Ion channels and the transduction of light signals

Authors


Correspondence: E. P.Spalding. Fax: 608 2627509; E-mail: spalding@facstaff.wisc.edu

ABSTRACT

Studies of biological light-sensing mechanisms are revealing important roles for ion channels. Photosensory transduction in plants is no exception. In this article, the evidence that ion channels perform such signal-transducing functions in the complex array of mechanisms that bring about plant photomorphogenesis will be reviewed and discussed. The examples selected for discussion range from light-gradient detection in unicellular algae to the photocontrol of stem growth in Arabidopsis. Also included is some discussion of the technical aspects of studies that combine electrophysiology and photobiology.

INTRODUCTION

Unlike an animal with an organ specialized for the purpose, cells in many parts of a plant body and at various stages of development are busy extracting important information from the light environment through photoreceptors and attendant signal transduction chains. Growth and development is altered as a consequence to adapt the plant morphologically and physiologically to the prevailing light conditions, which may include shading by soil or neighbouring plants, changing daylength, or spatial and temporal variations in the amount of radiation useful to photosynthesis ( Kendrick & Kronenberg 1994). This review will be primarily concerned with the evidence that ion channels play a part in bringing about some of the myriad reactions to these inconstant light conditions. Rather than an exhaustive coverage of all known instances, situations in which the evidence is multifaceted and the prognosis bright for future advancements will be emphasized.

As the various examples from algae to Arabidopsis are discussed, a reader may find it useful to keep in mind the general paradigm exemplified by the human eye – absorption of light by the photoreceptor rhodopsin activates an enzyme that degrades a ligand responsible for keeping a particular type of ion channel open ( Stryer 1991). An electrical consequence of the ensuing channel closure propagates along the optic nerve to where the brain interprets the signal. Before addressing whether plant photoreceptors and ion channels also work together to transduce light into cellular responses, I will briefly review what ion channels are, what they do, and how they are studied.

BACKGROUND

Ion channels are integral membrane proteins that facilitate the movement of their ionic substrates across the lipid bilayer, a structure that would otherwise present a formidable kinetic barrier to their movement. Because channels are passive transporters, the direction of transport is dictated by the difference in electrochemical potential of the substrate. This measure of free energy takes into account the contributions of differences in substrate concentration and electric potential across the membrane. The membrane potential (Vm) component of this energy gradient can be very significant in plant cells. As a consequence, the energetically ‘downhill’ direction for an ion may actually be up against a steep concentration difference. (Active transporters such as pumps and cotransporters are fundamentally different from channels in that they use a separate energy source to move a substance ‘uphill’, against its difference in electrochemical potential.) Channels switch between transporting and non-transporting (open and closed) states by a stochastic process referred to as gating, which may be influenced by a variety of factors depending on the type of channel. Certain channels are more likely to open when some perturbation causes Vm to cross a certain threshold value. Such channels are said to be voltage-dependent or voltage-gated, and several of these in plants are known in molecular-level detail ( Czempinski et al. 1999 ). Other channels are much more likely to open after binding a specific ligand such as Ca2+ or, in the case of human vision, cyclic GMP (cGMP). Genes homologous to animal cyclic-nucleotide gated channels are known in plants ( Schuurink et al. 1998 ; Leng et al. 1999 ; Köhler, Merkle & Neuhaus 1999), as is a Ca2+ -activated K+ channel ( Czempinski et al. 1997 ).

Ions flowing through a single open channel at a typical rate of 107 per second create an electric current on the order of 10−12 A. Should some stimulus affect the activity (open-probability) of a significant fraction of a particular type of channel, the resulting change in ionic current across the entire membrane could (i) alter the cytoplasmic concentration of the substrate ion and (ii) alter Vm. Theoretically, both of these consequences could play a role in signal transduction, and demonstrations of one or the other being induced by light may be taken as evidence of a role for channels in the response chain.

