|Home | About | Journals | Submit | Contact Us | Français|
A putative transport protein (Orf9) of alkaliphilic Bacillus pseudofirmus OF4 belongs to a transporter family (CPA-2) of diverse K+ efflux proteins and cation antiporters. Orf9 greatly increased the concentration of K+ required for growth of a K+ uptake mutant of Escherichia coli. The cytoplasmic K+ content of the cells was reduced, consistent with an efflux mechanism. Orf9-dependent translocation of K+ in E. coli is apparently bidirectional, since ammonium-sensitive uptake of K+ could be shown in K+-depleted cells. The upstream gene product Orf8 has sequence similarity to a subdomain of KTN proteins that are associated with potassium-translocating channels and transporters; Orf8 modulated the transport capacities of Orf9. No Orf9-dependent K+(Na+)/H+ antiport activity was found in membrane vesicles. Nonpolar deletion mutants in the orf9 locus of the alkaliphile chromosome exhibited no K+-related phenotype but showed profound phenotypes in medium containing high levels of amine-nitrogen. Their patterns of growth and ammonium content suggested a physiological role for the orf9 locus in bidirectional ammonium transport. Orf9-dependent ammonium uptake was observed in right-side-out membrane vesicles of the alkaliphile wild type and the mutant with an orf8 deletion. Uptake was proton motive force dependent and was inhibited by K+. Orf9 is proposed to be designated AmhT (ammonium homeostasis). Ammonium homeostasis is important in high-amine-nitrogen settings and is particularly crucial at high pH since cytosolic ammonium accumulation interferes with cytoplasmic pH regulation. Endospore formation in amino-acid-rich medium was significantly defective and germination was modestly defective in the orf9 and orf7-orf10 deletion mutants.
Monovalent cation transport systems of alkaliphilic Bacillus species have a crucial role in the central biological problem of these “extremophiles,” i.e., cytoplasmic pH homeostasis. Prokaryotic alkaliphiles such as Bacillus pseudofirmus OF4 grow optimally up to pH values of 10 to 11. The cells maintain a cytoplasmic pH that is over 2 pH units below the external pH during steady-state growth at such pH values. Moreover, they exhibit that large pH gradient, acid in, immediately after a sudden shift of pH 8.5-equilibrated cells to pH 10.5 (31-33, 51). A large body of evidence indicates that pH homeostasis in B. pseudofirmus OF4 and related alkaliphiles depends upon Na+. One reason for this dependence is that Na+/H+ antiporters are an indispensable part of the mechanism for accumulation of H+ in the cytoplasm relative to the external milieu (29, 31-33). The Mrp antiporter system of alkaliphilic Bacillus species is predicted to have a dominant role in alkaliphile pH homeostasis (19, 24, 29, 33), but supporting roles for two or more additional Na+/H+ antiporters, one of which is nhaC, have been indicated in B. pseudofirmus OF4 (25). The full complement of genes participating in this process has not yet been elucidated in any single alkaliphile. Nor has the full range of other physiological roles of the known alkaliphile antiporters been completely clarified. In nonalkaliphilic prokaryotes Na+/H+ antiporters have been suggested to have roles in Na+ resistance and cell volume regulation (40, 41, 54), initiation of sporulation (30), and spore germination (50, 52, 53).
When a new gene locus of B. pseudofirmus OF4 was found to contain a cistron encoding a putative member of a cation:proton family of transporters, CPA-2 (45), it was thus of interest to determine whether it encoded a Na+/H+ antiporter, as does napA of that family (54). Such an alkaliphile homologue might be involved in one or more of the abovementioned physiological processes in this extremophile or in functions that have been reported for other CPA-2 transporters. Apart from NapA, members of the CPA-2 transporter family encompass proteins of diverse catalytic capacities and roles including the iron transporter MagA (37); a Na+/H+-K+ antiporter, GerN, with a role in spore germination (50, 53) as well as a less characterized homologue, GrmA, also involved in spore germination (52); and several K+ efflux systems, KefC and KefB of Escherichia coli (6, 8, 13). The K+ efflux members of the CPA-2 family have been extensively studied and may be channels rather than secondary transporters (6, 8). The data presented here suggest that the product of the new alkaliphile gene can catalyze transmembrane fluxes of potassium ions and ammonium, with the latter being the physiologically important substrate of the transporter in the alkaliphile.
