears to be the major, but not the only, form of selenium exported by the liver to peripheral tissues. Testis and brain appear to have a greater dependence on selenoprotein P for their selenium than do other tissues. This raises the possibihty that selenoprotein P supplies selenium to tissues by more than one mechanism. Effects of selenoprotein P deletion on the brain In addition to having low brain selenium concentrations, Seppl'' mice were noted by both groups to develop neurological dysfunction [19,20]. The dysfunction was clinically obvious only when animals were fed <0.25 mg selenium per kg (low-selenium) diet [19,21]. The initial clinical manifestation was a wide, clumsy gait with a short stride. Some mice walked only backwards in a poorly coordinated way. As the condition progressed, the mice became spastic and uncoordinated. Episodes of hyperactivity occurred, sometimes with the animal colliding against the side of its cage. Finally, mice became so spastic that they could not move purposefully. They could not eat or drink adequately and weight loss occurred. The Kohrle group observed animals having seizures [27] but we did not. Once neurological impairment had occurred in Seppl'' mice fed a lowselenium diet, feeding them a high-selenium (1 mg selenium per kg) diet failed to reverse the neurological dysfunction but did prevent its worsening [21]. This suggests that structural damage to the brain and not a metabolic effect underlies the impairment. Neurotesting oiSeppl'' mice revealed that they had a shorter stride length than Seppl''''' mice, even when fed a high-selenium diet [21]. Thus, measurable neurological dysfunction was present in Seppl'' mice no matter how much selenium they were fed. This clinical finding correlates with the presence of degenerating axons in the brainstem and spinal cord of Seppl'' mice, regardless of their selenium intake [23]. In summary, Seppl'' mice fed a low-selenium diet had brain selenium levels and brainstem axonal degeneration that were indistinguishable fi-om Seppl'' mice fed a high-selenium diet. Yet, the high-selenium diet prevented the severe neurological dysfunction that occurred when the low-selenium diet was fed to these mice. These divergent observations hint at the complexity of selenium and selenoprotein functions in the brain.

116

Selenium: Its molecular biology and role in human health

Effects of selenoprotein P deletion on the testis Selenium is an essential dietary micronutrient for the maintainence of male fertility. Early studies established that epididymal spermatozoa of seleniumdeficient mice and rats possessed various tail defects, including abnormal angulations, fracture, and separation of the axial filaments and disorganization of the mitochondrial sheath [28-30]. These defects result in the reduction or complete absence of sperm motility. Selenium is concentrated in the midpiece segment [31], where it is incorporated into a 20-kDa structural protein of the disulfide bond-stabiUzed mitochondrial capsule, a complex associated with the outer mitochondrial membrane [32]. Recently proteomic analyses identified this polypeptide as phospholipid hydroperoxide glutathione peroxidase (GPX4), the major sperm selenoprotein [33]. GPX4 is synthesized during spermiogenesis, but it is during post-testicular sperm development in the epididymis that it becomes disulfide (or selenenylsulfide) bond cross-linked, to form the insoluble mitochondrial capsule. This process is accompanied by a loss of GPX4 catalytic activity. The function of the mitochondrial capsule is poorly understood, but reduced sperm GPX4 is linked to infertility in men [34]. Moreover, in animal models selenium deficiency is accompanied by an increased incidence of abnormally shaped sperm mitochondria [35], possibly reflecting an altered mitochondrial capsule. Since sperm selenium is predominantly present in the mitochondrial capsule, an unresolved problem has been to explain the structural basis for the array of flagellar defects that result from dietary selenium deficiency. Spermatozoa released from the seminiferous epithelium of the testis possess a full complement of organelles but they are functionally immature, lacking fertilizing ability and progressive motility [36]. These capacities develop during post-testicular maturation and are expressed by mature spermatozoa stored in the Cauda region of the epididymis. Recently we examined both selenium-deficient rats and mice to determine if a temporal development of flagellar structural abnormalities occurs during spermiogenesis and/or post-testicular matviration and to determine whether a common lesion leads to the extensive disruption of flagellar architecture [24,37]. Our studies suggest that in both species a common defect present in testicular spermatozoa of selenium deficient animals leads to progressive flagellar disruption as sperm traverse the epididymis (Figures 2 and 3). Comparision of testicular spermatozoa from selenium replete and from selenium deficient animals showed that both had the normal arrangement of flagellar cytoskeletal elements including the axoneme, connecting piece, outer dense fibers and fibrous sheath (Figure 2c). A difference between the groups in sperm flagellar morphology was detected in the organization of the mitochondria. Sperm mitochondria are organized into the mitochondrial sheath that surroimds the flagellar cytoskeletal elements over the full length

117

Deletion of selenoprotein P in the mouse

~7

mp

pP cimp PP

f

'•'•%.

•. ,-f

•

••

' •• """'Wi

t

^

Figure 2, See legend on facing page.

I

mp...

118

Selenium: Its molecular biology and role in human health

Figure 2. Photomicrographs and electron micrographs of spermatozoa from male rats with severe selenium deficiency, (a) Phase contrast photomicrograph of caput epididymal spermatozoon. Note the narrowing of the distal midpiece (dmp) that reflects the premature termination of the mitochondrial sheath, mp = midpiece; pp = principal piece, (b) Phase contrast photomicrograph of cauda epididymal spermatozoon. Note the abrupt narrowing of the distal midpiece. The underlying cytoskeletal fibers (arrows) have been extruded from this segment of the midpiece. (c) Electron micrograph showing cross sections at various levels of the flagellum of late spermatids. Note that in the midpiece segment the mitochondria (m) wrap around the central axoneme and outer dense fibers, which display normal architecture. However, in this selenium deficient animal the distal midpiece is atypical and lacks the overlying mitochondria, (d) Longitudinal section of caput epididymal spermatozoon. Note the abrupt termination of the mitochondrial sheath at the point of the last mitochondria. This termination is prior to the junction with the principal piece, leaving the distal midpiece without a mitochondrial sheath, (e) Electron micrograph showing cross sectioned flagella of Cauda epididymal spermatozoa. Note the extensive disruption of flagellar geometry in both the principal piece and midpiece segments, reflecting the extrusion of specific doublet microtubules and outer dense fibers.

of the midpiece segment. Typically the mitochondrial sheath terminates posteriorly at the ring-like annulus that is affixed to the proximal margin of the fibrous sheath, an organelle that is restricted to the principal piece segment. In selenium deficient rats and mice the mitochondrial sheath terminates prematurely so its distal end does not reach the annulus, resulting in exposure of the underlying outer dense fibers and axoneme (Figures 2a-d). Examination of spermatozoa fi^om the caput and cauda regions of the epididymis indicates a key role for the mitochondria-deficient, terminal segment of the midpiece in the flagellar disruption that is characteristic of selenium deficient animals. Spermfi-omthe caput epididymis of selenium deficient animals fi-equently display initiation of abnormal tail bending at the midpiece-principal piece junction, as well as the extrusion of axial fibers from the mitochondria-free segment of the midpiece. Electron microscopic analyses demonstrate that a specific subset of outer dense fibers and doublet microtubules is extruded at this site. This indicates that the absence of the overlying mitochondrial sheath generates a weak site in the midpiece permitting flagellar disintegration. Observations on sperm from the cauda epididymis of selenium deficient animals demonstrate that a progressive disruption of flagellar geometry occurs by specific extrusion of outer dense fibers and doublet microtubules numbers 4 through 7 fi-om the terminal midpiece. In rats this filament extrusion process ruptures the overlying plasma membrane so that the spermatozoa are nonviable and immotile (Figure 2b). However in mouse spermatozoa, which possess smaller outer dense fibers than the rat, the extruded fibers do not disrupt the plasma membrane but become interposed between it and the mitochondrial sheath (Figures 3a-d). As a result of the

Deletion of selenoprotein P in the mouse

O

i.J

119

••

•••••••••••••''"""""

j

^

" t ..?>'^/3

mp pp

^Tiw,

"^-pm

^

Figure 3. See legend on facing page.

120

Selenium: Its molecular biology and role in human health

Figures. Photomicrographs and electron micrographs of spermatozoa from selenium deficient SeppI*'* mice and Seppl''' mice fed a selenium adequate diet, (a) Phase contrast photomicrograph of cauda epididymal spermatozoa of a selenium deficient Seppl*'* mouse. Note the majority of spermatozoa display a hairpin flagellar bend at the midpiece-principal piece junction (arrows). One sperm with an extended flagellum is evident and shows narrowing of the distal midpiece (dmp) that results from premature termination of the mitochondrial sheath, (b) Phase constrast photomicrograph of cauda epididymal spermatozoa of & Seppl'' mouse showing that the majority of spermatozoa display a hairpin flagellar bend (arrows) like that seen in selenium deficient mice, (c) Electron micrograph showing a cross section through the hairpin-bend flagellum of a cauda epididymal sperm from a selenium deficient mouse. Note the profile of the midpiece (mp) and principal piece (pp) contained within an intact plasma membrane (pm). The principal piece shows a disorganized flagellar architecture. The arrows point to doublet microtubules and outer dense fibers that have been extruded from the terminal midpiece segment, (d) Electron micrograph showing hairpinned flagellum of a cauda epididymal spermatozoon from Seppl'' mouse and a number of principal pieces with disorganized flagellar architecture.

fiber extrusion, most mouse spermatozoa display a hairpin bend at the midpiece-principal piece junction (Figure 3a), but since the plasma membrane remains intact, many spermatozoa still exhibit a feeble motility. Seppl' mice display a male infertility phenotype [19]. Recently we demonstrated an absolute requirement for this serum selenoprotein in sperm development [24]. Spermatozoa from SeppV^ mice display a temporal development of flagellar abnormalities that appear identical to those seen in selenium deficient wild type mice. Testicular spermatozoa of SeppV^' mice possess a truncated mitochondrial sheath and cauda epididymal spermatozoa display a hairpin bend flagellar configuration and a pattern of axonemal disruption that is indistinguishable from that seen in selenium deficient wild type animals (Figures 3b and 3d). Moreover, the testis selenium level of the Seppl'' male is dramatically reduced and comparable to that of the selenium deficient wild type animal. These findings suggest that selenoprotein P plays a crucial role in selenium delivery to developing germ cells during spermiogenesis. Therefore, identification of the mechanisms that frafific selenium in this plasma selenoprotein through the blood-testis barrier to the developing germ cells are critical to our understanding of sperm development. Research needs Studies with the selenoprotein P null mouse have allowed the identification of several phenotypes related to absence of the protein. It is now clear that selenoprotein P plays an important role in whole-body selenium homeostasis. Brain and testis depend on it for their supply of selenium and cannot function normally when it is absent. Determination of the mechanism(s) by which selenoprotein P affects tissue selenium concentrations is needed. Are there receptors for it on certain cells?

Deletion of selenoprotein P in the mouse

121

Is its C terminus cleaved off and taken up as a source of selenium? Does it serve as a reservoir of selenium in the plasma and within such organs as the brain? Determination of whether selenoprotein P has an activity is needed. Does the apparent redox site in its A'^ terminus have enzymatic activity? Is it able to use the reducing equivalents implied by its selenocysteine and cysteine residues to defend the extracellular space against reactive chemical species? Does it bind to cells through its heparin-binding site? Does it perform an enzymatic function? Despite the advances made over the last few years in understanding selenoprotein P, many questions remain to be investigated. Acknowledgements The authors thank Virginia Winfrey for preparing Figures 2 and 3. Research by the authors is supported by NIH grants ES02497, DK58763 and HD44863. References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23.

RF Burk, PE Gregory 1982 Arch Biochem Biophys 213:73 MA Motsenbocker, AL Tappel 1982 Biochim Biophys Acta 719:147 RF Burk, KE Hill 2005 Annu Rev Nutr 25:215 Y Saito, T Hayashi, A Tanaka, Y Watanabe, M Suzuki, E Saito, K Takahashi 1999 J Biol Chem 274:2866-2871 S Himeno, HS Chittum, RF Burk \996 J Biol Chem 27\:\5769 RJ Hondal, S Ma, RM Caprioli, KE Hill, RF Burk 2001 J Biol Chem 276:15823 S Ma, KE Hill, RF Burk, RM Caprioli 2003 Biochemistry 42:9703 S Ma, KE Hill, RM Caprioli, RF Burk 2002 J Biol Chem 111: 12749 B Akesson, T Bellew, RF Burk 1994 Biochim Biophys Acta 1204: 243 V Mostert, I Lombeck, J Abel 1998 Arch Biochem Biophys 357:326 Y Xia, KE Hill, DW Byrne, J Xu, RF Burk 2005 Am J Clin Nutr 81:829 J-G Yang, KE Hill, RF Burk 1989 J Nutr 119:1010 BA Carlson, SV Novoselov, E Kumaraswamy, BJ Lee, MR Anver, VN Gladyshev, DL Hatfield 2004 J Biol Chem 279:8011 RF Burk, KE Hill, R Read, T Bellew 1991 Am J Physiol 261 :E26 H Sies, GE Arteel 2000 Free Radio Biol Med 28:1451 K Hill, R Burk 2001 Selenium: Its molecular biology and role in human health (Ed. Hatfield D) Kluwer Academic Publishers, Boston pi 23 V Mostert 2000 Arch Biochem Biophys 376:433 M Persson-Moschos 2000 Cell Mol Life Sci 57:1836 KE Hill, J Zhou, WJ McMahan, AK Motley, JF Atkins, RF Gesteland, RF Burk 2003 J Biol Chem 218:12640 L Schomburg, U Schweizer, B Holtmann, L Flohe, M Sendtner, J Kohrle 2003 Biochem 7370:397 KE Hill, J Zhou, WJ McMahan, AK Motley, RF Burk 2004 J Nutr 134: 157 U Schweizer, M Michaelis, J Kohrle, L Schomburg 2004 Biochem J 378:21 WM Valentine, KE Hill, LM Austin, HL Valentine, D Goldowitz, RF Burk 2005 Toxicol

PatholiiSlO 24. GE Olson, VP Winfl-ey, SK Nagdas, KE Hill, RF Burk 2005 Biol Reprod 73:201 25. U Schweizer, F Streckfuss, P Peh, BA Carlson, DL Hatfield, J Kohrle, L Schomburg 2005 Biochem J386:221

122 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37.

Selenium: Its molecular biology and role in human health RF Burk, KE Hill, AK Motley, M Hu, LM Austin 2004 FASEB J18: A849 U Schweizer, L Schomburg, NE Savaskan 2004 JNutr 134:707 KE McCoy, PH Weswig 1969 JNutr 98:383 E Wallace, H Calvin, G Cooper 1983 Gamete Res 4:377 AS Wu, JE Oldfield, LR ShuU, PR Cheeke 1979 Biol Reprod 20:793 DG Brown, RF Burk 1973 JNutr 103:102 HI Calvin 1978 J Exp Zool 204:445 F Ursini, S Heim, M Kiess, M Maiorino, A Roveri, J Wissing, L Flohe 1999 Science 285:1393 H Imai, K Suzuki, K Ishizaka, S Ichinose, H Oshima, I Okayasu, K Emoto, M Umeda, Y Nakagawa2001 Biol Reprod 64:674 E Wallace, G Cooper, H Calvin 1983 Gamate Res 4:389 J Bedford 1975 Handbook of Physiology: Endocrinoloy, Male Reproductive System eds R Greep and E Astwood Waverly Press, Washington, DC p303 GE Olson, VP Winfrey, KE Hill, RF Burk 2004 Reproduction 127:335

Chapter 11. reduction

Selenium and methionine sulfoxide

Hwa-Young Kim and Vadim N. Gladyshev Department of Biochemistry, University of Nebraska, Lincoln, Nebraska 68588, USA

Summary: Methionine residues in proteins can be readily oxidized to a diastereomeric mixture of methionine sulfoxides by reactive oxygen species, hi most organisms, methionine sulfoxides are reversibly and stereospecifically reduced back to methionine by two distinct classes of repair enzymes, methionine-iS-sulfoxide reductase (MsrA) and methionine-i?sulfoxide reductase (MsrB). Methionine sulfoxide reduction is thought to be an essential pathway that protects cells from oxidative stress and regulates protein function. This pathway is also implicated in delaying the aging process in organisms from yeast to mammals. The first selenoprotein identified using bioinformatics methods, SelR (also known as SelX or MsrBl), was recently found to be a selenocysteine-containing MsrB. In mammals, selenoprotein MsrBl is a major MsrB, while MsrB2 and MsrB3 contain cysteine in place of selenocysteine. It has been found that selenocysteine- and cysteine-containing MsrBs employ different catalytic mechanisms. Interestingly, a selenocysteine-containing form of MsrA was also described, but so far was only detected in green algae. Introduction Selenium is an essential trace element in humans and other mammals. It is cotranslationally incorporated into proteins in the form of the 21st amino acid, selenocysteine (Sec) [1-3]. The Sec-containing proteins, selenoproteins, are found in all three kingdoms of life. Twenty five selenoprotein genes have been identified in human and 24 in rodent genomes [4]. A limited number of selenoproteins have been characterized while functions and physiological roles of many other selenoproteins have yet to be determined. Among selenoproteins with known functions the majority are oxidoreductases, for example, glutathione peroxidase [5], thioredoxin reductase [6] and formate dehydrogenase [7]. Selenoproteins typically exhibit 100-1,000 fold higher enzyme activities than their cysteine (Cys) mutants or natural Cys-containing forms. A key reason for the use of Sec in biological systems is explained by this high catalytic activity of Sec-containing enzymes. For incorporation of Sec at in-frame UGA codons, cis- and trans-acting factors are required.

124

Selenium: Its molecular biology and role in human health

including Sec insertion sequence (SECIS) element, SECIS-binding proteins, ^j^^[Ser]Sec^ and Sec-specific elongation factor [3,8]. In 1999, initial bioinformatics methods were developed for identification of selenoprotein genes by searching for SECIS elements (see Chapter 9). The first selenoprotein identified using this approach was designated as selenoprotein R (SelR) [9]. Independently, it was described as selenoprotein X (SelX) [10]. Comparative genomic analyses were then used to link the function of SelR to the pathway of methionine sulfoxide reduction, and this prediction was verified experimentally [11]. Proteins can be oxidized by reactive oxygen species (ROS) generated in cells during stress and physiological processes. Methionine residues in proteins are among the most susceptible to oxidation by ROS and are converted by these species to methionine sulfoxide. A diastereomeric mixture of methionine-^-sulfoxide and methionine-/?-sulfoxide is generated by ROS because of the presence of chiral sulfur in methionine sulfoxides [12]. Generation of methionine sulfoxides may manifest significant structural and functional changes in proteins. However, methionine sulfoxide can be reduced back to methionine by repair enzymes, methionine sulfoxide reductases. Therefore, methionine sulfoxide reduction is thought to be an important pathway that protects cells against oxidative stress, regulates protein function, and delays the aging process [13-17]. This chapter will focus on methionine sulfoxide reduction in mammals and the role of selenium in this pathway. Methionine sulfoxide reductases Methionine sulfoxide reductases reduce firee and protein-bound methionine sulfoxides back to methionine in the presence of thioredoxin (Trx) or dithiothreitol (DTT) [18,19]. To catalyze the repair process, two distinct stereospecific enzymes are evolved. MsrA can only reduce the S epimer of methionine sulfoxide, whereas MsrB is specific for the R form of this amino acid (Figure 1). Most organisms from bacteria to humans contain MsrA and MsrB genes in their genomes. However, some hyperthermophiles and intracellular parasites do not have MsrA, MsrB, or both proteins [11]. While parasites have access to metabolic pathways of the host, the reason why certain organisms that live at high temperatures lack the methionine sulfoxide reduction system is not understood. MsrA and MsrB genes are clustered in several bacterial genomes and often form an operon. Furthermore, MsrA and MsrB activities are often detected in a single polypeptide formed via direct MsrA/MsrB fusion [11,20]. Single MsrA and MsrB genes are present in yeast (e.g., Saccharomyces cerevisiae) and many animals (e.g., Caenorhabditis elegans and Drosophila melanogaster) [11]. On the other hand, multiple MsrA and/or MsrB genes

Selenium and Methionine Sulfoxide Reduction

125

have been identified in the plant kingdom, for example in Arabidopsis thaliana [21,22] and Chlamydomonas reinhardtii [23]. Methionine-S-Sulfoxide COOH O CH-(CH2)2 MSrA

/

NH2

S^ CH3

Methionine COOH

reductant

^

CH-(CH2)2 NH2

s^

\^

oxidant(ROS)

CH3

MsrB COOH CH"(CH2)2

S^Q

NH2 CH3 Methionine-R-Sulfoxide

Figure 1. A pathway of methionine sulfoxide reduction. R and S diastereomers of methionine sulfoxide are formed directly or indirectly in free methionines and protein methionine residues in the presence of oxidants, such as hydrogen peroxide. Methionine-5-sulfoxides are reduced by MsrA and methionine-/{-sulfoxides by MsrB with reductants, such as thioredoxin.

MsrA was discovered decades ago, and its function, catalytic mechanism, and structure are well understood [18,19,24-29]. MsrB has only recently been identified and is currently being extensively characterized [11,30-34]. Mammalian methionine sulfoxide reductases Human and mouse genomes contain a single MsrA gene [35]. Mammalian MsrA has a typical N-terminal mitochondrial targeting peptide. Interestingly, however, this protein is located in cytosol and nucleus as well as in mitochondria [36-38]. Although molecular mechanisms responsible for targeting this protein to different cellular compartments are not fully understood, a recent study has shown that structural and functional elements of MsrA play a role in subcellular occurrence of this protein [38]. In contrast to a single MsrA gene [39,40], there are three MsrB genes in mammals [41]. MsrBl (also known as selenoprotein R or selenoprotein X) [9,10] is a selenoprotein in which Sec occupies the active site. This protein resides in the cytosol and nucleus. The other two MsrBs are homologous proteins in which Cys residues are present in place of Sec. MsrB2 (also known as CBS-1) [32,42] is a mitochondrial protein. Interestingly, human MsrB3 occurs in two protein forms, MsrB3A and MsrB3B. These two forms

126

Selenium: Its molecular biology and role in human health

are generated by alternative splicing of the first exon. The MsrB3A is targeted to the endoplasmic reticulum (ER), whereas MsrB3B is targeted to mitochondria [41]. However, there is no evidence for alternative splicing in the mouse MsrB3 gene. This protein contains the N-terminal ER signal followed by the mitochondrial signal sequence in a single coding region. This mouse MsrBS form resides in the ER [43]. These findings of multiple cellular locations of MsrA and MsrB suggest that different compartments in mammalian cells maintain the methionine reduction system to repair oxidized methionine residues. Physiological roles of methionine sulfoxide reductases Reversible interconversion between methionine and methionine sulfoxide residues is implicated in several biological processes. Previously proposed functions of methionine sulfoxide reductases include repair of damaged proteins, antioxidant function as scavengers of ROS, and regulation of protein function [44,45]. A number of published reports describe the role of methionine sulfoxide reductases in antioxidant defense. For example, overexpression of MsrA protected S. cerevisiae and human T cells against oxidative stress [46], and the corresponding homologs were implicated in the protection against ROS in many microorganisms, such as Neisseria gonorrhoeae [47], Staphylococcus aureus [48], and Helicobacter pylori [49]. Recently, MsrA was found to promote viability of lens cells [50] and retinal pigmented epithelial cells [51] by conferring resistance to oxidative stress. In addition, MsrA can play a protective role against hypoxia/reoxygenation-mediated neuronal cell injury [52]. Lens MsrBs were also found to play a role in resistance to oxidative stress [53]. In addition, A. thaliana MsrA was reported to repair oxidatively damaged proteins during dark by reducing methionine sulfoxides in these proteins [21]. Methionine sulfoxide reduction and aging Methionine sulfoxide reductases are directly implicated in regulation of the aging process. Research in this area has focused mainly on the effects of MsrA. Deletion of the MsrA gene in mice decreased the lifespan by 40% [54], whereas overexpression of MsrA in Drosophila extended it by 70% [55]. The data that MsrA regulates lifespan in a variety of organisms raise a possibility that MsrB may also play an important role in the aging process. In particular, an attractive hypothesis has been advanced that since MsrBl is a selenoprotein, dietary selenium supplementation may be used to increase expression of this protein. If MsrBl regulates lifespan, elevated levels of this protein due to increased dietary intake of selenium may promote longevity in certain organisms or genetic bacl^ounds.

Selenium and Methionine Sulfoxide Reduction

127

The role of MsrB in aging has recently been tested using S. cerevisiae as a model organism [56]. This protein was found to extend the yeast lifespan under caloric restriction conditions, whereas MsrA was most efficient under normal growth conditions. The lifespan extension required oxygen because it was found that neither MsrA nor MsrB regulate the lifespan of yeast cells grown under anaerobic conditions. In the future, deletion or overexpression of MsrB genes in mammals is needed to investigate the roles of these proteins in mammalian aging. Characterization of the effect of overexpression of MsrA on the lifespan in a mouse model also should be informative. In addition to regulating lifespan, methionine sulfoxide reductases were directly implicated in aging-related neurodegenerative diseases, such as Alzheimer's and Parkinson's diseases [57-59]. Selenoprotein forms of methionine sulfoxide reductases Although MsrBs are widely distributed in nature, the selenoprotein form of MsrB has only been found in vertebrates. In contrast, vertebrate MsrAs are Cys-containing proteins. Interestingly, a selenoprotein form of MsrA has also been identified, but could only be detected in C. reinhardtii, a unicellular green algae [23]. This Sec-containing MsrA has not been characterized. The observation that both classes of methionine sulfoxide reductase can occur in the form of selenoproteins suggests a catalytic advantage that Sec can offer in protection against oxidation of methionine residues. Catalytic properties and key role of selenium in MsrB Both bacteria and eukaryotes share basic features of the Sec insertion machinery, such as SECIS element. Sec tRNA, and SECIS-binding protein. However, a bacterial Sec insertion system also significantly differs from the mammalian system [3, 60-63]. For example, the conserved sequences and structures of SECIS elements are different in bacteria and eukaryotes. The location of these structures is also different: bacterial SECIS elements are present within coding regions, immediately downstream of Sec-coding UGA codons, whereas eukaryotic SECIS elements are located in the 3'untranslated regions. What is the role of the Sec residue in catalytic function of mammalian MsrBl? By site-directed mutagenesis, a bacterial SECIS element was introduced immediately downstream of the in-frame UGA Sec codon, and the recombinant Sec-containing MsrBl was expressed in Escherichia coli [41]. This recombinant selenoprotein MsrBl had four mutations (S99R, SIOOL, K102G, and F103P) compared to the wild type Sec-containing protein. The mutant exhibited ~800-fold higher enzyme activity than the corresponding Cys-containing form, indicating the essential role of Sec in this enzyme. In addition, a separate study reported that a recombinant Sec-

128

Selenium: Its molecular biology and role in human health

containing MsrBl expressed from a construct that contained the entire E. coli formate dehydrogenase H SECIS element in the coding region, had 100-200 fold higher specific activity than the corresponding Cys mutant [64]. However, natural mammalian Cys-containing enzymes, MsrB2 and MsrB3, also showed a high enzyme activity [41]. Although the primary function of MsrA and MsrB is to catalyze the reduction of protein-based methionine sulfoxides, these enzymes can also reduce free methionine sulfoxides, albeit with low activity. Consistent with this property, all three mammalian MsrBs exhibited the activity in the conversion of free methionine-i?-sulfoxide to methionine [41]. Some but not all MsrBs are zinc-containing proteins. This group of proteins includes all animal MsrBs, which contain a structural zinc coordinated by two CXXC motifs (two Cys separated by two other residues) [11,41]. Mutation of any zinc-coordinating Cys to Ser in Drosophila MsrB resulted in complete loss of metal and catalytic activity [34]. It has been suggested that zinc plays a structural role in metalloprotein MsrBs, and that this metal is not directly involved in the catalytic function [34,65]. Catalytic mechanisms of MsrB The reaction mechanism of MsrA has been characterized biochemically and is well supported by 3D structures [25-29]. Studies involving MsrBs from D. melanogaster, N. gonorrhoeae, and N. meningitides [24,34,66] revealed a similar catalytic mechanism. Interestingly, MsrA and MsrB folds are completely different [24,26-28,67]. MsrB functions in the following manner: 1) a catalytic Cys attacks sulfoxide moiety of the substrate resulting in the formation of sulfenic acid intermediate and concomitant release of methionine; 2) a resolving Cys attacks the sulfenic acid intermediate to form an intramolecular disulfide bond; and 3) a fully reduced enzyme is regenerated by reduction of the disulfide with Trx, a natural electron donor. DTT can also reduce the disulfide in in vitro assays. Multiple sequence alignments showed that -60% of known MsrBs contain the conserved resolving Cys. The remaining -40% of MsrB, including all three mammalian MsrBs, do not have this Cys. In addition, some bacterial MsrBs have only a single Cys in their sequences, suggesting the lack of any resolving Cys. How does the reaction proceed in these enzymes? Two alternative reaction mechanisms have been proposed. One is a direct reduction of the sulfenic acid intermediate by Trx. The other is the use of an alternative resolving Cys to form the intramolecular disulfide bond with the catalytic Cys. The second possibility is supported by the recent study of Xanthomonas campestris MsrB [68].

Selenium and Methionine Sulfoxide Reduction

129

Different sets of active site features in selenoprotein and nonselenoprotein MsrB Multiple sequence alignments reveal three highly conserved residues in MsrB sequences. These three residues are present in Cys-containing MsrBs, but absent in selenoprotein forms of this enzyme (Figure 2). It was previously found that these residues are part of the active site. Why did selenoprotein MsrBs evolve different residues in these positions? What are the roles of these residues in the catalytic function of selenoprotein and Cyscontaining MsrBs? 71 mouse MsrB 1 mouse MsrB2 human MsrB3 D. melanogaster S. cerevisiae A. thaliana E. coli M. thermautotropliicus A. lumefaciens M. tuberculosis S. enterica V. cholerae H. pylori H. influenzae N. gonorrhoeae

I

81

I

91

I

I

t

101

I

VSCGKCGHGLGHEFLNDGP •^ ••-KRGQSRFUIPS S SLKFVPK \A/CKQCEAHLGHyFP--DGP KPTGQRFCINSV/lLKFKPS TSCSQCGAHLGHIFD-DGP RPTGXRYCINSAALSFTPA VRCARCNAHMGHVFE-DGP KPTRSRYCINSASIEFVNA ICCARCGGHLGHVFEGEGWKQLLNLPKDTRHCVNSASLNI/KTSD INCAVCDGHLGHVHKGEGY STPGD3RLCVNSVSINFNRA IRCGNCDAHLGHVFP-DGP QPTG3RYCVNSASLRFTDG VLCARCDAHLGHVFD-DGP RPTGXRYCKNSAALKFIPR IRCAHCGGHLGHVFP-DGP I ' P T G : RYCTNGHSMVf'KPV VLCANCDSHLGHVFAGEGY -PTPTDXRYCINSISLRLVPG IRCGNCDAHLGHVFP-DGP QPTGFRYCVNSASLAFSDE IRCAACDSHLGHVFE-DGP KTTCLRFCVNSVSLIFNKK VLSRIGKAHLGHVFN^DGP KELGGLRYCINSAALRFIPL \'LSRAGNAHLGHVFD-DGP KDKGGRRYCINSASIKFIPL yRSRAADSHLGHVFP-DGP RJKGGIRYCINGASLKFIPL

Figure 2. Partial alignment of Sec- and Cys-containing MsrBs. Catalytic Cys (C) and Sec (U) residues are shown by an arrow. The conserved His, Val/Ile, and Asn residues in Cyscontaining proteins are indicated by arrowheads, and the corresponding residues (Gly, Glu, and Phe, respectively) in mouse selenoprotein MsrBl are indicated by bold letters. Numbering of amino acids is based on the mouse MsrBl sequence. Cys71 and Cys74 coordinate zinc, and the corresponding CxxC motif is conserved in many, but not all MsrBs. Accession numbers (GI) are as follows: Mus musculus MsrBl, 7305478; M. musculus MsrB2, 27753987; Homo sapiens MsrB3, 72534836; Drosophila melanogaster, 17944415; Saccharomyces cerevisiae, 6319816; Arabidopsis thaliana, 4115939; Escherichia coli, 15802192; Methanothermobacter thermautotrophicus, 15678738; Agrobacterium tumefaciens, 15888246; Mycobacterium tuberculosis, 15609811; Salmonella enterica, 16760604; Vibrio cholerae, 15642000; Helicobacter pylori, 3252888; Hemophilus influenzae, 16273361; Neisseria gonorrhoeae, 19526685.

It was recently found that the three residues uniquely conserved in Cyscontaining MsrBs are critical for enzyme activity in MsrB2 and MsrB3, yet introducing these residues into MsrBl inactivates this selenoprotein. Interestingly, when these residues are introduced into the Cys-containing mutant of MsrBl, the activity of this mutant increases several fold. Thus, the three residues are required for Cys-containing MsrB forms, but detrimental

Selenium: Its molecular biology and role in human health

130

for the Sec-containing forms. These data suggested that Sec- and Cyscontaining MsrBs evolve distinct sets of active site features to maximize their catalytic efficiencies. Different catalytic mechanisms between selenoprotein and nonselenoprotein mammalian MsrBs As discussed above, many selenoproteins have fully functional orthologs, wherein Cys replaces Sec. These Cys orthologs are often catalytically as efficient as selenoprotein forms. The reason why Sec is used in proteins if the Cys versions are sufficient for their functions is only beginning to be understood. The three mammalian MsrBs offer a great model system to address these questions which are central to the role of selenium in biology. We recently reported that selenoprotein and non-selenoprotein forms of mammalian MsrBs employ different catalytic mechanisms with respect to the regeneration of the fully reduced proteins (Figure 3) [69]. Sec-SeH HS-Cys Sec-SeOH SH-Cys J—-J^et-R-SO Met ^ k. Sec-containing f ^,.^\ \ ^ /selenenic acioj IVIsrBI 1 reoucea I WenmediaW

Sec-§e—S-Cys ». /6elenenyl»ulflcl^

Cys-containing r . ^ \

V ^

/^uitenicacwN

) ^ - ^

Uiermediatej

• (

reduced

1

Cys-SH

Cys-SH Cys-SOH ^-"-•--^et-R-SO Met ^ - - J - , ^ ^ IVIsrB2 & MsrB3 (reduced

Sec-SeH HS-Cys

Trx

Figure 3. Models for catalytic mechanisms of Sec- and Cys-containing MsrBs. In Seccontaining MsrBl, the catalytic Sec attacks the methionine-7?-sulfoxide (Met-R-SO) to form a selenenic acid intermediate, which interacts with a resolving Cys and forms a selenenylsulfide bond. This selenenylsulfide bond, subsequently, is reduced by Trx. In contrast, in Cyscontaining MsrB2 and MsrB3, a sulfenic acid intermediate can be directly reduced by Trx because the resolving Cys is dispensable.

As shown in Figure 3, the selenoprotein MsrBl requires a unique resolving Cys for recycling of the enzyme. A selenenic acid intermediate of MsrBl is engaged in a selenenylsulfide bond with resolving Cys located in the Nterminal part of the protein. Mutation of this Cys to Ser results in complete loss of enzyme activity in the Trx-dependent reaction but not in the DTTdependent reaction. Subsequently, the selenenylsulfide is reduced by a physiological electron donor, Trx. It appears that the selenenic acid intermediate cannot be reduced by Trx but is reducible by DTT. In contrast, mutational analyses revealed that the resolving Cys is not needed for Cys-

Selenium and Methionine Sulfoxide Reduction

131

containing MsrB2 and MsrB3, in which the sulfenic acid intermediate is likely directly reduced by Trx. Catalytic advantages and disadvantages of Sec-containing proteins compared to Cys-containing counterparts The data on Sec- and Cys-containing MsrBs suggested that Sec per se may result in a higher catalj^ic activity. Replacement of Cys with Sec in mammalian MsrB2 and MsrB3 increased the activity over 100 fold in the DTT-dependent reaction [69]. It has been previously reported that the catalytic activity of phospholipid hydroperoxide glutathione peroxidase can also be increased by replacing the active site Cys with Sec [70]. Thus, the enhanced catalytic activity of selenoproteins can explain, at least in part, the advantage of Sec over Cys. However, replacement of Cys with Sec may not only influence enzyme activity, but also alter protein function. For instance, a Sec-containing form of subtilisin is an efficient peroxidase rather than a protease [71]. The use of Sec in glutathione S-transferase also changed this protein to a peroxidase [72]. Are there any other advantages of the use of selenium besides enhancing enzyme activity? A recent study has shown that the presence of a unique Cterminal active site in Drosophila thioredoxin reductase (SCCS instead of GCUG in mammalian enzymes) can convert this protein into a highly active enzyme [73]. This study suggested that the advantages of selenoenzymes are in a broader range of substrates and flexibility of microenvironmental conditions in the active sites. It appears that the use of Sec can also result in catalytic disadvantages. Although the mutant Sec-containing forms of MsrB2 and MsrB3 are characterized by a 100-fold increased methionine sulfoxide reductase activity, regeneration of the active enzymes by natural electron donor (e.g., Trx) is not possible [69]. However, by introducing a unique resolving Cys, the selenoprotein form of MsrB3 can become as active as the natural Cyscontaining enzyme in the Trx-dependent reaction [69]. Evolutionary implications Most selenoprotein forms likely evolved by replacing catalytic Cys with Sec, which equips oxidoreductases with enhanced activity. However, this evolutionary process is more complex than a simple change of a Cys codon to TGA. The changes must also involve the generation of SECIS elements. Furthermore, Sec may evolve only in environments where selenium is present in sufficient levels and in organisms with active Sec insertion system. Clearly, although replacement of catalytic redox Cys with Sec may be expected to enhance protein function, other requirements should be satisfied. These and other factors may limit the use of Sec in biological systems.

132

Selenium: Its molecular biology and role in human health

Acknowledgments This work is supported by NIH AG021518 to VNG. References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36.

TC Stadtman 1996 Amu Rev Biochem 65:83 JF Atkins, RF Gesteland 2000 Nature 407:463 DL Hatfield, VN Gladyshev 2002 Mol Cell Biol 22:3565 GV Kryukov, S Castellano, SV Novoselov, AV Lobanov, O Zehtab, R Guigo, VN Gladyshev 2003 Science 300:1439 I Chambers, J Frampton, P Goldfarb, N Affara, W McBain, PR Harrison 1986 EMBO J 5:1221 VN Gladyshev, KT Jeang, TC Stadtman 1996 Proc Natl Acad Sci USA 93:6146 F Zinoni, A Birkmann, TC Stadtman, A Bock 1986 Proc Natl Acad Sci USA 83:4650 A Bock, K Forchhammer, J Heider, C Baron 1991 Trends Biochem Sci 16:463 GV Kryukov, VM Kryukov, VN Gladyshev 1999 J Biol Chem 274:33888 A Lescure, D Gautheret, P Carbon, A Krol 1999 J Biol Chem 274:38147 GV Kryukov, RA Kumar, A Koc, Z Sun, VN Gladyshev 2002 Proc Nad Acad Sci USA 99:4245-4250 C Jacob, GI Giles, NM Giles, H Sies 2003 Angew Chem Int Ed Al-Al^l H Weissbach, F Etienne, T Hoshi, SH Heinemann, WT Lowther, BW Matthews, G St. John, C Nathan, N Brot 2002 Arch Biochem Biophys 397:172 H Weissbach, L Resnick, N Brot 2005 Biochim Biophys Acta 1703:203 I Petropoulos, B Friguet 2005 Biochim Biophys Acta 1703:261 ER Stadtman, H Van Remmen, A Richardson, NB Wehr, RL Levine 2005 Biochim Biophys Acta 1703:135 J Moskovitz 2005 Biochim Biophys Acta 1703:213 N Brot, L Weissbach, J Werth, H Weissbach 1981 Proc Nad Acad Sci USA 78:2155 N Brot, H Weissbach 1983 Arch Biochem Biophys lli:ll\ B Ezraty, L Aussel, F Barras 2005 Biochim Biophys Acta 1703:221 U Bechtold, DJ Murphy, PM MuUineaux 2004 Plant Cell 16:908-919 CV Dos Santos, S Cuine, N Rouhier, P Rey 2005 Plant Physiol 138:909 SV Novoselov, M Rao, NV Onoshko, H Zhi, GV Kryukov, Y Xiang, DP Weeks, DL Hatfield, VN Gladyshev 2002 EMB0J1\-M%\ WT Lowther, H Weissbach, F Etienne, N Brot, BW Matthews 2002 Nat Struct Biol 9:348 WT Lowther, N Brot, H Weissbach, JF Honek, BW Matthews 2000 Proc Nad Acad Sci USA 97:463 WT Lowther, N Brot, H Weissbach, BW Matthews 2000 Biochemistry 39:13307 F Tete-Favier, D Cobessi, S Boschi-MuUer, S Azza, G Branlant, A Aubry 2000 Structure 8:1167 AB Taylor, DM Benglis Jr, S Dhandayuthapani, PJ Hart 2003 J Bacterial 185:4119 M Antoine, S Boschi-Muller, G Branlant 2003 J Biol Chem 278:45352 A Olry, S Boschi-Muller, M Marraud, S Sanglier-Cianferani, A Van Dorsselear, G Branlant 2002 J Biol Chem 277:12016 R Grimaud, B Ezraty, JK Mitchell, D Lafitte, C Briand, PJ Derrick, F Barras 2001 J Biol Chem 276:48915 S Jung, A Hansel, H Kasperczyk, T Hoshi, SH Heinemann 2002 FEBS Lett 527:91 F Etienne, D Spector, N Brot, H Weissbach 2003 Biochem Biophys Res Commun 300:378 RA Kumar, A Koc, RL Cemy, VN Gladyshev 2002 J Biol Chem 277:37527 A Hansel, SH Heinemann, T Hoshi 2005 Biochim Biophys Acta 1703:239 A Hansel, L Kuschel, S Hehl, C Lemke, HJ Agricola, T Hoshi, SH Heinemann 2002 FASEB J 16:911

Selenium and Methionine Sulfoxide Reduction 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73.

133

S Vougier, J Mary, B Friguet 2003 Biochem J373:531 HY Kim, VN Gladyshev 2005 Biochemistry 44:8059 J Moskovitz, H Weissbach, N Brot 1996 Proc Natl Acad Sci USA 93:2095 L Kuschel, A Hansel, R Schonherr, H Weissbach, N Brot, T Hoshi, SH Heinemann 1999 FEBSLettA56M HY Kim, VN Gladyshev 2004 MolBiol Cell 15:1055 W Huang, J Escribano, M Sarfarazi, M Coca-Prados 1999 Gene 233:233 HY Kim, VN Gladyshev 2004 Biochem Biophys Res Commun 320:1277 RL Levine, J Moskovitz, ER Stadtman 2000 lUBMB Life 50:301 ER Stadtman, J Moskovitz, RL Levine 2003 Antioxid Redox Signal 5:577 J Moskovitz, E Flescher, BS Berlett, J Azare, JM Poston, ER Stadtman 1998 Proc Natl Acad Sci USA 95:U01\ EP Skaar, DM Tobiason, J Quick, RC Judd, H Weissbach, F Etienne, N Brot, HS Seifert 2002 Proc Natl Acad Sci USA 99:10108 VK Singh, J Moskovitz 2003 Microbiology 149:2739 P Alamuri, RJ Maier 2004 MolMicrobiol 53:1397 M Kantorow, JR Hawse, TL Cowell, S Benhamed, GO Pizarro, VN Reddy, JF VieitcaaacxklOOA Proc Natl Acad Sci USA 101:9654 PG Sreekumar, R Kannan, J Yaung, CK Spee, SJ Ryan, DR Hinton 2005 Biochem Biophys Res Commun 334:245 O Yermolaieva, R Xu, C Schinstock, N Brot, H Weissbach, SH Heinemann, T Hoshi 2004 Proc Natl Acad Sci USA 101:1159 MA Marchetti, GO Pizarro, D Sagher, C Deamicis, N Brot, JF Hejtmancik, H Weissbach, M Kantorow 2005 Invest Ophthalmol Vis Sci 46:2107 J Moskovitz, S Bar-Noy, WM Williams, J Requena, BS Berlett, ER Stadtman 2001 Proc Natl Acad Sci USA 98:12920 H Ruan, XD Tang, ML Chen, ML Joiner, G Sun, N Brot, H Weissbach, SH Heinemann, L Iverson, CF Wu, T Hoshi 2002 Proc Natl Acad Sci USA 99:2748 A Koc, AP Gasch, JC Rutherford, HY Kim, VN Gladyshev 2004 Proc Natl Acad Sci USA 101:7999 SP Gabbita, MY Aksenov, MA Lovell, WR Markesbery 1999 JNeurochem 73:1660 C Schoneich 2005 Biochim Biophys Acta 1703:111 CB Glaser, G Yamin, VN Uversky, AL Fink 2005 Biochim Biophys Acta 1703:157 Z Liu, M Reches, I Groisman, H Engelberg-Kulka 1998 Nucleic Acids Res 26:896 KE Sandman, CJ Noren 2000 Nucleic Acids Res 28:755 ES Amer, H Sarioglu, F Lottspeich, A Holmgren, A Bock J Mol Biol 292:1003 F Zinoni, J Heider, A Bock 1990 Proc Natl Acad Sci USA 87:4660 S Bar-Noy, J Moskovitz 2002 Biochem Biophys Res Commun 297:956 A Olry, S Boschi-MuUer, H Yu, D Bumel, G Branlant 2005 Protein Sci 14:2828 A Olry, S Boschi-MuUer, G Branlant 2004 Biochemistry 43:11616 B Kauffmann, A Aubry, F Favier 2005 Biochim Biophys Acta 1703:249 F Neiers, A Kriznik, S Boschi-Muller, G Branlant lOOAJBiol Chem 279:42462. HY Kim, VN Gladyshev 2005 PLoSBiol 3:e375 S Hazebrouck, L Camoin, Z Faltin, AD Strosberg, Y Eshdat 2000 J Biol Chem 275:28715 IM Bell, ML Fisher, ZP Wu, D Hilvert 1993 Biochemistry 32:3754 HJ Yu, JQ Liu, A Bock, J Li, GM Luo, JC Shen 2005 J Biol Chem 280:11930 S Gromer, L Johansson, H Bauer, LD Arscott, S Rauch, DP Ballou, CH Williams Jr, RH Schirmer, ES Amer 2003 Proc Natl Acad Sci {/&4 100:12618

Chapter 12. Selenoprotein W in development and oxidative stress Chrissa Kioussi Department of Biochemistry and Biophysics, Oregon State University, Corvallis, OR 97331, USA

Philip D. Whanger Department of Environmental and Molecular Toxicology Oregon State University, Corvallis, OR 97331, USA

Summary: Dietary selenium deficiency in animals results in significant reduction of selenoprotein W (SeW) expression in their tissues. Selenium supplementation results in SeW accumulation predominantly in skeletal muscle and heart, suggesting a critical metabolic role for this protein in these tissues. SeW is expressed in the developing heart, muscle, brain, neural tube and otic ventricle during mouse development. SeW promoter studies indicated the presence of metal response element consensus sequences and binding sites for factors that suppress SeW expression. SeW exhibits glutathione-dependent redox properties in vivo. Aerobic growth conditions in microbes favor the production of SeW without the attached glutathione, whereas anaerobic growth conditions promote the glutathione-bound form. SeW levels are upregulated under oxidative stress in myoblasts. We suggest that SeW is a scavenger for reactive oxygen species such as hydrogen peroxide during muscle and nervous system development. Introduction Selenoprotein W (SeW) was originally reported as a missing component in selenium deficient lambs suffering fi-om white muscle disease [1]. This was confirmed by more detailed subsequent studies [2, 3]. When selenocysteine was identified as the form of selenium in a crude fi-action of SeW [4], this provided more convincing evidence that SeW was indeed a selenoprotein. The chemical form of selenium had been previously identified as selenocysteine in glutathione peroxidase (GPX; [5]), and the form of selenium is now known to be selenocysteine in all the mammalian selenoproteins [6]. SeW was first successfully purified fi-om rat muscle and multiple forms with and without glutathione (GSH) were found [7,8]. SeW, a member of the selenoprotein mammalian families, 25 in humans and 24 in

136

Selenium: Its molecular biology and role in human health

rodents, is 87 amino acids in length. SeW is highly conserved (83%) among mammalian species, and two species of the purified protein contain GSH bound to Cys36 [9]. TGA, the codon for selenocysteine, is located at position 13 in SeW cDNA. SeW is expressed at its highest levels in limb muscle, heart and brain in mammals. Expression of SeW in the brain is preserved in selenium deficient animals [10], which suggests its role in nervous system development and function. Recent studies indicated SeW as the molecular target of methylmercury. SeW is downregulated by methylmercury and its behavior is associated with intracellular GSH depletion in neuronal cells [11]. Three reviews have been written on SeW [12-14] and thus mostly recent information will be reported in this review. SeW promoter studies SeW promoter (-2090 bp fragment) contains a metal response element (MRE) just upstream of the transcriptional start site, -20bp, and binding sites for the transcription factors Spl, AP-1 and the glucocorticoid receptor (GR), genes that are involved in the transcriptional activation of metallothionein genes [15]. SeW promoter fragments, -404bp and -1265bp, with mutated and non-mutated MRE sites were cloned into luciferase reporter vectors and transfected into glial and muscle cells. Transfected cells were exposed to physiological and excess levels of cadmium, copper and zinc and assayed for luciferase reporter gene expression. SeW promoter activity was increased in presence of copper and zinc only in glial cells. Mutation of MRE sites abolishes SeW promoter response to metal exposure [16]. The -200/-400 bp promoter fragment exhibits the most activity [17]. Band shift assays showed specific binding of Spl to the Spl consensus sequence in the SeW promoter as well as to the MRE [18]. Spl was also shown to bind to GR-rich regions of the MRE sequences of other promoters and negatively regulates their expression [19]. SeW transcription might be also negatively regulated by the Spl/MRE interactions. Truncations of rat SeW promoter, 2090, 1265, 741 and 404 base pairs, of genomic DNA lying immediately upstream of the SeW coding sequence were also cloned into a luciferase reporter vector and assayed in muscle and brain cells. The smaller truncations stimulated greater promoter activity than the largest one, suggesting that the full length promoter contains binding sites for factors that suppress SeW expression [17]. Promoter activity of constructs of the SeW promoter ranging from 200 base pairs to 51 base pairs gradually decreased to zero in brain cells, but fell precipitously to zero in muscle cells. These results indicate that cell lines from various tissues respond differently to the identical truncated SeW sequence, and this behavior could contribute to tissue specific expression of SeW [10,20].

Selenoprotein Win development and oxidative stress

137

GSH and SeW Using rat mutant SeW where selenocysteine was replaced with cysteine, conditions were found which influence the binding of GSH to SeW [21]. Studies with the His-tagged recombinant form of rat mutant SeW revealed that aerobic growth conditions in microbes produced primarily a form without bound GSH, but anaerobic growth produced a GSH bound form. Thus, growth conditions can be used to advantageously obtain SeW predominantly with or without GSH for studies to explore the importance of this tripeptide with SeW. SeW in human tissues Selenium deficiency is associated with a cardiomyopathy in humans called Keshan disease, but it is not known whether SeW is a factor in this disorder. However, SeW levels are significantly lower in muscle and heart from aborted fetuses with low selenium status [22]. The presence of SeW in fetal tissues lends support for its function in development. For comparison, GPX activity was also found to be significantly lower in these tissues fi-om deficient fetuses. The highest levels of SeW in muscle and heart fi-om these fetuses are consistent with the patterns noted for adult human tissues [23]. SeW in development SeW is expressed in high concentration in skeletal muscle, heart and brain in primates [23], sheep [10] and muscle and brain in rats [24]. Selenium depletion and repletion affected SeW levels differently in various brain regions [25]. SeW levels in cerebellum, cortex and thalamus was not significantly affected by selenium depletion, but SeW levels increased only in thalamus with selenium repletion. RNA whole mount in situ hybridization assays (Figure 1) demonstrates high levels of SeW in the decidua and the newly implanted embryo, egg cylinder stage, at embryonic day 6 (e6), with stronger expression in the ectodermal cells. At embryonic day 8 (e8), SeW was expressed in ectodermderived tissues with strong presence in the neuroepithelium of the cephalic neural folds (future forebrain) and somites (future skeletal muscles). One day later, e9, SeW was detected at high levels in the nervous system, the forebrain and midbrain and otic pit and neural tub. At elO, SeW was expressed in the entire brain, forebrain, midbrain and diencephalons (future thalamus and hypothalamus), the neural tube and otic pit. In addition to the nervous system, SeW was expressed in the ventricles of reconstructed heart and in the proliferating area of limbs. At el 1, SeW was clearly expressed in the brain region, otic pit, heart and limbs. Expression profile during development indicated its strong presence in early development events such as implantation, placentation and gastrulation. During implantation the embryo is nourished by maternal fluids and blood.

138

Selenium: Its molecular biology and role in human health

This is the beginning of the intimate relationship between maternal and trophoblastic tissues, which will last until birth. Presence of SeW in the gastrulating embryo indicates that SeW is involved in the molecular network of the transition from the maternal directed to embryonic genome. SeW is speculated to be involved in the morphogenetic movements and the formation of the germ layers and induction of the major organ systems, such as nervous system, cardiovascular and skeletal musculature (Figure 1) [26].

m

m

op op

Nt

m

di

op

f

f

ii

e6

eS

e9

elO

di

h

ell

Figure 1. Expression Profile of SeW during development. SeW RNA whole mount hybridization assays in mouse embryos at day 6, 8, 9, 10 and 11. SeW was detected at the deciduas of the implanted embryo and e6 and at the ectodermal cells (asterisk) of the implanted embryo (brackets). At e8, SeW was expressed in cephalic neural folds (c) and the somites (s). At e9, SeW was strongly expressed in the forebrain (f) and midbrain (m) and neural tube and otic pit (op). At elO and e l l , SeW was expressed in the forebrain, diencephalons (di), midbrain, heart (h) and hindlimb (asterisk), respectively, and in the neural tube(Nt)atelO.

SeW in oxidative stress Oxidative stress is associated with miscarriages, a common placental-related disorder of pregnancy. Early embryogenesis occurs in a low oxygen environment, protecting differentiating cells from damaging free radicals and reactive oxygen species (ROS). Miscarriages can be due to deficient trophoblast invasion during implantation. Initially, these cells occlude the arteries, limiting maternal blood flow into the placenta. Once embryogenesis is complete, the maternal intervillous circulation becomes fully established, and intraplacental oxygen concentration rises threefold [27,28]. Selenoproteins are major players in vertebrate [29] and invertebrate [30] embryonic development by protecting the embryo from oxygen toxicity. Thus, selenium as selenoproteins appears to exert its protective effects by preventing oxidative damage and cellular redox imbalances caused by hydrogen peroxide, lipid hydroperoxides, and other ROS. Overexpression of SeW in ovary epithelial and lung cancer cells causes a markedly reduced sensitivity to hydrogen peroxide [31]. Expression of mutant SeWs, in which selenocysteine 13 or cysteine 36 was replaced by serine did not confer resistance to hydrogen peroxide, implicating the antioxidant activity of SeW

Selenoprotein Win development and oxidative stress

139

[32]. Selenium deficiency in Caco-2 cells resulted in a 73% reduction in SeW mRNA [33]. This reduction in SeW mRNA was similar to that observed for GPXl mRNA (60-80% decrease), but larger than GPX4 mRNA (15-25% decrease) and GPX2 mRNA (no decrease) [33]. SeW was localized in the cytoplasm and a small proportion was associated with the cell membrane [7].

A.


0,^

K^

'iJ^

nf

SeW 18S

Hours Post Treatment

Figure 2. SeW response to oxidative stress. A. SeW RNA levels detected by northern blot analysis in C2C12 murine myoblasts. Total RNA was prepared from cells vifere grown in presence of bovine serum (10%) and 300nM L-buthionine-[S,R]-sulfoxide (BSO) for 16 hrs before exposure to 125 \iM Wfii for 0. 0-5, 1, 2 for 24 hrs. Ethidium Bromide staining of the 18S RNA was used for loading quantitation. B. Graphic analysis of the SeW RNA levels in C2C12 myoblast after oxidative stress. SeW levels were increased in the presence of H2O2 with the highest response at 2 hrs. Twenty four hrs later SeW levels declined due to massive cell death.

SeW was expressed in proliferating C2C12 murine myoblasts. When myoblasts exposed to oxidative stress, after treatment with hydrogen peroxide and L-buthionine-[S,R]-sulfoxide, an inhibitor of de novo GSH synthesis, SeW RNA levels were upregulated, with the highest of four fold induction at 2 hrs post treatment (Figure 2). Similar response has been observed for glyceraldehyde-3-phosphate dehydrogenase, an enzyme involved in the glycolytic ATP biosynthesis [26]. SeW immediate response to oxidative stress by hydrogen peroxide in CHO and HI299 human lung

140

Selenium: Its molecular biology and role in human health

cancer cells [32] and murine myoblasts [26] suggest an antioxidant function for SeW which depends on both selenium and GSH. References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14.

15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33.

ND Pedersen, PD Whanger and PH Weswig 1969 Proc Pacific Slope Biochem Conf. p49 ND Pedersen, PD Whanger, PH Weswig, OH Muth 1972 Bioinorganic Chem 2:33 RS Black, MJ Tripp, PD Whanger, PH Weswig 1978 Bioinorganic Chem 8:161 MA Beilstein, MJ Tripp, PD Whanger 1981 J Inorganic Biochem 15:339 JW Forstrom, JJ Zakowski, AL Tappel 1978 Biochemistry 17:2639 GV Kryukou, S Castellano, SV Novoselov, AV Lobannov, O Zehtab, R Guigo, VN Gladyshev 2003 Science 300:1439 SC Vendeland, MA Beilstein, CL Chen, ON Jensen, E. Barofsky, PD Whanger 1993 J Biol Chem 268:17103 MA Beilstein, SC Vendeland, E. Barofsky, ON Jensen, PD Whanger 1966 J Inorganic Biochem 61:117 Q-P Gu, MA Beilstein, E. Barofsky, LW Ream, PD Whanger 1999 Arch Biochem Biophys 361:23 J-Y Yeh, QP Gu, MA Beilstein, NE Forsberg, PD Whanger 1997 JNutr 127:394 Y-J Kim, Y-G Chai, J-C Ryu 2005 Biochem Biophys Res Comm 330:1095 PD Whanger 2000 Cell Mol Life Sci 57:1846 PD Whanger 2002 Methods Enzymology 347:179 LW Ream, WR Vorachek, PD Whanger 2001 In: Selenium: Its molecular biology and role in human health. DL Hatfield ed Kluwer Academic Publishers, Boston/Dordrecht/London Chapter 12 pi 37 K Shinji, O Fuminori 1992 In S"' Metallothionein International Conference, Birkhaueuser, Basel, Switzerland p. 457 A Amantana, WR Vorachek, JA Butler, ND Costa, PD Whanger 2002 J Inorganic Biochem 91:356 LA Hooven, WR Vorachek, AB Bauman, JA Butler, LM Ream, PD Whanger 2005 J Inorganic Biochem 99: 2007 A Amantana, WR Vorachek, JA Butler, L. W. Ream, PD Whanger 2004 J Inorganic Biochem 9S:\513 Y Ogra, K Suzuki, P Gong, F Otsuka, S Koizumi 2001 J Biol Chem 276: 16534 Y Sun, JA Butler, NE Forsberg and PD Whanger 1999 Nutr Neurosci 2:222 AT Bauman et al 2004 Biochem Biophys Res Commun 313:308 JA Butler, Y. Xia, Y. Zhou, Y. Sun, PD Whanger 1999 The FASEB J 13:A248 Q-P Gu, Y Sun, LW Ream, PD Whanger 2000 Mol Cell Biochem 204:49 J-Y Yeh, MA Beilstein, JS Andrews, PD Whanger 1995 FASEB J 9:392 Y Sun, JA Butler, PD Whanger 2001 JNutr Biochem 12:88 L Loflin et al 2005 J Inorganic Biochem in press J Hempstock, E Jauniaux, N Greenwold, GJ Burton 2003 Hum Pathol 34:1265 L Poston, MT Raijmakers 2004 Placenta A:S72 RF Burk. 2002 Nutr Clin Care 5:75 N Morozova et al 2003 Genes Cells 8:963 DW Jeong, EH Kim, TS Kim, YW Chung, H Kim, lY Kim 2004 Mol Cells 17:156 DW Jeong, TS Kim, YW Chung, BJ Lee, lY Kim 2002 FEBS Lett 517:225 V Pagmantidis et al 2005 FEBS Lett 579:792

Chapter 13. The 15-kDa selenoprotein functional analysis and role in cancer

(Sepl5):

Vyacheslav M. Labunskyy and Vadim N. Gladyshev Department of Biochemistry, University of Nebraska, Lincoln, NE 68588, USA

Dolph L. Hatfield Section on the Molecular Biology of Selenium, Laboratory of Cancer Prevention, National Cancer Institute, National Institutes of Health, Bethesda, MD 20892, USA

Summary: The 15-kDa selenoprotein (Sepl5) was identified several years ago as a protein of unknown fiinction. In recent years, several lines of evidence implicated Sep 15 in the effect of dietary selenium in cancer prevention. These lines of evidence include: 1) protein expression patterns in normal and malignant cells; 2) identification of polymorphic sites that regulate Sep 15 levels and differentially respond to selenium supplementation; 3) location of the Sep 15 gene in the human genome; and 4) correlation between Sep 15 haplotype and susceptibility to cancer. Functional analyses revealed a specific interaction between Sep 15 and a protein folding sensor in the endoplasmic reticulum of mammalian cells and identified Sep 15 as a novel thioredoxin-like fold redox regulator. Sep 15 defines a new protein family that occurs in several organisms from green algae to mammals and also contains selenoprotein M (SelM) and a recently identified fishspecific selenoprotein Fepl5. Introduction Selenium, an important micronutrient, has been implicated in reducing the incidence of cancer in different animal models and human clinical trials [13]. As discussed in many chapters in this book, this trace element is incorporated into proteins in the form of selenocysteine and is often present at the active centers of selenium-containing enzymes. This chapter describes one such protein, designated the 15-kDa selenoprotein (Sepl5). Sepl5 was identified in 1998 by purifying and characterizing a protein fi-om human T cells [4]. This protein was subsequently implicated in the cancer prevention effect of dietary selenium [5], and more recently in regulation of redox homeostasis in the endoplasmic reticulum (ER) [6]. This chapter provides up to date information on Sep 15 and members of its family.

142

Selenium: Its molecular biology and role in human health

Potential role of SeplS in cancer prevention Several selenoproteins have been proposed as candidates that are responsible for the cancer prevention potential of selenium. Among these proteins is SeplS that has been implicated in mediating the chemopreventive effect of selenium in certain types of cancers, including liver [5], prostate [7], breast [8], and lung malignancies [9].

Exon 1

Exon 2

t=t

Intron 1

Exon 4

Exon 3

Intron 2

t

Intron 3

Exon 5

i=*

Intron 4

SECIS element )-^1125G/A ATG

TGA

TAA

811C/T .3'

1 80 248 312362 Exon Exon Exon Exon

1

2

3

4

1244

Exon 5

Figure 1. Structural organization of the human SeplS gene. The upper panel shows exonintron organization of the SeplS gene. Closed squares correspond to exons, and horizontal lines correspond to introns and flanking regions. The lower panel shows organization of the human SeplS cDNA sequence. The relative positions of the ATG initiation and the TGA Sec codons, the TAA termination signal and the single nucleotide polymorphisms (811C/T and 112SG/A) are indicated. The long horizontal line corresponds to the SeplS cDNA, and short vertical lines correspond to exon-exon junctions. Numbers under junction sites correspond to last nucleotides in preceding exons.

The following lines of evidence make an argument for a possible role of SeplS in cancer etiology. The human gene encoding SeplS is located on chromosome 1 at position p31, a locus commonly deleted or mutated in human cancers and implicated in tumor suppression [10,11]. SeplS expression was found to be significantly decreased in more than half of the tested cancer samples and tumor cell lines as compared to corresponding controls [5,9]. Genetic analysis of the human SeplS gene revealed the presence of two polymorphisms that show strong allelic association [S]. These two single nucleotide polymorphisms (SNPs) are located at positions 811 (C/T) and 112S (G/A) in the SeplS mRNA (see Figure 1). The latter polymorphism is, in fact, part of the SECIS element, an RNA structure in the 3'-untranslated region required for recognition of in-frame UGA codons as selenocysteine. This SNP appears to influence expression of SeplS in a

The 15-kDa selenoprotein (SeplS)

143

selenium-dependent manner. The A1125 variant resulted in a higher expression of the selenoprotein, but it could not efficiently respond to the addition of selenium in the cell culture medium [5,7,12]. Thus, this SNP directly influenced the expression levels of Sep 15, and the outcome was dependent on the selenium status. These data suggested that individuals differing in these alleles may not only differ in Sep 15 levels but also in the response to selenium dietary supplements. The polymorphisms have further relevance to cancer. The Al 125 form is prevalent in African Americans who are known to have a higher incidence of prostate cancer [7]. It also should be noted that the highest expression of Sep 15 is observed in the prostate [5]. It is conceivable that African Americans require higher levels of selenium to achieve the protective levels of Sep 15 expression. In addition, it was found that African Americans showed differences in allelic frequency in head, neck, and breast cancers and examples of the loss of heterozygosity at the Sep 15 locus also were observed [7]. Recent studies from other groups provide further evidence in support of these observations. For example, the A1125 form was found to be less responsive than the G1125 form to the selenium that showed growth inhibitory and apoptotic effects [9]. Interaction of SeplS with UDP-gIucose:glycoprotein glucosyltransferase During initial purification from a human T-cell line, Sep 15 was isolated in the denatured state [4]; however, native Sep 15 isolated from rat prostate and mouse liver co-purified with a protein of-160-240 kDa. The binding partner of Sepl5 was identified as UDP-glucose:glycoprotein glucosyltransferase (UGT), an ER chaperone and essential regulator of the calnexin cycle [6]. The calnexin cycle is a quality control pathway localized to the ER that specifically assists in the folding of N-linked glycoproteins [13]. In this quality confrol pathway, UGT functions as the folding sensor that recognizes unfolded or improperly folded glycoproteins (Figure 2). UGT catalyzes the transfer of a glucose moiety from UDP-glucose to the glycan core [14] that creates a retention signal and initiates binding of membrane-bound calnexin and its lumenal homologue calreticulin to the glycan [15-18]. This triggers the binding of ERp57 (a lumenal protein disulfide isomerase) to calnexin and calreticulin, and accelerates folding by catalyzing disulfide bond exchange [19-23]. The tight association of Sep 15 and UGT in 1:1 ratio (with apparent Kd of 20 nM) suggests that this selenoprotein may modulate the enzymatic activity of UGT and/or be involved in assessing structural fidelity [24]. This role is also supported by the observation that expression of SeplS is activated by the unfolded protein response (UPR), a signaling pathway activated in response to accumulation of unfolded proteins in the ER [25]. The UPR pathway results in enhanced expression of genes encoding proteins

144

Selenium: Its molecular biology and role in human health

that facilitate protein folding and help cells remove incorrectly folded and/or excess proteins in the ER.

• Glucose V- Glycan core

Oxidoreductase

Unfolded glycoprotein

Glucosidase

-S

Folded glycoprotein

Figure 2. Sep 15 and the quality control pathway. UGT recognizes unfolded or improperly folded glycoproteins and catalyzes the transfer of a glucose residue to the glycan. This creates a retention signal and initiates the calnexin/calreticulin cycle. A protein disulfide isomerase binds to calnexin and calreticulin and isomerizes incorrectly formed disulfide bonds. Finally, glucose residue is cleaved by glucosidase, and properly folded glycoproteins are exported form the ER. If not, they are recognized by UGT and the calnexin/calreticulin cycle repeats. The possible functions of Sep 15 in this pathway are shown by open arrows and question marks.

Sequence analysis of the Sep 15 family members identified two distinct domains within this selenoprotein: a C-terminal thioredoxin-like domain that contains the redox active CjtU motif and a novel cysteine-rich domain located in the N-terminal part of the protein (Figure 3). The latter domain exclusively mediates the interaction with UGT and contains six cysteine residues, which are highly conserved among Sep 15 proteins. In contrast to Sepl5, this domain is absent in its homologue selenoprotein M (SelM), and SelM does not interact with UGT. Mutational studies have shown that substitution of any of the six conserved cysteine residues with serines disrupts the interaction between Sep 15 and UGT. As four of these six cysteines form a pair of CxcC motifs, there was a possibility that these motifs were involved in coordination of metal ions. However, neither cysteine-rich domain itself coordinated metal ions, nor was the formation of Sepl5-UGT complex mediated by metals [24] suggesting that these conserved cysteine residues likely form intramolecular disulfide bonds. Site-directed mutagenesis of cysteine residues may disrupt proper arrangement of the

The 15-kDa selenoprotein

(Sepl5)

145

disulfide bonds and inhibit protein-protein interaction between Sep 15 and UGT.

Signal peptide

Sep15

UGT-binding domain C

CxxC

CxxCC --

Thioredoxin-liite domain CGU

ER

signal SelM

CGGU

Figure 3. Schematic illustration of the domain arrangement of Sep 15 and SelM. Both selenoproteins encode an N-terminal signal peptide (colored black); one selenocysteine (U) residue; and a thioredoxin-like domain (colored gray). The N-terminal extension of Sep 15 containing six highly conserved cysteine residues has been labeled as the UGT-binding domain.

Consistent with the ER localization of UGT, Sep 15 was also localized to the ER, and its N-terminal signal peptide, which was cleaved in the mature protein, was necessary for the ER translocation [6]. In contrast, the Cterminus of Sep 15 lacked a typical ER retention signal suggesting that location of Sep 15 in this cellular compartment is the result of the interaction with UGT via the N-terminal cysteine-rich domain. Thiol-disulfide oxidoreductase function The solution NMR structures of the thioredoxin-like domain of Sep 15 from Drosophila melanogaster and SelMfromMus musculus have been identified [26]. The structural studies revealed that both Sep 15 and SelM have a thioredoxin-like fold and suggested a thiol-disulfide oxidoreductase fiinction for these selenoproteins. Although the precise biological function of Sep 15 remains unknown, its surface-accessible CjcU motif, located in the Cterminal redox domain, may participate in the reduction, isomerization or oxidation of disulfide bonds and facilitate protein folding (Figure 2). The redox potential of Drosophila Sep 15, which naturally encodes a cysteine residue in place of the selenocysteine in the redox-active CxU motif, was measured to be -225 mV. This redox potential is higher than that of the strong disulfide reductant thioredoxin (-270 mV [27,28]), but is lower than that of the dithiol oxidase DsbA (-122 mV [29,30]). However, the redox potential of Sep 15 is within the range of the redox potentials of disulfide reshuffling enzymes, such as PDI (-175 mV [27]), indicating that Sep 15 may ftinction as a disulfide isomerase. Although the measured redox potential is close to that of the protein disulfide isomerase, the possibility cannot be excluded that biological activity of Sep 15 in vivo may be affected by proteinprotein interactions with other ER resident proteins and available

146

Selenium: Its molecular biology and role in human health

donors/acceptors of electrons. Furthermore, in contrast to protein disulfide isomerase and its broad substrate specificity, Sep 15 may service a restricted group of protein substrates as demonstrated for protein disulfide isomerase ERp57, which functions exclusively as an isomerase for partially folded glycoproteins that are bound to the chaperones calnexin and calreticulin [19,22,31]. In this case, UGT, which recognizes improperly folded glycoproteins, may serve as a primary binding site for Sep 15 protein substrates (analogous to the role of non-catalytic b'-domain of human protein disulfide isomerase). Expression pattern and regulation of SeplS expression by dietary selenium Expression of the Sep 15 gene was examined in several mouse and human tissues by northern blot and immunoblot analyses [5]. The highest levels of gene expression were observed in prostate, liver, kidney, testes and brain, while lower levels were found in lung, spleen and skeletal muscle. Dietary selenium has been shown to regulate the expression of selenoproteins and the abundance of corresponding mRNAs by acting at both transcriptional and translational levels. For example, selenium deficiency is known to result in up to a 99% decrease in the activity of glutathione peroxidase 1 (GPxl) and in up to a 90% decrease in the abundance of GPxl mRNA in livers of rats and mice. Similar to GPxl, SeplS expression was also regulated by selenium availabiUty in liver and kidney, although the changes in expression were less pronounced [26,32]. In contrast, Sep 15 expression in testes remained almost constant regardless of the selenium availability, consistent with previous observations that the expression patterns of selenoproteins localized to the liver and kidney are highly responsive to the availability of selenium, while the testes efficiently retain this element under selenium-deficient conditions [33,34]. SeplS selenoprotein family members Sep 15 is highly conserved from plants to humans [4] suggesting an evolutionary conserved physiological function among Sep 15 proteins. Another ER resident eukaryotic selenoprotein, SelM, a distant homolog of Sep 15, has recently been identified and characterized [35]. SelM and SeplS share 31% sequence identity in their proposed redox-active domains and encode an N-terminal signal peptide (see Figure 3), which is cleaved in the mature protein after translocation into the ER. In contrast to SeplS, the cysteine-rich UGT-binding domain is absent in SelM, and SelM is likely retained in the ER by a C-terminal retention signal (H/R/K-X-DL). This selenoprotein shows a different expression pattern than SeplS, with the highest expression in the brain. Although the function of SelM has not been firmly established, recent studies provide evidence that decreased expression

The 15-kDa selenoprotein (SepIS)

147

of this protein is associated with Alzheimer's disease and SelM may be involved in protection of neurons from oxidative damage [36]. Recently, a new selenoprotein distantly homologous to Sep 15 family members has been identified [37]. This selenoprotein was found only in fish and was designated Fepl5 (for fish Sepl5-like protein). As observed with SelM, Fepl5 contains a known ER retention signal (RDEL) at its C-terminus and was also localized to the ER. Fepl5 encodes a single selenocysteine residue. In contrast to Sepl5, in which Sec is organized into highly conserved CJCU motif, and SelM, in which Sec is present within CxcU motif, Fepl5 has valine in place of the conserved cysteine. Moreover, Fepl5 does not have any conserved cysteine residues, and in some of the Fepl5 sequences cysteines are not present at all. Thus, this selenoprotein has a distinct catalytic mechanism that may involve formation of selenenic acid or selenenylsulfide bond with a cysteine residue in another protein or low molecular weight thiol, such as glutathione [38]. Another unique feature of this novel selenoprotein is that it is present exclusively in selenocysteinecontaining form, while other members of the Sep 15 protein family exist in the form of both selenocysteine- and cysteine-containing proteins. Furthermore, the very narrow distribution of Fepl5 among fishes suggests that this protein may have a highly specialized function. Acknowledgments This work is supported by NIH grant CA080946 to VNG and the Intramural Research Program of the National Institutes of Health, National Cancer Institute, Center for Cancer Research to DLH. References 1.

Clark LC, Combs GF, Tumbull BW, Slate EH, Chalker DK, Chow J, Davis LS, Glover RA, Graham GF, Gross EG, Krongrad A, Lesher JL, Park HK, Sanders BB, Smith CL, Taylor JR 1996 JAMA 276:1957 2. Zhang Z, Kimura M, Itokawa Y \997 Biol TraceElem Res 57:U7 3. Ip C, Dong Y, Ganther HE 2002 Cancer Metastasis ^ev 21 .-281 4. Gladyshev VN, Jeang KT, Wootton JC, Hatfield DL 1998 J Biol Chem 273:8910 5. Kumaraswamy E, Malykh A, Korotkov KV, Kozyavkin S, Hu Y, Kwon SY, Moustafa ME, Carlson BA, Berry MJ, Lee BJ, Hatfield DL, Diamond AM, Gladyshev VN 2000 J Biol Chem 275:35540 6. Korotkov KV, Kumaraswamy E, Zhou Y, Hatfield DL, Gladyshev VN 2001 J Biol Chem 276:15330 7. Hu YJ, Korotkov KV, Mehta R, Hatfield DL, Rotimi CN, Luke A, Prewitt TE, Cooper RS, Stock W, Vokes EE, Dolan ME, Gladyshev VN, Diamond AM 2001 Cancer Res 61:2307 8. Nasr MA, Hu YJ, Diamond AM 2003 Cancer Ther 1:307 9. Apostolou S, Klein JO, Mitsuuchi Y, Shetler JN, Poulikakos PI, Jhanwar SC, Kruger WD, Testa JR 2004 Oncogene 23:5032 10. Apostolou S, De Rienzo A, Murthy SS, Jhanwar SC, Testa JR 1999 Cell 97:684 11. Cheung TH, Chung TK, Poon CS, Hampton GM, Wang VW, Wong YF 1999 Cancer 86:1294

148

Selenium: Its molecular biology and role in human health

12. Kumaraswamy E, Korotkov KV, Diamond AM, Gladyshev VN, Hatfield DL 2002 Methods Enzymol 347:187 13. Daniels R, Kurowski B, Johnson AE, Hebert DN 2003 Mol Cell 11:79 14. Hebert DN, Foellmer B, Helenius A 1995 Cell 81:425 15. Zapun A, Petrescu SM, Rudd PM, Dwek RA, Thomas DY, Bergeron JJ 1997 Cell 88:29 16. Ellgaard L, Riek R, Herrmann T, Guntert P, Braun D, Helenius A, Wuthrich K 2001 Proa Natl Acad SciUSA98:2l 33 17. Schrag JD, Bergeron JJ, Li Y, Borisova S, Hahn M, Thomas DY, Cygler M 2001 Mol Cell 8:633 18. Kapoor M, Srinivas H, Kandiah E, Gemma E, Ellgaard L, Oscarson S, Helenius A, Surolia A 2003 J Biol Chem 278:6194 19. Zapun A, Darby NJ, Tessier DC, Michalak M, Bergeron JJ, Thomas DY 1998 J Biol Chem 273:6009 20. Molinari M, Helenius A 1999 Nature 402:90 21. Molinari M, Helenius A 2000 Science 288:331 22. Frickel EM, Riek R, Jelesarov I, Helenius A, Wuthrich K, Ellgaard L 2002 Proc Natl Acad Sci US A99:1954 23. Leach MR, Cohen-Doyle MF, Thomas DY, Williams DB 2002 J Biol Chem 277:29686 24. Labunskyy VM, Ferguson AD, Fomenko DE, Chelliah Y, Hatfield DL, Gladyshev VN 2005 J Biol Chem 280:37839 25. Shen X, Ellis RE, Sakaki K, Kaufman RJ 2005 PLoS Genet 1 :e37 26. Ferguson AD, Labunskyy VM, Fomenko DE, Arac D, Chelliah Y, Amezcua CA, Rizo J, Gladyshev VN, Deisenhofer J 2005 J Biol Chem (in press) 27. Lundstrom J, Holmgren A 1993 Biochemistry 32:6649 28. Mossner E, Huber-Wunderlich M, Glockshuber R 1998 Protein Sci 7:1233 29. Inaba K, Ito K 2002 EMBO J 21:2646 30. Huber-Wunderlich M, Glockshuber R 1998 FoldDes 3:161 31. Pollock S, Kozlov G, Pelletier MF, Trempe JF, Jansen G, Sitnikov D, Bergeron JJ, Gehring K, Ekiel I, Thomas DY 2004 EMBO J 23:\020 32. Novoselov SV, Calvisi DV, Labunskyy VM, Factor VM, Carlson BA, Fomenko DE, Moustafa ME, Hatfield DL, Gladyshev VN 2005 Oncogene (in press) 33. Hill KE, Lyons PR, Burk RF 1992 Biochem Biophys Res Commun 185:260 34. Behne D, Hilmert H, Scheid S, Gessner H, Elger W 1988 Biochim Biophys Acta 966:12 35. Korotkov KV, Novoselov SV, Hatfield DL, Gladyshev VN 2002 Mol Cell Biol 22:1402 36. Hwang DY, Cho JS, Oh JH, Shim SB, Jee SW, Lee SH, Seo SJ, Lee SK, Kim YK 2005 Neurochem Res 30:1009 37. Novoselov SV, Hua D, Lobanov AV, Gladyshev VN 2005 Biochem J (in press) 38. Kim HY, Gladyshev VN 2005 PLoSBiol 3:e375

Chapter 14. expression

Regulation of glutathione peroxidase-1

Roger A. Sunde 1415 Linden Drive, University of Wisconsin, Madison, fVI 53705, USA

Summary: Studies on selenoprotein expression are revealing a hierarchy of selenoproteins with regard to the extent of regulation by dietary selenium status, and with regard to the dietary selenium requirement for maximal expression. In rats, liver glutathione peroxidase-1 (GPxl), liver type 1 deiodinase, and plasma selenoprotein P have the highest dietary selenium requirements to maintain plateau levels of selenoprotein. Liver GPxl, liver GPx4 and liver thioredoxin reductase-1 have the highest selenium requirements to maintain plateau levels of the mRNA. The 10-fold change in GPxl mRNA with changes in selenium status is the largest extent of regulation for any known selenoprotein. Regulation of selenoprotein translation by selenium, mediated by Sec-tRNA availability, also regulates the level of all selenoproteins. The unique and specific regulation of GPxl expression by selenium is mediated by GPxl mRNA stability, and involves nonsense-mediated mRNA decay. This regulation requires a functional SECIS element and a UGA codon, and the UGA must be followed by an intron. Our recent results indicate that GPxl mRNA in selenium-deficient rat liver is moderately abundant, and that GPxl mRNA increases more than 20-fold with selenium supplementation. We hypothesize that selenium regulation of GPxl expression is a major component of GPxl function in higher animals, and that in this role, GPxl serves as a biological selenium buffer that maintains modest selenium stores for future selenoprotein synthesis. This role of GPxl and the dramatic regulation of GPxl mRNA level by selenium status make GPxl mRNA an excellent molecular biology marker for assessment of selenium requirements. Introduction Glutathione peroxidase (GPx), the first discovered selenium-dependent enzyme [1], has been a major subject for research because of its abundance in higher animals, its apparent antioxidant role, and because the expression of GPx activity in rats and other animals is highly regulated by selenium status. Early studies thus found that glutathione peroxidase-1 (GPxl) activity in liver and red blood cells could be used to determine selenium

150

Selenium: Its molecular biology and role in human health

requirements [2]. These biochemical assays, including plasma glutathione peroxidase activity (GPx3) in humans, proved to be extremely useful for determining selenium status and for the establishing selenium requirements in animals [3-5] and in humans [6]. The discovery of other selenoproteins such as iodothyronine deiodinases, selenoprotein P, and thioredoxin reductases provides altemative biochemical parameters for assessing selenium status. The imcovering of the complete selenoproteome [7] now offers all 25 human selenoproteins as potential markers of selenium status. Choice of the most useful biomarkers will depend on ease and cost of use, and on the imderlying mechanisms that control the expression of these biomarkers. This chapter will review (i) our work characterizing selenium regulation of GPxl expression, (ii) the biochemical mechanism for this regulation, (iii) the potential role of this regulation in the biological function of GPxl, and (iv) the selection and use of molecular biology markers for assessing selenium requirements. Regulation of GPxl Expression Selenium regulation of GPxl expression is distinct from most markers of mineral nutrient status. In selenium deficiency, liver GPxl activity falls exponentially and dramatically, and it decreases to an undetectable level 21 days after weanling rats are switched to a selenium-deficient diet [2]. Similar but less dramatic falls in GPxl activity are observed in other tissues such as erythrocytes. Secondly, graded dietary supplementation of selenium raises GPxl activity sigmoidally such that it reaches a plateau at 0.1 \i/g diet; additional dietary selenium does not increase liver GPxl activity above this plateau (Figure 1). Lastly, the minimum dietary selenium requirement necessary to reach the GPxl activity plateau is remarkably constant across a wide range of species, suggesting that a common molecular mechanism is responsible for this tight regulation [4]. This regulation makes GPxl a most sensitive parameter for changes in selenium status over the deficient to adequate range [4], and has made GPxl arguably the parameter of choice for assessment of selenium status and selenium requirements. We set out in 1985 to examine the underlying mechanism responsible for the uniformity of dietary selenium requirements. Our first approach was to examine selenium regulation of GPxl protein as well as activity levels. Using anti-GPxl antibodies, we found that weanling rats fed a seleniumdeficient diet have a rapid exponential decline in GPxl protein as well as GPxl activity, with half-lives in liver of 5.2 and 2.8 days, respectively [8]. For selenium repletion, GPxl protein as well as activity requires larger doses and longer time periods than for maintenance [9]. We next found that selenium deficiency has a dramatic effect on GPxl mRNA levels in liver [10] which fall to approximately one-tenth of levels

Regulation of glutathione peroxidase-1 expression

151

found in selenium-adequate animals. In progressive selenium deficiency, we saw a coordinated dramatic exponential drop in GPxl mRNA {XVi = 3.2 d) as well as GPxl activity (t'A = 3.3 d) and GPxl protein (tVi =5.0 d) [11]. These experiments thus begin to substantiate that an underlying molecular mechanism is likely to be responsible for selenium regulation of GPxl mRNA levels [12]. 120

0.05 0.10 0.15 Dietary Se (^g/g diet) ^120" |100^ -J

80-

i 60E

» 40S

1^7"^ •

J f JJ

^

•S 20a: < n0.00

• • • • A 1

1 —

— 1 —

0.05 0.10 0.15 Dietary Se (fig/g diet)

GPX1 GPX4 • TR1 GPX3 • SelP r—*•

0.20

Figure 1. Dietary selenium regulation of selenoprotein activity and mRNA levels. Top: Relative activity levels of liver GPxl, liver GPx4, liver TRl, and plasma GPx3 in rats fed a selenium-deficient diet (<0.01 (ig selenium/g) and supplemented with graded levels of selenium. Bottom: Relative mRNA levels of liver GPxl, liver GPx4, liver TRl, kidney GPx3, and liver SelP in rats. Curves are logistic best-fit curves of data from several studies [14,15,17,19] expressed as a percent relative to rats fed 0.2 ng selenium/g diet.

For these animal studies, we use as our model, young, rapidly growing weanling rats fed selenium-deficient torula yeast diets (0.008 |j.g selenium/g)

152

Selenium: Its molecular biology and role in human health

or crystalline amino acid-based diets (0.002 i^g selenium/g), and supplemented with graded levels of selenium as Na2Se03 for 28 days. For male and female rats, both liver GPxl activity and mRNA levels respond sigmoidally to increasing dietary selenium concentration [13-15], with GPxl activity and mRNA reaching plateaus at 0.1 and 0.05 jig selenium/g diet, respectively (Figure 1, Table 1). Table 1. Selenium requirement hierarchy in rats Parameter

Growing rats'

Presnancv^

Lactation^

(Hg Se/g diet) Rat growth

<0.01 [15]

<0.01 [16]

<0.01 [16]

Liver SelP mRNA

<0.01 [13]

<0.01 [16]

<0.01 [16]

Liver GPX4 mRNA

<0.01 [14]

<0.01 [16]

<0.01 [16]

Kidney GPX3 mRNA

<0.01 [17]

-

-

Liver necrosis

0.04 [18]

-

-

Liver GPX4 activity

0.05 [14]

0.05 [16]

0.05 [16]

Liver GPXl mRNA

0.05 [15]

0.05 [16]

0.05 [16]

Liver TRl mRNA

0.05 [19]

--

-

Liver Dl mRNA

0.05 [20]

<0.01 [16]

<0.01 [16]

Plasma GPX3 activity

0.075 [15]

0.05 [16]

0.075-0.1 [16]

Liver TRl activity

0.075 [19]

~

~

RBC GPXl activity

0.1 [2]

0.05 [16]

0.05 [16]

Liver Dl activity

0.1 [20]

-

-

Liver GPXl activity

0.1 [15]

0.05-0.075 [16]

0.075-0.1 [16]

Liver Se

0.1 [15]

0.075 [16]

0.1 [16]

Plasma SelP

0.1 [21]

--

~

Growing rat dietary selenium requirements from references cited in brackets. Se-adequate weanling rats are fed these diets from weaning. Requirements are the minimum dietary selenium necessary for the indicated parameter to reach plateau levels. ^Dietary selenium requirements for pregnant and lactating rats fed the indicated diets from weaning for 10 weeks, and then bred to become pregnant.

