Occupation of the QB-binding pocket by a photosystem II inhibitor triggers dark cleavage of the D1 protein subjected to brief preillumination.

The D1 protein of the photosystem (PS) II reaction center turns over very rapidly in a light-dependent manner initiated by its selective and specific cleavage. The cleavage of D1 was studied by using a PS II inhibitor, N-octyl-3-nitro-2,4,6-trihydroxybenzamide (PNO8), as a molecular probe. The following results were obtained. (i) D1 was selectively cleaved into 23-kDa N-terminal and 9-kDa C-terminal fragments in complete darkness by PNO8 at a single site in a D-E loop connecting membrane-spanning helices D and E. (ii) The cleavage was markedly enhanced when PS II was illuminated for a brief period before the addition of PNO8 in darkness. (iii) The effect of preillumination was slowly lost during incubation in the dark, with a decay half-time of ∼1 h at 25°C. (iv) The light intensity of preillumination required for the cleavage was much lower than that required for O2 evolution. (v) The light-triggered cleavage of D1 was observed in thylakoids, PS II membranes, and PS II core particles, but not in purified PS II reaction centers. More than 60% of D1 was cleaved into the two fragments with no other by-products. (vi) The cleavage reaction revealed a marked pH dependence that was considerably different from that for inhibition of PS II activity. The results are interpreted as indicating that the binding of PNO8 to the QB-binding pocket triggers proteolytic cleavage of D1 that has been previously modified during illumination.

Photosystem (PS) 1 II is composed of at least 20 proteins, including the reaction center D1 and D2 proteins, embedded in thylakoid membranes (1). D1 is involved in the binding of all redox-active components but Q A quinone required for primary charge separation and electron transport from water to Q B quinone (2,3). When photosynthetic oxygen-evolving organisms are exposed to excess light with an intensity that oversaturates photosynthesis, electron transport is inhibited at the acceptor side of PS II (4 -7). In damaged PS II, D1 is selectively degraded, while other PS II proteins are generally preserved in an intact state (5)(6)(7)(8). These processes are termed photoinhibition. Under high light conditions, the translation rate of D1 mRNA is 50 -100 times greater than those of other PS II proteins, and defective PS II is repaired by inserting newly synthesized D1 into photoinhibited PS II (9 -11). PS II activity can remain uninhibited since the rates of both de novo synthesis of D1 and its reassembly are coordinated with the D1 degradation rate (12). Therefore, the rate of the light-dependent turnover of D1 may be determined by its degradation rate.
D1 is considered to be irreversibly modified during lightinduced inactivation of PS II to become a target of the following pathway toward proteolytic breakdown (5)(6)(7)(8). Active oxygen species are believed to be involved in damaging D1 both in vivo and in vitro (13)(14)(15)(16). However, what damage specifically leads to degradation of D1 and how it is selectively degraded remain unclear. Light-induced degradation of D1 is retarded by photoaccumulation of plastoquinol at the Q B site (17) or by the addition of herbicides such as 3-(3,4-dichlorophenyl)-1,1-dimethylurea and atrazine (9, 11, 18 -22). Therefore, it has been proposed that degradation of damaged D1 is regulated by conformation of the acceptor side of PS II, which is affected by Q B site occupation by plastoquinone or herbicides (23). Evidence for changes in the conformation of the Q B site induced by herbicide binding includes the apparent influence of herbicides on the accessibility of trypsin to the target arginine residue located near the Q B site in D1 (18,24). Therefore, it has been hypothesized that D1 with conformationally modified Q B sites is selectively degraded by a further proteolytic process.
Degradation of D1 is initiated by cleavage at specific sites, resulting in several distinct fragments. Extensive in vivo and in vitro studies have shown that the primary cleavage occurs at the loop domain connecting transmembrane helices D and E, yielding N-terminal and C-terminal fragments with apparent molecular masses of 23 kDa (25)(26)(27)(28)(29) and 10 kDa (27,28,30). However, the precise site of cleavage has not been determined. Since this loop domain serves as the binding site for Q B plastoquinone as well as various kinds of herbicides, primary cleavage in this region appears consistent with a conformational change around the Q B site, leading to subsequent degradation of photodamaged D1.
