Planta (1990)180:445~55
P l ~ n t ~ 9 Springer-Verlag1990
Potassium channel currents in intact stomatal guard cells: rapid enhancement by abscisic acid Michael R. Blatt Botany School, University of Cambridge, Downing Street, Cambridge CB2 3EA, UK
Abstract. Evidence of a role for abscisic acid (ABA) in signalling conditions of water stress and promoting stomatal closure is convincing, but past studies have left few clues as to its molecular mechanism(s) of action; arguments centred on changes in H +-pump activity and membrane potential, especially, remain ambiguous without the fundamental support of a rigorous electrophysiological analysis. The present study explores the response to ABA of K + channels at the membrane of intact guard cells of V i c i a f a b a L. Membrane potentials were recorded before and during exposures to ABA, and whole-cell currents were measured at intervals throughout to quantitate the steady-state and time-dependent characteristics of the K § channels. On adding 10 ~tM ABA in the presence o f 0.1, 3 or 10 m M extracellular K § the free-running membrane potential (Vm) shifted negative-going ( - ) 4 - 7 mV in the first 5 min of exposure, with no consistent effect thereafter. Voltage-clamp measurements, however, revealed that the K+-channel current rose to between 1.84- and 3.41-fold of the controls in the steadystate with a mean halftime of 1.1 ___0.1 min. Comparable changes in current return via the leak were also evident and accounted for the minimal response in Vm. Calculated at Vm, the K + currents translated to an average 2.65fold rise in K + efflux with ABA. Abscisic acid was not observed to alter either K +-current activation or deactivation. These results are consistent with an ABA-evoked mobilization o f K + channels or channel conductance, rather than a direct effect of the phytohormone on K +channel gating. The data discount notions that large swings in membrane voltage are a prerequisite to controlling guard-cell K + flux. Instead, they highlight a rise Present address." Department of Biochemistry and Biological Sciences, University of London, Wye College, Wye (Ashford), Kent TN25 5AH, UK Abbreviations: ABA = abscisic acid; EK= K § equilibrium potential; I-V = current-voltage (relation); Kff = extracellular K § (concentration); TEA=tetraethylammonium chloride; Vm=free-running membrane potential (difference)
in membrane c a p a c i t y for K + flux, dependent on concerted modulations of K+-channel and leak currents, and sufficiently rapid to account generally for the onset of K + loss from guard cells and stomatal closure in ABA.
Key words: Abscisic acid and guard cells - Membrane potential - Potassium channel - Stomatal guard cell Vicia - Voltage clamp
Introduction Abscisic acid (ABA) is thought to function as a physiological signal during conditions of water stress and, especially, to mediate stress-induced closure o f stomata in the leaves of higher plants (see Walton 1988; also Raschke 1987 for reviews). Abscisic acid is known to be synthesized during periods of drought stress and accumulates in the leaf tissues, including the guard cells (see Cornish and Zeevaart 1986; Behl and Hartung 1986; Harris et al. 1988); the extent o f ABA accumulation usually correlates inversely with stomatal aperture (see Weyers and Hillman 1979; also discussion in Harris et al. 1988); and exogenous ABA induces closure of stomata in leaves and epidermal peels when added to the transpiration stream or bathing medium at concentrations comparable to those found in situ (l-100 gM; see Mittelheuser and van Steveninck 1969; also Raschke et al. 1975). In addition; ABA is known also as an antagonist to K + salt accumulation and stomatal opening (Jones and Mansfield 1970; Hartung 1983). Yet, despite the implicit role in modulating solute transport across the guard cell and, generally, across the higher plant cell plasma membrane, very little is known about the site of ABA action or the transport process(es) affected. One popular notion (see Raschke 1977, 1987) maintains that ABA must reduce H+-ATPase activity and electrogenic H + extrusion across the plasma membrane,
446 hence lessening the driving force for K + uptake. Consistent with this view, net H + extrusion measured from a variety of tissues has been reported to decline in the presence of ABA (Rayle 1973; Lado et al. 1975; Gepstein et al. 1982). There are scattered accounts also of membrane depolarizations in ABA; furthermore, changes in membrane potential and net K § flux have been related sensibly to the prevailing equilibrium (diffusion) potential for K § (EK) and to the reduced driving force for K § uptake, at least in Nicotiana tabacum leaf discs (Kasamo 1981). Noteworthy, too, treatments with the fungal toxin fusicoccin appear to " p r o t e c t " many tissues from the effects of ABA (see Squire and Mansfield 1972; also Ballarin-Denti and Cocucci 1979); fusicoccin has been thought to act primarily by stimulating electrogenic H § extrusion via the H§ (see Marr6 1979; also Blatt 1988a for reviews), so these observations complement notions of H § pump down-regulation by ABA. There is also some evidence which might argue to the contrary. Hartung et al. (1980) reported that ABA treatments hyperpolarized (negative-going) membrane potentials recorded from Lemna gibba to values in excess of - 2 0 0 mV and well negative of all dominant diffusion regimes. Uptake of [14C]glucose and [14C]glycine, too, were stimulated following ABA pretreatments. Furthermore, unidirectional flux analyses using Commelina communis guard cells (MacRobbie 1981) revealed that ABA had little effect on K+(S6Rb +) and CI-(82Br-) influx, but led to a large and transient stimulation of the tracer effluxes. At least at face value, neither set of data appears consistent with a simple decline in H § pump activity, much less any singular explanation for ABA action at the membrane. The seemingly conflicting results aside, it must be stressed that membrane potential and net H § flux alone are unsatisfactory as guides to transport behavior(s) in H§ membranes, including the plant plasma membrane. In effect, both voltage and net H + flux are metabolic intermediates, being " p r o d u c t s " of primary electrogenic pumping and "substrates" for secondary and H+-coupled transporters. So, for example, membrane depolarization and reduced H § extrusion in ABA cold reflect a genuine decline in H § output by the pump; equally, the phenomenology could be a consequence of A B A activating H+-coupled transport or an efflux of weak acid (e.g. Van Kirk and Raschke 1978). Likewise, few clues to ABA action can be derived from unidirectional flux measurements in the absence of any detail on the mechanism(s) for ionic flux or changes in driving force. So, without knowing the component transport currents and their voltage dependence, neither the voltage nor the flux results documented for ABA can distinguish between a primary effect on the pump or on other transport processes, or both. Electrical recordings do offer a means of dissecting out and characterizing individual transport processes in vivo, provided that membrane current can be measured under voltage clamp. Recent voltage clamp studies with intact guard cells of Vicia have demonstrated that fusi-
M.R. Blatt: Abscisic acid and guard cell K § channels coccin inactivates outward-rectifying K § channels and K § (86Rb § efflux (Blatt and Clint 1989; Clint and Blatt 1989), in addition to its limited effects on primary H + pumping (Blatt 1988a). That channel gating is altered and the channels inactivated irreversibly in fusicoccin supports the idea that the channels mediate K § loss during stomatal closure; indeed, a central feature of fusicoccin toxicity is a loss of the capacity for stomata to close (Turner and Graniti 1969). The effects of fusicoccin on K § channel current and tracer flux also raise questions about the transduction of other chemical and environmental signals in stomatal response. It is conceivable that the ABA-induced K § (86Rb+) efflux reported the MacRobbie, and stomatal closure, might be a simple consequence of membrane depolarization activating the K § channels, whether the voltage shift be achieved by down-regulating the H § pump or by stimulating an anion efflux (see Schroeder and Hagiwara 1989). Equally plausible, however, ABA could have a more direct effect on K + flux, possibly including changes in K § channel gating (Schauf and Wilson 1987) and steady-state current characteristics. The present study explores the K § channel response to ABA of intact guard cells in epidermal peels from Vicia. Measurements were carried out at the end of the growing season (September-October) when guard cells showed little indication of primary pump activity, thus greatly simplifying analyses of whole-cell recordings. Adding ABA resulted in a rapid rise in the steady-state K § current measured under voltage clamp, but without an apparent change in voltage-dependent channel gating. These data indicate that ABA modifies the capacity for K + flux at the guard cell membrane, independent of any change in the free-running membrane potential (Vm) 1, and point to a concerted action of the phytohormone on ion transport pathways in the membrane distinct from the electrogenic H § pump.
