PhotosynthesisResearch 47: 145-156, 1996. (~) 1996KluwerAcademicPublishers. Printedin the Netherlands. Regular paper

Phosphatase activities in spinach thylakoid membranes - effectors, regulation and location Inger Carlberg & Bertil Andersson Department of Biochemistry, Arrhenius Laboratories for Natural Sciences, Stockholm University, S-10691 Stockholm, Sweden. Received22 August 1995;acceptedin revisedform4 December1995 Key words: light-harvesting complex, Photosystem II, protein phosphorylation

Abstract

The dephosphorylation of seven phosphoproteins associated with Photosystem II or its chlorophyll a/b antenna in spinach thylakoids, was characterised. The rates were found to fall into two distinct groups. One, rapidly dephosphorylated, consisted of the two subunits (25 and 27 kD) of the major light harvesting complex of Photosystem II (LHC II) and a 12 kD polypeptide of unknown identity. A marked correlation between the dephosphorylation of these three phosphoproteins, strongly suggested that they were all dephosphorylated by the same enzyme. Within this group, the 25 kD subunit was consistently dephosphorylated most rapidly, probably reflecting its exclusive location in the peripheral pool of LHC II. The other group, only slowly dephosphorylated, included several PS II proteins such as the D1 and D2 reaction centre proteins, the chlorophyll-a binding protein CP43 and the 9 kD PS II-H phosphoprotein. No dephosphorylation was observed in either of the two groups in the absence of Mg 2+-ions. Dephosphorylation of the two LHC II subunits took place in both grana and stroma-exposed regions of the thylakoid membrane. However, deposphorylation in the latter region was significantly more rapid, indicating a preferential dephosphorylation of the peripheral (or 'mobile') LHC II. Dephosphorylation of LHC II was found to be markedly affected by the redox state of thiol-groups, which may suggest a possible regulation of LHC II dephosphorylation involving the ferredoxin-thioredoxin system. Abbreviations: CP 43-43 kD chlorophyll a- binding protein; D1 and D2-reaction centre proteins of PS II; LHC II-light-harvesting complex of PS II; LHC 11-25-25 kD subunit of LHC II; LHC 11-27-27 kD subunit of LHC II; NEM-N-ethylmaleimide; PP2C-protein phosphatase 2C, PS II-H-psb H gene product Introduction

Reversible protein phosphorylation of thylakoid membranes has been studied extensively over the last decade (Bennett 1991; Allen 1992). At least 10-15 thylakoid proteins, most of which are associated with Photosystem II or its light-harvesting antenna, can be phosphorylated in a light dependent manner in intact plants, green algae, isolated chloroplasts or thylakoid membranes (Bennett 1991; Allen 1992). The phosphorylation is controlled by the redox state of the plastoquinone pool (Allen et al. 1981; Horton et al. 1981) in a process that involves the cytochrome b/fcomplex (Gal

et al. 1987; Bennett et al. 1988; Wollmann and Lemaire 1988; Coughlan 1988). A membrane bound 64 kD protein has been implicated as the kinase (Coughlan and Hind 1986; Gal et al. 1990) although more recent studies suggest this protein to be polyphenol oxidase (Hind et al. 1995; Sokolenka et al. 1995). Phosphorylation of the major light-harvesting complex of PS II (LHC II), the dominant phosphoprotein in the thylakoid membrane, is thought to be involved in regulating the distribution of excitation energy between the two photosystems (Allen et al. 1981; Barber 1982; Bennett 1991; Allen 1992). Under conditions favouring PS II excitation, which induces a

146 reduction of plastoquinone, LHC II becomes phosphorylated. A subpopulation of the phosphorylated LHC II is detached from PS II and migrates from the stacked grana regions where the majority of PS II is located, to the stroma-exposed thylakoids (Kyle et al. 1983; Larsson et al. 1983, 1985, 1987; Jennings et al. 1986; Anderson and Andersson 1988) and possibly the grana margins (Anderson 1989; Albertsson et al. 1990) where it may interact with PS I. In addition, phosphorylation of LHC II has been suggested to be important in protection against photoinhibition (Horton and Lee 1985), as well as in the process of long-term light acclimation of the light-harvesting antenna of PS II (Glick et al. 1987; Andersson and Styring 1991). The physiological significance of phosphorylation of the other phosphoproteins in the thylakoid membrane, is not yet known, even though a few effects on PS II function have been reported (Bennett 1991; Allen 1992). In particular, phosphorylation of the reaction centre protein D1 has been shown to influence its degradation and turnover (Callahan et al. 1990; Aro et al. 1992, 1993). The level of phosphorylation of a certain protein, does not only depend on the activity ofkinases, but also on the activity of the phosphatases involved. Despite the fact that the importance of protein phosphatases in the regulation of cellular processes, as well as the complexity in their control, is now well established for mammalian cells (Ballou and Fisher 1986; Cohen 1989), their role in plant cell regulation has only been recently recognised (MacKintosh and Cohen 1989; MacKintosh et al. 1991). Since the original observation by Bennett (Bennett 1980), surprisingly few studies on the dephosphorylation of the integral phosphoproteins of the thylakoid membrane, have been published (Sun et al. 1989, 1993; Kieleczawa et al. 1992; Sun and Markwell 1992; Silverstein et al. 1993; Elich et al. 1993; Cheng et al. 1994; Ebbert and Godde 1994). Many of the questions concerning the number of phosphatases, their location and regulation, as well as their identity, remain largely unanswered. In the present study we have characterised the in situ dephosphorylation of phosphoproteins in spinach thylakoid membranes, with respect to heterogeneity, effector requirements, membrane location and possible regulation. Evidence for the operation of two distinct dephosphorylating activities are presented. The activity responsible for the dephosphorylation of LHC II is shown to be present in all thylakoid regions, to have a high specificity for the peripheral pool of LHC

II and to be markedly affected by the redox state of thiol-groups.