Regulation of glutathione peroxidase-1 expression

153

Erythrocyte GPxl activity has a breakpoint but not a plateau at 0.1 \ig selenium/g diet in young, rapidly growing rats [13,15]. Longer studies (see pregnant and lactating rat study below) show that erythrocyte GPxl activity also reaches a defined plateau at 0.05 ^ig selenium/g diet [16], illustrating that even moderately-long repletion studies may over-estimate the selenium requirement because of the prolonged time necessary to fully reach a steadystate. Plasma glutathione peroxidase (GPx3) activity reaches a plateau breakpoint at 0.075 |xg selenium/g diet [15], and thus shows a distinct response curve for this kidney-derived plasma selenoenzyme. Kidney GPx3 mRNA levels, however, are not reduced in this Se-deficient rat model (Table 1) [17]. Regulation of expression of otiier selenoproteins In selenium-deficient rat liver, phospholipid hydroperoxide glutathione peroxidase (GPx4) activity only decreases to 41% of selenium-adequate levels, and the plateau breakpoint occurs at approximately 0.05 jig selenium/g diet [14]. In contrast, liver GPx4 mRNA is not significantly affected by dietary selenium, with the plateau breakpoint for GPx4 mRNA occurring before 0.013 |ag selenium/g diet (Figure 1). Rat thioredoxin reductase-1 (TRl) activity in rat liver falls to 10% of selenium-adequate levels, with a selenium requirement of 0.075 jig selenium/g diet [19]. In contrast, the pattern of TRl mRNA regulation is more similar to that of GPx4 mRNA (Table 1). Arthur and colleagues [20] found that liver iodothyronine deiodinase-1 (Dl) activity in selenium-deficient animals is 5% of that observed in selenium-adequate animals, and Dl mRNA levels are 50% of adequate levels. The selenium requirement for Dl activity is 0.1 |j,g selenium/g diet (Table 1). Others have reported a selenium requirement of 0.05 ng selenium/g diet [22]. The Dl mRNA response curve is hyperbolic rather than sigmoidal with a plateau breakpoint at 0.05 fig selenium/g diet [20]. Burk and Hill [24] found that plasma selenoprotein P (SelP) concentration in selenium deficiency drops to less than 10% of adequate animals [23], and selenium supplementation of deficient rats results in an increase in SelP concentration ahead of the increases in plasma GPx3 activity and liver GPxl activity. In our rat model [25], however, liver SelP mRNA levels are not decreased at all relative to selenium-supplemented controls (Table 1). This strongly suggests that SelP regulation is not significantly mediated by alterations in mRNA levels. Selenium regulation in pregnancy and lactation We have recently studied the selenium requirements in pregnant and lactating rats to determine selenium requirements in later periods of the life cycle [16]. Female weanling rats were fed a Se-deficient diet (<0.01 jig Se/g) or supplemented with graded levels of dietary selenium (0 to 0.3 \ig Se/g) for

154

Selenium: Its molecular biology and role in human health

>10 wk, bred, and euthanized on d 1, 12, and 18 of pregnancy and d 7 and 18 of lactation; selenium response curves were determined for 10 parameters. Growth, and mRNA levels for SelP, Dl and GPx4 are not decreased by selenium deficiency (Table 1). GPx4 activity requires 0.05 ng Se/g diet for maximum activity, similar to growing rats. Dietary selenium requirements for plasma GPx3 activity decrease 33% in pregnancy, but return in lactation to the growing-rat requirement. The selenium requirement for GPxl activity decreases 25% in pregnancy, but not in lactation. GPxl mRNA requires 0.05 Hg Se/g diet for maximum levels in both pregnancy and lactation, similar to growing rats. This study with adult rats affirms, for the most part, the hierarchy of selenium regulation observed in young, rapidly growing rats. Unexpectedly, Se-adequate levels of GPxl mRNA and activity declines during pregnancy and lactation to <40% and 50%, respectively, of nonpregnant levels [16]. This observation shows the need to fully understand biomarkers at all stages of the life cycle. It also illustrates that without this knowledge it may be dangerous to conclude in isolation that a reduced level of a biomarker must be a sign of low nutrient status. In summary, selenium regulation of selenoproteins in our well-regulated rat model reveals a hierarchy of selenium requirements ranging from <0.01 to 0.1 ng Se/g diet, depending on choice of selenium status parameter, with liver GPxl, liver Dl, and plasma SelP having the highest dietary selenium requirements to maintain levels of the selenoprotein. Liver GPxl, liver GPx4 and liver TRl have the highest selenium requirements to maintain levels of the mRNA (Table 1). The hierarchy with regard to the extent of selenium regulation of selenoprotein mRNA level is (difference in level between deficient and adequate, ranked from most to least): GPxl » Dl > TRl > SelP, GPx3, GPx4. Experiments focused on the underlying mechanism suggest that selenium regulation of translation is important for GPx4 and TRl/SelP patterns, whereas mRNA stability as well as translational control is important in selenium regulation of GPxl. Selenium regulation of GPxl mRNA stability The singular pattern of selenium regulation of GPxl is mediated by the dramatic effect of selenium status on GPxl mRNA levels. To investigate the cw-acting nucleic acid sequence requirements for selenium regulation of GPxl mRNA levels, we used Chinese hamster ovary cells transfected with chimeric GPxl DNA constructs in which specific regions of the genomic gpxl are mutated, deleted or replaced by comparable regions from unregulated GPx4 [26]. Transfected mRNA levels in cells grown in selenium-deficient or selenium-adequate media were determined by RNase protection assay (RPA). Analysis of chimeric GPxl/GPx4 constructs shows that the GPx4 SUTR can completely replace the GPxl 3TJTR in selenium regulation of GPxl mRNA! This indicates that the GPxl STJTR and SECIS

Regulation of glutathione peroxidase-1 expression

155

alone cannot fully explain the unique selenium regulation of GPxl mRNA. In contrast, replacement of the GPxl coding regions with corresponding GPx4 coding regions diminishes or eliminates selenium regulation of the transfected GPxl mRNA [26]. Further analysis of the GPxl coding region demonstrates that the GPxl Sec codon (UGA) and the GPxl intron sequences are required for full selenium regulation of transfected GPxl mRNA levels. Mutations which moved the GPxl Sec codon to three different positions within the GPxl coding region suggest that the mechanism for selenium regulation of GPxl mRNA requires a Sec codon within exon 1. Lastly, we found that addition of the GPxl STJTR to P-globin mRNA can convey significant selenium regulation to P-globin mRNA levels when a UGA codon is placed within exon 1. These studies clearly show that selenium regulation of GPxl mRNA requires a functional SECIS in the 3'UTR and a Sec codon followed by an intron [26]. Nonsense codons are known to destabilize many different mRNA species when located upstream from an intron (reviewed in [27]). Maquat and colleagues [28] demonstrated that GPxl mRNA degradation in rat hepatocytes occurs in the cj^oplasm and that this degradation is likely to occur via nonsense-mediated decay of mRNA. Media selenium concentration raises wildtype GPxl mRNA levels but does not affect mRNA levels when the UGA codon is replaced with the nonsense codon UAA or the cysteine codon UGC. Interestingly, GPx4 mRNA is susceptible to nonsense-mediated decay in vitro in fibroblasts or hepatocj^es, suggesting that there must be a mechanism that precludes nonsense-mediated decay of GPx4 in intact animals [29]. Brigelius-Flohe [30] using similar constructs in cells found that the GPxl BTJTR can also make the GPx4 coding region unstable. Collectively, these approaches [26,28-30] establish that GPxl mRNA is degraded by a nonsense-mediated mRNA decay mechanism when Sec is not available for translation. These studies further suggest that there are additional sequences in GPxl mRNA or there are additional factors in tissues that confer Se-sensitivity uniquely to GPxl mRNA. Selenoprotein transcript abundance We conducted a study in our well-regulated rat model using intact seleniumdeficient and selenium-adequate rats to evaluate the relative contribution of mRNA abundance versus translational efficiency to overall regulation of GPxl expression [31]. GPxl, GPx4 and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) transcripts per cell in rat liver were quantitated using RPA. Surprisingly, we found that GPxl transcripts in selenium deficiency are moderately abundant and similar in abundance to GAPDH and other selenoprotein mRNAs; selenium supplementation increases GPxl mRNA so that it is 30-fold higher than GAPDH mRNA (Figure 2). Secondly, in the same animals, we quantitated translation by assessing ^^Se

Selenium: Its molecular biology and role in human health

156

incorporation into gel-purified GPxl, GPx4 and TRl. We found that translational efficiency of GPxl mRNA is still half of that of GPx4, even when corrected for differences in transcript abundance. More importantly, translational efficiency of GPxl mRNA decreases to 4-6% in selenium deficiency. The net effect is that selenium regulation of GPxl mRNA stability appears to switch GPxl mRNA from nonsense-mediated degradation to translation. This regulatory switch ~ a selenium thermostat ~ helps to explain why GPxl expression is the optimum parameter for assessment of selenium status [31].

] -Se Male ] +Se Male -Se Female +Se Female

30000 O

*^ •S 20000 Q. U (0

c

(0

CDT

c

10000

ir

GPx1

GPx4

GAPDH

CDCD

SelP

Figure 2. Selenoprotein transcript number in liver. Molecules of mRNA/cell were calculated from the massfractionof each mRNA species [31] as determined by Ribonuclease Protection Analysis. Bars indicate the mean (n=3, + standard error of the mean) for GPxl, GPx4, GAPDH and Sel P mRNA abundance in Se-deficient (-Se) and Se-adequate (+Se) male and female rat liver. Bars sharing a common letter are not significantly different (P<0.01) by analysis of variance.

If GPxl mRNA is the only mammalian selenoprotein with a seleniumspecific switch controlling mRNA stability, what is the cause for the modest decreases in other selenoprotein mRNAs in selenium deficiency? The apparent selenium-specific down-regulation of mRNA levels for SelP, Dl and even GPx4 may be explained by ribosomal pausing that must occur when Sec-tRNA^*" concentrations are limiting for protein synthesis. The competition between termination factors and a limited supply of Sec-tRNA^"" would predispose these selenoproteins for early termination. And this idle mRNA could then be subjected to more rapid degradation via normal decapping and exonuclease hydrolysis. This suggests that there are two components involved in selenium regulation of mRNA levels: 1) nonspecific effects relative to whether or not sufficient selenium (Sec-tRNA^^) is available for translation; and 2) GPxl mRNA-specific selenium regulation.

Regulation of glutathione peroxidase-1 expression

157

It is, however, possible that differential translational efficiency of different mRNAs also plays a role. This continues to be an exciting area for future research. Selenium regulation of translation It appears that in many tissues, GPxl translation is most regulated and GPx4 least regulated, with the hierarchy (most to least selenium regulation): GPxl > Dl, SelP, TRl > GPx4. The easiest hypothesis to explain this hierarchy is that GPx4 mRNA is best optimized to compete for selenium incorporation when selenium is in limited supply. A number of observations, however, indicate that additional and overlapping factors are likely to play a role. Selenium concentration clearly appears to be the preeminent factor modulating Sec translation. Without selenium and therefore with insufficient Sec-tRNA , UGA will be interpreted as a stop codon, thus limiting selenoprotein translation. Selenium status also affects the relative levels of two isoforms of the Sec-tRNA^^" in mice (containing mcm^U base modification versus the methylated mcm^Um modification at position 34 within the anticodon of the tRNA). High selenium status raises the concentration of the methylated form and appears to enhance selenium incorporation into GPxl and GPx3 relative to GPx4 and TR [32]. A second set of differences in relative translation of selenoproteins involve the selenocysteine insertion sequence (SECIS) elements in the 3'UTR. Berry and colleagues [33,34] used Dl chimeric constructs with 3'-UTRs from different selenoprotein mRNAs, and later used additional chimera, and showed that different SECIS elements can alter the translational efficiency of selenoproteins. Some of this effect may specifically be due to the differential affinity of SECIS-binding Protein-2 (SBP2) for different SECIS elements [35]. In addition, distances of <111 nt between the UGA and the SECIS element curtail Sec translation [34]. We have also conducted experiments to systematically study the role of UGA position and context on GPxl expression. Wen et al [36] constructed a genomic GPxl expression vector that would result in over-expression of ^Se-labeled recombinant protein in COS-7 cells at higher levels than for endogenously encoded selenoproteins. When the GPxl UGA is mutated to cysteine (UGC) and ten separate codons are mutated to UGA, Wen et al found that UGA codons located in the middle of the open reading frame most efficiently direct Sec incorporation whereas UGA located <21 nt from the AUG-start or <204 nt from the SECIS element had reduced Sec incorporation. Collectively, these studies suggest that differences in Sec-tRNA^^"^ isoforms, SBP-2 affinity for different SECIS elements, UGA/SECIS spacing, and local UGA context and/or RNA secondary structure are all factors influencing the mechanism(s) that modulate Sec insertion during

158

Selenium: Its molecular biology and role in human health

translation. Biological selenium buffer So what is the survival advantage of having a peroxidase whose concentration is profoundly regulated by the status of its cofactor? The discovery that selenium was a component of GPxl logically explained the antioxidant functions of selenium [37]. The discovery of multiple GPx enzymes and genes, however, indicates that this perception was clearly an over-simplification. The demonstration that the GPxl knockout mouse is viable, and grows and reproduces the same as congenic wildtype controls [38], further indicates that GPxl is not the crucial antioxidant enzyme as originally proposed. The hierarchy of susceptibility of these enzymes to decreases due to dietary selenium deficiency shows that GPx4, SelP, TRl, and Dl all are protected from selenium deficiency relative to GPxl. Secondly, when mRNA changes are carefully compared relative to changes in GPxl mRNA, it is clear that GPxl mRNA levels fall by an order of magnitude or more, whereas mRNA levels for GPx4, GPx3, TRl, Dl and SelP tj^ically do not change or decrease by only a factor of 2. The unique and specific regulation of GPxl mRNA suggests that this may be an important aspect of the physiological role of GPxl. We propose that the important and major role of GPxl in liver, and perhaps other tissues is part of the homeostatic mechanism that keeps the free concentration of selenium low, that diverts selenium to more important biological functions of selenium in times of deficiency, and that can reversibly bind excess selenium over the deficient to adequate range. In other words, the major role of GPx is to serve as a "biological selenium buffer" [39]. The observation shown in Figure 2, that GPxl mRNA is not a low-abundant mRNA in Se-deficient liver, but rather that GPxl mRNA levels increase in Se-deficient liver to 30 times the level of GAPDH mRNA, further supports this role - GPxl mRNA increases only when other needs for selenium are met, thus facilitating safe storage of this otherwise toxic element. While mediated by a different mechanism, the result is similar to the zinc-induced transcriptional increase in metallothionein mRNA in response to excess zinc, or the de-repressed translation of ferritin mRNA in response to excess iron. And as in these cases, when dietary intake does not keep up with biochemical needs, turnover of the storage protein (GPxl) can release the element (selenium) to meet new needs. The impact of this selenium buffering capacity is that it facilitates an expansion of the dietary selenium range between selenium deficiency and selenium toxicity. This function is more appropriately called a biological selenium buffer, rather than a "selenium store" or "selenium sink" to indicate the dynamic homeostatic nature and to indicate GPxl's active role in modulating selenium flux between incorporation into other selenoenzjones

Regulation of glutathione peroxidase-1 expression

159

and incorporation into GPxl [4]. GPxl thus appears to be another dual-function protein. Proteins are increasingly being discovered to have multiple functions, just as GPx4 has a dual role as a peroxidase and as a structural protein in sperm [40]. The demonstration that GPxl has roles in protection against viral infection [41] or in protection against oxidative stress [42] shows that the peroxidase activity also has svirvival value. Molecular biology markers for assessing selenium status If evolution has programmed GPxl gene expression in higher animals to sense and store selenium when dietary intake exceeds biochemical demand, monitoring this function would be a logical means to assess selenium status of higher animals. Using this biomarker linked to the endogenous homeostatic mechanism should be better at assessing the overall selenium status of an organism versus use of plasma GPx3 which has an unclear role. Selenoprotein P also has merit as a biomarker, but as an inter-organ selenium transport protein, it may reflect immediate selenium intake rather than overall selenium status. Thus we have begun to evaluate human GPxl mRNA and other selenoprotein mRNA levels as a potential way to assess human selenium status. The advent of rapid molecular biology assays suggests that molecular biology markers will soon become important in assessing human health, including human nutrition. We now have found that human blood [43], just as rodent blood [44], has GPxl mRNA expression at levels comparable to the major tissues. We expect to find that GPxl mRNA is an especially effective molecular biology marker of human selenium status, but mRNA for one or more of the newly discovered selenoproteins [7] may also be useful molecular biomarkers of selenium status. References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10.

JT Rotruck, AL Pope, HE Ganther, AB Swanson, DG Hafeman, WG Hoekstra 1973 Science (Washington, DC) 179:588 DG Hafeman, RA Sunde, WG Hoekstra 1974 JNutr 104:580 National Research Council 1983 Selenium in Nutrition National Academy Press Washington DC RA Sunde 1997 Handbook of Nutritionally Essential Mineral Elements BL O'Dell, RA Sunde (eds) Marcel Dekker New York p 493 RA Sunde 2000 Biochemical and Physiological Aspects of Human Nutrition MH Stipanuk (ed) WB Sanders New York p 782 Food and Nutrition Board 2000 Dietary Reference Intakes for Vitamin C, Vitamin E, Selenium and Carotenoids National Academy Press Washington DC p 284 GV Kryukov, S Castellano, SV Novoselov, AV Lobanov, O Zehtab, R Guigo, VN Gladyshev 2003 Science 300:1439 SAB Knight, RA Sunde 1987 JNutr 117:732 SAB Knight, RA Sunde 1988 JNutr 118:853 MS Saedi, CG Smith, J Frampton, I Chambers, PR Harrison, RA Sunde 1988 Biochem Biophys Res Commun 153:855

160 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44.

Selenium: Its molecular biology and role in human health RA Sunde, MS Saedi, SAB Knight, CG Smith, JK Evenson 1989 Selenium in Biology and Medicine A Wendel (ed) Springer-Verlag Heidelberg, Germany p 8 RA Sunde 1990 Annu Rev Nutr 10:451 SL Weiss, JK Evenson, KM Thompson, RA Sunde 1997 J Nutr Biochem 8:85 XG Lei, JK Evenson, KM Thompson, RA Sunde 1995 J Nutr 125:1438 SL Weiss, JK Evenson, KM Thompson, RA Sunde 1996 J Nutr 126:2260 RA Sunde, JK Evenson, KM Thompson, SW Sachdev 2005 J Nutr 135:2144 JK Evenson, CR Mikus, SM Blake, AD Wheeler, RA Sunde 2003 FASEB y 17:Al 137 K Schwarz, CM Foltz \9SlJAm Chem Soc 79:3292 KB Hadley, RA Sunde 2001 J Nutr Biochem 12:693 G Bermano, F Nicol, JA Dyer, RA Sunde, GJ Beckett, JR Arthur, JE Hesketh 1995 Biochem J 3\\A25 JG Yang, KE Hill, RF Burk 19S9 J Nutr 119:1010 S Vadhanavikit, HE Ganther 1993 J Nutr 123:1124 RF Burk, KE Hill 1994 J Nutr 124:1891 KE Hill, PR Lyons, RF Burk 1992 Biochem BiophysRes Commun 185:260 RA Sunde, KM Thompson, JK Evenson, SL Weiss 1998 Proc Nutr Soc 57:155 A SL Weiss, RA Sunde 1998 RNA 4:816 LE Maquat 2001 Biofactors 14:37 PM Moriarty, CC Reddy, LE Maquat 1998 Mol Cell Biol 18:2932 X Sun, X Li, PM Moriarty, T Henics, JP LaDuca, LE Maquat 2001 Mol Biol Cell 12:1009 C MuUer, K Wingler, R Brigelius-Floh6 2003 Biol Chem 384:11 SW Sachdev, RA Sunde 2001 Biochem J251:S5l BA Carlson, XM Xu, VN Gladyshev, DL Hatfield 2005 J Biol Chem 280:5542 MJ Berry, L Banu, JW Harney, PR Larsen 1993 EMBO J 12:3315 GW Martin, III, JW Harney, MJ Berry 1996 RNA 2:171 DM Driscoll, PR Copeland 2003 Annu Rev Nutr 23:17 W Wen, SL Weiss, RA Sunde 1998 J Biol Chem 273:28533 WG Hoekstra 1975 Fed Proc 34:2083 A Spector, Y Yang, YS Ho, JL Magnenat, RR Wang, W Ma, WC Li 1996 Exp Eye Res 62:521 RA Sunde 1994 Selenium in Biology and Human Health RF Burk (ed) Springer-Verlag New York NY p 45 F Ursini, S Heim, M Kiess, M Majorino, A Roveri, J Wissing, L Flohe 1999 Science 285:1393 MA Beck, RS Esworthy, YS Ho, F-F Chu 1998 FASEB J 12:1143 Y Fu, JM Porres, XG Lei 2001 Biochem J1:687 JK Evenson, RA Sunde 2005 FASEB J \9:K\ti\5 JK Evenson, AD Wheeler, SM Blake, RA Sunde 2004 J Nutr 134:2640

Chapter 15. Selenoproteins of the glutathione system Leopold Floh6 MOLISA GmbH, UniversMtsplatz 2, D-39106 Magdeburg, Germany

Regina Brigelius-Flohe Department Biochemistry of Micronutrients, German Institute of Human Nutrition PotsdamRehbruecke (DIfE), Arthur-Scheunert-Allee 114-116, D-14558 Nuthetal, Germany

Summary: The protein family of glutathione peroxidases (GPx) is found throughout the entire life kingdoms. Five distinct molecular clades characterized by an active site selenocysteine residue may coexist in vertebrates. All selenocysteine-containing GPx types reduce hydroperoxides with rate constants k'l near sec", while the cysteine homologs are poor peroxidases. The scope of accepted hydroperoxides increases from the cytosolic and gastrointestinal type to the extracellular type and phospholipid hydroperoxide GPx, while the specificity for GSH declines in this order. Compelling evidence defines cGPx as a device to detoxify H2O2 and soluble hydroperoxides. Being dispensable for survival, cGPx is nevertheless essential to maintain hydroperoxide homeostasis, as demonstrated by mimicking the development of Keshan disease in cGPx^"'"' mice. Being unable to substitute for cGPx in challenged cGPx^''^ mice, the functions of other GPxs have to be sought in local regulation of peroxide-dependent processes, e. g., silencing leukotriene biosynthesis, dampening cytokinedependent NFKB activation, regulating apoptosis and sperm differentiation by PHGPx. The functional divergence within the GPx family is further underscored by different mechanisms of transcriptional control. Introduction Glutathione had been recognized to maintain the intracellular redox-balance long before its role as a substrate for the selenium-containing glutathione peroxidases was discovered. For more than a decade, GPx, which we now call the classical or cytosolic GPx (cGPx; GPxl), remained the only known mammalian selenoprotein. Accordingly, most of selenium deficiency syndromes in higher animals were tentatively attributed to the role of selenium in antioxidant defense. After the discovery of several distinct families of selenop-oteins this proposal can no longer be upheld [1].

162

Selenium: Its molecular biology and role in human health

Glutathione peroxidase activity may be associated with a variety of proteins. The reaction of GSH with ROOH may be catalysed by, e.g.,GSHS-transferases [2], selenoprotein P [3] and peroxiredoxins [4], i.e., proteins that do not share sequence homology with GPx. The 'real' GPxs constitute an old protein family that is distributed throughout the entire living kingdom. The family is clearly split into several molecular clades, which may or may not contain selenium [5]. Characteristics common to the GPx family such as kinetics and catalytic mechanism as well as the substrate specificities and distribution of the subtypes have been discussed in [5]. This chapter focuses on the emerging functional diversification of the GPx family. Enzymatic characteristics of different GPx types Thus far, all members of the GPx family display at least some glutathione peroxidase activity. In the tetrameric enzymes, the active site selenium is located in a valley formed by interacting subunits [6,7], whereas in the monomeric enzyme (PHGPx), the reaction centre is more accessible [8]. This peculiarity may explain the broad specificity of PHGPx for hydroperoxides comprising, apart from H2O2, phosphatidylcholine and cholesterolester hydroperoxides even when integrated into lipoproteins or biomembranes. In contrast, the specificity of cGPx and likely that of GI-GPx is restricted to H2O2 and soluble organic hydroperoxides [9,10]. pGPx adopts an intermediate specificity in acting on hydroperoxides of phospholipids, but not of cholesterol esters [11]. The specificity for GSH decreases in the order cGPx >GI-GPx >pGPx » P H G P x : 1) The k2 values, which reflect affinity to GSH are at least one order of magnitude higher for cGPx [12] than for pGPx [13,14] and PHGPx [15]; 2) pGPx is equally active with GSH, thioredoxin and glutaredoxin [16]; 3) PHGPx oxidises various dithiols [17], but nevertheless prefers GSH and acts on various protein thiols only when GSH becomes limiting [18-20] (see Chapter 28). 4) the P. falciparum GPx homolog clearly prefers thioredoxins over GSH [21]; and 5) the GPx homolog of Trypanosoma brucei uses tryparedoxin as reductant [22]. Differential response to selenium Glutathione peroxidases occupy extreme positions in the "hierarchy of selenoproteins'' , with cGPx ranking lowest [5]. The extracellular GPx behaves similar to cGPx [5]. PHGPx, is fairly resistant against selenium deficiency and is rapidly synthesized when selenium is replenished [5,23]. GI-GPx is most resistant to selenium deficiency [24]. The relative position of the selenoproteins within this hierarchy was believed to reflect their relative biological importance. The selenium-responsiveness of the four GPx types, however, conflicts with this assumption. GI-GPx and cGPx, reflecting the two extremes on the scale, have been knocked out without creating any overt pathological effects in unstressed mice [25-27].

Selenoproteins of the glutathione system

163

The molecular mechanisms leading to differential synthesis of selenoproteins under selenium restriction are far from being clear. As a rule, the ranking parallels the stability of the pertinent mRNA [24,28-30]. The instability of cGPx mRNA has been attributed to "cytoplasmic nonsensemediated decay'' , a phenomenon describing the elimination of mRNA species having a stop codon located at a certain distance upstream from a premRNA splice site [31]. However, the mRNAs of PHGPx and GI-GPx have their UGA in homologous positions and remain stable or are even increased in selenium deficiency. Differences in translational efficiencies do not account for the different stabilities either and maximum SECTS efficiencies do not correlate with the selenium-responsiveness of SECTS efficiencies [58]. Likely, the mRNAs are specifically stabilized by a factor binding to its SECTS in a selenium-dependent manner. The SECTS-binding protein SBP2 itself can hardly be responsible for the differential mRNA stabilities, since it is not known to directly interact with any cellular selenium compound. Also, the dissociation constants for the reaction of SBP2 with different GPx 3'UTRs, as measured by mobility shift assays [32] or by surface plasmon resonance (own unpublished data), are similar. Likely, the stabilizing factor is the eukaryotic SelB homolog EFsec [33,34] that may modulate the SECTS affinity of SBP2 in response to binding selenocysteyl-loaded tliNA^^"'^*'^ in the ribosomal super complex [33-35]. According to this hypothesis, selenocysteine-loaded tRNA[^''^^*'^ would reflect the cellular selenium status and EFsec is the selenium sensor (see Chapters 6 and 7). Se-independent regulation of glutathione peroxidases cGPx: The cGPx gene has been reported to contain an oxygen-responsive element in the 5'-flanking region, and the targeting binding protein 'OREBP' responding to oxygen tension has been characterized. Further, human cardiomyocytes [37] and endothelial cells [38] respond to hyperoxia with an up-regulation of cGPx. However, convincing data on a transcriptional response of gpx-1 to oxidative stress in general does not exist [39]. Typically, the capacity of the cGPx/GSH system is enhanced by induction of yglutamyl-cysteine synthetase upon exposure to peroxides or redox cyclers via the NrG/Keapl system [40,41]. Most consistently, cGPx is up-regulated by estrogens [42-45]. A typical estrogen-responsive element is, however, not detectable in gpx-1. The estrogen response of gpx-J is thought to result from estradiol-mediated activation of N F K B that targets a putative binding site of gpx-1 [43,45]. Tn neutrophils cGPx was shown to depend on the transcription factor PU.l. Binding of PU.l to putative PU.l sites in the 5' promoter and the 3' flanking region of gpx-1 was demonstrated by gel shift assays and transactivation [46]. The gpx-1 promoter was further reported to be activated by p53 in human osteogenic cell lines [47]. Tn RXR-negative mice cGPx expression was

164

Selenium: Its molecular biology and role in human health

reduced to 30%, as was that of y-glutamyl-cysteine synthetase, suggesting a regulation of cGPx via the retinoid X receptor [48]. At the translational level, cGPx biosynthesis is down-regulated by homocysteine, which may interfere with the selenium-dependent readthrough at the UGA codon [49]. Homocysteine-induced endothelial dysfunction can be ameliorated by overexpression of cGPx [50] and basal endothelial dysfunction is more severe in cGPx knockout mice [51]. pGPx: pGPx was found to be increased in plasma of patients with bowel disease and of mice with experimental colitis [52] as well as in the epithelial lining fluid in the lungs of asthmatic patients [53]. The increased levels of pGPx were considered to result from transcriptional up-regulation due to oxidative stress. A functional consensus sequence for the redox-regulated transcription factor AP-1 is located in the 5' promoter region of gpx-3 [53]. More recently, an alternative transcription start in gpx-3 was identified [54]. The pertinent promoter contains putative binding sites for SP-1 and the hypoxia-inducible factor-1 (HIF-1), a redox-sensitive metal response element and an antioxidant response element. Experimentally, hypoxia was demonstrated to strongly up-regulate gpx-3 expression in Caki-2 cells [54]. GI-GPx: Several caudal homeobox protein binding sequences and two retinoic acid responsive elements were identified in gpx-2 [55], and GI-GPx could be induced by retinoic acid in some (MCF-7), but not all (HT29) cells. Further, GI-GPx was down-regulated in hepatoma cells infected with hepatitis C virus subgenomic RNA. Inversely, induction of GI-GPx by retinoic acid suppressed the HCV replicon [56] suggesting a therapeutically interesting inverse relationship of GI-GPx levels and viral replication. In several microarray studies, gpx-2 transcripts were found to be elevated together with phase II enzyme transcripts upon exposure to the Nrf2 activator sulforaphane or to hyperbaric oxygen [57,58], In gpx-2 a ATG-proximal conserved ARE proved to be indispensable to endogenous and sulforaphaneinduced gpx-2 expression at the transcript and protein level. The relevance of the pertinent NrfZ/Keapl system was corroborated by enhancement of a gpA;2 promoter-driven reporter gene expression, by transfection with Nr£2 and suppression thereof by transfection with Keapl [59]. PHGPx: Related studies of regulatory phenomena do not yet yield a comprehensive picture that could explain the extraordinary tissue distribution, the differentiation-specific expression of PHGPx or the relative abundance of its isoforms. Functionality of putative hormone-responsive elements in gpx-4 could not be verified in somatic cell lines. Instead, the known dependence of testicular PHGPx on gonadotropic hormones [60] could be explained by abundant expression of PHGPx in spermatogenic cells that proliferate under stimulation by testosterone [61]. A selective upregulation of PHGPx in the bovine oviduct by 17p estradiol was, however, verified, although without a mechanistic explanation [62]. In rat casein-

Selenoproteins of the glutathione system

165

elicited neutrophils, PHGPx was found up-regulated. The results were interpreted as induced self-protection against the oxidant products of these cells. The effect could be mimicked by recombinant growth-regulated oncogene (GRO) and abrogated by anti-GRO antibodies [63]. Most recent studies focus on the expression of the nuclear isoform of PHGPx. Maiorino et al [64] and Moreno et al [65] demonstrated a basal promoter activity of the first intron oigpx-4 resulting in the nucleus-specific transcript, while Borchert et al [66] and Ufer et al [67] found no or marginal activity of this region and postulated that also the expression of nuclear PHGPx is essentially triggered by the promoter region upstream of the ATG codons that represent the translation starts of the cytosolic and the mitochondrial forms. Here, binding sites for SPl, nuclear factor Y (NF-Y) and members of the Smad family were identified [67], while expressiondepressing sites for EGRl and SREBPl were detected in the first intron [66]. Tramer et al [68] confirmed the promoter activity of the first intron and there identified the cAMP-response element modulator x (CREM-x) as an essential activator. According to these authors, the expression of nuclear PHGPx in spermatogenic cells is explained by high levels of CREM-t in pachytene spermatocytes and spermatids [68]. None of the quoted studies, however, explains the burst of normal PHGPx expression in spermatids that is evidently more important to sperm function than the faint expression of nuclear PHGPx (see Chapter 28). Functional diversification of glutatliione peroxidases Lessons from knock out and overexpression of gpx-1 The cGPx knock-out mouse reveals that cGPx is little else than an emergency device to cope with hydroperoxide challenge. Unchallenged cGFx^''^ mice developed normally, even grew faster and tolerated elevated oxygen tension [25]. The lack of any overt phenotype is not surprising in light of the clinical phenotype of patients with deficiencies in GSH regeneration. Such patients are normal as long as they are unchallenged with hydroperoxides and the related genetic defects were accordingly rated as "non-diseases'' [70]. The complete lack of cGPx in mice is tolerated like the more or less pronounced impairment of GSH regeneration in patients. Like these, the cGPx^"'"' mice are however highly susceptible to oxidative damage, as has been shown by exposure to redox-cycling herbicides [26,71,72] and to lipopolysaccharides that trigger oxidative burst in phagocytes [73]. The cGPx^"'"^ mouse also demonstrates that the seemingly normal life without cGPx is threatened by certain environmental hazards. Most importantly, the cGPx*''^ mouse proved to be a model for the classical human selenium-deficiency syndrome Keshan disease. When these mice were exposed to a non-virulent Coxsackie strain, the virus rapidly mutated into a

166

Selenium: Its molecular biology and role in human health

virulent form [74], as had been observed before with selenium-deficient mice [75]. Based on these studies, the debate whether Keshan disease is caused by selenium-deficiency itself or a viral infection [76] can likely be settled. A decrease in cGPx, the enzyme responding fastest to selenium deprivation, results in elevated steady-state-levels of hydroperoxides which in turn accelerate the mutation rate of usually benign virus strains and allow virulent mutants to become dominant. The experiments with oxidatively challenged cGPx^"'"' mice were performed in selenium-deprived and selenium-adequate animals and, as a rule, the selenium-status did not significantly alter the results. This surprising outcome allows two alternative interpretations: Either the selenium deficiency was not sufficient to decrease other selenoproteins to any relevant degree or, more likely, none of the remaining selenoproteins can efficiently substitute for cGPx in balancing a systemic oxidative challenge. The latter conclusion may be viewed as provocative in light of the occurrence of four more selenoperoxidases and of the metabolic link of thioredoxin reductases to peroxiredoxin-type peroxidases. Optimization of cGPx activity, e.g., by selenium supplementation or GPx mimics, is widely considered to be beneficial. Some observations, however, cast doubt on this view: 1) in HIV-infected tissue culture, cGPx overexpression enhanced viral spread [77]; 2) knockout of cGPx enhanced resistance against kainic acid-induced epileptic seizures [78]; 3) overexpression of cGPx reduces TNFa-induced NPKB activation in tissue cultures [79], as does selenium-supplementation [80,81]; 4) overexpression of cGPx promoted acetaminophen toxicity [82]; 5) hepatocytes of cGPx^'"^ mice, while being highly sensitive to oxidative stress, were largely protected against peroxynitrite challenge [83]; 6) strikingly, an increase of skin cancer tumor incidence in GPx overexpressing transgenic mice was induced by DMBA/TPA treatment, an experimental design commonly believed to involve ROS-induced/promoted carcinogenesis [84]; and 7) cGPx*'"^ mice consistently developed hyperglycemia and insulin resistance, as is typical of type II diabetes that is commonly believed to be facilitated by oxidative stress [85]. Evidently evolution has created a ratio of peroxide generating and detoxifying systems that meets the demands of most, though not all, endogenous and environmental conditions, and it would be unwise to disturb this delicate balance without proven need. pGPx and the extracellular peroxide tone pGPx has only a limited opportunity to counter a serious challenge by hydroperoxides, since the tiny concentrations of extracellular GSH or thioredoxin would quickly be consumed in absence of any regenerating system. The biological role of pGPx therefore still remains speculative. pGPx has been implicated in the reduction of lipid hydroperoxides LDL [3] and

Selenoproteins of the glutathione system

167

thus might be relevant to Steinberg's ideas on atherogenesis [86]. In view of the low steady-state level of peroxidized lipoproteins, the limited reduction capacity of pGPx may just suffice. The substrate specificity of pGPx, however, is not ideal for the reduction of peroxidized LDL, since it does not reduce peroxidized cholesterol esters [11]. Alternatively, the enzyme could reduce soluble lipid hydroperoxides, which have been implicated in the activation of cyclooxygenase, the key enzyme of prostaglandin synthesis [87]. Similarly, the activation of other lipoxygenases, which typically remain dormant in the absence of any hydroperoxides, could be prevented by pGPx, as occurs in vitro by cGPx [88] and PHGPx [89-92]. Similarly, the extracellular pGPx/GSH system could be regarded as a redox buffer that is required to discriminate between irrelevant and serious inflammatory stimuli [5]. Specifically, a fast acting peroxidase in a compartment with low reducing capacity could make the best use of the oxidative host defense machinery, which is indispensable for survival in a hostile environment, but selfdestructive when over-reacting [5]. GI-GPx as modulator of inflammatory responses of the intestine GI-GPx, because of its preferred localization in the intestine, has been proposed to prevent systemic access of food-bom peroxides, which could pre-exist in food or are generated by the intestinal flora or by the mucosa when metabolising xenobiotics [93]. Experimental evidence in support of this concept is scarce. GI-GPx^"'"' mice appear to have a normal phenotype [27], but mice deficient in both, cGPx and GI-GPx, showed retarded growth after weaning, suffered from ileocolitis starting at day 11 of age [94], and develop intestinal cancer [95]. Interestingly, the development of both, ileocolitis and tumors depended on, or was essentially aggravated by, colonization of the intestine by bacteria. A single allele of gpx-2 in the double transgenic mice was sufficient to ameliorate the pathologies, while a single allele of gpx-1 was ineffective [95-97]. The observations point to a pivotal role of GI-GPx in counteracting inflammatory responses elicited by the intestinal microflora. The intimate relationship of GI-GPx and the gastrointestinal flora is further evidenced by an induction of GI-GPx upon colonization of the intestine [96]. Pathological changes that occur in the double knockout suggest a mutual complementation of GI-GPx and cGPx. The enzymes could, e.g., cooperate in redox-regulated processes such as proliferation or apoptosis. Inhibition of apoptosis has been documented for the overexpression of cGPx [77] and PHGPx [98-100] and may be common to all selenoperoxidases. The intestinal epithelium displays a steep gradient of GI-GPx that declines from the proliferating stem cells in the lower crypts to the luminally exposed cells gradually undergoing apoptosis [101,102]. If this delicate balance of events is indeed regulated by GI-GPx, the absence of cGPx could unmask a " dysregulation'' that is still compensated in the isolated GI-GPx knock-

168

Selenium: Its molecular biology and role in human health

out. The intriguingly high concentration of GI-GPx in Paneth cells, which are involved in mucosal immunity and not in absorption, and its association in colon cells with vesicular structures [102] also suggest highly specialized functions that still remain to be elucidated. PHGPx in lipid peroxidation, inflammation, and differentiation PHGPx was discovered and characterized as an enzyme preventing progression of lipid peroxidation in biomembranes due to its unique ability to reduce hydroperoxo-groups in complex lipids [103]. In this context, it is synergistically supported by -tocopherol, which reduces lipid peroxy radicals to lipid hydroperoxides. The latter, if not reduced by PHGPx, would reinitiate free radical-mediated lipid peroxidation by Haber-Weiss- or Fentontype chemistry. Over the years, however, growing evidence indicated that protection of biomembranes against unspecific lipid peroxidation is possibly not the most important role of PHGPx. A knockout mouse model, which could shed light on the multiple roles of PHGPx, has not yet been obtained. Homozygous PHGPx^''^ mice died between day 7.5 and 8.5p.c.[104,105]. Specific functions of the enzyme are however evident from overexpression and selenium supplementation studies. With other glutathione peroxidases, PHGPx shares the ability to silence lipoxygenases [106,107], to inhibit apoptosis [98-100], and to suppress cytokine-induced N F K B activation [108]. In several cases, however, PHGPx was demonstrated to be the biologically most relevant regulator: 1) Seleniumdeficient rat basophilic leukemia cells as well as whole animals overproduce 5-lipoxygenase products comprising the potent pro-inflammatory leukotrienes [106]. Similarly, leukotriene biosynthesis is suppressed in transformed rat basophilic leukemia cells selectively overexpressing PHGPx [109]. PHGPx thus appears to be the principal selenoperoxidase in charge of silencing 5-lipoxygenase. 2) A moderate overexpression of PHGPx in the human ECV cell line completely abrogated interleukin-1-induced N F K B activation, while a huge variation of cGPx activity achieved by deprivation and re-supplementation of selenium did not [108]. N F K B activation induced by hydroperoxides is similarly suppressed by overexpression of PHGPx in rabbit aortic smooth muscle cells [99] and UV-stressed human skin fibroblasts [110]. 3) The pivotal role of PHGPx in dampening inflammation is further underscored by its ability to suppress COX-2 expression/activity, to prevent COX2 product-dependent malignant growth [111] and to inhibit VCAM-1 expression by inducing heme oxygenase-1 [112]. The dampening of inflammatory responses thus is common to all the GPxs, but is in part achieved by different, although complementary mechanisms. A potential cross-talk between PHGPx and GI-GPx in the management of exogenous stressors is outlined in Scheme 1. How the peculiar specificity of PHGPx in regulating inflammatory processes is achieved is unknown. It may

Selenoproteins of the glutathione system

169

be due to its preference for hydrophobic lipids or to subcellular micro compartmentation. The observation that

Mic riJC-iDla

EI«fctryip*iHBi DKidalwB stress e g 11.. 1 f*.i| mmibmem^

W

Toll lice

m

1

1

jr SH SM

ROOH

C»X-l.l.03Cl.«tl. ."• -•.VI"»:V:4

Cnrr; •iivi(.;r.'i!ysis;

Irifkirrnn" o i

AlherosclETO'ii'S

Scheme 1. Role of GI-GPx and PHGPx in inflammatory (right) and adaptive (left) responses. The N F K B system, which is activated by pro-inflammatory cytokines and bacterial toxins via Toll-like receptors TLR), is activated by reactive oxygen species, in particular by hydroperoxides, and dampened by selenium-containing peroxidases, in particular PHGPx. The Nr£2/Keapl system is activated by dietary electrophiles and moderate oxidative challenge, promotes ARE-mediated expression of phase 2 enzymes and GI-GPx, and is supposed to enforce resistance against inflammation and cancer.

PHGPx is sometimes found oxidatively cross-linked to itself or other proteins [18,19] suggests another possibility: PHGPx, by a reaction of its oxidized selenium with accessible protein thiols, could also act as a peroxidedependent thiol modifying agent. By an analogous reaction, PHGPx polymerizes in absence of GSH and can thereby be transformed into a

170

Selenium: Its molecular biology and role in human health

structural protein. This process has been shown to be pivotal to the differentiation of mammalian spermatids into mature spermatozoa as is described in detail in Chapter 28. Acknowledgements The preparation of this article was supported by the Deutsche Forschungsgemeinschaft DFG (grants Fl 61/12-3 and Bri 778/5-3). References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35.

L Floh6, JR Andreesen, R Brigelius-Flohe et al 2000 lUBMB Life 49:411 RF Burk, RA Lawrence 1978 Functions of glutathione in liver and kidney H Sies, A Wendel (Eds) Springer Verlag, Berlin, p 114 Y Saito, T Hayashi, A Tanaka et al 1999 J Biol Chem 274:2866 JW Chen, C Dodia, SI Feinstein et al 2000 J Biol Chem 275:28421 L Flohe, R Brigelius-Floh^ 2001 Selenium Its molecular biology and role in human health 1st "dition DL Hatfield (Ed) Kluwer Academic Publishers, Boston, Dordrecht, London, p 157 O Epp, R Ladenstein, A Wendel 1983 EurJBiochem 133:51 B Ren, W Huang, B Akesson et al 1997 JMol Biol 268:869 L Floh6, KD Aumann, R Brigelius-Floh6 et al 1993 Active Oxygen, Lipid Peroxides, and Antioxidants K Yagi (Ed) CRC press, Boca Raton, p 299 A Grossmann, A Wendel 1983 EurJBiochem 135:549 A Sevanian, SF Muakkassah-Kelly, S Montestruque 1983 Arch Biochem Biophys 223:441 Y Yamamoto, K Takahashi 1993 Arch Biochem Biophys 305:541 L Floh6, G Loschen, WA Gunzler et al 1972 Hoppe Seylers ZPhysiol Chem 353:987 RS Esworthy, FF Chu, P Geiger et al 1993 Arch Biochem Biophys 307:29 G Takebe, J Yarimizu, Y Saito et al 2002 J Biol Chem 277:41254 F Ursini, M Maiorino, C Gregolin 1985 Biochim Biophys Acta 839:62 M Bjoemstedt, J Xue, W Huang et al 1994 J Biol Chem 269:29382 A Roveri, M Maiorino, C Nisii et al 1994 Biochim Biophys ^cto 1208:211 C Godeas, F Tramer, F Micali et al 1996 Biochem Mol Med 59:118 F Ursini, S Heim, M Kiess et al 1999 Science 285:1393 M Maiorino, L Flohe, A Roveri et al 1999 BioFactors 10:251 H Sztajer, B Gamain, K-D Aumann et al 2001 J Biol Chem 276:7397 T Schlecker, A Schmidt, N Dirdjaja et al 2005 J Biol Chem 280:14385 F Weitzel, F Ursini, A Wendel 1990 Biochim Biophys Acta 1036:88 K Wingler, M Bocher, L Flohe et al 1999 EurJBiochem 259:149 YS Ho, JL Magnenat, RT Bronson et al 1997 J Biol Chem 272:16644 JB de Haan, C Bladier, P Griffiths et al \99% J Biol Chem 273:22528 RS Esworthy, JR Mann, M Sam et al 2000 Am J Physiol Gastrointest Liver Physiol 279:G426 RA Sunde, JA Dyer, TV Moran et al 1993 Biochem Biophys Res Commun 193:905 XG Lei, JK Evenson, KM Thompson et al 1995 JNutr 125:1438 G Bermano, JR Arthur, JE Hesketh 1996 FEBSLett 387:157 X Sun, PM Moriarty, LE Maquat 2000 EMBO J19:4734 JE Fletcher, PR Copeland, DM Driscoll et al 2001 RNA 7:1442 RM Tujebajeva, PR Copeland, XM Xu et al 2000 EMBO Rep I ASS D Fagegaltier, N Hubert, K Yamada et al 2000 EMBO J19:4796 AM Zavacki, JB Mansell, M Chung et al 2003 Mol Cell 11:773

Selenoproteins of the glutathione system 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87.

171

DB Cowan, RD Weisel, WG Williams et al 1993 J Biol Chem 268:26904 F Merante, SM Altamentova, DA Mickle et al 2002 Mot Cell Biochem 229:73 L Jomot, AF Junod 1995 5/oc/(em 7306:581 S Desaint, S Luriau, JC Aude et al 200475/0/ Chem 279:31157 M McMahon, K Itoh, M Yamamoto et al 2001 Cancer Res 61:3299 DM Krzywanski, DA Dickinson, KE lies et al 2004 Arch Biochem Biophys 423:116 RE Pinto, W Hartley 1969 Biochem J115:449 J Vina, C Borras, J Gambini et al 2005 FEBSLett 579:2541 ID Capel, AE Smallwood 1983 Res Commm Chem Pathol Pharmacol 40:367 C Borras, J Gambini, MC Gomez-Cabrera et al 2005 Aging Cell 4:113 SL Throm, MJ Klemsz 2003 JLeukoc Biol 74:111 M Tan, S Li, M Swaroop et al 1999 J Biol Chem 274:12061 Y Wu, X Zhang, F Bardag-Gorce et al 2004 Mol Pharmacol 65:550 DE Handy, Y Zhang, J Loscalzo 2005 J Biol Chem 280:15518 N Weiss, YY Zhang, S Heydrick et al 2001 Proc Natl Acad Sci USA 98:12503 MA Forgione, N Weiss, S Heydrick et al 2002 Am J Physiol Heart Circ Physiol 282;H1255 DM Tham, JC Whitin, HJ Cohen 2002 PediatrResS\M\ SA Comhair, PR Bhathena, C Farver et al 2001 FASEB J 15:70 C Bierl, B Voetsch, RC Jin et al 2004 J Biol Chem 279:26839 FF Chu, RS Esworthy, L Lee et al 1999 JNutr 129:1846 M Morbitzer, T Herget 2005 J Biol Chem 280:8831 RK Thimmulappa, KH Mai, S Srisuma et al 2002 Cancer Res 62:5196 HY Cho, SP Reddy, A Debiase et al 2005 Free Radic Biol Med 38:325 A Banning, S Deubel, D Kluth et al 2005 Mol Cell Biol 25:4914 A Roveri, A Casasco, M Maiorino et al 1992 J Biol Chem 267:6142 M Maiorino, JB Wissing, R Brigelius-Flohe et al 1998 FASEB 712:1359 J Lapointe, S Kimmins, LA Maclaren et al 2005 Endocrinology 146:2583 H Hattori, H Imai, A Hanamoto et al 2005 Biochem J2,2,9:219 M Maiorino, M Scapin, F Ursini et al 2003 J Biol Chem 278:34286 SG Moreno, G Laux, M Brielmeier et al 2003 Biol Chem 384:635 A Borchert, NE Savaskan, H Kuhn 2003 J Biol Chem 278:2571 C Ufer, A Borchert, H Kuhn 2003 Nucleic Acids Res 31:4293 F Tramer, A Vetere, M Martinelli et al 20045ioc/(emy383:179 PR Copeland, DM Driscoll 1999 J Biol Chem 274:25447 E Beutler 1983 Biomed Biochim Acta 42:S234 WH Cheng, YS Ho, BA Valentine et al 1998 JA^w?/-128:1070 Y Fu, WH Cheng, JM Porres et al 1999 Free Radic Biol Med ll-MS H Jaeschke, YS Ho, MA Fisher et al 1999 Hepatology 29:443 MA Beck, RS Esworthy, YS Ho etal 1998 f.^SESy 12:1143 MA Beck, Q Shi, VC Morris et al 1995 Nat Med 1:433 H Guanqing 1979 Chin Med J (Engl) 92:416 PA SandstrOm, J Murray, TM Folks et al 1998 Free Radic Biol Med 24:1485 D Jiang, G Akopian, YS Ho et al 2000 Exp Neurol 164:257 C Kretz-Remy, P Mehlen, ME Mirault et al 1996 J Cell Biol 133:1083 V Makropoulos, T Bruning, K Schulze-Osthoff 1996 Arch Toxicol 70:277 K Hon, D Hatfield, F Maldarelli et al 1997 AIDS Res Hum Retroviruses 13:1325 O Mirochnitchenko, M Weisbrot-Lefkowitz, K Reuhl et al 1999 J Biol Chem 274:10349 Y Fu, H Sies, XG Lei 2001 J Biol Chem 276:43004 YP Lu, YR Lou, P Yen et al 1997 Cancer Res 57:1468 JPMcClung,CARoneker,WMu etal 2004PracAfa/Z^carf5a t / S ^ 101:8852 D Steinberg 1997 J Biol Chem 272:20963 RJ Kulmacz, WE Lands 1983 Prostaglandins 25:531

172 88. 89. 90. 91. 92. 93. 94. 95. 96. 97. 98. 99. 100. 101. 102. 103. 104. 105. 106. 107. 108. 109. 110. 111. 112.

Selenium: Its molecular biology and role in human health M Haurand, L Flohe 1988 Biol Chem Hoppe Seyler 369:133 WC Chang 2003 JBiomed Sci 10:599 CJ Chen, HS Huang, WC Chang 2003 FASEBJMA 694 CJ Chen, HS Huang, SB Lin et al 2000 Prostaglandins Leukot Essent Fatty Acids 62:261 H KOhn, A Borchert 2002 Free Radic Biol Med 33:154 DA Parks, GB Bulkley, DN Granger 1983 Surgery 94:428 RS Esworthy, R Aranda, MG Martin et al 2001 Am J Physiol Gastrointest Liver Physiol 281:0848 FF Chu, RS Esworthy, PG Chu et al 2004 Cancer Res 64:962 RS Esworthy, SW Binder, JH Doroshow et al 2003 Biol Chem 384:597 FF Chu, RS Esworthy, JH Doroshow 2004 Free Radic Biol Med 36:1481 K Nomura, H Imai, T Koumura et al 1999 J Biol Chem 274:29294 R Brigelius-Floh6, S Maurer, K L6tzer et al 2000 Atherosclerosis 152:307 H Imai, T Koumura, R Nakajima et al 2003 Biochem J 371:799 FF Chu, RS Esworthy 1995 Arch Biochem Biophys 323:288 S Florian, K Wingler, K Schmehl et al 2001 Free RadRes 35:655 F Ursini, M Maiorino, M Valente et al 1982 Biochim Biophys Acta 710:197 H Imai, F Hirao, T Sakamoto et al 2003 Biochem Biophys Res Commun 305:278 LJ Yant, Q Ran, L Rao et al 2003 Free Radic Biol Med 34:496 F Weitzel, A Wendel 1993 J Biol Chem 268:6288 K Schnurr, J Belkner, F Ursini et al 1996 J Biol Chem 271:4653 R Brigelius-Flohe, B Friedrichs, S Maurer et al 1997 Biochem J 328:199 H Imai, K Narashima, M Aral et al 1998 J Biol Chem 273:1990 J Wenk, J SchUller, C Hinrichs et al 2004 J Biol Chem 279:45634 I Heirman, D Ginneberge, R Brigelius-Flohe et al 2005 Free Radic Biol Med, in press: A Banning, R Brigelius-Floh6 2005 Antioxid Redox Signal 7:889

Chapter 16. New roles of glutathione peroxidase-1 in oxidative stress and diabetes Xin Gen Lei Department of Animal Science, Cornell University, Ithaca, NY 14853, USA

Wen-Hsing Cheng Laboratory of Molecular Gerontology, National Institute on Aging, National Institutes of Health, Baltimore, MD 21224, USA

Summary: Glutathione peroxidase-1 (GPXl) was identified as an antioxidant enzyme in 1957 and as a selenoprotein in 1972. In the last ten years, ample data have been generated from several lines of GPXl knockout mice, showing an essential role of GPXl in defending against severe oxidative stress mediated by pro-oxidants. This protection is associated with attenuated oxidation of NADPH, NADH, lipid, and protein. When Sedeficient mice are under mild oxidative stress, minute amounts of GPXl activity in tissues are able to protect them against the pro-oxidant-induced lethality and hepatic aponecrosis. Strikingly, GPXl prevents apoptosis of hepatocytes caused by reactive oxygen species, but potentiates cell death caused by reactive nitrogen species. Mice overexpressing GPXl developed insulin resistance and obesity. These new data illustrate mixed roles of GPXl in coping with different types of oxidative stress, and suggest a possible deleterious impact of GPXl up-regulation on glucose metabolism. Introduction Selenium (Se) is of fundamental importance to human health, and performs its metabolic functions presumably in the form of selenoproteins. It was not until 1957 that Se was considered essential in the diet for rats [1]. Coincidentally, this was the same year that cellular glutathione peroxidase (glutathione: H2O2 oxidoreductase, EC 1.11.1.9, GPXl) was found to protect erythrocytes against oxidative hemolysis, as erythrocytes from Se-deficient rats were more prone to hemolysis upon hydrogen peroxide exposure than those from Se-adequate rats [2]. In the 1970s, Se was shown to be cofractioned with GPXl activity in Se-deficient rats after Se^^ administration [3,4]. Consequently, GPXl became the first identified selenoprotein and was

174

Selenium: Its molecular biology and role in human health

considered the biochemical mediator of Se in protecting erythrocytes against oxidative hemolysis. GPXl is expressed in virtually all tissues and represents the majority of body Se [5]. As a homotetramer of 88-kDa, the enzyme shares 30-60% sequence homology with other Se-dependent glutathione peroxidases. The human gpxl gene was mapped on chromosome 3ql l-3ql3.1[6]. The protein sequence of human GPXl is 86 and 83% homologous to its counterparts in mice and rats, respectively. Cloning of the mouse Gpxl gene in 1986 led to the identification of employing the stop codon UGA for translation of selenocysteine, the 21*' amino acid [7]. The process requires a stem-loop structure called a selenocysteine insertion sequence element in the 3' untranslated region of mRNA [8,9]. Nutritionally, Se deficiency in rats results in a 90% loss of liver GPXl mRNA and even greater loss of GPXl activity [10]. The sensitivity of GPXl activity to body Se fluctuation renders it a convenient and responsive biochemical marker to assess body Se status and dietary Se requirements. However, manipulating tissue GPXl activity by altering dietary Se intake in conventional animal models does not allow studies for specific or exclusive roles of GPXl, due to possible confounding effects of multiple selenoproteins. Several GPX mimics and inhibitors are available, but are not highly specific or completely satisfactory [11,12]. Thus, it was necessary to develop GPXl knockout [GPXl(-/-)] and overexpression [GPX1(+)] mice for an accurate assessment of the physiological functions of GPXl. Up to now, three lines of GPX1(-/-) mice [13-15], and at least two lines of GPX1(+) mice [16,17] have been generated. Effect of GPXl null or overexpression on selenoprotein expression Two features of GPXl argued against an essential role for the enzyme in vivo. First, GPXl activity was more susceptible to dietary Se deprivation than other selenoproteins [18,19]. Second, there was no apparent adverse effect on animal health when GPXl activity in tissues dropped to < 1% of controls. In addition, GPXl is the predominant biochemical form of Se in many tissues [5]. Thus, GPXl was proposed as a storage or buffer of body Se that could mobilize its own Se for the synthesis of other selenoproteins in Se deficiency, but takes up cellularfireeSe in Se adequacy [20,21]. To test this hypothesis, we determined expression of several selenoproteins and concentrations of total Se in tissues of GPX1(-/-) and GPX1(+) mice fed Se-deficient, -adequate, or -excessive diets for 5-13 weeks [17,22,23]. We observed that knockout of GPXl reduced liver Se concentration by 60% in Se-adequate mice [22,23]. However, knockout or overexpression of GPXl in mice did not alter the mRNA and (or) activity expression of GPX3, GPX4, thioredoxin reductase, and selenoprotein P in various tissues, irrespective of their dietary Se concentrations. If GPXl were truly a buffer of body Se, there should have been more Se available for a higher activity expression of GPX3,

New roles of glutathione peroxidase-1

175

GPX4, or thioredoxin reductase in the GPX1(-/-) than in the wild-type mice fed Se-deficient diets. Likewise, decreases in activity expression of these selenoenzymes in dietary Se deficiency could have been, at least, partially corrected by GPXl overexpression compared with wild-type controls. Clearly, our data do not support the buffer role of GPXl in Se partitioning for selenoprotein synthesis, but suggest that expression of GPXl is independent of other selenoproteins. Recent searching for selenoprotein genes among the human genome by Gladyshev and colleagues indicates that there are a total of 25 selenoproteins in humans and 24 in rodents [24]. It remains to be seen whether knockout or overexpression of GPXl affects the expression of the newly identified selenoproteins. Overall metabolic impact of GPXl null or overexpression When GPX1(-/-) mice were fed a Se-adequate diet in our animal facility, they were healthy and fertile, showing normal body weight gain, food intake, and no abnormal histology in various organs [13,23]. On the contrary, we observed insulin resistance and obesity in male, GPX1(+) mice 24 weeks of age, compared to their age-matched wild-type controls [25]. Although the GPX1(-/-) mice developed in another study showed a 20% reduction of body weight by 8 months of age compared to controls [15], the GPX1(-/-) mice from a third study did not display a body weight change [14]. Another reported phenotype of GPX1(-/-) mice is the age-related onset of cataracts. Reddy et al. [26] observed a higher incidence of cataracts in GPX1(-/-) mice (90%) than in wild-type mice (20%) of 15 months at age. However, Spector et al. [27] failed to detect cataracts in old GPX1(-/-) mice. Early research on GPXl function Overexpression of GPXl in cultured human cells enhanced their resistance to pro-oxidant-induced oxidative stress [28-30]. Approximately an 80% reduction in GPXl by antisense RNA in cells increased their sensitivity to selective genotoxic oxidants [31]. Apparently, these in vitro results show the biological potential of GPXl as an antioxidant enzyme, but could not be extrapolated as to its physiological function. The in vivo antioxidant roles of GPXl need to be verified using whole animal models. Paraquat and diquat have been used to study antioxidant roles of Se in several species [11,32,33]. Following acute exposures of animals to these compounds, the major target organ for diquat and paraquat is liver and lung, respectively [34]. Both pro-oxidants initiate reactive oxygen species (ROS) generation via the conversion of molecular oxygen into superoxide radicals [34,35]. The toxic superoxide radicals are converted into hydrogen peroxide by superoxide dismutase. Glutathione peroxidase and catalase then reduce hydrogen peroxide into water. Excessive hydrogen peroxide may react with iron to produce highly reactive hydroxyl fi-ee radicals that destroy large

176

Selenium: Its molecular biology and role in human health

biomolecules, leading to cellular dysfunction, multiple organ failures, and death in animals. Thus, the paraquat or diquat-mediated hydrogen peroxide production offers an excellent test of the in vivo antioxidant role of GPXl. It was clear that prior intraperitoneal (ip) injections of Se into Se-deficient rats prevented diquat or paraquat-induced lipid peroxidation and liver necrosis [32]. However, the metabolic mediator conferring the Se protection remained unclear. Because of differences in their repletion profiles after the Se administration, selenoprotein P, but not GPXl, was considered responsible for the protection by the injected Se [36]. In addition, selenoprotein P was found in endothelial cells of liver in rats [37] and was shown to exhibit peroxidase activity [38]. Although these data suggested a possible involvement of selenoprotein P in the Se protection against diquat or paraquat toxicity, the inherent limitations of the animal model could not allow a conclusion on the role of selenoprotein P or an exclusion of GPXl from the protection. In contrast, the recently-developed GPXl [13-15] or selenoprotein P knockout mice [39,40] provide us with unprecedented models for direct answers to these questions. Role of GPXl in severe oxidative stress We used GPX1(-/-) and GPX1(+) mice to determine the importance and contribution of GPXl, in relation to other selenoproteins, to the body defense against severe or lethal oxidative stress [14,41-44]. After an ip injection of paraquat at a dose of 50 mg/kg body weight, all GPX1(-/-) mice, regardless of their body Se status, died within 4-6 hours [41]. While the Se-deficient wild-type mice also died within 4-6 hours after paraquat injection, the Seadequate wild-type mice survived 3 days [41]. Apparently, GPXl was the mediator of body Se for the protection against the paraquat-induced lethality, and the survival time of mice was a function of tissue GPXl activity. Consistently, de Hann et al. [14] showed hypersensitivity of GPX1(-/-) mice to paraquat injection (ip) at a dose of 30 mg/kg body weight. In another experiment conducted by us, Se-adequate wild-tjqje mice survived an ip injection of diquat at a dose of 24 mg/kg body weight, whereas Se-adequate GPX1(-/-) and Se-deficient wild-type mice died within 3-4 hours [42]. When Se-adequate GPX1(+) mice were given an ip injection of paraquat at 125 mg/kg body weight, their mean survival time was 10-fold longer (59 vs. 5.8 h) than that of wild-type mice. Altogether, the resistance of mice to prooxidant-induced lethality was decreased by GPXl knockout, but enhanced by GPXl overexpression. Undoubtedly, GPXl is antioxidative in vivo and is important for the body to defend against severe, acute oxidative stress. Because GPX1(-/-) mice died of high doses of paraquat or diquat acutely, without showing tissue lesions seen in Se-adequate wild-type mice that survived much longer [41-43], we conducted time course experiments to determine redox status (ratios of NADPH/NADP and NADH/NAD [45]),

New roles of glutathione peroxidase-1

111

lipid peroxidation (F2-isoprostanes [46]), and protein oxidation (protein carbonyl [47]). Decreases in hepatic NADPH/NADP and NADH/NAD ratios occurred in all groups of mice at 1 hour post paraquat injection, and the drop was much sharper in GPXl-deficient mice than the GPXl-adequate mice [48]. Tissue F2-isoprostanes and protein carbonyl contents were also sharply increased after the paraquat injection in these GPXl-deficient mice. An increased liver F2-isoprostanes in GPXl-deficient mice injected with lethal doses of diquat preceded the plasma ALT activity rise, an indicator of liver injury [42]. Obviously, the GPXl protection against mouse lethality induced by high levels of paraquat or diquat was associated with the attenuated protein oxidation, lipid peroxidation, and redox shift. Likely, GPXl is important to impede the redox shift toward oxidation, driven by overproduction of ROS under severe oxidative stress, so that the NADPHdependent metabolic systems are not largely disturbed. Role of GPXl in mild oxidative stress Deprivation of Se potentiates paraquat toxicity in several species [11,12,32]. In rats, Se deficiency shifts the LD50 of paraquat from 30 to 10 mg/kg body weight and the target organ from lung to liver [33]. To determine the role of GPXl in coping with mild oxidative stress in Se deficiency vs. Se adequacy, we injected Se-deficient and Se-adequate mice with 12.5 mg of paraquat/kg body weight (Figure 1, [49]). Lrespective of the genotypes, all Se-adequate mice survived fi^om this insult and showed no rise in plasma ALT activity, whereas 90% of the Se-deficient mice died within 10 hours after the injection with approximately 1,000-fold increase in plasma ALT activity over the baseline. It was striking that an ip injection of Se (50 (Ag/kg body weight, as Na2Se03), at 6 hours before the paraquat injection, had little effect on the responses of Se-deficient GPX1(-/-) mice to the pro-oxidant insult, but reduced the mortality rate from 90 to 50% (P < 0.05) and plasma ALT activity from 24,000 to 8,300 U/L. Hepatic aponecrosis, the combined apoptosis and necrosis, was attenuated by the Se injection in the Se-deficient WT mice, but not in the Se-deficient GPX1(-/-) mice. All these key differences between the two genotypes were associated with only a 4% increase in tissue GPXl activity in the Se-deficient wild-type mice by the Se injection [49]. A time-course study indicated that this minute amount of GPXl activity repletion in the Se-deficient wild-type mice delayed the appearance and decreased the severity of the paraquat-mediated hepatic aponecrosis, compared with that in the Se-deficient GPX1(-/-) mice. Consistently, the former mice had lower levels of hepatic phosphor-c-Jun Nterminal kinase (phospho-JNK), p53, and phosphor-p53 than the latter. In contrast, the paraquat-mediated gene and/or protein expression of proapoptotic Bax, Bcl-w, and Bcl-Xs, cell survival/death factors GADD45, MDM2, c-Myc, and caspase-3 were up- regulated, but that of antiapoptotic

Selenium: Its molecular biology and role in human health

178

Se (50 i-ig/ka)

T

.ft.

*

\

ime (li)

Se-deficieiit

-6 61i Paraquat (12.5 mg/kg) 0 G;-;;

WT

4l

6 10

Assays Biochemical analyses Iimuiiiioliistopatliclogy Western analyses of p53- JKK. and p38 Iiuiiiuiiiocomplex kinase assay Microarray of apoprosis gene expression Figure 1. Model for assessing GPXl function in moderate oxidative stress. The small boxes represent injections of Se or paraquat. GPXl"'", GPXl knockout; WT, wild-type.

Bcl-2 was down-regulated in the GPX1(-/-) mice compared with wild-type mice. Interestingly, we conducted kinase assays using anti-JNK immunoprecipitates, and found that phosphor-JNK catalyzed phosphorylation of endogenous or purified p53 on Ser-15, and the event was promoted by the minute amount of GPXl activity repletion in the Sedeficient wild-type mice [50]. A previous study has shovwi that JNKl

New roles of glutathione peroxidase-1

179

depletion by antisense technology abolished Ap-induced p53 phosphorylation on Ser-15 [51]. Our findings suggest a modulating role of GPXl on the JNK-dependent p53 phosphorylation on Ser-15 under mild oxidative stress. Functional interactions between GPXl and vitamin E The nutritional essentiality of Se was initially recognized by its vitamin Esparing role in the prevention of liver injuries, and symptoms of Se deficiency are generally confounded with those of vitamin E deficiency [1]. Vitamin E is known as an antioxidant that quenches free radicals in biological membranes. While vitamin E was shown to inhibit paraquatinduced cell death and lipid peroxidation in cultured rat hepatocytes [52,53], it played only limited roles in diquat cytotoxicity [54,55]. Moreover, supplementing vitamin E did not alleviate acute oral paraquat lethality in chicks [56]. To determine whether high levels of dietary vitamin E replaced the protection of GPXl against paraquat-induced oxidative stress in mice, we challenged Se-adequate GPX1(-/-) and wild-type mice with an ip injection of paraquat (50 mg/kg body weight) after feeding these mice with various levels of dietary vitamin E (up to 100-fold of daily needs) [57]. Although high levels of dietary vitamin E attenuated the paraquat-mediated hepatic lipid peroxidation, mouse survival time or rate was affected by only the GPXl knockout, but not dietary vitamin E levels. Clearly, the protection conferred by GPXl against this lethal oxidative stress can not be replaced by high levels of dietary vitamin E. Contrasting roles of GPXl in coping with ROS vs. RNS Past research on the antioxidant roles of GPXl or any other enzymes has heavily skewed to ROS. However, reactive nitrogen species (RNS) are constantly generatedfirommetabolism such as the formation of peroxynitrite from superoxide anion and NO. Using purified bovine GPXl protein and a cell-free system [58], Sies et al. suggested that GPXl was a peroxynitrite reductase. To test the biological relevance of their view, we isolated hepatocytes from the GPX1(-/-) and wild-tj^je mice and treated them with diquat (a superoxide generator), S-nitroso-iV-acetyl-penicillamine (SNAP, a nitric oxide donor), 3-morpholinosydnonimine (SIN-1, a peroxynitrite generator), and peroxynitrite (a potent RNS) [59,60]. We measured DNA strand breaks, cytochrome c release, and caspase-3 activation as indicators of apoptosis. It was very striking that the diquat-induced apoptosis was significantly greater in the GPX1(-/-) than in wild-t5rpe cells, whereas the complete opposite was true for the peroxynitrite-induced apoptosis. The GPX1(-/-) cells were not more susceptible to the treatments of SNAP or SIN1 with diquat than wild-type cells [59,60]. Instead, there was less protein nitrotyrosine formation in the GPX1(-/-) cells treated with SNAP and diquat

180

Selenium: Its molecular biology and role in human health

than that in wild-type cells [60]. Although stimulated macrophages isolated from GPX1(-/-) mice produced more NO than those from the wild-type mice, and GPXl protected NO-associated protein carbonyl formation in these cells [61], our results do not support the notion that GPXl was a peroxynitrite reductase [58]. In fact, our data suggest that GPXl may potentiate RNSrelated oxidative sfress. This is completely opposite to its protection against ROS-related oxidative stress. Our view of the contrasting roles of GPXl in ROS vs. RNS-related oxidative stress is supported by a number of animal experiments [62-66]. It has been shown that GPXl protects against ischaemia/reperfusion injury [6264], virus-induced myocarditis [65], and endotoxemia [66]. Increased endotoxemia in GPX1(-/-) mice was in association with increases in the expression of genes whose products regulate levels of ROS [67]. Furthermore, GPX1(-/-) mice are more susceptible than wild-type mice to pro-oxidant-induced neurotoxicity [68,69], despite variations from studies with neurons isolated from GPX1(-/-) mice [70,71]. In contrast, toxicities induced by RNS-generated drugs were prevented or attenuated by GPXl null, but aggravated by GPXl overexpression. Mice overexpressing GPXl became more susceptible to acetaminophen-induced lethality and hepatic GSH depletion than wild-type mice [72]. Meanwhile, GPX1(-/-) mice were more resistant to acetaminophen-induced plasma ALT activity increases [73] and to kainic acid-induced mortality and seizures than wild-type mice [74]. It seems to be overly simplifying to call GPXl an antioxidant enzyme as its role in coping with any given stress may depend upon the nature of oxidants. Impacts of GPXl overexpression on insulin function It has been reported that Se exhibits insulin-mimetic property, and that there is a linkage between oxidative stress and diabetes [75]. Surprisingly, we found that GPX1(+) mice developed insulin resistance and obesity, along with hyperglycemia, hyperinsulinemia, and elevated plasma leptin [25]. After insulin stimulation, GPX1(+) mice exhibited attenuated phosphorylations of insulin receptor and Akt on Ser^°* and Ser'*^^, compared with wild-type controls. It is likely that overexpression of GPXl overquenched intracellular ROS, resulting in an accelerated dephosphorylation of proteins in the insulin cascade. A recent human study has shown a positive association between increases in erythrocyte GPXl activity and incidences of insulin resistance during pregnancy [76]. Although knockout of GPXl did not affect renal damage associated with type 1 diabetic nephropathy [77], our results imply a possible involvement of GPXl in type 2 diabetes. Perspectives The successful generation of the GPX1(-/-) and GPX1(+) mice have helped us in elucidating our understanding of GPXl regulation and fimction. From

New roles of glutathione peroxidase-1

181

the initial verification of m vivo antioxidant roles of GPXl to the more recent illustration of GPXl overexpression on insulin function, we have begun to get glimpses of the complex networks of GPXl actions. Unequivocally, GPXl is a bona fide antioxidant in vivo, and a regulator of ROS and RNS metabolism. It is important to recognize that metabolic functions of GPXl in oxidative stress are not unilateral. It will be fascinating to find out: 1) how GPXl exerts its differential roles in coping with ROS and RNS; 2) how GPXl modulates the nature of necrotic and apoptotic cell deaths, and 3) how GPXl overexpression intervenes in insulin signaling and function. Answers to these critical questions will definitely enrich our knowledge of GPXl biology and help in developing diagnoses and therapies related to oxidative injuries. Acknowledgements Research was supported in part by NIH grant DK53018 to X.G.L. References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16.

K Schwarz, CM Foltz 1957 J Am Biol Soc 79:3292 GC Mills 1957 J Biol Chem 229:189 L Flohe, WA Gunzler, HH Schock 1973 FEES Lett 32:132 JT Rotruck, et al. 1973 Science 179:588 D Behne, W Wolters 1983 JNutr 113:456 S Chada, MM Le Beau, L Casey, PE Newburger 1990 Genomics 6:268 I Chambers, et al. 1986 EMBO J5:\22\ Q Shen, FF Chu, PE Newburger 1993 J Biol Chem 268:11463 MJ Berry, et al. 1991 Nature 353:273 MS Saedi, et al. 1988 Biochem Biophys Res Commm 153:855 SD Mercurio, GF Combs Jr 1986 J Nutr 116:1726 SD Mercurio, GF Combs Jr 1986 Biochem Pharmacol 35:4505 YS Ho, et al. 1997 J Biol Chem 272:16644 JB de Haan, et al. 1998 J Biol Chem 273:22528 LA Esposito, et al. 2000 Free Radic Biol Med 28:754 O Mirochnitchenko, U Palnitkar, M Philbert, M Inouye 1995 Proc Natl Acad Sci USA 92:8120 17. WH Cheng, et al. 1997 JNutr 127:675 18. XG Lei, JK Evenson, KM Thompson, RA Sunde 1995 JNutr 125:1438 19. G Bermano, et a/. 1996 Biol Trace Elem Res 51:211 20. RF Burk 1991 FASEB J 5:2274 21. RA Sunde 1994 Selenium in Biology and Human Health Springer-Verlag, New York pp 45 22. WH Cheng, GF Combs Jr, XG Lei 1998 Biofactors 7:311 23. WH Cheng, et al. 1997 JNutr 127:1445 24. GV Kryukov, et al 2003 Science 300:1439 25. JP McClung, et al. 2004 Proc Natl Acad Sci USA 101:8852 26. VN Reddy, et al. 2001 Invest Ophthalmol Vis Sci 42:3247 27. A Spector, W Ma, RR Wang, Y Yang, YS Ho 1997 Exp Eye Res 64:477 28. ME Mirault, A Tremblay, N Beaudoin, M Tremblay 1991 J Biol Chem 266:20752 29. DM Hockenbery, ZN Oltvai, XM Yin, CL Milliman, SJ Korsmeyer 1993 Cell 75:241

182 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77.

Selenium: Its molecular biology and role in human health MJ Kelner, RD Bagnell, SF Uglik, MA Montoya, GT MuUenbach 1995 Arch Biochem Biophys m-AO SD Taylor, LD Davenport, MJ Speranza, GT MuUenbach, RE Lynch 1993 Arch Biochem Biophys 305:600 RF Burk, RA Lawrence, JM Lane 1980/ Clin Invest 65:1024 SZ Cagen, JE Gibson 1977 Toxicol Appl Pharmacol 40:193 LL Smith 1987 Hum Toxicol 6:31 JA Farrington, M Ebert, EJ Land, K Fletcher 1973 Biochim Biophys Acta 314:372 RF Burk, et al. 1995 //epato/ogv 21:561 JB Atkinson, KE Hill, RF Burk 2001 Lab Invest 81:193 Y Saito, et al 1999 J Biol Chem 274:2866 L Schomburg, et al 2003 Biochem J 2,10:2,91 KE Hill, et al 2003 J Biol Chem 278:13640 WH Cheng, et al.i 199S JNutr 128:1070 Y Fu, WH Cheng, JM Porres, DA Ross, XG Lei 1999 Free Radic Biol Med 27:605 Y Fu, WH Cheng, DA Ross, X Lei 1999 Proc Soc Exp Biol Med 222:164 RH Van, et al. 2004 Free Radic Biol Med 36:1625 H Witschi, S Kacew, KI Hirai, MG Cote 1977 Chem Biol Interact 19:143 JD Morrow, et al 1990 Proc Natl Acad Sci USA 87:9383 RL Levine, JA Williams, ER Stadtman, E Shacter 1994 Methods Enzymol 233:346 WH Cheng, YX Fu, JM Porres, DA Ross, XG Lei 1999 FASEB y 13:1467 WH Cheng, FW Quimby, XG Lei 2003 Free Radic Biol Med 34:918 WH Cheng, X Zheng, FR Quimby, CA Roneker, XG Lei 2003 Biochem J 370:927 MP Fogarty, EJ Downer, V Campbell 2003 Biochem J31\:1S9 MG Traber, H Sies 1996 Annu Rev Nutr 16:321 N Watanabe, Y Shiki, N Morisaki, Y Saito, S Yoshida 1986 Biochim Biophys Acta 883:420 MS Sandy, MD Di, MT Smith 1988 Toxicol Appl Pharmacol 93:288 L Eklow-Lastbom, L Rossi, H Thor, S Orrenius 1986 Free Radic Res Commun 2:57 GF Combs Jr, FJ Peterson 1983 JNutr 113:538 WH Cheng, BA Valentine, XG Lei 1999 JNutr 129:1951 H Sies, VS Sharov, LO Klotz, K Briviba 1997 J Biol Chem 272:27812 Y Fu, H Sies, XG Lei 2001 J Biol Chem 276:43004 Y Fu, JM Porres, XG Lei 2001 Biochem J 359:687 Y Fu, CC McCormick,C Roneker, XG Lei 2001 Free Radic Biol Med 3\:A50 N Maulik, T Yoshida, DK Das 1999 Mol Cell Biochem 196:13 T Yoshida, et al 1997 Circulation 96:11 PJ Crack, et al 2001 JNeurochem 78:1389 MA Beck, RS Esworthy, YS Ho, FF Chu 1998 FASEB J \2:\\43 H Jaeschke, YS Ho, MA Fisher, JA Lawson, A Farhood 1999 Hepatology 29:443 C Li, J Liu, MP Waalkes, H Jaeschke 2003 Toxicol Lett 144:397 P Klivenyi, et al 2000 J Neurosci 20:1 J Zhang, DG Graham, TJ Montine, YS Ho 2000 JNeuropathol Exp Neurol 59:53 K Nakamura, et al.200O JNeurochem 74:2305 JM Taylor, U Ali, RC lannello, P Hertzog, PJ Crack 2005 JNeurochem 92:283 O Mirochnitchenko, et al 1999 J Biol Chem 274:10349 TR Knight, A Kurtz, ML Bajt, JA Hinson, H Jaeschke 2001 Toxicol Sci 62:212 D Jiang, G Akopian, YS Ho, JP Walsh, JK Andersen 2000 Exp Neurol 164:257 OEzakil990y5jo/aeOT265:1124 X Chen, TO SchoU, MJ Leskiw, MR Donaldson, TP Stein 2003 J Clin Endocrinol Metab 88:5963 JB de Haan, et al 2005 Am J Physiol Renal Physiol 289:F544

Chapter 17. Selenoproteins of the thioredoxin system Ame Holmgren Medical Nobel Institute for Biochemistry, Department of Medical Biochemistry and Biophysics, Karolinska Institute, SE-171 77 Stockholm, Sweden

Summary: The three isoenzymes of mammalian thioredoxin reductase are dimeric selenoproteins containing an essential catalytically active selenocysteine (Sec) residue. In contrast to the enzymes from bacteria, yeast and plants, the mammalian enzymes are larger and entirely different in structure and mechanism. They are homologous to glutathione reductase, but with a C-terminal elongation of 16 residues containing the conserved Cterminal active site sequence -Gly-Cys-Sec-Gly. The active site is a selenneylsulfide formed from the conserved Cys-Sec sequence, which is reduced to a selenolthiol by electrons from the redox active disulfide of the other subunit, as revealed by a three-dimensional structure of the rat enzyme. The essential role of Sec in thioredoxin reductase explains the very broad substrate specificity including reduction of thioredoxin, selenite, dehydroascorbic acid and ascorbyl free radical, hydrogen peroxide and lipid hydroperoxides. The essential role of selenium in human thioredoxin reductases further explains roles of this trace element in cell growth via pleiofropic effects in reduction of thioredoxin with its multiple roles in elecfron transport to essential biosynthetic enzymes, thiol redox confrol of transcription factors, or in defense against oxidative sfress. Clinically used inhibitors of cell growth or inflammation like gold thioglucose are targeted to the reduced Sec residue of the enzyme. Introduction The thioredoxin system comprised of NADPH, thioredoxin (Trx) and the flavoprotein thioredoxin reductase (TrxR) is ubiquitously present from Archaea to man [1, 2]. Thioredoxin with a redox-active dithiol/disulfide is an electron donor for essential enzymes such as ribonucleotide reductase and a general protein disulfide reductase with numerous functions in confrol of infracellular redox potential, defense against oxidative sfress and signal fransduction by thiol redox control [2]. Thioredoxin reductases from mammalian cells and higher eukaryotes are selenoenzymes [3,4] and very different from the smaller selenium-independent enzymes of Archaea,

184

Selenium: Its molecular biology and role in human health

bacteria, yeast and plants [5]. This chapter will discuss reactions between selenium compounds and the thioredoxin system and some of the structurefunction relationships of mammalian thioredoxin reductases. General properties of thioredoxin systems All thioredoxin reductases reduce oxidized thioredoxin (Trx-S2) at the expense of NADPH [1,2] (Reaction 1). Reduced thioredoxin [Trx-(SH)2] is reoxidized by disulfides in proteins generating thiols (Reaction 2): TrxR Trx-(SH)2 + NADP^

Trx-S2 + NADPH + H^

Trx-(SH)2 + Protein-S2

spontaneous Trx-S2 + Protein-(SH)2

(1)

(2)

Generally, the Km-value for NADPH is low or in the range below 10 nM and that of Trx-S2 is typically from 1 to 3 (xM. Isolation and characterization of mammalian thioredoxin and thioredoxin reductase started about 30 years ago [6-8]. As shown in Table 1, there are some major differences between the thioredoxin systems of prokaryotes like E. coli and that of mammalian organisms.

Table 1. Properties of Thioredoxin Systems E. coli

Human

Thioredoxin

U,= 12,000 108 aa -CGPC-active site Trx-S2 stable upon storage

Mr= 12,000 104 aa -CGPC-active site +3 structural SH-groups, Trx-activity reversibly lost by additional disulfide formation upon aerobic storage

Tliioredoxin reductase

Mr= 70,000 2 subunits High specificity Stable

Mr= 114,000 or larger; 3 genes 2 subunits Broad specificity, selenoenzymes Labile to oxidation - reduction cycles

E. coli and mammalian cytosolic thioredoxins are homologous proteins with a conserved -Cys-Gly-Pro-Cys- active site. However, mammalian thioredoxin must be purified in the fully reduced form since they contain structural SH-groups which form additional disulfides upon oxidation. This may have autoregulatory function of thioredoxin activity in resting cells or

Selenoproteins of the thioredoxin system

185

upon oxidative stress, yet incompletely known in vivo. Thioredoxin reductases from mammalian cells have very different properties when compared with the enzymes from E. coli, yeast or plants (reviewed in [5]). The cytosolic en2yme has subunits with 55 kDa or larger instead of the 35 kDa in the E. coli enzyme with a known three-dimensional structure [5]. As will be described below, the mammalian enzyme also has a very broad substrate specificity entirely different from the generally subsfrate-specific enzymes only reducing Trx-S2 that are present in prokaryotes, yeast and plant cytosol. Selenium reduction by the thioredoxin system The fact that administration of selenium compounds like selenite (SeOs^) cause the inhibition of tumor cell proliferation in vivo and the knowledge that thioredoxin reductase appeared to be more highly expressed in malignant cells prompted us to start investigations on the reactions of selenium compounds with the mammalian thioredoxin system. Contrary to expectations, we discovered that selenite is a direct substrate for thioredoxin reductase as well as an efficient oxidant of thioredoxin [9,10]. With 200 ^M NADPH and 50 nM calf thymus thioredoxin reductase, addition of 10 ^M selenite caused oxidation of 40 jxM NADPH in 12 min and 100 jxM NADPH after 30 min demonstrating a direct reduction of selenite with redox cycling by oxygen [9,10]. This was demonstrated by incubation under anaerobic conditions where only 3 mol of NADPH was oxidized per mol of selenite according to Reaction 3:

SeOa'+ 3 NADPH + 3H^

TrxR -^ Se^'+ 3 NADP^ + 3 HjO

(3)

Addition of thioredoxin stimulated the reaction further since selenite rapidly reacts with Trx-(SH)2 to oxidize it to Trx-S2 [11-13]. Since glutathione reductase will not react with selenite [13], Reaction 3 should provide cells with selenide, a required precursor for selenophosphate and selenocysteine (Sec) synthesis [14]. Selenite and glutathione react to form selenodiglutathione (GS-Se-SG) which has been suggested to be a major metabolite of inorganic selenium salts in mammalian tissues [15]. Reduction of selenodiglutathione by NADPH and glutathione reductase was demonstrated by Ganther [16] and it has been proposed to be a source of selenide in cells as well as an inhibitor of neoplastic growth [17]. We synthesized GS-Se-SG [11,18] and discovered that this compoimd is a direct efficient substrate for mammalian thioredoxin reductase and a highly

186

Selenium: Its molecular biology and role in human health

efficient oxidant of reduced thioredoxin. Since GSSG is not a substrate for mammalian thioredoxin reductase [7,8] the insertion of the selenium atom in the GSSG molecule to form GS-Se-SG makes this molecule highly reactive with the enzyme. Reduction of GS-Se-SG to yield selenide by glutathione reductase requires two mol of NADPH. We found only the first stoichiometric reduction to be fast with GS-Se" as a product [11]. The second reaction was slow and relatively inefficient. These results strongly suggest that the major selenide generation in cells is via thioredoxin reductase and thioredoxin. Thus, in mammalian cells the selenoenzyme thioredoxin reductase is also responsible for the generation of selenide required for its own synthesis. An oxygen dependent and non-stoichiometric consumption of NADPH is given by the thioredoxin system in the presence of selenite, selenodiglutathione and selenocystine [9-11,18]. The latter oxidized form of Sec is an efficient substrate for mammalian thioredoxin reductase with a Km of 6 \iM [18]. The mechanism may be that the XSe" reacts with a dithiol (or selenolthiol) to catalyze oxidation according to Reaction 4:

XSe'+ R-(SH)2 + (0)

-^

XSe+R-Sz + HaO

(4)

The effect will be 02-dependent consumption of NADPH and the results demonstrating autoxidation of the selenium compounds provide an explanation for the lack of a free pool of Sec as well as the acute toxic effects of selenium compounds on cells, e.g., leading to oxidative stress and apoptosis. Substrate specificity of thioredoxin reductase Mammalian thioredoxin reductases display a surprisingly wide substrate specificity as first observed during purification [7,8]. This is in contrast to the smaller prokaryotic thioredoxin reductases, which do not use mammalian thioredoxins as substrates despite their conserved active sites and closely related three-dimensional structures [19]. As summarized in Table 2, a truly wide range of direct reductions are catalyzed by mammalian cytosolic thioredoxin reductases. Thioredoxin from E. coli is a substrate with a similar Kcat, but with a 15-fold higher Km-value (35 \xM) compared with the rat liver protein [8]. Mammalian cytosolic thioredoxins generally show full crossreactivity with thioredoxin reductases fi-om different mammalian sources and vice versa, hi many instances, fi-ee selenocyst(e)ine will stimulate reduction of substrates [24,13,23,29].

Selenoproteins of the thioredoxin system

187

Table 2. Reactions catalyzed by cytosolic mammalian thioredoxin reductases (rat, bovine and human) 5,5-dithiobis-(2-mtrobenzoic acid) reduction Thioredoxin-Sa reduction Protein disulfide isomerase (PDI) Selenite (SeOj^") and Sec reduction Selenodiglutathione reduction Nitrosoglutathione (GSNO) reduction Electron donor to plasma glutathione peroxidase H2O2 and lipid hydroperoxide reductase Reduction of alloxan and vitamin K NK-lysin disulfide reduction and inactivation of cytotoxic activity Lipoic acid and lipoamide reduction Reduction of dehydroascorbic acid Reduction of ascorbylfl-eeradical

[7] [8] [20] [ 10,13] [11] [21 ] [22] [23,24] [8,25] [26] [27] [28] [29]

Structure and mechanism of mammalian thioredoxin reductase Recent biochemical studies, sequencing and cloning of mammalian thioredoxin reductases have revealed that the enzymes are selenoproteins and entirely different from the corresponding enzymes in bacteria, yeast and plants (review in [5]). Stadtman and coworkers serendipitously discovered that human tumor cell thioredoxin reductase is a selenoprotein using labeling of selenoproteins with radioactive selenite [3]. This also explained [30] why a previously putative clone of the human enzyme [31] where the TGA codon for Sec was interpreted as the stop codon (Figure 1) gave no enzyme activity. The TGA will be acting as a stop codon in E. coli due to the species-specific machinery for synthesis of selenoproteins which is different in bacteria and mammalian cells [14]. By sequencing large parts of the cytosolic bovine enzyme, we also directly identified the C-terminal peptide as containing Sec. The bovine peptides were used to identify a rat cDNA clone which was sequenced [4]. The results showed a polypeptide chain with a high homology to glutathione reductase [4,23] including an identical active site disulfide (CVNVGC) (Figure 1) but with a 16-residue elongation containing the conserved Cterminal sequence, Gly-Cys-Sec-Gly. A Sec insertion sequence (SECIS) was identified in the 3'-untranslated region [4]. Furthermore, digestion of thioredoxin reductase by carboxypeptidase after reduction by NADPH released Sec with loss of activity; the oxidized form of the enzyme was resistant to carboxypeptidase digestion [4]. Redox titrations with dithionite and NADPH demonstrated that the mechanism of the human placenta enzyme is similar to that of lipoamide dehydrogenase and glutathione

188

Selenium: Its molecular biology and role in human health

reductase and distinct from the mechanism of thioredoxin reductase from E. coli [32]. The results also demonstrated that the Sec residue of human thioredoxin reductase is redox active and communicates with the redox active disulfide since more than 4 elecfrons per subunit are required to completely reduce the FAD of the oxidized enzyme. Furthermore, the Sec residue is alkylated with loss of activity only after reduction by NADPH [4,33,34]. The Sec residue is also the target of the irreversible inhibitor 1chloro-2,4-dinifrobenzene only after reduction by NADPH [35] as shown by peptide analysis [34].

CVNVGC

GCUG

r^v^: H jN -I

V FAD

NADPH

Intertaoe

|- COOH

a n-Al a - a y-Cys-Sec-Q y-Ter (human TrxR) CAG GOT GGC TGC TQA GGT TAA GCC CCA . . . CAG TOT GGC TGC TGA GGT TAA GCC CCA . . . a n - S e r - Q y-Cys-Sec-Q y-Ter ( r a t TrxR) Figure 1. Schematic structure (upper small, rectangular box) and C-terminal sequences (lower large, rectangular box) of human and rat thioredoxin reductases [4,30,31]. The N-terminal glutathione reductase-like active site disulfide (CVNVGC) is shown as well as the FAD, NADPH and interface domains. The active site is shown in the C-terminal region with GCUG denoting Gly-Cys-Sec-Gly. The lower part of the figure also shows the TGA codon encoding Sec.

