Passive Entry of CO2 and Its Energy-dependent Intracellular Conversion to HCO 3 − in Cyanobacteria Are Driven by a Photosystem I-generated ΔμH+ *

CO2 entry intoSynechococcus sp. PCC7942 cells was drastically inhibited by the water channel blocker p-chloromercuriphenylsulfonic acid suggesting that CO2 uptake is, for the most part, passive via aquaporins with subsequent energy-dependent conversion to HCO 3 − . Dependence of CO2uptake on photosynthetic electron transport via photosystem I (PSI) was confirmed by experiments with electron transport inhibitors, electron donors and acceptors, and a mutant lacking PSI activity. CO2 uptake was drastically inhibited by the uncouplers carbonyl cyanide m-chlorophenylhydrazone (CCCP) and ammonia but substantially less so by the inhibitors of ATP formation arsenate and N, N,-dicyclohexylcarbodiimide (DCCD). Thus a ΔμH+ generated by photosynthetic PSI electron transport apparently serves as the direct source of energy for CO2 uptake. Under low light intensity, the rate of CO2 uptake by a high-CO2-requiring mutant ofSynechococcus sp. PCC7942, at a CO2concentration below its threshold for CO2 fixation, was higher than that of the wild type. At saturating light intensity, net CO2 uptake was similar in the wild type and in the mutant IL-3 suggesting common limitation by the rate of conversion of CO2 to HCO 3 − . These findings are consistent with a model postulating that electron transport-dependent formation of alkaline domains on the thylakoid membrane energizes intracellular conversion of CO2 to HCO 3 − .

On illumination, many photosynthetic microorganisms maintain the concentration of dissolved CO 2 ([CO 2(dis) ]) in their surrounding medium below that expected at chemical equilibrium with HCO 3 Ϫ (1)(2)(3)(4)(5). This displacement of [CO 2(dis) ] from equilibrium can be observed in the absence of CO 2 fixation and is largely due to CO 2 uptake, intracellular conversion to HCO 3 Ϫ , and release of the latter into the medium (5,6). The reverse phenomenon has been described in Synechococcus WH 7803 (7) and Nannochloropsis sp. (8,9) where HCO 3 Ϫ uptake, internal conversion to CO 2 , and efflux of the latter result in elevated [CO 2(dis) ] in the medium. HCO 3 Ϫ transport systems, in Cyanobacteria, are probably located at the cytoplasmic membrane and are believed to be driven by ATP either directly (10) or possibly indirectly (11)(12)(13). CO 2 uptake has been observed to result in HCO 3 Ϫ accumulation in the cytoplasm where [CO 2(dis) ] is maintained below that expected at chemical equilibrium, and it has been inferred that a CA 1 -like activity is involved in its uptake and intracellular conversion to HCO 3 Ϫ (6, 14 -17). The location of the CAlike activity has not been identified and the mode of energization of the active HCO 3 Ϫ accumulation is not understood. Active transport of CO 2 across the plasmalemma has also been suggested (5,18), but it is difficult to distinguish this from diffusion of CO 2 across the plasma membrane with subsequent energy-dependent conversion to HCO 3 Ϫ (6,19). Passive entry of CO 2 across the membrane may occur via aquaporins (20), a possibility examined here by the application of a water channel blocker (WCB), p-chloromercuriphenylsulfonic acid. Use of this WCB prevented changes in cell volume and inactivation of PSI and PSII following osmotic stress in Synechococcus sp. strain PCC7942 (21).
Until recently, it was widely accepted that cyclic PSI activity plays the major role in energization of CO 2 uptake in Cyanobacteria (22,23). Some recent observations appear to conflict with this conclusion. In the ⌬ndhD1/D2 mutant of Synechocystis sp. strain PCC6803 lacking components of NAD(P)H dehydrogenase (NDH-1), oxidation of P700 was depressed, but CO 2 uptake was only slightly affected (24). On the other hand, in the ⌬ndhD3/D4 mutant (25) and the ⌬ndhD3 mutant of Synechococcus sp. strain PCC7002 (26) CO 2 uptake was drastically depressed with only a small effect on P700 oxidation (i.e. PSI cyclic electron transport, ET). We have therefore reexamined the involvement of PSI in CO 2 uptake and suggest how the former data may be reconciled.
