Two Types of Functionally Distinct NAD(P)H Dehydrogenases inSynechocystis sp. Strain PCC6803*

The ndhD gene encodes a membrane protein component of NAD(P)H dehydrogenase. The genome ofSynechocystis sp. PCC6803 contains 6 ndhDgenes. Three mutants were constructed by disrupting highly homologousndhD genes in pairs. Only theΔndhD1/ΔndhD2(ΔndhD1/D2) mutant was unable to grow under photoheterotrophic conditions and exhibited low respiration rate, although the mutant grew normally under photoautotrophic conditions in air. The ΔndhD3/ΔndhD4(ΔndhD3/D4) mutant grew very slowly in air and did not take up CO2. The results demonstrated the presence of two types of functionally distinct NAD(P)H dehydrogenases in Synechocystis PCC6803 cells. TheΔndhD5/ΔndhD6(ΔndhD5/D6) mutant grew like the wild-type strain. Under far-red light (>710 nm), the level of P700+was high in ΔndhD1/D2 and M55 (ndhB-less mutant) at low intensities. The capacity ofQ A (tightly bound plastoquinone) reduction by plastoquinone pool, as measured by the fluorescence increase in darkness upon addition of KCN, was much less inΔndhD1/D2 and M55 than inΔndhD3/D4 andΔndhD5/D6. We conclude that electrons from NADPH are transferred to the plastoquinone pool mainly by the NdhD1·NdhD2 type of NAD(P)H dehydrogenases.

The Type I NAD(P)H dehydrogenase complex (NDH-1) 1 in cyanobacteria is involved in both the respiratory and photosynthetic electron transport chains (1). The whole genome sequence data base for Synechocystis sp. PCC6803 has shown the presence of genes for 12 subunits of NDH-1 with the large, hydrophobic NdhB, NdhD, and NdhF subunits being core membrane components (2). The data base also reveals that ndhD and ndhF are present as gene families with six and three members, respectively (note that NdhF4 has homology to NdhD5 and has been designated as NdhD6 (3)), although most ndh genes are present as single copies. This suggests that several types of NDH-1 exist in cyanobacteria, each with different NdhD and/or NdhF subunits, and with each potential complex having differing functions (4 -6). In fact, of the five ndhD-less mutants, ⌬ndhD3 is the only mutant that displays the phenotype of slow growth at limiting CO 2 (i.e. 50 ppm CO 2 ) and reduced affinity for CO 2 uptake, whereas the other ndhDless mutants (⌬ndhD1, ⌬ndhD2, ⌬ndhD4, and ⌬ndhD5) do not show such phenotype (6).
It has been demonstrated that NDH-1 is essential for inorganic carbon (CO 2 and HCO 3 Ϫ ; designated C i ) transport in cyanobacteria (3)(4)(5)(6)(7)(8)(9)(10)(11), and it was assumed that ATP produced by NDH-1-dependent cyclic electron flow is essential to energize the C i transport (7, 12). However, a recent observation indicated that mutations in ndh genes lead to inhibition of CO 2 uptake rather than HCO 3 Ϫ uptake (6). This suggested that CO 2 uptake is energized differently from HCO 3 Ϫ uptake. The presence of an ATP-dependent HCO 3 Ϫ transporter in Synechococcus PCC7942 has been recently demonstrated (13). In an attempt to see if there are functionally distinct NDH-1 complexes, we constructed double mutants of Synechocystis sp. PCC6803 by disrupting highly homologous ndhD genes in pairs and analyzed their growth under various conditions, CO 2 uptake and redox levels of P700 and plastoquinone (PQ) pool. The results suggested the presence of two types of functionally distinct NDH-1 complexes, one essential for photoheterotrophic growth of the cells and the other essential for CO 2 uptake.

