Advances in Botanical Research, Vol. 15

Advances in

BOTANICAL RESEARCH VOLUME 15

Advances in

BOTANICAL RESEARCH Editor-in-Chief J. A. CALLOW

Department of Plant Biology, University of Birmingham, Birmingham, England

Editorial Board H. W. WOOLHOUSE W. D. P. STEWART E. G. CUTTER W. G. CHALONER

E. A. C. MAcROBBIE

John Innes Institute, Norwich, England Department of Biological Sciences, The University, Dundee, Scotland Department of Botany, University of Manchester, Manchester, England Department of Botany, Royal Holloway & Bedford New College, University of London, Egham Hill, Egham, Surrey, England Department of Botany, University of Cambridge, Cambridge, England

Advances in

BOTANICAL RESEARCH Edited by

J. A. CALLOW Department of Plant Biology University of Birmingham Birmingham, England

VOLUME 15

1988

ACADEMIC PRESS Harcourt Brace Jovanovich, Publishers

London San Diego New York Berkeley Boston Sydney Tokyo Toronto

ACADEMIC PRESS LIMITED 24/28 Oval Road London NW17DX

United States Edition published by ACADEMIC PRESS INC. San Diego, CA 92101

Copyright 0 1988by ACADEMIC PRESS LIMITED

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.

British Library Cataloguing in Publication Data Advances in botanical research.-Vol. 15 1. Botany-Serials 581'.05 ISBN 0-12-005915-0

Typeset by Paston Press, Loddon, Norfolk and printed in Great Britain by T.J. Press (Padstow), Cornwall

CONTRIBUTORS TO VOLUME 15

THOMAS BJORKMAN, Department of Botany, Universityof Washington, Seattle, Washington, U S A H. GRIFFITHS, Department of Biology, University of Newcastle, Newcastle upon Tyne, N E l 7RU, UK LEON V. KOCHIAN, US Plant, Soil and Nutrition Laboratory, USDAA R S , Cornell University, Ithaca, New York, U S A WILLIAM J. LUCAS, Department of Botany, University of California, Davis, California, U S A ROGER I. PENNELL, John Znnes Institute and A F R C Institute of Plant Science Research, Colney Lane, Norwich, NR4 7UH, U K

This Page Intentionally Left Blank

PREFACE

The major part of this volume of the Advances is devoted to three of the most enduring themes in plant physiology, the mechanisms by which plants perceive and respond to gravity (Bjorkman), Crassulacean Acid Metabolism (CAM) and its regulation (Griffiths), and the mechanisms of K+ uptake and transport in roots (Kochian and Lucas). All three have been reviewed many times, what justification is there for further treatments of these perennial favourites? In his evaluation of the various hypotheses that seek to explain gravity perception and response, T. Bjorkman seeks a more critical application of the laws of physics than hitherto. It appears, however, that a basic lack of information still limits our ability to advance sound models at the subcellular and molecular level. Ways of rectifying the situation are presented with particular emphasis on incorporating recent advances in the understanding of cellular signalling mechanisms. Much of the pioneering biochemical work on CAM and its regulation was carried out by the Newcastle Group, most notably involving M. Thomas and S . L. Ranson. In continuing the Newcastle tradition, H. Griffiths’ review is concerned with the ecophysiological aspects of CAM regulation as found in plants of very diverse terrestrial and aquatic habitats. The emphasis here is on the integrated study of the various primary aspects of CAM and its diverse secondary consequences. The author also seeks to rectify the conventional view that CAM is all about malate, by raising the, as yet, not entirely explained role of citrate. Despite a vast literature on potassium uptake and accumulation by plant roots, it seems that our understanding of the mechanisms involved and their regulation is still far from complete. In their highly authoritative review, L. V. Kochian and W. J. Lucas give appropriate consideration to the historical dimension in attempting a synthesis of the current models, but they are also concerned to look to the future to identify the most profitable lines of research through the integration of the physiological approach with advanced biophysical techniques such as “patch-clamping” , and with improved immunological and gene cloning methods for characterizing membranes and their protein components. vii

...

Vlll

PREFACE

Volume 15 is not entirely devoted to physiology. Although the cellular processes involved in the formation of microspores and megaspores are fairly well documented for flowering plants other groups of seed-bearing plants are more neglected and, taking Tuxus as his main example, R. Pennell outlines recent work on this subject. I thank all the authors for their endeavours and efforts to minimize the editor’s task. J. A. CALLOW

CONTENTS

CONTRIBUTORS TO VOLUME 15 . . . . . . . . . . PREFACE . . . . . . . . . . . . . . . . . . . .

V

vii

Perception of Gravity by Plants THOMAS BJORKMAN I. I1.

Objectives

. . . . . . . . . . . . . . . . . . .

Susception . . . . A . How Gravity Acts B . Thermal Motion

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

I11.

Transmission . . . . . A . Electrical Transmission B . Chemical Transmission

IV .

Perception . . . . . . . . A . Signal Transduction Overview B . Multiple Systems . . . . C . Statolith Sensors . . . . D . Nonstatolith Perception . .

V.

Integration and Conclusion

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . . . . . . . . . .

1 3 3 4

7 8 10 12 12 16 18 32 36

Crassulacean Acid Metabolism: a Re-appraisal of Physiological Plasticity in Form and Function H . GRIFFITHS I. I1.

Introduction

. . . . . . . . . . . . . . . . .

Modification and Regulation of Constitutive CAM . . . . . . A . Occurrence and Distribution of Constitutive CAM . . . .

ix

44 45 45

CONTENTS

X

B. C. D. E.

Characterization of the Die1 Cycle . . . . . . . . . Biochemical Regulation . . . . . . . . . . . . . Modification of CAM Phases by the Environment . . . . Plant Water Status: Is there Regulation both of and by Solute Accumulation? . . . . . . . . . . . . . . . . Stable Isotope Ratio Analysis . . . . . . . . . . . Regulation of Respiratory C 0 2 Recycling During CAM . . .

46 50 51

111.

Plasticity of Metabolic Response: Shades of CAM . . . . . . A . CharacteristicsofC,- CAMIntermediates . . . . . . . B . Occurrence, Distribution and Evolution . . . . . . . . C . Respiratory COz Recycling by C,- CAM Intermediates . . . D . Physiological Characteristics of the C,- CAM Transition . .

67 68 70 74 76

IV .

Significance of Respiratory COz Utilization During CAM . . . . A . RecyclingintheTerrestrialEnvironment . . . . . . . B . Recycling in the Aquatic Environment . . . . . . . . C . Organic Acid Speciation: The Newcastle Hypothesis Revisited

78 79 82 83

. . . . . . . .

85

F.

G.

V.

CAM: Development of Integrated Research

53 55 64

Potassium Transport in Roots LEON V . KOCHIAN and WILLIAM J . LUCAS

. . . . . . . . . . . . . . . . . .

94

I1 .

Plasma Membrane Transport of K+in Roots . . . . . . . . A . Early Work: thecarrier-Kinetic Approach . . . . . . . B . Are Root K+Fluxes Coupled to H+? . . . . . . . . . C . Uptake at High K+ Concentrations: the Linear Component . D . Summary . . . . . . . . . . . . . . . . .

94 94 99 120 128

111.

Redox-coupledPlasmalemmaTransportofK' . . . . . . . A . Influence of Exogenous NADH on K+ Influx . . . . . . B . Membrane Transport and the Wound Response . . . . . C . DevelopmentofanIntegratedNADHModel . . . . . .

129 130 131 132

IV .

Radial K+Transport to the Xylem . . . . . . . . . . . A . Site of K+ Entry into the Symplasm . . . . . . . . . B . Radial Pathway . . . . . . . . . . . . . . . C . Lag Phase in Xylem Loading . . . . . . . . . . . D . K+Transport into the Xylem . . . . . . . . . . . Regulation of K+ Fluxes within the Plant . . . . . . . . . A . Allosteric Regulation of K+Transport . . . . . . . . B . K+ Cycling within the Plant: an Integration of Regulatory Mechanisms . . . . . . . . . . . . . . . .

136 137 140 143 145 151 152

I.

V.

VI .

Introduction

Future Research and Prospects

. . . . . . . . .

163 169

CONTENTS

xi

Sporogenesis in Conifers ROGER I . PENNELL I.

. . . . . . . . . . . . . . . . . .

179

I1.

Sporogenesis in the Pollen-bearing Cone . . . . . . . . . A . The Archaesporium and Differentiation within the Sporangium B . Sporogenous Cells and Tapetum . . . . . . . . . . C . Meiosis . . . . . . . . . . . . . . . . . . D . ExinePatterningandtheFreeSporePeriod . . . . . .

181 181 182 183 185

I11.

Megasporogenesis . . . . . . . . . . . . . . . . A . The Origin of the Reproductive Cell Lineage within the Ovule B . Mitochondria, Plastids and Planes of Division within the Megaspore Mother Cell . . . . . . . . . . . . . . . C . Megaspore Viability . . . . . . . . . . . . . .

190 190 191 193

AUTHOR INDEX . . . . . . . . . . . . . . .

197

SUBJECT INDEX . . . . . . . . . . . . . . .

207

Introduction

This Page Intentionally Left Blank

Perception of Gravity by Plants

THOMAS BJORKMAN

Department of Botany KB-15 University of Washington, Seattle, Washington 98195, USA

I.

11.

111.

IV.

. . . . . . . . . . . . . . . . . . Susception . . . . . . . . . . . . . . . . . . A. How Gravity Acts . . . . . . . . . . . . . . B. Thermal Motion . . . . . . . . . . . . . . Transmission . . . . . . . . . . . . . . . . . A. Electrical Transmission . . . . . . . . . . . . Objectives

1

. . . . , .

3 3 4

. . 7 . . 8 B. Chemical Transmission . . . . . . . . . . . . . . 10 Perception . . . . . . . . Signal Transduction Overview B. Multiple Systems . . . . C. Statolith Sensors . . . . D. Nonstatolith Perception . . A.

V.

. .

Integration and Conclusion

. . . . . . . . . . . . . . .

. . .

. . . . . . . . . . . . . . .

. . . . .

. . . . . 12 . . . . . 12 . . . . . 16 . . . . . 18 . . . . . 32

. . . . . . . . . . .

36

I. OBJECTIVES An environmental cue widely used by plants to guide development is gravity. Although many things are known about plant responses to gravity, one fundamental aspect remains obscure, i.e. the means by which the physical stimulus of gravity is transduced into a physiological response which the plant can use to guide development. Copyright 01988 Academic Press Limited All rights of reproduction in any form reserved.

Advances in Botanical Research Vol. 15 ISBN 0-12-005915-0

1

2

T. BJORKMAN

In order to analyse that transduction, both the physics and physiology must be considered. Although the physical aspects have been considered before, notably by Audus (1979), some reiteration and expansion of earlier discussions is useful to interpret recent data and to indicate valuable concepts to pursue. In this chapter the intention is to give these physical laws deeper attention than they have received in the past. This discussion of physics and gravity perception should point out paradigms, some provocative, which will be helpful in developing new hypotheses of gravity perception. The physical behaviour of objects on a human scale is very different from that on the subcellular scale. It is easy, when imagining how gravity perception might work, to make erroneous assumptions about how various components will act. It is hoped to give the reader an intuitive grasp of how gravity acts on plant cells and how that may be turned into physiological information. The perception of other physical stimuli, light and sound, has recently been well characterized in animals. Metabolic syndromes in these transduction mechanisms may have parallels in gravity perception by plants. The experimental approaches which led to the elucidation of the animal transducers also provide a useful guide to promising future approaches. Gravitropism is the result of a series of events, and the terms used for each step are as defined by Hensel(1986a). Susception is the initial physical reaction by amass in the gravitational field. Perception is the conversion of the physical signal to a physiological one. Transmission is how the signal moves from the cells where gravity is sensed to those where growth occurs. The response is the differential growth which results in curvature. The term transduction is sometimes used to describe the step called perception, but here a more specific meaning will be used: the carrying over of energy or information from one form (or place) to another. By that definition, transduction occurs in each of the four steps of gravitropism. Gravity sensing (or gravisensing) is also a common term used to describe part of gravitropism, usually corresponding to susception and perception. There is not yet a consensus on the terminology describing the steps of gravitropism; as these steps become physiologically better defined, so will the words used to describe them. In order to discuss potential mechanisms of perception, gravity will first be considered from a physical perspective and related to a plant’s susception of gravity. Second, the types of signals which may be involved in transmission will be reviewed. In the main section, the means by which the information provided by susception can be transduced to the kinds of signals which may be transmitted will be considered. Some potential perception mechanisms will be evaluated in terms of their likelihood of performing as rapidly and as sensitively as does the true gravity-sensing system.

PERCEPTION OF GRAVITY BY PLANTS

3

11. SUSCEPTION The first step in gravitropism is a physical action of gravity on some element in the plant, which is called susception. Gravity is an attractive force on a mass; it is that force which a graviresponding plant must use to orient itself relative to the gravitational field. Although gravity acts on every atom in the plant, the number of relevant interactions is limited. This section will cover the action of the gravitational force and limitations to its detection. A. HOW GRAVITY ACTS

To analyse how a plant senses gravity one must know how gravity interacts with physical objects. Newton’s law of gravitation holds that two objects attract each other with a force proportional to the product of their masses and inversely proportional to the square of the distance between them:

Fg = Gmm’/r2 where m and m’ are the two masses and r the distance between them. G is the experimentally determined gravitational constant. The attractive gravitational force will tend to move two particles towards each other, which we can observe, for example, as the attraction between an apple and the earth. In the case of gravity sensing by plants on the earth’s surface, one particle is the earth, and the other is within the plant. The distance between the particles is the radius of the earth because the gravitational force of a sphere acts as if all the mass is at its centre. Thus the values for G, m and r are constant and the gravitational equation reduces to: Fg = (9.8 m s-’) m’. The gravitational force then depends on the mass (m‘)of whatever particle we are considering. A plant must sense gravity by detecting the attraction between the earth and the mass of some object associated with the plant, which will be referred to as the sensing particle. For that attraction to be detectable, the sensing particle must do work (in the thermodynamic sense) on something to cause a change in the physiological activity of the plant. Displacement of the particle in the gravitational field is required to convert gravitational potential energy to work. If the particle does not move, there is no energy to alter the physiology of the plant. A simple example of work, as defined in physics, is a mass moving against gravity. The work done is the force applied ( g x m ’ ) times the distance the mass is moved. This can be illustrated by a seesaw (Fig. 1). With mass A on the low side, if a larger mass, B, is placed on the high side it will exert more force and drop. As it drops, it loses gravitational potential energy and does work on mass A. The work done on mass A causes the

4

T. BJORKMAN

Fig. 1. Displacement of a mass in a gravitational field does work. (a) Mass A is at rest on a seesaw; a larger mass (B) is placed on the high side of the seesaw. (b) Mass B exerts more force and therefore moves the seesaw. As it descends, it loses potential energy (B x g X d ) and does work on mass A by raising it in the gravitational field. (c) The mechanical work done on mass A increases its gravitational potential energy by A X g X d. The work could also be converted to: (d) electrical energy by running a generator; or (e) chemical energy by running a reverse osmosis unit which separates solutes from water.

mass to rise, increasing its gravitational potential energy. The rising arm of the lever could also cause a generator to turn, creating electrical energy; or it could push the piston in a reverse osmosis unit, creating chemical potential energy by purifying water. Work can convert the gravitational potential energy to many other forms of energy, depending on the transduction mechanism. Therefore, a simple model of susception can lead to many possibilities for perception. Susception can occur in a number of possible ways: the sensor may do work either by being denser than the surrounding medium and sinking, or by being lighter and rising. The sensor may be inside or outside the cell. There are many objects within the cell which may move relative to each other due to their differing densities. The cell could perceive motion, displacement or position of the sensor. We will see what evidence there is for each of these things happening. B. THERMAL MOTION

The mass acted on by gravity is also being constantly agitated by collision with water and other molecules which have kinetic energy due to heat.

PERCEPTION OF GRAVITY BY PLANTS

5

This random thermal motion is commonly seen as Brownian motion. For the gravity sensor to be effective, it must be relatively insensitive to the random thermal energy but be very sensitive to changes in the direction of gravity. An important limiting factor for a sensing mechanism, then, is thermal noise. The magnitude of thermal energy on a particle is +kT in each dimension, where k is Boltzman’s constant (1.38 X lovz3J K-’)and T is the absolute J. The thermal temperature. At room temperature, f k T = 2 x agitation of a particle is independent of the particle’s mass, so the effect of gravity relative to the thermal noise is greater the more massive the particle. Thermal noise sets one lower limit on the minimum work which must be done by a sensor during susception. The rate of a chemical reaction is limited by the activation energy of the reaction. Thermal motion of the reactant provides the energy needed to overcome the activation energy of a chemical reaction. For a reaction caused by a mechanical stimulus, in contrast, the sensor should be selectively activated by the stimulus rather than thermal motion. This can be accomplished if the activating reaction has an activation energy high enough that it is rarely stimulated spontaneously. If the activation energy is high, the minimum stimulus must be correspondingly large. There is a trade-off between a sensor’s sensitivity and its selectivity. The effect of activation energy on the spontaneous reaction rate can be calculated (Fig. 2). The frequency of activation by thermal energy decreases very rapidly as the activation energy increases. The figure is based on the Arrhenius equation: Rate = Ae-E’RT.The light sensor in vision is rhodopsin, which is physically activated by the energy in photons. Rhodopsin is stimulated only very rarely by heat. If the enzyme reduced the activation energy for the reaction only two-fold, thermal activation could occur 10” times faster. Rhodopsin functions well in light perception because its activation energy is high enough to make spontaneous triggering very rare (about once every 1000 years per molecule), but is low enough that the light stimulus contains ample energy. By a similar approach, this figure can be used to estimate the activation energy of gravity perception. The rate of the first step in gravity perception by the thermal motion of the suscepting body must be much less than the rate during gravistimulation. The activation energy of that step must thereto fore be high enough that the thermal energy (hkT) produces activations for each one produced by a small gravistimulation. From Fig. 2 it can be determined that the activation energy of the first step of perception which fits this criterion is 3 4 x lO-”J. Thus the activation energy, and hence the amount of work required during susception, can be fairly precisely estimated. Presumably there would be many activating events per cell per second, each using about 4 X lo-’’ J. This estimate applies regardless of the specific mechanisms of susception and perception.

6

T.BJORKMAN

-5

-

-10

-

1/2 kT

Enzymatic reactions

W c

e

c .-0

t -150

p!

m

-0

-20

-25

-

Rhcdopsinactivation

Activation energy ( J 1 Fig. 2. The effect of activation energy on the relative rate of reaction. Enzymic reactions have activation energies clustered over a small range, yet one where the rate of a reaction can vary by many orders of magnitude. Spontaneous thermal activation of the light-sensing molecule rhodopsin occurs at a very low rate. A gravity sensor must be insensitive to thermal energy (IkT), so its activation energy would be expected to be 3 4 X lo-*' J. Intermediate marks on the abscissa are 2 and 5 times the order of magnitude. The formula for this curve is: log R = -E,/kT.

Although small effects of gravity can be amplified through various biochemical cascades, these amplifiers will not discriminate between the desired signal and noise. This lack of discrimination is why amplification alone does not provide the necessary sensitivity to detect weak stimuli. For an amplifier to be useful, the total signal must be filtered or averaged using some criterion to reduce the noise. The possibilities for such processing by the perceiving system will be discussed in a later section. The only measurable physiological response of a plant which yields information about susception is the presentation time, which is the threshold gravistimulation required to elicit a growth response. There is a reciprocity between the presentation time and acceleration when the acceleration is changed from 1 x g by centrifuging or clinostatting (Johnsson, 1965). A metabolic process would not exhibit this sensitivity to the gravitational force, so the presentation time must reflect the physical process of susception. The presentation time may also include some time

PERCEPTION OF GRAVITY BY PLANTS

7

required for the physiological steps of perception (Johnsson, 1965). The inverse relationship between the force and the presentation time is consistent with a threshold displacement of a sensor required to stimulate perception. This threshold may be thermal noise. Hair cells in the cochlea (the auditory receptor in animals) can detect extremely small stimuli through tuning and time-averaging the repetitive stimulus of a sound wave. The limit of perception corresponds to a motion of atomic dimensions (Harris, 1967), even though sensitivity is limited by thermal motion. A measure of the ability of the sensor in susception to overcome thermal noise is the minimum amount of stimulation which will produce a gravitropic response. Avena roots respond to 3 X 10-4g when stimulated as long as 68 h on a clinostat (Shen-Miller et al., 1968). In lettuce seedlings grown in centrifuges aboard the Salyut 7 space station, the shoots had a threshold response at 3 x 10-3g, and the roots at much lower gravity (Merkis et al., 1985). At 1 X g the presentation time can be as short as 7 s for Lepidium roots (Larsen, 1969). Such high sensitivity can be achieved only by signal averaging and with a substantial responding mass. The potential sensors in a plant are limited to those that are large enough to move a perceptible amount relative to thermal noise within the time it takes a plant to detect a gravistimulus. That much motion must be produced even by the very small stimuli to which plants are capable of responding.

111. TRANSMISSION Gravity sensing occurs only in certain regions of the plant, but gravitropic curvature rarely occurs in those cells which sense gravity. A signal which indicates how the responding cells must alter their growth passes from one group of cells to the other. Transmission has been excellently reviewed by Audus (1979). In roots, for example, gravity is sensed in the root cap, but growth occurs in the elongating zone of the root apex, several millimetres away (Darwin, 1899). In etiolated beans, gravity is sensed in the cotyledonary hook, but curvature occurs 2-3 cm lower in the stem (Hart and Macdonald, 1984; Verbelen et al., 1985). The sensing and responding cells may also be separated radially. In coleoptiles, gravity is sensed in the inner mesophyll, but growth is controlled by cells in and near the epidermis (Thimann and Schneider, 1938; Kutschera et al., 1987). Curvature can only be expressed in cells which are growing or which may be induced to grow; sensing cells need not be near a growing zone. A signal must therefore travel from one group of cells where gravity is sensed to another where growth is controlled. The signal may move symplasmically or apoplasmically;in either case the signal must leave the cell where it originates by crossing the plasma membrane.

8

T. BJORKMAN

Membrane transport is tightly regulated, and therefore is a good place to focus attention when trying to narrow down the type of signal which is elicited from the sensing cell. In roots, transmission of the gravistimulus appears to be through modulation of a growth-retarding factor in the elongating zone (Shaw and Wilkins, 1973), potentially an inorganic ion rather than a plant growth regulator (Mertens and Weiler, 1983). In the shoot it is likely to be modulation of a growth-stimulating factor (Dolk, 1933). In grass nodes, a growth initator moves to the epidermis (Kaufman and Dayanadan, 1984). Perception therefore occurs by a mechanism which causes this kind of modulation. Below, the biological mechanisms which elicit such signals are discussed and those which promise to be relevant to gravity perception are identified. Intercellular communication can be either electrical or chemical. The most dramatic example of bioelectrical communication is the nervous system of animals. Chemical communication is typified by hormones, second messengers and plant growth regulators. The nature of the transmitted signal indicates the kind of reaction which is the final step of perception. A. ELECTRICAL TRANSMISSION

Signals are often transmitted electrically, both in biological and engineered systems. A biological electrical signal can have many manifestations: an action potential, an electrical gradient or electrophoresis in an electrical field. Each of these manifestations could transmit a physiological signal. 1. Action Potentials Action potentials allow rapid communication. The signal for leaf folding in Mimosa pudica (sensitive plant) and Dionea muscipula (Venus flytrap) are carried by action potentials (Pickard, 1973a). In Mimosa, vibration causes an action potential which is transmitted to pulvini at the base of petioles. There, the action potential triggers ion fluxes which cause turgor changes, folding the leaf. In Dionea, stimulation of a trigger hair sends an action potential to the leaf base (Burdon-Sanderson, 1873; Benolken and Jacobson, 1970), initiating rapid growth (Williams and Bennett, 1982). Action potentials are rapid and transient changes in the membrane potential in response to a stimulus. The change in potential is due to increased permeability of an ion which is far from its equilibrium and which normally has a low permeability. The membrane potential then approaches the equilibrium potential for that ion. The potential difference between a stimulated cell and its neighbour is detected, presumably across plasmadesmata, and triggers an action potential in the second cell. This process continues down the line of excitable cells. Cells which can transmit an action potential are termed excitable because they respond actively to a stimulus.

PERCEPTION OF GRAVITY BY PLANTS

9

Action potentials have been measured in many plants, especially in association with rapid movements including growth (Pickard, 1973a), but there are no examples analogous to the processes in gravitropism. Apparent action potentials have been noted in gravitropic Allium roots (Berry and Hoyt, 1943) but not in Lepidium roots (Behrens and Gradmann, 1985) Action potentials remain a possibility for transmission of the gravitropic stimulus, but evidence to support this possibility is lacking.

2. Electrical Gradient A potential applied apoplasmically across a tissue can affect growth (Schrank, 1948; Evers and Lund, 1947; Moore et al., 1987). Gravistimulation does produce an electrical gradient (Schrank, 1947; Grahm and Hertz, 1964; Tanada and Vinten-Johansen, 1980; Behrens et al., 1982; Bjorkman and Leopold, 1987a), raising the possibility that curvature is the result of an electrical field. Although ion currents flowing in such an electrical field gradient have been proposed as guides to development in many tissues (Jaffe and Nuccitelli, 1974), there is no evidence that the field is sensed directly. A direct sensor would not be unheard of: magnetosensors, which sense magnetic fields, are known in biology (Lohmann and Willows, 1987). 3. Electrophoresis An electrical field applies an attractive force to charged particles; therefore, an apoplasmic electrical gradient across a tissue causes mobile ions to move across the tissue through the cell walls. The ions, redistributed by electrophoresis, may then influence growth or act allosterically to modulate growth regulators (Hasenstein and Evans, 1986). There is evidence against electrophoretic translocation of Ca2+ in the establishment of tissue asymmetry, and it is applicable to other ions as well. Curvature is towards the positive pole of an applied potential gradient in both positively gravitropic roots (Bjorkman, 1987) and negatively gravitropic shoots and coleoptiles (Schrank, 1948; Woodcock and Wilkins, 1969a). Transverse movement of Ca2+ is towards the slower-growing, positively charged side both in roots (Lee et al., 1983) and in coleoptiles (Slocum and Roux, 1983). Calcium ions are therefore moving against the electrical gradient; they are not being electrophoresed. Auxin anions would electrophorese towards the slower-growing side in both cases. That is compatible with its role as a growth inhibitor in roots, but there is apparently no auxin redistribution in roots (Mertens and Weiler, 1983). In shoots, auxin stimulates growth and is redistributed against its electrical gradient (Mertens and Weiler, 1983; Bandurski et al., 1984). Auxin would perhaps not be susceptible to apoplasmic electrophoresis in any case because at the low wall pH it would be in the uncharged, protonated form. Thus, existing data disprove such simple models of growth modulation through electrophoresis

10

T. BJORKMAN

of biologically active ions. Nevertheless, electrical polarization appears to be a common feature of the gravitropic response. An electrical gradient is established by differences in electrogenic ion transport across the plasma membrane. Electrogenic transport occurs when a charge is moved across a membrane without a balancing charge, resulting in a change in the electrical potential across the membrane. The electrogenic transport could be varied by changing the activity of an electrogenically ion-transporting ATPase, such as the proton pump. The membrane potential could either hyperpolarize or depolarize, depending on whether the activity of the pump were increased or decreased. Depolarization could also be effected by increasing the permeability to some ion which normally has a large electrochemical gradient across the plasmalemma, as in action potentials. Such a permeability increase would most likely result from opening of a specificion channel and ion movement through it down the electrochemical gradient without balancing counterions. Regulation of ion transport is an area of active research, and there are experimental means of determining how electrical effects of gravistimulation are generated. B. CHEMICAL TRANSMISSION

The unequal signal to the growth zone may be generated by direct chemical transport, with the electrical polarization which results from ion movement being incidental. Signal substances transmitted by this type of mechanism would be those which are actively transported by the cells using polarly distributed carriers for directional facilitated diffusion, or even specific ATP-driven transport proteins.

