Membrane Protein Transport, Volume 2 (Membrane Protein Transport) (Membrane Protein Transport)

MEMBRANE PROTEIN TRANSPORT A Multi-Volume Treatise Volume 2 • 1995

This Page Intentionally Left Blank

MEMBRANE PROTEIN TRANSPORT A Multi-Volume Treatise Editor:

STEPHEN S. ROTHMAN University of California San Francisco, California

VOLUME 2 • 1995

JAI PRESS INC. Greenwicli, Connecticut

London, England

Copyright © 7 995 JAI PRESS INC. 55 Old Post Road, No. 2 Greenwich, Connecticut 06836 JAI PRESS LTD. The Courtyard 28 High Street Hampton Hill, Middlesex TW12 IPD England All rights reserved. No part of this publication may be reproduced, stored on a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, filming, recording, or otherwise, without prior permission in writing from the publisher. ISBN: U55938-983-4 Manufactured in the United States of America

CONTENTS LIST OF CONTRIBUTORS THE ROLE OF MOLECULAR CHAPERONES IN TRANSPORT OF PROTEINS ACROSS MEMBRANES Elizabeth A. Craig, B. Diane Gambill, Wolfgang Voos, and Nikolaus Pfanner THE NUCLEAR PORE COMPLEX IN YEAST Paola Grandi and Eduard C. Hurt ATP BINDING CASSETTE TRANSPORTERS IN YEAST: FROM MATING TO MULTIDRUG RESISTANCE Ralf Egner, Yannick Make, Rudy Pandjaitan, Veronika Huter, Andrea Lamprecht, and Karl Kuchler

vii

1

29

57

THE APICAL SORTING OF GLYCOSYLPHOSPHATIDYLINOSITOL-LINKED PROTEINS Michael P. Lisanti, ZhaoLan Tang, Philipp E. Scherer, and Massimo Sargiacomo CAVEOLAE: PORTALS FOR TRANSMEMBRANE SIGNALING AND CELLULAR TRANSPORT Michael P. Lisanti, ZhaoLan Tang, and Massimo Sargiacomo NUCLEAR TRANSPORT OF URACIL-RICH SMALL NUCLEAR RIBONUCLEOPROTEIN PARTICLES Elisa Izaurralde, lain W. Mattaj, and David S. Goldfarb

97

111

123

vi

CONTENTS

PHOSPHORYLATION-MEDIATED REGULATION OF SIGNAL-DEPENDENT NUCLEAR PROTEIN TRANSPORT: THE "CcN MOTIF" David A. Jans

161

MEMBRANE PROTEIN TOPOGENESIS IN ESCHERICHIA COLI Gunnar von Heijne

201

MODEL FOR INTEGRATING P-TYPE ATPases INTO ENDOPLASMIC RETICULUM Randolph Addison andjialing Lin

215

NUCLEAR TRANSPORT AS A FUNCTION OF CELLULAR ACTIVITY Carl M. Feldherr and Debra Akin

237

INDEX

261

LIST OF CONTRIBUTORS Randolph Addison

Department of Biochemistry Health Science Center University of Tennessee Memphis, Tennessee

Debra Akin

Department of Anatomy and Cell Biology College of Medicine University of Florida Gainsville, Florida

Elizabeth A. Craig

Department of Biomolecular Chemistry University of Wisconsin Madison, Wisconsin

Carl M. Feldherr

Department of Anatomy and Cell Biology College of Medicine University of Florida Gainsville, Florida

B. Diane Cambill

Department of Biomolecular Chemistry University of Wisconsin Madison, Wisconsin

David S. Goldfarb

Department of Biology University of Rochester Rochester, New York

Paola Grandi

European Molecular Biology Laboratory Heidelberg, Germany

RalfEgner

Department of Molecular Genetics University and Biocenter of Vienna Vienna, Austria

VII

VIM

LIST OF CONTRIBUTORS

Eduard C. Hurt

European Molecular Biology Laboratory Heidelberg, Germany

Veronika Huter

Department of Molecular Genetics LJniversity and Biocenter of Vienna Vienna, Austria

Elisa Izaurralde

European Molecular Biology Laboratory Heidelberg, Germany

David A. Jans

Division for Biochemistry and Molecular Biology John Curtin School of Medical Research Australian National University Canberra, Australia

Karl Kuchler

Department of Molecular Genetics University and Biocenter of Vienna Vienna, Austria

Andrea Lamprecht

Department of Molecular Genetics University and Biocenter of Vienna Vienna, Austria

Jianling Lin

Department of Biochemistry Health Science Center University of Tennessee Memphis, Tennessee

Michael P. Lisanti

The Whitehead Institute for Biomedical Research Cambridge, Massachusetts

Yannick Mah6

Department of Molecular Genetics University and Biocenter of Vienna Vienna, Austria

lain W. Mattaj

European Molecular Biology Laboratory Heidelberg, Germany

Rudy Pandjaitan

Department of Molecular Genetics University and Biocenter of Vienna Vienna, Austria

List of Contributors

IX

Nikolaus Pfanner

Biochemisches Institut Universitat Freiburg Freiburg, Germany

Massimo Sargiacomo

Department of Hematology and Oncology Instituto Superiore di Sanita Rome, Italy

Philipp E. Scherer

The Whitehead Institute for Biomedical Research Cambridge, Massachusetts

ZhaoLan Tang

The Whithehead Institute for Biomedical Research Cambridge, Massachusetts

Gunnar von Heijne

Department of Biochemistry Arrhenius Laboratory Stockholm University Stockholm, Sweden

Wolfgang Voos

Biochemisches Institut Universitat Freiburg Freiburg, Germany

This Page Intentionally Left Blank

THE ROLE OF MOLECULAR CHAPERONES IN TRANSPORT OF PROTEINS ACROSS MEMBRANES

Elizabeth A. Craig, B. Diane Gambill, Wolfgang Voos, and Nikolaus Pfanner

I. Introduction 2 A. hsp70 Structure and Biochemical Properties 2 B. ThchsplOs of Saccharomyces cerevisiae 3 C. The Basics of Import into Mitochondria 4 II. A Role for CytosolichspTO in Protein Translocation into Organelles . . 4 III. Role ofMitochondrialhspTO in Protein Translocation 5 A. In vivo Analysis of a Mitochondrial hsp70 Mutant 5 B. Mitochondrial hsp70 Is Required for Translocation into the Matrix 7 C. Degree of Requirement of Ssc 1 p Activity in Transport Is Correlated with the Conformation of the Precursor 14 D. Model of hsp70 Action in Translocation Across Mitochondrial Membranes 18 Membrane Protein Transport Volume 2, pages 1-28 Copyright © 1995 by JAI Press Inc. All rights of reproduction in any form reserved. ISBN: 1-55938-983-4

2

E. A. CRAIG, B. D. GAMBILL, W. VOOS, and N. PFANNER

E. Requirements for Mitochondrial hsp70 in the Translocation of Proteins into the Matrix Compared to the Intermembrane Space 20 F. Components Other Than Mitochondrial hsp70 Required for Protein Movement across the Outer and Inner Membranes . . 22 IV. Role of ATP in Protein Translocation 23 V. Role of hsp70 of the Endoplasmic Reticulum in Protein Translocation 24 VI. Summary 25 References 25 I.

INTRODUCTION

A. hsp70 Structure and Biochemical Properties The process of transporting proteins from their site of synthesis in the cytosol into organelles such as the mitochondrion or endoplasmic reticulum is facilitated by a group of proteins known as the molecular chaperones. Among the best characterized molecular chaperones are the hsp70 proteins. The hsp70s are highly conserved polypeptides found in all species, with identity ranging from 50% between procaryotic and eucaryotic examples to as high as 99% between members of a HSP70 gene family in yeast (Boorstein et al, 1994). The hspTOs possess two biochemical activities: a weak ATPase activity and peptide binding activity (reviewed in Gething and Sambrook, 1992). The ATPase activity can be stimulated severalfold by binding the hsp70 to peptides or polypeptides. Biochemical and structural studies of mammalian cytosolic hsp70 have shown that an N-terminal 44-kDa proteolytic fragment has ATPase activity that cannot be stimulated by peptide binding. The structure of the N-terminal proteolytic fragment of mammalian HSP70 has been solved by X-ray crystallography (Flaherty et al., 1990). The roughly U-shaped structure consists of two lobes separated by a deep cleft in which ATP binds. The most highly conserved amino acids of the amino-terminal domain are found lining this cleft. The peptide binding activity resides in the carboxyl-terminal portion of the protein. It should be emphasized that there is no "consensus" binding site of hsp70s, inasmuch as they are able to interact with a wide variety of peptides. However, studies of a mammalian hsp70 indicated a preference for binding of peptides rich in aliphatic and hydrophobic amino acids that were at least seven residues in length (Flynn et al., 1991). The cycles of binding and release of peptide require the hydrolysis

Molecular Chaperones and Protein Transport

3

of ATP, implying an interaction between the C-terminal peptide binding domain and the N-terminal ATPase domain. The structure of the C-terminal domain of hsp70 has not been solved; however, two groups have proposed a structure similar to that of the well-characterized major histocompatibility complex class I antigen presenting (HLA) molecules, based on slight similarities in primary sequence and secondary-structure predictions (Rippmann et al., 1991; Flajnik et al., 1991). The HLA structure and the proposed hsp70 peptide binding region consist primarily of P sheets with peptides binding in an extended conformation in the groove bounded by a helices (Fremont et al., 1992). The universal abihty of hsp70s to undergo cycles of binding and release with short regions of peptides determines their role in a great variety of intracellular functions. In this chapter we will discuss how, based on analysis of mutants of the yeast Saccharomyces cerevisiae, hspTOs function in the translocation of proteins from the cytosol into mitochondria and the endoplasmic reticulum has begun to be elucidated. B. The hsp70s of Saccharomyces cerevisiae

Like other eucaryotes, the budding yeast Saccharomyces cerevisiae has multiple hsp70 proteins. This large gene family is divided into subfamilies that correlate with the subcellular localization of the gene products (reviewed in Craig et al., 1993). The SSA and SSB subfamilies of proteins are found in the cytosol. The SSB subfamily has two members that are not absolutely essential for growth. This subfamily of hspTOs is found in association with translating ribosomes, perhaps bound to the nascent polypeptide chain. The four SSA gene products are found dissolved in the cytosol. The SSA subfamily is essential; cells require at least one of the SSA proteins in substantial amounts for growth at any temperature. SSA proteins function to regulate the transcription of heat shock responsive genes (Stone and Craig, 1990) and as discussed below are important for efficient translocation of at least some proteins from the cytosol into the ER and mitochondria. The other two yeast hsp70s are organellar. Kar2p is an essential protein found in the lumen of the endoplasmic reticulum (Rose et al., 1989; Normington et al., 1989). The mitochondrial hsp70, Ssclp, is a soluble protein of the matrix and is essential for growth (Craig et al., 1987, 1989). As discussed below both organellar hsp70s are required for transport of proteins into their respective organelles. Here, we focus on the role of mitochondrial hsp70 in translocation since it is the subject of work from our laboratories.

4

E. A. CRAIG, B. D. GAMBILL, W. VOOS, and N. PFANNER

C. The Basics of Import into Mitochondria The biogenesis of mitochondrial proteins is a multistep process (reviewed in Pfanner et al., 1991; Click and Schatz, 1991). Precursor proteins synthesized in the cytosol with an N-terminal presequence and probably bound to cytosolic hsp70s associate with receptors on the mitochondrial surface. The presequence is then inserted into the outer membrane and then across the inner membrane triggered by the membrane potential, A\|/, across the inner membrane. The presequence is proteolytically removed by the processing protease of the mitochondrial matrix. Then the bulk of the polypeptide moves vectorially across the outer and inner membranes. Experimentally the translocation of precursor proteins across the inner membrane can be broken down into two steps: (1) the insertion of the presequence and (2) the translocation of the remainder of the polypeptide. The first step is dependent on Ay; the second is not. As described below Ssclp mainly functions in the second step, the movement of the bulk of the polypeptide into the matrix. A genetic approach has proved to be a very useful companion to biochemical analysis in understanding a number of physiological processes in both eucaryotes and procaryotes. The translocation of proteins from the cytosol into mitochondria has been no exception. Temperaturesensitive mutants have allowed the assessment of the effect of the absence of a single functional protein in vivo. When cells are grown under permissive conditions, the protein of interest is functional. Defects that are detected within a very short time after a shift to the nonpermissive temperature are likely a direct result of the inactivation of the protein. In this chapter we discuss the results of our combined genetic and biochemical analysis of the role of mitochondrial hsp70 in the translocation of proteins into mitochondria using temperature-sensitive hsp70 mutants.

IL A ROLE FOR CYTOSOLIC hsp70 IN PROTEIN TRANSLOCATION INTO ORGANELLES Several years ago it was suggested that cytosolic hspTOs of 5. cerevisiae were important for efficient translocation of at least some proteins from the cytosol to the lumen of the ER or matrix of the mitochondria. In vivo evidence came from the observation that strains depleted of the SSA proteins accumulated precursors of two proteins, the (3 subunit of the FJFQ ATPase, a mitochondrial protein, and a factor, a secreted protein, as the amount of SSA proteins dropped below the level found in wild-type

Molecular Chaperones and Protein Transport

5

(wt) cells (Deshaies et al., 1988). This initial experiment suffered from the fact that several hours elapsed after new expression of SSA proteins was stopped before a decrease in protein translocation was observed. Therefore it was possible that the effect of SSA depletion on translocation was an indirect one. However, recent experiments with temperature-sensitive SSA mutants have shown a defect in translocation of the same proteins within 10 minutes after a shift to the nonpermissive temperature, suggesting a direct effect (Becker and Craig, unpublished results). In vitro experiments supported the notion that SSA proteins are important for translocation. Translocation of radiolabeled proteins into microsomal vesicles and mitochondria was facilitated by the addition of SSA proteins (Chirico et al., 1988). Incubation of the precursor with urea substituted for the addition of SSA proteins in facilitating translocation, suggesting that the SSA proteins may aid in maintaining the precursor in a partially unfolded, translocation-competent conformation. This is an appealing idea because of the known peptide-binding activity of hsp70s and because of their propensity to bind to regions rich in hydrophobic residues as would be exposed in unfolded proteins. However, at this point there is no demonstration in vivo of a direct interaction of SSA proteins with precursors, although such interactions have been observed in reticulocyte lysates (Chirico, 1992).

III. ROLE OF MITOCHONDRIAL hsp70 IN PROTEIN TRANSLOCATION A. In vivo Analysis of a Mitochondrial hsp70 Mutant The first indication that mitochondrial hsp70 is required for the translocation of proteins from the cytosol into the matrix of mitochondria came from analysis of a temperature-sensitive mutant of S. cerevisiae (Kang et al., 1990). This mutant, called sscl-2, was obtained by directly mutagenizing SSCl DNA in vitro. Because of the suggestion that the cytosolic hsp70s, the SSA proteins, are involved in the translocation of at least some proteins from the cytosol into the endoplasmic reticulum and mitochondria, the sscl-l mutant was tested for its ability to carry out translocation at the nonpermissive temperature of 37°C. Specifically, the effect of a shift to 37°C on the conversion of the precursor form of mitochondrial proteins to the mature form was tested. The proteins in extracts made from cells 30 minutes after a shift from 23°C to 37°C were separated by electrophoresis and reacted with antibodies specific for

E. A. CRAIG, B. D. GAMBILL, W. VOOS, and N. PFANNER

*^

4- 4- ^ /

••••NliM

/ HSP60

f^f^

JL 23* C

37* C

Figure 1. Accumulation of precursor proteins in SSC1 mutants in vivo. Cultures of WT, ssc1'2 and ssc1-3 cells growing at 23°C were divided, and half of each culture was shifted to 37°C for 30 minutes prior to harvest. Protein extracts were fractionated by SDS-PAGE, electrotransferred to nitrocellulose filters, and probed with hsp60- and F^p-specific antiserum, p, precursor, m, mature.

mitochondrial proteins. Accumulation of the precursor form of hsp60 (Fig. 1) as well as a number of other proteins was observed. Since the protease that catalyzes the cleavage of matrix-destined mitochondrial proteins is located in the matrix, the lack of efficient cleavage in the sscl-2 mutant suggested a defect in the translocation process. Since initially cells were shifted to the nonpermissive temperature for 30 minutes prior to harvest of the cells, the effect on processing could easily have been due to a secondary rather than a direct effect of a mutation. To determine how rapidly the processing defect occurred after temperature shift, cells were labeled between 5 and 10 minutes after the shift and the

Molecular Chaperones and Protein Transport

7

P subunit of the FJFQ ATPase (FjP) was precipitated using specific antibodies. Whereas in wild-type (wt) cells all of the p subunit synthesized in this period was converted to the mature form, all of the detectable P subunit in the mutant was found in the precursor form, indicating a rapid decrease in precursor processing after the shift to the nonpermissive temperature. To ascertain the location of the precursor in the sscl-l cells, pulselabeled cells were fractionated prior to immunoprecipitation with FjPspecific antibody (Kang et al., 1990). After a 5-minute pulse, nearly all of the labeled subunit was associated with the mitochondria. However, it remained sensitive to added protease compared to matrix-localized proteins, indicating that it was associated with the mitochondrial surface. B. Mitochondrial hsp70 Is Required for Translocation into the Matrix Functional Ssclp Is Required for Translocation into Isolated Mitochondria

In vivo experiments are limited in allowing the determination of a specific defect in the translocation process; therefore we attempted to analyze translocation in sscl'2 mitochondria in a cell-free system (Kang et al., 1990). Mitochondria were isolated from wt and sscl-2 cells grown at 23°C. Prior to use in in vitro translocation assays the mitochondria were incubated at 37^C for 15 minutes to inactivate SSCl protein. Radiolabeled precursor proteins synthesized in the presence of [^^S] methionine in reticulocyte lysates were added to these isolated, energized mitochondria. Although efficient processing of the precursor proteins occurred, the proteins remained sensitive to exogenously added protease, indicating that the bulk of the proteins was not imported. In initial experiments similar results were obtained with precursors of Fe/S protein, the p subunit of the FJFQ ATPase, and cytochrome b2, suggesting that a decrease in SSCl function causes a general defect in translocation. The in vitro and in vivo results are in apparent contradiction. The in vivo experiments suggested a defect in translocation resulting in failure of insertion of the N-terminal targeting sequence across the inner membrane. However, the in vitro experiments suggested that the membrane potential-dependent insertion of the targeting sequence is not inhibited, whereas translocation of the remainder of the protein is defective. We think that the in vitro experiments provide a better indication of the role

8

E. A. CRAIG, B. D. GAMBILL, W. VOOS, and N. PFANNER

of mitochondrial hsp70 in translocation. An explanation of the apparent discrepancy may be the differences in the conditions. In an in vitro assay, radiochemical amounts of precursor are being imported; sites of import are in vast excess over the number of precursor proteins. However, in the in vivo experiment cells are labeled over a period of 5 minutes, 5 minutes after the shift to the nonpermissive temperature. The amount of precursor synthesized during the labeling time exceeds the number of import sites available. Therefore if translocation were blocked as indicated by the in vitro experiments with precursors "stuck" in an import site, the precursors synthesized after this point would be expected to accumulate in the unprocessed form. The observed association of precursors with mitochondria is probably due to interaction with receptors or other interactions with the outer surface of the outer membrane. Precursors labeled at later times after the temperature shift were found in the cytosol, suggesting that binding sites on the mitochondrial surface were saturated. Ssclp Interacts Directly with Precursor Proteins Since hsp70s were known to bind to unfolded proteins it was not unreasonable to propose that Ssclp's role in translocation might be due to its direct association with the precursor as it enters the mitochondria. An indication of a direct association came from co-immunoprecipitation experiments (Ostermann et al., 1990). Labeled proteins that were trapped during translocation into sscl'2 mitochondria at the nonpermissive temperature were precipitated by Ssclp-specific antibody. In agreement with these findings, cross-linking experiments identified Ssclp as a protein in extremely close proximity to translocating polypeptides in wild-type mitochondria (Scherer et al., 1990). To test the effect of the conformation of the precursor on its ability to be translocated into mutant mitochondria, the fusion protein Su9-DHFR was denatured by incubation in 8 M urea. The denatured preprotein was diluted directly into import reactions. Unlike partially folded precursor added directly from reticulocyte lysates, the import rates for the unfolded precursors into wt and sscl-2 mitochondria were very similar. These results suggested that Ssclp is involved in the unfolding of preproteins during translocation into mitochondria. In summary: initial studies indicated that mitochondrial hsp70 was required for translocation of a number of proteins into mitochondria. Initial insertion of the targeting sequence across the outer and inner membrane is not dependent upon hsp70 function. However, the translo-

Molecular Chaperones and Protein Transport

9

cation of the bulk of the protein across the membranes requires functional hsp70 and involves a direct interaction between the precursor protein and mitochondrial hsp70. Isolation and Analysis of a Second SSC1 Allele^ ssc1-3

In order to more fully understand the role of mitochondrial hsp70 in translocation, additional temperature-sensitive SSCl mutants were sought. One additional mutant, called sscl-3, was isolated by the same procedure used in the isolation of ssc 1-2 (Gambill et al., 1993) and was found to be defective in processing precursor proteins at the nonpermissive temperature of 37°C (Fig. 1). Surprisingly, sscl-2 and sscl-3 mitochondria had different properties with regard to the translocation of the fusion protein Su9-DHFR (dihydrofolate reductase) (Fig. 2). Su9-DHFR contains the presequence plus three amino acids of the mature portion of the Neurospora crassa F^-ATPase subunit 9 and the entire protein coding region of murine DHFR (Pfanner et al., 1987). Unlike most presequences the Su9 presequence contains two cleavage sites for the matrix-localized protease, after amino acids 35 and 66. Wild-type mitochondria efficiently cleaved Su9-DHFR to the mature form and translocated it to a proteaseresistant location. Consistent with the initial results described above, addition of Su9-DHFR to heat-treated sscl-l mitochondria resulted in cleavage to the mature form, but it did not result in translocation to a protease-resistant location. However, most of the Su9-DHFR added to sscl-3 mitochondria was only cleaved at the first cleavage site at amino acid 35 and not translocated to a protease-resistant location. This difference in precursor translocation into mitochondria suggested that the sscl-3 mitochondria might be more defective in translocation than sscl-2 mitochondria. To further address this difference, we compared the ability of the two types of mitochondria to import Su9-DHFR that was artificially denatured by incubation in 8 M urea (Gambill et al., 1993). Both sscl-2 and sscl-3 mitochondria efficiently translocated denatured precursor to a protease-protected location (Fig. 2). However, while the precursor was translocated across both the inner and outer membranes in sscl-2 mitochondria, it was translocated across only the outer membrane of the sscl-3 mitochondria. This was shown by subjecting the mitochondria to a hypotonic solution after import, causing rupture of the outer, but not the inner membrane (Fig. 3). After swelling, most of the protein imported into sscl-2 mitochondria was still resistant to exogenously added pro-

+ATP mm

^

r4

m

^

r4 en

^

iH

iH

^

iH

iH

^

»

»

^

CO

CO

OS

1

+PK

+ATP

(N ro CO

+PK Figure 2. Undenatured Su9-DHFR precursor accumulates in ssc^-2 and sscl-3 mitochondria in a protease-sensitive form, but after denaturation is translocated to a protease-resistant location. ^^S-labeled precursor of Su9DHFR synthesized in reticulocyte lysate was incubated with isolated mitochondria (25 ^g/lane) from WT, sscl-2, or ssc1-3 that had been preincubated for 15 minutes at 37°C to inactivate mutant Ssdp. Where indicated the mitochondria were treated with proteinase K (PK) after the import reaction. The reisolated mitochondria were analyzed by SDS-PAGE and fluorography. (A) The lysate was added directly to the mitochondria. (B) The reticulocyte lysate containing the labeled precursor was precipitated with ammonium sulfate and dissolved in 8 M urea to denature it (Kang et al., 1990). The precursor was then added directly to mitochondria; the final import mix had a concentration of 200 m M urea. Reprinted (Gambill et al., 1993) with permission. 10

Molecular Chaperones and Protein Transport

WT

Swelling PK

-

+ +

11

sscl-2

+

-

+ +

sscl-3

+

-

Figure 3. Urea-denatured Su9-DHFR is transported into the matrix of sscl-2 mitochondria but remains in the intermembrane space in sscl-3 mitochondria. Import was performed as described in Figure 2B. After import mitochondria were reisolated, a portion was diluted into a hypotonic solution (+ swelling), and a portion was diluted into an isotonic buffer containing 0.6 M sorbitol (-swelling). Where indicated (+ PK), proteinase K was added for 20 minutes after the dilution. The mitochondria were reisolated and analyzed by SDS-PAGE and fluorography. The treatment in a hypotonic solution results in the rupture of the outer but not the inner membrane, as confirmed by the localization of the intermembrane space protein cytochrome 62/ the inner membrane protein ADR/ATP carrier, and the matrix protein Ssd p. Taken from Gambill et al. (1993) with permission.

tease, indicating that the protein had been completely translocated into the matrix. But a large portion of the protein imported into sscl-3 mitochondria became susceptible. Therefore in sscl-3 mitochondria most of the Su9-DHFR protein accumulated as an intermediate that was cleaved once and spanned from the matrix, across the inner membrane into the intermembrane space. This type of intermediate has been termed the "intermembrane space intermediate" (Hwang et al., 1989; Rassow and Pfanner, 1991; Jascur et al., 1992; Pfanner et al., 1992). Even the small amount of mature form accumulated was exposed to the intermembrane space. It is possible that the failure of the sscl-3 mitochondria to translocate Su9-DHFR across the inner membrane in these experiments was due to the refolding of the precursor prior to crossing the inner membrane into a conformation that prevented its passage. To test this possibility translocation was carried out with mitochondria whose outer membranes had been disrupted (Gambill et al., 1993). Previously Schatz and co-workers

12

E. A. CRAIG, B. D. GAMBILL, W. VOOS, and N. PFANNER

(Hwang et al., 1989) had shown that such mitochondria, referred to as mitoplasts, were capable of translocating some preproteins. While denatured Su9-DHFR could be transported across the inner membrane of sscl-2 mitoplasts, almost no translocation into sscl-3 mitoplasts occurred. We concluded that unfolding of the polypeptide chain is not sufficient to allow translocation across the inner membrane of sscl-3 mitochondria. Since hsp70 function is required for translocation of proteins even when denatured, we concluded that mitochondrial hsp70 is a bonafide component of the mitochondrial translocation machinery. Comparison of the sscl-l and sscl-S Alleles In the in vitro translocation assays described above the sscl-S mutant displayed a much stronger phenotype than the sscl-2 mutant (Fig. 4). In the sscl-3 mutant only the first site of cleavage of Su9-DHFR was cleaved, whereas in the sscl-2 mutant both were cleaved. In addition, after denaturation Su9-DHFR only traversed the outer membrane, whereas complete translocation into the matrix was observed in the sscl-2 mutant. These in vitro results were somewhat surprising to us since our initial observations of the growth phenotypes of the two strains suggested that sscl-3 was a "weaker" allele than sscl-2. Upon the shift to 37°C sscl-3 is able to form small, barely visible colonies before growth ceases, whereas sscl-2 forms no visible colonies, sscl-3 cells can resume growth and form colonies at 23°C after being at 37°C for several days. While less than 1 day at 37°C prevents colony formation of sscl-2 at any temperature. Therefore the sscl-3 mutation is merely cytostatic at the nonpermissive temperature, whereas the sscl -2 mutation is cytocidal. This phenotypic difference is likely due to differences in the effect of the two mutations on the function of Ssclp. We tested the ability of the mutant hsp70s to interact with translocating polypeptides. Urea-denatured and labeled Su9-DHFR was incubated with wt, sscl-2, and sscl-3 mitochondria. The capacity of the mature protein to be co-immunoprecipitated by Ssclp antibody was tested. At short incubation times, Su9-DHFR was co-immunoprecipitated in both wt and sscl-2 mitochondria. The immunoprecipitation was about 70% efficient as compared to precipitation with DHFR antibody in wt mitochondria and about 80% as efficient in sscl-2 mitochondria. Interestingly, after longer times of incubation the DHFR could no longer be precipitated by Ssclp antibody in wt mitochondria, presumably because the DHFR protein had become properly folded. However, efficient precipitation was still observed in

.OM > IM

SSC1-2 (-urea)

SSC1-2 (+urea)

,0M ^ IM

SSC1-3 (-urea)

ssc1'3

(+urea)

Figure 4. Depiction of the location of the Su9-DHFR in ssc1-2 (A, B) and sscl-3 (C, D) mitochondria. As described in the text and shown in Figure 3. (A) When the precursor is not denatured by incubation in 8 M urea, both N-terminal cleavage sites of Su9-DHFR enter the matrix, but the bulk of the protein remains outside the outer membrane of ssc1-2 mitochondria. (B). After denaturation with urea the entire protein is imported into the matrix of ssc1-2 mitochondria. (C) In ssc1-3 mitochondria the N-terminus of Su9-DHFR is imported to a point that allows cleavage only at the first site. (D). Upon denaturation with urea, the protein enters the intermembrane space, but no further import into the matrix is detected. Asterisk (*) indicates the sites of cleavage of the matrix-localized processing protease. For diagrammatic purposes, cleavage is not shown.

13

14

E. A. CRAIG, B. D. GAMBILL, W. VOOS, and N. PFANNER

sscl-2 mitochondria, suggesting that the Sscl-2p remained bound to imported protein for a much longer time. Analysis of the interaction of Sscl-3p with precursor was more difficult to evaluate. Only a small amount of mature protein was formed in the sscl-3 mitochondria and only about 20% of that imported was able to be co-immunoprecipitated with Ssclp antibody. Thus it appears that Sscl-3p is strongly impaired in binding Su9-DHFR, whereas Sscl-2p fails to release from Su9DHFR. Although the exact effect of the two mutations on substrate protein binding and release will not be determined until purified protein is analyzed, the results from the analysis of the interaction of Su9-DHFR with the different SSCl proteins in isolated mitochondria suggests an explanation for the seemingly contradictory effects of the mutant proteins on translocation in the in vitro system and the effects on cell growth. The translocation assays suggest that sscl-3 mutant protein is less effective than sscl-2 in protein translocation into mitochondria. Our working model proposes that Sscl-2p still has substantial affinity for substrate proteins but is also defective in releasing the substrates that are bound. This residual binding activity would be sufficient for translocation under certain conditions, such as when the precursor protein is denatured. However, the defect in release indicated by prolonged association with imported proteins could be problematic in vivo. After a shift to the nonpermissive temperature, failure to release from proteins after binding might well interfere with mitochondrial processes more than a simple failure to bind, causing a lethal effect. Because it could not be efficiently immunoprecipitated with precursor proteins, Sscl-3p likely has a greatly decreased affinity for substrate proteins. C. Degree of Requirement of Sscl p Activity in Transport Is Correlated with the Conformation of the Precursor The results described above, concerning the failure of Su9-DHFR precursor to be imported across the inner membrane into sscl-3 mitochondria even when denatured by treatment with urea, indicate that Ssclp is necessary for translocation across the inner membrane. However, the ability of urea-denatured Su9-DHFR to cross the inner membrane of sscl-2 mitochondria implies an effect of the structure of the precursor on its ability to be translocated. In order to better understand the role of Ssc Ip on protein translocation we tested the ability of a variety of different precursors to be imported into isolated wt, sscl-2, and sscl-3

Molecular Chaperones and Protein Transport

^ PL, Jk a; r>. '22 vD T H CTJ

15

i-b 2(167)-DHFR

403020-

^m-b2(167)-DHFR

10-

• WT o sscl-2

^0)O ^-1 '^

as6

o sscl-3

»—t

01 f ()

^

5

1

1

15

20

1

10

Time (min) Figure 5. 62(167)-DHFR precursor is imported into sscl-2 and sscl-3 mitochondria as well as into wt mitochondria. Import experiments using )b2(167)-DHFR precursor were carried out as described in Figure 2. After the import reaction was carried out for various times, the mitochondria were treated with proteinase K to degrade any protein on the mitochondrial surface and were reisolated and analyzed by SDS-PAGE and fluorography. Taken with permission from Voos et al. (1993).

mitochondria (Voos et al., 1993). Surprisingly, we found a precursor that was imported into the three types of mitochondria with nearly identical efficiency (Fig. 5). The precursor, Z72(167)-DHFR, consisted of 167 N-terminal residues of the yeast cytochrome ^2 protein and the entirety of murine DHFR. The 167 residues of the cytochrome b^ protein included 80 residues of the presequence followed by 87 residues of the mature protein. Cytochrome ^2 is located in the intermembrane space; the 80 residues of the presequence contain the information for targeting to this compartment. In the process of import the precursor is cleaved once by the matrix processing protease to give an intermediate-sized form and then a second time on the outer side of the inner membrane by inner membrane protease I to yield the mature form. The processing of fe2(167)-DHFR into the intermediate and mature forms occurred at the same rate in wt, sscl-2, and sscl-3 mitochondria that had been incubated at 37°C for 15 minutes prior to import to inactivate the mutant hsp70s.

16

E. A. CRAIG, B. D. GAMBILL, W. VOOS, and N. PFANNER

I

rsi

ro

(N

ro

u

u

U

U

CO (/)

5 min

CO CO

(/) U5 1

U.

CO CO

15 min

Figure 6. Urea-denatured fa2(220)-DHFR is imported with similar efficiency into WT, sscl-2 and ssclS mitochondria. Reticulocyte lysate containing ^^S-labeled b2(220)-DHFR was precipitated with ammonium sulfate and dissolved in 8 M urea. The denatured precursor was diluted 40-fold into the import reaction containing mitochondria. Import into mitochondria was performed for 5 or 15 minutes, as described in Figure 2. Taken from Voos et al. (1993) with permission.

In all cases the mature form was protected against digestion with exogenous protease; however, it was susceptible to digestion after incubation with hypotonic buffer, indicating that the imported protein was appropriately localized to the intermembrane space. We had previously observed that the import of authentic cytochrome ^2 into sscl'2 mitochondria was partially inhibited (Kang et al., 1990). We repeated this result and determined that import of the full-length cytochrome 63 protein into sscl-3 mitochondria was almost completely blocked. To track down the reason for the difference in the translocation of authentic cytochrome 63 and the b2167-DHFR fusion, we tested other fusions that contained larger portions of cytochrome 62- It was possible that the DHFR moiety conferred the SSCl independence. However, when we tested two other fusions that contained larger portions (220 and 330 amino acids) of cytochrome ^2 linked to DHFR we found that their

Molecular Chaperones and Protein Transport

17

100en W

^^ffl

(IH

Q1

^H ^H

^

«-H

-^ o

— ^^^

CN g

50^ ^

H ^^^

I

«
o 9"

^^^ ^^§

,

s

M

n -

U

H

fN

ro

H

(N

^

'-^

rH

^

f^

(/3 C/3

t/i CO

u

u

u

CO CO

ro rH

u

c/3 CO

+ Hemin figure?. Import of 62(220)-DHFR into ssc/-2 and ssc^'3 mitochondria is inhibited in the presence but not the absence of hemin. Rabbit reticulocyte lysates were prepared in the presence of 3'5' cycl ic AMP rather than hemin. Imports were carried out as described in Figure 1. Where indicated (+ hemin), hemin was added to a final concentration of 10 |J.M. Taken from Voos et al. (1993) with permission.

import into sscl-l and sscl-3 mitochondria was inhibited (Voos et al, 1993). These results suggested that sequences between 167 and 220 conferred dependence on Ssclp. To test if the conformation of the preprotein was causing the hsp70dependence, the b2(220)-DHFR precursor was denatured in urea prior to addition to mitochondria. The denatured precursor was translocated into the three types of mitochondria at nearly identical rates, indicating that the conformational state of the longer fusion and the authentic cytochrome &2 prevented translocation (Fig. 6). Cytochrome h^ binds heme noncovalently; the heme-binding domain is located within the first 99 amino acids of the mature form of the protein. The b2( 167)-DHFR fusion, that is translocated into mutant mitochondria contains only 87 amino acids of the mature protein, whereas b2(220)-DHFR includes the entire heme-binding domain. Since reticulocyte lysates contain hemin we thought that the heme might bind to cytochrome h^ and stabilize the

18

E. A. CRAIG, B. D. GAMBILL, W. VOOS, and N. PFANNER

conformation of the heme-binding domain such that it would require unfolding in order to be translocated into the mitochondria. To test this idea, the b2(220)-DHFR fusion was synthesized in a reticulocyte lysate that contained 3^5' cyclic AMP rather than hemin (Ernst et al, 1976; Nicholson et al., 1987). Indeed, the fusion was imported into sscl-l and sscl-3 mitochondria as efficiently as into wt mitochondria (Fig. 7). Furthermore, addition of hemin to lysates that lacked hemin inhibited the translocation into the mutant mitochondria. Therefore we concluded that the dependence of cytochrome ^2 and the fusions with DHFR are dependent upon mitochondrial hsp70 in the in vitro translocation assays because of the conformation of the precursor protein formed in the reticulocyte lysates. D. Model of hsp70 Action in Translocation Across Mitochondrial Membranes Putting together the data collected from analysis of the sscl-2 and sscl-3 mitochondria, we have concluded that mitochondrial hsp70 is required for the translocation of precursor proteins into the mitochondrial matrix. However, the degree of dependence, at least in vitro, varies among different precursors. Some precursors require a low level of hsp70 action, as evidenced by their ability to be imported into sscl'2 mitochondria. This is particularly evident in the complete translocation of denatured precursors into sscl-2 but not into the more severely affected sscl-3 mitochondria. This lack of translocation of denatured precursors into sscl-3 mitochondria indicates that mitochondrial hsp70 is a critical component of the mitochondrial translocation machinery. However, the greater dependence of undenatured precursors on mt-hsp70 indicates that it also plays a role in the unfolding of the precursor on the cytosolic side of the membranes. How might mt-hsp70 act in the translocation and unfolding of a precursor protein? We propose a simple model (Fig. 8) that attributes action of mt-hsp70 in translocation and unfolding to its well-established peptide-binding activity. After translocation of the leader sequence across the outer and inner membranes, a process dependent upon a membrane potential and the targeting signal, mt-hsp70 binds to the precursor protein. In the simplest form of the model the movement of the protein across the membranes may simply be the result of Brownian motion. However, the binding of mt-hsp70 would prevent the movement of the translocating polypeptide back toward the cytosol by steric hin-

Molecular Chaperones and Protein Transport

19

X [1]

[2]

[3]

[4]

/

Figure 8. Model for mt-hsp70 action during translocation of precursor proteins into the matrix of mitochondria. (1) The precursor protein binds to import receptors on the outer surface of the mitochondria. c-hsp70 may be bound to the partially folded precursor. (2) The membrane potentialdependent step of import, the translocation of the N-terminus of the protein, occurs. m-hsp70 binds to the precursor and prevents any movement back toward the cytosol. (3) By Brownian motion (and possibly because of the action of some proteins) the precursor protein moves further into the matrix. (4) Another m-hsp70 binds to a site more toward the C-terminus of the protein and prevents movement back toward the cytosol. (5) More of the protein moves into the matrix as described in (3). (6) Another m-hsp70 binds as described in (4). (7) The entire protein moves into the matrix and translocation is complete. See text for discussion of the action of c-hsp70 versus m-hsp70.

drance. As more of the protein is exposed in the matrix, additional high-affinity binding sites for mt-hsp70 are exposed, and binding to these sites prevents movement back toward the cytosol. In a more complex model components of the translocation machinery in the membrane, such as MIM44, MIM23, and MIM17, could also, in part, provide the driving force of translocation (Dekker et al., 1993; Blom et al., 1993; Emtage and Jensen, 1993). Although this notion of binding of mt-hsp70 providing the driving force of translocation is fairly straightforward, how binding of an hsp70 inside the mitochondria could affect the "unfolding" of a protein on the

20

E. A. CRAIG, B. D. GAMBILL, W. VOOS, and N. PFANNER

cytosolic side is more difficult to understand. Translocation may be a dynamic process, a competition between unfolding on the cytosolic side and translocation into the matrix. The secondary and tertiary structures of the precursor could sterically inhibit the translocation process. A partially folded protein in the act of translocation, perhaps bound to cytosolic chaperones, likely undergoes slight changes in conformation that expose unfolded regions. The higher the degree of structure on the cytosolic side, perhaps the greater the need for binding of mt-hsp70. If a protein is denatured, a partially active mt-hsp70 with a lower affinity may be sufficient. Such a model predicts a competition between cytosolic and mitochondrial hsp70s since translocation into the matrix occurs, binding to the mitochondrial form is favored over binding to the cytosolic form. This could be accomplished in several ways. For example, there may be an inherent difference in binding affinities between the hspTOs or the availability of binding sites may be different on opposite sides of the membranes. If the preprotein is in a partially folded form on the cytosolic side but in an unfolded form as it enters the matrix, there would be inherent differences in sites exposed for chaperone binding. Since hsp70s are thought to have a higher affinity for sequences rich in hydrophobic residues and since these are more exposed in unfolded proteins and buried in folded ones, mt-hsp70 may very well have available higher-affinity binding sites than cytosolic hsp70. According to this model, the role of mt-hsp70 in translocation and unfolding would mechanistically be the same, but an hsp70 with a lower binding activity would be sufficient to drive the translocation of a denatured protein, but not of a protein with substantial tertiary structure. E. Requirements for Mitochondrial hsp70 in the Translocation of Proteins into the Matrix Compared to the Intermembrane Space In our analysis of the dependence of precursor proteins on mt-hsp70 we tested the import of shorter cytochromefc2-DHFRfusions (Voos et al., 1993). While fusions containing 151 and 84 amino acids of the cytochrome ^2 precursor were imported as efficiently into mutant mitochondria as wt mitochondria, fusions containing 55 or 47 amino acids were imported efficiently into wt and sscl-2 mitochondria but not sscl-3 mitochondria (Fig. 9). The sorting signal for the intermembrane space, a hydrophobic region preceded by a positively charged region, is located in the second half of the presequence, prior to amino acid 80 (van Loon

Molecular Chaperones and Protein Transport

21

100 H Q I

^

2

rH -^

O U

JD

O

•d

*

vo

d

50 H

o

I

P4 (N

fN

m

\i

U

u

5 mln

u CO CO

CO CO

CO CO

15 min

Figure 9. b2067)^^^-DhfR is Inhibited in import into ssc1-3, but not SSC/-2 mitochondria. The precursor of 62(167)^i9-DHFR was imported into isolated mitochondria as described in Figure 2.

et al., 1986; Haiti et al., 1987; KoU et al., 1992; Click et al., 1992). It seemed likely that the reason for the lack of import of the short fusions into sscl'3 mitochondria was due to the lack of a functional intermembrane space sorting signal. This idea was supported by the lack of import of afc2(167)-DHFRfusion having a deletion between amino acids 47 and 65, which is known to prevent sorting of the inner membrane space (Koll et al., 1992) and which also was not imported into sscl-3 mitochondria. In fact, in sscl-3 mitochondria the first cleavage of this fusion did not occur even if the precursor was first treated with urea, whereas cleavage and translocation to a protease-resistant location occurred in wt and sscl'2 mitochondria. &2(55)-DHFR, &2(47)-DHFR, and b^iiei)^^^ DHFR are known to be translocated into the matrix of the mitochondria because they lack a functional intermembrane space sorting signal. Therefore from the work discussed above, it is not surprising that their translocation into the matrix is dependent on Ssclp. However, the dependence of authentic cytochrome 62 and the longer DHFR fusions on Ssclp is more provocative since the sorting pathway of cytochrome Z?2 is currently the subject

22

E. A. CRAIG, B. D. GAMBILL, W. VOOS, and N. PFANNER

of controversy. According to the "stop-transfer" hypothesis, targeting to the inner membrane space is caused by the arrest of the translocating polypeptide in the inner membrane. In other words, the sorting signal prevents the protein from entering the matrix. If this model is correct in its simplest form it would not be at all surprising if cytochrome 63 import was independent of hsp70 and if deletion of the sorting signal rendered import hsp70-dependent, since the protein would be imported into the matrix. However, the longer fusions and authentic cytochrome ^2 ^ ^ dependent on hsp70, as described above because of the conformation of the heme-binding domain. According to the "stop-transfer" model (Glick et al., 1992a,b), Ssclp could only interact with the extreme N-terminus of the preprotein, but by this interaction it would be able to facilitate the unfolding of the protein on the other side of the membrane. Although this possibility cannot be excluded, it is easier for us to imagine this being accomplished by interaction of the protein with hsp70 along its entire length. Along these lines, the original "conservative sorting" model proposed (Hartl et al., 1987) that cytochrome ^2 was imported entirely into the matrix and then reexported to the inner membrane space. A revision of the model proposed recently (Koll et al., 1992) suggests that the import and export steps are coupled and that a sorting component may recognize the sorting signal and redirect translocation back across the inner membrane. Perhaps if a precursor protein has little tertiary structure then the import does not require hsp70; that is, interaction with the intermembrane space sorting machinery may be sufficient for import. However, if the conformation of the preprotein is more of a detriment to import, the additional action of hsp70 may be required, as it is for matrix-bound proteins. F. Components Other Than Mitochondrial hsp70 Required for Protein Movement across the Outer and Inner Membranes From the results described above it appears that mt-hsp70 is required for translocation of preproteins across the inner membrane regardless of the conformation of the precursor protein. However, in the mutant with very low mt-hsp70 activity proteins could be translocated across the outer membrane into the intermembrane space if they were denatured prior to addition to mitochondria. This dependence on denaturation to cross the outer membrane suggests that there is a component of the translocation machinery in addition to the hsp70 in the matrix that is

Molecular Chaperones and Protein Transport

23

required for translocation across the outer membrane. This component may well reside in the intermembrane space and act in conjunction with mt-hsp70. This same component or another may also be involved in the translocation of proteins destined for the intermembrane space as well as those translocated into the matrix. This may explain the independence of certain precursors, such asfo2(167)-DHFR,on Ssclp function. In cases where there is little conformational restriction the second component may be sufficient for translocation, whereas other precursors with more tertiary structure may require both components.

IV. ROLE OF ATP IN PROTEIN TRANSLOCATION It has been appreciated for some time that ATP is required for the translocation of proteins into the mitochondrial matrix. Although some proteins require ATP on the outside of the outer membrane, all precursors tested require ATP inside mitochondria (Hwang and Schatz, 1989; Neupert et al., 1990; Manning-Krieg et al., 1991). Since hsp70s are ATP binding proteins and ATP hydrolysis is required for cycles of peptide binding and release, it is reasonable to propose that ATP is required because mt-hsp70 is required for translocation. Although there may be other requirements for ATP in the translocation and folding processes in the matrix, the close similarity between the effects of ATP depletion and lack of hsp70 function on translocation strongly supports the idea that ATP is required for hsp70 action in vivo, as described below. ATP levels were lowered by treating isolated mitochondria with apyrase and inhibition of the mitochondrial ATP-synthase with oligomycin (Voos et al., 1993). Upon ATP depletion both wt and sscl-l mitochondria accumulated the intermediate form of Su9-DHFR in a protease-accessible form, just as did sscl-3 mitochondria in the presence of ATP. In addition, after ATP depletion the Su9-DHFR precursor could not be co-immunoprecipitated with Ssclp antibody in wt or sscl'2 mitochondria, just as in sscl-S mitochondria in the presence of ATP. Although there is agreement in the literature that cycles of peptide binding and release require ATP hydrolysis, the mechanistic details of this cycle are not yet resolved. In some studies ADP-bound hsp70 has a higher affinity for peptide binding (Palleros et al., 1991), whereas in others ATP is required for the stabilization of the hsp70-protein complex when other accessory proteins such as DnaJ or GrpE are present (Langer et al., 1992; Georgopoulos, 1992). The failure of Ssclp to bind the precursor protein in the absence of ATP (Manning-Krieg et al., 1991;

24

E. A. CRAIG, B. D. GAMBILL, W. VOOS, and N. PFANNER

Gambill et al., 1993) indicates that in at least this in v/vo-like situation, the ATP-bound form is required for efficient binding.

V. ROLE OF hsp70 OF THE ENDOPLASMIC RETICULUM IN PROTEIN TRANSLOCATION As with SSCl, temperature-sensitive alleles have been exploited to analyze the function of the hsp70 of the ER, Kar2p. When shifted to the nonpermissive temperature, kar2-159 strains accumulate the precursor forms of several proteins normally translocated across the ER membrane, indicating a direct involvement in protein translocation (Vogel et al., 1990). An in vitro system was used to assess more directly the role of Kar2p in the process (Sanders et al., 1992). A translocation intermediate that becomes jammed during transit was used to freeze the translocation apparatus, allowing a study of its components. An association between the jammed precursor and Sec61p was detected, indicating that the precursor interacts directly with Sec61p. Sec61p, an integral membrane protein, is known to be critical for protein translocation and proposed to be a critical component of the translocation apparatus. The effect of three different KAR2 alleles on the translocation of the precursor and its interaction with Sec61 was determined as a function of the temperature in the in vitro system. Microsomes were prepared from the three mutant strains grown at 23°C, and translocation assays were carried out over a range of temperatures. Although the alleles showed different levels of activity all showed decreased activity at higher temperatures, and microsomes from wt cells were unaffected. Two of the alleles, kar2-113 and kar2-159, resulted in a severe reduction in complex formation. However, the third allele, kar2-203, did not significantly reduce the formation of the Sec61p-precursor complex. The authors suggested that this difference indicates that Kar2p acts at two points in the translocation process, in the formation of the Sec61 p-precursor complex and in some undefined event important for translocation that occurs after the Sec61 p-precursor interaction. As is the case with Ssclp, immunoprecipitation experiments indicate a direct interaction of Kar2p and the precursor protein. This interaction is consistent with results from mammalian cells showing a transient interaction between proteins entering the ER and the mammalian equivalent of Kar2p, BiP (reviewed in Gething and Sambrook, 1992). From these experiments the roles of Ssclp in the mitochondria and Kar2p in the ER appear to be similar. Both bind to proteins as they emerge

Molecular Chaperones and Protein Transport

25

into the organelle. Both are needed for translocation for a variety of proteins and are thus involved in general protein translocation. Although the extent of the parallels between the roles of the two hsp70s remains to be seen, this initial work suggests that hsp70 function may be important for the biogenesis of proteins in other organelles and perhaps the cytosol as well. The SSB cytosolic hsp70s bind to translating ribosomes and are important for efficient translation. Evidence suggests that SSB proteins may interact with the nascent chain on the ribosome. Exit from the ribosome into the cytosol may be similar to entering organelles after traversing the mitochondrial or ER membranes, a process facilitated by hspTOs. VI. SUMMARY In summary, mitochondrial hsp70 is a critical component of the apparatus required for translocation across the inner membrane of mitochondria. hsp70 function is required even for the translocation of denatured proteins. However, in the case of precursors with secondary and tertiary structure there is a greater requirement for hsp70 function, presumably because of the requirement for at least partial unfolding prior to translocation across the membranes. These studies concentrated on the role of hsp70 in the translocation process. However, clearly hsp70 does not act alone. In bacteria, hsp70 (DnaK) acts with two other proteins, DnaJ and GrpE. It will be of interest to analyze the function of homologous proteins in mitochondrial import. In addition, components of the inner membrane and the intermembrane space likely play critical roles. In the next few years an understanding of the interplay among these components of the translocation apparatus should be attained. REFERENCES Blom, J., Kubrich, M., Rassow, J., Voos, W., Dekker, P.J.T., Maarse, A.C., Meijer, M., & Pfanner, K. (1993). The essential yeast protein MIM44 (encoded by MP/7) is involved in an early step of preprotein translocation across the mitochondrial inner membrane. Mol. Cell. Biol. Boorstein, W., Ziegelhoffer, T, & Craig, E.A. (1994). Molecular evolution of the hsp70 multigene family. J. Mol. Evol. 37, in press. Chirico, W.J. (1992). Dissociation of complexes between 70 kDa stress proteins and presecretory proteins is faciliated by a cytosolic factor. Biochem. Biophys. Res. Commun. 189,1150-1156. Chirico, W., Waters, M.G., & Blobel, G. (1988). 70Kheat shock related proteins stimulate protein translocation into microsomes. Nature 332, 805-810.

26

E. A. CRAIG, B. D. GAMBILL, W. VOOS, and N. PFANNER

Craig, E.A., Kramer, J., & Kosic-Smithers, J. (1987). SSCl, a member of the 70-kDa heat shock protein multigene family of Saccharomyces cerevisiae, is essential for growth. Proc. Natl. Acad. Sci. USA 84,4156-4160. Craig, E.A., Kramer, J., Shilling, J., Wemer-Washburne, M., Holmes, S., Kosic-Smither, J.. & Nicolet, CM. (1989). SSCl, an essential member of the 5. cerevisiae HSP70 multigene family, encodes a mitochondrial protein. Mol. Cell. Biol. 9,3000-3008. Craig, E.A., Gambill, B.D., & Nelson, R.J. (1993). Heat shock proteins: molecular chaperones of protein biogenesis. Microbiol. Rev. 57,402-414. Dekker, RJ.T., Keil, R, Rassow, J., Maarse, A.C., Pfanner, N., & Meijer, M. (1993). Identification of MIM23, a putative component of the protein import machinery. FEES Lett. 330, 66-70. Deshaies, R., Koch, B., Wemer-Washburne, M., Craig, E. & Schekman, R. (1988). A subfamily of stress proteins facilitates translocation of secretory and mitochondrial precursor polypeptides. Nature 332, 800-805. Emtage, J.L.T., & Jensen, R.E. (1993). MAS6 encodes an essential inner membrane component of the yeast mitochondrial protein import pathway. J. Cell. Biol. 122, 1003-1012. Ernst, v.. Levin, D.H., Singh Ranu, R., & London, LM. (1976). Control of protein synthesis in reticulocyte lysates: effects of 3'5' cyclic AMP, ATP, and GTP on inhibitions induced by heme-deficiency, double-stranded RNA, and a reticulocyte translation inhibitor. Proc. Natl. Acad. Sci. USA 73,1112-1116. Flaherty, K.M., DeLuca-Flaherty, C , & McKay, D.B. (1990). Three dimensional structure of the ATPase fragment of a 70K heat-shock cognate protein. Nature 346, 623-628. Flajnik, M., Canel, C , Kramer, J., & Kasahara, M. (1991). Hypothesis: which came first, MHC class I or class II. Immunogenetics 33, 295-300. Flynn, G., Pohl, J., Flocco, M., & Rothman, J. (1991). Peptide-binding specificity of the molecular chaperone BiR Nature 353, 726-730. Fremont, D., Matsumura, M., Strua, E., Petersen, P., & Wilson, I. (1992). Crystal structure of two viral peptides in complex with murine MHC class I H-2Kb. Science 257, 919-926. Gambill, B.D., Voos, W, Kang, RJ., Miao, B., Langer, T, Craig, E.A., & Pfanner, N. (1993). A dual role for mitochondrial heat shock protein 70 in membrane translocation of preproteins. J. Cell Biol. 123, 109-117. Georgopoulos, C. (1992). The emergence of the chaperone machines. Trends Biochem. Sci. 17, 294-299. Gething, M.-J., & Sambrook, J. (1992). Protein folding in the cell. Nature 355, 33-45. Glick, B. & Schatz, G. (1991). Import of protein into mitochondria. Annu. Rev. Genet. 25, 21-44. Glick, B., Beasley, E.M., & Schatz, G. (1992a). Protein sorting in mitochondria. Trends Biochem. Sci. 17,453-459. Glick, B.S., Brandt, A., Cunningham, K., Muller, S., Hallberg, R., & Schatz, G. (1992b). Cytochrome Cj and b2 are sorted to the intermembrane space of yeast mitochondria by a stop-transfer mechanism. Cell 69, 809-822. Hartl, F.-U., Ostermann, J., Guiard, B., & Neupert, W (1987). Successive translocation into and out of the mitochondrial matrix: targeting of proteins to the intermembrane space by a bipartite signal peptide. Cell 51, 1027-1037.

Molecular Chaperones and Protein Transport

27

Hwang, S. & Schatz, G. (1989). Translocation of proteins across the mitochondrial inner membrane, but not the outer membrane, requires nucleotide triphosphate in the matrix. Proc. Natl. Acad. Sci. USA 86, 8432-8436. Hwang, S., Jascur, T, Vestweber, D., Pon, L., & Schatz, G. (1989). Disrupted yeast mitochondria can import precursor proteins directly through their inner membrane. J. Cell Biol. 109, 487-493. Jascur, T., Goldenberg, D.P., Vestweber, D., & Schatz, G. (1992). Sequential translocation of an artificial precursor protein across the two mitochondrial membranes. J. Biol. Chem. 267, 13636-13641. Kang, PJ., Ostermann, J., ShHling, J., Neupert, W., Craig, E.A., & Pfanner, N. (1990). Hsp70 in the mitochondrial matrix is required for translocation and folding of precursor proteins. Nature 348, 137-143. Koll, H., Guiard, B., Rassow, J., Ostermann, J., Horwich, A., Neupert, W., & Hartl, F.-U. (1992). Antifolding activity of hsp60 couples protein import into the mitochondrial matrix with export to the intermembrane space. Cell 68,1163-1175. Langer, T, Lu, C , Echols, H., Flanagan, J., Hayer, M.K., & Hartl, FU. (1992). Successive action of DnaK, DnaJ, and GroEL along the pathway of chaperone-mediated protein folding. Nature 356, 683-689. Manning-Krieg, U.C., Scherer, P., & Schatz, G. (1991). Sequential action of mitochondrial chaperones in protein import into the matrix. EMBO J. 10, 3273-3280. Neupert, W., Hartl, U., Craig, E., & Pfanner, N. (1990). How do polypeptides cross the mitochondrial membranes. Cell 83,447-450. Nicholson, D., Kohler, H., & Neupert, W. (1987). Import of cytochrome c into mitochondria: cytochrome c heme lyase. Eur. J. Biochem. 164, 147-157. Normington, K., Kohno, K., Kozutsumi, Y, Gething, M.J., & Sambrook, J. (1989). S. cerevisiae encodes an essential protein homologous in sequence and function to mammalian BIP Cell 57,1223-1236. Ostermann, J., Voos, W., Kang, PJ., Craig, E.A., Neupert, W., & Pfanner, N. (1990). Precursor proteins in transit through mitochondrial contact sites interact with hsp70 in the matrix. FEBS Lett. 277, 281-284. Palleros, D.R., Welch, W.J., & Fink, A.L. (1991). Interaction of hsp70 with unfolded proteins: effects of temperature and nucleotides on the kinetics of binding. Proc. Natl. Acad. Sci. USA 88,5719-5723. Pfanner, N., Tropschug, M., & Neupert, W. (1987). Mitochondrial protein import: nucleotide triphosphates are involved in conferring import-competence to precursors. Cell 49, 815-823. Pfanner, N., Sollner, T, & Neupert, W. (1991). Mitochondrial import receptors for precursor proteins. Trends Biochem. Sci. 16, 63-67. Pfanner, N., Rassow, J., van der Klei, I.J., & Neupert, W. (1992). A dynamic model of the mitochondrial protein import machinery. Cell 68,999-1002. Rassow, J. & Pfanner, N. (1991). Mitochondrial preproteins en route from the outer membrane to the inner membrane are exposed to the intermembrane space. FEBS Lett. 293, 85-88. Rippmann, F, Taylor, W., Rothbard, J., & Green, N.M. (1991). A hypothetical model for the peptide binding domain of hsp70 based on the peptide binding domain of HLA. EMBOJ. 10, 1053-1059.

28

E. A. CRAIG, B. D. GAMBILL, W. VOOS, and N. PFANNER

Rose, M.D., Misra, L.M., & Vogel, J.P. (1989). KAR2, a karyogamy gene, is the yeast homolog of the mammaUan BiP/GRP78 gene. Cell 57, 1211-1221. Sanders, S., Whitfield, K., Vogel, J., Rose, M., & Schekman, R. (1992). Sec61p and BiP directly facilitate polypeptide translocation into the ER. Cell 69, 353-366. Scherer, P., Krieg, U., Hwang, S., Vestweber, D., & Schatz,G. (1990). A precursor protein partially translocated into yeast mitochondria is bound to a 70kd mitochondrial stress protein. EMBO J. 9, 4315-4322. Stone, D.E. & Craig, E.A. (1990). Self regulation of 70 kilodalton heat shock proteins in Saccharomyces cerevisiae. Mol. Cell. Biol. 10, 1622-1632. van Loon, A.p., BrandU, A.W., & Schatz, G. (1986). The presequences to two imported mitochondrial proteins contain information for intracellular and intramitochondrial sorting. Cell 44, 801-812. Vogel, J.R, Misra, L.M., & Rose, M.D. (1990). Loss of BiP/grp78 function blocks translocation of secretory proteins in yeast. J. Cell Biol. 110, 1885-1895. Voos, W., Gambill, B.D., Guiard, B., Pfanner, N., & Craig, E.A. (1993). Presequence and mature parts of preproteins strongly influence the dependence of mitochondrial protein import on heat shock protein 70 in the matrix. J. Cell Biol. 123,119-126.

THE NUCLEAR PORE COMPLEX IN YEAST

Paola Grandi and Eduard C. Hurt

I. Introduction A. Structural Analysis of Nuclear Pore Complexes in Higher Eukaryotes B. Components of Vertebrate Nuclear Pores C. In Vitro Systems to Study Nuclear Transport II. Structureof the Yeast Nuclear Pore Complex III. Componentsof the Yeast Nuclear Pore Complex IV. Experimental Approaches to Studying Nuclear Pore Structure and Function in Yeast A. Biochemical Approach B. Synthetic Lethality as a Genetic Approach to Studying the Nuclear Pore Complex C. /n V/rra Nuclear Transport Assays V. Conclusions VI. Summary References

Membrane Protein TVansport Volume 2, pages 29-56 Copyright © 1995 by JAI Press Inc. All rights of reproduction in any form reserved. ISBN: 1-55938-983-4

29

30 31 32 33 35 36 39 39 43 47 49 51 51

30

PAOLA GRAND! and EDUARD C. HURT

I. INTRODUCTION Although nuclear pores were already identified at the end of the 1940s, when ultrastructural preparation techniques and electron microscope analysis of nuclei became available (Callan et al., 1949), the molecular architecture of nuclear pore complexes as well as their role in nucleocytoplasmic transport are still only poorly understood. Nuclear pores are supramolecular complexes of more than 120,000 kDa (Reichelt et al., 1990; Hinshaw et al, 1992) that in eukaryotic cells span the nuclear envelope at sites where the two nuclear membranes merge (Unwin and Milligan, 1982). Interestingly, nuclear pores of evolutionary distant organisms such as yeast and mammals have very similar structural appearance in the electron microscope (Maul, 1977; Allen and Douglas, 1989; Rout and Blobel 1991). The universal function of nuclear pores is to allow the exchange of molecules between the nucleus and the cytoplasm: ions, metaboUtes, and small macromolecules (exclusion size between 40 and 60 kDa) (Bonner, 1975) can generally diffuse through the pore channel, whereas for larger macromolecules exceeding the exclusion limit of the functional pore diameter (karyophilic proteins, RNPs, preribosomal particles) and also for the 21 kDa histone HI, an active transport mechanism is required (for reviews, see in this series and Dingwall and Laskey, 1986;Silver, 1991;Bataill6etal., 1990; Dingwall, 1991; Izaurralde and Mattaj, 1992; Osborne and Silver, 1993; Breeuwer and Goldfarb, 1990). In the case of active transport, specific targeting signals are required on the transported molecule: nuclear localization signal (NLS) on karyophilic proteins (Dingwall and Laskey, 1991; Silver, 1991; Osborne and Silver, 1993), monomethyl cap on the RNA moiety of exported RNPs (Hamm and Mattaj, 1990), and trimethyl cap and Sm proteins for snRNP import (Hamm et al., 1990; Fischer and Liihrmann, 1990; Fischer et al., 1991, 1993). It is assumed that these signals serve for the interaction with specific carrier proteins (NLS binding proteins, CAP binding proteins, RNA binding proteins) involved in the transport reaction. In addition to these transport signals, ATP and physiological temperatures (Newmeyer and Forbes, 1988; Akey and Goldfarb, 1989; Garcia-Bustos et al., 1991b) are indispensable. Moreover, nuclear transport exhibits saturation kinetics (Goldfarb et al., 1986; Silver et al., 1989). Finally, it has been shown in the Xenopus system that a given nuclear pore can be crossed in either direction by different substrates (Feldherr and Akin, 1989).

The Nuclear Pore Complex in Yeast

31

A. Structural Analysis of Nuclear Pore Complexes in Higher Eukaryotes

Electron microscopical techniques have been successfully employed to dissect the structural complexity of the nuclear pore complex (NPC) (Unwin and MiUigan, 1982; Akey, 1989; Reichelt et al., 1990; Hinshaw et al., 1992; Akey and Radermacher, 1993). This revealed that NPCs are organized in distinct substructures that are arranged with an octagonal symmetry. Two coaxial rings on each side of the nuclear membrane embrace the compact "plug-spoke" complex. The spokes protrude toward the central pore channel where the plug or transporter is located. It has been postulated that the plug might be involved in controlling the active, NLS-driven movement of substrates between nucleus and cytoplasm by means of a "double-iris" mechanism that would open or close the pore channel in response to precise stimuli (Akey, 1990). Diffusion of ions and small molecules is thought to occur through the eight peripheral channels that are formed between the spokes (Hinshaw et al., 1992) or, alternatively, between the transporter and the inner part of the spokes (Akey and Radermacher, 1993). Despite the bilateral symmetry of rings, spokes, and plug, a basketlike structure protruding from the nucleoplasmic ring of the NPC toward the nucleoplasm was first visuaUzed at the EM in isolated Xenopus oocyte nuclear envelopes (Ris, 1989; Jamik and Aebi, 1991; Goldberg and Allen, 1992). This structure is formed by eight filaments attached to the nuclear ring of the pore and terminating in a smaller ring inside the nucleoplasm. These distal rings are inserted into a fibrous network interposed between the nuclear lamina and the nucleoplasm: the nuclear envelope lattice (NEL) as seen by scanning electron microscopy (Goldberg and Allen, 1992). On the cytoplasmic ring of the NPC, granules of the size of the ribosomes were seen in negatively stained membrane-bound NPCs (Unwin and MiUigan, 1982; Stewart, 1992). With less drastic sample preparation techniques, short filaments have been detected instead (Jamik and Aebi, 1991). It is possible that collapsing of these filaments onto the cytoplasmic ring might give rise to the granules seen in the earlier experiments (Stewart, 1992). The functional role of the baskets, the NEL, and the cytoplasmic filaments is not known, although it has been postulated that they might be involved in, respectively, export of RNPs and import of proteins (Jamik and Aebi, 1991; Akey, 1992).

32

PAOLA GRANDI and EDUARD C. HURT

B. Components of Vertebrate Nuclear Pores The initial characterization of nuclear pore proteins was achieved by raising antibodies against rat liver nuclear envelope fractions enriched for NPCs (Gerace et al, 1982; Davis and Blobel, 1986). A nuclear membrane-spanning protein, called gp210 (Gerace et al., 1982), and a family of eight to ten related and conserved glycoproteins termed nucleoporins (Davis and Blobel, 1986; Snow et al., 1987) were thus identified. Gp210 and the nucleoporins were localized at the nuclear pore by indirect immunofluorescence and immunoelectron microscopy (Davis and Blobel, 1986; Gerace et al., 1982; Greber et al., 1990). With this latter technique the nucleoporins were located in the pore channel and at the cytoplasmic/nucleoplasmic periphery of the NPC, whereas anti-gp210 gold-coupled antibodies decorated the pore-membrane boundary. Nucleoporins in higher eukaryotes bear N-acetyl glucosamine residues and therefore bind wheat germ agglutinin (WGA) (Holt et al., 1987). Interestingly, both WGA and monospecific anti-nucleoporin antibodies block active nuclear transport when injected into Xenopus oocytes (Finlay et al., 1987; Park et al., 1987). In addition, polyclonal antibodies against p62, one of the more abundant and widely distributed nucleoporins (Davis and Blobel, 1986), were used to deplete nuclear reconstitution extracts (Dabauvalle et al, 1990; Finlay and Forbes, 1990; Finlay et al., 1991). When these fractions were used in the Xenopus nuclei reconstitution system, active nuclear transport was blocked. Moreover, a mRNA encoding a monoclonal antibody directed against the luminal domain of gp210 was expressed in cultured rat cells: the antibody was internalized into the perinuclear space and could bind to gp210 inside the perinuclear lumen, thereby blocking nuclear transport (Greber and Gerace, 1992). These findings demonstrate that nucleoporins and gp210 are essential components of the nuclear import machinery. Only a few nucleoporins from higher eukaryotes have been cloned so far: p62 has been identified and sequenced from rat, human, mouse, and Xenopus {^6%) (Starr etal., 1990; Carmo-Fonsecaetal., 1991; Cordeset al., 1991). Three distinct domains can be identified in this protein: i) In the amino-terminal domain, many peptide sequences of the GFSFG type are tandemly repeated. Similar repeats have also been found in the yeast nucleoporins NSPl (Nehrbass et al., 1990) and NUPl (Davis and Fink, 1990). ii) Adjacent to these repeats, there is a Ser/Thr-rich domain that is glycosylated by N-acetyl glucosamine residues, iii) A carboxy-termi-

The Nuclear Pore Complex in Yeast

33

nal domain organized in hydrophobic heptad repeats is potentially able to form coiled-coil interactions (Steinert and Roop, 1988; Lupas et al., 1991). Immunoprecipitation of p62 from the pool of WGA-binding proteins in rat liver nuclei revealed that this protein is tightly associated with two other proteins of 58 and 54 kDa (Finlay et al., 1991). Similarly, p68 (the p62 homologue in Xenopus) was found to be part of a complex of approximately 250 kDa when egg extracts were run on a sucrose gradient and probed with a specific anti-p68 antibody (Dabauvalle et al., 1990). Recently, a new nucleoporin has been identified in rat liver nuclei; its cDNAhas been cloned by using one of the anti-nucleoporin monoclonal antibodies generated previously in the same laboratory (Davis and Blobel, 1986; Sukegawa and Blobel, 1993). The primary sequence of the protein, termed NUP153, shows peptide repeats as found in p62, NSPl, and NUPl, but additionally four zinc finger motifs. NUP153 binds to rat genomic DNA only in the presence of zinc ions, and by immunoelectron microscopy it has been localized exclusively on the nucleoplasmic side of the NPCs. Therefore, it was speculated that NUP153 has a role in the three-dimensional organization of the genome, probably in positioning transcribable genes in close proximity to the pores (Blobel, 1985; Sukegawa and Blobel, 1993). C. In Vitro Systems to Study Nuclear Transport In order to identify soluble factors necessary for nuclear import and to analyze directly the step(s) of this process in which nucleoporins are involved, two different in vitro systems have been employed in higher eukaryotes. In the first system, functional nuclei are formed upon addition of Xenopus egg extracts to DNA or isolated rat liver nuclei (Newmeyer et al., 1986a,b). These in vitro assembled/reconstituted nuclei are surrounded by a double nuclear membrane studded with pores and are able to enter one complete round of DNA replication. Import of nuclear proteins through NPCs of the reconstituted nuclei is cytosol and ATP dependent. Formation of nuclear pores functional in the import reaction is abolished when p62 or other nucleoporins are immuno-depleted from the extracts (Dabauvalle et al., 1988, 1990; Finlay and Forbes, 1990). The other in vitro system makes use of mammalian cells whose cytoplasmic membrane has been permeabilized with digitonin (Adam et al., 1990). The role of NLS-binding proteins (NBPs) and other cytosolic

34

PAOLA GRANDI and EDUARD C. HURT

factors in nuclear import was mainly studied in this system. A 55-kDa cytoplasmic NBP could stimulate nuclear accumulation of proteins when added to the permeabilized cells (Adam et al., 1990; Adam and Gerace, 1991). Antibodies against a yeast NBP phosphoprotein of 70 kDa (NBP70) inhibit nuclear import of proteins in various in vitro assays, including semipermeabilized Drosophila tissue culture cells (Stochaj and Silver, 1992). A more detailed screen for nuclear import factors using digitonin-permeabilized cells was initiated by Moore and Blobel (Moore and Blobel, 1992). Two fractions purified from Xenopus cytosol by ion-exchange chromatography were added to the permeabilized cells together with a fluorescently labeled NLS-BSA conjugate as a synthetic karyophile for nuclear import. One fraction (fraction A) induced accumulation of the reporter protein at the nuclear periphery, and the other (fraction B) promoted the translocation of the bound karyophile into the nucleoplasm (Moore and Blobel, 1992). At least one active component of this latter fraction has been recently identified as the small GTP-binding protein Ran/TC4, and although GTP hydrolysis is required for nuclear import, the specific function of this Ras-related nuclear protein is not yet known (Moore and Blobel, 1993). The involvement of members of the heat shock protein family in nuclear transport has been discovered after depletion of these proteins from cytosolic fractions active in the in vitro transport assay. Removal of hsc70 and hsp70 caused block of active nuclear transport in permeabilized HeLa cells, and addition of the purified or bacterially expressed proteins fully restored it (Shi and Thomas, 1992). This may indicate that proteins of the heat shock family are among the soluble factors involved in the translocation step across the nuclear pores. Moreover, bacterially expressed and purified hsc70 can bind to the wild-type SV40 NLS but less well to a mutant NLS, suggesting that hsc70 may act as a shuttling carrier for NLS-containing nuclear protein (Goldfarb, 1992; Imamoto et al., 1992). Despite progress in the study of the biochemistry of some nucleoporins and soluble factors required for nuclear transport, a more complete analysis of the individual components of the nuclear pores, such as cloning of their genes, gene disruption, and creation of mutants, still represents considerable problems in higher eukaryotes. Since nuclear pore proteins are likely involved in the transport of substrates across the pore channel, availability of mutants of these proteins would allow an in vivo inspection of their function.

The Nuclear Pore Complex in Yeast

35

The lower eukaryote Saccharomyces cerevisiae represents a convenient system to study a huge organellar-like structure such as the nuclear pore complex on a molecular level. The ease of yeast genetics allows the identification of pore components in the living cell and the analysis of their physical and genetic interaction. On the other hand, yeast is also amenable to biochemical assays that are necessary to complement the genetic approaches. In this chapter we will first introduce what is known about the 5. cerevisiae nuclear pores in terms of structure and function and describe the yeast nuclear pore proteins identified so far. We will then emphasize two complementary approaches we undertook in our laboratory to identify novel nuclear pore proteins in yeast and study not only their individual function, but also how these nucleoporins are linked to each other both by physical interactions or/and overlapping functions. Finally, we describe an in vitro system for nuclear protein transport that was recently developed in P. Silver's laboratory which allows the in vitro analysis of nuclear pore protein mutants and thus could be complementary to the in vivo nuclear transport analysis.