Materials and methods

Spinach thylakoids were prepared as described earlier (Andersson et al. 1976) and resuspended in 50 mM Tricine pH 7.8, 0.1 M sorbitol, 20 mM NaCI and 5 mM MgC12 (incubation medium). Thylakoid membranes were phosphorylated in the incubation medium, at a chlorophyll concentration of 0.4--0.5 mg/cm a in the presence of 0.4 mM ATP containing (7-32p)-ATP (0.02 mCi/mg chlorophyll), by illumination at 300 #mol photons m -2 s -1 at room temperature for 10 min. In order to study the dephosphorylation process, phosphorylated thylakoid membranes were quickly spun down at maximal speed in an Eppendorf centrifuge, resuspended in incubation medium and kept at room temperature in the dark. Dephosphorylation was stopped by the addition of icecold 10 mM EDTA and 10 mM NaF. The specific 32p-content of various proteins was analysed by SDS-PAGE and autoradiography. Samples were resolved according to (Laemmli 1970) using a 12-22% polyacrylamide gradient in the separation gel. Membranes were solubilised by the addition of sample buffer according to (Laemmli 1970), but to a final concentration of 4% SDS and 10% fl-mercaptoethanol and incubated at 37 °C for 20 min. This procedure was found to minimise the amount of aggregation at the top of the stacking gel. The gels were exposed to film and the autoradiograms were analysed by laser densitometry. Thylakoid subfractionation into grana and stroma exposed membranes was performed essentially as described in (Carlberg et al. 1992) with minor modifications. Two ml of thylakoid membranes at a chlorophyll concentration of 0.5 mg/cm 3 were mixed with an equal volume of 0.4% digitonin at room temperature. After 30 s of incubation the solubilisation was stopped by a 10-fold dilution with ice-cold 10 mM sodium phosphate, pH 7.8 containing 0.1 M sucrose, 5 mM MgC12, 5 mM NaC1 and 10 mM NaF (dilution buffer). Subsequent to an initial centrifugation at 1000 x g for 3 min, the grana membranes were isolated by a centrifugation at 10000 x g for 8 min. The grana pellet was further purified as described in (Cline 1988). After resuspension in 160 mm 3 of dilution buffer, 25 mm 3 of 10% Triton X-100 was added, and after 60 s of solubilisation, the membranes were diluted with 1

147 cm 3 of dilution buffer. Samples were then again spun at 1000 x g for 3 min and 10000 x g for 8 min. The grana fractions obtained by this procedure had a chl a/b ratio of approximately 1.9. The supernatant from the first 10000 x g centrifugation was spun at 40000 x g for 30 min and the stroma thylakoid membranes were finally collected by centrifugation at 140000 x g for 70 min, giving a stroma thylakoid fraction with a chl a/b ratio of 5-6. Chlorophyll a/b ratios were determined according to Arnon (1949).

120

A 1001 8O

20

Results

0 1.2

I

1

t

I

I

I

I

I

I

I

I

In situ dephosphorylation of LHC H and PS H polypeptides 0.6

Upon protein phosphorylation of spinach thylakoid membranes at least twelve polypeptides are phosphorylated. The most conspicuous of these are the two subunits of the light harvesting complex (LHC I1-25 and LHC II-27), the 9 kD PS II-H (Michel and Bennett 1987), a 12 kD polypeptide of unknown identity (Bennett et al. 1987; Peter and Thornber 1991; Bhalla and Bennett 1987), as well as at least two polypeptides in the 30 kD region, identified as the D1 and D2 proteins (Michel et al. 1988) and one or two in the 40 kD region, presumably CP43 (Michel et al. 1988). All of these (except the 12 kD phosphoprotein, for which information is not yet available) are phosphorylated on threonine residues in the stroma exposed N-terminal region of the protein (Allen 1992). Among the minor labelled bands are two of high molecular mass, around 62 and 54 kD, respectively, one around 23 kD and at least one or two in the range of 15-18 kD. Dephosphorylation of phosphorylated spinach thylakoid membranes was followed in the dark. Figure 1A depicts the rate of disappearance of radioactive phosphate from the seven major phosphoproteins (the D1and D2-proteins are not well resolved in the gel system used and analysed as one phosphoprotein band) and shows that these were dephosphorylated at markedly different rates. The rates fall into two distinct groups, with the LHC 11-25, LHC I1-27 and the uncharacterised 12 kD phosphoprotein forming one in which the proteins were all rapidly dephosphorylated. After 15 min incubation in the dark, less than 50% of the label of LHC I1-25 remained. LHC 11-27 and the 12 kD phosphoprotein both appear to be dephosphorylated at a rate slightly but significantly lower (75-80%) than that of LHC 11-25. After 30 min at least 75-90% of the label

0

0

15 Time (min)

I

30

Figure 1. Rate of in situ dephosphorylation of six phosphoproteins in the spinach thylakoid membrane. (A) O, 9 kDa PS II-H; [ ] , CP43; A, DI/D2; m , 12kDa; A, LHC II-27; 0 , LHC II-25. Phosphorylated, washed thylakoids were incubated in the dark. Samples were withdrawn at specified timepoints and the relative 32p content was determined. (B) The ratio between phospho-LHC II-25 and phospho-LHC II-27 during dephosphorylation. Each point gives the mean 4- S.D. of seven independent experiments.

had disappeared from all of the peptides in this group. The difference between the rates of dephosphorylation of LHC II-25 and LHC II-27 is verified in Figure 1B, which shows the decrease in ratio between the two phosphoproteins during dephosphorylation. The 9 kD psbH phosphoprotein together with the D 1 and D2 and the CP43 were only slowly dephosphorylated under the present conditions. Even after prolonged incubation times, such as 2.5 h, these polypeptides retain a phosphorylation level of at least 40-50% of the initial amount. The possible relatedness between the different dephosphorylation reactions shown in Figure 1A was analysed. To this end, the initial level of phosphorylation of a specific protein and its rate of dephosphorylation were compared to LHC II-27 and plotted as in Figure 2. A correlation with a straight line in this plot, would indicate that the phosphoprotein analysed and LHC II-27 are both substrates for the same enzyme, while a lack of correlation indicates that the two dephosphorylation events are independent. It is

148 25

the lines in Figure 2A and B indicate that both LHC II-25 and the 12 kD phosphoprotein are slightly better dephosphorylated than LHC 11-27 (slopes 1.22 and 1.25, respectively). No correlation is observed between the dephosphorylation of LHC I1-27 and that of the PS IIphosphoproteins, as judged by the scattered appearance of the plot for the D1/D2 (Figure 2C). CP43 and the 9 kD phosphoprotein gave the same type of picture. This means that the PS II phosphoproteins do not appear to be dephosphorylated in the same enzymological process as phospho-LHC I1-27.