The following is a very brief overview of how these electrical and chemical consequences of channel activity are measured. The Vm of plant cells can be measured continuously in real time by inserting a salt-filled glass microelectrode into a cell and recording the electric potential difference (voltage) between it and a reference electrode located extracellularly ( Fig. 1a). The value of Vm for the plasma membrane of plant cells is generally between –150 and –200 mV ( Fig. 2). In some instances, a non-invasive method employing surface-contact electrodes can be used to monitor changes in Vm ( Fig. 1b). A shift in Vm to more negative values is a hyperpolarization whereas a positive shift is a depolarization ( Fig. 2). Proof that a change in Vm is due to a change in the activity of a particular type of ion channel is best obtained by patch-clamping the membrane ( Fig. 1c). The development of this powerful technique revolutionized the field by enabling channel activity to be studied in the presence of controlled chemical and electrical gradients, with submillisecond resolution, and in either a whole cell membrane or a small patch containing even a single channel molecule ( Neher 1992).

Figure 1.

Electrophysiological methods used to study photosensory transduction. (a) A salt-filled microelectrode is inserted into a cell. The preparation may be an isolated cell, an intact seedling as shown, a leaf, or a pulvinus, etc. A grounded reference electrode must make contact with the preparation, typically through an electrolyte bathing solution or saline agar (checkered block). The electric potential difference between the two electrodes is the membrane potential (Vm), which is measured by an electrometer (E). A computer arranged to administer the light treatment and continuously record Vm will register changes induced by a light treatment with high resolution (e.g. Spalding 1995). This is a relatively non-invasive procedure. It has been employed in many studies of light signal transduction, such as ( Racusen 1976; Nishizaki 1987; Elzenga et al. 1995 ; Parks et al. 1998 ; Stahlberg & Van Volkenburgh 1999). (b) Changes in Vm but not the absolute value of Vm can be measured with electrodes that contact the surface of a plant organ via saline agar. The recorded signal is often referred to as surface potential or Vs and light-induced changes in it reflect changes in Vm of the illuminated cells. This non-invasive technique can be used even on the tiny seedlings of Arabidopsis ( Newman & Briggs 1972; Hartmann 1975; Cho & Spalding 1996). (c) The patch-clamp technique allows the direct study of ion-channel activity by measuring the current created as ions flow through open channels. Typically, protoplasts are prepared from the tissue of interest. A clean pipette containing a defined electrolyte is pressed against the membrane and coerced to seal to it by suction. The opening and closing of a channel residing in such a patch can be studied by monitoring the change in current flowing across the patch with a voltage-clamp amplifier. Light treatments may be applied to the cell to test the effects on the channel activity. This approach has been used in studies of light sensing ( Lew et al. 1990 ; Cho & Spalding 1996; Elzenga & Van Volkenburgh 1997). The patch of membrane containing the channel may be excised from the cell, permitting studies in precisely controlled ionic and electrical gradients. Agents that may regulate channel activity as second messengers can be tested by applying them to the cytoplasmic side of excised patches. The whole-cell configuration, useful for studying all the active channels in a membrane at once, may be created by rupturing the patch of membrane to make the interior of the pipette continuous with the cytoplasm of the cell. The effect of photomorphogenic light stimuli on active transporters such as pumps can be studied using this configuration ( Assmann et al. 1985 ; Serrano, Zeiger & Hagiwara 1988).

Figure 2.

Depolarization of the plasma membrane induced by blue light in an Arabidopsis hypocotyl. Vm was measured continuously as described in Fig. 1a. From a resting potential in darkness of –188 mV, a 20 s pulse of blue light delivered via a computer-controlled shutter induced a transient shift in Vm of approximately 40 mV before returning to near its initial resting potential. The response is transient even when induced by continuous blue light, although the kinetics will vary depending on fluence rates. The depolarizing phase is largely, if not completely, due to the activation of anion channels. The inset shows that the response is clearly underway within 3 s of the shutter opening to begin the light treatment. (These are previously unpublished results of M.H. Cho and E.P. Spalding.).

Changes in concentration of the transported ion can be measured electrophysiologically by inserting ion-selective microelectrodes into the appropriate cellular compartment. Measurements in the cell wall, cytoplasm, and vacuole have been accomplished but all aspects of this technique are challenging and these days it is used less frequently than optical methods that rely on ion-specific fluorescent dyes or chemiluminescent proteins such as aequorin. All of the above techniques for measuring voltages, currents, and ion concentrations have been used in studies of light-signal transduction, as will be shown below.

