Halotolerant Cyanobacterium Aphanothece halophytica Contains an Na (cid:1) /H (cid:1) Antiporter, Homologous to Eukaryotic Ones, with Novel Ion Specificity Affected by C-terminal Tail*

Recently, a cyanobacterium Synechocystis sp. PCC 6803 has been shown to contain an Na (cid:1) /H (cid:1) antiporter gene homologous to plants ( SOS1 and AtNHX1 from Arabidopsis ) and mammalians ( NHE s from human) but not to Escherichia coli ( nha A and nha B). Here, we examined whether a halotolerant cyanobacterium Aphanothece halophytica has homologous genes. It turned out that A. halophytica contains an Na (cid:1) /H (cid:1) antiporter homologous to plants, mammalians, and some bacteria ( nha P from Pseudomonas and synnha P from Synechocystis ) but with novel ion specificity. Its gene product, ApNhaP (Na (cid:1) /H (cid:1) antiporter from Aphanothece halophytica ), exhibited the Na (cid:1) /H (cid:1) antiporter activity over a wide pH range between 5 and 9 and complemented the Na (cid:1)

In a previous study (9), we showed that a cyanobacterium Synechocystis sp. PCC 6803 contains an Na ϩ /H ϩ antiporter, SynNhaP, 1 homologous to eukaryotic and prokaryotic (NhaP from Pseudomonas aeruginosa) ones but not to the NhaA, NhaB, and ChaA. It was also shown that the SynNhaP contains a conserved Asp 138 in the transmembrane (TM)-spanning region and a relatively long C-terminal hydrophilic tail, which are important for the carrier activity. The long C-terminal tails are believed to play a role in the regulation of transport activity in animals (2,10). These facts suggest that the cyanobacterial antiporters would provide a model system for the study of the structural and functional properties of eukaryotic Na ϩ /H ϩ antiporters.
To date, only a few functional residues, especially the residues involved in the cation transport, have been identified in Na ϩ /H ϩ antiporter proteins. The importance of Asp 138 in SynNhaP (9) and Asp 133 , Asp 163 , and Asp 164 in NhaA (11,12) has been reported. The Na ϩ /H ϩ antiporter has been thought to exchange specifically between H ϩ and Na ϩ or Li ϩ , but some antiporters exhibited low exchange activity between H ϩ and Li ϩ (13). The E. coli ChaA has been reported to have an exchange activity between Ca 2ϩ and H ϩ as well as Na ϩ and H ϩ at alkaline pH (14), but this has not been examined in detail. It is not clear by which factors ion specificity is determined. To identify the conserved sequences for the cation transport in Na ϩ /H ϩ antiporters, the cloning of an antiporter with novel ion specificity is a prerequisite.
Aphanothece halophytica is a halotolerant cyanobacterium that can grow in a wide range of salinity conditions from 0.25 to 3.0 M NaCl (15,16). A. halophytica accumulates an osmoprotectant glycine betaine at high salinity (15). The DnaK protein of A. halophytica has been shown to contain a longer C-terminal segment than other DnaK/Hsp 70 family members (17) and exhibit extremely high protein folding activity at high salinity (18). Transformation of tobacco with the DnaK from A. halophytica was shown to enhance the tolerance for salt (19). There-* This work was supported in part by grants-in-aid for scientific research from the Ministry of Education and Science and Culture of Japan, the High-Tech Research Center of Meijo University. 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.
The nucleotide sequence(s) reported in this paper has been submitted to the DDBJ/GenBank TM  fore, it was of interest to examine whether A. halophytica contains a unique Na ϩ /H ϩ antiporter. Here, we isolated an Na ϩ /H ϩ antiporter gene, homologous to eukaryotic ones but with novel ion specificity. It was also shown that the ion specificity was affected by the C-terminal tail.