A

0

05 K

I

r

[

0.5

I

1.5

0

,~

R=0.858

O.35

O.2-5 ~-

Effectors of LHC H dephosphorylation

0.2 0.15 0.1 0.05

1

R=0.967

o

I

I

I

I

I

0.05

0.l

0.15

0.2

0.25

0.3

O.2 C O 0.13

0

0

0

~,

o.~

~ 0.~ R=0.103

0

I

I

0.1

0.2

O I 03

I

I

I

0.4

05

0.6

O-----0.7

0.8

Figure 2. Relationship between the rate of dephosphorylation relative to LHC 11-27 and the initial level of phosphorylation relative to LHC II-27. The relative initial rate of dephosphorylation of(A), LHC II-25; (B), phospho 12kDa; (C), D1/D2 compared to that of LHC II-27 in a specific experiment was calculated and plotted against the relative initial level of phosphorylation compared to that of LHC II27 in the same specific experiment. A value of 1 on either axis means that the rate of dephosphorylation of a specific phosphoprotein or the amount of that specific phosphoprotein is equal to that of LHC II-27. A good correlation with a straight line indicates an interrelationship between the two processes evaluated. The estimated correlations (r = correlation coefficient) are shown in each case.

clearly seen (Figure 2A), that the higher the relative initial level of phospho-LHC 11-25 compared to phosphoLHC 11-27, the higher its relative rate of dephosphorylation compared to that of phospho-LHC II-27. A similar behaviour is seen for the 12 kD phosphoprotein (Figure 2B), strongly indicating that these three phosphoproteins are dephosphorylated by the same phosphatase. Analysed in this manner, the slopes of

Whereas dephosphorylation in pea thylakoids has been found to show high rates of activity and high stability, the same has not generally been true for spinach thylakoid membranes (Horton and Foyer 1983; Kieleczawa et al. 1992). Also, in this work it has been observed that while most preparations showed high levels of activity towards phospho-LHC II, preparations exhibiting no or very low in situ phosphatase activity were occasionally obtained. These previous and present observations indicate that the activity is relatively unstable, easily lost from the membrane or physiologically down-regulated. In the present study, we found that spinach thylakoid membranes having a high LHC II phosphatase activity (around 50% of the label disappearing in 15 min, equivalent to an activity of approximately 100 pmole Pi released min-l mg chl-1), could be extensively washed in either destacking (no Mg 1+) or high salt (up to 1 M NaC1) media without significant loss of activity, indicating that the enzyme is not easily released from the membrane. Preparations devoid of LHC II phosphatase activity, however, could not be reactivated by any of a variety of treatments tested, including various stimulating effectors such as metal ions, thiol-reducing agents (see below) or compounds that would reoxidise the plastoquinone pool. Consequently it seems that spinach thylakoid membranes with low phosphatase activity are irreversibly inactivated during the preparation or down-regulated in vivo due to the metabolic condition of the plant. In membranes with intermediate level of activity, on the other hand, dephosphorylation could often be stimulated by incubation with millimolar concentrations of DTI' as shown in Figure 3. The degree of activation was observed to be quite variable and within ten separate thylakoid preparations it was found to

149 11

2

12°f

211

lOOl 8O 8

i

6O o~

0.5 i

4O 0

A

B

C

0

Figure 3. DTT-activationof the dephosphorylatianof LHC II. Phos-

phorylated, washedthylakoidmembraneswereincubatedin the dark for 30 minutes in the absenceor presence of 5 mM DTT. (A), low; (B), intermediate; and (C), high stimulatory effect by DTT.Black, stimulation, by DTT;grey,relativerate of dephosphorylationbefore DTT-addition. The data given are the mean 4- S.D. of three (A) and (B) and four (C) independent experimentsrespectively.

range from no effect to a more than two-fold increase. In Figure 3 the data are grouped into A (low), B (intermediate) and C (high) according to the relative stimulatory effect of DTT. Included in Figure 3 are also the relative rates of dephosphorylation before addition of DTT, corresponding to each group. A slight negative relationship between the control rate and the stimulatory effect could be observed. These data show that the activity responsible for the dephosphorylation of LHC II is directly or indirectly influenced by the redox state of thiol groups for full activity. No marked influence by thiol-reducing agents on the activity for the slowly dephosphorylated group of phosphoproteins belonging to PS II was observed, while the dephosphorylation of the 12 kD phosphoprotein was stimulated in a way similar to LHC II dephosphorylation. In order to more unequivocally establish the possible role of thiol-groups in the dephosphorylation of LHC II, other sulphhydryl-directed reagents were tested for their effect on the phosphatase activity. Figure 4 shows the effect of the thiol-oxidising agent iodosobenzoate on the dephosphorylation of the two LHC II subunits. Iodosobenzoate strongly inhibited their dephosphorylation with an 150 of around 5 mM. The inhibition obtained by the lower concentrations of iodosobenzoate (1 and 2 mM) could be totally reversed by the addition of 10 mM DTT (Figure 4), indicating that the oxidation of a dithiiol to the corresponding disulphide had occurred. After incubation with higher concentrations of iodosobenzoate, only a partial reactivation was obtained, which could indicate that iodosobenzoate at elevated concentrations have multiple or unspecific

2O I

0 0

I

I

I

2 4 6 8 Iodosobensoate ( mM )

10

Figure 4. Inhibition of LHC II dephosphorylation by iodosobenzoate. Phosphorylatedthylakoid membraneswere incubated, with iodosobenzoateat the indicatedconcentrationsfor 15 min. The membranes were subsequentlywashed and incubated for 30 min in the dark in the absence, &, or presenceof 10 mM DTT,A.

effects on the dephosphorylation process. Apart from oxidising dithiols to disulphides, iodosobenzoate can also oxidise single thiolgroups to sulphenic, sulphinic or sulphonic acids. Attempts to verify the participation of a vicinal dithiol by inhibition with arsenite (5 mM) were unsuccessful. However, arsenite may be too hydrophilic to significantly affect a membrane protein. The LHC II phosphatase activity was also inhibited by the alkylating thiol-group reagent Nethylmaleimide (NEM), consistent with an involvement of thiol-groups in LHC II dephosphorylation. However, the inhibitory effect of NEM was less pronounced than that of iodosobenzoate, 8 mM gave only 12% inhibition, while the same concentration of iodosobenzoate inhibited 76%. This in turn indicates that an oxidation, presumably to a disulphide, is more deleterious for the phosphatase activity than the attachment of a bulky alkylgroup. No apparent differences in the response to the various thiol-group reagents were observed between LHC II-25 and LHC 11-27 dephosphorylation. The dephosphorylation of the 12 kD phosphoprotein was inhibited by both iodosobenzoate and NEM, but to a somewhat lesser extent than that of LHC II. Due to the very low control rates for the dephosphorylation of PS II core associated phosphoproteins in the spinach thylakoid membrane, inhibitory effects by NEM or iodosobenzoate could not be established. Dephosphorylation in the spinach thylakoid membrane was 95-100% inhibited by the addition of 10 mM

150 .o

1.2

g

LHCII-25

!