EXPERIMENTAL PROCEDURES
Growth Conditions-Wild-type and mutant cells of Synechocystis sp. PCC6803 were grown at 30°C in BG11 medium (14) buffered with 20 mM Tes-KOH (pH 8.0) and bubbled with 3% CO 2 in air (v/v) or air (about 400 l of CO 2 liter Ϫ1 ). Solid medium was BG11-supplemented with 1.5% agar, 5 mM sodium thiosulfate, and 20 mM of the same buffer. Continuous illumination was provided by fluorescent lamps, generating photosynthetically active radiation of 60 mol of photons m Ϫ2 s Ϫ1 .
Construction of Mutants-Construction of single ndhD mutants, e.g. slr0331 (ndhD1), slr1291 (ndhD2), sll1733 (ndhD3), sll0027 (ndhD4), and slr2007 (ndhD5)-less mutants, has been described in a previous study (6). Descriptions of these mutants and the slr2009 (ndhD6)-less mutant have been deposited on the Web (CyanoMutants), where the drug resistance cassette used for each inactivation and the sites of insertion into the target genes are shown. The constructs used to generate the single mutants were used to transform various appropriate single mutants of Synechocystis sp. PCC6803 to generate the double mutants, i.e. ⌬ndhD1/⌬ndhD2 (⌬ndhD1/D2), ⌬ndhD3/⌬ndhD4 (⌬ndhD3/D4), and ⌬ndhD5/⌬ndhD6 (⌬ndhD5/D6). M55 is the mutant constructed by inserting a Km r cartridge at the BamHI site in ndhB, as described previously (7). The mutated ndhD genes in the transformants were segregated to homogeneity (by successive streak purification) as determined by polymerase chain reaction amplification.
Determination of Growth Characteristics-Wild-type and mutant strains grown under 3% CO 2 were collected and resuspended in fresh BG11 medium to an OD 730 nm of 1.0, 0.1, or 0.01. 2 l of the cell suspensions was spotted onto BG11 agar plates buffered at various pHs in the absence or presence of 10 M 3-(3,4-dichlorophenyl)-1,1-dimethylurea (DCMU) and 5 mM glucose. The plates were incubated under 3% CO 2 in air (v/v) or air for 5 days with continuous illumination by fluorescent lamps under a photosynthetically active radiation intensity of 60 mol of photons m Ϫ2 s Ϫ1 . The OD 730 nm was measured using a recording spectrophotometer, model UV2200 (Shimadzu Co., Kyoto, * This study was supported by a Grant-in-aid for Scientific Research (B) (2) (12440228), by a grant for "Research for the Future" Program (JSPS-RFTF97R16001), and by a grant from the Human Frontier Science Program (RG0051/1997M). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Japan).
Assay of Respiratory Activity-Oxygen uptake was measured in the dark with a Clark type oxygen electrode (Hansatech, United Kingdom). Cells were suspended in 20 mM Tes-KOH (pH8.0) containing 15 mM NaCl, and the activities of respiration were measured as the rates of oxygen uptake 3 min after addition of 5 mM glucose, as described by Mi et al. (12).
CO 2 Exchange Measurements-The wild-type and mutant cells grown under 3% CO 2 in air (v/v) were bubbled with air for 18 h in the light and then harvested. Cells were suspended in 25 ml of 20 mM Hepes-KOH buffer, pH 7.0, containing 15 mM NaCl at a chlorophyll level of 4.3 g ml Ϫ1 and placed in a reaction vessel (15,16). CO 2 exchange of the cell suspension was measured at 30°C using an open gas analysis system, which records the rate of CO 2 exchange as a function of time. Air containing 340 l of CO 2 liter Ϫ1 was passed into the reaction vessel at a flow rate of 1.0 liter min Ϫ1 , the exchanged gas was dried, and then the CO 2 concentration was analyzed using an infrared gas analyzer.
Measurements of the Redox Level of P700 -Changes in levels of P700 ϩ were monitored by measuring the absorbance at 820 nm on an amplitude-modulation (PAM) chlorophyll fluorometer with a dual wavelength P700 unit (ED-700DW-T), as described by Schreiber et al. (17). The monitoring conditions were the same as those described by Mi et al. (12). The oxidation of P700 was induced by illumination with far-red (FR) light (Ͼ710 nm) or a 50-ms pulse of saturating white light (1500 watts m Ϫ2 ). Wild-type and mutant cells used for P700 measurements were grown under 3% CO 2 in air (v/v) to late logarithmic phase followed by incubation in darkness for 32 h.
