INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

In cyanobacteria the type I NAD(P)H dehydrogenase complex (NDH-1) is a proton-pumping complex that has been shown to be involved in both the respiratory and photosynthetic electron transport chains (25). NDH-1 acts as a plastoquinone oxidoreductase with NADH or NADPH as a substrate. There is evidence that NDH-1 is located in the cytoplasmic membrane as well as the thylakoid membrane, but evidence for location in the cytoplasmic membrane has not been consistent. NDH-1 is composed of 12 recognized subunits, with the large, hydrophobic NdhB, NdhD, and NdhF subunits being core membrane components. The availability of a whole-genome database for Synechocystis sp. strain PCC 6803 has shown that most ndh genes are present as single copies; however, ndhD and ndhF are present as multiple copies with five and four members, respectively (note that NdhF4 has homology with NdhD5 and could be counted as an NdhD homolog). There is a marked degree of protein sequence divergence within the NdhD and NdhF families, and this has led to suggestions that several NDH-1 complexes may exist in cyanobacteria, each with different NdhD and/or NdhF subunits and with each potential complex having differing functions (21, 24). One main role for NDH-1 in the thylakoid membrane is to participate in cyclic electron flow around photosystem I and to pump protons into the lumen, thereby contributing to pH-driven ATP generation at the expense of NADPH (11-13). However, other roles for NDH-1 are possible.

It has been demonstrated that NDH-1 is essential for inorganic carbon (C_i) transport in cyanobacteria (4, 5, 10, 16-18, 21). Inactivation of ndhB or ndhL in Synechocystis sp. strain PCC 6803 greatly reduced the activities of CO₂ and HCO₃ uptake, and in the past it has been assumed that the NDH-1-dependent cyclic electron transport supplies ATP to drive the C_i uptake (16-18). However, the question of why ATP produced by linear electron transport in these mutants does not drive the C_i transport, remains unresolved, and it was not certain whether ATP energizes the uptake of CO₂ and HCO₃. The finding that the cmp operon encodes an ATP binding cassette transporter for HCO₃ clearly indicates the presence of at least one type of ATP-dependent HCO₃ transporter (22). Li and Canvin (8) reported that HCO₃ uptake is supported by linear electron transport while CO₂ uptake is supported by cyclic electron transport, based on the observation of differential effects of electron transport inhibitors and acceptors on uptake of the two carbon species. The results suggested that ATP produced by noncyclic electron transport energized HCO₃ transport. Although the involvement of NDH-1 in CO₂ uptake is clearly demonstrated, little is known about the mechanism of CO₂ uptake and how it is energized.

Some ndh mutants, such as M55 (ndhB), show major inhibition of both CO₂ and HCO₃ uptake (16-18), whereas other mutants, such K22 and A41 (ndhD3), show an effect largely on high-affinity CO₂ uptake (5). However, the previously described M55 mutant of Synechocystis sp. strain PCC 6803 is a highly disruptive insertional mutant of the single-copy ndhB gene (16-18), and it could be argued that removal of the NdhB protein could cause nonassembly of all NDH-1 complexes and production of secondary phenotypes. A less disruptive approach is to mutagenize ndhB with multiple single-base mutations and to select for mutants where the NdhD protein is still capable of assembly but lacks most functionality. To help understand the role of NDH-1 in C_i transport, we investigated whether random-site mutations within the ndhB gene of Synechocystis sp. strain PCC 6803 inhibit the uptake of CO₂ and HCO₃ to the same extent. We show in this paper by measurement of CO₂ and HCO₃ uptake in these mutants that CO₂ uptake is much more strongly inhibited than HCO₃ uptake, a result more consistent with the effect of ndhD3 mutations in Synechococcus sp. strain PCC 7002.

