NdhV subunit regulates the activity of type-1 NAD(P)H dehydrogenase under high light conditions in cyanobacterium Synechocystis sp. PCC 6803

The cyanobacterial NAD(P)H dehydrogenase (NDH-1) complexes play crucial roles in variety of bioenergetic reactions. However, the regulative mechanism of NDH-1 under stressed conditions is still unclear. In this study, we detected that the NDH-1 activity is partially impaired, but the accumulation of NDH-1 complexes was little affected in the NdhV deleted mutant (ΔndhV) at low light in cyanobacterium Synechocystis sp. PCC 6803. ΔndhV grew normally at low light but slowly at high light under inorganic carbon limitation conditions (low pH or low CO2), meanwhile the activity of CO2 uptake was evidently lowered than wild type even at pH 8.0. The accumulation of NdhV in thylakoids strictly relies on the presence of the hydrophilic subcomplex of NDH-1. Furthermore, NdhV was co-located with hydrophilic subunits of NDH-1 loosely associated with the NDH-1L, NDH-1MS′ and NDH-1M complexes. The level of the NdhV was significantly increased at high light and deletion of NdhV suppressed the up-regulation of NDH-1 activity, causing the lowered the photosynthetic oxygen evolution at pH 6.5 and high light. These data indicate that NdhV is an intrinsic subunit of hydrophilic subcomplex of NDH-1, required for efficient operation of cyclic electron transport around photosystem I and CO2 uptake at high lights.

uptake 8 . In contrast to the crystal structure of Complex I 9 , NDH-1 from cyanobacteria is speculated to possess an oxygenic photosynthesis-specific (OPS) domain 10 comprised of NdhL-NdhO identified in Synechocystis 6803 11,12 . Several NDH subunits function in stabilization of NDH-1. NdhP and NdhQ were found in the purified NDH-1L complex from T. elongatus 13 . NdhP is involved in the respiratory and cyclic electron flow 14 and stabilization of the NDH-1L complex 15 . NdhQ is also essential for stabilization of the large complex of NDH-1 16 . NdhS participates in the activity of cyclic electron flow around PSI in Arabidopsis 17 or in cyanobacteria 18 , and serves as the Fd docking site domain, accepting electrons from Fd in chloroplasts 19 . Furthermore, He et al. 20 have found that the NDH-1L complex interacts with Fd via the subunit NdhS in Thermosynechococcus elongatus.
Cyanobacteria utilize both CO 2 and HCO 3 − as carbon species 21 . The dissolved inorganic carbon is pH-dependent. With increasing pH, the ratio of HCO 3 − to CO 2 continues to rise 22 . In contrast, the water with acid pH is favorable for providing a sufficient CO 2 supply. Cyanobacteria possess a CO 2 -concentrating mechanism (CCM) that enables to raise the concentration of inorganic carbon (Ci, HCO 3 − and CO 2 ) at the carboxylation site to a high level for efficient CO 2 fixation despite the low affinity of their Rubisco for CO 2 23,24 . To date, five inorganic transporters have been found including two Na + -dependent HCO 3 − transporters (BicA and SbtA), one ATPase-dependent HCO 3 − transporter (BCT1), and two CO 2 -uptake NDH-1 complexes in Synechocystis 6803 and other cyanobacterial strains [24][25][26] . As one of CO 2 uptake systems, NDH-1MS complex, consisting of NdhD3, NdhF3 and CupA (ChyY), is inducible at limiting Ci conditions and has a higher uptake affinity for CO 2 [27][28][29] . Further research showed that the proteins encoded by NdhF3/NdhD3/CupA/CupS formed a small complex NDH-1S, in which CupA and a small protein CupS were identified as subunits of cyanobacterial NDH-1 by proteomic analysis 30,31 . NDH-1S and NDH-1M form an NDH-1MS complex which has been isolated from a Thermosynechococcus elongatus strain in which the C terminus of NdhL has been tagged with 6-His. This complex is easily dissociated into NDH-1M and NDH-1S complexes 31 . As a homologous gene of cupA, cupB (chyX) is involved in constitutive CO 2 uptake system encoded by NdhD4/NdhF4/CupB which formed a small complex NDH-1S′ 28,29 . Based on the purification of a 450 kDa complex contained both NdhH and CupB protein, it has been suggested that the complex is NDH-1MS′ located in the thylakoid membranes 32 .