E. coli DH5α (Gibco-BRL) was used for routine cloning procedures. E. coli strains EP432 (ΔnhaA ΔnhaB) (43) and KNabc (ΔchaA ΔnhaA ΔnhaB) (39) were used for screening for Na+ resistance. Potassium-uptake-deficient E. coli TK2420 (kdp kup trk) (14) was grown on a defined medium (15) to determine the effects of various concentrations of KCl. E. coli strains were routinely grown at 37°C on LBK medium at pH 7.5 (17). B. pseudofirmus OF4 811 M (wild type used in this study) was grown on semidefined medium at either pH 7.5 or pH 10.5. At pH 7.5 the medium buffer consisted of 0.1 M MOPS (morpholinepropanesulfonic acid)-NaOH, 1 mM potassium phosphate, 135 mM NaCl, and 10 mM MgSO4. At pH 10.5 or 9.5 the medium buffer consisted of 100 mM Na2CO3-NaHCO3, 1 mM potassium phosphate, and 0.1 mM MgSO4. The media were supplemented with 0.1% (wt/vol) yeast extract, 1% STS trace salts (20), and 50 mM sodium malate. For sporulation, the same pH 10.5 medium was used, except that the 50 mM sodium malate was replaced by 33.3 mM sodium malate and 16.6 mM sodium glutamate. The alkaliphile was also grown on QA medium, a minimal, defined medium in which the yeast extract was replaced with 1 μg of thiamine/ml, 1 μg of biotin/ml, 10 μg of l-methionine/ml, 0.1% l-alanine, and 0.1% l-glutamine. The plasmids used in this study were low-copy-number pMW118 (Nippon Gene, Totama, Japan), pGEM3zf(+) (Promega), and shuttle vector pBK36 (obtained from K. Zen) (11).
A partial orf9 gene was first identified in a clone from a B. pseudofirmus OF4 DNA library that had been prepared from a partial mbo digest of chromosomal DNA. This clone was isolated as part of an earlier study (23) but was not reported or further studied until the orf9 sequence and additional downstream DNA sequence were available. This was achieved using a λ phage BamHI library containing B. pseudofirmus OF4 genomic DNA as used elsewhere (55). Briefly, a 32P-labeled DNA probe was employed for the screening of the library for the orf9 gene locus. The probe was a 350-bp PCR product corresponding to the sequence covering the C-terminal end of Orf8 and the N-terminal part of Orf9 (nucleotides [nt] 6672 to 7022 from U89914.2.0) with primers YM25/6672F and YM25/7022R; these and all primers used for cloning and reverse transcription-PCR (RT-PCR) are listed in Table Table1.1. The positive candidates, designated λ-M25, were identified by standard Southern blotting procedures and confirmed by PCR. The insert of λ-M25 was sequenced by primer walking on both strands, and additional PCR products were prepared and sequenced as necessary to clarify or confirm the sequence thus obtained. The sequencing employed an Applied Biosystems (Foster City, Calif.) 373 DNA sequencer at the Biotechnology Center at Utah State University, Logan, and the DNA Sequencing Core Facility at the Mount Sinai School of Medicine. The DNA sequences were aligned and analyzed for location of possible open reading frames (ORFs) by using the Gene Runner 3.05 program (Hastings Software, Hastings, N.Y., 1994) and the ORF Finder program from the National Center for Biotechnology Information (NCBI) web service (http://www.ncbi.nlm.nih.gov/Sitemap/index.html#ORFFinder). The putative ORFs were analyzed using the BLAST (2) network service from the NCBI (http://www.ncbi.nlm.nih.gov/) and the National Institute of Genetics Center for Information Biology and DNA Data Bank of Japan (http://www.ddbj.nig.ac.jp). The secondary structure and the free energy calculations of RNA stem-loops were analyzed using a network service (http://www.bioinfo.rpi.edu/applications/mfold/old/rna/form3.cgi).
