,DVANCES IN
igronomy
V O L U M6 2E
Advisory Board Martin Alexander
Ronald Phillips
Cornell University
University...
37 downloads
1176 Views
17MB Size
Report
This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form
,DVANCES IN
igronomy
V O L U M6 2E
Advisory Board Martin Alexander
Ronald Phillips
Cornell University
University of Minnesota
KennethJ. Frey
Larry P. Wilding
Iowa State University
Texas A&M University
Prepared in cooperation with the
American Society of Agronomy Monographs Committee P. S. Baenziger Jon Bartels Jerry M. Bigham M. B. Kirkham
William T. Frankenberger,Jr., Chairman David H. Kral Dennis E. Rolston Sarah E. Lingle Diane E. Storr Kenneth J. Moore Joseph W. Stucki Gary A. Peterson
DVANCES IN
Edited by
Donald L. Sparks Department of Plant and Soil Sciences University of Delaware Newark, Delaware
ACADEMIC PRESS San Diego London Boston New York
Sydney Tokyo Toronto
This book is printed on acid-free paper. @ Copyright 0 1998 by ACADEMIC PRESS All Rights Reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the Publisher. The appearance of the code at the bottom of the f i s t page of a chapter in this book indicates the Publisher's consent that copies of the chapter may be made for personal or internal use of specific clients. This consent is given on the condition, however, that the copier pay the stated per copy fee through the Copyright Clearance Center, Inc. (222 Rosewood Drive, Danvers, Massachusetts 01923). for copying beyond that permitted by Sections 107 or 108 of the U.S. Copyright Law. This consent does not extend to other kinds of copying, such as copying for general distribution, for advertising or promotional purposes, for creating new collective works, or for resale. Copy fees for pre-1998 chapters are as shown on the title pages. If no fee code appears on the title page, the copy fee is the same as for current chapters. 0065-21 13/98 $25.00
Academic Press a division of Harcourt Brace & Company 525 B Street, Suite 1900, San Diego, California 92101-4495, USA http://www.apnet.com
Academic Press Limited 24-28 Oval Road, London NWI 7DX, UK http://www.hbuk.co.uk/ap/ International Standard Book Number: 0-12-000762-2
PRINTED IN THE U"FD STATES OF AMERICA 97 98 9 9 0 0 01 0 2 B B 9 8 7 6
5
4
3 2 1
Contents CONTRIBUTORS ........................................... PREFACE .................................................
vii ix
USINGATOMICFORCEMICROSCOPY TO STUDY SOIL MINERALREACTIONS Patricia A. Maurice and Steven K. Lower I. Introduction.. ............................................ 11. Fundamentals of AFM. ..................................... 111. Examples of Applications. ................................... IV Some New Frontiers in AFM Research ........................ References ...............................................
1 3 1s 37 40
PLANTGROWTH-REGULATING SUBSTANCES INTHE RHIZOSPHERE: MICROBIAL PRODUCTION AND FUNCTIONS Muhammad Arshad and William T. Frankenberger, Jr. I. Rhizosphere as a Site of Plant-Microbe Interactions. . . . . . . . . . . . . . II. Plant Growth-Regulating Substances. .........................
HI. Sources of PGRs .......................................... n! Biochemistry of Microbial Production of PGRs . . . . . . . . . . . . . . . . . v. Production of PGRs by Rhizosphere Microorganisms ............ VI. Metabolism of PGRs in Soil ................................. VII. Ecological Significance of PGRs Produced in the Rhizosphere . . . . . VIII. Conclusions .............................................. Appendix: Abbreviations .................................... References ...............................................
46 46 50 51 68 104 110 12 1 123 125
LONG-TERM TRENDS OF CORN YIELDAND SOILORGANIC MATTERINDIFFERENT CROPSEQUENCES AND SOIL TREATMENTS ON THE MORROW PLOTS FERTILITY Susanne Aref and Michelle M. Wander I. Introduction and History of the Morrow Plots . . . . . . . . . . . . . . . . . . 11. CornYield ............................................... V
153 163
vi
CONTENTS
I11. Soil Variables: Soil Organic Matter. pH. P. and K . . . . . . . . . . . . . . . . N. Conclusions: Lessons from the Morrow Plots . . . . . . . . . . . . . . . . . . . Appendix: Abbreviations .................................... References ...............................................
181 191 194 19.5
USING GENOTYPE-BY-ENVXRONMENT INTERACTION FOR CROPCULTWAR DEVELOPMENT Manjit S. Kang I. Introduction .............................................. I1. Implications of GE Interaction in Breeding ..................... I11. Causes of GE Interaction ................................... w. Ways of Dealing with GE Interaction ......................... v: Stability Statistics: Concepts and Usefulness .................... VI. How to Exploit or Minimize Interaction ....................... w. Conclusions .............................................. References ...............................................
200 208 212 223 225 234 240 241
MODELING CARBON AND NITROGEN PROCESSES IN SOILS Jean-Alex E. Molina and Pete Smith Introduction and Historical Background ....................... Model Description......................................... Model Validation .......................................... Model Applications ........................................ Conclusions and Future Work ............................... References ...............................................
253 2.56 275 28.5 288 290
INDEX. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
299
I. 11. I11.
N. V.
Contributors Numbers in parentheses indicate the pages an which the authors’cantributians begin
SUSANNE AREF (153), Department of Crop Science, University oflllinois, Urbana, Illinois 61801 M A R S H A D (45), Department of Soil Science, University of Agriculture, Faisalabad, Pakistan WILLIAM T. FRANKENBERGER, JR. (49, Department of Soil and Environmental Sciences, University of California,Riverside, California 92521 MANJIT S. KANG (199), Department of Agronomy, Louisiana Agricultural Experiment Station, Louisiana State UniversityAgricultural Center, Baton Rouge, Louisiana 70803 STEVEN K. LOWER (l), Department of Geology, Kent State University, Kent, Ohio 44242 PATRICIA A. MAURICE (l), Department of Geology, Kent State University, Kent, Ohio 44242 JEAN-ALEX E. MOLINA (2 5 3), Depa-ent o f Soil, Water,and Climate, University of Minnesota, Saint Paul, Minnesota Y5108 PETE SMITH ( 2 5 3), Soil Science Department, IACR-Rothamsted, Harpenden, HertfordsbireALJ 2JQ, United Kingdom MICHELLE M. WANDER (153), Department ofNatural Resources and Environmental Sciences, Universityof Illinois, Urbana,Illinois 61 801
vii
This Page Intentionally Left Blank
Preface Volume 62 continues the tradition of including cutting-edge reviews that have been the hallmark of Advances in Agronomy for the past 48 years. In the most recent Citation Index, Advances in Agronomy continues to maintain a very high ranking among all agricultural publications. I am very pleased to announce that Dr. Ronald L. Phillips, a distinguished plant geneticist at the University of Minnesota, has joined the Advisory Board of Advances in Agronomy. Dr. Phillips replaces Dr. E. J. Kamprath, who recently retired after an illustrious career at North Carolina State University. On behalf of the publisher and other members of the Advisory Board, we thank Dr. Kamprath for his dedicated service and valuable contributions. Chapter 1 in Volume 62 is a comprehensive and thoughtful review on the use of atomic force microscopy (AFM) in studying soil mineral reactions. AFM is an innovative in situ microscopic technique that is revolutionizing the study of reactions at the soil mineral-water interface and is of great interest to many in the field of agronomy. Topics covered in Chapter I include fundamentals of AFM; examples of applications to electric double-layer forces, mineral growth and dissolution, and particulate studies; and some new frontiers in AFM research. Chapter 2 is an outstanding review of plant growth-regulating (PGR) substances in the rhizosphere with emphasis on microbial production and functions. Coverage includes the rhizosphere as a site of plant-microbe interactions; plant growth-regulating substances and their sources; biochemistry of microbial production of PGRs; production of PGRs by rhizosphere microorganisms; metabolism of PGRs in soil; and ecological significance of PGRs produced in the rhizosphere. Chapter 3 is an informative, historical overview of the Morrow Plots at the University of Illinois. Historical trends in corn yield and soil variables (e.g., soil organic matter, pH, phosphorus and potassium contents) are extensively covered. Chapter 4 discusses using genotype-by-environment interaction in crop cultivar development. Implications of genotype-by-environment interaction in breeding, causes of genotypeby-environment interaction and ways to deal with it, stability statistics, and ways to exploit or minimize interaction are all discussed in a clear and comprehensive manner. The final chapter, “Modeling Carbon and Nitrogen Processes in Soils,” provides an historical background on these very important processes. Model description, validation, and applications are discussed, with the latest developments presented. I am grateful for the authors’ thoughtful and contemporary reviews. DONALD L. SPARKS ix
This Page Intentionally Left Blank
USING ATOMICFORCEM~CROSCOPY TO STUDYSOIL REACTIONS Patricia A. Maurice and Steven K. Lower Department of Geology Kent State University Kent, Ohio 44242
I. Introduction 11. Fundamentals of AFM A. Basic Operating Principles B. Tip-Sample Interactions In. Examples of Applications A. Atomic-Scale AFM B. AFM as a Probe of Double-Layer Forces C. In Situ AFM Studies of Mineral Growth and Dissolution D. Ex Situ Studies: Particulate Imaging Iv.Some New Frontiers in AFM Research References
I. INTRODUCTION Within the realm of modern science, many of the most challenging fields of endeavor involve interfuces. Not only are interfaces highly complex environments in and of themselves, they also require interdisciplinary research in a time of increased specialization. Soils are the quintessential interfaces at the earth’s surface, for they are the links connecting the atmosphere, the hydrosphere, the geosphere, and the biosphere. Soil particles, with their high surface-area-to-volume ratios, present practically infinite environments for interaction. Yet, the structures, chemical compositions, and chemical reactivities of soil mineral surfaces remain but poorly understood. Over the past few decades, new techniques and approaches have evolved to probe the complex natures of mineral surfaces and surficial interactions. Sophisticated surface-sensitive technologies, such as X-ray photoelectron spectroscopy (XPS), secondary-ionmass spectroscopy (SIMS), low-energy electron diffraction 1 Adwuncrsin Agronomy, Volume 62 Copyright 0 1998 by Accademic Press. All rights of reproduction in any form reserved. 0065-21 13/98 $25.00
2
PATRICIA A. MAURICE AND STEVEN K. LOWER
(LEED), and Auger electron spectroscopy (AES), become even more complex when they are applied not to pristine single crystals but to the “dirty” real world of soil environments. Whenever a new technique is applied to problems of soilmineral reactivity, it is essential for the soil science community to conduct critical evaluations focused on the unique requirements of actual soils. One of the most exciting new interface techniques applied by soil scientists is atomic force microscopy (AFM; also called scanning force microscopy, SFM). AFM was developed in 1985 (Binnig et al., 1986), and applications by geochemists and soil chemists began to appear in the literature by the early 1990s (e.g., Hartman et al., 1990; Hochella et al., 1990; Gratz et al., 1991; Johnsson et al., 1991). Briefly, AFM works by rastering a sample underneath a sharp tip that is attached to or part of a cantilever. A variety of forces, as described in Section I.B, cause the tip to deflect as different surface features pass beneath it. By monitoring this deflection, a three-dimensional map of the sample surface can be constructed. Although the tip may be used to map out a variety of surficial properties, the most commonly used modes of AFM result in a map of surface microtopography. A number of factors make AFM uniquely applicable to studies of soil surface structure and reactivity: ( I ) when used properly, AFM is for the most part nondestructive; (2) under most operating conditions, micron- to nanometer-scale resolution is easily attainable; (3) under ideal operating conditions, molecular- to atomic-scale resolution may be achieved; (4) surfaces may be imaged in air or immersed in liquids, including aqueous solutions (vacuum AFMs are also available); (5) a nanometer- to micron-scale portion of the surface may be imaged repeatedly such that reaction progress can be monitored; and (6)sample preparation is generally minimal. In most early work, AFM did not appear to produce true atomic-scale images of complex mineral surfaces. However, in 1993, Ohnesorge and Binnig demonstrated that AFM could provide true atomic-scale resolution on the mineral calcite, using small attractive forces on the order of lo-“ N. In addition to atomic-scale imaging, AFM has been used successfully to probe forces at the mineral-water interface (Ducker et al., 1991, 1992; Weisenhorn et al., 1992); to measure directly the kinetics of growth, dissolution, heterogeneous nucleation, and redox processes (e.g., Hillner et al., 1992a,b; Dove and Hochella, 1993; Gratz et al., 1993; Manne etal., 1994; Maurice et al., 1995; Junta-Rosso and Hochella, 1996; Jordan and Rammensee, 1996); to visualize sorption of macromolecular organic substances and hemimicelles (Manne et al., 1994); to image soil aggregates (Maurice, 1996); and to determine clay particle thicknesses and morphology of clay-sized particles (Lindgreen et al., 1991; Maurice et al., 1995; Friedbacher et al., 1991; Blum and Eberl, 1992; Blum, 1994;Nagy, 1994; Zhou et al., manuscript under preparation). Many additional avenues of research certainly remain to be explored. The goal of this chapter is to present a critical review of the state-of-the-art of AFM as applied to research on the structure, chemical composition, and chemical
USING ATOMIC FORCE MICROSCOPY
3
reactivity of soil particle surfaces. First, basic operating principles are reviewed. Second, tip-sample interactions are discussed, including forces between the tip and the sample, special considerations in the new tapping-mode AFM (TMAFM), and tip shape considerations. A variety of common artifacts are discussed, with the dual purpose of alerting new users to potential pitfalls and of enabling nonusers to evaluate more fully AFM results. Third, examples of applications to studies of soil chemistry-particularly soil particle chemistry-are presented. Finally, several “new frontiers” in AFM research are discussed. The goal is to provide the reader with a sense of the ever-increasing capabilities of AFM and to point out some of the potential problems and limitations that need to be recognized, addressed, and eventually overcome.
II. FUNDAMENTALS OF AFM A. BASICOPERATING PRINCIPLES We work primarily with commercial atomic-force microscopes manufactured by Digital Instruments (DI; Nanoscope I1 and Nanoscope 111). The discussion that follows therefore is biased toward the Digital Instruments machines. The discussion should be applicable to most commercial AFMs, although some of the details and terminology may vary, depending on the manufacturer. Detailed reviews of the basic operating principles of AFM were provided by Hochella (1990, 1995), Eggleston (1994), and Maurice (1996). A brief review of the basic operating principles is provided here as background for further discussion of image collection and interpretation. AFM design is elegantly simple (Fig. 1). A sample is mounted on a piezoelectric tube, which allows the sample to be precisely scanned under a sharp tip that is attached to or part of a cantilever. Stable motion on a scale of less than 1.0 A is possible with a well-built piezoelectric scanner. Deflection of the cantilever as the sample is scanned under it can be monitored by a variety of mechanisms, the most common of which is via an “optical lever” consisting of laser light reflected off the end of the cantilever toward a photodiode detector (Fig. I). AFM imaging may be conducted in either the constant-force mode (also known as “height” mode), in which a feedback loop is used to adjust the height of the sample to keep the cantilever deflection constant during sample scanning, or the so-called constantheight mode (also known as “force” mode), in which the response of the z-piezo is more sluggish so that the height of the sample remains more nearly constant, and the cantilever changes its deflection as the sample is scanned under it. The constant-height mode is not appropriate for surfaces with a large amount of topographic relief, because the sample surface may be damaged as the z-piezo fails to
4
PATRICIA A. MAURICE AND STEVEN K. LOWER
Figure 1 Schematic illustration of the atomic force microscope, showing that the essential elements are the tip, the cantilever, the detector (laser-cantilever-photodiode system), the piezoelectric scanner, the computer control, and the sample.
compensate for large changes in relief and the tip drags against relatively large topographic features. As the surface is scanned, the computer keeps track of the position of the x-y scanner, controllable to the Angstrom level and integrates these data with either the z-piezo movement (height mode or constant-force mode) or the cantilever deflection (force mode or constant-height mode). Individual scan lines are compiled into an image. The number of scan lines may vary with machine and application but is commonly 512. Since each scan line contains 5 12 points, a complete image consists of a grid of 5 12 points on a side. Each point represents a unique set of xyz coordinates. One area of the sample surface may be scanned repeatedly, providing a sequential series of images. Scan speed must be varied as a function primarily of image size. In general, larger images (greater than 1 micron on a side) require slower scan speeds (less than 10 Hz) so that the tip may properly track large-scale surface features. Too fast a scan speed may result in smeared-out edges in the scan direction; an example is shown in Fig. 2. Smaller images (less than 1 micron) require increasingly faster scan speeds, primarily to account for thermal and piezoelectric drift. A 1 pm- (or
USING ATOMIC FORCE MICROSCOPY
5
Figure 2 Scanning too fast over sharp features (in this case, submicron-scale hematite particles) can result in “tail of the comet” type structures in the direction of scan. In essence, the signal is smeared by a scan speed that is too fast, resulting in a tip that cannot properly track features. Scan direction from right to left. Image by J. Forsythe (Kent State Univ.). Scan area = 2.00 p n on a side.
larger) scale image often takes approximately 1 to 3 minutes to complete; a nanometer-scale image can be completed in less than a minute. We have found that good TMAFM images of particles (discussed later) take much longer to complete than do contact-mode images. We routinely run micron-scale images at less than or equal to 1 Hz, so that a single image frame can take 10 minutes or longer to complete. Depending upon the size of the particular scan head used, the microscopist may move from one area to another within a small region of space (e.g., a common scan head permits imaging anywhere within an area 16 microns on a side). The microscopist also may zoom in to an area of interest, although thermal and piezoelectric drift and piezoelectric nolinearity may limit precise “zoom” capabilities. This zoom capability does not merely blow up the area of interest; rather, the zoom area is imaged with higher resolution. The computer control must go through a variety of data-processing procedures to compile an image that is true to sample structure and interpretable to the human eye. Blum (1994) discusses in detail planefit corrections for sample tilt (slope) and flattening procedures that are required for proper interpretation of rough surfaces. A variety of real-time high and low pass filters also may be applied. Such filters
6
PATRICIA A. MAURICE AND STEVEN K. LOWER
often are needed to remove vibrational noise during molecular- to atomic-scale imaging in air or solution. However, they need to be used with extreme caution because they do remove some portion of the overall signal (Blum, 1994). Use of real-time filters, and of postprocessing filters, as well, should be reported in publications, and images with and without filtering should be compared. For microtopographic images, we recommend that real-time filters not be used, since they seldom are needed at larger scales and the fact that they remove a portion of the overall signal means that resulting images cannot be used for precise measurements. If real-time filters appear to be needed, the AFM probably needs to be moved to a quieter room or placed in an isolation chamber. Postprocessing application of low-pass filters, of planefit filters to account for sample tilt, and of flatten filters to correct for problems associated with variable topography are common practice but should always be reported. In general, we have found that AFMs work best if they are placed in a room located on bedrock, without windows, far removed from other equipment, and with temperature and humidity control. Basement closets often work well, if they are not too damp. Frequently encountered problems include too much traffic in and around the room; noisy heating and air conditioning; telephone noises; and poor humidity control, which can lead to the tip “sticking” to the surface because of thick layers of adsorbed water or to buildup of static electricity. Room temperatures of approximately 68°C tend to work well; warmer temperatures seem to result in increased thermal drift, whereas cooler temperatures can be uncomfortable when the microscopist sits still for hours on end. Vibrational noise, which can be a major problem for atomic-scale imaging, generally can be corrected for by placing the AFM on a cement block suspended from elastic cords. A variety of isolation boxes also are available from the manufacturer. Eggleston (1994) showed how a ringing telephone can ruin a high-resolution image. We use a telephone with a blinking light and turn the ringer off.
B. TIP-SAMPLE INTERACTIONS 1. Forces between the Tip and the Sample Because AFM works by mechanical interaction between the tip and the sample, it is important to understand the nature and magnitude of tip-sample interaction forces. An excellent review of tip-sample interaction forces was provided by Eggleston (1994); a brief review is provided here. Most AFM imaging to date has been conducted in the so-called repulsive mode, also known as the contact mode, wherein repulsive forces between the tip and the sample dominate. The main repulsive force is responsible for the original name of the instrument. This is the so-called atomic force that occurs between any two
USING ATOMIC FORCE MICROSCOPY
7
given atoms (e.g., on the tip and sample surfaces) when the electron clouds of the atoms begin to overlap. These repulsive forces have often been called Pauli-exclusion forces. Repulsive forces also may arise from solvation or hydration forces that occur because water near hydrophilic surfaces is structured (Israelachvili, 1992). When the tip and the sample are brought into close contact during atomicforce microscopy, resistance occurs; hence apparent repulsion arises as the structured water molecules on the surfaces of the tip and the sample are pushed away. Van der Waals (VdW) interaction forces also need to be considered. As described by Eggleston (1994), VdW forces are long-range, relatively weak forces that are generally attractive but can become repulsive in some media. Repulsive or attractive electrostatic forces also may come into play. As described in Section III.B, surface scientists are using AFM as a probe of double-layer forces by systematically varying tip and sample materials and solution conditions. During imaging in air, capillary adhesion, which is a relatively strong attractive force, can become important. This capillary force results from formation of a meniscus made up of water and adventitious organic contaminants sorbed on the surface of the tip and the sample (Weisenhorn et al., 1992). The capillary force is large and has been estimated to be on the order of N or greater (Weisenhorn et al., 1989). When the tip and the sample are completely immersed in water or another liquid, a meniscus does not form and hence the capillary forces are absent. Due to limitations imposed by the capillary forces, the minimum force that can be achieved by AFM in air tends to be on the order of lop7 N. By working in solution, the overall tracking forces can be reduced by two to three orders of magnitude. Weihs et al. (1991) noted that the adhesive force between the tip and the sample decreased with decreasing tip radius. Capillary forces were reduced for samples in vacuum or in low-humidity environments (Thundat, 1993). Lateral frictional forces also must be taken into account as the sample is rastered beneath the tip. Frictional forces have been shown to vary on an atomic scale and with temperature, scan velocity, relative humidity, and tip and sample materials (Delawski and Parkinson, 1992; Overney eta]., 1992). Scan direction also can be important. Frictional forces tend to be greatest when scanning parallel to the long axis of the tip-cantilever system (0" on Nanoscopes), due to flexure of the cantilever. Rotating scan direction to be perpendicular to the long axis (90") often reduces lateral frictional forces and can be helpful in imaging particles that otherwise tend to be plucked from substrate surfaces. As opposed to contact-mode, TMAFM (DI, 1995; Zhong et al., 1993; Prater et al., 1995) is a relatively new technique that allows high-resolution topographic imaging of soft, adhesive, or fragile samples because it overcomes problems associated with friction, adhesion, and electrostatic forces (Prater et al., 1995). In T M A W , a piezoelectric driver is used to excite the cantilever into resonance oscillation. The tip is thus caused to vibrate and to contact the sample surface numerous times for each data point. When imaging in air, the cantilever oscillation
8
PATRICIA A. MAURICE AND STEVEN K. LOWER
is damped when the tip contacts the water layer and the sample surface, but the large vibration amplitude gives the cantilever sufficient energy to overcome the surface tension of the adsorbed water layer. TMAFM uses the root-mean-squared (RMS) of the cantilever deflection feedback symbol to keep the cantilever vibration amplitude constant by adjusting the piezo height. The change of voltage applied to the z piezo reflects the topography of the sample (Digital Instruments AFM Manual, 1993). The force between tip and sample is only to lop9 N. TMAFM has a large, linear operating range that makes for highly stable vertical feedback, allowing routine reproducible sample measurements (Prater et al., 1995). Although originally TMAFM could not be performed in solution, a new TMAFM fluid cell has been developed. TMAFM does not appear to be capable of producing atomic-scale images on most surfaces. However, TMAFM is the technique of choice for imaging rough andlor easily deformable surfaces at scan sizes on the order of a few tens of nanometers and larger.
2. Tip Size and Shape Considerations Finite tip size and shape are responsible for many of the major artifacts in AFM images. To a certain degree, the resolution of an image is dependent on the quality of the tip. Blum (1994) and Maurice (1996) have reviewed imaging artifacts caused by different tip shapes in imaging environmentalparticles. Eggleston (1994) summarized tip-sample interactions and resulting artifacts on nm-scale images. Typical contact-mode imaging, including atomic-resolution imaging, is generally conducted using standard silicon nitride (Si,N,) probes that are integrated tips-cantilevers. These probes may also be used for TMAFM in solution, although we have found that specialized force modulation etched silicon probes tend to work better. As supplied (from DI), each Si,N, probe contains two cantilever lengths and two widths, which are usually referred to as thick- and thin-legged. Thus, four different cantilever geometries result, with four different possible force (spring) constants. Different geometries may work better for atomic-scale versus micron-scale imaging or for different surfaces. Hence, it is worthwhile to try different cantilever geometries and see which works best for a particular application. A fall, 1996, DI Web page (http://www.di.com)gives the following specifications for standard Si,N, probes: Force (or spring) constants: 0.58,0.32,0.12,0.06 N/m* Nominal tip radius of curvature: 20-60 nm Cantilever lengths: 100 and 200 pm Cantilever configuration: V-shaped Reflective coating: gold Shape of tip: square pyramidal Tip half angle: 35" (*Actual values can vary substantially.)
USING ATOMIC FORCE MICROSCOPY
9
For tapping mode, etched silicon probes generally are used. Each probe contains only one integrated tip-cantilever. DI reports the following specifications for etched silicon tapping-mode probes: Force (or spring) constants: 20-100 N/m Resonant frequency: 200-400 kHz Nominal tip radius of curvature: 5-1Onm Cantilever length: 125 p,m Cantilever configuration: single beam Reflective coating: uncoated Tip half angle: 17” side, 25’ front, 10” back B a r r h et al. (1997, in press) report that these probes have a solid angle of between 20 and 50”, but that this solid angle decreases to about 20” within 200 nms of the end of the probe. Blum (1994) reported that for vertical features less than -30 nm, such as lowrelief steps, the radius of curvature of the tip limits the resolution. This problem can make low-relief features and small particles either appear to be too broad or cause them to disappear altogether (Barrett, 1991; Gnffith and Grigg, 1993). Wilson et d.(1996) demonstrated that contact-mode AFM images of biomolecules and other structures on the order of 10s of nanometers in height often are enlarged by as much as 25% due to the finite size of the tip. Note that the nominal radius of curvature for Si,N, tips is greater than for etched silicon tapping-mode tips. These are “nominal” radii, because the radius of curvature may vary depending upon defects in manufacturing, and because tips tend to abrade with use. Although the nominal radius is less for tapping-mode tips, this does not result in higher resolution at the near-atomic scale. At present, the tapping procedure itself appears to limit resolution to the scale of 10s to 100s of nanometers. Blum (1994) also reported that for imaging larger-scale features (vertical features greater than -30 nm), it is the shape of the tip (tip half angle) that limits resolution. For large-scale features, tip shape can “convolve” with surface features, often producing misleading results (the quotation marks are used because this is not a true convolution, since it is nonlinear). When the tip rides over a sharp feature on a sample surface, the sides of the tip often contact the edges of features before the apex of the tip comes into contact. The solid angle (pyramidal) of a standard Si,N, tip is --55”, and according to Blum (1994), the steepest vertical features that can be accurately imaged are -62.5’. Note that tapping-mode tips have asymmetrical tip half angles. Hence, different scan angles may result in notably different resolution on topographic features. Typical tapping-mode tips are sharper overall than contact-mode tips, and we have found that resolution of microtopographic features and submicron-scale particles tends to be better using tapping mode. Often, tip-related artifacts can be difficult to identify as such; hence, the expe-
10
PATRICIA A. MAURICE AND STEVEN K. LOWER
rience of the microscopist becomes invaluable. Excellent reviews, such as Griffith and Grigg (1993), are helping to alert microscopists to problems, limitations, and potential solutions. Once a particular artifact, such as a reverse-tip image (discussed later), is properly identified, the microscopist is alerted that similar structures imaged on other samples may also be artifacts. In the extreme case of very sharp features on sample surfaces, a reverse image of the tip may result (Oden et al., 1992). The process controlling such reverse-tip images is illustrated in Fig. 3(a). Such reverse-tip images show structures related to the tip shape, primarily pyramidal features with triangular facets. Convolution may result in an entire reverse-tip image, or more in subtle effects along step or particle edges, or wherever a sample surface feature is steeper than the tip structure. This problem can be particularly frustrating in the imaging of particles, as shown in Figs. 3(b) and 3(c). Figure 3(b) shows “convolution” along the termination of a needle of hydroxypyromorphite (HPY). Apparently, the end of the nee-
Figure 3 (a) Schematic illustration of how tip shape can “convolve” with sample shape if surface features are sharper than the tip. (b) An example of tip-sample “convolution.” Here, the end of a hydroxypyrornorphite (HPY) needle is probably oriented perpendicular to the substrate, resulting in convolution with the tip (see arrow). Scan area = 1.20 p m on a side. (c) Reverse-tip images that occurred upon imaging submicronsized particles of hydroxylapatite (HAP). Numerous reverse-tip images are scattered across the image. Scan area = 3.21 p n on a side. TMAFM images.
USING ATOMIC FORCE MICROSCOPY
11
dle was sticking up from the surface. Figure 3(c) occurred while imaging submicron-size particles of hydroxylapatite (HAP) and contains reverse-tip images. In Figs. 4(a) and 4(b), AFM images of hematite particles (a) are compared with TEM images of the same sample (b). Convolution with tip shape tends to obscure the
Figure 4 TMAFM images in air of hematite particles deposited on pc membrane filters (a) are compared with a TEM image of a different subsample from the same sample (b). The left-hand image in (a) shows height data; the right-hand image shows amplitude data. Pores are apparent on the filterpaper substrate in the AFM images. Scan area of each image is 2.35 pm on a side. Length of scale bar in (b) is 100 nm. “Convolution” of sample and AFM tip shape tends to obscure the hexagonal-to-rhombohedra1 shapes of the particles, which are apparent in the TEM image. Such tip-sample convolution artifacts are common during imaging of particle aggregates, which tend to have rough surfaces. TEM image courtesy S. Traina, Ohio State University.
12
PATRICIA A. MAURICE AND STEVEN K. LOWER
hexagonal to rhombohedra1 shapes of the particles, which are apparent in the TEM image. Such tip-sample convolution artifacts are extremely common during imaging of particle aggregates, which tend to have rough surfaces. Tip-sample convolution also can lead to more subtle artifacts, such as sloping step and particle edges. For example, Fig. 5 (cross section) shows a distorted kaolinite edge surface caused by convolution of the tip shape with the vertical kaolinite edge (Zhou et al., manuscript in preparation). The kaolinite crystal should have a vertical edge. Due to convolution, the edge of the kaolinite particle is a reverse mirror half-tip shape. Note that no distortion occurs on the basal-plane surface (001). This type of artifact does not affect the measurement of particle thickness; particle diameter can be measured from the top of the particle, but not from the bottom. Kepler and Gewirth (1994) noted that different types of tips will result in different shapes of artifacts; for example, they found square pyramidal tip-sample convolution artifacts with a Digital Instruments Nanoscope tip-cantilever but triangular artifacts with a Park Scientific Ultralever tip-cantilever. The Nanoscope Si,N, tips ideally have rectangular pyramidal shapes, whereas the Park Scientific Ultralever tips ideally have a triangular shape. Finally, frictional forces between tip and sample may result in frictional tip-sample convolution artifacts; such artifacts generally have a roughly L-shaped outline as the tip drags across a sharp surface feature. Tip dimensions also can limit the ability of the tip to image narrow, deep features, such as steep-sided pits. In such a case, the pit will appear to be V-shaped, since the edges of the tip meet the sides of the pit before the tip reaches the bottom. If pit depth exceeds maximum z (height) range, the pit may appear to be too shallow and flat bottomed (off-scale). In attempts to circumvent tip-shape artifacts, tips have been developed with high aspect ratios or smaller radii of curvature (Kado et al., 1992; Keller et al., 1992). Unfortunately, sharp tips tend to be fragile. As an alternative, a number of researchers have proposed methods and algorithms for improving images using deconvolution procedures. As pointed out by Griffith and Grigg (1993), the term deconvolution is not strictly correct because the tip and sample interact in a nonlinear fashion; nevertheless, the term will be used here for simplicity. Some researchers have relied on purely mathematical treatments in which the geometry is assumed to be an idealized shape, such as spherical; however, more precise methods are possible. Markiewicz and Goh (1994) back-calculated the tip shape by imaging samples (polybead-amino microspheres) of known size and geometry, recognizing that the AFM image will be a convolution of tip and sample shapes. Calculated tip shapes were then used to deconvolute additional samples. Wilson et al. (1996) used 14-nm gold spheres in a similar procedure and found good agreement between back-calculated or “restored” tip shapes and actual SEM images of
USING ATOMIC FORCE MICROSCOPY
13
Figure 5 TMAFM image in air of a kaolinite particle on muscovite mica. Cross section shows how the particle edges are distorted due to tip-sample convolution; however, accurate particle diameter and thickness data can be obtained. Vertical distance AB (15.9 nm) is the thickness of the particle; horizontal distance, CD (1 12.5 nm), is the diameter; vertical distance EF (0.22 nm), is the height difference on the mica substrate at points E and F. From Zhou et al., manuscript in preparation.
14
PATRICIA A. MAURICE AND STEVEN K. LOWER
Figure 6 TMAFM image in air of a hydroxylapatite (HAP) surface. The three-pronged “birdsfoot-like’’ features that appear scattered on the surface, all with the same orientation, are not real surface features but rather the result of a “dirty” or imperfect tip. Scan area is 5.00 p,m on a side.
tips. Deconvolution procedures such as these need to be calibrated and standardized and necessary algorithms incorporated into AFh4 software. Irregular tip shapes can pose additional problems. On the atomic scale, multiple (more than single-atom) tips can lead to a confusing array of apparent atomicscale surface structures. The effects of multiple tips on atomic-scale imaging have been the subjects of numerous investigations. Such investigations should continue in light of Ohnesorge and Binnig’s (1993) reevaluation of atomic-scale imaging. At microtopographic scales, tips with irregular structures can lead to equally erroneous and potentially misleading results. This tends to be an especially major problem when imaging soft and deformable surfaces from which material can be scraped up and deposited on the tip. Figure 6 shows a surface that was imaged with a “dirty” tip. Small bird-foot-shaped features appear scattered across the surface, with the same orientation. These features are not part of the surface, but rather the result of convolution between an imperfection or “dirt” on the tip and surface features. Figure 7 shows apparent doubling of particles that results from a double tip.
USING ATOMIC FORCE MICROSCOPY
15
Figure 7 TMAFM image in air of hydroxypyromorphite (HPY) needles deposited on pc membrane filter paper. The apparent doubling of features is the result of a double tip shape. Scan area is 1.10 pm on a side.
m. EXAMPLES OF APPLICATIONS A. ATOMIC-SCALE AFM Mineral surface structure is seldom or perhaps never an exact extension of the bulk; rather, we can expect some relaxation and/or reconstruction to occur at most surfaces. Until recently, mineral surface structure had to be inferred, for the most part, due to a lack of surface structural techniques. Although TEM on microtomed sections may provide surface and near-surface structural data, some sample disturbance may occur, and the sample must be analyzed in vacuum. LEED can give symmetry and spacing information, although the diffraction pattern is a statistical representation that does not work optimally for small particles, such as clays, or for heterogeneous surfaces. The advent of AFM during the mid- to late 1980s raised considerable excitement within the community of scientists studying mineral surfaces and surfacerelated phenomena, because AFM appeared to be a simple and straightforward
16
PATRICIA A. MAURICE AND STEVEN K. LOWER
technique for determining atomic-scale surface structure under environmentally appropriate conditions, e.g., on particles in air or immersed in aqueous solutions. During the late 1980s and early 1990s,a number of articles were published showing the apparent molecular- to atomic-scale structures of several important soil minerals. However, it was quickly noted that these images for the most part showed perfectly ordered periodic arrays, with none of the small-scale defects, such as monoatomic steps and kink sites, commonly present in STM images (e.g., Eggleston and Hochella, 1992). This eventually led many researchers to suspect that, like all techniques, atomic-scale AFM imaging could be subject to a variety of potential pitfalls. In 1993, Ohnesorge and Binnig published a landmark paper that is helping to redefine atomic-scale AFM imaging. These authors argued that most pre- 1993supposed atomic-scale imaging had been conducted with too high a force between the tip and the sample. Given that the typical force constant for a chemical bond is on the order of 100 Nm-', tip-sample interaction forces greater than 100 nN, as commonly encountered in repulsive-mode (contact-mode) imaging in air,would displace the atoms in a bond by nm distances (Eggleston, 1994).Such strong forces would likely break bonds and damage the sample surface. Additionally, strong forces may result in the tip being driven into the sample such that the contact point between the tip and sample consists of several atoms. Images with periodicity related to crystallographic structure can occur at high contact forces, but the images do not show individual atoms, and high forces can distort relative atomic positions. Ohnesorge and Binnig (1993) argued that these problems could be overcome by imaging in solution and in the attractive mode. In a study of calcite structure, they demonstrated that true atomic-scale resolution, including imaging of monoatomic steps, could be achieved only when the estimated net repulsive loading force between the tip and the sample was less than or equal to N. According to Ohnesorge and Binnig (1993), the four keys to true atomic-scale resolution are using sharp tips, imaging in solution, systematically regulating forces, and, when possible, imaging in the attractive-forceregime. Their true atomic resolution images on calcite were obtained with tip-attractive forces of -lo-" N. Images collected in air or with relatively high forces may still be valuable in that the observed structures should be related to crystallographic parameters, even though individual atoms are not accessible. However, previous work on molecular- to atomic-scale imaging of mineral surfaces should be reevaluated and perhaps improved upon with the help of these guidelines. Clear, regular, repeatable atomic-scale images are needed to evaluate atomicscale surface structure. Unfortunately, atomic-scale AFM imaging is at present extremely difficult. Thus, many noisy, irregular images have been published. One common problem is that images often are published containing just one isolated regular unit cell, or several unit cells, surrounded by irregular features that may be
USING ATOMIC FORCE MICROSCOPY
17
either noise or noncrystallinematerial. However, reproducibility both within a single image and from one image area to another is essential. AFM studies of atomic-scale surface structure should include imaging at numerous locations, on a number of samples or subsamples, with a variety of tips, in different media, and with systematic variation of imaging forces. Feedback oscillations with a periodicity that could be confused with atomic-scale structure can occur if the gains are set too high. It is therefore important to confirm that apparent atomic-scale structure is “real” by systematically altering scan rate, scan size, scan angle, scan direction, and gains. Although atomic-scale imaging is time-consuming, such painstaking work is required. Publication of unfiltered images and transform plots is important for proper image interpretation.Two-dimensional fast Fourier-transform filtering can remove noise without introducing extra spots, but unfiltered images are needed to evaluate the reliability of the data. Transform plots succinctly summarize data and are useful for determining artifacts such as drift and double-tip effects. Combination of AFM with other techniques such as STM, LEED, XPS, XRD (X-ray diffraction), and TEiM is crucial, especially considering that mineral surfaces tend to be heterogeneous and that structural data may be ambiguous. There is a critical need within the AFM community to address sampling statistics and to set standards for image quality and reproducibility.
B. AFM AS A PROBE OF DOUBLE-LAYER FORCES The use of AFM as a probe of double-layer forces is one of the most exciting developments from the standpoint of mineral-water interface geochemistry. Prior to AFM development,the most widely regarded technique for measurement of surface forces was the Israelachvili-Adamsforce apparatus, which works by measuring the force between two cylindrical macroscopic surfaces. Israelachvili and Adams (1978) measured the forces between two muscovite sheets at close approach, in solutions of different pH and ionic strength. A limitation of their experiments was the use of multiple beam interference to measure surface roughness. Visible light interference probably is not sensitive enough to detect small-scale features of limited lateral extent, e.g., ultrafine particles and small pits. Johnsson ef al. (1992) showed that muscovite may develop small etch pits and other surface features rapidly upon exposure to aqueous solution. Additionally, muscovite is brittle and may form fine fractures upon bending (into cylindrical form). AFM circumvents these problems because surface microtopography may be measured at the nanometer scale, although sample drift makes it difficult to define the exact area of measurements. Butt et al. (1995) recently reviewed AFM force measurements in liquid environments. Tip-sample interaction forces can be measured by recording force curves, wherein the deflection of the cantilever is monitored as the tip approaches
18
PATRICIA A. MAURICE AND STEVEN K. LOWER
the sample, the tip and sample come into contact with one another, and the tip is subsequently retracted. Ducker et al. (199 1, 1992) succeeded in probing the forces between a tip consisting of a silica sphere and a planar silica surface oxidized to a depth of -30 nm, immersed in solutions of different ionic strength. The authors collected force-vs-tip-sample displacementdata in solutions of ionic strength 10-4 to lo-' M.To convert these measurements to force-vs-distancedata, zeros of force and distance had to be defined. The zero of deflection was chosen where the deflection was constant, i.e., with the spherical tip and sample far apart. The zero of distance was defined based on the point at which the photodiode output became a linear function of displacementof the sample, i.e., where the sphericalparticle was in contact with the surface. The resulting force-vs-distance curves generally were in agreement with double-layertheory. However, deviations were observed at very short distances. The authors noted that these deviations could be due to hydration forces, i.e., relatively ordered water bound to the mineral surface, but that the potential effects of surface roughness could complicate the interpretation. Experiments by Ducker et al. (1991, 1992) pioneered the measurement of colloidal forces between colloidal-sizedprobes (tips) of different compositions and various sample surfaces. Butt (1994) extended this technique to measure the forces on a colloidal-sized glass particle as it entered an air bubble or a water droplet. Weisenhorn et al. (1992) recorded force-vs-distancecurves for interactions between a Si3N4tip and mica in KCl solutions at pH -6.6 and ionic strengths of 0.1-30 mM. They found that the force curves showed repulsive behavior in the noncontact regime, presumably due to double-layerforces. The repulsion decayed exponentially with increasing distance, as expected from double-layer theory. Plots of the Debye length (UK)versus inverse square root of the concentration ~ with increased ionic strength, as from 0.1 to 30 mM KCl showed that 1 / decayed expected from double-layer theory; the measured slope of 0.308 compared favorably with the predicted value (0.305). Butt et al. (1991a,b) showed that electrostatic tip-sample forces depend strongly on both pH and salt concentration. They showed that the pH and/or salt concentration may be adjusted so that attractive VdW forces are in effect cancelled out by repulsive electrostatic forces. Thus, careful choice of solution may allow one to minimize potentially damaging adhesive tip-sample interaction forces. Radmacher et al. (1994) used laterally resolved force curves to study the adsorption of organic (lysozyme)molecules on mica. They showed that the adhesion forces between the Si3N, tip and the mica surface were different from the forces between the tip and the adsorbed molecules. These results demonstrate the potential for development of a form of atomic force spectroscopy based on force-curve characteristics (see Section IV). Manne el al. (1994) utilized noncontact-mode AFM, based on double-layer repulsion between tip and sample, to image surfactant hemimicelles on highly oriented pyrolitic graphite (HOPG). Hence, the double layer can be utilized to image delicate structures.
USING ATOMIC FORCE MICROSCOPY
19
C . IN SITU AFM STUDIESOF MINERALGROWTH AND DISSOLUTION 1. Fundamentals of in Situ AF’M Until recently, most studies of mineral-water interface reactions have relied primarily on macroscopic observations, e g , monitoring the concentrations of reactants or reaction products in solution. For example, dissolution experiments traditionally are conducted over a range of saturation states, the concentrations of constituents released to solution are measured at various intervals, and a curve is fit to the data. The “order” of this curve often is taken to give an indirect indication of the reaction mechanism. In reality, simple curve fitting does not provide direct evidence of the type of reaction mechanism and may result in misleading interpretations (e.g., Inskeep and Bloom, 1985; Rimstidt and Dove, 1986; Shiraki and Brantley, 1995).Hence, kinetic studies need to incorporate some means of directly documenting a process as it occurs at the mineral surface. In siru AFM is being used to fill this void in mineral-water interface chemistry by permitting researchers to directly monitor changes in mineral surface microtopography over the course of reaction in aqueous solution, at micron to subnanometer scales. Indeed, the potential exists for watching changes in atomic-scale features, such as movements of monoatomic steps, over the course of reactionalthough this potential is not yet fully realized. By monitoring changes in microtopography in response to varying reaction parameters, such as time, saturation, and pH, researchers can directly support or refute hypotheses developed based on macroscopic observations or modeling. Whereas macroscopic observations give information about overall reactions, AFM provides insight into surface heterogeneity and variations in reaction rates and “mechanisms.” AFM has advantages over other ‘‘insitu” microscopic techniques. Scanning tunneling microscopy (STM) also may be used in solution, but its use is generally limited to conductors or semiconductors,and the tunneling current may promote electrochemical reactions. Environmental SEM (ESEM) may be used on hydrated surfaces but not on surfaces fully immersed in solution; hence, reaction progress cannot be monitored. Dove and Chermak (1994) provide a comprehensive review of the application of in situ AFM to studies of mineral-water interface geochemistry.Herein, we provide a more concise review, focusing on experimental considerations and potential artifacts. AFM offers distinct advantages by providing extremely high-resolution imaging in solution, while generally having little or no effect on reaction progress (Dove and Hochella, 1993). However, the technique is subject to a variety of limitations and pitfalls. First, by its very nature, AFM is a microscopic technique. Bosbach er al. (1996) studied the influence of different electrolyte solutions on the growth ki-
20
PATRICLA A. MAURICE AND STEVEN K. LOWER
netics of gypsum, and they were able to decipher different growth “mechanisms.” They demonstrated that growth rates and “mechanisms” varied on a microtopographic scale, suggesting that bulk growth rates of certain minerals cannot be easily predicted based on microscopic observations. In situ experiments are difficult and time-consuming; hence, it is difficult to obtain a statistically meaningful sampling of surface reactivity. To put microscopicAFM results in context, they always need to be coupled with macroscopic experiments. Indeed, we recommend coupling in situ (in solution) imaging on mm-scale crystals with ex situ imaging of particulates prior to and following reaction, to try to get a handle on the effects of surface heterogeneity (e.g., Maurice et al., 1995). Second, because AFM images are collected as lines of information that eventually form an image, very fast or very slow reactions may be difficult to study in real time. Dove and Platt (1996) estimate that AFM can be used to observe the “real-time” growth and dissolution of monolayer surfaces occumng at rates of between and mol me2s-*. Mineral surfaces with reaction rates outside of this range can be adjusted to fit the AFM-compatible range by carefully controlling pH, saturation, and so on. For example, Bosbach and Rammensee (1994) were able to observe the dissolution of gypsum surfaces by using partially saturated solutons that lowered the reaction affinity, thereby decreasing reaction rates. Reaction rates that are slower than the AFM-compatible range may be increased by adjusting pH or saturation state, although oftentimes these changes are insufficient. At times, it may be easier to rely on ex situ observations of the mineral surface before and after reaction, using a statistical sampling regime. One problem with this technique is that surface microtopography of some minerals may change considerably on drying. In some cases, in situ AFM may be used over a period of hours or days to observe the general progress of a surface reaction, although it is not always possible to image one area for prolonged periods of time. Third, as commonly used, AFM provides structural, micromorphologic, and microtopographic information. Other techniques, such as AES and XPS, must be used to determine surface composition. Fourth, in situ AFM is only useful for imaging processes that occur at the surface of the subject mineral. Processes that occur strictly in solution, such as homogeneous nucleation in solution, cannot be directly accessed by AFM.However, AFM can be used to obtain indirect evidence of homogeneous nucleation (e.g., Lower et al., manuscript in preparation). Fifth, although AFM is for the most part a nondestructive imaging technique, some surface damage can occur. As described in Section 1I.B.1, forces are minimized when working in solution. Nevertheless, damage of the sample surface can occur during contact-mode imaging, especially during small-region scans (<1 pm) and when soft features are present; for example, over the course of dissolution or during growth of an amorphous precipitate. Maurice et al. (1995) presented an example of scan-induced sample erosion during a dissolution experiment. In
USING ATOMIC FORCE MICROSCOPY
21
this experiment, an etch pit was observed to nucleate and grow on the basal-plane surface of hematite immersed in citric acid. The original pit shape was hexagonal. However, near the end of the experiment, the scan area was decreased (from 10 pm on a side to <1 pm) to the region immediately surrounding the pit. During this procedure, the pit edges began to align with the scan direction and the image frame, suggesting that scan-induced erosion occurred during the small-region scans. Scan-induced erosion also was observed during small scan imaging of the growth of “active” Cr-hydroxide precipitate on the surface of hematite (Eggleston et al., manuscript in preparation; Maurice-Johnsson, 1994).Marchant et al. (1991) showed that 1-2 pm scans disturbed the structure of von Willebrand Factor (a blood plasma glycoprotein) sorbed to mica, whereas -8 p m scans did not produce a noticeable disturbance. The reason for the enhanced small-region scan erosion is currently under debate. It has been suggested that enhanced erosion during small-region scans is a result of the rapid scan rate needed to compensate for drift at high resolution (Barrett, 1991). However, slowing scan rate does not always resolve the problem. In situ growth experiments often show isolated holes or pits that appear on the sample surface and remain stable over long periods of time, even when several steps migrate over the holes (Hillner et al., 1992a,b; Bosbach and Rammensee, 1994).These features have been observed during both small- (less than 1 pm)and large- (greater than 5 p,m) region scans. Whether or not they are scan-induced, perhaps by frictional mechanisms (Delawski and Parkinson, 1992), remains to be determined. Scan-induced damage often results in features that align with the scan direction and the image frame. Reaction rates may appear to vary from a small-scan (e.g., less than 1 pm) to a large-scan area (e.g., 16 pm). To evaluate scan-induced artifacts, in situ reaction studies should include careful calibration and monitoring of forces, systematic alteration of scan angle, periodic disenablement of the y-scan direction, periodic change in scan size, and occasional tip withdrawal and reengagement (Maurice et al., 1995).Additionally, scan-induced erosion should be suspected whenever there is evidence of the tip sticking to the sample surface, e.g., when the zero of height is reset, requiring application of a flatten filtering procedure. Finally, it should be understood that surface erosion may not be limited to small-region scans, particularly when the sample surface is soft and easily deformable or when imaging forces have not been minimized. Thus, although AFM is a powerful technique for studying growth or dissolution in situ, the experiments must be performed with caution. TMAFM should be less likely to erode sample surfaces. However, TMAFM does not appear to give as high a resolution as contact-mode AFM; indeed, it can be difficult to obtain resolution below several tens of nanometers during in situ TMAFM experiments. Additionally, if force modulation tips are used, in situ TMAFM can be quite expensive. Despite these current drawbacks, with time,
-
22
PATRICIA A. MAURICE AND STEVEN K LOWER
TMAFM probably will become the method of choice for most in situ experiments of changes in surface microtopography.
2. The in Sitzl Experiment Researchers have used a variety of different approaches to in situ experiments, and the best approach depends upon the particular problem at hand. The fluid cell provided by Digital Instruments (Fig. 8) consists of a square of glass with inlet and outlet openings, a clip for the integrated probe cantilever-tip, and a circular groove made to fit an O-ring that is used to seal the glass cell to the underlying sample or sample stub. A small amount of an appropriate grease can be used to help prevent leakage from around the O-ring. The system may be used in either static, pulsedflow, or continuous-flow conditions. One common experimental setup is shown in Fig. 9. In this experimentalsetup, which was used by Maurice et al. (1995),a freshly fractured mm-scale piece of specular hematite was placed in the AFM fluid cell. The sample was monitored first in deionized water. Then, an oxalic acid solution was added and allowed to drip through the fluid cell using gravity feed from a syringe, at a very slow flow rate of only 1 mL hr- *.The slow flow rate helped to provide stable images; faster flow rates may result in fluid convection or turbu-
-
Figure 8 Cross-sectional view of the Digital Instruments AFM fluid cell. Using this cell, the tip and sample are completely immersed in solution.
USING ATOMIC FORCE MICROSCOPY
23
Figure 9 A gravity-flow-through fluid-cell experiment with the Digital Instruments AFM. Flow rate can be modified using a small clamp on the inlet or outlet tube. These tubes can be clamped off entirely to run in static conditions. Adapted from Maurice er al., 1995.
lence, which enhances drift and degrades cantilever stability. Fast flow rates also may result in bubbles being introduced into the fluid cell. The effluent from the cell can be collected and analyzed, although sample volumes tend to be low. Dove and Hochella (1 993) used a similar setup to image calcite dissolution and growth features. They began their experiments in a flow-through reactor and transferred the calcite crystals to the AFM fluid cell, along with solution at the appropriate saturation state. The experiment was then run as a closed (static) system, thus mimiclung a batch reactor. One problem was that the surfaces tended to dry during transport between the reactor and the AFM fluid cell, resulting in changes in microtopography. In the case of reactive surfaces, it may be important to try to prevent the surfaces from drying completely during transport. Alternatively, allowing a surface to dry can sometimes provide nuclei for growth experiments. Temperature control is another important consideration. Dove and Chermak ( 1994) reported that energy transfer from the AFM laser resulted in temperature increases of as much as 8°C in the fluid cell over six hours of continuous imaging at room temperature (22°C). Jordan and Rammensee (1996) modified the fluid cell to permit temperature control and precise temperature monitoring. Thus, it may be
24
PATRICIA A. MAURICE AND STEVEN K. LOWER
possible to prevent unwanted temperature increases. In the absence of such modifications, we recommend that the experiment may be stopped periodically, and the laser light turned off, to help prevent excessive warming of the fluid cell. The difficulty of this approach is that it is not always possible to return to the same spot upon re-engagement. Jordan and Rammensee’s (1996) experimental apparatus also opens the door for a range of kinetics experiments, although it should be noted that increased temperature can be expected to lead to enhanced thermal drift that may limit the range of viable experimental temperatures. When single mm-scale crystals are imaged in the fluid cell, it is important to provide an appropriate substrate if the typical steel sample stub might be reactive in the solution of choice. We have used either a piece of glass coverslip, a piece of mylar, or a piece of parafilm glued to the steel stub. We have also succeeded in imaging powders on the order of 200 pm or more in diameter in the TMAFM fluid cell, by dispersing the powders on the surface of double-sided cellophane tape attached to the sample stub (Lower et al., manuscript in preparation). At this point, we are assuming that the tape does not add reactive products to the experiment. Dove and Chermak (1994) discuss imaging clay-sized particles in solution and note that Nagy (Dove and Chermak, 1994, personal communication) recommends using a small puff of air from a compressed air source to disperse clay particles over epoxy. They note that although this method does have limitations, it appears to work. In general, we have found that fluid cell experiments tend to show a large amount of drift during initial imaging (see Fig. lo), but this drift generally settles down after 15 minutes. Thus, it is often best to begin experiments with “nonreactive” solution and then introduce the reactive solution of interest after the drift has died down and good images can be produced. Many experiments also benefit from first monitoring surface microtopography in a “nonreactive” solution and then watching the changes that occur upon introduction of the solution of interest. Unfortunately, changing solutions generally requires withdrawing-disengaging the tip from the sample surface for a brief interval. Upon reengagement, the image area often changes. We have found that if we work with a scan head allowing images 16 pm on a side, we can generally return to an area that at least overlaps the original image area. Several other considerations are important for in situ imaging. First, it is important to avoid introducing air bubbles into the fluid cell since they not only can negate the experiment (if the air bubble overlies the scan area) but also can complicate the laser signal if they occur within the path of the laser light. Second, because solutions and the glass of the fluid cell have different refractive indices than air, the laser, the mirror, and the photodetector generally require substantial realignment. Third, it is imperative to prevent leakage of the fluid cell onto the AFM since electrical shortages could ensue. We generally keep a substantial supply of rolled up Kimwipes ready to blot up even the smallest leak. If a leak develops, it
-
USING ATOMIC FORCE MICROSCOPY
25
Figure 10 An example of drift. The top of the image is drifted, resulting in elongation of features and some curvature. Drift may occur at any time but often is more pronounced within the first few minutes of imaging, when the scan direction turns around from upscan to downscan (or vice versa) or when the tip is having difficulty tracking the sample surface and is close to disengagement. Scan area is 3.1 I pm on a side.
is important to turn the AFM off immediately and allow it to dry completely before proceeding. Other researchers have developed plastic aprons that fit over the AFM base to protect from leaks. As the sample reacts in the fluid cell, reaction products or bits of torn-up surface may accumulate on the tip and result in a multitude of interesting tip-sample “convolution” artifacts. Multiple images are commonplace (e.g., Maurice, 1996). Reaction products also may accumulate on the cantilever; for example, we have found that supersaturated solutions may result in accumulation of nucleated materials on the cantilever, leading to destabilization of the imaging process (Lower et al., manuscript in preparation). In general, the fluid cell tends to be relatively difficult to use, and fluid-cell experiments can be tedious and time-consuming. However, the benefits of working in solution greatly outweigh the drawbacks.
3. Examples of in Situ Studies AFM has been used to study mineral surface growth (e.g., Dove and Hochella, 1993; Bosbach et al., 1996), dissolution (e.g., Gratz et al., 1991; Hillner et al.,
26
PATRICIA A. MAURICE AND STEVEN K. LOWER
1992a,b; Bosbach and Rammensee, 1993; Gratz et al., 1993; Maurice et al., 1995; Putnis et al., 1995), and heterogeneous nucleation processes (e.g., Junta and Hochella, 1994; Eggleston et al., manuscript in preparation) in situ since the early 1990s.A few examples will be discussed herein. One of the most enlightening fluid-cell experiments to date was conducted by Dove and Hochella (1993). ‘fhese authors used fluid-cell AFM to study near-equilibrium calcite growth and dissolution. Calcite fragments were reacted for 1 to 2 days in a flow-through reaction chamber and then transferred to the AFM fluid cell, along with aliquots of the same solutions. Dove and Hochella (1993) observed that when solution saturation states relative to calcite were greater than 1 or 2, precipitation began with the formation of surface nuclei that spread, coalesced, and continued to grow. A transition to a spiral-growth-type mechanism did not occur until after nearly 2 hours. Dove and Hochella (1993) also found that upon introduction of a phosphate-containing solution into calcite growth media, smooth straight-edged steps transitioned into widely spaced, jagged-edged steps. This observation was consistent with predictions by Stumm and Leckie (1970) and Berner and Morse (1974) that phosphate may bind to step edges and poison further growth. Hence, the in situ microscopic studies by Dove and Hochella were able to directly verify predictions based on macroscopic and theoretical considerations. Jordan and Rammensee (1996) used in situ AFh4 to study the kinetics of brucite (Mg(OH),) dissolution on the (001) surface. They used a gravimetric liquid-feed system that consisted of silicone tubes and allowed in situ exchange of solution while scanning the sample surface. A flow rate of 2.5 mL min-’ was used. Experiments were conducted at elevated temperature (to 35°C) by immersing the liquid-supplying tube in water in a thermostated vessel next to the microscope. A micro-NTC-resistor was placed in the outflow section of the fluid cell. Using this experimental setup, Jordan and Rammensee (1996) were able to monitor the retreat of steps across the brucite (001) surface and the growth of etch pits. From these data, they calculated dissolution rates as a function of temperature. The resulting Arrhenius plot was used to calculate an activation energy of 60.2 ? 12 W mol-’ for dissolution on the (001) surface. This activation energy is within the realm of surface-controlled dissolution. However, the authors observed some interstep interaction, which suggests that surface diffusion might contribute to the rate-controlling process, as well. Jordan and Rammensee’s (1996) pathbreaking work opens the door for a wide range of studies on the kinetics of mineral growth and dissolution. Eggleston et al. (manuscript in preparation) used in situ contact-modeAFM and XPS to study the formation of “active” Cr(II1) hydroxide (ACH) precipitate (e.g., Giovanoli et al., 1973) on a hematite surface by surface-catalyzed reduction of aqueous Cr(V1). The subsequent mobility of Cr bound in ACH depends on gradual aging of the reactive precipitate toward less reactive forms with less solidsolution interfacial area. The results of Eggleston et al. (manuscript in preparation)
USING ATOMIC FORCE MICROSCOPY
27
showed that ACH aging proceeds by layer-by-layer growth from an initial thin precipitate layer. The rate of growth (or dissolution) of a given layer depends on its position in a stack of layers and on the distance between the steps bounding each layer. However, clear evidence was seen for influences on AFM scan processes on the dissolution and growth of ACH. The tip appeared to scrape material off the surface, leading to alignment of some features with the scan direction. Use of XPS was important for determining reaction mechanism. XPS showed Cr(II1) on the reacted surface; however, these results were not unambiguous since XPS may itself cause photoreduction of Cr(V1) to Cr(II1). Nevertheless, Eggleston et al. believed that at least some, or perhaps all, of the Cr(II1) was present before XPS analysis. These experiments demonstrated the importance of combining different surficial techniques and of carefully analyzing in situ images for potential tip-related effects on apparent growth and dissolution.
D. Ex SZTUSTUDIES:PARTICULATE IMAGING 1. Fundamentals of ex Sits Particulate AFM Although in situ imaging of mm- to cm-scale crystals can provide direct mechanistic and kinetic data, soil chemistry ultimately revolves around the properties of soil particles. Imaging of particles can provide nanometer-scale data on particle surface microtopography under environmentally appropriate conditions, i.e., air or, in some cases, aqueous solution. Because AFM provides three-dimensional data, the technique also can be used to determine particle micromorphology, leading to calculations of aspect ratios and ultimately to estimations of external geometric surface areas. Particles can be imaged ex situ prior to and following reaction and changes in surface features; e.g., formation of etch pits can be directly linked to macroscopic data. AFM offers distinct advantages over most other imaging techniques in that it provides nanometer- to potentially atomic-scale three-dimensional surface images under environmental conditions, without the need for surface coatings, and in a nondestructive manner. Despite these capabilities and advantages, particle imaging by AFM is neither simple nor straightforward, and it entails a number of difficulties and disadvantages. First, AFM is by its very nature a surface technique. As such, it cannot probe the internal surfaces of intact particles, including internal pore spaces. Second, due to the finite shape of the tip, tip-sample “convolution” artifacts abound and may be more or less severe depending upon the particle micromorphology as well as sample preparation. Third, particle imaging requires some method of stabilizing the particles onto an imageable surface. Friedbacher et al. (1 99 1) showed that particulate material could be imaged by pressing powdered samples into pellets. However, this technique has two disadvantages: (1)
28
PATRICIA A. MALJlUCE AND STEVEN K. LOWER
particles often overlap, obscuring some features; and (2) the rough surfaces of pressed powders may be subject to significant tip-convolution artifacts. Hence, correct imaging of isolated, single particles requires use of a sample substrate.
2. Preparation of Samples Choice of substrate material is of paramount concern and may vary with the specific application. Maurice (1996) listed the following criteria for a good substrate material: (1) the substrate needs to be flat and regular on a scale greater than the scale of the particles of interest; (2) if the particles to be imaged are to be deposited from solution, the substrate must be unreactive; (3) the substrate must be readily imaged, and have consistent nanometer- to micron-scale topography; (4) the atomic-scale structure and nanometer-scale microtopography of the substrate should be distinguishable from that of the material of interest; ( 5 ) the substrate should not be easily deformed by the scanning process; and (6) particles should adhere well. Maurice et al. (1995) and Maurice (1996) showed that Nucleopore polycarbonate (pc) membrane filters originally developed for SEM work can serve as excellent substrate materials for imaging particle microtopography and micromorphology. These filter membranes provide several distinct advantages: (1) the substrate can be easily distinguished from mineralogic particles due to the filter pores and the fibrous nature of the filter materials; (2) the filters can be used both to separate solid from solution and to image the solid fraction; and (3) most mineralogic particles adhere well to the filter surfaces (Maurice, 1996). However, pc membranes have several disadvantages. First, the pc membrane surface is not as flat on an atomic to molecular scale; and, hence, the membranes are not suitable for particle thickness measurements and detailed measurements. As described later, muscovite mica is recommended for quantitative analysis of particle thicknesses. Also, the frictional properties of the filter membrane may be different from the frictional properties of mineralogic specimens, leading to frictional artifacts, such as inverted topography. Frictional artifacts generally occur only during conventional contact-modeAFM imaging and not during TMAFM imaging. Generally, we have found that pc membranes retain materials far smaller than their reported pore sizes. Muscovite mica is perhaps the most commonly used substrate for AFM imaging in air, and it has the advantage of containing large molecularly flat areas and frictionalproperties similar to clay materials.However, it exhibits two major drawbacks. First, muscovite is reactive in water (Lin and Clemency, 1981) such that muscovite surfaces exposed to water for even short periods of time may contain etch pits and secondary precipitates (Johnsson et al., 1992; Blum, 1994). Second, muscovite surface structure resembles the surface structures of clay minerals. U1trafine particles of muscovite caused by imperfect cleavage may be confused with
USING ATOMIC FORCE MICROSCOPY
29
the particles of interest, and molecular-scale imaging cannot be used to differentiate the substrate and sample materials. Nevertheless, muscovite often is the substrate of choice for measurements of environmental particles. It should be noted that in this section, we are specifically speaking about mounting particles for imaging in air; particles cannot be expected to remain attached to substrates, such as pc membranes and muscovite mica, when immersed in solution. Blum (1994) used contact-mode AFM to analyze the morphology of smectiteillite particles deposited on muscovite. After deposition, the solution was allowed to evaporate, leaving the particles attached to the surface. Blum (1994) found that ultrasonic dispersion of the suspension was necessary to isolate individual particles, although ultrasonic dispersion may lead to cleavage parallel to the basal plane. With careful attention to sample preparation and imaging, he concluded that AFM can be used to determine the heights of clay particles with an accuracy of 2 1.5 A for particles on the order of several nm in thickness and x y dimensions greater than 100 nm. Height accuracy for thicker andor less laterally extensive particles was reported to be +3.0 nm. Particle thickness measurements were found to agree well with measurements on the same samples by TEM and XRD. Maurice et al. (1995) and Lower et al. (manuscript in preparation) have imaged particles on pc membrane filters. Sample preparation is simple, but results are optimized by following a few careful procedures. Successful imaging requires deposition of an extremely low particle density on the filter membrane. High particle densities result in particle overlap, which obscures some features, and in diverse orientations may lead to increased tip-sample convolution. Hence, we try to start with a suspension containing less than 0.5 g of solids per liter of solution. We generally use 13-25 mm diameter, 0.1 pm filter membranes and syringe filters, and filter no more than -0.5-2 mL of suspension through the filter membranes. Depending upon the exact nature of the experiments, e.g., how reactive the particles are, we sometimes put through a small amount of either deionized water or another appropriate solution to rinse and then force a small amount of air from a clean syringe through the filter membrane to remove potential reaction products that otherwise might precipitate upon drying. A gentle vacuum also may be used. After the filter membrane is removed from the filter housing, the membranes should be allowed to continue drying under a protective dust cover. Subsequent to filtration and drying, an optical microscope should be used to ensure that individual particles or small aggregates are dispersed across the membrane. We have found that most samples should be imaged within a few hours to a few days after filtration, because particles may detach from the membrane surfaces. After the sample membrane has dried, the membrane can be cut into pieces a few mms on a side, and mounted to a stainless steel sample stub with double-sided cellophane tape. We generally dab the sample surface with a clean piece of Parafilm to ensure that air bubbles are removed from beneath the tape and sample. This must be performed gently to prevent potential damage of the sample surfaces.
30
PATRICIA A. MAURICE AND STEVEN K. LOWER
Subsequent to AFM imaging, these same samples can be coated with Au or C and analyzed by SEM and/or EDS.
3. Imaging Procedures Once the sample has been prepared, there are several factors to consider before intiating the imaging procedure. TMAFM reduces the frictional drag between tip and sample surfaces and is hence a better choice than contact-modeAFM for imaging of particles. A binocular optical microscope attachment is a vital asset to the successful imaging of particles, because a specific location can be chosen. For imaging of isolated particles with a size too small to be detected with an optical microscope, we generally choose an area of the sample-substrate surface that looks under the optical microscope to be essentiallyfree of particles and then move from place to place to try to find isolated particles. This is a laborious process. For imaging of aggregates, we have found that the edges of the aggregates are easiest to image because features tend to lie relatively flat, and particle density is reduced. It should, of course, be recognized that the edges of the aggregates may have different structure than the bulk of the aggregates. As described later, combination of AFM with SEM and TEM can sometimesprovide greater insight into particle and aggregate structures. Traditionally,AFM data are collected in the so-called height mode. This mode of operation provides the most accurate height measurements. However, we found that the amplitude mode (for TMAFM) complements the height mode, and it can be particularly useful for rough, aggregate surfaces. The amplitude mode, which may be collected simultaneously with the height mode data, is essentially a record of the error in the height image. It should be noted that amplitude mode for TMAFM is analogous to “deflection” images collected in contact-modeAFM (see Blum, 1994). The height image is a record of lateral changes in the z-axis piezoelectric cylinder and directly represents the microtopography of a sample surface.All quantitative height information should be gathered from this image. The amplitude image is a record of lateral changes in the photodiode voltage and reflects the magnitude of lag between the tip’s deflection by a topographic feature and the movement of the z-axis piezoelectrodeto maintain a constant force between the tip and the sample. The amplitude image displays the rate of change in surface microtopography. The magnitude of the tip deflection is a function of the steepness of the microtopography and is greatest in regions where the sample is roughest. For aggregates of particulate materials, which tend to be rough, the amplitude image is perhaps the most important part of the “real-time” imaging procedure because it provides the investigator with an easily interpretable representation of the surface. Figure 11 provides a comparison of height- and amplitude-modedata. Even using low particle densities and pc membrane filters, imaging of particles
Figure 11 Comparison of TMAFh4 "height-mode'' (a) and "amplitude-mode'' (b) images of particulates deposited on pc membrane filter (see pores in upper-right corner). The images were collected simultaneously using TMAFM in air on samples of hydroxylapatite (HAP) reacted with 100 mg LPb at pH 6,22"C. The needles in these images are of the mineral hydroxypyromorphite (HPY), which formed upon reaction. While the height-mode image gives vertical data, the amplitude-mode image is clearer, while maintaining horizontal scale. For explanation of sample chemistry and structure, see section III.D.4. Scan area of each image is 2.30 Fm on a side.
'
32
PATRICIA A. MAURICE AND STEVEN K. LOWER
often proves difficult. The tip-convolution problems described earlier need to be considered. Additionally, particles can be dislodged from the surface, resulting in noisy images and potential damage to the tip. When imaging particles, it is generally best to start with a scan region less than 1 pm on a side, and to slowly increase the scan size. Imaging with a very slow scan speed (less than 1 Hz) also can prove helpful. Scanning at 90”may result in less drag between the tip and the sample. In general, TMAFM is much more reliable than contact-mode AFM for particle imaging. Finally, imaging particles within a few days of sample preparation generally works best because particles tend to detach from substrate materials with time.
4. Pb “Sorption”to Hydroxylapatite
AFM can be used as an ex situ technique for imaging particulates subjected to batch growth or dissolution experiments. In so doing, micromorphological and microtopographical AFM data may be used to help identify solid phases and to elicit such information as the types of reaction mechanisms (e.g., surface vs diffusion controlled). However, imaging of particulates presents a challenge because AF’M traditionally does not provide chemical data. Therefore, solid phases must be identified primarily based on morphology. To successfully use AFM to identify solid phases, one should take a systematic approach. First, reactant and product materials should be characterized as fully as possible by XRD or EDS and other chemical techniques, such as infrared (IR) or Raman Spectroscopies. Second, it is essential to compile a large catalog of AFM images of the solid reactant. These images of reaction “blanks” may be compared with images of the reaction product to identify newly formed solid phases or changes in the original reactant. Unless there are distinct morphological forms for various solid phases, it will be difficult to use AFM to identify individual phases in heterogeneous samples. Finally, AFM should be compared with established forms of microscopy, such as SEM and TEM.As with any new emerging technique, there is a critical period of “validation.” AFM can produce numerous artifacts. Comparing AFM images to SEM or TEM images of the same material will provide the necessary documentation to ensure that AFM images are real. Additionally, different microscopic techniques offer complementary information. As part of a systematic approach to developing AFM application to studies of reaction products, TMAFM was applied to studies of Pb sorption to the mineral hydroxylapatite (Ca,(PO,),OH) (HAP) (Lower et al., manuscript in preparation). Aqueous solutions of Pb, ranging in concentration from 0-500 mg L-I, were reacted with 0.5 g L- of HAPin batch experiments that proceeded for 2 hours. XRD analysis of the reaction product revealed that a new phase, hydroxy-pyromorphite (Pb,(PO,),OH) (HPY) was produced. TMAFM analyses of the reaction product
’
USING ATOMIC FORCE MICROSCOPY
33
Figure 12 Surface of an HAP particle reacted at pH 6, 22”C, in the absence of Pb, and imaged with TMAFM in air. Scan area is 2.00 pn on a side.
filtered onto a 0.1 km pc membrane showed two distinct particle morphologies: elongate needles and clusters of material (Fig. 11). The chemical data provided by XRD, as well as a comparison of AFM images of the reaction product to a catalog of images of HAP reaction blanks (Fig. 12), allowed us to conclude that the needles were HPY and the clusters of material were HAP. Next, we conducted SEM/EDS analyses of the same samples subjected to AFM (Fig. 13). In general, we found that AFM shows good agreement with results using SEM. On the one hand, AFM provides higher resolution and more detail; on the other hand, SEM gives more information about larger-scale features and association of phases. EDS provides chemical information not available by AFM. SEM results showed that the tips of HPY needles generally were pointed; yet, they appeared to be rounded in AFM images. This suggests that even the sharp TMAFM tips can “convolve” with tip edges, distorting details of micromorphology. The ability to identify solid phases based on their morphology, as well as the increased resolution afforded by AFM, allowed us to make several hypotheses concerning the types of reaction mechanisms controlling Pb sorption to HAP (Lower et al., manuscript in preparation). At initial Pb concentrations of 10-500 mg L-’, needles of HPY form concurrently with HAP dissolution. At lower concentrations
34
PATRICIA A. MAURICE AND STEVEN K. LOWER
Figure 13 SEM image of HAP (platey structures) and HPY (needle-like structures) following reaction of HAP with 100 mg L-‘ Pb at pH 6, 22°C. This image agrees well with Fig. 11, which was from the same sample. SEM permits imaging at a larger scale, which helps to put features in context. Also, tip-convolutionfeatures are not a problem with SEM. However, TMAFM provides higher resolution and greater detail of the surface structures. Scale bar is 5 p,m.
of initial Pb, fewer needles of HPY were dispersed across the pc membrane (Fig. 14). At higher concentrations of Pb, numerous intergrown needles of I-IF’Y were detected with AFM (Fig. 15 on p. 36). These needles were smaller than the needles produced at lower initial Pb concentrations. The small size and intergrown nature of HPY needles at high concentrations (Fig. 15) suggests rapid homogeneous nucleation. At intial Pb concentrations less than 10 mg L-I (Lower et al., manuscript in preparation), we saw evidence of heterogeneous nucleation in the form of needles appearing to grow “epitaxially” on the HAP. This suggests that both mechanisms may be responsible for Pb sorption, although the importance of these different mechanisms may vary with solution conditions (Lower et al., manuscript in preparation). In situ studies are underway to help further our understanding of growth mechanisms. These experiments demonstrate the potential usefulness of AFM in studies of complex reaction “mechanisms.” By using AFM simultaneously with other techniques and designing scientifically relevant, carefully controlled laboratory studies, scientists are able to better understand multidimensional soil processes. In so doing, one is able to make powerful statements concerning natural systems.
USING ATOMIC FORCE MICROSCOPY
35
Figure 14 TMAFM image in air of HAP and HPY reaction products following reaction of HAP with 10 mg L-’ Pb at pH 6, 22°C. Based on XRD, microtopographic, and micromorphologic analysis, we believe that there are only a few scattered needles of HPY (see twin in upper-left portion of image) and that the bulk of the material imaged is HAP. Some filter-membrane substrate appears in this image (see pores). Tip-sample convolution artifacts appear near the upper-central portion of this image. Scan area is 3.64 p,rn on a side.
5. Measurements of Kaolinite Particles: Comparison of AsMeasurements Kaolinite is one of the most abundant minerals in earth-surface environments (Moore and Reynolds, 1989). Because kaolinite occurs primarily as clay-sized particles with large surface-area-to-volume ratios, it may strongly influence the distribution of species in solid-solution systems at the earth’s surface. Additionally, kaolinite is widely used in ceramic, paper, and coating-pigment industries. The performance of kaolinite in such industrial applications depends largely on the mineral’s surface properties. Surficial properties affect interaction with chemical compounds and media, as well as particle-to-particle interactions. Thus, study of kaolinite’s surface characteristics has important implications for both fundamental and applied research. Zhou et al. (manuscript in preparation) used AFM to compare the particle mor-
36
PATRICIA A. MAURICE AND STEVEN K. LOWER
Figure 15 TMAFM image in air of reaction products following reaction of HAP with 500 mg L-' Pb at pH. 22°C. The intergrown nature of the HPY needles suggests rapid homogeneous nucleation. Similar aggregates were observed with SEM, although details of the structure were difficult to discern with that method. We cannot determine from this image whether rounded structures are HAP or the ends of needles. XRD, however, did not detect any HAP remaining in this sample after reaction. Scan area is 2.00 pm on a side.
phology, surface microtopography, and estimated geometric surface areas of two American Clay Minerals Society standard kaolinites, Kga-lb (well crystallized) and Kga-2 (poorly crystallized). Qualitatively, the two kaolinite standards were found to have notably different surface microtopographies and micromorphologies. Particles of Kga-1b generally had crisp hexagonal outlines and basal-plane surfaces showing clear crystallographx control. In contrast, particles of Kga-2 tended to have more rounded hexagonal outlines, with curved edge steps and irregular basal-plane surfaces. AFM also was used quantitatively to measure grain size, thickness, and size distribution of the two samples. The average diameter and thickness were found to be 785 and 58 nm, respectively, for Kga-lb; and 447 and 45 nm, respectively, for Kga-2. Three dimensional data were used to estimate geometric specific surface areas (As), as shown in Table I. The average values of As of the two kaolinites, calculated from AFM measurements, were intermediate between relatively lower values determined by BET surface-area analysis, thermogravimetric analysis (TGA), and relatively higher values measured by XRD. Mode values determined by AFM
USING ATOMIC FORCE MICROSCOPY
37
Table I Comparison of Specific Surface Areas of Two Kaolinites Based on Different Methods Specific surface area (As) (m*/g)
Sample"
AFM mean
AFM mode
BET
XRD
TGA
Ratio edgebasal plane AselAsb
BETh ~
Kga-lb Kga-lbc Kga-2 Kga-2c
21 27 30 37
13 16 22 24
12.5 12.6 22.4 22.2
52.1 48.5 56.7 57.5
64.4 62.6 71.6 72.5
12.6
8.5-10.05
21.7
20.C24.0
Mean
Mode
~~
0.17 0.17 0.18 0.19
0.10 0.10 0.12 0.14
"Sample cleaned in 1 M NaCl adjusted to pH 3 wth HCI, followed by deionized water washings. bData from van Olphen and Fripiat (1979).
were close to BET and TGA values, suggesting that the AFM samples might have been subject to systematic sampling error. Values determined by XRD were double to triple the values determined by AFM, BET, and TGA, suggesting that XRD may measure small domains rather than entire crystals. Overall, the As values determined by the different methods could lead to substantial differences in estimated surface reactivities.AFM surpassed the other methods by providing additional detail in the form of grain size and As distributions (see Zhou et al., manuscript in preparation), as well as estimations of edge vs basal-plane surface areas (see Table I and Zhou, 1996). However, one major problem of AFM was that numerous artifacts precluded application of a strict point-counting regime. Hence, some sampling bias was inevitable.
n! SOME NEW FRONTIERS J NAFM RESEARCH One of the most promising new frontiers in AFM research appears to be chemical sensing. Chemical sensing relies on the fact that AFM force measurements are extremely sensitive to the properties of the tip and the sample. A variety of techniques may be used to distinguish chemically distinct surfaces. For example, Burnham et al. (1990) used force-curvemeasurements to differentiatebetween two similar samples covered with organic monolayers. One surface had -CH, group terminations, and the other had -CF, group terminations; these different terminations resulted in different tip-sample interaction forces. Chemical modification of AFM tip surfaces is opening a whole new realm of chemical sensing. For example, Frisbie et al. (1994) deposited various functionalized thiol monolayers on gold-coated tips and samples. By systematically varying the functional group terminations on the tips and the samples, the authors were able to map out different
38
PATRICIA A. MAURICE AND STEVEN K. LOWER
adhesion and friction properties between various combinations of carboxyl (-COOH) vs methyl (-CH,) functional groups. Hence, although AFM cannot give direct chemical determinations, it can be used to distinguish materials with different physicochemical properties, such as different hydrophobic-hydrophilic characteristics, surface charges, and charge densities. Another new technique is known as phase imaging. Phase imaging (see DI Web page, http://www.di.corn) is an extension of TMAFM that maps the phase of the cantilever oscillation during the scan. The phase of cantilever oscillation may vary with composition, adhesion, friction, viscoelasticity, and potentially other properties of sample materials. Hence, phase imaging can help to differentiate different materials on a sample surface. Generally, phase data are displayed next to TMAFM height information, so that the two types of data can be compared. Force modulation imaging is another method that may be useful for distinguishing materials with different stiffnesses. According to the DI Web page, in force modulation imaging, the sample is scanned with a small vertical ( 2 ) oscillation (modulation) that is significantly faster than the scan rate. The force on the sample is modulated about the setpoint scanning force in such a way that the average force on the sample is equivalent to that in contact-mode AFM. When the probe is brought into contact with a sample, the surface resists the oscillation and the cantilever bends. Under the same applied force, a stiff area on the sample surface will deform less than a soft area. Stiffer surfaces cause greater resistance to the vertical oscillation, and thus results in greater bending of the cantilever. Thus, the variation in the amplitude of cantilever deflection gives a measure of the relative stiffness of the surface. Microtopographical information (“height” data) is collected simultaneously with the force modulation data so that the two may be compared. Within the realm of soil surface chemistry, two of the most important new frontiers of research involve imaging humic substances and soil bacteria. Both of these research areas require use of specialized techniques, because humic substances and bacteria are soft materials. Namjesnik-Dejanovic and Maurice (1997) imaged soil and stream humic and fulvic acids deposited on muscovite surfaces. The humic substances were deposited on muscovite from solution, allowed to dry, and the samples were imaged in air by TMAFM. Spheroids, branched-chain structures, fibrous networks, and perforated sheets were observed at successively higher concentrations of humic substances. Results agreed well with results of previous studies utilizing TEM and SEM on similarly prepared samples (e.g., Stevenson and Schnitzer, 1982). Maurice et al. (1996b) imaged soil bacteria on Fe(II1) oxide surfaces, using ex situ TMAFM and in situ contact-mode AFM. Grantham and Dove (1996) used in situ TMAFM to investigate the adhesive properties of soil bacteria. One major problem with the humic-substances research (Namjesnik-Dejanovic and Maurice, 1997) is that humic substances can be expected to have different con-
USING ATOMIC FORCE MICROSCOPY
39
figurations when adsorbed in solution than when dried and imaged in air. Recent developments in AFM techniques should make in situ imaging of sorbed humic substances possible. In situ TMAFM is one technique that may allow such imaging by permitting low-force imaging in solution. Contact-mode imaging in solution may also be possible, with the help of force-minimization techniques. Use of soft cantilevers may prove beneficial. Some of the long Si cantilevers provided by DI (450 km length) may have spring constants less than 0.06 N/m; sharpened Si,N, probes with spring constants -0.04 Nlm also may be used. Specialized tips may also be important. For example, recent images with Buckey-balls attached to cantilevers have provided high resolution on polymer surfaces. Use of chemically modified tips (chemical imaging) may help to map out hydrophobic-hydrophilic materials, or areas with different adhesion and laterial forces, and should be useful for imaging humic materials. Because the elastic moduli of soft materials, such as bacteria and humic substances, should be quite different from substrates such as muscovite, force-modulation imaging should also prove useful. Phase imaging can be used to map surface hardness or elastic modulus. Phase-imaging techniques have been used to image soft mesophase (liquid crystal) structures, and should therefore be applicable to bacteria and humic substances, as well. These are just a few of the recent frontiers in AFM. Techniques and applications continue to advance at a rapid rate, and keeping track of recent advances is one of the greatest challenges for the AFM practitioner. One of the best ways for keeping up-to-date is the World Wide Web. DI, for example, maintains an award-winning Web page, including recent application notes, up-to-date press releases, and search engines for a variety of AFM-related topics. AFM discussion groups on the Internet are also useful, especially for the neophyte user.
ACKNOWLEDGMENTS
Our research would not be possible without the continued help and support of our collaborators. Research on Pb “sorption” to hydroxylapatite is being conducted in collaboration with S. Traina (Ohio State University) and E. Carlson (Kent State University, KSU). Research on microbial interactions with iron oxides is being conducted in collaboration with G. Sposito (University of California, Berkeley), L. Hersman (Los Alamos National Laboratories), and J. Forsythe (KSU). Research on kaolinite measurements is being conducted in collaboration with D. Eberl (US. Geological Survey), M. Jaroniec (KSU), and Q. Zhou (KSU). R. Klouda (Liquid Crystals Institute, KSU), assisted with SEM imaging of HAPIHPY. We thank the Water Resources Institute of KSU, the National Science Foundation, the Petroleum Research Fund of the American Chemical Society, the U.S. Department of Energy, and the LACOR program of Los Alamos National Laboratories for funding various aspects of our AFMrelated research.
40
PATRICIA A. MAURICE AND STEVEN K. LOWER
REFERENCES Barrett, R. C. (1991). “Development and applications of atomic force microscopy.” Ph.D. Dissertation, Stanford University, Stanford, CA. Barrh, V., GBlvez, N., Hochella, M. F., Jr., and Torrent, .I.(1997). Epitaxial overgrowth of goethite on hematite synthesized in phosphate media: A scanning force and transmission electron microscopy study. Am. Mineralogist. In press. Berner, R. A., and Morse, J. W. (1974). Dissolution kinetics of calcium carbonate in sea water IV. Theory of calcite dissolution. Amer. J. Sci. 274, 108-134. Binnig, G.,Quate, C. F., and Gerber, C. (1986). Atomic force microscope. Phys. Rev. Len. 56,930-933. Blum, A. E., and Eberl, D. D. (1992). Determination of clay particle thicknesses and morphology using scanning force microscopy. In “Water-rock interaction” (Y.K. Kharaka and A. S. Meast, eds.), Vol. VII, pp. 133-136. Balkema, Amsterdam. Blum, A. E. (1994). Determination of illite/smectite particle morphology using scanning force microscopy. In “Scanning Probe Microscopy of Clay Minerals” (A. E. Blum and K. Nagy, eds.), pp. 172-202. Clay Minerals SOC.,Boulder, CO. Bosbach, D., and Rammensee, W. (1994). In situ investigation of growth and dissolution on the (010) surface of gypsum by scanning force microscopy. Geochim. Cosmochim. Acta 58,843-849. Bosbach, D., Jordan, G., and Rammensee, W. (1995). Crystal growth and dissolution kinetics of gypsum and fluorite: An in situ scanning force microscope study. EUKJ. Mineral. 7,267-276. Bosbach, D., Junta-Rosso, J. T., Becker, U., and Hochella, M. F., Jr. (1996). Gypsum growth in the presence of background electrolytes studied by scanning force microscopy. Geochim. Cosmochim. Actu 60,3295-3304. Burnham, N. A., Domingue, D. D., Mowery, R. L., and Colten, R. J. (1990). Probing the surface of monolayer films with an atomic force microscope. Phys. Rev. Lett. 64, 1931-1934. Butt, H. J. (1991a). Electrostatic interaction in atomic force microscopy. Biophys. J. 60,777-785. Butt, H. J. (1991b). Measuring electrostatic, van der Waals, and hydration forces in electrolyte solutions with an atomic forcer microscope. Biophys. J. 60, 1438-1444. Butt, H. J. (1994). A technique for measuring the force between a colloidal particle in water and a bubble: An atomic force microscopy study. Lungmuir 11, 156-162. Butt, H. J., Jaschke, M., and Ducker, W. (1995). Measuring surface forces in aqueous electrolyte solution with the atomic force microscope. Bioelecrrochem. and Bioenergetics. Delawski, E., and Parkinson, B. A. (1992). Layer-by-layer etching of two-dimensional metal chalcogenides with the atomic force microscope. J. Am. Chem. Soc. 114, 1661-1667. Digital Instruments (DI). (1995). “Nanovations.” Digital Instruments Newsletter. Dove, P. M., and Chermak, J. A. (1994). Mineral-water interactions: Fluid cell applications of scanning force microscopy. In “Scanning Probe Microscopy of Clay Minerals” (A. E. Blum and K. Nagy, eds.), pp. 149-169. Clay Minerals SOC.,Boulder, CO. Dove, P. M., and Hochella, M. F., Jr. (1993). Calcite precipitation mechanisms and inhibition by orthophosphate: In situ observations by scanning force microscopy. Geochim. Cosmochim. Actu 57, 705-714. Dove, P. M., and Platt, F. M. (1996). Compatible real-time rates of mineral dissolution by atomic force microscopy (AFM). Chem. Geol. 127,331-338. Ducker, W. A., Senden, T. J., and Pashley, R. M. (1991). Direct measurement of colloidal forces using an atomic force microscope. Nature 353,239-241. Ducker, W. A., Senden, T. J., and Pashley, R. M. (1992). Measurement of forces in liquids using a force microscope. Langmuir 8,183 1-1836. Eggleston, C. M. ( 1994). High resolution scanning probe microscopy: Tip-surface interaction, artifacts, and applicatons in mineralogy and geochemistry. In “Scanning Probe Microscopy of Clay Minerals” (A. E. Blum and K. Nagy, eds.), pp. 1-90. Clay Minerals Soc.,Boulder, CO.
USING ATOMIC FORCE MICROSCOPY
41
Eggleston, C. M., and Hochella, M. F., Jr. (1992). The structure of the hematite (001 ] surfaces by scanning tunneling microscopy: Image interpretation, surface relaxation, and step structure. Am. Mine,: 77,911-922. Eggleston, C. M., Maurice, P. M., Stumm, W., Hauert, R., and Kraack, M. (n.d.). Direct, in-situ observation of layer-by-layer growth of “active” Cr(OH),(H,O), from Cr(V1) reduction onto a hematite surface: Precipitate aging and evidence for surface diffusion. Manuscript in preparation. Friedbacher, G., Hansma, P. K., Ramli, E., and Stucky, G . D. (1991). Imaging powders with the atomic force microsope: From biominerals to commercial materials. Science 253, 1261-1263. Frisbie, C. D., Rozsnyai, L. F., Noy, A., Wrighton, M. S., and Lieber, C. M. (1994). Functional group imaging by chemical force microscopy. Science 265,207 1-2073. Giovanoli, R., Stadelmann, W., and Feitknecht, W. (1973). Uber kristallines Chrom(II1)hydroxid. I. Helv. Chim. Acta 56,839-847. Grantham, M. E., and P. M. Dove. ( I 996). Bacterial-mineral surface interactions: Investigations using fluid tapping mode atomic force microscopy. Geochim. Cosmochim. Acta 60,2473-2480. Gratz, A. J., Hillner, P. E., and Hansma, P. K. (1993). Step dynamics and spiral growth on calcite. Geochim. Cosmochim. Acta. 57,491495. Gratz, A. J., Manne, S., and Hansma, P. K. (1991). Atomic force microscopy of atomic-scale ledges and etch pits formed during dissolution of quartz. Science 51, 1343-1346. Griffith, J. E., and Grigg, D. A. (1993). Dimensional metrology with scanning probe microscopes. J. A@. Phys. 74, R834109. Hartman, H., Sposito, G., Yang, A,, Manne, S., Gould, S. A. C., and Hansma, P. K. (1990). Molecularscale imaging of clay mineral surfaces with the atomic force microscope. Clays Clay Miner 38, 337-342. Hillner, P. E., Gratz, A. J., Manne, S., and Hansma, P. K. (1992a). Atomic-scale imaging of calcite growth and dissolution in real-time. Geology 20,359-362. Hillner, P. E., Manne, S., Gratz, A. J., and Hansma, P. K. (1992b). AFM images of dissolution and growth on a calcite crystal. Ultramicroscopy 42-44, 1387-1393. Hochella, M. F., Jr. (1990). Atomic structure, microtopography, composition, and reactivity of mineral surfaces. In “Mineral-Water Interface Geochemistry” (M. F. Hochella, Jr., and A. F. White, eds.), pp. 87-132. Miner. SOC.Am., Washington, DC. Hochella, M. F., Jr. (1995). Mineral surfaces: Their characterization and their chemical, physical and reactive nature. In “Mineral Surfaces” (D. J. Vaughan and R. A. D. Pattrick, eds.), pp. 17-60. Chapman and Hall, New York. Hochella, M. F., Jr., Eggleston, C. M., Elings, V. B., and Thompson, M. S. (1990). Atomic structure and morphology of the albite (010)surfaco: An atomic-force microscope and electron diffraction study. Am. Mineral. 75,723-730. Inskeep, W. P., and Bloom, P. R. (1985). An evaluation of rate equations for calcite precipitation kinetics at pC0, less than 0.01 atm and pH greater than 8. Geochim. Cosmochim. Acta. 49, 2165-2180. Israelachvili, J. N. (1992). “Intermolecular and Surface Forces.” 2nd ed. Academic Press, San Diego, CA. Israelachvili, J., and Adams, G. E. (1978). Measurements of forces between two mica surfaces in aqueous electrolyte solutions in the range 0-100 nm. J. Chem. SOC. Faruduy Trans. I, 74,975-1001. Johnsson, P. A., Eggleston, C. M., and Hochella, M. F., Jr. (1991). Imaging molecular-scale structure and microtopography of hematite with the atomic force microscope. Am. Mineral. 76,1442-1445. Johnsson, P. A,, Hochella, M. F., Jr., Parks, G. A., Blum, A. E., and Sposito, G. (1992). Direct observation of muscovite basal-plane dissolution and secondary phase formation: An XPS, LEED, and SFM study. In “Water-Rock Interaction” (Y. K. Kharaka and A. S. Maest, eds.), Vol. VII, pp. 159-162. Balkema. Amsterdam.
PATRICIA A. MAURICE AND STEVEN K. LOWER Jordan, G., and Rammensee, W. (1996). Dissolution rates and activation energy for dissolution of brucite (001): A new method based on the microtopography of crystal surfaces. Geochim. Cosmochim. Acfa. 60,5055-5061. Junta, J. L., and Hochella, M. F., Jr. (1994). Manganese (II) oxidation at mineral surfaces: A microscopic and spectroscopic study. Geochim. Cosmochim. Acta. 58,49854999. Junta-Rosso J. L. and Hochella M. F., Jr. (1996). The chemistry of hematite (001} surfaces. Geochim. Cosmochim. Acta. 60,305-314. Kado, H., Yokoyama, K., and Tohda, T. (1992). A novel ZnO whisker tip for atomic force microscopy. Ultramicroscopy 42-44, 1659-1663. Keller, D., Deputy, D., Alduino. A,, and Luo, K. (1992). Sharp, vertical-walled tips for SFM imaging of steep or soft samples. Ulframicroscopy42-44,148 1-1489. Kepler, K. D., and Gewirth, A. A. (1994). In situ AFM and STM investigation of electrochemical hydride growth on Ge(l10) and Ge(l1 I ) surfaces. Sudace Sci. 303, 101-113. Lin, F. C., and Clemency, C. V. (1981). The kinetics of dissolution of muscovites at 25°C and 1 atm CO, partial pressure. Geochim. Cosmochim. Acta 45,571-576. Lindgreen, H., Gamaes, J., Hansen, P. L., Besenbach, F., Laegsgaard, E., Stensgaard, I., Gould, S. A., and Hansma, P. K. (1991). Ultrafine particles of North Sea illitelsmectite clay minerals investigated by STM and AFM. Am. Mineral. 76, 1218-1222. Lower, S. K., Maurice, P. A,, Traina, S. J., and Carlson, E. H. (n.d.). Pb “sorption” to hydroxylapatite: Evaluation of ex-situ AFM applications to growth, dissolution, and heterogeneous nucleation studies. Manuscript in preparation. Manne, S., Cleveland, J. P., Gaub, H. E., Stucky, G. D., and Hansma, P. K. (1994). Direct visualization of surfactant hernimicelles by force microscopy of the electrical double layer. Langmuir 10,4409. Marchant, R. E., Lea, A. S., Andrade, J. D., and Bockenstedt, P. (1991). Interactions of von Willebrand factor on mica studied by atomic force microscopy. J. Colloid Znte$ace Sci. 148,261-272. Markiewicz, P., and Goh, M. C. (1994). Atomic force microscopy probe tip visualization and improvement of images using a simple deconvolution procedure. Langmuir 10,5-7. Maurice, P. A. (1996). Application of atomic force microscopy in environmental colloid and surface chemistry. Colloids and Su$aces A 107,57-75. Maurice, P., Forsythe, J., Hersman, L., and Sposito, G. (1996). Application of atomic-force microscopy to studies of microbial interactions with hydrous Fe(II1)-oxides. Chem. Geol. 132, 3 3 4 3 . Maurice, P. A., Hochella, M. F., Jr., Parks, G. A., Sposito, G., and Schwertmann, U.(1995). Evolution of hematite surface microtopography upon dissolution by simple organic acids. Clays Clay Minec 43,29-38. Maurice-Johnsson, P. A. (1993). “Hematite dissolution in natural organic acids.” Ph.D. Dissertation, Stanford University, Stanford, CA. Maurice-Johnsson. P. A. (1994). Use of AFM in experimental surface geochemistry. Miner Mag. 58A, 575-576. Moore, D. M., and Reynolds, R. C., Jr. (1989). “X-ray Diffraction and the Identification and Analysis of Clay Minerals.” Oxford University Press, New York. Nagy, K. (1994). Application of morphological data obtained using scanning force microscopy to quantification of fibrous illite growth rates. In “Scanning Probe Microscopy of Clay Minerals” (A. E. Blum and K. Nagy, eds.), pp. 203-239. Clay Minerals SOC.,Boulder, CO. Namjesnik-Dejanovic, K., and Maurice, P. A. (1997). Atomic force microscopy of soil and stream fulvic acids. Colloids and Sugaces A 120,77-86. Oden, P. I., Nagahara, L. A., Graham, J. J., Pan, J., Tao, N . J . , Li, Y., Thundat, T. G.,Derose, J. A., and Lindsay, S . M. (1992). Atomic force and scanning tunneling microscopy observations of whisker crystals and surface modification on evaporated gold films. Ultramicroscopy 42,580-586. Ohnesorge, F., and Binnig, G. (1993). True atomic resolution by atomic force microscopy through repulsive and attractive forces. Science 260, 1451-1456.
