Evidence for K ÷ channel control in Vicia guard cells

The Plant Journal (1995) 8(2), 187-198

Evidence for K ÷ channel control in Vicia guard cells coupled by G,proteins to a 7TMS receptor mimetic

Fiona Armstrong t and Michael R. Blatt* Department of Biological Sciences, University of London, Wye College, Wye, Kent TN25 5AH, UK

Summary

To explore the involvement of a class of seven-transmembrane-span (TTMS) receptors in cellular signalling, a synthetic analogue (mas7) of the amphipathi¢ tetredeca- psptide mestoperan was used to mimic hormonal stimulus in guard calls of Vicia faba. The ability for mes7 to substitute for an activated receptor complex was assayed by the effect on guard cell ion channel activities in the absence of any hormonal stimulus. Currents carried by inward- (/K,in) and outward- (/K,out) rectifying potassium channels were determined under voltage clamp conditions before, during, and after exposure to mas7. The dominant effect of mes7 was to inactivate/K, in within 30 sec of application. By contrast, /K,out was largely unaffected under these conditions. The effect of mas7 on IK, in was both concentration- and voltage-dspendent. At any one clamp voltage, mas7 inactivation showed Michenlian beheviour, with a mean/~ of 0.05 _+ 0.02 I~M at -240 mV. Increasing rues7 concentration also shifted the voltage for half-maximal activation of the current negative, with 0.5 IJM mes7 effecting a -13 _+ 2 mV displacement and lengthening the halftime for activation of the current by up to threefold. By contrast, the non-amphipathic analogue of mes7, masCP, had no appreciable effect on the steady-state current or its activation kinetics; nor was the poly-cation polylysine able to substitute for mes7 in its action on the K ÷ channels. Application of the non-hydrolyseble analogue of GDP, GDP-~-S, either by iontophoresis or by diffusion from the microelectrode, effectively blocked mes7-inducad inactivation of/K,in. These, and additional results provide in vivo evidence for the involvement of G-protein-linked 7TMS receptors in the regulation of membrane transport in a higher plant call.

Introduction

Classical concepts of cellular signalling have emerged from studies of animal cells and appear to be remarkably conserved throughout both eukaryotes and prokaryotes.

Received 17 October 1994; revised 5 April 1995; accepted 9 May 1995. *For correspondence (fax +44 1233 813140). tPresent address: Pflanzenphysiologie, Universit~it Salzburg, Hellbrunner- strasse 34, A-5020 Salzburg, Austria.

Typically, signals such as hormones are coupled to a receptor protein at the external surface of the plasma membrane, with subsequent signal transduction via associ- ation of the receptor/agonist complex with a heterotrimeric guanine nucleotide binding protein (G-protein). The G- protein is activated by exchange of bound GDP for GTP and dissociates into cc and ~ subunits, which propagates the signal leading to the regulation, among others, of ion channels.

In plants, evidence to support the existence of G-proteins similar to those identified in animals is gathering. As well as higher molecular-weight heterotrimeric G-proteins, low molecular-weight monomeric GTP-binding proteins have been found, similar to the Ras, Rho and Ypt/Rab type involved in cellular 'housekeeping' functions (Bednarek et al., 1994; Matsui et al., 1989; Terryn et al., 1993). Plant G-proteins identified at the gene level include GPA1 from Arabidopsis ( M a e t al., 1990), TGA1 from tomato (Ma eta/., 1991), both encoding Ga subunits, and ~-subunit-like proteins from Chlamydomonas and Arabidopsis (Schloss, 1990; Weiss et al., 1994). To date, G-proteins of the heterotrimeric class have been implicated in a diverse array of physiological processes crucial for plant growth and development. From responses to fungal elicitors (Vera- Estrella et al., 1994), hormonal signalling (Zaina et al., 1990), phototransduction (Bossen et al., 1990; Romero and Lam, 1993; Warpeha et al., 1991), secondary messenger regulation (Dillenschneider et al., 1986), ionic homeostasis (Allan et al., 1989), and control of ion channels (Fairley- Grenot and Assmann, 1991; Li and Assmann, 1993), to enzyme regulation (Legendre et al., 1993) all have been shown to be influenced by G-proteins. (For a review of GTP-binding proteins in plants, see Terryn et al. (1993)).

Evidence for G-protein function in vivo comes from a relatively small number of model cell systems, including the stomatal guard cell. Guard cells control stomatal aperture through changes in cellular turgor, processes that are driven by fluxes of potassium and other ions, and these occur in response to environmental stimuli such as light, CO2 and phytohormones. By far the best characterized responses are those to auxin and abscisic acid (ABA). Perception of either hormone initiates a signalling cascade which entails changes in cytoplasmic-free [Ca 2÷] ([Ca2+]i) (Blatt et al., 1990; McAinsh et al., 1990) and pH (PHi) (Blatt, 1992; Blatt and Armstrong, 1993) and converges on two classes of potassium channels, at least one anion conductance, and the H÷ATPase (Blatt and Thiel, 1993).

In the case of ABA, current through a class of inwardly rectifying K ÷ channel (IK,in) is reduced, contributing to a

187

188 Fiona Armst rong and Michael R. B/art

pattern of changes in transport that precipitates stomatal closure. This channel activates upon sufficient membrane hyperpolarization, and normally provides a major pathway for K + influx during guard cell swelling and stomatal opening. IK, in is sensitive to pH i (Blatt, 1992; Blatt and Armstrong, 1993) and to changes in [Ca2+]i, with increasing [Ca2+]i inactivating current (Lemtiri-Chlieh and MacRobbie, 1994). In contrast, the outwardly rectifying potassium channels (IK,out) of these cells, responsible for K + efflux during stomatal closure, appear to be remarkably insensitive to Ca 2÷ (Lemtiri-Chlieh and MacRobbie, 1994). The latter channels are highly sensitive to pH i , with acid-going pH i serving to all but eliminate K + channel activity, and alkalini- zations increasing the current (Blatt and Armstrong, 1993). During ABA-induced stomatal closure, IK,ou t conductance is enhanced, but can be blocked by acid-going phi (Blatt and Armstrong, 1993), indicating a second-messenger role for cytoplasmic pH.

Both classes of K ÷ channels are also subject to regulation by phosphatases which may play a role in hormonal control of guard cell transport. In the case of the outward rectifier, a 1/2A-type phosphatase was implicated in channel regulation from studies with okadaic acid (Thiel and Blatt, 1994), and Luan et al. (1993) have shown that the [Ca2+]i-depend - ence of IK,in can be affected by Ca2+-dependent protein phosphatases. A phosphatase may also contribute to ABA sensing of guard cells in stomatal regulation. The gene product of abil, one of several loci responsible for reduced ABA responsiveness and a wilty phenotype in Arabidopsis thaliana, is now known to encode a 2C-type (serine/threon- ine) protein phosphatase that includes a putative Ca 2÷- binding domain (Leung et al., 1994; Meyer et al., 1994). Thus, an effective signalling machinery is in place within the guard cell allowing a graded coupling of external signals to volume changes and stomatal aperture.

More elusive is the nature of the putative receptors that coordinate and initiate signals to propagate within this signalling network. Almost exclusively, animal G-proteins are coupled to a family of seven-transmembrane-span (7TMS) receptors (Dohlman, 1991). Such receptors link agonist binding at the external surface of the membrane to intracellular G-proteins and subsequent signal generation. A common architecture for 7TMS receptors isolated thus far is emerging, and their ubiquitous nature underlines their importance in cellular signalling cascades. Indirect evidence exists to support the concept of G-protein linked 7TMS receptors in plant signal transduction (see Terryn et al., 1993). However, to date, firm evidence linking 7TMS receptors to G-protein function and cellular physiology has been lacking.