MATERIALS AND METHODS
A. halophytica Culture Conditions-A. halophytica cells were grown photoautotrophically in BG11 medium plus 18 mM NaNO 3 and Turk Island salt solution as described previously (15)(16)(17) except that the NaCl concentration of the culture medium was adjusted to a range from 0.25 to 2.5 M as desired. Cotton-plugged 500-ml conical flasks containing 200 ml of medium each were used and shaken on a reciprocal shaker without supplementation of condensed CO 2 gas. The culture flasks were incubated at 28°C under continuous fluorescent white light (30 microeinsteins m Ϫ2 s Ϫ1 ).
Isolation of A. halophytica Na ϩ /H ϩ Antiporter-A. halophytica genomic DNA was isolated as described (17). Partially degenerate oligonucleotides were designed based on two highly conserved polypeptide regions among several Na ϩ /H ϩ antiporters (9). The forward primer, NP-F, was designed after the polypeptide stretch Phe-Leu-Pro-Pro-Leu-Leu-Phe-Glu-Ala (residues 73 to 81 in SynNhaP). The reverse primer, NP-R, contains the complementary sequence corresponding to the stretch Glu-Gly-Glu-Ser-Leu-Phe-Asn-Asp-Gly (residues 160 to 168 in SynNhaP). The sequences of all the primers are shown in Fig. 1A. Amplified DNA fragments of an expected size (ϳ0.3 kilobase pairs) were obtained and sequenced. Using the determined nucleotide sequence, the adjacent unknown regions of DNA were amplified by the inverse polymerase chain reaction (PCR) method (20). For this, DNA fragments obtained by partial digestion with Sau3AI or complete digestion with EcoRI or AseI were size-fractionated by agarose gel electrophoresis. The fragments from 1-5 kilobase pairs were self-ligated using a ligation kit (Takara Shuzo Co., Shiga, Japan), and then the specific DNA fragment was amplified and sequenced. By repeating this, the nucleotide sequence of about 2.4 kilobase pairs, which covers the whole sequence of the Na ϩ /H ϩ antiporter gene (apnhaP) of A. halophytica, was determined. The DNA sequence was determined using an ABI310 genetic analyzer (Applied Biosystems, Foster City, CA) and analyzed with the DNASIS program (Hitachi Software Engineering, Kanagawa, Japan).
Construction of Expression Plasmids-The coding region of apnhaP was isolated by the PCR reaction. The forward primer, ANP-NcoI, contains the start codon ATG and NcoI site. The reverse primer, ANP-HindIII, contains the HindIII restriction site. The amplified fragment was ligated into NcoI/HindIII sites of the pTrcHis2C plasmid. The resulting plasmid, pANhaP, encodes the ApNhaP fused in-frame to six histidines and was transferred first to E. coli DH5␣ cells and then to TO114 cells in which nhaA, nhaB, and chaA genes were deleted (21).
Construction of Chimera between ApNhaP and SynNhaP-A chimeric gene that encodes the TM region of ApNhaP (residues Met 1 -Ile 401 ) and the C-terminal cytosolic region of SynNhaP (residues Gln 401 -Ser 527 ) was constructed as follows (Fig. 1). The nucleotides corresponding to the TM region of ApNhaP and C-terminal region of SynNhaP were amplified by the forward/reverse primer sets, ANP-NcoI/ANP-N-R and SNP-C-F/SNP-C-R, respectively. The template for the former nucleotide was the pANhaP, and that for the latter was the pSNhaP in which the synnhaP gene was ligated into the NcoI/EcoRI site of pTrcHis2C (9). The forward primer, SNP-C-F, contains the sequence corresponding to the polypeptide stretches Gln 398 -Thr-Val-Ile 401 of ApNhaP and Gln 401 -Phe-Val-Leu 404 of SynNhaP, and the reverse primer, ANP-N-R, is the complementary sequence of SNP-C-F. The reverse primer, SNP-C-R, contains the EcoRI site just before the stop codon of pSNhaP. Upon mixing both PCR products, a chimeric gene apsynnhaP was amplified using the forward ANP-NcoI and reverse SNP-C-R primers and ligated into the NcoI/EcoRI-digested sites of pTrcHis2C, which generated the plasmid pASNhaP. For the construction of a chimeric gene that encodes the TM region of SynNhaP (residues Met 1 -Thr 400 ) and the C-terminal region of ApNhaP (residues Glu 402 -Glu 521 ), the respective nucleotides were amplified by the following forward/reverse primer sets, SNP-N-F/SNP-N-R and ANP-C-F/ ANP-HindIII. Here, the reverse primer, SNP-N-R, is the complementary sequence of ANP-C-F. Upon mixing both PCR products, a chimeric gene synapnhaP was amplified using the forward SNP-N-F and reverse ANP-HindIII primers and ligated into the NcoI/HindIII-digested sites of pTrcHis2C, which generated the plasmid pSANhaP.
