|Home | About | Journals | Submit | Contact Us | Français|
Attempts to identify members of the antiporter complement of the alkali- and saline-adapted soda lake alkaliphile Alkalimonas amylolytica N10 have used screens of DNA libraries in antiporter-deficient Escherichia coli KNabc. Earlier screens used Na+ or Li+ for selection but only identified one NhaD-type antiporter whose properties were inconsistent with a robust role in pH homeostasis. Here, new screens using elevated pH for selection identified three other putative antiporter genes that conferred resistance to pH ≥ 8.5 as well as Na+-resistance. The three predicted gene products were in the Calcium:Cation Antiporter (CaCA), Cation:Proton Antiporter-2 (CPA2) and Cation:Proton Antiporter-1 (CPA1) families of membrane transporters, and were respectively designated Aa-CaxA, Aa-KefB and Aa-NhaP, reflecting homology within those families. Aa-CaxA conferred the poorest Na+-resistance and also conferred modest Ca2+-resistance. Aa-KefB and Aa-NhaP inhibited growth of a K+ uptake-deficient E. coli mutant (TK2420), suggesting that they catalyzed K+ efflux. For Aa-NhaP, the reversibility of the growth inhibition by high K+ concentrations depended upon an organic nitrogen source, e.g. glutamine, rather than ammonium. This suggests that NH4+ as well as K+ efflux is catalyzed by Aa-NhaP. Vesicles of E. coli KNabc expressing Aa-NhaP, which conferred the strongest alkali-resistance, exhibited K+/H+ antiport activity in a pH range from 7.5 to 9.5, and with an apparent Km for K+ of 0.5 mM at pH 8.0. The properties of this antiporter are consistent with the possibility that the soda lake alkaliphile uses K+(NH4+)/H+ antiport as part of its alkaline pH homeostasis mechanism and part of its capacity to reduce potentially toxic accumulation of cytoplasmic K+ or NH4+, respectively, under conditions of high osmolarity or active amino acid catabolism.
Bacteria that grow in soda lakes must be adapted to both high pH and high concentrations of Na+, as well as stresses that are secondary to these major ones, such as low water activity and scarcity of some micronutrients (Grant & Tindall, 1986; Grant, 2004; Jones et al., 1998). For example, Lake Chahannor, a soda lake in the Inner Mongolia Autonomous region of China, has reported pH values in a range from 9.5-10.2, contains > 5 M Na+, high concentrations of Cl- and CO32-, very low levels of Mg2+ and undetectable levels of Ca2+ (Ma et al., 2004; Zheng, 1992). Alkalimonas amylolytica N10 is a halotolerant, alkaliphilic Gram-negative (γ-Proteobacterium) isolate from this lake (Ma et al., 2004). In an earlier study, we sought to identify A. amylolytica N10 genes encoding cation/proton antiporters that might participate in alkaline pH homeostasis and Na+-resistance (Liu et al., 2005). In other bacteria, a complement of several distinct secondary cation/proton antiporters usually contributes to alkaline pH homeostasis (Padan et al., 2005). Na+(Li+)/H+ antiporters often play dominant roles in both this function and Na+(Li+) –resistance (Padan et al., 2005). K+/H+ antiporters have also been proposed to participate in alkaline pH homeostasis (Kakinuma & Igarashi, 1995; Nakamura et al., 1984; Nakamura et al., 1992; Plack & Rosen, 1980) and recently the genes encoding such antiporters have been identified in both non-marine and marine bacteria (Fujisawa et al., 2005; Lewinson et al., 2004; Radchenko et al., 2006a; Radchenko et al., 2006b). Na+/H+ and K+/H+ antiporters can exchange cytoplasmic cations for external protons to achieve a cytoplasmic pH significantly less alkaline than the external pH (Padan et al., 2005). Genomic data from the haloalkaliphilic archaeon Natronomonas pharaonis from African soda lakes (Falb et al., 2005) suggests the presence of: one Na+/H+ antiporter of the Cation:Proton Antiporter-1 (CPA1) family and two more from the NhaC family in addition to four K+/H+ antiporters of the Cation:Proton Antiporter-2 (CPA2) family and one from the Cation:Proton Antiporter-3 (CPA3) family, as classified in the Transporter Classification system (Ren et al., 2007; Saier et al., 2006). However, the cation specificity, properties and physiological functions of these transporters have yet to be studied experimentally.