Recently, White et al. (1993) showed that the mastoparan analogue mas7 could stimulate GTP-7-S binding to pea and maize microsomal membranes. Mastoparan and its synthetic analogue mas7, are toxins commonly used as

diagnostics for the existence of G-protein-coupled 7TMS receptors. Mastoparan is a tetradecapeptide and was ori- gina!ly isolated from wasp venom (Hirai et al., 1979). Upon transfer from aqueous to lipid environment, mastoparan undergoes a conformational change, assuming a helical formation (Higashijima et al., 1984). In this form, mastoparan has structural and functional similarities to the third intracellular loop of 7TMS receptors (Higashijima et al., 1988; Weingarten et al,, 1990), a cytoplasmic domain that is thought to be important in determining G-protein interaction. Mastoparan effectively mimics this domain and stimulates GTPase activity (i.e. exchange of GDP for GTP) independent of the usual agonist/receptor interaction. This study utilizes mas7 to probe receptor/G-protein coupling in guard cells of Vicia faba. We show that mas7 effects a selective reduction of current through IK, in, but not through IK.out, and that this response can be ascribed to G-protein- mediated modulation via secondary messengers.

Results

mas7 and K ÷ channel currents

To explore the effect of mas7, membrane currents were recorded under voltage clamp at regular intervals before, during and after treatments with the toxin. Figure 1 shows currents recorded from one guard cell, and details the current carried by the inwardly rectifying K ÷ channels (IK, in). The current was apparent upon clamping the cell from a holding potential of -100 mV to potentials between -120 and -250 mV and then to a tailing potential o f - 3 0 mV (Figure la). The K + current was characterized by a slow rise in clamp current magnitude, with halftimes of about 100 msec near -200 inV. Steady-state characteristics for IK, in were determined from the trajectories, as shown in Figure l(a), by first subtracting the instantaneous current recorded at the beginning of each voltage step from the steady-state current, measured at the end of the step. Figure l(b) shows the results of this current subtraction and highlights IK, in dependence on clamp voltage. Figure l(b) also shows that the effect of mas7 was to reduce IK, in in a concentration-dependent manner. By contrast, voltage- clamp measurements at voltages positive of -100 mV and steady-state data for whole-cell currents demonstrated that mas7 had no consistent effect on the outwardly rectifying K ÷ channels (IK, out), nor did it affect background membrane conductance (Figure 2; see also Figure 1). For the cell shown in Figure 1, 0.1 tiM mas7 reduced IK,in by 46% at -250 mV, while 0.5 tiM and 1.0 llM effected reductions of 69% and 80%, respectively, at this voltage. Inactivation by mas7 was Michaelian at any one voltage (Figure 3), and comparable results were obtained in all 15 other cells treated with mas7. Pooled results from all guard cells

a

/ oe_.

o , - - ~

Guard cell K ÷ channel modulation by a 7TMS receptor mimetic 189

V/mY ~ cm'2 -250 -2OO -150 - IO0 '

o.e 1

I ] o4 -40

• 1 0 2

. 0.0 4 0 V/mY

I/1~ ¢m "=

V/mV

50

q- - t - 50

, - 2 5

. . 5 0

• -75

' -100

Figure 2. Whole-cell current response to mas7. Steady-state current are from the same guard cell as in Figure 1, gathered using a bipolar staircase pulse protocol to scan through the accessible voltage range before (0) and during (V) exposure to 0.5 ItM mas7. The dashed line indicates the background current estimated by linear extrapolation from points between EK (-71 mV) and -120 mV (see Blatt, 1992). Currents above (positive-going) and below (negative-going) this line correspond to IK,out and Ix,in, respectively. Note the absence of an appreciable change in Ix,ou t or in background current during challenge with mas7.

Figure 1. Inactivation of IK,in b y mas7. Data from one Vicia guard cell bathed in 5 mM CaZ+-MES with 10 mM KCI. Cell parameters: surface area, 1.8 × 10 -s cm2; volume 4.9 pl; stomatal aperture, 7 pro. (a) Current trajectories of the inward rectifier (Ix, in) recorded before (e) and during exposure to 0.1 (•), 0.5 (~') and 1.0 (I-I) pM mas7. Clamp cycle voltages (above): conditioning, -100 mV; test (n = 8), -120 to -250 mV; tailing, -30 mV. Scale: vertical, 250 mV or 25 ~A cm-2; horizontal, 500 msec. Note the absence of any increase of background (instantaneous) current at the beginning of each test clamp step. (b) Steady-state current-voltage relations for IK,in taken from currents recorded at the end of the voltage pulse scan in (a) after subtracting the instantaneous background currents recorded less than 5 msec into the test pulse at each voltage. Data shown are for control (e), 0.1 (~7), 0.5 (V) and 1.0 (I-I) IIM mas7. Lines are calculated from the fitted conductances using equations (1) and (2) (see text). Inset: plot of steady-state conductance as a function of clamp voltage for control (O), 0.1 (V), 0.5 (V), and 1.0 pM mas7 (I-1). Conductances fitted jointly to the data in (a) and (b) using equations (1) and (2) (see text) yielded a common apparent gating charge, 6, of 1.43 _+ 0.08. Vertical lines mark voltages for half-maximal activation, Vlrz: -163 (O), -175 (V), -177 (T) and -183 mV (D).

y ie lded an apparent Ki for mas7 inact ivat ion of 0.05 ___ 0.02

IIM at - 2 4 0 mV (Figure 3).

masT, IK, in voltage-dependence and kinetics

One of the most str iking effects of mas7, apart f rom the

reduct ion of IK,in w a s a dramat ic increase in the hal f t imes for current act ivat ion (Figure 4). Add ing 0.5 IIM mas7 increased the act ivat ion hal f t ime by approx imate ly threefold. The increase was concentra t ion-dependent on a stat- istical basis, but largely independent of c lamp vol tage as could be determined by normal iz ing the act ivat ion half-

t imes in mas7 to their cor responding values in the control (Figure 4, inset). Half t imes for deact ivat ion, as calculated

f rom tail current measurements, were not s igni f icant ly affected by mas7 (Table 1).

The effect of mas7 on IK, in act ivat ion impl ied a signi f icant inf luence on the vo l tage-dependence for the steady-state current. Analysis of the currents, such as shown in Figure

1, suppor ted this predict ion. Figure l (b , inset) shows the

conductance gK, in calculated f rom the data in Figure l (a) and plot ted as a funct ion of c lamp vol tage. Conductances

were de termined assuming an ohmic dependence on membrane vol tage so that

gK = IK/(V-EK) (1)

As shown, the data have been f i t ted jo int ly to the Bol tzmann funct ion

gK.max gK = (2)

1 + eSF(Vl/2- V)/RT

where gK,max is the conductance at saturat ing negat ive

membrane vol tage, ~ is the apparent gat ing charge, Vt/2 is the characterist ic ha l f -maximal act ivat ion vol tage and F, R

and Thave thei r usual meanings. Satisfactory f i t t ings were

obta ined only when V1/2 and gmex were a l lowed to vary between mas7 t reatments, and y ie lded a concentrat ion-

dependent shift in the vol tage for ha l f -maximal act ivat ion of -11 _+ 2.8 in 0.1 I~M, - 1 3 +- 1.8 for 0.5 pM, and - 1 9 -+ 3.1 mV for 1.0 IIM mas7. Thus, mas7 not only reduced the max ima l K + channel current, but effect ively displaced

channel gat ing to increasingly negat ive vol tages. Quali tat- ively s imi lar results were obta ined on a cell-by-cell basis

190 Fiona Armstrong and Michael R. Blatt

U

0,4

u

U u 0,4 u u t.o i i IBBTI0~ -=no -==

v/mv Figure 3. Inactivation of IK,in as a function of mast concentration at a clamp voltage of -240 mV. Data points are means + SE for all 15 cells challenged with mas7 alone (0), and following cytoplasmic loading with nucleotides from electrodes containing 5-10 mM GDP-I~-S (~7, n=13) and ADP-~-S (17, n=5). Data for mas7 treatments (0) were fitted to a Michaelis-Menten function, giving an apparent Kv2 of 0.05 -+ 0.02 ~M and maximum inactivation of 80%. Roughly equivalent Fntings were obtained when the maximum inactivation was constrained to a value of 100%.

from all 15 other V. faba guard cells challenged with mas7 (data not shown).