Material and methods Plant culture and experimental protocol. Viciafaba L., cv. (Bunyan) Bunyard Exhibition, was grown on vermiculite with Hoagland's Salts medium and epidermal strips were prepared as described before from newly expanded leaves taken four to six weeks after sowing (Blatt 1987a, b). Measurements were carried out in rapidly flowing solutions (10 ml/min ~20 chamber volumes/min) containing 5 mM Ca 2+-Hepes (4-(2-hydroxyethyl)-l-piperazineethansesulfonic acid), pH 7.4 (5 mM Hepes buffer titrated to pH 7.4 with Ca(OH)2), final [Ca2+] ~1 mM). Potassium salts, tetraethylammonium chloride (TEA) and abscisisc acid (ABA) were included as required. Ambient temperatures were 20-22~ C. Surface areas and volumes of impaled cells were calculated assuming a cylindrical geometry (Blatt 1987a). The orthogonal dimensions (diameter, length) of the cells, and stomatal apertures, were measured with a calibrated eye-piece micrometer. Cell dimensions in these experiments typically varied over 10-14 ~tm (diame-
1 For reasons detailed previously (Blatt 1987a; see also Parsons and Sanders 1989), electrical recordings from the guard cells are assumed to reflect the characteristics of the plasma membrane alone. I use the terms guard 'cell' and 'membrane' potential, current and conductance in this context.
M.R. Blatt: Abscisic acid and guard cell K § channels ter), and 30-35 gm (length). Estimated surface areas thus were 1.2.10-5-1.9 910 -5 cm z and cell volumes were 2.2-5.3 pl. Electrical. Mechanical, electrical and software design have been described in detail (Blatt 1987a, b). However, two new modifications were introduced in these studies. First, the epidermal peels were affixed with gentle contact to the glass bottom of the experimental chamber after coating the chamber surface with an optically clear and pressure-sensitive, silicone adhesive (No. 355 medical adhesive, Dow Corning, Brussels, Belgium). Second, all operations were carried out on a Zeiss IM inverted microscope (Zeiss, Oberkochert, FRG) fitted with Nomarski Differential Interference Contrast optics. With the epidermal peels now horizontally mounted (normal to the optical axis), the microelectrodes were brought in at an angle 33~ off the horizontal, thus providing an entirely free field of view. Recordings were obtained by the two-electrode method using double-barrelled microelectrodes. The electrodes were filled with 200 mM K§ pH 7.2 to minimize salt leakage and saltloading artifacts (Blatt 1987a), and were coated with paraffin or Sylgard polymer (Dow Coming) to reduce electrode capacitance. Current-voltage relations were made by voltage clamp under microprocessor control. Steady-state current-voltage (I-V) relations were determined by clamping cells to a bipolar staircase of command voltages (Blatt 1987 b). Steps alternated positive and negative from Vm (typically 20 bipolar pulse-pairs) and were separated by equivalent periods when the membrane was clamped to V~. The current signal was filtered by a six-pole Butterworth filter at 1 or 3 kHz ( - 3dB) before sampling, and currents and voltages were recorded during the final 10 or 20 ms of each pulse. For time-dependent characteristics, current and voltage were sampled continuously at 1, 2 or t0 kHz while the clamped potential was driven through cycles of 1-4, programmable pulse steps. Recordings at 10 kHz (above the Nyquist limit) were restricted by available data storage space to 200-ms "windows" within each cycle; otherwise data taken at all frequencies gave similar results. No attempt was made to compensate for the series resistance (Rs) to ground (Hodgkin et al. 1952). Estimates for Rs indicated that it was unlikely to pose a serious problem in measurements of clamp potential, despite the often high resistivity of the bathing media (=2.5 kf2.cm for 5 mM Ca2+-Hepes with 0.1 mM KC1). In practice, Rs is likely determined by the cell wall which might place values near 1-2 MY2 (Briggs et al. 1961 ; Blatt 1988 b). Cell input resistances near Vm ranged between 1 and 4 GO and, in the worst case for the K + conductance maximum near 0 mV, (slope) input resistance never fell below 110 MO. Hence, the maximum voltage error likely was 1-2% of the clamp potential, or less than 2 mV on clamping a cell to 0 mV. Numerical analysis. Guard cell I-V relations were fitted to polynomials, using a non-linear, least-squares algorithm (Marquardt 1963; Jennings et al. 1988). In most cases, 8th- (occasionally 10th-) order polynomials coped with the rapid changes in slope at positive potentials. Exponential fittings, likewise, were by non-linear, leastsquares. Chemicals and solutions. Tetraethylammonium chloride, ( + ) A B A and the pH buffer Hepes, were from Sigma Chemical Co. (St. Louis, USA). Otherwise, all chemicals were Analytical Grade from BDH (Poole, Dorset, UK). Where appropriate, results are reported as the mean + standard error of (n) observations.
447
evidence of primary pump activity as could be determined from the current-voltage (I-V) and conductancevoltage (G-V) characteristics of the cells, from their voltage response to external pH and K + concentration (K~-), or from their voltage and conductance behaviour during metabolic blockade with cyanide (compare Blatt 1987 a, b with Blatt 1988b and Blatt and Clint 1989). Under these conditions, free-running membrane potentials (Vm) lay positive to the K + equilibrium potential (EK) and K + channel current contributed to charge balance at Vm (Blatt 1988b; Clint and Blatt 1989). Comparable results were obtained in the present study; values for Vm and G m ( = slope conductance at Vm) were - 1 4 4 _ 5 mV and 64_ 15 ktS"cm- 2 (n = 8) in 5 mM Ca 2 § pH 7.4, with 0.1 mM KCI, and Vm showed a near-Nernstian dependence on K~- at concentrations about 1 mM and above [54___2mV/K~- decade between 1-30mM K~ (n = 6)]. The mean stomatal aperture at the start of impalements was 7.0 _ 0.3 ~tm (n = 8). The ensemble K § channel current also dominated the steady-state I-V relations of the guard cells (see also Blatt 1988 b). Clamping the membrane to voltages away from Vm revealed large, outward-rectifying (positive) currents at potentials near and positive to Vm (Fig. 1 a, open symbols) which developed sigmoidally with time (see Fig. 2) and which could be blocked rapidly and reversibly by adding 10 mM TEA. The background, or ' leak' current remaining in TEA (Fig. 1 a, filled symbols) was essentially linear at potentials negative to approx. - 1 0 0 mV and rectified only slightly at more positive voltages. In Fig. 1 the K + current is indicated by the shaded difference between the two I-V curves, _ TEA. [Tetraethylammonium chloride also introduced an additional leak current, possibly related to the anion added (see Blatt and Clint 1989). So, for purposes of estimating K + channel current near Vm -- where K + currents were small and errors introduced by the leak were proportionately large - the leak could be approximated by extrapolating a line fitted to data points ( - T E A ) over the voltage range in which the channels were deactivated (Fig. 1 a, inset; see Blatt 1988b).] It is noteworthy that the free-running potentials of the guard cells lay positive to EK (determined from K § channel tail current reversal potentials; see Blatt 1988 b). Exposures to TEA resulted in rapid and reversible positive-going shifts in Vm (Fig. 1 a, inset above), indicatng that current through the K § channels contributed appreciably to charge balance in the absence of the channel blocker. Again, an estimate for the K § current present under free-running conditions could be obtained by subtracting currents _ TEA at Vm in the absence of TEA. The data in Fig. 1 a (inset, arrow) yielded a value of 0.5 [ . t A ' c m - 2 ( = 5 . 2 p m o l K + . c m - 2 . s - 1 ) ; extrapolating a linear leak fitted to points negative from EK gave a similar, if marginally smaller figure.
Results
Steady-state characteristics. Previous experiments indicated that guard cells examined at the end of the growing season (September-November) generally showed little
Abseisic acid. Although charge (K +) transport through the K + channels is heavily biased in favour of net K + efflux (positive current) from the guard cells in 0.1 mM K~- (Fig. 1) and also at higher concentrations (Blatt
.5o
448
M.R. Blatt: Abscisic acid and guard cell K § channels
a -R4
T/IjA
-2
-60
cm
-ABA
0-'-:..,..,,
-150
-150
r
+ABA
-2.=
0"7.. m
T I n A err~2
b 40 pA cm 2
IIIIIIIIIIlllLII
I/INIII .mm,.m.