1. Ion Pumping Directional active transport of ions is widespread in plants to serve their mineral nutrition needs. The systems which are in place to distribute mineral ions in the plant may be adapted to transmit information, or similar independent transport systems could be dedicated to this role. Applied Ca2+ can cause curvature in corn roots, even if they are decapped (Lee et al., 1983), suggesting that an apoplasmic Ca2+gradient may be the transmitted signal. The increased apoplasmic Ca2+ concentration may sensitize the tissue to auxin which is present in growth-inhibiting concentrations (Hasenstein and Evans, 1986; Salisbury et al., 1985). It may also act directly on the wall to inhibit growth (Cleland and Rayle, 1977). Apoplasmic Ca2+ ions therefore fulfil the requirements of a signal substance. Transverse Ca2+fluxes have been measured (Lee et al., 1983) and wall calcium has been shown to redistribute (Slocum and ROUX,1983) following gravistimulation. The redistribution of only one other inorganic ion, K+,has been measured. Although K+ redistribution is important in pulvinus move-

PERCEPTION OF GRAVITY BY PLANTS

11

ment in seismonastic Mimosa, it does not generate the response of the pulvini to gravistimulation (Roblin and Fleurat-Lessard, 1987). If Ca2+ions are electrogenically pumped towards the slower-growingside of a tissue, that side will become electropositive relative to the faster-growing side. Such polarization has been notoriously difficult to measure (Woodcock and Wilkins, 1969b). Reliable measurements in coleoptiles show that polarization is a consequence, not a cause, of auxin redistribution (Grahm, 1964). Apparent upward currents in root caps following gravistimulation (Behrens et al., 1982; Bjorkman and Leopold, 1987a) imply that the upper side is positive relative to the lower. This would be consistent with electrogenic cation transport towards the upper side. However, Ca2+move towards the lower side of root tips (Lee et al., 1983). A calcium-transporting ATPase has been discovered on the plasma membrane (Rasi-Caldogno etal., 1987), but the effect on the charge distribution is unclear because the Ca2+ are exchanged for protons in an unknown stoichiometry. Unfortunately, there is not quite enough information available to determine whether direct pumping of an ion may be the transmitter of the gravity signal.

2. Growth Regulator Pumping Polar (or directional) transport of growth-regulating substances would more directly effect differential growth than ionic messengers. Evidence is weak for growth regulator redistribution in roots (Mertens and Weiler, 1983; Jackson and Barlow, 1981), but good for above-ground parts (Pickard, 1985) with some exceptions (Trewavas, 1981). Polar transport of auxin occurs through an auxin-translocating carrier in the plasma membrane, down an electrochemical gradient maintained through protonation of indoleacetate with protons supplied by the plasma membrane proton pump (Rubery and Sheldrake, 1974). The net result is ATP-driven auxin transport. The transport is polar because the IAA carrier is located only on one flank of the cell. This transport requires Ca" transport from the auxin sink (Niedergang-Kamien and Leopold, 1957), though the biochemical basis for this requirement is unknown. A gradient across the tissue may be established through two kinds of cellular response. Either the top and bottom of each sensing cell respond differently, with each sensing cell behaving similarly, or the cells in the upper part of the tissue may act differently from those in the lower part. The membrane potential of sensing cells in Lepidium root caps respond to gravistimulation; those on the lower side hyperpolarize slowly whereas those on the upper side depolarize rapidly yet transiently (Behrens et al., 1985). This is a clear example of different responses by cells in different parts of the sensitive region. This result is expected if the cell is organized so that the flank towards the epidermis differs from the side towards the centre of the tissue (Volkmann and Sievers, 1979). In grass nodes, such an organiza-

12

T. BJORKMAN

tion is believed to induce auxin transport only from cells on the lower part of a horizontal node (Wright and Osborne, 1977). Whether an electrical gradient is established through change in electrogenic ion transport or a chemical gradient is established by polar transport of some substance, modulation of transport across the plasma membrane is necessary. The tissue gradient can be established either by asymmetric transport across each sensing cell or by cells on the bottom of the tissue responding differently from those on top. Because there is intercellular signalling, in all cases a change in plasma membrane transport would be involved. Based on presently available data we cannot definitely eliminate any of the transmission methods: ion channels, plasmadesmata, proton pump modulation, action potentials, polar transport of ions or growth regulators, and activation of a specific porter.

IV. PERCEPTION The section on susception described ways in which perception could be triggered. The section on transmission described the type of signal produced by perception. Taking the inferences about perception from those sections into account, this section will establish how perception is likely to work and will consider some mechanisms which may safely be rejected. The mass with which gravity interacts in susception must not only detect gravity, but must cause some physiological response. The effect of perception is to transmit a polar signal across the pIasma membrane of some or all of the cells in the sensing tissue. The transduction from an intracellular stimulus to an extracellular signal, perception, could occur through a direct interaction between the sensing mass and a cell component which elicits the transmitted signal. More likely, it is mediated through a cascade of intracellular messengers to trigger transmission of a signal. An intermediate reaction cascade amplifies the signal and provides an opportunity for modulation of the message by other cellular processes. To consider the ways susception could produce a transmitted signal, biological signal transduction in general will be discussed, and then compared to what we know about gravity perception. Some hypotheses which have been proposed for gravity perception in plants will also be discussed. A. SIGNAL TRANSDUCTION OVERVIEW

As a basis for discussing the mechanism of gravity perception in plants, it is worthwhile to first review signal transduction in some sensory systems which are better understood. There may be analogous patterns among the sensory systems which can be useful in predicting perception of gravity by plants.

PERCEPTION OF GRAVITY BY PLANTS

13

1, Other Sensory Systems

The photoreceptors in eyes of invertebrate animals are probably the sensory system which is best characterized at the physiological level (Steive, 1986). Perception of light begins with the absorption of photons by rhodopsin in the plasma membrane of retinal rod cells. Activation of a rhodopsin molecule J (Yau et al., 1979), 100 times the thermal noise in each requires 2 x dimension, so spontaneous triggering is rare. A photon contains 4 X J, easily activating the rhodopsin (Yau et a l . , 1979). The activated rhodopsin causes hydrolysis of polyphosphatidylinositol in the plasma membrane to yield inositol triphosphate in the cytoplasm. The inositol triphosphate stimulates cGMP release, amplifying the signal. The cGMP binds to ion channels, causing their opening. In a dark-adapted cell, a single photon causes the opening of 1000 channels. The resulting depolarization stimulates neurotransmitter secretion, The neurotransmitter binds to the adjacent neuron and causes it to trigger. The initial stimulus produces a second messenger which starts an amplifying cascade to produce a larger stimulus to the transmission apparatus. Light also causes the cytoplasmic Ca” concentration in the rod cells to rise. This increase was previously thought to be part of the perception sequence. However, the release of CaZt from endoplasmic reticulum by inositol triphosphate (Payne and Fein, 1987) serves to desensitize the cell and is in fact the means of light adaptation; the gain of the cascade is attenuated at least 1000-foldfrom the dark-adapted state (Yau et al., 1986). The cytoplasmic Ca2+ concentration serves as a short-term memory of previous light levels, not as an amplificaton step of signal transduction. The transduction of sound in the ear is less well established, but is relevant because it occurs through a mechanotransducer, as must gravity perception. A modified form of the sound-detecting cells is used by animals to sense acceleration and maintain balance. In this vestibular system, a calcified sphere called a statolith or an otolith provides the mechanical stimulus in place of sound waves. In hearing, sound waves reaching the ear enter the cochlea, a remarkable fluid-filled resonance chamber. The cochlea acts as a spectrum analyser by establishing standing waves at different positions for each wavelength. This initial filtering greatly simplifies the amount of information which must be elicited from each sensing cell, since the location of the cell identifies the wavelength, and the intensity of the signal corresponds to the intensity of the sound at that wavelength. The hair cells of the cochlear epithelium have rod-like microtubule bundles which are displaced by the sound waves in the cochlear fluid. The displacement of the bundles is believed to open stretch-sensitive K+ channels in the epithelial membrane (Hudspeth, 1985). The influx of K+ causes cell depolarization, which in turn triggers neurotransmitter release into the synapse with the adjacent auditory neuron.

14

T. BJORKMAN

The cytoplasmic Ca2+concentration increases during stimulation of the hair cells, as it does in retinal rod cells. The sequence of ion movements has been described by Hudspeth (1985). Calcium ions enter from the extracellular pool through voltage-gated Ca2+ channels. These channels open during the depolarization caused by K+ influx. The Ca2+in turn open Ca-gated K+ channels to the inside of the epithelium where the K+ concentration is low, allowing the excess K+ to leave the cell and restore the membrane potential. In effect, the rise in Ca2+ serves to attenuate the effect of the stimulus. What parallels may be drawn between these sensory systems and gravity perception in plants? A plant gravity sensor perceives a mechanical stimulus as do the hair cells in both hearing and balance. The vestibular system uses the inertia of a large mass (a statolith) to detect acceleration. The stimulus is amplified to produce a large signal out of the cell. There is also a means for resetting the conditions in the cell when the stimulus ceases. The perceiving cells transmit a signal to other cells where the final response is elicited. Signalling of this type predominantly uses ion channels to create a rapid response (Methfessel and Sakmann, 1986). A difference between these sensors and plant gravity perception is that nerve cells do not have a plant analogue, so the nature of the transmitted signal will be different. Although the mechanotransducer in Dionea trigger-hair bases produces an action potential which may be considered analogous to nervous transmission, the evidence for an action potential being the elicited signal in gravity perception is weak. Also, evidence that cyclic nucleotides have a physiological role in higher plants is not strong, so amplification of the stimulus must occur by some other means. There is one clear commonality between the signals leaving gravity-perceiving cells and those discussed above, and that is polar transmembrane transport. 2. Calcium Ions in Transduction In the sensory transducers just described, Ca2+ play an ancillary role. There are also cases where the ion is central in the transduction chain. The intracellular action of Ca2+ in regulation is often mediated by Ca-binding proteins, whereby it acts as a second messenger. The best known of these proteins is calmodulin (CaM). Calmodulin, when activated by Ca2+,often regulates enzymes by stimulating their phosphorylation. In this manner, an amplifying cascade is established. Calcium+almodulin can also directly regulate ion transport proteins (Dieter, 1984). In the sensory transducers discussed above, Ca2+ attenuated the signal, but this action was not through CaM. The possible role of Ca-CaM as an amplifier in gravity perception also deserves discussion. Many enzyme reactions in plants are known to be under Ca-CaM control (Dieter, 1984). Typically, the cytosolic Ca2+concentration is below 1 PM. If

PERCEPTION OF GRAVITY BY PLANTS

15

the concentration is elevated beyond this by some stimulus, Ca2+ binds CaM. This complex can then bind other enzymes in a regulatory manner. Perception of light by phytochrome involves Ca-CaM (Roux et al., 1986). One example is the rotation of the chloroplast of Mougeotia (Wagner et al., 1984). Red light stimulates phytochrome in or near the plasma membrane, inducing a Ca2+influx from the extracellular medium. This Ca2+binds CaM, in turn stimulating a microfilament-associated protein, perhaps myosin light-chain kinase. The actomyosin then contracts differentially, according to the relative irradiance on phytochrome in the cell, to reorient the chloroplast. There is good biochemical evidence that Ca-CaM also regulates NAD kinase, NAD-quinate oxidoreductase and ion transport proteins (Dieter, 1984). The first two enzymes catalyse reactions unlikely to be part of gravity perception, though ion transport can be. The broad range of Ca-CaM-regulated reactions in animals suggests that there are many more reactions to be discovered in plants, and that these will have many essential functions. There is some evidence that CaM may be involved in gravitropism. The evidence is of two kinds: localization of CaM and effects of CaM inhibitors. Immunocytochemical labelling of CaM has identified particularly high concentrations of CaM in gravity-sensingcells of root caps and shoot apices (Roux and Dauwalder, 1985;Lin et al., 1986). In maize roots which require light induction to become positively gravitropic, the CaM activity and graviresponsivenessrise in parallel (Stinemetz and Evans, 1988). Calmodulin inhibitors block gravitropism (Bjorkman and Leopold, 1987b;Stinemetz and Evans, 1987). Furthermore, CaM inhibitors block the bioelectrical response of maize roots associated with gravity perception (Bjorkman and Leopold, 1987b). Signal amplification in perception could occur through Ca-CaM interaction. There is also evidence that gravity perception requires CaM activity. More rigorous experiments are necessary to reveal whether CaM has a direct role in signal amplification during gravity perception. 3. Phosphoinositides in Transduction Signal transduction can also involve phosphoinositides, which were discussed in relation to vision above. There is evidence for the phosphoinositide pathway being present in plants [for review see Poovaiah et al. (1987)l. The phosphoinositide pathway is generally initiated by a receptor in the plasma membrane which causes phosphatidylinositol bisphosphate to be cleaved into inositol triphosphate (IP,) and diacylglycerol by phospholipase C. The IP3 causes Ca2+ release from the endoplasmic reticulum with the consequences described in the discussion of CaM. Diacylglycerol stimulates protein kinase C, which can regulate enzymes by phosphorylation. Protein kinase C has been shown to turn on a class of ion channels which are

16

T. BJORKMAN

completely silent in the unstimulated cell (Strong et al., 1987). Through this pathway there are numerous ways to establish polar transport by a sensing cell, but each requires activation of some receptor in the plasma membrane. There are no experiments to date which have tested the involvement of phosphoinositides in gravity perception. Signal transduction is accomplished in biological systems in a number of ways. Calcium ions, which have been associated with many aspects of gravitropism, have different roles in different transduction apparatus. In the following sections, I will discuss ways gravity perception may work, using these common transduction pathways as guides.

B. MULTIPLE SYSTEMS

From an engineering standpoint the most efficient way to detect the direction of gravity is to use the displacement by an object which is attracted by gravity-something heavy. Sensors which work in this way are called statoliths. Efficiency and simplicity are two reasons why the statolith theory (Haberlandt, 1900) has been widely accepted for gravity sensing. There are, however, data which are difficult to reconcile with a statolith theory. In fungi, though gravisensing is poorly studied, it is known that the graviresponse is rapid and that there are no detectable sedimenting bodies (Burnett, 1976). The moss Physcornitrella responds to gravity, but no statoliths are apparent (Jenkins et al., 1986). In these, statoliths may have gone unobserved if they are in unexpected places. Bean seedlings provide an example where a simple observation could have failed to find statoliths because the gravisensitive region has been difficult to identify, Etiolated bean shoots perceive gravity using statoliths in the cotyledonary hook (Verbelen et al., 1985), but when de-etiolated they perceive gravity using statoliths along much of the hypocotyl (Heathcote, 1981); in both instances they curve 2-3 cm below the cotyledons. Nevertheless, gravity can apparently be perceived in the absence of statoliths. Nonstatolith gravity sensing has been proposed in higher plants also. Making amyloplasts too light to sediment by depleting the starch usually eliminates gravisensitivity (Haberlandt, 1902; Iversen, 1969), but there are exceptions. In starch-depleted, excised wheat coleoptiles a graviresponse is observable aft& 5 h (Pickard and Thimann, 1966). The lighter plastids may have taken longer to have an effect than the normal amyloplasts, which produced a response in excised coleoptiles within 2 h, or a slower alternative mechanism could have been responsible for the graviresponse. A starch-free Arabidopsis mutant (Caspar et a l . , 1985) appears to be a more interesting exception because it responds more quickly. However, changing the intensity of the gravistimulus produces a response fully consistent with the statolith theory (Sack and Kiss, 1988).

PERCEPTION OF GRAVITY BY PLANTS

17

In discussions of graviresponsive organs which apparently lack statoliths, the assumption is often made that the same gravisensing system is operating as in organs which do have statoliths (Pickard and Thimann, 1966; Moore and McClelen, 1985). This may not be a safe assumption. The presence of a second mechanism for gravitropism in addition to a statolith mechanism has been proposed (Shen-Miller and Hinchman, 1974); it would not be a novel case of a plant having more than one mechanism for getting an important task done. Gravitropism is essential for many organisms, and it would not be surprising if plants and other organisms are “overbuilt” for responding to gravity, as suggested by George Malacinski (personal communication), because failure to respond to gravity would often have lethal consequences. There may be a sensing mechanism which evolved before plastids which is potentially present in all cells. In organisms with specialized organs and diverse cellular constituents, an efficient statolith-based gravisensing system may operate. While amyloplasts act as statoliths in angiosperms, the green alga Cham uses membrane-bound barium sulphate crystals, and both invertebrates and vertebrates use crystalline organic calcium complexes as statoliths in their vestibular (balance) systems. Jellyfish contain calcium sulphate statoliths to detect their orientation relative to the gravitational field. The wide occurrence of statoliths is perhaps an example of convergent evolution. If there are parallel gravity-sensing systems present in an organism and the statolith sensor is disabled, the “primitive” system may take over. Characterizations of systems with disabled or absent statoliths have concentrated on the rate and extent of curvature, an indicator of how growth is controlled (Moore and McClelen, 1985; Caspar et al., 1985). However, there are no data which indicate whether the gravity-perceiving step has characteristics similar to those in the normal, statolith-containing plants. To learn something about the susception of gravity, the most interesting difference in apparent nonstatolith systems, susception (and perception) must be studied rather than growth. The parameters to measure would be the site of perception, sensitivity to small stimuli, the presentation time and whether the reciprocity rule (Johnson, 1965) holds. There are very few data on gravity sensing in organisms which appear never to use statoliths, so there is very little basis for guessing how they might perceive gravity. A critical survey of such mechanisms would be very welcome. Whatever the mechanism is, it must obey the physical laws discussed in this chapter. If an organism truly has no sedimenting intracellular component, the inclination would be to look for relatively large (>20 pm) structures outside the cell, or interactions between cells. If there are indeed parallel gravity-sensing mechanisms, it is important to determine to which mechanism a measured parameter applies. Data gathered from two gravity-sensing systems probably cannot be reconciled as consistent with a single mechanism.

18

T. BJORKMAN C. STATOLITH SENSORS

1. Identifying the Stutolith The most obvious candidates for sensing bodies are those which sediment. Sedimentation of amyloplasts is well documented, and has been the focus of studies on statolith sensors in higher plants. Most cellular components are unable to move freely, being secured by the cytoskeleton. Components which move by cytoplasmic streaming do so by attachment to the cytoskeleton (Wagner, 1979). In concentrating on amyloplasts, have we overlooked &her components of the cell which could act as statoliths in addition to, or instead of, amyloplasts?

Zntrurnernbrune statolith. If intercellular communication is accomplished by effecting a change in the membrane properties, the plasma membrane may be a good place to look for a sensing body. An intramembrane sensor would be a protein, most probably a transport protein capable of migrating to the bottom of the cell. If a membrane protein could act as a statolith, sedimentation could occur without microscopically detectable changes. Since proteins are more dense than lipids, they would move towards the bottom and not be buoyed to the upper side (Fig. 3a). The sedimentation rate for a protein in the membrane can be calculated from Stokes' Law, using a measured diffusion coefficient for proteins in the membrane (5 x m2 s-') (Schlessinger et ul., 1977) and a typical value for the density of a protein (1.33) and of the plasma membrane (1.03). The time for such a protein to settle half a cell diameter (5 pm) would be 88 years! Because equilibrium occurs in much less than 88 years, the sedimentation equilibrium is more appropriate to consider than sedimentation kinetics. By this method, we find the difference in concentration of a freely diffusing protein at the bottom of the cell (c) divided by the mean concentration (co):

Cleo = M ( l - @)gh/2RT = 6 x lo-'

where M = molecular weight (25 kg mol-'); i j = specific volume of prom3 kg-'); Q = density of plasma membrane (1.10 X lo3 kg tein (7 x m); m-3); g = gravity (9.8 m s - ~ ) ; h = half the height of the cell (5 X R = gas constant (8 J K-'mol-'); and T = temperature (300 K). Thus, if there were lo7 molecules of the relevant protein in the membrane there would only be one more molecule in the bottom half of the cell than in the top half at equilibrium. In comparison, a hair cell in the ear contains about 300 transducer channels (Hudspeth, 1985). Clearly gravity will not cause transport proteins to sediment in the plane of the membrane, so sedimentation of proteins will not influence gravitropism.

PERCEPTION OF GRAVITY BY PLANTS

19

Fig. 3. Hypothetical gravity susception and perception by proteins in the plasma membrane. (a) Dense transport proteins sediment to the lower part of the cell; unequal transport of the substrate creates a gradient. (b) Ion channels are pulled out of the plane of the membrane by gravity, and are active only in this position. This channel would be active only on the lower flank of the cell. Neither mechanism is energetically possible.

In addition to moving in the plane of the plasma membrane, a membrane protein can move across the membrane bilayer. A protein in a horizontal membrane would tend to be pulled out of the plane of the membrane by gravity. If that displacement were great enough to expose a catalytic site, on an ion channel for example, it could act as a gravity sensor (Fig. 3b). As described above, the force of gravity on a single protein is very small, whereas intrinsic membrane proteins are held in position by hydrophobic forces greater than gravity, the thermal motion greatly exceeds the effect of gravity, and the membrane potential also has a stronger influence on the position of the protein than gravity. Gravity therefore cannot be perceived by displacement of proteins across the plane of the membrane. Membrane components are clearly too small to be effective sedimenting bodies. To find a statolith, it will be necessary to look at larger components, inside the cell. Zntrucellulur stutoliths. To consider which intracellular components could act as statoliths, we will further consider the Stokes equation. The settling rate depends on the difference between the density of the particle and the density of the medium, the particle’s volume, and the effective viscosity of the medium. The density of amyloplasts is approximately 1.5 x lo3kg m-3; that of proteins, 1.3 X lo3kg m-3; of mitochondria, 1.2 x lo3kg m-3; and of Golgi bodies, 1.1X lo3kg m-3 (Audus, 1962).The cytoplasm is non-Newtonian, so the viscous drag is difficult to calculate. The effective viscosity and density of the cytoplasm would, however, act similarly on all organelles.

20

T. BJBRKMAN

The total work potentially done by a sedimenting particle is the force of gravity times half the cell diameter (for a 90” rotation). Because the cytoplasm must be displaced, the effective sedimenting mass is not density times volume, but density difference times volume. For any amyloplast, force = density difference x volume x gravity, = [(1.50 x lo3 kg m-’) - (1.03 x lo3 kg m-’)I x (1.4 x lo-’’ m’) x (9.8 m s-’) = 4.92 x 10-14 N Potential energy = force x distance, so where distance =

N) x energy = (4.9 x = 4.9 x 10-19 J

m,

m)

The gravitational potential energy in one sedimentable amyloplast is, therefore, 250 times the thermal noise (4kT = 2 x J), about 15 times the stimulus estimated from Fig. 2, and similar to the energy in a photon. A change in angle of 90”was used in this example but plants detect stimuli far smaller. To account for the observed sensitivity, a 90” rotation should generate a response well above the detection limit. The amount of energy involved in amyloplast sedimentation seems reasonable for its proposed role as a statolith. Mitochondria are smaller (0.5 pm) and less dense (1.2 g cm-’) than amyloplasts. They are unlikely to be sensors (Audus, 1962), losing only 1.1 x J (0.3kT) of potential energy in a 10 pm fall. Thermal agitation has a larger effect on mitochondria1motion than does gravity. If a perceiving structure were triggered by such a low energy, it would be triggered by thermal motion far more often than by gravistimulation, even with large movements. In fact, a mitochondrion would not sediment 10 pm within the presentation time: they have not been observed to sediment at all. Stokes’ law predicts that other organelles are too small or light to sediment detectably within the presentation time, and microscopy confirms this (Griffiths and Audus, 1964; Edwards and Pickard, 1987). Sedimenting amyloplasts have been the focus of most studies of the gravity sensor. In addition to the obvious sedimentation of amyloplasts, the great attention given to them is due to the presence of sedimenting amyloplasts at the site of gravisensitivity. Amyloplasts are plastids in which starch storage is the overriding function. Starch storage serves to maintain an energy source without the osmotic consequences of accumulating small molecules such as hexoses and sucrose. There are many instances of amyloplasts serving only for storage, such as in potato tuber cells and in bean parenchyma, but such amyloplasts maintain a fixed position in the cell (Verbelen et al., 1985). However, in the gravisensing cells of angiosperms,

PERCEPTION OF GRAVITY BY PLANTS

21

such as the node of grass stems, the root cap, and the gravisensing part of seedling shoots and coleoptiles, there are amyloplasts which sediment in response to gravity. If a maize root is decapped, the quiescent centre will rapidly produce amyloplasts, but gravisensitivity returns only after a change in the cytoplasm which allows them to sediment (Hillman and Wilkins, 1982). The amyloplasts in the gravisensing region are unlike amyloplasts elsewhere in that they are larger, multigranular, will not produce chlorophyll, and have more RNA (Gaynor and Galston, 1983). All gravisensing regions in higher plants normally have sedimenting amyloplasts, and amyloplasts sediment only in gravisensing cells. Rhizoids of the green alga Chara contain membrane-bound crystals of barium sulphate whose sedimentation produces curved growth (Sievers and Schroter, 1971; Schroder, 1904). These statoliths are able to move in a restricted zone in the growing tip of each rhizoid. When the rhizoid is turned from a vertical position, the statoliths sediment rapidly, and collect at the side wall. The requirements of a statolith are that it be massive enough for displacement to be detected against thermal noise and that it sediments rapidly. The barium sulphate-filled vesicles of the green alga Chara meet these requirements, and there may be other kinds in plants. However, in higher plants, the amyloplast is by far the best and apparently only reasonable candidate for a statolith.