Uncoupling K ÷ channel inactivation with GDP-~-S

There are principally three possible explanations for the effect of mas7 on IK, in. The first explanation is that mas7 indeed acts as a receptor-mimetic and, by stimulating GDP/ GTP exchange of a G-protein, initiates a signalling cascade that culminates with the inactivation of IK.in. The second explanation invokes a direct interaction of mas7, by virtue of its high density of positive charges, with the mouth of the K ÷ channel to effect a block of IK,in. Finally, mas7 might act independently of any G-protein activation to influence downstream signalling elements that affect K ÷ channel gating. Each explanation makes a number of experiment- ally testable predictions.

In the first case, preventing GDP/GTP exchange and G- protein activation should effectively block the mas7-evoked signal and, hence, IK, in inactivation. Equally, non-amphipathic analogues of mas7 would be expected to lack physiological activity. As a positive test for coupling via G- protein activation, the non-hydrolysable analogue of GDP, GDP-~-S, was used which could be expected to lock putative G-protein intermediates in the inactivated state by irreversible binding to the G¢ subunit. It was reasoned that by curtailing G~ activation, any mas7 signal transmitted in this manner would also be blocked. Guard cells were loaded with GDP-13-S, either by passive diffusion into the cytoplasm from the voltage-following barrel of the standard, double-barrelled microelectrodes or by iontophoretic injection of the nucleotide. In the latter case guard cells were impaled with four-barrelled microelectrodes and two

4oo

J el

200

Rgure 4. Halftimes for the activation of IK, in plotted as a function of voltage. Activation halftimes (tat2) calculated from single-exponential fittings to current traiectories (cf. Figure I) before treatment (O) and in the presence of 0.5 I~M mas7 (I'), following pretreatment with GDP-~S (V) and subsequent challenge with 0.5 p.M mas7 (Q). Note that GDP-J3-S had a negligible effect on activation kinetics but significantly reduced the effect of subsequent mas7 treatments. Data for 0.5 ~M mas7 alone differ significantly from all other treatments at the 95% confidence level, inset: voltage-independence of mas7 action on tl/2. Data replotted relative to the control and cross- referenced by symbol. Data are means _+ SE of 15 (!1', 0) and 13 (V, [Z) cells, respectively.

Table 1. Halftimes (t1/2) for deactivation of /K, in as a function of mas7 and GDP-~-S treatments a

Treatment tv2 (msec)

Control 6.3 -+ 0.6 (4) 0.5 I~M mas7 7.7 -+ 1.6 (4) GDP-~-S (~1 mM) 12.5 -+ 2.1 (4) GDP-~-S (~<1 mM) + 0.5 tIM mas7 19.9 -+ 1.9 (7)

aHalftimes determined from tail currents recorded at -30 mV following conditioning clamp steps to -240 mV. Means -+ SE of (n) trials taken from fittings to a single exponential (Blatt, 1992). Analyses were limited to the first 150 msec to avoid contributions from IK,out at this voltage.

barrels were used for injection, one containing nucleotide in a 10 mM K+-acetate carrier electrolyte.

Figure 5 summarizes the results from one guard cell loaded with GDP-~-S by iontophoresis. Current trajectories were recorded under voltage-clamp before and during injection of GDP-~-S, and during subsequent exposure to 0.5 pM mas7. Prior to injection, GDP-~-S was retained in the microelectrode by applying a 0.25 nA backing current, and nucleotide injection was effected with a 0.25 nA current over 2 min. It is apparent from the current trajectories (Figure 5a) and the steady-state current-voltage relations (Figure 5b) that injection of GDP-~-S itself had no measurable effect on IK,in. However, after loading the cell with GDP- B-S, iK, in was insensitive to mas7 treatment. Comparable results were obtained in all 13 guard cells loaded with GDP-13-S prior to mas7 application. For seven guard cells loaded by diffusion, partial and temporary declines in IK, in

were observed when mas7 was added in the first 15 rain following impalements. In each case, however, washing


a

L

v . , , ,

V/mV -250 -200 -150 /- I/pA cm ~

-100 • / . / / ---=P

-10

-20

-30

-4O

Figure 5. Protection of IK. in from mas7-evoked inactivation by prior cytoplasmic loading with GDP-~-S. Data from a single Vicia guard cell bathed in 5 mM Ca2÷-MES with 10 mM KCI and challenged with 0.5 ~tM mas7. Cell parameters: surface area, 2.1 x 10 -s cm2; volume 5.1 pl; stomatal aperture, 12 p.m. Cytoplasmic loading was by iontophoresis with 0.25 nA over 2 min. Maximum estimate for final [GDP-~-S] = 1 mM given a cytoplasmic volume of 0.5 pl and assuming a transport number of 0.03, which accounts for the ratio of GDP-~-S to total charged species in the electrode. (a) Current trajectories for IK,in following GDP-~-S injection (0), in the presence of 0.5 pM mast (V) and after washing mas7 from the bath (,). Clamp voltages: conditioning, -100 mV, test (n = 8), -120 to -240 mV, tailing, -30 mV. Scale: vertical, 250 mV or 10 p~ cm-2; horizontal, 500 msec. (b) Current-voltage relations for IK,in recorded at the end of the voltage pulse scan in (a) after subtracting the instantaneous background currents recorded less than 5 msec into the test pulse at each clamp voltage. Data cross-referenced to (a) by symbol.

mas7 f rom the bath resulted in comple te recovery of IK,in and later chal lenges w i th mas7 showed the current to

develop comple te insensit iv i ty to the toxin. The protect ive effect of GDP-I~-S is in stark contrast to the effect of mas7

alone, to which /K, in showed a dramat ic reduct ion w i th v i r tual ly no recovery in every case (see Figures 1-3). We

noted that GDP-~-S alone effected an increase in the hal f t ime for IK, in deact ivat ion (Table 1), a l though it had no signif icant effect on act ivat ion hal f t imes (Figure 4).

However, cy top lasmic loading w i th GDP-I3-S tempered the

effect of mas7 on the kinetics for IK,in act ivat ion (Figure 4). By contrast, the current retained its sensit iv i ty to mas7 in f ive guard cells pre loaded wi th ADP-I~-S, demonst ra t ing

the nucleot ide specif ici ty of the response (Figure 6). (A calculat ion of the nucleot ide load for the exper iment in Figure 5 gave a value of I mM, given a cytop lasmic vo lume of 0.5 pl and assuming a t ranspor t number of 0.03 for

a

I

o " ~ - ~ ' ~ ~ . . , " . - ' : ~ . ~ L

b

V I m v -250 -200 -150 -100 t

: , : - - - - ' , /> '

-20

.-40

Figure 6. MasT-evoked inactivation of /K, i n following prior cytoplasmic loading with ADP-13-S. Data from a single Vicia guard cell bathed in 5 mM Ca2+-MES with 10 mM KCI and challenged with 0.5 IIM mas7. Cell parameters: surface area, 1.7 x 10 ~ cm2; volume 4.5 pl; atomatal aperture, 9 Izm. Cytoplasmic loading was by iontophoresis with 0.30 nA over 2.8 min. Maximum estimate for final [ADP-~-S] = 2 mM calculated as in Figure 5. (a) Current trajectories for IK,in following ADP-I~-S injection (O), in the presence of 0.5 p.M mas7 (V) and after washing mas7 from the bath (T). Clamp voltages: conditioning, -100 mV, test (n = 8), -120 to -240 mV; tailing, -30 mV. Scale: vertical, 250 mV or 20 pA cm-2; horizontal, 500 msec. (b) Current-voltage relations for IK,in recorded at the end of the voltage pulse scan in (a) after subtracting the instantaneous background currents recorded less than 5 msec into the test pulse at each clamp voltage. Data cross-referenced to (a) by symbol.