NIN NINI
400 ms Fig. 2. Voltage-dependent K + current activation and deactivation. Data from one Vicia guard cell before and 5.6 rain after adding 10 I.tM ABA to the bathing medium. Currents were recorded in 10 mM K~ to slow activation (Blatt 1988 b) with clamp steps (8) to voltages between - 6 0 and + 50 mV preceded by conditioning steps to - 1 5 0 mV from the holding potential (= Vm: --56 mV, - A B A ; - 6 1 mV, +ABA) and followed by steps to -150 mV to record deactivation. Sampling frequency, 2 kHz. The principal effect of ABA was in amplifying the time-dependent (K +) current activated by positive-going and deactivated by negative-going voltage steps. Cell parameters: surface area, 1.8.10-5 cm2; volume, 3.9 pl; stomatal aperture, 7.2 gm
,v -200
- 100
1O0
Fig. 1 a, b. Steady-state current-voltage (I-V) relations and (ensemble) K + channel current from one Vicia guard cell in 5 mM Ca 2§ Hepes, pH 7.4, with 0.1 mM KC1. a Scans (bipolar staircase of 150-ms steps, 12 s total duration) were run immediately before (o, V,.=--143mV) and 35s after (e, V m = - 8 4 m V ) adding 10 mM TEA to the bathing medium. Data points fitted to 8th-order polynomials. Potassium current as a function of (clamped) membrane potential was found as the (shaded) difference between the two curves, i ( - T E A ) i(+TEA). Cell parameters: surface area, 1.6.10-5 .cmZ; volume, 3.5 pl; stomatal aperture, 4.2 gm. At 0 mV the shaded (difference) current was equivalent to approx. 560 pA over the whole cell surface. Inset (above): voltage trace with period in TEA (shading) and times of I-V scans (carats, cross-referenced by symbol) as indicated. Inset (below): exploded view of the curves with the K + current at Vm ( - T E A ) indicated by the arrow. The dotted line was extrapolated from a linear fitting to data points between EK (=--186 mV) and - 2 2 0 mV. The slight downward displacement of the +TEA curve from this line may relate to the C1- anion added with the channel blocker. Potassium equilibrium potentials (EK) were determined as the K § channel tail current reversal potential (compare Blatt 1988b). b K + current replotted from a (shading)
1988b), v o l t a g e d e p e n d e n c e a l o n e offers few insights into t h e p h y s i o l o g i c a l c o n t r o l s on i o n flux. A net o r e n h a n c e d K + effiux in A B A m i g h t be achieved s i m p l y b y shifting Vm positive a l o n g the v o l t a g e axis, a n d c o u l d be realized by an increase in the ( i n w a r d - d i r e c t e d ) leak current. F o r e x a m p l e , a ( + ) 2 3 m V shift w o u l d suffice to raise the K + flux f o u r f o l d w i t h o u t a n y c h a n g e in the s t e a d y - s t a t e I-V profile for the e n s e m b l e K + c h a n n e l c u r r e n t s h o w n in Fig. 1. A l t e r n a t i v e l y (or in a d d i t i o n !), A B A c o u l d affect the I-V c h a r a c t e r i s t i c for the K + current itself, so as to increase the capacity for K + flux n e a r Vm. In the absence o f a n y c h a n g e in l e a k current, the effect in this case c o u l d be expected to h y p e r p o l a r i z e ( n e g a t i v e - g o i n g ) Vm. TO e x p l o r e these possibilities, m e m b r a n e p o t e n t i a l s o f g u a r d cells (eight e x p e r i m e n t s ) a n d the v o l t a g e - a n d t i m e - d e p e n d e n t features o f their K + c u r r e n t s were e x a m ined before, d u r i n g a n d after e x p o s u r e s to 10 g M A B A . C u r r e n t - v o l t a g e scans were run at intervals over 2 0 -
M.R. Blatt: Abscisic acid and guard cell K + channels
449 t~/2/ms
I /juA c m 2 90
/~
,o/
aoo
/
/
fO O
/
30o/
/
100
10
- 12 0
~
f~/~/f/i /
...., ...'-'"" V / m V
65
~, ......... '..................... 2 6 " ~" -10
Fig. 3. Instantaneous (leak) and net steady-state currents from Fig. 2. Data points are for currents recorded before (open symbols) and after (/~lled symbols) adding ABA. Instantaneous currents (squares) were determined 2 ms following positive-going steps from - 1 5 0 mV; steady-state currents (circles, K + channel plus leak currents) were taken at the end of each clamp step between - 6 0 and + 50 inV. Points at - 1 5 0 mV are currents taken at the end of the conditioning voltage steps
100
ms L
+ 5o _ . . . . ~
-IsK
"
0-'-~
| 3 0 pA cf~ 2
_
_
-1~~
C'
[
- a'0
'
' -10
b 1
25 m i n following each i m p a l e m e n t before A B A was added to ensure a stable baseline o f K + current. M e a s u r e m e n t s were carried o u t in 0.1, 3.0 a n d 10 m M K~-, all with c o m p a r a b l e results. I n each case, Vm shifted n e g a t i v e - g o i n g d u r i n g the first 5 rain in A B A [range ( - ) 4 - 7 mV] with n o consis-
b
'
3
'
50
V /mY
Fig. 5. Insensitivity to ABA of the voltage-dependence for K + current activation. Halftimes (tl/z) for data taken in 10 mM Kg from the same cell as in Fig. 4. Steps to the voltages indicated were preceded by clamping to a conditioning voltage of - 150 mV. Clamp cycles were initiated before (o) and after 8.3 (m) and 11.2 (o) min in 10 gM ABA. Where necessary to accommodate current scatter, halftimes (+SE) were taken after first fitting the data to 8th-order polynomials (see Fig. 4) by non-linear least-squares (Material and methods). A joint fitting to a single exponential (solid line) gave an e-fold rise in tl/2 per (-)71.3 mV
Table l. Insensitivity to ABA of activation and deactivation kinetics for Vicia K + channel current. Summarized are data (n = 5) sampled at 2 kHz during exposures to 10 mM Kg which slows voltagedependent activation (see Blatt 1988b). In each case, activation at +50 mV was preceded by a 800-ms conditioning step to - 150 mV and deactivation was followed during a subsequent step to this same voltage (see Fig. 2). Halftimes for activation were determined by eye after subtracting instantaneous (leak) currents; for deactivation, data points were fitted satisfactorily to single, falling exponentials (Material and methods). Exposure times to 10 gM ABA were > 5 min in each case
3 0 0 ms
Fig. 4. Insensitivity of K + current activation to ABA. Data from one Vicia guard cell in 10 mM K~- before and 8.3 min after adding I0 gM ABA. Currents recorded at 2 kHz during voltage clamp cycles as in Fig. 3. Holding potentials (Vm), --58 mV ( - A B A ) and - 6 5 mV (+ABA). The time-dependent current obtained at + 50 mV before adding ABA was fitted to an 8th-order polynomial for clarity, and then expanded on the ordinate axis by a factor of 2.35 to superimpose on the current recorded in ABA. Halftime for activation, 109 ms. Inset: raw current data and corresponding voltage steps. Cell parameters: surface area, 1.4.10- s cm 2; volume 3.1 pl; stomatal aperture, 6.7/am
'
tl/2 (ms)
Activation (+ 50 mV) Deactivation ( - 1 5 0 mV)
-- ABA
+ ABA
110 + 4 10.8 + 0.2
108 + 3 10.8 ___0.3
tent c h a n g e thereafter. C l a m p i n g the m e m b r a n e to voltages n e a r a n d positive to 0 mV, however, revealed the full i m p a c t o f A B A o n the K + current. D a t a for one g u a r d cell in 10 m M K~- before, a n d 5.6 m i n after a d d i n g A B A are s h o w n in Fig. 2. C l a m p steps were preceded by a c o n d i t i o n i n g voltage ( - 150 mV) to deactivate the K + current. Stepping the voltage to values positive from Ez ( = - 7 9 m V for this cell in 1 0 r a M K~-, d a t a n o t shown) t h e n gave a n i n s t a n t a n e o u s j u m p in (leak) current followed by a t i m e - d e p e n d e n t , rising c o m p o n e n t
450
M.R. Blatt: Abscisic acid and guard cell K + channels
a
-2
I / p A cm
70 f ~176176149
? ? 7 50
/11
?
