Planta

Planta (1992)187:26-36

(c~ Springer-Verlag1992

Compartmental distribution and redistribution of abscisic acid in intact leaves II. Model analysis Stefan Slovik* and W o l f r a m H a r t u n g

Julius von Sachs-Institut fiir Biowissenschaften, Lehrstuhl Botanik I, UniversitS,t Wfirzburg, Mittlerer Dallenbergweg 64, W-8700 Wfirzburg, Federal Republic of Germany Received 19 October 1990; accepted 9 October 1991

Abstract. A computer model written for whole leaves and described in the preceding publication (Slovik et al. 1992, this volume) has been developed for calculating the distribution and fluxes of weak acids or bases amongst different leaf tissues and their compartments, considering m e m b r a n e transport, transpiration-driven mass transport, symplasmic and apoplasmic diffusion, and metabolic turnover rates in specified compartments. The model is used to analyse flux equilibria and the transport behaviour of the p h y t o h o r m o n e abscisic acid (ABA) in unstressed and stressed leaves. We compare experimental data of unstressed Valerianella locusta L. leaves and expectations based on the detailed analysis of the data. (i) The mean daily influx of A B A into the leaf lamina via the xylem sap is a b o u t 10 nmol 9m - 2 day-1. It is balanced by the sum of an export of ABA via the phloem sap (0.7%), possibly also by a basipetal ABA transport in the petiole p a r e n c h y m a of young leaves (up to 18%), by an irreversible conjugation of ABA (0.4-4%) and by net degradation of ABA in the leaf lamina (80-95%). (ii) The estimated kinetic parameters of this net degradation are for the mesophyll apoplasm: apparent Kin= 3.7 n M and Vmax= 12.9 nmo1 - m -3 9 s -1, or for the mesophyll cytosol: apparent K m--8.1 n M and Vm,x= 32.3 nmol 9m -3 9 s - 1. (iii) The dynamic ABA concentration in the phloem sap o f Valerianella is 2.8 nM. This is only 5.5% of the static A B A equilibrium concentration in excised leaves or 70% o f the ABA concentration in the mesophyll apoplasm, and it equilibrates within a few hours after source concentrations in the mesophyll a p o p l a s m are changed under stress. Thus, the phloem sap is a flexible medium for transporting 'new p h y t o h o r m o n e information' from the lamina to the shoot and roots. (iv) Measured compartmental ABA concentrations are close to calculated equilibrium concentrations in unstressed leaves. We conclude that model calculations are close to reality. (v) p H gradients within .

Abbreviations: ABA = abscisic acid; CON = ABA conjugates; HABA = neutral ABA species 9 To whom correspondence should be addressed

the apoplasm influence the apoplasmic distribution of ABA. Its concentration is maximally about twofold higher in guard-cell walls relative to the mesophyll apoplasm. (vi) Unexpectedly, all compartmental equilibrium concentrations of ABA in the leaf lamina depend on plasm a l e m m a conductances for undissociated ABA and on the transport properties of the plasmodesmata. This is a consequence of the cyclic diffusion pathway: mesophyll c y t o s o l - mesophyll plasmalemma mesophyll apoplasm - epidermal apoplasm epidermal plasmalemma - epidermal cytosol - plasmodesmata - mesophyll cytosol (in this direction), if there are different apoplasmic or cytosolic p H values in both tissues. The cyclisation rate is 42 fmol 9 s- 1. m - 2 leaf area, which corresponds to a turnover time = 11.0 h for the total ABA content within the leaf lamina. A decrease of the epidermal plasmalemma conductance by 90 % yields a threefold ABA concentration in the guard-cell free space. (vii) C o m p a r t m e n t a l relaxation-time coefficients are estimated and summarised for all leaf tissues and its major compartments. They range from 1.5 min for chloroplasts up to 3.3 d for mesophyll vacuoles. (viii) The highest ABA concentration, which can be expected in any leaf compartment, is 7 m M in the guard-cell cytoplasm of certain plant species. (ix) We employed circadian changes (equal day + night, 12 h each = equinoctium) of the stromal p H +0.3 in C3 plants, and for Crassulacean acid metabolism (CAM) plants, additionally, vacuolar pH-+_ 2.5 changes, and calculated the consequences for ABA redistribution within the lamina. In plants of both photosynthesis types, the ABA concentration in guard-cell walls is only 1.5 times higher in the night relative to the day. We conclude that stomata m a y not be regulated by ABA in a night-day regime. The influence of the extreme vacuolar p H changes on ABA distribution is small in C A M plants for two reasons: the ABA content in C A M mesophyll vacuoles is low (maximum 2.7% of the total ABA mass per unit leaf area) and there is only a 6.5-fold increase of the mole fraction of undissociated ABA when the the vacuolar p H is lowered from 5.5 to 3.0 (importance of the absolute pKa = 4.75 of ABA).

S. Slovik and W. Hartung: ABA distribution in leaves: II. Model analysis Key words: Abscisic acid (compartmentation in leaves) - Crassulacean acid metabolism plant - C3 plant - Computer model (ABA compartmentation) - Leaf (ABA compartmentation) - Phytohormone compartmentation

Introduction In order to understand the influence of phytohormones on plant responses and developmental physiology, it is necessary to know the p h y t o h o r m o n e concentrations (and ratios) in tissues and cell compartments at any time. However, it is impossible to examine experimentally the time course of p h y t o h o r m o n e concentrations (in our case of abscisic acid (ABA)) in all leaf compartments simultaneously. To overcome this difficulty, we deduce the distribution and redistribution of ABA on the basis o f the physical and chemical properties o f ABA and the anatomical, biophysical and biochemical properties of the plant tissues and its cell compartments in order to integrate all experimental data to a physiological sourcesink network, which can be studied by model calculations. In the preceding communication (Slovik et al. 1992), we have described a computer model which permits the analysis of ABA distribution amongst different leaf compartments and the kinetics o f its redistribution when conditions are changed. The model considers an entire leaf consisting of chlorenchyma cells, epidermal cells, guard cells, phloem and xylem. It formulates membrane transport processes of the weak acid ABA (molecule permeation, anion transport, carrier transport), symplasmic (plasmodesmata) and apoplasmic (cell wall) diffusion processes, mass transport of water-dissolved ABA (cell wall, xylem, phloem) and ABA metabolism including its conjugation. We tried to consider all presently available quantitative information on ABA physiology in leaves in the context o f a coherent mathematical network. Most of the essential morphological and physiological input data were determined for Valerianella loeusta L. We expect the mathematical network of the model to be close to reality. In this communication, we compare measured and calculated ABA concentrations, we estimate kinetic parameters o f ABA metabolism, we present compartmental relaxation-time coefficients, and we demonstrate the redistribution of ABA in the unstressed model leaf employing a diurnal light-dark regime for C3 and Crassulacean acid metabolism (CAM) plants. Results and discussion

Equilibration of the model leaf, disregarding ABA metabolism (case 1). In Table 1, the measured concentrations of ABA in different leaf cell compartments are compared with equilibrium concentrations (case 1), which can be calculated employing the model constants described by Slovik et al. (1992). There is no agreement between measured concentrations and data which are calculated assuming only flux equilibration (i) within the model leaf