USING ATOMIC FORCE MICROSCOPY
43
Overney, R. M., Meyer, E., Frommer, J., Brodbeck, D., Liithi, R., Howald, L., Giintherodt, H. J., Fujihira, M., Takano, H., and Gotoh, Y.(1992). Friction measurements on phase-separated thin films with a modified atomic force microscope. Nature 359, 133-135. Prater, C. B., Maivald, P. G., Kjoller, K. J., and Heaton, N. G. (1995). Tapping-mode imaging application and technology. Digital Instruments Nanonotes. Putnis, A,, Junta-Rosso, J. L., and Hochella, M. F., Jr. (1995). Dissolution of barite by a chelating ligand: An atomic force microscopy study. Geochim. Cosmochim. Acfa 59,46234632. Radmacher, M., Fritz, M., Cleveland, J. P., Walters, D., and Hansma, P.K. (1994). Imaging adhesion forces and elasticity of lysozyme adsorbed on mica with the atomic force microscope. Langmuir 10,3809-3814. Rimstidt, J. D.. and Dove, P. M. (1986). Mineral/solution reaction rates in a mixed flow reactor: Wollastonite hydrolysis. Geochm. Cosmochim.Acta. 50,2509-25 16. Shiraki, R., and Brantley, S. L. (1995). Kinetics of near-equilibrium calcite precipitation at 100°C: An evaluation of elementary reaction-based and affinity-based rate laws. Geochim. Cosmochim. Acra 59, 1457-1471. Stevenson, I. L., and Schnitzer, M. (1982). Transmission electron microscopy of extracted fulvic and humic acid. Soil Sci. 133, 179-185. Stumm, W., and Leckie, J. 0. (1970). Phosphate exchange with sediments: Its role in the productivity of surface waters. In “Advances in Water Pollution Research 2” (S. H. Jenkins, ed.) pp. 26/1-26/16. Pergamon, New York. Thundat, T., Zheng, X.-Y.,Chen, G. Y., and Warmack, R. J. (1993). Role of relative humidity in atomic force microscopy imaging. Surf: Sci. 294, L939-L943. Weihs, T. P., Nawaz, Z., Jarvis, S. P., and Pethica, J. B. (1991). Limits of imaging resolution for atomic force microscopy of molecules. Appl. Phys. Lett. 59,35363538. Weisenhorn, A. L., Hansma, P. K., Albrecht, T. R., and Quate, C. F. (1989). Forces in atomic force microscopy in air and water. Appl. Phys. Lett. 54,2651-2653. Weisenhorn, A. L., Maivald, P., Butt. H. J., and Hansma, P. K. (1992). Measuring adhesion, attraction, and repulsion between surfaces in liquids with an atomic-force microscope. Phys. Rev. B 45, 11226-1 1232. Wilson, D. L., Kump. K. S., Eppell, S. J., and Marchant, R. E. (1997). Lungmuir: In press. Zhong, Q., Inniss, D., Kjoller, K., and Elings, V. B. (1993). Fractured polymer/silica fiber surface studied by tapping mode atomic force microscopy. Surf: Sci. 290, L688-L692. Zhou, Q., Maurice, P. A,, Eberl, D. D., and Jaroniec, M. (n.d.). Comparison between the surface characteristics of two Georgia kaolinites. Manuscript in preparation.
This Page Intentionally Left Blank
PLANTGROUTTH-REGULATING SUBSTANCES IN THE RHIZOSPHERE: MICROBIAL PRODUCTION AND FUNCTIONS Muhammad Arshadl and William T. Frankenberger,Jr.* 'Department of Soil Science University of Agriculture Faisalabad, Pakistan 2Departmentof Soil and Environmental Sciences University of California Rverside, California 9252 1
I. Rhizosphere as a Site of Plant-Microbe Interactions 11. Plant Growth-Regulating Substances A. Auxins B. Gibberellins C . Cytokinins D. Ethylene E. Abscisic Acid 111. Sources of PGRs IV. Biochemistry of Microbial Production of PGRs A. Auxins B. Gibberellins C . Cytokinins D. Ethylene E. Abscisic Acid V; Production of PGRs by Rhizosphere Microorganisms A. Plant Growth-Promoting Rhizobacteria B. Free-Living Diazotrophs (Azotobacterand Azospiellum) C . Rhizobium D. Mycorrhizal Symbiosis E. Pathogenesis VI. Metabolism of PGRs in Soil A. Auxins B. Gibberellins c. Cytokinins D. Ethylene E. Abscisic Acid
45 Advances in Agronomy. Volume 62
Copyright 0 1998 by Academic Press. All rlghts of reproduction In any form reserved 0065-2113/98 $25.00
46 MUHAMMAD ARSHAD AND WILLIAM T. FRANKENBERGER,JR. VII. Ecological Significance of PGRs Produced in the Rhizosphere A. PGRs Produced by an Inoculum B. Precursor-Inoculum Interactions C. Exogenous Application of Physiological Precursors of PGRs VIII. Conclusions Appendix: Abbreviations References
r.
RHIZOSPHERE AS A SITE OF PLANT-MICROBE INTERACTIONS
The rhizosphere is that portion of the soil under the direct influence of the roots of higher plants. It is considered the most intense ecological habitat in soil in which microorganisms are in direct contact with plant roots. The root system of all higher plants is associated with a distinct, diverse community of metabolically active soil microbiota that carry out biochemical transformations. Rhizosphere microorganisms may have specific associations (mutualistic, antagonistic, or a combination thereof) with plants through which they exert their influence on plant growth. Several possible mechanisms of plant-microbe interactions are summarized in Fig. 1 (Frankenberger and Arshad, 1995).Among these, the production of biologically active metabolites, particularly the plant growth regulators by rhizosphere microbiota are considered one of the most important mechanisms of action through which the rhizosphere microbiota affect plant growth directly after being taken up by the plant or indirectly by modifying the rhizosphere environment. Microbial production of PGRs in the rhizosphere and its subsequent influence on plant growth is the major focus of this chapter.
11. PLANT GROWTH-REGULATING SUBSTANCES Plant growth-regulating substances are naturally occurring organic compounds that influence physiological processes in plants at concentrations far below those at which nutrients or vitamins could affect these processes. When produced endogenously in plants, plant physiologists refer to them as “plant hormones” or “phytohormones”; the term plant growth regulators (PGRs) is being used by the agrochemical industry for synthetic compounds with plant growth-regulatory properties. However, no strict restriction has been imposed by scientists working in this field on the interchangeability of these terms. In this
PLANT GROWTH REGULATORS IN THE RHIZOSPHERJ?
47
Figure 1 Possible plant-microbe interactions affecting plant growth (from Frankenberger and Arshad, 1995, used with permission).
chapter, the primary focus is on plant hormones produced (exogenously) by rhizosphere microbiota-thus we will use the term PGRs. There are five classes of well-established PGRs, namely auxins, gibberellins, cytokinins, ethylene, and abscisic acid. It is not our intent to discuss in depth the biochemistry of these PGRs in plants (for this information, see Frankenberger and Arshad, 1995);however, introductory information is necessary to understand the foundation of this subject.
48 MUHAMMAD ARSHAD AND WILLIAM T. FRANKENBERGER,JR.
A. AUXINS Several naturally occurring auxins found in plants include indole-3-acetic acid (IAA), its halogenated derivatives (e.g., 4-chloroindole-3-acetic acid [.I-Cl-IAA]), and indole-3-butyric acid (IBA). Amide and ester IAA conjugates are also present and are believed to be storage forms of auxins in plants. Some of the intermediates of tryptophan (TRP) conversion into IAA-including indole-3-acetamide (IAM); tryptophol (TOL), also known as indole-3-ethanol, and tryptamine (TAM)-also possess auxin activity. In addition to these indolic auxins, various phenolic compounds (e.g., phenylacetic acid and phenylacetamide) and others in plants have low auxin activity. IAA (Fig. 2) is the most physiologically active auxin in most plants. Initially, TRP was considered the sole precursor for IAA biosynthesis, but recent studies have revealed that plants that cannot synthesize TRP can make IAA de novo (Normanly et al., 1993; Wright et al., 1991). The presence of more than one TRP-based pathway of IAA biosynthesis in higher plants has been reported (Bandurski et al., 1995; Frankenberger and Arshad, 1995).
B. GIBBERELLINS Gibberellins (GAS) are tetracyclic diterpenoid acids with an ent-gibberellane ring system (Fig. 2). Although the most widely recognized gibberellin is GA, (gibberellic acid), which is a fungal product, the most active GAin plants is GA,, which is primarily responsible for stem elongation (Davies, 1995). Over 89 GAS are known to date, and those are numbered GA, through GA,, in approximate order of their discovery. For more information and their chemical structure, see Frankenberger and Arshad (1995). Mevalonic acid (MVA) is believed to be the primary precursor of GAS biosynthesis in plants (Sponsel, 1995). Most GASknown today are not physiologically active; however, they play an important role by serving as possible precursors for the physiologically active GAS.
c. CYT0KI"S Cytokinins are N6-substituted aminopurines, including ribosides, ribotides, and glucosides. These are adenine derivatives characterized by their ability to induce cell division in tissue culture in the presence of auxins. The most common cytokinin in plants is zeatin (Fig. 2), which is converted to other cytokinins (McGaw and Burch, 1995). Over 40 cytokinins have been characterized in plant tissues; however, they greatly differ in their biological activity (Frankenberger and Arshad, 1995). They exist in plants as free-base forms or as tRNA constituents.
PLANT GROWTH REGULATORS IN THE RHIZOSPHERE
Indole-3-acetic acid (IAA)
OH
HO
Gibberellane skeleton of gibberellins
Zeatin
H
/c-c \-
H
/H
\
H
Ethylene
Abscisic acid Figure 2 Structures of compounds representing five classes of PGRs.
49
50 MUHAMMAD ARSHAD AND WILLIAM T FFWNKENBERGER,JR.
D. ETHYLENE Ethylene (Fig. 2) is synthesized from methionine (MET) in many plant tissues, mostly in response to stress. It is the only hydrocarbon with a pronounced effect on plants. In addition to its recognition as a “ripening hormone,” ethylene (C,H4) is involved in other developmental processes, from germination of seeds to senescence of various organs. The pathway of MET to C,H, in plant systems is now well established with S-adenosylmethionine (SAM) and 1-aminocyclopropane-1carboxylic acid (ACC) serving as intermediates (MET + SAM +ACC + C2H4). However, in microorganisms, more than one pathway of C,H4 biosynthesis has been reported.
E. ABSCISIC ACID Abscisic acid (ABA) is a sesquiterpene (Fig. 2) derived from MVA. Although its exogenous application can inhibit growth in the plant, ABA appears to act as much as a promoter (e.g., storage protein synthesis in seeds) as an inhibitor, and a more open attitude toward its overall role in plant development is warranted (Davies, 1995). The biosynthetic pathway of ABA in higher plants and by some fungi has been characterized and reveals that, like GAS, MVA serves as the substrate for ABA synthesis (Zeevaart and Creelman, 1988; Frankenberger and Arshad, 1995; Walton and Li, 1995).
111. SOURCES OF PGRs Higher plants are capable of synthesizing all five major classes of PGRs. Details about these endogenously produced plant hormones can be found in many recently published reviews (Bandurski et al., 1995; Sponsel, 1995; McGaw and Burch, 1995; McKeon et al., 1995; Walton and Li, 1995; Kutacek et al., 1988; Moore, 1989; Crozier, 1983; Sembdner et al., 1980;Takahashi, 1986).Another potential source of PGRs is the soil microbiota, particularly rhizosphere microorganisms (Frankenberger and Arshad, 1995; Costacurta and Vanderleyden, 1995; Patten and Glick, 1996; Arshad and Frankenberger, 1993). The presence of PGRs in soil has also been well demonstrated (see Nieto and Frankenberger, 1990a; Arshad and Frankenberger, 1993; Frankenberger and Arshad, 1995). However, the soil pool of these PGRs might have partially originated from plants released into the rhizosphere as root exudates and/or partially synthesized by soil microbiota in situ. Microflora that are able to produce PGRs in vitro are present in appreciable numbers in the rhizosphere of plants. Barea et al. (1976) found that among 50 bac-
PLANT GROWTH REGULATORS IN THE RHIZOSPHEFE
51
terial isolates obtained from the rhizosphere of various plants, 86, 58, and 90% produced auxins, GAS,and kinetin-like substances, respectively. Kampert and colleagues (1975a) found that 55% of the bacteria and 86% of the fungi isolated from the rhizosphere of Pinus silvestris were capable of producing GASand similar substances. Sarwar and Kremer (1995a) screened 16 different bacterial isolates (belonging to the genus Enterobacter, Xanthomonas, Pseudornonas,Alcaligenes, and Agrobacterium spp.) originating from the rhizosphere of various plants and 4 isolates of Pseudomonas spp. from bulk soil capable of auxin production in vitro. They found that all of the rhizosphere isolates were much more efficient producers of auxins than the soil isolates. In another study, Sarwar and Kremer (1995b) reported the ability of 16 rhizobacteria, out of a total of 70 isolates, could metabolize L-TRP into auxins. The amount of auxin produced ranged from 4.0 to 86.1 mg IAA-equivalent L- culture medium. Phosphate-solubilizingbacteria belonging to the genus Pseudomonas and Acinetobacter isolated from the rhizosphere of wheat and rye were found to produce auxins ranging from 0.01 to 3.98 mg IAAequivalents per liter culture medium (Leinhos and Vacek, 1994). Similarly, Fuentes-Ramirez and associates (1993) reported that all 18 strains of the plant growth-promoting rhizobacterium (PGPR) Acetobacter diazotrophicus from 13 cultivars of sugarcane had the ability to produce IAA ranging from 0.14 to 2.42 Fg IAA m I - I . They suggested that since this PGPR is found within the plant tissues, its ability to synthesize IAA could promote sugarcane growth. By using bioassays, Mansour et al. (1 994) tested 24 soil isolates of thallobacteria belonging to the genus Streptomyces for their potential to produce PGRs. All were found capable of producing auxins, GAS, and cytokinins but varied greatly in their production. The auxins produced ranged from 10.5 to 39.0, whereas GASwere in the range of 1.21 to 13.3, and cytokinins varied between 2.5 to 14.9 pg mL-'. Plantassociated methylotrophs (Long et al., 1996), thermophilic cyanobacteria (Reimers et al., 1994), and phototrophic purple bacterium (Serdyuk et al., 1995) have been reported capable of producing cytokinins in vitro. Arshad and Frankenberger (1989) observed that corn rhizosphere contained appreciable numbers of microorganisms capable of synthesizing C,H,.
'
IV BIOCHEMISTRY OF MICROBIAL PRODUCTION OF PGRs A. AUXINS The pathway of TRP-dependent IAA biosynthesis is illustrated in Fig. 3. In microorganisms, auxin formation has been'shown to occur in the absence of TRP, but its presence stimulates auxin production, indicating that some microorganismsuse
OH
mcH2CHC~H 1
H
Indde-34actic acid
(IW H Indole-3-ethanol (Tau
H Indole-3-pyrwic acid (WA)
Tryptophan
Indole-3-acetaldehyde
Q-CH,CH2NH*
H Indole-3-acetic acid
(IW
H Tryptarnine (TAM)
H Indole-3-acetarnide (IAW
Figure 3 Various routes of TRP-dependent IAA biosynthesis present both in higher plants and in microorganisms.
PLANT GROWTH REGULATORS IN THE RHIZOSPHERE
53
TRP as an auxin precursor. Various intermediates detected in a TRP-supplemented medium reveal that auxin production by microorganisms proceeds via more than one pathway, with a single bacterium strain sometimes operating more than one pathway (Sequeira and Williams, 1964; Kuo and Kosuge, 1970). Kuo and Kosuge (1970) found that IAA was produced by Pseudornonas syringae pv. savastanoi when supplied with TRP, IPyA, IAM, and IAAld, but not when supplied with TAM. IAM was the most effective precursor in IAA formation with whole cells of P syringae pv. savastanoi. However, IAA was not detected when IAM was supplied to sonicated cells. Only TRP and IAAld were converted into IAA in cellfree extracts, leading to the conclusion that more than one pathway exists in the conversion of TRP into IAA. The biochemistry and molecular aspects of various pathways of IAA biosynthesis in microorganisms have been critically reviewed by Gaudin et al. (1994), Costacurta and Vanderleyden (1995), and Patten and Glick (1996). According to Patten and Glick (1996), the level of expression of IAA in bacteria depends on the biosynthetic pathways; location of genes involved (either chromosomal or plasmid DNA) and their regulator sequences; and the presence of enzymes that can convert active IAA into an inactive, conjugated form. There is increasing evidence that various microorganisms can produce IAA by the formation of IPyA, derived from L-TRP by an a-ketoglutarate-dependent transaminase reaction. This IPyA pathway is considered to be a major IAA pathway in plants. The instability of IPyA in the extraction solvent (Schwarz and Bitancourt, 1957; Stowe, 1955; Srivastava, 1964; Frankenberger and Poth, i987b) has hindered the elucidation of this pathway in microbial biosynthesis of IAA. Badenoch-Jones et al. (1982a) identified IPyA as an intermediate in the conversion of TRP to IAA in culture supernatants of Rhizobium strains by GC-MS. Since then, TRP-derived IAA production through the IPyA pathway by Rhizobiurn and Azospirillurn species has been convincingly demonstrated by using state-of-theart techniques (Kaneshiro et al., 1983; Ruckdaschel et al., 1988, 1990; Ruckdaschel and Klingmuller, 1992; Zimmer et al., 1988; Crozier et al., 1988). By using HPLC-UV spectrum analysis, Martens and Frankenberger (199 1) identified IPyA among the catabolites of TRP in the culture of Pseudornonas sp. Frankenberger and Poth (1988) also reported the transformation of L-TRPto IPyA in a cellfree extract of a rhizobacterial isolate of Festuca octoflora Walt. Recently, Koga et al. (1994) purified L-TRP aminotransferase from Enterobacter cloacae, which catalyzes the conversion of TRP to IPyA. The presence of IPyA was also detected by El-Abyad et al. (1994) in the culture medium of an actinomycete, Streptornyces griseojiavus. The second step in this pathway is decarboxylation of IPyA into IAAld (Fig. 3). The gene encoding TPyA decarboxylase has been cloned and sequenced from Enterobacter cloacae (Koga et al., 1991a,b, 1992; Zimmer et al., 1994), Enterobacter agglornerans, Klebsiella oxytoca DSM 3539, Klebsiella aerogenes DSM681 and Pantoca agglornerans IMETll328( Z i m e r et al., 1994), and Azospirillurn brasilense (Costacurta et al., 1994). Koga (1995) has comprehen-
54 MUHAMMAD ARSHAD AND WILLIAM T.FRANKENBERGER,JR. sively reviewed the structure and function of IPyA decarboxylase involved in IAA synthesis. Koga et al. (1 994) indicated that IPyA decarboxylase is the ratelimiting step in this pathway. This enzyme has a very high level of specificity for the substrate, IPyA (Koga et al., 1992). One of the side reactions is that IPyA may also be transformed into ILA. Among other studies, ILA was found as an intermediate of TRP metabolism by several cultures of bacteria (Libbert et al., 1970a,b,c; Zimmer et al., 1988; Crozier et al., 1988; Costacurta and Vanderleyden, 1995). Labeled ILA was identified as a metabolite of labeled TRP added to soil, indicating the ability of soil-indigenousmicroflora to carry out this transformation (Martens and Frankenberger, 1993a). Another side reaction of this IAA pathway is the conversion of IAAld into TOL, a transformation regarded to be a regulatory reaction in IAA synthesis. Numerous studies have reported the detection of TOL among TRPmetabolites in microbial cultures (Rigaud, 1970a,b;Badenoch-Jones et al., 1982b; Hartmann and Glombitza, 1967; Libbert et al., 1970a; Crozier et al., 1988; Ruckdaschel and Klingmiiller, 1992; Ernstsen et al., 1987). Oxidation of IAAld leads to IAA formation (Kuo and Kosuge, 1970). TAM has been identified as an intermediate in TRP-dependent IAA formation by several microorganisms (Hirata, 1958; Crady and Wolf, 1959; Schwinn, 1965; Sridhar et al., 1968). TRP is converted into TAM by pyridoxal phopshate-dependent TRP decarboxylase. Perley and Stowe (1966b) reported that Bacillus cereus possesses a TRP decarboxylase that catalyzed the conversion of TRP into TAM. TAM is metabolized into IAAld by an amine oxidase. The sequence of reactions involved in this pathway is shown in Fig. 3. The TAM pathway is also operative in Azospirillum spp. (Hartmann et al., 1983; Ruckdaschel and Klingmuller, 1992). The addition of TAM to soils resulted in a considerable percentage (18%) in the conversion to IAA (Martens and Frankenberger, 1993a), revealing the ability of soil indigenous microflora to derive IAA from TAM. It can be concluded that the TAM pathway of IAA biosynthesis is widespread in many microorganisms. However, direct conversion of TRP to IAAld by a TRP side-chain oxidase has also been reported by Takai and Hayaishi (1987) and by Oberhlinsli et al. (1991) in a Pseudomonas sp. The most well-studied pathway in the microbial conversion of TRP to IAA involves the oxidative decarboxylation of TRP to IAM and then hydrolysis of IAM to yield IAA (Fig. 3). Recently, this pathway has also been found functional in plant tissues of trifoliate orange (Kawaguchi et al., 1993). P. syringae pv. savastanoi was the first isolated bacterium known to produce IAM from TRP. Subsequently, this pathway was also found in other pathogens, includingAgrobacteruim tumefaciens (Follin et al., 1985; InzC et al., 1984; Klee er al., 1984; Van Onckelen et al., 1985, 1986; Schroder et al., 1984; Thomashow et al., 1986),Agrobucterium rhizogenes (Camilleri and Jouanin, 1991), and Erwinia herbicolu pv. gypsophilae (Manulis er al., 1991a,b). Recent studies have shown that this pathway is also operative in nonpathogenic m.icroorganisms. IAA production derived from
PLANT GROWTH REGULATORS IN THE RHIZOSPHEFE
55
TRP by the IAM pathway has been recently demonstrated with Azospirillum brasilense sp.7 (Bar and Okon, 1993;Prinsen et al., 1993; Costacurta et al., 1992), whereas Sekine et al. (1988, 1989) have confirmed unequivocally the ability of Bradyrhizobium japonicum to produce IAA from TRP via IAM pathway. Similarly, spontaneous mutants (5-methyl tryptophan resistant) of R. leguminosarum biovar vicie, which have a DNA sequence with high homology to the bum gene showing IAM hydrolase activity, were isolated (Kawaguchi et al., 1990). By using HPLC and mass spectrometry, Prikryl et al. (1985) detected IAM among the metabolites derived from TRP by Pseudomonas putida and R fluorescens. They suggested that IAA biosynthesis by these bacteria involves two different metabolic pathways. The first one consists of decarboxylation of TRP to IAM, which is subsequently hydrolyzed to IAA. The second one involves IPyA as an intermediate, which is also common in plants. Recently, Martens and Frankenberger (199 1, 1993a) reported IAM as a TRP metabolite in the cultures of a soil bacterium (Pseudomonas sp.) and rhizobacteria (A. tumefaciens, P. aeruginosa) isolated from the grass Festuca octoflora Walt. Similarly, by using HPLC, Kravchenko et al. (1991) also detected IAM among the catabolites of TRP in the culture supernatant of Xanthomorzas, Flavobacterium, and Enterobacter aerogenes. Labeled TRP was converted into IAM by P. syringae pv. savastanoi and catalyzed by a single enzyme (Magie et al., 1963). TRP appears to be an inducer for this enzyme (Kuo and Kosuge, 1969). The conversion of IAM to IAA is believed to occur by hydrolysis. Comai and Kosuge (1980) reported that the synthesis of IAA, being coded by a plasmid in P. syringae pv. savastanoi, is catalyzed by two enzymes: TRP-2monoxygenase (TOM) that decarboxylates TRP to IAM,and IAM hydrolase that cleaves IAM to IAA and ammonia. Later studies confirmed that P. syringae pv. savastanoi contains two plasmid-borne genes that encode the enzymes TOM and IAM hydrolase (Comai and Kosuge, 1982; Comai et al., 1982; Yamada et al., 1985; Chernin et al., 1984). The transformation of IAM to IAA may be catalyzed by a membrane-bound enzyme, since sonicated cells of P. syringae pv. savastanoi failed to carry out this reaction (Kuo and Kosuge, 1970). Molecular studies revealed substantial homology between the plasmid-borne genes that coded for IAA synthesis in P. syringae pv. savastanoi and A. tumefaciens (Yamada et al., 1985). After detecting IAM among the TRP catabolites in soil, Frankenberger and coworkers proposed that TRP-derived IAA synthesis in soil also occurs by the IAM pathway, in addition to other pathways (Frankenberger and Brunner, 1983; Martens and Frankenberger, 1993a,b). An additional bacterial IAA pathway from TRP involves indole-3-acetaladoxime and indole-3-acetonitrile (IAN). A gene encoding for a nitrilase, which catalyzes the conversion of IAN to IAA, has been cloned and sequenced in Alcaligenes faecalis (Kobayashi et al., 1993). Later, Kobayashi et al. (1995) reported the presence of a nitrile hydratase and an amidase in A. tumefaciens and Rhizobium. The enzyme, nitrile hydratase, stoichiometrically catalyzes the hydration of IAN
56 MUHAMMAD ARSHAD AND WILLIAM T. FRANKENBERGER,JR. into IAM. Kobayashi et al. (1995) proposed a scheme revealing the conversion of IAN to IAA via IAM (IAN + IAM + IAA). IAA is metabolized readily by microorganisms to several products. The reactions leading to these products include oxidation, conjugation, and hydroxylation of IAA. The major microbial catabolic products of IAA seem to be indole-3carboxylic acid (ICA) and indole-3-aldehyde. Egebo et al. (1991) and Nielsen et al, (1988) reported that IAA catabolism by 8.japonicum is substrate inductive and 0, dependent. Opening of the indole ring is catabolized by tryptophan 2,3dioxygenase. Egebo et al. (1991) identified anthranilic acid as the terminal degradation product. Martens and Frankenberger (1993a) have demonstrated the mineralization of IAA, IPyA, IAM, and TOL in soils by measuring CO, recovery from an amended soil. This implies that soil-indigenous microflora have the ability to further degrade auxins into unknown substances. The formation of IAAconjugates (ester and amide) by microorganisms is also common. The conjugate, 3-indoleacetyl-E-L-lysine, was formed by R syringae pv. savastanoi when incubated with IAA and lysine (Hutzinger and Kosuge, 1968a,b; Kuo and Kosuge, 1969). Kosuge et al. (1983) and Comai and Kosuge (1980) reported that TRP-derived IAA formation by strains of R syringae pv. savastanoi represents only a transitory intermediate since IAA was rapidly conjugated to form IAA-lysine.
B. GIBBERELLINS Much of the progess in understanding the pathway of GA metabolism has been made with cultures of Gibberellafujikuroi. The steps involved in GA biosynthesis can be grouped into three stages: (1) conversion of MVA to ent-kaurene, (2) conversion of ent-kaurene to GA,,-7-aldehyde (GA,,Ald), and (3) conversion of GAl,-7-aldehyde to C,,-GAS and interconversion of GAS (Fig. 4). MVA is first converted into the key intermediate, isopentenyl pyrophosphate (iPP), which, in turn, is converted into geranylgeranyl pyrophosphate (GGPP) by stepwise condensation. Fall and West (1971) have purified and characterized kaurene synthetase from Fusarium moniliforrne. The most striking feature of this enzyme system is the existence of two inseparable activities, Aand B. ActivityAconverts GGPP into the bicyclic copalyl pyrophosphate (CPP), and activity B carries out cyclization of CPP to yield ent-kaurene. Two processes, including oxidation along with subsequent hydroxylation and ring contraction, are believed to be involved in conversion of ent-kaurene to GA1,-7-aldehyde (GA,,-Ald). entKaurene is oxidized stepwise at C-19 to form ent-kaurenol, ent-kaurenal, and entkaurenoic acid, with the latter becoming hydroxylated to ent-7a-hydroxykaurenoic acid. ent-7a-hydroxykaurenoic acid serves as a triple branch point and can be metabolized either by oxidative P-ring contraction yielding GA,,-Ald, or by 66-hydroxylation yielding ent-6a,7a-dihydroxykaurenoicacid, or by 6a-hydroxylation yielding lactonic 7P-hydroxykaurenolide (West, 1973). Only GA,,-Ald is
PLANT GROWTH REGULATORS INTHE RHIZOSPHERE CH3,
57
,
CHGOOH
H0Ic\
CH2CHaOH
Hsvalonic acid (MVA)
ant- kaurene
t
*
GA24
GAii 4 . ) GA47
GAo Figure 4 Biosynthesis of GAs in various fungi (Gibberellafujikuroi, Sphaceloma manihoticola, and Phaeosphuria sp. L487) (adapted from Sassa, 1992).
the precursor of GAS,which indicates that ring contraction separates the GASfrom other tetracyclic diterpenes. GA ,,-7-aldehyde is the first compound possessing an em-gibberellane skeleton but lacks a carboxyl group at C-6. It is converted to all the major GAS mainly through GA1,-7-aldehyde (CAI,-Ald) by G. fujikuroi (Cross et al., 1968; Bearder et al., 1973; Hedden et al., 1974). It serves as a branch-
58 MUHAMMAD ARSHAD AND WILLIAM T. FRANKFLNBERGER, JR. ing point leading to either the 3-nonhydroxylated GAS via GA,,, such as GA, and GA,,, or through GA,,-Ald to the 3P-hydroxylated GAS, such as GA, and GA,, (Evans and Hanson, 1975; Hedden et al., 1974). 3P-Hydroxylation seems to be the first reaction taking place immediately after ring contraction (Bearder et al., 1975). Some C,,-GAS may act as substrates for the synthesis of other C,,-GAS by G.fijikuroi. As evident from Fig. 4, GA, and GA, are the first two C,,-GAS formed by G. fijikuroi. Recently, Fernindez-Martin et al. (1995) investigated GAS biosynthesis in gib mutants of G. fijikuroi defective at various steps of synthesis. Their study indicated that GA, is derived from GA,, whereas GA, is derived from GA,, by the fungus. Recently, Sassa and his co-workers reported GA-production by the fungus Phaeosphaeria spp. and investigated the metabolic route from GA, to GA,, comparing it with G. fujikuroi and Sphaceloma manihoticola. They detected a wide spectrum of GAS (GA,, GA,, GA,, GA,,, GA,,, GA,,, GA,,, and GA,,) in the culture filtrate of Phaeosphaeria sp. L487 (Sassa, 1992; Sassa et al., 1989; 1994; Kawaide and Sassa, 1993; Kawaide et al., 1995). Feeding experiments revealed that this fungus metabolized GA, to GA, via GA, and GA,,, with the GA, route being more efficient (Kawaide et al., 1995).
c. cYT0KI"s Unlike auxins, extensive studies characterizing the biosynthetic pathway of cytokinins in soil microorganisms have not been conducted, although several studies have reported microbial production of cytokinins in vitro. Among the proposed pathways, adenine (ADE) is considered the starting point (substrate) for cytokinin production in both plants and microorganisms (Dickenson et al., 1986; Murai, 1980; Nieto and Frankenberger, 1989a). Nieto and Frankenberger (1989a) reported that ADE, adenosine-5'-monophosphate(AMP), isopentyl alcohol, isopyrene, and 3-methyl-2-buten-1-01 served as precursors of cytokinin synthesis by Azotobacter chroococcum. They identified a number of cytokinins in the culture media of A. chroococcum enriched with ADE and isopentyl alcohol. In plants, zeatin (Z) serves as the central point for the synthesis of other cytokinins (McGaw and Burch, 1995), but in microorganisms it is not clear whether or not the cytokinins identified were derived from 2 or directly from a precursor. The primary cytokinins, isopentenyl adenine (iP), isopentenyladenosine (iPA), or their phosphorylated derivatives, are hydroxylated by a yet undefined enzyme to form the highly biologically active t-Z derivatives (Letham and Palni, 1983; Barry et al., 1984; Heinemeyer et al., 1987). Studies with the phytopathogen A. tumefaciens have provided evidence of the presence of ipt in T-DNA, which codes for the A2-isopentenyl pyrophosphate: AMP-2-isopentenyl transferase. This enzyme catalyzes the synthesis of [9R-5'P]iP from AMP and 2-iPP (AMP + 2iPP -3 [9R-5'P]iP) (Barry et al., 1984). This en-
PLANT GROWTH REGULATORS IN THE RHIZOSPHERE
59
zyme has also been characterized in cell-free preparation of the slime mold, Dictyostelium discoideum (Taya et al., 1978). Moms (1995) has proposed a scheme for direct and indirect synthesis of cytokinins in prokaryotes. According to the proposed scheme, the major direct route in bacteria (Fig. 5A) is via condensation of dimethylallylpyrophosphate (DMAPP) and 5’-AMP to form isopentenyladenosine5’-phosphate ([9R-5’P]iP. Subsequent dephosphorylation and deribosylation andor side-chain hydroxylation (Fig. 5B) can lead to the production of [9R]iP, iP, [9R-5’P]Z, [9R]Z, and Z. The condensing enzyme is a prenyl transferase, dimethylallylpyrophosphate: 5-AMP transferase (DMAPP:AMP transferase or isopentenyl transferase). Genes encoding for DMAPP:AMP transferase have been isolated from A. tumefaciens, A. rhizogenes, I? syringae pv. savastanoi, I? solanacearum, Rhodococcus facians, and Erwinia herbicola pv. gypsophilae (Morns, 1986; Regier et al., 1989; Akiyoshi et al., 1989; Crespi et al., 1992; Lichter et al., 1995).An alternative indirect biosynthetic route (Fig. 5C) is via isopentenylated tRNA. Almost all organisms isopentenylate some adenine residues in subpopulation of their tRNA. The enzyme responsible is a DMAPP:tRNA transferase encoded by the m i d gene (Moms et al., 1993; Gray et al., 1992). Recently, by using a m i d - deletion-insertion mutant of A. tumefaciens, Gray et al. (1996) unequivocally demonstrated the release of iP into an extracellular medium synthesized by the activity of tRNA:isopentenyl transferase encoded by the bacterial miaA gene. They hypothesized that this tRNA-mediated synthesis may also account for cytokinin production by other plant-associated bacteria, such as rhizobia, that have been reported to secrete similarly low levels of nonhydroxylated cytokinins (Gray et al., 1996).Thus, isopentenylated tRNA is a potential source of free cytokinins by the excision process illustrated in the Fig. 5C (Morris, 1995). Recently, Schwartzenberg et al. (1994) reported further metabolism of [9R]iP and [9R]Z by rhizosphere microflora into two unknown metabolites, “X’ and “Y,” respectively. They found no metabolism when sterile in vitro culture seedlings were incubated with [9R]iP or [9R]Z, indicating that microorganisms are necessary for the biogenesis of “X’ and “Y.”
D. ETHYLENE The biochemistry of microbial production of C2H4is very complex, since C2H, can be derived from a variety of compounds. Most of the literature indicates that microbial biosynthesis of C,H4 occurs through pathway(s) different from those of higher plants (MET 3 SAM + ACC 3 C,H4). The pathways of C,H, biosynthesis identified in higher plants and microorganisms are illustrated in Fig. 6 (on p. 62). This subject has been recently reviewed by Fukuda and Ogawa (1991), Arshad and Frankenberger (1992), and Frankenberger and Arshad (1995). A wide range of compounds has been reported as possible precursors for C,H,
60 MUHAMMAD ARSHAD AND WILLIAM T FRANKENBERGER,JR.
A
Direct Synthesis
bH i)H
OH OH
Ioopntenyladeno.lne 5 -monophosphate
Adonoelno I'-monoph~phate
B
DMAPP + 5'AMP
IP
I
[9R-5'P]Z
PI
Rlbose
Figure 5 Possiblecytokinin biosynthetic pathways in prokaryotes. (A) Direct synthesis catalyzed by DMAPPAMP transferase encoded by the A. fumejuciensgenes ipr or tzs. (B) Putative metabolic interconversions following direct isopentenyladenosine5'-monophosphate synthesis. (C) Indirect synthesis via M A . Adenine residues adjacent to the anticodon triplet are isopentenylated by the DMAPP: tRNA prenyl transferase encoded by m i d . Subsequent cleavage of the nucleoside bond leads to the formation of free iP (from Moms, 1995, used with permission).
generation by different microbial isolates, including alcohols, sugars, organic acids, amino acids, Krebs cycle acids, methionine analogs, and phenolic compounds of humic nature (see Frankenberger and Arshad, 1995). However, MET has been found the most favorable substrate. Most bacteria and fungi generally produce C,H, only in the presence of MET (Primrose, 1976; Billington et al., 1979; Lynch, 1972; Ince and Knowles, 1985,1986; Fukuda et ul., 1989a; Arshad and Frankenberger, 1989). However, a partially purified ethylene-forming enzyme (EFE) isolated from I? digitutum utilized a-ketoglutarate (KGA) as an immediate
PLANT G R O W T H REGULATORS IN T H E RHIZOSPHERE
C
61
Indirect Synthesis (mIW 3'
DMAPP
A
A-37
leA-37
iP
Figure 5-Continued
precursor of C,H4, and no C,H, was released when it was substituted with MET (Fukuda et al., 1986). The use of glucose as a direct C,H, substrate in microbial biosynthesis of C,H, is an issue of controversy. Several workers believe that glucose supplies an energy source that stimulates microbial production of C,H, but does not directly serve as an C2H4substrate (Lynch and Harper, 1974a; Chalutz et al., 1977; Fukuda et al., 1989a; Arshad and Frankenberger, 1989). However, a number of other studies have revealed the production of C,H, by microbial cultures when grown in a medium in which glucose was the only organic nutrient (Swart and Kamerbeek, 1977; Dasilva et al., 1974; Wang et al., 1962; Pazout et al., 1981; Goto et al., 1985; Hahm et al., 1992). Fukuda et al. (1988) proposed a biosynthetic pathway for C2H4 generation by f? digitaturn that revealed that glucose is the starting point and is converted into KGA, serving as an immediate precursor of C2H,. Although most soil microorganisms can synthesize C,H4 from MET, the pathway does not appear to be the same as that of higher plants, because: (1) SAM and ACC have not yet been isolated in C,H,-producing microbial cultures; (2) studies
NHz CH7v-CHz-CHz-CH-COOH I
0
2- 0x0-4-methylthiobutyricacid (KMBA)
Methionine (MET)
other substrate
I+,++HOOC-CHz-CHz-
Ethylene (ClUR)
CO-COOH
a-ktoglutaric acid
(KGN Figure 6 Pathways of C2H4biosynthesis identified in (A) higher plants, (B) E. coli and C. albidus. and (C) I? digitaturn and I? syringae pv. phaseolidola.
PLANT GROWTH REGULATORS IN THE RHIZOSPHERE
63
have not demonstrated microbial utilization of SAM or ACC as C,H, precursors (however, according to Li et al., 1992, a transformed Escherichia coli generates C,H, by a pathway operative in higher plants); and (3) inhibitors of C,H, biosynthesis in higher plants are not effective in inhibiting MET-dependent C,H, production by microbial cultures. Studies with E. coli, Cryptococcus albidus, I? digitatum, and I? syringae pv. phaseolicola revealed that there are mainly two biosynthetic pathways of C,H, production in microorganisms. In one pathway, C,H, is produced from MET via 2-0x0-4-methylthiobutyric acid (KMBA) and in the other pathway a-ketoglutaric acid (KGA) (which may or may not be derived from glucose) serves as a substrate for C,H, formation. Unlike plants, MET-derived C,H, production in some microorganisms involves KMBA as an intermediate. Primrose (1976) suggested that methional or KMBA were possible intermediates in MET-derived C,H, production by E. coli; however, more evidence supports the involvement of KMBA rather than methional (Primrose, 1977). Later, Billington et al. (1979) confirmed the presence of KMBA in culture fluids of a diverse group of C,H,-producing bacteria, as well as some yeast, when supplemented with MET. Ince and Knowles (1985,1986) further confirmed the findings of Primrose (1977) that KMBA serves as an intermediate in MET-dependent C,H, biosynthesis in E. coli. They demonstrated that the conversion of MET into KMBA is catalyzed by a soluble transaminase enzyme. 2hydroxy-4-methylbiobutyric acid (HMBA) was also a product but did not serve as an intermediate. Both KMBA and HMBA were formed before optimal production of C,H, occurred. C,H, produced was derived from the C-3 and C-4 atoms of MET. Other studies have also demonstrated the conversion of MET into C,H, through KMBA in E. coli cultures (Mansouri and Bunch, 1989; Shipston and Bunch, 1989). The enzyme involved in the conversion of MET to KMBA has not been well characterized. Ince and Knowles (1986) are of the view that the enzyme involved is an NAD+ or N A D P linked dehydrogenase. However, inclusion of aketo acids resulted in stimulation of C,H, formation, which suggests that a transaminase is operative in the transformation of MET into KMBA. Fukuda et al. (1989a) reported that KMBA, a deaminated product of MET, accumulated in the culture filtrate of C. albidus. They partially purified a cell-free C,H,-forming enzyme (EFE) from C. albidus in a supernatant fraction obtained by centrifugation. This system required KMBA, NADH, EDTA, Fe(III), and 0, for C,H, production. It seems likely that C. albidus produces C2H4through a pathway similar to that of E. coli. In another study, Fukuda and co-workers (1989~) found that EFE is an NADH-Fe(II1)EDTA oxidoreductase. Ogawa et al. (1990) concluded that production of C2H, requires the presence of a specific transaminase that catalyzes the formation of KMBA depending on NADH:Fe(IIII)EDTA oxidoreductase, which is present in all microorganisms. Studies have unequivocally demonstrated that KGA serves as an immediate precursor of C,H, in some microorganisms. Wang et al. (1962) reported the active
64 MUHAMMAD ARSHAD AND WILLIAM T. F-NBERGER,
JR.
conversion of I4C-labeled glucose, glycine, alanine, glutamic acid, and aspartic acid into C,H, by I? digitatum. Chou and Yang (1973) conducted a comprehensive study by feeding radio-labeled Krebs cycle intermediates and glutamate to f! digitatum. They concluded that both KGA and glutamic acid are the most efficient precursors of C,H, biosynthesis by this fungus and that C,H, is derived from C, and C, of these substrates. Chou and Yang (1973) concluded that KGA must be the branching point at which the pathway of C,H, biosynthesis breaks off from the Krebs cycle. On the other hand, some studies have demonstrated that P digitatum is also capable of using MET as a C,H, precursor (Primrose, 1977; Chalutz et al., 1977; Chalutz and Lieberman, 1978; Billington et al., 1979). It was shown that f! digitatum produces C,H, by two different pathways, depending on whether the fungus is cultured under static or shake conditions (Chalutz et al., 1977; Chalutz and Lieberman, 1978). Glutamate and KGA serve as C,H4 precursors in static cultures, whereas MET serves as a precursor and inducer of C,H, production in shake cultures. Chalutz et al. (1977) further indicated that MET is the precursor of C,H4 produced by living fungal cells but not by filtrates. Fukuda et al. (1986) also reported that cell-free extracts of I? digitatum did not use MET to produce C,H,. An C,H,-forming system was isolated that utilized KGA as an immediate precursor when the reaction mixture was incubated on a shaker (Fukuda et al., 1986). Since incubation in studies conducted by Fukuda et al. (1986, 1989b) was carried out on a shaker and KGA served as a precursor, these findings are contrary to those reported by Chalutz et al. (1977) and Chalutz and Lieberman (1978), who stated that KGA or glutamate served as C,H, precursors of P digitatum in static cultures only. Ethylene production by I? syringae pv. phaseolicola has been investigated by many workers (Goto et al., 1985; Goto and Hyodo, 1987; Nagahama et al., 1991a,b; Fukuda et al., 1992). Goto et al. (1985) reported that the bacterium effectively produced C2H, from amino acids such as glutamate, aspartate, and their amides. In another study, Goto and Hyodo ( 1987)characterized the cell-free C,H,forming system of I? syringae pv. phaseolicola (Kudzu strain). Ethylene was most effectively produced from KGA at 0.5 mM followed by glutamate and then L-histidine at 5-10 mM. Nagahama et al. (1991a,b) studied the formation of C,H, in vitro by an extract off! syringae pv. phaseolicola PK2. The components of this system, with the exception of the stirnulatory effects of L-histidine, were similar to those of a system derived from the C,H,-producing, plant pathogenic fungus I? digitatum, which also produced C2H4in vitro in a reaction dependent on KGA.
E. ABSCISIC ACID Unlike many of the other phytohormones, soil microbiota have not been investigated extensively for their ability to synthesize ABA. The first authentic demon-
PLANT GROWTH REGULATORS IN THE RHIZOSPHERE
65
stration of ABA biosynthesis by the fungus Cercospora rosicola Passerini was reported by Assante et al. (1977). Since then, ABA synthesis in fungi has been well characterized. Efforts have been made to elucidate the ABA biosynthetic pathway in three related fungi, namely C. rosicola, C. cruenta, and C.pini-densiforae.Although these fungi use acetate and MVA as substrates for ABA biosynthesis, there is some evidence that variation, particularly at latter stages, occurs in their biosynthetic pathways. As shown in Fig. 7, 1I-deoxy ABA, 1’,4’-dihydroxy-y-ionylideneacetic acid (1 ’,4’-diH-y-AA), and 1’P’-t-diol ABA serve as immediate precursors of ABA in C. rosicola, C. cruenta, and C. pirzi-densiJorae,respectively. Strong evidence suggests that C. rosicola synthesizes ABA via an isoprenoid pathway (Bennett et al., 1981; Neill et al., 1981, 1982b; Neill and Horgan, 1983; Al-Nimri and Coolbaugh, 1990). When this fungus was supplied with labeled acetate or MVA, ABA was synthesized in a relatively high yield. Moreover, labeled farnesyl phosphate (FP) and farnesyl pyrophosphate (FPP) were also converted into ABA upon feeding to C.rosicola culture (Bennett et al., 1984), confirming the involvement of the isoprenoid pathway. A radioactive metabolite of C. rosicola was identified as 1’-deoxyABA from a medium amended with labeled MVA (Neill et al., 1981; 1982b; Horgan et a/., 1983). When 1 -deoxy ABA was re-fed to the fungus, it was converted to ABA in good yield, revealing that it may be a naturally occurring fungal metabolite (Neill et al., 1981, 1982a; Horgan et al., 1983). No detectable conversion of the trans isomer of 1’-deoxyABA to either ABA or transABA was found. This implies that 1’-deoxy ABA is an immediate precursor of ABA in C. rosicola. Other studies have confirmed the role of 1‘-deoxy ABA as a precursor of ABA biosynthesis by C. rosicola (Neill et a/., 1987; Norman et al., 1985). Later, Al-Nimri and Coolbaugh (1991) isolated and partially characterized a cell-free extract from C. rosicola that catalyzed the conversion of a labeled 1 deoxy ABA into ABA. Most of the work on intermediates in the ABA biosynthetic pathway of fungi has involved C,, compounds with a carbon skeleton similar to that of ABA. It is strongly suggested that 4’-hydroxy-a-ionylidene acetic acid (4’-OH-a-AA) is the immediate precursor of 1’-deoxy ABA in C. rosicola (Horgan et al., 1983; Neill et al., 1987; Norman et al., 1985). When *H-labeled a-ionylidene ethanol (a-ET) and a-ionylideneacetic acid (a-AA) were supplied to C. rosicola, both compounds were converted into 1I-deoxy ABA and ABA. However, the step between FPP and a-ET remains unclear. The conversion of FPP to ABA requires transformations at several positions of the molecule (Neill and Horgan, 1983). However, the instability of intermediates makes it difficult to obtain direct evidence for their involvement in ABA biosynthesis by the fungus. A possible alternative approach would be to isolate the enzymes involved in these steps. Neill et al. (1987) reported that C. rosicola converted cis and trans 1 ’ , 4 I diols of ABA into ABA, demonstrating that the required enzymes do not exhibit a high degree of substrate specificity. Similarly, small amounts of ethyl-1 ’,4’-t-diol ABA were also detected. I-
66 MUHAMMAD ARSHAD AND WILLIAM T. FRANKENBERGER, JR.
ti h
Y
h iz
PLANT GROWTH REGULATORS IN THE RHIZOSPHERE
67
Since the first report on ABA production by C. cruenta in 1982, Oritani and his co-workers have conducted several studies to characterize the biochemistry of ABA biosynthesis by this fungus (Oritani et al., 1982; 1984; Oritani and Yamashita, 1985; 1987; Ichimura et al., 1983). Labeled MVA lactone administered to C. cruenta was incorporated into ABA (Oritani and Yamashita, 1985), revealing that MVA serves as a substrate for ABA biosynthesis. The radioactivity of labeled MVA was first incorporated into (-)4’-hydroxy-y-ionylideneaceticacid (4’-OH-y-AA), which was further metabolized to 1‘-deoxy ABA and (+)-1’,4‘dihydroxy-y-ionylideneaceticacid (1 ’,4’-diH-y-AA). However, recently, Yamamot0 et al. (1996) suggested that another isomer of 4’-OH-y-AA with an equatorial hydrogen at the 1 ’-position is the key for further oxygenation to ABA. The 1 ’,4’-diH-y-AA was converted to ABA with a high incorporation ratio by the fungus (Oritani and Yamashita, 1985; Kitagawa et al., 1995). The production of 1’deoxy ABA and 1’,4’-diH-y-AA in relatively high amounts in culture broth of the fungus was also found in other studies (Oritani et al., 1982; 1984). Oritani and Yamashita (1987) confirmed the presence of intermediates, (+)4’-OH-y-AA and 1,4’-diH-y-AA, in the culture broth of C. cruenta. y-ionylidene derivatives isolated from the mycelium of C. cruenta included cis-y-ionylidene ethanol (y-ET) and cis-y-ionylideneacetic acid (y-AA) (Oritani et al., 1985). Feeding experiments have demonstrated that the derivative 4’-OH-y-AA accumulates in the early culture period and then gradually decreases, whereas, l ’,4’-diH-y-AA and ABA are formed later. Furthermore, when labeled 4’-OH-y-AA was supplied to C. cruenfa, the radioactivity was incorporated into ABA in high yields. Based upon these observations, Oritani and Yamashita (1985) proposed that y-AA served as the first cyclic intermediate in a biosynthetic pathway of ABA in C. cruenta. It is obvious from the proposed scheme (Fig. 7) that both 1’,4’-diH-y-AA and 1’-deoxy ABA serve as immediate precursors for ABA biosynthesis. Inhibitors of carotenoid biosynthesis did not affect the accumulation ofABAby C. cruenta, so it is assumed that a direct pathway from FPP is involved (Oritani and Yamashita, 1985). However, the metabolic steps immediately after the formation of FPP are unknown. Few studies have been conducted to elucidate the biosynthetic pathway of ABA in C. pini-denstyorue (Okamoto et al., 1987; 1988a,b). Okamoto et al. (1988a) detected l’A’-t-diol ABA, 4‘-OH-a-AA, 1I-deoxy ABA, and ABA as the endogenous metabolites of this fungus. Through feeding experiments, they confirmed that 1’,4’-t-diol ABA is not a metabolite of ABA, and both 1’,4’-t-diol ABA and 1’-deoxy ABA were converted into ABA by C. pini-densiforae. However, the former served as a more efficient precursor than the latter, which indicates that the major route of ABA biosynthesis is via 1’,4’-t-diolABArather than 1I-deoxy ABA. Their results suggested that like C. rosicola, biosynthesis of ABA by C. pini-densiforae proceeds via successive oxidations of a-ionylidene intermediates (Okamoto et al., 1988a). Okamoto et al. (1988b) found that C. pini-densgora converted a-ETmore easily into 1’,4’-t-diol ABA. However, a-AA was more easily converted into 1’-
68 MUHAMMADARSHAD AND WILLIAM T. FRANKENBERGER,JR. deoxy ABA than 1‘,4’-t-diol ABA. The major in vivo route includes conversion of WET into 1’,4’-t-diolABA via (1 ’R)-4’S-OH-a-AA, and the resulting 1’-4’-t-diol ABA is oxidized to ABA. 1‘4‘4-diol ABA has also been isolated from cultures of Botrytis cinerea (Hirai et al., 1986), Cercosporu theae, C. fici (Okamoto et al., 1988a), and C. rosicolu (Neill et al., 1987). Labeled 1’,4’-t-diolABAfed to B. cinereu and C. rosicola was converted into ABA (Hirai et al., 1986; Neill et al., 1987), but labeled ABA was not converted into 1A-t-diol ABA (Hirai et al., 1986), suggesting that 1,4-t-diol ABA is not a metabolite of ABA, but a possible precursor.
V PRODUCTION OF PGRs BY RHIZOSPHERE MICROORGANISMS Since several studies have demonstrated the ability of rhizosphere microflora to synthesize PGRs in vitro, indirect inferences have been drawn about the involvement of these PGRs in plant-microbe interactions.
A. PLANT GROWTH-PROMOTING RHIZOBACTERIA The term plant growth-promoting rhizobacteria (PGPR) encompasses all bacteria that inhabit plant roots and exert a positive affect by any mechanism, ranging from a direct influence (e.g., increased solubilization and uptake of nutrients or production of PGRs) to an indirect effect (e.g., pathogen suppression, such as biocontrol, production of siderophores, or antibiotics) (Kloepper et al., 1986, 1989; Davison, 1988; Lambert and Joos, 1989; Glick et al., 1994a,b). Beneficial free-living rhizosphere bacteria are usually referred to as PGPR (Kloepper et al., 1989); or by the Chinese as “yield-increasing bacteria” (YIB). In addition to the well-known Azotobacter and Azospirillum spp., a number of other bacteria may be considered as PGPR, including various species of Pseudomonas, Acetobacter, Burkholderia, Alcaligenes, Klebsiella, Enterobacter, Herbaspirillum, Xanthomonas, and Bacillus (Brown, 1974; Elmerich, 1984; Tang, 1994; Okon and Labaudera-Gonzalez, 1994; Glick, 1995; Costacurta and Vanderleyden, 1995; Frankenberger and Arshad, 1995). Noel et al. (1996) reported that seed inoculation with Rhizobium leguminosarum effective or defective in N,-fixation significantly increased the root growth of canola and lettuce under gnotobiotic conditions and concluded that this bacterium should be considered as a PGPR. Schroth et al. (1984) recognized the importance of PGRs produced by rhizosphere microflora and the susceptibility of plant roots to bacterial colonization.
PLANT GROWTH REGULATORS IN THE RHIZOSPHERE
69
Since the majority of the PGPRs known today are capable of producing PGRs in vitro, several workers have reported that PGRs are most likely involved in plant-microbe interactions (see Frankenberger and Arshad, 1995). However, production of PGRs by these PGPR in vivo in the rhizosphere of host plants has not been investigated thoroughly. Moreover, a direct relationship between the ability of PGPR to produce PGRs and growth promotion caused by PGRs has not been found consistent. Azotobacter and Azospirillum have been thoroughly investigated among the PGPR, thus these bacteria will be discussed in detail in the proceeding section.
B. FREE-LIVING DIAZOTROPHS (AZOTOBACTER AND AZOSPIRILLUM) An extensive amount of work shows that inoculation with Azotobacter and Azospirillum spp. leads to significant positive effects on plant growth and development. Initially it was assumed that fixation of dinitrogen by these diazotrophs would provide a significant input to the nitrogen economy of plants. Stimulation of plant growth was later attributed to the production of biologically active compounds by these free-living diazotrophs (see Arshad and Frankenberger, 1993; Frankenberger and Arshad, 1995). Production of PGRs by these bacteria and alteration in plant growth were suggested as plausible explanations by several workers (Tien et al., 1979; Jain and Patriquin, 1985; Barbieri et al., 1986, 1988; Fallik et al., 1989; Avivi and Feldman, 1982; Kolb and Martin, 1985; Barbieri and Galli, 1993; Mahmoud et al., 1984; Azcdn e? al., 1978; Hussain e? al., 1987; Pati et al., 1995). It is now well established that various Azotobacter and Azospirillim spp. are capable of synthesizing PGRs, including auxins, gibberellins, and cytokinins, in culture media (Table I). The ability to produce auxins is so widespread in Azotobacter spp. that 13 of 14 (93%) strains isolated from the roots of berseem, cotton, maize, pea, and wheat produced auxins (Apte and Shende, 1981). Indole compounds and other biologically active substances have been reported in sterile soil inoculated with A. chroococcum (Elwan and El-Naggar, 1972). Similarly, several studies have demonstrated in vitro production o f auxins and other PGRs by Azospirillum spp. According to Hubbell et al. (1979), production of physiologically active quantities of IAA and other phytohormones is a phenotypic characteristic of several Azospirillum strains. TRP-derived IAA production by A. brasilense has been demonstrated unequivocally by use of HPLC, immunoassays, and GC-MS (Tien et al., 1979; Reynders and Vlassak, 1979; Barbieri et al., 1986; Crozier et al., 1987, 1988; Muller et al., 1989; Hartmann e?al., 1983).Azospirillum spp. of various strains differ in their capacity to excrete IAA in culture media, being depen-
70 MUHAMAW) ARSHAD AND WILLIAM T. FRANIENBERGER, JR. Table I Production of PGRs by Azotobacter and Azospirillurn
PGRs" Azotobacter beijernickii CLS A. beijerinckii A. chroococcum A. chroococcum A. chroococcum A. chroococcum A. chroococcum A. chroococcurn A. paspali A. vinelandii A. vinelandii A. vinelandii A. vinelandii Azotobacter spp. Azospirillum brasilense A. brasilense A. brasilense A. brasilense A. lipoferum A. lipoferum Azospirillum sp.
Method of detection
Reference
HPLC-UV, bioassay
Nieto and Frankenberger (1989a) Auxins, GLS PC, UV, bioassay Azc6n and Barea (1975) GLS, GA,, IAA PC, bioassay Brown and Burlingham (1968) GLS PC, UV, bioassay Martinez-Toledo et al. (1988) GLS PC, bioassay Salmeron et al. (1990) CLS HPLC-UV, bioassay Martinez-Toledo et al. (1988) t-Z, [9R]Z, iP, Nieto and Frankenberger HPLC-UV, bioassay (diH)[9R]Z (1989a) IAA ELISA Miiller et al. (1989) CLS, IAA, GLS PC, bioassay Barea and Brown (1974) CLS HPLC-UV, bioassay Nieto and Frankenberger (1989a) t-Z, [9R]iP TLC, ELISA, bioassay Taller and Wong (1989) IAA TLC, PC, bioassay Lee et al. (1970) IAA, GLS PC, UV, bioassay Gonzalez-Lopez et al. (1986) IAA, GLS PC, UV spectrophotometry, Mahmoud et al. ( 1984) bioassay CLS, GLS TLC, HPLC, bioassay Tien et al. (1979) iP, [9R]iP, Z Sephadex LH-20, RIA Horemans et al. (1986) GA,, iso-GA,, GC-MS Janzen et al. (1992) GA, IAA RIA Martin et al. (1989) IAA RIA Martin et al. (1989) GA,, GA,, GC-MS. GC-SIM Bottini et al. (1989) iso-GA, GLS HPLC, UV, bioassay Hubbell et al. (1979) ~
~~
For abbreviations, see Appendix.
dent upon culture conditions, growth stage, and genetic makeup and substrate (TRP) concentration (Horemans and Vlassak, 1985; Horemans et al., 1986; Fallik et al., 1989; Miiller et al., 1989; De Francesco et al., 1985; Hartmann et al., 1983;Bar and Okon, 1992; Okon ef al., 1991). Physiological and molecular stud-
PLANT GROWTH REGULATORS IN THE RHIZOSPHERE
71
ies have been conducted to elucidate the pathway(s) involved in IAA biosynthesis by Azospirillum spp. Up to now, efforts to isolate azospirilla mutants that are completely unable to synthesize IAA after chemical or Tn mutagenesis have failed (Marocco et al., 1983; Abdel-Salam and Klingmuller, 1987; Ruckdaschel and Klingmuller, 1992; Barbieri et al., 1986). This has led to the suggestion that Azospirillum possesses either more than one copy of genes involved in IAA biosynthesis or more than one IAA synthetic pathway. By using an IAA mutant and its transconjugant restored for IAA production, Prinsen et aZ. (1993) reported multiple IAA biosynthetic pathways in A. brasilense, including the IAM pathway, a second TRP-dependent pathway, and a TRP-independent pathway. Metabolites such as TAM (Hartman et al., 1983), IAA, TOL, indole-3-methanol (IM) (Crozier et al., 1988), IAM (Costacurta et al., 1992; Bar and Okon, 1992; Prinsen et al., 1993), IPyA, IAAld (Costacurta et al., 1994), and anthranilate (Katzy et al., 1990) have been detected in the culture medium of Azospirillum. TRP-dependent IAA formation by A. brasilense through the IPy A, IAM, and TAM pathways has been demonstrated by various workers. Zimmer et al. (1988) detected TRP, IAA, ILA, IAAld, IPyA, and indole-3-propionate in the culture medium of A. bradense, revealing an active transamination pathway of IAA synthesis. Crozier et al. (1988) demonstrated the presence of IAA, ILA, TOL, and IM as A. brasilense metabolites of TRP. Similarly, Ruckdiischel et al. (1988, 1990) and Ruckdaschel and Klingmuller (1992) detected intermediates of the transamination pathway of IAA synthesis in A. lipoferum. Costacurta and associates (1994) conducted studies on molecular cloning and sequence analysis of A. brasilense and found conversion of IPyA to IAAld and TOL, which clearly demonstrates that IAA production in A. brasilense is mediated by IPyA decarboxylase. The presence of the TAM pathway in Azospirillum has also been suggested (Hartmann et al., 1983; Ruckdaschel and Klingmuller, 1992). However, the enzymes and genes involved in IPyA and TAM pathways have not yet been characterized. Recent studies have also provided unequivocal evidence for the presence of the IAM pathway of IAA synthesis in A. brasilense (Prinsen et al., 1993; Bar and Okon, 1993; Costacurta et al., 1992). Costacurta et al. (1992) found that a Tn mutant (SpM79 18) of A. brasiZense Sp6 produced high amounts of IAM and very low amounts of IAA, whereas its one exconjugant SpM7918 (p0.2) synthesized five times more IAA than the mutant, preventing the accumulation of IAM. They concluded that A. brasilense contains a functional gene that encodes the conversion of IAM to IAA (Costacurta et al., 1992). Anthranilate is an intermediate in TRP biosynthesis and catabolism, and recent molecular studies have demonstrated a role of anthranilate in regulation of IAA synthesis by Azospirillum (Zimmer and Elmerich, 1991; Zimmer et al., 1991). Feeding experiments revealed that A. brasilense Sp245 with anthranilate or indole results in higher IAA production when compared with cultures supplied with TRP (Costacurta and Vanderleyden, 1995). By using different species and mutants of Azospirillum,
72 MUHAMMAD ARSHAD AND WILLIAM T FRANKENBERGER,JR. Zimmer et al. (1991) found that TRP-dependent IAA production is probably repressed by anthranilate, since A. irakense KA3 released 10 times less IAA into the medium than did A. brasilense Sp7, whereas anthranilate accumulation was 55 times greater in the former. Recently, Costacurta and Vanderleyden (1995) thoroughly discussed the biochemistry of IAA synthesis in Azospirillum spp. Although the IAM, IPyA, and TRP-independent pathways for IAA biosynthesis have been identified in Azospirillum, IAA production and its regulation need to be studied further in order to understand the role played by different IAA pathways in the interaction with plants. In addition to auxins,Azotobacter and Azospirillum also synthesize GAS(Table I). Janzen et al. (1992) reported that A. bransilense produced GAS,not only in pure culture, but also in mixed cultures (co-cultured with Trichoderma). MartinezToledo et al. (1988) reported that the addition of root exudates obtained from Zea mays stimulated the growth of Azotobacter spp. and doubled the amount of plant hormones (auxins, GAS,cytokinins)excreted in vitro by this bacterium. This stimulation in hormone production was more apparent with root exudates obtained from a 30-day-old plant than from those of a 7-day-old plant, indicating that changes in the composition of root exudates can influence microbial production of phytohormones in the rhizosphere. Interestingly, an A. chroococcum strain, H23, isolated from the root of Z. mays solubilized inorganic phosphate in nitrogen-free medium amended with 0.5% glucose. Nitrogen fixation and production of auxins, GAS,and cytokinins by this strain were correlated with the amount of phosphate solubilized (Salmeron et al., 1990). Thus, this bacterium may influence plant growth indirectly by enhancing the availability of nutrients, such as phosphorus, and directly through production of PGRs. Production of cytokinins in cultures of Azotobacter and Azospirillum is also listed in Table I. It is obvious from most investigations that cytokinins produced by these bacteria have not been characterized rigorously. Only three studies carried out by Horemans et al. (1986), Nieto and Frankenberger (1989a), and Taller and Wong (1989) provide the identity of the cytokinins produced in cultures of A. brasilense, A. chroococcum, and Azotobacter vinelandii, respectively. The mixed culture of A. brasilense and Arthrobacter giacomelloi showed higher productivity of gibberellinsand cytokininsthan the single culture (Cacciari etal., 1989).This may imply that a rhizosphere microbial population consisting of different species may stimulate phytohormone production in the presence of introduced inocula of Azospirillum or Azotobacter or both. There is only one report on Azospirillum production of ABA. By using a radioimmunoassay, Kolb and Martin ( 1985) detected ABA in the culture medium of Azospirillum brasilense Ft 326. There is substantial evidence that plants can respond to inoculation with Azotobacter or Azospirillum spp. active in production of PGRs when inoculated in the rhizosphere (Table 11).
Table II Secondary Metabolites of Free-LivingDiaztrophs and Their Effects on Plant Growth PGRs detected"
Q
w
Plant
Azotobacter spp.
IAA, GLS
Barley
A. biejernickii
IAA, GLS, CLS
Medicago
A. chroococcum
m,GA,
Tomato
A. chroococcum
IAA, GLS
Wheat
A. paspali
IAA, GLS, CLS
Lycopersicon esculentum, Paspalum notaturn, Triticum vulgare, Lactuca sativa, Centrosema pubescens, Lolium perenne
A. vinelandii
IAA, GLS, CLS
Tomato
A. vinelandii
IAA, GLS, CLS
Lavandula, Lycopersicon esculentum
Responses The metabolites showed stimulatory effect on plant height and dry weight. Cell-free supernatants and whole bacterial cultures behaved as pure hormones (LAA, GA,, kinetin) in improving dry weight and infection Exogenous application of IAA and GA, at an amount similar to those present in bacterial cultures produced an effect similar to inoculations.
Reference Mahmoud er al. (1984) Axon et al. (1 978)
Brown et al. (1968)
Five-day-old crude culture increased the root and short length, most likely through production of PGRs. Inoculation affected plant growth and development significantly. Since there was no N,fixation, the pronounced effect was attributed to PGRs (IAA, GLS, CLS).
Pati et al. ( 1995) Barea and Brown ( 1974)
Treating roots with bacterial cultures accelerated plant growth and increased the yield of fruit. Effects were most likely caused by plant hormones (IAA, GLS, CLS). Cell-free supernatants and whole bacterial cultures behaved as pure hormones in improving dry weight and infection.
Azcdn and Barea (1975)
Azcdn et al. (1978)
continues
Table II-Continued PGRs detected"
Plant
Azospirillum brasilense
IAA,ABA
Beta vulgaris ssp. iricla and wheat
A. brasilense
IAA
Maize
A. brasilense
IAA, GLS, CLS
Pearl millet
A. brasilense
IAA, GLS, CLS
Pearl millet, sorghum
Azospirillum SP.
not determined
Wheat
A. brasilense
Responses Root elongation of B. vulgaris spp. cicla was stimulated and the number of lateral roots was increased in response to inoculation. Exogenous application of IAA to wheat plants caused a similar response. Since the inocula produced IAA in high amounts, growth promotion caused by inoculation was speculated to be due to IAA excretion. Inoculated roots had higher amounts of both free and bound IAA as compared to the control. The amounts of free IAA significantly increased in the inoculated roots two weeks after sowing. GC-MS analysis revealed the presence of both IAAand IBA in the two-weekold inoculated seedling roots. Combination of IAA,GA,, and kinetin produced changes in root morphology similar to those produced by the inoculum. Combination of IAA, GA,, and kinetin produced changes in root morphology similar to those produced by the inoculum. Bacterial secretion of PGRs was the major factor affecting plant growth, whereas fixed nitrogen was minimal
IAA completely and NO, partly substituted for inoculation in an assay where the increase in dry weight of intact wheat roots was determined after an incubation for 10 days.
Reference Kolb and Martin ( 1985)
Fallik et al. (1989)
Tien et al. (1979) Hubbell et al. (1979) Avivi and Feldman ( 1982) Zimmer et al. (1988)
A. brasilense
IAA
Wheat
Maize
A. lipoferum
A. brasilense
IAA
Wheat Medicago sativa
A. brasilense
IAA
Wheat
A. brasilense (low and high IAA producing strains)
IAA
Wheat
Azornonas rnacrocytogenes
IAA, GLS
Wheat
For abbreviations, see Appendix.
Inoculation with a wild-type strain active in IAA production caused an increase in number and length of lateral roots. A Nif strain (a low producer of IAA) did not affect root development. Inoculation significantly enhanced the root growth, and GA, was identified in a free acid fraction from the inoculated roots, whereas in the noninoculated roots GA, was detected after hydrolysis, indicating that inoculation affected GA status of maize seedlings roots. Inoculation and exogenous application of IAA had similar effects on total root length to number of root hairs of wheat and on nodule numbers in Medicago sativa, suggesting that IAA is an important factor in the effects observed after inoculation. The number and length of roots were significantly increased by inoculation with IAA-producing strains, whereas strains unable to produce IAA did not cause the same effect. Treatment with different concentrations of IAA showed a dose effect on the root system development. Inoculation of wheat with the tri.5-induced mutant: A. brasilense SpM7918, a very low IAA producer, promoted less root development than did the wild-type strain Sp6. Five-day-old crude culture increased the root and shoot length, most likely through PGRs production.
Barbieri et al. (1986) Fulchieri et al. (1993)
Martin et al. (1989)
Barbieri et al. (1988)
Barbieri and Galli (1993) Pati et al. (1995)
76 MUHAMMAD ARSHAD AND WILLIAM T. F-MERGER,
JR.
c. RHZZOBZUM Several studies have demonstrated that Rhizobium spp. are capable of producing PGRs (Table 111). With the use of state-of-the-art analytical techniques, IAA production has been detected in free-living cultures of Rhizobium, concomittantly with other indole compounds, including L A , IPyA, IAld, TOL, ICA, and N-acetyl-LTRP (Wang et al., 1982; Badenoch-Jones et al., 1982a,b; Kaneshiro et al., 1983; Ernsten et al., 1987; Sekine et al., 1988). Usually, TRP serves as the physiological precursor for IAA synthesis, with a number of other metabolic products released, including IAA derivatives. However, Rhizobium spp. are also capable of synthesizing IAA in the absence of TRP, but the exogenous application of TRP increases IAA production severalfold (Dullaart, 1970; Wang et al., 1982; Wheeler et al., 1984; Ernstsen et al., 1987; Atzorn et al., 1988; Kittell et al., 1989). The transformation of TRP into IAA by rhizobia via both IPyA and IAM pathways has been demonstrated. Tan-colored mutants of R. japonicum exhibited the ability to catabolize TRP into IPyA and IAA when grown in a glutamate-limitedmedium (Kaneshiro et al., 1983). Detection of IPyA in rhizobia (R. trifolii and R. leguminosarum) extracts (Badenoch-Jones et al., 1982b) implies that IAAformation most likely involves the conversion of TRP to IAA via IAAld, with IPyA serving as an intermediate. Kittell et al. (1989) also reported IAA synthesis via IPyA (transamination) pathway in R. meliloti. Production of IAA is believed to proceed by this route in most higher plants (Bandurski et al., 1995). Ernstsen et al. (1987) reported the conversion of labeled TOL into IAA and IM by R. phaseoli. Metabolism of TRP to TOL, which is eventually converted to IAA, most likely results by a reaction involving IAAld and, thus, may be an important metabolite involved in the regulation of IAA biosynthesis (Sandberg et al., 1987; Sembdner et al., 1980). Metabolic studies with rhizobial cultures in the presence of isotopically labeled substrates have demonstrated the conversion of L-TRPto IAA, TOL, and IM; TOL to IAA and IM; and IAA to IM (Emstsen et al., 1987). Labeled [3H]IAM was not converted to IAA nor was it detected as a metabolite of L - [ ~ - ~ H TRP ] (Ernstsen et al., 1987). Contrarily, Sekine et al. (1988,1989) have demonstrated the ability of Bradyrhizobium spp. to convert TRP into IAA via IAM.By using a system involving conversion of a-naphthalene acetamide (NAM) to a-naphthalene acetic acid (NAA)-a process analogous to the conversion of IAM into IAA-Sekine et al, (1988) screened several species of Rhizobium and Bradyrhizobium for IAM hydrolase activity. They found that Bradyrhizobium spp. were capable of transforming NAM into NAA, and, by using GC-MS analysis, they confirmed the presence of IAM and IAA in a TRP-supplemented culture medium of B. japonicum, but no IAM was detected in rhizobia cultures. This study strongly suggests that production of IAA in Bradyrhizobium spp. involves the same pathways that operate in P. syringae pv. savastanoi and A. tumefaciens. In another study, by using molecular cloning of genes for IAM-hydrolase, Sekine et al. (1989) confirmed the presence of this pathway of IAA formation from TRP in B. japonicum. The structural gene
Table 111
PGRs Detected in Rhizobium Cultures
Rhizobium species Bradyrhizobium japonicum B. japonicum
(5-methyltryptophan resistant mutants) R. japonicum (tan-mutant) (wild-type) R. leguminosarurn
(nodulating and nonnodulating) R. legurninosarum (nodulating and nonnodulating) R. leguminosarum (N,-fixing and nonN,-fixing) R. leguminosarum (nodulating and (nonnodulating) R. lupini
Method of detection
Product" NAA
IAM,IAA TRP, IAA, IAA IPyA, M A IAA
HPLC, GC-MS HPLC, GC-MS GC
Sekine et al. (1988)
TLC-UV, GLC-MS TLC-UV. GLC-MS
Kaneshiro et al. ( I 983) Kaneshiro et al. (1983) Badenoch-Jones et al. (l982b)
IAld, TOL, ICA, HPLC, GC-MS IGA, ILA, IGoxA, N-Ac-TRP, IPyA, IAA IAA GC-MS
Badenoch-Jones et al. (1983)
IAA
HPLC, GC-MS
Wang et al. (1 982)
IAA, TOL
PC, TLC, spectrofluorometry HPLC-UV
Dullaart (1 970)
HPLC, GC-MS IAld, IM HPLC, EIA, GC-MS
Ernstsen et al. (1987) Atzorn et al. (1988)
Wheeler et al. ( 1984) Ernstsen e f al. ( 1987) Bulard et al. (1963) Badenoch-Jones et al. (1982b)
R. phaseoli
IAA
(nodulating, nonnodulating, N,fixing, non-N,fixing) R. phaseoli
IAA
HPLCkcintillation
IAA, TOL, IM, IAld IAA, IAN
GC-MS
IAAld, TOL, ICA, IGA, ILA, IGoxA, N-Ac-TRP, IPyA, IAA
HPLC, GC-MS
R. trifolii
(nodulating and nonnodulating)
Badenoch-Jones et al. (1982a)
TLC-HPLC/ scintillation
IAA
R. trifolii
Hunter and Kuykendall(1990)
IAAsp
R. meliloti (wild type and mutants) R. phaseoli
R. phaseoli
Reference
IAA, TOL
PC
Kittell et al. (1 989)
continues
77
Table III-continued Method of detection
Productn
Rhizobium species
Reference
R. trgolii
IAA
GC-MS
(nodulating and nonnodulating) R. japonicum
Badenoch-Jones et al. (1982a)
GLs, GA,
Pc,TLC,
Katznelson and Cole (1965) Katznelson and Cole ( I 965) Katznelson and Cole ( 1965) Atzorn ei al. ( 1988) Williams and de Mallorca (1982) Sturtevant and Taller (1989)
R. leguminosarum R. meliloti
bioassay PC, TLC, bioassay PC, TLC, bioassay HPLC, RIA, GC-MS TLC, bioassay
GLs, GA,, GA, GLs, GA,
R. phaseoli
R. trifolii
B. japonicum
t-[9R]Z [2MeS 9R]Z, iP
Bradyrhizobium sp. (lupini) Rhizobium spp.
iP, Z, [MeSIZ
Rhizobium spp.
R. fredii
iP, [9R]iP, Z, [9R]Z, C-Z, c-[9R]Z, (diH)Z, (diH)[9R]Z iP, WeSlZ, Z
R. meliloti
iP, CLS
R. meliloti
[9R]iP, [MeSIZ
R. leguminosarum
iP, 2, [MeSIZ
R. loti
iP, [9R]Z
R. phaseoli
iP, 2, [MeSIZ
R. trifolii
iP, [9R]iP, [9R]Z, [MeSIZ
a
Sephadex LH-20, ELISA, HPLC-UV, bioassay Sephadex LH-20, bioassay Sephadex LH-20, bioassay GC-MS, radioimmunoassay
iP, [9R]iP
Sephadex LH-20, bioassay
pc, radioimmunoassay Sephadex LH-20, bioassay Sephadex LH-20, bioassay Sephadex LH-20, bioassay Sephadex LH-20, bioassay Sephadex LH-20, bioassay
For abbreviations, see Appendix.
78
Taller and Sturtevant (1991) Sturtevant and Taller (1989) Upadhyaya etal. (1991b)
Sturtevant and Taller (1989) Hua et al. ( 1980) Taller and Sturtevant (1991) Taller and Sturtevant (1991) Taller and Sturtevant ( 1991) Tailer and Sturtevant (1991) Taller and Sturtevant (1991)
PLANT GROWTH REGULATORS IN THE RHIZOSPHERE
79
for IAM hydrolase (bum)has a high degree of similarity with iaaH of l? syringae pv. savastanoi and the tms-2 gene of A. tumefacians. Activity of the IAM pathway was not, however, observed in fast-growing Rhizobiurn spp. (Ernstsen et al., 1987; Selune et al., 1988). One possible explanation for this could be that the genes for the IAM pathway in Rhizobium spp. are suppressed in the free-living state in the absence of the plant symbiont. Recently, Kawaguchi et al. (1990) investigated the validity of this hypothesis and supported the premise that the activity of IAM hydrolase is suppressed in the free-livingstate in R. Zeguminosarum biovar viciae K5. They detected IAM by HPLC in a suspension of 5-methyl TW-resistant variants supplemented by a large amount of TRP (Kawaguchi et al., 1990).The absence of the IAM pathway in most plants and its presence in bradyrhizobia could be of great use in understanding the role of this bacterial endophyte in nodulation and its relative contribution in elevated levels of auxins in nodules. Rhizobia are also capable of catabolizing auxins, particularly IAA. Some evidence suggests that the following transformation may occur: IAA -+ IM -+ IAld -+ ICA (Sandberg et al., 1987). In R. phaseoli, IM was the major metabolite of [2’-l4C]1AAdegradation, whereas IAld, being an unstable intermediate, did not accumulate. ICA has been identified in extracts of R. trifolii and R. leguminosarum grown in a TRP-supplemented medium (Badenoch-Jones et al., 1982b). However, ICA was not detected in R. phaseoli as either an IAA catabolite or as an endogenous indole metabolite. Thus, further work is needed to determine the catabolic pathway of IAA in rhizobia. Rhizobium spp. have not been investigated extensively for their ability to synthesize GAS.Only a few studies report the presence of GASas secondary metabolites of rhizobia in culture media (Table 111). In most studies, bioassays have been employed to detect GAS after separation by paper chromatography (PC) or thinlayer chromotography (TLC), and, only recently, Atzorn et al. (1988) for the first time used state-of-the-art techniques (HF’LC, radioimmunoassay [RIA], GC-MS) to detect GASin cultures of various strains of R. phaseoli. They successfully identified GA, and GA, as the major GAS and GA, and GA,, as minor constituents excreted by wild-type and mutant strains effective or defective in nodulation and nitrogen fixation. These results imply that rhizobia possess the genetic makeup and metabolic activity to synthesize GAS. Various Rhizobium spp. have been found capable of synthesizing cytokinins in culture media (Table IN),but many of these substances have not been well characterized. Sturtevant and Taller (1989) reported the speciation of cytokinins present in cultures of B. japonicum by using HPLC-UV and enzyme-linked immunosorbent assay (ELISA) as indicated in Table III.Although various Rhizobium spp. differ in their efficiency in producing cytokinins (Phillips and Torrey, 1972;Sturtevant and Taller, 1989; Taller and Sturtevant, 1991), the significance of these differences for nodule formation is not known. By using a bioassay, HPLC, or an immunoassay, Taller and Sturtevant (1991) analyzed the cytokinin content in culture media of
80 MUHAMMAD ARSHAD AND WILLIAM T. FRANKENBERGER,JR. eight major cross-inoculationgroups of rhizobial strains. They found that all strains examined produced at least two cytokinin-active compounds, with total cytokinin activity ranging from one to several micrograms of kinetin equivalents per liter of culture filtrates. There were both qualitative and quantitative differences between the rhizobial species. Similarly, Upadhyaya et ul. (1991b) found substantial differences in the ability of two Rhizobium strains (ANU240 and IC 3342) in releasing iP and Z in their culture supernatants, analyzed by GC-MS and RIA.Puppo and Rigaud (1978) believe that the presence of both symbiotic partners, the plant and bacteria, are required for cytokinin synthesis, i.e., a direct consequence of symbiosis. They failed to detect any cytokinin activity in hydroponic cultures without the symbiotic partner. Akiyoshi et ul. (1987) could not confirm cytokinin production by rhizobia. Sturtevant and Taller (1988) also speculated that other plant compounds may influence cytokinin production by rhizobia, in terms of speciation and amounts. They found that 2-methylthiozeatin ([2MeSJZ) increased approximately 20-fold when B. juponicum was grown in the presence of a soybean extract. Similarly, in another study, Taller and Sturtevant (1991) observed that, in a defined medium, B. juponicurn produced ribosylzeatin ([9R]Z), Z and [MeSIZ. With the addition of a soybean extract, Z disappeared and there was a significant increase in [MeSIZ. An analysis of the control medium containing the soybean extract showed mostly [9R]Z and a trace of Z but no [MeSIZ. Thus, the additional [MeSIZ did not come directly from the soybean extract. They proposed that cytokinins in the soybean extract may have been derived from precursors for cytokinin synthesis by B. juponicum. However, they confirmed that [9R]Z was not metabolized by B. juponicurn. This may imply that when B. rhizobiurn is in symbiosis, its ability to produce cytokinins may be altered both quantitatively and qualitatively. Beard and Harrison (1992) reported the ability of Rhizobiurn sp. to produce C2H4and detected ACC in the culture medium. Exogenous application of ACC resulted in a large burst of C,H4 in the rhizobial culture, indicating an active EFE system. However, no other study has confirmed the ability of Rhizobiurn to derive C,H, from ACC.
1. Production of PGRs, Nodulation, and Plant Growth A root nodule is a unique and highly organized structure that develops as a result of the symbiotic relationship between a microsymbiont (e.g., Rhizobium) and plants (e.g., legume). This symbiosis is essential for the supply of atmospheric N to plants fixed by the rhizobia. It has been claimed that 90% of the plant’s requirements of N may come through this association (Drevon, 1983). It has been over 50 years since it was first proposed that PGRs (e.g., auxins) secreted by the microsymbiont Rhizobium may be a regulator in the development of N-fixing nodules. Recent work has been conducted to characterize PGR synthesis by rhizobia in vitro, but the importance of microbially derived PGRs in nodule development
PLANT GROWTH REGULATORS INTHE RHIZOSPHERE
81
is still subject to controversy. Recently Hirsch and Fang (1994) have critically reviewed the relationship between PGRs and nodulation. Based upon information regarding the presence of PGRs in nodules and the role of inhibitors of hormone transport in inducing nodule-like structures, they proposed a regulatory scheme indicating the involvement of PGRs in nodule development (Fig. 8). a. Auxins Wang et al. (1982) detected IAA in the culture medium of nodulating and nonnodulating strains of R. leguminosarum by GC-MS, even in the absence of an exogenous supply of TRP, but they could not establish any causal relationship between the ability to nodulate peas and production of IAA. Badenoch-Jones et al. (1982a) further confirmed the findings of Wang ef al. (1982) by demonstrating the presence of IAA in the culture supernatants of several rhizobia strains, including those mutants defective in various stages of nodule formation. Flavonoids activate nod gene expression in rhizobia, resulting in the synthesis of nod signals that trigger nodule organogenesis in the host plant (Peters et al., 1986; Zaat et al., 1987; Phillips et al., 1994). Prinsen et al. (1991) have demonstrated that nod-inducers (flavonoids) also stimulate the production of IAA,suggesting that nodule morphogenesis may be controlled by the highly specific nod signal in combination with phytohormones, such as auxins, released by rhizobia. This premise has been supported by others, who hypothesized that the changes in the phytohormone balance are a necessary requirement to elicit nodule formation (Hirsch et al., 1989; Long and Cooper, 1988) and that expression of nod genes affects the phytohormone balance (Schmidt et al., 1991). Hirsch et al. (1989) suggested that auxin transport inhibitors mimic the activity of compound(s) made after the induction of rhizobial nod genes. These molecular studies confirm the undefined role of auxin in nodule organogenesis. There is now substantial evidence that the nodules of the leguminous plants Erythrina indica, Sesbania grandgora, Pterocarpus santalinus, Crotalana retusa, Clitoria tematea, Lmse esculenta, Phaseolus aureus Roxb. var. mungo, and Crotaria refuge all contain higher levels of MA, GAS, and cytokinins than do their respective roots (Bhowmick and Basu, 1984; Bhattachruyya and Basu, 1991; Roy and Basu, 1991).The origin of these high amounts of MA in root nodules is still controversial. A mutant of R. japonicurn (producing 30 times more auxin than the wildtype strain) significantly enhanced nodulation of Glycine m a ,whereas the parental strain required supplementation with sucrose, CaCO,, EDTA, nicotinic acid, and glutamate for effective nodulation (Kaneshiro and Kwolek, 1985). Milic et al. (1993) compared the effects of two strains of B. japonicum (strain 2b produced -8 times more IAA than did strain 1) on growth, nodulation, and nitrogenase activity of soybean. They reported that inoculation with strain 2b resulted in better growth and higher mass production,with more nitrogenase activity and Ncontent in plant tissues. Although the number of nodules per plant was less, nodule
82 MUHAMMAD ARSHAD AND WILLIAM T. FRANKEMERGER, JR.
nodulation stops Shoot-derived autoregulatory factor Nodule primordium
Figure 8 Possible involvement of various PGRs in nodulation. Cytokinin is produced in the root tips. Auxin, produced in the shoot tip, is basipetally transported down the stem. Rhizobia synthesize the nod factor, which affects both root hair deformation and cell divisions that foreshadow the nodule primordia. Although rhizobia also produce auxin, cytokinins, and gibberellins, it is not clear whether any of these are required for nodule primordia formation. Exogenous cytokinin can elicit nodule formation, and nodules contain high levels of cytokinin and auxin. Although auxin is involved in some way with nodulation, its role is as yet unclear. Ethylene and shoot-derived autoregulatory factors inhibit nodule development. The stimulus to produce the shoot-derived factor is synthesized in the root and translocated to the shoot. During pod fill, nodulation stops (from Hirsch and Fang, 1994, used with permission).
PLANT GROWTH REGULATORS IN THE RHIZOSPHERE
83
dry weight was greater in plants inoculated with strain 2b than with strain 1. Interestingly, when the eluate of IAA spots from TLC plates obtained from strain 2b were added to strain 1 inoculum, plant growth was dramatically increased (2.25-fold increase in dry weight). They concluded that JAA and other PGRs have a significant role in the effectiveness of B. juponicum strains by affecting growth, nodulation, and N,-fixation. Studies with B. juponicum TRP auxotrophs show that those with blockages in the TRP pathway before tryptophan synthase do not nodulate (Wells and Kuykendall, 1983). Also, mutants of B. japonicum with enhanced TRP catabolism stimulated plant growth and nodulation when applied to soybean seedlings grown in aerated water, most likely by supplying required growth components (Kaneshiro and Nicholson, 1989). These catabolic mutants characteristically produced large amounts of indole compounds during degradation of TRP (Kaneshiro et al., 1983; Hunter and Kuykendall, 1990), and it has been suggested that these indoles are involved in enhanced nodulation (Kaneshiro et al., 1983). After screening various mutants of B. juponicurn altered in TRP metabolism, Hunter (1994) found that one strain accumulated higher amounts of anthranilic acid, IPyA, ILA, and IAA and increased the number and mass of nodules in soybean compared to the wild-type strain, which produced very low amounts of these compounds. Hunter ( 1989) examined the bacteroids from nodules containing low amounts of IAA of plants inoculated with a parental strain (a poor producer of IAA) and from those containing large amounts of IAA inoculated with methyltryptophanresistant B. juponicurn mutants (a rich producer of IAA). He found that IAA production by the wild-type bacteroids was 90 pmol mg- protein h-' and that of the mutant bacteroids was 620-690 pmol mg-' protein h-'. This indicates that bacteroid-produced IAA accumulates in the nodules. This suggests that nodular bacteroids produce IAA symbiotically and that bacterially-produced IAA accumulates in the nodule (Hunter, 1989). However, the possibility that relatively large quantities of IAA in nodules might also result from alterations in indole metabolism by the plant tissue itself in response to the rhizobial infection cannot be excluded. From this perspective, nodule formation might be analogous to the crown-gall disease and other phytopathogenic phenomena (Frankenberger and Arshad, 1995).
'
b. Gibberellins The role of GAS in nodule formation is still uncertain. Studies have shown that nodules of legumes, including Phaseolus vulguris and Pisurn sutivum (Radley, 1961), Lupinus luteus (Dullaart and Duba, 1970), Glycine mux (Williams and de Mallorca, 1982), Phaseolus lunutus (Evensen and Blevins, 1981; Dobert et al., 1992b), Phaseolus aureus (Dangar and Basu, 1987), Lens esculenta (Dangar and Basu, 1984). Crotalaria retusa L. (Bhattacharyya and Basu, 1991), and Vignu unguiculuta (Dobert et al., 1992a,b), contain substantially higher amounts of GAS than the roots. Based on the ability of R. juponicum to synthesize and release GAS in culture medium and the greater concentration of GAS in root nodules than in
84 MTXUMMAD ARSHAD AND WILLIAM T. FRANKENBERGER, JR. roots, Williams and de Mallorca (1982) suggested that rhizobial GA production may contribute to the nodule GA content. Other studies have shown that GAS have little role in nodulation. Dangar and Basu (1987) reported that the formation and growth of nodules in I? aureus were not controlled only by the nodules hormones. Atzorn et al. (1988) reported no difference in the GA content of nodules and roots of F! vulgaris when detected by HPLC, RIA, and GC-MS, although the microsymbiont was capable of producing GASin pure culture. Studies with lima bean (Phaseolus lunatus L.) and cowpea (@pa unguiculata) inoculated with two strains of Bradyrhizobium (strain 127E14,which lacks constitutive nitrate reductase activity, and strain 127E15,which contains constitutive nitrate reductase activity) provided evidence that root nodules may contribute GAS to the host plant and evoke a plant response (Evensen and Blevins, 1981; Triplett et al., 1981; Dobert et al., 1992a,b,c). Triplett et al. (1981) found that 5 weeks after inoculation, plants treated with Bradyrhizobiurn strain 127814 were significantly taller (due to internodal elongation) and had a greater number of leaves than those inoculated with strain 127E15. This response appeared to be hormone-mediated. Application of GA, to the apex of plants inoculated with strain 127E15caused an increase in plant height similar to that observed in untreated plants inoculated with strain 127E14. Furthermore, application of CCC and AMO-1618 (inhibitors of GA biosynthesis) inhibited 127E14-inducedinternodal elongation (plant height). Triplett et al. (1981) concluded that increased growth caused by inoculation with strain 127E14 could be the result of increased GA synthesis in lima bean nodules. However, they did not test for in vitro production of GASby this strain. Similar results were obtained by Evensen and Blevins (1981) in which they reported up to 50 times greater quantities of GA-like substances in nodules of plants inoculated with 127E14 compared with those inoculated with 127E15. Similarly, Dobert et al. (1992a) observed that cowpea plants inoculated with strain 127E14were 23% taller than those inoculated with strain 127E15after six weeks of growth. Cowpea plants treated at the apex with exogenous GA, or GA, responded by increasing internodal length when compared with controls. They indicated that, as in lima beans, the rhizobial-induced growth response observed in cowpeas may be in response to an imbalance in the levels of GA-like substances within the plants. GibberellinsGA, , GA,, GA,, GA,,, GA,,, GA,,, and GA, were found in nodule tissue associated with strain 127E14. In another study, Dobert et al. (1992b) found that lima bean plants inoculated with Bradyrhizobiurn strain 127E14 displayed a period of marked internodal elongation that was not observed in plants inoculated with other compatible bradyrhizobia, including strain 127815. When strain 127E14 nodulated an alternative host, cowpea, a similar, although less dramatic, growth response by the bacteria was observed. By using deuterated internal standards and GC-MS, they quantified the levels of GA,, GA,,, GA,,, and GAU in nodules and stems of two varieties of lima bean (bush and pole) and one variety of cowpea that were inoculated with either strain 127814 or 127E15. In nodules formed by strain 127E14 on
PLANT GROWTH REGULATORS IN THE RHIZOSPHERE
85
lima bean, endogenous levels of GA,, and GA,, were 10 to 40 times higher (35 to 88 ng g-’ dry weight) than those found in nodules formed by strain 127E15 (2.2 to 3.9 ng g-’ dry weight). Relative amounts of GA, were also higher (4- to 11-fold) in 127E14 nodules, but this increase was less pronounced. The rhizobial-induced increase of these GASin the nodule occurred in both pole and bush varieties of lima bean and seemed to be independentof host morphology.Regardless of rhizobial inoculum, levels of the “bioactive” GA, in the nodule (0.3-1.1 ng g-’ dry weight) were similar. In cowpea nodules, a similar, although smaller, difference in GA content due to the rhizobial strain was observed. The concentration of GA, in lima bean stems was generally higher than that observed in the nodule, whereas concentrations of other GASmeasured were lower. In contrast with the nodule, GA concentrations in lima bean stems were not greater in plants inoculated with strain 127E14, and in some, the slower growing plants inoculated with strain 127E15 actually had higher levels of GA, ,, GA,,, and GA,. Thus, there were major differencesin concentrations of the precursors to GA, in nodules formed by the two bacterial strains, which were positively correlated with the observed elongation growth. These results support the hypothesis that the rhizobial strain can modify the endogenous status of the symbiotic system. This alteration in GA balance within the plant, presumably, underlies the observed growth response. By using HPLC and GC-MS, Dobert et al. (1992~)confirmed the presence of GAS GA,, GA,, GA,,, GA,,, GA,,, and GAM in root nodules formed by Bradyrhizobiurn strain 127E14 on lima bean. Gibberellins GA,, GA,, GA,,, GA,,, and GA, were also identified in the lima bean stem tissue. c. Cytokinins The origin, role, and fate of cytokinins and their involvement in nodule development and formation have been studied thoroughly, but the results are still inconclusive. Circumstantial evidence suggests that cytolunins may play a critical role in root nodule development and functioning. Evidence supporting this hypothesis includes: (1) higher cytokinin activity in nodules than in roots; (2) the ability of the microsymbionts to release cytokinins in culture media, (3) initiation of polyploid mitosis in the presence of auxins, a characteristic of an early phase in nodule development; and (4) the formation of pseudonodules or induction of early nodule formation in response to externally applied cytokinins. Root nodules of both legumes and nonlegumes generally contain elevated levels of cytokinins. Recent studies have demonstrated that the microsymbiont Rhizobium can contribute substantially to the elevated cytokinin pool of the host (Upadhyaya et al., 1991a,b). Upadhyaya et al. (1991a,b) compared Rhizobium strain IC3342 with that of the normal strain ANU240 for their ability to produce cytokinins by using HPLC and radioimmunoassay (RIA) in culture media. They found that IC3342 was highly efficient in producing iP and Z, which were 26 and 8 times greater, respectively, than the concentrationsproduced by ANU240. Upadhyaya et al. (1991b) also analyzed the xylem sap of pigeon pea nodulated by the
86 MUHAMMAD ARSHAD AND WILLIAM T. FRANKENBERGER,JR. IC3342, a Tn5-induced curl- mutant (ANU3003), a normal wild-type cur- strain IHP(100), and a nonnodulated nitrate control plant. They found that [9R]Z and (diH)[9R]Z concentrations in xylem sap from plants with leaf curl symptoms were 7 to 9 times higher than those in the sap from nodulated plants showing no symptoms. The sap from plants (showing no symptoms) nodulated by a curl- mutant had [9R]Z and (diH)[9R]Z concentrations comparable with those in the normal plants. However, the Z content varied considerably in the sap of plants inoculated with ANU3003 and IHP100. They proposed that it is likely that such a level of [9R]Z and (diH)[9R]Z overproduction would be sufficient to cause abnormal effects on shoot development. They further suggested that the difference in the type of cytokinins released by strain IC3342 within the culture supernatants and that detected in the plant sap inoculated with this strain might be due to modification (reduction and especially ribosylation) of rhizobial-produced cytokinin bases in the nodules or root tissues. Similarly, Kumar-Rao et al. (1984) found that effective nodulation (caused by Rhizobium strain IC3342) is a prerequisite for the development of a leaf curl syndrome in pigeon pea. It was suggested that rhizobia upset the hormonal balance, and overproduction of cytokinin was responsible for this abnormal growth symptom. Cytokinins have also been shown to specifically induce the expression of Enod2 (early nodulation genes) in Sesbania roots (Dehio and deBruijin, 1992). Cooper and Long (1994) demonstrated the morphogentic rescue of Rhizobium meliloti nodulation mutants by t-Z secretion. They reported that a Rhizobium mutant harboring pTZS plasmid carrying a constitutive t-Z secretion (tzs) gene from A. tumefaciens T37 suppressed the nod- phenotype of Rhizobium nodulation mutants such that the mutant harboring pTZS stimulated the formation of nodule-like structures on alfalfa roots. Both the pattern of induced cell division and the spatially restricted expression of an alfalfa nodule-specific marker gene (MsEnod2) in pTZS-induced nodules support the conclusion that localized cytokinin production produces a phenocopy of nodule morphogenesis. Recently, Bauer et al. (1996) conducted a comprehensive study to understand signals in nodule initiation. For this, they generated a transgenic alfalfa carrying the promoter of an early nodulation gene, MsEnodl2A, fused to the reporter gene, gusA. They reported that inoculation with R. meliloti or treatment with purified nod factors induced the GUS activity in the cell division foci of the inner cortex, and GUS-stainingpatterns in nodules and roots were in agreement with the activation of the endogenous MsEnodl2A gene. Treatment of roots with purified nod factors and cytokinins induced similar patterns of cortical cell division, GUS staining, and amyloplast accumulation. Like the nodfactor responses, the cytokinin responses required photosynthesis and limiting combined nitrogen supply. Thus, cytokinins and nod factors may share elements of their signal transduction pathways to the inner root cortex. Based upon these observations, they proposed a model (Fig. 9) of the possible involvement of cytokinins in coordinating plant metabolism with nodule initiation (Bauer et al., 1996).