2. Statolith Action When susception occurs by sedimentation of a large organelle, what part of sedimentation is actually perceived? There are several aspects of sedimentation which could be detected. These are the position ( x ) , displacement (dx), velocity (dxldt) and acceleration (dxldt2). The statolith balance systems in animals are optimized to detect acceleration. Plants, being stationary, do not need a bioaccelerometer to correct their travel as animals do. The other three aspects of sedimentation can be considered, however. In this section, how each of these aspects of statolith sedimentation might be the one which is perceived will be discussed. The concept of a pressure sensor is included with displacement, because a pressure transducer works by measuring the displacement of something with specified elastic behaviour. 3. Statolith Motion If statolith velocity is the parameter which is detected, perception would be due to an effect which depends on conditions varying with time. The statolith could affect the structural or the electrical conditions in the cell. Cytoskeletal shear. The change in gravitational potential energy when an amyloplast sediments was calculated above. How much of the potential

22

T. BJORKMAN

.. '.

Statolith

Mechanotransducer

Intracellular signal

Transmitter

Fig. 4. If a statolith is doing work on a transducer, it will sediment more slowly than predicted by the viscosity of the medium. The transducer could then modulate the cell's communication to other cells.

energy is available to stimulate the transducer? Some portion of the potential energy is dissipated as heat by the viscous drag of the cytoplasm, leaving the rest available to do work to trigger perception. An analogy is illustrated in Fig. 4.A ball falls through a viscous medium, and a string attached to it turns a shaft on a generator-the analogue of the transducer. The gravitational potential energy is converted to work which is converted to electrical energy in this example. The electrical energy then modulates the signal from the transmitter in proportion to the sedimentation rate. Knowing the mass and volume of the ball and the viscosity of the medium, one can calculate how fast the ball would be expected to fall if it were not attached to a transducer. If it is attached to the transducer, it will descend more slowly. Can this analogy be used to determine whether sedimenting amyloplasts may be doing work as they sediment? The observed rate of sedimentation of an amyloplast, 5-20 p m min-' (Sack et al., 1985a), reflects an apparent cytoplasmic viscosity of 350-1400 cp (water is 1 cp at 20°C) if the cell is considered analogous to a falling-ball viscometer. This viscosity is similar to that of glycerol. Is it appropriate for cytoplasm, or would an unrestrained particle fall faster? The rheological properties of the cytoplasm are believed to be determined largely by the cytoskeleton (MacLean-Fletcher and Pollard, 1980). Studies of the rheology of actin and microtubule solutions reveal a very peculiar behaviour. The shear force is essentially constant at all shear velocities; i .e. the apparent viscosity decreases in inverse proportion to the amount of stress (Buxbaum et al., 1987). An object moving rapidly through such a medium will tend to cause its viscosity to decrease, a phenomenon called shear thinning. For a fluid with dynamic viscosity characteristics like

PERCEPTION OF GRAVITY BY PLANTS

23

those of actin solutions, the behaviour of a falling ball viscometer is difficult to predict, but changes in viscosity would tend to be exaggerated (Rockwell et al., 1984). Therefore, even if the rheological nature of the cytoplasm were known, the expected sedimentation kinetics of amyloplasts could not be calculated accurately. Nevertheless, qualitative assessments can be made. Experiments on isolated cytosol with a falling-ball viscometer reveal an apparent viscosity of about 1 cp for the sol state, and >lo00 cp for the gel state (MacLean-Fletcher and Pollard, 1980). Since the viscometer used a metal ball of different size and density from an amyloplast, the shear forces are different and therefore the viscosity values are not quantitatively comparable. Nevertheless, these measurements indicate that amyloplasts in vivo sediment more slowly than isolated amyloplasts would in a sol state cytosol in vitro. The possibility remains that amyloplasts are restrained, and may thereby modulate the perceiving mechanism. An interesting observation is that the amyloplasts slow down as they fall. In corn root statocytes, the initial rate is 19 pm min-', but after moving about 2 pm, they have slowed to 4 p m min-' (Sack et al., 1985a). Shear thinning of the cytoplasm by Brownian motion of the amyloplast may cause the cytoplasm to have a lower effective viscosity in the region just around the amyloplast. The amyloplast would sediment faster in this thinned region, and slow down when it leaves this region (Fig. 5a). This may in part be the explanation for the observed decrease in the sedimentation rate during the early part of an amyloplast's fall. A region of thinned cytoplasm around the amyloplasts could indicate the gravity vector if it perturbed cell metabolism, which is dependent on the cytoskeletal structure. If this is the case, statolith motion in response to gravistimulation is not detected. Rather, Brownian motion would be used to create a signal which indicates the statolith position. Alternatively, the obscured slowing could be explained if amyloplasts, like chloroplasts (Witztum and Parthasarathy, 1985), are directly attached to the cytoskeletal matrix. During the initial part of the fall, the amyloplast would not be restrained, but after a short distance, the cytoskeletal elements would become taut and slow the descent of the amyloplast (Fig. 5b). Through this restraint, the motion of the amyloplasts could modulate a signal through localized perturbations as discussed above. Still another way to account for the slowing of the amyloplasts is compression of the cytoskeletal matrix as the amyloplasts settle on it (Fig. 5c). In Chara statoliths are held in position away from the tip of the rhizoid by a cytoskeletal matrix. If the rhizoid is treated with colcemid, the statoliths descend to the extreme apex (Friedrich and Hertel, 1973). If such a compression is involved in perception, displacement rather than motion is detected; this mechanism is discussed later.

24

T. BJORKMAN t=O

t =5s

Fig. 5. Models which could explain the slowing of sedimenting amyloplasts. The left column is immediately after gravistimulation, the right column is about five seconds later. (a) Shear caused by Brownian motion of the amyloplasts will reduce the effective viscosity of microfilaments in the cytoplasm. They will therefore sediment more rapidly through this layer than through the remaining matrix. (b) Hypothesized cytoskeletal connections to amyloplasts will become taut after some displacement. (c) Cytoskeletal elements may be compressed below the sedimenting amyloplast, and increase the effective viscosity.

It is difficult to make more than general suggestions about the significance of amyloplast sedimentation kinetics based on available data. These models could be tested by measuring the sedimentation kinetics when the statocytes are reinverted. A comparison of these kinetics with sedimentation kinetics of isolated amyloplasts through appropriate actin suspensions in vitro would make it easier to assess the likelihood of connections between amyloplasts and the cytoskeleton in detecting amyloplast sedimentation.

25

PERCEPTION OF GRAVITY BY PLANTS

Electricalfield. An amyloplast carries a charge (Sack and Leopold, 1982), so one can envisage its motion being detected from the changing electrical field. An electrical generator works through electrical and magnetic fields moving past a coil. Bandurski et al. (1985) proposed that the amyloplasts deform the electrical field around plasmadesmatal openings and thereby open these intercellular connections, if they are voltage-sensitive. The distortion of the field around the amyloplast would only extend as far as the Debye layer-about 2 nm (Starzak, 1984); even with electron microscopy, this distance is indistinguishable from actual contact, which does not occur (Perbal, 1978; Sack and Leopold, 1985). Also, the potential energy in the charge separation caused by the negatively charged amyloplast moving in the cytoplasm is much less than the potential energy due to gravity. This would be inefficient conversion of the potential energy to work in an instance where almost all the energy is required for sensing to occur within the presentation time. If none of the potential energy of the amyloplast is converted to work as it sediments, then the sensing must occur when the amyloplast contacts the bottom flank of the cell. The energy available to do work on a sensor by this scheme is the kinetic energy in the moving amyloplast. That energy is also easily calculated using some of the values cited above: kinetic energy depends on the mass and the velocity. If mass = 2 x

kg

and velocity = 20 pm min-' = 3.3 x 1 0 - ~ m

s-l

then kinetic energy = 1/2 x mass x (velocity)' = (0.5) x (2 x kg) x (3.3 x = 1.1 x 10-27 J

m s-')

The kinetic energy in a moving amyloplast is four million times less than the thermal noise. This comparison shows that amyloplasts sediment so slowly that there is simply not enough kinetic energy for the motion of an amyloplast to be detected. Although physiological models to detect statolith motion can be described, energetic evaluations rule them out. Perception based on statolith motion would also fail to explain the continued differential growth after sedimentation is complete until the tissue is again in its preferred orientation. Perception must come about through detection of some parameter of sedimentation other than motion of statoliths.

26

T. BJORKMAN

4. Statolith displacement An obvious event in gravity-sensing cells is the movement of statoliths from one position in the cell towards another when the tissue is displaced. The indicator of orientation in the gravitational field could be the change in statolith position. Specifically, the distance statoliths move from their normal (e.g. tissue vertical) position, or the amount some structure is displaced by statoliths, could be the graded stimulus which elicits a response. A statolith exerts a force on any structure with which it comes into contact. The pressure is detected by displacement of an elastic structure which responds in proportion to the amount of displacement. A pressure sensor is therefore a special case of displacement perception. The effect of thermal noise on a receptor can be reduced if the receptor averages the input over a period of time. If statolith sedimentation does work on a receptor by displacing it, the displacement required for triggering should be high enough that the triggering by gravity occurs much more frequently than triggering by thermal agitation of the statolith. This can be accomplished by averaging the signal over time. Although the probability of a large thermal displacement increases as the observation period increases, the effect of gravity increases more during statolith sedimentation. This can

Fig. 6. The effect of integration time on the ability to distinguish the effect of gravity on a statolith from the effect of thermal noise. The left axis is statolith motion due to each force. The right axis is the amount of work done by the corresponding displacement of an amyloplast. Indicated along the right axis is the work required for activation of sensors on vision [rhodopsin (Yau ef al., 1979)] and hearing [hair cell (Corey and Hudspeth, 1983)] as well as thermal noise ( M )Values . used in this graph were estimated from data in Sack etal. (1985a): r] = 300 cp = 0.3 Pa, r = 1 pm, sedimentation rate = 4 p m min-'. The uncertainty of each line is about three-fold due to imprecision in the estimation of one of the parameters and assumptions about the behaviour of the cytoplasm.

PERCEPTION OF GRAVITY BY PLANTS

27

be calculated from Einstein’s equation of Brownian motion (Einstein, 1907):

a

where is the net distance moved, z is the integration time, 7 is the viscosity of the medium and r the radius of the particle. The relative effects of sedimentation due to gravity and random thermal motion can be seen in Fig. 6. An integration of several seconds is needed for sedimentation to be the dominant source of the signal, and for work done by sedimentation to be equivalent to that which triggers other sensitive sensors. Evidence of such integration is that gravistimulation need not be imposed continuously. If short intermittent stimulations are frequent enough, they have the same effect as the same amount of stimulation given continuously (Pickard, 1973b). One-second stimulations every five seconds are summed; halfsecond stimulations must be repeated every second to be summed. Significantly, the smaller stimuli are “remembered” for a shorter time. The perception mechanism probably averages stimuli over a time period of one or a few seconds, and the minimum stimulus involves a displacement of at least 100 nm.

Cytoskeleton stretching. The lack of contact between amyloplasts and the plasma membrane (Witztum and Parthasarathy, 1985; Heathcote, 1981) suggests that an indirect interaction causes a signal to be passed out of the cell. This indirect interaction could be through cytoskeletal members attached to amyloplasts transmitting, by tension, energy to receptors in a membrane. In the extreme case, where the amyloplasts are completely restrained, they would do no work at all. Amyloplasts are often restrained to this extent in cells which do not function as gravity sensors. The cytoskeleton is certainly a promising agent for transmitting the stimulus, but not by immobilizing the amyloplasts, although they would be partially restrained. This concept has been raised by Larsen (1969) who proposed that the amyloplasts function as pendula attached to the distal part of the statocyte by the cytoskeleton. Stimulus transmission by the cytoskeleton has also been proposed by Shen-Miller and Hinchman (1974) and by Friedrich and Hertel (1973). If the cytoskeleton is attached to a stretch-sensitive Ca2+channel in the amyloplast envelope, displacement of amyloplasts could cause tension in the stationary cytoskeleton. There is a precedent for microfilament-plastid interaction (Witztum and Pathasarathy, 1985). The tension could cause a conformational change in the channel to which the cytoskeleton is attached, analogous to the way the membrane potential causes a conformational change in voltage-gated channels, by exerting a force on dipoles in the protein. This conformational change decreases the amount of energy needed

28

T. BJORKMAN

to open the channel (Honig et al., 1986). Amyloplasts contain large amounts of Ca2+(Chandra et al., 1982). More frequent opening of a Ca2+channel in the amyloplast membrane would cause a locally elevated Ca2+concentration. This region of higher Ca2+would be a directional signal when amyloplasts are only in the lower part of the cell. A Ca2+channel in the amyloplast membrane, attached to microfilaments anchored at the end of the cell where the amyloplasts are in vertical tissues, would cause a rise in Ca2+in the lower part of the cell. A localized release of Ca2+would result in an elevated Ca2+ concentration only in a restricted region of the cytoplasm (Keith et al., 1985; Brownlee and Wood, 1986; Weir et al., 1987). Should such channels be at the other end of the cytoskeletal connection, in the plasma membrane, sedimenting amyloplasts would not create a directional signal. In that case the local rise in intracellular Ca2+ would occur in the same location in the cell regardless of the direction of stimulation. A locally higher Ca" concentration around displaced amyloplasts could stimulate ion transport across the plasma membrane in that region of the cell. Elevated cytoplasmic Ca2+ may stimulate ion transport proteins directly or through CaM (Dieter, 1984). The directional ion transport is the type of signal which the perceiving cell would be expected to elicit. If sedimenting amyloplasts do work on an ion channel via microfilaments or microtubules, they can elicit a physiological asymmetry by stimulating channels in only one part of the cell. The operation of this type of microfilament-membrane channel interaction has been found (Horwitz et al., 1986) and is likely to be common (Geiger, 1985). Lawton et al. (1986) noted disruption of microtubules around amyloplasts at the onset of sedimentation. The interaction of the cytoskeleton with channels is a promising candidate for perceiving mechanical stimuli.

Displacement of endoplasmic reticulum. Volkmann and Sievers (1979) have proposed that perception is by the interaction of amyloplasts with the endoplasmic reticulum. In that case, amyloplasts would begin to act only after reaching the endoplasmic reticulum at the lower surface of the cell. The presentation time could then be interpreted as including the time required for the amyloplasts to reach a position where they had an effect. Amyloplasts sediment rapidly, as fast as 40 p m min-' in Taraxacum stalks (Clifford and Barclay, 1980). Sack et al. (1985a,b) measured both the presentation time and the sedimentation of amyloplasts in two different kinds of statocytes-the root cap and the coleoptile of Zea mays. In both cases, the amyloplasts sedimented rapidly enough that the first ones had reached the new lower flank of the cell within the presentation time. Sedimentation of amyloplasts thus occurs in an appropriate time period for perception to occur through interaction with the endoplasmic reticulum. The motion of amyloplasts at the new lower flank of the cell also bears

PERCEPTION OF GRAVITY BY PLANTS

29

on this possibility. Heathcote (1981) observed that amyloplasts in Phaseolus hypocotyls slow down when they are about 1 p m from the lower wall of the cell. The slowing may be caused by amyloplasts deforming the endoplasmic reticulum which underlies the plasma membrane. Heathcote makes the comment that during sedimentation, the amyloplasts “appeared to be slowed by invisible cytoplasmic structures”. Observations of live tissue, like these and also those of Sack (Sack et al., 1985a,b; Sack and Leopold, 1985), are very helpful when considering the interactions of statoliths with the perception mechanism. Volkmann’s observation that graviresponse is proportional to pressure prompted him to propose that the endoplasmic reticulum senses pressure directly (Volkmann, 1974). If the endoplasmic reticulum is deformed elastically, Hooke’s law holds that displacement will be proportional to pressure. Hence Volkmann’s data support displacement detection equally well. Sievers et al. (1984) propose that the signal which is elicited when amyloplasts settle on the endoplasmic reticulum is intracellular Ca2+ released from the endoplasmic reticulum, a storage site for CaZf in the cell. In their model, statocytes depolarize the cell as a result of Ca2+ release when amyloplasts deform the distal beds of endoplasmic reticulum. The proposed involvement of a specific endoplasmic reticulum structure is supported by the observation that a pea mutant in which the endoplasmic reticulum is uniformly distributed is not graviresponsive (Olsen and Iversen, 1980). The endoplasmic reticulum is arranged so that maximum amyloplast contact is with beds on the outer side of the cell (Juniper and French, 1970), explaining the depolarization only of cells on the lower side of the tissue (Behrens et al., 1982). The product of this Caz+-induced depolarization of the lower cells is an electrical asymmetry across the root cap. The initial perceiving step remains difficult to explain, namely how amyloplasts cause Ca2+ release from the endoplasmic reticulum. Amyloplasts deform the membrane extensively, bringing the endoplasmic reticulum cisternae into contact (Volkmann and Sievers, 1979). Perhaps this intermembrane interaction can induce Ca2+ release. The question of how amyloplast action elicits a molecular response is unanswered for this specific model, but the model provides a good tool to answer this question.

5. Statolith Position Finally, the position of a statolith as the detected aspect of sedimentation will be considered. The argument that statolith position can be detected seems to go against the preceding discussion of thermal noise and of work done during sedimentation, but those restrictions still apply though in a somewhat different way. To detect the position of a statolith there would be a sensitive area on the lower surface of the cell which recognizes the statoliths on some basis other than mass. The statolith would still need to be

30

T. BJORKMAN

massive enough to sediment to the bottom surface without being extensively agitated by Brownian motion. Also, a high-affinity recognition site could use electrical or chemical potential energy rather than gravitational for the recognition (Volkmann, 1974), but the gravitational force would have to be large enough to break the attraction when the tissue is reoriented. The specific recognition site would be near the outer flank of the cell. Although there is no contact between the amyloplast and the plasma membrane (Perbal, 1978; Heathcote, 1981), there are several structures just inside the plasma membrane with which the amyloplast may interact: one or more layers of endoplasmic reticulum (Juniper and French, 1970; Volkmann, 1974; Sievers and Hensel, 1982) held in place by microfilaments (Hensel, 1984,1985, 1986b); highly stable cortical microtubules (Kakimoto and Shibaoka, 1986); a meshwork of microfilaments (Parthasarathy, 1985) not involved with cytoplasmic streaming (Derksen et al., 1986); desmotubules which extend into the cytoplasm. There are many structures of importance with which statoliths may interact, but it is unclear which are involved in gravity perception. Electrostatic attraction. If electrical potential energy is used to trigger perception, it could be through electrostatic attraction between the charged amyloplast membrane (Sack and Leopold, 1982) and charged sites on the sensitive surface. The electrical field created by the surface charge of a membrane decays rapidly away from the membrane surface, being negligible more than 1 or 2 nm from the surface (Starzak, 1984). An electrostatic interaction is not a long-distance one on a cellular scale, requiring an approach microscopically indistinguishable from contact. Without postulating characteristics of the binding site, it is impossible to calculate whether the energetics of electrostatic binding are reasonable, and there is not enough information for productive speculation about the properties of such a binding site. One generalization which can be applied is that an electrostatic interaction must be strong enough to elicit a reaction, but weak enough to allow the gravitational force on the statolith to move it away if the tissue is reoriented. There is evidence for electrostatic effects of gravistimulation. In maize roots, the surface charge of the plasma membrane changes (Pilet, 1985). This change may affect the transport properties of the charged membrane through screening of substrates and through electrostatic effects on transport proteins within the membrane (Mdler and Lundborg, 1985). On the other hand, the plasma membrane surface charge could be altered by changes in the extracellular Ca2+ activity (Moller and Lundborg, 1985), which also occurs on gravistimulation (Lee et al., 1983). Ligand binding. The location of statoliths in a sensing cell may also be detected by releasing chemical potential energy if exothermic binding occurs. The energy to trigger a physiological change would be released by a

PERCEPTION OF GRAVITY BY PLANTS

31

reaction such as ligand binding. As a simplistic example, if a ligand on the amyloplast envelope binds to a receptor on the endoplasmic reticulum, an associated ion channel could be caused to open. Cytological evidence suggests that amyloplasts approach the endoplasmic reticulum more closely than they do the plasma membrane (Perbal, 1978). Ligand-activated channels are a common type (Hille, 1984); though ligands are usually small molecules which diffuse quickly in the cytoplasm, none have been found attached to a large organelle. One difficulty in using ligand binding as a perception mechanism in gravity sensing, as with electrostatic binding, is that the binding energy would make it hard for the amyloplasts to come loose again. If the binding is exothermic and the contribution of gravity is energetically negligible, the force of gravity on the amyloplast will be insufficient to release it from the binding site. A direct role of statolith position in altering growth to produce curvature has been proposed in Chum rhizoids (Sievers and Schroter, 1971). An important distinction must be made between this alga and angiosperms, however. Sensing and response occur in the same cell, with the statoliths sedimenting to the exact position in the cell where active growth is occurring. Thus no transmission step is necessary, and perception is presumably simpler. Sievers and Schroter (1971) suggested that the role of the statoliths is to prevent Golgi vesicles from fusing with the plasma membrane and thereby preventing wall growth and membrane expansion in that region of the cell. This contention is strengthened by the observation that growth at the cell apex of vertical roots stops if the statoliths are caused to settle there due to reduced turgor (Sievers and Schroter, 1971), or by treatment with the anti-microtubule agent colcemid (Friedrich and Hertel, 1973). This very straightforward action of statolith position works well in a tip-growing cell, but is difficult to apply to multicellular responses. A statolith acting in this manner must be large in order to remain at the lower flank of the cell despite thermal agitation. Specifically, as described in Section 11, it must take about 3x J to move the statolith away in order for thermal displacement to be insignificant, yet for a change in the gravity vector to move the statolith. Specific chemical or electrostatic interactions are intriguing possibilities. A combination of these attractions with displacement due to gravity would form the basis of such an interaction. There are very few data which are useful in evaluating amyloplast binding, although those which suggest that amyloplast position is important support the possibility indirectly. Based on microscopic evidence, any such direct interaction would most likely be with endoplasmic reticulum or cytoskeletal elements, rather than the plasma membrane or its components. The identity of the relevant statolith action can thus be limited somewhat. The kinetic energy of statolith motion is too small to be perceived. Statolith position appears to act directly in the case of Cham rhizoids. In multicellular tissues, displacement is much more likely. The total displacement of a

32

T. BJORKMAN

statolith on gravistimulation may not be relevant; rather, it may be the displacement of some transducer, occurring only during a part of sedimentation. The relevant physical displacement may appear as if statolith position were detected, as in the model in Lepidium root caps involving endoplasmic reticulum beds. The larger response to a sliding action across a cell flank than to just sedimentation to it (Iversen and Larsen, 1971) implies that a signal is elicited by deformation of a structure on the lower cell flank. Just as there are several types of biological statoliths, there may be more than one way to perceive their action. For multicellular plants, gravity perception by statoliths is likely to be through the work done by statolith displacement.

D. NONSTATOLITH PERCEPTION

The preceding discussion of statolith action involves fairly straightforward principles and can refer to a large body of published work because the statolith theory has dominated research in gravitropism. Gravity sensing without statoliths would necessarily be more subtle. There is nevertheless strong evidence that it does occur and the question of how must be addressed. The ability of organisms to detect subtle signals easily exceeds our ability to explain it. A remarkable example is the marine mollusc Tritionia which, without any ferromagnetic particles, detects not only the earth’s magnetic field but also the phase of the moon, from the bottom of Puget Sound (Lohmann and Willows, 1987)! 1 . Pressure Differential As an alternative to a sedimenting body, Pickard and Thimann (1966) have proposed that the weight of a cell’s protoplasm stimulates the sensor by exerting pressure on the side of the cell towards gravity. The rationale for this hypothesis is that the protoplasm is more massive than any substituent of the cell and therefore can exert more force on a sensor. A discussion of this mechanism may be found in Audus (1979). The resultant change in pressure across the cell plasma membrane and wall, higher on the lower side, and lower on the upper side, would be instantaneous. This pressure difference can be calculated; it is the density of the cytoplasm times gravity times the diameter of the cell. Thus, for a cell 10 p m in diameter, the pressure will be 0.1 Pa higher at the bottom than at the top. For a large stimulation (90’ rotation), the pressure change at the new lower side would be 5 X Pa. Plants respond to stimuli that are at least 100 times smaller. The pressure change would have to be detected against the background turgor pressure, which is typically 5-15 X 10’ Pa. To complicate matters further, the turgor pressure is not static, but is constantly changing with the evaporative demand on the plant and, to a smaller extent, with fluctuations in solute exchange in and out of the cell. In pea stems held in a uniform, humid

PERCEPTION OF GRAVITY BY PLANTS

33

environment, the turgor pressure varied over a range of 5 X lo4 Pa in the period of 1 min (Cosgrove and Steudle, 1981; Cosgrove and Cleland, 1983). To detect a change in pressure due to gravity against a static background pressure at least ten million times larger, and a rapid fluctuation in that pressure one million times larger, would require an amplifier which would selectively amplify the signal to overcome the noise with only a few seconds of sampling time. Such an amplifier would be unprecedented in both biology and engineering. 2. Membrane Tension The differential volume change resulting from changes in the osmotic potential inside or outside the cell result in changes in the plasma membrane tension (tangential vector). Because the elastic modulus of the membrane is high, the tension is much more sensitive than the turgor to changes in volume. For reasons very similar to those responsible for the difference in protoplasmic pressure, gravity would cause a difference in membrane tension between the upper and lower surface of the cell. Is there a way this difference in tension could have metabolic consequences leading to cell polarization? Guharay and Sachs (1984) have discovered a tension-sensitive ion channel which would respond to the tension changes resulting from a change in volume. It represents an attractive candidate for the elusive turgor-sensing mechanism. A direct effect of pressure has been proposed as a turgor pressure sensor (Coster and Zimmermann, 1976), which might be relevant to gravity sensing, but the direct effect of pressure on membrane permeability is relevant only at pressures of lo8 to lo9 Pa (Aldridge and Bruner, 1985). A change in turgor always involves a change in cell surface area because cell walls are elastic. This surface area change causes large changes in the membrane tension (Wolfe and Steponkus, 1981) which, through a tension-sensitive channel, may be the basis for turgor sensing. Although the cell surface area would not change with reorientation in a gravitational field, there would be a differential membrane tension between the top and bottom of the cell. This tension differential has been proposed as a gravity sensor (Edwards and Pickard, 1987). Could the energy from the difference in tension be sufficient to overcome the activation energy for changing the state of the channel? The tension-sensitive ion channel in patch-clamped tobacco protoplast membranes (Falke et al., 1986) opens at a pressure difference of 2500 Pa. If the radius of the patch pipette is 0.5 pm and there is one channel per patch, then the tension for opening and activation energy can be calculated: Tension = 1/2(pressure difference) X (radius of curvature) = 1/2(2.5 x lo3Pa)(5 x lo-’ m) = 625 p N m-l