II. STRUCTURE OF THE YEAST NUCLEAR PORE COMPLEX In most eukaryotes, nuclear pores disassemble during mitosis concomitant with the onset of nuclear envelope breakdown. In contrast, yeast undergoes a closed mitosis in which the nuclear envelope remains intact during mitotic division and accordingly nuclear pores do not disassemble (Jordan et al., 1977; Byers, 1981). Nuclear pores in yeast are formed throughout the whole cell cycle, but bursts of pore biogenesis are observed both in the Gj phase and before mitosis (Jordan et al., 1977). Although nuclear pores tend to appear like distinct organelles within the nuclear envelope, it is difficult to define the limits of a nuclear pore complex. This is due to the fact that NPCs, as shown for higher eukaryotic cells, are completely embedded in the double nuclear membrane and are attached to several other macromolecular structures, such as the nuclear lamina on the inner side of the nuclear membrane (Aebi et al., 1986) and filamentous cytoskeletal elements on the cytoplasmic side (CarmoFonseca et al., 1987). This is also the case for yeast NPCs, which are tightly attached to filamentous structures reminiscent of a nuclear lamina or a nuclear scaffold (Allen and Douglas, 1989). It is evident that isolated yeast nuclei have filamentous structures 6-10 nm thick attached to the

36

PAOLA GRANDI and EDUARD C. HURT

outer nuclear membrane, mainly at the level of the pores. If these nuclei are then treated with nucleases and salt and subjected to differential sedimentation, NPCs can be separated from most of the nuclear material. In the enriched fraction, pore complexes still appear to be directly associated with a protein meshwork organized into filaments with an estimated diameter of 8-11 nm (Allen and Douglas, 1989). These nuclear pore-attached structures resemble intermediate filaments and could represent the yeast nuclear lamina. Although a typical nuclear lamina structure could not be visualized by EM on LIS/digitonin-extracted nuclei (Cardenas et al., 1990), cross-reactivity of yeast nuclear proteins with turkey anti-lamin and anti-p58 (lamin B receptor) antibodies has suggested the existence of yeast lamins (Georgatos et al., 1989). Pore complexes can be further purified using a combination of detergents, reducing agents, and sucrose density gradient centrifugation. When these samples are spread on EM grids and negatively stained, they still have a typical nuclear pore appearance but are devoid of filamentous attachments. As in the case of higher eukaryotes, the nuclear pore complexes in yeast appear to be organized into eight subunits symmetrically arranged around a central channel; this channel is sometimes occupied by an as yet unidentified structure, possibly the transporter or collapsed filaments. Yeast NPCs are generally poorly preserved during biochemical isolation procedures; therefore single rings are often observed to be detached from the pore, and sometimes spike-like structures can be visualized protruding from individual rings and converging toward the central channel (Allen and Douglas, 1989). Rout and Blobel (1991) developed an isolation method that allows the purification of large quantities of NPCs from yeast on the basis of both negative stain-electron microscopy and coenrichment of characterized nuclear pore protein antigens. The purified nuclear pore complexes have a disc-shaped structure, with the central transporter surrounded by an annulus of octagonal symmetry. It is estimated that the purified pore preparation contains between 50 and 100 individual pore proteins.

III. COMPONENTS OF THE YEAST NUCLEAR PORE COMPLEX Conservation between yeast and higher eukaryote NPCs is not only observed at the ultrastructural level but also extends to its molecular components. In fact, monoclonal antibodies that recognize mammalian

The Nuclear Pore Complex in Yeast

37

nucleoporins also cross-react with several yeast nuclear pore proteins (Aris and Blobel, 1989; Davis and Fink, 1990; Wente et al., 1992). On the basis of this cross-reactivity, the yeast nuclear pore protein NUPl could be cloned; by indirect immunoflorescence NUPl was localized at the nuclear periphery, and it was shown that it is essential for yeast cell growth (Davis and Fink, 1990). The nuclear pore protein NSPl was identified and cloned with the help of polyclonal antibodies raised against the insoluble fraction of yeast nuclei (Hurt, 1988). The nuclear pore location of NSPl was determined by immunoelectron microscopy using affinity-purified antiNSPl antibodies (Nehrbass et al, 1990). The gene of another pore protein, NUP2, has been isolated through screening of a Xgtl 1 yeast expression library with a monoclonal antibody against mammalian nucleoporins followed by low-stringency Southern hybridization using a degenerate oligonucleotide encoding a peptide that contains the FSFG motif, a repeated amino acid sequence present in the yeast nucleoporins NSPl and NUPl (Loeb et al., 1993). Sequence comparison between mammalian and yeast nuclear pore proteins revealed a similar domain organization and, in addition, the central domains of these nuclear pore proteins contain highly conserved repeat sequences. The conserved repeat motif is the GFSFG pentapeptide within yeast NSPl, NUPl, and NUP2 and mammaUan p62 and NUP153 (Hurt, 1988; Davis and Fink, 1990; Loeb et al., 1993; Sukegawa and Blobel, 1993). These repeats (up to 28 in the NUPl protein (Davis and Fink, 1990)) represent the epitope recognized by the monoclonal antibody Mab 414 (Davis and Fink, 1990) and therefore are thought to be present in all the mammalian nuclear pore proteins reactive with this monoclonal antibody. Recently, several new nuclear pore proteins have been cloned and characterized in yeast by using either immunological or genetic approaches. The genetic method that was based on synthetic lethality will be discussed in more detail later in this review. In the immunological approach, a panel of monoclonal antibodies raised against rat liver NPC-lamina fractions has been screened for nuclear pore labeling on yeast spheroplasts through indirect immunofluorescence and immunoelectron microscopy (Wente et al., 1992). Mab 192 was finally chosen to clone the corresponding nuclear pore proteins from a yeast expression library. Three proteins were identified and termed NUP49, NUPlOO, and NUPl 16 from their calculated molecular weights (Wente et al., 1992). Sequence comparison revealed that

38

PAOLA GRANDI and EDUARD C. HURT

NUPlOO and NUPl 16 have an overall similarity of 65%, with the highest homology in their carboxy-terminal domains. The amino-terminal domain of all three proteins is characterized by repeat sequences of the GLFG type and responsible for the reactivity with Mabl92. A small number of more degenerate repeats of the GLFG type are also found in the amino-terminal part of NSPl and carboxy-terminal end of NUPl. Therefore, all of the nucleoporins identified in yeast fall into two distinct families according to the conserved repeat sequences: FSFG or GLFG. For both types of repeats a (3 structure is predicted, but no specific function has been shown so far for these highly conserved domains. There is preliminary evidence that they may interact with NLS-binding proteins (i.e., NBP70) (Stochaj and Silver, 1992; P. Silver, personal communication). The carboxy-terminal domains of NSPl and NUP49 represent the part of the proteins essential for cell viability. Moreover, the NSPl carboxyterminal domain is 50% similar and 30% identical to the carboxy-terminal domain of vertebrate p62 protein (Carmo-Fonseca et al., 1991), and both domains show a heptad repeat organization. Interestingly, the NUP49 carboxy-terminal domain is also organized in heptad repeats (Wimmer et al., 1992). We do not yet know the specific function performed by these essential heptad-repeated carboxy-terminal domains; in the intermediate filament proteins the heptad repeats are essential for coiled-coil interactions and allow the formation of dimers (Steinert and Roop, 1988). NSPl is involved in nuclear import of NLS-containing proteins (Mutvei et al., 1992; Nehrbass et al, 1993). A single amino acid substitution (E706P) in the carboxy terminal domain of NSPl causes a temperature-sensitive phenotype (Nehrbass et al., 1990, 1993). When nuclear transport of NLS-containing fusion proteins is examined in these mutants, chimeric proteins like Mata2-lacZ and Pho2-lacZ accumulate in the cytoplasm at the restrictive temperature (37°C), whereas at the permissive temperature (23°C) these reporters are transported into the nucleus (Nehrbass et al., 1993). NSPl is also involved in nuclear pore biogenesis since in cells depleted of the NSPl protein, the number of newly synthesized pores decreases, as observed by freeze-fracture EM analysis (Mutvei et al., 1992). Although the number of cloned nuclear pore proteins in yeast increases steadily, very little information is available so far as to how these different pore components interact with each other. Do they directly

The Nuclear Pore Complex in Yeast

39

interact, and if so, are they arranged in defined subcomplexes involved in distinct steps of nuclear transport? To answer these questions, it is necessary to develop new approaches for the study of the known nucleoporins and the identification of novel components of the NPC.

IV. EXPERIMENTAL APPROACHES TO STUDYING NUCLEAR PORE STRUCTURE AND FUNCTION IN YEAST In the following, we discuss three different approaches undertaken with the yeast S. cerevisiae (i.e., biochemical, genetic, and in vitro reconstitution studies), which have the potential, if combined with each other, to reveal a great deal about the role of nuclear pore proteins in nucleocytoplasmic transport. Because NPCs have a complex polypeptide composition, the understanding of the role of a given pore protein depends on the understanding of how this pore protein is integrated into the NPC structure and how it physically and functionally interacts with other components. As a candidate protein on which to start the analysis of nuclear pore protein interaction, we concentrated on the yeast NPC protein NSPl. A. Biochemical Approach One can assume that the 100 or so different NPC proteins are organized into nuclear pore substructures. Such substructures may either correspond to building blocks of the NPC or perform individual functions at the pore (e.g., being involved in protein transport into the nucleus or RNA export out of the nucleus, interaction with cytoplasmic or nuclear structures etc.). To dissect such substructures, a biochemical approach was begun; NSPl was shown to be localized at the NPC (Nehrbass et al., 1990) and involved in protein transport into the nucleus. The occurrence of heptad repeats within the essential NSPl carboxy-terminal domain makes it likely that this pore protein binds to other pore components by means of coiled-coil interactions (Steinert and Roop, 1988; Lupas et al., 1991). If so, purification of the NSPl protein under nondenaturing conditions may allow identification of copurifying components that are stably attached to NSPl. Since nuclear pore proteins are not abundant (we estimate that NSPl occurs in a concentration of 0.02% of all cellular proteins (Grandi et al., 1993)), one requires efficient purification methods to obtain chemical quantities of this pore protein in order to perform further biochemical studies. We therefore tagged NSPl with

40

PAOLA GRANDI and EDUARD C. HURT

IgG-binding sequences derived from Staphylococcus aureus protein A and started its affinity purification. The protein A tag binds with high affinity to IgG-Sepharose resin and thus allows a one-step purification of protein A fusion proteins (Moks et al., 1987). Several IgG binding units of protein A (one binding unit is called the Z domain and consists of 58 amino acids) can be tandemly fused to the protein of interest (Moks et al., 1987). The tandem arrangement of these Z domains increases the binding efficiency to the IgG-Sepharose column. It is our experience with different protein A fusion proteins that Z domains fold independently of the attached reporter protein, thus allowing both efficient affinity purification by IgG Sepharose chromatography and correct folding of the attached passenger protein. The functionality of a given protein A-nucleoporin fusion protein can be tested directly in vivo by expression of the chimeric genes in yeast cells in which the genomic copy of this nucleoporin has been disrupted. The high affinity of the Z domain for rabbit and human IgG allows the detection of the fusion proteins both on immunoblots and by immunofluorescence. On the other hand, protein A does not react as well with mouse IgG or IgM, which therefore makes it possible to perform double immunofluorescence staining. In particular, four Z domains were fused to the carboxy-terminal domain of NSPl and placed under the control of the authentic NSPl promoter (Grandi et al, 1993). This fusion gene was then expressed in cells in which the chromosomal NSPl has been destroyed. The chimeric protein complemented the otherwise nonviable nspl null mutant, demonstrating that the essential function of NSPl is not altered by attaching four Z domains derived from S. aureus protein A. The localization of the protein A-NSPl fusion protein at the NPC was shown by indirect immunofluorescence using fluorescendy conjugated IgGs as the detection reagent, and on Westem blots it was direcdy detected by probing with rabbit IgG conjugated to horseradish peroxidase. Under these conditions, only the protein A fusion protein and none of the endogenous yeast proteins were reactive. To affinity purify NSPl and its interacting components yeast cells functionally expressing the protein A-NSPl fusion protein were converted into spheroplasts and lysed in a non-ionic detergent buffer (Grandi et al., 1993). Under these conditions the tagged nucleoporin was completely solubilized. We do not know whether nuclear pores desintegrate under the lysis conditions chosen, but disruption of the entire nucleus does certainly occur under these conditions, as judged from the release

The Nuclear Pore Complex in Yeast

41

of nuclear and nucleolar marker proteins. The solubilized material was then loaded onto a IgG-Sepharose column, and after several washing steps the bound fraction was eluted by lowering the pH. In this one-step affinity purification procedure an efficient recovery of the tagged NSPl protein could be achieved, the yield being approximately 50-60% of the loaded protein. According to this high recovery of the protein A-NSPl fusion protein, we estimate that between 10 and 20 NSPl molecules are present per nuclear pore complex, a value that comes close to those reported for mammalian nuclear pore proteins such as p62 (Snow et al., 1987). The analysis of the fraction eluted from the IgG-Sepharose column by SDS-polyacrylamide gel electrophoresis followed by silver or Coomassie blue staining revealed a simple protein profile of a few prominent bands (Figure 1). Besides the tagged NSPl fusion protein, four major proteins were specifically recovered in the eluted fraction.

..ProtA-NSPl •p54 '-NUP49

Figure 1. A NSPl -containing NPC subcomplex. The purified NSP1 complex was analyzed by SDS-PAGE followed by silver staining according to the method of Grandi et al. (1993). The positions of NIC96, p80, ProtANSP1, p54, and NUP49 as well as of two main degradation products (*) of ProtA-NSPI are indicated. The location of this complex at the nuclear pore complex is shown in a schematic way.

42

PAOLA GRANDI and EDUARD C. HURT

Two of these proteins were identified as GLFG-containing nucleoporins, NUP49 and p54, since these two bands exhibit reactivity with Mabl92, previously shown to be specific for the GLFG nucleoporin protein family (Wenteetal., 1992). Another protein copurifying with NSPl was a polypeptide in the 90 kDa range, so far of unknown structure and function. Since we could obtain sufficient quantities of this protein for microsequencing, it was possible to design degenerate oligonucleotide primers and clone its gene by polymerase chain reaction. The gene encodes a novel protein of 96 kDa and thus was named NIC96 (for nucleoporin interacting component). NIC96 does not contain repeat sequences of the FSFG or GLFG type, but has heptad repeats in its amino-terminal domain similar to those found in the carboxy-terminal domains of NSPl and NUP49. Subsequently, we could show that NIC96 is an essential nuclear pore protein that is in physical interaction with both GLFG and FSFG nucleoporins. In a complementary experiment, we tagged NIC96 with IgG binding sequences from protein A and affinity purified the fusion protein under nondenaturing conditions. Together with protein A-NIC96, NSPl, NUP49, and p54 copurify with IgG-Sepharose chromatography (unpublished results). In conclusion, this biochemical analysis has shown that NSPl, NIC96, NSP49, and p54 exist in a multimeric complex that can be extracted from the nuclear pores under the chosen lysis conditions (Figure 1). Further biochemical studies can now be performed to map all of the interactions among the individual components of this complex: cross-linking experiments and biochemical reconstitution of the complex by use of the in vitro translated nucleoporins may help to understand the actual order of assembly in this nuclear pore subcomplex. It is also of considerable interest to understand the biogenesis of this subcomplex and ultimately of the NPC itself in the living cell; pulse/chase experiments combined with subcellular fractionation should help to reveal the stages in NSPl-complex assembly. Furthermore, if nucleoporins different from those of the NSPl-complex are organized in similar NPC subcomplexes, it might be possible to purify and characterize them and to clone the genes of these additional components. A step-by-step procedure will therefore allow the identification of more and more components of the nuclear pore and simultaneously an understanding of how they interact with each other.

The Nuclear Pore Complex in Yeast

43

B. Synthetic Lethality as a Genetic Approach to Studying the Nuclear Pore Complex A genetic analysis of the nuclear pore complex has the potential to complement and/or extend the biochemical characterization of this large 125-MDa organelle. In particular, the network of in vivo interactions between various nuclear pore proteins may be unraveled by genetic methods, and it may be possible to study not only pore proteins that physically interact but also those that share an overlapping function. In yeast, different genetic approaches can be undertaken in order to study the functional interaction of proteins in the living cell (Botstein and Fink, 1988). In the past, extragenic suppressor approaches were successful that aim for a gain of function of a mutant gene exhibiting a conditional phenotype (for a review see Broach, 1986). In the case of the nuclear pore complex and its structural components, extragenic suppressors (single-copy or high-copy suppressors) of temperature-sensitive mutants of nuclear pore proteins have not been isolated so far (E. Hurt, unpublished results). However, the approach of synthetic lethality could be successfully applied to the nuclear pore proteins NSPl (Wimmer et al, 1992) and NUP49 (V. Doye, unpublished results). A synthetic lethal phenotype can be the result of combining mutant alleles of two different genes, whereas each mutant allele alone still allows the cells to grow (Huffaker et al., 1987; Bender and Pringle, 1991). It was therefore suggested that synthetic lethality provides genetic proof that gene products physically interact with each other or share overlapping functions. We reasoned that the nuclear pore protein NSPl is functionally active within a network of protein-protein interactions at the nuclear pore. Accordingly, one could screen for synthetic lethal mutants in which the combination of a mutant nspl allele with another mutated gene (called NSP-X) would cause cell death, whereas each mutant allele alone would not impair cell viability (Figure 2). We used a visual red/white colony sectoring system in yeast (Huffaker et al., 1987; Kranz and Holm, 1990; Bender and Pringle, 1991; Costigan et al., 1992) to identify nuclear pore mutants with a synthetic lethal phenotype. The screen for synthetic lethal mutations was performed in a genetic background of a mutated NSPl protein; the mutation chosen exhibits a single amino acid substitution (L640S) within the essential NSPl carboxy-terminal domain. The mutant strain is viable at permissive temperatures (e.g., at 23 °C and 30°C) but stops growing at 37°C. A screening yeast strain was constructed that lacks the chromosomal gene

PAOLAGRANDI and EDUARD C. HURT

44

•

A A A •• •• •• •• •• !NSP-X

^m ^H

^i^^

••••• •••••

•••••

•• •• •• ••

'V5-^ ^ NSP-X

•• •• •• •• •• ••

j

.

Synt

'\^-y <%nsp-X

Figure 2. A model to explain synthetic lethality. The yeast cell is depicted as a building that consists of bricks, including the essential nuclear pore protein NSP1 and a putative interacting component NSP-X. (A) Both NSPl and NSP-X are intact. (B) NSP1 is mutated in its carboxy-terminal domain (nsp1); since NSP-X is still intact, the stability of the building is not affected (viability). (C) Both nspl and nsp-X are mutated; under these conditions, the building becomes unstable and finally collapses (synthetic lethality, SL). (D) Mutation of an unrelated brick that is in neither physical nor functional interaction with nspl does not affect the stability of the building (viability).

of NSPl but is complemented by two different NSPl alleles offered to the cell on two different plasmids: (1) the wild-type NSPl linked to a plasmid that confers red color to yeast colonies and (2) a mutated nspl gene supplied on a second plasmid that gives rise to a white colony appearance if the red color plasmid is missing. Thus, two different NSPl proteins are expressed and the cell can use one or the other for cell growth. In fact, since the hypothetical component NSPl-X is intact in the starting situation, both the wild-type and the mutated NSPl proteins give rise to functional interaction with NSP-X without affecting viability at temperatures between 20 °C and 30 °C. Accordingly, this screening strain forms red/white sectoring colonies because it can afford to lose the

The Nuclear Pore Complex in Yeast

45

wild-type NSPl, which is linked to the red color plasmid. The screening strain was mutagenized by UV light to generate random mutations within the yeast genome with the hope of hitting one or the other gene, which can give rise to synthetic lethality. Synthetic lethals of nspl were then identified on plates as red, nonsectoring colonies. The appearance of red colonies could mean that cells of a colony cannot afford to lose the wild-type NSPl gene, which is inserted into the red color plasmid. However, there can be other scenarios that cause red colony appearance and are not due to synthetic lethality, but these false positives can be easily identified by further genetic tests (Bender and Pringle, 1991). Among the 200,000 screened colonies, 29 red colonies finally remained that were synthetic lethal mutants of NSPL To clone the genes that give rise to the synthetic lethal phenotype, we started with one of the many red mutants, mutant SL32, and transformed it with a yeast genomic DNA library. If transformants take up a plasmid that contains the corresponding wild-type gene of NSP-X, they regain the capability to lose the wild-type NSPl gene and the colony will exhibit a red/white sectoring phenotype. By complementation of SL32, a new pore protein called NSPl 16 could be cloned (Wimmer et al., 1992). TheNSP116 gene not only restored red/white sectoring in SL32, but also in a further five SL mutants, suggesting that they belong to the same complementation group. Subsequently, one of the remaining SL mutants not complemented by the NSPl 16 gene was taken, and the transformation procedure with the yeast genomic library was repeated. In this case, another gene designated NSP49 could be isolated that complemented this and other four SL mutants. Meanwhile, several other genes complementing some of the remaining SL mutants were isolated by this "subtractive" cloning procedure. The two genes that were most frequently found in the SL screen encode novel nucleoporins NSP49 and NSPl 16. Like the previously described nucleoporins NSPl (Nehrbass et al., 1990), NUPl (Davis and Fink, 1990), and NUP2 (Loeb et al., 1993), NSP49 and NSP116 contain many repeat sequences, but with a new repetitive sequence motif GLFG, which classifies them as a subclass of nucleoporins. Double immunofluorescence of yeast cells using antibodies against NSPl and a monoclonal antibody recognizing GLFG nucleoporins (Wente et al, 1992) reveals colocalization of NSPl and the GLFG nucleoporins at the rim of the nuclear envelope (Wimmer et al., 1992). Therefore, NSP49 and NSPl 16 may not only functionally overlap but also physically interact with NSPl at the nuclear pores. In fact, this could be shown

46

PAOLA GRANDI and EDUARD C. HURT

biochemically for NSP49, which is in a multimeric complex with NSPl (see above). Interestingly, both nuclear pore proteins NSPl 16 and NSP49 were independently cloned by screening an expression library using a monoclonal antibody against mammalian nuclear pore proteins. Wente and Blobel called these nucleoporins NUPl 16 and NUP49 (Wente et al., 1992). To avoid problems of nomenclature, we will also use the names NUP116andNUP49. What are the other remaining complementation groups of the synthetic lethal screen? Interestingly, two other synthetic lethal strains were complemented by the NIC96 gene (H. Tekotte, unpublished results). The NIC96 protein physically interacts with NSPl and constitutes a nuclear pore subcomplex together with NUP49 (see also above). Two further SL mutants are complemented by a gene that encodes a novel nucleoporin with GLFG repeats designated NUP145 (E. Fabre, manuscript in preparation). Several other genes complementing further SL mutants have also been cloned and are in the process of being characterized. To summarize, we have obtained 29 SL mutants that belong in at least seven different complementation groups, but the number is likely to be higher, because the wild-type genes of several SL mutants have not been cloned so far. Recently, synthetic lethality was demonstrated between NUPl and NUP2, as well as between NSPl and NUP2 (Loeb et al., 1993). In this case, different mutant alleles of the three nucleoporins were co-expressed in the cell and analyzed for loss of viability (Loeb et al., 1993). The NUP2 gene is not essential for cell viability, but when a nup2 null mutant is combined with a NUPl N-terminal truncation, a synthetic lethal phenotype results at any temperature. This phenotype is rescued by the 175amino acid amino-terminal domain of NUPl, which in a background of wild-type NUP2 is not essential. Similarly, the combination of a null nup2 with a modified NSPl gene that only expresses the carboxy-terminal domain of the protein is lethal to the cells; however, this SL phenotype can be rescued by the expression of the NUP2 amino-terminal domain (Loeb et al., 1993). Surprisingly, this part of NUP2 does not contain nucleoporin typical repeat sequences such as FSFG or GLFG repeats. However, the fact that mutations simultaneously present in two or more nucleoporins cause cell death indicates that these proteins might have overlapping functions or that they interact directly with each other. These physico-functional interactions among nucleoporins are clearly essential for cell viability.

The Nuclear Pore Complex in Yeast

47

The results of the synthetic lethal screen have shown that the conceptual basis of our genetic approach is sound in the sense that (i) novel nucleoporins can be identified with this method and (ii) a genetic network of interaction between NSPl and other members of nuclear pore protein complex exists. It has also been proved for the first time that the nuclear pore complex is amenable to genetic approaches. For the future, we expect that genetics will be exploited as a powerful way to study the nuclear pore complex in the living cell. C. In Vitro Nuclear Transport Assays

In vitro nuclear import systems that faithfully reproduce the nuclear transport reaction of the living cell are now available for yeast. Such a cell-free nuclear localization system can be used to study the requirement of factors for nuclear transport, to analyze the primary defect of nuclear transport mutants, and to complement the deficiency of these mutants by the addition of active fractions isolated from wild-type yeast cells. In addition, antibodies against nuclear pore proteins and against factors required for nuclear transport could be tested in the in vitro assay for their inhibitory effect on nuclear protein accumulation. The first two reports on in vitro nuclear transport assays in yeast used purified yeast nuclei (KaUnich and Douglas, 1989; Garcia-Bustos et al., 1991b). KaUnich and Douglas (1989) showed that radiolabeled SV40 large T-antigen and Xenopus nucleoplasmin were efficiently imported into isolated nuclei and protected against extemally added trypsinagarose. The nuclear uptake required an intact NLS on the imported molecule and was ATP, calcium, time, and temperature dependent but did not require cytosol. The import of these substrates was inhibited after trypsin treatment of the outer nuclear membrane. This suggests that the yeast nuclear envelope recognizes NLSs from heterologous proteins with a receptor machinery that is similar to the higher eukaryotic nuclear import system. In a similar assay, Hall and co-workers tested authentic yeast nuclear proteins like the transcription factors MCMl and STE12 for import into purified yeast nuclei (Garcia-Bustos et al, 1991b). To identify the receptors that bind nuclear import substrates in vitro, an endogenous yeast nuclear protein, MCMl, was tested for binding to the nuclear envelope of purified but ATP-depleted yeast nuclei (Garcia-Bustos et al., 1991a). It was found that MCMl binds to high-affinity binding sites of the nuclear envelope in a saturable fashion. The dissociation constant for binding is 0.5 |iM, with approximately 10-30 binding sites

48

PAOLA GRANDI and EDUARD C. HURT

per nuclear pore complex. Binding of MCMl could be competed for by two other nuclear proteins, STE12 and SWI5, but not by the nucleolar proteins NOPl or RAPl. These data indicate that different receptor(s) for nuclear proteins might be present in limited amounts at the nuclear pore complex. In the in vitro assays described above, the exact location of imported nuclear proteins was not shown by immunofluorescence staining of the purified nuclei. Recently, an in vitro system for nuclear protein import was developed in yeast that faithfully measures nuclear transport and shows the same requirements as observed under in vivo conditions (Schlenstedt et al., 1993). In this in vitro assay, a visual demonstration of nuclear accumulation of the transport substrate was achieved by fluorescence microscopy, starting with yeast spheroplasts that were permeabilized by slow freezing-thawing. Semi-open yeast cells have holes in the cytoplasmic membrane and accordingly have lost a considerable amount of the cytoplasmic content, but the nucleus remains intact with a double nuclear membrane and nuclear pore complexes. Macromolecules (such as antibodies or nuclear transport substrates) can penetrate into the semi-open cells and thus gain access to the cellular interior, but the nuclear membrane remains as a barrier for further diffusion into the nucleus. When a fluorescently labeled nuclear reporter protein (the NLS of SV40 was coupled to fluorescently tagged human serum albumin (HSA)) is added to the semi-intact yeast cells, nuclear accumulation of the NLS conjugate was observed, but only if ATP and cytosol were present. Thus, nuclear transport had occurred through the nuclear pore complexes in an active process. If a mutated NLS that failed to be imported into the nucleus in the in vivo assays was used, no import was observed. In the absence of ATP and cytosol, only NLS-dependent binding at the nuclear envelope was seen. This suggests a two-step mechanism for nuclear transport in yeast that is ATP-independent: binding to the nuclear envelope followed by cytosol- and ATP-dependent translocation across the nuclear pore complexes. The same two-step mechanism was found in mammalian (Adam et al., 1990) andDrosophila (Stochaj and Silver, 1992) semi-intact cells and Xenopus reconstituted nuclei (Newmeyer et al., 1986). When the nuclear transport substrate is first bound to the nuclear periphery in the absence of ATP, it colocalizes with the nuclear pore protein NSPl as revealed by confocal immunofluorescence microscopy, suggesting a docking reaction at the nuclear pore complex. This bound NLS conjugate behaved like a translocation intermediate, because it could be chased into the nucleus upon addition of

The Nuclear Pore Complex in Yeast

49

cytosol and ATP. At least one factor required for in vitro nuclear transport was shown to be a NEM-sensitive protein within the added cytosol. Interestingly, cytosol from HeLa cells could substitute for the yeast cytosol, suggesting that the cytosol-derived nuclear import factor(s) are conserved between evolutionary distant species such as yeast and humans. A role of yeast nuclear pore proteins in the in vitro nuclear import reaction could be shown by two independent criteria. First, import (but not binding) was inhibited by pretreating semi-open cells with antibodies against the nuclear pore protein NSPl. These antibodies were shown to bind to the cytoplasmic site of the nuclear pore complexes, thereby blocking translocation through the pores. Second, semi-open cells prepared from a temperature-sensitive nspl mutant that in vivo no longer imports nuclear reporter proteins (Nehrbass et al., 1990,1993) were also defective in the in vitro nuclear transport reaction, even in the presence of ATP and cytosol derived from wild-type cells. Binding of the nuclear transport substrate to the nuclear envelope, however, was still observed, although it was reduced to about 30% when compared to wild-type cells. In a similar way, a thermosensitive mutation in another nucleoporin, nup49, completely blocked nuclear transport in the mutant semi-open cells. Another yeast mutant defective in nucleocytoplasmic transport npl3 was subjected to the in vitro import assay. NPL3 encodes a nuclear protein with homology to RNA-binding proteins that shuttle between the nucleus and the cytoplasm (Bossie et al., 1992). Semi-intact cells prepared from npl3 mutants and shifted to the nonpermissive temperature stop importing NLS-HSA transport substrates. In summary, these data show that the reconstituted nuclear transport system makes it possible to analyze nuclear transport mutants under in vitro conditions and to distinguish between a defect in binding to the nuclear envelope and import into the nucleus. Furthermore, the in vitro system is amenable to biochemical manipulation (e.g., depletion or addition of import factors), which is a valuable completion of the analysis of the in vivo phenotype of nuclear transport mutants.

V. CONCLUSIONS In our laboratory, we favor the yeast S. cerevisiae as an experimental system for studying the NPC and propose combining the approaches available in the yeast system for identification of the nuclear pore components, for the study of their interactions, and for analysis of their in vivo role in pore function. These approaches include genetics, bio-

50

PAOLA GRANDI and EDUARD C. HURT

chemistry, and in vitro reconstitution. No other eukaryotic organism currently offers such a palette of methods to be used for studying the NPC. The approach of synthetic lethality led to the identification of novel nuclear pore proteins; only a few of these proteins are in physical (direct) interaction with each other, as confirmed by biochemical analysis. The other newly identified proteins might represent components that have only transient physical interaction or, on the other hand, have functional overlaps. The high number of complementation groups found in the synthetic lethal screen seems not to account for only direct protein-protein interactions among all of these nucleoporins. We have complemented our genetic analysis with biochemical methods. It has emerged that affinity purification of protein A-tagged nucleoporins is a powerful approach, making it possible to obtain large amounts of purified pore proteins. This allows us to make a physicochemical characterization of these nuclear pore components. Furthermore, since affinity purification is performed under nondenaturing conditions, the identification of NPC subcomplexes and the nearestneighbor relationship analysis can be performed. In the case of NSPl, a distinct subcomplex of the NPC could be isolated that is now useful for further in vitro studies, including the reconstitution of the subcomplex and its assembly into the NPC. By tagging other nuclear pore proteins that were isolated in the genetic screen, we found that some occur in other subcomplexes with a unique polypeptide composition (C. Wimmer, unpublished data). Using the biochemical and the genetic approach described, we aim to "walk" through the whole NPC and map the structural and functional interactions of its components. This method is an alternative method to the biochemical isolation of whole NPCs from yeast and successive analysis of the proteins present in this preparation (Rout and Blobel, 1991). All of the cloned proteins of the NSPl subcomplex are essential for cell viability. It is interesting to note that these nuclear pore proteins (NSPl, NUP49, NIC96) all exhibit heptad repeats in their sequences that represent domains essential for viability. These domains could be involved in coiled-coil interactions, by forming of either homo- or heterodimers and thereby holding the NPC subcomplex together. In addition, these heptad repeats could have a role in targeting the individual partners at the nuclear envelope and facilitate assembly into the NPC. Generation of mutant pore proteins that have amino acid exchanges within the heptad repeats would aid in the understanding of how these

The Nuclear Pore Complex in Yeast

51

different pore proteins physically interact with each other, how they become incorporated into the nuclear pore complex, and how they perform their function at the NPC. With the in vitro nuclear transport system now available in yeast, we can directly study the role of a given nuclear pore protein in nuclear transport. The kinetics of transport may be determined and different substrates can be analyzed. Finally, nuclear transport may be reconstituted in mutated strains by adding purified components from wild-type cells. This would provide a proof, which is otherwise difficult to show in vivo, that these components have a direct role in nuclear transport. In summary, we have outlined in this review biochemical and genetic methods of studying the nuclear pore complex in the yeast S. cerevisiae. It will be the combination of these different systems, including in vitro reconstitution and in vitro nuclear transport studies, that should give further insight into the mechanism of how nuclear pore proteins perform their function in the living cell. VI.

SUMMARY

Nuclear pore complexes have been known for many years to mediate nucleocytoplasmic transport, despite the fact that little information is available about the pore components required for this process. Progress has now been made in identifying new nuclear pore components by a genetic and biochemical analysis of the nuclear pore complex (NPC) in the yeast S. cerevisiae. This combination of biochemistry and genetics has the potential to unravel structure/function relationships of nuclear pore proteins (nucleoporins) in the living cell. Furthermore, a recently developed in vitro assay for nuclear protein accumulation makes it possible to analyze the specific defect of nuclear transport mutations and to rescue the deficiency by the addition of purified components. REFERENCES Adam, S.A. & Gerace, L. (1991). Cytosolic proteins that specifically bind nuclear localization signals are receptors for nuclear import. Cell 66, 837-847. Adam, S.A., Marr, R.S., & Gerace, L. (1990). Nuclear protein import in permeabilized mammalian cells requires soluble cytoplasmic factors. J. Cell Biol. I l l , 807-816. Aebi, U., Cohn, J., Buhle, L., & Gerace, L. (1986). The nuclear lamina is a meshwork of intermediate type filaments. Nature 323, 560-564. Akey, C.W. (1989). Interactions and structure of the nuclear pore complex revealed by cryo-electron microscopy. J. Cell Biol. 109, 955-970.

52

PAOLA CRANDI and EDUARD C. HURT

Akey, C.W. (1990). Visualization of transport-related configurations of the nuclear pore transporter. Biophys. J. 58, 341-355. Akey, C.W. (1992). The nuclear pore complex. Curr. Opin. Struct. Biol. 2,258-263. Akey, C.W. & Goldfarb, D.S. (1989). Protein import through the nuclear pore complex is a multistep process. J. Cell Biol. 109, 971-982. Akey, C.W. & Radermacher, M. (1993). Architecture of the Xenopus nuclear pore complex revealed by three-dimensional cryo-electron microscopy. J. Cell Biol. 122,1-19. Allen, J.L. & Douglas, M.G. (1989). Organization of the nuclear pore complex in Saccharomyces cerevisiae. J. Ultrastruct. Mol. Struct. Res. 102,95-108. Aris, J.P. & Blobel, G. (1989). Yeast nuclear envelope proteins cross-react with an antibody against mammalian pore complex proteins. J. Cell Biol. 108,2059-2067. Bataille, N., Helser, T, & Fried, H.M. (1990). Cytoplasmic transport of ribosomal subunits microinjected into the Xenopus laevis oocyte nucleus: a generalized, facilitated process. J. Cell Biol. I l l , 1571-1582. Bender, A. & Pringle, J.R. (1991). Use of a screen for synthetic lethal and multicopy suppressee mutants to identify two new genes involved in morphogenesis in Saccharomyces cerevisiae. Mol. Cell. Biol. 11, 1295-1305. Blobel, G. (1985). Gene gating: a hypothesis. Proc. Natl. Acad. Sci. USA82,8527-8529. Bonner, WM. (1975). Protein migration into nuclei. I. Frog oocyte nuclei in vivo accumulate microinjected histones, allow entry to small proteins, and exclude large proteins. J. Cell Biol. 64, 421^30. Bossie, M.A., DeHoratius, C , Barcelo, G., & Silver, P. (1992). A mutant nuclear protein with similarity to RNA binding proteins interferes with nuclear import in yeast. Mol. Biol. Cell 3, 875-893. Botstein, D. & Fink, G.R. (1988). Yeast: an experimental organism for modem biology. Science 240, 1439-1443. Breeuwer, M. & Goldfarb, M. (1990). Facilitated nuclear transport of histone HI and other small nucleophilic proteins. Cell 60, 999-1008. Broach, J.R. (1986). New approaches to a genetic analysis of mitosis. Cell 44, 3-4. Byers, B. (1981). Cytology of the yeast life cycle. In: The Molecular Biology of the Yeast Saccharomyces cerevisiae: Life and Inheritance (Strathem, J.N., Jones, E.W., and Broach, J.R., eds.), pp. 59-96. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY. Callan, H.G., Randall, J.T., & Tomlin, S.G. (1949). An electron microscope study of the nuclear membrane. Nature 163, 280-281. Cardenas, M.E., Laroche, T, & Gasser, S.M. (1990). The composition and morphology of yeast nuclear scaffolds. J. Cell Sci. 96,439-450. Carmo-Fonseca, M., Cidadao, A.J., & David-Ferreira, J.F. (1987). Filamentous crossbridges link intermediate filaments to the nuclear pore complexes. Eur. J. Cell Biol. 45, 282-290. Carmo-Fonseca, M., Kern, H., & Hurt, E.C. (1991). Human nucleoporin p62 and the essential yeast nuclear pore protein NSPl show sequence homology and a similar domain organization. Eur. J. Cell Biol. 55, 17-30. Cordes, V., Waizenegger, I., & Krohne, G. (1991). Nuclear pore complex glycoprotein p62 of Xenopus laevis and mouse: cDNA cloning and identification of its glycosylated region. Eur. J. Cell Biol. 55, 3 1 ^ 7 .

The Nuclear Pore Complex in Yeast

53

Costigan, C , Gehrung, S., & Snyder, M. (1992). A synthetic lethal screen identifies SLKl, a novel protein kinase homolog implicated in yeast cell morphogenesis and cell growth. Mol. Cell. Biol. 12, 1162-1178. Dabauvalle, M.-C. & Scheer, U. (1991). Assembly of nuclear pore complexes in Xenopus egg extract. Biol. Cell 72, 25-29. Dabauvalle, M.C., Benavente, R., & Chaly, N. (1988). Monoclonal antibodies to a M^. 68,000 pore complex glycoprotein interfere with nuclear protein uptake in Xenopus oocytes. Chromosoma 97, 193-197. Dabauvalle, M.-C, Loos, K., & Scheer, U. (1990). Identification of a soluble precursor complex essential for nuclear pore assembly. Chromosoma 100, 56-66. Davis, L.I. & Blobel, G. (1986). Identification and characterization of a nuclear pore complex protein. Cell 45, 699-709. Davis, L.I. & Fink, G.R. (1990). The NUPl gene encodes an essential component of the yeast nuclear pore complex. Cell 61, 965-978. Dingwall, C. (1991). Transport across the nuclear envelope: enigmas and explanations. BioEssays 13, 213-218. Dingwall, C. & Laskey, R.A. (1986). Protein import into the nucleus. Annu. Rev. Cell Biol. 2, 367-390. Dingwall, C. & Laskey, R.A. (1991). Nuclear targeting sequences—a consensus? Trends Biochem.Sci. 16,478-481. Feldherr, CM. & Akin, D. (1990). EM visualization of nucleocytoplasmic transport processes. Electron Microsc. Rev. 3,73-86. Finlay, D.R. & Forbes, D.J. (1990). Reconstitution of biochemically altered nuclear pores: transport can be ehminated and restored. Cell 60, 17-29. Finlay, D.R., Newmeyer, D.D., Price, T.M., & Forbes, D.J. (1987). Inhibition of in-vitro nuclear transport by a lectin that binds to nuclear pores. J. Cell Biol. 104,189-200. Finlay, D.R., Meier, E., Bradley, P, Horecka, J., & Forbes, D.J. (1991). A complex of nuclear pore proteins required for pore function. J. Cell Biol. 114, 169-183. Fischer, U. & Ltihrmann, R. (1990). An essential signaUng role for the m3G cap in the transport of Ul snRNPto the nucleus. Science 249, 786-790. Fischer, U., Darzynkiewicz, E., Tahara, S.M., Dathan, N.A., Ltihrmann, R., & Mattaj, I.W. (1991). Diversity in the signals required for nuclear accumulation of U snRNPs and variety in the pathways of nuclear transport. J. Cell Biol. 113, 705-714. Fischer, U., Sumpter, V., Sekine, M., Satoh, T, & Luhrmann, R. (1993). Nucleo-cytoplasmic transport of U snRNPs: definition of a nuclear location signal in the Sm core domain that binds a transport receptor independently of the m3G cap. EMBO J. 12, 573-583. Garcia-Bustos, J.F., Wagner, P., & Hall, M.N. (199 la). Nuclear import substrates compete for a limited number of binding sites. Evidence for different classes of yeast nuclear import receptors. J. Biol. Chem. 266, 22303-22306. Garcia-Bustos, J.F., Wagner, P., & Hall, M.N. (1991b). Yeast cell-free nuclear protein import requires ATP hydrolysis. Exp. Cell Res. 192, 213-219. Georgatos, S.D., Maroulakou, I., & Blobel, G. (1989). Lamin A, lamin B, and lamin B receptor analogues in yeast. J. Cell Biol. 108, 2069-2082. Gerace, L., Ottaviano, Y, & Kondor-Koch, C (1982). Identification of a major polypeptide of the nuclear pore complex. J. Cell Biol. 95, 826-837.

54

PAOLA GRANDI and EDUARD C. HURT

Goldberg, M.W. & Allen, T.D. (1992). High resolution scanning electron microscopy of the nuclear envelope: demonstration of a new, regular, fibrous lattice attached to the baskets of the nucleoplasmic face of the nuclear pores. J. Cell Biol. 119, 1429-1440. Goldfarb, D.S. (1992). Are the cytosolic components of the nuclear, ER, and mitochondrial import apparatus functionally related? Cell 70,185-188. Goldfarb, D.S., Gari^py, J., Schoolnik, G., & Kornberg, R.D. (1986). Synthetic peptides as nuclear localization signals. Nature 322, 641-644. Grandi, P., Doye, V., & Hurt, E.C. (1993). Purification of NSPl reveals complex formation with "GLFG" nucleoporins and a novel nuclear pore protein NIC96. EMBO J. 12,3061-3071. Greber, U.F. & Gerace, L. (1992). Nuclear protein import is inhibited by an antibody to alumenalepitopeof a nuclear pore complex glycoprotein. J. Cell Biol. 116,15-30. Greber, U.F., Senior, A., & Gerace, L. (1990). A major glycoprotein of the nuclear pore complex is a membrane-spanning polypeptide with a large luminal domain and a small cytoplasmic tail. EMBO J. 9, 1495-1502. Hamm, J. & Mattaj, I.W. (1990). Monomethylated cap structures facilitate RNA export from the nucleus. Cell 63,109-118. Hamm, J., Darzynkiewicz, E.,Tahara, S.M., & Mattaj, I.W. (1990). The rimethylguanosine cap structure of Ul snRNA is a component of a bipartite nuclear targeting signal. Cell 62, 569-577. Hinshaw, J.E., Carragher, B.O., & Milligan, R.A. (1992). Architecmre and design of the nuclear pore complex. Cell 69, 1133-1141. Holt, G.D., Snow, CM., Senior, A., Haltiwanger, R.S., Gerace, L., & Hart, G.W. (1987). Nuclear pore complex glycoproteins contain cytoplasmically disposed 0-linked N-acetylglucosamine. J. Cell Biol. 104, 1157-1164. Huffaker, T.C., Hoyt, M.A., & Botstein, D. (1987). Genetic analysis of the yeast cytoskeleton. Annu. Rev. Genet. 21,259-284. Hurt, E.C. (1988). A novel nucleoskeletal-like protein located at the nuclear periphery is required for the life cycle of Saccharomyces cerevisiae. EMBO J. 7,4323-4334. Imamoto, N., Matsuoka, Y, Kurihara, T, Kohno, K., Miyagi, M., Sakiyama, R, Okada, Y, Tsunasawa, S., & Yoneda, Y (1992). Antibodies against 70-kD heat shock cognate protein inhibit mediated nuclear import of karyophilic proteins. J. Cell Biol. 119,1047-1061. Izaurralde, E. & Mattaj, I.W. (1992). Transport of RNA between the nucleus and the cytoplasm. Semin. Cell Biol. 3, 279-288. Jamik, M. & Aebi, U. (1991). Toward a more complete 3-D structure of the nuclear pore complex. J. Struct. Biol. 107, 291-308. Jordan, E.G., Severs, N.J., & Williamson, D.H. (1977). Nuclear pore formation and the cell cycle in Saccharomyces cerevisiae. Exp. Cell Res. 104, 446^49. Kalinich, J.F. & Douglas, M.G. (1989). In vitro translocation through the yeast nuclear envelope. J. Biol. Chem. 264, 17979-17989. Kastner, B., Bach, M., & Luhrmann, R. (1990). Electron microscopy of small nuclear ribonucleoprotein (snRNP) particles U2 and U5: evidence for a common structure-determining principle in the major U snRNP family. Proc. Natl. Acad. Sci. USA 87, 1710-1714.

The Nuclear Pore Complex in Yeast

55

Kranz, J.E. & Holm, C. (1990). Cloning by function: an alternative approach for identifying yeast homologs of genes from other organisms. Proc. Natl. Acad. Sci. USA 87, 6629-6633. Loeb, J.D.J., Davis, L.L, & Fink, G.R. (1993). NUP2, a novel yeast nucleoporin, has functional overlap with other proteins of the nuclear pore complex. Mol. Biol. Cell 4, 209-222. Lupas, A., Van Dyke, M., & Stock, J. (1991). Predicting coiled coils from protein sequences. Science 252, 1162-1164. Maul, G.G. (1977). The nuclear and cytoplasmic pore complex: structure, dynamics, distribution and evolution. Int. Rev. Cytol. 6, 75-186. Moks, T, Abrahms^n, L., Osterlof, B., Josephson, S., Ostling, M., Enfors, S.-O., Persson, I., Nilsson, B., & Uhlen, M. (1987). Large-scale affinity purification of human insulin-like growth factor I from the culture medium of E. coli. Biotechnology 5, 379-382. Moore, M.S. & Blobel, G. (1992). The two steps of nuclear import, targeting to the nuclear envelope and translocation through the nuclear pore, require different cytosohc factors. Cell 69, 939-950. Moore, M.S. & Blobel, G. (1993). The GTP-binding protein Ran/TC4 is required for protein import into the nucleus. Nature 365, 661-663. Mutvei, A., Dihlmann, S., Herth, W., & Hurt, E.C. (1992). NSPl depletion in yeast affects nuclear pore formation and nuclear accumulation. Eur. J. Cell Biol. 59, 280-295. Nehrbass, U., Kern, H., Mutvei, A., Horstmann, H., Marshallsay, B., & Hurt, E.C. (1990). NSPl: a yeast nuclear envelope protein localized at the nuclear pores exerts its essential function by its carboxy-terminal domain. Cell 61, 979-989. Nehrbass, U., Fabre, E., Dihlmann, S., Herth, W., & Hurt, E.C. (1993). Analysis of nucleo-cytoplasmic transport in a thermosensitive mutant of the nuclear pore protein NSPl. Eur. J. Cell Biol. 62, 1-12. Newmeyer, D.D. & Forbes, D.J. (1988). Nuclear import can be separated into distinct steps in vitro: nuclear pore binding and translocation. Cell 52, 641-653. Newmeyer, D.D. & Forbes, D.J. (1990). An N-ethylmaleimide-sensitive cytosolic factor necessary for nuclear pore protein import: requirement in signal-mediated binding to the nuclear pore. J. Cell Biol. 110,547-557. Newmeyer, D.D., Finlay, D.R., & Forbes, D.J. (1986a). In vitro transport of fluorescent nuclear protein and exclusion of non-nuclear proteins. J. Cell Biol. 103, 20912102. Newmeyer, D.D., Lucocq, J.M., Burglin, T.R., & De Robertis, E.M. (1986b). Assembly in vitro of nuclei active in nuclear protein transport: ATP is required for nucleoplasmin accumulation. EMBO J. 5, 501-510. Osborne, M.A. & Silver, P.A. (1993). Nucleocytoplasmic transport in the yeast Saccharomyces cerevisiae. Annu. Rev. Biochem. 62,219-254. Park, M.K., D*Onofrio, M., Willingham, M.C., & Hanover, J.A. (1987). A monoclonal antibody against a family of nuclear pore proteins (nucleoporins): 0-linked N-acetylglucosamine is part of the immunodeterminant. Proc. Natl. Acad. Sci. USA 84, 6462-6466. Reichelt, R., Holzenburg, E.L., Buhle, E.L., Jamik, M., Engel, A., & Aebi, U. (1990). Correlation between structure and mass distribution of nuclear pore complex components. J. Cell Biol. 110, 883-894.

56

PAOLA GRANDI and EDUARD C. HURT

Ris, H. (1989). Three-dimensional imaging of cell ultrastructure with high resolution low voltage SEM. Inst. Phys. Conf. Sen 98, 657-662. Rout, M.P. & Blobel, G. (1991). A biochemical enrichment procedure for components of the yeast nuclear pore complex. J. Cell Biol. 115,458a. Schlenstedt, G., Hurt, E.G., Doye, V., & Silver, R (1993). Reconstitution of nuclear protein transport with semi-intact yeast cells. J. Cell Biol. 123, 785-798. Shi, Y. & Thomas, J.O. (1992). The transport of proteins into the nucleus requires the 70-kilodalton heat shock protein or its cytosolic cognate. Mol. Cell. Biol. 12, 2186-2192. Silver, RA. (1991). How proteins enter the nucleus. Cell 64,489-497. Silver, P., Sadler, I., & Osborne, M.A. (1989). Yeast proteins that recognize nuclear localization sequences. J. Cell Biol. 109, 983-989. Snow, CM., Senior, A., & Gerace, L. (1987). Monoclonal antibodies identify a group of nuclear pore complex glycoproteins. J. Cell Biol. 104, 1143-1156. Starr, CM., D'Onofrio, M., Park, M.K., & Hanover, J.A. (1990). Primary sequence and heterologous expression of nuclear pore glycoprotein p62. J. Cell Biol. 110, 1861-1871. Steinert, P.M. & Roop, D.R. (1988). Molecular and cellular biology of intermediate filaments. Annu. Rev. Biochem. 57, 593-625. Stewart, M. (1992). Nuclear pore structure and function. Cell Biol. 3,267-277. Stochaj, U. & Silver, P.A. (1992). A conserved phosphoprotein that specifically binds nuclear localization sequences is involved in nuclear import. J. Cell Biol. 117, 473-482. Sukegawa, J. & Blobel, G. (1993). A nuclear pore complex protein that contains zinc finger motifs, binds DNA, and faces the nucleoplasm. Cell 72, 29-38. Unwin, P.N. & Milhgan, R.A. (1982). A large particle associated with the perimeter of the nuclear pore complex. J. Cell Biol. 93, 63-75. Wente, S.R., Rout, M.P., & Blobel, G. (1992). Anew family of yeast nuclear pore complex proteins. J. Cell Biol. 119, 705-723. Wimmer, C, Doye, V., Grandi, R, Nehrbass, U., & Hurt, E. (1992). A new subclass of nucleoporins that functionally interacts with nuclear pore protein NSPl. EMBO J. 11,5051-5061. Wozniak, R.W. & Blobel, G. (1992). The single transmembrane segment of gp210 is sufficient for sorting to the pore membrane domain of the nuclear envelope. J. Cell Biol. 119,1441-1449.

ATP BINDING CASSETTE TRANSPORTERS IN YEAST: FROM MATING TO MULTIDRUG RESISTANCE

Ralf Egner, Yannick Mahe, Rudy Pandjaitan, Veronika Huter, Andrea Lamprecht, and Karl Kuchler

I. Introduction 58 11. The Family of Yeast ABC Proteins 59 III. Export of a-Factor Pheromone Requires a Dedicated ABC Transporter 65 A. The Lipid Modified C-terminus of a-Factor May Represent a Secretion Signal 65 B. Biosynthesis and Turnover of the Ste6 a-Factor Transporter . 74 IV. Multidrug and Heavy Metal Resistance Mediated by ABC Transporters 77 A. Biosynthesis and Turnover of Yeast Multidrug Transporters . 83 V. Conclusion and Perspectives 87 Acknowledgments 88 References 89 Membrane Protein IVansport Volume 2, pages 57-96 Copyright © 1995 by JAI Press Inc. All rights of reproduction in any form reserved. ISBN: 1-55938-983-4

57

58

RALFEGNERetal.

I. INTRODUCTION ABC (ATP binding cassette) transporters comprise a novel superfamily of membrane transport proteins that are found operating from bacteria to humans (reviewed in Higgins, 1992). More than 150 ABC-sequences are now deposited in the NCBI databank and their number keeps growing rapidly. The hallmark characteristics of most ABC proteins include the presence of two highly conserved domains for ATP-binding (ABC) and two membrane domains each containing usually six membrane spanning a-helices (TMS). These four domains are normally arranged in a (TMS^ABC)2 or (ABC-TMS6)2 configuration, but "half-size" transporters with an TMS5-ABC or ABC-TMS^ topology as well as other topologies are also frequently found (Higgins, 1992; Kuchler and Thomer, 1992b). Although the structural organization has been highly conserved throughout evolution, most ABC proteins appear to be of rather limited or dedicated substrate specificity. As a consequence, ABC transporters are implicated in a remarkable variety of transport processes, including the transmembrane transport of ions, heavy metals, carbohydrates, anticancer drugs, amino acids, oligopeptides, steroids, glucocorticoids, mycotoxins, antibiotics, pigments, and proteins (Higgins, 1992; Kuchler and Thomer, 1992b; Kuchler, 1993). However, the mechanism by which transport of such a substrate and size diversity can be achieved, while each ABC transporter maintains selectivity for its particular substrate, represents an intriguing and yet unsolved problem. One of the most prominent ABC proteins known to exist in animal cells is the cystic fibrosis transmembrane conductance regulator, CFTR (Riordan et al., 1989), which is associated with cystic fibrosis, a fatal genetic disease occurring most frequently among Caucasians (Harris and Argent, 1993). Furthermore, expression of the mammalian P-glycoprotein or Mdrl (for multiple drug resistance) is associated with a multidrug resistance (MDR) phenotype in cultured cells and in many tumors (Borst et al., 1993; Gottesman and Pastan, 1993; Kane, 1995). The hallmark of MDR, which poses a major impediment to successful cancer chemotherapy, is resistance to a single agent to which tumors or cultured cells are exposed, and the concomitant development of cross-resistance to a variety of structurally and functionally unrelated drugs (Gottesman and Pastan, 1993). Additional ABC proteins of medical importance include P-glycoproteins of the malarial parasite Plasmodium (Cowman and Karcz, 1993) and from Leishmania (Callahan and Beverley, 1991) implicated in chloroquine and heavy metal resistance, respectively. Recent

ABC Transporters in Yeast

59

findings suggest that peroxisomal diseases such as adrenoleukodystrophy and Zellweger syndrome may be linked to defective transporters of the ABC protein family (Gartner et al, 1992; Gartner and Valle, 1993; Mosser et al., 1993). Finally, antigen presentation of viral peptide antigens requires the action of a heterodimeric Tapl/Tap2 ABC transporter of the ER membrane that imports antigenic peptides from the cytoplasm into the ER lumen before their cell surface presentation in association with the MHC class I proteins (NeeQes and Momburg, 1993; Townsend and Trowsdale, 1993; Hill and Ploegh, 1995). The lower eukaryote yeast has quickly outpaced all other organisms in terms of number of different ABC proteins existing in one species. While only a few years ago the Ste6 a-factor transporter (Kuchler et al., 1989; McGrath and Varshavsky, 1989) was the only known yeast ABC protein, there are at least 23 different yeast ABC proteins known today and their number appears to be ever-growing. Thus, the scope of this review is to summarize these most recent developments and discuss the functions of yeast ABC transporters. In particular, we will recapitulate the mechanism of a-factor pheromone export, which is mediated by the Ste6 ABC transporter and proceeds via a novel mechanism distinct from the classical secretory pathway (Kuchler et al, 1989; Kuchler, 1993; Kuchler et al, 1994). Furthermore, we will discuss a newly emerging subfamily of yeast ABC transporters, all members of which appear to be linked to high copy mediated multidrug and/or heavy metal resistance phenomena. 11. THE FAMILY OF YEAST ABC PROTEINS The discovery that the Ste6 a-factor pheromone transporter represents a close homologue of the mammalian multidrug resistance P-glycoproteins (Kuchler et al., 1989; McGrath and Varshavsky, 1989) motivated many laboratories to search for other yeast genes that are functionally and/or structurally related to STE6. Indeed, experimental strategies such as PCR cloning, low stringency hybridization screening, as well as various genetic screens have led to the identification of several new yeast ABC protein genes. In addition, the joint efforts of many laboratories to systematically sequence the yeast genome has led to the identification of additional novel yeast ABC proteins. In Table 1, all currently known yeast ABC proteins are listed. Based on their functions and, where applicable, on the phenotype of the gene disruption, we suggest to group all yeast ABC protein genes into four classes. First, class I contains ABC

Table I . The Family of Yeast ATP Binding Cassette Proteins Classification Class I

ABC Protein Ste6

Class I1 Stsl/Pdr5/ YdrlILem 1

Snq2 0 0

Yen Ycfl Pmd 1 (S. pombe) Hmt 1 (S. pombe) Hba2 (S. pombe) Cdr 1 (Candida) Class I11 Atml Ssh2/Pall/Pxal

YKL741

Substrate

Length (aa)

Topology

Localiu~tion

Mutant Phenotype

Chromosome

Homologue

References

a-factor

1290 (TMS6-ABC),

PM,GV,ES? Sterile

XI

MdrlMlyB

Kuchler et al. (1989) McGrath & Varshavsky (1989)

Drugs, steroids

1511 (ABC-TMS,),

PM

XV

whitelbrown/ scarlet

Drugs, steroids Drugs Cadmium

1501 (ABC-TMS,),

PM

IV

1044 (TMS,-ABC,)? 1515 (TMS6-R-ABC),

N

whitelbrownl scarlet Mdrl (?) Cftr

Balzi et al. (1994). Bissinger & Kuchler ( 1994), Hirata et al. (1994), Kralli et al., (1995) Servos et al. (1993)

Drugs

1362 (TMS6-ABC),

?

Ste6/Mdrl

Nishi et al. (1992)

III

MdrllSte6

Ortiz et al. (1992)

?

NCBI-609264a

XVI

whitelbrownl scarlet whitelbrownl scarlet Mdrl Pmp7OIALDp

XI

Pmp7OIALDp

Viable but drug HS

? ?

694 TMS6-ABC 758 TMS6-ABC

Viable but drug HS Ribo? Cyto? Essential Vacuole Viable but Cd HS ? Viable but drug HS Vacuole Viable but Metal HS ? Viable but drug HS ? Viable but drug HS Mitochondria Slow growth Peroxisome Viable but oleate

?

853 TMS,-ABC

Peroxisome? ?

Heavy metals Brefeldin A

830 TMS6-ABC 1530 (ABC-TMS6)*

Drugs

1501 (ABC-TMS,),

?

? ?

Sandbaken et al. (1990) Szczypka et al. (1994)

Prasad et al. (1995) Leighton & Schatz (1995)~ Kuchler et al. (1992), Shani et al. (1995). Swartzman et al. (1995) Bossier et al. (1994)

Class IV SshlhIdl2 Mdl l Adp l

?

? ?

Viable Viable

XI1

III

MdrlRapllTap2 Kuchler et al. (1992), Dean et al. (1994) Mdrl/Tapl/Tap2 Dean et al. (1994) whitelbrown/ Pumelle et al. (1991) scarlet Snq2 NCBI-I42685 1 MRP ? NCBI-549649 unknown ORF NCBI-603269

C.el.

Snq2 Sts l StsllSnq2 Notes:

-

m A

NCBI-604162' d.e c

' ~ u r iand Rose, personal communication 'several novel ABC sequences identified in reference 67 were not included. '~arrell,personal communication. d ~ a r l and e Wolfe, personal communication. e . B~ssingerand Kuchler, unpublished results. In cases where ABC sequences were deposited in the database, the NCBI sequence identification numbers to allow for sequence remeval PM, plasma membrane; Ribo, Ribosome, Cyto, Cytoplasm, GV,Golgi vesicle, ES, endosome; TMS, transmembrane segment; ABC, ATP-binding cassette; HS, hypersensitivity; R, regulatory domain.

62

RALFEGNERetal.

proteins whose physiological function and natural substrates are known. The only ABC transporter in this class is the Ste6 a-factor transporter whose function is extensively discussed in Section III. Second, overexpression of all class II transporters (see Section IV) is associated with multidrug resistance (MDR), drug hypersensitivity (DHS) or heavy metal resistance (HMR) phenomena. Third, the group of class III transporters comprises ABC protein genes whose chromosomal deletions do result in detectable phenotypes although the in vivo functions of the respective proteins are unknown. Finally, class IV comprises ABC protein genes of unknown function that do not exhibit an obvious phenotype upon gene disruption. The majority of class III and class IV transporters were either identified fortuitously during systematic sequencing projects or cloned by PCR using degenerate oligonucleotides designed according to the evolutionary conserved ABC domains. For instance, we have used PCR to identify and clone two novel yeast ABC protein genes, SSHl and SSH2 (for sterile six homologues), both of which are closely related to mammalian P-glycoproteins (Kuchler et al., 1992). SSHl is identical to MDL2 (for Mdrlike), which was independently isolated by another group using a similar approach (Dean et al., 1994). In addition, the same group has cloned another ABC transporter, Mdll, that is closely related to but different to both Sshl/Mdl2 and Ssh2 (Dean et al., 1994). Cells carrying chromosomal deletions of either the SSH1/MDL2, MDLl or SSHl gene are viable under normal growth conditions and exhibit no obvious phenotypes. The functions of Sshl/Mdl2 and Mdll and their cellular locaUzation are unknown at present. All three ABC proteins are typical half-size transporters with a predicted TMS^-ABC topology. Sshl/Mdl2 is most closely related to mammalian P-glycoproteins such as Mdrl (Kuchler et al., 1992), yet its overexpression is not associated with a MDR phenotype. A possible explanation could be that appropriate drug substrates for Sshl/Mdl2 and Mdll have not been identified. Alternatively, although this explanation seems unlikely because they are half-size transporters, Sshl/Mdl2 or Mdll could represent the yeast homologues of the mammalian Mdr2 P-glycoproteins, whose expression is also not associated with MDR phenomena despite an apparent 85% identity of Mdrl and Mdr2 (Borst et al., 1993; Gottesman and Pastan, 1993). The closest homologue of Ssh2 in animal cells are the peroxisomal Pmp70 (Kamijo et al., 1990; Gartner et al, 1992; Gartner and Valle, 1993) and ALDp (Mosser et al., 1993) proteins, both of which have been implicated in peroxisomal genetic diseases such as Zellweger syndrome

ABC Transporters in Yeast

63

and X-linked adrenoleukodystrophy, respectively. Both diseases are associated with apparent defects in peroxisome function and/or organelle assembly and with the demyelination of the nervous system that is paralleled by the accumulation of very-long-chain fatty acids in peroxisomes (Aubourg et al., 1993; Gartner and Valle, 1993). Interestingly enough, yeast Assh2 cells, despite being viable under normal growth conditions, are unable to grow on oleate as the sole carbon source (Swartzman et al., 1995). Growth of yeast cells on oleate requires functional peroxisomes, and the Ssh2 transcript is only detectable in oleate-grown cells, suggesting Ssh2 to be a constituent of the yeast peroxisomal membrane. Indeed, localization experiments have shown that Ssh2 is a peroxisomal membrane protein and therefore SSH2 has been renamed PALI (Swartzman et al., 1995). PALI is identical to PXAl; which was cloned independently by another group (Shani et al., 1995). PALI/PXAl and another closely related gene, YKL741, which was identified during the systematic yeast genome sequencing project (Bossier et al., 1994), could form heterodimers to comprise a functional transporter in the peroxisomal membrane. Thus, as proposed for mammaUan Pmp70 and/or ALDp, yeast Ssh2/Pall/Pxal and YKL741 could play a role in peroxisome biogenesis and/or organelle proliferation (Shani et al., 1995; Swartzman et al., 1995). It will be very important to determine if mammalian Pmp70 and ALDp, when functionally expressed in yeast, are able to restore growth of Assh2 cells on oleate as the sole carbon source. ABC proteins have also been speculated to be involved in the biogenesis and assembly of organelles such as mitochondria and chloroplasts (Kuchler and Thomer, 1990,1992b). Indeed, chloroplasts of Marc/ianr/a polymorpha do contain an ABC protein, MbpX, but its function there is not known at the moment (Umesono et al., 1988). The laboratory of Gottfried Schatz has recently set out to search for yeast ABC protein genes required for mitochondrial biogenesis and eventually for protein import into mitochondria (Hannavy et al., 1993). PCR cloning allowed for the isolation of at least ten different PCR-fragments some of which were corresponding to novel ABC protein sequences (Leighton and Schatz, 1995), others turned out to be identical to already cloned ABC genes such as Adpl (Pumelleetal., 1991) and Sshl (Kuchler etal., 1992). Out of three novel ABC protein genes that gave a phenotype upon gene disruption, at least one, ATMl (for ABC transporter of mitochondria) was indeed found to encode a mitochondrial protein (Leighton and Schatz, 1995). Deletion of the ArM7 gene leads to a slow-growth phenotype on rich medium and cessation of growth on minimal media. Atml is a

64

RALFEGNERetal.

694-residue TMS^-ABC half-size ABC transporter that is, based on cell fractionation and immunofluorescence experiments, localized in the inner mitochondrial membrane with the ABC domain facing the mitochondrial matrix (Leighton and Schatz, 1995). Protein import of most, but not all, mitochondrial polypeptides is thought to occur at contact sites between inner and outer membrane requiring ATP-hydrolysis in the mitochondrial matrix (Pfanner et al., 1992; Cyr et al, 1993; Hannavy et al., 1993; Price and Vemer, 1993). The inner membrane localization of Atml could still be consistent with a function for Atml in mitochondrial protein import, since one can speculate that Atml could be a structural and energy-providing component of the import site (Hannavy et al., 1993). The architecture of such a pore is likely to form a hydrophilic proteinaceous "channel" across both mitochondrial membranes to allow for protein import to occur. Notably, puromycin, a known substrate for mammalian P-glycoproteins (Gottesman and Pastan, 1993), is a potent inhibitor of protein import as it inhibits ATP-hydrolysis in the matrix (Price and Vemer, 1993). Similar to mitochondrial protein import, it was proposed that the bacterial HlyB/HlyD transporter for hemolysin forms a protein pore stretching both bacterial membranes at the membrane junctions through which the toxin is released to the medium (Koronakis and Hughes, 1993), although in hemolysin export the direction of protein transport is the opposite of mitochondrial import. An alternative possibility would be that Atml could also be responsible for export of substrates from the mitochondrial matrix rather than import (Leighton and Schatz, 1995). If Atml plays a role in uptake of substances into yeast mitochondria, Atml would be the first example for an eukaryotic ABC transporter whose transport direction would be similar to the one found in certain bacterial ABC proteins functioning in nutrient uptake (Higgins, 1992). While the function of Atml is still unknown at present, in vitro studies will certainly show if Aatml mitochondria are characterized by a defective protein import machinery or if the slow-growth phenotype of Aatml cells is the result of another yet unidentified transport function of Atml. Finally, very little is known about the majority of the class IV ABC proteins listed in Table 1, as many of them were identified during the course of the yeast genome sequencing project or by low stringency hybridization. In many cases neither a subcellular localization nor a phenotype associated with the corresponding gene disruption, if any, is known at the moment. The detailed molecular analysis of these genes is impatiently anticipated.

ABC Transporters in Yeast

65

III. EXPORT OF a-FACTOR PHEROMONE REQUIRES A DEDICATED ABC TRANSPORTER Despite its deceptively simple life style as a unicellular eukaryotic organism, Saccharomyces cerevisiae has three distinct cell types. Mating or conjugation of two haploid cell types, MAJa and MATa, results in the formation of a diploid a/a cell (Sprague and Thomer, 1992). This sexual reproduction cycle requires the action of extracellular peptide hormones known as mating pheromones a-factor and a-factor, that are released from haploid MATa and MATa cells, respectively (Sprague and Thomer, 1992). The pheromone produced by MATa cells, a-factor, is a 13-residue peptide that is proteolytically liberated from a larger glycosylated precursor polypeptide and secreted via the classical ER-Golgi secretory pathway (Sprague and Thomer, 1992). Mating pheromone a-factor is a post-translationally modified 12-residue lipopeptide whose C-terminal cysteine residue is both famesylated and carboxymethylated (Anderegg et al., 1988). The signal triggering these modifications was identified as the so-called CAAX box (where C is cysteine, A is an aliphatic residue and X may be any amino acid) found at the extreme C-terminus of all prenylated proteins (Glomset et al., 1990), including both yeast and mammalian Ras proteins (Schafer and Rine, 1992) and the unmodified a-factor precursors (Brake et al., 1985). Biogenesis of biologically active a-factor requires several gene products including proteases, a famesyl transferase activity, a methyltransferase, a dedicated pheromone transporter and, of course, the two structural genes, MFa7 and MFa2 encoding the two known pheromone precursors (Brake et al., 1985). A hypothetical working model for a-factor secretion by haploid MAJa cells is depicted in Figure 1. The a-factor export machinery is likely to reside in the plasma membrane, although so far only Ste6 could be functionally locaHzed to the plasma membrane (Kuchler et al., 1993). However, most of the steady state levels of Ste6 appear to be present in intracellular Golgi-like compartments as determined by immunofluorescence experiments (Kuchler et al., 1993; KoUing and Hollenberg, 1994) and biochemical fractionation studies (KoUing and Hollenberg, 1994). Hydrophobicity analysis of the STEM gene (Ashby et al., 1993) encoding a carboxymethyltransferase (StemeMarr et al., 1990; Hrycyna et al., 1991; Ashby et al., 1993; Sapperstein et al., 1994), suggested that Stel4 is membrane associated, but a plasma membrane localization has not yet been demonstrated. Notably, the proteolytic processing activities required for a-factor maturation have

Extracellular

a-factor

farnesyl \ 0.CH3

Plasma Membrane

Cytoplasm

pro-a-factor Figure 1. A hypothetical working model for yeast mating pheromone a-factor export. The a-factor secretion apparatus may assemble in the plasma membrane as a transport-competent translocation complex. Proa-factor is synthesized in the cytoplasm as an inactive precursor that is farnesylated at the C-terminal CAAX box by a heterodimeric farneysltransferase (Rami, Ram2). Farnesylation targets the precursor to the plasma membrane where it undergoes proteolytic removal of the three C-terminal residues of the CAAX box. The gene encoding the protease responsible for CAAX box processing is yet unidentified (?). Modification of the CAAX box is completed after carboxymethylation of the C-terminal cysteine by a methyltransferase (Stel4) which is putatively localized in the plasma membrane. It has not been established when exactly the proteolytic cleavage at the a-factor N-terminus occurs. However, it may precede the actual translocation step, which is mediated by the Ste6 ABC transporter in the plasma membrane.

66

ABC Transporters in Yeast

67

been characterized only at the biochemical level (Ashby et al., 1992; Hrycyna and Clarke, 1992) but the corresponding genes have not yet been identified. Initially, a-factor is synthesized on cytoplasmic polysomes as a unglycosylated precursor that lacks a hydrophobic signal peptide but has a hydrophilic N-terminal extension (Brake et al., 1985; Kuchler et al., 1989). Immediately after its biosynthesis, pro-a-factor is lipid-modified at the CAAX box by the heterodimeric Raml/Ram2 famesyltransferase (Goodman et al., 1990; He et al., 1991). CAAX box famesylation most likely targets the pheromone to the plasma membrane for further maturation. The three C-terminal amino acids of the CAAX box are then clipped by proteolysis, and the pheromone is carboxymethylated by Stel4 (Steme-Marr et al., 1990; Hrycyna et al., 1991). This order of events in a-factor maturation was also confirmed in vitro by using a-factor maturation assays (Marcus et al., 1990). At present, it is not clear why only fully processed a-factor can be efficiently secreted from cells (Kuchler et al., 1989; Berkower and Michaelis, 1991), but one may assume that a-factor release is coupled to its intracellular processing, because very little, if any, mature a-factor can be detected intracellularly (Kuchler et al., 1989). Genetic and biochemical evidence suggested that bioactive a-factor is translocated from the cytoplasm across the plasma membrane (Kuchler et al., 1993) to the extracellular space in an ratelimiting export step mediated by Ste6 (Kuchler et al., 1989). Export of a-factor bypasses the classical ER-Golgi secretory pathway, for extracellular pheromone is still being produced when MATsi cells carrying temperature-sensitive secretion-defective (sec) mutations are shifted to the restrictive temperature (Kuchler et al., 1989; McGrath and Varshavsky, 1989). Thus, a-factor secretion does not require any functional vesicular secretory compartment (Pryer et al., 1992) and therefore represents a novel route for protein secretion in eukaryotic cells (Kuchler, 1993; Kuchler etal., 1994). A. The Lipid Modified C-terminus of a-Factor May Represent a Secretion Signal The secretion signal that drives a-factor export is poorly understood. It has been speculated that the hydrophilic N-terminal extension of pro-a-factor promotes pheromone release to the medium, but the existence of many mutations within the N-terminal leader sequence with no obvious phenotype seems to contradict this idea (Michaelis, 1993). In

68

RALFEGNERetal. farnesyl

L

a-factor

pro-a-factor

N-H^COCHs

N-l

21

WJCVlA-C

7X

...SSEKKDNYIIKGVF...