27

lamina and (ii) between the lamina and the remaining plant corm, i.e. assuming ABA-flux equilibrium at the petiole. Case 1 disregards any metabolic processes of ABA in the leaf besides reversible ABA conjugation. A realistic mean transpiration rate per unit leaf area of Valerianella plants grown in our greenhouse is 3.2 m m o l . m - Z - s -1 in the light (relative humidity R H --60 %, leaf temperature Tleaf= 25 ~ C, air temperature Tair=25 ~ C, stomatal water conductance C n 2 o = 0 . 6 c m - s - 1 ; or: R H = 1 0 0 % , Tleaf=25 ~ C, Tair = 20 ~ C , C H 2 o = 1.0 c m ' s-1). Assuming these transpiration data, the ABA import flux per unit leaf area is about 240 fmol 9m - z . s-1 into the lamina via the xylem sap (ABAx= 4.1 nM). It is balanced by an ABA-export flux in the phloem only if the ABA concentration in the phloem sap is 377 n M (flow rate 6 3 3 - 1 0 - 6 m 3 "s -1 - m -2 leaf area; cf. Slovik et al. 1992). This concentration is in equilibrium with an ABA concentration of 543 nM in the mesophyll apoplasm. However, at equilibrium, all other compartmental ABA concentrations are much too high (Table 1, case 1) compared to available measurements. The calculated total ABA content per unit leaf area is about 290 times higher than determined. We conclude that the phloem sap is quantitatively an unimportant sink for ABA within the leaf lamina. As there are no saturatable ABA-accumulating processes at the phloem-cell plasmalemma of Plantago maius (Baier and Hartung 1991), we are forced to conclude that in situ the leaf lamina of Valerianella is not in ABA-flux equilibrium at the petiole. Either (i) there is indeed a net import and accumulation of ABA while the leaf ages, or (ii) there must be potential sinks for ABA within the leaf besides the phloem sap (degradation o f ABA, polar parenchymatic ABA export at the petiole, irreversible ABA conjugation). Concerning the first explanation, we assume a mean transpiration rate of 3.2 m m o l - m -2 9 s -1 in the light (cf. above). When day and night are of equal length, (with closed stomata in the dark) the mean ABA influx into the lamina is 0 . 5 . 8 6 4 0 0 s - d - 1 . 2 4 0 f m o l . m - 2 - s -1 = = 10.4 nmol 9 m -z 9d -1. At flux equilibrium of ABA at the petiole we expect a total ABA content of 475 n m o l . m -z leaf area (Table 1, case 1), i.e. within maximally about 6.5 weeks the leaf lamina should approximate the mentioned equilibrium. This is about the age of Valerianella leaves when harvested for experiments, but we could measure only a total content of free A B A = 2 . 6 8 _ 0 . 9 6 n m o l . m -2 leaf area. As the phloem sap compensates only about 0.7 % of the ABA import via the xylem (cf. below), we conclude that there must indeed be other, more potent sinks for ABA somewhere in the leaf. Let us now consider the basipetal transport of ABA from the leaf lamina to the shoot, as can be studied in the petiole parenchyma of young leaves (D6rffiing and B6ttger 1968). We have re-examined the existence o f a polar ABA transport in young leaf petioles of Phaseolus, but found no basipetal transport at all (data not shown). D6rffiing and B6ttger measured a basipetal transport velocity of 30 mm 9h - 1, which balances only about 18% of the mean daily ABA import rate of 120 fmol 9m - 2 . S-- 1 ( 4 . 1 nM ABAmw in the mesophyll apoplast, mean parenchyma area in a cross section of

28

S. Slovik and W. Hartung: ABA distribution in leaves: II. Model analysis

Table 1. Measured compartmental concentrations of ABA and its hydrolysable conjugates (CON) in Valerianella loeusta leaves, and equilibrium concentrations of ABA (nM) as calculated using the model. We discuss four different cases. (1) relaxation to equilibrium without any ABA net synthesis or net degradation in the lamina; (2) equilibrium for a constant (measured) ABA concentration of 4.1 n M in the mesophyll apoplasm; (3) equilibration with a constant (measured) ABA concentration of 9.0 n M in the mesophyll cytosol; (4) equilibration as in case 3, but with ABA conductances of the epidermal plasmalemma and tonoplast being only as small as in mesophyll cells. Additionally, we summarise the total ABA and C O N mass per unit leaf area (nmol 9m-2). Case 1 : Equilibrium of the model leaf described by Slovik et al. (1991) if net metabolism of ABA in the lamina (besides its conjugation in vacuoles) is absent; Compartment"

ABA accumulates until the phloem export of ABA compensates the import of ABA via the xylem. Case 2: The ABA concentration in the mesophyll apoplasm of 4.1 n M is kept constant. This is achieved by a mean daily ABA degradation rate of 6.8 nmol - m -3 9 s -1 in the mesophyll apoplasm of unstressed leaves. Case 3: The ABA concentration in the mesophyll cytosol of 9.0 n M is kept constant. This is achieved by a mean daily ABA degradation rate of 17.0 n m o l . m - 3 . s-1 in the mesophyll cytosol of unstressed leaves. The ABA conductances of the epidermal tonoplast and plasmalemma are about the same as in guard cells: 250 - 10- 9 m 9s - t at the plasmalemma and 13 9 10 9 m - s -x at the tonoplast. Case 4: As for case 3, but the ABA conductances of the epidermal tonoplast and plasmalemma are only as small as in mesophyll cells: 25 9 10 -9 m 9 s 1 at the plasmalemma and 5 - 10 -9 m 9 s -~ at the tonoplast

ABA concentration (nM) Determined b

Calculated Case 1

Case 2

Case 3

Case 4

419 415 543

3.16 3.13 4.1

3.57 3.54 7.29

9.59 9.52 11.6

Apoplasm G u a r d cells Epidermal cells Mesophyll cells

(4.1) (4.1) 4.1

Cytosol Mesophyll Epidermis Guard cells

9.0 55.1 426.4

5 780 6 950 29 600

43.5 52.4 222.5

9.0 41.4 251.8

9.0 28.2 676.0

35.7 467.6

22 900 32 500

173.0 244.0

35.7 276.1

35.7 741.2

2.8 4.8 5.6 3.8 195.7

631 1 079 711 483 20 400

4.74 8.1 5.38 3.66 172.6

1.17 2.00 4.26 2.90 195.3

0.99 1.69 2.98 2.03 523.3

377

2.85

5.06

8.05

475.0 153.3

3.60 1.15

1.34 0.37

1.65 0.29

Chloroplasts Mesophyll Guard cells Vacuoles Mesophyll ABA CON Epidermis ABA CON G u a r d cells ABA Phloem Sieve-tube sap Leaf ABA [nmol 9 m -2] CON [nmol 9m -2]

(12.0) 1.64 b 0.75 b

a See Slovik et al. (1992); ABA denotes free abscisic acid, CON denotes hydrolysable ABA conjugates b Values in parentheses are only estimations; see Appendix in Slovik et al. (1992). The sum of the tabulated values is recalculated on a leaf-area basis. Our independently measured contents in Valerianella locusta are (2.68 -4-0.96) nmol ABA 9 m - 2 leaf area (n = 3), and (1.13 _-4-0.54)nmol C O N . m - z leaf area (n = 2)