120

LHCII-27

100

8O

0.8

~'

-~

6o

~,,

4o

i

0.6

~d

0.4

0.8-

relative rate

0.4 20

"~

i

0.2 0

0

0

1

2

3

4

5

Figure5. Mg2+ requirementfordephospborylationofLHCII. Phosphorylatedthylakoidswereresuspendedin incubationmedium(30 mM TricinepH 7.9, 0.1 M sorbitol,5 mM MgCl2)withthe following alterations: 1. control;2. withoutMg2+; 3. withoutMg2+, 0.25 M NaCl added; 4. without Mg2+, 2 mM EDTAadded; 5. treated as sample3 but subsequentlywashedand resuspendedin incubation mediumwith 5 m M M g 2+ . The sampleswereincubatedfor 30 rain in the dark, whereafterthe 32p contentwas determined.LHC II-25, black and LHC II-27,grey.

NaF but not by 5 #M okadaic acid, the potent inhibitor of type 1 and 2A phosphatases (Biajolan and Takai 1988). Dephosphorylation was found to be inhibited 50-60% by millimolar concentrations of ATP, a property shared by many phosphatases (Ballou and Fisher 1986; Cohen 1989), but not previously reported for thylakoid membranes. Metal ion dependence A partial stimulation of dephosphorylation in pea thylakoids by Mg 2+ was reported previously by Bennett (1980) and has later been confirmed by others (Sun et al. 1989; Kieleczawa 1992; Sun and Markwell 1992). However, since Mg 2+ ions strongly influence the organisation of the thylakoid membrane (Barber 1982), and since metal ion dependence has been an important characteristic in the classification of phosphatases (Cohen 1989; Ballou, and Fisher 1986), we wanted to elucidate the Mg 2+ dependence of dephosphorylation in spinach thylakoids in more detail. Figure 5 shows that resuspension ofphosphorylated thylakoid membranes in a medium containing no added Mg 2+ reduced the rate of dephosphorylation of LHC 11-25 and LHC 11-27 by approximately 60%, compared to a medium containing 5 mM Mg 2+. If, however, 0.25 M NaC1 or 2 mM EDTA was added to thylakoids resuspended in the medium without Mg 2+, only residual

I

I

1

2

I

I

I

3

4

5

6

M g 2+ (raM)

Figure 6. The Mg 2+ dependence of the rate of dephosphorylation of phospho-LHC II. Phosphorylated thylakoids were first washed in incubation medium without MgCI2, containing 0.25 M NaCI and subsequently in medium without MgCi2 or NaCI. They were finally resuspended in the incubation medium containing MgCI2 at the indicated concentrations and incubated in the dark: for 30 min. Inserted is a Scatchard plot of the same data.

activity remained (Figure 5). The inactivating effect of NaCI remains even if the membranes are subsequently transferred to a medium without Na +. The activity could be restored by the readditon of Mg 2+, showing that the effect of Na + is reversible. Figure 6 shows the rate of dephosphorylation of the two LHC II polypeptides as a function of increasing Mg2+-concentration. It can be seen that extensive dephosphorylation was obtained already at very low concentrations of Mg 2+. When analysed in a Scatchard plot (Figure 6, insert), the dependence was linear and a half saturation value of 0.25 mM for the stimulating effect of Mg 2+ on LHC II dephosphorylatian could be calculated. Consistently, LHC 11-27 was found to respond slightly more rapidly than LHC 11-25. These observations clearly demonstrate an absolute dependence on Mg 2+ ions for the dephosphorylation of the LHC II. The loss of activity in the absence of Mg 2+ cannot be explained by secondary effects due to destacking of the thylakoid membrane, since the halfsaturation value of 0.25 mM is far below the concentrations required for induction of thylakoid stacking and a background level of 0.25 M Na + is known to maintain grana stacking (Barber 1982). Dephosporylation of the 12 kD phosphopeptide also required Mg 2+, but at somewhat higher concentrations than LHC II dephosphorylation. Dephosphorylation of the phosphoproteins associated with PS II on the other hand required much higher Mg 2+ concentrations (above 10 mM) for maximal activity and no

151 120

120

A 100'

100~

~80

80 60

40

40

20

20

0 0

10

20

30

40

Time (m/n)

50

O' 0

I0

20

30

40

50

Time (m/n)

Figure 7. Dephosphorylation of phospho-LHC II in the grana and the stroma-exposed thylakoid membranes. Phosphorylated washed thylakoid membranes were incubated in the dark. At specified timepoints, samples were withdrawn, solubilised with digitonin, whereafter the different subfractions representing grana and stroma-exposed membranes were collected by centrifugation and the 32p content of LHC II determined. (A), total phospho-LHC II: grana, O; stroma-exposed membranes, @. (B), LHC II-25: grana, A; stroma-exposed membranes, • and LHC 11-27: grana, © ; stroma exposed membranes, @. Each point gives the mean 4- S.D. of four independent experiments.

dephosphorylation could be observed below 1-2 mM Mg 2+. Consequently, in this case we can not totally exclude some influence of membrane stacking for maximal dephosphorylation rates. Other divalent cations were also tested for their capacity to restore dephosphorylation of LHC II in Mg2+-depleted thylakoid membranes. In the presence of a high concentration of NaC1 (0.25 M, to maintain stacking), addition of Mn 2+ and Ca 2+ at 2 mM, resulted in about half the rate of LHC II dephosphorylation as compared to the same concentration of Mg 2+. In contrast, addition of Zn 2+ or Co 2+ did not have any effect on the activity. If Ca 2+ or Mn 2+ were added simultaneously with Mg 2+, the rate of dephosphorylation was decreased, indicating that both ions compete with Mg 2+. In membranes that had not been subjected to Na+-pretreatment, and consequently retained some activity, addition of Ca 2+ or Mn 2+ was found to be inhibitory. The apparent total dependence upon Mg 2+ as well as the insensitivity to okadaic acid characterising the phosphatase activity involved in LHC II dephosphorylation suggest that it could be a PP2C phosphatase (Ballou and Fisher 1986; Cohen 1989), No physiological function has yet been ascribed to PP2C and it does not belong to the same gene family as the other phosphatases (Tamura et al. 1989). Recently, the first membrane bound PP2C was isolated and sequence comparisons show the PP2C family to be a highly diverged family (Klumpp et al. 1994). Eventhough PP2C

has been reported to not be present in chloroplasts (MacKintosh and Cohen 1989), this question may not yet have been settled.