The first attempts to identify genes encoding antiporters in DNA libraries from A. amylolytica N10 utilized a screen in the triple antiporter mutant of Escherichia coli strain KNabc (ΔnhaAΔnhaBΔchaA), in which resistance to Na+ or Li+ at pH 7.5 was the basis for identifying candidate plasmids that carry antiporter genes of interest (Liu et al., 2005). This is the most frequently used protocol for isolating genes encoding Na+(Li+)/H+ antiporters from bacteria whose genomes have not been sequenced and which lack a genetic system that allows screening in the natural host via transpositional mutants that are alkali-sensitive (Padan et al., 2005). The initial screens in E. coli KNabc resulted in identification of only one antiporter, an NhaD-type Na+(Li+)/H+ antiporter with an optimum pH ≥ 9.5, exceeding the limits of the assay to determine definitively (Liu et al., 2005). In addition, the optimum [Na+] was unusually high at about 600 mM. These properties suggested that Aa-NhaD may serve an emergency role, i.e. when acute cytoplasmic alkalinization and adverse Na+ elevation have occurred, while other yet to be identified antiporters have important roles in alkaline pH homeostasis and Na+–resistance (Liu et al., 2005). Many halotolerant microorganisms are adapted to high cytoplasmic concentrations of Na+ while others maintain cytoplasmic Na+ levels that are much lower than the high external concentrations (Ventosa et al., 1998). However, even if A. amylolytica N10 is adapted to high cytoplasmic Na+, it seemed unlikely that it grows robustly with both a high cytoplasmic [Na+] and a cytoplasmic pH well above 9 since elevated pH exacerbates Na+-toxicity and elevated Na+ compromises growth at high pH (Padan et al., 2005).
In the current study, screening of a DNA library of A. amylolytica N10 was carried out using the same mutant E. coli KNabc host but using selection based on alkali-resistance, with a view towards broadening the search for antiporters that might contribute to alkali- and/or Na+-resistance. Since E. coli KNabc is deficient in K+(Ca2+)/H+ as well as Na+(Li+)/H+ antiport, this selection might identify antiporters with different cation specificities or Na+/H+ antiporters with a higher affinity for Na+ than Aa-NhaD. Indeed, Aa-NhaD was not identified in this screen but three other genes were identified whose products are predicted to come from three different cation/proton antiporter families of the Transporter Classification system (Ren et al., 2007; Saier et al., 2006). The capacity of each gene to complement several different growth phenotypes was studied in mutant E. coli hosts that have well-defined properties. One of the antiporters, designated Aa-NhaP, was then characterized further in assays of antiport activity in membrane vesicles.
The bacterial strains and plasmids used in this study are shown in Table 1. Alkalimonas amylolytica N10 was grown at 37 °C, with shaking, in Horikoshi I medium (Horikoshi, 1991). E. coli strains DH5α, KNabc and TK2420 were routinely grown in LBK medium at pH 7.5 (Goldberg et al., 1987). In LBC medium, 6g/L of choline-chloride (40 mM) replaced the KCl that is present in LBK; the pH of LBC was adjusted by additions of HCl or BTP. For growth experiments on E. coli KNabc in LBK and LBC at different pH values and levels of added NaCl, cells were grown at 37 °C, with shaking, for 9 hours, after which the A600 was recorded. When antibiotics were added to the medium for selections or plasmid maintenance, they were added at the following concentrations: 100 μg ampicillin ml-1; 50 μg chloramphenical ml-1; and 50 μg kanamycin ml-1.