Valerianella p e t i o l e s = 0 . 6 4 9 10 . 3 m 2 9 m 2 l e a f a r e a ) . E v e n i f a p o l a r A B A t r a n s p o r t e x i s t e d , it is n o t s u f f i c i e n t to compensate the mean ABA influx via the xylem. There m u s t b e o t h e r A B A s i n k s in t h e l a m i n a . N o w w e d i s c u s s the 'sink capacity' of an irreversible net conjugation of ABA, for which there are experimental indicat i o n s ( W e i l e r 1 9 8 0 ; K a i s e r et al. 1985). A g a i n , w e assume a mean ABA influx rate into the lamina of 10.4 n m o l 9 m - 2 9 d - a (cf. a b o v e ) , a n d - as o b s e r v e d - a m o r e o r less c o n s t a n t A B A c o n t e n t p e r u n i t l e a f a r e a w h i l e t h e l e a v e s age. I f all A B A i m p o r t e d v i a t h e x y l e m s a p w e r e t o b e c o n j u g a t e d i r r e v e r s i b l y i n Valerianella, t h e n t h i s m e a n A B A flux y i e l d s a n a c c u m u l a t i o n o f A B A c o n j u g a t e s w i t h i n 150 d o f u p t o 1.55 ~tmol - m -2. T h i s is m o r e t h a n a 2 0 0 0 - f o l d i n c r e a s e o f t h e c o n j u g a t e c o n -

t e n t p e r u n i t l e a f a r e a (cf. T a b l e 1). W e i l e r ( 1 9 8 0 ) o b served an accumulation of total ABA conjugates only up t o a f a c t o r o f 8.6 w i t h i n five m o n t h s f o r Betula papyrifera o r 10.4 i n 5.5 m o n t h s f o r Acer pseudoplatanus. T h i s is a m e a n a c c u m u l a t i o n f a c t o r o f 9 i n five m o n t h s . T h e s e t r e e s , g r o w n i n t h e field, m u s t h a v e h a d e v e n h i g h e r t r a n s p i r a t i o n r a t e s t h a n Valerianella g r o w n i n o u r g r e e n house. Thus, we conclude that about 100.9/ 2 0 0 0 = 0 . 4 % o f t h e i m p o r t flux o f A B A i n t o t h e l e a f lamina can be conjugated irreversibly. We have also found a net synthesis rate of hydrolysable ABA conj u g a t e s o f 1.1 g m o l 9 m - 2 d -~ w h e n a n A B A c o n c e n t r a t i o n o f 10 ~tM is f e d v i a t h e p e t i o l e o f e x c i s e d Valerianella l e a v e s ( d a t a n o t s h o w n ) . T h e e m p l o y e d c o n c e n t r a t i o n is a b o u t 2.5 9 103 t i m e s h i g h e r t h a n t h e A B A c o n c e n t r a t i o n

S. Slovik and W. Hartung: ABA distribution in leaves: II. Model analysis in the mesophyll apoplasm in situ. Assuming that the conjugation rate is linearly correlated with the ABA concentration in the mesophyll apoplast, we estimate an in-situ conjugation rate o f 0.44 nmol - m -z . d-1. This is about 4.2% o f the ABA import rate into the Valerianella leaf lamina. We conclude that the conjugation o f ABA within the leaf is only a minor sink for ABA and is therefore unimportant. We summarise the relative importance o f the 'ABA sinks' in the leaf so far discussed; phloem sap: 0.7%; basipetal transport: 18%; ABA conjugation: 0.4 to 4%. Thus, import must be balanced by net degradation of ABA, which needs to compensate for about 80% o f the ABA imported via the xylem sap in young leaves and for as much as 95% in mature leaves. So far, we have discussed the flux balance of ABA between the leaf and the shoot of Valerianella. Now we regard the flux balance of ABA within the leaf lamina and its compartments.

Equilibration of the model leaf, keeping the measured apoptasmic ABA concentration in the mesophyll constant (ease 2. Table 1). We now calculate for ABA and its conjugates all compartmental equilibrium concentrations, which can be expected if the ABA concentration in the mesophyll apoplasm is kept constant at the measured value o f 4.1 nM. In this case there is fair agreement between measured and calculated concentrations. F o r epidermal cells, the differences are small. The simplest way o f explaining these findings is to conclude that in epidermal cells the employed compartmental p H values and volumes are close to reality, and that the distribution of ABA at the epidermal plasmalemma and tonoplast can indeed be calculated by the Henderson-Hasselbalch equation, i.e. the permeability o f the membranes o f epidermal cells to A B A - is low. Additionally, there is no fast net synthesis or net degradation o f ABA in epidermal cells o f unstressed leaves. Indeed, stripped epidermes o f Commelina show no accumulation and no loss o f ABA (Dtrffiing et al. 1980). However, it should be stressed that our conclusions on epidermal cells are preliminary. Experimental data for epidermal cells are incomplete. It is possible to calculate exactly the measured bulk ABA concentration in epidermal strips on the basis o f numerous other, but more complex, postulates. We propose the simplest explanation. F o r guard cells, which cannot synthesise and catabolise ABA at all (Dtrffiing 1983; Behl and Hartung 1986; Lahr and Raschke 1988), the ratio between measured and calculated ABA concentrations is higher, but is does not exceed a factor of about 1.9. As there is good agreement for guard-cell vacuoles, we think that the deviation by a factor o f 1.9 in the guard-cell cytosol and chloroplasts is not o f consequence: for experiments, guard cells were 'isolated', but in the model they are integrated into the lamina. The equilibration o f the model allows us to estimate the ABA concentration in the epidermal apoplasm, which is about the same as in guard-cell walls, but only 76 % of the ABA concentration in the mesophyll apoplasm. The concentrations are not identical even at equilibrium because the apoplastic p H values are different in different tissues, but the influence o f p H gradients within the apoplasm on

29

its ABA gradients is small. We conclude that apoptasmic ABA concentrations, measured in the exudate o f a pressure chamber (mesophyll apoplasm), also approximate apoplasmic ABA concentrations in epidermal and guardcell walls, but the deviation can be up to a factor o f 2 (cf. case 3 below). Model equilibration also yields an estimation of the unknown ABA concentration in the streaming phloem sap o f Valerianella, which is 2.85 nM. This is 0.7 % relative to the equilibrium ABA concentration (and ABA-export rate) in the phloem sap as calculated for case 1, Table 1. Finally we compare measured and calculated ABA concentrations in the chlorenchyma cell compartments (Table 1, case 2). The calculated vacuolar ABA concentration is only 1.7 times higher than measured, but the calculated ABA concentrations in the mesophyll cytosol and chloroplasts are 4.8 times higher. The last deviation is considered to be important: we conclude there must be a sink for ABA in the mesophyll cytosol o f Valerianella, i.e. there must be a small cytosolic net degradation o f ABA, which indeed could not be prevented completely in the presence o f Tetcyclacis (inhibitor o f ABA degradation) during compartmentaleffiux-analysis experiments (cf. Daeter and Hartung 1990). Still, isolated protoplasts o f Hordeum vulgate do not metabolise ABA (Loveys and Robinson 1987). Alternatively, there may be degradation o f ABA in the leaf apoplasm, as has been postulated by Loveys and Robinson (1987). Indeed, we have found apoplasmic ABA degradation in Gossypium hirsutum and Valerianetla loeusta (data not shown). As epidermal cells do not show net degradation of ABA (cf. above), there are only two compartments where the postulated ABA net degradation (cf. above) can be localised (cf. Slovik et al. 1992): either in the mesophyll cytosol, or in the apoplasm (or in both). We calculate both compartmental alternatives separately and discuss the consequences for the equilibrium concentrations o f ABA in the model-leaf compartments. One alternative has already been presented in case 2 by assuming a constant ABA concentration o f 4.1 n M in the mesophyll apoplasm, which artificially constitutes an apoplasmic 'ABA-stat-metabolism'. The other alternative is presented in case 3 (Table 1). Without apoplasmic ABA degradation the equilibrium concentration o f ABA in the mesophyll apoplasm would increase by a factor o f more than 100 (cf. case 2 relative to case 1 in Table 1). In order to keep the measured apoplasmic ABA concentration of 4.1 nM constant, we postulate ABA degradation in the mesophyll apoplasm of 0.5. 240 fmol 9m -z 9s -1 = 120 fmol 9m -z 9 s -1 at equinoctium (compare above), an a m o u n t which just compensates the mean daily ABA influx into the unstressed leaf lamina via the xylem. The necessary catabolic rate in the mesophyll apoplasm is 120 fmol 9 m -z - s-i/(17.7 . 10 -6 m 3 9m -2) = 6.8 nmol - m -3 9s -1 in the mesophyll apoplasm o f unstressed leaves. The denominator is the mesophyll apoplasm volume per unit leaf area (cf. Slovik et al. 1992). This apoplasmic rate is, on a compartmentvolume basis, about 150-million times smaller than a typical photosynthetic rate in the chloroplast stroma (1.0 mol CO2 9m -3 9 s-1). This rate is sufficient to catabolise the measured total ABA content per unit leaf area