Chlorophyll Fluorescence Measurements-Time-based fluorescence measurements were performed on a dual modulation kinetic fluorometer (model FL-100, Photon Systems Instruments, Brno, Czech Republic) interfaced with a computer (18). The duration of the measuring flashes was 5 s, and the measurements were performed at room temperature. Wild-type and mutant cells were harvested in the logarithmic growth phase and suspended in fresh BG11 at a chlorophyll concentration of 2 g ml Ϫ1 . The fluorescence level (F 0 ) was measured for 15 s with the measuring flash of 1-s intervals, and then KCN was injected to a final concentration of 1 mM. The measuring flash was kept on at 10-s intervals for the duration of the 6-min time course. The measuring flash itself did not have a noticeable actinic effect. The fluorescence yield values were recorded as a function of incubation time with KCN.
Other Methods-Unless otherwise stated, standard techniques were used for DNA manipulation (19). Pigments in the cells were extracted in methanol, and the concentration of chlorophyll in the extract was determined (20).

RESULTS
Six ndhD Genes-The Synechocystis sp. PCC6803 genome contains six ndhD genes that were denoted as ndhD1 (slr0331), ndhD2 (slr1291), ndhD3 (sll1733), ndhD4 (sll0027), ndhD5 (slr2007), and ndhD6 (slr2009) (3). Table I summarizes the amino acid sequence similarity of the products of these genes. NdhD1, NdhD3, and NdhD5 showed the highest homology to NdhD2, NdhD4, and NdhD6, respectively. Double mutants were constructed by disrupting highly homologous ndhD genes in pairs, i.e. ndhD1/ndhD2, ndhD3/ndhD4, and ndhD5/ndhD6. Growth Characteristics-To explore how the inactivation of ndhD genes affects the growth characteristics of the cells, growth of the mutant strains was examined under photoautotrophic conditions on solid BG11 medium buffered at various pHs in the presence of 3% CO 2 or air as well as under photoheterotrophic conditions on the same medium (pH 8.0) containing 5 mM glucose and 10 M DCMU with air. There was a striking difference between the growth characteristics of ⌬ndhD1/D2 and ⌬ndhD3/D4. The ⌬ndhD1/D2 mutant grew as fast as the wild-type at pH levels of 8.0 and 7.0 in air but was unable to grow under photoheterotrophic conditions (Fig. 1, middle column panels). In contrast, ⌬ndhD3/D4 grew very slowly in air, although the mutant grew as fast as the wild-type under photoheterotrophic conditions. These results demonstrated that there are at least two types of functionally distinct NDH-1 complexes in Synechocystis. One type of NDH-1 is essential for the photoheterotrophic growth of cells, and the other type is essential for growth under low CO 2 concentrations. No significant difference was observed between the wild-type and ⌬ndhD5/D6 strains in their growth characteristics under the conditions examined. Both ⌬ndhD1/D2 and ⌬ndhD3/D4 grew more slowly than the wild-type at pH 6.5 even under 3% CO 2 (Fig. 1, panels C). The growth of ⌬ndhD5/D6 was also slower than the wild-type at pH 6.5 in air. Thus, all the ndhD genes appear to be needed for cells to grow under acidic conditions.
The panels on the right column in Fig. 1 show the growth of four single ndhD-less mutants and M55 in air under photoautotrophic and photoheterotrophic conditions. The ⌬ndhD1 mutant exhibited partially the phenotype of ⌬ndhD1/D2 and grew poorly under photoheterotrophic conditions. On the other hand, ⌬ndhD3 exhibited slow growth in air. However, none of these single mutants showed the severe phenotypes of the double mutants ⌬ndhD1/D2 and ⌬ndhD3/D4. Double mutants constructed by inactivating ndhD genes in different combinations did not show the phenotype of ⌬ndhD1/D2 or ⌬ndhD3/D4 (data not shown). As expected, M55 exhibited the phenotypes of both ⌬ndhD1/D2 and ⌬ndhD3/D4 and did not grow under air as well as under photoheterotrophic conditions (Fig. 1, right  column panels).