GDP-J~-S in the carr ier electrolyte. This value represents a theoret ical m a x i m u m and assumes, in the t ranspor t num-

ber, a charge-weighted equivalence in mobi l i t ies between GDP-I~-S, K ÷ and the acetate anion. Hence, the t rue nucleo-

t ide load was probab ly closer to 100 IIM.)

masCP, TEA and K ÷ channel block

Al though the abi l i ty of GDP-~-S to protect IK,in f rom mas7 inact ivat ion argues in favour of G-prote in-dependent chan-

nel modula t ion, the results do not rule out addi t ional direct effects of the o l igopept ide on the channels. If mas7 effected

channel block directly, interact ing at the channel mouth (Obermeyer et aL, 1994), then s imi lar ly charged o l igopep-

t ide analogues of mas7 wou ld be expected to block IK, in, i rrespective of any amphipathic i ty . As a check against

a V I m V

.25o b

-200 -150

192 Fiona A rms t rong and Michael R. Blatt

~ cm "2 .IOD

I i / /

: -10

-20

.30

-4O

- . 5 0

- -60

b v/mv -25O -2O0 -150

I I p A a n "=

-100 ; F / /d

- -25

.-El(}

• -75

-100

- -125

• -150

Figure 7. Insensitivity of IK, in to masCP, and lack of IK,in protection from mas7 by concurrent TEA exposure. Data from two separate Vicia guard cells bathed in 5 mM Ca2+-MES with 10 mM KCI. Cell parameters in both: surface area, 2.0 x 10 -5 cm2; volume 5.1 pl; stomatal aperture, 10 I~m. (a) Steedy-state current-voltage relations of the inward rectifier (IK,in) recorded before (e) and during (~7) exposure to 0.1 pM masCP. (b) Steady-state current-voltage relations of the inward rectifier (/K,in) recorded before (0), during (~7) exposure to 30 mM TEA plus 0.1 I~M mas7, and following mas7 and TEA wash from the bath (T). The guard cell was treated first with 30 mM TEA to give partial block of /K, in and then subsequently challenged with 0.1 pM mas7 in the presence of TEA.

charge interactions between the K + channels and mas7, guard cells were also challenged with masCP. MasCP is an analogue of mas7 and like mas7, comprises 14 amino acids but with an additional positive charge (Leu substituted at position 6 with Lys). As a result, masCP is predicted not to form an amphipathic helix at the lipid interface and it does not mimic the third cytoplasmic loop of the 7TMS receptors (Higashijima et al., 1988). In vitro masCP, unlike mas7, is ineffective at promoting GTP-7-S binding to plant membrane fractions (White et al., 1993). Thus, masCP was expected to lack physiological activity in G-protein activation, but might substitute for mas7 in direct, charge- dependent channel block.

When masCP was tested over the same concentration range as mas7, masCP failed to evoke any consistent change in IK, in. Indeed, in three of 15 cells tested IK, in was enhanced 10-15% in the presence of 0.5 IIM masCP compared with currents recorded from the same cells prior to treatments. Figure 7(a) shows the effect of 0.1 pM

masCP--a concentration close to the Ki for mas7--on the current-voltage relations of one Vicia guard cell.

Guard cells were also challenged with mas7 in the presence of TEA which blocks K ÷ channels by binding within the mouth of the channel pore. We reasoned that if mas7 interacted with the K ÷ channel by entering the channel mouth to block the K + current, then TEA might compete with mas7 for binding and prevent inactivation. So in these experiments, the guard cells were first exposed to 10 mM TEA and mas7 was added subsequently in the presence of TEA. Channel current was assayed for the effect of mas7 after first washing the peptide and then TEA from the bath.

Figure 7(b) shows the effect of TEA on the current- voltage relations of a single Vicia guard cell. TEA quickly and effectively reduced IK, in. Experiments with TEA alone indicated that wash-out was equally rapid (see also Blatt, 1992), taking only 1-2 rain for IK, in tO recover fully (results not shown). This is in contrast to the inactivation of IK, in by mas7, which was largely irreversible (see Figure 1 and accompanying discussion). Application of mas7 and TEA together effected no further reduction of IK, in compared with TEA alone. However, upon washing out mas7 and then TEA from the bath, the current did not recover (Figure 7b). Thus, mas7 could effect inactivation of IK, in even when the channel mouth was occluded by TEA. Again, quantitatively similar results were obtained in six other cells, These, and the results with masCP argue strongly against a direct interaction between the K + channels and mas7.

Discriminating mas7/G-protein from non-specific interactions in K ÷ channel control

The third explanation for mas7 action on IK, in entails non- specific interactions with other signalling elements that control the K + channels. Mastoparan and related analogues are known to affect Ca 2÷ binding with calmodulin (Ohki eta/., 1991) and to alter the activities of C-type phospholip- ases (Drobak and Watkins, 1994; Wallace and Carter, 1989). These actions are typically evident only at concentrations around 10 pM and above. So, the very low K i for mas7 inactivation of IK, in--and ability of GDP-~-S to protect IK,in from mast inactivation--argues against such indirect effects as the dominant mode of mas7 action. None the less, the data cannot rule out additional secondary effects, and we note that some Ca2+-dependent protein kinases are sensitive to polycations even at nanomolar concentrations (see Roberts and Harmon, 1992).

As a check against secondary interactions of these kinds, guard cells were exposed to low molecular-weight (3-5 kDa) polylysine alone, or to mas7 together with neomycin sulphate, an inhibitor of phospholipase C (Faddis and Brown, 1993; Legendre eta/., 1993). Figure 8 shows the


a

L o v -~ - . . . . =r| ,l=

b I/pA ¢m "2

VImV -250 -200 -150 -10G

-40

- -60

Figure 8. Insensitivity of IK,in to polylysine. Data from a single Vicia guard cell bathed in 5 mM Ca2+-MES with 10 mM KCI and challenged with 1 pM polylysine (3-5 kDa MW). Cell parameters: surface area, 2.3 x 10 -s cm2; volume 5.5 pl; stomatal aperture, 11 pm. (a) Current trajectories for IK.in before addition (0), in the presence of 1 ~M polylysine (V) and after washing polylysine from the bath (T). Clamp voltages: conditioning, -100 mV; test (n = 8), -120 to -240 mV; tailing, -30 mV. Scale: vertical, 250 mV or 25 ~ cm-2; horizontal, 500 msec. (b) Current-voltage relations for /K, in recorded at the end of the voltage pulse scan in (a) after subtracting the instantaneous background currents recorded less than 5 msec into the test pulse at each clamp voltage. Data cross-referenced to (a) by symbol.

current trajectories and steady-state current-voltage relations from one guard cell challenged with 1 pM polylysine. On its own, the polycation had no appreciable effect on s teady-s ta te IK,in or on its activation kinetics. In three of five cells tested, higher concentration (5-10 pM) of polylysine led to reductions less than or equal to 30% in the steady-state current. However, in no case was there a measurable increase in the halftimes for activation of IK, in and in every cell the response was reversible on washing the polylysine from the bath. Neomycin sulphate, at a concentration of 100 ~LM, also failed to block the action of subsequent exposures to 0.5 ~LM mas7 in five guard cells. In the latter case, mas7 evoked a mean inactivation of 54 - 6% at -240 mV, comparable to the action of the toxin on its own (see Figure 3).

sulphate would argue against a pivotal function of phospholipase C- and inositol trisphosphate-mediated release of intracellular Ca 2+, alternative pathways for Ca 2+ entry to the cytoplasm have been mooted (MacRobbie, 1992). Less still is known of the origins for changes in pH i , although it is clear that physiological stimuli such as auxin and ABA influence the cytoplasmic-free [H+]. Increasing [Ca2+]i (Schroeder and Hagiwara, 1989) and alkaline-going pH i both inactivate /K, in (Blatt and Armstrong, 1993), and either of these changes could mediate its inactivation by mas7. Thus, the question arises whether changes in either PHi or [Ca2+]i could transduce the signal downstream of a mas7/G-protein interaction. To investigate the role of these second messengers downstream of the mas7 stimulus, experiments were designed to buffer changes in [Ca2÷]i and pHi. Again, we reasoned that buffering any mas7- evoked changes should curtail any signal transmission via these ionic intermediates.

