1. Mechanism of proton uptake at the QB-site of photosystem II S. Padden, J. Minagawa, A. Kanazawa, Govindjee, and A.R. Crofts.

2. PS II structure showing pathway of electron flow on photoactivation Coordinates taken from 1fe1.pdb – (Witt et al.)

3. Fluorescence yield for reactions of Photosystem II P: P680, Sn: Mn-Cluster State, Z: TyrZ, QA bound quinone, QBsoluble quinone Reactants State Fluorescence Quencher SnZPQAQB Open Low QA  h SnZP+QA-QB Closed Low P+  (20 ns) SnZ+PQA-QB Closed High None  (~30 s for n=1) S+n+1 Z P QA- QB Closed High None  + H+ (~200 s) S+n+1ZPQAQB-(H+) Open Low QA  h S+n+1ZP+QA-QB- Closed Low P+  (200 ns) S+n+1Z+PQA-QB- Closed High None + H+ (~400 s) S+n+2ZPQAQBH2 Open Low QA

4. Kinetics of electron transfer from QA- measured from the fluorescence yield decay 1st flash - DCMU 2nd flash 1st flash Kinetics of electron transfer after the first flash can also be measured by dephasing of binary oscillation. After correction, kinetic trace after 1st flash can be well fit by three components of ~150 ms, ~2 ms, and >100 ms.

5. Determination of the Apparent Equilibrium Constant (Kapp) for the Sharing of One Electron in the Quinone Acceptor Complex as a function of pH (in pea chloroplasts) Rate of the back-reaction from QA- to the Mn-cluster as a function of pH (DCMU present) Rate of the back-reaction when QA and QB share the one electron present in the complex, as a function of pH (by rephasing with variable delay after flash 1). Apparent equilibrium constant (Kapp) for the electron between QA and QB (Kapp = (kQA / kQB) – 1, where kx is rate constant for backreaction from QA or QB). Kapp measured directly from extent of the decay kinetics at 10 ms, after correction for slow (> 100 ms) phase. Curves in C, D are calculated from values: pK on the oxidized complex at 6.0 - 6.4 pK on the one-electron reduced complex at 7.9-8.0 Values for Kapp varied between 3.5 and 95. Neither pK was due to pK on QA, as this has a value outside this range (at 8.9).

6. Scheme for the two-electron gate of PS II 70% 30% The path for most functional centers over the neural pH range (70% with QB bound) is the light-blue reaction followed by the green. The remainder (QB-site vacant) follow the yellow then green path. Robinson and Crofts (1983) had observed that the equilibrium constant for sharing an electron between QA and QB following a single flash (Kapp) was pH dependent. Crofts et al. (1987) suggested that the two pKs needed to describe the curve were most simply interpreted as reflecting a shift in pK of a single residue, induced by the coulombic effect as an electron was transferred to QB, possibly His-252.

7. Stable at pK 7 before flash On the basis of a tertiary structural model of the QB-site, Crofts et al. (1987) suggested that D1H252 was a likely candidate for proton binding, and pointed out that this would stabilize the electron on QB-. Effectively, the pK is shifted on formation of the semiquinone, and binding of a proton then stabilizes the QB-. In this lecture studies of mutant strains with this key residue changed will be presented. The results support the model, and their relevance to the mechanism of the QB-site will be discussed. Stable at pK 7 after flash

8. Model of the QB-site The tertiary model is based on the known structure of the bacterial reaction center from B. viridis The important residues are highlighted. In this talk the residue of focus are D1H252 and D1S264, the ‘front’ and ‘back end’ of the serine hydrogen bond switch respectively (Wraight, 2004)

9. QB-site of Photosystem II taken from pdb 3BZ1 (Guskov et al. , 2009) Green surf and transparent coil is the D1 protein, ice-blue surf is the D2 protein, plastoquinone is drawn as ‘licorice’ with the prominent residues drawn as ball and stick. Yellow is carbon, blue is nitrogen, and red is oxygen. Hydrogen bonds are white dashes with distance in white text