DNA isolation, cloning and restriction analyses were carried out by standard methods (Sambrook et al., 1989). For construction of a DNA library, chromosomal DNA was isolated from A. amylolytica N10 and partially digested with Sau3AI. Fragments in the range of 3-8 kb were isolated and ligated to BamHI-digested pGEM-3Zf(+) cloning vector (Promega). The ligation mixture was used to transform E. coli DH5α under selection on LBK-ampicillin plates (pH 7.5) at 37°C. After amplification, the mixed plasmid pool from about 5 ×106 colonies was collected, and used to transform E. coli KNabc. The transformant pool was screened for colony-forming cells at pH 8.5 to 9.0 at 37°C on LBK- ampicillin-kanamycin plates. Plasmids in transformants from colonies that arose on these plates were re-tested. After confirmation of their ability to support growth under the same screening condition, complete sequencing was conducted on the inserts of five complementing plasmids that had distinct restriction patterns. These plasmids were designated as pNAK7-11; the sizes of their inserts were, respectively, 4113 bp, 3566 bp, 5412 bp, 3782 bp and 8473 bp. The plasmid insert sequences were further confirmed to match the sequence of PCR products from chromosomal DNA of A. amylolytica N10. DNA sequencing was conducted in the Mount Sinai School of Medicine DNA Core Facility. Computer analyses were performed by using the Gene Runner 3.05 program (Hastings Software, Hastings, N.Y., 1994). Putative open-reading frames were identified by ORF Finder program and homologue searches were conducted using BLASTP (Altschul et al., 1990) and the network services of the National Center for Biotechnology Information (NCBI). The secondary structure and the free energy calculations of putative stem-loop structures were analyzed using the method of Zuker (Zuker, 2003) (http://www.bioinfo.rpi.edu/applications/mfold/old/rna/form3.cgi).
Since part of the pNAK9 insert contained a 2.3 kb fragment that was identical to that in pNAK10, the sequences were appropriately merged. The DNA sequences of the four plasmid inserts were deposited in GenBank under accession numbers: DQ649017 for pNAK7; DQ649018 for pNAK8; DQ649019 for pNAK9/10; and DQ649020for pNAK11. Three of these plasmids, pNAK7, pNAK9/10 and pNAK11, contained genes that were candidates for cation/proton antiporters. These three genes were studied further and their putative products were deposited in GenBank with the following designations and accession numbers: Aa-CaxA, ABG37980; Aa-KefB, ABG37986; and Aa-NhaP, ABG37987.
The three putative antiporter gene ORFs, including their native Shine-Dalgarno (SD) sequences, were cloned behind the T7 promoter of pGEM-3Zf(+) (Promega). The PCR primers listed in Table 1 were used to amplify the genes from A. amylolytica N10 chromosomal DNA by PCR reactions that used the TaKaRa Taq polymerase kit according to the manufacture's instructions. The Aa-caxA gene was cloned as follows: PCR with primer pair pNAK7/EF and pNAK7/XR1 was conducted so that the forward primer contained an EcoRI site upstream of the putative SD sequence and the reverse primer contained an XbaI site downstream of the stop codon of the Aa-caxA coding sequence. The 1.6 kb PCR product was digested by EcoRI and XbaI restriction enzymes, ligated to pGEM-3Zf(+) vector, linearized with the same enzymes and used to transform E. coli DH5α. The resulting recombinant plasmid was designated as pYW136. Similar strategies were used to construct plasmid pYW138 for cloning Aa-kefB with primer pair pNAK9/SF1 (SacI) and pNAK9/BR1 (BamHI), and to construct plasmid pYW139 containing Aa-nhaP using primers pNAK11/EF1 (EcoRI) and pNAK11/XR1(XbaI). The sequence of all the recombinant plasmids was confirmed to be correct.