Na ϩ /H ϩ Antiporter Activity-The Na ϩ /H ϩ antiporter activity was examined on everted membrane vesicles prepared from the cells grown in LBK medium (Luria broth with KCl instead of NaCl) as described (9,22). Briefly, E. coli cells were harvested by centrifugation at 3,000 ϫ g for 10 min at 4°C and then washed with a suspension buffer TCDS (10 mM Tris-HCl (pH 7.5), 0.14 M choline chloride, 0.5 mM dithiothreitol, and 0.25 M sucrose). The pellet was suspended in 10 ml of TCDS buffer and applied to a French pressure cell (4,000 p.s.i.). Then, the solution was centrifuged at 12,000 ϫ g for 10 min at 4°C. The supernatant was centrifuged at 110,000 ϫ g for 60 min at 4°C, and the pellet was suspended in 600 l of TCDS buffer. The antiporter activity was esti- mated from the changes of ⌬pH (transmembrane pH gradient) by the addition of salt to the reaction mixture, which contained 10 mM Tris-HCl (titrated with HCl to the indicated pH), 5 mM MgCl 2 , 0.14 M choline chloride, 1 M acridine orange, and membrane vesicles (50 g of protein) in a volume of 2 ml. The ⌬pH was monitored with acridine orange fluorescence, which was obtained by excitation at 492 nm and emission at 525 nm. Before the addition of salt, Tris-DL-lactate (2 mM) was added to initiate fluorescence quenching (Q) due to respiration. Lactate energizes the vesicle and accumulates H ϩ intravesiculary, which causes the accumulation of dye and fluorescence quenching (Q). Upon the addition of salt (5 mM), fluorescence increases because of the excretion of H ϩ by antiporters and dequenched fluorescence (⌬Q) were monitored. Then, NH 4 Cl (25 mM) was added to dissipate ⌬pH.
Computer Analysis and Other Methods-The hydropathy profile of the deduced amino acid sequence was predicted according to the method of Kyte and Doolittle (23). The possible TMs of the ApNhaP sequence was deduced by a computer program TopPredII (24). SDS-PAGE and immunoblotting were carried out as described previously (17). Protein was determined by the method of Lowry et al. (25). An antibody raised against 6ϫ histidine (His 6 tag) was obtained from R&D Systems Inc. (Minneapolis, MN).

RESULTS
Cloning of Na ϩ /H ϩ Antiporter from A. halophytica-Using the mixed oligonucleotides of two highly conserved regions among several NhaP-type antiporters, a fragment of the Na ϩ /H ϩ antiporter gene was amplified and sequenced. Then, the upstream and downstream regions of the fragment were amplified by the inverse PCR method (20). The nucleotide sequence of the entire 2431 bp was determined ( Fig. 2A). The sequence analysis revealed one possible open reading frame, apnhaP, spanning positions 796 to 2361. The predicted gene product consists of 521 amino acids with a molecular mass of 56,881 Da (ApNhaP). The upstream sequence of the first Met revealed the presence of Ϫ10 (ATGAAT) and Ϫ35 (TACACT) consensus sequences (Fig. 2B). At further upstream regionspanning positions 185 to 676, another open reading frame encoding a low potential cytochrome c associated with the photosystem II (cytochrome c 550 ) was found. The amino acid sequence of cytochrome c 550 deduced from the nucleotide sequence showed high homology to those from cyanobacteria, algae, and plants (data not shown).