We have recently suggested a working hypothesis according to which CO 2 uptake by Cyanobacteria and its intracellular conversion to HCO 3 Ϫ may be energized by photosynthetic electron transport via the formation of alkaline domains on the stromal face of the thylakoid membrane (6). Catalyzed conversion of CO 2 to HCO 3 Ϫ in these domains would maintain an inward diffusion gradient for CO 2 . Results presented are con-sistent with a prediction of this model, i.e. that CO 2 uptake would depend on a ⌬H ϩ rather than on ATP hydrolysis.

MATERIALS AND METHODS
Synechococcus sp. PCC7942, Synechocystis PCC6803, and mutants thereof were grown on BG-11 medium (27) supplemented with 20 mM Hepes-NaOH, pH 8.0, and with 5 mM glucose in the case of the Synechocystis PCC 6803 mutant ⌬psaA/B (28). The cultures were aerated with either high or low CO 2 concentration (5% CO 2 in air or 1:1 mixture of air and CO 2 -free air, respectively), at 30°C and light intensity of 100 mol photons m Ϫ2 s Ϫ1 . The cells were harvested during the log phase of growth and resuspended in growth media.
Gas exchange measurements were performed with a membrane inlet mass spectrometer (Balzers QMG 421) as described earlier (7). Changes in the concentration of one gas affects the signal obtained for the others and, if not corrected for, may lead to erroneous interpretation of the results. Simultaneous measurements of argon and nitrogen concentrations were therefore used to correct for variations or drifts in the system due to biological formation or consumption of O 2 and CO 2 , small changes in the rate of stirring or temperature, and in the gas consumption by the mass spectrometer. The latter was minimized by using silicon tubing with a small surface area (7). The cells were placed in a temperature-controlled chamber (2.8 ml) at 30°C and illuminated with two optic fibers at the desired light intensity. The various chemicals supplied to the cell suspension were introduced via a special injection port. Fig. 1 may serve as an example of the correct interpretation of curves obtained using the closed membrane introduction mass spectrometry chamber. Upon illumination of Synechococcus sp. strain PCC7942 cells the dissolved CO 2 concentration ([CO 2(dis) ]) in the medium declined steeply even prior to the onset of net O 2 evolution (see Fig. 1, panels A and B). Note that the slope of the curve relating external [CO 2(dis) ] to time (Panel B) cannot be taken as a direct indication of the initial rate of CO 2 removal by the cells because CO 2 is being formed continuously in the solution by net dehydration of HCO 3 Ϫ . Moreover, even though HCO 3 Ϫ concentration is virtually constant under the conditions of this experiment, net dehydration rate is not constant but rises because of the decline in CO 2 hydration rate as [CO 2(dis) ] drops. Consequently, the further [CO 2(dis) ] deviates from chemical equilibrium with HCO 3 Ϫ the lower the rate of CO 2 hydration, and therefore the higher the rate of net dehydration. At plateaus in the curve, where the CO 2 concentrations are relatively constant, the net rate of CO 2 uptake by the cells will be equal to the net rate of CO 2 formation by dehydration of HCO 3 Ϫ in the medium.

Displacement of [CO 2(dis) ] from Chemical Equilibrium upon
Illumination-The extent of [CO 2(dis) ] displacement from equilibrium was strongly affected by light intensity (Figs. 1 and 3). Raising the latter from 85 mol photons m Ϫ2 s Ϫ1 (Fig. 1, 2(dis) ] curve at high light intensity ( Fig. 1, panel C) and the following downward slope at low light (panel D) most likely reflect changes in the light-driven ET rate via the photosystems due to adjustments in the efficiency of energy transfer from the phycobilisomes. 2 Upon darkening ( Fig. 1, panel E), the [CO 2(dis) ] rose rapidly to the equilibrium value as the net rate of CO 2 uptake fell below the dehydration rate. The [CO 2(dis) ] frequently rose transiently above that expected at equilibrium probably due to formation of CO 2 from HCO 3 Ϫ in the intracellular C i pool and leak of the former to the medium (2). Note that in the second half of the period of high illumination (panel C), the rate of O 2 evolution rose as the ambient concentration of CO 2 increased, i.e. slower net CO 2 uptake. Moreover, while the ambient [CO 2(dis) ] declined (during the second half of panel D) indicating a rising rate of CO 2 uptake, the rate of O 2 evolution remained constant. If CO 2 fixation accounted for alterations in [CO 2(dis) ], one would have expected that changes in the CO 2 uptake curve to be the mirror image of those in the O 2 evolution curve, but that is not the case. These data provide supporting evidence for the conclusion that CO 2 uptake does not solely reflect CO 2 fixation and may occur even in its absence (5,6).