JJ

IL.la::pro-a.factor

N-|s:>:^i7bNS>^^^^|"liTjj^Bjc ^^ ^ ^ J C V I A - C ...SSEKKDNYnKGVF..

IL-la::a-facloi

N - ( ^ ^ 1 7 0 ^ ^ ^ ^ N^ s » <j r - v i A .r

...LIAPSLlKGVF...

N^M^^C Figure 2. Structure of murine interleukin-1 a-yeast a-factor chimeras expressed in yeast. The amino acid residues around the junctions of the heterologous fusion proteins are indicated in single letter code. Putative proteolytic processing sites in pro-a-factor are marked by arrows. The numbers in boxes indicate the length of individual domains provided by each protein. Dashed box, IL-1a; white box, a-factor precursor sequence; black box, mature a-factor.

order to test if the modified C-terminal CAAX box of a-factor comprises a secretion signal, we have fused the a-factor coding sequence including its CAAX box to the C-terminus of heterologous proteins such as murine interleukin-la (IL-la). Mature murine IL-la and IL-1 p, two cytokines involved in a variety of immunoregulatory responses (Dinarello, 1992),

ABC Transporters in Yeast

69

when expressed in yeast normally remain cytosolic, unless a hydrophobic signal peptide is attached to the N-terminus (Baldari et al., 1987). By contrast, we have constructed chimeric fusion proteins comprised of mature murine IL-la and both a-factor and pro-a-factor attached to the C-terminus of IL-la (Figure 2). We have chosen IL-la for the chimeric constructs because like yeast a-factor, both IL-la and IL-lp lack a hydrophobic signal peptide, and both cytokines are secreted from mammalian cells bypassing the classical secretory pathway (Rubartelli et al., 1990). The chimeric "pheromones" were expressed from a YEp352based multicopy plasmid under the control of the STE6 promotor in a yeast strain which lacks both structural genes for a-factor (Kuchler, in preparation). Thus, a MATa Amfiil Afma2 cell is absolutely sterile, because it does not produce any intracellular a-factor pheromone that is essential for mating (Michaelis and Herskowitz, 1988). Immunoblotting of soluble and membrane fractions demonstrated that, unlike mature IL-la, the chimeric pheromones efficiently and quantitatively associate with a membrane fraction, presumably because their terminal CAAX box is famesylated in vivo and therefore mediates membrane insertion (Figure 3). Surprisingly though, both semiquantitative and quantitative mating assays (Kuchler et al., 1993) demonstrated that the chimeric pheromones have retained at least some a-factor bioactivity, since they are able to mediate conjugation, albeit at low levels, with tester cells of increased MFal

1T

S

in

IL-la::a-f

IL-lo::pro-a-f

M 1T S

1T S

-

^

M

'

IL-la

1 T

S

M 1

1

Figure 3. Murine interleukin-1 a-yeast a-factor chimeras associate with a membrane fraction. The chimeric pheromones schematically depicted in Figure 2 were expressed in yeast strain CT300, which is a isogenic m^2;;ura3(FOA) derivative of strain SMI 229 (MATa ura3 Ieu2 his4 trp1 can! AT7/a/;;/.E(J2/T7&2::L//?/43) kindlyprovided by S. Michaelis (Michaelis and Herskowitz, 1988). The preparation of cell free extracts (T), membrane fractions (M) and soluble high speed supernatant fractions (S) for SDS-PAGE analysis and immunoblotting was carried out as described elsewhere (Kuchler et al., 1993). Immunoblots were probed with polyclonal anti-IL1a antibodies purchased from Genzyme.

RALFEGNERetal.

70 MAT a sst2 STE6-^

Aste6

© © ® © © ® © © ©©

Figure 4. Murine interleukin-1a-yeast a-factor chimeras have mating pheromone bioactivity. An identical number of isogenic wild-type (STE6'^) and mutant {Aste6) cells expressing chimeric pheromones in strain CT300 were grown on agar plates together with appropriate control strains. A semi-quantitative patch mating assay (Kuchler et al., 1993) was used to score the strains for their abil ity to conjugate with the supersensitive RC757 mating tester strain (MATa sst2-1) suitable to detect low levels of extracellular mating pheromones (Kuchler et al., 1993). Resulting diploid colonies were selected by growth on minimal medium after replica-plating of mating mixtures. 1, MFal; 2, I L - l a ; 3, IL-la::pro-a-factor; 4, IL-1a::a-factor, K l , MFa1 in plasmid Yep352, K2, STE6 + MFal in plasmid Yep352.

pheromone sensitivity (Figure 4). Because the extremely sensitive mating assay allows for detection of very low levels of extracellular pheromone, the chimeric pheromones must be at the extracellular space. Cell lysis cannot explain the appearance of extracellular a-factor chimeras, because it is solely dependent on a functional Ste6 transporter, and does not occur in the isogenic Aste6 mutant strain (Figure 4). Another possible explanation for the appearance of pheromone in the medium could be unspecific proteolysis of the chimeras that would generate aberrant intracellular a-factor forms recognized and transported by Ste6. This explanation seems unlikely for several reasons. First, immunoblotting using both anti-a-factor and anti-IL-la antibodies demonstrated that all chimeric pheromone constructs are stable and do not give rise to unspecific degradation products because their mobility in

ABC Transporters in Yeast

71

i\

t\ anti-IL-la

anti-a-factor

ISkDa —

18 kDa

16kDa--.

-

-

' 1

t

1

»

Figure 5. Murine interleukin-1a-yeast a-factor chimeras are stably expressed. Total cell-free extracts were prepared from yeast strain CT300 expressing various constructs from yVlFa/, IL-1a, IL-1a::a-factor, and IL1a::pro-a-factor harboring YEp352-based expression plasmids. Equivalent amounts of extracts were subjected to SDS-PAGE and immunoblotted (Kuchler et al., 1993) using both polyclonal anti-IL-la and polyclonal anti-a-factor antibodies (Kuchler et al., 1989).

SDS-acrylamide gels is exactly as expected from their primary sequence (Figure 5). Second, proteolysis can only be considered in the IL-la::pro-a-factor fusion, but not in the IL-la::a-factor chimera, as the latter lacks all protease cleavage sequences required for pro-a-factor precursor processing (Figure 2). Thus, one would expect that cells expressing the ILla::pro-a-factor chimera mate at almost wild-type levels, since all processing sites of pro-a-factor are also present in the chimeric protein (Figure 2). However, mating of IL-la::pro-a-factor strains with wildtype tester strains is about 1000-fold lower than the control, presumably because proteolytic processing in the IL-la::pro-a-factor molecule is largely debiUtated, most likely due to a sterical inhibition of the processing protease(s). By contrast, the amino acid sequence around the junction in the IL-la::a-factor fusion protein has no homology whatsoever to processing sites in the a-factor precursor sequence and, thus, is unlikely to be cleaved by any of the processing proteases (Figure 2). These results demonstrate that the size requirements of Ste6 for pheromone export may be more relaxed than previously anticipated, since Ste6 is apparently able

72

RALFEGNERetal.

to transport even an enlarged "a-factor pheromone" such as the chimeric IL-la: :a-factor protein that is about 17 kDa in size as compared to 3 kDa of mature a-factor (Figure 2). Although we cannot firmly rule out the unlikely possibility that the amino acid sequence of mature a-factor itself represents a plasma membrane targeting signal, our results suggest that a fully modified C-terminal CAAX box in a-factor may represent a pheromone secretion signal, as it can target an otherwise cytosolic, heterologous fusion protein to the extracellular space. Indeed, a C-terminal CAAX box was demonstrated to be able to target bacterial protein A to the plasma membrane in mammalian MDCK cells (Hancock et al., 1991). Moreover, famesylation, but also carboxymethylation, were shown to be essential for plasma membrane association of Ras proteins (Hancock et al., 1991). The famesyl moiety of a-factor has also been speculated to mediate protein-protein and possibly protein-lipid interactions (Epand et al., 1993). Thus, the hydrophobic famesyl tail of the pheromone could facilitate initial contact of mating cells (see also later). This is confirmed by results that show that the famesyl moiety of a-factor is essential for mating pheromone bioactivity (Marcus et al., 1991). Interestingly, an in vivo mating restoration assay demonstrated that synthetic a-factor (Xue et al., 1989), when exogenously added to conjugating yeast cells, was able to partially restore mating only if added to mixtures containing MATa, Amjul Amjk2 and MATa cells, but not when added to MATa Amfsil Amfal AsteO and MATa cells (Marcus et al., 1991). This finding indicates that Ste6, in addition to its essential role in a-factor pheromone export, could have further functions in mating. For instance, one could speculate that Ste6 actually translocates a-factor to the cell surface and loosely binds the pheromone on the cell surface, thereby presenting a-factor to MATa cells. The existence of point mutations in hydrophilic extracellular loops which abrogate Ste6 function further supports this idea (Huter and Kuchler, unpublished observations). The hypothetical model for the a-factor export assumes that the predicted Ste6 topology, which places both ABC domains on the cytoplasmic face of the plasma membrane, is correct (Kuchler et al., 1989). Such a Ste6 topology was also indirectly suggested by the existence of linker insertion and point mutants in predicted intracellular hydrophilic loops which resulted in a non-functional Ste6 transporter, although we do not know at present if these mutations cause a mislocalization of the Ste6 transporter (Huter and Kuchler, unpublished observations). Those results do not come entirely unexpected, as one can appreciate that

ABC Transporters in Yeast

73

pheromone interaction with Ste6 would presumably take place on the inner side of the plasma membrane, requiring distinct structural features of the cytoplasmic loops in Ste6. We have recently begun to analyze the Ste6 membrane topology in more detail using limited protease digestion experiments of functional Ste6 derivatives that carry specific factor Xa protease cleavage sites in predicted intra- and extracellular hydrophilic loops. So far, our results obtained for at least two TMS in the first half of Ste6 are consistent with the predicted (ABC-TMS6)2 topology for the Ste6 transporter (Huter and Kuchler, unpublished results). The binding of the a-factor pheromone to Ste6 could also take place within the phospholipid bilayer, since famesylated a-factor is extremely hydrophobic and readily associates with phospholipid bilayers even in the absence of Ste6 (Kuchler, unpublished observations). Moreover, the famesyl moiety of a-factor has been shown to promote association with artificial phospholipid bilayers in vitro (Epand et al., 1993), and the famesyl moiety can mediate membrane association of the chimeric IL-la::a-factor fusion proteins (Figure 3). Hence, lateral diffusion of pheromone in the membrane could bring a-factor to the transporter before its extrusion to the medium. A somewhat similar model for drug binding and extrusion, the "molecular vacuum cleaner" was suggested for the P-glycoprotein drug efflux pump (Gottesman and Pastan, 1993; Kane, 1995). In this model, it was proposed that hydrophobic compounds, in contrast to hydrophilic cytotoxic substances, actually never enter the cell. Instead, drugs would partition in the phospholipid bilayer due to their hydrophobicity, and lateral membrane diffusion and interaction of drugs with certain TMS a-helices of Mdrl would precede ATP-dependent drug pumping by Mdrl (Gottesman and Pastan, 1993; Kane, 1995). Strikingly, several mutations in predicted TMS of P-glycoproteins lead to non-functional transporters or to a different substrate specificity (Borst et al., 1993; Gottesman and Pastan, 1993) which strongly supports the vacuum cleaner model for the function of Mdrl. In the case of a-factor, one may speculate that accessory proteins such as plasma membrane bound famesyl receptors could facilitate the interaction of a-factor with its export machinery and/or membrane association of the maturing pheromone. Indeed, there is preliminary evidence from cell fractionation experiments that most of the membrane associated famesylated pheromone chimeras (Figure 3) are in fact found in the plasma membrane fraction, implying the existence of potential famesyl receptors in yeast (Kuchler, unpubUshed observations). Similarly, heterologous protein A: :CAAX box fusion proteins are exclusively targeted

74

RALFEGNERetal.

to the plasma membrane in mammalian MDCK cells, but not to other cellular membranes, which prompted the authors to propose the existence of accessory plasma membrane receptors for prenylated proteins (Hancock et al., 1991b). However, so far, high affinity binding sites for prenylated proteins have only been identified in rat liver microsomal membranes (Thissen and Casey, 1993), but not in the plasma membrane. B. Biosynthesis and Turnover of the Ste6 a-Factor Transporter

The expression of the Ste6 transporter appears to be under hormonal control, as a-factor arrested cells express almost 10-times higher levels of Ste6 (Kuchler et al., 1993). In vegetatively growing cells, Ste6 is present predominantly in intracellular vesicles that subtend the plasma membrane (Kuchler et al., 1993; Kolling and Hollenberg, 1994). Exposure of MAra cells to a-factor, however, induces Ste6 expression and concomitantly leads to an apparent redistribution of Ste6 to the cell surface (Kuchler et al., 1993). Simultaneously, Ste6 is localized in a highly polarized fashion in the tip of the mating projection (Kuchler et al., 1993). The mating projection (also known as "schmoo") is always directed toward the mating partner and it is actually induced by pheromones (Madden et al., 1992). Schmoo formation, which indicates that the mating signal transduction cascade is functional and that mating partners are arrested in the G^ phase of their cell cycles (Sprague and Thomer, 1992), induces a number of biochemical changes and precedes actual fusion of haploid cells that occurs exclusively at the schmoo tips (Madden et al., 1992; Sprague and Thomer, 1992). The polarized Ste6 localization results in a pronounced anisotropy of a-factor secretion around the tip of the projection (Kuchler et al., 1993). In addition, the hydrophobicity of a-factor results in a steep pheromone concentration gradient reaching from the cell surface into the medium. Such a delocalized pheromone signal, and the highest possible levels of extracellular pheromone due to increased Ste6 expression, is believed to be necessary for courtship during mating partner discrimination (Jackson and Hartwell, 1990). The Ste6 transporter was recently shown to be an extremely shortlived protein with an apparent half-life of 10-15 minutes (Berkower et al., 1994; Kolling and Hollenberg, 1994). In addition, the metabolic stability of Ste6 is dramatically increased in a pep4 vacuolar protease mutant strain (Berkower et al., 1994; Kolling and Hollenberg, 1994), suggesting that endocytosis delivers Ste6 from the cell surface to the

ABC Transporters in Yeast

75

vacuole. Indeed, vacuolar degradation of Ste6 was recently found to be dependent on a functional endocytic pathway (Berkower et al., 1994; Kolling and Hollenberg, 1994). Likewise, constitutive and pheromoneinduced endocytosis of the pheromone receptors Ste3 and Ste2 has been shown to be responsible for uptake of pheromone receptors via endosomes and their rapid turnover in the vacuole (Davis et al., 1993; Rohrer et al., 1993). Notably, it has been demonstrated that the rapid turnover of MATa2, a short-lived transcriptional repressor of the yeast mating pathway, is triggered by ubiquitination (Chen et al., 1993), and the rapid degradation of MATa2 requires a functional cytoplasmic proteasome (Richter-Ruoff et al., 1994). Similarly, Ste6 is ubiquitinated in vivo (Kolling and Hollenberg, 1994) and Ste6 quantitatively accumulates in an ubiquitinated form in the plasma membrane of end4 mutant cells (Kolling and Hollenberg, 1994) which are blocked in the first step of endocytosis, the formation of early endosomes (Raths et al., 1993). However, in contrast to MATa2, cellular turnover of Ste6 does not require a functional proteasome, since the Ste6 half-life is not affected in prel'l mutant cells harboring a non-functional proteasome (R. Egner and K. Kuchler, unpublished result, R. Kolhng, personal communication). Instead, we would speculate that ubiquitination of cell surface proteins such as Ste6 and maybe others (see also Section IV) represents a signal for endocytosis (Riezman, 1993) and subsequent vacuolar delivery for terminal degradation (Knop et al., 1993). The rapid turnover of Ste6 also provides a plausible explanation for the polar localization of Ste6 in the plasma membrane of pheromone responsive cells. While cellular Ste6 present before a-factor exposure is degraded with a half-life of 10-15 min (Kolling and Hollenberg, 1994), newly synthesized and pheromone-induced Ste6 molecules are deposited mainly at the site of membrane growth along the projection tip (Kuchler et al., 1993), thereby resulting in a predominant schmoo-tip localization of Ste6, but also in an anisotropy of a-factor secretion (Kuchler et al., 1993). It seems unlikely though that the polar localization of Ste6 in the mating projection is caused by either an increased Ste6 degradation or increased Ste6 endocytosis rate in areas outside the schmoo-tip of pheromone responsive cells (Berkower et al., 1994), as there is preliminary evidence that the tumover of Ste6 does not change in a-factor pheromone-induced cells when compared to normal cells (R. Kolling, personal communication). However, the intracellular vesicular staining of Ste6 appears to decrease in pheromone arrested cells, implying

76

RALFEGNERetal.

enhanced surface delivery of Ste6 upon pheromone induction (Kuchler et al., 1993). We have previously interpreted the intracellular vesicular staining in normal cells to be due to exocytic secretory organelles such as Golgi or secretory vesicles that contain Ste6 en route to the plasma membrane (Kuchler et al., 1993). This idea is also confirmed by a recent report demonstrating a Golgi-localization for Ste6 (Kolling and Hollenberg, 1994). Likewise, Ste2, the a-factor receptor of MAJa cells, shows an identical localization pattern as Ste6, both in normal and in pheromone treated cells (Jackson et al., 1991). However, we would argue that due to the endocytic delivery of Ste6 to the vacuole, one cannot exclude the possibility at the moment that the vesicular spot and dot-like staining of Ste6 in uninduced yeast cells (Kuchler et al., 1993; KoUing and Hollenberg, 1994) could be due to the presence of Ste6 in both endocytic compartments and exocytic secretory organelles. From the energetic point of view, it seems clear that the actual a-factor translocation step across the plasma membrane must require energy. Indeed, Ste6 has been shown to be an ATP binding protein in vitro (Kuchler et al., 1993), and mutations in the ABC domains that impair ATP binding also debilitate Ste6 function (Berkower and Michaelis, 1991), suggesting that ATP hydrolysis powers pheromone export. In addition, both ABC domains are required for Ste6 function, because very elegant in vivo severing experiments of Ste6 demonstrated that only both halves of Ste6 co-expressed in the same cell, but not the N-terminal TMS^-ABC or C-terminal ABC-TMS^ half alone, give rise to a functional pheromone transporter (Berkower and Michaelis, 1991). It is not known, however, if ATP binding and/or ATP hydrolysis also serves a structural function for the architecture of a transport complex in the membrane, as was previously shown for certain bacterial ABC transporters (Koronakis and Hughes, 1993). Taken together, both Ste6 halves and their ABC domains must tightly interact in vivo to form a putative membrane pore through which pheromone extrusion can occur. Similarly, the two half-size transporters for viral peptide antigens, Tapl and Tap2, form heterodimers in the ER membrane to create a functional antigen transporter which translocates 8-12 residue peptides into the ER lumen (Hill and Ploegh, 1995). The physical interaction of Tapl and Tap2 must be very strong, since it can only be disrupted by SDS, because both proteins always co-purify as a complex, and because the heterodimer Tapl/Tap2 complex can be immunoprecipitated using monoclonal antibodies recognizing either Tapl or Tap2 (Meyer et al., 1994). Although it

ABC Transporters in Yeast

77

seems reasonably clear that Ste6 mediates the actual a-factor transport step across the plasma membrane, it cannot be ruled out at the moment that additional proteins perhaps dynamically associate with Ste6 to build up the a-factor translocation complex. To be able to more precisely define the a-factor binding domains in Ste6, and to define the function of other proteins such as Stel4 and Raml/Ram2 required for a-factor biogenesis, more Ste6 transporter and processing mutants will have to be isolated and characterized. These studies will be greatly facilitated by the availabiUty of a reconstituted in vitro a-factor transport system that will help to answer many of the open questions concerning the molecular mechanism of a-factor export mediated by the Ste6 pump. Ste6 is a typical member of the ABC transporter superfamily being most closely related both in predicted membrane topology and in primary sequence to the mammalian P-glycoproteins (Kuchler et al., 1989; McGrath and Varshavsky, 1989). In fact, recent results show that human Mdrl and mouse Mdr3 when functionally expressed in yeast (Kuchler and Thomer, 1992) can even partially compensate for the loss of Ste6 function (Kuchler et al., 1992a; Raymond et al., 1992). Although these findings demonstrate a functional conservation between mammalian Mdrl and yeast Ste6, overexpression of Ste6 in yeast is generally not associated with compound multidrug resistance phenomena. However, yeast has been reported to contain numerous MDR determinants, some of which have been shown to be encoded by ABC transporter genes, while others belong to the major facilitator superfamily of membrane transporters (Balzi and Goffeau, 1994).

IV. MULTIDRUG AND HEAVY METAL RESISTANCE MEDIATED BY ABC TRANSPORTERS It has been demonstrated that overexpression of certain ABC proteins in prokaryotes and eukaryotes is linked to drug and antibiotic resistance phenomena (Higgins, 1992). Yeast is no exception to that, as overexpression of all class II ABC proteins listed in Table 1 is linked to either MDR, DHS or HMR phenomena. Most of the class II ABC genes were identified by virtue of their ability to confer a high-copy mediated MDR and/or HMR phenotype, although in some cases expression of an ABC protein leads to DHS rather than MDR. For instance, overexpression of the putative Yef3 elongation factor leads to hypersensitivity to the aminoglycoside antibiotics such as hygromycin and paromomycin (Sandbaken et al., 1990). The 1044-residue essential Yef3 protein represents an

78

RALFEGNERetal.

interesting exception to the general ABC topology because Yef3 appears to harbor two consecutively fused ABC domains. It is not clear whether and how many, if any, a-helical TMS Yef3 may have (Sandbaken et al, 1990). The interaction of Yef3 with hygromycin and paromomycin, both of which increase translational errors, implies that Yef3 may be locaUzed on ribosomes or in the ER membrane, but this has not been established. The function of Yef3 may be to provide mechanical energy as generated during ATP hydrolysis to drive mRNA translocation during protein translation at ribosomes (Sandbaken et al., 1990). MDR and HMR phenomena have also been described in the fission yeast Schizosaccharomycespombe. For instance, Pmdl, which is closely related to Ste6 and mammalian Mdrl, is able to confer resistance to leptomycin B and to other unrelated cytotoxic compounds (Nishi et al., 1992). However, despite being closely related to yeast Ste6, Pmdl does not appear to be required for sexual reproduction in fission yeast, since pmdl mutant cells do not show any defects in conjugation (Nishi et al., 1993). The Hmtl half-size transporter is yet another recently discovered fission yeast ABC protein (Ortiz et al., 1992). Hmtl, a 830-residue ABC protein of the vacuolar membrane, appears to be most closely related to P-glycoproteins from parasites like Leishmania (Callahan and Beverley, 1991), since Hmtl overexpression mediates heavy metal tolerance in fission yeast (Ortiz et al., 1992). Finally, Hba2 from S. pombe was recently identified as a brefeldin A resistance determinant in an attempt to isolate new genes which affect the secretory pathway by interfering with the inhibitory action of this drug (Turi and Rose, personal communication). An exciting new addition to the family of yeast ABC proteins is the 1515-residue YCFl gene product (for Yeast Cadmium Factor), a homologue of the human cystic fibrosis transmembrane conductance regulator (Szczypka et al., 1994). As in CFTR, a highly charged R-domain, which is believed to be required for the regulation of the chloride channel activity in CFTR (Harris and Argent, 1993), is also found in Ycfl (Szczypka et al., 1994). Ycfl was cloned in a multi-copy based genetic screen in which yeast cells were selected based on their ability to grow in the presence of increased cadmium concentrations (Szczypka et al., 1994). Ycfl is a vacuolar membrane protein, but its physiological function is unknown at present. One can envisage, however, that cadmium detoxification presumably involves vacuolar sequestration of intracellular metal ions in the vacuole. Thus, Ycfl, like fission yeast Hmt 1, could be responsible for sequestering intracellular cadmium to the

ABC Transporters in Yeast

79

vacuole thereby leading to cadmium tolerance. Indeed, Ycf 1 expression appears to be regulated by stress response factors such as Yapl, as yapl mutants exhibit cadmium hypersensitivity (Wemmie et al., 1994), confirming a role for Ycf 1 in heavy metal detoxification. In any case, it will be of utmost interest to determine the in vivo role of Ycf 1, since this could potentially lead to insights as to the nature of additional functions and/or in vivo substrates of the human CFTR transporter. There are at least two yeast ABC transporters whose overexpression leads to a MDR phenotype comparable to the one observed for overexpression of P-glycoproteins in higher eukaryotic cells. The recently cloned SNQ2 gene encodes a 1501-residue ABC protein associated with a MDR phenotype conferring cross-resistance to several mutagens (Servos et al., 1993). The SNQ2 gene was originally discovered in an attempt to characterize yeast mutants that are hypersensitive to DNA-damaging agents (Haase et al., 1992). In a somewhat similar approach, we have searched for genes that can mediate resistance to xenobiotic compounds such as sporidesmin, a highly toxic epidithiodioxopiperazine mycotoxin produced in the spores of the fungus Phytomyces chartarum (Jordan and Cordiner, 1987). Its toxicity is believed to be mediated by superoxide and hydroxyl radicals generated during intracellular thiol oxidation of reduced sporidesmin. Selection of yeast transformants able to grow in the presence of high concentrations of sporidesmin allowed the cloning of a sporidesmin resistance gene, STSl (for sporidesmin toxicity suppressor) (Bissinger and Kuchler, 1994), which encodes a 1511-residue protein most closely related to Snq2 (35 % identical over the entire length. Figure 6). Moreover, Sts 1 is highly homologous to Drosophila white and brown and scarlet, three previously identified ABC transporters putatively required for eye pigmentation and tryptophan transport in the fly (O'Hare et al., 1984; Dreesen et al, 1988; Ewart et al., 1994) (Figure 6). Evolutionary tree analysis indicated that the C-terminal Stsl ABC domain is more closely related to a certain subfamily of ABC proteins (white, brown, scarlet, Ste6, CFTR, Snq2), while the N-terminal ABC domain of Sts 1 appears to be more closely related to the P-glycoproteins, suggesting that STSl may have evolved through gene fusion rather than gene duplication events (Bissinger and Kuchler, 1994). Interestingly, the domain organization of both Stsl and Snq2 is reversed as compared to other full-size P-glycoproteins, because the first ABC domain is positioned at the N-terminus of the polypeptides (Figure 7). A similar domain organization was predicted for white, brown and scarlet (O'Hare et al., 1984; Dreesen et al., 1988; Ewart et al., 1994).

Figure 6. The yeast Stsl multidrug transporter is an ABC protein. Stretchesfrom the C-terminal and N-terminal ATP-binding domains of Stsl (Stsl-C and Stsl-N) representing the Walker A, Walker B and ABC signature domains are aligned to the corresponding regions of other ABC proteins including Snq2, Adpl, white, brown, Ste6, Mdrl, HlyB. Amino acid positions are numbered according to the primary sequence of each protein with the methionine being 1. Amino acids identical to either the C-terminal or N-terminal ABC domain are shaded if conserved in at least four proteins.

ABC Transporters In Yeast

81

COOH

Figure 7. Predicted transmembrane topology of Stsi and Snq2. The membrane topology as predicted from the hydrophobicity analysis of 5757 and SNQ2 is shown. Solid black lines represent the polypeptide chains. Putative transmembrane segments are shown as vertical black bars. The two hydrophilic ABC-domains are marked by the black ovals and "ATP.'' Dotted oval balls indicate potential N-linked carbohydrate. Hydrophobicity analysis of SNQ2 allows for the prediction of a similar (ABC-TMS^)2 topology.

However, while white, brown and scarlet are half-size ABC-TMS^ transporters, Sts 1 and Snq2 are typical full-size ABC proteins (Figure 7). A putative (ABC-TMS6)2 counterpart of Stsl and/or Snq2 in mammalian cells has not yet been discovered. Chromosomal deletions of STSI or SNQ2 lead to viable cells of both mating types, indicating that the genes are not essential for cell growth (Servos et al., 1993; Balzi et al., 1994; Bissinger and Kuchler, 1994; Hirata et al., 1994). This suggests functional redundancy of Stsl and/or Snq2 in yeast, which is also supported by results from both low stringency Southern and Northern analysis hinting the existence of additional yeast genes homologous to STSI and SNQ2 (Kuchler and Bissinger, unpublished results). We have identified at least two additional transporters on chromosome XV which we have tentatively named Dosl and Dos2 (for daughter of Stsl) as they appear to be closely related to Stsl and Snq2 (Bissinger and Kuchler, unpublished results). Dos2 is very likely to be identical to the PdrlO ABC transporter which was identified

82

RALFEGNERetal.

during the course of the yeast genome sequencing project on chromosome XV at approximately the same map position as Dos2 (Parle and Wolfe, personal communication). Moreover, besides their homology to white/brown/scarlet, several of the newly identified yeast ABC transporters such as YIL013C, YKR103w and YER036p are most closely related to either Snq2 or Stsl (Table 1), implying that they may share functional features with Stsl and/or Snq2. While Astsl cells exhibit supersensitivity to a number of structurally unrelated drugs (Bissinger and Kuchler, 1994), Asnq2 cells are hypersensitive to a different set of compounds (Servos et al., 1993; Hirata et al, 1994). In fact, the MDR patterns conferred by Stsl and Snq2 are largely distinct (Meyers et al., 1992; Servos et al., 1993; Bissinger and Kuchler, 1994) with very Uttle overlap (Hu*ata et al., 1994), indicating a quite different substrate specificity for both transporters. However, this does not exclude the possibility that they have a common function in vivo, since the physiological substrates of neither Snq2 nor Stsl have been determined. Although a Asnq2 Astsl double mutant does not have an obvious detrimental growth phenotype, it exhibits a high degree of hypersensitivity to a large number of different drugs due to unexplained synergistic effects of the double deletion mutation (Egner and Kuchler, unpublished results). Drug and non-drug substrates of Stsl identified to date include sporidesmin, cycloheximide, chloramphenicol, rhodamine 123, fluconazole, itraconazole, ketoconazole, erythromycin, fluphenazine, sulphometuron methyl, antimycin, myconazole, lyncomycin, compactin, staurosporine, cerulenin, estradiol, dexamethasone (Balzi and Goffeau, 1994; Balzi et al., 1994; Bissinger and Kuchler, 1994; Hirata et al., 1994; Kralli et al., 1995) (Egner and Kuchler, unpubUshed results; Mah6 and Kuchler, in preparation), and confirmed substrates for Snq2 comprise 4-nitroquinoline-N-oxide,N-methyl-N'-nitro-N-nitrosoguanidin, triazoquinone, 0-phenanthroline, sulphometuron methyl, staurosporine, fluphenazine, corticosterone and corticoaldosterone (Servos et al., 1993; Balzi and Goffeau, 1994; Hirata et al., 1994; Kralli et al, 1995). It should be noted that the yeast STSl gene was independently cloned by several laboratories as the PDR5 (for pleiotropic drug resistance) gene, which was originally characterized in a cycloheximidebased mutant screen and mapped to chromosome XV (Leppert et al., 1990), and by yet another group as the YDRl multidrug resistance determinant (Hirata et al., 1994). A close homologue of yeast Stsl, Cdrl, was recently discovered in Candida albicans (Prasad et al., 1995). Expression of Cdrl in the yeast

ABC Transporters in Yeast

83

Saccharomyces is not only able to confer a multidrug resistance phenotype (Sanglard, personal communication; Prasad et al., 1995), it also complements stsl deletion mutations and restores wild-type drug sensitivity profiles similar to the one found in STSl cells (Sanglard et al., submitted). In fact, functional cloning approaches in yeast using genomic libraries from Candida albicans allowed for the isolation of several novel Candida multidrug resistance genes of the ABC protein family (Sanglard and Kuchler, unpublished observations). Amplification and overexpression of CDRl is one of the causes for MDR in Candida isolates from AIDS patients, and may in fact be largely responsible for the failure of antifungal treatment of reoccurring oropharyngeal candidiasis in HIV patients (Sanglard et al., submitted). A. Biosynthesis and Turnover of Yeast Multidrug Transporters

The observed MDR phenotype of cells overexpressing Stsl and Snq2 raises the possibility that these proteins may be components of an endogenous defense or detoxification system for toxic metabolites. Such a function would suggest the plasma membrane to be the normal cellular location for Sts 1 and Snq2. The Sts 1 drug pump was indeed found mainly in membrane preparations enriched for plasma membrane vesicles (Balzi et al., 1994; Decottignies et al., 1994) isolated frompdrl mutants which apparently overexpress Stsl (Balzi et al., 1994). Furthermore, protein microsequencing of Sts 1 purified from these plasma membrane fractions yielded an N-terminal protein sequence identical to the one encoded by the STSl gene (Balzi et al., 1994). The plasma membrane localization of Stsl was also confirmed by both refined subcellular fractionation procedures and indirect immunofluorescence experiments, which demonstrated that Stsl is a resident plasma membrane protein in yeast (Egner et al., submitted). Likewise, we have recently also demonstrated a plasma membrane localization for the Snq2 transporter (Mahe and Kuchler, in preparation). However, the plasma membrane localization of Stsl appears to be only transient, as we have recently demonstrated that Stsl is constitutively endocytosed from the cell surface and delivered to the vacuole for terminal degradation (Egner et al., submitted). Strikingly, like the Ste6 pheromone transporter, Stsl is a rather short-lived protein with an apparent half-life of about 60-90 minutes, and like Ste6, Stsl is ubiquitinated in vivo (Egner and Kuchler, in preparation). However, ubiquitination of Stsl does not tag the transporter for degradation by the cytoplasmic proteasome, since the half-life of Stsl is unaffected in pre

84

RALFEGNERetal.

7-7 mutants defective in proteasome mediated proteolysis (Egner et al., submitted). Although the function of ubiquitination of Stsl and Ste6 is not clear at the moment, we believe that ubiquitination may somehow represent a signal for removal of cell surface proteins by endocytosis (Egner et al., submitted). As pointed out above, overexpression of yeast multidrug transporters exhibits the hallmarks of P-glycoprotein-mediated MDR in animal cells. Yet again, as in animal cells, the physiological function(s) of the yeast multidrug transporters remains an enigma. A possible in vivo role for yeast Stsl and/or Snq2 in intracellular steroid or phospholipid transport as recently demonstrated for human Mdr2 (Smit et al., 1993; Ruetz and Gros, 1994a; Smith et al., 1994) is suggested from very interesting results obtained in two entirely different experimental setups. First, the group of Keith Yamamoto has been exploiting yeast genetics to study the ability of the rat glucocorticoid receptor (GR) to respond to steroid hormone analogs and activate transcription in yeast, so as to identify factors in the steroid signal transduction pathway (Garabedian and Yamamoto, 1992). There are several important differences in the GR activity in yeast and mammalian cells, the most important one being the fact that dexamethasone, a strong agonist in mammalian cells, is a very weak agonist in yeast. Thus, a genetic screen was devised to identify yeast mutants that would show increased sensitivity specifically to dexamethasone, but not to others, in an attempt to identify genes that modulate ligand efficacy of the glucocorticoid receptor (GR). Reporter genes such as P-Gal were placed under the transcriptional control provided by GR-responsive elements (GREs) and mutants were isolated that showed enhanced expression of the reporter genes in response to dexamethasone. The corresponding wild type genes were cloned by introducing CEN-based plasmids that reduced expression again back to the level seen in wildtype cells. This way, a gene named LEMl (for ligand efficacy modulator) was isolated. DNA sequence comparison demonstrated that LEMl is allelic to STSl (Kralli et al., 1995). These results demonstrate that Stsl can modulate intracellular dexamethasone availability by effluxing the ligand that was otherwise taken up. In a similar approach, SNQ2 was isolated as a modulator of GR function in yeast (Kralli et al., personal communication). However, in contrast to Stsl that appears to transport preferentially dexamethasone, overexpression of Snq2 seems to have a broader effect, because it can efflux many other steroid ligands (Kralli et al., personal communication).

ABC Transporters in Yeast

85

Second, yeast genetics was used to analyze the phenomenon of transcriptional interference or squelching (Gilbert et al., 1993). As in mammalian cells, the hormone binding domain of estrogen receptor (ER) can subject the activity of a heterologous protein, which is fused to it, to hormonal control (e.g., estradiol). Thus, a chimeric transcriptional activator consisting of the ER hormone binding and DNA binding domain plus the activation domain of viral protein 16 (VP16) stimulates both episomal and integrated reporter genes exclusively in the presence of steroid hormones. When a chimeric VP16-human ER is expressed in yeast at high levels, the presence of estradiol in the medium is normally highly toxic and leads to cell death caused by squelching (Gilbert et al., 1993). Consequently, the search for suppressor mutants led to the repeated isolation of a pdrl allele as a squelching suppressor, because this mutant is apparently capable of efficiently modulating intracellular steroid availability. Intriguingly, it has been shown XhdXpdrl mutant cells overexpress STSl mRNA (Meyers et al., 1992) and thus, the observed suppression of estradiol toxicity may be due to elevated levels of the S ts 1 pump. Indeed, we have recently tested this idea and demonstrated by immunoblotting using polyclonal antibodies against Stsl and Snq2 that not only Stsl, but also Snq2 is dramatically overexpressed in pdrl-3 mutants (Mahe and Kuchler, in preparation). Moreover, analysis of estradiol mediated squelching in Astsl Asnql double mutants, as well as in the respective single mutants, revealed a marked increased sensitivity to estradiol as compared to the isogenic wild-type strain, demonstrating for the first time a function of Stsl and Snq2 in steroid transport in vivo (Mahe and Kuchler, in preparation). Taken together, these results suggest that both Stsl and Snq2 are able to modulate intracellular concentrations of many steroid analogs by pumping them out of cells. Based on these findings, we would speculate that a possible in vivo function of Stsl/Leml and/or Snq2 could be a role in maintaining or regulating lipid homeostasis of cellular membranes, so as to regulate membrane fluidity and permeability (Lipowsky, 1991). It seems plausible, that a mechanism must exist by which membrane fluidity and rigidity is controlled, since this is a prerequisite for regulated transport processes across membranes or vesicle formation which, in turn, regulates growth, proliferation and cellular cross-talk (Lipowsky, 1991). Interestingly, a detailed analysis of the sterol composition of various yeast membranes demonstrated that the plasma membrane represents the cellular membrane with the highest sterol content (Zinser et al., 1993). It will be of great interest to test

86

RALFEGNERetal.

whether Astsl and Asnq2 mutants have an altered sterol composition in the plasma membrane. The expression level of yeast Stsl/Snq2 multidrug transporters in the plasma membrane could be responsible for maintaining a desired sterol concentration and/or composition in the membrane. This could be achieved by binding of various sterols to the hydrophobic TMS domains of Stsl/Pdr5/Leml or Snq2, similar to the proposed vacuum cleaner model for the interaction of hydrophobic drugs and steroids with mammalian P-glycoproteins (Gottesman and Pastan, 1993; Kane, 1995). Indeed, mammalian P-glycoproteins have been demonstrated to interact with steroids, and are able to mediate transcellular transport of steroids and phospholipids in vitro and in vivo (Yang et al., 1989; Qian and Beck, 1990; Saeki et al., 1992; Ueda et al., 1992; Smit et al.,1993; van Kalken et al., 1993; Smith et al., 1994). The availability of numerous mutants in both the yeast ergosterol(Chambonetal., 1991; Casey etal., 1992;Zinser et al., 1993) and phospholipid biosynthetic pathway (Nikoloff and Henry, 1991) will allow to design future experiments to test this hypothesis for the yeast multidrug transporters. Moreover, these experiments will be greatly facilitated using an in vitro vesicle transport system isolated from late sec mutants blocked in exocytic vesicle fusion (Nakamoto et al., 1991), in which uniformly oriented unilamellar vesicles can be accumulated by simply shifting the cells to the restrictive temperature (Nakamoto et al., 1991; Pryer et al., 1992). These tightly sealed vesicles can be purified by conventional biochemical techniques (Ruetz and Gros, 1994b) and used for in vitro transport studies. Finally, it should be noted that compound MDR phenomena in yeast are subject to control by a highly complex network of various different transcription factors which appear to regulate both ABC transporters and other membrane transporters, most of which belong to the major facilitator superfamily (Balzi and Goffeau, 1994). For example, PDRl (Balzi et al., 1987), PDR3 (Delaveau et al., 1992; Ruttkay et al., 1992), CADI (Wu et al., 1993), the allelic SNQ3/YAP1/PAR1 genes (Moye-Rowley et al, 1989;Hertleetal., 1991 ;Schnell and Entian, 1991),PD/?7,andPD/?9 (Dexter et al., 1994) encode putative transcription factors imphcated in both drug and heavy metal resistance phenomena. Transcription oiSTSl was demonstrated to be under control of transcription regulatory proteins such as Pdrl and Pdr3 (Katzmann et al., 1994). Genetic experiments have shown that pdrl mediated cycloheximide resistance requires the presence of STSl and that STSl mRNA levels were elevated in a drug-resistant/7Jr7-i mutant (Meyers et al., 1992). By contrast, in Apdrl Apdr3

ABC Transporters in Yeast

87

double mutants, STSl mRNA synthesis is completely abolished (Katzmann et al., 1994). In conclusion, numerous yeast genes may comprise a distinct and rather large gene family responsible for the coordinate regulation and phenotypic expression of MDR and HMR phenomena. It will be interesting to determine if and which yeast MDR genes (both ABC and non-ABC genes) interact in common metabolic pathways to exert their normal cellular functions. Although this field of research is just at the beginning, we would nonetheless expect and anticipate that we will soon be able to identify such pathways and identify metaboUc cross-roads and/or the precise targets of transcriptional regulators by exploiting yeast genetics and synthetic lethality screens.

V. CONCLUSION AND PERSPECTIVES The lower eukaryote yeast has become a evolutionary storehouse for a number of functionally diverse ABC proteins, many of which appear to have structural and/or functional homologues in mammalian cells. This suggests that this lower eukaryote represents a valuable model system for the molecular genetic analysis of both endogenous and heterologous ABC proteins. In bacteria, the purification of individual ABC proteins and their functional reconstitution into proteoliposomes has substantially contributed to our understanding how these bacterial ABC transport systems work in vivo. The Ste6 a-factor transporter represents an ideal candidate protein for similar studies because its physiological substrate is known. Likewise, the yeast multidrug transporters Snq2 and Stsl, as well as many others listed in Table 1, are also excellent candidates for in vitro studies, since compelling evidence from other systems already suggests possible in vivo functions for the yeast multidrug transporters. Similar studies can be carried out on mammalian ABC proteins, since they can be functionally expressed in yeast (Kuchler and Thomer, 1992a; Raymond et al., 1992; Ueda et al., 1993). In addition, in cases where a functional complementation between yeast and animal ABC proteins has been demonstrated (Kuchler et al., 1992; Raymond et al., 1992), the construction of chimeric ABC transporters can be utilized for a moleculargenetic structure-function analysis to determine domains of functional importance in ABC proteins from different organisms. Such a strategy was employed by creating chimeric CFTR-Ste6 transporters enabling the identification of functional domains in CFTR (Teem et al., 1993). Moreover, the recent discovery of yeast Ycfl has further strengthened the importance of yeast as a model system for structure-function studies on

88

RALFEGNERetal.

human CFTR and other mammaUan ABC proteins associated with the pathology of human disease. In addition, once a functional assay for an ABC protein can be developed, one can utilize cloning by function to isolate in yeast the corresponding cDNAs from mammalian cells in cases where mammalian counterparts of yeast ABC proteins are unknown. Finally, the accessibility of yeast genetics enables the isolation and characterization of unlimited numbers of transport mutants of any ABC transporter, since expression of mutant ABC proteins in yeast can be monitored by phenotypic selection or other functional assays. Moreover, it goes without saying that much more structural work will have to be done to allow for uncovering the molecular mechanism of ABC protein mediated transport. It may be feasible, and possible, to subject ABC proteins to both 2D and 3D crystallization attempts, once they become available in larger quantities and in highly purified form. Ultimately, these strategies will result in a dissection of functional domains in individual transporters and facilitate to uncover their molecular mechanism of action. The fact that some of the yeast ABC proteins have been isolated independently in several different laboratories, sometimes using entirely different approaches and/or experimental strategies, suggests that saturation may have been reached concerning the number of different yeast ABC protein genes. However, we would argue that the functional diversity of bacterial ABC transporters (Higgins, 1992) implies that we can anticipate even more ABC proteins to be discovered in both yeast and in mammalian systems in the near future. About 60% of the yeast genome has been sequenced to date, revealing more than two dozen different ABC protein genes. If statistics holds true, one can predict at least some 50 different ABC transporters in yeast. The definite number of yeast ABC transporters is in sight, as the yeast genome sequencing project is anticipated to be finished by the end of 1996. The pace by which new ABC transporters are being discovered in many organisms including yeast, sets the stage for high expectations and certainly promises new and exciting findings on this remarkable family of transporters in many years to come.

ACKNOWLEDGMENTS We are deeply indebted to Jeremy Thomer for communicating unpublished results. Thanks to David Chaplin for providing the murine IL-la cDNA and to Cindy Trueblood and Susan Michaelis for yeast strains CT300 and SM1229, respectively. We would like to thank our collaborators Ralf Kolling and

ABC Transporters in Yeast

89

Dominique Sanglard for many helpful ideas and interesting discussions. The personal communications of unpublished results by many colleagues are thankfully acknowledged. The critical reading of the manuscript by Franz Wohlrab is highly appreciated. Work in my laboratory on ABC proteins is supported by grants from the Austrian National Bank (project OENB 4486) and by funds from the Austrian Science Foundation (project P-09537 and P-10123). R.E. is supported by a postdoctoral fellowship from the "Deutsche Forschungsgemeinschaft." R.P. was supported in part by a fund from the Commission of Oncology of the University of Vienna, Medical Faculty.

REFERENCES Anderegg, R.J., Betz, R., Carr, SA., Crabb, J.W., & Duntze, W. (1988). Structure of the Saccharomyces cerevisiae mating hormone a-factor: identification of S-famesyl cysteine as a structural component. J. Biol. Chem. 263, 18236-18240. Ashby, M.N., King, D.S., & Rine, J. (1992). Endoproteolytic processing of a famesylated peptide in vitro. Proc. Natl. Acad. Sci. (USA) 89,4613^617. Ashby, M.N., Errada, PR., Boyartchuk, V.L., & Rine, J. (1993). Isolation and DNA sequence of the STEM gene encoding famesyl cysteine carboxyl methyltransferase. Yeast 9, 907-913. Aubourg, R, Mosser, J., Douar, A.M., Sarde, CO., Lopez, J., & Mandel, J.L. (1993). Adrenoleukodystrophy gene: unexpected homology to a protein involved in peroxisome biogenesis. Biochimie 75, 293-302. Baldari, C , Murray, J.A., Ghiara, R, Cesareni, G., & Galeotti, C.L. (1987). Anovel leader peptide which allows efficient secretion of a fragment of human interleukin-ip in Saccharomyces cerevisiae. EMBO J. 6, 229-234. Balzi, E., Chen, W., Ulaszewski, S., Capieaux, E., & Goffeau, A. (1987). The multidrug resistance gene PDRl from Saccharomyces cerevisiae. J. Biol. Chem. 262,1687116879. Balzi, E. & Goffeau, A. (1994). Genetics and biochemistry of yeast multidrug resistance. Biochim. Biophys. Acta 1187,152-162. Balzi, E., Wang, M., Leterme, S., van Dyke, L., & Goffeau, A. (1994). PDR5, a novel yeast multidrug resistance conferring transporter controlled by the transcription regulator PD/?7. J. Biol. Chem. 269, 2206-2214. Berkower, C , Loayza, D., & Michaelis, S. (1994). Metabolic instability and constitutive endocytosis of STE6, the a-factor transporter of Saccharomyces cerevisiae. Mol. Biol. Cell 3, 633-654. Berkower, C. & Michaelis, S. (1991). Mutational analysis of the yeast a-factor transporter STE6, a member of the ATP binding cassette (ABC) protein superfamily. EMBO J. 10, 3777-3785. Bissinger, P. & Kuchler, K. (1994). Molecular cloning and expression of the Saccharomyces cerevisiae STSl gene product: a yeast ABC-transporter conferring mycotoxin resistance. J. Biol. Chem. 269,4180-4186. Borst, P., Schinkel, A.H., Smit, J.J., Wagenaar, E., van Deemter, L., Smith, A.J., Eijdems, E.W., Baas, F. & Zaman, G.J. (1993). Classical and novel forms of multidrug

90

RALFEGNERetal.

resistance and the physiological functions of P-glycoproteins in mammals. Pharmacol. Ther. 60, 289-299. Bossier, P., Femandes, L., Vilela, C , & Rodrigues-Pousada, C. (1994). The yeast YKL741 gene situated on the left arm of chromosome XI codes for a homologue of the human ALD protein. Yeast 10,681-686. Brake, A.J., Brenner, C , Najarian, R., Laybourn, P., & Merry weather, J. (1985). Structure of genes encoding precursors of the yeast peptide mating pheromone a-factor. In: Protein Transport and Secretion, Gething, M.-J., ed., pp. 103-108. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York. Callahan, H.L. & Beverley, S.M. (1991). Heavy metal resistance: a new role for P-glycoproteins in Leishmania. J. Biol. Chem. 266, 18427-18430. Casey, W.M., Keesler, G.A. & Parks, L.W. (1992). Regulation of partitioned sterol biosynthesis in Saccharomyces cerevisiae. J. Bacteriol. 174, 7283-7288. Chambon, C , Ladeveze, V., Servouse, M., Blanchard, L., Javelot, C , Vladescu, B., & Karst, F. (1991). Sterol pathway in yeast. Identification and properties of mutant strains defective in mevalonate diphosphate decarboxylase and famesyl diphosphate synthetase. Lipids 26, 633-636. Chen, P., Johnson, P., Sommer, T, Jentsch, S., & Hochstrasser, M. (1993). Multiple ubiquitin-conjugating enzymes participate in the in vivo degradation of the yeast MATo2 repressor. Cell 74, 357-369. Cowman, A.F. & Karcz, S. (1993). Drug resistance and the P-glycoprotein homologues of Plasmodium falciparum. Sem. Cell Biol. 4, 29-35. Cyr, D.M., Stuart, R.A., & Neupert, W. (1993). A matrix ATP-requirement for presequence translocation across the inner membrane of mitochondria. J. Biol. Chem. 268, 23751-23754. Davis, N.G., Horecka, J.L., & Sprague Jr., G.F. (1993). cis- and rraw^-acting functions required for endocytosis of the yeast pheromone receptors. J. Cell Biol. 122,53-65. Dean, M., AUikmets, R., Gerrard, B., Stewart, C , Kistler, A., Shafer, B., Michaelis, S., & Strathem, J. (1994). Mapping and sequencing of two yeast genes belonging to the ATP-binding cassette superfamily. Yeast 10,377-383. Decottignies, A., Kolaczkowski, M., Balzi, E., & Goffeau, A. (1994). Solubilization and characterization of the overexpressed PDR5 multidrug resistance nucleotide triphosphatase of yeast. J. Biol. Chem. 269, 12797-12803. Delaveau, T, Jacq, C , & Perea, J. (1992). Sequence of a 12.7 kb segment of yeast chromosome II identifies a PDR-\\Yt gene and several new open reading frames. Yeast 8, 761-768. Dexter, D., Moye-Rowley, W.S., Wu, A.L., & Golin, J. (1994). Mutations in the yeast PDR3, PDR4, PDR7 and PDR9 pleiotropic (multiple) drug resistance loci affect the transcript level of an ATP binding cassette transporter encoding gene, PDR5. Genetics 136, 505-515. Dinarello, C.A. (1992). The biology of interleukin-1. Chem. Immunol. 51,1-32. Dreesen, T.D., Johnson, D.J., & Henikoff, S. (1988). The brown protein of Drosophila melanogaster is similar to the white protein and to components of active transport complexes. Mol. Cell. Biol. 8, 5206-5215. Epand, R.F, Xue, C.B., Wang, S.H., Naider, F, Becker, J.M., & Epand, R.M. (1993). Role of prenylation in the interaction of the a-factor mating pheromone with phospholipid bilayers. Biochemistry 32, 8368-8373.

ABC Transporters in Yeast

91

Ewart, G.D., Cannell, D., Cox, G.B., & Howells, A.J. (1994). Mutational analysis of the traffic ATPase (ABC) transporters involved in uptake of eye pigment precursors in Drosophila melanogaster. Implications for structure-function relationships. J. Biol. Chem. 269, 10370-10377. Garabedian, M.J. & Yamamoto, K.R. (1992). Genetic dissection of the signaling domain of a mammalian steroid receptor in yeast. Mol. Biol. Cell. 3,1245-1257. Gartner, J. & Vallee, D. (1993). The 70 kDa peroxisomal membrane protein: an ATPbinding cassette transporter protein involved in peroxisome biogenesis. Semin. Cell Biol. 4, 45-52. Gartner, J., Moser, H., & Vallee, D. (1992). Mutations in the 70 kd peroxisomal membrane protein gene in Zellweger syndrome. Nature Genet. 1,16-20. Gilbert, D.M., Heery, D.M., Losson, R., Chambon, R, & Lemoine, Y. (1993). Estradiolinducible squelching and cell growth arrest by a chimeric VP16-estrogen receptor expressed in Saccharomyces cerevisiae: suppression by an allele of PDRl. Mol. Cell. Biol. 13,462-472. Glomset, J.A., Gelb, M.H., & Famsworth, C.C. (1990). Prenyl proteins in eukaryotic cells: a new type of membrane anchor. Trends Biochem. Sci. 15,139-142. Goodman, L.E., Judd, S.R., Famsworth, C.C, Powers, S., Gelb, M.H., Glomset, J.A., & Tamanoi, F. (1990). Mutants of Saccharomyces cerevisiae defective in the famesylation of Ras proteins. Proc. Natl. Acad. Sci. USA 87,9665-9669. Gottesman, M. & Pastan, I. (1993). Biochemistry of multidrug resistance mediated by the multidrug transporter. Ann. Rev. Biochem. 62, 385-427. Haase, E., Servos, J., & Brendel, M. (1992). Isolation and characterization of additional genes influencing resistance to various mutagens in the yeast Saccharomyces cerevisiae. Curr. Genet. 21,319-324. Hancock, J.F., Cadwallader, K., & Marshall, C.J. (1991a). Methylation and proteolysis are essential for efficient membrane binding of prenylated p21K-ras(B). EMBO J 10, 641-646. Hancock, J.R, Cadwallader, K., Paterson, H., & Marshall, C.J. (1991b). A CAAX or a CAAL motif and a second signal are sufficient for plasma membrane targeting of ras proteins. EMBO J. 10,4033-4039. Hannavy, K., Rospert, S., & Schatz, G. (1993). Protein import into mitochondria: a paradigm for the translocation of polypeptides across membranes. Curr. Opin. Cell Biol. 5, 694-700. Harris, A. & Argent, B.E. (1993). The cystic fibrosis gene and its product CFTR. Semin. Cell Biol. 4, 37-44. He, B., Chen, R, Chen, S.Y., Vancura, K.L., Michaelis, S., & Powers, S, (1991). RAM2, an essential gene of yeast, and RAMI encode the two polypeptide components of the famesyltransferase that prenylates a-factor and Ras proteins. Proc. Natl. Acad. Sci. USA 88, 11373-11377. Hertle, K., Haase,E., & Brendel, M. (1991). The SNQ3 gene of Saccharomyces cerevisiae confers hyper-resistance to several functionally unrelated chemicals. Curr. Genet. 19,429-433. Higgins, C.F. (1992). ABC-Transporters: From microorganisms to man. Ann. Rev. Cell Biol. 8,67-113.

92

RALFEGNERetal.

Hill, A. & Ploegh, H. (1995). Getting the inside out: the transporter associated with antigen processing (TAP) and the presentation of viral antigen. Proc. Natl. Acad. Sci. USA 92, 341-343. Hirata, D., Yano, K., Miyahara, K., & Miyakawa, T. (1994). Saccharomyces cerevisiae YDRl, which encodes a member of the ATP-binding cassette (ABC) superfamily, is required for multidrug resistance. Curr. Genet. 26, 285-294. Hrycyna, C.A. & Clarke, S. (1992). Maturation of isoprenylated proteins in Saccharomyces cerevisiae. Multiple activities catalyze the cleavage of the three carboxy-terminal amino acids from famesylated substrates in vitro. J. Biol. Chem. 267, 10457-10464. Hrycyna, C.A., Sapperstein, S.K., Clarke, S.,& MichaeHs,S. (1991). TYit Saccharomyces cerevisiae STEM gene encodes a methyltransferase that mediates C-terminal methylation of a-factor and RAS proteins. EMBO J. 10, 1699-1709. Jackson, C.L. & Hartwell, L.H. (1990). Courtship in Saccharomyces cerevisiae: both cell types choose mating partners by responding to the strongest pheromone signal. Cell 63, 1039-1051. Jackson, C.L., Konopka, J.B., & Hartwell, L.H. (1991). S. cerevisiae a-pheromone receptors activate a novel signal transduction pathway for mating partner discrimination. Cell 67, 389-402. Jordan, T.W. & Cordiner, S.J. (1987). Fungal epipolythiodioxopiperazine toxins have therapeutic potential and roles in disease. Trends Pharmacol. Sci. 8,144-149. Kamijo, K., Taketani, S., Yokota, S., Osumi, T, & Hashimoto, T. (1990). The 70-kDa peroxisomal membrane protein is a member of the Mdr (P-glycoprotein)-related ATP-binding protein superfamily. J. Biol. Chem. 265,4534-4540. Kane, S.E. (1995). Muhidrug resistance of cancer cells. In: Advances in Drug Research, B. Testa and U.A. Meyer, eds. Academic Press Ltd., San Diego, in press. Katzmann, D.J., Burnett, RE., Golin, J., Mah^, Y, & Moye-Rowley, W.S. (1994). Transcriptional control of the yeast PDR5 gene by the PDR3 gene product. Mol. Cell. Biol. 14,4653-4661. Knop, M., Schiffer, H.H., Rupp, S., & Wolf, D.H. (1993). Vacuolar/lysosomal proteolysis: proteases, substrates, mechanisms. Curr. Opin. Cell Biol. 5, 990-996. Kolling, R. & Hollenberg, C.R (1994). The ABC-transporter Ste6 accumulates in the plasma membrane in a ubiquitinated form in endocytosis mutants. EMBO J. 13, 3261-3271. Koronakis, V. & Hughes, C. (1993). Bacterial signal peptide-independent protein export: HlyB-directed secretion of hemolysin. Semin. Cell Biol. 4, 7-15. Kralli, A., Bohen, S.R, & Yamamoto, K.R. (1995). LEMl, an ATP-binding cassette transporter, selectively modulates the biological potency of steroid hormones. Proc. Natl. Acad. Sci. USA 92,4701-4705. Kuchler, K. (1993). Unusual routes of protein secretion: the easy way out. Trends Cell Biol. 3,421-426. Kuchler, K. & Thomer, J. (1990). Membrane translocation of proteins without hydrophobic signal sequences. Curr. Opin. Cell Biol. 2, 617-624. Kuchler, K. & Thomer, J. (1992a). Functional expression of human mdrl cDNA in Saccharomyces cerevisiae. Proc. Natl. Acad. Sci. USA 89, 2302-2306.

ABC Transporters in Yeast

93

Kuchler, K. & Thomer, J. (1992b). Secretion of peptides and proteins lacking hydrophobic signal sequences: the role of ATP-driven membrane translocators. Endocrine Rev. 13,499-514. Kuchler, K., Sterne, R.S., & Thomer, J. (1989). Saccharomyces cerevisae STE6 gene product: a novel pathway for protein export in eukaryotic cells. EMBO J. 8, 3973-3984. Kuchler, K., Swartzman, E.R., & Thomer, J. (1994). A novel mechanism for transmembrane translocation of peptides: the Saccharomyces cerevisiae Ste6 Transporter and export of the mating pheromone a-factor. Curr. Top. Membranes 41,19-41. Kuchler, K., Dohlman, H., & Thomer, J. (1993). The a-factor transporter {STE6 gene product) and cell polarity in Saccharomyces cerevisiae. J. Cell Biol. 120, 12031215. Kuchler, K., Goransson, M., Visnawathan, M., & Thomer, J. (1992). Dedicated transporters for peptide export and intercompartmental traffic in yeast. Cold Spring Harbor Symp. Quant. Biol 57, 579-592. Leighton, J. & Schatz, G. (1995). An ABC transporter of the inner mitochondrial membrane is required for normal growth of yeast. EMBO J. 14, 188-195. Leppert, G., McDevitt, R., Falco, S.C., Van Dyk, T.K., Ficke, M.B., & Golin, J. (1990). Cloning by gene amplification of two loci conferring multiple drug resistance in Saccharomyces. Genetics 125,13-20. Lipowsky, R. (1991). The conformation of membranes. Nature 349,475-481. Madden, K., Costigan, C , & Snyder, M. (1992). Cell polarity and morphogenesis in Saccharomyces cerevisiae. Trends Cell Biol. 2, 22-29. Marcus, S., Caldwell, G.A., Miller, D., Xue, C.B., Naider, P., & Becker, J.M. (1991). Significance of C-terminal cysteine modifications to the biological activity of the Saccharomyces cerevisiae a-factor mating pheromone. Mol. Cell. Biol. 11,36033612. Marcus, S., Caldwell, G.A., Xue, C.B., Naider, P., & Becker, J.M. (1990). Total in vitro maturation of the Saccharomyces cerevisiae a-factor lipopeptide mating pheromone. Biochem. Biophys. Res. Commun. 172,1310-1316. McGrath, J.P. & Varshavsky, A. (1989). The yeast STE6 gene encodes a homologue of the mammalian multidrug resistance P-glycoprotein. Nature 340,400-404. Meyer, T.H., van Endert, RM., Uebel, S., Ehring, B., & Tampe, R. (1994). Functional expression and purification of the ABC transporter complex associated with antigen processing (TAP) in insect cells. FEBS Lett. 351,443^47. Meyers, S., Schauer, W., Balzi, E., Wagner, M., Goffeau, A., & Golin, J. (1992). Interaction of the yeast pleiotropic drug resistance genes PDRl and PDR5. Curr. Genet. 21,431-436. Michaehs, S. (1993). STE6, the yeast a-factor transporter. Semin. Cell Biol. 4,17-27. MichaeUs, S, & Herskowitz, L (1988). The a-factor pheromone of Saccharomyces cerevisiae is essential for mating. Mol. Cell. Biol. 8,1309-1318. Mosser, J., Douar, A., Sarde, C , Kioschis, P., Feil, R., Moser, H., Poustka, A., Mandel, J., & Aubourg, P. (1993). Putative X-linked adrenoleukodystrophy gene shares unexpected homology with ABC-transporters. Nature 361,726-730. Moye-Rowley, W.S., Harsham, K.D., & Parker, C.S. (1989). Yeast YAPl encodes a novel form of the Jun family of transcriptional activator proteins. Genes Dev. 3,283-292.

94

RALFEGNERetal.

Nakamoto, R.K., Rao, R., & Slayman, C.W. (1991). Expression of the yeast plasma membrane [H+]ATPase in secretory vesicles. A new strategy for directed mutagenesis. J. Biol. Chem. 266, 7940-7949. Neefjes, J.J. & Momburg, F. (1993). Cell biology of antigen presentation. Curr. Opin. Immunol. 5, 27-34. Nikoloff, D.M. & Henry, S.A. (1991). Genetic analysis of yeast phospholipid biosynthesis. Annu. Rev. Genet. 25, 559-583. Nishi, K., Yoshida, M., Horinouchi, S., & Beppu, T. (1993). Mating of the fission yeast occurs independently of prndl-^ gene product, a structural homologue of budding yeast STE6 and mammalian P-glycoproteins. Arch. Microbiol. 160,162-165. Nishi, K., Yoshida, M., Nishimura, M., Nishikawa, M., Nishiyama, M., Horinouchi, S., & Beppu, T. (1992). A leptomycin B resistance gene of Schizosacchawmyces pombe encodes a protein similar to the mammalian P-glycoproteins. Mol. Microbiol. 6, 761-769. 0*Hare, K., Murphy, C., Levis, R., & Rubin, G.M. (1984). DNA sequence of the white locus of Drosophila melanogaster. J. Mol. Biol. 180,437-455. Ortiz, D.F., Kreppel, L., Speiser, D.M., Scheel, G., McDonald, G., & Ow, D.W. (1992). Heavy metal tolerance in the fission yeast requires an ATP-binding cassette-type vacuolar membrane transporter. EMBO J. 11, 3491-3499. Pfanner, N., Rassow, J., van der Klei, I., & Neupert, W. (1992). Adynamic model of the mitochondrial protein import machinery. Cell 68,999-l(X)2. Prasad, R., de Wergifosse, P., Goffeau, A., & Balzi, E. (1995). Molecular cloning and characterization of a novel multidrug resistance gene of Candida albicans, CDRL Curr. Gen. 27, 320-329. Price, J. & Vemer, K. (1993). Puromycin inhibits protein import into mitochondria by interfering with an intramitochondrial ATP-dependent reaction. Biochim. Biophys. Acta 1150, 89-97. Pryer, N.K., Wuestehube, L.J., & Schekman, R. (1992). Vesicle-mediated protein sorting. Annu. Rev. Biochem. 61,471-516. Purnelle, B., Skala, J., & Goffeau, A. (1991). The product of the YCR105 gene located on the chromosome III from Saccharomyces cerevisiae presents homologies to ATP-dependent permeases. Yeast 7, 867-872. Qian, X.D. & Beck, W.T. (1990). Progesterone photoaffinity labels P-glycoprotein in multidrug-resistant human leukemic lymphoblasts. J. Biol. Chem. 265, 1875318756. Raths, S., Rohrer, J., Crausaz, F, & Riezman, H. (1993). end3 and end4\ two mutants defective in receptor-mediated and fluid-phase endocytosis in Saccharomyces cerevisiae. J. Cell Biol. 120, 55-65. Raymond, M., Gros, P., Whiteway, M., & Thomas, D.Y (1992). Functional complementation of yeast ste6 by a mammalian multidrug resistance mdr gene. Science 256, 232-234. Richter-Ruoff, B., Wolf, D.H., & Hochstrasser, M. (1994). Degradation of the yeast MATo2 transcriptional regulator is mediated by the proteasome. FEBS Lett. 354, 50-52. Riezman, H. (1993). Yeast endocytosis. Trends Cell Biol. 3,273-277. Riordan, J.R., Rommens, J.M., Kerem, B.-S., Alon, N., Rozmahel, M., Grzelczak, Z., Zielenski, J., Lok, S., Plavsic, N., Chou, J.-L., Drumm, M.L., lanuzzi, M.C.,

ABC Transporters in Yeast

95

Collins, F.S., & Tsui, L.-C. (1989). Identification of the cystic fibrosis gene: cloning and characterization of complementary DNA. Science 245,1066-1073. Rohrer, J., Benedetti, H., Zanolari, B., & Riezman, H. (1993). Identification of a novel sequence mediating regulated endocytosis of the G protein-coupled a-pheromone receptor in yeast. Mol. Biol. Cell 4, 511-521. Rubartelli, A., CozzoHno, F., Talio, M., & Sitia, R. (1990). A novel secretory pathway for Interleukin-ip, a protein lacking a signal sequence. EMBO J. 9, 1503-1510. Ruetz, S. & Gros, R (1994b). Functional expression of P-glycoproteins in secretory vesicles. J. Biol. Chem. 269,12277-12284. Ruetz, S. & Gros, R (1994a). Phosphatidylcholine translocase: a physiological role for the mdr2 gene. Cell 77,1071-1081. Ruttkay, N.B., Obemauerova, M., & Subik, J. (1992). High-level resistance to cycloheximide resulting from an interaction of the mutated pdr3 and cyh genes in yeast. Curr. Genet. 22, 337-339. Saeki, T, Shimabuku, A.M., Ueda, K., & Komano, T. (1992). Specific drug binding by purified lip id-reconstituted P-glycoprotein: dependence on the lipid composition. Biochim. Biophys. Acta 1107,105-110. Sandbaken, M., Lupisella, J.A., Di-Domenico, B., & Chakraburtty, K. (1990). Protein synthesis in yeast. Structural and functional analysis of the gene encoding elongation factor 3. J. Biol. Chem. 265,15838-15844. Sapperstein, S., Berkower, C , & Michaelis, S. (1994). Nucleotide sequence of the yeast STEM gene, which encodes famesyl cysteine carboxyl methyltransferase, and demonstration of its essential role in a-factor export. Mol. Cell Biol. 14, 14381449. Schafer, W. & Rine, J. (1992). Protein prenylation: genes, enzymes, targets, and functions. Ann. Rev. Genet. 26,209-237. Schnell, N. & Entian, K.D. (1991). Identification and characterization of a Saccharomyces cerevisiae gene (PARI) conferring resistance to iron chelators. Eur. J. Biochem. 200,487-493. Servos, J., Haase, E., & Brendel, M. (1993). Gene SNQ2 of Saccharomyces cerevisiae^ which confers resistance to 4-nitroquinoline-N-oxide and other chemicals, encodes a 169 kDa protein homologous to ATP-dependent permeases. Mol. Gen. Genet. 236, 214-218. Shani, N., Watkins, RA., & Valle, D.L. (1995). PXAl, a putative S. cerevisiae homolog of the human adrenolykodystrophy gene. Mol. Biol. Cell 5, A42. Smit, J.J.M., Schinkel, A.H., Oude Elferink, R.P.J., Groen, A.K., Wagenaar, E., van Deemter, L., Mol, C.A.A.M., Ottenhoff, R., van der Lug, N.M.T., van Room, M.A., vander Kalk, M.A., Offerhaus, G.J.A., Bems, A.J.M., & Borst, R (1993). Homozygous disruption of the murine mdr2 P-glycoprotein gene leads to a complete absence of phospholipid from bile and to liver disease. Cell 75,451-462. Smith, A.J., Timmermans, H.J.L., Roelofsen, B., Wirtz, K.W., van Blitterswijk, W.J., Smit, J.J., Schinkel, A.H., & Borst, R (1994). The human MDR3 P-glycoprotein promotes translocation of phosphatidylcholine through the plasma membrane of fibroblasts from transgenic mice. FEBS Lett. 354, 263-266. Sprague, G.F. & Thomer, J. (1992). Pheromone response and signal transduction during the mating process of Saccharomyces cerevisiae. In: The Molecular Biology of the

96

RALFEGNERetal.

Yeast Saccharomyces cerevisiae. Second Edition, pp. 657-744. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York. Steme-Marr, R.E., Blair, L.C., & Thomer, J. (1990). Saccharomyces cerevisiae STEM gene is required for COOH-terminal methylation of a-factor mating pheromone. J. Biol. Chem. 265, 20057-20060. Swartzman, E.E., Viswanathan, M.N., Emerick, A.E., & Thomer, J. (1995). S. cerevisiae PALI gene product, homologous to human ALDp and Pmp70, is required for peroxisome function in vivo. Mol. Biol. Cell 5, A2789. Szczypka, M.S., Wemmie,J.A.,Moye-Rowley,W.S.,&Thiele,D.J. (1994). Ayeastmetal resistance protein similar to human cystic fibrosis transmembrane conductance regulator (CFTR) and multidrug resistance-associated protein. J. Biol. Chem. 269, 22853-22857. Teem, J.L., Berger, H.A., Ostedgaard, L.S., Rich, D.R, Tsui, L.C., & Welsh, M.J. (1993). Identification of revertants for the cystic fibrosis delta F508 mutation using STE6-CFTR chimeras in yeast. Cpll 73, 335-346. Thissen, J.A. & Casey, RJ. (1993). Microsomal membranes contain a high affinity binding site for prenylated peptides. J. Biol. Chem. 268, 13780-13783. Townsend, A. & Trowsdale, J. (1993). The transporters associated with antigen presentation. Semin. Cell Biol. 4, 53-61. Ueda, K., Okamura, N., Hirai, M., Tanigawara, Y, Saeki, T, Kioka, N., Komano, T, & Hori, R. (1992). Human P-glycoprotein transports Cortisol, aldosterone, and dexamethasone, but not progesterone. J. Biol. Chem. 267, 24248-24252. Ueda, K., Shimabuku, A. M., Konishi, H., Fujii, Y, Takebe, S., Nishi, K., Yoshida, M., Beppu, T, & Komano, T. (1993). Functional expression of human P-glycoprotein in Schizosaccharomyces pombe. FEBS Lett. 330, 279-282. Umesono, K., Inokuchi, H., Shiki, Y, Takeuchi, M., Chang, Z., Fukuzawa, H., Kohchi, T, Shirai, H., Ohyama, K., & Ozeki, H. (1988). Structure and organization of Marchantia polymorpha chloroplast genome. II. Gene organization of the large single copy region from rps'12 to atpB. J. Mol. Biol. 203, 299-331. van Kalken, C.K., Broxterman, H.J., Pinedo, H.M., Feller, N., Dekker, H., Lankelma, J., & Giaccone, G. (1993). Cortisol is transported by the multidrug resistance gene product P-glycoprotein. Br. J. Cancer 67, 284-289. Wemmie, J.A., Wu, A.L., Harshman, K.D., Parker, C.S., & Moye-Rowley, W.S. (1994). Transcriptional activation mediated by the yeast AP-1 protein is required for normal cadmium tolerance. J. Biol. Chem. 269, 14690-14697. Wu, A., Wemmie, J.A., Edgington, N.P., Goebl, M., Guevara, J.L., & Moye-Rowley, W.S. (1993). Yeast bZip proteins mediate pleiotropic drug and metal resistance. J. Biol. Chem. 268, 18850-18858. Xue, C.B., Caldwell, G.A., Becker, J.M., & Naider, F (1989). Total synthesis of the lipopeptide a-mating f2iC\ox of Saccharomyces cerevisiae. Biochem. Biophys. Res. Commun. 162,253-257. Yang, C.H., DePinho, S.G., Greenberger, L.M., Arceci, R., & Horwitz, S.B. (1989). Progesterone interacts with P-glycoprotein in multidrug-resistant cells and in the endometrium of gravid uterus. J. Biol. Chem. 264, 782-788. Zinser, E., Paltauf, F , & Daum, G. (1993). Sterol composition of yeast organelle membranes and subcellular distribution of enzymes involved in sterol metabotism. J. Bacteriol. 175, 2853-2858.