MOTILE GREEN ALGAE

An exciting example to take up first is the green alga Chlamydomonas. Light gradients guide the swimming of this unicellular organism into conditions optimal for its photosynthesis. Its eyespot organelle contains a photoreceptor, the chemical nature of which was first glimpsed when retinal was shown to restore wild-type phototaxis to a blind Chlamydomonas mutant ( Foster et al. 1984 ). A retinal-binding protein was subsequently purified from the eyespot and the corresponding cDNA was shown to have regions of sequence similar to invertebrate opsins ( Deininger et al. 1995 ). Although ‘chlamyopsin’ differs topologically from animal opsins, which interact with signal-transducing G-proteins through an arrangement of seven membrane-spanning helices, it is fair to say that this alga senses light with a photoreceptor that shares structural features with human rhodopsin. Do ion channels function downstream of chlamyopsin as they do in rhodopsin-based vision? The answer is a resounding yes. Faster even than the closing of cGMP-gated channels in the human retina, the absorption of light by chlamyopsin triggers the opening of a Ca2+ -permeable channel ( Litvin, Sineshchekov & Sineshchekov 1978; Harz & Hegemann 1991). The inward Ca2+ current activates so quickly (less than 50 μs lag time) that the channel may be directly induced to open by a light-induced conformational change in the receptor ( Holland et al. 1996 ). It has even been speculated that the receptor and channel activities are contained in the same molecule. Expressing chlamyopsin in a Xenopus oocyte and testing for retinal-dependent currents induced immediately by light could directly test this intriguing idea (for a precedent, see Khorana et al. 1988 ). Studies of closely related algal species indicate that other channels also participate in the mechanism ( Braun & Hegemann 1999), with a light-activated G-protein serving an intermediary role in the slower components ( Calenberg et al. 1998 ) as in animal vision. Because the full range of modern experimental approaches may be employed to study these cells, there can be little doubt that the details will be resolved. How relevant will these details be to light-sensing mechanisms in land plants? The placement of Chlamydomonas and its allies in the green algal clade least related to the presumed progenitors of land plants (Graham & Wilcox 2000) does not augur well for a high degree of conservation. Nonetheless, careful searching of genomic databases for opsin-like sequences is warranted because there are rapid blue-light responses in Arabidopsis for which no receptor has yet been identified ( Parks, Cho & Spalding 1998).

BROWN ALGAE

Zygotes of the brown algae Fucus and Pelvetia provide another example of single-celled systems with a light-sensing mechanism that appears to involve channels. Asymmetric illumination with blue light causes the rhizoid to emerge from the shaded side of the spherical zygote ( Kropf 1997; Brownlee & Bouget 1998). The earliest-detected event following the light treatment is the focusing of inward ionic currents on the region in which the rhizoid will later initiate. This change in current distribution could be produced by a clustering of channels at the rhizoid-initiation site, by the local activation of channels that are uniformly distributed, or a combination of both. The preponderance of evidence indicates that Ca2+ carries at least a share of the current entering the site of rhizoid initiation ( Robinson & Jaffe 1975; Robinson 1996; Love, Brownlee & Trewavas 1997), resulting in locally higher cytoplasmic Ca2+ concentrations detectable with fluorescent dyes and confocal microscopy ( Pu & Robinson 1998). A blue-light treatment that establishes polarity in the zygotes has been shown to increase the concentration of cGMP ( Robinson & Miller 1997), the ligand that activates the Ca2+-permeable channels in our retinas. This finding, plus the demonstration that zygotes contain the same retinal isomer as chlamyopsin ( Robinson et al. 1998 ), raises the possibility that the photosensory mechanism in these brown algae is similar to that operating in motile green algae. Because the fucoid zygote and its rhizoid can be studied with electrophysiological techniques and is a well-established cell biological system, studies of its light responses could reveal important and generally applicable details.