Dependence of CO 2 Uptake on Light Intensity-To distinguish between the effects of light intensity on CO 2 uptake and on CO 2 fixation, we used the high-CO 2 -requiring mutant of Synechococcus PCC7942, IL-3 (32) which maintains the [CO 2(dis) ] below CO 2 /HCO 3 Ϫ equilibrium even at CO 2 concentrations lower than its threshold for net CO 2 fixation (6). High-CO 2 -grown cells of Synechococcus PCC 7942 and of mutant IL-3 were exposed to a range of light intensities in the membrane introduction mass spectrometry chamber. The cells were provided with 1 mM C i , sufficient to saturate photosynthesis in the case of the wild type but too low to enable CO 2 -dependent O 2 evolution in the case of the mutant. At light intensities below 200 mol photons m Ϫ2 s Ϫ1 , CO 2 uptake by IL-3 was considerably faster than in the wild type ( Fig. 2). At higher light intensities, the rates of net CO 2 uptake by the mutant and Synechococcus PCC7942 were similar (Fig. 2, inset). The rates of net CO 2 uptake declined when cells of Synechococcus or mutant IL-3 were exposed to light intensity higher than 600 The inset provides the results obtained over the entire range of light intensities examined. Cell density corresponded to 4 g Chl/ml, and C i concentration was 1 mM, sufficient to saturate CO 2 fixation by the wild type but below the threshold level required in the mutant. Temperature was 30°C.
Inhibition of CO 2 Uptake by a WCB-Addition of the WCB p-chloromercuriphenylsulfonic acid to a cell suspension of high-CO 2 -grown Synechococcus PCC7942 resulted in severe, almost complete, inhibition of net CO 2 uptake by over 90% (as calculated from the CO 2 concentration at the plateau attained after the addition of the WCB, Fig. 3A). Photosynthetic O 2 evolution was also severely depressed. To distinguish between a direct effect of the WCB on CO 2 uptake and a possible indirect effect due to the decline in CO 2 fixation, we applied iodoacetamide (IAC) that completely inhibits CO 2 fixation (5, 6, 29 -31). We also made use of the high-CO 2 -requiring mutant of Synechococcus PCC7942, IL-3 in which the light-saturated rate of CO 2 uptake is similar to that of its wild type (Fig. 2) even at CO 2 concentrations lower than its threshold for net CO 2 fixation (6). In the presence of IAC the rate of net CO 2 uptake only declined by about 20% (Fig. 3B), although O 2 evolution was completely suppressed (not shown) providing further evidence that displacement of [CO 2(dis) ] from equilibrium may occur irrespective of whether or not CO 2 is fixed. Addition of the WCB either to IAC-treated wild type or to IL-3 cells inhibited CO 2 uptake almost completely (Fig. 3B). The possibility that WCB inhibition of CO 2 uptake and fixation reflected severe unspecific damage to the cells was examined by raising the concentration of C i in the medium. Normal photosynthetic rates (306 mol O 2 evolved mg Ϫ1 Chl h Ϫ1 ) were observed when WCB-treated Synechococcus cells were supplemented with 20 mM C i . These data indicate that a reduced availability of CO 2 to otherwise fully functional photosynthetic machinery led to the inhibition of CO 2 fixation by p-chloromercuriphenylsulfonic acid. Interestingly, Synechocystis PCC6803 is far less inhibited by the WCB than Synechococcus PCC7942 for a reason yet unknown.
Dependence of CO 2 Uptake on Photosynthetic Electron Trans-port via PSI-The functional linkage between photosynthetic ET and CO 2 uptake was examined with the aid of electron acceptors, donors, and electron transfer inhibitors. In addition to Synechococcus PCC7942 we also examined Synechocystis PCC6803, which exhibits a similar displacement of [CO 2(dis) ] from equilibrium upon illumination and where mutants impaired in PSI activity are available. As previously reported (33), inhibition of linear electron flow by DCMU abolished CO 2 uptake (Fig. 4A). However, addition of duroquinol that donates electrons to plastoquinol and reduces cytochrome b6f thus priming light-driven PSI electron flow, reestablished CO 2 uptake but not CO 2 fixation (completely inhibited by DCMU). These results provide further evidence that generation of CO 2 / HCO 3 Ϫ disequilibrium is not compulsorily linked to CO 2 removal by the carboxylation reactions but does require PSI activity (Fig. 4A).