The recombinant plasmids pYW136, pYW138 and pYW139 were transformed into Na+(K+)(Ca2+)/H+ antiporter-deficient E. coli KNabc and K+ uptake-deficient strain E. coli TK2420. The recombinant plasmid pGerN that contains an Na+/H+-K+ antiporter from Bacillus cereus (Southworth et al., 2001) was used as a positive control and the plasmid vector pGEM-3Zf(+) was used as a negative control; GerN confers robust Na+- and alkali-resistance upon E. coli KNabc. For studies of complementation of the Na+- and alkali-sensitive phenotypes of E. coli KNabc, the test and control transformants were cultured overnight in LBK, pH 7.5, after which the absorbance of each culture was adjusted to A600=0.5 before 10-fold serial dilutions were made with the same medium. One μl of each dilution was dotted on LBK plates with different pH values and concentrations of added NaCl. For assays of Ca2+-resistance, conducted in E. coli KNabc on plates, Tris-E medium (pH 8.0) containing 100 mM added CaCl2 was used (Brockman & Heppel, 1968; Waditee et al., 2004). A control transformant of E. coli DH5α that has a wild-type chaA gene was used as the positive control. Since E. coli ChaA has Ca2+/H+ antiport activity, it supports calcium-resistance (Ivey et al., 1993; Ohyama et al., 1994). Pre-cultures were grown and diluted the same way as described above. The plates for these sets of assays were incubated for the times indicated in the legend to Figure 3. Tests were also conducted of the complementation or exacerbation of the high K+ requirement of K+ uptake-deficient E. coli TK2420. The inocula for these growth experiments was grown in LBK (pH 7.5) and then diluted 200-fold for the growth assays in liquid minimal medium containing a range of added KCl concentrations (Epstein et al., 1993). In some experiments, the (NH4)2SO4 that is the sole nitrogen source in this minimal medium was replaced by 5 mM glutamine. After 16 hours, the A600 was recorded. All complementation assays were carried out in duplicate in at least two separate trials.
Everted membrane vesicles were prepared from E. coli KNabc transformants by the method described by Rosen (Rosen, 1986). Briefly, cell pellets from one liter of culture were washed twice with 25 ml of TCDG buffer containing: 10 mM Tris-chloride (pH 7.5), 140 mM choline-chloride, 0.5 mM dithiothreitol, 10% glycerol. Cell suspensions in 20 ml batches were used for disruption after addition of a protein inhibitor tablet (Roche), 1 mM phenylmethylsulfonyl fluoride and a trace amount of DNase I (Roche). The cell suspensions were passed through a French Press Cell under 10,000 p.s.i. pressure. The broken cell suspensions were subjected to two low speed centrifugations before the vesicles were collected by ultracentrifugation at 250,000 g (Beckman Ti60 rotor), 4°C for 1 hour. The vesicles were suspended in 1 ml of TCDG buffer. The protein concentration was measured by the Lowry method using lysozyme as the standard (Lowry et al., 1951). Monovalent cation/H+ antiport assays of the everted membrane vesicles were carried out in a 2 ml reaction mix containing 10 mM BTP buffer containing 140 mM choline-chloride, 5 mM MgCl2, 1 μM acridine orange, and 50 μg of vesicle protein ml-1 that were pre-mixed at room temperature. The pH values are indicated for particular experiments; the buffer pH was adjusted with sulfuric acid. Antiport activity was monitored by a fluorescence assay in which acridine orange (AO) was used as a probe of the ΔpH, acid inside the everted vesicles, that is first established by respiration upon addition of an electron donor (2.5 mM Tris-D-lactate) to the reaction mixture. Quenching of the AO fluorescence is observed as the probe is internalized in parallel with the development of this ΔpH, and then antiport is assessed as the % dequenching of fluorescence that occurs when cation is added to the reaction mix (Goldberg et al., 1987). Fluorescence was monitored using a Shimadzu RF-5301PC spectrofluorophotometer. Background values from the vector control plasmid vesicle preparations were measured and subtracted. At the end of the assay, 25 mM NH4Cl was added to establish the baseline as the ΔpH dissipated. At least two assays were conducted on two independent vesicle preparations in each experiment.