10. Mutant strains We have constructed mutants with D1H252 changed to aspartate (D1H252D), lysine (D1H252K), asparagine (D1H252N) and glutamine (D1H252Q). Molecular engineering was performed using PCR-cassette mutagenesis in a designed, intron-free psbA gene (encoding the D1 protein of photosystem II), using Chlamydomonas reinhardtii as a model for the plant system. Only the D1H252D mutant was able to grow photosynthetically, and cells evolved O2 rates 40 - 60% those observed in cells with wild-type protein. Both D1H252N and D1H252Q showed some electron transfer from QA- to QB following the first actinic flash after dark-adaptation, but both were severely inhibited on the second flash. D1H252K was severely inhibited even after the first flash, but this seems to be a kinetic effect, since thermoluminescence experiments suggest that QB- is formed after flash activation.

11. Rate of oxygen evolution in wild-type and mutants strains The graph shows that the D1H252D strain is capable of evolving oxygen at 45% - 60% compared to wild-type. The D1H252K, D1H252N, and D1H252Q strains show a greatly inhibited rate of O2-evolution. In strain D1H252K, a very weak rate of O2-evolution was detected Light on

12. Assay of electron transfer kinetics by measurement of fluorescence yield decay 2.0 An actinic flash is given at time zero, and a series of weak measuring flashes (<1% actinic) at the times indicated. This experiment tracks the ‘state’ of the reaction center, through the kinetics of decay of the high fluorescence yield due to the state QA-. As electrons are transferred from QA- to QB (or QB-) the centers ‘re-open’, and the fluorescence yield declines. pBA158 WT 1.0 0 10 ms O ms

13. D1H252 mutant fluorescence yield decayA: D1H252D; B: D1H252K; C: D1H252N; D: D1H252Q In the N and Q mutants the first flash suggests electron transfer but no stabilization. In the K mutant the electron transfer is decreased but occurs at a much reduced rate as seen in the TL data. In the D mutant, the electron transfer reverses phase when compared to WT. In the K, N and Q mutants the second electron transfer is reduced compared to WT, because protonation of the QB- state cannot occur. The D mutant rate is faster than the 1stflash.

14. Table 1: Kinetic parameters determined from first flash of the fluorescence decay curves of the WT and mutant strains of C. reinhardtii

15. Binary Oscillation of Wild-Type The back-reaction of QB- can be followed by observing the fraction of centers remaining in the QB-state, by measurement of the rephasing of the binary oscillation of the extent of decay at ~150 ms. After one actinic flash the electron on QB- will back-react with the Mn-Cluster If an actinic flash is delivered before the back-reaction has occurred then the 150 ms fluorescence yield (Fv(150)) will reflect electron transfer to the QB--state (i.e. will be high compared to that after the first flash) If the back-reaction has occurred, then the Fv(150)will be low (due to S1QB state). At a characteristic time, 50% of PSII will be ‘open’ and 50% will be ‘closed’ (~12 sec from the graph) so the oscillation is lost. This time is taken to be the t 1/2 for the back reaction from QB- to S2+ (Flashes after the 1st.)

16. D1H252 mutant binary oscillationA: D1H252D; B: D1H252K; C: D1H252N; D: D1H252Q The D mutant shows a weak binary oscillation expected for a functional 2-electron gate. However, since the second electron transfer is faster than the first, at short times the binary oscillation shows a reverse pattern. With a dephasing time( t1/2 )about 1 s The K and N mutants do not have any binary oscillation, while the Q has a binary oscillation apparent, with a dephasing time (t1/2) in the 2 to 3 second range.