Direct evidence for the role of PSI in CO 2 uptake was obtained by use of a ⌬psaA/B mutant of Synechocystis PCC6803 that lacks a functional PSI (28). The mutant was unable to displace [CO 2(dis) ] from equilibrium even when supplied with DMBQ, an efficient electron acceptor of PSII (Fig. 4B). The high DMBQ-dependent O 2 evolution (290 mol O 2 mg Ϫ1 Chl h Ϫ1 , Fig. 4B) indicated significant PSII activity in the mutant. This mutant obviously is neither able to fix CO 2 nor to perform cyclic electron flow in the light. Thus linear electron flow from PSII via plastoquinone to DMBQ, which does not involve PSI, is not sufficient to drive light-dependent CO 2 uptake.
Addition of the electron acceptors DMBQ or methyl viologen (MV) that draw electrons from PSII and PSI, respectively thus inhibiting cyclic electron flow via PSI, resulted in net CO 2 extrusion in Synechococcus. The [CO 2(dis) ] at steady state was higher than expected at CO 2 /HCO 3 Ϫ equilibrium (Fig. 4, C and  D). Addition of CA to the medium reestablished the chemical equilibrium value (Fig. 4C). Similar results were obtained when MV was supplied to mutant M55 of Synechocystis PCC6803 in which ndhB, the encoding subunit II of NAD(P)H dehydrogenase NDH-1, was inactivated (not shown). Net CO 2 efflux of the magnitude observed in the presence of MV is highly likely to be the consequence of net HCO 3 Ϫ uptake followed by intracellular conversion to CO 2 and leak of the latter to the medium (7,23,34). These results suggest potential reversibility of the direction of the C i cycling. The ⌬psaA/B mutant did not evolve CO 2 in the presence of DMBQ (Fig. 4B) for a reason not understood.
Energy Requirement for the C i Cycling Activity-To distinguish between ATP hydrolysis and ⌬H ϩ as the direct source of energy for CO 2 uptake, we have examined the effects of drugs that specifically inhibit ATP synthesis as well as uncouplers that dissipate the ⌬H ϩ . Arsenate and DCCD inhibit the formation of ATP at the substrate level and proton gradient driven synthesis respectively while hardly affecting ⌬H ϩ (35,36). The uncouplers CCCP and ammonia, on the other hand, abolish the generation of ⌬H ϩ .
Photosynthetic O 2 evolution gradually ceased following addition of arsenate to a cell suspension of Synechococcus PCC7942 (Fig. 5A). The slowness of the response probably reflects the rate of arsenate uptake. Although the ambient [CO 2(dis) ] rose after about 100 s, it was still below the CO 2 / HCO 3 Ϫ equilibrium value after photosynthesis had halted completely. The further rise in [CO 2(dis) ] after addition of CA confirmed the lack of CO 2 /HCO 3 Ϫ equilibrium due to net CO 2 uptake. Addition of DCCD brought about a rapid rise in [CO 2(dis) ] (Fig. 5B), but as in the case of arsenate treatment, [CO 2(dis) ] was still well below the equilibrium value after O 2 evolution (CO 2 fixation) had ceased. Following the DCCD treatment, the rate of CO 2 uptake, calculated from the plateaus in the curve before and after the addition of DCCD, declined by FIG. 3. The effect of a water channel blocker on net CO 2 uptake by Synechococcus PCC 7942 and its mutant IL-3. Cell density corresponded to 4 g Chl/ml, and light intensity was 160 mol photons m Ϫ2 s Ϫ1 . Other conditions were as in Fig. 1. A, inhibition of net CO 2 uptake following the addition of the water channel blocker (WCB) p-chloromercuriphenylsulfonic acid (21) to Synechococcus PCC 7942. B), rates of net CO 2 uptake by Synechococcus PCC 7942 and mutant IL-3 following iodoacetamide (IAC, 3.3 mM provided in the dark 8 min prior to illumination) and of WCB treatments. The data are presented as a percentage of those exhibited by the wild type (270 mol CO 2 mg Ϫ1 Chl h Ϫ1 ). 54% (from 240 to 112 mol CO 2 absorbed mg Ϫ1 Chl h Ϫ1 ). At the same time, net O 2 evolution (235 mol O 2 evolved mg Ϫ1 Chl h Ϫ1 ) was replaced by respiratory O 2 uptake (67 mol O 2 absorbed mg Ϫ1 Chl h Ϫ1 ). Supply of MV led to a slight rise in [CO 2(dis) ] but the rate of increase was much lower than that observed in the absence of DCCD (Fig. 5B) possibly because of lack of sufficient ATP to drive HCO 3 Ϫ uptake (and consequently CO 2 extrusion). Subsequent CA supply sharply increased the [CO 2(dis) ] again indicating that the latter was below the equilibrium value due to net CO 2 uptake. On the addition of the proton conductor CCCP, the rate of O 2 evolution rose transiently (Fig. 5C), most probably because of stimulation of electron transport due to dissipation of the trans-thylakoid ⌬H ϩ . Photosynthetic O 2 evolution then ceased, but O 2 uptake at a rate similar to that of dark respiration was detectable. The [CO 2(dis) ] rose almost immediately upon addition of CCCP, briefly overshooting equilibrium value. The [CO 2(dis) ] level trace resembles that observed when photosynthetic ET is halted by darkening (see Fig. 1). Similar results were obtained using ammonium chloride (not shown).