THE APICAL SORTING OF GLYCOSYLPHOSPHATIDYLINOSITOLLINKED PROTEINS

Michael P. Lisanti/ ZhaoLan Tang, Philipp E. Scherer, and Massimo Sargiacomo

I. II. III. IV. V. VI.

Introduction GPI Biosynthesis and Transfer to Protein Polarized Sorting of GPI-Linked Proteins in Epithelia Caveolar Clustering of GPI-Linked Proteins A Model for Caveolar Clustering and Assembly Concluding Remarks Acknowledgments References

Membrane Protein Transport Volume 2, pages 97-110 Copyright © 1995 by JAI Press Inc. All rights of reproduction in any form reserved. ISBN: 1-55938-983-4

97

98 99 100 103 104 106 106 107

98

M.P. LISANTI, Z. TANG, RE. SCHERER, and M. SARGIACOMO

I. INTRODUCTION As an alternative to hydrophobic amino acids, many cells, from yeast to human, use a glycosylated form of phosphatidylinositol to anchor proteins to the cell surface (Low, 1989; Cross, 1990; Ferguson, 1991; Hooper, 1992). Several terms have been coined to describe this anchoring mechanism. They include: GPI (glycosylphosphatidylinositol) anchor-

Inhibitors PMSF DFP

ManN

0-11

FGIc Ampho

EthN-P-6 Man a 1,2 Man a 1,6 Man a 1,4 GlcN Inos-P- DAG

A/C/H Mutant Connplementation Groups Figure 1. Inhibitors of GPI biosynthesis and genetic defects in GPI anchoring correspond to similar points along the GPI-synthesis pathway. Inhibitors: PMSF inhibits the addition of the terminal ethanolamine-P group, but paradoxically it did not prevent GPI anchoring of the protein product (Masterson and Fergsuson, 1991). ManN prevents the formation of a 1,2 linkages and consequently blocks GPI anchoring of the protein product, resulting in secretion (Lisanti et al., 1991). FGIc inhibits the synthesis of dolichol-P-mannose, the donor for all three mannose residues within the GPI core, but is also a potent inhibitor of protein synthesis (Schwarz et al., 1989). 0 - 1 1 , a myristic acid analog, is selectively incorporated into bloodstream trypanosomal GPI, as they use exclusively myristate as their GPI-fatty acid component (Doering et al., 1991). Incorporation of 0-11 results in selective toxicity (vacuolated, nondividing, nonmotile trypanosomes), whereas mammalian cells are unaffected. DFP and ampho act in a manner similar to that of PMSF and FGIc, respectively, but are only active in a cell-free system. Of the known GPI inhibitors, only

(continued)

Calveolar Clustering and GPi-Linked Proteins

99

Figure 1. (continued) ManN prevents GPI anchoring in living cells without interfering with the synthesis of the protein product. Additionally, ManN is the only inhibitor that has been shown to work in both trypanosomal and mammalian cell systems. All other inhibitors were evaluated solely for their effects on trypanosomal GPI. A^ufanfs; Thy-1 (aCPI-linked protein)-negativeT-cell mutants defective in GPI biosynthesis have been assigned to six complementation groups (designated A, B, C, E, F, H). These defects map to points along the GPI synthesis pathway that are similar to where GPI inhibitors are thought to act. This high correspondence suggests that these steps represent key control points in the synthesis pathway and that inhibitors can be used to mimic the mutant phenotype. Such mutants have recently been used in an expression cloning strategy to identify the enzymes involved in GPI biosynthesis (Inoue et al., 1993; Miyata et al., 1993). Abbreviations: PMSF, phenylmethanesulfonyl fluoride; DFP, diisopropyl fluorophosphate; ManN, mannosamine, 2-amino-2-deoxy-D-mannose; FGIc, 2-fluoro-2-deoxy-D-glucose; ampho, amphomycin; 0 - 1 1 , 10-[propoxy]decanoic acid.

ing; glypiation; PIG (phosphatidylinositol-glycan) tailing; and "greasy foot." Proteins with diverse roles in cellular functioning contain this glycophospholipid tail, from parasite coat proteins (VSG, T. brucei) to hydrolytic enzymes (alkaline phosphatase) and signal transducing receptor molecules (CNTF receptor and CD14-LPS-receptor). Despite this diversity, certain unifying features have emerged from the study of GPI-linked proteins. All GPI anchors studied to date contain a conserved glycan core structure, composed of ethanolamine, phosphate, mannose, glucosamine, and inositol (Figure 1) (Ferguson, 1991), and GPI-linked proteins are found clustered in caveolae, a specialized domain of the plasma membrane (Anderson, 1993). The latter observation may also explain why GPI-linked proteins are selectively transported to the apical surface of polarized epithelial cells, with GPI acting as a dominant apical trafficking signal (Lisanti et al,, 1993; Sargiacomo et al., 1993).

II. GPI BIOSYNTHESIS AND TRANSFER TO PROTEIN GPI is synthesized by sequential glycosylation in a series of discrete steps from the lipid precursor, phosphatidylinositol (Menon et al., 1988; Doering et al., 1990). Donors for these reactions include UDP-GlcNAc,

100

M.P. LISANTI, Z. TANG, P.E. SCHERER, and M. SARCIACOMO

Dol-P-Man, and phosphoethanolamine (Menon and Stevens, 1992). These biosynthetic reactions appear to take place on the cytoplasmic face of the endoplasmic reticulum, necessitating the existence of a "flipase" to transfer the completed precusor to the ER lumen (Vidugiriene and Menon, 1993). In accordance with these studies, candidate enzymes involved in GPI biosynthesis have recently been cloned by complementation of cells defective in various steps of the GPI biosynthetic pathway, and analysis of their cDNAs suggests that they assume a cytoplasmic orientation (Inoue et al., 1993; Miyata et al., 1993). Specific GPI synthesis inhibitors have also been identified that act at key points along the pathway (Figure 1) (Schwarz et al., 1989; Menon et al., 1990; Lisanti et al., 1991; Masterson and Fergsuson, 1991). After completion, GPI precursors are transferred to protein in the lumen of the ER, in a reaction that involves coordinated proteolytic removal of the signal for GPI attachment. GPI addition is directed by a short C-terminal hydrophobic peptide domain that has many properties in common with N-terminal signal sequences (Caras and Weddell, 1989). The GPI addition signal requires two main features: a hydrophobic stretch of suitable length (13 amino acids or greater) and a favorable upstream cleavage/attachment site (co, usually a small amino acid, such as gly, ala, ser, asp, cys, and asn) (Caras et al., 1989; Moran et al., 1991). Amino acids immediately adjacent to the cleavage site at the co + 1 and CO + 2 positions also appear to have specific requirements (Kodukula et al., 1993). The enzyme responsible for this cleavage and attachment, the putative GPI transpeptidase, remains unknown.

III. POLARIZED SORTING OF GPI-LINKED PROTEINS IN EPITHELIA Epithelial cells line the body cavities and form a protective barrier that regulates the passage of ions and macromolecules from the outside world (e.g., lumen of the intestine or renal tubules) to the circulation. To perform this selective barrier transport function, epithelial cells are equipped with two distinct plasma membrane domains, termed apical (luminal) and basolateral (serosal), that are separated by a tight seal, the zonula occludens or tight junction. The sorting of different classes of plasma membrane proteins to these distinct domains is therefore critical for maintaining bodily homeostasis. An important clue to the mechanism by which polarized sorting takes place comes from the study of GPI-linked proteins. Because both endo-

Calveolar Clustering and CPI-Linked Proteins

101

Table 1. Recombinant Expression of GPI-Linked Proteins in Polarized MDCK Cells Protein Ectodomain

Anchoring Mechanism

Polarity

DAF, FLAP, or Thy-1

Native, GPI-linked anchorless, secreted

Apical Apical

hGH or bacterial endoglucanase E

Native, secreted fusion protein, GFI-linked

Nonpolarized Apical

gD-lorVSVG

Native, transmembrane anchorless, secreted fusion protein, GFI-linked

Basolateral Basolateral Apical

N-CAM

Transmembrane; Id 180 kDa transmembrane; sd 140 kDa GPI-linked; ssd 120 kDa

Basolateral Basolateral Apical

Note: See Soole et al. (1992) for bacterial endoglucanase E and Lisanti and Rodriguez-Boulan (1992) for other specific references,

genous and transfected GPI-linked proteins are selectively targeted to the apical surface of renal and intestinal epithelial cells (Lisanti et al, 1988, 1989,1990; Brown et al, 1989; Powell et al., 1990; Wilson et al, 1990), it has been suggested that GPI may act as a signal for apical transport. In dramatic support of this hypothesis, 1) recombinant transfer of GPI to a basolateral or unsorted antigen leads to apical localization (see Table 1) (Brown et al., 1989; Lisanti et al., 1989,1991; Soole et al., 1992); and 2) inhibition of GPI biosynthesis with mannosamine prevents apical sorting and leads to unpolarized secretion of a non-GPI-linked protein product (Lisanti et al., 1991). Taken together, these data indicate that GPI anchoring can actively target the attached protein at distinct domains of the plasma membrane. This targeting event appears to be independent of tight junctions, inasmuch as GPI-linked proteins are sorted to the free surface (equivalent of the apical surface) and excluded from the attached surface (equivalent of the basolateral surface) of epithelial cells prevented from forming cell-cell contacts (Lisanti et al., 1990). Because glycosphingolipids are also selectively apically sorted in epithelia (van Meer et al., 1987; van Meer and Simons, 1988), one hypothesis is that both glycolipids and GPI-linked proteins are sorted by co-clustering in the trans-Golgi network (TGN) (Lisanti and RodriguezBoulan, 1990; Lisanti etal., 1990). However, because neither glycolipids

102

M.P. LISANTI, Z. TANG, P.E. SCHERER, and M. SARGIACOMO

Table 2. Comparison of the Polarity of Recombinant GPI-Anchored Proteins, Glycosphingolipids, and Transmembrane Proteins Expressed inMDCKandFRTGells Cell Surface Distribution Molecule

Anchoring Mechanism

MDCK

FRT

DAF gD-1 gD-l-DAF

Native, GPI-linked Transmembrane Fusion, GPI-linked

Apical Basolateral Apical

Nonpolarized Basolateral Basolateral

Glucosyl-ceramide

Ceramide moiety

Apical

Basolateral

Influenza HA DPP IV VSV G protein L-CAM Na-KATPase TfR

Transmembrane Transmembrane Transmembrane Transmembrane Transmembrane Transmembrane

Apical Apical Basolateral Basolateral Basolateral Basolateral

Apical Apical Basolateral Basolateral Basolateral Basolateral

Note: See text for specific references.

nor GPI-linked proteins span the membrane, it is more likely that they are recognized by a lectin-like transmembrane adaptor molecule that transmits addressing signals to the cytoplasm (Lisanti and RodriguezBoulan, 1992; Hannan et al., 1993). In support of both these hypotheses, GPI-linked proteins become insoluble in non-ionic detergents (Hooper and Turner, 1988), such as Triton X-100, during polarized sorting to the apical surface (Brown and Rose, 1992), implying the formation of membrane microdomains. Insolubility appears to be selective for GPIlinked proteins, as many apical and basolateral transmembrane antigens remain Triton-soluble (reviewed in Rodriguez-Boulan and Powell, 1992). This sorting event also appears to be independent of the polarized sorting of transmembrane proteins, as two distinct epithelial cell lines have been identified that selectively missort GPI-linked proteins but correctly sort both apical and basolateral transmembrane antigens (Table 2) (Lisanti et al., 1990; Zurzolo et al., 1990; Hannan et al., 1993; Sargiacomo et al., 1993; Zurzolo et al, 1993). These observations suggest that GPI-linked proteins may be transported by a distinct subclass of carrier vesicles (most likely caveolae), separated from the bulk of transmembrane protein traffic.

Calveolar Clustering and GPi-Linked Proteins Table 3.

103

Properties of Purified MDCK Caveolae

I. Appearance A. Size B. Shape Transmission Rotary shadowing II. Composition A. Caveolar protein Total Cell surface Newly synthesized B. Identified protein components Caveolin GPI-linked protein Hetero-trimeric G proteins Gs Gi,2 Gi,3 Gq/11 Nonreceptor tyrosine kinases and their substrates c-Yes Annexin II

50-100 nm in diameter Curved or flat membrane bilayers; granular in some sections Collections of 4-6 nm particles arranged in concentric rings up to 50-100 nm in diameter % total cell protein; enrichments relative to total cell-lysates 0.05% 0.5-0.8%; 15-20-fold enriched 0.005%; 5-7.5-fold enriched 2,000-fold enriched 240-fold enriched 25-fold enriched 75-fold enriched 175-fold enriched 60-fold enriched

50-fold enriched 400-fold enriched

IV. CAVEOLAR CLUSTERING OF GPI-LINKED PROTEINS After transport to the cell surface, GPI-linked proteins are selectively excluded from clathrin-coated pits (Bretscher et al., 1980) but are found clustered in another membrane micro-invagination—^known as caveolae or plasmalemmal vesicles (Anderson et al., 1992; Anderson, 1993). Caveolar clustering of GPI-linked proteins is cholesterol dependent, as cholesterol depletion or incubation with cholesterol-binding antibiotics disrupts caveolar ultrastructure and disperses GPI-linked proteins (Rothberg et al.,1990, 1992). Similar to GPI-linked proteins, glycosphingolipids are also found clustered in plasmalemmal caveolae (Montesano et al., 1982). As caveolae concentrate both glycolipids and GPI-linked proteins (apically sorted molecules), we have suggested that intracellular

104

M.P. LISANTI, Z. TANG, RE. SCHERER, and M. SARGIACOMO

caveolae located within the TGN could be responsible for the apical recognition and transport of GPI-linked proteins during their biogenesis (Lisanti et al., 1990). Several lines of evidence now support this hypothesis. We have recently purified caveolae from cultured epithelial cells based on their selective insolubility in Triton and we find that a model GPI-linked protein and caveolin, a caveolar marker protein, are dramatically enriched over 150-fold, whereas the bulk of cellular proteins are excluded (Table III; Sargiacomo et al., 1993). These purified caveolae also contain cytoplasmically oriented signahng molecules, such as heterotrimeric G-proteins, that have recently been shown to play a dual role in both regulating the budding of transport vesicles and in cell surface signal transduction events (Sargiacomo et al., 1993). As caveolin is an integral membrane protein and undergoes phosphorylation on serine and tyrosine residues (Glenney, 1989; Glenney and Zokas, 1989; Sargiacomo et al., 1993), we have proposed that caveolin could play the role of an adaptor molecule in coupling GPI-linked proteins to cytoplasmic sorting machinery. Perhaps other transmembrane components of caveolae could play this role as well. In addition, we find that caveohn exists as a hetero-oligomeric protein complex with other integral membrane proteins—forming putative caveolar assembly units (Lisanti et al., 1993). These caveolin hetero-oligomers interact with a model GPI-linked protein in a pH- and cholesterol-dependent manner. As the TGN is shghtly acidic (Anderson and Pathak, 1985), GPI-linked proteins would be expected to interact or "cluster" with caveolin hetero-oligomers at the level of the Golgi—explaining the Triton insolubility of GPI-linked proteins during apical transport. In further support of the hypothesis that caveolar clustering underlies the apical sorting of GPI-linked proteins, two independent cell lines that selectively missort GPI-linked proteins either fail to express caveolin or form incomplete caveolin hetero-oligomers (Lisanti et al., 1993; Sargiacomo et al., 1993). It remains to be elucidated which member of the caveolin hetero-oligomer recognizes the GPI anchor structure. Such recognition could be regulated through caveolin phosphorylation.

V. A MODEL FOR CAVEOLAR CLUSTERING AND ASSEMBLY As caveolin exists as a hetero-oligomeric protein complex with other integral membrane proteins, these hetero-oligomers may represent the assembly units of caveolae (Lisanti et al., 1993). The close packing of

caveolin hetero-oligomers

kinase domain regulatory domain

J-P receptor tyrosine C kinases

Figure 2. The protein exclusion-lipid inclusion model of caveolar assembly. We propose that caveolae are constructed from integral membrane protein complexes, i.e., caveolin hetero-oligomers. These hetero-oligomers self-associate and cluster with specific lipids (GPI, glyco-sphingolipids (GSLs), and cholesterol) while excluding other noncaveolar transmembrane proteins. During or after assembly, lipid-modified (i.e., nonreceptor tyrosine kinase, hetero-trimeric G-proteins, etc.) or lipld-binding (i.e., annexin II) signaling molecules "plug into" these protein-poor, lipid-rich membrane domains. Caveolae may also assemble around cell surface receptors to transduce signals to the cytoplasm.

105

106

M.P. LISANTI, Z. TANG, RE. SCHERER, and M. SARGIACOMO

these integral membrane protein complexes should then exclude other transmembrane proteins, while including specific lipid components (GPI, glycosphingolipids, and cholesterol) that promote their self-association. Lipid-modified cytoplasmic signaling molecules (nonreceptor tyrosine kinase, hetero-trimeric G-proteins, and others) could then insert or "plug into" these protein-poor, lipid-rich membrane domains (Figure 2). This protein exclusion-lipid inclusion model of caveolae formation also has implications for cell surface signal transduction, as the P-adrenergic receptor and the insulin receptor have been localized to plasmalemmal caveolae (Goldberg et al., 1987; Dupree et al., 1993). In this regard, receptors may not "enter" preformed caveolae, but caveolae may "assemble around" cell surface receptors, allowing these receptors to associate with cytoplasmic signaling molecules. In accordance with this model, the P-adrenergic receptor is lipid-modified (palmitoylated), and this modification is essential for coupling the receptor to G-proteins (O'Dowd et al., 1989).

VI. CONCLUDING REMARKS GPI-linked proteins are selectively transported to the apical surface of polarized epithelial cells (Lisanti and Rodriguez-Boulan, 1990), whereas plasmalemmal caveolae are present on both epithelial surfaces (Rothberg et al., 1992; Dupree et al, 1993). A solution to this apparent contradiction comes from studying the biogenesis of caveolae. Our recent results suggest that apically destined caveolae form intracellularly within the Golgi complex, whereas basolateral caveolae form at the level of the basolateral membrane (Lisanti et al, 1993). As apical proteins are sorted intracellularly within the TGN and at the level of the basolateral endosome (Rodriguez-Boulan and Powell, 1992), this scheme is consistent with the known biogenetic pathways followed by apical membrane proteins, including GPI-linked proteins.

ACKNOWLEDGMENTS We thank Harvey Lodish for his patience and advice, and members of the Buratowski, Fink, Lodish, Matsudaira, and Young labs for thoughtful discussions. This work was supported in part by a grant from the W.M. Keck Foundation to the Whitehead Fellows Program (M.RL.) and an NIH FIRST

Calveolar Clustering and GPi-Linked Proteins

107

Award (GM-50443) to M.P.L. P. E. Scherer is funded by a Long-term Fellowship from the European Molecular Biology Organization.

NOTE 1. Corresponding author

REFERENCES Anderson, R.G.W. (1993). Plasmalemmal caveolae and GPI-anchored membrane proteins. Curr. Opin. Cell Biol. 5, 647-652. Anderson, R.G.W. & Pathak, R.U. (1985). Vesicles and cistemae in the trans-Golgi apparatus of human fibroblasts are acidic compartments. Cell 40, 635-643. Anderson, R.G.W., Kamen, B.A., Rothberg, K.G., & Lacey, S.W. (1992). Potocytosis: sequestration and transport of small molecules by caveolae. Science 255,410-411. Bretscher, M.S., Thomson, J.N., & Pearse, B.M.F. (1980). Coated pits act as molecular filters. Proc. Natl. Acad. Sci. USA 77,4156-4159. Brown, D. & Rose, J.K. (1992). Sorting of GPI-anchored proteins to glycolipid-enriched membrane subdomains during transport to the apical cell surface. Cell 68, 533544. Brown, D.A., Crise, B., & Rose, J.K. (1989). Mechanism of membrane anchoring affects polarized expression of two proteins in MDCK cells. Science 245, 1499-1501. Caras, I.W. & Weddell, G.N. (1989). Signal peptide for protein secretion directing glycophospholipid membrane anchor attachment. Science 243, 1196-1198. Caras, I.W, Weddell, G.N., & Williams, S.R. (1989). Analysis of the signal for attachment of a glycophospholipid membrane anchor. J. Cell Biol. 108, 1387-1396. Cross, G.A.M. (1990). Glycolipid anchoring of plasma membrane proteins. Annu. Rev. Cell Biol. 6, 1-39. Doering, T.L., Masterson, W.J., Hart, G W , & Englund, PT. (1990). Biosynthesis of the GPI membrane anchors. J. Biol. Chem. 265, 611-614. Doering, T.L., Raper, J., Buxbaum, L.U., Adams, S.R, Gordon, J.I., Hart, G W , & Englund, P.T. (1991). An analogue of myrystic acid with selective toxicity for African trypanosomes. Science. 252,1851-1854. Dupree, P, Parton, R.G., Raposo, G , Kurzchalia, TV., & Simons, K. (1993). Caveolae and sorting of the trans-Golgi network of epithelial cells. EMBO J. 12,1597-1605. Ferguson, M.A.J. (1991). Lipid anchors on membrane proteins. Curr. Opin. Struct. Biol. 1,522-529. Glenney, J.R. (1989). Tyrosine phosphorylation of a 22 kD protein is correlated with transformation with Rous sarcoma virus. J. Biol. Chem. 264, 20163-20166. Glenney, J.R. & Zokas, L. (1989). Novel tyrosine kinase substrates from Rous sarcoma virus transformed cells are present in the membrane cytoskeleton. J. Cell Biol. 108, 2401-2408. Goldberg, R.I., Smith, R.M., & Jarett, L. (1987). Insulin and 0C2-macroglobuHn-methylamine undergo endocytosis by different mechanisms in rat adipocytes. I. Comparison of cell surface events. J. Cell. Physiol. 133, 203-212.

108

M.P. LISANTI, Z. TANG, P.E. SCHERER, and M. SARGIACOMO

Hannan, L.A., Lisanti, M.R, Rodriguez-Boulan, E., & Edidin, M. (1993). Correctlysorted molecules of a GPI-anchored protein are clustered and immobile when they arrive at the apical surface of MDCK cells. J. CeD Biol. 120, 353-358. Hooper, N.M. (1992). More than just a membrane anchor. Curr. Biol. 2, 617-619. Hooper, N.M. & Turner, A.J. (1988). Ectoenzymes of the kidney microvillar membrane: differential solubilization by detergents can predict a GPI membrane anchor. Biochem. J. 250, 865-869. Inoue, N., Kinoshita, T, Orii, T, & Takeda, J. (1993). Cloning of a human gene, PIG-F, a component of GPI anchor biosynthesis, by a novel expression cloning strategy. J. Biol. Chem. 268, 6882-6885. Kodukula, K., Gerber, L., Amthauer, R., Brink, L., & Udenfriend, S. (1993). Biosynthesis of GPI-anchored membrane proteins in intact cells: specific amino acid requirements adjacent to the site of cleavage and GPI-attachement. J. Cell Biol. 120, 657-664. Lisanti, M.P. & Rodriguez-Boulan, E. (1990). Glycophospholipid membrane anchoring provides clues to the mechanism of protein sorting in polarized epithelial cells. Trends Biochem. Sci. 15, 113-118. Lisanti, M.P & Rodriguez-Boulan, E. (1992). Polarized sorting of GPI-linked proteins in epithelia and membrane microdomains. In: GPI Membrane Anchors (de Almeida, M.L.C., ed.), pp. 170-196. Academic Press, New York. Lisanti, M.R, Sargiacomo, M., Graeve, L., Sahiel, A.R., & Rodriguez-Boulan, E. (1988). Polarized apical distribution of glycosyl-phosphatidylinositol anchored proteins in a renal epithelial cell line. Proc. Natl. Acad. Sci. USA 85,9557-9561. Lisanti, M.R, Caras, I.W., Davitz, M.A., & Rodriguez-Boulan, E. (1989). A glycophospholipid membrane anchor acts as an apical targeting signal in polarized epithelial cells. J. Cell Biol. 109, 2145-2156. Lisanti, M.R, Caras, I.W., Gilbert, T., Hanzel, D., & Rodriguez-Boulan, E. (1990a). Vectorial apical delivery and slow endocytosis of a glycolipid-anchored fusion protein in transfected MDCK cells. Proc. Natl. Acad. Sci. USA 87,7419-7423. Lisanti, M.R, Le Bivic, A., Saltiel, A.R., & Rodriguez-Boulan, E. (1990b). Preferred apical distribution of glycosyl-phosphatidylinositol (GPI) anchored proteins: a highly conserved feature of the polarized epithelial cell phenotype. J. Membr. Biol. 113,155-167. Lisanti, M.R, Rodriguez-Boulan, E., & Saltiel, A.R. (1990c). Emerging functional roles for the glycosyl-phosphatidylinositol (GPI) membrane protein anchor. J. Membr. Biol. 117, 1-10. Lisanti, M.R, Caras, I.W., & Rodriguez-Boulan, E. (1991a). Fusion proteins containing a minimal GPI-attachment signal are apically expressed in transfected MDCK cells. J. Cell Sci. 99, 637-640. Lisanti, M.R, Field, M.C., Caras, I.W., Menon, A.K., & Rodriguez-Boulan, E. (1991b). Mannosamine, a novel inhibitor of glycosyl-phosphatidylinositol incorporation into proteins. EMBO J. 10, 1969-1977. Lisanti, M.R, Tang, Z.-L., & Sargiacomo, M. (1993). Caveolin forms a hetero-oligomeric protein complex that interacts with an apical GPI-linked protein: implications for the biogenesis of caveolae. J. Cell Biol. 123, 595-604. Low, M.G. (1989). Glycosyl-phosphatidylinositol: a versatile anchor for cell surface proteins. FASEB J. 3, 1600-1608.

Calveolar Clustering and GPi-Linked Proteins

109

Masterson, W.J. & Ferguson, M.A.J. (1991). Phenylmethanesulfonyl fluoride inhibits GPI anchor biosynthesis in African trypanosomes. EMBO J. 10,2041-2045. Menon, A. & Stevens, V.L. (1992). Phosphatidylethanolamine is the donor of the ethanolamine residue linking a GPI-anchor to protein. J. Biol. Chem. 267,1527715280. Menon, A.K., Mayor, S., Ferguson, M.A.J., Duszenko, M., & Cross, G.A.M. (1988). Candidate glycophospholipid precursor for the glycosyl-phosphatidylinositol membrane anchor of Trypanosoma brucei variant surface glycoproteins. J. Biol. Chem. 263, 1970-1977. Menon, A.K., Mayor, S., & Schwarz, R.T. (1990). Biosynthesis of GPI lipids in Trypanosoma brucei: involvement of mannosyl-phosphoryldolichol as the mannose donor. EMBO J. 9,4249-4258. Miyata, T, Takeda, J., lida, Y, Yamada, N., Inoue, N., Takahashi, M., Maeda, K., Kitani, T., & Kinoshita, T. (1993). The cloning of PIG-A, a component in the early step of GPI-anchor biosynthesis. Science 259, 1318-1319. Montesano, R., Roth, J., Robert, A., & Orci, L. (1982). Non-coated membrane invaginations are involved in binding and internalization of cholera and tetanus toxins. Nature 296, 651-653. Moran, P., Raab, H., Kohr, W.J., & Caras, I.W. (1991). Glycophospholipid membrane anchor attachment: molecular analysis of the cleavage/attachment site. J. Biol. Chem. 266, 1250-1257. O'Dowd, B.F, Hnatowich, M., Caron, M.G., Lefkowitz, R.J., & Bouvier, M. (1989). Palmitoylation of the human beta-2 adrenergic receptor: mutation of cys-341 in the carboxyl tail leads to an uncoupled non-palmitoylated form of the receptor. J. Biol. Chem. 264, 7564-7569. Powell, S., Lisanti, M.P., & Rodriguez-Boulan, E. (1990). Thy-1 expresses two signals for apical localization in epithelial cells. Am. J. Physiol. 260, C715-C720. Rodriguez-Boulan, E. & Powell, S.K. (1992). Polarity of epithelial and neuronal cells. Annu. Rev. Cell Biol. 8, 395-427. Rothberg, K.G., Ying, Y, Kamen, B.A., & Anderson, R.G.W. (1990). Cholesterol controls the clustering of the glycophopholipid-anchored membrane receptor for 5-methyltetrahydrofolate. J. Cell Biol. I l l , 2931-2938. Rothberg, K.G., Heuser, J.E., Donzell, W.C, Ying, Y, Glenney, J.R., & Anderson, R.G.W. (1992). Caveolin, a protein component of caveolae membrane coats. Cell. 68, 673-682. Sargiacomo, M., Sudol, M., Tang, Z.-L., & Lisanti, M.P. (1993). Signal transducing molecules and GPI-linked proteins form a caveolin-rich insoluble complex in MDCK cells. J. Cell Biol. 122, 789-807. Schwarz, R.T., Mayor, S., Menon, A.K., & Cross, G.A.M. (1989). Biosynthesis of the glycolipid membrane anchor of T brucei variant surface glycoproteins: involvement of Dol-P-Man. Biochem. Soc. Trans. 17,746-748. Soole, K.L., Hazlewood, G., Gilbert, H., & Hirst, B. (1992). A GPI-anchor can target a bacterial enzyme to the apical surface of polarized epithelial cells. Biochem. Soc. Trans. 21,42S. van Meer, G. & Simons, K. (1988). Lipid polarity and sorting in epithelial cells. J. Cell. Biochem. 36, 51-58.

110

M.P. LISANTI, Z. TANG, RE. SCHERER, and M. SARGIACOMO

van Meer, G., Stelzer, E.H.K., Wijnaendts-van Resandt, R.W., & Simons, K. (1987). Sorting of sphingolipids in epithelial (Madin-Darby canine kidney) cells. J. Cell Biol. 105,1623-1635. Vidugiriene, J. & Menon, A.K. (1993). Early lipid intermediates in GPI-anchor assembly are synthesized in the BR and located in the cytoplasmic leaflet of the ER membrane bUayer. J. Cell Biol. 121, 987-996. Wilson, J.M., Fasel, N., & Kraehenbuhl, J.-R (1990). Polarity of endogenous and exogenous GPI-anchored membrane proteins in MDCK cells. J. Cell Sci. 96, 143-149. Zurzolo, C, Nitsch, L., Rodriguez-Boulan, E., & Lisanti, M.P. (1990). Reversed polarity of GPI-anchored proteins in a polarized thyroid cell Une. J. Cell Biol. Ill, 327a. Zurzolo, C , Lisanti, M.R, Caras, I.W., Nitsch, L., & Rodriguez-Boulan, E. (1993). GPI-anchored proteins are preferentially targeted to the basolateral domain in Fischer rat thyroid epithelial cells. J. Cell Biol. 121, 1031-1039.

CAVEOLAE: PORTALS FOR TRANSMEMBRANE SIGNALING AND CELLULAR TRANSPORT

Michael P. Lisanti/ ZhaoLan Tang, and Massimo Sargiacomo

I. II. III. IV. V.

Introduction Caveolae: Signaling Organelles Caveolar Transport Processes Role in Human Disease Concluding Remarks Acknowledgments References VI. Added in Proofs

Ill 113 115 117 118 119 119 121

I. INTRODUCTION Caveolae, also known as plasmalemmal vesicles, are 50-100-nm flaskshaped vesicular organelles located at the plasma membrane and perhaps Membrane Protein Transport Volume 2, pages 111-122 Copyright © 1995 by JAI Press Inc. All rights of reproduction in any form reserved. ISBN: 1-55938-983-4

111

112

MICHAEL P. LISANTI, ZHAOLAN TANG, and MASSIMO SARGIACOMO

Figure 1. Purified caveolae. See Sargiacomo et al. (1993).

within the Golgi complex (Figure 1). Consistent with their locahzation, evidence has been presented that caveolae represent a specific subclass of vesicular carriers that shuttle between the plasma membrane and the trans-Golgi network, along both endocytic and exocytic routes. Their distinctive cytoplasmic appearance, described as striped bipolar structures, makes them stand out from the rest of the plasma membrane in rapid-freeze, deep-etch micrographs (Peters et al., 1985). Originally described by Yamada (1955) and Bruns and Palade (1968), they were first implicated in the rapid transcytosis (15-60 sec) of small tracer molecules and albumin across capillary endothelial cells. However, their role as signaling organelles has gone unappreciated until recently.

Caveolae, Transmembrane Signaling, and Cellular Transport

113

Here we suggest that caveolae play a central role in normal and pathogenic signaling events and that they represent a focal point for a variety of human diseases. Perhaps, like clathrin-coated vesicles, different subpopulations of caveolae exist to perform these diverse functions, each with its distinct accessory proteins that could be cell type specific.

il. CAVEOLAE: SIGNALING ORGANELLES As caveolae concentrate a large number of specific signaling molecules, they appear to play the role of a molecular switchboard in directing a variety of different transmembrane signaling events. These include proteins that are critical for maintaining calcium homeostasis, G-protein receptor signaling, and members of the Src family of nonreceptor tyrosine kinases and their substrates (Fujimoto, 1993; Fujimoto et al., 1993; Sargiacomo et al., 1993; summarized in Table 1). Many of these molecules are lipid modified (myristoylated, palmitoylated, etc.), suggesting that such modifications could function to target signaling molecules to caveolae. In support of this hypothesis, v-Src lacking N-terminal myristate fails to phosphorylate caveolin, a caveolar marker protein, and prevents cellular transformation (Glenney, 1989). Calcium fluxes occur as a general response to a variety of cell surface stimuli and are thought to be regulated by inositol signaling pathways. As the plasma membrane calcium pump (for calcium export) and the IP3 receptor (for calcium import) are concentrated within caveolae (Fujimoto, 1993; Fujimoto et al., 1993), these organelles appear to play a major role in regulating intracellular calcium levels. This role may also extend to excitation-contraction coupling during muscle contraction and calcium-triggered exocytosis, as caveolae form part of the T-tubule network in muscle cells (North et al., 1993) and contain annexin II, a calcium-dependent phospholipid-binding protein involved in both exocytic and endocytic fusion events (Sargiacomo et al., 1993). Caveolar regulation of calcium homeostasis could also explain the observation that targeted disruption of the o-Src proto-oncogene in mice leads to defects in Ca^"^ metabolism, such as osteopetrosis (Soriano et al., 1991). Caveolin was first identified as a major v-Src substrate and later as a caveolar marker protein (Glenney, 1989; Glenney and Zokas, 1989; Rothberg et al., 1992). Now known to be an integral membrane phosphoprotein, caveolin could play the role of an adaptor molecule, linking other caveolar components (GPI-linked protein, glycolipids, etc.) to cytoplasmic signaling molecules (Sargiacomo et al., 1993). As c-Yes (a

Table 1. Functions of Caveolar Membrane Domains Function Signal Transduction Ca mobilization/homeostasis G-protein coupled signaling Nonreceptor protein tyrosine kinases and their substrates Transport Protein/peptide transport

Clustering of lipids

Uptake of small molecules via potocytosis Uptake of nucleic acids Disease Processes Cholera (intractable diarrhea)

Muscular dystrophy Malaria {P. vivax) Diabetes Scrapie Cellular transformation

Molecules Concentrated in Caveolae IP3-receptor-Ca^"' influx, Ca^"^-ATPaseCa^"^ efflux p-Adrenergic receptor, hetero-trimeric G proteins (G3,G.2, G. 3, Gq/ii) c-Yes, v-Src substrates (caveolin, annexin II) Endocytosis/transcytosis of albumin, and protein tracers (ferritin, HRP, myoglobin, and small heme containing peptides) across endothelia Glyco-lipids (Gj^^ and other gangUosides, via cholera and tetanus toxin binding) GPI-linked proteins (e.g. folate receptor, 5' nucleotidase, alkaline phosphatase, etc.) Cholesterol (3-P-hydroxy sterols via filipin/nystatin binding) Folate, adenosine, others, via probenecidsensitive anion transporter Naked plasmid DNA Cholera toxin B subunit-glycolipid binding A subunit-ADP ribosylates G^ Dystrophin; associated with T-tubule network "Schuffner's dots" (malarial stippling of RBCs) corresponds to increased number of caveolae in infected cells Insulin receptor clustering; ~10-fold increase in the number of caveolae during adipocyte differentiation GPI-linked prion protein (PrP) Caveolin, originally identified as the major v-Src substrate in RSVtransformed cells

114

Caveolae, Transmembrane Signaling, and Cellular Transport

115

Src-like nonreceptor tyrosine kinase) and other v-Src substrates (annexin II) are concentrated in caveolae (Sargiacomo et al., 1993), modulation of caveolae by v-Src phosphorylation could explain its transforming properties. In direct support of this hypothesis, activation of G^, a caveolar component (Sargiacomo et al, 1993), dramatically increases the tumorigenicity of v-Src-transformed cells (Gottesman et al., 1984). Similarly, an activating mutant of G.2, another caveolar component, leads to cellular transformation (Pace et al, 1991). In addition to being the probable target of v-Src transformation, caveolae may also represent the organellar site where G-protein coupled receptors go to signal. In support of this proposal, the P-adrenergic receptor, a G^-coupled receptor, and a variety of hetero-trimeric G-proteins are all concentrated within caveolae (Dupree et al., 1993; Sargiacomo et al., 1993). In this regard, we envision that many G-proteincoupled receptors will be found to have a regulated affiUation (Ugandinduced) with caveolae. Because C-terminal palmitoylation of a region of the p-adrenergic receptor is critical for G-protein-receptor coupling, and this site is conserved among most receptors of this class (O'Dowd et al., 1989), we propose that this simple Upid modification could serve as a regulatable signal to target this class of receptors to caveolae. A prediction of this hypothesis is that G-protein effector molecules, such as adenylate cyclase, should also be concentrated in caveolae. In support of this proposal, overexpression of c-Src dramatically enhances p-adrenergic receptor-induced accumulation of cAMP (Bushman et al., 1990). Finally, "cross-talk" observed between these "different" signaling pathways could easily be explained by their close physical proximity (colocalization) within caveolae. III. CAVEOLAR TRANSPORT PROCESSES The role of caveolae in the transcellular transport of protein molecules was first studied in endothelial cells. Palade and co-workers elegantly demonstrated that a series of both large and small fluid-phase tracer molecules (small heme-peptides, myoglobin, ferritin, horseradish peroxidase, dextrans, and glycogen) were rapidly transported across capillary endothelial cells via caveolae (plasmalemmal vesicles) within 15-60 sec (Simonescu et al., 1975). Later, gold-tagged serum albumin was shown to undergo specific receptor-mediated transcytosis in a similar manner (Ghitescu et al., 1986). Inasmuch as serum albumin is a carrier for steroid hormones, fatty acids, and thyroid hormones, caveolae may

116

MICHAEL P. LISANTI, ZHAOLAN TANG, and MASSIMO SARGIACOMO

also play a major role in the transport and processing of these circulating substances. Since the capillary network of the nervous system forms the so-called blood-brain barrier, these endothelial transport studies have implications for designing carrier molecules to allow transport of specific drug molecules across this barrier. In addition to fluid-phase transport, caveolae specifically concentrate certain classes of lipid molecules. These include cholesterol, sphingolipids, and GPI-linked molecules. The identification of caveolae as cholesterol-rich domains was first made with cholesterol-binding antibiotics, such as filipin and nystatin, as probes that specifically recognize the 3-p-hydroxy group of sterols (Orci etal., 1981; Rothbergetal., 1992). Caveolar clustering of sphingo-lipids was detected via the specific binding and internalization of ganglioside-binding bacterial toxins (cholera and tetanus) (Montesano et al., 1982). More importantly, however, caveolae provide a vehicle for these toxins to invade the cell. In light of the signaling role of caveolae, it is interesting to note that sphingo-lipids may also serve as the precursors for the generation of novel second messengers, such as ceramide (Raines et al., 1993). These results suggest that caveolae may be specialized for the transport of cholesterol and sphingo-lipids, while excluding the majority of transmembrane plasma membrane molecules (Lisanti et al., 1993; Sargiacomo et al., 1993). GPI-linked molecules represent a diverse group of cell surface proteins and glyco-conjugates. Certain members of this group are hydrolytic enzymes (alkaline phosphatase), whereas others are receptors for small molecules (folate receptor) or growth factors (CNTF receptor; CD 14, LPS receptor) involved in transmembrane signaling events (Anderson et al., 1992; Hooper, 1992). Because neither glyco-lipids nor GPI-linked proteins span the membrane, adaptor molecules, such as caveolin, must transmit signals from these lipid-linked molecules to the cytoplasm. In addition, another class of GPI-linked molecules is not protein bound and has been implicated in insulin-signaling events as an intracellular second messenger (Saltiel and Cuatrecasas, 1988). These results fit well with the observations that caveolae are signaling organelles and contain a probenecid-sensitive anion transporter for the uptake of small molecules (for example, folate or insulin-generated second messengers; this process has been termed "potocytosis"; Kamen et al., 1991). These studies also have implications for understanding the sorting of caveolar components in polarized epitheUal cells. Because both glyco-lipids and GPIlinked proteins are concentrated in caveolae and both are selectively

Caveolae, Transmembrane Signaling, and Cellular Transport

117

transported to the apical surface of epithelial cells, caveolae may form intracellularly—at the level of the ER or Golgi—and first undergo vectorial transport to the apical surface (Lisanti et al., 1993). The ability of caveolae to sample the extracellular environment is not limited to the uptake of lipid and protein molecules, but also extends to nucleic acids. Because naked plasmid DNA is specifically concentrated and internalized by caveolae (Wolff et al., 1992), they may represent the organellar site responsible for the uptake and transmission of extracellular DNA. These findings could have implications for designing new strategies for gene therapy.

IV. ROLE IN HUMAN DISEASE Because caveolae appear to be organelles specialized for transmembrane signaUng and cellular transport, perhaps it is not surprising that they may play a role in a diverse group of human diseases. In cholera and whooping cough, bacterial toxins are generated that recognize cell surface glycolipids and selectively modify hetero-trimeric G-proteins (G^ for cholera toxin; G. for pertussis toxin), thereby hijacking normal signal transduction pathways (Milligan, 1988). Because caveolae concentrate both targets of these toxins (glyco-Upids and G-proteins; Sargiacomo et al., 1993), these toxins are ideally designed to act via caveolae. Caveolae are very abundant in both in fat and muscle cells. During adipocyte differentiation, the number of caveolae increases at least 10-fold, and certain studies indicate that the insulin receptor is localized to these structures (Fan et al., 1983; Goldberg et al., 1987). In muscle cells, dystrophin, a protein that is deficient in Duchenne's muscular dystrophy, colocalizes to the same sarcolemmal domains as caveolae, and both are associated with the T-tubule network (North et al., 1993), which regulates calcium levels during muscle contraction. This localization is not surprising given the role that caveolae play in calcium homeostasis in other cell types. Caveolae may also participate in the formation of neuromuscular synapses, since they are specifically concentrated within the postsynaptic region of the smooth muscle cell's plasma membrane (Wheater et al., 1984). The relevance of their synaptic localization can now be appreciated given their dual roles in signaling and transport. Taken together, these studies suggest that caveolae may be important in the pathogenesis of certain types of diabetes and muscular dystrophies.

118

MICHAEL P. LISANTI, ZHAOLAN TANG, and MASSIMO SARGIACOMO

Scrapie and other transmissible spongiform encephalopathies are thought to be caused by the prion protein (PrP), an endogenous GPIlinked cellular protein (Chesebro and Caughey, 1993). The exact mechanism by which the infectious form of this protein is transmitted and replicates remains controversial. However, the infectious form is protease resistant and insoluble. One possible mechanism for transmission could involve caveolae. Because GPI-linked proteins are concentrated at the cell surface within caveolae, the prion protein could be shed within caveolae, accounting for the size of prion particles (50-100 nm). Infection could then be mediated by the reinsertion of these "shed" caveolae into the plasma membrane of another cell. The isolation of caveolae as Triton-insoluble complexes could account, in part, for the observed insolubility of the infectious prion particles. Caveolae may be important in other human diseases as well, such as malaria, in which malarial stippling of red blood cells (Schuffner's dots) corresponds ultrastructurally to an increased number of caveolae in infected cells (Lanners, 1991). Given the observations that caveolae are signaling organelles (Sargiacomo et al., 1993) and that caveolin is the major v-Src substrate in RSV-transformed cells (Glenney, 1989; Glenney and Zokas, 1989), caveolae could represent a focal point for intervening or preventing cellular transformation. These and future studies on caveolar functioning could provide the key to understanding a variety of human diseases, including cancer.

V. CONCLUDING REMARKS Recent advances in molecular membrane biology indicate that caveolae and hetero-trimeric G-proteins play a heretofore unimagined role both in cell communication and cellular transport. Traditionally, the role of G-proteins was confined to transmembrane signaling. Now it is known that they also regulate the budding of trans-Golgi-derived transport vesicles. Conversely, the role of caveolae was confined to cellular transport, but has now been expanded to include signaling functions. Because the same molecular machinery appears to participate in both of these processes, it seems that transmembrane signaling and cellular transport are deeply interrelated. In this regard, caveolae—for the first time—^provide an organellar or compartmental framework for organizing the complexity of transmembrane signaling events.

Caveolae, Transmembrane Signaling, and Cellular Transport

119

ACKNOWLEDGMENTS We thank Harvey Lodish for his patience and advice, and members of the Buratowski, Fink, Lodish, Matsudaira, and Young laboratories for thoughtful discussions. This work was supported in part by a grant from the W.M. Keck Foundation to the Whitehead Fellows Program (M.P.L.), and an NIH FIRST Award (GM-50443) to M.P.L.

NOTE 1. Corresponding author

REFERENCES Anderson, R.G.W., Kamen, BA., Rothberg, K.G., & Lacey, S.W. (1992). Potocytosis: sequestration and transport of small molecules by caveolae. Science 255,410-411. Bruns, R.R. & Palade, G.E. (1968). Studies on blood capillaries. I. General organization of blood capillaries in muscle. J. Cell Biol. 37, 244-276. Bushman, W.A., Wilson, L., Luttrell, D., Moyers, J., & Parsons, S.J. (1990). Overexpression of c-src enhances beta-adrenergic-induced cAMP accumulation. Proc. Natl. Acad. Sci. USA 87, 7462-7466. Chesebro, B. & Caughey, B. (1993). Scrapie agent replication without the prion protein? Curr. Biol. 3, 696-698. Dupree, P, Parton, R.G., Raposo, G., Kurzchalia, TV., & Simons, K. (1993). Caveolae and sorting of the trans-Golgi network of epithelial cells. EMBO J. 12,1597-1605. Fan, J.Y., Carpentier, J.-L., van Obberghen, E., Grunfeld, C , Gordeh, P., & Orci, L. (1983). Morphological changes of the 3T3-L1 fibroblast plasma membrane upon differentiation to the adipocyte form. J. Cell Sci. 61, 219-230. Fujimoto, T (1993). Calcium pump of the plasma membrane is localized in caveolae. J. CellBiol. 120, 1147-1157. Fujimoto, T, Nakade, S., Miyawaki, A., Mikoshiba, K., & Ogawa, K. (1993). Localization of inositol 1,4,5,-triphosphate receptor-like protein inplasmalemmal caveolae. J.CellBiol. 119,1507-1513. Ghitescu, L., Fixman, A., Simonescu, M., & Simonescu, N. (1986). Specific binding sites for albumin restricted to plasmalemmal vesicles of continuous capillary endothelium: receptor-mediated transcytosis. J. Cell Biol. 102,1304-1311. Glenney, J.R. (1989). Tyrosine phosphorylation of a 22 kD protein is correlated with transformation with Rous sarcoma virus. J. Biol Chem. 264, 20163-20166. Glenney, J.R. & Zokas, L. (1989). Novel tyrosine kinase substrates from Rous sarcoma virus transformed cells are present in the membrane cytoskeleton. J. Cell Biol. 108, 2401-2408. Goldberg, R.I., Smith, R.M., & Jarett, L. (1987). InsuUn and alpha-2 macroglobulin undergo endocytosis by different mechanisms in rat adipocytes. J. Cell. Phys. 133, 203-212. Gottesman, M., Roth, C , Vlahakis, G., & Pastan, I. (1984). Cholera toxin treatment stimulates tumorigenicity of RSV-transformed ceils. Mol. Cell Biol. 4,2639-2642.

120

MICHAEL P. LISANTI, ZHAOLAN TANG, and MASSIMO SARGIACOMO

Hooper, N.M. (1992). More than just a membrane anchor. Curr. Biol. 2, 617-619. Kamen, B.A., Smith, A.K., & Anderson, R.G.W. (1991). The folate receptor works in tandem with a probenecid-sensitive anion carrier. J. Clin. Invest. 87,1442. Lanners, H.N. (1991). Ultrastructure of erythrocytes from monkeys infected with P. vivax. Parasitol. Res. 77, 395^01. Lisanti, M.R, Tang, Z.-L., & Sargiacomo, M. (1993). CaveoUn forms ahetero-oligomeric protein complex that interacts with an apical GPI-linked protein: implications for the biogenesis of caveolae. J. Cell Biol. 123, 595-604. Milligan, G. (1988). Techniques used in the identification and analysis of function of pertussis toxin-sensitive guanine nucleotide binding proteins. Biochem. J. 255, 1-13. Montesano, R., Roth, J., Robert, A., & Orci, L. (1982). Non-coated membrane invaginations are involved in binding and internalization of cholera and tetanus toxins. Namre 296, 651-653. North, A.J., Galazkiewicz, B., Byers, T.J., Glenney, J.R., & Small, J.V. (1993). Complementary distributions of vinculin and dystrophin define two distinct sarcolemma domains in smooth muscle. J. Cell Biol. 120, 1159-1167. O'Dowd, B.F., Hnatowich, M., Caron, M.G., Lefkowitz, R.J., & Bouvier, M. (1989). Palmitoylation of the human beta-2 adrenergic receptor: mutation of cys-341 in the carboxyl tail leads to an uncoupled non-palmitoylated form of the receptor. J. Biol. Chem. 264, 7564-7569. Orci, L., Montesano, R., Meda, R, Malaisse-Lagae, R, Brown, D., Perrelet, A., & VassalU, P. (1981). Heterogeneous distribution of filipin-cholesterol complexes across the cistemae of the Golgi apparams. Proc. Natl. Acad. Sci. USA 78,293-297. Pace, A.M., Wong, Y.H., & Bourne, H.R. (1991). A mutant alpha subunit of Gy2 induces neoplastic transformation of Rat-1 cells. Proc. Natl. Acad. Sci. USA 88, 70317035. Peters, K.-R., Carley, W., & Palade, G.E (1985). Endothelial plasmalemmal vesicles have a characteristic striped bipolar surface structure. J. Cell Biol. 101,2233-2238. Raines, M.A., Kolesnick, R.N., & Golde, D.W. (1993). Sphingomyelinase and ceraminde activate mitogen-activated protein kinase in myeloid HL-60 cells. J. Biol. Chem. 268, 14572-14575. Rothberg, K.G., Heuser, J.E., Donzell, W.C, Ying, Y, Glenney, J.R., & Anderson, R.G.W. (1992). Caveolin, a protein component of caveolae membrane coats. Cell 68, 673-682. Saltiel, A.R. & Cuatrecasas, P. (1988). In search of a second messenger for insulin. Am. J.Phys.255,Cl-ll. Sargiacomo, M., Sudol, M., Tang, Z.-L., and Lisanti, M.P. (1993). Signal transducing molecules and GPI-linked proteins form a caveolin-rich insoluble complex in MDCK cells. J. Cell Biol. 122, 789-807. Simonescu, N., Simonescu, M., & Palade, G.E. (1975). Permeability of muscle capillaries to small heme-peptides: Evidence for the existence of patent transendothelial channels. J. Cell Biol. 64, 586-607. Soriano, P., Montgomery, C , Geske, R., & Bradley, A. (1991). Targeted disruption of the c-src proto-oncogene leads to osteopetrosis in mice. Cell 64, 693-702.

Caveolae, Transmembrane Signaling, and Cellular Transport

121

Wheater, P.R., Burkitt, H.G., & Daniels, V.G., eds. (1984). Functional Histology, p. 94. Churchhill Livingstone, London. Wolff, J.A., Dowty, M.E., Jiao, S., Repetto, G., Berg, R.K., Ludtke, J., Williams, R, & Slauterback, D.B. (1992). Expression of naked plasmids by cultured myotubes and entry of plasmids into T tubules and caveolae of mammalian skeletal muscle. J. CellSci. 103, 1249-1259. Yamada, E. (1955). The fine structure of the gall bladder epithelium of the mouse. J. Biophys. Biochem. Cytol. 1,445-458.

VI. ADDED IN PROOFS The following is a partial list of caveolae-related papers that have been published by our laboratory since this review was first written. Chun, M., Liyanage, U.K., Lisanti, M.R, & Lodish, H.F. (1994). Signal transduction of a G-protein coupled receptor in caveolae: Colocalization of endothelin and its receptor with caveolin. Proc. Natl. Acad. Sci., USA 91,11728-11732. Koleske, A.J., Baltimore, D., & Lisanti, M.R (1995). Reduction of caveolin and caveolae in oncogenically transformed cells. Proc. Natl. Acad. Sci., USA 92, 1381-1385. Kuliawat, R., Lisanti, M.R, & Arvan, P. (1995). Polarized distribution and delivery of plasma membrane proteins in thyroid follicular epithelial cells. J. Biol. Chem. 270, 2478-2482. Li, S., Okamoto T., Chun, M., Sargiacomo, M., Casanova, J.E., Nishimoto, L, & Lisanti, M.R (1995). Evidence for a regulated interaction between hetero-trimeric G proteins and caveolin. J. Biol. Chem. 270,15693-15701. Lisanti, M.R, Scherer, P., Tang, Z.-L., & Sargiacomo, M. (1994). Caveolae, caveolin and caveolin-rich membrane domains: a signalling hypothesis. Trends Cell Biol. 4, 231-235. Lisanti, M.R, Scherer, RE., Tang, Z.-L., Kubler, E., Koleske, A.J., & Sargiacomo, M. (1995). Caveolae and human disease: transcytosis, potocytosis, signalling and cell polarity. Sem. Dev. Biol. 6,47-58. Lisanti, M.R Scherer, RE., Vidugiriene, J., Tang, Z-L., Vosatka, A.H., Tu, Y-H., Cook, R.F., & Sargiacomo, M. (1994). Characterization of caveolin-rich membrane domains isolated from an endothelial rich source: implications for human disease. J.CellBiol. 126,111-126. Lisanti, M.R, Tang, Z.-L., Scherer, P., & Sargiacomo, M. (1995). Caveolae purification and GPI-linked protein sorting in polarized epithelia. Meth. Enzymol. 250, 655668. Sargiacomo, M., Scherer, RE., Tang, Z.-L., Kubler, E., Song, K., Sanders, M., & Lisanti, M.R (1995). Oligomeric structure of caveolin: implications for caveolae membrane organization. Proc. Natl. Acad. Sci., USA, in press. Sargiacomo, M., Scherer, RE., Tang, Z.-L., Casanova, J.E., & Lisanti, M.R (1994). In vitro phosphorylation of caveolin-rich membrane domains: identification of an associated serme kinase activity as a casein kinase Il-like enzyme. Oncogene 9, 2589-2595.

122

MICHAEL P. LISANTI, ZHAOLAN TANG, and MASSIMO SARCIACOMO

Scherer, P.S., Tang, Z.-L., Chun, M., Sargiacomo, M., Lodish, H.F., & Lisanti, M.P. (1995). Caveolin isoforms differ in their N-terminal protein sequence and subcellular distribution: identification and epitope mapping of an isoform-specific monoclonal antibody probe. J. Biol. Chem. 270, 16395-16401. Scherer, P., Lisanti, M.P., Baldini, G., Sargiacomo, M., & Lodish, H.F. (1994). Induction of caveolin during adipogenesis and association of GLUT4 with caveolin-rich vesicles. J. Cell Biol. 127, 1233-1243. Tang, Z.-L., Scherer, P.E., & Lisanti, M.P. (1994). The primary sequence of murine caveolin reveals a conserved consensus site for phosphorylation by protein kinase C. Gene 147, 299-300.

NUCLEAR TRANSPORT OF URACIL-RICH SMALL NUCLEAR RIBONUCLEOPROTEIN PARTICLES

Elisa Izaurraide, lain W. Mattaj, and David S. Goldfarb

I. II. III. IV. V. VI. VII. Vni. IX. X.

Introduction The Nuclear Import of Proteins TheNuclearlmportofMacromolecular Assemblies Composition and Function of Spliceosomal U snRNPs The RNA Constituents of Spliceosomal snRNPs The Protein Components A. The Common Proteins B. The Specific Proteins Assembly Pathway of Spliceosomal U snRNPs Nuclear Export of U snRNPs The Role of the Cap Structure in UsnRNA Export Characterization of the Nuclear Cap Binding Activity Involved in U snRNA Export

Membrane Protein Transport Volume 2, pages 123-159 Copyright © 1995 by JAI Press Inc. All rights of reproduction in any form reserved. ISBN: 1-55938-983-4

123

124 126 128 132 132 133 133 135 137 138 139 140

124

ELISA IZAURRALDE, IAIN W. MATTAJ, and DAVID S. GOLDFARB

XL Nuclear Import of UsnRNPs XII. The Requirement of the Trimethyl Cap Structure for U snRNA Nuclear Import Is Not General XIII. A Unifying Model for the Role of the Trimethyl Cap and the Core Domain in Signaling U snRNP Nuclear Import XrV. Nuclear Import of U6snRNP after Mitosis XV. The Role of Cap Structures in Signaling RNA Subcellular Localization XVI. Inhibitors of Nuclear Transport XVII. Kinetic Evidence for Multiple Nuclear Targeting Pathways in U snRNP Import XVIII. Multiple U snRNP Import Pathways References

142 144 146 147 147 148 151 152 153

I. INTRODUCTION Uracil-rich small nuclear ribonucleoprotein particles (U snRNPs) are RNA-protein complexes involved in a large variety of nuclear functions, including pre-ribosomal RNA and pre-messenger RNA processing, 3' end formation of some classes of messenger RNA, and ribosome assembly. Many U snRNPs have not been functionally characterized, and it is therefore likely that the list of cellular processes in which these particles play a role will be extended in the future.

Figure 1. The pathway of synthesis and assembly of spliceosomal sRNPs in higher eukaryotes. The U4 snRNA is a representative of the polymerase ll-transcribed snRNAs. After transcription, these snRNAs are transported to the cytoplasm, where they assemble with the Sm proteins. The RNA cap is then hypermethylated before the newly synthesized snRNP is transported back to the nucleus. RNA post-transcriptional modification and assembly with snRNP specific proteins are likely to occur in both the nuclear and cytoplasmic compartments. In contrast, the synthesis of the polymerase Ill-transcribed U6 snRNA is strictly nuclear. U4 and U6 snRNA associate by base-pairing in the same snRNP. The solid lines represent DNA (when indicated) or RNA, the stippled line ihe 3' extension of the precursor RNA. The ovals and circles represent the common (Sm) or specific proteins. Pol II and III are RNA polymerases II and III, respectively. m^'G, the monomethyl guanosine cap structure, ^mG, the trimethyl guanosine cap, and \|/ represent internal base modifications.

125

1 26

ELISA IZAURRALDE, IAIN W. MATTAJ, and DAVID S. GOLDFARB

Based on their subnuclear localization U snRNPs can be divided into two broad classes: nucleolar (sno) and nucleoplasmic (sn) U RNPs. Nucleolar U snoRNPs are believed to play a role in ribosomal RNA processing and/or ribosome assembly. However, direct evidence of involvement in these processes has been obtained in only a few cases (i.e., U3, U8, U14). The major nucleoplasmic U snRNPs (i.e., Ul, U2, U4/U6, and U5) are involved in pre-mRNA splicing and are designated "spliceosomal U snRNPs." One of the minor nucleoplasmic U snRNPs (U7) has been shown to play a role in the formation of 3' ends on histone mRNAs. Because of their relative abundance in mammalian cells, the early availability of a collection of autoimmune sera directed against some of their constituent proteins, and the later generation of monoclonal antibodies suitable for large-scale purification, spliceosomal U snRNPs are the best characterized members of the family. The nucleocytoplasmic transport of these particles has been extensively studied and will be the main subject of this review. However, we will, where necessary, refer to aspects of the transport behavior of other classes of U snRNP. The nucleocytoplasmic trafficking of all the U snRNPs we will describe, except U6 snRNP, involves a transport circuit that begins with the export of newly transcribed U snRNAs to the cytoplasm and then, following U snRNP assembly and modification, ends with the import of the U snRNPs into the nucleus where they function (see Figure 1). An outstanding issue in the field of nuclear transport is the novel nature of the signal that directs the import of mature U snRNPs. We will begin by discussing general properties of the nuclear import apparatus that are relevant to U snRNP import, focusing particularly on specializations of the targeting apparatus that are adapted to mediate the import of oligomeric import substrates such as U snRNPs. Aspects of U snRNA export will be discussed in later sections.

11. THE NUCLEAR IMPORT OF PROTEINS The mechanism of protein import to the nucleus serves as a good paradigm for understanding the import of other macromolecules such as U snRNPs. The import of proteins is believed to occur predominantly by a post-translational mechanism, although some cotranslational import may occur. As in other protein targeting pathways the nuclear import of proteins can be divided into an initial targeting phase, where the nuclear localization signal (NLS)-containing protein is brought to the nuclear

Nuclear Transport of U snRNPs

127

envelope, and a subsequent ATP-dependent translocation stage, where the substrate moves through the nuclear pore complex (Garcia-Bustos et al., 1991; Forbes, 1992). NLS-containing proteins are recognized in the cytoplasm by the targeting apparatus after their synthesis, their reappearance in the cytoplasm during mitotic nuclear envelope breakdown, or their egress from the nucleus during a cycle of shuttling (Borer et al., 1989; Goldfarb, 1991). In situ (Breeuwer and Goldfarb, 1990) and in vitro (Adam and Gerace, 1991) studies indicate that the karyophile (nuclear-targeted macromolecule) is rapidly bound in the cytoplasm by a NLS receptor that commits the karyophile (nuclear-targeted macromolecule) to a mediated import pathway. This commitment step is particularly important for smaller karyophiles such as histone HI (21 kDa), which, if left uncomplexed, would diffuse through 10-nm aqueous channels in the nuclear pore complex (Breeuwer and Goldfarb, 1990; Forbes, 1992; Hinshaw et al., 1992). The 10-nm diffusion channels may allow some smaller macromolecules that lack NLSs, and hence are not bound by the targeting apparatus, to cross the nuclear envelope by passive diffusion. The nuclear transport of any macromolecule that exceeds the diameter of the diffusion pore, about 50 kDa for globular proteins, occurs exclusively by NLS-mediated facilitated transport. U snRNPs are definitely too large to cross the nuclear envelope by diffusion. Translocation across the nuclear envelope is catalyzed by the nuclear pore complex (Garcia-Bustos et al, 1991; Forbes, 1992). The structure and function of the yeast (Hurt vol2) and mammalian (Aebi vol3) pore complexes are topics of other articles in this series and will not be described here in any depth. Access to the pore complex is controlled mainly through targeting pathways that deliver substrates to the nuclear envelope. However, the functional efficiency of the pore complex itself can vary with cellular activity (see below). In the following sections we will focus mainly on targeting pathways, although selected aspects of translocation will be presented where relevant. Whereas studies in living cells have revealed the basic steps along the protein import pathway, the development of cell-free nuclear transport assays (Adam et al., 1990; Newmeyer and Wilson, 1991) has led to the functional characterization of transport factors, beginning with NLS receptors (see Yamasaki and Lanford, 1992). The best characterized putative NLS receptor is P54/56, a pair of polypeptides whose NLS binding activity was discovered in rat liver extracts, but whose function in targeting was demonstrated with protein purified from bovine eryth-

128

ELISA IZAURRALDE, IAIN W. MATTAJ, and DAVID S. GOLDFARB

rocyte extracts (Adam et al., 1989; Adam and Gerace, 1991). Specifically, P54/56 can be efficiently cross-linked to wild-type NLS sequences but has significantly lower affinity for NLS point mutations that do not function in transport (Adam et al., 1989). Also, P54/56 is the only NLS-binding protein that has been shown to function during targeting in cell-free nuclear transport assays (Adam and Gerace, 1991). Cell fractionation studies indicate that several other factors probably act in concert with NLS receptors to target proteins at the nuclear envelope, some of which are sensitive to N-ethylmaleamide, a sulfhydryl-modifying reagent (Forbes, 1992). Other factors, hsc70/hsp70 chaperone proteins (Imamoto et al., 1992; Shi and Thomas, 1992) and Ran/TC4 (Moore and Blobel, 1993; Melchior et al., 1993), a 26 kDa Ras-like GTPase, have been linked to protein import. The identification of a GTPase as a factor in nuclear transport is intriguing because GTPases have been shown to function as conformational switches in the intracellular targeting of membrane vesicles and the SRP-mediated targeting of signal-peptide-containing proteins at the endoplasmic reticulum membrane (Goldfarb, 1994). P54/56 and hsc70 may accompany the karyophile through the nuclear pore into the nucleus, presumably to be recycled back to the cytoplasm for further rounds of import (Okuno et al., 1993). Finally, it is likely that hsc70 is required for the import of some, but not all, karyophiles. Targeting is completed when the karyophile is delivered to docking sites at the pore complex. The initial docking site is probably at the periphery of the pore complex (Akey and Goldfarb, 1989), perhaps on filaments that protrude into the cytoplasm (Cordes et al., 1993, and references therein). Little is known about any of the steps that occur after initial docking, including the transfer of the karyophile from peripheral docking sites to the central translocation channel. Ultimately, ATP-dependent translocation is believed to occur through the 70-nm-deep central channel, which can dilate to accommodate larger substrates (Akey and Goldfarb, 1989).

III. THE NUCLEAR IMPORT OF MACROMOLECULAR ASSEMBLIES A key feature of U snRNPs is the fact that neither the individual protein or snRNA components of the U snRNP, when present alone in the cytoplasm, are karyophilic. As described below U snRNP import requires the assembly of U snRNP constituents to generate a karyophilic particle.

Nuclear Transport of U snRNPs

129

In this section we will introduce aspects of nuclear transport that are peculiar to oligomeric complexes. Another characteristic of U snRNP nuclear trafficking is the passage of the U snRNA, first out of the nucleus and then back again as an assembled, processed U snRNP. This kind of circuit is analogous to the trafficking of ribosomal proteins, which are initially imported to the nucleus, assembled into ribosomal subunits in nucleoli, and then exported back to the cytoplasm as 40S and 60S subunits. During pre-mRNA export, proteins associated with the nuclear pre-mRNA are thought to leave the nucleus as part of the pre-mRNP; they are displaced from the RNA in the cytoplasm and re-imported back to the nucleus, presumably for another round of transport (Pinol-Roma and Dreyfuss,1992). Thus a good fraction of nucleocytoplasmic trafficking consists of proteins and RNAs that pass through the pore complex either twice, in the case of U snRNAs and ribosomal proteins, or repeatedly, as in the case of shuttling proteins like certain pre-mRNA binding proteins. Implicit in these trafficking circuits is the capacity of the nuclear pore complex to selectively transport intact macromolecular assemblies in both directions. Although it is possible, even likely, that these simple ckcuits are controlled by positive or negative feedback between import and export steps, nothing yet is known about the coordination of bidirectional transport. Therefore the import and export legs of U snRNP nuclear trafficking circuits will be considered here as two separate, noninteracting reactions. As U snRNPs are assembled and covalently modified in the cytoplasm (see below), it is possible that conformational changes occur that are crucial for signaling import. Whereas the translocation of proteins across the endoplasmic reticulum and mitochondrial membranes requires that the polypeptides be unfolded as they cross the membrane(s), fully folded polypeptides and quaternary associations are largely retained during passage across the nuclear envelope. The demonstrable import of U snRNPs and export of ribosomal subunits (Bataille et al., 1990) are strong evidence that hetero-oligomeric complexes cross the nuclear envelope intact. However, this fact does not necessarily imply that the local tertiary conformation of the karyophile is irrelevant to nuclear transport. The role of hsc70 in nuclear transport may be to ensure the adequate presentation of the NLS to the targeting receptors (Goldfarb, 1992). It will be interesting to learn if there is a role for hsc70 in U snRNP assembly and import. Because most karyophiles exist as oligomeric complexes, NLS copy number effects are an important, albeit poorly understood, feature of

130

ELISA IZAURRALDE, IAIN W. MATTAJ, and DAVID S. GOLDFARB

nuclear transport. NLS combinatory effects may well be a factor in U snRNP import. In most carefully studied cases, each subunit of an oligomeric protein contains its own independently functioning NLS sequence. This is true, regardless of whether the polypeptide is imported as a homo-oligomer or a hetero-oligomer, or in association with an RNA or DNA. The presence of NLSs within each subunit of a macromolecular assembly may allow the independent import of unassociated, excess, or replacement subunits. Experimentally, only one NLS is required for the import of multimeric proteins. For example, when antibodies directed against nuclear proteins are microinjected into the cytoplasm of tissue culture cells, they will accumulate in nuclei via their association with the NLS-containing protein. Viral DNAs and RNAs probably enter nuclei by association with karyophilic proteins (see below). A classical case is that of nucleoplasmin, an abundant pentameric protein located in the nuclei of Xenopus oocytes. The C-terminal NLS of the nucleoplasmin polypeptide can be proteolytically removed without disrupting the pentamer to produce a series of proteins with 0-5 NLSs. Pentamers having between one and five NLSs were still transported into oocyte nuclei, although the rate of import was less for those with fewer NLSs (Dingwall et al, 1982). Additive effects of NLS sequences on import have also been observed for mutated NLSs that are too weak to function independently. In one study, a partially defective NLS was inserted at three different sites within the primary sequence of a tetrameric cytoplasmic protein, chicken pyruvate kinase (CPK) (Roberts et al., 1987). CPK tetramers containing only one defective NLS per monomer were localized mainly to the cytoplasm; tetramers containing the defective NLS in all three locations (12 NLSs per tetramer) were imported at levels comparable to those found in wild-type nuclear proteins. Taken together, these studies indicate that the activity of NLSs can be additive, both in instances where a single polypeptide contains multiple NLSs and where the number of NLSs increases upon subunit oligomerization. The mechanism of NLS additivity is likely to involve avidity effects that increase the stability of the NLS-receptor-karyophile complex. The effect of NLS copy number on import rate and efficiency has been modeled by studying variously sized NLS-colloidal gold particles injected into mammalian tissue culture cells (Dworetzky et al., 1988; Feldherr and Akin, 1990). Here, the import of the larger particles was accelerated by increasing the number of NLSs adsorbed to their surface. Very large gold particles (>20 nm) were imported only when associated

Nuclear Transport of U snRNPs

131

with high numbers of NLSs. The impUcation from these studies is that NLS copy number influences the efficiency of translocation through the pore complex. Thus NLS copy number effects may act during both targeting and translocation stages of import. U snRNPs are relatively large assemblies and may, therefore, require multiple NLSs for efficient import. The import of RNA and DNA viruses is vaguely similar to the import of U snRNPs because the import of the nucleic acid component is mediated by associated proteins. During the import of adenovirus (Greber et al, 1993) and HIV-1 (Bukrinsky et al., 1993b), capsid proteins must be removed before the particle can be imported. Like U snRNPs, therefore, these viruses have latent periods in the cytoplasm before they become karyophilic. However, the mechanisms of viral and U snRNP latency are probably dissimilar. Being extremely large, most karyophilic viruses appear to display multiple NLSs in the form of associated NLS-containing polypeptides. A recent example is the import of HIV-1 (Bukrinsky et al., 1993a). Import of one HIV-1 strain is mediated by a NLS-containing gag-matrix protein. The mutation of the gag-matrix NLS inhibited the nuclear transport of the 160S HIV-1 preintegration complex. Perhaps as many as 100-200 copies of the gag-matrix protein are normally associated with the preintegration complex. Among all the different polynucleic acids that are imported into nuclei, both in nature and in the laboratory, including viruses, retrotransposible elements, plasmids, and transforming DNAs, the role of the m^'^'^GpppN cap of U snRNAs is unique in its apparent role as a factor in signaling import (see below). Although the fundamental mechanics of nuclear transport are probably conserved among all eukaryotes, quantitative differences in import rates are commonly observed among different cell types. As will be described below, the particulars of U snRNP import in somatic cells can differ significantly from their transport in oocytes. Furthermore, the capacity of the transport apparatus can vary according to the physiological state of the cell. ProUferating cells import larger colloidal gold particles more efficiently than cells that have been maintained in a nonprolifcrating state for many days (Feldherr and Akin, 1990). This effect is a property of the nuclear envelope and not the cytoplasm. Studies using heterokaryons between proliferating and nonproliferating cells showed that the nonproliferating nucleus retained its poor import capacity even when bathed in the cytoplasm of a proliferating cell; the proliferating nucleus continued to import efficiently (Feldherr and Akin,

1 32

ELISA IZAURRALDE, IAIN W. MATTAJ, and DAVID S. COLDFARB

1993). The role of cell activity in nuclear transport is the topic of another article in this series (Feldherr, vol3). We hope that the preceding paragraphs will provide a general context for the specific discussion of U snRNP nuclear trafficking. In the next section we will summarize some aspects of U snRNP composition, assembly, and function. These topics have been reviewed recently (Liihrmann et al., 1990; Andersen and Zieve, 1991; Mattaj et al, 1993). For more extensive background information, interested readers are referred to these articles.