LAND PLANTS

Blue light and stem growth

The hypocotyls of etiolated seedlings extend rapidly in darkness but are inhibited abruptly by blue light as part of the developmental shift toward expansion of photosynthetic organs ( Cosgrove 1994). High resolution measurements of growing Arabidopsis seedlings have shown that two genetically separable phases of inhibition comprise this blue-light response ( Parks et al. 1998 ). The first is mediated by an unknown UVA/blue-specific photoreceptor after a lag time of approximately 30 s. The second, dependent upon the CRY1 photoreceptor, begins 45–60 min after the onset of blue light. A very large membrane depolarization always precedes the onset of even the rapid phase of growth inhibition, evincing a role for ion channels in this photomorphogenic response ( Spalding & Cosgrove 1989). The lag time for the depolarization can be less than 10 s and the response is transient; Vm returns to near its initial value within minutes whether the response is evoked by a pulse or continuous radiation. An example of the membrane depolarization induced by blue light in an etiolated Arabidopsis hypocotyl is shown in Fig. 2. Mutants lacking the cryptochrome 1 photoreceptor (CRY1) (hy4) display a depolarization that is only 30% of the wild-type magnitude, indicating that CRY1 along with at least one other photoreceptor quickly acts to alter ionic currents across the plasma membrane ( Parks et al. 1998 ). A current carried by Clflowing out of the cell as well as inhibition of the proton pump was considered a likely cause of the depolarization for a variety of pharmacological and thermodynamic reasons ( Spalding & Cosgrove 1992). Using the patch-clamp technique, a Clchannel residing in the plasma membrane of an etiolated Arabidopsis hypocotyl cell was identified and indeed found to be activated by blue light ( Cho & Spalding 1996). Blocking the channel with micromolar concentrations of the channel-blocker NPPB (5-nitro-2-(3-phenylpropylamino)-benzoic acid) blocked the depolarization and the onset of the CRY1-dependent phase of growth inhibition ( Cho & Spalding 1996; Parks et al. 1998 ). Thus, it seems that the excitation of CRY1 by blue light leads to the activation of anion channels at the plasma membrane within seconds, resulting in membrane depolarization ( Fig. 3). It is somewhat surprising that the initial ionic events appear not to be causally linked to the rapid phase of inhibition that they most closely precede.

Figure 3.

Initiation of membrane depolarization induced by blue light (BL) in wild-type and hy4 seedlings measured with surface-contact electrodes. The Vs signal is of opposite polarity compared to Vm ( Fig. 2). The depolarization is affected by the mutation to a statistically significant degree approximately 11 s after the onset of irradiation (asterisk), indicating that excitation of the CRY1 receptor contributes early to the activation of plasma membrane anion channels. (This is a reanalysis of data in Parks et al. 1998 ).

The rapid activation of anion channels at the plasma membrane by CRY1 may seem at odds with the nuclear localization of this photoreceptor when expressed in onion epidermal cells ( Cashmore et al. 1999 ). However, Guo et al. (1999) found that the majority of CRY1 was not present in the nuclear fraction of Arabidopsis plants. The rapidity of its action at the plasma membrane indicates that a functionally significant fraction of CRY1 is present in the cytoplasm of Arabidopsis hypocotyl cells. Alternatively, a CRY1-dependent signal could emanate from the nucleus to influence channel gating within the first few seconds of illumination. The second cryptochrome in Arabidopsis (CRY2) appears to be more completely partitioned to the nucleus than CRY1 ( Guo et al. 1999 ; Kleiner et al. 1999 ). It has not yet been determined whether plants lacking CRY2 are impaired with respect to the activation of anion channels, or at what point in time the cry2 mutation affects inhibition of hypocotyl growth.

The activation of anion channels by blue light also manifests itself as a transient shrinkage of protoplasts prepared from growing stems ( Wang & Iino 1997, 1998). As a proxy for the depolarization, the shrinkage response may offer some experimental advantages. For example, protoplasts are more readily treated with pharmacological agents and ion-specific fluorescent dyes than are intact hypocotyls. Exploitation of the shrinkage response may reveal details of the depolarization mechanism that would be more difficult to obtain by electrophysiology.