The essential role of selenium in the catalytic activities of mammalian thioredoxin reductase was revealed by characterization of recombinant enzymes with Sec mutations [23]. This was done by removing the Sec insertion sequence in the rat gene and changing the Sec498 encoded by TGA to Cys or Ser codons by mutagenesis. The truncated protein having the Cterminal dipeptide deleted, expected in selenium deficiency, was also engineered. All three mutants were successfully overexpressed in E. coli and purified to homogeneity with 1 mol of FAD per monomeric subunit. All three mutant proteins rapidly generated the A540 absorbance resulting from the thiolate-flavin charge transfer complex characteristic of mammalian TrxR. Only the Sec498 Cys enzyme showed catalytic activity in reduction of thioredoxin, with a 100-fold lower Kcat and a 10-fold lower K^ compared to

189

Selenoproteins of the thioredoxin system

the wild type rat enzyme. The pH-optimum of the Sec-containing wild type enzyme was 7 whereas the Sec498 to Cys mutant showed a pH optimum of 9. This strongly suggested the involvement of the low pKa Sec selenol in the enzyme mechanism. Also selenium was required for hydrogen peroxide reductase activity [23]. Thus, selenium is required for the catalytic activities of thioredoxin reductase explaining the essential role of this trace element in cell growth. Based on the homology to glutathione reductase we proposed a model of mammalian thioredoxin reductase (Figure 2).

NADPH

domain

Interface

16 aa with

elongation Cys-SeCys

I'AD

i'AD

domain

domain 16 aa elongation with Cys-SeCys

Thioredoxin

dotnilin

NADPH

domain

Reductase

Figure 2. Structural model of mammalian thioredoxin reductase based on the homology to glutathione reductase. The 16-residue C-terminal extension with the active site is shown as well as the head to tail arrangement of the subunits in the dimer. Taken from [36]. The FAD, NADPH and interface domains are shown (see also Figure 1).

The enzyme is a head to tail dimer with the 16-residue elongation in principle taking the place of GSSG in glutathione reductase. The active site of the enzyme is a selenolthiol in its reduced form and a selenenylsulfide formed from the conserved cysteine-Sec sequence in the oxidized form [36]. The selenenylsulfide was isolated by peptide sequencing and also confirmed by mass spectrometry [36]. Mechanisms of the enzyme have also been postulated involving a reductive half-reaction similar to that of glutathione reductase leading to reduction of the active site disulfide (Figures 1 and 2). Electrons are thereafter transferred from the redox-active dithiols to the selenenylsulfide of the other subunit generating the selenolthiol. Characterization of the Cys mutant enzyme revealed that the selenium atom

190

Selenium: Its molecular biology and role in human health

with its larger radius is critical for the formation of the unique selenenylsulfide [36] since the C-terminal dithiol stays reduced in the Cys mutant [36]. Similar results confirming these data have also been obtained by others [37]. The structure of the enzyme has been solved by X-ray crystallography after the Cys mutant enzyme has been crystallized [38]. Crystal structure of the Sec498Cys mutant of rat TrxRl in complex with NADPH was determined to 3.0-A resolution. The overall structure is similar to that of glutathione reductase, including the conserved amino acid residues that bind the cofactors FAD and NADPH. The redox active disulfide in the N-terminal portion of the enzyme is identical to that of glutathione reductase. Residues directly binding the substrate glutathione disulfide in glutathione reductase are conserved despite the fact that glutathione disulfide is not a substrate for thioredoxin reductase [38]. The 16-residue Cterminal tail, a unique feature on mammalian thioredoxin reductases, folds in such a way that it can approach the active site disulfide of the other subunit in the dimer (see schematic drawing in Figure 2). A unique feature of the Sec498Cys mutant of rat TrxRl is that the thiols do not form a disulfide [38]. A model of the complex of TrxR read Trx allows docking of oxidized thioredoxin to the structure without large conformational changes [38]. This is in great contrast to the large conformational change required for the prokaryotic Cys-residue enzymes [5]. The model suggests specific interactions between residues in thioredoxin (D60, D61 and K72 and corresponding charges in TrxR forming electrostatic interactions). The Xray structure particularly explains how the 16-residue C-terminal extension conserved in all three mammalian isoenzymes of thioredoxin reductase. It extends the electron transport chain from the catalytic disulfide to the enzjone surface, enabling reaction with Trx and a range of other substrates (Table 2). It acts to prevent the enzyme from acting as a glutathione reductase by blocking access to the redox active disulfide. The results of the X-ray study [38] strongly suggest that mammalian thioredoxin reductase evolved from a glutathione reductase scaffold rather than from its prokaryotic coimterpart. Such an evolutionary switch will render cell growth dependent on selenium in the form of Sec and it may have advantages for cells using reactive oxygen species like hydrogen peroxide in cell signaling. Isoenzymes of thioredoxin reductase Apart from the cytosolic thioredoxin reductase, TrxRl, two additional genes encoding novel forms of human and mouse selenoprotein thioredoxin reductases have been identified [39]. One is a mitochondrial enzyme [40,41] and the other thioredoxin-glutathione reductase carrying an N-terminal glutaredoxin domain; the latter is preferentially expressed in testis. All these enzymes have extensions in the N-terminal region but share the C-terminal

Selenoproteins of the thioredoxin system

191

active site sequence. Additional complexity is given by the identification of enzymes with mRNA variants differing in the 5'-untranslated region [42] and by 5'-exon splicing [43]. The nematode C. elegans contains two homologues related to mammalian thioredoxin reductase, one with Cys and the other with Sec. The Sec containing enzyme with 74 kDa subunits is the major selenoprotein in C. elegans. Medical aspects of selenium in thioredoxin reductase Human thioredoxin reductase is a general reducing enzyme with a wide substrate specificity contributing to cellular redox homeostasis and is a major pathophysiological factor and drug target. Together with thioredoxin, it is involved in prevention, intervention and repair of damage caused by hydrogen peroxide-based oxidative stress. As a selenite reducing enzyme with a selenol containing active site human thioredoxin reductase plays a central role in selenium physiology. A range of human diseases and conditions are now known or suspected to be related to the activity and function of thioredoxin reductase (in-depth review in [45]). This involves diseases like reumatoid arthritis, Sjogren's syndrome, AIDS and malignancies. The close homology between human thioredoxin reductase and glutathione reductase has lead to the realization that several clinically used drugs like nitrosurea derivatives are targeted to thioredoxin reductase [45]. Furthermore, studies on the regulation of thioredoxin reductase mRNA [46] and the development of specific inhibitors for use in antitumour therapy [47-49] make the enzyme a major target for drug development. In this context it will be important in future studies to establish also the role of the glutathione-dependent glutaredoxin system [50,51] which is an alternative non-selenium pathway of transferring electrons to essential biosynthetic reactions like ribonucleotide reductase. Thus, determining if a malignant cell is dependent on the thioredoxin system or the glutaredoxin system should be essential in drug selection in tumor therapy. The fact that the thioredoxin system is ubiquitous and present in quite highly variant forms in pathogenic bacteria makes the enzyme a particularly attractive drug target. There is a surprising diversity in the structure and mechanism of the enzyme in several important pathogenic bacteria (reviewed in [45]). This may lead to the development of specific inhibitors of bacterial infections as in Lepra, parasitic diseases and malaria. Treatment of an inflammatory disease like reumatoid arthritis with drugs like gold thioglucose and aiu-anofin which are strong inhibitors of thioredoxin reductase likely occur by binding to the reduced Sec residue in the enzyme. Since thioredoxin reductase is involved in central biosynthetic reactions and defense against oxidative stress via thioredoxin, it has a high priority in the hierarchy of synthesis of selenoproteins. Of particular interest is whether

192

Selenium: Its molecular biology and role in human health

the truncated en2yme expected in selenium deficiency is present in cells. This may be of great importance for understanding the effects of selenium supplementation as an anticancer agent [52]. The modification of the Sec residue in TrxR or the truncated enzyme gives rapid induction of cell death [53]. A number of clinically used anticancer compounds, including alkylating and platinum-containing drugs, inhibit thioredoxin reductase, but not glutathione reductase [54]. Obviously, thioredoxin reductase is a novel and important molecular target for cancer therapy [55]. Thioredoxin reductase gene targeting Reactive oxygen species (ROS) are generated as by products of the respiratory chain or by NADPH oxidases. ROS are implicated in the pathogenesis and pathophysiology of a variety of human diseases, such as cardiovascular and degenerative disorders and cancer. ROS is also implicated in cellular signaling. Peroxiredoxins working together with thioredoxins and thioredoxin reductases are controling the levels of reactive oxygen species and free radicals. A complete thioredoxin system, including thioredoxin reductase, Trx and peroxiredoxin (Prx III) is present in mitochondria. To address the fimction of mitochondrial thioredoxin reductase (TrxR2), a ubiquitous Cre-mediated inactivation of TrxR2 was shown to be associated with death at embryonic day 13 [56]. TrxR2" embryos are smaller and severely anemic and showed increased apoptosis in liver [56]. Also, the size of hematopoietic colony cultures ex vivo was dramatically reduced. TrxR2-deficient embryonic fibroblasts showed high sensitivity to endogeneously produced oxygen radicals when glutathione synthesis was inhibited [56]. Also, the ventricular heart wall of the mitochondrial thioredoxin reductase knockout embryos was thinner and the proliferation of cardiomyocytes was decreased. Cardiac specific ablation of TrxR2 resulted in fetal cardiomyopathy with symptoms similar to those of Keshan disease and Friedreich's ataxia [56]. Thus, mitochondrial thioredoxin reductase plays an essential role in hematopoiesis, heart development and heart function [56]. A similar study on the cytoplasmic thioredoxin reductase (TrxRl), using a conditionally target deletion of the Txnrdl gene showed that the gene was essential for embryogenesis [57]. Ubiquitous Cre-mediated inactivation of Txnrdl leads to early embryonic lethality [57]. Embryos of the homozygous mutant displayed severe growth retardation and failed to turn [57]. Also, Txnrdl-deficient embryonic fibroblasts do not proliferate in vitro in line with growth impairment. Surprisingly, in contrast, ex v/vo-cultured embryonic Txnrdl-deficient cardiomyocytes were not effected and mice with a heart specific inactivation of Txnrdl developed normally and appeared healthy [57]. The conclusion from these studies is that the TrxRl

Selenoproteins of the thioredoxin system

193

enzyme is essential during embryogenesis in most developing tissues except for the heart. Obviously, the role of thioredoxin, thioredoxin reductase to provide electrons for the synthesis of deoxyribonucleotide by ribonucleotide reductase is the suspected reason for the essential role of TrxRl. The association of TrxRl with proliferation make this enzyme a specifically interesting drug target for cancer therapy. The combined effects of the studies on gene targeting [56,57] stress the importance of inhibiting TrxRl specifically without affecting mitochondrial TrxR2. Drugs affecting TrxR2 will also have serious cardiac side effects [56]. Selective RNA interference may in the future be used to selectively target TrxRl, since drugs are likely to effect both the cytoplasmic and mitochondrial thioredoxin reductases with their identical active sites. Acknowledgement The research support from the Swedish Medical Research Council (3529), the Swedish Cancer Society and the K A Wallenberg foundation is gratefully acknowledged. References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27.

A Holmgren 1985 Annu Rev Biochem 54:237 ES J Am6r, A Holmgren 2000 Eur J Biochem 161:6102 T Tamura, T C Stadtman 1996 Proc Natl Acad Sci USA 93:1006 L Zhong, ESJ Amer, J Ljung, F Aslund, A Holmgren 1998 J Biol Chem 273:8581 CH Williams Jr, LD Arscott, S Muller, BW Lennon, ML Ludwig, P-F Wang, DM Veine, K Becker, RH Schirmer 2000 Eur J Biochem 267:6110 NE EngstrSm, A Holmgren, A Larsson, S Soderhall 1974 J Biol Chem 249:205 A Holmgren 1977 J Biol Chem 252:4600 MLuthman, A Holmgren 1982 SiocAe/««//y 21:6628 A Holmgren, S Kumar 1989 Selenium in Biology and Medicine A Wendel (Ed) Springer-Verlag, Berlin, 47 S Kumar, M BjOmstedt, A Holmgren 1992 Eur J Biochem 207:435 M BjSmstedt, S Kumar, A Holmgren 1992 J Biol Chem 267:8030 X Ren, M Bjdmstedt, B Shen, M Ericson, A Holmgren 1993 Biochemistry 32:9701 M BjOmstedt, S Kumar, L Bjorkhem, G Spyrou, A Holmgren 1997 Biomed Environ Sci 10:271 TC Stadtman 1996 Annu Rev Biochem 65:83 HS Hsieh, HE Ganther 1975 Biochemistry 14:1632 HE Ganther 1971 Biochemistry 10:4089 RJ Shamberger 1985 Mutat Res 154:29 M BjOmstedt, S Kumar, A Holmgren 1995 Methods Enzymol 252:219 A Holmgren 1995 Structure 3:239 J Lundstrdm, A Holmgren 1990 J Biol Chem 265:9114 DNikitovic,AHolmgrenl996J5/o/C/iem271:19180 M Bjfimstedt, J Xue, W Huang, B Akesson, A Holmgren 1994 J Biol Chem 269:29382 L Zhong, A Holmgren 2000 J Biol Chem 275:18121 M Bjfimstedt, M Hamberg, S Kumar, J Xue, A Holmgren 1995 J Biol Chem 270:11761 AHolmgren,C Lyckeborg 19S0 Proc Natl Acad Sci USA 77:5149 M Andersson, A Holmgren, G Spyrou 1996 J Biol Chem 271:10116 ESJ Amer, J Nordberg, A Holmgren 1996 Biochem Biophys Res Commun 225:268

194 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56.

57.

Selenium: Its molecular biology and role in human health JM May, S Mendiratta, KE Hill, RF Burk 1997 J Biol Chem 272:22607 JM May, CE Cobb, S Mendiratta, KE Hill, RF Burk 1998 J Biol Chem 273:23039 VN Gladyshev, K-T Jeang, TC Stadtman 1996 Proc Natl Acad Sci USA 93:6146 PY Gasdaska, JR Gasdaska, S Cochran, G Powis 1995 FEBSLetters 373:5 LD Arscott, S Gromer, RH Schirmer, K Becker, CH Williams Jr 1997 Proc Natl Acad Sci USA 94:9621 SN Gorlatov, TC Stadtman 1998 Proc Natl Acad Sci USA 95:8520 J Nordberg, L Zhong, A Holmgren, ES J Am6r 1998 J Biol Chem 273:10835 ESJ Amer, M Bjfimstedt, A Holmgren 1995 J Biol Chem 270:3479 L Zhong, ESJ Am6r, A Holmgren 2000 Proc Natl Acad Sci USA 97:5854 SR Lee, S Bar-Noy, J Kwon, RL Levine, TC Stadtman, SG Rhee 2000 Proc Natl Acad Sci USA 97:2521 T Sandalova, L Zhong,Y Lindqvist, A Holmgren, G Schneider 2001 Proc Natl Acad Sci USA 98:9533 Q-A Sun, Y Woo, F Zappacosta, K-T Jeang, BJ Lee, DL Hatfield, VN Gladyshev 1999 J Biol Chem 274:24522 SR Lee, JR Kim, KS Kwon, HW Yoon, RL Levine, A Ginsburg, SG Rhee 1999 J Biol Chem 274:4722 A Miranda-Vizuete, AE Damdimopoulos, JR Pedrajas, J-A Gustafsson, G Spyrou 1999 Eur J Biochem 261:405 A-K Rundief, M Carlsten, MMJ Giacobini, ESJ Am^r 2000 Biochem J 347:661 QA Sun, F Zappacosta, VM Factor, PJ Wirth, DL Hatfield, VN Gladyshev 2001 J Biol Chem 276:3106 VN Gladyshev, M Krause, X-M Xu, KV Korotkov, GV Kryukov, Q-A Sun, BJ Lee, JC Wootton, DL Hatfield 1999 Biochem Biophys Res Commun 259:244 K Becker, S Gromer, RH Schirmer, S Muller 2000 Eur J Biochem 267:6118 DL Kirkpatrick, S Watson, M Kunkel, S Fletcher, S Ulhag, G Powis 1999 Anticancer Drug Res 5A21 JR Gasdaska, JW Harney, PY Gasdaska, G Powis, MJ Berry 1999 J Biol Chem 274:25379 MM Berggren, JF Mangin, JR Gasdaska, G Powis 1999 Biochem Pharmacol 57:187 G Powis, DL Kirkpatrick, M Angulo, A Baker 1998 Chem Biol Interact 111-112:23 A Holmgren 1999 Redox Regulation of Cell Signaling and its Clinical Application L Packer, J Yodoi (Eds) Marcel Dekker, New York, 279 A Holmgren 1989 J Biol Chem 264:13963 S Gromer, JH Gross, 2002 J Biol Chem 277:9701 K Anestal, ESJ Am6r 2003 J Biol Chem 278:15966 A-B Witte, K Anestal, E Jerremalm, H Ehrsson, ESJ Amer 2005 Free Radio Biol Med 39:696 P Nguen, RT Awwad, DDK Smart, DR Spitz, D Gius 2005 Cancer Letters in press M Conrad, C Jakupoglu, SG Moreno, S Lippl, A Banjac, M Schneider, H Beck, AK Hatzopoulos, U Just, F Sinowatz, W Schmal, KR Chien, W Wurst, GW Bomkamm, M Brielmeier 2004 Mol Cell Biol 24: 9414 C Jakupoglu, GKH Przemeck, M Schneider, SG Moreno, N Mayr, AK Hatzopoulos, M Hrab6 de Angelis, W Wurst, GW Bomkamm, M Brielmeier, M Conrad 2005 Mol Cell Biol 25:1980

Chapter 18. Mitochondrial and cytosolic thioredoxin reductase knockout mice Marcus Conrad and Georg W. Bomkamm Institute of Clinical Molecular Biology and Tumor Genetics, GSF-Research Centre for Environment and Health, 81377 Munich, Germany

Markus Brielmeier Department of Comparative Medicine, GSF-Research Centre for Environment and Health, 85764 Neuherberg, Germany

Summary: To address the role of the thioredoxin system in redox regulation of apoptosis and proliferation, mice with targeted deletions of both the cytosolic (Txnrdl) and the mitochondrial (Txnrd2) thioredoxin reductases were generated. These two selenoproteins are key enzymes governing the activities of cytosolic and mitochondrial thioredoxins, respectively, which are, in turn, implicated in a variety of cellular functions, such as cell-cell communication, proliferation and apoptosis. Ubiquitous and heart-specific inactivation revealed widely non-redundant functions of Txnrdl and Txnrd2. A significant drop in cell proliferation rates throughout the embryo (except in the heart), but not increased apoptosis, was the underlying cause of embryonic death of Txnrdl knockout embryos at E10.5. Perturbed cardiac development and increased apoptosis of fetal blood cells in the liver caused severe anemia, growth retardation and embryonic death (E13.5) in Txnrd2 knockout embryos. Cardiac-tissue restricted inactivation of Txnrd2 led to biventricular dilated cardiomyopathy and postnatal death; in contrast heartspecific inactivation of Txnrdl had no apparent effect on the viability of the knockout mice. In conclusion, Txnrdl contributes to cell proliferation, whereas Txnrd2 is rather involved in apoptosis regulation. Introduction Three distinct thioredoxin reductases are known in mammals, each encoded by individual genes. Thioredoxin reductase 1 (Txnrdl) is primarily localized in the cytosol [1,2], thioredoxin reductase 2 (Txnrd2) in mitochondria (Txnrd2) [3],and thioredoxin reductase 3 (Txnrd3), also called thioredoxinglutathione-reductase (TGR), is mainly expressed in testis [4]. Thioredoxin reductases are homodimeric flavoproteins with each subunit of approximately 54-58 kDa in size, members of the pyridine nucleotide-disulfide

196

Selenium: Its molecular biology and role in human health

oxidoreductase family and possess two N- and C-terminally located interacting redox-active centres [5,6]. Txnrd3 contains an N-terminal glutaredoxin-like domain giving the enzyme an additional protein-disulfide isomerase function [7]. NADPH/H^ is a cofactor of thioredoxin reductases and reducing equivalents are first transferred to the prosthetic group FAD, from where they are passed to the N-terminal -CVNVGC- catalytic centre of one subunit and subsequently to the C-terminally located redox-active selenenylsulfide of the other subunit [8-11]. Selenocysteine (Sec), which is part of a conserved (-GCUG) motif, is crucial for Txnrd function [12-14]. Due to the easily accessible C-terminal catalytic centre, thioredoxin reductases have a broad range of substrates including hydrogen peroxide, selenite, lipoic acid, NK-lysin, ascorbate, and ubiquinone [15-17]. Cytochrome C was recently shown to be a Txnrd2 substrate [18]. But the main substrates of thioredoxin reductases are thioredoxins (Txn). Thioredoxins are small redox-reactive proteins and involved in numerous physiological processes including cell-cell communication, redox metabolism, proliferation, and apoptosis [19]. For instance, Txn exert a cytokine-like influence on blood cells [20], modulate the activity of redoxregulated transcription factors, such as N F - K B [21] and AP-1 [22], are putatively involved in DNA synthesis, and efficiently protect cells from oxidative damage by acting through peroxiredoxins [16,23]. Several gene targeting approaches in mice have been performed to investigate the participation of the thioredoxins in development and adult physiology. Deletion of either cytosolic thioredoxin (Txnl) or mitochondrial thioredoxin (Txn2) revealed both genes are indispensable for murine embryonic development [24,25]. In Txnl knockout mutants, early embryonic death (E6.5) is associated with dramatically reduced proliferation of the inner mass cells. Txn2-deficient embryos develop exencephaly, show markedly increased apoptosis, and die during midgestation around El0.5. Moreover, in the chicken B cell line DT40, Txn2 is critically involved in the regulation of mitochondria-dependent apoptosis [26]. Heart-specific overexpression of dominant negative Txnl was shown to be associated with increased oxidative sfress and cardiac hypertrophy in mice [27]. While the catalytic mechanisms and structural properties of thioredoxin reductases have been extensively studied in the past, little is known about the individual contribution of the different thioredoxin reductases in living organisms, or about possible redundancies among the different redox systems. A number of reports have linked the thioredoxin/thioredoxin reductase-system to cell proliferation, cancer development, angiogenesis, invasiveness, and drug resistance of tumor cells [28,29], and diseases, such as rheumatoid arthritis. All of these argue for the development and potential therapeutic use of Txnrd inhibitors. Various chemicals, such as 1,3-bis-(2chloro-ethyl)-l-nitrosourea, antirheumatic gold compounds, cisplatin, and a

Mitochondrial and cytosolic thioredoxin reductase knockout mice

197

number of other platinum and organotellurium compounds have been shown to inactivate Txnrd [6,30-32]. However, the presence of several thioredoxin reductases with virtually identical redox-centres may impose a great challenge for the development of drugs selectively inhibiting the activity of one isoform without affecting the other ones. Therefore, mice with targeted deletions of either Txnrd 1 or Txnrd2 might thus prove a most valuable tool to decipher the contribution of the thioredoxin/thioredoxin reductase network in physiology, pathophysiology and disease development. Mouse models with conditional alleles for Txnrdl and Txnrd2 Anticipating that loss of Txnrdl or Txnrd2 might be associated with embryonic death, as already observed in Txnl and Txn2 null mice, and to be able to investigate their functions in specific organs and at defined time points, mice with conditional alleles were established. Txnrdl and Txnrd2 are encoded by genes spanning regions of approximately 39 [33] and 53 kb [34], and are composed of 15 and 18 exons, respectively. The last exon of the Txnrdl gene encodes the final 22 amino acids including the C-terminally located Sec-containing redox-centre. The 1.7 kb long 3' untranslated region additionally contains the Sec insertion sequence (SECTS) element essential for co-translational Sec incorporation at the UGA codon [35], AU-rich mRNA instability elements, and the endogenous transcription termination signal. AU-rich elements have been found in several cytokines and protooncogenes and are responsible for rapid mRNA turnover. Txnrd2 mRNA lacks AU-rich elements. The Sec codon of Txnrd2 is encoded by exon 17 whereas the SECIS element and the poly-adenylation signal are localized on the last exon of the Txnrd2 gene. The gene targeting strategies for both genes aimed at flanking the last exon(s) with loxP sites, which upon Cre-mediated removal, leads to nonfunctional alleles. Only the last exon of Txnrdl was flanked by loxP sites, whereas in case of Txnrd2, the last 4 exons were flanked by loxP sites. The fit-flanked neomycin phosphotransferase gene (neo) required for selection of homologous recombination in embryonic stem (ES) cells was placed downstream of the genes. Several considerations were taken into account to specifically target the 3' regions of both genes: (i) extensive alternative first exon usage, as reported for both Txnrdl [36,37] and Txnrd2 [36], may restrict gene targeting of the 5' regions; (ii) the 5' region of Txnrd2 overlaps with the first exon of the catechol-o-methyltransferase gene [34]; (iii) mutational and biochemical modification of the Sec codon UGA and the Sec moiety [12-14], respectively, as well as deletion of the SECIS element [38] result in inactivation of Txnrd activity. Homologous recombination of both targeting constructs in mouse ES cells and Flp-mediated removal of the neo gene in the loxP flanked (floxed) Txnrd2 alleles was performed [39,40].

198

Selenium: Its molecular biology and role in human health

Backcross of chimeric animals [39] to C57BL/6J mice gave rise to germline transmission of the floxed Txnrdl [41] and Txnrd2 alleles [42]. Embryonic lethality of both Txnrdl and Txnrd2 knockout mice Floxed mice were crossed to congenic C57BL/6J Cre deleter mice resulting in deletion of the floxed alleles in any tissue including the germline [43]. Txnrdl^'" and Txnrd2^'' mice are viable, fertile, show no overt phenotype, and have a normal life span compared to wild-type littermates. Mice with either the floxed or the deleted alleles were backcrossed on a C57BL/6J background to obtain congenic mice. Intercross of hemizygous Txnrdl and Txnrd2 knockout mice never resulted in viable homozygous mutant mice. Genotyping of embryos dissected from hemizygous intercrosses at different days of gestation revealed that the expected Mendelian ratio was maintained up to E10.5 for Txnrdl and E13.5 for Txnrd2; resorption of Txnrdl knockout embryos was frequently observed between gestational days 9.5 and 10.5. RTPCR analysis using embryonic mRNA isolated at E9.5 or El3.5 did not yield products with mRNA's from knockout embryos with primer pairs specific for the deleted exons. However, faint products were obtained with primer pairs covering the central regions indicating very low levels of truncated Txnrdl and Txnrd2 messages. Embryonic expression profile of Txnrdl and Txnrd2 To better understand the basis of the null phenotypes of Txnrdl and Txnrd2 knockout embryos, the expression profile of both genes was studied. At E8.5, Txnrdl is expressed throughout the entire embryo with the exception of the primitive heart with the highest levels detected in neuronal tissues such as the developing forebrain and the rhombomeres. At E9.5, Txnrdl expression is confined to the neural tube, the forebrain, branchial arches, somites, and to the limb buds. At E10.5, Txnrdl is present in developing somites, in the apical ectodermal ridge of the limb buds, in the first and second branchial arches, and in the lateral edges of the nasal pit. Thus, Txnrdl shows a complex and dynamic expression pattern in early developmental stages. To study embryonic Txnrd2 expression at E12.5, sections of mice with a conditional lacZ knock-in into the Txnrd2 locus were used. Anti-lacZ immunohistochemistry of hemizygous Txnrd2 lacZ knock-in embryos revealed strong expression in the embryonic heart, especially in the myocardium and atrial walls, and to a lower extent in the embryonic liver. These findings reflect Txnrd2 expression with mRNA data obtained from adult tissues [44], and associate Txnrd2 function with organs characterised by high metabolic activity. This may further corroborate a crucial role for Txnrd2 in the control of harmful intracellular reactive oxygen species. In short, the distinct expression patterns of Txnrdl and Txnrd2, especially the

Chapter 19. function

Selenium, deiodinases and endocrine

Antonio C. Bianco and P. Reed Larsen Thyroid Section, Division of Endocrinology, Diabetes and Hypertension, Department of Medicine, Brigham and Women's Hospital and Harvard Medical School, Boston, MA 02115, USA

Summary: The three iodothyronine deiodinases catalyze the initiation (Dl, D2) and termination (D3) of thyroid hormone effects in vertebrates. A 3dimensional model predicts that these enzymes share a similar structural organization and belong to the thioredoxin (TRX) fold superfamily. Their active center is a selenocysteine-containing pocket defined by the pi-al-p2 motifs of the TRX fold and a domain that shares strong similarities with the active site of iduronidase, a member of the clan GH-A fold of glycoside hydrolases. While Dl and D3 are long-lived plasma membrane proteins (tl/2 10-12 h), D2 is an endoplasmic reticulum resident protein that is inactivated by selective conjugation to ubiquitin, a process that is mediated by WSB-1, a Hedgehog-inducible gene. Remarkably, D2 ubiquitination is reversible and activity restored after deubiquitination by the pVHL-interacting deubiquitinating enzymes (VDUl and VDU2). Deiodinases play a major role in development as well as in adults as critical players in thyroid hormone homeostasis, particularly during hypo- and hyperthyroidism. In addition to playing an important part in energy homeostasis, changes in deiodinase activity explain the alterations in thjroid economy observed during illness and in the recently described syndrome of consumptive hypothyroidism. Introduction The three deiodinases, enzymes that activate thyroxine (T4) and inactivate both T4 and T3, are present in all vertebrates. Their relevance resides in the fact that T4 is a long-lived (tl/2 is ~7 days in humans) pro-hormone molecule that must be activated by deiodination to the short-lived biologically active T3 (tl/2 is ~1 day) in order to initiate thyroid hormone action. T3 modulates gene expression in virtually every vertebrate tissue through ligand-dependent transcription factors, the thyroid hormone receptors. The deiodination of T4 to T3 occurs in the phenolic (outer or 5')ring of the T4 molecule and is catalyzed by two iodothyronine deiodinases, i.e. Dl and D2. As a counter point to the activation pathway, T4 activation

208

Selenium: Its molecular biology and role in human health

can be prevented and T3 can be irreversibly inactivated by deiodination of the tyrosyl (inner or 5)-ring, a reaction catalyzed by D3 (and Dl). In both experimental animals and humans, the coordinated changes in the expression and activity of these enzymes ensures thyroid hormone homeostasis and the constancy of T3 production, constituting a major mechanism for adaptation to changes in the ingestion of iodine, starvation and changes in environmental temperature (reviewed in [1]). The study of animals with deficiency of Dl (C3H mouse) or targeted disruption of D2 (Dio2''") or D3 (D3'") genes has not only confirmed but revealed new intricacies about the critical role played by these enzymes in thyroid hormone homeostasis [2-5]. The 3D structure of the deiodinases is conserved The three deiodinase proteins (Dl, D2 and D3) show considerable similarity (~ 50 % sequence identity). All are integral membrane proteins of 29-33 kDa, and have regions of high homology in the area surrounding the active center [6-8]. Insights into the structures of these proteins were obtained through protein modeling using hydrophobic cluster analysis (HCA) [9]. Based on the HCA analysis it is clear that the three deiodinases share a common general structure composed of a single trans-membrane segment, which is present in the N-termini of Dl, D2 and D3, and several clusters, typical of a-helices or P-strands, corresponding to core secondary structures of the deiodinase globular domains. A striking common feature is the presence of the thioredoxin (TRX) fold, defined by the paP and ppgt motifs. It is interesting that, within the canonical TRX fold, the relationship between the pap and ppa motifs is locally interrupted by interfering elements. These sequences correspond to distinct secondary structure elements added to the canonical TRX fold core, a feature also observed in other proteins of the TRX fold family [10]. A unique aspect of the deiodinases, however, is that one of these highly conserved intervening elements shares similarities with a-L-iduronidase (IDUA; 47 % identity with Dl and D3, 60% with D2), a lysosomal enzyme that cleaves a-linked iduronic acid residues fi-om the nonreducing end of glycosaminoglycans [11]. The 3D general model of the deiodinases predicts that the active center is a pocket defined by the pi-al-P2 motifs of the TRX-fold and the IDUA-like insertion. A striking feature of this pocket is the rare amino acid selenocysteine (Sec), critical for the deiodination reaction catalyzed by all three deiodinases. This was first identified when the rat Dl cDNA became available, the analysis of which revealed the presence of the Sec encoded by UGA, which is recognized in the vast majority of mRNAs as a STOP codon [12]. However, a specific RNA stem-loop immediately downstream of the UGA codon allows for the Sec incorporation in the STOP codon. This structure is termed the Sec Insertion Sequence, or SECIS element, which is present in the deiodinases and all other selenoproteins [13].

Selenium, deiodinases and endocrine function

209

The ubiquitination pathway inactivates D2 D2 is considered the critical homeostatic T3-generating deiodinase due to its substantial physiological plasticity (see [1] for review). A number of transcriptional and post-translational mechanisms have evolved to ensure limited expression and tight control of D2 levels, which is inherent to its homeostatic function. The D2 mRNA in higher vertebrates is more than 6 kb in length, containing long 5' and 3' untranslated regions (UTRs). The D2 5'UTRs are greater than 600 nucleotides and contain 3-5 short open reading frames (sORFs), which reduce D2 expression by as much as 5-fold [14]. Alternative splicing is another mechanism that regulates D2 level as mRNA transcripts similar in size to the major 6- to 7-kb D2 mRNAs, but not encoding an active enzyme, are present in both human and chicken tissues. D2 levels can also be regulated by AUUUA instability motifs located in the 3'UTR of D2 mRNA as deletion of 3.7-kb from this region increases D2 activity ~3.8-fold due to an increase in D2 mRNA half-life [14]. D2 activity/mRNA ratios are variable, indicating that there is significant post-translational regulation of D2 expression [15]. In fact, the decisive D2 property that characterizes its homeostatic behavior is a short half-life (-40 min) [16] that can be further reduced by exposure to physiological concentrations of its substrate, T4, and in experimental situations, reverse T3 or even high concentrations of T3 [16-22]. This constitutes a rapid, potent generalized regulatory feedback loop that efficiently controls T3 production and intracellular T3 concentration based on how much T4 is available. Important metabolic pathways often contain key rate-limiting enzymes whose half-lives can be modified by selective proteolysis. This process is mediated by the ubiquitin (Ub)-proteasome system by which target proteins are marked for destruction by conjugation to Ub, a ~8kDa protein. The ubiquitinated proteins are subsequently recognized and degraded by the proteasomes [23,24]. Indeed, ubiquitination and proteasomal degradation are deeply implicated in the post-translational regulation of D2 activity. The first evidence was obtained in GH4C1 cells in which the half-life of the endogenous D2 was noted to be stabilized by MG132, a proteasome inhibitor [25]. Substrate-induced loss of D2 activity was also inhibited by MG132 in such cells, indicating that both pathways affecting loss of D2 activity were mediated by the proteasomes. This implies that the loss of D2 activity is, at least partially, due to proteolysis, a premise that was confirmed after the levels of immunoprecipitable labeled D2 were shown to parallel D2 activity, both under basal conditions and after exposure to substrate [26]. Selection of specific proteins for proteolysis is usually achieved at the level of Ub conjugation, a process that involves recognition of amino acidsequences in the target protein by the ubiquitinating enzymatic machinery. The first step is activation of Ub by ATP, a process catalyzed by the El enzyme. The next step, target recognition, is coordinated by the combined

210

Selenium: Its molecular biology and role in human health

actions of a series of Ub-conjugating enzymes (E2s) and Ub-ligases (E3s). Individual E2s are involved in different cellular processes and, therefore, in the ubiquitination of different classes of substrate proteins. In the case of D2, we have identified UBC6 and UBC7, two E2s that specifically assist in the transfer of activated Ub to D2 [27,28]. E3s, on the other hand, are more abundant and with no overt sequence homology, are thought to be largely responsible for the high degree of specificity of protein ubiquitination [24]. Using a yeast-two hybrid system to screen a brain library we identified a novel D2-interacting protein, WSB-1, which is a SOCS-box-containing WD-40 protein of unknown function that is induced by hedgehog signaling in embryonic structures during chicken development [29]. We subsequently showed that WSB-1 acts as an E3 ubiquitin ligase for D2. The WD-40 propeller of WSB-1 recognizes an 18amino acid loop in D2 that confers metabolic instability, while the SOCSbox domain mediates its interaction with an ubiquitinating catalytic core complex, modeled as Elongin BC-Cul5-Rbxl (ECS^^°"'). In the developing tibial growth plate, hedgehog-stimulated D2 ubiquitination via ECS^^^"' induces parathyroid hormone related peptide (PTHrP), thereby regulating chondrocyte differentiation. Thus, ECS*^^"' mediates a novel mechanism by which "systemic" thyroid hormone can effect local control of the hedgehogPTHrP negative feedback loop and thus skeletogenesis [29]. Using the same yeast two hybrid system that identified WSB-1, we identified D2 as the only known specific substrate of VDUl and VDU2 [30], which in turn are the first ubiquitin-specific processing proteases (UBP) known to specifically deubiquitinate an ERAD substrate. These results show that protein recognition is not only involved in the E3-mediated ubiquitination process but also in the deubiquitination pathway catalyzed by UBPs. Both VDUs are downstream targets for ubiquitination by pVHL E3 ligase, and VHL mutations that disrupt the interaction between the VDUs and pVHL abrogate their ubiquitination [31,32]. Although hundreds of UBP enzymes have been cloned, only a few examples of substrate recognition by UBP enzymes have been reported and, to our knowledge, none are ERresident proteins [33-37]. Confocal studies indicate that both VDUs colocalize with D2, itself an integral membrane ER-resident protein. Although present in the particulate fi-action and not in cytosol, it is not clear, based on their hydrophobic profile, whether VDUl/2 are integral membrane proteins [30]. Their physical colocalization with D2, however, provides the opportunity for catalysis and D2 deubiquitination. Thus, due to the intrinsic inefficiency of the selenoprotein synthesis, the availability of a reversible ubiquitination-dependent mechanism to control the activity of D2 constitutes an advantage that allows for rapid control of thyroid hormone activation. The finding that VDUl and VDU2 are coexpressed with D2 in many human tissues, including brain, heart and skeletal

Selenium, deiodinases and endocrine function

211

muscle [1,31], indicates that the importance of this mechanism may extend well beyond thermal homeostasis to include brain development, cardiac performance, glucose utilization and energy expenditure. Role of deiodinases in thyroid hormone homeostasis T3 can be produced by two different and relatively independent sources, namely the direct product of the thyroid secretion or as a result of extrathyroidal deiodination of T4. The relative contributions of the two sources of T3, thyroid secretion and T4 5' deiodination, can be quantified by determining the T4 to T3 conversion rate, which is on average about 36 % [38]. Hence, with a normal T4 production rate of 110 nmol Id, 40 nmol of T3 are produced by peripheral deiodination of T4 and the remaining 10 nmol are secreted directlyfi-omthe thyroid gland. Extrathyroidal T3 can derive from T4 via two different deiodination pathways, namely Dl or D2. To quantitate the role of Dl in catalyzing the production of plasma T3, it is informative to review the results of two studies performed in patients with primary hypothyroidism receiving fixed doses of exogenous T4 [39,40]. In these patients, administration of PTU (1000 mg/day) caused about a 25 % decrease in serum T3. In a third study, the production of labeled plasma T3 from T4 was not reduced in patients given 1200 mg/day of PTU [41]. Results of these three studies favor the concept that, except for during hyperthyroidism, Dl-catalyzed T3 production is not a major component of exfrathyroidal T3 production in euthyroid humans. In a more recent study, we modeled in vitro the in vivo situation and calculated the rate of T4 to T3 conversion by intact cells fransiently expressing Dl or D2 at low (2 pM), normal (20 pM), and high (200 pM) free T4 concentrations. Deiodinase activities were then assayed in cell sonicates. The ratio of T3 production in cell sonicates (catalytic efficiency) was multiplied by the tissue activities reported in human liver (Dl) and skeletal muscle (D2). From these calculations, we predicted that in euthyroid humans, D2-generated T3 is 29 nmol/d, while that of Dl-generated T3 is 15 nmol/d, from these major deiodinase-expressing tissues. The total estimated extrathyroidal T3 production, 44 nmol/d, is in close agreement with the 40 nmol T3/d based on previous kinetic studies. D2-generated T3 production accounts for approximately 71% of the peripheral T3 production in hypothyroidism, but Dl for approximately 67% in thyrotoxic patients [42]. Deiodinases mediate tissue-specific control of thyroid hormone action Given the generalized metabolic sensitivity to thyroid hormone documented during hypo- and hyperthyroidism, one would anticipate a major physiological role of this hormone in energy homeostasis. However, serum T3 concentration is remarkably constant, thus precluding a major role of T3 in the basal metabolic rate (BMR) variations observed after a meal or during

212

Selenium: Its molecular biology and role in human health

sleep. In the last 20 years, light was shed on this problem by studies demonstrating that in some tissues the cellular actions of thyroid hormone are determined not only by serum T3. Thyroid hormone action is initiated through its binding to nuclear receptors, which are high-affinity nuclear T3 binding proteins that regulate transcription of T3-dependent genes. Receptor occupancy is determined by the affinity of the receptor for T3 and the T3 concentration in the nucleus. These values are such that, at normal serum T3 concentration, the contributionfi^omserum T3 alone results in an approximately 50% saturation of thyroid hormone receptors in most tissues. However, tissues expressing D2 have an additional source of T3 contributed by the conversion of intracellular thyroxine (T4) to T3 [43,44]. As a result, receptor saturation can reach as high as 100 %, with more than half of this T3 produced locally [4547]. While we still do not understand all the intricacies of this system, we do know that generation of T3 by D2 occurs in the perinuclear region, a cellular compartment with preferential access to the nucleus. This is in contrast to Dl, which is localized to the plasma membrane, from which the T3 produced more readily enters the plasma [48]. Thus, for cells lacking D2, intracellular thyroid status is determined predominantly by the serum T3 concentration. In contrast, cells expressing D2 have the ability to generate intracellular T3 fi-om T4. Thus, cells expressing D2 have two potential sources of nuclear T3: plasma T3, or T3(T3), and T3(T4). On the other hand, D3 is localized to the plasma membrane and to undergo recycling in the endosomes. Remarkably, D3 expression causes cell hypothyroidism as a result of its inactivating effect on thyroid hormone [49], thus creating a virtual barrier that prevents thyroid hormone from entering the cell nucleus. Thus, despite steady serum T3 levels, intracellular thyroid status varies along a wide range according to the type and level of deiodinase expression. Thyroid hormone receptor saturation is expected to be minimum in cells expressing D3 and maximum in cells expressing D2. In addition, because of the plasticity of deiodinase expression, particularly that of D2, receptor saturation of a single cell type might change rapidly and dramatically without affecting serum T3. Thus, deiodinase expression in metabolically active tissues is a potent mechanism by which energy dissipation can be controlled. Significance of D2 to adaptive tliermogenesis in Iiumans The potential role of D2 in human energy homeostasis has been ignored because human newborns grow less dependent on BAT thermogenesis, and adult humans, unlike small mammals, do not have substantial amounts of BAT [50]. However, since the cloning of the human D2 cDNA and the finding of cAMP-inducible D2 mRNA and activity in human skeletal muscle [51,52], the role of D2 in controlling human adaptive thermogenesis has been

Selenium, deiodinases and endocrine function

213

revisited. Thyroid hormone per se is known to increase energy expenditure in skeletal muscle (see [53] for review) and could also regulate local energy homeostasis through its interaction with the SNS. Accordingly, human skeletal muscle is under the influence of the thyroid-adrenergic synergism and an increase in local cAMP production is known to activate glycolytic enzymes, sarcolemmal Na^/K^ pumps, phospholamban, and voltage-sensitive and sarcolemmal Ca^^ channels [54,55], resulting in increased glucose uptake and utilization [56,57]. The expression of GLUT4, the insulinresponsive-glucose transporter that mediates glucose metabolism in skeletal muscle, is also up-regulated by thyroid hormone [58]. Various studies support a previously unrecognized role of D2 in determining the thyroid status and metabolic rate of the skeletal muscle, analogous to its role in BAT. Earlier experimental studies of humans [59] have consistently found diet-induced changes in serum thyroid hormones that could be explained by changes in D2 activity. As an example, the increase in BMR observed in subjects fed a high carbohydrate diet is typically associated with an increase in the serum T3/T4 ratio [59], a condition that is also observed in adult subjects chronically treated with terbutaline, a Padrenergic receptor (P-AR) stimulator [60]. This indicates the existence of a relevant cAMP-dependent T4-to-T3 conversion pathway in humans that plays a role in energy homeostasis. That this pathway is predominantly through D2 is supported by the finding that the D2 gene is up-regulated several fold by adrenergic stimulators and cAMP [61]. Studies of patients receiving T4 replacement at various dosages have shown a direct correlation of the BMR with free T4 and inversely with serum TSH but not with serum T3 [62]. Together, these data indicate that D2-produced T3 in skeletal muscle might be a significant determinant of energy expenditure in humans. Recent studies describing a Dio2 polymorphism in which a threonine (Thr) change to alanine (Ala) at codon 92 (D2 Thr92Ala) provides is additional support to a role of D2 in glucose uptake and utilization. Of note, in humans, skeletal muscle is the primary site of insulin-dependent glucose disposal [63]. Remarkably, this Dio2 polymorphism was associated with an -20 % lower rate of glucose disposal in obese women than in non-obese women [64]. In addition, the fi-equency of the variant allele was found to be increased in some ethnic groups, such as Pima Indians and MexicanAmericans, with a higher prevalence of insulin resistance [64]. The possible role of the D2 Thr92Ala polymorphism on insulin resistance was also investigated in patients with type 2 diabetes Mellitus (DM2). Studies of these patient offers a practical approach to the investigation of energy expenditure because they require intense metabolic monitoring and are subjected to a detailed scrutiny of fuel utilization. In accordance with the previous study in obese individuals, individuals homozygous for the variant allele have an increased insulin-resistance index as assessed by the HOMA

214

Selenium: Its molecular biology and role in human health

(homeostasis model assessment) index. The increased insulin resistance observed in the DM2 patients homozygous for the Ala allele could be explained by a decrease in D2 activity, as has been found in thyroid and skeletal muscle samples from individuals with this genotype [65]. A lower D2 activity would decrease D2-generated T3 in skeletal muscle and create a state of relative intracellular hypothyroidism, decreasing the expression of genes involved in energy utilization, such as GLUT4, leading to insulin resistance. Supporting this hypothesis is the remarkable finding that the UCPl knock-out mouse develops, as a compensatory mechanism, increased D2 activity in white adipose tissue [66], stressing the importance of understanding the D2-generated T3-dependent thermogenic mechanisms. Changes in iodothyronine deiodination during fasting or illness It has been recognized for decades that there are significant changes in the concentrations of circulating thyroid hormones during illness or starvation in human plasma. Despite numerous studies, there remains much controversy regarding both the precise etiology of these changes and what, if anything, should be done therapeutically regarding them [67-69]. The hallmark of these changes is a decrease in circulating free T3 and an increase in total reverse T3, although there is not complete agreement even on these most basic changes [70-72]. The similarity of the changes in illness to those of fasting or caloric deprivation suggest that the decrease in thyroid hormone activation is a beneficial physiological response designed to reduce metabolic rate and conserve energy during stress [73]. In patients postcoronary artery bypass grafting, however, there is disagreement about the effectiveness of T3 supplementation with one study showing a positive effect [74,75]. The changes in circulating thyroid hormones and TSH during illness are a continuum with progressively more abnormalities with more severe illness. Patients with mild illnesses, such as after uncomplicated surgery, or who are fasting, generally have a reduction of up to 50% in circulating T3, a reciprocal increase in serum reverse T3 and no changes in serum T4 or TSH [72,76]. With moderately severe illness, the clearance of T4 is slowed while T4 secretion persists, leading to increases in free T4 accompanied by further decreases in serum T3 and increases in reverse T3. When the abnormalities in T3 and reverse T3 were first described, the initial assumption was that these reciprocal changes reflected diversion from T4 activation to inactivation. This raised the possibility that the changes could be attributed to an alteration specificity of Dl from 5' to 5 T4 deiodination since this is the only deiodinase with the capacity to catalyze both outer and inner ring deiodination of T4. Subsequent studies indicated that the elevation in reverse T3 was due to a reduction in the clearance of this T4 byproduct and its production rate is unchanged as long as T4 remains

Selenium, deiodinases and endocrine function

215

normal [76]. This indicates that the tissues in which reverse T3 is produced from T4, largely by the action of D3, are processing T4 normally during illness or fasting. On the other hand, since the principal pathway for reverse T3 clearance is via Dl, these results indicate either that the Dl enzyme or its co-factor is reduced, or that the uptake of reverse T3 into Dl-expressing tissues is impaired [77]. Decreased transport of reverse T3 into the Dlcontaining liver or kidney during fasting or illness has been attributed to either ATP depletion or interference with reverse T3 transport by competing substances circulating in the plasma [67]. In moderate to severe illness, the serum T3 can fall to 20-30 % of baseline. About 20% of T3 in human plasma derives from the thyroid with the T3 derived from extrathyroidal Dl and D2 catalyzed T4 5' monodeiodination accounting for about 25 and 55% of the plasma T3, respectively. Since TSH, and therefore T3 secretion, is not suppressed unless illness is prolonged and/or severe, the severely reduced T3 in ill patients is due to decreased peripheral T4 deiodination by Dl, D2 or both. The fact that the fall in T3 substantially exceeds what we can currently assign to Dl (~30 %) suggests that T3 generation by D2 must also be inhibited. With respect to Dlcontaining tissues, T4 uptake into the rapidly equilibrating pool, primarily liver and kidney, is significantly reduced in obese patients on a 240 kcal diet and similar observations have been made in uremia [78]. This can explain the decrease in T3 production via Dl and, again, either inhibition of T4 transport by unknown circulating compounds or by ATP depletion could be to blame [67]. In addition, entry of T4 into the slowly equilibrating pool, likely to be the one in which D2-catalyzed T3 production occurs, is also reduced in obese patients on a hypocaloric diet [78]. In addition to reduced T4 transfer, a second important consideration with respect to D2-catalyzed T4 to T3 conversion is the rapid proteolysis of D2 through the ubiquitin-proteasome pathway. Thus, persistent D2 synthesis is required to maintain D2 levels normal. Protein synthesis is impaired during fasting or in severe illness. It is tempting to speculate that a rapid fall in D2 protein can explain the abrupt decrease in plasma T3 associated with these conditions. The possibility that D3 action is also increased during illness has also been considered. In a study that determined serum thyroid hormone levels and the expression of Dl, D2, and D3 in liver and skeletal muscle from deceased intensive care patients, liver Dl was down-regulated and D3 was induced in liver and skeletal muscle [79]. Dl and D3 mRNA levels corresponded with enzyme activities, suggesting regulation of the expression of both deiodinases at the pre-translational level. Liver Dl was down-regulated and D3 (which is not present in liver and skeletal muscle of healthy individuals) was induced, particularly in disease states associated with poor tissue perfusion. These observations may represent tissue-specific ways to reduce thyroid hormone bioactivity during illness [79].

216

Selenium: Its molecular biology and role in human health

D3 overexpression in hemangiomas D3 activity in the normal uteroplacental unit regulates the transfer of maternal thyroid hormone to the fetus [80]. D3 is expressed in multiple fetal structures, but the endometrium and the placenta are the only normal tissues known to express high levels of D3 activity in the mature human. D3 has also been found in vascular anomalies, in human brain tumors, and in some malignant cell lines. These studies have led to the categorization of D3 as an oncofetal protein, but recent data indicate that postnatal expression can be reactivated in normal tissues during critical illness [79]. D3 expression at high levels occurs in infantile hemangiomas [81]. If these tumors are sufficiently large, the rate of thyroid hormone inactivation can exceed the maximal rate of thyroid hormone synthesis. The first patient documented with this condition was 3 months old, presenting with severe hypothyroidism with an elevation in serum TSH, undetectable serum T4 and T3 concentrations and high reverse T3 and thyroglobulin. The relationship between infantile hemangiomas and D3 expression is especially significant since it identifies a previously unrecognized cause of hypothyroidism, which usually occurs at a critical age for neurological development. While extensive hepatic hemangiomas can be fatal, a significant fraction of these infants survive due to therapy and the natural tendency of these tumors to regress. Accordingly, these patients may require replacement with large quantities of thyroid hormone in addition to therapy directed at their hemangiomas. Thyroid hormone treatment is also imperative to prevent the complication of irreversible mental retardation later in life. Hemangiomas produce high quantities of basic fibroblast growth factor, which has been shown to activate the expression of D3 in rat glial cells via ERK activation [82]. It seems likely that this is one mechanism for the high D3 expression in these tumors. Using nontransformed human cells, it has been shown that TGF-beta stimulates transcription of the hDio3 gene via a Smad-dependent pathway. Combinations of Smad2 or Smad3 with Smad4 stimulate hDio3 gene transcription only in cells that express endogenous D3 activity, indicating that Smads are necessary but not sufficient for D3 induction. TGF-beta induces endogenous D3 in human cell types such as fetal and adult fibroblasts from several tissues, hemangioma cells, fetal epithelia, and skeletal myoblasts. Maximum stimulation of D3 by TGF-beta also requires MAPK and is synergistic with phorbol ester and several mitogens known to signal through transmembrane receptor tyrosine kinases but not with estradiol. These data reveal a previously unrecognized interaction between two pluripotent systems, TGF-beta and thyroid hormone, both of which have major roles in the regulation of cell growth and differentiation [83].

Selenium, deiodinases and endocrine function

217

Dl overexpression contributes to the relative excess of T3 production in hyperthyroidism It has been well established that the production rate of T3 and its circulating concentration is about 2-fold higher relative to that of T4 in hyperthyroid patients [84]. This is reflected in the markedly greater elevation in free T3 than in free T4 in such patients. Since the human Diol promoter is T3responsive, one would anticipate that Dl activity or mRNA would be significantly increased in hyperthyroid patients. This has been demonstrated in Graves' thyroid tissue and in mononuclear leukocytes of patients with Graves' disease [85-87]. It would be expected that PTU, a drug that blocks Dl, but not D2 activity, would have a greater effect on plasma T3 production in thyrotoxic than euthyroid individuals in whom Dl activity should be increased and D2 activity reduced. That PTU inhibits T4 to T3 conversion was demonsfrated in a series of patients comparing the acute changes in serum T3 between Graves' patients freated with a combination of iodide and PTU with a group of similar patients treated with methimazole and iodide [84]. These results indicate that a PTU-inhibitable process, Dl-catalyzed T4 to T3 conversion, is more active in the hyperthyroid than the euthyroid subject in which PTU causes 0-25 % decrease in T3. This has led to the recommendation that large doses of PTU or other agents which block T4 to T3 conversion, such as iopanoic or ipodipic acid, be used in the acute treatment of the severely hyperthyroid individual [88-91]. A paradoxical observation in the Graves' thyroid is that thyroidal D2 mRNA and activity is increased despite systemic thyrotoxicosis [92]. This is due to the effect of the thyroid immunostimulator to activate the cAMPdependent hDio2 promoter, which must overwhelm the negative transcriptional effect of T3 on hDio2. Furthermore, the presence of D2 activity in the Graves' or TSH-stimulated human thyroid raises the possibihty that a portion of the excess T3 secretion in Graves' disease results from intrathyroidal T4 to T3 conversion catalyzed by D2 [92]. References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11.

AC Bianco, D Salvatore, B Gereben, MJ Berry, PR Larsen 2002 Endocr Rev 23:38 MJ Berry et al 1993 J Clin Invest 92:1517 MJ Schneider et al 2001 Mol Endocrinol 15:2137 A Hernandez et al 2003 In 85th Annual Meeting of the Endocrine Society Philadelphia, PA LA de Jesus et al 2001 J Clin Invest 108:1379 MJ Berry, JD Kieffer, JW Harney, PR Larsen 1991 J Biol Chem 266:14155 W Croteau, SL Whittemore, MJ Schneider, DL St Germain 1995 J Biol Chem 270:16569 C Buettner, JW Harney, PR Larsen 2000 Endocrinology 141:4606 I Callebaut et al 2003 J Biol Chem 278:36887 JL Martin 1995 Structure 3:245 PM Coutinho, B Henrissat 1999 Server at http://afinbcnrs-mrsfr/~cazy/CAZY/indexhtml

218

12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 31. 32. 33. 34. 35. 36. 37. 38.

39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62.

Selenium: Its molecular biology and role in human health

MJ Berry, L Banu, PR Larsen 1991 Nature 349:438 MJ Berry et al 1991 Nature 353:273 B Gereben, A KoUar, JW Harney, PR Larsen 2002 Mol Endocrinol 16:1667 Burmeister, LA, Pachucki, J, DL St Germain 1997 Endocrinology 138:5231 DL St Germain 1988 Endocrinology 122:1860 JL Leonard, MM Kaplan, TJ Visser, JE Silva, PR Larsen 1981 Science 214:571 RJ Koenig et al 1984 Endocrinology 115:324 JE Silva, JL Leonard 1985 Endocrinology 116:1627 Y Halperin, LE Shapiro, MI Surks 1994 Endocrinology 135:1464 JL Leonard, et al 1984 Endocrinology 114:998 MJ Obregon, PR Larsen, JE Silva 1986 Endocrinology 119:2186 O Coux, K Tanaka, AL Goldberg 1996 Annu Rev Biochem 65:801 A Hershko, A Ciechanover 1998 Annu Rev Biochem 67:425 J Steinsapir, JW Harney, PR Larsen \99S J Clin Invest 102:1895 J Steinsapir et al 2000 £nrfocnno/ogy 141:1127 BW Kim et al 2003 Mol Endocrinol 17:2603 D Botero et al 2002 Mol Endocrinol 16:1999 M Dentice et al 2003 J Clin Invest 112:189 Z Li, X Na, D Wang, SR Schoen, EM Messing, G Wu 2002 J Biol Client 277:4656 Z Li et al 2002 Biochem Biophys Res Commun 294:700 S Taya, T Yamamoto, M Kanai-Azuma, SA Wood, K Kaibuchi 1999 Genes Cells 4:757 N Gnesutta et al 2001 J Biol Chem 276:39448 S Taya et al 199% J Cell Biol 142:1053 X Chen, B Zhang, JA Fischer 2002 Genes Dev 16:289 H Ideguchi et al 2002 Biochem J i6T.87 PR Larsen, TF Davies, ID Hay 1998 In Williams Textbook of Endocrinology JD Wilson, DW Foster, HM Kronenberg, PR Larsen, editors Philadelphia: WB Saunders Co 389-515 DL Geffher, M Azukizawa, JM Hershman 1915 J Clin Invest 55:224 M Saberi, FH Sterling, RD Utiger 1975 J Clin Invest 55:218 JS LoPresti et al 1989 J Clin Invest 84:1650 AL Maia, BW Kim, SA Huang, JW Harney, PR Larsen 2005 J Clin Invest 115:2524 JE Silva, PR Larsen 1977 Science 198:617 JE Silva, JL Leonard, FR Crantz, PR Larsen 1982 y Clin Invest 69:1176 JE Silva, TE Dick, PR Larsen 1978 Endocrinology 103:1196 JE Silva, PR Larsen 1978 7 Clin Invest 61:1247 AC Bianco, JE Silva 1987 Endocrinology 120:55 MM Baqui et al 2000 Endocrinology 141:4309 FW Wassen et al 2002 Endocrinology 143:2812 K Bruck, 1998 In Fetal neonatal physiology RA Polin, WW Fox (eds) Philadelphia WB Saunders Co 676 D Salvatore, T Bartha, JW Harney, PR Larsen 1996 Endocrinology 137:3308 W Croteau et al 1996 J Clin Invest 98:405 L de Meis 2001 fi/osci/Jep 21:113 YT Yang, MA McElligott 1989 Biochem J26l:\ SJ Roberts, RJ Summers 1998 Eur J Pharmacol 348:53 FD McCarter et al 2001 J Surg Res 99:235 LG Fryer et al 2002 Biochem J 363:167 O Ezaki 1997 Biochem Biophys Res Commun 241:1 JE Danforth et al 1979 J Clin Invest 64:1336 K Scheidegger et al 1984 J Clin Endocrinol Metab 58:895 T Bartha et al 2000 £Wocnno/ogv 141:229 H al-Adsani, LJ Hoffer, JE Silva 1997 J Clin Endocrinol Metab 82:1118

Selenium, deiodinases and endocrine function 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. 88. 89. 90. 91. 92.

E Ferrannini et al 1999 Eur J Clin Invest 29:842 D Mentuccia et al 2002 Diabetes 51:880 LH Canani et al 2005 J Clin Endocrinol Metab 90:3472 X Liu et al 2003 J Clin Invest 111:399 R Docter et al 1993 Clin Endocrinol 39:499 B Mclver, CA Gorman 1997 Tliyroid 7:125 LJ De Groot 1999 J Clin Endocrinol Metab 84:151 J Faber, K Siersbaek-Nielsen 1996 Clin Chim Acta 256:115 IJ Chopra 1998 Uyroid 8:249 MM Kaplan et al 1982 Am J Med 72:9 DF Gardner, MM Kaplan, CA Stanley et al. 1979 N Engl J Med 300:579 JD Klemperer et al 1995 N EnglJ Med 333:1522 E Bennett-Guerrero et al 1996 JAMA 275:687 Z Eisenstein et al \978 J Clin Endocrinol Metab 47:889 EM Kaptein et al 1983 7 Clin Endocrinol Metab 57:181 JTM van der Heyden et al \9&6 Am J Physiol 25\:E156 RP Peelers et al 2003 J Clin Endocrinol Metab 88:3202 SA Huang 2005 Thyroid 15:875 SA Huang et al 2000 N EnglJ Med 343:185 S Pallud et al 1999 Endocrinology 140:2917 SA Huang et al 2005 Mol Endocrinol 19:3126 J Abuid, PR Larsen 1974 J Clin Invest 54:201 H Ishii et al 1981 7 Clin Endocrinol Metab 52:1211 M Sugawara et al 1984 Metabolism 33:332 M Nishikawa et al 1998 Biochem Biophys Res Commun 250:642 SY Wu, TP Shyh, I Chopra et al 1982 J Clin Endocrinol Metab 54:630 SY Wu et al 1982 J Clin Endocrinol Metab 54:630 H Burgi et al 1976 J Clin Endocrinol Metab 43:1203 MS Croxson, TD Hall, JT Nicoloff 1977 J Clin Endocrinol Metab 45:623 D Salvatore, H Tu, JW Harney, PR Larsen 1996 J Clin Invest 98:962

219

Chapter 20. Biotechnology of selenocysteine Linda Johansson and Elias S. J. Amer Medical Nobel Institute for Biochemistry, Department of Medical Biochemistry and Biophysics, Karolinska Institute, SE-171 77 Stockholm, Sweden

Summary: In this chapter we describe strategies to produce synthetic selenoproteins, with a focus on recombinant selenoprotein production in E. coli. We further discuss the possible use of selenocysteine (Sec) in proteins for biotechnological applications. Such applications are based upon either the introduction of a selenium isotope, with specific characteristics such as highenergy radioactivity (as for ^'Se and ^^Se) or nuclear spin (^^Se), or on the high reactivity of the nucleophilic Sec residue enabling site-specific conjugation-based applications with electrophilic ligands. Utilization of Sec insertion for protein purification or detection purposes has recently been demonstrated in a number of different experimental systems and we envision that further such applications will be developed in the near future. Introduction Production of synthetic selenoproteins, although far fi-om trivial, is possible by utilizing various chemical or genetic approaches. Such production may be highly useful in enabling detailed studies of selenoproteins in general. Features of selenocysteine (Sec) not shared by any of the other 20 common amino acids, such as the selenium atom chemistry, isotope abundance and high nucleophilicity, make it tempting to introduce Sec into nonselenoproteins for use in Sec-targeted biotechnological applications. For that purpose, we developed a tetrapeptide motif for recombinant proteins expressed in E. coli containing a Sec-residue, named a Sel-tag [1], which will be described at the end of this chapter. Methods to obtain selenoproteins and numerous theoretically possible applications based upon the presence of a Sec residue in protein are summarized in Figure 1, illustrating the true biotechnological potential unleashed by the use of Sec. Production of selenoproteins In order to use Sec in proteins for Sec-based biotechnological applications, the first step is, naturally, to obtain the selenoprotein to serve as the basis for such applications. Purifying native selenoproteins fi-om natural sources (Step A, Figure 1) is ofl;en laborious, time-consuming and results in low product

222

Selenium: Its molecular biology and role in human health

yields. Transfection of eukaryotic cells for eukaryotic selenoprotein overproduction is an alternative, but has hitherto also resulted in low yields [2,3]. One technique for producing selenoproteins is to synthetically incorporate a Sec residue into a target protein using chemical substitution reactions (Step B, Figure 1). Different methods for chemical synthesis of selenoproteins have been described, involving either native chemical ligation or expressed protein ligation [4-7]. In general, those methods utilize chemical synthesis of free Sec, chemical synthesis of peptides containing the Sec residue and, finally, ligation of the selenopeptide with a target protein. The reactivity of free Sec makes these methods chemically demanding and they require many steps in the synthesis of Sec-containing polypeptides. Nonetheless, it is likely that both protein ligation methods and recombinant selenoprotein production in E. coli (described next) will become useful techniques for the production of selenoproteins, the choice of which will be dictated by the requirements of each specific application.

Overproducing ducing systems

Natural sources |

Chemical synthesis of selenoproteins

Purification of native selenoproteins

^

Recombinant synthetic selenoprotein production

• /

Radical trapping in EPR studies

Labeling w. selenium radionuclides

Directed folding bydiseienide formation NMR studies

Sec-targeted fluorescence labeling Sec-targeted labeling w. §-emitters

PET studies

Phase determination in X-ray crystallography Sec as source for dehydroalanlne for DHA-targeting with nucleophlllc agents Sec-targeted chemical ligation with electrophlles

Figure 1. Strategies to obtain selenoproteins (A-C) and the different possible Sec-based applications discussed in this chapter [1-11].

Biotechnology of selenocysteine

223

In 1992, it was shown for the first time that E. coli had the capacity to overexpress recombinant selenoproteins in a study involving bacterial formate dehydrogenase H [8]. However, the direct expression of heterologous recombinant proteins in E. coli was not considered possible, due to the fact that most selenoprotein genes would not be compatible with the bacterial Sec incorporation machinery (see Chapter 2 on prokaryotic selenoprotein synthesis machinery). One possibility to circumvent this problem for incorporating Sec into a recombinant protein expressed in E. coli is to use a cysteine (Cys) auxotrophic strain, which allows substitution of Cys with Sec by adding a selenium source to the sulfur-deficient growth medium [9]. This method is, however, not suitable if the protein of interest contains several Cys residues and the aim is to substitute only one residue. In 1999, we succeeded in the site-specific incorporation of Sec into mammalian selenoprotein thioredoxin reductase (TrxR) in high yields in E. coli (Step C, Figure 1) [10]. The open reading frame of rat TrxR was fused with an engineered variant of the bacterial SECIS element, enabling targeted co-translational Sec-incorporation using the bacterial selenoprotein synthesis machinery. This was possible to achieve with a maintained TrxR amino acid sequence due to a penultimate position of the Sec residue in TrxR. That allowed engineering of a functional SelB binding motif of the SECIS element to be positioned outside of the open reading fi-ame, thereby not interfering with the encoded amino acid sequence of the expressed gene. To achieve this a termination codon (TAA, or UAA in the mRNA) was inserted in the lower stem of the SECIS element downstream of the Sec-encoding TGA [10]. That approach proved to be a successful strategy and furthermore showed that the entire bacterial SECIS element did not need to be translated to be functional. We also showed that co-overexpression of the selA, selB, and selC genes, which are key components of the bacterial Sec-incorporation machinery, resulted in even higher selenoprotein yields. We achieved a production yield of 20 mg TrxR per liter of bacterial culture having approximately 25% of the native enzyme activity [80]. We have later shown that the specific activity of the recombinantly produced TrxR directly correlated to the ratio of full-length and UGA-truncated proteins [11]. The efficiency of Sec insertion in E. coli is known to be dependent upon several factors. Because the SelB elongation factor must form a quaternary complex with GTP, selenocystenyl-tRNA^^*^ and the SECIS-element, the stoichiometry between these factors is of major importance [12]. Furthermore, there is a competition between the SelB elongation factor and the bacterial release factor 2 (RF2, the prfB gene product) terminating translation at UGA codons [13]. By assessing different production conditions for the synthesis of mammalian TrxR in E. coli, we found that when expression was induced in late exponential phase, an unexpectedly large upregulation of the Sec incorporation efficiency increased total yield to 40 mg TrxR per liter of

224

Selenium: Its molecular biology and role in human health

bacterial culture and to about 50% Sec content [11]. We suggested that this may have been due to a more efficient SelB function in comparison to RF2 in eariy stationary phase, but this proposition has not yet been experimentally verified. The Sec-containing full-length TrxR can easily be obtained in a pure form by a single-step purification over phenylarsine oxide (PAO) sepharose [1], with a final yield of about 15-20 mg selenoprotein obtained from a liter of bacterial culture. Production of recombinant selenoproteins in E. coli that carry an internal Sec residue requires engineering of a bacterial-type SECIS element within the open reading frame of the recombinant selenoprotein gene. In most cases, this necessitates introduction of point mutations in the protein to have the mRNA compatible with a functional SECIS element. In essence, the sequence of four to seven amino acid residues starting four positions downstream of the Sec must be restricted in such approach for selenoprotein production [10,14]. In spite of these limitations, the strategy has been used for production of a GPx variant [15], a methionine sulfoxide reductase B [16,17] and a Sec-containing glutathione S-transferase [18]. The results firom these studies show, in spite of the limitations of the technique, that recombinant production of selenoproteins with internal Sec-residues in E. coli is technically possible and may indeed become useful for certain applications. Recently an intriguing possibiUty was presented by Gladyshev and coworkers, showing that a distant SECIS element present in the 3'untranslated region can guide Sec-incorporation in bacteria as long as the SECIS element is structurally close to the UGA codon, albeit the efficiency was very low [19]. For further details on the general principle of expressing recombinant selenoproteins in E. coli see earlier reviews on the subject [14,20]. Applications based upon radiolabeling or detection of the selenium atom The Sec-residue can be utilized for introduction of selenium radionuclides, resulting in residue-specific biologically controlled radiolabeling of selenoproteins (Appl. 1, Figure 1). For recombinant selenoproteins expressed in E. coli this is achieved simply by adding the radionuclide to the bacterial growth medium. Provided that excess Cys is added to block nonspecific incorporation into Cys or Met residues, this results in a highly residuespecific labeling of the Sec moiety [21]. The selenium radionuclide most commonly used for this purpose is the gamma emitter ^^Se, which is currently commercially available in the form of [^^Se]selenite with high specific radioactivity (~2500 mCi/g) from the Research Reactor Center, University of Missouri-Columbia, USA. The ^^Se isotope has a high energy (0.86 MeV) and a half-life of 120 days and it is suitable for analysis by autoradiography, liquid scintillation or detection with gamma counters. Consequently, selenoproteins with ^^Se-labeled Sec can serve the basis for

Biotechnology of selenocysteine

225

many different applications in basic science based upon detection of radioactive proteins, such as metabolic tracking and turnover studies or as radiolabeled antigens in Radioimmuno Assays. Detection of incorporated ^^Se also constitutes a useful validation method for demonstration of successful expression of recombinant selenoproteins mE. coli [10,16,18]. Another selenium radionuclide with potential biomedical importance is the positron emitter ^^Se (ti/2=7.1 hour), which can be produced in good yields [22] and which has been postulated to be suitable for use in positron emission tomography (PET) studies in humans [23]. PET is a non-invasive clinical method for detection of trace amounts of compounds labeled with positron emitters, which can be used to localize and quantify positron decays over time and thereby enables studies of biochemical and physiological processes in real-time in humans. The clinically used positron emitters exhibit short half-times and thus it is important to have fast labeling techniques of the ligand to be employed. For production of synthetic selenoproteins labeled with ^^Se, recombinant production in E. coli would likely be too slow. Possibly chemical synthesis of ^^Se-labeled selenoprotein ligands could be developed. A different approach to produce positron emitting protein ligands for PET is, however, to use the reactivity of a Sec residue for a Sec-targeted reaction with electrophilic agents containing positron emitting radionuclides. There are also naturally occurring stable selenium isotopes of biotechnological importance, e.g., the ^^Se isotope having a nuclear spin of 1/2, which makes it possible to be used for high-resolution NMR spectroscopy (Appl. 9, Figure 1) (reviewed in [24]). ''^Se has therefore been introduced into selenomethionine (SeMet) residues and subsequently used for NMR determinations of proteins expressed in E. coli [25]. A ^^Se-GPX could also be enriched from lamb having been fed a 5 months diet with ^^Se and the protein was then used for NMR studies [26]. Recently, synthetic methods to produce L-[^^Se]-Sec have been developed, which could expand the further use of ^^Se-labeled selenoproteins for NMR determinations [27]. The selenium atom can also be utilized in X-ray crystallography (Appl. 8, Figure 1), and SeMet introduction into proteins for multiwavelength anomalous diffraction (MAD) to solve the phasing problem is a well established method [28]. The development of methods for introducing Sec residues into proteins has recently been used for double labeling techniques infroducing both SeMet and Sec [29], or Sec alone [30], for further facilitated phase determinations. The different methods to produce synthetic selenoproteins as discussed above should hence be possible to use as an aid for protein structure determinations with X-ray crystallography.

226

Selenium: Its molecular biology and role in human health

Applications based on tlie cliemical reactivity of Sec A Sec residue exhibits quite different properties compared to its sulfur containing analogue, Cys [31]. The most obvious difference is that the pKa for Sec is much lower than for Cys (5.2 vs. 8.3) [32]. Consequently, at physiological pH the selenol of Sec is mainly in its anionic selenolate form, while the thiol of a Cys residue is typically protonated, making Sec significantly more reactive than Cys. Sec is also a stronger nucleophile than Cys. The majority of the hitherto characterized selenoproteins are enzymes, where the Sec residue is essential for the catalytic activity since its reactivity is generally employed in the catalysis. An illustration of the qualitative differences between Sec-containing oxidoreductases and their Cyscontaining counterparts is the significantly lower activity of the latter [33,34], although changes in an active site microenvironment may activate Cys residues, in certain cases, to approach the reactivity of Sec [35]. The high reactivity of Sec can form the basis for numerous technological applications. By studying synthetic Cys- or Sec-containing peptides, the redox potentials of the Sec-Sec, Sec-Cys and Cys-Cys couples have been determined at pH 7 to -381 mV, -326 mV and -180 mV, respectively [36]. These redox potentials result in a preferential diselenide formation if two Sec residues are present among additional cysteines and it was therefore postulated that a pair of Sec residues could be introduced in place of two Cys residues in a protein, to allow for targeted diselenide generation and thereby directed correct folding of a protein [36] (Appl. 10, Figure 1). This has indeed been demonstrated with the synthetically produced endothelin-1 peptide where the introduction of two Sec-residues in place of Cys directed the correct oxidative folding of the peptide, although it contained two additional Cys residues [37]. Alternatively, for studies of folding intermediates or transition states during catalysis of redox reactions, introduction of a single Sec residue at specific sites in proteins or enzymes could be utilized for the trapping of otherwise transient disulfides or thiolate-targeted intermediates. Electron paramagnetic resonance (EPR) spin trapping [38] is a technique for the direct detection of radical species and could be used to detect formation of selenenyl radical formation in a selenoprotein (Appl. 11, Figure 1). By introducing a Sec residue into an enzyme, it may perhaps be possible to change the specificity or function of the enzyme, thus constituting a potential for a directed evolution of enzymatic activity (Appl. 5, Figure 1). This has indeed been demonstrated by the production of the artificial selenoenzyme selenosubtilisin, where the change of an active site Ser residue to a Sec residue in the serine protease subtilisin, resulted in conversion of the enzyme to a glutathione peroxidase mimic with peroxidase activity [39,40]. A similar approach has been reported in a study that changed a Ser to a Sec in a monoclonal antibody resulting in GPX activity [41,42], although in that case

Biotechnology of selenocysteine

227

it is not clear why the targeted Ser residue should have been extraordinarily active as compared to other serine residues in the protein. A Cys residue in the active site of glyceraldehyde 3-phosphate dehydrogenase (GAPDH), when changed to a Sec residue, also revealed a novel peroxidase activity [43]. It is clear that not all cases of Sec introduction into the active sites of enzymes must yield peroxidase activity, e.g., as shown in the production of a Sec-containing variant of GST [18]. However, another GST isoform that was substituted with Sec indeed seemed to possess novel peroxidase activity [44]. Considering the different types of reactions catalyzed by native selenoproteins it is possible that production of synthetic selenoproteins may also be utilized for engineering of other types of reaction catalysts than those supporting peroxidase reactions, such as halogen transfer reactions (cf, the thyroid hormone deodinases). It was reported that the selenium atom can be lost from Sec during purification of selenoproteins, forming dehydroalanine [45]. This type of conversion can also be catalyzed deliberately by a chemoselective oxidation of Sec. The thereby formed electrophilic dehydroalanine moiety can then be used as precursor for peptide conjugation reactions with nucleophilic ligands (Appl.7,Figurel)[4]. Conjugation-based applications The nucleophilic properties of a Sec residue can be utilized for selective selenolate-targeting using electrophilic compounds (Appl. 2,3,6, Figure 1). For instance, there are many commercially available electrophilic thiolate-, and thereby selenolate-reactive probes, which can be used for fluorescence labeling preferentially at a Sec residue. A Sec-specific reaction that avoids targeting of less reactive Cys residues becomes possible when using short reaction times and low pH (described in more detail below in the discussion of Sel-tagged based applications). Analogously, it is possible to use electrophilic iodoacetamido-biotin reagents for specific Sec-labeling and thereby subsequent use of anti-biotin antibodies for detection [46]. We also recently demonstrated the use of electrophilic compounds containing short-lived positron emitters, such as "C, for labeling of Sec residues [1]. This is promising for use in the field of PET studies and is discussed further in the next section describing applications based upon a Sel-tag. Applications based on a Sel-tag We have recently developed a Sel-tag, which is a Sec-containing multifunctional tetrapeptide motif for recombinant proteins expressed in E. coli [1]. By fusing the C-terminal tetrapeptide of TrxR, -Gly-Cys-Sec-Gly, as a Sel-tag for non-selenoproteins we could take advantage of selenium biochemistry, which could, in theory, be utilized for all Sec-based

228

Selenium: Its molecular biology and role in human health

applications discussed above. In addition, this motif, corresponding to the four last amino acids of the natural mammalian selenoprotein TrxR [47-49], is redox active. When reduced, the Sec residue becomes easily targeted by electrophilic compounds, but when oxidized it forms a selenenylsulfide with its neighboring Cys residue and thereby becomes essentially inert to alkylating agents [50]. Thus, due to presence of the Cys residue in the Seltag, the normally highly reactive selenium atom of Sec could be protected in the oxidized state of the protein as a result of the selenenylsulfide bond formed between the Sec and Cys residues within the Sel-tag. The selenolthiol motif, when reduced, e.g., by DTT, may furthermore serve the basis for purification of Sel-tagged proteins using a one-step purification technique with PAO sepharose. The PAO-sepharose had previously been utilized for purification of proteins containing vicinal dithiols [51-53]. We found that the affinity of a selenolthiol to PAO was much stronger than that of a dithiol and proteins bound to PAO sepharose through dithiols can thereby be eluted using DTT while the selenolthiol-containing Sel-tagged protein can not. The highly specific arsine oxide chelator dimercaptopropanol sulfonic acid (DMPS) could be used to elute the selenolthiol containing proteins [1]. In fact, for several Sel-tagged proteins, we have found the PAO sepharose purification procedure to be essentially equivalent in both yield and specificity to the commonly used purification of His-tagged proteins over Nickel columns. The Sel-tag has already been utilized for a number of biotechnological Sec-based applications. We have labeled a Sel-tagged mite allergen, Der p 2, with the gamma-emitting isotope ^^Se and utilized it for an in vivo study, tracking the ^^Se-labeled allergen in a mouse model for allergy [54]. We have also demonstrated that a selective selenolate-targeting of the Sec residue in Sel-tagged Der p 2 can be achieved with an electrophilic fluorescent probe for fluorescence labeling, in spite of the presence of six additional Cys residues in the protein [1]. This could be accomplished by incubating the reduced Sel-tagged protein for a short duration with the electrophilic fluorescent compound at a low pH in the presence of excess DTT, thereby scavenging labeling of Cys residues while allowing the highly reactive selenolate to become labeled. In the same study we showed that the Seltagged human vasoactive intestinal peptide (VIP) can be labeled with a fluorescent compound without interfering with its ability to bind to its native VIP-receptor[l]. We reasoned that a similar principle as for residue-specific fluorescent labeling could be used to label a Sec residue with electrophilic compounds containing short-lived positron emitters. This was indeed possible utilizing ["C]-methyl iodide, which gave an efficient "C-labeling of Sel-tagged proteins [1]. To label proteins or peptides with short-lived isotopes is generally a difficult task and this particular Sel-tag application could thereby

Biotechnology of selenocysteine

229

become highly useful as a general method for generating radiolabeled protein ligands suitable for PET imaging studies. Concluding remarks The recent development of efficient production methods for synthesis of selenoproteins has opened new possibilities to use these proteins for Secbased biotechnological applications. As outlined in this chapter, such applications employ the unique features of Sec, the only natural seleniumcontaining amino acid. The biotechnological potential of these techniques is vast and we trust that the biomedical field involving biotechnology of Sec will expand rapidly in the near future. References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32.

L Johansson, C Chen, J Thorell, A Fredriksson, S Stone-Elander, G Gafvelin, ESJ Amer 2004 Nature Methods 1 ;61 MJ Berry, GW Martin 3rd, R Tujebajeva et al 2002 Methods Enzymol 347:17 A Mehta, CM Rebsch, SA Kinzy, JE Fletcher, PR Copeland 2004 J Biol Chem 279:37852 MD Gieselman, Y Zhu, H Zhou, D Galenic, WA van der Donk 2002 Chembiochem 3:709 RJ Hondal, RT Raines 2002 Methods Enzymol ZAl-.lQ L Moroder 2005 J Pept Sci 11:187 R Quaderer, A Sewing, D Hilvert 2001 Helvetica ChimicaActa 84:1197 GT Chen, MJ Axley, J Hacia, M Inouye 1992 Mol Microbiol 6:781 S Muller, H Senn, B Gsell, W Vetter, C Baron, A Bock 1994 Biochemistry 33:3404 ESJ Am6r, H Sarioglu, F Lottspeich, A Holmgren, A B6ck 1999 J Mol Biol 292:1003 O Rengby, L Johansson, LA Carlson et al 2004 Appl Environ Microbiol 70:5159 P Tormay, A Sawers, A Bock 1996 Mol Microbiol 21:1253 JB Mansell, D Guevremont, ES Poole, WP Tate 2001 EMBO J20:7284 ESJ Amer 2002 Methods Enzymol 347:226 S Hazebrouck, L Camoin, Z Faltin, AD Strosberg, Y Eshdat 2000 J Biol Chem 275: 28715 S Bar-Noy, J Moskovitz 2002 Biochem Biophys Res Commun 297:956 HY Kim, VN Gladyshev 2004 Mol Biol Cell 15:1055 Z Jiang, ESJ Amer, Y Mu et al 2004 Biochem Biophys Res Commun 321:94 D Su, Y Li, VN Gladyshev 2005 Nucleic Acids Res 33:2486 L Johansson, G Gafvelin, ESJ Amer 2005 Biochim Biophys Acta, in press S Muller, J Heider, A B6ck 1997 Arch Microbiol 168:421 M Fassbender, D de Villiers, M Nortier, N van der Walt 2001 Appl Radiat hot 54:905 R Bergmann, P Brust, G Kampf, HH Coenen, G Stocklin 1995 Nucl Med Biol 22:475 H Duddeck 1995 Progress in NMR Spectroscopy 27:1 JO Boles, WH ToUeson, JC Schmidt et al 1992 J Biol Chem 267:22217 P Gettins, BC Crews 1991 J Biol Chem 266:4804 E Stocking, J Schwarz, H Senn, M Salzmann, L Silks 1997 J Chem Soc Perkin Trans 1:2442 WA Hendrickson, JR Horton, DM LeMaster 1990 EMBO J 9:1665 MP Strub, F Hoh, JF Sanchez, JM Strub, A Bock, A Aumelas, C Dumas 2003 Structure (Camb) \ 1:1359 JF Sanchez, F Hoh, MP Strub, A Aumelas, C Dumas 2002 Structure 10:1363 C Jacob, GI Giles, NM Giles, H Sies 2003 Angew Chem Int Ed Engl 42:4742 RE Huber, RS Criddle 1967 Arch Biochem Biophys 122:164

230 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54.

Selenium: Its molecular biology and role in human health S Bar-Noy, SN Gorlatov, TC Stadtman 2001 Free Radio Biol Med 30:51 L Zhong, A Holmgren 2000 J Biol Chem 215:\S121 S Gromer, L Johansson, H Bauer et al 2003 Proc Natl Acad Sci USA 100:12618 D Besse, N Budisa, W Kambrock et al 1997 Biol Chem 378:211 S Pegoraro, S Fiori, S Rudolph-Bohner, TX Watanabe, L Moroder 1992, J Mol Biol 284:779 MJ Davies, CL Hawkins 2004 Free Radic Biol Med 36:1072 IM Bell, ML Fisher, ZP Wu, D Hilvert 1993 Biochemistry 32:3754 ZP Wu, D Hilvert 1989 J Am Chem Soc Wl :4513 L Ding, Z Liu, Z Zhu, G Luo, D Zhao, J Ni 1998 Biochem J 332 (Pt 1):251 GM Luo, ZQ Zhu, L Ding, G Gao, QA Sun, Z Liu, TS Yang, JC Shen 1994 Biochem BiophysRes Commun 198:1240 S Boschi-Muller, S Muller, A Van Dorsselaer, A Bock, G Branlant 1998 FEBS Lett 439:241 HJ Yu, JQ Liu, A Bock, J Li, GM Luo, JC Shen 2005 J Biol Chem 280:11930 S Ma, RM Caprioli, KE Hill, RF Burk 2003 J Am Soc Mass Spectrom 14:593 KE Sandman, CJ Noren 2000 Nucleic Acids Res 28:755 VN Gladyshev, KT Jeang, TC Stadtman 1996 Proc Natl Acad Sci USA 93:6146 L Zhong, ESJ Amer, A Holmgren 2000 Proc Natl Acad Sci USA 97:5854 L Zhong, ESJ Am6r, J Ljung, F Aslund, A Holmgren 1998 J Biol Chem 273:8581 J Nordberg, L Zhong, A Holmgren, ESJ Am6r 1998 J Biol Chem 273:10835 RD Hoffman, MD Lane \992 J Biol Chem 267:14005 E Kalef, PG Walfish, C Gitler 1993 Anal Biochem 212:325 GY Zhou, M Jauhiainen, K Stevenson, PJ Dolphin 1991 JChromatogr 568:69 L Johansson, L Svensson, U Bergstr5m et al 2005 FEBS J 272:3449

Part III

Selenium and human health

Chapter 21. Selenium, selenoproteins and brain function Ulrich Schweizer Neurobiology of Selenium, Neuroscience Research Center and Institute for Experimental Endocrinology, Charite-Universitatsmedizin Berlin, Charite Campus Mitte, D-WJJ7 Berlin, Germany

Lutz Schomburg Institute for Experimental Endocrinology, Charite-Universitatsmedizin, Berlin, Charite Campus Mitte, D-10JJ7 Berlin, Germany

Summary: After the discovery of selenium (Se) as an essential trace element, direct evidence that Se plays a role in brain function remained relatively scarce for many years. This was probably due to the remarkable stability of brain Se levels during times of dietary Se restriction in experimental animals. In these experiments, activities of the first known Sedependent enzymes, e.g. glutathione peroxidase (GPX), thioredoxin reductase (TrxR), and deiodinase (Dio), were also little changed in the brains of rodents fed Se-deficient diets for extended periods of time. Thus, the lack of spontaneous neurological deficits seemed to exclude an important role for Se in brain function. This notion remained largely unchanged despite the purification of selenoprotein P (SePP) from serum as a neurotrophic factor for cultured neurons and the finding that selenite is an essential component of media for in vitro culture of central neurons. Only later experiments revealed that Se-deficiency exacerbated the outcome of neurological disease in certain animal models. Oxidative stress is considered to play a role in neurodegenerative processes, and GPX 1-transgenic mice provided the first molecular proof for an involvement of selenoproteins in such conditions. Then, gene targeting of SePP led to clear-cut spontaneous neurological deficits in Se-deficient animals and placed SePP at center stage for the priviledged Se supply to the brain. Whether impaired expression of selenoproteins in human brain contributes to the incidence or severity of neurodegenerative disease remains to be established. Still, available evidence already suggests that selenoproteins are playing important roles for brain development, function, and disease in mice - and also most likely in humans.