Respiration-The rates of respiration of the wild-type and mutant cells in the presence of glucose are summarized in Table II. Glucose enhanced the respiration rates by 50 -400%.  Constant rates of oxygen uptake were attained 0.5 ϳ 2 min after addition of glucose and continued until oxygen in the cell suspension was exhausted. Out of the three double mutants, only ⌬ndhD1/D2 showed significantly reduced respiration rate, which was about one-fourth the wild-type activity but was 40% higher than that of M55. The respiration rates of ⌬ndhD3/D4 and ⌬ndhD5/D6 were not significantly different from the rate of the wild-type. CO 2 Uptake- Fig. 2 shows changes in the rates of CO 2 exchange of the wild-type and mutant cells upon switching the light on and off, as measured using the gas analysis system (15,16). When the suspension of the wild-type cells was illuminated, the rate of CO 2 uptake increased and reached the maximum level in 20 s. An efflux of CO 2 from the cells was observed immediately after turning the light off. Similar CO 2 exchange profiles have been observed with the wild-type cells of Synechococcus PCC7942 (15,16) and Anabaena variabilis (21) and have been shown to reflect the CO 2 uptake in the light by a CO 2concentrating mechanism and the release of intracellular C i as CO 2 after darkening. The ⌬ndhD3/D4 mutant did not take up CO 2 at all in the light, indicating that the CO 2 uptake system is impaired in this mutant. A similar result has been observed with M55 (7) and was confirmed in this study (Fig. 2). The ⌬ndhD1/D2 and ⌬ndhD5/D6 mutants showed similar CO 2 exchange profiles to the wild-type, and their CO 2 uptake activities were not significantly different from the activity in the wild-type. The ⌬ndhD3 mutant showed about 30% of the CO 2 uptake activity in the wild-type under our experimental conditions (340 l of CO 2 liter Ϫ1 ).
Redox Level of P700 under FR Light-In cyanobacteria, NADPH produced by photosystem I (PSI) reaction donates electrons to the PQ pool via NDH-1, thus constituting NDH-1dependent PSI cyclic electron flow (12). The donation of electrons from NADPH to the PQ pool can be observed by measuring the redox changes of P700. Fig. 3 shows the redox levels of P700 in the wild-type and mutant cells as a function of the intensity of FR light. The values in this figure are normalized based on the assumption that P700 was completely oxidized when the cell suspension was illuminated with white flash light (50-s duration) superimposed on strong FR light (98 eq m Ϫ2 s Ϫ1 ). When the cell suspension of M55 or ⌬ndhD1/D2 was illuminated with strong FR light, P700 was oxidized completely and no further oxidation of P700 was observed when the white flash light was superimposed on the FR light. In contrast, oxidation of P700 in the wild-type and other mutants was not complete with the strong FR light alone and further oxidation of P700 was observed with the flash light. P700 in M55 and ⌬ndhD1/D2 was highly oxidized even under weak FR light, whereas the oxidation levels of P700 in the wild-type and other mutants were low under the same conditions. At 3.3 eq m Ϫ2 s Ϫ1 of FR light, the levels of P700 ϩ in wild-type, ⌬ndhD1/D2, ⌬ndhD3/D4, ⌬ndhD5/D6, and M55 strains were 2, 56, 7, 10, and 72%, respectively. Evidently, in mutants M55 and ⌬ndhD1/D2 the reduction of P700 is strongly inhibited and the transport of electrons from NADPH to the PQ pool, and thereby to P700, is mediated mainly by the NdhD1⅐NdhD2 type of NDH-1 complexes. The higher oxidation levels of P700 in ⌬ndhD3/D4 and ⌬ndhD5/D6 than in the wild-type under weak FR light indicate that the NdhD3⅐NdhD4 and NdhD5⅐NdhD6 NDH-1 complexes also have a role in mediating transfer of electrons from NADPH to the PQ pool. The capacity of electron donation from NADPH to the PQ pool by these NDH-1 complexes, however, appears to be much less than that mediated by the NdhD1⅐NdhD2 NDH-1 complexes.