The homology search revealed that ApNhaP is highly homologous to the Na ϩ /H ϩ antiporters from eukaryotes (SOS1, NHEs, AtNHX1, and NHX1) and prokaryotes (NhaP, SynN-haP, and SynNhaP2) as shown in Fig. 3. Here, SynNhaP2 is the second NhaP-type antiporter found in Synechocystis sp. PCC 6803 (9). ApNhaP showed the highest homology to SynNhaP (ϳ61% identity in amino acids). Analysis of the hydropathy plot (23) and the TM prediction program (24) predicted 11 putative TM segments in ApNhaP (Fig. 3A). The Asp 139 in ApNhaP was the conserved ionic amino acid in the membrane-spanning region whose importance for the exchange activity has been shown by the mutagenesis analysis of SynNhaP (9). The ApNhaP contained a long hydrophilic C-terminal tail, which is the case in eukaryotic Na ϩ /H ϩ antiporters but not in the NhaP from P. aeruginosa. These data indicate that a halotolerant cyanobacterium A. halophytica contains an Na ϩ /H ϩ antiporter homologous to eukaryotic ones.
ApNhaP Complements Na ϩ -sensitive E. coli Mutant-To characterize the molecular properties of ApNhaP, the apnhaP gene was ligated into the pTrcHis2C plasmid. The resulting plasmid, pANhaP, as well as pSNhaP, was expressed in the E. coli host cells (TO114) in which NhaA, NhaB, and ChaA were absent (9,21). Western blotting analysis of the membrane fractions revealed that the E. coli cells transformed with pANhaP and pSNhaP exhibited a single cross-reaction band corresponding to ϳ53 kDa, whereas the E. coli cells transformed with the vector alone (pTrcHis2C) did not show any band (data not shown). These results indicate that the ApNhaP could be expressed and assembled in E. coli membranes. Fig. 4A shows that the E. coli TO114 cells transformed with pANhaP can grow with a similar rate to that of the TO114 cells transformed with pTrcHis2C and pSNhaP in LBK medium at pH 7.0 and 37°C. However, because of the absence of Na ϩ /H ϩ antiporter genes (nhaA, nhaB, and chaA) in TO114 cells, the E. coli cells transformed with pTrcHis2C could not grow in the presence of 0.2 M NaCl (Fig. 4B). In contrast, the E. coli cells transformed with pANhaP or pSNhaP could grow (Fig. 4B). Interestingly, the TO114 cells transformed with pANhaP grow slightly faster than those transformed with pSNhaP (Fig. 4B). Fig. 4B suggests the different complementation ability between ApNhaP and SynNhaP. Thus, we examined other conditions that cause more clear differences in the complementation ability in these two antiporters. Fig. 5A shows that at the growth temperature of 30°C, the TO114 cells transformed with pANhaP could grow slightly faster than those transformed with pSNhaP at alkaline pH in LBK medium. Under these conditions, ApNhaP could complement the salt-sensitive phenotype of TO114 cells when the growth medium contained 0.2 M NaCl, whereas SynNhaP could not (Fig. 5B). TO114 cells transformed with pANhaP could even grow in the presence of 0.5 M NaCl (Fig. 5C). These differences in complementation ability were not observed at lower pH values (data not shown). The above results indicate that ApNhaP could complement more efficiently than SynNhaP at alkaline pH.
Novel Ion Specificity of ApNhaP Measured by Activity in the Everted Membrane Vesicles-To examine directly the antiporter activity of ApNhaP, the everted membrane vesicles were prepared, and their antiporter activities were monitored by measuring the lactate-induced fluorescence quenching (Q) and salt-induced fluorescence dequenching (⌬Q). As shown in the upper panel of Fig. 6, the dequenching (⌬Q x100/Q) was observed upon the addition of NaCl in the ApNhaP-expressing cells but not in the control (pTrcHis2C) cells, indicating that ApNhaP has Na ϩ /H ϩ antiporter activity. Fig. 6A shows that the dequenching of fluorescence by ApNhaP was observed over a wide pH range between 5 and 9, which is similar to dequench- ing by SynNhaP (9) but quite different from that by E. coli NhaA. In E. coli NhaA, the antiporter activity could not be observed below pH 7.5, whereas the activity increased with increasing pH (1).
