The sequence of electron transfers which shuttles holes from P-680+ to the site of water oxidation has been studied extensively in both oxygen-evolving and inhib- ited preparations and was reviewed in detail by Van Gorkom [14]. In general, there is a reasonably good understanding of these processes in untreated preparations (Table 4). The development of inside-out thylakoids and 0,-evolving PS I1 par- ticles has provided a number of new ways by which to inhibit the system. Electron transfer following these newer treatments is less well understood, but the now wider array of inhibition methods is providing a finer understanding of the role of the various PS I1 components in electron transfer and water splitting.
3.1. Electron transfer in the untreated PS IIIOEC
The photosynthetically wasteful electron/hole recombination between P-680+ and QA occurs with a half time of -100 ps [180]. Thus efficient photosynthesis re-
TABLE 4
Kinetic parameters associated with several events in the water-splitting process
P-680' -+ P-680 23 ns
z+ Z' < 3 ps
z++ z -50 ps
s, -+ S n i , 30 ps
H' release 250 ps
0, release -
(amount) (1)
s, + s2 s, -+ s3 s3+ s, Refs.
23 ns 50, 250 ns 50, 250 ns 186
< 3 p s < 3 p s < 3 p s 193, 195 -50 p~ -400 ps 1 /.LS 193, 198
110 ps 350 p~ 1.3 msa 121,201
- 200 ps 1.2 rns 272
(1) (2)
- - = 1 ms 199
( 0 )
a Oxygen release is somewhat slower in PS I1 membrane fragments than in chloroplasts or in algae;
see Ref. 14.
quires that either P-680+ is reduced or QA is oxidized in considerably shorter times.
The observation that QA transfers electrons to subsequent acceptors in the 100-500
ps time range [181] indicates that the reduction of P-680+ must be in the sub-ps range. Owing to this fast reduction time and to the fact that electron donation to P-680' is slowed in preparations which do not evolve 0, (Section 3.2), establish- ing the functionally relevant time course of P-680+ reduction proved difficult.
Mauzerall [182] and Van Best and Mathis [183] showed that this occurs in the ns range and Witt and co-workers extended this approach substantially [184-1861. The reduction of P-680' is multiphasic and dependent on the number of oxidizing equivalents stored in the water-splitting ensemble. For the states So and S1, P-680+
is reduced with a major phase of 20 ns and a minor phase of 35 ps; for S, and S3 the prevalent phases are 50 and 260 ns, with an additional minor phase of 35 ps [187]. The amplitude of the 35 ps phase is S-state-dependent and is maximal for S2 and S,. In the interpretation of Witt and co-workers, the 35 ps component is a consequence of rapid equilibria on the donor side of PS I1 and corresponds to for- ward charge separation-conserving electron transfer. Van Gorkom [ 141 suggests that the 35ps component arises from back reaction in PS I1 p centers (see Chapter 4), a view which is consistent with Eckert and Renger's conclusion that charge storage does not occur as a result of the 35 ps P-680+ decay phase [188].
To explain the S-state dependence of the sub-ps decay phases, Witt and co- workers postulated two donors operating in series between the OEC and P-680 and an equilibrium between the first and P-680' that depends on the net charge on the OEC. In this model, the second donor corresponds to the EPR-detectable Z spe- cies [ 1851. The question of the number of endogenous donors between P-680' and the O E C was reviewed lucidly by Bouges-Bocquet in 1980 [189]. Thus far the only donor for which there exists direct spectroscopic evidence is the Z species dis- cussed above, although under nonphysiological conditions Cyt b-559 [ 1901 caro- tenoid [191] and chlorophyll [192] may donate to P-680+ (see Chapter 4). In 0,- evolving preparations, the rise of Z+ has been measured by EPR with 3 ps time resolution and faund to be instrument-limited [ 1931. This observation may be ra- tionalized by either one or two intermediate donor models. In inhibited prepara- tions, however, the data clearly favor only a single intermediate. The rise of Z+
parallels the decay of P-680' in inhibited PS I1 particles [194,195] and in CI--de- pleted preparations only two reducing equivalents, one from the S , -+ S2 transi- tion, the other from the oxidation of Z, can be readily extracted from the donor side [196,197]. Van Gorkom has suggested an interesting rationalization for some of these observations within the two-intermediate framework [ 141. However, given the data in inhibited systems and the lack of direct, physical evidence for donors other than Z, it appears at present that the one-intermediate model is minimally sufficient to explain the bulk of the existing data.