The effects of various herbicides and PS II inhibitors on the stability of D1 were studied in an attempt to demonstrate that its degradation can be induced in the absence of photoinhibitory conditions through changes in the conformation of the D-E loop domain. Recently, we demonstrated that under dark conditions, D1 is selectively and specifically cleaved into two fragments with apparent molecular masses of 23 and 9 kDa when the Q B site is occupied by PNO8, a highly potent PS II inhibitor that specifically inhibits electron transport between Q A and Q B (31). PNO8 belongs to the category of phenol-type PS II inhibitors (32) because it contains the characteristic phenol nucleus and other structural features essential for phenol-type inhibitors (33). Furthermore, PS II of atrazine-resistant plants has increased sensitivity both to phenol-type inhibitors and to PNO8 due to replacement of Ser-264 by Gly in the D1 protein (32). PNO8 has a quite unique chemical structure due to its phloroglucinol nucleus. In a previous study, PNO8-dependent cleavage of D1 was suppressed at low temperatures and by inhibitors of serine-type proteases, but was not inhibited by the absence of oxygen (31). The 23-and 9-kDa fragments are also generated as degradation products by photoinhibitory illumination in the absence of PNO8 (27,28). Therefore, binding of PNO8 to the Q B site seems to induce conformational changes that are similar to those induced by photodamage in the D-E loop domain, triggering proteolytic cleavage. However, the population of D1 cleaved by PNO8 was limited to a minor fraction of the total PS II, despite PNO8 binding at Q B sites in all PS II centers (31). This suggests that PNO8-dependent cleavage of D1 requires certain specific conditions.
In this study, the selective and specific cleavage of D1 induced by PNO8 is further characterized. It is demonstrated that PNO8 binding to the Q B site leads to cleavage of the fraction of D1 that has been modified during illumination before the addition of PNO8. Possible mechanisms to explain the effects of preillumination are discussed. In addition, the site of PNO8-induced cleavage in D1 was studied to further characterize the proteolytic mechanism.

EXPERIMENTAL PROCEDURES
Preparations and Plant Materials-Thylakoid membranes were prepared from market spinach as described previously (31). They were resuspended in 400 mM sucrose, 15 mM NaCl, 5 mM MgCl 2 , and 40 mM MES/NaOH (pH 6.5) and stored in liquid nitrogen. BBY-type O 2 -evolving PS II membranes were prepared from spinach and seedlings of wheat by solubilization with Triton X-100 as described previously (34). O 2 -evolving PS II core particles were fractionated by centrifugation following solubilization of the PS II membranes with n-heptyl thioglucoside according to the method of Enami et al. (35) with modifications. PS II reaction center complexes were prepared from PS II membranes following Triton X-100 solubilization according to the method of Nanba and Satoh (3). Samples were stored in liquid nitrogen until use. After thawing, samples were washed and suspended in assay buffer containing 400 mM sucrose, 20 mM NaCl, and 40 mM MES/NaOH (pH 6.5). All procedures were carried out under a dim green light or in complete darkness, unless otherwise noted. The samples were then incubated in the dark for at least 2 h at 0°C, after which further treatments were applied.
Treatment with PNO8 and with Lysylendopeptidase-The sample suspension was incubated with PNO8 in complete darkness at 25°C. PNO8 was dissolved in dimethyl sulfoxide, and this was added to the sample such that the final concentration of dimethyl sulfoxide in the suspension mixture was Ͻ1% (v/v). For preillumination, the sample suspension at 25°C was illuminated with white light (50 mW/cm 2 unless otherwise noted) filtered through heat-absorbing and neutral density filters prior to the addition of PNO8 in the dark. The sample concentrations during PNO8 treatment were 400, 200, 80, and 4 g of Chl/ml for thylakoid membranes, PS II membranes, O 2 -evolving core particles, and reaction center complexes, respectively. The following buffers were used to treat the samples with PNO8 at various pH values: 40 mM MES/NaOH for pH 4.5-6.8, HEPES/NaOH for pH 7.1-7.5, and Tricine/NaOH for pH 8.0 -8.5, supplemented with 400 mM sucrose and 20 mM NaCl. After incubation, the treated membranes were quickly frozen and stored in liquid nitrogen until protein analysis. Digestion of wheat PS II membranes by lysylendopeptidase was carried out according to the methods described by Miyao et al. (15). A 120-l sample of a suspension consisting of 200 g of Chl/ml of PS II membranes was supplemented with 6 l of 1.0 M Tris, 1.8 l of 0.2 M EDTA, 18 l of 1% (w/v) SDS, and 30 l of 4 units/ml lysylendopeptidase (Acromobacter; Wako) and then was incubated at 30°C for 1 h in the dark to allow digestion.