Synechocystis sp. strain PCC 6803 possesses five ndhD genes (3), with sll1733 (designated ndhD3) being homologous to ndhD3 in Synechococcus sp. strain PCC 7002. We describe in this paper the growth characteristics of mutants constructed by knocking out each of the ndhD genes in Synechocystis and show that ndhD3 has a specific role in inducing high-affinity CO₂ uptake. Another purpose of the present work was to make some comparisons among the suite of ndhD mutants, as well as existing mutants such as M55 and new multiple-point mutants of ndhB, and attempt to provide a consensus view of the role of NDH-1 in C_i uptake in cyanobacteria.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Growth conditions. WT and mutant cells of Synechocystis sp. strain PCC 6803 were grown at 30°C in BG11 medium (26) buffered with 20 mM N-Tris(hydroxymethyl)methyl-2-amino-ethanesulfonic acid (TES)-KOH (pH 8.0 and 7.0) or 2-(N-morpholino)ethanesulfonic acid (MES)-KOH (pH 6.5) and aeration with 3% (vol/vol) CO₂ in air or with air. The solid medium was BG11 supplemented with 1.5% agar, 5 mM sodium thiosulfate, and 20 mM the same buffer. When air containing 50 or 20 ppm CO₂ was used, the air was passed through Wako lime (Wako Co., Tokyo, Japan) packed in a metal tube (5-cm diameter and 20-cm length) at a flow rate of approximately 200 ml/min. Continuous illumination was provided by fluorescent lamps generating photosynthetically active radiation of 60 µmol of photons/m²/s.

Construction of mutants. A clone, pUCEE5.1, derived from pUCEK11.8 (16), was used for random mutagenesis of ndhB. A kanamycin resistance cartridge (Km^r) cassette was inserted at the BglII site downstream of ndhB. Insertion of the cassette at this site did not change the phenotype of WT Synechocystis cells transformed with this construct (16). A fragment of 606 bp between the BamHI and SpeI sites was removed from the ndhB gene, and fragments containing the same DNA region were synthesized by a PCR method that introduces random mutations (7) and then reinserted between the two restriction sites. The sequences upstream of the BamHI site and downstream of the SpeI site were used to design the PCR primers, and pUCEE5.1 was used as a template. A library thus constructed, containing various mutated ndhB genes, was used to transform the WT cells of Synechocystis sp. strain PCC 6803 into Km^r mutants, using the protocol of Williams and Szalay (29). The transformants were spread on agar plates containing BG11 medium and kanamycin (10 µg/ml) buffered at pH 8.0, and the plates were incubated in air. Eight mutants were rescued as colonies that grew slowly under air. The mutated ndhB in the transformants was segregated to homogeneity (by successive-streak purification) as determined by PCR amplification. The ndhB gene in each mutant was amplified by PCR and used as a DNA-sequencing template.

Silicone oil-filtering centrifugation. The uptake of ¹⁴CO₂ and H¹⁴CO₃ was measured in high-CO₂-grown cells (aerated with air for 18 h in the light) using the silicone oil-filtering centrifugation method (28). The cells were suspended in BG11 medium buffered with 20 mM TES-KOH (pH 8.0) at a chlorophyll (Chl) concentration of 20 µg/ml. C_i uptake was initiated by the addition of ¹⁴CO₂ or H¹⁴CO₃ in the light and terminated by centrifugation.

Determination of growth characteristics. WT and mutant strains grown under 3% CO₂ were collected and resuspended in fresh BG11 medium to optical densities at 730 nm (OD₇₃₀s) of 1.0, 0.1, and 0.01. Two microliters of the cell suspensions was spotted onto BG11 agar plates buffered at various pHs. The plates were incubated under 3% (vol/vol) CO₂ in air or under air for 5 days with continuous illumination by fluorescent lamps at a photosynthetically active radiation intensity of 60 µmol of photons/m²/s. The OD₇₃₀ was measured with a recording spectrophotometer, model UV2200 (Shimadzu Co., Kyoto, Japan).

Electrophoresis and Western analysis. An antibody was raised against partial NdhB fused to glutathione-S-transferase (GST). The ndhB gene amplified by PCR was digested with MvoI and MbaI, and a fragment of 226 bp (from base 419 to 644 as numbered from the initiation codon of ndhB) was excised from a gel after electrophoresis of the digest. The fragment was blunted and ligated to the SmaI site of pUC119. The insert DNA was excised with EcoRI and BamHI and was ligated to pGEX-3X (Pharmacia, Uppsala, Sweden). The construct was used to transform Escherichia coli. GST-NdhB (partial) formed inclusion bodies, which were isolated, solubilized with 5% sodium dodecyl sulfate (SDS), and electrophoresed by SDS-polyacrylamide gel electrophoresis. A prominent band of GST-NdhB (partial) at 33 kDa was cut out from the gels and mashed with a pestle and mortar to be injected into rabbits.