Location of the phosphatase activity As mentioned above, extensive washings of the thylakoid membrane at high ionic strengths did not significantly decrease the phosphatase activity, indicating that the process is integral to the membrane. When intact chloroplasts were lysed in a minimal volume, and phosphorylation and dephosphorylation were followed in the complete, 'semi-intact' system, no changes in the rate of dephosphorylation were observed compared to control thylakoids, where the stromal proteins had been removed by centrifugation. This further corroborates the notion that the phosphatase activity that we observe is membrane bound. It also indicates that the very low activity observed for the PS II core associated phosphoproteins in the thylakoid membrane cannot simply be explained by the loss of a stromal phosphatase. LHC II is composed of two main populations (Kyle et al. 1983; Larsson et al. 1983, 1985, 1987; Jennings et al. 1986; Anderson and Andersson 1988). There is an inner pool containing the LHC 11-27 subunit exclusively, which is tightly bound to the PS II core, and there is an outer or peripheral pool that is enriched in the LHC I1-25 subunit (ratio 25/27 of 1/2 as compared to 1/4 for the whole antenna (Larsson et al. 1987)).

152 Upon phosphorylation, part of the mobile LHC 11-25 rich, outer LHC II antenna migrates from the grana to the unstacked stroma-exposed membranes (Kyle et al. 1983; Larsson et al. 1983, 1985, 1987; Jennings et al. 1986; Anderson and Andersson 1988), while the phosphorylated PS II core proteins, as well as the inner antenna, remain in the stacked grana regions. It was consequently of interest to determine whether dephosphorylation of the LHC II polypeptides occurred both in the grana and the stroma-exposed membranes. Phosphorylation and dephosphorylation carmolt easily be followed in isolated thylakoid subfractions due to the low levels of phosphorylation generally obtained after rupturing the membranes (Larsson, UK, unpublished observations). Moreover, isolation of stroma thylakoid membranes subsequent to phosphorylation is too time-consuming to allow the dephosphorylation in the purified subfractions to be monitored. To circumvent these experimental problems, we tried to address the question of phosphatase location at the level of the intact thylakoid membrane. Prephosphorylated membranes were incubated in the dark to allow dephosphorylation to proceed, at specified time points, samples were withdrawn and immediately subfractioned using digitonin. The grana and stroma thylakoid fractions were collected by differential centrifugation and the relative content of radioactive label in the various phosphoproteins was determined. As is clearly seen in Figure 7A, dephosphorylation of LHC II takes place in both grana and stromaexposed thylakoids. However, the radioactive label in LHC II disappeared significantly faster in the latter membrane region. Moreover, Figure 7B shows that the higher dephosphorylation rate observed in the stromaexposed membranes compared to the grana is mainly due to a more rapid loss of radiolabel from LHC II27. The dephosphorylation of LHC 11-25, on the other hand, is only marginally faster in the stroma-exposed membranes than in the grana. When comparing the dephosphorylation of LHC 11-25 and LHC 11-27, as described for intact thylakoids in Figure 2, we find that while the dephosphorylation of LHC II-25 is strongly favoured (slope = 1.7, r = 0.96) in the grana membranes, in the stroma exposed regions, the LHC II25 and LHC 11-27 are dephosphorylated equally well (slope = 0.97, r = 0.96). Inclusion of 10 mM NaF in the experiments inhibited the disappearance of radiolabelled phosphate from both membrane regions, ruling out the possibility that continued lateral migration, to any major part, was cre-

120 100

80 60 40 20 I

0 0

10

I

I

20 30 Time (mitt)

I

40

50

Figure 8. Remigration of dephosphorylatedLHC II from stromaexposed to grana thylakoids. Phosphorylated,washed thylakoids were incubated in the dark. At specifiedtimepoints, samples were withdrawn and subjected to digitonin fractionation. The stroma exposed membraneswere collectedand the chl a/b- value and the 32p contentof LHC II was determined. 100%representsthe maximal changein these parameters(dephosphorylation,93% decrease; change in chl a/b, 1.4 units) within the time of the experiment. Dephosphorylation• ; changein chi a/b ratio, O.

ating an apparent disappearance from the grana membranes. The PS II polypeptides (CP43, D 1/D2 and 9 kD) do not migrate in response to phosphorylation under normal physiological conditions (Larsson et al. 1987). Accordingly, only residual amounts of the D1, D2 and 9 kD phosphoproteins were found in the stromaexposed thylakoids, and this location did not affect their rate of dephosphorylation. Dephosphorylation of LHC H in relation to the remigration to the grana

In order to resume its association with PS II, phosphoLHC II in the stromaexposed membranes has to become dephosphorylated and subsequently migrate back to the appressed grana regions. This remigration process can be experimentally followed by subfractionation of thylakoid membranes and observed as an increase in the chl a/b ratio in the isolated stromaexposed membranes. In Figure 8 a comparison of the rate of this increase in chl alb ratio with the rate of dephosphorylation is depicted, showing that the two rates are clearly different. From a set of four independent experiments, initial rates (expressed as percent of total change min -1) were calculated to 2.6 + 0.8 for the change in chl alb ratio and 4.6 4- 1.3 for the dephophorylation, revealing a two-fold difference in the time-scale of the two processes.

153 Discussion In the present investigation the in vitro protein phosphatase activity in spinach thylakoid membranes was followed. The dephosphorylation of the seven main phosphoproteins was found to fall into two distinct groups, one rapidly dephosphorylated (the two LHC II subunits and the 12 kD phosphoprotein) and one only very slowly dephosphorylated (PS II core subunits), indicating the operation of two separate processes. The marked correlation between the dephosphorylation of LHC 11-25, LHC 11-27 and the 12 kD phosphoprotein clearly suggest that these are most likely dephosphorylated by the same phosphatase. In addition, our data also indicate that this phosphatase activity is not responsible for the dephosphorylation of the PS II phosphoproteins. Recently, it was reported that the dephosphorylation in the pea thylakoid membrane is also kinetically heterogenous (Silverstein et al. 1993). It was subsequently suggested that it is catalysed by one single phosphatase as judged from the inhibition by a synthetic 15 aminoacid LHC II phosphopeptide (Cheng et al. 1994), even if the authors did not exclude the participation of multiple enzymes. If the phosphorylation of the PS II proteins serves a reversible regulatory function in vivo, efficient means for dephosphorylation should also exist. A different situation, without need for rapid dephosphorylation, could be anticipated if the purpose of the phosphorylation is to target a protein for, i.e. degradation, as was originally suggested for the D1 protein (Callahan et al. 1990) but which has been questioned recently (Aro et al. 1992). In the isolated spinach thylakoid membrane, the rate of dephosphorylation of the PS II proteins appears too low to be considered physiologically significant, in contrast to that of LHC II dephosphorylation. The low rate of PS II dephosphorylation that we do observe, most likely reflects residual, unspecific or maybe down regulated activity. In the latter case, one can of course not totally rule out a possible participation of the same protein as the one responsible for the LHC II dephosphorylation (cf. Cheng et al. 1994). It would however still imply that the two processes are in some way mechanistically different. In the present study we were not able to increase the low rate of dephosphorylation of these phosphoproteins by any of a large variety of treatments including, light, different effectors of electron transport, metal ions and reductants. It has recently been reported that dephosphorylation of PS II core proteins in Spirodela is light-dependent in vivo (Elich et al.