The new library of A. amylolytica N10 DNA in pGEM-3Zf(+) was first screened at pH7.5 for clones that complemented the Na+(Li+)-sensitive phenotype of E. coli KNabc, i.e. with the same screen that had yielded Aa-NhaD from an earlier library in pUC18 (Liu et al., 2005). A large proportion of the complementing clones were found to contain Aa-nhaD and no other membrane transporter candidates were identified in this screen. We therefore sought to devise a screen based on alkali-resistance. Preliminary growth experiments were conducted on the untransformed host E. coli KNabc strain to establish conditions for the screen. Both alkali-resistance and Na+-resistance of E. coli KNabc are expected to be enhanced by the presence of added K+ (Harel-Bronstein et al., 1995; Padan & Krulwich, 2000; Padan et al., 2005; Radchenko et al., 2006b). However a side-by-side test of these expectations had not been presented. Assays of the E. coli KNabc growth yield were conducted as a function of pH and added NaCl in Luria Broth (LB) in which added NaCl is replaced with KCl (LBK) vs LB with choline chloride (LBC) replacing added NaCl or KCl. As expected, growth inhibition by Na+ was more pronounced as the growth pH was raised (Fig. 1). Moreover, at every pH tested, the triple antiporter mutant was more sensitive to growth inhibition by Na+ in LBC than in the K+-replete LBK medium. In the absence of added Na+, addition of K+ led to only a modest increase in growth yield at pH 7.0-7.5, but K+ addition strongly increased the growth yield at high pH values. At pH 8.5-9.0, no growth was observed in the absence of added K+ even when Na+ was also omitted from the medium. At pH 9.0, only a very low level of growth of E. coli KNabc was observed under this condition. The screening conditions chosen to identify antiporters from the A. amylolytica N10 were LBK medium at pH 8.5 and 9.0. These conditions might select for antiporters that can use the K+ or Ca2+ as an efflux substrate to couple to the H+ uptake that lowers the alkali-sensitivity of E. coli KNabc. The screen might also select for Na+/H+ antiporters that are different from Aa-NhaD in being able to use the low Na+ content of LBK to achieve H+ uptake in support of alkali-resistance. Five plasmids with distinct inserts were isolated from transformants that grew under these screening conditions and conferred resistance in a re-screening experiment under the same conditions. One of these plasmids, pNAK7, was from the pH 8.5 screen while the other four plasmids, pNAK8-11, were from the pH 9.0 screen.
Complete sequence and homology analyses revealed that two of the plasmid inserts, those of pNAK9 and 10 were overlapping gene sequences. As shown in the diagram of the inserts in Fig. 2, there were candidate genes for cation/H+ antiporters in pNAK7, pNAK9/10 and pNAK11 but not in pNAK8. pNAK8 is predicted to encode a member of the SulP transporter family whose most closely characterized homologue is a bicarbonate transporter from cyanobacteria (Price et al., 2004). This raises the interesting possibility that bicarbonate transport could be used by prokaryotes for alkaline pH homeostasis as well as for 1-carbon homeostasis. Significant exploration will be required to test this possibility and it is not pursued here. Further study focused on the candidate cation/H+ antiporter genes in pNAK7, pNAK9/10 and pNAK11. Each of these three genes is listed with a proposed name in Table 2, which shows the two closest homologues on the basis of sequence similarity followed by the two closest homologues for which there are experimental data about structure and/or function. In all instances, the two closest homologues have not yet been characterized.
The candidate gene of interest in the insert from pNAK7 (4.1kb) was a putative trans-membrane protein of 319 amino acids that has strong homology to the K+-dependent Ca2+/Na+ exchangers from the cation/Ca2+ (CaCA) transporter super-family (Cai & Lytton, 2004; Saier et al., 1999). It was designated Aa-CaxA because the related CAX and YRBG branches of the CaCA family contain the closest bacterial homologues among a large group of transporters that are involved in calcium exchange reactions (Cai & Lytton, 2004). Aa-CaxA shows significant sequence similarity (42% identity) to the putative E. coli Na+/Ca2+ exchange protein YrbG, whose membrane topology was determined but whose transport activity has not yet been reported. Aa-CaxA also shows a comparable hydropathy profile and acidic motif but less sequence similarity to two cyanobacterial transporters, Syn-CAX (23% identity, see Table 2) and Ap-CAX (20% identity), for which Ca2+/H+ antiporter activity has been reported (Waditee et al., 2004). It shows slightly less similarity (18% identity), as do other members of the CaCA family (Cai & Lytton, 2004), to the ChaA antiporter of E. coli that has been shown to catalyze Na+(Ca2+)(K+)/H+ antiport (Ivey et al., 1993; Radchenko et al., 2006a).