Immunoblot Analysis-SDS-PAGE for immunoblot analysis was carried out according to Laemmli (36) with a 13% (w/v) polyacrylamide separation gel containing 5 M urea (1-mm thickness). The sample was solubilized in 60 mM Tris, 60 mM dithiothreitol, 3% (w/v) lithium dode-cyl sulfate, 0.003% (w/v) bromphenol blue, and 6% (w/v) sucrose at 0°C for 30 min in the dark. The separated polypeptides were electroblotted onto a nitrocellulose membrane with a pore size of 0.2 m (Schleicher & Schuell), as described by Towbin et al. (37), in the presence of 0.05% (w/v) SDS with a semidry-type blot apparatus (Bio-Rad) at 10 V at room temperature for 2 h. D1 and its fragments were probed with four different rabbit antisera, raised against D1 from spinach (provided by Dr. M. Ikeuchi, University of Tokyo) and against synthetic oligopeptides corresponding to Thr-227-Glu-235, Phe-239 -Asn-247, and Glu-333-Ala-344 of spinach D1. Each oligopeptide was designed to have an extra cysteine residue at its N terminus, through which a peptide was conjugated to maleimide-activated keyhole limpet hemocyanin (Pierce). Immunoreacted polypeptide bands were immunodetected using an alkaline phosphatase-conjugated goat antibody raised against rabbit IgG (Jackson ImmunoResearch Laboratories, Inc.) and visualized by reaction with nitro blue tetrazolium chloride and bromochloroindolyl phosphate (Sigma). Densitometric determination of immunoreacted bands was carried out using an image scanner (ATTO densitograph system).
Measurement of PS II Activity-O 2 evolution was measured at 25°C using a Clark-type oxygen electrode placed in assay buffer supplemented with 0.5 mM phenyl-p-benzoquinone as an electron acceptor.
Preparation of PNO8 -PNO8 was synthesized and separated using a silica gel column (38). PNO8 was then highly purified by high performance liquid chromatography with an octadecyl silica gel column (Pegasil ODS, Senshu Pack) using acetonitrile/H 2 O/formic acid (87.5:12.4:0.1, v/v) as the solvent system. The structure of synthesized PNO8 was confirmed by 1 H NMR, IR, and mass spectroscopy. No derivatives were detected during storage in dimethyl sulfoxide. Fig. 1 shows the effects of preillumination on PNO8-induced cleavage of D1. PS II membranes were illuminated under white light at 50 mW/cm 2 for 0.5 s in the absence of PNO8. PNO8 was added, and the membranes were further incubated for 60 min in the dark. The treated membranes were then subjected to SDS-PAGE, and D1 and its cleavage products were detected immunologically using D1 antiserum. As reported previously, incubation in the dark of membranes with PNO8 but without preillumination induced cleavage of D1, yielding two fragments of 23 and 9 kDa (Fig. 1, lane c) (31). The band of the 23-kDa fragment appeared as a doublet due to interference of the psbS gene product (39) superimposed on the fragment. When the PS II membranes were illuminated prior to the addition of PNO8, the quantity of the two bands of the fragments increased markedly to at least five times greater than the quantity induced without preillumination (lane d). Since the positions of the fragments in the preilluminated membranes were exactly the same as those without preillumination and no new fragments were detected in the preilluminated membranes, probably in both cases, D1 is cleaved at the same site. D2 was not cleaved by PNO8 treatment even after preillumination (data not shown). As shown in lanes c and d, PNO8 treatment induced the formation of a heterodimer between D1 and D2 revealed as a 60-kDa band, and preillumination also stimulated accumulation of the heterodimer. It is worthwhile to note that photoinhibitory illumination also leads to the formation of this heterodimer band (8,14). Fig. 2 shows the effect of preillumination on the cleavage of D1 induced by PNO8 in thylakoid membranes, PS II membranes, O 2 -evolving PS II core particles, and PS II reaction center complexes. Preillumination was as described for Fig. 1. Preillumination stimulated the accumulation of the two fragments in the thylakoid membranes (Fig. 2, lane c) and core particles (lane i) as in the PS II membranes (lane f), but no other new fragments were induced by preillumination. In contrast, D1 was not cleaved in the reaction center complexes, even in the presence of PNO8, irrespective of preillumination. It is unlikely that loss of D1 cleavage is due to the absence of O 