17. DCMU Back Reaction: t1/2 of QA-/S2 The back reaction with (black) and without (red) DCMU (10 µM), a known inhibitor of electron transfer from QA- to QB This reaction measured the timing of the back reaction from the reduced QA- to the oxidized S2 state of the OEC from which you gather the half-time for the kinetics of the QA- back reaction. pBA158 WT + DCMU

18. DCMU Back Reaction of D1H252 MutantsA: D1H252D; B: D1H252K; C: D1H252N; D: D1H252Q

19. Table 2: Parameters of the two electron gate in wild type and mutant strain of C. reinhardtii

20. Thermoluminescence of Wild-Type Thermoluminescence is the fluorescence emission as an electron in the QA-QB or QAQB- state returns to S2 via P680*. The temperature of the peak is an indication of the depth of the trap. Cells were illuminated by a saturating flash before freezing to –70oC, and the glow curve measured while warming at a set rate. The peak at ~15oC seen in the presence of DCMU (the Q-band) reflects the smaller energy gap for promotion to the P* state from QA-. In the absence of DCMU, the band is at higher temperature (38-40oC), because QB- has a higher redox potential. B-band Q-band

21. Thermoluminescence of H252D In the absence of DCMU, the curve appears to have several components, with a peak at ~20oC, and a significant component at ~39oC With DCMU, the maximum is shifted to a normal Q-band, and the high temperature band is diminished. Possibly, the shift in band maximum is due to the presence of a B-band from a lower potential form of QB-, consistent with the smaller value for Kapp(lower Em for QB). B-band? Q-band

22. Thermoluminescence of H252Q The H252Q mutation has no obvious effect on the Q-band, and gives a B-band at 32o, somewhat lower in temperature than that of wild-type, suggesting a lower Em for QB, compatible with a loss of stabilization by protonation. B-band Q-band

23. Thermoluminescence of H252K The H252K mutation shows three interesting effects: The Q-band is normal, but the B-band is shifted down to 30o. Both bands are seen at equal amplitude in the absence of DCMU DCMU addition (1 mM) only partially eliminates the B-band. These suggest that some electron transfer to QB occurs in this strain, albeit slowly; that QB has a lower potential than in wild-type; and that DCMU binding is inhibited Q-band B-band

24. Table 3: Determination of maximal peak height at temperature of the thermoluminescence bands

25. pH Analysis of fluorescence decay for the WT (top) and D1H252D mutant (bottom)

26. pH analysis of the WT and D1H252D mutant black (WT), red (D)

27. Changes in parameters for the QB-site in a mutant of A. hybridus with Ser –264 changed to glycine (atrazine resistant strain) Note changes in pK in mutant strain The results suggested that Ser-264 might interact with a dissociable group involved in stabilizing QB-.

28. Serine Hydrogen Bond Switch Photosystem II from T. elongatus (cyanobacteria) (pdb: 3BZ1) Bacterial reaction Center from R. sphaeroides (purple bacteria) (pdb: 1AIG)

29. Modeling of the D1H252 mutations in the QB-site of PSII A: pBA158 (WT); B: D1H252D; C: D1H252K; D: D1H252N1st; E: D1H252N2nd,; F: D1H252Q

30. Conclusions The model of the photosystem II acceptor side based on the bacterial reaction center has provided a useful starting point for mutagenesis studies of the mechanism of the two-electron gate. D1H252 plays an important role in the mechanism, and without it the ‘serine hydrogen bond switch’ is non-functional. The properties observed in the mutant strains support the suggestion that protonation of D1H252 serves to stabilize the QB-state and is responsible for the first proton transfer to the anionic semiquinone. The change to aspartate (H252D strain), which maintains a dissociable group at the position, but with a different pK and initial charge, also maintains a limited two-electron gate function. Replacement with glutamine (H252Q) or asparagine (H252N), which retain the polarity of the histidine, but not the proton dissociation, are functional in transfer of the first electron, but defective in transfer of the second electron, possibly because protonation is required before the second electron transfer. Replacement by lysine (H252K) led to greatly inhibited electron flow on both flashes. However, the thermoluminescence data suggest that electron transfer occurred on the first flash, and the very weak O2 evolution may indicated a much inhibited but functional two electron gate. We suggest a mechanism in which the H+ taken up by protonation of D1H252 after the first electron transfer, is necessary for proton transfer to the QA-QB- state before the second electron transfer can occur. It is unlikely that Ser-264 is an important ligand to quinone, because the change to glycine does not strongly affect quinone binding in plants. It is likely that Ser-264 H-bonds to His-252 (‘the serine hydrogen bond switch’), and provides a relay for proton transfer to the reduced site. This H-bond modifies the pK values for D1H252.