While there is uncertainty as to whether Z is the immediate donor to P-680' in 0,-evolving preparations, there is general agreement that Z + is the immediate ox- idant of the O EC manganese. Babcock et al. used time-resolved E PR to show that the reduction of Zt is S-state dependent [198]. During the S, + S, and S, + S4
-+ S,, transitions. Zt is reduced in 400 ps and 1 ms, respectively. The latter time corresponds to the overall rate-limitation in 0, evolution [199]. This indicates that electron transfer, and not H,O/O, chemistry, is rate-limiting in water splitting, consistent with Sinclair and Arnason's observation of a lack of a significant deu- terium isotope effect in O2 evolution [200]. Z+ reduction during the S,, -+ S , and S , -+ S2 transitions was suggested to occur in times fast relative to their -100 ps time resolution. Boska and Sauer provided support for this by showing fast (=50 ps) decay transients in the Z+ reduction kinetics under steady-state conditions [ 1931.
These results argue against parallel donor models in which Zf functions only with S2 and S, and an alternate intermediate serves S , and S, [189]. Dekker et al. [201]
have monitored the S-state transitions directly in their time-resolved optical work and have found complementary times for the S-state transitions (Table 4; the slight slowing down of the final step is due to the fact that these measurements were made in PS I1 particles [14]). These results are significant not only for their kinetic im- portance but also because they provide support for the clever but complicated de- convolution procedures used to obtain optical difference spectra for the various S states. The interpretation of these data continues to be controversial, however.
Results from Witt's group [202] support the pattern of charge accumulation in the Leiden model but disagree with those of Dekker et al. [201] in the assignment of absorption spectra, particularly on S,, + S , [186]. Lavergne [203] and Renger and Weiss [121] have conflicting proposals as well (Section 4 . 2 ) . The latter group has also determined the ZtS, reaction kinetics and finds good agreement with the other optical and EPR data [121].
With the exception of the S4 state, which reacts in -- 1 ms t o produce O,, the higher S states are remarkably stable. The early literature on this topic has been reviewed [204]. Both S2 and S, survive for several seconds in the dark following their formation. These times were originally measured in electrode oxygen-evo- lution experiments (e.g. Refs. 2 4 ) . Recently, thermoluminescence has been used as a probe of higher S-state decay [205-21)8], and Inoue, Rutherford and co-work- ers extended early work implicating the redox state of the PQ pool in S-state deac- tivation by determining the S2Q, and S3QB and the S,Q, and S3QA recombina- tion times [205,211]. These results correlate with earlier 0, electrode and delayed luminescence measurements [204,209,210]. This group has also used the technique
to study protonation of both donors and acceptors to PS I1 [211].