B. pseudofirmus OF4 and selected deletion mutants thereof (see below) were pregrown in pH 9.5 medium at 30°C overnight and then diluted 1:10 into pH 10.5 medium at 30°C. When the culture reached approximately 140 Klett units, growth was stopped by adding a half volume of precooled 50 mM NaN3 into the cultures. RNA isolation was performed as described previously (56). RT was performed on 5 μg of total RNA with a SuperScript First-Strand Synthesis System for an RT-PCR kit from Invitrogen (catalog no. 11904-018). PCR was performed using the HotStarTaq MasterMix kit (catalog no. 203443) from Qiagen. The primers used for RT-PCR are shown in Fig. Fig.11 and Table Table11.
PCR was performed according to the instructions accompanying Platinum Pfx DNA polymerase kits from Invitrogen (catalog no. 11708-021). The primers used for cloning were as follows: for orf8, 8SalF and orf8BR1; for orf9, 9SalF and orf9BR2; for orf8-orf9, 8SalF and orf9BR2. The PCR products were digested with SalI and BamHI restriction enzymes and inserted into the cloning vector pMW118 SalI-BamHI sites. After amplification of this plasmid in E. coli, the fragment was transferred into shuttle expression vector pBK36 under the control of the ermC promoter and a Shine-Dalgarno sequence on the vector. The alkaliphile DNA fragments in the recombinant plasmids contained only the complete coding regions for the specific ORF(s) indicated. All the constructs were confirmed by DNA sequencing.
A series of in-frame deletion mutants of orf7, orf8, orf9, orf10, orf8-orf9, and orf7-orf10 was constructed. For the in-frame deletion of orf7, two PCRs were conducted using the Platinum Pfx DNA polymerase kit from Invitrogen. The primer pair for the first PCR was 7BF and 7SR. 7BF was the forward primer containing a BamHI site. 7SR was the reverse primer containing a SacI site (the positions of the restriction sites are shown in Table Table1).1). The PCR product generated by 7BF-7SR corresponded to the orf9 locus sequence from nt 5074 to 5778, which is a 704-bp product extending from sequence upstream of orf7 through the sequence encoding the first six amino acids (MEKDEL) of Orf7. The primer pair for the second PCR was 7SF and 7KR. 7SF was the forward primer containing a SacI site. 7KR was the reverse primer containing a KpnI site. The PCR product generated by 7SF-7KR corresponded to the orf9 locus sequence from nt 6198 to 6832, a 634-bp product encompassing the coding sequence for the final 3 amino acids (SES) of Orf7 through sequence downstream of orf7. After restriction enzyme digestions, the two PCR fragments were linked together by the SacI site. The resulting construct contained an in-frame deletion of orf7 that is missing 140 codons and retains the first six and last three codons. The two linked PCR fragments were cloned into pGEM3zf(+) vector between the BamHI and KpnI sites. After the inserted fragment was confirmed by sequence analysis, this fragment was released from the pGEM3zf(+) vector by BamHI and KpnI digestion and recloned into a temperature-sensitive shuttle vector plasmid, pG+host-4 (Ermr; Appligen, Pleasanton, Calif.), between its BamHI and KpnI sites.