234

Selenium: Its molecular biology and role in human health

Introduction In 1957, Schwarz and Foltz identified Se as an essential trace element for rats. When maintained on a Se-deficient diet, animals developed liver necrosis, but could be rescued with a preparation called "factor 3" which was shown to contain Se [1]. Although subsequent studies revealed that the initial Se-deficient diet was also vitamin E-deficient, Se was now recognized as an essential trace element and no longer simply regarded as a potential environmental toxin. Unlike in most other organs, brain Se levels remained quite stable during dietary Se restriction [2-4]. Only one report demonstrated spontaneous neurological symptoms in Se-deficient mammals, i.e. "leg crossing", in Balb/c mice maintained on a Se-deficient diet [5], a finding which may be strain-specific since similar observations were not reported in other strains of mice or rats. Oxidative stress, i.e. a disproportionate increase of reactive oxygen species leading to the oxidation of cellular constituents like proteins, DNA, and lipids, is thought to contribute to the cellular damage during excitotoxicity and pathogenesis of neurodegenerative disorders [6-8]. Given the reactions catalyzed by known selenoenzymes, it is conceivable that selenoproteins modulate the outcome of neurological disease in animal models. GPxl degrades hydrogen peroxide, the product of superoxide dismutase (SOD). GPx4 degrades phospholipid hydroperoxides thereby potentially protecting cellular membranes fi-om oxidative damage. Methionine sulfoxides are formed fi-om protein-bound methionine during oxidative stress and can be reduced by methionine sulfoxide reductase (Msr) A and MsrB; the latter being also known as selenoprotein R. Mammalian TrxR accepts a wide range of substrates including hydrogen peroxide modulated redox-sensitive transcription factors. In fact, in animal models of neurodegenerative disease Se-deficiency generally exacerbated the neurological and histological damage and a simple Se-supplementation most often proved beneficial. Stroke During stroke or hypoxia/ischemia-reperfusion (HI), a dramatic increase in reactive oxygen species occurs that is believed to trigger molecular events culminating in increased apoptosis, necrosis, and neuroinflammation that may further increase neuronal cell loss and subsequently lead to memory impairment and motor incoordination [8-11]. Since stroke is among the leading causes of disabilities in aging Western societies, efficient treatments are urgently needed. Superoxide radicals (O2") liberated during HI are degraded by SOD. The role of SOD in tissue protection is clearly illustrated after HI by increased cell death in mice with reduced SODl activity [12,13] and in mice protected from HI by over-expression of SODl [14]. During catalysis, SOD consumes O2" and produces the less reactive hydrogen peroxide (H2O2). However, in the presence of Fe^^, H2O2 may decompose

Selenium, selenoproteins, and brain function

235

forming the highly reactive hydroxyl radical (Fenton reaction). Thus, GPX enzymes represent the second line of defense against reactive oxygen species. Based on its ubiquitous expression and ease of assay, GPXl is the best-studied member of the family of Se-dependent GPX isozymes. Neurons and astrocytes express GPXl, but most cytosolic GPX activity is localized in astrocytes [15,16]. In addition, an immunohistochemical study revealed GPX induction in astrocytes surrounding the infarcted area [17]. As expected, transgenic mice over-expressing GPXl are protected against HI in the brain. After HI, indices of cell death are reduced in GPXl-transgenic (tg) animals and infarct volume is significantly reduced. Accordingly, neurological deficits are mitigated in GPXl-tg animals as compared to wild-type controls [18,19]. In contrast, mice made genetically deficient for GPXl exhibit a drastically increased infarct volume, increased neurological deficits, and more pronounced neuronal cell death [20]. Similar observations have been made in myocardial HI models (reviewed in [21]). Thus, GPXl definitely acts as a protective enzyme during HI in vivo. In vitro. Furling et al. have shown that GPXl over-expression protects hippocampal slices from transient hypoxia and maintains electrophysiological properties, e.g., LTP induction [22]. These findings therefore underline the detrimental role played by H2O2 during HI and suggest that anti-oxidative therapy might represent a rational treatment for stroke. It should be stressed that cellular damage after stroke occurs not only in the immediate center of the infarcted area due to shortage of oxygen and energy substrates, but the damaged area grows for several days - even after the thrombus has been dissolved. The underlying mechanisms and contributions of vascular and immune cells, microglia, astrocytes, neurons and oligodendrocytes are not within the scope of this chapter, but the temporal delay of cellular demise has been regarded as an open window of opportunity in which anti-oxidative treatment may still be beneficial. It is known that GPX activity is induced while serum Se falls after stroke [23]. Interestingly, among men free of stroke at the outset, low serum Se was associated significantly with stroke mortality [24], but a larger study could not support this finding [25]. In some cases of rare familial childhood stroke, plasma GPX activity was reportedly only 50% in affected children as compared to their siblings. Interestingly, plasma GPX activity correlated inversely with platelet P-selectin expression and NO-induced aggregation in vitro suggesting a mechanistic link between thrombotic stroke and plasma GPX [26]. In animal models of HI, Se supplementation appeared protective [27,28]. The organoselenium compound ebselen exhibits weak GPX activity and also interacts with and possibly potentiates the Trx-TrxR system. The drug has therefore been tested in preclinical and clinical trials for the treatment of stroke. However, despite early encouraging results in rodent models, ebselen has not met these expectations in clinical trials [29].

236

Selenium: Its molecular biology and role in human health

Unfortunately, less is known on the role of GPX4, the phospholipid hydroperoxide-specifc GPX, in HI in the brain. But cardiomyocytes infected with GPX4-expressing adenovirus showed protection of lipids, electron transport chain complex IV function and cellular damage after HI [30]. The Trx-TrxR system is likely involved in protection against HI. Overexpression of Trx in mice protected them against focal cerebral ischemia [31]. Recently, conditional knock-out mice for TrxRl and TrxR2 have been established [32,33] and it will be interesting to define the roles of these enzymes during HI. Interestingly, there seem to exist functional differences among the individual isoenzymes in vivo, since targeted disruption of TrxR2, but not of TrxRl, in the heart leads to dilated cardiomyopathy, a condition in which oxidative stress has also been implicated. Parkinson's Disease Oxidative stress is also believed to be involved in the pathogenesis of sporadic Parkinson's disease (PD) [7]. In this neurodegenerative disorder dopaminergic neurons located in the substantia nigra perish and thus the striatum is deafferented fi-om its main dopaminergic input. The resulting progressive dysfunction of the basal ganglia leads to a characteristic movement disorder which can pharmacologically be treated for some time, but ultimately has devastating consequences. While in some rare cases of familial PD the underlying gene defects have been identified, the cause of idiopathic PD which constitutes the vast majority of cases is still unknown. Some investigators suggest that endogenous or exogenous neurotoxins may contribute to nigral cell death. For example, hydrogen peroxide can be formed fi"om dopamine by the action of monoamine oxidase or by autoxidation of dopamine [34]. Signs of increased oxidative stress have been reported in substantia nigra fi'om PD brains, but due to the delayed tissue processing normally occuring after death of a patient, some of the oxidative damage may have taken place after death complicating interpretation of the results. However, consistent with a crucial role for oxidative stress, an increased amount of iron in substantia nigra was demonstrated in PD [35,36]. Moreover, modulation of iron availability in animal models of PD attenuated the susceptibility of nigral neurons to cell death [37] suggesting that the Fenton reaction contributes to dopaminergic neuron degeneration. Thus, a reduction of H2O2 levels by selenoenzymes might help reduce nigral cell demise. Several pharmacological models for PD in rodents and primates have been established, among them methamphetamine- (MA), 6-hydroxydopamine- (6-OHDA), methyl-phenyl-tetrahydropyridine- (MPTP) and diiminoproprionitrile (DIPN)-induced dopaminergic neurodegeneration. Cellular lesions in 6-OHDA, MA, and MPTP models of PD are increased in Se-deficient rats and can be reduced by Se administration [38,39]. Pretreatment of mice with Se in the DIPN model resulted in reduced lipid

Selenium, selenoproteins, and brain function

237

peroxidation products and inhibited neurobehavioural alterations in a dosedependent manner [40]. Striatal dopamine depletion was prevented by Se supplementation in the MA model [41]. Induction of Se-dependent enzymes like GPXl via Se administration dose-dependently attenuated neurodegeneration in 6-OHDA treated rats [42] and MA-treated rats and mice [39,43]. These Se-dependent protective effects are likely, at least partially, mediated via modulation of GPXl activity, because mice that over-express GPXl are protected against 6-OHDA induced nigral degeneration [44]. In addition, although brain Se levels are only mildly changed by Se-deficient diet in wild-type rodents, it has been observed that GPXl activity, a sensitive marker for Se status in other systems, is changed during Se-deficiency in the brain including substantia nigra and striatum [38,45]. In the MPTP model, GPXl-knockout mice showed increased sensitivity [46] providing further evidence for a protective role of GPXl in PD-like neurodegeneration - and for the involvement of H2O2 in this process. Thus, results from animal models for PD suggest that Se and Se-dependent enzymes, likely GPXl, protect nigral neurons. It should be noted that two conceptually different situations have been discussed: Decreased GPXl activity (or decreased Se levels) increase the susceptibility against neurotoxins, while transgenic overexpression of GPXl protects against neurodegeneration even in the presence of normal Se levels. Simply increasing dietary Se levels over the recommended dietary requirement was not reported to afford protection. Several studies aimed to correlate serum Se levels or blood cell GPX activity in patients with PD, but the results were inconsistent [47-50]. Moreover, Se determination in cerebrospinal fluid did not reveal differences between PD patients and controls [47,51]. It should be noted, however, that there is no strict correlation between blood and brain Se levels (see below). Determination of Se protein expression in human brain samples using immunohistochemical or activity assays, therefore, should be more informative. Damier et al reported increased GPX immunoreactivity in nigral astrocytes in PD [52], but direct measurement of enzymic activity in one study [53] revealed a 20% reduction in several brain regions from PD patients including substantia nigra, basal ganglia, and cerebral cortex, while in another study [48] no changes were observed. Among the many biochemical measures of oxidative stress in PD brain samples, loss of GSH is the most reproducible [7]. Interestingly, the amount of GSSG is not considerably changed and it is unknown how and where GSH is lost. Thus, while the GPX/GSH system proved particularly protective in models of PD, and oxidative stress is likely involved in PD, there is no clear data whether this system is disturbed in the brains of PD patients and whether or not observed changes in GPX and GSH levels are the cause or consequence of the disease process [7].

238

Selenium: Its molecular biology and role in human health

Epilepsy Induction of oxidative stress by local application of Fe^^ to the rodent cortex leads to epileptiform EEG discharges and was used as a model for epilepsy [54]. Using this model, Rubin and Willmore showed that antiperoxidant treatment, including selenite, reduced the histopathological tissue damage as well as epileptic discharges. [55,56]. Conversely, reduced dietary Se intake increased seizure frequency in rats treated with the excitotoxin kainic acid and increased hippocampal cell loss [4]. Thus, it appears as if Se status affects the response to experimentally induced seizures although it has never been shown that Se treatment increased brain Se directly. Rather, some investigators found that brain Se levels are stable during dietary Se depletion, although a reduction of brain GPXl activity has been noted [4,39,45]. While more studies are needed to clarify this issue, it should be noted that seizures and muscle weakness have been observed in patients that had become severely Se-deficient during total parenteral nutrition [57,58] at a time when Se was not sufficiently included into dietary formulations. Probably the best evidence to link Se with epilepsy is given by two unrelated clinical reports. In the first study, Weber [59] showed that children with a form of intractable seizures had low blood Se and low plasma GPx activity. Dietary Se supplementation led to normalization of plasma Se and seizure activity was subsequently reduced. The study was initially not well received [60]. Several years later an unrelated study reported similar findings with children suffering from infractable seizures [61]. Again, the patients manifested low plasma Se and low plasma GPx activity. Dietary supplementation with Se compounds normalized plasma Se levels and the seizures subsequently responded to anticonvulsant therapy. Most importantly, upon cessation of Se freatment in one patient, seizures resumed but responded again to Se supplementation. Unfortunately, when the patient was again referred to another hospital, Se freatment was interrupted and the patient died in status epilepticus. Some of the patients described in the two reports above suffered from classical Se-deficiency symptoms like brittle hair, white nails, weak muscle, and in one case the knee joint exhibited morphological changes reminiscent of Kashin-Beck osteopathy. At present it is not clear whether Se deficiency was the primary reason for the seizures or, probably more likely, Se deficiency mediated the refractoriness of the seizures towards anticonvulsant therapy. In addition, the cause for the observed Se deficiency is not known. We hypothesised that mutations in genes involved in Se metabolism may underlie an impaired capacity to accumulate dietary Se or to retain Se in the body. Respective novel poljmiorphisms are currently under investigation (U. Schweizer, unpublished). Along these lines it should be noted that mice made genetically deficient for SePP display spontaneous seizures and a movement phenotype that depends on dietary Se availability [62-65] (see below and Chapter 10).

Selenium, selenoproteins, and brain function

239

Since Se-deficiency rendered rats more susceptible to the neurotoxin kainic acid [4], it was expected that transgenic over-expression of GPXl in mice protects them from seizures induced by kainic acid injection. Surprisingly it was found that GPXl over-expression increased hippocampal cell death and seizure intensity in transgenic animals [66]. The authors showed that increased GPXl activity led to increased extracellular GSSG levels after kainic acid treatment and speculated that this may activate the NMDA receptor. This explanation is conceivable as GSSG application is known to alter NMDA-R responses. GPX activity is increased in pilocarpine-treated rats before status epilepticus [67] and GSH/GSSG decreased in seizure prone mice [68], but it remains to be determined whether this is a protective adaptation or cause of the seizure. Clearly, this finding implicates that even generally protective enzyme systems need to be well-balanced and finelytuned into the physiological context. It should be noted that epilepsy is a frequent disorder probably afflicting almost 1 in 100 persons during their lifetime. The diverse manifestations of epilepsy are reflected by a plethora of possible causes as far as they have been identified. We do not believe that Se deficiency is a major cause of epilepsy in humans. However, there seems to exist a subset of Se-deficient patients who may respond favorably to Se administration. Other neurodegenerative disorders As in many other neurodegenerative disorders, oxidative stress has been implicated in the pathology of amyofrophic lateral sclerosis (ALS), which is known in the US as Lou Gehrig's disease'. Li this invariably fatal condition, spinal and cranial motomeurons, which innervate muscle fibers degenerate leading to weakness of the limbs, loss of speech, and respiratory disfress. In many cases there is also involvement of principal neurons in the primary motor cortex. Similar as in other major neurodegenerative disorders, most cases occur sporadically, but there are also cases of familial ALS (FALS) that helped identify some underlying gene defects in genes encoding SODl, alsin, and the androgen receptor. At first, oxidative stress seemed to cause ALS, since most patients suffering from FALS carry missense mutations in the SODl gene [69]. Accordingly, transgenic mice carrying human SODl mutant genes develop an ALS-like disease. Although questioned from the beginning, the popular notion that a loss-of-function mutation of SODl occurs in FALS has not been substantiated and more recent evidence rather established an unrelated gain-of-function mechanism. Still, markers of oxidative sfress are reportedly elevated in man and mice carrying SODl mutations. Regarding Se or selenoenzymes in ALS, however, we are not

' Famous US sportsman. New York Yankees 1925-1939, died in 1941 from ALS

240

Selenium: Its molecular biology and role in human health

aware of data from animal models. The few clinical studies performed did not show a correlation of Se or plasma GPX with disease. Alzheimer's disease (AD) is a relatively common neurodegenerative disorder that develops at an advanced age. Cortical atrophy affects principal neurons and cholinergic neurons in the cortex and hippocampus. The disease is characterized by memory loss and diminished higher brain fimctions [70]. So called amyloid plaques are a pathological hallmark of the disease. These plaques consist of misfolded/aggregated protein. A fragment, Ap, of APP (amyloid precursor protein) is proteolytically generated and a major constituent of amyloid plaques. Certain alleles of APP, the E4 allele of apolipoprotein E, and the presenilin genes which encode polytopic fransmembrane proteins of the y-secretase complex are associated with AD. Since AD is such a common and regrettably devastating disease, an enormous amount of data have been accumulated regarding AD which cannot be reasonably dealt with here. It should be noted that only very recently a physiological function of APP has been elucidated in detail [71]. With respect to Se or selenoproteins, not many studies have been undertaken and the published ones did not yield a consistent picture. Several authors tested Se levels in AD brain [72-74] or plasma samples [75]. Two studies reported plasma and/or erythrocyte GPX measurements in AD patients [75,76]. Again, the resulting picture was not clear. Given the large variations of plasma Se and plasma GPX activity in human populations, it may not be possible to find clear associations with AD in small studies. We may expect more insight from the PREADVISE study which accompanies the SELECT trial. Li this longitudinal study over 5-12 years, patients are treated with Se and/or vitamin E or placebo. Neurological investigations are directed to uncover any protective effects of Se on the occurence of AD in the study population. The rationale for this study, again, is the notion that oxidative stress may be involved in AD pathogenesis. Interestingly, levels of Trx fall while expression of TrxR increases in the AD brain [77]. It is, however, conceivable that oxidative stress is rather a consequence of AD pathology related to protein misfolding or activation of microglial cells [8] and not the reason - and thus selenoproteins may modulate disease progression without the need to postulate a causal role for Se in AD incidence. Huntington's disease (HD) results from the expansion of a CAG repeat (encoding the amino acid glutamine) in a gene whose inheritance was dominantly linked to HD. The encoded protein, huntingtin, is as yet of unknown function and aggregates in response to enlarged polyglutaminerepeats. Although huntingtin is expressed in many neuronal cell types, mutant huntingtin initially causes a remarkably specific degeneration of small spiny GABAergic neurons in the striatum, but later the patients succumb to a devastating neurodegenerative disease. We have found no clinical data linking HD and Se status or selenoproteins and only one such

Selenium, selenoproteins, and brain function

241

study in a pharmacological animal model. Injection of rats with quinolinic acid leads to excitotoxicity, lipid peroxidation, neuronal death, and neurological symptoms similar to HD. Se supplementation of rats in that model was protective [78]. Friedreich's ataxia (FA) is one of the most common forms of autosomal recessive ataxia [79]. About 97% being due to expansion of a GAA trinucleotide repeat in intron 1 of the FRDA gene which impairs its expression. FA is characterized by degeneration of large sensory neurons and spinocerebellar tracts, cardiomyopathy and increased incidence of diabetes. The disorder is usually manifest before adolescence and is generally characterized by uncoordination of limb movements, dysarthria, and impairment of position and vibratory senses. The FRDA gene encodes the small mitochondrial protein frataxin that is involved in iron import into mitochondria and antioxidant defense [80]. Frataxin deficiency leads to mitochondrial iron overload, respiratory chain imbalance and increased ROS generation leading to oxidative damage. Moreover, a direct involvement of frataxin in ROS detoxification, activation of GPx and elevation of reduced thiols has recently been demonsfrated [81]. Because of the overlap in the antioxidant biochemistry of Se and the increased iron-induced mitochondrial peroxidation, a treatment of FA patients with Se might prove successful, but no data are presently available, yet. There is substantial evidence implicating oxidative stress in the pathophysiology of neurodegenerative disorders and thus it is conceivable that Sedependent enzyme systems may be involved in protection against neurodegeneration. A major problem for clinical studies trying to link neurodegenerative disorders with Se metabolism lies in the recruitment of a sufficiently large homogenous study cohort. Since most cases of neurological disease occur sporadically with unclear etiology and since even in one given disease several underlying gene defects have been identified, it may be hard to find such correlations in small groups of patients with heterogeneous etiology of the disease. While we are still lacking suitable surrogate markers for Se or selenoenzyme levels in the human brain, we are left with post mortem analyses and nutritional and/or transgenic animal models. Transgenic selenoprotein-deficient mouse models with neurological disease Selenoprotein P (SePP) is the major Se containing protein in human and rodent plasma [82] (see Chapter 10). Among selenoproteins it is unique since it contains more than one Sec residue (10 in man and mouse, 12 in cattle, and 17 in zebrafish). SePP plasma levels closely respond to dietary Se intake (see Chapter 35). Because of these and other properties, it was hypothesized from early on that SePP is a Se fransport protein [83].

242

Selenium: Its molecular biology and role in human health

We and the group of R. Burk and K. Hill have independently inactivated the gene encoding SePP in the mouse and shown that SePP-KO mice displayed a neurological phenotype including seizures, a movement disorder, and axonal degeneration [62,63,65,84]. The neurological phenotype of SePPKO mice depends on dietary Se supply and can completely be abrogated if SePP-KO mice are supplemented with sufficient amounts of Se [64,65]. This finding contrasts sharply with the stability of brain Se levels and brain function during dietary Se restriction in rodents [2,4]. Thus, it appears as if, in the absence of SePP, the brain becomes sensitive to circulating Se implicating SePP as a crucial part of the mechanism that maintains preferential supply of Se to the brain [85]. We wondered at which time during development the lack of SePP manifests itself in neurological dysfunction. In oder to study this question, we have supplemented heterozygous SePP breeder pairs at different time points with selenite in their drinking water and found that SePP-KO mice can be rescued if Se treatment started from mating or from birth onwards. When Se supplementation started from weaning, after the phenotype (movement disorder and seizures) had already occurred, the mice were stabilized at the status quo, but the phenotype was not ameliorated (Figure 1).

1

1

noSe

1

phenotype 1

efrrnnn

Se Se Se

^•1

Se malting

strong/stable progressive

bi ith

wea ling

adulthood

Figure 1. Experimental design for selenite supplementation of SePP-KO mice during development. Complete rescue of Se-dependent neurological phenotypes was observed if Se supplementation was initiated early at mating or birth, but was not achieved if started after weaning. Phenotypes developed progressively if supplementation was terminated at any time.

These findings suggest that a Se-dependent developmental process is taking place between birth and weaning that underlies the neurological defects. Since in rodents the cerebellum develops mainly from birth until

Selenium, selenoproteins, and brain function

243

weaning and since it controls movement co-ordination, we hypothesized that cerebellar development may be impaired in SePP-KO mice. A hallmark of hypothyroidism is incomplete cerebellar development. Granule cells, the most abundant neuronal cell type in the brain, are produced in the external germinal layer and migrate along radial glial fibers through the Purkinje cell monolayer towards their final destination, the inner granule cell layer. During developmental hypothyroidism migration of granule cells is delayed and arborization of Purkinje cell dendritic trees impaired. Thus, we studied thyroid hormone levels and deiodinase activities in SePP-KO mice in brain and other tissues and analysed cerebellar development morphologically. To our surprise, we could neither demonstrate delayed cerebellar granule cell migration nor incomplete dendritic tree development of Purkinje cells in SePP-KO cerebellum. In addition, activity of type 2 deiodinase, the major T4 activating deiodinase in brain, was unaltered [86]. Thus, we could not support local hypothyroidism as the cause of the movement disorder of SePP-KO mice. In addition, mice which were first supplemented with Se and thus phenotypically rescued, developed the movement phenotype after withdrawal of supplemental Se (Figure 1). Thus, we conclude that the movement phenotype of SePP-KO mice is not a result of developmental disturbance of cerebellar development, but may rather be a consequence of functional impairment or neurodegeneration due to Se depletion of the brain. There is indeed evidence that SePP is required for neuronal survival. Following the identification of the nerve growth factor (NGF) as a survival factor for sympathetic and sensory neurons [87], the search started for factors that could promote the survival of spinal and cortical neurons. Since it was known that factors contained within fetal calf serum (FCS) helped to maintain primary central neurons in culture, Kaufman and Barrett fractionated FCS and partially purified a protein unlike other known neurotrophic factors [88]. In a follow-up paper, Yan and Barrett [89] purified the corresponding factor to homogeneity and identified it as SePP. Despite its extraordinary potency in promoting neuronal survival, the issue was not followed much further and the search for neurotrophic factors remained centered on homologs of NGF and ligands for receptor tyrosine kinases. The study of SePP as neuronal survival promoting factor was probably loosing attraction after the development of serum-fi'ee culture conditions [90]. Thus it remained largely unappreciated in the field that Se was found indispensible for serum-fi'ee neuron culture. In fact, culture supplements like "N2", "B24", and the nowadays commercially available "B27" contain selenite as a Se source (approx. 50 nM). In a series of elegant studies, Takahashi and coworkers have shown that SePP can act as a Se supply protein [91,92], and in retrospect, it seems conceivable that SePP in FCS acts as the main Se source for cultured neurons in the absence of inorganic selenite.

244

Selenium: Its molecular biology and role in human health

These in vitro studies suggested that other Se sources, like selenite, can substitute for SePP in order to promoter cellular function. Since SePP is expressed in most tissues including the brain, but most abundantly in liver, we wanted to determine the role of hepatically secreted SePP in contrast to locally expressed SePP in the brain. To this end, we made use of a mouse model in which selenoprotein expression is abrogated in a hepatocytespecific manner [93]. When we compared these mice with our SePP-KO mice, we demonstrated that Alb-Cre/Trsp"'" mice, although containing similarly reduced amounts of Se in plasma as SePP-KO mice, did not exhibit the same neurological deficits, but rather appeared entirely normal [94]. Brain Se and GPX activity remained unaltered in Alb-Cre/Trsp*"" mice, while kidney GPX and plasma GPX (which is secreted from kidney) were reduced similarly as in SePP-KO mice. We concluded from this study that local SePP expression in the brain is necessary for normal brain function, possibly because of a local transport or storage role of SePP within the CNS, while hepatically derived SePP supplies Se to other organs like, e.g., the kidney. Since SePP-KO mice can be rescued with inorganic Se supplements, plasma SePP is likely not the only chemical form of Se taken up by the brain.

,

•'••

•••

• • • •

•

n. >

;

*4c

A X %. i

Figure 2. Cerebellar development is deranged in Tal-Cre/Trsp*"" mice. A sagittal section through the cerebellum at postnatal day 12 was stained with an antibody directed against parvalbumin as a marker for Purkinje neurons. There is marked Purkinje cell loss (open arrow) and some Purkinje cells are found at ectopic locations (solid arrow). Purkinje cell dendrites are stunted, partially disordered, if present at all.

Since metabolic disorders sometimes result in seizures, we wanted to verify that the neurological deficits in SePP-KO mice are primarily related to brain Se deficiency. To probe the role of selenoproteins for neuronal development and function, we inactivated selenoprotein biosynthesis in a neuron-specific manner using mice expressing Cre recombinase under control of the tubulin-

Selenium, selenoproteins, and brain function

245

a l promoter (Tal-Cre) [95]. In this model, Cre-mediated recombination starts after the final mitosis and neuronal differentiation, such that the target gene, in our case Trsp, is inactivated early-on in the life of respective neurons. Tal-Cre/Trsp"'*' mice are bom at the expected frequency indicating that there is no appreciable embryonic lethality. After about one week of age, however, the knockouts appear growth retarded and stop gaining weight. They do not attain postural control and die before the age of two weeks. While the forebrain seemed morphologically normal, we detected severe cerebellar hypoplasia. The folding of the cerebellar cortex and most of the volume of the cerebellum derives from the massive generation of granule cells during postnatal cerebellar development. Thus it is conceivable that the hypoplasia is associated with a reduction of granule cell number. We are currently in the process of determining whether granule cell proliferation is reduced or whether a fraction of differentiated granule cells dies in the absence of functional selenoproteins. Granule cell proliferation depends, among other signals, on sonic hedgehog secretion from Purkinje cell dendrites. Given the striking patchy lack of Purkinje cells and the thinning of the external germinal layer overlying zones of Purkinje cell loss, one may conclude that reduced proliferation of granule cells partly derives from the lack of a growth signal (Figure 2). On the other hand, the selenoprotein TrxR is implicated in DNA synthesis, and the massive proliferation of granule cell precursors in the external germinal layer might be impaired because of the lack of TrxR. It will be interesting to compare Tal/Trsp*"" mice with neuron-specific mutants for essential selenoproteins and thus assign specific roles in brain development and function to individual selenoproteins. References 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

K Schwarz, CM Foltz 1957 Nutrition 15:255 D Behne, H Hilmert, S Scheid, H Gessner, W Elger 1988 Biochim Biophys Acta 966:12 G Bermano, F Nicol, JA Dyer, RA Sunde, GJ Beckett, JR Arthur, JE Hesketh 1995 BiochemJ 311:425 NE Savaskan, AU Brauer, M Kuhbacher, lY Eyupoglu, A Kyriakopoulos, O Ninnemann, D Behne, R Nitsch 2003 FASEB 717:112 E Wallace, HI Calvin, GW Cooper 1983 Gamete Res 4:377 JT Coyle, P Puttfarcken 1993 Science 262:689 P Jenner 2003 Ann.Neurol 53 Suppl 3: S26 JK Andersen 2004 Nat Med 10 Suppl: S18 U Dimagl, RP Simon, JM Hallenbeck, 2003 Trends Neurosci 26:248 PJ Crack, JM Taylor 2005 Free Radic Biol Med 38:1433 PH Chan 2004 Neurochem Res 29:1943 T Kondo, AG Reaume, TT Huang, E Carlson, K Murakami, SF Chen, EK Hoffman, RW Scott, CJ Epstein, PH Chan 1997 J Neurosci 17:4180 GW Kim, T Kondo, N Noshita, PH Chan 2002 Stroke 33:809 G Yang, PH Chan, J Chen, E Carlson, SF Chen, P Weinstein, CJ Epstein, H Kamii. 1994 5/ro)te 25:165 H Savolainen 1978 Res Commum Chem Pathol Pharmacol 21:173

246 16 17 18 19 20 11 22 23 24 25 26 27 28 29 30 31 32

33

34 35 36 37

38 39 40 41 42 43 44 45 46

Selenium: Its molecular biology and role in human health G Trepanier, D Furling, J Puymirat, ME Mirault 1996 Neuroscience 75:231 S Takizawa, K Matsushima, Y Shinohara, S Ogawa, N Komatsu, H Utsunomiya, K Watanabe 1994 J Neurol Sci 122:66 N Ishibashi, O Prokopenko, M Weisbrot-Lefkowitz, KR Reuhl, OMirochnitchenko 2002 Brain Res Mol Brain Res 109:34 N Ishibashi, O Prokopenko, KR Reuhl, O Mirochnitchenko 2002 J Immunol 168:1926 PJ Crack, JM Taylor, NJ Flentjar, J de Haan, P Hertzog, RC lannello, I Kola 2001 J NeurochemlS:\3S9 U Schweizer, L Schomburg 2005 lUBMB Life In press. D Furling, O Ghribi, A Lahsaini, ME Mirault, G Massicotte 2000 Proc Natl Acad Sci USA 97:4351 C Zimmermann, K Winnefeld, S Streck, M Roskos, RL Haberl 2004 Eur Neurol 51:157 J Virtamo, E Valkeila, G Alfthan, S Punsar, JK Huttunen, MJ Karvonen 1985 Am J Epidemiol 122:276 WQ Wei, CC Abnet, YL Qiao, SM Dawsey, ZW Dong, XD Sun, JH Fan, EW Gunter, PR Taylor, SD Mark 2004 Am J Clin Nutr 79:80 G Kenet, J Freedman, B Shenkman, E Regina, F Brok-Simoni, et al 1999 Arterioscler Thromb Vase Biol 19:2017 MA Ansari, AS Ahmad, M Ahmad, S Salim, S Yousuf, T Ishrat, F Islam 2004 Biol Trace Elem Res 101:73 R Gupta, M Singh, A Sharma 2003 Pharmacol Res 48:209 AR Green, T Ashwood 2005 Curr Drug Targets CNS Neurol Disord 4:109 JM Hollander, KM Lin, BT Scott, WH Dillmann 2003 Free Radic Biol Med 35:742 Y Takagi, A Mitsui, A Nishiyama, K Nozaki, H Sono, Y Gon, N Hashimoto, J Yodoi 1999 Proc Natl Acad Sci USA 96:4131 M Conrad, C Jakupoglu, SG Moreno, S Lippl, A Banjac, M Schneider, H Beck, AK Hatzopoulos, U Just, F Sinowatz, W Schmahl, KR Chien, W Wurst, GW Bomkamm, M Brielmeier 2004 Mol Cell Biol 24:9414 C Jakupoglu, GK Przemeck, M Schneider, SG Moreno, N Mayr, AK Hatzopoulos, MH de Angelis, W Wurst, GW Bomkamm, M Brielmeier, M Conrad 2005 Mol Cell Biol 25:1980 J Sian, M Gerlach, MB Youdim, P Riederer 1999 J Neural Transm 106:443 E Sofic, P Riederer, H Heinsen, H Beckmann, GP Reynolds,G Hebenstreit, MB Youdim 1988 J Neural Transm 74:199 D Kaur, JK Andersen 2002 Aging Cell 1:17 D Kaur, F Yantiri, S Rajagopalan, J Kumar, JQ Mo, R Boonplueang, V Viswanath, R Jacobs, L Yang, MF Beal, D DiMonte, I Volitaskis, L EUerby, RA Chemy, AI Bush, JK Andersen 2003 Neuron 37:899 HC Kim, WK Jhoo, DY Choi, DH Im, EJ Shin, JH Suh, RA Floyd, G Bing 1999 Brain /?e* 851:76 H Kim, W Jhoo, E Shin, G Bing 2000 Brain Res 862:247 S al Deeb, K al Moutaery, GW Bruyn, M Tariq 1995 J Psychiatry Neurosci 20:189 SZ Imam, GD Newport, F Islam, W Slikker, SF Ali 1999 Brain /fes 818:575 KS Zafar, A Siddiqui, I Sayeed, M Ahmad, S Salim, F Islam 2003 JNeurochem 84:438 V Sanchez, J Camarero, E O'Shea, AR Green, MI Colado 2003 Neuropharmacology 44:449 JC Bensadoun, O Mirochnitchenko, M Inouye, P Aebischer, AD Zum 1998 Eur J Neurosci 10:3231 A Castano, J Cano, A Machado 1993 J Neurochem 61:1302 P Klivenyi, OA Andreassen, RJ Ferrante, A Dedeoglu, G Mueller, E Lancelot, M Bogdanov, JK Andersen, D Jiang, MF Beal 2000 J Neurosci 20:1

Selenium, selenoproteins, and brain function 47

48 49 50 51

52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78

79

247

MV Aguilar, FJ Jimenez-Jimenez, JA Molina, I Meseguer, C Mateos-Vega, MJ Gonzalez-Munoz, F de Bustos, C Gomez-Escalonilla, M Ort-Pareja, M Zurdo, MC Martinez-Para 1998 J Neural Transm 105:1245 RJ Marttila, M Roytta, H Lorentz, UK Rinne 1988 J Neural Transm 74:87 J Kalra, AH Rajput, SV Mantha, K Prasad 1992 MolCell Biochem 110:165 P Johannsen, G Velander, J Mai, EB Thorling, E Dupont 1991 7 Neurol Neurosurg Psychiatry 54:679 I Meseguer, JA Molina, FJ Jimenez-Jimenez, MV Aguilar, CJ Mateos-Vega, MJ Gonzalez-Munoz, F de Bustos, M Orti-Pareja, M Zurdo, A Berbel, E Barrios, MC Martinez-Para 1999 J Neural Transm 106:309 P Damier, EC Hirsch, P Zhang, Y Agid, F Javoy-Agid 1993 Neuroscience 52:1 SJ Kish, C Morito, O Homykiewicz 1985 Neurosci Lett 58:343 LJ Willmore 1990 Epilepsia 31 Suppl 3:S67 JJ Rubin, LJ Willmore 1980 Exp Neurol 67:472 LJ Willmore, JJ Rubin 1981 Neurology 31:63 CL Kien, HE Ganther 1983 ^m y Clin Nutr 37:319 KM Brown, JR Arthur 2001 Public Health Nutr 4:593 GF Weber, P Maertens, XZ Meng, CE Pippenger 1991 Lancet 337:1443 U Schweizer, AU BrSuer, J K6hrle, R Nitsch, NE Savaskan 2004 Brain Res Brain Res RevA5:\(A VT Ramaekers, M Calomme, D Vanden Berghe, W Makropoulos 1994 Neuropediatrics 25:217 L Schomburg, U Schweizer, B Holtmann, L Flohe, M Sendtaer, J Kohrle 2003 Biochem J110-391 KE Hill, J Zhou, WJ McMahan, AK Motley, JF Atkins, RF Gesteland, RF Burk 2003 J Biol Chem 218:13,640 U Schweizer, M Michaelis, J Kohrle, L Schomburg 2004 Biochem J21&:2\ KE Hill, J Zhou, WJ McMahan, AK Motley, RF Burk 2004 J Nutr 134:157 R Boonplueang, G Akopian, FF Stevenson, JF Kuhlenkamp, SC Lu, JP Walsh, JK Andersen 2005 Exp Neurol 192:203 MI Bellissimo, D Amado, DS Abdalla, EC Ferreira, EA Cavalheiro, MG NaffahMazzacoratti 2001 Epilepsy Res 46:121 M Hiramatsu, A Mori 1981 Neurochem Res 6:301 DR Rosen, T Siddique, D Patterson, DA Figlewicz, P Sapp, A Hentati, D Donaldson, J Goto, JP O'Regan, HX Deng 1993 Nature 362:59 N Durany, G Munch, T Michel, P Riederer 1999 Eur Arch Psychiatry Clin Neurosci 249 Suppl 3:68 P Soba, S Eggert, K Wagner, H Zentgraf, K Siehl, S Kreger, A Lower, A Langer, G Merdes, R Paro, CL Masters, U MuUer, S Kins, K Beyreuther 2005 EMBO J 24M24 CR Comett, WR Markesbery, WD Ehmann 1998 Neurotoxicology 19:339 CR Comett, WD Ehmann, DR Wekstein, WR Markesbery 1998 Biol Trace Elem Res 62:107 D Wenstrup, WD Ehmann, WR Markesbery 1990 Brain Res 533:125 I Ceballos-Picot, M Merad-Boudia, A Nicole, M Thevenin, G Hellier, S Legrain, C Berr 1996 Free Radic Biol Med 20:579 C Jeandel, MB Nicolas, F Dubois, F Nabet-Belleville, F Penin, G Cuny 1989 Gerontology 35:275 MA Lovell, C Xie, SP Gabbita, RW Markesbery 2000 Free Radic Biol Med 28:418 A Santamaria, R Salvatierra-Sanchez, B Vazquez-Roman, D Santiago-Lopez, J VilledaHemandez, S Galvan-Arzate, ME Jimenez-Capdeville, SF Ali 2003 J Neurochem 86:479 MB Delatycki, R Williamson, SM Forrest 2000 J Med Genet 37:1

248 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95

Selenium: Its molecular biology and role in human health M Babcock, D de Silva, R Oaks, S Davis-Kaplan, S Jiralerspong, L Montermini, M Pandolfo, J Kaplan 1997 Science 276:1709 SA Shoichet, AT Baumer, D Stamenkovic, H Sauer, AF Pfeiffer, CR Kahn, D MuUerWieland, C Richter, M Ristow 2002 Hum Mol Genet 11:815 RF Burk, KE Hill (2005) Annu Rev Nutr 25: 215 MA Motsenbocker, AL Tappel 1982 Biochim BiophysActa 719:147 WM Valentine, KE Hill, LM Austin, HL Valentine, D Goldowitz, RF Burk 2005 Toxicol Pathol 33:570 L Schomburg, U Schweizer, J K6hrle 2004 Cell Mol Life Sci 61:1988 L Schomburg, C Riese, M Michaelis, E Griebert, MO Klein, R Sapin, U Schweizer, J K6hrle 2005 Endocrinology In press. R Levi-Montalcini 1987 EMBO y 6:1145 LM Kaufman, JN Barrett 1983 Science 220:1394 J Yan, JN Barrett 1998 JNeurosci 18:8682 GJ Brewer, CW Cotman 1989 Brain Res 494:65 Y Saito, K Takahashi 2002 EurJBiochem 269:5746 Y Saito, N Sato, M Hirashima, G Takebe, S Nagasawa, K Takahashi 2004 Biochem J 381:841 BA Carlson, SV Novoselov, E Kumaraswamy, BJ Lee, MR Anver, V Gladyshev, DL Hatfield 2004 J Biol Chem 279:8011 U Schweizer, F Streckfuss, P Pelt, BA Carlson, DL Hatfield, J Kohrle, L Schomburg 2005 5/oc/iemJ386:221 V Coppola, CA Barrick, EA Southon, A Celeste, K Wang, B Chen, E Haddad, J Yin, A Nussenzweig, A Subramaniam, L TessaroUo 2004 Development 131:5185

Chapter 22. Selenium as a cancer preventive agent Gerald F. Combs, Jr. Grand Forks Human Nutrition Research Center, USDA-ARS, Grand Forks, ND 58202, USA

Junxuan Lii Hormel Institute, University of Minnesota, Austin, MN 55912, USA

Summary: Most epidemiological studies have shown inverse associations of selenium (Se) status and cancer risk; almost all experimental animal studies have shown that supranutritional exposures of Se can reduce tumor yield; and each of the limited number of clinical intervention trials conducted to date has found Se treatment to be associated with reductions in cancer risks. The known metabolic functions of Se, which appear to be discharged by a fairly small number of selenoproteins may not fully explain these effects, particularly those observed in response to Se-supplementation of non-deficient subjects. Emerging evidence indicates anticarcinogenic roles of at least some selenoproteins, namely, those involved in antioxidant protection (the glutathione peroxidases), redox regulation (the thioredoxin reductases) and hormonal regulation of metabolism (iodothyronine 5'deiodinases). The fact that abundant empirical evidence has shown anticarcinogenic effects of Se in individuals with apparently full selenoenzyme expression suggests other mechanisms with relevance to non-deficient populations. Certain Se-metabolites (hydrogen selenide, methylselenol, seleno-diglutathione) can be anti-carcinogenic by inhibiting cell proliferation, stimulating cell death by apoptosis, and inhibiting neoangiogenesis. Therefore, while the hypothesis remains plausible that Sedeprivation may increase cancer risk by compromising selenoprotein expression, there is strong support for the hypothesis that supranutritional exposures to Se can reduce cancer risk. These hypotheses are not mutually exclusive, and it is likely that Se can function as a cancer preventive agent through both nutritional and supranutritional mechanisms. Emergence of a selenium-cancer link The nutritional essentiality of Se was recognized in the late 1950's when the element was found to be the active principle in liver that could replace vitamin E in the diets of rats and chicks for the prevention of vascular,

250

Selenium: Its molecular biology and role in human health

muscular and/or hepatic lesions [1]. The first suggestion that Se may be anticarcinogenic came a decade later and was based on empirical observation of an inverse relationship of cancer mortality rates and forage crop Se contents in the United States [2,3]. The body of scientific evidence that has subsequently been developed indicates that, indeed, Se can play a role in cancer prevention. Epidemiological Evidence The epidemiological literature on Se and cancer has been reviewed [4-7]. Most, but not all, of this literature has found Se status to be inversely associated with cancer risk. Prospective cohort studies in several countries have all shown cancer cases to have significantly lower mean pre-diagnostic serum Se levels than controls [8-15]. Negative associations have been found for various parameters of Se status and risks to cancers or pre-cancerous lesions of the bladder [8], brain [16], esophagus [17], lung [18-20], head and neck [21], ovary [9], pancreas [22], thyroid [23], stomach [24,25], melanoma [26], prostate [27] and colon [28]. Animal Model Evidence Studies with animal tumor models have shown that Se treatment can reduce tumor yields. Some years ago. Combs [29] estimated that, of what was then more than 100 studies in which tumor production and/or preneoplastic endpoints had been measured, two-thirds showed that supranutritional Se doses reduced the incidences of such outcomes, with half showing reductions of 50% or more. Further studies have demonstrated similar reductions in tumor yields (see [30]) or experimental metastases [31,32]. Four studies have found selenite treatment to enhance tumorigenesis; but the interpretation of these is not straightforward, as three [33-35] found increases in tumors at one site to be accompanied by reductions at another site, and one [36] found such enhancement only when the carcinogen was administered in a certain way. Clinical Trial Evidence Several clinical trials have been conducted to determine the efficacy of Se in reducing cancer risk in humans. Yu et al [37,38] reported that, after 8 yrs, the incidence of primary liver cancer (PLC) was 35% lower in a community using selenite-enriched table salt compared to non-treated communities which showed no changes in PLC incidence; however, their data were not analyzed statistically. They also reported lower PLC incidence among hepatitis surface antigen-positive subjects randomized to a Se-enriched yeast treatment (200 meg Se per day) in comparison to a placebo group (0 vs. 5 cases, P<0.05). A series of trials was conducted in 1984-1991 in Lianxian, Henan Province, an area with a high prevalence of esophageal cancer [39-

Selenium as a cancer preventive agent

251

41]. The first trial involved 3698 subjects randomized to an array of mixed treatments; subjects receiving Se-enriched yeast (50 meg Se per day), vitamin E and beta-carotene showed modest reductions in total (9%) and stomach cancer mortality (21%). However, subjects with confirmed esophageal dysplasia in a companion study [39] showed no significant treatment effects on any cancer-related endpoint. In a large trial (29,584 subjects) [42,43], the same group noted a 13% reduction in total and cancer mortality in the group treated with Se-enriched yeast (50 meg Se per day) plus vitamin E and beta-carotene. Krishnaswamy et al [44,45] reported that smokers with precancerous oral lesions randomized to a twice-weekly supplement containing vitamin A, riboflavin, zinc and Se-enriched yeast (100 meg/day for 6 months, followed by 50 meg/day for 6 months.) showed a significantly greater rate of complete remission of lesions compared to those in the control group (57% vs. 8%, P<0.05). Subjects without lesions in the treated group showed fewer new lesions than those in the placebo group (12% vs. 38%, P<0.02). However, these latter studies were complicated by the use of a combination of agents. The strongest evidence of anti-cancer efficacy of Se in humans comes fi-om the Nutritional Prevention of Cancer (NPC) Trial [46-51]. That was a double-blind, randomized, placebo-controlled clinical trial designed to test the hypothesis that a daily oral dose of Se (200 mcg/d as Se-enriched yeast) could reduce the rate of recurrent non-melanoma skin cancer in a high-risk group of 1312 older (63.2± 10.1 yrs) Americans. The results, for the first 10 years of the trial (average 6.4 yrs follow-up) [46,47] showed no significant effects of Se-treatment on the incidences of either of the primary endpoints of the study: basal or squamous cell carcinoma of the skin. They did, however, show significant differences in risks to total cancer incidence, total cancer deaths, incidences of carcinomas of the prostate, lung, colon-rectum, and total non-skin. Analyses of the complete trial results (average 7.9 yrs follow-up) [48-51] support the strongest protective effects previously detected: Se-treatment was associated with reduced risks to total cancer incidence (RR=0.63) and incidences of carcinomas of the prostate (RR=0.51) and colon-rectum (RR=0.46); but no effect on lung cancer incidence (RR=0.70, P=0.18). The new analysis seemed to refute the previous conclusion that Se-treatment did not affect the primary endpoint of nonmelanoma skin cancer; however, the analysis using a sub-set of 1250 patients with confirmed baseline blood Se levels support the original finding [46], and showed that while Se-treatment did not affect risk of basal cell carcinomas (BCC) it appeared to delay the diagnosis of the first BCC [48,51]. With the increased statistical power of the complete data set, the more recent analyses also found Se-treatment to be associated with increased risks to both squamous cell carcinomas (SCC) (RR=1.31, P=0.005) and total non-melanoma skin cancers (RR=1.22, P=0.004). However, it is doubtful

252

Selenium: Its molecular biology and role in human health

that such effects can be attributed to the Se-treatment as cancers diagnosed soon after patient randomization to treatment will have resulted from cellular events that occurred prior to that time, perhaps years earlier. Thus, it is significant that analyses of cancer outcomes diagnosed after two years of treatment showed no significant treatment effect on SCC incidence (RR=1.21,P>0.05)[39]. The complete NPC Trial data show that, for men with plasma prostate specific antigen (PSA) concentrations <4 ng/ml, Se-treatment was associated with a 65% reduction in prostate cancer risk (P=0.01) [50], while for men with PSA>4 ng/ml, there was no significant effect of treatment (RR=0.88, P=0.86). Neither did Se-treatment reduce elevated PSA values or affect the clinical stage or incidence of advanced prostate cancers. These findings are consistent with protection by Se in early stage(s) of carcinogenesis; however, there was no indication that Se-treatment affected the stage of prostate disease among men with that diagnosis. Protection by Se was significant only for subjects who entered the trial with relatively low baseline plasma Se levels [51]. Those with plasma Se<106 ng/ml, i.e., in the lowest tertile of the cohort, showed the strongest treatment effect (RR=0.14, P=0.002), while subjects in the middle tertile of plasma Se, 107-123 ng/ml, showed a modest but still protective effect (RR=0.39, P=0.03). Subjects in the highest tertile (plasma Se >123 ng/ml) showed no significant treatment effect (RR=1.20, P=0.66). That this effect was not due to diagnostic bias was indicated by the results of simulations of diagnoses projected from comparable biopsy rates in both treatment groups, which showed significant protection by Se in the lowest plasma Se tertile group despite a generally attenuated cancer incidence (50). It is significant that the protective effect of Se was observed in the NPC Trial in a population that, with baseline plasma Se levels of 114+23 ng/ml, was not deficient in Se [46]. Only 2 subjects had levels below 80 ng/ml, which Neve [52] observed as the level above which healthy adults show no further increases in Se-dependent glutathione peroxidase activities in response to supplemental Se. Those blood levels suggest an average intake of at least 85 meg Se per day, or at least 155% of the RDA [53]. That subjects entering the trial with plasma Se levels <120 ng/ml showed the greatest risks of subsequent cancer as well as the sfrongest protective effects of Se-yeast supplementation would suggest that level as an appropriate upper limit for eligibility for future cancer prevention trials using Se. Mechanisms of anti-carcinogenicity It is highly relevant to public health considerations to determine whether the apparent relationship of Se and cancer risk involves increased risks due to Se-deficiency, reduced risks associated with adequate to luxus Se status, or whether both types of relationship may occur. That Se deficiency could

Selenium as a cancer preventive agent

253

increase cancer risk might be expected on the basis of the known functions of selenoenzymes in antioxidant protection (glutathione peroxidases, GPXs), and redox regulation (thioredoxin reductases, TRs) and, thus, in the metabolic defense against carcinogenic free radicals. Mutagenic oxidative stress is generally thought to be a major factor in the initiation of human carcinogenesis, as the electron-rich DNA bases are susceptible to electrophilic attack by reactive oxygen species (ROS) including superoxide radical, hydrogen peroxide, singlet oxygen, hydroxyl radical, and electrophilic metabolites of xenobiotics and other reactive intermediary metabolites. These can cause genetic damage and the production of mutant oncogenes and tumor suppressor genes as well as epigenetic changes that affect expression. Evidence that Se may protect against such changes includes findings that the induction of skin tumors by either ultraviolet irradiation [54-56] or phorbol esters [36] varied inversely with skin GPX activity in animal models, and that protection by selenite against (2oxoproyl)amine-induced intrahepatic cholangiocarcinomas in Syrian golden hamsters correlate with the restoration of hepatic GPX activity [57]. Alternatively, that anti-carcinogenic effects of Se have been observed under conditions of maximal selenoprotein expression would suggest the existence of other anti-carcinogenic mechanisms most likely involving Semetabolites. On this point, a large body of evidence indicates that Se intake in excess of the nutritional requirement can inhibit and/or retard tumorigenesis in experimental animals. Anti-tumorigenically effective Seexposures in animal models (e.g., at least 1.5 mg/kg diet) have typically been an order of magnitude greater than those required to prevent clinical signs of Se deficiency or to support the maximal expression of known selenoproteins (less than 0.2 mg/kg diet). Accordingly, it is significant that Se-supplements were effective in reducing cancer risks in the NPC Trial [46] few, if any, subjects of which had nutritionally limiting Se intakes as judged by their baseline plasma Se levels', hi fact, subjects entering the NPC Trial with plasma Se levels below ca. 120 ng/ml showed higher risks of subsequent cancer as well as the strongest apparent protective effects of Se-yeast supplementation^ [46,58]. Available evidence addresses the cancer-impacts of supranutritional Se exposures far more completely than it does those of Se-deficiency. Thus, while it remains plausible that Se-deprivation may enhance tumorigenesis, it is well documented that at least some forms of Se can, in supranutritional 'The cohort level was 114+23 ng/ml; very few subjects had levels below 80 ng/ml, the level N6ve [52] found to be the upper limit for GPX responses to supplemental Se in healthy adults. These levels suggest an average daily intake of at least 85 meg Se/day, or at least 155% of the RDA [53]. ^Those findings would suggest that the plasma Se level of 120 ng/ml might be appropriate as an eligibility upper limit for future cancer prevention trials.

254

Selenium: Its molecular biology and role in human health

doses, reduce cancer risk. Accordingly, while anti-carcinogenic functions of selenoproteins cannot be excluded, it is probable that one or more Secompounds/metabolites can function in directly anti-carcinogenic ways. Individually and collectively, these two general mechanisms would appear to underlie the anti-carcinogenic effects described for Se in cellular antioxidant protection, carcinogen metabolism, gene expression, immune surveillance, cell cycle/death regulation and neo-angiogenesis (see reviews [4,5,7,59]). Metabolic Bases for Selenium Anti-Carcinogenesis Roles of Selenoenzymes. Because the etiologies of at least some cancers are believed to involve mutagenic oxidative stress, the antioxidant GPXs and the redox-regulatory TRs are expected to have anti-carcinogenic impact by way of genetic and epigenetic pathways: by removing DNA-damaging H2O2 and lipid hydroperoxides; by blocking the production of reactive oxygen species and malonyldialdehyde; by regulating the redox signaling system that is critical to the growth of many cancers [59-61]. Such mechanisms may underlie the protection by Se against both the carcinogenic and cytotoxic effects of UV-irradiation, which are thought to be due to the oxidative stress of H2O2 generated photochemically. This has been observed at doses within the range (0.1-0.5 mg/kg diet) of maximization of selenoprotein expression [54-56,62-66]. As essential components of the iodothyronine 5-deiodinases (TDIs), Se is important in the regulation of thyroid hormone metabolism, and may thereby affect cancer cell growth. Several findings point to this possibility: thjroid hormones were shown to oppose the proliferative action of estrogen on breast cancer [67], breast cancer incidence was higher in areas of endemic iodine deficiency than in non-endemic areas [68], and breast cancer patients had lower plasma thyroid hormone (T3) levels as well as a tendency toward lower toenail Se concentrations than controls [69]. Nevertheless, the epidemiological evidence relating thyroid hormone status and breast cancer risk is not consistent [69-74]. Selenium, presumably through the actions of antioxidant selenoproteins (GPXs and/or TRs), has been shown to modulate p53 activity by redox modification of cys275,277 mediated by Ref-1 (75), enhancing repair of drug- or radiation-induced DNA damage (76). Because p53 is known to suppress the expression of the angiogenic factor VEGF (77) and to induce the expression of the angiogenesis suppressor thrombospondin-1 (78), a selenoprotein-mediated increase of p53 activity could play a pivotal role in switching off angiogenesis in early lesions. The association of allelic variation in selenoprotein expression with cancer risk and cancer cell responses to Se suggest the involvement of at least some selenoproteins in cancer protection. For example, a single nucleotide polymorphism (SNP) at codon 198 of human GPX-1, resulting in a

Selenium as a cancer preventive agent

255

substitution of leucine for proline, has been associated with increased risks of cancers of the lung [79], breast [80], head and neck [81], and bladder [82], although not basal cell carcinoma of the skin [83]. Recent studies indicate that the 198-leucine genotype may be less responsive to Se exposure than the 198-proline genotype [79], suggesting that the increased cancer susceptibility of individuals with that allele may be due to their increased need for Se for maximal GPX-1 activity. Polymorphism has also been identified in the promoter [84] of selenoprotein P (SelP) [85]; however, its frequency was similar in groups of healthy controls and colorectal adenoma patients [84], although the malignant tissues showed reduced SelP expression compared to adjacent normal tissue [86,87]. Reduced expression of the 15 kDa selenoprotein (Sepl5) has been observed in malignant liver and prostate [88], as well as in malignant mesothelioma cells which also showed resistance to growth-inhibitory and pro-apoptotic effects of Se [89]. Two Sep 15 SNPs, at positions 811 (C/T) and 1125 (G/A) have been identified, both in the selenocysteine insertion sequence (SECIS) element located in the 3'-UTR, which affect the incorporation of Se into selenoproteins [90,91]. Thus, it is likely that one or more groups of selenoproteins may play important anticarcinogenic roles. It would follow, therefore, that anticarcinogenic roles of selenoproteins would be limited under conditions of insufficient Se supply, i.e., in populations with deficient or marginal Se status. Correction of nutritional Se-deficiency might, therefore, be expected to have anti-carcinogenic effects by increasing selenoprotein expression (see Figure 1). That possibility not withstanding, the documented efficacy of Sesupplementation in reducing cancer risk in non-deficient individuals (i.e., at supranutritional intakes) would suggest the involvement of other anticarcinogenic mechanisms that function in addition to mechanisms involving selenoproteins, and that these may be important under apparently Seadequate conditions. Roles of Se-Metabolites. There is evidence of anti-carcinogenic activities for several intermediary metabolites of Se. These include selenodiglutathione (GSSeSG), the reductive metabolite of the oxidized inorganic salts (selenite, selenate); hydrogen selenide (H2Se), the common intermediate of that reductive pathway and the catabolism of selenoamino acids; and the methylated metabolites of selenide ([CH3]xSe) that have hitherto been thought of only as excretory forms of the element. The anti-carcinogenic activities attributable to each of these metabolites are summarized in Figure 1. The product of the thiol-dependent reduction of selenite, GSSeSG, would appear to be relevant only under conditions of exposure to selenite and/or selenate. That metabolite has been shown to block protein synthesis by inhibiting eukaryotic initiation factor 2 [93], to inhibit ribonucleotide

256

Selenium: Its molecular biology and role in human health

reductase [94], to serve as an efficient oxidant of thioredoxin [95], and to suppress the mRNAs for several GPX isoforms [96]. That these effects can be anti-carcinogenic is suggested by findings that GSSeSG can inhibit DNAbinding of the transcription factor AP-1 [97], inhibit cell proliferation [93,96,98-100], enhance apoptosis [99], and be more effective than either selenite or selenomethionine in inhibiting the growth of Ehrlich ascites tumors in mice [101]. Both selenite and GSSeSG have been shown to induce apoptosis in vitro; however, unlike selenite, GSSeSG did not induce widespread tyrosine phosphorylation characteristic of oxidative stress [102]. Because GSSeSG is unstable under physiological conditions it is unlikely to accumulate in cells, breaking down instead to glutathione selenol (GSSeH) a n d H2Se.

>I'DNA, ^RNA, >lprotein synthesis tapoptosis, 4'AP-l, >lNF-kB

NazSeO.

•-GPXsJRs —•• ^ROS, tredox control SeMet SeCys

t02", tH202, tDNA SSBs, S/G2-arrest, >lpolyamines, tapoptosis iPKC, i endothelial MMP, ^epithelial VEGF

CHjSeOjH CHjSeCN CHjSeCys

Gi arrest, tcaspase-mediated apoptosis (CH3)2Se (breath)

(w/o genotoxicity)

I. (CH3)3Se^ (urine) Figure 1. Se-metabolites apparently active in cancer prevention (after [92]). Abbreviations: SeMet, selenomethionine; SeCys, selenocyeteine; CH3Se02H, methyl-seleninic acid; CHsSeCN, methylselenocyanate; CHsSeCys, Se-methylseleno-cysteine; GPXs, glutathione peroxidases; TRs, thioredoxin reductases; ROS, reactive oxygen species; SSBs, DNA single strand breaks; PKC, protein kinase C; MMP-2, matrix metalloproteinase-2; VEGF, vascular endothelial growth factor; PSA, prostate specific antigen; AR, androgen receptor.

Hydrogen selenide appears to be an important player in Se-anticarcinogenesis by way of its further metabolism. Its oxidative metabolism produces superoxide anion (O2 ) and H2O2, the formation of which induces DNA single-strand breaks leading to S phase/G2 cycle arrest and cell death by

Selenium as a cancer preventive agent

257

apoptosis [103-107]. This mechanism would appear to mediate seleniteinduced apoptosis, as the genotoxic and pro-apoptotic effects of selenite on leukemia, mammary or prostate cancer cells have been shown to be blocked by a superoxide dismutase or its mimetics [103,108,109], but not by an hydroxyl free radical scavenger [110]. Further, catalase added to the cell culture medium blocked the induction of cell death by selenite [111]. In addition, H2Se can be methylated to produce a string of metabolites that, although being readily excreted, include some that are anti-carcinogenic. Ip, Ganther and coworkers [112-120] have produced strong experimental evidence that the anti-tumorigenic effects of Se are mediated by methylselenol (CHsSeH) or its derivatives (see Figure 1). They found that the CHaSeH-precursors selenobetaine (CH3Se02H) and methyl-selenocysteine (CHsSeCys) are anti-carcinogenic in the 7,12-dimethylbenzanthracene (DMBA)-induced rat mammary tumor model, each being somewhat more efficacious than selenite. In contrast, dimethyl selenoxide, which is metabolized to dimethylselenide ([CH3]2Se) and very rapidly excreted in the breath, was very poorly chemo-preventive, and the rapidly excreted urinary metabolite trimethylselenonium ([CHsJsSe^) was completely ineffective. Further work has shown that the CHsSeH-precursors methylselenocyanate (CHsSeCN) and CHsSeCys can each inhibit mammary cell growth, arresting cells in the Gi or early S phase and inducing apoptosis [106,107,118-122], The latter effect is caspase-dependent [123], as methyl-Se induced apoptosis involves at least three caspase-dependent actions: mitochondrial release of cytochrome C, cleavage of poly(ADP-ribose), and DNA nucleosomal fragmentation. Selenite-induced cell death, in contrast, is independent of these death proteases [109,121,123,124]. That methyl-Se can cause caspasedependent apoptosis in cell lines that do not contain functional p53 [124] suggests that its pro-apoptotic action is independent of p53. This was also evident in a recent study in which methyl-Se induced apoptosis of p53positive, LNCaP cells was found not to involve a change in p53 activation [125]. Se-Methylselenocysteine has been shown to inhibit the cell cycle regulatory enzymes CDK2 and protein kinase C (PKC) [126,127]. Unlike the proximal H2Se-precursors, CHaSeH-precursors potently inhibit the expression of matrix metalloproteinase (MMP-2) in vascular endothelial cells and of vascular endothelial growth factor (VEGF) in cancer cells [121,122,128,129], critical components of the angiogenic response, suggesting that methyl-Se inhibits cellular proliferation and survival of activated endothelial cells by inhibiting neo-angiogenesis. Sub-apoptotic concentrations of methyl-Se have been shown to reduce androgen receptor protein expression [130] and to inhibit androgen-stimulated PSA promoter transcription [130-132], to reduce PSA expression and secretion [130], and to cause rapid PSA degradation [130]. These findings suggest a unique

258

Selenium: Its molecular biology and role in human health

mechanistic basis for the apparent sensitivity of the prostate to Seanticarcinogenesis [46,47]. Because methyl-Se compounds can be demethylated ultimately to feed the H2Se-exchangeable metabolic pool (see Figure 1), both CHsSeCys and dimethylselenoxide can support GPX expression [116]. Despite that phenomenon, evidence indicates that CHaSeH and its precursors have anticarcinogenic actions independent of those associated with the H2Se pool. Ip et al [112-117] found that arsenic, which competitively inhibits both the methylation of H2Se and the demethylation of CHsSeH (and the analogous di- and tri-methylated species) greatly reduced the anti-tumorigenic effects of selenite while enhancing those of selenobetaine or methylselenocysteine (CHjSeCys) which yields CHjSeH metabolically. Specifically, CHaSeHprecursors were shown to lack the genotoxic (DNA single-strand breaks [106,107,118,133] or DNA-oxidative damaging [134]) effects of selenite or selenide. The anti-carcinogenic activities of the methylated Se-metabolites and synthetic Se-compounds are likely related to reactions with critical proteins as well as to redox cycling, which effects may selectively impact the transformed phenotype. Ganther [128] described ways in which Secompounds may affect cellular proteins: through the formation of selenotrisulfide (-S-Se-S-) and selenylsulfide (-S-Se-) bonds and the catalysis of disulfide bonds formation/ dissolution, which would affect the activities of many enzymes with critical sulfhydryl groups; and through the formation of diselenide bonds (-Se-Se-) affecting the activities of selenoproteins which have SeCys residues at their active centers. Selenium-induced inhibition, presumably due to one or more of these reactions, has been demonstrated for a variety of relevant enzymes: ribonuclease [135], Na,K-ATPase [136], PKC [127,137,138]. Inhibition of PKC would be particularly important, as that enzyme system is known both to activate nuclear transcriptional factors and to bind phorbol ester-type tumor promoters. The inhibition of PKC by a Semetabolite such as CHsSeH would be expected to trigger a number of downstream effects including cell cycle arrest, apoptosis and angiogenic switch regulation. Evidence for at least some of these effects has been reported in response to the CHsSeH-precursors: decreased cdk2 kinase activity [126]; decreased DNA synthesis and elevated gadd gene expression [107]; inhibition of vascular endothelial MMPs and VEGF expression [139]. Thus, it appears that Se- doses large enough to support high, steady-state concentrations of CHsSeH can effect anti-carcinogenesis by inhibiting critical redox-sensitive factors including PKC and, probably, NF-kB [140] and AP-1, thus, impairing tumor cell metabolism and transformation. These effects would appear to be fairly targeted to certain factors, rather than involving wider perturbations in cellular redox control. After all, Semetabolites are typically present in tissues in much lower (nano- to micro-

Selenium as a cancer preventive agent

259

molar) concentrations than those (miUimolar) of thiols. In fact, susceptibility to redox modification by Se-attack seems to be limited to structures containing clustered cysteinyl residues [137,138]. Many of the effects of Se-compounds on cell proliferation may result from their abilities to form catalytically active, redox-cycling intermediates. Selenite, diselenides and the oxidation product of H2Se, selenium dioxide, for example, can each react with GSH to produce the selenolate ion (RSe) [141-143]. In the presence of GSH and molecular oxygen, RSe" can cycle continuously to generate Oi' and H2O2. This redox cycling is thought to be the basis of Se-toxicity, and it is possible that it may also contribute to anticarcinogenesis. Spallholz et al (144) found that dimethyldiselenide ([CH3Se]2) was the most catalytically active of a series of 19 Se-compounds^ in its ability to generate in vitro O2" in the presence of GSH and O2. They attributed this activity to CHsSeH produced by the reduction of ([CH3]Se)2 presumably generating the radical anion CHsSe'; however, it remains to be determined whether such catalytically active species can be generated intracellularly as the result of the metabolism of proximal (e.g., CHsSeCys, CH3Se02H) and/or upstream (e.g., SeMet, SeCys) precursors. There is no evidence that the common forms of Se in foods and feedstuffs, the selenoamino acids selenomethionine (SeMet) and selenocysteine (SeCys), are directly anticarinogenic. However, each can be metabolized first to H2Se and, then, to CHsSeH (see Figure 1). That conversion occurs directly for SeCys, which cannot be used directly in general protein synthesis; it is catabolized by a lyase to yield H2Se. The process is not direct for SeMet, which can enter the general protein pool as a mimic of methionine (Met). In fact, the conversion of SeMet from either dietary or proteinturnover sources necessarily involves its first being converted to SeCys by the Met-transsulfuration pathway. For this reason, most studies have found SeMet to be generally less anti-carcinogenically efficacious than SeCys or selenite [145-149], as would be expected in short-term studies and, particularly, under conditions of limiting Met supply. However, under steady-state conditions effected by long-term use, and particularly with highMet diets, one would expect the anti-tumorigenic efficacy of SeMet to approach that of SeCys and selenite. A number of synthetic Se-compounds have also been found to be anticarcinogenic. Ip et al [149] tested a series of alkylselenocyanates (H[CH2]xSeCN) using the DMBA-induced murine mammary carcinogenesis model, finding that anti-carcinogenic efficacy varied directly with increasing ^In addition to ([CH3]Se)2, these included nine other catalytically active Se-compounds: selenite, selenium dioxide, selenocystine, selenocystamine, diselenopropionic acid, diphenyldiselenide, dibenzyldiselenide, pXSC and 6-propylselenouracil; and nine Se-compounds that were not catalytically active: elemental Se, selenate, SeMet, CHaSeCys, selenobetaine, dimethylselenoxide, selenopyridine, TPSe and potassium selenocyante.

260

Selenium: Its molecular biology and role in human health

chain length up to five carbons. The same group [150] also showed that allyl-selenocysteine, which is expected to yield allylselenol, a fairly hydrophobic metabolite, is more anti-carcinogenic than the corresponding alkylseleno-cysteine. Several aryl selenocyanates have also been found to be anti-tumorigenic. The more effective of these are benzylselenocyanate [151153], p-methoxybenzyl-selenocyate [152],/;-phenylselenocyanate [152-156]. These compoimds are thought to undergo initial metabolism through arylselenol, which may explain their similar responses to the alkylselenocyanates and other CHsSeH-precursors. Each induces apoptosis of cancer cells in vitro without inducing DNA single strand breaks. When compared to selenite on a molar basis, these forms are not only less effective in supporting GPX expression but also less toxic; yet, they offer comparable anti-tumorigenic efficacy [157,158]. It would appear that the anticarcinogenic efficacies of these synthetic Se-compounds are related to their relative lipophilicities and, thus, to uptake/retention by transformed cells. Accordingly, their anti-tumorigenic efficacies would appear to be affected by dietary fat intakes, being enhanced by the use of low-fat diets [159]. That anti-carcinogenicity need not involve selenoprotein expression is again evidenced, this time by triphenylselenonium chloride (TPSe), which is antitumorigenic at fairly high levels of exposure (dietary EC5o=15 ppm for preventing DMBA-induced mammary cancer [158]). The Se in TPSe is tightly bonded to three unsubstituted benzene rings rendering it unavailable to metabolism, ineffective in supporting GPX expression in the Se-deficient rat, and without adverse effects on rat growth at dietary levels as high as 200 ppm [160]. Conclusion Increasing evidence shows that Se-compounds can inhibit and/or delay carcinogenesis in animal models and reduce the risks for at least some kinds of cancer in humans. These effects may involve the protective, nutritional functions of Se as an essential constituent of a number of metabolically important selenoenzymes; such functions may be compromised in Sedeficient individuals. Recent evidence suggests that allelic variants of some selenoproteins may be related to cancer risk. In addition to such effects, certain Se-metabolites, notably methyl-Se compounds, appear to inhibit carcinogenesis through mechanisms unrelated to the nutritional functions of Se and at doses greater than necessary for such functions, i.e, at supranutritional levels of exposures. Thus, the emerging picture is of Se as a nutrient that functions in anti-carcinogenesis in two ways: as an essential constituent of metabolically important selenoproteins, and as a source of anticarcinogenic metabolites. Because selenoprotein expression appears to be optimized at lower levels of Se exposure than are necessary to support the

Selenium as a cancer preventive agent

261

latter functions, minimization of cancer risk appears to require Se intakes greater than those required for maximal selenoprotein expression. References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40.

K Schwarz, CM Foltz 1957 J Am Chem Soc 79:3292 RJ Shamberger, DV Frost 1969 Can Med Assn 7104:82 J Shamberger, CE Willis 1971 Clin Lab Sci 2:211 GF Combs Jr, WP Gray 1998 Pharmacol Ther 79:179 GF Combs Jr, LC Clark 1999 in Nutritional Oncology D Heber, GL Blackburn VLW Go (eds) Academic Press Inc New York 215 P Kneckt 2002 in Handbook of Antioxidants E Cadenas, L Packer (eds) Marcel Dekker Inc New York 665 PDWhanger 2004 S/-yiV«/r 91:11 KJ Helzlsouer et al 1996 J Nat Cancer Inst 88:32 KJ Helzlsouer, GW Comstock, JS Morris 1989 Cancer Res 49:6144 FJ Kok, AM De Bruijn, A Hofman, R Vermeeren, HA Valkenburg 1987 Am J Epidemiol 125:12 JT Salonen et al 1985 fir Med J 290:417 J Nayini, K El-Bayoumy, S Sugie, LA Cohen, BS Reddy 1989 Carcinogen 10:509 W Willett et al 1983 Lancet 2:130 Van Den Brandt et al 1993 J Nat Cancer Inst 85:224 AMY Nomura, J Lee, GN Stemmemann, GF Combs Jr 2000 Cancer Epidemiol Biomarkers Prev 9:883 P Philipov, K Tzatchev 1988 Zentrabl Neurochir 49:344 K Jaskiewicz, WFO Marasas, JW Rossouw, FE Van Niekerk, EWP Heinetech 1988 Cancer 62:263 L. Geardsson, D Brune, IGF Nordberg, PO Wester 1985 Br J Industrial Med Al:(,\l H Miyamoto et al 1987 Cancer 60:1159 P Knekt et al 1990 J Nat Cancer Inst 82:864 T Westin et al 19S9 Arch Otolaryngol Head Neck Surg 115:1079 PGJ Bumey, GW Comstock, JS Morris 1989 J Clin Nutr 49:895 E Glattre et al 1989 Int J Epidemiol 18:45 CP Caygill, K Lavery, PA Judd, MJ Hill, AT Diplock 1989 Food Addit Contam 6:359 P Knekt et al 1988 Int J Cancer 42:846 U Reinhold, H Blitz, W Bayer, KH Schmidt 1989 Acta Derm Venerol 69:132 K Yoshizawa et al 1998 J Nat Cancer Inst 90:1219 LC Clark et al 1993 Cancer Epidemiol Biomarkers Prev 2:41 GF Combs Jr 1989 Nutrition and Cancer Prevention T Moon, M Micozzi (eds) Marcel Dekker New York, 389 C l p 1998 yA^M/r 128:1845 L Yan, J A Yee, MH McGuire, GL Graef 1997 Nutr Cancer 28:165 L Yan, JA Yee, D Li, MH McGuire, GL Graef 1999 Anticancer Res 19:1327 J Ankerst, H Sjogren \9S2 Int J Cancer 29:701 RD Dorado, EA Porta, TM Aquino 1985 Hepatol 5:1201 H Nakadaira, T Ishizu, M Yamamoto 1996 Cancer Lett 106:279 JP Perchellet, NL Abney, RM Thomas, YL Guislan, EM Perchellet 1987 Cancer Res 47:477 SY Yu, YJ Zhu, WG Li 1997 Biol Trace Elem Res 56:1 \7 SY Yu, YJ Zhu, QS Huang, CZ Wang, QN Zhang 1991 Biol Trace Elem Res 29:289 JY Li et al 1993 J Nat Cancer Inst 85:1492 PR Taylor, B Li, S Dawsey 1994 Cancer Res 54:2029s

262 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53.

54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85.

Selenium: Its molecular biology and role in human health WJ Blot, JY Li, PR Taylor, W Guo, SM Dawsey, B Li 1995 Am J Clin Nutr 62:14248 WJ Blot et al 1993 J Nat Cancer Inst 85:1483 WJ Blot 1997 Proc Soc Exp Biol Med 216:291 K Krishnaswamy, MP Prasad, TP Krishna, VV Annapuma, GA Reddy 1995 Eur J Cancer 31:41 MP Prasad, MA Makunda, K Krishnawamy 1995 Eur J Cancer 31B:155 LCClarketal 1996y.4/«iWe^^i5oc 276:1957 LC Clark et al 1998 Brit J Urol 81:730 AJ Duffield-Lillico, et al 2002 Cancer Epidem Biomarkers Prev 11:630 ME Reid et al 2002 Cancer Epidem Biomarkers Prev 11:1285 AJ Duffield-Lillico et al 2003 Br J Urol 91:608 AJ Duffield-Lillico et al 2003 J Nat Cancer Inst 95:1477 J N^ve 1995 J Trace Elements Med Biol 9:65 Panel on Dietary Antioxidants and Related Compounds 2000 Dietary Reference Intakes for Vitamin C, Vitamin E, Selenium and Beta-Carotene and other Carotenoids. National Academy Press Washington DC KE Burke, GF Combs Jr, EG Gross, KC Bhuyan, H Abu-Libdeh 1992 Nutr Cancer 17:123 BC Pence, E Pelier, DM Dunn 1994 J Invest Dermatol 102:759 AM Diamond, P Dale, JL Murray, DJ Grdina 1996 Mutat Res 356:147 Y Kise et al 1991 Nutr Cancer 16:153 GF Combs Jr, LC Clark, BW Tumbull 2001 Proc 7'* Internat Symp Selenium Biol Med 152 J Lii, C Jiang 2005 Antioxidants Redox Signaling 7:1715 M Berggren et al 1996 Anticancer Res 16:3459 GPowisetal 1996^nricancerZ)n
Selenium as a cancer preventive agent 86. 87. 88. 89. 90. 91. 92. 93. 94. 95. 96. 97. 98. 99. 100. 101. 102. 103. 104. 105. 106. 107. 108. 109. 110. 111. 112. 113. 114. 115. 116. 117. 118. 119. 120. 121. 122. 123. 124. 125. 126. 127. 128. 129. 130. 131. 132. 133. 134.

263

H Mork et al 2000 Nutr Cancer 37:108 OH Al-Taie et al 2004 Nutr Cancer 48:6 D Behne et al 1997 Biomed Environ Sci 10:340 S Apostolou et al 2004 Oncogene 23:5032 E Kumaraswamy et al 2000 J Biol Chem 275:35540 YJ Hu et al 2001 Cancer Res 61:2307 PC Raich, J Lu, HJ Thompson, GF Combs Jr 2001 Cancer Invest 19:540 LN Vemie 1987 Proc 3rd Internat Symp Selenium Biol Med GF Combs Jr, JE Spallholz, OA Levander, JE Oldfield (Eds) AVI Publ Co Westport, CT 1074 G Spyrou, M Bjomstedt, S Skog, A Holmgren 1996 Cancer Res 56:4407 M Bjomstedt, S Kumar, A Holmgren 1992 J Biol Chem 267:8030 L Wu, J Lanfear, PR Harrison 1995 Carcinogenesis 16:1579 M Bjomstedt, S Kumar, L Bjorkhem, G Spyrou, A Holmgren 1997 Biomed Environ Sci 10:271 LN Vemie, CJ Hamburg, WS Bont 1981 Cancer Lett 14:303 J Lanfear, JJ Flemming, L Wu, G Webster, PR Harrison 1994 Carcinogenesis 15:1387 BC Pence, M Stewart, L Walsh, G Cameron 1996 Proc 6th Internat Symp Selenium Biol Med ILSI, Beijing 82 KA Poirier, JA Milner 1983 J Nutr 113:2147 PR Harrison et al 1997 Biomed Environ Sci 10:235 C Jiang, Z Wang, H Ganther, J Lu 2002 Mol Cancer Ther 1:1059 J La 2001 Adv Exp Med Biol'^91:n\ J Lii, C Jiang 2001 Nutr Cancer 40:64 J Lii, Pei, C Ip, D Lisk, H Ganther, HJ Thompson 1996 Carcinogenesis 17:1903 M Kaeck et al 1997 Biochem Pharmacol 53:921 WZhong,TDOberley2001 Cancer Res 61:107\ C Jiang, Z Wang, H Ganther, J Lii, 2002 Mol Cancer Therap 1:1059 J Lu, M Kaeck, C Jiang, AC Wilson, HJ Thompson 1994 Biochem Pharmacol 47:1531 Z Zhu et al 1996 Biol Trace Elem Res 54:123 C Ip, H Ganther 1990 Cancer Res 50:1206 C Ip, H Ganther \99\Carcinogenesis 12:365 C Ip, HE Ganther 1992 Carcinogenesis 13:1167 C Ip, HE Ganther 1992 JInorgan Biochem 46:215 C Ip, C Hayes, RM Budnick, HE Ganther 1991 Cancer Res 51:595 S Vanhanavikit, C Ip, HE Ganther 1993 Xenobiotica 23:731 J L u e t a l l 9 9 5 Biochem Pharmacol 252:119 C Ip, S Vadhanavkit, H Ganther 1995 Carcinogenesis 16:35 C Ip, Z Zhu, HJ Thompson, D Lisk and HE Ganther 1999. Anticancer Res 19:2875 Z Wang, C Jiang, H Ganther, J Lu 2001 Cancer Res 61:7171 Z Wang, C Jiang, J Lu 2002 Mol Carcinog 34:113 C Jiang, Z Wang, H Ganther, J Lu 2001 Cancer Res 61:3062 T Kim, U Jung, DY Cho, AS Chung 2001 Carcinogenesis 22: 559 H Hu, C Jiang, G Li, J Lu 2005 Carcinogenesis 26:1374 R Sinha, D Medina 1997 Carcinogenesis 18:1541 R Sinha et al 1999 Cancer Lett 146:135 HE Ganther 1999 Carcinogenesis 20:1657 C Jiang, H Ganther, J Lu 2000 Mol Carcinog 29:236 SD Cho et al 2004 Mol Cancer Ther 3:605 Y Dong, SO Lee, H Zhang, J Marshall, AC Gao, C Ip 2004 Cancer Res 64:19 H Zhao, ML Whitfield, T Xu, D Botstein, D Brooks 2004 Mol Biol Cell 15:506 AC Wilson, HJ Thompson, PJ Schedin, NW Gibson, HE Ganther 1992 Biochem Pharmacol 43:1 \31 MS Stewart, RL Davis, LP Walsh, BC Pence 1997 Cancer Lett 117:35

264

Selenium: Its molecular biology and role in human health

135. 136. 137. 138. 139. 140. 141. 142. 143. 144. 145. 146. 147. 148. 149. 150. 151. 152. 153. 154. 155.

HE Ganther, C Corcoran 1969 Biochem 8:2557 PL Bergad, WB Rathbun \9^6 Curr Eye Res S:9\9 R Gopalakrishna, U Gundimeda, ZH Chen \991&Arch Biochem Biophys 348:25 R Gopalakrishna, ZH Chen, AU Gundimed 1997b Arch Biochem Biophys 348:37 C Jiang, W Jiang, C Ip, H Ganther, J. Lu 1999 Mol Carcinogenesis 26(4):213 AV Gasparian et al 2002 Mol Cancer Therap 1:1079. J Chaudiere, O Courtin, J Leclaire 1992 Arch Biochem Biophys 296:328 JE Spallholz 1997 Biomed Environ Sci 10:260 MS Stewart, JE Spallholz, KH Neldner, EC Pence 2000 Free Radic Biol Med 26: 42 JE Spallholz, BJ Shriver, TW Reid 2001 Nutr Cancer 40:34 C Ip, G White 1987 Carcinogenesis 8:1763 C Ip, C Hayes 1989 Carcinogenesis 10:921 HJ Thompson, C Ip 1991 Biol Trace Elem Res 30:163 B Siwek et al 1994 Arch Toxicol 68:246 C Ip, S Vadhanavkit, H Ganther 1995 Carcinogenesis 16:35 C Ip, Z Zhu, HJ Thompson, D Lisk, HE Ganther 2000 Anticancer Res 19:2875 BS Reddy, S Sugie, H Maruyama, K El-Bayoumy, P Marra 1987 Cancer Res 47:5901 Z Ronai et al 1995 Int N Cancer 63:428 V Adler et al 1996 Carcinogenesis 17:1849 BProkopczyk, etal 1996 Carcinogenesis 17:7^9 C Ip, H Ganther 1993 Selenium in Biology and Human Health RF Burk (Ed) SpringerVerlag New York 170 K El-Bayoumy, CV Rao, BS Reddy 2001 Nutr Cancer 40:18 C Ip et al 1994 Carcinogenesis 15:187 BS Reddy etal 1997 J Nat Cancer Inst S9:506 C Ip, H Thompson, HE Ganther 1994 Carcinogenesis 15:2879 C Ip, HJ Thompson, HE Ganther 1998 Anticancer Res IS:9

156. 157. 158. 159. 160.

Chapter 23. Peering down the kaleidoscope of thiol proteomics and unfolded protein response in studying the anticancer action of selenium Ke Zu, Yue Wu, Young-Mee Park and Clement Ip Departments of Cancer Chemoprevention [KZ, YW, CI] and Cellular Stress Biology [YMPJ, Roswell Park Cancer Institute, Buffalo, NY 14263, USA

Summary: A monomethylated selenium metabolite called methylseleninic acid (MSA) has recently been shown to cause global thiol redox modification of proteins. These changes represent a form of cellular stress due to protein misfolding or unfolding. When this occurs in the endoplasmic reticulum (ER), a process known as unfolded protein response (UPR) is orchestrated to repair the damage or commit the cells to apoptosis if the rescue effort becomes inadequate. Treatment of PC-3 human prostate cancer cells with MSA leads to activation of three signature ER stress transducers and increased expression of UPR target genes, as exemplified by GRP78 and GADD153. GRP78 is part of the damage control mechanism, while GADD153 is a transcription factor associated with growth arrest and apoptosis. Transfection with GRP78 largely negates the apoptotic effect of MSA and GADD153 induction because an abundance of GRP78 allows cells to cope better with ER stress. Conversely, knocking down GRP78 by siRNA magnifies the cell growth arrest effect of MSA and GADD153 expression. Collectively, the research findings support the idea that UPR plays an important role in mediating the anticancer activities of selenium. Introduction Numerous dietary and synthetic compounds have been shown to be useful for cancer chemoprevention. The chemical structures of these compounds are very diverse, but most of them express their anticancer effects through growth arrest and apoptosis. Mechanistic studies reveal that certain cell cycle and death effector molecules are commonly implicated, even though these preventive agents differ greatly in their lipophilicity, reactivity, or association with known receptors. The above information suggests to us that there might be some innate events triggered by these chemicals to either activate or suppress a defined set of signaling cascades which ultimately lead to modulation of cell growth. Cellular stress could conceivably be a causal

266

Selenium: Its molecular biology and role in human health

phenomenon. There are many forms of cellular stress specific to different organelles - nucleus, mitochondria, endoplasmic reticulum, lysosome, etc [1]. The specificity is related to the kinds of damage sustained by each organelle, the sensors deployed to recognize the damage, the local as well as the distal damage response. Broadly speaking, stress response is characterized by a balance between rescue/survival and death; the latter is activated when stress is too severe or prolonged. The fate of a cell population under stress depends on which side the balance tilts. The rationale for using metliylseleninic acid The anticancer activities of selenium have been well documented in both in vivo and in vitro experimentations. Selenomethionine is a first-generation organic selenium compound that was studied extensively in the early 1990s. As proposed originally by Ip and coworkers, the metabolism of selenomethionine to methylselenol (CHsSeH) is critical for the chemopreventive effect [2]. hi mammals, liver and kidney are the major organs for selenium metabolism. The pathway of converting selenomethionine to methylselenol requires five different enzymatic steps. Tissues such as breast and prostate have a low capacity to produce methylselenol. For this reason, cultured breast or prostate cells generally are growth inhibited by selenomethionine only when it is present at levels of 100-400 jxM in the medium. These concentrations of selenium are much higher than the physiological concentrations of 3-5 i^M found in plasma. Methylselenol is highly reactive, and cannot be tested as is in vitro. In order to overcome this obstacle, a stable metabolite, methylseleninic acid (CHsSeOaH, abbreviated to MSA), was developed specifically for cell culture studies [3,4]. MSA is water-soluble and is taken up into cells very easily. Once inside, it is quickly reduced to methylselenol by GSH or NAPDH through non-enzymatic reactions. In vitro studies with human breast or prostate cancer cells showed that exposure to MSA results in cell cycle block, DNA synthesis suppression, and apoptosis [5-7]. Are these effects due to stress, and if so, what kind of stress is caused by MSA? This chapter summarizes our most recent research aimed at addressing the above question. Monomethylated selenium as a protein redox modulator At physiological pH, methylselenol is present as anionic methylselenolate. Most protein cysteine residues would not react with methylselenolate because they generally exhibit a pKa value of 8.0-8.5. But some protein cysteine residues exist as thiolate anion at neutral pH, because their pKa value is lowered as a result of the influence of neighboring nucleophilic groups, and the thiolate is often stabilized by salt bridges to positively charged moieties. By virtue of the negative charge, thiolates have enhanced

Selenium andER stress

267

reactivity and are called "reactive thiols". Under oxidative stress condition, a reactive thiol could lose an electron to become a thiyl radical which can then react with methylselenolate to form a selenenylsulfide intermediate (Figure lA). The latter is susceptible to attack by a second thiol to form an intramolecular disulfide. As proposed previously by Ganther [8], methylselenolate also readily reacts with protein disulfides and converts them to sulfliydryl groups (Figure IB). Depending on the reduction potential of the redox couple, it is apparent that interaction with methylselenolate could result in either a gain or a loss of reactive thiols.

CHjSe-

•SHjo]^ |—S-

•SH

•SH

|— s-

J

^

— S-SeCH,

-^ —S

B GSH I + CH3;Se—S

•SH

CH,Se

^X

•S-SeCH,

GSH GSSG

-SH — S-SG

V_y^

•SH h-SH

Figure 1. Biochemistry of protein thiol redox modification by methylselenolate.

Recently, the time-dependent change in thiol proteome profile was studied in MSA-treated PC-3 human prostate cancer cells by using a reactive thiol display method linked to MALDI-TOF and ESI-tandem mass spectrometry [9]. An outline of this procedure is illustrated in Figure 2. Briefly, proteins with reactive thiols are labeled with a thiol specific reagent, iodoacetamide, which is biotinylated (the abbreviation of this reagent is BIAM). After separation by 2-D gel electrophoresis, the protein spots are visualized by peroxidase-conjugated streptavidin. The signal intensity of all protein spots in the control and treated samples is then quantified and analyzed. Following in-gel trypic digestion, the proteins are identified by mass spectrometry. Suffice it to note that this methodology is limited to the detection of high abundance proteins. With few exceptions, low abundance proteins, proteins with extreme pi, or proteins with poor solubility, are generally not well resolved. In reality, the number of redox-sensitive proteins is likely to be

268

Selenium: Its molecular biology and role in human health

underestimated. Of the total of 194 BIAM-labeled proteins on the display, nearly half showed either a loss or a gain of reactive thiols. The kinetics of change was examined by clustering analysis with the use of the Self Organizing Map (SOM) algorithm. Essentially, the proteomics dataset was transformed to a plot of the treatment to control signal ratio at each time point. Four distinct patterns of redox changes emerged. Cluster 1, which consists of 100 proteins or ~51% of the total, exhibited very little change in reactive thiol labeling intensity throughout the course of MSA treatment.

(J?)—s (H)— sx Protein solution

X BIAM labeling

T ( P ? ) — S-IAM-Biotin

(P2)—sx X 2D Gel/ HRP-conjugated streptavidin

Ohr

24 hr 0.5 hr

12 hr 6hr ^ri ' • ' ^

3i

aft vO'

A^^

I

MALDI-TOF / ESI-tandem mass spectrometry

Protein Identification

Figure 2. Schematics of reactive thiol labeling by BIAM and the identification of proteins by mass spectrometry. The triangles and circles denote the same protein spots of which the intensity of BIAM labeling is decreased or increased, respectively, subsequent to MSA treatment. Adapted from [9].