The Redox State of the PQ Pool-The redox state of the PQ pool in thylakoids was monitored indirectly by determining the chlorophyll fluorescence yield in darkness upon addition of KCN (inhibiting oxidase activity). The fluorescence yield depends on the redox state of Q A that is in redox equilibrium with the PQ pool. Therefore, the rate of reduction of the PQ pool can be monitored by measuring the chlorophyll fluorescence yield after oxidation of the PQ pool is blocked by KCN in the dark (22).
The wild-type cells exhibited a rapid increase in the fluorescence yield upon addition of KCN, reflecting an increase in the Q A Ϫ level due to PQ pool reduction. When the wild-type and mutant cells were illuminated in the presence of DCMU, the PCC6803 Cells were suspended in 20 mM Tes-KOH (pH 8.0) containing 15 mM NaCl. Oxygen consumption of the cell suspension was measured in darkness 3 min after addition of glucose (final concentration of 5 mM). Results are shown as means Ϯ SD (n ϭ 4 -7).

FIG. 3. Effects of the FR light intensity on the oxidation level of P700 in the wild-type and mutant cells.
The P700 ϩ levels were normalized assuming that P700 was completely oxidized when a 50-s pulse of saturating white light (1500 watts m Ϫ2 ) was given to the cell suspension superimposed to strong FR light (98 eq m Ϫ2 s Ϫ1 ). Wild-type (f), ⌬ndhD1/D2 (Ⅺ), ⌬ndhD3/D4 (⅜), ⌬ndhD5/D6 (‚), M55 (q). fluorescence yield increased to the level attained with the wildtype cells upon addition of KCN in the dark. The increase of the fluorescence yield was not observed with the M55 cells upon addition of KCN, indicating that NDH-1 is involved in this phenomenon. This explanation contradicts the report that the rapid increase in the fluorescence yield upon addition of KCN reflects electron donation to the PQ pool mediated by succinate: quinol oxidoreductases (SDH) (22). The ndhD mutants exhibited an increase in the fluorescence yield after addition of KCN, but the rate of the increase in ⌬ndhD3/D4 and ⌬ndhD5/D6 was slower than that in the wild-type strain. The rise in fluorescence yield in ⌬ndhD1/D2 was very slow but was much faster than in M55. Thus, the capacity of Q A reduction by PQ pool in ⌬ndhD1/D2 is much less than in ⌬ndhD3/D4 and ⌬ndhD5/D6 mutants but much more than in M55. This confirms the conclusion obtained by measuring the redox level of P700 that the reduction of the PQ pool is mediated mainly by the NdhD1⅐NdhD2 NDH-1 complexes and that the NdhD3⅐NdhD4 and NdhD5⅐NdhD6 NDH-1 complexes also transfer electrons from NADPH to PQ pool. DISCUSSION The ability of cells to grow under photoheterotrophic conditions strongly depends on their respiratory activity. The ⌬ndhD1/D2 and M55 mutants, which exhibited low respiration rates, were unable to grow under photoheterotrophic conditions ( Fig. 1 and Table II). Reduction of the PQ pool and P700 was also strongly inhibited in these mutants (Figs. 3 and 4). This indicated that the NdhD1⅐NdhD2 type of NDH-1 complexes mediates the transport of electrons from NADPH to the PQ pool and thereby to P700, thus constituting a PSI-dependent cyclic electron transport pathway. The NdhD3⅐NdhD4 and NdhD5⅐NdhD6 types of NDH-1 complexes also contribute to this reaction, as seen from the higher oxidation levels of P700 in ⌬ndhD3/D4 and ⌬ndhD5/D6 than in the wild-type under FR light (Fig. 3). The contribution of these complexes of NDH-1 to the PSI-dependent cyclic electron transport appears to be much smaller than that of the NdhD1⅐NdhD2 NDH-1 complexes (Figs. 3 and 4, Table II).