Previous studies have shown that the cyclic electron flow around PS I mediated by NDH functions in protection against stressed conditions. Ogawa 2 has found that deletion of NdhB gene resulted in the unable survival phenotype of Synechocystis 6803 under low CO 2 (air) condition, suggesting that the NDH-1 functions in inorganic carbon transport. The activity of cyclic electron flow around PSI mediated by NDH-1 has been found to be enhanced under stressed conditions, including low CO 2 33 , strong light 34 . However, how NDH-1 participates in the regulation of photosynthesis is still unclear.
Recently, NdhV is reported to function in stabilization of chloroplast NDH-like complex in Arabidopsis 35 and heat tolerance in Synechocystis 6803 36 . However, the mechanism of NdhV in response to stressed conditions is not clear. Here we report that cyanobacterial NdhV localizes in the hydrophilic subcomplex of NDH-1MS′ , NDH-1L and NDH-1M. Further analysis showed that NdhV functions in regulation of NDH-1 activity for efficient operation of cyclic electron flow around PS I and CO 2 uptake in Synechocystis 6803 in response to high lights.

Results
Deletion of NdhV partially arrests NDH-1 activity. To reveal the function of NdhV in cyanobacteria, we inactivated the cyanobacterial NdhV protein by inserting a kanamycin resistance (kan R ) cassette into its coding region (Fig. 1A). The PCR analysis of the NdhV locus using the ndhV-up-F and ndhV-Dn-R primers showed a band of 1.5 Kbp in the WT and 2.7 Kbp in the ΔndhV mutant which contained an extra Kanamycin resistance cassette of about 1.2 Kbp; it confirmed the complete segregation of the ΔndhV allele in this mutant (Fig. 1B).
Post-illumination increase in Chl fluorescence was explained as the reduction of plastoquinone (PQ) by the electrons from photoreductants accumulated in the stroma or cytosol during illumination, which reflects cyclic electron transport around PSI mediated by NDH-1 37,38 in cyanobacteria and by chloroplast NDH in higher plants 39 . By comparison with WT, the post-illumination increase in Chl fluorescence was partly impaired in two lines of ndhV deleted mutants ΔndhV-1 and ΔndhV-2 (Fig. 1C), suggesting that NdhV contributes to the NDH-1 activity.
Growth of ΔndhV was suppressed under high light and low pH conditions. In addition to cyclic electron flow around PSI, NDH-1 is also involved in CO 2 uptake 2 . To examine whether NdhV participates in CO 2 uptake, we compared the growth phenotype between WT and ΔndhV under various conditions, low (40 μ mol photons m −2 s −1 ) or high light (300 μ mol photons m −2 s −1 ), low (below pH 7.5), or high pH (above pH 7.5), high (2%) or low CO 2 (0.04%). With the pH increasing to higher than 7.0, the concentration of HCO 3 − is predominant but that of CO 2 becomes minor, while at the low pH below 6.5 condition, the concentration of HCO 3 − becomes minor and that of CO 2 is predominant. The growth phenotype was almost the same under pH 8.0 either at low light or high light ( Fig. 2A-D). The growth of ΔndhV was slightly slow at low pH, low light, low CO 2 (Fig. 2B) or high light, high CO 2 (Fig. 2C) and more evidently under high light, low CO 2 (Fig. 2D). To confirm the difference, we further compared the growth rate in the liquid culture conditions bubbling with 2% CO 2 at pH 6.5 at low light or high light. The growth rate was identical between ΔndhV and wild type under low light condition in liquid (Fig. 2E), consistent with the results on agar ( Fig. 2A). However, the growth of ΔndhV was slower than WT under the high light condition (Fig. 2F), suggesting that NdhV plays an important role under the stressed conditions. In contract, the ndhB defective mutant (M55) in