30

S. Slovik and W. Hartung: ABA distribution in leaves: II. Model analysis

of 1.64 nmol 9m 2 (Table 1) within a turnover time of about 3.8 h. The maximum ABA-degradation rate of Phaseolus vulgaris leaves, which can be observed during recovery after a wilting stress, corresponds to a turnover time o f 2.0 h (Pierce and Raschke 198t). If Valerianella leaves have a similar minimal turnover time, i.e. maximum degradation rate relative to the total ABA mass per unit leaf area, then it is possible to estimate the apparent Km and the Vmaxof the ABA degradation process(es) in the mesophyll apoplasm. It should be stressed that we estimate only apparent parameters assuming MichaelisMenten kinetics', but we make no assumptions with regard to mechanisms. For case 2 the mesophyll apoplasm is the only ABA-catabolising compartment in the leaf lamina. Vm,x is 3.8 h/2.0 h = 1.9 times higher than the rate V = 6 . 8 nmol 9m -3 9 s -~, which is necessary to compensate the mean ABA influx into the unstressed lamina, i.e. Vmax=12.9 nmol 9m - 3 . s-1. The Km=3.7 nM can be estimated by employing the rearranged Michaelis-Menten formula, where [ABAm,Lq=4.1 nM is the equilibrium concentration of ABA in the mesophyll apoplasm: Km = [ A B A ~ ] e q . ( - ~ -

1)

(Eq. 1.1)

These two kinetic constants are preliminary estimations, calculated on the basis of experimental data, but indirectly. It was not necessary to employ the model. Because of a lack o f measured compartmental kinetic data for ABA metabolism, they cannot as yet be compared with our estimations. Seventy percent o f the total apoplast volume o f Valerianella is mesophyll apoplasm (cf. Slovik et al. 1992). Therefore we attributed the total net degradation per unit leaf area to the mesophyll apoplasm. Since, for case 2, the ABA concentration within the apoplasm is more or less uniform (Table 1), it would, alternatively, be possible to attribute 70% of the apoplasmic degradation rate to the total free space.

Equilibration of the model leaJ~ keeping the measured cytosolic ABA concentration in the mesophyll constant (case 3, Table 1). Alternatively, the postulated net ABA degradation may be localised in the mesophyll-cell cytosol. We will now keep the measured cytosolic ABA concentration of 9.0 nM artificially constant in the chlorenchyma cells and investigate equilibration. The compartmental equilibrium concentrations now also approximate the measured data (Table 1, case 3). The maximal deviation is for mesophyll vacuoles, where the measured ABA concentrations is about 2.4 times higher than that calculated. A factor of 2.4 may not be too important, but we have experimental evidence that the prim, o f Valerianella mesophyll vacuoles is indeed about 0.1 units higher when determined from the ABA distribution across the tonoplast (Daeter and Hartung 1991) than when determined by 3~p-nuclear magnetic resonance measurements (data not shown). Both arguments are in favour of small but appreciable ABA-anion permeability at the mesophyll tonoplast. The necessary A B A - conductance at the tonoplast in order to over-

come the difference between 'case 3' and measured data is only about 0.1. 10 -9 m - s -1, i.e. 2% of the conductance of neutral ABA species (HABA) at the tonoplast o f mesophyll cells (calculation not shown). Compared with case 2, in case 3 there is a moderate ABA concentration increase in all apoplasmic spaces, in guard-cell compartments and in phloem cells. The ABA concentration increases in the mesophyll apoplasm by a factor o f 1.8 until the HABA gradient across the plasmalemma becomes high enough (in equilibrium) to feed ABA from the mesophyll apoplasm into the catabolising mesophyU cytosol at the same rate as ABA is imported into the leaf lamina via the xylem sap. The differences between the equilibria of cases 2 and 3 are too small in unstressed leaves to get an idea o f where the postulated net degradation o f ABA may occur. Both cases are close to measured data. Deviations can be simply explained by the fact that all measured compartmental concentrations and membrane conductances are not determined using the same Valerianella leaf, or even plant. Keeping this in mind, the deviations are indeed rather small. The measured data set for Valerianella is compatible with model calculations. Therefore, we conclude that in situ the compartmental distribution o f ABA is close to the expected equilibrium in unstressed leaf laminae o f Valerianella. In fact, the maximum compartmental relaxation-time coefficient (Table 2 below) is 79 h, i.e. within a few days after the appearance of a mature leaf at the corm, all leaf compartments should be close to ABA-flux equilibrium. We realise that it is possible to substitute the real leaf o f t/'alerianella by its model leaf in order to study in detail its complex properties, not only experimentally but also numerically (cf. below; cf. Slovik and Hartung 1991, 1992). In homology to case 2 it is possible to estimate the apparent K m and Vm,x constants of the cytosolic ABA-catabolising enzyme(s) if, alternatively, the mesophyll cytosol were the only catabolising compartment in the leaf lamina. In order to keep the measured cytosolic ABA concentration of 9.0 nM constant, there must be in the mesophylt cytosot a defined catabolic rate which just compensates the mean daily ABA influx of 1 2 0 f m o l . m - 2 . s - 1 into the unstressed leaf lamina (cf. above). The necessary catabolic rate in the mesophyll cytosol is 120 fmol. m -2 9s-1/(7.07 9 10 -6 m 3 - m -2) = = 17.0 n m o l . m -3 9 s -1 in the mesophyll cytosol of unstressed leaves. The denominator is the mesophyll cytosol volume per unit leaf area (cf. Slmdk et al. 1992). Again Vm~xis 3.8 h/2.0 h --- 1.9 times higher than the rate V = 17.0 nmol 9m - 3 . s-1 which is necessary to compensate the mean ABA influx into the unstressed lamina, i.e. Vmax=32.3 n m o l . m - 3 - s -1. The Kin=8.1 nM can be estimated by employing Eq. 1.1. [ABAm~]eq=9.0 nM is the measured concentration of ABA in the mesophyll cytosol. Finally, we must take into account that, at least in excised stressed leaves, there is also synthesis of ABA (cf. Pierce and Raschke 1981). However, no information exists on turnover rates of ABA in unstressed leaves. For simplicity's sake, we assume here that the total (not net) ABA-synthesis rate in the unstressed model-leaf cytosol is zero, i.e. the only ABA source for the whole lamina is

S. Slovik and W. Hartung: ABA distribution in leaves: II. Model analysis the xylem sap, but this may not be corrent even for unstressed leaves.