RT-PCR analysis. RNA from air-grown Synechocystis cells was extracted by the method of Aiba et al. (1) and then subjected to reverse transcriptase (RT) PCR (14). Primers specific to the 3' ends of ndhF3, ndhD3, sll1734, and sll1735, respectively, were used for the RT reaction. The subsequent PCR was performed with a pair of primers specific to ndhF3.

Mass spectrometric measurements. Measurements of the initial CO₂ uptake rate and kinetics for CO₂ and HCO₃ uptake under steady-state conditions were performed according to the method of Badger et al. (2).

Other methods. Unless otherwise stated, standard techniques were used for DNA manipulation (9). Pigments in the cells were extracted in methanol, and the Chl concentration in the extract was determined (20).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Isolation of ndhB mutants and sites of mutations. Eight mutants that grew poorly in air on agar plates buffered at pH 8.0 were rescued after mutations were introduced into ndhB of WT Synechocystis sp. strain PCC 6803 by low-fidelity PCR (7). Sequencing of impaired ndhB genes in these mutants (designated B1 to B8) revealed that B7 and B8 were identical to B1. Amino acid residues substituted in the B1 to B6 mutants are summarized in Table 1. There were substitutions of 2 to 8 amino acid residues in each of the ndhB mutants.

CO₂ and HCO₃ uptake in the ndhB mutants. Figure 1 shows the time courses of CO₂ uptake (left panel) and HCO₃ uptake (right panel) for WT, B1, and M55 cells grown at pH 8.0 under 3% CO₂ and then transferred to aeration with air for 18 h in the light. The activity of CO₂ uptake was very low in the B1 and M55 mutants, but the activity of HCO₃ uptake in B1 was as high as that in the WT, whereas M55 possessed half the WT activity. High-CO₂-grown cells of these mutants also showed very low activity of CO₂ uptake, indicating that the low activity is not a result of aeration with air in the light (data not shown). Thus, CO₂ uptake was inhibited preferentially in these mutants. This was confirmed by the results summarized in Fig. 2, where the activities of CO₂ and HCO₃ uptake in the ndhB mutants are shown as percentages of the WT activities. Although the extent of inhibition varied, CO₂ uptake was more strongly inhibited than HCO₃ uptake in all the mutants examined. Thus, CO₂ uptake is strictly dependent on NDH-1 whereas HCO₃ uptake proceeds even in the absence of this enzyme. However, the result was that HCO₃ uptake was partly inhibited in the mutants, although the extent of inhibition was at most 50%, even in M55. This suggests that the low activity of HCO₃ transport is a secondary effect caused by the absence of NDH-1 or that NDH-1 has a supplemental role in the transport of HCO₃.

View larger version (22K): [in this window] [in a new window]   FIG. 1.   Time courses of uptake of 14CO2 (left panel) and H14CO3 (right panel) into WT (circles), B1 (squares), and M55 (triangles) cells, measured by the silicone oil-filtering centrifugation method. The cells were grown under 3% CO2 and then aerated with air for 18 h in the light at pH 8.0 (solid curves) or pH 7.0 (dashed curves). The cells were suspended in BG11 medium containing 15 mM NaCl buffered with 20 mM HEPES-KOH, pH 8.0. The concentrations of 14CO2 and H14CO3 were 7.9 and 61 µM, respectively.

View larger version (23K): [in this window] [in a new window]   FIG. 2.   Amount of Ci taken up by the mutant cells during incubation with 14CO2 (hatched bars) for 10 s or with H14CO3 (open bars) for 20 s in the light. Each value is shown as a percentage of the value obtained for WT cells (259 nmol of CO2/mg of Chl per 10 s and 473 nmol of HCO3/mg of Chl per 20 s), and the error bars indicate standard deviations (n = 3). Cells grown at pH 8.0 were suspended in the same medium as for Fig. 1.

Growth characteristics of the ndhB mutants. To explore how the inhibition of CO₂ and HCO₃ uptake affects the growth characteristics of the cells, growth of the mutant strains was examined on solid BG11 medium buffered at various pHs under 3% CO₂ or air. At pHs 8.0 and 7.0 under 3% CO₂, all the mutants except M55 grew as fast as the WT (Fig. 3A and B). M55 grew more slowly than the WT at pH 8.0 and was unable to grow at pH 6.5 under the high-CO₂ conditions (Fig. 3A to C). The B1 and B2 mutants grew more slowly at pH 6.5 in 3% CO₂ (Fig. 3C). All the mutants except B4 and B6 grew more slowly than the WT even at pH 8.0 under air (Fig. 3D) and very slowly at pH 7.0 (Fig. 3E). Most of the mutants were unable to grow at pH 6.5, and the growth of B6 was relatively poor under these conditions (Fig. 3F). The ability of the mutants to grow under atmospheric levels of CO₂, especially at pH 7.0 or 6.5, was correlated with their activity for CO₂ uptake (Fig. 2 and 3).