34

T. BJORKMAN

Energy = (tension) = (6.25 x =

X

(area of patch) N m-') x 1/2(4 n)(5x

m)'

J

The activation energy for channel opening is about lo-'' J per channel, occurring at a tension of 625 pN m-'. For comparison, the resting tension of a plant cell membrane is about 100 p N m-l and the critical tension for lysis is 4000 p N m-' (Wolfe and Steponkus, 1981). The difference in tension due to gravity, based on a turgor difference of 0.05 Pa and a cell radius of 5 p m , is about 0.25 pN m-'. Even if the energy in the differential tension caused by gravity across a whole cell's membrane ( W . mirabilis L (r, 1 4 5 )

AH'"

CO, Light

Recycling' 613C

Dark" Releaseb Uptakeb %

6H+

%d

%

Conditions

Reference

PAR (pmol m-'s-l)

0 9 15 24

40 42 35 27

+

+ + +

118 48 45 30

-25.3 -24.3 -24.2 -20.5

46 26 70 (46 24)

35 24 0

24

830 220 30

-24.7

+

6D

200 shade 401 Epiphytic 400 100 sun

Winter et al. (1983)

Ting and Burk (1983)

100

70

Field

-18.5

to

von Willert et al. (1982)

-19.5

A izoaceae Mesem bryanthemum crystallinum (r, 3 ) Cactaceae Pereskia aculeata (r, 2)

L

L

0

0

448 552

150 165

78 101

20 24

0

-

0

-

2708 664 1164

207 87

0 33 149 40 222

-30 -24 -26 to - 16

100 78 100 101 -27.@

-446

+H,O +400 mM NaCl NaCl+ H 2 0 Field (wet + dry)

Winter (1974) Winter and Liittge

+H,O -H20

Rayder and Ting

(1976)

Winter et al. (1978)

(1981)

Sternberg et al. (1984b) continued

TABLE 111-Continued C 0 2 Light succuFamilyandspecies" lence"

AH'O

Dark" Releaseb Uptakeb %

Cactaceae (continued) P. grandifiora (r, 2 )

L

35 50

Clusiaceae Clzuia rosea (r, 2)

L

18 60

L

57 62 202 350 0 73 73 94 76 855 330

13 32 0 175 0 0 22 38 25 150 120

39

11 0

Crassulaceae Kalanchoe unifIora (r,2,5> Sedum acre ( r , 3 ) Sedum telephium (3)

Sempervivum montanum (r, 3) Gesneriaceae Condonanthe crassifolia

(r,112,5)

L L

L

L

Recycling'

co2

44

0 0

-

-

59 11

100 100

6H+

613C Omd

-

+ -

+ + + + -

-

840 308 43 35 10 0 15 1042 310 43 4

%

(31) (0) 202 0 0 73 29 18 26 355 90

Conditions

Reference

+HzO -H20

35 50 -17.9

(54) (0) 100 0 0 loo 40 20 44 65 27

6D

I

gn: : :

Epiphytic

'"'"}Epiphytic -H,O -27.5 -27.0

+HiOSpring - H 2 0 Summer +H20 :C3

-25.06

+HZO: C&AM -H20: C3-CAM -H20: CAM -H20: CAM 300 p n o l m - 2 s 900 PAR

I

- 18.0

-24*

+13'

+H20}Epiphytic -HzO

Ting et al. (1985a)

Schafer and Luttge (1986) Schuber and Kluge (1981) Lee and Griffiths (1987) Griffiths (unpublished) Wagner and Larcher (1981) Guralnick et al. (1986) 'Tingetal. (1985a)

Piperacaceae Peperomia camptotricha (I,

112,5)

Portulaceae Portulacaria afra (I,

1/29 5 )

36 55 93 80 60 220 36 200

7 48 5 0

180 50 56 1 0

15 9 7

26 46 69

60

Young -27.7'

+23'

-27.5' -17.5'

+H,O - H 2 0 1week -H,O 2 weeks

-10'

August February April

Sipes and Ting (1985) Sternberg et al. (1984a)-

'

Guralnick et al. (1984) Sternberg et al. (1984b) Mooney et al. (1977) Martin and Zee (1983)

'

' Talinum calycinum (r, 3 )

Vitaceae cissus quadrangularis (r, 112,5) Bromeliaceae Guzmania monostachia ( I ) Nidularium innocenti (r, 3 )

Liliaceae Yucca gloriosa (r, 3 ) For notes, see Table 11.

400 180

0 104

1020 37

33 78

0 78

70 89 15 17

2 7

210

0

100 400 0 0

-27.8

+H,O -H,O 5 days

121 0

loo (50)

-25.3 -19.4 -17.8

Leaf Stem Old stem

Ting et al. (1983)

33

94

66

18

3

-26.76 -26.5' -31.5' -24'

Sun Sun Shaded 10°C, 16 h dark period

'Smith et al. (1986a) 'Smith et al. (1985)

-22.0

Field

Martin et al. (1982)

2

100 210

Griffithset al. (1986)

McWilliams (1970)

'Medina et al. (1977)

74

H. GRIFFITHS

Griffiths, 1987), which is in common with many other temperate succulent families such as the Sempervivoideae (Osmond et al., 1975; Wagner and Larcher, 1981; Earnshaw et al., 1985) and Portulacaceae (Martin and Zee, 1983; see also Ting, 1985). There also seems to be an example of convergent evolution in two families of predominantly neotropical distribution, the Piperaceae and Gesneriaceae (Sipes and Ting, 1985; Ting et al., 1985a; Guralnick et al., 1986). Many species in both families are epiphytic [up to 20% in the Piperaceae (Ting et al., 1985a)], and have similar leaf structure and distribution of WSP. Additionally, both families display a variant of CAM in which recycling of respiratory COz predominates [CAM cycling (Ting et al., 1985a; Guralnick et al., 1986; Table III)]. The C,-CAM transition has also been described for deciduous climbers in the Vitaceae [e.g. Cissus trifoliata (Olivares et al., 1984); C . quadrangularis (Ting et al., 1983; Virzo de Santo et al., 1983)], whereby both stems and leaves may show a varying degree of CAM depending on water status. Finally, a most interesting recent development has been the identification of CAM in hemi-epiphytic stranglers in the genus Clusia (Tinoco-Ojanguren and Vazquez-Yanes, 1983; Ting et al., 1985b). This is the first example of CAM in a woody dicotyledenous tree, and the regulation of organic acid metabolism (i.e. malic and citric acid) is currently receiving more attention (Luttge, 1988; Popp and Liittge, personal communication; Section 1V.C).

C. RESPIRATORY COZ RECYCLING BY C3-CAM INTERMEDIATES

Having demonstrated that the magnitude of AH' in constitutive CAM can be related to the extent of respiratory COz recycling under a range of environmental conditions, a similar analysis is made in this section for those plants usually identified as C,-CAM intermediates (Table 111;see also Table 11). The plasticity of the CAM response in these plants may entail considerable daytime COz uptake, without any ready separation into the light period phases of constitutive CAM. Data on COz uptake in the light have been integrated and are presented as the magnitude of net uptake and/or release (Table 111). Recycling, where possible, has been calculated as described previously (Section II.G, Table 11), and values of 613C and 6D have been included where applicable (see Section 1I.F). Although it was not possible to calculate recycling for Pyrrosia confluens, there is clearly considerable plasticity of response by sun and shade populations (Winter et al., 1983; Table 111). Daytime C 0 2 uptake was related to PAR, but the concomitant dark C 0 2 uptake in each of the treatments may not have been directly related to AH', indicating a contribution from recycling (Table 111). For the controversial Welwitschia, two contrasting studies have suggested that respiratory COz recycling is the major feature of

CRASSULACEAN ACID METABOLISM

75

any CAM-like adaptation. In laboratory studies AH' and AC02 were both measurable (Ting and Burke, 1983), but in the field a possible (malate citrate) fluctuation was accompanied by C 0 2 uptake only in daytime (von Willert et al., 1982; von Willert, 1985). Despite the many publications featuring M. crystallinum, it was with difficulty that comparable AH+ and ACOz could be obtained, by applying correction factors calculated from Winter and Luttge (1976) to data published by Winter (1974) for similarly treated plants (Table 111). In the C3 mode, H' does not fluctuate, and there is marked daytime C 0 2 uptake. When the plants are grown with 400 mol mP3 NaCl, there is a gradual decrease in daytime COz uptake (Winter, 1974; Winter and Luttge 1976) and increases in AH' and ACOz, with 33% recycling. On returning to control (salt-free) solution, CAM is retained with an increase in AH', ACOz and daytime C 0 2 uptake; recycling also increased in both percentage and absolute terms (Table 111). During CAM induction, 613C of new plant material shifts from -30 to -24%,, but studies in the field have perhaps shown that the CAM transition is more important, with Bplantvarying from -26 to -16%, as drought stress is prolonged (Winter et al., 1978). However, the respiratory C 0 2 contribution to AH' has not been measured for field-grown populations. In the Portulacaceae, there are also several species which have been described as inducible CAM plants (Ting and Hanscom, 1977; Guralnick et al., 1984a,b). However, as pointed out by Winter (1985), this does not represent true CAM induction but a changing proportion of light :dark C 0 2 uptake, with recycling probably contributing to the AH' (Guralnick et al., 1984a,b; Table 111). The characterization of CAM in Pereskia spp. has been made by Rayder and Ting (1981) and Diaz and Medina (1984). During progressive drought stress, AH' increases in field-grown plants of Pereskia guamacho (Diaz and Medina, 1984), and in laboratory experiments on two other species there was no net C 0 2 uptake at night (i.e. 100% recycling) by P. aculeata and P . grandiflora (see also Nobel and Hartsock, 1986). This is a unique response, since in all other C3-CAM intermediates examined, there is some net C 0 2 uptake at night under certain conditions (see Table 111). One recurrent theme in Table I11 is the number of epiphytes which are C3-CAM intermediates: Kalanchoe uniflora induces CAM in response to drought stress, although recycling ceases (Schafer and Luttge, 1986). Thus 50% of the AH' is derived from respiratory C 0 2 before stress, but then the more usual generic characteristics of Kalanchoe, with little or no recycling, are shown as plant water status decreases (Table 111; see also Table 11). This is in contrast to many of the other epiphytes and C3-CAM intermediates in general [with the exception of S. acre (Schuber and Kluge, 1981)], although rates of recycling could not be properly calculated for any of the Piperaceae or Gesneriacae (Table 111). In the Bromeliaceae, Guzmania monostachia

+

76

H. GRIFFITHS

(subfamily Tillandsioideae) was originally thought to be a classical inducible CAM plant (Medina et a f . , 1977), but field measurements showed little uptake of C 0 2during the dark period (corresponding to 94% recycling), and uptake of COz early in the light period (Griffiths et a f . , 1986; Table 111). Other field studies did correlate AH' with 613C (Smith et al., 1985; Table 111), and more work is required to fully evaluate the CAM status of this plant. Using combined data of McWilliams (1970) and Medina et al. (1977), it appears that Nidularium innocenti (subfamily Bromelioideae) is also a C,-CAM intermediate, with a small AC02 and AH' being accompanied by a C3-like 613C ratio (Table 111). Using similar criteria, Yucca gforiosa also appears to be a C3-CAM intermediate (Martin et a f . ,1981; Table 111). Sedum spp. and Sempervivum montanum also show a range of recycling characteristics, with recycling decreasing under high PAR in S. montanum (Wagner and Larcher, 1981). However, there is a decrease with stress in S. acre (Schuber and Kluge, 1981), and a whole range of values depending on the stage of CAM induction in S . telephium (Lee and Griffiths, 1987; Table 111). The nature of the regulation of +CAM transitions is discussed in more detail in the following section. Finally, it is worthwhile re-emphasizing that many of these C3-CAM intermediates seem to be related to the type or degree of succulence (see Section 1I.B). It is essential that future work takes consideration of succulence, both in terms of variation in plant water status during experimental procedures, and in terms of the respiratory C 0 2 recycling characteristics found in different families. D. PHYSIOLOGICAL CHARACTERISTICS OF THE C&AM

TRANSITION

In view of the phenotypic variation found in all CAM plants, there may perhaps be a simple explanation for the apparent diversity of C3-CAM intermediates, as exemplified by the proliferation of terminologies. Many studies have not used a sufficient range of environmental variables over a long enough period of plant growth and development to account for the possible range of CAM characteristics. There have only been a few integrated measurements of the full C,-CAM transition, collated in Table 111, and it is worthwhile highlighting the complete studies on Sedum acre (Kluge, 1977; Schuber and Kluge, 1981), Portulacaria afra (Guralnick etaf.,1984a,b; Guralnick and Ting, 1986) and M . crystaffinum(Winter, 1985). Figure 2 illustrates the range of physiological characteristics which occur during the complete +CAM transition by Sedum tefephium (Lee and Griffiths, 1987; Table 111). Plants were grown with an intermittent water supply so as to create C3-CAM intermediates; one group was then continuously watered, while the water supply was withdrawn from the other group. Measurements were made after 24 h and then again at 5-day intervals, and included C 0 2 uptake, AH' and xylem sap tension (pressure chamber, data in MPa).

77

CRASSULACEAN ACID METABOLISM

a 100 0 40 Dusk xylem 0.36 sap tension 0 0.50 (MPa) Recycling %

-

n

I

x

12

0

a

20 0.29

0

0

0.50

O0.50

12

12

0

44 a0.26

I

I

I

I

24

a 0

24

12

12

I

24

12

Fig. 2. The transition between C, and CAM by Sedum telephium in response to water supply. C3-CAMintermediate plants were watered (0)or drought-stressed(0), with measurement of C 0 2 uptake, titratable acidity and xylem sap tension being made at intervals (days 1,s and 10) following the imposition of the watering regimes. (a) day 1; (b) day 5; (c) day 10. Redrawn from Lee and Griffiths (1987).

During the initial 24-h sampling period, for both + and - water treatments, the AH' was 74 pmol H+ (g fresh wt)-' (Fig. 2a). Those undergoing continued drought showed net C 0 2uptake at night, while the watered plants had C3 gas exchange characteristics. After 5 days, the rewatered plants maintained a relatively constant level of titratable acidity, while those undergoing drought increased AH' to 94 pmol H' (g fresh wt)-' (Fig. 2b). After 10days, the background level of titratable acidity had fallen to 35 pmol H' (g fresh wt)-' (i.e. fully C,), but the plants undergoing drought were obviously stressed and AH' was reduced to 76 pmol H' (g fresh wt)-' (Fig. 2c). At each stage, the dusk values of xylem sap tension reflected improved water status in the rewatered plants, with the C3transition being represented by (0.364.26 MPa); in the CAM mode, xylem sap tension was constant at 0.5 MPa. As part of this progression from C3 to CAM in S. telephiurn it is possible to identify a number of the specific intermediate types described in Section III.A, including latent CAM (Fig. 2b, water), CAM cycling (Fig. 2a, - water) and perhaps a shift towards CAM idling (Fig. 2c, - water). How much simpler, therefore, to utilize the AH+ in conjunction with the proportion of respired C 0 2to describe the entire transition from C3to CAM (Lee and Griffiths, 1987). What, then, is the stimulus for the transition between C3and CAM? Plant

+

78

H. GRIFFITHS

water status is most simply implicated, and there was a change in xylem sap tension in S . telephium as CAM was repressed (Fig. 2; Lee and Griffiths, 1987). Ting and Rayder (1982) proposed that the mechanism may be controlled by stomata, and cite the high background levels of organic acids in P. afra as being poised for CAM (i.e. latent CAM?). In S. tefephium,there were similar high background levels of titratable acidity but also daytime C02uptake (Fig. 2b), and C 0 2uptake also occurred in the light following a large AH' (Fig. 2a). Daytime C02 uptake immediately following organic acid accumulation (i.e. effectively during phase 111) was observed in nearly all of the studies listed in Table 111, and the regulation of stomata1 aperture during deacidification urgently requires clarification, particularly as Nishio and Ting (1987) have suggested that daytime C4carboxylation may occur in Peperomia. The observation that root anoxia or low temperature induces CAM in M . crystallinurn (Winter, 1985) also points to a mesophyll-based stimulus. Ting and Rayder (1982) also suggested a role for ABA in CAM induction, but this has been discounted by Winter (1985) since ABAper se did not induce CAM in M . crystaflinum. A further suggestion, made by Lee and Griffiths (1987), was that the increasing background levels of organic acids could parallel osmotic adjustment in C3plants (Teeri et a f . , 1986). Thus as drought stress continues, the organic acids start to fluctuate diurnally, with the C 0 2being mainly derived from respiratory C 0 2 . This could also be related to the potential for water uptake during dewfall as x increases (Ruess and Eller, 1985; Liittge, 1987,1988; Smith et a f . ,1987; Section 1I.E). Investigations in our laboratory are currently directed towards characterizing this transition for S . tefephium.Preliminary observations in watercultured plants have shown that a AH' is induced by PEG within 24 h of changing the rooting media, and that levels of proline are three times higher in water culture + PEG, and six times higher in soil-grown plants with CAM, when compared to well-watered controls (Lee and Griffiths, unpublished results). At any event, it seems that any inductive stimulus is likely to be found in the mesophyll, but much more work is required to characterize the physiology and biochemistry of the C,-CAM transition before the inductive process can be identified.

IV. SIGNIFICANCE OF RESPIRATORY CO2 UTILIZATION DURING CAM Recycling has been shown above to be a recurring phenomenon during CAM. An attempt is now made to relate the significance of this apparently energetically futile process to the regulation of CAM under natural conditions in terrestrial and aquatic environments. Perhaps a close parallel can

CRASSULACEAN ACID METABOLISM

79

be seen between recycling in CAM and photorespiration in C3plants: both seem in teleological terms to have little purpose, but are ubiquitous nevertheless (Osmond et al., 1980). A. RECYCLING IN THE TERRESTRIAL ENVIRONMENT

To date, only Winter et al. (1986b) have attempted a systematic study of respiratory COz recycling, although it has always been quantified by P. S . Nobel. Using two stem succulents (0.vulgaris and M . woodsii) and a leaf succulent (K.daigrernontiana), COzuptake and acidification (also in COzfree air) were measured under a range of temperatures and imposed drought stress (Winter et al., 1986b). Some of the results have been summarized in Table I11 (see Section III.C), but in general terms they showed that recycling of respiratory COz conserved carbon at high temperatures, with -20% of carbon being lost at 30°C (Winter et al., 1986b). Intriguingly, under drought stress AH' was greater at high night temperatures, but net carbon balance was negative (Winter et al., 1986b). Results in Table IV are presented from a series of experiments on two species of epiphytic bromeliads in which distribution is clearly related to altitudinal zonation and rainfall in Trinidad (Griffiths, unpublished results). A . nudicaulis is one of the most widely distributed CAM epiphytes, throughout most of the rainfall zones, but at upper altitudes it occurs less frequently and A . fendleri predominates (Smith et al., 1986a). A . nudicaulis forms a tank from a tight rosette of leaves, as opposed to the looser rosette and less succulent leaves of A . fendleri. Field measurements in Trinidad indicated that these two species also differ markedly in terms of expression and regulation of CAM activity, and this had in part been correlated to the degree of succulence (Griffiths et al., 1986). Plants were grown in a humid glasshouse and then transferred to growth chambers. The regulation of CAM, gas exchange and water relations were studied under two light regimes (100 and 300 pmol m-'s-l PAR), and a range of night temperatures (12,18 and 25°C). Phase I COz uptake was significantly greater at higher PAR under the 25/25 regime, but under low PAR ACOzwas similar under both temperature regimes (Table IV). For both species, ACOz increased with temperature under high PAR, but was maximal at 18°C under low PAR. Our previous studies have shown that under natural conditions A . fendleri has a lower percentage of recycling than A . nudicaulis (Griffiths et al., 1986). This also holds for these laboratory-grown plants, although under some night temperature regimes (under low PAR) recycling is similar for the two species in percentage terms, but in absolute terms dH+ was generally greater in A . nudicaulis (Table IV).

TABLE IV Respiratory C 0 2 recycling by two epiphytic bromeliuds (Aechmea fendleri and A . nudicaulis) in response to variations in PAR and night temperature

Night temp. (“C)

Dusk-dawn titratable acidity AH+ (mol m-3)

Net C 0 2 uptake dark period (m mol kg-’)

250

12 18 25

191 249 254

100

12 18 25

250

PAR (pmol m-’s-’) Aechmea fendleri

Aechmea nudicaulis

100 ~

~~

~~

Data of Griffiths, unpublished.

Recycling of respired C 0 2

Net dark respiration (pmol O2m-2 s-l)

Calculated respiration

%

6H’

86.3 123.4 134.6

10 1 0

18 2 0

0.40 1.83 3.67

0.09 0.01

242 233 243

62.3 110.0 88.2

49 6 27

117 13 67

0.52 1.56 3.77

0.60 0.07 0.34

12 18 25

160 163 250

47.3 55.2 73.1

41 32 38

65 104

0.40 2.84 5.36

0.71 0.57 1.13

12 18 25

113 123 144

27.0 58.1 45.6

52 6 37

59 7 53

0.73 2.50 5.29

0.64 0.07 0.58

53

CRASSULACEAN ACID METABOLISM

81

At 25°C under high PAR, A . fendleri took up more C 0 2 than could be accounted by the AH’; at 18”C,there was nearly the exact 2 : 1stoichiometry of AH’ : C 0 2 uptake, and at 12°C only 10% of AH’ was derived from respired C 0 2 . Under low PAR, there was much more recycling by A . fendleri at 12 and 25°C (49% and 27% respectively), but 18°C seemed optimal with only 6% recycling. A . nudicaulis showed a similar response, with 3241% recycling under high PAR, but also near-optimal6Yo recycling at 18°C under low PAR. A theoretical rate of respiration can be calculated from the discrepancy between C 0 2uptake at night and the observed degree of acidification (Griffiths et al., 1986, 1988; Luttge and Ball, 1987;Lee et al., 1988). This is shown in Table I1 in comparison with the measured rate of respiration, the latter being determined with the Hansatech leaf electrode at each of the night temperatures utilized. In only three cases ( A . fendleri, 12”C,low PAR; A . nudicaulis, 12”C,high and low PAR) does the theoretical respiration rate approach that measured with the O2 electrode. This indicates that the rates of recycling of respired C 0 2 , derived indirectly from differences in C o t uptake and acidification at night, are not unreasonable in physiological terms and emphasizes the importance of recycling to plant carbon balance. This also validates the field data for many bromeliad species (Griffiths et al., 1986). Xylem sap tension did not significantly alter between treatments (data not shown), and although bromeliad stomata may respond directly to humidity (Lange and Medina, 1979; Griffiths et al., 1986), the variation in recycling shown in Table IV can be ascribed to a direct interaction between temperature, respiration and PAR supply. Osmond et al. (1980) were the first to suggest that the recycling of respired C 0 2 during CAM may have a role in preventing photoinhibition by maintaining photosystem stability. Fluorescence kinetics (variable and low temperature) demonstrated that photochemical stability was maintained after six months of drought stress in CAM-idling plants (Osmond, 1982). It has also been demonstrated that CAM plants show photoinhibition (Nobel and Hartsock, 1983; Martin et al., 1986; Table II), and more recently that 0. basilaris may be photoinhibited throughout the year in Death Valley, California (Adams et al., 1987). There are three characteristic phenotypic variations of Bromelia humilis found under natural conditions which have been described as “yellow exposed”, “green exposed” and “green shade” (Lee et al., 1988; see also Medina et al., 1986). Measurements of variable fluorescence in the field on transplanted populations of the three phenotypes have also demonstrated that photoinhibition may be a limiting factor in the exposed plants under natural conditions (Lee et al., 1988). It would be premature to suggest that recycling is only related to photoinhibition, particularly as many laboratory studies are carried out well below PAR saturation (see Tables 11,111 and IV). In A . fendleri and A . nudicaulis there is clearly an interaction between PAR and night temperature which

82

H. GRIFFITHS

results in reduced recycling at 18°C; the significance of these processes is as yet unresolved. In view of the 1: 1stoichiometry between AH' (estimated as Amal) and An (Smith and Luttge, 1985; Smith et al. , 1986b), it is unlikely to be due to any major changes in organic acid speciation (see Section 1V.C). What is clear, however, is that integrated studies which consider leaf energy status (Koster and Winter, 1985), quantum yield (Adams et al., 1986; Adams et al., 1987) and regulation of respiratory processes in all CAM phases (Luttge and Ball, 1987) are now required. B . RECYCLING IN THE AQUATIC ENVIRONMENT

The activity of CAM as a biochemical C02-concentrating mechanism has been shown to be a response to C 0 2 limitation in the aquatic habitat (for review see Keeley, 1987; Raven, 1984; Raven et al., 1987). Although the parallel between this form of CAM and that in terrestrial plants has already been drawn (Section III.A), there are further similarities in terms of the extent of respiratory C 0 2recycling (Richardson et al., 1984; Madsen, 1987). Madsen has shown that internal C 0 2 levels within the lacunae of the submerged aquatic Littorella unifroru are maintained well above bulk water equilibration concentrations, equivalent to 1-2% C 0 2in air. Table V shows the data for the regulation of gas exchange (C02and O2by both roots and shoots) in L . unifZora grown under two PAR regimes, 50 and TABLE V Recycling of respiratory C 0 2and net carbon balance of Littorella uniflora under two P A R regimes Recycling PAR (pmol m-2 s-')

AH' (pmol (g fresh wt)-')

ACOza

%

6H'

50

42

+8

60

26

300

117

+17

71

83

Dark respirationb 29 49

Balance of gas exchange by intact plants" Roots PAR (pmol m-2 s-l)

50 300

Shoots

co2

+12 +17

-3

+0.3

Whole plant co2

0 2

+8

+11

+ 17

+20

'Determined by changes in bulk medium concentration around intact plants (pmol (g fresh wt)-') in a 12-h dark period. + = net C 0 2or 02 uptake; - = net C 0 2 or O2release. bDeterminedby O2electrode with leaf slices (pmol (g fresh wt)-'). Data of Robe and Griffiths, unpublished.