which both the PSI-cyclic electron flow and CO 2 uptake were inactivated, could hardly grow up at pH 6.5 under either low light or high light conditions (Fig. 2E,F), similar to the observation by Ogawa (1991 Suppression of the activity of CO 2 uptake in ΔndhV. To confirm whether the CO 2 uptake activity was affected by deletion of NdhV, we compared the rate of CO 2 uptake in WT, ΔndhV, and M55 in which only NDH-1S is active in the CO 2 acquisitive systems, using a portable open-flow gas exchange system under high lights and high or low CO 2 conditions. Cell suspensions were placed on the BG-11 agar plate to measure the CO 2 uptake rate. The rate of CO 2 uptake was lowered by about 10% in ΔndhV than that in WT at 100 μ mol photons m −2 s −1 either under high CO 2 or low CO 2 conditions (Fig. 3A), and at 300 μ mol photons m −2 s −1 under low CO 2 , but not under high CO 2 condition (Fig. 3B). In contrast, the rate of CO 2 uptake was almost completely suppressed in M55 (Fig. 3A,B) suggesting the loss of function of both NDH-1MS and NDH-1MS′ in the mutant. These results demonstrated that deletion of NdhV suppresses the CO 2 uptake activity under high lights and CO 2 limitation conditions. The proton gradient across thylakoid membranes was lowered in ΔndhV. It has been suggested that cyanobacterial NDH-1 provides ATP for CO 2 uptake 2 . To investigate whether NdhV is involved in this process, the proton gradient across thylakoid membranes, which drives ATPase to synthesis ATP was compared between WT and ΔndhV. Quinacrine (QA) fluorescence quenching can be used for the determination of ΔpH across the thylakoid membrane for intact cells of Synechocystis 6803 40 . The QA fluorescence quenching was decreased by about 20% in ΔndhV mutants and 40% in M55 compared with those in WT (Fig. 3C), suggesting that NdhV is involved in regulation of proton gradient across thylakoid membrane.
NdhV localizes to the thylakoid membrane and its accumulation relies on the electron donor domain of NDH-1M of NDH-1MS′, NDH-1L. The previous study demonstrated that Arabidopsis NdhV is mainly localized to the thylakoid membrane 35,36 . As a homologous protein, the localization of the cyanobacterial NdhV was investigated with a polyclonal antibody against the recombinant cyanobacterial NdhV. Western blotting analysis detected a band with a molecular mass of about 15 kDa (the theoretical molecular mass of mature NdhV is ~18 kDa) in the thylakoid and also in the stromal fractions of WT, but absent in those of ΔndhV (Fig. 4A), confirming that the cyanobacterial NdhV is mainly localized to the thylakoid membrane.
In cyanobacteria, a total of 4 NdhD isoforms are present and they are located in different NDH-1 complexes. Given that NDH-1L, NDH-1MS, NDH-1MS′ were respectively defective in ΔndhD1/D2, ΔndhD3 and ΔndhD4 mutant in Synechocystis 6803 30 , suggesting that NDH-1M is stable in those NdhD deleted mutants, we checked the amount of NdhV in different NdhD deleted mutants. As shown in Fig. 4B, there was no much difference in the amount of NdhV among WT and the various NdhD deleted mutants, suggesting that NdhV is present in NDH-1M complex.  (Table S1). (C) Monitoring of NDH-1 activity using chlorophyll fluorescence analysis. The top curve shows a typical trace of chlorophyll fluorescence in the WT of Synechocystis 6803. The cells (OD 730 around 0.4) supplemented with 10 mM NaHCO 3 were used for the measurement. After the sample was exposed to the actinic light (AL, 100 μ mol photons m −2 s −1 ) for 90 s, AL was turned off, and the transient increase in chlorophyll fluorescence level was recorded, which was used to ascribe NDH-1 activity. The inset shows magnified traces from the box area.