Selenium and ER stress

269

Cluster 2, which consists of 60 proteins or - 3 1 % of the total, responded rapidly with an averaged 8-fold decrease of thiol labeling intensity as early as 30 min after exposure to MSA. The decrease was stable for at least 24 hr. Cluster 3 consists of 19 proteins or 10% of the total; this cluster showed a gradual loss of reactive thiols starting at 30 min after MSA treatment and reached a nadir at 3 hr. Interestingly, it recovered rather quickly, and the label intensity returned towards the control value at 6 hr. Cluster 4 consists of 15 proteins or 8% of the total; in this cluster there was a significant gain of reactive thiols between 2 to 3 hr, followed by a rapid recovery to normal at 6 hr. Based on the clustering analysis, it is evident that MSA causes protein redox changes almost instantaneously. Some of the damage is repaired rather quickly, but some persists much longer, e.g. cluster 2. Why are the cluster 2 proteins so resistant to repair, and what might be the lingering effect of this occurrence? Presently we do not have an answer to either question. Biological implication of redox-modified proteins Not all thiol-reactive proteins were successfully identified by mass spectrometry; only 85 out of a total of 194 were confirmed through the NCBI database, and 39 of them were in clusters 2 to 4. These proteins are found to be distributed in various subcellular compartments, including the nucleus, cytosol, mitochondria, endoplasmic reticulum, and lysosome. Their ubiquity is consistent with the pervasive nature of a small selenium metabolite capable of striking many sensitive targets anywhere in a cell. Redox regulation of proteins has a profound effect on their activities. The concept of selenium modifying thiol/disulfide interchange with accompanying alterations of activities has been reported previously with protein kinase C and p53 [10,11]. Without doing similarly sophisticated biochemistry with all 39 proteins in our study, there is no way to predict how their activities will be affected. In looking at this problem fi*om a different angle, we are intrigued by the idea that thiol redox modification is likely to result in protein misfolding or unfolding. Newly synthesized proteins are particularly vulnerable before they are properly folded in the endoplasmic reticulum (ER). Thus it is highly plausible that MSA may produce ER stress due to the accumulation of misfolded/unfolded proteins in the lumen of the ER. To prove that this phenomenon is actually taking place in the ER will not be an easy task. However, there is a fairly extensive literature on ER stress and ER stress response which we could borrow to study the effect of selenium. ER stress and unfolded protein response In addition to redox alterations, causal factors of ER stress include suppression of protein glycosylation and disruption of calcium homeostasis [12]. These perturbations all lead to protein misfolding or unfolding, which

270

Selenium: Its molecular biology and role in human health

in turn initiates a series of transducer pathways as a self-protective mechanism. This so called unfolded protein response (UPR) is characterized by an immediate stoppage of new protein synthesis and growth arrest, followed by adaptive survival, or apoptosis if the rescue effort is exhausted [12-14]. The UPR is mediated primarily by one protein chaperone, BiP/GRP78, and three transmembrane ER stress transducers: PKR-like ER kinase (PERK), activating transcription factor 6 (ATF6), and inositol requiring 1 (IREl), as shown in Figure 3. to the unstressed ER, BiP/GRP78 binds to the ER luminal domains of the transducers and keeps them inactive in sequestration. Upon sensing the accumulation of unfolded proteins, BiP/GRP78 dissociates from its clients and translocates to the ER lumen to help protein folding [15,16]. Once released from sequestration, the three fransducers undergo activation through different mechanisms and cause the up-regulation of a number of transcription factors.

ER

MasUi-

reuiilalor

GRP78

i:U Ndi'ss ll'illlsdllL'i'l'S

Transcription factors

Nucleus

Induction of ER stress target genes

Figure 3. Schematics of activation of ER stress transducer pathways. PERK is activated by oligomerization and autophosphorylation. Activated PERK phosphorylates eukaryotic initiation factor 2a (eIF2a), thereby shutting off general protein translation [17]. On the other hand, phospho-

Selenium and ER stress

271

eIF2a increases the protein translation of ATF4, an ER stress transcription factor. After dissociation from BiP/GRP78, ATF6 translocates to the Golgi, where the active form, p50ATF6, is generated by proteolysis [18,19]. As a transcription factor, p50ATF6 binds to the ER stress response element (ERSE), resulting in the induction of ER stress target genes, including BiP/GRP78, XBPl, CHOP/GADD153, and p58"''^ [20-22]. Similar to PERK, IREl is fully activated by dimerization and autophosphorylation. The sitespecific endoribonuclease activity of IREl mediates the removal of a 26nucleotide intron from XBPl mRNA [23-25]. The spliced form of XBPl subsequently binds to the ERSE and up-regulates the transcription of BiP/GRP78 and EDEM (ER degradation-enhancing alpha-mannosidase-like protein), the latter is meant to accelerate the degradation of misfolded proteins [26]. In summary, a circuitry of signaling molecules is functioning cooperatively to alleviate the burden of unfolded/misfolded proteins, increase the processing capacity of the ER, and accelerate protein degradation. When the survival response fails to adapt under severe ER stress, cells will eventually undergo self-destruction, although the mechanism behind this decision has not been elucidated. Induction of UPR signaling by MSA All three UPR fransducer pathways, PERK-eIF2a, ATF6 and IREl-XBPl, are up-regulated rapidly when PC-3 cells are freated with MSA [27]. In general, the increase peaks between 3-6 hr and gradually returns to the basal level by 24 hr. Thus the time course of UPR activation is in step with that of protein redox changes. As expected, the expression of UPR target genes, such as BiP/GRP78 and CHOP/GADD153, is also significantly stimulated. BiP/GRP78, together with PERK-eIF2a, are the survival/rescue molecules. BiP/GRP78 is charged with the task of maintaining proteins in a foldingcompetent state, while PERK-eIF2a signaling is meant to block new protein synthesis in order to reduce the accumulation of misfolded/unfolded proteins. If the rescue undertaking is insufficient to mend the injury, the decision to commit to apoptosis may be triggered. CHOP/GADD153 is a key growth arrest and pro-apoptotic franscription factor that is closely identified with ER stress. Additionally, caspase-12 and caspase-7 are the unique ER stressassociated caspases which might play a role in pushing cells towards apoptotic death [28-30]. It is interesting to find that a higher concenfration of MSA is required to elicit the apoptotic markers (CHOP/GADD153, caspase-12 and caspase-7) compared to the rescue markers (phospho-PERK, phospho-eIF2a, BiP/GRP78 and GRP94). In general, the apoptotic markers increase proportionally over a wide range of MSA doses, while the rescue markers fail to keep pace with an increasing challenge from MSA [31]. Cells clearly have a calibrated capacity to manage sfress. It is thus reasonable that

272

Selenium: Its molecular biology and role in human health

low doses of MSA stimulate preferentially the rescue arm of the UPR, while high doses of MSA lead to the assembly of the apoptotic machinery. Consequence of BiP/GRP78 over-expression or knockdown in modulating the effect of MSA Given that BiP/GRP78 is like the rheostat of the UPR circuit board, this molecule was transiently over-expressed in PC-3 cells to increase the availability of BiP/GRP78 in the ER [31]. The ability of MSA to up-regulate phospho-PERK, phospho-eIF2a, GRP94, CHOP/GADD153, caspase-12, caspase-7 and cleaved PARP is significantly muted in the BiP/GRP78 overexpressing cells (Figure 4A). The apoptotic response to MSA is dampened as well (Figure 4B). A generous supply of free BiP/GRP78 allows cells to cope better with ER stress, thereby improving the odds for survival and negating the commitment to apoptotic death. The inference is that depending on how well cells are able to up-regulate BiP/GRP78, different cell types may have different thresholds for the apoptosis tripwire. With the use of a coimmunoprecipitation assay, we also discovered that MSA stimulates the dissociation of BiP/GRP78 from caspase-7. To our knowledge, this is a new mechanism by which selenium recruits caspases in addition to the more traditional way of activating the intrinsic and extrinsic death signaling pathways [7,32]. In summary, the data from the BiP/GRP78 over-expression experiments strongly suggest that MSA acts through UPR, at least in part, to induce programmed cell death. no 5 |JM selenium selenium

B. • rrock-transfected

transfection O phospho-PERK

16 - • GRFV8-transfected

0.81 12

phospho-elF2a 0.66 GRP94

a 8

CHOP/GADD153

° 4

cleaved caspase-12 0.62

0-

cleaved caspase-7 0.40 cleaved PARP 0.69

Figure 4. Over-expression of GRP78 attenuates the induction of ER stress markers (panel A) and apoptosis (panel B) by MSA. * Significantly different from the respective control value (P<0.05). Adapted from [31].

The flipside to the above approach is to suppress the induction of BiP/GRP78 by RNA interference in order to intensify the stress level in the

273

Selenium andER stress

ER. Interestingly, the three transducer pathways respond differently to BiP/GRP78 knockdown [27]. In the BiP/GRP78 siRNA transfected cells, the increase of phospho-PERK and phospho-eIF2a by MSA is attenuated, while the induction of p50ATF6 is maintained for a longer period of time. In contrast, IRE 1-mediated XBPl splicing is not affected by reduced expression of BiP/GRP78. The above information suggests that the signaling intensity of each transducer could be fine-tuned depending on BiP/GRP78 availability. In the presence of MSA, the expression of CHOP/GADD153, which is implicated in growth arrest and apoptosis, is raised even higher by BiP/GRP78 knockdown (Figure 5A). The effect of MSA on several cell cycle regulatory molecules (p21^^, CDKl and CDK2) is also magnified in a manner consistent with enhanced cell growth arrest (Figure 5B). Additional experiments strongly indicated that CHOP/GADD153 up-regulates the expression of p21^ in a p53-independent manner (PC-3 cells are p53-null). Collectively, the above findings support the idea that UPR plays a positive role in mediating the induction of growth arrest by MSA.

B.

A. MSA

6hr

0 lir

+

Scramble siRNA + G R P 7 8 SiRNA

-

C- 40%

+

+

5 20%-

+

E-

c 5 -20%

p2l^'" ( DKI

GAI'DII

24 h r

"^ o%-

CHOI'

( l)K2

+

• Scramble siRNA IGRP78siILNA

12 h r

D —-

"S -40% -60%

L

GO/Gl

S

G2/M

Figure 5. BiP/GRP78 knockdown enhances the effect of MSA on the expression of CHOP and cell cycle regulatory molecules (panel A) and cell cycle block (panel B). *P<0.05. Adapted from [27].

Food for thought ER stress has been studied mostly in neuropathology, such as Parkinson's disease and Alzheimer's disease [33]. Scanty information is available regarding UPR and cancer. As far as we are aware, this is the first in-depth perspective of how selenium, as a cancer chemopreventive agent, causes ER stress, which in turn triggers a cascade of molecular events leading to growth arrest and apoptosis. The key elements of this new concept are capsulized in the diagram of Figure 6. Our laboratory had previously investigated the gene expression profile in MSA-treated PC-3 cells by oligonucleotide array

274

Selenium: Its molecular biology and role in human health

analysis. Hundreds of genes are affected, among which is a large group of cell cycle regulatory genes [6,34]. To date, we have made the connection linking CHOP/GADD153, p21^'^, CDKl and CDK2 in the context of UPR. This is probably just the tip of the iceberg. The ER is also equipped with the machinery for apoptosis. As noted earlier, caspase-12 is an ER specific caspase that is activated by the IRE1-TRAF2 complex [28]. Caspase-7, -8 and -4 are also known to be recruited to the ER during stress [30,35-37]. Additionally, the IRE1-TRAF2 complex can bind to ASKl in order to activate JNK [38]. JNK/p38MAPK activation is intimately involved in stress-induced apoptosis. The death signal might be amplified by the crosstalk between the ER and the intrinsic mitochondrial pathway [12]. Admittedly, we have only touched on a small subset of stress response signals sensitive to selenium. There are many more stress transducers and downstream effectors which have yet to be uncovered. Redox Activities of Selenium

Protein Tliiol/Disulfide Interchange

Unfolded Protein Response

Modulation of ER Stress Target Genes and Downstream Effectors

Growth Arrest

Cell Death

Figure 6. A flow diagram showing the action of selenium from redox activities to the outcome of growth arrest and cell death.

The induction of UPR by selenium is unlikely to be cell type specific in the whole organism. That is to say, proteins in normal cells are vulnerable to redox modification by selenium as well. Here lies the caveat. Different cells may have different abilities to manage stress. The selectivity of selenium as a chemopreventive agent has been well documented [4]. A recent study reported that pre-administration of selenium increases the therapeutic

Selenium and ER stress

275

efficacy of irinotecan in the nude mice tumor xenograft model [39]. On the other hand, selenium is highly protective of normal cells, and is able to overcome the dose-limiting toxicity of the drug. The above finding implies that selenium may favor survival response in normal cells, but facilitates apoptotic response in cancer cells, hideed, the protective role of preconditioned ER stress response against cytotoxicity has been reported in normal cells [40,41]. Many factors may impact on ER stress response to selenium. For example, the microenvironment may determine whether the outcome of UPR is survival or death. Hypoxia is a known inducer of ER stress [42,43]. The hypoxic condition of a solid tumor could sensitize cancer cells to selenium. Our laboratory has preliminary data (unpublished) showing that hypoxia significantly enhances selenium induction of apoptosis. Genetic background may be another factor in determining whether the balance tips towards survival or death in response to the same stress signal. There has been a flurry of activities investigating the ability of selenium to inhibit the expression of androgen and estrogen receptors [44-48]. Androgen and estrogen signalings are critically important in the development of prostate and breast cancers, respectively, because they are positive regulatory pathways for differentiation and growth. A distinguishing feature of the down-regulation of both hormone receptors by selenium at the mRNA level is that it happens very quickly, within a matter of hours. Current effort is focused on identifying a minimal promoter region responsible for the inhibitory effect of selenium, and characterizing the transcription factor(s) associated with this region. Modulation of transcription factor expression is a hallmark of cellular stress response. It is not at all uncommon to find one transcription factor increasing or decreasing the expression of a second transcription factor, and that the process is repeated down the line. Shutting off the differentiation and growth stimulatory pathways in times of stress would make a great deal of sense. The answer to the question of whether selenium-mediated down-regulation of androgen and estrogen receptors is or is not related to stress could be rather illuminating. References 1. KF Ferri, G Kroemer 2001 Nature Cell 5jo/ 3:E255-E263 2. C Ip 1998 yA^«i/-128:1845 3. C Ip, HJ Thompson, Z Zhu, HE Ganther 2000 Cancer Res 60:2882 4. C Ip, Y Dong, HE Ganther 2002 Cancer Metastasis Rev 21 il&l 5. Y Dong, HE Ganther, C Stewart, C Ip 2002 Cancer Res 62:708 6. Y Dong, H Zhang, L Hawthorn, HE Ganther, C Ip 2003 Cancer Res 63:52 7. K Zu, C Ip 2003 Cancer Res 63:6988 8. HE Ganther 1999 Carcinogenesis 20:1657 9. EM Park, KS Choi, SY Park, K Zu, Y Wu et al 2005 Cancer Genom Proteom 2:25 10. R Gopalakrishna, Z-H Chen, U Gundimeda 1997 Arch Biochem Biophys 348:37 11. ML Smith, JK Lancia, TI Mercer, MR Kelley, C Ip 2004 Anticancer Res 24:1401

276

Selenium: Its molecular biology and role in human health

1 2 . x Shen, K Zhang, RJ Kaufman 2004 J Chem Neuroanat 28:79 13. Y Ma, LM Hendershot 2004 J Chem Neuroanat 28:51 14. H Kadowaki, H Nishitoh, H Ichijo 2004 J Chem Neuroanat 28:93 15. A Bertolotti, Y Zhang, LM Hendershot, HP Harding, D Ron 2000 Nat Cell Biol 2:326 16. CY Liu, HN Wong, JA Schauerte, RJ Kaufman 2002 J Biol Chem 277:18346 17. HP Harding, Y Zhang, D Ron 1999 Nature 397:271 18. X Chen, J Shen, R Prywes 2002 J Biol Chem 277:13045 19. T Okada, K Haze, S Nadanaka, H Yoshida, NG Seidah et al 2003 J Biol Chem 278:31024 20. H Yoshida, K Haze, H Yanagi, T Yura, K Mori 1998 J Biol Chem 273:33741 21. H Yoshida, T Okada, K Haze, H Yanagi, T Yura et al 2000 Mol Cell Biol 20:6755 22. R van Huizen, JL Martindale, M Gorospe, NJ Holbrook 2003 J Biol Chem 278:15558 23. H Yoshida, T Matsui, A Yamamoto, T Okada, K Mori 2001 Cell 107:881 24. M Calfon, H Zeng, F Urano, JH Till, SR Hubbard et al 2002 Nature 415:92 25. K Lee, W Tirasophon, X Shen, M Michalak, R Prywes et al 2002 Genes Dev 16:452 26. H Yoshida, T Matsui, N Hosokawa, RJ Kaufman, K Nagata, K Mori 2003 Dev Cell 4:265 27. K Zu, T Bihani, A Lin, YM Park, K Mori, C Ip 2006 Oncogene in press 28. T Nakagawa, H Zhu, N Morishima, E Li, J Xu, BA Yankner, J Yuan 2000 Nature 403:98 29. T Yoneda, K Imaizumi, K Oono, D Yui, F Gomi et al 2001 J Biol Chem 276:13935 30. RV Rao, E Hermel, S Castro-Obregon, G del Rio, LM Ellerby et al 2001 J Biol Chem 276:33869 31. Y Wu, H Zhang, Y Dong, Y-M Park, C Ip 2005 Cancer Res 65:9073 32. H Hu, C Jiang, C Ip, YM Rustum, J Lu 2005 Clin Cancer Res 11:2379 33. J Lehotsky, P Kaplan, E Babusikova, A Strapkova, R Murin 2003 Physiol Res 52:269 34. H Zhang, Y Dong, H Zhao, JD Brooks, L Hawthorn et al 2004 Cancer Genom & Proteom 2:97 35. A Jimbo, E Fujita, Y Kouroku, J Ohnishi, N Inohara et al 2003 Exp Cell Res 283:156 36. DG Breckenridge, M Nguyen, S Kuppig, M Reth, GC Shore 2002 Proc Natl Acad Sci USA 99:433 37. T Katayama, K Imaizumi, T Manabe, J Hitomi, T Kudo, M Tohyama 2004 J Chem Neuroanat 28:67 38. F Urano, X Wang, A Bertolotti, Y Zhang, P Chung, HP Harding, D Ron 2000 Science 287:666 39. S Cao, FA Durrani, YM Rustum 2004 Clin Cancer Res 10:2561 40. CC Hung, T Ichimura, JL Stevens, JV Bonventre 2003 J Biol Chem 278:29317 41. K Bednard, N MacDonald, J Collins, A Cribb 2004 Basic Clin Pharmacol Toxicol 94:124 42. C Koumenis, C Naczki, M Koritzinsky, S Rastani, A Diehl et al 2002 Mol Biol Cell 22:7405 43. S Tajiri, S Oyadomari, S Yano, M Morioka, T Gotoh et al 2004 Cell Death Differ 11:403 44. Y Dong, SO Lee, H Zhang, J Marshall, AC Gao, C Ip 2004 Cancer Res 64:19 45. Y Dong, H Zhang, AC Gao, JR Marshall, C Ip 2005 Mol Cancer Ther 4:1 46. SO Lee, N Nadiminty, XX Wu, W Lou, Y Dong et al 2005 Cancer Res 65:3487 47. YM Shah, A Kaul, Y Dong, C Ip, B Rowan 2005 Breast Cancer Res Treat 92:239 48. YM Shah, M Al-Dhaheri, Y Dong, C Ip, FE Jones, BG Rowan 2005 Mol Cancer Ther 4:1239

Chapter 24. Genetic variation among selenoprotein genes and cancer Alan M. Diamond and Rhonda L. Brown Department of Human Nutrition, University of Illinois at Chicago, Chicago, IL 60612, USA

Summary: Selenium is an essential trace element shown to prevent cancer in animals. In humans, selenium status has been shown to be inversely associated with cancer risk, and clinical studies have indicated a potential cancer prevention benefit fi-om supplementing diets with low levels of selenium as well. It is likely that many of the benefits of selenium are associated with consequential effects on the levels or activity of seleniumcontaining proteins, or selenoproteins, which contain selenium as the amino acid selenocysteine. Twenty five selenoproteins are encoded by the human genome. Genetic variations among these genes may be associated with cancer risk, as has been indicated for the anti-oxidant selenoprotein glutathione peroxidase 1 (GPx-1). Genetic variation among selenoprotein genes also has facilitated the recognition of allelic loss during cancer development, where losing one of two gene copies presumably reduces the ultimate levels of the corresponding protein and the loss of a protective function. Genetic variations among other selenoprotein genes have been documented, but the biological consequences, if any, have yet to be defined. Introduction The human genome project has provided information about the genetic differences among individual human beings. Each member of our species can be distinguished by approximately 1% of the size of the genome, or 30 million nucleotide positions. The vast majority of these variations are likely to be without biological consequence, as they occur in regions of DNA without apparent function, or they do not influence the levels or regulation of the final gene-product. However, many of these genetic polymorphisms will account for the differences in apparent and non-apparent traits that distinguish among the individuals that make up the collective population. Some fraction of these polymorphisms is likely to have a significant impact on human health, influence disease susceptibility and our response to the environment. Selenium has been implicated in a broad range of human health issues, including fertility, immune response, aging and susceptibility to viral

278

Selenium: Its molecular biology and role in human health

infection [1-3]. In addition, dietary intake of selenium has been shown to influence cancer incidence in animal studies, and a similar role has been indicated in humans based on epidemiological studies showing an inverse relationship between selenium in the diet and cancer risk. It is likely that many of the biological properties of selenium are mediated through its role as a constituent of selenoproteins, and polymorphic variations among selenoprotein genes are likely to contribute to the risk of cancer. The association of these particular genetic variations with increased cancer risk and the value of using polymorphic positions within selenoproteins to follow allelic loss during cancer development is the focus of this chapter. Selenium and Cancer Supplementation of the diet of rodents with low, non-toxic levels of selenium has been shown to effectively reduce cancer incidence [4,5]. The published reports have included mice, rats and hamsters as experimental subjects and have shown selenium supplementation to be effective as a chemopreventive for most organs examined. Remarkably, selenium has been shown to be effective in preventing cancer initiated by a wide variety of carcinogens, including both chemicals and radiation. The volume and consistency of this literature has raised significant interest in the use of selenium as a chemopreventive in humans as well. There is considerable human epidemiology indicating a role for selenium in cancer prevention. These studies have generally revealed an inverse relationship between dietary selenium levels and the risk of specific cancers, although several studies have failed to establish that relationship. While some reports have indicated that selenium status is an independent indicator of general cancer risk and mortality [6,7], several others have shown a risk of specific types of cancers are associated with low selenium status. Low selenium status has been reported to be a risk factor for lung cancer in both a nested case-control study [8,9] and by meta analysis [10]. Another study, using a cross-sectional analysis in the study design, indicated that higher selenium levels were associated with reduced risk of esophageal adenocarcinoma among individuals diagnosed with Barrett's esophagus [11]. While both a British and an Austrian study could not find an association between selenium status and the risk of prostate cancer [12,13], several other studies have reported a significantly higher risk for men with the lowest selenium status [14,15], with another study reporting that low selenium status is associated with a four to five fold increase in the risk of prostate cancer [16]. These results, indicating an inverse association between selenium intake and prostate cancer, were recently substantiated by meta analysis [17]. Similarly, several studies have indicated that selenium is protective against colon cancer. A case control study involving 1,048 incidence cases of colon

Genetic variation among selenoprotein genes and cancer

279

cancer and 688 population-based controls indicated a statistically significant inverse correlation between toenail selenium levels and the risk of colon cancer for both genders (OR, 0.42 at a P = 0.009) [18]. A significant inverse association between selenium levels and the incidence of large adenomatous polyps, after adjusting for confounding variables in patients less than 60 years of age (OR=0.17, p=0.029), has also been reported [19] and this conclusion was recently substantiated by an analysis of pooled data fi-om three independent studies [20]. Selenium intake was also inversely associated with bladder cancer risk in women, but not men [21]. In contrast to the studies summarized above, others have failed to find a benefit to higher selenium levels as it reflects the risk of cancers at other sites, including breast [22-24] and cervical [25], and basal cell carcinoma [26]. In addition to epidemiology indicating that reduced selenium levels may be associated with increased cancer risk, data has also been reported indicating the benefits of selenium supplementation in reducing cancer incidence. This had been presented for cancers of several types, including prostate [27,28] and gastrointestinal tract [15], a conclusion supported by a systematic review and meta-analysis of studies examining the effects of selenium supplementation on the incidence of gastrointestinal cancers [29]. Selenoproteins It is likely that many of the effects of selenium are mediated through its role as a constituent of selenium-containing proteins. Twenty five selenoproteins are present in the human genome 24 in the mouse) and all contain selenium as the amino acid selenocysteine (Sec) [30,31]. Sec is incorporated cotranslationally during selenoprotein synthesis in response to in-frame UGA codons in the mRNA for these selenoproteins, and Sec insertion requires dedicated translation factors including a sec tRNA and elongation factor, in addition to the RNA element in the 3'-untranslated portion of the mRNA that directs Sec incorporation in response to all in-frame UGA codons [32,33]. In mammalian cells, this process is highly regulated and responsive to selenium availability, both at the levels of RNA stability and translation [34,35]. Genetic evidence for a role for GPx-1 in cancer etiology Glutathione peroxidase (GPx) represents a family of four related selenoproteins, each one carrying a single selenium atom in the form of Sec at the enzyme's active site [36-38]. The family includes cytosolic GPx, GPx1; gastrointestinal GPx, GPx-2; plasma GPx, GPx-3 and the phospholipid hydroperoxide GPx, GPx-4 [39]. Of these, evidence indicating that allelic variants of GPx-1 are associated with the risk of cancer has been presented. Several polymorphic positions within the GPx-1 gene were reported by Moscow et al. [40], and that which determines whether a proline or leucine

280

Selenium: Its molecular biology and role in human health

is present in the corresponding protein at codon 198 has been imphcated with cancer risk. Ratnasinghe et al. examined this polymorphism among participants of a cancer prevention study conducted in Finland [41]. Analysis of DNA obtained from blood obtained prior to lung cancer diagnosis versus matched controls indicated an association between the leu allele and lung cancer risk [41]. This analysis revealed that individuals that were either heterozygous or homozygous for the leu allele were at greater risk of contracting lung cancer, with odds ratios of 1.8 and 2.3, respectively. The risk of particular subtypes of lung cancer was also examined, with the reported risk estimates for squamous cell carcinoma and adenocarcinoma being statistically significant, while that for small cell carcinoma of the lung was not. Similarly, a Japanese study examined GPx-1 genotypes obtained from blood among patients with bladder cancer and compared these to data obtained from a matched, control group [42]. In the examined population, there were no individuals with the leu/leu genotype; however, the pro/leu heterozygous genotype was shown to be at 2.6-fold greater risk for the disease as compared to those with the pro/pro genotype. While these two studies reported a significant association between the leu allele and cancer risk, others have failed to detect such a relationship for other cancer types, including cancers of the colon, breast, prostate and skin [43-45]. It is noteworthy that the pro/leu allele has also been associated with a significant increase in carotid artery disease, as compared to the pro/pro allele [46]. An additional polymorphism in the GPx-1 gene has been examined; an inframe variation in the number of GCG repeats occurring at positions 338349 of the corresponding cDNA and resulting in either 5,6 or 7 alanines in the protein product. Using data obtained from the Ontario Familial Breast Cancer Registry, Knight et al. reported a significant association with the risk of breast cancer in premenopausal women carrying the five-alanine repeat encoding allele, as compared to those bearing the 6- or 7-alanine repeat poljonorphism [44]. The significance of this association was lost when postmenapausal women were included in the analysis. In contrast, no association was found between the number of alanine repeats in the GPx-1 gene and the risk of young onset prostate cancer, although some indication of increased risk among men who were homozygous for the 6 alanine-encoding allele was reported [47]. The data addressing the association between polymorphisms in the GPx-1 gene and cancer risk are summarized in Table 1. The two polymorphisms in the GPx-1 gene discussed above, resulting in either a proline or leucine at codon 198, or variations in the number of alanines in the repeats, occur within the coding region of the GPx-1 gene, and therefore may have consequences with regard to enzyme function. This has been investigated by generating GPx-1 expression constructs that encoded cDNAs differing only by a leucine or proline at codon 198 and

Genetic variation among selenoprotein genes and cancer

281

transfecting these into human breast carcinoma cells that express only marginal endogenous GPx-1 levels [48]. Analysis of the transfectants indicated that the resulting proline-containing allele produces higher GPx enzyme activity following the supplementation of the culture medium with selenium as compared to the leucine-containing protein. No correlation was reported between GPx-1 enzyme levels in erythrocytes and genotype in a Swedish population, although issues relating to other polymorphisms in the GPx-1 gene, selenium status and cell-type specificity were not addressed [49]. However, it is interesting that a comparison of the proteins generated using expression constructs producing either the alanine(6)/leul98 or the alanine(5)/198proline products indicated that the former had a 40% lower enzyme activity than the later [46].

Table 1. Association between polymorphisms in the GPx-1 gene and cancer risk"

Polymorphism Pro/Leu 198

GCG Repeat

Tumor Type Lung Bladder Breast Skin Colorectal Prostate'' Breast

Association'' Y Y N N N Y Y

Reference [411 [42] [44] [43] [47] [47]

G/A(-592), T/CC+Z)" ^Epidemiological studies assessing whether there is an association between a polymorphism in the GPx-1 gene and cancer risk ''Indicates whether an association between a polymorphism and cancer risk was indicated: Y, yes; N no The trend did not reach statistical significance ""No information about associations of these polymorphisms and cancer risk

Genetic variations among GPx-1 alleles also permit the detection of loss of one of two copies of the gene dimng tumor development. When this occurs, allelic loss can be detected if the locus is heterozygous in the germline and only one of the two differing GPx-1 genes can be detected in DNA obtained from tumor samples. Loss of heterozygosity (LOH) is a commonly used tool to observe genetic loss during cancer development, often indicating the loss of a gene that provides some protection against transformation. LOH at the GPx-1 locus was first reported by Moscow et al. in 1994, who indicated a 70% heterozygosity frequency in DNAs obtained from cancer-fi-ee

282

Selenium: Its molecular biology and role in human health

individuals, while only 10 of 64 DNA samples obtained from lung cancers showed the presence of both alleles [40]. LOH at the GPx-1 locus was subsequently reported in breast cancer as well [48]. In good agreement with the data of Moscow et al., the heterozygosity index at this locus in the cancer-free population was 75%, but only 39% in the DNA of 74 breast cancer samples. Using a similar approach in measuring GPx-1 LOH in tumor vs. normal tissue, a 42% reduction in the heterozygosity frequency was observed in DNA obtained from cancers of the head and neck as compared to that from cancer-free individuals [50]. In this same study, DNA was genotyped from 3 sets of samples obtained from the same patients: from peripheral lymphocytes, tumor and microscopically normal tissue adjacent to the tumors. GPx-1 LOH occurred in all three of these sample sets, with only one allele being evident in the analysis of tumor DNA while two different alleles were present in the DNA from lymphocytes. Importantly, LOH was also observed in the tumor margin in two of three sets, consistent with the concept that this area, although appearing non-malignant, represents a "field" of pre-cancerous tissues from which tumors develop. This observation also indicates that GPx-1 LOH is likely to be an early event in cancer development. The data on GPx-1 LOH is summarized in Table 2. LOH is likely to have a gene dosage effect, reducing the number of genes capable of ultimately producing the corresponding protein. This suggestion has been supported by data indicating that lung cancers that have undergone GPx-1 LOH had reduced GPx-1 enzyme activity and higher levels of 8hydroxydeoxyguanosine, an oxidative lesion known to be mutagenic [51].

Table 2. Loss of heterozygosity at the GPx-1 locus in human tumors

Tumor Type Lung Breast Head and Neck Colon

LOH" 55% 36% 42% 42%

Ref. [41] [48] [13] [63]

"LOH indicates the difference in the percentage heterozygosity for control, cancer-free individuals and the heterozygosity index reported for the indicated tumor type

SeplS Sep 15 is a conserved selenoprotein that is highly expressed in the prostate, liver, brain, kidney and testis [52]. The gene is polymorphic in humans, existing as either C*'"G"^^ or X^'VA"^^ haplotypes. These polymorphic positions are located in the 3'-untranslated portion of the gene, with the 1125 position being located in that region of the gene encoding the SECIS

Genetic variation among selenoprotein genes and cancer

283

element, a regulatory sequence within the 3'-untranslated region that determines the recognition of the in-frame UGA triplets of selenoprotein mRNAs as the codon for Sec. Using a specialized reporter plasmid designed to measure SECIS function, it was determined that the polymorphism at the 1125 position was functional, with X^^'/A"^^ containing gene being less responsive to the addition of selenium to the culture media [53,54]. Using mesothelioma cells, it was shown that cells carrying the X^'VA"^^ haplotype were less responsive to the growth inhibitory and apoptosis promoting effects of selenium [55]. LOH at the Sep 15 locus also occurs in human tumors. By genotj^jing DNAs from cancer-free individuals and from either cancers of the head and neck or breast, a significant difference in the allele frequency in both cancer types among African Americans was observed, indicating a reduction in the heterozygosity in the cancer DNAs [56]. An example of LOH in a sample set of normal DNA and that obtained from a supraglotis tumor was also reported. Using microsatelite markers for heterozygosity along chromosome 1, the human chromosome where Sep 15 resides, it was also shown that only a marker tightly linked to Sep 15 displayed LOH in breast tumors, while other markers along the same chromosome did not [57]. This indicated that the loss of a Sep 15 allele is a common event in breast cancer, and that the loss of Sep 15 or another important gene very tightly linked to it is a critical step in tumorigenesis. The Sep 15 genetic analysis also indicated a striking difference in Sep 15 allele frequency between Caucasians and African Americans, with there being an approximately 5-fold higher incidence of T^'VA''^^, the same allele that is likely to be less responsive to available selenium availability, in African Americans [56]. Polymorphisms in other selenoprotein genes While direct relationships between poljmiorphisms within the genes for other selenoproteins and cancer etiology have not been described to date, several variations with potential biological significance have been reported. The gene for another member of the GPx family, GPx-4, has been shown to contain a polymorphism in the 3'-untranslated region in the vicinity of that gene's SECIS element that results in a T or C at nucleotide position 718 [58]. Unlike GPx-1, GPx-4 can use phospholipid hydroperoxides as subsfrate and has the second function of serving as a structural component of the mitochondrial capsule of spermatozoa. The GPx-4 polymorphism was associated with differences in the levels of the products of 5-lipoxygenase, indicating that this variation may have an effect on immune function [58]. Several polymorphisms have also been reported in a third member of the GPx family, GPx-2. GPx-2 is predominantly expressed in tissues of the gasfrointestinal fract and no association between polymorphisms in that gene

284

Selenium: Its molecular biology and role in human health

and colon cancer were observed [59]. Additional polymorphisms in the selenoprotein genes for Selenoprotein P and type 2 iodothyronine deiodinase were not shown to be associated with cancer risk, although they have been linked to other human diseases [60-62]. Conclusion Genetic data indicate that the alleles for specific selenoprotein genes are associated with cancer risk, and that allelic loss of several of these genes is a common event in the development of several cancer types as well. Given the benefits of selenium in the prevention of cancer, there is increasing likelihood that several of these benefits will occur as a consequence of selenium's effects on influencing the levels and/or activity of several of these proteins. References 1. 2. 3. 4. 5. 6.

K Brown, J Arthur 2001 Public Health Nut 4:593 M Ryan-Harshman, W Aldoori 2005 Can J Diet Pract Res 66:98 MP Raytnan 2000 Lancet 356:233 K El-Bayoumy (Ed), The role of selenium in cancer prevention JB Lippincott Co, Philadelphia 1991 CIpl998yiV«/r 128:1845 NT Akbaraly, J Amaud, I Hininger-Favier, V Gourlet, AM Roussel, C Berr 2005 Clin

ChemSUim 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20.

M Komitzer, F Valente, D De Bacquer, J Neve, G De Backer 2004 Eur J Clin Nutr 58:98 TJ Hartman, PR Taylor, G Alfthan, R Fagerstrom, J Virtamo, SD Mark, M Virtanen, MJ Barrett, D Albanes 2002 Cancer Causes Control 13:923 P Knekt, J Mamiemi, L Teppo, M Heliovaara, A Aromaa 1998 AmerJEpidem 148:975 H Zhuo, AH Smith, C Steinmaus 2004 Cancer Epidemiol Biomarkers Prev 13:771 RE Rudolph, TL Vaughan, AR Kristal, PL Blount, DS Levine, PC Galipeau, LJ Prevo, CA Sanchez, PS Rabinovitch, BJ Reid 2003 J Natl Cancer Inst 95:750 K Lipsky, R Zigeuner, M Zischka, L Schips, K Pummer, P Rehak, G Hubmer 2004 C/ro/ogv 63:912 NE Allen, JS Morris, RA Ngwenyama, TJ Key 2004 Br J Cancer 90:1392 H Li, MJ Stampfer, EL Giovannucci, JS Morris, WC Willett, JM Gaziano, J Ma 2004 / Natl Cancer Inst 96:696 PA van den Brandt, MP Zeegers, P Bode, RA Goldbohm 2003 Cancer Epidemiol Biomarkers Prev 12:866 JD Brooks, EJ Metter, DW Chan, LJ SokoU, P Landis, WG Nelson, D Muller, R Andres, HB Carter 2001 J Urol 166:2034 M Etminan, JM Fitzgerald, M Gleave, K Chambers 2005 Cancer Causes Control 16:1125 P Ghadirian, P Maisonneuve, C Perret, G Kennedy, P Boyle, D Krewski, A Lacroix 2000 Cancer Detect Prev 24:305 F Femandez-Banares, E Cabre, M Esteve, MD Mingorance, A Abad-Lacruz, M Lachica, A Gil, MA GassuU 2002 Am J Gastroenterol 97:2103 ET Jacobs, R Jiang, DS Alberts, ER Greenberg, EW Gunter, MR Karagas, E Lanza, L Ratnasinghe, ME Reid, A Schatzkin, SA Smith-Warner, K Wallace, ME Martinez 2004 J Natl Cancer Inst 96:1669

Genetic variation among selenoprotein genes and cancer 21. 22. 23. 24. 25.

26. 27.

28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47.

48. 49.

285

DS Michaud, I De Vivo, JS Morris, E Giovannucci 2005 Br J Cancer 93:804 P Ghadirian, P Maisonneuve, C Perret, G Kennedy, P Boyle, D Krewski, A Lacroix 2000 Cancer Detect Prev 24:305 M Garland, JS Morris, MJ Stampfer, GA Colditz, VL Spate, CK Baskett, B Rosner, FE Speizer, WC Willett, DJ Hunter 1995 J Natl Cancer Inst 87:497 PA van den Brandt, RA Goldbohm, P van't Veer, P Bode, E Dorant, RJ Hermus, F Sturmans \994 Am J Epidemiol 140:20 FE Thompson, BH Patterson, SJ Weinstein, M McAdams, VL Spate, RF Hamman, RS Levine, K Mallin, PD Stolley, LA Brinton, JS Morris, RG Ziegler 2002 Cancer Causes Control 13:517 SA McNaughton, GC Marks, P Gaffiiey, G Williams, AC Green 2005 Cancer Causes Control 16:609 LC Clark, GFJ Combs, BW Tumbull, EH Slate, EH Chalker, J Chow, LS Davis, RA Glover, GF Graham, EG Gross, A Krongrad, JLJ Lesher, HK Park, BBJ Sanders, CL Smith, JR Taylor 1996 JAMA 276:1957 A DufTield-Lillicoe, M Reid, J Tumbull, GJ Combs, E Slate, L Fischbach, J Marshall, L Clarke 2002 Cancer Epidemiol Biomarkers Prev. 11:630 G Bjelakovic, D Nikolova, RG Simonetti, C Gluud 2004 Lancet 364:1219 GV Kryukov, VM Kryukov, VN Gladyshev 1999 J Biol Chem 274:33888 GV Kryukov, S Castellano, SV Novoselov, AV Lobanov, O Zehtab, R Guigo, VN Gladyshev 2003 Science 300:1439 DL Hatfield, VN Gladyshev 2002 Mol Cell Biol 22:3565 R Tujebajeva, P Copeland, X-M Xu, B Carlson, J Harney, D Driscoll, D Hatfield, M Berry 2000 EMBO Rep\A52, G Bermano, F Nicol, JA Dyer, RA Sunde, GJ Beckett, JR Arthur, JE Hesketh 1995 BiochemJ?,\\-A25. DM Driscoll, PR Copeland 2003 Annu Rev Nutr 23:17 D Behne, A Kyriakopoulos 2001 Annu Rev Nutr 21:453 JR Arthur 2000 Cell Mol Life Sci 57:1825 GT MuUenbach, A Tabrizi, BD Irvine, GI Bell, JA Tainer, RA Hallewell 1988 Protein Eng 2:239 R Brigelius-Flohe 1999 Free Radic Biol Med 27:951 JA Moscow, L Schmidt, DT Ingram, J Gnarra, B Johnson, KH Cowan 1994 Carcinogen 15:2769 D Rataasinghe, JA Tangrea, MR Andersen, MJ Barrett, J Virtamo, PR Taylor, D Albanes 2000 Cancer Res 60:6381 Y Ichimura, T Habuchi, N Tsuchiya, L Wang, C Oyama, K Sato, H Nishiyama, O Ogawa, T Kato 2004 J Urol 172:728 R Hansen, M Saebo, CF Skjelbred, BA Nexo, PC Hagen, G Bock, IM Bowitz Lothe, E Johnson, S Aase, IL Hansteen, U Vogel, EH Kure 2005 Cancer Lett 229:85 JA Knight, UV Onay, S Wells, H Li, EJ Shi, IL Andrulis, H Ozcelik 2004 Cancer Epidemiol Biomarkers Prev 13:146 U Vogel, A Olsen, H Wallin, K Overvad, A Tjonneland, BA Nexo 2004 Cancer Epidemiol Biomarkers Prev 13:1412 T Hamanishi, H Furuta, H Kato, A Doi, M Tamai, H Shimomura, S Sakagashira, M Nishi, H Sasaki, T Sanke, K Nanjo 2004 Diabetes 53:2455 Z Kote-Jarai, F Durocher, SM Edwards, R Hamoudi, RA Jackson, A Ardem-Jones, A Murkin, DP Deamaley, R Kirby, R Houlston, DF Easton, R Eeles 2002 Prostate Cancer Prostatic Dis 5:\89 YJ Hu, AM Diamond 2003 Cancer Res. 63:3347 L Forsberg, U de Faire, SL Marklund, PM Andersson, B Stegmayr, R Morgenstem 2000 Blood Cells Mol Dis 26:423

286 50. 51. 52. 53

54.

55.

56.

57. 58. 59. 60. 61. 62. 63.

Selenium: Its molecular biology and role in human health YJ Hu, ME Dolan, R Bae, H Yee, M Roy, R Glickman, L Kiremidjian-Schumacher, AM Dx&moMlWiA Biol Trace Elem Res 101:97 LJ Hardie, JA Briggs, LA Davidson, JM Allan, RF King, GI Williams, CP Wild 2000 Carcinogen 21:167 VN Gladyshev, K Jeang, JC Wootton, DL Hatfield 1998 J Biol Chem 273:8910 E Kumaraswamy, A Malykh, KV Korotkov, S Kozyavkin, Y Hu, SY Kwon, ME Moustafa, BA Carlson, MJ Berry, BJ Lee, DL Hatfield, AM Diamond, VN Gladyshev 2000 J Biol Chem 275:35540 Y Hu, KV Korotkov, Mehta, R Mehta, DL Hatfield, CN Rotimi, A Luke, TE Prewitt, RS Cooper, W Stock, E Vokes, ME Dolan, VN Gladyshev, AM Diamond 2001 Cancer Res 61:2307 S Apostolou, JO Klein, Y Mitsuuchi, JN Shetler, PI Poulikakos, SC Jhanwar, WD Kruger, JR Testa, HJ Wu, C Lin, YY Zha, JG Yang, MC Zhang, XY Zhang, X Liang, M Fu, M Wu 2004 Oncogene 22:119 YJ Hu, KV Korotkov, R Mehta, DL Hatfield, CN Rotimi, A Luke, TE Prewitt, RS Cooper, W Stock, EE Vokes, ME Dolan, VN Gladyshev, AM Diamond 2001 Cancer Res 61:2307 MA Nasr, YJ Hu, AM Diamond 2003 Cancer Therapy 1:293 S Villette, JA Kyle, KM Brown, K Pickard, JS Milne, F Nicol, JR Arthur, JE Hesketh 2002 Blood Cells Mol Dis 29:174 OH Al-Taie, N Uceyler, U Eubner, F Jakob, H Mork, M Scheurlen, R Brigelius-Flohe, K Schottker, J Abel, A Thalheimer, T Katzenberger, B Illert, R Melcher, J Kohrle 2004 Nutr Cancer 48:6 TW Guo, FC Zhang, MS Yang, XC Gao, L Bian, SW Duan, ZJ Zheng, JJ Gao, H Wang, RL Li, GY Feng, D St Clair, L He 2004 J Med Genet 41:585 DA Chistiakov, KV Savost'anov, RI Turakulov 2004 Mol Genet Metab 83:264 D Mentuccia, L Proietti-Pannunzi, K Tanner, V Bacci, TI Pollin, ET Poehlman, AR Shuldiner, FS Celi 2002 Diabetes 51:88 Y Hu, RV Benya, RE Carroll, AM Diamond 2005 J Nutr (In Press)

Chapter 25. Selenium and viral infections Melinda A. Beck Department of Nutrition, University of North Carolina at Chapel Hill, Chapel Hill, NC 27599, USA

Summary: Viruses which infect a selenium-deficient host have been shown to mutate. Two very different viruses, coxsackievirus and influenza virus, repeatedly evolve the same specific stable viral mutations when Se-deficient mice are infected with these viruses. The mutations resulted in increased virulence of both viruses, hi addition, in a human population with low Se status, the live attenuated vaccine strain of poliovirus had increased mutations relative to individuals with higher selenium status. Decreased Se is also associated with immune impairment, with both chemokine and cytokine levels altered. These studies demonstrated that the Se status of the host can profoundly influence the genome of viral pathogen, leading to a new viral strain. Thus, host nutritional status should be considered when studying the mechanisms underlying the evolution of emerging viruses and may assist in predicting new viral outbreaks and devising new strategies to limit the emergence and spread of these pathogenic forms. Se-deficiency and coxsackievirus The first hint that a deficiency in Se may have an effect on a viral infection came from China in the early 1930's. A cardiomyopathy developed predominately in women and children living in specific geographic regions of China in which the soil was deficient in Se [1]. Because food was grown and consumed locally, people living in areas with Se-deficient soils became Se-deficient and developed an endemic cardiomyopathy termed Keshan disease. Keshan disease is characterized by foci of necrosis throughout the myocardium with the lesions exhibiting varying degrees of cellular infiltration and calcification. However, because the disease occurred with both a seasonal and annual incidence and because not everyone who was Sedeficient developed the disease, scientists in China suspected a virus infection in addition to Se deficiency was necessary for the development of Keshan disease. Using both serological and viral isolation techniques, coxsackie B virus antibodies and coxsackie B viruses were found in the blood and tissues of Keshan disease victims [2]. ^dependent of Se status, these viruses are

288

Selenium: Its molecular biology and role in human health

known etiologic agents of viral-induced myocarditis and suspected agents of dilated cardiomyopathy. There are 6 group B coxsackieviruses, and coxsackievirus B3 and B4 are most commonly associated with myocarditis. Indeed, these two coxsackieviruses were the most common viral finding in Keshan disease victims. Within the past several years investigators have utilized RT-PCR techniques on archival tissue from Keshan disease victims to further confirm the association of coxsackievirus and Se-deficiency with Keshan disease [3]. hi order to further characterize the effect of Se-deficiency on coxsackievirus infection, an animal model was utilized. The mouse model of coxsackievirus-induced myocarditis has been well characterized and development of myocarditis in the infected mouse is quite similar to what is seen in infected humans [4]. Using the mouse model, mice were fed Sedeficient grains from Keshan endemic areas. Control mice were fed grains from non-Keshan areas. Following the feeding, mice were infected with a strain of coxsackievirus B4 that was originally isolated from the heart of a Keshan disease patient. Following infection, mice fed the Se-deficient grains developed severe myocarditis, whereas those fed the Se-adequate grains did not develop severe disease [5]. To further expand upon this earlier work, purified diets either sufficient or deficient in Se were fed to mice for 4 weeks to induce Se-deficiency. Mice were then infected with either a myocarditic sfrain of coxsackievirus B3 (CVB3/20) or an avirulent CVB3 strain (CVB3/0). As shown in Figure 1, the normally myocarditic strain of B3, CVB/20, induced a greater level of pathology in the Se-deficient mice and the normally avirulent strain of B3, CVB3/0, induced heart pathology only in the Se-deficient animals. Thus, a deficiency in Se could greatly exacerbate the pathogenesis of the virus infection [6,7]. To further understand the effects of the Se-deficiency on viral pathogenesis, the immune response to the virus in both Se-deficient and Seadequate mice was studied. Many nutritional deficiencies are associated with impaired immunity, and a deficiency in Se has been found to increase the susceptibility to parainfluenza 3 virus in lambs and infectious bovine rhinofracheitis virus in steers [8,9]. Neufralizing antibody titers, which are important in viral clearance, were found to be equivalent between Se-adequate and Se-deficient mice. However, the ability of spleen cells to proliferate to both mitogen and specific CVB3 antigen was greatly diminished in the Se-deficient mice, though natural killer cell activity was not affected by the Se deficiency. Thus, immune impairment occurred as a result of the Se deficiency [6,7]. Cardiac inflammation is the hallmark of myocarditis and there are many different signals which direct the influx of the inflammatory cells to the site

Selenium and viral infections

289

I

Se+

OS

a.

Se-

o ra •2 O

CVB3/20

CVB3/0

Infecting Virus Strain Figure 1. Heart pathology post CVB3 infection of Se-adequate or Se-deficient mice. Each bar represents the mean +/- S.D. pathology score of 10 individual animals. Cardiac pathology was scored at 10 days post infection as follows: 0 = no lesions; 1 = foci of mononuclear cell inflammation associated with myocardial cell reactive changes without myocardial cell necrosis; 2 = inflammatory foci clearly associated with myocardial cell reactive changes; 3 = inflammatory foci clearly associated with myocardial cell necrosis and dystrophic calcification; 4 = extensive inflammatory infiltration, necrosis and dystrophic calcification.

of infection. Chemokines or chemoattractant cytokines, are secreted by immune cells, fibroblasts and endothelial cells. The release of chemokines provides a chemical gradient by which immune cells are drawn to the site of infection. The requirement for the involvement of chemokines in coxsackievirus-induced myocarditis was demonstrated with macrophageinflammatory protein l a (MlP-la) knockout mice. These mice were completely protected from coxsackievirus-induced myocarditis. Although virus could be recovered from the hearts of infected KO mice, the infection failed to induce an inflammatory response due to the lack of MlP-la [10]. This finding was particularly interesting because the production of other chemokines could not compensate for a lack of MlP-la production. To determine if chemokine responses were altered in the CVB3/0 infected Se-deficient mice, hearts were assayed for mRNA expression of RANTES, eotaxin, MCP-1, MlP-la and MIP-ip. No differences were found in the expression of these chemokines except for MCP-1. MCP-1 expression

290

Selenium: Its molecular biology and role in human health

peaked on day 7 post infection in the Se-adequate mice, but peaked on day 10 in the Se-deficient mice [11]. This delay in MCP-1 expression may have led to the development of myocarditis in the CVB3/0-infected Se-deficient mice. In addition to chemokines, a number of pro-inflammatory cytokines also play a critical role in the development of cardiac inflammation post coxsackievirus infection. However, no differences were found in the cardiac mRNA expression of IL-1, IL-2, IL-4, IL-5, IL-6 and IL-15 between infected Se-adequate and Se-deficient mice [11]. However, there were significant differences with respect to mRNA expression of IFN-y. The hearts of Sedeficient mice produced very little IFN-y mRNA compared with Se-adequate mice. IFN-y is an antiviral cytokine and is also important for the activation of macrophages. Although a number of immune differences were found between CVB3infected Se-adequate and Se-deficient mice, there was also the possibility that the virus itself had undergone changes due to the Se-deficiency. To test this possibility, CVB3/0 virus isolated from the hearts of either Se-deficient or Se-adequate mice were used to infect Se-adequate mice. If the change in virulence of the CVB3/0 virus was due to host effects caused by the Sedeficiency, then the mice infected with virus recovered from either Seadequate or Se-deficient mice should not develop myocarditis. However, the Se-adequate mice infected with virus recovered from Se-deficient mice did, in fact, develop myocarditis. This finding demonstrated that the virus itself had mutated as a result of replicating in a Se-deficient host. To confirm this finding, CVB3/0 viruses recovered fi-om Se-deficient and Se-adequate mice were sequenced and compared with the input CVB3/0 sequence. No nucleotide changes were found in the virus isolated fi"om Se-adequate mice. However, six nucleotide changes (4 of which led to amino acid changes, and two in the non-translated region of the virus) were found in the virus isolated fi-om the Se-deficient mice [12]. Each of these nucleotide changes are identical to nucleotide sequences found in myocarditic strains of CVB3. A seventh nucleotide found in myocarditic strains of CVB3 was not changed in the virus obtained from the Se-deficient animals. Thus, the newly virulent CVB3/0 strain isolated fi-om Se-deficient mice contained 6 of 7 nucleotides associated with virulence. This finding confirmed that the phenotype change of the normally avirulent CVB3/0 into a virulent virus in the Se-deficient host was due to a change in the viral genome. This was the first report of a specific nutritional deficiency driving changes in a viral genome. Oxidative Stress and Se-deficiency The finding that host Se-deficiency could induce viral mutations led us to determine the mechanism for this phenomenon. Se is an essential component of the peroxide-destroying enzyme glutathione peroxidase

Selenium and viral infections

291

(GPX). Therefore, one function of Se is as an antioxidant. There are five isozymes of GPX, and GPX-1, or classical GPX, is found in almost all cells of the body. A deficiency of Se in the diet will lead to a decrease in the activity of glutathione peroxidase. Indeed, GPX-1 levels were decreased 5fold in mice fed the Se-deficient diets for 4 weeks. To determine if the antioxidant properties of Se were involved in the mutation of the CVB3 virus, GPX-1 knockout (KO) mice were utilized. These mice have normal levels of Se, but lack GPX-1 activity, although the other 4 GPX isozymes are intact. Under non-stressed conditions, these mice grow and reproduce normally. GPX-1 KO and wildtype control mice were infected with the avirulent strain of CVB3, CVB3/0. A little over 50% of the GPX-1 KO mice developed myocarditis [13]. As expected, none of the wildtype mice developed myocarditis. The results of this experiment suggested that the change in virulence of the virus was due to the antioxidant properties of Se as a component of glutathione peroxidase. To determine if viral mutation occurred in the infected GPX-1 KO mice similar to what occurred in the Se-deficient mice, virus was sequenced from the hearts of infected GPX-1 KO mice with and without myocarditis as well as from infected wildtype mice. Virus obtained from GPX-1 KO mice that developed pathology had 7 nucleotide changes, 6 of which were identical to the changes found in the Se-deficient mice [13]. An additional nucleotide change was found, which corresponded to the nucleotide change found in myocarditic strains of the virus. No changes in the viral genome were found in infected GPX-1 KO mice which did not develop pathology. In addition, virus isolated from wildtype mice did not mutate. Thus, the genome changes were associated with pathology and a lack of GPX-1 activity may be the mechanism by which a deficiency in Se leads to viral mutations. Further evidence that oxidative stress may be the mechanism by which a Se deficiency can drive changes in a viral genome includes studies with mice fed a diet deficient in vitamin E. Vitamin E is a lipid-soluble vitamin that acts as a free radical scavenger. Thus, similar to Se, it functions as an antioxidant, although by a very different mechanism. Mice fed a diet deficient in vitamin E developed myocarditis when infected with the avirulent CVB3/0 strain [14]. As for Se deficiency, the phenotype change in the virus was due to viral mutations. Furthermore, the addition of fish oil to the diet (a pro-oxidant stimulator) also enhanced the pathogenicity of the virus [14]. Finally, feeding mice a diet with excess iron also increases the virulence of the CVB3/0 virus [15]. Excess iron is associated with increased hepatic liver peroxidation and increased oxidative stress. All of these observations taken together suggest a common mechanism of oxidative stress. Thus, a lack of Se in the diet leads to increased host oxidative stress.

292

Selenium: Its molecular biology and role in human health

which in turn can enhance the abihty of a virus to mutate under these increased stress conditions. Influenza virus and Se-deficiency Each year, worldwide infection with influenza virus results in a great deal of morbidity and mortality. In the United States alone, over 36,000 deaths and 120,000 hospitalizations occur each year as a consequence of infection with influenza virus and its complications [16]. Because new strains of influenza virus arise each year with surface changes that aid in escape from immune detection, exposure to last's year influenza strain may not necessarily protect from the new viral strain. [17]. Decisions on which strains to include in the influenza vaccine are made each year 6-8 months in advance of the influenza season, therefore the efficacy of the vaccine is dependent on the correct choice of strains to include in the vaccine. Thus, year to year, there is variation in the degree of protection that the influenza vaccine provides. The precise mechanism(s) by which influenza viruses mutate each year is not well understood. Small changes (antigenic drift) in the surface proteins of the virus, the hemagglutinin (HA) and the neuraminidase (NA), which are involved in the ability of the virus to bind to and be released from host cells, are primarily responsible for the yearly epidemics. Large changes in these two proteins (antigenic shift) due to reassortment of gene segments are responsible for worldwide pandemics. Both the HA and NA are involved in virulence of the virus. Because the HA and the NA are external proteins, their exposure to the immune system is thought to be a driving force for inducing viral mutations. The influenza virus matrix protein (M) is also involved in virulence. However, because the M protein is internal and therefore not exposed to the host antibody response, it is thought to be more stable and exhibits few nucleotide changes over time [18]. In order to determine if Se deficiency could influence the pathogenicity of influenza virus, a mouse model of influenza virus infection was utilized. As for coxsackievirus, the mouse has been used for influenza viral studies for over 50 years and provides an excellent animal model for examining the pathogenicity of influenza virus. Like CVB3, influenza virus is an RNA virus. However, unlike CVB3, it has a segmented, negative-sense genome and is in a different viral order (Orthomyxoviridae) than coxsackievirus (Picomaviridae). Mice were fed a diet either deficient or adequate in Se for 4 weeks, beginning at weaning. Following the feeding period, mice were inoculated intranasally with Influenza A Bangkok/1/79. This virus strain causes a mild interstitial pattern of pneumonia in infected mice. At days 0 (uninfected) 4, 6, 10 and 21 days post infection, mice were sacrificed and their lungs were removed for histopathological analysis. As shown in Figure 2, at days 4, 6, 10, and 21 days post infection, the lung pathology was much more severe in

Selenium and viral infections

293

the Se-deficient mice as compared with Se-adequate mice [19]. Thus, Sedeficient mice are much more susceptible to developing severe pathology when infected with a mild strain of influenza virus.

CD

o o CO

2

-

>>

Se-adequate

D) O O

m Se-deficient

x: -I—'

CO Q-

O) 3

Days Post Infection Figure 2. Lung pathology post influenza vims infection in Se-adequate (solid bars) and Sedeficient (hatched bars) mice. Each bar represents the mean +/- the S.D. of 10 animals. Lung pathology was scored without knowledge of the experimental variables and was graded semiquantitatively according to the relative degree (from lung to lung) of mononuclear cell infiltrate. Bars marked with * indicate statistically significant difference between groups at that time period (p>.001)

The immune response to influenza is characterized by 2 stages, an innate stage and a cell mediated stage. The initial, innate response, involving natural killer (NK) cells, dendritic cells and macrophages is essential for directing the subsequent cell-mediated response. We examined key aspects of the innate response in Se-deficient mice. Specifically, interferon (IFN)-a and IFN-P mRNA, which help control viral replication and activate a host of anti-viral genes, were reduced in Se-deficient influenza infected mice (Figure 3). IFN-y, which is produced by NK cells early in infection, was also reduced in Se-deficient infected mice (Figure 3). Interestingly, NK cell activity is unaffected in Se-deficient mice (data not shown). Together, these

294

Selenium: Its molecular biology and role in human health

data indicate that Se-deficient mice have impaired, early anti-viral cytokine responses.

IFN-a

IFN-P

IFN-Y

Figure 3. Quantitative RT-PCR was performed for the cytokines and normalized to G3PDH. The data are expressed as arbitrary units +/- SEM. Se deficiency resulted in 2-4 fold decrease in production of IFN-a, IFN-P and IFN-y, at 24h p.i.

In order for the lung inflammation to occur in the infected influenza mice, a co-ordinate production of chemokines must occur. This process was altered in the Se-deficient mice. Chemokine mRNA levels for RANTES, MlP-la, MIP-P and MCP-1 were highest on days 4 and 5 post infection for the Se-adequate mice, and then began to decrease, whereas these chemokines were highest at later time points for the Se-deficient mice [19]. Clearly, Se deficiency leads to an increase in influenza-induced histopathology which is associated in part with altered chemokine and IFN expression. Because host Se deficiency induced changes in the coxsackievirus genome, we reasoned that the increased virulence of the influenza virus in the Sedeficient mice may also be due to changes in the viral genome. Virus was recovered from the lungs of Se-deficient and Se-adequate mice and all 8 viral RNA segments were sequenced and compared with the sequence of the input strain. Surprisingly, few changes were found in the HA and NA segments of the virus, which are associated with a high mutation rate. Changes in the HA and NA segment were random, and found in viruses obtained from both Seadequate and Se-deficient mice. In stark contrast, however, the M gene

Selenium and viral infections

295

contained multiple mutations [20]. As shown in Table 1, three separate isolates from 3 individual Se-deficient mice all had identical mutations in 29 positions. One of the 3 isolates from a Se-deficient mouse had an additional 6 mutations. None of these changes were seen in viruses obtained from the Se-adequate mice. Thus, as for coxsackievirus, influenza virus replicating in a Se-deficient host undergoes rapid genetic change, resulting in a more virulent virus which can now cause disease even in a host with normal Se status. Poliovirus and Se deficiency A recent study has reported that subjects in the United Kingdom with low Se status (< 1 i^Mol/L) have a decreased immune response to poliovirus vaccination [21]. Of particular note, they found an increased mutation rate of the vaccine strain of vaccine virus which had been shed in the feces. Supplementation with Se of the low Se status population enhanced the immune response and lowered the number of mutations found in the shed vaccine strain of the virus. Thus, low Se status was associated with increased mutation rate of the live attenuated poliovirus vaccine strain when compared with vaccinated individuals supplemented with Se. To assess the mutation rates, the investigators utilized temporal temperature gradient electrophoresis (TTGE). Although this technique can identify mutations occurring in the genome, it does not provide information on which specific nucleotides were altered. The Broome et al. study supports the hypothesis that polio vaccination of individuals with low Se status may lead to increased mutations in the vaccine strain of virus. This area of research is particularly relevant in view of recent findings that attenuated poliovirus vaccine sfrains have circulated and reverted to virulence in several areas of the world where undernutrition is prevalent [22]. Selenium and other viruses Infection with human immunodeficiency virus (HIV) results in a loss of CD4+ helper T cells and subsequent immune dysfunction leading to increased opportunistic infections. In addition, oxidative stress increases during an HIV infection. A number of studies have examined the relationship between specific nutritional factors and disease progression and survival of HIV infected individuals. Se status of HIV infected individuals has also been studied. In developed countries (France and the US), 3 studies have demonstrated that lower serum Se levels are associated with an increased risk of mortality from HIV [23-25]. In a study of HFV infected pregnant women in Tanzania, low selenium status was found to be associated with accelerated HIV disease progression [26].

296

Selenium: Its molecular biology and role in human health

Table 1. Comparison of nucleotide sequences of influenza A/Bangkok/1/79 M gene of the infecting virus and of virus isolated from Se-adequate (Se+) and Se-deficient (Se-) mice.

Nucleotide Position 136 205 238 309 322 325 328 331 334 370 371 406 439 454 455 502 503 524 525 544 566 567 568 610 619 652 655 667 669 670 677

Infecting I Virus A G G G A C A A T A G C A C C

cA G G A C C G A G C G G G A G

Se status of host virus isolated from: Se± Se+ Se+ Se; Soz Sez A G G G A C A A T A G C A C C C A G G A C C G A G C G G G A G

A G G G A C A A T A G C A C C C A G G A C C G A G C G G G A G

A G G G A C A A T A G C A C C C A G G A C C G A G C G G G A G

C A G A C T G C C C T T G A C T C A G C T C A G A T A A G G A

C A G A C T G C C C T T G A C T C A G C T C A G A T A A G G A

C A A A C T G C C C T T G A A T C A A C T T A G A T A A A G A

AA Change

RtoK

AtoS

AtoT TtoA

AtoT

AtoT

Se levels have also been inversely correlated with hepatitis B virus infection. Lifection with hepatitis B virus is a major health problem throughout the world. In addition, chronic hepatitis B infection is thought to

Selenium and viral infections

297

be a significant factor in most hepatocellular carcinomas, a highly malignant neoplasm with a high mortality rate. A study from Taiwan [27] demonstrated that mean Se plasma levels were significantly lower in hepatocellular carcinoma patients, as compared with individuals testing positive for hepatitis B virus. A further study from Qidong county in China [28] demonstrated a protective effect of Se supplementation in a population at high risk of developing primary liver cancer due to a high prevalence of hepatitis positive individuals. Conclusion Low host selenium status has been shown to be important in driving viral mutations. This increase in viral mutations in a Se-deficient host may be due to an increase in oxidative stress status, as virus which replicated in GPX-1 knockout mice also mutated. Emerging viruses are either newly arisen viruses or are viruses that are rapidly expanding their range. Understanding the mechanisms underlying the evolution of emerging viruses is critical to predicting new viral outbreaks and devising new strategies to limit the emergence and spread of these new pathogenic forms. Data from the Se studies demonsfrates that host Se status is a driving force for emergence of new viral variants. These observations suggest a new area for research, namely the interaction of host nufrition and viral evolutionary processes. The precise mechanism(s) by which a deficiency in Se leads to mutations in a viral genome remains to be determined. References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17.

BQ Gu 1983 Chin Med J 96:251 C Su, C Gong, J Li, L Chen, D Zhou, Q Jin 1979 Chin MedJ59:466 LQ Ren, XJ Li, GS Li, ZT Zhao, B Sun, F Sun 2004 World J Gastroenterol 10:3299 JF Woodruff 1980 Am J Pathol 101:427 J Bai, S Wu, K Ge, X Deng, C Su 1980 Acta Acad Med Sin 2:29 MA Beck, PC Kolbeck, LH Rohr, Q Shi, VC Morris, OA Levander \994 J Infect Dis 170:351 MA Beck, PC Kolbeck, LH Rohr, Q Shi, VC Morris, OA Levander 1994 J Med Virol 43:166 JK Reffett, JW Spears, TT Brown Jr 1988 J Animal Sci 66:1520 JF Reffett, JW Spears, TT Brown Jr 1988 JNutr 118:229 DN Cook, MA Beck, T Coffman, SL Kirby, JF Sheridan, IB Pragnell, O Smithies 1995 5«e«ce 269:1583 MA Beck, CC Matthews 2000 Proc Nutr Soc 59:1 MA Beck, Q Shi, VC Morris, OA Levander 1995 Nat Med 1:433 MA Beck, RS Esworthy, Y-S Ho, F-F Chu \99SFASEBJ \2:l\43 MA Beck, PC Kolbeck, LH Rohr, Q Shi, VC Morris, OA Levander 1993 J Nutr 124:345 MA Beck, Q Shi, VC Morris, OA Levander 2005 Free Radic Biol Med 38:112 CB Bridges, SA Harper, K Fukuda, TM Uyeki, NJ Cox, JA Singleton 2003 Morb Mortal Wkly Rep 52:1 BR Murphy, RG Webster 1996 Fields Virology BN Fields (ed) Lippincott-Raven Philadelphia PA pi 397

298 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28.

Selenium: Its molecular biology and role in human health AC Ward 1997 Virus Genes 14:187 MA Beck, HK Nelson, Q Shi, P Van Dael, EJ Schiffrin, S Blum, D Barclay, OA Levander 2001 F^SESy 10.1096/fj.00-072ige HK Nelson, Q Shi, P Van Dael, EJ Schriffrin, S Blum, D Barclay, OA Levander, MA Beck. 2001 FASEBJ\0M9m].Q\-Q\\5f)e CS Broome, F McArdle, J A Kyle, F Andrews, NM Lowe, CA Hart, JR Arthur 2004 J Clin Nutr m:\54 OM Kew, RW Sutter, EM de Gourville, WR Dowdle, MA Pallansch 2005 Ann Rev Microbiol 59:5S7 MK Baum, G Shor-Posner, S Lai, G Zhang, H Lai, MA Fletcher, H Sauberlich, JB Page 1997 y.4/D5 15:370 A Campa, MK Shor-Posner, F Indacochea, G Zhang, H Lai, D Asthana, GB Scott, MK Baum 1999 y^/DS 20:508 J Constans, JL Pellegrin, C Sergeant, M Simonoff, L Pelegrin, H Fleury, B Leng, C Conri 1995 y4/DS 10:392 R Kupka, GI Msamanga, D Spiegelman, S Morris, F Mugusi, DJ Hunter, WW Fawzi 2004 yiVMfr 134:2556 M-W Yu, I-S Homg, K-H Hsu, Y-C Chiang, Y-F Liaw, C-J Chen 1999 Am J Epidemiol 150:367 SY Yu, YJ Zhu, WG Li 1997 Biol TraceElem Res5(,:\\l

Chapter 26. Role of selenium in HIV/AIDS Marianna K. Baum and Adriana Campa Florida International University, Stempel School of Public Health, Department of Dietetics and Nutrition, U200SW8th Street, Miami, Florida 33199, USA

The advent of Highly Active Antiretroviral Therapy (HAART) in the late 90s has transformed HIV infection from a deadly condition into a chronic, manageable viral infection in developed countries [1]. The developing world, however, accounts for 96% of the global HIV-l infections, and in most of these countries, antiretrovirals are not yet widely available. The number of persons living with Human Immuno-Deficiency Virus (HIV) infection and Acquired Immuno-deficiency Syndrome (AIDS) worldwide has been estimated to be approximately 40 million [2], and this figure includes approximately 5 million people who acquired HIV in 2004. In the same period, approximately 3.1 million adults and children died from AIDS, and 14,000 new individuals are still infected daily, a number that lessens hopes for a rapid solution to this pandemic [2]. The gap between developed and developing countries in the control of the pandemic and treatment of infected persons is growing, and one of the factors fueling the epidemic in poor countries is malnufrition. Moreover, protein-energy malnutrition (PEM), and the accompanying and aggravating micronufrient deficiencies, are already an overwhelming health problem and still the main cause for immune disturbances in poor countries [3,4]. Sub-Saharan Africa, where the greatest growth in severe and generalized malnutrition has occurred in the last two decades [5], is also the region in which 12 out of 44 countries have more than 10% prevalence of HIV in the adult population [6]. Numerous studies have demonstrated that nutritional deficiencies accelerate HIV disease progression and decrease survival [7-16]. Moreover, nutrient deficits interfere with the effectiveness of antiretrovirals by delaying the recuperation of the immune system and aggravating side-effects attributed to treatment [17-20]. Selenium appears to have a multifactorial role in HIV-l infection. Selenium status affects HIV disease progression and mortality [14-16] through various potential mechanisms. In two recent studies, deficiency of selenium has been associated with elevated measures of HIV infectivity [21,22], and therefore, with increased potential to transfer the infection.

300

Selenium: Its molecular biology and role in human health

Selenium is required for the function of gluthatione peroxidase, a biological antioxidant that protects against oxidative stress. Other selenoproteins may also act as antioxidants by the incorporation of selenocysteine in their molecules [23]. In HIV-infected persons, dietary selenium intake was strongly associated with reduced measures of oxidative stress [24]. Adequate selenium status may also be essential for controlling viral emergence and evolution [25,26]. In addition, selenium may enhance resistance to infection through modulation of both cellular and humoral immunity. Plasma selenium levels are associated with interleukin production and subsequent changes in Thl/Th2 cytokine responses [27,28]. Other nutritional factors interact with selenium status and are important in HIV-1 disease progression and mortality. These factors include disease stage, nutritional status at the onset of the disease, types of treatment and compliance, and secondary infections that may act independently or in combination. Treatment of malnutrition, and the accompanying micronutrient deficiencies, thus, requires a carefully individualized approach. This chapter will review the role of selenium in HFV-l disease progression, morbidity and mortality, as well as the factors that may affect these relationships. Selenium and immunity in HIV Selenium has been shown to affect the immune process [29]. In vivo and in vitro studies suggest that selenium may act at different levels of immune function. In animal models, selenium deficiency was shown to impair the ability of phagocytic neutrophils and macrophages to destroy antigens, and selenium status was associated with humoral immune response [30]. In humans, Broome and colleagues [31] found that in a population of sixty-six healthy participants who were marginally deficient in selenium (<1.2 |.imol/L), supplementation with 100 |ig of selenium, as compared to a placebo, increased the cellular immune response, but did not affect the humoral response. In Broome's study, selenium supplementation increased plasma selenium concentrations, the body exchangeable selenium pool, lymphocyte phospholipids and cytosolic glutathione peroxidase activities. After stimulating the immune system of the participants in the study with poliovirus vaccine, those supplemented with selenium showed faster clearance of the poliovirus and less mutations in its reverse transcriptase polymerase chain reaction products than those who received a placebo [31]. Our studies in a cohort of HrV+ drug users [32] suggest that selenium deficiency is significantly correlated with manifestations of herpes infections. Herpes infections are of particular significance in HIV-infected individuals since the co-infection is associated with faster disease progression [32,33].

The role of selenium in HIV/AIDS

301

Selenium status also appears to be associated with the cytokine profile. Producing the appropriate pattern of cytokines depends on adequate activation of T helper (Th) cells by an infections agent [34]. Thl cells are especially effective when cellular response to antigens, such as viruses, is needed. Th2 cells are helpers for B cells, and appear to be adapted to support antibody response and defense against parasites. In HIV-1 infection, when a strong cellular response is critical to control the virus, a change to a Th2 pattern of cytokine production may be harmful. In in-vitro models, selenium regulates levels of interleukin-2 (IL-2), the cytokine responsible for the earliest and most rapid expansion of T lymphocytes. Selenium enhances IL-2 production in a dose-dependent manner. The mechanism of selenium action appears to occur through the increased expansion of high-affinity receptors [35]. In animal and human studies, selenium supplementation has shown to increase the cellular immune response through the increased activity of natural killer (NK) cells, earlier expansion and proliferation of TIjmiphocytes, increase in CD4+ T-helper cells, increased production of interferon y, increased high-affinity interleukin-2 receptors, improved Thl response to RNA viruses, delayed hypersensitivity skin responses, and stimulation of vaccine induced immunity [31,36-38]. In HIV infected patients, selenium deficiency has been significantly correlated with total lymphocyte counts. Plasma selenium levels have been positively correlated with CD4 cell counts and CD4/CD8 ratio, and have been inversely correlated with Pa-microglobulin and thymidine-kinase activity [39]. The pro-inflammatory cytokines associated with HFV-l infection include interleukin-1, tumor necrosis factor alpha (TNF-a), interleukin-6 and interferon [40]. TNF-a, a cytokine prominent in the pathogenesis of anorexia and cachexia in chronic diseases, seems to be affected by selenium status [41]. TNF-induced HIV replication has been suppressed by selenium supplementation, and appears to be related to selenoprotein synthesis, especially in the glutathione and thioredoxin systems [42]. Look et al [39] demonstrated in HIV-infected patients that plasma selenium levels are inversely correlated with levels of soluble TNF type n receptors. Selenium supplementation, therefore, may have the potential to reduce TNF receptors and prevent some of the adverse effects of high TNF circulating levels, such as wasting and Kaposi's sarcoma. Evidence for a selenium-cytokine mechanism of action has also been described in studies indicating the potential of selenium to decrease neuropathogenesis through suppression of interleukin-induced HIV-l replication, neuronal apoptosis, and blood brain barrier damage [39,42-44]. Although supplementation with omega-3 fatty acid in cancer patients has shown some promise as anti-inflammatory for preventing or ameliorating

302

Selenium: Its molecular biology and role in human health

cancer cachexia through its effect on the production of tumor necrosis factor and interleukin-6 [45], studies in AIDS patients, however, have not been conclusive [46]. Increased production of cytokines in AIDS interferes with metabolic processes in the liver, leading to increased gluconeogenesis, proteolysis, lipolysis, and consequently, loss of body fat stores. Selenium levels have been inversely correlated with yet another cytokine, interleukin-8 (IL-8) [39]. Two possible mechanisms have been advanced. The first proposes that the increased oxidative stress in HIV infection is caused by elevated IL-8 levels, which exhausts the available selenium to protect cells against the inflammatory response. The second, supported by in vitro studies, proposes that selenium, in the glutathione peroxidase system, can inhibit IL-8 release by endothelial cells [43], a mechanism impaired by selenium deficiency. HIV-1 infection, wasting and metabolic syndrome AIDS has been characterized by malnutrition and wasting since the early years of the epidemic [47]. Despite the improvement on survival achieved by antiretroviral therapy [48-51], loss of weight and lean body mass remains independent predictors of mortality [52-54]. Even modest weight loss of 5% of total weight within a six-month period has been found to be an independent predictor of opportunistic infections and shorter survival time, after adjusting for highly active antiretroviral therapy (HAART), and baseline CD4 counts [52,55-57]. Moreover, long-term HAART, has not been found to prevent a relatively elevated (18 to 40%) prevalence of wasting among some groups of HIV-infected patients [53,58-60]. The immunological consequences and clinical manifestations of AIDS are similar to those associated with starvation [54,61-63]. In agreement with this observation, wasting, in our cohort of 119 drug users in Miami, was found to be strongly and significantly associated with food insecurity (lack of access to food regularly) and viral load, after controlling for HAART. Although more than 90% of those wasting received HAART, the exceedingly high mean viral load was interpreted as a sign of treatment failure [54]. The lifestyle and socio-economic factors that were associated with wasting in this group were alcohol abuse, illicit cocaine use, and inability to hold a full or part time job [54]. The wasting syndrome in AIDS is accompanied by loss of a large proportion of fat free mass, a process that is more closely associated with starvation [64] than with the cachectic/catabolic pathways demonstrated in other conditions. The similarities in immunological abnormalities observed in protein-energy malnutrition and HIV-1 infection include decreased T lymphocyte and CD4 cells, reduced secretory IgA, and impaired primary and secondary delayed cutaneous hypersensitivity responses [65-67]. However,

The role of selenium in HIV/AIDS

303

HlV-wasting remains an extremely complex phenomenon, which involves many other aspects of host defenses and immunological functions, well beyond the socio-economic factors that may affect HIV-wasting, and its similarities with starvation, [68]. Moreover, this relationship appears to be complicated by the nutritional and metabolic side-effects of HAART. Nausea and anorexia were frequent in HIV-1 infected patients before the advent of HAART, and remain a jfrequent side-effect of the treatment. The disease may produce an imbalance between cholecystokinin and betaendorphins, both modulators of patterns of consumption, that may play a role in reduced intake and wasting in HIV-1 infected subjects [69]. Among the factors that may contribute to the onset of HIV-wasting are the type of therapy used, alteration in appetite caused by opportunistic infections, cytokine profile, and psychological changes produced by socioeconomic factors and the disease process [54] In addition, both lack of appetite and lower intake have been associated with lesions in the gastrointestinal tract caused by the HIV itself or opportunistic infections [70,71]. The contribution of measures of body composition and resting energy expenditure (REE) to wasting in HIV infection is still controversial. REE showed a modest but significant increase in asymptomatic HlV-seropositive men, which was largely balanced by increases in caloric intake [72,73]. During secondary infections, thus, substantial hypermetabolism has been foimd to be a significant factor in weight loss [74]. In HIV-l infected patients, diarrhea and destruction of the gastrointestinal lining by HIV or secondary infections contribute to malabsorption, an important factor in micronutrient deficiencies. Altered levels of blood lipids have been documented in HFV-asymptomatic and symptomatic individuals. Hypocholesterolemia and hypertriglyceridemia were the mostfi^equentlyreported lipid disturbances before the initiation of potent antiretroviral treatment [75-77]. These lipid levels were associated with viral activation and the accompanying immune response, especially cytokine production [40]. With the advent of HAART, the clinical course, prognosis, and survival of HIV infected patients have greatly improved. Its prolonged use, particularly of the protease inhibitors (PI), however, has been associated with adverse changes in lipid profiles and increase in coronary heart disease (CHD) risk [78-80], and with a number of metabolic abnormalities collectively named metabolic syndrome [81-83]. A number of studies have reported metabolic abnormalities and increased CHD risk in HlV-seropositive cohorts [84-89]. In contrast, our studies in Miami showed that HAART with or without PI did not significantly impact the 10-year CHD risk estimate or metabolic syndrome in a cohort of 119 HIV-positive chronic drug users. Of the total cohort, 71% were on HAART, and 36.5% were receiving PI. The estimated effect of PI, however, was positively and

304

Selenium: Its molecular biology and role in human health

significantly related to triglyceride levels (effect estimate=95.81; 95% CI:39.40, 152.21; p<0.01), a risk factor for CHD, after controlling for age, gender, smoking, viral load, CD4 cell count, and BMI [Unpublished data, Baum et al.]. HIV-1 and micronutrient deficiencies Although the effect of the HIV-l virus on the immune system is the primary cause of the immunodeficiency associated with this disease, micronutrient deficiency further contributes in a significant and reversible manner to the observed immune dysfunction [90]. Nutrient deficiencies, independent of immune function and HAART accelerate disease progression and mortality by influencing the degree of oxidative stress and viral expression, contributing to morbidity. Oxidative stress has been reported during the early and advanced stages of HIV-l infection [91]. The accumulation of Reactive Oxygen Species (ROS) may contribute in several ways to disease progression. Oxidative stress has been linked to HIV programmed cell death (apoptosis) of T-lymphocytes [92], to alterations in the HIV-l promoter that may produce progression to AIDS in patients with latent HIV [93], to the development of AIDS Kaposi sarcoma [94], and to increased risk of coronary heart disease in those receiving treatment, especially Pl-containing treatments [95]. Oxidative stress may contribute to neural damage and to the onset of sjanptoms of neural deficiencies and even to HIV-encephalopathy [96]. In addition, oxidative stress may induce alterations in the interleukin profile, contributing to the immune dysregulation and increased viral replication observed during the progression of HIV-l infection to ADDS [97,98]. Oxidative damage observed in chronic infections are aggravated by other changes in the antioxidant defense system in plasma as well as in various tissues, including changes in levels of ascorbic acid, B vitamins, tocopherols, carotenoids, selenium and glutathione, [42-44,99,100]. Alterations in nutritional status are widespread among various HIV-l infected cohorts including asymptomatic and symptomatic homosexual men, drug users, and children [33,101]. The interaction between nutrition and HIV-l infection creates a fatal vicious cycle. Beyond the devastating effects of the disease on nutritional status, each population has characteristic socioeconomic, psychological and physiological factors that modulate and aggravate these effects. In turn, nutritional status has the potential for influencing the rate of disease progression, and mortality in ways that are closely related to the specific dietary practices and nutrition-related risk factors of each population. Some of the examples of population-specific nutritional risk factors found in our cohorts are gender distribution, rate of obesity, effects of type of drug abuse, competing nutritional demands placed

The role of selenium in HIV/AIDS

305

by growth in children, pregnancy, breastfeeding or co-morbidity in women, and characteristics of Hfe style [33,101,102]. In HIV-infected individuals, who suffer from an underlying immunodeficiency disorder, defining the role of nutrition as cause or effect is difficult. Our studies have evaluated the influence of specific nutrient deficiencies upon immune function, disease progression and mortality by following large numbers of HTV-l infected subjects longitudinally, and carefully documenting nutritional, immunological, and health status. Interactions between immune fimction and specific nutrient deficiencies in HFV-l infection have been documented to occur with parameters of protein (albumin and prealbumin) [103] and lipid status [104], trace elements (selenium and zinc) [8,9,33,100], and vitamins A, E, Be, and B ^ [7-9,10,12]. In addition to the well recognized relationships between wasting and disease progression and mortality in HIV-1 infection, several studies have explored the relation of specific nutrient deficiencies to HIV-l related morbidity and mortality [10,12,14-16]. Development of low plasma levels of vitamins A, Bg and B12 has been associated with faster disease progression, whereas normalization of vitamin A, vitamin B12, and zinc levels has been linked to slower disease progression [9]. Moreover, Tang and colleagues [12] have reported a nearly twofold increase in risk of progression to AIDS in HTV-l infected subjects with low serum vitamin B12 concentrations, and both vitamin A deficiency and wasting have been associated with increased mortality in HFV-l seropositive drug users [105]. Fawzi and colleagues [106] recently reported that supplementation of 1078 HIV-positive pregnant women in a double-blind, placebo-controlled clinical trial in Tanzania, those supplemented with a nutritional formula that included B-vitamins, vitamin C and vitamin E had significantly higher CD4+ and CD8+ cell counts and significantly lower viral loads, as compared to those receiving placebo. In addition, multivitamin supplementation was also associated with reductions in the relative risk of death related to AIDS (RR=0.73; p=0.09), and decreased the relative risk of progression to WHO stage 4 (RR=0.50; p=0.02). m v and selenium deficiency While several micronutrient deficiencies may contribute to HFV disease progression, selenium deficiency has been strongly and independently associated with mortality in HFV/AIDS [1415]. The prevalence of selenium deficiency increases from 2% to 4% in asymptomatic individuals to 75% in Stage IV AIDS [14,107,108]. In our longitudinal studies, selenium deficiency was the only nutrient deficiency that was found to be independently associated with mortality and disease progression in multivariate statistical models that also included CD4 cell count at baseline

306

Selenium: Its molecular biology and role in human health

and over time, protein status, deficiency of vitamins A and B^, and zinc. This finding was corroborated by three independent studies in cohorts of HIV-infected homosexual men, drug users and children. [14,33,101]. Selenium deficiency has showTi a significant effect on disease progression, even when controlling for deterioration of overall nutritional status utilizing plasma levels of albumin and specific nutrients, and disease status with CD4 cell count. Despite the importance of other antioxidants in immune function as protection against oxidative stress, only selenium deficiency was an independent predictor of survival. The dramatic, increased risk of mortality (10.8, p=0.002) with selenium deficiency may be related not only to its role in maintaining immune competency, but also in modulating viral expression and protecting against the oxidative damage caused by the HIV infection [109-113]. In an HIV-infected cohort in Canada, dietary intake of selenium was found to be strongly and inversely correlated with plasma malondialdehyde (MDA), a measure of oxidative stress (R2=0.12, p=.36) [24] Selenium deficiency has also been associated with increased HIV infectivity in two large trials in Afiica. In Kenya, Baeten and colleagues [21] tested 318 HFV-l seropositive women for selenium status and quantified HIV-l DNA in vaginal and cervical specimens in a cross-sectional study. After adjusting for CD4 count and vitamin A deficiency, selenium deficiency was associated with a threefold increase in HIV-DNA genital shedding (adjusted odds ratio = 2.9, 95% confidence interval: 1.0-8.8, p =0.05). In a similar cohort that was followed by the same research group in Kenya, double-blind supplementation with selenium or placebo in 400 HIV1-seropositive women, who were not receiving antiretroviral treatment, resulted in higher CD4 (+23 cells/muL, P = 0.03) and CDS (+74 cells/muL, P = 0.005) counts compared to those in the placebo group, but genital shedding was increased [114]. In a more recent and larger study in Tanzania, Kupka and colleagues [22] measured plasma selenium in 670 HIV-positive pregnant women between 12-27 weeks of gestation and followed the motherchild pairs prospectively for 24 months after delivery. They concluded that low plasma selenium levels were associated with increased risks of fetal and child mortality, and intrapartum mother-to-child transmission. Selenium supplementation The high risk of HIV-related mortality associated with selenium deficiency underscores the strong association of selenium deficiency with HIV disease progression, morbidity, infectivity and mortality and suggests the importance of maintaining optimal selenium status in HIV-l infected men and women. Studies of supplementation with selenium have resulted in striking findings [115-117]. In an HIV infected individual, supplementation with selenium

The role of selenium in HIV/AIDS

307

may help to increase the enzymatic defense systems [118], and has been linked to increased blood selenium levels and an improvement in general health [119-121]. Symptoms associated with selenium deficiency similar to the cardiac complications observed in Keshan Disease in China, [122], have been described in a child with HIV/AIDS and, upon supplementation with selenium (4 fig/kg), the symptoms improved [123]. Rather than decreasing the importance of antioxidant chemoprevention, chronic HAART is creating new research challenges for the role of antioxidants in HIV-1 disease. In our preliminary research in HFV-l infected chronic drug users receiving HAART, low selenium levels have been significantly associated with hyperglycemia, whereas thrombocytopenia has been significantly related to selenium deficiency, but not to HAART [124]. Lipodystrophy, hyperlipidemias and insulin resistance in patients receiving HIV Pis [125] may increase the long-term risk of oxidative damage associated with development of atherosclerosis and coronary heart disease [126], with antioxidant research in HIV/AIDS becoming an even more pressing matter. Supplementation with 100 )ig of selenium daily, compared to placebo, has produced significant improvement in the immune response to viral vaccine in mildly selenium-deficient and otherwise healthy participants in a British study [31]. In cancer research, long term selenium supplementation has produced significant results on the incidence of, and mortality from, carcinomas of several sites [127]. In a double-blind, placebo controlled study, healthy participants received a nutritional dose (200 )Lig/daily) of selenium, regardless of their plasma selenium status. The blinded phase of the trial was stopped early, as the selenium-treatment group exhibited a 51% reduction in total cancer mortality and 41% reduction in total cancer incidence as compared to the placebo group. In other studies, nutritional supplementation of selenium has been shown to significantly reduce the incidence of primary liver cancer in China [128], and provide significantly greater resistance to Aflatoxin Bl-induced carcinogenic damage in lymphocytes isolated from healthy human subjects administered daily selenium [129]. The chemoprevention trials in the United States and China have demonstrated that prolonged supplementation with nutritional doses of selenium (50 to 200 |ag/daily) is safe, with a low risk of toxicity. In a 30-month, double-blind, placebo-control trial of selenium supplementation in 186 HIV-positive drug users in Miami, supplementation with 200 fig of selenium yeast slowed disease progression as measured by CD4 cell counts. Patients supplemented with selenium also showed significantly increased vigor, less anxiety [130], decreased hospitalizations and lower cost of health care compared to those receiving a placebo [131].

308

Selenium: Its molecular biology and role in human health

Short-term supplementation of 200(ig of selenium for 6 weeks in a cohort of 400 women in Kenya demonstrated similar positive results on CD4 and CDS cell counts when compared to those receiving a placebo [114] In summary, the cumulative findings from selenium supplementation trials suggest that administration of selenium to HIV-l seropositive persons at nutritional levels is safe, and may be an effective method to slow disease progression, reduce morbidity, enhance survival, and might even improve the effectiveness of treatment by stimulating immvine reconstitution. Large selenium supplementation trials in HIV-l infected individuals are now underway in the United States and abroad. References

1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34.