CRASSULACEAN ACID METABOLISM

83

300 pmol m-2 s-l (Robe and Griffiths, unpublished). In terms of the regulation of CAM, AH’ and AC02 were greater under high PAR, and although recycling expressed as a percentage is similar, the absolute 6H’ is much greater under high PAR (60-71%, 26-83 pmol H+ (g fresh wt)-’), in parallel with the higher rate of net dark respiration (Table V). The overall carbon balance of the plant is determined by a preponderance of root over shoot C 0 2 uptake, with C 0 2 being lost from the leaves of the lower light treatment. There was, however, an excellent agreement between total C 0 2 uptake (as measured by infrared gas analysis) and O2 (as measured by O2 electrode) for the intact plant. A constant concentration of C 0 2 was maintained within the leaf lacunae throughout day and night, at 1.03 and 0.86 mol m-3 (low and high PAR, respectively; data not shown). The significance of this C02-concentrating mechanism can easily be understood for isoetids growing in “vernal pools” which undergo large diel fluctuations in 02,C 0 2 , pH and temperature (Keeley, 1981, 1983, 1987). These are in marked contrast to the relatively constant physicochemical conditions in large oligotrophic water bodies where isoetids are also commonly found, e.g. Loch Brandy, Tayside, Scotland (Richardson etal., 1984; and also Keeley, 1987). Just as the same criteria for CAM apply to plants in terrestrial and aquatic habitats, so perhaps does the rationale behind the extent of respiratory C 0 2 recycling. In any event, the unique plasticity of CAM shows remarkably uniform responses when analysed comparatively in this way, for a pathway which has evolved independently so many times throughout the plant kingdom. C. ORGANIC ACID SPECIATION: THE NEWCASTLE HYPOTHESIS REVISITED

One of the most widely quoted reviews originally defining CAM is that of Ranson and Thomas (1960), who characterized the biochemical regulation of the diel cycle (which we now recognize as the CAM phases) but perhaps failed to stress the ecophysiological adaptations now inherent in our understanding of CAM. Perhaps we could now explain the then puzzling discrepancy in malic acid : C 0 2uptake as resulting from the recycling of respiratory c02. One other common theme throughout the review of Ranson and Thomas concerned the role of citric acid and other organic acid fluctuations, but because they “behaved erratically” the authors went on to “focus attention primarily on . . . malate” (Ranson and Thomas, 1960), the role of citric acid being put aside in favour of calculations of malic acid stoichiometry and energetics. However, an example was given whereby citric acid could account for up to 25% of the AHf in Bryophyllum calycinum (Pucher et al., 1949). At the same time, Wolf (1960) also reviewed CAM, and cited his

84

H. GRIFFITHS

work where 4 6 8 0 % of AH' could be accounted for by citric acid in the same species (Wolf, 1939). Clusia rosea (Ting et al., 1985b), Welwitschia mirabilis (von Willert, 1985) and S. acre (Schuber and Kluge, 1981) have also shown the potential for citric acid accumulation during CAM. Luttge has suggested that citric acid has often been ignored because the strict stoichiometry between malate, C 0 2 and 2H' had been demonstrated for several species (see Luttge, 1988). Then, as now, Ranson and Thomas were unable to perceive any rationale behind the biochemistry or bioenergetics of citric acid accumulation. Luttge (1988) makes a detailed analysis of the energetic costs of malic versus citric acids, and points out that there is no advantage in terms of net C 0 2 gain if hexoses or glucans are used as carbon reserves via citrate synthase. In energetic terms, citric acid is more favourable because synthesis and storage produces net ATP, as compared to malic acid (Luttge, 1988; Section 1I.C). The only function for citric acid as an osmoticum [in terms of enhanced water uptake (Luttge, 1987, 1988; Ruess and Eller, 1985; von Willert and Brinckmann, 1985; Smith et al., 1987)] could be when it is derived from storage glucans rather than hexoses (Luttge, 1988). Citric acid could be important in terms of respiratory C 0 2recycling, which seems predominantly to conserve carbon during CAM. Unfortunately, few studies have measured the combined contribution of malic + citric acid to nocturnal acidification [see Ong et ul. (1986) where shikimic acid fluctuations rather than citric acid are concomitant with malic acid in ferns]. These deficiencies are amply demonstrated in Tables I1 and 111, whereby all calculations were based on the 2H': lmal: 1 C 0 2stoichiometry. Based on our current understanding of the regulation of citrate metabolism, it is only possible for citric acid to result in net C 0 2 fixation if acyl-CoA has been derived from fatty acids by P-oxidation. Current work at Newcastle has not only indicated that P-oxidation of fatty acids can occur in plant mitochondria (Thomas and Wood, 1986; Wood et al., 1986), but has also emphasized the role of carnitine as a transmembrane fatty acid carrier. Carnitine has been shown to be important in the regulation of fatty acid transfer in both mitochondria and chloroplasts (Wood et al., 1984), and the full implications for cellular biochemical control are still being elucidated. What, then, would result if carnitine was fed to K . crenatu leaves in the transpiration stream? Figure 3 shows the results from one of a series of experiments, whereby 20 mol mP3 D/L carnitine was added to the water supply of detached leaves of K . crenata for 24 h prior to measurement of titratable acidity during the dark period (Brown, Griffiths, Thomas and Wood, unpublished results). It can be seen that carnitine significantly enhanced the rate of acidification, AH' and deacidification (Fig. 3), but it should be noted that no such response was found when 10 mol m-3 D/L carnitine was supplied (data not shown). Although an investigation into the role of citric acid in CAM is currently

CRASSULACEAN ACID METABOLISM

85

I

24.00

06.00

12.00

Fig. 3 . Stimulation of acidification by carnitine in detached leaves of Kulunchoe crenata. Petioles were incubated for 24 h prior to sampling in 50 mol m-3 phosphate buffer (I = control) or in 20 rnol m-3 carnitine in 50 mol m-3 phosphate buffer (0 + carnitine). Unpublished results of Brown, Griffiths, Thomas and Wood.

being undertaken at Newcastle in terms of regulation by mitochondrial, chloroplastic and microbody reactions, we have at the moment no conventional explanation for the data shown in Fig. 3; it is presented simply to show that there is still much to learn about the regulation of organic acid synthesis and decarboxylation during CAM.

V. CAM: DEVELOPMENT OF INTEGRATED RESEARCH The recent review by Luttge (1987) extolled the virtues of CAM research as “exemplifying the need for integration in ecophysiological work”, and the synthesis presented in this chapter certainly bears out this view. An attempt has been made to re-appraise the regulation of CAM, as found in a wide range of terrestrial and aquatic habitats, and it is hoped that the initial “reductionist” aim of simplifying and unifying our interpretation of CAM has been achieved. It is also hoped that a number of avenues of future research interest have been identified: an essential feature of any future work will be no longer simply to examine CAM activity at a single level, but to compare environmental interactions with regard to both the phenotypic and genotypic plasticity in CAM. More than ever, there is a need for rigorous comparative approaches to as many components of CAM as possible, e.g. AH’, organic acid speciation, biochemical regulation of carboxylation and respiration, gas exchange and leaf-cell water relations (e.g. Osmond et al., 1982; Osmond, 1984,1987). With CAM thus unified in

86

H.GRIFFITHS

theory, but diverse in form and function, we may be able to evaluate more fully the many facets of this intriguing metabolic adaptation.

ACKNOWLEDGEMENTS Financial support from NERC, The Nuffield Foundation and The Royal Society is gratefully acknowledged. I would also like to thank the following colleagues who have stimulated discussion and have allowed access to unpublished data: J. Brown, Dr H. S. J. Lee, Professor U. Luttge, Dr C. Martin, W. Robe, Dr J. A. C. Smith, Dr D. R. Thomas and Dr C. Wood. I am also indebted to Lynn Wilson for processing the ms with alacrity and efficiency, and to M. Green, N. M. Griffiths and C. S. Hetherington for technical assistance.

REFERENCES Acevedo, E., Badillo, I. and Nobel, P. S. (1983). Plant Physiol. 72,775-780. Adams, W. W. 111, Noshida, K. and Osmond, C. B. (1986). Plant Physiol. 81, 297-300. Adams, W. W. 111, Smith, S. D. and Osmond, C. B. (1987). Oecologia 71,221-228. AndrC, M., Thomas, D. A., von Willert, D. J. and Gerbaud, A. (1979). Planta 147, 141-144. Bartholomew, D. P. (1982). In “Crassulacean Acid Metabolism” (I. P. Ting and M. Gibbs, eds), pp. 278-294. American Society of Plant Physiologists, Rockville. Beardall, J., Griffiths, H. and Raven, J. A. (1982). J . Exp. Bot. 33,729-737. Benzing, D. H. (1980). “The Biology of the Bromeliads.” Mad River Press Inc., Eureka, California. Benzing, D. H. and Renfrow, A. (1971). Bull. Torrey. Bot. Club 98,322-327. Benzing, D. H., Friedman, W. E., Peterson, G . and Renfrow, A. (1983). Am. J. Bot. 70,121-133. Black, C. C., Carnal, N. W. and Kenyon, W. H. (1982). In “Crassulacean Acid Metabolism” (I. P. Ting and M. Gibbs, eds), pp. 51-68. American Society of Plant Physiologists, Rockville. Boston, H. L. and Adams, M. S. (1986). Oecologia 68,615622. Brulfert, J. and Queiroz, 0. (1982). Planta 154,339-343. Brulfert, J., Kluge, M., Guerrier, D. and Queiroz, 0. (1987). Planta 170,92-98. Chen, S . 4 . and Black, C. C. (1983). Plant Physiol. 71,373-378. Cockburn, W. (1983). Plant Cell Environ. 6,275-279. Cockburn, W. (1985). New Phytol. 101,3-24. Cockburn, W.,Ting, I. P. andSternberg,L. 0.(1979). PlantPhysiol. 63,1029-1032. Cockburn, W., Goh, C. J. and Avadhani, P. N. (1985). Plant Physiol. 77,53-86. Deleens, E. and Garnier-Dardart, J. (1977). Planta 135,241-248. Deleens, E. and Queiroz, 0. (1984). Plant Cell Environ. 7,279-283. Deleens, E., Treichel, I. and O’Leary, M. H. (1985). Plant Physiol. 79,202-206. de Luca, P., Alfani, A. and Virzo de Santo, A. (1977). Bot. Gaz. 138,474-478. Diaz, M. and Medina, E. (1984). In “Eco-fisiologica de Plantas CAM” (E. Medina, ed.), pp. 98-113. CIET (IVIC-UNESCO), Caracas. Earnshaw, M. J . , Carver, K. A. and Lee, J. A. (1985). Oecologia 67,486-492.

CRASSULACEAN ACID METABOLISM

87

Earnshaw, M. J., Carver, K. A. and Charlton, W. A. (1987). P h t a 170,421432. Edwards, G. E. and Walker, D. A. (1983). “C3, C4: Mechanisms and Cellular and Environmental Regulation of Photosynthesis.” Blackwell, Oxford. Eller, B. M. and Ruess, B. R. (1986). J. Plant Physiol. 125,295-309. Evans, J. K., Sharkey, T. D., Berry, J. A. and Farquhar, G. D. (1986). Aust. J. Plant Physiol. 13,281-292. Farquhar, G. D. (1980). In “Carbon Dioxide and Climate: Australian Research” (G. I. Pearman, ed.), pp. 105-110. Australian Academyof Science, Canberra. Farquhar, G. D. and Richards, R. A. (1984). Aust. J . Plant Physiol. 11,539-552. Farquhar, G. D., O’Leary, M. H. and Berry, J. A. (1982). Aust. J. Plant Physiol. 9, 121-137. Fischer, A. and Kluge, M. (1984). Planta 160,121-128. Friemert, V . ,Kluge, M. and Smith, J. A. C. (1986). Physiol. Plant. 68,353-358. Fu, C. L. and Hew, C. S. (1982). Bot. Caz. 294-297. Griffiths, H. (1984). In “Eco-fisiologia de Plantas CAM (E. Medina, ed.), pp. 45174. CIET (IVIC-UNESCO), Caracas. Griffiths, H. (1988). In “Phylogeny and Physiology of Epiphytes” (U. Luttge, ed.). Springer, Berlin. (in press). Griffiths, H. and Smith, J. A. C. (1983). Oecologia 60,176-184. Griffiths, H., Luttge, U., Stimmel, K.-H., Crook, C. E., Griffiths, N. M. and Smith, J. A. C. (1986). Plant Cell Environ. 9,385-393. Griffiths, H., Smith, J. A. C., Luttge, U., Cram, W. J. C., Lee,H. S. J.,Medina, E., Popp, M., Schafer, C. and Stimmel, K.-H. (1988). New Phytol. (in press). Groenhof, A. C., Bryant, J. A. and Etherington, J. R. (1986). Ann. Bot. 57, 689-695. Guralnick, L. J. and Ting, I. P. (1986). Oecologia 70,85-91. Guralnick, L. J., Rorabaugh, P. H. and Ting, I. P. (1984a). Plant Physiol. 75, 454-457. Guralnick, L. J., Rorabough, P. H. and Ting, I. P. (1984b). Plant Physiol. 76, 643-646. Guralnick, L. J., Ting, I. P. and Lord, E. M. (1986). Am. J. Bot. 73,336-345. Hartsock, T. L. and Nobel, P. S. (1976). Nature 262,574-576. Heun, A.-M., Gorham, J., Luttge, U., Wyn-Jones, R. G. (1981). Oecologia 50, 66-72. Hew, C. S., Ng, Y.W., Wong, S. C., Yeoh, H. H. and Ho, K. K. (1984). Physiol. Plant. 60,154-158. Holtum, J. A. M., O’Leary, M. H. and Osmond, C. B. (1983). Plant Physiol. 71, 602-609. Holtum, J. A. M., Simmons, R. A., Roeske, C., Comins, H. and O’Leary, M. H. (1984). In “Advances in Photosynthesis Research” (C. Sybesma, ed.), Vol. 111, pp. 545-548. Dr W. Junk, Dordrecht. Keeley, J. E. (1981). A m . J . Bot. 66,1079-1084. Keeley, J. E. (1983). Oecologia 58,57-62. Keeley, J. E. (1987). In “The Ecology of Amphibious and Intertidal Plants” (R. M. Crawford, ed.). Cambridge University Press. Keeley, J. E. and Bowes, G. (1982). Plant Physiol. 20,1455-1458. Keeley, J. E., Osmond, C. B. and Raven, J. A. (1984). Nature 310,634-635. Kemp, P. R. and Gardetto, P. E. (1982). Oecologia 55,149-156. Kerby, N. W. and Raven, J. A. (1985). In “Advances in Botanical Research” (J. A. Callow and H. W. Woolhouse, eds.), pp. 71-123. Academic Press, London. Kluge, M. (1977). Oecologia 29,77-83. Kluge, M. and Ting, I. P. (1978). “Crassulacean Acid Metabolism.” Springer, Berlin.

88

H . GRIFFITHS

Kluge, M., Fischer, K. and Buchanan-Bollig, B. (1982). In “Crassulacean Acid Metabolism” (I. P. Ting and M. Gibbs, eds), pp. 31-50. American Society of Plant Physiologists, Rockville. Koster, S. and Winter, K. (1985). Plant Physiol. 79,520-524. Lange, 0 .L. and Medina, E. (1979). Oecologia 40,357-363. Lange, 0 . L. and Zuber, M. (1980). Flora 170,529-553. Lee, H. S. J. and Griffiths, H. (1987). J. exp. Bot. 38,834-841. Lee, H. S. J., Smith, J. A . C., Cram, W. J., Griffiths, H., Luttge, U., Medina, E., Popp, M., Schafer, C. and Stimmel, K.-H. (1988). New Phytol. (in press). Littlejohn, R. 0. and Ku, M. S. B. (1984). Plant Physiol. 74,1050-1054. Luttge, U. (1986). Planta 168,287-289. Liittge, U. (1987). New Phytol. 106,593-629. Luttge, U. (1988). Plant Cell Environ. (in press). Luttge, U. and Ball, E. (1979). J . Membr. Biol. 47,401-422. Luttge, U. and Bzll, E. (1980). Plant Cell Environ. 3, 195-200. Luttge, U. and Ball, E. (1987). Plant Physiol. Biochem. 25,3-10. Luttge, U. and Clarkson, D. (1985). Prog. Bot. 47,73-86. Luttge, U. and Nobel, P. S . (1984). Plant Physiol. 75,804-807. Luttge, U. andsmith, J. A. C. (1984a). J . Membr. Biol. 81,149-158. Liittge, U. and Smith, J. A. C. (1984b). In “Salinity Tolerance in Plants: Strategies for Crop Improvement” (R. C. Staples and G. H. Toenniessen, eds), pp. 125150. Wiley and Sons, New York. Luttge, U., Smith, J. A . C. and Marigo, G. (1982). In “Crassulacean Acid Metabolism” (I. P. Ting and M. Gibbs, eds), pp. 69-91. American Society of Plant Physiologists, Rockville. Luttge, U., Stimmel, K.-H., Smith, J. A. C. and Griffiths, H. (1986). Plant. Cell Environ. 9,377-383. Madsen, T. V. (1987). J . Exp. Bot. 38,367-377. Martin, C. E. and Adams, W. W. I11 (1987). Photosynth. Res. 237-244. Martin, C . E. and Zee, A . K. (1983). Plant Physiol. 73,718-723. Martin, C. E., Christiansen, N. L. and Strain, B. R. (1981). Oecologia 49,322-328. Martin, C. E., Lubbers, A. E. andTeeri, J. A. (1982). Bot. Gaz. 143,491-497. Martin, C. E., Eades, C. A. and Pitner, R. A. (1986). Plant Physiol. 80,23-26. McWilliams, E. (1970). Bot. Gaz. 131,285-290. Medina, E. (1974). Evolution 28,677-686. Medina, E. (1982). 2. Pflanzenphysiol. 107,251-258. Medina, E. (1984). In “Eco-fisiologia de Plantas CAM” (E. Medina, ed.), pp. 5272. CIET (IVIC-UNESCO), Caracas. Medina, E. and Delgado, M. (1976). Photosynthetica 10,155-163. Medina, E. and Osmond, C. B. (1981). Aust. J. Plant Physiol. 8,641-649. Medina, E. and Troughton, J. H. (1974). Plant Sci. Lett. 2,357-364. Medina, E., Delgado, M., Troughton, J. H. and Medina, J. D. (1977). Flora 166, 137-152. Medina, E., Olivares, E. and Diaz, M. (1986). Oecologia 70,441-446. Mook, W. G., Bommerson, J. C. and Staverman, W. H. (1974). Earth Plan. Sci. Lett. 22, 169-176. Mooney, H. A., Troughton, J. H. and Berry, J. A. (1977). Oecologia 30,295-306. Nalborczyk, E., La Croix, L. J. and Hill, R. D. (1975). Can. J. Bot. 53,1132-1138. Neales, T. F. (1975). In “Environmental and Biological Control of Photosynthesis” (R. Marcelle, ed.), pp. 299-310. Dr W. Junk, The Hague. Neales, T. F. and Hew, C. S. (1975). Planta 123,303-306. Neales, T. F., Sale, P. J. M. and Meyer, C. P. (1980). Aust. J . Plant Physiol. 7, 375-385.

CRASSULACEAN ACID METABOLISM

89

Nishio, J. N. and Ting, I. P. (1987). Plant Physiol. 84,600-604. Nobel, P. S. (1976). Plant Physiol. 58,576582. Nobel, P. S . (1982). In “Crassulacean Acid Metabolism” (I. P. Ting and M. Gibbs, eds), pp. 260-277. American Society of Plant Physiologists, Rockville. Nobel, P. S. (1983). Plant Cell Environ. 6,711-719. Nobel, P. S. and Hartsock, T. L. (1978). Plant Physiol. 61,510-514. Nobel, P. S . and Hartsock, T. L. (1983). Plant Physiol. 71,71-75. Nobel, P. S. and Hartsock, T. L. (1984). Physiol. Plant. 60,98-105. Nobel, P. S. and Hartsock, T. L. (1986). Plant Physiol. 80,913-917. Nobel, P. S., Luttge, U., Heuer, S. and Ball, E. (1984). Plant Physiol. 75,799-803. O’Leary, M. H. (1981). Phytochemistry 20,553-587. O’Leary, M. H. and Osmond, C. B. (1980). Plant Physiol. 66,931-934. O’Leary, M. H., Treichel, I. and Rooney, M. (1986). Plant Physiol. 80,578-582. Olivares, E., Urich, R., Montes, G., Coronel, I. and Herrera, A. (1984). Oecologia 61,358-362. Ong, B. L., Kluge, M. and Friemert, V. (1986). Plant Cell Environ. 9,547-557. Osmond, C. B. (1978). Ann. Rev. Plant Physiol. 29,379-414. Osmond, C. B. (1982). In “Crassulacean Acid Metabolism” (I. P. Ting and M. Gibbs, eds), pp. 112-127. American Society of Plant Physiologists, Rockville. Osmond, C. B. (1984). In “Advances in Photosynthesis Research” (C. Sybesma, ed.), Vol. 111, pp. 557-564. Dr W. Junk, Dordrecht. Osmond, C. B. (1987). New Phytol. 106 (supplement), 161-175. Osmond, C. B., Ziegler, H., Stichler, W. and Trimborn, W. (1975). Oecologia 18, 209-217. Osmond, C. B., Bender, M. M. and Burris, R. H. (1976). Aust. J . Plant Physiol. 3 , 787-799. Osmond, C. B., Nott, D. L. and Firth, P. M. (1979). Oecologia 40,331-350. Osmond, C. B., Winter, K. and Powles, S. B. (1980). In “Plant Responses to Water and High Temperature Stress” (N. C. Turner and D. J. Kramer, eds), pp. 139-154. Wiley, New York. Osmond, C. B., Winter, K. and Ziegler, H. (1982). In “Encyclopaedia of Plant Physiology”, Vol. 12B, “Physiological Plant Ecology 11” (0. L. Lange, P. S. Nobel, C. B. Osmond and H. Ziegler, eds), pp. 479-549. SpringerVerlag, Berlin. Pearman, G. I. (1980). In “Carbon Dioxide and Climate: Australian Research” (Pearman, G . I., ed.), pp. 11-20. Australian Academy of Science, Canberra. Pucher, G. W., Vickery, H. B., Abrahams, M. S. and Leavenworth, C. S. (1949). Plant Physiol. 24,610-620. Ranson, S . L. and Thomas, M. (1960). Ann. Rev. Plant Physiol. 11,81-100. Raven, J. A. (1984). “Energetics and Transport in Aquatic Plants.” A. R. Liss, New York. Raven, J. A. (1986). New Phytol. 104,25-77. Raven, J. A. (1987). In “Biochemistry of Plants” (D. D. Davies, ed.), Vol. 13, pp. 127-180. Academic Press, London. Raven, J. A., Beardall, J. and Griffiths, H. (1982). Oecologia 53,68-78. Raven, J. A., MacFarlane, J. J. and Griffiths, H. (1987). In “The Ecology of Amphibious and Intertidal Plants” (R. M. Crawford, ed.), pp. 129-149. Cambridge University Press. Rayder, L. and Ting, I. P. (1981). Plant Physiol. 69,139-142. Rayder, L. and Ting, I. P. (1983a). Plant Physiol. 72,606-610. Rayder, L. and Ting, I. P. (1983b). Photosynth. Res. 4,203-211.

90

H.GRIFFITHS

Richardson, K., Griffiths, H., Reed, M. L., Raven, J. A. and Griffiths, N. M. (1984). Oecologia 61,115-121. Ritz, D., Kluge, M. and Veith, H. J. (1986). Planta 167,284-291. Ruess, B. R. and Eller, B. M. (1985). Planta 166,57-66. Rundel, D. W., Rundel, J. A., Ziegler, H. and Stichler, W. (1979). Oecologia 38, 45-50. Rygol, J., Buchner, K.-H., Winter, K. and Zimmermann, U. (1986). Oecologia 69, 171-175. Sale, P. J. M. and Neales, T. F. (1980). Aust. J. Planf Physiol. 7,363-373. Schafer, C. and Liittge, U . (1986). Oecologia 71,127-132. Schuber, M. and Kluge, M. (1981). Oecologia 50,82-87. Sinclair, R. (1984). Plant Physiol. 77,5943. Sipes, D. L. and Ting, I. P. (1985). Plant Physiol. 77,59-63. Smith, F. A. and Raven, J. A. (1979). Ann. Rev. Plant Physiol. 30,289-311. Smith, J. A. C. (1984). In “Eco-fisiologia des Plantas CAM” (E. Medina, ed.), pp. 30-51, CIET (IVIC-UNESCO), Caracas. Smith, J. A. C. and Liittge, U. (1985). Planta 163,272-283. Smith, J. A. C. andNobel, P. S. (1986). J. Exp. Bot. 37,1044-1053. Smith, J. A. C., Uribe, E. G., Ball, E., Heuer, S. and Liittge, U. (1984a). Eur. J . Biochem. 141,415-420. Smith, J. A. C., Uribe, E. G., Ball, E. and Liittge, U. (1984b).Planta 162,299-304. Smith, J. A. C., Griffiths, H., Bassett, M. and Griffiths, N. M. (1985). Oecologia 67, 475-485. Smith, J. A. C., Griffiths, H. andLiittge, U. (1986a). Plant Cell Environ. 9,359-376. Smith, J. A. C., Griffiths, H., Liittge, U., Crook, C. E., Griffiths, N. M. and Stimmel, K.-H. (1986b). Plant Cell Environ. 9,395-410. Smith, J. A. C., Schulte, P. J. and Nobel, P. S. (1987). Plant Cell Environ. 10, 639-648. Smith, T. L. and Eickmeier, W. G. (1983). Oecologia 56,374-380. Spalding, M. H., Schmitt, M. R., Ku, J. B. and Edwards, G. E. (1979). Plant Physiol. 63,738-743. Sternberg, L., De Niro, M. J. and Ting, I. P. (1984a). Plant Physiol. 74,104-107. Sternberg, L., De Niro, M. J. and Johnson, H. B. (1984b). Plant Physiol. 74, 537-561. Steudle, E., Smith, J. A. C. and Luttge, U. (1980). Plant Physiol. 66,1155-1163. Stimmel, K.-H. (1984). In “Eco-fisiologia de Plantas CAM” (E. Medina, ed.), pp. 73-82. CIET (IVIC-UNESCO), Caracas. Szarek, S. R. and Troughton, J. H. (1976). Plant Physiol. 52,539-541. Szarek, S . R., Johnson, H. B. and Ting, I. P. (1973). Plant Physiol. 52,539-541. Teeri, J. A. (1982a). In “Biochemical Aspects of Evolutionary Biology” (M. H. Nitacki, ed.), pp. 93-130. University of Chicago Press, Chicago. Teeri, J. A. (1982b). In “Crassulacean Acid Metabolism” (I. P. Ting and M. Gibbs, eds), pp. 224-259. American Society of Plant Physiologists, Rockville. Teeri, J. A. (1984). Oecologia 64,68-73. Teeri, J. A., Tonsor, S. J . and Turner, M. (1981). Oecologia 50,367-369. Teeri, J. A., Turner, M. and Gurevitch, J. (1986). Plant Physiol. 81,678-680. Thomas, D. R. and Wood, C. (1986). Planta 168,261-266. Ting, I. P. (1985). Ann. Rev. Plant Physiol. 36,595-622. Ting, I. P. and Burk, J. A. (1983). Plant Sci. Lett. 32,279-285. Ting, I. P. and Gibbs, M. (1982). “Crassulacean Acid Metabolism.” American Society of Plant Physiologists, Rockville. Ting, I. P. and Hanscom, Z. (1977). Plant Physiol. 59,511-514.