A remarkable aspect of the higher S-state stability is that their redox potentials are necessarily high (at least +0.6 V and more likely in the +0.9 V range), which requires that stabilization is achieved by kinetic, not thermodynamic, means. This apparently involves simple limitation of access of endogenous and most exogenous reductants to the manganese ensemble, with the 17, 23 and 33 kDa peripheral polypeptides playing a role in this shielding mechanism. Velthuys has developed a useful assay of this phenomenon by monitoring optical changes which result from TMPD reduction of donor side components; a significant decrease in the second- order rate constant is observed in oxygen-competent vs. oxygen-inhibited PS I1 particles [212,213]. Similar access limitation has been shown for the exogenous do- nor, benzidine, in salt-washed preparations and the 33 kDa polypeptide has been implicated in this process [27]. The importance of the 17 and 23 kDa polypeptides in the shielding mechanism has been demonstrated by their role in limiting access of H 2 0 2 to the O E C in inside-out thylakoids [214]. One class of compounds, li- pophilic anions (also termed ADRY reagents by Renger, who has characterized these most extensively (e.g. Ref. 215)) appears to be able to overcome the acces- sibility barrier and deactivate higher S states effectively. Models in which the li- pophilic anion acts as a direct reductant [216] or as a redox-inactive catalyst [215]
have been proposed.
Chloride depletion provides additional stabilization of the higher S states [172,217] and Velthuys showed that NH, and methyl amines similarly increase the stability of S, and S, [218,219]. Homann et al. interpreted the altered thermolu- minescence properties of Cl--depleted samples to indicate that C1- displacement provides a thermodynamically stabilized higher S state i.e., the redox potential of the state is lowered [217]. Ghanotakis et al. had invoked a similar thermodynamic stabilization argument for the amine case to rationalize their finding that NH, blocks benzidine electron donation to the higher S states in NaC1-washed PS I1 particles [27]. Thus, a plausible explanation for inhibition of O2 evolution by C1- depletion or by amine binding is that in the treated systems the higher S states are structur- ally modified and no longer thermodynamically able to oxidize water. Ono et al.
have reported an experiment which shows that the S2 state is indeed modified by C1- depletion, but which suggests that the relationship between structure and ther- modynamics may be more complicated than indicated above [220]. In the absence of the 17 and 23 kDa polypeptides, the Sz multiline signal may only be generated if sufficient CI- is present [221,222]. However, if C1- is added in the dark after single flash excitation of CI--depleted, 17 kDa and 23 kDa polypeptide-depleted membranes, the multiline signal spontaneously regenerates. Signal regeneration by dark C1- addition shows the prior formation of a modified S,; the spontaneity of multiline formation suggests that the modified and regenerated S2 states are not thermodynamically far apart. Several factors influence this result, however, in- cluding assay temperature and C1- concentration, and a more detailed consider- ation of these seems appropriate.
In the original formulation of the S-state model, Kok et al. postulated that the dark resting state of the OEC was a 3:l mixture of S , and So, respectively [3]. This
unusual distribution has provoked considerable speculation as to its origin [204]
although no definitive model has emerged. Several developments have sparked re- newed interest in this phenomenon. Beck et al. have shown that the characteristics of the multiline EPR signal are dependent upon the dark incubation time prior to its generation [128]. This suggests that S,, and S, states are subject to long-yrm reorganization, either in the manganese ensemble directly or in the protein and lipid matrix. Plijter et al. found that the dark SOIS, distribution is pH sensitive, at least in PS I1 particles, and that at high pH almost 100% S,, is favored in dark- adapted material [ 2 2 3 ] . The observed p K , for this phenomenon. 7.6. correlated with a similar pK, for a structural modification in PS 11 which allows access of trypsin t o the O E C [224]. Alternatively. at lower pH the S, state appears to occur in nearly 100% of the centers [225]. The apparent 3:1 ratio, which is determined from the oxygen yield pattern. was attributed to the interception of oxidizing equivalents by the reduced form of the stable E PR Signal 11 species, D, during the initial flashes in the sequence. Earlier EP R data had shown that this species reacts with the state S, or S, to form D i and S, or S2, respectively [226]. Rutherford has correlated these observations to suggest that D + functions to oxidize So to S , in the dark so as to prevent accumulation of the lowest S state [17]. Brudvig has made a similar suggestion and, as noted above, points out that the lower S states may be expected to be more labile to metal loss (personal communication).