PLANT GROWTH REGULATORS IN THE RHIZOSPHERE Absence of nitrate
Cytokinlns
NOD factors
I
/
\
I
Sensitizationto cytoklnin action
87
e
#'
Rhlzodermis
\
Endodermis Pericycls ~
~~~
Photosynthesis-
--
-
Saccharose
vascular tissua
Figure 9 Model representing possible links between nod factor and cytokinin actions and the carbon-nitrogen metabolism during nodule initiation. A schematic view of a longitudinal root section with divided cortical cells is represented. Root cortical cells are sensitized in the absence of combined nitrogen, thus allowing exogenously applied nod factor and cytokinins to elicit cortical cell divisions, amyloplast deposition, and early nodulin gene expression. MsEnodlZA probably is indirectly induced once cortical cell division starts. It is not known whether cytokinins act successively or in parallel to nod factors. Carbon-nitrogen metabolites accumulating in the root under nitrogen limitation and/or plant hormones indirectly induced by the metabolic changes may further control induction of cortical cell division. Dashed arrows indicate multiple events that may occur (from Bauer et al., 1996, used with permission).
d. Ethylene In a recent study, Hunter (1992) compared C2H4 formation and ACC concentrations in inoculated (B. japonicum) vs uninoculated soybean plants grown hydroponically. He reported that the presence of nodules on the roots increased the amount of C,H, produced by the soybean. Similarly,higher levels (greater than 2fold) of ACC were found in nodulated roots than in uninoculated, and ACC decreased in the roots when nodules were removed. The nodules contained the highest concentration of ACC (greater than 4-fold) compared with the uninoculated roots. Nodules formed with an ineffective B. japonicum strain had less ACC then an effective strain. Hunter (1992) concluded that nearly all (greater than 95%) of the C,H, produced by nodules was of plant origin, with little produced by the bacteroids. Similarly, C,H, production and ACC concentrations were also evaluated
88 MUHAMMAD ARSHAD AND WILLIAM T. FRANKENBERGER,JR. in inoculated vs uninoculated bean (Phaseoh vulgaris L.) seedlings (Beard and Harrison, 1992). Preliminary results indicated that the pattern of C,H4 and ACC production in the various segments was similar for inoculated and uninoculated plants. However, the level of a conjugated metabolite of ACC, N-malonyl-ACC (MACC), was greatly increased in nodulated seedlings. They reported the ability of Rhizobium sp. to produce C,H4 and detected ACC in the culture medium. Exogenous application of ACC resulted in a large burst of C2H, in the rhizobial culture, indicating an active EFT system. Given these observations, Beard and Harrison (1992) suggested that nodulation may not change the overall C2H4 biosynthesis in bean seedlings but may perhaps alter the level of conjugation of ACC to MACC. It is not yet known whether rhizobial-infected nodules contribute C2H4 or ACC to the plant tissues. However, it is interesting that this Rhizobium spp. has the capacity to produce C2H4by the same biosynthetic pathway as found in higher plants (Beard and Harrison, 1992). By using inhibitors of C,H4 formation or action (aminoethoxyvinylglycine [AVG], Ag+, Co2+),many workers investigated whether or not endogenous C,H4 regulates nodulation. Several workers found that the exogenous application of these compounds resulted in increased nodulation and decreased endogenous C2H, synthesis, revealing a role of C2H, in nodule formation (Peters and CristEstes, 1989; Fearn and LaRue, 1991; Guinel and LaRue, 1991, 1992; Lee and LaRue, 1992a,b,c; Stokkermans et al., 1992). On the other hand, a study conducted by Stokkermans et al. (1992) revealed that the use of AVG, Co2+,and Ag+ failed to restore nodulation in a soybean line carrying the Rj4 allele, which restricts nodulation upon inoculation with B. japonicum strain USDA 6 1. Generally, it is considered that either C2H4suppresses nodulation or has no apparent role in nodule formation. However, recently, Arshad et al. (1993) reported that exogenous application of L-MET (a precursor of C,H4) to soil significantly influenced nodulation of Albizia lebbeck L. They found 1.88- and 4.48-fold increases in nodule numbers and nodule dry weight, respectively, compared with the g kg-' soil of L-MET.They attribcontrol, in response to the application of uted this positive effect of L-MET to be microbially derived C2H4 within the rhizosphere or an interaction of L-MET-dependent microbially-derived C,H4 with other phytohormones. e. Abscisic Acid Like other PGRs, it is unclear whether ABA plays a regulatory role in nodule formation. However, several studies have shown that ABA levels in nodules are usually much greater than the ABA contents of roots. Williams and de Mallorca (1982) detected cis, trans-ABA in roots and root nodules of soybean by GC and a wheat embryo bioassay. They reported that ABA was 12- to 15-fold more concentrated in the nodule tissue (2.21 nM g-' dry weight) than in the root system (0.18 mMg- I). Similarly,higher levels of ABA were found in developing pea root
PLANT GROWTH REGULATORS IN THE WIZOSPHERE
89
nodules than in uninfected roots (Charbonneau and Newcomb, 1985). Basu and co-workers reported that root nodules of Erythrina indica, Sesbania grand$ora, Pterocarpus santalinus, Lense esculenta, Clitoria ternatea, Crotaliaria retusa, Samanae saman, and Phaseolus aereus Roxb. var. mungo contained elevated levels of ABA-like substances and other phytohormones when compared with their respective roots (Bhowmick and Basu, 1984; Dangar and Basu, 1984, 1987; Chattopadhyay and Basu, 1989; Roy and Basu, 1992; Bhattacharyya and Basu, 1991). The level of ABA-like substance increased with age, and, unlike other phytohormones that were maximum in mature nodules, it was highest in the old nodules (Dangar and Basu, 1984, 1987; Bhattacharyya and Basu, 1991) since ABA is a senescence-phasehormone. The ability of a microsymbiont to synthesize ABA has not been investigated thoroughly except for studies conducted by Williams and de Mallorca (1982) and Dangar and Basu (1987). By using GC and a bioassay, Williams and de Mallorca (1982) reported that R. japonicum was unable to produce ABA in culture medium, whereas Dangar and Basu (1987) detected ABA and other phytohormones produced by a Rhizobium culture by TLC and a bioassay. However, chromatography and bioassays used for detection of ABA do not provide unequivocal proof that rhizobia is capable of producing ABA. Elevated levels of ABA found in root nodules might be obtained directly from the host. A number of plant-microbe interactions are possible: (1) infection by the microsymbiont somehow alters the metabolism of endogenous ABA; (2) ABA is released from its conjugates more actively, resulting in more accumulation within the nodule tissue; or (3) ABA may be transported from other parts of plants to the nodules. Alternatively, the host may provide a precursor of ABA to the microsymbiont; this has not yet been investigated in culture medium.
D. MYCORRHIZAL SYMBIOSIS Mycorrhizae involves a unique symbiotic association between plant roots and infecting fungi. This association often increases growth and yield of many crops by enhanced nutrient uptake, resistance to drought and salinity, and increased tolerance to pathogens (Rhodes and Gerdemann, 1978a,b,c;Hayman, 1983; Dahne, 1982; Abbott and Robson, 1982; Bethlenfalvay et al., 1985; Meyer and Linderman, 1986; Kucey, 1987). However, it has been suggested that PGRs released by the infecting fungi may also contribute to enhanced plant growth. Recently, Beyrle (1995) has comprehensively reviewed the role of phytohormones in the function and biology of mycorrhiza. Several studies have revealed that auxin production is widespread among many mycorrhizal fungi (Table IV). By using state-of-the-art analytical techniques, some recent studies have unequivocally confirmed the ability of mycorrhizal
Table IV PGRs Detected in Cultures of Mycorrhizal Fungi
Fungi Amanita caesaria
IAA
A. frostiana A. muscaria A. rubescens Basidiomycetes spp. (nonsporulating) Boleyus badius B. californicus
IAA IAA IAA Auxin-like
B. felleus B. granulatus B. luieus B. varietano Cephanosporum ucremonium C. glutineum Cenococcum graniforme Hebeloma anthracophilum H. circinans
H. crustulinifurme H. c,ylindrosporum ( I 1 isolates)
H. edurum
H. hiemale H. longicauduni H. mesophaeum
H. rudicicosum H. sarcophyllum
H. sinapizans
Method of detection
Auxina
IAA IAA IAA IAA IAA
PC, bioassay PC, bioassay GC-MS PC, bioassay PC, bioassay GC-MS PC, bioassay PC, bioassay PC, bioassay PC, bioassay PC, bioassay
IAA, indoles auxin-like
PC, bioassay
auxin-like
PC, bioassay
IAA
GC-MS
IAA, ICA, IAld IAA, ICA, IAld IAA
HPLC
IAA, ICA, IAlD IAA, ICA, IAld IAA, ICA IAld IAA IAA, ICA, IAld IAA, auxin-like IAA, ICA, lAld IAA, ICA, IAld IAA, ICA, IAld
HPLC GC-MS HPLC HPLC-MS HPLC HPLC, MS HPLC GC, bioassay HPLC HPLC HPLC
Reference Ulrich ( 1960) Ulrich (1960) Ek et a/. (1983) Ulrich (1960) Strzelczyk et al. (1977) Ek et al. (1 983) Ulrich (1960) Ulrich (1960) Ulrich (1960) Ulrich (1960) Ulrich (1960) Strzelczyk et al. ( 1977) Strzelczyk et al. ( 1977) Ek er al. (1983) Gay and Debaud (1987) Gay and Debaud (1987) Ek et al. (1983) Gay and Debaud (1987) Gay and Debaud (1987) Gay and Debaud (1987) Gay et a/. (1989) Gay and Debaud (1987) Strzelczyk et a/. ( 1992) Gay and Debaud (1987) Gay and Debaud ( 1987) Gay and Debaud ( 1987) continues
90
Table IV-continued Method of detection
Fungi
Auxin"
H. spoliatum
IAA, ICA, IAld
HPLC
Gay and Debaud (1987)
H. subsaponaceurn
HPLC
k c a r i a laccata Mycelium radicis
IAA, ICA, IAld IAA Auxin-like
GC-MS PC, bioassay
Gay and Debaud (1987) Ek er al. (1983)
Parillus ini~olurus
Auxin-like
PC, bioassay
Phlegmacium caesiostramineum I? cephulixum
5-OH-IAA
PC, bioassay
IAA IAA 5-OH-IAA
PC, bioassay
P corrosum I? elotum I? fuscumaculatum I? glaucopus I? infractum I? orechalcium I? purpurascens Pisolirhus finctorius Rhizopogon lureolus Scleroderma auratiurn
IAA IAA IAA IAA IAA IAA Auxin IAA
PC. bioassay PC, bioassay PC, bioassay PC, bioassay PC, bioassay PC, bioassay PC, bioassay TLC, HPLC, ELISA, GC-MS Bioassay bioassay
IAA IAA IAA
GC-MS GC-MS bioassay
GC-MS GC-MS RIA
A. rubcscens Basidiomycefes spp.
IAA IAA [9R]iP, PRIZ Z. [9R]Z CLS
Boletus eleguiis
[9R]Z. Z, iP
column chromatography, (Sephadex LH-20) bioassay
S. bovinus S. luteus S. variegatus
S. variegatus
Trichlomn imbricatu Amaniru rnuscaria
PC, bioassay PC. bioassay
Reference
Strzelczyk er al. ( 1977) Strzelczyk and Pokojska-Burdziej (1984) Horak ( 1964) Horak ( 1964) Horak (1964) Horak ( 1964) Horak ( 1964) Horak ( 1964) Horak ( 1964) Horak (1964) Horak ( 1964) Frankenberger and Poth ( I 9874 Ruddwska ( I 982) Tomaszewski and Wojciechowska ( 1 974) Ek et al. (1983) Ek et al. (1983) Tomaszewski and Wojciechowska ( 1974) Ek et nl. (1983) Ek et al. (1983) Kraigher e f a/. (1991) Miller (1968) Kampert and Strzelczyk (1978) Ng et ul. (1 982)
continues 91
Table IV-continued Fungi
Method of detection
Auxin"
Reference
E. edulis var. pinicola Cenococcum granifone Cephanlospordum acremonium Glomus mosseae
Z
PC, bioassay
Gogala, 1970
CLS, [9R]iP, ~ 9 ~ 1 2 CLS
GC, bioassay
CLS
PC, bioassay
Hebeloma crustuliniforme H.mesophaeum
CLS, [9R]iP, PRIZ CLS, [9R]iP, PRIZ [9RliP, [9R]Z [9R]iP, [9R]Z CLS, GS [9R]iP, [9R]Z CIS
GC, bioassay
HPLC-RIA RIA TLC, uv RIA PC, bioassay
[9R]iP, [9R]Z
GC, bioassay RIA RIA GC, bioassay
R. ochraceoncbens
[9R]iP, [9R]Z [9RliP, [9R]Z CLS, [9R]iP, PRIZ CLS
R. roseolus
1-Z, t-[9R]Z
R. vinicolor Suillus bovinus
[9R]iP, [9R]Z CLS, [9R]iP, PRIZ 1-Z, t-[9R]Z
PC, W, MS, bioassay RIA GC, bioassay
Strzelczyk et al. (1992) Kampert and Strzelczyk (1978) Barea and Azc6nAguilar (1982) Strzelczyk et al. (1992) Strzelczyk et al. ( 1 992) Kraigher et al. (1991) Kraigher et al. (1991) Ho (1987b) Kraigher etal. (1991) Kampert and Strzelczyk (1978) Strzelczyk etal. (1985) Kraigher e? al. (199 1) Kraigher et al. (1991) Strzelczyk and Kampert ( 1987) Crafts and Miller (1974) Miller (1967)
Laccaria bicolur L. laccata L. laccata L. proxima Mycelium radicis atruvirens Paxillus involutus
l? involutus Pisolithus tinctorius Rhizopogon luteolus
S.cotiturnatus S.luteus S. punctipes
Thelephora terrestris 1: terresrris
Waematoloma fasciculare
[9R]iP, [9R]Z t-Z, [9R]Z [9R]iP, [9R]Z iP, [9R]iP, Z, [9RlZ CLS
PC, bioassay
GC,bioassay
TLC, UV, MS, bioassay RIA TLC, PC, UV, MS, bioassay RIA HPLC-RIA PC, bioassay
For abbreviations, see Appendix. 92
Kraigher et al. (1991) Strzelczyk and Kampert ( 1987) Miller (1971) Kraigher etal. (1991) Crafts and Miller ( 1974) Kraigher et al. (1991) Kraigher et al. (1991) Kampert and Strzelczyk (1978)
PLANT GROWTH REGULATORS IN THE RHIZOSPHERE
93
fungi to release auxins into their culture media (Rouillon et al., 1986; Frankenberger and Poth, 1987a; Gay et al., 1989; Gay and Debaud, 1987; Barroso et al., 1986; Ek et al., 1983). Ek et al. (1983) identified IAAin the culture media of 17 of 19 species or isolates of ectomycorrhizal fungi. Ho (1 987a) reported that all 8 isolates of Pisolithus tinctorius obtained from various sources were capable of producing auxins in culture media but varied substantially in their ability to release IAA. Similarly, Gay and Debaud (1987) screened 12 species of Hebeloma and 11 strains of H. cylindrosporum and reported that all were IAA producers, but there were large differences among species as well as strains in producing IAA. Some of the mycorrhizal fungi are capable of releasing small amounts of auxins in the absence of TRP; however, the presence of TRP enhances IAA production by a majority of the mycorrhizal isolates (Strzelczyk and Pokojska-Burdziej, 1984; Kampert and Strzelczyk, 1975; Horak, 1964; Haselwandter, 1973; Strzelczyk et al., 1992). Exogenous application of IAA or IBA alone or in combination with other plant hormones improves the development of mycorrhizae (Gunze and Hennessy, 1980; Azcdn et al., 1978; Dutra et al., 1996). By using auxin mutants of Hebeloma cylindrosporum, Gay et al. (1994) reported that auxin overproducer mutants exhibited an increased mycorrhizal activity. Several studies have demonstrated increased auxin content (hyperauxiny) in response to mycorrhizal infection, which may indicate a role of auxin in symbiosis. Sherwood and Klarman (1980) found that Pinus virginiana seedlings inoculated with Amanita rubescens contained higher concentrations of auxins than did noninoculated seedlings. Mitchell et al. (1986) reported that the level of IAA in short-leaf pine (Pinus echinata Mill.) roots was increased (three-fold) by inoculation with the ectomycorrhizal fungus Pisolithus tinctorius, indicating that hyperauxiny is associated with mycorrhizal symbiosis. Recently, Seagel and Linderman (1994) investigated the variation in conifer seedling growth, survival, and endogenous IAA content as influenced by inoculation with ectomycorrhizal fungi capable of producing IAA and C,H, in vitro. They reported that the capacity for in vitro production of IAA and C,H, by mycorrhizal fungi was partially correlated to levels of endogenous root IAA and root growth both with inoculated greenhouse-grown and field-transplanted nursery-grown seedlings. They concluded that although the degree of correlation varied with fungal isolates as well as conifer species, IAA mediated changes in root growth, and survival could be indirectly attributed to in vitro fungal PGR production. However, further studies are needed to investigate hyperauxiny in mycorrhizal plants to establish a relationship between auxin and mycorrhiza establishment. Very little work has been conducted on the detection of GAS released by mycorrhizal fungi. Gogala (1970) detected GA-like compounds in culture medium of the mycorrhizal fungus Boletus edulis var. pinicolus as well as in the fungal mycelium and fruiting body. Ho (1987b) screened 9 genera, 24 species, and 46 isolates of mycorrhizae for PGR production and found that they all varied greatly in their ability to release PGRs, including GA-like substances. In another study, Ho
94 MUHAMMAD ARSHAD AND WILLIAM T. FRANKENBERGER,JR.
(1987a) found that all 8 isolates of Pisolithus tinctorius were capable of producing plant hormones, including GAS, in varying amounts. Similarly, significant quantities of GA-like compounds were detected in filtrates of the ectomycorrhizal fungus Thelephora terestris (Hanley and Greene, 1987). Strzelczyk et al. (1975) found GA-like substances produced by the fungi Basidiomycetes, Mycelium radices atrovirens, and Cephalosporium glutineum isolated from the mycorrhizae of pine. Four of seven mycorrhizal fungi, including Suillus bovinus, Hebeloma mesophaeum, Cenococcum graniforme, and an ectomycorrhizal fungus, MrgX, were capable of producing GA-like substances in media with a pH less than 7.0 (Strzelczyk et al., 1992). In the postculture media of C. graniforme and H. mesophaeum, GA,-acetate was detected by GC. Similarly, Ho and Trappe (1992) detected extracellular growth regulators, included GAS, cytokinins, and IAA, in the culture of two Tricholoma magnivelare isolates. One of the most commonly observed differences between mycorrhizal-infected and noninfected plants is the increased size of infected plants in early stages of plant growth. This growth difference has usually been attributed to improved phosphorus nutrition in infected plants, but recent studies indicate the involvement of plant hormones in this host-parasite interaction. The ability of a microsymbiont to synthesize plant hormones, including GAS, in vitro favors this hypothesis. Furthermore, Allen et al. (1982) observed that infection by the mycorrhizal fungi Glomus fasciculatus resulted in significantly increased GA activity in the leaves of Boutelona gracilis, with a tendency for decreased GA activity in the roots. They suggested that these changes in GA levels may substantially alter the physiology of the host. With use of GC-MS, Clapperton et al. (1985) compared GAS in slender wheat grass (Agropyron trachycaulum) infected with the fungus Glomus aggregatum with that of noninfected plants. Infected plants had more endogenous GASthan did noninfected ones; however, contrary to Allen et al. (1982) findings, roots had a greater quantity of GA-like substances. Several mycorrhizal fungi have been screened and shown to be capable of producing cytokinins in vitro (Table IV). It is unclear whether these fungi that are capable of producing cytokinins also do so in association with the host plant, because these two habitats differ greatly in factors to which the symbiont is exposed. The role of these cytokinins can be considered from two different points of view relative to their ecological impact: (1) their role in mycorrhizae formation, and (2) their role in changing the cytokinin balance in the plant as a mycorrhizal response. No direct, unequivocal evidence indicates that cytokinins are a prerequisite for the formation of mycorrhizae. However, higher cytokinin levels in mycorrhizal plants have been reported, but the source of increased cytokinin levels in mycorrhizalinfected plants is somewhat unresolved. Allen et al. (1980) were the first to report higher cytokinin activity in mycorrhizal plants compared with noninfected (control) plants. They found 57 and 111% greater cytokinin activity in leaves and roots, respectively, in mycorrhizal, Boutelona gracilis-infected plants over control
PLANT GROWTH REGULATORS IN THE RHIZOSPHERE
95
plants. Similarly, Thiagarajan and Ahmad (1994) reported significantly greater cytokinin content (156%) in mycorrhizal roots of 45-day-old cowpeas compared to nonmycorrhizal roots. Several other studies confirmed these findings and provided evidence that inoculation with mycorrhizal fungi results in increasing the endogenous cytokinin contents of host plants (Edriss et al., 1984; Dixon et al., 1988a,b; Baas and Kuiper, 1989; Dixon, 1989; Danneberg et al., 1992). These alterations in the endogenous cytokinin pool may partly regulate many of the physiological responses of mycorrhizal-infected plants. Recently, two studies have demonstrated the influence of mycorrhizae andor Rhizobium on cytokinin activity in drought-stressed alfalfa (Goicoechea et al., 1995; 1996). Goicoechea et al. (1995) detected the cytokinin contents of alfalfa leaves: ( 1) inoculated with Glomus fascuenlatum and Rhizobium meliloti (MR); (2) inoculated with only Rhizobium (RP); ( 3 ) inoculated with only mycorrhizae (MN); and (4) noninoculated (NP), grown under both normal and drought-stressed conditions. They found that NP plants contained the highest total cytokinin concentration under normal conditions, but exposure to drought caused a drastic drop in cytokinin levels and decreased the number of stems as well as increased the leaf tissue senescence. By contrast, stressed symbiotic plants (RP, MN, and MR) showed higher green leaf weight than NP plants due to the delay of leaf senescence and an increase in stem number, in addition to maintaining leaf cytokinin levels during drought. They suggested that mycorrhizal symbiosis plays an important role in maintaining cytokinin levels under drought conditions, which may explain different plant responses to drought compared to nonmycorrhizal plants. Similar observations were made with respect to root cytokinins (Goicoechea et al., 1996). Symbiotic plants showed significantly higher root and nodule activity during drought conditions, which may be related to less decline in cytokinin content of mycorrhizal plants compared to nonmycorrhizal plants. These studies indicate that mycorrhizal symbiosis has a role in maintaining endogenous cytokinin levels that subsequently evoke a plant response. Driige and Schonbeck (1992) found significant growth responses of shoots and roots, caused by mycorrhizal infection, that were preceded by higher [9R]Z levels in specific organs. In one experiment, they observed a coincident increase in biomass and [9R]Z level. Driige and Schonbeck (1992) concluded that the enhanced internal cytokinin levels improved photosynthesis and growth of mycorrhizal flax. They reported that the improved growth of mycorrhizal flax was basically caused by the enhanced cytokinin production in mycorrhizal roots; however, interaction with other phytohormones cannot be ruled out. Edriss et al. (1984) investigated the cytokinin activity in nonmycorrhizal and mycorrhizal (Glomusfasciculatum, G. etunicatum, and Gigaspora heterogama)infected sour orange (Citrus aurantium L.) seedlings under three phosphorus levels applied in imgation water. They concluded that the enhancement of cytokinin production was associated with mycorrhizal infection, rather than with increased
96 MUHAMMAD ARSHAD AND WILLIAM T. FRANKENBERGER,JR. P uptake. Dixon (1989) observed increases in both cytokinin levels and P content in citrus seedlings in response to vesicular-arbuscular (VA) mycorrhizal infection. He suggested that a minimum level of P is required to sustain cytokinin activity, and cytokinins may facilitate P utilization, thereby affecting cell growth. The elevated cytokinin flux of inoculated seedlings is often associated with improved P nutrition and a significant increase in seedling biomass (Dixon et al., 1988b). Cytokinin transport from the root to shoot is also influenced by VA mycorrhizal symbiosis. Seedlings of citrus inoculated with G.fasciculatum or G. mosseae yielded a greater flux of Z, (diH)Z, and [9R]Z than did noninoculated seedlings. The flux of [9R]Z was significantly greater than that of Z in inoculated seedlings (Dixon et al., 1988b). Interestingly, the greater xylem flux of cytokinins (Z and [9R]Z) in VA mycorrhizal seedlings than in nonmycorrhizal seedlings is consistent with reports of in vitro production of Z and [9R]Z by the mycorrhizal fungi (Slankis, 1973; Crafts and Miller, 1974; Barea and Azcon-Aguilar, 1982). This may imply that if these cytokinins are not directly supplied by the fungal symbiont, the plant’s endogenous level may be stimulated by the mycorrhizal infection. Graham and Linderman (1980) examined 23 different ectomycorrhizal fungi for their ability to synthesize C2H4.They found that all produced C,H, when grown in a medium containing DL-MET.Graham and Linderman (1980) also reported that C,H, was produced by aseptically grown Douglas fir seedlings inoculated with Cenococcum geophilum, Hebeloma crustuliniforme, and Laccaria laccata, and the appearance of C,H, coincided with the formation of mycorrhizae. Lateral root formation was stimulated by inoculation with these three fungi but was inhibited by Pisolithus tinctorius. After confirming the C,H,-producing ability of L. laccata S238A, DeVries et al. (1987) noted an apparent correlation between C,H, production and morphological effects, such as stimulation of lateral root formation by this fungus. However, they found a poor correlation between high C,H, levels in vivo and low dichotomy. Rupp et al. (1989a) examined the role of C,H, in mycorrhizae formation and root development on axenically grown seedlings of Pinus mungo. They observed that mycorrhizal formation by L. laccata and I? tinctorius in a defined medium was associated with increases in C,H, production four- and two-fold greater than the controls (medium plus fungus minus seedlings) by L. laccata and P: tinctorius, respectively, 5 weeks after the seedlings were inoculated. Silver thiosulfate, an inhibitor of C,H, action, decreased mycorrhizal formation by L. laccata but had no significant effect on mycorrhizal formation by I? tinctorius. The authors speculated that endogenous C,H, may have influenced mycorrhizal formation and associated changes in root morphology. Similarly, Livingston (1991) conducted experiments to study the ability of two laccarial ectomycorrhizal fungi, L. laccata and L. bicolol; in producing MET-dependent C,H, with development of mycorrhizae. Although the laccarial isolates produced C,H,, it was not related to their ability to form ectomycorrhizae. Studies on the biochemistry of C,H,-biosynthesis by ectomycorrhizal fungi re-
PLANT GROWTH REGULATORS IN THE RHIZOSPHERE
97
veals that it is different from that of higher plants (Rupp et al., 1989b; Livingston, 1991). Rupp et al. (1989b) found that L. laccata and H. crustiliniforme produced C2H4 when grown in MET-amended media but failed to yield C2H4upon replacing MET with ACC. Moreover, AVG, an inhibitor of C,H, biosynthesis in higher plants, did not inhibit MET-induced C,H, production by these fungi. Livingston (1991) confirmed that the addition of MET promoted C2H, generation by L. luccata and L. bicolor, but ACC did not have any promotion effect. Production of ABA by mycorrhizal fungi has not been demonstrated as yet; however, few studies have investigated the alteration in ABA levels in mycorrhizal-infected plants. Danneberg et al. (1992) determined concentrations of ABA and other phytohormones in mycorrhizae (infected by VAM fungus, Glomus isolate TJ and in a nonmycorrhizal control and found that concentrations of ABA, both in shoots and roots, were always higher in mycorrhizal plants than in control plants throughout the growth period assayed. Cytokinin levels increased only at the last stage of growth, whereas the auxin content did not change. They suggested that the long-term effects ofABAcould involve regulation of plant-myconhizal symbiosis (Danneberg et ul., 1992). Contrarily, Allen et al. (1982) found that infection with a vesicular-arbuscular fungus, Glomus fasculatus, resulted in decreased ABA levels in the leaves but remained unchanged in roots of Bouteloua grucilis grown in a defined medium. Altered ratios of gibberellins, ABA, and cytokinins (Allen et al., 1980, 1982; Dixon et al., 1988a,b) may be responsible for these major changes in the physiology of the host plant with mycorrhizal interactions. Although some factors-such as widespread ability of mycorrhizal fungi to produce PGRs in culture media and induction of mycorrhizal-like changes in response to exogenous application of PGRs-favor the speculation that fungal PGRs might have a role in establishment of the symbiotic relation and in the physiology of mycorrhizal plants, the physiology of PGRs released by the symbiont and the role of these metabolites in the symbiotic association are still poorly understood. The interaction between phytohormones themselves may explain some of the discrepencies found in determining the physiological role of the individual PGR in mycorrhiza establishment. However, the changes in the balance of all five major plant hormones have yet to be examined for any plant species, and plant development stages in vesicular-arbuscular mycorrhizae-infected plants are also critical. Increased understanding of the hormonal interaction between mycorrhizae and plants could be of great ecological benefit to the agriculture industry.
E. PATHOGENESIS Factors such as the widespread ability of pathogens to produce PGRs in vitro and hyper levels observed in some infected hosts (Sequeira, 1973; Gruen, 1959;
98 MUHAMMAD ARSHAD AND WILLIAM T FRANKENBERGER,JR. Brian, 1957; Van Andel and Fuchs, 1972; Pengelly and Meins, 1982, 1983; Liu and Kado, 1979; Liu et al., 1982; Mahadevan, 1984; Fett et al., 1987; Ishikawa et al., 1988; Morris, 1986, 1995) favor the premise that some PGRs may play a role in pathogenicity. The abnormal growth responses that are frequently associated with the invasion of plants by pathogenic microflora are sometimes attributed to increased PGRs production. However, few studies have made an attempt to separate the contribution of both partners (pathogen and host) to the PGR changes occurring at various stages of pathogenesis. The suggested mechanisms for hyper levels of PGRs in diseased plants may include the following: (1) the pathogen synthesizes PGRs, which increases the endogenous PGR level in the infected tissues; (2) the infected plant reacts and generates PGRs in excess; (3) the pathogen releases other substances that inactivate enzymes involved in degradation of PGRs; (4) the plant increases the transport of PGRs to the infection site; (5) degradation and conjugation of PGRs is being altered in the infected plants; and (6) bound PGRs are released by the host in response to infection. The pathogenic bacteria A. tumefaciens and I? syringae pv. savustanoi are known to incite tumors and galls in dicotyledonousplants and provide the classic examples of involvement of PGRs (auxins and cytokinins) in pathogen-plant interactions. The hairy-root disease of dicotyledons caused by Agrobacterium rhizogenes is also indicative of auxin involvement. Similarly, “Bakanae” disease of rice caused by Gibberellafujikuroi represents a classic example of involvement of GASin pathogenesis.These systems have been studied extensively and will be discussed in detail in the following sections.
1. Agmbactm’um tlcmefaciens Agrobacterium tumefaciens causes crown gall tumor, a neoplastic disease of dicotyledonous plants. The tumors are hormone-autotrophicin culture and contain elevated levels of auxins and cytokinins (see Moms, 1986, 1995; Moms et af., 1986). A biological characteristic that distinguishes the tumor cell from the normal cell is its capacity for uncontrolled or autonomous growth. This autonomy appears to be a result of continued production and/or accumulation of greater-thannormal amounts of intercellular phytohormones by the tumor cells, leading to uncontrolled growth accompanied by cell division. Auxin production by the pathogen has been rigorously characterized, and the genes responsible for auxin biosynthetic enzymes have been identified. This bacterium produces auxin from TRP via IAM pathway (Follin et al., 1985; InzC et al., 1984; Klee et al., 1984; Van Onckelen et al., 1985, 1986; Schroder et al., 1984; Thomashow et al., 1984,1986). The first enzyme involved is TRP-2-monooxygenase (TOM) encoded by the tmsl locus, and the second is IAM hydrolase encoded by trns2 (see Morris, 1986,1995).In addition to auxins, cytokinins may also be involved in tumorigenesiscaused by A. tumefaciens (Scott and Horgan, 1984;Miller, 1974; Lippincott and Lippincott, 1976). Levels of free cytokinins and cytokinin-
PLANT GROWTH REGULATORS IN THE RHIZOSPHERE
99
glucosides were markedly elevated in Datura and Nicotiana crown gall tissues compared to nontransformed tissues (Palni et al., 1983b; Scott and Horgan, 1984). Crown gall tumor tissues often have an unusually high level of diversity in endogenous cytokinins, with over 15 cytokinins identified by GC-MS (Palni et al., 1985). Ondrej et al. (1991) demonstrated more than a 100% increase in the level of endogenous cytokinins in tobacco plants transgenic with constructed A. tumefaciens. Similarly, the tissue- and organ-specific overproduction of cytokinins was observed in transgenic tobacco plants injected by A. tumefaciens encoded for the cytokinin gene (Li et al., 1992). A comparison of in vitro cytokinin production by virulent and arivulent strains of A. tumefaciens reveals that cytokinins released by the pathogen may have a substantial role in tumor formation. In different studies, cytokinin production by virulent strains of A. tumefaciens was several-fold greater than the production by avirulent strains (Einset e f al., 1977; Greene, 1980; KaissChapman and Morris, 1977). An extensive body of evidence indicates that A. tumefaciens incites crown gall tumors by transferring a fragment (T-DNA) of a large endogenous plasmid, the Ti plasmid (the tumor-inducing plasmid), to the host. The integration of T-DNA into plant genomes results in the transformation of the host cells. All virulent strains of A. tumefaciens were found to have Ti plasmids (Zaenen et al., 1974j, which encode for the enzymes involved in hormone production (Morris et al., 1982; Schroder et al., 1984; Barry et al., 1984; Thomashow et al., 1984; Hammerschlag et al., 1989) and is the tumor-inducing principle (TIP). The plasmid may either carry (1) structural genes for enzymes involved in auxin and cytokinin synthesis; (2) a regulatory gene that would change specifically the normal control of hormone metabolism; or (3) a regulatory gene with pleiotropic effects, causing an increase in the synthetic activities of transformed cells. Details of the molecular mechanism underlying tumor formation by A. tumefaciens can be found in various reviews (Morris, 1986, 1995; Nester et al., 1984; Gaudin et al., 1994; Patten and Glick, 1996; Costacurta and Vanderleyden, 1995). Recently, Gafni et al. (1995) reported that incubation ofA. tumefaciens C58 with IAA, stimulated (up to 10-fold) the bacterial virulence. They suggested that excretion of IAA by transformed cells stimulates bacterial virulence mechanism(s) encoded by the Ti plasmid, the chromosome, or both. In some plant species, tumor growth cannot be explained on the basis of elevated levels of IAA and/or cytokinins alone (Weiler and Spanier, 1981j. There appears to be some relationship between the cytokinin-auxin ratio and primary tumor morphology (Akiyoshi et al., 1983).
2. l? syringae pv. savastanoi Pseudomonas syringae pv. savastanoi induces tumorous outgrowth, known as galls or knots on woody species, including olive, oleander, and privet. Extensive studies have been conducted to investigate the determinants of virulence involved
100 MUHAMMAD ARSHAD AND WILLIAM T. FRANKENBERGER,JR. in this pathogen-host interaction (see Morris, 1986, 1995). It has been concluded that a hormonal imbalance in response to infection may be the main cause of pathogenicity (Comai and Kosuge, 1982, 1983; Comai et al., 1982; Smidt and Kosuge, 1978; Silverstone et al., 1990; Moms, 1986, 1995). Iacobellis et al. (1994) confirmed that expression of IAA alone is sufficient to initiate the development of knots on oleander, whereas cytokinins are necessary for the full expression of disease symptoms (determining knot size). Molecular studies confirmed that the bacterium contains two plasmid-borne genes, iaaM and iaaH, which encode the enzymes TRP-2-monooxygenase (TOM) and IAM hydrolase, respectively, which act in concert to produce IAA from TRP via IAM (Comai and Kosuge, 1980; 1982; Comai et al., 1982;Yamada et al., 1985; White and Ziegler, 1991). The genes iaaM and iaaH maintain a substantial homology with their counterparts tmsl and rms2 from A. tumefaciens (Yamada et d., 1985). By using different parent and mutant (defective in IAA synthesis) strains, Smidt and Kosuge (1978) demonstrated that the isolates that accumulated IAA caused formation of galls, whereas those that did not accumulate IAA failed to incite gall formation on oleanders, establishing a clear connection between IAA production by the pathogen and its virulence. Molecular studies revealed that the mutant defective in IAA production had lost virulence, whereas overproduced IAA exhibited enhanced virulence (Comai and Kosuge, 1980; Comai and Kosuge, 1982; Comai et al., 1982). Cytokinin production by Z? syringae pv. savastanoi has also been reported by many workers (Surico et al., 1975, 1985; MacDonald et al., 1986; Morris et al., 1986). Surico et al. (1975, 1985) detected cytokinins in addition to IAA in cultures of F! syringae pv. savastanoi. They found that knots were formed on olive or oleander only by strains that produced cytokinins in addition to IAA. Identity of the cytokinins produced by Z? syringae pv. savastanoi was recently confirmed, which indicates that there is a diversity of cytokinins, including methyl derivatives synthesized (MacDonald et al., 1986; Morris et al., 1986; Moms, 1986; Akiyoshi et al., 1987). Akiyoshi et al. (1987) demonstrated that Z? syringae pv. savastanoi strains are prolific producers of t-Z (up to 1 mg L-l). Morris et al. (1986) identified a plasmid-borne gene (ptz)encoding cytokinin biosynthesis. Expression of ptz in E. coli caused secretion of Z, [9R]Z, iP, and [9R]iP and has considerable functional homology with tzs and ipt of A. tumefaciens. In contrast to the crown gall tumors caused by A. tumefaciens, P. syringae pv. savastanoi produces galls without transformation of the host cells by transferring its DNA. Rather, it exerts its pathogenic effects by virtue of close association with the host cells, at least in part, by secreting high amounts of phytohormones (Comai et al., 1982; Morris, 1986, 1995). F! syringae pv. savastanoi, which causes brown spot of bean, also synthesizes IAA, at least in part, via IAM pathway (White and Ziegler, 1991). By using various mutants of this pathogen for their effects on bean plants, Mazzola and White (1994) concluded that P. syringae pv. savastanoi-derived IAA is involved in the
PLANT GROWTH REGULATORS IN THE RHIZOSPHERE
101
regulation of plant growth and in the expression of other factors that affect the host-pathogen interaction.
3 . Agrobacterizcm rhizogenes The hairy root syndrome is a neoplastic disease of dicotyledonous plants induced by A. rhizogenes. Numerous adventitious roots sprout from the site of infection and bear strong resemblance in many respects to crown gall tumors induced by A. tumefaciens. These similarities include: (1) the presence of genetic elements responsible for bacterial virulence on large plasmids (Moore et al., 1979; White and Nester, 1980; Costantino et al., 1981; Melchers and Hooykaas, 1987; Binns and Thomashow, 1988); (2) transfer and integration of a fragment of root-inducing plasmid (Ri), i.e., T-DNA in the genome of the infected plant cells (Chilton et al., 1982; White er al., 1982, 1985; Willmitzer et al., 1982; Spano et al., 1982; Shen er al., 1990; Cardarelli et al., 1987); (3) presence of auxin synthetic genes in the T-DNA of some Ri plasmids (Huffman et al., 1984; Cardarelli et al., 1987; Camilleri and Jouanin, 1991); and (4) the plasmids share strongly homologous vir genes, the genetic determinants that direct T-DNA mobilization (Risuleo et al., 1982; Hooykaas et al., 1984). Furthermore, Camilleri and Jouanin (1991) reported a significant similarity between the auxin biosynthesis genes of A. rhizogenes (auxl and aux2) with that ofA. tumefaciens (tmsl and tms2) and of P. syringae pv. savastanoi (iaaM and iaaH). The preceding facts favor the hypothesis that TDNA-directed auxin synthesis in transformed cells may be involved in the development of disease symptoms. 4. Colynebacterizcmfascians
Corynebucteriumfascians causes leafy galls, a disease known as “witch’s broom” or fasciation. Fasciations lead to several abnormalities, including the outgrowth of lateral buds, abnormally small leaves, and thickened internodes. Symptoms similar to those caused by C.fascians infection can be produced by treating plants with cytokinins (Roussaux, 1966; Sachs and Thimann, 1967; Thimann and Sachs, 1966). Studies on hormone levels in infected tissues further support the fact that C.fascians has the capacity to supply cytolunins to the host plant. Using pea as a test plant, Murai et al. (1980) compared five strains of C.fascians for virulence and cytokinin production. They found that the highly virulent strains were the most efficient producers of cytokinins, whereas an avirulent strain was the least effective. Recently, Crespi er al. (1992) reported that fasciation induction by the phytopathogen Rhodococcus fascians depended upon a linear plasmid encoding a cytokinin synthase. The cytokinins produced by C. fascians in culture have been speciated as iP, [9R]iP, c-Z, r-Z, [9R]Z, c-[2MeS]Z, and [2MeS]iP (Armstrong et al., 1976). The cytokinins iP and Z were the most active biologically, whereas the cis isomer of iP was produced in the most quantity.
102 MUHAMlMAD ARSHAD AND WILLIAM T FRANKENBERGER,JR.
Many investigators have concluded that disease symptoms by C. fascians are a result of the exogenous supply of cytokinins to the plant surface released by the infected bacterium, demonstrating a plant-microorganism relationship in which there is a direct effect of microbial cytokinins on the plant (see Greene, 1980). However, the possibility that the bacterium stimulates excess production of host cytokinins as well as other endogenous growth regulators cannot be excluded. Hormonal imbalance may also contribute to pathogenicity. 5. Others
Plant pathologists discovered GAS while studying “Bakanae” disease of rice. In this disease, infection of a rice plant with G. fujikuroi results in abnormal elongation of plants because of the contribution of GAS to their host by the fungus (see Frankenberger and Arshad, 1995, for details). After comparing various mutants of G. fujikuroi for virulence, Takenaka et al. (1992) observed a decrease in virulence due to decreased production of GAS by the mutants. Other studies conducted to assess the role of GAS in pathogenesis often lack information on the type of GAS involved because most investigators relied on bioassays to detect GAS. Also, little emphasis has been placed on monitoring the concomitant GA changes with the physiological and biochemical alterations of infected plants. Since many plant pathogens have been shown to produce C,H4 in culture, it has been suggested that many symptoms of disordered growth in diseased plants are a result of excessive C2H4production. Based upon some typical symptoms associated with C,H4 action, such as formation of adventitious root initials, epinasty, synthesis of specific enzymes, abscission, stunting, chlorosis, necrosis, and finally death, numerous studies now indicate the involvement of C,H4 in pathogenesis. However, the relationship between microbial (pathogen) production of C,H, and plant disease is somewhat complex, which makes it very difficult to asses the role of C,H4 as a causal agent in disease development. Further complexities of pathogenesis involve factors such as C2H4interactions with other plant hormones, particularly with auxin; the effects of C2H4on the physiology and metabolism of plants in the absence of obvious growth effects; and secretions of toxins by the parasite, inducing C,H4 synthesis by the host. An in-depth discussion on C2H4 in host-pathogen relationshipshas been provided by Archer and Hislop (1 973, Boller (1991), and Frankenberger and Arshad (1995). Depending upon the system, C2H4 may either induce a defense response in the host or it may act synergistically with the pathogen to accelerate symptom development (Bell, 1981; David et al., 1986; Ecker and Davis, 1987; Elstner, 1983;Goto and Hyodo, 1985; Stall and Hall, 1984). Nevertheless, production of C,H4 at an accelerated rate during various stages of pathogenesis is a commonly observed phenomenon. Several studies indicate that host-parasite interactions can trigger C,H4 production, which coincides with or precedes the development of disease symptoms and subsequent physiological
PLANT GROWTH REGULATORS IN THE RHIZOSPHERE
103
changes in plants, or C2H4 outburst occurs at later stages of disease development. It is difficult to be conclusive about the origin of increased C2H4generation in infected hosts; it is either a host metabolite or a pathogen product. Chalutz (1979) presented convincing evidence that C2H4production in I? digitatum rot of orange and lemon fruit is largely contributed by the fungus. But Coleman and Hodges (1986) suggested that the pathogen may stimulate C2H4 production from the host tissue early in pathogenesis; whereas, later, during the infection process, the pathogen may also directly contribute C2H4 and ACC. Similarly, Achilea et ul. (1985a,b) suggested that C2H4 production in the healthy fruit was of plant origin; whereas markedly enhanced production of C2H4 by I? digitatum in the infected grapefruit was mostly or entirely of fungal origin. On the other hand, studies have shown indirectly that increased C2H4production from infected tissues may be a plant metabolic product stimulated by infection. From experiments in which plants were treated with bacterial or fungal components, so-called elicitors, Boller (1991) concluded that most of the C,H4 synthesized in bacterial or fungal diseases is of plant origin. Toppan et al. (1982) revealed that enhanced C,H4 in melon infected by Colletotrichum lagenarium was suppressed by specific inhibitors of the plant C2H4 biosynthetic pathway, namely L-canaline and AVG, implying that C,H4 biosynthesis was a plant metabolic product. Similar observations were noted by Bashan (1994), who investigated the correlation between symptom expression and C2H4 production in leaf blight of cotton caused by Alternaria macrospora and A. alternatu alone and in combination. Similarly, several studies have shown that high rates of C2H, production triggered by infection were suppressed by AVG in pea pods infected by Fusarium solani f. sp. pisi (Mauch et al., 1984); parsley cells infected by Phytophthora megasperma (Chappell et ul., 1984); melon seedlings treated with Colletotrichum lagenarium (Roby et al., 1985, 1986), and swingle citrumelo leaves infected by Xunthomonas compestris pv. citri (Dutta and Biggs, 1991). This may imply that the pathogen accelerates C2H, production from the host. Although all of the known fungi capable of producing ABA are phytopathogens, little work has been conducted by plant pathologists to investigate the role of ABA in pathogenesis. Studies have shown that ABA influences the development of some phytopathogens as well as some diseases, and infection may lead to altered endogenous ABA levels; however, it is not possible to draw a conclusion about the definite role of ABA in pathogenesis. Recently, Fraser (1991) has critically reviewed the relationship between ABA and plant responses to pathogens. It has been shown that infection leads to altered ABA levels in host tissues (see Frankenberger and Arshad, 1995). The increases in endogenous ABA levels in response to infection suggest a role of ABA in pathogenesis. However, no efforts have been made to establish a “cause and effect” relationship, except for a recent study conducted by Yoshida and Takano (1992). They screened several phytopathogen fungi for their ability to synthesize ABA in vitro and selected two ABA-producing fungi, including Botrytis cinereu (host, Tulipa gesneriana) and Stagonospora curtisii (hosts,
104 M -
ARSHAD AND WILLIAM T. FRANKENBERGER,JR.
Hippeastrum hybridum and Narcissus tuzettu),for an infection study. Infection with S. curtisii stimulated the breakdown of chlorophyll in narcissus leaves and inhibited the development of the scape of H.hybridum. The loss of chlorophyll in narcissus was stimulated when the detached leaves were treated with the culture filterate of the fungus or by an ABA solution. This indicates that the expression of symptoms of diseased leaves infected by the fungus is closely related to production of ABA by the fungus itself. Tuomi et al. (1 993) studied the interaction of ABA- and/or MA-producing fungi (B. cinerea, Cludosporium cludosporioides, and Aureobasidiurn pulluluns) in the Sulix leaves. They reported that the content of ABA, and to a lesser degree that of IAA, showed a positive correlation with the frequency of infection by the hormone-producing organisms. However, they were of the view that fungal production of ABA or IAA might not be the significant contributor to the hormonal contents of Sulix leaves. Similarly, Kettner and Dorffling (1995) studied the biosynthesis and metabolism of ABA in leaves of two near-isogenic lines of tomato (wild type and ABA-deficient mutant) infected with two virulent strains of B. cinereu. One strain produced ABA, whereas the other did not. They found that levels of endogenous ABA in leaves of tomato infected with ABA-producing strain rose more than ten-fold in Ailsa Craig (wild type) and two-fold injuccu (ABA-deficient mutant), whereas infection with a nonproducing fungal strain resulted in a five-fold increase in ABA in Ailsa Craig and no increase injaccu. They concluded that at least four processes control the level of ABA in wild-type tomato leaves infected with B. cinerue: (1) the host stimulates fungal ABA biosynthesis; (2) the fungus releases ABA or its precursor; (3) infection stimulates plant ABA biosynthesis; and/or (4) the fungus inhibits metabolism of plant ABA. In brief, except for a few cases it is not possible to explain precisely the role of PGRs in pathogenesis. It is not clear whether any alterations in PGR concentrations in the infected plant are due to interference by the pathogen with the host’s PGR’s biosynthetic machinery. Although in vitro production of PGRs by pathogens has been well documented, in most cases in vivo microbial contribution of PGRs to the host has not been well characterized.
VI. METABOLISM OF PGRs IN SOIL A. A m s Since a large number of soil and rhizosphere microorganisms have been reported to produce auxins, indirect inferences have been made about the presence of auxins in soil in amounts high enough to be biologically active. By using the Avenu bioassay, Stewart and Anderson (1 942) reported that fertile soils contain greater amounts of auxin than less fertile soils. The addition of TRP as an auxin
PLANT GROWTH REGULATORS IN THE RHIZOSPHERE
10s
precursor substantially increases the IAA levels in soils (Chandramohan and Mahadevan, 1968; Pumshothaman et ul., 1973, 1974; Frankenberger and Brunner, 1983; Wohler et al., 1990; Martens and Frankenberger, 1991; 1993a,b; Sarwar et al., 1992; Lebuhn and Hartmann, 1993; Lebuhn et al., 1994). In most cases, exogenous application of TRP is required to have detectable levels of IAA in soils (Martens and Frankenberger, 1993a; Lebuhn and Hartmann, 1993; Sarwar et al., 1992). Several workers reported substantial variations in the capacity of various soils to produce auxins upon addition of TRP (Sarwar et al., 1992; Martens and Frankenberger, 1991; 1993a,b; Chandramohan and Manhadevan, 1968). HPLC analysis showed that IAA concentrations in six California soils ranged between 1 to 13 mg kg-' soil upon addition of 225 mg kg-' soil of L-TlU? (Martens and Frankenberger, 1991). Several studies have demonstrated that auxin production in soil is a biotic process (Purushothaman et al., 1974; Sarwar et al., 1992; Martens and Frankenberger, 1993a; Lebuhn and Hartmann, 1993). IAA is not detected in sterile soils amended with a sterile-autoclaved solution of TRP. Moreover, inoculation with pretested, efficient IAA-producing actinomycetes resulted in IAA accumulation in soil (Purushothaman et al., 1974). IAA synthesis was greater in the nonsterile soil compared with the actinomycete-inoculated sterile soil, indicating that the soil's indigenous microflora were active in IAA production. This clearly demonstrates that auxin production in soil is widespread and is primarily a biotic phenomenon. Bacterial isolates from fertile soils possess the ability to synthesize comparatively much more IAA than isolates from nonfertile soils (Dahm et al., 1977). Narayanaswami and Veerraju (1969) compared IAA formation in nonrhizosphere and rhizosphere soils, both amended with DL-TRP;they reported more IAA accumulation in the rhizosphere than in bulk soil and attributed this difference to the high population of bacteria, fungi, and actinomycetes in the rhizosphere soil. Rossi et al. (1984) found that the auxin content of free bulk soil and of maize rhizosphere varied with mineral fertilizerand atrazine treatments in relation to seasonal weather conditions and the plant vegetative cycle. The highest auxin content in the rhizosphere was at maize plant emergence. The auxin level was usually higher in the rhizosphere than in the free bulk soil, probably as a consequence of an increased microbial population and/or accelerated metabolism due to the presence of root exudates. Miiller et al. (1989) reported the presence of IAA and other phytohormones in corn rhizosphere soil. They observed a high concentration gradient of IAAexisting from the rhizoplane and adjacent rhizosphere toward the bulk soil. This gradient leveled off as the distance from the rhizosphere increased. They also observed a high level of IAA in the rhizosphere of corn inoculated with A. chroococcum compared to an uninoculated control. Similarly, Lebuhn and Hartmann (1993) suggested the presence of higher auxin content in rhizosphere soils because of root colonization with azospirilla and rhizobia capable of excreting auxin without the addition of TRP, whereas bulk (nonrhizosphere) soil has only traces of IAA without the exogenous addition of
106 MUHAMMAD ARSHAD AND WILLIAM T FRANKENBERGER,JR. TRP. In another study, Lebuhn et al. (1994) reported that potential (W-dependent) microbial production of IAA was predominant in equilibrated fresh soils and dependent on the soil nutrient content and the size and metabolic status of the soil biomass. They observed a shift from potential IAA production to TOL and anthranilic acid production in response to drying-rewetting stress. They proposed that the release of TRP, microbial auxins, and the shift toward TOL production function as stimulants for root developmentwere induced by environmental fluctuations (Lebuhn et al., 1994). Martens and Frankenberger (199313) found that the addition of three amino acids-L-alanine, L-asparagine, and L-lysine, which have been reported to be present in root exudates-stimulated L-TRP-derived formation of IAA in two soils. They also found that growing rhizobacterial isolates of Festuca octoflors Walt. on minimal salt agar amended with L-lysine, L-asparagine, and L-alanine stimulated their capacity to convert TFV into IAA. We postulated that the presence of these IAA-stimulating amino acids in root exudates may explain why rhizosphere soil has a higher content of auxin when compared with root-free soil. Auxin production in soil occurs through different pathways. Frankenberger and Brunner (1983) unequivocally demonstrated the presence of both IPyA and IAM in addition to IAA in soil amended with L-TRP. Later, Martens and Frankenberger (1991, 1993b) reported IAM as one of the TRP-metabolites in cultures of various bacterial isolates of soil. Similarly, when labeled L-TRP was added to soil, TRPcatabolites, including IAM, L A , IAA, TOL, and IAld, were identified (Martens and Frankenberger, 1993a).These studies clearly demonstrate that IAA formation from TRP in soil occurs at least via two pathways: (1) TRP is converted to IAA by deaminating to IPyA followed by decarboxylation to IAAId, which is further oxidized to IAA; and (2) TRP is decarboxylated to IAM, which is hydrolyzed to IAA. The addition of IAAld, a direct precursor of IAA in the IPyA pathway, resulted in little or no IAA formation in the five soils tested, whereas addition of IAM or TAM resulted in large percentages of conversion to IAA (20 and 18%, respectively) (Martens and Frankenberger, 1993b). This implies that the IAM pathway of MA-formation from TRP dominates over the I Q A pathway in these soils. However, Lebuhn and Hartmann (1993) found only traces of IAM in one soil and no IAM in another soil, both amended with TRP. It is highly likely that in some soils one pathway of IAA production may dominate over others and vice versa. In addition to auxin production, degradation of auxins also occurs in soil. Parker-Rhodes (1940) first suggested that both the auxin synthesizing and destroying factors involving microorganisms are present in soil. When IAA was applied directly to soil, its presence declined rapidly (Hamence, 1946; Chandramohan and Mahadevan, 1968; Strzelczyk and Karwowska, 1969). Recent studies have confirmed the early work on auxin catabolism in soil. Martens and Frankenberger (199 1) observed that IAA concentration reached a maximum in TRP-amended soils after incubation for 5 days and then decreased with increasing incubation time. In another comprehensive study, they investigated the stability of various auxins in
PLANT GROWTH REGULATORS IN THE RHIZOSPHERE
107
soil (Martensand Frankenberger, 1993a).Adsorption-desorptionisotherms showed a low affinity of auxin derivatives (5-OH-IAA, IAM, ILA, IAA, TOL, and IAld) for soil colloides. Mineralization of these derivativesvaried from soil to soil as measured by CO, evolution; however, an average of five soils showed 29,40,45, and 43% mineralization of TOL, IAA, IPyA, and IAM, respectively, after 5 days of incubation. Degradation of L-TW and auxin derivatives followed a first-order kinetic reaction and the measured half life (t,,,) ranged from an average of 25 hours for L-TRPto over 127 hours for IAM with the five soils. Martens and Frankenberger (1993a) suggested that the more stable auxins may have a greater effect on plant growth and yield when compared with auxins of low stability. In brief, the presence of auxins in soil is of ecological importance, particularly when it is present in the rhizosphere for plant uptake. In v i m studies have revealed that specific nutrients affect IAA production by microbial isolates. In soil, nutrients are often limiting. Future research is needed to study environmental factors affecting auxin metabolism in soil.
B. GIBBERELLINS Gibberellins are the least-studied PGRs in soil systems, most likely due to the difficulties associated with extraction and analytical detection. However, few studies have demonstrated GASsynthesis in soil. By using TLC and a bioassay, Elwan and El-Naggar (1972) demonstratedthe synthesis of GAs and gibberellin-like substances in sterile soil inoculated with A. chroocaccum. The presence of plant hormones including GAS in root exudates has been demonstrated by Zupancic and Gogala (1980). Moreover, exogenous application of root exudates to A. chroococcum media stimulated GA synthesis by two-fold (Martinez-Toledo et al., 1988), implying that root exudation in the rhizosphere can enhance microbial synthesis of GAS in sztu. Rossi et al. (1984) found greater amounts of GA-like compounds in maize rhizosphere than in nonrhizosphere soil, particularly during seedling emergence. Different organic amendments applied to soil increased the number of free-living nitrogen fixers and production of physiological active substances including GAS or GA-like substances (Moszherin, 1978). Ota (1976) studied the transformationsof free and bound GASin soil in a series of experiments.He found rapid conversion of conjugate GASinto correspondingfree GAS,showing that free GASare reasonably stable in soil.
c. CYT0KI"S Since many soil microorganisms are potential producers of cytokinins (as discussed in previous sections) and root exudates of many plants contain cytokinins
108 MUHAMMAD ARSHAD AND WILLIAM T. FRANKENBERGER,JR. (Jeske, 1976; Prasad et aL, 1980; Kuriger and Agrios, 1977), many investigators have speculated that cytokinins are present in soil. By using the soybean callus bioassay, van Staden and Dimalla (1976) reported the presence of cytokinin-like activity in soils. Gonzales-Lopez et at. (1986) investigated cytokinin synthesis in dialyzed soil medium inoculated with A. vinelandii and detected appreciably high amounts of cytokinins. Analysis of rhizosphere soil for cytokinin-like activity at various stages of the maize growth cycle revealed that cytokinin-active metabolites were present in soil at greater amounts during the flowering stage (Rossi et aL, 1984). The cytolunin concentration is usually greater in the rhizosphere than in free soil, most probably due to an increased microbial population or accelerated metabolism due to the presence of root exudates. Miiller et al. (1989) reported the presence of cytokinins, including [9R]iP, [9R]Z, and (diH)[9R]Z, in the rhizosphere of corn seedlings grown in sterile media and demonstrated that inoculation with A. chroococcum resulted in a several-fold increase in cytokinin content of the rhizosphere. Their study further revealed that concentrations of [9R]iP and (diH)[9R]Z were greatest in the rhizosphere soil but sharply declined with distance away from the plant roots. Cytokinin biosynthesis in soil by rhizosphere microorganisms may increase with the addition of physiological precursors. By using HPLC-UV and a bioassay, Nieto and Frankenberger (1989b) conducted a study to establish the effect of the purine adenine (ADE); the isoprenoid isopentyl alcohol (IA); and a cytokinin-producing bacterium A. chroococcum, on cytokinin biosynthesis in an Arlington soil (coarse-loamy,mixed thermic Haplic Durixeralf).Various concentrationsof ADE, g IA kg-' soil, and A. chroococseparately and in combination with 8.81 x cum (ATCC 9043), were applied as soil treatments. Cytokinin production was evident as early as 2 days and reached a maximum at 5 days posttreatment. An application of 1.35 x lop4 g ADE kg-I soil, 8.81 x g IA kg-' soil, and A. chroococcum enhanced cytokinin production of [9R]Z and t-Z in this soil up to 1S O - and 1.39-fold,respectively, in comparison with controls containing no precursors. Production of [9R]Z (78.0 pg kg-I) and t-Z (22.1 pg kg-' soil) in the presence of ADE and IA was 1.35- and 2.44-fold greater, respectively, with the inoculation of A. chroococcum, when compared with the indigenous microbiota alone. This study is the first to reveal that microbial biosynthesis of cytokinins can be enhanced by applying physiological precursors to soil.
D. ETHYLENE Several field and laboratory investigations have shown that soil is a source of C2H, (see Arshad and Frankenberger, 1992,1993; and Frankenberger and Arshad, 1995).Many studies also demonstratethat soil is a potential sink for C2H4(Abeles et al., 1971; deBont, 1976; Cornforth, 1975; Smith et al., 1973; Arshad and
PLANT GROWTH REGULATORS IN THE RHIZOSPHERE
109
Frankenberger, 1990~).The net amount of C2H4present in soil represents a balance of C2H4production minus C2H4catabolized by the soil microbiota. Moreover, it should be noted that C2H4is also lost from soil to the atmosphere by outward diffusion. Soils vary greatly in their potential to synthesize and accumulate C2H4in the gas phase (Goodlass and Smith, 1978;Babiker and Pepper, 1984;Hunt et al., 1982;Arshad and Frankenberger, 1991b).A number of factors govern C2H4 accumulation in soil and include organic matter, soil pH, nitrogen content, trace elements, aeration, waterlogging, and the redox potential of soil. Both biological and chemical processes contribute to C2H4accumulation in soil. However, C2H4production primarily as a result of microbial activity is the major source of soil C2H4.Studies have demonstrated that sterilization of soil by irradiation or by autoclaving does not completely inhibit C2H4accumulation in soil, indicating that at least some fraction of C2H4is released as a result of chemical reactions (Rovira and Vendrell, 1972; Smith and Restall, 1971; Babiker and Pepper, 1984; Frankenberger and Phelan, 1985a; Harvey and Linscott, 1978; Arshad and Frankenberger, 1990b,c). Arshad and Frankenberger ( 1990b) reported that autoclaved soil treated with L-ethionine resulted in copious amounts of C2H4,indicating that L-ethionine can be decomposed nonenzymatically to C2H4.These results may imply that some compounds present in native soil organic matter could yield C2H4 through nonenzymatic transformations (Smith et al., 1978; Arshad and Frankenberger, 1991b). We have found that the addition of Fe(I1) at greater than or equal to 100 mg kg - * soil promotes the abiotic production of C2H4in soil via an as yet unknown mechanism (Arshad and Frankenberger, 1991b). On the other hand, nonsterilized soils amended with various organic compounds release C2H4 in magnitudes several-fold greater than that released when the same soils are sterilized and treated with a sterile solution of these amendments, suggesting a major role of soil microbiota in production of C,H4 (Babiker and Pepper, 1984; Frankenberger and Phelan, 1985a;Arshad and Frankenberger, 1990b,c, 1991b). Similarly, antibiotic treatments of soil reduces C2H4 synthesis, indicating microbial involvement in C2H4 production (Arshad and Frankenberger, 1990b,c; Frankenberger and Phelan, 1985a).Although it is now well accepted that under natural conditions microorganisms are the primary producers of C2H4in soil, a small fraction may also be contributed by plant roots. In flooded soils, water acts as a barrier to the escape of C2H4produced in roots (Schneider and Musgrave, 1992). The C2H4-producingability of soils is thought to arise from a diverse group of microflora (Lynch, 1972; Lynch and Harper, 1974a,b; Babiker and Pepper, 1984; Arshad and Frankenberger, 1989). The involvement of any specific group of heterotrophs in C2H4 generation from soil is highly dependent on the nature of the substrates available and prevailing environment factors (Arshad and Frankenberger 1990b,c). Thus, it is highly likely that both bacteria and fungi contribute to C2H4formation in soil, but their relative efficiency is dependent on the prevailing soil conditions and the nature of substrates available. No other microorganism known today is capable of producing C2H4from MET
110 MUHAMMAD ARSHAD AND WILLIAM T. FRANKENBERGER,JR. via ACC (a pathway operative in plants) except only one report of Rhizobium (Beard and Harrison, 1992). However, the addition of ACC to soil resulted in an outburst of C,H, production (Frankenberger and Phelan, 1985a,b; Arshad and Frankenberger, 1990b), indicating the ability of soil microflora to generate C,H, from ACC.
E. ABSCISICACID There are very few studies on the accumulation of ABA in soil. Miiller et al. (1989) demonstrated that rhizosphere microflora were unable to produce ABA in culture media; however, ABA and other phytohormones were detected in maize rhizosphere and adjacent bulk soil. They found a high concentration gradient of ABA from the rhizoplane to rhizosphere and bulk soil. This gradient decreased as the distance from the rhizosphere increased. They also reported that the ABA concentrations in the rhizosphere soil changed during the growth period and were correlated with a particular growth stage or an external influence. Since ABA appears during a period of severe water stress, the roots could have possibly released ABA to the rhizosphere soil. Miiller et al. (1989) demonstrated that the ABA content in the rhizosphere of corn seedlings inoculated with A. chroococcurnwas several-fold greater than the rhizosphere of uninoculated sterile seedlings grown in sterile media. However, the root exudates could have possibly released some precursor(s) for ABA synthesis by A. chroococcum. Miiller et al. (1989) reported that growth effects of bacteria on various maize root parameters could be sufficiently described by the combined effect of IAA and ABA. Hartung et al. (1990) found ABA in aqueous extracts of desert soils. Assuming that the ABA extracted was dissolved in soil water, the concentration may be as high as 40 nM. In a compost soil, ABA is metabolized rapidly and as a result is kept at low concentrations. In a desert soil, which was very rich in NaCl and CaCl,, metabolism is reduced, especially when the soil water content is low; thus, ABA concentrations can be relatively high (Hartung and Davies, 1991).
VII. ECOLOGICAL SIGNIFICANCE OF PGRs PRODUCED IN THE RHIZOSPHERE Production of PGRs is widespread among many soil and rhizosphere microorganisms. Many of these microorganisms synthesize more than one PGR. Numerous studies have shown an improvement in plant growth and development in response to seed or root inoculation with various microbial inoculants. It has been proposed that PGRs released as secondary metabolites by these inoculants may
PLANT GROWTH REGULATORS IN THE RHIZOSPHERE
1 11
contribute to plant growth-promoting effects, but only a few studies have directly demonstrated this cause-and-effect relationship. Some of these studies have been discussed in preceding sections on free-living diazotrophs (Azotobacter and Azospirillium,) Rhizobium and mycorrhizae.
A. PGRs PRODUCED BY AN INOCULUM Loper and Schroth (1986) reported that two bacterial strains belonging to Enterobacteriaceae were capable of producing copious amounts of IAA, leading to reduced plant root elongation and an increased shoot-root ratio in sugar beet when applied as seed inoculants. A significant linear relationship was observed between IAA accumulation of the rhizobacterial strains and decreased root elongation and increased shoot-root ratios of the sugar beet seedlings. l? syringae pv. savastanoi strains active in IAA production also caused a significant decrease in root elongation and increase in shoot-root ratios, whereas strains deficient in IAA production had no effect. Similarly, Duberkowsky et al. (1993) used two isogenic PseudomonasJEuorescens strains differing in IAA production to inoculate softwood cuttings. They observed that bacterial IAA production had a stimulatory effect on the root development of black currant softwood cuttings, whereas for sour cherry, it was inhibitory. They also found that the size of the population of the inoculated strain on the root surface of the cuttings correlated with the effect observed, indicating a cause-and-effect relationship. Gutierrez-Maiiero et al. (1996) detected auxin-like compounds in culture filterate of two PGPR, Bacillus pumilus and B. licheniformis (1.736 and 1.790 mg L-' culture growth medium), and they compared the effects of exogenously supplied auxins and the bacterial culture filtrate on the growth of Alnus glutinosa. Results revealed that application of IAA at 1.8 mg L-I had similar effects on aerial length, aerial surface, and number of leaves of both nodulated and nonnodulated plants of A. glutinosa. Total nitrogen was comparable in the case of nodulated plants, but it was higher in bacterial-metabolites-treated nonnodulated plants than in IAA treated plants. Recently, Xie et al. (1996) compared the effects of a wild-type strain of Pseudomonasputida GR 12-2 and its mutants, which overproduce IAA (up to four times more than its parent strain), on root elongation of canola. They found that the mutant that produced the highest amount of IAA inhibited root elongation. This was most likely due to the production of more IAA than the threshold level required, or the increased IAA may have stimulated synthesis of ACC, which resulted in higher levels of C,H,, causing inhibition to root growth. Haahtela et al. (1990a) found that inoculation with different enterobacterial isolates or treatment with a cell free ethyl acetate extract (pH 7.0) significantly increased the number of root hairs in Poa pratenesis. The bioactive compound causing increased root hair production was identified as IAA by GC-MS. Recently, Noel etal. (1996) conducted a study in which they used
112 MUHAMMAD ARSHAD AND WILLIAM T. FRANKENBERGER,JR. parent and mutant strains of Rhizobium leguminosarum (effective and defective in N, fixation) to inoculate the seeds of canola and lettuce (both nonlegumes). They observed that inoculation with certain strains significantly increased the root growth of seedlings under gnotobiotic conditions. Since auxotrophic Rhizobium mutants requiring TRP (the IAA precursor) or adenosine (the cytokinin precursor) did not promote growth to the extent of the parent strain, they concluded that the growth-promoting effect appears to be direct, with the possible involvement of PGRs, IAA, and cytokinins. Oberhansli et al. (1990) found that a wild-type strain of F! juorescens produced 10-fold more IAA than did its mutant in the presence of 1mM L-TRP, When wheat plants grown in artificial soil (pH 6.2) and fertilized with ammonium-N were inoculated with both strains, there were significant increases in the shoot-root ratios inoculated with the wild-type strain compared with the mutant strain. Holl et al. (1988) investigated the effect of inoculation with Bacillus polymyxa (a soil diazotroph) on crested wheat grass, perennial rye grass, and white clover. Plant growth responses (root and shoot dry weights, root-shoot ratios, and seedling emergence) varied from slightly negative to highly positive. Based upon the characteristics of the inoculum studied, Holl et al. (1988) concluded that production of plant growth-promoting compounds similar to IAA by the bacterium were the likely stimuli for the observed increase in plant productivity. Grapelli and Rossi (198 1) detected IAA, cytokinin-like, and gibberellin-like substances in culture filtrates of Arthrobacter sp. and reported that treating Lactuca sativa seedlings with these bacterial hormones resulted in increased plant development. Grayston et al. (1990) found that inoculation with various microbial isolates promoted wheat growth in both sterile and nonsterile systems, but the effect was cultivar specific. Supernatants of some rhizobacteria were found to be as effective as live cells in achieving growth promotion. IAA was detected in the culture supernatant of these isolates, indicating that phytohormone production may be the mechanism of action for the enhanced plant growth response (Grayston et al., 1990). A recent comprehensive study conducted by Muller et al. (1989) revealed that morphogenetic effects caused by rhizosphere bacteria may be a result of different ratios of plant hormones produced by roots as well as by rhizosphere bacteria. By considering the potential of various associative N,-fixing bacteria to derive auxins from TRP and the presence of TFW in wheat exudates, Kravchenko et al. (1991) concluded that some strains of N,-fixing bacteria can produce physiologically active auxin concentrations. A free-living bacterium, Panroea agglomerans, isolated from the phyllosphere of winter wheat increased the yield and root length of winter wheat upon inoculation. This bacterium was also found to be a potential producer of auxins and cytokinins (Scholz-Seidel and Ruppel, 1992). This may imply that production of PGRs by this bacterium is responsible for creating the physiological response in winter wheat. Wilkinson et al. (1994) compared the effects of seven isolates of orchid-associated bacteria with that of pure hor-
PLANT GROWTH REGULATORS IN THE RHIZOSPHERE
113
mones on mycorrhizal-assisted germination of the terrestrial orchid Pterostylis vitta. Each species of bacteria produced IAA in vitro, whereas the mycorrhizal fungus did not. They concluded that enhancement of symbiotic germination development may have resulted either from production of IAA by the bacteria and/or by the induction of endogenous hormones in the orchid by the metabolites of the bacteria andor mycorrhizal fungus (Wilkinson et al., 1994). The discovery of GAS as a class of natural hormones represents a classic example of the microbe-plant interaction. However, this interaction leads to pathogenicity. The effects of soil GASreleased by a nonpathogen on the physiology of host plants has been studied by some workers. Bottini et al. (1989) identified GA,, GA,, and iso-GA, in gnotobiotic cultures of Azospirillurn lipoferurn strain A1 op 33 by GC-MS. Later, they found that roots of maize seedlings inoculated with this strain and two other strains (A1 iaa 320 and ATCC 29708) contained GA, in a free acid fraction, whereas in the noninoculated roots GA, was detected after hydrolysis of a fraction expected to contain glucosyl conjugates (Fulchieri et al., 1993). This provided direct evidence that A. lipoferurn inoculation affects the GAS status of maize seedling roots. Thus, GAS may be involved in the beneficial effects of Azospirillum spp. in seedling root growth of Graminae. Similarly, Yakushkina and Tarasenko (1975) reported that corn plants cultivated under sterile conditions had low contents of auxins and gibberellin-likesubstances with subsequentless growth than the control (nonsterile) plants. Reinfection of sterilized seeds with epiphytic microorganisms led to an increase in endogenous hormones (including GAS) and increased growth rate. Microbially produced cytokinins may play an important role in stimulating plant growth and development. Barea et al. (1976) found that 90% of 50 bacterial isolates from several plants belonging to different genera were potential producers of cytokinin-likesubstances. Such results strongly suggest that the rhizosphere, a site rich in carbon flow with root exudates and microorganisms, would be an ideal habitat for cytokinin synthesis.After screening rhizobacterial strains for cytokinin production and P solubilization, Young et al. (1992) examined their influence on wheat and found significant yield increases upon inoculation with strains that had either phosphate solubilizing activity or overproduced (diH)[9R]Z in vitro. Similarly, Scholz-Seidel and Ruppel (1992) detected cytokinins and auxins in the supernatant of a free-living bacterial strain, Pontoea agglomarans, isolated from the phyllosphere of winter wheat that was able to increase winter wheat yield and root length in sterile and nonsterile experiments. They speculated that the increased growth of inoculated plants might be related to the hormone production ability of the inoculant. Reddy et al. (1991) screened various rhizobacteria for their biocontrol activity. They reported that bacterial strains with biocontrol activity induced root elongation on cucumber and tomato and also produced (diH)[9R]Z and iPA, suggesting a role for PGRs in the bacterial influence on plant growth. Hoflich (1992) also observed stimulated growth and yield of winter wheat, winter rape, oil
114 MUHAMMAD ARSHAD AND WILLIAM T FRANJCENBERGER,JR. radish, mustard, and peas in pot and field experiments in response to inoculation with PseudomonasPuorescens (isolated from rhizosphere of winter wheat) and attributed the effects to be hormone-like, including cytolunins and auxins. In another study, Hoflich and Weise (1992) reported that the combined inoculation of a peaplant specific Rhizobium leguminosarum biovar viceae strain together with a red clover strain, R. leguminosarum biovar trifofic,increased the shoot dry matter, root growth, grain yield, nodulation, N, fixation, and leghemoglobin content in the nodules of pea in field experiments. Since the red clover strain cannot produce nodules on pea plants, they speculated that the growth promotion was a result of the water-soluble cytokinins released by this strain. Ethylene produced by soil microorganisms, released by plant roots, and formed abiotically from soil organic matter contributes to the C,H, pool of soil. Since C,H, is biologically active within an extremely low concentration range, it is most likely that C,H, present in the soil atmosphere surrounding the roots can affect plant growth and crop productivity. This hypothesis is supported strongly by the plant response to an exogenous source of C,H, applied to the roots. Freytag et al. (1972) reported that soil-injected C,H, improved the yields of cotton and sorghum. Similarly, Eplee (1975) showed that C,H, injected into soil triggered suicidal germination of witchweed up to 90%, suggesting a role for C,H, in weed eradication. The unequivocal proof that C,H, uptake by plant roots can move to the shoots was demonstrated by Jackson and Campbell (1975), who reported rapid movement of labeled C,H, in tomato. Recently, Glick and his co-workers have conducted studies to determine the mechanism responsible for growth-promoting effects of a PGPR, Pseudomonas putida GR12-2 (Glick et al., 1994a,b; Hall et af., 1996;Xie et al., 1996).They discovered that I! putida GR 12-2 contains the enzyme ACC deaminase (Jacobson et al., 1994; Glick et al., 1994a,b). This enzyme hydrolyzes ACC, the immediate precursor of C,H, in higher plants (Yung et al., 1982). Glick and his associates investigated the role of ACC deaminase in growth-promoting activity of this bacterium by using its ACC deaminase mutants (Glick et al., 1994a,b; Hall et al., 1996). Glick et al. (1 994a,b) reported that only the wild-type cells of F? putida GR 12-2, but not any of the ACC deaminase mutants, promoted root growth of developing canola seedlings under gnotobiotic conditions. In another study, similar observations were made in the case of lettuce, tomato, and wheat, in addition to canola (Hall et al., 1996). Hall et al. (1996) also used an inhibitor of C,H,, AVG, for comparison and reported similar effects of AVG and the wild-type bacterium. All of these results have led Glick and co-workers to postulate that following binding of I? putida GR 12-2 to the seed coat, this enzyme (ACC deaminase) might act to stimulate plant growth (in particular, root elongation) by sequestering and then hydrolyzing ACC from germinating seeds, thereby lowering the endogenous level of ACC and, hence, the level of C,H, in seeds, which subsequently results in growth promotion (Fig. 10).This model was later confirmed by comparing the ef-
elongation and proliferation
"\
IAA
+- - IAA
ACC
+ACC
1
1
ACC oxidase
Ethylene
ACC deaminase
x- Ketobutyrate
ammonia
Root elongation
Seed or root
Bacterium
Figure 10 Model explaining PGPR stimulation of plant root elongation. In this model, IAA synthesized by a bacterium that is bound to the surface of either the seed or root of a developing plant is taken up by the plant and, in conjunction with the endogenous plant IAA, can then either stimulate cell proliferation and/or elongation or the activity of the enzyme ACC synthase. In the latter instance, ACC synthesis within the plant is increased. A significant portion of this ACC may be exuded from plant roots or seeds (along with other small molecules normally present in seed or root exudates), actively taken up by the bacterium, and hydrolyzed by the enzyme ACC deaminase to a-ketobutyrate and anmonia. The bacterium causes the plant to synthesize more ACC than the plant would otherwise need. thereby providing the bacterium with a unique source of nitrogen (i.e., ACC). A direct consequence of lowering the level of ACC within the plant (either the endogenous level or the IAA-stimulated level) results in a lowering of the amount of C,H, within the plant and a decreased extent of C,H, inhibition of plant seedling root elongation (from Hall el ul., 1996, used with permission).
1 16 MUHAMMAD ARSHAD AND WILLIAM T. FRANKENBERGER, JR. fects of overproducingIAA mutants of P. putida GR 12-2 to that of the wild type strain on canola root elongation (Xie et al., 1996).They found that the mutant that produced the highest amount of IAA inhibited root growth of canola most likely by stimulating ACC synthesis, which resulted in a higher endogenous C,H, level. Glick et al. (1995) has suggested that this trait, i.e., presence of ACC deaminase, can be used as a rapid-screening technique for isolating PGPR. Few studies have demonstrated indirectly the role of ABA in symbiotic associations such as mycorrhizae and nodulation, as discussed previously. Miiller et aE. (1989) has demonstrated the vital role of ABA and other phytohormones present in the rhizosphere and in plant-microbe interactionsaffecting plant growth and development. ABA, particularly in combination with other phytohormones, can be used for improvement of plant growth, particularly under water-deficit conditions.
B. PRECURSOR-INOCULUM INTERACTIONS The beneficial effects of inoculation with PGPR on plant growth are well documented in the existing literature; however, there is a lack of consistency in reproducibility. To improve the effectiveness of inoculation and enhance reproducibility, Frankenberger and co-workers developed an approach based upon the hypothesis that inoculation in the presence of a specific physiological precursor is more effective than inoculation alone (Frankenberger and Arshad, 1995; Arshad and Frankenberger, 1993). By using this approach, the synthesis of a particular PGR in the rhizosphere can be controlled through precursor-inoculum interactions, evoking a physiological response. Several published studies have demonstrated the success of this approach, as discussed later. Low availability of TRP could be the most limiting factor for auxin production in the rhizosphere of plants. Root exudates are the only natural source of TRP for rhizosphere microflora. Recently, detectable amounts of TRP in root exudates of some but not in all varieties of wheat have been reported (Kravchenko et al., 1991; Martens and Frankenberger, 1992, 1994). This indicates that not all plants can release adequate quantities of TRP into the rhizosphere for microbial production of auxins. Auxin production in the root zone is limited by the genetic and physiological properties of both the microorganism(s) and the plant(s). Thus, effective utilization of microbial-produced auxins requires careful selection of both partners. Moreover, the presence of other amino acids (e.g., glutamine, a-ketoglutaric acid, and pyruvic acid) and growth factors may affect auxin production in the rhizosphere, and rhizosphere microflora may be capable of catabolizing auxins. Thus, intensive research is needed to explore auxin production in the rhizosphere. TRP is considered the physiological precursor of auxin for both plant and soil microorganisms. Its addition to soil has been shown to increase auxin content and
PLANT GROWTH REGULATORS IN THE RHIZOSPHERE
1I 7
influence physiological effects on plant growth via microbially-derived auxins. Frankenberger and his co-workers have conducted many experiments to evaluate this hypothesis. After confirming L-TRP-dependent IAA synthesis by an ectomycorrhizal fungus, Pisolithus tinctorius, using TLC, HPLC, ELISA, and GC-MS, Frankenberger and Poth (1987a) studied the influence of L-TRP and the fungus as an inoculum on Douglas fir (Pseudotsuga rnenziesii).They found that I? tinctorius stimulated the growth of potted seedlings of Douglas fir only when supplied with L-TRP. Analyses of variance for seedling height, stem diameter, shoot-root dry weight, and root-shoot ratio showed significant differences among treatments. There was basically no difference in growth between the inoculated and uninoculated treatments in the absence of L-TRP; however, the addition of a dilute solution of L-TRP with I? tinctorius promoted a dramatic response. The higher concentration of L-TRP applied at M inhibited the dry weights of the seedlings, which is a typical auxin response. Root examination revealed that the mycelial inoculum of I? tinctorius was highly effective in forming ectomycorrhizae on the seedlings. However, the effectiveness of this isolate on plant growth was not much different from that of the noninoculated stock seedlings unless L-TRP was provided in the nanogram-to-microgram range as a precursor of IAA. Zahir et al. (1997) also investigated the effectiveness of precursor-inoculum interactions. After confirming the ability of Azotobacter to catabolize TRP into auxins in vitro, they studied the effect of Azotobacter inoculation, both in the presence and absence of TRP, on potato yield under fertilized conditions. They reported that the combined application of Azotobacter inoculation and TRP was more effective than their application alone in increasing the tuber and straw yield of potato. Similarly, Arshad et al. (unpublished data) found that the combined application of Azotobacter (capable of producing auxins) and L-TRPincreased wheat yield much more than their use alone. Sarwar and Kremer (1995b) employed the precursorinoculum approach for the biological control of weeds. They reported that Enterobacter taylorae with high L-TRP-derived auxin producing potential (72 mg LIAA equivalents) inhibited the root growth of field bindweed seedlings by 77.4%. Application of L-TRP ( l O P 5 M ) alone reduced the root growth by 18.3%,whereas supplementing L-TRP(at lops M> with the inoculum further inhibited root growth (by 90.5%). They suggested that providing L-TRP with the selected auxin-producing inoculum may be a practical biological control strategy against weeds. A similar response was observed in the case of other plants, including red clover, wheat, velvet leaf, pigweed, green foxtail, morning glory, corn, and soybean, most likely because of higher-than-optimum required production of auxins as a result of precursor-inoculum interactions (Sarwar and Kremer, 1995b). These four studies demonstrate the superiority of precursor-inoculation interactions over inoculation alone for improving crop yields. Few studies have been conducted on searching for physiological precursors for
'
118 MUHAMMAD ARSHAD AND WILLIAM T. FRANKENBERGER,JR. microbial biosynthesis of GAS. MVA is known to be the precursor of GA biosynthesis in plants and G. fujikuroi; however, it is not specific for GASbecause many other metabolites are derived from MVA. It may not serve as a universal substrate for synthesis of GASby other microbes. Moreover, its high cost restricts its use as an agrochemical to enhance GA synthesis in the rhizosphere of plants. Although GA, is commercially available to improve crop yield and quality, it is not widely used because of its expense. It is of prime importance to screen economical physiological precursors to be used as soil amendments to promote GA synthesis in the rhizosphere. Nieto and Frankenberger (1990b) investigated the effect of adenine (ADE), isopentyl alcohol (IA), and the cytokinin-producing bacteriumA. chroococcum on the morphological plant characteristics of Raphanus sativus (radish) in sand under axenic-inoculated conditions and in soil under glasshouse and field conditions. The combined application of ADE, IA, and the bacterial inoculum enhanced the dry weight of root and shoot tissues, leaf area, and chlorophyll a content to a much greater degree than in the presence or absence of the cytokinin precursors (ADE or IA) or the bacterium alone. Enhanced plant growth observed under axenic conditions upon the addition of ADE and IA indicated that the plant has the ability to assimilate and utilize ADE and IA for growth and metabolism. Although the addition of the inoculum without precursors was also stimulatory, growth promotion was not stimulated to the same degree as in the presence of the two precursors and A. chroococcum. Greater enhancement of plant growth was observed following the application of ADE, IA, and A. chroococcum, which was attributed primarily to the increase in microbial production of cytokinins within the rhizosphere. In another study, Nieto and Frankenberger (1991) evaluated the effect of ADE, IA, and A. chroococcum applied to soil on the vegetative growth of Zea mays studied under glasshouse conditions.A combined treatment of ADE, IA, and the bacterial inoculum enhanced the vegetative growth of maize to a greater degree than did the application of ADE plus IA; ADE plus A. chroococcum; or ADE, IA, or A. chroococcum alone. The dry mass of the root and shoot tissues was enhanced up to 5.6-and 5.0-fold, respectively, in comparison with the controls; however, the root-shoot ratios were similar. The increase in shoot height, internodal distance, and stem and leaf width over the controls under the optimal treatment (ADE, IA, and A. chroococcum) were 2.07-, 2.82-, 1.46-, and 1.70-fold,respectively. . The ecological significance of microbially produced C2H4 under nonwaterlogged conditions has not been evaluated until recently by Arshad and Frankenberger (1988). By using a C2H4 bioassay, the classical “triple” response in etiolated pea seedlings, Arshad and Frankenberger (1988) demonstrated that C2H4of microbial origin can affect plant growth. They showed that MET-dependent C,H4 produced by an inoculum, Acremonium falciforme, or by soil-indigenous microflora caused the classical “triple” response in etiolated pea seedlings, including reduction in elongation, swelling of the hypocotyl, and a change in direction of
PLANT GROWTH REGULATORS IN THE RHIZOSPHERE growth (horizontal). A similar response was observed C,H, gas.
119
direct application of the
C. EXOGENOUS APPLICATION OF PHYSIOLOGICAL PRECURSORS OF PG& The addition of physiological precursors of various PGRs to soil results in the synthesis of PGRs in much higher amounts. Thus, the addition of these precursors to the rhizosphere can stimulate PGRs production by indigenous rhizosphere microflora and evoke a plant response. A new research field has now been established with the use of precursors of PGRs applied to soil for the betterment of crop production. Frankenberger et al. (1990) reported the physiological response of radish (R. sativus) to L-TRP applied to soil under optimal nutritional conditions. They observed a positive effect of L-TRP on growth parameters of radish when applied at low concentrations at the seedling stage, comparable to that of selected auxins (IAMand ILA). Foliar application of L-TRP had no effect, indicating that the site of entry of TRP and/or its microbial-transformed metabolites appears to be through the root zone. High concentrations of L-TRP had negative effects, a response usually observed in the presence of high auxin levels. The plant response was more pronounced under high fertility status, which excluded any nutritional effect of LTRP. Sarwar and Frankenberger (1994) compared the effects of soil-applied LTRP with auxins (IAM, TOL, and IAA) on growth of corn (Z. mays). They found that L-TRP applied at an appropriate concentration could have positive effects on corn growth comparable to pure auxins (TOL and IAA). Phenylacetic acid (PAA) is another known auxin, and L-phenylalanine (PHE) has been reported as the physiological precursor of PAA. Sarwar and Frankenberger ( 1995)reported that L-PHE as a soil amendment promoted microbial production of PAA and phenylpyruvic acid. The addition of PHE to the uniform seedlings of maize caused a significant effect in vegetative growth parameters most likely via its conversion into PAA by rhizosphere microbiota. Similarly, increased growth rates and increased yields in two of three varieties of wheat were observed in response to L-TRP (lops to lo-’ M) treatment applied to the root zone (Martens and Frankenberger, 1994). Since there was poor direct uptake of labeled L-TRP compared to labeled IAA by wheat seedling roots, the positive response was attributed to microbial production of auxins within the rhizosphere upon the addition of the precursor, L-TRP. Yield response of watermelon (Citrullus lanatus) and cantaloupe (Cucumis melo) to LTRP applied to soil was evaluated in the field (Frankenberger and Arshad, 199 1a). The fruit yield (number and weight of melons) was enhanced by low concentrations of L-TRP added at the seedling stage. The positive effect of L-TRP on melon yield was most apparent at the later stages of growth (third harvest), which fur-
120 MUHAMMAD ARSHAD AND WILLIAM T FRANKENBERGER,JR. ther reveals a physiological effect. Frankenberger and Martens (1995) found that exogenous application of auxins (IAM, TAM) and L-TRP to soil significantly increased the number of harvested Tiffany melons, increased both weight and harvested number of Picnic melons, and increased the uniformity of the harvested melons in both varieties compared with control plants. Similarly, application of LTRP to soil significantly improved the yield of bell peppers (Capsicum annuurn) (Frankenberger and Arshad, 1991b). The effect was more obvious at the third harvest. The number of fruit and fresh weight of fruit per plant were highest at low concentrations (ng g-') of L-TRP applied. The direct application of L-TRP in solution to soil surrounding the seedlings established in the field is often unsuccessful in promoting the growth of crops. The yield of watermelon, cantaloupe, and bell pepper was significantly increased only when L-TRP was added to contained cells and the seedlings were temporarily limited in space for root proliferation, forcing the uptake of L-TRP and/or its metabolites before transplanting. Recently, Arshad et al. (1995a,b) investigated the effect of soil-applied L-TRPon growth and yield of cotton and soybean. They reported a significant positive growth response to the soil-applied auxin precursor. The hypothesized mechanisms of action of exogenously applied L-TRPon plant growth include: (1) substrate-dependent auxin production in soil by the indigenous microflora, (2) uptake directly by plant roots followed by metabolism within their tissues, and/or (3) a change in the balance of rhizosphere microflora affecting plant growth (Frankenberger et al., 1990; Frankenberger and Arshad, 1991a,b). However, recent studies have demonstrated that TRP may not serve as a physiological precursor of auxins in higher plants (Wright etal., 1991; Baldi et al., 1991). Moreover, Martens and Frankenberger (1992, 1994) reported poor uptake of labeled TRP by the wheat seedling roots compared to labeled IAA. Thus, the hypothesis of substrate-dependent microbially derived auxin in the rhizosphere affecting plant growth and development seems to be a more plausible explanation. Soil-applied GA, (lO-'O to M) has been found to significantly promote plant height and root growth of dwarf maize (Frankenberger and Arshad, 1995), revealing that GAS present in soil surrounding the roots can be of ecological significance regarding plant growth: In another study, soil application of GA, ( to lop3 M) four weeks prior to seedling transfer significantly enhanced the plant height of dwarf maize (Frankenberger and Arshad, 1995). This indicates that GAS in soil may remain available for plant utilization for a reasonably long period. The agriculture industry could possibly benefit by utilizing economic substrates of cytokinins and optimizing the environment that stimulates in vivo synthesis of cytokinins. Arshad et al. (1995b) studied the effect of ADE applied to soil on soybean yield and reported a significant positive response. Arshad and Frankenberger (1990a) found that soil application of L-MET affected the vegetative growth and resistance to stem breaking (lodging) of two cultivars of corn. Similarly, soil treatment with L-ethionine resulted in a significant
PLANT GROWTH REGULATORS IN THE RHIZOSPHERE
12 1
epinastic response, enhanced fruit yield, and early fruit formation and ripening in tomato (Arshad and Frankenberger, 1990a).In another study,Arshad and co-workers (1993) found that Albiziu lebbeck L. responded positively to low to moderate concentrations of L-MET applied to soil. Growth parameters, including plant height, girth, dry weight of roots, total biomass, number of nodules, and dry weight of nodules, were promoted significantly in response to various L-METconcentrations. Similarly, plant N, P, and K concentrations and their uptake were also enhanced by treatment with L-MET.A significant quadratic dose-response relationship was found in all cases when each individual parameter was regressed against log [L-MET]excluding the control. Similarly, in another study, a significant yield response was observed with soybean exposed to L-METapplied to soil (Arshad et ul., 1995b). The proposed mechanism of action was substrate-dependent C,H, production by the indigenous soil microflora.
VIII. CONCLUSIONS This chapter clearly indicates that production of PGRs is widespread among rhizosphere microflora, including both pathogenic and beneficial microorganisms. However, conditions in the rhizosphere are often quite variable, and many factors, such as temperature, pH, availability of nutrients, and composition and amount of root exudates, can affect the synthesis of PGRs by microorganisms associated with plant roots. Because substrate(s) concentrations in root exudates vary with plant species, it seems reasonable to suggest that production of physiologically active levels of PGRs by microorganisms within the rhizosphere can be regulated by the plant with which they are associated. Less-than-optimalor unfavorable conditions may lead to little or no synthesis of PGRs in the root zone by the rhizosphere microorganisms.Additional input to the plant’s PGRs pool by rhizosphere microorganisms could modify endogenous hormones to either optimal or supraoptimal levels, resulting in the modification of plant growth or pathogenesis (Patten and Glick, 1996). Since the endogenous level of PGRs varies with the age of the plant, the plant response to exogenously supplied PGRs would depend on the time of exposure. Moreover, the amount of microbially derived PGRs available to influence the plant is subject to several levels of regulation. Microbial hormone synthesis within plant cells may result in high concentrations of PGRs, whereas secretion by bacteria living outside plant cells may be less available since they would be subject to diffusion, adsorption, and degradation. It is the extent to which microbial sources of PGRs modify endogenous levels of free and bound PGRs in the plant that ultimately determines whether the relationship is beneficial or deleterious. Optimal levels of PGRs often enhance growth, whereas supraoptimal levels may elicit a disease response.
122 MUHAMMAD ARSHAD AND WILLIAM T. FRANKENBERGER,JR. Despite the fact that piants are capable of synthesizing phytohormones, they may also respond to exogenous applications during certain growth phases and under certain cultivation conditions. Plants may not have the capacity to synthesize sufficient endogenous phytohormones for optimal growth and development under suboptimal climatic and environmental conditions. Moreover, plants may store excess amounts of phytohormones, including exogenous sources as conjugates for later use. The effects of PGRs have been elucidated largely from exogenous applications (Arshad and Frankenberger, 1993; Frankenberger and Arshad, 1995). It should be emphasized that PGRs do not act alone but rather interact with one another in a variety of complex ways (Barendse and Peeters, 1995). Thus, the plant response to exogenously supplied PGRs is regulated by the net balance of PGRs supplied as well as the net balance of endogenous PGRs. The use of microbial inoculants such as PGPR is a now common practice in many regions of the world. Among the proposed mechanisms of action of PGPR, production of PGRs in the rhizosphere is considered the most plausible one. During recent years, advances have been made in detecting and characterizing PGRs produced by various microbial isolates in vitru. Many studies have shown a direct correlation between the plant response and prolific production of PGRs by microorganisms, particularly for symbiotic and parasitic associations. However, it is difficult to interpret and understand such interactions. These studies reveal that PGRs of microbial origin could be of valuable ecological significance to the agricultural industry. Microbially released PGRs are not only economical but also provide a continuous supply that may prove better than a one-time application of synthetic PGRs. There are numerous unresolved important aspects concerning the involvement of PGRs of microbial origin in evoking physiological responses in plants. Further research is needed to study environmental factors affecting in situ synthesis of PGRs in the rhizosphere, relations between concentrations of PGRs in the rhizosphere and plant responses, the influence of exogenous PGRs in a nutrient-limited matrix, stability and kinetics of PGR production and microbial degradation in soil, physicochemical adsorption and chemical tranformations of PGRs that promote their synthesis in the rhizosphere soil, agronomic practices that promote the synthesis and stability of PGRs in the soil-root environment and enhance the availability and uptake of PGRs by plant roots, and interaction of rhizosphere PGRs with other biologically active substances and their effects on microbial activities. Further work in this discipline may provide a better understanding of the mechanisms of action of microbially derived PGRs and their interactions with plants. Moreover, development of hormone-deficient plant mutants can provide opportunities to clearly define the role of microbially produced PGRs in plan-microbe interactions. An understanding of these aspects can aid in the utilization of microbial PGRs for the betterment and benefit of sustainable agriculture.