To examine the ion specificity of ApNhaP, the dequenching was measured upon the addition of different cations. In contrast with most Na ϩ /H ϩ antiporters, which could catalyze the exchange between Li ϩ and H ϩ , ApNhaP showed virtually no activity of the Li ϩ /H ϩ antiporter at various pH values (Fig. 6B). Moreover, it was found that ApNhaP could exhibit Ca 2ϩ antiporter activity at neutral or alkaline pH (Fig. 6C), whereas only small or no antiporter activity was observed for K ϩ /H ϩ or Mg 2ϩ /H ϩ , respectively (data not shown). These results indicate that ApNhaP is an antiporter with novel ion specificity.
ApNhaP Complements Ca 2ϩ -sensitive but Not Li ϩ -sensitive E. coli Mutant-Since the above results indicate that ApNhaP has an exchange activity between Ca 2ϩ and H ϩ , but not between Li ϩ and H ϩ , we further examined whether these ion specificities could be demonstrated in the complementation of salt-sensitive phenotypes of E. coli mutant cells. As shown in Fig. 7A, the TO114 cells expressing ApNhaP could not grow in LBK medium containing 4 mM LiCl at pH 8.0, whereas the TO114 cells expressing SynNhaP could grow. In contrast, Fig.  7B indicates that the TO114 cells expressing ApNhaP could grow well in the TrisE medium containing 0.1 M CaCl 2 at pH 8.0 (26), but the TO114 cells expressing SynNhaP grew slowly. In both cases, the control cells (vector only) could not grow (Fig.  7, A and B). These results are consistent with the results of the Li ϩ /H ϩ and Ca 2ϩ /H ϩ antiporter activities of ApNhaP and SynNhaP (Fig. 6, B and C), although the growth rate of SynN-haP-expressing cells in Ca 2ϩ -containing medium appeared to be slower than that anticipated from the Ca 2ϩ /H ϩ exchange activity of SynNhaP (Figs. 7B and 6C).
C-terminal Domain Plays a Role for Ion Specificity of ApN-haP and SynNhaP-Previously, it was shown that the partial deletion of the C-terminal tail decreased the Na ϩ /H ϩ antiporter activity of SynNhaP (9). Although ApNhaP showed the highest homology to SynNhaP, the charges on the C-terminal tail of these two antiporters are significantly different, 22 basic and 14 acidic amino acids in ApNhaP and 15 basic and 24 acidic amino acids in SynNhaP. Therefore, we constructed the chimeric antiporters, ASNhaP and SANhaP, in which the long C-terminal tails of ApNhaP and SynNhaP were replaced with those of SynNhaP and ApNhaP, respectively, as shown in Fig.  8. The exchange activities of chimeras were measured using the everted membrane vesicles. Fig. 9 shows that the ASNhaP chimera exhibited comparable Na ϩ /H ϩ , Li ϩ /H ϩ , and Ca 2ϩ /H ϩ exchange activities with those of the parental ApNhaP, which were shown in Fig. 6. Interestingly, the Li ϩ /H ϩ exchange activity, which was virtually nondetectable in ApNhaP (Fig. 6B), could be clearly observed in the ASNhaP chimera (Fig. 9B). On the other hand, the SANhaP chimera showed reduced exchange activities of both Na ϩ /H ϩ and Li ϩ /H ϩ as compared with the parental SynNhaP ( Fig. 9 versus Fig. 6). Furthermore, the Ca 2ϩ /H ϩ exchange activity in the SANhaP chimera appeared to be pH-dependent, i.e. increased activity was observed at neutral or alkaline pH (Fig. 9C), which was in contrast with that observed for the parental SynNhaP, showing relatively unchanged activity irrespective of the pH (Fig. 6C). These results suggest that the C-terminal region plays a role for the ion specificity of ApNhaP and SynNhaP, although some activity data do not reconcile with this viewpoint. For example, the SANhaP chimera had low Na ϩ /H ϩ exchange activity compared with that of ApNhaP and SynNhaP (Figs. 9A and 6A), and the Ca 2ϩ /H ϩ exchange activity of ASNhaP chimera was pH-dependent, which is the same as that for the parental ApNhaP but different from that for the SynNhaP (Fig. 9C). These results suggest that the ion specificities of ApNhaP and SynNhaP are also affected by the structures in TM regions, as is assumed in most cases of Na ϩ /H ϩ antiporters (1).