To buffer pHi, guard cells were treated with butyrate which imposes an acid load, effectively clamping pH i to a more acid value. Previous studies had shown that acid loading stimulates IK, in but otherwise has no effect on its competence to respond to exogenous stimuli (Blatt and Armstrong, 1993). Thus, if mas7 inactivated IK, in through pH i alkalinization, preloading with butyrate would be expected to relieve its effect on IK,in. Figure 9(a) shows the response of IK, in in one guard cell to 10 mM butyrate and to subsequent addition of 0.5 pM mas7 against this background of butyrate. Acid-loading with butyrate effected an increase in IK, in when compared with the control. However, addition of mas7 led to inactivation of the current none the less, both compared with the control and with the butyrate treatment, and the inactivation of IK,in was roughly equivalent on a relative basis when compared with guard cells challenged with mas7 alone (see Figure 1). Similar results were obtained in experiments with five other guard cells, giving a mean relative inactivation of IK,in at -250 mV of 0.50 -+ 0.18 in butyrate compared with the value of 0.54 with mas7 alone (see Figure 3).

To buffer cytoplasmic-free [Ca2÷], three guard cells were preloaded with 50 mM EGTA, introduced by diffusion from the voltage-following barrels of two-barrelled microelectrodes (Blatt et aL, 1990). Loading EGTA had a negligible effect on IK,in in each case (see also Blatt et aL, 1990). Figure 9(b) shows that a comparable degree of inactivation was observed in IK,in on adding 0.5 pM mas7, even after prior loading with EGTA for 20 min (compare with Figure 1), and the current failed to recover on washing mas7 from the bath.

How is the mas7 signal transduced?

Calcium and pH have established second messenger roles in guard cells and, while the experiments with neomycin

Discussion

While inward-rectifying K ÷ channels in guard cells of V. faba had previously been shown to be subject to regulation by

a

V/mV -250 -200 -150


II1~1~ c m "z

-1oo : / / /

.20

-4O

-6o

-0O

-1oo

V/mV -250 -200 -150

I/I~A an ~ -100. /

• /

-25

-50

-'/5

-100

-125

-150

Figure 9. Effect of buffering intracellular pH and calcium prior to application of mas7. Data from two Vicia guard cells, bathed in 5 mM CaZ+-MES with 10 mM KCI. Cell paramaters (a): surface area, 1.8 x 10 -s cmZ; volume 4.7 pl; stomatal aperture, 6 p.m; (b): surface area, 1.9 x 10 -s cm2; volume 4.9 pl; stomatal aperture, 8 lim. (a) Current-voltage relations for/K,i, recorded before (T) and after (V) acid loading by application of 10 mM Na+-butyrate and with subsequent addition of 0.5 liM mas7 in the presence of 10 mM Na+-butyrate (e). The acid load effected an increase of IK,in (see also Blatt, 1992; Blatt and Armstrong, 1993), but mas7 reduced the current beyond both the control and butyrate- treated levels. (b) Current-voltage relations for IK,in recorded before (e) and during (V) exposure to 0.5 llM mas7 after first loading the guard cell with 50 mM EGTA to buffer cytoplasmic-free [CaZ+]. Loading in this experiment was by diffusion over a period of 17 rain prior to challenge with 0.5 p.M mas7, and data gathered after washing mas7 from the bath are included (T). Compare with Figure 1.

putative G-proteins (Fairley-Grenot and Assmann, 1991), their nature and potential receptor coupling has remained unknown. Our data now strongly suggest that in vivo the K ÷ channels are coupled through G-proteins to a class of 7TMS-like receptors. Three principal observations bear on this conclusion: (i) mas7 inactivation of the K + channels was specific to /K,in, the toxin being without consistent or significant effect on IK,out or background membrane conductance (Figures 1 and 2); (ii) cytoplasmic loading with GDP-I~-S, which could be expected to uncouple putative G-protein signals by irreversibly binding and inactivating G~, was effective in protecting IK, in against mas7-evoked inactivation and could not be substituted with equivalent ADP-I~-S loads (Figures 5 and 6); (iii) an inactive mas7 analogue, masCP, and polylysine were without effect on

the current while pretreatments with TEA, neomycin sulphate and cytoplasmic Ioadings with the Ca 2+ buffer EGTA failed to prevent the inactivation of IK,in in subsequent mas7 exposures (Figures 7-9). The first point rules out any non-specific detergent-like action of mas7 on the plasma membrane (cf. Okumura eta/., 1981; Weidman and Winter, 1994); such effects would be visible as increases in 'leak' current against the very low background membrane conductance near the free-running potential (typically 10-20 llS cm -2 or 1-3 G~, comparable with background conductances of patch recordings from plant membranes). The third point argues against the predominance of secondary interactions of mas7 with downstream elements, including Ca2+-dependent protein kinases (Roberts and Harmon, 1992), phospholipase C and Ca2+-related events (Drobak and Watkins, 1994; Ohki et al., 1991; Wallace and Carter, 1989). However, the strongest argument for a mas7/ G-protein-mediated pathway is the specificity for and GDP- I~-S antagonism of K + channel inactivation (Figures 5 and 6).

Indeed, these results underline the parallels found in the signalling strategies utilized by both animals and plants. Thus far, studies which have investigated the roles of G- proteins in guard cell signalling have generally concen- trated on putative targets, and not on possible receptor types. Although we cannot provide conclusive evidence that the receptor has seven transmembrane domains, and can only allude to the nature of the region interacting with the G-protein(s), evidence from animal cells suggest that the 7TMS receptor class is one which is predominantly linked to heterotrimeric G-proteins. To date, there is only one report of a receptor with a single TMS domain (Okamoto et al., 1990), which none the less retained the crucial regions thought to be important for G-protein interaction. Our data imply that even if the functional region, which mas7 mimics, is not part of 7TMS receptor per se, it must be adequate both for interaction with the G-protein, and for subsequent signal transduction in the absence of agonist.

Nature of Ga with which mast interacts

In animal cells, mas7 is thought to interact preferentially with members of the Gao/Gai family of heterotrimeric G- proteins (Higashijima et al., 1988, 1990; Weingarten et al., 1990), and mastoparan has been shown to have a lower sensitivity for G~q than for Gag/Gal. Mas7 is thought to stimulate Gai, which would result in a decrease of Ca 2+ via phospholipase C-dependent IP 3 hydrolysis. Previously, Gai as well as Gos(or Gaq) classes had been implicated in control of IK,in, on the basis of sensitivity to pertussis toxin (PTX) and cholera toxin (CTX) (Fairley-Grenot and Assmann, 1991). Both toxins inactivate IK, in, and it was suggested, therefore, that at least two classes of G~ might control the K + channels, the first inactivated by PTX (Gai-

Guard cell K + channel modulat ion by a 7TMS receptor mimet ic 195

like) and the second stimulated by CTX (Gas-like). Following this model, mas7 would be predicted to stimulate a Gai- type protein, causing a decrease in Ca 2÷ and thereby enhancing IK,in- Our results suggest that this is not the case. As yet, interactions determining receptor/Ga subunit specificity are not well defined, especially in plant cells. Therefore the question remains open as to the nature of the native G~ with which mas7 interacts. The two heterotrimeric G¢ subunits so far identified in plants have been assigned to the G~i class on the basis of sequence homology (Ma et al., 1990, 1991). However, the situation is further complicated by the lack of PTX-ribosylation sites in the two cloned Ga subunits. Thus, it is to be anticipated that the structure/function relationships determined by individual G-proteins may not be conserved between plants and animals, and only tentative suggestions can be made at this time as to the nature of the Ga with which mas7 interacts.

An alternative explanation for the apparent Ca 2+- (and, perhaps PHi-) independence of the masT-evoked response is that the effect is mediated via GI~ subunits, and not by a Ga-linked pathway. Evidence is gathering which suggests that signalling functions are not restricted to G¢ subunits, and that GI~ may have roles in signal transduction beyond merely stabilizing the Ga subunit (Brown, 1993). Thus, the situation can be envisaged whereby receptor stimulation of the G-protein leads to both the generation of 'active' G~ subunits, and also GI~ subunits, which are then free to interact with effector molecules. One of the few roles assigned thus far to GI~ subunits in animal cells is the regulation of K + channels (Reuveny et a1.,1994; Wickman et a/., 1994), and it will be of great interest to determine whether they have an analogous role in plant signalling.