?
?
,o.o.o'~176
10
OS . 0 4
V /mV
-250 . _=e,%~176 . ""~176176 t '@i~@i@@~mi~m'~'~ -100
:
100
-10
b
-2
T tOA cr~2 80
I /pA crn 0
12 ' 40
0
,
! o
8
-2~5 ,~
/
!
/./ .(
-21o
0
4
.........
....S . / "
V/mY
'
if/':
IK, V~ (I- t A ' c m - 2)
- ABA
+ ABA
0.49 -I-0.09
1.30 -I-0.18
,o-
d
p/
Table 2. Enhancement by ABA of K + current at the free-running potential, IK. Vm' Data from five Vicia guard cells before and after > 5 rain exposure to 10 pM ABA determined under steady-state voltage clamp (see Fig. 6). Analyses of measurements in 0.1, 3 and 10 mM K~ gave comparable currents at V,,, and the results are pooled accordingly
t,
Fig. 6a, b. Steady-state current-voltage response to ABA. Data from one Vicia guard cell in 0.1 mM K~- before (o) and 9.3 min after (o) adding 10 I.tM ABA to the bathing medium. Currents were averaged over the final 20 ms during each 150-ms step of bipolar staircase scans (Material and methods) and were decomposed into K § channel and leak components by linear extrapolations of the leak from points negative to EK (=--178 mV, data not shown) as described in Fig. 1. a Whole-cell I-V data. b Detail of the data in a with the component K § channel (solid and dashed lines) and leak currents (dotted and dot-dashed lines) as indicated ( - A B A , solid and dotted lines; +ABA, dashed and dot-dashed lines). Arrows indicate K + channel (1') and leak (+) currents at Vm -I-ABA. Potassium current at Vm: -ABA, 0.85 IxA'cm-2; +ABA, 1.64 ixA.cm-2. Inset: K + currents determined by subtracting I-V curves in TEA (not shown) before (solid line) and after (dashed line) adding ABA. Cell parameters: surface area, 1.7-10-5 cm2; volume, 4.4 pl; stomatal aperture, 8.1 Ixm.Currents in ABA were 765 pA at 0 mV and 1.1 nA at the maximum near + 60 mV as the K + current activated (see Blatt 1988b). Subsequently clamping to - 1 5 0 mV, negative from EK, showed an inward (negative) current which decayed over approx. 50 ms as the K + current deactivated. After 5.6 min in A B A the same clamp protocol yielded time-
dependent (K + current) components which were amplified roughly 2.4-fold at each voltage. Figure 3 summarizes the instantaneous and steady-state I-V characteristics for the data in Fig. 2. One intriguing feature of this A B A response was an apparent lack of effect on the voltage-dependence of channel gating. Abscisic acid enhanced the K + current in purely scalar fashion, that is, without any significant change in the time course for current activation or deactivation. Figure 4 illustrates this point for current activation with data taken in 10 m M K~- from another cell using the same pulse protocol; scaled to a c o m m o n ordinate, the time-dependent currents at + 50 mV superimpose with a c o m m o n halftime for current rise of 109 ms. The current measured in A B A was 2.35-fold greater in magnitude than the control, however (Fig. 4, inset). Halftimes for current rise as a function of clamp potential and A B A exposure are summarized in Fig. 5 and Table 1, including data from Fig. 2. Complementary results were obtained also in 0.1 and 3 m M K~, although lowering K~ concentration, itself, accelerated current activation (Blatt 1988b). Halftimes for current deactivation at - 1 5 0 m V (3 and 1 0 m M K +) and - 2 2 0 m V (0.1 m M K~) were 10.8 and 8.1 ms, respectively, and are comparable to values determined from previous experiments under similar conditions (Blatt 1988b; Blatt and Clint 1989); in ABA, deactivation varied in all cases by less than i 2% of the controls and without any consistent trend (Table 1).
Membrane potential and the leak. The instantaneous currents recorded on stepping positive from - 150 mV (see Figs. 2, 3) and - 220 mV also indicated a limited change in leak current with ABA. In fact, the near-constant Vm in the face of a rise in K + current anticipated this observation; an appreciable rise in current directed out of the cell otherwise might have driven Vm well negative in the face of a constant leak. However, to confirm this expectation and to determine the extent to which the K + current was enhanced at Vm, steady-state I-V curves were determined using a bipolar staircase protocol and currents were time-averaged at the end of each voltage step to improve the small-signal resolution for currents near Vm (see Material and methods). Potassium channel and leak currents near V,, were assessed by linear extrapolations, as outlined in Fig. 1 Figure 6 shows measurements from one cell which were taken in 0.1 m M K~-. Equivalent analyses carried
M.R. Blatt: Abscisic acid and guard cell K § channels out on data f r o m four other cells gave similar results, and K + currents at Vm + A B A are summarized in Table
451
T /~A
a
c m -2
140
2. From Fig. 6 it will be seen that adding ABA was
followed by a 2.3-fold rise in K + current at any one voltage, and Vm was displaced 4 m V negative from the control ( = - 136 mV) as indicated by the V-axis intercepts for the data points (Fig. 6 b). Now, however, this limited change in Vm could be tied directly to a concomitant rise in the (inward-directed) leak which balanced the K § current. At Vm ( - A B A ) , the K + current rose from 0.85 laA.cm -2 in the control (Fig. 6b, solid line) to 1.94 IxA.cm -2 after 9.3 min in A B A (Fig. 6b, dashed line); but this enhancement was nearly matched by changes in current return via the leak (Fig. 6b: dotted line, - A B A ; dot-dashed line, + A B A ) . So, charge balance was maintained with little change in the free-running potential (at Vm, im= 0 = i~:+ iL, where iK is the K + current and the leak current, iL, is the sum o f currents passing through all other charge transport pathways). By contrast, had the leak remained unaffected. A B A would have driven Vm negative by 1 3 m V to - 1 4 9 m V where the leak in the control ( - A B A , dotted line) and the K + current in ABA (solid line) were equal in magnitude and opposite in sign. A b s c i s i c - a c i d t i m e course. A singular feature of the physi-
ological response to A B A is the speed with which stomata close in the presence of the phytohormone. The most detailed time courses available (e.g. Cummins et al. 1971; Raschke et al. 1975; Weyers and Hillman 1979) indicate that guard cells must begin to lose salt and osmotic pressure within a few tens o f seconds to minutes following exposure to ABA, although 1 0 4 0 min m a y be required before the response is complete and the stom a t a close fully. So, it could be expected that the K § current should follow a similar time course and rise rapidly to a new quasi-steady-state in ABA, if the K § channels were to qualify as a notional pathway for K § flux during stomatal closure 2. Because of the speed with which the cells clearly did respond to the p h y t o h o r m o n e , data for times early on following ABA addition, again, were determined using a bipolar staircase protocol. Whole-cell and K § currents from one guard cell are shown in Fig. 7 with the corresponding response timecourse in Fig. 8. In this case adding l0 ~tM A B A resulted, at any one clamp potential, in an approx. 2.1-fold rise in K + current and conductance (Fig. 7b, inset) which was essentially complete within 6 min (tl/2, 1.08 min). C o m p a r a b l e results were obtained in the other experiments, giving a pooled (n = 6) z Published time courses are for ABA-induced stomatal closure or for stomatal (gaseous) conductance. These parameters index the solute content and, broadly speaking, the K + concentration in the guard cell. As a measure of the rate of cation movement, the K § current must relate to the derivative of these ABA time courses, that is, to the rate of change in K § concentration with time. Hence, the rise in the K § current should correspond to the onset of stomatal closure, and the current should attain at a new quasi-steady-state consistent with the longer times required to complete closure.
