1
Charge stabilization and recombination in Photosystem II containing the D1′ protein product of the psbA1 gene in Synechocystis 6803

2
We have studied the process of charge stabilization and recombination in various engineered strains of Synechocystis 6803 in which either the whole D1 protein of the Photosystem II reaction center was replaced by the D1′ product of the psbA1 gene, or prominent point mutations of D1′ were introduced into the normal D1.

3
Both the up-shifted thermoluminescence peak positions and increased time constants of flash-induced chlorophyll fluorescence components, that can be assigned to charge recombination, indicated that the energetic stabilization of the S2QA and S2QB charge pairs is increased in D1′ containing Photosystem II.

4
Similar stabilization of S2QA was induced by the Phe186Leu, Phe186Ala and Phe186LeuPro162Ser substitutions in the normal D1 protein.

5
However, these replacements also affected the function of the QA − QB two-electron gate by slowing down the QA to QB electron transfer step and decreasing the redox gap between QA and QB.

6
We conclude that in the D1′-containing PSII centers and in the Phe186 mutants, Em(S2/S1) is decreased by 20–25 mV.

7
In the Phe186 mutants, this affect is accompanied by a 40–50 mV decrease of Em(QB/QB), which provides an example of pronounced long-range influence of PSII donor side modifications on the QA − QB binding region.

8
However, this acceptor side effect is compensated by the other substitutions of the D1′ protein, which restores efficient electron flow between QA and QB, indicating that D1′-containing PSII centers might have a physiological function under certain environmental conditions.

Introduction

9
PSII catalyses the light-driven oxidation of water, evolving one molecule of O2 and four protons per two H2O molecules oxidized.

10
The reaction center complex is made up by the heterodimer of the D1 and D2 proteins, that bind or contain the redox cofactors involved in light induced electron transport.

11
The site of water oxidation is situated on the lumenal side of the thylakoid membrane, containing four Mn ions and the redox active tyrosine Tyr-Z.1–5

12
One Ca2+ and one Cl ion are also required for catalytic activity and appear to be located in the vicinity of the Mn cluster.2

13
The structure of PSII has recently been determined by X-ray crystallography at 3.5–3.8 Å resolution,6–8 which facilitates structure-based understanding of the PSII function.

14
The D1 subunit of the PSII reaction center is encoded by the psbA gene.

15
In higher plants and algae psbA typically exists as a single-copy gene, whereas in cyanobacteria it belongs to small gene families.9,10

16
In Synechocystis 6803 there are three psbA genes: psbA1, psbA2 and psbA.311

17
The coding regions of psbA2 and psbA3 show 99% identity and encode and identical D1 protein.12,13

18
In contrast, the psbA1 gene is largely different, and shows a 75 and 85% nucleotide and amino identity respectively, to the coding regions of psbA2 and psbA.39

19
This corresponds to a 54 amino acid difference between the polypeptides encoded by psbA1 and psbA2/psbA.314

20
Not only the amino acid sequences, but also the expression of these genes are largely different.

21
Both psbA2 and psbA3 are expressed in a light-regulated way with psbA2 responsible for more than 90% of psbA transcripts under normal light conditions.15,16

22
In contrast, no conditions have been identified under which the psbA1 gene is transcribed.9,15,16

23
However, under the influence of the psbA2 promoter the coding region of psbA1 is transcribed and the produced protein results in a functional PSII complex.17

24
It was found that of the 54 amino acid substitutions in the deduced sequence of the psbA1 gene, 8 are at positions that are otherwise conserved in all other D1 proteins.14,17

25
One of these changes, F186L, is particularly noteworthy since F186 has been implicated in the electron transfer step between Tyr-Z and P680 from computer assisted structural modeling.18

26
However, studies on site-directed mutants have demonstrated that this residue is not indispensable for PSII function.19

27
The effect of site-directed mutations of F186,20 as well as of the replacement of D1 by D1′ on PSII electron transport has been addressed in previous studies.19

28
However, the changes induced in the kinetics and energetics of charge separation and recombination has not been studied in detail.

29
Here we used flash-induced thermoluminescence and chlorophyll fluorescence yield measurements to study the functioning of PSII in strains where either the whole D1 protein was replaced with D1′ or F186 was changed to Leu and Ala.

30
Our results show that in D1′-containing PSII, the energetic stabilization of the S2QA and S2QB states was increased while in strains containing D1 with the F186 substitutions not only the donor side characteristics but also the QA to QB electron transfer step were affected.

Materials and methods

Cells

31
Synechocystis sp.

32
PCC 6803 cells were propagated in BG-11 growth medium containing 5 mM glucose in a rotary shaker at 30 °C under a 5% CO2-enriched atmosphere.