View larger version (73K): [in this window] [in a new window]   FIG. 3.   Effects of pH and CO2 concentration on the growth of the WT and mutants on agar plates. The WT and ndhB mutant cells of Synechocystis were pelleted by centrifugation and resuspended in BG11 medium, pH 8.0, 7.0, or 6.5. Two microliters of the cell suspensions, OD730 values of 1.0 (upper row of each panel), 0.1 (middle row), and 0.01 (lower row), were spotted on agar plates containing BG11 medium buffered at pH 8.0, pH 7.0, and pH 6.5. The plates were incubated under 3% (vol/vol) CO2 in air (A to C) or under air (D to F) for 5 days at 60 µmol of photons of photosynthetically active radiation/m2s1.

View larger version (13K): [in this window] [in a new window]   FIG. 4.   Effect of pH on growth rates of WT, B1, and M55 in liquid BG11 medium during aeration with 3% (vol/vol) CO2 in air (closed symbols) or under air (open symbols). Growth rates are shown by the µ values. Doubling time (in days), 0.693/µ.

Identification of NdhB. The antibody raised against GST-NdhB (partial) cross-reacted with several bands in the total cell membranes of the WT but most strongly with the band at 55 kDa, the molecular mass of NdhB estimated from the DNA database translation (Fig. 5, left lane). No cross-reacting band was present at 55 kDa in the membranes of M55 (right lane), indicating that the cross-reacting band in the WT is NdhB. Western analysis with the total cell membranes of B1 using the same antibody gave a band at 55 kDa (Fig. 5, middle lane). The density of the band in B1 membranes was lower than that in the WT membrane. The results indicated that NDH-1 is present in the B1 mutant, although the protein may not be fully functional.

View larger version (67K): [in this window] [in a new window]   FIG. 5.   Immunoblot of NdhB in the total cell membranes of Synechocystis sp. strain PCC 6803 WT, B1, and M55. Samples (10 µg of proteins) were solubilized at room temperature and boiled for several minutes and were run in a 10% gel. The antibody against GST-NdhB (partial) was used for immunoblotting.

Growth of ndhD mutants under limiting CO₂ conditions. ndhD1, ndhD2, ndhD3, ndhD4, and ndhD5 mutants grew as fast as the WT at pH 7.0 in air (data not shown). Measurement of the growth rates of these mutants at pH 6.5 and 50 ppm CO₂ revealed that the ndhD3 mutant grew at about a third of the rate of WT cells, although the rest of the ndhD mutants showed virtually the same growth rate as the WT (Table 2). ndhF3 and sll1734 mutants constructed by inactivating the genes upstream and downstream of ndhD3, respectively, showed the same growth characteristics as the ndhD3 mutant at pH 6.5 and 50 ppm CO₂, but inactivation of sll1735 downstream of sll1734 did not have a significant effect on the growth rate under these conditions. These results suggested that ndhD3, ndhF3, and sll1734 have a common and specific effect on the growth of cells under limiting CO₂ conditions.

Expression of ndhD3 and neighboring genes. The products of RT reactions with primers specific to the 3' ends of ndhF3, ndhD3, sll1734, and sll1735 were used as templates for the subsequent PCR with a set of primers specific to ndhF3. The PCR with three different templates gave the same products (Fig. 6, lanes 1 to 3). This indicated that ndhF3, ndhD3, and sll1734 are expressed together as an operon. No PCR product was found when a primer specific for sll1735 was used for the RT reaction (lane 4), indicating that the gene is expressed independently of the upstream genes.

View larger version (8K): [in this window] [in a new window]   FIG. 6.   RT-PCR analysis of total RNA from Synechocystis cells showing the expression of ndhF3, ndhD3, and sll1734 as an operon. RNA was prepared from cells grown at pH 8.0 under 50 ppm CO2. PCR amplification was performed with a set of primers specific to ndhF3 (ATTATCTGGCTAGTACC and GAATAGCTAAGAAAGGC), and the product was 1.8 kbp. The templates used for PCR were cDNAs which were synthesized by reverse transcription of mRNA by addition of RT with primers specific to ndhF3 (lane 1), ndhD3 (lane 2), sll1734 (lane 3), and sll1735 (lane 4), respectively.