An important question to answer is how CRY1 activates anion channels within seconds. Patch-clamp experiments on membrane patches isolated from Arabidopsis hypocotyl cells demonstrated that increasing cytoplasmic Ca2+ to concentrations greater than 1 μM activates the anion channel ( Lewis et al. 1997 ). But does that mean that blue light activates the anion channel in vivo by increasing cytoplasmic Ca2+? This question was addressed by monitoring cytoplasmic Ca2+ photometrically in transgenic plants expressing aequorin, a jellyfish protein that emits photons at a Ca2+-dependent rate. However, no changes in cytoplasmic Ca2+ were detected before, during or after channel-activation induced by blue light ( Lewis et al. 1997 ). An increase in Ca2+ induced by cold shock was readily observed, as well as its expected effect on the anion channel. Thus, Ca2+ can activate the anion channel in vitro and in vivo but blue light appears to activate it some other way. Redox reactions initiated by electron transfer from an excited CRY1 flavoprotein to an unknown receiver molecule ( Cashmore et al. 1999 ), perhaps the anion channel itself, would seem to be a possible mechanism. The observation that the activity of Arabidopsis K+ channels can be affected by their redox state adds some plausibility to this speculation ( Spalding et al. 1992 ).

Another blue light response in Arabidopsis seedlings that appears to depend upon anion channel-activation is the accumulation of anthocyanins. Blocking the channel with NPPB inhibited anthocyanin accumulation induced by blue light ( Noh & Spalding 1998). Experiments showed that NPPB did not prevent the up-regulation of key biosynthetic enzymes that normally results from blue-light treatment, nor did the drug inhibit the vacuolar transport step of the accumulation process. Rather, the data indicate that blue light acts through anion channels to post-translationally modify one or more of the earliest steps in the pathway to increase the biosynthetic capacity ( Noh & Spalding 1998). Induction of the pathway by blue light and its modification by an anion-channel-dependent mechanism both appear necessary for the normal accumulation of anthocyanins.

The evidence that anion channels play an early and important transduction role in the blue-light responses of seedling stems is matched in certain respects by findings made in expanding leaves of pea and bean. White light activates a Ca2+-dependent anion channel in the plasma membrane of growing leaf mesophyll cells ( Elzenga & Van Volkenburgh 1997), consistent with the membrane depolarization and ion fluxes also observed in these cells upon illumination ( Elzenga, Prins & Van Volkenburgh 1995; Shabala & Newman 1999). However, the effects on leaf and hypocotyl growth are opposite. Light promotes the expansion of leaf cells by ‘loosening’ the cell wall through acidification ( Stahlberg & Van Volkenburgh 1999) but inhibits growth and ‘tightens’ cell walls in hypocotyls ( Cosgrove 1994). No explanation of how anion-channel activation affects cell-wall yielding properties in any system has been advanced yet, although separation of the two processes by 45–60 min in hypocotyls ( Parks et al. 1998 ) indicates that the connection is not direct.

Candidate genes for light-activated anion channels

A deeper understanding of the role of anion channels in the control of growth by light may not come until molecular genetic approaches can be included in studies of the channel. At present, the genetic basis of the light-activated anion channel has not been determined, and no anion channel mutants are known. Plants clearly have genes related to at least three families of genes encoding Cl channels in animals. The present challenge is to attribute a channel known only by electrophysiological parameters to a gene known only by sequence. There follows a summary of genes known to encode Cl channels for the purpose of introducing present candidates for the light-activated anion channel of plants.

The large superfamily of proteins known as the ATP-binding cassette (ABC) transporters has a well-known Cl channel as one of its members, the CFTR channel mutated in people suffering from the terrible disease, cystic fibrosis. Plant ABC transporters of the MRP subclass display a limited stretch of sequence similar to CFTR’s characteristic R domain but no gene obviously belonging to the CFTR subfamily has been reported yet in plants ( Rea et al. 1998 ). Expression of an Arabidopsis MRP (AtMRP2) in Xenopus oocytes (an expression system frequently used to study the activities of cloned ion channels) did not produce a new Cl conductance in the oocyte membrane (E.P. Spalding, unpublished observations). That negative result, the probable localization of AtMRP1 and AtMRP2 to the tonoplast, and their demonstrated glutathione-S-conjugate pumping activity ( Lu, Li & Rea 1997; Lu et al. 1998 ) argue against an MRP-type ABC transporter being the light-activated anion channel at the plasma membrane.