IV. COMPOSITION AND FUNCTION OF SPLICEOSOMAL U snRNPs Spliceosomal U snRNPs are composed of one (Ul, U2, and U5) or two (U4/U6) short uracil rich RNAs (U snRNAs) in association with a set of proteins. Some of the proteins are common to the various snRNPs, and others are specific for individual snRNPs. The particles are named according to the RNA molecules that they contain, either Ul, U2, U5, or U4 and U6 snRNA. In the U4AJ6 snRNP particle, U4 snRN A extensively base-pairs with U6 snRNA. A tri-snRNP complex composed of U4/U6 and U5 snRNPs has also been described. Formation of the tri-snRNP is functionally essential for splicing activity, but this form of the snRNPs seems to be more labile to isolation conditions than are the individual U4/U6 or U5 snRNPs. In fact, all spliceosomal snRNPs interact with each other, and with additional factors, in a sequential order to form a functional splicing complex, the spliceosome, on the precursor mRNA. Most of these interactions are transient and change during the dynamic process of pre-mRNA splicing (see Moore et al., 1993, for a recent comprehensive review).

V. THE RNA CONSTITUENTS OF SPLICEOSOMAL snRNPs With the exception of U6, spliceosomal U snRNAs are transcribed by RNA polymerase II. Polymerase Il-transcribed U snRNAs are synthesized as precursors with extended y ends, which are removed during the assembly of U snRNPs (Elicieri, 1974; Madore et al., 1984). Our knowledge of the activities involved in this process is still limited and will be discussed below in more detail. According to the polymerase involved in their transcription, U snRNAs differ at their 5' ends. Polymerase Il-transcribed U snRNAs, in

Nuclear Transport of U snRNPs

133

common with other pol II transcripts, have a monomethyl guanosine cap (m^GpppN) added cotranscriptionally (Reddy and Bush, 1988, and references therein). This cap structure will be subsequently hypermethylated during the transient cytoplasmic phase of U snRNP assembly (see below). U6, the only spliceosomal snRNA transcribed by polymerase III, carries a y-methyltriphosphate at its 5' terminus (Singh and Reddy, 1989). In addition to the modifications of their 5' and 3' termini, U snRNAs possess multiple modified internal nucleotides (Reddy and Bush, 1988). It is not yet clear in which cellular compartment these post-transcriptional modifications occur and whether association with snRNP proteins is required for intemal modification. The functional significance of these modifications has also not been studied. VI. THE PROTEIN COMPONENTS A. The Common Proteins The four spliceosomal U snRNPs, Ul, U2, U5, and U4/U6, have at least nine polypeptides in common, with molecular masses ranging from 9 to 69 kDa (Table 1; see Wolin and Walter, 1991; Will et al., 1993, and references therein). The latest member of this group, a polypeptide with an apparent molecular mass of 69 kDa, was only recently described (Hackl et al., 1993), probably because its interaction with snRNP particles appears to be more labile than that of the other common proteins. The common proteins are also named core proteins or, because of their interaction with sera of the anti-Sm serotype, Sm proteins (Lemer et al., 1979). The particles formed by association of the core proteins with individual U snRNAs are called core particles to distinguish them from the mature particles, which contain, in addition, specific proteins. The core proteins are made and stored in the cytoplasm in the form of protein complexes that associate with the U snRNAs after their export from the nucleus (Eliceiri, 1974; De Robertis et al., 1982). Apparently, these proteins do not contain an exposed nuclear localization signal of the type found on karyophilic proteins, and their nuclear migration takes place only upon their binding to the U snRNAs (Zeller et al., 1983; Mattaj and De Robertis, 1985). In Xenopus oocytes, the Sm proteins are stored in very large amounts in the cytoplasm in the absence of snRNAs for use in early embryogenesis (Zeller et al., 1983). Although there is little apparent storage of common snRNP proteins in the cytoplasm of somatic cells, the pathway of snRNP assembly is the same as in oocytes. Zieve

134

ELISA IZAURRALDE, IAIN W. MATTAJ, and DAVID S. GOLDFARB

Table 1. The Common Proteins of Spliceosomal U snRNPs

Name

Apparent molecular (kDa)

G F E Dl D2

9 11 12 16 16.5

D3 B B'

18 28 29

mass

69 Note: The core proteins are named according to their electrophoretic mobility. The molecular masses are according to Will et al. (1993).

and co-workers have described three different core protein complexes in murine cells: a complex with a sedimentation coefficient of 6S, composed of D, E, F, and G core proteins with a suggested stoichiometry of D2EFG (note that these results were obtained before it became clear that there are three distinct D-sized proteins, Dl, D2, and D3; see Table 1). This particle seems to be the first protein complex that assembles with snRNA (Fischer et al., 1985; Zieve and Sauterer, 1990). However, since additional core proteins have been identified subsequent to these initial studies, the snRNP assembly pathway should be reinvestigated in more detail. A 20S protein complex was also found, composed of the D core protein in association with a 70-kDa protein. Whether this 70-kDa protein corresponds to the recently described 69-kDa core polypeptide is not yet known. The B/B' core proteins sedimented at around 5S-6S, at l i s , and at 20S; however, it is not clear whether this heterogeneity reflects self-association or interaction with other proteins. The core proteins interact in a sequential order with a partially conserved RNA structural motif that is designated the Sm binding site or domain A and is present in U1, U2, U4, and U5, but not in U6 snRNAs (Branlant et al., 1982). The core of the Sm binding site consists of a single-stranded region containing the consensus sequence PuA(U)^GPu, with n > 3. This core is usually flanked by hairpin loop structures (Jacob et al., 1984). In spite of this apparent conservation, different Sm binding

Nuclear Transport of U snRNPs

135

sites are not equivalent, and the contribution of neighboring RNA sequences and structural elements to the binding of Sm proteins varies in different snRNPs (Jarmolowski and Mattaj, 1993). Nevertheless, it appears at present that exactly the same set of core proteins interact with each of the individual RNAs. Indeed, electron microscopic studies have revealed that U1, U2, U4, and U5 core particles are morphologically very similar and form a spherical body about 8 nm in diameter (Kastner et al., 1990). This basic structural element of the snRNPs has been called the Sm core domain. Interestingly, the Sm binding site of U7 snRNAs from various organisms does not correspond well to the consensus sequence listed above. This results in inefficient assembly with the core proteins (Grimm et al., 1993). This inefficient interaction, paradoxically, has been shown to be essential for the assembly of functional U7 snRNPs (Grimm etal., 1993). Core proteins play an important role in snRNP assembly and subcellular localization. In particular, hypermethylation of the cap structure of U snRNAs, transport of the assembled particles into the nucleus, and possibly some post-transcriptional modifications of the RNA strictly depend on binding of the core proteins to the Sm domain (Mattaj and De Robertis, 1985; Mattaj, 1986). No evidence for a direct functional role of these proteins in splicing has been provided. Indeed, it has been shown in vitro, in an extract of human cell nuclei, that a U4AJ6 snRNP lacking stably bound core proteins is functionally active in splicing complementation (Wersig and Bindereif, 1992). B. The Specific Proteins

Apart from the common proteins, each snRNP contains unique proteins (see Table 2). Ul- and U2-specific proteins were initially identified by the use of autoimmune antibodies. For example, antibodies against the U1 -specific proteins 70K, A and C, are found in the serum of patients suffering from systemic lupus erythematosus and mixed connective tissue diseases. Autoantibodies to the U2-specific proteins A' and B'^ can also be detected in patients suffering from autoimmune diseases. Specific proteins associated with U5 and the triple U4/U6.U5 snRNP, as well as additional U2-associated proteins, were identified by biochemical fractionation (Will et al., 1993, and references therein). The protein composition of purified snRNPs varies according to the ionic conditions used during the isolation procedure. At high salt, U2 snRNP sediments with a coefficient of 12S and is composed of the core proteins

136

ELISA IZAURRALDE, IAIN W. MATTAJ, and DAVID S. GOLDFARB Table 2. The Specific Proteins of Spliceosomal U snRNPs Ul (12S) C(22) A (34) 70K (70)

U2 (17S) B " (28.5) A'(31)

35 53 60 66 92 110 120 150 160

U5 (20S)

15 40 52 100 102 116 200 205

U4/U6.U5 (25S)

15 40 52 100 102 116 200 205 + 15.5 20 27 60 60 60 90

in association with the two U2-specific proteins A' and B''. At low ionic strength U2 sediments at 17S. This shift in the sedimentation coefficient is due to the presence of nine additional proteins (Table 2). Similarly, the triple snRNP, which has a sedimentation coefficient of 25S, is observed at low salt concentration and contains, in addition to the proteins already associated with the individual U5 and U4/U6 snRNPs, five additional polypeptides specific for the tri-snRNP complex. It is likely that other proteins transiently associate with the snRNPs during assembly of the particles (chaperones), their transport across the nuclear pore (transport receptors), or splicing (non-snRNP splicing factors). In contrast with the common proteins, whose nuclear accumulation is strictly dependent on their recruitment into U snRNP core particles, some of the specific proteins move to the nucleus separately from the rest of the snRNP (Feeney et al., 1989; Jantsch and Gall, 1992; Kambach and Mattaj, 1992,1994). Examples of this are the two Ul snRNP-specific proteins Ul A and UlC and the two U2-specific proteins U2A' and U2B'^ Nuclear targeting of these proteins appears to be mediated by the mechanism common to karyophilic proteins. However, the nuclear localization signals of Ul A and U2B'' proteins, which were subject to a more detailed analysis, are unusual in primary structure and do not correspond to either of the

Nuclear Transport of U snRNPs

137

"canonical" types of NLS previously defined for nuclear proteins (see below). VIL ASSEMBLY PATHWAY OF SPLICEOSOMAL U snRNPs Assembly of U snRNPs takes place in the cytoplasm where, as discussed above, pools of core proteins exist in the form of various complexes (see Figure 1). Newly synthesized pol II U snRNAs are exported into the cytoplasm and associate with the common proteins before migrating back to the nucleus (EUcieri, 1974; Mattaj and De Robertis, 1985; Zieve et al., 1988). The major exception to this pathway is the U6 snRNA, which lacks a binding site for the core proteins and appears not to leave the nucleus after transcription (Vankan et al., 1990). The interaction of U6 with U4 snRNA to form the double U4AJ6 snRNP occurs in the nucleus, upon re-entry of U4 core particles (Vankan et al., 1990; Wersig et al., 1992). In most cell types examined, there is an excess of U6 RNA over U4 snRNA, perhaps to ensure efficient U4/U6 assembly. During the transient cytoplasmic phase, the monomethyl cap structure of U snRNAs is converted into a trimethyl cap (m^'^'^GpppN) by the addition of two methyl groups at position 2 of the guanosine. The identity of the methylase is not known, but the observation that binding of the Sm proteins is a prerequisite for its activity suggests that the enzyme must recognize, in addition to the cap structure, some determinants on the snRNP core particle (Mattaj, 1986). It is also during assembly that trimming of the 3' ends and modification of the internal nucleotides take place. Processing of the extended 3' termini is initiated in the nucleus, since on the appearance of Ul, U2, and U4 snRNAs in the cytoplasm, most of the 3' additional nucleotides have been trimmed (Zieve and Sauterer, 1990). However, the remaining few additional 3' nucleotides seem to be processed after binding of the core proteins, which is dependent on the RNA cap structure and on the occurrence of transport (Neuman de Vegvar and Dahlberg, 1990; Yang et al., 1992). Thus, accurate 3' end formation appears to be coupled to, and might affect the efficiency of, transport of U snRNPs through the nuclear pore complex and seems to involve interaction with both termini of the U snRNAs (Neuman de Vegvar and Dahlberg, 1990; Yang et al., 1992). Precise 3' end formation, on the other hand, is not essential for reaccumulation of U snRNAs in the nucleus after snRNP assembly (Konings and Mattaj, 1987; Jarmolowski and Mattaj, 1993).

138

ELISA IZAURRALDE, IAIN W. MATTAJ, and DAVID S. GOLDFARB

In the next sections we focus our attention on the steps of the pathway that involve the transport of U snRNPs back and forth through the nuclear pore complex. It should be noted that, although the assembly pathway described above is well supported by experiments carried out in various vertebrate cell types, the same pathway has not yet been demonstrated to be universal among eukaryotes. Furthermore, the more detailed knowledge of signals and mechanisms described below is largely based on experiments carried out in a single, highly specialized cell type, the Xenopus laevis oocyte, because of the ease with which RNA transport can be examined and manipulated in these cells.

VIII. NUCLEAR EXPORT OF U snRNPs Many of the data on U snRNA export have been obtained using a Ul snRNA mutant, UlASm (previously called UlAD), which carries a substitution of six nucleotides in its Sm binding site (Hammet al., 1987). Unlike the wild-type RNA, whose cytoplasmic residence is transient, the mutant RNA accumulates in the cytoplasm after export, since it is no longer able to interact with the core proteins. This uncoupling of export from re-import makes UlASm RNA suitable for export studies. If export takes place, the RNA is in the cytoplasm, whereas if export is blocked (i.e., by inhibitors, modification of the RNA, or the use of antibodies) the RNAremains in the nucleus. UlASm snRNA export, in common with rRNA, tRNA, and mRNA export, exhibits the characteristics of a receptor-mediated process, such as temperature dependence and saturability (Jarmolowski et al., 1994). One of the first insights into the U snRNA export mechanism was provided by the observation that after transcription U6 snRNA is not exported into the cytoplasm, unlike pol II U snRNAs (Vankan et al., 1990). Since U6 snRNA is the only spliceosomal RNA transcribed by RNA polymerase III, it was important to determine whether it was the polymerase responsible for the transcription of a particular RNA, which determines whether this RNA will be exported or retained into the nucleus. This question was investigated by Hamm and Mattaj (1990), who compared the export of UlASm RNA synthesized by pol II with an artificial UlASm RNA whose transcription was driven by the RNA polymerase III pro motor derived from the U6 gene. The pol III UlASm RNA was retained in the nucleus, whereas the pol II transcript, as expected, accumulated in the cytoplasm. Since the only difference between these two RNAs was the presence of the cap structure at the 5' end

Nuclear Transport of U snRNPs

139

of the pol II transcript as opposed to the triphosphate 5' end present on the pol III transcript, the role of the cap structure in Ul snRNA export was investigated in more detail. First, it was shown that UlASm snRNA export could be competitively inhibited by a cap analog, the dinucleotide m^GpppG. Second, the export of an in vitro synthesized RNA, carrying a trimethyl cap structure, was found to be slower than that of the monomethyl-capped RNA. This study provided the first indication that the cap structure played a role in RNA export. The observation that the export of a capped RNA could be inhibited by a large excess of the cap analog strongly suggested that the cap structure was recognized directly by one or more factors that participate in Ul snRNA export. In the competition experiment, this factor was titrated by the excess of cap analog, resulting in inhibition of transport. The observation that a trimethyl capped RNA was still exported, but with slower kinetics, indicated that the physiological cap structure (m'^GpppN) was not absolutely required for RNA export, but that it would accelerate the export rate of the RNA. There are two plausible explanations for this observation. One is that an RNA with a trimethyl cap can still be recognized by the cap binding factor(s), but with a lower affinity, resulting in slower kinetics of export. The second is that interaction with the cap binding factor is not essential for export, but affects export rate when present on the RNP export substrate (see below). More recent studies on RNA export in Xenopus oocytes have confirmed the conclusions of this work and further showed that the role of the cap structure is not restricted to U snRNA export, but can also be extended to mRNA export (Dargemont and Kiihn, 1992; Jarmolowski et al., 1994). It appears, however, that although factors that recognize the cap structure are essential for the export of U snRNAs from oocyte nuclei, their effect on mRNA is less important, being confined to affecting the rate of export (Jarmolowski et al., 1994).

IX. THE ROLE OF THE CAP STRUCTURE IN U snRNA EXPORT The role of the cap structure in U snRNA export was further investigated by Jarmolowski et al. (1994). By microinjection of increasing quantities of m vitro synthesized, unlabeled Ul ASm snRNA competitor into oocyte nuclei these authors could progressively inhibit export of labeled Ul ASm RNA transcripts. The inhibition was observed when the competitor RNA had a m^GpppG cap but not an ApppG. These experiments show not only

140

ELISA IZAURRALDE, IAIN W. MATTAJ, and DAVID S. GOLDFARB

that Ul snRNA export is a saturable process but furthermore, the export factor that is first saturated by increasing quantities of UlASm RNA substrate requires the cap structure in order to recognize RNA. The essential role of the cap structure was confirmed by the observation that a capped RNA, not related in sequence to Ul snRNA, was also able to inhibit Ul snRNA export in a concentration-dependent manner. Similar results were obtained when the export of ASm mutant versions of U2 and U5 snRNAs were analyzed. Increasing amounts of capped Ul snRNA progressively inhibited export of these two U snRNAs, whereas an ApppG-capped Ul snRNA was not inhibitory at the maximal concentration tested. Export of U2 and U5 snRNAs could also be competed with by increasing amounts of the unrelated RNA mentioned above when an m^GpppG cap structure was present at its 5' end. Together, these results strongly suggest that a cap recognition factor(s) is generally involved in U snRNA export. As previously observed for the RNA with a trimethyl cap (Hamm and Mattaj, 1990), an RNA with the nonphysiological ApppG cap structure could still be exported from the nucleus but with slower kinetics, when injected in the absence of a competitor. This result can best be reconciled with the competition results in terms of relative affinities of the cap binding factor involved in U snRNA export. The prediction would be that the cap binding factor should be able to interact with the RNAs with nonphysiological cap structures, but with a reduced affinity, resulting in slower export kinetics. This prediction was strongly supported by the observation that export of RNAs with nonphysiological cap structures can also be competitively inhibited by m^GpppG capped RNAs (Jarmolowski et al., 1994, and unpublished data). It is not yet clear, however, whether the cap binding factor is able to recognize nonphysiological caps directly, although with a lower affinity, or is recruited onto the RNP export substrate via protein-protein interactions. In conclusion, these results suggest that the physiological cap is not required for export per se, but it has a kinetic effect on this process. In contrast, the cap recognition factor appears to be essential.

X. CHARACTERIZATION OF THE NUCLEAR CAP BINDING ACTIVITY INVOLVED IN U snRNA EXPORT Having established the role of the cap structure in U snRNA export and the involvement of a cap recognition factor in this process, the next step was the identification of such a factor. Several nuclear and cytoplasmic

Nuclear Transport of U snRNPs

141

cap binding proteins (CBPs) had been described in extracts of mammalian, plant, and yeast cells (Sonenberg, 1981; Patzelt et al., 1983; Rozen and Sonenberg, 1987; Ohno et al, 1990), the best studied being the eukaryotic translation initiation factor 4E (eIF-4E; for reviews see Shatkin, 1985; Rhoads, 1988; Sonenberg, 1988). It was important to determine whether the cap binding activity (CBA) involved in nuclear transport was related to any of these CBPs and in particular to the translation initiation factor eIF-4E, which, in spite of its cytoplasmic function, is also found in low concentration in the nucleus (Lejbkowicz et al., 1992). To investigate this question a series of cap analogs, whose affinity for human eIF-4E had previously been determined, were tested for their ability to inhibit nuclear export of Ul ASm snRNA transcripts in Xenopus oocytes (Izaurralde et al., 1992). Five dinucleotide cap analogs, m^GpppG, Et^GpppG, m^'^GpppG, m^'^'^GpppG, and GpppG, and the mononucleotide m^'^GMP were tested by microinjection into oocytes transcribing the Ul ASm gene. In oocytes injected with either m^GpppG or Et^GpppG, UlASm transcripts were retained in the nucleus. Injection of GpppG, m^'^GpppG, m^'^'^GpppG or m^'^GMP did not affect the cellular distribution of UlASm transcripts, which accumulated exclusively in the cytoplasm. Subsequently, similar results were obtained when instead of using dinucleotide competitors, the various cap analogs were incorporated into an unrelated RNA, which was then used as a competitor (Jarmolowski and Izaurralde, unpublished data). These competition experiments showed that, in addition to 7methyl diguanosine triphosphate, the only N-7 substituted cap analog able to inhibit nuclear export of Ul snRNA transcripts was 7-ethyl diguanosine triphosphate. The dinucleotide m^'^GpppG was not able to inhibit Ul snRNA export, whereas it has been shown to be more efficient than Et^GpppG in binding to eIF-4E and inhibiting protein synthesis (Darzynkiewicz et al., 1989; Carberry et al., 1990). In addition, neither m^'^GMP nor other mononucleotides that bind tightly to eIF-4E in vitro inhibited nuclear export of UlASm transcripts (unpublished data). These results demonstrated that the CBA involved in RNA export does not exhibit the same binding specificity as eIF-4E. In contrast, an 80-kDa nuclear cap binding protein (CBP80) from human cell nuclei initially identified and purified by Ohno et al. (1990) did exhibit the same cap analog recognition specificity as the CBA involved in Ul snRNA nuclear export (Izaurralde et al., 1992). The only dinucleotides shown to efficiently inhibit binding of CBP80 to a capped RNA probe in vitro were m^GpppG and Et^GpppG. All other dinucleotides and mononucleotides

142

ELISA IZAURRALDE, IAIN W. MATTAJ, and DAVID S. GOLDFARB

tested were inefficient competitors. Whether CBP80, or other nuclear CBPs, plays a role in U snRNA export should soon be clarified through isolation and characterization of the proteins and their cognate genes.

XL NUCLEAR IMPORT OF U snRNPs The presence of large pools of snRNP proteins in the cytoplasm of Xenopus oocytes has been of enormous advantage for the study of U snRNP assembly and transport in vivo. In vitro synthesized U snRNAs injected into the oocyte cytoplasm are assembled into snRNP core particles by association with the Sm proteins and subsequently transported into the oocyte nuclei. Therefore, the Xenopus oocyte offers the possibility of uncoupling U snRNA import from export and directly studying the requirements for U snRNP nuclear uptake. Recently, a major breakthrough in the study of U snRNP import has been achieved with the development of an in vitro reconstitution system for snRNP particles (Sumpter et al., 1992) and the adaptation of preexisting in vitro and in vivo systems for protein nuclear import to the study of U snRNP import (Fischer et al., 1993, 1994; Marshallsay and Luhrmann, 1994). The availability of these two systems made possible the study of U snRNP assembly and transport in mammalian cells. Comparison and synthesis of the results emerging from the amphibian and mammalian systems will be essential to unraveling the basic mechanisms involved in this process. U snRNP nuclear import, in common with import of karyophilic proteins, exhibits the characteristics of a receptor-mediated process, such as temperature dependence and saturability (Fischer et al, 1993,1994). Earlier studies of U snRNA import had shown that mutation of the Sm binding site, but not other regions of U2 snRNA, including those required for stable binding of U2-specific proteins, affected the nuclear accumulation of U2 snRNA (Mattaj and De Robertis, 1985). Binding of Sm proteins was also found to be required for cap hypermethylation (Hernandez and Weiner, 1986; Mattaj, 1986). Because binding of the Sm proteins was essential for both nuclear migration and cap trimethylation, it was not clear whether the trimethyl guanosine cap structure, the Sm proteins, or both elements were necessary for nuclear migration. Subsequently, it has been shown that, in Xenopus oocytes, hypermethylation of the cap structure and binding of Sm proteins are both required to direct the assembled Ul and U2 snRNPs back to the nucleus. Evidence for the involvement of the trimethyl cap (m^'^'^GpppN) in signaling the import of Ul snRNP came from various experiments

Nuclear Transport of U snRNPs

143

(Fischer and Liihrmann, 1990; Hamm et al., 1990). First, after injection into the cytoplasm, Ul snRNAs with ApppG 5' termini were not transported to the nucleus, unUke Ul snRNAs with an intact trimethyl cap. All of these RNAs were, however, apparently able to assemble normally with both the core and the specific Ul snRNP proteins. Second, Ul snRNA transport was specifically inhibited after removal of its 5' terminus via RNase H cleavage mediated by a DNA oligonucleotide complementary to the Ul 5' end. Third, Ul snRNP nuclear import was competitively inhibited by coinjection into Xenopus oocytes of a large excess of the dinucleotide m^'^'^GpppG but not of the m^GpppG dinucleotide. Furthermore, antibodies directed against the trimethyl cap, or oxidation of the ribose ring of the trimethyl cap by treatment with periodate, severely inhibited nuclear transport of Ul snRNPs. These data clearly established that a factor(s) recognizing the trimethyl cap of Ul snRNA is required to ensure movement of the assembled snRNP back to the nucleus. This factor can be specifically saturated by the trimethyl cap analog but not by the monomethyl dinucleotide. Interestingly, in contrast to the inability of mononucleotides to inhibit nuclear export, the mononucleotide m^'^'^GTP was able to inhibit Ul snRNA import with the same efficiency as the trimethyl dinucleotide m^'^'^GpppG (unpublished data), suggesting a differential mode of recognition of monomethyl and trimethyl cap structures by their respective "receptors." The same authors also showed that the trimethyl cap is not the only feature of the snRNPs recognized by the cellular import machinery. In vitro synthesized Ul snRNA mutants unable to bind to the core proteins but carrying a trimethyl cap structure were not transported into the nucleus after injection into the oocyte cytoplasm. These experiments clearly demonstrated that the Sm proteins not only ensure hypermethylation; they also participate directly in the nuclear targeting of Ul snRNA. Confirmation, and even stronger evidence, of the role of Sm proteins as part of the U snRNP nuclear localization signal came from a series of elegant experiments described by Fischer et al. (1993) using in vitro assembled snRNP core particles. These particles were shown to accumulate in oocyte nuclei with the same kinetics and requirements similar to those of core particles assembled in vivo by injecting U snRNAs into the oocyte cytoplasm. By taking advantage of the availability of these two alternative in vivo and in vitro assembly systems, it was possible to petform competition studies. Oocytes were pre-injected with increasing amounts of unlabeled U snRNAs with different cap structures. These RNAs associate, as expected, with the large pool of Sm proteins present

144

ELISA IZAURRALDE, IAIN W. MATTAJ, and DAVID S. GOLDFARB

in the oocyte cytoplasm to form competitor core particles. Subsequently, radioactivelly labeled, in vitro assembled particles containing trimethyl capped RNAs were injected into the cytoplasm and their nuclear accumulation monitored. Inhibition of import of labeled particles was observed even when the competitor particles lacked the trimethyl cap structure. Therefore, like the dinucleotide m^'^'^GpppG, the core particles are able to saturate a limiting component required for U snRNP nuclear accumulation. This putative U snRNP transport receptor recognizes the nuclear localization signal in the core particles independently of the cap structure. In summary, the nuclear localization signal of Ul snRNP in Xenopus oocytes results from the contribution of two elements, the trimethyl cap structure on the RNA and the bound core proteins. The nature of the receptor(s) interacting with the two elements of this bipartite nuclear localization signal is not yet known. It is also not known which features of the core particles are recognized in this process (see below).

XII. THE REQUIREMENT OF THE TRIMETHYL CAP STRUCTURE FOR U snRNA NUCLEAR IMPORT IS NOT GENERAL The requirement of the trimethyl cap structure for Ul snRNP import is not general for all spliceosomal U snRNPs or for all cell types. In Xenopus oocytes, U2 snRNP behaves like Ul and its transport can be inhibited by the dinucleotide m^'^'^GpppG, but U4 and U5 are not as sensitive as Ul or U2 snRNPs to the presence of the trimethylated dinucleotide (Fischer et al, 1991). Furthermore, unlike Ul and U2, whose nuclear import strictly depends on their having a trimethyl cap, U4 and U5 have a much less stringent requirement and are transported into the nucleus, although with slower kinetics, even when the nonphysiological ApppG cap structure is present at their 5' termini. Thus, although the trimethyl cap structure is not essential for nuclear uptake in the case of U4 and U5 snRNPs, its presence does affect the rate of import (Fischer etal., 1991). This observation suggests that the cap functions as an accessory signal to increase the efficiency of nuclear import. In the absence of the cap, the nuclear localization signal present on the Sm core domain would be sufficient to allow nuclear accumulation of the U5 snRNP, but not of the Ul or U2 snRNPs. The possibility that the different core domains are not functionally equivalent can be excluded by the observation that U5

Nuclear Transport of U snRNPs

145

snRNP transport can be inhibited by an excess of Ul snRNP core particles and vice versa, suggesting that their transport is mediated by a common receptor that recognizes similar features on the Sm core domain of both snRNPs (Fischer et al., 1993). The reason for the differential requirement for the cap structure is not yet completely understood. Possible clues were provided by the finding that a minimal Ul snRNP particle reconstituted in vitro on an RNA lacking the 5'-terminal stem loops of Ul RNA (stem loops I to III) was targeted to the nucleus, in spite of the absence of a trimethyl cap (Fischer et al., 1993), and that precise deletion of stem loops I and II from Ul snRNA resulted in an snRNP whose nuclear transport was trimethyl cap independent (Jarmolowski and Mattaj, 1993). Thus, in some ill-defined way, the structure of the 5^ region of Ul snRNA is responsible for the trimethyl cap requirement of U1 snRNA transport. It is not obvious how to interpret these observations in mechanistic terms. The absolute requirement for a trimethyl cap structure for Ul and U2 in the oocyte system is not maintained in somatic cells. This observation was made by microinjection of either Ul snRNAs or in vitro assembled Ul snRNP particles into the cytoplasm of Vero cells (African green monkey kidney cells) and mouse fibroblast 3T3 cells (Fischer et al., 1994). In these somatic cells a Ul particle with an ApppG cap instead of the trimethyl cap was imported into the nucleus. The ApppG-capped snRNP was accumulated in the nucleus at a reduced rate, however, indicating that the trimethyl cap, although dispensable, retained a signaling role for nuclear targeting of Ul snRNP. The finding that Ul snRNP particles were transported in a very similar way in the two somatic cell types tested suggested that this behavior may apply generally for all somatic cells, irrespective of their origin. It will be necessary, however, to test Xenopus somatic cells before accepting the conclusion that the difference is between somatic and germ cells rather than between species. Thus, in somatic cells the cap appears to function as an accessory signal to increase the efficiency of Ul snRNP nuclear import, as was the case for U4 and U5 snRNPs in the oocyte system. Perhaps the most remarkable result obtained so far in this context is the observation that, in vitro, somatic cell extracts confer trimethyl cap independence on Ul snRNP accumulation in the nucleus, whereas oocyte extracts confer cap dependence (Marshallsay and Liihrmann, 1994). Thus, the difference between the cell types reflects functional differences in cytoplasmic factors required for snRNP nuclear transport (see below).

146

ELISA IZAURRALDE, IAIN W. MATTAJ, and DAVID S. GOLDFARB

XIII. A UNIFYING MODEL FOR THE ROLE OF THE TRIMETHYL CAP AND THE CORE DOMAIN IN SIGNALING U snRNP NUCLEAR IMPORT Since the Sm core domain is both necessary and, for some U snRNPs or in certain cell types, sufficient to direct U snRNP particles into the nucleus, it is likely that an absolutely essential snRNP NLS is located on the core domain and is created by the interaction of the Sm proteins with the U snRNA. Several possibilities can be envisaged; for example, a nuclear localization signal may preexist in the core proteins and become exposed upon binding to the RNA. It is also possible that binding of the core proteins may induce a conformational change in the RNA and that some features on the RNA rather than the proteins are recognized by factors that mediate nuclear import. These two possibilities are not mutually exclusive, and the nuclear localization signal may be generated by the contribution of the two constituents of the core snRNP, the core proteins, and the U snRNA. The trimethyl cap may act as an accessory signal to increase the efficiency of nuclear transport. In oocytes, transport of Ul snRNP is much less rapid than in somatic cells, and therefore the role of the cap may become essential because of a simple difference in transport efficiency. In somatic cells, the role of the cap would be limited to an accessory function since transport is more efficient than in oocytes and therefore, even a suboptimal NLS may be recognized by the transport receptor. The higher efficiency of the transport apparatus in somatic cells versus oocytes could be due to a stronger interaction of the receptor with the Sm core domain or, alternatively, be a consequence of higher receptor concentration in these cells or the presence of different forms of, or perhaps even entirely different, receptors in the two cell types. The observation that the concentration of Ul snRNP required to saturate its own transport in somatic cells is 5-10-fold higher than in the oocytes supports the possibility that effective receptor concentration is higher in somatic cells (Fischer et al., 1994). Note that it is not yet clear for either cell type whether there are two distinct snRNP receptors or one unique factor that recognizes simultaneously both parts of the bipartite snRNP NLS. In mechanistic terms the trimethyl cap may increase the affinity of the receptor recognizing the Sm domain by interacting directly with this receptor, with an additional factor, or directly with one or several of the Sm proteins bound to the U snRNA to influence the formation of the core NLS. Clearly, a major goal for future studies will be the identification of

Nuclear Transport of U snRNPs

147

the receptor system that recognizes the bipartite nuclear targeting signal and investigation of the nature of this recognition.

XIV. NUCLEAR IMPORT OF U6 snRNP AFTER MITOSIS The U snRNP import mechanism described above is Ukely to be valid for the reimport of the particles into the nucleus after mitosis as well as during interphase. In the case of U6 snRNP, which does not have a transient cytoplasmic phase, a nuclear import signal was nevertheless shown to exist, presumably to ensure the reaccumulation of the free U6 snRNP in the nucleus after mitosis. The import of U6 snRNP particles has been investigated by microinjection into the oocyte cytoplasm of in vitro transcribed U6 RNA (Hamm and Mattaj, 1989). A region of U6 snRNA (nucleotides 20-24) required for nuclear migration has been proposed to be the binding site of a U6 protein. This protein has not yet been identified. However, the observation that the nuclear targeting signal of U6 particles appears to be similar in nature to the NLSs of karyophilic proteins leads to the suggestion that U6 RNA may be directed into the nucleus via a NLS carried by this protein. The role of the 5' terminus of U6 RNA in signaling nuclear import has also been addressed. Nuclear accumulation of U6 transcripts does not require the presence of a y-methyl triphosphate at the 5' end of the U6 RNA. U6 RNAs having several different cap structures accumulate in the nucleus, the single exception being U6 transcripts with a monomethyl cap structure (m'^GpppG), which are retained in the cytoplasm. Cytoplasmic retention of these monomethyl-capped U6 RNAs could be overcome by coinjection of the dinucleotide m^GpppG, suggesting that a cap binding activity was responsible for anchorage of the RNA in the cytoplasm (Fischer et al., 1991).

XV. THE ROLE OF CAP STRUCTURES IN SIGNALING RNA SUBCELLULAR LOCALIZATION The observation that a U6 snRNA with a m^GpppG cap structure is retained in the cytoplasm, and that this retention can be relieved by coinjection of m^GpppG dinucleotide (Fischer et al., 1991), points to the existence of several different roles for cap structures in targeting RNAs to different subcellular compartments. The m^GpppN cap structure seems to be part of the RNA export signal in the nucleus, whereas in the cytoplasm, it promotes cytoplasmic retention. To allow U snRNPs to

148

ELISA IZAURRALDE, IAIN W. MATTAJ, and DAVID S. GOLDFARB

migrate efficiently back to the nucleus this structure has to be modified by the hypermethylation described for the pol II U snRNPs. Indeed, cytoplasmically microinjected U6 RNA with a trimethyl cap is imported into the nucleus. In this case the trimethyl cap is not a positive signal for import, since this process is insensitive to the presence of a large excess of the dinucleotide m^'^'''GpppG. This finding demonstrates that hypermethylation reverses the anchorage in the cytoplasm caused by the monomethyl cap. Therefore hypermethylated caps can act in some cases as an essential part of the nuclear import signal (e.g., in Ul and U2 import in oocytes) and influence the rate of import in other snRNPs and cell types, but more generally, hypermethylation may prevent interaction with cytoplasmic cap binding proteins to overcome the resulting anchorage of the RNA in the cytoplasm.

XVI. INHIBITORS OF NUCLEAR TRANSPORT In this section we will focus on the uses of transport inhibitors to elucidate aspects of nuclear transport. Generally, inhibitors that reduce the function of the nuclear pore complex inhibit the rate of all mediated import and export processes. The passive diffusion of molecules through the 10-nm aqueous channels is affected by only a single inhibitory antibody (see below). In contrast, reagents that act on targeting steps are usually more selective and may only inhibit the transport of one kind of substrate. We will deal first with general inhibitors of nuclear transport. The depletion of a cell's energy charge with metabolic inhibitors such as azide and 6-deoxyglucose arrests protein import at a stage after targeting. Thus, fluorescence-labeled substrates accumulate at the external surface of the nuclear envelope of cells depleted of ATP. Where studied, the export of macromolecules, such as ribosomal subunits or mRNA, is inhibited by ATP depletion (Battaille et al., 1990; Dargemont and Kiihn, 1992). Studies such as these have led researchers to conclude that bidirectional translocation through the nuclear pore complex requires phosphate bond energy (Forbes, 1992). The energy-requiring step(s) in translocation have not, however, been identified. Translocation, but not targeting, is also inhibited by chilling (Richardson et al., 1988). Sensitivity to chilling does not necessarily indicate that transport is blocked at an ATP-requiring step. Most enzyme-catalyzed reactions are slowed when incubated on ice, and the nuclear pore complex can certainly be viewed as an enzyme catalyst. Rather, chilling sensitivity indicates that the translocation reactions have a significant positive

Nuclear Transport of U snRNPs

149

energy of activation. Targeting stages, because they are insensitive to both ATP depletion and chilling, apparently involve only simple diffusion and binding reactions. The role of phosphate bond energy in nuclear transport has recently become complicated by observations in cell-free nuclear transport assays. The addition of nonhydrolyzable GTP analogs inhibited targeting and GDP analogs inhibited translocation (Moore and Blobel, 1993; Melchior et al., 1993). These effects have been attributed to the RanyTC4 GTPase (Goldfarb, 1994, and references therein) and raise the possibility that GTP hydrolysis instead of, or in addition to, ATP hydrolysis is required for transport. Competitive inhibition of transport can be achieved by injecting high concentrations of import or export substrates that titrate or saturate the transport apparatus. A distinguishing feature of competition kinetics is that under conditions of saturation, the transport apparatus runs at maximum velocity. Under saturating conditions of maximum transport rate, the mole percentage of available substrate that is transported per unit time is decreased, a condition that provides for competition among substrates. Saturation kinetics has been shown for tRNA (Zasloff, 1983), U snRNA (Jarmolowski et al., 1994), mRNA (Dargemont and Kiihn, 1992; Jarmolowski et al., 1994) and ribosome export (Bataille et al, 1990), and protein and U snRNP import (Goldfarb et al., 1986; Fischer et al., 1993). Saturation kinetics is a classical characteristic of receptormediated processes. For protein and U snRNP import, saturation appears to occur during an early targeting step, presumably at the level of NLS recognition in the cytoplasm. The use of competition kinetics to explore U snRNP targeting is described below. Two related classes of inhibitors, lectins and anti-nucleoporin antibodies, bind directly to pore complex nucleoporins. Wheat germ agglutinin (WGA) and related lectins bind to a large number nucleoporins that are modified in the cytoplasm with O-linked N-acetylglucosamine (OGlcNAc) (Forbes, 1992). Immunogold labeling studies showed that WGA, which inhibits both import and export, binds to both sides of the nuclear pore complex (Starr and Hanover, 1992). The specificity of WGA action is uncertain because many other intracellular proteins in addition to the GlcNAc-containing nucleoporins are also modified with GlcNAc residues and their binding by WGA could cause pleiotropic effects. This problem is partially abrogated by the fact that WGA, a dimer, has significantly higher avidity for clustered GlcNAc residues such as those found at the pore complex than for the isolated GlcNAc residues found on most other proteins.

1 50

ELISA IZAURRALDE, IAIN W. MATTAJ, and DAVID S. GOLDFARB

The inhibitory anti-nucleoporin monoclonal antibodies bind to groups of nucleoporins that contain common epitopes (Featherstone et al., 1988; Forbes, 1992; Starr and Hanover, 1992). These nucleoporin subgroups represent subsets of those bound by WGA, which is more general, so it is likely that the two inhibitors work by the same mechanism. However, not all anti-nucleoporin antibodies inhibit nuclear transport. The mechanism of action of WGA and anti-nucleoporin antibodies may be due to steric effects. Steme-Marr et al. (1992) demonstrated that WGA interferes with the binding of uncharacterized transport factors to the pore complex. Also, because WGA and anti-nucleoporin antibodies are bivalent, it was suggested that they inhibit transport by cross-linking adjacent pore complex subunits and mechanically restricting their activity (Akey and Goldfarb, 1989). Neither WGA nor anti-nucleoporin antibodies inhibit the passive diffusion of small microinjected proteins or dextrans. Thus these reagents do not occlude the 10-nm diffusion channels. Second-generation anti-nucleoporin antibodies, with specificities for single nucleoporins, have been developed and should serve as more specific probes into pore complex function (Cordes et al, 1993; Wilken et al., 1993). Interestingly, some karyophiles are significantly less sensitive to WGA inhibition than others. The import of U6 snRNA, for example, is several times more sensitive to WGA than is Ul snRNAimport (Michaud and Goldfarb, 1992). Variation in sensitivity to WGA might reflect the use of different transport factors or disparate pore complex docking sites. Alternatively, the differences may be due to steric factors and/or biochemical characteristics of the karyophiles such as ionic charge or hydrophobicity. A unique inhibitor of nuclear transport is a monoclonal IgG directed against a linear peptide epitope of gp210, an integral membrane protein that is thought to function as a membrane anchor for the annulus of the pore complex (Greber et al., 1990). When mRNA encoding anti-gp210 IgG was microinjected into tissue culture cells, the mRNA was translated and the nascent IgG chains secreted into the lumen of the ER, where they bound gp210 and inhibited transport (Greber and Gerace, 1992). The fact that an antibody could inhibit nuclear transport by binding to the lumenal portion of a protein, situated across the nuclear membrane and at the extreme periphery of the pore complex, suggests that the entire 125-MDa pore complex structure is conformationally active during transport. The anti-gp210 monoclonal is unique in that it inhibits both NLS-mediated transport and passive diffusion through the 10-nm diffusion channels.

Nuclear Transport of U snRNPs

151

XVII. KINETIC EVIDENCE FOR MULTIPLE NUCLEAR TARGETING PATHWAYS IN U snRNP IMPORT Most nuclear proteins contain amino acid sequences that are at least vaguely related to the archetypal SV40 large T-antigen NLS (PKKKRKV) or the nucleoplasmin NLS, which comprises two domains of basic amino acids separated by a spacer of 10 amino acids (KRPAATKKAGQAKKKK)(Garcia-Bustosetal., 1991;Forbes, 1992). The single basic domain T-antigen NLS will target proteins to nuclei of yeast (Silver et al., 1989) and Tetrahymena (White et al., 1989), demonstrating that fundamental aspects of nuclear targeting are conserved among eukaryotes. Robbins et al. (1991) suggested that the T-antigen NLS may be an extreme single-ended variant of the bipartite nucleoplasmin NLS, and it is likely that the double basic domain NLS motif is most common. The import of synthetic peptide NLS-serum albumin conjugates, such as P(Lys)-BSA, which is based on the SV40 T-antigen NLS, was shown to be kinetically saturable in Xenopus oocytes (Goldfarb et al., 1986). The microinjection of saturating concentrations of P(Lys)-BSA was later shown to compete with the import of most nuclear proteins, indicating that each of these proteins is imported by the same targeting apparatus, regardless of whether they carry single or double basic domain NLSs (Michaud and Goldfarb, 1993). Thus, even though the amino acid sequences of basic domain NLSs vary considerably, sequence differences probably reflect only quantitative variation in NLS "strength" rather than differences in the targeting apparatuses that mediate their import. Although these studies are consistent with a single predominant targeting apparatus, they could not rule out the existence of multiple NLS receptors having relatively broad overlapping sequence specificities for different NLS subfamilies. Early competition kinetic experiments between basic domain NLScontaining proteins and U snRNPs indicated that these two classes of karyophiles use distinct targeting apparatuses (Michaud and Goldfarb, 1991). Here, the import of U2 snRNP was unaffected by saturating concentrations of P(Lys)-BSA. The converse competition experiment, using free m^'^'^GpppG cap dinucleotide, showed that the saturation of cap-dependent import had no effect on basic domain NLS import (Fischer et al., 1991; Michaud and Goldfarb, 1992). Together, these studies provided the kinetic basis for the idea that the m^'^'^GpppN cap directed U snRNPs along a distinct targeting pathway. One important interpreta-

152

ELISA IZAURRALDE, IAIN W. MATTAJ, and DAVID S. GOLDFARB

tion of these data is the conclusion that the saturable steps in import lay upstream of the nuclear pore within the targeting pathways (Goldfarb and Michaud, 1991). The idea that the m^'^'^GpppN cap was a novel signaUng moiety was directly supported by experiments demonstrating that the import of Ul snRNP into oocyte nuclei depended on the m^'^'^GpppN cap, as discussed in preceding sections. More recent results, also described above, indicated that the m^'^'^GpppN cap is not always essential for import in somatic cells or in oocytes. However, whereas the nuclear import of Ul-type snRNPs in somatic cells is less stringently dependent on the m^'^'^GpppN cap than in oocytes, it still appears to follow a pathway that is independent of the basic domain NLS targeting pathway. This conclusion came from studying import in tissue culture cells that expressed a cytoplasmically anchored mutant SV40 T-antigen (FS T antigen) (van Zee et al., 1993). FS T antigen is membrane-bound and appears to inhibit the nuclear import of basic domain NLS-containing karyophiles by sequestering targeting factors that bind to its NLS. High levels of cytoplasmic FS T antigen interfered with the nuclear import of basic domain NLS-containing proteins, but not with the import of Ul snRNP. It will be interesting to learn if the import of a microinjected ApppG-capped Ul snRNA, which is imported slowly in tissue culture cells, is sensitive to a FS T antigen import block. A key to resolving the role of 5' cap structures in U snRNP import will be the biochemical identification and functional characterization of the m^'^'^GpppN cap binding protein.

XVIII. MULTIPLE U snRNP IMPORT PATHWAYS Complexity in U snRNP targeting pathways was discovered when competition analyses were extended to U snRNPs that differed from the Ul-type U snRNPs in their biochemical composition and cap structure. In oocytes, the import of Ul, U2, U4, and U5 snRNPs (Ul-type snRNPs) was distinguished from the import of basic domain-containing proteins by the competitive effects of free m^'^'^GpppG cap dinucleotide and P(Lys)-BSA. U6 snRNP showed the converse behavior, being saturable by P(Lys)-BSA but not by m^'^'^GpppG, much like NLS-containing proteins. These results suggest that U6 snRNA import is mediated by a NLS-containing U6 snRNA-binding protein. Because U6 snRNP is assembled in the nucleus and is present in the cytoplasm only during mitosis (or following microinjection), the cell's predominant protein import pathway appears to serve the trafficking needs of U6 snRNP.

Nuclear Transport of U snRNPs

153

U3 snoRNP contains a 3mGpppG cap but is not complexed by the Sm proteins. In oocytes U3 snoRNP is imported by a cap-independent pathway (Baserga et al., 1992), and the import of SmGpppG-capped U3 snRNA is not competed with by cap dinucleotide. The cap independence of U3 snRNA import may be related to the fact that U3 snoRNA does not bind Sm proteins, which are essential for the import of Ul-type U snRNPs. Thus the unique import properties of U3 snoRNA reinforce the conclusion that both the cap and Sm proteins play physiological roles in Ul-type snRNP import. The import of U3 snoRNP is also not competed with by saturating concentrations of P(Lys)-BSA, suggesting that it is not imported by the basic domain NLS pathway used by U6 snRNP (Michaud and Goldfarb, 1992). These results suggest that U3 snoRNP may be imported by a specialized targeting apparatus, although it is unclear why this would be necessary. In conclusion, kinetic competition studies suggest that U snRNPs may be imported along three distinct import pathways, depending on the cap structure of the snRNA and the proteins that comprise the karyophilic U snRNP. These studies, and those of van Zee et al. (1993), predict that the import of Ul-type snRNPs will be mediated by transport factors that are distinct from those used to import NLS-containing proteins. The test of this hypothesis will probably come from in vitro fractionation and reconstitution studies of U snRNP import (Marshallsay and Liihrmann, 1994). This powerful approach will allow the identification of cytoplasmic components that direct the nuclear transport of U snRNPs. Cell extract-dependent in vitro transport systems should also lead to the elucidation of the exact roles of the 3mGpppG cap structure and Sm proteins in Ul-type snRNP import. REFERENCES Adam, S.A. & Gerace, L. (1991). Cytosolic proteins that specifically bind nuclear location signals are receptors for nuclear import. Cell 66, 837-847. Adam, S.A., Lobl, T.L., Mitchell, M.A., & Gerace, L. (1989). Identification of specific binding proteins for a nuclear location sequence. Nature 337, 276-279. Adam, S.A., Steme-Marr, R.E., & Gerace, L. (1990). Nuclear protein import in permeabihzed mammalian cells requires soluble cytoplasmic factors. J. Cell Biol. I l l , 807-816. Akey, C. & Goldfarb, D.S. (1989). Nuclear import through the nuclear pore complex is a multi-step process. J. Cell Biol. 109, 971-982. Akey, C.W. & Radermacher, M. (1993). Architecture of the Xenopus nuclear pore complex revealed by three-dimensional cryo-electron microscopy. J. Cell Biol. 122, 1-19.

154

ELISA IZAURRALDE, IAIN W. MATTAJ, and DAVID S. GOLDFARB

Andersen, J. & Zieve, G.W. (1991). Assembly and intracellular transport of snRNP particles. Bioessays 13, 57-64. Baserga S.J., Gilmore-Hebert M., & Yang E.W. (1992). Distinct molecular signals for nuclear import of the nucleolar snRNA, U3. Genes Dev. 6,1120-1130. Bataille, N., Helser, T, & Fried, H.M. (1990). Cytoplasmic transport of ribosomal subunits microinjected into the Xenopus laevis oocyte nucleus: a generalized, facilitated process. J. Cell Biol. HI, 1571-1582. Borer, R.A., Lehner, C.F., Eppenberger, H.M., & Nigg, E.A. (1989). Major nucleolar proteins shuttle between nucleus and cytoplasm. Cell 56, 379-390. Branlant, C , Krol, A., Ebel, J.R, Lazar, E., Haendler, B., & Jacob, M. (1982). U2 RNA shares a structural domain with Ul, U4 and U5 RNAs. EMBO J. 1, 1259-1265. Breeuwer, M. & Goldfarb, D.S. (1990). Facilitated nuclear transport of histone HI and other nucleophilic proteins. Cell 60, 999-1008. Bukrinsky, M., Haggerty, S., Dempsey, M., Sharova, N., Adzhubei, A., Spitz, L., Lewis, P., Goldfarb, D., Emerman, M., & Stevenson, M. (1993a). A nuclear localization signal within HIV-1 matrix protein that governs infection of non-dividing cells. Nature 365,666-669. Bukrinsky, M.I., Sharova, N., McDonald, T.L. Pushkaraskaya, T. Tarpley, W.G., and Stevenson, M. (1993b) Association of HIV-1 integrase, matrix, and reverse transcriptase antigens with viral nucleic acids following acute infection. Proc. Natl. Acad. Sci. USA 90, 6125-6129. Carberry, S.E., Darzynkiewicz, E., Stepinski, J., Tahara, S.M., Rhoads, R.E., & Goss, D.J. (1990). A spectroscopic study of the binding of N-7-substituted cap analogs to human protein synthesis initiation factor 4E. Biochemistry 29, 3337-3341. Cordes, V.C, Reidenbach, S., Kohler, A., Sniurman, N., van Driel, R., & Franke, W.W. (1993). Intranuclear filaments containing a nuclear pore complex protein. J. Cell Biol. 123,1333-1344. Dargemont, C. & Kuhn, L.C. (1992). Export of mRNA from microinjected nuclei of Xenopus leavis oocytes. J. Cell Biol. 118, 1-9. Darzynkiewicz, E., Stepinski, J., Ekiel, I., Goyer, C , Sonenberg, N., Temeriusz, A., Jin, Y, Sijuwade, T., Haber, D., & Tahara, S. M. (1989). Inhibition of eukaryotic translation by nucleoside 5'-monophosphate analogues of mRNA 5'-Cap: changes in N7 substituent affect analogue activity. Biochemistry 28, 4771-4778. De Robertis, E.M., Lienhard, S., & Parisot, R.F (1982). Intracellular transport of microinjected 5S and small nuclear RNAs. Nature 295, 572-577. Dingwall, C , Shamick, S.V., & Lasky, R.A. (1982). A polypeptide domain that specifies migration of nucleoplasmin into the nucleus. Cell 30,449-453. Dworetzky, S.I., Lanford, R.E., & Feldherr, CM. (1988). The effect of variations in the number and sequence of targeting signals on nuclear uptake. J. Cell Biol. 107, 1279-1288. Eliceiri, G.L. (1974). Short-lived, small RNAs in the cytoplasm of HeLa cells. Cell 3, 11-14. Featherstone, C, Darby, M.K., & Gerace, L. (1988). A monoclonal antibody against the nuclear pore complex inhibits nucleocytoplasmic transport of protein and RNA in vivo. J. Cell Biol. 107, 1289-1297.

Nuclear Transport of U snRNPs

155

Feeney, R.J., Sauterer, R.A., Feeney, J.L., & Zieve, G.W. (1989). Cytoplasmic assembly and nuclear accumulation of mature small nuclear ribonucleo-protein particles. J. Biol. Chem. 264, 5776-5783. Feldherr, CM. & Akin, D. (1990). The permeability of the nuclear envelope in dividing and nondividing cell cultures. J. Cell Biol. I l l , 1-8. Feldherr, CM. & Akin, D. (1991). Signal-mediated nuclear transport in proliferating and growth-arrested BALB/c 3T3 cells. J. Cell Biol. 115, 933-939. Feldherr, CM. & Akin, D. (1993). Regulation of nuclear transport in proliferating and quiescent cells. Exp. Cell Res. 205, 179-186. Fischer, U. & Ltihrmann, R. (1990). An essential signaling role for the m3G cap in the transport of Ul snRNPto the nucleus. Science 249, 786-790. Fisher, D.E., Conner, G.E., Reeves, W.H., Wisniewolski, R., & Blobel, G. (1985). Small nuclear ribonucleoprotein particle assembly in vivo: demonstration of a 6S RNAfree core precursor and posttranslational modification. Cell 42, 751-758. Fischer, U., Darzynkiewicz, E., Tahara, S.M., Dathan, N.A., Liihrmann, R., & Mattaj, I.W. (1991). Diversity in the signals required for nuclear accumulation of U snRNPs and variety in the pathways of nuclear transport. J. Cell Biol. 113, 705-714. Fischer, U., Sumpter, V., Sekine, M., Satoh, T, & Luhrmann, R. (1993). Nucleo-cytoplasmic transport of U snRNPs: definition of a nuclear location signal in the Sm core domain that binds a transport receptor independently of the m3G cap. EMBO J. 12, 573-583. Fischer, U., Heinrich, J., van Zee, K., Fanning, E., & Luhrmann, R. (1994). Nuclear transport of Ul snRNP in somatic cells: differences in signal requirement compared with Xenopus laevis oocytes. J. Cell Biol. 125, 971-980. Forbes, D.J. (1992). Structure and function of the nuclear pore. Annu. Rev. Cell Biol. 8, 495-528. Garcia-Bustos, J., Heitman, J., & Hall, M.N. (1991). Nuclear protein localization. Biochim. Biophys. Acta 1071, 83-101. Goldfarb, D.S. (1991). Shuttling proteins go both ways. Curr. Opin. Biol. 1, 212-214. Goldfarb, D.S. (1992). Are the cytosolic components of the nuclear, ER, and mitochondrial import apparatus functionally related? Cell 70, 185-188. Goldfarb, D. (1994). A GTPase cycle for nuclear transport. Curr. Biol. 4,57-60. Goldfarb, D. & Michaud, N. (1991). Pathways for the nuclear transport of proteins and RNAs. Trends Cell Biol. 1,20-24. Goldfarb, D.S., Gariepi, J., Schoohiik, G., & Romberg, R.D. (1986). Synthetic peptides as nuclear localization signals. Nature 322, 641-644. Greber, U.F. & Gerace, L. (1992). Nuclear protein import is inhibited by an antibody to a lumenal epitope of a nuclear pore complex glycoprotein. J. Cell Biol. 116,15-30. Greber, U.F., Senior, A., & Gerace, L. (1990). A major glycoprotein of the nuclear pore complex is a membrane-spanning polypeptide with a large lumenal domain and a small cytoplasmic tail. EMBO J. 9, 1495-1502. Greber, U.F, Willetts, M., Webster, R, & Helenius, A. (1993). Stepwise dismantling of adenovirus during entry into cells. Cell 75,477-486. Grimm, C , Stefanovic, B., & Schumperli, D. (1993). The low abundance of U7 snRNA is partly determined by its Sm binding site. EMBO J. 12, 1229-1238.

1 56

ELISA IZAURRALDE, IAIN W. MATTAJ, and DAVID S. GOLDFARB

Hackl, W., Fischer, U., & Liihrmann, R. (1994). A69-kD protein that associates reversibly with the Sm core domain of several spliceosomal snRNP species. J. Cell Biol. 124, 261-272. Hamm, J. & Mattaj, I.W. (1989). An abundant U6 snRNPfound in germ cells and embryos of Xenopus laevis. EMBO J. 8,4179-4187. Hamm, J. & Mattaj, I.W. (1990). Monomethylated cap structures facilitate RNA export from the nucleus. Cell 63,109-118. Hamm, J., Kazmaier, M., & Mattaj, I.W. (1987). In vitro assembly of Ul snRNPs. EMBO J. 6, 3479-3485. Hamm, J., Darzynkiewicz, E., Tahara, S.M., & Mattaj, I.W. (1990). The trimethylguanosine cap structure of Ul snRNA is a component of a bipartite nuclear targeting signal. Cell 62, 569-577. Hernandez, N. & Weiner, A.M. (1986). Formation of the 3'-end of Ul snRNA requires compatible snRNA promoter elements. Cell 47, 249-258. Hinshaw, J.E., Carragher, B.O., & Milligan, R.A. (1992). Architecture and design of the nuclear pore complex. Cell 69, 1133-1141. Imamoto, N., Matsuoka, Y, Kurihara, T, Kohno, K., Mihagi, M., Sakiyama, F, Okada, Y, Tsunasawa, S., & Yoneda, Y (1992). Antibodies against 70-kD heat shock cognate protein inhibit mediated nuclear import of karyophilic proteins. J. Cell Biol. 119,1047-1062. Izaurralde, E., Stepinski, J., Darzynkiewicz, E., & Mattaj, I.W. (1992). A cap binding protein that may mediate nuclear export of RNA polymerase Il-transcribed RNAs. J.CellBiol. 118,1287-1295. Jacob, M., Lazar, E., Haendler, B., Gallinaro, H., Krol, A., & Branlant, C. (1984). A family of small nucleoplasmic RNAs with common structural features. Biol. Cell 51,1-10. Jantsch, M.F & Gall, J.G. (1992). Assembly and localization of the Ul-specific snRNP C protein in the amphibian oocyte. J. Cell Biol. 119,1037-1046. Jarmolowski, A. & Mattaj, I.W. (1993). The determinants for Sm protein binding to Xenopus Ul and U5 snRNAs are complex and non-identical. EMBO J. 12, 223-232. Jarmolowski, A., Boelens, W C , Izaurralde, E., & Mattaj, I.W (1994). Nuclear export of different classes of RNA is mediated by specific factors. J. Cell Biol. 124, 627-635. Kambach, C. & Mattaj, I.W (1992). Intracellular distribution of the Ul Aprotein depends on active transport and nuclear binding to Ul snRNA. J. Cell Biol. 118,11-21. Kambach, C. & Mattaj, I.W. (1994). Nuclear transport of the U2 snRNP-specific U2B" protein is mediated by both direct and indirect signalling mechanisms. J. Cell Sci. 107,1807-1816. Kastner, B., Bach, M., & Liihrmann R. (1990). Electron microscopy of small nuclear ribonucleoprotein (snRNP) particles U2 and U5: evidence for a common structure-determining principle in the major U snRNP family. Proc. Natl. Acad. Sci. USA 87, 1710-1714. Konings, D. A.M. & Mattaj, I.W (1987). Mutant U2 snRNAs of Xenopus which can form an altered higher order RNA structure are unable to enter the nucleus. Exp. Cell Res. 172,329-339.

Nuclear Transport of U snRNPs

157

Lejbkowicz, R, Goyer, C , Darveau, A., Neron, S., Lemieux, R., & Sonenberg, N. (1992). A fraction of the mRNA 5' cap-binding protein, eukaryotic initiation factor 4E, localizes in the nucleus. Proc. Natl. Acad. Sci. USA 89, 9612-9616. Lemer, E.A., Lemer, M.R., Janeway, C.A., Jr., & Steitz, J.A. (1979). Monoclonal antibodies to nucleic acid-containing cellular constituents: probes for molecular biology and autoimmune disease. Proc. Natl. Acad. Sci. USA 78,2737-2741. Liihrmann, R., Kastner, B., & Bach, M. (1990). Structure of spliceosomal snRNPs and their role in pre-mRNA splicing. Biochim. Biophys. Acta 1087,265-292. Madore, S.L., Wieben, D.E., & Pederson, T. (1984). Intracellular site of Ul small nuclear RNA processing and ribonucleoprotein assembly. J. Cell Biol. 98, 188-192. Marshallsay, C. & Liihrmann, R. (1994). In vitro nuclear import of snRNPs: cytosolic factors mediate mjG-cap dependence of Ul and U2 snRNP transport. EMBO J. 13,222-231. Mattaj, I.W. (1986). Cap trimethylation of U snRNA is cytoplasmic and dependent in U snRNP protein binding. Cell, 46,905-911. Mattaj, I.W. & De Robertis, E.M. (1985). Nuclear segregation of U2 snRNA requires binding of specific snRNP proteins. Cell 40, 111-118. Mattaj, I.W., Boelens, W., Izaurralde, E., Jarmolowski, A., & Kambach, C. (1993). Nucleocytoplasmic transport and snRNP assembly. Mol. Biol. Rep. 18, 79-83. Melchior, P., Paschal, B., Evans, J., & Gerace, L. (1993). Inhibition of nuclear protein import by hydrolyzable analogues of GTP and identification of the small GTPase Ran/TC4 as an essential transport factor. J. Cell Biol. 123,1649-1660. Michaud, N. & Goldfarb, D.S. (1991). Multiple pathways in nuclear transport: the import of U2 snRNP occurs by a novel kinetic pathway. J. Cell Biol. 112, 215-224. Michaud, N. & Goldfarb, D.S. (1992). Microinjected U snRNAs are imported to oocyte nuclei via the nuclear pore complex by three distinguishable targeting pathways. J. Cell Biol. 116,851-862. Michaud, N. & Goldfarb, D.S. (1993). Most nuclear proteins are imported by the same pathway. Exp. Cell Res. 208,128-136. Moore, M.S. & Blobel, G. (1993). The two steps of nuclear import, targeting to the nuclear envelope and translocation through the nuclear pore, require different cytosolic factors. Cell 69,939-950. Moore, M.S. & Blobel, G. (1992). The GTP-binding protein Ran/TC4 is required for protein import into the nucleus. Nature 365, 661-663. Moore, M.J., Query, C.C, & Sharp, PA. (1993). Splicing of precursors to messenger RNAs by the spliceosome. In: The RNA World (Gesteland, R., and Atkins, J.F., eds.), pp. 303-357. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY. Neuman de Vegvar, H.E. & Dahlberg, J.E. (1990). Nucleocytoplasmic transport and processing of small nuclear RNA precursors. Mol. Cell Biol. 10, 3365-3375. Newmeyer, D.D. & Wilson, K.L. (1991). Egg extracts for nuclear import and nuclear assembly reactions. Methods Cell Biol. 36, 599-626. Patzelt, E., Blaas, D., & Kuechler, E. (1983). CAP binding proteins associated with the nucleus. Nucleic Acids Res. 11,5821-5835. Pinol-Roma, S. & Dreyfuss, G. (1992). Shuttling of pre-mRN A binding proteins between the nucleus and cytoplasm. Nature 355, 730-732. Ohno, M., Kataoka, N., & Shimura, Y. (1990). A nuclear cap binding protein from HeLa cells. Nucleic Acids Res. 18,6989-6995.

1 58

ELISA IZAURRALDE, IAIN W. MATTAJ, and DAVID S. GOLDFARB

Okuno, Y., Imamoto, H., & Yoneda, Y. (1993). 70 kD heat-shock cognate protein colocalizes with karyophilic proteins into the nucleus during their transport in vitro. Exp. Cell Res. 206, 134-142. Reddy, R. & Busch, H. (1988). Small nuclear RNAs: RNA sequences, structure and modifications. In: Structure and Function of Major and Minor Small Nuclear Ribonucleoprotein Particles (Birnstiel, M. L., ed.), pp. 1-37. Springer-Verlag, New York. Rhoads, R.E. (1988). Cap recognition and the entry of mRNA into the protein synthesis initiation cycle. Trends Biochem. Sci. 13,52-56. Richardson, W.D., Mills, A.D., DDworth, S.M., Laskey, R.A., & Dingwall, C. (1988). Nuclear protein migration involves two steps: rapid binding at the nuclear envelope, followed by slower translocation through nuclear pores. Cell 52, 655-664. Robbins, J., Dilworth, S.M., Laskey, R.A., & Dingwall, C. (1991). Two interdependent basic domains in nucleoplasmin nuclear targeting sequence: identification of a class of bipartite nuclear targeting sequence. Cell 64, 615-623. Roberts, B.L., Richardson, W.D., & Smith, J.A. (1987). The effect of protein context on nuclear location signal function. Cell 50,465-475. Rozen, F. & Sonenberg, N. (1987). Identification of nuclear cap specific proteins in HeLa cells. Nucleic Acids Res. 15,6489-6500. Shatkin, A.J. (1985). mRNA cap binding proteins: essential factors for initiation of translation. Cell 40, 223-224. Shi, Y & Thomas, J.O. (1992). The transport of proteins into the nucleus requires the 70-kilodalton heat shock protein or its cytosolic cognate. Mol. Cell Biol. 12, 2186-2192. Silver, R, Sadler, I., & Osborne, M.A. (1989). Yeast proteins that recognize nuclear localization sequences. J. Cell Biol. 109, 983-989. Singh, R. & Reddy, R. (1989). y-Monomethyl phosphate: a cap structure in spliceosomal U6 small nuclear RNA. Proc. Natl. Acad. Sci. USA 86, 8280-8283. Sonenberg, N. (1981). ATP/Mg"*"*"-dependent cross-linking of cap binding proteins to the 5' end of eukaryotic mRNA. Nucleic Acids Res. 9, 1643-1656. Sonenberg, N. (1988). Cap-binding proteins of eukaryotic messenger RNA: functions in initiation and control of translation. Prog. Nucleic Acids Res. Mol. Biol. 35, 174-207. Starr, CM. & Hanover, J.A. (1992). Structure and function of nuclear pore glycoproteins. In: Nuclear Trafficking (Feldherr, CM., ed.), pp. 175-201. Academic Press, New York. Steitz, J.A., Black, D.L., Gerke, V., Parker, K.A., Kramer, A., Frendewey, D., & Keller, W. (1988). Functions of the abundant U snRNPs. In: Structure and Function of Major and Minor Small Nuclear Ribonucleoprotein Particles (Birnstiel, M. L., ed.), pp. 115-154. Springer-Verlag, New York. Steme-Marr, R., Blevitt, J.M., & Gerace, L. (1992). 0-linked glycoproteins of the nuclear pore complex interact with a cytosolic factor required for nuclear protein import. J. Cell Biol. 116,271-280. Sumpter, V, Kahrs, A., Fischer, U., Komstadt, U., & Luhrmann, R. (1992). In vitro reconstitution of Ul and U2 snRNPs from isolated proteins and snRNA. Mol. Biol. Rep. 17, 1-6.

Nuclear Transport of U snRNPs

159

Terns, M.P., Lund, E., & Dahlberg, J.E. (1992). 3'-end-dependent formation of U6 small nuclear ribonucleoprotein particles in Xenopus oocyte nuclei. Mol. Cell. Biol. 12, 3032-3040. Vankan, P., McGuigan, C , & Mattaj, I.W. (1990). Roles of U4 and U6 snRNAs in the assembly of splicing complexes. EMBO J. 9, 3397-3404. van Zee, K., Dickmanns, A., Fischer, U., Ltihrmann, R., & Fanning, E. (1993). A cytoplasmically anchored nuclear protein interferes specifically with the import of nuclear proteins but not Ul snRNA. J. Cell Biol. 121, 229-240. Wersig, C. & Bindereif, A. (1992). Reconstitution of functional mammalian U4 small nuclear ribonucleoprotein: Sm protein binding is not essential for splicing in vitro. Mol. Cell. Biol. 12, 1460-1468. Wersig, C , Guddat, U., Pieler, T., & Bindereif, A. (1992). Assembly and nuclear transport of the U4 and U4/U6 snRNPs. Exp. Cell Res. 199, 373-377. White, E.M., AUis, CD., Golfarb, D.S., Srivastva, A., Weir, J.W., & Gorovsky, M.A. (1989). Nuclear specific and temporally restricted localization of proteins in Tetrahymena macronuclei and micronuclei. J. Cell Biol. 109, 1983-1992. Wilken, N., Kossner, U., Senecal, J.-L., Scheer, U., & Dabauvalle, M.-C. (1993). Nupl 80, a novel nuclear pore complex protein localizing to the cytoplasmic ring and associated fibrils. J. Cell Biol. 123,1345-1354. Will, C.L., Behrens, S.-E., & Ltihrmann, R. (1993). Protein composition of mammalian spliceosomal snRNPs. Mol. Biol. Rep. 18,121-126. Wolin, S.L. & Walter, P (1991). Small ribonucleoproteins. Curr. Opin. Struct. Biol. 1, 251-257. Yamasaki, L. & Lanford, R.E. (1992). Nuclear localization signal receptors. Trends Cell Biol. 2,123-127. Yang, H., Moss, M.L., Lund, E., & Dahlberg, J.E. (1992). Nuclear processing of the 3'-terminal nucleotides of pre-Ul RNAin Xenopus laevis oocytes. MoL Cell. Biol. 12, 1553-1560. Zasloff, M. (1983). tRNA transport from the nucleus in aeukaryotic cell: carrier-mediated translocation process. Proc. Nad. Acad. Sci. USA 80,6436-6440. Zeller, R., Nyffenegger, T, & De Robertis, E.M. (1983). Nucleocytoplasmic distribution of snRNPs and stockpiled snRNA-binding proteins during oogenesis and early development in Xenopus laevis. Cell 32,425-434. Zieve, G.W. & Sauterer, R.A. (1990). Cell biology of the snRNP particles. CRC Crit Rev. Biochem. Mol. Biol. 25, 1-26. Zieve, G.W, Sauterer, R.A., & Feeney, R.J. (1988). Newly synthesized small nuclear RNAs appear transiently in the cytoplasm. J. Mol. Biol. 199, 259-267.