The absence of the rapid increase in the fluorescence yield in ⌬ndhD1/D2 and M55 upon addition of KCN suggested that this fluorescence increase reflects the electron donation to the PQ pool mediated by the NdhD1⅐NdhD2 NDH-1 complexes (Fig. 4). This contradicts the report by Cooley et al. (22) that the rapid fluorescence increase reflects the reduction of the PQ pool mediated by SDH. It appears possible that inactivation of SDH led to decreased formation of NADPH, thereby decreasing the NDH-1-dependent reduction of the PQ pool. Adversely, inacti-vation of NDH-1 might lead to the decrease of succinate formation. Inactivation of ndhB or ndhD1 plus ndhD2 had much stronger effect on the fluorescence increase than the inactivation of sll1625 plus sll0823 ( Fig. 4; also see Fig. 6 in Ref. 22), and hence, we prefer the former possibility.
Among the single and double ndhD mutants constructed in this and previous studies (5,6), only ⌬ndhD3/D4 showed the accentuated phenotype of M55 (i.e. complete loss of CO 2 uptake activity). It is evident that inability of these mutants to grow under low CO 2 conditions is a result of inactivation of their CO 2 uptake activity (Figs. 1 and 2) and that only the NdhD3⅐NdhD4 NDH-1 complexes are essential for CO 2 uptake. We have previously proposed a model in which ATP produced by NDH-1dependent PSI cyclic electron transport drives the "transport" of CO 2 (23). However, inactivation of ndhD1 and ndhD2, which are essential to the PSI cyclic electron transport, did not have any significant effect on CO 2 uptake (Fig. 2). Thus, this previous model appears to be unlikely.
The mechanism of CO 2 uptake and the role of NdhD3⅐NdhD4 NDH-1 complex types in CO 2 uptake are not known. The CO 2 uptake reaction is postulated to be an energy-dependent unidirectional conversion of CO 2 to HCO 3 Ϫ (11, 24). Kaplan and Reinhold proposed a model in which an "alkaline pocket" produced in the light functions as a converter of CO 2 to HCO 3 Ϫ (11). The information presented in this paper does allow for limited speculation on how a particular type of NDH-1 complex might be specifically involved in CO 2 uptake. A vectorial carbonic anhydrase-like reaction could be closely associated with the NdhD3⅐NdhD4 NDH-1 complex type such that OHϪ ions might be produced in a "localized pocket" and used to drive the conversion of CO 2 to HCO 3 Ϫ . Such a pocket could be produced at the site(s) of protonation of NADP and/or PQ, and a relatively high electron transfer rate would be required to account for the high CO 2 uptake rate. The low rate of electron transfer from NADPH to PQ pool in the ⌬ndhD3/D4 mutant would suggest that the NdhD3⅐NdhD4 NDH-1 complexes donate electrons mainly to an alternate electron acceptor. Further studies are in progress to identify the electron acceptor.
In a previous paper, the following phenotypes were reported to characterize the ndhB-less mutant (M55) of Synechocystis sp. strain PCC6803 (7). 1) It grows very slowly under air and is unable to take up CO 2 , 2) it is unable to grow under photoheterotrophic conditions, and 3) it has a low respiration rate. The first phenotype was observed in the ⌬ndhD3/D4 mutant described in this study and the second and third phenotypes were found for the ⌬ndhD1/D2 mutant (Figs. 1 and 2, Table II).
Thus, it appears evident that there are two types of functionally distinct NDH-1 complexes in Synechocystis sp. PCC6803, and M55 exhibits the phenotype of both types of the mutants. The presence of the homologues of the six ndhD genes in the genome of Anabaena sp. PCC7120 (the Web, CyanoBase) and Synechococcus sp. PCC7002 (4) strongly suggests that the presence of the two types of functionally distinct NDH-1 complexes is common among various strains of cyanobacteria.