33
The intensity of white light during growth was 40 μE m−2s−1.

34
Cells in the exponential growth phase (A580 of 0.8–1) were used.

35
All site directed mutants were constructed in the psbA2 gene of Synechocystis sp.

36
PCC 6803 as described previously.15,20

37
In order to express the D1′ protein the coding region of the gene psbA1 was fused with the psbA2 promoter and was introduced into the PSII-deficient strain of Synechocystis 6803 in which the psbA2 and psbA3 genes had been inactivated as described earlier17.

Thermoluminescence

38
This was measured with a home-built apparatus as in21 in filter paper containing 50 μg Chl.

39
Dark adapted samples were excited by single turnover saturating flashes (1 Hz) at +5 or −10 °C as shown in the respective figure legends.

40
After excitation of the samples, TL was detected at a heating rate of 20 °C/min.

Flash-induced fluorescence relaxation kinetics

41
Flash-induced increases and the subsequent decay of chlorophyll fluorescence yield were measured by a double-modulation fluorometer (PSI Instruments, Brno)22 in the 150 μs to 100 s time range.

42
The sample concentration was 5 μg Chl ml−1.

43
Multicomponent deconvolution of the measured curves was done by using a fitting function with two exponential and one hyperbolic component as described earlier:23F(t) − F0 = A1exp(−t/T1) + A2exp(−t/T2) + A3/(1 + t/T3),where F(t) is the variable fluorescence yield, F0 is the basic fluorescence level before the flash, A1A3 are the amplitudes, T1T3 are the time constants.

44
The non-linear correlation between the fluorescence yield and the redox state of QA was corrected for by using the Joliot model24 with a value of 0.5 for the energy-transfer parameter between PS II units.

Results

Relaxation kinetics of flash-induced fluorescence

45
In dark-adapted samples illumination with a single saturating flash forms QA, which results in a rapid rise of variable fluorescence.

46
The subsequent relaxation of fluorescence yield reflects different reoxidation routes of QA.25–27

47
In the control AR1 cells, relaxation of the fluorescence yield after the flash is dominated by a fast component (T1 ≈ 320 μs, A1 ≈ 75%), which is related to the reoxidation of QA by QB (Fig. 1, Table 1).

48
The middle phase (T2 ≈ 4 ms, A2 ≈ 17%) arises from QA reoxidation in centers, which had an empty QB site at the time of the flash and have to bind a PQ molecule from the PQ pool.

49
Finally the slow phase (T3 ≈ 12 s, A3 ≈ 8%) reflects QAQB reoxidation through equilibrium with QAQBvia back reaction with the S2 state.

50
Replacement of D1 with D1′ induced a minor increase of T1 and T2, and a 2-fold increase in T3.

51
In all of the studied F186 mutants, the fluorescence relaxation became slower in the ms time scale and faster in the seconds time scale.

52
The initial amplitude of the fluorescence signals corresponds to the differences observed in the steady state oxygen yield.20

53
Analysis of the fluorescence curves revealed an increase of T1 (600–800 μs) and decrease of A1 (40–50%) accompanied by the decrease of T3 (2–4 s) and increase of A3 (30–37%) in the F186 mutants (Table 1).

54
When PQ binding to the QB site is blocked by DCMU the reoxidation of QA occurs via charge recombination with donor side components.

55
In cells with fully functional water-oxidizing complex the fluorescence relaxation is dominated by a hyperbolic component arising from the recombination of QA with the S2 state of the water-oxidizing complex (Fig. 2).

56
Fast decaying fluorescence can be observed even in the presence of DCMU, when the donor side is impaired by mutations28–30 or inhibited by UV-B light.23

57
This fast fluorescence decay is expected to reflect charge recombination between QA and Tyr-Z˙.23,28–30

58
Analysis of the decay curve in the presence of DMU resulted in a fast phase (T1 ≈ 2.7 ms, A1 ≈ 5%) and a slow phase (T3 ≈ 0.45 s, A3 ≈ 95%) in the AR1 cells (Table 1).

59
In the A1K strain the minor fast phase was unaffected, but T3 increased 2-fold.

60
Similar increase of T3 was induced by the F186L and F186LP162S mutations, accompanied by the decrease of A3, to 90–92%, and increase of A1, to 8–10%, (Table 1).

Thermoluminescence characteristics

61
The main thermoluminescence band, the so-called B band arises from S2QB charge recombination after single flash excitation.31,32

62
Its peak position is around 20 °C in the control AR1 strain (Fig. 3).

63
In the A1K strain the B band is upshifted to 33 °C, confirming the increased energetic stability of the S2QB charge pair indicated by the fluorescence data.