J Roberts 1996 Brit Med J 312:796 AIDS Epidemic Update: December 2004 UNAIDSAVHO 2004 www.unaids.org/ B Lindtjom 1990 Trop Geogr Med 42:365 NS Scrimshaw 1986 FedProc 45:2421 M Lipton 2001 Proceedings of the Nutritional Society 60:203 C Lau, AS Muula 2004. Croat Med J 45:402 MK Baum, G Shor-Posner, Y Lu et al 1995 AIDS 9:1051 MK Baum 1996 Nutrition 12:124 MK Baum, G Shor-Posner 1998. Nutrition Reviews 56(S1-S2):58 AM Tang, NMH Graham, AJ Saah 1996 Am J Epidemiol 143:1244 RDSemba, NMH Graham, TWaleskaetal 1993. Arch Intern Med 153:2149 AM Tang, NMH Graham, RD Semba et al 1997 AIDS 11:613 H Lai, S Lai, Shor-Posner G et al. 2001 JAcquir Immune Defic Syndr.27-.56 MKBaumetal. \997 J Acquir Immune Defic Syndr Hum Retroviral 15:370 R Kupka, GI Msamanga, D Spiegelman et al. 2005 Eur J Clin Nutr 59:1250 R Kupka, Msamanga GI, Spiegelman D et al. 2004 J Nutr. 134(10):2556-60 G Famularo, SM Moretti, S Marcellini S et al. 1997 ^/DS 11:185 P. Aukrust, F Muller 1999 Nutrition 15:165 S Moretti, G Famularo, S Marcellini S. 2002. Antioxid Redox Signal. 4(3):391 MK Baum, JJ Javier, Mantero-Atienza et al. 1991. J Acquir Immune Defic Syndr 4:1218 JM Baeten, SB Mostad , MP Hughes, et al. 200UAcquir Immune Defic Syndr. 26:360 R Kupka, M Garland, G Msamanga, et al. 2005 J Acquir Immune Defic Syndr 39:203 RR Jamason, BA Carlson, M Butz, et al. 2002. J Nutr 132:1830 JM McDermid et al 2002. J Acquir Immune Defic Syndr 29:158 MA Beck, A Lavander 1998 Annu Rev Nutr 18:93 MA Beck, J Handy, OA Levander. 2004 Trends Microbiol 12:417-23. MK Baum, MJ Miguez-Burbano, A Campa et al 2000 J Infect Dis 182(Suppl 1 ):S69 P Kidd. 2003 Altern Med Rev 8:223-46. NS Scrimshaw 1997 Am J Clin Nutr 66:464S JESpallholz 19SI Adv Exper Med Biol 135A3 CS Broome, F McArdle, JAM Kyle et al. 2004. Am J Clin Nutr. 80:154 MJ Miguez-Burbano, A Campa, G Shor-Posner et al STI and the Millennium, A Joint Meeting of the ASTDA and the MSSVD, Baltimore, Maryland, May 3-6, 2000. MK Baum 2001 JInsti Rosell. Proc of the Salt Lake City Symp. Assoc Bioinorg Sci p43 TR Mosman, S Sad 1996 Immunol Today 17:138

The role of selenium in HIV/AIDS

35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85.

309

M Roy et al 1993 Proc Soc Exp Biol Med 202:295 RC McKenzie, TS Rafferty, GJ Beckett 1998. Immunol Today 19:342 L Kiremidjian-Schumacher, M Roy, HI Wishe et al 1994 Biol Trace Elem Res AIMS M Roy et al 1993 Proc Soc Exp Biol Med 202:295 MP Look, JK Rocstroh, GS Rao et al 1991Biol Trace Elem /?e.s 56:31 U Keller 1993 Supportive Care in Cancer 1:290 PA Haslett 1998 Semin Oncol 25:53 K Hori, D Hatfield, F Maldarelli et al \991 AIDS Res Hum Retroviruses 13:1325 M Moutet, P d'Alessio, P Mlette et al 1998 Free Radic Biol Med 25:270 C Sappey et al 1994 AIDS Res Human Retrovir 10:1451 S Endres, R Ghorbani, VE Kelley et al 1989 N EnglJ Med 320:265 MK Hellerstein, K Wu, M McGrath et al 1996 JAcquir Immune Defic Syndr 11:258 D Serwadda, RD Murgewa, NK Sewankamboo, et al 1985 Lancet 2:849 Simply stated...are people still wasting? Res Initiat Treat Action 1998 4:15. RRoubennoffetal2002 Am J Physiol Endocrinol Metab 2S3:E13S GS Reiter 1996 AIDS Clin Care 8:89-91,93,96 DP Kotler 2000 JAcquir Immune Defic Syndr 25 Suppl 1 :S81 DA Wheeler et al 1998 JAcquir Immune Defic Syndr Hum Retrovirol 18:80 AM Tang, J Forrester, D Spiegelman et al 2002 J Acquired Immune Defic Syndr 31:230 A Campa, Z Yang, S Lai et al. 2005. Clin Infect Dis. 41:1179 JP Palenicek, NM Graham, YD He et al 1995 J Acquired Immune Defic Syndr 10:366 WD DeWys, C Begg, PT Lavin et al 1980 ^w J Med 68:683 U Suttmann, J Ockenga, O Selber et al 1995 JAcquir Immune Defic Syndr 8:23 Z Yang et al 2004 Intematl HIV/AIDS Conf, Bangkok, Thailand, July 7-13, 2004 CA Wanke, M Silva, TA Knox et al 2000 Clin Infect Dis 31:803 LM Hodgson, H Ghattas, H Pritchitt et al 2001 AIDS 15:2341 RS Beach, PF Laura 1985 Ann Int Med 99:565 RH Gray 1983 Am JPubl Health 73:1332 VK Jain, RK Chandra 1984 Nutr Res 4:537 NlJPatonetal \991 J Acquir Immune Defic Syndr Hum Retrovirol \A:\\9 CG Neumann, GJ Lawlor, ER Strehm et al 1975.4m J Clin Nutr 28:89 S Cunningham-Rundless \982 Am J Clin Nutr 35:]02 AS Fauci 1988 Science 239:717 C Grundfeld, K Feingold 1992 N Engl J Med 327:329 F Amalich, P Martinez, A Hemanz et al 1997 AIDS 11:1129 A Schwenk, B Berger, D Wessel et al AIDS 7:1213 ED Schuartz, JB Greene 1992 Seminar in Liver Disease 12; 142 RD Sharpstone, CP Murray, HM Ross et al AIDS 10:1377 MJT Hommes, JA Romijin, E Endert 1991 Am J Clin Nutr 54-3\\ C Grunfeld, M Pang, L Shimizu et a\\992 Am J Clin Nutr 55:455 C Grunfeld , D Kotler, R Hamdeh. 1989 Am J Med 86:27 G Shor-Posner, A Basit, Y Lu 1993 Am J Med 94:515 R Zangerle, M Sarcletti, H Gallati 1994 J Acquired Immune DefSynd 7:1149 ML Gougeon, L Penicaud, M Fromenty et al 2004 Antivir Ther 9:161 HH Schmidt, G Behrens, J Genschel et al 1999 Antivir Ther 4:163 G Behrens, A Dejam, RE Schmidt 1999 AIDS 13:F63 National Cholesterol Education Program (NCEP). 2001 JAMA 287:356 HB Brewer Jr. 2003 Clin Cardiol 26(4 Suppl 3):III19 RL Talbert. 2003 . Am J Health Syst Pharm 60:S3 B Isomaa, P Almgren, T Tuomi et al 2001 Diabetes Care 24:683 C Grunfeld, M Pang, W Doerrler et al 1992 J Clin Endo Met lA: 1045

310

86. 87. 88. 89. 90.

Selenium: Its molecular biology and role in human health

A Blum, V Hadas, M Burke et al 2005 Clin Cardiol 28:149 BF Asztalos, EF Schaefer, KV Horvath et al 2006 Atherosclerosis 184:72 BE Hurwitz, NG Klimas, MM Llabre et al. 2004 Cardiovasc Toxicol 4:303 JS Currier, A Taylor, F Boyd et al 2003 J Acquir Immune Defic Syndr 33:506 RK Chandra ed 1992 Nutrition Immunology ARTS Biomedical Publishers and Distributors, St. John's, Newfoundland, Canada pp 241 91. A Favier, C Sappey, P Leclerc et al 1994 Chem Biol Interact 91:165 92. TS Dobneyer et al 1997 Free Radical Biology & Medicine 22:775 93. GW Pace , CD Leaf 1995 Free Radical Biol Med 19:523 94. SR Mallery, RT Bailer, CM Hohl et al 1995 Journal of Cellular Biochemistry 59:317 95. BE Hurwitz, NG Klimas, MM Llabre et al 2004 Cardiovasc Toxicol 4:303 96. S Dewhrst, HA Gelbord, SM Fine 1996 Molecular Medicine Today 2:16 97. F Muller, AM Svardal, P Aukurust 1996 Am J Clin Nutrit 63:242 98. P Aukurist, AM Svardal, F Muller 1995 Blood 86:258 99. E. Bemasconi, M Uhr, L Magenta, et al 2001 AIDS 15:1081 100. RS Beach, E Mantero-Atienza, G Shor-Posner et al 1992 AIDS 6:701 101. A Campa et al 1999 J Acquir Immune Defic Syndr Hum Retrovirol 20:508 102. RD Semba, WT Caiaffa, NMH Graham et al 1995 J Infect Dis\l\:\\96 103. CM Huang, M Ruddel, JE Ronald 1988 Clinical Chemistry 34:1957 104. Grunfeld C et al 1992 J Clin Endocrinol Metab lA: 1045 105. RD Semba, WT Caiaffa, NMH Graham et al 1995 7 Infect Z)w 171:1196 106. WW Fawzi,, GI Msamanga, D Spiegelman et al. 2004 N EnglJ Med 351:23 107. ACampaetal \999 J Acquir Immune Defic Syndr Hum Retrovirol 2Q:5Q% 108. C Avellana, B Dousset, T May et al 1995 Biol Trace Elem Res 47:133 109. EW Taylor, CS Ramanathan, RG Nadimpalli et al 1995 Antiviral Res 26:A271 110. M Witted ed Comput Med Public Health and Biotech 1: Singapure: World Sci pp 285 111. EW Taylor, CS Ramanathan 1996 J Orthomol Med 10:131 112. EW Taylor, RG Nadimpalli, CS Ramanathan 1997 Biol Trace Elem Res 56:63 113. BM Dworkin, WS Rosenthal, GP Wormser, et al 1988 Biol Trace Elem Res 20:86 114. RS McClelland et al 2004 J Acquir Immune Defic Syndr 37:1657 115. LC Clark, GS Combs, BW Tumbull 1996 FASEB 10:550 116. JYLi,PRTaylor,BLietal 1993 J Natl Cancer Inst S5:\ 492 117. WJ Blot, JY Li, PR Taylor et al 1993 J Natl Cancer Inst 85:1483 118. MC Delmas-Beauvieux, E Peuchant, A Coucouron et al 1996 Am J Clin Nutr 64:101 119. A Cirelli, M Ciardi, C De Simone et al 1991 Clin Biochem 24:211 120. L Olmsted, GN Schrauzer, M Flores-Arce et al 1989 Biol Trace Elem Res 20:59 121. GN Scharauzer, J Sacher 1994 Chem Biol Inter 91:199 122. Keshan Disease Research Group 1979 Chin MedJ91Al\ 123. AL Kavanaugh-McHugh et al 1991 JPEN JParenter Enteral Nutr 15:347 124. MJ Miguez-Burbano et al 2001 Abstract No 8662 FASEB 2001, Oriando, Florida 125. A Carr, K Samaras, S Burton et al 1998.4/Z)512:F51 126. BHalliwell 1995 Am J Clin Nutr 6\:670S 127. LC Clark, GF Combs, BW Tumbull et al 1996 JAMA 276:1957 128. SY Yu, YJ Zhu, WG Li et al 1991 Biol Trace Elem Res 29:289 129. SY Yu YJ Zhu, WG Li 1998 Biol Trace Elem Res 15:231 130. G Shor-Posner, R Lecusay, MJ Miguez et al 2003 Int J Psychiatry Med 33:55 1 3 1 . x Burbano, MJ Miguez-Burbano, K McCoUister K et al 2002 HIV Clin Trials 3:483

Chapter 27. Effects of selenium on immunity and aging Roderick C. McKenzie Laboratory for Clinical and Molecular Virology, Royal Dick Veterinary School, University of Edinburgh, Summerhall, Edinburgh EH9 IQH, UK

Geoffrey J. Beckett Department of Clinical Biochemistry, University of Edinburgh, Combined Laboratories, The Royal Infirmary of Edinburgh, 51 Little France Cresdent, Edinburgh, Scotland, EH16 4SA, UK

John R. Arthur Division of Vascular Health, Rowett Research Institute, Bucksburn, Aberdeen, Scotland, AB21 9SB, UK

Summary: An adequate selenium intake is necessary for the optimum function of both cellular and humoral immune processes. This chapter addresses the roles of selenium in immunity including influences on eicosanoid metabolism and modulation of adhesion molecule and cytokine expression. The effects of selenium on humoral and cell-mediated immunity are also reviewed. The involvement of selenium in the onset of immunemediated diseases is considered along with consequent effects on aging. Introduction The immune system relies on many processes including the deployment of cells generating oxidative stress as a defence against microbial pathogens, coordinated regulation of adhesion molecules and the expression of soluble mediators such as eicosanoids and cytokines and their receptors. Selenium has the potential to influence such immunity at many stages. The immune system uses the generation of reactive oxygen species for microbiocidal activity [1]; for example, in the burst reaction of the neutrophils. In limited doses, release of reactive oxygen species generates inflammation and destroys microbial invaders, but chronic production of these reactive species may lead to oxidative damage to the host. The host defends against this using antioxidant systems. Selenium was first postulated to be immunologically important when radiolabelling experiments in dogs showed incorporation of selenium into leukocytes. Further studies showed that selenium was incorporated into the selenoenzyme glutathione peroxidase

312

Selenium: Its molecular biology and role in human health

(GPx) [2.3]. Additional indirect evidence for a role of selenium in immune function was its incorporation into immune-important organs such as spleen, liver and lymph nodes [4]. Throughout the 1970's and the 1980's, research into the immunostimulatory effects of selenium increased dramatically as summarised by Spallholz et al [5]. Immuno-protective roles for selenium acting through antibody and complement responses were demonstrated. Selenium was shown to augment these responses to both natural and experimental immunogens such as tetanus toxoid, typhoid toxin, sheep red blood cells, and immunoglobulins (see [5] and references therein). Aspects of immunity were altered in selenium-deficient hosts. These include: defective neutrophils function, stimulation of non-specific immunity in rabbits, increases in antibody titres to bacterial and mycotic antigens, increased H2O2 release during neutrophil phagocytosis, decreased neutrophil numbers, decreased antibody response to sheep red blood cells, decreased neutrophil fungicidal activity, inactivation of NADPH-dependent generation of superoxide by granulocytes, reduced natural killer cell activity, and increased mortality due to candidiasis. Selenium injection or supplementation enhances vaccine immunity against malaria, increases antibody-producing B-cell numbers, increases T-celldependent antibody production (T-cell helper), increases selenium concentration in Ijonph nodes and immune tissue, increases neutrophil numbers, increases Ij^nphocyte GPx activity, and increases concentration of selenium in neutrophils (see [5] and references therein). These observations support a role for the trace element in the immune system. In some cases, however, high levels of selenium supplements decreased immunity [5]. This probably reflects the need for optimal forms and levels of selenium for immunostimulatory properties. Modulation of the immune system by selenium probably involves one or more of the following mechanisms: 1) Detoxification of organic hydroperoxides and hydrogen peroxide. 2) Regulation of the balance of activity in the eicosanoid synthesis pathways, leading to preferential sjTithesis of leukotrienes and prostacyclins over thromboxanes and prostaglandins. 3) Down regulation of cytokine and adhesion molecule expression. 4) Upregulation of interleukin-2 receptor expression, leading to enhanced activity of lymphocytes, natural killer and lymphokine activated killer cells. 5) Regulation of metabolism and cell function by the > 25 selenoprotein genes identified in mammalian systems (see Chapter 9). Seleno-enzymes as peroxynitrite reductases Protection against oxidative damage is the mechanism by which selenium is likely to exert some protective effects on immunity. Nitric oxide (NO) has

Effects of selenium on immunity and aging

313

microbiocidal effects, yet under oxidative conditions in which superoxide (O2") is produced concomitantly (by neutrophils and mononuclear phagocytes), the highly reactive and destructive oxidant, peroxynitrite (ONOO) is produced. Experiments performed in vitro show that the addition of selenocysteine and selenomethionine protects plasmid DNA from ONOOmediated damage [6]. Selenomethionine, selenocysteine and the GPx mimic ebselen were more effective protectants than selenite, showing specificity in the reaction [7]. However, selenoenzymes including GPx and TR also protect from ONOO-mediated damage [8-9]. The roles of the GPx and TR families as a peroxynitrite reductase need to be further elucidated by studies in intact cells. Selenium and eicosanoid metabolism The eicosanoids include the leukotrienes, the thromboxanes, the prostaglandins and the lipoxins (see [10] for review). Selenium (mediated through GPxl and GPx4) probably has anti-inflammatory effects, preventing the release of inflammatory mediators through reduction of organoperoxides which mediate production of active leukotrienes. The conversion of arachidonic acid to prostaglandin GG2 is catalysed by cyclo-oxygenase. This enzyme requires a minimal level of peroxide to function. However, if peroxide levels in the cell are high, cyclo-oxygenase activity is also inhibited [11,12]. Despite the fact that reduction of the hydroperoxyeicosatetraenoic acids to hydroxyeicosatefraenoic acids requires the reductive power of peroxidases, the resultant products are generally proinflammatory. Anti-inflammatory activity of selenium may be explained by the ability of selenoenzymes to inhibit the 5- and 15-lipoxygenase enzymes, which convert arachidonic acid to the 5-hydroperoxyeicosatetraenoic acid precursor of the leukotrienes [13,14]. The conversion of selenite to selenide (which inhibits lipoxygenase) appears to be catalysed by a reaction of NADPH with TR [15]. Selenium deficiency also leads to decreased leukotriene B4 synthesis, this leukotriene impairs the functions and mobility of phagocyctes [16]. Another important effect of selenium deficiency is the disturbance in the balance of production of the pro-coagulant thromboxanes and the anticlotting prostacyclins [17,18]. This could underlie the prevalence of atherosclerosis in populations which have low dietary selenium intake (see [19] for review). Evidence suggests that selenium supplementation can only protect against atherosclerosis in populations whose selenium intake is below the recommended daily allowance [20]. Platelet GPxl activity is sensitive to the effects of selenium deficiency in humans, this being associated with increased aggregation, thromboxane B2 production and the synthesis of lipoxygenase-derived products. In such people, selenium supplementation

314

Selenium: Its molecular biology and role in human health

increases platelet GPxl activity and decreases hyperaggregation (see [11] for review). Effect of selenium on adhesion molecules and cytokines Pro-inflammatory cytokines, such as tumor necrosis factor-a and interleukin1, induce many of the adhesion molecules which are upregulated in inflammation. Existing evidence is more consistent with a selenium effect on adhesion molecule expression through regulation of cytokine release. In general, selenium-deprived cells or endothelium from selenium-low individuals have a higher constitutive expression of adhesion molecules, and selenium supplementation decreases their expression. Endothelial cells obtained from asthmatic patients, had significantly higher constitutive expression of P-selectin, vascular adhesion molecule-1, E-selectin and intercellular adhesion molecule than cells from normal subjects. However, after 3 months of selenium supplements a significant decrease in vascular adhesion molecule-1 and E-selectin expression was observed [21]. This was also confirmed by treatment of cultured endothelial cells with 6 nM to 48 nM selenium. Similarly, bovine endothelial cells grown under selenium-deficient conditions and stimulated with tumor necrosis factor-a, had higher levels of E-selectin, P-selectin expression and intercellular adhesion molecule-1, which was manifested by greater adherence of neutrophils [22]. A GPx mimic inhibited the expression of intercellular adhesion molecule-1 and vascular adhesion molecule-1 and GPx analogs prevented tumor necrosis factor-a-stimulated expression of P-selectin and E-selectin, as well as tumor necrosis factor-a and interleukin-1-stimulated interleukin-8 release in human endothelial cells [23, 24]. Oxidative stress induces several pro-inflammatory cytokines including: interleukin-1, interleukin-6, interleukin-8 and tumor necrosis factor-a possibly through the activation of the transcription factors AP-1 and N F - K B [25]. Pre-incubation of keratinocytes with selenium abrogated upregulation of the mRNAs for interleukin-6 and interleukin-8 [26] in response to ultraviolet radiation B, a potent environmental oxidative stress. IL-10 immunostaining in murine keratinocytes was also suppressed by pretreatment with selenomethionine before UVB treatment [27]. The speciation of selenium is probably important for the frace element to exert optimal effects. For example, selenomethionine seems a better protectant than selenite against ultraviolet radiation B-induction of cytokines. Since GPxl depletes reduced glutathione, cytokine release into the culture fluid of endothelial cells increased after treatment with selenite [28]. Also, selenite (but not selenomethionine) supplementation of BALB/c mice increased the release of interleukin-1 and tumor necrosis factor-a from phytohaemagluttinin-P stimulated splenic macrophages [29]. This could result from the pro-oxidant effects of high doses of selenite. Moreover

Effects of selenium on immunity and aging

315

selenite and selenocystamine, but not selenomethionine, increased oxidative stress, oxidative DNA damage and apoptosis in keratinocytes [30]. There are few reported studies on the effects of cytokines on selenium metabolism. Treatment of the liver cell line HepG2 with interleukin-ip, tumor necrosis factor-a, or interferon-y had no effect on selenoprotein P expression. However, transforming growth factor-P (100 pM for 48 h) led to a 21% decrease in expression of selenoprotein P mRNA [33]. Incubation with transforming growth factor-P also down regulated mRNA and activities of GPxl and catalase [33]. Effects of selenium on cell-mediated and humoral immunity Studies of the effects of selenium on cell-mediated immune cells and on antibody production have been reviewed up to 1990 [5]. Many of these studies were performed by veterinary researchers and used simultaneous supplements of selenium and vitamin E, since the nutrients can act synergistically, and may substitute for each other. This is a factor that needs to be borne in mind when interpreting these early studies. In rats, selenium deficiency decreased IgG production slightly, but had no effects on IgA production. However, IgM production was greatly decreased and lowered further by vitamin E deficiency. Selenium supplementation partially ameliorated the vitamin E deficiency-induced decreases in IgA and IgG [34]. IgG levels were higher in cows given 120 ng/kg selenium and calves fi"om these cows had higher post suckle serum IgG levels. Thus, maintaining optimal selenium intake may promote health of offspring as well as of the mothers. The effect of vitamin E and selenium supplements on the immune responses of domestic animals has been reviewed by Finch and Turner in 1996 [35]. Selenium-enriched diets given to poultry improve their antibody responses to Salmonella and aflatoxin vaccination [36]. A combination of vitamin E and selenium supplements in the diet increased both antibody titres to Newcastle disease virus and gave maximum gain in body weight [37]. In sheep vaccinated against Chlamydia psittaci, which causes abortion, injection of selenium (0.1 mg/kg) alone increased the Chlamydia antibody response. However, this was decreased if coadministered with vitamin E [38]. Selenium supplementation improves responses in most studies of cellmediated immunity. Selenium deficiency in rats results in decreased candidiacidal activity in neutrophils and impairs survival after Staphylococcus aureus infection [39]. Polymorphonuclear cells also provide defense against mastitis in cattle. Supplementation of such cells in vitro with selenium and vitamin E increased superoxide production and migration after stimulation with phorbol esters [40 Selenite enhances chemotaxis of macrophages. Murine infection with the parasite Trypanosoma cruzi was treated with selenium at 0 ppm, 2 ppm, 4

316

Selenium: Its molecular biology and role in human health

ppm, 8 ppm, or 16 ppm as sodium selenate in drinking water [41]. Sixty four days after infection, the mice without selenium supplements had all died. But 60% of the animals in groups supplemented with 4 and 8 ppm selenium survived. Survival was much less in the group fed 16 ppm, indicating an optimal dose is important for selenium protection. Studies with experimental and agricultural animals all support a role for selenium, albeit not always consistent, in maintaining components of the immune system and the ability to resist parasitic infection [42-44]. The importance of selenium and other nutrients in maintaining resistance to viral infection has been proved in a number of studies mainly with mice [44]. As well as decreasing selenium status, knockout of glutathione peroxidase 1 also enhances the negative effects of viruses in mice [44]. Effects of selenium on interleukin-2 receptor and lymphocytes Selenium augments the performance of both T- and B-lymphocytes and perhaps the common effect is through the up-regulation of the interleukin-2 receptor a and p subunits which results in a greater number of high affinity interleukin-2 receptors in mice [45] and humans [46]. This is accompanied by enhanced proliferation and differentiation into cytotoxic effector cells [47]. Selenium supplementation in humans (200 )J.g/day for eight weeks) also up-regulates the activity of cytotoxic T-cells (118%), natural killer cells (82%) [47] and down-regulates the activity of suppressor T-cells. The lytic capabilities of natural- and lymphokine activated-killer cells in humans, were increased, purportedly by up-regulation of interleukin-2 receptors [48]. In rats given selenite (0.5 ppm, 2.0 ppm or 5.0 ppm) in the water supply, the response of natural killer cells was boosted in the group receiving 0.5 and 2.0 ppm selenite, but cell activity in the 5.0 ppm group was similar to that of unsupplemented animals. Antibody synthesis was not significantly increased, but fell in the group given 5.0 ppm selenium. Production of prostaglandin E2 was decreased at all selenium doses [49]. Similarly, an inhibitory effect of selenite on natural killer cell activity and lymphokine activated killer cell activity was seen in human lymphocytes supplemented in culture with 0.8 fig/ml selenite. However, this is a very large dose and may be toxic. In the same study, lymphocyte proliferation to T-cell mitogens was suppressed by selenium in the range of 0.5-1.0 fig/ml [50]. Activated T-cells have increased activity of the enayme selenophosphate synthetase, which is essential for the synthesis of selenocysteine, an obligatory step in selenoprotein synthesis [51]. This is consistent with higher concentrations of selenium found in immune-active tissues such as spleen, lymph nodes and liver [4]. The importance of selenium in maintaining cellular immunity was further supported by studies in uremic patients [52]. Patients had lower plasma selenium concentrations than controls. Supplementation with 500 ^g of selenium thrice weekly for three months

Effects of selenium on immunity and aging

317

was followed by 200 ng/day for the next three months. Although no change in lymphocyte numbers or subpopulations was observed, delayed-type hypersensitivity responses (to phytohaemoagluttinin) were significantly higher in the selenium-supplemented group after 6 months, compared with their own pre-experimental levels and a placebo group. The augmented responses dropped to pre-supplementation values 3 months after ceasing selenium supplementation. The overall conclusion was that selenium supplementation could be beneficial in uremic patients. On the other hand, there was no improvement with selenium and zinc supplements on the delayed-type hypersensitivity responses in elderly patients, despite an improved humoral immune response to influenza vaccination [53]. Selenium and immune-mediated disease in humans From the findings listed above, it may be expected that selenium would have beneficial effects on inflammatory conditions such as rheumatoid arthritis. However, there are few blinded, controlled trials exploring such possibilities. Epidemiological studies indicate that serum selenium levels are correlated with disease state. However, reports do not agree on association.s of disease with GPx activities in rheumatoid arthritis patients. This could be because sub-forms of the disease were not categorized in some studies. Moreover, in some variants of the disease neutrophil GPx activity was not increased by dietary selenium supplementation. Thus lack of effect of selenium on arthritic symptoms in some studies may reflect an already adequate selenium status. The results of several studies have been summarized (see [19,54] for review). Nevertheless, some studies have shown that low selenium status may be a risk factor for rheumatoid factor-negative (but not rheumatoid factor-positive) arthritis [55]. Because selenoproteins can be acute phase reactants, a decrease in plasma selenium is not necessarily associated with loss of antioxidant function. In Crohn's disease patients -in which immune activation may be mediated by reactive oxygen species- there was a negative correlation between plasma selenium and soluble interleukin-2 receptor and erythrocyte sedimentation rate. The soluble interleukin-2 receptor concentration is positively correlated with the degree of immune activation [56]. Oxidative stress and micronutrient deficiencies have been identified along with selenium deficiency and decreased red cell GPx activity as risk factors for the development of asthma (see [57] for review). In juvenile asthma patients with intrinsic disease, 100 \x.g selenite/day improved their clinical symptoms [58]. Protection against asthmatic wheeze has been found in adult asthma patients in England [19] and in asthmatic children in New Zealand [59]. The reasons for this are not clear, but atopic asthmatics have low platelet and red blood cell GPx activities. However, GPx activity was higher in eosinophils in both normal and asthmatic subjects (but was not different

318

Selenium: Its molecular biology and role in human health

between these groups) than in neutrophils from both groups. The higher GPx activity of eosinophils may have prolonged their survival at inflammatory sites, thus driving the inflammatory process [60]. An increase in the incidence of asthma has recently been noted in the Western World and more studies into potential benefits of increased selenium intake in asthmatics are warranted. Clinical trials on sepsis and systemic inflammatory response syndrome patients suggest that these patients have low plasma selenium and GPx activity. Two independent prospective studies showed a beneficial therapeutic response in patients given selenium supplements [61]. Selenium supplementation of otherwise healthy human subjects can also have beneficial effects on immune related processes. In UK subjects with blood selenium levels of approximately 1 (iM, pre-treatment with selenium has improved polio virus handling after vaccination. This was associated with an augmented cellular immune response manifest as increased interferon-y gamma and IL-2 and IL-10 production [62]. Other studies that have increased selenium intake in human volunteers have resulted in improved activation and proliferation of B lymphocytes [63]. Lipoxygenase activity in lymphocytes is also associated with a small nucleotide polj^norphism in GPx4 [64]. In the macrophage cell line J774.1 selenium in vitro enhances phagocytosis, degranulation and production of superoxide after stimulation. Additionally, TNFa, IL 1 and IL 6 release was enhanced when compared with selenium-deficient cells [65]. Selenium and the skin The skin is the body's largest organ and as well as its main interface with the environment. Skin is continually exposed to many stresses due to the products of commensal organisms on the surface as well as the oxidative stress and cell damage caused by exposure to ultraviolet radiation. Both selenomethionine and selenite at nanomolar concentrations can protect keratinocytes, melanocytes and fibroblasts from UV-induced cell death and apoptosis. The processes behind these effects include inhibition of oxidative DNA damage, lipid peroxidation, apoptosis, suppression of inflammatory and immune suppressive cytokine release, and modulation of p53 activity (see [66] for review). Selenium can prevent UVB-induced skin tumors in hairless mice, (see [66] for discussion); it remains to be seen if the protective effect against skin cancer in mice also operates in humans. In skin, selenium deficiency in vivo may impair immunity by decreasing infiltration of Langerhans cells after stimulation by UV irradiation [67]. Selenium and aging Throughout life, cells accumulate oxidative damage -"the oxidative theory of aging" - (see [68] for review) and the aging lymphocyte population fails to

Effects of selenium on immunity and aging

319

expand as effectively on antigenic challenge, resulting in damage to both mitochondrial and nuclear DNA. Apart from lipid peroxidation, there is an accumulation of carbonyl moieties on protein, both types of lesions being produced by oxidative stress. Due to oxidative metabolism, mitochondria accumulate damage, which helps release more reactive oxygen species, exacerbating the process [68]. For example, treatment of fibroblasts with non-lethal doses of hydrogen peroxide activates a senescence program, which leads to growth cessation. Thus, a role for GPxl and other selenoproteins in slowing cellular damage and the aging process is possible. The efficiency of the immune system declines with age and the elderly are more prone to infections than young or middle aged adults. Generally, the response to antigen challenge and the ratio of effector to naive T-cells decreases, along with a decrease in the ratio of CD4 to CDS T-cells and decreased ratio of CDS" to CD5^ B-cells [69]. There is also decreased ability of macrophages and monocytes to destroy microbes. For example, aged mice produced weak interferon-y and interleukin-2 responses to the parasite Trypanosoma musculi [70]. Decreased proliferation of spleen lymphocytes to allogeneic or mitogen stimulation from aged mice was restored by dietary selenium supplements [71]. The mechanism appeared to be via upregulation of the interleukin receptor. Selenium supplementation in vitro enhanced the previously depressed chemotactic and cytokine release capabilities of polymorphonuclear cells from elderly donors [72]. In elderly humans, low blood selenium and erythrocyte GPxl activity was correlated with lower triiodothjmnine (T3) to thyroxine (T4) ratios, mainly due to raised T4 concenfrations, and was seen with advancing age [73,74]. Selenium supplementation decreased the serum T4 concentration. The agerelated decline in T3:T4 ratios was ascribed to impaired iodothyronine-5'deiodinase activity. Impaired T4 to T3 conversion will affect general metabolism, including immunity. Longevity in areas of the world which had selenium-rich soils was noted by Foster and Zhang in 1995 [75]. Less people over 80 years of age were found in areas where the selenium-deficiency diseases Kashin-Beck and Keshan disease were endemic [75]. A hypothesis has been put forward which proposes that the areas of the world which have higher lifespan than (national) average are areas where soil selenium is high, but mercury content, which sequesters selenium, is lowest [76]. In humans, cancer is a disease associated principally with old age, pointing to an age-dependent decrease in the efficiency of the immune system to detect and desfroy tumors. This may be due to a decrease in the effectiveness of natural killer cells allied to nutritional deficiencies. In a free-living elderly Italian women (ages 90-106 years of age), the percentage of natural killer cells in the circulation was related to serum selenium content [77]. An adequate selenium intake may maintain GPx activities in aging cells. Increasing the intracellular hydrogen peroxide content by blocking GPx

320

Selenium: Its molecular biology and role in human health

activity with buthionine sulfoximine and inhibiting catalase activity by aminotriazol treatment raised the levels of collagenase mRNA [78]. Collagenase-1 activity contributes to connective tissue damage, which is a feature of tumor expansion, inflammatory disease and photo-aging. GPxl activity is lower in neutrophils from human volunteers over 65 years of age compared with cells from younger volunteers [79]. The enzyme from the elderly group had a decreased V max compared with that in neutrophils from younger donors (21-34 years of age). Furthermore, in the cells from the young group, the affinity (Km) of the enzyme for its substrate increased on neufrophil activation which, did not occur in the elderly group. Finally, telomere length decreases with age in peripheral leukocytes and this is accelerated by oxidative stress in fibroblasts [80]. A role has been proposed for GPxl in the maintenance of telomere length. The rate of telomere shortening and carbonyl group accumulation was inversely correlated with GPxl activity in fibroblasts [80]. Furthermore, experiments with human breast cells which were fransfected with DNA constructs to produce lines that had differing GPxl expression, revealed an important role for the enzyme in the protection against oxidative-induced mitochondrial DNA damage. Lines contained 100 times differing GPx activities. Exposure to 25 nM menadione for 1 hour caused approximately three-fold more single-sfrand breaks and 8-oxo-deoxyguanosine residues in the low GPx lines [81]. As well as GPx, the mitochondrial TR [82] may play an important role in limiting oxidative damage in immune cells. References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19.

JC Fantone, PA Ward 1982.4m J Pathol 107:397 JT Rotruck, AL Pope, HE Ganther, AB Swanson, DG Hafeman, WG Hoekstra 1973 Science 179:588 L Flohe, WA Gunzler, HH Schock 1973 FEBS Lett 32:132 RC Dickson, RH Tomlinson 1967 Clin Chim Acta 16:311 JE Spallholz, LM Boylan, HS Larsen 1990 Ann NY Acad Sci 587:123 I Roussyn, K Briviba, H Masumoto, H Sies 1996 Arch Biochem Biophys 330:216 K Briviba, I Roussyn, VS Sharov, H Sies 1996 Biochem 7319:1315 H Sies, VS Sharov, LO Klotz, K Briviba \997 J Biol Chem 272:27812 GE Arteel, K Briviba, H Sies 1999 Chem Res Toxicol 12:264 GR Davies, DS Rampton 1997 Euro J Gastroenterol Hepatol 9:1033 D Vitoux, P Chappuis, J Amaud, M Bost, M Accominotti, AM Roussel 1996 Annates De Biologic Clinique 54:181 MJ Pamham, E Graf 1987 Biochem Pharmacol 36:3095 O Werz, D Steinhilber 1996 Eur J Biochem 242:90 C Schewe, T Schewe, A Wendel 1994 Biochem Pharmacol 48:65 M Bjomstedt, B Odlander, S Kuprin, HE Claesson, A Holmgren 1996 Biochem 35:8511 C Gairola, HH Tai 1985 Biochem Biophys Res Commun 132:397 YZ Cao, CC Reddy, LM Sordillo 2000 Free Radical Biol Med 28:381 M Meydani 1992 Biol Trace Element Res 33:79 MP Rayman 2000 Lancet 356:233

Effects of selenium on immunity and aging 20. 21. 22. 23. 24. 25. 26.

27. 28. 29. 30. 31. 32. 33. 34.

321

JK Huttenen 1997 Biomed Environment Sci 10:220 M Horvathova, E Jahnova, F Gazdik 1999 Biol Trace Element Res 69:15 JF Maddox, KM Aheme, CC Reddy, LM Sordillo 1999 JLeuco Biol 65:658 P D'Alessio, M Moutet, E Coudrier, S Darquenne, J Chaudiere 1998 Free Radical Biol Med 24:979 M Moutet, P D'Alessio, P Malette, V Devaux, J Chaudiere 1998 FreeRadical Biol Med 25:270 G Powis, JR Gasdaska, A Baker 1997 Adv Pharmacol 38:329 RC McKenzie, TS Rafferty, GJ Beckett, JR Arthur 2001 in Selenium: Its molecular biology and role in human health DL Hatfield (ed) Kluwer Academic Publishers Chapter 21:257 TS Rafferty, C Walker, JA Hunter, GJ Beckett, RC McKenzie 2002 Br J Dermatol 146:485 R Tolando, A Jovanovic, R Brigelius-Flohe, F Ursini, M Maiorino 2000 Free Radical Biol Med 2^:919 VJ Johnson, M Tsunoda, RP Sharma 2000 Arch Environ Contamination Toxicol 39:243 MS Stewart, JE Spallholz, KH Neldner, BC Pence 1999 Free Radical Biol Med 26:42 GN Schrauzer 2000 JNutr 130:1653 JE Spallholz 1994 Free Radical Biol Med 17:45 V Mostert, I Dreher, J Kohrle, J Abel 1999 FEBS Letts 460:23 S Bauersachs, M Kirchgessner, BR Paulicks 1993 J Trace Elements Electrolytes Health

&Disl:Ul 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57.

JM Finch, RJ Turner 1996 Res Vet Sci 60:97 SM Hegazy, Y Adachi 2000 Poultry Sci 79:331 BK Swain, TS Johri, S Majumdar 2000 Brit Poultry Sci 41:287 N Giadinis, G Koptopoulos, N Roubles, V Siarkou, A Papasteriades 2000 Comp Immunol Microbiol Infec Dis 23:129 R Boyne, JR Arthur, AB Wilson 1986 J Comp Pathol 96:379 N Ndiweni, JM Finch 1996 Vet Immunol Immunopathol 51:67 CD Davis, L Brooks, C Calisi, BJ Bennett, DM McElroy 1998 J Parasitol 84:1274 JA Rooke, JJ Robinson, JR Arthur 2004 J Agric Sci 142:253 A Smith, KB Madden, KJ Au Yeung, A Zhao, J Elfrey, F Finkelman, O Levander, T Shea-Donohue, JF Urban 2005 JNutr 135:830 MA Beck, J Handy, OA Levander 2004 Trends in Microbiol 12:417 M Roy, L Kiremidjianschumacher, Hi Wishe, MW Cohen, G Stotzky 1992 Proc Soc Exp Biol Med 200:26 M Roy, L Kiremidjianschumacher, Hi Wishe, MW Cohen, G Stotzky 1994 Biol Trace Element Res 41:103 L Kiremidjianschumacher, M Roy, HI Wishe, MW Cohen, G Stotzky 1996 Biol Trace Element Res 4l:\\5 L Kiremidjianschumacher, M Roy, HI Wishe, MW Cohen, G Stotzky 1996 Biol Trace Element Res 52:227 LD KoUer, JH Exon, PA Talcott, CA Osboume, GM Heningsen 1986 Clin Exp Immunol 63:570 MP Nair, SA Schwartz 1990 Immunopharmacol 19:177 MJ Guimaraes, D Peterson, A Vicari et al 1996 Proc Natl Acad Sci USA 93:15986 M Bonomini, S Forster, F Derisio et al 1995 Nephrol Dialysis Transpl 10:1654 F Girodon, P Galan, AL Monget et al. 1999 Arch Internal Med 159:784 U Tarp 1995 Analyst 120:877 P Knekt, M Heliovaara, K Aho, G Alfthan, J Mamiemi, A Aromaa 2000 Epidemiology 11:402 JM Reimund, C Hirth, C Koehl, R Baumann, B Duclos 2000 Clinical Nutr 19:43 LS Greene 1995 J Am Coll Nutr 14:317

322

58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82.

Selenium: Its molecular biology and role in human health

L Hasslemark, R Malgren, U Zetterstorm 1993 Allergy 48:30 R Shaw, K Woodman, J Crane et al. \994 N Zealand J Med 107:387 NLA Misso, DJ Peroni, DN Watkins, GA Stewart, PJ Thompson 1998 J Leuk Biol 63:124 R Gartner, M Angstwurm 1999 Medizinische Klinik 94:54 CS Broome, F McArdle, JAM Kyle, F Andrews, NM Lowe, CA Hart, JR Arthur, MJ Jackson 2004 Am J Clin Nutr 80:154 WC Hawkes, DS Kelley, PC Taylor 2001 Biol Trace Element /Jes 81:189 S Villette, JAM Kyle, KM Brown, K Pickard, JS Milne, F Nicol, JR Arthur, JE Hesketh 2002 Blood Cells Molecules & Diseases 29:174 N Safir, A Wendel, R Saile, L Chabraoui 2003 Clin Chem Lab Med 41:1005 RC McKenzie 2000 Clin Exp Dermatol 25:1 TS Rafferty, M Norval, A El-Ghorr, GJ Beckett, JR Arthur, F Nicol, JAA Hunter, RC McKenzie 2003 Biol Trace Elem Res 92:161 T Finkel, NJ Holbrook 2000 Nature 408:239 BM Lesourd 1997 Medizinische Klinik 66:S478 JW Albright, JF Albright 1998 Exp Gerontol 33:13 M Roy, L Kiremidjianschumacher, Hi Wishe, MW Cohen, G Stotzky 1995 Proc Soc Exp Biol Med 209:369 MT Ventura, E Serlenga, C Tortorella, S Antonaci 1994 Cytobios 11:115 Olivieri, D Girelli, M Azzini et al 1995 Clinical Sci 89:637 Olivieri, D Girelli, AM Stanzial, L Rossi, A Bassi, R Corrocher 1996 Biol Trace Element Res 51:31 HD Foster, LP Zhang 1995 Sci of Total Environ 170:133 HD Foster 1997 Medical Hypoth 48:355 G Ravaglia, P Forti, F Maioli et al 2000 Am J Clin Nutr 71:590 P Brenneisen, K Briviba, M Wlashek, J Wenk, K ScharfetterKochanek 1997 Free Radical Biol Med 22:515 Y Ito, O Kajkenova, RJ Feuers et al 199?, J Gerontol Series A-Biol Sci Med Sci 53:M169 V Serra, T Grune, N Sitte, G Saretski, T Von Zglinicki 2000 Ann NY Acad Sci 908:327 J Legault, C Carrier, P Petrov et al 2000 Biochem Biophys Res Com 272:416 S Watabe, Y Makino, K Ogawa et al 1999 Eur J Biochem 264:74

Chapter 28. Selenium and male reproduction Matilde Maiorino, Antonella Roveri, Fulvio Ursini Department of Biological Chemistry, University of Padova, Viale G. Colombo, 3, 1-35121 Padova, Italy

Regina Brigelius-Flohe Department Biochemistry of Micronutrients, Gennan Institute of Human Nutrition PotsdamRehbruecke (DIfE), Arthur-Scheunert-Allee 114-116, D-14558 Nuthetal, Germany

Leopold Flohe MOLISA GmbH, Universitatsplatz 2, D-39106 Magdeburg, Germany

Summary: Selenium deficiency has long been documented to result in impaired male fertility of rats, mice and boars. The prominent feature of selenium-deficient spermatozoa is a distorted architecture of the mid piece, where normally the mitochondria are embedded into a keratinous matrix called the mitochondrial capsule. This material, which contains most of the selenium of sperm, is composed of oxidatively cross-linked proteins, a major component being the selenoprotein phospholipid hydroperoxide glutathione peroxidase (PHGPx). PHGPx is abundantly synthesized in round spermatids under indirect control of testosterone. In late phase of spermatogenesis, the active soluble peroxidase is transformed into an enzymatically inactive structural protein by an oxidative process that is not understood in detail. Likely, it involves oligomerization of PHGPx itself, cross-linking of PHGPx with the sperm mitochondrion-associated cysteine-rich protein (SMCP) and other cysteine-rich proteins and selenadisulfide reshuffling with or without the aid of thioredoxin-glutathione reductase. Introduction The potential relevance of selenium to the reproductive system in livestock, laboratory animals and humans has been considered for at least five decades [1]. Impaired reproductive abilities due to selenium deficiency were reported for both sexes. In cows, cystic ovarian disease [2] and retained placenta [3-5] appear to respond to selenium supplementation; infertility of ewes may be associated with selenium deficiency [6,7]; and a selenium-deficient diet

324

Selenium: Its molecular biology and role in human health

resulted in reduced egg production and embryonic survival in hens that could be normalized by selenium supplementation [8]. The biochemical basis of these disturbances in female reproductive ability remains elusive. In contrast, a molecular basis for the impaired spermatogenesis, as was first reported for rats [9-12], mice [13,14] and boars [15], is emerging. Function and morphology of sperm in selenium deficiency Data on the influence of selenium on human reproduction are scarce and contradictory (compiled in [1,16]). Thus, the role of selenium in human fertility must be largely inferred from studies in laboratory animals. Interestingly, testes have the ability to accumulate selenium and to retain this trace element even during substantial selenium deficiency [17,18]. Specific alterations of sperm were first seen in rats depleted of selenium for two generations [9,10]. In mice, these alterations increased through successive generations of selenium deprivation [13,14]. Therefore, impairment of male fertility carmot reasonably be expected to result fi-om transient variations of selenium supply. In severe and prolonged selenium deficiency, male rats and mice become sterile as spermatogenesis is arrested. The seminiferous epithelium is degenerated and the lumen of the testicular tubules has the appearance of being more or less devoid of sperm [13, 19]. Clearly, this kind of azoospermia or aspermia mimics a block in cell division. The functional and morphological alterations of spermatozoa, observed in less severe selenium deprivation, are more discrete. In rats, the prominent feature is reduced sperm motility leading to impaired fertilization capacity [11]. Sperm motility is less affected by selenium deprivation in mice [13]. In both species, abnormal sperm morphology is observed [11,12,14]. Characteristically, the mid-piece of the spermatozoon, that harbours the helix of mitochondria embedded in a keratin-like matrix, appears structurally disturbed, fuzzy or broken. The sperm tail, consequently, appears distorted, and isolated sperm heads and tails are often seen. A particularly weak point in the seleniumdeficient rat sperm mid-piece appears to be an impaired fusion of the annulus with the mitochondrial sheath, which leads to flagellar disorganization and disruption [20]. Interestingly, the mid-piece is precisely the part of the spermatozoon where Brown and Burk [17] found most of the selenium accumulated when ^^Se was injected into rats. Within the midpiece, selenium proved to be primarily associated with a cysteine- and proline-rich protein in rodents [21,22] and bulls [23]. Cloning of this 'mitochondrial capsule selenoprotein (MCS)', howev er, revealed that it was not a selenoprotein [24,25]. It therefore was re-named "sperm mitochondrion-associated cysteine-rich protein (SMCP) [24]. After this investigation, the search for the real selenoprotein(s) in sperm became revitalized.

Selenium and male reproduction

325

Selenoproteins in the male genital system Pulse-labeling experiments with ^^Se in selenium-deprived rats show specific selenium incorporation into a variety of testicular and epididymal proteins. They cover a wide range of apparent molecular masses when separated on SDS-polyacrylamide gels [26,27]. Some of bands on gels could be tentatively identified as their molecular weights correspond to those of cytosolic glutathione peroxidase (cGPx), phospholipid hydroperoxide GPx (PHGPx), mitochondrial and cytosolic thioredoxin reductases [28], and selenoprotein P (SelP). A band with an apparent MW of 34 KDa, which was only seen in testis [27] and contained in sperm nuclei [29], could be attributed to a PHGPx variant (snGPx or nPHGPx) with a chemically distinct N-terminus [30] that results from the use of an alternative transcription start, representing a nuclear targeting sequence [30], and it is driven by an alternative promoter within the first intron in gpx-4 [31]. More recently, a thioredoxin reductase variant with a fused glutaredoxin sequence (TGR), that displays glutathione, thioredoxin reductase and protein isomerase activities, was shown to be particularly abundant in elongating spermatids at the site of the mitochondrial sheath formation [32]. A band corresponding to a 15 KDa selenoprotein is detected in prostate epithelium [29]. Selenoprotein P is a secreted extracellular protein with multiple selenocysteine residues. It was surprisingly found to be expressed in Leydig cells of mice by means of/« situ hybridization [33]. The promoter region of the SelP gene contains putative SRY sites that are presumed to bind the sexdetermining region product of the Y chromosome [33]. The specific role of SelP in testis is unknown. SelP (-/-) mice, however, have a low testicular selenium content that results in infertility due to impaired sperm motility [34,35]. The observations point to a role of testicular SelP in assuring selenium supply to the seminiferous epithelium. Glutathione peroxidase activities have been repeatedly measured in testis and epididymis. Yet most of the early investigations did not differentiate between the different types of GPx (reviewed in [36]) and thus have been practically useless in elucidating a role of selenium in fertility. The data may be compromised by summing up GPx activities of the selenoperoxidases, of the androgen-responsive cysteine homologs of extracellular GPx [37] and even of GSH-S-transferases [38]. The presence of cytosolic GPx in testis was shown by cGPx activity measurement after separation from PHGPx and by in situ hybridization [39,40]. cGPx has been implicated in antioxidant defence in Leydig cells that are presumed to produce H2O2 during steroid hormone synthesis [41]. There appears to be a general feeling that the seminiferous and mature sperm also require a particularly efficient protection against oxidative stress [42-44]. cGPx, as the selenoperoxidase most efficient in H2O2 reduction, would indeed be the enzyme of choice to meet this demand [36]. There is, however,

326

Selenium: Its molecular biology and role in human health

no indication of any predominance of cGPx in particular sites of the genital system. In situ hybridization studies display an uncharacteristic low level of cGPx mRNA in rat testis [39] and cGPx activities are accordingly low [39,45]. Finally, a specific role of cGPx in male reproduction can be ruled out as cGPx knock-out mice develop and reproduce normally [46]. In contrast, PHGPx is abundantly present in rat testis [43,47-49], but only after puberty [47,48]. The peripuberal increase in testicular PHGPx can be prevented by hypophysectomy and restored by application of chorion gonadotropin [47] indicating a hormonal control of PHGPx gene expression. Attempts to verify the hormonal control of PHGPx gene transcription by means of reporter gene constructs, however, have been unsuccessful [39]. In addition, testosterone or forskolin did not directly activate transcription and inhibition by 17-(3-estradiol could not be detected in hormone-responsive T47D and MCF7 cells [39]. Testosterone and forskolin also did not enhance PHGPx activity when added to decapsulated testes [39]. An explanation for these seemingly contradictory results was provided by in situ hybridization: PHGPx mRNA was seen to be predominantly expressed in a cell layer of the seminiferous epithelium representing the round spermatids [39]. Later studies with isolated spermatogenic cells confirmed the preferential expression in round spermatids [50]. The thickness of the spermatid layer reflecting proliferation of the germ epithelium is controlled by testosterone that is provided by gonadotropin-stimulated Leydig cells. When the Leydig cells are selectively destroyed, for example, by ethane dimethane sulfonate, the spermatid layer shrinks with some delay, and the PHGPx content of whole testis decreases in parallel to the disappearance of spermatids [39]. The hormonal control of PHGPx in testis, thus, is an indirect one. The PHGPx gene is not regulated by hormone action, but the transcribing cell type depends on testosterone. The burst of PHGPx gene transcription is reflected by a high content of PHGPx protein detected by immune histochemistry [47] and high PHGPx activity measured with the specific substrate phosphatidylcholine hydroperoxide [39,47,51,52]. PHGPx mRNA declines with elongation of spermatids and it is no longer detectable in spermatozoa. Immunostained PHGPx declines similarly, but remains faintly visible in spermatozoa. In contrast, PHGPx activity becomes almost undetectable in mature epididymal spermatozoa [51]. The PHGPx protein, however, is still present in spermatozoa, but as an enzymatically inactive, densely packed material contributing to constitute a considerable portion of the mitochondrial capsule of spermatozoa [51, 52]. The PHGPx protein can be solubilized out of this keratinous material by strong reduction and chaotropic agents and detected by MALDI-TOF mass spectrometry or Westem blotting [51]. Prolonged preincubation with 0.1 M DTT or mercaptoethanol even leads to the recovery of enzymatic activity [51].

Selenium and male reproduction

327

The nuclear form of PHGPx was originally considered to be a spermspecific protein ('sperm nuclei GPx', snGPx) but its messenger was later shown to be present in various somatic cells [31,53]. It was believed to be essential for chromatin condensation and thus pivotal for spermatogenesis [30]. Targeted deletion of nPHGPx, in fact, led to a defective chromatin condensation and head instability. Surprisingly, however, these defects were limited to spermatozoa collected from the epididymis, and nPHGPx -/- mice remained fully fertile [54]. A pivotal function of nPHGPx in male reproduction can therefore be ruled out. Since cytosolic PHGPx is also detectable in spermatid nuclei [55], the latter may compensate for the nPHGPx missing in the knockout mice. Impact of PHGPx moonlighting on sperm maturation The puzzling switch of PHGPx from an active peroxidase in spermatogenic cells to an enzymatically inactive protein in spermatozoa raises the question of what this new example of 'moonlighting' [56-58] might mean in the context of male fertility. The chemical process leading to the transformation of the soluble active peroxidase to a structural protein, although unknown in detail, is an oxidative one. The keratinous capsule material resists solubilizers like guanidine or sodium dodecyl sulfate, unless disulfides are reduced. Only upon the reductive treatment, monomeric PHGPx can be recovered. Inversely, when total sperm proteins are reductively solubilized and exposed to H2O2 in the absence of low molecular weight thiols, they readily form high molecular weight aggregates containing PHGPx. Purified PHGPx is also polymerized by H2O2 in the absence of GSH [51]. However, the product thus formed is a linear oligomer of PHGPx molecules having the active site selenium reacted with cysteine 148 on the back side of another PHGPx molecule [59]. These PHGPx oligomers easily dissolve with physiological concentrations of GSH. Taken together, the observations indicate that PHGPx, when oxidized by hydroperoxides in the absence of GSH, not only reacts with its exposed thiols but also with other protein thiols, and thereby becomes cross-linked probably via Se-S bonds. In enzymological terms, the proposed reaction simply means that the selenolate function of the ground state PHGPx is oxidised by ROOH to a first intermediate, which is a selenenic acid derivative (E-SeOH) (see Chapter 15]. The first intermediate then reacts with protein -SH instead of GSH to form the second intermediate (E-Se-S-Prot). Regeneration of the ground state enzyme (E-Se) that is commonly carried out by a second GSH, proceeds only slowly with this intermediate. The inactive, insoluble oxidation products of PHGPx might be thus considered as alternate substrate dead-end intermediates. Such reactions likely occur in late spermatogenesis. The transition from round to elongated spermatids is paralleled by a decrease of GSH and protein thiols [42,60-62].

328

Selenium: Its molecular biology and role in human health

Evidently, the loss of GSH in this phase of spermatogenesis forces PHGPx into the alternate substrate pathway. Beyond these basics, little else can be said with certainty about the transformation process. The mechanism leading to the pivotal disappearance of GSH in late spermatogenesis is obscure. Interestingly, not only GSH, but also GSSG and mixed disulfides derived from GSH become undetectable. This could result from increased GSH metabolism or, more likely, from GSH oxidation followed by the release of GSSG [63,64] and subsequent extracellular degradation by y-glutamyl transpeptidase, which is absent in spermatogenic cells, but abundant in adjacent testicular and epididymal tissue [65]. Again, the source of the required oxidation equivalents that becomes up regulated at the specific point of sperm differentiation remains elusive. Some of the reaction partners that are required to build up the mitochondrial capsule have been recently identified in the mitochondrial capsule by HPLC-ESI-MS/MS: SMCP fragments, voltage dependent anion channel (VDAC2) and three types of keratins (complex I, acidic, kbl type II, and k5) [66]. Among those, SMCP with its 30% cysteine residues is the most likely candidate to react with PHGPx. This assumption is corroborated by strict co-localization of SMCP and PHGPx in the mitochondrial midpiece [67]. Moreover, peptides with adjacent cysteine motifs, which are abundant in SMCP, proved to be excellent substrates of PHGPx [66]. The cys-cys motifs of SMCP classifies this protein as a "high sulfur keratin-associated protein' ' (KAP), thus offering the intriguing possibility to bridge PHGPx/SMCP copolymers with keratins in the mid piece architecture. VDAC-2 finally has been shown to be linked to the outer dense fibers and to the mitochondria in bovine sperm [68]. The interplay of these capsule components is far from being clear, but a plausible sequence of reactions might be: i) oxidation of cys-cys motifs in SMCP by PHGPx; ii) reshuffling of disulfide and (selena-) disulfide bonds with or without the aid of TGR, which can act as a protein disulfide isomerase [32]; and iii) fixation of crosslinked proteins to the fibrous sheath and the mitochondrial surfece via SMCP moieties and/or VDAC2. Attempts to corroborate the essentiality of the capsule components by inverse genetics yielded ambiguous results. Knock-out of PHGPx resulted in prenatal fatality [69,70]. Interim data obtained along this line, however, were revealing [71]. Nine male chimeric mice with genotype +/- having more than 50% PHGPx reached maturity. They were fertile, but among 190 offspring not a single homozygous or hemizygous mouse deficient in PHGPx could be identified. This observation suggested that not even hemizygous cells could contribute to the male germ line. The chimeric mice did not display any obvious phenotype apart from mosaic-like disturbance of the testes: While parts of the testis looked normal, others, obviously derived from hemizygous cells, were markedly altered. In the tubules of affected areas, only few

Selenium and male reproduction

329

morphologically disturbed spermatozoa were detectable. The alterations, e.g, distorted tails, fuzzy and broken mid pieces, and isolated heads strongly mimic pathologies seen in selenium deficiency. Thus far, the results comply with our presumption that PHGPx is indispensable for the integrity of the mid piece architecture. Another observation, however, suggests that PHGPx must indeed have a dual function in spermatogenesis. In the affected testicular tissue of chimeric mice, the seminiferous epithelium looked degenerated and severely disorganized. These results could, however, not be reproduced in another strain of mice in which hemizygous PHGPx +/- mice remained fertile (M. Conrad, personal communication). Similarly, targeted disruption of SMCP resulted in infertility and asthenozoospermia in some mice, but only in marginal disturbance of spermatogenesis in others [72]. The sensitivity of the SMCP knock-out to the genetic background clearly points to the presence of a back up system that may or may not substitute for its cross-linking function. Conclusions and outlook Of the known selenoproteins, cGPx appears dispensible to fertility. One or the other thioredoxin reductase, because of the link of thioredoxin to nucleic acid metabolism, may be indispensable to sustain proliferation of the seminiferous epithelium. Thioredoxin reductase deficiency may therefore account for the complete arrest of spermatogenesis observed in rodents deprived of selenium for several generations. PHGPx proved to be pivotal for rodent spermatogenesis. In early spermatogenic cells, it is present as an active peroxidase and may regulate proliferation and/or differentiation. During maturation, it builds up the capsular architecture by oxidizing SMCP adjacent cysteine residues to cystine that finally undergo reshuffling and possibly proofreading by TGR, which also is a selenoprotein [32]. In mature spermatozoa, PHGPx represents a structural component of the mitochondrial capsule. The relevance of other selenoproteins to male fertility remains to be established. The dual role of PHGPx during sperm maturation appears to apply to other mammals including man, but not to non-mammalian vertebrates or metazoa [73]. The clinical relevance of sperm PHGPx content to fertility is supported by preliminary studies on subjects with fertility problems. The PHGPx activity of the sperm samples, as measured after reactivation, correlated positively with functional parameters indicative of fertiUty [74,75]. PHGPx genes of infertile subjects showed a trend towards a higher content of single nucleotide polymorphisms (SNPs) than in the controls, but most of the SNPs were not associated with infertility [76]. Some multiple mutations of gpx-4 were only observed, however, in infertiles. Taken together, mutations in gpx-4 can at best be considered a rare cause of infertility. Large scale and well designed studies will be required to determine how fi-equently PHGPx deficiency accounts for impaired fertility

330

Selenium: Its molecular biology and role in human health

and whether such deficiencies are due to insufficient selenium supply, defect of PHGPx, defects in the PHGPx gene, or an altered regulation of PHGPx biosynthesis. References 1.

2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38.

Subcommittee on Selenium, Committee on Animal Nutrition, Board on Agricolture, National Research Council 1983 Selenium in nutrition. National Academy Press, Washington, DC JA Harrison, DD Hancock, HR Conrad 1984 J Dietary Sci 61:12,1 N Trinder, CD Woodhouse, C Renton 1969 Vet Rec 85:550 EC Segerson, GJ Riviere, HL Dalton, MD Whitacre 1981 J Dairy Sci 64:1833 WR Julien, HR Conrad, JE Jones, AL Moxon 1976 J Dairy Sci 64:1833 ED Andrews, WJ Hartley, AB Grant 1968 AT Z Vet J16:3 WJ Hartley, AB Grant 1961 Fed Proc 20:679 AH Cantor, ML Scott 1974 Poult Sci 53:1670 KE McCoy, PH Weswig 1969 JNutr 98:383 ASH Wu, JE Oldfield, OH Muth, PD Whanger, et al 1969 Proc West Soc Am Soc An Sci 20:85 SH Wu, JE Oldfield, PD Whanger, PH Weswig 1973 BiolReprod 8:625 AS Wu, JE Oldfield, LR Shull, PR Cheeke 1979 Biol Reprod 20:793 E Wallace, HI Calvin, GW Cooper 1983 Gamete Research 4:377 E Wallace, GW Cooper, HI Calvin 1983 Gamete Research 4:389 CH Liu, YM Chen, JZ Zhang, MY Huang, et al 1982 Acta Vet Zootech Sinica 13:73 GN Schrauzer 1998 Selen: neue Entwicklungen aus Biologic, Biochemie und Medizin, Johann Ambrosius Barth Verlag, Huethig GmbH, Heidelberg, Leipzig DG Brown, RF Burk 1973 J Nutr 103:102 DBehne.THofer-Bosse 1984 y^«
Selenium and male reproduction 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76.

331

M Maiorino, JB Wissing, R Brigelius-Flohe, F Calabrese, et al 1998 FASEB 712:1359 A Zini, PN Schlegel 1997 Fertil Steril 68:689 V Peltola, I Huhtaniemi, T Metsa-Ketela, M Ahotupa 1996 Endocrinology 137:105 F Bauche, MH Fouchard, B Jegou 1994 FEBSLett 349:392 F Tramer, F Rocco, F Micali, G Sandri, et al 1998 Biol Reprod 59:753 M Maiorino, F Ursini 2002 Biol Chem 383:591 L Flohe 1989 In: D Dophin, R Paulson and O Awamocic (Eds). Glutathione: Chemical, biochemical and medical aspects - Pari A, John Wiley & Sons, Inc., New York p 643 YS Ho, JL Magnenat, RT Bronson, J Cao, et al 1997 J Biol Chem 272:16644 A Roveri, A Casasco, M Maiorino, P Dalan et al 1992 J Biol Chem 161:6U1 A Giannattasio, M Girotti, K Williams, L Hall, et al 1997 J Endocrinol Invest 20:439 F Weitzel, F Ursini, A Wendel 1990 Biochim Biophys Acta 1036:88 K Mizuno, S Hirata, K Hoshi, A Shinohara, et al 2000 Biol Trace Elem Res 74:71 F Ursini, S Heim, M Kiess, M Maiorino, et al 1999 Science 285:1393 M Maiorino, L Flohe, A Roveri, P Steinert, et al 1999 Biofactors 10:251 A Borchert, NE Savaskan, H Kuhn 2003 J Biol Chem 278:2571 M Conrad, SG Moreno, F Sinowatz, F Ursini, et al 2005 Mol Cell Biol 25:7637 M Maiorino, P Mauri, A Roveri, L Benazzi, et al 2005 FEBSLett 579:667 J Piatigorski 1998 Ann New York Acad Sci 842:7 CJ Jeffery 1999 Trends Biochem Sci 24:8 P Tompa, C Szasz, L Buday 2005 Trends Biochem Sci 30:484 P Mauri, L Benazzi, L Floh6, M Maiorino, et al 2003 Biol Chem 384:575 R Shalgi, J Seligman, NS Kosower 1989 Biol Reprod 40:1037 J Seligman, NS Kosower, R Shalgi \991 Biol Reprod A()-30\ HM Fisher, RJ Aitken 1997 J Exp Zool 277:390 H Sies, TP Akerboom 1984 Methods Enzymol 105:445 SC Lu, WM Sun, J Yi, M Ookhtens, et al 1996 J Clin Invest 97:1488 J Seligman, GL Newton, RC Fahey, R Shalgi, et al 2005 JAndrol 26:629 M Maiorino, A Roveri, L Benazzi, V Bosello, et al 2005 J Biol Chem 280:38395 K Nayemia, M Diaconu, G AumuUer, G Wennemuth, et al 2004 Mol Reprod Dev 67:458 KD Hinsch, V De Pinto, VA Aires, X Schneider, et al 2004 J Biol Chem 279:15281 UK Laemmli 1970 Nature 227:680 LJ Yant, Q Ran, L Rao, H Van Remmen, et al 2003 Free Radic Biol Med 34:496 M Conrad, U Heinzelman, W Wurst, GW Bomkamm, et al 2000 7th International Symposium on Selenium in Biology and Medicine, Venice, Italy, October 1 -5 K Nayemia, IM Adham, E Burkhardt-Gottges, J Neesen, et al 2002 Mol Cell Biol 22:3046 M Maiorino, A Roveri, L Flohe, F Ursini 2000 7th International Symposium on Selenium in Biology and Medicine, Venice, Italy, October 1 -5 C Foresta, L Flohe, A Garolla, A Roveri, et al 2002 Biol Reprod 67:967 H Imai, K Suzuki, K Ishizaka, S Ichinose, et al 2001 Biol Reprod 64:674 M Maiorino, V Bosello, F Ursini, C Foresta, et al 2003 Biol Reprod 68:1134

Chapter 29. Mouse models for assessing the role of selenoproteins in health and development Bradley A. Carlson, Xue-Ming Xu, Rajeev Shrimali, Aniruddha Sengupta, Min-Hyuk Yoo, Nianxin Zhong and Dolph L. Hatfield Molecular Biology of Selenium Section, Laboratory of Cancer Prevention, Center for Cancer Research, National Cancer Institute, National Institutes of Health, Bethesda, MD 20892, USA

Robert Irons' and Cindy D. Davis Nutritional Science Research Group, Division of Cancer Prevention, National Cancer Institute, National Institutes of Health, Bethesda, MD 20892, USA

Byeong Jae Lee Laboratory of Molecular Genetics, Institute of Molecular Biology and Genetics, School of Biological Sciences, Seoul National University, Seoul 151-742, Korea

Sergey V. Novoselov and Vadim N. Gladyshev Department of Biochemistry, University of Nebraska, Lincoln, NE 68688, USA

Summary: Mouse models have been generated to assess the roles of selenoproteins involved with housekeeping tasks and/or stress-related phenomena in development and health. Each mouse model has taken advantage of the fact that the synthesis of all selenoproteins is dependent on the expression of two selenocysteine (Sec) tRNA^^*"^^ ^'^ isoforms that differ from each other by a single methyl group on the ribosyl moiety at position 34. The endogenous Sec tRNA^^^*^'^^" population was selectively altered by generating mouse models involving 1) transgenic animals carrying mutant or wild type Sec tRNA'^'^^^'^ transgenes, 2) conditional knockout animals carrying a floxed Sec tRNA'^'^^^' gene that was targeted for removal in specific tissues and organs using loxP-Cre technology and 3) transgenic/standard knockout animals carrying mutant or wild type transgenes and a knockout of the Sec tRNA'^''^" gene wherein the animal's survival is dependent on the transgene. These mouse models perturbed selenoprotein expression, often in a protein- and tissue-specific maimer, permitting us to better assess their function in health and development. 'RI is also affiliated with the Molecular Biology of Selenium Section, Laboratory of Cancer Prevention, CCR, NCI, NIH, Bethesda, MD 20892.

334

Selenium: Its molecular biology and role in human health

Introduction Numerous human clinical trials examining the role of selenium in health have been initiated recently, or have already been completed. Perhaps the most famous of those studies that have been completed is that of Clark et al [1]. These investigators reported in 1996 that supplementing the diet with 200 micrograms of selenium/day reduced the incidence of prostate and colon cancers by more than 50% and lung cancer by about 35%. This study provided the basis for numerous subsequent human clinical trials involving selenium. For example, the largest human clinical trial ever undertaken, the Selenium and Vitamin E Cancer Prevention Trial (SELECT [2]), involves more than 35,000 males in determining whether selenium and vitamin E may play a role in protecting men against prostate cancer. Another human clinical trial recently undertaken involves 1960 individuals in determining whether selenium may have a role in preventing the occurrence of secondary lung tumors in patients with a previous surgically removed non-small cell lung lesion [3]. These trials are being carried out at the cost of hundreds of millions of dollars with little or no understanding of how selenium acts metabolically to bring about potential health benefits. It is important therefore to generate mouse models to determine how selenium acts at the molecular level in promoting better health and to determine whether small molecular weight selenocompounds and/or selenoproteins are the responsible agents. Selenoproteins are dependent on Sec tRNA^^*'^'^*'^ for their expression as discussed in detail in Chapter 3 (see also [4,5]). In addition, the Sec tRNA^^^'^^^*'' population in mammalian cells is composed of two major isoforms that differ from each other by a 2'-C>-methylribosyl moiety at position 34 which is designated Um34 (Chapter 3 and [4,5]). The two isoforms, designated mcm^U and mcm^Um, have different roles in selenoprotein synthesis wherein mcm^U appears to be largely responsible for the expression of housekeeping selenoproteins and mcm'Um largely responsible for selenoproteins involved in stress-related phenomena [6,7]. By perturbing the expression of Sec tRNA^^^"^'^" in mice, selenoprotein synthesis can be altered in such a way that their roles in development and health can be readily studied. In fact. Sec tRNA'^"^^'" synthesis can be modulated in various ways in generating different mouse lines so that the roles of selenoproteins involved in housekeeping tasks and in stress-related phenomena can be elucidated. We have generated several mouse models for examining the role of selenoproteins in development and health. These mouse models involve generating transgenic, standard knockout, conditional knockout or a combination of transgenic/knockout mice encoding either wild type or a mutant Sec tRNA'^''^*^ gene or transgene. It should also be noted that the characteristics and properties of a number of individual selenoproteins have been examined, including being targeted for removal, to

Mouse models in health and development

335

assess their roles in health and development (see the Chapters in Part II. Selenium-containing proteins). Mouse models for elucidating tlie role of selenoproteins in development and health To generate mouse models for elucidating selenoprotein function, we took advantage of the fact that selenoprotein synthesis is dependent on the presence of Sec tRNA^^*^'^'. Our mouse models involve the introduction of a transgene, designated trsp', or manipulation of the Sec tRNA^^^*^'^" gene, designated trsp and they are summarized as follows: 1) transgenic mice carrying wild type or a mutant trsp' [8]; 2) conditional knockout mice carrying a floxed trsp that can be specifically targeted for removal using loxP-Cre technology [9]; and 3) standard knockout/transgenic mice containing trsp knockout, designated Atrsp, and carrying wild type or mutant trsp transgenes in which the survival of the animal is dependent on the transgene [6,7]. A fourth mouse model is also briefly discussed wherein transgenic/conditional knockout mice carrying two transgenes, a Cre transgene under control of a promoter targeted for a specific organ or tissue, and a wild type or mutant trsp' in addition to the targeted (floxed) trsp (B.A. Carlson, M.E. Moustafa, R. Shrimali, M. Rao, N. Zhong, S. Wang, L. Feigenbaum, B.J. Lee, V.N. Gladyshev and D.L. Hatfield, submitted). The mouse models are shown in Table 1. They were prepared for assessing the roles of selenoproteins in health and development as well as defining the roles of the two Sec tRNA^^^'^^^'^' isoforms, mcm^U and mcm^Um, in selenoprotein synthesis. Since the role of mcm^U, which is largely involved in the expression of housekeeping selenoproteins, and of mcm'Um, which is largely involved in the expression of selenoproteins involved in stress-related phenomena, has been discussed in Chapter 3 and detailed elsewhere [5-7], the emphasis of the present chapter is on generating mouse models for examining the roles of selenoproteins in health and development. It should be noted that some of our mouse models may also be used to elucidate the role of small molecular weight selenocompounds in disease prevention. Transgenic mouse models Transgenic mice are generated by introducing one or more copies of the wild type or mutant trsp into the mouse genome. The resulting mice are then bred to obtain a stable breeding population (i.e., mice that are homozygous for the transgene). Mutations were made at either position 37 [8] or at postion 34 in Sec tRNA^^''^^" for preparing mutant transgenes. The base at position 37 in the fully maturated tRNA'^'^^^*^ is isopentenyladenosine (i*A), while that at position 34 is mcm^U (see Chapter 3 or [4,5]). The base at position 37 was changed to G (A37->G37) and that at position 34 to A (T34->A34). Importantly, neither of these two, very different, mutant tRNAs contain

336

Selenium: Its molecular biology and role in human health

Um34, the methyl group located at the 2'-hydroxylribose on mcm^U [10]. An A in the wobble position of the anticodon in any tRNA is normally converted to I. About 65% of the A34 mutant Sec tRNA'^''^^" is converted to 134 in mouse liver (B.A. Carlson, M.E. Moustafa, R. Shrimali, M. Rao, N. Zhong, S. Wang, L. Feigenbaum, B.J. Lee, V.N. Gladyshev and D.L. Hatfield, submitted), and interestingly, the ACA and ICA anticodons in A34 Sec tRNA'^''^^'' decode UGU and UGU, UGC (UGU and UGC are codons for cysteine) and UGA (UGA is the codon for Sec), respectively. Table 1. Mouse models and their uses. Type

Genotype"

Uses*

Transgenic

trsp'

Transgenic

G37trsp'

Transgenic Standard KO" Conditional KO'*

A34trsp' Atrsp Atrsp'

Transgenic/ standard KO'' Transgenic/ conditional KO'' (liver)

trsp'/Atrsp

Determine if Sec tRNA^^^'J'"'' is limiting [8] SV rescue in standard KO'' [6,7] SV replacement in targeted tissues and organs'' Rescue of SPs*^ in KC'mice Replacement of SPs in conditional KC'mice Roles in muscle adaptation [9] and cancer risk in prostate,^ mammary^ gland and colon^ SP'' replacement only SP" rescue with trsp' & trsp'G3>7 KO** trsp in various tissues and organs using promoters that are tissue and organ specific SP'^ rescue in KO"* mice [6,7]

G37 trsp'/Atrsp trsp'/Atrsp''

Partial SP'^ rescue in KO'' mice [6,7] SP'^ replacement in liver^

G31 trsp'/Atrsp/^

SP*^ replacement in liver^

"Genotype designations used for mouse models (see text). *Uses - the various uses of the mouse models with accompanying references (see also text). "SP - selenoprotein(s). ''KO - knockout. *AM Diamond, personal communication. -'R Irons, BA Carlson, DL Hatfield, C Davis, submitted. ^ A Carlson, ME Moustafa, R Shrimali, M Rao, N Zhong, S Wang, L Feigenbaum, BJ Lee, VN Gladyshev and DL Hatfield, submitted.

Mouse models in health and development

337

Initially, the resulting transgenic mice encoding multiple copies of trsp' and G37trsp' (see Table 1) were examined [8]. This study provided the first example of transgenic mice engineered to contain functional tRNA transgenes. The fact that over-expression of wild type Sec tRNA^^^'^'^^*' due to the extra copies of trsp' did not appear to influence selenoprotein synthesis in the organs and tissues examined suggested that the levels of the Sec tRNA^^'^^^'^' isoforms were not limiting in protein synthesis (reviewed in [4,5]). However, mice carrying G37trsp' had a pronounced effect on selenoprotein expression and the effect occurred in a protein- and tissuespecific manner [8]. The most and least affected selenoproteins were glutathione peroxidase 1 (GPxl) and thioredoxin reductase 1 (TRl), respectively, and the organs which manifested the most and least affect on selenoprotein synthesis were liver and testes, respectively. The mutant Sec tRNA'^''^'" product from G37trsp' lacked the i*A base modification and altered the levels of the two host Sec tRNA'^"^^^'' isoforms is such a manner that the Um34 species was reduced and the mcm^U species was enriched. As the amount of the mutant tRNA increased with increasing numbers of transgenes, the amount of the Um34 species and the amount of some selenoproteins, and in particular, GPxl also decreased. The correlation in reduction of the Um34 isoform and certain selenoproteins led us to propose that the Um34 modification is responsible for the expression of several selenoproteins that are involved in the lower echelon of selenoprotein hierarchy expression [5-7]. Interestingly, many of the selenoproteins that are expressed in the lower echelon of selenoprotein hierarchy and are sensitive to selenium status, such as GPxl, serve largely stress-related functions, while those that are expressed in the upper echelon of selenoprotein hierarchy and are less sensitive to selenium status, such as TRl, serve largely housekeeping functions. Those members associated with stress-related phenomena are the ones dependent on the Sec tRNA'^"'^^° Um34 modification for their expression (see also Chapter 3 and [5-7]). As shown in Table 1, the G37trsp' mice have been used in several different studies. The studies have shown that these mice, which are deficient in selenoproteins involved in stress-related phenomena, have an enhanced 1) skeletal muscle adaptation after synergist ablation and following exercise [9], 2) incidence of prostate malignancy when the mice also carry an oncogene directed to this tissue (see legend to Table 1), and 3) incidence of breast malignancy when the mice also carry an oncogene directed to this tissue (see legend to Table 1). In addition, we have observed that G'i7trsp' mice a significantly greater number of azoxymethane (AOM)-induced aberrant crypt formations (preneoplastic lesions in the colon) than wild type mice (R Irons, BA Carlson, DL Hatfield and CD Davis, submitted). Supplementing the diets of the AOM treated G37trsp' and wild type mice with 0.1 and 2.0 i^g/g selenium significantly reduced the incidence of preneoplastic lesions in both

338

Selenium: Its molecular biology and role in human health

mouse lines. These observations are the first to provide evidence that both selenoproteins and low molecular weight selenocompounds play a role in the cancer protective effects of selenium. In another study involving Sec tRNA^^^''^^'^ mutant transgenic mice, we examined the effect of A34trsp', which also lacked Um34, on selenoprotein synthesis (B.A. Carlson, M.E. Moustafa, R. Shrimali, M. Rao, N. Zhong, S. Wang, L. Feigenbaum, B.J. Lee, V.N. Gladyshev and D.L. Hatfield, submitted). The effects of A34trsp' on down regulating stress-related selenoprotein expression were similar to those of G31trsp'. Since both these mutants lack Um34, these observations provided further evidence that Um34 is responsible governing the expression of those selenoproteins involved in stress-related phenomena. Transgenic/knockout mouse models Since the standard knockout of trsp is embryonic lethal [9,11], it appeared that this mutant could not be used for further study of selenoprotein expression. However, we devised a means of rescuing selenoprotein expression by crossing heterozygous trsp knockout mice with homozygous trsp' transgenic mice and breeding the offspring to obtain a line of mice lacking trsp (Atrsp) that was dependent on the trsp' for survival [6,7]. One advantage of rescuing a knockout mouse with a wild type or mutant transgene is that the number of transgenes, and therefore, the levels of the corresponding gene product can be maintained at normal or elevated amounts depending on the transgene copy number. Rescuing with 20 copies of the wild type transgene enriched the Sec tRNA'^"^^''' population several fold, but little or no effect on selenoprotein expression in various tissues or organs was observed [6,7]. These observations provided further evidence that Sec tRNA^^"^^*^ is not limiting in selenoprotein biosynthesis (reviewed in [4,5]). Rescue of the Atrsp mice with the mutant transgene, G37trsp' afforded us with an opportunity of obtaining a mouse line with a mutant transgene wherein there is no background of host Sec tRNA'^'^'^^^*'' and selenoprotein expression is therefore totally dependent on the mutant tRNA. Gil trsp yielded a tRNA that lacked two base modifications, i*A37 and Um34 (see above, Chapter 3 and [6,10]) and mice rescued with G37trsp' lacked several selenoproteins including glutathione peroxidases 1 and 3, SelR and SelT [6,7]. Interestingly, we were not successful in rescuing selenoprotein synthesis in Atrsp mice with A34trsp' (see reference to Carlson et al above). As discussed in greater detail in Chapter 3, the novel regulation of several selenoproteins involved in stress-related functions occurs at the level of translation.

Mouse models in health and development

339

Conditional knockout mouse models Removal of trsp from the mouse genome is embryonic lethal [9,11] which prevented further study of the standard trsp knockout per se in selenoprotein expression. We therefore prepared the conditional knockout of trsp using loxP-Cre technology [9]. The removal of floxed trsp in mouse mammary epithelium [9] and in liver hepatocytes [12] was examined, trsp was targeted for removal in mammary epithelium using transgenic mice carrying the Cre recombinase gene under the control of the mouse mammary tumor virus long terminal repeat promoter or the whey acidic protein promoter. Neither Cre promoter was effective in complete removal of trsp in mammary epithelial cells, but the Sec tRNA^^^"^^^^*^ population was substantially reduced to alter selenoprotein expression in a protein specific manner [9]. In liver, however, the targeted removal of floxed trsp with transgenic mice carrying the Cre recombinase under the control of the albumin promoter was virtually complete [12]. Surprisingly, the mice survived without selenoprotein expression in hepatocytes which comprise about 85% of the liver cell mass. Selenoprotein P (SelP), which is the only known selenoprotein with multiple Sec residues (see Chapters 9, 10 and 21), would seem to be largely made in the liver and transported to other organs and tissues as its level was reduced about 75% in plasma of the selenoproteinless liver knockout mice. These mice appeared phenotypically normal until about 24 hours before death and death appeared to be due to severe hepatocellular degeneration and necrosis with concomitant necrosis of peritoneal and retroperitoneal fat [52]. Although most animals lacking selenoprotein expression in their liver died within two to three months in this initial study, these animals may be kept alive for extended periods of time on a diet enriched in other nutrients (U. Schweizer, L. Schomburg and J. Kohrle, personal communication). This is an important observation since these animals live much longer on a different diet and can be subjected to various environmental agents to study the role of selenoproteins in liver function and health. As selenium has been implicated in heart disease (see Chapter 25) and immune function (Chapter 27), we examined the role of selenoproteins in cardiovascular disease and the immune system. By targeting the removal of trsp in either endothelial cells or myocytes in skeletal and heart muscle, we have elucidated the role of selenoproteins in cardiovascular disease (R. Shrimali, J.A. Weaver, G.R. Miller, B.A. Carlson, S.V. Novoselov, E. Kumaraswamy, V.N. Gladyshev and D.L. Hatfield, submitted). Removal of selenoprotein expression in endothelial cells was embryonic lethal. 14.5-dayold embryos had numerous abnormalities including necrosis of the central nervous system, subcutaneous hemorrhage and erythrocyte immaturity. Loss of selenoprotein expression in myocytes, however, manifested no apparent phenotype until about day 12 after birth, when affected mice developed decreased mobility and an increased respiratory rate, followed by death

340

Selenium: Its molecular biology and role in human health

within a few hoxirs. Although there was no evidence of inflammation of the skeletal muscle in mice that lacked trsp in myocytes, they had moderate to severe myocarditis with inflammation extending into the mediastinum.. Targeted removal of trsp in endothelial cells demonstrated an essential role of selenoproteins in their development, while targeted removal of trsp in myocytes demonstrated an essential role of selenoproteins in proper function of cardiac muscle. These studies also showed a direct connection between the loss of selenoprotein expression and cardiovascular disease (R. Shrimali, J. A. Weaver, G.R. Miller, B.A. Carlson, S.V. Novoselov, E. Kumaraswamy, V.N. Gladyshev and D.L. Hatfield, submitted). The conditional knockout of trsp is an important mouse model as it provides a means of elucidating the roles of selenoproteins in health, development and/or function in tissues or organs for which there is a specific promoter. Promoters that function to express Cre early in development of a specific tissue or organ can be used to assess the role of selenoproteins in development of that tissue or organ. On the other hand, promoters that function to express Cre after the tissue or organ is developed can be used to assess the role of selenoproteins in proper function of that organ or tissue. Furthermore, promoters that function to express Cre either early or late in development of a specific tissue or organ, and provided the animal survives for a long period of time following trsp knockout, can be used to examine the animal's ability to handle various forms of stress (e.g., viral or bacterial infection, carcinogen(s), cancer driver gene(s), etc). Those animals that survive for long periods of time, while lacking selenoprotein expression in a specific tissue or organ, can also be used to assess the role of small molecular weight selenocompounds in health. Therefore, the conditional knockout of trsp in specific tissues and organs is an important tool for assessing the role of selenoproteins in health and development and this model can also be used to assess the role of small molecular weight selenocompounds in protecting the animal against environmental stresses. Transgenic/conditional knockout mouse models Although the rescue of mice encoding a A/rs/j mice with G37trsp' provides a novel model for studying the role of housekeeping and stress-related selenoproteins in health and the role of the two Sec tRNA'^"^'^*^ isoforms in governing selenoprotein expression [6,7], this mouse model permit us to focus on the animal as a whole and not on not on the roles of these components in individual organs and tissues. However, removal of selenoprotein expression by targeting floxed trsp with a specific promoter Cre and replacing selenoprotein expression with a mutant trsp' permits us to assess the role of housekeeping and stress-related selenoproteins, as well as the role of both isoforms, in health and/or proper function of specific organs or tissues. Selenoprotein removal was targeted in liver of the floxed trsp

Mouse models in health and development

341

mouse using the albumin Cre promoter and selenoprotein expression was replace with either G37trsp' or A34trsp' (B.A. Carlson, M.E. Moustafa, R. Shrimali, M. Rao, N. Zhong, S. Wang, L. Feigenbaum, B.J. Lee, V.N. Gladyshev and D.L. Hatfield, submitted). The pattern of replacing selenoprotein expression, wherein several selenoproteins associated with stress were not recovered, was similar with either the G37 and A34 mutant transgenes. The fact that the two mutant Sec tRNA^^'"^'^'^ isoforms govern selenoprotein expression in virtually an identical manner without the influence of host wild type Sec tRNA^^''^*'' demonstrates that Um34 is responsible for the synthesis of stress-related selenoproteins. Importantly, this mouse model will most certainly provide insights into one of the central questions in the selenium field which is "What are the contribution of selenoproteins versus low molecular weight selenocompounds in the cancer chemopreventive effects of selenium and other health benefits of this trace element?" Both these mutant transgenes replaced housekeeping selenoprotein expression, but not stress-related selenoprotein expression. Acknowledgements This work was supported by the Intramural Research Program of the National Institutes of Health, National Cancer Institute, Center for Cancer Research. References 1. 2. 3. 4. 5. 6. 7. 8.

9.

10.

11. 12. 13.