Scientific RepoRts | 6:28361 | DOI: 10.1038/srep28361 To further confirm the association of NdhV with NDH-1M, the accumulation of NdhV in the mutants defective in different Ndh subunits was investigated by immunoblot. Figure 4C shows that NdhV was detected neither in NdhB defective mutant M55, in which both the NDH-1L and NDH-1M were disassembled 30 nor in NdhM deleted mutant ΔndhM in which all the NDH-1 complexes were disassembled 41 . The results indicate that the accumulation of NdhV relies on NDH-1M and the complexes containing NDH-1M, including NDH-1L, NDH-1MS and NDH-1MS′ .
We further checked the accumulation of cyanobacterial NdhV in NdhS deleted mutant of Synechocystis 6803 (ΔndhS). NdhV was hardly detected in ΔndhS (Fig. 4C), consistent with the result obtained in Arabidopsis ndhS/ crr31 mutant (Fan et al.) 35 or in Synechocystis 6803 (Gao et al.) 36 , suggesting that the stability of NdhV is dependent on the electron donor domain of NDH-1 in cyanobacteria. Furthermore, we checked the accumulation of NdhV in the mutant defective in the hydrophilic subunits of NDH-1. The amount of NdhV was reduced significantly in NdhI partly deleted mutant (ΔndhI U ) (Fig. 4C). The result indicates that the accumulation of NdhV is dependent on the presence of the hydrophilic subcomplex of NDH-1related to its electron donor domain.
NdhV is loosely bound in NDH-1 complexes and its absence hardly affects the accumulation of other Ndh subunits and assembly of NDH-1 complexes under normal growth conditions. To investigate whether the accumulation of the NDH-1 complexes is affected in ΔndhV, thylakoid protein complexes isolated from ΔndhV and WT were separated by a 5-13% gradient blue-native PAGE (BN-PAGE) followed by 2D SDS-PAGE for the immunoblotting analysis. The results show that in the wild type, NdhV presents as a free form and the bands corresponding to NDH-1L and NDH-1M complexes were detected using the antibody of NdhK, and the bands corresponding to NDH-1MS′ and NDH-1S′ complex could also be detected using the antibody of the key component CupB (Fig. 5A). However, in ΔndhV, these bands were almost the same as those in wild type, indicating that deletion of NdhV hardly affected the assembly of NDH-1 complexes, including NDH-1L, NDH-1MS′ and NDH-1M (Fig. 5A), in accordance with the results in Arabidopsis that the NdhV is a protein easily dissociated from NDH-1 complexes during sample preparation and BN-gel electrophoresis. To overcome this limitation, we performed the chemical crosslinking experiments using the thylakoids isolated from WT. The result shows that NdhV co-migrates with the NdhI and NdhK subunits in the NDH-1M, NDH-1L, NDH-1MS′ and a supercomplex around 1000 kDa, probably due to the aggregation of the NDH-1 (Fig. 5B). These results also confirm that NdhV is a subunit of cyanobacterial NDH-1 but easily dissociated from NDH-1 complexes. To know whether NdhV affects the accumulation of other subunits of NDH-1, we checked the amount of several subunits in different parts of cells from ΔndhV. The results show that there were no obvious difference of the amount of NdhH, I, K, M either in total proteins, or in the supernatant or in the thylakoid membrane between the WT and ΔndhV cultured at pH 6.5 (Fig. 1SA) or at pH 8.0 (Fig. 1SB).