The [CO 2(dis) ] was below CO 2 /HCO 3 Ϫ equilibrium even when the internal ATP was largely exhausted as indicated by the cessation of O 2 evolution following the arsenate or DCCD treatments (Fig. 5, A and B). On the other hand, the uncouplers abolished both CO 2 uptake and fixation (Fig. 5C). These data suggest that CO 2 uptake depends on a ⌬H ϩ and that direct involvement of ATP hydrolysis in the displacement of [CO 2(dis) ] below equilibrium is therefore unlikely.

DISCUSSION
Severe inhibition of CO 2 uptake by the aquaporin blocker (Fig. 3) suggests that these channels form a major route for CO 2 entry to high-CO 2 -grown Synechococcus cells. Because passage of CO 2 through aquaporins is presumably passive either by diffusion or by mass flow together with water molecules, CO 2 transport mediated by specific membrane entities would appear to play a minor role, if any, unless the aquaporin blocker also specifically inhibits these entities.
Results presented in this work confirm that net CO 2 uptake by Cyanobacteria is not directly linked to CO 2 fixation and may proceed in its absence (Figs. 1, 2, and 4). Under low, but not high light intensity, CO 2 uptake by mutant IL-3, which had been maintained at CO 2 concentration lower than its threshold for CO 2 fixation, was faster than in the wild type (Fig. 2). This may reflect consumption of NADPH and/or dissipation of ⌬H ϩ to support CO 2 fixation in the wild type. As discussed below, CO 2 uptake may well be driven by electron transport-dependent ⌬H ϩ (6). At saturating light intensity, net CO 2 uptake was similar in the wild type and in mutant IL-3 suggesting common limitation by the rate of conversion of CO 2 to HCO 3 Ϫ . In view of the queries recently raised as to the role of PSI as the major energy source for CO 2 uptake (24), we summarize the evidence in favor of this role obtained in the present investigation as follows. 1) CO 2 /HCO 3 Ϫ disequilibrium consequent on net CO 2 uptake was formed even in the presence of DCMU when reduced duroquinone (Fig. 4A) or dithiothreitol (not shown) was added. 2) C i cycling was absent in the ⌬psaA/B mutant, lacking PSI activity even in the presence of DMBQ, which enabled a high flow of electrons via PSII (Fig. 4B); 3) Synechocystis PCC6803 mutant M55 or the Synechococcus PCC7942 mutant N5 in which ndhB had been inactivated (37,38) are defective in cyclic PSI ET. Although they exhibit photosynthetic carbon fixation when supplied with elevated CO 2 levels (38), they are unable to displace the CO 2 /HCO 3 Ϫ equilibrium. 4) Draining electrons from PSI by means of artificial acceptors switched Synechococcus PCC7942 from net CO 2 uptake to net HCO 3 Ϫ uptake (Fig. 4, C and D 2(dis) ] from equilibrium even after addition of 2,5-dimethyl-p-benzoquinone (DMBQ), which accepts electrons from PSII. Cell density corresponded to 10 g Chl/ml. C, inversion of CO 2 /HCO 3 Ϫ disequilibrium after supply of DMBQ to Synechococcus PCC 7942 and return to equilibrium after supply of carbonic anhydrase (CA). Cell density corresponded to 15 g Chl/ml. D, inversion of CO 2 /HCO 3 Ϫ disequilibrium after supply of methyl viologen (MV) to Synechococcus PCC 7942, abolition of disequilibrium in the dark, recreation of disequilibrium in a subsequent light period, and return to equilibrium after supply of CA. Cell density corresponded to 8 g Chl/ml. These observations together with those reported elsewhere (23,25,39) provide a very strong case for the central role of PSI ET in driving CO 2 uptake. The question is, therefore, how these observations can be reconciled with the contrasting effects of inactivation of NDH-1 components on CO 2 uptake and on P700 oxidation in Synechocystis PCC6803 (24). Another problem is that the results presented here indicate that a ⌬H ϩ generated by photosynthetic ET serves as the direct source of energy for CO 2 uptake (Fig. 5). This finding is consistent with the predictions of our model (6), but the exclusive dependence of CO 2 uptake on PSI ET has to be reconciled with the accepted notion that both linear and cyclic PSI ET lead to the formation of a ⌬H ϩ across the thylakoid membrane. Both problems are addressed in the following discussion.