This Page Intentionally Left Blank

PHOSPHORYLATION-MEDIATED REGULATION OF SIGNAL-DEPENDENT NUCLEAR PROTEIN TRANSPORT: THE ''CcN MOTIF^^

David A. Jans

I. Introduction II. Nuclear-Cytoplasmic Processes in the Eukaryotic Cell A. The Nuclear Pore Complex as a Molecular Sieve B. Nuclear Localization Signals III. Methodological Considerations in Analyzing Nuclear Transport rV. Regulated (Conditional) Nuclear Entry V. SV40 Large Tumor Antigen and the CcN Motif VI. Mechanisms of Regulation of Signal-Dependent Nuclear Protein Transport A. Cytoplasmic Retention Factors

Membrane Protein IVansport Volume 2, pages 161-199 Copyright © 1995 by JAI Press Inc. All rights of reproduction in any form reserved. ISBN: 1-55938-983-4

161

162 163 163 165 168 172 172 180 180

162

DAVID A. JANS

B. NLS Masking C. The CcN Motif: Positive and Negative Regulation by Dual Phosphorylation VII. Variants of the CcN Motif: Other Kinases Able to Regulate NLS-Dependent Nuclear Protein Import A. Signal Transduction-Responsive NLSs B. Cell-Cycle-Dependent NLSs VIII. Conclusion and Future Prospects Acknowledgments References

183 186 187 187 189 190 191 192

I. INTRODUCTION Since the sequence responsible for subcellular localization of the SV40 large tumor antigen (T-ag) was defined, research in the field of nuclear protein transport has largely revolved around the idea of a nuclear localization signal (NLS) being exclusively responsible for nuclear localization. NLSs have been defined for a multitude of nuclear proteins, and specific NLS-binding proteins or receptors have been identified. It has become increasingly clear, however, that NLS function can be regulated, with phosphorylation as the main mechanism controlling NLS-dependent nuclear localization of a number of proteins, including transcription factors (TFs) such as N F - K B , c-rel, v-jun, the a-interferonregulated factor ISGF-3, the yeast TF SWI5, and other nuclear proteins. Recent analysis indicates that even the archetypal NLS-containing T-ag is subject to regulated (conditional) nuclear localization. The regulatory module for T-ag nuclear transport (the "CcN motif) has been identified, in which the NLS ("N") determines ultimate subcellular destination, a casein kinase II site ("C") 15 amino acids N-terminal to the NLS determines the rate of nuclear import, and a cdk (cyclin-dependent kinase) site ("c") immediately adjacent to the NLS regulates the absolute amount of T-ag maximally accumulated in the nucleus. Since C and c sites can be identified in the vicinity of the NLS of a variety of other physiologically important proteins, it is likely that the CcN motif represents a conserved structural domain conferring conditional NLS-dependent transport on many TFs. Variants of the CcN motif include "SRNs" (signal transduction-responsive NLSs) and "CDNs" (cell-cycle-dependent NLSs), where kinases other than those acting at the CcN motif specifically regulate nuclear import of particular proteins through phosphorylation at site(s) close to their respective NLSs. Cytoplasmic reten-

Phosphorylation-Mediated Nuclear Transport

163

tion factors, intra- and intermolecular NLS masking, and direct NLS masking by phosphorylation are some of the better understood mechanisms by which phosphorylation specifically regulates nuclear transport. The regulation of nuclear transport through such mechanisms appears to be common to eukaryotic cells from yeast and plants to higher mammals. II. NUCLEAR-CYTOPLASMIC PROCESSES IN THE EUKARYOTIC CELL A. The Nuclear Pore Complex as a Molecular Sieve The fact that eukaryotic cells possess a nucleus has important implications for cellular function, since the existence of a nuclear compartment means that the genetic information, the DNA, is separated from the site of protein synthesis, the cytoplasm, by a double membrane structure, the nuclear envelope (Dingwall and Laskey, 1992). Gene transcription and translation accordingly take place in different subcellular compartments, meaning that specific transport events between the two are necessary for a eukaryotic cell to be functional: mRNA must make its way from the nucleus to the cytoplasm in order to be translated into protein; and proteins that are needed in the nucleus (structural nuclear proteins such as histones and lamins, the enzymes required for DNA replication, gene transcription, activation of gene expression, etc.) have to be transported from their site of synthesis in the cytoplasm into the nucleus. All transport goes through the nuclear envelope-localized nuclear pore complexes, very large (about 10^ kDa) almost organelle-like structures composed of at least 30 distinct protein components. Between 100 and 5 x 10^ pore complexes are present per nucleus, depending on the metabolic and differentiation state of the cell, with more complexes present in active or differentiating cells. As its name suggests, the nuclear pore complex has a pore or molecular sieve function whereby molecules smaller than 40-45 kDa can diffuse freely between cytoplasm and nucleus (Miller et al., 1991; Silver, 1991). B. Nuclear Localization Signals Proteins larger than 45 kDa require a nuclear localization signal (NLS) in order to be targeted to the nucleus.^ Through mutation/deletion analysis and/or their abihty to target a normally cytoplasmically localized protein to the nucleus, NLSs have been defined as the sequences neces-

164

DAVID A. JANS

sary and sufficient for nuclear localization. They function through recognition/ligand-receptor-like interactions rather than through binding to chromatin or nuclear structures, etc. (see Miller et al., 1991; Silver, 1991). Proteins larger than 45 kDa microinjected into the nucleus remain nuclear whether they possess an NLS or not, demonstrating that NLSs do not function as nuclear retention signals (see Miller et al., 1991; Silver, 1991). The archetypal NLS is that of the SV40 large tumor antigen (T-ag) (Pro-Lys-Lys-Lys-Arg-Lys-Val^^^). It is both sufficient and necessary for nuclear localization of T-ag (Kalderon et al., 1984a,b; Lanford and Butel, 1984) and functional in targeting normally cytoplasmic carrier proteins to the nucleus whether present within the coding sequence of the carrier protein, or as a peptide covalently coupled to the carrier. The mutation of a single amino acid residue within this sequence (Lys^^^ to either Thr or Asn) abolishes its nuclear targeting function. The T-ag NLS, with its concentration of basic amino acids preceded by a P-tum, has served as the basis of the search via homology for NLSs in other nuclear proteins (Garcia-Bustos et al., 1991; Miller et al., 1991). Another class of NLS— the "bipartite NLS" with two clusters of basic amino acid residues separated by an intervening 10-12-amino acid spacer whose length appears to be critical—^has more recently been established as a variant of the T-ag NLS archetype (Dingwall and Laskey, 1991, 1992; Robbins et al., 1991); examples are those for the Saccharomyces cerevisiae transcription factor (TF) SWIS and the Xenopus laevis nuclear phosphoproteins nucleoplasmin and N1N2 (see Tables 1 and 2). A further class of NLS is based on the yeast MAT a2 NLS (Asn-Lys-Ile-Pro-Ile-LysAsp^ see Garcia-Bustos et al., 1991; Miller et al., 1991). NLSs have been hypothesized to be recognized by receptor/transporter proteins called NLS-binding proteins (NLSBPs), which may play a direct role in active nuclear transport. Although their precise specificity and subcellular location are still unresolved, NLSBPs have been identified by a number of research groups (Yoneda et al., 1988; Adam et al., 1989;Bendittetal., 1989; Li and Thomas, 1989; Meier and Blobel, 1990; Lee at al., 1991; Stochaj et al., 1991). Direct evidence has been provided that the molecular chaperonin and heat shock-related HSP70/HSC70 protein is both necessary for NLS-dependent nuclear transport in reconstituted systems and capable of specifically binding NLSs (Imamoto et al., 1992; Shi and Thomas, 1992).

Phosphorylation-Mediated Nuclear Transport

165

III. METHODOLOGICAL CONSIDERATIONS IN ANALYZING NUCLEAR TRANSPORT A drawback of most nuclear transport studies to date relating to the identification of NLSs is the nature of the experiments and techniques used, whereby eukaryotic cells are generally transfected with a proteinexpressing plasmid construct and then scored qualitatively for subcellular localization 24-^8 hours later using indirect immunofluorescence in fixed cells. Apart from difficulties in interpreting the results through the fact that fixation procedures may artifactually affect subcellular localization, such an approach provides minimal quantitative information with respect to how much protein is in the nucleus. It also provides no information with respect to the rate of nuclear import, since analysis is performed on cells at steady state, and the subcellular localization is clearly a function not only of intrinsic nuclear transport properties, but also of the transfection efficiency and the level of expression of the transfected DNA. In addition, the high levels of protein expression in the cells analyzed are often not physiological, especially in the case of TFs (e.g., N F - K B ) , which are normally present in very low amounts. Some of the above drawbacks can be avoided by examining the transport of carrier proteins such as bovine serum albumin carrying covalently attached NLS peptide microinjected into living cells, where information may be obtained on the initial rate of nuclear transport. However, the carrier protein constructs used to date have generally involved multiple NLS peptides, thus being far from physiologically relevant in terms of the ratio of NLS to molecular weight/size. This has been shown to be of significance in several studies (Nelson and Silver, 1989; Yoneda et al., 1992). Alternative methods such as subcellular fractionation can provide quantitative data if performed in the correct manner, but suffer from the caveat that redistribution of proteins may occur during the fractionation process. Technical advances in the application of confocal laser scanning microscopy (CLSM) have made the quantitative estimation of fluorescence possible, meaning that relative protein concentrations in different subcellular compartments can be measured in vivo (Peters, 1986; Jans, 1992). The use of microinjection of fluorescently labeled protein in conjunction with CLSM enables the quantitative analysis of nuclear import in single living cells (see Figure 1 A), and it has recently proved possible to resolve temporally the nuclear transport process and deter-

DAVID A. JANS

166

NUCLEAR TRANSPORT IN VIVO IN MICROINJECTED HTC CELLS I Cells fused with PEG i to yield polykaryons

mm Cells grown in standard Wm medium/petri dish without antibiotics

&^ii»i&»@

CKII-SITE-DEPENDENT £ik2-K-INHIBITABLE NLS-DEPENDENT

CLSM

NUCLEAR T R A N S P O R T IN V I T R O IN M E C H A N I C A L L Y P E R F O R A T E D H T C CELLS Cells grown on washed Dincoverslips intracellular buffer

CKII-SITE-DEPENDENT £d£2-K-INHIBITABLE NLS-DEPENDENT

WMMMm

1m .4V. ^'^^ ~ Cytosol lAF-labelled Protein

^ Coverslip removed to leave behind apical cell membrane

Figure 1. Systems to measure the kinetics of signal-dependent nuclear protein transport in vivo (A) and in vitro (B), using quantitative confocal laser scanning microscopy.

mine transport kinetics (Rihs and Peters, 1989; Rihs et al., 1991; Jans et al, 1991; Jans and Jans, 1994; see Section V below). The application of in vitro or reconstituted nuclear import assay systems has enabled an initial assault on understanding the steps of

Phosphorylation-Mediated Nuclear Transport

167

Figure 1. (continued) (C) Signal dependence of nuclear transport in the respective systems for microinjected HTC cell polykaryons (37 °C, 15 and 40 minutes, respectively, for a T-ag(amino acids 111-135)-P-galactosidase(amino acids 9-1028) fusion protein (left) and a fluorescein-labeled 70-kDa dextran (right), respectively) (/), and for mechanically perforated HTC cells (23°C, 75 minutes for T-ag (amino acids 111-135)-|i-galactosidase (amino acids 9-1028) fusion proteins without (left) and with (right) a Thr^^^ mutation in the NLS rendering the NLS inactive) (2). Proteins were labeled with 5-iodo-acetamido-fluorescein (lAF), and HTC cells were fused with polyethylene glycol 1 hour before microinjection as described previously (Rihs and Peters, 1989). Microinjection was performed with a Narshige IM200 microinjectorand Zeiss micromanipulator, and images in the case of both assay systems were obtained with an MCR 600 Bio-Rad CLSM.

nuclear protein import, and defining them mechanistically with respect to NLS-binding proteins, including HSP70/HSC70 (Adam et al., 1989; Stochaj et al., 1991; Shi and Thomas, 1992).^ In conjunction with CLSM, in vitro nuclear transport assays can be applied in a quantitative approach to define the components and kinetics of the steps involved in regulating NLS-dependent nuclear transport. We have established an in vitro system based on cells of the HTC rat hepatoma cell line that have been mechanically perforated (see Figure IB) (Jans et al., 1991), using a modification of a procedure first used to investigate intracellular vesicle trafficking

168

DAVID A. JANS

(Simons and Virta, 1987). Consistent with the findings from other mv/rro assay systems based on differential permeabilization of the plasma membrane and the nuclear envelope by the detergent digitonin (Adam et al., 1989; Stochaj et al., 1991; Shi and Thomas, 1992; Moore and Blobel, 1993; Adam and Adam, 1994), signal-dependent nuclear transport in mechanically perforated HTC cells can be shown to involve at least two stages (Jans et al., 1991), as first reported by Newmeyer and Forbes (1988). Both stages are dependent on cytosolic factors (unpublished results; Jans et al, 1991; see Figure IB): an ATP-independent docking or binding stage at the nuclear envelope/nuclear pore complex (see Figure 4), and an ATP-dependent active transport step. Signal (NLS) dependence of nuclear transport in vivo and in vitro is illustrated in Figure IC. Analysis of the rates and maximal extent of nuclear transport (see below) indicate quite clearly that nuclear transport is not an all-or-none phenomenon exclusively determined by the possession or absence of an NLS—^rather, as soon as kinetic measurements are quantitatively performed, it becomes clear that other sequences function in a regulatory fashion to modulate NLS-dependent nuclear transport in a very precise way.

IV. REGULATED (CONDITIONAL) NUCLEAR ENTRY Whereas some proteins such as histones and nucleolin appear to be constitutively targeted to the nucleus, others only seem to be targeted to the nucleus under specific conditions, often being mostly in the cytoplasm (see Miller et al., 1991; Schmitz et al., 1991). TFs regulating nuclear gene expression are not different from other proteins in terms of their being synthesized in the cytoplasm and therefore being subject to specific mechanisms regulating nuclear protein import. The advantages of a conditionally cytoplasmic location for a TF include the potential to control its nuclear activity through the regulation of its nuclear uptake, and its direct accessibility to cytoplasmic signal-transducing systems. The nuclear translocation of various TFs (Lenardo and Baltimore, 1989; Shirakawa and Mizel, 1989; Govind and Steward, 1991; Moll et al., 1991; Schmitz et al., 1991) and oncogene products (Gusse et al., 1989; Van Etten et al., 1989; Capobianco et al, 1990; Roux et al., 1990) has been shown to accompany changes in the differentiation or metabolic state of eukaryotic cells precisely, indicating that nuclear protein import can be a key control point in regulating gene expression and signal transduction.

Table I . Examples of Reaulated Nuclear Protein Transport Protein

Kinase/stimulus

Effect on Nuclear Transport (Sequences ~nvolved)~

A. Signal transduction-related T-ag CKll c-re1 PK-A NF-KB

Lamin B2 GR

c-jos Cofilin

rNFIL-6 ISGF-3

NF-AT

CKll site increases the rate of nuclear transport by -40-fold (a) (s"'s"~DDE-10 a.a. spacer-PKKKRKV). PK-A site enhances nuclear localization (b), whereby mutation of SZMabolishes nuclear localization ( R R P S ' ~ ~a.a. -~~ spacer-KAKROR). Phosphorylation (of 1KB or NF-KB subunits) results in unmasking of ~ 5 0 1 ~ NLSs 6 5 and NF-KB nuclear localization (c) IL-1cONFa E - ~spacer-ORKROK ~ human p65 R R P S ~ ~ ~ D R E L S a.a. E - I ~spacer(human p50/p110 R R K S ~ ~ ~ - D L E T Sa.a. PK-C (PK-A/elF2K) EEKRKR). PK-C-mediated phosphorylation inhibits nuclear transport (d) ( R S ~ ~ ~ S ~ " R G K R R R I E ) . PK-C Hormone binding by GR releases it from a cytoplasmic complex with HSP90 and results in its nuclear localization (e). GH Nuclear retention of GR is impaired during the G2 stage of the cell cycle, concomitant with altered phosphorylation PP-2A? of GR. The inhibition of phosphatase PP-2A by okadaic acid similarly leads to inefficient nuclear retention of GH-occupied GR, as does transformation by the v-mos oncogenic kinase, implying that site-specific (cell-cycledependent) dephosphorylationof GR is involved in hormonedependent nuclear translocation (f). (AconstitutiveNLS RKCLQAGMNLEARKTKK (amino acids 479-495 of the rat sequence) is repressed by a second ligand bindingdependent NLS function, which includes amino acids 600-626 and 696-777.) (g) Phosphorylation of c-jos reverses its binding to a putative inhibitor protein and retention in the cytoplasm (h) PK-Nsemm (KRRIRRIRNKMAAAKCRNRRRL-200 a.a. S ~ ~ C ~ ~ - R K G S ~ ~ ~ S S " ) . Heat shock induces unmasking of the cofilin NLS via dephosphorylation at a consensus multifunctional calmodulindeMultifunctional pendent protein kinase site, resulting in its nuclear localization (i) (RKSS~~TPEEKKRKKA). calmodulin dept. protein kinasemeat shock The adenylate c clase activator forskolin stimulatesphosphorylation of rNF1L-6 and its translocation to the nucleus cAMPIPK-A? Z CKPSKKPS' ??I. Interferon-induced tyrosine phosphorylation of the three ISGF3a subunits is required to effect their association and Interferon a/$ translocation to the nucleus (k).' tyrosine kinase ca2+-dependentactivation of calcineurin results in dephosphorylation of cytoplasmicNF-AT, and its nuclear localization ca2+ (1). Through different mechanisms, the immunosuppressants FKS06lcyclosporin inhibit the phosphatase and thereby FK506/cyclosporin induce nuclear translocation of NF-AT. (continued)

Table 1. Continued Prorein

d

U 0

Kinase/stimulus

Effect on Nuclear Transport (sequences ~nvolved)~

B. Cell cycledependent/developmental T-ag cdWcdc2-K cdc2-K-mediated phosphorylation reduces the maximal level of nuclear accumulation, probably through increasing the affinity of binding to a cytoplasmic retention factor (m) ( S T ~ ~ P K K K R K V ) . SWIS cdWCDC28 Phosphorylation-mediatednuclear exclusion through NLS masking--mutation of the CDC28-serines to alanine results 109 - a.a. s p a c e r - ~ ~ ~ ~ in corlstitutive nuclear localization (n) ( s ~ ~ ~ P s K RKDGTSSVSSS~PIK). lodestar CKII? Nuclear exclusion until prophase (DESS~%DS~~~DS%DKNKKRK) (0) v-jun cdk? Cell-cycledependent determination of the rate of nuclear localization conferred by Sa8 (faster during G2); c-jun possesses cZ4' and is constitutively nuclear ( A S ~ ~ ~ R K R(p). KL) dorsal toll Relocalization from the cytoplasm to the nucleus during developmentin a ventral-todorsal gradient, involving multiple PK-A/c~~+? components of the roll si al transduction pathway, including the cytoplasmic retention factor cacrus, and "releasing factors" pelle and rube.8" Phosphorylation of dorsal leads to release from cactus and nuclear localization, whereby activated roll receptor and increased cytoplasmicca2+ appear to be involved (q) ( R R P S ~ ' * a.a. - ~ ~spacer -RRKROR). AdenovirusS ?? Nuclear targeting up to the early neumla stage of Xenopus embryonic development through a "developmentally Ela protein regulated NLS (drNLS) (r) (~~NLs':amino acids 142-182:m- 20 a.a. spacer - MCSLCYMRTCGM!!, distinct from the constitutive C-terminal NLS: a.a. 285-289: KRPRP). RB-1 cdWcdc2-K Cellcyclede ndent hype~hospho~lation reduces nuclear association ("nuclear tethering") (s) (Putative CcN motif: human ADMYLS %v&' where P1 is a putative CKll site, and serines 608 and 612 cdk sites).'

STIR.

Notes:

a.a.. amino acids; U - 1 4 interleukin l a ; TNFa, tumor necrosis factor a ; eIF2K. heme regulated initiation factor 2 kinase; GR, glucocorticoid receptor; GH. glucocorticoid hormone; PP-2A, protein phosphatase type 2A; ISGF-3, interferon-stimulated gene factor; rNFlL-6, rat nuclear factor induced by IL-6;NF-AT,nuclear factor of activated T cells; RB-I. pl loRb,the product of the "retinoblastoma-susceptibility factor'' tumor suppressor gene. single letter amino acid code is used. NLSs are underlined, and phosphorylatedresidues are numbered accordingto their residue number in the respective protein. References are: (a) Ribs et al. (19911; Jans et al. (1991); Jans and Jans (1994); (b)Mosialos et al. (1991); (c) Shirakawa and Mizel(1989); Lenardo and Baltimore (1989); Ghosh and Baltimore (1990); Henkel et al. (1992); Zabel et al. (1993); Beg et al. (1992); (d) Hennekes et al. (1993); (e) Picard et al. (1988); (f) De Franco et al. (1991); Hsu et al. (1992); Qi et al. (1989); (g) Cadepond et al. (1992); (h) Roux et a]. (1990); Auwerx et al. (1990); Tratner et al. (1992); (i) Nishida et al. (1987); Ohta et al. (1989); (i)Metz and Ziff (1991); (k) Pellegrini and Schindler (1993); Schindleret al. (1992); (1) Liu (1993); (m)Jans et al. (1991); (n) Moll et al. (1991); (0) Girdham and Glover (1991); (p) Chida and Vogt (1992); (q) Govind and Steward (1991); Whalen and Steward (1993); Kubota et al. (1993); (r) Standiford and Richter (1992); (s) Templeton et al. (1991); Templeton (1992).

~

~

1

~

b~egulationof nuclear translocation of rNFIL-6 may be cell type specific, since CAMP-mediatedstimulation is required for its nuclear localization in pheochromocytomacells, whereas rNFIL-6 is constitutively nuclear in HeLa cells (Metz and Ziff, 1991). 'The 91-kDa ISGF-3a subunit is also a componentof the interferon y-induced GAF (gamma-activated factor) TF (Shuai eta]., 1992; see Pellegrini and Schindler, 1993).Tyrosine phosphorylation induces nuclear localization of the 91-kDa ISGF-3a subunit in this signaling pathway as well. %he C-terminus (LQISNISIST~'~) of dorsal is implicated in cytoplasmic retention, since its deletion results in nuclear localization of the dorsal gene product. Proteolysis of dorsal or cactus may also be involved. e~omologoussequences have been identified in the rat (amino acids 594-6231. mouse, and human GH binding domains of GR (see Standiford and Richter. 1992). which are known to include a part of the ligand binding-dependent NLS (Picard et a]., 1988; Cadepond et al., 1992). A ' 'bipartite" NLS purported to participate in nuclear localization of mouse RB has been identified: KRSAEFFNPPKPL='~~-~~ ,.a. S ~ ~ C ~ ~ - Swhere ~ ~ ~s~~~ I GisEa , putative CKII site) (Zacksenhaus et al., 1993).

1 72

DAVID A. JANS

TFs able to undergo inducible nuclear import include the glucocorticoid receptor (GR) (Picard et al., 1988), the a-interferon-regulated factor ISGF-3 (interferon-stimulated gene factor 3) (Schindler et al., 1992), the nuclear w-jun oncogenic counterpart of the AP-1 transcription complex member c-jun (Chida and Vogt, 1992), the yeast mating switch/HO endonuclease promoter regulator SWI5 (Moll et al., 1991), the Drosophila melanogaster morphogQu dorsal (Govind and Steward, 1991), and the nuclear factor N F - K B (Lenardo and Baltimore, 1989; Shirakawa and Mizel, 1989; Schmitzetal., 1991). A number of physiological examples of regulated NLS-dependent nuclear protein localization, where the signal and regulatory sequences responsible have been identified to some extent, are listed in Table 1. Examples of signal transduction pathways from extracellular hormonal signal to the nucleus ultimately leading to changes in the regulation of gene expression include the GR, where, upon glucocorticoid hormone (GH) binding, cytoplasmic receptor translocates to the nucleus to modulate gene transcription by direct binding to specific DNA sequences called GH response elements (GREs) (Picard et al., 1988); another example is the hormone-stimulated response to elevated cAMP levels. The latter results in translocation of the cAMP-dependent protein kinase (PK-A) catalytic (C-) subunit from the cytoplasm to the nucleus (Nigg et al., 1985; Meinkoth et al., 1990; Pearson et al., 1991), where it phosphorylates and thereby modulates the activities of nuclear TFs such as the cAMP response element (CRE) binding protein (CREB). In a fashion similar to that of PK-A, the MAP- (ERK) and RSK-kinases also enter the nucleus upon mitogenic stimulation and activation (Chen at al., 1992), whereas the 82-kDa a isoform of the Ca^Vphospholipid-dependent kinase (PK-C) relocalizes to the nuclear envelope upon activation (Leach et al., 1989).^ Hormonally induced nuclear transport of activated kinases is a commonly exploited means of communicating signals to the nucleus in many signal transduction and developmental pathways.

V. SV40 LARGE TUMOR ANTIGEN AND THE CcN MOTIF As stated in Section II.B, nuclear localization of T-ag is dependent on amino acids 126-132. Mutations within the T-ag NLS abolish nuclear targeting (Kalderon et al., 1984a,b; Lanford and Butel, 1984; see Figure IC). Using quantitative CLSM and the systems shown in Figure 1, A and B, to measure nuclear transport kinetics at the single cell level both in vivo and in vitro, we have demonstrated that NLS-dependent nuclear

Phosphorylation-Mediated Nuclear Transport

173

transport of T-ag is regulated by phosphorylation. Sites for the physiologically important casein kinase II (CKII) and the cdk (cyclin-dependent kinase) cdc2-kinase (crfc2-K), which are close to the T-ag NLS, regulate the kinetics of nuclear transport of T-ag, probably through the modulation of specific protein-protein binding affinities and recognition events (Rihs and Peters, 1989; Jans et al., 1991; Rihs et al., 1991; Jans and Jans, 1994). The CKII site (the serine at position 112) increases the rate of NLS-dependent nuclear import, so that maximal accumulation within the nucleus occurs within 20 minutes, contrasting with the 10 hours taken when CKII phosphorylation is prevented through deletion or mutation of the CKII site (Rihs et al., 1991; Jans and Jans, 1994) (see Figure 2A). In contrast to the enhancing effect of the CKII site, phosphorylation by cdc2'K (at threonine 124) inhibits nuclear transport, drastically reducing the level of maximal nuclear accumulation (Jans et al., 1991) (see Figure 2B). We have named this regulatory module for T-ag nuclear transport the "CcN motif," comprising CKII ("C") and cdk/cdc2K ("c") sites, and the NLS ("N") (Jans et al., 1991) (see Figure 2). Aspartic acid (see Figure 3) can substitute functionally for phospho-serine/threonine in the case of both phosphorylation sites of the T-ag CcN motif (Jans etal, 1991; Jans and Jans, 1994), implying that the phosphorylated sites probably represent recognition signals in themselves that are recognized by components of the cellular nuclear transport machinery. The existence of a complex regulatory system for S V40 T-ag nuclear localization, namely the CcN motif involving two different phosphorylation sites, demonstrates the existence of specific mechanisms regulating nuclear entry. Clearly, the NLS is not the sole determinant conferring nuclear localization in an all-or-none phenomenon; rather, the kinetics of NLS-dependent nuclear localization are quantitatively regulated by phosphorylation sites in the vicinity of the NLS. Through phosphorylation at the signal transduction-responsive CKE"^ site positively regulating the rate of nuclear transport, and phosphorylation at the cell cycle-dependent cdk/cc/c2-K site negatively regulating the maximal extent of accumulation, the level of T-ag present in the nucleus can be precisely regulated, presumably exactly as required with respect to the eukaryotic cell cycle and stages of the viral lytic cycle. Significantly, a variety of nuclear proteins in addition to T-ag possess putative CcN motifs (Jans et al., 1991), implying a general role for the CcN motif in regulating nuclear protein transport. Some of these are shown in Table 2; others include the Drosophila "Notch" group of genes.

The CKII Site Enhances the Rate of Nuclear Transport

Fn/c

CKII site intact

C cdc2-K-phosphorylation Inhibits Nuclear Transport

Unphovphorylated

cdc2-pho«phorylated

i^ -J

0

5

I

I

I

i

I

I

10

15

20

25

30

35

40

Time (min)

Nuclear Transport is completely dependent on the NLS

N 174

Phosphorylation-Mediated Nuclear Transport

175

including Notch itself and Enhancer of Split and human homologs (EUisen et al., 1991; Stifani et al., 1992; Fortini et al., 1993); the yeast TF SNF2 (Laurent et al, 1993) and human and Drosophila homologs (Chiba et al., 1994); the family of interferon a- or y-induced TFs, including IFi204 and IFil6 (Choubey and Lengyel, 1992; Trapani et al, 1992); the Swi4 family of mismatch repair enzymes (Fleck et al., 1992), bovine poly-A polymerase (Wahle et al., 1991), yeast DNA topoisomerase II (Shiozaki and Yanagida, 1992), the DNA repair helicase ERCC6 (Troelstra et al., 1992), and the protein tyrosine phosphatase PEP (Matthews et al., 1992). Putative CcN motifs can also be found in the sequences of proteins localizing in the plant cell nucleus (Klimczak et al., 1992; Raikhel, 1992; Tinland et al., 1992; Hicks and Raikhel, 1993) (see Table 2), underlining the universal nature of NLSs and nuclear protein transport across the animal/plant kingdom, as well as of the latter's regulation by phosphorylation. This universality is further supported by the functionality of the T-ag NLS in Drosophila embryos (Stein and Jans, unpublished observations, yeast, and plant cells, etc. (see Raikhel, 1992; Shiozaki and Yanagida, 1992); of the plant cell NLS of Agrobacterium tumefaciens VirD2 protein in yeast cells (Tinland et al., 1992); and of the cdk/CDC28kinase-regulated yeast NLS of SWI5 in rat and monkey cells (Jans et al., 1994). In the latter case, we have been able to show not only that the SWI5-NLS is functional in nuclear targeting, but also that the cdk

Figure 2. Regulation of nuclear protein transport by the CcN motif. Nuclear transport kinetics of fluorescently labeled T-ag (amino acids 111-135)-[J-galactosidase (amino acids 9-1028) fusion protein derivatives in microinjected HTC cells (see Figure 1 A) were measured by quantitative CLSM. Fn/c represents fluorescence quantitated in the nucleus relative to that in the cytoplasm (i.e., fold accumulation in the nucleus). Regulation/dependence of nuclear protein transport by/on the CcN motif component CKII site (A), cdcl-Y. site (B), and NLS (C) is shown. The CKII site mutant (substitution of Ser^^^^^^^ by Ala, and replacement of the CKII recognition site Asp-Asp-Glu by Asn-Asn-GIn) and deletion (of amino acids 111-119, including the CKII site) derivatives shown have been described previously (Jans and Jans, 1994; Rihsetal., 1991). /n wfro phosphorylation with purified HeLa cdc2-kinase(stoichiometry of phosphorylation atThr^^"* of 1.4 mol P/mol tetramer) was performed as described in Jans et al. (1991). Results identical to those shown are obtained in mechanically perforated HTC cells (not shown, Jans et al., 1991).

Negative charge at the CKII site increases the nuclear transport rate Fn/c

Negative charge at the cdc2-site reduces maximal nuclear accumulation 20 r

Fn/c WT {Thrl24)

12 h

10

15

20

25

40

Time (min)

Figure 3. Negative charge at the CcN motif phosphorylation sites is mechanistically important for the regulation of nuclear transport. Nuclear transport kinetics of fluorescently labeled T-ag (amino acids 111-135)-^galactosidase(amino acids 9-1028) fusion protein derivatives in microinjected HTC cells were measured as in Figure 2. Asp at position 112 (Gly at position 111) (above) increases the rate of nuclear transport (compare to transport in the absence of a CKII site, with Gly at position 112 (and Ala at position 111), similar to CKII phosphorylation at the site (Ser^^^) (Jans and Jans, 1994). Asp at position 124 (below) reduces the extent of maximal accumulation, thus simulating cdc2-k phosphorylation of Thr^^^ (Jans et al., 1991); compare to Figure 2B. Comparable results to those shown are obtained in mechanically perforated HTC cells (not shown; Jans et al., 1991). 176

Table 2. Potential and Confirmed CKll and cdk-Kinase Phosphorylation Sites in the Vicinity of the NLS of Various Nuclear Proteins proteina SV40 T-ag Polyoma T-ag

CKII Site($) (Signal Transduction)

cdk Sire($) (Cell Cycle)

S'"S"~DDE (a) SStS6l"TD ES~~'ENE

PPK (b) P R T ' PVS ~~ ATz7' PPK SS~"PQP (e) PST "'RKP TS~PRS SPS'~~PTS(i) APDT'~~PEL (k) P A V S ~ ~ ~ P(k) LL T4 7 4 ~ ~ ~ T4 7 9 ~ ~ ~ T4 4 0 ~ ~ ~ TS~~PVR

Human p53 Human c-myc Lamin N C Mouse c-abl N

D G S ~ ~ ~ L(m) ND Human B-myb Human c-myb

ax^ poly A polymerase (bovine) ERCC6 (human)

S4" FLD s5'7m

S~DNDDIEVES~~DEE (n) DS I4OSS1 4 2 ~ ~ 1 (n) 4 4 DS~~'SLD DDS~~~EESD

~ ~ ~ ~

T'~~s'~%EE S~~EDELEE~ S ' 7 ' ~ ~ ~ ~ S~~TKE E T ~ ~ K S E E T ' ~ ~(0) LDE S~~~K (0)G E

~

NU PKKKRKV'" (c) VSRKRPRP'" P P K K A R E D ~(d) ~~ PQPKKKP 319 (f) PAAKRVKLD"~(g) RQRRNELKRS'~~(~) S V T K K R K L E ~(j) '~ SALIKKKKKMAP~~' (1)

LKRQRKRR~'~ LKK~KQ~~~ KRAHHNALERKRR~~ Q S R K K L R M ' ~(n) ~ KRK-12 a.a. spacer-^^^^^^^^ RRWNKLRLQDKEKRLK~~' KRR- I I a.a. spacer-^^^^^'^^^^ RRRRRR~ KRKSMREEETGVKKSKAAK'94f KRKKSTKEKAGPKGSK'~~~ SKKKKNAK~~ (continued)

Table 2. (Continued) CKII Site(s) (Signal Tramduction)

cdk Site(s) (Cell Cycle)

S'~~PAD S ~ ~ ~ D Q D DSS~~~HYDS~"~DGD DVS~~~NED S~'~~PVD S'~~PAK

RTRS~~~PL

REKYRTR-I0 a.a. s p a c e r - ~ K R R KCsh~ ~ ~

PSS~~~PR

RGTDKRR-9 a.a.

proteina Drosophila gro TLE- I TAN-1' Plectrin (rat)

S'~~PTKK S'~~PPK S"~PLK

Yeast SW15 (S. cerevisiae)

s1362s 1

K R S ~ ~ ~ P(s) RK S S S ~ ~ P (s) IK 3

6

3

~

~

DPS 1 4 5 5 1458 ~ ~ ~ MDEPS'463~~~~

~

spacer-^^^^^^^'^'^

s ~ ~ ~ ~ P H KRAKDLKARRKK~'~~ S~'GSE SPAKKPKV~* (see p)

X. laevis Nucleoplasmin

DNA topoisomerase Il (S. pombe)

NLS

~

~

ST~~'~PK S'~"PIR

KRPAATKKAGQAKKKKL'~~ (q)

R K K R K - ~ ~ ~ . ~ . ~ ~ ~ C(r)~ ~ - K K S K Q E ~ ~ KKYENVVIKRSPRKRGRPRK~~~~ (s) KSRKR-12a.a. spacer-^^^^^^'^^^ RKRPTRR'~~'

Plant Opaque-2 (maize) G B F l (broccoli)

ES'~~NRE S ' ~ G N D A S ' ~ ~ H S(u) DE S'~~SDENDE

Agrobacterium rurnefaciens VirE2 protein VirD2 protein

S%TE DES~'QS'~DDD

(octopine)

EQDT"~~RDD

RAIKTKYGSDTEIKLKSK"~~ (v)

EYLSRKGKLEL~~ (w) S~PKR

K R P R D R H D G E L G G R K R A R ~ ~(x) ~~

Notes: a ~ h single e letter amino acid code is used. The consensus site for C m comprisesan acidic amino acid three residues C-terminal to the phosphorylatable SA', with an elevated number of acidic residues in the vicinity of the site increasing its affinity for CKlI (Krebs et al., 1988); cdks phosphorylate SA' residues N-terminal to proline residues, with a basic amino acid one or two residues C-terminal to the proline. References apply as indicated to confirmed NLS, and CKJI- and cdc2-K sites, with the phosphorylated/phosphorylatableSA' residues numbered: (a) Gfisser et al. (1988); (b)McVey at al. (1989); (c) Kalderon et at. (1984a,b);Lanford and Butel (1984); (d) Richardson et al. (1986); (e) Bischoff et al. (1990); (f) Chelsky et al. (1989); Addison et al. (1990); (g) Dang and Lee (1988); (h) Liischer et al. (1989); (i) Ward and Kirschner (1990); (i)Loewinger and McKeon (1988); (k) Kipreos and Wang (1990); (1) Van Etten et al. (1989); Sawyers et al. (1994); (m) Liischer et al. (1990); (n) F'rendergast et al. (1992); (0) Jin and Burakoff (1993); (p) Chelsky et al. (1989); Wiche et al. (1993); (q) Robbins et al. (1991); (r) Kleinschmidt and Seiter (1988); (s) Moll et al. (1991); (t) Varagona et al. (1992); (u) Klimczak et al. (1992); (v) Citovsky et al. (1992); (w) linland et al. (1992); (x) Howard et at. (1992).

b~denticalregions are found in the mouse Max correlate Myn (see F'rendergast et al., 1992). 'The NLSs of SWIS, nucleoplasmin, and the C-terminus of VirD2, as well as of the proteins above containinga "spacer" of 10-12 amino acids (a.a.), constitute '*bipartite-type"NLSs (see Section 1I.B) (Dingwall and Laskey, 1991, 1992; Robbins et at., 1991). 'Ihe alternatively spliced p70s6kvariant significantly lacks the putative N-terminal NLS and is not nuclear localized (Reinhard et at., 1994). %his putative CKJI site may not be necessary for nuclear targeting-see Reinhard et al. (1994). r~omologousregions are found in the related protein IFi202 (Choubey and Lengyel, 1992). g~omologousregions are found in the bovine FKBP25 (the FK506 binding protein) sequence. h~imilarsequences are found in TLE-2 and -3, the human homologues of the enhancer of split complex genes (Groucho) (Stifani et al., 1992; Ellisen et al., 1991). i Other members of the Notch family of proteins including hN also retain CcN motifs (Stifani et al., 1992; Fo~tiniet at., 1993).

180

DAVID A. JANS

site-mediated regulation of its function (see The relldorsal Family, below) operates in mammalian cells. This argues strongly for phosphorylation and the CcN motif and related sequences being universal in terms of regulating nuclear protein import. Notably in the case of the TF SWI5, we have been able to show that the cdk sites regulate the extent of maximal nuclear accumulation (Jans et al., 1995), exactly as crfc2-k phosphorylation does in the case of T-ag, although the mechanism of the phosphorylation-mediated effect appears to be different (compare SV40 Large T-Antigen and NLS Masking by Phosphorylation, below). Whereas Table 2 shows that an array of proteins other than T-ag possess CcN motifs, only a few have as yet been experimentally examined in terms of the regulation of their nuclear import kinetics. Several examples of cell cycle-dependent (cdk) phosphorylation inhibiting nuclear transport are given in Table 1, but the demonstration of CKII-mediated effects on the transport rate has thus far been restricted to T-ag. This is almost certainly due to the technically demanding nature of the quantitative analysis at the single, cell level required (see Section III). VI. MECHANISMS OF REGULATION OF SIGNAL-DEPENDENT NUCLEAR PROTEIN TRANSPORT A. Cytoplasmic Retention Factors The rel/dorsal Family One mechanism of regulating nuclear protein import is that of cytoplasmic retention, whereby a cytoplasmically localized "anchor" protein or retention factor specifically binds an NLS-containing protein and prevents it from migrating to the nucleus (Hunt, 1989) (see Figure 4). A cytoplasmic retention function has been described for the NF-KB-binding inhibitor protein IKB (also known as MAD-3), where phorbol ester (or other) treatment induces release of the NLS-carrying NF-KB p65 subunit from an inactive cytoplasmic complex with IKB in order to combine with the mature NF-KB p50 subunit and migrate to the nucleus (Lenardo and Baltimore, 1989; Shirakawa and Mizel, 1989; Schmitz et al., 1991). Phosphorylation of IKB by PK-C, PK-A, or the heme-regulated eIF2K (initiation factor 2 kinase) prevents its association with N F - K B in vitro (Ghosh and Balumore, 1990). Roles similar to that of IKB have been proposed for cactus (which shows sequence homology to the N F - K B inhibitor protein IKB) in negatively regulating nuclear localiza-

Reguiation of SV40 Large T-antigen Nuclear cdk (cdc2-K)

Cytoplasm

Accelerated Docking (?)

Nucleus

Figure 4. Regulation of SV40 T-ag nuclear transport by phosphorylation and the CcN nnotif. A simpi istic representation of some of the steps involved in the regulation of NLS-dependent nuclear transport of T-ag by cdk (cdc2-k) and CKII phosphorylation sites is shown. Phosphorylation by cdc2-k appears to result in cytoplasmic retention of T-ag, probably by increasing the affinity for a cytoplasmic anchor protein (see SV40 Large T Antigen, below; and Jans et al., 1991); this effect appears to be titratable. The CKII site is necessary for an accelerated rate of import into the nucleus (steady state within 15-20 minutes as opposed to approximately 10 hours in its absence), probably through increasing the affinity and kinetics of docking at the nuclear envelope/nuclear pore complex. The NLSBP (possibly HSC70/HSP70)-mediated interactions shown are speculative—CKII phosphorylation may conceivably regulate interaction with NLSBPs directly (see Section VI.C). Negative charge is mechanistically important for both cdk and CKII site phosphorylations in the nuclear transport process (Jans et al., 1991; Jans and Jans, 1994), implying that the phosphorylated sites constitute targeting signals themselves, recognized alone or in concert with the NLS, and demonstrating that dephosphorylation at either site is not mechanistically involved in the transport process.

181

182

DAVID A. JANS

tion of the Drosophila morphogen dorsal (Govind and Steward, 1991) and for the p4(y^^ protein (avian homologue of IKB), which may retain the c-rel proto-oncogene in the cytoplasm in analogous fashion. In the case of the former, phosphorylation of dorsal has been shown to effect its release from cactus and nuclear translocation (Whalen and Steward, 1993); dorsal is constitutively nuclear in the absence of cactus (Roth et al., 1989; Steward, 1989). Other IKB family members include lKBY(a discrete gene product identical to the 70-kDa C-terminus of the NF-KB p50 precursor pi 10 (Inoue et al., 1992) and the proto-oncogene bcl-3 (Wulczyn et al., 1992), both of which can bind the NF-KB p50 subunit in vitro and may function to retain it in the cytoplasm in vivo. All of the lytRlcactus family members contain five to seven ankyrin repeats, structural elements involved in protein-protein interactions. Some of these appear to be directly involved in cytoplasmic retention since deletion of ankyrin repeat 7 together with part of 6 inactivates bcl'3 binding of NF-KB p50 (Wulczyn et al., 1992); and deletion analysis of pi 10 indicates that a short sequence including ankyrin repeat 6 and an adjacent acidic sequence are responsible for cytoplasmic retention of the protein (Blank et al., 1991). The mechanism of cytoplasmic retention appears to operate through masking of the NLS in the case of the IKBS (see Nuclear Factor NF-KB and Intra- and Intermolecular Masking, below). Glucocorticoid Receptor and HSP90 The heat shock protein HSP90 plays a cytoplasmic anchor role in retaining the GR in the cytoplasm in the absence of GH (Picard et al., 1988). Hormone binding by GR directly results in dissociation of the complex with HSP90, and NLS-dependent nuclear translocation of the receptor, which is then active for DNA binding and transcription regulation (Picard et al., 1988). More recent results suggest that cell-cycle-dependent dephosphorylation of the GR, probably by phosphatase PP-2A, may also be involved in this nuclear translocation (Hsu et al., 1992; see Table 1). cAMP'Dependent Protein Kinase Catalytic and Regulatory Subunits As alluded to in Section IV, the PK-A C-subunit translocates to the nucleus upon dissociation from the PK-A holoenzyme complex, which

Phosphorylation-Mediated Nuclear Transport

183

occurs subsequent to hormonal stimulation and binding of cAMP by the regulatory (R-) subunit (Nigg et al., 1985; Meinkoth et al., 1990; Pearson et al., 1991). The PK-A R-subunit can be regarded as playing the role of a cytoplasmic anchor, since it functions to retain the C-subunit in the cytoplasm (in the vicinity of the perinuclear Golgi in the case of the type II R-subunit and type II PK-A holoenzyme) (Pearson et al., 1991) in the absence of cAMP-mediated stimulation. c-fos Proto-oncogene

Whereas y-fos is constitutively nuclear, nuclear localization of its c-fos proto-oncogenic counterpart appears to be dependent on cAMPmediated or serum stimulation (Roux et al., 1990). Phosphorylation of c-fos, possibly by PK-A, appears to effect its nuclear translocation by reversing its retention in the cytoplasm, where it is bound to a putative labile inhibitor protein (Roux et al., 1990). SV40 Large T Antigen

Cytoplasmic anchoring appears to be the mechanistic basis of the crfc2-K-mediated inhibition of T-ag nuclear import described in Section V. Maximal inhibition of transport of tetrameric T-ag-p-galactosidase fusion proteins (containing four copies each of NLS and the cdk/cdc2-K site, respectively) is effected by a stoichiometry of phosphorylation of one (rather than four) at the cdc2-K site (Jans et al., 1991), which indicates that one phosphorylated cdc2-K site is sufficient to retain the protein in the cytoplasm, even in the presence of three non-cdc2-K phosphorylated CcN motifs. Phosphorylation presumably increases the affinity of the specific interaction between T-ag and its cytoplasmic anchor. It appears that the inhibitory effect of cdc2-K phosphorylation (or aspartic acid substitution) of Thr^^'* can be oyercome by increasing the concentration of cytosolic T-ag fusion protein (unpublished results), implying that there may be a finite (titratable) cellular level of the cytoplasmic factor capable of retaining Thr^^'^-phosphorylated T-ag in the cytoplasm. B. NLS Masking

A number of proteins possessing apparently functional NLSs when assayed out of context (e.g., as a peptide coupled to a carrier protein) are predominantly cytoplasmic, which is proposed to be due to inaccessibil-

184

DAVID A. JANS

ity or masking of their NLSs. Masking may take place through interaction with another protein (e.g., a factor binding to the NLS itself) or conformational effects whereby the NLS is masked by other parts of the molecule. Phosphorylation is an efficient and potentially rapidly responsive means of regulating/effecting NLS accessibility. Nuclear Factor NF-KB and Intra- and Intermolecular Masking The active (nuclear) form of NF-KB is composed of the p50 and p65 (relA) protein components that contact specific DNA sequences as a homo- or heterodimer (Lenardo and Baltimore, 1989; Schmitz et al., 1991). Both subunits are homologous within a 300 amino acid NLS-containing sequence required for DNA binding and dimerization, which is shared with members of the rel oncogene family and the D. melanogaster developmental control gene dorsal Initial models (see The vd/dorsal Family, above) for the regulation of nuclear localization of the NF-KB TF revolved around the observation that phorbol ester (or other)^ treatment effected release of the NF-KB p65 subunit from an inactive cytoplasmic complex with the cytoplasmic retention factor IKB ( M A D - 3 ) , in order to be transported in NLS-dependent fashion to the nucleus to activate K light chain gene transcription (Lenardo and Baltimore, 1989; Shirakawa and Mizel, 1989; Ghosh and Baltimore, 1990; Schmitz et al., 1991). A dual mechanism involving NLS masking of both NF-KB components has more recently been favored (Beg et al, 1992; Henkel et al, 1992; Zabel et al, 1993), whereby the C-terminus of the pi 10 precursor of N F - K B p50 (which also exists in several cell types as a discrete gene product called iKBy, purported to be functionally comparable to MAD-3; Inoue et al, 1992) has been shown to function to retain the NF-KB p50 subunit in the cytoplasm through intramolecular masking of its NLS. Antibodies specific to the NLS recognize p50 but not pi 10, implying that the NLS is masked in the larger precursor (Henkel et al, 1992). As discussed above (The rd/dorsalFamily), the C-terminus of pi 10, which is absent from p50, appears to mask the NLS directly through specific ankyrin repeat sequences (Blank et al, 1991), one mechanism of unmasking of the NLS being through proteolysis of the pi 10 C-terminus (Blank et al, 1991; Fan and Maniatis, 1991; Riviere et al, 1991). Intermolecular masking of its NLS by IKB (MAD-3) appears to be the mechanism of cytoplasmic retention of NF-KB p65 (Zabel et al, 1993), whereby deletion or mutation of the p65 NLS eliminates binding to IKB

Phosphorylation-Mediated Nuclear Transport

185

(Beg et al., 1992). The mechanism of release from cytoplasmic retention appears to be through IKB degradation (see Beg et al., 1993) induced by PK-C phosphorylation (probably by the PK-C^ isotype, known to be activated by tumor necrosis factor a, one of the physiological stimuli for N F - K B ; Diaz-Meco et al., 1994). Nuclear localization of the NF-KB T F is thus dually regulated by specific intra- and intermolecular masking of the respective NLSs of the two N F - K B subunits. Signal transduction-triggered phosphorylation regulates the masking events precisely to enable rapid response in terms of TF nuclear translocation and gene induction. NLS Masking by Phosphorylation

NLS masking can also be directly effected by phosphorylation, whereby phosphorylation close to or within an NLS masks or inactivates it either through charge or conformational effects. SWI5. Cell-cycle-dependent nuclear exclusion of the 5. cerevisiae TF SWI5 involved in mating type switching is effected by phosphorylation by the cdk/CDC28-kinase (Moll et al., 1991), the yeast homologue of C(ic2-K. Three cdk/CDC28-kinase sites, one of which is within the spacer of the SWI5 bipartite NLS (see Tables 1 and 2), are proposed to inhibit nuclear localization by inactivating or masking the function of the SWI5 NLS through charge or conformational effects (Moll et al., 1991). At anaphase, CDC28 kinase activity falls and SWI5 is dephosphorylated to effect nuclear entry and activation of transcription of the HO mating switch endonuclease; removal by mutation of the CDC28 kinase sites results in constitutive nuclear localization. Lamin B2. NLS-dependent nuclear transport of lamin B2 is inhibited by phosphorylation at two PK-C sites N-terminally adjacent to the NLS (see Table 1) (Hennekes et al, 1993), both in vivo in response to phorbol ester stimulation and in vitro. Negative charge close to the NLS presumably inactivates it, in a fashion analogous to that of the cdk/CDC28 phosphorylation-mediated inactivation of the SWI5 NLS described above. Cofilin, Nuclear translocation of the actin-binding protein cofilin occurs upon heat shock treatment and is accompanied by dephosphorylation at a consensus multifunctional calmoduUn-dependent protein kinase site adjacent to the putative cofilin NLS (see Table 1; Nishida et al..

186

DAVID A. JANS

1987; Ohta et al., 1989). Phosphorylation at this site is proposed to mask or inactivate the function of the NLS (Ohta et al., 1989). C. The CcN Motif: Positive and Negative Regulation by Dual Phosphorylation We have established that the CKII and cdc2-K. phosphorylation sites appear to function completely independently of one another in terms of both regulating nuclear transport and influencing phosphorylation at the other site (Jans et al., 1991); CKII phosphorylation does not enhance nuclear import by reducing phosphorylation at the inhibitory cdc2-K site; nor does Thr^^"^ phosphorylation inhibit nuclear transport by impairing Ser^^^ phosphorylation (Jans et al., 1991). Both phosphorylations almost certainly occur in the cytoplasm (see Scheidtmann et al., 1982; Jans et al., 1991; Jans and Jans, 1994), implying that the regulatory events determining nuclear import kinetics largely take place in the cytoplasm (see Figure 4). This is consistent with the cytosolic requirement for signal dependent nuclear transport in vitro (see Section III). Nuclear phosphorylation at either kinase site to effect retention in the nucleus is clearly not the mechanism of nuclear accumulation. Aspartic acid can substitute functionally for phospho-serine/threonine in the case of both phosphorylation sites of the T-ag CcN motif (Jans et al., 1991; Jans and Jans, 1994; see Figure 3), implying that the phosphorylated sites may represent signals in themselves that are recognized by components of the cellular nuclear transport machinery (see also Jans and Jans, 1994). The results with aspartic acid substituted proteins clearly demonstrate that dephosphorylation/phosphatase activity at either of the CcN motif kinase sites can have no direct role in the nuclear transport process; i.e., nuclear dephosphorylation at either the CKII or cdcl-k site is not involved in T-ag retention in the nucleus (Jans and Jans, 1994). Whereas cytoplasmic retention (see SV40 Large T Antigen, above) would appear to be the basis of the inhibition of T-ag nuclear transport via the cell-cycle-regulated ct/c2-K, the mechanism of CKII-phosphorylation-mediated enhancement of the rate of nuclear transport is not clear. Phosphorylation (or aspartic acid substitution) at the CKII site increases negative charge at the site, and it is intriguing to speculate that the combination of concentrations of negative (phosphorylated or Asp-substituted CKII site) and positive (NLS) charge has a mechanistic role. The molecular basis of this may be analogous to that proposed for eukaryotic proteasome component proteins, which possess both putative NLS and

Phosphorylation-Mediated Nuclear Transport

187

"cNLS" (series of Asp and Glu residues complementary to the NLS) sequences, whereby a masking interaction between the two regions is proposed (Grziwa et al., 1992). It may be a worthwhile exercise to examine the sequences of nuclear proteins that lack a CKII site for regions of negative charge in the vicinity of the NLS, which may conceivably represent, in conjunction with the NLS, a signal for constitutively rapid nuclear transport similar to Asp^^^-substituted T-ag fusion proteins (Jans and Jans, 1994). Another possible explanation of how the CKII site exerts its effect on the NLS-dependent nuclear import of SV40 T-ag proteins is that it facilitates the interaction between the T-ag NLS and NLSBPs, possibly HSP70/HSC70 (Imamoto et al., 1992; Shi and Thomas, 1992). Interestingly in this regard, the spacing between the CKII site and NLS in T-ag is 10 amino acid residues, which coincides with the 10-12-amino acid spacer between the two arms of basic residues required for efficient binding to bipartite NLSs (Dingwall and Laskey, 1992). It is intriguing to speculate that the CKII site, in conjunction with the T-ag NLS, is directly involved in NLSBP interaction, whereby phosphorylation may strongly regulate the affinity of binding. Preliminary evidence (unpublished) from our in vitro system implies that CKII site phosphorylation may increase the rate of interaction ("docking") at the nuclear envelope/nuclear pore complex, which is likely to be NLSBP mediated, and hence is consistent with this idea. Figure 4 is a simplistic representation of some of the events envisaged to be involved in the regulation of T-ag nuclear import by the CcN motif, involving cdc2-K regulation of cytoplasmic retention and CKH regulation of the kinetics of docking at the pore complex. Vll. VARIANTS OF THE CcN MOTIF: OTHER KINASES ABLE TO REGULATE NLS-DEPENDENT NUCLEAR PROTEIN IMPORT A. Signal Transduction-Responsive NLSs CKII clearly regulates T-ag nuclear import kinetics (Rihs et al., 1991; Jans and Jans, 1994), and the presence of CKII sites in the CcN motifs identified in other nuclear localized proteins implies significance in regulating the nuclear uptake of many other proteins. Since its activation is regulated by growth and proliferative signals, it is possible that one of the major physiological roles of CKII is to regulate protein transport to

188

DAVID A. JANS

the nucleus in response to such hormonal signals. However, this putative central role of CKII does not preclude the possibility that the kinases, which participate in hormonally regulated signal transduction pathways, may play a similar role. Evidence that this may indeed be the case already exists for PK-A, PK-C, and the multifunctional calmodulin dependent protein kinase, where phosphorylation sites for these kinases, close to or within the NLS of the respective proteins, regulate NLS function. PK-A is directly implicated in positively regulating nuclear locahzation of the c-rel proto-oncogene (see Table 1), with spacing (22 amino acids) between the PK-A site and NLS having been demonstrated to be of importance (Gilmore and Temin, 1988; Garson and Kang, 1990). In a further parallel with T-ag and CKII, aspartic acid at the PK-A site in c-rel simulates PK-A phosphorylation at the site in terms of the effect on subcellular localization (Mosialos et al., 1991). Results indicate that phosphorylation at the PK-A site enhances nuclear translocation of c-rel. It is noteworthy in this regard that all of the xtMdorsal family TFs possess a homologous region including the NLS and PK-A site (see Schmitz et al., 1991), and significantly, the region of dorsal to which the cytoplasmic retention factor cactus binds has been localized to this region (Kidd, 1992). PK-A phosphorylation of c-rel may decrease its affinity for its specific cytoplasmic retention factor and thereby effect its nuclear translocation. It is an intriguing possibility that the apparent nuclear transport regulatory function of the PK-A site near the NLS of c-rel may be conserved in other rel/dorsal family members. A further example of putative PK-A regulation of nuclear localization is that of rNFIL-6, which translocates to the nucleus upon the elevation of intracellular cAMP levels in rat PC 12 pheochromocytoma cells (Metz and Ziff, 1991). A consensus PK-A site resides adjacent to the putative rNFIL-6 NLS (see Table 1). As for PK-A, evidence thus far implies that both PK-C and the multifunctional calmodulin-dependent protein kinase are signal transduction-responsive kinases able to regulate nuclear protein import (see NLS Masking by Phosphorylation). NLS-dependent nuclear transport of lamin B2 is inhibited by PK-C phosphorylation at sites N-terminally adjacent to the NLS (see Table 1) (Hennekes et al., 1993), whereas nuclear localization of the actin-binding protein cofilin is similarly negatively regulated by a multifunctional calmodulin-dependent protein kinase site N-terminally adjacent to the NLS (see Table 1) (Ohta et al., 1989).

Phosphorylation-Mediated Nuclear Transport

1 89

Clearly, kinases other than CKII can regulate NLS-dependent nuclear transport in response to hormonal signals in a fashion mechanistically similar to the T-ag CcN motif. One can regard these kinase sites together with the NLSs as variants of the CcN motif, where a hormonally regulated signal other than those activating CKII can enhance or inhibit nuclear localization of the appropriate TF or other proteins of a signal transduction pathway. Many other kinases will presumably be demonstrated to be able to function in the same way to regulate nuclear protein transport according to mitogenic, differentiation, and developmental signals. Such constellations of signal transduction responsive kinase sites and NLSs ("N") can be called SRNs (signal transduction-responsive NLSs). A number of SRN variants may well reside in the list of proteins known to exhibit regulated (conditional) nuclear transport in Table 1, where demonstration of this will require direct experimentation of a quantitative nature. B. Cell-Cycle-Dependent NLSs As listed in Table 1, cdks regulate the nuclear transport of a variety of proteins, the best demonstrated examples being cdc2'k and T-ag and CDC28 kinase and SWI5. This regulation appears to affect the maximal amount of nuclear protein accumulated rather than its rate of transport to the nucleus, although an exception appears to be v-jun transport to the nucleus, where cdk (?)-mediated phosphorylation within the NLS is responsible for determining the rate of nuclear transport in cell-cycle-dependent fashion (Chida and Vogt, 1992). Cell cycle dependence of nuclear entry may also be effected directly by kinases other than cdks, although their respective activities may well be directly regulated by cdks. CKII, for example, may be the kinase responsible for nuclear exclusion of the Drosophila lodestar protein through multiple phosphorylation sites in the vicinity of a putative NLS, whereas dorsal nuclear relocalization during development probably involves a Ca^^-regulated kinase and possibly PK-A (see Section VILA; Table 1). NLSs whose function appears to be modulated by cell cycle phosphorylation can be called CDNs (cell-cycle-dependent NLSs). Clearly, the combination of a variety of SR and (directly or indirectly) CD kinases provides a plethora of possible ways to regulate the nuclear translocation of proteins very precisely, in terms of both the amount of protein transported to the nucleus and the rate at which it is imported, as

190

DAVID A. JANS

required during cellular processes such as development, differentiation, transformation, and regulation of cellular metabolism.

VIIL CONCLUSION AND FUTURE PROSPECTS It is quite clear now from a number of studies that the structure and function of NLSs is essentially the same across the plant and animal kingdom, with NLSs from plants, yeast, and higher mammals being essentially identical. NLSs functional in plants are active in nuclear targeting in yeast cells; yeast NLSs are functional in mammalian cells; and the SV40 T-ag NLS appears able to confer nuclear localization on appropriate, normally cytoplasmic carrier proteins in mammalian, insect, yeast, and plant cells. Furthermore, there is the implication that the regulation of NLS activity, primarily by phosphorylation, also appears to be universal in eukaryotic cells. Phosphorylation regulates NLS-dependent nuclear localization of many different proteins and in a variety of cell types. The clear implication is that the signals for conditional NLS-dependent nuclear protein transport are potentially as universal as the NLS signals themselves. Hand in hand with this is the fact that the mechanisms of regulating nuclear protein transport, such as cytoplasmic retention factors, intra- and intermolecular NLS masking, the masking of NLSs by phosphorylation, etc., also appear to be conserved across eukaryotes (e.g., the conservation of structure and function of the IKB, p40'^^', and cactus cytoplasmic retention factor proteins, as well as their specific TF partners, despite their phylogenetically diverse origins). The nuclear targeting regulatory module of the CcN motif, comprising the NLS, CKII site, and cdk site, appears to be conserved in many nuclear localizing proteins from mammals, yeast, insects, and plants (see Table 2). Several other regulatory module variants for NLS-dependent nuclear protein transport appear to exist, whereby kinase sites close to the NLS, other than those for CKII and cdk, modulate NLS function through signal transduction-regulated (SRNs) and cell-cycle-regulated (CDNs) NLSs. Both of these are likely to be as universal as CcN motifs in terms of distribution and function. In terms of future prospects, the rapidly increasing knowledge with respect to regulated nuclear protein transport may be applied to medically pertinent problems, such as the possibility of using NLSs to target molecules of interest to the nucleus. Signals for tightly regulated, conditional nuclear localization responding to hormonal or cell-cycle-dependent stimulation in a variety of cell types (e.g., CcN motifs, CDNs,

Phosphorylation-Mediated Nuclear Transport

191

SRNs) may enable precise cueing of the nuclear localization of relevant proteins and other molecules according to need. This has potential application in gene therapy and gene targeting through facilitating the directed transport of DNA molecules to the nucleus in mammalian (or plant) cells, and thereby potentially increasing transfection/homologous recombination efficiencies (see Rosenkranz et al., 1992; Jans, 1994). Alternatively, toxic molecules such as photosensitizers might be efficiently targeted to sensitive subcellular sites such as the nucleus in order to kill tumor cells (Akhlynina et al., 1993, 1995). The apparent universality of NLSs, and more importantly, the apparent universaUty of the mechanisms of regulation of nuclear protein targeting and CcN motifs and variants thereof, would imply that these approaches are assured of success, above all through their potential widespread appUcability in terms of cell type and protein/gene.

ACKNOWLEDGMENTS Patricia Jans is thanked for helpful discussions, and the Clive and Vera Ramaciotti Foundation is gratefully acknowledged for supporting the work on the CcN motif.

NOTES 1. Several proteins larger than 45 kDa, although lacking an identifiable NLS, have been shown to be predominantly localized in the nucleus. The mechanism appears to be through association with NLS-bearing proteins and cotransport into the nucleus ("piggy back" transport) (e.g., Zhao and Padmanabhan, 1988). 2. Using reconstituted nuclear transport systems, a role in nuclear envelope binding ("docking") has been proposed for a 97-kDa protein together with an NLSBP (Adam and Adam, 1994); and a role in active import into the nucleus has been implicated for the 25-kDa GTP-binding protein Ran/TC4 (and possibly the interacting components RCCl and RanBP-1; see Moore and Blobel, 1993,1994; Melchior et al., 1993). 3. Deletion of either the regulatory or kinase domains of PK-Ca leads to its constitutive nuclear localization (Eldar et al., 1992), purported to be through the unmasking of normally inaccessible NLS sequences in the intact molecule (see Section VLB for examples of NLS masking). Activation of PK-Ca upon phorbol ester binding effects conformational changes that appear to relieve the intramolecular NLS masking in the nonactivated molecule and regulate kinase activity. 4. CKII is regulated by proliferative (as well as other) stimuli such as epidermal growth factor and insulin (Krebs et al., 1988; Meisner and Czech, 1991). 5. Physiological stimuli for NF-KB activation/nuclear localization include the cytokine interleukin (IL)-la. This is interesting in the context of NF-KB, rel, and dorsal

192

DAVID A. JANS

homologies since the toll receptor in the dorsal pathway has homology to the IL-1 receptor. Common nuclear signaling mechanisms are clearly implicated.

REFERENCES Adam, E.J.H. & Adam, S.A. (1994). Identification of cytosolic factors required for nuclear location sequence-mediated binding to the nuclear envelope. J. Cell Biol. 125,547-555. Adam, S.A., Lobl, Th.J., Mitchell, M.A., & Gerace, L. (1989). Identification of specific binding proteins for a nuclear location sequence. Nature 227, 276-279. Addison, C , Jenkins, J.R., & Sturzbecher, H-W. (1990). The p53 nuclear localisation signal is structurally linked to a p34^^^^ motif. Oncogene 5,423-426. Akhlynina, T.V., Yachmenev, S.V., Jans, D.A., Gulak, P.V., Rosenkranz, A.A., Serebryakova, N.V., & Sobolev, A.S. (1993). The use of intemalizable derivatives of chlorin e6 for increasing its photosensitizing activity. Photochem. Photobiol. 58,45-48. Akhlynina, T.V., Rosenkranz, A. A., Jans, D.A., & Sobolev, A.S. (1995). Insulin-mediated intracellular targeting enhances the photodynamic activity of chlorin e6. Cancer Res. 55, 1014-1019. Auwerx, J., Staels, B., & Sassone-Corsi, P. (1990). Coupled and uncoupled induction of fos and jun transcription by different second messengers in cells of hemapoietic origin. Nucleic Acids Res. 18, 221-228. Beg, A.A., Ruben, S.M., Scheinman, R.I., Haskill, S., Rosen, C.A., and Baldwin, A.S., Jr. (1992). IKB interacts with the nuclear localization sequences of the subunits of NF-KB: a mechanism for cytoplasmic retention. Genes Dev. 6, 1899-1913. Beg, A.A., Finco, T.S., Nantermet, RV., & Baldwin, A.S., Jr. (1993). Tumor necrosis factor and interleukin-1 lead to phosphorylation and loss of iKBa: a mechanism for NF-KB activation. Mol. Cell. Biol. 13, 3301-3310. Benditt, J.O., Meyer, C , Fasold, H., Barnard, F.C., and Riedel, N. (1989). Interaction of a nuclear location signal with isolated nuclear envelopes and identification of signal-binding proteins by photoaffinity labelling. Proc. Natl. Acad. Sci. USA 86, 9327-9331. Bischoff, J.R., Friedman, RN., Marshak, D.R., Prives, C , & Beach, D. (1990). Human p53 is phosphorylated by p60-cdc2 and cyclin B-cdc2. Proc. Natl. Acad. Sci. USA 87, 4766^770. Blank, V., Kourilsky, P., & Israel, A. (1991). Cytoplasmic retention, DNA binding and processing of the NF-KB p50 precursor are controlled by a small region in its C-terminus. EMBO J. 10,4159-4167. Cadepond, F , Gasc, J-M., Delahaye, F , Jibaud, N., Schweizer-Groyer, G., SegardMaurel, I., Evans, R., & Baulieu, E-E. (1992). Hormonal regulation of the nuclear locahzation signals of the glucocorticosteroid receptor. Exp. Cell Res. 201, 99108. Capobianco, A.J., Simmons, D.L., & Gilmore, T.D. (1990). Cloning and expression of a chicken c-rel cDNA: unlike p59^''^^', p68^"'^^' is a cytoplasmic protein in chicken embryo fibroblasts. Oncogene 5, 584-591. Chelsky, D., Ralph, R., & Jonak, G. (1989). Sequence requirements for synthetic peptide-mediated translocation to the nucleus. Mol. Cell. Biol. 9, 2487-2492.

Phosphorylation-Mediated Nuclear Transport

193

Chen, R-H., Samecki, C , & Blenis, J. (1992). Nuclear lcx:alization and regulation oferkand rsk-Qncoded protein kinases. Mol. Cell. Biol. 12, 915-927. Chiba, H., Muramatsu, M., Nomoto, A., & Kato, H. (1994). Two human homologues of Saccharomyces cerevisiae SWI2/SNF2 and Drosophila brahma are transcriptional coactivators cooperating with the estrogen receptor and the retinoic acid receptor. Nucleic Acids Res. 22, 1815-1820. Chida, K. & Vogt, RK. (1992). Nuclear translocation of viral Jun but not of cellular Jun is cell cycle dependent. Proc. Nad. Acad. Sci. USA 89,4290-4294. Choubey, D. & Lengyel, P. (1992). Interferon action: nucleolar and nucleoplasmic localization of the interferon-inducible 72-kD protein that is encoded by the Ifi204 gene from the gene 200 cluster. J. Cell Biol. 116, 1333-1341. Citovsky, V., Zupan, J., Wamick, D., & Zambryski, P. (1992). Nuclear localization of Agrobacterium VirE2 protein in plant cells. Science 256, 1802-1804. Dang, C. V. & Lee, W. M. F. (1988). Identification of the human c-myc protein nuclear translocation signal. Mol. Cell. Biol. 8, 4048-4054. De Franco, D.B., Qi, M., Borror, K.C., Garabedian, M.J., & Brautigan, D.L. (1991). Protein phosphatase types 1 and/or 2A regulate nucleocytoplasmic shuttling of glucocorticoid receptors. Mol. Endocrinol. 5,1215-1288. Diaz-Meco, M.T., Dominguez, I., Sanz, L., Dent, P., Lozano, J., Municio, M.M., Berra, E., Hay, R.T., Sturgill, T.W., & Moscat, J. (1994). ^PKC induces phosphorylation and inactivation of iKB-a in vitro. EMBO J. 13, 2842-2848. Dingwall, C. & Laskey, R.A. (1991). Nuclear targeting sequences—a consensus ? Trends Biochem. Sci. 16,478-481. Dingwall, C. & Laskey, R.A. (1992). The nuclear membrane. Science 258,942-948. Eldar, H., Ben-Chaim, J., & Livneh, E. (1992). Deletions in the regulatory or kinase domains of protein kinase C-a cause association with the cell nucleus. Exp. Cell Res. 202, 259-266. Ellisen, L.W., Bird, J., West, D.C., Lee Soreng, A., Reynolds, T.C., Smith, S.D., & Sklar, J. (1991). TAN-1, the human homolog of the Drosophila Notch gene, is broken by chromosomal translocations in T lymphoblastic neoplasms. Cell 66, 649-661. Fan, C-M. & Maniatis, T. (1991). Generation of p50 subunit of NF-KB by processing of pl05 through an ATP-dependent pathway. Nature 354, 395-398. Fleck, O., Michael, H., & Heim, L. (1992). The swi4"^ gene of Schizosaccharomyces pombe encodes a homologue of mismatch repair enzymes. Nucleic Acids Res. 20, 2271-2278. Fortini, M.E., Rebay, I., Caron, L.A., & Artavanis-Tsakonas, S. (1993). An activated Notch receptor blocks cell-fate commitment in the developing Drosophila eye. Nature 365, 555-557. Garcia-Bustos, J., Heitman, J., & Hall, M.N. (1991). Nuclear protein localization. Biochim. Biophys. Acta 1071, 83-101. Garson, K. & Kang, C-Y. (1990). Mapping of the functional domains of the \-rel oncogene. Oncogene 5, 1431-1434. Ghosh, S. & Baltimore, D. (1990). Activation in vitro of NF-KB by phosphorylation of its inhibitor IKB. Nature 344, 678-682. Gilmore, T.D. & Temin, H.M. (1988). \-rel oncoproteins in the nucleus and cytoplasm of transform chicken spleen cells. J. Virol. 62,733-741.

194

DAVID A. JANS

Girdham, C.H. & Glover, D.M. (1991). Chromosome tangling and breakage at anaphase result from mutations in lodestar, a Drosophila gene encoding a putative nucleoside triphosphate-binding protein. Genes Dev. 5,1786-1799. Govind, S. & Steward, R. (1991). Dorsoventral pattern formation in Drosophila. Trends Genet. 7, 119-125. Grasser, F.A., Scheidtman, K-H., Tuazon, P.T., Traugh, T.A., & Walter, G. (1988). In vitro phosphorylation of SV40 large T-antigen. Virology 165, 13-22. Grziwa, A., Dahlmann, B., Cejka, Z., Santarius, U., & Baumeister, W. (1992). Localization of a sequence motif complementary to the nuclear localization signal in proteasomes from Thermoplasma acidophilum by immunoelectron microscopy. J. Struct. Biol. 109, 168-175. Gusse, M., Ghysdael, J., Evan, G., Soussi, T., & Mechali, M. (1989). Translocation of a store of maternal cytoplasmic c-myc protein into nuclei during early development. Mol. Cell. Biol. 9, 5395-5403. Henkel, T, Zabel, U., van Zee, K., Muller, J.M., Fanning, E., & Baeuerle, P.A. (1992). Intramolecular masking of the nuclear location signal and dimerization domain in the precursor for the p50NF-KB subunit. Cell 68, 1121-1133. Hennekes, H., Peter, M., Weber, K., & Nigg, E.A. (1993). Phosphorylation of protein kinase C sites inhibits nuclear import of lamin B2. J. Cell Biol. 120, 1293-1304. Hicks, G.R. & Raikhel, N.V. (1993). Specific binding of nuclear localization sequences to plant nuclei. Plant Cell 5, 983-994. Howard, E.A., Zupan, J.R., Citovsky, V, & Zambryski, P (1992). The VirD2 protein of A. tumefaciens contains a C-terminal bipartite nuclear localization signal: implications for nuclear uptake of DNA in plant cells. Cell 68, 109-118. Hsu, S-C, Qi, M., & de Franco, D.B. (1992). Cell cycle regulation of glucocorticoid receptor function. EMBO J. 11, 3457-3468. Hunt, T. (1989). Cytoplasmic anchoring proteins and the control of nuclear localization Cell 59, 949-951. Imamoto, N., Matsuoka, Y, Kurihara, T, Kohno, K., Miyagi, M., Sakiyama, F, Okada, Y, Tsunasawa, S., & Yoneda, Y (1992). Antibodies to 70kD heat shock cognate protein inhibit mediated nuclear import of karyophilic proteins. J. Cell Biol. 119, 1047-1061. Inoue, J., Kerr, L.D., Kakizuka, A., & Verma, I.M. (1992). iKBy, a 70 kd protein identical to the C-terminal half of pi 10 NF-KB: a new member of the IKB family. Cell 68, 1109-1120. Jans, D.A. (1992). The mobile receptor hypothesis revisited: a mechanistic role for hormone receptor lateral mobility in signal transduction. Biochim. Biophys. Acta 1113,271-276. Jans, D.A. (1994). Nuclear signalling pathways for polypeptide ligands and their membrane-receptors? FASEB J. 8, 841-847. Jans D.A. & Jans, P. (1994). Negative charge at the casein kinase II site flanking the nuclear localization signal of the S V40 large T-antigen is mechanistically important or enhanced nuclear import. Oncogene 9, 2961-2968. Jans, D.A., Ackermann, M., Bischoff, J.R., Beach, D.H., & Peters, R. (1991). p34''^''^mediated phosphorylation at T "^ inhibits nuclear import of SV40 T-antigen proteins. J. Cell Biol. 115, 1203-1212.