Equilibration of the model leaf with changed HABA conductance, keeping the measured cytosolic ABA concentration in the mesophyll constant (case 4. Table 1). We repeated the equilibration demonstrated in case 3 (constant ABA concentration in the mesophyll cytosol), but changed the epidermal membrane conductances for HABA. Unexpectedly, membrane conductances are capable of influencing flux equilibria. This is possible only if the plasmodesmata between mesophyll cells and epidermal cells are open, and if the epidermal apoplasm (or cytosol) has a p H value other than that of the mesophyll apoplasm (or cytosol). Then, there is a latent circulation of ABA within the lamina from the epidermal cell wall into the epidermal cytoplasm, and from there via the plasmodesmata to the chlorenchyma cytosol, where ABA permeates the plasmalemma. Finally, ABA diffuses from the mesophyll apoplasm back into the epidermal cell wall. This diffusion cycle is balanced. At flux equilibrium there is no accumulation of ABA in any compartment. The running cycle redistributes protons within the leaf because uncharged permeant H A B A is a weak acid. Weak acids acidify alkaline compartments while being trapped in them (Laisk et al. 1988). The energy demand of ABA-cycling is supplied by a small percentage of endergonic compartmental pH-stat mechanisms. For case 2, the latent cycling rate of 41.6 fmol 9 s - 1 . m - z leaf area corresponds to a 'turnover time' of only 11.0 h (ABA content = 1.64 nmol 9m - z leaf area; simulation data not shown). This is the same order of magnitude as turnover times for ABA metabolism (cf. Pierce and Raschke 1981 and above). As diffusive ABA flow is effective, there is only a very small accumulation o f ABA in epidermes by means of transpiration-driven waterflow within leaf laminae. Even if water were to evaporate exclusively at guard-cell surfaces in the model leaf (i.e. if there were a 20-fold mass-flow of ABA to guard cells; see Appendix o f Slovik et al. 1992), then there would be only a 1.14-fold ABA concentration in guard-cell walls. The epidermal ABA concentration would remain almost unchanged under these circumstances and the ABA-cycling rate between mesophyll and epidermis would increase only by 3 % (calculations not shown). We conclude that (i) it is impossible to markedly accumulate ABA 'automatically' in guard-cell walls if transpiration increases and (ii) that the influence on ABA redistribution of the localisation of water evaporation in the substomatal cavity in leaves is small. The latent cycling of ABA is the consequence o f an apoplasmic p H difference (pH,w = 6.0 in epidermal cell walls and pHmw= 6.5 in mesophyll cell walls; cf. Slovik et al. 1992), if plasmodesmata are open (Slovik and Hartung 1991). The magnitude and direction o f the ABA-cycling rate depends on the magnitude and sign of the p H difference between the two tissues. It becomes zero only if the pH difference becomes zero between the two tissues in both compartments separated by the membrane under consideration. In other words, if there are pH gradients along any membrane in one or both adjacent compartments, then there must be a cyclic

31

diffusion of ABA for physical reasons. Under these conditions, the different flux equilibria as expected solely, for example, at the plasma membranes o f both tissues are modified by the processes of ABA diffusion within the apoplast and via plasmodesmata. F r o m Table 1, case 4 we deduce that the compartmental ABA concentrations at equilibrium can change by a factor o f up to about 3 after changing the plasmalemma conductance by a factor of 0.1. Firstly, the apoplasmic ABA concentrations, fed via the ABA import o f the xylem sap, must rise as a consequence of decreasing membrane conductances. Secondly, the ABA concentrations in the phloem and in guard-cell compartments rise to equilibrium. The ABA concentration in epidermal cell compartments decreases because the influx of ABA from the xylem via the epidermal cell plasmalemma becomes smaller after decreasing its ABA permeability. In summary, it is possible to cause a redistribution of the p h y t o h o r m o n e ABA to a new flux equilibrium within all leaf compartments just by changing the membrane conductances in a defined tissue (e.g. epidermis; cf. Slovik and Hartung 1991). It is not known whether such membrane effects are involved in, for example, signal chains of stress perception. It is worth mentioning that the ABA concentration in the guard-cell apoplast increases by a factor of 2.7 (Table 1, case 4). Also the ABA concentration in the phloem sap increases. Similar (but slow) redistribution effects can be obtained by changing plasmodesmata conductances in the model leaf between different tissues (Slovik and Hartung 1991). It is practically impossible to study this experimentally.

Relaxation-time coefficients to equilibrium. The shapes of compartmental relaxation curves to a final ABA-flux equilibrium are usually very similar to exponential functions (or to a sum of them). Instead of visualising numerous similar graphs for all compartments, we summarised the final equilibria in Table 1, and some selected compartmental relaxation-time coefficients (dimension: min or h) in Table 2. Relaxation-time coefficients (x) vary considerably among different tissues and compartments. The fastest process is the relaxation of chloroplasts with = 1.5 min (Table 2, case 3 or 4). This can also be observed experimentally (Heilmann et al. 1980). The redistribution of ABA between vacuoles and the rest o f the leaf is very slow, with ~=3.3 days for mesophyll-cell vacuoles. This slow relaxation explains why Kaiser et al. (1985) found no detectable tonoplast permeability for ABA within a few hours for isolated vacuoles. The processes for net diffusion of ABA within the apoplasm among different tissues are not very fast (20 min) under the selected conditions. The phloem sap has a x value o f about 30-50 min. We conclude that the ABA concentration in the phloem sap, which is a function of the apoplasmic ABA concentration in the leaf lamina (source, cf. below), should be a rather fast and sensitive medium for transport of information to the stem and roots (generally sinks) of a plant (but see Fig. 1 for excised leaves). It should be stressed that the values o f 9 are very sensitive to p H changes (cf. Cowan et al. 1982). In Table 2 we summarised x values for 10-fold (relative to the final equilibrium) compartmental ABA concentrations. As

32

S. Slovik and W. Hartung: ABA distribution in leaves: II. Model analysis

Table 2. Relaxation-time coefficients (~) o f all c o m p a r t m e n t s for A B A and its conjugates (CON). We present three cases, corresponding to cases 2-4 in Table 1 : Case 2, the A B A concentration is kept c o n s t a n t in the mesophyll apoplasm. The final equilibrium is summarised in Table 1, case 2. Case 3, the A B A concentration is kept constant in the mesophyll cytosol (cf. Table 1, case 3). Case 4, again, the A B A concentration is kept constant in the mesophyll cytosol, but the m e m b r a n e conductances in epidermal cells are n o w only as small as in mesophyll cells (cf. Table 1, case 4). The values o f 9 = 1.443 - T1/2, determined graphically on the

basis of the half time T1/2, are only estimations because the relaxation curves (not shown) are not ideal exponential functions. The relaxation started with exactly a 10-fold equilibrium concentration (of. Table 1) in all compartments other than the compartment in which the ABA concentration is kept constant Leaf compartment a