10.2 rnin ilillll 101~i
ooo D
=' o~
,llo',~ =o'e /,e
4. I rain
1.7 rain
i'ft'~o~176176176 Control
I
. .p
P l ~ n t ~ 9 Springer-Verlag1990
Potassium channel currents in intact stomatal guard cells: rapid enhancement by abscisic acid Michael R. Blatt Botany School, University of Cambridge, Downing Street, Cambridge CB2 3EA, UK
Abstract. Evidence of a role for abscisic acid (ABA) in signalling conditions of water stress and promoting stomatal closure is convincing, but past studies have left few clues as to its molecular mechanism(s) of action; arguments centred on changes in H +-pump activity and membrane potential, especially, remain ambiguous without the fundamental support of a rigorous electrophysiological analysis. The present study explores the response to ABA of K + channels at the membrane of intact guard cells of V i c i a f a b a L. Membrane potentials were recorded before and during exposures to ABA, and whole-cell currents were measured at intervals throughout to quantitate the steady-state and time-dependent characteristics of the K § channels. On adding 10 ~tM ABA in the presence o f 0.1, 3 or 10 m M extracellular K § the free-running membrane potential (Vm) shifted negative-going ( - ) 4 - 7 mV in the first 5 min of exposure, with no consistent effect thereafter. Voltage-clamp measurements, however, revealed that the K+-channel current rose to between 1.84- and 3.41-fold of the controls in the steadystate with a mean halftime of 1.1 ___0.1 min. Comparable changes in current return via the leak were also evident and accounted for the minimal response in Vm. Calculated at Vm, the K + currents translated to an average 2.65fold rise in K + efflux with ABA. Abscisic acid was not observed to alter either K +-current activation or deactivation. These results are consistent with an ABA-evoked mobilization o f K + channels or channel conductance, rather than a direct effect of the phytohormone on K +channel gating. The data discount notions that large swings in membrane voltage are a prerequisite to controlling guard-cell K + flux. Instead, they highlight a rise Present address." Department of Biochemistry and Biological Sciences, University of London, Wye College, Wye (Ashford), Kent TN25 5AH, UK Abbreviations: ABA = abscisic acid; EK= K § equilibrium potential; I-V = current-voltage (relation); Kff = extracellular K § (concentration); TEA=tetraethylammonium chloride; Vm=free-running membrane potential (difference)
in membrane c a p a c i t y for K + flux, dependent on concerted modulations of K+-channel and leak currents, and sufficiently rapid to account generally for the onset of K + loss from guard cells and stomatal closure in ABA.
Key words: Abscisic acid and guard cells - Membrane potential - Potassium channel - Stomatal guard cell Vicia - Voltage clamp
Introduction Abscisic acid (ABA) is thought to function as a physiological signal during conditions of water stress and, especially, to mediate stress-induced closure o f stomata in the leaves of higher plants (see Walton 1988; also Raschke 1987 for reviews). Abscisic acid is known to be synthesized during periods of drought stress and accumulates in the leaf tissues, including the guard cells (see Cornish and Zeevaart 1986; Behl and Hartung 1986; Harris et al. 1988); the extent o f ABA accumulation usually correlates inversely with stomatal aperture (see Weyers and Hillman 1979; also discussion in Harris et al. 1988); and exogenous ABA induces closure of stomata in leaves and epidermal peels when added to the transpiration stream or bathing medium at concentrations comparable to those found in situ (l-100 gM; see Mittelheuser and van Steveninck 1969; also Raschke et al. 1975). In addition; ABA is known also as an antagonist to K + salt accumulation and stomatal opening (Jones and Mansfield 1970; Hartung 1983). Yet, despite the implicit role in modulating solute transport across the guard cell and, generally, across the higher plant cell plasma membrane, very little is known about the site of ABA action or the transport process(es) affected. One popular notion (see Raschke 1977, 1987) maintains that ABA must reduce H+-ATPase activity and electrogenic H + extrusion across the plasma membrane,
446 hence lessening the driving force for K + uptake. Consistent with this view, net H + extrusion measured from a variety of tissues has been reported to decline in the presence of ABA (Rayle 1973; Lado et al. 1975; Gepstein et al. 1982). There are scattered accounts also of membrane depolarizations in ABA; furthermore, changes in membrane potential and net K § flux have been related sensibly to the prevailing equilibrium (diffusion) potential for K § (EK) and to the reduced driving force for K § uptake, at least in Nicotiana tabacum leaf discs (Kasamo 1981). Noteworthy, too, treatments with the fungal toxin fusicoccin appear to " p r o t e c t " many tissues from the effects of ABA (see Squire and Mansfield 1972; also Ballarin-Denti and Cocucci 1979); fusicoccin has been thought to act primarily by stimulating electrogenic H § extrusion via the H§ (see Marr6 1979; also Blatt 1988a for reviews), so these observations complement notions of H § pump down-regulation by ABA. There is also some evidence which might argue to the contrary. Hartung et al. (1980) reported that ABA treatments hyperpolarized (negative-going) membrane potentials recorded from Lemna gibba to values in excess of - 2 0 0 mV and well negative of all dominant diffusion regimes. Uptake of [14C]glucose and [14C]glycine, too, were stimulated following ABA pretreatments. Furthermore, unidirectional flux analyses using Commelina communis guard cells (MacRobbie 1981) revealed that ABA had little effect on K+(S6Rb +) and CI-(82Br-) influx, but led to a large and transient stimulation of the tracer effluxes. At least at face value, neither set of data appears consistent with a simple decline in H § pump activity, much less any singular explanation for ABA action at the membrane. The seemingly conflicting results aside, it must be stressed that membrane potential and net H § flux alone are unsatisfactory as guides to transport behavior(s) in H§ membranes, including the plant plasma membrane. In effect, both voltage and net H + flux are metabolic intermediates, being " p r o d u c t s " of primary electrogenic pumping and "substrates" for secondary and H+-coupled transporters. So, for example, membrane depolarization and reduced H § extrusion in ABA cold reflect a genuine decline in H § output by the pump; equally, the phenomenology could be a consequence of A B A activating H+-coupled transport or an efflux of weak acid (e.g. Van Kirk and Raschke 1978). Likewise, few clues to ABA action can be derived from unidirectional flux measurements in the absence of any detail on the mechanism(s) for ionic flux or changes in driving force. So, without knowing the component transport currents and their voltage dependence, neither the voltage nor the flux results documented for ABA can distinguish between a primary effect on the pump or on other transport processes, or both. Electrical recordings do offer a means of dissecting out and characterizing individual transport processes in vivo, provided that membrane current can be measured under voltage clamp. Recent voltage clamp studies with intact guard cells of Vicia have demonstrated that fusi-
M.R. Blatt: Abscisic acid and guard cell K § channels coccin inactivates outward-rectifying K § channels and K § (86Rb § efflux (Blatt and Clint 1989; Clint and Blatt 1989), in addition to its limited effects on primary H + pumping (Blatt 1988a). That channel gating is altered and the channels inactivated irreversibly in fusicoccin supports the idea that the channels mediate K § loss during stomatal closure; indeed, a central feature of fusicoccin toxicity is a loss of the capacity for stomata to close (Turner and Graniti 1969). The effects of fusicoccin on K § channel current and tracer flux also raise questions about the transduction of other chemical and environmental signals in stomatal response. It is conceivable that the ABA-induced K § (86Rb+) efflux reported the MacRobbie, and stomatal closure, might be a simple consequence of membrane depolarization activating the K § channels, whether the voltage shift be achieved by down-regulating the H § pump or by stimulating an anion efflux (see Schroeder and Hagiwara 1989). Equally plausible, however, ABA could have a more direct effect on K + flux, possibly including changes in K § channel gating (Schauf and Wilson 1987) and steady-state current characteristics. The present study explores the K § channel response to ABA of intact guard cells in epidermal peels from Vicia. Measurements were carried out at the end of the growing season (September-October) when guard cells showed little indication of primary pump activity, thus greatly simplifying analyses of whole-cell recordings. Adding ABA resulted in a rapid rise in the steady-state K § current measured under voltage clamp, but without an apparent change in voltage-dependent channel gating. These data indicate that ABA modifies the capacity for K + flux at the guard cell membrane, independent of any change in the free-running membrane potential (Vm) 1, and point to a concerted action of the phytohormone on ion transport pathways in the membrane distinct from the electrogenic H § pump.