3.2. Electron transfer in the PS IIIOEC following inhibition
With the progress made recently in understanding the interplay between the pe- ripheral polypeptides and the required small ions, chloride, calcium and man- ganese, the effects of PS I1 oxidizing side inhibition have become both more clear and more complicated. A good deal of effort is currently being spent defining the loci of inhibitory treatments, particularly when manganese is not released by the inhibition. The situation when manganese is released is more clear and is dealt with first.
A number of PS I1 inhibitor treatments , such as Tris-washing and NH,OH ex- traction, had been well established as releasing Mn from the O E C [109]. Aker- lund, Anderson and co-workers showed that these treatments also released the peripheral polypeptides [71]. Mathis and co-workers showed that the ns phases in P-680+ reduction that occur in untreated preparations are replaced by ps com- ponents in inhibited samples [180]. The predominant decay phase is pH depend- ent, -2 ps at p H 8 and -45 ps at pH 5 . and is attributed to P-680+ reduction by a secondary donor, usually designated D , in optical work, which is identical to the EPR-detectable Z species [180,194,195]. The bulk pH and salt concentration de- pendencies of this phase implicate local membrane pH in influencing its time course [227]. The decreased rate of electron transfer to P-680f following Tris inhibition has been explained as indicating a shift of -120 mV in the Z+lZ redox potential [228]. Slower reduction phases, -100 ps and pH independent, may also be de- tected after inhibition and correspond to P-680' QA recombination. Ford and Ev- ans have pointed out that the rate of the recombination depends on the redox state
of Z [229]. Warden and Goldbeck found that another donor competes with Z in reducing P-680' in these preparations. Its redox potential (240 mV) precludes a direct role for this species in charge accumulation and the authors tentatively iden- tify it as Cyt b-559 [230]. Only one treatment so far, NH,OH at modest concen- tration, is known to inhibit electron transfer specifically between Z and P-680 [29,231]; interestingly, this inhibition is reversible.
With the 17, 23, and 33 kDa polypeptides and Mn removed, reduction of Z t occurs at the expense of endogenous reductants. Exogenous donors, when added, accelerate this reduction in a straightforward second-order kinetic process that provides the principal electron entry site to PS I1 in inhibited preparations. The efficiency of the donor increases as its hydrophobicity increases, reflecting the hy- drophobic environment of the Zt species [232].
A variety of treatments have been developed which allow more delicate manip- ulation of the polypeptide/ionic cofactor composition of the PS II/OEC. With care, manganese is usually retained in its binding site(s) following these treatments, at least some of the S-state transitions occur, and partial or full reactivation of 0, evolution may be achieved. Thus it is now possible to remove either the 17 kDa polypeptide selectively or the 17 and 23 kDa polypeptides with salt-washing, with loss of 0,-evolution activity, provided CI- and Ca2+ concentrations are sufficiently low [7,8]. The latter preparations have been characterized most extensively. The reduction time of P-680+ shifts into the ps time range under 'multiple turnover' conditions [233], although evidence from optical [32,234], EP R [27,222,235], lu- minescence [236], thermoluminescence [237] and TMPD oxidation [213] indicates that the lower S-state transitions proceed. The CI- and Ca'+ requirements for the S-state transitions are likely t o be different, however. In the absence of both CI- and Ca2+, the system apparently proceeds only to a modified S, state upon illu- mination [220]. The multiline EPR is absent [238,239], but may be induced by CI- addition either before [221,222] or after [220] photon absorption without reacti- vation of 0,. In CI--sufficient, Ca2+-deficient preparations the system appears un- able to advance beyond Z+S, to split water (e.g. Ref. 236). An understanding of electron transfer in these preparations is evolving but controversies continue. The origin of the disputes lies, no doubt, in the complexity of the interactions. Two factors may be particularly bothersome. (1) Even in the absence of the 17 and 23 kDa polypeptides, there is evidence of tight and/or residual Ca2+ binding [82,156,234]. Care must be taken to deplete this site prior to assay. Recent reports indicate, for example, that multiline EPR formation may require both Ca2+ and C1- [43,222], contrary to earlier conclusions that only CI- is necessary. (2) There are likely to be multiple binding sites, particularly for CI-, in the O E C and these may have different pH and CI- concentration dependencies (see below). The con- flicting reports [217,220] on the possibility of forming a modified S3 state in C1-- deficient preparations may be a manifestation of this phenomenon.