Downstream coupling to IK, in

One of the most striking effects of mas7 on IK,in was on the activation time constant, tl/2 (Figure 4) and a corresponding shift in the half-activation voltage (Figure 1). This effect is consistent with a reduction in the open probability of the channel (Po) and cannot be explained simply in terms of changes in conductance or numbers of available channels. Indeed, Wu and Assmann (1994) have recently described the effects of GDP-I~-S and GTP-7-S, at the single channel level, on the kinetics of IK, in activation and deactivation. They found that the effects of the nucleotides were wholly attributable to alterations in Po, rather than any change in channel conductance or permeation. It is of interest that a similar behaviour was found in the effect of mastoparan on the maxi Ca2+-activated K+channel of bovine chromaffin cells. Glavinovic et aL (1992) resolved the effect of mastoparan at the single channel level and found that the peptide reduced Po in a concentration- dependent manner. However, the effect of mastoparan was

complex, having a more pronounced effect on certain defined subcomponents of opening.

Although it is apparent that mas7 can inactivate IK, in in a manner analogous to that observed with hormones such as abscisic acid and auxin, the pathway of signal transduction is not clear from these studies. Given the effect of mas7 on the kinetics of the channel opening, and the fact that pHi and [Ca2+]i also effect changes in gating of IK, in (Blatt, 1992; Schroeder and Hagiwara, 1989), it was somewhat surprising that an unequivocal role could not be found for either. Our results suggest that alkaline-going changes in pHi are unlikely to contribute to mas7-induced signalling. Even when the cytoplasmic pH of the guard cells was clamped with butyrate, mas7 treatments resulted in the inactivation of IK, in. Nor could IK,in inactivation be prevented by buffering changes in [Ca2+]i . In this latter instance a simple interpretation is more difficult, however. A similar approach for buffering changes in guard cell [Ca2÷]i has been adopted previously with success (Blatt et al., 1990). None the less, EGTA is known to be poor in chelating rapid changes in [Ca2+], and is often unable to contend with the rapidity of agonist-induced Ca 2+ transi- ents. There is evidence for a mastoparan-evoked, IP3-gated calcium spike in animal cells (Mousli et al., 1990; Okano eta/., 1985), and mastoparan also stimulates IP 3 production in Chlamydomonas reinhardtii (Quarmby et al., 1992), suggesting that its effect on calcium release is not confined to animal cells.

The lack of putative ionic signalling intermediates raises the tantalizing possibility that mas7 signal transduction might be membrane delimited, such that no cytoplasmic signalling intermediates are required for interaction between a G-protein subunit and the K + channel (see Brown, 1993, for a review). Recent evidence suggests that a membrane-delimited pathway for G-proteins in K ÷ channel coupling does exist in guard cells (Wu and Assmann, 1994). Indeed, there are parallels between the mechanism of inactivation of IK, in by mas7 and the effect of guanine nucleotides on the channel. A membrane- delimited pathway of G-protein/channel coupling would be unlikely to necessitate gross changes in [Ca2+]i or pHi, as localized changes might be sufficient to effect transduction. However, the evidence at present does not confirm any direct physical coupling between the G-protein and the channel.

Yet another interpretation of mas7 action on IK,in, and one which could support coupling via GI~ ~, invokes mas7 stimulation of phospholipase A 2. The resulting generation of lipid messengers could explain K ÷ channel modulation without recourse to the ionic messengers. In animal cells one of the few assigned signalling roles of GI~, along with the control of K + channels, is to activate a cytoplasmic phospholipase A 2 (PEA2; Kim et a/., 1989; Kurachi eta/., 1989; Reuveny et a/., 1994). Stimulation of PLA 2 would


generate lysolipids and fatty acids such as arachidonic

acid. Indeed, arachidonic acid and other fatty acids have

been recently implicated in the modulat ion of guard cell K + channels (Lee et al., 1994), and certain species of lysolipids may exert similar control on the channels (Arm- strong and Blatt, unpublished observations). Furthermore,

in this context mastoparan has been shown to stimulate PEA 2 in Zucchini hypocotyls in relation to auxin-induced H+ATPase activity and growth. Regulation of all PLA 2

isozymes thus far identified has shown a strong dependence on [Ca2+]i, but a recent report suggests that the

[Ca ]i-independence (Nalefski enzymaticact iv i tycan show 2+ •

et al., 1994). It may therefore be significant that, in the

guard cells, mas7 appears to regulate the K÷channels independent of changes in cytoplasmic-free [Ca 2+] and pHi.

Experimental procedures

Plant culture and experimental protocol

Plants of Vicia faba var. bunyard were grown in 10 cm 2 pots at 16-20°C with a photoperiod of 16 h light: 8 h dark under artificial lighting, and watered with Hoaglands solution (Blatt, 1987b). For experiments, plants approximately 4-6 weeks old were used, with only the youngest newly expanded leaves chosen. Leaves were harvested just before each experiment. Epidermal strips were taken from the abaxial surface and mounted on a glass chamber that had been precoated with an optically clear, pressure-sensitive, silicone adhesive (No. 355 medical adhesive, Dow Coming, Mid- land, MI, USA). Measurements were conducted with the strips submerged in rapidly flowing solutions (10 ml min -1 ~ 20 chamber volumes min-1). Ambient temperatures during experiments were 20o-22°C.

All operations were carried out on a Zeiss IM inverted micro- scope (Zeiss, Oberochen, Germany) fitted with Nomarski Differen- tial Interference Contrast optics. Surface areas and volumes of impaled cells were calculated assuming a cylindrical geometry (Blatt, 1987a). The orthogonal dimensions (diameter, length) of the cells were measured with a calibrated eye-piece micrometer. Cell dimensions typically varied over 10-14 liM (diameter) and 35-45 llM (length). Estimated surface areas were thus 1.9x 10 -5- 3.1×10-5 cm 2 and cell volumes were 2.7-7.8 pl.

Standard bath medium was prepared with 5 mM 2-(N-morpho- lino)ethanesulphonic acid (MES, pK a = 6.1). The buffer was titrated to its pK a (=6.15) with Ca(OH) 2 (final [Ca 2+] ~- 1 mM). Potassium chloride (KCI), tetraethylammonium chloride (TEA) and butyrate were added as required. Butyrate was added after titrating with NaOH to the specified pH. Mas7 and masCP were included as required in a solution of the same buffers and salts via a 5 ml syringe placed at the side of the chamber. Exposures in these cases were carried out under stopped-flow conditions. Trials carried out without peptide additions showed that stopped flow alone had no appreciable effect on the K ÷ currents.

Microelectrodes

Recordings were obtained using two- and four-barrelled microelectrodes (Blatt, 1987a, 1991; Blatt and Slayman, 1987) coated with paraffin to reduce electrode capacitance. Current-passing and voltage-recording barrels were filled with 200 mM K+-acetate

(pH 7.5) to minimize salt leakage and salt-loading artefacts associ- ated with the CI- anion (Blatt, 1987a; Blatt and Slayman, 1983). For injections of non-hydrolysable nucleotides, four-barrelled microelectrodes were used, of which one barrel was filled with 100 llM or 1 mM nucleotide and backfilled with 10 mM K+-acetate. Current through this barrel was balanced by filling a second barrel with 50 mM K+-acetate. The remaining two barrels (for current- passing and voltage-following) were filled with 200 mM K +- acetate, as above. In all cases, connection to the amplifier head- stage was via a 1 M KCI/Ag-AgCI halfcell. A matching halfcell and 1 M KCI-agar bridge served as the reference (bath) electrode.

Electrical

Mechanical, electrical and software design have been described in detail previously (Blatt, 1987a, 1987b, 1990, 1991). Membrane currents were recorded with a voltage clamp under microproces- sor control using a WyeScience AD/DA converter, voltage clamp and tiP amplifier system (WyeScience, Wye, Kent, UK). Steady- state current-voltage (I-V) relations were determined by clamping cells to a bipolar staircase of command voltages (Blatt, 1987b). Steps alternated positive and negative from the resting potential, Vrn, typically a total of 20 bipolar pulse pairs, and were separated by equivalent periods when the membrane was clamped to V m. The current signal was filtered using a six pole Butterworth filter at 1 or 3 kHz (-3 dB) before sampling, and currents and voltages were recorded during the final 10 msec of each pulse.