Material and methods Plant culture and experimental protocol. Viciafaba L., cv. (Bunyan) Bunyard Exhibition, was grown on vermiculite with Hoagland's Salts medium and epidermal strips were prepared as described before from newly expanded leaves taken four to six weeks after sowing (Blatt 1987a, b). Measurements were carried out in rapidly flowing solutions (10 ml/min ~20 chamber volumes/min) containing 5 mM Ca 2+-Hepes (4-(2-hydroxyethyl)-l-piperazineethansesulfonic acid), pH 7.4 (5 mM Hepes buffer titrated to pH 7.4 with Ca(OH)2), final [Ca2+] ~1 mM). Potassium salts, tetraethylammonium chloride (TEA) and abscisisc acid (ABA) were included as required. Ambient temperatures were 20-22~ C. Surface areas and volumes of impaled cells were calculated assuming a cylindrical geometry (Blatt 1987a). The orthogonal dimensions (diameter, length) of the cells, and stomatal apertures, were measured with a calibrated eye-piece micrometer. Cell dimensions in these experiments typically varied over 10-14 ~tm (diame-
1 For reasons detailed previously (Blatt 1987a; see also Parsons and Sanders 1989), electrical recordings from the guard cells are assumed to reflect the characteristics of the plasma membrane alone. I use the terms guard 'cell' and 'membrane' potential, current and conductance in this context.
M.R. Blatt: Abscisic acid and guard cell K § channels ter), and 30-35 gm (length). Estimated surface areas thus were 1.2.10-5-1.9 910 -5 cm z and cell volumes were 2.2-5.3 pl. Electrical. Mechanical, electrical and software design have been described in detail (Blatt 1987a, b). However, two new modifications were introduced in these studies. First, the epidermal peels were affixed with gentle contact to the glass bottom of the experimental chamber after coating the chamber surface with an optically clear and pressure-sensitive, silicone adhesive (No. 355 medical adhesive, Dow Corning, Brussels, Belgium). Second, all operations were carried out on a Zeiss IM inverted microscope (Zeiss, Oberkochert, FRG) fitted with Nomarski Differential Interference Contrast optics. With the epidermal peels now horizontally mounted (normal to the optical axis), the microelectrodes were brought in at an angle 33~ off the horizontal, thus providing an entirely free field of view. Recordings were obtained by the two-electrode method using double-barrelled microelectrodes. The electrodes were filled with 200 mM K§ pH 7.2 to minimize salt leakage and saltloading artifacts (Blatt 1987a), and were coated with paraffin or Sylgard polymer (Dow Coming) to reduce electrode capacitance. Current-voltage relations were made by voltage clamp under microprocessor control. Steady-state current-voltage (I-V) relations were determined by clamping cells to a bipolar staircase of command voltages (Blatt 1987 b). Steps alternated positive and negative from Vm (typically 20 bipolar pulse-pairs) and were separated by equivalent periods when the membrane was clamped to V~. The current signal was filtered by a six-pole Butterworth filter at 1 or 3 kHz ( - 3dB) before sampling, and currents and voltages were recorded during the final 10 or 20 ms of each pulse. For time-dependent characteristics, current and voltage were sampled continuously at 1, 2 or t0 kHz while the clamped potential was driven through cycles of 1-4, programmable pulse steps. Recordings at 10 kHz (above the Nyquist limit) were restricted by available data storage space to 200-ms "windows" within each cycle; otherwise data taken at all frequencies gave similar results. No attempt was made to compensate for the series resistance (Rs) to ground (Hodgkin et al. 1952). Estimates for Rs indicated that it was unlikely to pose a serious problem in measurements of clamp potential, despite the often high resistivity of the bathing media (=2.5 kf2.cm for 5 mM Ca2+-Hepes with 0.1 mM KC1). In practice, Rs is likely determined by the cell wall which might place values near 1-2 MY2 (Briggs et al. 1961 ; Blatt 1988 b). Cell input resistances near Vm ranged between 1 and 4 GO and, in the worst case for the K + conductance maximum near 0 mV, (slope) input resistance never fell below 110 MO. Hence, the maximum voltage error likely was 1-2% of the clamp potential, or less than 2 mV on clamping a cell to 0 mV. Numerical analysis. Guard cell I-V relations were fitted to polynomials, using a non-linear, least-squares algorithm (Marquardt 1963; Jennings et al. 1988). In most cases, 8th- (occasionally 10th-) order polynomials coped with the rapid changes in slope at positive potentials. Exponential fittings, likewise, were by non-linear, leastsquares. Chemicals and solutions. Tetraethylammonium chloride, ( + ) A B A and the pH buffer Hepes, were from Sigma Chemical Co. (St. Louis, USA). Otherwise, all chemicals were Analytical Grade from BDH (Poole, Dorset, UK). Where appropriate, results are reported as the mean + standard error of (n) observations.
447
evidence of primary pump activity as could be determined from the current-voltage (I-V) and conductancevoltage (G-V) characteristics of the cells, from their voltage response to external pH and K + concentration (K~-), or from their voltage and conductance behaviour during metabolic blockade with cyanide (compare Blatt 1987 a, b with Blatt 1988b and Blatt and Clint 1989). Under these conditions, free-running membrane potentials (Vm) lay positive to the K + equilibrium potential (EK) and K + channel current contributed to charge balance at Vm (Blatt 1988b; Clint and Blatt 1989). Comparable results were obtained in the present study; values for Vm and G m ( = slope conductance at Vm) were - 1 4 4 _ 5 mV and 64_ 15 ktS"cm- 2 (n = 8) in 5 mM Ca 2 § pH 7.4, with 0.1 mM KCI, and Vm showed a near-Nernstian dependence on K~- at concentrations about 1 mM and above [54___2mV/K~- decade between 1-30mM K~ (n = 6)]. The mean stomatal aperture at the start of impalements was 7.0 _ 0.3 ~tm (n = 8). The ensemble K § channel current also dominated the steady-state I-V relations of the guard cells (see also Blatt 1988 b). Clamping the membrane to voltages away from Vm revealed large, outward-rectifying (positive) currents at potentials near and positive to Vm (Fig. 1 a, open symbols) which developed sigmoidally with time (see Fig. 2) and which could be blocked rapidly and reversibly by adding 10 mM TEA. The background, or ' leak' current remaining in TEA (Fig. 1 a, filled symbols) was essentially linear at potentials negative to approx. - 1 0 0 mV and rectified only slightly at more positive voltages. In Fig. 1 the K + current is indicated by the shaded difference between the two I-V curves, _ TEA. [Tetraethylammonium chloride also introduced an additional leak current, possibly related to the anion added (see Blatt and Clint 1989). So, for purposes of estimating K + channel current near Vm -- where K + currents were small and errors introduced by the leak were proportionately large - the leak could be approximated by extrapolating a line fitted to data points ( - T E A ) over the voltage range in which the channels were deactivated (Fig. 1 a, inset; see Blatt 1988b).] It is noteworthy that the free-running potentials of the guard cells lay positive to EK (determined from K § channel tail current reversal potentials; see Blatt 1988 b). Exposures to TEA resulted in rapid and reversible positive-going shifts in Vm (Fig. 1 a, inset above), indicatng that current through the K § channels contributed appreciably to charge balance in the absence of the channel blocker. Again, an estimate for the K § current present under free-running conditions could be obtained by subtracting currents _ TEA at Vm in the absence of TEA. The data in Fig. 1 a (inset, arrow) yielded a value of 0.5 [ . t A ' c m - 2 ( = 5 . 2 p m o l K + . c m - 2 . s - 1 ) ; extrapolating a linear leak fitted to points negative from EK gave a similar, if marginally smaller figure.
Results
Steady-state characteristics. Previous experiments indicated that guard cells examined at the end of the growing season (September-November) generally showed little
Abseisic acid. Although charge (K +) transport through the K + channels is heavily biased in favour of net K + efflux (positive current) from the guard cells in 0.1 mM K~- (Fig. 1) and also at higher concentrations (Blatt
.5o
448
M.R. Blatt: Abscisic acid and guard cell K § channels
a -R4
T/IjA
-2
-60
cm
-ABA
0-'-:..,..,,
-150
-150
r
+ABA
-2.=
0"7.. m
T I n A err~2
b 40 pA cm 2
IIIIIIIIIIlllLII
I/INIII .mm,.m.