Another ABC transporter that has had Cl channel activity ascribed to it is P-glycoprotein, also known as MDR ( Valverde et al. 1992 ). The MDR group of ABC transporters is better known for pumping drugs out of tumour cells, rendering them resistant to therapeutic chemicals. For a while, an MDR molecule was considered bifunctional, capable of active drug pumping as well as passive channel-like transport of Cl ( Higgins 1995). This interpretation was revised to the view that mammalian MDR regulates the activity of a separate Cl channel rather than forms a channel itself ( Hardy et al. 1995 ). However, well-documented evidence, inconsistent with any connection between MDR and Cl channels, engendered a fully dissenting view that is presently holding sway ( Morin et al. 1995 ; Tominaga et al. 1995 ). Nonetheless, Arabidopsis definitely contains MDR-like molecules and presumably they transport something ( Dudler & Hertig 1992). Overexpression and antisense expression of one MDR homolog (Atpgp1) affects hypocotyl length specifically in light-grown plants. Over-expression created taller hypocotyls, the opposite was observed in antisense plants, and the effect was greater under red light than blue light ( Sidler et al. 1998 ). The mechanism by which expression level of an MDR affects hypocotyl growth rate in light is unknown, but the possibility that it involves changes in ion permeability of the plasma membrane deserves further attention.

The CLC class of membrane proteins indisputably functions as Cl channels in animal cells, where some members are located in the plasma membrane and others in intracellular membranes ( Jentsch et al. 1999 ). Genes clearly in the CLC family are present in plants although to date it has not been convincingly demonstrated that any actually functions as a Cl channel ( Lurin et al. 1996 ; Hechenberger et al. 1996 ). Any plant CLC-type gene that can be shown to encode a Cl channel and is found to reside at the plasma membrane in planta would be a candidate for the light-activated channel(s). The electrophysiological properties of the heterologously expressed channel should be similar to those of the channel activated by light if the two channels are one and the same. Also, both should be similarly blocked by micromolar concentrations of NPPB. Even better evidence that a particular gene encodes the light-activated channel would be a demonstration that a T-DNA insertion allele, perhaps obtained by reverse genetics ( Krysan, Young & Sussman 1999), impairs the light-induced electrical responses.

Osmoregulation in motor cells

Volume changes in motor cells are responsible for the reversible movements of pulvini and stomata, and occur as an osmotic consequence of altered solute fluxes. They are demonstrably affected by light. For example, blue light induces swelling of guard cells ( Assmann 1993) and pulvinar extensor cells ( Satter et al. 1988 ) causing stomata to open and Samanea leaflets to lift, respectively. Channels conduct the K+ influx that helps drive the volume changes but it is not strictly correct to assign these K+ channels a signal transducing role in the movement response to blue light. The evidence to date indicates that blue light activates the plasma membrane H+-ATPase, producing an outward current that hyperpolarizes the membrane ( Assmann, Simoncini & Schroeder 1985; Shimazaki, Iino & Zeiger 1986; Amodeo, Srivastava & Zeiger 1992; Kinoshita & Shimazaki 1999). This tips the K+ electrochemical potential gradient more steeply inward, which could increase the rate of passive K+ uptake through voltage-dependent channels that are more likely to open as the membrane hyperpolarizes. Mutant analysis indicates that the photoreceptor responsible for the blue-light-induced stomatal opening uses the carotenoid zeaxanthin as a chromophore ( Zeiger & Zhu 1998).

To reverse the process and have K+ leave the cell, the direction of the gradient in K+ electrochemical potential must switch from inward to outward. For this to happen, Vm must change. The reigning idea is that the opening of Cl channels and inhibition of the H+-ATPase depolarizes the membrane to an extent that makes K+ efflux a passive process mediated by outward-rectifying channels ( Ward, Pei & Schroeder 1995).

The guard cell has proven to be a powerful system for studying the control of transport processes at the plasma membrane as well as hormone signal transduction involving ion transport. With well-defined light responses and the advantages of Arabidopsis genetics, this electrophysiologically tractable cell should continue to be fertile ground for studies of early events in light-signal transduction.