For time-dependent characteristics, current and voltage were sampled continuously at 1 kHz while the clamped potential was driven through one to eight cycles of four programmable pulse steps. In all cases the holding potential was set to Vm at the start of the clamp cycle. No attempt was made to compensate for the series resistance (R s) to ground. Estimates for R s indicated that it was unlikely to pose a serious problem in measurements of clamp potential (Blatt, 1990, 1992), despite the often high resistivity of the bathing media (=2.5 k.Q cm for 5 mM Ca2+-MES with 0.1 mM KCI).

Numerical analysis

Data analysis was carried out by non-linear, least-squares (Marquardt, 1963) and, where appropriate, results are reported as the mean -+ standard error of (n) observations.

Chemicals and solutions

The pH buffer, MES, butyric acid, TEA, EGTA, ADP-I~-S and GDP- ~S were from Sigma Chemical Co. (St Louis, USA). Mas7 and masCP were gifts of Dr P.A. Millner, University of Leeds. Otherwise, all chemicals were analytical grade from BDH Ltd (Poole, Dorset, UK).

Acknowledgements We thank Dr Paul Millner for his support and helpful discussions, and Dr Ulrike Homann for comments on the manuscript. This work was possible with the aid of equipment grants from the Gatsby Charitable Foundation, the Royal Society and the Univer- sity of London Central Research Fund. F.A. was the recipient of a Sainsbury Studentship.

Guard cell K ÷ channel modula t ion by a 7TMS receptor mimet ic 197

References

Allan, E., Dawson, A., Drobak, B. and Roberts, K. (1989) GTP causes calcium release from a plant microsomal fraction. Cell. Signalling, 1, 23-29.

Bednarek, S.Y., Reynolds, T.L., Schroeder, M., Grabowski, R., Hengst, L., Gallwitz, D., and Raikhel, N. (1994) A small GTP- binding protein from Arabidopsis thaliana functionally complements the yeast YPT6 null mutant. Plant Physiol. 164, 591-596.

Blatt, M.R. (1987a) Electrical characteristics of stomatal guard cells: the contribution of ATP-dependent, 'electrogenic' transport revealed by current-voltage and difference-current-voltage analysis. J. Membr. Biol. 98, 257-274.

Blatt, M.R. (1987b) Electrical characteristics of stomatal guard cells: the ionic basis of the membrane potential and the consequence of potassium chloride leakage from microelectrodes. Planta, 170, 272-287.

Blatt, M.R. (1990) Potassium channel currents in intact stomatal guard cells: rapid enhancement by abscisic acid. Planta, 180, 445-455.

Blatt, M.R. (1991) Ion channel gating in plants: physiological implications and integration for stomatal function. J. Membr. Biol. 124, 95-112.

Blatt, M.R. (1992) K + channels of stomatal guard cells: characteristics of the inward rectifier and its control by pH. J. Gen. Physiol. 89, 615-644.

Blatt, M.R. and Slayman, C.L. (1983) KCI leakage from microelectrodes and its impact on the membrane parameters of a nonexcitable cell. J. Membr. Biol. 72, 223-234.

Blatt, M.R. and Slayman, C.L. (1987) Role of 'active' potassium transport in the regulation of cytoplasmic pH by nonanimal cells. Proc. Natl Acad. Scl, USA, 84, 2737-2741.

Blatt, M.R. and Armstrong, F. (1993) K ÷ channels of stomatal guard cells: abscisic acid-evoked control of the outward rectifier mediated by cytoplasmic pH. Planta, 191,330-341.

Blatt, M.R. and Thiel, G. (1993) Hormonal control of ion channel gating. Ann. Rev. Plant Physiol. Mol, Biol. 44, 543-567.

Blatt, M.R., Thiel, G. and Trentham, D.R. (1990) Reversible inactivation of K ÷ channels of Vicia stomatal guard cells following the photolysis of caged inositol 1,4,5-trisphosphate. Nature, 346, 766-769.

Bossen, M.E, Kendrick, R.E. and Vredenberg, W.J. (1990) The involvement of a G-protein in phytochrome-regulated, Ca 2÷- dependent swelling of etiolated wheat protoplasts. Physiol. Plant. 80, 55-62.

Brown, A.M. (1993) Membrane-delimited cell signaling complexes: direct ion channel regulation by G-proteins. J. Membr. Biol. 131, 93-104.

Dillenschneider, M., He~herington, A.M., Graziana, A., Alibert, G., Berta, P., Haeisch, J. and Ranjeva, R. (1986) The formation of inositol phosphate derivatives by isolated membranes of Acer pseudoplatinus is stimulated by guanine nucleotides. FEBS Lett. 208, 413-417.

Dohlman, H.G. (1991) Model systems for the study of seven- transmembrane-segment receptors. Ann. Rev. Biochem. 60, 653--688.

Drobak, B. and Watkins, P. (1994) Inositol-l,4,5-trisphosphate production in plant cells--stimulation by the venom peptides, melittin and mastoparan. Biochem. Biophys. Res. Commun. 205, 739-745.

Faddis, M.N. and Brown, J.E. (1993) Intracellular injection of heparin and polyamines: effects on phototransduction in Limulus ventral photoreceptors. J. Gen. Physiol. 101,909-931.

Fairley-Grenot, O.K. and Assmann, S.M. (1991) Evidence for G-

protein regulation of inward potassium ion channel current in guard cells of fava bean. Plant Cell, 3, 1037-1044.

Glevinovic, M.I., Jeehi, A. and Trifaro, J.M. (1992) Mastoparan blockade of currents through Ca2+-activated K + channels in bovine chromaffin cells. Neuroscience, 50, 675-684.

Higashijima, T., Wakamatsu, K., Saito, K., Fujino, M., Nakajima, 1". and Miyazawa, T. (1984) Molecular aggregation and conformational change of wasp venom mastoparan as induced by salt in aqueous solution. Biochem. Biophys. Acta, 802, 157-161.

Higeshijima, T., Uzo, S., Nakajima, T. and Rose, E.M. (1988) Mastoparan, a peptide toxin from wasp venom, mimics receptors by activating GTP-binding regulatory proteins (G proteins). J. Biol. Chem. 263, 6491-6494.

Hirai, Y., Yauhara, K., Yoshids, H., Nakajima, T., Fujino, M. and Kiteds, C. (1979) A new mast cell degranulating peptide 'mastoparan' in the venom of Vespula lewisii. Chem. Pharm. Bull. 27, 1942-1944.

Kim, D., Lewis, D.L., Graziadsi, L., Neer, E.J., Bar-Sagi, D. and Clapham, D. (1989) G-protein I~¢-subunits activate the cardiac muscarinic K + channel via phospholipase A 2. Nature, 337, 557-560.

Kurachi, Y., Ito, H., Sugimoto, T., Shimizu, T., Miki, I. and Ui, M. (1989) Arachidonic acid metabolites as intracellular modulators of the G-protein-gated cardiac K ÷ channel. Nature, 337, 555- 557.

Lee, Y.S., Lee, H.J., Crain, R.C., Lee, A. and Korn, S.J. (1994) Polyunsaturated fatty-acids modulate stomatal aperture and 2 distinct K + channel currents in guard cells. Ceil. Signal. 6, 181-186.

Legendre, L., Yueh, Y.G., Crain, R., Haddock, N., Heinstein, RE and Low, P.S. (1993) Phospholipase C activation during elicitation of the oxidative burst in cultured plant cells. J. Biol. Chem. 268, 24 559-24 563.

Lemtiri-Chlieh, E and MacRobbie, E.A.C. (1994) Role of calcium in the modulation of Vicia guard cell potassium channels by abscisic acid: a patch-clamp study. J. Membr. Biol. 137, 99-107.

Leung, J., Bouvier-Durand, M., Morris, P-C., Guerrier, D., Chefdor, F. and Giraudat, J. (1994) ArabidopsisABA response gene ABII: features of a calcium-modulated protein phosphatase. Science, 264, 1448-1452.