NIN NINI
400 ms Fig. 2. Voltage-dependent K + current activation and deactivation. Data from one Vicia guard cell before and 5.6 rain after adding 10 I.tM ABA to the bathing medium. Currents were recorded in 10 mM K~ to slow activation (Blatt 1988 b) with clamp steps (8) to voltages between - 6 0 and + 50 mV preceded by conditioning steps to - 1 5 0 mV from the holding potential (= Vm: --56 mV, - A B A ; - 6 1 mV, +ABA) and followed by steps to -150 mV to record deactivation. Sampling frequency, 2 kHz. The principal effect of ABA was in amplifying the time-dependent (K +) current activated by positive-going and deactivated by negative-going voltage steps. Cell parameters: surface area, 1.8.10-5 cm2; volume, 3.9 pl; stomatal aperture, 7.2 gm
,v -200
- 100
1O0
Fig. 1 a, b. Steady-state current-voltage (I-V) relations and (ensemble) K + channel current from one Vicia guard cell in 5 mM Ca 2§ Hepes, pH 7.4, with 0.1 mM KC1. a Scans (bipolar staircase of 150-ms steps, 12 s total duration) were run immediately before (o, V,.=--143mV) and 35s after (e, V m = - 8 4 m V ) adding 10 mM TEA to the bathing medium. Data points fitted to 8th-order polynomials. Potassium current as a function of (clamped) membrane potential was found as the (shaded) difference between the two curves, i ( - T E A ) i(+TEA). Cell parameters: surface area, 1.6.10-5 .cmZ; volume, 3.5 pl; stomatal aperture, 4.2 gm. At 0 mV the shaded (difference) current was equivalent to approx. 560 pA over the whole cell surface. Inset (above): voltage trace with period in TEA (shading) and times of I-V scans (carats, cross-referenced by symbol) as indicated. Inset (below): exploded view of the curves with the K + current at Vm ( - T E A ) indicated by the arrow. The dotted line was extrapolated from a linear fitting to data points between EK (=--186 mV) and - 2 2 0 mV. The slight downward displacement of the +TEA curve from this line may relate to the C1- anion added with the channel blocker. Potassium equilibrium potentials (EK) were determined as the K § channel tail current reversal potential (compare Blatt 1988b). b K + current replotted from a (shading)
1988b), v o l t a g e d e p e n d e n c e a l o n e offers few insights into t h e p h y s i o l o g i c a l c o n t r o l s on i o n flux. A net o r e n h a n c e d K + effiux in A B A m i g h t be achieved s i m p l y b y shifting Vm positive a l o n g the v o l t a g e axis, a n d c o u l d be realized by an increase in the ( i n w a r d - d i r e c t e d ) leak current. F o r e x a m p l e , a ( + ) 2 3 m V shift w o u l d suffice to raise the K + flux f o u r f o l d w i t h o u t a n y c h a n g e in the s t e a d y - s t a t e I-V profile for the e n s e m b l e K + c h a n n e l c u r r e n t s h o w n in Fig. 1. A l t e r n a t i v e l y (or in a d d i t i o n !), A B A c o u l d affect the I-V c h a r a c t e r i s t i c for the K + current itself, so as to increase the capacity for K + flux n e a r Vm. In the absence o f a n y c h a n g e in l e a k current, the effect in this case c o u l d be expected to h y p e r p o l a r i z e ( n e g a t i v e - g o i n g ) Vm. TO e x p l o r e these possibilities, m e m b r a n e p o t e n t i a l s o f g u a r d cells (eight e x p e r i m e n t s ) a n d the v o l t a g e - a n d t i m e - d e p e n d e n t features o f their K + c u r r e n t s were e x a m ined before, d u r i n g a n d after e x p o s u r e s to 10 g M A B A . C u r r e n t - v o l t a g e scans were run at intervals over 2 0 -
M.R. Blatt: Abscisic acid and guard cell K + channels
449 t~/2/ms
I /juA c m 2 90
/~
,o/
aoo
/
/
fO O
/
30o/
/
100
10
- 12 0
~
f~/~/f/i /
...., ...'-'"" V / m V
65
~, ......... '..................... 2 6 " ~" -10
Fig. 3. Instantaneous (leak) and net steady-state currents from Fig. 2. Data points are for currents recorded before (open symbols) and after (/~lled symbols) adding ABA. Instantaneous currents (squares) were determined 2 ms following positive-going steps from - 1 5 0 mV; steady-state currents (circles, K + channel plus leak currents) were taken at the end of each clamp step between - 6 0 and + 50 inV. Points at - 1 5 0 mV are currents taken at the end of the conditioning voltage steps
100
ms L
+ 5o _ . . . . ~
-IsK
"
0-'-~
| 3 0 pA cf~ 2
_
_
-1~~
C'
[
- a'0
'
' -10
b 1
25 m i n following each i m p a l e m e n t before A B A was added to ensure a stable baseline o f K + current. M e a s u r e m e n t s were carried o u t in 0.1, 3.0 a n d 10 m M K~-, all with c o m p a r a b l e results. I n each case, Vm shifted n e g a t i v e - g o i n g d u r i n g the first 5 rain in A B A [range ( - ) 4 - 7 mV] with n o consis-
b
'
3
'
50
V /mY
Fig. 5. Insensitivity to ABA of the voltage-dependence for K + current activation. Halftimes (tl/z) for data taken in 10 mM Kg from the same cell as in Fig. 4. Steps to the voltages indicated were preceded by clamping to a conditioning voltage of - 150 mV. Clamp cycles were initiated before (o) and after 8.3 (m) and 11.2 (o) min in 10 gM ABA. Where necessary to accommodate current scatter, halftimes (+SE) were taken after first fitting the data to 8th-order polynomials (see Fig. 4) by non-linear least-squares (Material and methods). A joint fitting to a single exponential (solid line) gave an e-fold rise in tl/2 per (-)71.3 mV
Table l. Insensitivity to ABA of activation and deactivation kinetics for Vicia K + channel current. Summarized are data (n = 5) sampled at 2 kHz during exposures to 10 mM Kg which slows voltagedependent activation (see Blatt 1988b). In each case, activation at +50 mV was preceded by a 800-ms conditioning step to - 150 mV and deactivation was followed during a subsequent step to this same voltage (see Fig. 2). Halftimes for activation were determined by eye after subtracting instantaneous (leak) currents; for deactivation, data points were fitted satisfactorily to single, falling exponentials (Material and methods). Exposure times to 10 gM ABA were > 5 min in each case
3 0 0 ms
Fig. 4. Insensitivity of K + current activation to ABA. Data from one Vicia guard cell in 10 mM K~- before and 8.3 min after adding I0 gM ABA. Currents recorded at 2 kHz during voltage clamp cycles as in Fig. 3. Holding potentials (Vm), --58 mV ( - A B A ) and - 6 5 mV (+ABA). The time-dependent current obtained at + 50 mV before adding ABA was fitted to an 8th-order polynomial for clarity, and then expanded on the ordinate axis by a factor of 2.35 to superimpose on the current recorded in ABA. Halftime for activation, 109 ms. Inset: raw current data and corresponding voltage steps. Cell parameters: surface area, 1.4.10- s cm 2; volume 3.1 pl; stomatal aperture, 6.7/am
'
tl/2 (ms)
Activation (+ 50 mV) Deactivation ( - 1 5 0 mV)
-- ABA
+ ABA
110 + 4 10.8 + 0.2
108 + 3 10.8 ___0.3
tent c h a n g e thereafter. C l a m p i n g the m e m b r a n e to voltages n e a r a n d positive to 0 mV, however, revealed the full i m p a c t o f A B A o n the K + current. D a t a for one g u a r d cell in 10 m M K~- before, a n d 5.6 m i n after a d d i n g A B A are s h o w n in Fig. 2. C l a m p steps were preceded by a c o n d i t i o n i n g voltage ( - 150 mV) to deactivate the K + current. Stepping the voltage to values positive from Ez ( = - 7 9 m V for this cell in 1 0 r a M K~-, d a t a n o t shown) t h e n gave a n i n s t a n t a n e o u s j u m p in (leak) current followed by a t i m e - d e p e n d e n t , rising c o m p o n e n t
450
M.R. Blatt: Abscisic acid and guard cell K + channels
a
-2
I / p A cm
70 f ~176176149
? ? 7 50
/11
?