2 evolution capability in the complexes because preillumination enhances PNO8-dependent D1 degradation in PS II membranes depleted of the manganese cluster (data not shown). It has been reported that the reaction center complexes bind some phenol-type herbicides, diuron and atrazine, with somewhat lower affinity compared with thylakoids (40). It therefore might be assumed that PNO8 binds to the reaction center complexes, but does not induce the cleavage of D1. Perhaps this is due to large perturbations on the acceptor side of PS II that occur during preparation, but we cannot rule out the possibility that PNO8 is incapable of binding to the Q B site in the reaction center complexes. Alternatively, some proteinaceous factor re-sponsible for D1 degradation may be absent in these preparations. Fig. 3 shows the dependence of the PNO8-dependent D1 cleavage on the light intensity of preillumination. The quantity of the 23-kDa fragment induced by preillumination was plotted against the light intensity. The quantity of the 23-kDa fragment band increased rapidly with increasing light intensity and reached saturation at ϳ100 mW/cm 2 . At this intensity, O 2 evolution by PS II membranes was only 8% saturated. Clearly, the light intensity required for PNO8-dependent cleavage is much lower than that needed for O 2 evolution. It has been proposed that active oxygen species including 1 O 2 , superoxide, and H 2 O 2 are responsible for the damage to D1 (13)(14)(15)(16). It might be assumed that such species are generated during preillumination and are responsible for the stimulation of the PNO8dependent cleavage of D1. However, O 2 does not directly participate in the mechanism of D1 cleavage induced by PNO8 (31). Active oxygen species are generated at the acceptor side (41) or at the reaction center of PS II (16) due to reduction of molecular oxygen or charge recombination. Since their yields increase in proportion to the rate of the electron transport reaction in PS II and show the same dependence on light intensity as does O 2 evolution, active oxygen species do not play a direct role in the preillumination effect.

RESULTS
In the above experiments, PNO8 was added to the membrane suspension immediately after preillumination. The lifetime for predisposition to PNO8-dependent D1 cleavage was estimated by illuminating the membranes for 10 s, incubating the membranes in the dark for various periods of time, and then adding PNO8. As shown in Fig. 4, the stimulation effect of preillumination persisted for a surprisingly long time even at room temperature and decayed exponentially, with a decay half-time of ϳ1 h at 25°C (Fig. 4, inset). No D1 fragments were detected following preillumination or during incubation in the dark in the absence of PNO8. Fig. 5A shows the effect of repetitive illumination on the PNO8-dependent cleavage of D1. PS II membranes were preilluminated for 0.5 s and then incubated in the dark after the addition of PNO8. During incubation in the dark, the sample suspension was illuminated intermittently every 30 min for 0.5 s. The quantity of the 23-kDa fragment that accumulated during each 30-min incubation period in the dark was nearly the same after the first and second illuminations, decreased progressively after each successive illumination, and showed virtually no change after the fifth illumination. The quantity of D1 decreased to ϳ40% of the original after the fifth illumination, but no fragments other than the 23-and 9-kDa fragments were detected. It should be noted that no fragments were detected even after incubation for 6 h in the absence of PNO8 (data not shown). Fig. 5A also shows the accumulation of the 23-kDa fragment following a single preillumination. The 23-kDa fragment appeared with no lag time and reached a constant level after 150 min, in good agreement with the kinetics for its formation in the absence of preillumination (31). This indicates that preillumination does not stimulate the rate of the cleavage reaction, but instead increases the number of PS II susceptible to PNO8 treatment. As shown in Fig. 5B, the quantity of the 23-kDa fragment that accumulated during repetitive illumination showed an inversely proportional relationship to the quantity of D1. It therefore can be concluded that the PNO8-dependent cleavage of D1 at a single site is solely the direct cause of the decrease in the quantity of D1.