Case 2

Case 3

Case 4

27.7 22.8 -

19.7 min 18.3 min 9.51 min

24.0 min 22.4 min 17.0 min

5.76 h 2.51 h 6.67 h

28.0 min 2.34 h

35.8 rain 2.60 h

5.79 h 6.77 h

1.52 min 2.46 h

1.46 rain 2.71 h

Cell wall G u a r d cells Epidermis Mesophyll

min min

Cytosol Mesophyll Epidermis G u a r d cells Chloroplasts Mesophyll G u a r d cells Vacuole Mesophytl A B A CON Epidermis A B A CON Guard cellsABA

78.6 39.3 22.85 31.6 16.2

h h h h h

38.8 38.8 20.0 20.0 12.6

h h h h h

39.1 39.1 49.5 49.5 12.7

h h h h h

Phloem

Sieve-tubesap

27.1 min

44.7 min

54.5 min

a A B A denotes abscisic acid, C O N denotes hydrolysable A B A conjugates

60 ................f,

,

,

I

,

,

, --

50

0

40

r

o 20 -e 10

Phloem sap ] 0

10

20

30

40

50

60

Time [h] Fig. 1. Relaxation kinetics o f the A B A concentration in the phloem sap after excision o f the leaf lamina at hour 1.0. The measured A B A concentration in the mesophyll a p o p l a s m is assumed to be constant

depends also on the relaxation flux direction and on the absolute initial ABA concentrations in some compartments, Table 2 just summarises approximate values for special cases, but not system constants, which characterise the ABA leaf model for any given case.

The effect of excising leaves from the shoot. The lamina compartment most sensitive to the abscision o f a leaf from the shoot is the phloem sap. As sieve-tube cells usually close their wounds with callose soon after leaf excision, the velocity of the phloem-sap flow becomes zero. Figure 1 shows the time course of ABA accumulation in the phloem sap after excising the lamina 'in air' at hour 1.0. The relaxation-time coefficient, ~, is about 7.5 h at a constant apoplasmic ABA concentration, and therefore higher than after changing compartmental ABA concentrations as presented above. The dynamic steady-state ABA concentration (cases 2-4 in Table 1) in the streaming phloem sap is roughly only about 5-10% of the static equilibrium concentration of ABA (50.8 nM), which can be expected in a non-streaming phloem sap after equilibration. The ABA concentration in the mobile phloem sap is somewhat smaller than the ABA concentration in the mesophyll apoplasm. It is linearly 0.7 times the ABA concentration in the mesophyll apoplasm. This factor is dominated by the phloem-cell plasmalemma conductance to undissociated H A B A and by the pH values of the mesophyll apoplasm and the phloem sap. Equilibration of the model leaf, keeping constant the very high apoplasmie ABA concentrations and low apoplasmic pH values which can be measured in certain species. In Table 3, apoplasmic ABA concentrations, measured in the exudate of petioles or stems (pressure chamber, enzyme-linked immunosorbent assay), are summarised for 17 plant species. The ABA concentrations range from 4 nM in unstressed Valerianella locusta leaves up to 2.8 laM in stressed Gossypium hirsutum leaves, i.e. they cover almost 3 orders o f magnitude for different species. After equilibration the total ABA content per leaf area and the ABA concentration in most lamina compartments are linear functions of the apoplasmic equilibrium concentration o f ABA. Additionally, apoplasmic pH values may be as low as 5.2 in leaves of certain species (Pfanz and Dietz 1987; some Commelinaceae). At this low apoplasmic pH the mole fraction o f undissociated ABA is 15 times higher than at pH 6.5 (unstressed Valerianella mesophyll apoplasm). Therefore, the highest ABA concentration which maximally can be expected in any stressed-leaf compartment, is about 15. 7 0 0 ~ 1 0 000 times higher than the highest compartmental concentration in Table 1 (cases 2-4), i.e. 104. 0.7 g M ~ 7 m M in the guard-cell cytoplasm, or for example, 36 gmol 9m -2 leaf area. This is a theoretical limit. Indeed, high guard-cell ABA concentrations (1-2 mM) were found for Commelina communis (Brinckmann et al. 1990), but - in contrast to Valerianella locusta - there may be an ABA-transport carrier in the guard-cell plasmalemma of some species because the mentioned theoretical limit is valid only if simultaneously there is an extremely high ABA con-

S. Slovik and W. Hartung: ABA distribution in leaves: II. Model analysis Table 3. The range of typical apoplasmic ABA concentrations in pressure-chamber exudates of leaves and stems of different unstressed and drought-stressed plant species. The pH of most of the saps ranged between 6.2 and 6.6; Zebrina and Prosopis, however, had a more acid apoplasmic fluid (pH 5.2-5.5). A~ values (MPa) are differences in the water potential between drought-stressed and unstressed tissues

33

Species

Organ

ABA concentration (nM) Unstressed Stressed

A~ (MPa)

Valerianelta locusta Helleborus viridis Zebrina pendula Asteriscus pygmaeus Phaseolus coccineus Xanthium strumarium a Harnmada scoparia Prunus dulcis Helianthus annuus b Anastatica hi&oehuntiea Zygophyllum dumosum Gossypiurn hirsutum r Quercus engelmannia Prosopis 91andulosa Quercus agrifolia Craterostigma plantagineum Chilopsis linearis

leaf leaf stem stem root leaf stem stem leaf stem stem leaf stem stem stem flower axis stem

4 10 50 80 20 100 100 t 00 100 10 100 150 250 500 530 290

0.2 1.2 1.4 0.6 -~ 1.0 1.1 b 1.3 1.1 1.0 n.d. 1.6 1.3

30 100 120 230 570 650 700 1500 2200 2480 1800 2830 5300 8000 9800

Cornish and Zeevaart (1985); stressed 1.2 MPa, unstressed 'turgid' b Gollan (1987); soil water content 'turgid' =0.2 g. g-i, 'stressed'=0.06 g 9g-1. Hartung et al. (I 988). All other data are our own measurements using the same techniques as described in Hartung et al. (1988)

centration and a low p H in the leaf lamina apoplasm. Usually, also the A B A content per unit leaf area is much smaller than that given above and it is extremely variable a m o n g different plant species. This has consequences concerning the postulated ' A B A receptor' at or in guard cells. We conclude that this ' A B A receptor' m a y be sensitive to ABA concentration changes O[ABA]/& rather than to absolute ABA concentrations. I f in fact the A B A receptor is a 'differential sensor', then it automatically adapts to changed ABA concentrations in leaf c o m p a r t ments. Circadian redistribution o f A B A among all l e a f cell comp a r t m e n t s in C 3 plants. In Fig. 2 we present the changes of all compartmental A B A concentrations in the leaf lamina of Valerianella, which can be expected employing

model calculations on the basis o f measured data, if there is a p H decrease in the stroma o f mesophyll and guardcell chloroplasts in the d a r k o f only 0.3 units (vice versa in the light; U. Heber, personal communication). We assume daily light and dark periods o f equal length. In the night, the mole fraction o f undissociated A B A increases in the chloroplast stroma (mesophyll: factor 2.0). Therefore, chloroplasts become a source of ABA. The redistribution into m o s t other leaf c o m p a r t m e n t s is rather fast (Fig. 2; Table 2; Cowan et al. 1982). Vacuoles are irrelevant as sinks for A B A because the A B A conductances o f tonoplast m e m b r a n e s are rather small. All other leaf c o m p a r t m e n t s are supplied with A B A almost uniformly (Fig. 2). As there are no p l a s m o d e s m a t a between adult guard cells and epidermal cells, the guard-cell apoplasm is assumed to be the most likely c o m p a r t m e n t in which stomatal water conductance will change in response to a changed A B A concentration in the local histological environment. In the dark, the apoplasmic ABA concentration in the guard-cell wall is expected to