Phosphorylation-Mediated Nuclear Transport

195

Jans, D.A., Moll, T, Nasmyth, K., & Jans, P. (1995). Cyclin-dependent kinase-site regulated signal-dependent nuclear localization of the yeast transcription factor SWI5 in mammalian cells. J. Biol. Chem. 270, 17064-17067. Jin, Y.J. & Burakoff, S.J. (1993). The 25-kDa FK506-binding protein is localized in the nucleus and associates with casein kinase II and nucleolin. Proc. Natl. Acad. Sci. USA 90, 7769-7773. Kalderon, D., Richardson, W.D., Markham, A.F., & Smith, A.E. (1984a). Sequence requirements for nuclear location of simian virus 40 large-T-antigen. Nature 311, 33-38. Kalderon, D., Roberts, B.L, Richardson, W.D., & Smith, A.E. (1984b). A short amino acid sequence able to specify nuclear location. Cell 39,499-509. Kidd, S. (1992). Characterization of the Drosophila cactus locus and analysis of interactions between cactus and dorsal proteins. Cell 71, 623-635. Kipreos, E.T. & Wang, J.Y.J. (1990). Differentiation phosphorylation of c-Abl in cell cycle determined by cdc2 kinase and phosphatase activity. Science 256, 217-220. Kleinschmidt, J.A. & Seiter, A. (1988). Identification of domains involved in nuclear uptake and histone binding of nuclear protein N1/N2. EMBO J. 7, 1605-1614. Klimczak, L.J., Schindler, U., & Cashmore, A.R. (1992). DNA binding activity of the Arabidopsis G-box binding factor GBFl is stimulated by phosphorylation by casein kinase II from broccoli. Plant Cell 4, 87-98. Krebs, E.G., Eisenman, R.N., Kuenzel, E.A., Litchfield, D.W., Lozeman, F.J., Luscher, B., & Summercom, J. (1988). Casein kinase II as a potentially important enzyme concerned with signal transduction. Cold Spring Harb. Symp. Quant. Biol. 53, 77-84. Kubota, K., Keith, F.J., & Gay, N.J. (1993). RelocaUzation of Drosophila dorsal protein can be induced by a rise in cytoplasmic calcium concentration and the expression of constitutively active but not wild type Toll receptors. Biochem. J. 296,497-503. Lanford, R.E. & Butel, J.S. (1984). Construction and characterization of an SV40 mutant defective in nuclear transport of T-antigen. Cell 37, 801-813. Laurent, B.C., Yang, X., & Carlson, M. (1993). An essential Saccharomyces cerevisiae gene homologous to SNF2 encodes a helicase-related protein in a new family. Mol. Cell Biol. 12, 1893-1902. Leach, K.L., Powers, E.A., Ruff, V.A., Jaken, S., & Kaufmann, S. (1989). Type 3 protein kinase C localization to the nuclear envelope of phorbol ester-treated NIH 3T3 cells. J. Cell Biol. 109, 685-695. Lee, W.-C, Xue, Z., & Melese, T. (1991). The nsrl gene encodes a protein that specifically binds nuclear localization sequences and has two RNA binding motifs. J. Cell Biol. 113,1-12. Lenardo, M.J. & Baltimore, D. (1989). NF-KB: a pleiotropic mediator of inducible and tissue-specific gene control. Cell 58, 227-229. Li, R. & Thomas, J.O. (1989). Identification of a human protein that interacts with nuclear localization signals. J. Cell Biol. 109, 2623-2632. Liu, J. (1993). FK506 and cyclosporin: molecular probes for studying intracellular signal transduction. Trend Pharmacol. Sci. 14, 182-188. Loewinger, L. & McKeon, F. (1988). Mutations in the nuclear lamin proteins resulting in their aberrant assembly in the cytoplasm. EMBO J. 7, 2301-2309.

196

DAVID A. JANS

Luscher, B., Kiinzel, E.A., Krebs, E.G., & Eisenman, R.N. (1989). Myc oncoproteins are phosphrylated by casein kinase IIEMBO J. 8, 1111-1119. Luscher, B., Christenson, E., Litchfield, D., Krebs, E.G., & Eisenman, R.N. (1990). Myb DNA binding inhibited by phosphorylation at a site deleted during oncogenic activation. Nature 344, 517-522. Matthews, R.J., Bowne, D.B., Flores, E., & Thomas, M.L. (1992). Characterization of hematopoietic intracellular tyrosine phosphatases: description of a phosphatase containing an SH2 domain and another enriched in proline-, glutamic acid-, serine-, and threonine-rich sequences. Mol. Cell. Biol. 12, 2396-2405. McVey, D., Brizuela, L., Mohr, I., Marshak, D.R., Gluzman, Y, & Beach, D. (1989). Phosphorylation of large tumor antigen by cdc2-simulates SV40 DNA replication. Nature 341, 503-507. Meier, U.T. & Blobel, G. (1990). A nuclear localization signal binding protein in the nucleolus. J. Cell Biol. I l l , 2235-2245. Meinkoth, J.L., Ji, Y, Taylor, S.S., & Feramisco, J.R. (1990). Dynamics of the distribution of cyclic-AMP-dependent protein kinase in living cells. Proc. Natl. Acad. Sci. USA 87,9595-9599. Meisner, H. & Czech, M.P. (1991). Phosphorylation of transcriptional factors and cell-cycle-dependent proteins by casein kinase II. Curr. Opin. Cell Biol. 3, 474483. Melchior, P., Paschal, B., Evans, J., & Gerace, L. (1993). Inhibition of nuclear protein import by nonhydrolyzable analogues of GTP and identification of the small GTPase Ran/TC4 as an essential transport factor. J. Cell Biol. 123,1649-1659. Metz, R. & Ziff, E. (1991). cAMP stimulates the C/EBP-related transcription factor rNFIL-6 to translocate to the nucleus and induce c-fos transcription. Genes Dev. 5,1754-1766. Miller, M., Park, M.Y, & Hanover, J. A. (1991). Nuclear pore complex: structure function and regulation. Physiol. Rev. 71, 909-949. Moll, T, Tebb, G., Surana, U., Robitsch, H., & Nasymth, K. (1991). The role of phosphorylation and the CDC28 protein kinase in cell cycle-regulated nuclear import of the S. cerevisiae transcription factor SWI5. Cell 66, 1-20. Moore, M.S. & Blobel, G. (1993). The GTP-binding protein Ran/TC4 is required for protein import into the nucleus. Nature 365, 661-663. Moore, M.S. & Blobel, G. (1994). AG protein involved in nucleocytoplasmic transport: the role of Ran. Trends Biochem. Sci. 19, 211-216. Mosialos, G., Hamer, P., Capobianco, A.J., Laursen, R.A., & Gilmore, T.D. (1991). A protein-kinase-A recognition sequence is structurally linked to transformation by p59''"'*^' and cytoplasmic retention of p68'^"*'''. Mol. Cell Biol. 11, 5867-5877. Nelson, M. & Silver, PA. (1989). Context affects nuclear protein localization in Saccharomyces cerevisiae. Mol. Cell Biol. 9, 384-389. Newmeyer, D.D, & Forbes, D.J. (1988). Nuclear import can be separated into distinct steps vitro: nuclear pore binding and translocation. Cell 52, 641-653. Nigg, E.A., Hilz, H., Eppenberger, H.M., & Dulty, F (1985). Rapid and reversible translocation of the catalytic subunit of cAMP-dependent protein kinase type II from the Golgi complex to the nucleus. EMBO J. 4, 2801-2807.

Phosphorylation-Mediated Nuclear Transport

197

Nishida, E., lida, K„ Yonezawa, N., Koyasu, S., Yahara, I., & Sakai, H. (1987). Cofilin is a component of intranuclear and cytoplasmic actin rods induced in cultured cells. Proc. Nad. Acad. Sci. USA 84, 5262-5266. Ohta, Y, Nishida, E., Sakai, H., & Miyamoto, E. (1989). Dephosphorylation of cofilin accompanies heat shock-induced nuclear accumulation of cofilin. J. Biol. Chem. 264,16143-16148. Pearson, D., Nigg, E.A., Nagamine, Y, Jans, D.A., & Hemmings, B.A. (1991). Mechanisms of cAMP-mediated gene induction: examination of renal epithelial cell mutants affected in the catalytic subunit of the cAMP-dependent protein kinase. Exp. Cell Res. 192,315-318. Pellegrini, S. & Schindler, C. (1993). Early events in signalling by interferons. Trends Biochem. Sci. 18,338-342. Peters, R. (1986). Fluorescence microphotolysis to measure nucleocytoplasmic transport and intracellular mobility. Biochim. Biophys. Acta 864, 305-359. Picard, D., Salser, S.J., & Yamamoto, K.R. (1988). Amovable and regulable inactivation function within the steroid binding domain of the glucocorticoid receptor. Cell 54, 1073-1080. Prendergast, G.C., Hopewell, R., Gorham, B.J., & Ziff, E.B. (1992). Biphasic effect of Max on Myc cotransformation activity and dependence on amino- and carboxyterminal Max functions. Genes Dev. 6, 2429-2439. Qi, M., Hamilton, B.J., & De Franco, D. (1989). \-mos oncoproteins affect the nuclear retendon and reutDization of glucocorticoid receptors. Mol. Endocrinol. 3,12791288. Raikhel, N.V. (1992). Nuclear targeting in plants. Plant Physiol. 100,1627-1632. Reinhard, C , Fernandez, A., Lamb, N.J.C., & Thomas, G. (1994). Nuclear localization of p85^^^: functional requirement for entry into S phase. EMBO J. 13,1557-1565. Richardson, W.D., Roberts, B.L., & Smith, A.E. (1986). Nuclear location signals in polyoma virus large-T. Cell 44,77-85. Rihs, H.-P. & Peters, R. (1989). Nuclear transport kinetics depend on phosphorylationsite-containing sequences flanking the karyophilic signal of the simian virus 40 T-antigen EMBO J. 8,1479-1484. Rihs, H.-R, Jans, D.A., Fan, H., & Peters, R. (1991). The rate of nuclear cytoplasmic protein transport is determined by a casein kinase II site flanking the nuclear localization sequence of the SV40 T-antigen. EMBO J. 10,633-639. Riviere, Y, Blank, V., Kourilsky, P., & Israel, A. (1991). Processing of the precursor of NF-KB by the HIV-1 protease during acute infection. Nature 350,625-626. Robbins, J., Dilworth, S.M., Laskey, R.A., & Dingwall, C. (1991). Two interdependent basic domains in nucleoplasmin nuclear targeting sequence: identification of a class of bipartite nuclear targeting sequence. Cell 64, 615-623. Rosenkranz, A.A., Yachmenev, S.V., Jans, D.A., Serebryakova, N.V., Murav'ev, V.I., Peters, R., & Sobolev, A.S. (1992). Receptor-mediated endocytosis and nuclear transport of a transfecting DNA construct. Exp. Cell Res. 199, 323-329. Roth, S., Stein, D., & Nusslein-Volhard, C. (1989). A gradient of nuclear localization of the dorsal protein determines dorsoventral pattern in the Drosphila embryo. Cell 59, 1189-1202.

198

DAVID A. JANS

Roux, P., Blanchard, J.-M., Fernandez, A., Lamb, N., Jeanteur, Ph., & Piechaczyk, M. (1990). Nuclear localization of c-fos, but not v-fos proteins, is controlled by extracellular signals. Cell 63, 341-351. Sawyers, C.L., McLaughlin, J., Goga, A., Hvlik, M., & Witte, O. (1994). The nuclear tyrosine kinase c-Abl negatively regulates cell growth. Cell 77, 121-131. Scheidtmann, K.-H., Echle, B., & Walter, G. (1982). Simian virus 40 large T-antigen is phosphorylated at multiple sites clustered in two separate regions. Virology 44, 116-133. Schindler, C , Shuai, K., Prezioso, V.R., & Darnell, J.E., Jr. (1992). Interferon-dependent tyrosine phosphorylation of a latent cytoplasmic transcription factor. Science 257, 809-813. Schmitz, M.L., Henkel, T., & Baeuerle, PA. (1991). Proteins controlling the nuclear uptake of NF-KB, Rel, and dorsal. Trends Cell Biol. 1, 130-137. Shi, Y. & Thomas, J.O. (1992). The transport of proteins into the nucleus requires the 70-kilodalton heat shock protein or its cytosolic cognate. Mol. Cell Biol. 12, 2186-2192. Shiozaki, K. & Yanagida, M. (1992). Functional dissection of the phosphorylated termini of fission yeast DNA topoisomerase II. J. Cell Biol. 119, 1023-1036. Shirakawa, F. & Mizel, S.B. (1989). In vitro activation and nuclear translocation of NF-KB catalyzed by cyclic AMP-dependent protein kinase and protein kinase C. Mol. Cell Biol. 9, 2424-2430. Shuai, K., Schindler, C , Prezioso, V.R., & Darnell, J.E., Jr. (1992). Activation of transcription by IFN-gamma: tyrosine phosphorylation of a 91-kD DNA binding protein. Science 258, 1808-1812. Silver, PA. (1991). How proteins enter the nucleus. Cell 64,489-497. Simons, K. & Virta, H. (1987). Perforated MDCK cells support intracellular transport. EMBO J. 6, 2241-2247. Standiford, D.M. & Richter, J.D. (1992). Analysisof a developmentally regulated nuclear localization signal in Xenopus. J. Cell Biol. 118, 991-1002. Steward, R. (1989). Relocalization of the dorsal protein from the cytoplasm to the nucleus correlates with its function. Cell 59, 1179-1188. Stifani, S., Blaumueller, CM., Redhead, N.J., Hill, R.E., & Artavanis-Tsakonas, S. (1992). Human homologs of a Drosophila enhancer of Split gene product define a novel family of nuclear proteins. Nature Genet. 2, 119-127. Stochaj, U., Osborne, M., Kurihara, T, & Silver, PA. (1991). A yeast protein that binds nuclear localization signals: purification, localization, and antibody inhibition of binding activity. J. Cell Biol. 113, 1243-1254. Templeton, D.J. (1992). Nuclear binding of purified retinoblastoma gene product is determined by cell cycle-regulated phosphorylation. Mol. Cell. Biol. 12,435-443. Templeton, D.J., Park, S.H., Lanier, L., & Weinberg, R.A. (1991). Nonfunctional mutants of the retinoblastoma protein are characterized by defects in phosphorylation, viral oncoprotein association, and nuclear tethering. Proc. Natl. Acad. Sci. USA 88, 3833-3837. Tinland, B., Koukolikova-Nicola, Z., Hall, M.N., & Hohn, B. (1992). The T-DNA-linked VirD2 protein contains two distinct functional nuclear localization signals. Proc. Natl. Acad. Sci. USA 89, 7442-7446.

Phosphorylation-Mediated Nuclear Transport

199

Trapani, J., Browne, K.A., Dawson, MJ., Ramsay, R.G., Eddy, R.L., Shows, T.B., White, RC, & Dupont, B. (1992). A novel gene constitutively expressed in human lymphoid cells is inducible with interferon-yin myeloid cells. Immunogenetics 36, 369-376. Tratner, L, Ofir, R., & Verma, I.M. (1992). Alteration of a cyclic AMP-dependent protein kinase phosphorylation site in the c-Fos protein augments its transforming potential. Mol. Cell. Biol. 12, 998-1006. Troelstra, C , van Gool, A., de Wit, J., Vermeulen, W, Bootsma, D., & Hoeijmakers, J.H.J. (1992). ERCC6, a member of a subfamily of putative helicases, is involved in Cockayne's sydrome and preferential repair of active genes. Cell 71,939-953. Van Etten, R.A., Jackson, R, & Baltimore, D. (1989). The mouse type IV c-abl gene product is a nuclear protein, and activation of transforming ability is associated with cytoplasmic localization. Cell 58, 669-678. Varagona, M.J., Schmidt, R.J., & Raikhel, N.V. (1992), Nuclear localization signal(s) required for nuclear targeting of the maize regulatory protein opaque-2. Plant Cell 4,1213-1227. Wahle, E. Martin, G., Schitz, E., & Keller, W. (1991). Isolation and expression of cDNA clones encoding mammalian poly(A)polymerase. EMBO J. 10,4251-4257. Ward, G.E. & Kirschner, M.W. (1990). Identification of cell cycle-regulated phosphorylation sites on nuclear lamin C. Cell 61,561-577. Whalen, A.M. & Steward, R. (1993). Dissociation of the dorsal-cactus complex and phosphorylation of the protein correlate with the nuclear localization of dorsal. J. Cell Biol. 123, 523-534. Wiche, G., Gromov, D., Donovan, A., Castanon, M.J., & Fuchs, E. (1993). Expression of plectrin mutant cDNA in cultured cells indicates a role of COOH-terminal domain in intermediate filament association. J. Cell Biol. 121, 607^19. Wulczyn, F.G., Naumann, M., & Scheidereit, C. (1992). Candidate proto-oncogene bcl-3 encodes a subunit-specific inhibitor of transcription factor NF-KB. Nature 358, 597-599. Yoneda, Y., Imamoto-Sonobe, N., Matsuoka, Y, Iwamoto, R., Kiho, Y, & Uchida, T. (1988). Antibodies to asp-asp-asp-glu-asp can inhibit transport of nuclear proteins into the nucleus. Science 242, 275-378. Yoneda, Y, Semba, T., Kaneda, Y, Noble, R.L., Matsuoka, Y, Kurihara, T, Okada, Y, & Imamoto, N. (1992). A long synthetic peptide containing a nuclear localization signal and its flanking sequences of SV40 T-antigen directs the transport of IgM into the nucleus efficiently. Exp. Cell Res. 201, 313-320. Zabel, U., Henkel, T., dos Santos Silva, M., & Baeuerle, RA. (1993). Nuclear uptake control of NF-KB by MAD-3, an IKB protein present in the nucleus. EMBO J. 12, 201-211. Zacksenhaus, E., Bremner, R., Phillips, R.A., & Gallic, B.L. (1993). A bipartite nuclear localization signal in the retinoblastoma gene product and its importance for biological activity. Mol. Cell. Biol. 13,4588-^599. Zhao, L-J. & Padmanabhan, R. (1988). Nuclear transport of adenovirus DNA polymerase is facilitated by interaction with preterminal protein. Cell 55, 1005-1015.

This Page Intentionally Left Blank

MEMBRANE PROTEIN TOPOGENESIS IN ESCHERICHIA COLI

Gunnar von Heijne

I. Intrcxiuction II. Mechanisms of Membrane Protein Assembly A. The "Positive Inside" Rule B. The Sec Machinery C. The Membrane Potential III. Helix-Helix Packing in a Lipid Environment IV. Predictionof Topology and Structure V. Artificial Membrane Proteins VI. Conclusions and Outlook References

201 202 202 205 206 208 209 210 211 211

I. INTRODUCTION How does the amino acid sequence of a membrane protein dictate its insertion into the membrane and how does it determine the final threedimensional fold? This deceptively simple question has gradually moved

Membrane Protein IVansport Volume 2, pages 201-214 Copyright © 1995 by JAI Press Inc. All rights of reproduction in any form reserved. ISBN: 1-55938-983-4

201

202

GUNNARVONHEIJNE

to the forefront of membrane protein research, and it now excites workers vyith interests ranging from bacterial genetics and cell biology to protein structure prediction and modeling. As will be reviewed in this chapter, our own work has largely focused on membrane protein assembly in the inner membrane of the bacterium Escherichia coli and has led to the formulation of the "positive inside" rule—a shorthand intended to capture the essential role played by positively charged amino acids in determining the topology of proteins spanning the cytoplasmic membrane. Although all of the mechanistic details of the membrane protein assembly processes in various cell types and organelles are still far from being worked out, our understanding is already sufficiently advanced to allow good predictions of protein topology to be made directly from the amino acid sequence, and to suggest how "artificial" membrane proteins of a predetermined topology may be designed. II. MECHANISMS OF MEMBRANE PROTEIN ASSEMBLY A. The "Positive Inside" Rule At the most simple level, the structure of an integral membrane protein can be characterized by its membrane topology, that is, by a diagram showing which parts are exposed on each side of the membrane. The first question that must be addressed when trying to understand the entire assembly process is thus whether the formation of the topology corresponds to a distinct step in the assembly pathway, or whether the formation of the topology and the formation of the entire 3D structure are inseparable processes. Fortunately, there is now ample evidence in support of the former view, both from biophysical studies on the reformation of the native structure of the multi-spanning protein bacteriorhodopsin from fragments that individually retain the same topology as in the intact molecule (Popot et al., 1987; Kahn and Engelman, 1992; Kahn et al., 1992; Kataoka et al., 1992), and from molecular genetic studies where pieces of inner membrane proteins have been expressed and shown to insert into the membrane separately while still being able to ultimately assemble into functional molecules (Zen et al., 1994). Thus, the concept of a two-step folding pathway (Popot and de Vitry, 1990; Popot and Engelman, 1990), where the membrane-spanning segments are first inserted individually and fold into stable transmembrane a-heli-

Membrane Protein Topogenesis

203

ces and only later pack together to form the full 3D structure, is now well established. If the topology is thus more than an abstract representation of the structure and in fact corresponds to a real intermediate on the folding pathway, it may be fruitful to ask whether there is some simple relation between the primary amino acid sequence and the topology. Indeed, statistical studies of membrane proteins of known topology from a range of different cellular membranes have demonstrated that there is a strong correlation between the location of positively charged amino acids and the topology; such residues abound in segments of the chain that are not translocated across the membrane, but are scarce in translocated segments (von Heijne, 1986; von Heijne and Gavel, 1988; Gavel et al., 1991; Gavel and von Heijne, 1992). In particular, this "positive inside" rule holds exceptionally well for E. coli inner membrane proteins, an observation that prompted us to undertake a series of experimental studies intended to clarify the role of basic amino acids in controlling the topology of bacterial inner membrane proteins in vivo. Partly by serendipity, our first studies were conducted using E. coli leader peptidase (Lep), a protein that spans the inner membrane twice with an NQ^J-CQU^ topology (Figure 1). It proved possible to redesign this molecule such that it inserted in the opposite topology (Nij^-C^j^) by

periplasm

Lep wt

Figure 1. Topology of wild-type Lep [left), a deletion mutant where most of the charged residues in the cytoplasmic loop have been removed {middle), and a mutant with an "inverted'' topology [right) that was constructed by adding four positively charged lysine residues to the N-terminus of the deletion mutant (von Heijne, 1989).

204

GUNNARVONHEIJNE

periplasm

cytoplasm

Figure 2. Topology of a "frustrated" construct where three positively charged residues have been placed in the first and second loops. This forces the molecule to adopt a "leave-one-out" topology with only three transmembrane segments. The construct was made by fusing two gene fragments encoding the membrane domains of two Lep mutants in tandem (Gafvelin and von Heijne, 1994).

simply removing most of the positively charged amino acids normally present in the loop connecting the two transmembrane segment while at the same time adding a few positively charged lysines to the N-terminal tail (von Heijne, 1989). This result opened the way for studies where the topological effects of both positively and negatively charged residues could be analyzed in detail (Nilsson and von Heijne, 1990; Andersson and von Heijne, 1991; Andersson etal., 1992; Andersson and von Heijne, 1993a), and ultimately, together with data from other groups (Boyd and Beck with, 1990; Dalbey, 1990), provided ample experimental verification of the positive inside rule. More recently, we have extended our studies of the topological effects of positively charged residues to proteins with up to four transmembrane segments (Gafvelin and von Heijne, 1994). An interesting result was that molecules where the location of the positively charged residues was intentionally chosen to "frustrate" the molecule (Figure 2) inserted into the inner membrane with "leave-one-out" topologies, that is, the requirement that all highly charged loops should remain in the cytoplasm forced at least one putative transmembrane segment to stay out of the membrane. From these and other results, we concluded that the insertion process is locally determined and probably proceeds by the independent

Membrane Protein Topogenesis

205

insertion of individual "helical hairpins" (Engelman and Steitz, 1981) composed of two neighboring transmembrane segments and a connecting loop. B. The Sec Machinery Periplasmic and outer membrane E. coli proteins normally make use of the so-called Sec machinery for translocation across the inner membrane. The secretory pathway involves cytoplasmic chaperones such as Ffh, GroES/EL, DnaK, and SecB; the SecAAf/E/G complex in the inner membrane; and the additional inner membrane components SecD and SecF (Schatz and Beckwith, 1990; Wickner et al., 1991; Luirink and Dobberstein, 1994; Nishiyama et al., 1994; Pogliano and Beckwith, 1994). However, certain inner membrane proteins are able to insert efficiently into the membrane even under conditions when the function of one or the other of the Sec components is severely compromised. In particular, this is true for the "inverted" form of Lep (von Heijne, 1989). Translocation of the C-terminal periplasmic domain in wild-type Lep, in contrast, requires a fully functional Sec machinery (Wolfe and Wickner, 1984; Wolfe etal., 1986). A possible relation between the length of a translocated segment and its dependence on the Sec machinery for translocation was suggested by a statistical study of the content of positively charged residues in periplasmic loops as a function of loop length (von Heijne and Gavel, 1988; von Heijne, 1994). Although most periplasmic loops tend to be short and contain only few arginines and lysines, loops longer than -60 residues often contain a much higher frequency of such residues, similar to the overall frequency found in periplasmic proteins. A possible explanation would be that the mechanism of translocation is different for short and long loops, and that this leads to different restrictions on the amino acid composition of the loops. This hypothesis was confirmed in a study where the length of the periplasmic loop in "inverted" Lep molecules was successively increased from -25 to -65 residues (Andersson and von Heijne, 1993b). For this series of molecules, the SecA dependence of translocation was found to increase in proportion to the loop length, up to a limiting value of about 55 residues. A qualifying remark should be added at this point: different loops of similar lengths may show different degrees of Sec dependence, depending on their amino acid composition and on the surrounding transmem-

206

GUNNARVONHEIJNE

brane segments. Thus, when the positively charged residues in a long periplasmic loop placed in an "inverted" Lep construct were successively mutated to uncharged residues, the degree of SecA dependence dropped proportionately (Andersson and von Heijne, 1994b). In another series of experiments, we recendy tested the SecA dependence of the translocation of a 180-residue-long periplasmic loop in the MalF protein and found that whereas translocation of this loop when in its normal sequence context is only marginally affected by perturbation of SecA function, the SecA dependence is dramatically increased when the loop is placed out of context in an "inverted" Lep construct (Saaf et al, 1994). In all but one of the cases that we have looked at, the translocation of long loops (>60 residues) has been clearly affected by perturbation of the Sec machinery, although to varying degrees. We have encountered only one exception to the rule that the translocation of long periplasmic segments is Sec dependent: the N-terminal tail of the ProW inner membrane protein (Whitley et al., 1994b). ProW spans the inner membrane seven times, and its N-terminal, 100-residuelong tail is located in the periplasm. Interestingly, the N-tail contains only three positively charged residues (but 12 negatively charged ones), suggesting that this particular tail, although long, is under the same selective pressure as the normally much shorter. Sec-independent loops. Indeed, translocation of the ProW N-tail was found to be independent of SecA within the accuracy of our measurements, and it was completely blocked by the introduction of extra positively charged residues. Generalizing from this one example, it is possible that the Sec machinery can only translocate polypeptide stretches located C-terminally to a translocation signal, and that N-terminally located segments may only be translocated by a Sec-independent mechanism regardless of their length. So far, we have thus found that Sec dependence tends to correlate with the length of the translocated segment, with its content of positively charged residues, and with its location within the protein. The issue of what a high versus low Sec dependence means in mechanistic terms is entirely unclear, however, and may only be possible to address using reconstituted in vitro systems where the different Sec components can be physically removed. C. The Membrane Potential What, finally, may be the mechanistic basis for the positive inside rule? An attractive possibihty is that the membrane electrochemical

Membrane Protein Topogenesis

207

periplasm

cytoplasm

Figure 3. A Lep mutant that adopts different topologies depending on whether a transmembrane electrochemical potential is present during the insertion step. In the presence of a potential (i.e., in the absence of the protonophore CCCP), the negatively charged N-terminal tail is translocated in preference to the loop carrying one positively and one negatively charged residue (left). In the absence of a potential, in contrast, the loop is more easily translocated than the N-terminal tail {middle). If the function of SecA is blocked by the addition of sodium azide, translocation of the Sec-dependent C-terminal domain is prevented but the topology is not affected (right) (Andersson and von Heijne, 1994a).

potential (acidic and positive toward the periplasm) may act against the translocation of positively charged residues. To test this notion, we expressed various Lep-derived constructs both in normal cells and under conditions where the electrochemical potential had been dissipated by addition of the protonophore CCCP (Andersson and von Heijne, 1994a). To our surprise, certain constructs were found to insert with different orientations when expressed in the absence or the presence of CCCP, (Figure 3), and their insertion behavior could in all cases be explained by a model which postulates that the electrochemical potential promotes the translocation of negatively charged residues and works against the translocation of positively charged ones, thus introducing an asymmetry into the topological effects of positively and negatively charged amino acids. Although this study must be followed up, it provides a first hint that interactions between the membrane potential and charged residues in the nascent polypeptide may explain the positive inside rule.

208

GUNNARVONHEIJNE

IIL HELIX-HELIX PACKING IN A LIPID ENVIRONMENT According to the two-step folding model, the transmembrane helices are first formed individually in the membrane, and only then pack together into the final helix-bundle structure. The energetics of helix-helix interactions in a lipid environment are largely unknown, and hence prediction of the 3D structure of integral membrane proteins is in its infancy. In part, this is the result of the lack of high-resolution 3D structures for membrane proteins; to date, such information is available only for three helix-bundle membrane proteins: the photosynthetic reaction center (Deisenhofer et al., 1985), bacteriorhodopsin (Henderson et al, 1990), and a light-harvesting polypeptide (Kiihlbrandt et al., 1994). These difficulties have prompted the development of altemative means of obtaining 3D structure information. In particular, methods based on molecular genetic techniques seem quite promising. The most complete analysis of this kind has so far been carried out on the glycophorin A homo-dimer, a single-spanning plasma membrane protein where dimerization is driven by interactions between the two transmembrane helices. Saturation mutagenesis has identified seven critical residues on one face of the helix that are necessary and sufiicient for dimerization (Lemmon et al, 1992,1994); all of the other residues in the helix can be replaced by leucines with only minor effects on the monomer-dimer equilibrium. Both small nonpolar and large hydrophobic residues are found among the critical residues, and it seems that dimerization is driven by the tight fit between the two complementary helix surfaces rather than by a few specific interactions between, for example, charged or polar residues embedded in the hydrophobic helices. Another recent method is based on the idea that cysteine residues introduced into neighboring transmembrane helices by site-directed mutagenesis may form disulfide bridges if located close to each other. This approach has so far been tried on two E. coli inner membrane proteins, the Tar chemoreceptor (Pakula and Simon, 1992), and leader peptidase (Whitley et al, 1993). In both cases it was found that the observed pattern of disulfides could be interpreted in terms of a 3D model, but the method has yet to be tested on a known 3D structure. Other techniques that can yield distance information are, for example, various energy transfer measurements (Lakey et al, 1993; Adair and Engelman, 1994), and the analysis of second-site revertants of nonfunctional point mutations in transmembrane helices (Sahin-Toth and Kaback, 1993).

Membrane Protein Topogenesis

209

IV. PREDICTION OF TOPOLOGY AND STRUCTURE To predict the topology of a membrane protein from its amino acid sequence, one needs (i) to predict the approximate location of all the transmembrane segments, and (ii) to decide on the orientation of the molecule in the membrane. Transmembrane segments are usually identified by the aid of a hydrophobicity plot, where the average hydrophobicity over a 15-20-residue-long window is plotted as a function of position in the chain. Various optimization schemes intended to perfect the underlying hydrophobicity scale have been proposed (Klein et al., 1985; Edelman and White, 1989; Edelman, 1993), but many of the early rough-and-ready scales perform almost as well, and the choice of scale is more a matter of taste than rational deliberation. None of the available hydrophobicity analysis algorithms allow a clear distinction between transmembrane and nontransmembrane segments to be made, however. Thus, in many cases one is left with one or more uncertain candidate transmembrane segments that can be combined into a number of different topologies. In such cases, the positive inside rule can be of help in deciding which of the possible topologies is the most likely one; for each topology, one simply calculates the difference in the number of positively charged residues between the two sides of the structure (counting only short loops of 60 residues or less, but always including the N-terminal tail, irrespective of its length; see above) and then chooses the topology with the highest difference as the best guess. This procedure has been automated (Claros and von Heijne, 1994) and works very well for bacterial inner membrane proteins (von Heijne, 1992) and somewhat less well for eukaryotic membrane proteins (Sipos and von Heijne, 1993). Amore refined version of this basic approach has recently been developed (Jones et al., 1994); to what extent this method performs better than the simple version is not clear. As noted above, the prediction of optimal helix-helix packing interactions is still an unsolved problem, although a few early attempts have been made. A systematic study of all possible ways of arranging the seven transmembrane helices in bacteriorhodopsin on a lattice, with an optimization criterion based on the maximization of contacts between polar residues and the minimization of contacts between evolutionarily highly variable residues, places the observed arrangement near the top (Taylor et al., 1994). Amore detailed calculation on bacteriorhodopsin where the helical axes are kept fixed in space but where each helix is allowed to

210

GUNNARVONHEIJNE

rotate around and slide along its axis also produces an optimal structure close to the correct one (Tuffery et al., 1994). The arguably most realistic prediction attempted so far has been carried out for the glycophorin A dimer, where a simulated annealing molecular dynamics protocol was used to identify low-energy structures (Treutlein et al, 1992; Lemmon et al, 1994). Again, the structure identified by the mutagenesis approach described above was among the ones of lowest energy. Taken together, these studies suggest that good, generally applicable methods of making predictions of the 3D structure of integral membrane helix-bundle proteins may not be far off.

V. ARTIFICIAL MEMBRANE PROTEINS As is clear from the previous sections, a rather simple picture of the sequence factors dictating membrane protein insertion is emerging: hydrophobic stretches of amino acids drive the insertion and end up forming transmembrane helices, and positively charged residues flank-

periplasm

cytoplasm

Figure 4. Topology of a de novo designed inner membrane protein with four highly simplified transmembrane segments. Each transmembrane segment is composed only of leucine and alanine residues, and the topology is controlled by appropriately placed lysine residues (Whitley et a l , 1994a).

Membrane Protein Topogenesis

211

ing the hydrophobic segments control their orientation in the membrane. As a stringent test of this idea, and as a first step toward the construction of completely novel membrane proteins, we recently decided to attempt the design and expression of artificial membrane proteins composed of highly simplified transmembrane segments and appropriately located, positively charged residues. So far, proteins with up to four transmembrane segments composed only of leucines and alanines have been successfully expressed in E, coli and have been shown to insert efficiently and with the predicted topology into the inner membrane (Figure 4) (Whitley et al., 1994a). These proteins do not appear to be toxic to the cell and can be produced in reasonable yields, providing the first indication that the in vivo expression of de novo designed membrane proteins is a viable idea.

VI. CONCLUSIONS AND OUTLOOK Over the past few years, studies reviewed here and elsewhere (von Heijne, 1994) have led to a fairly advanced understanding of how integral membrane proteins insert into membranes, both from the point of view of the sequence determinants that encode the topogenic information and in terms of the mechanism of integration. Based on this understanding, rather reliable prediction schemes that allow the topology of a protein to be deduced directly from its amino acid sequence have been developed, and completely novel membrane proteins have been successfully designed and expressed in vivo. In the next few years, it should be possible to further decipher the role of the Sec machinery during the insertion process by studies both in vitro and in vivo. Also, the corresponding processes in eukaryotic cells (e.g., integration into the ER membrane and into the inner membrane of mitochondria) must be studied in a systematic way. Perhaps the most exciting prospect for the coming years, however, is the possibility of designing artificial membrane proteins that may provide new and powerful tools for studying important biological processes such as ion channel function and electron transport phenomena. REFERENCES Adair, B.D. & Engelman, D.M. (1994). Glycophorin A helical transmembrane domains dimerize in phospholipid bUayers: a resonance energy transfer study. Biochemistry 33, 5539-5544.

212

GUNNARVONHEIJNE

Andersson, H. & von Heijne, G. (1991). A 30-residue-long "export initiation domain" adjacent to the signal sequence is critical for protein translocation across the inner membrane of Escherichia coli. Proc Natl. Acad. Sci. USA 88,9751-9754. Andersson, H. & von Heijne, G. (1993a). Position-specific Asp-Lys pairing can affect signal sequence-function and membrane protein topology. J. Biol. Chem. 268, 21389-21393. Andersson, H. & von Heijne, G. (1993b). 5^c-dependent and sec-independent assembly of E. coli inner membrane proteins: the topological rules depend on chain length. EMBO J. 12,683-691. Andersson, H. & von Heijne, G. (1994a). Membrane protein topology: effects of A^H^ on the translocation of charged residues explain the "positive inside" rule. EMBO J. 13,2267-2272. Andersson, H. & von Heijne, G. (1994b). Positively charged residues influence the degree of sec-dependence in protein translocation across the E. coli inner membrane. FEBS Lett. 347,169-172. Andersson, H., Bakker, E., & von Heijne, G. (1992). Different positively charged amino acids have similar effects on the topology of a polytopic transmembrane protein in Escherichia coli. J. Biol. Chem. 267, 1491-1495. Boyd, D. & Beckwith, J. (1990). The role of charged amino acids in the localization of secreted and membrane proteins. Cell 62,1031-1033. Claros, M.G. & von Heijne, G. (1994). TopPred II: an improved software for membrane protein structure prediction. CABIOS 10, 685-686. Dalbey, R.E. (1990). Positively charged residues are important determinants of membrane protein topology. Trends Biochem. Sci. 15,253-257. Deisenhofer, J., Epp, O., Miki, K., Huber, R., & Michel, H. (1985). Structure of the protein subunits in the photosynthetic reaction centre of Rhodopseudomonas viridis at 3A resolution. Nature 318, 618-624. Edelman, J. (1993). Quadratic minimization of predictors for protein secondary structure: application to transmembrane a-helices. J. Mol. Biol. 232,165-191. Edelman, J. & White, S.H. (1989). Linear optimization of predictors for secondary structure: application to transbilayer segments of membrane proteins. J. Mol. Biol. 210, 195-209. Engelman, D.M. & Steitz, T.A. (1981). The spontaneous insertion of proteins into and across membranes: the helical hairpin hypothesis. Cell 23,411-422. Gafvelin, G. & von Heijne, G. (1994). Topological "frustration" in multi-spanning E. coli inner membrane proteins. Cell 77,401-412. Gavel, Y. & von Heijne, G. (1992). The distribution of charged amino acids in mitochondrial inner membrane proteins suggests different modes of membrane integration for nuclearly and mitochondrially encoded proteins. Eur. J. Biochem. 205,12071215. Gavel, Y., Steppuhn, J., Herrmann, R., & von Heijne, G. (1991). The positive-inside rule applies to thylakoid membrane proteins. FEBS Lett. 282,41-46. Henderson, R., Baldwin, J.M., Ceska, T.A., Zemlin, P., Beckmann, E., & Downing, K.H. (1990). A model for the structure of bacteriorhodopsin based on high resolution electron cryo-microscopy. J. Mol. Biol. 213, 899-929.

Membrane Protein Topogenesis

213

Jones, D.T., Taylor, W.R., & Thornton, J.M. (1994). A model recognition approach to the prediction of all-helical membrane protein structure and topology. Biochemistry 33, 3038-3049. Kahn, T. & Engelman, D. (1992). Bacteriorhodopsin can be refolded from two independently stable transmembrane helices and the complementary five-helix fragment. Biochemistry 31, 6144-6151. Kahn, T.W., Sturtevant, J.M., & Engelman, D.M. (1992). Thermodynamic measurements of the contributions of helix-connecting loops and of retinal to the stability of bacteriorhodopsin. Biochemistry 31, 8829-8839. Kataoka, M., Kahn, T.W., Tsujiuchi, Y, Engelman, D.M., & Tokunaga, F. (1992). Bacteriorhodopsin reconstituted from two individual helices and the complementary 5-helix fragment is photoactive. Photochem. Photobiol. 56, 895-901. Klein, P., Kanehisa, M., & DeLisi, C. (1985). The detection and classification of membrane-spanning proteins. Biochim. Biophys. Acta 815,468^76. Kiihlbrandt, W., Wang, D.N., & Fujiyoshi, Y. (1994). Atomic model of plant light-harvesting complex by electron crystallography. Nature 367, 614-621. Lakey, J.H., Duche, D., Gonzalezmanas, J.M., Baty, D., & Pattus, F. (1993). Ruorescence energy transfer distance measurements: the hydrophobic helical hairpin of colicinA in the membrane bound state. J. Mol. Biol. 230,1055-1067. Lemmon, M.A., Flanagan, J.M., Treutlein, H.R., Zhang, J., & Engelman, D.M. (1992). Sequence specificity in the dimerizatibn of transmembrane a-helices. Biochemistry 31,12719-12725. Lemmon, M. A., Treutlein, H.R., Adams, RD., Briinger, A.T., & Engelman, D.M. (1994). A dimerization motif for transmembrane a-helices. Nature Struct. Biol. 1,157-163. Luirink, J. & Dobberstein, B. (1994). Mammalian and Escherichia coli signal recognition particles. Mol. Microbiol. 11,9-13. Nilsson, I.M. & von Heijne, G. (1990). Fine-tuning the topology of apolytopic membrane protein role of positively and negatively charged residues. Cell 62, 1135-1141. Nishiyama, K., Hanada, M., & Tokuda, H. (1994). Disruption of the gene encoding pl2 (SecG) reveals the direct involvement and important function of SecG in the protein translocation of Escherichia coli at low temperature. EMBO J. 13, 32723277. Pakula, A.A. & Simon, M.I. (1992). Determination of transmembrane protein structure by disulfide cross-linking: the Escherichia coli Tar receptor. Proc. Natl. Acad. Sci. USA 89, 4144-4148. Pogliano, J.A. & Beckwith, J. (1994). SecD and SecF facilitate protein export in Escherichia coli. EMBO J. 13, 554-561. Popot, J.-L. & de Vitry, C. (1990). On the microassembly of integral membrane proteins. Annu. Rev. Biophys. Biophys. Chem. 19, 369-403. Popot, J.-L., Gerchman, S.-E., & Engelman, D.M. (1987). Refolding of bacteriorhodopsin in lipid bilayers. A thermodynamically controlled two-stage process. J. Mol. Biol. 198,655-676. Popot, J.L. & Engelman, D.M. (1990). Membrane protein folding and oligomerization: the 2-stage model. Biochemistry 29,4031-4037. Saaf, A., Andersson, H., Gafvelin, G., & von Heijne, G. (1994). SecA-dependence of the translocation of a large periplasmic loop in the E. coli MalF inner membrane protein is a function of sequence context. Mol. Memb. Biol. 12, 209-215.

214

GUNNARVONHEIJNE

Sahin-Toth, M. & Kaback, H.R. (1993). Properties of interacting aspartic acid and lysine residues in the lactose permease of Escherichia coli. Biochemistry 32, 1002710035. Schatz, P.J. & Beckwith, J. (1990). Genetic analysis of protein export in Escherichia coli. Annu. Rev. Genet. 24, 215-248. Sipos, L. & von Heijne, G. (1993). Predicting the topology of eukaryotic membrane proteins. Eur. J. Biochem. 213, 1333-1340. Taylor, W.R., Jones, D.T., & Green, N.M. (1994). A method for a-helical integral membrane protein fold prediction. Proteins Struct. Funct. Genet. 18, 281-294. Treutlein, H.R., Lemmon, M.A., Engelman, D.M., & Briinger, A.T. (1992). The glycophorin A transmembrane domain dimer: sequence-specific propensity for a righthanded supercoil of helices. Biochemistry 31, 12726-12732. Tuffery, P., Etchebest, C., Popot, J.L., & Lavery, R. (1994). Prediction of the positioning of the seven transmembrane a-helices of bacteriorhodopsin: a molecular simulation study. J. Mol. Biol. 236, 1105-1122. von Heijne, G. (1986). The distribution of positively charged residues in bacterial inner membrane proteins correlates with the transmembrane topology. EMBO J. 5, 3021-3027. von Heijne, G. (1989). Control of topology and mode of assembly of a polytopic membrane protein by positively charged residues. Nature 341,456-458. von Heijne, G. (1992). Membrane protein structure prediction: hydrophobicity analysis and the positive-inside rule. J. Mol. Biol. 225,487-494. von Heijne, G. (1994). Membrane proteins: from sequence to structure. Annu. Rev. Biophys. Biomol. Struct. 23, 167-192. von Heijne, G. & Gavel, Y. (1988). Topogenic signals in integral membrane proteins. Eur. J. Biochem. 174, 671-678. Whitley, P., Nilsson, L., & von Heijne, G. (1993). 3-Dimensional model for the membrane domain oiEscherichia coli leader peptidase based on disulfide mapping. Biochemistry 32, 8534-8539. Whitley, P., Nilsson, I., & von Heijne, G. (1994a). De novo design of integral membrane proteins. Nature Struct. Biol. 1, 858-862. Whitley, P, Zander, T, Ehrmann, M., Haardt, M., Bremer, E., & von Heijne, G. (1994b). Sec-independent translocation of a l(X)-residues long periplasmic N-terminal tail in the E. coli inner membrane proteins ProW EMBO J. 13,4658-4661. Wickner, W, Driessen, A.J.M., & Hartl, F.U. (1991). The enzymology of protein translocation across the Escherichia coli plasma membrane. Annu. Rev. Biochem. 60, 101-124. Wolfe, P.B. & Wickner, W. (1984). Bacterial leader peptidase, a membrane protein without a leader peptide, uses the same export pathway as pre-secretory proteins. Cell 36, 1067-1072. Wolfe, P.B., Rice, M., & Wickner, W (1985). Effects of two sec genes on protein assembly into the plasma membrane of Escherichia coli. J. Biol. Chem. 260,1836-1841. Zen, K.H., Mckenna, E., Bibi, E., Hardy, D., & Kaback, H.R. (1994). Expression of lactose permease in contiguous fragments as a probe for membrane-spanning domains. Biochemistry 33, 8198-8206.

MODEL FOR INTEGRATING P-TYPE ATPases INTO ENDOPLASMIC RETICULUM

Randolph Addison and Jialing Lin

I. II. III. IV. V.

Introduction An Overview ofA^^Mrc?^/76>ra Plasma Membrane H"*"ATPase . . . Experimental Design Integration Model ofA^^Mra5/7ora Plasma Membrane H'^ATPase . . A General Integration Mechanism for P-Type ATPases and for Porters Acknowledgment References

215 216 218 226 231 232 233

r. INTRODUCTION The Neurospora plasma membrane proton-translocating ATPase (H*^ ATPase) transduces the chemical energy of ATP hydrolysis into a transmembrane proton-motive force (Scarborough, 1976). The latter in turn

Membrane Protein IVansport Volume 2, pages 215-235 Copyright © 1995 by JAI Press Inc. All rights of reproduction in any form reserved. ISBN: 1-55938-983-4

215

216

RANDOLPH ADDISON and JIALING LIN

functions to drive the uptake and extrusion of various ions, metabolites, and nutrients through chemiosmotic coupling devices called porters (Mitchell, 1967). The Neurospora W ATPase forms a kinetically competent phosphoryl-enzyme intermediate during the catalytic cycle (Dame and Scarborough, 1980). Thus, this enzyme shares properties with the NaVK"", W/K", and the Ca^"" ATPases of higher eukaryotic cells. The purposes of this article are to discuss how manipulations at the cDNA level have been used to gain insights into the topogenesis of the Neurospora W ATPase, to propose a plausible model for its integration into the membrane, and to propose how this model can be applied to the integration of other polytopic integral membrane proteins.

II. AN OVERVIEW OF NEUROSPORA PLASMA MEMBRANE H^ ATPase The physiological role of the plasma membrane H"*^ ATPase from Neurospora and yeast was reviewed previously (Serrano, 1989). C. L. Slay man and associates (1970) provided the first experimental evidence for an electrogenic ATPase in the plasma membrane of Neurospora, They postulated that this electrogenic ATPase is a proton pump. It became possible to demonstrate this directly when Neurospora plasma membrane was isolated in high yield and purity by using a concanavalin A method (Scarborough, 1978). Experimental results with everted vesicles of the plasma membrane containing fluorescein isothiocyanate-labeled dextran provided direct proof that the electrogenic ATPase is a proton pump (Scarborough, 1980). Subsequently, the plasma membrane H^ ATPase was solubilized by detergent and purified to homogeneity (Addison and Scarborough, 1981; Bowman et al., 1981). The H"" ATPase has a molecular weight of approximately 105,000 when resolved by polyacrylamide gel electrophoresis in the presence of sodium dodecyl sulfate. Biochemical characterization of the H"^ ATPase showed that the catalytic sequence involves a phosphoryl-enzyme intermediate (Dame and Scarborough, 1980), that vanadate is a potent inhibitor of ATP hydrolysis by the ATPase (Addison and Scarborough, 1981), and that the enzyme undergoes significant conformational changes during its catalytic cycle (Addison and Scarborough, 1982). Reconstituted proteoliposomes containing only the hydrolytic moiety of the H"^ ATPase provided evidence that the 105-kDa moiety coupled ATP hydrolysis to the generation of a proton-motive force (Scarborough and Addison, 1984).

Topogenesis of Neurospora ht ATPase

217

Biochemical experiments dearly established that the Neurospora and yeast plasma membrane H"^ ATPase have features in common with the NaVK^ and Ca^"^ ATPases, implying that cation-motive ATPases function via a similar mechanism (Dame and Scarborough, 1981; Mitchell, 1981). During recent years, the genomic DNA and the cDNA of the cation-motive ATPases were cloned and sequenced (Hesse et al., 1984; MacLennan et al., 1985; ShuU et al., 1985; Addison, 1986). A comparison of the deduced amino acid sequences of the ATPases revealed highly conserved regions (Addison, 1986). This observation combined with the biochemical data strongly supported the earlier conjectures that cation-motive ATPases operate through a common mechanism. This implies that all ion-motive ATPases evolved from a common ancestral cation-motive pump. To understand the molecular mechanism of these remarkable cationmotive ATPases, membrane topological models were proposed that were based mainly on hydropathy plots of the ATPases' primary sequence. These models predict that these proteins contain either 8- or 10-transmembrane segments with approximately 80% of the polypeptide chain on the cytoplasmic side of the membrane and less than 5% localized to the exoplasmic side of the membrane. By using different techniques, the putative energy-transduction domain, the ATP-binding and hydrolytic domains, and the amino and carboxyl termini have been localized to the cytoplasmic side of the membrane (Green, 1989; Clarke et al., 1990; Scarborough and Hennessy, 1990). The latter observation implies that these enzymes have an even number of transmembrane segments. Site-directed mutagenesis of the cDNAof the cation-motive ATPases is being used to generate mutants that are in turn used to gain insights into the structure-function relationships of these energy transducers (MacLennan, 1990). The membrane topological models as predicted by the hydropathy plots are used to select specific residues to mutate to determine their role as specific ligand-binding residues and their role in the energy transduction mechanism in general. These plots are, however, ambiguous for regions that span the membrane (Fasman and Gilbert, 1990; Jahnig, 1990). For instance, three topological models of the Neurospora plasma membrane H"*^ ATPase exist that have either 8,10, or 12 segments (Addison, 1986; Hager et al, 1986; Rao et al., 1991). Clearly, x-ray diffraction analysis of crystals of these cation-motive ATPases will resolve these problems and will provide reliable models of these enzymes. However, a major impedance to achieving this goal is the inability to obtain crystals of cation-motive ATPases that are suitable for

218

RANDOLPH ADDISON and JIALING LIN

x-ray structural analysis. Nevertheless, reliable models are needed to assist in designing and interpreting mutagenesis studies. To achieve this, an experimental technique is required that can verify whether a predicted transmembrane segment is within the lipid bilayer.

Hi. EXPERIMENTAL DESIGN An approach to verifying whether a stretch of residues is within the lipid bilayer is to study the topogenesis of polytopic integral membrane proteins. The molecular mechanism by which proteins obtain their characteristic arrangement across the osmotic barrier is poorly understood, however. Two fundamentally different models of protein integration have been proposed. One model proposes that a stretch of hydrophobic residues inserts directly into the hydrophobic interior of the lipid bilayer (Engelman and Steitz, 1981). Experimental observations have demonstrated that a stretch of hydrophobic residues is capable of partitioning spontaneously into liposomes or detergent micelles. This model, however, cannot explain the membrane selectivity of proteins— why are proteins inserted only into certain membranes and not others? Nor can this model explain the role of integral membrane proteins of the endoplasmic reticulum that are apparently essential for the integration event. The other model proposes that protein integrates into membranes by the decoding of topogenic sequences within proteins by proteinaceous effectors (Blobel, 1980). The characteristic asymmetrical arrangement of proteins in the membrane is achieved by the decoding of topogenic signals that are able to initiate and stop translocation across the membrane. This model, referred to as the sequential insertion model, was originally proposed by Blobel. It has been modified by experiments that unveiled proteins which are involved in the integration event, as well as by insights provided from in vitro studies of the integration of proteins into membranes. Blobel classified integral membrane proteins on the basis of the number of times the polypeptide chain spanned the membrane: (a) monotopic integral membrane proteins have their hydrophilic residues exposed on only one side of the membrane; (b) bitopic integral membrane proteins span the membrane once and have their hydrophilic residues exposed to both sides of the membrane; (c) polytopic integral membrane proteins span the membrane multiple times. In the sequential insertion model, a topogenic signal (termed a "starttransfer signal") would initiate translocation of carboxyl terminal resi-

Topogenesis of Neurospora hf ATPase

219

dues and a stop-transfer signal (another topogenic signal) would halt the translocation process. The topogenic signals are subsequently displaced (unspecified) from the putative integration machinery into the lipid bilayer. This model is attractive because it allows for the manipulation of the cDNA of a poly topic integral membrane protein by cut-and-paste types of experiments to assay for the presence of topogenic sequences as well as designing on the cDNA level a fusion protein to test the ability of a stretch of hydrophobic residues to function as a topogenic signal. That is, it is possible to engineer into the cDNA of a particular protein a DNA fragment coding a presumed topogenic signal. RNA transcripts generated frorn this template are then used to program an in vitro translation system supplemented with translocation competent microsomes. Integration or translocation of the resultant proteins into microsomes is readily determined by fractionation methods or by monitoring processing of the product, that is, N-linked glycosylation. Therefore, to study the topogenesis of the Neurospora H^ ATPase, RNA transcripts of the H^ ATPase were translated in a Neurospora in vitro system supplemented with microsomes isolated from Neurospora (Addison, 1987,1990). The H^ ATPase was integrated into microsomes. This was demonstrated by cosedimentation of the in vitro synthesized fj+ ATPase with microsomes after extracting the translation mixture with 0.1 M Na2C03 (pH .11.5). Although the H"" ATPase was integrated into microsomes, the results provided no insights into the integration process and provided no insights into the role, if any, of topogenic signals. To localize putative topogenic signals, the cDNA of the H"^ ATPase was restricted by restriction endonucleases within codons to generate RNA transcripts coding truncated H"^ ATPase of different lengths (Addison, 1991, 1993). These transcripts were translated in a Neurospora in vitro system. The products integrated into microsomes in a process that required nucleotides and microsomal proteins. Although these early results early established the presence of topogenic signals in the H^ ATPase, they provided no information about the exact position of or the number of signals involved in the integration of the truncated proteins into the microsomes. It proved difficult to remove putative transmembrane segments and test them individually or in tandem because of the absence of unique restriction endonuclease sites at suitable positions within the cDNA of the H"^ ATPase. To overcome this problem, we used the polymerase chain reaction (PCR) to construct on the cDNA level fusion proteins of H^ ATPase (Lin and Addison, 1994a).

RANDOLPH ADDISON and JIALING LIN

220 H+ATPase cDNA

T3 promoter

T3RNA polymerase

RNA transcripts I Neurospora in-vitro translation i system Fusion Proteins

Figure 1. Constructing the fusion proteins. cDNA fragments encoding the hydrophilic residues of the amino (rectangular box with diagonal lines) or carboxyl (rectangular box with dots) terminus and the putative transmembrane segments (solid rectangular boxes) were obtained using PCR and the Neurospora H"^ ATPase cDNA as a template. cDNA fragments encoding the invertase segment with consensus N-linked glycosylation sites (rectangular box with two dash lines) were obtained using PCR and yeast invertase cDNA as a template. In the cDNA of invertase, the gray box represents the region encoding the signal sequence. cDNA fragments were ligated into pBluescript II KS+ plasmid. The recombinant plasmid digested with PvuW was used as a template to generate RNA transcripts using T3 RNA polymerase. The transcripts were translated in a Neurospora in vitro system. Other details are in the text.

Figure 1 depicts the construction of fusion proteins of the Neurospora plasma membrane H^ ATPase. As diagrammed, DNA fragments were generated from the cDNA of the H"^ ATPase by using oligonucleotide primers that were complementary to the first 15 bases at the ends of the region that is to be amplified by PCR. The cDNA fragment coding

Topogenesis of Neurospora hf^ ATPase

221

hydrophilic residues 1-115 of the H^ ATPase and 170 bp of the 5' untranslated sequence was obtained using an oligonucleotide sense primer with a recognition sequence for HindlU and an antisense primer with recognition sequences for Pstl/BamHl. The DNA fragment coding hydrophilic residues 879-920 of the H"^ ATPase, a stop codon, and 15 bp of the 3' untranslated sequence was obtained using an oligonucleotide sense primer with a recognition sequence for Xbal and an antisense primer with a recognition sequence for Sacll. cDNA fragments coding the putative transmembrane segments were obtained using oligonucleotide sense primers with recognition sequences for Pstl/BamHl and antisense primers that had recognition sequences for EcoRVBamHl, cDNA fragment coding residues 73-132 of the secretory form of invertase was obtained using an antisense primer with a recognition sequence for Xbal and a sense primer that had recognition sequences for EcoRl/BamHL This invertase segment contains three consensus Nlinked glycosylation sites that will serve as a reporter of translocation into the microsomes. To ensure complete digestion of the cDNA fragments by the restriction endonucleases, three base pairs were added to the 5' end of each oligonucleotide primer. The addition of two recognition sequences in some primers gave us flexibility in generating certain combinations of putative transmembrane segments. For the PCR, singlestranded DNA of the cDNA was generated from the recombinant phagemid pBluescript II KS4- (Stratagene). PCR products were generated using the Perkin Elmer's GeneAmp kit and the Thermal cycler 880 following standard protocols. PrimerErase Quick push columns (Stratagene) were used to remove excess primers. The products were digested by restriction endonucleases, resolved on an agarose gel, and excised and eluted from the gel using Glassmilk or Glassfog (BIO 101, La JoUa, CA). cDNA fragments were ligated into pBluescript II KS+. DNA sequence of the constructs was confirmed by DNA sequencing using chain-terminating dideoxyl nucleotides. As mentioned, topological models of the Neurospora plasma membrane H"^ ATPase are based almost exclusively on hydropathic plots of the primary sequence that predict either 8- or 10-transmembrane segments (Addison, 1986; Hager et al., 1986). This method identifies stretches of hydrophobic residues that are long enough to span the membrane as a-helices. Another model of the H^ ATPase based on the exhaustive trypsin digestion of the reconstituted H^ ATPase proposes that the enzyme has 12-transmembrane segments (Rao et al., 1991). The 8and 10-segment models agree that there are four transmembrane seg-

RANDOLPH ADDISON and JIALING LIN

222

Table 1. Amino Acid Sequences of the Neurospora H^ ATPase Used in Constructing the Fusion Proteins N-terminus sequence: MADHSASGAPALSTNIESGKFDEKAAEAAAYQPKPKVEDDEDEDIDALIEDL ESHDGHDAEEEEEEATPGGGRVVPEDMLQTDTRVGLTSEEVVQRRRKYGLN 1-115 QMKEEKENHFLKPA AM 116-138 M1

GSFLGFFVGPIQFVMEG A AVLA AGLEF

M2

EWVDFGVICGLLLLNAVVGFV

M3

GSGIGTILLILVIFTLLIVWVSSFYEF * * ** AILEFTLAITIIGVPVGLPAVVTTTMAVGAAYLA *

141-160 292-314

M4

322-354 688-713

M5

AMYAYVVYRIALSIHLEIFLGLWIAIL

M6

ASLNIELVVFIAIFADVATLAIAYDNA

M7

ALWGMSVLLGVVLAVGTWITVTTMYA *

807-826

M8

AWLIFITRANGPFWSSIPSWQ

827-847

M9

ALSGAIFLVDILATCFTIWGWF

852-878

716-741 755-779

M10 ATSIVAV VRIWIFSFGIFCIMGG V Y YIL Invertase sequence: GSFWGHATSDDLTNWEDQPIAIAPKRNDSGAFSGSMVVEYNNTSGFF ** NDTIDPRQRCVAIWTSR

73-132

C-terminus sequence: QDSVGFDNLMHGKSPKGNQKQRSLEDFVVSLQRVSTQHEKSQ

879-920

Notes: The hydrophilic residues of the amino terminus of the H"^ ATPase are followed by the sequence of a putative transmembrane segment (M), which in turn is followed by the invertase segment and the hydrophilic residues of the carboxyl terminus of the enzyme. Amino acids marked with an asterisk were introduced during construction of the fusion proteins. The numbers of inclusive residues are given at the end of each sequence.

Topogenesis of Neurospora ht ATPase

223

ments in the amino terminal third, but disagree about the number of segments in the carboxyl terminal third of the protein. The 12-segment model has six transmembrane segments in the amino terminal third and six in the carboxyl third of the enzyme. With the protein fusion method it now became possible to verify the number of transmembrane segments in H"^ ATPase. This method should also provide means for accurately determining the nature of the topogenic signals required for integrating H"^ ATPase into the membrane. The amino acid sequences used in the fusion proteins are listed in Table 1. Constructs with a putative transmembrane segment were designated pM followed by a number that indicated the position of the segment within a model. As mentioned, in the sequential insertion model, oddnumbered transmembrane segments initiate translocation of the carboxyl terminal residues across the membrane, and even-numbered segments stop the transfer process. Therefore, if the transmembrane segments are correctly predicted, an odd-numbered segment engineered into the fusion protein should initiate transfer of the invertase segment into the lumen of the microsomes, and when this segment is combined with an evennumbered segment, glycosylation of the invertase segment should be blocked (see Figure 2). Using transmembrane segments as predicted in the 10-segment model (Table 1), we generated constructs containing M3, M5, or M7. The resultant proteins were glycosylated. These transmembrane segments functioned as predicted by the sequential insertion model. Experimental results from digestion of the membrane-integrated products by proteases, from fractionation of the translation mixtures on step-sucrose gradient under alkaline conditions, and from immunoprecipitation of the digested membrane-integrated products using polyclonal antibodies directed against the amino- or carboxyl-terminal residues showed that the proteins were integrated into microsomes with the amino terminus on the cis side and the carboxyl terminus on the trans side of the membrane (Lin and Addison, 1995a,b). In contrast, constructs containing Ml or M9 were not glycosylated. These constructs were also not associated with microsomes after extracting the translation mixtures at pH 11.5. Constructs containing M2, M4, M6, or M8 did not associate with microsomes. These segments are predicted to function as stop-transfer signals. Only MIO, a predicted stop-transfer segment, initiated translocation of the invertase segment into microsomes. Stop-transfer signals can sometime function as start-transfer signals. Therefore, it was not surprising that MIO could do so in this in vitro.system.

RANDOLPH ADDISON and JIALING LIN

224

I

II

Figure2. Membranetopology of fusion proteins. Fusion proteins with one transmembrane segment (so//drecfangu/ar box) have the membranetopology of a bitopic integral membrane protein (I) with the carboxyl terminus and the invertase segment localized in the lumen of the microsomes. In the lumen, the invertase segment (rectangular box with diagonal lines) will be glycosylated. The arrows represent oligosaccharide side chains. Fusion proteins with two transmembrane segments will have a membrane topology of a polytopic integral membrane protein (II). In this membrane topology, the invertase segment remains on the cytosolic side of the membrane.

We noticed, however, that any putative transmembrane segments containing a negatively charged residue did not initiate transfer of the invertase segment into microsomes. These segments include Ml, M2, M4, M6, and M9 (see Table 1). M8 and MIO each contain a positively charged residue. MIO initiated transfer of the carboxyl terminal residues into microsomes; M8 did not. The latter also contains two prolines. It is possible that this combination of residues impaired the ability of this segment to function individually in the construct. One transmembrane segment, M5, functioned individually in the construct although it contains a negatively and a positively charged residue. These charged residues are seven residues apart. It is possible that these residues are able to form a salt bridge that would mask the ionic charges and create a functional topogenic signal. Charged residues in the transmembrane segments are presumed to function as cation-binding sites (Clarke et al., 1989). If so, these residues serve an essential function for cation-motive

Topogenesis of Neurospora tt ATPase

225

ATPases. However, the presence of these charged residues in the segments impaired their ability to integrate into the membrane. Insight into this apparent paradox was provided when we engineered additional constructs containing two or more transmembrane segments. For instance, neither Ml nor M2 functioned in the constructs. In contrast, construct pMlM2 integrated into microsomes as a polytopic integral membrane protein. This implies that combining these segments created a topogenic signal for integration. The segments apparently acted concertedly in the integration event. Identical results were observed for the combination of M3, M4; M5, M6; M7, M8; and M9, MIO. In the aqueous milieu, these closely juxtaposed transmembrane segments must associate. This association is favored by the short stretch of residues between the segments and by hydrophobic force that gives a hydrophobic interaction energy of about 20 kcal/mol of paired segments (Engelman and Steitz, 1981). In the case of pMlM2, when the number of residues between the segments was gradually lengthened, the integration effectiveness of the resultant products decreased (Lin and Addison, 1995a). This implies that the proximity of the segments is crucial in forming the topogenic signal. In contrast, when residues between the segments in pM3M4 and in pM7M8 were lengthened, the intervening residues were transferred to the lumen of the microsomes and the termini of the proteins were localized to the cis (cytosolic) side of the membrane. Although some of the transmembrane segments of the H^ ATPase behaved as predicted by the sequential insertion model, others did not. Are these other segments integrated into the membrane by a novel mechanism? If so, this implies that integrating the H"^ ATPase into the membrane requires multiple integration machineries. In pM3M4 and pM7M8 the transmembrane segments functioned as predicted by the sequential insertion model; these segments are separated, however, by a short stretch of hydrophilic residues in the native H"^ ATPase and in the constructs. In this closely juxtaposed position, will the segments act concertedly or individually in integrating the H"^ ATPase into the membrane? When we replaced M2 in the construct pMlM2 by M4, the resultant protein integrated into microsomes as a polytopic integral membrane protein. This implies that the putative integration machinery recognizes some general structural feature in the paired segments. In pMlM2, three hydrophilic residues separate the segments; in pMlM4 and in pM3M4, eight residues separate the segments. Therefore, it is likely that M3 and M4 form a topogenic signal similar to that formed by M1 and M2. If this

226

RANDOLPH ADDISON and JIALING LIN

is true, this implies that integrating the H"^ ATPase into microsomes uses only one type of topogenic signal for integration and only a single integration machinery. This topogenic signal consists of two transmembrane segments. Since this topogenic signal is required for integrating the H"*^ ATPase into microsomes, we will refer to it as an integration signal. Also, our experimental results localized four transmembrane segments in the amino-terminal third and six in the carboxy-terminal third of H^ ATPase. These data support a 10-segment model for H"*^ ATPase.

IV. INTEGRATION MODEL OF NEUROSPORA PLASMA MEMBRANE H^ ATPase Before continuing to a plausible integration model for H^ ATPase, it might be useful to make a few remarks about two main types of molecular recognition mechanisms in living organisms. One type is described for enzymatic reactions: enzymes bind stereospecific substrates and catalyze the rearrangement of chemical groups or atoms. The other type is described for receptor binding: a receptor binds stereospecific molecules or stereospecific areas of biological polymers. Information is transmitted in the form of conformational changes to other region(s) of the receptor, or the resultant signal is transduced into another signal through the interactions of the receptor-ligand complex with other biological molecules. Chemical equilibrium in the cell is expressed using the Gibbs free energy AG' at constant pressure and temperature. Generally, the difference between AG° measured from standard states of reactants and products and AG' measured from the conditions under which the reaction is conducted is termed the thermodynamic drive, or driving force (AG*), of a particular reaction. This is expressed by the simple relationship: AG* = jdG The change in G* is the integral or sum over all increments of dG in different states of the reaction. This is best explained by giving as an example the transfer of a hydrophobic polymer from an aqueous milieu (phase 1) to a hydrophobic milieu (phase 2). The product of the moles of polymer in phase 1 times the difference in the chemical potential of the polymer in the standard states and in the states from which the reaction is conducted will be termed dG^^^ ^ The same type of calcula-

Topogenesis of Neurospora ht ATPase

227

tion for the polymer in phase 2 will give dG^^^^ 2- Therefore, the difference in G* would be the difference between Gphase2 ^^^ Gp^ase i- This AG* represents a measure of the hydrophobicity of the polymer and measures the system's ability to adopt a thermodynamically stable structure in phase 2 by minimizing interactions with water in phase 1. Integration of transmembrane segments (TS) into the membrane from the aqueous phase is viewed as a receptor-mediated process: TS

^^

^^

• TS

RTS

In this scheme, the receptor (R) is situated between the aqueous compartment (on the left), that is, the cytosol or the aqueous channel of the putative integration machinery, and the lipid bilayer (on the right), the hydrophobic phase. For the reaction to go from left to right, the overall sign of AG* must be negative. The difference between TS on the left and TS on the right of the diagram need involve only changes in the configuration or the conformation of secondary and ionic or noncovalent bonds such as those participating in the ionization or hydration of TS (these reactions are not indicated above). These changes may require an input of energy. This will make an intermediate dG positive. As stated, the overall sign of AG* must be negative for the reaction to go from left to right. This can be accomplished if the dG from the configuration changes is much smaller than the overall AG* from transferring a pair of a helices from one phase to another. Alternatively, the reaction could be coupled to an exergonic reaction. Our objective in introducing the foregoing arguments is to emphasize that integrating proteins into the lipid bilayer is similar to other basic biological recognition processes in living organisms. To integrate proteins into membranes, an effector of this process must function at the interface of and within two distinct compartments. Although integration has not received as much interest and investigation as the translocation process, we feel that an understanding of the details of the integration mechanism is just as interesting as that of translocation and that integration can be analyzed also by dissecting the integration process into

228

RANDOLPH ADDISON and JIAUNG LIN

separate catalytic and structural units. Studying components of the integration process may provide insights into novel protein-protein interactions that may be interesting in a much broader context. Based on the experimental results reviewed briefly in the above sections, a plausible model of how the Neurospora H^ ATPase is integrated into membrane can be formulated. The key features of this model are shown in Figure 3. The nature of the topogenic signal in this model is based on experimental evidence. But the binding of this signal to the integration machinery and its subsequent spatial displacement to the lipid bilayer are yet unsupported by any data. Therefore, this part of the model must remain conjectural. Since trypsin treatment of microsomes blocked integration (Lin and Addison, 1995a), we postulate that integration of H^ ATPase into microsomes is a receptor-mediated event. Integration begins with the binding of a pair of transmembrane segments to a binding cleft positioned at the interface of subunits of a multisubunit complex localized in the membrane. Reasons for placing the binding site here will be given below. This multisubunit complex could be identical to that used for translocating proteins across the membrane. When the pair of segments is integrated into the membrane, there is no translocation. Therefore, a ribosome engaged in synthesizing H^ ATPase will not form a tight ribosome-membrane junction. The osmotic barrier across the membrane is maintained because binding of the topogenic signal to the integration machinery will not gate the opening of a channel in the membrane. After binding to the receptor in the membrane, the next major consideration is the spatial displacement of the bound topogenic signal from the machinery to the lipid bilayer. The bound ligand is a pair of transmembrane segments that apparently remain unmodified upon binding to the putative receptor. Therefore, why would the receptor release a substrate for which it has a high affinity? As mentioned, enzymes (E) have a high affinity for their substrates (S). There is a finite amount of substrate released after binding. The overall reaction is forced through the ES complex to the formation of the transition state (Segel, 1975). The substrate is eventually modified. The resultant product(s) is released from the enzyme because the product(s) has a lower binding affinity with the enzyme than the substrate. So why would a pair of bound and apparently unmodified transmembrane segments be released from the receptor? Insight into one possibility of how the segments are released is provided by experimental evidence. Integration into microsomes requires ATP and GTP (Lin and Addison, 1995a). The latter may be required for a SRP-SRP receptor-

Topogenesis of Neurospora Hf^ ATPase

229

Cytosol Membrane Lumen Figure 3. Integration nnodel of P-Type ATPases. Pair of transmembrane segments (solid rectangular boxes) is decoded by a soluble effector (R1), which in turn targets the resultant complex at the membrane, where R1 interacts with its cognate receptor (R2). The integration signal is transferred to the integration machinery (X/Y). Other proteins, represented by a question mark, may be required for this step. Alternatively, the integration signal interacts directly with the machinery, bypassing R1. The binding site for the integration signal in the machinery is depicted at the interface of subunits X and Y. The binding site could be composed of identical subunits of the complex. This possibility is not depicted in the above diagram. The pair of segments is channeled to the membrane (two horizontal lines).

mediated targeting event, analogous to that found in canine pancreas microsomes (Figure 3). But what is the ATP used for? One possibility is that binding and/or hydrolysis of ATP could be used to drive the channeling of the ligand from the receptor to the lipid bilayer. Alternatively, another, more attractive possibility is that in the aqueous milieu the pair of transmembrane segments has a secondary or tertiary structure that is different form that in a hydrophobic milieu. This conjecture is supported

230

RANDOLPH ADDISON and JIALING LIN

by experimental evidence demonstrating a milieu-specific configuration of hydrophobic peptides (Li and Deber, 1993). If this is true, we postulate that the pair of segments that binds to the putative receptor either directly from the cytosolic compartment or through a SRP-SRP receptor-mediated targeting pathway has a specific conformation that is subsequently altered at the receptor-binding site. This could occur in an ATP-dependent fashion, analogous to the action of heat shock proteins, or the putative receptor itself could induce specific alteration in the conformation of the segments, analogous to the binding of antigens to antibodies. In either case, the net effect of these conformational maturations is to alter the affinity of the substrate, a pair of transmembrane segments, for the receptor. This will drive the ligand from the receptor down a hydrophobic gradient to the lipid bilayer. Since we have posited that the integration machinery will have an aqueous channel, the receptor will be positioned between two phases: the aqueous phase of the channel and the hydrophobic phase of the lipid bilayer. This will create a hydrophobic potential difference across the receptor, and this will serve as the driving force for channeling the pair of segments from the receptor to the lipid bilayer. This integration mechanism, however, will explain the integration of four of the possible five pairs of transmembrane segments of H^ ATPase into the membrane. The exception to this model is the pair M7, M8. These segments are separated by 27 residues. In the 10-segment model, these residues are localized on the exoplasmic side of the membrane. Translocation of these across the membrane can be accommodated within the framework of the above model, however. We have postulated that the substrate for the receptor is a pair of segments. Therefore, the binding of a single segment to the receptor will not elicit reactions as outlined above. M7 will bind and remain associated with the receptor. Consequently, the ribosomes will become tightly associated with the putative machinery and eventually will form a tight membrane junction. Polypeptide synthesis will continue. The growing chain will evoke the opening of the gated channel and will be transferred to the lumenal side of the membrane. Upon entering the aqueous compartment of the channel, M8 will bind to the receptor next to M7, concomitantly evoking specific conformation changes in the resultant pair of segments and the channeling of these to the lipid bilayer. As mentioned, in the sequential insertion model (Blobel, 1980), the odd-numbered transmembrane segments initiate translocation of the carboxy-terminal residues into the lumen of the endoplasmic reticulum

Topogenesis of Neurospora H^ ATPase

2 31

and the even-numbered segments stop the translocation event. A series of these events resuhs in the asymmetrical arrangement of the protein in the membrane. This model is a direct function of the decoding of individual transmembrane segments. The integration model outlined herein differs from the sequential insertion model in three major ways: (a) the topogenic signal for integration is a pair of transmembrane segments; (b) the putative integration machinery does not disassemble in order to release the segments into the lipid bilayer; and (c) a putative start-transfer signal does not gate the channel, but instead simply holds the ribosomes in proximity to the machinery. Subsequent events will lead to the formation of a tight ribosome membrane junction. The growing polypeptide chain opens the channel to the lumenal side of the membrane. This conjecture is supported by the fact that proteins without a signal sequence are translocated across the endoplasmic reticulum (Wiedmann et al., 1994) and by the fact that the segment of nascent secretory proteins localized in the aqueous compartment of the gated channel is not accessible from the lumenal compartment to probes (Crowley et al., 1994). Taken together, this suggests that the channel is not gated by the signal sequence but by a segment of the nascent polypeptide chain.