The results also show that the effect of repetitive illumination is cumulative with respect to stimulating the cleavage of D1. Since the light intensity of each illumination was near the saturation level for cleavage, the cumulative nature of the preillumination effect may imply that modification of D1 can be reversibly and repeatedly effected by illumination, thus resulting in a state that is cleaved upon binding of PNO8. The first illumination was in the absence of PNO8, but successive illuminations were in the presence of PNO8, a potent PS II inhibitor. This indicates that preillumination stimulates the PNO8dependent cleavage of D1 even when electron transfer from Q A Ϫ to Q B is interrupted. This is consistent with the finding that preillumination stimulated the cleavage of D1 in core particles that had been depleted of a plastoquinone molecule at the Q B site, as shown in Fig. 2. Fig. 6 shows the pH dependence of the cleavage of D1 and the inhibition of O 2 evolution activity by PNO8. In this experiment, PS II membranes were incubated with PNO8 for 60 min with no preillumination to minimize any possible effect of pH on the preillumination process. D1 cleavage, as measured by the formation of the 23-kDa fragment band, was relatively low above pH 7 and increased steeply below pH 7, reaching a plateau at around pH 4.5. This is not ascribed to a pH-dependent change in the capability of PNO8 to bind to the Q B site since the potential of PNO8 as a PS II inhibitor showed a constantly high level between pH 4.5 and 7.0 and decreased above pH 7.0. Preillumination was found to stimulate the PNO8-dependent D1 cleavage at pH 5.0 as well as at pH 6.5 (data not shown). Throughout the pH range examined in this study, only the 23and 9-kDa fragments were detected as cleavage products of D1, indicating that the cleavage site remains the same and that the pH dependence of the formation of the 23-kDa fragment reflects a cleavage reaction at a single site.
The cleavage site acted on by PNO8 in D1 was further studied by limited proteolysis using lysylendopeptidase, as shown in Fig. 8. D1 from spinach does not contain a lysine residue, while D1 from wheat contains one at position 238 in the D-E loop connecting helices D and E (42). When D1 from wheat is cleaved at Lys-238 by digestion with lysylendopeptidase, two polypeptides with molecular masses of 26.0 and 12.0 kDa as determined by amino acid sequences should be produced (42). After digestion of control membranes by the peptidase, the band corresponding to native D1 completely disappeared, and two new bands are observed with apparent molecular masses of 20.5 and 12.5 kDa (Fig. 8A, lane c). Since anti-(227-235) reacted with the 20.5-kDa band, but not with the 12.5-kDa band (Fig. 8B, lane c), the origins of the 20.5-and 12.5-kDa bands can be assigned to the 26.0-kDa N-terminal and 12.0-kDa C-terminal polypeptides, respectively. Incuba-tion of wheat membranes with PNO8 resulted in the formation of the two bands with apparent molecular masses identical to those of the PNO8-induced fragments of spinach D1 (Fig. 8A,  lane b). Since anti-(227-235) reacted with the 23-kDa band, but not with the 9-kDa band (Fig. 8B, lane b), it can be concluded that PNO8 induced cleavage of wheat D1 at the same site as it did for spinach D1.
After digestion with the peptidase, the 23-kDa fragment disappeared (Fig. 8, A, lane d; and B, lane d), but the 9-kDa fragment was not affected (Fig. A, lane d). The disappearance of the 23-kDa fragment implies that this fragment contained Arg-238 and was digested to the 26.0-kDa N-terminal polypeptide and a residual small polypeptide, which was not resolved  (lanes a, c, e, and g) or with (lanes b, d, f, and h) 10 M PNO8. No preillumination was carried out. D1 and its degradation products were detected with antisera raised against the whole D1 protein (lanes a and b) and against synthetic oligopeptides corresponding to Thr-227-Glu-235 (lanes c and d), Phe-239 -Asn-247 (lanes e and f), and Glu-333-Ala-344 (lanes g and h) of D1. The sample quantity loaded in each SDS-PAGE well was 1 g of Chl. Some bands found both in the untreated control and PNO8-treated membranes were due to nonspecific reactions of the antiserum with PS II proteins.

FIG. 8. Characterization of fragments of D1 induced by PNO8 treatment by lysylendopeptidase digestion.