1993). No light-stimulation was found in isolated thylakoid membranes, indicating that the integrity of the chloroplast, stromal factors or both are of importance for this stimulation. This would also be in agreement with the observations (Silverstein et al. 1993) that the phosphatase in isolated pea thylakoids is redox independent. The PS II core associated phosphoproteins (D l/D2, CP43 and 9 kD) are almost exclusively situated in the grana stacks and it could be argued that they are less accessible to a phosphatase than phospho-LHC II which is distributed also to the stroma exposed membranes (Kyle et al. 1883; Larsson et al. 1983, 1985, 1987; Jennings et al. 1986; Anderson and Andersson 1988) and possibly to the margins (Anderson 1989; Albertsson 1990). Speaking against this, are the observations that destacking of the membrane or redistribution of D1/D2 or phospho-9kD to the stroma exposed membranes do not increase their rate of dephosphorylation. Upon phosphorylation, part of the LHC II25 enriched peripheral pool, migrates to the nonappressed stroma-exposed membrane regions. The inner LHC II-pool stays attached to PS II which remains in the grana regions, except under special conditions such as heat or triggering for D 1-protein degradation (Aro et al. 1993). Since, except for some 5-10% of LHC II-27, all radioactive phosphate in LHC II disappears from the thylakoid membrane and only about half of all of phospho-LHC II migrates away from the grana appressions, one has to assume the existence of phosphatase activity in both appressed and stroma exposed thylakoid regions. This assumption is clearly supported by the present subfractionation experiments, which in addition show that the dephosphorylation is more efficient in the stroma-exposed membranes than in the grana. This can predominantly be explained by a markedly higher rate of dephosphorylation of LHC 11-27 in this membrane region. The explanation for the relatively low rate for LHC 11-27 in the grana membranes, is most likely to be found in the heterogeneous organisation of LHC II. The inner pool of LHC II contains only LHC 11-27 and is tightly associated with the PS II complex while the peripheral or mobile pool contains both LHC 11-27 and LHC 11-25. Presumably, the inner LHC 11-27 is less accessible for dephosphorylation than the LHC I1-27 in the peripheral pool. This latter notion is also supported by increased relative rates of LHC 11-27 dephosphorylation under destacking conditions. The similar rates of dephosphorylation of LHC 11-25 and LHC 11-27 in the stroma-exposed

154 regions indicate that the two phosphoproteins as such are equally good substrates for the phosphatase. The grana margins have recently been suggested to be a distinct membrane domain (Anderson 1989; Albertsson et al. 1990) and to possibly be enriched in the kinase involved in phosphorylation (Gal et al. 1990; Yu et al. 1992). It should be emphasised that our studies do not exclude a possible involvement of the margin derived membranes in the dephosphorylation observed for the grana thylakoid fraction. The observation that the rate of dephosphorylation was always faster than the rate of remigration of LHC II to the appressed thylakoid regions, strongly indicate that the latter process is the rate limiting step in the reassociation of the peripheral pool of LHC II with PS II. This observation as well as the relatively long times involved, indicate a long distance lateral diffusion of LHC II and not only a structural reorganisation within the thylakoid margins. The kinase(s) responsible for the light-dependent protein phosphorylation in the thylakoid membrane, is regulated via the redox-state of the plastoquinone pool (Allen et al. 1981; Horton et al. 1981), and more specifically via a plastoquinol bound to the reduced cytochrome blfcomplex in the vicinity of the Rieske Fe-S protein (Vener et al. 1995). In contrast, cytochrome blfis not required for the phosphorylation of the 9 kD polypeptide and possibly also the D1 and D2 proteins (Gal et al. 1987; Bennett et al. 1988; Wollmann and Lemaire 1988; Coughlan 1988). In light of this control of the kinase activity, as well as the accumulating knowledge concerning strict regulation of phosphatase activities in general (Ballou and Fisher 1986; Cohen 1989), it seems likely that thylakoid protein dephosphorylation, is also under regulation (Allen 1992). The recent findings (Elich et al. 1993) show that this indeed is the case for the dephosphorylation of the D1 phosphoprotein, and our present study supports the possibility of regulatory control also in the case of LHC II dephosphorylation. The observation that the LHC II phosphatase activity is dependent upon the thiol redox-state as judged by the effect of several reagents, makes it tempting to suggest a possible involvement of thiol groups in regulation. This does not necessarily contradict the observations indicating that the phosphatase activity is redox independent (Silverstein et al. 1993), since there could be kinetic restrictions in the redox reactions of the thiols involved. A solubilised and partially purified phosphatase from spinach thylakoid membranes has been found to be totally dependent on the presence

of DTT (A. Vener and B. Andersson, unpublished), indicating that the phosphatase itself may be sensitive to thiol-oxidation. The phosphatase isolated from pea thylakoids was also reported to be slightly sensitive to alkylating agents (Kieleczawa et al. 1992). However, it should be stressed that the dephosphorylation could be dependent upon the thiol redox-state of thylakoid components, other than the phosphatase itself, such as the phosphoprotein substrate or a regulatory component. LHC II does contain one single Cys residue, which is deeply buried in the lipid bilayer (Btirgi et al. 1987) and consequently less likely to be the one influenced in the present studies. Possible regulatory factors are yet to be identified. Interestingly, the protein phosphorylation in the thylakoid membrane has been reported to be sensitive to thiol alkylating reagents (Millner et al. 1982) and we have preliminary evidence suggesting that the activity of the protein kinase is also strongly affected by the thiol redox-state (I. Carlberg and B. Andersson, unpublished). The thiol-status has for a long time been known to influence the activity of the thylakoid coupling factor (CFI) (Petrack et al. 1965; Mills et al. 1980) and has more recently been shown to affect the non-photochemical quenching in intact chloroplasts (Demmig-Adams et al. 1990; Neubauer 1993). The possibility of connections between these events and the dephosphorylation of LHC II cannot be excluded. The potential involvement of thiol groups in the regulation of reversible thylakoid protein phophorylation would most likely imply the participation of the ferredoxin-thioredoxin system (Cseke and Buchanan 1986; Buchanan 1991) at some stage in the process. If so, this could provide a link between protein phosphorylation in the thylakoid membrane and the Calvin cycle reactions in the chloroplast stroma (Foyer et al. 1990).