The merged sequences from the inserts of pNAK9 and 10 contained a cation/H+ antiporter candidate from a different transporter family. The predicted gene product has 661 amino acids and homology to the members of cation: proton antiporter family-2 (CPA2) (Saier et al., 1999). This family includes well characterized transport proteins, such as the glutathione-gated K+ efflux Kef B and KefC systems from E. coli that have some channel-like properties (Booth et al., 2005). The CPA2 family also contains transporters for which antiport activity has been experimentally supported, e.g. Na+/H+ antiporter NapA from Enterococcus hirae (Waser et al., 1992), the Na+/H+-K+ antiporter GerN from Bacillus cereus (Southworth, 2001), and the Fe2+/H+ antiporter MagA from Magnetospirillum sp. strain AMB-1 (Nakamura et al., 1995). Among the CPA-2 proteins that have been functionally characterized, the protein from A. amylolytica N10 most closely resembles E. coli Kef B (29%) (Table 2) and hence was designated Aa-KefB.
The insert in pNAK11 was the largest of the inserts (Fig. 2) and contained a gene encoding a protein that was designated Aa-NhaP. This predicted product has 574 amino acids and homology to members of the monovalent cation:proton antiporter-1 (CPA1) family (Saier et al., 1999). The closest homologues among characterized CPA1 antiporters are YcgO (41%) from E. coli and NhaP2 (40%) from Vibrio parahaemolyticus. YcgO plays a role in cell volume regulation at low osmolarity and was suggested to translocate Na+ and K+ (Verkhovskaya et al., 2001). Vp-NhaP2 is exclusively a K+/H+ antiporter whose proposed function is protection against adversely high concentrations of K+ at alkaline pH (Radchenko, 2006b). Aa-NhaP has much less similarity (24% identity) to MjNhaP, an Na+ (Li+)/H+ antiporter that is only active at pH 7.0 or below (Hellmer et al., 2002).
The individual Aa-caxA, Aa-kefB, and Aa-nhaP genes, cloned in pGEM-3Zf (+) with their own ribosomal binding sites behind the T7 promoter (in plasmids pYW136, pYW138 and pYW139, respectively), were tested for their ability to complement several phenotypes of E. coli mutant strains. The empty vector was the negative control and the B. cereus Na+/H+-K+ antiporter GerN (Southworth et al., 2001) was a positive control for all but one of the plate assays. E. coli DH5α, which has a wild-type chaA gene, was the control for the Ca2+-sensitivity assay.
The complementation profile for pYW136, expressing Aa-caxA, in assays conducted in E. coli KNabc, was: modest complementation of Na+-sensitivity at pH 7.5 and no complementation at pH 8.0 (Fig. 3A,B); significant complementation of alkali-sensitivity in the absence of added Na+ (Fig. 3C); and modest complementation of Ca2+-sensitivity (Fig. 3D). Given the undetectable levels of Ca2+ and very low levels of Mg2+ in the natural environment of A. amylolytica N10, it is unlikely that a robust efflux protein for these important micronutrients has an important physiological role. Possibly the low complementation activity reflects this. Alternatively, it is possible that under physiological conditions in the natural host, Aa-CaxA is used to accumulate Ca2+ in exchange for cytoplasmic Na+; the modest beneficial effect on alkali-sensitivity could be an indirect effect of lowering cytoplasmic [Na+]. A test of the Aa-caxA effect in the K+ uptake-deficient strain E. coli TK2420 showed that Aa-caxA modestly enhanced growth of the mutant in the presence of limiting [K+] (Fig. 4A). This is consistent with the possibility that the exchange catalyzed by Aa-CaxA can use K+ as part of the coupling ion complement, as has been demonstrated for some other antiporters such as B. cereus GerN (Southworth et al., 2001), B. subtilis CzcD (Guffanti et al., 2002) and Aphanotece halophytica NapA1-1 (Wutipraditkul et al., 2005). In this scenario Aa-CaxA might be a Na+/Ca2+-K+ exchanger, but it is also possible that the beneficial effect of Aa-CaxA at limiting K+ is an indirect effect of reduced cytoplasmic Na+ rather than actual K+ uptake. Attempts to demonstrate Ca2+ or Mg2+/H+ antiport by assays described for cyanobacterial Syn-Cax (Waditee et al., 2004) were negative, with or without added K+ at a range of alkaline pH values (data not shown).