CRASSULACEAN ACID METABOLISM

91

Ting, I. P. and Rayder, L. (1982). In “Crassulacean Acid Metabolism” (I. P. Ting and M. Gibbs, eds), pp. 193-207. American Society of Plant Physiologists, Rockville. Ting, I. P. and Sipes, D. (1985). In “Nitrogen Fixation and C0,Metabolism” (P. W. Ludden and J. E. Burris, eds), pp. 371-378. Elsevier, Holland. Ting, I. P., Sternberg, L. 0. and De Niro, M. J. (1983). Plant Physiol. 71,677-679. Ting, I. P., Bates, L., Sternberg, L. 0. and De Niro, M. J. (1985a). Plant Physiol. 78,246-249. Ting, I. P., Lord, E. M., Sternberg, L., Da, S. L. andDe Niro, M. J. (1985b). Science 229,969-971. Tinoco-Ojanguren, C. and Vazquez-Yanes, C. (1983). Bol. SOC.Bot. Mex. 45, 150-153. Tomlinson, P. B. (1969). In “Anatomy of the Monocotyledons” (C. R. Metcalfe, ed.), 111. Commelinales-Zingiberales. Oxford University Press, Oxford. Virzo de Santo, A., Alfani, A., Russo, G. and Fioretto, A. (1983). Bot. Gaz. 144, 342-346. Virzo de Santo, A., Ligrone, R . , Alfani, A., Fioretto, A. and Russo, G. (1984). Plant Cell Environ. 7,105-112. von Willert, D. J. (1985). In “Advances in Botanical Research” (J. A. Callow and H. W. Woolhouse, eds), pp. 157-191. von Willert, D. J. and Brinckmann, E. (1985). Ber. Deutsch Bot. Ges. Bd. 98, 455-464. von Willert, D. J., Eller, B. M., Brickmann, E . and Baasch, R. (1982). Oecologia 55,21-29. Wagner, J. and Larcher, W. (1981). Oecologia 50,88-93. Walker, D. B., Gysi, J., Sternberg, L. and De Niro, M. J. (1983). Science 220, 419-421. Winter, K. (1974). Ber. Deutsch Bot. Ges. Bd. 86,467-476. Winter, K. (1982). In “Crassulacean Acid Metabolism’’ (Ting, I. P. and Gibbs, M., eds), pp. 153-169. American Society of Plant Physiologists, Rockville. Winter, K. (1985). In “Photosynthetic Mechanisms and the Environment” (J. Baxter and N. R. Baker, eds), pp. 329-387. Elsevier, Amsterdam. Winter, K. and Luttge, U. (1976). In “Ecological Studies”, Vol. 19, “Water and Plant Life” (0.L. Lange, L. Kappen and E.-D. Schultze, eds), pp. 323-334. Springer-Verlag, Berlin. Winter, K. and Schramm, M. J. (1986). Plant Physiol. 82,173-178. Winter, K. and Tenhunen, J. D. (1982). Plant Physiol. 70,1718-1722. Winter, K. and von Willert, D. J. (1972). 2. Pflanzenphysiol. 267,166-170. Winter, K., Troughton, J. H., Evanari, M., Lauchli, A. and Luttge, U. (1976). Oecologia 25,125-143. Winter, K., Luttge, U., Winter, E. and Troughton, J. H. (1978). Oecologia 34, 225-237. Winter, K., Edwards, G. E. and Holtum, J. A. M. (1981). Plant Physiol. 68,355-357. Winter, K., Wallace, B. J., Stocker, G. C. and Roksandic, Z. (1983). Oecologia 57, 129-141. Winter, K., Medina, E., Garcia, V., Mayoral, M. L. and Muniz, R . (1985). J . Plant Physiol. 113,73-78. Winter, K., Osmond, C. B. and Hubick, K. T. (1986a). Oecologia 68,224-230. Winter, K., Schroppel-Maier, G. and Caldwell, M. M. (1986b). Plant Physiol. 81, 390-394. Wood, C., Noh, Hj Jalil, J., MacLaren, I . , Yong, B. C. S . , Ariffin, A., McNeil, P. H., Burgess, N. and Thomas, D. R. (1984). Planra 161,255-260.

92

H.GRIFFITHS

Wood, C., Burgess, N. and Thomas, D. R. (1986). Planta 167,54-57. Woodhouse, R. M.,Williams, J. G . and Nobel, P. S. (1980). Am. J . Bot. 67, 1179-1 185. Wolf, J. (1939). Plunta 29,450467. Wolf, J. (1960). in “Encyclopaedia of Plant Physiology 12” (Ruhland, W . , ed.), pp. 809-889. Springer, Berlin. Zeiger, E. (1983). Ann. Rev. Plant Physiol. 34,441475. Ziegler, H.,Osmond, C. B . , Stichler, W. and Trimborn, P. (1976). Planta 128, 85-92.

Potassium Transport in Roots

LEON V . KOCHIANa and WILLIAM J . LUCASb a

U.S.Plant. Soil and Nutrition Laboratory. USDA.ARS. Cornell University. Ithaca. New York. USA Department of Botany. University of California. Davis. California. USA

I. I1.

Introduction

. . . . . . . . . . . . . . . . .

Plasma Membrane Transport of K+ in Roots . . . . . . . A . Early Work: the Carrier-Kinetic A proach . . . . . . B . Are Root K+ Fluxes Coupled to H ? . . . . . . . . C . Uptake at High K+ Concentrations: the Linear Component D . Summary . . . . . . . . . . . . . . . .

P

I11.

Redox-coupledPlasmalemmaTransportof K+ . . A . Influence of Exogenous N A D H on K+Influx . B Membrane Transport and the Wound Response C. DevelopmentofanIntegratedNADHModel .

V.

VI .

.

.

.

.

.

94 94 99 120 128

. . . . . . . . . . . . . . .

129 130 131 132

. . . . . . . . .

136 137 140 143 145

Regulation of K+ Fluxes within the Plant . . . . . . . . . A . Allosteric Regulation of K+ Transport . . . . . . . . B . K+ Cycling within the Plant: an Integration of Regulatory . . . . . . . . . . . . . . . . Mechanisms

151 152

.

IV .

94

Radial K+ Transport to the Xylem . . A . Site of K+ Entry into the Symplasm B . Radial Pathway . . . . . . C . Lag Phase in Xylem Loading . . D . K+ Transport into the Xylem . .

Future Research and Prospects

. . . .

. . . .

. . . .

. . . .

. . . . .

. . . .

. . . .

. . . .

. . . .

. . . . . . . . . . .

Advances in Botanical Research Vol . 15

. . . .

163 169

Copyright 0 1988 Academic Press Limited All rights of reproduction in any form reserved .

ISBN 0-12-005915-0

93

94

L. V. KOCHIAN AND W. J. LUCAS

I. INTRODUCTION The involvement of Kf in a number of plant functions (enzyme activation, maintenance of turgor, protein synthesis, etc.), along with the relatively high permeability of plant cell membranes to K’, has been the basis for the study of K+ transport as a model system for plant ion transport. Numerous investigations have been conducted over the past 40 years, aimed at furthering our understanding of the mechanism(s), energetics, and cellular location of K+ transport systems in roots. Despite the surfeit of literature in these areas, it can be said with some confidence that the processes by which K+ ions are transported into and across the root are still far from being fully resolved. This is due, in part, to the complex nature of the root, with its various cell types and tissues. Additional complications arise from the fact that the individual cells are coupled (both electrically and physiologically) via plasmodesmata, and also simply from the complex nature of individual cells, with their multiple compartments and subcompartments. In the first part of this chapter, we will outline and summarize what is known (or at least thought to be known) concerning mechanistic aspects of K+ transport in roots. Once this background has been established, the regulation of these transport processes can be discussed in the context of the integration of these cellular processes at the organ and whole plant level.

11. PLASMA MEMBRANE TRANSPORT OF K+ IN ROOTS A. EARLY WORK: THE CARRIER-KINETIC APPROACH

Over the past half century, most of the studies that have been conducted concerning ion absorption by plants have generally utilized three basic types of plant material: the giant algae such as Chum, Nitellu, and Vuloniu, slices cut from storage tissue of beet, potato, and carrot, and either intact roots or roots excised from seedlings. The use of excised roots as research material can be traced to the classical paper of Hoagland and Broyer (1936). They found that the roots of barley seedlings grown in dilute salt solutions exhibited extremely high initial rates of ion accumulation. Because radioisotopes had not yet been introduced, the ability of these roots to maintain high rates of accumulation made them very useful experimental material. Hence, the now well-known “low-salt roots” characterized by low salt content, high sugar content, and a large capacity for ion transport, became widely used in research. It was during this era that many of the basic concepts of membrane transport were developed. Much of the work conducted with low-salt roots concerned the absorption of K+ and other alkali cations (for the reasons discussed above). Although the concept of the lipid bilayer nature of the plasmalemma had not yet been fully developed, many researchers were

95

POTASSIUM TRANSPORT IN ROOTS

beginning to realize that plant (and animal) cells were bounded by a non-aqueous layer that was relatively impermeable to electrolytes (Osterhout, 1931). Furthermore, the rapid absorption of Kf by plant cells led a number of investigators to hypothesize that special structures must exist that would facilitate K+ uptake across this outer boundary (Jacobsen and Overstreet, 1947; Jacobsen et al. , 1950; Osterhout, 1950,1952). The concept that, in the plasmalemma, specific carriers were involved in K+ uptake was developed further by Epstein and co-workers. When K+(86Rb+)uptake from dilute solutions into excised barley roots was studied, the kinetic profile approximated a rectangular hyperbola. This response was quite similar to that found in classical enzyme kinetic studies [Fig. 1 (see low concentration range)]. Epstein and Hagen (1952) were the first to apply Michaelis-Menten enzyme kinetics to ion transport. They postulated that specific alkali cation transport systems operated in a fashion analogous to substrate-specific enzymes. They went on to show, for the uptake of K+ and other alkali cations, that at higher external concentrations their observed kinetics deviated from classical Michaelis-Menten form (Leggett and Epstein, 1956; Epstein et af., 1963). Saturation was attained at low external K+ concentrations, but then at higher concentrations the curves appeared to reach a second level of saturation (Fig. 1). This biphasic pattern was labelled the “dual isotherm of uptake” by Epstein, and was hypothesized to be due to the operation of two separate classes of carriers in the plasmalemma. In the low K+ concentration range ( 1mM

Em

>EK

E, i w l t h IK'I,

t

K'influx via uniport

Fig. 6. Schematic model developed by Cheeseman and Hanson (1979b) to explain the role of the electrogenic H+-ATPase in Kf transport across the corn root plasmalemma when the external K+ concentration is within the mechanism I and I1 range. Note that within the mechanism I1 K+ concentration range, it is proposed that external K+ has a stimulatory effect on the rate at which H t are pumped out by the ATPase. E, is the electrogenic component of the Emcontributed by the H+-ATPase.

and is insensitive to ATPase inhibitors. Essentially, what Cheeseman and co-workers were proposing was a model that incorporates both types of coupling illustrated in Fig. 4. At low external K+ levels, active Kf uptake would occur via a K+-H+ exchange ATPase (pathway 2 of Fig. 4). At higher external Kf concentrations, passive K+ uptake would be mediated by a K+ channel, with the driving force arising primarily from the electrical potential difference developed across the plasmalemma by a H+-ATPase. Recently, evidence in support of this model was presented following a re-evaluation of the electrogenic nature of K+-H+ fluxes in corn roots (Thibaud et al., 1986). It should be noted that if Kf uptake is active at low K+ levels, then the K+ transport system must be coupled either directly to an ATPase (K+-ATPase or K+-H+ exchange ATPase), or to the transmembrane proton electrochemical gradient (PMF) via a Kf-H+ co-transport system. Indirect coupling (pathway 1, Fig. 4)can only be invoked for the passive K+ uniport. Early evidence for the existence of a Kf-transporting ATPase came from work which correlated Kf influx in corn, wheat, oat, and barley roots with K+-stimulated ATPase activity located in a microsomal membrane preparation isolated from these roots (Fisher and Hodges, 1969; Fisher et al., 1970). The putative plasmalemma ATPase was further characterized in subsequent studies (Hodges et al., 1972; Leonard and Hodges, 1973; Leonard and Hotchkiss, 1976). The enzyme showed an acidic pH optimum for activity, and a requirement for Mg2+,and was further stimulated by

106

L. V. KOCHIAN AND W. J. LUCAS

"0

10

20

30

40

50

20

40

60

80

1c

I

0

KCI concentration ( mM Fig. 7. Correlation between the kinetics of 42K+influx into excised oat roots (a) and the stimulation of a microsomal ATPase activity as a function of K+ present in the reaction medium (b). ATPase activity was determined by the amount of Pi released. Insets show 42K+ influx and Pi released over the concentration range of 0.01-0.35 mM KCI. Data from Leonard and Hodges (1973).

monovalent cations, particularly K+ and Rb'. Leonard and Hodges (1973) showed that the complex kinetics for K+ absorption, in oat roots, were quite similar to the kinetics observed for K+ stimulation of ATPase activity (Fig. 7). It seemed appropriate, therefore, to correlate K+ transport in plants with the H+/K+-ATPase of the gastric mucosa and the Na+/K+ATPase of animal cells, since these systems all exhibited cation-induced stimulation and transported K+ directly (Hodges, 1976; Cantley, 1981; Faller et al., 1982; Leonard, 1984; Briskin, 1986a). The correlation between plasma membrane-associated, K+-stimulated ATPase activity and K+ influx, and the similarity between the sequence for monovalent cation stimulation of ATPase activity (K' > NH: > Rb+ > Cs+ > Li') and the specificity of monovalent cation uptake into roots (Sze and Hodges, 1977), has been used as further evidence that this ATPase is involved in K+ uptake, putatively as a K+-H+ exchange system. It has also been demonstrated that microsomal membrane vesicles isolated from

POTASSIUM TRANSPORT IN ROOTS

107

tobacco callus, and believed to be plasmalemma in origin, exhibited a stimulation of K+-ATPase activity in the presence of nigericin (which facilitates electroneutral K+-H+ exchange), or valinomycin plus a protonophore (Sze, 1980). These results again suggest that the plasmalemma ATPase may be involved in K+-H+ exchange. Recent work with isolated plasma membranes from corn roots (Briskin and Leonard, 1982a,b), oat roots (Vara and Serrano, 1983), and red beet storage tissue (Briskin and Poole, 1983a,b), has demonstrated that the plasmalemma ATPase forms a covalently bound, phosphorylated intermediate during the course of ATP hydrolysis. Briskin and Leonard (1982a,b) showed that K+ mildly stimulated the breakdown of this phosphorylated intermediate, using a crude preparation from corn roots, and concluded that this stimulation is further evidence for a K+ transport function by the ATPase. An analogy was again drawn with the reaction mechanism for the Na+/K+-ATPaseof animal cells, which also exhibits a K+ stimulation of dephosphorylation, and transports K+ directly (Cantley, 1981). However, it should be noted that the Na+/K+-ATPase phosphoenzyme intermediate has a much higher turnover rate, and the influence of K+ on the rate of dephosphorylation is considerable. Similar K+ effects on the dephosphorylation step were obtained with the plasmalemma ATPase isolated from red beet (Briskin and Poole, 1983a,b; Briskin and Thornley, 1985;Briskin, 1986b). However, in these studies the authors appear to place less emphasis on ascribing a K+ transport function for the ATPase. In contrast to the above studies, Vara and Serrano (1983) could find no effect of K+ on the phosphorylation and dephosphorylation of the plasmalemma ATPase from oat roots. This result led Serrano (1984) to propose that the ATPase is not a K+-dependent enzyme involved in K+ transport. As he points out, this interpretation is supported by the observation that, in the studies of the corn root microsomal membranes, the turnover of the phosphoenzyme was quite slow, and the K+ stimulation of the phosphoenzyme breakdown was much too small to support the concept of a direct role for K+ in the reaction kinetic scheme for this enzyme. Recently, Spanswick and Anton (1988) presented a model describing the kinetic characteristics of a H+-translocatingATPase purified from isolated plasmalemma vesicles of tomato roots (Fig. 8). They observed that the kinetics for the ATPase were similar to those for other cation-transporting ATPases that form phosphorylated intermediates. Their model does not ascribe a K+ transport role to the ATPase. Instead, stimulation of the breakdown of the phosphorylated intermediate by cytoplasmic K+ appears to be sufficient to explain the K+-stimulation of ATPase activity, without invoking a K+ transport role for the ATPase. It was proposed that K+ transport occurs by a separate mechanism operating in conjunction with the H+-ATPase. There are additional criticisms that can be directed at the work published in support of a K+ transport function for the plant plasmalemma ATPase.

L. V. KOCHIAN AND W. J. LUCAS

108

t

K:Vi

-“-f E2-p

Fig. 8. A reaction scheme for the plasma membrane ATPase proposed by Spanswick and Anton (1988). The binding of H+ at the cytoplasmic face of the plasma membrane occurs via reaction 4 to give HE1, subsequent hydrolysis of ATP via reaction 1 produces HE, P and from this state the H+ is released to the outer surface of the plasma membrane, via reaction 2. Potassium is proposed to act from the cytoplasmic face of the membrane to accelerate the dephosphorylation of E2-P, to return the ATPase to the state in which it can again bind H+. (Vi stands for inorganic vanadate which acts as a competitive inhibitor on the E2complex.)

-

First, the degree of K+ stimulation of ATPase activity is quite low when compared with total ATPase activity, and the K+ concentrations often used to achieve this stimulation are excessively high (50 mM) when considering Kf concentrations likely to be present in the soil solution or root apoplasm. These results again suggest that the K+effect may be a general salt effect, or activation by cytoplasmic K+, and not evidence for Kf requirement by a K+ transporting ATPase. Additionally, K+ stimulation is observed only at pH values below 7 (Leonard and Hodges, 1973; Leonard and Hotchkiss, 1976). These considerations led Serrano (1984), in his quite comprehensive review of fungal and higher plant plasmalemma ATPases, to state, “Thus, the plasma membrane ATPase is not a potassium-dependent enzyme although it is specifically stimulated by this ion”. An alternative view can be taken which would fit both mechanistic views. The dominant transport protein in the plant plasmalemma could be the H+-translocatingATPase, which might not be expected to exhibit Kf-stimulated activity. However, a Kf-ATPase could also be present, at much lower levels. This could explain the small K+ stimulation of ATPase activity and phosphoenzyme turnover, which would be measured against a large background of activity due to the H+-ATPase. It is not unrealistic, in the light of work from animal and bacterial systems, to suggest that a number of

POTASSIUM TRANSPORT IN ROOTS

109

different ATPases might be contained in the various plant membranes, operating in response to different signals and participating in various cell functions. The resolution of this controversy may ultimately come from demonstrations of ATP-dependent K+ and H+ transport in purified and reconstituted systems (Briskin, 1986a). Much of the early work on plasma membrane ATPases shared the rather common problem of contamination by other membranes. For example, it was only recently that ATP-dependent transport was demonstrated with membrane fractions isolated from plant roots. Using microsomal membranes isolated from corn, several different groups have now been able to demonstrate the existence of an ATP-dependent proton pump (Sze and Churchill, 1981; Churchill and Sze, 1983; DuPont et al., 1982; Mettler et al., 1982; Stout and Cleland, 1982; Bennett and Spanswick, 1983). However, it was unclear at the time whether the membrane vesicles exhibiting H+ transport originated from the plasmalemma. It appears that much of the early examples of ATP-dependent H+ pumping were being conducted with tonoplast vesicles (Mettler et al., 1982; Serrano, 1984). It has now been amply demonstrated that at least two distinct ATPases were contained in these microsomal membrane preparations: a K+-stimulated, vanadate-sensitive, NO;-insensitive ATPase that is plasma membrane in origin, and a C1--stimulated, vanadate-insensitive, NO;inhibited ATPase that is located in the tonoplast (Marrb and Ballarin-Denti, 1985). ATP-dependent Hf pumping has been demonstrated with both types of membrane vesicles. Recent improvements in plant membrane fractionation techniques have allowed for the isolation of fairly pure membrane fractions. Hence, it is now possible to reconstitute partially purified preparations of plasmalemma ATPases into proteoliposomes (Vara and Serrano, 1982; O'Neill and Spanswick, 1984). However, an unequivocal demonstration of ATP-dependent K+transport is still lacking. When partially purified oat root (Vara and Serrano, 1982) and red beet storage tissue plasmalemma ATPases (O'Neill and Spanswick, 1984) were reconstituted into liposomes, only H+ transport was observed. Vara and Serrano demonstrated ATP-dependent Hf pumping with inverted liposomes containing K+-free solutions (Fig. 9), which suggested to them that the ATPase did not mediate K+-H+ exchange. Additionally, it can also be seen in Fig. 9 that introducing K+ produced an instantaneous stimulation of H+ transport, which would be expected if K+ was acting on the active site of the ATPase facing the external medium (cytoplasmic side of inverted vesicles). They interrupted this result as further evidence against a K+-H+ transport function, since a lag in the stimulation by Kf would be expected as K+ diffused into the vesicle and became available for transport. Because of the demonstrated H+-pumping capacity of these membranes, and a lack of K+ transport ability, the present consensus of most researchers

110

L. V. KOCHIAN AND W. J. LUCAS

ATP

KCI\

ATP K W \

,_

I

Fig. 9. Changes in 9-amino-6-chloro-2-methoxyacridine fluorescence upon energization of Kf-free proteoliposomes prepared from oat root plasma membrane ATPase. Assay medium contained either 25 mM MgS04 (A, B and C), or 25 mM Mg(N03)2(D). Tris-ATP (1.25 mM), K2S04(25 mM), KCI (50 mM), KN03 (50 mM), imidazole-HC1 (pH 6.5,20 mM), and gramicidin D (5 pg ml-') were added as indicated. The initial rate of quenching expressed as per cenrof total fluorescence min-' is indicated. Data from Vara and Serrano (1982).

in this area appears to be that the plant plasmalemma ATPase acts as an electrogenic Hf transport system rather than a K+/H+-ATPase. However, this is an area of research in which rapid progress is being made, and it may well be that this consensus will be modified in the near future. Very recently, evidence suggesting ATP-dependent K+ transport in plasmalemma vesicles isolated from red beet storage tissue has been presented (Giannini et af., 1987). In this study, inverted plasmalemma vesicles were loaded with 86Rb+-labelled K+ solutions by a freeze-thaw technique. Although these vesicles were leaky for K+, the addition of ATP stimulated Kf efflux over and above that observed in the absence of ATP. The ATP-dependent efflux was completely inhibited by vanadate, but only partially inhibited by carbonyl cyanide N-chlorophenylhydrazone (CCCP), which should have abolished the H+ electrochemical gradient. Giannini et al. suggest that K+ transport may be mediated by two systems; one system would be indirectly coupled to the H+ gradient, while the other would be directly coupled to the ATPase, as a K+-translocating ATPase. It should be noted that the results of Giannini et af. do not support the K+-H+ exchange hypothesis. As noted by Serrano (1984), H+ ionophores

POTASSIUM TRANSPORT IN ROOTS

111

should not influence the transport function of ATPases that directly transport K+, either as a Kf-ATPase or as a K+/H+-ATPase.However, since H+ ionophores dissipate both the chemical and electrical components of the protonmotive force generated by a Hf-ATPase, any indirect coupling of K+ transport to a H+-ATPase would be significantly inhibited. This would include K+ uniports driven by the electrical component of the PMF and H+-K+ co-transport. Evidence for the co-transport of K+ with H+ will be discussed in Section II.B.4. 3. Indirect Coupling: Electrophoretic K + Uniport A number of researchers have argued that K+ uptake is not directly coupled to the H+-translocatingATPase, but is associated indirectly with the activity of the H+ pump through a K+ transport system driven by the electrical component of the PMF (Pitman et al., 1975; Marrb, 1979 and references therein). Pitman and co-workers were the first to propose this type of electrophoretic coupling, following studies on the effects of fusicoccin (FC) on K+ and H+ fluxes in low-salt and salt-saturated barley roots (Pitman et al., 1975). Their studies were generally conducted at high external K+ levels (5 mM). In low-salt roots, the FC-stimulated H+ efflux was similar, whether the roots were exposed to 5 mM KCl or 5 mM NaC1. However, in saltsaturated roots, which have a greater passive permeability to K+ than Na+ (in relation to low-salt roots), H+ efflux was stimulated to a greater degree in KC1 solutions. Hence, they suggested that FC acted to stimulate passive K+ influx, through the increased electrical gradient created when H+ extrusion was enhanced by FC. Subsequent studies, which further support the hypothesis of an electrical coupling between the H+ efflux and K+ influx, are based on the application of FC to various plant tissues at K+ concentrations (>1 mM) where uptake is considered to be passive (Marr5, 1977, 1979). It has been consistently observed that under these conditions FC stimulates both H+ efflux and K+ uptake, and elicits a hyperpolarization of Em.A large body of circumstantial evidence has been accumulated, indicating that FC acts directly on the plasmalemma H+-ATPase to stimulate H+ translocation (MarrP, 1979). Since, in either the direct or indirect mode of coupling, a stimulation of H+ efflux should enhance K+ uptake, it is not possible to differentiate between the two models simply through the application of FC. Obviously, other approaches must be used in conjunction with FC experiments. Marrb (1979), in his review on FC, has summarized the observations in support of indirect (electrophoretic) coupling. They are as follows: 1. The stoichiometry of the FC-stimulated K+-H+ exchange is often quite close to 1: 1 (at least for corn roots and pea stem segments), particularly when corrections are made to account for H+ consumption during anion uptake. Although this observation can be used in support of either mode of

112

L. V. KOCHIAN AND W. J. LUCAS

coupling, it appears to support indirect coupling when considered in conjunction with the following observations. 2. Under the conditions in which FC-stimulated fluxes are generally studied, Kf is in passive equilibrium across the plasmalemma (Pitman etal., 1975; Cocucci et al., 1976). Hence, any stimulation of net Kf influx would involve an increase in passive uptake. 3. It has been demonstrated that lipophilic cations, such as tributylbenzylammonium and tetraphenylphosphonium, can substitute for K+ in its role in promoting FC-stimulated H+ efflux (Bellando et al., 1979; Bellando and Trotta, 1980). Presumably, these cations cross the plasmalemma nonspecifically, by permeating the lipid bilayer. When this occurs, it has been shown that Em is depolarized and H+ efflux is stimulated. Therefore, it has been argued that the apparent dependency of H+ efflux on K+ is due to a K+ influx-induced depolarization of Em,which would activate the electrogenic H+-translocating ATPase. Conversely, chemical modifiers that stimulate the proton pump (such as FC) would hyperpolarize the membrane potential and increase the driving force for passive, electrophoretic K+ influx. Interpretation of FC experiments is based on the premise that FC acts specifically on the plasma membrane ATPase to increase its activity. The rapidity of the FC effects on plasmalemma ion transport, and the similar inhibitions of FC-stimulated Kf and H+ fluxes, and ATPase activity, by known plasmalemma ATPase inhibitors, are often used in support of this premise (Marrb et al., 1974a,b). Additionally, [3H]-FC has been shown to bind specifically to a protein component of the plasmalemma-enriched fraction isolated from corn coleoptiles (Dohrmann et al., 1977), and FC is known to stimulate ATPase activity associated with plasmalemma-enriched membrane fractions (Beffagna et al., 1977). However, caution must be exercised when interpreting these data. Several groups have solubilized the putative FC-binding protein from plasmalemma-enriched membrane fractions of corn coleoptile (Pesci et al., 1979; Tognoli et al., 1979) and oat roots (Stout and Cleland, 1980). In each case, it has been shown that the FC-binding protein can be separated from the Kf-stimulated plasmalemma Hf-translocating ATPase. Although it has been suggested in each of the above studies that the FC-binding protein could be a subunit of a multisubunit ATPase, the possibility exists that FC may act at sites totally separate from the Hf-ATPase. In a recent electrophysiological study on Viciafaba guard cells, Blatt (1987) obtained evidence that FC may not act on the pump, but via an effect on the related co-transport processes. Additionally, his current-voltage data indicated that FC may act to block a Kf channel involved in K+ efflux from guard cells. Such a mode of action could also explain the FC effects on other plant cell membranes, and it may be necessary to conduct a complete re-evaluation of these FC data. The use of the electrophoretic coupling model to explain Kf uptake over all K+ concentrations is subject to a number of criticisms, some of which

113

POTASSIUM TRANSPORT IN ROOTS

have been rather eloquently summarized (Glass and Siddiqi, 1982; Siddiqi and Glass, 1984). They point out that K+ influx often greatly exceeds H+ efflux, an observation that we have also stressed (Kochian et al., 1987,1988; Newman et al., 1987). This is a feature which has been presented in work dealing with K+ and H+ fluxes, although it has often gone unmentioned in the texts of these papers (e.g. Poole, 1974; Pitman et al., 1975; Lado et al., 1976; Lin and Hanson, 1976). It is difficult to reconcile a model where K+ influx is energetically dependent on the membrane potential (which, in turn, is dependent primarily on H+ efflux), when the H+ fluxes are usually much smaller than the associated K+ fluxes. Sometimes this disparity has been explained on the basis that the measured, “apparent” net H+ efflux is considerably smaller than the true unidirectional efflux, due to processes that also consume H+. These processes might include anion-H+ co-transport, passive “leaks” that would tend to return H+ into the cell, and H+ captured by the cell wall. However, Glass and Siddiqi (1982) conducted a careful assessment of the contribution of these processes to the underestimation of “true” H+ efflux, and concluded that for their system (low-salt barley roots) the measured net H+ efflux is a good approximation of the unidirectional flux. Furthermore, the data presented in Fig. 10 indicate that H+ efflux may be dependent upon external K+, or K+ influx, per se. Clearly, as stressed by Glass and Siddiqi (1982), this is the converse of what would be expected if K+ uptake were dependent upon H+ extrusion. They also noted that at

-0

-0

E,

f

X

2 1

5 -

r

iii

+

I

+ Y

0

0 0.01

0.1

10

1

Potassium Concentration

(

rnol rn-3

)

Fig. 10. Potassium influx (0, A) and H+ efflux (0,A ) as a function of K’ (K,SO,) in the bathing medium. Experiments were conducted on two varieties of barley, var. Fergus (0, 0) andvar. Conquest ( A , A). Data from Glass and Siddiqi (1982).