Wheat PS II membranes were illuminated at 25°C for 10 s, followed by incubation in the dark at 25°C for 60 min after the addition of 10 M PNO8, and then were digested with lysylendopeptidase. D1 and its degradation products were detected with antisera raised against the whole D1 protein (A) and against synthetic oligopeptides corresponding to Thr-227-Glu-235 of the D1 protein (B). Untreated membranes before (lane a) and after (lane c) digestion and PNO8-treated membranes before (lane b) and after (lane d) digestion with lysylendopeptidase are shown. by the SDS-PAGE system used in this study. With the SDS-PAGE system containing high concentrations of urea, all the hydrophobic proteins of PS II show that a linear relationship exists between the logarithm of relative mobility and molecular mass deduced from the amino acid sequences (43). We found in our gel system that the two peptides induced by digestion with lysylendopeptidase also participated in the linear relationship. As determined from the difference of the migration distance in the gel between the bands for the 23-kDa fragment and the 26.0-kDa N-terminal peptide, the molecular mass of the 23-kDa fragment is larger by 2.3 kDa than that of the 26.0-kDa N-terminal polypeptide. Taking into account the amino acid sequence of D1, it can be concluded that cleavage occurs around Leu-258, although the exact site is not yet known due to the relatively large ambiguity in determining the molecular mass from a protein band on SDS-PAGE. DISCUSSION This study demonstrates that the cleavage of D1 by PNO8, an inhibitor of electron transport between Q A and Q B , is greatly stimulated by illumination of the sample material before the addition of PNO8, although light is not necessary for the cleavage process. The effect of preillumination decayed exponentially in the dark, and very small quantities of degradation products were induced by PNO8 treatment after prolonged incubation in the dark, as shown in Fig. 4. This implies that PS II is altered as a result of preillumination and that binding by PNO8 can induce the cleavage of D1 only in PS II subjected to preillumination. A straightforward interpretation of the effect of preillumination is that some redox-active component is either photoreduced or photo-oxidized during preillumination, and this plays a direct role in the cleavage of D1 by PNO8. To assess this possibility, we attempted to identify redox-active components that show a relaxation course after illumination coinciding with the dark decay for the PNO8-dependent cleavage of D1 shown in Fig. 4. The effect of preillumination was still pronounced in the PS II core preparation (Fig. 2), in which the electron transfer from Q A Ϫ to Q B was largely retarded presumably due to the lack of a plastoquinone molecule in the Q B pocket and the PS II membranes depleted of the manganese cluster by NH 2 OH treatment (data not shown). It can therefore be concluded that the redox reactions of Q B and the manganese cluster do not achieve the effect of preillumination. Y Z ϩ , Q A Ϫ , Pheo Ϫ , and P680 ϩ are also precluded as candidates since their lifetime is much shorter than that of the preillumination effect (2). Y D ϩ is a radical species present on the donor side of PS II and could be formed during preillumination (2). Although Y D ϩ could be considered as a candidate with respect to its stability, quantitative analysis by EPR showed that 70 -80% Y D is still in an oxidized Y D ϩ form in the dark-adapted PS II membranes (data not shown). Y D ϩ cannot, therefore, account for the 5-fold increase in D1 cleavage following preillumination (Fig. 1). Cytochrome b 559 is also not a likely candidate since it is not reduced or oxidized by illumination at ambient temperatures in preparations with a functional manganese cluster (44). Chlorophyll and carotenoid molecules can be oxidized by PS II under certain conditions (44), but not under the present experimental conditions. Furthermore, no new EPR signal was observed in the PS II membranes following preillumination, indicating that no new organic radical(s) had been formed (data not shown). Thus, these two pigments can both be excluded.
The non-heme iron present on the acceptor side of PS II is usually present in reduced form under darkness, but it has been reported to be partially oxidized gradually by oxygen in PS II membranes and then re-reduced upon illumination (45). It has been proposed that non-heme iron in a reduced form plays a role in the cleavage of D1 via an active oxygen species in vitro system (15). In the present study, however, Ͼ80% of the non-heme iron was in a reduced state even after dark adaptation as detected by means of low temperature EPR and Fourier transform IR spectroscopy (data not shown). It is therefore unlikely that the preillumination effect is related to the photoreduction of the non-heme iron.