The exchange activity data of chimeras were further substantiated by the complementation analysis. As shown in Fig.  10, the ASNhaP chimera could complement the Na ϩ -, Li ϩ -, and Ca 2ϩ -sensitive phenotypes of the E. coli mutant, whereas the SANhaP chimera could hardly complement the Na ϩ -and Li ϩsensitive phenotype of the E. coli mutant. The growth rate of E. coli mutant cells expressing SANhaP was slower than that anticipated from the Ca 2ϩ /H ϩ exchange activity of the SANhaP chimera (Figs. 10C and 9C). DISCUSSION We could isolate an Na ϩ /H ϩ antiporter gene homologous to eukaryotic ones from a halotolerant cyanobacterium A. halophytica. The A. halophytica antiporter gene exhibits the highest homology to the synnhaP and encodes a polypeptide consisting of 521 amino acids. The hydropathy plot and TM prediction analysis of ApNhaP suggested the presence of 11 TM segments and a relatively long C-terminal tail in cytosolic space. One of the important ionic amino acids, Asp 139 in ApN-haP, was conserved in the membrane-spanning region.
Based on the findings that the antiporter-deficient E. coli TO114 mutant cells became salt-tolerant by transformation with the apnhaP gene (Fig. 4) and also by the direct observation of Na ϩ /H ϩ antiporter activity in the transformant membrane vesicles (Fig. 6), it was concluded that the apnhaP encodes the Na ϩ /H ϩ antiporter ApNhaP. The most striking functional feature of ApNhaP is its novel ion specificity. The ApNhaP did not show the Li ϩ /H ϩ antiporter activity (Fig. 6B) but did show the Ca 2ϩ /H ϩ antiporter activity (Fig. 6C). These conclusions were also substantiated by the observations that the ApNhaP did not complement the Li ϩ -sensitive phenotype of the E. coli mutant (Fig. 7A) but complemented the Ca 2ϩsensitive phenotype of the E. coli mutant (Fig. 7B).
It has been reported that the E. coli ChaA has proton/cation exchange activity with Na ϩ or Ca 2ϩ but not with Li ϩ or K ϩ (14,21), which is essentially the same ion specificity as that of ApNhaP. However, the ChaA did not show any homology to the ApNhaP. The ChaA has the acidic motif Glu 200 -His-Glu-Asp-Asp-Ser-Asp-Asp-Asp-Asp 209 conserved in several Ca 2ϩ -binding proteins such as calsequestrin, calreticulin, and Na ϩ /Ca 2ϩ exchanger (14), whereas the ApNhaP lacks the acidic motif. The hydropathy plot of ChaA suggests the absence of a long hydrophilic C-terminal tail seen with the ApNhaP (data not shown). The ApNhaP did not show any homology to the vacuolar Ca 2ϩ /H ϩ exchangers from yeast (27) and plants (28). All these data suggest that the ApNhaP is an Na ϩ /H ϩ antiporter with a different structure from that of ChaA.