Phytochrome

The early suggestion that phytochrome exerts its effects by first altering the permeability of the plasma membrane to ions ( Hendricks & Borthwick 1967) received promising support from subsequent demonstrations of red-light-induced changes in membrane potential ( Newman & Briggs 1972; Racusen & Galston 1980; Racusen & Galston 1983), and the curious Tanada effect ( Tanada 1967). Not a great deal has been learned in the meantime about phytochrome-mediated effects on transport processes at the plasma membrane, despite their potential importance as primary transducing steps. However, intriguing evidence turns up frequently enough to keep alive the idea that phytochrome exerts some of its effects by altering ion channel activity. The gametophytic stage of the moss Physcomitrella patens, filamentous in form and obscure in nature, offers perhaps the best-studied example. Irradiation with red light promotes the formation of side-branch initials that develop into the leafy, haploid structures most recognizable as a moss ( Cove et al. 1978 ). This bud induction process requires the presence of Ca2+ in the medium ( Schumaker & Dietrich 1997). A series of studies showed that red light, almost certainly acting through phytochrome, induces an inward Ca2+ current within a few seconds that begins to depolarize the membrane while simultaneously increasing cytoplasmic Ca2+ concentration ( Ermolayeva et al. 1996 ; Ermolayeva, Sanders & Johannes 1997). Whether as a result of the depolarization, the increase in Ca2+ , or perhaps some other event, anion channels are activated and this furthers the depolarization ( Ermolayeva et al. 1997 ). An increase in the K+ permeability of the membrane occurs during the depolarization, presumably due to the voltage-dependent opening of K+ channels ( Johannes, Ermolayeva & Sanders 1997). This would contribute to the membrane repolarization, which occurs during the next few minutes with kinetics that depend upon the length and/or fluence of the inductive treatment ( Ermolayeva et al. 1996 ). Pharmacological evidence indicates that these substantial though transient effects on ionic currents are mechanistically related to the developmental changes that occur over the ensuing 3 days. Treating the filamentous cells with agents shown to inhibit Ca2+ and Cl channels in the protonemal preparation inhibited bud formation ( Ermolayeva et al. 1997 ). Physcomitrella patens is amenable to an array of physiological, electrophysiological and molecular genetic techniques. Indeed, it is becoming established as a model system for studying plant development because of some practical advantages it offers over Arabidopsis, not least of which is homologous recombination at a frequency reasonable enough to make reverse genetics feasible ( Reski 1999). Thus, one may look forward to interdisciplinary studies aimed at understanding how photoreceptors influence the activity of ion channels to achieve information transduction.

Will the results from such studies with moss be applicable to flowering plants? Probably yes, in some instances, because similar though smaller depolarizations induced by red and reversed or inhibited by far-red light have been detected in oat coleoptiles ( Newman & Briggs 1972; Racusen 1976) and the flexor cells of pulvini ( Racusen & Satter 1975). The phytochrome-mediated changes in Vm in oat coleoptiles may be an electrical manifestation of a transduction mechanism that also brings about the unrolling of the primary leaf wrapped within the coleoptile. This unrolling response in cereal seedlings is the result of differential cell expansion across the juvenile leaf. It requires the presence of Ca2+ in the medium ( Viner, Whitelam & Smith 1988), and may be studied at the cellular level by measuring the light-induced swelling of protoplasts prepared from the primary leaf. The protoplast swelling response also requires Ca2+ in the medium ( Bossen et al. 1988 ) and is inhibited by agents known to block Ca2+ channels in other systems ( Tretyn, Kendrick & Bossen 1990). The interpretation of these results is that photoconversion of Pr to Pfr resulted in a change in Ca2+ channel activity at the plasma membrane, and the resulting increase in cytoplasmic Ca2+ triggered cell swelling through an unspecified process. The red-light-induced increase in cytoplasmic Ca2+ predicted by this orthodox model was indeed detected in etiolated wheat leaf protoplasts by measuring fluorescence from a Ca2+-specific dye ( Shacklock, Read & Trewavas 1992). Furthermore, the same study demonstrated that artificially increasing cytoplasmic Ca2+ by photolytically releasing it from caged forms caused protoplast swelling in the absence of red light.