be a b o u t 1.5 times higher relative to illuminated leaf laminae (Fig. 2). In fact, we have determined a 1.6-fold increase (6.3 ~ 10.5 riM) in the ABA concentration in the xylem sap o f darkened Helianthus annuus leaves after darkening (means o f 81 exudations with a pressure b o m b ; 95% significance using Student's t-test; data not shown). In Helianthus leaves there is a slight pH-shift in the xylem sap (pH 6.5 ~ p H 6.7; 99 % significance) after darkening, which supports trapping o f ABA in the apoplasm. Is a 1.6-fold increase o f the A B A concentration in the free space o f guard-cell walls o f C3 plants sufficient to induce stomatal closure in the dark? C o w a n et al. (1982) discussed the same question assuming a pHc decrease in the chloroplast stroma o f 0.5 units in the dark. They and we assume that an increase in the apoplasmic ABA concentration smaller than at least two-fold is not sufficient to induce stomatal closure in a short time range. Circadian redistribution o f A B A between all l e a f cell comp a r t m e n t s in C A M plants. In Fig. 2 we compare the

expected ABA redistribution in C A M leaves with that in C3 leaves. Neglecting model calculations, we would have expected there to be a dramatic increase of the ABA concentration in the guard-cell a p o p l a s m at night as a consequence o f (i) a pHc decrease (ApHo= - 0 . 3 ) in the chloroplast stroma, accompanied by (ii) a dramatic p H decrease (ApHm,,= - 2 . 5 ) in mesophyll vacuoles o f C A M plants. In contrast to our expectations, the circadian amplitude o f the A B A concentration in the guard-cell wall is rather small. After some redistribution o f ABA in the first night, the apoplasmic A B A concentration in guard-cell walls is only a b o u t 1.5 times higher in the dark relative to the ABA concentration in the light. Therefore, in both C A M plants and C3 plants the apoplasmic ABA concentration is governed by the ApHo in the chloroplast

34

S. Slovik and W. Hartung: ABA distribution in leaves: II. Model analysis

1S0

-

100

I

I

.

/

l

--

I

--

l

I

I

!

t#

I

|

I

I

I

I

t

I~ - "t

!

l

I

8 6

I

4

SO Ipidermal cutoplasm

0

I

,

I

~

guard

(ABA)

I

J

.

I

I

cell

,

I

-.

(ABA)

will

,

I

,

I

6 4

S epidermal

0

I

LJ

6

(ABA)

vacuole

,

I

,

#"

,

,' -

in

I

,

,

,

epidermal

I

I

.

I

I

=

I

2

(ABA)

apoplaem

,

I

C 0 .M

I

t

,

I

0

I

10 8 6 4 2 0

~--

4

2 C m

o c 0 o

omll

phloem

L

0

meaophwll apoplaam

(ABA)

Iip

I

,

I

,

I

I

I

I I

I

l I

,

I

,

I

,

I

I

(ABA)

i

100

400 l

I

I

I

$0

200 guard

mesophull

m

0

r

I

i

I

cutosol

i

cell

(ABA)

cwtoeol

(ABA)

I

0

i

I

.

I

,

,

I

I

,

I

300

6

m E

4

L

200 meeophull

i

m E

0

0 O

vacuole

(ABA)

i t

2

n

0 8

10

C

2

I

guard

,

I

,

I

l

r

-

200

I

,

I

.

I

cell

,

I

,

100

(ABA)

vacuole

I

,

I

F-

0 400

100

200

mmsophwll c h l o r o p l a s t

(ABA)

0 I

,

I

i

I

,

I

guard

I

cell

,

chloroplast

I

,

(ABA)

I

=

I

10

0 8

6 S

maeophwll

vacuole

4

(CON)

l l

0

I

0

epidermal u i c u o l e ,

I

1

,

I

2

=

I

I

3

0

Time

[days]

,

Fig. 2. Circadian redistribution of ABA between all leaf lamina compartments for C3 plants ( ) or CAM plants ( ............) as expected on the basis o f model calculations. Initial compartmental concentrations are equilibrium concentrations summarised in Table 1, case 2. Both model calculations (C3 and CAM) start at sunset. We employed the model parameters as defined by Slovik et al. (1992), but there is no import of ABA via the xylem ([ABA]x=0 nM). There is no net degradation somewhere in the model leaf lamina. For C3 plants, we present the redistribution of ABA after just changing the pHc in the chloroplast stroma of

I

1

,

I

2

2

(CON)

,

I

0

3

mesophyll cells and guard cells ( A p r i l = - - 0 . 3 in the dark and ApHc=0.3 in the light). At sunset and sunrise, this ApHc shifts to the corresponding final pH c value in the stroma within 1 h. For CAM plants, additionally (!) the pHmv in the mesophyll vacuole shifts linearly from 5.5 to 3.0 in the dark within 6 h after sunset. pHmv = 3.0 remains constant for the last 6 h in the dark. Starting at sunrise, the pHmv increases to 5.5 within 6 h and remains constant at 5.5 for the remaining 6 h in the light until sunset, where the PHmv again starts to become as low as 3.0 within 6 h and so on. We compare a mesophytic CAM leaf and a C 3 leaf

S. Slovik and W. Hartung: ABA distribution in leaves: II. Model analysis stroma, but the stomata of C 3 plants usually close in the dark while C A M plants open their stomata at night. Apparently, an increase of the A B A concentration in the guard-cell wall of only 1.5-fold may not be sufficient to induce stomatal closure in C A M plants. Rather, in a day/night regime the intercellular CO2 concentration, for example, seems to be more important in regulating the stomatal aperture than moderate changes of apoplasmic A B A concentrations (cf. Raschke 1975; Radin et al. 1988). Finally, we ask why the dramatic p H decrease in mesophyll vacuoles of C A M plants in the night has only small effects on the compartmentation of ABA. The mean daily concentration of A B A in guard-cell walls is only about 1.5 1.6 times higher in C A M plants relative to Ca plants (cf. Fig. 2). This is a consequence of the smaller mean daily pHmv = 4.25 in the mesophyll vacuole of the C A M leaf. The mole fraction of undissociated A B A at the mean pHmv = 4.25 (CAM) is 30 times higher than at pHmv= 6.33 (C3 ; pKa of A B A = 4.75). Therefore, the vacuolar relaxation time x is only 3.3 d (C3, Table 2) divided by 30 = 2.6 h (CAM). In fact, there is an 'oscillation' of the ABAmv concentration in mesophyll vacuoles, i.e. relaxation to equilibrium is fast enough (Fig. 2). The ratio of the mole fraction of undissociated H A B A at pHmv=3.0 relative to pHmv=5.5 is only 6.5. This factor is about the quotient of the daily maximum and minim u m ABAmv concentration is mesophyll vacuoles (Fig. 2). The small influence of ApH__+2.5 is the consequence of the absolute p K a o f ABA. In the evening, the maxim u m ABAmv concentration in mesophyll vacuoles o f the model C A M plants is only about 0.9 nM. Calculated on a leaf-area basis (concentrations in Fig. 2, volumes in Table 1 of Slovik et al. 1992), there is only 2.7% of the total A B A mass in the mesophyll vacuoles (maximum). Therefore, only (6.5-1)/6.5 times 2.7% = 2.3 % of the total leaf A B A is redistributed by a vacuolar ApHmv• Thus, after redistribution in the night, all other leaf compartments increase their A B A concentrations only by a factor of [ 0 . 0 2 3 + ( 1 - 0 . 0 2 7 ) ] / ( 1 - 0 . 0 2 7 ) = 1.02 relative to the illuminated C A M leaf. As a consequence of (i) the absolute pKa value of A B A and (ii) of the small vacuolar ABAmv content of the C A M leaf, the influence of the circadian vacuolar ApHmv is negligible. Finally we consider succulent C A M plants. The relaxation time increases by a factor of about 1.5 relative to nonsucculent leaves. This is the consequence of the increased vacuole volume/tonoplast surface ratio, which is estimated assuming 96% vacuole volume and 4% cytoplasm volume (K. Winter, Institut f/Jr Botanik, University of Wfirzburg, personal communication). N o w the vacuolar relaxation time is 1.5-2.6 h ( a b o v e ) = 3 . 9 h, i.e. only somewhat higher. All 'non-mesophyll-vacuole' compartments increase their A B A content by a factor of 1.05 in the dark (by vacuolar p H shifts). The accumulation factor of A B A in guard-cell walls of C A M plants in the night remains at about 1.5 (caused by stroma p H shifts), but the mean daily A B A content in guard-cell walls is now 2.5-3.0 times higher in succulent C A M leaves relative to non-succulent C3 leaves. These general conclusions do not much depend on the employed membrane permeabilities to A B A of the C3 plant Valerianella