Addition of high concentrations of both CaZt and CI- to 17, 23 kDa depleted preparations restores water splitting (Section 2.3). In the reactivated samples, however, there are changes in the behavior of the system: (a) the miss parameter, a , is about double its value in untreated controls [213,234] (but see Ref. 236), (b)
the 0,-release reaction is slowed by a factor of about 4, although electron transfer remains limiting [234], and (c) Cyt b-559 [240] remains in its low-potential form [241,242]. Readdition of the polypeptides in addition to Ca2+ and C1- restores high- potential b-559 (D.F. Ghanotakis, unpublished), but does not affect either a or the slowed 0, release [234].
Characterization of electron transfer in preparations depleted of all three pe- ripheral polypeptides but which retain manganese has begun [213,237]. The data thus far are contradictory, but indicate that in these preparations some of the S- state transitions occur as well.
Several PS I1 inhibitory treatments exert their effects only when the O E C is in its higher S states. These include C1- depletion, F- treatment, incubation at al- kaline p H and amine inhibition. It appears now that all four of these treatments are related to the chloride requirement for O2 evolution, as F-, O H - and amines are competitive inhibitors (with respect to Cl-) of water splitting (Section 2.3).
Recent data suggest two different sites for CI- in the O E C - one at the S, state, the other in facilitating the water-splitting S; .+ So transition. The S, site is dem- onstrated by the fact that only two reducing equivalents may be extracted from the donor side of PS I1 following CI- depletion under certain conditions [196,197,220].
Chloride titration of the restoration of water-splitting activity [ 1761, oxygen-re- lease phase lag measurements [ 1751 and EPR and 0,-evolution measurements [ 1241 provide evidence for the second Cl--sensitive site at the S3 -+ S4 -+ So transition.
Damoder et al. have found that the CI- binding constant for this site is roughly an order of magnitude less than for the site controlling the advance beyond the modified S, state; moreover, they find a slightly different effectiveness for the var- ious anions which substitute for C1- at these two sites. Baianu et al. [177] may have provided an estimate of the binding constant for the higher-affinity site (-0.5 mM).
The concept of two sites of C1- action does not necessarily imply distinct physical structures, however; the S, and the S, + S4 +. So loci may be the same molecular site but with binding constant parameters altered by, for example, a valence change nearby.
Although the picture which is emerging for C1- in PS I1 is beginning to clear, ambiguities and disagreements remain (e.g., compare Ref. 124 with Ref. 220). Part of the confusion may arise from the competition between OH- and C1- for the sites involved. Discrepancies in the literature regarding the effect of pH on the in- hibition of electron transfer in PS I1 may also reflect this competition. Briantais et a1 [243] had implicated the S2 state as the inhibition site at high pH. More recent work on the S, multiline E P R signal, almost all of it in PS I1 membrane fragments, has yielded contradictory results. Cole et al. note reversible inhibition of multiline formation at pH 7.5 [244] and a pH titration of 02-evolution inhibition similar to that reported by Sandusky et al. [78]. Damoder and Dismukes [245] and Beck et al. [246], however, see normal multiline formation at this pH. Cole et al. suggest that the reversible inhibition reflects OH-/CI- competition, which implies that the extent of alkaline inhibition will be a function of CI- concentration. Such CI- ef- fects have been reported (e.g., Refs 247,248) and recently Vass et al. [249] have shown such a C1- protecting effect at high pH. They associated the OH--sensitive