The second messenger role of Ca2+, strongly supported by the above evidence and in the chloroplast rotation response of the filamentous green alga Mougeotia ( Roux, Wayne & Datta 1986), may not be ubiquitous. For example, there have been no reports of changes in cytoplasmic Ca2+ concentration in Arabidopsis hypocotyl cells in response to red light, even though the phytochrome system profoundly affects their growth with a rapidity consistent with such a mechanism ( Parks & Spalding 1999). Neither is there any electrophysiological evidence that phytochrome signalling in hypocotyls involves changes in ion fluxes. For example, red light does not change Vm in Arabidopsis hypocotyls (E.P. Spalding, unpublished observations). This is unfortunate because the biophysical approaches one would use to elucidate the details would be powerfully augmented by the available array of Arabidopsis mutants affected in phytochrome signalling.

If a photomorphogenic stimulus does not induce a change in Vm, a role for ion channels in the transduction pathway is not ruled out. The lack of an effect on Vm does mean, however, that approaches other than the electrophysiological may be more productive. The seedling hypocotyl may be a case in point. Although no changes in Ca2+ concentration have been detected in the cytoplasm of hypocotyl cells, micro-injecting Ca2+ into phytochrome-deficient tomato hypocotyls mimicked phytochrome-mediated effects on genes encoding components of the photosynthetic apparatus responses ( Neuhaus et al. 1993 ; Bowler et al. 1994 ). The finding that artificially increasing cytoplasmic calcium could trigger a subset of phytochrome responses led the authors to suggest that photoconversion of phytochrome normally produces an increase in cytoplasmic Ca2+ concentration to effect downstream responses. Changes in cytoplasmic Ca2+ can result from changes in flux across the plasma membrane, but also and perhaps even usually, they occur as a result of release from internal stores such as the vacuole. A Ca2+ current across an endomembrane such as the tonoplast may not directly change Vm recorded by an intracellular microelectrode, reconciling the micro-injection results with the lack of a red-light-induced voltage change in hypocotyls. Nonetheless, acceptance of the mechanism indicated by the micro-injection studies should depend upon a demonstration that the cytoplasmic concentration of Ca2+ in hypocotyl cells indeed changes in response to red light.

The same experimental approach of micro-injecting potential second messengers into hypocotyl cells of the phytochrome-deficient aurea mutant of tomato pointed to a role for cGMP in the photocontrol of chalcone synthase expression, a key enzyme in the anthocyanin biosynthetic pathway ( Bowler et al. 1994 ; Wu et al. 1996 ). Should phytochrome phototransformation actually lead to an increase in the levels of cGMP in hypocotyls, the cyclic nucleotide-gated channels identified in Arabidopsis and barley ( Schuurink et al. 1998 ; Leng et al. 1999 ; Köhler et al. 1999 ) could be among the targets for this ubiquitous second messenger. This possibility returns the reader back to the beginning, to the vertebrate vision paradigm.

Some of the more exciting recent developments in phytochrome signalling point to a mechanism of action that does not have an obvious requirement for changes in ionic currents or concentrations. For example, after light-dependent migrations of phytochromes A and B into the nucleus ( Kircher et al. 1999 ; Yamaguchi et al. 1999 ), these photoreceptors may influence gene expression fairly directly by binding to a transcription factor ( Ni, Tepperman & Quail 1999). Although these impressive advances further solidify the view that changes in gene expression are key to photomorphogenesis, they do not reduce the possibility that changes in ionic fluxes are also necessary. Instead, it may be useful to consider physiologically generated ionic changes as integral aspects of at least some mechanisms that culminate in altered gene expression. For example, ionic events of the sort discussed above may predispose the cytoplasm for the nuclear import of phytochrome, or for the better translation of light-induced transcripts into altered growth and development. The next decade of research may well reveal that physiological processes such as ionic currents are required to co-ordinate, complement, augment, and perhaps in some instances cause those changes in gene expression that bring about normal photomorphogenesis.

ACKNOWLEDGMENTS

This work was supported by grants from the National Aeronautics and Space Administration/National Science Foundation Network for Research on Plant Sensory Systems (IBN-9416016) and the National Science Foundation (IBN-9974585).

Ancillary