PLANT GROWTH REGULATORS IN THE RHIZOSPHERE
123
APPENDIX: ABBREVIATIONS a-AA ABA ACC ADE y-AA AMO- 1618 AMP auxl and aux2 AVG bum C*H, 4-C1-IAA
ccc
CLS CPP DMAPP (diH) [9R]Z (diH)Z 1',4'-diH-y-AA 1',4'-diol ABA 1'-deoxy ABA a-ET EDTA EFE y-ET ELISA FP FPP G*3
GAS GC-MS GGPP GLS 4'-OH-a-ET 4I-OH-y-AA 4I-OH-y-AA
a-ionylideneacetic acid abscisic acid I -aminocyclopropane-1 -carboxylic acid adenine y - iony lideneacetic acid 2-isopropyl-4-diimethylamino-5-methylphenyl1-piperidhe carboxylate methyl chloride adenosine-5'-monophosphate plasmid-borne gene of Agrobacterium rhizogenes encoding for auxin biosynthesis aminoethoxyvinylglycine plasmid-borne gene encoding for IAM hydrolase activity ethylene 4-chloroindole-3-acetic acid 2-chloroethyltrimethyl ammonium chloride (chlormequat) cytokinin-like substance copalyl pyrophosphate dimethylallyl pyrophosphate dihydrozeatin riboside (ribosyl dihydrozeatin) dihydrozeatin 1',4'-dihydroxy-y-ionylideneaceticacid 1 ',4'-diol abscisic acid 1'deoxy abscisic acid a-ionylidene ethanol ethylenediaminetetraacetic acid ethylene-forming enzyme y-ionylidene ethanol enzyme-linked immunosorbent assay farnesyl phosphate farnesyl pyrophosphate gibberellic acid gibberellins gas chromatography-mass spectrometry geranylgeranyl pyrophosphate gibberellin-like substance 4'-hydroxy-a-ionylidene ethanol 4'-hydroxy-y-ionylidene acetic acid 4'-hydroxy-y-ionylideneaceticacid
124 MUHAMMAD ARSHAD AND WILLIAM T FRANKENBERGER,JR. HMBA HPLC I AAld IA IAA iaaH iaaM
IAAsp IAld IAM IAN IBA ICA IGA IGoxA LA IM iP iP [9R]iP PA iPP ipt I4rA a-KB KGA JSMBA MACC [2MeS]iP [2MeS 9R]Z [MeSIZ MET MVA NAA NAD NADH NADP NADPH NAM
2-hydroxy-4-methylbiobutyric acid high-performanceliquid chromatography indole-3-acetaldehyde isopentyl alcohol indole-3-acetic acid plasmid-borne gene of Pseudomonas syringae pv. savastonoi encoding for indole-3-acetamidehydrolase plasmid-borne gene of €? syringae pv. savastonoi encoding for tryptophan-2-monooxygenase indole-3-acetyl-~-aspartate indole-3-aldehyde indole-3-acetamide indole-3-acetonitrile indole-3-butyricacid indole-3-carboxylicacid indole-3-glycolic acid indole-3-glyoxylicacid indole-3-lactic acid indole-3-methanol isopentenyl adenine isopentenyl adenine isopentenyl adenosine isopentyl alcohol isopentenyl pyrophosphate plasmid-borne gene of Agrobacterium tumefaciens encoding cytokinin biosynthesis indole-3-pyruvic acid a-ketobutyrate a-ketoglutaric acid 2-0x0-4-methylthiobutyric acid N-malonyl-1-aminocyclopropane-1-carboxylic acid 2-methylthio isopentenyl adenine 2-methylthiozeatin riboside methylthiozeatin methionine mevalonic acid a-naphthalene acetic acid nicotinamide adenine dinucleotide nicotinamide adenine dinucleotide, reduced nicotinamide adenine dinucleotide phosphate nicotinamide adenine dinucleotide phosphate, reduced a-naphthalene acetamide
PLANT GROWTH REGULATORS IN THE RHIZOSPHERE PAA PC PGPR PGRs PHE RIA SAM T-DNA TAM Ti plasmid TIP TLC tmS I tmS2
TOL TOM TRP tzs VA mycorrhiza Z r9~1z
12 5
phenylacetic acid paper chromatography plant growth-promoting rhizobacteria plant growth regulators phenylalanine radioimmunoassay S-adenosylmethionine transfer element of DNA tryptamine tumor-inducing plasmid tumor-inducing principle thin-layer chromatography plasmid-borne gene of Agrobacterium tumefaciens encoding for tryptophan-2-monooxygenase plasmid-borne gene of Agrobacterium tumefaciens encoding for indole-3-acetamide hydrolase tryptophol (indole-3-ethanol) tryptophan-2-monooxygenase tryptophan plasmid-borne gene of Agrobacterium tumefaciens encoding cytokinin vesicular-arbuscular mycorrhizae zeatin ribosylzeatin
REFERENCES Abbott, L. K., and Robson, A. D. (1982). The role of vesicular arbuscular mycorrhizal fungi in agriculture and the selection of fungi for inoculation. Ausr. J. Agric. Res. 33,389408. Abdel-Salam, M. S., and Klingmuller, W. (1987).Transposon Tn5 mutagenesis in Azospirillum Iipoferum: Isolation of indole acetic acid mutants. Mol. Gen. Gene?.210, 165-170. Abeles, F. B., Craker, L. E., Forrence, L. E., and Leather, G. R. (1971). Fate of air pollutants: Removal of ethylene, sulfur dioxide and nitrogen dioxide by soil. Science 173,914-916. Achilea, 0..Chalutz, E., Fuchs, Y.,and Rot, I. (1985a). Ethylene biosyntbesis and related physiological changes in Penicillium digitatum-infected grapefruit (Cirrus paradisi). Physiol. Plant Pathol. 26,125-134. Achilea, O., Chalutz, E., Fuchs, Y., and Rot, I. (1985b). The contribution of host and pathogen to ethylene biosynthesis in Penicilliurn digitaturn-infected citrus fruit. Physiof. Plant Pathol. 2 7 , 5 5 4 3 . Akiyoshi, D. E., Moms, R. O., Hinz, R., Mischke, B. S.,Kosuge, T., Garfinkel, D. J., Gordon, M. P., and Nester, E. W. (1983). Cytokinidauxin balance in crown gall tumors is regulated by specific loci in the T-DNA. Proc. Natl. Acad. Sci. USA 8 0 , 4 0 7 4 1 1. Akiyoshi, D. E., Regier, D. A., and Gordon, M. P. (1987). Cytokinin production by Agrobacterium and Pseudomonas spp. J . Bacteriol. 169,4242-4248.
126 MUHAMMAD ARSHAD AND WILLIAM T FRANKENBERGER,JR. Akiyoshi, D. E., Regier, D. A,, and Gordon, M. P. (1989). Nucleotide sequence of the tu gene from Pseudomonas solanacearum strain K60. Nucleic Acid Res. 17,8886. Al-Nimri, L. F., and Coolbaugh, R. C. (1990). Inhibition of abscisic acid biosynthesis in Cercospora msicola by triarimol. J. Plant Growth Regul. 9,221-225. Al-Nimn, L. F., and Coolbaugh, R. C. (1991). Conversion of 1 ’-deoxy 2H-ABA to ’H-ABA in cellfree extracts from Cercospora rosicola. J. Plant Growth Regul. 10,6346. Allen, M. F., Moore, T. S., Jr., and Christensen, M. (1980). Phytohormone changes in Bouteloua gracilis infected by vesicular-arhuscular mycorrhizae: I. Cytokinin increases in the bost plant. Can. J. Bot. 58,371-374. Allen, M. F., Moore, T. S., Jr., and Christensen, M. (1982). Phytohormone changes in Bouteloua gracilis infected by vesicular-arbuscular mycorrhizae. 11. Altered levels of gibberellin-like substances and abscisic acid in the host plant. Can. J. Bot. 60,468-471. Apte, R., and Sbende, S. T. (1981). Studies on Azotobacter chroococcum. IV. Seed bacterization with strains of Azotobacter chroococcum and their effect on crop yields. Zentralbl. Baktiol. Abt. II. Abt. 136,637-640. Archer, S. A,, and Hislop, E. C. (1975). Ethylene in host-pathogen relationships. Ann. Appl. Biol. 81, 121-126. Armstrong, D. J., Scarbrough. E., and Skoog, F. (1976). Cytokinin in Corynebacteriumfascians cultures. Isolation and identification of 6-(4-hydroxy-3-methyI-cis-2-butenylamino)-2-methylthiopurine. Plant Physiol. 58,749-752. Arshad, M., and Frankenberger, W. T., Jr. (1988). Influence of ethylene produced by soil microorganisms on etiolated pea seedlings. Appl. Environ. Microbiol. 54,2728-2732. Arshad, M., and Frankenberger, W. T., Jr. (1989). Biosynthesis of ethylene by Acremoniumfafcifome. Soil Biol. Biochem. 21,633-638. Arshad, M., and Frankenberger, W. T., Jr. (1990a). Response of Zea mays L. and Lycopersicon esculentum to the ethylene precursors, L-methionine and L-ethionine, applied to soil. Planr Soil 122, 219-227. Arshad, M., and Frankenberger, W. T., Jr. (1990b). Ethylene accumulation in soil in response to organic amendments. Soil Sci. SOC.Am. J. 54,1026-1031. Arshad, M., and Frankenberger, W. T., Jr. (1990~).Production and stability of ethylene in soil. Biol. Fen. Soils 10,29-34. Arshad, M., and Frankenberger, W. T., Jr. (1991). Effects of soil properties and trace elements on ethylene production in soils. Soil Sci. 151,377-386. Arshad, M., and Frankenberger, W. T., Jr. (1992). Microbial biosynthesis of ethylene and its influence on plant growth. Adv. Microbial Ecol. 12,69-I 11. Arshad, M., and Frankenberger, W. T., Jr. (1993). Microbial production of plant growth regulators. In “Soil Microbial Ecology” (B. F. Metting, ed.), pp. 307-347. Marcel Dekker, New York. Arshad, M., Hussain, A., and Shakoor, A. (1995a). Effect of soil applied L-tryptophan on growth and chemical composition of cotton. J. Plant Nutrition 18,317-329. Arshad, M., Hussain, A., and Zahir, 2. A. (1995b). Effect of soil applied precursors of phytohormones on growth and yield of soybean. PGRSA Quarterly 23,6349. Arshad, M., Hussain, A., Jared, M., and Frankenberger, W. T., Jr. (1993). Effect of soil applied Lmethionine on growth, nodulation and chemical composition of Albizia lebbeck L. Plant and Soil 148, 129-135. Assante, G., Merlini, L., and Nasini, G . (1977). (+)-Abscisic acid, a metabolite of the fungus Cercospora msicola. Experientia 33, 1556-1557. Atzorn, R., Crozier, A., Wheeler, C. T., and Sandberg, G . (1988). Production of gibberellins and indole-3-acetic acid by Rhizobium phaseoli in relation to nodulation of Phaseoius vulgaris roots. Planta 175,532-538. Avivi, Y., and Feldman, M. (1982). The response of wheat to bacteria of the genus Azospirillum. Israel J. Bot. 31,237-245.
PLANT GROWTH REGULATORS INTHE RHIZOSPHERE
127
Azcbn, R., and Barea, M. (1975). Synthesis of auxins, gibberellins and cytokinins by Azotobacter vinelandii and Azotobacter beijerinckii related to effects produced on tomato plants. Plant Soil 4 3 , 6 0 9 4 19. Azcbn, R., Azc6n, G., De Aguilar, C., and Barea, J. M. (1978). Effects of plant hormones present in bacterial cultures on the formation and responses to VA endomycorrhiza. New Phyrol. 80, 359-364. Baas, R., and Kuiper, D. (1989). Effects of vesicular-arbuscular mycorrhizal infection and phosphate on Plantago major ssp. pleiosperma in relation to internal cytokinin concentrations. Physiol. Plant. 76,211-215. Babiker, H. M., and Pepper, I. L. (1984). Microbial production of ethylene in desert soils. Soil Biol. Biochem. 16,559-564. Badenoch-Jones, J., Rolfe, B. G., and Letham, D. S . (1983). Phytohormones, Rhizobium mutants, and nodulation in legumes. Plant Physiol. 73,347-352. Badenoch-Jones, J., Summons, R. E., Djordjevic, M. A,, Shine, J., Letham, D. S., and Rolfe, B. G. (1982a). Mass spectrometric quantification of indole-3-acetic acid in Rhizobium culture supernatants: Relation to root hair curling and nodule initiation. Appl. Environ. Microbiol. 44,275-280. Badenoch-Jones, J., Summons, R. E., Entsch, B., Rolfe, B. G., Parker, C. W., and Letham, D. S. (1 982b). Mass spectrometric identification of indole compounds produced by Rhizobium strains. Biomed. Mass Spec. 9,429-437. Baldi, B. G., Maher, B. R.,Solvin, J. P., and Cohen, J. D. (1991). Stable isotope labeling, in vivo,Of Dand L-tryptophan pools in Lemna gibba and the low incorporation of label into indole-3-acetic acid. Planf Physiol. 95, 1203-1208. Bandurski, R. S., Cohen, J. D., Solviri, J. P.. and Reinecke, D. M. (1995). Hormone biosynthesis and metabolism. B 1 auxin biosynthesis and metabolism. In “Plant Hormones, Physiology, Biochemistry, and Molecular Biology” (P. I. Davies, ed.), pp. 39-65. Kluwer Academic Pub. Dordrecht, The Netherlands. Bar, T., and Okon, Y.(1992). Induction of indole-3-acetic acid synthesis and possible toxicity of troptophan in Azospirillum brasilense Sp7. Symbiosis 13, 191-198. Bar, T., and Okon, Y. (1993). Tryptophan conversion to indole-3-acetic acid via indole-3-acetamide in Azospirillum brasilense Sp7. Can. J. Microbiol. 39,8 1-86. Barbieri, P., and Galli, E. (1993). Effect of wheat root development of inoculation with an Azospirillum brasilense mutant with altered indole-3-acetic acid production. Res. Microbiol. 144,69-75. Barbieri, P., Bemardi, A,, Galli, E., and Zanetti, G. (1988). Effects of inoculation with different strains of Azospirillum brasilense on wheat root development. In Azospirillum IV Genetics, Physiology, Ecology (W. Klingmiiller, ed.), pp. I8 1-1 88. Springer-Verlag, Berlin, Heidelberg, Germany. Barbieri, P., Zanelli, T., Galli, E., and Zanetti, G. (1986). Wheat inoculation with Azospirillum brasislense Sp6 and some mutants altered in nitrogen fixation and indole-3-acetic acid production. FEMS Microbiol. Lett. 36,87-90. Barea, J. M., and Azcbn-Aguilar, C. (1982). Production of plant growth-regulating substances by the vesicular-arbuscular mycorrhizal fungus Glomus mosseae. Appl. Environ. Micmbiol. 43, 810-813. Barea, J. M., and Brown, M. E. (1974). Effects on plant growth produced by Azotobacterpaspafi related to synthesis of plant growth regulating substances. J. Appl. Bacteriol. 37,583-593. Barea, J. M., Navarro, E., and Montoya, E. (1976). Production of plant growth regulators by rhizosphere phosphate-solubilizing bacteria. J. Appl. Bacteriol. 40, 129-134. Barendse, G. W. M. and Peeters, J. J. M. (1995). Multiple hormonal control in plants. Acta Bot. Necrl. 44,3-17. Barroso, J., Neves, H. C., and Pais, M. S. (1986). Production of indole-3-ethanol and indole-3-acetic acid by the mycorrhizal fungus of Ophrys lurea (Orchidaceae). New Phyfol. 103,745-749. Barry, G. F., Roger, S. G., Fraley, R. T., and Brand, L. (1984). Identification of a cloned cytokinin biosynthetic gene. Proc. Nut. Acad. Sci. LISA 81,4776-4780.
128 MUHAMMAD ARSHAD AND WILLIAM T. FRANKENBERGER,JR. Bashan, H. (1994). Symptom expression and ethylene production in leaf blight of cotton caused by Alrernaria macmspora and Altemaria alternata alone and in combination. Can. J. Botany 72, 1574-1 579. Bauer, P., Ratet, P., Crespi, M. D., Schultze, M., and Kondorosi, A. (1996). Nod factors and cytokinins induce similar cortical cell division, amyloplast deposition and MsEnodlM expression patterns in alfalfa roots. Plan? J. 10,91-105. Beard, R., and Harrison, M. A. (1992). Effect of inoculation on ethylene production in beans. Plant Physiol. 99 (Suppl.), 66. Bearder, J. R., Hedden, P., MacMillan, J., and Wels, C. M. (1973). Gibberellin biosynthesis in the mutant B 1 4 1 a of Gibberellafujikuroi. J. Chem. SOC.Chem. Commun. 717-778. Bearder, J. R., MacMillan, J., Wels, C. M., and Phinney, B. 0. (1975). The metabolism of steviol to 13hydroxylated ent-gibberellanes and ent-kauranes. Phytochem. 14, 1741-1748. Bell,A. A. (1981). Biochemical mechanisms of disease resistance.Annu. Rev. Plant Physiol. 32,21-81. Bennett, R. D., Norman, S. M., and Maier, V. P. (1981). Biosynthesis of abscisic acid from [1,2W21acetate in Cercospora rosicola. Phytochem. 20,2343-2344. Bennett, R. D., Norman, S. M., and Maier, V. P. (1984). Biosynthesis of abscisic acid from farnesol derivatives in Cercospora msicola. Phytochem. 23, 1913-1915. Bethlenfalvay, G. J., Ulrich, J. M., andBrown, M. S. (1985). Plant response to mycorrhizal fungi: Host, endophyte, and soil effects. Soil Sci. SOC.Am. J. 49, 1164-1 168. Beyrle, H. (1995). The role of phytohormones in the function and biology of mycorrhizas. In “Mycorrhiza, Structure, Function, Molecular Biology and Biotechnology” (A. Varma and B. Hock, eds.), pp. 365-390. Springer-Verlag, Heidelberg, Germany. Bhattacharyya, R. M., and Basu, P. S. (1991). Bioproduction of different phytohomones in root nodules of Crotalaria retusa L. and by its symbiont. Ind. J. Exp. Biol. 29,664667. Bhowmick, P. K., and Basu, P. S. (1984). Contents of hormones, and indoleacetic acid metabolism in root nodules of Erythrina indica Lamk., Sesbania grandgora Pers., and Pterocarpus santalinus Linn. Biochem. Physiol. PJanzen. 179,455462. Billington, D. C., Golding, B. T., and Primrose, S. B. (1979). Biosynthesis of ethylene from methionine. Biochem. J. 182,827-836. Binns, A. N., and Thomashow, M. F. (1988). Cell biology ofAgrobacterium infection and transformation of plants. Annu. Rev. Microbiol. 42,575-606. Boller, T. (1991). Ethylene in pathogenesis and disease resistance. In “The Plant Hormone Ethylene” (A. K. Mattoo and J. C. Suttle, eds.), pp. 293-3 14. CRC Press, Boca Raton, FL. Bottini, R., Fulchieri, M., Pearce, D., and Pharis, R. P. (1989). Identification of gibberellins A,, A,, and iso-A, in cultures of Azospirillum lipoferum. Plant Physiol. 90,4547. Brian, P. W. (1957). The effects of some microbial metabolic products on plant growth. In The Biological Action of Growth Substances. Symp. SOC.Exptl. Biol. 11,166-182, Cambridge Univ. Press, Cambridge, UK. Brown, M. E. (1974). Seed and root bacterization. Annu. Rev. Phytoparhol. 12, 181-197. Brown, M. E., and Burlingham, S. K. (1968). Production of plant growth substances by Azotobacter chroococcum. J. Gen. Microbiol. 53, 135-144. Brown, M. E., Jackson, R. M., and Burlingham, S . K. (1968). Effects produced on tomato plants, Lycopersicum esculentum by seed or root treatment with gibberellic acid and indole-3yl-acetic acid. J. Exp. Bot. 19,544552. Bulard, C., Guichardon, B., and Rigaud, J. (1963). Mise en evidence de substances de nature auxinique synthetisees par Rhizobium cultive en presence de tryptophane. Ann. Inst. Pasteur 105,150-157. Cacciari, I., Lippi, D., Pietrosanti, T.,and Pietrosanti, W. (1989). Phytohormone-like substances produced by single and mixed diaztrophic cultures of Azospirillum and Arthrobactel: Plant Soil 115, 151-1 53. Camilleri, C., and Jouanin, L. (1991). The TR-DNA region carrying the auxin synthesis genes of the
PLANT GROWTH REGULATORS IN THE RHIZOSPHERE
129
Agrobacterium rhizogenes agropine-type plasmid pRiA4: Nucleotide sequence analysis and introduction into tobacco plants. Mol. Plant-Microbe Interaction. 4, 155-162. Cardarelli, M., Spano, L., Mariotti, D., Mauro, M. L., Vansluys, M. A., and Costantino, P. (1987). The role of auxin in hairy root induction. Mol. Gen. Genet. 208,457-463. Chalutz, E. (1979). No role for ethylene in the pathogenicity of Penicillium digitatum. Physiol. Plant Pathol. 14,259-262. Chalutz, E., and Lieberman, M. (1978). Inhibition of ethylene production in Penicillium digitaturn. Plant Physiol. 61, 1 11-114. Chalutz, E., Lieberman, M., and Sisler, H. D. (1977). Methionine-induced ethylene production by Penicillium digitaturn. Plant Physiol. 60,402406, Chandramohan. D., and Mahadevan, A. (1968). Indole acetic acid metabolism in soils. Curr. Sci. 4, 112-1 13. Chappell, J., Hahlbrock, K., and Boller, T. (1984). Rapid induction of ethylene biosynthesis in cultured parsley cells by fungal elicitor and its relationship to the induction of phenylalahine ammonialyase. Planta 161,475-480. Charbonneau, G . A., and Newcomb, W. (1985). Growth regulators in developing effective root nodules of the garden pea (Pisum safivum L.). Biochem. Physiol. Pflanzen. 180,667-681. Chattopadhyay, K. K., and Basu, P. S. (1989). Studies on root nodules of leguminous trees: 11. The hioproduction of different phytohormones in root nodules of Samanea saman (Jacq.) Merril and by its rhizobial microsymbiont. Biochem. Physiol. Pjlanzen. 184,387-394. Chernin, L. S., Lobanok, E. V., Fomicheva, V. V., and Katel, N. A. (1984). Crown gall-suppressive IncWR plasmids cause a decrease in auxin production by Agrobacterium turnefaciens. Mol. Gen. Genet. 195, 195-199. Chilton, M. D., Tepfer, D. A., Petit, A., David, C., Casse-Delbart, F., and TempC, J. (1982). Agrobacterium rhizogenes inserts T-DNA into the genomes of the host plant root cells. Nature 295, 432-434. Chou, T. W., and Yang, S. F. (1973). The biogenesis of ethylene in Penicillium digitaturn. Arch. Biochem. Biophys. 157,73-82. Clapperton, M. J., Koshioka, M., and Pharis, R. P. (1985). The effect of infection by a vesicular arbuscular mycorrhizal fungus on the gibberellin content of slender wheat grass. Plant Physiol. 77, 79. Coleman, L. W., and Hodges, C. F. (1986). The effect of methionine on ethylene and I-aminocyclopropane-I-carboxylic acid production by Bipolaris sorokiniana. Phyropathol. 76,85 1-855. Comai, L., and Kosuge, T. (1980). Involvement of plasmid deoxyribonucleic acid in indoleacetic acid synthesis in Pseudomonas savasfanoi. J. Bacieriol. 143,950-957. Comai, L., and Kosuge, T.(1982). Cloning and characterization of iaaM, a virulence determinant of Pseudornonas savastanoi. J . Bacteriol. 149,4046. Comai, L., and Kosuge, T. (1983). Transposable element that causes mutations in a plant pathogenic Pseudomonas sp. J. Bacteriol. 154, 1162-1 167. Comai, L., Surico, G., and Kosuge, T. (1982). Relation of plasmid DNA to indoleacetic acid production in different strains of Pseudomonas syringae pv. savasroni. J. Gen. Microbiol. 128, 2157-2 163. Cooper, J. B., and Long, S. R. (1994). Morphogenetic rescue of Rhizobium melilofi nodulation mutants by trans-zeatin secretion. Plant Cell 6, 2 15-225. Cornforth, I. S. (1975) The persistence of ethylene in aerobic soils. Plant Soil. 42, 85-96. Costacurta, A,, and Vanderleyden, J. (1995). Synthesis of phytohormones by plant-associated bacteria. Critical Rev. Microbiol. 21, 1-18. Costacurta, A., Keijers, V., and Vanderleyden, J. (1994). Molecular cloning and sequence analysis of an Azospirillum brasilense indole-3-pyruvic decarboxylasegene. Mol. Gen. Genet. 243,463472. Costacurta, A,, Prinsen, E., van Onckelen, H., Michiels, K., Vanderleyden, J., and Nuti, M. P. (1992).
130 MUHAMMAD ARSHAD AND WILLIAM T. FRANKENBERGER,JR. IAA synthesis in Azospirillum brasilense Sp6: Analysis of a mutant impaired in IAM-hydrolase. Symbiosis 13, 151-158. Costantino, P., Mauro, M. L., Micheli, G.,Risuleo, G., Hooykaas, P. J. J., and Schilperoort, R. A. (1981). Fingerprinting and sequence homology of plasmids from different virulent strains of Agrobacterium rhizogenes. Plasmid. 5, 170-1 82. Crady, E. E., and Wolf, F. T. (1959). The production of indole acetic acid by Taphrina deformans and Dibotryon morbosum. Physiol. Plant. 12,526-533. Crafts, C . B., and Miller, C. 0. (1974). Detection and identification of cytokinins produced by mycorrhizal fungi. Plant Physiol. 54,586-588. Crespi, M., Messers, E., Caplan, A. B., van Montagu, M., and Desomar, J. (1992). Fasciation induction by the phytopathogen Rhodococcus facians depends upon a linear plasmid encoding cytokinin synthase gene. EMEO Journal 11,795-804. Cross, B. E., Galt, R. H. B., and Norton, K. (1968). The biosynthesis of the gibberellins-U. Tetrahedron. 24,231-237. Crozier, A. (1983). Introduction: The occurrence and structure of gibberellins. In “The Biochemistry and Physiology of Gibberellins” (A. Crozier, ed.), Vol. I , p. 1-18. Praeger, New York. Crozier, A., Arruda. P., Jasmim, J. M., Monteiro, A. M., and Sandberg, G.(1988). Analysis of indole3-acetic acid and related indoles in culture medium from Azospirillum lipoferum and Azospirillum brasilense. Appl. Environ. Microbiol. 54,2833-2837. Crozier, A., Sandberg, G.,Wheeler, C. T., Ernstsen, A,, Atzorn, R., Monteiro, A. M., Jasmim, J. M., and Armda, P. (1987). Phytohormone production by nitrogen fixing bacteria. In “Physiology and Biochemistry of Auxins in Plants” (M. Kutacek, R. Bandurski, and J. Krekule, eds.), pp. 67-75. SPB Academic Publ., The Hague, The Netherlands. Dahm, H., Sitek, J. M., and Strzelczyk, E. (1977). Synthesis of auxins by bacteria isolated from the roots of pine seedlings inoculated with rusty forest soil. Polish J. Soil Sci. 10, 131-137. Dahne, H.-W. (1982). Interaction between vesicular-arbuscular mycorrhizal fungi and plant pathogens. Phytopathol. 72, 1115-1119. Dangar, T. K.,and Basu, P. S. (1984). Seasonalchanges and metabolism of plant hormones in root nodules of Lens sp. Eiol. Plant. 26,253-259. Dangar, T. K., and Basu, P. S. (1987). Studies on plant growth substances, IAA metabolism and nitrogenase activity in root nodules of Phaseolus aureus Roxb. var. mungo. Eiol. Plant 29,350-354. Danneberg, G.,Latus, C., Zimmer, W., Hundeshagen, B., Schneider-Poetsch.H., and Bothe, H. (1992). Influence of vesicular-arbuscularmycorrhiza on phytohormonebalances in maize (Zea mays L.). J. Plant Physiol. 141,33-39. Dasilva, E. J., Henriksson, E., and Henriksson. L. E. (1974). Ethylene production by fungi. Plant Sci. Lett. 2,6346. David, A. B., Bashan, Y., and Okon, Y. (1986). Ethylene production in pepper (Capsicum annuum) leaves infected with Xanthomonas campestris pv. vesicatoria. Physiol. Mol. Plant Pathol. 29, 305-3 16. Davies, P. J. (1995). The plant hormones: their nature, occurrence, and functions. In “Plant Hormones: Physiology, Biochemistry, and Molecular Biology” (P. J. Davies, ed.), 2nd ed., pp. 1-12. Kluwer Academic Publishers. Dordrecht, The Netherlands. Davison, J. (1988). Plant beneficial bacteria. EioiTechnology 6,282-286. De Francesco, R., Zanetti, G.,Barbieri, P., and Galli, E. (1985). Auxin production by Azospirillum brasilense under different cultural conditions. In “Azospirillum m:Genetics, Physiology, Ecology” (W. Klingmuller, ed.), pp. 109-1 15. Springer-Verlag, Berlin, Heidelberg, Germany. deBont, J. A. M. (1976). Oxidation of ethylene by soil bacteria. Anronie van Leeuwenhoek. 42,59-71. Dehio, C., and delruijn, F. J. (1992). The early nodulin gene SrEnod2 from Sesbania rostrata is inducible by cytokinin. Plant J. 2, 117-128. DeVries, H. E., Mudge, K. W., and Lardner, J. P. (1987). Ethylene production by several ectomycor-
PLANT GROWTH REGULATORS IN THE RHIZOSPHERE
131
rhizal fungi and effects on host root morphology. In “Mycorrhizae in the Next Decade” (E. M. Sylvia, L. L. Hung, and J. H. Graham, eds.), p. 245. Inst. of Food and Agric. Sci., Univ. of Florida, Gainesville. Dickinson, J. R., Forsyth, C., and Van Staden, J . (1986). The role of adenine in the synthesis of cytokinins in tomato plants and in cell free root extracts. Plant Growth Regul. 4,325-334. Dixon, R. K. (1989). Cytokinin activity in Citrus jambhiri seedlings colonized by mycorrhizal fungi. Agric. Ecosystems Environ. 29, 103-106. Dixon, R. K., Garrett, H. E., and Cox, G. S. (1988a). Cytokinin activity in Citrus jambhiri Lush. seedlings colonized by vesicular-arbuscular mycorrhizal fungi. Trees. 2 , 3 9 4 . Dixon, R. K., Garrett, H. E., and Cox, G. S. (1988b). Cytokinins in the root pressure exudate of Citrus jambhiri Lush. colonized by vesicular-arbuscular mycorrhizae. Tree Physiol. 4,9-18. Dobert, R. C., Rood, S. B., and Blevins, D. G. (1992a). Gibberellins and the Rhizobium-legume symbiosis. I. Endogenous GAS of lima bean (Phaseolus lunarus L.) stems and nodules. Plant Physi01. 98,221-224. Dobert, R. C., Rood, S. B., and Blevins, D. G. (1992b). Rhizobial-induced increase in internode length and identification of endogenous GAS of cowpea (Vigna unguiculutu [L.] Walp) stems and nodules. J. Pianr Growth Regul. 11,155-164. Drevon, J. J. (1983). Various organisms that fix nitrogen. In “Technical Handbook on Symbiotic Nitrogen Fixation, LegumeRhizobium,” FAO, United Nations. I Biol. 2, 1 4 . Driige, U., and Schonbeck, F. (1992). Effect of vesicular-arbuscular mycorrhizal infection on transpiration, photosynthesis and growth of flax (Cinus usitatissimum L.) in relation to cytokinin levels. J. Plant Physiol. 141,4048. Dubeikovsky, A. N., Mordukhova, E. A., Kochetkov, V. V., Polikaqova, F. Y., and Boronin, A. M. (1993). Growth promotion of blackcurrant softwood cuttings by recombinant strain Pseudomonas Juorescens BSP53a synthesizing an increased amount of indole-3-acetic acid. Soil Biol.Biochem. 25,127771 28 1 . Dullaart, J. (1970). The bioproduction of indole-3-acetic acid and related compounds in root nodules and roots of Lupinus luteus L. and by its rhizobial symbiont. Acta Bot. Neerl. 19,573-61 8 . Dullaart, J., and Duba, L. I. (1970). Presence of gibberellin-like substances and their possible role in auxin bioproduction in root nodules and roots of Lupinus luteus L. Actu Bot. Neerl. 19,877-883. Dutra, P. V., Abad, M., Alrnela, V., and Agusti. (1996). Auxin interaction with the vesicular-arbuscular mycorrhizal fungus Glomus intraradices Schenck & Smith improves vegetative growth of two citrus rootstocks. Scientia Hort. 66,77-83. Dutta, S., and Biggs, R. H. (1991). Regulation of ethylene biosynthesis in citrus leaves infected with Xanthomonas compestris pv. citri. Physiol. Plant. 82,225-230. Ecker, J. R., and Davis, R. W. (1987). Plant defense genes are regulated by ethylene. Proc. Nut. Acud. Sci. USA 84,5202-5206. Edriss, M. H., Davis, R. M., and Burger, D. W. (1984). Influence of mycorrhizal fungi on cytokinin production in sour orange. J. Am. SOC.Horf. Sci. 109,587-590. Egebo, L. A., Nielsen, S. V. S., and Jochimsen, B. U. (1991). Oxygen-dependent catabolism of indole3-acetic acid in Bradyrhizobium japonicum. J. Bacteriol. 173,48974901. Einset, T. W., Greene, E. M., Skoog, F., Doyle, E., and Hanson, R. S. (1977). Cytokinin production by strains of Agrobacterium tumefaciens. Plant Physiol59 (Suppl.), 109. Ek, M., Ljungquist, P. O., and Stenstrom, E. (1983). Indole-3-acetic acid production by mycorrhizal fungi determined by gas chromatography-mass spectrometry. New Phytol. 94,401407. El-Abyad, M., El-Sayed, M. A,, El-Shanshoury, A. R., and Farid, M. (1994). Optimization of culture conditions for indole-3-pyruvic acid production by Strepfomyces griseoJavus. Can. J. Microbiol. 40,754-760. Elmerich, C. (1984). Molecular biology and ecology of diazotrophs associated with non-leguminous plants. Bioh’echnology 2,967-978.
132 MUHAMMAD ARSHAD AND WILLIAM T. FRANKENBERGER,JR. Elstner, E. F. (1983). Hormones and metabolic regulation in disease. In “Biochemical Plant Pathology” (J. A. Callow, ed.), pp. 415-434. Wiley & Sons, New York. Elwan, S. H., and El-Naggar, M. R. (1972). Stimulatory effects of Azotobacter chmococcum exudates in culture and soil on its nitrogen fixation. Z. Mikrobiol. 12, 107-1 14. Eplee, R. E. (1975). Ethylene, a witchweed seed germination stimulant. Weed Sci. 23,433436. Emstsen,A., Sandberg, G., Crozier, A., and Wheeler, C. T. (1987). Endogenous indoles and the biosynthesis and metabolism of indole-3-acetic acid in cultures of Rhizobium phaseoli. Plunta 171, 422-428. Evans, R., and Hanson, J. R. (1975). Studies in terpenoid biosynthesis. XIII. The biosynthetic relationship of the gibberellins in Gibberellafujikuroi. J. Chem. SOC.Perkin. I, 663-666. Evensen, K . B., and Blevins, D. G. (1981). Differences in endogenous levels of gibberellin-like substances in nodules of Phaseolus lunatus L. plants inoculated with two Rhizobium strains. Plant Physiol. 68, 195-198. Fall, R. R., and West, C. A. (1971). Purification and properties of kaurene synthetase from Fusarium moniliforme. J. Biol. Chem. 246,6913-6928. Fallik, E., Okon, Y., Epstein, E., Goldman, A., and Fischer, M. (1989). Identification and quantification of IAA and IBA in Azospirillum brusilense-inoculated maize roots. Soil Biol. Biochem. 21, 147-153. Fearn, J. C., and LaRue, T. A. (1991). Ethylene inhibitors restore nodulation to sym 5 mutants of Pisum sativum L. cv Sparkle. Plant Physiol. 96,239-244. Fernhndez-Martin, R., Reyes, F., Domenech, C. E., Cabrera, E., Bramley, P. M., Barrero, A. F., Avalas, J., and Cerdi-Olmedo, E. (1995). Gibberellin biosynthesis in gib mutants of Gibberella fujikuroi. J. Biol. Chem. 270, 14970-14974. Fett, W. F., Osman, S. F., and Dunn, M. F. (1987). Auxin production by plant-pathogenic pseudomonads and xanthomonads. Appl. Environ. Microbiol. 53, 1839-1 845. Follin, A,, Inzt, D., Budar, F., Genetello, C., Van Montagu, M., and Schell, J. (1985). Genetic evidence that the tryptophan-2-monooxygenase gene of Pseudomonas suvastunoi is functionally equivalent to one of the T-DNA genes involved in plant turnour formation by Agrobucterium tumefaciens. Mol. Gen. Genet. 201, 178-185. Frankenberger, W. T., Jr., and Arshad, M. (1991a). Yield response of watermelon and muskmelon to Ltryptophan applied to soil. Hort. Sci. 26,35-37. Frankenberger, W. T., Jr., and Arshad, M. (1991b). Yield response of Cupricum annuum to the auxin precursor, L-tryptophan applied to soil. PGRSA Quarterly 19,23 1-240. Frankenberger, W. T., Jr., and Arshad, M, (1995). “Phytohormones in Soils: Microbial Production and Function,” p. 503. Marcel Dekker, New York. Frankenberger, W. T., Jr., and Brunner, W. (1983). Method of detection of auxin-indole-3-acetic acid in soils by high performance liquid chromatography. Soil Sci. SOC.47,237-241. Frankenberger, W. T., Jr., and Martens, D. A. (1995). Use of plant hormones for crop improvement. US.Patent and Trade Ofice. Patent Pending. Frankenberger, W. T.. Jr., and Phelan, P. J . (1985a). Ethylene biosynthesis in soil. I. Method of assay in conversion of I-aminocyclopropane-I-carboxylic acid to ethylene. Soil Sci. Soc. Am. J. 49, 1416-1422. Frankenberger, W. T., Jr., and Phelan, P. J. (1985b). Ethylene biosynthesis in soil. 11. Kinetics and thermodynamics in the conversion of I-aminocyclopropane-1-carboxylicacid to ethylene. Soil Sci. SOC.Am. J. 49, 1422-1426. Frankenberger, W. T., Jr., and Poth, M. (1987a). Biosynthesis of indole-3-acetic acid by the pine ectomycorrhizal fungus Pisolithus tinctorius. Appl. Environ. Microbiol. 53,2908-291 3. Frankenberger, W. T., Jr., andPoth, M. (1987b). Determination of substituted indole derivatives by ion suppression-reverse phase high performance chromatography. Anal. Biochem. 165,300-308. Frankenberger, W. T., Jr., and Poth, M. (1988). L-Tryptophan transaminase of a bacterium isolated from the rhizosphere of Festuca octofloru (Graminae). Soil Biol. Biochem. 20,299-304.
PLANT GROWTH REGULATORS IN THE RHIZOSPHERE
1 33
Frankenberger, W. T., Jr., Chang, A. C., and Arsbad, M. (1990). Response of Raphanus sativus to the auxin precursor, L-tryptophan applied to soil. Plant Soil 129,235-241. Fraser, R. S. S. (1991). ABA and plant responses to pathogens. In “Abscisic Acid, Physiology, and Biochemistry” (W. J. Davies and H. G. Jones, eds.), pp. 189-199. Bios Scientific, Oxford. UK. Freytag, A. H., Wendt, C. W., and Lira, E. P. (1972). Effects of soil-injected ethylene on yields of cotton and sorghum. Agron. J. 64,524-526. Fuentes-Ramirez, L. E., Jimenez-Salgado, T., Abarco-Ocampo, I. R., and Caballero-Mellado, J. (1 993). Acetobacrer diazotrophicus; an indoleacetic acid producing bacterium isolated from sugarcane cultivars of Mexico. Plant Soil 154, 145-150. Fukuda, H., and Ogawa. T. (1991). Microbial ethylene production. In “The Plant Hormone Ethylene” (A. K. Mattoo and J. C. Suttle, eds.), pp. 279-292. CRC Press, Boca Raton, FL. Fukuda, H., Fujii, T., and Ogawa, T. (1986). Preparation of a cell-free ethylene-forming system from Penicillium digitatum. Agric. Biol. Chem. 50, 977-98 1. Fukuda, H., Fujii, T., and Ogawa, T. (1988). Production of ethylene by a growth-suppressed mutant of Penicillium digitatum. Biotech. Bioeng. 31,620-623. Fukuda, H., Kitajima, H., Fujii, T., Tazake, M., and Ogawa, T. (l989b). Purification and some properties of a novel ethylene-forming enzyme produced by Penicillium digitaturn. FEMS Microbiol. Lett. 59, 1-6. Fukuda, H., Ogawa, T., Ishihara, K., Fujii, T., Nagahama, K., Osmata, T., Inoue, Y., Tanase, S.. and Morino, Y.(1992). Molecular cloning in Escherichi coli, expression and nucleotide sequence of the gene for ethylene-forming enzyme of Pseudomonas syringae pv. phaseolicola PK2. Biochem. Biophys. Res. Commun. 188,826-832. Fukuda, H., Takahashi, M., Fujii, T , and Ogawa, T. (1989a). Ethylene production from L-methionine by Cryptococcus albidus. J. Ferment. Bioeng. 67, 173-175. Fukuda, H., Takahashi, M., Fujii, T., Tazaki, M., and Ogawa, T. (1989~).An NADH:Fe(III) EDTA oxidoreductase from Cryptococcus albidusr An enzyme involved in C,H, production in vivo? FEMS Microbiol. Lett. 60, 107-1 12. Fulchieri, M., Lucangeli, C., and Bottino, R. (1993). Inoculation with Azospirillum lipoferum affects growth and gibberellin status of corn seedling roots. Plant Cell Physiol. 34, 1305-1309. Gafni, Y., Icht, M., and Rubinfeld, B.-Z. (1995). Stimulation of Agrobacterium tumefaciens virulence with indole-3-acetic acid. Lett. Appl. Microbiol. 20,98-101. Gaudin, V., Vrain, T., and Jouanin, L. (1994). Bacterial genes modifying hormonal balances in plants. Plant Physiol. Biochem. 32,ll-29. Gay, G., and Debaud, J. C. (1987). Genetic study on indole-3-acetic acid production by ectomycorrhizal Hebeloma species: inter- and intraspecific variability in homo- and dikaryotic mycelia. Appl. Microbiol. Biotechnol. 26, 141-146. Gay, G., Normand, L., Marmeisse, R., Sotta, B., and Debaud, J. C. (1994). Auxin overproducer mutants of Hebeloma cylindrosporum Romagnesi have increased mycorrhizal activity. New Phytol. 128,645-657. Gay, G., Rouillon, R., Bemillon, J., and Favre-Bonvin, J. (1989). IAA biosynthesis by the ectomycorrhizal fungus Hebeloma hiemale as affected by different precursors. Can. J. Bot. 67,2235-2239. Glick, B. R. (1995). The enhancement of plant growth by free-living bacteria. Can. J. Microbiol. 41, 109-1 17. Glick, B. R., Jacobson, C. B., Schwarze, M. M. K., and Pastemak, J. J. (1994a). Does the enzyme 1aminocyclopropane- I -carboxylate deaminase play a role in plant growth-promotion by Pseudomonas putida GR12-2? In “Improving Plant Productivity with Rhizosphere Bacteria” (M. H. Ryder, P. M. Stephens, and G. D. Bowen, eds.), pp. 156152. Commonwealth Scientific and Industrial Research Organization, Adelaide, Australia. Glick, B. R., Jacobson, C. B., Schwarze, M. M. K., andpasternak, J . J. (1994b). 1-AminocyclopropaneI-carboxylate deaminase mutants of the plant growth-promoting rhizobacterium Pseudomonas putida GR12-2 do not stimulate canola root elongation. Can. J. Microbiol. 40, 91 1-915.
134 MUHAMMAD ARSHAD AND WILLIAM T. FRANKENBERGER,JR. Gogala, N. (1970).Einfluss der natiirlichen Cytokinine von Pinus sylvestris L. und andere Wuchsstoffe auf das Mycelwachsturn von Boletus edulis var. pinicolus. Vitf. Oesterr: Bot. Z. 188,321-333. Gogala, N. (1971). Growth substances in mycorrhiza of the fungus Boletus pinicola vitt. and the pine tree Pinus-sylvestris L. Razprave. 14, 123-202. Goicoechea, N., Antolin, M. C., Stmad, M., and Shchez-Diaz, M. (1996). Root cytokinins, acid phosphatase and nodule activity in drought-stressed mycorrhizal or nitrogen-fixing alfalfa plants. J. Exper: Botany 47,683-686. Goicoechea, N., Doltzal, K., Antolin, M. C., Stmad, M., and Sfinchez-Diaz, M. (1995). Influence of mycorrhizae and Rhizobium on cytokinin content in drought-stressed alfalfa. J. Exper: Botany 46, 1543-1 549. Gonzalez-Lopez, J., Salmeron, V., Martinez-Toledo, M. V., Ballesteros, F., and Ramos-Cormenzana, A. (1986). Production of auxins, gibberellins and cytokinins by Azotobacter vinelandii ATCC 12837 in chemically-defined media and dialized soil media. Soil Eiol. Biochem. 18, 119-120. Goodlass, G.. and Smith, K. A. (1978a).Effect of pH, organic matter content and nitrate on the evolution of ethylene from soils. Soil Biol. Biochem. 10,193-199. Goto, M., and Hyodo, H. (1985).Role of extracellular polysaccharides of Xanthomonas campestris pv. cirri in the early stage of infection. Ann. Phytopathol. SOC.Japan. 51,22-31. Goto, M., and Hyodo, H. (1987). Ethylene production by cell-free extract of the Kudzu strain of Pseudomonas syringae pv. phaseolicola. Plant Cell Physiol. 28,405414. Goto, M., Ishida, Y,Takikawa, Y., and Hyodo, H. (1985). Ethylene production by the Kudzu strain of Pseudomonas syringae pv. phaseolicola causing halo blight in Pueraria lohata (Wild) Ohwi. Plant Cell Physiol. 26, 141-150. Graham, J. H., and Linderman, R. G. (1980). Ethylene production by ectomycorrhizal fungi, Fusarium oxysporum f. sp. pini, and by aseptically synthesized ectomycorrhizae and Fusarium-infected Douglas fir roots. Can.J. Micmbiol. 26, 1340-1347. Grappelli, A,, and Rossi, W. (1981).The effect of phytohormones produced by Arrhrobacter sp. on the phosphatase activity in plant roots. Folia Micmbiol. 26, 137-141. Gray, J., Gelvin, S. B., Meilan, R., and Moms, R. 0. (1996).Transfer RNA is the source of extracellular isopentenyladenine in a Ti-plasmidless strain of Agrobacterium tunefaciens. Plant Physiol. 110,431438. Gray, J., Wang, J., and Gelvin, S. B. (1992).Mutation of the miaA gene Agrobacterium tumefaciens results in reduced vir gene expression. J. Bacteriol. 174, 1086-1098. Grayston, S. J., Stephens, J. H., and Nelson, L. M. (1990).Field and greenhouse studies on growth promotion of spring wheat inoculated with co-existent rhizobacteria. In “Abstract, 2nd Intl. Workshop on Plant Growth-Promoting Rhizobacteria,” 4-7. International Union of Biological Sciences. Greene, M. (1980).Cytokinin production by microorganisms. Bot. Rev. 46,25-74. Gruen, H. E. (1959).Auxins and fungi. Annu. Rev. Plant Physiol. 1 0 , 4 0 5 4 0 . Guinel, F. C., and LaRue, T. A. (1992). Ethylene inhibitors partly restore nodulation to pea mutants El07 (brz). Plant Physiol. 99,515-518. Guinel, F. C., and LaRue, T. A. (1991). Light microscopy study of nodule initiation in Pisum sativum L. CV Sparkle and in its low-nodulating mutant E2 (sym 5). Plant Physiol. 97, 1206-121 1. Gunze, C. M. B., and Hennessy, C. M. R. (1980).Effect of host-applied auxin on development of endomycomhizae in cowpeas. Trans. Br: Mycol. SOC.74,247-25 1. Cutierrez-Maiiero, F. J. G., Acero, N., Lucas, J. A., and Probanza, A. (1996). The influence of native rhizobacteria on European alder (Alnus glutinosa (L.)Gaertn.) growth. Plant and Soil 182,67-74. Haahtela, K., Ronkko, R., R. Laakso, R., Williams, P. W., and Korhones, T.K.(1990a).Root-associated Enterobacter and Klebsiella in Poapratensis: Characterization of an iron-scavenging system and a substance stimulating root hair production. Mol. Plant Microbe Interact. 3,358-365. Hahm, D. H., Kwak, M. Y.,Bae, M., and Rhee, J. S. (1992).Effect of dissolved oxygen tension on microbial ethylene production in continuous culture. Biosci. Biotech. Biochem. 56, 1146-1 147. Hall, J. A,, Peirson, D., Ghosh, S., and B. R. Glick. (1996).Root elongation in various agronomic crops
PLANT GROWTH REGULATORS I NTHE RHIZOSPHERE
13 5
by the plant growth promoting rhizobacterium Pseudomonns putidu GR 12-2. Israel J. ffanr Sci. 44,37-42. Hamence, J. H. (1946). The determination of auxins in soils, including a note on synthetic growth substances. Analyst 71, 111-1 16. Hammerschlag, F. A., Owens, L. D., and Smigocki, A. C. (1989). Agrobacferium-mediated transformation of peach cells derived from mature plants that were propagated in vitro. J. Am. SOC.Horf. Sci. 114,508-5 10. Hanley, K. M., and Greene, D. W. (1987). Gibberellin-like compounds from two ectomycorrhizal fungi and the GA, response on Scotch pine seedlings. HorrScience 22,591-594. Hartmann, A., Singh, M., and Klingmuller, W. (1983). Isolation and characterization of Azospirillum mutants excreting high amounts of indoleacetic acid. Can. J. Microbiol. 29,916923. Hartmann, T., and Glombitza, K.-W. (1967). Der Truptophanabbau bei Rhizobium leguminosarum. Arch. Mikrobiol. 56, 1-8. Hartung, W., and Davies, W. J. (1991). Drought-induced changes in physiology and ABA. In “Abscisic Acid, Physiology, and Biochemistry” (W. J. Davies and H. G. Jones, eds.), pp. 63-79. Bios Scientific, Oxford, UK. Hartung, W., Heilmeier, H., H. Warfinger, H., Kettemann, I., and Schulze, E.-D. (1990). Ionic content and abscisic acid relationships of Anasratica hierochuntica L. under arid conditions. Israel J. Bof. 39,373-382. Harvey, R. G., and Linscott, J. J. (1978). Ethylene production in soil containing quackgrass rhizomes and other plant materials. Soil Sci. SOC.Am. J. 42,721-724. Haselwandter, K. (1973). Indol-(wuchs)-stoffbeziehungen zwischen Mikroorganismen aus der Rhizosphi4re und einem Mykorrhizapilz von Pinus cembra L. Arch. Mikrobiol. 94,259-268. Hayman, D. S. (1983). The physiology of vesicular-arbuscular endomycorrhizal symbiosis. Can. J. Bog. 61,94463. Hedden, P., MacMillan, J., and Phinney, B. 0. (1974). Fungal products. Part XII. Gibberellin A,,-aldehyde, an intermediate in gibberellin biosynthesis in Gibberellafujikuroi. J. Chem. SOC.Perkin. I, 587-592. Heinemeyer, W., Buchmann, I., Tonge, D. W., Windass, J. D., Moerbe, J., Weiler, E. W., Botz, T., and Schroder, J. ( 1987). Two Agrobacrerium tumefaciens genes for cytokinin biosynthesis: T. plasmid-coded isopentenyltransferase adapted for function in prokaryotic or eukaryotic cells. Mol. Gen. Genet. 210,257-262. Hirai, N., Okamoto, M., and Koshimizu, K. (1986). The 1’,4’-rrans-diol of abscisic acid, a possible precursor of abscisic acid in Bofryfiscinerea. Phyrochem. 25, 1865-1868. Hirata, S. (1958). Studies on the phytohormone in the malformed portion of the diseased plants. 5 . On the auxin formation by Proromyces fungi. Ann. Phyfopathol. SOC.Japan 23,145-149. Hirsch, A. M., and Fang, Y. (1994). Plant hormones and nodulation: What’s the connection?Planr Molecular Biol. 26,5-9. Hirsch, A. M., Bhuvaneswari, T. V., Torrey, J. G., and Bisseling, T. (1989). Early nodulin genes are induced in alfalfa root outgrowths elicited by auxin transport inhibitors. Proc. Nufl.Acad. Sci. USA 86,1244-1248. Ho, I. (1987a). Comparison of eight Pisolifhusfunctorius isolates for growth rate, enzyme activity and phytohormone production. Can. J. For Res. 17,31-35. Ho, I. (l987b). Enzyme activity and phytohormone production of a mycorrhizal fungus, Laccana laccata. Can. J. For: Res. 17,855-858. Ho, I., and Trappe, J. M. (1992). Enzymatic activity and phytohormone production of Trichloma magnivelare (Peck) Preadreed, The “American Matsutake.” Indian J. Mycol. PI. Parhol. 22, 6 6 7 2 . Holl, F. B., Chanway. C. P., Turkington, R., and Radley, R. A. (1988). Response of crested wheatgrass (Agropyron crisfarurn L.), perennial ryegrass (lolium perenne) and white clover (TrifoZium repens L.) to inoculation with Bacillus polumyxa. Soil Eiol. Biochem. 20, 19-24. Hooykaas, P.J. J., Hofker, M., den Dulk-Ras, H., and Schilperoort, R. A. (1984). A comparison of vir-
13 6 MUWMMAJI A R S W AND WILLIAM T. FRANKENBERGER,JR. ulence determinants in an octopine Ti plasmid, a nopaline Ti plasmid, and an Ri Plasmid by complementation analysis of Agrobacteriurn tumefaciens mutants. Plusmid 11, 195-205. Horak, E. (1964). Die bildung von IES-derivaten durch ectotrophe mykorrhizapilze (Phlegmacium spp.) von Picea abies Karsten). Phytopathol. Z. 51,491-515. Horemans, S., and Vlassak, K. (1985). Production of indole-3-acetic acid by Azospirillurn brasilense. In “Azospirillum111, Genetics, Physiology, Ecology” (W. Klingmiiller, ed.), pp. 98-108. SpringerVerlag, Berlin, Germany. Horemans, S., Koninck, K. D., Neuray, J., Hermans, R., and Vlassak, K. (1986). Production of plant growth substances by Azospirillum sp. and other rhizobacteria. Symbiosis 2,341-346. Horgan, R., Neill, S. J., Walton, D. C., and Griffin, D. H. (1983). Biosynthesis of abscisic acid. Biochem. SOC. Trans. 11,553-557. Hoflich, G . (1992). Interrelationships between phytoeffective Pseudomonas bacteria and the growth of crops. Zentralbl. Mikrobiol. 147, 182-191. Hoflich, G., and Weise, I. (1992). Improvement of efficiency of an Rhizobium-inoculation on pea through a combined inoculation of Rhizobium leguminosarum biovar. viceae with R. leguminosarum biovar. trifolii. Zentralbl. Mikrobiol. 147,378-387. Hua, S.-S., Brandon, D. L., Fuller, G., Corse, J. W., and Layton, L. L. (1980). Cytokinin synthesis in Rhizobium spp. In Abst,: Annu. Meet. Am. Soc. Microbiol. SO, 127. Hubbell, D. H., Tien, T. M., Gaskins, M. H., and Lee, J. (1979). Physiological interactions in the Azospirillum-grass root association. In “Associative N,-Fixation” (P. B. Vose and A. P. Ruschel, eds.), pp. 1 4 . CRC Press, Boca Raton, FL. Huffman, G. A,, White, F. E, Gordon, M. I?, and Nester, E. W. (1984). Hairy root inducing plasmid: physical map and homology to tumour-inducing plasmids. J. Bacteriol. 157,269-276. Hunt, P. G., Matheny, T.A., Campbell, R. B., and Parsons, J. E. (1982). Ethylene accumulation in southeastern coastal plain soils: Soil characteristics and oxidative-reductive involvement. Commun. Soil Sci. Plant Anal. 13,267-278. Hunter, W., and Kuykendall, L. D. (1990). Enhanced nodulation and nitrogen fixation by a revertant of a nodule-defective Bradyrhizobium japonicum tryptophan auxotroph. Appl. Environ. Microbiol. 56,2399-2403. Hunter, W. J. (1989) Indole-3-acetic acid production by bacteroids from soybean root nodules. Physi01. Plant. 76,31-36. Hunter, W. J. (1992). Production of the plant hormone ethylene by soybean roots and nodules. In “Abstract, 92nd Mtg. Am. Soc.Microbiol,” p. 307. American Society of Microbiology,Washington, DC. Hunter, W. J. (1994). Increased nodulation of soybean by a strain of Bradyrhizobium with altered tryptophan metabolism. Lett. Appl. Microbiol. 18,34@342. Hussain, A,, Arshad, M., Hussain, A., and Hussain, F. (1987). Response to maize (Zea mays) and Azotobacter inoculation under fertilized and unfertilized conditions. Biol. Fertil. Soils 4,73-77. Hutzinger, O., and Kosuge, T. ( 1 968a). Microbial synthesis and degradation of indole-3-acetic acid. HI. The isolation and characterisation of indole-3-acetyl-I-lysine. Biochemisty 7,601405. Hutzinger, O., and Kosuge, T.(1968b). 3-lndole-acetyl e-1-lysine, a new conjugate of 3-indoleacetic acid produced by Pseudoznonas savastonoi. In “Biochemistry and Physiology of plant growth substances” (EWightman and M. Setterfield, eds.), pp. 183-194. Runge Press, Ottawa, Canada. Iacobellis, N. S., Sisto, A., Surico, G., Evidente, A., and DiMaio, E. (1994). Pathogenicity of Pseudomonas syringae subsp. savastanoi mutants defective in phytohormone production. J. Phytopathol. 140,238-248. Ichimura, M., Oritani, T., and Yamashita, K. (1983). The metabolism of (2Z,4E)-a-ionylideneacetic acid in Cercospora cruenta, a fungus producing (+)-abscisic acid. Agric. Biol. Chem. 47, 1895-1900. Ince, J. E., and Knowles, C. J. (1986). Ethylene formation by cell-free extracts of Escherichia coli. Arch. Microbiol. 146, 151-158.
PLANT GROWTH REGULATORS IN THE RHIZOSPHERE
137
Ince, J. E., and Knowles, C. J. (1985). Ethylene formation by cultures of Escherichia coli. Arch. Microbiol. 141,209-213. InzC, D., Follin, A., van Lijsebettens, M., Simoens, C., Genetello, C., Van Montagu, M., and Schell, J. (1984). Genetic analysis of the individual T-DNA genes of Agrobacterium tumefaciens: further evidence that two genes are involved in indole-3-acetic acid synthesis. Mol. Gen. Genet. 194, 265-274. Ishikawa, K., Kamada, H., Yamaguchi, I., Takahashi, N., and Harada, H. (1988). Morphology and hormone levels of tobacco and carrot tissues transformed by Agrobacterium tumefaciens. I. Auxin and cytokinin contents of cultured tissues transformed with wild-type and mutant Ti plasmid. Plant Cell Physiol. 29,461466. Jackson, M. B., and Campbell, D. J. (1975). Movement of ethylene from roots to shoots, a factor in the responses of tomato plants to waterlogged conditions. New Phytol. 74,397406. Jacobson, C. B., Pasternak, J. J., and Glick, B. R. (1994). Partial purification and characterization of 1aminocyclopropane- I-carboxylate deaminase from the plant growth promoting rhizobacterium Pseudomonasputida GR12-2. Can. J. Microbiol. 40, 1019-1025. lain, D. K., and Patriquin, D. G. (1985). Characterization of a substance produced by Azospirillurn which causes branching of wheat root hairs. Can. J. Microbiol. 31,206210. Janzen, R. A., Rood, S. B., Dormaar, J. F., and McGill, W. B. (1992). Azospirillum brasilense produces gibberellin in pure culture and on chemically-defined medium in co-culture on straw. Soil Biol. Biochem. 24, 1061-1064. Jeske, G. (1976). Role of roots and leaves in cytokinin synthesis. P,: Kern. Kauk. Lesn. Poznam Tow. Przyi Nauk. 41,133-156. Kaiss-Chapman, R. W., and Morris, R. 0. (1977). Cytokinin production, plasmid status and virulence in Agrobacterium tumefaciens. Plant Physiol. 59 (suppl.), 109. Kampert, M., and Strzelczyk, E. (1975). Synthesis of auxins by fungi isolated from the roots of pine seedlings (Pinus silvestris L.) and from soil. Acta Microbiol. folonica 7,223-230. Kampert, M., and Strzelczyk, E. (1978). Production of cytokinins by mycorrhizal fungi of pine (Pinus silvestris L.). Bull. Acad. Pol. Sci. Ser: Sci. Biol. II 26,499-503. Kampert, M., Strzelczyk, E., and Pokojska, A. (1975). Production of gibberellin-like substances by bacteria and fungi isolated from the roots of pine seedlings (Pinus silvestris L.). Microbiol. Polonica. Acta. 7, 157-166. Kaneshiro, R., Slodki, M. E., and Plattner, R. D. (1983). Tryptophan catabolism of indolepyruvic and indoleacetic acids by Rhizobium japonicum L-259 mutants. Cur,: Microbiol. 8,301-306. Kaneshiro, T., and Kwolek, W. F. (1985). Stimulated nodulation of soybeans by Rhizobium japonicum mutant (B- 14075) that catabolizes the conversion of tryptophan to indole-3yl-acetic acid. Plant Sci. 42, 141-146. Kaneshiro, T., and Nicholson, J. J. (1989). Tryptophan catabolism by tan variants isolated from enriched cultures of bradyrhizobia. Curt Microbiol. 18,57. Katznelson, H., and Cole, S. E. (1965). Production of gibberellin-like substances by bacteria and actinomycetes. Can. J. Microbiol. 11,733-741. Katzy, E. I., Isipenco, A. D., Egorenkov, D. A., Zhuravleva, E. A,, Panasenko, V. I., and Ignatov, V. V. (1 990). Involvement of Azospirillum brasilense plasmid DNA in the production of indole-3-acetic acid. FEMS Left. 72, I . Kawaguchi, M., Fujoika, S., Sakurai, A., Yamaki, S. T., and Syono, K. (1993). Presence of a pathway for the biosynthesis of auxin via indole-3-acetamide in trifoliata orange. Plant Cell Physiol. 34, 121-128. Kawaguchi, M., Sekine, M., and Syono, K. (1990). Isolation of Rhizobium legminosarum biovar. viciae variants with indole-3-acetamide hydrolase activity. Plant Cell Physiol. 3 1 , 4 9 4 5 5 . Kawaide, H., and Sassa, T. (1993). Accumulation of gibberellin A, and the metabolism of gibberellin Ag to gibberellin A, in a Phaeosphaeria sp. L487 culture. Biosci. Biotech. Biochem. 57, 1403-1405.
13 8 MUHAMMAD ARSHAD AND WILLIAM T. FRANKENBERGER,JR. Kawaide, H., Sassa, T., and Kamiya, Y. (1995). Plant-like biosynthesis of gibberellin A, in the fungus Phaeosphaeria sp. L487. Phytochem. 39,305-3 10. Kettner, J., and Dorflling, K. (1995). Biosynthesis and metabolism of abscisic acid in tomato leaves infected with Bottytis cinerea. Planta 196,627-634. Kitagawa. Y., Yamamoto, H., and Oritani, T. (1995). Biosynthesis of abscisic acid in the fungus Cercospora cruenra: Stimulation of biosynthesis by water stress and isolation of a transgenic mutant with reduced biosynthetic capacity. Plant Cell Physiol. 36,557-564. Kittell, B. L., Helinski, D. R., and Ditta, G.S. (1989). Aromatic aminotransferase activity and indoleacetic acid production in Rhizobium melilori. J. Bacteriol. 171,5458-5466. Klee, H., Montoya, A,, Horodyski, F., Lichtenstein, C., Garfinkel, D., Fuller, S., Flores, C., Peschon, J., Nester, E., and Gordon, M. (1984). Nucleotide sequence of the tms genes of the pTiA6NC octopine Ti plasmid: ’ h o gene products involved in plant tumorigenesis. Proc. Natl. Acad. Sci. USA 81, 1728-1732. Kloepper, J. W., Lifshitz, R., and Zablotowicz, R. M. (1989). Free-living bacterial inocula for enhancing crop productivity. Trends Biotechnol. 7 , 3 9 4 . Kloepper, J. W., Scher, F. M., Laliberte, M., and Tipping, B. (1986). Emergence-promoting rhizobacteria: Description and implications for agriculture. In “Iron, Siderophores, and Plant Disease” (T. R. Swinburne, ed.), pp. 155-164. Plenum, New York. Kobayashi, M., Izui, H., Nagasawa, T., and Yamada, H. (1993). Nitralase in biosynthesis of the plant hormone indole-3-acetic acid from indole-3-acetonitrile: Cloning of the Alcaligenes gene and sitedirected mutagenesis of cystein residues. Proc. Natl. Acad. Sci. USA 90,247-25 1. Kobayashi, M., Suzuki, T., Fujita, T.,Masuda, M., and Shimizu, S. (1995). Occurrence of enzymes involved in biosynthesis of indole-3-acetic acid from indole-3-acetonitrile in plant-associated bacteria of Agrobacterium and Rhizobium. Proc. Natl. Acad. Sci., USA 92,714-71 8 . Koga, J., Adachi, T., and Hidaka, H. (1991a). IAA biosynthetic pathway from tryptophan via indole3-pyruvic acid in Enterobacter cloacae. Agric. Biol. Chem. 55,701-706. Koga, J., Adachi, T., and Hidaka, H. (1991b). Molecular cloning of the gene for indolepyruvate decarboxylase from Enterobacter cloacae. Mol. Gen. Genet. 226, 10-16. Koga, J. (1995). Structure and function of indolepyruvate decarboxylase, a key enzyme in indole-3acetic acid biosynthesis. Biochem. Biophys. Acra 1249, 1-1 3. Koga, J., Adachi, T.,and Hidaka. H. (1992). Purification and characterization of indolepyruvate decarboxylase. J. Biol. Chem. 267, 15822-15828. Koga, J., Syono, K., Ichikawa, T., and Adachi, T. (1994). Involvement of !.-tryptophan aminotransferase in indole-3-acetic acid biosynthesis in Enrerobacter cloaceae. Biochem. Biophys. Actu 1209,241-247. Kolb, W., and Martin, P. (1985). Response of plant roots to inoculation with Azospirillum brasilense and the application of indole acetic acid. In “Azospirillum111: Genetics, Physiology, Ecology” (W. Klingmuller, ed.), pp. 215-221. Springer-Verlag. Berlin, Heidelberg. Kosuge, T., Comai, L., and Glass, N. L. (1983). Virulence determinants in plant-pathogen interactions. In “Plant Molecular Biology” (R. B. Goldberg, ed.), pp. 167-177. Alan R. Liss, New York. Kraigher. H., Grayling, A., Want, T. L., and Hanks, D. E. (1991). Cytokinin production by two ectomycorrhizal fungi in liquid culture. Phytochem. 30,2249-2254. Kravchenko, L. V., Borovkov, A. V., and Pshikvil, Z. (1991). Possibility of auxin synthesis by association-forming nitrogen-fixing bacteria in the rhizosphere of wheat. Mikmbiol. 60,647-658. Kucey, R. M. N. (1987). Increased phosphorus uptake by wheat and field beans inoculated with a phosphorus-solubilizing Penicillium bilaji strain and with vesicular-arbuscular mycorrhizal fungi. Appl. Environ. Microbiol. 53,2699-2703. Kumar-Rao, J . V. D. K., Dart, P. J., and Kiram, U.M. (1984). Rhizobium induced leaf roll in pigeon pea (Cajanus cajan (L.) Millsp.) Soil Biol. Biochem. 16,89-91. Kuo, T. T., and Kosuge, T. (1969). Factors influencing the production and further metabolism of indole-3-acetic acid by Pseudomonas savastanoi. J. Gen. Appl. Microbiol. 15,5 1 4 3 .
PLANT GROWTH REGULATORS IN THE RHIZOSPHERE
13 9
Kuo, T. T., and Kosuge, T. (1970). Role of aminotransferase and indole-3-pyruvic acid in the synthesis of indole-3-acetic acid in Pseudomonas savastanoi. J. Gen. Appl. Microbiol. 16, 191-204. Kuriger, W. E., and Agrios, G. N. (1977). Cytokinin levels and kenetin-virus interactions in tobacco ringspot virus infected cowpea plants. Phytopathology 67,604-609. Kutacek, M. (1988). Enzymes of auxin synthesis in plants and tissue cultures. In “Physiology and Biochemistry of Auxins in Plants” (M. Kutacek, R. S. Bandurski, and J. Krekule, eds.), pp. 57-64. SBA Academic Pub., The Hague, Netherlands. Lambert, B., and Joos, H. (1989). Fundamental aspects of rhizobacterial plant growth promotion research. Trends Biotechnol. 7,215-219. Lebuhn, M., and Hartmann, A. (1993). Method for determination of indole-3-acetic acid and related compounds of L-tryptophan catabolism in soils. J. Chromatogr: 629,255-266. Lebuhn, M., Heilmann, B., and Hartmann, A. (1994). Effects of dryinghwetting stress on microbial auxin production and L-tryptophan catabolism in soils. Biol. Fertil. Soils 18,302-3 10. Lee, K. H., and LaRue, T. A. (1992a). Inhibition of nodulation of pea by ethylene. Plant Physiol. 99 (suppl.): 108. Lee, K. H., and LaRue, T. A. (1992b). Ethylene as a possible mediator of light and nitrate-induced inhibition of nodulation of Pisum sativum L. cv. Sparkle. Plant Physiol. 100, 1334-1338. Lee, K. H., and LaRue, T. A. (1992~).Exogeous ethylene inhibits nodulation of Pisum sativum L. cv. Sparkle. Plant Physiol. 100,1759-1763. Lee, M., Breckenridge, C., and Knowles, R. (1970). Effect of some culture conditions on the production of indole-3-acetic acid and a gibberellin-like substanceby Azorobacrer vinelandii. Can. J. Microbiol. 16, 1325-1330. Leinhos, V., and Vacek, 0.(1994). Biosynthesis of auxins by phosphate-solubilizing rhizobacteria from wheat (Triticum aestivurn) and rye (Secale cereale). Microbial. Res. 149,31-35. Letham, D. S., and Palni, L. M. S. (1983). The biosynthesis and metabolism of cytokinins. Annu. Rev. Plant Physiol. 34, 163-197. Li, N., Wiesman, Z., Liu, D., and Mattoo, A. K. (1992). A functional tomato ACC synthase expressed in Escherichia coli demonstrates suicidal inactivation by its substrate S-adenosylmethionine. FEBSLett. 306, 103-107. Libbert, E., Brunn, K., Drawert, A,, and Schroder, R.(1970~).Pathways of IAA production from tryptophan by plants and by their epiphytic bacteria: Metabolism of indolepyruvic acid and indolelactic acid. Biol. Plant 12,246-255. Lihbert, E., Fischer, E., Drawert, A., and Schroder, R. (1970a). Pathways of IAA production from tryptophan by plants and by their epiphytic bacteria: A comparison. 11. Establishment of the tryptophan metabolites, effects of a native inhibitor. Physiol. Plant. 23,278-286. Libbert, E., Schroder, R., Drawert, A., and Fischer, E. (1970b). Pathways of IAAproduction from tryptophan by plants and by their epiphytic bacteria: A comparison. 111. Metabolism of tryptamine, indoleacetalaldehyde, indoleethanol, and indoleacetamide, effects of a native inhibitor. Physiol. Plant 23,287-293. Lichter, A,, Barash, I., Valinsky, L., and Manulis, S. (1995). The genes involved in cytokinin biosynthesis in Erwinia herbicola pv. gypsophilae: Characterization and role in gall formation. J. Bacteriol. 177,44574465. Lippincott, J . A,, and Lippincott, B. B. (1976). Morphogenetic determinants as exemplified by the crown gall disease. In “Encyclopedia of Plant Physiology” (R. Heitefuss and P. H. Williams, ed.), New Series, Vol. 4, pp. 356388. Springer-Verlag, Berlin, Germany. Liu, G. T., and Kado, C.I. (1979). Indoleacetic acid production: A plasmid function of Agrobacteriurn tumefaciens C58. Biochem. Biophys. Res. Commun. 90, 171-178. Liu, S.-T., Perry, K. L., Schardl, C. L., and Kado, C. I. (1982). Agrobacteriurn Ti plasmid indoleacetic acid gene is required for crown gall oncogenesis. Proc. Nat. Acad. Sci. USA 79,2812-2816. Livingston, W . H. (1991). Effect of methionine and 1-aminocyclo-propane-carboxylicacid on ethylene production by Luccaria bicolor and L. laccata. Mycologia 83,236241.
140 MUHAMMADARSHAJI AND WILLIAM T. FRANKENBERGER,JR. Long, R. L. G., Holland, M. A,, Stebbins, N., Moms, R. 0..and Polacco, J. C. (1996). Evidence for cytokinin production by plant-associated methylotmphs. Plant Physiol. 111,89. Long, S. R., and Cooper, J. (1988). Overview of symbiosis. In “Molecular Genetics of Plant-Microbe Interactions” (R. Palaciss and D. P. S. Verma, eds.), pp. 163-178. APS Press, St. Paul, MN. Loper, J, E., and Schroth, M. N. (1986). Influence of bacterial sources of indole-3-acetic acid on root elongation of sugar beet. Phytopathol. 76,386-389. Lynch, J. M. (1972). Identification of substrates and isolation of microorganisms responsible for ethylene production in soil. Nature 240,4546. Lynch, J. M., and Harper, S. H. T.(1974a). Formation of ethylene by a soil fungus. J. Gen. Microbiol. 80,187-195. Lynch, J. M., and Harper, S. H. T. (1974b). Fungal growth rate and formation of ethylene in soil. J. Gen. Microbiol. 85,91-96. MacDonald, E. M. S., Powell, G. K., Regier, D. A., Glass, N. L., Roberto, F., Kosuge, T., and Morris, R. 0. (1986). Secretion of zeatin, ribosylzeatin, and ribosyl-1”-methylzeatin by Pseudomonas savastanoi: Plasmid-coded cytokinin biosynthesis. Plant Physiol. 82,742-747. Magie, A. R., Wilson, E. E., and Kosuge, T. (1963). Indoleacetamide as an intermediate in the synthesis of indoleacetic acid in Pseudomonas savastanoi. Science 141,128 1-1282. Mahadevan, S., ed. (1984). “Growth Regulators, Microorganisms and Diseased Plants,” p. 466. Oxford and IBH Pub., New Delhi, India. Mahmoud, S. A. Z., Ramadan, E. M., Thabet, F. M., and Khater, T. (1984). Production of plant growth promoting substances by rhizosphere microorganisms. Zentralbl. Mikrobiol. 139,227-232. Mansour, F. A., Ildesuguy, H. S., and Hamedo, H. A. (1994). Studies on plant growth regulators and enzyme production by some bacteria. Qatar Univ. Sci. J. 14,281-288. Mansouri, S., and Bunch, A. W. (1989). Bacterial ethylene synthesis from 2-0x0-4-thiobutyric acid and from methionine. J. Gen. Microbiol. 135,2819-2827. Manulis, S., Gafni, Y., Clark, E., Zutra, D., Ophir, Y., and Bashan, I. (1991b). Identification of a plasmid DNA probe for detection of Enviniu herbicola pathogenic on Gypsophila paniculara. Phytopathol. 81,5657. Manulis, S . , Valinski, L., Gafni, Y., andHershenhorn, J. (1991a). Indole-3-acetic acid biosynthetic pathways in Envinia herbicola in relation to pathogenicity on Gypsophila paniculata. Physiol. Mol. Plant Pathol. 39,161-171. Marocco, A,, Bazzicalupo, M., and Perenzin, M. (1983). Forage grasses inoculation with gentamycine and sulfaguanidine resistant mutants of Azospirillum brasilense. In “Azospirillum II: Genetics, Physiology, Ecology” (W. Klingmiiller, ed.), pp. 149-158. Springer-Verlag, Berlin, Germany. Martens, D. A., and Frankenberger, W. T., Jr. (1991). On-line solid-phase extraction of soil auxins produced from exogenously-applied tryptophan with ion-suppression reverse-phase HPLC analysis. Chromatogr: 32,417422. Martens, D. A., and Frankenberger, W. T., Jr. (1992). Assimilation of 3‘-’4C-indole-3-acetic acid and tryptophan by wheat varieties from nutrient media. In “Proc. 19th Ann. Mtg. Plant Growth Regul. Soc.Am,” pp. 99-100. Plant Growth Regul. Soc. of Amer. Martens, D. A., and Frankenberger, W. T., Jr. (1993a). Stability of microbial-produced auxins derived from L-tryptophan added soil. Soil Sci. 155,263-271. Martens, D. A., and Frankenberger, W. T., Jr. (1993b). Metabolism of tryptophan in soil. Soil Biol. Biochem. 25,1679-1687. Martens, D. A., and Frankenberger, W. T., Jr. (1994). Assimilation of exogenous 2’-14C-indole acetic acid and 3‘-14C-tryptophanexposed to the roots of three wheat varieties. Plant Soil 166,281-290. Martin, P., Glatzle, A., Kolb, W., Omay, H., and Schmidt, W. (1989). N,-fixing bacteria in the rhizosphere: Quantification and hormonal effects on root development. Z. P’anzenernah,: Bodenk. 152, 237-245. Martinez-Toledo, M. V., de la Rubia, T., Moreno, J., and Gonzalez-Lopez, J. (1988). Root exudates of
PLANT GROWTH REGULATORS IN THE RHIZOSPHERE
141
Zea mays and production of auxins, gibberellins and cytokinins by Azofobacrer chroococcum. Planr Soil 110, 149-152. Mauch, F., Hadwiger, L. A., and Boller, T.(1984). Ethylene: Symptoms, not signal for the inducation of chitinase and P-1,3-glucanase in pea pods by pathogen and elicitors. Plant Physiol. 76, 607-6 1I , Mazzola, M., and White, F. F. (1994). A mutation in the indole-3-acetic acid biosynthesis pathway of Pseudomonus syringae pv. syringae affects growth in Phaseolus vulgaris and syringomycin production. J. Bucteriol. 176, 1374-1382. McGaw, B. A., and Burch, L. R. (1995). Cytokinin biosynthesis and metabolism. In “Plant Hormone Physiology, Biochemistry and Molecular Biology” (P. J. Davies, ed.), 2nd ed., pp. 98-1 17. Kluwer Academic Pub., Dordrecht, The Netherlands. McKeon, T.A., Fernandez-Maculet, J. C., and Yang, S. F. (1995). Biosynthesis and metabolism of ethylene. In “Plant Hormone Physiology, Biochemistry and Molecular Biology” (P. J. Davies, ed.), 2nd ed., pp. 118-139. Kluwer Academic Pub., Dordrecht, The Netherlands. Melchers, S., and Hooykaas, P. J. J. (1987). Virulence ofAgrobacterium. Oxford Sum. Plant Mol. Cell Bid. 4, 114-119. Meyer, J. R., and Linderman, R. G. (1986). Response of subterranean clover to dual inoculation with vesicular-arbuscular mycorrhizal fungi and a plant growth-promoting bacterium, Pseudomonas putida. Soil Biol. Biochem. 18, 185-190. Milic, V., Z. Saric, Vorosbayanyi, I., and Mrkovacki, N. (1993). Relation between the content of growth regulators and effectiveness of Brudyrhizobium japonicum. Symbiosis. 15, 183-193. Miller, C. 0. (1967). Zeatin and zeatin riboside from a mycorrhizal fungus. Science 157, 1055-1057. Miller, C. 0.( 1968). Naturally-occurring cytokinins. In “Biochemistry and Physiology of Plant Growth Substances” (F. Wightman and F. Setterfield, eds.), p. 33-35. Runge Press, Ottawa, Canada. Miller, C. 0. (1971). Cytokinin Production by mycorrhizal fungi. In “Ectomycorrhizae. Proc. North Am. Conf. on Mycorrhizae” (E. Hacskagho, ed.), pp. 168-174. U.S. Forest Service, Washington, DC. Miller, C. 0. (1974). Ribosyl-rrans-zeatin, a major cytokinin produced by crown gall tumor tissue. Proc. Natl. Acad. Sci. USA 71, 334-338. Mitchell, R. J., Garrett, H. E., Cox, G. S., and Atalay, A. (1986). Boron and ectomycorrhizal influences on indole-3-acetic acid levels and indole-3-acetic acid oxidase and peroxidase activities of Pinus echinara Mill. roots. Tree Physiol. 1, 1-8. Moore, L., Warren. G.. and Strobel, G. (1979). Involvement of a plasmid in the hairy root disease of plants caused by Agrobacterium rhizogenes. Plasmid 2 , 6 17-626. Moore, T. M. (ed.). (1989). “Biochemistry and Physiology of Plant Hormones,” 2nd ed., p. 330. Springer-Verlag, New York, Berlin, Heidelberg. Morris, R. 0. (1986). Genes specifying auxin and cytokinin biosynthesis in phytopathogens. Am. Rev. Plant Physiol. 37,509-538. Morris, R. 0. (1995). Molecular aspects of hormone synthesis and action. E l . Genes specifying auxin and cytokinin biosynthesis in prokaryotes. In “Plant Hormone Physiology, Biochemistry and Molecular Biology” (P. J. Davies, ed.), 2nd ed., pp. 318-339. Kluwer Acad. Pub., Dordrecht, The Netherlands. Moms, R. O., Akiyoshi, D. E., MacDonald, E. M. S., Moms, J. W., Regier, D. A., and Zaerr, J. B. (1982). Cytokinin metabolism in relation to turner induction by Agmbacterium tumefaciens. In “Plant Growth Substances” (P. F. Wareing, ed.), pp. 175-183. Academic Press, New York. Morris, R. 0..Blevins, D. G., Dietrich, J. T.,Durley, R. C., Gelvin, S. B., Gray, J., Hommes, N. G., Kaminek, M., Matthews, L. J., Meilan, R., Reinbott, T. M., and Luissayavedra-Soto, L. (1993). Cytokinins in plant pathogenic bacteria and developing cereal grains. Ausr. J. Plant Physiol. 20, 621-637. Moms, R. 0..Powell, G. K., Beaty, J. S., Durely, R. C., Hommes, N. G., Lica, L., and MacDonald, E.
142 MUHAMMADARSHAD AND WILLIAM T. FRANKENBERGER, JR. M. S. (1986). Cytokinin biosynthetic genes and enzymes from Agrobacterium tumefaciens and other plant-associated prokaryotes. In “Plant Growth Substances” (M. Bopp, ed.), pp. 185-196. Springer-Verlag, Berlin, Germany. Moszherin, N. M. (1978). Influence of organic supplementson the process of biological fixation of nitrogen in soil and the producing of physiologically active substances.Akad. Nauk SSSR. Sibirskoe Otdelenie. Izvestiia. Seri. Biol. Nauk. pp. 23-26. Murai, N. (1980). Cytokinin biosynthesis and its relationship to the presence of plasmids in strains of Corynebacterium fascians. In “Metabolism and Molecular Activites of Cytokinins” (J. Guern and C. Peaud-Lenoel,eds.), pp. 1-16. Springer-Verlag, Berlin, Germany. Murai, N., Skoog, F., Doyle, M. E., and Hanson, R. S. (1980). Relationships between cytokinin production, presence of plasmids, and fasciation caused by strains of Corynebacferiumfascians. Proc. Natl. Acad. Sci. USA 77,619423. Miiller, M., Deigele, C., and Ziegler, H. (1989). Hormonal interactionsin the rhizosphereof maize (Zea mays L.) and their effects on plant development. Z. PJanzen. Bodenk. 152,247-254. Nagahama, K., Ogawa, T., Fujii, T., Tazaki, M., Goto, M., and Fukuda, H. (1991a). L-Arginine is essential for the formation in vitro of ethylene by an extract of Pseudomonas syringae. J. Gen. Micmbiol. 137, 1641-1646. Nagahama, K., Ogawa, T., Fujii, T., Tazaki,M., Tanase, S., Morino, Y., and Fukuda, H. (1991b). Purification and properties of an ethylene-formingenzyme from fseudomonas syringae pv. phaseolicola PK2. J. Gen. Microbiol. 137,2281-2286. Narayanaswami, R.,and Veerraju, V. (1969). IAA synthesis in paddy soil as influenced by ammonium sulphate fertilization. Cur,: Sci. 38.517-518. Neill, S . J., and Horgan, R.(1983). Incorporation of omega-ionylideneethanol and omega-ionylidene acetic acid into abscisic acid by Cercospora rosicola. Phytochem. 22,2469-2472. Neill, S. J., Horgan, R., Lee, T. S., and Walton, D. C. (1981). 3-methyl-5-(4’-oxo-2‘,6’,6’-trimethylcyclohex-2’-en-l‘-yl)-2,4 pentadienoic acid, a putative precursor of abscisic acid from Cercospora rosicola. FEBS Lett. 128,30-32. Neill, S. J., Horgan, R.,Walton, D. C., and Griffin, D. (1982a). Biosynthesis of abscisic acid. In “Plant Growth Substances. Proc. 1lth Intl. Conf. Plant Growth Substances” (P. F. Wareing, ed.), pp. 315-321. Academic Press, London. Neill, S. J., Horgan, R., Walton, D. C., and Lee, T. S. (1982b).The biosynthesis of abscisic acid in Cercospora rosicola. Phytochem. 21,6145. Neill, S . J., Horgan, R.,Walton, D. C., and Mercer, C. A. M. (1987). The metabolism of a-ionylidene compounds by Cercospora rosicola. Phytochem. 26,25 15-25 19. Nester, E. W., Gordon, M. P., Amasino, R. M., and Yanofsky, M. F. (1984). Grown gall: A molecular and physiological analysis. Ann. Rev. Plant Physiol. 35,387413. Ng, P. P., Cole, A. L. J., Jameson, P. E., and McWha, J. A. (1982). Cytokinin production by ectomycorrhizal fungi. New fhysiol. 91,5742. Nielsen, S. V. S., Egebo, L. A., and Jochimsen, B. (1988). Bradyrhizohial indoleacetic acid metabolism and its significance for root nodule development. In “Molecular Genetics of Plant-Microbe Interactions” (R. Palacios and D. Verma, eds.), pp. 151-152. APS Press, St. Paul., MN. Nieto, K. F., and Frankenberger, W. T., Jr. (1989a). Biosynthesis of cytokinins by Azotobacter chroococcum. Soil Biol. Eiochem. 21,967-972. Nieto, K. F., and Frankenberger. W. T., Jr. (I989b). Biosynthesis of cytokinins in soil. Soil Sci. Soc. Am. J. 53,735-740. Nieto, K. F.,and Frankenberger, W. T., Jr. (1990a). Microbial production of cytokinins. In “Soil Biochemistry” (J. M. Bollag and G. Stotzky, eds.), Vol. 6.. pp. 191-248. Marcel Dekker, New York. Nieto, K. F., and Frankenberger, W. T., Jr. (1990b). Influence of adenine, isopentyl alcohol and Azofobarter chmococcum on the growth of Raphanus sariva (radish). Plant Soil 127, 147-156. Nieto, K. F., and Frankenberger, W. T., Jr. (1991). Influence of adenine, isopentyl alcohol and Azotobarter chroococcum on the vegetative growth of Zea mays. Plant Soil 135,213-221.
PLANT GROWTH REGULATORS IN THE RHIZOSPHERE
143
Noel, T. C., Cheng, C., Yost, C. K., Pharis, R. P., and Hynes. M. F. (1996). Rhizobium leguminoserum as a plant-growth rhizobacterium: Direct growth promotion of canola and lettuce. Can. J. Microbiol. 42,279-283. Norman, S. M., Poling, S. M., Maier, V. P., and Nelson, M. D. (1985). Ionylideneacetic acids and abscisic acid biosynthesis by Cercospora rosicola. Agric. Biol. Chem. 49.23 17-2324. Norrnanly, J., Cohen, I. D., and Fink, G. R. (1993). A tryptophan auxotroph reveals two indole-3-acetic acid biosynthetic pathways in Arabidopsis thaliana. Proc. Nat. Acud. Sci. USA 90, 10355-10359. Oberhansli, T., Dtfago, G., and Haas, D. (1990). Influence of tryptophan side chain oxidase of a Pseudomonasjuorescens strain on indole-3-acetic acid (MA) production and on the suppression of take all of wheat and black root rot of tobacco under gnotobiotic conditions. In “Abstract, 2nd Intl. Workshop on Plant Growth-Promoting Rhizobacteria.” International Union of Biological Sciences. Oberhansli, T., Dtfago, G., and Haas, D. (1991). Indole-3-acetic acid (IAA) synthesis in the biocontrol strain CHAO of Pseudomonasjuorescence: Role of tryptophan side chain oxidase. J. Gen. Microbiol. 137, 2273. Ogawa, T., Takahashi, M., Fujii, T., Tazaki, M., and Fukuda, H. (1990). The role of NADH:Fe(III)EDTA oxidoreductase in ethylene formation from 2-keto-4-methylthio-butyrate. J. Ferment. Bioeng. 69,287-29 1. Okamoto, M., Hirai, N., and Koshimizu, K. (1987). The biosynthesis of abscisic acid in Cercospora pini-densifiorue. In “Proc. 14thAnn. Mtg. Plant Growth Regul. SOC.Am., pp. 45-50. Plant Growth Regul. SOC.Am. Okamoto, M., Hirai, N., and Koshimizu, K. (1988a). Biosynthesis of abscisic acid in Cercosporapinidensijorae. Phytochem. 27,2099-2103. Okamoto, M., Hirai, N., and Koshimizu, K. (1988b). Biosynthesis of abscisic acid from a-ionylidemethanol in Cercospora pini-denstjlorae. Phytochem. 27,3465-3469. Okon, Y., and Labandera-Gonzglez, C. A. (1994). Agronomic applications of Azospirillum. In “Improving Plant Productivity with Rhizosphere Bacteria” (M. H. Ryder, P. M. Stephens, and G. D. Bowen, eds.), pp. 27&278. Commonwealth Scientific and Industrial Research Organization, Adelaide, Australia. Okon, Y.,Bar, T., Tal, S., and Zaady, E. (1991). Physiological properties of Azospirillum brasilense involved in root growth promotion. In “Nitrogen Fixation” (M. Polsinelli, R. Naterassi, and M. Vincenzini, eds.), pp. 113-125. Kluwer Acad. Pub. Norwell, MA. Ondrej, M., Bavrina, T. V., Dudko, N., Hrouda, M., Krekule, J., Lozhrukova, V. N., Machackova, I., Siedlova, F., and Vlasak, I. (1991). Transgenic tobacco plants with T-DNA phytohormone synthesis genes. 5iol. Plant. 3 3 , 4 0 4 8 . Oritani, T., and Yamashita, K. (1985). Biosynthesis of (+)-abscisic acid in Cercosporu cruentu. Agric. Biol. Chem. 49,245-249. Oritani, T., and Yamashita, K. (1 987). Isolation and structure of 4’-hydroxy-y-ionylideneaceticacids from Cercospora cruenta, a fungus producing (+)-abscisic acid. Agric. Biol. Chem. 51,275-278. Oritani, T., Ichimura, M., and Yamashita, K. (1982). The metabolism of analogs of abscisic acid in Cercosporu cruentu. Agric. 5iol. Chem. 46, 1959-1960. Oritani, T., Ichimura, M., and Yamashita, K. (1984). A novel abscisic acid analog, (+)-(22,4E)-S( I ’,4’-dihydroxy-6’,6’-dimethyl-2’-methylenecyclohexyl)-3-methyl-2,4-pentadienoic acid, from Cercosporu cruenta. Agric. Biol. Chem. 48, 1677. Oritani, T., Nitxu, M., Kato, T., and Yamashita, K. (1985). Isolation of (22, UE)-y-ionylideneethanol from Cercospora cruenta, fungus producing (+)-abscisic acid. Agri. Biol. Chem. 49,2819-281 1. Ota, T. (1976). Transformation of Gibberellins in Soil-Especially, on the decomposition of bound type gibberellin. Sci. Rep. Yokohama Nat. Univ. Sec. 2: Biol. Geol. Sci. pp. 77-89. Palni, L. M. S., and Horgan, R. (1983b). Cytokinin biosynthesis in crown gall tissue of Vincu rosea: Metabolism of isopentenyladenine. Phyrochem. 22, 1597. Palni, L. M. S.. Tay, S. A. B., Nandi, S. K., Pianca, D. J., de Klerk, G. J. M., Wong, 0. C., Letham,
I44 MUHAMMAD ARSHAD AND WILLIAM T FRANKENBERGER,JR. D. S., and MacLeod, J. K. (1985). Cytokinin biosynthesis in plant tumor tissues. Biol. Plant. 27, 195-203. Parker-Rhodes, A. F. (1940). Preliminary experiments on the estimation of traces of heteroauxin in soils. J. Agric. Sci. 30,654-67 1. Pati, B. R., Sengupta, S., and Chandra, A. K. (1995). Impact of selected phyllospheric diazotrophs on the growth of wheat seedlings and assay of the growth substances produced by the diazotrophs. Microbiol. Res. 150, 121-127. Patten, C. L., and Glick, B. R. (1996). Bacterial biosynthesis of indole-3-acetic acid. Can. J. Microbial. 42,207-220. Pazout, J., Wurst, M., and Vancura, V. (1981). Effect of aeration on ethylene production by soil bacteria and soil samples cultivated in a closed system. Plant Soil 62,431-437. Pengelly, W. L., and Meins, J. F. (1982). The relationship of IAA content and growth of crown-gall tumor tissues of tobacco (Nicotiana tabacum cultivar Turkish) in culture. Diflerentiation 21,27-3 1. Pengelly, W. L., and Meins, J. F. (1983). Growth, auxin requirement, and indole-3-acetic acid content of cultured crown-gall and habituated tissues of tobacco. Diflerentiation 25, 101-105. Perky, J. E., and Stowe, B. B. (1966b). On the ability of Taphrina deformans to produce indoleacetic acid from tryptophan by way of tryptamine. Plant Physiol. 41,234-237. Peters, N. K., and Crist-Estes, D. K. (1989). Nodule formation is stimulated by the ethylene inhibitor aminoethoxyvinylglycine.Planr Physiol. 91,690-693. Peters, N. K., Frost, J. W., and Long, S. R. (1986).A plant flavone, luteolin, induces expression of Rhizobium meliloti nodulation genes. Science 233,977-980. Phillips, D. A., and Torrey, .I. G.(1972). Studies on cytokinin production by Rhizobium. Plant Physi01. 49, 11-15. Phillips, D. A., Dakora, F. D., Sande, E., Joseph, C. M., and Zon, J. (1994). Synthesis, release, and transmission of alfalfa signals to rhizobial symbionts.Plant Soil 161,6940. Prasad, S. V., Kumar, A. V., Ramaiah, V. V., Rao, K. V. R., Rao, K. N., and Rao, G. R. (1980). Cytokinin-like substances in leachates of green and black gram (Phaseolus radiatus L. and ?? mungo L.). Indian J. Exp. Biol. 18, 159-165. Prikryl, Z., Vancura, V., and Wurst, M. (1985). Auxin formation by rhizosphere bacteria as a factor of root growth. Biologia Plantancm 27(2-3), 159-163. Primrose, S. B. (1976). Ethylene-forming bacteria from soil and water. J. Gen. Microbiol. 97,343-346. Primrose, S . B. (1977). Evaluation of the role of methional, 2-keto-4-methyl-thiobutyric acid and peroxidase in ethylene-formationby Escherichia coli. J. Gen. Microbiol. 98,519-528. Prinsen, E., Chauvaux, N., Schmidt, J., John, M., Wieneke, U., DeGreaf, J., Schell, J.. and Van Onckelen, H. (1991). Stimulation of indole-3-aceticacid production in Rhizobium by flavonoids. FEBS Lett. 282,53-55. Prinsen, E., Costacurta,A., Mchiels, K., Vanderleyden, J., and Van Onckelen, H. (1993). Azospirillum brasilense indole-3-aceticacid biosynthesis: Evidence for a non-tryptophan dependent pathway. Molec. Plant-Microbe Interaction 6,609-61 5 . Puppo, A., and Rigaud, J. (1978). Cytokinins and morphological aspects of French-bean roots in the presence of Rhizobium. Plant Physiol. 42,202-206. Purushothaman, D., Balaraman, K., and Oblisami, G.(1973). Indole acetic acid metabolism in soil as influenced by pesticide application. Cur%Sci. 42,356-366. Purushothaman, D.. Marimuthu, T., Venkataramanan, C. V., and Kesavan, R. (1974). Role of actinomycetes in the biosynthesis of indole acetic acid in soil. Cur,: Sci. 43,413414. Radley, M. (1961). Gibberellin-likesubstances in plants. Nature 191,684-685. Reddy, M. S., Campbell, S. E., Young, S. E., and Brown, G.L. (1991). Role of rhizobacteria in greenhouse mix for the suppression of damping-off of cucumber seedlings caused by Pythium ultimum. Can. J. Plant Pathol. 13,284. Regier, D. A., Akiyoshi, D. E., and Gordon, M. P. (1989). Nucleotide sequence of the tzs gene for Agmbacterium rhizogenes strain A4. Nucleic Acids Res. 17,8885.
PLANT GROWTH REGULATORS IN THE RHIZOSPHERE
145
Reimers, D. B., Taller, B. J., and Stevens, S. E., Jr. (1994). Cytokinin production by the thermophilic cyanobacterium Mastigocladus laminosus UTEX 1931. In “Proc. 94th Gen. Mtg. Am. Soc. Microbiol.,” p. 305. American Society of Microbiology, Washington DC. Reynders, L., and Vlassak, K. (1979). Conversion of tryptophan to indoleacetic acid by Azospirillum brasilense. Soil Biol. Biochem. 11,2771-2775. Rhodes, L. H.,and Gerdemann,J. W. (1978a). Translocation of calcium and phosphate by external hyphae of vesicular-arbuscular mycorrhizae.Soil Sci. 126, 125-26. Rhodes, L. H., and Gerdemann,J. W. (1978b). Hyphal translocation and uptake of sulphur by vesicular-arbuscularmycorrhizae of onions. Soil Biol. Biochem. 10,355-60. Rhodes, L. H., and Gerdemann, J. W. (1978~). Influence of phosphorus nutrition on sulfur uptake by vesicular-arbuscular mycorrhizae of onions. Soil Biol.Biochem. 10,361-64. Rigaud, J. (1970a).La biosynthesede I’acide indolyl-3-acetiqueen liaison avec le metabolismedu tryptophol et de l’indolyl-3-acetaldehydechez Rhizobium. Plant. Physiol. 23, 171-178. Rigaud, J. (1 970b). L‘acide indolyl-3-lactique et son metabolisme chez Rhizobium. Arch. Mikmbiol. 72,297-307. Risuleo, G., Battistoni, P., and Costantino, P. (1982). Regions of homology between tumorigenic plasmids from Agrobacterium rhizogenes and Agrobacterium tumefaciens. Plasmid 7,45-5 1. Roby, D., Toppan, A,, and Esquerrb-Tugayb, M. T.(1985). Cell surfaces in plant-microorganism interactions. V. Elicitors of fungal and of plant origin trigger the synthesis of ethylene and of cell wall hydroxyproline-richglycoproteinin plants. Plant Physiol. 77,700-704. Roby, D., Toppan, A., and Esquerrb-TugayC. M. T. (1986). Cell surfaces in plant-microorganism interactions. VI. Elicitors of ethylene from Colletotrichum lagenanum trigger chitinase activity in melon plants. Plant Physiol. 81,228-233. Rossi, W., Grapelli, A,, and Pietrosanti, W. (1984). Phytohormones in soil after atrazine application. Folia Microbiol. 29,325-329. Rouillon, R., Gay, G., Bemillon, J., Favre-Bonvin,J., and Bruchet, G. (1986).Analysis by HPLC-mass spectrometry of the indole compounds released by the ectomycorrhizalfungus Hebloma hiemale in pure culture. Can. J. Bor. 64,1893-1897. Roussaux, J. (1966). Sur la “fasciation” du Pois en presence de kinetine. C. R. Acad. Sci. Se,: D. 263, 1077-1 080. Rovira, A. D., and Vendrell, M. (1972). Ethylene in sterilized soil. Its significance in studies of interactions between microorganismsand plants. Soil Biol. Biochem. 4 , 6 3 4 9 . Roy, M., and Basu, P. S. (1991). Contents of hormones and indole acetic acid metabolism in root nodules of Clitoria ternatea L. Indian J. Exp. Biol. 30,835-838. Ruckdaschel, E., and Klingmiiller, W. (1992). Analysis of IAA biosynthesis in Azospirillum lipoferum and Tn5 induced IAA mutants. Symbiosis 13,123-131. Ruckdaschel, E., Kittell, B. L., Helsinki, D. R., and Klingmiiller, W. (1988). Aromatic amino acid aminotransferasesof Azospirillum lipoferum and their possible involvementin IAA biosynthesis. In “Azospirillum IV:Genetics, Physiology, Ecology” (W. Klingmiiller, ed.), pp. 49-53. SpringerVerlag, Berlin, Heidelberg. Ruckdaschel,E., Rarnchultz, S., Kraft, K., and Klingmiiller, W. (1990).Genetics and biochemical studies on the production of indole-3-acetic acid (MA) in Azospirillum lipofemm. In “Nitrogen Fixation” (M. Polsinelli, R. Materassi, and M. Vincenzini, eds.), Vol. 48, pp. 301-302. Kluwer Acad. Pub., Dordrecht, The Netherlands. Rudawska, M. (1982).Effect of various organic sources of nitrogen on the growth of mycelium and content of auxin and cytokinin in cultures of some myconhizal fungi. Acta Physiol. Plant. 4, 11-20. Rupp, L. A,, DeVries, H. E., 11, and Mudge, K. W. (1989b). Effect of aminocyclopropanecarboxylic acid and aminoethoxyvinylglycineon ethylene production by ectomycorrhizalfungi. Can. J. Bot. 67,483-485. Rupp, L. A,, Mudge, K. W., and Negm, F. B. (1989a). Involvement of ethylene in ectomycorrhiza formation and dichotomous branching of roots of mugo pine seedlings. Can. J. Bot. 67,477-482.