Acknowledgements This work was supported by grants from The Swedish Agricultural and Forestry Research Council, The Swedish Natural Science Research Council, and The G6ran Gustafsson Foundation for Natural Sciences and Medicine.

155

References Albertsson P-.~, Andreasson E and Svensson P (1990) The domain organization of the plant thylakoid membrane. FEBS Lett 273: 36-46 Allen JF (1992) Protein phosphorylation in regulation of photosynthesis. Biochim Biophys Acta 1098:275-335 Allen JF, Bennett J, Steinback KE and Arntzen CJ (1981) Chloroplast protein phosphorylation couples plastoquinone redox state to distribution of excitation energy between Photosystems. Nature 291:25-29 Anderson JM (1989) The grana margins of plant thylakoid membranes. Physiol Plant 76:243-248 Anderson JM and Andersson B (1988) The dynamic photosynthetic membrane and regulation of solar energy conversion. Trends Biochem Sci 13:351-355 Andersson B and Styring S (1991 ) Photosystem II: Molecular organization, function and acclimation. In: Lee CP (ed) Current Topics in Bioenergetics, Vol 16, pp 2-81. Academic Press, New York Andersson B,/~kerlund H-E and Albertsson P-/~ (I976) Separation of subchloroplast membrane particles by counter-current distribution. Biochim Biophys Acta 423:122-132 Arnon DI (1949) Copper enzymes in isolated chloroplasts. Polyphenoloxidase in Beta vulgaris. Plant Physiol 24:1-15 Aro E-M, Kettunen R and Tyystj~irviE (1992) ATP and light regulate D1 protein modification and degradation- role of DI* in photoinhibition. FEBS Lett 297:29-33 Aro E-M, Virgin I and Andersson B (1993) Photoinhbition of Photosysytem II. Inactivation, protein damage and turnover. Biochim Biophys Acta 1143:113-134 Ballou LM and Fischer EH (1986) Phosphoprotein phosphatases. In: Boyer P and Krebs EG (eds) The Enzymes, Vol 17, pp 311-361. Academic Press, Orlando Barber J (1982) Influence of surface charges on thylakoid structure and function. Annu Rev Plant Physiol 33:261-295 Bennett J (1983) Chloroplast pbosphoproteins. Evidence for a thylakoid-bound phosphoprotein phosphatase. Eur J Biochem 104:85-89 Bennett J (1991) Protein phosphorylation in green plant chloroplasts. Annu Rev Plant Physiol Plant Mol Biol 42:281-311 Bennett J, Shaw EK and Bakr S (1987) Phosphorylation ofthylakoid proteins and synthetic peptide analogs. FEBS Lett 210:22-26 Bennett J, Shaw EK and Michel H (1988) Cytochrome btlfcomplex is required for phosphorylation of light harvesting chlorophyll a/b complex II in chloroplast photosynthetic membranes. Eur J Biochem 171:95-100 Bhalla P and Bennett J (1987) Chloroplast phosphoproteins: phosphorylation of a 12-kDa stromal protein by the redox-controlled kinase of thylakoid membranes. Arch Biochem Biophys 252: 97-104 Bialojan C and Takai A (1988) Inhibitory effect of a marine-sponge toxin okadaic acid on protein phosphatases. Biochem J 256: 283290 Buchanan BB (1991) Regulation of CO2 assimilation in oxygenic photosynthesis: The ferredoxin / thioredoxin system. Arch Biochem Biophys 288:1-9 Btirgi R, Suter F and Zuber H (1987) Arrangement of the lightharvesting chlorophyll a/b protein complex in the thylakoid membrane. Biochim Biophys Acta 890:346-351 Callahan FE, Ghirardi ML, Sopory SK, Mehta MA, Edelman M and Mattoo AK (1990) A novel metabolic form of the 32 kDa-D1 protein in the grana-localized reaction center of Photosystem II. J Biol Chem 265:15357-15360

Carlberg I, Bingsmark S, Vennigerholz E Larsson UK and Andersson B (1992) Low temperature effects on thylakoid protein phosphorylation and membrane dynamics. Biochlm Biophys Acta 1099: 111-117 Cheng L, Spangfort MD and Allen JF (1994) Substrate specificity and kinetics of thylakoid phosphoprotein phosphatase reactions. Biochim Biophys Acta 1188:151-157 Cline K (1988) Light-harvesting chlorophyll a/b protein. Plant Physiol 86:1120-1126 Cohen P (1989) The structure and regulation of protein phosphatases. Annu Rev Biochem 58:453-508 Coughlan SJ (1988) Chloroplast thylakoid protein phosphorylation is influenced by mutations in the cytochrome bfcomplex. Biochim Biophys Acta 933:413-422 Coughlan SJ and Hind G (1986) Purification and characterization of a membrane-bound protein kinase from spinach thylakoids. J Biol Chem 261: 11378-11385 Cseke C and Buchanan BB (1986) Regulation of the formation and utilization ofphotosynthate in leaves. Biochim Biophys Acta 853: 43-63 Demmig-Adams B, Adams III WW, Heber U, Neimanis S, Winter K, Kriiger A, Czygan F-C, Bilger W and Bjtrkman O (1990) Inhibition of zeaxantin formation and of rapid changes in radiationless energy dissipation by dithiothreitol in spinach leaves and chloroplasts. Plant Pbysiol 92:293-301 Ebbert V and Godde D (1994) Regulation ofthylakoid protein phosphorylation in intact chloroplasts by the activity of kinases and phosphatases. Biochim Biophys Acta 1187:335-346 Elich TE, Edelman M and Mattoo AK (1993) Dephosphorylation of Photosystem II core proteins is light-regulated in vivo. EMBO J 12:4857-4862 Foyer C, Furbank R, Harbinson J and Horton P (1990) The mechanisms contributing to photosynthetic control of electron transport by carbon assimilation in leaves. Photosynth Res 25:83-100 Gal A, Shahak Y, Schuster G and Ohad I (1987) Specific loss of LHC II phosphorylation in the Lemna mutant 1073 lacking the cytochrome b61fcomplex. FEBS Lett 221:205-210 Gal A, Hauska G, Herrmann R and Ohad I (1990) Interaction between light harvesting chlorophyll-a/b protein (LHC II) kinase and cytochrome b6/f. J Biol Chem 265:19742-19749 Glick RE, Larsson UK and Melis A (1987) Differential phosphorylation of thylakoid membrane polypeptides during chloroplast adaptation to light quality. In: Biggins J (ed) Progress in Photosynthesis Reasearch, Vol II, pp 253-256. Martinus Nijhoff Publishers, Dordrecht Hind G, Marshak DR and Coughlan SJ (1995) Spinach thylakoid polyphenol oxidase: Cloning, characterization and relation to a putative protein kinase. Biochemistry 34:8157-8164 Horton P and Foyer C (1983) Relationship between protein phosphorylation and electron transport in the reconstituted chloroplast system. Biochem J 210:517-521 Horton P and Lee P (1985) Phosphorylation of chloroplast membrane proteins partially protects against photoinhibition. Planta 165: 37-42 Horton P, Allen J F, Black MT and Bennett J (1981) Regulation of phosphorylation of chloroplast membrane polypeptides by the redox state of plastoquinone. FEBS Lett 125:193-196 Jennings RC, Islam K and Zucchelli G (1986) Spinach-tylakoid phosphorylation: Studies on the kinetics of changes in Photosystem antenna size, spill-over and phosphorylation of lightharvesting chlorophyll a/b protein. Biochim Biophys Acta 850: 483--489