LC Clark etal 1996 y^M^ 276:1957 http://www.crab.org/select/ http://www.cancer.gov/clinicaltrials/ft-ECOG-5597 DL Hatfield, VN Gladyshev 2002 Mol Cell Biol 22:3565 DL Hatfield, BA Carlson, XM Xu, H Mix, VN Gladyshev 2006 Prog Nucl Acids Res Mol Biol (In press) BA Carlson, XM Xu, VN Gladyshev, DL Hatfield 2005 J Biol Chem 280:5542 BA Carlson, XM Xu, VN Gladyshev, DL Hatfield 2005 In Grosjean (Ed) Fine-Tuning of RNA Functions by Modification and Editing. Topics in Current Genetics Vol 12, p 431 ME Moustafa, BA Carlson, MA El-Saadani, GV Kryukov, QA Sun, JW Harney, KE Hill, GF Combs, L Feigenbaum, DB Mansur, RF Burk, MJ Berry, AM Diamond, BJ Lee, VN Gladyshev, DL Hatfield 2001 Mol Cell Biol 21:3840 TA Homberger, TJ McLoughlin, JK Leszczynske, DD Armstrong, RR Jameson, PE Bowen, ES Hwang, H Hou, ME Moustafa, BA Carlson, DL Hatfield, AM Diamond, KA Esser 2003 J Nutrition 133:3091 E Kumaraswamy, BA Carlson, F Morgan, K Miyoshi, GW Robinson, D Su, S Wang, E Southon, L TessaroUo, BJ Lee, VN Gladyshev, L Hennighausen, DL Hatfield 2003 Mol Cell Biol 23:1477 LK Kim, T Matsufuji, S Matsufuji, BA Carlson, SS Kim, DL Hatfield, BJ Lee. 2000 RNA 6:1306 MR Bosl, K Takaku, M Oshima, S Nishimura, MM Taketo 1997 Proc Natl Acad Sci USA 94:5531 BA Carlson, SV Novoselov, E Kumaraswamy, BJ Lee, MR Anver, VN Gladyshev, DL Hatfield 2004 J Biol Chem 279:8011

Chapter 30. Drosophila as a tool for studying selenium metabolism and role of selenoproteins Cristina Pallares, Florenci Serras and Montserrat Corominas Departament de Geneica, Universitat de Barcelona, Diagonal 645, 08028 Barcelona, Spain

Summary: The synthesis of selenoproteins is conserved in organisms ranging from bacteria to humans, including flies, and involves the differential decoding of the UGA stop codon as a selenocysteine (Sec). Drosophila genetics provides an excellent tool with which to investigate the synthesis machinery as well as the functions of specific selenoproteins. Mutations in some components of the selenoprotein synthesis machinery have opened new questions regarding their function. Only three selenoproteins have been identified so far in the fly genome, including sps2. Although mutants that lack the translation elongation factor selB/eEFsec are viable and fertile, the effects of mutations in the selD/spsl gene demonstrate that it is required for development and cell proliferation, hi this chapter, we review the recent advances on fly selenoproteins and their machinery of synthesis. Introduction For more than a century Drosophila melanogaster has been one of the most important genetic models used in biology. Few systems are so easy to manipulate or offer the powerful advantage of allowing researchers to approach any biological question at various points in the development of the organism using tools ranging from genetics to genomics or biochemistry. Due to the remarkable similarities between invertebrates and vertebrates at many levels, Drosophila biology has come to serve as a source of information for human biology and can be expected to have a direct impact on our understanding of human health. Analysis of selenium (Se) metabolism has been shown in recent years to be no exception. As an example, it is worth mentioning that the first mutation in a Drosophila gene involved in Se metabolism was isolated in a P-ZacFF transposon insertion mutant collection screen aimed at identifying genes involved in imaginal disc morphogenesis, a process with no a priori link with Se [1,2]. hi this screen, it was found that mutant flies for an enzyme involved in Se metabolism are lethal in larval/pupal stages, demonstrating the importance of Se in life. Studies on Se metabolism undertaken in a variety of systems have shown that this trace element is both essential and potentially toxic [3-5]. Se occurs

344

Selenium: Its molecular biology and role in human health

most commonly in the organism as a selenocysteine (Sec) residue in natural proteins referred to as selenoproteins [6]. Incorporation of Sec into selenoproteins involves an unusual translation step in which UGA—^normally a stop codon—specifies Sec insertion. This step requires a stem-loop element termed a Sec insertion sequence (SECIS) present in the 3' untranslated regions (3' UTRs) of mRNAs encoding eukaryotic selenoproteins [7,8]. While this represents the main route of specific incorporation, Se can also be found as selenomethionine (Sem) or Sec that has not been specifically incorporated [9]. As in other eukaryotes, the Sec synthesis machinery in Drosophila appears to be quite conserved when compared with that of prokaryotes, as shown by the identification of key homologous genes [1,1012]. In this chapter, we will review current understanding of the biology of selenoproteins and their synthesis in Drosophila, as well as addressing how flies have been used to approach functional studies. Sec synthesis machinery Se intake Se is mainly obtained fi-om the diet and, consequently, physiological levels are constrained by concentrations in the food source [13]. Selenocompounds are mainly obtained as selenoaminoacids, such as L-selenomethionine and LSec, and some inorganic compounds, mostly selenate. Also, a small proportion of the Se in the organism seems to comefi-omrecycling pathways [14]. Se coming fi-om diet or protein turnover can be metabolized by a Sec lyase present in some prokaryotes [15,16] and mammals [14,17], but this enzyme has not yet been identified in flies. Sec lyase catalyzes the conversion of Sec to L-alanine and elemental Se that can be reused for Sec synthesis. The strong dependence on Se coming from the diet raises questions about the role of Se metabolism. To address this, a simple defined medium was devised that supports the growth of adult Drosophila and requires Se supplementation for optimal survival [5]. Those experiments showed that a Se concentration in the fly medium of 10"*-10"^ M mimics normal culture and seems to be the most beneficial for life span. On the other hand, insufficient Se intake in Drosophila produces a clear reduction in survival and flies grown on a Se-depleted medium show more than 50% reduction in the numbers of eggs laid and, thus, reduced fertility. Furthermore, Se deficiency somehow results in down-regulation of the mRNAs encoding selenoproteins [5] as has been observed in mammals [18], revealing the role of Se in selenoprotein expression. Selenocystevl-tRNA The most remarkable feature of Sec synthesis is that it involves its own specific tRNA [19]. In flies, tRNA.SelC is present as a single-copy gene on chromosome 2 that generates a 90-nt tRNA, with some characteristic

Drosophila selenium metabolism and selenoproteins

345

nucleosides that are also present in vertebrates, namely 5 methylcarboxymethyluridine, N6-isopentenyladenosine, and pseudouridine at positions 34, 37, and 55, respectively [10,20]. Like its counterpart in higher organisms, Drosophila tRNA.SelC, also designated Sec-tRNA^^*"^'^", is first aminoacylated with serine and then modified to selenocysteinyl-tRNA, in a reaction that uses selenophosphate as the Se donor. Sec synthetase, the enzyme that catalyses this reaction, has not yet been identified in Drosophila [21]. Electrophoresis of Sec tRNA'^^^^" reveals two major bands expressed throughout development in Drosophila. In mammals, there are also two isoforms of Sec-tRNAf^'^^"" that differ by methylation of mcm^U to form methylcarboxymethyl-5'-uridine-2'-0 methylribose (mcm^Um); this methylation is considered to be the final step of maturation. Furthermore, this reaction depends on Se content, increasing with Se concentration, and seems to have regulatory implications [22]. This may not be the case in Drosophila, since the methylation process is not described and both pools have identical structures, probably corresponding to two stable conformations [20]. Selenophosphate synthetase The biosynthesis of Sec requires monoselenophosphate, the activated form of Se, in the cytoplasm. Selenophosphate synthetase catalyzes a reaction involving ATP and hydrogen selenide coming from the diet or Se delivery from recycling pathways to generate selenophosphate (the Se donor) together with AMP and orthophosphate [14, 15, 23]. In prokaryotes, this reaction is performed by the selD gene product [19]. This gene can encode either a selenoprotein (H. influenzae, M. jannaschii) or an ortholog that contains Cys {E. coli with Cys-17). The presence of this Cys or its equivalent Sec in the catalytic domain of SelD has been considered essential for its enzymatic activity in vitro [23-24]. More recently, it was suggested that replacement of the Cys-17 residue with serine (Ser) might render the enzyme able to use only Se delivered from the recycling pathway [15]. Like in vertebrates, two orthologs of selD have been described in Drosophila [1,11,12]. The Sec-containing form, Sps2, was the first selenoprotein identified in flies [12]. The second selenophosphate synthetase, Spsl or SelD, was separately described following the isolation of a cDNA clone [11] and after screening of a P-element insertion mutant collection to identify genes that affect cell proliferation [1]. This mutation, known as patufet or selD'""^, is lethal in homozygous flies at larval and pupal stages and leads to a marked disruption in the size and morphology of the imaginal discs and brain hemispheres. Drosophila SelD/Spsl has an arginine residue (threonine in mouse and human) instead of Cys or Sec [1]. Northern blot analysis reveals that this gene, located at position 50E on the second chromosome, is expressed

346

Selenium: Its molecular biology and role in human health

throughout development and in situ hybridization shows that low-level expression is ubiquitous up to early gastrulation, at which point expression is upregulated in the gut primordium and nervous system [1,11]. The expression in the brain and imaginal discs, the tissues that will contribute to adult structures, shows a dynamic pattern that appears to correlate with cell proliferation. Analysis of colocalization of selD/spsl expression with BrdU labeling reveals that expression levels are high in dividing cells and low or undetectable in non-dividing cells [25]. Interestingly, these highly proliferative tissues are the most damaged in patufet mutant flies [1]. More detailed studies of cell-cycle markers in the imaginal discs and brain of selD/spsI mutants have shown a reduction in the number of BrdUincorporating cells, fewer mitotic cells, as indicated by a reduction in histone H3 phosphorylation, and high levels of cyclin B expression, suggesting that cells could remain arrested in G2 phase [25]. Loss of function mutations in the selD/spsI gene trigger cell death, mostly through the caspase-dependent Dmp53/Rpr pathway, indicating that the gene is essential for cell viability [1,26]. Flies or cells that are homozygous for seliy""^ have an altered redox system, leading to a burst of reactive oxygen species (ROS) [25-26]. Genetic evidence shows that the initiator caspase DRONC is activated and the effector caspase DRICE is processed to commit mutant cells to die [26]. Moreover, in that study, it was also shown that ectopic expression of the inhibitor of apoptosis DIAPl rescues the viability of seHy""^ mutant cells. These observations indicate that ROS-induced apoptosis in Drosophila is mostly through the caspase-dependent Dmp53/Rpr pathway. Generation of genetic mosaics by mitotic recombination in flies is a powerful technique with which to address gene function, since it allows clones of homozygous mutant cells to be generated in a heterozygous background. This is especially useful in the case of lethal mutations such as seljy""^. Homozygous selDf""^ clones in the adult wing are rounded and smaller than control clones, they contain fewer cells and those cells are smaller, and they reveal some autonomous effects of the mutation, such as suppression of vein differentiation [1]. Non-autonomous effects were also observed in wild-type cells surrounding the clone, leading to abnormal differentiation of ectopic veins. Similar clonal phenotypes were observed in cells homozygous for loss-of-function mutations in genes involved in the Drosophila EGF receptor (DER) pathway [27]. These results, together with some evidence that ROS may play an important role in signal transduction [28], raise questions about how alteration of the redox balance caused by the mutation affects the Ras/MAPK signaling pathway. Both the rough eye phenotype and the ectopic wing veins induced in gain-of-function mutants of the Ras/MAPK pathway are clearly suppressed by the removal of one copy of the selenophosphate synthetase product [29]. The hypothesis that the

Drosophila

selenium

metabolism

and selenoproteins

2>A1

seliy""^ mutation selectively modulates the RAS/MAPK pathway through alteration of the redox balance is further supported by the finding that an increase in ROS caused by the amorphic catalase allele Caf', one of the main enzymes of the Drosophila antioxidant system, also reduces RAS/MAPK signaling [29]. The results of these experiments strongly suggest that accumulation of ROS should be substantially different between heterozygous flies for those mutations and wild-type organisms. Although haplo-insufficient selDf""^fliesdo not have an apparent phenotype when kept under normal laboratory conditions, a significant decrease in life span is observed when they are treated with oxidants [30]. In contrast, while increased amounts of superoxide dismutase (SOD) extend longevity [31], overexpression of spsl in motomeurons leads to a reduction in life span, possibly due to an accumulation of toxic precursors [30]. Finally, it is clear that selLf"^ causes an impairment of selenoprotein synthesis, as revealed by the failure to detect selenoproteins in protein extracts of mutant larvae grown in fly medium containing ^^Se [25]. Therefore, it is tempting to speculate that selenoproteins may be instrumental in maintaining a certain redox state in the cell, as has already been shown for several mammalian selenoproteins [9,32]. Despite the findings mentioned above, it should be noted that purified Drosophila selD/spsl expressed in E. coli does not catalyze selenidedependent ATP hydrolysis in vitro and does not complement a selD deficiency in bacteria [11]. Although at first sight these results appear to disagree with those of other studies using the human spsl gene, in which weak complementation of an E. coli mutation was observed [33], if we take into account the fact that organisms that possess one variant also contain the other we can consider the possibility that SelD/Spsl may have a different function in Se metabolism. It is possible that SelD/Spsl is only efficient in using Se delivered fi-om Sec recycling pathways, as seems to be the case in the human spsl gene fi-om human-lung adenocarcinoma cells [14] or in the E. coli selD(C17S) [15]. The other selenophosphate synthetase, Sps2, contains a Sec residue at the position equivalent to E. coli Cys-17, suggesting a possible autoregulatory role in Sec synthesis; this gene also carries a mammalian-type SECIS. Transgenic embryos expressing a luciferase reporter containing the 3'UTR of the sps2 gene showed significantly higher reporter activity than those lacking the sps2 3'UTR, demonstrating the functionality of the SECIS element present in this 3' terminal region [12]. The spsl gene, located at 2L/31D, is expressed as at least two transcripts: sps2-RA (1348 bp), which was described initially [12], and sps2-RB (1295bp), which is the product of an alternative splicing event in the 4rt exon that, according to FlyBase (www.flybase.org), generates a change in the open reading fi-ame. Both transcripts could encode putative selenoproteins due to

348

Selenium: Its molecular biology and role in human health

an in-frame TGA codon and the SECIS element. Both proteins contain the ATP/GTP binding site required for ATP hydrolysis during selenophosphate production [12]. Expression of the different transcripts has been assessed by both Northern blot analysis and in situ hybridization [12]. Northern blot analyses at different developmental stages revealed the presence of at least the larger transcript throughout development. In situ hybridization studies show a strong accumulation of maternal transcripts produced by nurse cells during oogenesis; these remain present at high levels up to the blastoderm stage (maternal effect). During the remainder of embryogenesis, zygotic expression occurs in a more restricted pattern and is first detected in the embryonic midgut primordium and later in a variety of tissues and organs including the gut and nervous system. A regulatory element that is thought to regulate cell-proliferation-related genes in Drosophila has been identified in the sps2 gene [34,35]. This DNA replication-related element (DRE) located downstream of the initiation site of the gene is essential for its transcription [35]. A transcription factor that specifically binds the DRE sequence has also been isolated in flies and ablation of this factor by double-stranded RNA interference (dsRNAi) experiments shows a significant decrease in dsps2 promoter activity [35]. Sec translational machinery Although much progress has been made in resolving the machinery associated with Sec translation in prokaryotes and vertebrates, less is known about the particular features of the system in Drosophila. Nevertheless, the conservation of some of the elements that have been identified suggests that the basic machinery would be the same. Briefly, the model proposed for eukaryotes includes a requirement for cis and trans factors that form a ribonucleoprotein complex known as a selenosome that functions to incorporate Sec at a UGA codon and thereby prevents translation being stopped. This selenosome complex will consist of at least a SECIS element in the 3 'UTR of the selenoprotein mRNA as a cis factor and a SECIS binding protein 2 (SBP2), a Sec-elongation factor (SelB/eEFsec), a Sec-tRNA '^''^ ^^ and the ribosome itself as trans factors. SBP2 is thought to interact with a conserved region of the SECIS element known as the quartet, with the 28S ribosomal RNA, and with SelB/eEFsec, which will also specifically bind the Sec-tRNA^^'^'^'''^, leading to co-translational incorporation of a Sec when a UGA triplet is encountered [36]. In the fruit fly. Sec insertion is directed, like in other eukaryotes, by the presence of a SECIS in the 3'UTR of the selenoproteins [8]. The Drosophila SECIS element contains the canonical characteristics found mostly in eukaryotic SECIS: the core structure with the quartet of non-Watson-Crick interacting base pairs, and the unpaired adenosines in the apical loop.

Drosophila selenium metabolism and selenoproteins

349

Moreover, this SECIS appears to belong to form 2 in all the identified selenoproteins because of the presence of an additional small stem-loop at the top of the SECIS element [5,12,38]. The trans factors that are known to be required include the Sec-tRNA gene [10] and SelB/eEFsec [37]. Drosophila selB/eEFsec, located at position 57E on Chromosome 2, was found by sequence homology in an in silico analysis of the genome [37]. Its single transcript seems to be expressed throughout development, at high levels during early embryogenesis, lower levels during larval stages, and then at increasing levels in pupae and adults. Because SelB/eEFsec is essential for selenoprotein biosynthesis, it is a suitable target to mutate as a model in which to study the effects of selenoprotein deficiency. selB/eEFsec knockout mutants have been generated by homologous recombination, giving rise to flies that are unable to synthesize SelB/eEFsec and consequently fail to decode the UGA codon as Sec. Moreover, in spite of the impairment of the Sec UGA-decoding mechanism, selB/eEFsec mutant animals develop into fertile flies. In addition, although most known selenoproteins in eukaryotes seem to be involved in antioxidative defense and redox metabolism, life-span studies in the selB/eEFsec mutant do not reveal a role in viability and the mutants do not show sensitivity to induced oxidative stress. These findings challenge the view that a Sec-based oxidative stress defense system was responsible for conserving the selenoprotein biosynthesis system over the course of evolution [37]. Selenoproteins In recent years, genome projects have become an extremely powerful tool through which to identify protein-coding genes. However, because of the non-standard use of the UGA codon, computational gene prediction methods were unable to identify selenoproteins in the sequence of eukaryotic genomes until recently. Only the identification of members of the synthesis machinery was possible based on homology with known genes, as mentioned earlier, for example, with Sps2, the first selenoprotein identified in flies [12]. Using a biochemical approach it was shown that metabolic labeling of flies with '^Se revealed three clear major bands [5,25]. According to the predicted molecular weight, the 42-43 kDa band could correspond to the Sps2 selenoprotein itself, but those experiments did not provide any information on the nature of the other bands. Two different in silico approaches have been used in an attempt to solve this problem. Briefly, one approach combined and improved existing gene prediction programs and developed a method that relies on the prediction of SECIS elements alongside the prediction of genes in which a strong codon bias characteristic of protein-coding regions extends beyond a TGA codon that interrupts the open reading fi-ame [38]. The other approach involved a computational screen to search for SECIS elements followed by

350

Selenium: Its molecular biology and role in human health

selenoprotein gene signature analyses [5]. Both screens led to the identification of three selenoprotein genes in fruit flies: the previously identified sps2 and two new selenoproteins named SelM (also known as SelH or BthD) and SelG (SelK or G-rich). Both proteins incorporate ^*Se when transfected in mammalian cells [38] and all of them were detected after metabolic labeling of flies with ^^Se [5]. These findings indicate that the genes encode true selenoproteins. Furthermore, both of the newly identified genes have paralogs that employ Cys instead of Sec as well as orthologs that use Sec or Cys in other organisms, including vertebrates [38]. SelM/BthD/SelH Selenoprotein SelM/BthD was the first component identified [5,38] in a new family that currently includes several Cys or Sec ortholog representatives in eukaryotes, including the uncharacterized human selenoprotein H [39]. This gene maps to position 12A8 on the X chromosome and has two distant paralogs, CG13186 and CG15147, the latter containing Cys instead of Sec [38]. The selM/BthD gene encodes a 30 kDa protein containing 249 amino acids and a Sec residue belonging to the CXXU motif near the N-terminus [5]. This motif, which is also found in both bacteria and animals, including the mammalian selenoproteins SelT, SelW, and SelH, is similar to the redox motif CXXC [22], suggesting a redox function, with the Sec possibly forming a selenenylsulfide bridge. SelM/BthD shows a dynamic expression pattern. High levels of transcript are detected in adult females, with abundant expression in the developing ovary. In contrast, the expression in males is very weak [5,40]. During early embryogenesis, both transcript and protein seem to display abundant ubiquitous expression, especially in the blastoderm, suggesting that there is a strong maternal contribution [38,40]. At late stages of embryogenesis selM/BthD expression accumulates in the developing salivary gland [40]. Finally, although selM/BthD mRNA appears to be more weakly expressed during larval stages [5,40], in situ hybridization reveals that the transcript is ubiquitously distributed in imaginal disc and larval brain [38]. A dynamic subcellular distribution has been detected using a specific antibody against SelM/BthD [40]. The protein distribution in various Drosophila tissues is cytoplasmic, with a particulate pattern observed in salivary glands. Immunolocalization studies in SL2 cells reveal a colocalization with a Golgi marker, suggesting a possible role in protein secretion or processing [40]. Two different RNAi strategies for silencing selM/BthD expression have been employed to show that loss of selM/BthD reduces viability, although with differing penetrance. Hypomorphic mutants generated by dsRNA injection in embryos exhibit dramatically reduced embryonic viability [41]. Moreover, the use of inducible duplex RNAi under the control of Gal4 drivers revealed that loss of selM/BthD interferes with salivary gland

Drosophila selenium metabolism and selenoproteins

351

morphogenesis and reduces animal viability [40]. SelM/BthD silencing decreases total anti-oxidant capacity in embryos and Schneider cells and increases lipid peroxidation. On the other hand, transient expression of selM/BthD in the cell line decreases lipid peroxidation, suggesting that this protein may have antioxidant functions [41]. SelG/SelK/G-rich Selenoprotein G/G-rich—^named on the basis of its 28% glycine residues—is a less well-known selenoprotein with a mass of 12 kDa. Although homologs of SelG/SelK containing Cys or Sec can be found in vertebrates [5,38,39], its function remains to be elucidated. Located at position 10F4 on Chromosome X, the selG/selK gene gives rise to a single transcript that encodes a 110amino acid selenoprotein, with a Sec residue at the C-terminal penultimate position, similar to some mammalian thioredoxin reductases [22]. SelG/SelK has a cysteine paralog, SelG-like [38]. The two genes appear in tandem, separated by only 320 bp, have the same exonic structure, and share 65% identity at the protein level. Northern blot analysis shows that it is expressed at all stages of fly development [5], while in situ hybridization reveals the distribution to be ubiquitous during embryonic stages [38]. One approach to elucidating selG/selK function has been the characterization of RNAi hypomorphic mutants in cells and embryos. Embryos microinjected with dsRNA corresponding to selG/selK display decreased viability, considered as a percentage of hatched larvae, in addition to morphological defects or developmental retardation [41]. On the other hand, studies in S2 cells have not revealed an effect of SelG/SelK on the redox system. Concluding remarks We would like to conclude this chapter by addressing some of the challenging questions that remain to be resolved. First, only three selenoproteins have been identified in Drosophila so far and it remains possible that the true number of selenoproteins will prove to be higher. Moreover, while interfering with mRNA encoding selG/selK and selM/BthD seems to reduce viability [39,40], the precise cellular function of these fly selenoproteins is unknown except in the case of Sps2, an enzyme involved in the synthesis of selenoproteins [12]. Second, it will be essential to elucidate the role of the various Cyscontaining paralog genes. Do they act as a functional backup? If so, to what extent are the selenoproteins and their paralog genes redundant? Also, it remains to be determined whether or not the two variants of selenophosphate synthetase, SelD/Spsl and Sps2, are redundant. Third, loss of function mutations in selD/spsl are lethal and result in a lack of selenoproteins [1,25]. However, eEFsec mutants, which lack

352

Selenium: Its molecular biology and role in human health

selenoproteins, are viable [37]. This raises the question of whether or not selenoproteins are essential for insect viability and to what extent fly selenoproteins contribute to redox balance or have adopted other functions. It is possible that, in addition to its role in Se metabolism, selD/spsl is also involved in processes that are essential for viability. Alternatively, the accumulation of Se compounds due to the lack of enzymatic activity could account for reduced viability. Drosophila genetics provides an opportunity to approach these questions using a variety of tools to create alleles for those genes and, beyond that, to study their role in metabolism, proliferation, growth and development. Transposable P-elements are still widely used as mutagenesis reagents and form the backbone of projects that seek to generate mutant insertions in every predicted gene in the fly genome. Molecularly mapped deletions have been generated at both a genome-wide and a custom-made level using genetically engineered vectors based on the FLP/FRT system [42]. Moreover, elements have been developed for a wide range of transgenic applications, including enhancer trapping, gene tagging, targeted misexpression, RNA interference delivered by the Gal4AJAS system and homologous recombination. Further genetic experiments will be required to reconcile these issues. References 1. 2.

3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17.

B Alsina, F Serras, J Bagufta, M Corominas 1998 Mol Gen Genet 151-.WZ F Roch, F Serras, FJ Cifuentes, M Corominas, B Alsina, M Amoros, A Lopez-Varea, R Hernandez, D Guerra, S Cavicchi, J Bagutia, A Garcia-Bellido 1998 Mol Gen Genet 257:103 K Schwarz, CM Foltz 1957 J Am Chem Soc 79:3292 OF Combs Jr, SB Combs 1986 The Role of Selenium in Nutrition Academic Press Inc New York pp 532 FJ Martin-Romero, GV Kryukov, AV Lobanov, BA Carlson, B J Lee, VN Gladyshev, DL Hatfield 2001 J Biol Chem 276:29798 DM DriscoU, PR Copeland 2003 Annu Rev Nutr 23:17 MJ Berry, L Banu, YY Chen, SJ Mandel, JD Kieffer, JW Harney, PR Larsen 1991 Nature 353:273 A Krol 2002 Biochimie 84:765 D Behne, A Kyriakopoulos 2001 Annu Rev Nutr 2\:453 BJ Lee, M Rajagopalan, YS Kim, KH You, KB Jacobson, DL Hatfield 1990 Mol Cell fi/o/10:1940 BC Persson, A Bock, H Jackie, G Vorbruggen 1997 J Mol Biol 274:174 M Hirosawa-Takamori, H Jackie, G Vorbruggen 2000 EMBO Rep 1:441 CB Allan, GM Lacourciere, TC Stadtman 1999 Annu Rev Nutr 19:1 T Tamura, S Yamamoto, M Takahata, H Sakaguchi, H Tanaka, TC Stadtman, K Inagaki 2004 Proc Natl Acad Sci USA 101:16162 GM Lacourciere, H Mihara, T Kurihara, N Esaki, TC Stadtman 2000 J Biol Chem 275:23769 GM Lacourciere, TC Stadtman 2001 Biofactors 14:69 H Mihara, T Kurihara, T Watanabe, T Yoshimura, N Esaki 2000 J Biol Chem 275:6195

Drosophila selenium metabolism and selenoproteins 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42.

353

PM Moriarty, CC Reddy, LE Maquat 1998 Mol Cell Biol 18:2932 TC Stadtman 1996 Amu Rev Biochem 65:83 X Zhou, SI Park, ME Moustafa, BA Carlson, PF Grain, AM Diamond, DL Hatfield, BJ Lee, 1999 J Biol Chem llA:mi9 H Romero, Y Zhang, VN Gladyshev, G Salinas 2005 Genome Biol 6: R66 DL Hatfield, VN Gladyshev 2002 Mol Cell Biol 22:3565 Z Veres, lY Kim, TD Scholz, TC Stadtman 1994 J Biol Chem 269:10597 lY Kim, Z Veres, TC Stadtman 1992 J Biol Chem 161:19650 B Alsina, M Corominas, MJ Berry, J Bagufla, F Serras 1999 J Cell Sci 112:2875 M Morey, M Corominas, F Serras 2003 J Cell Sci 116:4597 FJ Diaz-Benjumea, E Hafen 1994 Development 120:569 T Finkel 1998 Curr Opin Cell Biol 10:248 M Morey, F Serras, J Bagufla, E Hafen, M Corominas 2001 Dev Biol 238:145 M Morey, F Serras, M Corominas 2003 FEBS Lett 534:111 TL Parkes, A J Elia, D Dickinson, AJ Hilliker, JP Phillips, GL Boulianne, 1998 Nat Genet 19:171 S Gromer, JK Eubel, BL Lee, J Jacob 2005 Cell Mol Life Sci 61:1A\A SC Low, JW Harney, MJ Berry 1995 J Biol Chem 270:21659 A Matsukage, F Hirose, Y Hayashi, K Hamada, M Yamaguchi 1995 Gene 166:233 JS Jin, S Back, H Lee, MY Oh, YE Koo, MS Shim, SY Kwon, I Jeon, SY Park, K Back, MA Yoo, DL Hatfield, BJ Lee 2004 Nucleic Acids Res 32:2482 A Lescure, D Fagegaltier, P Carbon, A Krol 2002 Curr Protein Pept Sci 3:143 M Hirosawa-Takamori, HR Chung, H Jackie 2004 EMBO Rep 5:317 S Castellano, N Morozova, M Morey, M J Berry, F Serras, M Corominas, R Guigo 2001 EMBO Rep imi GV Kryukov, S Castellano, SV Novoselov, AV Lobanov, O Zehtab, R Guigo, VN Gladyshev 2003 &ie«ce 300:1439 SY Kwon, P Badenhorst, FJ Martin-Romero, BA Carlson, BM Paterson, VN Gladyshev, BJ Lee, DL Hatfield 2003 Mol Cell Biol 23:8495 N Morozova, EP Forry, E Shahid, AM Zavacki, JW Harney, Y Kraytsberg, MJ Berry, 2003 Genes Cells 8:963 E Ryder 2004 F Blows, M Ashbumer, R Bautista-Llacer, D Coulson, J Drummond, J Webster, D Gubb, N Gunton, G Johnson, CJ O'Kane, D Huen, P Sharma, Z Asztalos, H Baisch, J Schulze, M Kube, K Kittlaus, G Renter, P Maroy, J Szidonya, A RasmusonLestander, K Ekstrom, B Dickson, C Hugentobler, H Stocker, E Hafen, JA Lepesant, G Pflugfelder, M Heisenberg, B Mechler, F Serras, M Corominas, S Schneuwly, T Preat, J Roote, S Russell Genetics 167:797

Chapter 31. Selenoproteins in parasites Gustavo Salinas Catedra de Inmunologia, Facultad de Quimica-Facultad de Ciencias, Universidad de la Republica. Instituto de Higiene, Avda. A. Navarro 3051, Montevideo, CP 11600, Uruguay

Alexey V. Lobanov and Vadim N. Gladyshev Department of Biochemistry, University of Nebraska, Lincoln, NE 68588-0664, USA

Summary: Parasites, which cause an enormous burden in the population of the third world, are a diverse group of organisms, many of which are sensitive to oxidative stress imposed by their hosts. In recent years, several selenoprotein families, some with antioxidant properties, have been described and characterized in metazoan parasites. Glutathione peroxidase and thioredoxin glutathione reductase (TGR) appear to be essential selenoproteins in flatworms (phylum Platyhelminthes). TGR is the single enzyme that provides reducing equivalents to both thioredoxin and glutathione pathways, in contrast to hosts, which evolve parallel pathways. In roundworms (phylum Nematoda), selenoproteins have recently been described, revealing species differences in the Sec/Cys protein sets and the presence of an unusual SECIS element. Plasmodium sp, one of the most important protozoan parasites that affect humans, also decode Sec. The selenoprotein families encoded by Plasmodial genomes have neither Sec nor Cys homologs in their hosts, raising the possibility that targeting their selenoproteomes may provide new treatment strategies. Introduction Although significant research efforts have been made to study selenoproteins and selenocysteine insertion systems in humans and various model organisms, little has been reported in the literature regarding the utilization of selenium in eukaryotic parasitic organisms. This chapter focuses on the progress made in the characterization of selenoenzyme families in flatworms, the recent advances in the synthesis and utilization of selenoproteins in roundworms and protozoan parasites, and discusses why selenoproteins of platyhelminths and Plasmodia may represent interesting targets for chemo- or immune-prophylaxis.

356

Selenium: Its molecular biology and role in human health

Parasites: diverse organisms that face similar oxidative stress challenges Parasites live at least part of their lifecycle inside another organism (the host), which they exploit for their own survival and reproduction. This definition includes different types of infectious agents (viruses, bacteria, fungi, protozoa, helminths). However, for historical reasons, the term is most often reserved for 'protozoa' and 'helminths' organisms. Indeed, parasitology was identified as a separate research field during the exploration of the tropics and the establishment of 'tropical medicine' [1]. Both 'protozoa' and 'helminths' also include free-living organisms, and neither 'protozoa' nor 'helminths' are monophyletic; on the contrary, both groups are represented by highly divergent phyla. Nonetheless, this historical classification is not useless. These two groups of parasites are very different: protozoan are unicellular protists, which multiply quickly within the host, and are, in most cases, intracellular in habitat; in contrast, helminths are metazoan organisms with complex multicellular organization (with nervous system and reproductive organs), which undergo complex metamorphoses and migrations within the host. Table 1 presents the main features of the major human parasitic infections. Table 1. Major human parasites (Source: [2])

Protozoan parasites'* Species (Disease) Plasmodium sp (Malaria) Trypanosoma brucei (sleeping sickness ) Trypanosoma cruzi (Chagas disease*) Leishmania sp (Leishmaniasis)

Helminths parasites^ Species/Disease Schistosoma sp (Schistosomiasis or bilharzia^) Onchocerca volvulus (Onchocerciasis or river blindness ) Filariidae family (Lymphatic filariasis )

Phylum Apicomplexa

Death per year/DALYs' 1,124,000/42,280,000

Kinetoplastida

50,000/1,590,000

Kinetoplastida

13,000/649,000

Kinetoplastida

59,000/2,357,000

Phylum Plathyhelminthes

Death per year/DALYs 15,000/1,760,000

Nematoda

0/987,000

Nematoda

0/5,644,000

"DALYs: DisabiUty Adjusted Life Years (the number of healthy years of life lost due to premature death and disabiUty). Protozoan parasites include many diverse phyla, among them Apicomplexa and Kinetoplastida.

Selenoproteins in parasites

357

Distribution: mainly confined to poorer tropical areas of Africa, Asia and Latin America. More than 90% of malaria cases and the great majority of malaria deaths occur in ttopical Africa. Plasmodium falciparum is the main cause of severe clinical malaria and death. Distribution: 36 countries in sub-Saharan Africa *Distribution: Latin America f Distribution: Endemic in 88 countries on 4 continents. Two forms of the disease: cutaneous (caused by Leishmania major), and visceral (caused by L. donovani) Helminth parasites are contained in three phyla: Nematoda (roundworms), Platyhelminthes (flatworms) and Acantocephala (spiny-headed worms). Helminth infections are rarely fatal, but pose an enormous burden to human population in the tiopics Distribution: endemic in 74 developing countries with more than 80% of infected people living in sub-Saharan Africa Distribution: 35 countries in total. 28 in tropical Africa, where 99% of infected people live. Isolated foci in Latin America and Yemen. Distribution: Endemic in over 80 countries in Africa, Asia, South and Central America and the Pacific Islands. Three species are of significance, Wuchereria bancrofti, Brugia malayi and Brugia timori.

In Spite of the diversity of parasites, all face similar biological problems that relate to their parasitic lifestyle. Among them, the neutralization of the effector mechanisms deployed by the host immune system is of paramount importance. Resident macrophages and inflammatory-site phagocytic leukocytes (mostly neutrophils, but also monocytes and eosinophils, depending of the type of infection) are cells equipped to kill foreign organisms. They possess an oxidase system located in their plasma membrane, which becomes activated upon certain stimuli, for example, by interaction of cell receptors with antibodies bound to the foreign organism or with parasite molecular motifs (Figure la) [3]. Subsequently, 'respiratory burst' (increase in oxygen uptake not linked to respiration) takes place and produces superoxide anion and additional reactive oxygen species (ROS) [4]. Large amounts of nitric oxide (NO) are also produced by macrophages (and to a lesser extent by neutrophils) activated by a variety of immunological stimuli, such as y-interferon and tumor necrosis factor. NO reacts with superoxide to produce peroxynitrite and other reactive nitrogen species (RNS) (Figure lb) [5]. In addition, activated neutrophils and eosinophils release myeloperoxidase and eosinophil peroxidase, respectively, that catalyze the conversion of hydrogen peroxide and halides into hypohalous acids that are powerful oxidants and can form further damaging species [4]. Collectively, ROS and RNS are powerful oxidants and nitrating species: they can inactivate enzymes and initiate the process of lipid peroxidation and nitration, which leads to radical chain reactions that further damage membranes, nucleic acids and proteins (Figure Ic). These processes (and an additional arsenal of the host effector cells, such as hydrolytic enzymes) may

358

Selenium: Its molecular biology and role in human health

PRR

PAMP

(b) FDSand RNS

Activated leukocytes

M'"^''*+ 'NO,

ONOO Myeto peroxidase

I

HOCI

pKa~ 6,8jp

p,|^

R'OH + NO,

"s.C;0

ONOOH

i

oNOoco;

•OH + •NO

•NOp+ CO,

(c) Damage • Rotein oxidation and nitration:

R 9 , RSS?, R3DK R 3 0 0 H nittotyrosine , inactivatlon of Fe/Sclusteis (metal OMdation) • Lipid peroxidation and nitration: LDOK LNO,, IDONO^, propagation of radical ctiain reactions • DMA strand breaks

(d) Bizymatic defenses SOD GR< •<

dismutation of superoMde reduction of peroMdes

TBc: reductionof sulfonic acid Sbme TX: reduction of perD)^itrrte (^pair enzymes Tx reduction of protein disulTrdes Methionine sulfoxide reductase Sjifiredoidns

l-tO^-^HO RDOH-»RDH RSOH->FEH ONOOH-+NO, FSff?-* R3H R90OH^.R33H

Selenoproteins in parasites

359

Figure 1. Reactive oxygen and nitrogen species generated by the host immune response and antioxidant defenses, (a) Recognition of parasites by host leukocytes (such as macrophages, neutrophils and eosinophils) occurs by pattern recognition receptors (PRR) that bind to pathogen-associated molecular patterns (PAMPs), or through antibodies (Ig), and leads to activation of host immune cells. Upon activation, these cells produce superoxide ("02") and nitric oxide ('NO) radicals. 'NO is produced in the cytosol (but can cross membranes) by inducible nitric oxide synthase (iNOS); 02" is produced by a multi-component, membraneassociated NADPH oxidase. Superoxide is released towards the extracellular space in the case of non-phagocytosable parasites {e.g., worms), or towards the phagosome (topologically equivalent to the extracellular space) in the case of intracellular parasites {e.g., protozoans), (b) 'NO and 'O2" react at diffusible controlled rate to produce peroxynitrite (ONOO"). Peroxynitrite can react in one-electron oxidations {e.g., with transition metal centers), two electrons oxidations (of a given target), or with CO2, redirecting its reactivity. It also decomposes spontaneously into other ROS and RNS such as 'OH and •NO2. In addition, activated neutrophils and eosinophils release myeloperoxidase and eosinophil peroxidases, respectively, which catalyze the conversion of hydrogen peroxide and halides into hj^sohalous acids, (c) Collectively, these products can inactivate enzymes, damage membranes and nucleic acids, and ultimately kill the parasitic organisms. (D) Parasites' defenses include antioxidant enzymes that directly scavenge superoxide, decreasing peroxynitrite formation (superoxide dismutases), and hydrogen and organic peroxide reductases (GPx and TPx). Some TPx have also been shown to reduce peroxynitrite catalytically. Repair mechanisms include methionine sulfoxide reductase, thioredoxin, and sulfiredoxin among others. *R'H denotes a hydrocarbon chain, or alcohol (R'H=ROH), or a thiol R'H=RSH)

ultimately lead to killing parasitic organisms. Yet, well-adapted parasites cope with the oxidative stress imposed by the host's immune response by a series of cellular chemicals and antioxidant enzymes that directly neutralize ROS and RNS (Figure Id), and constitute important model organisms to study antioxidant defense. Several antioxidant enzymes found in parasites belong to selenoprotein families. Glutathione peroxidase: tlie first selenoenzyme described in parasites Glutathione peroxidase was the first selenoenzyme to be characterized from a parasite. A cDNA from the platyhelminth Schistosoma mansoni encoding a GPx with a TGA in-frame at the active site was cloned in the early 1990s [6]. The protein encoded by this gene has biochemical properties similar to mammalian phospholipid hydroperoxide glutathione peroxidase (PHGPx); its activity being higher with phosphatidyl choline hydroperoxide and other phospholipid hydroperoxides than with hydroperoxide substrates, such as cumene hydroperoxide and hydrogen peroxide [7]. GPx and superoxide dismutase, another antioxidant enzyme, co-localize in the tegument and gut epithelium of adult worms, which are the exposed interfaces of the parasite towards the host [8]. Additional evidence suggests that antioxidant enzymes, and GPx in particular, are vital for ROS neutralization and parasite survival within the host. Indeed, expression of GPx is developmentally regulated, with the highest levels present in the adult worm [8], the stage most resistant

360

Selenium: Its molecular biology and role in human health

to oxidative stress and immune elimination [9]. In addition, GPx expression is upregulated by hydrogen peroxide and xanthine/xanthine oxidase generated ROS [10]. Recently, a search for GPx in Expressed Sequence Tag databases (dbEST) of platyhelminths identified a second GPx (GPx2) in S. mansoni and S. japonicum [11]. GPx2 also encodes a Sec residue at the active site and possesses an N-terminal signal peptide, which targets this isoform to the extracellular compartment, suggesting that this secreted variant would be important for extracellular hydroperoxide removal, helping to protect the parasite in its immediate environment. In this study, a GPxl ortholog whose 3 '-untranslated region revealed the presence of a SECIS element was also identified in Echinococcus granulosus (another flatworm) transcriptome using the SECISearch algorithm (Chapter 9 and http://genome.unl.edu/SECISearch.html) [12]. In contrast to platyhelminths, the corresponding Cys-containing enzymes appear to occur in nematodes [13], as reviewed in [14]. Nevertheless, recent datamining of nematode dbEST revealed some exceptions (see below) [15]. Free-living nematode Caenorhabditis elegans has no Sec-containing GPx encoded in its genome [15]. GSH- and Trx-reduction pathways in platyhelminth parasites are controlled by a single selenoenzyme In most living organisms, there are two analogous and mutually supporting enzymic systems that provide antioxidant defense to cells: the glutathione (GSH) and the thioredoxin (Trx) systems (Figure 2) [16,17]. These systems have overlapping yet distinct targets. GSH, due to its reactivity and intracellular concentration, is one of the most important cellular antioxidants, being efficient in rescuing small disulfide molecules and in reacting directly with ROS. The major function of Trx is to maintain cysteine residues in substrate proteins in the reduced form. In addition to their direct function as antioxidants, GSH and Trx provide electrons to GPx and Trx peroxidase (TPx), respectively, which reduce hydrogen peroxide and organic hydroperoxides, and to methionine sulfoxide reductase, which is also an important antioxidant repair enzyme. GSH and Trx are usually reduced by GSH and Trx reductases (GR and TR), respectively, at the expense of NADPH oxidation. Recent characterization of these systems in platyhelminth parasites has shown that 'conventional' GR and TR are absent; instead, the GSH and Trx systems are intermingled with the enzyme thioredoxin glutathione reductase (TGR), which provides reducing equivalents to both pathways (Figure 2).

Selenoproteins in parasites

361

(a) Comparison of the SSH, Tnc and llntod Tnc43SH systems

(I) Glulathlons system NADPH + H* •

NADP*

X

GR red

|GlutathloneY 6rx Ysgrt - ^ ox W r e d - ^ '

|il) Thioredoxlii system HADPH + H*

HADP'

Y

W

rerf

-

^

ox V S » ^ '

ox TargetsIW (III) Linked Uiloredoxin-glnlathlone systems MADPH

HADP* +H*

(b) Domains and redox centre i in the primary stryctyre of 6 R , TrxR, TGR, Grx and Tnc CXXXXG

i] F«)

GR

1 FAO-WtMtinfl

CXXMXG

TR

FA0 CMC

TGR Gra

1

cxxxxc FAB

1

cxxc : SSH

:

GXXG

Trx

\

T^x c X

taSu

IK :,.:,..:.., domairia

V..J fi - =«

c f--?.

TR duniainy Trx t: G g;*™0*,_ C J Xi(\*^

fi Trx

r-

c- >^ '••I'-'

t

The

362

Selenium: Its molecular biology and role in human health

Figure 2. Linked thioredoxin-glutathione systems, (a) Comparison of thioredoxin, glutathione and linked thioredoxin-glutathione systems. The glutathione system comprises (i) GR, GSH and Grx, whereas the thioredoxin system consists of (ii) TR and Trx. In linked Trx-GSH systems (iii), TGR functionally replaces TR, GR and Grx, providing reducing equivalents to targets of both systems. In all systems, NADPH is the upstream donor of reducing equivalents, (b) Components of the thioredoxin and glutathione systems. Redox centers of GR, TR, TGR, Grx and Trx are indicated, as well as the FAD prosthetic group and the ligands NADPH and GSH. TR and TGR possess a C-terminal extension missing in GR, which contains the Cterminal GCUG redox-active motif TGR possesses an N-terminal Grx domain that is absent in TR and GR. The Grx and Trx domains contain the CXXC redox center. Grx, unlike Trx, binds GSH. (c) Schematic representation of electron flow in TGR. TGR, like GR and TR, is a homodimer, with monomers oriented in a head-to-tail manner. Electrons flow from NADPH to FAD, to the CX4C redox center, to the C-terminal GCUG redox center of the second subunit, to the CX2C redox center of the Grx domain of the first subunit, and to targets, including GSSG (left scheme). Alternatively, electrons can flow, presumably directly, from the GCUG redox center to Trx (right scheme). The model proposes a flexible hinge, which connects the TR and Grx domains. This organization allows electrons to flow to the "in built' Grx domain or to Trx. Parts (a) and (b) in the figure reprinted with modifications from [11] with copyright with permission from Elsevier.

This protein is a second selenoenzyme family that has been characterized in platyhelminth parasites (reviewed in [11]). TGR is an oxidoreductase shown to possess TR, GR and Grx activities, achieving its broad substrate specificity by a fusion between Grx and TR domains (Figure 2b); this domain fusion was originally described in a mouse testis TGR [18]. Experimental and in silico data support the proposition that TGR is the single enzyme responsible for recycling both oxidized Trx and GSH in platyhelminth parasites. Treatment of S. mansoni adult worm extracts with auranofin, a known inhibitor of Sec-containing TRs, resulted in complete inhibition of TR and GR activities [19]. In addition, TGR was the single protein isolated from Taenia crassiceps (also a flatworm) extracts as a result of tracing GR and TR activities [20]. Examination of EST databases from Schistosoma species, which covers more than 90% of the gene content of this organism [21], revealed cDNAs encoding TGR, but not conventional TR or GR [11]. The biochemical characterization of E. granulosus and T. crassiceps TGR indicated that the native enzyme shuttles elecfrons from NADPH to oxidized Trx (TR activity), GSSG (GR activity) and glutathionemixed disulfides (Grx activity). The stoichiometric inhibitory effect of auranofin on both GR and TR activities of TGR indicates that the Seccontaining C-terminal redox center participates in elecfron fransfer to GSSG and oxidized Trx [20,22]. In addition, TR and Grrx domains can function either in coupled reactions or independently. Conventional TRs neither bind GSH nor possess GR activity; thus, the N-terminal Grx domain of TGR would reduce GSSG, accepting elecfrons from the Sec-containing C-terminal redox center. The idea that the C-terminal redox center donates elecfrons to the fused Grx domain implies that the Grx domain of TGR would be linked

Selenoproteins in parasites

363

to the TR domains by a flexible hinge to allow reduction of the oxidized Trx (Figure 2c). It is interesting to note that T. crassiceps TGR showed a hysteretic behavior in enzymatic assays with GSSG at high concentrations; this observation led the authors to propose a model in which TGR would possess high and low affinity sites for glutathione [20]. Clearly, further biochemical characterization and structural data on this multifunctional enzyme are needed that will shed light on the mechanism of catalysis, hi addition, molecular characterization of the corresponding gene could also provide clues regarding the mechanism of generation of isoforms. hideed, the analysis of TGR in E. granulosus revealed two trans-spliced cDNAs derived from a single gene [22]. These variants code for mitochondrial (mt) and cytosolic (c) TGRs, containing identical Grx and TrxR domains, but differing in their N-termini. These variants derive from alternative initiation of transcription, followed by trans-splicing. Similarly, mtTGR and cTGR variants also derived fi"om a single gene have been identified in S. mansoni [11]. Collectively, the results from platyhelminth studies strongly suggest that TGR is the main pyridine-nucleotide thiol-disulfide oxidoreductase in these organisms, in contrast to their hosts, where there is some redundancy of mechanisms for recycling oxidized Trx and GSH. Very little has been published about these pathways in the other phylum of helminth parasites (Nematoda), and to the best of our knowledge, nothing is known about Sec/Cys-containing TR or TGR in parasitic nematodes. However, no single genome has yet been completed from metazoan parasites. Selenoproteins of nematode parasites: old families, unusual SECIS An in silica analysis of Caenorhabditis elegans and Caenorhabditis briggsae (free-living nematodes) genomes revealed that these organisms encode a single a selenoprotein, TR [15], corroborating earlier experimental data [23]. However, no experimental studies have yet been performed with selenoproteins from parasitic nematodes. Nevertheless, in a recent study [15], the existing nematode ESTs were searched for selenoprotein genes using SECISearch and by screening for homologs of known selenoproteins. These analysis identified selenoprotein homologs of selK, selT, selW, Sepl5, selenophosphate synthetase and GPx. Two interesting points were noted from these analyses. First, various nematodes encode different selenoproteins, and the distribution of selenoprotein families within this phylum is mosaic. Second, it was found that all detected nematode selenoprotein genes contained an unusual form of SECIS element, with G rather than a canonical A at the conserved position preceding the quartet of non-Watson-Crick base pairs [15].

364

Selenium: Its molecular biology and role in human health

Selenoproteins of protozoan parasites: waiting for surprises? Very little is known about selenoproteins from protozoan parasites. Recently, the presence of tRNA^^' was described in several species of the phylum Apicomplexa [24] (Lobanov et al., submitted). Plasmodium falciparum, which is the causative agent of malaria - the most overwhelming human parasitic infection, belongs to this phylum. The finding of tRNA^'" was consistent with the presence of putative EFsec and selenophosphate synthetase in P. falciparum and other Plasmodia. In addition, tRNA^*'' was observed in Toxoplasma, but not in Cryptosporidium parasites. Genomewide searches for SECIS elements in the six Plasmodium genomes revealed four selenoprotein genes. Interestingly, homology analyses of these proteins identified no hits outside Apicomplexa, suggesting that these selenoproteins do not exist in the apicomplexan hosts. These properties make the new selenoproteins attractive targets for anti-malaria drug development. The other reference in the literature to a parasite Sec-decoding protozoan is the description of a Cys-containing selenophosphate synthetase from Leishmania major [25]. Leishmania belongs to trypanosomatidae family, which also includes Trypanosoma brucei, and T. cruzi (Table 1), which are causative agents of disabling and fatal diseases in the poorest rural population of the third world [26]. Consistent with the finding of selenophosphate synthetase, recent bioinformatics analyses revealed three selenoprotein genes in several Trypanosoma genomes (Lobanov and Gladyshev, unpublished). Finally, no single reference could be found in the literature regarding a Sec-decoding amoebae, a traditional group of protozoa that include the parasitic amoebae of humans, Entamoebae histolytica. Parasite selenoproteins: drug or vaccine candidates? From a global perspective, the confrol of parasitic infections, which are a major cause of disability and mortality in many developing countries, remains as one of the most important challenges for medicine in the 21^' century [2]. Although there are safe and effective drugs to control some parasitic diseases, parasites can develop resistance to drugs rendering them ineffective, as it has been the case of certain antimalarial drugs [27]. Thus, effective vaccines and new drugs against parasitic organisms are needed. The task ahead is enormous considering that parasite and hosts are eukaryotic organisms; as yet, there is not a single vaccine for a human parasitic infection. Whether selenoproteins can be drug targets or generate immunity depends on premises that are not necessarily different from those for any other target protein: the validity of a drug target would rely on it being an essential protein, and sufficiently different from the host homolog(s) as to be selectively inhibited. Likewise, a good vaccine candidate should generate an

Selenoproteins in parasites

365

appropriate and selective immune response against the parasite, without inducing pathology to the host. In platyhelminths, TGR is an attractive pharmacological target because of the lack of redundant mechanisms (i.e., TR and GR) to provide reducing equivalents to essential enzymes. Inhibition of this enzyme could lead to impaired synthesis of DNA and antioxidant defenses, compromising parasite survival. TGR may also be a good vaccine candidate, since it is a large protein with a degree of identity to host enzymes below 60%. However, there are no studies regarding TGR as an immunogen. Contrary to TGR, there are promising studies on the use of GPx as a vaccine candidate. Vaccination of mice (not a natural host) against the platyhelminth S. mansoni with naked DNA constructs containing Sec-containing GPx showed significant levels of protection compared to a control group [28]. In this context, it is important to emphasize not only the fact that GPx appears to be important at the host parasite interface, but also that platyhelminth lack catalase and rely exclusively on GSH and Trx peroxidases for hydrogen peroxide removal. In the case of protozoan parasites, further studies are needed to identify and functionally characterize their selenoproteins. Nevertheless, it is highly significant that the four selenoproteins identified in Plasmodium sp have neither Sec nor Cys homologs in humans. Considering that Sec is usually located at the redox-active sites of enzymes, the selenol- and thiol-based redox systems may play vital an important role in the survival of protozoan parasites [29]. Finally, selenoproteins may be different to other proteins in one respect: electrophilic drugs, such as gold or platinum compounds, or alkylating agents that react preferentially with Sec over Cys may affect the parasite and the host to a different extent, depending on the relative importance of selenoproteins for the two organisms, and the presence/absence of Cyscontaining enzymatic back up systems. Acknowledgements This work has been supported by Fogarty International Research Collaboration Award TW006959 and Ministry of Education, Uruguay, PDT 29/171. References 1. 2. 3. 4. 5. 6.

K Warren 1988 The Biology of Parasitism PT Englund, A Sher (Ed) Alan R. Riss Inc New York 3 WHO The world health report -changing history. 2004 (http://www.who.int/whr/2004/en/report04_en.pdf) DH McGuinness, PK Dehal, RJ Pleass 2003 Trends Parasitol 19:312 BG Halliwell, JMC Gutteridge 1999 Free Radicals in Biology and Medicine Oxford University Press Inc New York R Radi, G Peluffo, MN Alvarez, M Naviliat, A Cayota 2001 Free Radic Biol Med 30:463 DL Williams, RJ Pierce, E Cookson, A Capron 1992 Mol Biochem Parasitol 52:127

366 I. 8. 9. 10. II. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21.

22. 23. 24. 25. 26.

27. 28. 29.

Selenium: Its molecular biology and role in human health M Maiorino, C Roche, M Kiess, K Koenig, D Gawlik, M Matthes, E Naldini, R Pierce, L Flohe 1996 EurJBiochem 238:838 H Mei, PT LoVerde 1997 Exp Parasitol 86:69 GM Mkoji, JM Smith, RK Prichard 1988 Int J Parasitol 18:661 UE Zelck, B Von Janowsky 2004 Parasitology 128:493 G Salinas, ME Selkirk, C Chalar, RM Maizels, C Fernandez 2004 Trends Parasitol 20:340 GV Kryukov, VM Kryukov, VN Gladyshev \999 J Biol Chem 274:33888 L Tang, K Gounaris, C Griffiths, ME Selkirk 1995 J Biol Chem 270:18313 K Henkle-Duhrsen, A Kampkotter 2001 MolBiochem Parasitol 114:129 K Taskov, C Chappie, GV Kryukov, S Castellano, AV Lobanov, KV Korotkov, R Guigo, VN Gladyshev 2005 Nucleic Acids Res 2005 ZTi-.llll A Holmgren 2000 Antioxid Redox Signal 2:811 PG Winyard, CJ Moody, C Jacob 2005 Trends Biochem Sci 30:453 QA Sun, L Kimarsky, S Sherman, VN Gladyshev 2001 Proc Natl Acad Sci USA 2001 98:3673 HM Alger, AA Sayed, MJ Stadecker, DL Williams 2002 Int J Parasitol 32:1285 JL Rendon, IP del Arenal, A Guevara-Flores, A Uribe, A Plancarte, G MendozaHemandez 2004 Mol Biochem Parasitol 133:61 S Verjovski-Almeida, R DeMarco, EA Martins, PE Guimaraes, EP Ojopi, AC Paquola, JP Piazza, MY Nishiyama, Jr., JP Kitajima, RE Adamson, PD Ashton, MF Bonaldo, PS Coulson, GP Dillon, LP Farias, SP Gregorio, PL Ho, RA Leite, LC Malaquias, RC Marques, PA Miyasato, AL Nascimento, FP Ohlweiler, EM Reis, MA Ribeiro, RG Sa, GC Stukart, MB Soares, C Gargioni, T Kawano, V Rodrigues, AM Madeira, RA Wilson, CF Menck, JC Setubal, LC Leite, E Dias-Neto 2003 Nat Genet 2003 35:148 A Agorio, C Chalar, S Cardozo, G Salinas 2003 J Biol Chem 2003 Apr 1111%: 12920 VN Gladyshev, M Krause, XM Xu, KV Korotkov, GV Kryukov, QA Sun, BJ Lee, JC Wootton, DL Hatfield 1999 Biochem Biophys Res Commm 259:244 T Mourier, A Pain, B Barrell, S Griffiths-Jones 2005 RNA 11:119 PC Jayakumar, VV Musande, YS Shouche, MS Patole 2004 DNA Seq 15:66 CM Morel, T Acharya, D Broun, A Dangi, C Elias, NK Ganguly, CA Gardner, RK Gupta, J Haycock, AD Heher, PJ Hotez, HE Kettler, GT Keusch, AF Krattiger, FT Kreutz, S Lall, K Lee, R Mahoney, A Martinez-Palomo, RA Mashelkar, SA Matlin, M Mzimba, J Oehler, RG Ridley, P Senanayake, P Singer, M Yun 2005 Science 309:401 TE Mansour Chemotherapeutic Targets in Parasites: Contemporary strategies 2002 T Mansour (ed) Cambridge University Press Cambridge 4 KA Shalaby, L Yin, A Thakur, L Christen, EG Niles, PT Lo Verde 2003 Vaccine 22:130 S MuUer, E Liebau, RD Walter, RL Krauth-Siegel 2003 Trends Parasitol 19:320

Chapter 32. Incorporating 'omics' approaches to elucidate the role of selenium and selenoproteins in cancer prevention Cindy D. Davis and John A. Milner Nutritional Science Research Group, Division of Cancer Prevention, National Cancer Institute, National Institutes of Health, Bethesda, MD 20892, USA

Summary: Epidemiologic and preclinical studies provide evidence that increasing dietary selenium (Se) may have cancer protective properties. However, variation in cancer incidence among and within populations with similar Se intake suggests that an individual's response may reflect interactions with genetic and/or environmental factors. The "omics" of nutrition (nutrigenomics, nutrigenetics, nutritional epigenomics, nutritional transcriptomics, proteomics and metabolomics) may assist in understanding the cellular and molecular events associated with the cancer protective effects of Se, as well as in identifying responders and non-responders. Approaches, that utilize transgenic and knockout mice with altered selenoprotein expression offer models to evaluate the importance of selenoproteins or small molecular selenocompounds in mediating the cancer protective effects of Se. While the challenges will be enormous, the potential rewards in terms of both cancer morbidity and mortality will be of equally great magnitude. Introduction Extensive evidence indicates that dietary selenium (Se) supplementation reduces the incidence of cancer in experimental animals. Adding Se to the diet or drinking water inhibits initiation and/or post-initiation stages of liver, esophageal, pancreatic, colon and mammary carcinogenesis and spontaneous liver and mammary tumorigenesis in several rodent models [reviewed in 1,2]. Similarly, ecological studies have usually found an inverse relationship between Se status and mortality from cancer of the large intestine, rectum, prostate, breast, ovary, lung, and leukemia [3]. Data from most case-control and cohort studies indicate its possible protective relationship with lung and prostate cancer, but data is not overly convincing for other cancer sites, including breast and colon/rectum [4]. A recent meta-analysis suggests that Se supplementation may afford some protection against lung cancer in populations where average Se levels are traditionally low [5]. Evidence

368

Selenium: Its molecular biology and role in human health

suggests that toenail Se may be a useful predictor of status [5]. Cohort studies also have identified low baseline serum or toenail Se concentrations as a risk factor for prostate cancer [6,7]. A recent intervention study provides the most compelling evidence for the protective effects of Se against cancer [8]. This randomized controlled trial was designed to test Se as a deterrent to the development of basal or squamous carcinomas. Secondary end-point analyses showed that the mineral resulted in a significant reduction in total cancer mortality (RR = 0.5), total cancer incidence (RR = 0.63), and incidence of lung (RR = 0.54), colorectal (RR = 0.42), and prostate (RR = 0.37) cancer [8]. Participants with baseline plasma Se concentration in the lowest two tertiles (<121.6 ng/ml) experienced reductions in total cancer incidence, whereas those in the highest tertile showed an elevated incidence (HR =1.20, 95% CI= 0.77-1.86) [9]. Re- analysis of incidence data through the end of the blinded clinical trial indicated that supplementation significantly reduced lung [10] and prostate cancer [11] only in individuals with the lowest baseline Se concentrations. While these data are intriguing, they document that not everyone benefits equally from Se supplementation.

Nufrlgenetlcs

DNA

Nutritional Epigenomlcs Itlutritional rranscrlptomlcs -

RNA

i

Phenotype

Selenium

Proteomics

lUetabolomics

•' Protein '•

»-f Metabolite'

Figure 1. Variation in the 'omics' of nutrition (nutrigenetics, nutritional epigenomics, nutritional transcriptomics, proteomics and metabolomics) likely contribute to the different phenotypic expression among individual's in response to dietary Se for cancer prevention (modified from 1).

Variation in cancer incidence among and within populations with similar Se intake suggests that an individual's response may reflect interactions with multiple genetic factors. Nutrigenomics, defined as the interaction between nutrition and an individual's genome or the response of an individual to

Incorporating 'omics' approaches

369

different diets, will likely provide important clues about responders and nonresponders (see Figure 1). Nutrigenomics encompasses an understanding about how the response to dietary components depend on an individual's genetic background or nutrigenetics, nutrient induced changes in DNA methylation and chromatin alterations or nutritional epigenetics, and nutrient induced changes in gene expression or nutritional transcriptomics [12]. Expanding the use of these evolving technologies will allow the identification of molecular sites of action of dietary Se and how these targets bring about a phenotypic change. Nutrigenetics Genetic polymorphisms may be partially responsible for variations in individual responses to Se. Individuals differ substantially in their ability to increase selenoprotein activity in response to additional dietary Se [13]. This inter-individual variation in selenoprotein levels may be a result of single nucleotide polymorphisms (SNPs) in genes encoding selenoproteins and may determine the efficacy with which individuals can incorporate Se into selenoproteins [13-16]. Furthermore, polymorphisms associated with cancer risk may indicate that these genes are involved directly with cancer etiology. Polymorphisms in selenoproteins, including glutathione peroxidase 1 (GPXl), glutathione peroxidase 4 (GPX4), selenoprotein P (SelP) and the 15 kDa selenoprotein (SEP 15), have been associated with increased cancer risk. The role of genetic variants of these proteins and cancer risk will be reviewed in Chapter 24 and will not be addressed here. It is possible that other polymorphisms including those associated with drug metabolism mediate the response to Se. One of the most important detoxification systems is the glutathione-5'-transferase (GST) family of enzymes. These enzymes are expressed in a wide variety of human tissues, including both normal and malignant tissues. The human GSTs consist of four main classesalpha (A), mu (M), pi (P) and theta (T) each of which is divided into one or more isoforms. Functional polymorphisms are known for the GST genes Ml, PI and Tl and they all lead to less active enzymes compared to the wild-type gene products [17]. GST has long been recognized for its involvement in Se metabolism and elimination [18]. In studies by Chen et al [19], plasma Se concentrations were inversely associated with aflatoxin Bj-albumin adduct levels, a biomarker for liver cancer susceptibiUty, only among individuals with the null genotype for GSTMl and GSTTl. Definitive information is needed whether individuals with these polymorphisms have increased concentrations of selected selenocompounds because of alterations in Se metabolism. Oxidative stress and mitochondrial DNA damage play important roles in carcinogenesis. Manganese-dependent superoxide dismutase (MnSOD) is a major enzyme that is responsive for the detoxification of reactive oxygen species in the mitochondria. This enzyme converts reactive oxygen species

370

Selenium: Its molecular biology and role in human health

to oxygen and hydrogen peroxide, and the latter is catalyzed into water by either catalase or the selenoprotein, glutathione peroxidase (GPXl). A polymorphism replacing valine (V) with alanine (A) at codon 16 in the mitochondrial targeting sequence of the human MnSOD gene alters the protein secondary structure and lowers transport of MnSOD into the mitochondria [20]. In the Physicians' Health study, men with the AA genotype for MnSOD were particularly sensitive to antioxidant status, with a 5-fold gradient in risk for aggressive prostate cancer among the four quartiles of plasma Se concentrations [21]. In contrast, differences in Se concentrations did not modify cancer risk in individuals with the V allele [21]. Similarly, women in the Shanghai Breast Cancer Study with a low Se intake had increased breast cancer risk with the AA genotype but not with the VA or VV genotype [22]. Because polymorphisms in both MnSOD and GPXl have been linked to cancer susceptibility and both are involved in the detoxification of reactive oxygen species, studies on the joint effects of the MnSOD and GPXl gene polymorphisms might provide further information on the role played by oxidative stress and the potential modification by dietary Se in cancer risk. Nutritional Epigenetics In addition to genomics, DNA methylation is an important epigenetic mechanism of transcriptional control. DNA methylation plays an essential role in maintaining cellular function, and changes in methylation patterns may contribute to the development of cancer. During tumor progression, the DNA becomes paradoxically hypomethylated, despite the presence of regional hypermethylation and an increase in DNA methyltransferase activity [23]. Global hypomethylation can result in chromosome instability and hypermethylation of promoter regions has been associated with transcriptional silencing. In fact, hypermethylation of promoter regions, is at least as common as DNA mutations as a mechanism for inactivation of classical tumor suppressor genes in human cancers [23,24]. Furthermore, a number of candidate tumor suppressor genes that are not commonly inactivated by mutation are transcriptionally silenced by this mechanism [24]. Preclinical and clinical studies suggest that part of the cancer protective effects associated with several bioactive food components may relate to DNA methylation patterns [23,26]. Se deficiency has been shown to decrease DNA methylation in Caco-2 cells and in rat liver and colon [27]. Thus, alterations in DNA methylation may be a potential mechanism whereby deficient dietary Se increases liver and colon tumorigenesis. However, supplemental Se has been found to decrease DNA methyltransferase protein expression and activity in vitro [27], but increase liver and colon DNA methyltransferase activity in vivo [28]. The explanation for these paradoxical results is unknown but may relate to

Incorporating 'omics' approaches

371

selenite binding to thiol groups and precipitating proteins in vitro. Nevertheless, future studies are warranted to investigate the effects of dietary Se on global and gene specific DNA methylation, DNA methyltransferase activity and cancer susceptibility, to better elucidate the role of alterations in DNA methylation on the cancer protective effects of dietary Se. Nutritional Transcriptomics Another important site of regulation is the rate of transcription of various genes. Subtle changes in gene expression, even at the single-cell level, can be measured by quantitative techniques such as real-time PCR and high density microarray analysis. Microarrays make it possible to assess the effect of Se on the expression of a large proportion of the whole genome in a high-throughput manner. Microarray analysis has been utilized to identify molecular targets of Se, the signaling pathways modified by Se, and the down-stream effects of Se in cell culture models. These studies demonstrate that Se can modify many and varied types of molecular targets. In the premalignant MCFIOAT human breast cancer cell line, a 200-gene membrane-based cDNA array was used to investigate the effect of Se on expression of genes associated with apoptosis and cell cycle regulation [29]. Se targets included GADD153, cyclin A, CDKl, CDK2, CDK4, CDC25, E2Fs, as well as the MAPK/JNK and phosphoinositide-3 kinase pathways. There is considerable overlap of the Se-modulated genes or signaling pathways identified in breast cancer cells and prostate cancer cells. Dong et al [30] utilized oligonucleotide array technology to gain insight into the gene expression changes that might influence the regulation of human prostate cancer cell growth by dietary Se. Of a total of 12,000 genes screened, over 2,500 genes were identified as responsive to Se treatment. Because of the large number of genes whose expression was modified by Se, the data were analyzed by cluster analysis in order to group genes according to similarities in their expression profiles across multiple time points. These genes fell into 12 clusters of distinct kinetic patterns of modulation by Se [30]. The response to Se is likely amplified by simultaneously influencing a multitude of targets. Microarray technology has also been utilized to examine the molecular events in animal models for studying cancer. Mice fed a Se deficient diet exhibited increased expression of genes involved in DNA damage processing, oxidative stress and cell-cycle control and decreased expression of genes involved in detoxification [31]. Supplementation with synthetic organoselenium compounds (p-XSC or />-XSeG) revealed reduced expression of Phase I enzymes, inflammatory mediators and cell cycle regulatory genes but increased expression of Phase II enzymes and apoptosis related genes [32]. Thus, the response to Se can be monitored in vivo and

372

Selenium: Its molecular biology and role in human health

should help in the interpretation of the molecular targets as well as candidate biomarkers of Se during intervention trials. Comparison of gene expression changes across species may be useful for prioritizing genes for biomarker validation. Comparative genomics has been utilized to compare Se-induced gene expression profiles between rat and human prostate cancer cells [33]. A subset of 154 genes was found to demonstrate a similar level of differential gene expression in response to Se treatment in both the human (PC3) and rat (PAII) prostate cancer cells lines. Data mining identified insulin like growth factor binding protein 3 (IGFBP3) and retinoic X receptor alpha (RXRalpha) as genes that were both upregulated with respect to Se and down-regulated during tumorigenesis [33]. Thus altered expression of IGFBP3 and RXRalpha could contribute to the mechanism of Se's protective effects against prostate cancer. Such evidence points to the need to examine them as variables in ongoing Se and prostate cancer prevention clinical trials. Transcriptomics has also been utilized to investigate the molecular changes that occur during tumor progression. These types of studies have demonstrated the importance of selenoproteins during the carcinogenic process. For example, Sep 15 is expressed at high levels in the normal liver and prostate, but at reduced levels in the corresponding malignant organs [16]. A series of prostate cancer cell lines were isolated at different stages of tumorigenesis and analyzed for alterations in gene expression profiles utilizing microarrays [34]. SelP was identified as a gene that correlated with tumorigenesis. Studies utilizing quantitative real-time reverse transcriptionPCR for SelP have revealed a similar down-regulation of the transcript of this gene in a subset of human prostate tumors, mouse tumors and prostate carcinoma cell lines [34]. Thus altered expression of SelP during prostate tumor progression likely occurs in many different species including humans. However, whether or not modulation of SelP occurs in other cancer sites remains unknown. Another approach that can help evaluate global changes in gene expression is subtractive hybridization. Suppression subtractive hybridization was used to identify genes differentially expressed in malignant mesothelioma cells compared to normal mesothelial cells [35]. SEP15, was isolated using this approach and subsequently shown to be down regulated in -60% of malignant mesothelioma cell lines and tumor specimens. Interestingly, malignant mesothelioma cells with down regulated Sep 15 or the 1125A variant were less responsive to the growth inhibitory and apoptotic effects of Se than malignant mesothelioma cells expressing the wild-type protein [35]. RNAi knockdown studies demonstrated that Sep 15 inhibition makes sensitive malignant mesothelioma cells more resistant to Se [35]. Collectively evidence suggests Sep 15 is another selenoprotein that may be involved in cancer etiology.

Incorporating 'omics' approaches

373

Proteomics and Metabolomics Several dietary components, such as Se, can influence enzymatic activity by modifying the translation of RNA to proteins, as well as posttranslational modifications. Variation in the proteome (all of the proteins produced by a species much as the genome is the entire set of genes) can have marked effects on the cellular phenotype. Proteomic analysis allows a point in time comparison of protein expression after a dietary intervention. Protein expression is not always correlated with mRNA expression. Several factors may account for the inconsistency between mRNA and protein expression. For example, alternative splicing may produce multiple proteins from a single gene. Post-franslational modification (i.e. glycosylation, phosphorylation, oxidation, reduction) and/or degradation may also generate multiple protein products originating from a single gene or a single transcript. These modified proteins often vary in their biological activities. Dietary Se has been shown to alter post-translational modifications of proteins and thus influence their activity. For example, methylseleninic (MSA) acid induces apoptosis in DU145 prostate cancer cells [36]. This induction was related to alterations in extracellular signal-regulated kinase (ERK) activity [36]. Western blot analysis revealed that MSA exposure did not affect ERK protein content per se but increased its phosphorylation [36]. Another Se-induced post-franslational regulation involves protein kinase C (PKC). PKC activity is dependent on the presence of reduced cysteine residues on the protein and methyl selenol is capable of oxidizing these residues thus down-regulating PKC activity [37]. Interestingly, reduction of the cysteine residues on PKC is accomplished by reduced thioredoxin, and thioredoxin reduction is reduced by the seloprotein thioredoxin reductase. Proteomic analysis is a powerfiil technology used to comprehensively inspect protein expression in bodily fluids, tissues and cells. Changes in serum proteomic patterns were evaluated in patients with clinically localized prostate cancer randomized to take Se, vitamin E, both or placebo (formulations identical to those in SELECT, the large Se and vitamin E prostate cancer prevention trial currently being conducted) for 3 to 6 weeks preprostatectomy [38]. Sera were collected from patients before and after dietary supplementation and age-matched, disease-free men served as confrols. Mass profiling of lipophilic serum proteins was conducted and mass specfra were analyzed using custom-designed software. The results indicated that in sera from patients with prostate cancer, Se and vitamin E combined, but neither alone, induced statistically significant proteomic changes associated with prostate cancer-free status [38]. Therefore, Se and vitamin E may have beneficial effects in individuals that already have prostate cancer.

374

Selenium: Its molecular biology and role in human health

Proteins whose expressions are altered during carcinogenesis can be characterized by comparing the proteomic profiles between a healthy or control sample and a malignant sample. Proteomic analysis reveals that Sebinding protein 1 (SBP-1) may be important in cancer progression. Although the mechanism by which Se binds to this protein and the exact function of SBP-1 are unknown, low levels of SBP-1 protein have been associated with lung cancer [39] and gastric adenocarcinoma [40]. Two-dimensional polyacrylamide gel electrophoresis (2-D PAGE) was used to identify proteins that were differentially expressed in squamous cell lung carcinoma, adenocarcinoma, large cell carcinoma and basaloid carcinoma [41]. SBP-1 protein expression was down-regulated in all lung cancer samples compared to non-tumorous lung tissue [41]. Furthermore, low levels of SBPl protein were observed in tumors from patients with poor survival. When comparing subgroups of tumors, both SBPl protein and mRNA levels were significantly decreased in poorly differentiated, T2-T4, and bronchus-derived tumors, indicating that the down-regulated expression of SBPl occurs in larger and potentially more aggressive lung adenocarcinomas [39]. Future studies investigating the function of this protein, whether it is a selenoprotein, and the basis for its reduced expression in cancer are warranted. Chronic activation of peroxisome prolifeator-activated receptor alpha (PPARa) is important in the induction of cell-specific responses, including the development of liver tumors. Alterations in protein expression patterns in livers of mice with chronic PPARa activation were delineated by using a proteomics approach [42]. An 18-fold decrease in Se-binding protein 2 accompanied PPARa stimulation [42]. These results suggest cross-talk between these proteins. Overall, proteomic analysis has identified several proteins, including selenoproteins, are linked to the carcinogenic process. One of the newest "omics" is metabolomics, which refers to the dose and temporal changes in cellular small molecular weight compounds which can be altered in response to dietary or drug treatments. Previous studies have shown that dietary Se can modulate small molecular weight compounds such as ketones [43] and fatty acids [44]. However, no studies have used a metabolomic approach to evaluate all of the biochemical changes following Se supplementation. It is possible that metabolomic analysis would allow for a better understanding of the mechanisms for the cancer protective effects of Se and for the identification of individuals, based on their metabolic abilities, who would benefit from Se for cancer prevention. Other Evolving Approaches Recent transgenic and knockout animal models are providing powerful insights that can help discern the mechanisms of action of bioactive food components. Many knockout models have been developed to investigate phenotypic changes when one or two selenoproteins are deleted. Deletion of

Incorporating 'omics' approaches

375

the TR2 gene [45] or GPX4 gene [46,47] is embryonic lethal. In contrast, SelP knockout mice have altered distribution of Se and respond differently to Se supplementation than wild-type mice [48,49]. Studies of GPXl knockout mice demonstrate that GPXl functions against acute oxidative stress [50] and studies with GPXl and GPX2 double knockout animals demonstrate that these mice are highly susceptible to bacteria-associated inflammation and cancer [51]. Thus, the GPX family of proteins probably does play a role in protecting against cancer. However, one limitation with this type of approach is that they only provide information about effects relative to an individual selenoprotein. Selenoprotein transgenic and conditional knockout models currently in development in mice will help clarify the role of Se and selenoproteins in cancer risk and prevention. For example, Moustafa et al [52] developed a transgenic model to assess the roles of selenoproteins. They generated a transgenic mouse line that carries a mutation at position 37 within the anticodon loop of Sec tRNA^^"^^". Normally, position 37 contains an adenosine that is modified to a much larger base, isopentenyladenosine (i^A). Transgenic mice that express Sec tRNA^^*''^*^ and lack i^A manifest a reduction in the level of numerous selenoproteins [52]. The reduction in selenoprotein synthesis occurs in a protein- and tissue-specific manner, whereby GPXl and TR3 are the most- and least-affected selenoproteins, respectively, and liver and testes are the most- and least-affected tissues, respectively, of those examined [52]. Furthermore, this mutation causes a reduction in one of the species of tRNA'^"^^^'', which is thought to be critical for stress-selenoprotein synthesis [53]. Because the levels of numerous selenoproteins are reduced, the i*A-deficient mouse provides a model system to determine the role of selenoproteins in cancer prevention and modulation of selenoprotein expression by dietary Se. hi fact, i^A-defiicient mice are more susceptible to carcinogen-induced aberrant crypt formation, a preneoplastic lesion for colon cancer, than wild-type mice (R Irons, BA Carlson, DL Hatfield, CD Davis, submitted). Furthermore, Se supplementation decreases aberrant crypt formation in these mice. These results suggest that alterations in selenoprotein synthesis do increase colon cancer susceptibility. It is known that removal of the Sec tRNA'^'^^'' gene by developing a knockout mouse is embryonic lethal [54,55]. One means of examining the roles of genes that are embryonic lethal is gene silencing on a conditional basis using Cre/loxP technology to create a tissue-specific conditional knockout of the gene. For example, when the Sec tRNA^^"^^" gene is flanked by loxP sites, the gene can be excised with the Cre recombinase. The Cre recombinase gene can then be expressed under the control of a tissue-specific promoter, which induces site-specific recombination between the loxP sites. The Cre/loxP system has been utilized to develop conditional knockouts of

376

Selenium: Its molecular biology and role in human health

the Sec tRNA'^"^^*" gene in the mammary epithehum [54] and liver [56]. In fact, expression of the tumor suppressor genes BRCAl and p53 was altered in the mammary epithelium of mice containing a conditional knockout of the Sec tRNA^^*'^'^"' gene, suggesting greater susceptibility to cancer [54]. Because these models allow the selective manipulation of Sec tRNA levels and selenoprotein expression, they provide the tools to conduct functional studies on the role of selenoproteins in cancer prevention. Future studies are needed to determine whether these mice respond differently to dietary Se, whether they are more sensitive to chemical carcinogens, and if so, can dietary Se compensate for the increased cancer susceptibility? Conclusions While there is intriguing evidence that Se is protective against cancer, evolving 'omics' approaches will help elucidate the role of Se and selenoproteins in cancer prevention, as well as the identification of those individuals who will or will not benefit fi-om Se supplementation. The identification of functional polymorphisms in the genes for several selenoproteins and the association of polymorphisms with cancer risk suggest that selenoproteins may mediate the beneficial effects of Se. In addition, allelic loss of selenoprotein genes during tumor development supports the possibility that deletion of these genes promotes carcinogenesis. Selenoprotein transgenic and conditional knockout mice will help clarify the importance of selenoproteins and/or small molecular weight selenocompounds in mediating the cancer protective effects of Se. Collectively, it is clear that Se can influence a number of key molecular events that are involved in cancer prevention. As the era of molecular nutrition unfolds, a greater understanding of how Se influences cancer will surely arise. Such information will be critical in the development of effective tailored strategies for reducing the cancer burden. References 1. 2. 3. 4.

5. 6. 7. 8. 9.

CD Davis, R Irons 2005 Current Reviews Nutrition Food Science (in press) CD Davis, JW Finley 2003 Functional Foods and Nutraceuticals in Cancer Prevention Iowa State Press, Iowa pp 55 GN Schrauzer, DA White, CI Schneider 1997 Bioinorg Chem 7:23 World Cancer Research Fund/American Institute for Cancer Research Food, Nutrition and the Prevention of Cancer: A Global Perspective American Institute for Cancer Research, Washington, DC H Zhuo, AH Smith, C Steinmaus 2004 Cancer Epidemiol Biomarkers Prev 13:771 EA Klein 2004 J Urol 171 :S50 H Li, MJ Stampfer, EL Giovannucci, JS Morris, WC Willett, JM Gaziano, J Ma 2004 J Natl Cancer Inst 96:696 LC Clark, GF Combs, BW Tumbull, EH Slate, DK Chalker et al 1996 JAMA 276:1957 AJ DufField-Lillico, ME Reid, BW Tumbull, GF Combs Jr, EH Slate, LA Fischbach, JR Marshall, LC Clark 2002 Cancer Epidemiol Biomarkers Prev 11:630

Incorporating 'omics' approaches 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45.

'hll

ME Reid, AJ Duffield-Lillico, L Garland, BW Tumbull, LC Clark, JR Marshall 2002 Cancer Epidemiol Biomarkers Prev 11:1285 AJ Duffield-Lillico, BL Dalkin, ME Reid, BW Tumbull, EH Slate et al Nutritional Prevention of Cancer Study Group 2003 BJUInt 91:608 CD Davis, JA Milner 2004 Mutation Res 551:51 KM Brown, K Pickard, F Nicol, GJ Beckett, GG Duthie, JR Arthur 2000 Clin Sci 98:593 YJ Hu, Diamond AM 2003 Cancer Res 63:3347 V Diwadkar-Navsariwala, AM Diamond 2004 JNutr 134:2899 E Kumarasv*ramy, A Malykh, KV Korotkov, S Kozyavkin, Y Hu et al 2000 J Biol Chem 275:35540 MJ Grubben, FM Nagengast, MB Katan, WH Peters 2002 Scand J Gastroenterol Suppl 234:68 MJ Christensen, BL Nelson, CD Wray 1994 Biochem Biophys Res Commun 202:271 SY Chen, CJ Chen, WY Tsai, H Ahsan, TY Liu, JT Lin, RM Santella 2000 Nutr Cancer 38:179 S Shimoda-Matsubayashi, H Matsumine, T Kobayashi, Y Nakagawa-Hattori, Y Shimizu, Y Mizuno 1996 Biochem Biophys Res Commun 226:561 H Li, PW Kantoff, E Giovannucci, MF Leitzmann, JM Gaziano, MJ Stampfer, J Li 2005 Cancer Res 65:1A9% Q Cai, XO Shu, W Wen, JR Cheng, Q Dai, YT Gao, W Zheng 2004 Breast Cancer Res 6:R647 CD Davis, EO Uthus 2004 Exp Biol Med 229:988 JW Miller, MR Nadequ, J Smith, D Smith, J Selhub 1994 Biochem J 298:415 SF DeCabo, J Santos, J Femadex-Piqueras 1995 Cytogenet Cell Genet 71:187 SA Ross 2003 Ann N Y Acad Sci. 983:197 CD Davis, EO Uthus 2002 JNutr 132:292 CD Davis, EO Uthus 2003 JNutr 133:2907 Y Dong, HE Ganther, C Stewart, C Ip 2002 Cancer Res 62:708 Y Dong, H Zhang, L Hawthorn, HE Ganther, C Ip 2003 Cancer Res 63:52 L Rao, B Puschner, TA Prolla2001 JNutr 131:3175 K El-Bayoumy, BA Narayanan, DH Desai, NK Nayayanan, B Pittman, SG Amin, J Schwartz, DW Nixon 2003 Carcinogenesis 24: 1505 M Schlicht, B Matysiak, T Brodzeller, X Wen, H Liu et al 2004 BMC Genomics 5:58 A Calvo, N Xiao, J Kang, CJM Best, I Leiva, MR Emmert-Buck, C Jorcyk, JE Green 2002 Cancer Res 62:5325 S Apostolou, JO Klein, Y Mitsuuchi, JN Shetler, PI Poulikakos, SC Jhanwar, WD Kruger, JR Testa 2004 Oncogene 23:5032 H Hu, C Jiang, G Li, J Lu 2005 Carcinogenesis 26:1374 R Gopalakrishna, Y Gundimeda 2001 Nutr Cancer 40:55 J Kim, P Sun, Y Lam, P Troncoso et al 2005 Cancer Epidemiol Biomarkers Prev 14:1697 G Chen, H Wang, CT Miller, DG Thomas, TG Gharib et al 2004 J Pathol 202:321 QY He, YH Cheung, SY Leung, ST Yuen, KM Chu, JF Chiu 2004 Proteomics 4:3276 LS Li, H Kim, H Rhee, SH Kim, DH Shin, KY Chung, KS Park, YK Paik, J Chang, H Kim 2004 Proteomics 4:3394 R Chu, H Lim, L Brumfield, H Liu, C Herring, P Ulintz, JK Reddy, M Davison 2004 A/o/Ce//5w/24:6288 M Yoshida 1991 y Nutr Sci Vitaminol 4:425 ML Dodge, RC Wander, Y Xia, JA Butler, PD Whanger 1999 J Trace Elem Med Biol 12:221 M Conrad, C Jakupoglu, SG Moreno, S Lippl, A Banjac et al 2004 Mol Cell Biol 24:9414

378 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56.

Selenium: Its molecular biology and role in human health H Imai, F Hirao, T Sakamoto, K Sekine, Y Mizukura, M Saito, T Kitamoto, M Hayasaka, K Hanaoka, Y Nakagawa 2003 Biochem Biophy Res Commun 305:278 LJ Yant, Q Ran, L Rao, H Van Remmen, T Shibatani, JG Belter, L Motta, A Richardson, TA Prolla 2003 Free Radic Biol Med 34:496 KE Hill, J Zhou, WJ McMahan, AK Motley, JF Atkins, RF Gesteland, RF Burk 2003 J Biol Chem 278:13640 L Schomburg, U Schweizer, B Holtmann, L Flohe, M Sendtner, J Kohrle 2003 Biochem J 370:397 JB De Haan, PJ Crack, N Flentjar, RC lannello, PJ Hertzog, I Kola 2003 Redox Rep 8:69 FF Chu, RS Esworthy, PG Chu, JA Longmate, MM Huycke, S Wilczynski, JH Doroshow 2004 Cancer Res 64:962 ME Moustafa, BA Carlson, MA El-Saadani, GV Kryukov, QA Sun et al 2001 Mol Cell 5(0/21:3840 BA Carlson, X-M Xu, VN Gladyshev, DL HatfiM J Biol Chem 280:5542 E Kumaraswamy, BA Carlson, F Morgan, K Miyoshi, GW Robinson et al 2003 Mol Cell Biol 23:Un MR Bosl, K Takaku, M Oshima, S Nishimura, MM Taketo 1997 Proc Natl Acad Sci USA 94:5531 BA Carlson, SV Novoselov, E Kumaraswamy, BJ Lee, MR Anver, VN Gladyshev, DL Hatfield 2004 J Biol Chem 279:8011

Chapter 33. Selenium-induced apoptosis Ick Young Kim, Tae Soo Kim and Youn Wook Chung Laboratory of Cellular and Molecular Biochemistry, School of Life Sciences and Biotechnology, Korea University, 1,5-Ka, Anam-Dong, Sungbuk-Ku, Seoul 136-701, Korea

Daewon Jeong BK21 HLS, Seoul National University, 28 Yeonkun-Dong, Chongno-Ku, Seoul 110-749, Korea

Summary: Selenium is both an essential and a toxic trace element. It has been reported to induce cell growth and cell proliferation, but also cell death by necrosis or apoptosis. The biological action of selenium is dependent on both its specific chemical form and concentration. At high levels, selenium induces oxidation and cross-linking of protein thiol groups and generation of reactive oxygen species, ultimately leading to cell death. Lideed, a shift to a more oxidizing environment induced by selenium is thought to be largely responsible for selenium-induced apoptotic cell death. Selenium compounds capable of oxidizing thiol groups and generating the superoxide anion (O2*") also trigger loss of the mitochondrial function and release of proapoptotic proteins, such as cytochrome c, from mitochondria into the cytosol. In addition, some selenium compounds activate caspase-3, which in turn contributes to morphological changes in the cell and DNA fragmentation characteristic of the late stage of apoptosis. The serial events of seleniuminduced apoptosis are thus thought to include the generation of an oxidizing intracellular environment followed by induction of mitochondrial dysfunction, cytochrome c release, caspase-3 activation, and DNA fragmentation. Introduction The major source of selenium in mammals is from the diet, and thus, the selenium status of an organism is reflected in the level of selenium in the environment [1]. The bioavailability and toxicity of selenium depend on its chemical form and concentration [2-4], with organic forms being more bioavailable and less toxic than inorganic forms [5,6]. Selenium toxicity was first described in livestock in the 1930s and was attributed to the consumption of selenium-accumulating plants over a long period [7]. Selenium poisoning due to consumption of inappropriate amounts of selenium in the diet causes blind staggers in cattle, alkali disease in horses

380

Selenium: Its molecular biology and role in human health

and cattle [8], poliomyelomalacia in pigs [9], and neurological disorders in humans [10]. These toxicities are thought to result from the ability of selenium to induce oxidative stress [7,11]- In this chapter, we discuss the mechanism of apoptosis that results from the induction of oxidative stress by selenium. Selenium toxicity Selenium circulates through the food chain, and its distribution in the environment, including water and food sources, depends on its concentration in soil [12]. The geographic distribution of selenium in soil varies greatly, with volcanic regions commonly possessing selenium-deficient soil and selenium-rich regions including the Western part of the United States, Ireland, Israel, and the Northern part of Australia [8,11,13]. Selenium-tolerant plants of the genuses Astragalus, Xylorrhiza, Oonopsis, and Stanleya contain selenium at levels (5 to 50 mg per kilogram of mass) that are 100 to 1000 times those in plants that do not accumulate this element [13]. Chronic selenosis can occur in livestock and humans as a result of long-term exposure to selenium levels in the milligram per kilogram of diet [8,11,1416]. Cattle that graze in regions with selenium-rich soil develop alkali disease or blind staggers, the former of which is characterized by lack of vitality, anemia, stifl&iess, deformed and sloughed hooves, and lameness, and the latter by weight loss, blindness, ataxia, disorientation, respiratory distress, and neurological symptoms [13]. The symptoms of chronic selenium poisoning in humans include loss of skin and nails, skin lesions, tooth decay, and nervous system disorders [10]. Safe levels of selenium exposure There is limited information about the safe levels of exposure to different chemical forms of selenium, including inorganic forms, such as selenite and selenate, as well as organic forms, such as selenomethionine and selenocysteine. However, the toxic effects of selenium depend on its chemical form [17]. The toxicity of selenium is also determined by its ability to penefrate cells, the ability of cells to remove this element once it has entered, the activity of a pathway for detoxification by methylation, the efficiency of bioconversion of inorganic forms to organic forms, cell type, as well as route, frequency, and duration of exposure. Given these many determinants of selenium toxicity, it is difficult to set safe levels for exposure to this essential element. In addition, rapidly dividing cells with a high level of infracellular thiols manifest an increased sensitivity to selenium [18]. In general, mammals with an intake of a milligram of selenium per kilogram of body mass per day and cultured cells freated with selenium at micromolar concenfration develop selenium toxicity [14,17,19,20]. On the basis of several recent studies, it has been

Selenium-induced apoptosis

381

recommended that the daily intake of selenium in adult humans should be ~100 )j,g and should not exceed 1 mg [21]. However, the current recommended daily allowance is discussed in detail in Chapter 35. Selenium-induced oxidative stress Selenium adversely affects cellular redox status by generating oxidative radicals and oxidizing thiols. In vitro, selenium compounds, such as selenite, selenium dioxide, diselenides, selenocystine, and selenocystamine, react with thiols, such as reduced glutathione (GSH) and L-cysteine and thereby generate the superoxide anion (02*~) [11,22-24]. Exposure of isolated mitochondria to certain selenium compounds, including selenite, selenocystine, selenocystamine, and selenodioxide, to GSH, or to both selenium compounds and GSH has also been shown to result in a marked increase in superoxide generation [25,26], as has the addition of exogenous GSH to cultured cells [23] or treatment of cells with selenite. Moreover, exposure of LNCaP or LAPC-4 human prostate cancer cells to selenite resulted in a decrease in intracellular GSH content and an increase in the amount of oxidized glutathione (GSSG). In addition, selenite toxicity in rat embryos was greatly reduced by depletion of endogenous GSH as a result of treatment with L-buthionine-[5',i?]-sulfoximine [27], which is a specific inhibitor of GSH synthesis. These various observations indicate that certain selenium compounds react with intracellular GSH to produce superoxide. SeSOj^f 4GSH->. I

H2O2 + Fe2+

• Fe^^ + HO* + HO^ "

GSSG^^T

GS-Se-SG GSH->. ^ ^ I GSSG-*-^! J^^ GS-SeH

•

r\

\ / A^Fe2+ + 2H+ O2 O^-^ Y^ ONOO^ --?--• - ^ NO,' + HO' NOI\ ^ f H+ H,Se • SeiJE

GSH GSSG Figure 1. Generation of ROS by reductive reaction of selenite with GSH. Reaction of selenite with GSH generates O2*", which either undergoes dismutation in a reaction catalyzed by superoxide dismutase (SOD) to produce H2O2 or liberates Fe^* from iron-sulfur clusters or Fe from ferritin, generating the highly reactive species HO' and HO" via the Fenton reaction [29]. H2O2 also reacts with Fe"^ to form 02*". Nitric oxide (NO*) reacts with O2*" to yield the toxic radicals NO2* and HO* via the intermediate ONOOH [30]. Thesefreeradicals can attack protein thiols [31,32] and DNA [33] and thereby induce apoptosis. Such reactions generate a more oxidizing environment in the cell and likely account for selenium toxicity.