114

L. V. KOCHIAN AND W. J. LUCAS

higher K+ levels (>1 mM), H+ efflux often levels off or declines slightly (see Fig. 10). This saturation is somewhat surprising, considering earlier published reports which specifically linked K+ and H+ fluxes at higher K+ concentrations (Pitman, 1970; Marr2, 1977). Because it was found that anion influx [SO:- in the case of Glass and Siddiqi (1982)l was greatly increased at higher substrate levels (>0.5 mM K2S04), these potentially anomalous results could be explained by a reduction in net H+ efflux due to the consumption of H+ during SO:- influx, if uptake were mediated by an anion-H+ co-transport. However, these data, and most of the results used in support of models coupling K+ and H+ fluxes, could be explained equally well by a H+ extrusion mechanism involved in the maintenance of charge balance (Glass and Siddiqi, 1982), as originally suggested by Ulrich (1941). The constraints of charge balance are most evident when low-salt roots are transferred from a CaS04medium to one containing various inorganic ions, whence they begin to increase their internal salt levels. The roots would then be in a transition between two regulatory states. Pitman (1970) proposed that the stimulated K+ uptake, H+ efflux, and excess cation absorption exhibited as salt status is increased, would be a reflection of this transitional stage. Siddiqi and Glass (1984) have extended this hypothesis, by suggesting that these transitional changes are the result of a concerted effort, by the root, to maintain electrical neutrality during the alteration of salt status. This attempt to maintain electroneutrality does not appear to be specific for H+ efflux. Under certain conditions, plants can use other ion transport processes in order to effect charge balance (Siddiqi and Glass, 1984). For example, both K+-stimulated Na+ efflux and NHi-stimulated K+ efflux have been observed in plant roots. Thus, the apparent coupling of K+ and H+ fluxes may reflect a more general feature of plants, i.e. that cation exchanges (H+, K+, Na', Ca2+etc.) may be needed to maintain charge balance during times when high rates of cation uptake occur (i.e. during the transition from low-salt to high-salt status). Finally, it seems unlikely that a single type of mechanism, or coupling mode, will suffice to account for K+ uptake in higher plants. This is particularly apparent when one considers the potentially wide range of soil K+ concentrations that a root may experience. There are many examples in the literature for plant, bacterial, and animal transport systems, where the uptake of a particular solute is mediated by two or more types of transport mechanisms. Since it appears that K+ uptake in higher plants is active at low K+ levels, and thermodynamically passive at higher concentrations, it seems plausible to speculate that two different K+ transport systems might exist. The existence of K+ channels has been known for many years in animal and bacterial membranes, and increasingly strong evidence for K+ channels in higher plants has been accumulating (Schroeder et al., 1984,1987; Bentrup et al., 1985; Kochian et al., 1985; Kolb et al., 1987). Hence, it is quite possible that K+ channels in the plasmalemma facilitate passive K+ uptake (or release) at higher (>0.5 mM) concentrations. The direction and magnitude

115

POTASSIUM TRANSPORT IN ROOTS

of these fluxes would depend upon the value of the membrane potential and on any voltage-dependent characteristics that these channels may possess. However, at low K+ concentrations, where active K+ uptake most probably occurs, a different type of transport system must operate. Uptake must either be directly coupled to an ATPase, as a K+-ATPase (or the less likely case of a K+/H+-ATPase),or active K+ uptake may be coupled to the H+ gradient, via a K+-H+ co-transport system. The K+-H+ co-transport hypothesis will be considered in the next section. Other Modes of Coupling: K + - H + Co-transport? Another approach that has been taken in order to simultaneously study K+ and H+ transport in roots involves the use of ion-selective microelectrodes for K+ and H+ (Kochian and Lucas, 1987; Kochian et al., 1988; Newman et al., 1987). An ion-selective microelectrode system was developed that could quantify and map the extracellular electrochemical potential gradients for K+, H+, and C1- along the roots of 4-day-old corn seedlings. From an analysis of the extracellular ion gradients, it is possible to simultaneously determine and monitor the net H+ and K+ fluxes associated with a few cells at the root surface (due to the small electrode tip diameter of 0.5-1.0 pm). Because this system provides a high degree of spatial (and temporal) resolution, it has proven to be a useful method for studying the coupling of K+ and H+ fluxes at the cellular level. A diagrammatic representation of the system used in these studies is shown in Fig. 11. Data collected with this system indicated that at any point along the root, large fluctuations in the fluxes (particularly H+ fluxes) occurred with time. On many occasions, H+ efflux was near zero while K+ influx was “normal”; the converse was also occasionally observed. Additionally, when repeated measurements were made at the same location on the root, both H+ efflux and K+ influx could vary significantly, and usually the variation of the two fluxes did not show any correlation. The data presented in Table I illustrate 4.

TABLE I Measurement of net H + and K+ influx in low-salt-grown corn roots K+ Influx

Time (min)

H+ Efflux (pmol (g fresh wt)-’ h-’)

(pmol (g fresh wt)-’ h-’)

0 3 8 12 20

1.87 0.76 0.00 0.00 1.12

2.92 3.18 2.52 2.89 2.27

Measurements were made with pH and K+ microelectrodes at a position 3.0 cm from the root apex at distances of 50 and 125 pm from the root surface. Bathing solution consisted of 0.1 mM K2S04 + 0.2 mM CaS04. Data taken from Kochian and Lucas (1987).

116

L. V. KOCHIAN AND W. J. LUCAS

€'=c E,v-€; 3

(- ,

+ c v-vo I,-I

6

f

Rootradius = 500pm

v, C O

( a )

E;

__

Ell+EKt

!I/

I

' Centre of corn roo1

I -

Portion of an intact root or root segment

1 5 p m j r M i c r o e l e c t r o d e tipslocated within15-20pm (b 1

_ -/ Cno,r////

root

'///////'Surface

of corn root

Fig. 11. Schematic representation of the ion-specific microelectrode system used by Newman et al. (1987) to measure the Kt and H+ electrochemical potential gradients which develop at the surface of intact corn roots. Individual voltage components are shown in (a): EL and E r represent the EMF of the half-cells for the reference and ion-selective electrode, respectively; E: is the EMF of the half-cell for the extracelluar electric potential electrode; E, is the EMF developed across the ion-exchange resin; V and V , are the extracellular potentials near the root surface and in the background bathing solutions, respectively. A scaled representation of the relative positions of Kt,Htand extracellular potential (EI) microelectrodes is shown in (b). Here the microelectrodes (0.5 pm tip diameter) ae shown at their closest position to the corn root surface.

this point. Here, five separate measurements of K+ influx and H+ efflux were made over a 20-min period, and it can be seen that H + efflux varied rather dramatically, while K+ uptake was relatively stable. From these studies no fixed stoichiometry was found for the two fluxes. Generally, net K+ uptake was significantly larger than H+ efflux; the K+ :H + flux ratio ranged from 3.6: 1 to 1:2.75. A similar lack of correlation between K+ and H+ fluxes was observed in barley roots (Glass and Siddiqi, 1982). Increasing the bathing medium pH resulted in a gradual, but small, increase in K+ influx (see Fig. 5 ) . However, H+ efflux did not exhibit a similar pH dependency, but rather appeared to vary independently of the pH value. In this situation, the K+:H+ flux stoichiometries varied from 1:1 to 12: 1, again with no apparent pH dependency. The response of the cortical cell membrane potential to changes in

117

POTASSIUM TRANSPORT IN ROOTS

7 -100 E Y

-0 -120.-

E -1400

c

0

0.2rnM CaS04

I-7Z-l 0.I rnM K2S04

a -1600)

-180-

n

E -200-

5

I

Time (min)

Fig. 12. Response of the cortical membrane potential of intact low-salt-grown corn roots to changes in the concentration of external K+ (as K2S04). (CaSO,, was present at 0.2 mM in all solutions.) Data from Newman et al. (1987).

external K+ was also investigated in these studies. Unlike the case in earlier studies (Cheeseman and Hanson, 1979a,b), the low-salt corn root membrane potential was extremely sensitive to external K+. Increasing the Kf concentration from 0.2 to 20 p~ caused a rapid 70 mV depolarization, from - 190 mV to - 120 mV (see Fig. 12). A further increase in K+ to 200 p~ only elicited a further 7 mV depolarization. Newman et al. (1987) made simultaneous measurements of Em and K+ fluxes in response to changes in external K+,using a three-microelectrode system (one to measure Em,and two K+ microelectrodes for K+ flux measurements). Both the kinetics (dependence on external K') for K+ influx and the depolarization of the membrane potential yielded similar (and quite low) apparent K m values in the 6-9 p~ range. These results indicated that low-salt corn roots possess a very high affinity K+ uptake system that is extremely electrogenic (Newman et al., 1987). In view of the highly electrogenic nature of this system, it is possible that other cations may be transported into the cell with Kf. Since these measurements were made in simple salt solutions (10 p~ K2S04 0.2 mM CaS04), H+ appears to be the most likely candidate for co-transport with K ' . In such a situation, active K+ influx would be achieved using the energy gained from the movement of H+ down their A&. Strong evidence for the existence of a K+-H+ co-transport system in fungi has been recently provided by work on Neurospora (Rodriguez-Navarro and Ramos, 1986; Rodriguez-Navarro et al., 1986;Blatt and Slayman, 1987; Slayman et al., 1988). Slayman and co-workers have described a high-affinity (K, = 1-10 p ~K+ ) uptake system, in K+-starved Neurosporu cells, that is highly electrogenic, as demonstrated by the response presented in Fig. 13. Measurements of internal and external K+, and Em,suggested that this uptake system may mediate active K+ uptake, and the system appears to be

+

118

L. V. KOCHIAN AND W. J. LUCAS ( 0

1

(b)

=-o

-I

100 urn Sdutlan

flow

~

IK'1,-

I

50~M

[K+l,

2

++

3 -304

5aM

,

4n - 2 1 8 +

-305

2014

1 2 0 9 + +

-307

5 0 ~ M

++

200g

Fig. 13. Effect of added extracellularKCon the membrane potential measured in low-K+ spherical cells of Neurospora crassa. (a) Diagram of the arrangement of the cell, the impaling electrode and a K+-floodingpipette for rapid introduction and removal of K+. (b) Depolarization of the Emwith 50 p~ K'. (c) Condensed record from a single cell, showing the Emresponse to four different K+ concentrations. Symbols I and I1 indicate cell impalement and electrode removal, respectively. Data from Rodriguez-Navarroet al. (1986).

coupled to the very active plasmalemma H+-ATPase of Neurospora. A net stoichiometry of one H+ out for one K+ in was demonstrated, which could be taken as evidence for K+-H+ exchange. However, current-voltage analysis conducted by Slayman's group indicated that the K+-associated inward current was twice that of the net K+ influx (see Fig. 14). Thus, one additional positive charge enters with every K+. In addition, the following points must be considered: (1) the H+-ATPaseoperates in parallel with the K+ uptake system; (2) almost every charge absorbed must be balanced by an extruded H+; and (3) only a single H+ is measured (released to the external solution) for every Kf taken up. Hence, the second charge coming in with the K+ must be a H+,which indicates that the high-affinity K+ uptake system operates as a K+-H+ symport (Rodriguez-Navarro et al., 1986; Slayman et al., 1988). Slayman and co-workers speculate that the data presented for higher plants could also be explained by this type of co-transport system coupled to the plasmalemma H+-ATPase. As discussed previously, it has been proposed that a similar co-transport system exists in the plasmalemma of corn root cells (Newman et al., 1987). Further support for this hypothesis comes from the observed similarities between the K+ uptake system seen in K+-starved Neurospora cells and that described above for low-salt corn roots (Kochian et al., 1987,1988; Newman et al., 1987). Both are very high affinity K+ uptake systems with almost identical K , values for K'. Furthermore, K+ uptake through both systems is

119

POTASSIUM TRANSPORT IN ROOTS

? C

2

A

30

0

40

80

120

External K + Concentration

200

160 (

p~

Fig. 14. Stoichiometry of the K+-Hf co-transport system of Neurosporu crassa. The smooth curves are Michaelis-Menten functions fitted to the two sets of data, using a common value of 14.9 ~ L MKt for the apparent K,. Separate values of V,,, are 15.3 k 1.0pmol cm-'s-' for the flux, and 30.1 f 1.6 pmol cm-2 s-' for the measured current. The stoichiometric ratio of the current to net K+ flux is very close to two. Data from Rodriguez-Navarroet af. (1986).

highly depolarizing (compare Figs 12 and 13). The similarities between the two systems, taken in conjunction with the lack of evidence, in low-salt corn roots, for a coupling between K+ influx and H+ efflux, strongly suggest that such a co-transport system could be operating in higher plants. On a kinetic basis, it can be reasoned that a K+-H+ co-transport system should reflect a sensitivity to changes in both extracellular and cytoplasmic pH values. Previous reports of increasing K+ influx, in response to decreasing H+ concentrations in the bathing medium, appear to be somewhat at variance with this prediction. However, in these reports, K+ influx was insensitive to external pH values from 6 to 9 (see Fig. 5 ) . To further investigate this hypothesis for K+ influx into corn roots, Kochian et al. (1987) used K+-selective microelectrodes and 86Rb+to measure K+ influx as a function of pH. Varying the external pH from 4 to 8 had no effect on either net K+ influx (measured with K+ microelectrodes) or unidirectional K+(86Rb+)influx. Additionally, Kochian et al. found that the K+-induced depolarizations of Em (with solution of 50 PM K+) also exhibited absolutely no pH dependence from pH 4 to 8. These results suggest that the K+-H+ co-transport system may have an extremely high affinity for H+. However, an alternative interpretation is that K+ influx is mediated by a different molecular mechanism, e.g. a K+-ATPase,which would not, apriori,show a sensitivity to external pH values.

120

L. V. KOCHIAN AND W. J. LUCAS

C. UPTAKE AT HIGH K+ CONCENTRATIONS: THE LINEAR COMPONENT

As discussed in Section II.A, studies on the kinetics of K+ uptake, for various plant tissues, have generally focused on the generation of complex and often discontinuous kinetic curves, via either the operation of multiple Michaelis-Menten transport systems (such as Epstein's mechanisms I and 11), or by complex, multisite carriers. However, there are numerous examples, particularly in association with studies involving animal and bacterial systems, of complex transport kinetics that appear to be due to the parallel operation of one or more saturable transport mechanisms plus a system that exhibits nonsaturating, or first-order uptake kinetics (linear component). For example, a linear transport component has been demonstrated for the uptake of amino acids and sugars in animal cells (Akedo and Christensen, 1962; Christensen and Liang, 1966; Munck and Schultz, 1969; Cohen, 1975, 1980; Debnam and Levin, 1975), and for the uptake of lactose (Maloney and Wilson, 1973), K+ (Rhoads et al., 1976; Epstein and Laimins, 1980) and amino acids (Wood, 1975; Iaccarino et al., 1978) in E. coli. In many of these studies, nonsaturating solute uptake was dismissed as a physiologically insignificant process, because it was considered to reflect passive diffusion across the lipid portion of the plasma membrane. However, in work on K+ uptake in E. coli, linear Kf uptake was hypothesized to be mediated by a transport protein, which was presumably a Kf channel (Rhoads etal., 1976; Epstein and Laimins, 1980). Furthermore, Christensen and Liang (1966) demonstrated that nonsaturating amino acid uptake in Ehrlich tumour cells was substrate-specific, and exhibited considerable sensitivity to pH and temperature. Thus, passive diffusion was discounted in favour of a more complex system involving a transport protein. In recent years, there has been an increasing interest in linear, nonsaturating solute uptake kinetics from studies involving plant tissues. First-order kinetics have been observed for the uptake of Fe2+in rice roots (Kannan, 1971), sucrose and 3-0-methylglucose in Ricinus cotyledons (Komor, 1977; Komor et al., 1977), sucrose in sugar beet leaf and petiole sections (Maynard and Lucas, 1982a,b), soybean cotyledons (Lichtner and Spanswick, 198l), Vicia leaves (Delrot and Bonnemain, 1981) and red beet vacuoles (Willenbrink and Doll, 1979), and sucrose, fructose, and glucose in Allium leaf discs (Wilson et al., 1985), and for amino acid transport in Lemna (Fischer and Luttge, 1980), and suspension-cultured tobacco cells (Blackman and McDaniel, 1978). Separation of the contribution made by the linear transport component from the saturable mechanisms has been achieved through the use of various inhibitors (Debnam and Levin, 1975; Polley and Hopkins, 1979; Maynard and Lucas, 1982b; Van Be1 et al., 1982). For K+ uptake into corn roots, Kochian and Lucas (1982a,b, 1983) have shown that the complex kinetics

121

POTASSIUM TRANSPORT IN ROOTS

could be resolved into saturable and first-order kinetic components, through the use of sulphhydryl modifiers. When corn root segments were subjected to a series of increasing N-ethyl maleimide (NEM) exposures (e.g. 0,10, and 30 s NEM exposures for the data illustrated in Fig. 3b), saturable K+ uptake was specifically inhibited and subsequently abolished, while linear uptake was relatively unaffected. Hence, it was suggested that the saturable and linear kinetic components represented separate K+ transport systems. Saturable K+ uptake would be analogous to Epstein's mechanism I; however, nonsaturating K+ uptake, which dominates uptake in the concentration range associated with Epstein's mechanism 11, appears to be distinctly different from the transport system described by Epstein and co-workers. Therefore, subsequent work was carried out in order to characterize this linear component for Kf uptake (Kochian and Lucas, 1984; Kochian et al., 1985); certain features of this transport system will now be discussed.

I . Anion Involvement in Nonsaturating K + Uptake Early work by Epstein et al. (1963) demonstrated that in barley roots K+ uptake by mechanism I1 was dramatically inhibited when C1- was replaced by SO:- in the uptake solution. Similarly, it was shown that linear K+uptake in corn roots was partially dependent on the presence of CI- in the uptake solution (Kochian et al., 1985). As shown in Fig. 15, replacing C1- in the

0

2

4

6

8

10

K+concentration (mM

Fig. 15. Influence of the accompanyinganion on K+(86Rb+)influx intocom root segments grown in 0.2 mM CaSO, (low-salt status). The first-order rate coefficients, k , in pmol (g fresh wt)-' h-' m K ' , for the linear component of K+ influx were as follows: 0.5 (Cl-), 0.21 (SO$-), and 0.24 (H2P0,). Data from Kochian et al. (1985).

122

L. V. KOCHIAN AND W. J. LUCAS

uptake solution with SO$-, H2PO; or NO; resulted in a 60% reduction in K+ uptake by the linear component in low-salt roots, while causing only a 15% reduction in the V,,, for the saturable system. In high-salt roots (grown on 5 mM KCl), a similar reduction in linear K+ uptake was seen, while saturable uptake was unaffected. This association between the linear component for K+ uptake and the presence of C1- was shown to be related to a coupling via the saturable C1-

Control

.

I , L

6 DIDS

-a

=

OCO”

o

. 2

4

6

8

10

8

10

CI-concentration (mM 1

0

2

4

6

K+concentration (mM)

Fig. 16. Effect of the anion transport inhibitor DIDS on the influx of wl- (a) and K+ (=Rb’) (b) into low-salt corn root segments. For K+ influx, k had values of 0.40 and 0.19 pmol (g fresh wt)-’ h-’ mM-’for the control and DIDS treatment, respectively. Data from Kochian etal. (1985).

POTASSIUM TRANSPORT IN ROOTS

123

acid influx process. Application of 1mM 4,4-diisothiocyano-2,2’-disulphonic stilbene (DIDS), which has been shown to be an anion transport inhibitor in red blood cells (Cabantchik et al., 1978), corn roots (Lin, 1981) and Chum (Keifer et al., 1982), abolished the saturable component for C1- influx into low-salt corn roots (Fig. 16a). In the presence of this inhibitor, the linear component of K+ uptake was suppressed to an identical degree to that seen when C1- was replaced in the uptake solution by other anions (compare Figs 15 and 16b). These results strongly suggest that the linear component of K+ influx in corn roots is, in some way, linked, at least partially, to saturable C1- uptake.

2. Involvement of K’ Channels? The effect of the quaternary ammonium salt, tetraethylammonium chloride (TEA), which has been shown to block K+ channels in excitable membranes, such as the plasma membrane of nerve fibres (Tasaki and Hagiwara, 1957; Armstrong, 1969), and in Chara (Keifer and Lucas, 1982), was studied on K+ uptake in low- and high-salt corn roots. In high-salt roots, TEA caused a dramatic (75%) and specific inhibition of the linear component of K+ influx (Fig. 17), which suggests that K+ channels may be involved. However, although low-salt roots possess a similar linear component for K+ influx, it was found to be insensitive to TEA, which seems at variance with the above interpretation. Uptake studies with [I4C]-TEAindicated that high-salt corn roots exhibited much higher rates of TEA accumulation than did low-salt roots, presumably via a transport system for quaternary ammonium salts

K+concentration (mM influx into high-salt-growncorn roots. Fig. 17. Influence of 10 mM TEA-CI on K+ (86Rb+) Roots were pretreated with a solution containing 10 mM TEA-C1,5 mM KCI, and 0.2 mM CaS04 for 30 min prior to K+ (“Rb’) uptake (10 mM TEA-CI was included in the uptake solutions). The first-order rate coefficients, k, for the linear component were 0.31 and 0.09 pmol (g fresh wt)-’ h-’ mM-’ for the control and TEA-C1, respectively. Data from Kochian et al. (1985).

124

L. V. KOCHIAN AND W. J. LUCAS

that has been demonstrated in higher plants (Michaelis et al., 1976). Hence, it was proposed that corn root tissue behaves in a similar manner to nerve axons, where it is necessary to inject the TEA into the axoplasm in order to block Kf channels (Tasaki and Hagiwara, 1957; Armstrong, 1969). Highsalt roots may be able to accumulate more TEA in the cytoplasm relative to low-salt roots. If this were the case, then K+ channels could be involved in nonsaturating K+ influx in both low- and high-salt roots; but only in high-salt roots would the cytoplasmic TEA concentration rise to a level high enough to block the putative Kf channels at the cytoplasmic face of the channel. In recent years, the application of the patch-clamp technique to plant cell membranes has provided direct evidence for the existence of K+ channels in higher plant cell membranes. For example, K+ channels have been demonstrated in the plasmalemma of guard cells (Schroeder et al., 1984, 1987), Samanea pulvinar cells (Moran et al., 1987) and corn root cells (Ketchum et al., 1987), and in the tonoplast of Chenopodium suspension cells (Bentrup et al., 1985) and barley mesophyll cells (Kolb et al., 1987). Therefore, it seems reasonable to speculate that one or more classes of K+ channels could operate in the plasmalemma of root cells, in order to facilitate passive K+ uptake at high external levels of K+. Although it is axiomatic that any transport system should have a finite transport capacity, it is possible that a system like an ion channel would not exhibit saturation kinetics under physiological conditions. As Cohen (1975) has pointed out, such a channel-mediated process should eventually saturate at high substrate concentrations. However, in his system (amino acid uptake into mouse brain slices), at high substrate levels, the medium changes from isotonic buffered saline to hypertonic buffered amino acid saline. Thus, any changes in transport kinetics may be due to changes in media composition. The same applies for K+ influx in corn roots. Potassium uptake has been studied from a range of concentrations up to 50 mM. At these high K+ levels, there was no significant change in nonsaturating K+ uptake (Kochian and Lucas, 1982a). Uptake was not studied from solutions of higher concentration, because it was felt that ionic and osmotic effects could alter membrane lipid/protein structure and make data interpretation tenuous.