The pYW138 plasmid in which Aa-KefB is encoded conferred significant Na+-resistance in E. coli KNabc at both pH 7.5 and 8.0 and also conferred significant alkali-resistance in LBK medium without added Na+ (Fig. 3A-C). Growth studies of the Aa-kefB transformant of E. coli TK2420 showed that this gene exacerbated the requirement for high [K+] in the medium since growth was lower than that of the empty vector control at < 40 mM added K+ (Fig. 4A). This is consistent with a K+ efflux capacity as found for other Kef proteins (Bakker et al., 1987). Perhaps Aa-KefB is an K+(Na+)/H+ antiporter, accounting for all the complementation data, but we have not yet found assay conditions that demonstrate such activity. We note that there is no obvious accessory protein encoded in an operon with Aa-KefB of the kind that have been identified for the subset of CPA-2 proteins that include E. coli KefB and KefC (Miller et al., 2000). Neither of the small ORFs upstream of Aa-kefB (Fig. 2) resembles the ancillary proteins of the E. coli KefB and KefC homologues. However, many members of the CPA-2 antiporter family for which antiport activity has been demonstrated (e.g. GerN and NapA) lack apparent auxiliary proteins (Southworth et al., 2001; Waser et al., 1992).
The plasmid pYW139 whose insert encodes Aa-NhaP, complemented both the Na+- and alkali-sensitive phenotypes of E. coli KNabc better than any of the other A. amylolytica N10 genes albeit not as strongly as the positive control (Fig. 3A-C). Most strikingly, Aa-NhaP strongly impaired growth of K+ uptake-deficient E. coli TK2420, inhibiting growth almost completely even in the presence of 40 mM added K+ (Fig. 4A). The failure to reverse the growth inhibition by Aa-NhaP by high concentrations of K+ indicated that the inhibition was not solely a result of robust Aa-NhaP-dependent K+ extrusion via a presumed antiport mechanism. We noted that the Aa-NhaP transformant of E. coli TK2420 grew identically to the control transformant in the undefined LBK maintenance medium. LBK contains an organic nitrogen source in the form of a mixture of amino acids and peptides, whereas the minimal medium used for complementation assays contains (NH4)2SO4 as the sole nitrogen source. Since numerous K+ transporters can also use NH4+ as a substrate (Buurman et al., 1991; Epstein, 2003; Wei et al., 2003), a plausible hypothesis for the growth inhibition is that NH4+ is also an efflux substrate for Aa-NhaP, perhaps even a preferred one relative to K+. Even in the presence of added K+ such an antiporter might inhibit growth of E. coli TK2420 in the complementation medium by using its NH4+/H+ antiport activity to exclude the NH4+ in the medium, thereby depriving the bacterium of a nitrogen source. If this is the case, growth of the E. coli TK2420 transformant with Aa-NhaP on high [K+] should be restored if the (NH4)2SO4 in the medium is replaced with glutamine. Indeed, when 5 mM glutamine was the nitrogen source, the adverse effect of Aa-NhaP on growth of E. coli TK2420 was still observed at low added [K+] but now higher concentrations of K+ reduced the growth inhibition at 16 hours; at 24 hours, growth of the Aa-NhaP was comparable to the control in the presence of 40 mM added K+ (Fig. 4B). The overall complementation pattern suggested that Aa-NhaP is a highly active K+(NH4+)/H+ antiporter that may be able to use Na+ as a less optimal efflux substrate.