Superoxide generated from reductive reaction of selenium with GSH is able to react with oxygen to produce additional reactive oxygen species

382

Selenium: Its molecular biology and role in human health

(ROS), including H2O2, HO', HO , ONOO", and NO2' (Figure 1). Moreover, these cascading reactions lead to GSH depletion as a result of export of cellular GSSG by a transporter. The inhibition of selenium-induced generation of ROS by antioxidants, including catalase, superoxide dismutase (SOD), and deferroxamine, results in inhibition of selenium-induced apoptosis [22]. These findings support the hypothesis that a low redox status induced by selenium plays an important role in selenium toxicity. Various mechanisms, including detoxification by methylation [28] and antioxidantbased defenses, operate in mammals and plants to combat such toxicity. Molecules targeted for selenium-induced thiol modification Thiol groups located at or near the active site or other domains of certain proteins play an essential role in enzymatic activity or in other functions such as the DNA binding activity of transcription factors such as N F - K B and AP-1. Selenium impairs the functions of such proteins by reacting with these essential thiol groups to form S-Se-S (selenotrisulfide) or S-Se adducts (Figure 2) [34]. In addition to N F - K B and AP-1, proteins shown to be targeted for such thiol modification by selenium include the Na^- and K^dependent ATPase, the glucocorticoid receptor, prostaglandin D synthase, lipoxygenase, glycerol-3-phosphate dehydrogenase, the protease caspase-3, and the protein kinases Cdk2, protein kinase C, and JNK [35-45]. Some of these proteins, including AP-1, N F - K B , caspase-3, Cdk2, protein kinase C, and JNK, are redox-regulated signaling molecules. Inhibition by selenium of signaling mediated by such proteins therefore likely affects various signaling pathways, including those that regulate apoptosis. Selenium-induced mitochondrial dysfunction Apoptosis is a genetically programmed type of cell death for unwanted cells and plays a critical role in embryonic development, immune regulation, and tumor regression. Two major apoptotic signaling pathways have been identified: the death receptor-mediated (extrinsic) pathway and the mitochondrial (intrinsic) pathway [46,47]. The extrinsic pathway of apoptosis is initiated by interaction of a ligand with its transmembrane death receptor (such as CD95, tumor necrosis factor receptor, and TRAIL receptor) and the consequent activation of membrane-proximal caspases (caspase-8 or -10). The intrinsic pathway is initiated by mitochondrial outer membrane permeabilization (MOMP), which results in the release into the cytosol of proteins normally localized in the space between the inner and outer mitochondrial membranes (including cj^ochrome c, procaspase-2, -3, and -9, as well as apoptosis-inducing factor). MOMP is promoted by the oxidation of protein thiol groups or an increase in the cytosolic concentration of Ca^^. Exposure of mitochondria isolated fi"om rat liver to selenite was found to induce the oxidation and cross-linking of protein thiol groups, a decrease in

Selenium-induced apoptosis

383

A^m, mitochondrial swelling, and the release of cytochrome c [25,26]. Selenite-induced MOMP was inhibited by the Ca^^ chelator EGTA and by cyclosporine A, a blocker of the mitochondrial permeability transition pore. iV-Ethylmaleimide (NEM), a monofunctional thiol oxidant that forms a stable thioether-bonded adduct, also prevented both the protein aggregation induced by selenite-dependent thiol cross-linking as well as mitochondrial swelling. In addition, dithiothreitol reversed the protein aggregation induced by selenite. These fmdings thus indicate that selenite-induced mitochondrial swelling is mediated by cross-linking of thiol groups (Figure 2) [48].

-s

?

1

1 -HS

-S-Se-S-

hf)

1 1 1 1 1 1 1 1 1 1 S

-SH

^

-SH

»,^

1 Ml

-SH

-SH

m-

1

,

^

u

%

U

-S-Set-

u

u

-SH 1

-S-NEM Figure 2. Reactions of selenium compounds witli protein thiol groups and serial events in selenium-induced apoptosis. Selenium reacts with thiol groups to generate an intramolecular disulfide (1), a selenotrisulfide (S-Se-S) (2), or a selenenylsulfide (S-Se) (3). Seleniuminduced apoptosis is mediated by such thiol reactions and consequent activation of the indicated mitochondrial pathway of cell death.

Numerous studies have provided the basis for the conclusion that selenium compounds are able to induce apoptosis. Inhibition of cell growth, loss of cell viability, or DNAfragmentationinduced by selenium compounds has been demonstrated in human leukemia HL-60 cells [49], human hepatic carcinoma HepG2 cells [50], human colonic carcinoma HT29 and SW480 cells [51], and human glioma A172 and T98G cells [52]. Exposure of HL-60 cells to methylselenocysteine resulted in a dose-dependent increase in the amount of intracellular ROS and subsequent apoptosis [53]. This apoptotic effect of methylselenocysteine was blocked by GSH, the GSH precursor A''acetylcysteine, or deferroxamine. Furthermore, treatment of HepG2 cells with selenite induced an increase in the amount of O2*" and serial apoptotic events including a change in ATm, cytochrome c release, caspase-3 activation.

384

Selenium: Its molecular biology and role in human health

and DNA fragmentation [54]. The accumulation of 02*~, the change in ASPm, and downstream apoptotic events induced by selenite were suppressed by A^acetylcysteine, the antioxidant 4-hydroxy-2,2,6,6-tetramethylpiperidine-iVoxyl, or the MOMP inhibitors cyclosporine A or trifluoperazine [54]. Various additional studies have also indicated that selenium-induced apoptosis is executed through a shift in cellular redox status, mitochondrial dysfunction, caspase-3 activation, and DNA fragmentation [55-58]. Conclusion In summary, selenium acts as a pro-oxidant and renders the intracellular environment more oxidizing by inducing the generation of ROS through reductive reaction with GSH, oxidizing thiols directly and indirectly via ROS, reducing the GSH/GSSG ratio, and depleting intracellular GSH. The oxidative stress induced by selenium is thought to be primarily responsible for selenium-induced apoptosis. The observations described in this chapter thus suggest that the pathway of selenium-induced apoptosis involves the induction of a shift in intracellular redox status, the cross-linking of thiol groups of mitochondrial proteins, a change in A^Fm, mitochondrial swelling, the release of cytochrome c from mitochondria into the cytosol, the activation of caspase-3, and DNA fi"agmentation. References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19.

K Wada 1985 Metal and Human: ecotoxicology and clinical medicine Asakura Shoten Tokyo pp 134 JC King 2001 yiV«
Selenium-induced apoptosis 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58.

385

GD Frenkel, A Walcott, C Middleton 1987 Mol Pharmacol 31:112 MP Rayman 2000 Lancet 356:233 HM Shen, CF Yang, CN Ong 1999 Int J Cancer 81:820 L Yan, JE Spallholz 1993 Biochem Pharmacol 45:429 MS Stewart, JESpallholz, KH Neldner, BC Pence 1999 Free Radic Biol Med 26:42 TS Kim, DW Jeong, BY Yun, lY Kim 2002 Biochem Biophys Res Commm 294:1130 TS Kim, BY Yun, lY Kim 2003 Biochem Pharmacol 66:2301 M Usami, H Tabata, Y Ohno 1999 Teratog Carcinog Mutagen 19:257 C Ip, HE Ganther 1990 Cancer Res 50:1206 ES Henle, S Linn 1997 J Biol Chem 111: 19095 S Liu, D Shia, G Liu, H Chen, S Liu, Y Hu 2000 Life Sci 68:603 M Bjomstedt, S Kumar, A Holmgren 1992 J Biol Chem 267:8030 S Kumar, M Bjomstedt, A Holmgren 1992 Eur J Biochem 207:435 HJ Thompson, R Strange, PJ Schedin 1992 Cancer Epidemiol Biomarkers Prev 1:5972 HE Ganther 1999 Carcinogenesis 20:1657 JY Kim, HS Park, SI Kang, EJ Choi, lY Kim 2002 Biochim Biophys Acta 1569:67 J Ghosh 2004 Biochem Biophys Res Commun 315:624 PL Bergad, WB Rathbun 1986 Curr Eye Res 5:919 Y Tashima, M Terui, H Itoh, H Mizunuma, R Kobayashi, F Marumo 1989 J Biochem (Tokyo) 105:358 F Islam, Y Watanabe, H Morii, O Hayaishi 1991 Arch Biochem Biophys 289:161 R Sinha, D Medina 1997 Carcinogenesis 18:1541 R Gopalakrishna, ZH Chen, U Gundimeda 1997 Arch Biochem Biophys 348:37 ML Handel, CK Watts, A deFazio, RO Day, RL Sutherland 1995 Prvc Natl Acad Sci U S A 92:4497 IY Kim, TC Stadtman 1997 Proc Natl Acad Sci USA 94:\ 2904 HS Park, SH Huh, Y Kim, J Shim, SH Lee, IS Park, YK Jung, lY Kim, EJ Choi 2000 J Biol Chem 275:8487 HS Park, E Park, MS Kim, K Ahn, lY Kim, EJ Choi 2000 J Biol Chem 275:2527 RW Johnstone, AA Ruefli, SW Lowe 2002 Cell 108:153 DR Green 2000 Cell 102:1 V Petronilli, P Costantini, L Scorrano, R Colonna, S Passamonti, P Bemardi 1994 J Biol Chem 269:16638 DY Cho, U Jung, AS Chung 1999 Biochem Mol Biol Int 47:781 Z Zhu, W Jiang, HE Ganther, C Ip, HJ Thompson 2000 Biochem Pharmacol 60:1467 MS Stewart, RL Davis, LP Walsh, BC Pence 1997 Cancer Lett 117:35 Z Zhu, M Kimura, Y Itokawa, T Aoki, J A Takahashi, S Nakatsu, Y Oda, H Kikuchi 1996 Biol Trace Elem Res 54:\2i U Jung, X Zheng, SO Yoon, AS Chung 2001 Free Radic Biol Med 31 ;479 HM Shen, CF Yang, WX Ding, J Liu, CN Ong 2001 Free Radic Biol Med 30:9 C Jiang, KH Kim, Z Wang, J Lu 2004 Nutr Cancer 49:174 L Zuo, J Li, Y Yang, X Wang, T Shen, CM Xu, ZN Zhang 2004 Ann Hematol 83:751 M Takahashi, T Sato, F Shinohara, S Echigo, H Rikiishi 2005 Int J Oncol 27:489 YW Chung, TS Kim, SY Lee, SH Lee, Y Choi, N Kim, BM Min, DW Jeong, lY Kim 2006 Toxicol Lett 160:143

Chapter 34. Selenoprotein mimics Junqiu Liu Key Laboratory for Supramolecular Structure and Materials, Jilin University, Changchun 130012, China

Guimin Luo Key Laboratory for Molecular Enzymology and Engineering, Jilin University, Changchun 130023, China

Summary: Selenium has a long history of association with human health and disease. This essential trace element exerts its important biological role in the form of selenocysteine in selenoproteins. To explore the structural and functional importance of selenium in selenoproteins and to apply this information to develop selenium-based medicine, significant efforts have been made in selenoprotein biomimetic chemistry. Because of the unique redox properties of selenium in mammalian selenoenzyme glutathione peroxidase (GPx), a number of organoselenium/tellurium compounds and selenoproteins were designed to mimic the natural GPx. Recently, mimics were developed for other important selenoenzymes, such as iodothyronine 5'deiodinase. Herein, several strategies that utilize chemical and biological techniques or the combination of both for the redesign of selenoenzyme structure and function are reviewed. It can be anticipated that as our understanding of the basic biology and biochemistry of selenoproteins increases, even more sophisticated approaches will be developed for the rational design of new selenoprotein mimics. Introduction In 1973, selenium was found to be a vital component of two bacterial enzymes, formate dehydrogenase [1] and glycine reductase [2]. More importantly, however, the biochemical role of selenium in mammals was established by the discovery that selenium is a crucial part of the active site of the antioxidant enzyme, glutathione peroxidase (GPx) [3,4]. Following this discovery, a number of selenoproteins have been identified in prokaryotic and eukaryotic cells [5,6, also see Chapter 9], including well studied selenoenzymes, such as thioredoxin reductases [7], iodothyronine deiodinases [8], selenophosphate synthetase [9], selenoprotein P (SeP) and selenoprotein W (SeW) [10,11]. Selenium is now viewed as an essential

388

Selenium: Its molecular biology and role in human health

trace element that exerts its physiological role as selenocysteine (Sec) residue in at least 25 distinct selenoproteins in mammals. Selenium is associated with human health and disease to a large extent as integral component of selenoproteins [5,6,12,13]. Thus, the pharmacology, biology and biochemistry of selenoproteins are of significant interest. To understand the structural and functional roles of selenium in selenoproteins and to apply them in the development of selenium-based medicine, selenoprotein biomimetic chemistry is particularly important [14-16]. Thus, many organoselenium compounds have been studied as biological models that are capable of simulating catalytic functions of natural enzymes [14-16]. Of these selenoproteins, mammalian selenoenzyme glutathione peroxidase has received particular attention due to unique redox properties of selenium in this enzyme [17,18]. During the past decade, a number of organoselenium/ tellurium compounds were synthesized to mimic the action of this important selenoenzyme [14-16]. Recent advances were made in the development of synthetic or semisjoithetic selenoproteins mimicking GPx function [16]. These selenoproteins consist of selenium-containing catalytic antibodies, bioimprinted proteins and modified enzymes. However, there are only a few reports on the design of other selenoenzyme models [19]. In this chapter, we review recent advances regarding the design of selenoprotein mimics through the synthesis of organoselenium compounds, transformation of proteins by chemical modification and the design by genetic engineering. Antioxidant selenoproteins It is well established that a variety of human diseases are associated with oxidative stress caused by reactive oxygen species (ROS), including Alzheimer's disease, myocardial infarction, atherosclerosis, Parkinson's disease, autoimmune diseases, radiation injury, emphysema, and sunburn [2022]. ROS are known mediators of intracellular signaling cascades. Excess production of ROS, however, may lead to oxidative stress, resulting in damage to DNA, lipids, and proteins, loss of cell function, and ultimately, apoptosis or necrosis. A balance between oxidation and the antioxidant function is vital for cell function, regulation, and adaptation to diverse stimuli [23]. To protect themselves from oxidative injury, aerobic organisms possess a complex and elaborate antioxidant defense system. This defense system includes both nonenzymatic antioxidants, such as glutathione (GSH), and antioxidant enzymes, such as superoxide dismutase (SOD) [24], catalase (CAT) [25], and GPx [17,18]. GPx (EC 1.11.1.19) catalyzes the reduction of a variety of hydroperoxides (ROOHs) using GSH as reductant. It plays a vital role in scavenging ROS and therefore protects mammalian cells from oxidative damage [26]. There are four well-characterized GPx selenoenzymes in mammals: the classical cellular GPx (cGPx) [3,4], phospholipid hydroperoxide GPx (PHGPx)

Selenoprotein mimics

389

[27], plasma GPx (pGPx) [28], and gastrointestinal GPx (GIGPx) [29]. Although cGPx, GIGPx, and pGPx are homotetramers, PHGPx is a monomer with a molecular size smaller than the other GPx [30]. In addition to GPx, other selenoproteins, such as SeP [31], SeW [32], and thioredoxin reductase [7], exert their functions as antioxidants. Of these enzymes, cGPx has thoroughly been studied and its crystal structure was refined to a 0.2-nm resolution [33]. Synthesis of organoselenium-based selenoenzyme mimics The rational design and synthesis of low molecular weight catalysts, which mimic natural enzyme function, may help elucidate enzyme structure and mechanism as well as serve as pharmaceuticals in the treatment of or protection fi-om disease. The design of GPx models represents successful examples in selenoprotein mimics. Insights into the structure and mechanism of native GPx resulted in the design of a series of GPx models. Ebselen, ebselen analogs, selenenamides, diselenides, a-phenylselenoketones, selenium-containing proteins, antibodies, and cyclodextrins, as well as their tellurium analogues, were demonstrated to catalyze the reduction of hydroperoxides (ROOH) in the presence of thiols [14-16]. Scheme 1 O

CC^^ rrN^>cooH I Se

\ =

Se'

NH

2 O -Se'

44

NO, 5

O

fA 7

6 OMe

NH Se'

€6

It was suggested that two conserved amino acid residues, tryptophan and glutamine, are part of a catalytic triad in which the selenol group of Sec is both stabilized and activated by hydrogen bonding with the imino group of tryptophan residue and with the amido group of the glutamine residue [33,34]. In the synthetic approach, the initial attempt is to synthesize simple organoselenium compounds in which the S^—N interaction observed in GPx is mimicked by placing N or O in close proximity to selenium. This can be accomplished by creating a mimic in which selenivim binds directly to a heteroatom such as nitrogen. Ebselen (2-phenyl-l,2-benzisoselenazol-3(2H)-

390

Selenium: Its molecular biology and role in human health

one) (structure 1), the first biologically active selenoorganic compound, represents an excellent example of GPx mimic [35,36]. Following the discovery of ebselen, several groups synthesized its analogs and derivatives (e.g., 2-9 in Scheme 1). In order to improve the reactivity and solubility of ebselen, a number of attempts have been made to modify its basic structure through substituent effects and isosteric replacements (Scheme 1) [37,38]. ie2 jmQ\

"ClN^J*

f^^^^h

J^Sefj-

KJ 10

KJ 11 Et

[jl^OH

Pi"

"OH

Se)2

Me

14

16

15

Another way of imitating the S—^N interaction in GPx is the design wherein the selenium is not directly bound to the heteroatom (N or O), but is located in close proximity to it. Thus, GPx models (e.g., 10-13) [39-41] having Se—N interactions and compounds (e.g. 14-17) possessing Se—O interactions were developed [14,16] (Scheme 2). However, the presence of an amino group in close proximity to selenium does not always play a positive role. Recently, Scheme 3 Me NMe,

18

19

H

C^&>^NMe,

^

^

^

20

Singh et al reported the GPx-like activity of a series of diaryl diselenides having intramolecular Se—N interactions [42]. It was found that the diselenides that have quite strong Se—N intramolecular interactions were less active, whereas the diselenides that contain a built-in basic amino group but no Se—N interactions showed excellent GPx activity [43]. The ferrocene-based diselenides (18-20 in Scheme 3) were reported to display much higher activities than those of the phenyl-based diselenides, and the enhancement in catalytic activity could be ascribed to the synergistic effect of redox-active ferrocenyl and internally chelating amino groups.

Selenoprotein mimics

391

By analogy to diselenides, ditellurides and related compounds were proposed as GPx mimics. Engman et al [44] was first to report that some of these compounds (e.g. 21-22) show GPx-like activity. Modifying the basic structure of diphenyl ditelluride based on substituent effects and isosteric replacements further impacts the reactivity of diphenyl ditelluride. More recently, Mugesh et al [45] directly compared the thiol peroxidase activity of several ditellurides (e.g., 23-26 in Scheme 4) with that of their selenium analogues. All ditellurides were found to be much more efficient catalysts than the corresponding diselenides in reducing H2O2 with PhSH as cosubstrate. Scheme 4 ^R

"O'^'K!}"' "O'^'-^'O'' 0-^=-'^^ 22

21

oci

^ 24

23

22-1

NMej

Teh

25

26

Scheme 5 X-X

0

0)_ I ' f 27,28 X = Se, Te

32

0

29-31 X =Se,Te,Sec o HNCH2(CH2NHCH2)„CH2-N^

Se-S(

SeOgH

Q

j[

J

SeOgH

33-35 n = 1-3

Cyclodextrins (CDs) are cyclic oligosaccharides containing a hydrophobic cavity, in which many complexes can be formed via host-guest chemistry [46]. They have extensively been exploited as enzyme models and molecular receptors [47]. To elucidate the effect of substrate recognition on catalysis, a series of cyclodextrin-derived organoselenium and organotellurium compounds (e.g. 27-35) were developed as GPx models (Scheme 5) [48-51].

392

Selenium: Its molecular biology and role in human health

The first model compound 29 was prepared by attaching a diselenide group to the CD primary face [48]. Attachment of a ditelluride group onto cyclodextrin resuhed in GPx models 28 (2-TeCD) and 30 (6-TeCD) [50]. The catalytic efficiency of 2-TeCD-catalyzed reduction of hydroperoxides by GSH was found to be 350,000-fold higher than that involving diphenyl diselenide (PhSeSePh). Selenoenzyme transformation by chemical modification Transformation of natural enzymes into selenoenzymes Although many low molecular weight GPx mimics are known, they possess serious disadvantages: low activity, low solubility in water, and in some cases, toxicity. In this regard, natural proteins may have advantage as protein macromolecules carry molecular information for both substrate recognition and efficient catalysis. Engineering proteins by genetic and chemical methods is a valuable strategy for introducing new functions into protein scaffolds. So far, three corresponding protein design methods have been described: site-directed mutagenesis, chemical modification and the combination of both. By using these strategies, natural enzymes, proteins and antibodies have been used successfully to construct efficient selenoenzyme models [16]. Scheme 6 ICH2—Q_'^!^L_/2|l/4(H

"^°2

/||^^XxO;H

The first example in the field of selenoenzyme design is the chemical conversion of the active site serine residue of the bacterial serine protease subtilisin (EC 3.4.21.14) into selenocysteine [52]. The hydroxyl group of Ser221 could be selectively modified to introduce distinct functional groups into the active site of subtilisin. Inspired by the earlier work on the first semisynthetic enzyme, thiolsubtilisin [53], Wu and Hilvert prepared selenosubtilisin by using a similar method [54] (Scheme 6). The semisynthetic selenoenzyme exhibited significant GPx-like redox activity. It catalyzed the reduction of a variety of hydroperoxides by 3-carboxy-4nitrobenzenthiol (ArSH). The reduction of ter^butyl hydroperoxide (^ BuOOH) by ArSH was at least 70,000-fold faster than the reaction catalyzed by diphenyl diselenide, a well-studied antioxidant [54,55]. Since selenosubtilisin has the same substrate binding pocket as subtilisin, it was possible to rationalize and even predict its substrate selectivity. Thus, a series of different racemic hydroperoxides was chemically synthesized by the Schreier group and subjected to selenosubtlisin-catalyzed reactions [56,57].

Selenoprotein mimics

393

All alkyl aryl hydroperoxides showed an enrichment of enantiomers. In a similar fashion, Liu et al prepared a selenotrypsin by converting the active site serine into Sec [58]. The study revealed that GSH is not a particularly good substrate for selenotrypsin. Nevertheless, the data showed that it was possible to convert an active site serine into Sec in various serine proteases. Recent studies showed that tellurium is an excellent alternative element for the construction of GPx models [14-16]. However, introducing tellurium into proteins is currently a challenge. Following selenosubtilisin, Liu and coworkers developed a methodology to introduce tellurium into the binding pocket of subtilisin and yielded a first semisynthetic telluroen2yme tellurosubtilisin (Scheme 6) [59]. Like natural GPx, tellurosubtilisin can catalyze the reduction of ROOH by thiols efficiently and acts as an excellent GPx mimic.

Table 1. Catalytic activities of selenoprotein GPx mimics (Data from ref. 21). selenoenzyme mimic

GPx activity (U/nmol)

ebselen

1

printed protein

100-800

catalytic antibody

1100-24300

selenoGST

2000-6200

natural GPX

5780

Although seleno/tellurosubtilisin and selenotrypsin were generated via covalent modification of naturally occurring enzymes [54,58,59], it is a great challenge to prepare highly efficient semisynthetic enzymes that can rival natural selenoenzymes. Recently, Luo et al developed a method to mimic the action of GPx by chemically modifying a naturally occurring enzyme glutathione transferase (GST, EC.2.5.1.18) [60]. Taking advantage of the highly specific GSH binding site of GST, seleno-GST(Se-GST) was generated by chemical mutation using a method described for preparation of selenosubtilisin [54]. The selenium-containing Se-GST can efficiently catalyze the reduction of hydrogen peroxide with an activity that is greater than that for some natural counterparts (Table 1) [60].

394

Selenium: Its molecular biology and role in human health

Transformation of natural proteins into selenoenzymes To create an efficient artificial enzyme, the affinity for the substrate in the enzjmie-substrate complex must be reasonably high, and the catalytic groups should be adjacent to the reactive group of the substrate. An alternative approach to artificially creating such binding sites is the molecular imprinting technique [61]. Biopolymers can also be used as an alternative backbone for the imprinting procedure. This innovation has led to the development of the bioimprinting technique for the synthesis of proteinbased binding and catalytic sites. An imprinted enzyme model with GPx activity has been developed by a combination of bioimprinting and chemical mutation [62], A';,S-bis-2,4dinitrophenyl-glutathione (GSH-2DNP), a GSH derivative, was synthesized and acted as a template molecule (36 in Scheme 7). In the bioimprinting process, the imprinted molecule was allowed to interact with denatured proteins (e.g., egg albumin) to form a new conformation via hydrogen bonds, ion pairing and hydrophobic interactions. Scheme 7 N02 N02

--

H^

'>N02 '

O - < 0 > - "COOH - ^ ^ N J YO" ^ COOH 36

The new conformation was then fixed using the cross-linker glutaraldehyde. After removal of the imprinting molecule by dialysis, the serine residues located at the binding sites of the inprinted proteins were activated using phenylmethene sulfonyl fluoride and then converted into Sec in the presence of NaHSe. The imprinted protein exhibits GPx activity and is 100-800 fold more active than ebselen (Table 1). Transformation of antibodies into selenoenzymes One important way for designing binding sites is the use of antibodies. A recognition site for enzyme substrate is easy to generate by using a standard monoclonal antibody (McAb) preparation technique. This strategy was widely applied in the design of catalytic antibodies using transition state analogs as haptens. Recently, Luo et al employed this strategy in the design of GPx mimics [63-65]. The authors used substrate analogs instead of transition state analogs as haptens in order to generate monoclonal antibodies with the substrate binding site. In the design of catalytic antibodies, the polar groups of substrate GSH were modified by different hydrophobic

Selenoprotein mimics

395

groups and the modified substrates were used as a series of haptens (Scheme 8). Scheme 8 NH2

o

H

6

O

S

ISIH-,

"

o

o

HO

os-^^^o

6^°^

O H ,-^0 O

NO2

H3CO-^^'^V^N'^^V"OCH3 O H NH2 40

37-39 R = H, CHj, CHjCHjCHjCHj

The substrate binding sites were first made by monoclonal antibody preparation technique using hydrophobically modified GSH and GSSG as haptens (37-40 in Scheme 8). Thus, not only is the hydrophobic cavity of the antibody similar to that of the active site of native enzymes, but the affinity of the antibody active site for the substrate can also be adjusted to that of the native enzyme. The catalytic Sec was then incorporated into the McAb by chemical modification of the serine residue (Scheme 9) [63]. Surprisingly, these catalytic antibodies exhibited remarkably high catalytic efficiency which could rival the natural enzyme (e.g., rabbit liver GPx) (Table 1). In order to produce pharmaceutical proteins and elucidate the reason why this novel catalyst exhibited high catalytic efficiency, a selenium-containing single chain antibody was prepared (Se-ScFv) [65]. Scheme 9

MaAb production

y - y I 'S-CHjOH

PMSF

f - y " lS-CH20S02CH2-/>

NaHSe

(^ " ljS-CH2SeH

Similarly, Ding and coworkers [19] prepared a selenium-containing catalytic antibody (Se-4C5) by converting the serine residues of monoclonal antibody 4C5 raised against thyroxine (T4) into Sec. Se-4C5 catalyzes the deiodination of T4 to 3,5,3'-triiodothyronine (T3) in the presence of dithiothreitol via a ping-pong mechanism, with a Vmax value of 270 pmol mg" ' min"'. Thus, Se-4C5 acted as a deiodinase mimic.

396

Selenium: Its molecular biology and role in human health

Transformation of proteins into selenoenzymes by genetic engineering Since Sec is encoded by a stop codon UGA, it is difficult to prepare selenoproteins by traditional recombinant DNA technology. The most suitable approach to bioincorporating selenium is the auxotrophic expression technique. In 1975, the first use of selenium in sulfur pathways in E.coli was reported by Cowie & Cohen [66]. Following this early work, there was considerable interest in the insertion of selenium analogs of sulfur-containing amino acids into proteins. Moroder and Budisa further developed this strategy and incorporated selenomethionine, telluromethionine and their isosteric analogs into proteins in order to solve the phase problem in protein X-ray crystallography [67,68]. Furthermore, Bock et al used a similar auxotrophic expression system and incorporated Sec into thioredoxin [69]. The biosynthetic substitution of the catalytically essential cysteine (Cysl49) of glyceraldehyde 3-phosphate dehydrogenase (GAPDH) by Sec led to selenoGAPDH that displayed GPx-like properties [70]. This selenoenzyme catalyzed the reduction of hydroperoxides with aryl thiols instead of GSH as this enzyme lacked a GSH-binding site. Similarly to the studies on seleno-GAPDH, Liu et al converted glutathione transferase (Lucilia cuprina LuGSTl-1) into selenoenzyme (seleno-LuGSTl-1) by means of auxotrophic expression [71]. The Ser9 in the GSH-binding site was mutated to cysteine and then biosynthetically substituted to selenocysteine in an auxotrophic expression system. This novel selenium-dependent enzyme exhibited high catalytic activity toward H2O2 in the presence of GSH, which was similar to that of the native GPx. For the first time, a seleniumcontaining enzyme with such remarkable GPx activity was generated by genetic engineering. It is now clear that the design of selenoprotein mimics plays important roles in understanding biochemical functions and reaction mechanisms. It is becoming apparent that selenoprotein mimics possess therapeutic potential against various diseases and that their fiinctions range from antioxidants to anticancer and antiviral agents. It can be anticipated that as our understanding of the basic biology and biochemistry of selenoproteins increases, future efforts will result in even more sophisticated approaches to the rational development of new selenoprotein mimics. References 1. 2. 3. 4. 5. 6. 7. 8.

JR Andreesen, L Ljungdahl 1973 J Bacterial 116:867 DC Turner, TC Stadtman 1973 Arch Biochem Biophys 154:366 L Flohe, EA GOnzler, HH Schock 1973 FEBSLett 32:132 JT Rotruck, AL Pope, HE Ganther, et al 1973 Science 179:588 GV Kryukov, S Castellano, SV Novoselov et al 2003 Science 300:1439 VN Gladyshev, GV Kryukov, DE Fomenko, DL Hatfield 2004 Annu Rev Nutr 24:579 T Tamura, TC Stadtman 1996 Proc Natl Acad Sci U.S.A. 93:1006 JR Arthur, F Nicol, GJBeckett 1990 Biochem J 111:52,1

Selenoprotein mimics 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57.

397

D Mustacich, G Powis 2000 Biochem y 346:1 MA Motsenbocker, AL Tappel 1984 J Nut 114:279 MA Beilstein, SC Vendeland, E Barofsky et al 1996 Inorg Biochem 61:117 TC Stadtman 1996 Ann Rev Biochem 65:83 L Floh6, 1989 in Glutathione: Chemical, Biochemical, and Medical Aspects, D Dolphin, R Poulson, O Avamovic,(eds) Wiley, New York, p644 GMugesh,MWWdu,HSies2001C/ieOT7?ev 101:2125 GMugesh,HB Singh 2002 ^ccC/iem/fes 35:226 G Luo, X Ren, J Liu, Y Mu, J Shen 2003 Curr Med Chem 10:1151 TC Stadtman 1991 J Biol Chem 266:16257 F Ursini, 1994 In Oxidative Processes and Antioxidants; R Paoletti (ed) Raven Press New York p25 G Lian, L Ding, M Chen et al IQQXBiochem Biophys Res Commun 283:1007 AY Sun, YM J Chen 1998 Swmaf&j 5:401 JM Mates, C Peter-Gomez, I Nunez de Castro 1999 Clin Biochem 32:595 S Cuzzocrea, DP Riley, A? Caputi, D Salvemini 2001 Pharmacol Rev 53:135 H Sies, 1985 In Oxidative Stress, H Sies (ed) Academic Press, London, pi L Benov, I Fridovich 1998 J Biol Chem 273:10313 H Aebi, 1974 In: Methods of enzymatic analyses, HU Bergmeyer (ed) Academic Press New York, p673 GC Mills 1957 J Biol Chem 229:189 A Roveri, M Maiorino, C Nissii, F Ursini 1994 Biochem Biophys Acta 1208:211 TR Pushpa-Rekha, AL Burdsall, LM Oleksa et al 1995 J Biol Chem 270:26993 RS Esworthy, KM Swiderek, YS Ho, FF Chu 199 SBiochem Biophys .4cto 1381:213 F Ursini, M Maiorino, M Valente et al 1982 Biochem Biophys Acta 710:197 Y Saito, T Hayashi, A Tanaka et al 1999 J Biol Chem 29:2866 MA Beilstein, SC Vendeland, E Barofsky et al \9% J Inorg Biochem 61:117 O Epp, R Ladenstein, A Wendel 1983 Eur J Biochem 733:51 B Ren, W Huang, B Akesson, RJ Ladenstein 1991 Mol Biol 268:869 AEP MUller, E Cadenas, P Graf H, Sies 1984 Biochem Pharmacol 33:3235 A Wendel, M Fausel, H Safayhi, G Tiegs, RA Otter 1984 Biochem Pharmacol 33:3241 C Lambert, R Cantineau, L Christiaens et al 1986 Bull Soc Chim Belg 23:59 L Engman, A Hallberg 1989 J Org Chem 54:2964 TG Back, BP Dyck 1997 J Am Chem Soc 119:2079 SR Wilson, PA Zucker, RRC Huang, A Spector 1989 J Am Chem SocWX :5936 M Iwaoka, S Tomoda \99AJAm Chem Soc 116:2557 G Mugesh, A Panda, HB Singh, NS Punekar, RJ Butcher 1998 Chem Commun 227 G Mugesh, A Panda, HB Singh et al 2001 J Am Chem Soc 123:839 L Engman, D Stem, lA Cotgreave, CM Andersson 1992 J Am Chem Soc 114:9737 G Mugesh, A Panda, S Kumar et al 2002 Organometallics 21:884 G Wenz 1994 Angew Chem Int Ed Engl 33:803 R Breslow, SD Dong 1998 Chem Rev 98:1997 JQ Liu, SJ Gao, GM Luo et al 1998 Biochem Biophys Res Common 247:397 JQ Liu, GM Luo, XJ Ren et al 2000 Biochim Biophys Acta 1481:222 ZY Dong, JQ Liu, SZ Mao et al 2004 J Am Chem Soc 126:16395 Y Liu, B Li, L Li, HY Zhang 2002 Helv Chim Acta 85:9 FSJr Markland, E Smith, 1971 In The Enzymes, PD Boyer (ed) Academic Press New York Vol III p 561 T Nakatsuka, T Sasaki, ET Kaiser 1987 J Am Chem Soc 109:3808 ZP Wu, D Hilvert 1990 J Am Chem Soc 112:5647 EB Peterson, D Hilvert, 1995Biochemistry 34:6616 D Haring, M Herderich, E Schuler, et al 1997 Tetrahedron Asymmetry 8:853 D Haring, E Schueler, W Adam, CR Saha-Moeller, P Schreier \999 J Org Chem 64:832

398 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71.

Selenium: Its molecular biology and role in human health JQ Liu, MS Jiang, GM Luo, GL Yan, JC Shen 1998 Biotechnol Lett 20:693 SZ Mao, ZY Dong, JQ Liu et al 2005 J Am Chem Soc 127:11588 XJ Ren, P Jemth, PG Board et al 2002 Chem Biol 97:89 G Wulff 1995 Angew Chem Int Ed Engl 34:1812 J Liu, G Luo, S Gao, K Zhang, X Chen, J Shen 1999 Chem Commun 199 GM Luo, ZQ Zhu, L Ding et al 1994 Biochem Biophys Res Commun 198:1240 L Ding, Z Liu, ZQ Zhu, GM Luo, DQ Zhao, JZ Ni 1998 Biochem J 2,2,2:251 XJ Ren, SJ Gao, DL You et al 2001 Biochem J 359:369 DB Cowie, GN Cohen, 1957 Biochim Biophys Acta 26:252 N Budisa, B Steipe, P Demange et al 1995 Eur J Biochem 230-JS& N Budisa, C Minks, FJ Medrano et al 1998 Proc Natl Acad Sci USA 95:455 S MuUer, H Senn, B Gsell, W Vetter, C Baron, A Bock 1994 Biochemistry 33:3404 S Boschi-Muller, S Muller, AV Dorsselaer et al 1998 FEBS Letters 439:241 HJ Yu, JQ Liu, A Bock, J Li, GM Luo, JC Shen 2005 J Biol Chem 280:11930

Chapter 35. Update of human dietary standards for selenium Orville A. Levander Beltsville Human Nutrition Research Center, U. S. Department of Agriculture, Agricultural Research Service, Beltsville, MD 20705, USA

Raymond F. Burk Division of Gastroenterology, Hepatology, and Nutrition, Department of Medicine, Vanderbilt University School of Medicine, Nashville, TN 37232, USA

Summary: An update of the human dietary standards for selenium is presented, including the year 2000 Dietary Reference Intakes (U.S.A.), the 1996 standards of the World Health Organization (WHO), and a recent relevant intervention trial carried out in China. Two criteria have been used by official bodies to set recommendations. One is the prevention of Keshan disease, the only proven selenium-responsive disease. The other is full expression of all selenoproteins as indicated by optimization of a plasma biomarker, either glutathione peroxidase activity or selenoprotein P concentration. An average per capita intake of 20 |xg selenium/day will prevent Keshan disease in a population but will not allow optimization (full expression) of selenoproteins. Using plasma glutathione peroxidase activity as the selenium biomarker to be optimized, the RDA for adults in the U.S.A. was set at 55 )ig in 2000. The recent trial in China utilized selenoprotein P as a biomarker and its results suggest that an upward revision of the current RDA will be needed. Even higher intakes of selenium have been postulated to prevent cancer. Intervention trials now underway in the U.S.A. are evaluating that possibility and the safety of large selenium supplements. Introduction In South Dakota during the 1930s, selenium was identified as the toxic agent in animal feeds and forages that caused the livestock poisoning known as "alkali disease" [1]. Plants that grew in certain areas of the Great Plains of the United States took up so much selenium from the selenium-rich soils that they became toxic to poultry and livestock. In cattle the disease is characterized by hair and hoof loss and a generalized emaciated appearance. For an extensive description of selenosis in farm animals, consult the monograph by Rosenfeld and Beath [2].

400

Selenium: Its molecular biology and role in human health

There were suggestions in the literature that selenium might have a beneficial effect under certain conditions (e.g., see the work of Pinsent with bacteria [3]), but the prevailing opinion was that selenium was a toxin that played no positive role in metabolism. Moreover, selenium compounds were thought to be carcinogenic. Then in 1957 Schwarz and Foltz [4] announced their discovery that traces of dietary selenium could protect vitamin Edeficient rats from developing liver necrosis. Very soon thereafter, seleniumresponsive diseases were found in a variety of economically important farm animals including turkeys, chickens, sheep, swine, and cattle [5]. The need for selenium in human nutrition was shown by Chinese scientists who demonstrated in 1979 that this essential trace mineral protected against Keshan disease, a cardiomyopathy affecting young children and women of child bearing age residing in low-selenium areas in China. This finding increased greatly the interest in the human selenium requirement [6]. The current and previous dietary standard for selenium, the year 2000 Dietary Reference hitake (DRI) [7] and the 1989 RDA [8], respectively, were both based on maximization (or optimization) of plasma glutathione peroxidase activity so their values were rather similar. A proposal to base the next selenium standard on the full expression of selenoprotein P would lead to a higher value for the standard and is evaluated below. This review concludes with a discussion about the possible use of selenium as a cancer chemoprevention agent. Such a practice, if justified by studies that are in progress and planned, might result in substantially elevated recommendations of selenium intake. RDAs - tenth edition (1989) hiterest in the possible beneficial effects of selenium in human health continued to grow well into the 1980s and this was reflected in the Tenth Edition of the RDAs [8]. Literature citations increased six-fold. This was due not only to the increasing number of papers dealing with selenium but also to the expressed desire of the RDA Committee to make the RDA book more "scientific" and one in which every step of the derivation of the RDA was "transparent" so that the logic and reasoning behind the derivation of the dietary standard was clear and open for everybody to see. Fortunately for selenium researchers, several studies from China allowed the RDA Committee to pinpoint human selenium requirements with increased precision such that it was possible to advance selenium to full RDA status for the first time. One group of studies examined the dietary selenium intake needed to prevent Keshan disease in regions of China where it was endemic [9]. The disease did not occur in those areas in which the selenium intake by adults was 17 jxg/day or more. Thus, 17 ^g/day was suggested as a minimum daily requirement based on disease prevention.

Update of human dietary standards for selenium

401

In another study, a "physiological" selenium requirement was determined by following increases in glutathione peroxidase activity in the plasma of men living in a Keshan disease area who were given graded doses of selenomethionine as a supplement over a period of several months [10]. At a total intake of 41 jig/day or more (11 p.g from diet), the plasma glutathione peroxidase activity became optimized. Therefore, it was concluded that Chinese men had a physiological requirement of 41 fig/day. In order to convert this figure to an RDA for North American males, it was necessary to apply a correction factor for differences in body weight (79/60) and to apply a safety factor (1.3) to allow for individual differences in requirement. Thus, the calculation for adult males became: 41 X 79/60 X 1.3 = 70^g/day For adult North American females, the calculation was: 41 X 63/60 X 1.3 = 55ng/day A more detailed explanation of the RDA calculations for adults was presented elsewhere [11]. Because of the lack of data, the RDAs for young adults also served as the basis of RDAs for the elderly. Likewise, because of the lack of data, RDAs for infants and children were based on adult values with extrapolations downward on the basis of body weight plus a factor arbitrarily allowed for growth. The RDA during pregnancy was calculated using a factorial technique based on the fetal accretion of selenium. The RDA during lactation provided sufficient selenium to avoid depletion of the mother and permit a satisfactory selenium content in the breast milk. The Tenth Edition discussed selenium toxicity only in general terms [8]. An episode of human selenosis in China was described in which hair loss and fingernail changes were observed on intakes approximating 5000 jig selenium/day. It was pointed out that sensitive and specific biochemical indices of selenium overexposure were not available and no attempt was made to establish a safe upper limit of selenium intake. World Health Organization (1996) In 1996, the World Health Organization (WHO) published its dietary standards for several trace elements, including selenium [12]. WHO has the responsibiUty for setting recommendations that apply to many different countries around the globe (United Nations member states) that have highly varied national diets. For that reason, the Organization tends to suggest nutrient intakes that are often somewhat lower than those set in the U.S.A. This also turned out to be true for selenium since large parts of the U. S. Great Plains, a major wheat production area, have soils that are rich in

402

Selenium: Its molecular biology and role in human health

selenium and relatively generous amounts of the trace element are incorporated into the food chain. Intakes of selenium exceeding 100 i^g/day are not uncommon in the U.S. and so meeting an RDA of 55 to 70 }ig/day is not difficult. On the other hand, meeting the 1989 RDA for selenium could be quite a challenge for some other coimtries. Many parts of China, for example, routinely consume much lower amounts of selenium in their diet [9] and New Zealanders rarely ingest such RDA levels [13]. Likewise, Finland had a low-selenium food supply before deciding in 1985 to add selenium to its fertilizers [14]. In fact, dietary surveys indicate that several European countries would have problems achieving intakes as high as the 1989 RDA, including Belgium, Denmark, France, Germany, United Kingdom, Slovakia, and Sweden [reviewed in [15]]. The selenium intake in Switzerland was somewhat higher because of the common use of North American wheat rich in selenium. So it is not surprising that WHO was reluctant to set a dietary standard that so many of its member states could not attain, especially in the absence of any evidence of signs of human selenium deficiency outside of China. The reader will recall that the rationale used by the 1989 RDA Committee for its selenium recommendation was full expression of plasma glutathione peroxidase activity. The WHO Committee decided that such full activity was probably not necessary for human health and that only two-thirds full activity of plasma glutathione peroxidase still afforded sufficient protection against oxidative stress. This conclusion was based on observations that blood cells metabolized hydrogen peroxide satisfactorily until their glutathione peroxidase activity fell to one-quarter or less of normal. Of course, if one selects a lower target glutathione peroxidase activity for the biochemical criterion of adequate nutriture, this allows a lower dietary standard to be proposed also. In this case, the WHO Committee (formal designation: Joint FAO/IAEAAVHO Expert Consultation on Trace Elements in Human Nutrition) came up with 40 and 30 |J.g/day for the lower limit of the safe range of population mean dietary selenium intake that would meet the normative requirement of most adult males and females, respectively. As defined by WHO, the normative requirement referred to the "level of intake that serves to maintain a level of tissue storage or other reserve that is judged by the Expert Consultation to be desirable" [12]. WHO also defined a basal requirement that referred to the "intake needed to prevent pathologically relevant and clinically detectable signs of impaired function attributable to inadequacy of the nutrient." For selenium, the basal requirement was taken from the quantity needed to protect against Keshan disease. The lower limit of the safe range of population mean dietary selenium intake that would meet the basal requirement of most adult males

Update of human dietary standards for selenium

403

and females was calculated to be 21 and 16 p-g/day, respectively, after adjusting for body weight.

Table 1. The 1996 WHO Lower Limits of the Safe Ranges (Basal and Normative) of Population Mean Intakes of Dietary Selenium (ng/day)

Life Stage

Aee fvears)

Basal

Normative

Infants

0-0.25 0.25-0.5 0.5-1.0

3 5 6

6 9 12

1-3 3-6 6-10

10 12 14

20 24 25

Males

10-12 12-15 15-18 18+

16 19 21 21

30 36 40 40

Females

10-12 12-15 15-18 18+

16 16 16 16

30 30 30 30

Pregnancy

18

39

Lactation 0-3 months 3-6 months 6-12 months

21 25 26

42 46 52

Children

Adapted from [12].

The WHO Committee also attempted to deal with the question of tolerances of high dietary selenium intakes. On the basis of considerable fieldwork with human selenosis in China, Yang and associates proposed 750850 ^g as a marginal level of daily safe dietary selenium intake [16], defined as "the level of selenium intake at which few individuals have functional signs of excessive intake and above which the tendency to exhibit functional signs is apparent and symptoms may first appear among ... susceptible individuals [whose] selenium intake [is] further increased". The Committee

404

Selenium: Its molecular biology and role in human health

took one-half of the average of this range because of the uncertainty surrounding the harmful dose of selenium for people to suggest a maximal daily safe dietary selenium intake of 400 |ig. Dietary reference intakes (2000) The new millennium saw a host of changes in the way that dietary standards for selenium (and many other nutrients) were handled in the U. S. A. [7]. First of all, selenium was grouped with a variety of so-called "dietary antioxidants" (vitamins C and E and the carotenoids) instead of with the trace elements where it had traditionally been put. This change made sense because selenium, due to its multitude of roles protecting against oxidative stress, really had more in common with the nutritional antioxidants than it did with a collection of various microminerals. Another substantial change was in the dietary standards themselves [7]. The general term "Dietary Reference Intakes" was used to describe not only the RDA, but also Adequate Intake (AI), Tolerable Upper Intake Level (UL), and Estimated Average Requirement (EAR). Each of these terms has a particular role in describing the dietary standards of a nutrient and it might be worthwhile to repeat here their meanings as presented by the Panel on Dietary Antioxidants and Related Compounds: Recommended Dietary Allowance {RDA): the average daily dietary intake level that is sufficient to meet the nutrient requirements of nearly all (97 to 98 percent) healthy individuals in a particular life stage (which considers age, and when applicable, pregnancy and lactation) and gender group. Adequate Intake (AI): a value based on observed or experimentally determined approximations or estimates of nutrient intake by a group (or groups) of healthy people that are assumed to be adequate—used when an RDA cannot be determined. Tolerable Upper Intake Level (UL): the highest level of daily nutrient intake that is likely to pose no risk of adverse health effects to almost all individuals in the general population. As intake increases above the UL, the risk of adverse effects increases. Estimated Average Requirement (EAR): a daily average nutrient intake value that is estimated to meet the requirement of half the healthy individuals in a life stage and gender group." Thus, the redefinition of the RDA echoes the 1989 version [8], which states that they are "... the levels of intake of essential nutrients that... are judged ... to be adequate to meet the known nutrient needs of practically all healthy persons." The AI is reminiscent of the "Estimated Safe and Adequate Daily Dietary Intake" which was a dietary standard to be used when insufficient data were available to posit an RDA. The UL represents the first formal attempt by an "RDA Committee" to establish a ceiling of intake for the nutrients being considered by the group. In the "Dietary Reference Intakes"

Update of human dietary standards for selenium

405

(what the handbook on dietary standards in North America is now called— the title is no longer "Recommended Dietary Allowances") an entire chapter is devoted to describing a model for the development of ULs for nutrients. The EAR occupies a critical place in the new dietary standards, for without it, there can be no RDA. These two entities are related by the equation: RDA = EAR + 2 SD where SD is the standard deviation of the EAR. If the SD was unknown, the 2000 Committee generally assumed a coefficient of variation of 10% for the EAR so that RDA = 1.2 X EAR The 2000 Committee based its EAR on 2 intervention trials designed to estimate selenium requirements by determining the intake needed to optimize plasma glutathione peroxidase activity. The first trial was carried out in China [10] and in fact was the same study that served as the basis for the 1989 RDA [8]. The selenium intake needed to optimize plasma glutathione peroxidase in that work was 41 (ig/day, which came to 52 j^g/day after adjustment for Western body weight. The second intervention trial was from New Zealand [13] and the 2000 Committee interpreted that research as suggesting an EAR of 38 ^ig/day. Although other interpretations of the New Zealand trial may be possible [15], the average of both the New Zealand and Chinese trials, 45 |ig/day, was selected as the EAR. The RDA for adult males then was calculated as 45 X 1.2 to yield 55 ng/day. Thus, by using a lower base requirement figure than the 1989 Committee (45 vs. 52 (ig/day after adjustment for body weight) and a smaller correction factor for individual variation (1.2 vs. 1.3), the 2000 Committee arrived at a lower RDA figure for adult males than the 1989 Committee (55 vs. 70 |ig/day). Given the reported greater susceptibility of women to develop Keshan disease, their RDA was also set at 55 |ig/day despite their smaller body weight. The 2000 Committee could find no data available to calculate an EAR for children or adolescents, so the RDAs for them were extrapolated from young adult values. Similarly, there were no data that specifically addressed the selenium requirement for elderly persons and the 2000 Committee found no information that suggested that the aging process impaired selenium absorption or utilization, so their RDA was the same as young adults. A major philosophical shift occurred in the way that requirements were presented for infants up to one year if age. Because "No functional criteria of selenium status have been demonstrated that reflect response to dietary intake in infants", the 2000 Committee rescinded the 1989 RDA, so to speak.

406

Selenium: Its molecular biology and role in human health

and replaced it with an AI based on the "mean selenium intake of infants fed principally with human milk." This fundamental change in viewing infant requirements was not limited to selenium. In fact, other nutrients have also been accorded only AI status for infants, including calcium, magnesium, vitamin D, thiamin, and riboflavin. The 2000 Committee selenium AI for infants for the first and second 6 months of life are 15 and 20 ng/day, respectively, up 50% and 33% from their 1989 counterparts, respectively. Using somewhat different assumptions, the 2000 Committee came up with RDAs for pregnancy and lactation that were slightly less than those set by the 1989 Committee (60 vs. 65 fxg/day and 70 vs. 75 |ig/day, respectively). Table 2. The 2000 Dietary Reference Intakes (DRI) for Selenium (^g/day)

Life Stage

Age

DRI (ng/day)

Infants

0-6 mo 7-12 mo

15* 20*

Children

1-3 y 4-8 y

20 30

Males

9-13 y 14-70 y >70y

40 55 55

Females

9-13 y 14-70 y >70y

40 55 55

Pregnancy

60

Lactation

70

Adapted from [7]; values with asterisk are AI, others are RDA.

Another innovation in the 2000 DRI was the establishment of an upper limit of intake. The UL for selenium was based on the criteria of h^ir and nail brittleness and loss due to dietary overexposure in a high-selenium region in China. Intakes of selenium from food sources were inferred from blood levels. A No-Observed-Adverse-Effect-Level (NOAEL) was calculated to be 800 |ig/day. An imcertainty factor of 2 was chosen to protect sensitive individuals, thereby leading to a UL of 400 fig/day for adults 19 years of age

Update of human dietary standards for selenium

407

and older, a figure in agreement with upper limits set by others [12]. Finding no evidence of teratogenicity or selenosis in infants of mothers consuming high but not toxic amounts of selenium, the 2000 Committee kept to 400 Jig/day UL for pregnant and lactating women. The UL for infants 0 to 6 months old consuming human breast milk exclusively was set at 45 |ig/day based on the lack of any adverse effects (NOAEL) reported in such infants consuming breast milk containing 60 |ig selenium/L. The ULs for older infants, children, and adolescents were extrapolated on the basis of body weights. Recent trial using Selenoprotein P and glutathione peroxidase as biomarkers A report appeared in 2005 that described a selenium intervention trial carried out in 2001 in a low-selenium area of China [17]. The trial was designed to determine the selenium intake needed to 'optimize' the plasma selenium biomarkers, glutathione peroxidase and selenoprotein P. These 2 selenoproteins are accessible representatives (biomarkers) of the entire family of selenoproteins [18]. Optimization of them is used as an indicator of optimization (full expression) of all the selenoproteins in the body. A given selenoprotein is 'optimized' when providing additional selenium does not result in an increase in its concentration. There is a 'hierarchy' of selenoproteins with respect to their claim on available selenium. This means that when the selenium available will not allow optimization (full expression) of all selenoproteins, the selenoproteins most essential to the organism receive selenium and selenoproteins that are less essential do not. Thus, the "least essential" selenoprotein will be the last to be optimized when selenium availability is increased from inadequate to adequate, according to this concept. Animal experiments have suggested that liver glutathione peroxidase is the lowest selenoprotein in the hierarchy [19]. Because human tissues cannot be sampled for routine studies and blood can, plasma selenoproteins have been used as the best available representatives of the selenoproteins in the body. The subjects studied were farmers in a low-selenium region of Sichuan Province. Their average dietary selenium intake was 10 |j,g per day. hiitial plasma glutathione peroxidase activity was 40% of tj^ical U.S. values and selenoprotein P concentration was 23% of typical U.S. values. Subjects were supplemented for 20 weeks with placebo or several dose levels of selenite or L-selenomethionine (henceforth selenomethionine). Glutathione peroxidase became optimized with a supplement of 37 i^g of selenium per day as selenomethionine. The same level of glutathione peroxidase was achieved with 66 |j,g per day as selenite. The selenomethionine results are close to those of the 1983 study in China [10] and the 1995 study in New Zealand [13] and essentially confirm their results.

408

Selenium: Its molecular biology and role in human health

Selenite has not been evaluated previously in this type of study and its lower bioavailability than that of selenomethionine is noteworthy. Selenoprotein P did not become optimized in this trial, even when 61 ng selenium per day was administered as selenomethionine. Thus, selenoprotein P is lower in the hierarchy of human selenoproteins than plasma glutathione peroxidase and is therefore the better biomarker for optimization of the selenoproteins in the body. It seems likely that use of selenoprotein P as a selenium biomarker will lead to an increase in the selenium dietary intake recommendations. However, the results in this study do not allow an estimation of the amount of selenium needed because optimization of selenoprotein P was not achieved. Moreover, it is possible that supplementation at the dose levels used in this study for more than 20 weeks will optimize selenoprotein P. Thus, a study that includes higher supplemental dose levels and a longer supplementation period is needed. This study has pointed out that different biomarkers are optimized at different selenium intakes and that therefore the choice of biomarker is important. Presently, selenoprotein P appears to be a better biomarker than plasma glutathione peroxidase activity. The study has confirmed the results of the 2 earlier studies that used plasma glutathione peroxidase. It indicates, however, that the current selenium recommendations, based on glutathione peroxidase, are likely to be too low. Another important finding of this study is the greater bioavailability of selenium in the form of L-selenomethionine than in the form of selenite. This has implications for formulation of selenium supplements and addition of selenium to foods. Selenium as a possible cancer chemoprevention agent In total, almost 150,000 individuals have participated in phase III nutritional intervention studies to prevent cancer [20]. Besides selenium, nutritional intervention agents have included beta-carotene, alpha-tocopherol, retinal, and various vitamin and mineral combinations. Several lines of evidence have suggested that selenium might be a suitable candidate for testing as a chemoprevention agent. First, experiments with laboratory animal models indicated that various selenium compounds protect against tumorigenesis under a variety of conditions [21]. Second, about half of the 36 epidemiological studies evaluated by the FDA implied some value of selenium against cancer [22]. Finally, the selenium intervention trial of Clark and colleagues [23] found that subjects given 200 jxg selenium/day in yeast form to prevent skin cancer had lower incidences of several cancers than did the placebo group. However, no effect on skin cancer was observed. Because of the positive results in these studies, the U.S. National Cancer Institute is sponsoring a large trial (32,400 men) called SELECT (the

Update of human dietary standards for selenium

409

Selenium and Vitamin E Cancer Prevention Trial) [24]. SELECT is a phase III randomized, placebo-controlled test of selenium (200 (j.g/day as Lselenomethionine) and/or vitamin E (400 lU/day) to prevent prostate cancer in U.S. men. The subjects in SELECT are presumed to be selenium replete, with full expression of their selenoproteins. Thus, if SELECT demonstrates a preventive effect of selenium and/or vitamin E on prostate cancer, it will indicate that there is a health-related function of selenium independent of selenoproteins. Studies to characterize the dose of selenium needed to achieve the chemopreventive effect would then be desirable to inform strategies of supplementation. SELECT will also evaluate safety. Further consideration of the use of selenium to prevent cancer will depend on its being safe. The Food and Nutrition Board set an Upper Limit of selenium intake in their 2000 DRJ's at 400 Jig/day [7]. Intakes of subjects in the SELECT trial will be in the 300+ Hg/day range and determination of the safety of that intake over years will be important. Orderly progression of selenium chemoprevention studies is important. Only when data are produced that show selenium to be safe and effective in cancer prevention can recommendations to the public be made. References 1. 2. 3. 4. 5.

6. 7. 8.

9. 10.

11. 12. 13. 14. 15. 16. 17.

AL Moxon, M Rhian 1943 Physiol Rev 23:305 I Rosenfeld, O Beath 1964 Selenium. Geobotany, biochemistry, toxicity, and nutrition Academic Press New York J Pinsent 1954 Biochem J 57:10 K Schwarz, CM Foltz 1957 JAmer Chem Soc 79:3292 Subcommittee on Selenium, Committee on Animal Nutrition, Board on Agriculture, National Research Council 1983 Selenium in Nutrition Revised Edition National Academy Press Washington p 174 Keshan Disease Research Group 1979 Chinese Medical Journal 92:471 Institute of Medicine 2000 Selenium. In: Dietary reference intakes for vitamin C, vitamin E, selenium, and carotenoids National Academy Press Washington pp 284-324 Subcommittee on the Tenth Edition of the RDAs, Food and Nutrition Board, National Research Council 1989 Recommended Dietary Allowances lO"" Edition National Academy Press Washington GQ Yang, KY Ge, JS Chen, XS Chen 1988 World Rev Nutr Diet 55:98 G-Q Yang, L-Z Zhu, S-J Liu, L-Z Gu, P-C Qian, J-H Huang, M-D Lu 1987 In: Selenium in biology and medicine.. Part B (eds. GF Combs Jr, JE Spallholz, OA Levander, JE Oldfield) AVI New York pp 589-607 OA Levander 1991 J Am Diet Assoc 91:1572 Trace Elements in Human Nutrition and Health. Report of a Joint FAO/IAEA/WHO Expert Consultation 1996 World Health Organization, Geneva pp 343 AJ Duffield, CD Thomson, KE Hill, S Williams 1999 Am J Clin Nutr 70:896 A Aro, G Alfthan, P Varo 1995 Analyst 120:841 MP Rayman 2000 Lancet 356:233 G Yang, S Yin, R Zhou, L Gu, B Yan, Y Liu, Y Liu 1989 J. Trace Elem. Electrolyes Health Dis y.\22 Y Xia, KE Hill, DW Byrne, J Xu, RF Burk 2005 Am J Clin Nutr 81:829

410

Selenium: Its molecular biology and role in human health

18. GV Kryukov, S Castellano, SV Novoselov, AV Lobanov, O Zehtab, R Guigo, VN Gladyshev 2003 Science 300:1439 19. J-G Yang, KE Hill, RF Burk \9S9 JNutr 119:1010 20. PR Taylor, P Greenwald 2005 J Clin Oncol 23:333 21. CIp 1998JiV«/r 128:1845 22. PR Trumbo 2005 JNutr 135:354 23. LC Clark, GF Combs Jr, BW Tumbull, EH Slate, DK Chalker, J Chow, LS Davis, RA Glover, GF Graham, EG Gross, A Krongrad, JL Lesher, HK Park, BB Sanders, CL Smith, JR Taylor 1996 JAMA 276:1957 24. SM Lippman, PJ Goodman, EA Klein, HL Pames, IM Thompson Jr., AR Kristal, RM Santella, JL Probstfield, CM Moinpour, D Albanes, PR Taylor, LM Minasian, A Hoque, SM Thomas, JJ Crowley, JM Gaziano, JL Stanford, ED Cook, NE Fleshner, MM Lieber, PJ Walther, FR Khuri, DD Karp, GG Schwartz, LG Ford, CA Coltman Jr 2005 J Natl Cancer Inst 97-M

Index

Acquired immunodeficiency syndrome, 299-310 immunity, 300-302 metabolic syndrome, 302-304 micronutrient deficiencies, 304-305 selenium deficiency, 305-306 selenium supplementation, 306-308 wasting, 302-304 Adhesion molecules, cytokines, 314-315 Aging methionine sulfoxide reduction, 126-127 selenium in, 318-320 AIDS. See Acquired immunodeficiency syndrome Aminoacyl-tRNA recycling, 83-95 Antibodies, mimics of selenoprotein, selenoenzyme transformation, 394-395 Apoptosis, selenium-induced, 379-385 mitochondrial dysfunction, selenium-induced, 382-384 oxidative stress, selenium-induced, 381-382 safe levels of exposure, 380-381 selenium toxicity, 380 thiol modification, selenium-induced, molecules targeted for, 382 Atom detection, selenium, 224-225 Bacteria biosynthesis of selenocysteine, 14-15 decoding UGA with selenocysteine, 24-25 E. coli SECIS element, in gene expression, 16-11 incorporation of selenocysteine by, 12 SECIS element interaction, 18-21 SelB, domain structure, 18-21 termination vs. readthrough, 11-14 translation factor, SelB, 15-18 tRNAS'^'', 12-13 Bioinformatics tools for, selenoprotein identification, 99-102 Bios)mthesis of selenocysteine, 14—15 Brain function

effects of selenoprotein P deletion, 115 epilepsy, 238-239 neurodegenerative disorders, 239-241 Parkinson's disease, 236-237 selenium selenoproteins, 233-248 stroke, 234-236 transgenic selenoprotein-deficient mouse models, 241-245 Cancer prevention, 249-264, 367-368 animal models, 250 clinical trials, 250-252 epidemiological evidence, 250 15-kDa selenoprotein, 141-148 in cancer prevention, 142-143 dietary selenium, 146 glycoprotein glucosyltransferase, 143-142 thiol-disulfide oxidoreductase function, 145-146 mechanisms, 252-254 metabolic bases, 254-260 metabolomics, 373-374 nutrigenetics, 369-370 nutritional epigenetics, 370-371 nutritional transcriptomics, 371-372 proteomics, 373-374 Se-metabolities, 255-260 selenium dietary standards, 408^09 selenoenzymes, 254—255 selenoprotein gene variation, 277-286 GPx-1 in cancer etiology, 279-282 polymorphisms, 283-284 selenoproteins, 279 Sepl5, 282-283 thiol proteomics, 265-276 BiP/GRP78 over-expression, 272-273 methylseleninic acid, 266 monomethylated selenium, as protein redox modulator, 266-269 redox-modified proteins, 269 unfolded protein response, ER stress and, 269-271

412

Selenium: Its molecular biology and role in human health

UPR signaling, 271-272 Catalytic mechanisms, methionine sulfoxide reduction, 128 Cell-mediated immunity, 315-316 Conditional knockout mouse models, 339-340 Conformation-specific SEClS-binding activities, SBP2, eukaryotic selenoprotein synthesis, 69-70 Coxsackievirus, 287-290 Cys-containing counterparts, methionine sulfoxide reduction. Sec-containing proteins compared, 131 Cytoplasmic supramolecular complex, supramolecular complexes, selenocysteine biosynthesis, isolation, 89-91 Cytosolic, mitochondrial thioredoxin reductase knockout mice, 195-206 phenotype Txnrdl knockout embryos, 199-200 Txnrd2 knockout embryos, 201-202 Txnrdl/Txnrd2 embryonic expression profile, 198-199 embryonic lethality, 198 heart-specific inactivation of, 202-204 mouse models with conditional alleles for, 197-198 Decoding selenocysteine, 39-50 genetic code, 4 8 ^ 9 phenotype, dynamic process of evolution, 45^6 Sec decoding common origin, 40-41 lost trait, 41-42 Sec-tRNA^'^'' synthesis, non-canonical mechanism, 46-48 selenophosphate synthetase, 42-45 Decoding selenophosphate synthetase trait, 41^2 Decoding UGA with selenocysteine, 24-25 Deiodinases, endocrine function, 207-219 adaptive thermogenesis, 212-214 Dl overexpression, h5fperthyroidism, 217 D3 overexpression in hemangiomas, 216 deiodinases conservation, 3D structure, 208 fasting, changes in iodothyronine deiodination, 214-215 illness, changes in iodothyronine deiodination, 214-215 thyroid hormone homeostasis, 211

tissue-specific control of thyroid hormone action, 211-212 ubiquitination pathway, D2 inactivation, 209-211 Detection of selenium atom, 224-225 Diabetes, 173-182 early research, 175-176 glutathione peroxidase-1 in, 173-182 early research, 175-176 insulin function, 180 metabolic impact, 175 selenoprotein expression, 174-175 insulin function, 180 metabolic impact, 175 Dietary standards, selenium, 399^10 cancer chemoprevention, 408^09 dietary reference intakes, 404-407 glutathione peroxidase, as biomarker, 407-408 RDAs, 400-401 selenoprotein P, as biomarker, 407^08 World Health Organization, 401-404 Domain structure, SelB, 18-21 Drosophila, selenoproteins, 343-353 SelG/SelK/G-rich,351 SelM/BthD/SelH, 350-351 synthesis machinery, 344-348 intake, 344 selenocysteyl-tRNA, 344-345 selenophosphate synthetase, 345-348 translational machinery, 348-349 Drug development, parasite selenoproteins, 364-365 E. coll SECIS element, in gene expression, 26-27 EEFSec, SBP2 interactions, eukaryotic selenoprotein synthesis, 66-67 Eicosanoid metabolism, 313-314 Endocrine function, deiodinases and, 207-219 adaptive thermogenesis, 212-214 Dl overexpression, hyperthyroidism, 217 D3 overexpression, hemangiomas, 216 deiodinases conservation, 3D structure, 208 fasting, changes, iodothyronine deiodination, 214—215 illness, changes, iodothyronine deiodination, 214-215 thyroid hormone homeostasis, 211 tissue-specific control of thyroid hormone action, 211-212

Index ubiquitination pathway, D2 inactivation, 209-211 Endogenous Sec factors, supramolecular complexes, selenocysteine biosynthesis, supramolecular complexes composed of, 92 En2ymes, natural, mimics of selenoprotein, selenoenzyme transformation, 392-393 Epilepsy, selenium selenoproteins, 238-239 Eukaryotic Sec bios}Tithesis, supramolecular complexes, selenocysteine biosynthesis, protein factors, 86-87 Eukaryotic selenocysteine tRNAs, 29-37 biosynthesis, 32-34 evolution, insertion machinery, 34—36 insertion machinery, 34-36 mammalian Sec tRNA[s«]s<:<:, 30-31 methylase, Um34, 31-32 occurrence, 34-36 Um34, 30-31 Eukaryotic selenoproteins, 104—105 binding proteins, 63-72 fish 15 kDa selenoprotein, 105 functional domain, 65 L30, conformation-specific SECIS-binding activities, 69-70 methionine-S-sulfoxide reductase, 104-105 N-terminal domain, 64—65 Plasmodium selenoproteins, 105 previously known functions, 68 protein disulfide isomerase, 105 ribosomal protein L30, 67-68 RNA-binding domain, 69 RNA-binding properties, 69 RNA-binding specificities, molecular bases, 70 SBP2, 64 eEFSec, 66-67 ribosome, 67 self association, 66 SECIS binding domain, 65-66 Sell, 105 Sel2, 105 SeI3, 105 Sel4, 105 selenoprotein J, 105 selenoprotein U, 105 UGA recoding, 68-69 eukaryotes, mechanism, 70-71 Evolution of selenocysteine, 39-50 genetic code, 48-49

413 phenotype, dynamic process of evolution, 45-46 Sec decoding, common origin, 40-41 Sec decoding trait, lost, 41-42 Sec-tRNA^" synthesis, non-canonical mechanism, 46-48 selenophosphate synthetase, 4 2 ^ 5 Extracellular peroxide tone, selenoproteins of glutathione system, pGPx and, 166-167 Fasting, changes in iodothyronine deiodination, 214-215 15-kDa selenoprotein, 103, 105 cancer prevention, 141-148 dietary selenium, 146 gene variation, 282-283 glycoprotein glucosyltransferase, 143-142 thiol-disulfide oxidoreductase fiinction, 145-146 dietary selenium, 146 glycoprotein glucosyltransferase, 143-142 thiol-disulfide oxidoreductase function, 145-146 Fly selenoproteins, 343-353 Formate dehydrogenase, 105 Formylmethanofiiran dehydrogenase, 105 Gene deletion, selenoprotein P, in mouse, 111-122 brain, effects of selenoprotein P deletion, 115 research outlook, 120-121 selenium disposition effects, 113-115 testis, effects of selenoprotein P deletion, 116-120 Genetic code, selenophosphate synthetase, 48^9 Genetic engineering, mimics of selenoprotein, proteins transformation, 396 Genital system, selenoproteins in, 325-327 Glutaredoxin, 106 Glutathione peroxidase, 102, 359-360 as biomarker, 407-408 Se-independent regulation, 163-165 cGPx, 163-164 GI-GPx, 164 pGPx, 164 PHGPx, 164-165 Glutathione peroxidase-1 biological selenium buffer, 158-159

414

Selenium: Its molecular biology and role in human health

early research, 175-176 expression regulation, 149-160 insulin function, 180 in lactation, 153-154 metabolic impact, 175 mild oxidative stress, 177-179 molecular biology markers, in selenium status assessment, 159 mRNA stability, 154-155 other selenoproteins, 153 in oxidative stress, diabetes, 173-182 in pregnancy, 153-154 ROS vs. RNS, coping with, 179-180 selenoprotein expression, 174-175 selenoprotein transcript abundance, 155-157 severe oxidative stress, 176-177 translation, selenium regulation of, 157-158 vitamin E, 179 Glutathione selenoproteins, 161-172 differential response to selenium, 162-163 enzymatic characteristics, 162 extracellular peroxide tone, pGPx and, 166-167 functional diversification, 165 inflammatory responses of intestine, GI-GPx as modulator of, 167-168 knock out and overexpression, 165-166 lipid peroxidation, inflammation, differentiation, PHGPx in, 168-170 Se-independent regulation, 163-165 cGPx, 163-164 GI-GPx, 164 pGPx, 164 PHGPx, 164-165 Glycoprotein glucosyltransferase, 15-kDa selenoprotein, in cancer, 143-142 Heart-specific inactivation, cytosolic, mitochondrial thioredoxin reductase, Txnrdl/Txnrd2 knockout mice, 202-204 Historical perspective of selenium, 1-6 HIV, 299-310 immunity, 300-302 metabolic syndrome, 302-304 micronutrient deficiencies, 304-305 selenium deficiency, 305-306 selenium supplementation, 306-308 wasting, 302-304 Hydrogenase, 105 Hyperthyroidism, deiodinases and, 217

Immunity, 315-316 roles of selenium in, 311-322 selenium in, 311-322 In vivo protein, selenocysteine biosjoithesis, supramolecular complexes, 88-89 Inflammatory responses of intestine, GI-GPx as modulator of, 167-168 Influenza virus, 292-295 Insertion machinery, tRNAs, eukaryotic selenocysteine, 34-36 Interleukin-2 receptor, lymphocytes, 316-317 Intestine, inflammatory responses, GI-GPx as modulator, 167-168 K-tum binding proteins, SECIS RNAs, 51-61 phylogenetic conservation, 57-60 Knockout mouse models, 338 L30, conformation-specific SECIS-binding activities, 69-70 Lactation, glutathione peroxidase-1 expression regulation in, 153-154 Lipid peroxidation, inflammation, differentiation, PHGPx in, 168-170 Male reproduction, 323-331 genital system, selenoproteins in, 325-327 PHGPx moonlighting, sperm maturation, 327-329 sperm in selenium deficiency, 324 Metabolic syndrome, AIDS/HIV, 302-304 Metabolomics in cancer, 373-374 Methionine-R-sulfoxide reductase 1, 103 Methionine-S-sulfoxide reductase, 104-105 Methionine sulfoxide reduction, 123-133 active site features, non-selenoprotein MsrB, 129-130 aging, 126-127 catalytic mechanisms, 128 Cys-containing counterparts, Sec-containing proteins compared, 131 evolutionary implications, 131 methionine sulfoxide reductases, 124-125 mammalian, 125-126 non-selenoprotein mammalian MsrBs, selenoprotein, 130-131 physiological roles of, 126 selenium in, 127-128 selenoprotein forms of, 127

415

Index Methylase, eukaryotic selenocysteine, Um34, 31-32 Methylseleninic acid, cancer and, 266 Micronutrient deficiencies, AIDS/HIV, 304-305 Mimics of selenoprotein, 387-398 antibodies, selenoenzyme transformation, 394^395 chemical modification, selenoenzyme transformation by, 392-396 genetic engineering, proteins transformation, 396 natural enzymes, selenoenzyme transformation, 392-393 natural proteins, selenoenzyme transformation, 394 organoselenium-based selenoenzyme mimics, 389-392 Mitochondrial, cytosolic thioredoxin reductase knockout mice, 195-206 phenot5rpe Txnrdl knockout embryos, 199-200 Txnrd2 knockout embryos, 201-202 Txnrdl/Txnrd2 embryonic expression profile, 198-199 embryonic lethality, 198 heart-specific inactivation of, 202-204 mouse models with conditional alleles for, 197-198 Mitochondrial dysfunction, selenium-induced, 382-384 Molecular biology markers, in selenium status assessment, 159 mRNA stability, glutathione peroxidase-1 expression regulation, 154-155 N-terminal domain, eukaryotic selenoprotein synthesis, 64-65 Nematode parasites, selenoproteins of, 363 Neurodegenerative disorders, selenium selenoproteins, 239-241 Non-canonical mechanism, Sec-tRNA^"' synthesis, 46-48 Non-selenoprotein methionine sulfoxide reduction, active site features, 129-130 Nuclear supramolecular complex, supramolecular complexes, selenocysteine biosjTithesis, isolation, 89-91 Nucleocytoplasmic shuttling, SBP2, EFsec, subcellular localization, 78 Nutrigenetics, in cancer, 369-370 Nutritional epigenetics, in cancer, 370-371

Nutritional transcriptomics, in cancer, 371-372 Organoselenium-based selenoenzyme mimics, mimics of selenoprotein, 389-392 Origin, selenophosphate synthetase, 40-41 Oxidative stress, 290-292, 381-382 glutathione peroxidase-1 in, 173-182 early research, 175-176 mild oxidative stress, 177-179 severe oxidative stress, 176-177 selenoprotein W in, 138-140 promoter studies, 136 Parasites, selenoproteins in, 355-366 drug development, parasite selenoproteins, 364—365 glutathione peroxidase, 359-360 nematode parasites, selenoproteins of, 363 oxidative stress, 356-359 platyhelminth parasites, selenoenzyme GSH-reduction pathways, 360-363 Trx-reduction pathways, 360-363 protozoan parasites, selenoproteins of, 364 vaccine development, parasite selenoproteins, 364—365 Parkinson's disease, selenium selenoproteins, 236-237 Peroxiredoxin, 106 Peroxynitrite reductases, seleno-enzymes as, 312-313 PHGPx moonlighting maturation of sperm, 327-329 sperm maturation, 327-329 Plasmodium selenoproteins, 105 Platyhelminth parasites, selenoenzyme GSH-reduction pathways, 360-363 Trx-reduction pathways, 360-363 PoKovirus, 295 Polymorphisms, selenoprotein gene, cancer and, 283-284 Pregnancy, glutathione peroxidase-1 expression regulation in, 153-154 Previously known ftmctions, eukaryotic selenoprotein synthesis, 68 Prokaryotes, selenium metabolism in, 9-28 Prokaryotic selenoproteins, 105-106 formate dehydrogenase, 105 formylmethanofuran dehydrogenase, 105 glutaredoxin, 106 HesB-like, 106 hydrogenase, 105

416

Selenium: Its molecular biology and role in human health

peroxiredoxin, 106 selenoprotein A, 106 selenoprotein B, 106 thioredoxin, 106 Protein disulfide isomerase, 105 Protein interactions, supramolecular complexes, selenocysteine biosynthesis. Sec biosynthesis/incorporation, 88 Proteins, natural, mimics of selenoprotein, selenoenzyme transformation, 394 Proteomics, in cancer, 373-374 Protozoan parasites, selenoproteins of, 364 Radiolabeling in selenocysteine biotechnology, 224-225 RDAs, selenium dietary standards, 400-401 Readthrough, vs. termination, 21-24 Reference intakes, dietary, selenium, 404-407 Reproduction, male, 323-331 genital system, selenoproteins in, 325-327 PHGPx moonlighting, sperm maturation, 327-329 sperm in selenium deficiency, 324 Ribosomal protein L30, 63-72 eukaryotic selenoprotein synthesis, 67-68 Ribosome eukaryotic selenoprotein synthesis SBP2 interactions, 67 supramolecular complexes, selenocysteine biosynthesis, Sec factors associated with, 92 RNA-binding domain, eukaryotic selenoprotein synthesis, 69 RNA-binding properties, eukaryotic selenoprotein synthesis, 69 RNA-binding specificities, molecular bases, 70 RNA interactions SECIS, K-tum binding proteins, protein motifs, 51-61 supramolecular complexes, selenocysteine biosynthesis. Sec biosynthesis/incorporation, eukaryotic protein, 87 Roles of selenium in immunity, 311-322 SBP2 conformation-specific SEClS-binding activities, 69-70 EFsec for selenoprotein synthesis, subcellular localization, 73-82

eukaryotic selenoprotein synthesis, 64 SECIS binding domain, eukaryotic selenoprotein synthesis, 65-66 Sel-tag, in selenocysteine biotechnology, 227-229 Selenocysteine antioxidant selenoproteins, 388-389 biotechnology, 221-230 catalytic advantages, disadvantages, 107 conjugation-based applications, 227 cytoplasmic supramolecular complex, isolation, 89-91 decoding common origin, 4 0 ^ 1 lost trait, 4 1 ^ 2 detection of selenium atom, 224—225 endogenous Sec factors, supramolecular complexes composed of, 92 eukaryotic Sec biosynthesis, protein factors, 86-87 eukaryotic Sec incorporation, protein factors, 87 evolution, 39-50 genetic code, 48-49 human health, selenium in, 84 incorporation strategies, 85-86 nuclear supramolecular complex, isolation, 89-91 phenotype, dynamic process of evolution, 45^6 protein interactions in. Sec biosynthesis/incorporation, 88 radiolabeling, 224-225 ribosomes. Sec factors associated with, 92 RNA interactions. Sec biosynthesis/incorporation, eukaryotic protein, 87 SECp43, 91-92 Sel-tag, applications based on, 227-229 selenophosphate synthetase, 4 2 ^ 5 selenoproteins, 84^85 production, 221-224 supramolecular complexes, 83-95 tRNA-channeling, Sec cycling, comparisons, 92-93 tRNA^'"^ synthesis, non-canonical mechanism, 46-48 in vitro protein, 88 in vivo protein, 88-89 Selenocysteyl-tRNA, 344-345 Selenoenzymes cancer, 254—255 as peroxynitrite reductases, 312-313

Index Selenophosphate synthetase, 39-50 decoding common origin, 4 0 ^ 1 lost trait, 4 1 ^ 2 genetic code, 48-49 phenotype, dynamic process of evolution, 45^6 selenophosphate synthetase, 4 2 ^ 5 tRNA^'^'^ synthesis, non-canonical mechanism, 46-48 Selenophosphate synthetase 2, 103 Selenoprotein A, 106 Selenoproteins, 106 Selenoprotein H, 104 Selenoprotein I, 104 Selenoprotein J, 105 Selenoprotein K, 104 Selenoprotein M, 104 Selenoprotein mimics, 387-398 antibodies, selenoenzyme transformation, 394-395 chemical modification, selenoenzyme transformation by, 392-396 genetic engineering, proteins transformation, 396 natural enzymes, selenoenzyme transformation, 392-393 natural proteins, selenoenzyme transformation, 394 organoselenium-based selenoenzyme mimics, 389-392 Selenoprotein N, 104 Selenoprotein O, 104 Selenoprotein P, 103 as biomarker, 407-408 gene deletion brain, effects of selenoprotein P deletion, 115 mouse, 111-122 research outlook, 120-121 selenium disposition effects, 113-115 testis, effects of selenoprotein P deletion, 116-120 Selenoprotein S, 104 Selenoprotein!, 104 Selenoprotein U, 105 Selenoprotein V, 104 Selenoprotein W, 103-104 in development, oxidative stress, 135-140 GSHand, 137 in human tissues, 137 in oxidative stress, 138-140 promoter studies, 136

All Selenoproteins, 99-110 15 kDa selenoprotein (Sec 15), 103 bioinformatics tools for, selenoprotein identification, 99-102 in cancer, gene variation, 279 eukaryotic selenoproteins, 104—105 fish 15 kDa selenoprotein, 105 methionine-S-sulfoxide reductase, 104-105 Plasmodium selenoproteins, 105 protein disulfide isomerase, 105 Sell, 105 Sel2, 105 Sel3, 105 Sel4, 105 selenoprotein J, 105 selenoprotein U, 105 formate dehydrogenase, 105 formylmethanofiiran dehydrogenase, 105 fiinctions, 106-107 glutaredoxin, 106 glutathione peroxidases, 102 HesB-like, 106 hydrogenase, 105 mammahan selenoproteins, 102-104 methionine-R-sulfoxide reductase 1, 103 methionine sulfoxide reduction, 130-131 peroxiredoxin, 106 production, 221-224 prokaryotic selenoproteins, 105-106 selenophosphate synthetase 2,103 thioredoxin, 106 thioredoxin reductases, 102-103 thyroid hormone deiodinases, 102 Selenoproteins of glutathione system, 161-172 differential response to selenium, 162-163 enzymatic characteristics, 162 extracellular peroxide tone, pGPx and, 166-167 glutathione peroxidases functional diversification, 165 Se-independent regulation, 163-165 inflammation, PHGPx in, 168-170 inflammatory responses of intestine, GI-GPx as modulator of, 167-168 knockout, 165-166 lipid peroxidation inflammation, differentiation, PHGPx in, 168-170 PHGPx in, 168-170 overexpression, 165-166 PHGPx in, 168-170

418

Selenium: Its molecular biology and role in human health

Selenoproteins of thioredoxin system, 183-194 gene targeting, 192-193 general properties, 184—185 isoenzymes, 190-191 medical aspects of selenium in, 191-192 selenium reduction, 185-186 structure, mechanism, 187-190 substrate specificity, thioredoxin reductase, 186-187 Selenoproteomes, 107-108 Self association, SBP2 interactions, eukaryotic selenoprotein synthesis, 66 Skin, selenium in, 318 Sperm maturation of, PHGPx moonlighting, 327-329 selenium deficiency, 324 Stroke, selenium selenoproteins, 234-236 Subcellular localization, SBP2 EFsec localization, 75-76,78-81 NES,77 NLS, 77 selenoprotein synthesis, 73-82 selenoproteins, mediated decay, 74-75 EFsec for selenoprotein synthesis, 73-82 localization, SBP2, 78-81 NLS, in SBP2 and EFsec, 77 nucleocytoplasmic shuttling, 78 selenoproteins, mediated decay, 74—75 Sulfoxide reduction, methionine, 123-133 active site features, non-selenoprotein MsrB, 129-130 aging, 126-127 catalytic mechanisms, 128 Cys-containing counterparts. Sec-containing proteins compared, 131 evolutionary implications, 131 methionine sulfoxide reductases, 124-125 mammalian, 125-126 non-selenoprotein mammalian MsrBs, selenoprotein, 130-131 physiological roles of, 126 selenium in, 127-128 selenoprotein forms of, 127 Supramolecular complexes, selenocysteine biosynthesis, 83-95 cytoplasmic supramolecular complex, isolation, 89-91 endogenous Sec factors, supramolecular complexes composed of, 92

eukaryotic Sec biosynthesis, protein factors, 86-87 eukaryotic Sec incorporation, protein factors, 87 human health, selenium in, 84 nuclear supramolecular complex, isolation, 89-91 protein interactions in. Sec biosynthesis/incorporation, 88 ribosomes. Sec factors associated with, 92 RNA interactions. Sec biosynthesis/incorporation, eukaryotic protein, 87 Sec incorporation strategies, 85-86 SECp43, 91-92 selenoproteins, 84-85 tRNA-channeling, Sec cycling, comparisons, 92-93 in vitro protein, 88 in vivo protein, 88-89 Synthetic selenoproteins, strategies to producing, 221-230 Termination vs. readthrough, ll-l'^ Testis, effects of selenoprotein P deletion, 116-120 Thiol-disulfide oxidoreductase function, 15-kDa selenoprotein, in cancer, 145-146 Thiol modification, selenium-induced, molecules targeted for, 382 Thiol proteomics, in cancer prevention, 265-276 BiP/GRP78 over-expression, 272-273 methylseleninic acid, 266 monomethylated selenium, as protein redox modulator, 266-269 redox-modified proteins, 269 unfolded protein response, ER stress and, 269-271 UPR signaling, 271-272 Thioredoxin, 106 reductases, 102-103 selenoproteins, 183-194 gene targeting, 192-193 general properties, 184—185 isoenzymes, 190-191 medical aspects of selenium in, 191-192 selenium reduction, 185-186 structure, mechanism, 187-190 substrate specificity, thioredoxin reductase, 186-187

Index Thyroid hormone, deiodinases, 102, 211-212 Toxicity, selenium, 380 Transcript abundance, selenoprotein, glutathione peroxidase-1 expression regulation, 155-157 Transgenic/conditional knockout mouse models, 340-341 Transgenic/knockout mouse models, 338 Transgenic mouse models, 335-338 Translation factor, SelB, 15-18 tRNAs biosjTithesis, 32-34 eukaryotic selenocysteine, 29-37 evolution, insertion machinery, 34—36 insertion machinery, 34-36 mammalian Sec tRNA'S'^'-lScc, 30-31 methylase, Um34, 31-32 occurrence, 34-36 selenocysteine biosynthesis, supramolecular complexes, Sec cychng, comparisons, 92-93

419 Um34, 30-31 TRNAS'='>, 12-13 Trx-reduction pathways, platyhelminth parasites, selenoenzyme, 360-363 UGA receding, 68-69 eukaryotes, mechanism, 70-71 Vaccine development, parasite selenoproteins, 364—365 Viral infections, 287-298 coxsackievirus, 287-290 influenza virus, 292-295 oxidative stress, 290-292 poliovirus, 295 Vitamin E, glutathione peroxidase-1, in oxidative stress, diabetes, 179 Wasting, HlV/AlDS, 302-304 World Health Organization, selenium dietary standard, 401^04