Li, W. and Assmann, S.M. (1993) Characterization of a G-protein- regulated outward K + current in mesophyll cells of Vicia faba L. Proc. Nat/Acad. Sci. USA, 90, 262-266.

Luan, S., Li, W., Rusnak, F., Assmann, S.M. and Schreiber, S.L. (1993) Immunosuppressants implicate protein phosphatase regulation of K ÷ channels in guard cells. Proc. Natl Acad. Sci. USA, 90, 2202-2206.

Me, H., Yanofsky, M.E and Meyerowitz, E.M. (1990) Molecular cloning and characterization of OPAl, a G-protein ~ subunit gene from Arabidopsis thaliana. Proc. Nat/Acad. ScL USA, 87, 3821-3825.

Ma, H., Yanofsky, M.F. and Hal, H.A. (1991) Isolation and sequence analysis of TGA1 cDNAs encoding a tomato G-protein (~ subunit. Gene, 107, 189-195.

MacRobbie, E.A.C. (1992). Calcium and ABA-induced stomatal closure. Proc. R. Soc. Lond. B. Biol. Scl, 338, 5-18.

Marquardt, D. (1963) An algorithm for least-squares estimation of nonlinear parameters. J. Soc. Ind. Appl. Math. 11,431-441.

Matsui, M., Sasamoto, S., Kunieda, T., Nomura, N. and Ishizaki, R. (1989) Cloning of ara, a putative Arabidopsis thaliana gene homologous to the ras-related gene family. Gene, 76, 313-319.

McAinsh, M.R., Brownlee, C. and Hetherington, A.M. (1990) Abscisic acid-induced elevation of guard cell cytosolic Ca 2+ precedes stomatal closure. Nature, 343, 186-188.

198 Fiona Armst rong and Michael R. Blatt

Meyer, K., Leube, M.P. end Grill, E. (1994) A protein phosphatase 2C involved in ABA signal transduction in Arabidopsis thaliana. Science, 264, 1452-1455.

Mousli, M., Bueb, J.-L., Bronner, C. and Landry, Y. (1990) G-protein activation: a receptor-independent mode of cationic amphiphilic neuropeptides and venom peptides. Trends PhysioL ScL 11, 358-362.

Nalefski, E.A., Sultzman, L.A., Martin, D.M., Kriz, R.W., Towler, RS., Knopf, J.L. and Clerk, J.D. (1994) Delineation of two functionally distinct domains of cytosolic phospholipase A 2, a regulatory Ca2+-dependent lipid-binding domain and a Ca 2+- independent catalytic domain. J. Biol. Chem. 269,18 239-18 249.

Obermeyer, G., Armstrong, F. and Blatt, M.R. (1994) Selective block by (x-dendrotoxin of the K ÷ inward rectifier at the Vicia guard cell membrane. J, Membr. Biol. 137, 249-259.

Ohki, S., Yazawa, M., Yagi, K. and Hikichi, K. (1991). Mastoparan binding induces Ca 2+ transfer between two globular domains of calmodulin--a 1H+-NMR study. J. Biochem. 110, 737-742.

Okemoto, 1"., Katada, T., Murayame, Y., Ui, M., Ogata, E. and Nishimoto, I. (1990) A simple structure encodes G-protein- activating function of the IGF-II/mannose 6-phosphate receptor. Cell, 62, 709-717.

Okano, Y., Takagi, H., Tohmatsu, T., Nakashima, S., Kuroda, Y., Saito, K. and Nozawe, Y. (1985) A wasp venom mastoparan- induced polyphosphoinositide breakdown in rat peritoneal mast-cells. FEBS Left. 188, 363-366.

Okumura, K., Inui, K.-I., Hirai, Y. and Nakajima, T. (1981). The effect of mastoparan on ion movement in black lipid membrane. Biomed. Res. 2, 450-452.

Quarmby, L.M., Yueh, Y.G., Cheshire, J.L., Keller, L.R., Snell, W.J. and Crain, R.C. (1992) Inositol phospholipid metabolism may trigger flagellar excision in Chlamydomonas reinhardtii. J. Cell Biol. 116, 737-744.

Reuveny, E., Slesinger, RA., Inglsse, J., Morales, J.M., Inguez- Lluhi, J.A., Lefkowitz, J.M., Bourne, H.R., Jan, Y.N. and Jan, L.Y. (1994) Activation of the cloned muscarinic potassium channel by G protein ~ subunits. Nature, 370, 143-145.

Roberts, D.M. and Harmon, A.C. (1992) Calcium-modulated proteins: targets of intracellular calcium signals in higher plants. Ann. Rev. Plant Physiol. Mol. Biol. 43, 375-414.

Romero, L.C. and Lain, E. (1993) Guanine nucleotide binding protein involvement in early steps of phytochrome-regulated gene expression. Proc. Nat/Acad. ScL USA, 90, 1455-1459.

Schloss, J.A. (1990) A Chlamydomonas gene encodes a G protein I~ subunit-like polypeptide. Mol. Gen. Genet, 221,443-452.

Schroeder, J.I. and Hagiwera, S. (1989) Cytosolic calcium regulates ion channels in the plasma membrane of Vicia faba guard cells. Nature, 338, 427-430.

Terryn, N., Van Montagu, M. and Inz6, D. (1993) GTP-binding proteins in plants. Plant MoL Biol. 22, 143-152.

Thiel, G. and Blatt, M.R. (1994) Phosphatase antagonist okadaic acid inhibits steady-state K + currents in guard ceils of Vicia faba. Plant J. 5, 727-733.

Vera-Estrella, R., Berkla, B.J., Higgins, V.J. and Blumweld, E. (1994) Plant defense response to fungal pathogens. Plant Physiol. 104, 209-215.

Wallace, M. end Carter, H. (1989) Effects of the wasp venom peptide, mastoparan, on a phosphoinositide-speciflc phospholipase C purifed from rabbit brain membranes. Biochim. Biophys. Acta, 1006, 311-316.

Warpeha, K.M.F., Harem, H.E., Rssenick, M.M. and Kaufman, L.S. (1991) A blue-light-activated GTP-binding protein in the plasma membranes of etiolated peas. Proc. Natl Acad. Sci. USA, 88, 8925-8929.

Weldman, P.J. and Winter, W.M. (1994) The G protein-activating peptide, mastoparan, and the synthetic NH2- terminal ARF peptide, ARFp13, inhibit in vitro Golgi transport by irreversibly damaging membranes. J. Cell Biol. 127, 1815-1827.

Weingarten, R., Ransnss, L., Mueller, H., Sklar, LA. and Brock, G.M. (1990) Mastoparan interacts with the carboxyl terminus of the alpha subunit of Gi. J. Biol. Chem. 265, 11 044-11 049.

Weiss, C.A., Gemast, C.W., Mukai, K., Hu, Y. and Me, H. (1994) Isolation of cDNAs encoding guanine nucleotide-binding protein ~-subunit homologues from maize (ZGB1) and Arabidopsis (AGBI). Proc. Natl Acad. Sci. USA, 91, 9554-9558.

White, I.R., Wise, A. and Millner, RA. (1993) Evidence for G- protein-linked receptors in higher plants: stimulation of GTP-7- S binding to membrane fractions by the mastoparan analogue mas7. Planta, 191,285-288.

Wickman, K.D., Iniguez-Uuhl, J.A., Davenport, RA., Taussing, R., Krapivinsky, G.P., Under, M.E., Gilman, A.G. and Clapham, D.E. (1994) Recombinant G-protein ~ subunits activate the muscarinic-gated atrial potassium channel. Nature, 365, 255- 257.

Wu, W.-H. and Assmann, S.M. (1994) A membrane-delimited pathway of G-protein regulation of the guard-cell inward K + channel. Proc. Natl Acad. Sci. USA, 91, 6310--6314.

Zaina, S., Reggiani, R. and Bertani, A. (1990) Preliminary evidence for involvement of GTP-binding protein(s) in auxin signal transduction in rice (Oryza sativa L.) coleoptile. J. Plant Physiol. 136, 653-658.

Documents

Evidence for K ÷ channel control in Vicia guard cells