Fluorescence-based assays of Aa-NhaP-dependent antiport activity used AO quenching to assess development of a pH gradient, acid in, across the everted vesicles when respiration was initiated (first downward arrow in the traces in Fig. 5). The % dequencing of the fluorescence in response to added cation was used to assess antiport activity. An antiport mechanism was testable for K+/H+ and Na+/H+ antiport in these fluorescence-based assays of everted membrane vesicles expressing either a control empty vector or pYW139; NH4+ cannot be tested as a substrate in these assays because its addition abolished the pH gradient that is used to monitor activity. Initial experiments were conducted at pH 7.5 to assess the cation specificity of antiport. As shown in Fig. 5, vesicles expressing Aa-NhaP exhibited clear fluorescence dequenching in response to K+ addition whereas the control exhibited none. Although addition of Na+ to the Aa-NhaP vesicles did not elicit a sharp upward dequench as is usually observed, there was a slow dequench that consistently exceeded that observed in the control vesicles; this may reflect a modest capacity to use Na+ as a substrate. No activity with Li+ was observed in comparison to the control. The pH profile for K+/H+ activity in the heterologous system exhibited an optimum at pH 7.5 (Fig. 6A), which would probably be non-physiological for the natural extremophile host. The cytoplasmic pH of A. amylolytica N10 has not yet been determined when it is growing at its optimal growth pH of 10. However, other extreme alkaliphiles exhibit cytoplasmic pH values near 8 in that range of outside pH (Krulwich, 1995; Krulwich et al., 2007; Yumoto, 2002). The broad pH optimum for Aa-NhaP and retention of significant activity at pH 9.5 (the limit for the assay) are consistent with a possible role in pH homeostasis in the natural host (Fig. 6A). The activity of Aa-NhaP was studied as a function of [K+] and exhibited Michaelis-Menten kinetics with an apparent Km of about 0.5 mM. This apparent Km is in the same range of low values found for well-studied Na+/H+ antiporters from E. coli that have important roles at elevated pH (Padan et al., 2001). Until a genetic system is available for A. amylolytica N10, proposed roles remain untested. Aa-NhaP could play a role in osmo-adaptation and could also function in alkali-resistance, playing an adjunct role to the Na+/H+ antiporters that are expected to play a major role in both alkali- and Na+-resistance. For example, once Na+/H+ antiporter activity has generated a substantial, inwardly-directed Na+ gradient and a lower cytoplasmic pH than external pH, Aa-NhaP could reduce the cytoplasmic pH further using the outwardly-directed K+ gradient to power or partially power H+ uptake. Finally, we hypothesize that Aa-NhaP also catalyzes NH4+/H+ antiport, a possibility that merits further examination. When cells are actively catabolizing amino acids at high pH, accumulation of cytoplasmic NH4+ poses a problem as it would be expected to inhibit central metabolic pathways as well as pH homeostasis itself. Exchange of cytoplasmic NH4+ for external H+ would mitigate these effects. It will be of interest and importance to use biochemistry and genetics to determine the extent to which transporters currently characterized as using K+ as a substrate also use NH4+ and to evaluate the physiological impact of the use of NH4+.
This work was supported in part by research grant GM28454 from the National Institute of General Medical Sciences (to TAK) and a grant from the Chinese Academy of Sciences (Knowledge Innovation Program KSCX2-SW-33) and from the Ministry of Sciences and Technology of China (863 program 2004AA214060; 973 program, 2003CB716001) (to YM).
The GenBank/EMBL/DDBJ accession number for the sequences of the A. amylolytica N10 gene products are: CaxA (ABG37980), KefB (ABG37986) and NhaP (ABG37987); DNA sequences of the larger cloned fragments containing these genes were deposited, respectively, as pNAK7 (DG649017), pNAK9/10 (DG649019) and pNAK11 (DG649020) and a fourth cloned fragment that was not further studied, pNAK8, was deposited as DQ649018.