To account for the effect of mutation in a component of NDH-1 on CO 2 uptake and on P700 oxidation, Klughammer et al., (26) suggested that in Synechococcus PCC7002 one type of NDH-1 essential for cyclic ET is located on the thylakoid, whereas another type engaged in CO 2 uptake is located on the cytoplasmic membrane. However, immunolocalization studies have demonstrated that a component of NDH-1, NdhB, critical for CO 2 uptake is exclusively located in the thylakoid membrane (25).
The possibility should be considered that only a small fraction of the PSI population is engaged in CO 2 uptake. Depression of CO 2 uptake in the ⌬ndhD3/⌬ndhD4 mutant (24) was in fact associated with some decline in cyclic PSI activity, possibly reflecting this small fraction of PSI. In the ⌬ndhD1/⌬ndhD2 mutant where cyclic PSI activity was largely depressed, CO 2 uptake was little affected presumably because the PSI units engaged in CO 2 uptake were still operating. The marked dependence of the PSI/PSII ratio on the growth conditions, particularly CO 2 concentration (22) and salinity (40), would be in accordance with such a suggestion. The observation that the distribution of PSI in Synechococcus PCC7942 is heterogeneous, indicated by the higher abundance of PSI in peripheral as compared with inner thylakoid membranes (41), also lends support to the notion that the PSI population may be functionally heterogeneous. This heterogeneity might relate to different routes for electron flow from NADPH to plastoquinone (42,43).
Different routes for electron flow might also be the basis of the differences in CO 2 uptake between the various ndhD mutants (24). Uptake of CO 2 was observed in each of the single mutants ⌬ndhD3 and ⌬ndhD4 but not in the double mutant (24). Ogawa et al. 3 concluded that two discrete systems for CO 2 uptake operate in Synechocystis PCC6803. This is based on the differing kinetic parameters for CO 2 uptake between mutants ⌬ndhD4 and ⌬ndhD3 and the inducibility of CO 2 uptake by low CO 2 conditions in the wild type and in mutant ⌬ndhD4 but not in mutant ⌬ndhD3. The NdhD3-and NdhD4-dependent CO 2 uptake systems may constitute alternative PSI-dependent routes for electron flow. Quantitative consideration shows that the residual rate of P700 oxidation (i.e. the PSI cyclic electron flux) in the ⌬ndhD1/⌬ndhD2 (24) was too low to account for the rate of CO 2 uptake in this mutant. This may indicate the presence of an alternative acceptor of electrons from the NdhD3-and NdhD4-dependent CO 2 uptake systems such as succinate:quinol oxidoreductases (44), a possibility currently being examined.
It has recently been proposed (6) that CO 2 is converted to HCO 3 Ϫ in alkaline domains on the stromal face of the thylakoid membrane (see the Introduction). Conversion of CO 2 to HCO 3 Ϫ in such domains maintains the inward diffusion gradient for CO 2 and a cytoplasmic CO 2 concentration below that of equilibrium with HCO 3 Ϫ . Withdrawal of electrons from plastoquinone by either DMBQ or MV (Fig. 4) would prevent the formation of these alkaline domains and thus also of CO 2 uptake. The exclusive reliance of CO 2 uptake on PSI-generated ⌬H ϩ probably involves an electron carrier yet to be identified. In view of the finding that NdhD3 and NdhD4 are essential components of two CO 2 uptake systems (Ogawa et al., submitted) but not of the respiratory electron path they are likely candidates for the formation of the alkaline domains during PSI ET. Another candidate is ferredoxin-NADPH reductase. Recent studies (43) demonstrated that linear ET was functional but that the plastoquinone-cytochrome b6f complex reductase step of cyclic PSI was defective in a mutant where the N-terminal of ferredoxin-NADPH reductase was truncated preventing its association with the thylakoids).