?
?
,o.o.o'~176
10
OS . 0 4
V /mV
-250 . _=e,%~176 . ""~176176 t '@i~@i@@~mi~m'~'~ -100
:
100
-10
b
-2
T tOA cr~2 80
I /pA crn 0
12 ' 40
0
,
! o
8
-2~5 ,~
/
!
/./ .(
-21o
0
4
.........
....S . / "
V/mY
'
if/':
IK, V~ (I- t A ' c m - 2)
- ABA
+ ABA
0.49 -I-0.09
1.30 -I-0.18
,o-
d
p/
Table 2. Enhancement by ABA of K + current at the free-running potential, IK. Vm' Data from five Vicia guard cells before and after > 5 rain exposure to 10 pM ABA determined under steady-state voltage clamp (see Fig. 6). Analyses of measurements in 0.1, 3 and 10 mM K~ gave comparable currents at V,,, and the results are pooled accordingly
t,
Fig. 6a, b. Steady-state current-voltage response to ABA. Data from one Vicia guard cell in 0.1 mM K~- before (o) and 9.3 min after (o) adding 10 I.tM ABA to the bathing medium. Currents were averaged over the final 20 ms during each 150-ms step of bipolar staircase scans (Material and methods) and were decomposed into K § channel and leak components by linear extrapolations of the leak from points negative to EK (=--178 mV, data not shown) as described in Fig. 1. a Whole-cell I-V data. b Detail of the data in a with the component K § channel (solid and dashed lines) and leak currents (dotted and dot-dashed lines) as indicated ( - A B A , solid and dotted lines; +ABA, dashed and dot-dashed lines). Arrows indicate K + channel (1') and leak (+) currents at Vm -I-ABA. Potassium current at Vm: -ABA, 0.85 IxA'cm-2; +ABA, 1.64 ixA.cm-2. Inset: K + currents determined by subtracting I-V curves in TEA (not shown) before (solid line) and after (dashed line) adding ABA. Cell parameters: surface area, 1.7-10-5 cm2; volume, 4.4 pl; stomatal aperture, 8.1 Ixm.Currents in ABA were 765 pA at 0 mV and 1.1 nA at the maximum near + 60 mV as the K + current activated (see Blatt 1988b). Subsequently clamping to - 1 5 0 mV, negative from EK, showed an inward (negative) current which decayed over approx. 50 ms as the K + current deactivated. After 5.6 min in A B A the same clamp protocol yielded time-
dependent (K + current) components which were amplified roughly 2.4-fold at each voltage. Figure 3 summarizes the instantaneous and steady-state I-V characteristics for the data in Fig. 2. One intriguing feature of this A B A response was an apparent lack of effect on the voltage-dependence of channel gating. Abscisic acid enhanced the K + current in purely scalar fashion, that is, without any significant change in the time course for current activation or deactivation. Figure 4 illustrates this point for current activation with data taken in 10 m M K~- from another cell using the same pulse protocol; scaled to a c o m m o n ordinate, the time-dependent currents at + 50 mV superimpose with a c o m m o n halftime for current rise of 109 ms. The current measured in A B A was 2.35-fold greater in magnitude than the control, however (Fig. 4, inset). Halftimes for current rise as a function of clamp potential and A B A exposure are summarized in Fig. 5 and Table 1, including data from Fig. 2. Complementary results were obtained also in 0.1 and 3 m M K~, although lowering K~ concentration, itself, accelerated current activation (Blatt 1988b). Halftimes for current deactivation at - 1 5 0 m V (3 and 1 0 m M K +) and - 2 2 0 m V (0.1 m M K~) were 10.8 and 8.1 ms, respectively, and are comparable to values determined from previous experiments under similar conditions (Blatt 1988b; Blatt and Clint 1989); in ABA, deactivation varied in all cases by less than i 2% of the controls and without any consistent trend (Table 1).
Membrane potential and the leak. The instantaneous currents recorded on stepping positive from - 150 mV (see Figs. 2, 3) and - 220 mV also indicated a limited change in leak current with ABA. In fact, the near-constant Vm in the face of a rise in K + current anticipated this observation; an appreciable rise in current directed out of the cell otherwise might have driven Vm well negative in the face of a constant leak. However, to confirm this expectation and to determine the extent to which the K + current was enhanced at Vm, steady-state I-V curves were determined using a bipolar staircase protocol and currents were time-averaged at the end of each voltage step to improve the small-signal resolution for currents near Vm (see Material and methods). Potassium channel and leak currents near V,, were assessed by linear extrapolations, as outlined in Fig. 1 Figure 6 shows measurements from one cell which were taken in 0.1 m M K~-. Equivalent analyses carried
M.R. Blatt: Abscisic acid and guard cell K § channels out on data f r o m four other cells gave similar results, and K + currents at Vm + A B A are summarized in Table
451
T /~A
a
c m -2
140
2. From Fig. 6 it will be seen that adding ABA was
followed by a 2.3-fold rise in K + current at any one voltage, and Vm was displaced 4 m V negative from the control ( = - 136 mV) as indicated by the V-axis intercepts for the data points (Fig. 6 b). Now, however, this limited change in Vm could be tied directly to a concomitant rise in the (inward-directed) leak which balanced the K § current. At Vm ( - A B A ) , the K + current rose from 0.85 laA.cm -2 in the control (Fig. 6b, solid line) to 1.94 IxA.cm -2 after 9.3 min in A B A (Fig. 6b, dashed line); but this enhancement was nearly matched by changes in current return via the leak (Fig. 6b: dotted line, - A B A ; dot-dashed line, + A B A ) . So, charge balance was maintained with little change in the free-running potential (at Vm, im= 0 = i~:+ iL, where iK is the K + current and the leak current, iL, is the sum o f currents passing through all other charge transport pathways). By contrast, had the leak remained unaffected. A B A would have driven Vm negative by 1 3 m V to - 1 4 9 m V where the leak in the control ( - A B A , dotted line) and the K + current in ABA (solid line) were equal in magnitude and opposite in sign. A b s c i s i c - a c i d t i m e course. A singular feature of the physi-
ological response to A B A is the speed with which stomata close in the presence of the phytohormone. The most detailed time courses available (e.g. Cummins et al. 1971; Raschke et al. 1975; Weyers and Hillman 1979) indicate that guard cells must begin to lose salt and osmotic pressure within a few tens o f seconds to minutes following exposure to ABA, although 1 0 4 0 min m a y be required before the response is complete and the stom a t a close fully. So, it could be expected that the K § current should follow a similar time course and rise rapidly to a new quasi-steady-state in ABA, if the K § channels were to qualify as a notional pathway for K § flux during stomatal closure 2. Because of the speed with which the cells clearly did respond to the p h y t o h o r m o n e , data for times early on following ABA addition, again, were determined using a bipolar staircase protocol. Whole-cell and K § currents from one guard cell are shown in Fig. 7 with the corresponding response timecourse in Fig. 8. In this case adding l0 ~tM A B A resulted, at any one clamp potential, in an approx. 2.1-fold rise in K + current and conductance (Fig. 7b, inset) which was essentially complete within 6 min (tl/2, 1.08 min). C o m p a r a b l e results were obtained in the other experiments, giving a pooled (n = 6) z Published time courses are for ABA-induced stomatal closure or for stomatal (gaseous) conductance. These parameters index the solute content and, broadly speaking, the K + concentration in the guard cell. As a measure of the rate of cation movement, the K § current must relate to the derivative of these ABA time courses, that is, to the rate of change in K § concentration with time. Hence, the rise in the K § current should correspond to the onset of stomatal closure, and the current should attain at a new quasi-steady-state consistent with the longer times required to complete closure.
10.2 rnin ilillll 101~i
ooo D
=' o~
,llo',~ =o'e /,e
4. I rain
1.7 rain
i'ft'~o~176176176 Control
I
. .p