64
The F186L and F186LP162S substitutions did not increase the peak position of the B band.

65
Rather, it was decreased by 3–4 °C.

66
In contrast, in the F186A strain the TL band has a component at around 20 °C and another component at around 30 °C (Fig. 3).

67
In the presence of the electron transport inhibitor DCMU that blocks the QA to QB electron transfer step the main TL band, called Q band, arises from the S2QA charge recombination.31,33

68
In the control AR1 strain the Q band appears at around 15 °C, but upshifted to 28 °C in the A1K mutant (Fig. 4).

69
A smaller extent of upshift was observed as a result of the F186A substitution, whereas the F186L and F186LP162S replacements had practically no effect on the peak temperature of the Q band.

Discussion

Modified charge stabilization at the donor side of PSII

70
The role of the D1′ protein in Synechocystis 6803 has been enigmatic and no experimental conditions have been identified under which it was transcribed.

71
However, when placed under the regulation of the psbA2 promoter a functional D1 protein and active PSII was produced.17

72
It was previously proposed that changes in the fluorescence relaxation kinetics in the presence of DCMU could reveal if the amino acid modifications in D1′ affect the PSII donor side.19

73
Our present data demonstrate that the replacement of D1 for D1′ indeed results in marked modification of charge stabilization energies of the S2QA and S2QB charge pairs.

74
By using the time constants and relative amplitudes of the fluorescence decay curves it was possible to quantify the energetic changes induced by the replacement of D1 with D1′, which resulted in 17 (S2QB) and 24 (S2QA) meV.

75
Considering the accuracy of the method these values reflect the same extent of stabilization for the two charge pairs, which could arise either from the energetic modification of donor side (S2 state) or acceptor side components (QA and QB).

76
Since it is not very likely that both Em (QA/QA) and Em (QB/QB) would be shifted by the same value we can conclude that the stabilization most likely originates from the about 20 mV decrease of Em(S2/S1).

77
Prominent amino acid changes in D1′ are the replacement of F186 with Leu and of P162 with Ser.

78
Our data show that introduction of point mutations of F186 into D1 is sufficient to induce the stabilization of S2QA when assessed by the increased time constant of the slow phase of fluorescence decay.

79
This stabilization is not influenced by the additional mutation of P162S or by the additional amino acid changes in D1′.

80
However, from these data we cannot conclude definitely that F186L is solely responsible for the stabilization of the S2 state in D1′, since a large number of amino acid changes also occur elsewhere in the D1 sequence.

81
The fast phase of fluorescence relaxation in the presence of DCMU arises from the recombination of QA with Tyr-Z˙ in PSII centers where electron donation from the Mn cluster to Tyr-Z is blocked.

82
The small increase of A1 in the F186L and F186LP162S mutants indicates that the Mn cluster is inactive in a fraction of PSII in these mutants.

Modified charge stabilization at the acceptor side of PSII

83
An important observation of our work was that different mutations of F186 slowed down the QA to QB electron transfer, which effect appears to be related to the decreased free energy gap between the two acceptors.

84
The free energy gap can be calculated from the ratio of the time constants of the S2QA and S2QB recombinations, reflected by the T3 values of fluorescence relaxation measured in the presence and absence of DCMU, respectively.

85
An alternative method is based on the relative amplitude of the slow phase measured in the absence of DCMU, which reflects the recombination of the S2 state with QBvia the QAQB ↔ QAQB equilibrium.

86
Therefore, A3 is proportional to the equilibrium constant of sharing the electron between QA and QB,27 from which the free energy gap between QB/QB and QA/QA can be obtained.

87
These calculations resulted in 65–85 meV for the AR1 strain (Table 2) in good agreement with the about 70 meV free energy gap obtained in higher plant thylakoids by fluorescence34 and by thermoluminescence measurements.35

88
Both methods of calculation showed that the QA − QB free energy gap is only 25–30 meV in the F186 mutants, which represents 40–50 mV decrease when compared to AR1.

89
Since the S2QA charge pair is stabilized roughly to the same extent (20–30 meV) in the F186 mutants of D1 as in the A1K strain we can conclude that the Em(S2/S1) is shifted to more negative values by 20–30 mV as a consequence of replacing F186 by L or A in the normal D1 form.

90
Therefore, the decreased free energy gap between QA and QB in these mutants should be caused by a 40–50 mV decrease of Em(QB/QB).

91
Although this acceptor site region is quite far from the site of the mutations at the donor side, there are a number of previous observations in the literature that show the transmission of donor-side modifications to the QA − QB region.