The up-regulation of NDH-1 activity at high lights was suppressed in ΔndhV. Previous studies have indicated that NDH-1 activity was up-regulated by high light 34 . To test whether NdhV is required for up-regulation of NDH activity under high light conditions, we compared the NDH-1 activity under different light intensities between WT and ΔndhV. Figure 6A shows that the NDH-1 activity, reflected by the transient increase in Chl fluorescence after termination of actinic light, increased with the increase of the intensity of actinic light in WT, but the up-regulation of NDH-1 activity in response to high lights was significantly suppressed in ΔndhV. The results indicate that the up-regulation of NDH-1 activity is almost suppressed when NdhV is absent. Furthermore, we compared the changes of the amount of Ndh subunits after adaptation to the high light between WT and ΔndhV. As shown in Fig. 6B, after the growth light was shifted to the high light (200 μ mol photons m −2 s −1 ) for 1 h, the amounts of NdhH and NdhK were evidently increased by about 40% and 35%, respectively, both in WT and ΔndhV. Meanwhile, the amount of NdhV was approximately 2-fould higher at the The quenching of QA fluorescence was induced by adding the cells sample to 2 ml reaction mixture after the background fluorescence reached steady state about 2 min after started the measurement. The QA fluorescence quenching of WT is 5.02%. Values are means ± SE of three independent measurements. Asterisk indicates significant differences (t-test, * P < 0.05 and * * P < 0.01).
high light in WT. The results suggest that the up-regulation of NdhV is required for regulation of NDH activity in response to high lights in cyanobacteria.
The capacity of photosynthesis was suppressed at the high light and low pH in ΔndhV. To check whether NdhV functions in protecting the cyanobacterium against the high light stress, we compared the capacity of photosynthesis between WT and ΔndhV in response to the high light, as reflected by photosynthetic O 2 evolution at higher pH (8.0) where the HCO 3 − is predominant and lower pH (6.5) where CO 2 is predominant. With the increase in the light intensity, the rate of photosynthetic O 2 evolution increased, but there was not much difference between WT and ΔndhV at pH 8.0 (Fig. 7A). However, the rate of photosynthetic O 2 evolution in ΔndhV was approximately 10% lower than that in WT at pH 6.5 at higher light intensities (Fig. 7B). The results suggest that NdhV is also required for photosynthetic CO 2 assimilation under the high light and low pH stressed conditions.

Discussion
It has been shown that mutation of NdhV caused completely the loss of the NDH activity in Arabidopsis 35 but only partly in cyanobacterium Synechocystis 6803 (Gao et al.) 16 , which is confirmed in our work (Fig. 1). The only partial suppression of the NDH-1 activity (Fig. 1C) by deletion of the cyanobacterial NdhV is probably attributed to the compensation of the activity from different types of NDH-1 in cyanobacteria 7 . On the other hand, similar with the chloroplast NdhV, the cyanobacterial NdhV is also demonstrated to localize in the fragile part of NDH-1 and may interact with the electron donor domain of NDH-1 by the observation of the severe suppression of the accumulation of NdhV in ΔndhS (Fig. 4C). In addition, the accumulation of NdhV was almost lost in the thylakoid membranes in M55 and in ΔNdhM (Fig. 4C) in which the hydrophilic subcomplex of NDH-1 complexes are degraded (He et al.) 20 , but it was not affected either in NdhD3/D4 defective mutant (Fig. 4B) in which only NDH-1L and NDH-1M exist, or in NdhD1/D2 defective mutant (Fig. 4B) in which NDH-1M, NDH-1MS′ are still present. Moreover, NdhV co-localized with Ndh subunits of NDH-1MS′ , NDH-1L and NDH-1M complexes (Fig. 5A). Considering the cells were cultured under high CO 2 culture condition, NDH-1MS would not be visible. Those results let us conclude that NdhV interacts with the electron donor domain of these complexes which contains NDH-1M as a skeleton 30 . Although defect in NdhS and other hydrophilic subunits affected the accumulation of NdhV (Fig. 4C), but deletion of NdhV didn't affect the accumulation of relevant NDH-1 subunits   nor the assembly of NDH-1 complexes (Fig. 1S), implying that NdhV is a peripheral subunit functioning in regulation of NDH-1.