Initial and steady-state rates of CO₂ uptake. Cyanobacteria exhibit a transient and rapid uptake of CO₂ following a dark-to-light transition. The CO₂ uptake during steady-state photosynthesis and the relative affinity for gross CO₂ uptake can be measured and calculated according to a method developed by Badger et al. (2). Using this method, the steady-state rate and initial rate of CO₂ uptake were measured and calculated under various C_i concentrations, allowing K_0.5(CO₂), the concentration of CO₂ required to reach the half-maximal uptake rate, to be determined. Figure 7 shows the K_0.5(CO₂) and V_max values for the steady-state (Fig. 7A) and initial (Fig. 7B) rates of CO₂ uptake obtained for the WT and ndhD3 mutant cells grown at 2% CO₂ or induced overnight at 20 ppm CO₂. There was no significant difference between the V_max values of the WT and mutant cells. In contrast, the K_0.5(CO₂) values for initial and steady-state CO₂ uptake were significantly higher in the mutant than in the WT, in cells both before and after 18 h of induction at 20 ppm CO₂. The affinity of steady-state CO₂ uptake for CO₂ did not change in the mutant after induction at 20 ppm CO₂, while the affinity increased about 2.5-fold in the WT after induction. Thus, ndhD3 appears to be essential for the induction of high-affinity CO₂ uptake. In Synechocystis sp. strain PCC 6803, the V_max of net CO₂ uptake, upon illumination, was similar to the steady-state rate. The activation of initial CO₂ uptake may not be as fast in this strain as in other cyanobacteria studied (e.g., Synechococcus sp. strain PCC 7942), leading to similar initial and gross steady-state rates.

View larger version (36K): [in this window] [in a new window]   FIG. 7.   The K0.5(CO2) and Vmax values for steady-state (A) and initial (B) CO2 uptake by the WT and ndhD3 mutant (M) strains of Synechocystis sp. strain PCC 6803 grown under 2% CO2 (shaded and hatched bars) and after 18 h of induction under 20 ppm CO2 (open bars). Each bar shows the average value of the results of four measurements, and the error bars indicate standard deviations.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

As demonstrated previously (16), CO₂ uptake was abolished in the mutant (M55) of Synechocystis sp. strain PCC 6803 that lacked ndhB. CO₂ uptake was also inhibited in other ndhB mutants (B1 to B6 [Fig. 2]) that had been generated by random-site mutagenesis. However, in contrast to the previous observation with M55, inhibition of HCO₃ uptake was not significant in these mutants and was at most 50%, even after complete inactivation of ndhB in M55 (Fig. 1 and 2). In other words, careful analysis of M55 indicates that inhibition of CO₂ uptake is the major effect and that inhibition of HCO₃ is a minor effect. In the less disruptive B1 to B6 mutants, the difference is more specifically related to CO₂ uptake, indicating that the complete loss of the NdhB protein in M55, and potential loss of all NDH-1 complexes, may lead to secondary effects on HCO₃ uptake. Moreover, mutants B1 and M55 grew as fast as the WT in liquid medium adjusted to pH 8.0 and gassed with air levels of CO₂ (Fig. 4), unlike the previous observation that M55 was unable to grow under these conditions (16). Since WT cells grew at pH 7.0 as fast as at pH 8.0, we did not pay much attention to the pH of the medium for growth of M55 in previous studies, and a medium of pH 7.0 was frequently used. The discrepancy between the present and previous observations possibly arose from a mistake in growing the M55 cells at pH 7.0. To test this possibility, we grew M55 at pH 7.0 under 3% CO₂ and then aerated with air overnight at the same pH. These M55 cells did not show HCO₃ transport activity (Fig. 1), in contrast to the result obtained with the mutant cells grown and induced at pH 8.0 (Fig. 1 and 2). Thus, the slow growth or nongrowth of M55 under air reported in previous studies appears to be due to a failure to present the mutant with the more favorable conditions of pH 8.0 medium. In contrast, the B1 mutant grew almost as well as the WT at pHs 6.5, 7.0, and 8.0 under high-CO₂ conditions (Fig. 4).