The data of Fig. 9 show that the exchange of long C-terminal tails of ApNhaP and SynNhaP with those of SynNhaP and ApNhaP greatly affected the ion specificity of the antiporter, especially that of the Li ϩ /H ϩ exchange activity. These data suggest that exchange of the long C-terminal tail could disrupt the structural elements of the antiporter that are critical for ion-specific binding. A remarkable difference between the Cterminal tails of ApNhaP and SynNhaP is the net charges, 22 basic and 14 acidic amino acids in ApNhaP as compared with 15 basic and 24 acidic amino acids in SynNhaP. An intriguing possibility is that the negative charges on the C-terminal tail of SynNhaP help the binding of Li ϩ, which has a smaller ion radius and consequently a higher positive charge density than Na ϩ , on the membrane-spanning region of the antiporter, thereby allowing the Li ϩ /H ϩ exchange activity. However, the data of Ca 2ϩ /H ϩ exchange activity between parental and chimeric antiporters (Figs. 6C and 9C) do not reconcile with the above viewpoint, suggesting that TM region(s) may also be important for the ion specificity of Na ϩ /H ϩ antiporters. Since the functional role of the C-terminal domain is relevant to its topology, the topological study on ApNhaP might be interesting. In fact, although the C-terminal domains of Na ϩ /H ϩ antiporters have generally been assumed to be entirely cytosolic, exposure of at least some portion of C-terminal domains to the periplasm has been demonstrated (29,30). The construction of various mutants with alteration in the C-terminal hydrophilic region of ApNhaP is needed to elucidate the mechanism under-lying the ion specificity of ApNhaP. In addition, the construction of additional chimera in which chaA from E. coli that lacks a C-terminal tail serving as a parental antiporter may give further insight into the role of the C-terminal tail.
It is worthwhile to note that there is an apparent disparity of the data between the exchange activity and the growth rate. The difference of growth rates between ApNhaP-and SynN-haP-expressing cells (Fig. 7B) was much greater than the difference of Ca 2ϩ /H ϩ antiporter activities between ApNhaP and SynNhaP (Fig. 6C). In the case of the complementation of Ca 2ϩ -sensitive E. coli cells (Fig. 7B), Ca 2ϩ must be excluded from the cells against the concentration gradient of Ca 2ϩ . This activity was much higher in ApNhaP-expressing cells as compared with SynNhaP-expressing cells. In contrast, for Ca 2ϩ /H ϩ exchange, the activity was measured using the everted membrane vesicles subjected to the concentration gradient of Ca 2ϩ . The difference in this activity between ApNhaP and SynNhaP was not so high (Fig. 6C). Thus, the experimental conditions are different between the exchange activity measurement and the complementation test, at least on the orientation of vesicles, Ca 2ϩ concentration gradient, and pH at the H ϩ binding site. These different experimental conditions might be the cause, at least partly, for the apparent disparity between the exchange activity and the growth rate, although the exact nature of this disparity is still unknown.
The data of Figs. 4 and 5 show that the complementation ability of ApNhaP is more effective than that of SynNhaP. This finding would suggest an interesting application of ApNhaP for the genetic engineering of salt-tolerant plants. Previously, we showed that the DnaK from A. halophytica exhibits in vitro much higher refolding activity than that of the DnaK from freshwater cyanobacterium (18). The transformation of tobacco plants by A. halophytica DnaK conferred salt tolerance (19) as well as high temperature tolerance (31). Since it has been reported that the transformation of Arabidopsis by the vacuoletype Na ϩ /H ϩ antiporter AtNHX1 from Arabidopsis could confer salt tolerance of Arabidopsis (32), further studies aimed at constructing the transgenic plants using ApNhaP would enable us to obtain improved salt-tolerant plants. Transfer of multiple salt-tolerant genes from A. halophytica into plants will provide an interesting example to construct the salt-tolerant plants.
In conclusion, we could isolate a eukaryotic Na ϩ /H ϩ antiporter gene from halotolerant cyanobacterium A. halophytica. ApNhaP exhibited the Na ϩ /H ϩ antiporter activity over a wide pH range. The ion specificity of ApNhaP is unique. The ApN-haP did not show any activity of the Li ϩ /H ϩ antiporter but had high Ca 2ϩ /H ϩ antiporter activity. The replacement of a long C-terminal tail of ApNhaP with that of Synechocystis altered the ion specificity of antiporter. The Na ϩ /H ϩ antiporter activity was activated by salt shock as well as by an osmoprotectant, betaine. Thus, ApNhaP would provide a unique system for the study of the ion specificity of eukaryotic Na ϩ /H ϩ antiporters and also an application for the genetic engineering of salttolerant plants.