35

locusta, because we mainly compare flux equilibria of C3 and C A M plants in a day/night regime. These equilibria depend on compartmental volumes and p H values, but not much on membrane conductances. The detailed kinetics in Fig. 1, however, would be modified if membranes o f C A M leaves possessed other A B A conductances such as those found in C3 leaves, but there is still a lack of ABA-conductance data for leaves of C A M plants.

We are grateful to Professor U. Heber (Lehrstuhl Botanik I, University of Wtirzburg, FRG) for stimulating discussions. This work has been performed within the research program of the Sonderforschungsbereich 251 (TP 3 and 4) of the University of Wfirzburg. It has been also supported by the Fonds der Chemischen Industrie.

References Baier, M., Hartung, W. (1991) Movement of abscisic acid across the plasmamembrane of phloem elements of Planta#o major. J. Plant Physiol. 137, 297-300 Behl, R., Hartung, W. (1986) Movement and compartmentation of abscisic acid in guard cells of Valerianella locusta; effects of osmotic stress, external H § concentration and fusicoccin. Planta 168, 360-368 Brinckmann, E., Hartung, W., Wartinger, M. (1990) Abscisic acid levels of individual leaf cells. Physiol. Plant. 80, 51-54 Cornish, K., Zeevaart, J.A. (1985) Movement of abscisic acid into the apoplast in response to water stress in Xanthium strumarium. Plant Physiol. 78, 623-626 Cowan, I.R., Raven, J.A., Hartung, W., Farquhar, G.D. (1982) A possible role for abscisic acid in coupling stomatal conductance and photosynthetic carbon metabolism in leaves. Aust. J. Plant Physiol. 9, 489-498 Daeter, W., Hartung, W. (1990) Compartmentation and transport of abscisic acid in mesophyll cells of intact leaves of Valerianella locusta. J. Plant Physiol. 136, 306-312 D6rffiing, K. (1983) Regulation der Stomaapertur - Ein Beispiel fiir die Bedeutung der Hormonsynthese, Metabolisierung, Kompartimentierung und Interaktion ffir einen hormonal gesteuerten Prozess. Hohenheimer Arb. 129, 102-120 D6rffling, K., B6ttger, M. (1968) Transport yon Abscisinsfiure in Explantaten, Blattstiel und Internodialsegmenten yon Coleus rheneltianus. Planta 80, 299-308 Drrffling, K., Tietz, D., Streich, J., Ludewig, M. (1980) Studies on the role of abscisic acid in stomatal movements. In: Plant growth substances 1979 (Proc. 10th Int. Conf. on Plant Growth Substances, Madison, Wis., USA), pp. 274-285, Skoog, F., ed. Springer, Berlin Heidelberg New York Gollan, T. (1987) Wechselbeziehungen zwischen Abscisinsfiure, Nfihrstoffhaushalt und pH im Xylemsaft, und ihre Bedeutung fiir die stomatfire Regulation bei Bodenaustrocknung, Ph.D. thesis, University of Bayreuth, FRG Heilmann, B., Hartung, W., Gimmler, H. (1980) Redistribution of abscisic acid between chloroplasts and cytoplasm of leaf cells and the permeability of the chloroplast envelope for abscisic acid. Z. Pflanzenphysiol. 97, 67-78 Kaiser, G., Weiler, E.W., Hartung, W. (1985) The intracellular distribution of abscisic acid in mesophyll cells - the role of the vacuole. J. Plant Physiol. 119, 237-245 Lahr, W., Raschke, K. (1988) Abscisic acid contents and concentrations in protoplasts from guard cells and mesophyll cells of Vicia faba L. Planta 173, 528-531 Laisk, A., Pfanz, H., Schramm, M.J., Heber, U. (1988) Sulfur-dioxide fluxes into different cellular compartments of leaves photosynthesizing in a polluted atmosphere. Planta 173, 230-240

36

S. Slovik and W. Hartung: ABA distribution in leaves: II. Model analysis

Loveys, B.R., Robinson, S.P. (1987) Abscisic acid synthesis and metabolism in barley leaves and protoplasts. Plant Sci. 49, 23 30 Pfanz, H., Dietz, K.J. (1987) A fluorescence method for the determination of the apoplastic proton concentration in intact leaf tissues. J. Plant Physiol. 129, 41-48 Pierce, M., Raschke, K. (1981 ) Synthesis and metabolism of abscisic acid in detached leaves of Phaseolus vulgaris L. after loss and recovery of turgor. Planta 153, 156-165 Radin, J.W., Hartung, W., Kimball, B.A., Mauney, J.R. (1988) Correlation of stomatal conductance with photosynthetic capacity of cotton only in a COz-enriched atmosphere: mediation by abscisic acid? Plant Physiol. 88, 1058-1062 Raschke, K. (1975) Simultaneous repuirement of carbon dioxide and abscisic acid for stomatal closing in Xanthium strumarium L. Planta 125, 243-259

Slovik, S., Hartung, W. (1991) Stress-induced redistribution kinetics of ABA in leaves: Model considerations. 14th Int. Conf. on Plant Growth Substances, Amsterdam, July 21-26, 1991, Lectures, in press Slovik, S., Baier, M., Hartung, W. (1992) Compartmental distribution and redistribution of abscisic acid in intact leaves. I. Mathematical formulation. Planta 187, 14-25 Slovik, S., Hartung, W. (1992) Compartmental distribution and redistribution of abscisic acid in intact leaves. III. Analysis of the stress-signal chain. Planta 187, 37-47 Weiler, E.W. (1980) Radioimmunoassay for the differential and direct analysis of free and conjugated abscisic acid in plant extracts. Planta 148, 262-272