The growth phenotype of both wild type and ∆ndhV cultured in liquid medium bubbling with 2% CO 2 at pH 6.5 and low light was almost the same (Fig. 2E), which was consistent with the previous observation (Gao et al.) 16 who therefore concluded that NdhV didn't play any role in CO 2 uptake. However, when ∆ndhV was cultured in agar plate in 2% CO 2 at pH 6.5 and low light, the growth was slightly suppressed (Fig. 2B), suggesting that the bubbling may increase the diffusion of HCO 3 − into the carboxysome even if it was not predominant, but the ratio of CO 2 to HCO 3 − to is still about 0.7 at pH6.5. In addition, the growth of ∆ndhV was evidently suppressed at high light when HCO 3 − or CO 2 were limited at low pH or low CO 2 (Fig. 2C,D,F) but not significantly at high light when the inorganic carbon is saturated at high pH or high CO 2 (Fig. 2C,D). The results suggest that NdhV is required for the efficient CO 2 uptake at high light especially under the conditions when the inorganic carbon is limited, which is further supported by the observation that the CO 2 uptake activity was evidently suppressed at the high light of 300 μ mol photons m −2 s −1 under low CO 2 but not under high CO 2 condition (Fig. 2C) and also by the suppression of photosynthetic O 2 evolution at the high light at low pH ( Fig. 7B) but not at high pH (Fig. 7A). Defect in the cyanobacterial NdhV caused partially (about 10%) suppression of the rate of CO 2 uptake (Fig. 3A), it might be attributed to the partly contribution of NdhV to the cyclic electron or respiratory electron transports which might function in providing transthylakoid membrane proton gradient (Fig. 3C) for ATP synthesis or alkaline pocket (Kaplan and Reinhold, 1999).
The previous study showed that both the cyclic electron flow around PSI and respiratory O 2 uptake mediated by NDH-1 was enhanced at high light (100 μ mol photons m −2 s −1 ), accordingly the photosynthetic O 2 evolution was also up-regulated 34 , suggesting that the up-regulation of NDH-1 is required for high activity of photosynthesis. The up-regulation of the NDH-1 activity induced by the high lights was suppressed when NdhV is deleted (Fig. 6A), suggesting that the cyanobacterial NdhV is involved in the regulation of NDH-1 activity in response to the high lights. The significant up-regulation of NdhV by the treatment of the high light in the wild type (Fig. 6B) implies its important role in the regulation of photosynthesis, as one of components related with the electron donor domain of NDH-1. The evident suppression of the rate of CO 2 uptake at the high light (100 μ mol photons m −2 s −1 ) either at high CO 2 or low CO 2 (Fig. 3A) further supports this mention. In addition, the suppression of the photosynthetic oxygen evolution (Fig. 7B) and the growth (Fig. 2D,F) at the high light when CO 2 was limited in the NdhV deleted mutant might be attributed to the loss of the up-regulation for NDH-1 activity under the stressed conditions (Fig. 6A). It has been indicated that cyclic electron flow around PSI plays an important role in providing ATP for carbon assimilation 42,43 . Actually, the enhancement of cyclic electron flow around PS I by treatment with low concentration of NaHSO 3 enables cyanobacteria to generate sufficient proton gradient across the thylakoid membrane, thereby increased the biomass of Synechocystis PCC 6803 44 . Therefore, in NdhV deleted mutant, the partial suppression of NDH-1 activity (Fig. 1C) would lower the building-up of proton gradient across thylakoid membranes (Fig. 3C), resulting in the decrease in the activity of CO 2 uptake (Fig. 3) and suppressed growth, especially when HCO 3 − and CO 2 were limited under high light conditions (Fig. 2D). In conclusion, our work demonstrate that the cyanobacterial NdhV subunit locates at the hydrophilic parts of NDH-1M and plays an essential role in regulation of NDH-1 activity for efficient operation of cyclic electron flow around PS I and CO 2 uptake especially at high light.