146 MUHAMMAD ARSHAD AND WILLIAM T. FRANKENBERGER,JR. Sachs, T., and Thimann, K. V. (1967). The role of auxins and cytokinins in the release of buds from dominance. Am. J. Bot. 54,136-144. Salmeron, V., Martinez-Toledo, M. V., and Gonzalez-Lopez, J. (1990). Nitrogen fixation and production of auxins, gibberellins and cytokinins by an Azorobacter chroococcum strain isolated from the root of Zea mays in the presence of insoluble phosphate. Chemosphere 20,417-422. Sandberg, G.,Crozier, A., and Ernstsen, A. ( I 987). Indole-3-acetic acid and relatedcompounds. In “The Principles and Practices of Plant Hormone Analysis” (L. Rivier and A. Crozier, eds.), Vol. 2, pp. 169-301. Academic Press, London. Sarwar, M., and Frankenberger, W. T., Jr. (1994). Influence of L-tryptophan and auxins applied to the rhizosphere on the vegetative growth of Zea mays L. Plant Soil 160,97-104. Sarwar, M., and Frankenberger, W. T., Jr. (1995). Fate of L-phenylalanine in soil and its effect on plant growth. Soil Sci. SOC.Am. J. 59, 1625-1630. Sarwar, M., and Kremer, R. J. (1995a). Determination of bacterially derived auxins using a microplate methods. Lett. Appl. Microbiol. 20,282-285. Sarwar, M., and Kremer, R. J. (1995b). Enhanced suppression of plant growth through production of L-tryptophan-derived compounds by deleterious rhizobacteria. Planr and Soil 172,261-269. Sarwar, M., Arshad, M., Martens, M. A., and Frankenberger, W. T., Jr. (1992). Tryptophan-dependent biosynthesis of auxins in soil. Plant Soil 147,207-215. Sassa, T. (1992). Recent studies on plant hormones and related substances produced by microorganisms. pp. 19-21. In “Proc. 19th Ann. Mtg. Plant Growth Regu. SOC.Am.” Plant Growth Regul. SOC.Am. Sassa, T., Kawaide, H., and Takarada, T. (1994). Identification of gibberellins A,, A,, and A,, from Phaeosphaeria sp. L487 cultured in a chemically defined medium. Biosci. Biotech. Biochem. 58, 438-439. Sassa, T., Suzuki, K., and Haneki, E. (1989). Isolation and identification of gibberellin A,, and A, from a fungus Phaeosphaeria. Agri. Biol. Chem. 53,303-304. Schmidt, J., John, M., Wieneke, U., Stacey, G., Rohrig, H., and Schell, J. (1991). Studies on the function of Rhizobium meliloti nodulation genes. Adv. Mol. Genet. Plant-Microbe Interaction I, 15&155. Schneider, R. W., and Musgrave, M. E. (1992). The soil atmosphere. In “Methods for Research on Soil Borne Phytopathogenic Fungi” (L. L. Singleton, J. D. Mikail, and C. M. Rush, eds.), pp. 215-224. APS Press, St. Paul, MN. Scholz-Seidel, C., and Ruppel, S. (1992). Nitrogenase and phytohormone activities of Pantoea agglumerans in culture and their reflection in cornbindtion with wheat plant. Zentralbl. Microbiol. 147, 3 19-328. Schroth, M. N., Loper, J. E., and Hildebrand, D. C. (1984). Bacteria as biocontrol agents of plant disease. In “Current Perspectives in Microbiol Ecology” (M. J . Klug and c. A. Reddy, eds.), pp. 361-369. Am. Soc. for Microbiol.. Washington, DC. Schroder, G.,Waffenschmidt, S., Weiler, E. W., and Schroder, J. (1984). The T-region of Ti plasmids codes for an enzyme synthesizing indole-3-acetic acid. Eul: J. Biochem. 138,387-391. Schwartzenberg, K. von, Bonnet-Masimbert, M., and Doumas, P. (1994). Isolation of two cytokinin metabolites from the rhizosphere of Norway spruce seedlings (Picea abies L. Karst.). Plant Growth Regul. 15, 117-124. Schwarz, K., and Bitancourt, A. A. (1957). Paper chromatography of unstable substances. Science 126, 607-608. Schwinn, F. J. (1965). Untersuchungen zum Wirt-Parasit-Verhaltnis bei der Kragenfaule des Apfelbaumes (Phytophthora cactorum) und zuihyer Behampfung. 11. Zur Rolle der Inddwuchsstoffe furden erreger. Phytopathol. Z. 54, 162-184. Scott, I. M., and Horgan, R. (1984). Mass spectrometric quantification of cytokinin nucleotides and glycosides in tobacco crown gall tissue. Planta. 161,345-356. Seagel, C. F., and Lindeman, R. G.(1994). Increases in endogenous IAAcontent of conifer roots mediated by mycorrhizal fungi and exogenously applied plant growth regulators. Plant Physiol. 105,143.
PLANT GROWTH REGULATORS IN THE RHIZOSPHERE
147
Sekine, M.. Ichikawa, T., Kuga, N., Kobayashi, M., Sakurai. A., and Syono, K. (1988). Detection of the IAA biosynthetic pathway from tryptophan via indole-3-acetamide in Bradyrhizobium spp. Plant Cell Physiol. 29,867-874. Sekine, M., Watanabe, K., and Syono, K. (1989). Molecular cloning o f a gene for indole-3-acetamide hydrolase from Bradyrhizobium japonicum. J. Bacteriol. 171, 1I1 8- 1724. Sembdner, G., Gross, D., Liebisch, H.-W., and Schneider, G.(1980). Biosynthesis and metabolism of plant hormones. In “Hormonal Regulation of Development. I. Molecular Aspects of Plant Hormones.” Encyclopedia of Plant Physiology, Vol. 9., pp. 28 1 4 4 . (J. MacMillan, ed.). SpringerVerlag, Berlin, Germany. Sequeira. L. (1973). Hormone metabolism in diseased plants. Ann. Rev. PIarir Physiol. 24, 353-380. Sequeira, L., and Williams, P. H. (1964). Synthesis of indoleacetic acid by Pseudomonas sohacearum. Phytopathol. 54, 1240-1246. Serdyuk. 0. P., Smolygina, L. D., Muzafarov, E. N., Adanin, V. M., and Arinbasarov, M. U. (1995). 4Hydroxyphenethyl alcohol-a new cytokinin-like substance from the phototrophic purple bacterium Rhodospirillum rubrum IR. FEBS Letters 365, 10-12. Shen, W. H., Davioud, E., David, C., Barbier-Brygoo, H., Tempe, J., and Guern, J. (1990). High sensitivity to auxin is a common feature of hairy root. Planr Physiol. 94,554560. Shenvood, J. L., and Klarman, W. L. (1980). IAA involvement in fungal protection of Virginia pine seedlings exposed to methane. Forest Sci. 26, 172-1 76. Shipston, N., and Bunch, A. W. (1989). The physiology of L-methionine catabolism to the secondary metabolite ethylene by Escherichia coli. J. Gen. Microbiol. 135, 1489-1497. Silverstone, S. E., Bostock, R. M., Gilchrist, D. G., and Kosuge, T. (1990). Effect of IAA production on in plantu survival of Pseudomonas savastanoi. Phytopathol. 80,972. Slankis, V. (1973). Hormonal relationships in mycorrhizal development. In “Ectomycorrhizae: Their Ecology and Physiology” (G. C. Marks and T. T. Kozlowski, eds.), pp. 23 1-298. Academic Press, New York. Smidt. M., and Kosuge, T. (1978) The role of indole-3-acetic acid accumulation by alpha methyl tryptophan-resistant mutants of Pseudomonas savastanoi in gall formation on oleanders. Physiol. Plant Paihol. 13, 203-214. Smith, A. M., Milham. P. J., and Morrison, W. L. (1978). Plant diseases: Soil ethylene production specifically triggered by ferrous iron. In “Microbial Ecology” (M. W. Loutit and J. A. R. Miles, eds.), pp. 329-336. Springer-Verlag, New York. Smith. K. A,. and Restall. S. W. E (1971 ). The occurrence of ethylene in anaerobic soil. J. Soil Sci. 22, 430443. Smith, K. A., Bremner, J. M., and Tabatabai, M. A. (1973). Sorption of gaseous atmospheric pollutants by soils. Soil Sci. 116,313-3 19. Spano, L., Pomponi, M., Costantino, P., van Slogteren, G. M. S., and Tempe, J. (1982). Identification ofT-DNAin the root inducingplasmid ofthe agropine type Agrobacrerium rhizogenes 1855. PIant Mol. Biol. 1, 291-300. Sponsel,V. M. [ 1995).The biosynihesis and metabolism of gibberellins in higher plants. In “Plant Hormones, Physiology, Biochemistry, and Molecular Biology” (P. J. Davies, ed.), 2nd ed., pp. 66-97. Kluwer Acad. Pub., Dordrecht, The Netherlands. Sridhar, R., Chandramahan, D.. and Mahadevan, A. (1968). The role of tryptamine in indole acetic acid synthesis by Fusarium oxysporum f. vasinfecrum and Pyricularia oryuie. Indian J. Exp. Bid. 6, 59-60. Srivastava, B. I. S. ( 1964). Investigations on the occurrence and biosynthesis of indole-pyruvic acid in plant tissues and bacteria. Plant Physiol. 39,78 1-785. Stall, R. E., and Hall, C. B. (1984). Chlorosis and ethylene production in pepper leaves infected by Xanthomonas compestris pv. vesicatoria. Phytopathol. 74, 373-375. Stewart, W. S., and Anderson, M. S. (1942). Auxins in some American soils. Bot. Gaz. 103,570-575. Stokkermans, T. J . W.. Sanjuan, J., Ruan, X., Stacey, G.. and Peters, N. K. (1992). Bradyrhizobiurn
148 MUHAMMAD ARSHAD AND WILLIAM T. F W N B E R G E R , JR. japonicum rhizobitoxine mutants with altered host-range on Rj4 soybean. Plant Physiol. 99 (Supp1J:llO. Stowe, B. B. (1955). The production of indoleacetic acid by bacteria. Biochem. J. 61,9-10. Strzelczyk, E., and Kampert, E. (1987). Effect of B-group vitamins on cytokinin-likesubstances production by ectomycorrhizal-fungi of pine (Pinus sylvesfiis L.) Symbiosis 3,135-1 46. Strzelczyk, E., and Karwowska, J. M. (1969). Microbial degradation of indoleaceticacid (IAA) in soil. Pol. J. Soil Sci. 2,59-64. Strzelczyk, E., and Pokojska-Burdziej, A. (1984). Production of auxins and gibberellin-likesubstances by mycorrhizal fungi, bacteria and actinomycetes isolated from soil and the mycorrhizo-sphere of pine (Pinus silvestris L.). Plant Soil 81, 185-194. Strzelczyk,E., Kampert, E., and Michalski, L. (1985). Production of cytokinin-likesubstances by mycorrhizal fungi of pine (Pinus silvesrris L.) in cultures with and without metabolites of actinomycetes. Acta Micmbiol. Pol. 34, 177-1 86. Strzelczyk,E., Pokojska, A,, and Kampert, M. (1992). The effect of pH on production of plant growth regulators by mycorrhizal fungi. Symbiosis 14,201-215. Strzelczyk,E., Sitek, J. M., and Kowalski, S. (1975). Production of gibberellin-like substancesby fungi isolated from mycorrhizae of pine (Pinus sihestris L.). Acta Microbiol. Pol. 7 , 145-153. Strzelczyk, E., Sitek, J. M., and Kowalski, S. (1977). Synthesis of auxins from tryptophan and tryptophan-precursors by fungi isolated from mycorrhizae of pine (Pinus silvestris L.). Acta Microbiol. POI. 26,255-264. Sturtevant,D., and Taller, B. J. (1988). Effect of seed extract on cytokinin production by Bradrhizobium. In “Physiology and Biochemistq of Plant-Microbe Interactions”(N. T. Keen, T. Kosuge, and L. L. Walling, eds.), pp. 64.Amer. SOC.Plant Physiologists,Rockville, MD. Sturtevant,D., and Taller, B. J. (1989). Cytokinin production by a Parasponia nodule bacterium. In “Abstract,89th Ann. Mtg. Am. SOC.Microbio1,”p. 300. American Society of Microbiology, Washington, DC. Surico, G., Evidente, A., Iacobellis, N., and Randazzo, G. (1985). A new cytokinin from the culture filterate of Pseudomonas synngae pv. savastanoi. Phytochem. 24, 1499-1 502. Surico, G., Sparapano, L., Legaro, P., Durbin, R. D., and Iacobellis, N. (1975) Cytokinin-likeactivity in extracts from the culture filterate of Pseudornonas savastanoi. Experientia 31,929-930. Swart, A., and Kamerbeek, G. A. (1977) Ethylene production and mycelium growth of the tulip strain of Fusarium oxysporum as influenced by shaking of and oxygen supply to the culture medium. Physiol. Plant. 3 9 , 3 8 4 . Tagushi, F.,Mizukami, N., Saito-Taki, T., and Hasegawa, K. (1995). Hydrogen production from continuous fermentation of xylose during growth of Clostridium sp. strain No. 2. Can. J. Micmbiol. 41,536-540. Takahashi, N. (ed.) (1986). Chemistry Plant Hormones. CRC Press, Boca Raton, FL. Takai, K., and Hayaishi, 0. (1987). Purification and properties of tryptophan side chain oxidase types I and I1 from Pseudomonas. Meth. Enzymol. 142,195-217. Takenaka, M., Hayashi, S. E., Ogawa, T., Kimura T., and Tanaka, T. (1992). Lowered virulence to rice plants and decreased biosynthesis of gibberellinsin mutants of Gibberellafijikuroi selected with perfurazoate. J. Pesticide Sci. 17,2 13-220. Taller, B. J., and Sturtevant,D. B. (1991). Cytokinin production by rhizobia. In “Current Plant Science and Biotechnology in Agriculture,Vol. 10. Advances in Molecular Genetics of Plant-Microbe Interactions, Vol. 1” (H. Hennecke and D. P. S. Verma, eds.), pp. 215-221. Kluwer Academic Pub., Dordrecht, Netherlands. Taller, B. J., and Wong, T.-Y. (1989). Cytokinins in Azotobacrer vinelandii culture medium. Appf. Envirun. Microbiol, 55,266-267. Tang, W. H. (1994). Yield-increasing bacteria (YIB)and biocontrol of sheath blight of rice. In “Improving Plant Productivity with Rhizosphere Bacteria” (M. H. Ryder, P. M. Stephens, and G. D.
PLANT GROWTH REGULATORS TN THE RHIZOSPHERE
149
Bowen, eds.), pp. 267-278. Commonwealth Scientific and Industrial Research Organization, Adelaide, Australia. Taya, Y., Tanaka, Y.,and Nishimura, S. (1978). 5’-AMP is a direct precursor of cytokinin in Dictyosrelium discoideum. Nature 271,545-547. Thiagarajan, T. R., and Ahmad, M. H. (1994). Phosphatase activity and cytokinin content in cowpeas (Vigna unguiculata) inoculated with a vesicular-arbuscular mycorrhizal fungus. Biol. Fertil. Soils 17,51-56. Thimann, K. V., and Sachs, T. (1966). The role of cytokinins in the “fasciation” disease caused by Corynebacterium fascians. Am. J. Bot. 53,73 1-739. Thomashow, L. S., Reeves, S., and Thomashow, M. F. (1984). Crown gall oncogenesis: Evidence that a T-DNA gene from the Agrobacterium Ti plasmid pTiAG encodes on an enzyme that catalyzes synthesis of indoleacetic acid. Proc. Natl. Acad. Sci. USA 81,5071-5075. Thomashow, M. F., Hugly, S., Buchholz, W. G., and Thomashow, L. S. (1986). Molecular basis for the auxin-independent phenotype of crown gall tumor tissue. Science 231,616-61 8. Tien,T. M., Gaskins, M. H., and Hubbell, D. H. (1979).Plant growth substances produced by Azospirillum brasilense and their effect on the growth of pearl millet (Penniseturn americanum L.). Appl. Environ. Micmbiol. 37, 1016-1024. Tomaszewski, M., and Wojciechowska, B. (1974). The role of growth regulators released by fungi in pine mycorrhizae. In “Plant Growth Substances.Proc. 8th Int. Conf. on Plant Growth Substances,” pp. 217-227. Hirokawa, Tokyo. Toppan, A,, Roby, D., and Esquerrk-TugayB, M. T. (1982). Cell surfaces in plan-microorganism interactions. 111. In vivo effect of ethylene on hydroxyproline-rich glycoprotein accumulation in the cell wall of diseased plants. Plant Physiol. 70, 82-86. Triplett, E. W., Heitholt, I. J., Evensen, K. B., and Blevins, D.G. (1981). Increase in internode length of Phaseolus lunatus L. caused by inoculation with a nitrate reductase-deficient strain of Rhizobium sp. Plant Physiol. 67, 1 4 . Tuomi, T., Ilvesoksa, J., Laakso, S., and Rosenqvist, H. (1993). Interaction of abscisic acid and indole3-acetic acid-producing fungi with Salix leaves. J. Plant Growth Regul. 12, 149-156. Ulrich, J. M. (1960). Auxin production by mycorrhizal fungi. Physiol. Plant. 13,429-443. Upadhyaya, N. M., Letham, D.S., Parker, C. W., Hocart, C. H., and Dart,P. J. (1991a). Do rhizobia produce cytokinins? Biochern. Intl. 24, 123-130. Upadhyaya, N. M., Parker, C. W., Letham, D. S., Scott, K. F., and Dart, P.J. (1991b). Evidence for cytokinin involvement in Rhizobium (lC3342)-induced leaf curl syndrome of pigeon pea (Cajanus cajan Millsp.). Plant Physiol. 95, 1019-1025. Van Andel, 0. M., and Fuchs, A. (1972). Interference with plant growth regulation by microbial metabolites. In “Phytotoxins in Plant Diseases” (R. K. S. Wood, A. Ballio, and A. Graniti, eds.), pp. 227-249. Academic Press, London. Van Onckelen, H. A,, Prinsen, P., Inze, D., Rudelsheim, P., van Lijsebettens, M., Follin, A,, Schell, J., van Montagu, M., and DeGreef, J. (1986). Agrobacteriurn T-DNA, gene I codes for tryptophan2-monooxygenase activity in tobacco crown gall cells. FEBS Lett. 198,357-360. Van Onckelen, H. A,, Rudelsheim, P., Inze, D., Follin, A., Messens, E., Horemans, S., Schell, J., Van Montagu, M., and DeGreef, J. (1985). Tobacco plants transformed with the Agrobacteriurn TDNA gene I contains high amounts of indole-3-acetamide. FEBS Len. 181,373-376. van Staden, J., and Dimalla, G. G. (1976). Cytokinins from soils. Planta 130,85-87. Walton, D. C., and Li, Y .(1995).Abscisic acid biosynthesis and metabolism. In “Plant Hormones, Physiology, Biochemistry, and Molecular Biology” (P. J. Davies, ed.), 2nd ed., pp. 140-157. Kluwer Acad. Pub., Dordrecht, The Netherlands. Wang, C. H., Persyn, A., and Krackov, J. (1962). Role of the Krebs cycle in ethylene biosynthesis. Nature 195, 13&1308. Wang, T. L., Wood, E. A,, and Brewin, N. J. (1982). Growth regulators, Rhizobium and nodulation in
150 MUHAMMAD ARSHAD AND WILLIAM T. FRANKENBERGER,JR. peas: Indole-3-acetic acid from the culture medium of nodulating and non-nodulating strains of R. legurninosarum. Planra 155,345-349. Weiler, E. W., and Spanier, K. (1981). Phytohormones in the formation of crown gall tumors. Planta. 153,326337. Wells, S. E., and Kuykendall, L. D. (1983). Tryptophan auxotrophs of Rhizobium japonicum. J. Bacteriol. 256, 1356-1358. West, C. A. (1973). Biosynthesis of gibberellins. In “Biosynthesis and Its Control in Plants” (B. V. Milborrow, ed.), pp. 473-482. Academic Press, New York, London. Wheeler, C. L., Crozier, A,, and Sandberg, G. (1984). The biosynthesis of indole-3-acetic acid by Frankia. Plant Soil 79,99-104. White, E F., and E. W. Nester. (1980). Relationship of plasmids responsible for hairy root and crown gall tumorigenicity. J. Bacteriol. 144,7 10-720. White, F. F., and Ziegler, S. F. (1991). Cloning of the genes for indoleacetic acid synthesis for Pseudomonas syringae pv. syringae. Mol. Plant. Microbe Interaction 4,207-2 10. White, F.F., Ghidossi, G., Gordon, M. P., and Nester, E. W. (1982). Tumor induction by Agrobucterium rhizogenes involves the transfer of plasmid DNA to the plant genome. Proc. Nut. Acad. Sci. USA 79,3193-3197. White, F. F., Taylor, B. H., Huffman, J. G. A,, Gordon, M. P., and Nester, E. W. (1985). Molecular and genetic analysis of the transferred DNA regions of the root inducing plasmid of Agmbacrerium rhizogenes. J. Bacterial. 164,3344. Wilkinson, K. G., Dixon, K. W., Sivasithamparam, K., and Ghisalberti, E. L. (1994). Effect of IAA on symbiotic germination of an Australian orchid and its production by orchid-associated bacteria. Plant Soil 159,291-295. Williams, P. M., and de Mallorca, M. S. (1982). Abscisic acid and gibberellin-like substances in roots and root nodules of Glycine may. Plant Soil 65, 19-26, Willmitzer, L., Sanchez-Serrano, J., Buschfeld, E., and Schell, J. (1982). DNA from Agrobacterium rhizogenes is transferred to and expressed in axenic hairy root plant tissues. Mol. Gen. Genet. 186, 1622. Wohler, I., Lebuhn, M., and Hartmann, A. (1990). Occurrence of indole-3-acetic acid and related compounds in forest and agricultural soils. In “Abstract 2nd Intl. Workshop on Plant GrowthPromoting Rhizobacteria,” p. 42. International Union of Biol. Sciences. Wright,A. D., Sampson, M. B., Neuffer, M. G., Michalezuk, L., Slovin, J. P., andCohen, J. D. (1991). Indole-3-acetic acid biosynthesis in the mutant maize Orange pericarp, a tryptophan auxotroph. Science 254,998-1000. Xie. H., Pasternak, J. J., and Glick, B. R. (1996). Isolation and characterization of mutants of the plant growth-promoting rhizobacteriurn Pseudomonas putidu GR12-2 that overproduce indoleacetic acid. Current Microbiol. 32,67-71. Yakushkina, N. I., and Tarasenko, A. A. (1975). Comparative study of hormone content and growth of seedlings cultivated under sterile and unsterile conditions. Sox Plant Physiol. 22, 129-133. Yamada, T., Palm, C. J., Brooks, B., and Kosuge, T.(1985). Nucleotide sequences of the Pseudomonas savastanoi indoleacetic acid genes show homology with Agrobacterium tumefaciens T-DNA. Proc. Nut. Acad. Sci. USA 82,6522-6526. Yamamoto, H., Sugiyama, T., and Oritani, T. (1996). Conformational analysis of abscisic acid analogs produced by Cercospora cruenra. Biosci. Biotech. Biochem. 60,750-754. Yoshida, R., and Takano, K. (1992). Abscisic acid produced by plant pathogenic fungi and symptom expression in diseased leaves of flowers. In “Proc. 19th Ann. Mtg. Plant Growth Regulator SOC. Am.,” pp. 159-164. Plant Growth Regu. SOC.Am. Young, S., Reddy, M. S., Brown, G., and Rennie, R. (1992). Biological activities induced by rhizobacteria and influence on spring wheat yield. Phytopathol. 82, 11 71. Yung, K. H., Yang, S. E, and Schlenk, F. (1982). Methionine synthesis from 5methylthioribose in apple tissue. Biochem. Biophys. Res. Commun. 104,771-777.
PLANT GROWTH REGULATORS IN THE RHIZOSPHERE
15 1
Zaat, S. A. J., Wijrelmans, C. A,, Spain, H. P., van Brussel, A. A. N., Okker, R. J. H., and Lugtenberg, B. J. J. (1987). Induction of the nodApromoter of Rhizobium leguminosarum Sym plasmid plUIJI by plant flavanones and flavones. J. Bacferiol. 169, 198-204. Zaenen, I., van Larabeke, N., Teuchy, H., Van Montagu, M., and Schell, J. (1974) Super-coiled circular DNA in crown gall inducing Agrobacterium strains. J. Mol. Biol. 86, 109-127. Zahir, A. Z., Arshad, M., Azam, M., and Hussain, A. (1997). Effect of an auxin precursor tryptophan and Azotobacter inoculation on yield and chemical composition of potato under fertilized condiI . Plant Nutrition (In press). tions. . Zeevart, J. A. D., and Creelman, R. A. (1988). Metabolism and physiology of abscisic acid. Ann. Rev. Plant Physiul. Plant Mol. Eiol. 39,439473. Zimmer, W., and Elmerich, C. (1991). Regulation of the synthesis of indole-3-acetic acid in Azospirillum. In “Advances in Molecular Genetics of Plant-Microbe Interactions” (H. Hennecke and D. P. S. Verma, eds.), Vol. 1, pp. 465468. KluwerAcad. Pub., London. Zimmer, W., Aparicio, C., and Elmerich, C. (1991). Relationship between tryptophan biosynthesis and indole-3-acetic acid production in Armpirillurn identification and sequencing of a trpGDC cluster. Mol. Gen. Genet. 229,41-51. Zimmer. W., Bundeshagen, B., and Niederau, E. (1994). Demonstration of the indolepyruvate decarboxylase gene homologue in different auxin-producing species of the Enterobacteriaceae. Can. J. Micmbiol. 40, 1072-1076. Zimmer, W., Roeben, K., and Bothe, H. (1988). An alternative explanation for plant growth promotion by bacteria of the genus Azospirillum. Planta 176,333-342. Zupancic, A,, and Gogala, N. (1980). The influence of root exudate auxins and gibberellins on the growth of Suillus variegutur mycelium. Actu Bot. Croat. 39,85-93.
This Page Intentionally Left Blank
LONG-TERMTRENDS OF CORNYIELD AND SOILORGANIC MATTER INDIFFERENT CROPSEQUENCES AND SOILFERTILITY TREATMENTS ON THEMORROW PLOTS Susanne Aref and Michelle M. Wander Department of Crop Sciences and Department of Natural Resources and Environmental Sciences University of Illinois a t Urbana-Champaign Urbana, Illinois 61801
I. Introduction and History of the Morrow Plots A. Introduction B. Morrow Plots Treatments and Their Historical Context C. Records and Data from the Morrow Plots 11. Corn Xeld A. Experimental Effects B. Significance of Technology, Planting Date, and Weather C . Connection with Illinois State Average Corn Yield ILL Soil Variables: Soil Organic Matter, pH, P, and K A. Soil Organic Matter B. pH, P, and K C. SOM: Interaction between Corn Xeld and Soil Fertility rv. Conclusions: Lessons from the Morrow Plots Appendix: Abbreviations References
I. INTRODUCTION AND HISTORY OF THE MORROW PLOTS
A. INTRODUCTION I. The Morrow Plots and Long-Term Experiments There are four field experiments in the United States that are over 100 years old. These include the Sanborn Field in Missouri, started in 1888; the Magruder Plots in Oklahoma, started in 1892;the Old Rotation Plots in Alabama, started in 1896; and 153 Adunnces in Agronomy, Volume 62 Copyright 0 1YY8 by Academic Press. All rights of reproduction in any form reserved 0065-21 13/98 $25.00
154
SUSANNE AREF AND MICHELLE M. WANDER
the oldest experiment of all, the Morrow Plots in Illinois, started in 1876. The Sanborn Field and the Morrow Plots experiments are both rotation, fertility, and manure experiments with corn as the primary crop. Wheat is grown in the Magruder Plots, and cotton is grown in the Old Rotation Plots. All four experiments have been designated National Historic Landmarks. The focus of these old experiments has been the productivity and sustainability of continually cultivated land. (For more information about long-term experiments, see Mitchel et al., 1991;Paul et al., 1997.)
2. Management of Long-Term Experiments The Morrow Plots are better known for their historic significance than for the scientific information they have produced. Experiments like the Morrow Plots that are National Historic Landmarks cannot be ended. Whether or not they merely survive as historic places or continue to be viable experiments will depend on how they are managed and perceived. In long-term studies, treatments are applied for a long enough time period to assess management impacts on the resource base; this aspect of long-term studies demands a certain permanence to the design and, in some ways, makes them inflexible. Careful consideration must be used in their management. During the life of historic experiments, some practices lose their practical relevance. Rash change can terminate treatments under study before valuable information is gained and/or confound subsequent findings. It is the viable long-term experiment that can be modified to include new questions as time passes without sacrificing the original experiment’s intents. The Morrow Plots provide an example of such an experiment. This trial has been modified only a few times. Remarkably, treatments were added without ruining the initial setup or losing sight of the original questions. The experiment has evolved to study not only the effects of rotations but also timely fertility treatments. Stewards of the Morrow Plots have embellished the initial design, adding or expanding questions at historically appropriate points in time. The original 10-plot experiment was intended to determine if corn yield would remain consistent after years of planting the same land (Hopkins et al., 1908).This soil exhaustion study began in 1876with three different crop rotations and three fertility treatments. By 1904 the experiment had been reduced to only three of the original rotation plots, continuous corn (CC),corn-oat (C-0) and corn-oat-hay (C-0-H), all of which had not been fertilized. At this time, the three rotation main plots were split, and manure, lime, and phosphorus were added to their southern halves. Until 1955, the southern half of the rotation plots were then managed in accordance with what was known as the Illinois system of permanent soil fertility. This was a “philosophy” of soil management advocated by University of Illinois soil scientists who had worked diligently to idenhfy practices to support agriculture on a permanent basis (Smith, 1925). The use of hybrids started in 1937, reflecting agronomic practices common at that time. Use of commercial fertilizers became common during the middle of the century. In 1955, when all six main
MORROW PLOTS
1.5.5
plots were in corn, a nitrogen-phosphate-potassium (NPK) treatment was added to part of the plots. This third phase of the trial would determine if, and how rapidly, productivity could be restored to previously untreated plots and how treatment history (untreated or manured) affected yield response. At this time, seeding rates were adjusted to match soil fertility levels. The fourth phase of the trial was begun twelve years later, in 1967. Soybean replaced oat in the two-year C-0 rotation, and a very high-level NPK treatment replaced manure applications on some plots. This last modification brought the C-0 rotation in line with regional cropping practices and would deteimine how very high levels of fertilizer application affected productivity.
3. Relevance of the Morrow Plots Overall trends and cumulative impacts of management systems are best studied through long-term experiments (Peck, 1989; Mitchell et al., 1991; Barnett et al., 1995). In the Morrow Plots, changing agricultural norms have competed with experimental need to sustain treatments long term. The result has been three surviving rotation plots that have been reduced in size and repeatedly subdivided. Even with their limitations, long-term experiments like the Morrow Plots provide us with the only reasonable empirical basis upon which we can evaluate agricultural sustainability (Steiner, 1995). The changing nature of the Morrow Plots’ treatments reflect the agricultural history of the region and show that managers of the plots have been interested in both short- and long-term aspects of productivity.
4. Objectives This chapter describes the history of the Morrow Plots and the effects of their experimental treatments on corn yield and soil variables. We report on all data available at this time. For the first time, statistical analyses of the complete yield and soil organic matter (SOM) record are presented. We consider trends in response to rotation and fertility treatments applied during the main phases of the Morrow Plots’ history. Yield responses simultaneouslyreflect changing soil productivity levels and the immediate effects of improved technology and weather. By analyzing soil variables, soil organic C, total N, pH, P, and K throughout the trial’s history, we study the long-term effects of treatments on the system’s productive potential.
B. MORROWPLOTSTREATMENTS AND THEW HISTORICAL CONTEXT
1. Phase 1: 18761903, Life and Times during Plot Establishment a. History of the University The University of Illinois was founded as the Illinois Industrial University in 1867, as Dean George E. Morrow put it, to “secure a wider and better education
156
SUSANNE AREF AND MICHELLE M. WANDER
for the industrial class,” noting that it was not to be “simply and only a collection of farms and work shops in which the manual labor of the farmer and mechanic is shown and practiced.” b. Establishing Treatments, 10 Plots The Morrow Plots experiment began in 1876 to produce results “suggestive to the practical farmer” (Illini Week,Oct. 26, 1989). C. W. Silver’spromptings directly inspired this kind of trial. Silver (1875) advocated that Illinois emulate trials that had been going on in Europe for a generation using American crops. Taking the suggestions of Silver, Manley Miles, then professor of agriculture, established the experiment that may have become the Morrow Plots. It is not certain that Miles, who left the university after only one year, established the trial (called Experiment 23) that became the Morrow Plots. It is known that the management of Experiment 23 fell to Morrow, who had also been inspired by Rothamsted during an 1879 sabbatical trip. Morrow successfully fought an uphill battle for the College of Agriculture against the prevailing view that one did not go to a university to study agriculture. Morrow also had to overcome the fact that his experiment was seen by farmers as too “academic” (TheIllinois Agriculturalist, Oct. 1937). The Morrow Plots are located at 40’06’15” north latitude and 88’13’32” west longitude, on a prime piece of real estate on the University of Illinois campus. Little is known about the history of the site prior to plot establishment. Furthermore, no records were kept between 1876 and 1887.The site was part of the North Farm, which had been in agricultural use for some time, possibly 40 years or more before the plots were established (Darmody and Peck, 1997). The original trial consisted of five acres divided into 10 plots. The initial experimental setup was nearly balanced. There were three CC (&a mays L.) rotation plots, one C-0 (Avena sativa) rotation plot, and six C-0-H rotation plots. Hay crops included sweet and red clovers (Melilotus alba and Trifoliumpratense),with occasional inclusion of soybean (Glycine m a ) or cowpea (Vigna ungiculata). Even though this was primarily a rotation experiment, fertilizer treatments were used on two of the CC plots. Plot 1 received manure, and plot 2 received P, K, and sulfurous ammonium (PKS) amendments. The third CC plot received no amendments. The fourth plot was the only plot in C - 0 rotation, so corn was only grown every other year in this rotation. The remaining six plots were devoted to the C-0-H rotation, which was initially planted in a 6-year cycle: corn-com-oat-hay-hay-hay. These plots were offset so that each part of the cropping cycle was represented every year (Table I, phase 1). All plots were moldboard plowed in the fall. There is some indication that catch crops were sporadically grown in the CC and C -0 rotations. In August 1895 an observatory (now also a National Historic Landmark) was built on two of the fertilized CC plots (1 and 2), leaving only the rotation portion of the original trial. Lack of interest in the experiment resulted in a breakdown in
Table I 'Ikeatments in the Four Phases of the Morrow Plots Rotation
Direction N-S
W
E
~
A
Phase 1: 1876-1903" Plot I Plot 2 Plot 3 Plot 4 Plots 5-10 Phase 2: 19041954b Plot 3
cc cc cc c-0 cc c-0
Plot 5
C-0-H
cc
Plot 4
c-0
Plot 5
C-0-H
Phase 4:1967-1996d Plot 3 Plot 4 Plot 5
cc c-0 C-0-H
C
D
M PKS U U U
C-0-H
Plot 4
Phase 3: 1955-1966< Plot 3
B
N S N S N S
U M with LrP U M with LrP U M with LrP
U M with LbP U M with LbP U M with LbP
N S N S N S
U M U M U M
U-NPK M-NPK U-NPK M-NPK U-NPK M-NPK
U MPS U MPS U MPS
U M U M U M
N S N S N S
U H-NPK
U-NPK M-NPK U-NPK M-NPK U-NPK M-NPK
U MPS
U M U M U M
U
H-NPK U H-NPK
U
MPS U MPS
"Phase I , then called Experiment 23, included 10 plots with three crop rotations: continuous corn (CC), corn-oats (CO), and corn-ats-hay (C-0-H). Plot 1 was amended with manure (M); plot 2 was amended with P, K, and sulphorus ammonium (PKS); plots 3-10 were unamended (U). bDuring phase 2, plots 3.4, and 5 survived and were divided into N W ,NE, SW, and SE plots. Manure lime and phosphorus were applied to all south plots. Beginning in 1904and ending in 1919, phosphorus was applied to SW plots as rock phosphate (M with LrP) and to SE plots as bonemeal (M with LbP). Beginning in 1920, P was added as triple superphosphate. CDuring phase 3, plots were subdivided in the east-west direction (A, B, C, D). Nitrogen, phosphorus, and potassium were applied to formerly unamended (U-NPK) and manure amended (M-NPK plots). At this time, the seeding density of all NPK fertilized plots and some manure amended plots (MPS) was increased. phase 4 a high-NPK treatment was applied to some formerly manured plots (H-NPK) and the 2-year C - 0 rotation was changed to a corn-soybean rotation.
158
SUSANNE AREF AND MICHELLE M. WANDER
crop sequence on the C-O-H plots (5-10) in 1897. Plots 6-10 were planted for the last time in 1901. In 1903, almost 30 years after the initial experiment was begun, plots 6-10 were seeded to lawn.
2. Phase 2: 1904-1954, Modification of the Experiment a. Reduction of Size and Number of Plots, Installation of Tiles In 1904, the three remaining plots, numbered 3, 4, and 5, were reduced in the west-east direction to one-fifth-acre plots. All other agricultural or horticultural use of the land around the Morrow Plots had been suspended due to the growth of the university. Additionally, before the 1904 field season, tile lines were installed between the north and south subplots of all rotation treatments (Odell et al., 1984a). b. Introduction of Manure-Lime-PhosphateTreatment A fertility treatment was added to the Morrow Plots to make them reflect agronomic practices of the time (Table I, phase 2). This modification allowed study of the effects of application of manure, lime, and phosphate (Hopkins, 1911; DeTurk, 1938;Stauffer et al., 1940;Lee and Bray, 1949).The three rotation main plots were each divided in fourths in the north-south and wesi-east direction, and a manure, lime, and phosphate treatment was added to the south plots. There was no difference between the two north plots (NW and NE)in each rotation, whereas 600 lbs rock phosphate was applied to the SW plot and 200 lbs steamed bonemeal was applied to the SE plot. Manure was applied before corn was grown. Roughly 4 tons of cured manure were applied annually to the CC plots and applied every other year to the C - 0 plots. Six tons of manure were applied to the C-O-H plots every 3 years, again when corn was grown. Barnyard manure was obtained from the Department of Animal Sciences stockpile and contained mostly (greater than 90%) cattle and some horse or poultry manure. It included some bedding material, since much of the stockpiled waste came from the University’s dairy operation. Phosphate was applied with manure on all plots until 1909. From then on, manure was still applied every year on the CC plots. Phosphate was applied every other year to the CC and C-0 plots and every third year to the C-O-H plots. Initially, limestone application rates were set at 200 lbs per acre, with this amount to be applied every 10 years. This was changed to a regular lime application of 3 tons per acre, applied once every 6 years along with phosphate. Rock phosphate and bonemeal were last applied in 1919 (Odell er al., 1984a). c. Introduction of Hybrids In 1937 corn hybrids were introduced to the plots. Until 1955 several hybrids were used each year. Between 1955 and 1989only one hybrid was seeded at a time
MORROW PLOTS
159
(except in 1964), with varieties being changed about every eight years. The frequency of hybrid replacement increased after 1990.
3. Phase 3: 1955-1966, Addition of N and K Fertility a. Introduction of NPK Treatment Fifty years after the first addition of a fertility treatment, the Morrow Plots were updated again. Commercial N, P, and K fertilizers, then an accepted practice, were added to the plots. Nitrogen was added as urea, P as triple superphosphate, and K as muriate of potash. This modification brought the Morrow Plots in line with agronomic practices of the day. With this experimental alteration it could be determined whether, and how fast, soil fertility was restored to depleted land and how manuring influenced productivity (Bruce, 1955; Russell, 1956; Bartholomew and Kirkham, 1960). The implementation of the new treatment required further division of the plots. They were split in the west-east direction-plot NW became NA and NB, plot NE became NC and ND, SW became SA and SB, and SE became SC and SD (Table I, phase 3). All B plots, with and without a history of manure application, now received the NPK treatment. Manure was no longer added to SB plots. Additionally, crop residues were now returned to the B plots after harvest (Ode11 et al., 1984a). b. Number of Kernels per Hill Seeding density was adjusted to match soil fertility levels. The untreated north sections of all plots had always been planted at a lower density than the manureamended south plots. For the CC plots, the density was 8,000 plants per acre in untreated plots vs 12,000 plants per acre in manured plots. For the two rotations, C0 and C-0-H, all plots had a planting population of 12,000 plants per acre. All fertilized plots were planted at the higher rate of 16,000 plants per acre. One manured plot in each rotation was changed in 1957 to the higher planting density of 16,000 plants per acre, creating a manured-plus (MPS) treatment (Table 11). Alteration of fertility and seeding rates in phase 3 resulted in a total of five treatments (untreated plots, U; manure amended, M; manured with higher seeding density, MPS; and NPK applied to previously untreated plots, U-NPK, and to previously manured plots, M-NPK), which were applied to all crop rotations.
4. Phase 4: 1967-1996, Addition of High Fertility and Substitution of Soybean a. Introduction of High-Level NPK Treatment The so-called green revolution, which occurred in the late 1950s and early 1960s inspired the introduction of a very high-level fertilizer (H-NPK) application in 1967 (Table I, phase 4). This sixth fertility treatment has been studied by several
Table I1 Treatments Applied since 1967 to North and South Regions of the Morrow Plots (amounts in pounds per acre) Plots Location
Treatment
CC rotation North plants per acre N" Pb
KC Manured
B
A
C
8000 -
24,000 200 >45 >336
24,000 300 >112 >560
24,000 200 >45 >336 -
D
-
south plants per acre N P K Manure C - 0 and corn-soybean rotation North plants per acre N P K Manure South plants per acre N
P K Manure C-0-H rotation North plants per acre N P K Manure south plants per acre N P K Manure
-
-
24,000 200 >45 >336 -
24,000 300 >I12 >560 -
24,000 200 >45 >336 -
12,000 -
24,000 200 >45 >336
-
-
24,000 300 >112 >560
24,000 200 >45 >336 -
12,000 -
-
"Nitrogen was applied as urea. hPhosphorus, applied as triple superphosphate since 1920, had been added systematically (40 Ibs per acre) since 1955. In 1967, application was changed to a maintenance-based approach. The NPK and H-NPK plots testing less than 45 or 112, respectively, were amended with 49 or 98 Ibs per acre, respectively, of P. 'Potassium, applied as muriate of potash, was also changed from a systematic (300 Ibs per acre) to a maintenance-based approach. During phase 4, the NPK and H-NPK plots testing lower that 336 and 560 were amended with 93 and 186 Ibs per acre K, respectively. dManure was applied systematically at 4 tons per acre every year in CC plots, 6 tons per acre every other year in C-0 and every thiid year in C-0-H plots between 1904 and 1970. The frequency and quanity of manure applied to the plots decreased between 1970 and 1996.
MORROW PLOTS
161
authors (Jones and Hinesley, 1972; Cescas and Tyner, 1976; Welch, 1976; Omueti and Jones, 1977;Mortvedt, 1986; Jones, 1992).The H-NPK treatment was applied to the SA plots (previously manured plots) in each rotation. In addition, a switch was made from systematic fertilizer application rates to application based upon soil test values. The NPK treatments (B plots) with soil test levels less than 45 or 336 lbs per acre, received 49 and 93 lbs per acre of P and K, respectively. The HNPK treatment (SA) plots with soil test levels less than 112 and 560 received 98 and 186 lbs per acre of P and K, respectively. The planting population in H-NPK plots was increased from 12,000 to 24,000 plants per acre. The planting density in the other NPK plots was increased from 16,000 to 24,000 plants per acre some years later. Beginning in 1967, crop residues were returned to all plots. b. Other Alterations To reflect current rotation practices in the region, the 2-year rotation was changed from a C-0 to a corn-soybean rotation. Also, until 1967, manure was applied on the basis of crop removal. After 1967, wet manure (estimated at 3 0 4 0 % moisture content) was supposed to be applied on a 3-ton per-acre per-year basis. The timing and amount of manure applied has been inconsistent, and records are limited. During the last decade manure has been applied to all rotations once every 3 years during the fall prior to corn production in the C-O-H rotation.
C. RECORDSAND DATAFROM THEMORROW PLOTS The earliest phases of very old experiments are rarely adequately recorded, and the Morrow Plots are no exception. First, though the Morrow Plots have been cultivated since 1876, there are no records of yield before 1888 and no soil samples available for the period before 1904.This means that the first 12 years of yield data are missing and that 28 years passed before any soil samples were taken. 1. Records
Fortunately, the Illinois Agricultural Experiment Station (AES) was started in 1887. This meant that from 1888, data from the Morrow Plots appeared in annual reports published by AES. The trial was then identified as Rotation Experiment 23. Handwritten records in the Morrow Plots field book provide information about the earliest phases of the trial. The first entry in the Morrow Plots field book mentions the observatory, which had been built in 1895. Whoever recounted the period from 1876 to 1912 must have copied some of the earliest records from another source. Morrow recorded the Morrow Plots treatment structure and layout in 1879; his record was copied into the official field book. Field practices and yield have continued to be recorded in field books. No other formalized record-keeping system has evolved.
162
SUSANNE AREF AND MICHELLE M. WANDER
a. Plant Production Variables Crop row widths have been 40 inches. Plants were seeded in hills until 1996. The plots have been plowed, fertilized, and manured mostly in the fall and disked just before planting. Initially, weeds, were controlled by hand pulling and later by application of herbicides. Corn has been harvested by hand and, until recently, was planted by hand with an extra kernel added to each hill to permit thinning. Planting densities were adjusted as described previously (Sections I.B.3.b and I.B.4.a). The yield record began in 1888 and has continued. Extensive weather records were available for the period from 1888to 1996.The Urbana Weather Station was established in 1888 as an independentproject to study air and soil temperatures. This trial, located 700 feet northwest of the Morrow Plots was terminated in 1897. The weather station was moved to a location adjacent to the plots and maintained until 1984. After this, the equipment was moved 2 miles south-southwest of the plots. b. Soil Variables The soil, which is aFlanagan silt loam (fine, montmorillonitic,mesic Aquic Argiudoll) developed under tall-grass prairie in Peoria loess lain over Wisconsin age, calcareous loam glacial till (Ferenbacheret al., 1984; Darmody and Peck, 1997). Soil carbon contents were available for samples collected from west and east plots in 1904, 1911, 1913, 1923, 1944, 1953, 1955, 1967-69, 1973, 1974, and 1980-1 995. Organic carbon was determined by a modified Schollenberger method. The method involves oxidation at 1758C for 90 sec using a solution of concentrated sulfuric acidpotassium dichromate (Schollenberger, 1945).Nitrogen values, believed to be based on macro-Kjeldahl analyses, were available for the same samples through 1973. Soil samples collected in 1969 were analyzed using a LECO CNS-2000 with N calibration based on a NIST coal standard. The N concentrations of samples analyzed in 1968 via Kjeldahl matched combustion-based determinations of samples collected in 1969. The N content of soils collected in 1969, 1973, 1974,1980,1986, and 1992 were then determined with the LECO CNS-2000. Soil pH, P, and K values were assessed beginning in 1967, since the H-NPK was introduced and fertilizer application rates were based on test levels. Soil pH was determined with a glass electrode (l:l, soil-water), P with Bray P1, and K with ammonium acetate extraction and atomic absorption.
2. Statistical Analysis The Morrow Plots experiment lacks the randomization and replication that has long been expected in field experiments. In long-term experiments, original randomization remains in place; hence, spatial differences that exist between plots become apparent over long time periods. With the Morrow Plots’ data, we investigate these differences as well as differences resulting from the applied treatments. The four experimentalphases were treated as a factor. Phase 1 included data col-
MORROW PLOTS
163
lected before 1904. Even though hybrid adoption did not change the experimental focus or design, this shift in practice constituted a large technological advance. Therefore, phase 2 was divided into subphases 2A and 2B to distinguish between periods before and after hybrid introduction. Phase 2A was the period 1904-1936, phase 2B was the period 1937-1954, phase 3 was the period 1955-1966, and phase 4 was the period 1967-1996. Unless otherwise indicated, analyses were performed using SAS proc mixed version 6.1 1 (SAS Institute Inc., SAS Campus Dr., Cary, NC 27513). Year and interactions with year were random effects, and rotation and phase were fixed effects. Other fixed factors (fertility treatment; north and south plots; and A, B, C, and D blocks) varied with models according to experimental question. Yearly averages of variables from all phase, rotation, and treatment combinations were used to assay overall effects on corn yield and soil pH, P, and K. Yield analysis was based on data collected between 1888 and 1996. Soil pH, P, and K analyses were based on data collected annually between 1967 and 1995. Different analyses were carried out to determine the effects of spatial location on corn yield, soil C and N contents, and C-N ratios. Rotation and individual plots were used to analyze corn yield in each phase. To assess spatial impacts on soil C, N, and C-N ratios, plots were analyzed using both north and south direction (which coincides with manure application) and east and west effects. Soil C and N contents and C-N ratios were analyzed using data collected from 0 to 15 cm in 1904, 1911, 1913, 1923, 1933, 1943,1953,1961, 1973,1974,1980, 1986, and 1992. To assess yield stability, mean yield, and standard deviation (s.d.) were considered. Yearly corn yields from each phase, rotation, and treatment combination were averaged. These averages were then used to get means and s.d. for each phase. Mean and s.d. were analyzed as factorials with three fixed factors-rotation, treatment, and phase-using SAS proc glm to determine the behavior of yield and yield variability. Correlation analyses between Morrow Plots’ corn yield and individual weather variables as well as Morrow Plots’ corn yield and Illinois state yield averages were carried out using SAS proc corr.
II. CORNYIELD
A. EXPERIMENTAL EFFECTS 1. Overall Effects, 1888-1996 a. General Trends The complete historical yield record for the Morrow Plots is shown in Fig. 1. Yield of CC plots has always been much lower than yield of C - 0 and C-O-H plots.
High-level NF'K added
NPK added Introduction of hybrids Manure added
OJ
225
I
7
-t
0 1880
High-level NPK added
C-o-H
d
.: 1890
1900
1910
1920
1930
1940
I
1950
1960
1970
1980
1990
2ooo
Year Treatment: o=H-NPK, A=M-NPK, o=MpS,o=M,r=U-NPK, D=U
Figure 1 Morrow Plots yield from 1888 to 1996 by the three rotations: continuous corn (CC), corn-oat (C-O),and corn-oat-hay (C-0-H). Verticallines indicate changes from one phase to the next. Phases were 1888-1903,1904-1936.1937-1954,1955-1966, and 1967-1996. Plots were U. untreated, M, manured; MPS, manured with higher planting density; U-NPK, previously untreated, now NPK treated; MNPK. previously manured, now NPK treated; H-NPK, previously manured, now high-level NPK treated. 164
I65
MORROW PLOTS
Manure addition increased yield. After an initial decrease in corn yield from phase 1 to phase 2, yield was stable until the introduction of hybrids in 1937. Corn yield then increased in the C-0-H rotation. Productivity did not immediately increase after hybrid introduction in the CC or C-0 rotations. Yield in the C - 0 plots started to increase in the late 1940s and in the CC plots in the early 1950s. Benefits of the longer rotation allowed first C-0-H and then C - 0 rotations to utilize the higher yield potential of hybrids. There were large statistical differences between corn yield of the fertilized and manured plots, and even larger differences between yield of the untreated and all other plots. In fertilized plots that had been manured (M-NPK and H-NPK), yield in the C - 0 rotation equaled yield in the C-0-H rotation. Corn yield in U-NPK plots, which were previously untreated, was 11 bu. per acre less in the C - 0 rotation than in the C-0-H rotation. Yield in the CC rotation was 15 bu. per acre less in the U-NPK than in the M-NPK and H-NPK plots. Higher planting density in manured plots increased yield compared to lower density M plots; planting density-based differences in yield were 13, 18, and 25 bu. per acre in CC, C-0, and C0-H plots, respectively. Between 1950 and 1975, which included the years of the green revolution, yield in the untreated C - 0 and C-0-H plots increased linearly. Yield in these plots was extremely stable. Yield in the untreated CC plots and all treated plots (CC, C-0, and C-0-H) also increased at this time but was more variable. Yield in the 1980s and thereafter was highly variable in all plots. During the mid- 1970s yield began to plateau. b. Rotation and Treatment Effects Yield means of all three rotations were significantly different from each other. The difference between the yield of the C - 0 and C-0-H rotations was only half of the difference between the yield of the CC and C-0 rotations. Based on the ANOVA F values, the effect of rotation on yield was only half as large as the effect of treatment on yield (Table In). The NPK treatment effects were not different from Table I11 Morrow Plots Corn Yield, 1888-1996: ANOVA Tests of fixed effects ANOVA Source Phase Rotation Treatment Rotation X treatmenta
NDF
Type 111 F
P>F
4 2 5 10
24.59 108.82 171.60 11.89
0.0001 0.0001 0.0001 0.0001
“‘X’ identifies an interaction effect.
166
SUSANNEAREF AND MICHELLE M. WANDER Table IV Morrow Plots Corn Yield, 1888-1996: Mean Comparison" (in bushels per acre)
Rotation
cc
c-0 C-0-H Phase 1 2A 2B 3 4
All
87.82~ 106.01b 116.08a
U 35.81~ 50.49b 69.15a
M
MPS
64.60~ 88.13b 98.63a
All
Treatment
102.3Sbc 82.26d 96.39~ 111.87b 123.64a
U M MPS U-NPK M-NPK H-NPK
77.39~ 106.22b 123.83a A11
5 I .82d 83.79~ 102.48b 123.19a 128.13a 130.41a
U-NPK
M-NPK
H-NPK
106.22~ 125.85b 136.9%
121.30b 129.12a 133.98a
121.05b 136.24a 133.94a
cc
c-0
35.81e 64.60d 77.39c 106.22b 12 L30a 121.05a
50.49e 88.13d 106.22~ 125.85b 129.12ab 136.24a
C-0-H 69.15d 98.63~ 123.83b 136.95a 133.98a 133.94~1
"Values within groups in columns followed by different letters are significantly different at the 5% level.
each other, but all other treatment differences were highly significant (Table IV). Yield of the U plots was significantly lower than yield of all other treatments. Yield of the M plots was significantly lower than yield of the MPS plots, which was significantly lower than yield of the NPK plots. There was a highly significant interaction between rotation and treatment, but it was only a tenth of the size of the rotation effect on yield. For U, M, MPS, and UNPK, yield increased significantly with longer rotation (Fig. 2). This was not the case for the M-NPK and H-NPK treatments, where yield was the same in C - 0 and C-0-H rotations. The yield of M-NPK and H-NPK plots was not significantly different in each rotation. Yield in the U-NPK plot was significantly lower (by 15 bu. per acre) than yield in M-NPK and H-NPK plots only in the CC system. The difference in yield response associated with the different rotations was due to the residual effect of manure. The H-NPK and M-NPK fertility treatments were applied to plots that were previously manured as opposed to the U-NPK treatment, which was applied to untreated plots. c. Phase Effects The phase yield means were all significantly different except in phases 1 and 2B. Yield decreased significantly from the first phase to the second (Table IV). Soil exhaustion was reflected in declining yield until the introduction of hybrids shifted production upward. Sustained increases in yield in later phases were due to alterations in cultural practices that included the use of hybrids and the addition of fertility.
167
MORROW PLOTS
Treatment: 0
=H-NPK
0
=M
A =M-NPK 0 =MPS
=U
2o 0
< cc
c-0
C-O-H
Rotation Figure 2 Average yield for each treatment and rotation in the Morrow Plots. Averages are leastsquares means from the ANOVA for the whole data set. See Fig. I for abbreviations.
2. Yield by Phases a. Phase 1: Original Experiment Yield reported between 1888 and 1903 was used in the analysis of the first experimental phase of the Morrow Plots (Fig. 3). Comparisons of yield in the CC plots revealed that the untreated and PKS-amended plots were similar and were significantly lower than yield in the manured plot. In the manured CC plot, phase 1 yield was in the same range as yield in the C-O-H plots (all of which were untreated). Statistical contrasts indicated yield differences between the C-O-H rotation plots and the PKS CC, U CC, and C - 0 plots were significant. Yield in the C0 plot fell between the manured CC plot yield and the other two CC plots yield and was not significantly different from either. This early attempt at commercial fertilization did not succeed. Yield response to manure-and not P, K, and S addition-indicated N was already limiting productivity in the plots (Bogue, 1963; Ode11 et al., 1984b).During the soil exhaustion phase of the trial, there was a slight downward trend in yield in all the plots, although the correlation (-0.17) between yield and year was not significant.
SUSANNE AREF AND MICHELLE M. WANDER
168
4
ab 8a
4
4
a .
bc
45.0
40.0
4
4
1 L A
D
35.0 30.0 0 1
D
2 3
4
5
6 7
8 9 10
Plot Number Figure 3 Average yield for each of the 10 plots in phase 1, 1888-1903. Rotations were CC, C-0, and C-0-H. Comparisons were among yield in the three CC plots, the C - 0 plot, and the average of the C-0-H plots.
b. Phase 2: Manure-Lime-Phosphate TREATMENT.Addition of manure i. EFFECTOF MANURE-LIME-PHOSPHATE immediately increased corn production. ANOVA results indicated that the relative importance of treatment and rotation factors was similar in phases 2A and 2B (Table V). Differences between yield in U and M plots were 20 bu. per acre (phase 2A) and 40 bu. per acre (phase 23) (Table VI and Fig. 4).These differences are comparable to the differences between yield in the CC and C-0-H rotations, which
Table V
Analysis of Yield by Phases, ZA,2B, 3, and 4 ANOVA Phase
2B
2A
Source Rotation Plot Rotation Xplot
DF 2 3 6
F
P>F
F
P>F
39.69 0.0001 68.99 0.0001 127.37 O.OOO1 199.29 0.0001 3.85 0.0015 15.05 0.0001
4
3 DF 2 7 14
F
P>F
F
P>F
128.69 0.0001 97.65 O.OOO1 140.48 0.0001 221.00 O.OOO1 7.85 0.0001 8.19 O.OOO1
169
MORROW PLOTS Table VI Analysis of Yield by Phases, 2A, 2B, 3, and 4: Mean Comparisons" (in bushels per acre) Phase 2A
2B
3
4
Rotation
cc c-0 C-0-H Plot Nw NE
sw SE
32.14~ 44.48b 56.30a
41.91~ 63.41b 80.47a
31.80~ 36.61b 52.85a 55.97a
36.72~ 43.93b 81.64a 85.42a
Plot NA
NB NC ND SA
SB
sc SD
77.19c 95.01b 108.09a
90.51~ 117.49b 132.78a
5 1.60e 122.37ab 53.19de 60.50d 103.27~ 128.16a 120.15b 108.20~
74.00d 149.51a 73.60d 78.13d 153.06a 154.15a 121.36b 104.95c
"Values within groups in columns followed by different letters are significantly different at the 5% level.
were 24 bu. per acre (phase 2A) and 40 bu. per acre (phase 2B). Within each phase, there was no significant yield difference between the manured SE and SW plots which had different P sources; however, there was a small but statistically significant yield difference between the untreated NE and NW plots. Since the plots were treated the same, the difference was due to spatial variation. Even though plot differences of 5-7 bu. per acre were observed, the larger differences of 20 and 40 bu. per acre between the north and south (untreated and manured) plots indicated these differences were due to treatment. ii. BONEMEAL AND ROCKPHOSPHATE TREATMENTS.The ANOVA of yield in phases 2A and 2B indicated plots SW and SE were not significantly different, suggesting the source of phosphorous (rock phosphate vs bonemeal) had no effect on crop yield. Note that phosphorous was applied with manure, which had a large effect on yield, preventing separate assessment of the P sources. c. Phase 3: Nitrogen-Lime-Phosphate-Potassium Treatment i. MANURE, ROTATION, AND PLOTDIFFERENCES. Recall that during phase 3 the NPK treatment was applied to the B block and the MPS treatment (increased planting density) was applied to the SC plots. There was a 50 bu. per acre yield
170
SUSANNE AREF AND MICHELLE M. WANDER
Figure 4 Phase and treatment yield means for each rotation. The NA, NC, and ND plots were untreated, the SD plot was manured, the SC plot was manured in all phases and had higher planting density in phases 3 and 4, the NB plot was untreated, and the SB plot was manured in phases 2A and 2B. Both plots received NPK in phases 3 and 4, the SA plot was manured in phases ZA, 2B. and 3 and received high-level NF'K in phase 4. The NA and NB yield was recorded as one NW yield, and the NC and ND yield were recorded as one NE yield in phases 2A and 2B. South plot needs were similar.
difference (not including B nor SC) between north (untreated) and south (manured) plots, which was larger than the 40 bu. per acre difference between north and south plots in phase 2B, indicating manure addition had an additive effect, with benefits to yield increasing with time (Table VI and Fig. 4). The 30 bu. per acre yield difference between the C-0-H and CC rotation observed in phase 3 was less than the 40 bu. per acre yield difference observed in phase 2B, indicating diminished effect of rotation when fertility treatments are considered. For the individual plots with the same treatment, there was a significant yield difference (9 bu. per acre) between the NA and ND plots, exhibiting the same spatially based trend observed in the untreated plots in phases 2A and 2B. The untreated plot yields were ranked: NA 5 NC 5 ND. As in earlier phases, yield in the manure amended SA and SD plots did not differ. Yield in previously untreated OF NPK TREATMENT. ii. IMMEDIATEEFFECT plots increased immediately after introduction of NPK fertilizers. ANOVA of the
MORROW PLOTS
171
1955 data showed that in one year, fertilizer application increased yield in previously untreated plots to the same level as manured plots. Untreated plots yielded an average of 50 bu. per acre, which was significantly less than the mean yield produced by any of the other treatments. Corn yield in fertilized treatments, which was 102 bu. per acre for M-NPK plots, 95 bu. per acre for U-NPK plots, and 94 bu. per acre for M plots, was statistically similar. Yield in the CC plots (75 bu. per acre) was significantly lower than yield of the C - 0 and C-0-H rotations (87 and 93 bu. per acre, respectively). iii. LONG-TERM EFFECTOF NPK TREATMENT.Within a few years, yield differences between M, MPS, U-NPK, and M-NPK fertility treatments became apparent. Yield in the untreated plots remained the lowest overall. The higher planting density in SC plots (MPS) produced significantly higher yield than manured plots with lower densities. Yield was highest in M-NPK, intermediate in U-NPK, and lowest in MPS plots. Only the difference between M-NPK and MPS plots was significant.The planting density in this period was the same for M-NPK and MPS; hence, the observed yield difference of 8 bu. per acre was due either to NPK application or to spatial variation in the plots (Table VI and Fig. 4). d. Phase 4: High-Level NPK Treatment i. COMPARISON OF H-NPK, M-NPK, AND U-NPK. The first year the highlevel soil fertility treatment was applied, associated yield was lower than yield of the other NPK treatments (in all rotations). The second time the H-NPK treatment was applied, which occurred 1,2, and 3 years later in the CC, C-0, and C-0-H rotations, respectively, the associated yield equaled or exceeded yield in the M-NPK and U-NPK plots. After this and until the late 1970s, corn yield in H-NPK plots was higher than M-NPK and U-NPK yield (Fig. 1). This was due to differences in plot planting density. During this time only the H-NPK treatment had a density of 24,000 plants per acre, whereas the M-NPK and U-NPK plots had 16,000 plants per acre. In the late 1970s, planting density in all NF'K plots was raised to 24,000 plants per acre. Over the last 20 years, yield in the H-NPK and M-NPK plots has been similar within rotations. With similar manure histories, the high-level fertility treatment (H-NPK) failed to increase yield over the M-NPK treatment. Despite the initial planting density advantage in H-NPK plots, mean comparisons for the entire phase showed no difference between H-NPK, M-NPK, and U-NPK (Table VI and Fig. 4). In phase 4, the M-NPK and U-NPK plots improved corn production by 27 bu. per acre, whereas the H-NPK brought the yield from the previous M plot from 103 bu. per acre to 153 bu. per acre. ii. IMPACT OF SUBSTITUTION OF SOYBEAN FOR OAT. There was no obvious difference in corn yield before and after the change from oat to soybean. Corn yield increased steadily in all three rotations during phase 3 and during the first part of phase 4 (Fig. I). Note that even though soybean was now in the rotation, we continue to use the C-0 notation to identify the 2-year rotation.
172
SUSANNE AREF AND MICHELLE M. WANDER
MANUREAPPLICATION PRACTICES.In phase 4, iii. IMPACTOF CHANGING yield in M and MPS (higher planting density) plots did not change, whereas yield in all other plots (even in untreated plots) increased significantly. Yield increases ranged between 20 and 27 bu. per acre (Table VI and Fig. 4). In phase 3, yield in the C-0 M and CC U-NPK plots was the same, indicating comparable productivity (Fig. 4; phase statisticsnot shown). Likewise, yield in the C-0-H M, C - 0 MPS, CC M-NPK, C - 0 M-NPK and U-NPK plots was the same. In phase 4, yield in the C-0-H M, C-0 MPS, and C U-NPK plots was the same, as was yield in the C-OH MPS and the CC M-NPK plots. Records indicate that manure application frequency declined in CC and C-0 rotations during phase 4.Nutrient supply may have limited yield in M and M P S plots. Interestingly, the difference between U and M yield declined from 50 bu. per acre in phase 3 to 30 bu. per acre in phase 4. The overall difference between C-0-H and CC plot yield was 40 bu. per acre in phase 4. OF WEATHER.Besides experimental inputs and soil condiiv. SIGNIFICANCE tion, yield was dependent on weather. As mentioned earlier, the H-NPK plots were not more productive than the M-NPK plots. In fact, in very dry years, yield in HNPK plots was very low (Fig. 1). During the 1988 drought, the untreated C-0-H plot yield exceeded the H-NPK plot yield. Osmotic stress associated with high fertilizer rates may have been the cause.
3. Yield Stability Mean yield and sd representingrotation by treatment combinationswithin phases are shown in Fig. 5. Throughout the course of the trial, yield sd has declined and then increased, indicating periods of high and low yield stability. Yield means were discussed in Sections 1I.A. 1 and II.B.2 and are not discussed further here. For reference, results from ANOVA for yield mean are included with those for sd in Tables VII and VIII (on p. 174). Treatments had the greatest effect on yield sd. Rotation effects were only significant through interaction with experimental phase. The phase effect on sd was highly significant; the correspondingmean square was almost twice that of the interaction with rotation. Only the phase and treatment main effects on sd are discussed here. The contributions to stable yield made by adoption of various technologiesduring phases 2B and 3 cannot be separated from the excellent weather conditions that prevailed at this time. The years of the green revolution had an average maximum temperature of 85.2"F and an average precipitation of 4.13 inches in July. This was 0.9"F lower and 0.5 inches higher than the averages of the other years of the Morrow Plots' existence. The variances of these weather variables were also lower during this period than in all other years. Relatively consistent and slightly lower temperatures and higher precipitation in July may be associated with the desirable weather conditions that are generally believed to have occurred during the green revolution years in the Midwest. The re-
173
MORROW PLOTS 180 -r
X
A A
100 80
x A
A
30
f t
:L 20
0
1
3
2A 2B
4
- 0
1
2A 2B
3
4
180 7
. .
120
60 40
0
20 -0
1
-----
--
3
X
4
! . .
n
45 X
ii
C-0-H
.
30
0
2A 2B
-
180 160-140 120 100 80 -60 40..
20 0
ls
0
30 25
I
0
A A
o
la
5u 1
2A
2B
3
4
Phase
Treatment: U=U, I = M , A=U-NPK, k M - N P K , +=MPS, X=H-NPK Figure 5 Treatment effect on phase yield (a) and yield sd (b) in each rotation. Rotations were CC, C-0, and C-0-H.
duction in yield stability during phase 4 was associated with higher planting densities in some plots and by more variable weather. Averaged over years, higher phase means had higher sd. However, residual analyses of the yield data indicated that no relationship existed between yield size
SUSANNE AREF AND MICHELLE M. WANDER
174
Table VII Overall Yield Mean and Standard Deviation:ANOVA Yield mean Source Phase Rotation Phase X rotation Treatment
Yield sd
DF
MS
Fvalue
P >F
MS
Fvalue
P >F
4 2 8 5
2129.55 3597.84
22.51 38.03
O.OOO1 0.0001
5440.75
57.51
0.0001
50.18 16.76 31.10 332.44
4.66 1.56 2.88 30.84
0.0052 0.2289 0.0178 0.0001
and sd. In general, yield s.d. increased with the amount of fertility applied. The sd of plots was ordered: U < M 5 MPS < U-NPK IM-NPK < H-NPK. The very high sd of H-NPK plots yield indicates that not only did this treatment fail to increase yield, it lowered yield stability significantly.
B. SIGNIFICANCE OF TECHNOLOGY, PLANTING DATE, AND WEATHER 1. Impact of Hybrid Introduction Use of hybrids and other technologies resulted in a general increase in yield during phases 2B and 3 (Fig. I). Correlation between year and yield for U and M plots
Table VIII Overall Yield Mean and Standard Deviation Comparisons"(amounts in bushels per acre) Phase
ID 1
2A 2B 3 4
mean 93.95~ 71.56d 94.75~ 109.05b 124.69a
Rotationh sd
ID
24.41a 23.00ab 20.79b 19.50b 24.27a
CC C-0 C-0-H
mean 83.78~ 102.86b 113.36a
Treatment
ID U M MPS U-NPK M-NPK H-NPK
mean 48.94d 84.55~ 101.28b 117.04a 122.19a 126.00a
sd 11.21d 17.25~ 18.85~ 25.59b 25.24b 26.22a
"Values within groups in columns followed by different letters are significantly different at the 5% level. "Rotations were not significantly different for sd.
175
MORROW PLOTS
within rotation was used to statistically corroborate the effects of hybrids. Since hybrid use started in the Morrow Plots in 1937 and the quality of manure was changed in 1967, only data within this period were used in the analyses. All correlations between yield and year, except for the U plot in the C-0-H rotation (a = lo), were highly significant.This indicates that a strong positive linear relationship existed between yield and year. The correlation coefficients between yield and year of U plots in CC, C-0, and C-0-H rotations were 0.62***, 0.70**, and 0.45NS, respectively. Coefficients of M plots in CC, C-0, and C-0-H rotations were 0.71****, 0.76***, and 0.78**, respectively.Although hybrid introduction increased yield both in U and in M plots, the higher correlation between yield and year in M plots demonstrated how fertility magnified the benefits of hybrid adoption. The greater number of yield observations in the CC rotation led to a greater degree of significance associated with a smaller correlation coefficient.
2. Impact of Planting Date During phase 1, planting date varied between 120 and 147 Julian days (Fig. 6). The range in date narrowed during phase 2A from 125 to 140Julian days. In phase 2B the planting dates were progressively later. During subsequent phases, planting occurred earlier and earlier until the end of phase 4. The trend toward very early planting dates was dramatically reversed in the early 1990s. Planting dates have been progressively delayed during the 1990s. A positive relationship existed between planting date and precipitation in May, 160
Introduction Of hybrids
M ~ U E
0
rn'
- .-
T
?
I
-110 105 100
---
I
i
I
NPK added
I
I
I
I
[,-bi I
i
I
SUSANNE AREF AND MICHELLE M. WANDER
176
and a negative relationship existed between planting date and May maximum temperature. Correlations between April weather and planting date were not significant. When planting was not completed by early May, too much moisture and to a lesser degree low temperatures delayed planting. There was no significant correlation between planting date and average yield; however, years with very late planting dates tended to have low yield.
3. Important Weather Components: Correlations with Yield Correlation analysis was used to study the relationship between weather and corn yield. The weather variables considered were temperature, modified growing degree days (GDD), precipitation, the product of temperature by precipitation (TxPFT), and snow fall. Monthly averages of these variables were obtained from the previous fall through the growing season (September to September). During the growing season, highly significant correlations existed between the yearly average of treatment yield means and temperature (negative),GDD (negative), precipitation (positive), and TxPPT (positive) in July and temperature (negative) in August. Though planting date was significantly correlated with May weather, May and June weather variables were not significantlycorrelated with yield. Table IX con-
Table IX Yearly Average Yield Correlations with Weather Variables Using 108 Years of Data (1888-1996) ~
Temperature Maximum January July August Minimum April Growing degree days Total
r
-0.245f -0.347" -0.335" 0.192' 0.224'
"Previous Dec. bPreviousyear's total. 'Previous 2 years' total. dprevious Nov. 'Significant at the 5% level. fsignificant at the 1% level. "Significant at the 0.1%level. hSignificant at the 0.01%level.
Precipitation As rain Decembe? Totalb Total' July As snow January Total
r
0.200' 0.258f 0.236' 0.42Ih 0.246' 0.2 12'
~~
Interaction of precipitation With mean temp. January With minimum temp. Novembef' Totalb July Total
r
-0.200' 0.248f 0.251f 0.43 1 0.3 128
MORROW PLOTS
177
tains only the significant correlations. Similar observations have been made by others (Smith, 1914;Thompson, 1969; Offutt et al., 1987; Dixon et al., 1994). Pre-season monthly weather variables were not as highly correlated with yield as were July and August variables. April minimum temperature was the only spring variable that had a significant correlation (positive) with yield. In January, significant negative correlations existed between average yield and temperature and TxPF’T as well as a positive correlationbetween average yield and snow fall. Highly significant positive correlations existed between yield and the previous year’s November TxPPT and December precipitation. Total yearly GDD, snow fall, and TxPPT were all significantly and positively correlated with yield. The previous year’s totals of precipitation and TxPPT were positively correlated with yield as was total precipitation from 2 years earlier. The correlation coefficients for precipitation totals from 1 and 2 years earlier were of the same magnitude. Although growing season variables are most often considered in yield models, this data shows that weather in the previous winter, fall, and even year can influence yield. Van der Pauw (1966) indicated that not only growing season weather had an effect on yield but also noted the effect of rainfall in previous periods on soil factors.
C. CONNECTION WITH ILLINOIS STATEAVERAGECORNYIELD To use the Morrow Plots as a model of yield trends in Illinois, total yield for the state was compared to Morrow Plots mean yield for each phase-rotation-treatment combination (Fig. 7). In general, there was a positive correlation between these mean yields and corn yield in Illinois (Table X). Treatment correlations varied from phase to phase. We assume the treatments with the highest correlation in each phase best reflect agronomic practices in use in Illinois at that time. During phase 1, when all plots were untreated, only yield in the CC plots was significantly correlated with Illinois yield. In phase 2A, U plot yield in all three rotations was more highly correlated with Illinois yield than was yield in M plots. This changed in phase 2B; at this time yield in U plots was not correlated with Illinois yield. Yield in M CC and C-0-H plots was significantly correlated with production in Illinois. These results suggest that prior to hybrid adoption, production in untreated plots and Illinois farms was similar. By the end of phase 2, Illinois corn production was most like production in manured plots seeded with hybrids. The impact of commercial fertilizer application on yield correlations was assessed by combining data from phases 3 and 4. Between 1955 and 1995, the correlation between NPK-treated plot yield and average Illinois yield became the most significant; these highly significant correlation coefficients ranged between 0.63 1 and 0.89 1. A graph of Illinois and NPK-treated plots’ yield shows how similar production has been in the last 40 years (Fig. 8). Interestingly, yield in M and
225 200
175 150 125
100 75 50 25 0
225
T
200
175
9
150
0
i loo
125
j
P,
3 F
I5 50
25 0
225 -
C-0-H
200 -_
A
Ilit
.*
175 .-
J:
150 ._ 125 .-
loo .15 --
50 -25 -0
I
0
20
40
60
80
100
120
140
160
Illinois Yield (bu. per acre)
Treatment: o=H-NPK, x=M-NPK, +=MPS, x =M, A=U-NPK, w=U Figure 7 Comparison of Morrow Plots yield and Illinois average yield, 1888-1995
180
179
MORROW PLOTS Table X
Correlation between Illinois and Morrow Plots Yield in Each Phase, Treatment and Rotation Correlation coefficients Rotation
cc c-0
C-0-H
Phase 1
2A 2B 3-4h 1 2A
2B 3-4 1 2A 2B 3-4
n
U
16 33 18 41 9 16
0.685" 0.530" -0.064n 0.208P 0.2798 0.623" 0.1679 0.647" 0.3458 0.745d 0.6619 0.8 18'
9 21 4
I1 6 14
M
MPS
U-NPK
M-NPK
H-NPK"
0.386' 0.504' 0.05 18
0.1178
0.63 1'
0.701f
0.734'
0.4 199 0.084"
0.189"
0.8288
0.827f
0.738"
0.732' 0.927" 0.198Y
0.4858
0.891f
0.888)
0.760d
0.4138
"Data from H-NPK plots occurred in phase 4 only. hPhases 3 and 4 were combined. U-NPK and M-NPK correlation coefficients in phase 4 alone were 0.606' and 0.61 1' in CC, 0.743d and 0.783' in C-0. and 0.878' and 0.866d in C-0-H, respectively. "Significant at the 5% level. "Significant at the 1% level. 'Significant at the 0.1% level. 'Significant at the 0.01% level. 8Not significant at the 5% level.
MI'S plots was not significantly correlated with Illinois yield. When phases 3 and 4 were analyzed separately, there were positive but nonsignificant correlations in phase 3. During phase 4, yield stagnated in M and MPS plots, failing to keep pace with increasing state yield. Additionally, yield in U plots was not correlated with Illinois yield in the CC rotation, was highly correlated with yield in the C - 0 rotation, and was very highly correlated with yield in the C-0-H rotation. Even though yield was much lower in U plots, trends in C - 0 and C-0-H rotations were similar to trends in Illinois. In phases 3 and 4, yield of the Morrow Plots NPK treatment has been higher than Illinois yield, with a few exceptions (Fig. 8). In 1980, 1988, and 1995, yield in CC plots was lower than Illinois yield. In 1980 and 1988 drought occurred, and in 1995 planting was delayed by excess precipitation in May, which was followed by dry conditions in June and July. In 1993 yield in the U-NPK CC plot was substantially lower than Illinois yield. The cause for this is unknown; however, it has been suggested that gray squirrels damage may have been greatest in this plot, which is close to a large grassed area.
SUSANNE AREF AND MICHELLE M. WANDER
180 5zz
00Z SLI
OSI W
g 8
521
3 f=
001
B
3 a
SL
5
0s SZ
f
I I
I I
I I
I I
I
I
I
I
I
I
I
I
I I
0
1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000
Year o = Illinois average yield H-NPK
M-NPK
U-NPK
cc C-0
C-0-H Figure 8 Comparison of Morrow Plots yield and Illinois average yield, 1955-1995.
MORROW PLOTS
181
III. SOIL VARIABLES:SOIL ORGANIC MATTER, PH, P, AND K A. SOILORGANIC MATTER 1. Changes in C, N, and C-N
a. Phase Effects The soil C and N contents of the Morrow Plots have declined during the history of the trial (Fig. 9). The C and N record begins with samples collected in 1904 when phase 1 had ended. The initial C and N contents of the plots is unknown. The C and N contents of the adjacent grass boarder, which has never been disturbed, ranged between 39.4 and 26.7 g kg-' in 1990. These values are probably lower than those found in the initial plots (Darmody and Peck, 1997).We used data from 1904 (phase I), 1913,1923,1933 (phase2A), 1943,1953 (phase 2B), 1961(phase3), 1973,1980, 1986, and 1992 (phase 4) to assess overall rotation, phase, and plot effects on soil C, N, and C-N ratios. Statistically significant C losses occurred through all phases of the experiment (Table XI). Significant losses of N occurred between phases 1 and 2A and between phases 2A and 2B. During phase 3, N did not change. Soil N then increased significantly during phase 4 of the experiment. This could have been caused by addition of samples collected from B and D plots (see Section II.A.l.c), by change in N analyses methods (see Section II.A.2.a), andor by soil erosion. Darmody and Peck (1997) recently reported that the plots are now 15 cm lower that the adjacent grass border. While aggradation of borders probably contributes to some of this difference, erosion may be lowering the plow layer into subsurface horizons, which may have higher clay and fixed ammonium contents. Changing soil C and N contents are associated with significant changes in soil C-N ratios. Initially, soil C-N ratios increased; the phase 1 C-N ratio was significantly lower than the phase 3 ratio. During phase 4, C-N ratios decreased significantly, falling below 12.0 for the first time. Decreased C-N ratios could be associated with increased humification of SOM, increased abundance of fixed N (Stevenson, 1994), and/or inconsistent analyses, sampling, or tillage patterns. b. Effect of Rotation After experimental phase, crop rotation has had the most significant effect on soil C and N contents (Fig. 10; Table XI). Average SOM contents were highest in the C-0-H, intermediate in the C-0, and lowest in the CC rotations. Rotation has had no impact on average soil C-N ratios. The positive effect of rotation length on SOM contents has been reported before but to our knowledge has not been analyzed statistically (Ode11 et al., 1984a,b; Darmody and Peck, 1997). The interaction between rotation and experimentalphase will be discussed in Section II.A.2.a.
(a)
35
30
t 1910
(b)
35
r
20
-
15
-
1920
1930
1940
1950
1960
1970
1980
1990
oo -o -o -
c-o ~~
1910 35
1920
1930
1940
1950
1960
1970
1980
1990
-
1910
1920
1930
1940
1950
Year
1960
1970
1980
1990
MORROW PLOTS
183
Table XI
Soil Carbon and Nitrogen Contents and C-N Ratios: 0-15 cm"
Effect
ID
Carbon (g kg-' soil)
Nitrogen (g kg-' soil)
C-N ratio
Phaseh
1 2A 2B 3 4
25.6a 24.4b 22.6~ 21.8d 20.1e
2.06a 1.92b 1.74cd 1.72~ 1.75d
12.42a 12.72ab 12.91b 12.69ab I 1.42~
Rotation
cc c-0 C-0-H
19.2a 23.0b 26%
I .55a 1.84b 2.12c
12.36a 12.46a I2.48a
Direction
North south West East
20.9a 24.9b 20.8a 25.0b
I .68a 2.00b 1.20a 1.98b
12.38a 12.48a 12.23a 12.63b
Plot'
A
18.6a 19.7a 21.6b 235
1.66a 1.76a 1.84ab 1.96b
11.16a 11.22a 11.68b 12.06~
B C D
"Values within groups in columns followed by different letters are significantly different at the 5% level or less. %less specified, data includes values from soil samples collected in 1904, 1911, 1913,1923,1933, 1943, 1953, 1961, 1973,1974,1980, 1986,and 1992from west and east plots only. 'Values in plots A, B, C, and D are from phase 4 (1973, 1974, 1980, 1986, and 1992) only.
c. Spatial Heterogeneity In a nonreplicated trial like the Morrow Plots, care must be taken to investigate and acknowledge inherent variability of the experimental unit. Aware of spatial variability in the field, researchers have discussed C and N contents associated with individual plots, noting rotation and treatment effects on loss or gain occurring over time. Ode11 et al. (1984a) and Darmody and Peck (1997) both reported a west-east gradient in soil C and N contents. Our analysis confirms this trend, indicating C,N,and C-N ratios all increase significantly from western A plots to 4
Figure 9 Trends in Morrow Plots soil carbon contents, 1904-1992. Soils (0-15 cm) were collected from north (N) and south (S) plots on west (A and B) and east (C and D) sides of the field. Rotations were CC. C-0, and C-0-H.
184
SUSANNE AREF AND MICHELLE M. WANDER
Figure 10 Average carbon and nitrogen contents of soils under CC, C - 0 , and C-0-H rotations, during different phases of the Morrow Plots trial.
185
MORROW PLOTS
eastern D plots (Table XI). The wider soil C-N ratio of eastern plots indicates SOM in plots D and C was less humified than SOM in plots A and B. This spatial difference in C-N ratios was also noted by Ode11 et al. (1984a). Fortunately, the spatial gradient in soil C and N contents is perpendicular to, and therefore does not confound, the main rotation treatments. Manure application to the southern half of all plots, which occurred immediately after the soil record began, also created a spatial pattern in soil C and N contents (Tables XI and XU).Manure application increased average C and N contents in S plots relative to N plots but had no effect on soil C-N ratios. Like rotation, effects of manure application were not obscured by the west-east gradient. The west to east increase in SOM complicates the interpretation of phase ef-
Table XI1 Effect of Phase and Manure Application on Soil C,N,and C-N Ratios
cc Phase
North
c-0 south
C-0-H
south
North
North
south
28.7 25.5 23.8 22.7 22.2
29.3 29.0 29.7 28.1 27.5
2.02 2.06 1.99 2.00 1.93
2.25 2.02 1.80 1.79 1.97
2.35 2.29 2.27 2.21 2.40
12.7 12.9 13.0 12.5 11.6
12.7 12.6 13.1 12.6 11.3
12.4 12.7 13.0 12.7 11.4
Carbon (g kg-' soil) 1
2A 2B 3 4
22.5" 19.6 15.3 15.1 14.7
22.2 23.6 20.5 20.6 18.8
25.0 22.2 20.2 19.3 18.5
25.6 26.5 26.3 25.1 22.0
Nitrogen (g kg-' soil) 1
2A 2B 3 4
1.79 1.54 1.26 1.21 1.28
1.90 1.82 1.57 1.56 1.61
2.07 1.77 I .54 1.52 1.59 C-N ratio
1
2A 2B 3 4
12.6 12.7 12.1 12.5 11.5
11.9 12.9 13.0 13.2 11.6
12.3 12.6 13.1 12.7 11.7
"All samples were collected from 0-15 cm. After phase 1, manure, lime, and phosphate were applied to south plots. Only samples collected at least 6 years after phases began were included in phase means analyses. Sample numbers from each phase by treatment and location (north and south) combination varied considerably, preventing meaningful use of least-squaresmeans for mean comparison.
186
SUSANNE AREF AND MICHELLE M. WANDER
fects. Unfortunately, soil samples available from phases 1 through 3 were collected from west and east sides of the field, not from A, B, C, and D plots. Only the specific plot origins of samples collected during phase 4 are known. Phase effects on N were analyzed using only A and C plots to avoid spatial effects. Phase 4 increases in N were still statistically significant (Table XI). If the increases in N were due primarily to the inclusion of samples from the eastern plots, the C-N ratios should have increased and not declined.
2. Trends in C, N, and C-N Ratios a. SOM Interaction between Rotation and Phase Although most of the variation in soil C and N contents was explained by phase and crop rotation, the interaction between these factors was also significant. Figure 10 shows the relative changes in SOM content within rotation and phase and reveals that longer crop rotations lost less C and N. The CC rotation led to the largest and most rapid loss of C and N; overall, this rotation lost 5.9 and 0.39 g of C and N per kg soil, respectively, compared to losses of 4.8 and 0.27 g from the C-0 soil and 4.1 and 0.11 from the C-0-H soil. Moreover, the cumulative loss since 1904 of C was 26.5, 18.9, and 14.1%, and of N was 21.3, 13.3, and 4.8%, in the CC, C-0, and C-0-H soils, respectively. The amount of C and N lost between phases 1 and 2A from all three rotations’ soils was similar (Fig. 10). Losses of C and N occurring between 2A and 2B were significantly larger from the CC than from the C-0 or C-0-H soils. The most rapid loss of SOM from the CC soil coincided with the adoption of corn hybrids. The C contents of the CC plots remained constant during phases 2B and 3 and then decreased significantly again in phase 4. Even though less C was lost from the C - 0 and C-0-H soils than from the CC soils, their C contents declined significantly during all phases of the trial. Like the CC soils, C-0 soil N contents remained unchanged after phase 2B. High levels of above-ground productivity in the CC plots were not reflected by SOM contents; and although soil C losses did cease during phase 3, they accelerated again during phase 4 even as yields continued to increase. The magnitude of C loss during phase 4 from the C - 0 plots was similar to that lost from the CC plots. The return of crop residues to all plots, which began in phase 4, should have curtailed SOM losses. Declines could have been associated with the substitutionof soybean for oat in the C-0 rotation and/or the application of the H-NPK treatment to previously manured plots as well as reduced manure application frequency.As previously noted, overall soil N concentrationsincreased significantlyduring phase 4. This increase in N occurred in CC and C-0-H soils; however, only the N increase in C-0-H soils was statistically significant (Fig. 10). Again, this may have been due to erosion. The C-0-H plot, which lies on the south edge of the field, may have suffered more soil loss than the other two rotation plots. It has the highest elevation and drains, along with all the other plots, toward the
MORROW PLOTS
187
northwest edge of the field (Darmody and Peck, 1997). Another explanation may be the methods used to determine N. Even though combustion-basedanalyses produced average soil N contents that were similar to values obtained by wet oxidation techniques, an interaction may have occurred between SOM contents and N recovery. Work is underway to clarify this matter. b. SOM: Interaction between Treatment and Phase When rotation plots were split in two at the end of phase 1, the average SOM contents of north and south plots were similar (Table XII). The average soil C and N contents of phase 2A manure-amended plots (south) were notably higher than untreated (north) plots. Thereafter, a steady loss of C and N from CC soils was apparent in both north and south plots. Similar losses of SOM from C - 0 or C-0-H soils occurred only in untreated plots. The combination of longer rotation and manure application stabilized or even increased soil C and N contents. During phase 3 and phase 4, southern B plots and southern A plots, respectively, ceased to be manured. After this, means of south plots included data from manured and previously manured subplots. Accordingly, average soil C contents of the southern C-0 and C-0-H plots begin to decline in phases 3 and 4. The C-N ratios of north and south plots remained similar during all phases of the trial, suggesting manure application did not have a systematic effect on SOM composition.
B. PH, P,AND K 1. pH All pH, P, and K values are based on samples collected during phase 4 of the experiment. Both rotation and fertility treatment had significant effects on soil pH (Table XIII). The average soil pH of the CC plots was significantly lower than that of the C - 0 and C-0-H plots. Overall, the pH of untreated soil, which was similar in all three rotations, was significantly lower that the pH of all amended soils. In addition, the pH of the M soils was significantly higher than the pH of the M-NPK amended soils. The acidifying effect of fertilizer application was most expressed in the CC H-NPK plots. These soils had an average pH value of 6.01 compared to values of 6.32 and 6.37 in comparably fertilized C - 0 and C-0-H plots.
2. Phosphorus Fertility treatments had the greatest impact on soil P content (Table XIII). The H-NPK soils’ P levels were the highest, followed by the M and M-NPK, MPS, and U-NPK, and U soils. Soil P levels decreased with increasing rotation length (CC > C - 0 > C-0-H), reflecting the frequency of fertilizer application. One exception
188
SUSANNE AREF AND MICHELLE M. WANDER Table XIII Treatment Effect on Soil pH, P, and K Levels during Phase 4“ Soil Fertility Treatment
All rotations
cc
c-0
6.12A” 6.01ad 6.36~ 6.23bc 6.50d 5.40cd 6.19b
6.20B 6.32~ 6.3% 6.20b 6.51d 5.55e 6.29bc
C-0-H
~~
All treatments H-NPK M M-NPK MPS U U-NPK
6.24A“ 6.39B 6.24A 6.47B 5.48C 6.23A
6.20B 6.37~ 6.45cd 6.29bc 6.39cd 5.50e 6.22b
P (Ibs per acre) All treatments H-NPK M M-NPK MPS U U-NPK
97.9A 56.5B 56.9B 48.2C 11.3D 48.4C
63.OA 104.5a 89.9b 54.od 65% 13.5f 50.7de
49.9B 100.5a 42.4e 50.5de 43.oe 13.0f 50.7d
46.7C 88.7b 37.2e 66.0~ 36.0e 8.1f 43.9e
K (Ibs per acre) All treatments H-NPK M M-NPK MPS U U-NPK
352.4A 270.OB 277.4B 257.2C 214.7D 269.1BC
297.9A 388.3a 312b 290cd 289d 223.7h 283.0d
281.8B 373.6a 274.lde 293.3cd 258.3f 204.7j 286.6de
240.8C 295.6cd 223.2h 248.lfg 224.4h 215.6i 237.8gh
Values are means from samples collected (0-15 cm) annually between 1969 and 1995. ”Means for rotation treatments within “All Treatments” row that are followed by different capitalized letters are significantly different at the 5% level. CMeansfor fertility treatmentswithin “All Rotations”column that are followed by different capital letters are significantly different at the 5% level. w e a n s for treatment by rotation interactions that are followed by different lowercase letter within row or column are significantly differentat the 5% level.
MORROW PLOTS
189
to this ranking of rotations resulted from their interaction with the M-NPK fertility treatment; P levels were significantly higher in the C-0-H than in the C - 0 and CC rotation soils.
3. Potassium Soil K levels were influenced most by soil fertility treatments (Table XIII). As expected, the H-NPK treatment significantly increased K test levels. The IS contents of soils amended with manure and/or NPK were similar. The average K levels of unamended soils was 2 I5 lbs per acre, a rather high value for soils cropped continuously for over a century. Crop rotation had the same effect on K as it had on P; soil K contents decreased in order of increasing rotation length (CC > C-0 > C-0-H).
C. SOM: INTERACTION BETWEEN CORNYIELD AND SOILFERTILITY 1. Yield and SOM A century of data from the Morrow Plots reveals that corn yield has increased as SOM contents have declined; this does not indicate that SOM has a negative impact on yield. Throughout the experiment, yield in plots with higher SOM contents has been higher than yield in plots with low SOM contents. Yield trends have shifted upward in concert with technology adoption (hybrids, pesticides, and commercial fertilizers). Technology-based increases in yield potential can occur despite losses in the soils’ inherent productive capacity. Cassman and Pingali (1996) argue that diminishing soil productive capacity can reduce yield increases caused by improved technology adoption. To assess the effect of Morrow Plots fertility practices on soil productivity, we considered the changes in corn yield and soil C contents associated with the phases of the trial (Fig. 11). Both the loss of fertility and the benefits of technology were manifest in U plot yield and soil C trends. Initially, losses in SOM, and therefore soil nutrients, were reflected by decreased yield. We speculate that introduction of hybrids and improved disease, weed, and insect control led to later increases in the yield of U plots. After this, corn yield increased despite continually declining SOM contents. Yield during phase 2A, which was already higher, increased markedly in C0 - H but not in CC or C-0 plots. During phase 3, the lagging yield response was overcome to some degree by untreated CC and C-0 plots. During phase 4,average yield in untreated C-0 and C-0-H increased, while yield in CC plots did not. This may indicate that, once again, soil productivity is limiting yield in the U CC plots.
17 SUSANNE AREF AND MICHELLE M. WANDER
190
b. Manured Plots
a. Untreated Plots
b
150
4 4
150
&
I 7 l
h
B a
125
t
125
.4
0 4
.
e4
loo
'
75
'
50
1.50
1.75
2.00
2.25 U
cc c-0 C4-H
0
U-NPK
w v 0
2.50
2.75
3.00
1.50
I
l
l
1
1
1
1.75
2.00
2.25
2.50
2.75
3.00
M
WH-NPK
Soil Carbon Content (%)
c c o C - O V C-O-H
0
v
MP
0 8
0
Figure 11 Changes in corn yield and soil carbon contents by phase. (a) Untreated and previously untreated, now NPK treated plots; (b) manured, manured with higher planting density, and previously manured, now NPK and H-NPK treated plots. Numbers adjacent to symbols identify means from the four phases of the trial.
Manure amendment led to immediate increases in SOM content and to yield increases in both the C - 0 and C-0-H rotations. Initially, corn yield did not respond in the CC plots, indicating factors other than nutrition limited productivity in this rotation. During phase 2, soil C contents dropped or remained the same in CC, C0, and C-0-H plots as yield increased. Changes in soil C contents, which may be explained by altered manuring practices, were not reflected by decreasing yield response. During phase 3, yield continued to increase and higher planting densities (MPS) increased relative yield response. There was no change in yield during phase 4 in M and ME'S plots. At this time, soil data indicated SOM increased in M and decreased in MPS plots, Changes in SOM contents probably reflected plot effects more than anything else, since M plots were on the east and the MPS plots were on the west side of the field. Fertilizer application led to greater phase 3 yield response in U-NPK plots than in M-NPK plots. This was most notable in the CC rotation, which was the most nutrient stressed. During phase 4, CC yield response lagged behind response in the
MORROW PLOTS
191
C - 0 and C-0-H rotations in U-NPK, M-NPK, and H-NPK plots. It is possible that the relatively low SOM contents of CC plots hindered yield response in that rotation. Although not significantly different, the H-NPK treatment was associated with the greatest soil C losses.
2. Yield and pH, P, and K Information about soil pH, P, and K was only available for phase 4 of the trial. In general, low pH, P, and K test values were associated with low yield. However, differences in yield of the three rotations were not explained by overall differences in soil chemistry. The pH of the C - 0 and C-0-H rotations was higher, while P and K contents were lower. We conclude that higher yield in longer rotations was tied to SOM-dependent benefits and not mineral nutrition.