3. Root Salt Status and Nonsaturating K + Uptake Uptake into low-salt barley roots, at low external K+ levels, is highly specific for K', while at higher K+ concentrations (>0.5 mM), Na+ can competitively inhibit K+ influx (Epstein et al., 1963). A similar response was also observed with low-salt corn roots. Inclusion of 3 mM NaCl in the uptake solution (K+ concentration was from 0.1 to 10 mM) caused a 50% inhibition of the linear component for K+ uptake, while saturable uptake was unchanged (Kochian et al., 1985). The interesting feature is that within the mechanism I1 concentration range, high-salt corn roots exhibit a much higher selectivity for K+ influx over other cations (Pitman, 1967, 1970; Pitman et al., 1968; Kochian et al., 1985). If we accept that the linear component represents the

POTASSIUM TRANSPORT IN ROOTS

125

operation of a Kf channel, these data indicate that some aspect of tissue salt status modifies the channel characteristics from relatively nonspecific to highly K+-specificin high-salt roots. It will be interesting to see whether this prediction can be confirmed using the patch-clamp technique.

4. Alternative Explanation f o r the Linear Component Sanders (1986) has developed a generalized model to explain biphasic or otherwise complex transport kinetics, based on the co-transport of a solute with a driver ion. For Kf influx into roots, a K+-H+ co-transport system would be driven by the inwardly directed A& generated by the H+-ATPase. In Sander’s reaction kinetic model, the assumption is made that both solute (K’)and driver ion (H’) can bind randomly to the carrier, and the limitation is made that the carrier can cross the membrane only as the fully loaded complex (influx) and can return to the outside free of ligand. The reaction kinetic scheme for this model is shown in Fig. 18. Numerical analysis

Fig. 18. Membrane transport modelled on the basis of a reaction kinetic scheme. (a) Reaction kinetic scheme for random binding of solute (S) and H+ to a membrane-bound carrier (X)which catalyses the transport of S across the membrane. Carrier is represented as transporting positive charge in the loaded form. (b). As in (a), but with loaded carrier being neutral and charge transfer occurring on the unloaded form of the carrier. (c) Generalized reaction kinetic scheme for the charged and uncharged models. Concentration (density) of carrier state “j” is designated as Ni, with rate constants (not shown) from carrier state i tostate j designated kii.Redrawn from Sanders (1986), with permission.

126

L. V. KOCHIAN AND W. J. LUCAS

simulation of this model demonstrated that it generates biphasic kinetics under conditions resembling those present in many uptake experiments. The utility of this model is that it makes a number of predictions that are experimentally testable. These predictions are (in terms of a K+-H+ cotransport system): 1. As [H'], is increased to saturating levels, the kinetics for K+ uptake should change from a complex biphasic to a monophasic profile. 2. At saturating [H'],, increasing levels of internal K+ should result in an uncompetitive inhibition of K+ influx.

Fig. 19. Comparison between the experimental data (W) obtained for 6-deoxyglucose influx into Chorella, as a function of bathing medium pH (Komor and Tanner, 1975), and the theoretical simulation of Sanders' (1986) reaction kinetic model. The solid lines represent a simulation of the charged carrier model, simplified for very negative Em and internal 6-deoxyglucose concentration set at zero. Data from Sanders (1986).

POTASSIUM TRANSPORT IN ROOTS

127

3. As [H’l0 is decreased, the relative contribution to the uptake isotherm made by the “apparent” low-affinity K+ transport should increase.

As Sanders (1986) noted, there are few published examples of data concerning pH effects on complex isotherms for solute uptake in plants that permit his predictions to be properly tested. However, such kinetic isotherms do exist for the H+-sugar co-transport system of Chlorella (Komor and Tanner, 1975). As shown in Fig. 19, these data give a reasonable fit to the first prediction of the Sanders model. However, if one then examines the pH dependency of the detailed kinetics for sucrose uptake into sugar beet leaves obtained by Maynard and Lucas (1982a), it is apparent that as the external pH is decreased from 9.0 to 4.0, both the saturable and nonsaturable components of sucrose uptake are stimulated. Clearly, these data do not fit the first prediction of the Sanders (1986) model. Furthermore, it is difficult to reconcile K+ uptake into high-salt corn roots with this Sanders model, since, in this system, it was possible to speciJically abolish saturable K+ uptake with NEM, while leaving nonsaturable uptake relatively unaffected (Fig. 3b). Conversely, it was possible to specifically inhibit the linear transport component either by replacing C1- in the uptake solution with other anions (Fig. 15), or by treating the roots with the K+ channel blocking agent, TEA (Fig. 17). At present it is not obvious how these perturbations, which appear to specifically resolve the complex kinetics for K+ uptake into separate kinetic components, can be explained by the Sanders reaction kinetic model. 5. Physiological Role for Nonsaturating K + Uptake? It has been noted that nonsaturating solute uptake often occurs over a concentration range which exceeds the levels normally experienced by roots in the soil, and so it is difficult to assign a physiological role for such a transport system. Reisenauer (1966) has pointed out that the majority of soil Kf is below 2 mM; yet, in corn roots, nonsaturating K+ uptake becomes significant above 1 mM external K’. What then would be the physiological relevance of a transport system that operates at substrate levels rarely experienced by the plant? The answer to this question may come from transport studies conducted on E. coli and Neurospora. In both organisms, it has been well documented that multiple carrier systems are often involved in the transport of a single solute. These organisms tend to combine constitutive, low-affinity,high-capacity transport systems with derepressible high-affinity systems; glucose and phosphate uptake into N . crassa are excellent examples of this strategy (Lowendorf et al., 1974; Scarborough, 1970). These systems give the organism the adaptive advantage of most effectively obtaining nutrients whose concentrations may vary considerably over a period of time. A root growing through the soil may often experience extremely low K+ levels. Therefore, a high-affinity system may be necessary in order for the

L. V. KOCHIAN AND W. J. LUCAS

128

plant to satisfy its requirements for this essential nutrient. However, if the growing root encounters a localized region of high soil K', the high-affinity system may not have the capacity to utilize excess K'. In such a situation, a low-affinity, high-velocity system would allow the plant to effectively utilize these pockets of high soil K+. Furthermore, as soil water content (potential) declines, the root system must respond in terms of osmotic adjustment. If only a high-affinity K+ transport system were present, the root would not be able to take advantage of the increase in the K+ concentration of the soil solution that would occur along with the decline in soil moisture. The presence of a high-velocity system may provide the plant with an advantage in terms of its response to water stress. In this way, the plant may maintain relatively efficient methods of dealing with its varying environment. D. SUMMARY

A summary of the mechanisms by which K+ may enter the symplasm of the root is given in Fig. 20. Uptake at low external K+ levels is facilitated by a

H+ f

H't-

K'Channel

-

-

- -

~

F e e d b a c k on c a t i o n specificity

Cell Wall

Fig. 20. Schematic representation of possible K+ transport systems operating at the plasmalemma to facilitate K+ uptake at both low and high external K+ concentrations. At low K+ levels, a high-affinity system is hypothesized to be either a K+-ATPase, or a H+-K+ co-transport system, coupled to the H+-ATPase. Both systems would be subject to feedback control by internal K+. Under high K+ levels, nonsaturating K+ uptake involves a K+channel transport. The degree of K+ specificity exhibited which is, in some way, coupled to anion (CI-) by this channel may be influenced by cytoplasmic K+ levels.

POTASSIUM TRANSPORT IN ROOTS

129

very high affinity transport system that is strongly depolarizing, and can transport K+ against its electrochemical potential gradient. Several possible models appear to fit the published data equally well. In our evaluation, the most likely candidates would either be a K+-translocating ATPase or, as in Neurospora, a Kf-Hf symport coupled to the H+-translocating ATPase. The possibility exists that both types of active transport could be functioning in parallel. This transport system would appear to be subject to kinetic and thermodynamic control, and would also be subject to feedback (allosteric) regulation from Kf contained in an internal compartment (see Section V. A). Potassium uptake at higher external levels would be due to a lower-affinity transport system that would transport K+ passively, down its electrochemical potential gradient. This system could be a K+ channel, or a less specific cation channel, which is coupled, in some way, to a saturable C1uptake system which may function as a H+-Cl- co-transport system (Sanders, 1980b; Jacoby and Rudich, 1980). This system does not appear to be subject to feedback regulation by internal K+ levels, but does exhibit changes in substrate specificity as internal salt levels are altered.

111. REDOX-COUPLED PLASMALEMMA TRANSPORT OF K+ The influence of exogenous electron donors and acceptors has been studied in a variety of plant cell types (Bienfait, 1985; Lin, 1985; Liittge and Clarkson, 1985; Moller and Lin, 1986). Although much of this work has concerned the role of plasmalemma electron transport (redox) systems in Fe” reduction at the root surface (Luttge and Clarkson, 1985; Mdler and Lin, 1986), several studies have addressed the more general topic of the involvement of plasma membrane redox systems in the energetics of solute transport into plant cells. Crane and co-workers (Craig and Crane, 1980, 1981; Misra et al., 1984) investigated the effects of exogenous NADH and ferricyanide on membrane transport and cell growth in suspension cells of carrot. They reported that carrot suspension cells can oxidize exogenous NADH with a concomitant increase in O2consumption (30% stimulation), and that this oxidation results in a stimulation (-60%) of K+ influx. Misra et al. (1984) postulated that a plasma membrane redox system operates in close association with the H+-translocating ATPase of this membrane to exert an influence on K+ transport. In experiments conducted on corn root cortical protoplasts, Lin (1982a,b, 1984) reported that addition of NADH tripled O2consumption, and caused a two- to three-fold increase in K+ influx, a marked stimulation of H+efflux and a moderate (20 mV) hyperpolarization of the E m . Although the extent of the NADH influence was less, Lin reported that similar results were obtained on excised corn root segments. Consistent with the results of earlier workers, Lin (1984) proposed a model in which the oxidation of

130

L. V. KOCHIAN A N D W. J. LUCAS

exogenous NADH would result in the establishment of a transplasma membrane H+ gradient. The relationship between this H+ gradient and K+ influx remains to be established (but see Rubinstein and Stern, 1986). A. INFLUENCE OF EXOGENOUS NADH ON K+ INFLUX

Almost all of the above-mentioned NADH studies were conducted at one particular K+ concentration, usually 0.2 or 1.0 mM. However, Lin (1984) showed that the addition of 1.5 mM NADH to corn root protoplasts had its main effect on K+ influx within the mechanism I range. In these protoplast experiments the apparent K,,, for K+ influx (0.3 mM) was not affected by NADH, but the V,,, underwent a four-fold stimulation. Uptake of K+ over the higher concentration range did not appear to be affected by exogenous NADH. Although in some cases a stimulatory effect of exogenous NADH has been observed on K+ uptake, a clear discrepancy exists between these reports and the findings of Kochian and Lucas (1985) and Thom and Maretzki (1985). Figure 21 illustrates the inhibitory effect that 1.5 mM NADH had on K+ influx into corn root segments (Kochian and Lucas, 1985). At low external K+ concentrations (C0.5 mM)) influx was inhibited by 80%, and a similar degree of inhibition of K+ influx into protoplasts, prepared from sugar cane suspension cells, was reported by Thom and Maretzki (1985). (Leucine, arginine and 3-0-methyl glucose influx into

, I" r

ol

Kt concentration ( mM 1

Fig. 21. Influence of 1.5 mM NADH on @Rb+ influx into high-salt-grown corn root segments. (A similar response was observed in NADH experiments conducted on low-salt corn roots.) Data from Kochian and Lucas (1985).

POTASSIUM TRANSPORT IN ROOTS

131

these protoplasts were similarly inhibited by NADH.) Interestingly, K+ influx into corn roots was also inhibited by the electron acceptor ferricyanide (Kochian and Lucas, 1985; Rubinstein and Stern, 1986). It has been suggested that these two chemicals elicit this inhibition by separate mechanisms, based on the different K+influx kinetics elicited after exposure to NADH or ferricyanide, and the different time courses for recovery from inhibition. B. MEMBRANE TRANSPORT AND THE WOUND RESPONSE

Kochian and Lucas (1985) proposed that exogenous NADH reacts with the outer surface of the plasmalemma to signal a tissue wound response. Hanson and co-workers have shown that corn roots are a particularly sensitive tissue, and that physical handling of the roots, imposition of cold shock etc., can elicit a wound response (Leonard and Hanson, 1972; Gronewald and Hanson, 1980; Chastain and Hanson, 1982; Zocchi and Hanson, 1982). This response is generally characterized by a reduction in K+ influx and a stimulation of efflux, a reduction in H+ efflux, and a depolarization of the Em.A recovery period of about 4 h is usually required before the “wounded” tissue returns to control levels of physiological functioning. It was shown that NADH does not affect K+ influx if the corn root segments have not yet recovered from excision-associated wounding (Kochian and Lucas, 1985). Additionally, if cycloheximideis included in the recovery medium, K+ influx does not return to the control level, and again NADH has no effect on K+ influx. These results indicate that activation of a wound response blocks the NADH effect. Kochian and Lucas (1985) demonstrated that in this state no NADH-stimulated O2consumption can be detected. Only in the recovered state could they measure a 30% stimulation of O2 uptake upon addition of 1.5 mM NADH. Interestingly, although addition of NADH perturbed K+ influx (and efflux), it continued to be oxidized by these perturbed root segments. As mentioned earlier, when the redox system in corn roots (or protoplasts) is stimulated by NADH, net apparent H+ efflux was reported to increase (Lin, 1984). However, Kochian and Lucas (1985) found that NADH elicited a significant decrease in H+ efflux in both washed corn root segments and intact roots (see also Lucas and Kochian, 1987). In some cases an alkalinization of the external medium was observed after addition of NADH, which implies net apparent H+ influx. A similar situation has also been reported for sugar cane protoplasts (Thorn and Maretzki, 1985). Transferring NADH-treated roots to fresh control solution (minus NADH) resulted in an almost immediate recovery of net apparent H+ efflux to pretreatment values (Kochian and Lucas, 1985). This imbalance between recovery of H+ efflux and K+ influx has also been observed with roots that

132

L. V. KOCHIAN AND W. J. LUCAS

are recovering from excision wounding (Gronewald and Hanson, 1982; Kochian and Lucas, 1985). Wounding the root elicits a depolarization of the Em,a response which is in direct contrast to the NADH-induced hyperpolarization of the potential reported by Lin (1982a). In recent studies conducted by Lucas and Kochian (1987), NADH was found to cause a significant depolarization of the corn root Em.In some cases the potential remained in this depolarized state, while in others it repolarized very slowly towards the pre-NADH resting potential. However, consistent with the response of H+ efflux discussed above, removal of NADH always resulted in a repolarization of the potential. Certainly, then, the NADH influence on K+ influx, H+ efflux, and the Em is consistent with the hypothesis that exogenous application of this redox reagent elicits some form of wound response in corn root tissue, while in protoplasts and suspension-cultured cells its application appears to stimulate these physiological processes. C. DEVELOPMENT OF AN INTEGRATED NADH MODEL

Although there is a clear conflict concerning the reported effects of exogenous NADH on plasmalemma transport of K+, there is little doubt that bona fide redox systems are located within this membrane (De Luca et al., 1984; Rubinstein et al., 1984; Buckhout and Hrubec, 1986; Bottger and Liithen, 1986; Macri and Vianello, 1986;Pupillo et al., 1986; Luster et al., 1987). The challenge is to develop a working model of the way in which these putative redox systems interact with the various transport systems functioning within the plasmalemma of plant cells. Such a model would have to account for these reported extremes in tissue response to NADH. In the present context, we will confine our attention to the effects of NADH on H+ and K+ fluxes and the membrane potential. Firstly, the inability of freshly cut, or wounded, corn roots to oxidize exogenous NADH must be explained in terms of the inability of NADH to further inhibit K+ influx, once the K+ transport system has received the wound response “signal(s)”. In analysing the NADH response we have, perforce, assumed that Kf transport occurs via a K+-H+ co-transport system. Zocchi el al. (1983) have shown that cold shock treatment of corn roots increased the phosphorylation of microsomal membrane proteins. These changes were correlated with a decrease in ATPase activity and, since the major changes in phosphorylated proteins occurred within the 92 to 100 kDa range, they concluded that one effect of injury is the blockage of H+ efflux via phosphorylation of the plasma membrane H+-ATPase. This hypothesis is consistent with the observed changes in adenine nucleotide content of corn roots during the injury recovery period (Zocchi et al., 1983). The critical step is the mechanism by which tissue injury is “sensed” by the cells and transduced into protein phosphorylation.

POTASSIUM TRANSPORT IN ROOTS

133

Several possible control systems appear to be tenable. Calcium influx has been shown to increase upon wounding (Zocchi and Hanson, 1982; De Quintero and Hanson, 1982), and the resultant change in the cytosolic Ca2+ level may activate a Ca2+-calmodulin system. One consequence of this activation could be the stimulation of a specific protein kinase. Alternatively, the turnover of phosphatidylinositol may function as a biochemical signal transduction system in roots in a manner analogous to its function in animal tissues (Berridge, 1983;Berridge and Irvine, 1984;Nishizuka, 1984). In animal cells, various external stimuli activate a phospholipase C on the inner plasma membrane surface, which then hydrolyses phosphatidylinositol 4,5-bisphosphate to release myoinositol 1,4,5-triphosphate (IP,) into the cytosol and diacylglyceride (DG) which remains within the membrane. IP3 can stimulate numerous biochemical and biophysical processes, including the release of internally sequestered Ca2+ and activation of certain kinase systems. Recent studies on plant tissues have demonstrated the presence and metabolic turnover of IP, (Boss and Massel, 1985; Morse et al., 1986; Sandelius and Sommarin, 1986; Rincon and Boss, 1987). These reports, along with the central role played by IP, in regulatory phenomena in animal tissues, provided the impetus to incorporate the IP3 cycle as a central component in our NADH “wound” model outlined in Fig. 22. When the alternative agonist membrane receptor (MR) has been stimulated by a wound “signal”, the membrane transmitter (MT) activates the signal transduction system (STS; phospholipase C), and this produces an increase in the level of IP3. The consequences of this increase in IP3are:

1. Activation of a protein kinase which results in phosphorylation of a protein subunit of the H+-ATPase near the Pi release site. This causes a reduction in H+ efflux; the effect will depend on the extent of phosphorylation and the dephosphorylation capability of the cell. 2. Ca2+ release, which may or may not exert its effects on cellular metabolism via Ca*+-calmodulin(CaCM). 3. Direct or indirect effects of IP3 on the gating (G) of ionic channels in both the plasma membrane and tonoplast. An important consequence of this wound activation of MT is that in this state the transmission system is incapable of accepting alternative stimuli. As indicated in Fig. 22, we suggest that an external NADH-oxidizing site can function to stimulate the MT system. Evidence for the presence of this NADH oxidation system comes from a reinterpretation of Lin’s trypsin data (Lin, 1982b; see also Buckhout and Hrubec, 1986). Although the SDSpolyacrylamide gel electrophoretogram of the TCA-precipitated protein obtained from the supernatant of trypsin-treated corn root protoplasts indicates a high level of endogenous protease activity, Lin (1982b) was able to show that the supernatant was capable of NADH oxidation. If endomem-

134

L. V. KOCHIAN AND W. J. LUCAS

EXOGENOUS

NADH

+2H+

I

I

+

H'

POTASSIUM TRANSPORT IN ROOTS

135

brane contamination (mitochondria and endoplasmic reticulum) was absent (but see Komor et al., 1987), this NADH oxidation may be attributed to the presence of a 42 kDa polypeptide released by mild trypsin action (Lin, 1982b, 1984). This protein appears to be able to oxidize NADH and reduce O2 (provided that O2is present); these properties are inconsistent with the NADH redox model proposed by Lin (1984). Note that in the model presented in Fig. 22, exogenous NADH will result in the following: 1. A stimulation of the MT and STS, with the consequences outlined above. 2. A continued stimulation of O2consumption, even though the cell has received a wound signal. 3. Proton consumption at the outer surface of the plasmalemma; depending on the degree to which the total H+-ATPase system is inhibited, this may give rise to alkalinization of the root surface. 4. Continual exogenous NADH oxidation is required to maintain the enhanced IP3 level; removal of the exogenous NADH results in a decline in IP3with the rate being dependent on the levels and/or activation states of the enzymes involved in its breakdown and resynthesis to phosphatidylinositol 4,5-biphosphate. These properties would account for all of our experimental observations on the perturbative influence of exogenous NADH on K+ fluxes in corn roots and corn root segments (Kochian and Lucas, 1985;Lucas and Kochian, 1987). How, then, is it possible to explain the NADH-mediated stimulation of K+ influx observed in corn root segments, corn root protoplasts (Lin, 1982a, 1984) and carrot suspension cells (Misra el al., 1984)? Based on recent reports of plasmalemma-bound electron transport systems, we have included two forms of NADH redox systems in our model (Fig. 22). We propose that these systems would function only when the protoplast, cell or tissue is not in a “wound” state. The protoplast is the easiest system to

Fig. 22. Model summarizing the various effects of NADH on K+ influx into corn roots (or protoplasts). Two transplasmalemma NADH redox systems are shown. The Lin (1984) model (top) utilizes endogenous or exogenous NADH and transports both electrons and H+. The Rubinstein and Stern (1986) model transfers only electrons across the plasmalemma. In this model, generation of the H+ in the cytoplasm, by the oxidation of NAD(P)H, is compensated for by transport out of the cell by the H+-ATPase. In both models, stimulation of K+ uptake would occur by H+-K+ co-transport. We propose that the NADH-induced “wound response” exhibited by corn, is mediated by the coupling of a membrane-bound signal transduction system (STS) to the IPS cycle. Two agonist membrane receptors (MR) are postulated, one being an NADH oxidase [the 44 kDa protein from Lin (1984)], and the other being an alternative receptor mediating other types of wound response. Binding to the MR activates the STS (phospholipase C) via a membrane transmitter (MT). The resultant release of IP3 acts via the activation of protein kinases andor the release of Ca2+, to modify transport proteins (H+ATPase, cation channels) at the plasmalemma and tonoplast. Redrawn from Lucas and Kochian (1988).

136

L. V. KOCHIAN AND W. J. LUCAS

rationalize, because lysis during preparation could release proteases that actually remove the NADH MR protein. Buckhout and Hrubec (1986) also reported that washing their isolated plasmalemma preparation (with or without salt) resulted in a partial loss of NADH reductase as well as oxidase activity. In this state, the protoplasts may oxidize NADH (either exogenously supplied or via the cytosol) to cause an increase in the A& across the plasmalemma by direct transport of H+ (Lin, 1984; Bottger and Luthen, 1986) or by coupling to the H+-ATPase (Rubinstein and Stern, 1986). The IP, recycling system may also vary from cell to cell, or from one tissue to another, as a consequence of the physiological prehistory of the plant system. Since lithium acts as an inhibitor of IP, conversion to inositol and hence its resynthesis to phosphatidylinositol 4,5-bisphosphate, growing cultures in the presence of lithium may reduce the sensitivity of the IP, regulatory system, thereby allowing the NADH redox system to function. Introduction of myoinositol should re-establish the wound response. A further test of the IP,-regulation/wounding hypothesis may be achieved using FC. Chastain and Hanson (1982) have shown that in wounded corn roots FC can override the endogenous control mechanism that inhibits the plasmalemma H+-ATPase. It will be of interest to see whether FC can negate the perturbative influence that exogenous NADH has on K+ influx into roots. Purification of these putative redox systems of the plasmalemma should greatly assist the elucidation of their function. Questions of sidedness, direction of electron transport, and natural substrates (donors and acceptors) are all important issues that need to be resolved. Reconstitution of these proteins into liposomes may provide an unambiguous answer to the question of whether or not they are important in energizing K+ transport into roots.

IV. RADIAL K+ TRANSPORT TO THE XYLEM A symplasmic pathway for K+ movement, from the epidermis to the xylem parenchyma, was first proposed by Crafts and Broyer (1938). In their hypothesis, the cortex was the site where nutrients, like K’, were taken up into the symplasm by active, 02-dependent processes. Movement to the stele was via plasmodesmata, and release into the xylem vessels was by diffusion, or leakage, since the compact tissue of the stele was thought to be low in oxygen. Some 38 years later, Anderson (1976), in his review on solute transport across the root, concluded that, “In the stele, the most common proposal is that the parenchyma either leaks or secretes solutes which then diffuse or are swept along into the vessels.” Figure 23 is a diagrammatic representation of the pathway discussed by Anderson (1976). Based on Anderson’s reviews, it

POTASSIUM TRANSPORT IN ROOTS Epidermis

Cortex

~

Endodermis

near- unity 0 value

w

137

e

low d value

7

External

-

APOPl and water fluxes

probably turgor drlven

Arrows show @ fluxes

Apo$asmic barrier

Fig. 23. Diagrammatic sketch of the root in cross-sectional view showing what Anderson (1976) considered to be the “usually accepted mechanisms of ion and water through-putto the xylem vessels.” Stippled areas represent the cytosol which, in this Anderson model, also extends in the xylem vessel. From Anderson (1976).

would appear that over the period from 1938 to 1976 the only improvement to the Crafts and Broyer hypothesis was the concept that solutes might be secreted into the xylem vessels. However, the mechanistic basis for this “secretion” remained poorly defined. Note also that Fig. 23 suggests that the xylem vessels may be symplasmically connected to the surrounding parenchyma and contain parietal cytoplasm. For those who are quick to be critical of this concept, we should point out that the controversy of whether living xylem vessels are engaged in the release of K+ to the translocation stream is still unresolved (Lauchli et al., 1978; McCully et al., 1987). Clearly, the complexity of the root has made it difficult to obtain a rigorous test of the Crafts and Broyer hypothesis. However, experiments of the past decade have provided some new insights. A. SITE OF K+ ENTRY INTO THE SYMPLASM

1. Epidermis or Cortex? Most current textbooks on ion transport and plant physiology discuss K+ movement into the cortex in terms of two parallel pathways, namely the apoplasmic and symplasmic routes. However, there is still some question as to the relative importance of each component (Anderson, 1976; Lauchli, 1976; Van Iren and Boers-van der Sluijs, 1980; Kochian and Lucas, 1983). Using a theoretical model for Rb’ uptake into barley roots, Bange (1973) found that the model only generated reasonable profiles when transport was

138

L. V. KOCHIAN AND W. J. LUCAS

restricted to the epidermis. The studies of Van Iren and Boers-van der Sluijs (1980) offered experimental support for this hypothesis. These workers investigated K+ uptake by applying the assumption that plasmolysis would sever plasmodesmata, thereby symplasmically isolating each cortical and epidermal cell of a barley root. Autoradiographic localization of 86Rbf, accumulated by the root following plasmolysis, was generally limited to the root periphery. These results suggested that when roots are exposed to low K+ concentrations (