V. A GENERAL INTEGRATION MECHANISM FOR P-TYPE ATPases AND FOR PORTERS Since the Neurospora plasma membrane H"^ ATPase, as mentioned in the introduction, is a member of a family of cation-motive ATPases, it is reasonable to assume that all P-type ATPases will have a common integration mechanism. That is, the topogenesis of these energy transducers is achieved by a receptor-mediated, energy-dependent integration of pairs of transmembrane segments into the membrane. Can the results obtain from studying Neurospora H^ ATPase be applied to the integration of other polytopic membrane proteins? Porters, that is, uniporters, antiporters, and symporters, are polytopic integral membrane proteins involved in the import or export of sugars, metabolites, ions, and peptides across membranes. They have an even number of transmembrane segments, with most of these containing charged residues that function as specific ligand-binding sites. They have few residues on the exoplasmic side of the membrane and few residues between the transmembrane segments. Accordingly, porters share fundamental features with P-type ATPases. It is reasonable to assume that pairs of transmem-

232

RANDOLPH ADDISON and JIALING LIN

brane segments play a fundamental role in integrating porters into the membrane. If this is true, it is possible that all polytopic integral membrane proteins have a fundamentally similar integration mechanism that varies only in the length of the polypeptide chain translocated across the membrane and in the number of segments integrated into the membrane. The concepts that we have tried to convey in our description of how a P-type ATPase is integrated into the membrane are the key role of a pair of transmembrane segments in the integration event, the role of the receptor in mediating the conformational maturation of the bound transmembrane segments, and the channeling of the modified segments down a hydrophobic potential gradient. Integration into the membrane occurs as the natural consequent of the intrinsic asymmetrical arrangement of the receptor in the membrane that leads to a vectorial nature in the way the substrate, a pair of transmembrane segments, binds to and is released from the binding site. We emphasized, also, that effectors of the integration process are thermodynamically stable structures and are not assembled from various components to form the integration machinery, nor are they disassembled to release the segments into the membrane. Whether this model is ultimately proved correct, it does provide a reasonable explanation of how transmembrane segments containing charged and/or proline residues integrate into the lipid bilayer. In contrast, the sequential insertion model cannot provide a reasonable explanation of the cooperative interaction of transmembrane segments in the integration event or of how segments containing charged residues can integrate into the membrane. It has been hoped by some that a structural description of polytopic integral membrane proteins would provide valuable insight into the nature of the topogenesis signals that are essential for their integration into the lipid bilayer. So far this type of data has provided no insight into the nature of the signals used to integrate these proteins into the membrane. Instead, the information that is currently available suggests specific interactions among the transmembrane segments and has shown that only a small percentage of the surface of the segments is in direct contact with the lipid bilayer (Rees et al., 1989). It is hoped that attempts to confirm or deny the principles as outlined herein will lead to progress in understanding the molecular mechanism of integration.

ACKNOWLEDGMENT Supported in part by National Science Foundation grant DCB-9108460.

Topogenesis of Neurospora H^ ATPase

233

REFERENCES Addison, R. (1986). Primary structure of the Neurospora plasma membrane H"^ ATPase deduced from the gene sequence. J. Biol. Chem. 261,14896-14901. Addison, R. (1987). Secretory protein translocation in a Neurospora crassa in vitro system: hydrolysis of a nucleoside triphosphate is required for posttranslational translocation. J. Biol. Chem. 262, 17031-17037. Addison, R. (1990). Studies on the sedimentation behavior of the Neurospora crassa plasma membrane H"^ ATPase synthesized in vitro and integrated into homologous microsomal membranes. Biochim. Biophys. Acta 1030,127-133. Addison, R. (1991). GTP is required for the integration of a fragment of the Neurospora crassa H"*" ATPase into homologous microsomal vesicles. Biochim. Biophys. Acta 1065, 130-134. Addison, R. (1993). The initial association of a truncated form of the Neurospora plasma membrane H"*" ATPase and of the precursor of yeast invertase with microsomes are distinct processes. Biochim. Biophys. Acta 1152,119-127. Addison, R. & Scarborough, G.A. (1981). Solubilization and purification of the Neurospora plasma membrane H"*" ATPase. J. Biol. Chem. 256,13165-13171. Addison, R. & Scarborough, G.A. (1982). Conformational changes of the Neurospora plasma membrane H"*" ATPase during its catalytic cycle. J. Biol. Chem. 257, 10421-10426. Blobel, G. (1980). Intracellular protein topogenesis. Proc. Natl. Acad. Sci. USA 77, 1496-1500. Bowman, B.J., Blasco, F., & Slayman, C.W. (1981). Purification and characterization of the plasma membrane ATPase of Neurospora crassa. J. Biol. Chem. 256,1234312349. Clarke, D.M., Loo, T.W., Inesi, G., & Maclennan, D.H. (1989). Location of high affinity Ca "^-binding sites within the predicted transmembrane domain of the sarcoplasmic reticulum Ca^'^-ATPase. Nature 339,476^78. Clarke, D.M., Loo, T.W., & Maclennan, D.H. (1990). The epitope for monoclonal antibody A20 (amino acids 870-890) is located on the lumenal surface of the Ca^"*" ATPase of sarcoplasmic reticulum. J. Biol. Chem. 265,17405-17408. Crowley, K.S., Liao, S., Worrell, V.E., Reinhart, G.D., & Johnson, A.E. (1994). Secretory proteins move through the ER membrane via an aqueous, gated pore. Cell 78, 461-471. Dame, J.B. & Scarborough, G.A. (1980). Identification of the hydrolytic moiety of the Neurospora plasma membrane H"*" ATPase and demonstration of a phosphoryl-enzyme intermediate in its catalytic mechanism. Biochemistry 19, 2931-2937. Dame, J.B. & Scarborough, G.A. (1981). Identification of the phosphorylated intermediate of the Neurospora plasma membrane H"^ ATPase as P-aspartyl phosphate. J. Biol. Chem. 256,10724-10730. Engelman, D.M. & Steitz, T.A. (1981). The spontaneous insertion of proteins into and across membranes: the helical hairpin hypothesis. Cell 23,411-422. Fasman, G.D. & Gilbert, W.A. (1990). The prediction of transmembrane protein sequences and their conformation: an evaluation. Trends Biochem. Sci. 15, 89-92. Green, N.M. (1989). ATP-driven cations pumps: alignment of sequences. Biochem. Soc. Trans. 17,972-974.

234

RANDOLPH ADDISON and JIALING LIN

Hager, K.M., Mandala, S.M., Davenport, J.W., Speicher, P.W., Benz, EJ., & Slayman, C.W. (1986). Amino acid sequence of the plasma membrane ATPase of Neurospora crassa: deduction from genomic and cDNA sequences. Proc. Natl. Acad. Sci. USA 83, 7693-7697. Hesse, J.E., Wieczorek, L., Altendorf, K., Reicin, A.S., Dorus, E., & Epstein, W. (1984). Sequence homology between two membrane transport ATPases, the Kdp-ATPase of Escherichia coli and the Ca ATPase of sarcoplasmic reticulum. Proc. Natl. Acad. Sci. USA 81,4746-4750. Jahnig, F. (1990). Structure predictions of membrane proteins are not that bad. Trends Biochem. ScL 15,93-95. Li, S-C. & Deber, CM. (1993). Peptide environment specifies conformation: helicity of hydrophobic segments compared in aqueous, organic, and membrane environments. J. BioL Chem. 268, 22975-22978. Lin, J. & Addison, R. (1994a). Topology of the Neurospora plasma membrane H"*" ATPase: localization of a transmembrane segment. J. Biol. Chem. 269,3887-3890. Lin, J. & Addison, R. (1995a). A novel integration signal that is composed of two transmembrane segments is required to integrate the Neurospora plasma membrane H"*" ATPase into microsomes. J. Biol. Chem. 270,6935-6941. Lin, J. & Addison, R. (1995b). The membrane topology of the carboxyl terminal third of the Neurospora plasma membrane H"*" ATPase. J. Biol. Chem. 270,6942-6948. MacLennan, D.H. (1990). Molecular tools to elucidate problems in excitation contraction coupling. Biophys. J. 58, 1355-1365. MacLennan, D.H., Brandle, C.J., Korezak, B., & Green, N.M. (1985). Amino-acid sequence of a Ca + Mg -dependent ATPase from rabbit muscle sarcoplasmic reticulum, deduced from its complementary DNA sequence. Nature 316,696-7(X). Mitchell, P. (1967). Translocation through natural membranes. Adv. Enzymol. 29,33-87. Mitchell, R (1981). In: Of Oxygen, Fuels, and Living Matter (Semenza, G., ed.). Part I, pp. 1-56. John Wiley and Sons, Chichester, England. Rao, U.S., Hennessey, J.P., & Scarborough, G.A. (1991). Identification of the membraneembedded regions of the Neurospora crassa plasma membrane H"*" ATPase. J. Biol. Chem. 266, 14740-14746. Rees, D.C., Komiya, H., Yeates, TO., Allen, J.R, & Feher, G. (1989). The bacterial photosynthetic reaction center as a model for membrane proteins. Annu. Rev. Biochem. 58, 607-633. Scarborough, G.A. (1976). Proton translocation catalyzed by the electrogenic ATPase in the plasma membrane of Neurospora. Proc. Natl. Acad. ScL USA 73,1485-1488. Scarborough, G.A. (1978). Isolation of plasma membrane from Neurospora. Methods CellBiol. 20, 117-133. Scarborough, G.A. (1980). The Neurospora plasma membrane ATPase is an electrogenic pump. Biochemistry 19, 2925-2931. Scarborough, G.A. & Addison, R. (1984). On the subunit composition of the Neurospora plasma membrane H"^ ATPase. J. Biol. Chem. 259,9109-9114. Scarborough, G.A. & Hennessy, J.R, Jr. (1990). Identification of the major cytoplasmic regions of the Neurospora crassa plasma membrane H"*" ATPase using protein chemical techniques. J. Biol. Chem. 265, 16145-16149. Segel, I.H. (1975). In: Enzyme Kinetics: Behavior and Analysis of Rapid Equilibrium and Steady State Enzyme Systems. John Wiley & Sons, New York.

Topogenesis of Neurospora ht ATPase

235

Serrano, R. (1989). Structure and function of plasma membrane ATPase. Annu. Rev. Plant Physiol. Plant Mol. Biol. 40, 61-94. Shull, G.E., Schwartz, A., & Lingrel, J.B. (1985). Amino-acid sequence of the catalytic subunit of the (Na"*" + K'^)ATPase deduced from a complementary DNA. Nature 316,691-695. Slayman, C.L., Lu, C.Y-H., & Shane, L. (1970). Correlated changes in membrane potential and ATP concentrations in Neurospora. Nature 226, 274-276. Wiedmann, B., Sakai, H., Davis, T.A., & Wiedmann, K. (1994). A protein complex required for signal-sequence-specific sorting and translocation. Nature 370,434440.

This Page Intentionally Left Blank

NUCLEAR TRANSPORT AS A FUNCTION OF CELLULAR ACTIVITY

Carl M. Feldherr and Debra Akin

I. Introduction II. Macromolecular Transport across the Nuclear Envelope III. Regulation of Transport across the Nuclear Envelope A. Regulation of Specific Proteins B. Changes in Transport Capacity IV. Conclusions Acknowledgments References

237 238 240 240 241 254 255 256

I. INTRODUCTION Considering the strategic location of the nuclear envelope in eukaryotic cells, it could play an important role in regulating the transcription and translation of genetic information by modulating the nucleocytoplasmic exchange of macromolecules. In this report, we discuss the data relating to this fundamental question, specifically, whether the capacity for

Membrane Protein Transport Volume 2, pages 237-259 Copyright © 1995 by JAI Press Inc. All rights of reproduction in any form reserved. ISBN: 1-55938-983-4

237

238

CARL M. FELDHERR and DEBRA AKIN

signal-mediated nuclear transport varies as a function of cellular activity. Before considering these results, however, it is necessary to review the basic mechanisms by which large molecules or molecular aggregates are able to cross the nuclear envelope. Since the overall transport process as well as specific elements of the transport machinery have been the subject of numerous recent reviews (e.g., Forbes, 1992; Izaurralde and Mattaj, 1992; Newmeyer, 1993; Fabre and Hurt, 1994; Feldherr and Akin, 1994a; Pante and Aebi, 1994), only those aspects of the exchange process relevant to regulation are summarized in the next section.

II. MACROMOLECULAR TRANSPORT ACROSS THE NUCLEAR ENVELOPE The exchange of macromolecules across the envelope occurs through the nuclear pores, by either passive diffusion or signal-mediated transport. Diffusion occurs through a 90-100-A channel located at the center of each pore (Paine et al., 1975; Peters, 1986) and, perhaps, at peripheral sites along the pore margins (Hinshaw et al., 1992). The rates of diffusion are inversely related to the dimensions of the permeant molecules; for example, dextran with a hydrodynamic diameter of 24 A will equilibrate between the nucleoplasm and cytoplasm within 30 minutes after being injected into Xenopus oocytes, whereas 46 A dextran requires approximately 15 hours for equilibration. When the diameter increases to 71 A, the equilibration time increases to approximately 7 days (Paine et al., 1975). Signal-mediated transport of macromolecules across the envelope is a multistep process that is initiated by specific targeting signals. Signals that direct nuclear import are generally referred to as nuclear localization signals (NLSs) and have been identified in over 50 proteins (Richter and Standiford, 1992; BouHkas, 1994). Although there is no consensus sequence, NLSs can be divided into two general categories, simple and bipartate. The simple signals are short basic domains containing fewer than ten amino acids, whereas the bipartate signals comprise two basic regions separated by ten or more spacer amino acids. The large T antigen NLS (PKKKRKV; Kalderon et al, 1984) and the nucleoplasmin NLS (AVKRPAATKKAGQAKKKKLDGG; Robbins et al., 1991), respectively, are prototypes of these two signal categories. There is reason to believe that the signal content of specific proteins can play a regulatory role in nuclear import. In this regard, it has been demonstrated that NLSs are not all equally effective in transport. Fur-

Nuclear Transport and Cellular Activity

239

thermore, both the rate of nuclear import and the effective size of the transport channel (see below) are dependent on the number of available signals, which can vary among different proteins. The significance of signal content in nuclear protein uptake has been reviewed in detail by Feldherr and Akin (1994a). The NLSs function by interacting with cytoplasmic transport factors that are thought to direct the signal-containing permeant molecules to docking sites along the cytoplasmic surface of the pores. The nature of the cytoplasmic factors involved in this process is currently under investigation in several laboratories. So far, the following proteins have been identified as transport components; a 54~55-kDa polypeptide that functions in NLS binding (Adam and Gerace, 1991), a 97-kDa polypeptide (Adam and Adam, 1994), and Ran/TC4, which is a 25-kDa GTPase (Melchior et al., 1993; Moore and Blobel, 1993). Whether the complex between the permeant and cytoplasmic factors dissociates at the pore surface or enters the nucleus prior to dissociation is not known. After binding to the pore surface, translocation across the envelope occurs through the central region of the pore complex, the same area that functions in diffusion. However, in the presence of an NLS the central channel is able to dilate from approximately 90 A to over 230 A in diameter in normal proliferating cells. The existence of a central transport channel is supported by both functional and morphological studies (i.e., Feldherr et al., 1984; Akey and Radermacher, 1993). Other than the fact that the translocation step is energy dependent (Newmeyer and Forbes, 1988; Richardson et al., 1988) and the degree of dilation of the central channel is variable and related to signal number (Dworetzky et al., 1988), there is no information regarding the mechanism of exchange through the pore complex. Based largely on its molecular mass (approximately 125 x 10^ kDa), it has been estimated that the pore complex is made up of over 100 different proteins, referred to as nucleoporins. Of this hypothetical number, at least 12 vertebrate nucleoporins have been identified. These include (1) a family of 0-linked glycoproteins that bind wheat germ agglutinin (WGA) and contain FXFG repeat sequences, (2) transmembrane proteins that extend from the pore complex into the lumen of the nuclear envelope, and (3) at least one protein that is associated with the fibrillar elements that extend from the cytoplasmic surface of the pores (Starr and Hanover, 1992; Fabre and Hurt, 1994). At least three nucleoporins in yeast also contain FXFG repeats; however, it is not known whether these homologs of the vertebrate proteins are also glycosylated.

240

CARL M. FELDHERR and DEBRA AKIN

In addition, a second subfamily of yeast nucleoporins has been identified; the proteins that make up this group contain up to 33 copies of a GLFG motif (Wente et al, 1992). It has been shown in vertebrate cells that WGA or antibodies against the pore glycoproteins block signal-mediated nuclear import, but not passive diffusion. Based on a report by Steme-Marr et al. (1992) that 0-linked glycoproteins can bind essential cytoplasmic transport factors, it is possible that these substances act as receptors. In yeast, at least two of the nucleoporins (one each in the FXFG and GLFG subfamilies) are nonessential, although deletions are lethal when accompanied by mutations of other nucleoporins. These results suggest that specific pore proteins may have overlapping functions. There are also indications that the transmembrane proteins can modulate both mediated and passive transport through the pores (Greber and Gerace, 1992). Hi. REGULATION OF TRANSPORT ACROSS THE NUCLEAR ENVELOPE A. Regulation of Specific Proteins Cells have developed several strategies for controlling the nucleocytoplasmic distribution of specific regulatory proteins. These include (1) anchoring to cytoplasmic elements, (2) masking the NLS by binding to transport inhibitors, and (3) phosphorylation at sites flanking the NLS. These mechanisms are illustrated by the following examples. In the inactive state, the catalytic subunits of the cAMP-dependent kinase (PKA) are anchored to regulatory polypeptides associated with the Golgi. In the presence of cAMP, the catalytic subunits, but not the regulatory elements, are released from their anchorage sites and enter the nucleus. As the cAMP concentration decreases, the inactive PKA complex reforms at the Golgi (Nigg et al., 1985). Masking of the NLSs is required for the cytoplasmic retention of the transcription factors NF-KB, Rel, and dorsal (Govind and Steward, 1991; Schmitz et al., 1991; Blank et al., 1992; Grimm and Baeuerle, 1993). The inhibitors that bind these proteins can be released by several different stimuli (e.g., kinases or second messengers), thus initiating the nuclear import process. The cell-cycle-dependent transcription factor SWI5 in yeast is localized in the cytoplasm of cells in S, G2, and M. During these phases of division, three serine residues associated with the NLS are phosphorylated. Translocation of SWI5 into the nucleus, which occurs during Gj, is

Nuclear Transport and Cellular Activity

241

accompanied by dephosphorylation of the serines (Moll et al., 1991). Similarly, there is evidence that the effectiveness of the SV40 large T NLS is related to the phosphorylation of casein kinase II sites (which increases transport) or cdc2 kinase sites (which decreases the level of import) in the region flanking the signal (Rihs and Peters, 1989; Jans et al., 1991; Rihsetal., 1991). B. Changes in Transport Capacity

In addition to the above mechanisms that affect the nucleocytoplasmic exchange of specific macromolecules, it is conceivable that more general changes in signal-mediated transport, which are the direct result of alterations in the transport machinery itself, occur during different stages of cellular activity. This possibility was initially considered by Feldherr and Akin (1990), who investigated the nuclear import of nucleoplasmincoated gold (NP-gold) in proliferating, quiescent, and differentiating 3T3-L1 cells. The NP-gold was introduced into the cytoplasm by microinjection. The cells were fixed 30 minutes later and subsequently examined with a transmission electron microscope (Figure 1). Nucleoplasmin is a karyophilic protein found in Xenopus oocytes. It is com-

^' -

;1

Figure 1. An electron micrograph of a proliferating 3T3-L1 cell showing the intracellular distribution of NP-gold (arrows) 30 minutes after injection into the cytoplasm. By this time, the mean N/C gold ratio was 0.67. N, nucleus; C, cytoplasm. Bar, 0.5 |xm.

CARL M. FELDHERR and DEBRA AKIN

242

• Cytoplasm C3 Nuclei, proliferating 0 Nuclei, confluent

40-80

80-120

120-160

200 +

Particle Size (A) Figure 2. The size distribution of NP-gold particles that entered the nuclei of proliferating and confluent 3T3-L1 cells 30 minutes after injection. The size of the particles that were initially injected into the cytoplasm and were available for transport is also shown. From Feldherr and Akin (1990).

posed of five 22 kDa subunits, each of which contains one copy of a well-characterized bipartate NLS (see above), which is highly effective in nuclear targeting. By adsorbing the nucleoplasmin to colloidal gold, it is possible not only to determine the relative rates of nuclear import (by counting gold particles in equal areas of nucleoplasm (N) and cytoplasm (C) and calculating the N/C ratio), but also to establish the functional size of the transport channels (by measuring the particles that entered the nucleus). It was found in these studies that the N/C gold ratio obtained for 110-270-A gold particles (total diameter including the 30-A protein coat) was significantly greater in proliferating as compared to quiescent cells (0.67 vs. 0.09, respectively). In addition to the decrease in relative import, there was a significant decrease in the size of the particles that were able to enter the nuclei of quiescent cells (see Figure 2), reflecting a change in the functional diameter of the transport channel from approximately 230 A to 190 A. After differentiation of the quiescent cells to adipocytes, nuclear import of NP-gold returned to the proliferating cell level. It could be demonstrated that the results were not due to either

Nuclear Transport and Cellular Activity

243

changes in pore number or differences in the intracellular migration patterns of the larger gold particles during different physiological states, leading to the conclusion that alterations had occurred in the transport process itself. A significant decrease in the relative import of large NP-gold particles was also observed when proliferating BALB/c 3T3 cells entered GQ after growing to confluence or after serum depletion (Feldherr and Akin, 1991). In fact, the functional decrease in pore size in these cells was as great as 100 A, compared to approximately 40 A in 3T3-L1 cells. In these studies it was also found that the decrease in import rate that occurs during growth arrest, at least in BALB/c cells, is size dependent; thus, the N/C ratio of small NP-gold particles (50-80 A total diameter) did not change significantly as proliferating cells became quiescent. These results (i.e., the reduced uptake of large but not small gold particles) are consistent with the interpretation that transport in quiescent cells is limited primarily by a decrease in the exclusion size limit of the exchange channels. The above results show that nuclear transport capacity can vary as a function of cellular activity. It has been suggested (Feldherr and Akin, 1991) that one of the more important consequences of the observed changes would be a decrease in the efflux of RNP particles, since they have dimensions that can exceed the functional size of the pores in quiescent cells. This would explain, at least in part, the decrease in both cytoplasmic rRNA and protein synthesis that is known to occur during quiescence (Levine et al., 1965; Johnson et al., 1974, 1976). Since growth factors are required to maintain the proliferating state, it is possible that they are also involved in modulating nucleocytoplasmic transport. This was investigated by studying the effects of PDGF, IGF-1, and epidermal growth factor (EGF) on the recovery of nuclear transport capacity as quiescent BALB/c cells reentered the growth cycle (Feldherr and Akin, 1993). PDGF induces competence in GQ cells, and subsequent treatment with EGF and IGF-1 is necessary for progression through the cell cycle (Yang and Pardee, 1986; Ross et al., 1987). Nuclear transport in serum-depleted cells (cultured in 0.5% calf serum for 4 days) was assayed, using NP-gold, at intervals of 6, 12,18, and 24 hours after the addition of either growth factors or 10% serum. The results can be summarized as follows. Recovery of transport activity, after the addition of 10% calf serum, occurred gradually and was complete after 18-24 hours. Partial recovery was achieved by adding growth factors individually at physiological concentrations in either 0% or 0.5% calf serum.

244

CARL M. FELDHERR and DEBRA AKIN

Complete recovery in the absence of serum required both EGF and IGF-1. Surprisingly, PDGF acted as an antagonist when added to this mixture, suggesting that in combination these factors could function as either positive or negative regulators of transport. Whether growth factors act by altering the levels of translation of specific transport components or by causing posttranslational modifications (e.g., by activating kinases or phosphatases) remains to be determined. Cell Fusion

Experiments

To determine whether the differences in nuclear transport capacity between proliferating and quiescent cells are due to cytoplasmic agents (specifically, ATP or transport factors that are known to be necessary for nuclear import) or variations in the properties of the nuclear envelope, Feldherr and Akin (1993) studied nuclear import of NP-gold in mechanically fused cells. Proliferating and serum-starved BALB/c 3T3 cells were superimposed and simultaneously punctured with a microneedle. The cells were left undisturbed for 50 to 60 minutes after fusion to allow thorough mixing of the cytoplasmic components. They were then injected with an NP-gold fraction containing 110-270-A particles (total diameter), in order to analyze nuclear import. The results are shown in Table 1. It was initially found, in control studies, that the experimental procedure had no detectable effect on nuclear import; thus, when growtharrested cells were fused together, nuclear transport remained at a level normally observed for quiescent cells. Similarly, fusion of proliferating Table 1. Nuclear Uptake of Nucleoplasmin-Gold in Fused BALB/c Cells N/C Ratio (Mean-v SE) Fusion Quiescentquiescent Proliferatingproliferating Proliferatingquiescent

No. of Fusions 2 3 8

Quiescent 0.23 + 0.02(416)*' — 0.30 + 0.02(1727)

Proliferating

Significance^

—

—

1.67 + 0.06(2114)

—

1.36 + 0.06(2531)

s

Notes: ^ s indicates that there is a significant difference {P < 0.01) in NP-gold uptake in quiescent versus proliferating nuclei. From Feldherr and Akin (1993). Number of particles counted in parentheses. Source: From Feldherr and Akin (1993).

Nuclear Transport and Cellular Activity

245

cells did not alter nuclear import of NP-gold. In the heterokaryons, it was found that the difference in transport capacity between proliferating and quiescent nuclei was maintained after cell fusion. Since the nuclei of the fused cells shared the same cytoplasm, it was deduced that the differences in transport were due to variations in the properties of the nuclear envelope (presumably at the level of the pores), rather than changes in the availability or activity of cytoplasmic components. In addition to a reduction in transport capacity, the number of NP-gold particles associated with the pores was also significantly lower in quiescent versus proliferating nuclei, suggesting a difference in the effective number of pore complex receptors. This could be an important factor in regulating transport if the number of receptor-signal interactions determines the functional pore size, as suggested by Dworetzky et al. (1988). The Effects of Cell Shape

A number of investigators have demonstrated a relationship between cell shape and metabolic activity. In anchorage-dependent cells, for example, conversion from an extended to a more spherical state is accompanied by a significant change in nuclear function, specifically a decrease in the synthesis of DNA and RNA (Ingber et al., 1987; Ingber and Folkman, 1989). To determine whether signal-mediated nuclear transport capacity is similarly affected, Feldherr and Akin (1993) investigated the nuclear import of NP-gold in B ALB/c cells as a function of shape. Cells in different configurations were obtained according to the method reported by Ireland et al. (1987). In the initial step of this procedure, coverslips are coated with poly-HEMA (poly-2-hydroxyethyl methacrylate), which forms a nonadhesive coat that is unable to support cell attachment and growth. Palladium is then deposited on the polyHEMA through a copper mask that contains circular openings ranging in diameter from 22 to 80 |Lim. Cells that are plated on these surfaces attach only to the palladium; those adhering to the small domains remain rounded, whereas cells that attach to the larger areas of palladium are extended. The relative uptake rates of two different size NP-gold fractions (50-80 A and 110-270 A, total diameter) into the nuclei of rounded and extended cells is shown in Table 2. These results demonstrate that transport capacity is reduced in rounded cells, and, consistent with the changes detected in quiescent cells, the decrease is size dependent. Assuming that cell geometry is a reflection of cytoskeletal organization, transport capacity could be modulated by dynamic changes in the

246

CARL M. FELDHERR and DEBRA AKIN Table 2. The Effect of Cell Shape on Nuclear Transport N/C

Experiment A. Rounded Extended B. Rounded Extended Notes:

No. of cells NP'GoldSize 19 19 19 19

110-270 110-270 50-80 50-80

Mean ± SE

0.52 + 0.09 (644)*' 1.59 + 0.14(872) 1.51+0.15(2724) 1.48 + 0.11 (2041)

Significance^

s NS

NP-gold import is s, significantly different (P < 0.01) between rounded and extended cells. NS, not significant. Number of particles counted in parentheses.

Source: From Feldherr and Akin (1993).

interactions between the nuclear surface and elements of the cytoskeleton. For example, the functional dimensions, and perhaps the molecular organization of the pore complex could be altered by tension applied to the nuclear envelope. Intermediate filaments might play a key role in this regard, since they are known to interact with the envelope (Goldman et al., 1986) and, perhaps, the pore complex itself (Carmo-Fonseca et al., 1987). Comprehensive models suggesting mechanisms by which the cytoskeleton could alter nuclear activity in general and transport in particular have been proposed by Ingber and Folkman (1989) and Hanson and Ingber (1992), respectively. Alternatively, the transport effects could be dependent on an association between the cytoskeletal elements and regulatory substances such as transcription factors and kinases. It has been suggested that the release and activation of these regulatory agents might result from changes in the molecular organization of the cytoskeleton (Ben-Ze'ev, 1991). If the latter mechanism is functional, the alterations in the transport apparatus might be similar to those produced by growth factors, which would also be expected to involve transcriptional and post-translational changes. In fact, the two processes could be interrelated, since there is evidence that modifications in the assembly of the cytoskeleton can be induced by growth factors (Herman and Pledger, 1985). Increases in Transport Capacity in SV40'Transformed Cells

The results described above established that decreases in metabolic activity (as a result of either growth arrest or shape changes) can be accompanied by a reduction in signal-mediated transport capacity. This

Nuclear Transport and Cellular Activity Table 3.

Experiment SVT2 BALB/c controls pSVSneotransformed BALB/c controls

247

Nucleoplasmin-Gold Uptake in Transformed Cells

% Particles^ No. of No. of Particles 80200Cells Measured 200A 320A

N/C

Mean ± SE

Significance

Size

N/C

12 10

432 675

65.7 95.7

34.3 4.3

1.8010.15(814)'^ 1.0010.07(829)

s

s

25

707

82.8

17.2

2.1110.15(1228)

s

s

22

428

94.6

5.4

0.9910.16(1173) —

—

^ Particle measurements do not include the 30-A protein coat, s, significantly different (P < 0.01) than controls. ^ Number of particles counted in parentheses. Source: From Feldherr et al. (1992). Notes:

raises the possibility that an increase in transport capacity (above the normal proliferating cell level) might occur during periods of enhanced cellular activity. In this regard, Feldherr et al. (1992) analyzed nuclear transport in SV40-transformed BALB/c cells (SVT2 cell line) and BALB/c 3T3 cells that were transfected with the early region of the S V40 genome (pSV3neo plasmid) that encodes both the large T and small t antigens. These cell lines exhibited more rapid growth rates than nontransformed cells. In addition, there were statistically significant increases, compared to normal proliferating BALB/c controls, in both the nuclear uptake of NP-gold and the functional diameter of the transport channel (see Table 3). The latter value increased by approximately 40 A. Overall, the changes in transport triggered by SV40 could function to expedite DNA replication, thereby facilitating the production of virions in permissive cells. In nonpermissive cells, this could contribute to the increased division rate and help bring about the transformed state. Since, in most instances, large T alone is sufficient for transformation, we examined its effect on signal-mediated transport. In these experiments, purified large T was injected into the cytoplasm of proliferating BALB/c cells and nuclear transport was assayed either 1 or 6 hours later with NP-gold. The effect of injecting a mutant form of large T, (cT)-3, was also tested. This mutant, which contains asparagine rather than lysine at position 128, is deficient in nuclear translocation and is retained in the cytoplasm. Control cells were either untreated or preinjected with

248

CARL M. FELDHERR and DEBRA AKIN

bovine serum albumin before the transport capacity was measured. The results, given in Table 4, established that a 6-hour pretreatment with large T can increase transport to the level observed in transformed cells; however, there was no significant change in gold import after 1 hour. Interestingly, the (cT)-3 mutant had no effect on nuclear transport when injected into the cytoplasm, but caused a significant increase when introduced into the nucleoplasm. Considering both the time dependence and the requirement for a functional NLS, it can be concluded that large T does not act directly on the transport apparatus (which is presumably localized in the cytoplasm) but indirectly, by mediating the activity of an effector protein(s). In order to identify the large T domain responsible for enhancing transport, Feldherr et al. (1994) investigated the effects of 12 different mutants on the nuclear uptake of NP-gold. The locations of the lesions, in relation to known functional domains of large T, are shown in Figure 3. Plasmids encoding the mutant proteins were injected into the nuclei of BALB/c cells, and nuclear transport was assayed approximately 18 hours later. In all but two cases, there was a significant increase in nuclear transport. Both of the inactive mutants (1061 and 2809) were also unable to bind the tumor suppressor protein p53, suggesting that the transport activity and p53 binding colocalize to the same domain. This was further supported by the finding that simultaneous injection of a plasmid that overexpresses wild-type p53 (PC53-SN3) prevented the increases in nuclear uptake normally caused by the expression of active forms of the large T antigen. In view of the above findings, additional experiments were performed to establish the effect of p53 on NP-gold import in cells that were not expressing large T. Nuclear injection of the plasmid pC53-SN3 had no effect on transport capacity; however, plasmid pC53-Cx33, which expresses a mutant form of p53, significantly increased NP-gold import (see Table 5, part A). The fact that excess wild-type p53 did not alter nuclear uptake probably indicates that the transport system was already saturated by endogenous p53. The effect of the mutant protein, on the other hand, was likely due to a decrease in the effective concentration of endogenous p53. Thus, Milner and Medcalf (1991) reported that mutant and wild-type p53 form inactive oligomers, the net result being a decrease in the active form of p53. More direct evidence that decreasing the intracellular p53 concentration enhances nuclear transport was obtained by investigating the effects of two anti-p53 monoclonal antibodies, designated pAb421 and pAbl22. In separate experiments, the

Table 4. The Effect of Large T Antigen on Nucleoplasmin-Gold Uptake

NO.

Experiment

BSA injection, 6 hr Large T injection, 1 hr Large T injection, 6 hr (cT)-3 injection, cyto, 6 hr (cT)-3 injection, nuc, 6 hr BALBIc untreated controls Notes:

of

cells 16 16 19 16 18 16

No. of Particles Measured 363 402 545 47 1 454 519

80-200.4 %.4 91.0 75.6 94.5 85.2 95.8

a Particle measurements do not include the 30-A protein coat.

s, significantly different (P < 0.01) than controls; NS, not significant. Number of particles counted in parentheses. Source: Fmm Feldherr et al. (1992).

2~320.4 3.6 9.0 24.4 5.5 14.8 4.2

Mean k SE 0.63 k 0.07 (915)' 0.72 k 0.07 (1369) 2.03 f 0.1 1 (1492) 1.10k0.11 (1188) 1.52 0.10 (479) 0.71 k0.12 (1061)

Size

N/C

NS NS s NS s

NS NS

-

s

NS s -

Pola (1-82)

Rb (102-115) — NLS (126-132)

pBSVcT

DNA binding (131-259)

395HN ,

402DE 402DH

>

p53 binding (351-450)

•

408F4 2809 ATPase (418-627)

p53 binding (533-626)

I Host Range

I (682-708)

Figure 3. A map of large T showing the locations of the mutations (listed on the left) in relation to known functional domains of the antigen (listed on the right). A M 76, 1061, and 2465 are deletion mutations. 395HN, 402DE, 402DH, and pBSVcT are single amino acid substitutions. The remainder are insertion mutations. Further details describing the nature of the lesions are given in Feldherr et al. (1994). 250

Table 5. The Effect of p53 on Nuclear Transport No.of Emeriment bd

2

Cells

No. of

% particlesa

Particles Measured

8&200A

4.8 12.2 3.6

200-320A

A. pC53-SN3 ~C53-Cx3~ BALBIc controls

18 15 15

400 41 1 353

95.2 87.8 96.4

B. pAb421 Mouse IgG2a controls

15 15

372 340

87.5 95.4

12.5 4.6

pAb122 Mouse lgG2a controls

30 16

687 333

85.4 94.9

14.6 5.1

Notes:

'particle measurements do not include the 30-A protein coat. s, significantly different (P < 0.01) than controls; NS, not significant. Number of particles counted in parentheses.

Source: From Feldherr et al. (1994).

N/C Mean _+ SE

significanceb Size

N/C

252

CARL M. FELDHERR and DEBRA AKIN

antibodies were injected into the cytoplasm of BALB/c 3T3 cells, and transport was analyzed 1 or 2 hours later. Mouse IgG2a was injected into control cells. As shown in Table 5, part B, both anti-p53 antibodies significantly increased the N/C ratios of NP-gold, as well as the functional diameters of the pores. These results are consistent with, but do not prove, the view that large T modulates nuclear permeability by sequestering p53, which, in turn, acts as a transport inhibitor. p53 could function either by activating genes that encode specific repressors, or by blocking the expression of enhancing factors. The Increase in Transport Is Initiated by a Cytoplasmic Factor In unpublished experiments, Feldherr and Akin prepared a 100,000 x g cytoplasmic supernatant from SV40-transformed cells, using the general procedures described by Adam et al. (1992), and studied its effect on nuclear permeability in normal proliferating B ALB/c 3T3 cells. The supernatant was microinjected into the cells, and NP-gold was injected 10 minutes later. Even after this relatively short time interval, there was a significant increase in both the number and size of particles that were able to enter the nucleoplasm, as compared to untreated controls. Injecting a similar cytoplasmic fraction prepared from nontransformed BALB/c cultures had no effect on NP-gold import. It is unlikely that these results were due simply to the retention of large T antigen in the transformed cell extract. As mentioned above, the injection of purified large T (at concentrations that can easily be detected by immunofluorescence) requires over an hour to significantly alter transport capacity. The extract, on the other hand, causes an increase within 10 minutes, and large T is not present in sufficient amounts to be detected by immunofluorescence procedures. It is also necessary to consider whether the enhancing factor(s) in transformed cells correspond to one or more of the known cytoplasmic factors that normally mediate transport (the 54-56-kDa and 97-kDa proteins and Ran/TC4). This could be the case if one of the factors was rate limiting and was produced in considerably larger amounts after transformation. However, based on the results obtained with transformed cell extracts this seems unlikely. It is doubtful that a sufficient amount of material was injected in these experiments (approximately 5% of the cell volume) to significantly increase the concentration of any of the three transport factors. Although the possible involvement of these proteins cannot be formally ruled out, the results are more consistent with the

Table 6. Nuclear Uptake of Nucleoplasmin-Gold during the Cell Cycle % Particles in Nucleusa

N/C Ratio

Measured

80-200A

2O0-320A

Mean f SE

Size

18 20

473 449

85.5 95.4

14.5 4.6

1.44 + 0.08 (960)' 0.55 + 0.04 (566)

s

2 hr. post-anaphase BALBIc control

8 13

341 535

91.8 95.7

8.2 4.3

0.89 + 0.17 (631) 0.48 + 0.07 (788)

NS

NS

4 hr. post-anaphase

20 11

444

5.0 4.9

0.54 + 0.06 (677) 0.55 + 0.04 (566)

NS

448

95.0 95.1

NS

BALBIc control

6 hr. post-anaphase BALBIc control

18 19

478 561

94.6 95.4

5.4 4.6

0.63 + 0.05 (451 ) 0.57 + 0.06 (1116)

NS

NS

8 hr. post-anaphase

14 19

384 561

94.3 95.4

5.7 4.6

0.48 + 0.06 (412) 0.57+0.06(1116)

NS

NS

BALBIc control

12 hr. post-anaphase BALBIc control

19 13

599 535

95.8 95.7

4.2 4.3

0.31 + 0.03 (1506) 0.48 + 0.07 (788)

NS

NS

18 hr. post-anaphase BALBIc control

9 13

380 481

94.5 95.6

5.5 4.4

0.65 + 0.05 (614) 0.59 + 0.10 (788)

NS

NS

21 hr. post anaphase BALBIc control

20 16

629 573

98.7 95.8

1.3 4.2

0.98 + 0.06 (2635) 0.86 + 0.05 (1754)

s

NS

No. of

No.ofPanicles

cell;

I hr. post-anaphase BALBIc controld

Emerimenr

w

Notes:

a Particle measurements do not include the 30A protein coat.

s, significantly different (P< 0.01) than controls; NS, not significant.

'Number of particles counted in parentheses.

Control cells were randomly selected from non-synchronized, proliferating cultures. Source: From Feldherr and Akin (1994b).

significanceb

N/C

-

254

CARL M. FELDHERR and DEBRA AKIN

view that the enhancing factor(s) has a catalytic effect and could possibly be a kinase or phosphatase. Signal-Mediated Transport During the Cell Cycle Studies performed on dividing BALB/c 3T3 cells by Feldherr and Akin (1994b) demonstrated that changes in nuclear transport capacity not only occur during extensive, long-term modifications in metabolic activity (such as those that take place following growth arrest and transformation) but also accompany more transient cellular processes. In this investigation, cells in different stages of the division cycle were obtained by first performing shake-offs to increase the proportion of cells in mitosis. These cells were then plated on coverslips, and those entering anaphase were scored and followed individually for up to 21 hours. At various intervals from 1 to 21 hours after the start of anaphase, nuclear transport was assayed with NP-gold. The nuclear import data are shown in Table 6. There was a significant increase in both the N/C ratio and the diameters of the transport channels 1 hour after the onset of anaphase. EM analysis of the nuclear envelopes in early Gj cells and import studies showing that large gold particles coated with PVP (which lacks an NLS) were excluded from the nucleoplasm demonstrated that the transport increase detected at 1 hour was not due to incomplete reformation of the nuclear envelope, but rather to differences in the properties of the transport machinery. The observed increase in nuclear import at this early time in the cell cycle could play a significant role in the reformation of the nucleus following division. At the 21-hour time point there was a decrease in the size of the particles that were present in the nucleus, but no change in the N/C gold ratio. The effect of these transport changes on the division process is not known, but the results do suggest that nuclear import rates and channel size can be regulated independently under certain circumstances. No differences in transport capacity, compared to control cells, were detected at other times in the cell cycle; however, it is possible that transient changes occurred but were not detected. IV. CONCLUSIONS Signal-mediated nuclear import of macromolecules is a multistep process that is initiated by specific nuclear targeting signals that interact with cytoplasmic receptors. The receptor-substrate complex then migrates to

Nuclear Transport and Cellular Activity

255

the cytoplasmic surface of the pore complex, where it binds to a docking protein(s). Docking to the pore surface activates an energy-dependent translocation process that results in the movement of the targeted molecule to the nucleoplasmic surface of the pore. This is followed by release into the nucleoplasm, which represents the final step in the process. It is generally assumed that nuclear efflux utihzes the same general mechanism; however, differences are likely to exist with respect to the specific signals and their receptors. In this review, evidence is presented demonstrating that nuclear transport is a dynamic process that can increase or decrease in effectiveness in response to changes in cell function. Furthermore, it appears that regulation of transport capacity is accomplished by several different mechanisms that can act at different steps in the import process. Consistent with the decrease in protein and RNA synthesis that occur during growth arrest, there is a significant decrease in macromolecular exchanges across the envelope. Cell fusion studies, performed with arrested and proliferating cells, demonstrated that the changes affecting transport occur at the level of the envelope, presumably within the pores. The differences in transport could be related to the number or binding capacity of receptors at the pore surface. A size-dependent decrease in nuclear import also occurred when cells were prevented from spreading, implicating the cytoskeleton in transport regulation. In contrast to the above results, an increase in the rate of cell growth after SV40 transformation was accompanied by an increase in both the number and size of NP-gold particles that are able to enter the nucleus. Preliminary studies indicate that the increase in nuclear import in transformed cells is dependent on cytoplasmic enhancing factors. There are indications that enhancer activity is triggered by large T antigen. Although the evidence is not conclusive, it is likely that large T functions by sequestering p53, which appears to act as a transport suppressor. Since an increase in transport capacity was also detected during the cell cycle, it is interesting to speculate that transport enhancers and suppressors are normal components of cells, and that altering the balance between these factors can significantly alter transport capacity.

ACKNOWLEDGMENTS The authors would like to thank Dr. Robert Cohen for critically reviewing the manuscript.

256

CARL M. FELDHERR and DEBRA AKIN

REFERENCES Adam, E. & Adam, S. (1994). Identification of cytosolic factors required for nuclear location sequence-mediated binding to the nuclear envelope. J. Cell Biol. 125, 547-555. Adam, S. & Gerace, L. (1991). Cytosolic proteins that specifically bind nuclear location signals are receptors for nuclear import. Cell 66, 837-847. Adam, S., Steme-Marr, R., & Gerace, L. (1992). Nuclear protein import using digitoninpermeabilized cells. Methods Enzymol. 219,97-110. Akey, C. & Radermacher, M. (1993). Architecture of the Xenopus nuclear pore complex revealed by three dimensional cryo-electron microscopy. J. Cell Biol. 122,1-19. Ben-Ze'ev, A. (1991). Animal cell shape changes and gene expression. Bioessays 13, 207-212. Blank, V., Kourilsky, R, & Israel, A. (1992). NF-kB and related proteins: rel/dorsal homologies meet ankyrin-like repeats. Trends Biochem. Sci. 17,135-140. Boulikas, T. (1994). Putative nuclear localization signals (NLS) in protein transcription factors. J. Cell. Biochem. 55, 32-58. Carmo-Fonseca, M., Cidadao, A., & David-Ferreira, J. (1987). Filamentous crossbridges link intermediate filaments to the nuclear pore complexes. Eur. J. Cell Biol. 45, 282-290. Dworetzky, S., Lanford, R., & Feldherr, C. (1988). The effects of variations in the number and sequence of targeting signals on nuclear uptake. J. Cell Biol. 107,1279-1287. Fabre, E. & Hurt, E. (1994). Nuclear transport. Curr. Opin. Cell Biol. 6, 335-342. Feldherr, C. & Akin, D. (1990). The permeability of the nuclear envelope in dividing and nondividing cell cultures. J. Cell Biol. Ill, 1-8. Feldherr, C. & Akin, D. (1991). Signal-mediated nuclear transport in proliferating and growth-arrested BALB/c 3T3 cells. J. Cell Biol. 115,933-939. Feldherr, C. & Akin, D. (1993). Regulation of nuclear transport in proliferating and quiescent cells. Exp. Cell Res. 205,179-186. Feldherr, C. & Akin, D. (1994a). Role of nuclear trafficking in regulating cellular activity. Int. Rev. Cytol. 151,183-227. Feldherr, C. & Akin, D. (1994b). Variations in signal mediated nuclear transport during the cell cycle in BALB/c 3T3 cells. Exp. Cell Res. 215, 206-210. Feldherr, C , Kallenbach, E., & Schultz, N. (1984). Movement of a karyophilic protein through the nuclear pores of oocytes. J. Cell Biol. 99, 2216-2222. Feldherr, C , Lanford, R., & Akin, D. (1992). Signal mediated nuclear transport in simian virus 40 transformed cells is regulated by large tumor antigen. Proc. Natl. Acad. Sci. USA 89, 11002-11005. Feldherr, C, Cole, C , Lanford, R., & Akin, D. (1994). The effects of SV40 large Tantigen and p53 on nuclear transport capacity in BALB/c cells. Exp. Cell Res. 213, 164-171. Forbes, D. (1992). Structure and function of the nuclear pore complex. Annu. Rev. Cell Biol. 8,495-527. Goldman, R., Goldman, K., Green, K., Jones, J., Jones, S., & Yang, H. (1986). Intermediate filament networks: organization and possible functions of a diverse group of cytoskeletal elements. J. Cell Sci. 5, 69-97.

Nuclear Transport and Cellular Activity

257

Govind, S. & Steward, R. (1991). Dorsoventral pattern formation in Drosophila: signal transduction and nuclear targeting. Trends Genet. 7, 119-125. Greber, U. & Gerace, L. (1992). Nuclear protein import is inhibited by an antibody to a lumenal epitope of a nuclear pore complex glycoprotein. J. Cell Biol. 116,15-30. Grimm, S. & Baeuerle, P. (1993). The inducible transcription factor NF-KB: structurefunction relationship of its protein subunits. Biochem. J. 290, 297-308. Hanson, L. & Ingber, D. (1992). Regulation of nucleocytoplasmic transport by mechanical forces transmitted through the cytoskeleton. In: Nuclear Trafficking (Feldherr, C, ed.), pp. 71-89. Academic Press, San Diego. Herman, B. & Pledger, W. (1985). Platelet-derived growth factor induced alterations in vincuUn and actin distribution in BALB/c-3T3 cells. J. Cell Biol. 100,1031-1040. Hinshaw, J., Carragher, B., & Milligan, R. (1992). Architecture and design of the nuclear pore complex. Cell 69,1133-1141. Ingber, D. & Folkman, J. (1989). Tension and compression as basic determinants of cell form and function: utilization of a cellular tensegrity mechanism. In: Cell Shape: Determinents, Regulation and Regulatory Role (Stein, W., and Bronner, F, eds.), pp. 3-31. Academic Press, San Diego. Ingber, D., Madri, J., & Folkman, J. (1987). Endothelial growth factors and extracellular matrix regulate DNA synthesis through modulation of cell and nuclear expansion. In Vitro Cell. Dev. Biol. 23,387-394. Ireland, G., Dopping-Hepenstal, R, Jordan, R, & O'Neill, C. (1987). Effect of patterned surfaces of adhesive islands on the shape, cytoskeleton, adhesion and behavior of Swiss mouse 3T3 fibroblasts. J. Cell Sci. 8,19-33. Izaurralde, E. & Mattaj, I. (1992). Transport of RNA between nucleus and cytoplasm. Semin. Cell Biol. 3, 279-288. Jans, D., Ackermann, M., Bischoff, J., Beach, D., & Peters, R. (1991). p34^^^^-mediated phosphorylation at T124 inhibits nuclear import of SV-40 T antigen proteins. J. Cell Biol. 115,1203-1212. Johnson, L., Abelson, H., Green, H., & Penman, S. (1974). Changes in RNA in relation to growth of the fibroblast. I. Amounts of mRNA, rRNA, and tRNA in resting and growing cells. Cell 1,95-100. Johnson, L., Levis, R., Abelson, H., Green, H., & Penman, S. (1976). Changes in RNA in relation to growth of the fibroblast. IV. Alterations in the production and processing of mRNA and rRNA in resting and growing cells. J. Cell Biol. 71, 933-938. Kalderon, D., Roberts, B., Richardson, W., & Smith, A. (1984). A short amino acid sequence able to specify nuclear location. Cell 39,499-509. Levine, E., Becker, Y, Boone, C , & Eagle, H. (1965). Contact inhibition, macromolecular synthesis and polyribosomes in cultured human diploid fibroblasts. Proc. Natl. Acad. Sci. USA 53, 350-356. Melchior, F, Paschal, B., Evans, J., & Gerace, L. (1993). Inhibition of nuclear protein import by nonhydrolyzable analogues of GTP and identification of the small GTPase Ran/rC4 as an essential transport factor. J. Cell Biol. 123,1649-1659. Milner, J. & Medcalf, E. (1991). Cotranslation of activated mutant p53 with wild type drives the wild type p53 protein into the mutant conformation. Cell 65,765-774.

258

CARL M. FELDHERR and DEBRA AKIN

Moll, T., Tebb, G., Surana, U., Robitsch, H., & Nasmyth, K. (1991). The role of phosphorylation and the CDC28 protein kinase in cell cycle-regulated nuclear import of the S. cerevisiae transcription factor SWI5. Cell 66, 743-758. Moore, M. & Blobel, G. (1993). The GTP-binding protein RanA^C4 is required for protein import into the nucleus. Nature 365, 661-663. Newmeyer, D. (1993). The nuclear pore complex and nucleocytoplasmic transport. Curr. Opin. Cell Biol. 5, 395-407. Newmeyer, D. & Forbes, D. (1988). Nuclear import can be separated into distinct steps in vitro: nuclear pore binding and translocation. Cell 52, 641-653. Nigg, E., Hilz, H., Eppenberger, H., & Dutley, F. (1985). Rapid and reversible translocation of the catalytic subunit of cAMP-dependent protein kinase type II from the Golgi complex to the nucleus. EMBO J. 4, 2801-2806. Paine, P., Moore, L., & Horowitz, S. (1975). Nuclear envelope permeability. Nature 254, 109-114. Pante, N. & Aebi, U. (1994). Towards understanding the three-dimensional structure of the nuclear pore complex at the molecular level. Curr. Opin. Struct. Biol. 4, 187-196. Peters, R. (1986). Fluorescence microphotolysis to measure nucleocytoplasmic transport and intracellular mobility. Biochim. Biophys. Acta 864,305-359. Richardson, W., Mills, A., Dilworth, S., Laskey, R., & Dingwall, C. (1988). Nuclear protein migration involves two steps: rapid binding at the nuclear envelope followed by slower translocation through nuclear pores. Cell 52, 655-664. Richter, J. & Standiford, D. (1992). StrucUire and regulation of nuclear localization signals. In: Nuclear Trafficking (Feldherr, C , ed.), pp. 90-121. Academic Press, San Diego. Rihs, H. & Peters, R. (1989). Nuclear transport kinetics depend on phosphorylation-sitecontaining sequences flanking the karyophilic signal of the simian virus 40 T antigen. EMBO J. 8, 1479-1484. Rihs, H., Jans, D., Fan, H., & Peters, R. (1991). The rate of nuclear cytoplasmic protein transport is determined by the casein kinase II site flanking the nuclear localization sequence of the SV40 T-antigen. EMBO J. 10, 633-639. Robbins, J., Dilworth, S., Laskey, R., & Dingwall, C. (1991). Two interdependent basic domains in nucleoplasmin nuclear targeting sequence: identification of a class of bipartite nuclear targeting sequences. Cell 64, 615-623. Ross, R., Raines, E., & Bowen-Pope, D. (1987). The biology of platelet derived growth factor. Cell 46,155-169. Schmitz, M., Henkel, T, & Baeuerle, P. (1991). Proteins controlling the nuclear uptake of NF-KB, Rel and dorsal. Trends Cell Biol. 1, 130-137. Starr, C. & Hanover, J. (1992). Structure and function of nuclear pore glycoproteins. In: Nuclear Trafficking (Feldherr, C, ed.), pp. 175-202. Academic Press, San Diego. Steme-Marr, R., Blevitt, J., & Gerace, L. (1992). O-linked glycoproteins of the nuclear pore complex interact with a cytosolic factor required for nuclear protein import. J. Cell Biol. 116,271-280. Wente, S., Rout, M., & Blobel, G. (1992). A new family of yeast nuclear pore complex proteins. J. Cell Biol. 119, 705-723.

Nuclear Transport and Cellular Activity

2 59

Yang, H. & Pardee, A. (1986). Insulin-like growth factor regulation of transcription and replicating enzyme induction necessary for DNA synthesis. J. Cell Physiol. 127, 410-416.

This Page Intentionally Left Blank

INDEX

ABC proteins, 58-96 {see also "ATP binding cassette...") list of, 60-61 ABC transporters in yeast, 57-96 a-factor pheromone, export of, 6577 ABC-domain, 72 CAAX-box, 65-74 chimeric fusion proteins, 69-74 cytokines, 69 efflux pump, 73 ER-Golgi secretory pathway, 67 farnesyltransferase, 67 Golgi-localization, 76 heterodimers, 76 hydrophilic loops, 72 lipopeptide, 65 MATa2,15 mating pheromones a-factor and a-factor, 65-77 "molecular vacuum cleaner," 73 multidrug resistance phenomena, 77 P-glycoproteins, mammalian, 77 peptide antigens Tapl and Tap2, 76 peptide hormones, 65 pheromone secretion signal, 6774 pheromones, 67-74 261

phospholipid bilayers, 73 plasma membrane receptors, 74 polar localization of Ste6, 75-76 precursors, 65 prenylated proteins, 65, 74 proteasome, 75 resistance phenomena, 77 Saccharomyces cerevisiae, 65-77 "schmoo," 74 secretion signal driving a-factor export, 67-74 Ste6, 65-77 Ste6 a-factor transporter, biosynthesis and turnover of, 74-77 vesicular staining, 76 ABC proteins, yeast, family of, 5964 ALDp, 63 Atml, 63-64 chloroplasts, 63 DHS, 62 HMR, 62 mitochondria, 63-64 MRD, 62 PALI, 63 Pmp70, 63 puromycin, 64 SSHl and SSH2, 62-63 Ste6, 59-77 YKL741, 63

262

ABC-transporters as mediators for multidrug and heavy metal resistance, 77-87 ABC-motif, 80 Candida albicans, 82-83 Cdrl, 82-83 CFTR, 78 cytoplasmic proteasome, 83 dexamethasone, 84 Dos! and Dos2, 81-82 drug substrates, 82 effluxing ligand, 84 endocytosis, 84' facilitator superfamily, 86 Hba2, 78 HMR phenomena, 87 Hmtl,78 LEML 84 major facilitator superfamily, 86 MDR phenotype, 79, 82 Mdr2, 84 membrane fluidity, 85 multidrug transporters, biosynthesis and turnover of, 83-87 non-drug substrates, 82 PDR5. 82 pdrl-3 mutants, 85 permeability, 85 Pmdl,78 sec mutants, 86 SNQ2 gene, 79-87 sporidesmin, 79 squelching, 85 steroid signal transduction pathway, 84 steroid transport, 85 sterol content, 85-86 stress response factors, 79 STSl. 79-87 transcriptional interference, 85 transport of steroids and phospholipids, 86 vesicles, 86

INDEX

Walker A and Walker B, 80 Yap 1,79 YCFl, 78 YDRl, 82 Yef3 elongation factor, 77-79 conclusion and perspectives, 87-88 chimeric transporters, construction of, 87 endogenous ABC proteins, 87 functional reconstitution, 87 heterologous ABC proteins, 87 in human disease, 88 mammalian, 87 number of transporters in yeast, 88 purification, 87 introduction, 58-59 a-factor transporter, 59 ABC transporters, 58-96 adrenoleukodystrophy, 59, 6263 cystic fibrosis transmembrane conductance regulator (CFTR), 58 metal resistance, 58-59 multidrug resistance (MDR) phenotype, 58 P-glycoproteins, 58 peroxisomal diseases, 59 pheromone export, 59 TMS, 58 tumors, 58 Zellweger syndrome, 59, 62-63 multidrug and heavy metal resistance mediated by ABCtransporters, 77-87 {see also ".. .ABC-transporters...") Adrenoleukodystrophy, 59, 62-63 Aspartic acid, 186-187 ATPases, model for integrating into endoplasmic reticulum {see ^/5(9"P-type...")

Index

ATPase activity, 1-28 (see also "Molecular chaperones...") role of in protein translocation, 23-24 pBluescript II KS-f (Stratagene), 221 Caveolae as portals for transmembrane signaling and cellular transport, 111-121 concluding remarks, 118 human disease, role in, 117-118 cholera, 117 dystrophin, 117 insuling receptor, 117 malaria, 118 scrapie, 118 Schuffner's dots, 118 whooping cough, 117 introduction, 111-113 at plasma membrane, 111-112 purified caveolae, 112 in signaling events, 113 transcytosis, 112 as signaling organelles, 113-115 calcium homeostasis, 113 caveolin, 113-115 functions of caveolar membrane domains, 114 G-protein receptor signaling, 113 Src family of nonreceptor tyrosine kinases, 113 v-Src substrate, 113-115 transport processes, 115-117 in endothelial cells, 115-116 GPI-linked molecules, 116 potocytosis, 116 Caveolar clustering and GPI-linked proteins, 97-110 (see also "GPI-linked proteins...") CcN motif, 161-199 (see also "Phosphorylation-mediated regulation...")

263

CDNs, 162, 1^9-190 (see also "Phosphorylation-mediated regulation...") Cellular activity, nuclear transport as function of, 237-259 conclusions, 254-255 docking protein, 255 nuclear efflux, 255 receptor-substrate complex, 254-255 introduction, 237-238 macromolecular transport across nuclear envelope, 238-240 cytoplasmic transport factors, 239 nuclear localization signals (NLS), 238 (see also "Nuclear localization...") through nuclear pores, 238 nucleoporins, 239 passive diffusion, 238 signal-mediated transport, 238 regulation of transport across nuclear envelope, 240-254 alterations in transport machinery, 241 in BALB/c 3T3 cells, 246-252 cell fusion experiments, 244-245 cell shape, effects of, 245-246 changes in transport capacity, 241-254 cytoplasmic factor, increased transport and, 252-254 enhancing factors, 252 epidermal growth factor (EGF), effects of, 243 IGF-1, effects of, 243 NP-gold, investigation of, 241254 p53, effect of on NP-gold import, 248-252 PDGF, effects of, 243 phosphorylation, 240

264

signal-mediated transport during cell cycle, 254 of specific proteins, 240-241 SV40-transformed cells, increased transport capacity in, 246-252 transport inhibitors, 240 CFTR, 58 (see also "ATP binding cassette...") Chaperones, molecular, 1-28 (see also "Molecular chaperones...") CLSM, 165-166 Confocal laser scanning microscopy, 165-166 Cystic fibrosis transmembrane conductance regulator (CFTR), 58 (see also "ATP binding cassette...") Duchenne's muscular dystrophy, 117 {see also "Caveolae...") Escherichia coli, membrane protein topogenesis in, 201-214 (see also "Membrane protein...") Glassfog, 221 Glassmilk, 221 GPI-linked proteins, caveolar clustering and, 97-110 caveolar clustering, of, 103-104 caveolin hetero-oligomers, 104, 105 cholesterol-dependent, 103 plasmalemmal vesicles, 103, 106 Triton insolubility, 104 concluding remarks, 106 epithelia, polarized sorting in, 100103 apical plasma membrane domain, 100

INDEX

basolatera plasma membrane domain, 100 FRT cells, 102 glycosphingolipids, 101, 102 luminal plasma membrane domain, 100 MDCK cells, 102-103 plasma membrane domains, 100 serosal plasma membrane domain, 100 trans-Golgi network (TGN), 101 transfected GPI-linked proteins, 101 zonula occludens, 100 GPI biosynthesis and transfer to protein, 99-100 signal for GPI attachment, 100 introduction, 98-99 abbreviations, 99 glypiation, 99 GPI biosynthesis, 98, 99-100 GPI-synthesis pathway, 98, 106 "greasy foot," 99 inhibitors, 98-99 mutants, 99 PIG tailing, 99 unifying features, 99 model for clustering and assembly, 104-106 hetero-trimeric G-proteins, 106 lipid-modified signaling molecules, 106 protein exclusion-lipid inclusion, 106 tyrosine kinase, 106 Heat shock protein, 34, 182 hsp 70 proteins, 1-28 (see also "Molecular chaperones...") mitochondrial, 5-23 Intermembrane space intermediate, 11

Index

KAR2 alleles, 24 Membrane protein topogenesis in Escherichia coh\ 201-214 artificial, 210-211 conclusions and outlook, 211 helix-helix packing in lipid environment, 208 cysteine residues, 208 energy transfer measurements, 208 glycophorin A homo-dimer, 208 introduction, 201-202 mechanisms of assembly of, 202207 CCCP, 207 "helical hairpins," 204-205 membrane potential, 206-207 peptidase (Lep), 203-204 "positive inside" rule, 202-205 sec machinery, 205-206 topology and structure, prediction of, 209-210 for glycophorin A dimer, 210 hydrophobicity plot, 209 Mitoplasts, 12 Molecular chaperones, role of in protein transport across membranes, 1-28 ATP, role of, 23-24 cytosolic hsp 70, role of, 4-5 SSA proteins, 4-5 endoplasmic reticulum, role of hsp70 of, 24-25 KAR2 allleles, 24 introduction, 2-4 ATPase activity, 2 genetic approach, 4 hsp 70 structure and biochemical properties, 2-3 import into mitochondria, basics of, 4

265

Saccharomyces cerevisiae, hsp 70s of, 3 X-ray crystallography, 2 mitochondrial hsp 70, role of, 5-23 Z?2(167)-DHFR, 15-22 co-immunoprecipitation experiments, 8 components, other, required for protein movement, 22-23 "conservative sorting" model, 22 hemin, 17-18 hsp 70 action in translocation across mitochondrial membranes, model of, 18-20 intermembrane space intermediate, 11 matrix, required for translocation into, 7-14 matrix, translocation of proteins into compared to intermembrane space, 20-22 mitoplasts, 12 mutant, in vivo analysis of, 5-7 and precursor proteins, 8-9 Sscl'2, 5-14 Sscl-3, isolation and analysis of, 9-12 Sscl-2 and Sscl-3 alleles, comparison of, 12-14 Ssclp transport activity, precursor conformation and, 14-18 "stop-transfer" hypothesis, 22 Su9-DHFR, 8-18 summary, 25 NLS-binding proteins (NLSBPs), 33-34, 164 Nuclear envelope lattice (NEL), 31 Nuclear localization signal (NLS), 30, 126-127, 162, 163-164, 238 bipartite, 164, 238 NLS-binding proteins (NLSBPs), 164

266

nucleoplasmin simple, 238 SV40 large tumor antigen, 164, 238 Nuclear pore complex (NPCs) in yeast, 29-56 components, 36-39 coiled-coil interactions, 38, 39 FSFG, 37, 38 GFSFG pentapeptide, 37 GLFG, 38, 45 heptad repeats, 38 immunological approach, 37 Mabl92, 37-38,42 NSP1,37, 38, 39-49 nuclear pore biogenesis, 38 NUP1,37 NUP2, 37 NUP49, 37-38, 42 NUPlOO, 37-38 NUPl 16, 37-38 repeat sequences, 37 conclusions, 49-51 experimental approaches to study of, 39-49 affinity-purification, 40-41 biochemical, 39-42 fluorescence microscopy, use of, 48 genetic, 43-47 GLFG, 45 Mabl92,42 MCM 1,47-48 NIC96, 42, 46 NPL3, 49 NSPl, concentration on, 39-49 NSP49, 42, 45-47 NSPl 16, 45-47 NSP-X, 43-45 NUPl, 46 NUP2, 46 NUP49, 42, 46, 49 NUP116,46

INDEX

p54, 42 protein A fusion proteins, 40 red/ white colony sectoring, 43-45 SDS-PAGE analysis, 41 STE12, 47-48 synthetic lethality, 43-47 thermosensitive mutation, 49 in vitro nuclear transport assays, 47-49 Z domain, 40 introduction, 30-35 digitonin, 33-34 function of, 30 GTP-binding protein, 34 heat shock protein, 34 in higher eukaryotes, structural analysis of, 31 NLS-binding proteins (NBPs), 33-34 nuclear envelope lattice (NEL), 31 nuclear localization signal (NLS), 30 nuclear pore proteins, 32 nuclear transport and in vitro systems, 33-35 nucleoporins, 32-33 NUP153, 33 permeabilized HeLa cells, 34 "plug-spoke" complex, 31 Ran/TC4, 34 Saccharomyces cerevisiae, 35-51 in vertebrates, components of, 32-33 wheat germ agglutinin (WGA), 32-33 in Xenopus oocyte nuclear envelopes, 31, 32 as molecular sieve, 163 structure, 35-36 closed mitosis, 35 nuclear lamina, 36 summary, 51 Nucleoporins, 32-33, 239

Index

P-type ATPases, model for integrating into endoplasmic reticulum, 215-235 experimental design, 218-226 amino acid sequences used in construction of fusion proteins, 222 fusion proteins, construction of, 220-226 integration signal, 226 membrane selectivity, question of, 218 polymerase chain reaction (PCR), use of, 219 proteinaceous effectors, 218 sequential insertion model, 218219 "start-transfer signal," 218-219, 223 "stop-transfer signal," 219, 223 integration mechanism, general, for P-type ATPases and porters, 231-232 transmembrane segments, pairs of, role of, 231-232 integration model of Neurospora plasma membrane H^ATPase, 226-231 for enzymatic reactions, 226 M7, M8 as exception, 230 for receptor binding, 226 introduction, 215-216 H^ATPase, 215 porters, 216 Neurospora H^ATPase, 216-218 concanavalin A method, 216 electrogenic ATPase as proton pump, 216 models, 217 molecular mechanism, 217 vanadate, 216

267

porters, integration mechanism for, 231-232 {see also ".. .integration mechanism...") Perkin Elmer's Gene Amp kit, 221 Phosphorylation-mediated regulation of signal-dependent nuclear protein transport, 161-199 conclusion and future prospects, 190-191 eukaryotic cell, nuclearcytoplasmic processes in, 163-164 {see also ".. .nuclear-cytoplasmic...") future prospects, 190-191 gene therapy, potential for, 191 introduction, 162-163 casein kinase II, 162, 173 CcN motif, 162 cell-cycle-dependent NLSs (CDNs), 162 cyclin-dependent kinase, 162, 173 cytoplasmic retention factors, 162-162 intra-and intermolecular NLS masking, 163 NLS masking, 163 nuclear localization signal (NLS), 162 signal transduction-responsive NLSs (SRNs), 162 SV40 large tumor antigen, 162 mechanisms of regulation of, 180187 aspartic acid, 186-187 c-fos proto-oncogene, 183 cAMP-dependent protein kinase catalytic and regulatory subunits, 182-183 CKII phosphorylation, 186-187 cofilin, 185-186

268

cytoplasmic retention factors, 180-183 dual phosphorylation, positive and negative regulation by, 186-187 glucocorticoid receptor and HSP90, 182 intermolecular masking, 184185 intramolecular masking, 184 lamin B2, 185 NF-(kappa)B, 184-185 NLS masking, 183-186 NLS masking by phosphorylation, 185-186 NLSBP interraction, 187 PK-A R-subunit, 182-183 rel/dorsal family, 180-182 SV large T antigen, 183 SW15. 185 methodological considerations in analysis, 165-168 confocal laser scanning microscopy (CLSM), 165-166 stages, two, 168 in vitro measurement, 166-168 in vivo measurement, 165-166 nuclear-cytoplasmic processes in eukaryotic cell, 163-164 bipartite NLS, 164 NLS-binding proteins (NLSBPs), 164 nuclear localization signals, 163164 nuclear pore complex as molecular sieve, 163 SV40 large tumor antigen, 164 regulated (conditional) nuclear entry, 168-172 SV40 large tumor antigen and CcN motif, 172-180, 181

INDEX

variants on CcN motif, 187-190 cdks, 189-190 CDNs, 189-190 cell-cycle-dependent NLSs, 189190 CKII, 189 PK-A, 188 PK-C, 188-189 rNFIL-6, 188 signal-transduction-responsive NLSs, 187-189 SRNs, 189 Plasmalemmal vesicles, 111-121 {see also "Caveolae...") Polymerase chain reaction (PCR), 219 "Positive inside" rule, 202-205 {see also "Membrane protein...") PrimerErase Quick, 221 Saccharomyces cerevisiae, 35-51 {see also "Nuclear pore complex...") hsp 70s of, 3 {see also "Molecular chaperones...") Schuffner's dots, 118 SDS-PAGE,41 Sec machinery, 205-206 Squelching, 85 SRNs, 162 {see also "Phosphorylation-mediated regulation...") SSA proteins, 4-7 {see also "Molecular chaperones...") Sscl'2 mutant, 5-14 {see also "Molecular chaperones...") Sscl'3, 9-12 {see also "Molecular chaperones...") "Sto-transfer" model, 22 Stratagene, 221 Su9-DHFR, 8-18 {see also "Molecular chaperones...")

Index

Thermal cycler, 221 Trans-Golgi network (TGN), 101, 111-121 Transcriptional interference, 85 U snRNPs, nuclear transport of, 123-159 cap binding activity, nuclear, in U snRNA export, 140-142 cap binding proteins, (CBPs), 141 cap structure, role of in U snRNA export, 139-140 cap structures, role of in signaling RNA subcellular localization, 147-148 import pathways, multiple, 152153 inhibitors, 148-150 anti-gp210 monoclonal, 150 anti-nucleoporin antibodies, 149-150 ATP depletion, 148 chilling, 148-149 lectins, 149 nucleoporins, 149 saturation kinetics, 149 wheat germ agglutinin (WGA), 149-150 introduction, 124-126 classes, two, 126 nucleolar (sno), 126 nucleoplasmic (sn), 126 "spliceosomal," 126, 132 macromolecular assemblies, nuclear import of, 128-132 gag-matrix protein, 131 HlV-1, 131 hsc70, 129 NLS, 129-131 nucleoplasmin, 130 in oligomeric complexes, 129 U snRNA, passage of, 129

269 nuclear export of, 138-139 cap structure, 139 nuclear import of, 142-144 model for role of trimethyl cap and core domain in signaling, 146-147 Sm core domain, 146 nuclear import of U6 snRNP after mitosis, 147 nuclear targeting pathways, multiple, in import, 151-152 proteins, nuclear import of, 126128 ATP-dependent translocation stage, 127-128 hsc70, 128 karyophile, 127-128 NLS receptors, 127-128 nuclear localization signal (NLS), 126-127 nuclear pore complex, 127 spliceosomal, 132-138 assembly pathway of, 137-138 composition and function of, 132 core proteins, 133-135 monomethyl cap structure, 137 protein components, 133-137 proteins, common, 133-135 proteins, specific, 135-137 RNA constituents of, 132-133 Sm binding site, 134-135 Sm core domain, 135, 146 Sm proteins, 133, 142-144 trimethyl cap, 137 trimethyl cap structure for U snRNA nuclear import, 144-145 model for role of with core domain in signaling nuclear import, 146-147

270

Vanadate, 216 Wheat germ agglutinin (WGA), 3233, 149 X-ray crystallography, 2 Xenopus oocyte nuclear envelope, 31,32 nuclei reconstitution system, 32

INDEX

Yeast, ATP binding cassette transporters in, 57-96 {see also "ATP binding cassette...") Yeast, nuclear pore complex in, 2956 {see also "Nuclear pore complex...") Zellweger syndrome, 59, 62-63