The mutation of ndhB had much less effect on HCO₃ uptake than on CO₂ uptake (Fig. 1 and 2), consistent with the observation that HCO₃ transport appears to be driven by linear electron transport (8). The activity of HCO₃ uptake in M55 was much lower than the WT rate (Fig. 1 and 2). This could be due to secondary effects of stress caused by exposing the cells to light under low-CO₂ conditions in the absence of CO₂ uptake. The possibility that NDH-1 has a secondary role in HCO₃ transport in supplying additional ATP by an NDH-1-dependent cyclic electron transport around photosystem I cannot be ruled out (11-13). The CO₂ uptake reaction is postulated to be an energy-dependent unidirectional conversion of CO₂ to HCO₃ (reviewed in reference 4). However, this reaction may be ATP independent, since CO₂ uptake is not operational in the presence of linear electron transport alone (8).

The inability of the ndhB mutants to grow at pH 7.0 or 6.5 under atmospheric levels of CO₂ correlated with their low capacities for CO₂ uptake (Fig. 2 and 3). Thus, the inorganic carbon source is mainly supplied by CO₂ uptake under these conditions. These mutants grew as fast as the WT in liquid medium at pH 8.0 under air, which indicated that at this pH, HCO₃ transport was a sufficient carbon source for the cells. On agar plates, the mutant colonies grew more slowly than the WT at pH 8.0 under air.

The presence of five ndhD genes in the Synechocystis sp. strain PCC 6803 genome (3) and their differential expression under different CO₂ concentrations (21) suggested that there are multiple types of NDH-1 complexes with different functions. In this series of mutants, the ndhD3 mutant was the only one that displayed the phenotype of slow growth at limiting CO₂ (i.e., 50 ppm CO₂) and reduced affinity for CO₂ uptake, whereas the other mutants lacking ndhD (ndhD1, ndhD2, ndhD4, and ndhD5 mutants) did not show this mutant phenotype (Table 2 and Fig. 7). The K_0.5 value for CO₂ uptake was higher in the ndhD3 mutant than in the WT (Fig. 7), which may explain the low growth rate of ndhD3 at pH 6.5 and 50 ppm CO₂ (Table 2). Similar results have recently been reported with a mutant of Synechococcus sp. strain PCC 7002 in which the ndhD3 gene was inactivated (5). None of the ndhD mutants showed the accentuated phenotype of M55 (i.e., complete loss of CO₂ uptake activity). Recently, we found that an ndhD3-ndhD4 double mutant did not show any CO₂ uptake activity, although the mutant grew as fast as the WT under photoheterotrophic conditions at atmospheric levels of CO₂ (unpublished results). It appears evident that there are functionally distinct multiple NDH-1 complexes and that an NDH-1 complex having ndhD3 or ndhD4 as a subunit is essential to CO₂ uptake. The exact function of this NDH-1 complex in CO₂ uptake is not known and is being explored.

The information presented in this paper, however, does allow for limited speculation on how a particular type of NDH-1 complex might be specifically involved in CO₂ uptake but not HCO₃ uptake. One previous model for CO₂ uptake in cyanobacteria suggested that a vectorial carbonic-anhydrase-like moiety located within a notional plasma membrane-based transporter could function in the transport of CO₂ by providing a localized source of hydroxyl ions to drive the conversion of CO₂ to HCO₃ followed by release of HCO₃ on the cytosolic face of the transporter (23). In this model, the CO₂ transporter would operate as a type of active, facilitated diffusion and the necessary OH ions were envisaged to be supplied by some component of the respiratory electron transport chain. A new speculative model has arisen (4) that postulates that a type of facilitated diffusion could operate at the level of the thylakoid and result in active accumulation of HCO₃. Here, a vectorial carbonic-anhydrase-like reaction could be closely associated with an NDH-1 complex such that OH ions might be produced in a "localized pocket" and used to drive the conversion of CO₂ to HCO₃. In this model, initial entry of CO₂ into the cell would be passive, whereas direct HCO₃ uptake would occur via separate transporters, such as the cmp transporter, BCT1 (22). The results of this paper are consistent with this new model in that elimination or mutation of all potential NDH-1 complexes (i.e., ndhB modification) or modification of specific subcomplexes containing NdhD3 or NdhD4 result in partial or complete loss of CO₂ uptake capacity without major effects on HCO₃ uptake. The recent finding that the NdhB protein is missing from highly purified cytoplasmic membranes of WT cells (results not shown) but is present in thylakoid preparations (Fig. 5) is also consistent with this new thylakoid-based model for CO₂ uptake.