92
Such examples are the different mutations of D1-D34230 and D1-H332,30 removal of the 33 kDa OEC protein36 and inactivation of the Mn cluster.37,38

93
The decreased free energy gap between QA and QB is also consistent with the slower fast phase (increased T1) of fluorescence decay in the absence of DCMU.

94
This fluorescence component represents electron transfer from QA to QB in PSII centers, which contain bound QB.

95
The decreased free energy gap corresponds to decreased driving force that leads to slower rate of the electron transfer reaction.

96
It is also important to note that the decreased Em(QB/QB) seen in the F186 mutants is almost completely compensated by the additional mutations in D1′.

97
This effect may have a physiological significance since slow QA to QB electron transfer would decrease photosynthetic efficiency and lower the chance for the survival of the cells which have the D1′ protein in PSII.

98
The increased stabilization energy of the S2QA and S2QB recombinations in the PSII centers that contain the D1′ protein was confirmed by the increased peak positions of the Q and B bands, respectively.

99
The parallel shift of the two bands supports the idea that the decreased redox potential of the S2 state is responsible for the stabilization.

100
The B band is downshifted almost to the position of the Q band in the F186L and P162SF186L mutants, which is also consistent with the destabilization of QB shown by the fluorescence results.

101
In the F186A mutant a significant part of the B band is seen at the position of the Q band in the absence of DCMU, which is in agreement with the retarded QA to QB electron transfer shown by the fluorescence data.

102
It is of note, however, that the peak position of the Q band is somewhat lower in the F186L and P162SF186L mutants than in the AR1 strain although the increased time constant of fluorescence decay indicates almost the same extent of stabilization for S2QA as for seen in the A1K strain (Table 2), whose peak temperature is almost 20 °C higher.

103
The reason for this effect is not fully understood at present, but might be related to the modification of non-radiative recombination pathways that influence the TL characteristics39,40.

The function of F186 in PSII

104
From the positioning of the aromatic ring of F186, predicted from computer assisted structural modeling, it was proposed that this residue might facilitate electron transfer between Tyr-Z and P680.18

105
Although the recent crystal structure of PSII8 confirms that F186 is indeed located halfway between Tyr-Z and P680, previous studies on site directed mutants have shown that this residue is not absolutely required for donor side electron transfer.19,20

106
Several mutations of F186 are photoautrophic and able to evolve oxygen.20

107
However, the flash pattern of the oxygen yield is largely distorted and the probability of misses increased.19

108
This effect could partly be explained by the retarded QA to QB electron transfer, which facilitates S2QA recombination, and partly by donor-side induced modifications in the F186 mutants.

109
Such donor side effect has already been suggested by Funk et al19. and was demonstrated by our results.

110
From the present and previous data it is obvious that F186 is not indispensable for PSII activity.

111
However, it is involved in the regulation of charge stabilization in the water-oxidizing complex, and, surprisingly, in the regulation of efficient electron transfer between QA and QB.

The enigmatic role of the D1′ protein

112
The utilization of D1 protein copies with different amino acid sequences and differently regulated promoters appears to be a common strategy of cyanobacteria to cope with changing environmental conditions.

113
A well studied example is Synechococcus 7942 that encodes two different D1 polypeptides (D1:1 and D1:2) by three psbA genes.10

114
D1:1 is replaced by D1:2 under stress induced by high light,41,42 cold43 and UV-B radiation44 as an important part of the defence/acclimation process.

115
In Synechocystis 6803 the transcription of the psbA2 and psbA3 genes, which encode identical D1, is also regulated differentially under high light9,15 and UV radiation,45,46 helping to increase the psbA transcript level under conditions that require rapid PSII repair.

116
Although the sequence of D1′ does not show higher similarity to the stress-inducible D1:2 form of Synechococcus 7942 than to the other cyanobacterial D1 proteins it seems likely that the original function of D1′ was also related to environmental adaptation.

117
The apparent silence of the psbA1 gene most likely indicates that the putative environmental conditions, which required the presence of D1′ did not represent significant evolutionary pressure in the recent history of Synechococystis 6803 that led to the loss of the psbA1 promoter activity.

118
However, the fact that D1′ containing PSII centers are fully functional despite the large number of amino acid substitutions might indicate that under special conditions the D1′ protein could be expressed.

119
This idea is strengthened by our observation that slow electron transfer between QA and QB in the F186L mutant is fully restored in the D1′ containing PSII centers, indicating that the unfavorable effect of the F186L substitution on the acceptor side electron transport is compensated by the other amino acid substitutions of D1′ making the D1′ containing cells photosynthetically efficient.