Tv. CONCLUSIONS: LESSONS FROM THE MORROW PLOTS The changing nature of the Morrow Plots’ treatments reflects the agricultural history of the Corn Belt. During the existence of the plots, the dominant agricultural practices in the region have evolved from low-intensity systems that exploited highly fertile virgin soils, to systems that relied on rotation, manure, lime, and phosphorus application, to more intensive systems with simplified rotations that include hybrids grown in higher densities with increased commercial inputs. Substantive changes in the agricultural norms of the region have been mirrored by changes in the Morrow Plots. The treatments that were once focal, now serve as invaluable controls. The continuous corn treatment, which began as the most extreme treatment in the soil exhaustion trial, was compared with what was then considered to be the best management practice, crop rotation. Now it is the manured C-0-H rotation that is an extreme treatment, returning much more organic matter and nutrients to the soil than the mineral fertilized CC or corn-soybean systems. Initially Morrow Plots researchers were concerned about the depletion of native soil fertility. In the mid-to-late 1800s Jethro Tull and others observed sustained crop yield on newly plowed soils. This led the U.S. Bureau of Soils to promulgate the theory that “practically all soils contain sufficient plant food for good crop yields [and that] . . . this [nutrient] supply will be indefinitely maintained” (Hopkins, 1906). University of Illinois scientists argued vehemently against this assertion (Davenport, 1908). Results from phase 1 of the Morrow Plots and other experiments had already proved that soil fertility, N in particular, was limiting production in Illinois (Bogue, 1963). During phase 1 of the trial, corn yield and SOM contents were directly correlated and both were decreasing.
192
SUSANNE AREF AND MICHELLE M. WANDER
At this time, researchers wanted to determine whether and how productivity could be maintained on a permanent basis. The fertility treatment that was added during phase 2 of the experiment included the key elements of the Illinois System of Permanent Soil Fertility. Although not an absolute system, Illinois scientists then recommended the use of raw rock phosphate, crushed limestone, nitrogen from legume crops, and the liberation of K and other mineral nutrients from soils through biological processes (Smith, 1925). The phosphate trial that was applied to the west and east sides of the plots did not indicate that yield was influenced by source of P (rock phosphate or bonemeal). However, P treatments were only maintained from 1904 to 1912, and their effects on yield were overshadowed by the dramatic effect on yield caused by the accompanying manure and lime application. In phase 2, the manure, lime, and phosphate treatment immediately increased corn yield in all rotations. During this phase, manure treatment slowed SOM losses in the CC and C-0 rotations and halted SOM losses in the C-O-H rotation. Corn yield in the manured CC plots has generally been comparable to yield in the untreated CO-H plots, indicating that manure application or rotation alone can sustain reasonable levels (corn yield greater than 100 bu. per acre) of productivity long term. When hybrids were introduced during phase 2, corn yield and soil productivity were, to a degree, decoupled. The dramatic benefits of improved technologies were demonstrated as yield, which was still the highest in the C-O-H and lowest in the CC rotation, increased dramatically in all plots. Differences in timing of yield response exhibited by the three rotations reflected the nonnutrient-based benefits of higher SOM contents. Researchers now wondered about the limits to productivity. During phases 3 and 4 they would seek to determine whether application of commercial fertilizers and amendments and the intensified planting of improved hybrids would not only sustain, but also increase, productivity thresholds. Application of commercial fertilizers immediately increased yield in all three rotations. Soil condition continued to influence yield, but the effect was not as dramatic as the effect of fertilizers. Soil properties did significantly influence yield potential in continuous corn. During phase 4,the CC U-NPK plots yielded 12 bu. per acre less than the previously manured CC H-NPK plots and 14 bu. per acre less than CC M-NPK plots. In C-0 and C-O-H rotations, yield in U-NPK, M-NPK, and H-NPK plots was similar, indicating that rotation alone improved soil condition sufficiently to allow the full benefit of fertilization to be realized. The higher planting densities that were adopted during phases 3 and 4 increased yield when fertility was adequate. However, higher densities were associated with reduced yield stability when factors like bad weather limited productivity. Corn was planted as early as possible, a strategy that was most successful during the 1970sand 1980s. In the 1990s,years with very wet springs delayed planting. When planting was not completed by early May, the delay was generally caused by too much moisture and/or low temperatures. Weather in July and August was very
MORROW PLOTS
193
highly correlated with yield. Weather from the previous fall and winter also affected yield. The very high levels of fertilizer introduced in 1967 did not improve production and reduced yield stability. Although the H-NPK treatment is no longer seen as a relevant treatment, it serves as an extreme. Results show that in overfertilized systems, yield is not increased and is less stable than plots fertilized in accordance with university recommendations. The negative environmental and economic costs of the H-NPK treatment make its continuation undesirable. Ironically the findings of these plots, which were initially seen as too academic, have not been published widely in academic journals; they have, however, been used frequently in anecdotes. Despite the fact that next to nothing has been reported on Morrow Plots’ yield, production trends have been used as a sort of benchmark, reflecting conditions in the state. The early correlation between U plot yield and Illinois yield coincided with a period of soil mining. It was not until 1937 that yield in manured plots was more highly correlated with Illinois yield than yield in untreated plots. This may indicate that it took a while before there was widespread adoption of rotation and/or use of inputs like manure and lime. The high correlation between Illinois yield and yield in the NPK plots during the last 40 years suggests that these plots reflect state norm. Special attention should be paid to the CC and corn-soybean rotation, because these systems are so widely used. The U-NPK fertilized plots may best reflect field conditions in the region. Declining soil condition is indicated by production trends; during phase 4, yield in U-NPK CC and C-0 plots was 41 and 15 bu. per acre less than yield in the U-NPK C-0-H plot. Yield in the fertilized corn-soybean rotation may have exceeded yield in comparable CC plots because of immediate rotation effects or because of its history of rotation with oats. The long-term effect of the switch to soybean from oat in the C-0 rotation will not become clear until more time has passed. Decreases in soil C contents accelerated during phase 4 in both the CC and C- 0 rotations despite the return of residues to all plots during this phase. Reduction in the frequency of manure application to the M plots may have been a contributing factor. The questions asked during the various phases of the trial indicate how issues and expectations in agriculture evolve. When viewed as a whole, the story chronicled by the Morrow Plots is that of the effect of management practices and technological innovations on corn yield and SOM. Fertilizer application and pest control measures have increased corn yields. In all but the longest rotation, SOM levels continue to fall. However, declining inherent productivity has not been noticed; even in the most SOM-depleted soils, technological innovations have continued to increase yield. Despite the fact that yield responses have been greater where SOM is conserved, long crop rotations and manure are not widely used. When crop yield is the sole factor considered, use of these kinds of soil-building practices may not be competitive. If the relationship between SOM and soil quality, which includes soil’s ability to regulate water flow and/or its ability to act as
194
SUSANNEAREF AND MICHELLE M. WANDER
an environmental filter, is considered, maintenance of organic matter and all it represents may become an imperative. Changes in the Morrow Plots’ treatments have provided an empirical basis upon which we can evaluate the various phases in our agricultural history. Findings have shown that production is greater where soil condition is maintained, that use of mineral fertilizers have off-set to a large extent losses in soil productivity, and that excessive use of mineral fertilizers is undesirable from a production standpoint alone.
APPENDIX: ABBREVIATIONS
cc c-0 C-0-H NPK PKS Phase 1 Phase 2A
Phase 2B Phase 3
Phase 4 Nw NA NB
NE NC ND
sw
SA SB
Continuous Corn Corn-oat rotation Corn-oat-hay rotation N, P, and K treatment P, K, and sulphorous ammonium treatment 1876-1903, original 10-plot experiment (Experiment 23). 1904-1936, comparison of manure, lime, and phosphorous prior to hybrids was introduced; plots were split in the north-south direction for manure application and in the west-east direction for source of phosphorous 1937-1954, comparison of manure, lime, and phosphorous after hybrids was introduced 1955-1966, fertilizers and higher planting density introduced; plots were split again in the west-east direction, with B plots receiving fertilizers 1967-1996, H-NPK added to a previously manure plot (SA) Northwest subplot untreated from 1876; split into NA and NB in 1955 Western sub-subplot remained untreated after 1955 Easter sub-subplot; NPK added in 1955 Northeast subplot untreated from 1876; split into NC and ND in 1955 Western sub-subplot remained untreated after 1955 Eastern sub-subplot remained untreated after 1955 Southwest subplot manured from 1904; split into SA and SB in 1955 Western sub-subplot manured 1955-1967; received highlevel NPK from 1967 Eastern sub-subplot; NPK added in 1955
MORROW PLOTS
SE
sc SD SOM U M MPS U-NPK M-NPK H-NPK
195
Southeast subplot manured from 1904; split into SC and SD in 1955 Western sub-subplot remained manured, with higher planting density after 1955 Eastern sub-subplot remained manured after 1955 Soil organic matter Untreated plots Manure-treatedplots Manured plots with higher planting density NPK treatment applied to previously untreated plots NPK treatment applied to previously manured plots High-level NPK treatment applied to previously manured plots
ACKNOWLEDGMENTS We thank Ted Peck for his invaluable assistance. He has devoted a great deal of effort to the preservation of the Morrow Plots records and samples. Without these materials, this paper could not have been written. Dr. Peck also provided insights into the history of the Plots that would have otherwise been missed. Additionally, we thank Bob Dunker for providing yield records and information about plot management and Bob Darmody for comments on the first draft. We thank Audrey Bryan and Wayne Wendland of the Illinois State Water Survey for providing us the weather data. Finally, we acknowledge Xueming Yang for his work on the soil sample inventory and his assistance in the laboratory.
REFERENCES Barnett, V., Payne, R., and Steiner, R., (eds.) (1995). “Agricultural Sustainability, Economic, Environmental, and Statistical Considerations.” Wiley & Sons, London. Bartholomew, W. V., and Kirkham, D. (1960). Mathematical Descriptions and Interpretations of Culture Induced Soil Nitrogen Changes, pp. 471477. “Proc. 7% Int. Cong. of Soil Sci.,” Madison, Wl. Bogue, A. G. (1963). “From Prairie to Corn Belt: Farming on the Illinois and Iowa Prairies in the Nineteenth Century.” Univ. of Chicago Press, Chicago, IL. Bruce, R. R. (1955). An instrument for the determination of soil compactibility. Soil Sci. SOC.Am. Proc. 19,253-257. Cassman, K. G., and Pingali, P. L. (1995). Extrapolating trends from long-term experiments to farrners’ fields: The case of irrigated rice systems in Asia. In “Agricultural Sustainability, Economic, Environmental, and Statistical Considerations,” (V. Barnett, R. Payne, and R. Steiner, eds.), pp. 63-84. Wiley & Sons, London. Cescas, M. P., and Tyner, E. H. (1976). Rate of rock phosphate disappearance for the Morrow Plots. Ann. Agron. 27,891-924. Darmody, R. G., and Peck, T. R. (1997). Soil organic carbon changes through time at the University of Illinois Morrow Plots. In “Soil Organic Matter in Temperate Agroecosystems: Long-Term Experiments in North America” (E. A. Paul, K. Paustion, E. T. Elliott, and C. V. Cole, eds.), pp. 161-169. CRC Press, Boca Raton, FL.
196
SUSANNE AREF AND MICHELLE M. WANDER
Davenport, E. (1908). “The Status of Soil Fertility Investigations.” Univ. of Illinois, Agric. Exp. Sta. Circ. 123. Urbana, IL. DeTurk, E. E., Bauer, F. C., and Smith, L. H. (1927). “Lessons from the Morrow Plots.” Univ. of Illinois, Agric. Exp. Sta. Bull. 300. Urbana, IL. DeTurk. E. E. (1938). Changes in the soil of the Morrow Plots which have accompanied long-continuous cropping. Soil Sci. SOC. Proc. 3,83-84. Dixon, B. L., Hollinger, S. E., Garcia, P. and Tirupattur, V. (1994). Estimating corn yield response models to predict impacts of climate change. J. Agric. Res. Econ. 19(I), 58-68. Hopkins, C. G. (1906). “The Duty of Chemistry to Agriculture,” Univ. of Illinois, Agric. Exp. Sta., Circ. 105. Urbana, IL. Hopkins, C. G., Readhimer, J. E., and Eckhardt, W. G. (1908). “Thirty Years of Crop Rotations on the Common Prairie Soil of Illinois,” Univ. of Illinois, Agric. Exp. Sta., Bull. 125. Urbana, IL. Hopkins, C. G. (1911). “Methods and Results of Ten Years Soil Investigations in Illinois.” Univ. of Illinois, Agric. Exp. Sta., Circ., 149. Urbana, IL. Jones, R. L., and Hinesley, T. D. (1972). Total mercury content in Morrow Plot soils over a period of 63 years. Soil Sci. SOC.Am. Proc. 36,921-923. Jones, R. L. (1992). Uranium and phosphorus content in Morrow Plot soils over 82 years. Commun. Soil Sci. Plant Anal. 23,67-73. Lee, C. K., and Bray, R. H. (1949). Organic matter and nitrogen content of soils as influenced by management. Soil Sci. 68,203-212. Mitchell, C. C., Westerman, R. L., Brown, J. R., and Peck, T. R. (1991). Overview of long-term agronomic research. Agron. J. 8 3 , 2 4 2 9 . Mortvedt, J. J. (1986). Cadmium levels in soils and plant tissues from long-term soil fertility experiments in the United States. Trans. 13th Intl. Sac. ofsoil Sci. 870-871. O d d , R. T., Melsted, S. W., and Walker, W. M. (1984). Changes in organic carbon and nitrogen of Morrow Plot soils under different treatments, 1904-1973. Soil Sci. 137, 160-171. Odell, R. T., Walker, W. M., Boone, L. V., and Oldham, M. G. (1984). “The Morrow Plots: A Century of Learning.” Univ. of Illinois, Agric. Exp. Sta., Bull. 775 Urbana, IL. Offutt, S. E., Garcia, P., and Pinar, M. (1987). Technological advance, weather, and crop yield behavior. N . Cen. J. Agric. Econ. 9 , 4 9 4 3 . Omueti, J. A. I., and Jones, R. L. (1977). Fluorine content of soil from Morrow Plots over a period of 67 years. Soil Sci. SOC.Am. J. 41,1023-1024. Paul, E. A., Paustion, K., Elliott, E. T., and Cole, C. V. (1997). “Soil Organic Matter in Temperate Agroecosystems: Long-Term Experiments in North America,” CRC Press, Boca Raton, FL,. Peck, T. R. (1989). Morrow Plots: Long-Term University of Illinois field research plots, 1876 to present. In “Proc. of the Sanborn Centennial,” pp. 49-53. Russell, M. B. (1956). All the way back in one year. Plant Food Rev. 2,18-19. Silver, C. W., (1875). Abstract of the results of the field experiments by Lawes and Gilbert, Rothamsted, England. Illini. IV(5). 129. Smith, J. W. (1914). The effect of weather upon the yield of corn. Monrhly Weather Rev. 42,78-87. Smith, L. H. (1925). The Illinois system of permanent soil fertility in the light of twenty-five years of investigation. Univ. of Illinois Agric. Exp. Sta., Circ. 289. Urbana, IL. Stauffer, R. S., Muckenhim, R. J., and Odell, R. T. (1940). Organic carbon, pH and aggregation of the soil of the Morrow Plots as affected by type of cropping and manurial addition. J. Am. SOC.Agron. 32,819-832. Steiner, R. A. (1995). Long-term experiments and their choice for the research study. “Agricultural Sustainability: Economic, Environmental and Statistical Considerations” (V. Barnett, R. Pagne, and R. Steiner, eds.), pp. 15-22. Wiley & Sons, Chichester, England. Stevenson, F. J. (1994). “Humus Chemistry: Genesis, Composition, Reactions,” 2nd ed. Wiley & Sons, New York.
MORROW PLOTS
197
Thompson,L. M. (1969). Weather and technology in the production of corn in the U.S. corn belt. Agron. J. 61,453-456. Van der Paauw, F. (1966). Role of the organic cycle in fluctuations of crop yield. In “Soil Chemistry and Fertility,” (G. V. Jacks, ed.), pp. 125-129. International Society of Soil Science. Communications I1 & IV. Aberdeen. Welch, L. F. (1976). The Morrow Plots-100 years of research. Ann. Agron. 27,881-890. Welch, L. F., Melsted, S. W., and Oldham, M. G. (1976). Lessons from the Morrow Plots. Illinois Research, 1 8 , 3 4 .
This Page Intentionally Left Blank
USINGGENOTYPE-BY-ENVIRONMENT INTERACTION FOR CROPCULTIVAR DEVELOPMENT Manjit S. Kang Department of Agronomy Louisiana Agricultural Experiment Station Louisiana State University Agncultural Center Baton Rouge, Louisiana 70803-2 110
I. Introduction h Genotype, Environment, and Phenotype B. Environmental Influence on Heredity C. Genotype-by-Environment Interaction II. Implications of GE Interaction in Breeding A. Impact on Breeding B. Multienvironment Performance Evaluation 111. Causes of GE Interaction A. Environmental Stress on Genome B. Biotic Stresses C. Abiotic Stresses D. Phenotypic Plasticity A? Ways of Dealing with G E hteraction V. Stability Statistics: Concepts and Usefulness A. Static vs Dynamic B. Types of Stability C. Stability Statistics D. Simultaneous Selection for Yield and Stability E. Merits of Emphasizing Stability during Selection F. Contribution of Environmental Variables to Stability G. Stability Variance for Unbalanced Data VI. How to Exploit or Minimize Interaction A. Breeding for Resistance-Tolerance to Stress Factors B. Breeding for Stability-Reliability of Performance C. Measure Interaction a t Intermediate Growth Stages D. Early Multienvironment Testing E. Optimal Resource Allocation VII. Conclusions References
199 Advancer in Aronmv. Volume 62 " < Copyright 0 1998 by Academic Press. All rights of reproducdon in any form resewed. 0065-2113/98 $25.00
2 00
MANJIT S. KANG
I. INTRODUCTION Breeders and geneticists continually strive to broaden the genetic base of crop species to prevent problems associated with genetic vulnerability. With emphasis on broadening the genetic base and unpredictable climatic factors encountered at different sites and/or years, differential responses are expected of improved cultivars-strains in different environments. These differential genotypic responses to different environments are collectively called genotype-by-environment(GE) interaction. Genotype-by-environmentinteraction is a universal issue that relates to all living organisms, from humans to plants to bacteria. There are many facets to GE interaction; voluminous literature on the issue exists (Annicchiarico and Perenzin, 1994; Byrne et al., 1987; Dean, 1995; Helms, 1993; Ottaviano ef al., 1991; Van Oosterom et al., 1993) . Published works on this ubiquitous topic, not only as journal articles but also as conference-symposia proceedings and books from around the world, attest to its importance (Kang, 1990;Gauch, 1992;Rao etal., 1993;Kang and Gauch, 1996; Cooper and Hammer, 1996). In the past, emphasis has been on statistical measurements of differential performance of genotypes in different environments and on statistical methodologies to characterize genotypes as stable (consistent performance) or unstable (inconsistent performance) across environments. However, this is akin to treating the symptoms, not the causes of the problem. The relationship between statistics and GE interaction can be viewed as that between a lamppost and a drunk,it is for support, not illumination. Caligari (1993) appositely pointed out: “It is an area that can easily be dominated by statistics, which while being an essential element, must be seen as a tool rather than an end in itself. If viewed in a more integrated way, G x E is an area which perhaps needs some interesting new insights which could provide some fascinating opportunities.” In my opinion, GE interaction is primarily a crop breeding issue, not strictly a biometrical one. Crop breeders are interested in knowing how much of the selection progress achieved in one environment can be carried over to another environment. I intend to treat the subject from the standpoint of crop breeding and genetics. Crop breeding is the enterprise of providing genetic solutions to impaired plant productivity that arises from changes in climatic and edaphic factors, the altered spectrum of pests, changes in economic and consumer demands, and government policies (Scowcroft, 1988). Plant breeding is both an art and a science (Jensen, 1983).Art refers to personaljudgments and decisions made by a researcher, whereas science includes knowledge and application of genetic principles, biochemistry, plant pathology, soils, crop ecology and physiology, and statistics. The role of a crop improvement program is to develop high-yielding, profitable cultivars for sustainable production in target areas by managing genetic variability and generating new genetic combinations. A successful cultivar must possess various desired traits-high economic yield, desired value-added traits, and resistance-tolerance to various environmental stresses and pests.
USING GE INTERACTION FOR CROP DEVELOPMENT
201
Since 1990, GE interaction has received focused attention from crop breeders, geneticists, production agronomists,and biometricians (Kang, 1990; Gauch, 1992; Rao et al., 1993; Prabhakaran and Jain, 1994; Kang and Gauch, 1996; Cooper and Hammer, 1996).Earlier significant works on this issue are those of Matsuo (1975) and Byth and Mungomery (1981). Statistical aspects of stability of crop yield were discussed at a symposium (Rao et al., 1988). In addition, bibliographies on GE interaction have been provided by Crossa (1990), Aastveit and Mejza (1992), Denis et al. (1996), and Denis and Gower (1996). The bibliography of this chapter, though extensive, is not exhaustive. To better understand the issue, I begin with definitions of genotype, environment, and phenotype, and I discuss the influence of environment on heredity that may result in modification of a phenotype and lead to GE interaction. Some terms from ecological genetics will be used and explained.
A. GENOTYPE, ENVIRONMENT, AND PHENOTYPE 1. Genotype
Genotype (G) refers to an individual’s genetic makeup. It is the nucleotide sequences of DNA (a gene or genes) that are transmitted from parents to offspring. A gene may be defined as a segment of chromosome (DNA) that is sufficiently short for it to last long enough to function as a significant unit of artificial selection (Dawkins, 1978;Baker, 1984).Genotype may refer to one gene locus (bb,Bb, BB), two gene loci (aabb, aaBb, aaBB, etc.), or multiple genes (AABBCC,aabbccdd, AaBbCcDdEe, etc.). Genotypes may be characterized as homozygous (bb, aaBB, AABBCC, etc.), heterozygous (Bb, AaBb, AaBbCc, etc.), or hemizygous (a haploid situation, e.g., A, a, B, b, AB, Ab, aB, ab, etc.). A genotype by itself is of no consequence, because nucleotide sequences may be artificially produced and stored in a test tube in a cooler without any change.
2. Phenotype and Environment Phenotype (P) refers to physical appearance or discernible traits of an individual, which may be observable at a physical, morphological, anatomical, or biochemical level. It is dependent on expression of a genotype in an environment (E). Environment may be defined as the total of circumstances surrounding an organism or a group of organisms. Phenotype may be expressed as tall, short, red, liguleless, prolineless, etc. Identical or similar phenotypes do not necessarily breed true (for example, phenotypically Bb and BB may not be distinguishable). Thus, the genotype of an individual does not unambiguously determine its phenotype. The genotype may specify a range of phenotypic expressions that are called the norm of reaction (Redei, 1982; Brandon, 1990). Therefore, genotype and environment
2 02
MANJIT S. KANG
are the two main ingredients that make up a phenotype. Gene expression is environmentally induced and regulated. The gene product is a polypeptide chain (protein that may have enzymaticfunctions). If an enzyme's activity (phenotype or trait of interest) is environment sensitive, norms of reaction (or an array of phenotypes) are observed. Environmental factors may be available in optimal, suboptimal, or superoptimal supplies. The better the characterizationof the environment,the better our understanding of the relationship between crop performance and environment would be. In general, a phenotype can be expressed as follows if GE interaction is not important: P = G + E.
B. ENVIRONMENTAL INFLUENCE ON HEREDITY The genetic constitutionof an individual does not change from one environment to another, unless the environment is such to induce a mutation. Therefore, any phenotypic variation (norms of reaction) for a specific genotype is attributable to the environment.Even qualitative traits, which are controlled by one or two genes and assumed to exhibit a heritability of loo%, are subject to environmental modification. A few important examples follow: The sun-red gene in maize (Zea mays L.) produces red kernels when they are exposed to light; in the absence of light, kernels remain white (Bums, 1969). The henbane plant (Hyoscyamus niger) requires both a prolonged period of low temperature (vernalization) and daily illumination of more than 16 hours to develop flower primordia; in the absence of either of these environmental conditions, plants remain vegetative indefinitely (Redei, 1982). The darker color of the extremities of Himalayan rabbits is due to the temperature-sensitive alleles (chchor chc combinations) (Bums, 1969). Gardeners manipulate the soil pH to grow hydrangeas (Singleton, 1967).When hydrangeas are grown on acidic soil, they produce beautiful blue flowers, but on alkaline soil, they produce off-white, faintly pink, less attractive flowers. Another excellent example of an environmental influence on heredity is the reversal of dominance. In Arabidopsis, the co allele controls flowering time and is recessive under continuous illumination, but it behaves as dominant when the daily light cycle is reduced to eight hours (Redei, 1982).Reversal of dominance also was reported in squash (Shifriss, 1947). When plants with green fruits and vine-type stems were crossed with those with yellow fruits and bushy growth, the F, segregation ratio at flowering was 9 green, bushy : 3 green, vine : 3 yellow, bushy : l yellow vine; whereas, at maturity, a reversal of dominance was observed ( 1 green, bushy : 3 green, vine : 3 yellow, bushy : 9 yellow vine). In wheat (Triticum aestivurn L.), a gene for frost resistance is located on chromosome 5A. The additive-to-dominance ratio changed from one freezing temperature to another. At a high freezing temperature (- 10"C), frost resistance was dominant; however, as the temperature decreased, the direction of dominance
203
USING GE INTERACTION FOR CROP DEVELOPMENT
was reversed, and at the lowest freezing temperature (- 14"C), frost sensitivity was dominant (Sutka and Veisz, 1988). These examples clearly point out that dominance of a gene is not always absolute and that environment can and does modify phenotypic expression of a genotype. Now, it should be easy to understand how quantitativetraits (controlled by polygenes) can be modified by the environment. Quantitative traits exhibit continuous variation patterns that are due to polygenic control andor environmental or nongenetic factors. Because of the considerable influence of environment, quantitative traits generally have relatively low heritability. As environmental influence increases, reliability of phenotype as an indicator of genotype is reduced. Modification occurs at the biochemical level, i.e., gene product (protein) level. Recently, Dean (1995) provided evidence of GE interactions among activities of permeases; however, there was little or no GE interaction among galactosides' activities.
C. GENOTYPE-BY-ENVIRONMENT INTERACTION Genotype-by-environmentinteraction is said to be present when different cultivars or genotypes respond differently to diverse environments, and for GE to be detected via statistical procedures, at least two genotypes (cultivars) must be evaluated in at least two environments. A basic model that includes GE interaction is
P
=
G
+ E + GE.
+
This model can be written from a statistical standpoint as: Po = p + Gi Ej + (GEjU.It follows from this model that for a given genotype, there can be many phenotypes depending upon the environment. Simmonds (1981) defined G, E, and GE effects, for a 2-genotypes x 2-environments case, as follows: Genotype-environment
El
G1
a b A,=b-a
E2
Difference (E effect) ~
G2 Difference (G effect)
C
d A,=d-c
~
A,=c-u A,=d-b
GE interaction: (A2 - A,) = (A, - AJ or (d - b) - (c - a ) = (d - c ) - (6 a)or(A, + A,) = (A2 + A3)or(c - a) + (d - c ) = (d - 6 ) + (b - a).Thegenotype effect, A3, represents change (or influence) due to genotypes in environment El, and A, is the change due to genotypes in environment E2. The environmental effect, A,, represents change due to environments for genotype G1, and A, is the change due to environments for genotype G2. Total effect (7') = G + E + GE = (d - a); or GE = T - G - E. A distinction must be made between GE interaction and genotype-environ-
2 04
MANJIT S. KANG
mental correlation (covariance). The latter occurs if genotypic and environmental effects are not independent. There is an interaction if the difference between the average phenotypic value for two genotypes changes in different environments, but there is a correlation if particular genotypes tend to be associated with positive, and other genotypes with negative, environmental effects (Crow, 1986; Doolittle, 1987). Interaction signifies nonadditivity, whereas correlation means a preferred genotype is provided a better environment, and vice versa. Vp = V, + V, VGE 2 Cov,,, where V, denotes phenotypic variance; V,, genetic variance; VE, environmental variance; VGE, GE interaction variance; and Cov,,, covariance between genotype and environment. The 2 Cov,, entity is present in addition to VGE. Crow (1986) explained genotype-environment correlation as follows: “Suppose a measurement, z, is the sum of two components, x and y. Thus, we write z = x y. Then the variance of z is given by V, = Vx Vy + 2 Cov,. If x and y are independent, then Covv = 0.” There are many consequences and challenges experienced by breeders and geneticists when GE interactions are present; numerous researchers have voiced their concerns. McKeand et al. (1990) observed, “The breeder is faced with developing separate populations for each site type where genotypic rankings drastically change and/or with selecting genotypes which generally perform well over many sites. The first approach will yield greater genetic gains, but costs will most likely be increased. The second approach is less expensive but gains are also less.” Medina (1992) stated, “Genotype-environment interaction is of major importance to the plant breeder in developing improved varieties because the relative rankings of varieties grown over a series of environments may differ statistically, causing problems in plant selection.” Meredith (1984) pointed out: “Genotype x environment interactions are important to geneticists and breeders because the magnitude of the interaction component provides information concerning the likely area of adaption of a given cultivar. The relative magnitudes of the interaction, error, and genotypic components are useful in determining efficient methods of using time and resources in a breeding program.” Busey (1983) indicated that a lack of consistency in genotype performance across locations or years provided additional information for the breeder and that in addition to justifying the need for additional broad-based testing in different environments, the degree of inconsistency could help predict the variability expected among different farms. Denis and Gower (1996) advised plant breeders to consider GE interaction to avoid missing a variety whose average performance was poor but which performed well when grown in specific environments or to avoid selecting a variety whose average performance was good but which performed poorly when grown in a particular environment. They advocated the use of a biadditive model (a subclass of bilinear models) that would underpin better informed decisions on variety recommendation and genotype selection.
+
+
+
+
USING GE INTERACTION FOR CROP DEVELOPMENT
205
Generally, GE interaction is detected via statistical procedures, such as analysis of variance (ANOVA). Lewontin (1974) and Gupta and Lewontin (1982) argued that the GE interaction term in ANOVA was less useful than was the norm of reaction (Via, 1984b; Mazer and Schick, 1991), i.e., the graphical representation of the response of each genotype to a change in the environment; however, a shortcoming of the norms of reaction is that they cannot be statistically tested. An alternative approach could be to examine gene-environment interactions mathematically in the context of genetic correlations between the expressions of a character across environments (Yamada, 1962; Falconer, 1981; Via, 1984a; Kang et al., 1984). Falconer (1981) advised that a trait measured in two different environments should be regarded not as one trait but as two traits because physiological mechanisms in different environments would be, to some extent, different, and, consequently, loci required for high performance of a particular trait could also be, to some extent, different. Knowledge of the relationship between genotype and phenotype in different environments helps one to make accurate predictions of the response to selection in species that inhabit spatially or temporally heterogeneous environments (Mazer and Schick, 1991). If the phenotypic expression of a genotype for a trait were dependent on growing conditions (i.e., sufficient phenotypic plasticity is present), measures of its heritability would also vary across growing conditions (Mazer and Schick, 1991). Mazer and Schick (1991) explained various magnitudes of interaction as follows (Fig. 1): If the phenotypic ranks of two genotypes change across environments (e.g., A vs. C , D, or E), the genotype favored by selection will also differ between environments. However, if ranks remain unchanged (e.g., A vs. B, or B vs. D), but the magnitude of inter-genotypic differences in phenotype increases significantly across environments, the estimates of heritability and predicted phenotypic response to selection will increase (given constant phenotypic variance, Vp), whereas the genotype favored by selection remains unchanged. If both the phenotypic ranks and the degree of expressed additive genetic variance (V,) vary across environments, both the rate of phenotypic evolution and the particular genotypes favored in each environment may vary as well. Secondly, since phenotypic variance due to G x E interaction figures prominently in the denominator of the ratio that defines heritability (V,/[V, t V, + V,,]), this source of variance may preclude consistent measures of heritability for traits exhibiting high phenotypic plasticity in spatially or temporally heterogeneous environments. In such environments, heritability estimates will depend strongly on where, when, and among which genotypes this parameter is measured. Thirdly, if the relationship between genotype and phenotype varies across environments in a deterministic manner that results in predictable changes in her-
MANJIT S. KANG
2 06
Genotype x environment interoctions
2
I Environment
A vs B A vs C A vs D A vs E
Magnitude of intergenotypic difference
Genotypic ranks
Direction of environmental modification
Increases Remains the same Increases Increases
Remain the same Reverse Reverse Reverse
Opposite Opposite Same Opposite
Figure 1 Types of GE interaction. The relationship between genotype and phenotype may change across environments in two major ways that are not mutually exclusive. First, the magnitude of intergenotypic variation may change; second, the relative phenotypic rank of different genotypes may change across environments. The columns beldw the illustration indicate four comparisons between pairs of genotypes and the type of GE interaction that they demonstrate. (Reprinted from Mazer and Schick, 1991, with permission from the Genetical Society of Great Britain.)
itability with growing conditions, then it may be possible to identify the conditions under which phenotypic differences among genotypes are most likely to predict the conditions in nature under which evolutionary change by natural or artificial selection may most rapidly occur. Fourthly, many theoretical models make assumptions about the constancy of parameters such as the heritability of fitness and genetic correlations among fit-
USING GE INTERACTION FOR CROP DEVELOPMENT
207
ness components. For example, models of the phenotypic evolution of quantitative characters under multivariate natural selection assume that genetic and phenotypic covariance matrices remain relatively constant between generations (Lande, 1976, 1979, 1980; see Turelli, 1988, for further discussion). The application of such models to populations in natural habitats requires that the expressed heritabilities of fitness-related traits remain constant in spatially and temporally heterogeneous environments. Studies of the nature and strength of G x E interactions enable evolutionists to evaluate whether these assumptions apply generally to wild species in nature. Finally, one proposed mechanism for maintaining additive genetic variance within populations requires the presence of strong G x E interaction (Via and Lande, 1987). If the relative phenotypic rank of genotypes with respect to individual fitness changes across environments, significant V, may be maintained even in the presence of strong selection and a heritability of fitness expressed in each environment. Under conditions of frequent shifts in the ranks of different genotypes in distinct microhabitats, different genotypes would be favored by natural selection in different environments, thus, maintaining an array of genotypes in a heterogeneous environment.* Breeders-agronomists usually test a diverse array of genotypes in diverse environments, which implies that GE interactions are to be expected. Genotype-by-environment interactions can be grouped into two broad categories: crossover and noncrossover interactions; a brief discussion of each follows.
I. Crossover (Qualitative) Interaction Differential response of cultivars to diverse environments is referred to as crossover interaction when cultivar ranks change across environments. In Fig. I, genotype A vs C, A vs D, and A vs E, B vs E, C vs D, C vs E, and D vs E comparisons represent crossover or qualitative interactions. A main feature of crossover interaction is intersecting lines. If the lines do not intersect, there is no crossover interaction.
2. Noncrossover (Quantitative)Interaction These interactions represent changes in magnitude of genotype performance (quantitative), but rank order of genotypes across environments remains unchanged; i.e., genotypes that are superior in one environment maintain their superiority in other environments. Noncrossover interactions may mean that cultivars ‘Mazer, S. J. and Schick, C. T. (1991). Constancy of population parameters for life history and floral traits in Raphanus sativus L. I. Norms of reaction and the nature of genotype by environment interactions. Here&& 67,143-156; reprinted with permission of Blackwell Science Ltd.
208
MANJIT S. KANG
are genetically heterogeneous but test environments are homogeneous, or genotypes are genetically homogeneous but environments are heterogeneous. All identical genotypes grown in constant (ideal) environments would perform consistently. Any variance from the ideal environment leads to GE interaction. In crop cultivar development, the crossover interaction is more important than noncrossover interaction (Baker, 1990). According to Gregorius and Namkoong (1986), the former is not only nonadditive in nature but also nonseparable. Lack of crossover type of interaction for quantitative trait loci (QTL) even in the presence of significant GE interaction has been reported (Lee, 1995; Beavis and Keim, 1996). For further discussion on crossover and noncrossover interaction, refer to Crossa et al. (1996), Cornelius et al. (1996), and Beavis and Keim (1996).
II. IMPLICATIONS OF GE INTERACTION IN BREEDING Genotype-by-environmentinteraction is an important subject in quantitative genetics as related to plant breeding (Wricke and Weber, 1986; Hallauer, 1988; Jensen, 1988; Borojevic, 1990; Kang, 1994). Agricultural researchers have long been cognizant of the various implications of GE interactions in breeding programs (Yates and Cochran, 1938; Dickerson, 1962; Comstock and Moll, 1963; Finlay and Wilkinson, 1963; Allard and Bradshaw, 1964). Variation among genotypes in phenotypic sensitivity to the environment (GE interaction) presents a real problem for breeders, as it may necessitate the development of locally adapted varieties (Falconer, 1952).If no one genotype has superiority in all situations, GE interaction indicates the potential for genetic differentiation of populations under prolonged selection in different environments (Via, 1984a). Thus far, agricultural production has kept pace with the world’s population growth mainly because of innovative ideas and efforts of agricultural researchers. The world’s population is expected to double in the next 40 to 50 years (Lee, 1995). The key to doubling agricultural production is increased efficiency in utilization of resources (increased productivity per acre and per dollar), and that includes a better understanding of GE interaction and ways to exploit it. From evolutionary biologists’ perspective, GE interactions are important in maintaining genetic variation in and adaptation of species. Therefore, these interactions present both problems and opportunities for geneticists. The understanding achieved in evolutionary or ecological genetics can be applied to crop breeding since the common thread in both areas is selection-natural selection in the former and artificial selection in the latter. The importance of GE interaction can be stated according to Gauch and Zobel (1996): “Were there no interaction, a single variety of wheat (Triticum aestivum L.) or corn (Zea mays L.) or any other crop would yield the most the world over,
USING GE INTERACTION FOR CROP DEVELOPMENT
209
and furthermore the variety trial need be conducted at only one location to provide universal results. And were there no noise, experimental results would be exact, identifying the best variety without error, and there would be no need for replication. So, one replicate at one location would identify that one best wheat variety that flourishes worldwide.” Breeders and geneticists are aware, however, that this type of ideal situation does not exist in reality. A crop cultivar development program has two major components: crop breeding methodology and performance evaluation of improved cultivars in target areas. Implications of GE interaction for these areas are discussed in the following sections.
A. IMPACT ON BREEDING 1. Genetic (G), Environmental Q, and GE Interaction Contributions The importance of GE interactions can be seen from relative contributions of new cultivars and improved management to yield increases from direct comparisons of yields of old and new varieties in one trial (Silvey, 1981). Genetic improvements account for about 50% of the total gains in yield per unit area for major crops during the past 50 to 60 years (Simmonds, 1981; Silvey, 1981; Duvick, 1992, 1996).The remainder of yield gains is attributable to improved management and cultural practices. Barley yield data from the United Kingdom (1946-1977; mean yield for 1946 is 2.3 t/ha and for 1977 is 3.9 tha) indicated environmental contribution to be 10-30% and genetic contribution to be 3040%; the remaining 2 5 4 5 % of yield gain was attributed to GE interaction (Simmonds, 1981). For wheat for the same period (1946-1977; mean yield for 1946 is 2.4 t/ha and for 1977 is 4.7 t/ha), yield gain was attributed as follows: 40-60% to E, 2 5 4 0 % to G, and 15-25% to GE interaction (Simmonds, 1981). Genotype-by-environmentinteractions confound precise partitioning of the contributions of improved cultivars and improved environment-technology to yield (Silvey, 1981).
2. Proliferation of Breeding Stations Genotype-by-environmentinteraction occurs during and impacts all stages of a breeding program and has enormous implications in allocation of resources. A large GE interaction could mean establishment of two full-fledged breeding stations in a region instead of one, thus requiring increased input of resources (manpower, land, and money). For example, Louisiana and Texas have rice breeding stations within 150 miles of each other. If there were no GE interaction for all sets of rice breeding lines (strains) evaluated at LouisianaAgricultural Experiment Station, Crowley, Louisiana, and Texas Agricultural Experiment Station, Beaumont, Texas, during all selection stages, we could eliminate one station.
2 10
MANJIT S. KANG
3. Reduced Progress from Selection Heritability of a trait plays a key role in determining genetic advance from selection. As a component of the total phenotypic variance (the denominator in any heritability equation), GE interaction affects heritability negatively. The larger the GE interaction component, the smaller the heritability estimate; thus, progress from selection would be limited. Frey (1 990) indicated that success of mass selection depended on the heritability of the selected trait, the presence of additive gene action for the trait, and minimal confounding due to GE interaction. Selection response is negatively affected by large environmental and GE interaction components of phenotypic variation. Inbreeding, nonadditive genetic effects, and GE interactions cause parent-offspring (PO) regression estimates of heritability obtained from one pair of PO generations to differ from those obtained from another pair of generations (Gibson, 1996). Importance of heritability has been extensively discussed by Nyquist (1991).A recent report pointed out that selection for resistance to fusarium wilt in red clover was temperature dependent because a genotype-by-selection environment(temperature)interaction was detected (Venuto et al., 1996). Such studies are important for applied breeding programs.
4. Increased Cost of Cultivar Testing A large GE interaction reflects the need for testing cultivars in numerous environments (locations and/or years) to obtain reliable results. If the weather patterns and/or management practices differ in target areas, testing must be done at several sites representative of the target areas. If cultivars are to be used in marginal areas or fringes of a crop (specific adaptation), testing must begin in those areas as early as possible.
5. Gene Loss Due to Limited Testing in Early Selection Stages Kang (1993a) discussed the disadvantagesof discarding genotypes evaluated in only one environment in early stages of a breeding program. The discarded genotypes could have the potential to do well at another location or in another year. Thus, some potentially useful genes could be “lost” due to limited testing. An example from six-row barley illustrates this point well. A total of 288 barley lines were evaluated in the Magreb countries and in yield trials by ICARDA (International Center for Agricultural Research in Dry Areas) at three locations (Ceccarelli et al., 1994). Of the 103 lines selected at ICARDA and 154 lines at the Magreb, only 49 were selected at both locations. LeClerg (1966) suggested that in the early stages of a testing program, it would be advisable to have only one replication per location, with as many locations as possible. This could be easily achieved in hybrid crops, such as maize, or clonally
USING GE INTERACTION FOR CROP DEVELOPMENT
2 11
propagated crops, such as sugarcane and potato. However, in nonclonally propagated crops, a particular genotype from segregating progeny of a cross can be evaluated in only one environment. In such cases, a strategy should be not to plant all seed of the segregating genetic material in one year or at one location, rather a small aliquot of seed from each cross should be planted in several environments. This suggested strategy could circumvent the problems associated with augmented designs in initial selection stages, as each environmentwould represent a “replication.” In the case of clonally propagated material, one can reduce the plot size but increase the number of test environmentsat the earliest possible selection stage. This strategy would be effective in exploiting and preserving genetic variation. A breeding program should be designed to identify individual genotypes that would not only have superior performance but also display stable performance across environments at as early a selection stage as possible (Kang and Martin, 1987). Plant breeders can afford to eliminate or “lose” only those genes that show no promise in any environment.
B. MULTIENV~R~NMENT PERFORMANCE EVALUATION Performance evaluation is the second component of a breeding program. Testing done in one environment provides only limited information. For example, if 10 genotypes are evaluated in four replications in one environment, ANOVA will be as follows: Source
df
Mean square
Replications Genotypes Error
3
RMS GMS EMS
9 27
No GE interaction can be detected from a single-environment evaluation. If the test were repeated in another environment (location or year), ANOVA would be as follows:
.
Source
df
Mean square
Environments (E) Replications within E Genotypes (G) GE interaction Error
1 6 9 9 54
EnvMS R(Env)MS GMS GEnvMS EMS
212
MANJIT S. KANG
Multienvironment testing provides additional useful information; i.e., a GE interaction component can be estimated. In addition, multienvironment testing yields better estimates of variance components and heritability. Therefore, GE interaction need not be perceived only as a problem.
1. Difficulty in Identifying Superior Cultivars across Environments As the magnitude of a significant interaction between two factors increases, the usefulness and reliability of the main effects are correspondinglydecreased. Since GE interaction reduces the correlation between phenotypic and genotypic values, the difficulty in identifying truly superior genotypes across environments is magnified.
2. Increased Costs for Breeders and Seed Companies Obviously, the cost of cultivar evaluation increases as additional testing is carried out. However, with additional test environments, a breeder-agronomist can identify cultivars with specific adaptation as well as those with broad adaptation, which will not be possible from testing in a single environment. Multienvironment testing makes it possible to identify cultivars that perform consistently from year to year (small temporal variability) and those that perform consistently from location to location (small spatial variability). Temporal stability is desired by and beneficial to growers, whereas spatial stability is beneficial to seed companies and breeders. Stability of performance can be ascertained via stability statistics (Lin et al., 1986; Kang, 1990; Kang and Gauch, 1996).
III. CAUSES OF GE INTERACTION To be able to understand GE interaction and utilize it effectively in breeding programs, as much information as possible is needed on the factors responsible for differential response of genotypes to variable environments. A factor may be present at optimal, suboptimal, or superoptimal levels. When present at a level other than optimal, it represents a stress. According to Baker (1988), differences in the rate of increase in response of genotypes at suboptimal levels would reflect differences in efficiency, and differences in the rate of decrease at superoptimal levels would reflect differences in tolerance. For example, when water is at suboptimal levels (drought), water-use efficient genotypes, i.e., those with increased growth response relative to other genotypes, can be identified.At superoptimallevels (flooding), one can identify plants that are flood tolerant. Therefore, without the presence of stresses, genotype attributes, such as efficiency and tolerance, cannot be identified and investigated. In this section, the effects of environmental
.
USING GE lNTERACTION FOR CROP DEVELOPMENT
213
stress on plant genome in general and biotic and abiotic factors that may be responsible for GE interactions are considered.
A. ENVIRONMENTAL STRESS
ON
GENOME
Understanding of plant and animal stress responses is essential because of predicted global environmental changes and their impact on production of food and fiber. Stress, in the context of biological organisms, is a physiological response to the effect of an adverse environmental factor(s). Plants respond to a variety of environmental cues: nutrients, toxic elements, and salts in the soil solution, gases in the atmosphere, light of different wavelengths, mechanical stimuli, gravity, wounding, pests, pathogens, and symbionts (Crispeels, 1994). Although progress in environmentally regulated signal transduction has been made, further research is warranted in this important area. We need to understand the effects of various stresses on the genetic makeup of organisms before we can tackle the issues relative to GE interactions. Environmental stresses have been shown to elicit specific responses at the DNA level in a number of organisms.A differentiated cell expresses an array of genes required for its stable functioning and metabolic roles (Scandalios, 1990). In response to severe environmental changes, a genome can respond in a rapid and specific manner by selectively regulating (increasing or decreasing) the expression of specific genes. Plants have incorporated a variety of environmental signals into their developmental pathways that have provided for their wide range of adaptive capacities over time (Scandalios, 1990). In a monograph titled Genomic Responses to Environmental Stress, environmental stresses, such as oxidative stress, pathogenicity, temperature shifts, light, and anaerobiosis, were examined (Scandalios and Wright, 1990). Some stress factors directly affect genomes and induce heritable changes, which may even be adaptive under the stress that caused them. For example, Petunia cells selected in tissue culture for resistance against the herbicide glyphosate showed a 20-fold amplification of the gene for EPSP synthase (Shah et al., 1986). Bachmann (1993) pointed out that: (1) genomic size, aside from the coding content of DNA, had phenotypic effects, which played a role in organismic adaptation, especially under stress conditions; (2) heritable variation in genome size within species was observable as phenotypic variation; and (3) dramatic changes in genome size could occur quickly, sometimes within a generation. Qualitative similarities-differences among plants seem to be unaffected by differences in genome size (Hutchinson et al., 1979; Price et al., 1986). However, interspecific variation in DNA amounts is correlated with various quantitative properties of cells, and these may secondarily affect quantitative characters of the whole plant (Bennett, 1973, 1987; Cavalier-Smith, 1985a,b; Bachmann et al., 1985). Highly
2 14
MANJIT S. KANG
significant differences of up to 32% in DNA content were found in meristems of seedlings from 35 natural populations of hexaploid Festuca arundinacea (Ceccarelli et al., 1992). In cultivated maize, variation in genome size has been reported to be as high as 38.8% (Laurie and Bennett, 1985; Rayburn et al., 1985). Maize lines from higher latitudes of North America had nuclear DNA amounts significantly lower than those from lower latitudes (Rayburn etal., 1985). Raybwn and Auger (1990) determined the nuclear DNA content of 12 southwestern U.S. maize populations collected at various altitudes and observed a significant positive correlation between genome size and altitude. Higher amounts of DNA at higher elevation have also been found in teosinte (Laurie and Bennett, 1985). Cullis (1990) reviewed DNA rearrangements in response to environmental stress. Plants can show morphological changes in response to environmental factors. Classic cases of phenotypic plasticity include sun vs. shade leaves (Vogel, 1968), responses to herbivory (Hendrix, 1979; Lubchenko and Cubit, 1980), and competition (Turkington, 1983). The extent of phenotypic and genotypic variability is highest under conditions of stress (Parsons, 1988). Any shift from the environment to which an organism has become adapted increases the rate at which underlying mechanisms generate variation (Cullis, 1990).The examples of genomic changes in plants appear to occur by a number of basic mechanisms, viz., amplification, deletion, rearrangement, and transposition (Flavell, 1980, 1982; Cullis, 1990). The activity of transposable elements in maize, either when infected with brome mosaic virus (Mottinger et al., 1984) or during tissue culture (Lee and Phillips, 1987),was in a subset and not in all transposable element families equally. So, there may be selection for some parts of the genome that are more labile. The relationship between environmental stress and genetic reorganizations that may underpin the development of new adaptations is not well established at this time. In a stressful environment, under which organisms were severely restricted but able to survive for at least a limited time, a high rate of mutations induced by the same environment could generate the genomic reorganization underlying major adaptive shifts (Cullis, 1990). The limitation of these organizations to subsets of the genome, particularly those containing major genes controlling polygenically determined quantitative traits, would increase the number of viable variants compared with random reorganization of the genome (Cullis, 1990). If adaptive, combinations arise, the environment is no longer regarded as stressful for them. Herrera-Estrella and Simpson ( 1990) investigated influences of environmental factors on genes involved in photosynthesis. They indicated the different levels at which environmentalfactors have been proved or suggested (those marked with a "?') to act. Some factors (light, circadian factors, nitrogen stress?, iron stress?, chloroplast factors) have been shown to influence transcription initiation, others (light?, nitrogen stress?, iron stress?, cytokinins) influence mRNA stability or mRNA translatability. The mechanism of regulation may vary from one species to another (Herrera-Estrella and Simpson, 1990).
USING GE INTEMCTION FOR CROP DEVELOPMENT
2 15
Acclimation can be viewed as an adaptive response of the plant that enables it to survive an environmental stress. The metabolic and cellular changes that occur during acclimation are complex and can be considered quantitative in that the level of response may depend on both the temperature scale and the genotype of the plant involved (Hughes et al., 1993). This property of temperate plants is termed phenotypic plasticity. Features of phenotypic plasticity are: (1) modification of phenotype occurring in a single genotype, ( 2 ) responsiveness to changes in environment, and ( 3 ) changes having an adaptive value that allow an individual to withstand environmental stress (Smith, 1990). Plants have mechanisms by which they are able to acclimate to a number of environmental stresses (Key and Kosuge, 1984; Lange et d., 1981; Levitt, 1980). Thomashow ( 1990) reviewed molecular genetic mechanisms of cold acclimation in higher plants. Knowledge of molecular genetics of cold acclimation has lagged behind that of other plant responses triggered by environmental stresses, such as heat shock, anaerobic conditions, and light. However, it has been established that the COT gene in Arabidopsis codes for a 0.45-kb transcript that enhances cold tolerance more than 10-fold in acclimated plants. The level of the transcript remains high as long as plants are kept in the cold but falls to low or undetectable levels within 14 hours of returning the plants to control temperatures. The cold-regulated accumulation of COT transcripts involves both transcriptional and post-transcriptional control mechanisms (Thomashow, 1990). Research in this area has been directed to elucidating the transcriptional and post-transcriptional regulatory mechanisms involved in controlling cor gene expression, determining how plants “sense” cold temperatures and pass on this information into altered gene expression, and establishing the link between low-temperature-regulated and abscisic acid (ABA)-regulated expression of cor genes.
B. BIOTICSTRESSES Biotic stress factors are a major limitation to plant productivity and a dominant element in plant ecology and evolution (Higley et al., 1993). Biotic stresses and interactions among them and/or with abiotic factors remain poorly understood; however, they have significant relevance to GE interactions in plants. Plants may respond to pathogen infection by inducing a long-lasting, broadspectrum, systemic resistance to subsequent infections (Ryals et al., 1994). The induced disease resistance has been variously referred to as physiological acquired immunity, induced resistance, or systemic acquired resistance (SAR). The SAR is part of the defense response of a plant that is attacked by a pathogen or pest (Ryals et al., 1994). Plants’ response to a localized attack is to cause a signal to be transmitted to other plant parts where defense genes, such as the pathogenesis-related proteins and hydroxyproline-rich glycoproteins are induced. The SAR is distinct
216
MANJIT S. KANG
from preexisting resistance mechanisms, such as physical barriers or protein crosslinlung, and also from other inducible resistance mechanisms, such as phytoalexin biosynthesis, the hypersensitive response, and ethylene-induced physiological changes (Ryals et af, 1994). The SAR defense reaction is believed to be mediated by salicylic acid, which spreads systemically throughout the plant (Ryals et al., 1994). Superoxide dismutases (SOD) or catalases are understood to have an important protective function, as they can inactivate reactive oxygen species (Polle and Rennenberg, 1993; Ryals et al., 1994). Chen et al. (1993) isolated a cDNA encoding salicylic acid-binding protein that encoded a catalase isozyme. This catalase converts H,O, to water and 0,, thus inhibiting accumulation of reactive oxygen species that may act as secondary messengers to induce SAR gene expression (Ryals et al., 1994).
1. Differential Resistance-Tolerance of Cultivars to Plant Pests Differences in insect and disease resistance among genotypes can be associated with stable or unstable performance across environments. For example, Baker (1990) and Gravois et al. (1990) implicated disease resistance as a factor that contributed to GE interaction in crops.
2. Hereditary Nutrient Uptake Ability of Cultivars In the context of GE interaction, very little attention has been given to the important issue of genetic control of mineral uptake, transport, and metabolism in plants. Differential ion uptake by plants (Myers, 1960) and physiological genetics of plant nutrition (Epstein, 1972, 1976) have been discussed. Pope and Munger (1 953a,b) described two nutritional disorders controlled by single gene loci (susceptibility to magnesium deficiency and boron deficiency in celery). Iron deficiency of a soybean introduction was found to be controlled by a single recessive gene (Devine, 1982). Copper deficiency in triticale was associated with hairy peduncle ( H p ) gene located on chromosome 5R (Graham, 1982). Additional single gene controls for nutrient uptake and/or efficiency are given in Blum (1988).
3. Competition and Reproductive Adjustment Genotype-by-environmentinteraction could be due to differential survival rates among genotypes in competition with other genotypes of the same crop species or with weed species. Genetic and environmental factors and their interactions affect the number of seeds each genotype produces and the proportion of seeds of each genotype that reaches maturity (Allard, 1960). To survive in competition (stress), a plant may reduce the number of seeds generated by producing a smaller number of viable seeds. Such reproductive adjustment may be differential among genotypes.
USING GE INTERACTION FOR CROP DEVELOPMENT
2 17
4. Tolerance to Herbicides and Allelopathy Allelopathy is defined as the direct or indirect harmful or beneficial effects of one plant on another through the production of chemical compounds that escape into the environment (Miller, 1996). Differential responses among genotypes to herbicides and allelochemicalscould result in GE interaction. Nelson (1996) summarized the proceedings of Allelopathy in Cropping Systems symposium. Crop plants are often stressed by carryover herbicide or postemergence herbicide applications (Einhellig, 1996). Simultaneously with herbicide stress, weed-crop or crop-crop allelopathy can occur (Einhellig, 1996). Plant breeders should be able to alter plant resistance to the allelochemicals (Miller, 1996).
5. Water-Use Efficiency, Nutrient-Use Efficiency, and Radiation-Use Efficiency Under stress due to suboptimal levels of water, nutrients, and solar radiation, it should be possible to identify genotypes that are efficient or inefficient in using the respective resource. Woodend and Glass (1993) demonstrated the presence of GE interaction for potassium-use efficiency in wheat.
C. ABIOTIC STRESSES Signal transduction in plants is not as well understood as it is in animals. However, there are strong indications that plants use a variety of pathways to communicate information both within and between cells. The diversity of potential pathways being suggested indicates that there are a large number of possible ways in which plants can perceive and transduce environmental stresses and changes, be they biological, chemical, or physical in nature (Leigh, 1993). The major abiotic stresses include atmospheric pollutants, soil stresses (such as salinity, acidity, and mineral toxicity and deficiency), temperature (heat and cold), water (drought and flooding), and tillage operations (Blum, 1988; Unsworth and Fuhrer, 1993; Clark and Duncan, 1993; Specht and Laing, 1993). Genetic variation exists for plant responses to the preceding stress factors. Unsworth and Fuhrer (1993) pointed out that considerable potential exists for breeding for tolerance to air pollutants. Differential response of genotypes to these stresses could be a cause for GE interaction.
1. Differential Heat-Shock Responses Rapid temperature changes, particularly those toward the upper end of the adaptation range for individual plant species, can produce dramatic changes in the pattern of gene expression. Heat-shock responses are plants’ protective measures
218
MANJIT S. KANG
against potentially lethal, rapid-rate, upward departures from optimal temperature (Pollack et al., 1993).Tolerance of protein synthesis and seedling growth to a previously lethal high temperature can be induced by prior short exposure to a sublethal high temperature. During such a preparatory pretreatment, the synthesis of a specific set of proteins-the heat-shock proteins (HSPs)-occurs from mRNA that is newly transcribedin response to high temperature. In the meantime, the synthesis of normal cellular proteins is reduced or shut down. This process is detectable within minutes of the onset of stress (Ougham and Howarth, 1988). Heat-shock proteins are induced at different temperatures in different species. The rule of thumb is that temperature must be approximately 10°C higher than the optimal temperature for a particular species. In sorghum and millet, HSPs are induced at 45°C; in temperate grass Loliurn ternulentum, HSPs are induced at 35°C; and in the snow fungus Fusarium nivale, HSPs are triggered at 25°C. The heatshock response is controlled at both the transcriptional and the translational levels. HSP gene expression has also been detected in field-grown soybeans and other plants (Kimpel and Key, 1985; Burke et af.,1985). HSPs were expressed to a much greater extent in cotton grown in nonirrigated fields where leaf temperatures were 10°C higher than those in irrigated fields. Heat-shock protein studies have provided insight into the effects of sudden changes in temperature on plants (Nagao et al., 1990). Plants in the field are often exposed to high temperatures on a daily basis, occurring for a few hours at midday and recurring each day, and this can have a detrimental effect (Pollack et al., 1993). Experiments with pearl millet seedlings showed that although seedlings were able to synthesize HSPs on their first exposure to high temperature, on subsequent days this ability was lost (Howarth, 1991).Pollack et al. (1993) indicated that a seedling emerging from the soil in the field was likely to be exposed to a wide range of temperatures. In tropical areas, at midday, a seedling might be exposed to a root zone temperature of 40"C, a soil surface temperature of 60"C, and an air temperature of 45°C. Within the diurnal cycle, these temperatures might drop to 25,20, and 21"C, respectively. For a grass or cereal seedling whose shoot meristem is near the soil surface, this has considerable consequences for the growth and survival of the seedling (Peacock, 1975).HSPs may be important for plants' survival in such wide fluctuationsin temperature by stabilizingthe meristematic tissue. Seedlingsin very young developmental stages may be particularly vulnerable, as they may not have developed a full transpiration stream and an effective cooling system.
2. Oxidative Stress A common feature of different stress factors is an increased production of reactive oxygen species in plant tissues, but their mode of action varies depending on whether oxidants are generated outside (e.g., by oxidizing air pollutants) or inside (e.g., high radiation, low temperatures, or nutrient deficiency) a plant cell (Polle and Rennenberg, 1993).To improve plant protection from oxidative damage, it is
USING GE INTERACTION FOR CROP DEVELOPMENT
2 19
important to understand both the mode of action of different stress factors and the critical physiological properties that limit ameliorative mechanisms at the subcellular level (Polle and Rennenberg, 1993). The significance of antioxidant compounds and enzymes for removal of oxidants in different cellular compartments is dependent on the concentration in the given compound and the rate constant for the oxidant. Superoxide dismutase is the most efficient (high rate constant) known scavenger for superoxide radicals (Polle and Rennenberg, 1993). The enzyme occurs in different isozymes in different subcellular compartments (Asada and Takahashi, 1987). Chloroplasts contain SOD, ascorbate, and glutathione at concentrations of 10, 10-50, and 3 pM, respectively, which contributes to scavenging of 0; in the order 10:1-5: 1 (Polle and Rennenberg, 1993). In maize, Malan eb ul. (1990) studied the F, hybrid of two inbred lines, which resulted in increased activity of two antioxidant enzymes (SOD and glutathione reductase). They observed improved protection from oxidative stress. Singly, neither enzyme was effective in developing stress resistance.A balanced increase in antioxidants was needed to obtain increased stress resistance. Scandalios (1 990) summarized plant responses to environmental stress, pointing out that activated oxygen species (endogenous, by-products of normal metabolism, and exogenous, triggered by environmental factors) were highly reactive molecules capable of causing extensive damage to plant cells. The effects of oxidative stress can range from simple inhibition of enzyme function to the production of random lesions in proteins and nucleic acids and the peroxidation of membrane lipids. He further pointed out that loss of membrane integrity due to peroxidation, together with direct damage to enzymatic and structural proteins and their respective genes, can cause decreased mitochondria1 and chloroplast functions, whch, in turn,lower plant’s ability to fix carbon and to properly utilize the resulting products. This decrease in metabolic efficiency results in reduced yield. To understand and be able to explain GE interaction, the preceding factors must be considered. It is usually recognized that different cultivars respond differently to environmental stresses. This is precisely the reason why we see interaction of different genotypes with different environments. The term “stress” has to be judged from the perspective of an organism, not from that of a scientist. Cushman et al. (1990, p. 197), in a review on gene expression during adaptation to salt stress, stated: “Yet halophytes exist and thrive under conditions of high salinity, which suggests that life under salt water stress is not necessarily life on edge.” Hulobacteriurn has evolved such a high adaptation to a saline environment that its ribosomes are functional only in the presence of 3 to 4 M KCl, in contrast to 1 mM Mg in nonhalophytic Escherichiu coli (Conte, 1973). With the advances being made in the area of biotechnology (molecular genetics), it should be possible to minimize or reduce the effects of stress by identifying, cloning, and incorporating into cultivars, appropriate gene(s>from other organisms. The genetic or metabolic processes affected by some of the known environmental factors are tabulated to provide an overview of plant responses to stresses (Table I).
Table I Plant-Environment Interactionsas Related to Adaptation and Acclimation to the Environment: Biochemical and MetabolicAspects Environmental factor
NO,- and light
Variation in light High light
Water deficit
Water logging
Affected genetic or metabolic process; effect on growth Transcription of nitrate reductase gene. Circadian rhythm of NR gene expression in tobacco (NR mRNA increases at night and decreases in day). Glutamine (acts as repressor) involved in circadian expression of NR gene. Gln accumulation correlated with low N R mRNA accumulation. Light: causes rapid disappearance of phosphoenolpyruvatecarboxylase (PEPC) kinase activity and rapid dephosphorylation of PEPC and favors malate removal (similar to high temperature).(Crassulacean acid metabolism (CAM) PEPC). , for PEPC, and favors malate Light: causes phosphorylation of PEPC, reduces K formation. Fusicoccin effects similar to those of light (Guard cell P E E ) . Darkness: causes dephosphorylation of P E E , and favors malate removal. Supplies energy for photosynthesis; immediate effect on metabolism. Turnover of ascorbate (triggers synthesis and breakdown). Long exposure to high inadiance increases ascorbate pool, particularly at low temperatures, which increases the excess excitation energy. High ascorbate favors photo-oxidative damage. Decreased stomatal conductance; limits CO, supply and changes balance between photosynthesis and photorespiration. hhibition of growth and leaf expansion. Alteration of gene expression;ABA synthesis triggered. Cell-wall protein composition is altered. Plant response proline accumulation. Photo-inhibition and photo-oxidation. Water deficit could be caused by drought, heat, cold, and salt; glycinebetaine, polyols (polyhydric alcohols) such as sorbitol, and cyclitols accumulate in stressed organisms. Polyols serve as carbon storage and translocation compounds. Inhibition of aerobic respiration under hypoxia and induction of fermentation pathways as a means of energy production and NAD regulation. Ethylene production stimulated under hypoxia in maize roots. Putrescine (endogenous regulator) produced limits formation of aerenchyma by suppressingaction of ethylene rather than its synthesis. Pyruvate decarboxylase(PDC) genes in maize seedlings induced anaerobically. Two anaerobically induced genes in maize (in contrast to all other known genes of this type) do not seem to encode proteins with a glycolytic function.
Reference Cheng et af. (1986), Galangau et al. (1988). Deng et al. (1991)
N m o et QZ. (1995)
Nimmo et al. (1995) Pearcy (1990) Mishra et QZ. (1993), Smirnoff (1993) Lawlor (1995)
Popp and Smirnoff (1995)
Ratcliffe (1995)
Peschke and Sachs (1993) Peschke and Sachs (1994)
Temperature
Ambient CO, decrease Light and available CO, Pi (inorganic P)
N-nutrition Long-term N-deprivation Molecular oxygen (0,) Superoxide radical (O,-)
Expression of alcohol dehydrogenase and PDC genes in the anoxia-tolerant plant Acorus calarnus L. regulated at transcription, translation, and posttranslation levels. Oxidative damage to plants because of elevated level of soil Fe(I1) formed because of low redox potential. High iron levels in the cytosol or apoplast could result in increased hydroxyl radical formation (and, thus, in lipid peroxidation).
Bucher and Kuhlemeir (1993)
Metabolic rate; differential effect on enzymes and pathways. Carbohydrate metabolism, photosynthesis, and photorespiration. Sink-source relationships change differentially by effect of temperature.
Leegood ( 1995)
High temperature: causes rapid disappearance of phosphoenolpyruvate carboxylase (PEPC) kinase activity and rapid dephosphorylation of PEPC. reduces apparent Kl for malate, and favors malate removal (low temperature effects are opposite). (CAM PEPC). Rapid increase in rate of N assimilation by modulation of nitrate reductase (shoots and roots). Sucrose phosphate synthase (SPS)
Nimmo ef al. ( I 995)
Respiratory pathways; P, stimulon (complete set of genes that are co-regulated by P,). Plants respond adaptively to P, deprivation via induction of alternative pathways of glycolysis and mitochondria1 electron transport. SPS Decreased photosynthesis, biomass, and growth. Inhibition of photosynthesis (competes with CO, for RuBP carboxylase) and random production of free radicals. Membrane lipid peroxidation, enzyme inactivation, depolymerization of polysaccharides, reaction with H,O, to form OH', aging, and autoimmune diseases.
Theodorou and Plaxton (1 995)
Hydrogen peroxide (H2Oz)Inhibition of CO, fixation, marking of some proteins for proteolytic degradation, oxidation of flavonols and sulfhydryls, mutagenesis, and inactivation of lightactivated Calvin-cycle enzymes. Hydroxyl radical (OH.) Most potent oxidant known. Causes DNA lesions, protein degradation, peroxidation of membrane lipids, ethylene production (also implicated in rheumatoid arthritis). Singlet oxygen (lo,) Mutagenesis, membrane lipid peroxidation, and photoinhibition of amino acids.
UV-B (29C320 nm) Freezing and desiccation
Biologically active radiation (of concern because of depletion of ozone layer). Causes additive effect on photo-inhibition caused by photosynthetically active radiation. Oxidative and free-radical damage by physical disruption of cell structure.
Hendry and Brocklebank (1985)
Kaiser and Forster (1989), Pace ef al. ( 1990)
Rideout et al. ( 1994) Scandalios (1990). Smirnoff ( 1995) Scandalios ( 1 9 9 0 ~Smirnoff (1995) Scandalios (1990), Smirnoff (1995) Scandalios (1990), Smirnoff (1995) Scandalios (1990), Smirnoff (1995) Bornman and SundbyEmanuelsson (1995) McKersie er al. (1993)
0
2
24
P
n
3
3n
8%
r
5:
3 N N c
222
MANJIT S. KANG
Many of the environmental factors that fluctuate are associated intimately with metabolic processes (Smirnoff, 1995). Metabolic processes involve chemical reactions that are mediated by enzymes. Enzymes being products of gene expression (transcription and translation), biochemical and physiological processes are integrated with genetics.
D. PHENOTYPIC PLASTICITY 1. What Is Phenotypic Plasticity? Phenotypic plasticity was defined by Bradshaw (1965) as the amount by which the expressions of individual characters of a genotype are changed by different environments. Genotype-by-environment interaction is equivalent to genetic variation in phenotypic plasticity (Via and Lande, 1985). Phenotypic plasticity is extremely common in plants and is often expressed in stressful environments-for example, production of cleistogamous flowers (that do not open and ensure selffertilization) in dry periods compared to open flowers in other periods, and the loss of leaves under climatic stress (Bradshaw, 1965). Phenotypic plasticity encompasses short-term reversible responses, acclimation responses, and irreversible changes during development (Hoffmann and Parsons, 1991). High levels of phenotypic plasticity for physiological traits occur in species from habitats experiencing environmental fluctuations, whereas levels of plasticity for morphological traits are relatively low in species from stressed environments (Hoffmann and Parsons, 1991).
2. Genetics The degree of expression of phenotypic plasticity is under genetic control; populations and species show different levels of plasticity for the same character in response to the same environmental variables (Bradshaw, 1965; Schlichtling, 1986). Mazer and Schick (1991) demonstrated, by examining norms of reaction, that genetic variation (additive genetic variance) existed for phenotypic plasticity. Therefore, phenotypic plasticity is a trait that can be manipulated through breeding. Marshall and Jain (1968) suggested that the amount of phenotypic plasticity of a species would be inversely related to its genic heterozygosity.Adaptation to variable environments may be accomplished either by means of genetic variation or phenotypic plasticity, and the least heterozygosity should be the most plastic (Schlichting and Levin, 1984). In Arabidopsis thaliana, the degree of heterozygote advantage (heterosis) for growth increased with stress for all environmental variables (temperature, light intensity, mannitol concentration) except nutrient concentration (Pederson, 1968).
USING GE INTERACTION FOR CROP DEVELOPMENT
223
Four explanations have been proposed to account for changes in heterosis with stress (Hoffmann and Parsons, 1991): 1. Parental lines are highly inbred, and heterosis largely represents a recovery from inbreeding depression. Heterosis will increase with stress if inbreeding depression is more pronounced in adverse environments than in favorable ones. This explanation does not account for heterosis in the many crosses where parental strains are not inbred. 2. A second possibility is that parental lines reach their optimal performance at different levels of a single environmental variable or a combination of variables (Knight, 1973). Hybrid genotypes are assumed to have an intermediate optimum between the two parents (additive gene action). The hybrid will perform better than or equal to the midparental value across environments, and there will be heterosis by environment interactions because the hybrid will outperform its parents at some levels of the environmental variable. One environmental variable is not enough to generate increased heterosis in stress environments. Knight (1973) considered different levels of a second environmental variable that interacts with the first variable. Different levels of the first and second environmental variables can result in heterosis under extreme conditions, or heterosis under optimum conditions, or cases where hybrid performance increases or decreases relative to midparental value. 3. Heterozygotes have an advantage under stressful conditions but not in favorable environments. Heterozygote advantage may, therefore, be attributed to the fact that these genotypes are more successful at countering fluctuating conditions (greater homeostasis) (Lerner, 1954). This explanation may relate to some agricultural studies in which stressful conditions cause decreased productivity. Environmental variability may increase as conditions for agricultural productivity deteriorate (Blum, 1988). 4. Langridge (1968) explained that heat-sensitive enzymes were the most common consequences of mutations that do not inactivate the enzyme, and that some of the mutations are expressed only at high temperatures with complete dominance in the heterozygote.
IV WAYS OF DEALING WITH GE INTERACTION Eisemann et al. (1 990) provided critical insights into the phenomenon of GE interaction from a breeder’s perspective. They indicated that future progress in analyzing differences in genotypic adaptation in crop improvement programs would require plant breeders to pay more attention to influences of environmental factors. They advocated closer cooperation between crop breeders and other disciplines in integrated studies designed to understand the biological basis of geno-
224
MANJIT S . KANG
typic adaptation. Eisemann et al. (1990) listed three ways of dealing with GE interactions in a breeding program: (1) ignore them, i.e., by using genotypic means across environments even when GE interaction exists; ( 2 ) avoid them; or (3) exploit them in breeding objectives. All agree that interactions should not be ignored when they are significant and of the crossover type. The second way of dealing with these interactions, i.e., avoiding them, involves minimizing the impact of significant interactions. One approach is to group similar environments (mega-environments) via a cluster analysis. With environments being more or less homogeneous, genotypes evaluated in them would not be expected to show crossover interactions. By clustering environments, potentially useful information may be lost. For example, if 10 genotypes were evaluated in 100 environments, a genotype or genotypes that had stable performance across 100 environments could be identified. However, if the 100 environments were grouped into four mega-environments (25 environments each), we might have 4 different sets of genotypes to evaluate in the 4 mega-environments. Thus, we would not be able identify the genotype that had the potential to yield consistently across 100 environments. Such an approach may be useful to exploit narrow adaptation, however, if broad adaptation (ability to cope with the uncertainty of the environment) were the goal, clustering would not be a viable approach. This could also increase the number of breeding stations to accommodate several megaenvironments. One of the objectives of the CIMMYT (InternationalMaize and Wheat Improvement Center) maize and wheat breeding programs is to identify genotypes with broad adaptation (i.e., stable performance across diverse environments) at many international sites. Such an objective cannot be achieved by restricting (clustering) test environments. The third approach encompasses stability of performance across diverse environments by analyzing and interpreting genotypic and environmental differences. This approach allows researchers to identify the causes of GE interaction and provide the opportunity to correct the problem. If the cause for unstable performance of a genotype were known, either the genotype could be improved by genetic means or proper environment (inputs and management) could be provided to enhance its productivity. A genotype that performs consistently (high yielding) across many environments would possibly possess broad-based, durable resistances-tolerances to the biotic and abiotic environmental factors that it encountered during development. Broad adaptation and stability of performance (reliability) across environments are admirable objectives to conserve resources. To achieve greater success, crop environments need to be characterized as fully as possible. The more the breeders know about the crop environment, the better job they can do in developing cultivars with wide adaptability or judiciously targeting appropriate cultivars to production environments. Breeders and geneticists need to study the effects of environmental factors, and changes therein, on crop genotypes and apply appropriate and effective screens to identify suitable germplasm.
USING GE INTERACTION FOR CROP DEVELOPMENT
225
In the next section, the concepts of stability are presented. Methodology for identifying stable genotypes and environmental factors that may be responsible for stable or unstable performance is also given.
V. STABILITY STATISTICS: CONCEPTS AND USEFULNESS
A. STATICvs DYNAMIC The static concept means that a genotype has a stable performance across environments and there is no among-environments variance. This would mean that a genotype would not respond to high levels of inputs such as fertilizer. This type of stability would not be beneficial for the farmer. This type of stability has also been referred to as biological concept of stability (Becker, 1981), which is equivalent to type 1 stability in Lin et al. (1986). The dynamic concept means that a genotype has a stable performance, but for each environment, its performance corresponds to the estimated level or predicted level. There would be agreement between the estimated or predicted level and the level of actual performance (Becker and Leon, 1988). This concept has been referred to as the agronomic concept (Becker, 1981), which is equivalent to type 2 stability in Lin et al. (1986).
B. TYPES OF STABILITY Lin et al. (1986) defined four groups of stability statistics. Group A is based on deviation from average genotype effect (DG), group B on GE interaction term (GEI), and groups C and D on either DG or GEI. Groups A and B formulas represent sums of squares, and those of groups C and D represent regression coefficient or regression deviation. They integrated type 1, type 2, and type 3 stabilities with the four groups: group A was regarded as type 1 , groups B and C as type 2, and group D as type 3 stability. In type 1 stability, a genotype is regarded as stable if its among-environment variance is small; in type 2, a genotype is regarded as stable if its response to environments is parallel to the mean response of all genotypes in a test; and in type 3 stability, a genotype is regarded as stable if the residual mean square from the regression model on environmental index is small (Lin et al., 1986). Lin and Binns (1988) proposed type 4 stability concept on the basis of predictable and unpredictable nongenetic variation; the predictable component related to locations and the unpredictable component related to years. Lin and Binns
226
MANJIT S. KANG
(1988) suggested the use of a regression approach for the predictable portion and the mean square for years-within-locations for each genotype as a measure of the unpredictable variation. The latter was called type 4 stability statistic.
C. STABILITYSTATISTICS Several methods have been developed to analyze GE interaction (Lin et a/., 1986; Becker and Leon, 1988; Kang, 1990; Kang and Gauch, 1996; Weber et al., 1996). The earliest approach was the linear regression analysis (Mooers, 1921; Yates and Cochran, 1938). Finlay and Wilkinson (1963), Eberhart and Russell (1966), and Tai (197 1) popularized variations of the regression approach, assuming an expected linear response of yield to environment. Merits and demerits of several methods were discussed by Kang and Miller (1984). Kang et al. (1 987b) reported on the relationship between Shukla’s stability variance (Shukla, 1972) and Wricke’s ecovalence (Wricke, 1962) and concluded that these measures identically ranked cultivars for stability (rank correlation coefficient = 1.OO). These types of measures are useful to breeders and agronomists because they provide contribution of each genotype in a test to total GE interaction. They also can be used to evaluate testing locations by identifying those locations with similar GE interaction pattern (Glaz et al., 1985). Dashiell et al. (1994) evaluated the usefulness of several stability statistics for simultaneously selecting for high yield and stability of performance in soybean. Fernandez (1991) also evaluated stability statistics for similar purposes. Kang and Magari (1996) discussed new developments in phenotypic stability analyses. Recently, other statistical methods that have received attention are Pattern Analysis (DeLacy et al., 1996), the Additive Main effects and Multiplicative Interaction (AMMI) model (Gauch and Zobel, 1996), the Shifted Multiplicative Model (SHMM) (Crossa et al., 1996; Cornelius et al., 1996), nonparametric methods of Hiihn (1996) that are based on cultivar ranks, probability of outperforming a check (Eskridge, 1996), and Kang’s rank-sum method (Kang, 1988,1993b). The methods of Huhn (1996) and Kang (1988, 1993b) integrate yield and stability into one statistic that can be used as a selection criterion.
D. SMUETANEOUSSELECTIONFOR YIELDAND STABILITY Several methods of simultaneous selection for yield and stability and relationships among them were discussed by Kang and Pham (1991). Kang (1993b) discussed the motivation for emphasizing stability in selection processes. He enumerated consequences to growers of researchers committing type I (rejecting the null hypothesis when it is true) and type II errors (accepting the null hypothesis
USING GE INTERACTION FOR CROP DEVELOPMENT
227
when it is false) relative to selection on the basis of yield alone (conventional method, CM) and that on the basis of yield and stability. Generally, type I1 errors constitute the most serious risk for growers (Glaz and Dean, 1988; Johnson et al., 1992). The development and use of the yield-stability statistic (YS,) demonstrated the significance and wisdom of incorporating stability in selecting genotypes tested across a range of environments (Kang, I993b). A computer program (STABLE) for calculating this statistic is available free of charge (Kang and Magari, 1995). The stability component in YS, is based on Shukla’s ( 1972) stability-variance statistic (ui2).Shukla (1 972) partitioned GE interaction into components, one corresponding to each genotype, and termed each component as stability variance. Lin et al. (1986) classified:u as type 2 stability, meaning that it was a relative measure dependent on genotypes included in a particular test.
E. MERITS OF EMPHASIZING STABILITY DURING SELECTION Growers would prefer to use a high-yielding cultivar that performs consistently from year to year (temporal adaptation) and might be willing to sacrifice some yield if they are guaranteed, to some extent, that a cultivar would produce consistently from year to year (Kang et al., 1991). The guarantee that a cultivar would perform consistently would be in statistical terms, based on type I and type I1 error rates for a selection criterion that integrates both yield and stability. 1. Type II Error Rates (p) for Yield Comparisons
Consider a one-tailed null hypothesis: H,: p12 po,where p, is mean yield of a genotype and po is overall mean yield of all genotypes. The alternative hypothesis is: Ha: p,
MANJIT S. KANG
228
Table Il Type I1 Error Rate (p(1))at Different Levels of Type I Error Rate (a(l)),and Minimal Detectable Difference (6) for Pairwise Comparisonsamong Overall Yield Means” a( 1) level 6 (Mg ha-’)
0.001
0.01
0.05
0.10
0.25
0.2 0.4 0.5 0.6 0.8 1.o 1.2 1.4 1.6 1.8 2.0
0.993 0.967 0.940 0.890 0.723 0.486 0.251 0.097 0.027 0.005 0.00 1
0.956 0.857 0.777 0.674 0.425 0.206 0.074 0.019 0.003 0.000 0.000
0.845 0,649 0.527 0.401 0.189 0.066 0.016 0.003 0.000 0.000 0.000
0.743 0.505 0.382 0.271 0.108 0.030 0.006 0.001 0.000 0.000 0.000
0.515 0.278 0.181 0.111 0.032 0.006 0.001 0.000 0.000 0.000 o.oO0
“Adapted from Kang (1993b).
ror risk would be reduced by choosing a higher level of a.A higher level of a would not be as harmful to growers as a higher level of p. At a realistic value of 6 of about 0.5 Mg ha-’=l LSD at 4 2 ) = 0.1 (a(2) = 0.1 corresponds to a(1) = 0.05 in Table II), the risk of committing a type I1 error is 52.7%. Researchers should strive for (1-p) (power of a test) in the range of 0.70 to 0.80. Therefore, choosing an a( 1) between 0.10 and 0.20 would be appropriate. Carmer (1976) presented arguments for using a(2) in the range of 0.20 to 0.40 for LSD for making pairwise multiple comparisons between observed means from crop performance trials. Carmer’s suggestion is equivalent to setting a(1)in the range of 0.10 and 0.20, when one-tailed hypotheses are tested. If a( 1) = 0.25 and S were in the range of 1.0 to 1.4, the type I1 error rate would be almost zero (Table 11).
2. Type JI Error Rates for Stability-Variance Statistic For hypothesis testing with respect to GE interaction or the stability-variance statistic, comparisons are made between GE interaction means. A two-tailed null hypothesis is that mean yield of a genotype ( p I )in environment 1 (el) = its mean yield in each of the remainder environments (e2 to e l 6 in this case) or u2i = 0. The following H , is tested: no:k I e l= kle2 - kle3- kle4- . . . - k l e l 6 , i.e., u2,=0
Ha: ~
+
l e l ple2
+ kIe3 + kle4 + . . . + ~
i.e.9 c21+ 0.
1 ~ 1 6 3
USING GE INTERACTION FOR CROP DEVELOPMENT
229
If H, is rejected when it is true (type I error), the erroneous conclusion is that genotype performance is inconsistent or unstable across environments. Consequently, in calculation of YS,, that genotype would be penalized for instability, and by so doing, the chances of that genotype being rejected in the selection process would increase. The consequence of a type I error would be that the grower might miss using a stable genotype and might or might not suffer an economic loss, depending on the alternative genotype chosen. On the other hand, if H, is accepted when it is false (type I1 error), the erroneous conclusion is that genotype performance is stable or consistent across test environments. Consequently, an unstable genotype would not be penalized for instability in computing the YS, statistic. The consequence of this decision could be disastrous for growers, since they would expect high yield from the genotype on their farms, but actually the genotype could perform poorly. This implies that a higher penalty for unstable performance, i.e., a greater emphasis on stable performance, would not necessarily be harmful to growers. Thus, Huhn’s S; and S:, and Kang’s rank-sum statistic (Kang and Pham, 1991), which place greater emphasis on stability, are useful for simultaneously selecting for yield and stability. Simultaneous selection for yield and stability reduces the probability of committing type I1 errors. The combined rate of committing a type I1 error in this case will be the product of p for comparisons of overall yield mean (Table 11) and p for comparisons of GE interaction means (Table HI).For example, for 6 = 0.5 for overall mean yield comparisons and 6 = 1.6 for GE interaction means, the combined probability of type I1 error would be 0.627,0.3 16, and 0.18 1 at o! = 0.01, 0.05, and 0.10, respectively. The combined probability of committing a type I1 error for both yield and stability would be negligible (0.030 x 0.309 = 0.01) if Table III Type I1 Error Rate (p(1))at Different Levels of Type I Error Rate ( 4 2 ) )and Minimal Detectable Difference (6) Por Pairwise Comparisonsamong Genotype x Year Interaction Yield Means“ a(2) level 8
(Mg ha-’)
0.0 1
0.05
0.10
0.5 1.2 1.4 1.6
0.980 0.902 0.860 0.807 0.743 0.666
0.924 0.753 0.677 0.600 0.5 13 0.429
0.865 0.640 0.557 0.474 0.390 0.309
1.a
2.0
“Adapted from Kang (1993b).
230
MANJIT S. KANG
6 = 1.0 at a = 0.10 for mean yield comparisons and 6 = 2.0 at (Y = 0.1 for GE interaction means.
F. CONTRIBUTION OF ENVIRONMENTAL VARIABLES TO STABILITY Yield stability or GE interaction for yield is a complex issue. Evidence suggests that yield stability is genetically controlled, but the amount of genetic variation is related to the specific statistics used for stability evaluation as well as to environments. Yield stability depends on yield components and other plant characteristics, such as resistance to pests and tolerance to environmental stress factors. By determining factors responsible for GE interaction or stability-instability, breeders can improve cultivar stability. If instability was caused by susceptibility to a disease, breeding for resistance to the disease should reduce losses in disease-inducing environments and increase genotype stability. Methods of assessing contributions of weather variables and other factors (covariates) that contribute to GE interaction are available (Shukla, 1972; Denis, 1988; Magari et al., 1997;Van Eeuwijk et al., 1996). Contributions of different environmental variables to GE interaction have been reported in several articles (Saeed and Francis, 1984; Kang and Gorman, 1989; Kang et al., 1989; Gorman et al., 1989; Rameau and Denis, 1992). Based on maize hybrid yield trials conducted in multiple years and locations in Louisiana, Kang and Gorman (1989) removed from GE interaction heterogeneity due to maximum and minimum temperatures, rainfall, and relative humidity. In the following linear model, GE interaction is explained in term of the covariate used, as shown by Shukla (1972), YLJk .. =
+ ai + 8..rJ + pk + bkzi + cjk,
(1)
where Y8,kis observed trait value, p is grand mean, ai is environmental effect, 8, is blocks-within-environmentseffect, pk is cultivar effect, b, is the regression coefficient of the kth genotype’s yield in different environments, Zi is an environmental covariate, and cUkis experimental error. The environmental index (difference between mean of all genotypes in ith environment and grand mean) removed 9.61% of total GE interaction as heterogeneity. Heterogeneity removed via other covariates was low (Table IV). Rainfall during the growing season removed 1.4% of the total GE interaction, whereas the amounts removed by minimum and maximum temperature and relative humidity were negligible. Similar result have been reported for sorghum (Gorman et al., 1989), soybean (Kang et al., 1989), and other maize trials (Magari and Kang, 1993). Considering the complexity of GE interaction, the relative contribution of a single environmental variable might be very small in comparison with the total number of variables affecting it. When a number of environmental variables are considered, the combination of
USING GE INTERACTION FOR CROP DEVELOPMENT
231
Table IV Heterogeneity Removed from GE lnteractron via Environmental Covariates in Maize to Minimize Interaction" Mean squares
Source
df
Environmental index
Minimum temperature
Rainfall
GE interaction Heterogeneity Residual Pooled error
176 16 160 56 1
2.8" 2.8 2.8" 1.8
2.8" 0.01 3. I b 1.8
2.8" 0.4 3.0" 1.8
Adapted from Kang and Gorman (1989). "Significant at 0.01 probability level. 'I
two or more variables would remove more heterogeneity from GE interaction than individual variables do. Methods developed by Van Eeuwijk et al. (1996) may be helpful for this purpose. Magari et al. (1997) determined contributions of individual environmental factors or combinations thereof to GE interaction. The GE interaction was explained as: 6, = c A, where Sk is vector of effects for GE interaction for the kth genotype, cis the vector of environmental variable (matrix for more than one variable), and A is the linear regression coefficient (vector of linear regression coefficients for more than one variable). Magari et al. (1997) identified precipitation as the single most important environmental factor that contributed to GE interaction for ear moisture-loss rate in maize. They identified precipitation plus growing degree days from planting to black-layer maturity (GDD-BL) and relative humidity plus GDD-BL as the two-factor combinations that explained the largest amount of GE interaction.
G. STABILITY VARIANCE FOR UNBALANCED DATA In analyzing GE interactions, plant breeders often strive to grow all genotypes in all environments, thus producing a balanced data set. However, sometimes this is not possible, especially when a wide range of environments or long-term trials are considered. Also, the number of replications may not be equal for all genotypes when experimental plots are discarded for one reason or another. In such cases, plant breeders must deal with unbalanced data. Searle (1987) classified unbalancedness as planned unbalanced data and missing observations. Both categories of unbalancedness may occur, but planned unbalancedness (a situation when, for different reasons, one does not have data for all genotypes in all environments) is more difficult to handle. Researchers have
232
MANJIT S. KANG
used different approaches for studying GE interaction in unbalanced data (Freeman, 1975; Pedersen et al., 1978; Zhang and Geng, 1986; Gauch and Zobel, 1990; Rameau and Denis, 1992; Piepho, 1994). Usually environmental effects are considered as random and cultivar effects as fixed. Inference on random effects using least squares, in the case of unbalanced data, is not appropriate because information on variation among random effects is not incorporated (Searle, 1987). For this reason, mixed model equations (MME) are recommended (Henderson, 1975). Consider a simple case with ‘‘t” genotypes, “s” environments, and “r” replications. The general linear model for this analysis is (Kang and Magari, 1996)
Y.. v k = /.I.
+ a,+ p, + Yik + Eijk,
(2)
where Yiikis observed trait value, p is grand mean, oli is environmental effect, p, is genotype effect, yik is GE interaction, and eijkis experimental error. Genotypes are considered as fixed and environments as random. GE interaction effect will be considered as random (interaction between a random factor and a fixed factor is random). The response variable has an expected mean of (k + pk) and a variance of (uZe uZge+ u2w), where u2eis environmental variance, uZge is GE interaction variance, and u2wis experimental error variance. For calculating stability variances, Shukla (1972) partitioned GE interaction into ‘‘t’genotype, x environment components. These variances are unbiased and have minimum variance among all possible quadratic unbiased estimators (Shukla, 1972). The values of u? can be negative because they are calculated as the differences of two statistically dependent sums of squares. Although negative variance estimates are not uncommon, this is a negative feature of this approach. Computation of u? is impossible from unbalanced data, but genotype, x environment variance components (uZg(,)Jcan be estimated using the maximum-likelihood approach. The general linear model for randomized complete block design experiments conducted in different environments is
+
= /.I.
+
(yi
+ 8v. . + pk + Yik + Eijk,
(3)
Using matrix notation, Eq. (3) can be written as y = 1/.1. + X p
+ Wa + U0 + Z,Z,a, + E,
(4)
where, y is vector of observed yield data; 1 is vector of ones; X is design matrix for fixed effects (genotypes); p is vector of genotype effects; W and a are, respectively, a design matrix for and a vector of environmental effects; U and 8 are, respectively, a design matrix for and a vector of replications within environment effects; Z, and a, are, respectively, a design matrix for and a vector of genotype by environment interaction effects; and E is the vector of residuals. Equation (4) can be solved using Henderson’s (1975) MME. The levels of random factors are generally assumed to be independent.
USING GE INTERACTION FOR CROP DEVELOPMENT
233
To demonstratethis procedure, an unbalanced subset of Louisiana maize hybrid yield trial data on 11 hybrids grown at four locations (Alexandria, Baton Rouge, Bossier City, St. Joseph) for 4 years (1985-1988) was analyzed. All hybrids were not included in all years and/or locations. Experiments were conducted using arandomized complete-block design with four replications per location. A few experimental plots were discarded during the course of the experiments for various reasons. For calculating phenotypic stability variances, restricted maximum likelihood (REML) using the EM-type algorithm was employed (Patterson and Thompson, 1971).All variances were computed by iterating on MME and tested for H,:u2g(k)e = 0. For the mixed model ( 3 ) (random environments and replications within environment, and fixed hybrids) and unbalanced data, the variances of random effects affect the estimation of fixed effects (hybrid mean yields) (Searle, 1987).For this reason, MME was used, which gave best linear unbiased estimators (BLUE) for hybrids (fixed effects) and best linear unbiased predictors (BLUP) for random effects (Henderson, 1975). Hybrid yields, presented in the form of BLUE (Table V), were the solutions for hybrid means from the MME equations. Hybrid 6 was the highest yielding, although not statistically different (p > 0.23) from hybrids 9 and 3. For stability interpretation, u2g(k,e has the same properties as ui2.It can be regarded as a type 2 stability statistic (Lin ef al., 1986) or as a statistic related to agronomic concept of stability (Becker, 1981). High values of u2g(k)e indicated that genotypes were not stable or that they interacted with the environments. The (Table V) were statisticalIy not different from zero for both 0.05 values of u2g(k)e Table V Best Linear Unbiased Estimators (BLUE) and Stability Variances for Maize Hybrids from Performance Trials in Louisiana”
1
2 3 4 5 6 7 8 9 10 11
6.90 6.39 7.12 6.53 6.61 7.38 6.53 6.81 7.28 6.19 5.68
0.29 0.24 0.28 0.31 0.69 0.29 0.76 1.10h 0.40 0.42 0.46
“Used with permission of CRC Press, Boca Raton, FL. “Statistically different from zero at the 0.05 probability level.
234
MANJIT S. KLWG
and 0.01 probability levels, except u2g(8)e = 1.10 that was statistically different from zero at 0.05 probability level. Kang (1993b) also reported hybrid 8 (SB 1827) to be unstable with a probability greater than 99%. For calculating stability variances, REML was used because it can be effectively adapted to unbalanced data (Searle, 1987). Although the procedure requires data to have a multivariate normal distribution (Patterson and Thompson, 1971), Searle et al. (1992) indicated that in many circumstances this assumption was not seriously incorrect. The method is translation invariant, and REML estimators of u2g(k)e will always be positive if positive starting values are used (Harville, 1977). The REML methodology is generally preferred to maximum likelihood estimates because it considers the degrees of freedom for fixed effects for calculating error. Calculation of REML stability variances for unbalanced data allows one to obtain a reliable estimate of stability parameters, it also overcomes the difficulties of manipulating unbalanced data (Magari and Kang, 1996).
VI. HOW TO EXPLOIT OR MINIMIZE INTERACTION The amount of literature available on GE interaction is a strong indicator that it is present universally and cannot be avoided. Therefore, the best approach for breeders and geneticists would be to understand the nature and causes of GE interaction and to try to minimize its deleterious implications and exploit its beneficial potential through appropriate breeding, genetic, and statistical methodologies (Kang and Gauch, 1996). The following sections outline some of the important strategies for accomplishing this.
A. BREEDINGFOR RESISTANCE-TOLERANCE TO STRESS FACTORS 1. Correct Genetic Cause(s) of GE Interaction Resistance or tolerance to any type of stress, biotic or abiotic, is essential for stable performance (Khush, 1993; Duvick, 1996). Sources of increased crop productivity include enhanced yield potential, heterosis, modified plant types, improved yield stability, gene pyramiding, and exotic and transgenic germplasm (Khush, 1993). It is important to identify the factor(s) that are responsible for GE interaction. Suppose we have two maize genotypes (one resistant and the other susceptible to European corn borer, ECB) that were grown in two environments and that we could attribute differential performance of the genotypes in the two environments (crossover interaction) to ECB damage in one environment and its ab-
USING GE INTERACTION FOR CROP DEVELOPMENT
235
sence in the other environment by using damage ratings as a covariate. To minimize or remove the interaction caused by the ECB damage, a gene confemng resistance to ECB (e.g., the Bt gene) could be inserted in one of the two inbred parents of the susceptible genotype. This could be achieved by traditional backcross breeding, a DNA-based, marker-assisted backcrossing, or transformation. If the causes (e.g., disease resistance-susceptibility) for interaction are traits with monogenic or &genic inheritance, solutions are relatively easy. Breeders must know the inheritance of resistance-tolerance to stresses and have resistant germplasm available. For traits with more complex inheritance (e.g., quantitative traits), population improvement through recurrent selection to counter one or more stress factors may be necessary before parental lines for hybrids or cultivars are developed.
2. Characterize Genotypes and Environments To alleviate GE interaction concerns caused by stresses, breeders need to know as much about the various characteristics of genotypes as possible. They also need to characterize environments (micro and macro) as fully as possible. Knowledge of environmental aspects, such as soil characteristics, ranges of weather variables, and stresses that plant materials will be exposed to, is a prerequisite to exploiting the beneficial potentials of the genotypes and environments and to targeting appropriate cultivars to specific environments.