The above considerations suggest that none of the redox states and/or redox reactions of any PS II components are directly related to the effect of preillumination. A remaining explanation for the effect of preillumination is that the redox reaction somehow changes D1 so that cleavage can take place upon PNO8 binding. If this is the case, then the putative change in D1 must persist after complete relaxation of the redox reaction itself. Since O 2 evolution activity was unaffected by preillumination, the quantity of D1 irreversibly photodamaged during preillumination is too small to account for the amount of D1 cleavage by PNO8. It has been proposed that photodamaged D1 is either protected from or tagged for degradation (46,47) when it is phosphorylated in vivo, although D1 phosphorylation does not seem to influence PS II activity. Our results are consistent with regulation of D1 degradation by its modification. Preillumination may therefore induce an alteration that does not affect PS II functions, but that makes D1 susceptible to the cleavage. However, an irreversible covalent modification of D1 should not be ascribed to the preillumination effect since there is a dark decay of susceptibility to degradation. The possibility cannot be ruled out, however, that preillumination influences a PS II component other than D1 and that this indirectly affects D1 protein cleavage via PNO8. In previous report, we proposed that the conformation of the Q B site occupied by the PNO8 molecule would be similar to that induced by photoinhibitory treatment, and both allow cleavage of D1 presumably through a similar proteolytic process (31). Taking the above into consideration, it is reasonable to assume that some kind of conformational change in D1 in response to preillumination is required for the degradation of D1 with a modified Q B site induced either by photodamage or PNO8 binding. Further studies are required to assess whether lightdependent D1 phosphorylation or other modifications are involved in both its PNO8-induced degradation and light-dependent turnover.
The site of D1 cleavage promoted by PNO8 has been studied by immunodetection using antibodies to specific amino acid sequences of D1 (Fig. 7) and by the lysine-specific proteolysis of D1 from wheat with lysylendopeptidase (Fig. 8). The results indicate that D1 is cleaved at a single site located in the loop between helices D and E exposed on the stromal side of the thylakoid membranes, yielding a 23-kDa N-terminal fragment and a 9-kDa C-terminal fragment. Based on the change in apparent molecular mass of the N-terminal fragments when digested with lysylendopeptidase, it can be concluded that D1 is cleaved in the vicinity of Leu-258. This region contains the membrane parallel helix, de, and a loop connection between the de loop and helix E (48). Many amino acid residues in this region are implicated in the binding of the Q B quinone and various types of herbicides (49). These residues can also potentially participate in the binding of PNO8 and are thus candidates for the cleavage site. The site for D1 cleavage upon PNO8 treatment is in good agreement with that proposed for its degradation during photoinduced turnover and under strong light photoinhibition (25,26).
As shown in Fig. 5, Ͼ60% of the total D1 protein was cleaved by PNO8, but no other fragments except for the 23-and 9-kDa fragments were detected, indicating that no second-side cleavage takes place. This high specificity suggests that an enzymatic process is involved in the cleavage reaction induced by PNO8, consistent with our previous observation that the PNO8dependent cleavage of D1 is suppressed to 40ϳ60% by several inhibitors of serine-type proteases (31). It should be noted, however, that suppression by the inhibitors does not necessarily prove the direct involvement of a protease in the cleavage process by PNO8 since only partial inhibition is achieved. Nevertheless, there is no D1 degradation in reaction center complexes (Fig. 2), from which putative protease (or modification enzyme) may have been removed.
As shown in Fig. 6, the pH dependence of the PNO8-induced cleavage of D1 was considerably different from that of inhibition of PS II activity. Inhibition of PS II began to decrease above pH 7.0, whereas D1 cleavage was inhibited at lower pH values. Based upon the chemical properties of the PNO8 molecule with a nitro group in a phloroglucinol nucleus, the pK a value of hydroxyl groups in a phloroglucinol nucleus should be lowered to pH 7-8 compared with that in a phloroglucinol nucleus with no strong electron-withdrawing group. Thus, the decrease in the capability of PNO8 to act as a PS II inhibitor in alkaline conditions can be attributed to deprotonation of the particular hydroxyl group, which might influence the binding or inhibitory activity of PNO8 at the Q B site. The pK a value for D1 cleavage estimated from the pH dependence curve was found to be 5.8, and this value agrees with values reported for functional histidine residues of various enzymes (50). This implies that a histidine residue is involved when D1 is cleaved upon PNO8 binding. It is interesting to note that in a family of serine-type proteases, there exists a catalytic triad consisting of serine, histidine, and aspartate residues; the histidine residue functions as an activator of the serine residue and serves as a binding site for a water molecule during the course of proteolysis (50). A good possibility is that in the D-E loop domain of D1, there exists a special arrangement of amino acid residues that conform the structure functionally coincident with that of the catalytic triad in serine-type protease. In fact, several of these residues of the triad exist in the D-E loop, although there is no amino acid sequence homologous to that of serine-type proteases in D1. Perhaps a structural change induced in the Q B site by photodamage, attack of active oxygen, and PNO8 binding conformationally activates the putative triad, and cleavage of D1 at the D-E loop is autocatalytic.