156 Kieleczawa J, Coughlan SJ and Hind G (1992) Isolation and characterization of an allcaline phosphatase from pea thylakoids. Plant Physioi 99:1029-1036 Klumpp S, Hanke C, Donella-Deana A, Beyer A, Kellner R, Pinna LA and Schultz JE (1994) A membrane-bound protein phosphatase type 2C from Paramecium tetraurelia. J Biol Chem 269: 32774-32780 Kyle DJ, Staehelin LA, and Arntzen CJ (1983) Lateral mobility of the light- harvesting complex in chloroplast membranes controls excitation energy distribution in higher plants. Arch Biochem Biophys 222:527-541 Laemmli UK (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227:680--685 Larsson UK and Andersson B (1985) Different degrees of phosphorylation and lateral mobility of two polypeptides belonging to the light-harvesting complex of Photosystem II. Biochim Biophys Acta 809:396--402 Larsson UK, Jergil B and Andersson B (1983) Changes in the lateral distribution of the light harvesting chlorophyll-a/b-protein complex induced by its phosphorylation. Eur J Biochem 136: 25-27 Larsson UK, Sundby C and Andersson B (1987) Characterization of two different subpopulations of LHC II: Polypeptide composition, phosphorylation pattern and association with Photosystem II. Biochim Biophys. Acta 894:59--68 MacKintosh C and Cohen P (1989) Identification of high levels of type 1 and type 2A protein phosphatases in higher plants. Biochem J 262:335-339 MacKintosh C, Coggins J and Cohen P (199I) Plant protein phosphatases. Biochem J 273:733-738 Michel HP and Bennett J (1987) Identification of the phosphorylation site of an 8.3 kDa protein from Photosystem II of spinach. FEBS Lett 212:103-108 Michel HP, Hunt DE Shabanowitz J and Bennett J (1988) Tandem mass spectrometry reveals that three Photosystem II proteins of spinach chloroplasts contain N-acetyl-O-phosphothreonine at their NH2 termini. J Biol Chem. 263:1123-1130 Millner PA, Widger WR, Abbott MS, Cramer WA and DiUey RA(1982) The effect of adenine nucleotides on the inhibition of the thylakoid protein kinase by sulfhydryl-directed reagents. J Biol Chem 257:1736-1742

Mills JD, Mitchell P and Schiirmann P (1980) Modulation of the coupling factor ATPase activity in intact chloroplasts. The role of the ferredoxin system. FEBS Lett 112:173-177 Neubauer C (I993) Multiple effects of dithiothreitol on nonphotochemical fluorescence quenching in intact chloroplasts. Plant Physiol 103:575-583 Peter GF and Thomber JP (1991) Biochemical composition and organization of higher plant Photosystem II light-harvesting pigmentproteins. J Biol Chem 266:16745-16754 Petrack B, Craston A, Sheppy F and Farron F (1965) Studies on the hydrolysis of adenosine triphosphate by spinach chloroplasts. J Biol Chem 240:906-914 Silverstein T, Cheng L and Allen JF (1993) Chloroplast thylakoid protein phosphatase reactions are redox-independent and kinetically heterogeneous. FEBS Lett 334:101-105 Sokolenka A, Fulgosi H, Gal A, Ohad I, Altschmied L and Herrmann RG (1995) The 64 kDa polypeptide of spinach may not be the LHC II kinase but a lumen-located polyphenol oxidase. FEBS Lett 371:176-180 Sun G and Markwell J (1992) Lack of types 1 and 2A protein serine(P)/threonine(P) phosphatase activities in chloroplasts. Plant Physiol 100:620-624 Sun G, Bailey D, Jones MW and Markwell J (1989) Chloroplast thylakoid protein phosphatase is a membrane surface-associated activity. Plant Physiol 89:238-243 Sun G, Sarath G and Markwell J (1993) Phosphopeptides as substrates for thylakoid protein phosphatase activity. Arch Biochem Biophys 304:490-495 Tamura S, Lynch KR, Lamer J, Fox J, Yasui A, Kikuchi K, Suzuki Y and Tsuiki S (1989) Molecular cloning of rat type 2C (IA) protein phosphatase mRNA. Proc Natl Acad Sci USA 86:1796--1800 Vener AV, van Kan PJM, Gal A, Andersson B and Ohad 1 (1995) Activation/deactivation cycle of redox controlled thylalkoid protein phosphorylation. J Biol Chem 270:25225-25232 Wollmann F-A and Lemaire C (1988) Studies on kinase controlled state transitions in Photosystem II and bffmutants from Chlamydomonas reinhardtii which lack quinone binding proteins. Biochim Biophys Acta 933:85-94 Yu S-G, Stefansson H and Albertsson P-,~ (1992) Localization of 64 kDa LHC II- kinase in the thylakoid membrane from spinach. In: Murata N (ed) Reasearch in Photosynthesis, Vol I, pp 283-286. Kluwer Academic Publishers, Dordrecht