Materials and Methods
Cell culture conditions. The Synechocystis 6803 wild type and mutant cells were cultured in BG-11 medium 45 buffered with Tris-HCl (5 mM, pH 8.0) at 30 °C under 50 μ mol photons m −2 s −1 . The cells were bubbled with 2% v/v CO 2 in air. The BG-11 solid medium used was BG-11 supplemented with 1.5% agar, Tris-HCl (10 mM, pH 8.0), and 0.3% Na 2 S 2 O 3 . The illumination was provided by fluorescence lamps at 50 μ mol photons m −2 s −1 . According to culturing requirements of wild type and mutant cells, the BG-11 medium was buffered with 10 mM TES-KOH at several other pH conditions.
Construction of ΔndhV mutant. The upstream and downstream regions of sll0272 (NdhV) and a DNA fragment encoding a kanamycin resistance (Kan R ) cassette were firstly amplified by PCR using specific oligonucleotide primers (Table S1). Then the PCR products of upstream of sll0272 and Kanamycin resistance (Kan R ) cassette were used as templates to obtain the second PCR products(SP1), and the PCR products of downstream of sll0272 and Kan R cassette were used as templates to obtain the second PCR products(SP2). The third round PCR product containing the upstream and downstream regions inserted by a Kan R cassette were amplified by PCR using SP1 and SP2 as templates. Then the final products were ligated to the pMD19T (Fig. 1A), to make the construct for transforming the WT cells of Synechocystis 6803 as described by Williams and Szalay 46 . The transformants were spread on agar plates containing BG-11 medium and Kanamycin (10 μ g.ml −1 ) buffered at pH 8.0, the BG-11 solid medium plates were incubated in 2% v/v CO 2 in air. The mutated NdhV in the transformants was segregated to homogeneity as determined by PCR amplification and immunoblotting.
The measurement of chlorophyll fluorescence. The transient increase in chlorophyll fluorescence after actinic light had been turned off was monitored by means of using a PAM Chl fluorometer (Walz, Effeltrich, Germany), emitter-detector-cuvette assembly (ED-101US) and unit 101ED as previously described 37,38 . Oxygen exchange. The evolution of oxygen under different light intensities were determined in cultural BG-11 medium that contained the mid-logarithmic cell of Synechocystis 6803 at about 2.5 μ g chlorophyll ml −1 with a Clark-type oxygen electron. The whole monitoring process was performed at 30 °C.

CO 2 uptake measurement. CO 2 uptake was measured with a portable open-flow gas exchange system
Li-6400 (LI-COR Biosciences). Air temperature of the leaf chamber was maintained at 30 °C, the photosynthetically active radiation (PAR) was 100 or 300 μ mol photons m −2 s −1 and the flow rate of the air in the measuring chamber was 50 μ mol s −1 . CO 2 concentration was controlled at 2% (v/v in air) or 0.04%. The cells of WT and ΔndhV strains grown at 30 °C in BG-11 medium at 2% CO 2 were collected at mid-logarithmic stage by centrifugation and re-suspended in fresh BG-11 medium. They were concentrated into 100 of OD 730 , and 30 μ l of each cell suspension was placed on the agar plate as a spot incubated in the growth chamber before CO 2 uptake measurement. Each cell spot on agar was cut out into a square of 1 cm × 1 cm and put on a cover glass to measure the CO 2 uptake rate. Same size of agar was used as control. Measurements were repeated three times and the averages were recorded.