3. Identify and Select for Molecular Markers Associated with Stable Responses Economically important characters in crop species are generally quantitative in nature. Phenotypic plasticity or GE interactions are under genetic control. For improving quantitative traits, breeders need to know what genetic factors are involved, where they are located, and what type of inheritance they exhibit. Recent advances in molecular genetics have provided some of the best tools for obtaining a better insight into the molecular mechanisms associated with GE interactiona “trait” with relatively low heritability. One of the best ways to manipulate quantitative traits with low heritability is to employ molecular markers, such as restriction-fragment length polymorphisms (RFLPs), and determine their association with quantitative trait loci (QTL). These developments have paved the way for investigating QTL-by-environment interaction (QxE) (Beavis and Keim, 1996), which will ultimately provide a better genetic understanding and possible regulation of this phenomenon. Regions of plant genomes that provide stable responses across diverse environments can be identified by determining linkage of QTL to the abundance of RFLPs, which should make possible for breeders to manipulate QTL in the same fashion as single genes that control qualitative traits. This
236
i"JIT
S. KANG
could substantially reduce breeding and evaluation time, as marker-assisted selection for desirable QTL would be more reliable than phenotypic selection for quantitative traits. Greater genetic progress can be achieved by using Mendeliantype markers to map QTL. For many years, geneticists have sought methods by which simply inherited markers could be used to select for linked qualitative and quantitative traits. The use of RFLPs is ideal, as they are easily identifiable,codominant in expression, distributed throughout the genome, and do not affect phenotype. Other molecular markers may also be suitable for this purpose (see Lee, 1995). It is highly desirable to identify QTL for a complex trait (e.g., high yield) that are expressed in a number of environments. I propose an approach for analyzing GE interactions that integrates quantitative genetic and molecular genetic methodologies: Let us suppose that the quantitative trait under investigation is grain yield in maize. The GE issue would need to be addressed at the parental (inbred line) level, not at the hybrid level. At present, private seed companies are the major developers of inbred lines for commercial hybrid production. Let us say that they produced a diallel cross involving 11 inbred lines (55 F, crosses) and that they evaluated performance of the F, crosses at I0 locations over three years (30 environments). A highly significant GE interaction was detected for grain yield and yield components. General and specific combining ability effects for the 11 inbred lines were estimated. Suppose GCA was more important than SCA. Now, assign stability-variance values to all F, crosses. Select one parent with the largest positive GCAeffect and highest stability (i.e., consistent association with F, crosses that had low or nonsigniJcant stability-variance) and the other parent with opposite characteristics (the largest negative GCA effect and a consistent association with F, crosses with signiJcant stability-variance). Develop F, and F, generations for this cross. Grow the F, and F, progeny at several locations. Take a representative sample at each location. Collect, on an individual plant basis, DNA and data for yield and yield components. Identify RFLP markers associated with putative QTL for high yield and low yield (the same thing should be done with any yield component). Repeat the test for two more years. This, in essence, would represent simultaneous selection for yield (or yield component) and stability. With the increasing impact of molecular biology on breeding and genetics, we are at the dawn of a field that can be called molecular plant breeding. Molecular approaches are being incorporated at various levels. For example, the traditional field of cytogenetics has been transformed into molecular cytogenetics. I expect that molecular biology (including molecular genetics, biochemistry, and plant physiology) will play an enhanced role in breeding crop species and overcoming the constraints imposed on genotypes by their interaction with environmental factors. For example, cloning of genes for cold tolerance obtained from cold-tolerant plant species and inserting these genes into cold-sensitive crop species could overcome stress imposed by cold climate on the latter. We might expect differential expressivities of single genes in different genetic backgrounds.
USING GE INTERACTION FOR CROP DEVELOPMENT
237
B. BREEDINGFOR STABILITY-RELIABILITY OF PERFORMANCE Evans (1993) pointed to the need for developing new cultivars with broad adaptation to a number of diverse environments (selection for adaptability) and the need of farmers to use new cultivars with reliable or consistent performance from year to year (reliability), Smith et al. (1990) pointed out that genetic improvement for low input conditions would require capitalizing on GE interactions; furthermore, slower or limited gains in low input or stress environments suggested that conventional high input management of breeding nurseries and evaluation trials might not effectively select genotypes with improved performance under low input levels. Rosielle and Hamblin (1981) examined theoretical aspects of selection for yield in stress and nonstress environments. They showed that selection for tolerance to stress generally reduced mean yield in nonstress environments and that selection for mean productivity generally increased mean yields in both stress and nonstress environments. Bramel-Cox (1996) reviewed relevant literature on breeding for reliability of performance in unpredictable environments. To be reliable, a stability statistic must be based on a large number of environments (>lo). Information on stability is usually available in the final stages of a breeding program when replicated tests can be conducted. From the standpoint of individual growers, stability across years (temporal) is most important. However, the average life of a commercial cultivar is less than 10 years. Therefore, it would be impractical to recommend cultivars to growers on the basis of 10 years’ data. However, a breeder could test cultivars or lines for 10to 15 years and identify those that have temporal stability. Crosses could then be made among the most stable cultivars to develop source material (germplasm) that would be utilized for developing inbred lines or pure lines. Therefore, extensive cultivar testing across years is a precursor to cultivar development. Progress from selection would depend on the heritability or repeatability of the stability statistic used. In general, heritability or repeatability of stability statistics for yield is low (Eagles and Frey, 1977; Becker and Leon, 1988; Pham and Kang, 1988). Heritability may be improved by increasing the number of test environments. Stability of cultivars would be enhanced if multiple resistances-tolerances to stress factors were incorporated into the germplasm used for cultivar development. Recurrent selection should be practiced to improve germplasm for quantitative traits. If every cultivar (different genotypes) possessed equal resistance-tolerance to every major stress encountered in diverse target environments, GE interaction would be reduced. Conversely, if genotypes possessed differential levels of resistance (a heterogeneous group) and, somehow, we could render all target environments as homogeneous as possible, GE interaction would again be reduced. Since we do not have any control over unpredictable environments from year to year, the best approach would be the former, i.e., empowering genotypes with attributes, such as stress resistance-tolerance. This should help minimize GE interaction and increase stability of performance.
238
MANJIT S. KANG
Stability analyses can be used to identify durable resistance to disease pathogens (Jenns et al., 1982). If a cultivar-by-pathogen isolate interaction exists, it would be necessary to identify a cultivar that has general resistance instead of specific resistance. Another strategy would be to employ indirect selection for the trait of interest. For example, it may be easier to select for stability of a major component of a complex trait. If seed size were a more stable trait than yield, it might be worthwhile to improve stability for seed size. Eberhart and Russell (1 969) reported that twoeared maize cultivars had lower regression coefficients and deviations from regression (i.e., more stable) than did single-eared maize. A major trait usually represents the final phenotypic expression of a complex developmental process during growth, which can be investigated via the path analysis (Tai, 1990). Investigation of sequential relationship between yield and yield components could provide insights into GE interaction (Tai, 1990). Kang et al. (1987a) considered it worthwhile to examine whether stability of one trait was correlated with stability of another trait. If stability (stability variance, ecovalence, or any other stability statistic) of two traits were reasonably positively correlated, concurrent selection for stabilities of the two traits might be possible. They found that for both plant-cane (PC) and ratoon (RT) crops of sugarcane, stability of tons per hectare of sugar (THS) could be predicted from stability of tons per hectare of cane (THC) ( r s = 0.78** to 0.88** for combined PC and RT for two series of clones), and to a lesser extent, from stability of stalk number (rs= 0.73** to 0.77** for combined PC and RT for two series of clones). Stability of THC could be reasonably well predicted from stability of stalk number (rs= 0.78** to 0.88** for combined PC and RT for two series of clones). Stalk number is much easier to measure than THC or THS. For indirect selection to be effective, the following requirements should be met (Sherrard et al., 1985; Jones, 1992): Yield component or biochemical or physiological trait must be easier to assess than yield itself. A causal correlation must exist between the character and yield in the field. There must be heritable variation for the character. Screening test should be simple, accurate, economical, and rapid-preferably capable of being used at seedling stage at any time of year. Attributes that can be screened include: morphological and anatomical (e.g., plant height, leaf size, or stornatal frequency), compositional (e.g., protein or lysine content, ABA content as a test for drought tolerance), process rates (e.g., photosynthesis, respiration, or vernalization), and process control (e.g., enzyme activity or stomatal aperture and its behavior). Biochemical bases of plant breeding have been discussed (Neyra, 1985, 1986). Various authors in the volumes edited by Neyra (1985, 1986) presented a comprehensive survey of progress and knowl-
USING GE INTERACTION FOR CROP DEVELOPMENT
239
edge of biochemical processes with greater potential for the development of superior cultivars; for example, photosynthesis, photorespiration, nitrate assimilation, biological nitrogen fixation, and starch and protein syntheses.
C. MEASURE INTERACTION AT INTERMEDIATE GROWTH STAGES Acrop is exposed to variable environmental factors throughout the growing season. Generally, researchers investigate the causes of GE interaction at the final harvest stage. It may be critical to investigate GE interaction by subdividing the entire growing season (say, 120 days) into two intervals of 60 days each. Environmental variables should be recorded and growth measurements taken at the end of each period. This would help determine what effect, if any, the environmental variables from an earlier period had on the final yield. Could they explain GE interaction encountered at the final harvest? Another useful strategy would be to record yield at one or two intermediary stages to determine the stage at which GE interaction is first encountered and/or is most pronounced. This could provide a better understanding of the dynamic process of yield formation.
D. EARLY MULTIENWRONMENT TESTING Usually, there is a shortage of seed at the earliest stages of breeding, which prevents extensive testing. However, in a clonally propagated crop, such as sugarcane or potato, one stalk of sugarcane or one tuber of potato can be divided into at least two pieces and planted in more than one environment. Similarly, in other crops, if one has 20 kernels, one could plant 10 seeds each in two diverse environments. In the absence of a GE interaction, one would obtain a better evaluation of the genotypes, but if GE interaction were present, one would obtain information about consistency or inconsistency of performance of genotypes early in the program. This strategy would prevent gene loss or genetic erosion that could occur if testing were done only in one environment and would also result in an increased breeding effort without a corresponding increase in expenditure of resources.
E. OPTIMAL RESOURCE ALLOCATION Genotype-by-environment interaction can be employed to judiciously allocate resources in a breeding program (Pandey and Gardner, 1992; Magari et al., 1996). Carter et al. ( I 983) estimated that at a low level of treatment x environment interaction (10% of error variance), testing in at least two environments was necessary
240
MANJIT S. KANG
to detect treatment differences of 20%, and it required at least seven environments to detect smaller ( 10%) treatment differences for growth analysis experiments in soybean. With a larger magnitude of interaction, a larger number of environments would be needed for a given level of precision in treatment differences. Magari et al. (1996) used multienvironment (different planting dates) data for ear moisture loss rate in maize that exhibited planting date x genotype interaction. Differences among hybrids depended on planting dates. Relative efficiency for the benchmark protocol (1 1 plants per replication, three replications, and three planting dates) was regarded as the reference value (100%).The relative efficiency for five plants per plot in four replications and three planting dates was equivalent to that for the benchmark protocol. A relative efficiency of 100% also could be achieved with a sample of four planting dates, three replications, and three to four plants per plot. When the number of replications was increased to four in each of four planting dates, only two plants per plot were needed to achieve a relative efficiency of 100%. The number of planting dates (environments) was found to be a critical factor in determining the precision of an experiment. In forest tree crops, Matheson and Cotterill (1990) determined losses of potential gain as a result of GE interaction as
c = I-[(&,
+ $)”/
(u2C
+
(T&
+ u2)’/.],
where C is loss of potential gain, u * is~genetic variance component, u2GE is GE variance component, and u2 is error variance. Matheson and Raymond (1 984) described several ways in which effects of GE interaction on breeding programs could be estimated.
VII. CONCLUSIONS The primary goal of plant breeding programs is to develop productive cultivars. In the past, GE interactions have been shown to contribute to yield increases (Simmonds, 1981). Currently, integration of DNA-based markers with traditional plant breeding methodology provides a powerful selection tool. Genotype-by-environment interaction is and should be regarded as a genetic-breeding issue, which can be exploited via breeding and molecular genetic techniques. Identification of genomic regions associated with stability of performance across target environments would help breeders understand and possibly regulate the GE interaction phenomenon. Breeders must maintain, develop, and utilize germplasm of a broad genetic base for cultivar development. A lack of GE interaction can imply a lack of genetic diversity, which can be disastrous because of associated genetic vulnerability of a crop to disease epidemics, insect infestations, or other factors. A lack of GE inter-
USING GE INTERACTION FOR CROP DEVELOPMENT
241
action can also imply that the environments in which diverse genotypes were evaluated were more or less homogeneous. Genotype-by-environment interactions may be minimized by modifying genetic constitutions of cultivars, i.e., confemng upon them resistance-tolerance to different stresses to which they would likely be exposed. Various biotic and abiotic stresses have been implicated as causes of GE interactions, since gene expression is influenced by fluctuations in environmental variables. Causes of GE interaction must be identified on a cultivar-by-cultivarbasis, as precisely as possible. Once this has been accomplished, efforts need to be directed toward improving cultivars through “gene therapy.” The greater the number of biotic-abiotic stresses cultivars grown in a region are resistant to, the more stable-reliable their performance would be. Stresses provide opportunities for identifying and selecting genotypes that are efficient users of suboptimal levels of inputs or tolerant of superoptimal levels of inputs. Because an emphasis on stability during the selection process would reduce harmful type I1 errors, it is essential to incorporate stability of performance during all phases of selection. When crossover GE interactions are encountered, cultivars must be categorized as “specifically adapted” or “broadly adapted” and targeted according to suitable environments. Simultaneous selection for yield and stability of performance is a desirable goal. Two of the most suitable methods of simultaneously selecting for yield and performance stability are Huhn’s Si3 and Kang’s rank-sum (1988) statistics. Genotype-by-environmentinteractions offer opportunities for breeders to plan breeding programs, allocate resources efficiently, and identify molecular markers associated with stable cultivar performance. In the future, greater emphasis would need to be placed on understanding signal transduction in plants in response to environmental stresses, which should lead to a better understanding of the relationship between crop performance and environment. Important biochemical or physiological pathways in plants are mediated by enzymes that are the products of transcription and translation of specific genes. Genotype-by-environmentinteraction should be investigated at physiological or biochemical levels. Therefore, cooperation among breeders, geneticists, molecular biologists, biochemists, physiologists, and statisticians is deemed essential.
REFERENCES Aastveit, A. H., and Mejza, S. (1992). A selected bibliography on statistical methods for the analysis of genotype x environment interaction. Biuletyn Oceny Odmian 25,83-97. Allard, R. W. (1960). “Principles of plant breeding.”John Wiley & Sons, New York. Allard, R. W., and Bradshaw, A. D. (1964). Implications of genotype-environment interactions in applied breeding. Crop Sci. 4,503-508.
242
MANJIT S. KANG
Andrews, D. J., and Kumar, K. A. (1992). Pearl millet for food, feed, and forage. Adv. Agron. 48, 89-139. Annicchiarico, P., and Perenzin, M. (1994). Adaptation patterns and definition of macro-environments for selection and recommendation of common wheat genotypes in Italy. Plant Breeding 113, 197-205. Asada, K., and Takahashi, M. (1987). Production and scavenging of active oxygen in photosynthesis. In “Photoinhibition” (D. Kyle, C. Osmond, and C. Arntzen, eds.), pp. 227-287. Elsevier, Amsterdam. Bachmann, K. (1993). Effects of stress on the genome. In “Plant adaptation to environmental stress” (L.Fowden, T. Mansfield, and J. Stoddart, eds.), pp. 275-292. Chapman & Hall, New York. Bachmann, K., Chambers, K. L., and Price, H. J. (1985). Genome size and natural selection: Observations and experiments in plants. In “The evolution of genome size” (T.Cavalier-Smith, ed.), pp. 267-276. Wiley, Chichester, UK. Baker, R. J. (1984). Quantitative genetic principles in plant breeding. In “Gene manipulation in plant improvement” (J. P. Gustafson, ed.), pp. 147-176. Plenum Press, New York. Baker, R. J. (1988). Differential response to environmental stress. In “Proceedings of the Second International Conference on Quantitative Genetics” (B. S Weir, E. J. Eisen, M. M. Goodman, and G. Namkoong, eds.), pp. 492-504. Sinauer, Sunderland, MA. Baker, R. J. (1990). Crossover genotype-environmental interaction in spring wheat. In “Genotype-byenvironment interaction and plant breeding” (M. S. Kang,ed.), pp. 42-51. Louisiana State Univ. Agric. Center, Baton Rouge, LA. Beavis, W. D., and Keim, P. (1996). Identification of quantitative trait loci that are affected by environment. In “Genotype-by-environment interaction” (M. S. Kang and H. G. Gauch, Jr., eds.), pp. 123-149. CRC Press, Boca Raton, FL. Becker, H. C. (198 I). Correlations among some statistical measures of phenotypic stability. Euphyricu 30,835-840. Becker, H. C., and Leon, J. (1988). Stability analysis in plant breeding. Planr Breeding 101, 1-23. Bennett, M. D. (1973). Nuclear characters in plants. Bmokhaven Symp. Biol. 25,344-366. Bennett, M. D. (1987). Variation in genomic form in plants and its ecological implications. New Phyrol. 106 (Suppl.), 177-200. Blum, A. (1988). “Plant breeding for stress environments.” CRC Press, Boca Raton, FL. Bomman, J. F., and Sundby-Emanuelson, C. (1995). Response of plants to UV-B radiation: Some biochemical and physiological effects. In “Environment and plant metabolism” (N. Smirnoff, ed.), pp. 245-262. Bios Publ., Oxford, UK. Borojevic, S. (1990). “Principles and methods of plant breeding.” Elsevier, Amsterdam. Bradshaw, A. D. (1965). Evolutionary significance of phenotypic plasticity. Adv. in Genet. 13,115153. Bramel-Cox, P. J. (1996). Breeding for reliability of performance across unpredictable environments. In “Genotype-by-environment interaction” (M. S. Kang and H. G. Gauch, Jr., eds.), pp. 309-339. CRC Press, Boca Raton. FL. Brandon, R.N. (1990). “Adaptation and environment.” Princeton Univ. Press, Princeton, NJ. Bucher, M., and Kuhlemeier, C. (1993). Long-term anoxia tolerance. Multi-level regulation of gene expression in amphibious plant Acorus calumus L. Plant Physiol. 1 0 3 , 4 4 1 4 8 . Burke, J. J., Hatfield, J. L., Klein. R. R., and Mullet, J. E. (1985). Accumulation of heat-shock proteins in field grown cotton. Plant Physiol. 78,394398. Bums, G. W. (1969). “The science of genetics.” 2nd ed. Macmillan, New York. Busey, P. (1983). Management of crop breeding. In “Crop breeding” (D.R. Wood, K. M. Rawal, and M. N. Wood, eds.), pp. 31-54. h e r . SOC. Agron. and Crop Sci. SOC.of Am., Madison, WI. Byme, T., Drennan, D., and Harding, J. (1987). Genotype x environment interaction bias in heritability estimates for cut-flower yield of greenhouse gerberas. J. Am. SOC.Hort. Sci. 112,724727.
USING GE INTERACTION FOR CROP DEVELOPMENT
243
Byth, D. E., and Mungomery, V. E. (eds.) (1981). “Interpretation of plant response and adaptation to agricultural environments.” Aust. Inst. Agric. Sci., Queensland Branch, Australia. Caligari. P. D. S. (1993). GxE studies in perennial tree crops: Old, familiar friend or awkward, unwanted nuisance. In “Genotype-environment interaction studies in perennial tree crops” (V. Rao, I. E. Henson, and N. Rajanaidu, eds.), pp. 1-1 I. Int. SOC.for Oil Palm Breeders, and Palm Oil Res. Instit. of Malaysia, Kuala Lumpur, Malaysia. Carmer. S. G . (1976). Optimal significance levels for application of the least significant difference in crop performance trials. Crop Sci. 16,95-99. Carter, T. E., Jr., Burton, J. W., Cappy. J. J., Israel, D. W., and Boerma, H. R. (1983). Coefficients of variation, error variances, and resource allocation in soybean growth analysis experiments. Agron. J. 75,691496. Cavalier-Smith, T. (1985a). Eucaryote gene number, non-coding DNA and genome size. In “The evolution of genome size” (T. Cavalier-Smith, ed.), pp. 69-103. Wiley, Chichester, UK. Cavalier-Smith, T. (198%). Cell volume and the evolution of eukaryotic genome size. In “The evolution of genome size” (T. Cavalier-Smith, ed.), pp. 105-184. Wiley, Chichester, UK. Ceccarelli, M. Falistocco, E., and Cionini, P. G. (1992). Variation of genome size and organization within hexaploid Festuca arundinacea. Theor Appl. Genet. 83,273-270. Ceccarelli, S . , Erskine, W., Hamblin, J., and Grando, S. (1994). Genotype by environment interaction and international breeding programmes. Expl. Agric. 30, 177-1 87. Chen, Z., Silva, H., and Klessig, D. (1993). Active oxygen species in the induction of plant systemic acquired resistance by salicylic acid. Science 262, 1883-1 886. Cheng, C. L., Dewdney, J.. Kleinhofs, A,, and Goodman, H. M. (1986). Cloning and nitrate induction of nitrate reductase mRNA. Pmc. Nat. Acad. Sci. USA 83,6825-6828. Chrispeels, M. J. (ed.). (1994). Introduction to “Signal transduction in plants: A collection of updates.” Am. SOC.Plant Physiologists, Rockville, MD. Clark, R. B., and Duncan, R. R. (1993). Selection of plants to tolerate soil salinity, acidity, and mineral deficiencies. In “International crop science I” (D. R. Bruxton, R. Shibles, R. A. Forsberg, B. L. Blad, K. H. Asay, G. M. Paulsen, and R. F. Wilson, eds.), pp. 371-379. Crop Sci. SOC.of America, Madison, WI. Comstock, R. E., and Moll, R. H. (1963). Genotype-environment interactions. I n “Statistical genetics and plant breeding” (W. D. Hanson and H. F. Robinson, eds.), pp. 164-194. NAS-NRC Pub. 982, Washington, D.C. Conte, F. (1973). “Biochemical adaptation.” Univ. of Chicago Press, Chicago. Cooper, M., and Hammer, G. L. (eds.) (1996). “Plant adaptation and crop improvement.” CAB International, Wallingford, U.K., ICRISAT, Patancheru, India, and IRRI, Manila, Philippines. Cornelius, P. L., Crossa, J., and Seyedsadr, M. S. (1996). Statistical tests and estimates of multiplicative models for GE interaction. In “Genotype-by-environment interaction” (M. S. Kang and H. G. Gauch, Jr., eds.), pp. 199-234. CRC Press, Boca Raton, FL. Crossa, J. (1990). Statistical analyses of multilocation trials. Adv. Agron. 44,55-85. Crossa, J., Cornelius, P. L., and Seyedsadr, M. S. (1996). Using the shifted multiplicative model cluster methods for crossover GE interaction. In “Genotype-by-environment interaction” (M. S. Kang and H. G. Gauch, Jr., eds.), pp. 175-198. CRC Press, Boca Raton, FL. Crow, J. F. (1986). “Basic concepts in population, quantitative, and evolutionary genetics.” W. H. Freeman, New York. Cullis, C . A. (1990). DNA rearrangements in response to environmental stress. In “Advances in genetics” (J. G. Scandalios and T. R. F. Wright, eds.), pp. 73-97. Academic Press, New York. Cushman, J. C., De Rocher, E. J., and Bohmert, H. J. (1990). Gene expression during adaptation to salt stress. In “Environmental injury to plants” (F.Katterman, ed.), pp. 173-203. Academic Press, San Diego, CA.
244
MANJIT S. KANG
Dashiell, K. E., Ariyo, 0.1. and . Bello, L. (1994). Genotype x environment interaction and simultaneous selection for high yield and stability in soybeans (Glycine m a (L.) Men.). Ann. Appl. Biol. 124,133-139. Dawkins, R. (1978). “The selfish gene.” Granada Publishing Ltd., London. Dean, A. M. (1995). A molecular investigationof genotype by environmentinteractions.Genetics 139, 19-33. DeLacy, I. H., Cooper, M., and Basford, K. E. (1996). Relationships among analytical methods used to study genotype-by-environment interactions and evaluation of their impact on response to selection. In “Genotype-by-environment interaction” (M. S. Kang and H. G. Gauch, Jr., eds.), pp. 51-84. CRC Press, Boca Raton, FL. Deng, M. D., Moureaux, T., Cherel, I., Boutin, J. P., and Caboche, M. (1991). Effects of nitrogen metabolism on the regulation and circadian expression of tobacco nitrate reductase. Plant Physiol. Biochem. 29,239-247. Denis, J.-B. (1988). Two-way analysis using covariates. Stafistics19, 123-132. Denis, J.-B., and Gower, J. C. (1996).Asymptotic confidence regions for biadditive models: Interpreting genotype-environment interactions.Appl. Statist. 45,479493. Denis, J.-B., Gauch, H. G., Jr., Kang, M. S., Van Eeuwijk, E A., and Zobel, R. W. (1996). Bibliography on genotype-by-environment interaction. In ‘Genotype-by-environmentinteraction” (M. S. Kang and H. G. Gauch, Jr., eds.), pp. 405409. CRC Press, Boca Raton, FL. Devine, T. E. (1982). Genetic fitting of crops to problem soils. In “Breeding plants for less favorable environments” (M. N. Christensen and C. F. Lewis, eds.), p. 143. Wiley Interscience, New York. Dickerson, G. E. (1 962). Implicationsof genetic-environmentalinteraction in animal breeding. Animal Prod. 4,4744. Doolittle, D. P. ( I 987). “Population genetics: Basic principles.” Springer-Verlag, New York. Duvick, D. N. (1992). Genetic contributions to advances in yield of U. S. maize. Maydicu 37,69-79. Duvick, D. N. (1996). Plant breeding, an evolutionary concept. Crop Sci. 36,539-548. Eagles, H. A,, and Frey, K. J. (1977). Repeatability of the stability-varianceparameter in oats. Crop Sci. 17,253-256. Eherhart, S. A., and Russell, W. A. (1966). Stability parameters for comparing varieties. Crop Sci. 6, 3640. Eberhart, S . A., and Russell, W. A. (1969). Yield stability for a 10-line diallel of single-cross and double-cross maize hybrids. Crop Sci. 9,357-361. Einhellig, F. A. (19%). Interactionsinvolving allelopathy in cropping systems. Agron. J. 88,886893. Eisemann, R. L., Cooper, M., and Woodruff, D. R. (1990).Beyond the analytical methodology, better interpretationand exploitation of GE interaction in plant breeding. In “Genotype-by-environment interaction and plant breeding” (M. S. Kang, ed.), pp. 108-1 17. Louisiana State Univ. Agric. Center, Baton Rouge, LA. Epstein, E. (1972). “Mineral nutrition of plants: Principles and perspectives.” John Wiley, New York. Epstein, E. (1976). Genetic potential for solving problems of soil mineral stress: Adaptation of crops to salinity. In “Plant adaptation to mineral stress in problem soils” (M. J. Wright, ed.), pp. 73-82. Cornell Univ. Agric. Experiment Station, Ithaca, NY. Eskridge, K. M. (1996). Analysis of multiple environment trials using the probability of outperforming a check. In “Genotype-by-environment interaction” (M. S. Kang and H. G. Gauch, Jr., eds.), pp. 273-307. CRC Press, Boca Raton, FL. Evans, L. T. (1993). “Crop evolution, adaptation and yield.” Cambridge Univ. Press, New York. Falconer, D. S. (1952). Selection for large and small size in mice. J. Genet. 51,470-501. Falconer, D. S. (1981). “Introduction to quantitative genetics.” 2nd ed. Longman Press, London. Fernandez, G. C. J. (1991). Analysis of genotype x environment interactionby stability estimates. Hort. Science 26,947-950.
USING GE INTERACTION FOR CROP DEVELOPMENT
245
Finlay, K. W., and Wilkinson, G. N. (1963). The analysis of adaptation in a plant breeding programme. Aust. J. Ag,: Rex 14,742-754. Flavell, R. B. (1980). The molecular characterization and organization of plant chromosomal DNA. Ann. Rev. Plant Physiol. 31,569-596. Flavell, R. B. (1982). Amplification, deletion and rearrangement: Major sources of variation during species divergence. In “Genome evolution” (G. A. Dover and R. B. Flavell, eds.), pp. 301-324. Academic Press, New York. Freeman, G. H. (1975). Analysis of interactions in incomplete two-way tables. Applied Statistics 24, 46-55. Frey, K. J. ( I 990). Mass selection. In “Genotype-by-environment interaction and plant breeding” (M. S. Kang, ed.), pp. 28-39. Louisiana State Univ. Agric. Center, Baton Rouge, LA. Galangau, F., Daniel-Vedele, F., Moureaux, T., Dorbe, M. F., Leydecker, M. T., and Caboche, M. (1988). Expression of leaf nitrate reductase genes from tomato and tobacco in relation to lightdark regimes and nitrate supply. Plant Physiol. 88,383-388. Gauch, H. G., Jr. (1992). “Statistical analysis of regional yield trials: AMMI analysis of factorial designs.” Elsevier, Amsterdam. Gauch, H. G., Jr. and Zobel, R. W. (1990). Imputing missing yield trial data. Theon Appl. Genet. 70, 753-76 1. Gauch, H. G., Jr., and Zobel, R. W. (1996). AMMI analysis of yield trials. In “Genotype-by-environment interaction” (M. S. Kang and H. G. Gauch, Jr., eds.), pp. 85-122. CRC Press, Boca Raton, FL. Gibson, P. T. (1 996). Correcting for inbreeding in parent-offspring regression estimates of heritability with non-additive and genotype x environment effects present. Crop Sci. 36,594-600. Glaz, B., and Dean, J. L. (1988). Statistical error rates and their implications in sugarcane clone trials. Agron. J . 80,560-562. Glaz, B., Miller, J. D.. and Kang, M. S. (1985). Evaluation of cultivar-testing locations in sugarcane. Theo,: Appl. Genet. 71,22-25. Gorman, D. P., Kang, M. S., and Milam, M. R. (1989). Contribution of weather variables to genotype x environment interaction in grain sorghum. Plant Breeding 103,299-303. Graham, R. D. (1982). Breeding for nutritional characteristics in cereals. Adv. Plant Nut,: 1,57. Gregorius, H. R., and Namkoong, G. (1986). Joint analysis of genotypic and environmental effects. Theo,: Appl. Genet. 72,4 13422. Gravois, K. A., Moldenhauer, K. A. K., and Rohman, P. C. (1990). GE interaction for rice yield and identification of stable, high-yielding genotypes. In “Genotype-by-environment interaction and plant breeding” (M. S. Kang, ed.), pp. 181-188. Louisiana State Univ. Agric. Center, Baton Rouge, LA. Gupta, A. G.,and Lewontin, R. C. (1982). A study of reaction norms in natural populations of Drosophila pseudoobscura. Evolution 36,934-948. Hallauer, A. R. ( I 988). Genotype-environment interaction. In “Proceedings of the Second International Conference on Quantitative Genetics” (B. S. Weir, E. J. Eisen, M. M. Goodman, and G. Namkoong, eds.), pp. 488-491. Sinauer, Sunderland, MA. Harville, D. A. (1977). Maximum-likelihood approaches to variance component estimation and to related problems. J. Arne,: Stat. Assoc. 72,320-340. Helms, T. C. (1993). Selection for yield and stability among oat lines. Crop Sci. 33,423426. Henderson, C.R. (1975). Best linear unbiased estimation and prediction under a selection model. Biometrics 31,423A7. Hendnx, S. D. (1979). Compensatory reproduction in biennial herb following insect defoliation. Oecologia 42,107-118. Hendry, G. A. F., and Brocklebank, K. J. (1985). Ion-induced oxygen radical metabolism in waterlogged plants. New Phytol. 101, 199-206.
246
MANJIT S. KANG
Herrera-Estrella, L., and Simpson, J. (1990). Influence of environmental factors on photosynthetic genes. In “Advances in genetics” (J. G. Scandalios and T. R. F. Wright, eds.), pp. 133-163. Academic Press, New York. Higley, L. G., Browde. J. A., and Higley, P. M. (1993). Moving toward new understandings of biotic stress and stress interactions. In “International crop science I” (D. R. Bruxton, R. Shibles, R. A. Forsberg, B. L. Blad, K. H. Asay, G. M. Paulsen, and R. F, Wilson, eds.), pp. 749-754. Crop Sci. Soc. of Am., Madison, W1. Hoffmann, A. A,, and Parsons, P. A. (1991). “Evolutionary genetics and environmental stress.”Oxford Univ. Press, New York. Howarth, C. J. (1991). Molecular responses of plants to an increased incidence of heat shock. Plant Cell Environ. 14, 831-841. Hughes, M. A,, Dunn, M. A., Zhang, L., Pearce, R. S., Goddard, N. J., and White, A. J. (1993). Longterm adaptation and survival. In “Plant adaptation to environmental stress” &. Fowden. T. Mansfield, and J. Stoddart, eds.), pp. 251-262. Chapman and Hall, New York. Huhn, M. (1996). Nonparametric analysis of genotype x environment interactions by ranks. In “Genotype-by-environment interaction” (M. S. Kang and H. G. Gauch, Jr., eds.), pp. 235-271. CRC Press, Boca Raton, FL. Hutchinson, J., Rees, H.. and Seal, A. G. (1979). An assay of the activity of supplementary DNA in Lolium. Heredity 43,411-42 1. Jenns, A. E., Leonard, K. J., and Moll, R. H. ( I 982). Stability analyses for estimating relative durability of quantitative resistance. Theor: Appl. Genet. 63, 183-192. Jensen, N. F. (1983.) Crop breeding as a design science. In “Crop breeding” (D. R. Wood, K. M. Rawal, and M. N. Wood, eds.), pp. 21-29. Amer. SOC.Agron., and Crop Sci. SOC.Am., Madison, “1. Jensen, N. F. (1988.) “Plant breeding methodology.” John Wiley & Sons, New York. Johnson, J. J., Alldredge, J . R., Ullrich, S. E., and Dangi, 0. (1992). Replacement of replications with additional locations for grain sorghum cultivar evaluation. Crop Sci. 32,4346. Jones, H. G. (1992). “Plants and microclimate: A quantitative approach to environmental plant physiology.” 2nd ed. Cambridge Univ. Press, Cambridge, UK. Kaiser, W. M., and Forster, J. (1989). Low CO, prevents nitrate reduction in leaves. Plant Physiol. 91, 970-974. Kang, M. S. (1988). Arank-sum method for selecting high-yielding, stable corn genotypes. Cereal Res. Comrnun. 16, 113-115. Kang, M. S. (ed.) (1990). “Genotype-by-environment interaction and plant breeding.” Louisiana State Univ. Agric. Center, Baton Rouge, LA. Kang, M. S. (1993a.) Issues in GE interaction. In “Genotype-environment interaction studies in perennial tree crops” (V. Rao, I. E. Hanson, and N. Rajanaidu, eds.), pp. 67-73. Palm Oil Res. Inst. of Malaysia, Kuala Lumpur, Malaysia. Kang, M. S. (1993b). Simultaneous selection for yield and stability in crop performance trials: Consequences for growers. Agron. J. 85,754-757. Kang, M. S. (1994). “Applied quantitative genetics.” M. S . Kang Publisher, Baton Rouge, LA. Kang, M. S., and Gauch, H. G.,Jr. (eds.) (1996). “Genotype-by-environment interaction.” CRC Press, Boca Raton, FL. Kang, M. S., Glaz, B., and Miller, J. D. (1987a). Interrelationships among stabilities of important agronomic traits in sugarcane. Theor: Appl. Genet. 74,310-3 16. Kang, M. S., and Gorman, D. P.(1989). Genotype x environment interaction in maize. Agmn. J. 81, 662-664. Kang, M. S., Gorman. D. P., and Pham, H. N. (1991). Application of a stability statistic to international maize yield trials. Theor: Appl. Genet. 81, 162-165. Kang, M. S., Harville, B. G.. and Gorman, D. P. (1989). Contribution of weather variables to genotype x environment interaction in soybean. Field Crops Res. 21,297-300.
USING GE INTERACTION FOR CROP DEVELOPMENT
247
Kang, M. S., and Magari, R. (1995). STABLE: Basic program for calculating yield-stability statistic. Agron. J. 87,276-277. Kang, M. S., and Magari, R. (1996). New developments in selecting for phenotypic stability in crop breeding. In “Genotype-by-environment interaction” (M. S. Kang and H. G. Gauch, Jr., eds.), pp. 1-14. CRC Press, Boca Raton, FL. Kang, M. S., and Martin, F. A. (1987). Areview of important aspects of genotype-environmental interactions and practical suggestions for sugarcane breeders. J. Am. SOC.Sugar Cane Technol.7,3638. Kang, M. S., and Miller. J. D. (1984). Genotype x environment interactions for cane and sugar yield and their implications in sugarcane breeding. Crop Sci. 24,435-440. Kang, M. S., Miller, J. D., and Darrah, L. L. (1987b). A note on relationship between stability variance and ecovalence. J. Hered. 78, 107. Kang, M. S., Miller, J. D., and Tai, P. Y. P. (1984). Clonal and individual repeatability of agronomic traits in sugarcane. J. Am. SOC.Sugar Cane Techno. 3, 22-27. Kang, M. S., and Pham, H. N. (1991). Simultaneous selection for high yielding and stable crop genotypes. Agron. J. 83, 161-165. Key, J. L., and Kosuge, T. (1984). “Cellular and molecular biology of plant stress.” Liss, New York. Khush, G. S. (1993). Breeding rice for sustainable agricultural systems. In “International Crop Science I” (D. R. Buxton, R. Shibles, R. A. Forsberg, B. L. Blad, K. H. Asay, G. M. Paulsen, and R. F. Wilson, eds.), pp. 189-199. Crop Sci. SOC.Am., Madison, WI. Kimpel, J. A,, and Key, J. L. (1985). Presence of heat-shock mRNAs in field grown soybeans. Plant Physiol. 79, 672-678. Knight, R. ( I 973). The relation between hybrid vigour and genotype-environment interactions. Theor: Appl. Genet. 43,3 1 1-3 18. Lande, R.(1976). Natural selection and random genetic drift in phenotypic evolution. Evolution 30, 3 14-334. Lande, R. (1979). Quantitative genetic analysis of multivariate evolution applied to brain: Body size and allometry. Evolution 33,402-416. Lande, R. ( I 980). Genetic variation and phenotypic evolution during allopatric speciation. Am. Nut 116,463479. Lange, 0. L., Nobel, P. S., Osmond, C. B., and Ziegler, H. (eds.) (1981). “Physiological plant ecology. I. Responses to the physical environment.” Vol. 12, Springer-Verlag, Berlin and New York. Langridge, J. (1968). Thermal responses of mutant enzymes and temperature limits to growth. Mol. Gen. Gener. 103, 116-126. Laurie, D. A., and Bennett, M. D. (1985). Nuclear DNA content in the genus Zea and Sorghum: Intergeneric, interspecific and intraspecific variation. Heredity 55,307-3 13. Lawlor, D. W. (1995). The effects of water deficit on photosynthesis. In “Environment and plant metabolism” (N. Smirnoff, ed.), pp. 129-160. Bios Publ., Oxford, UK. LeClerg, E. L. (1966). Significance of experimental design in plant breeding. In “Plant breeding” (K. J. Frey, ed.), pp. 243-313. Iowa State Univ. Press, Ames, IA. Lee, M. (1995). DNA markers and plant breeding programs. Adv. Agron. 55,265-344. Lee, M., and Phillips, R. L. (1987). Genomic rearrangements in maize induced by tissue culture. Genome 29,122-128. Leegood, R. C. (1 995.) Effects of temperature on photosynthesis and photorespiration. In “Environment and plant metabolism” (N. Smirnoff, ed.), pp. 45-62. Bios Publ., Oxford, UK. Leigh, R. A. (1993). Perception and transduction of stress by plant cells. In “Plant adaptation to environmental stress” (L. Fowden, T. Mansfield, and J. Stoddart, eds.), pp. 223-237. Chapman and Hall, New York. Lerner, I. M. (1954). “Genetic homeostasis.” Oliver and Boyd, Edinburgh, UK. Levitt, J . (1980). “Responses of plants to environmental stress: Chilling, freezing, and high temperature stresses.” 2nd ed. Academic Press, New York.
248
MANJIT S. KANG
Lewontin, R. C. (1974). Annotation. The analysis of variance and the analysis of causes. Am. J. Hum. Genet. 2 6 , 4 0 0 4 1 1. Lin, C. S., and Binns, M. R. (1988). A method of analyzing cultivar x location x year experiments: A new stability parameter. Theor:Appl. Genet.76,425430. Lin, C. S., Binns, M. R., and Lefkovitch, L. P. (1986). Stability analysis: Where do we stand? Crop Sci. 26,894-900. Lubchenko, J., and Cubit, J. (1980). Heteromorphic life histories of certain marine algae as adaptations to variations in herbivory. Ecology 61,676-687. Magari, R., and Kang, M. S. (1993). Genotype selection via a new yield-stability statistic in maize yield trials, Euphytica 70, 105-1 11. Magari, R., Kang, M. S., and Zhang, Y. (1996). Sample size for evaluating field ear moisture loss rate in maize. Maydica 41, 19-24. Magari, R., Kang, M. S., and Zhang, Y.(1997). Genotype by environment interaction for ear moisture loss rate in corn. Crop Sci. 37,774-779. Malan, C., Greyling, M., and Gressel, J. (1990). Correlation between CuZn superoxide dismutase and glutathione reductase and environmental and xenobiotic stress tolerance in maize inbreds. Plant Sci. 69, 157-166. Marshall, D. M., and Jain, S. K. (1968). Phenotypic plasticity in Avena futua and A. barbara. Am. Nut. 102,457467. Matheson, A. C., and Cotterill, P. P. (1990). Utility of genotype x environment interactions. For Ecol. Management 30, 159-174. Matheson, A. C., and Raymond, C. A. (1984). The impact of genotype x environment interactions on Australian I? radiata breeding programs. Aust. For: Res. 14, 11-25. Matsuo, T.(1975). “Adaptability in plants.” Univ. of Tokyo Press, Tokyo. Mazer, S. J., and Schick, C. T. (1991). Constancy of population parameters for life history and floral traits in Raphanus sativus L. I. Norms of reaction and the nature of genotype by environment interactions. Heredity 67, 143-156. McKeand, S. E., Li, B., Hatcher, A. V., and Weir, R. J. (1990). Stability parameter estimates for stem volume for loblolly pine families growing in different regions in the southeastern United States. Forest Sci. 36, 10-17. McKersie, B. D., Chen, Y., de Beus, M., Bowley, S. R., Bowler, C., Inze, D., D’Halluin, K., and Botterman, J. (1993). Superoxide dismutase enhances tolerance of freezing stress in transgenic alfalfa (Medicago sativus L.). Planr Physiol. 103, 1155-1 163. Medina, R. C. (1992). Some exact conditional tests for the multiplicative model to explain genotypeenvironment interaction. Heredio 69, 128-1 32. Meredith, W. R., Jr. (1984). Quantitative genetics. In “Cotton and cotton improvement” (R.J. Kohel and C. F. Lewis, eds.), pp. 131-150. Am. SOC.Agron. and Crop Sci. SOC.Am., Madison, WI. Miller, D. A. (1996). Allelopathy in forage crop systems. Agmn. J. 88,854-859. Mishra, N. P., Mishra, R. K., and Singhal, G. S. (1993). Changes in activities of anti-oxidant enzymes during exposure of intact wheat leaves to strong visible light at different temperatures in the presence of protein synthesis inhibitors. Plant Physiol. 102,903. Mooers, C. A. (1921). The agronomic placement of varieties. J. Am. SOC.Agron. 13,337-352. Mottinger, J. P., Johns, M. A., and Freeling, M. (1984). Mutations of the Adhl gene in maize following infection with barley stripe mosaic virus. Mol. Gen. Genet. 195,367-369. Myers, W. M. (1960). Genetic control of physiological processes: Consideration of differential ion uptake by plants. In “Radioisotopes in the biosphere” (R. S. Caldecott and L. A. Snyder, eds.), pp. 201-226. University of Minnesota, Minneapolis. Nagao, R. T., Kimpel, J. A., and Key, J. L. (1990). Molecular and cellular biology of the heat-shock response. In “Advances in genetics” (J. G. Scandalios and T. R. F. Wright, eds.), pp. 235-274. Academic Press, New York.
USING GE INTERACTION FOR CROP DEVELOPMENT
249
Nelson, C. J. (1996). Allelopathy in cropping systems: Foreword. Agmn. J. 80,853. Neyra, C. A. (ed.). (1985). “Biochemical basis of plant breeding. Vol. I: Carbon metabolism.” CRC Press, Boca Raton, FL. Neyra, C. A. (ed.). (1986). “Biochemical basis of plant breeding. Vol. II: Nitrogen metabolism.” CRC Press, Boca Raton, FL. Nimmo, H. G., Carter, P. J., Fewson, C. A., Nelson, J. P. S., Nimmo, G. A., and Wilkins, M. B. (1995). Regulation of malate synthesis in CAM plants and guard cells; effects of light and temperature on the phosphorylation of phosphoenolpyruvate carboxylase. In “Environment and plant metabolism” (N. Smimoff, ed.), pp. 35-44. Bios Publ., Oxford, UK. Nyquist, W. E. (1991). Estimates of heritability and prediction of selection response in plant populations. Crit. Rev. Plant Sci. 10,235-322. Ottaviano, E., Sari-Gorla, M., Pe, E., and Frova, C. (1991). Molecular markers (RFLPs and HSPs) for the genetic dissection of thermotolerance in maize. Theol: Appl. Genet. 81,713-719. Ougham, H. J., and Howarth, C. J. (1988). Temperature shock proteins in plants. In “Plants and temperature symposium of the society for experimental biology” (S. P. Long and F. J. Woodward, eds.), pp. 259-280. Company of Biologists, Cambridge, UK. Ouyang, Z., Mowers, R. P., and Jensen, A. (1995). Cluster analysis for genotype x environment interaction with unbalanced data. Crop Sci. 35,1300-1305. Pace, G. M., Volk, R. J., and Jackson, W. A. (1990). Nitrate reduction in response to C0,-limited photosynthesis. Relationship to carbohydrate supply and nitrate reductase activity in maize seedlings. Plant Physiol. 92, 286-292. Pandey, S., and Gardner, C. 0. (1992). Recurrent selection for population, variety, and hybrid improvement in tropical maize. Adv. Agmn. 48, 1-87. Parsons, P. A. (1988). Evolutionary rates: Effects of stress upon recombination. Biol.J. Linn. Soc. 35, 49-68. Patterson, H. D., and Thompson, R. (1971). Recovery of inter-block information when block sizes are unequal. Biometrika 58,545-554. Peacock, J. M. (1975). Temperature and leaf growth in Lolium perenne. 11. The site of temperature perception. J. Appl. Ecol.12, 115-123. Pearcy, R. W. (1990). Sunflecks and photosynthesis in plant canopies. Annu. Rev. Plant Physiol. Mol. Biol. 41,421-453. Pedersen, A. R., Everson, E. H., and Grafius, J. E. (1978). The gene pool concept as basis for cultivar selection and recommendation. Crop Sci. 18,883-886. Pederson, D. G. (1968). Environmental stress, heterozygote advantage and genotype-environment interaction in Arabidopsis. Heredity 23, 127-138. Pescbke, V. M., and Sachs, M. M. (1993). Multiple pyruvate decarboxylase genes in maize are induced by hypoxia. Mol. Gen. Genet. 240,206-212. Pescbke, V. M., and Sachs, M. M. (1994). Characterization and expression of transcripts induced by oxygen deprivation in maize (Zea rnayJ!L.).Plant Physiol. 104,387-394. Pham, H. N., and Kang, M. S. (1988). Interrelationships among and repeatability of several stability statistics estimated from international maize trials. Cmp Sci. 28,925-928. Piepho, H.-P. (1994). Missing observations in analysis of stabilityHeredity72,141-145 (correction73,58). Pollack, C. J., Eagles, C. F,, Howarth, C. J., Schunmann, P. H. D., and Stoddart, J. L. (1993). Temperature stress. In “Plant adaptation to environmental stress” (L. Fowden, T. Mansfield, and J. Stoddart, eds.), pp.109-132. Chapman and Hall, New York. Polle, A., and Rennenberg, H. (1993). Significance of antioxidants in plant adaptation to environmental stress. In “Plant adaptation to environmental stress’’ (L. Fowden, T.Mansfield, and J. Stoddart, eds.), pp. 263-273. Chapman and Hall, New York. Pope, D. T., and Munger, H. M. (1953a). Heredity and nutrition in relation to magnesium deficiency chlorosis in celery. Proc. Am. Soc. Hon.Sci. 61,472-480.
2 50
MANJIT S. KANG
Pope, D. T., and Munger, H. M. (1953b). The inheritance of susceptibility to boron deficiency in celery. Proc. Am. SOC. Horr. Sci. 61,481486. Popp, M., and Smirnoff, N. (1995). Polyol accumulation and metabolism during water deficit. In “Environment and plant metabolism” (N. Smirnoff, ed.), pp. 199-215. Bios Publ., Oxford, UK. Prabhakaan, V. T., and lain, J. P. (1994). “Statistical techniques for studying genotype-environment interactions.”South Asian Publishers, New Delhi, India. Price, H. J., Chambers, K. L., Bachmann, K., and Riggs, I. (1986). Patterns of mean nuclear DNAcontent in Microseris douglusii (Asteraceae)populations. Bot. Guz. 147,496-507. Rameau, C., and Denis, J.-B. (1992). Characterizationof environmentsin long-term multi-site trials in asparagus, through yield of standard varieties and use of environmentalcovariates. Plunr Breeding 109, 183-191. Rao, T. D. P., Rao, D. V. S., and Rai, S. C. (1988). Symposium on “Statistical aspects of stability of crop yields. J. Ind. SOC.Agr: Statistics 60,70-79. Rao, V., Henson, I. E., and Rajanaidu, N. (eds.). (1993). “Genotype x environmentinteraction in perennial tree crops.” Int. SOC.of Oil Palm Breeders and Palm Oil Res. Instit. of Malaysia, Kuala Lumpur, Malaysia. Rayals, J., Uknes, S., and Ward, E. (1994). Systemic acquired resistance. Plant Physiol. 104, 1109-1 112. Ratcliffe, R. G. (1995). Metabolic aspects of the anoxic response in plant tissue. In “Environment and plant metabolism” @ Smirnoff, I. ed.), pp. 111-127. Bios Publ., Oxford, UK. Rayburn, A. L., and Auger, J. A. (1990). Genome size variation in Zeu mays ssp. Mays adapted to different altitudes. Theos Appl. Gener. 79,470-474. Rayburn, A. L., Price, H. J., Smith, J. D., and Gold, J. R. (1985). C-banded heterochromatin and DNA content in Zea mays. Am. J. Bot. 72, 1610-1617. Redei, G. P. (1982). “Genetics.” Macmillan, New York. Rideout, J. W., Chaillou, S., Raper, C. D., Jr., and Morot-Gaudry, J. F.(1994). Ammonium and nitrate uptake by soybean during recovery from nitrogen deprivation. J. Exp. Boz. 45,23-33. Rosielle, A. A., and Hamblin, J. (1981). Theoretical aspects of selection for yield in stress and nonstress environments. Crop Sci. 21,943-946. Saeed, M., and Francis, C. A. (1984). Association of weather variables with genotype x environment interaction in grain sorghum. Crop Sci. 24, 13-16. Scandalios, J. G . (1990). Response of plant antioxidantdefense genes to environmentalstress. In “Advances in genetics” (J. G. Scandalios, and T. R. F.Wright, eds.), pp. 1-41. Academic Press, New York. Scandalios, J. G., and Wright, T. R. F.(eds.) (1990). Genomic responses to environmentalstress. Adv. Genet. 28, xi. Schlichting, C. D. (1986). The evolution of phenotypic plasticity in plants. Ann. Rev. Ecology Sysrematics 17,667493. Schlichting,C . D., and Levin, D. A. (1984). Phenotypic plasticity of annual phlox: Tests of some hypotheses. Am. J. Bot. 7,252-260. Scowcroft, W. R. (1988). Genetic manipulation in crops: A symposium review. In “Genetic manipulation in crops,” pp. 13-17. International Rice Research Institute, Cassell Tycooly, Philadelphia. Searle, S. R. (1987). “Linear models for unbalanced data.” John Wiley & Sons, New York. Searle, S. R., Casella, G., and McCulloch, C. E. (1992). “Variance components.”John Wiley & Sons, New York. Shah, D. M., Horsch, R. B., Klee, H. J., Kishore, G. M., Winter, J. A., Turner, N. E., Hironaka, C. M., Sanders, P. R., Gassere, C. S., Aykent, S., Siegel, N. R., Rogers, S. G., and Fraley, R. T. (1986). Engineering herbicide tolerance in transgenic plants. Science 233,478-481. Sherrard,J. H., Lambed, R. J., Below, F. E., Dunand, R. T., Messmer, M. I.,Willman, M. R., Winkels,
USING GE INTERACTION FOR CROP DEVELOPMENT
25 1
C. S., and Hageman, R. H. (1985). I n “Biochemical basis of plant breeding, vol. 11: Nitrogen metabolism” (C. A. Neyra, ed.), pp. 109-130. CRC Press, Boca Raton, FL. Shifriss, 0. (1947). Developmental reversal of dominance in Cucurbitupepo. Proc. Am. SOC.Horf.Sci. 50,330-346. Shukla, G. K. (1 972). Some statistical aspects of partitioning genotype-environmental components of variability. Herediry 29, 237-245. Silvey, V. (1981). The contribution of new wheat, barley and oat varieties to increasing yield in England and Wales 1947-78. J. Nut. Inst. Agric. Bof. 15,399-412. Simmonds, N. W. (1981). Genotype (G). environment (E) and GE components of crop yields. Expl. Agric. 17,355-362. Singleton, R. W. (1967). “Elementary genetics.” Van Nostrand, New York. Smimoff, N. (1993).The role of active oxygen in the response of plants to water deficit and desiccation. New Phyfol. 125,27-58. Smimoff, N. (1995). Antioxidant system and plant response to the environment. I n “Environment and plant metabolism” (N. Smirnoff, ed.),pp. 217-243. Bios Sci. Publ., Exeter, UK. Smith, H. (1990). Signal perception, differential expression within multigene families and the molecular basis of phenotypic plasticity. Plant Cell Environ. 13,585-594. Smith, M. E., Coffman, W. R., and Barker, T. C. (1990).Environmental effects on selection under high and low input conditions. I n “Genotype-by-environment interaction and plant breeding” (M. S. Kang, ed.), pp. 261-272. Louisiana State Univ. Agric. Center, Baton Rouge, LA. Specht, J. E., and Laing, D. R. (1993). Selection for tolerance to abiotic stressesdiscussion. In “International crop science I” (D. R. Bruxton, R. Shibles, R. A. Forsberg, B. L. Blad, K. H. Asay, G. M. Paulsen, and R. F. Wilson, eds.), pp.381-382. Crop Sci. SOC.of America, Madison, WI. Sutka, J., and Veisz, 0. (1988). Reversal of dominance in a gene on chromosome 5A controlling frost resistance in wheat. Genome 30,313-317. Tai, G. C. C. (1971). Genotypic stability analysis and its application to potato regional trials. Crop Sci. 11, 184-190. Tai, G. C. C. (1990). Path analysis of genotype-environment interactions. In “Genotype-by-environment interaction and plant breeding” (M. S. Kang, ed.), pp. 273-286. Louisiana State Univ. Agric. Center, Baton Rouge, LA. Tanksley, S. (1993). Mapping polygenes. Annual Rev. Genetics 27,205-233. Theodorou, M. E., and Plaxton, W. C. (1995). Adaptations of plant respiratory metabolism to nutritional phosphate deprivation. In “Environment and plant metabolism” (N. Smirnoff, ed.), pp. 79-109. Bios Publ., Oxford, UK. Thomashow, M. F. 1990. Molecular genetics of cold acclimation in higher plants. In “Advances in genetics” (J. G. Scandalios and T. R. F. Wright, eds.), pp. 99-131. Academic Press, New York. Turelli, M. (1988). Phenotypic evolution, constant covariances and the maintenance of additive variance. Evolution 42, 1342-1 347. Turkington, R. (1983). Plasticity in growth and pattern of dry matter distribution of two genotypes of Trifolium repens grown in different environments of neighbours. Can. J. Bot. 61,21862194. Unsworth, M. H., and Fuhrer, J. (1993). Crop tolerance to atmospheric pollutants. In “International crop science I” (D. R. Bruxton, R. Shibles, R. A. Forsberg, B. L. Blad, K. H. Asay, G. M. Paulsen, and R. F. Wilson, eds.), pp. 363-370. Crop Sci. SOC.of America, Madison, WI. Van Eeuwijk, F. A., Denis, J.-B., and Kang, M. S . (1996).Incorporating additional information ongenotypes and environments in models for two-way genotype by environment tables. In “Genotypeby-environment interaction” (M. S. Kang and H. G. Gauch, Jr., eds.), pp. 1549. CRC Press, Boca Raton, FL. Van Oosterom, E. J., Kleijn, D., Ceccarelli, S., and Nachit, M. M. (1993). GE interactions of barley in the Mediterranean region. Crop Sci. 33,669-674.
252
MANJIT S. KANG
Venuto, B. C., Smith, R. R., and Grau, C. R. (1996). Temperature dependent reaction and selection response to two red clover populations to fusarium wilt. Crop Sci. 36, 1477-1481. Via, S. (1984a). The quantitative genetics of polyphagy in an insect herbivore. I. Genotype-environment interaction in larval performance on different host plant species. Evolution 38,881-895. Via, S. (1984b). The quantitativegenetics of polyphagy in an insect herbivore. 11. Genetic correlations in larval performance within and among host plants. Evolution 38,896-905. Via, S., and Lande, R. (1985). Genotype-environment interaction and the evolution of phenotypic plasticity. Evolution 39,505-522. Via, S., and Lande, R. (1987). Evolution of genetic variability in a spatially heterogeneous environment: Effects of genotype-environment interaction. Genet. Res. 49, 147-156. Vogel, S. (1968). “Sun leaves” and “shade leaves”: Differences in convective heat dissipation. EcologY49,1203-1204. Weber, W. E., Wricke, G., and Westermann, T. (1996). Selection of genotypes and prediction of performance by analyzing GE interactions.In “Genotype-by-environment interaction” (M. S. Kang and H. G. Gauch, Jr., eds.), pp. 353-371. CRC Press, Boca Raton, F L Woodend, J. J., andGlass, A. D. M. (1993). Genotype-environment interaction and correlationbetween vegetative and grain production measures of potassium use-efficiency in wheat (Zaeativum L.) grown under potassium stress. Plant and Soil 1 5 1 , 3 9 4 . Wricke, G. (1962). Uber eine Methode zur Erfassung der okologischen Streubreite in Feldversuchen. 2. Pfanzenzuchr 47,92-96. Wricke, G., and Weber, W. E. (1986). “Quantitative genetics and selection in plant breeding.” Walter de Gruyter, New York. Yamada, Y.(1962). Genotype by environment interaction and genetic correlation of the same trait under different environments. Jap. J. Genet. 37,498-509. Yates, F., andcochran, W. G. (1938). The analysis of groups of experiments.J. Agric. Sci. 28,556-580. Zhang, Q., and Geng, S. (1986). A method of estimating varietal stability for long-term trials. Theos Appl. Genet. 71,8 1&8 14.
MODELINGCARBON AND NITROGEN PROCESSES INSOILS Jean-Alex E. Molina' and Pete Smith' 'Department of Soil, Water, and Climate University of Minnesota Saint Paul, Minnesota 55108 2Soil Science Department IACR-Rothamsted Harpenden, Hertfordshire, ALS 2JQ United Kingdom
I. lntroduction and Historical Background 11. Model Description A. The First Models B. Current Models C. Modifiers of Rate Constants 111. Model Validation A. The Need to Critically Evaluate Models B. Validation or Calibration? C. Nontracer Data: Long-Term Simulations D. Tracer Data W. Model Applications A. Using Models as a Research Tool B. Using Models to Improve Agronomic Efficiency and Environmental Quality C. Using Models to Estimate Global C in SOM and Fluxes of C under Global Environmental Change V Conclusions and Future Work References
I. INTRODUCTION AND HISTORICAL BACKGROUND Soil carbon and nitrogen are found in a great diversity of chemicals. Carbon and nitrogen compounds do not occur haphazardly but are connectedby a web of transformations controlled mostly by microbial activity that is influenced by the physical and chemical conditions of the soil. A century ago, however, the soil was 253 Advanrer in Agronumy, ~ ' o ~ u m62 e
Copyright 0 1998 by Academic Press. All rights of reproduction in any form reserved. 0065-2113/98 $25.00
2 54
JEAN-ALEX E. MOLINA AND PETE SMITH
viewed, even by reputed scientists, as a mysterious environment: that portion of the earth, the crust, that was most exposed to the influence of the sun, the moon, and the planets; the receptacle of dead organisms; the matrix that fed plants and synthesized minerals; and the site of spontaneous generation. Before reviewing models of carbon and nitrogen transformations in soil, we would like to remind the reader that countless experiments and the elaboration and destruction of many theories were required to demystify the soil. In the short historical panorama that follows, our intention is to show with selected examples how issues related to carbon and nitrogen processes in soil were linked to fundamental scientific inquiries. A comprehensive story of the evolution of soil science can be found in Boulaine (1989), Krupenikov (1993), and Gorham (1991). In 1652, the physician Jean-Baptiste Van Helmont completed an experiment performed in the garden of his residence at Brussels, Belgium. For 5 years he followed the growth of a willow tree in a large pot containing 250 lb of soil. The tree’s dry mass increased from 5 lb to 169 lb 3 oz. The soil mass did not change. Van Helmont concluded that water was responsible for the plant mass increase. Three centuries later, the same issue would be raised, not about a tree in a pot, but about the earth’s global carbon mass balance; and as the correct answer eluded Van Helmont because of methodological difficulties in the detection of mass variations in air and 250 lb of soil, so are we confronted with the challenge to identify the origin of the “carbon leak” in a biosphere with 1400 Pg of soil organic matter (Cheng and Molina, 1995). By 1780, Antoine Laurent Lavoisier had purchased 1500 ha of prime agricultural land. He required of his farmers that a record be kept of agronomic practices performed, yields obtained, and the animal load maintained on each plot. Thus Lavoisier developed an appreciation for natural cycles and described in an unpublished document (Boulaine, 1989) the process that would be later known as the mineralization-immobilizationturnover: Plants draw in from the air which surrounds them, from the water and in general from the mineral kingdom, the materials required for their organization.Animals feed on plants or from other animals which have themselves fed on plants, so that the materials they are made of are always, in the last analysis, obtained from air and from the mineral kingdom. Finally, the fermentation, the putrefaction, and the combustion bring back continuously to the air of the atmosphere and to the mineral kingdom the principles which had been borrowed by the plants and the animals. By what process is nature directing this wonderful circulation between the two kingdoms? How can she synthesize substances which are combustible, fermentable, and putrescible with mixtures which do not have any of these properties? These are impenetrable mysteries. One foresees, however, that, since combustion and putrefaction are the means used by nature to give back to the mineral kingdom the material which she draws from to form
MODELING CARBON AND NITROGEN PROCESSES IN SOILS 255 plants and animals, the plant and animal generating processes must be the inverse operations of combustion and putrefaction. [Translated from the French] By the end of the nineteenth century the intuitive theory of Lavoisier on the natural cycling of carbon was finally established on a firm chemical basis. MacBride showed that the combustion of vegetation and animals produced carbon dioxide, the “fixed gas” of Van Helmont. At about the same time, photosynthesis was described: plants incorporate carbon dioxide (Percival) only under the influence of the sun (Ingen-Housz), a process that was associated with the release of a gas (Sennebier) identified by Priestley as oxygen. Nevertheless, many philosophers and scientists, such as Hassenzfratz, who was for some time the director of Lavoisier’s laboratory, remained convinced that plant carbon originated from soil. The term humus was vulgarized by the German Thaer who claimed in 1802 that “since humus was the result of life, so it is also the condition of life. It gives nourishment to organized bodies; without it, there could not be individual life.” Even the great chemist Berzelius ( 1779-1 848), who introduced the distinction between inorganic and organic chemistry and performed the first chemical fractionation of humus, did not rule out the role of humus as a source of plant carbon. Indeed, how could humus not be the principal source of plant carbon when soil is the most active site of spontaneous generation, a theory not dismissed until 1861 by Pasteur. In 1837, the Liverpool branch of the British Association for the Advancement of Science requested from the German chemist Von Liebig a book that would be published 3 years later-Chemistry in Its Relations with Agriculture and Plant Growth-to become the driving force behind the chemical fertilizer industry. Liebig established that plants can assimilate potassium and phosphorus from inorganic forms. He also set the stage for the debate about nitrogen transformations in soil by stating that “plants in general receive their nitrogen from the atmosphere as ammoniac.” By contrast, J. B. Lawes and J. H. Gilbert, the founders of Rothamsted Experiment Station in the UK, and J. B. Boussingault in France, demonstrated that crops derived N from soil. Agronomists knew that manure enriched soil and that nitrate was a good source of nitrogen for plants. By 1830, several thousand tons of nitrate were imported by European nations from Chile. What was the best source of nitrogen for plants: Ammoniac or molecular nitrogen in the air, organic molecules, or ammonium salts and nitrates? As it became apparent that all these compounds were present in soil, the issue of their stability, concentration, and transformation was raised. Gradually, the nitrogen processes in soil were identified: nonsymbiotic and symbiotic nitrogen fixation, mineralization and immobilization, nitrification, denitrification, and leaching. The microbial agents of the biological processes were isolated; the climatic and soil conditions controlling each transformation were identified, and the nitrogen cycle took shape (Waksman and Starkey, 1931; Winogradsky, 1949; Russell, 1973; Alexander, 1977).
256
JEAN-ALEX E. MOLINA AND PETE SMITH
II. MODEL DESCRIPTION A. THEFIRSTMODELS For this review, modeling of C and N processes in soil will be understood as the quantitative expression of changes in C and N concentrations caused by enzymatic reactions. For each identified or hypothetical chemical in soil, it will be computed as the rate of C and N input minus the rate of loss. This approach is based on the ability to write an equation to express the velocity of chemical reactions. The method was first used by Wilhemy in 1859 when he investigated the inversion of sugar in aqueous acid solutions. Berthelot and St. Gilles, studying the hydrolysis of esters, introduced the concept of chemical equilibrium, which was shown by Guldsberg and Waage to be dynamic rather that static. The application of the principle of dynamic equilibrium to processes in soil was described by Nihforoff (1936) in the first issue of the “Soil Science Society of America Proceedings”: “Assuming temporarily that the entire mass of organic residues is composed of substances which decompose uniformly and that the rate of decomposition remains the same from the beginning to the end of the process, one may express the process of humus accumulation by . . . the difference between the amount of humus newly formed and the amount reduced to the end product of mineralization.” Jenny (1941), expanding on the work of Salter and Green, modeled net changes in total soil N as the rate of N losses minus the rate of N fixation from air: dNldt = - k,N
+ k2
(1)
The rate of N fixation was assumed to be constant (k2); the rate of N losses was linearly linked to the concentration of soil organic N (N) by a proportionality constant (k,), called specific rate or rate constant. The integrated rate equation was used to express changes in N concentration as a function of time.
N = N, - (N, - No) e--kir where N, is the equilibrium level of N in soil equal to kJk, (dN/dt = 0). Values for k, and k2 (k, = 0.0608 yr-*, and k, = 38 kg N ha-‘ yr-’) were obtained by fitting Eq. (2) against data describing the gradual loss of soil N upon cultivation without N fertilizer addition. The net rate equation for change in soil organic matter (SOM) (r) proposed by Henin and Dupuis (1945) dy/dt
=
( K , w - K2 y )
describes the equilibrium dynamics between SOM decay and formation. W is the organic matter input (plant debris) expressed as apercentage of the total SOM content y. The organic input is transformed to the SOM with a rate constant K , , called
MODELING CARBON AND NITROGEN PROCESSES I NSOILS 257
by Henin the isohumic coefficient. K2 is the rate constant of SOM decay. Values for K , and K2 were obtained by curve fitting of the integrated net rate equation
(l/K2) I n [ ( K , w - K 2 y ) / ( K ,w - K2 Yo)] = -t
(3)
where Yo is the initial SOM content at time t = 0. Various simplified forms of Eq. ( 3 ) were considered to comply to the particulars of experimental conditions. For example, K , was set to zero when crop residues were not returned and organic additions were assumed to be negligible. The mean value for K2 was 0.0105 yr-l. It
g :2;
k4 :a; 0
lO\
lo,
2
:o; ii n n :; ;A i ,: AA :
3 2 d
: A ! : A :
@ ' r
C B
~
4 ;
9)
: A \
0
; A : ' A :
vs
+
Holtzand Vandecaveye
X
Barbier
I Snyder
: + :
C
l
.o_ c 3
r' p
B k
o
:
4
0
0 0 0 0 0 0 0
0
og 0
c y ( o o 3 % z N cv
*
I
8
!O
6 E3
;A / + I
0-
I;"
Woburn
o Allway A Swanson
3
0 c)
q
0
0
a Q
0
I
U'\,
0 0
0
0 0
l
4
', 6..*
A x, A',
0 0
l
h
I
=
a o c u c o N
my
z
I
I
I
I
a w n )
3
~ 3
0
0
0
a
o L 1 0
0
N
d d
0
d
0
'&--e------&--Q
0
n)
I
~
2 K 2 Value
~
9 0 0 0 9 0
0
Figure 1 Frequency distributionof the rate constant K2for SOM decay Henin and Dupuis, (1945). (Essai de bilan de la rnatiEre organique du sol. Ann. Agron. 15, 17-29. 0 INRA; reprinted with kind permission of INRA Editions.)
~
~
258
JEAN-ALEX E. MOLINA AND PETE SMITH
ranged from 0.004 to 0.04 yr-l, and 56% of the data was clustered in the interval 0.008-0.012yr-' (Fig. 1).The isohumic coefficient was obtainedfor various plant debris. It varied considerably. Calibration showed that 10-80% of added organic material was converted to SOM within 1 year. Olson (1963) quantitatively related the amount of organic material found on the forest floor to the litter production: dXldt = L - kX
where X is the level of dead organic matter at time t that decays with the rate constant k, and L the annual litter fall production per square meter of ground surface. Assumed to have reached a steady state (dX/dt = 0), a k value was computed from the slope of the linear relationship between measured L and X for stands of evergreen forests located in various locations with contrasting climatic conditions. Values for k were found to range from 0.0156 y-' (Sierra Nevada mountains) to 4.0 yr- I (Ghana, Congo). Olson also discussed nonsteady-state situations: decay with no production and accumulation with discrete annual litter fall. One universal feature of soils that have not been recently amended with N-depleted organic debris is to release inorganic N when incubated under conditions favorable to microbial activity. Stanford and Smith (1972) measured inorganic N accumulation in soils incubated for 8 weeks. They modeled the kinetics of net N mineralization by assuming that the rate of soil organic N decay was proportional to the level of mineralizable organic N concentration (N): dNldt = -kN
(4)
or upon integration, N , = No (1 - e-kr)
(5)
where N, is the accumulated inorganic N at time t, and N o the potentially mineralizable organic N at t = 0. The shape of the exponential curve generated by Eq. ( 5 ) is similar to sections of either a parabola
N,
=A
No tl'*
(6)
or a hyperbola N,
= No ( t / ( B
+ t))
or 1/N, = l/No
+ @/No) (llt)
(7)
where A and B are arbitrary coefficients used to control the curvature of the conic curves. Equations (6) and (7) were validated by Stanford and Smith's (1972) observations that inorganic N accumulated linearly with the square root of time; and that the inverses of inorganic N and of time were linearly related. N o was com-
MODELING CARBON AND NITROGEN PROCESSES I N SOILS 259 puted by extrapolation to I/f = 0 of the experimental data plotted on Y = 1/N, and x = l/t coordinates. The potentially mineralizable nitrogen No varied greatly from
5 to 40% of total soil N. However, the k values remained remarkably constant (2.808 t 0.468 yr-’) in spite of the diversity of soils considered.
B. CURRENTMODELS Early models simulated the SOM as one homogeneous pool. This limitation was imposed on modelers by the mathematical difficulty to integrate a system of differential equations. Beek and Frisel(1973) used a computer to obtain a numerical solution to a system of net rate equations. Their model considered two organic pools: the microbial biomass and humus. As computers became more accessible, more models with as many pools as needed to account for all the processes of carbon and nitrogen transformations in soil were published. The 1996 edition of the CAMASE register (Plentinger and Penning de Vries, 1996) describes 200 agroecosystems models, of which 98 include components describing processes in soil. The CAMASE register can be accessed through the Internet at: http://www.co.dlo.nl/camase. A workshop was held in 1990 at the Department of Fertilization and Plant Nutrition of the Institute for Soil Fertility Research, Haren, The Netherlands. The emphasis was on models of N turnover in soil-crop research. Fourteen models were compared. Proceedings of the workshop were published in a special issue of Fertilizer Research (Vol. 27, 1991). The Soil Organic Matter Network (SOMNET) is another source of information about models of soil transformations. It was compiled in 1995 for the Global Change and Terrestrial Ecosystems (GCTE) project of the International Geosphere-Biosphere Programme (IGBP). A key goal of GCTE Task 3.3.1. “concerns soil organic matter, a major carbon store and key component of soil fertility.” One objective of GCTE Core Research Project 2 is “to identify, facilitate development of, and apply simulation models to accomplish the objectives of Task 3.3.1.” A worldwide survey gave 27 operational SOM models (Table I) and 70 available long-term experimental field datasets. Information about the models and the long-term field experiments was published (P. Smith et al., 1996c) and is accessible through the Internet at: http://yacorba.res.bbsrc.ac.uk/cgi-bin/somnet. Evaluation of nine SOMNET models (CANDY, CENTURY, DAISY, DNDC, Hurley-ITE, NCSOIL, RothC-26.3, SOMM, and VERBERNE) was performed against some long-term data sets during an Advanced Research Workshop held at IACR-Rothamsted, UK (McGill, 1996; Powlson et al., 1996; P. Smith et al., 1997b,c). A cursory look at the models shows that they reflect a great variety of understanding and interpretation of C and N processes in soil. It is this diversity that will be illustrated in the next sections.
Table I
List of Models Registered with SOMNET in 1996 Model name and referencea
Resolution Spatial*
ANIMO P,F, C, R, N, Lv Ritjema et al. (1995) P, F, C, L20 CANDY Franko et al. (1995) CENTURY P,F.R,N, G. L1 Parton and Rasmussen ( 1994)
Temporal'
Factors affecting- decayrate constantsd
D, W, Mo
T, W. H, N, 0
D, Y
T, W, N, CI
M
T, W, N, C1, H, Ti
CNSP McCaskill and Blair (1990)
F. Lv
D
T, W, N, H
D3R Douglas et al. (1995) DAISY Jensen et al. (1994) DNDC Li ei al. (l994a) ECOSYS Grant (1995)
F, L1
D
T, W, N, Cv, Ti
P,F,C,L5
H
T, W, N, C1
P, L10
H, D
T, W, N, C1, Ti
S, P, F, Lv
Mi, H
T, W, N, 0, c1,c v
F, LIO
D
T, W, N, H,C1, Ce, Cv
P, F, L5
D
T, w.H, N, cv
P, Lv
D
T, W N
P,F,R,Ll
Y
N, H, C1, Cv
P, F, L1
Mi
T, W, N
EPIC Bryant et al. ( 1992) FERT Kan et al. (1993) GENDEC Moorhead and Reynolds (1991) HUMUS Shevtsova and Volodarskayab (1991) ITE FOREST Thornley and
Remarks
Amount of inert SOM function of soil texture One layer (0-20 cm) of organic pools, six layers for nitrate leaching Tracer C and N-NO; simulated Lignin in debris and soil texture affect decay of debris Decay rate constant modified by phosphorus and sulfur content, and digestibility of grass and clover Model undergoing modifications by merging with model GRASSGRO. Rate constants based on degree-days
Minutes used for transport equations, hours for biological transformations
Temperature and water factored in through a hydrothermic coefficient
continues
2 60
Table I-continued Model name and referencea Cannel(l994) NAM SOM Ryzhova ( I 993) NCSOIL Molina (1 996) O’LEARY O’Leary (1994) Q-SOIL Bosatta and &reen (1995)
ROTH-26.3 Coleman and Jenkinson (1996) SOCRATES Oades ( I 995) SOMM Chertov and Komarov (1996) SUNDIAL Smith et al. (1995) VERBERNE Verbeme et a!. ( 1990) VOYONS Andr6 et al. ( 1992)
WAVE Vanclooster etal. (1992)
Resolution Spatialb
Temporal‘
Factors affecting decay rate constantsd
Remarks
s, P, L1
Y
T, W, CI, Cv
A single layer, 100 cm deep
s, L1
D
T, W, N, H, CI, Ti
Time step, day or fraction of day, controlled by user Tracer C and N are simulated
P,F, L10
D
T, W, N, CL Ti
P,F, Lv
Y
T, W, N
P, F, C, R, N, G, L1
M
T, W, C1, Cv
P, L1
W
T, W, N,Cv, Ce
P,F, G , L2
D
T, W, N
Soil fauna affects organic matter decay
P, F, L12
W
T, W, CI
Tracer N is simulated
P,F, Lv
D
T, W, N, CI
S, P, F, Lv
D, W, Mo
T, W, CI
P,F, C, R,
D
T, W, N
Lv
Quality, a distribution function, defines the accessibility of substrates to decomposers The model is solved by analytical rather than numerical methods Tracer total C and N are simulated Tracer C is simulated
A user-friendly frame for various models (BIOMASSE, CENTGLOB, DECO, ECOSIMP2, MITO) Variable time step (> day) function of soil moisture
aFor more references and contact person, see http://yocorba.res.bbsrc.ac.uk/cgi-bidsomnet bS = microsite, P = plot, F = field, C = catchment, R = regional, N = national, G = global, Lx = nb. x of soil layers; Lv = nb. of soil layers controlled by user. ‘Mi = minutes, H = hours, D = days, W = weeks, Mo = months, Y = years. 9= temperature, W = water, H = pH, N = nitrogen, 0 = oxygen, C1 = clay, Ce = cation exchange capacity, Cv = cover crop, Ti = tillage. 261
2 62
JEAN-ALEX E. MOLINA AND PETE SMITH
1. Soil Organic Pools The network of C and N fluxes between the models’ pools forms the model structure. Each SOM pool is characterized by its position in the model’s structure and its decay rate or the rate at which its concentration decreases when there is no input. Decay rates are usually expressed by first-order kinetics with respect to the concentration (C) of the pool dCldt = -kC
(8)
where t is the time. The rate constant k of first-order kinetics is related to the time required to reduce by half the concentration of the pool when there is no input. The pool’s half-life ( h = (ln2)/k),or its turnover time (T = l/k) are sometimes used instead of k to characterize a pool’s dynamics: the lower the decay rate constant, the higher the half-life, the turnover time, and the stability of the organic pool. The diversity of soil organic pools and decay rate constants of the SOMNET models is shown in Table 11.
2. Model Structure a. C Flow Model structures are built on the assumption that SOM pools are the rings of a chain linked by C flows, thus forming a catenary sequence of substrates (Van Veen et al., 1981). The sequence represents C going from plant and animal debris to the microbial biomass, then to soil organic pools of increasing stability (Fig. 2 on p. 265). Some models break this simple linear flow with feedback loops to account for catabolic and anabolic processes and microbial successions. Microbial successions are simulated by C flow issued from one microbial pool feeding back into itself (Fig. 3 on p. 266). It has a great impact on the soil dynamics as it recycles C and N in a short half-life pool with the release of the stable end products of catabolic processes at each turnover (e.g., CO,). This scheme, first used by Beek and Frisel(1973), is found in the models RothC-26.3, NCSOIL, and VERBERNE, but it is not present in CENTURY where the microbial biomass pool is fed by the decomposition of plant debris, the SLOW and PASSIVE soil organic pools, but not by itself. The output flow from an organic pool is usually split. It is directed to a microbial biomass pool, another organic pool, and, under aerobic conditions, to CO,. This split simulates the simultaneous anabolic and catabolic activities and growth of a microbial population feeding on one substrate. Two parameters are required to quantify the split flow. They are often defined by a microbial (utilization) efficiency and stabilization (humification) factor that control the flow of decayed C to the biomass and humus pools, respectively. The sum of the efficiency and humification factors must be less than 1 to account for the release of CO,. Other forms
Table II
Pools and Rate Constants (for First-Order Kinetics at Optimum Decay Conditions) Found in SOMNET Models to Represent Organic Debris (OD), and Soil Organic Matter (SOM) Model
OD, rate constant (yr-I)
ANIMO
Fresh organic matter, 1 .8
CANDY
OD (UP10 6 pools), 0.05-0.4
CENTURY
Structural OD, 3.9 Metabolic OD. 15
CNSP
Dung Urine Uneaten green plant Uneaten dead plant Treading damaged organic debris Unprotected digestible herbage Unprotected indigestible herbage Clay protected, indigestible herbage Unprotected recalcitrant herbage Protected recalcitrant herbage Surface “fast” OD N rich > 1.5% N poor Burried “fast” OD N rich > I S % N poor Surface “slow” OD Burried “slow” OD Recalcitrant added OD (AOMO) Structural, OD (AOMI), 1.8 Metabolic, OD (OM2), 18 Structural, root debris (ARMI), 2.5 Metabolic, root debris ( A w l ) , 25
D3R
DAISY
DNDC
Very labile litter Labile litter Resistant litter. 7.3
ECOSYS
Protein OD Carbohydrate OD Cellulose OD Lignin OD Soluble OD
SOM, rate constant (yr-’) Root exudates, 36 Dissolved organic matter, 2.9 Stable humus, 0.2 Active organic matter, 0.14 Stable organic matter, 0.05 Inen organic matter
263
Surface microbes, 6.0 Soil microbes, 7.3 Slow soil organic matter, 0.2 Passive soil organic matter, 0.0045 Unprotected and protected biomass
Soil microbial biomass (SMB I), 0.36 Soil microbial biomass (SMB2). 3.6 SOM dead native (SOMl), 0.001 SOM dead native (SOM2), 0.05 Inert SOM (SOMO) Labile microbial mass, 120 Resistant microbial mass, 15 Labile humads (labile active humus), 58 Resistant humads, 2.2 Passive humus Soluble SOM Adsorbed SOM Microbial SOM Microbial residues Active SOM Passive SOM Particulate SOM continues
Table II-continued Model
OD, rate constant (yr-’)
SOM, rate constant (yr-l)
ECOSYS (cont.) Soluble OD (cont.) EPIC
OD
FERT
OD
GENDEC
Labile OD, 73 Hollocelluiose OD, 29 Lignin (recalcitrants) OD, 3.6 OD “Fast” plant litter Feces OD Foliage, branch-stem, roots, litter OD OD I OD I1 OD 111
HUMUS ITE FOREST
NAM SOM NCSOIL
O’LEARY
Surface residue, 0.84
Q-SOIL ROTHC-26.3
OD Decomposable plant material (DPM), 10 Resistant plant material (RPM), 0.3 Decomposable plant material Resistant plant material
SOCRATES
SOM Pool I, labile (microbial biomass), 120 Pool I, recalcitrant (microbial biomass), 15 Pool II, 2.2 Pool I1 (tillage), 58 Pool 111,
SOMM
Undecomposed litter
SUNDIAL
OD, 8.3
VERBERNE
Decomposable material (DPM) Structural material (SMP) Resistant material (RPM)
VOYONS
Metabolic OD, 14-18 Structural OD, 4-5 Lignin with OD Soil litter pool, 0.3-73 Soil manure pool, 0.36-13
WAVE
Acetate Methane Active SOM Stable SOM Microbial biomass Humus Live microbiota Dead microbiota, labile, 73 Dead microbiota, recalcitrant, 0.36 SOM Biomass SOM (dead)
264
Microbial biomass (unprotected) Microbial biomass (protected) Humus (stable pool) Litter impregnated by humic substances Humic substances of mineral top soil Microbial biomass, 0.66 Humus, 0.21 Nonprotected biomass (BIOMN) Protected biomass (BIOMP) Nonprotected SOM (NPOM) Physically protected SOM (POM) Active SOM, 7.3 Slow SOM, 0.2 Passive SOM, 0.0067 Soil humus pool, 0.025-0.36
MODELING CARBON AND NITROGEN PROCESSES IN SOILS 265
m
I
-Fungi Transformation by and Acarrna
k,
-
F CHS, complex of humus substances with undecomposed organic debris
-
I
k , Consumption by micro (meso) fauna
-
k,
I
k, Consumption by meso fauna (earthworms)
J.0
H
.
- humus bonded
~~
k6
of parametizations have been proposed. For example, the decay of each organic pool of the model RothC-26.3 is directed to the pools BIO, HUM, and CO, (Fig. 4 on p. 267). The decay flow is initially split between CO, and the biomass plus humus pools (COJ(BI0 + HUM), a ratio the value of which is functionally related to the soil clay content. Then, 46% of the BIO plus HUM flow is directed to the BIO pool (Coleman and Jenkinson, 1996). Some models split input flows to the labile and resistant forms of the recipient pool. This approach requires the specification of a partition parameter. For example, the surface litter is split into structural and metabolic portions in the model CENTURY; the model DAISY splits in half the C flow from added organic debris pool (AOMI) to the microbial pools SMBl and SMB2.
x
Debris Primary Decomposers Microbial Succession
03 Biomass
Pool 111
I Biomass
Ql
1
Biomass
0 3
Pool 111
1 Figure 3 Model NCSOIL, structure of C flow. Example of microbial successions. (A) simplification of the C flow by grouping each microbial succession into one biomass pool with C feeback loop; (B) final form of the model after collapsing the microbial biomass pools into Pool I (adapted from Molina, 1996. Description of the Model NCSOIL; with kind permission of Springer-Verlag).
MODELING CARBON AND NITROGEN PROCESSES IN SOILS 267
Organic
L
J
U
RPM : Resistant Plant Material DPM :Deoomposabte Plant Material 810 : Microbial Biomass
HUM :Hurnified OM IOM :Inert Organic Matter
Figure 4 Model RothC-26.3, structure of C flow (from Coleman and Jenkinson, 1996. Rothmodel for the turnover of carbon in soil; with kind permission of Springer-Verlag).
26.3-A
Not all decay processes induce microbial growth. The model CENTURY assumes that the lignin fraction of plant debris flows directly into the SLOW pool with the release of CO,, but without any concomitant contribution to the microbial pools, thus simulating the activity of exoenzymes (Fig. 5 on p. 268). The model VERBERNE assumes that dead microbes are composed of three fractions: nonprotected microbial material, the same material but protected against immediate decomposition, and slowly decomposing (recalcitrant) compounds (Fig. 6 on p. 269). These fractions are degraded by microbes with the formation of CO, and microbial biomass. However, it is also assumed that C flows from the recalcitrant to the protected pools and from the protected to the stable soil organic pools, without CO, production or microbial growth. Exoenzyme activities are not simulated in the models RothC-26.3 and NCSOIL. The microbial biomass C represents only a small percentage of total organic C. It is, however, dynamically important, since it recycles with a short half-life and is fed by most catabolic processes (except those representing exoenzyme activity). In spite of this central function, models use different representations of the microbial pool and associated flow webs. The model RothC-26.3 represents the microbial biomass with one pool (BIO). In a previous version (Jenkinson et al., 1987), a distinction was made between the autochthonous (BIOA) and zymogenous biomass (BIOZ). NCSOIL uses two pools differentiated by their decay rate constant. CENTURY distinguishes between the microbes responsible for the decomposition of the surface litter and the soil microbes that feed on the root litter and SOM. The greatest complexity is found in the model VERBERNE, which as-
268
JEAN-ALEX E. MOLINA AND PETE SMITH SURFACE UTTER
ROOT UlTER
Figure 5 Model CENTURY, structure of C flow (from Parton et al., 1993. Observations and modeling of biomass and soil organic matter dynamics for the grassland biome worldwide. Global Biogeochem. Cycles 7,785-809;with kind permission of the American Geophysical Union).
sumes that each soil has the ability to protect some of the microbial population from death. The protected microbes have a death rate of 0.5% per day. The nonprotected biomass dies at a higher rate (70% per day). Dead microbes are then split among three fractions, each with a different rate constant: the nonprotected microbial material, the same material but protected against immediate decomposition, and slowly decomposing (recalcitrant) compounds. The simulation of C flows in the model Q-SOIL differs from the concept of catenary sequence found in the other models ( & - e nand Bosatta, 1996). It considers one SOM pool that decays with a feedback loop into itself. The model is represented by a single rate equation. It is assumed that the dynamics of the SOM pool evolve as substrates and dead microbes interact. The SOM pool is divided into an infinite number of components, each characterized by its “quality” with respect to degradability as well as effect on the physiology of the decomposers. Quality (4)
MODELING CARBON AND NITROGEN PROCESSES IN SOILS 269
-c-flows -0-
N-flows
Figure 6 Model VERBERNE, structure of C and N flows. (Reprinted from Soil B i d . Biochem. 17, Van Veen et al.,Turnover of carbon and nitrogen through the microbial biomass, 747-756, copyright 0 1985, with kind permission from Elsevier Science Ltd, The Boulevard, Langford Lane, Kidlington OX5 1GB, UK.)
is quantified by a distribution function (D (q,q’),called the dispersion function; Z.,D(q,q’) = I), which expresses the fraction of decayed SOM of quality q’ changed by microbial assimilation and microbial death into SOM of inferior quality q. The rate equation for the model Q-SOIL (dp(q,t)lat) represents the dynamics of each SOM component of quality q. At any time t, the rate of feedback input into the SOM of quality q is equal to the sum of the contributions from each SOM component of quality q’
2 70
JEAN-ALEX E. MOLINA AND PETE SMITH
where D,(q’,t) is the time-dependent decay rate of the decomposers of quality 9‘. Accordingly, the models’ rate equation is dp(q,tYdt =
- Dc(q,r)/e(g) + Zq D(q,q‘) * D,(q’,t)
where p(q,t) is the concentration of the SOM of quality q, and D,.(q,t) is the decay rate of SOM of quality q, with a microbial efficiency e(q). The rate equation is quality dependent. Exact solutions to the rate equations are obtained analytically. Assumptions and simplifications required to derive the integral forms over q and t are discussed for various particular situations (e.g., Bosatta and Agren, 1994). b. N Flow In most models, N and C fluxes between organic pools are parallel, carried by organic molecules assumed to have the C-N ratio of the decaying pool (CNJ. As N enters the microbial pool, the amount of N mineralized or immobilized is controlled by the C-N ratio (CN,) and the efficiency factor (EFFAC) of the microbial pool and CNo. Comparison of the rates of N supplied and needed to maintain a constant CN, during microbial growth gives a criterium to distinguish between N mineralization, (dC/dt)/CNo > (dC/dt)*EFFAC/CN,
(9)
(dC/dt)/CNo < (dC/dt)*EFFAC/CN,
(10)
or immobilization,
where dC/dr is the rate of C flux from the decaying pool (Beek and Frisel, 1973). When the decaying pool is microbial (microbial successions), the scheme always induces N mineralization, since the influence of EFFAC (in the order of 0.6) is greater than that of the C-N ratios for the microbial pools (e.g., CNo = CN, = 6 for bacteria). This model accounts for net N mineralization but fails to simulate the N mineralization-immobilization turnover (MIT), whereby both processes occur simultaneously even in the absence of exogenous sources of C, as demonstrated with tracer I5N (Broadbent, 1966). If, however, N flows between the biomass pools per NH,+exclusively (Janson, 1958), instead of directly with organic compounds, a quantitative representation of the MIT is obtained (Molina et al., 1990; Nicolardot et al., 1994b). This model can also be viewed as similar to that of Kirkham and Bartholomew (1955), whereby the “unavailable atoms” box is replaced by several boxes serially linked to represent microbial successions. Whether N is carried by organic compounds or NH,+when microbes are feeding on organic debris is still debated. Similarly, whether immobilization feeds on NH,‘ or NO; and the extent to which NH,+is preferentially used by nitrification over immobilization are important issues that are not properly documented (Barak et al., 1990b; Hadas et al., 1987; Hadas et al., 1992; Jarvis et al., 1996). Nitrification is simulated by some models in great detail. Since the autotrophic
MODELING CARBON AND NITROGEN PROCESSES IN SOILS 27 1
nitrifiers obtain their energy from the oxidation of NH,' and NO;, the quantitative description of the transformations can be directly linked to the nitrifiers' growth for which much information is available. Grant (1994) linked growth to the difference between CO, reduced by the energy of NH,+oxidation and the energy expanded for anabolic processes. Laudelout ( 1976) expressed the rate of nitrification and nitrifiers' growth by combining the Michaelis-Menten kinetics dCldt = ( k , * E
* C)/(C, + C )
to the Monod growth equation dBldr
= (y ,
* C)/(C, + C ) ) * B
and assumed that the enzyme concentration (0responsible for the oxidation of NH,' (or NO;) at concentration C was proportional to the number of nitrifiers ( B ) E=k,*B
Values assigned to the parameters C ,, C,, and,,y , for the ammonium and nitrite oxidizers are obtained from the literature. The determination of k , is avoided by the introduction of the molar yield R
= dBldC =
ymax/(k,*k2)
shown to be constant and easier to measure than k,. Notice that for high concentrations of substrate, C # C, + C , and its rate of oxidation can be expressed in terms of molar yield as dCldt = (Y,~,,/R)B
(11)
It should be recognized that these approaches are based on the behavior of physiologically homogeneous nitrifiers population growing in liquid cultures, or soil columns continuously perfused with nutritive solutions. Sabey et al. (1969) showed that nitrification in soil proceeds as a zero-order reaction rather than exponentially. A stochastic interpretation of this kinetics was proposed whereby nitrification results from the cumulative effect of short pulses of ammonium oxidation from microbial clusters (Molina, 1985).Most models describe nitrification in soil as zero-, or first-order kinetics with respect to NH,'. Some models treat nitrification as an instantaneous process since it occurs more rapidly than net mineralization, which controls the limiting rate of NHd formation. This approach fails when N input is large and sudden as would happen when N fertilizer is added. 0therwise, it is valid for models with large computational time steps (2month) and avoids the unresolved issue of competition between immobilization and nitrification for NH,'. Denitrification in soil is driven by the energy derived from the decay of organic compounds. Under reduced partial pressure of 0,, the electron flow reduces successively NO;, NO;, and N,O with the formation of the end product N, (Betlach
272
JEAN-ALEX E. MOLINA AND PETE SMITH
and Tiedje, 1981). Most models simplify the expression of the denitrification rate by multiplying the decay rates of organic pools by a constant that expresses NNO; reduced per unit of organic C decayed. Different values are found for the constant, which can be computed on the basis of the stoichiometric equation for denitrification, obtained experimentally by relating CO, production to NO; consumed, or estimated by calibration of the model (Beauchamp et al., 1989). More elaborate models exist that consider the influence of the denitrifiers, 0,, and simulate the production of N,O. Leffelaar and Wessel (1988) make explicit the rate of reduction (electron consumption, dEJdt) of 0, ( i = 1) by strict aerobes, or NO;, NO;, and N,O ( i = 2, 3,4) by denitrifiers. This rate is computed by assuming that it is equal to the rate of electron produced during microbial growth dEJdt
=
(y?/Rimax) B
+ mi(Ei/E)B
whereby the first and second terms represent the electron flow during microbial growth (refer by analogy to Eq. [ 111) and maintenance (EJE being the fraction of electron used by the acceptor i, and mithe microbial maintenance coefficient), respectively. A double Monod equation with two half-saturation constants (Kc, Ke) y.I = ymax 1 [C/(Kc
+ C ) )(E/(K, + E ) ]
expresses the growth rate constant at low C and E substrate levels. Values for all the constants of the model were obtained by calibration. This approach is used in the model DNDC to simulate N,O production over large areas (Li et al., 1994b). Denitrification is maximum under anaerobiosis. As the partial pressure of 0, increases, a fraction of the electron flow is captured by 0,, thus reducing the rate of denitrification. Conversely, nitrification-the process that provides NO; for denitrification-is maximum under aerobiosis. In terms of quantitative simulation, it is critical to identify the range of partial pressure of 0, within which nitrification and denitrification occur simultaneously, albeit at reduced rates.
C. MODIFIERS OF RATE CONSTANTS Rate “constants” (k, e.g., Eqs. [4] and [S]) are constant for a given set of biotic and abiotic conditions. For nonoptimum environmental circumstances, the simplest way to modify the maximum value of k is by multiplication by a reduction factor p-the infamous fudge factor-ranging from 0 to 1.The environmentalfactors considered by the SOMNET models are shown in Table I. The interaction between several biotic and abiotic conditions (i = 1 , 2 . . .) should be ideally quantified by one reduction factor p obtained by a factorial experiment, which is seldom performed. In practice, p is set equal to either the product (p = II p i ) or the lowest value of the pi.An original approach presented in the model CANDY
MODELING CARBON AND NITROGEN PROCESSES W SOILS 273
consists of the multiplication of the time step by the reduction factor to define a biologically active time (Franko et al. 1995). There is no accepted terminology to distinguish between the maximum (k) and reduced (pk) values of rate constants. In general, values given with models' descriptions correspond to optimum conditions (Table I). Otherwise, the reader has to be aware of the experimental context. Values given by Stanford and Smith (Eq. 5) are maximum. Other rate constants provided in this review correspond to suboptimal conditions.
1. Temperature, Oxygen, and Water The literature abounds with studies of the effect of temperature on microbially mediated transformations in soil, either expressed as a reduction factor or the Arrhenius equation (e.g., Stanford ef al., 1973; Laudelout et al., 1974; Nyhan, 1976; Addiscott, 1983; McClaugherty and Linkins, 1990; Nadelhoffer er al., 1991). Nicolardot (1994a) cautions against the use of an overall relationship to characterize different transformations made of several intermediate steps. Water and oxygen have a major effect on the microbial physiology. The simulation of 0, concentrations in soil requires some consideration of the spatial heterogeneity of the soil-in particular, the distribution and size of soil aggregates. Several models of 0, distribution in soil aggregates have been proposed (Grant, 1991; Sierra and Renault, 1996). It is possible to avoid the complexity of 0, transport models by defining the extent of anaerobiosis on the basis of soil pore space filled with water (WFPS). There is an optimum soil water content for aerobic microbial activity whereby the rates of 0, and substrate diffusion are balanced for maximum microbial activity (Skopp er al., 1990). Doran ef al. (1988) showed that the optimum was the same for soils of various textural classes and corresponded to about 60% WFPS. Denitrification was negligible up to 70-75% WFPS, then increased exponentially to reach a maximum at 90% WFPS. Maximum rates of nitrification occurred at 44-59 and 57-72% WFPS for coarse and medium-to-fine textured soils, respectively.
2. Clay and Nitrogen Soil clay content and total SOM are correlated. Various schemes simulate the effect of clay on rate equations to obtain SOM accumulation. In the model CANDY, the fraction of inert C is equal to the fraction of clay plus fine silt (I 6 pm) multiplied by a factor ranging from 0.4 to 0.6 (Franko et al., 1995). In the model RothC-26.3, the effect of clay is quantified by the ratio CO,/(BIO + HUM) = 1.67 (1.85 + 1.60exp (-0.0786% clay), which increases the flow of C into the BIO + HUM pools in soils with higher clay content (Coleman and Jenkinson, 1996). Van Veen et al. (1985) contrasted the dynamic of ''C-glucose and l5N-NH:
2 74
JEAN-ALEX E. MOLINA AND PETE SMITH
between clay and sandy loam soils. The observed kinetics of total and tracer N-inorganic, C-biomass, and C-residual were accounted in the VERBERNE model with two texture-dependent factors: the capacity of a soil to protect microorganisms (MAXPB) and the fraction of microbial products remaining in the vicinity of survivors (FOPV). MAXPB was calibrated to 75 and 28 mg C 100 g-’ dry soil, and FOPV to 0.7 and 0.5 for the clay and sandy loam soils, respectively. The preferential accumulation of SOM in soil with high clay content can be simulated by lowering the rate constants of SOM pool decay. Functional relationships between the reduction factor and the clay content or CEC, derived from the results of Sorensen (1975), have been introduced in the models DAISY and RothC-26.3. At the highest clay content, the rate constants are reduced by 60%. The model CENTURY distinguishes between the effect of the sand, silt plus clay, and clay content of soils. Clay controls the stabilization of C from the slow to the passive SOM pools. The decay of soil microbes is modified by the silt plus clay content, whereas the split of C from decayed soil microbes to leached C and passive SOM pool is controlled by the sand and clay content, respectively. N is an essential element for microbial. growth, which will be maximal when enough N is assimilated to maintain the microbial C-N ratio (refer to Eqs. [9] and [ 101.Plant debris with high N content (e.g., legumes) decay more rapidly that those with low N content. Hunt (1978) expressed the exponential decay of debris by assuming that they were composed of two pools: A
=
Soexp(-k,f)
+ (l-So)
exp(-k,f)
(12)
where A is the fraction ( 5 1 ) of debris left at time f from 1 unit of initial material [So+ (1 --So)]decaying with rate constants k , = 0.044 d-’ and k2 = 0.00071 d-l. A good fit between experimental and simulated data for a wide variety of material (e.g., casein, blue grass, oak leave, cellulose) was obtained by relating Soto the N-C ratio of the original debris:
So = 0.070 + 1.11 (N/C)”3
(13)
Hunt selected the N-C ratio to calibrate Eq. (12) but mentioned that lignin could also have been used. These decay kinetics should be used with caution.According to Eq. (13), any material without N will decompose very slowly and at the same rate. Glucose, cellulose, and lignin decay at different rates. The contribution of inorganic N present in soil or provided by net N mineralization should also be considered. The influence of N on decay rates is treated differently in the model NCSOIL. The rate constant is determined for nonlimiting levels of N. Under N stress, the rate constant is reduced by a factor (pN)that is a function of a variable with the units of a C-N ratio (Acm): the rate (no N stress) of debris decay per day divided by N available (inorganic N already present plus rates of N mineralization from the debris and net mineralization from the SOM pools per day). Calibration of the reduction factor pNshowed that its lowest value (0.2) was reached for Acm > 120. Subsequent adjustments,if required, are achieved by reducing the
MODELING CARBON AND NITROGEN PROCESSES IN SOILS 275
microbial efficiency, thus increasing the fraction of decayed C dissipated as CO, (Molina et al., 1983).
III. MODEL VALJDATION A. THENEEDTO CRITICALLY EVALUATE MODELS There has been much discussion recently on how to validate models. It has been argued from a philosophical point of view, that in the same way as a scientific hypothesis cannot be proved, a model cannot be validated in absolute terms (Addiscott et al., 1995).A hypothesis might be rejected if a body of evidence arises that is contrary to the predictions of that hypothesis. But even if no evidence is found to contradict this hypothesis, it cannot be regarded as proved. Similarly, if a model performs well in many different environments and conditions, one cannot exclude the possibility that in any new situation the model may perform less wellin this sense a model can never be validated. Instead, it is more correct to describe a model as validatedfor use in a particular situation. Some authors prefer not to use the term validated at all (Addiscott et al., 1995; J. U. Smith et al., 1996). There is, however, a scientific and societal need to provide information on how widely applicableagiven model is and on the uncertaintiesassociated with that model’s predictions, especially when the implications reach beyond the scientific community into wider society (e.g., when predicting the consequences of global climate change). A model that has met criteria for good model performance in a wide variety of situations can be used with more confidencethan a model that has not. In these situations, the rather absolute term Validated is often applied. The term validated does not mean that the model is 100%accurate in any conceivable situation, instead it merely means that the model has been widely tested and has shown acceptable performance. The term perhaps conveys the wrong message about the confidence we have in the predictions of our models, unless we define the term carefully. Validation is nevertheless in common use, so we will also use it here after giving our definition. Our definition is probably the same as that intended by most people who use the term, i.e., the extensive evaluation of a model to assess its performance by comparing it (quantitatively)to measured data in a range of situations. Many attempts to evaluate models in the past have relied upon a visual-graphical comparison of simulated values with measured data. Although this has a very important role in model evaluation and comparison, there is a pressing need to attempt to quantify model performance so that subjectivitycan be minimized. Workers such as Loague and Green (199 1)and Hagen et al. (1993) have suggested methods that can be used to this end, and a recent review (with some novel approaches) of quantitative methods to evaluate and compare models was provided by J. U. Smith et al. (1996).
276
JEAN-ALEX E. MOLINA AND PETE SMITH
B. VALIDATION OR CALIBRATION? As explained previously, validation is the demonstration that a model is anchored to the reality of measured data. The linkage between the experimental information and theoretical models is an increasingly debated issue as the misuse of models to predict future agronomical and ecological events has become apparent (Christensen, 1996; Elliot et al., 1996). A model with n pools is defined by a system of n differential or net rate equations:
,...Ym,Zm+I ,...Z,K) dZjldt = Fj(Y1,...Ym,Zm+ ,,...Z,,,K) dYidr
= Fi(Y,
(14)
+
where Y,(i = 1 . . .m) and Zjcj = m 1 . . . n) are the elemental concentrationsof (1) the m pools, which are experimentally recognized (e.g., total C, microbial biomass, CO,, NH,‘,NO;); and (2) the n-m hypothetical pools, which have not been experimentally defined. K is a vector constant made of the parameters of the processes (e.g., rate constants,efficiency factors, C-N ratios). The interdependency between the pools represented by the model’s structure is expressed by making each net rate equation dependent on all the concentrations and parameters of all the pools. After the parameters and initial (t = to) concentrations have been measured and introduced in the model, validation of the model is obtained by a good fit between simulated and experimental data for events happening after the initial time (t > to) (Molina et al., 1994; see also J. U. Smith et al., 1996).The validation process is hindered by the fact that initial values of the speculated pools and some parameters are impossible to measure. Instead, the unknown data are obtained by calibration of the model against the measured concentrations (Yj)for t > to. The use of the experimental information to both validate and calibrate weakens the validation process. Calibration and validation should be performed against different sets of data contrasted by sites and treatments. Too often, calibration and validation are mingled within a single set of data.
(3.)
C. NONTRACER DATA:LONG-TERM SIMULATIONS Soil inorganicN status can change dramaticallyover the course of a year. For most purposes, when evaluating a model for its simulation of soil inorganic N, a dataset of a few years duration can be considered long term. Areasonable test of a model’s ability to simulate N turnover is if it successfully simulates aspects of N tumovere.g., after inorganic fertilizer application during plant growth and after incorporation of crop residues for a few years. Evaluations of the performance of the N components of models rarely use datasets longer in duration than a few years (e.g.,
MODELING CARBON AND NITROGEN PROCESSES IN SOILS 277
de Willigen, 1991). For C turnover in SOM and for total soil N, however, much longer runs of evaluation data are required (P. Smith et al., 1996a).We will outline some of the methodological difficulties associated with the use of long-term soil C data to evaluate models. Glendining and Poulton (1996), in a recent review on this subject, identified a number of potential problems with long-term soil C data, including modification in sampling protocol, changes in the number and distribution of samples taken, problems of spatial variability and bulking protocols, inadequate consideration of changes in soil bulk density during an experiment, soil movement between plots, incomplete records, and changes in methods of measuring soil C. If we take differences in methods for measuring soil C as an example of a potential problem for model evaluation, we see that many different methods are available, including dry combustion, wet combustion, hot-water soluble C, Tinsley method, Walkley-Black method, and Tyurin method (Glendining and Poulton, 1996). Some attempts have been made to compare these methods (e.g., Kalembasa and Jenkinson, 1973), revealing that the methods can result in quite different estimates of soil C. Nelson and Sommers (1982) and Tiessen and Moir (1993) provide comprehensive reviews of methods available for soil C determination. Compounding the analytical problem is the measurement of different fractions of soil C. In some experiments total C is measured in others organic C; and in some, especially in eastern Europe, “humus” is measured (e.g., Yakimenko, 1996), which corresponds to what is more commonly referred to as SOM. All of these methods are used for estimating soil C, and many different definitions of soil C are represented among different experimentsparticipatingin SOMNET(I? Smith et al., 1996c),demonstratingthat care is needed when interpreting the long-term measurementsfor comparison with models. However, Glendining and Poulton (1996) conclude that despite their limitations, well-monitored long-term experiments are invaluable for testing the validity of SOM models but should always be interpreted in light of any known limitations. There have been many studies reported in which models have been tested against data from long-term field experiments. Two examples from many include the testing of CENTURY against the long-term Pendleton Residue Management Experiment (Parton and Rasmussen, 1994) and the testing of the Rothamsted Carbon Model against the Rothamsted Classical Experiments (e.g., Jenkinson, 1990). Often such studies act not only to test the model, but also help to improve our understanding of the fundamental processes underpinning C and N turnover models. The testing of models using long-term datasets has assumed a new importance with the advent of modeling the effects of global environmental change. Since changes in soil organic C occur slowly, an essential first step in evaluating the utility of models for this kind of application is to test them against long-term datasets (Powlson et aZ., 1996). In the most comprehensive exercise of its kind to date, nine SOM models were evaluated and compared using datasets from seven long-term experiments representing different climates, management practices, ecosystems, and land uses within the temperate climatic zone (P. Smith et al., 1997b). Few of
278
JEAN-ALEX E. MOLINA AND PETE SMITH
the models were able to simulate soil organic C under all land uses (arable, grassland, and forestry). The models were compared quantitatively using methods described in J. U. Smith et al. (1996) and fell into two groups-one containing six models with significantly lower root mean square errors than the other group, which contained three models. At least some of the differences between models were accounted for by different levels of site-specific calibration. Among the main conclusions were that (1) models still require much work if they are to be used in a truly predictive fashion in the future (i.e., most still require considerable site-specific calibration), and (2) other aspects of C turnover, such as shorter-term evolution of and sensitivity to CO,, also need to be considered in the future (P. Smith et al., 1997~).
D. TRACERDATA Several methodologies have been used to compute SOM decay rate constants with tracer data (Paul and Van Veen, 1978). It is convenient to categorize them on the basis of the number of soil samples required (one, two, or multiple samples).
1. One Sample The method is based on the detection of I4C. The mean age (or mean residence time, MRT) of the sample is computed from the radiocarbon decay equation MTR
= -k
ln(MAo)
(15)
where k = 0.0001209 yr-I is the 14C decay rate constant (half-life: 4730 yr), A and A. are the measured 14C-specificactivity of the sample and the modern standard, respectively. The ratio MAo is often expressed in terms of the per mil enrichment (S140/oo = 1OOO(A -Ao)/AO). The half-life of the SOM can be computed directly from its MRT if the SOM of the sample is assumed to be represented by the SOM pool of the model described by Jenny (Eqs. [l] and [2] apply to N but are also valid for C or dry mass). This model has the peculiarity of presenting a relationship between the decay rate constant ( k , ) and the MRT (k, = l/MRT) if the SOM pool is in a steady state with respect to the yearly inputs (It2). It can also be shown that the level of the SOM equilibrium value Ne (refer to Eq. [2]), k , , and k, are related (Bartholomew and Kirkham, 1960)
(Ne= k,/k,)
(16)
Equation (15) has to be adjusted to account for exogenous I4C input, particularly for soils sampled after 1954 that have been enriched with 14C as a consequence of residual bomb effect. Several corrections have been proposed to account for this complication (O’Brien, 1984; Harkness et al., 1991; Hsieh, 1993). The effect of I4C discrimination during microbial decomposition of plant debris is not
MODELING CARBON AND NITROGEN PROCESSES I NSOILS 279 important and does not warrant correction of the MRT (Campbell et al., 1967; Scharpenseel, 1971; Martel and Paul, 1974). The method can be applied to isolated soil fractions. Humic acids, residues from 6M acid hydrolysis, and material associated with coarse clay have been found to be the most stable, with rate constants ranging from 0.0005 to 0.0009 yr-’ . Acid hydrolyzable C, materials associated with fine clay, and dissolved organic C from lakes, streams, and wetlands decay rapidly (k 5 0.001 yr-l). Residues from 0 S M acid hydrolysis and materials associated with fine or coarse clay have intermediate stabilities.Variable rate constants have been reported for humic and fulvic acid fractions (Martel and Paul, 1974; Martel and Lasagne, 1977;Anderson and Paul, 1984; Schiff et al., 1990).
2. Two Samples Hsieh (1992, 1996) assumed that the SOM in the samples was made of two pools. The I4C mass balance equations for soils from two sites (“paired-plots”) sampled at the same time but contrasted by different cultivation practices are: A,lC,l = AdCsI
+ AaI(C,,
- CSJ
A,2C,2 = As2Cs2 + Arr2(Ct2 - CS2) where A and C represent the 14C-specificactivities (14C-C ratio) and concentrations of the pools, respectively. The indices stand for the total SOM (t),and stable (s)and active (a) SOM pools, at sites (cultivation) 1 and 2. The activity of the active pool is computed by Aa = (C,,A,, - Cry4r2)4Ctl- C,) if it is assumed that (1) cultivation does not affect the 14C-specificactivity of the active and stable pools (A, = A,, = Au,; and As, = As2), and (2) cultivation increases or decreases the active pool Ca = (C, - C,) without significantly affecting that of the stable pool (C,, = C,,; but C,, # Cf2). The same approach is used to compute the activity and concentration of the stable pool. It suffices to select a second set of paired-plots with different specific activity Arrand A’,’ caused, for example, by different 14C input from the bomb effect. The corresponding mass balance equations are + A$, - C,) Arc, = A’,C’, = AsCs + A’JC‘, - C,)
from which As and CScan be computed if it is assumed that 14C addition by the bomb effect is felt by the active pool only. The pools’ MRT are computed from the radiocarbon decay Eq. (15) corrected for the bomb effect (Hsieh, 1993). Alternatively, the stable pool size and specific activity can be measured from the I3C-I2C method, which will be considered in the next section.
2 80
JEAN-ALEX E. MOLINA AND PETE SMITH Table Ill Amount and Stability of SOM Represented by a Model with Two Pool@ Decay rate constant (yr-’)b
Site
Covedmanagement
Belize
Illinois
Missouri
Sugarcane 15-20 yr Citrus orchard, >40 yr Forest Morrow Plots Continuous corn, no fertilizer Corn-oat-clover, no fertilizer Com-oat-clover, fertilizer Kentucky bluegrass Sanborn fields Continuouscorn, no fertilizer Continuous corn, fertilizer
Active pool
Stable pool
0.125-0.200
0.0026-0.0040
Amount of active pool (% of total C) 23-30 5.5 48-54
0.0149
0.00034 15 35 65 64
0.0294
0.0012 11 39
“From Hsieh, 1992, 1993, 1996. hDecay rate constants do not change with type of cover or management; rate constants are not at their maximum values but reflect site conditions.
Some results obtained by Hsieh’s approach are showed in Table 111. The active pool most likely encompassesmany labile fractions such as the microbial biomass, as well as some organic debris, since fractions smaller than 1 mm were included in the samples. It is important to remember that these MRT values do not demonstrate the presence of two SOM pools in soil; instead, they were computed on the basis of a model made of two SOM pools. This method has not been applied to isolated soil fractions from paired-plots.
3. Multiple Samples Changes in the activity of naturally occumng and introduced 14C, as well as in the concentration of stable 13C isotopes, have been used to document the kinetics of SOM in multiple samples. The decrease in I4C activity of soils amended with tagged substrates documents several processes: the kinetics of substrate decay and the transfer of the tracer into other SOM fractions,each with its particular dynamic. Decay rate constants of added plant debris computed on the basis of one-pool models are not stable but decrease with time as the data reflect the stabilization of the tracer into nonaccounted pools of greater stability. A more accurate interpretation of the data is rendered with multi-SOM pool models. Jenkinson et al. (1991) collected from the literature data from 3 1 experiments
MODELING CARBON AND NITROGEN PROCESSES I N SOILS 28 1
performed under varied climatic conditions and documenting the long-term decay of 14C-tagged plant material. Simulation of these kinetics by the Rothamsted model gave computed data that accounted for 83% of the experimental variance, the worse fit occurring with two soils with high levels of allophanes. The model's parameters and initial values associated to the SOM pools were calibrated by an independent set of data collected at Rothamsted.The variability in the data could be accounted for with differences in the climatic data and by adjusting the ratio of decomposable to resistant plant material. Extraction of C and N biomass by the fumigation procedure and the determination of its 14Cactivity and 15Nconcentration has given a new dimension to kinetic studies of C and N processes in soil. Oades (1988) has reviewed the literature about the stabilizing influence of clay on the microbial biomass. Similar conclusions were obtained by considering the decay rates of amino acids and chemicals found in the soluble fraction of weak acid hydrolysis after tagging with I5N inorganic N (Kelly and Stevenson, 1985). Interpretation of microbial kinetic tracer data cannot be done with one-pool models. Few multipool models have been tried on these data. Van Veen et al. (1985) justified several features of the VERBERNE model on the basis of such microbial kinetics. Nicolardot et al. (1994b) tested the model NCSOIL against the kinetics of tracer and total CO,, biomass, and inorganic N in three soils (Table IV). There is a wealth of experimental data in the literature that documents the kinetics of 14C introduced in soil with specific substrates (e.g., Martin and Haider, 1986; Buyanosky et al., 1987; Sgrensen, 1987; Amato and Ladd, 1992). These data should be used to calibrate and improve the multipool models. Information about the half-life of stable SOM pools can be obtained from the analysis of data documenting changes in I4C activity that resulted from the bomb effect or I3C stable isotopes. O'Brien and Stout (1977,1984) have documented the distribution of 14C-12C ratios in soil profiles over pasture and forest between 1959 and 1973 when the bomb effect was intense. They developed a multilayered model that took into consideration input from the I4C via the bomb effect and a soil C diffusivity factor between the layers. They computed a decay rate constant of 0.016 yr- for the SOM and a diffusivity of 13 cm2 * yr-l. By considering the variations of 14C activity in the soil profile, it was inferred that 16% of the SOM was very old, with a MRT of 7000 yr. Comparison between wormless and worm-infested soils showed that earthworms increased the level of metabolically active C, the SOM decay rate, and the soil diffusivity. The isotopic I3C-l2C ratio, expressed as 813 %r7
= 1000 [(13C/'2C) - (13C0/'2CO)]/ (13CO/12CO)
is different in C , plants (Si3: -30 to -20) and C, plants (814,: - 18 to -8) (Bender, 1968). Soils that have supported C, then C, plants (or vice versa) have intermediate Si3%0. Linearity between SI3 %O and the fraction X4 of C, material can be
Table IV
Measured (M) and Simulated (S) Data in the Glucose and Cellulose k t m e n t s for the First 56 Days of Incubation
Treatments
c0,-'4c
c0,-C"
Soils
Bi~rn.-'~C
Biorn.-C
Inorg.-N
Inorg.-I5N
Days
M
S
M
s
M
S
M
s
M
s
M
5
0 7 14 28 56
0.0 3M.6 511.7 685.3 951.5
0.0 390.8 494.2 663.1 969.9
0.0 241.7 269.8 293.7 3 18.6
0.0 287.6 308.7 331.O 359.3
426.2 659.5 546.6 504.2 413.7
426.2 539.6 494.3 466.9 424.4
0.0 167.0 128.8 120.8 86.9
0.0 177.8 137.5 98.1 54.4
33.0 14.2 22.6 30.6 53.6
33.0 13.2 24.0 37.3 60.9
33.0 5.5 7.8 8.5 10.5
33.0 3.8 6.5 9.0 12.5
0 7 14 28 56
0.0 355.7 520.5 712.2 988.5
0.0 347.5 483.4 658.1 966.5
0.0 2 10.4 281.7 325.0 358.4
0.0 244.3 297.9 326.1 356.0
426.2 680.5 483.9 460.0 389.5
426.2 535.2 505.3 473.3 427.5
0.0 131.8 88.4 69.4 50.6
0.0 173.5 148.6 104.5 57.6
33.8 17.4 24.8 34.1 52.7
33.8 15.5 22.8 35.7 58.0
33.8 8.6 10.7 12.5 12.4
33.8 8.3 9.1 11.4 14.6
0 7 14 28 56
0.0 354.0 416.2 513.1 660.5
0.0 307.2 386.9 512.9 633.4
0.0 254.9 267.1 280.8 294.8
0.0 242.5 268.1 295.1 315.0
249.6 478.8 358.3 331.7 297.3
249.6 436.3 391.1 359.5 335.6
0.0 129.7 101.1 98.1 75.3
0.0 215.6 166.7 118.9 86.9
73.7 49.6 52.4 57.8 66.9
73.7 42.0 51.4 62.7 72.9
73.7 35.9 35.8 35.7 37.6
73.7 33.6 37.6 41.4 44.3
Grkux-les-Bains Glucose
Cellulose
Mons-en-Cham& Glucose
Cellulose 0 7 14 28 56
0.0 283.0 398.3 516.7 673.9
0.0 267.3 399.7 540.5 670.1
0.0 197.1 255.6 287.6 314.6
0.0 202.6 280.8 322.7 343.0
249.6 561.0 365.1 317.6 293.4
249.6 369.1 372.9 349.4 326.0
0.0 166.7 106.3 16.5 74.0
0.0 148.4 148.5 108.9 77.0
72.7 65.2 66.5 71.5 75.9
72.7 54.2 55.4 65.0 75.6
72.7 54.7 50.6 54.7 49.0
72.7 47.1 43.8 45.9 48.5
0 7 14 28 56
0.0 344.8 441.9 583.0 754.2
0.0 374.2 442.2 556.2 766.9
0.0 254.7 302.1 321.1 339.1
0.0 319.1 337.0 356.0 380.1
187.4 382.9 322.9 286.6 269.3
187.4 330.6 309.6 302.6 286.4
0.0 127.9 113.3 94.7 72.0
0.0 151.3 117.1 83.6 46.4
30.5 7.5 11.9 19.8 35.2
30.5 8.1 14.6 22.9 38.6
30.5 3.0 4.1 5.3 9.1
30.5 3.5 5.9 8.2 11.6
0 7 14 28 56
0.0 245.0 391.1 559.2 137.5
0.0 251.7 386.1 522.9 741.3
0.0 171.3 253.8 306.6 347.8
0.0 202.6 280.8 322.7 354.4
187.4 457.3 329.5 277.8 246.6
187.4 327.6 34 1.O 327.9 299.8
0.0 169.5 96.6 70.9 46.0
0.0 148.4 148.5 108.9 59.8
31.3 18.4 19.9 27.2 44.8
31.3 9.9 10.0 18.7 35.9
31.3 13.1 11.4 12.7 16.8
31.3 7.1 5.2 7.4 11.2
Louvain Glucose
N W m
Cellulose
‘TO,-C, C02-14C, biomass C, and biomass 14C are expressed in mg C * kg-’ dry soil; inorganic N and inorganic ISN in rng N * kg-’ dry soil. Reprinted from Soil Biol. Biochem. 26, Nicolardot et al., C and N fluxes between pools of soil organic matter, 235-243; copyright 0 1994, with kind permission from Elsevier Science Ltd, The Boulevard, Langford Lane, Kidlington OX5 1GB. UK.
2 84
JEAN-ALEX E. MOLINA AND PETE SMITH
assumed, since the 12Cisotope is about 100 times more abundant than I3C:
x4= (8"-
s:3)/(8143- 8;')
x, = 1 - x4 where X, is the fraction of SOM of C, origin in the soil sample (Gem et al., 1985). The effects of isotopic discrimination during the decay of plant debris have been discussed by Balesdent et al. (1987) and Agren et al. (1996). Veldkamp and Weitz (1994) reported that spatial variability was the main source of error on S13%0data in pasture and forest soils. The gradual replacement of SOM of forest origin by that derived from sugarcane has been documented in Brazil by Cem (1985). The kinetics of SOM change were measured three times (0, 12, and 50 yr after planting of sugarcane). In addition, each soil sample was fractionated to obtain the decay rate constant of three granulometric fractions (0-50 prn organo-mineral complex; 50-200 pm plant de-
1960 1980 1970 1080
TIME ( years ) Figure 7 Changes in amount and origin of soil organic C (0-20 cm depth) accompanying longterm cultivation of fertilizedcorn on soil previously cultivated to C-3 forage for 61 years (from Balesdent et QL, 1988. Soil organic matter turnover in long-term field experiments as revealed by carbon-13 natural abundance. Soil. Sci. Soc. Am. J. 52,118-124; with kind permission of the Soil Science Society
of America).
MODELING CARBON AND NITROGEN PROCESSES IN SOILS 285 bris and sand; 200-2000 pm plant debris). Plant debris of forest origin decayed very rapidly ( k = 0.1 yr-') until total removal. They were replaced by sugarcane debris, which attained steady-state level about 12 years after the breaking of the forest. The 0-50 and 50-200 pm fractions converged toward common steady-state levels, the material of forest origin decaying exponentially ( k = 0.025 yr-l) for the first 10-20 yr, but with a lower rate constant thereafter. A similar study was performed in Missouri with soils from the Sanborn Field (Balesdent et al., 1988; Wagner, 1991).It involved the succession of three covers: native prairie (C,), forage or wheat (C,), and corn (C,) (Fig. 7). In the top 20 cm, the soil of native prairie origin decayed rapidly ( k = 0.06- 0.10 yr-l) for the first 10-15 yr, then stabilized with a material that decayed linearly with time ( k = 0.0032 yr-* under cropped wheat with tillage), and more slowly under no tilled timothy ( k = 0.0012 yr-l). The SOM C from forage (C, plants) became progressively more resistant and decayed with an initial k = 0.089 yr-l, which decreased to 0.049 yr-' after 36 yr of pasture. Higher decay rate constants were computed for the C , plants by dividing the equilibrium level by the annual rate of C input (Eq. [ 161).Values of 0.21,0.22, and 0.16 yr-' were obtained for wheat stubble and straw, wheat, and timothy stubble, respectively. Corn stubble, which has lower lignin content than wheat, degraded more rapidly ( k = 0.53 yr-I). These higher values may reflect rates of annual input underestimated by C input from roots. Two widely used agronomic crops-corn and soybean, C,, and C, plants, respectively-can be used to study the SOM dynamics by E l 3 analysis. Huggins et al. (1997) have considered 15 treatments of continuous corn, continuous soybean, and corn-soybean rotations of various length. Although the total SOM did not change over 10 years, differences in E l 3 between the rotations were observed. Isotopic 13C-12Cratios from one soil sample are not sufficient to compute decay rate constants.They have, however, given insights about variation in SOM stability with the soil profile (Schwartz et al., 1986; Volkoff and Cerri, 1987), soil particulate fractions (Balesdent et al., 1988; Martin et al., 1990; Cambardella and Elliott, 1992), the activity of earthworms (Martin et al., 1992), and tillage (Balesdent et al., 1990).
W. MODEL APPLICATIONS A. USINGMODELSAS
A
RESEARCHTOOL
Models are hypotheses of the dynamics of C and N in soil. For example, whether N is carried between SOM pools as NH,+or as organic compounds are options that translate into two models with different structures. The corresponding two sets of simulated output data can be used to test the hypotheses at a given probability
2 86
JEAN-ALEX E. MOLINA AND PETE SMITH
level of acceptance or rejection. Thus, using experimental data documenting increased net N mineralization with increasing initial levels of NH: in the absence of a readily available source of C (Broadbent, 1965), Molina et al. (1990) showed that the model representingthe direct transfer of N with organic compounds should be rejected. Various statistical tests are available to compare simulated to experimental data. Rejection of a model can be based on a failed t-test on a null mean for the difference distribution between simulated and experimental data. Another approach is based on the biological limits imposed on parameters values. For example, microbial efficiency factors should be in the range 0-1, and C-N ratios in the range 5-12. Values for these parameters can be obtained by model calibration against experimental data: any computed value that would fall outside the acceptable range would indicate a faulty model (Barak et al., 1990a). As experimental data can help to decide which of two models representing mutually exclusive hypotheses is acceptable, models can be used to screen among experimental procedures those that give data relevant to C and N processes (Molina er al. 1994). Juma and Paul (1984) compared the I5N enrichment and extractibility ratio of NH,+mineralized during incubation and released by various extraction procedures to determined which of the procedures identified the potentially mineralizable N. A sensitivity analysis identifies the relative importance of the models’ parameters and initial conditions on output data. The questionnaire mailed to identify SOMNET models asked how “the sensitivity analyses have been performed.” About half of the modelers addressed this issue and indicated that the sensitivity analysis had been performed on some exogenous variables (such as temperature), and initial levels (such as percentage of crop residue incorporated).Only one modeler (Ryzhova, 1993) mentioned having performed a sensitivity analysis on the SOM pools’ turnover (model NAM SOM).
B. USINGMODELS TO IMPROVE AGRONOMIC EFFICIENCY AND ENVIRONMENTAL QUALITY Six years ago, De Willigen (1991) summarized the results of a comparison study of 14 models of soil-crop systems: Seven years ago the Institute for Soil Fertility Research organized a similar workshop. . . . In discussing the results, de Willigen and Neeteson (1985) came to the conclusion that the main problems were caused by the modelling of the soil microbiological processes. For the present workshop it seems that a similar conclusion can be drawn: the main difficulties in modelling the nitrogen turnover in the soil-crop system lie in the description of the soil processes, and of these the biological processes appear to pose the most serious problems.
MODELING CARBON AND NITROGEN PROCESSES IN SOILS 287 Agricultural producers do not have much confidence in the use of soil-crop models. The influence of biological processes on N fertilization is estimated with areaspecific empirical formulae. For example, in western Minnesota the amount of fertilizer (lb per acre-') required (N,,,) is computed by the formula
Nrec= (1.2 * YG) - STN - Npc whereby YG (bu. per acre-') is the realistic yield goal, STN is the amount of NO; (lb per acre - I ) in the top 24 inches of soil, and Npc is the nitrogen credited for previous crops in rotation. Values for Npc are given by tables that itemize leguminous crops grown 1 and 2 years before in the rotation (Rehm et al., 1994). Different formulae are used in other areas of the midwestern United States (e.g., Schlegel and Havlin, 1995). The need for process-oriented models that account for MIT, N losses by denitrification, and the interaction between crop residues, manure, and the SOM will intensify, since more producers are investing in the machinery equipped with the technology for global positioning and site-specific application (Robert et al., 1991). A potentially productive use of soil-crop models will be to simulate the behavior of multi- and inter-cropping systems to replace complex, long-term, and costly experiments. An interface between the models CERES-wheat and SOYGRO has been designed to investigate water stress on the physiological maturity of wheat and the emergence of soybean (Parsch et al., 1991). This study did not factor in soil biological processes.
C. USINGMODELS TO ESTIMATE GLOBAL C IN SOM AND FLUXES OF C UNDER GLOBAL ENVIRONMENTAL CHANGE In addition to their use in the laboratory, and at plot, field, and watershed (catchment) scales, models of C and N turnover have also been applied at regional and global scales. An example of a regional scale application is the use of the CENTURY model to predict the effects of alternative management practices and policies in agroecosystems of the central United States (Donigan et al., 1994). In this study, CENTURY was coupled with soil and meteorological databases with GIS capability to estimate carbon sequestration potential for 44% of the land area of the United States. It was demonstrated that conservation tillage practices and the use of cover crops could be used to increase soil C over about 40 years (Donigan et al., 1994). The Rothamsted Carbon Model has also been linked to GIS spatial databases to estimate the potential effects of climate and land-use change on soil C and CO, emissions from natural ecosystems in New Zealand. This study showed that the combined effects of ecosystem degradation and climate change could lead to significant net CO, releases over 40 years (Parshotam et al., 1996).
288
JEAN-ALEX E. MOLINA AND PETE SMITH
C and N turnover models have also been used at the global scale. Post et al. (1982; 1985) have used such models (e.g.. Rothamsted Carbon Model) to determine global terrestrial pool sizes and the distribution of C and N. C and N turnover models have also been used to estimate the effects of climate change (rising temperature and increasing CO,) on global soil organic C stocks. Models used include CENTURY (Parton et al., 1987; Potter et al., 1993; Schimel et al., 1994), the Rothamsted Carbon Model (Post et al., 1996), Nakane (Goto et al., 1993), Osnabruck Biosphere Model (Esser, 1990), IMAGE 2 (Goldewijk et al., 1994), and TEM (Melillo et al., 1995).An overview of the findings from these studies can be found in Post et al., 1996). Recent initiatives to evaluate the accuracy of C and N turnover models for use in global environmental change research include the Global Change and Terrestrial Ecosystems (GCTE) Soil Organic Matter Network (SOMNET P. Smith et al., 1996b,c; Gregory and Ingram, 1996; Powlson et al., 1997), which has collected detailed information and data on models and long-term experiments from around the world. This network provides a framework in which C and N turnover models can be evaluated for their suitability for global environmental change research. As described earlier, a comparison of the performance of nine SOM models when simulating datasets from seven long-term experiments has recently been completed (P. Smith et al., 1997c) and will be published with details of the complete exercise (Powlson et al., 1996; P. Smith et al., 1997b). The ability of six models (CENTURY, DNDC, ECOSYS, EPIC, RothC, and SOCRATES)to scale up simulated net C storage from site-specific to regional basis has been investigated by Izaurralde et al. (1996).
V. CONCLUSIONS AND FUTURE WORK The simulation of C and N processes in soil presents a paradox: Why do models with different numbers of SOM pools, with different half-lives, and organized in different structures give simulated kinetics that closely follow measured data? The answer is in the distinction between “good fit” of data and correct representation of the processes. Referring to Eq. (14), the number of unknown parameters and initial values is large enough with respect to the experimental data to ensure that calibration will find a set of values to provide a good fit. That is not to say that these models are empirical, since their parameters and initial values have a range of possible values that is restricted by their biological connotation. But the proof that they are a correct representation of C and N processes in soil has to be established with rigor. In particular, a strict adherence to the distinction between validation and calibration should be followed; and the use of tracer 13C, 14C,and I5N elements to distinguish between the kinetics of young and old SOM should reduce
MODELING CARBON AND NITROGEN PROCESSES IN SOILS 289 the risk of declaring a process-oriented model validated-even for one site-when it is not. We have come a long way since Jean-Baptiste Van Helmont’s measurements of tree growth in a pot of soil in 1652. Although there is much yet to learn about C and N transformations in the soil, it is fair to say that most of the major processes involved in C and N transformation are now understood, at least well enough to build mathematical models of the soil that are able to describe and predict transformations in the soil system reasonably well. We need to deal with a new set of problems in the future: namely, the use of our models in a more predictive manner. Many of our C and N models were built to structure our knowledge, to test our understanding, or to explain observed results. We are increasingly asked to use our models to make predictions about the future, be that to provide N fertilizer recommendations to farmers at the field or farm scales or to provide predictions of the likely environmental effects of land use or climate change at the national, regional, or global scales for policymakers. These types of questions present quite different problems from those confronted in the past. There is a pressing need to further improve our understanding so that we can use our models for predictive purposes, without the need for site-specific calibration. Related to using models predictively is the need to quantify the confidence we have in the predictions of our models, both for scientific rigor and for public accountability. To do this we also need to evaluate the performance of our models in an objective and quantitative way and to attempt to quantify the degree of uncertainty associated with our model predictions. A further major challenge that relates to the predictive use of models for the purpose of global change was raised by P. Smith et al. (1997a) and can only be solved by experimental approaches. The problem is that current models simulate global changes, for example, changes in temperature, by shifting along a single process response curve. This may occur through a purely physico-biochemical response (e.g., a QlO response) or may implicitly involve the soil biota (e.g., through increasing or decreasing the size of the microbial biomass). In either case, there is an assumption that the response of the system will remain predictable within calibrated limits, i.e., that the response surface remains the same. Current models assume that this is the case. The problem arises when global change affects the soil system in such a way that the system no longer responds to future change in the predicted way; i.e., there may be a change from one response surface to another. If global change affects the soil in this way, the future rate of a process may, in certain circumstances,be radically different from that predicted. Only through an improved understanding of how the soil system will respond to global change will we be able to confidently model future global change. This gap in our knowledge necessitates more research into the effect of global change on the soil system’s future response to perturbation (P. Smith et al., 1997a).
290
JEAN-ALEX E. MOLINA AND PETE SMITH
A number of other scientific problems remain unresolved. Both Christensen (1996) and Elliott et al. (1996) highlight a need for reconciling the conceptual pools we use in models with measurable fractions within the soil. Although our current models can be shown to perform adequately in many situations, this remains a challenge for C and N modelers in the future. Another area that has received little attention in the past is the integration of the soil biota into the process-oriented models upon which this review has focused. Models such as those of Hunt et al. (1 987) and De Ruiter er al. (1994) describe C and N transformations in terms of trophic interactions between functional groups in the detrital food web. Although some combined models have used both processand food-web-oriented approaches (e.g., McGill et al., 198l), two other recent reviews have called for a more systematic approach to the inclusion of soil biota in C and N models (McGill, 1996; Paustian, 1994). The role of soil biota in C and N models is discussed in more detail in Paustian (1994) and P. Smith et al. (1997a). Paustian (1994) also stresses the importance of modular programming methods to allow different models to be integrated. Despite recent progress, we are still not in the position whereby modelers can swap process description modules with each other. More progress on this front will enable better flow of information among different modeling groups and the improvement of our models of C and N transformations in soil.
REFERENCES Addiscott, T.M. (1983).Kinetics and temperature relationships of mineralization and nitrification in Rothamsted soils with differing histories. J. Soil Sci. 34,343-353. Addiscott, T. M., Smith, J. U.,and Bradbury, N. J. (1995).Critical evaluation of models and their parameters. J. Env. Qual. 24,803-807. Agren, G . I., Bosatta, E. ( 1996).“Theoretical Ecosystem Ecology-Understanding Element Cycles.” Cambridge University Press, Cambridge. Agren, G. I., Bosatta, E., and Balesdent, J. (1996). Isotope discrimination during decomposition of organic matter: A theoretical analysis. Soil Sci. SOC.Am. J. 60, 1121-1126. Alexander, M. (1977).Introduction to soil microbiology. Wiley and Sons, New York. Amato, M., and Ladd, J. N. (1992).Decomposition of I4C-labelled glucose and legume material in soils: Properties influencing the accumulation of organic residues C and microbial biomass C. Soil Biol. Biochem. 24,455464. Anderson, D. W., and Paul, E. A. (1984).Organo-mineral complexes and their study by radiocarbon dating. Soil Sci. SOC.Am. J. 48,298-301. Andrk, M.,Thikry, J. M., and Cournac, L. (1992).ECOSIMP model: Prediction of CO, concentration changes and carbon status in closed ecosystems. COSPAR. Balesdent, J., Mariotti, A,, and Boisgontier, D. (1990).Effect of tillage on soil organic carbon mineralization estimated from I3C abundance in maize fields. J. Soil Sci. 41,587-596. Balesdent, J., Mariotti, J. A., and Guillet, B. (1987).Natural 13C abundance as a tracer for studies of soil organic matter dynamics. Soil Biol. Biochem. 19,25-30. Balesdent, J., Wagner, G. H., and Mariotti, A. (1988).Soil organic matter turnover in long-term field experiments as revealed by carbon-13 natural abundance. Soil Sci. Soc. Am. J. 52, 118-124.
MODELING CARBON AND NITROGEN PROCESSES IN SOLS 291 Barak, P., Molina, J. A. E., Hadas, A., and Clapp, C. E. (1990a). Optimization of an ecological model with the Marquardt algorithm. Ecol. Model 51,251-263. Barak, P., Molina, J . A. E., Hadas, A., and Clapp, C. E. (l990b). Mineralization of amino acids and evidence of direct assimilation of organic nitrogen. Soil Sci. SOC.Am. J. 54,769-774. Bartholomew, W. V., and Kirkham, D. (1960). Mathematical description and interpretations of culture induced soil nitrogen changes. Trans. 7rh Incl. Cong. Soil Sci. 2,47 1-477. Beauchamp, E. G., Trevors, J. T., and Paul, J. W. (1989). Carbon sources for bacterial denitrification. Adv. Soil Sci. 10, 113-1 42. Beek, J., and Frissel, M. J. (1973). Simulation of nitrogen behaviour in soils. Pudoc, Wageningen, The Netherlands. Bender, M. M. (1968).Mass spectrometric studies of I3C variations in corn and other grasses. Radiocarbon 10,468-472. Betlach, R. M., and Tiedje, J. M. (1981). Kinetic explanation for accumulation of nitrite, nitric oxide, and nitrous oxide during bacterial denitrification. Appl. Environ. Microbiol. 42, 1074-1084. Bosatta E., and Agren, G. I. (1994). Theoretical analysis of microbial biomass dynamics in soils. Soil Biol. Biochem. 26,143-148. Bosatta, E., and Agren, G. I. (1995). Theoretical analyses of the interactions between inorganic nitrogen and soil organic matter. Eur: J. of Soil Sci. 76, 109-1 14. Boulaine, J. (1989). “Histoire des ptdologues et de la science des sols.” Institut, National de la Recherche Agronomique, Paris. Broadbent, F. E. (1965). Effect of fertilizer nitrogen on the release of soil nitrogen. Soil Sci. SOC.Am. Proc. 29,692-696. Broadbent, F. E. (1966). Interchange between inorganic and organic nitrogen in soils. Hilgardia 37, 165-1 80. Bryant, K. J., Benson, V. W., Kiniry, J. R., Williams, J. R., and Lacewell, R. D. (1992). Simulating corn yield response to irrigation timings: Validation of the EPIC model. J. Prod. Agric. 5(2), 237-242. Buyanosky, G. A., Kucera, C. L., and Wagner, G. H. (1987). Comparative analyses of carbon dynamics in native and cultivated ecosystems. Ecol. 68,2023-203 1. Cambardella, C. A,, and Elliott, E. T. (1992). Particulate soil organic-matter changes across a grassland cultivation sequence. Soil Sci. SOC. Am. J. 56,777-783. Campbell, C. A,, Paul, E. A., Rennie, D. A., and McCallum, K. J. (1967).Factors affecting the accuracy of the carbon-dating method in soil humus studies. Soil Sci. 104,81-85. Cem, C., Feller, C., Balesdent, J., Victoria, R., and Plenecassagne, A. (1985). Application du tracage isotopique nature1 in I3C B I’Ctude de la dynamique de la m a t h organique dans les sols. C. R. Acad. Sci. Ser. II, 423-428. Cheng, H. H., and Molina, J. A. E. (1995).In search of bioreactive soil organic carbon: The fractionation approaches. In “Soils and Global Changes” (R. Lal, J. Kimble, E. Levine, and B. A. Stewart, eds.), pp. 343-350. Lewis Pub., Boca Raton, FX. Chertov, 0. G., and Komarov, A. S. (1996). SOMM-a model of soil organic matter and nitrogen dynamics in terrestrial ecosystems. In “Evaluation of Soil Organic Matter Models Using Existing, Long-Term Datasets” (D. S. Powlson, P. Smith, and J. U. Smith, eds.), pp. 231-236. NATO AS1 Series, I, Vol. 38. Springer-Verlag, Berlin. Christensen, B. T. (1996).Matching measurable soil organic matter fractions with conceptual pools in simulation models of carbon turnover. In “Evaluation of Soil Organic Matter Models Using Existing, Long-Term Datanets” (D. S. Powlson, P. Smith, and J. U.Smith, eds.), pp. 143-159. NATO AS1 Series, I, Vol. 38. Springer-Verlag. Berlin. Coleman, K., and Jenkinson, D. S. (1996). RothC-26.3. A model for the turnover of carbon in soil. In “Evaluation of Soil Organic Matter Models Using Existing, Long-Term Datasets” (D.S. Powlson, P. Smith, and J. U. Smith, Eds.), pp. 237-246. NATOASI Series, I, Vol. 38. Springer-Verlag. Berlin. De Ruiter, P. C., Neutel, A. M., and Moore, J. C. ( I 994). Modelling food webs and nutrient cycling in agro-ecosystems. Trends Ecol. and Evol. 9,378-383.
292
JEAN-ALEX E. MOLINA AND PETE SMITH
de Willigen, P. (1991). Nitrogen turnover in the soil-crop system; comparison of fourteen simulation models. Fert. Res. 27, 141-149. de Willigen, P., and Neeteson, J. J. (1985). Comparison of six simulation models for the nitrogen cycle in the soil. Fert. Res. 8, 157-171. Donigan, A. S., Jr., Barnwell, T. 0.. Jr., Jackson, R. B., IV,Patwardhan, A. S., Weinrich, K. B., Rowell, A. L., Chinnaswamy, R. V., and Cole, C. V. (1994). “Assessment of Alternative Management Practices and Policies Affecting Soil Carbon in Agroecosystems of the Central United States.” U.S. EPA Report EPA/600/R-94/067, Athens. Doran, J. W., Mielke, L. N., and Stamatiadis, S. (1988). Microbial activity and N cycling as regulated by soil water-filled pore space. Proc. 1lrhCon$ lntl. Soil Tillage Res. Org. 1,49-54. Douglas, C . L. Jr., Rickman, R. W., and Waldman, S. E. (1995). “ D 3 R A residue management aid. In “Columbia Basin Agricultural Research Annual Report,” pp. 96-98. Spec. Rpt. 946. Agric. Exp. Sta. Oregon St. Univ. and USDA-ARS. Elliott, E. T., Paustian, K., and Frey, D. D. (1996). Modeling the measurable or measuring the modelable: A hierarchical approach to isolating meaningful soil organic matter fractionations. In “Evaluation of Soil Organic Matter Models Using Existing, Long-Term Datasets” (D. S. Powlson, P. Smith, and J. U. Smith, eds.), pp. 161-179. NATO AS1 Series, I, Vol. 38. Springer-Verlag. Berlin. Esser, G. (1990). Modelling global terrestrial sources and sinks of CO, with special reference to soil organic matter. I n ‘‘Soils and the Greenhouse Effect” (A. F. Bouwman, ed.), pp. 247-261. Wiley and Sons, New York. Franko, U., Oelschlagel, B., and Scheck, S. (1995). Simulation of temperature-, water-, and nitrogen dynamics using the model CANDY. Ecol. Model. 81,213-222. Glendining, M. J., and Poulton, P. R. (1996). Interpretation difficulties with long term experiments. In “Evaluation of Soil Organic Matter Models Using Existing, Long-Term Datasets” (D. S. Powlson, P. Smith, and J. U. Smith, eds.), pp. 99-109. NATO AS1 I, Vol. 38. Springer-Verlag. Berlin. Goldewijk, K. K., van Minnen, J. G., Kreileman, G. J. J., Vloedbeld, M., and Leemans, R. (1994). Simulating the carbon flux between the terrestrial environment and the atmosphere. Water, Air and Soil Pollution 76, 199-230. Gorham, E. (1991). Biogeochemistry: Its origin and development. Biogeochem. 13, 199-239. Goto, N., Sakoda, A., and Suzuki, M. (1993). Modelling soil carbon dynamics as a part of the carbon cycle in terrestrial ecosystems. Ecol. Model. 74, 183-204. Grant, R. F. (1991). A technique for estimating denitrification rates at different soil temperatures, water contents and nitrate concentrations. Soil Sci. 152,41-52. Grant, R. F. (1994). Simulation of ecological controls on nitrification. SoilBiol. Biochem. 26,305415. Grant, R. F. (1995). Dynamics of energy, water, carbon and nitrogen in agricultural ecosystems: Simulation and experimental validation. Ecol. Model. 81, 169-181. Gregory, P. J., and Ingram, J. S. I. (1996). The contribution of the soil organic matter network (SOMNET) to GCTE. In “Evaluation of Soil Organic Matter Models Using Existing, Long-Term Datasets” (D. S. Powlson, P. Smith, and J. U. Smith, eds.), pp. 13-26. NATO ASI, I, Vol. 38, Springer-Verlag, Berlin. Hadas, A., Molina, J. A. E., Feigenbaum, S., and Clapp, C. E. (1987). Simulation of nitrogen-I5 immobilization by the model NCSOIL. Soil Sci. SOC.Am. J. 51, 102-106. Hadas, A., Sofer, M., Molina, J. A. E., Barak, P., and Clapp, C. E. (1992). Assimilation of nitrogen by the soil microbial population: NH; versus organic N. Soil Biol. Biochem. 24, 137-143. Hagen. N., Kleeberg, H. -B., and Niekamp, 0. (1993). Parameter estimation for ecosystem models with special regard to target functions. Model. Geo-Biosphere Proc. 2,29-325. Harkness, D. D., Hamson, A. F., and Bacon, P. J. (1991). The potential of bomb I4C measurements for estimating soil organic matter turnover. In “Advances in Soil Organic Research: The Impact on Agriculture and the Environment” (W. S. Wilson, ed.), pp. 239-251. R. SOC.Chem., Cambridge. Henin, S., and Dupuis, M. (1945). Essai de bilan de la matikre organique du sol. Ann. Agmn. 15,17-29.
MODELING CARBON AND NITROGEN PROCESSES IN SOILS 293 Hsieh, Y. P. (1992). Pool size and mean age of stable soil organic carbon in cropland. Soil Sci. SOC.Am. J. 56,461464. Hsieh, Y.P. (1993). Radiocarbon signature of turnover rates in active soil organic carbon pools. Soil Sci. SOC.Am. J. 57, 1020-1022. Hsieh, Y. P. (1996). Soil organic carbon pools of two tropical soils inferred by carbon signatures. Soil Sci. SOC.Am. J. 60,1117-1121. Huggins, D. R., Clapp, C. E., Allmaras, R. R., Lamb, J. A,, and Layese, M. F. (1997). Dynamics of soil organic carbon in corn-soybean sequences as estimated from natural 13Cabundance. Soil Sci. SOC. J. (in press). Hunt, H. W. (1978).A simulation model for the decomposition in grasslands. In “Grassland Simulation Model” (G. S. Hinis, ed.), pp. 155-183. Ecological Studies 26, Springer-Verlag. Berlin. Hunt, H. W., Coleman, D. C., Ingham, E. R., Ingham, R. E., Elliott, E. T., Moore, J. C., Rose, S. L., Reid, C. P. P., and Morley, C. R. (1987).The detrital food web in a shortgrass prairie. Biol. Fertif. Soils 3,57-68. Izarraulde, R. C., McGill, W. B., Jans-Hammermeister, D. C., Hanger-Kozyra, K. L., Grant, R. F., and Hiley, J. C. (1996). Development of a technique to calculate carbon fluxes in agricultural soils at the ecodistrict level using simulation models and various aggregation methods. Interim Report No. 01586-3-1280/01-XSG.Univ. of Alberta, Edmonton, Canada. Janson, S. L. (1958). Tracer studies on nitrogen transformations in soil with special attention to mineralization-immobilization relationships. Ann. Roy. Agl: Coll. Sweden 24, 101-361. Jarvis, S. C., Stockdale, E. A,, Shepherd, M. A,, and Powlson, D. S. (1996). Nitrogen mineralization in temperate agricultural soils: Processes and measurement. Adv. Agron. 57, 1 87-235. Jenkinson, D. S. (1990). The turnover of organic carbon and nitrogen in soil. Phil. Trans. of the Royal SOC.,London B 329,361-368. Jenkinson, D. S., Adams, D. E., and Wild, A. (1991). Model estimates of CO, emissions from soil in response to global warming. Narure 351,304-306. Jenkinson, D. S., Hart, P. B. S., Rayner, J. H., and Pany, L. C. (1987). Modelling the turnover of organic matter in long-term experiments at Rothamsted. INTECOL Butletin 15, 1-8. Jenny, H. (1941). “Factors of Soil Formation. A System of Quantitative Pedology.” McGraw-Hill, New York. Jensen, C., Stougaard, B., and Ostergaard, H. S. (1994). Simulation of nitrogen dynamics in farmland areas of Denmark (1989-1993). Soil Use and Mgt. 10, 111-1 18. Juma, N. G., and Paul, E. A. (1984).Mineralizable soil nitrogen: Amounts and extractibility ratios. Soil Sci. SOC.Am. J. 48,7640. Kalembasa, S. J., and Jenkinson, D. S. (1973). Acomparative study of titrimetric and gravimetric methods for the determination of organic carbon in soil. J. Sci. Fd. Agric. 24, 1085-1090. Kan, N. A,, and Kan, E. E. (1991).Simulation model of soil fertility. Physiol. Biochem. of Culr. Planrs 23(1), 2-16. (In Russian). Kelley, R., K., and Stevenson, F. (1985). Characterization and extractability of immobilized I5N from the soil microbial biomass. Soil Bid. Biochem. 17,517-523. Kirkham, D., and Bartholomew, W. V. (1955).Equations for following nutrient transformations in soil, utilizing tracer data: II. Soil Sci. SOC. Proc. 18, 189-192. Krupenikov, I. A. (1993). History of soil science from its inception to the present. Russian Translation Series 98. A. A. Balkema, Brookfield, VT. Laudelout, H., Lambert, R., Fripiat, J. L., and Pham, M. L. (1974). Effet de la temperature sur la vitesse doxydation de l’ammonium en nitrate par des cultures mixtes de nitrifiants. Ann. Microbiol. (Inst. Pasteur) 125B, 75-84. Laudelout, H., Lambert, R., and Pham, M. L. (1976). Influence de pH et de la pression partielle d‘oxygkne sur la nitrification. Ann. Microbiol. (Inst. Pasteur) 127,367-382. Leffelaar, P. A,, and Wessel, W. W. (1988). Denitrification in a homogeneous closed system: Experiment and simulation. Soil Sci. 146, 335-349.
2 94
JEAN-ALEX E. MOLINA AND PETE SMITH
Li, C., Frolking, S., and Harriss, R. (1994a). Modeling carbon biogeochemistry in agricultural soils. Glob. Biogeochem. Cycles 8,237-254. Li, C., Frolking, S., and Harriss, R. (1994b). Modeling nitrous oxide emissions from agriculture: A Florida case study. Chemosphere 28, 1401-1415. Loague, K.. and Green, R. E. (1991). Statistical and graphical methods for evaluating solute transport models: Overview and application. J. Conraminant. Hydrol. 7,5 1-73. Martel, Y.A., and Lasalle, P. (1977). Radiocarbon dating of organic matter from a cultivated topsoil in Eastern Canada. Can. J. Soil Sci. 57, 375-377. Martel, Y. A., and Paul, E. A. (1974). The use of radiocarbon dating of organic matter in the study of soil genesis. Soil Sci. SOC.Arne,: Proc. 38,501-506. Martin, J. P., and Haider, K. (1986). Influence of mineral colloids on the turnover rates of soil organic carbon. In “Interaction of Soil Minerals with Natural Organics and Microbes” (P. M. Huang and M. Schnitzer, eds.), pp. 283-304. Soil Sci. SOC.Am. Spec. Publication No. 17, Madison, WI. Martin, A., Mariotti, A,, Balesdent, J., Lavelle P., and Vuattoux, R. (1990). Estimate of organic matter turnover rate in a savanna soil by I3C natural abundance measurements. Soil Biol. Biochem. 22, 517-523. Martin, A., Mariotti, A., Balesdent, J., and Lavelle, P.(1992). Soil organic matter assimilation by a geophagous tropical earthworm based on V3C measurements. Ecol. 73,118-128. McCaskill, M., and Blair, G. J. (1990). A model of S, P, and N uptake by a perennial pasture. II. Calibration and prediction. Ferf. Res. 22, 173-179. McClaugherty, C. A., and Linkins, A. E. (1990). Temperature responses of enzymes in two forest soils. Soil Biol. Biochem. 22,29-33. McGill, W. B. (1996).Review and classification of ten soil organic matter (SOM) models. In “Evaluation of Soil Organic Matter Models Using Existing, Long-Term Datasets” (D. S. Powlson, P. Smith, and J. U. Smith, eds.), pp. 111-132. NATO AS1 Series, I, Vol. 38. Springer-Verlag. Berlin. McGill, W. B., Hunt, H. W., Woodmansee, R. G., and Reuss, J. 0. (1981).PHOENIX, a model of the dynamics of carbon and nitrogen in grassland soil. In “Terrestrial Nitrogen Cycles. Processes, Ecosystem Strategies and Management Impacts” (F. E. Clark and T. Roswall, eds.), pp. 49-1 15. Ecol. Bull. No. 33, Stockholm. Melillo, J. M., Kicklighter, D. W., McGuire, A. D., Peterjon, W. T., and Newkirk, K. M. (1995). Global change and its effect on soil organic carbon stocks. In “Role of Nonliving OrganicMatter in the Earth’s Carbon Cycle” (R. G. Zepp and C. H. Sonntag, eds.), pp. 175-189. Wiley and Sons, New York. Molina, J. A. E. (1985). Components of rates of ammonium oxidation in soil. Soil Sci. SOC.Am. J. 49, 603409. Molina, J. A. E. (1996). Description of the model NCSOIL. In “Evaluation of Soil Organic Matter Models Using Existing, Long-Term Datasets” (D. S. Powlson, P. Smith, and J. U. Smith, eds.), pp. 269-274. NATO AS1 Series, I, Vol. 38. Springer-Verlag. Berlin. Molina, J. A. E., Clapp, C. E., Shaffer, M. J., Chichester, F. W., and Larson, W. E. (1983). NCSOIL, a model of nitrogen and carbon transformations in soil: Description, calibration, and behavior. Soil Sci. SOC. Am. J. 47,85-91. Molina, J. A. E., Hadas, A., and Clapp, C. E. (1990). Computer simulation of nitrogen turnover in soil and priming effect. Soil Biol. Biochem. 22,349-353. Molina, J. A. E., Cheng, H. H., Nicolardot, B., Cbaussod, R., and Houot, S. (1994). Biologically active soil organics: A case of double identity. In “Defining Soil Quality for a Sustainable Environment” (J. W. Dorm, J. W. Coleman, D. C. Bezdicek, and B. A. Stewart, ed.), pp. 169-177. Soil Sci. SOC.Am. J. Special Publication No. 35. Madison, WI. Moorhead, D. L., and Reynolds, J. F. (1991). A general model of litter decomposition in the northern Chichuahuan Desert. Ecol. Model. 56, 191-218. Nadelhoffer, K. J., Giblin, A. E., Shaver, G. R., and Laundre, J. A. (1991). Effects of temperature and substrate quality on element mineralization in six arctic soils. Ecol. 72,242-253.
MODELING CARBON AND NITROGEN PROCESSES IN SOILS 295 Nelson, D. W., and Sommers, L. E. (1982). Total carbon, organic carbon and organic matter. In “Methods of Soil Analysis. Part 2” (A. L. Page el al.. eds.), pp. 539-579. ASNSSSA, Madison, WI. Nicolardot, B., Fauvet, G., and Cheneby, D. (1994a). Carbon and nitrogen cycling through soil microbial biomass at various temperatures. Soil Biol. Biochem. 26,253-26 1. Nicolardot, B., Molina, J. A. E., and Allard, M. R. (1994b). C and N fluxes between pools of soil organic matter: Model calibration with long-term incubation data. Soil Biol. Biochem. 26, 235-243. Nikiforoff, C. C. (1936). Some general aspects of the chernozem formation. Soil Sci. SOC.Am. Proc. 1,333-342. Nyhan, J. W. (1976). Influence of soil temperature and water tension on the decomposition rate of carbon-I4 labeled herbage. Soil Sci. 121,442-449. Oades, J. M. (1988). The retention of organic matter in soils. Biogeochem. 5,35-70. Oades, J. M. (1995). Klasmozems--organic matter. Aust. J. Soil Res. 33,43-57. O’Brien, B. J. (1984). Soil organic carbon fluxes and turnover rates estimated from radiocarbon enrichment. Soil Biol. Biochem. 16, 115-120. O‘Brien, B. J., and Stout, J. D. (1977). Movement and turnover of soil organic matter as indicated by carbon isotope measurements. Soil Biol. Biochem. 10,309-3 17. O’Leary, G. J. (1994). Soil water and nitrogen dynamics of dryland wheat in the Victorian Wimmera and Malle. PhD. Thesis, Univ. of Melbourne. Olson, J. S. (1963). Energy storage and the balance of producers and decomposers in ecological systems. Ecol. 44,322-331. Parsch, L. D., Cochran, M. J., Trice, K. L., and Scott, H. D. (1991). Biophysical simuation of wheat and soybean to assess the impact of timeliness on double-cropping economics. In “Modeling Plant and Soil Systems” (J. Hanks and J. T. Ritchie. eds.), pp. 51 1-534. Agronomy No. 31., Am. SOC. of Agron., Madison, WI. Parshotam, A., Tate, K. R., and Giltrap, D. J. (1995). Potential effects of climate and land-use change on soil carbon and CO, emissions from New Zealand’s indigenous forests and unimproved grasslands. Weather and Climate 15,3-12. Parton, W. J., and Rasmussen, P. E. (1994). Long-term effects of residue management in wheatlfallow: Century model simulations. Soil Sci. SOC.Am. J. 58,53&536. Parton, W. J., Schimel, D. S., Cole, C. V., and Ojima, D. (1987). Analysis of factors controlling soil organic levels of grasslands in the Great Plains. Soil Sci. Sac. Am. J. 51, 1173-1 179. Parton, W. J., Scurlock, J. M. 0..Ojima, D. S., Gilrnanov, T. G., Scholes, R. J., Schimel, D. S., Kirchner, T., Menaut, J. -C., Seastedt, T., Garcia Moya, E., Apinan, K., and Kinyamario, J. I. (1993). Observations and modeling of biomass and soil organic matter dynamics for the grassland biome worldwide. Global Biogeochem. Cycles 7,785-809. Paul, E. A,, and Van Veen, J. A. (1978). The use of tracers to determine the dynamic nature of organic matter. Trans. ilth Inteni. Congress Soil Sci. 3,61-102. Paustian, K. (1994). Modelling soil biology and biochemical processes for sustainable agricultural research. In “Soil Biota. Management in Sustainable Farming Systems” (C. E. Panburst, B. M. Doube, V. V. S. R. Gupta, and P. R. Grace, eds.), pp. 182-193. CSIRO Info. Set, Melbourne. Plentinger, M. C., and Penning de Vries F. W. T. (eds.), (1996). CAMASE, Register of agro-ecosysterns models. Ver. II. DLO-Research Institut for Agrobiology and Soil Fertility (AB-DLO), Wageningen, The Netherlands. Post, W. M.. King, A. W., and Wullschleger, S. D. (1996). Soil organic matter models and global estimates of soil organic carbon. In “Evaluation of Soil Organic Matter Models Using Existing, LongTerm Datasets” (D. S. Powlson, P. Smith, and J. U. Smith, Eds.), pp. 201-222. NATO AS1 I, Vol. 38. Springer-Verlag. Berlin. Post, W. M., Emanuel, W. R., Zinke, P. J., and Staggenberger, A. G. (1982). Soil carbon pools and world life zones. Narure 298, 156-159.
296
JEAN-ALEX E. MOLINA AND PETE SMITH
Post, W. M., Pastor, J., Zinke, P. J., and Staggenberger, A. G. (1985). Global patterns of soil nitrogen storage. Nature 317,613-616. Potter, C. S.. Randerson, J. T., Field, C. B., Matson, P. A,, Vitousek, P. M., Mooney, H. A., and Klooster, S. A. (1993). Terrestrial ecosystem production: A process model based on satellite and surface data. Global Biogeochem. Cycles 7,811-841. Powlson, D. S., Smith, P., and Smith, J. U. (1996). “Evaluation of Soil Organic Matter Models Using Existing, Long-Term Datasets.” NATO ASI, I, Vol. 38, Springer-Verlag, Berlin. Powlson, D. S., Smith, P., Coleman, K., Smith, J. U., Glendining, M. J., Korschens, M., and Franko, U. (1997). A European network of long-term sites for studies on soil organic matter. Soil and Ellage Res. (in press). Rehm, G., Schmitt, M., and Munter, R. (1994). Fertilizing corn in Minnesota. Publication FO-3790-C, Minnesota Extension Service, Univ. of Minnesota, Saint Paul. Ritjema, P. E., Groenendijk, P.,Kroes, J. G., and Roest, C. W. J. (1995). Formulation of the nitrogen and phosphorus behaviour in agricultural soils, the ANIMO model. Report 30. Winand Staring Centre, Wageningen, The Netherlands. Robert, P. C., Thompson, W. H., and Fairchild, D. (1991). Soil specific anhydrous ammonia management system. In “Proceedings of the Symposium for Automated Agriculture for the 21th Century,” pp. 418426. Am. Soc. of Agric. Eng. Russell, J. E. (1973). “Soil Conditions and Plant Growth.” Longman, London. Ryzhova, I. M. (1993). Analysis of sensitivity of soil-vegetation systems to variations in carbon turnover parameters based on a mathematical model. Euras. Soil Sci. 25,43-50. Sabey, B. R., Frederick, L. R., and Bartholomew, W. V. (1969). The formation of nitrate from ammonium nitrogen in soils. IV. Use of the delay and maximum rate phases for making quantitative predictions. Soil SOC.Am. Proc. 33,276278, Scharpenseel, H. W. (1971). Radiocarbon dating of soils; problems, troubles and hopes. In “Paleopedology” (D. H. Yalon, ed.), pp. 77-88. Israel University Press. Schiff, S. L., Trumbore, S. E., and Dillon, P. J. (1990). Dissolved organic carbon cycling in forested watersheds: A carbon isotope approach. Wafer Resour: Res. 26,2949-2957. Schimel, D. S., Braswell, B. H., Jr., Holland, E. A., McKeown, R. Ojima, D. S., Painter, T. H., Parton, W. J., andTownsend, J. R. (1994). Climatic, edaphic, and biotic controls over storage and turnover of carbon in soils. Global Biogeochem. Cycles 8,279-293. Schlegel, A. J., and Havlin, J. L. (1995). Corn response to long-term nitrogen and phosphorus fertilization. J. Prod. Agn’c. 8, 181-185. Schwartz, D., Mariotti, A., Lanfranchi, R., and Guillet, B. (1986). I3C/l2C ratios of soil organic matter as indicators of vegetation changes in the Congo. Geodema 39,97-103. Shevtsova, L. K., and Volodarskaya, I. V. (1991). Long-term fertilization influence on humus balance and its quality parameters in chemicalization in agriculture. All Russian Inst. for Fert. and Agric. Soil Sci. (unpublished manuscript, in Russian). Sierra, J., and Renault, P.(1996). Respiratory activity and oxygen distribution in natural aggregates in relation to anaerobiosis. Soil Sci. SOC.Am. J. 60, 1428-1438. Skoop, J., Jawson, M. D., and Doran, J. W. (1990). Steady-state aerobic microbial activity as a function of soil water content. Soil Sci. SOC.Am. J. 54,1619-1625. Smith, J. U., Bradbury, M. J., and Addiscott, T. M. (1995). SUNDIAL: Simulationof nitrogen dynamics in arable land. A use-friendly, PC-based version of the Rothamsted nitrogen turnover model. Agron. J. 88,3843. Smith, J. U., Smith P.,and Addiscott, T. M. (1996). Quantitative methods to evaluate and compare soil organic matter (SOM) models. In “Evaluation of Soil Organic Matter Models Using Existing, Long-Term Datasets” (D. S. Powlson, P. Smith, and J. U. Smith, eds.), pp. 181-200. NATO ASI, I, Vol. 38. Springer-Verlag. Berlin. Smith, P., Powlson, D. S., and Glendining, M. J. (1996a). Establishing a European GCTE Soil Organic Matter Network (SOMNET). In “Evaluation of Soil Organic Matter Models Using Existing,
MODELING CARBON AND NITROGEN PROCESSES IN SOILS 297 Long-Term Datasets.” (D. S. Powlson, P. Smith, and J. U. Smith, eds.), pp. 81-97. NATO ASI, I, Vol. 38. Springer-Verlag, Berlin. Smith, P., Powlson, D. S., Smith, J. U., and Glendining, M. J. (1996b). The GCTE SOMNET. A global network and database of soil organic matter models and long-term datasets. Soil Use and Mgt. 12, 104. Smith, P., Smith, J. U., and Powlson, D. S. (eds.) (1996~).“Soil Organic Matter Network (SOMNET): 1996 Model and Experimental Metadata.” GCTE Report No. 7. GCTE Focus 3 Office, Wallingford, UK. Smith, P.,Andrkn, 0..Brussaard, L., Dangerfield, M., Ekschmitt, K., Lavelle, P., andTate, K. (1997a). Soil biota and global change at the ecosystem level: Describing soil biota in mathematical models. Global Change Biology (in press). Smith, P., Powlson, D. S., Smith, J. U., and Elliott, E. T. (eds.). (1997b). “Evaluation and Comparison of Soil Organic Matter Models Using Datasets from Seven Long-Term Experiments.” Geodema, Special Issue (in press). Smith, P., Smith, J. U., Powlson, D. S., McGill, W. B., Arah, J. R. M., Chertov, 0. G., Coleman, K., Franko, U., Frolking, S., Jenkinson, D. S., Jensen, L. S., Kelly, R. H., Klein-Gunnewiek, H., Komarov, A., Li, C., Molina, J. A. E., Mueller, T., Parton, W. J., Thornley, J. H. M., and Whitmore, A. P. (1997~).A comparison of the performance of nine soil organic matter models using seven long-term experimental datasets. In “Evaluation and Comparison of Soil Organic Matter Models Using Datasets from Seven Long-Term Experiments” (P. Smith, D. S. Powlson, J. U. Smith, and E. T. Elliott, eds.). Geode-, Spec. Issue (in press). S@rensen,L. H. (1975). The influence of clay on the rate of decay of amino acid metabolites synthesized in soils during decomposition of cellulose. Soil Biol. Biochem. 7 , 171-177. SBrensen, L. H. (1987). Organic matter and microbial biomass in a soil incubated in the field for 20 years with ‘‘C-labelled barley straw. Soil Biol. Biochem. 1 9 , 3 9 4 2 . Stanford, G., and Smith, S. J. (1972). Nitrogen mineralization potentials of soils. Soil Sci. Soc. Am. Proc. 36,465-472. Stanford, G., Frere, M. H., and Schwaninger, D. H. (1973). Temperature coefficient of soil nitrogen mineralization. Soil Sci. 115,321-323. Thomley, J. H. M., and Cannell, M. G. R. (1994). Predictions of the effects of climate and management change on temperate grassland. J. ofAgric. Sci. 123, 151-152. Tiessen, H., and Moir, J. 0. (1 993). Total and organic carbon. In “Soil Sampling and Methods of Analysis” (M. R. Carter, ed.), pp. 187-199. Canadian Society of Soil Science, Lewis Pub., Boca Raton, FL. Van Veen, J. A., Ladd, J. N., and Amato, M. (1985). Turnover of carbon and nitrogen through the microbial biomass in a sandy loam and a clay soil incubated with [I4C(U)]glucose and [15N] (NH,) ,SO, under different moisture regimes. Soil Biol. Biochem. 17,747-756. Van Veen, J. A,, McGill, W. B., Hunt, H. W., Frissel, M. J., and Cole, C. V. (1981). Simulation models of the terrestrial nitrogen cycles. In “Terrestrial Nitrogen Cycles. Processes, Ecosystem Strategies and Management Impacts.” (F. E. Clark and T. Rosswell, eds.), pp. 2 5 4 8 . Ecol. Bull., Vol. 33, Stockholm. Vanclooster, M., Vereecken, H., Diels, J. Hyusmans, F., Verstraete, W., and Feyen, J. (1992). Effect of mobile and immobile water in predicting nitrogen leaching from cropped soils. Model. Geo-Biosphere Proc. 1 , 2 3 4 0 . Verberne, E. L. M., Hassink, J., de Willigen, P., Groot, J. J. R., and van Veen, J. A. (1990). Modelling soil organic matter dynamics in different soils. Netherlands J. Agric. Sci. 38,221-238. Volkoff, B., and Cerri, C. C. (1987). Carbon isotopic fractionation in subtropical Brazilian grassland soils. Comparison with tropical forest soils. Plant and Soil 102,27-3 1. Wagner, G. H. (1991). Using the natural abundance of ”C and I5N to examine soil organic matter accumulated during 100 years of cropping. Int. Atomic Energy Agency SM-313/8, pp. 261-268. Waksman, S. A., and Starkey, R. L. (1931). The soil and the microbe. Wiley, New York.
2 98
JEAN-ALEX E. MOLINA AND PETE SMITH
Winogradsky, S . (1949). Microbiologie du sol, problbmes et mbthodes. Masson, Paris. Yakimenko, E. (1996). Soil evolution under dry meadows in a boreal climate: The Moscow dry-meadow experiment. In “Evaluation of Soil Organic Matter Models Using Existing, Long-Term Datasets” @. S. Powlson, P. Smith, and J. U. Smith, eds.), pp. 419-422. NATO ASI, I, Vol. 38. Springer-Verlag, Berlin.
Index A Abiotic stress, genotype-by-environment interactions and, 217-222 Ahscisic acid activity, 50 in Azospirillum, 72 metabolism in soil, 110 microbial synthesis, 64-68 mycorrhizae and, 97 in pathogenesis, 103-104 root nodules and, 88-89 ACC, see 1-Aminocyclopropane-1 -carboxylic acid ACC deaminase, in plant-microbe interactions, 114116 Acclimation, environmental stress and, 2 I5 Adenine in cytokinin synthesis, 58 exogenous application, 120 in precursor-inoculum interactions, 118 soil cytokinin levels and, 108 Agricultural production, world population and, 208 Agrobacterium rhizogenes, auxin and, 101 Agrobacterium tumefaciens, plant growth regulators and, 98-99 Agronomic concept, of genotype, 225 Agronomy, soil-crop models and, 286-287 Alfalfa, cytokinin levels in, mycorrhizae and, 95 Allelopathy, genotype-by-environment interactions and, 217 I-Aminocyclopropane-1 -carboxylic acid in ethylene synthesis, 50 in root nodules. 87-88 soil ethylene levels and, 1 I0 Ammonium, in soil nitrogen modeling, 27027 1 Arabidopsis cold tolerance in, 215 heterosis in, 222 Artifacts, in atomic force microscopy, 9-14
Atomic force microscopy artifacts in, 9-14 atomic forces in, 6-7 atomic-scale surface studies, 15-17 mineral growth and dissolution, in situ studies, 19-27 mineral-water interface studies, 17-18 new techniques in, 37-39 operating principles, 3-6 soil particle studies, 2-3,27-37 tapping-mode, 3,5,7-8,21-22.30-32, 32-34 tip-sample interactions, 6 1 4 Atomic forces, in atomic force microscopy, 6 7 Auger electron spectroscopy, in mineral surface studies, 2,20 Auxin in Agrobacterium rhizogenes, 101 in Azospirillum, 69-72 in Azotobacter; 69 in Bradyrhizobium, 76.79 crown gall tumor and, 98-99 forms, 48 metabolism in soil, 104-107 microbial catabolism, 56 microbial synthesis, 5 1-56 mycorrhizae and, 89.93 in plant-microbe interactions, 111-1 13 precursor-inoculum interactions and, 1 16117 precursors, exogenous application, 119120 in Pseudomonas syringae pv. savastanoi, 100-101 in Rhizobium, 76.79 root nodules and, 81-83 Azospirillum auxin in, 69-72 gibberellin in, 113 Azotobacter auxin in, 69 cytokinin synthesis in, 58-59
299
3 00
INDEX
in precursor-inoculum interactions, 117118
B Bacillus, auxin in, 111 Bacteria, see also Microorganisms; Plant growth-promoting rhizobacteria; Plant microbe interactions atomic force microscopy and, 38 nitrogen-fixing, auxin in, 112 Bakanae disease, 102 Barley, gene loss and, 2 10 Biotic stress, genotype-by-environment interactions and, 215-217 Bonemeal, in long-term field experiments, 158, 169 Botrytis, abscisic acid in, 103-104 Bradyrhizobium auxin and, 76.79-80 cytokinin and, 80 Breeders, evaluation costs and, 212 Breeding, see Crop breeding Breeding stations, need for, 209 Brucite, dissolution studies, 26
C C3 plants, in SOM dynamics, 28 1,284-285 C4 plants, in SOM dynamics, 281, 284-285 Calcite, growth and dissolution studies, 26 CANDY model, 272-273 Capillary adhesion, in atomic force microscopy, 7 Carbon, see Soil carbon; Soil organic matter I4Carbon, in evaluation of SOM modeling, 278-285 13C:'2C ratio, in soil organic matter, 281,284 CENTURY model, 262,265,267,274.287 Cercospora, abscisic acid synthesis in, 65-68 Chemical fertilizer industry, origin of, 255 Chemical sensing, atomic force microscopy and, 37-38 Chromium(II1) hydroxide, growth and dissolution studies, 26-27 CIMMYT breeding program, 224 Citrus, cytokinin levels, mycorrhizae and, 95-96 Clay, relationship with soil organic matter, 273-274 Climate change, global, soil transformations modeling and, 288-289
C:N ratio in long-term field experiments, 181-1 87 in soil nitrogen modeling, 270 Cold tolerance, in Arabidopsis, 215 Colloidal forces, atomic force microscopy and, 18 Competition, genotype-by-environment interactions and, 2 16 Convolution, in atomic force microscopy, 912 cor gene, in Arabidopsis, 215 Corn environmental stress and, 214 genotype-by-environment interactions in, 230-23 1 in Morrow Plots, history of, 154, 156-161 SOM dynamics and, 285 stability in, 238 Corn hybrids, in long-term field experiments, 158-159, 174-175, 192 Corn-oat-hay rotation, in long-term field experiments, 154, 156-158, 161, 186-187 Corn-oat rotation, in long-term field experiments, 154, 156-158, 186 Corn-soybean rotation, in long-term field experiments, 161, 171 Corn yield, in Morrow Plots compared to Illinois state trends, 177-180 in high-level NPK treatment, 171-172 hybrids and, 174-1 75 in manure-lime-phosphate treatment, 68169 in nitrogen-lime-phosphate-potassiumtreatment, 169-171 in original experiment, 167 overall experimental effects, 163-166 planting date and, 175-176 soil organic matter and, 189-191 weather effects, 176-1 77 yield stability, 172-174 Corynebacteriumfascians, cytokinin in, 101-102 Cowpea, root nodules in, 84-85 Crop breeding defined, 200 early multienvironment testing in, 239 efficient program design, 239-240 environment characterization and, 224 evaluation of intermediate growth stages in, 239 genetic diversity and, 240
301
INDEX genotype-by-environment interactions and, 201,204,208-212,223-225, 240-24 1 molecular techniques in, 235-236 for performance stability, 237-239 stability statistics and, 225-234 for stress resistance, 234-236 unbalanced data sets in, 231-234 Crop environment, characterization for breeding, 224 Crop rotation, in long-term field experiments, 154-158, 161, 191 effects on soil organic matter, 181, 186-187 effects on yield, 165-166, 170-171 Crossover interaction, in genotype-by-environment interactions, 207 Crown gall tumor, plant growth regulators and, 98-99 Cryptococcus, ethylene synthesis in, 63 Cultivar development, see also Crop breeding early multienvironment testing in, 239 efficient program design and, 239-240 environmental clustering and, 224 evaluation of intermediate growth stages in, 239 genotype-by-environment interactions and, 209-2 12 performance stability and, 237-239 stability statistics and, 225-234 Cultivation, effects on SOM, 279,285 Cytokinin in Corynebacteriumfascians, 101 102 crown gall tumor and, 98-99 exogenous, physiological effects, 20-121 forms, 48 metabolism in soil, 107-108 microbial synthesis, 58-59 mycorrhizae and, 94-96 in plant-microbe interactions, 113-1 14 precursor-inoculum interactions and, 1 18 in Pseudomonas syringae pv. savastanoi, 100 in rhizobacteria, 72, 79-80 root nodules and, 85-86
D DAISY model, 265,274 Decomposition rates, effects of nitrogen on, 274-275 Deconvolution, in atomic force microscopy, 12, 14
Denitrification, in soil nitrogen modeling, 27 1-272 Diazotrophs, free-living, plant growth regulators in, 69-75 Dimethylallylpyrophosphate,in cytokinin synthesis, 59 Disease resistance, genotype-by-environment interactions and, 216 DNA, effects of environmental stress on, 2 13-2 14 DNDC model, 272 Double-layer forces, atomic force microscopy and, 17-18 Drought, plant cytokinin levels, mycorrhizae and, 95 Dynamic concept, of genotype, 225
E Earthworms, effects on soil transformations, 28 1 Electron diffraction, in mineral surface analysis, 1-2.15, 17 Electron microscopy, in mineral analysis, 15, 17, 19,32 Enrerobacferiaceae, auxin production in, 111 Environment, see also Genotype-by-environment interaction characterization, in crop breeding, 224 effects on cultivar stability, 230-231 influence on heredity, 202-203 phenotypic expression and, 201-202 yield improvements and, 209 Environmental scanning electron microscopy, in mineral analysis, 19 Environmental stress, effects on genome, 2 1 3-21 5 Error rates, in stability statistics, 227-230 Escherichia coli, ethylene synthesis in, 63 Ethylene activity, 50 metabolism in soil, 108-1 10 microbial synthesis, 59-64 mycorrhizae and, 9 6 9 7 in pathogenesis, 102-103 in plant-microbe interactions, 114-1 16, 118-119 precursors, exogenous application, 120-1 2 1 in Rhizobium, 80 root nodules and, 87-88 Evaluation, in cultivar development, 21 1-212
3 02
INDEX
Evolutionary biology, genotype-by-environment interactions and, 208 Exudate, root soil cytokinin levels and, 107-108 soil gibberellin levels and, 107 tryptophan in, 116
F Fasciation, cytokinin and, 101-102 Fertilizer estimation of application rates, 287 in long-term field experiments, 159-161, 170-172, 192-193 Fertilizer industry, origin of, 255 Field experiments, long-term, see also Morrow Plots in United States, 153-154 Filters, in atomic force microscopy, 5-6 Fluid cell, in atomic force microscopy, 22-25 Forage, SOM dynamics and, 285 Force modulation imaging, atomic force microscopy in, 38 Forest soil modeling of, 258 SOM dynamics and, 284-285 Free-living diazotrophs, plant growth regulators in, 69-75 Friction, in atomic force microscopy, 7 Fungi, see also Mycorrhizae abscisic acid synthesis in, 65-68
G Gene loss, in breeding, 2 10-21 I Genetic diversity, in crop breeding, 240 Genetics phenotypic plasticity and, 222-223 in yield improvements, 209 Genome, effects of environmental stress on, 213-215 Genotype defined, 201 phenotypic expression and, 201-202 stability and, 225-226 static versus dynamic, 225 Genotype-by-environment interaction in crop breeding, 201, 204,208-212, 223-225,240-24 1 program design and, 239-240
for stress resistance, 234-236 while developing cultivar stability, 237-239 crossover, 207 defined, 200 distinguished from genotype-environment correlation, 203-204 early multienvironment testing and, 239 environmental variables and, 230-23 1 during growth stages, 239 modeling of, 203 noncrossover, 207-208 phenotypic plasticity and, 222-223 stability statistics and, 225-234 statistical detection, 205 stress and, 2 12-222 unbalanced data sets in, 23 1-234 variability in, 205-207 Genotype-environment correlation, 203-204 Gibberella in Bakanae disease, 102 gibberellin synthesis in, 5 6 5 8 Gibberellin in Bakanae disease, 102 exogenous, physiological effects, 120 forms, 48 metabolism in soil, 107 microbial synthesis, 5 6 5 8 mycorrhizae and, 93-94 in plant-microbe interactions, 113 precursor-inoculum interactions and, 118 in rhizobacteria, 72.79 root nodules and, 83-85 Global climate change, soil transformations modeling and, 288-289 Global Climate Change and Terrestrial Ecosystems Project, 259,288 Glomus plant abscisic acid levels and, 97 plant cytokinin levels and, 95-96 Glucose, in ethylene synthesis, 61 Glutamate, in ethylene synthesis, 64 Growth in mycorrhizal plants, 94 plant growth regulator precursors and, 119-121
H Hairy root syndrome, auxin and, 101 Hay, see Corn-oat-hay rotation
INDEX Heat-sensitivity, heterosis and, 223 Heat-shock genotype-by-environment interactions and, 217-218 proteins, 218 Hebeloma, auxin and, 93 Hematite, atomic force microscopy and, 111,.
IL
Herbicides, genotype-by-environment interactions and, 2 17 Heredity, environmental influences, 202-203 Heritability, genotype-by-environment interactions and, 2 10 Heterosis, stress and, 222-223 Homeostasis, heterosis and, 223 Hormones, see Plant growth regulators Humic substances, atomic force microscopy and, 38-39 Humus, origin of term, 255 Hybrids corn, in long-term field experiments, 158-159, 174-175, 192 heterosis and, 223 Hydroxylapatite atomic force microscopy and, 11 sorbtion of lead to, 32-34 Hydroxypyromorphite atomic force microscopy and, 10-1 1 in sorbtion of lead to hydroxylapatite, 3234 Hyperauxiny, in mycorrhizal plants, 93
I iaaH gene, in Pseudomonas, 100 iaaM gene, in Pseudomonas, 100 Illinois, corn yield trends, 177-180, 193 Inbreeding depression, heterosis and, 223 Indirect selection, genotype-by-environment interactions and, 238-239 Indole-3-acetaldehyde, 54 Indole-3-acetamide in auxin synthesis, 54-56, 76.79 production in soil, 106 Indole-3-acetic acid, see Auxin Indole-3-acetonitrile, 55-56 Indole-3-ethanol in auxin synthesis, 54 production in soil, 106 Indole-3-lactic acid, 54
303
Indole-3-pyruvic acid pathway of auxin synthesis, 53-54 production in soil, 106 Inoculants, plant growth regulators and, 110-119 Interfaces, in soil, 1 lsopentyl alcohol, in precursor-inoculum interactions, 11 8 Isoprenoid pathway, in abscisic acid synthesis, 65 Israelachvili-Adams apparatus, 17
K Kaolinite, surface analysis, 12, 35-37 ent-Kaurene, 56 Kaurene synthetase, 56 a-ketoglutaric acid, 63-64 KMB A, see 2-0x0-methylthiobutyric acid
L Laccuria, ethylene and, 96-97 Lavoisier, Antoine Laurent, 254-255 Lead, sorbtion to hydroxylapatite, 32-34 Legumes, microbial plant growth regulators and, 8 1-89 Liebig, Justis von, 255 Light-use efficiency, genotype-by-environment interactions and, 217 Lima bean, root nodules in, 84-85 Lime, in long-term field experiments, 158, 168-171, 192 Long-term field experiments, see also Morrow Plots in United States, 153-154 Low-energy electron diffraction, in mineral surface analysis, 1-2, 15, 17
M Maize, see Corn Manure, in long-term field experiments, 158, 161, 192 effects on soil organic matter, 185 effects on yield, 165, 168-170, 172 Manure-lime-phosphate treatment, in long-term field experiments, 168-169, 192 Mean residence time, 278 Methionine in ethylene synthesis, 50.60-64 exogenous, physiological effects, 120-121
3 04
INDEX
Mevalonic acid in abscisic acid synthesis, 50, 67 in gibberellin synthesis, 48, 56, 1 I8 Mica, in atomic force microscopy, 18, 28-29 Microbial kinetics, studies with tracers, 28 1 Microorganisms, see also Plant-microbe interactions; Rhizobacteria abscisic acid synthesis in, 64-68 auxin catabolism in, 56 auxin synthesis in, 51-56 cytokinin synthesis in, 58-59 ethylene synthesis in, 59-64 gibberellin synthesis in, 56-58 root nodules and, 80-89 in soil ethylene production, 109-110 in soil nitrogen modeling, 270-271 in soil transformations modeling, 262, 265-270,272-275 as sources of plant growth regulators, 50-
planting date and, 175-176 soil organic matter and, 189-191 yield stability and, 172-174 data analysis and, 162-163 management history, 154-161, 191-194 record keeping in, 161-162 soil organic matter in, 181-187, 189-191, 193- I94 soil pH in, 187, 191 soil phosphorus in, 187, 189, 191 soil potassium in, 189, 191 Muscovite mica in atomic force microscopy, 28-29 in Israelachvili-Adam apparatus, 17 Mutations environmental stress and, 214 heterosis and, 223 Mycorrhizae, plant growth regulators and, 89-97
51
Miles, Manley, 156 Mineralization-immobilization turnover, 254-255,270 Minerals in siru studies of growth and dissolution, 19-27 surface analysis, 1-2, 15-17 Mineral uptake, genotype-by-environment interactions and, 216 Mineral-water interactions double-layer forces in, analysis of, 17-18 in situ studies, 19-27 Mixed model equations, for unbalanced data sets, 232-234 Molecular plant breeding, 236 Morrow, George E., 155-156 Morrow Plots corn yield in compared to Illinois state trends, 177-180, 193 effects of weather on, 172, 176177, 192-1 93 high-level NPK treatment, 171-172 hybrids and, 1 7 4 1 7 5 manure-lime-phosphate treatment, 168-169 nitrogen-lime-phosphate-potassium treatment, 169-171 original experiment, 167 overall experimental effects, 163-166
N NCSOIL model, 262,267,274 Nitrification, in soil nitrogen modeling, 27& 27 1 Nitrite, in soil nitrogen modeling, 271 Nitrogen, see also Soil nitrogen effects on soil transformations rate constants, 274-275 mineralization, modeling of, 258-259 I5Nitrogen, in studies of microbial kinetics, 28 1 Nitrogen-fixing bacteria, auxin in, 112 Nitrogen-fixing nodules, plant growth regulators and, 80-89 Nitrogen-lime-phosphate-potassium treatment, in long-term field experiments, 169-1 7 1 Nitrous oxide, modeling of soil production, 272 nod factors auxin and, 8 1 in root nodule formation, 86 Nodules, nitrogen-fixing, plant growth regulators and, 80-89 Noncrossover interaction, in genotype-by-environment interactions, 207-208 Norms of reaction, 205 defined, 201 environmental influences and, 202-203
INDEX NPK fertilizer, in long-term field experiments, 159-161, 170-172, 192-193 Null hypothesis, in stability statistics, 2 2 6 230 Nutrient uptake, genotype-by-environment interactions and, 216 Nutrient-use efficiency, genotype-by-environment interactions and. 217
0 Oats, see Corn-oat-hay rotation; Corn-oat rotation Oxidative stress, genotype-by-environment interactions and, 218-219 2-0x0-methylthiobutyric acid, in ethylene synthesis, 63 Oxygen, distribution in soil, 273
P Pantoea. effects on winter wheat, 112-1 I3 Pathogenesis, plant growth regulators and, 97-104 Pauli-exclusion forces, in atomic force microscopy, 7 Penicilliurn, ethylene and, 103 Performance evaluation, in cultivar development, 211-212 Petunia, environmental stress and, 213 pH, in long-term field experiments, 187, 191 Phase imaging, atomic force microscopy in, 38-39 Phaseohs, root nodules in, 84-85 Phenotype defined, 201-202 environmental influences and, 202-203 Phenotypic plasticity environmental stress and, 215 genotype-by-environment interactions and, 222-223 Phenylalanine, exogenous, physiological effects, 119 Phosphate, in long-term field experiments, 158, 168-171, 192 Phosphorus in long-term field experiments, 187, 189, 191 mycorrhizal plants and, 95-96
305
Photosynthesis, environmental stress and, 214-215 Phytohormones, see Plant growth regulators Pisolithus auxin and, 93 ethylene and, 96 in precursor-inoculum interactions, 117 Plant breeding, see Crop breeding Plant growth-promoting rhizobacteria, see also Azospirillum; Azoiobacter; Rhizobium as inoculants, 110-116, 122 plant growth regulators and, 68-69 precursor-inoculum interactions and, 11& 1 I9 types, 68 Plant growth regulators, see also specijc types in Azotobacter and Azospirillum, 69-75 exogenous application, 122 metabolism in soil, 104-110 microbial synthesis, 51-68 mycorrhizae and, 89-97 pathogenesis and, 97-104 in plant growth-promoting rhizobacteria, 68-69 in plant-microbe interactions, 121-122 precursor-inoculum interactions and, 116119 precursors, exogenous application, 119-121 in Rhizobiurn, 76-89 in rhizosphere, 50-51, 104-110, 121 in root nodules, 80-89 sources of, 50-5 1 types, 4 6 5 0 Planting date, effect on yield in long-term field experiments, 175-1 76 Planting density, see Seeding density Plant-microbe interactions in mycorrhizal symbiosis, 89-97 plant growth regulators in, 110-116, 121, 122 precursor-inoculum interactions and, 116119 in rhizosphere, 46,68-69 in root nodules, 80-89 Plasmid, tumor-inducing, auxin and, 99 Polycarbonate membranes, in atomic force microscopy, 28-30 Potassium, in long-term field experiments, 189, 191 Prairie, SOM dynamics and, 285
3 06
INDEX
Precursor-inoculuminteractions, 116-1 19 Pseudomonas putida, ACC deaminase in, 114, 116 Pseudomonas spp., auxin in, 53,55, 111 Pseudomonas syringae pv. savastanoi, plant growth regulators and, 99-101
Q Q-SOILmodel, 268-270 Quantitative trait loci, in crop breeding, 235-236
R Rate constants, in soil transformations models, 272-275 Reproductive adjustment, genotype-by-environment interactions and, 216 Restricted maximum likelihood, in unbalanced data sets, 233-234 Restriction-fragmentlength polymorphism, in crop breeding, 235-236 Rhizobacteria,see also Azospirillum; Azotobacter; Rhizobium as inoculants, 110-1 16, 122 plant growth regulators and, 68-69 precursor-inoculuminteractions and, 116119 types, 68 Rhizobium auxin in, 76-79, 112 cytokinin in, 79-80 ethylene in, 80 gibberellin in, 79 root nodule formation and, 80-89 Rhizosphere plant growth regulators in, 50-51, 104-110, 121 plant-microbe interactions in, 46,68-69 Rock phosphate, in long-term field experiments, 158, 169, 192 Root exudate soil cytokinin levels and, 107-108 soil gibberellin levels and, 107 tryptophan in, 116 Root nodules, plant growth regulators and, 8Cb 89 RothC-26.3 model, 262,265,267,273-274, 281,287
S Salt stress, genotype-by-environmentinteractions and, 219 Scanning electron microscopy,in mineral analysis, 32 Scanning force microscopy,see Atomic force microscopy Scanning tunneling microscopy, in mineral analysis, 16, 17, 19 Secondary-ion mass spectroscopy,in mineral surface analysis, 1 Seed companies,evaluation costs and, 212 Seeding density, in long-term field experiments, 159-161, 192 Selection, see also Crop breeding gene loss and, 211-212 genotype-by-environmentinteractions and, 210 indirect, 238-239 stability and, 227-230 for yield and stability, 226-227 Silicon nitride probes, in atomic force microscopy, 8-9 Silver, C.W., 156 Soil, plant growth regulators in, 104-1 10 Soil aggregates, oxygen distribution and, 273 Soil carbon, see also Soil organic matter historical approaches to, 254255 in long-term field experiments, 181187 modeling, 262-270,287-289. see also Soil transformations modeling variability in due to methodology,277 Soil-crop models, 286-287 Soil nitrogen historical approaches to, 255 in long-term field experiments, 181187 mineralization-immobilizationturnover, 256255,270 modeling, 256,270-272,276. see also Soil transformations modeling Soil organic matter '3C:'2C ratio in, 281,284 in forest to sugarcanetransition, 2 8 6 285 in long-term field experiments, 181-187, 189-191, 193-194
INDEX modeling, 256-258.262-270 evaluation with tracer data, 278-285 long-term evaluations, 277-278 relationship with clay, 273-274 Soil Organic Matter Network global climate change and, 288 soil transformations models and, 259-272 rate constants in, 272-275 sensitivity analysis and, 286 Soil organic pools, in soil transformationsmodeling, 262 Soil particles, see also Minerals atomic force microscopy and, 2-3, 27-37 methods of surface analysis, 1-2 Soil pH, in long-term field experiments, 187, 191 Soil phosphorus, in long-term field experiments, 187, 189, 191 Soil pores, water-filled, 273 Soil potassium, in long-term field experiments, 189,191 Soil science, history of, 253-255 Soil transformations modeling agronomic applications, 286287 confounding calibration and validation in, 276 current models, 259-272 early models, 256259 evaluation with tracer data, 278-285 global scale application, 288 long-term evaluations, 276278 model structure, 262-272 modularization, 290 predictive uses, 289 rate constants in, 272-275 regional scale application, 287 in research, 285-286 role of soil biota in, 290 soil organic pools in, 262 validation of, 275,288-289 SOM, see Soil organic matter Sour orange, cytokinin levels in, mycorrhizae and, 95-96 Soybean, see also Corn-soybean rotation SOM dynamics and, 285 Specific surface area, in kaolinite, 36-37 Spectroscopy, in mineral analysis, 1.2, 17.20, 27 Stability biological concept of, 225
307
crop breeding and, 237-239 environmentalvariables and, 230-23 1 during selection, 227-230 selection for yield and, 226-227 types, 225-226 Stability statistics, 225-234 Stability-variancestatistic, 228-230 Stagonospora, abscisic acid in, 103-104 Static concept, of genotype, 225 Stress, genotype-by-environmentinteractions and, 2 12-223 Stress resistance, in crop breeding, 234-236 Sugarcane indirect selection and, 238 SOM dynamics and, 284-285 Superoxide dismutase, stress resistance and, 219
T Temperature effects on atomic force microscopy, 2324 effects on soil transformationsrate constants, 273 genotype-by-environmentinteractions and, 2 17-21 8 Tillage, see Cultivation Tomato, abscisic acid levels, pathogens and, 104
Tracers, in evaluation of soil transformations models, 278-285 Transmission electron microscopy, in mineral analysis, 15, 17, 32 Transposableelements, environmental stress and, 214 Tryptamine, in auxin synthesis, 54 Tryptophan in auxin synthesis, 48.51-55 exogenous, physiologicaleffects, 119-120 in precursor-inoculum interactions, 1 16117 in root exudates, 116 soil auxin levels and, 104-107 Tryptophol in auxin synthesis, 54 production in soil, 106 Tull, Jethro, 191 Tumor-inducing plasmid, auxin and, 99 Tumor-inducing principle, 99
3 08
INDEX
U University of Illinois, Morrow Plots and, 155-156
Witch’s broom, cytokinin and, 101-102 World population, agricultural production and, 208
X V Van der Waals forces, in atomic force microscopy, 7 Van Helmont, Jean-Baptiste, 254 VERBERNE model, 262,267-268,274 Vesicular-arbuscular mycorrhizae, plant cytokinin levels and, 96 Vigna, root nodules in, 84-85
W Water, see also Mineral-water interactions effects on soil transformations rate constants, 273 Water-filled pore space, 273 Water-use efficiency, genotype-by-environment interactions and, 217 Weather, effects on yield in long-term field experiments, 172, 176-177, 192-193 Weed control, precursor-inoculum approach, 117 Wheat, SOM dynamics and, 285 Winter wheat, effects of Pantoea on, 112-1 13
X-ray diffraction, in mineral analysis, 17.32, 3637 X-ray photoelectron spectroscopy, in mineral analysis, 1, 17.20.27
Y Yield, see also Corn yield improvements, contributions to, 209 Yield stability factors in, 230 selection for, 226-227 Yield stability statistics, 227-228
Z Zeatin, 48 in cytokinin synthesis, 58-59 in myconhizal plants, 95-96 in Rhizobium,80 root nodules and, 85-86 soil cytokinin levels and, 108
This Page Intentionally Left Blank
I S B N 0-12-0007b2-2