Isolation of crude thylakoid membranes. Synechocystis 6803 thylakoid membranes were isolated as described by Gombos et al. 47 with some modifications as follows. The cell cultures (1L) were centrifuged at the logarithmic phase and then resuspended in 20 ml disruption buffer [10 mM HEPES-NaOH, 5 mM sodium phosphate, pH 7.5, 10 mM MgCl 2 , 10 mM NaCl, and 20% v/v glycerol)], broken by vortexing with glass beads (150-212 μ m) for three times at 70 Hz for 30 s with 30 s interval cooling using a Tissuelyser-48 system (Shanghi Jingxin). And then the lysis was centrifuged at 5000 g for 5 min at 4 °C to remove glass beads and unbroken cells. The crude thylakoid membranes and supernatant were obtained by centrifugation of the homogenate at 20 000 g for 30  Cross-linking of the thylakoid membranes. The cross-linking assay was performed as described previously 48 . Thylakoid membranes were resuspended in 1 ml of 20 mM Hepes-KOH, pH 8.0, 5 mM MgCl 2 (0.5 mg chlorophyllml −1 ) and incubated with 2.5 mM DSP (Dithiobis[succinimidyl propionate]) for 30 min at room temperature in the dark. After cross-linking, reactions were quenched with addition of 60 mM Tris-HCl, pH 7.5 for 15 min. Thylakoids were pelleted after centrifuged again at 20 000 g for 30 min at 4 °C and solubilized at 0.5 mg of chlorophyll ml −1 in a buffer containing 20 mM Hepes-KOH, pH 8.0, 200 mM NaCl, 1.2% Triton X-100, 1 mM PMSF for 30 min on ice. Insoluble thylakoids were removed by centrifugation at 22 000 g for 10 min at 4 °C and the supernatant was transferred to new tubes for further analysis.
Electrophoresis and immunoblotting. BN-PAGE of Synechocystis 6803 membranes was performed as described previously 48 with slight modifications. The chlorophyll a concentration of the membranes solubilized in buffer C was diluted at 0.5 mg ml −1 , then the 1/20 vloume of 20% n-dodecyl-β -D-maltoside (DM) was added. After incubation on ice for 20 min and centrifugation at 20 000× g for another 15 min, the supernatants were supplemented with 1/10 volume of BN sample buffer (5% Serva Blue G, 100 mM BisTris-HCl, pH 7.0, 30% w/v sucrose, 500 mM ε -amino-n-caproic acid and 10 mM EDTA). Solubilized membranes were then applied to a 1.0 mm thick 5-13% acrylamide gradient gel. Samples of 5 μ g chlorophyll a were loaded on the gel. Electrophoresis was performed at 4 °C by increasing the voltage gradually from 50-200 V during the 5 h run. For electrophoresis in the second dimension, the lanes of the BN gel were cut out and incubated in 1× SDS sample loading buffer containing 2% β -mercaptoethanol for 30 min. SDS-PAGE of the lanes of the BN gel was performed on a 15% polyacrylamide gel as described previously 49 .
For immunoblotting, the proteins were electrotransferred to a polyvinylidene difluoride (PVDF) membrane (Immobilon-P; Millipore) and detected using protein-specific antibodies with the ECL assay kit (BioRad) according to the manufacturer's protocol. Antibody against the NdhV protein of Synechocystis 6803 was raised in our laboratory. To amplify the NdhV gene, primer sequences were listed in Table S1. The PCR product was ligated into vector pET51b (Novagen). The plasmid was used to transform E.coli strain BL21 (DE3) pLysS for expression. Polyclonal antibody was raised in a rabbit from purified recombinant protein. The antibodies against CupB, NdhH, NdhI, NdhK, and NdhM were previously raised in our laboratory. Quinacrine (QA) fluorescence quenching. Fluorescence of Quinacrine (QA) at 503 nm using the PAM chlorophyll fluorometer (Maxi-version, Walz, Effeltrich, Germany) attached with a US-370 emitter with an emission peak at 375 nm and a PM-101/D detector as described previously 50 . Cells were harvested at logarithmic phase and suspended in reaction mixture contained 5 mM Tris/MES (pH 8.0), 0.3 M mannitol, 2 mM DTT, 5 mM D-Glucose, 1.5 μ M ATP, 2.5 μ M QA with a final chlorophyll concentration of 150 μ g/ml. The quenching of QA fluorescence was induced by adding the cells sample to 2 ml reaction mixture after the background fluorescence became stable about 2 min after started the measurement.