The pG+host-4-orf7 deletion plasmid was used to transform B. pseudofirmus OF4 protoplasts as described previously (25). In brief, the DM3 medium plates with 0.15 μg of erythromycin/ml were used for selection of the regenerated transformants at 30°C for 4 days. Single-crossover candidates were obtained by shifting the growth temperature from 30 to 42°C on complex medium plates in the presence of 0.6 μg of erythromycin/ml. They were confirmed by PCR with one primer (7FBF or 10RKR) in the chromosomal region outside that used in the PCR for generation of the deletion vector and the other primer on the erm gene within the plasmid vector (EM2753 or EM2471). Only the positive candidates gave products of the predicted size. Double-crossover candidates were then selected as colonies that lost the Ermr phenotype at 42°C. Further PCRs with the primer pair 7FBR-10RKR, corresponding to chromosomal regions outside those used for the deletion construct, were used to confirm that the deletion mutant product was the right size.
Similar strategies were used for the construction of in-frame deletions of orf8, orf9, orf10, orf8-orf9, and orf7-orf10. The primer pairs for upstream fragments of orf7, orf8, orf9, and orf10 were 7BF-7SR, 8BF-8SR, 9BF-9SR, and 10BF-10SR, respectively. The primer pairs for downstream fragments of orf7, orf8, orf9, and orf10 were 7SF-7KR, 8SF-8KR, 9SF-9KR, and 10SF-10KR, respectively. The orf8-orf9 in-frame deletion was constructed by ligation of the upstream PCR fragment of orf8 with the downstream PCR fragment of orf9, and the orf7-orf10 in-frame deletion plasmid was constructed using the upstream PCR fragment of orf7 with the downstream PCR fragment of orf10. All the constructs were confirmed by sequence analysis. All the in-frame deletion mutants were confirmed by PCR and designated Δ7, Δ8, Δ9, Δ10, Δ8-9, and Δ7-10.
E. coli TK2420 cells were grown in LBK medium to the late exponential phase and were depleted of K+ by the Tris-EDTA treatment described by Nakamura et al. (38). After K+ depletion, the cells were washed with and resuspended in the defined medium with 0.5% glycerol as carbon source instead of the standard glucose (15). The final suspension contained 0.5 mg of cell protein/ml in a volume of 15 ml. KCl was added to 25 mM, and samples (1 ml) were taken at various times for measurement of K+ uptake as described by others (13). Briefly, the samples were centrifuged for 1 min in a microcentrifuge, the supernatant was removed by aspiration, and the pellet was resuspended in 1 ml of water. After being boiled for 5 min, cell debris was removed by centrifugation and potassium was measured by flame photometry with a Bacharach Coleman Model 51 Ca flame photometer according to the instructions of the manufacturer. Flame photometry was also used to assess the K+ content of the E. coli TK2420 transformant expressing orf9 versus that with a plasmid control. The cells were grown overnight in LBK medium, harvested by centrifugation, and diluted into the defined medium containing 25 mM added K+ for 30 min before K+ was measured; an additional control involved treatment of the two transformants, under precisely the same conditions, with valinomycin (to a final 10 μM concentration).
Everted membrane vesicles of E. coli KNabc and EP432 transformants, with control vector or expressing orf8 and/or orf9, were prepared as described by Ambudkar et al. (3). These vesicles were used for fluorescence assays of either Na+/H+ or K+/H+ antiport activity monitored with acridine orange that were performed by the method of Goldberg et al. (17). Right-side-out (RSO) membrane vesicles of comparable E. coli TK2420 transformants were prepared by the method of Kaback (27). The exchange of radioactive rubidium for Na+ was assayed as previously described (50) using these RSO vesicles. RSO vesicles were also prepared from wild-type and mutant strains of B. pseudofirmus OF4 by the same method cited above, except that EDTA was omitted from the protocol; the vesicles were prepared in 50 mM sodium phosphate buffer, pH 7.5. Ammonium uptake by the vesicles was assayed after addition of ammonium sulfate to a final concentration of 1 mM. Where indicated, 1 mM potassium phosphate and/or 10 mM sodium ascorbate and 0.1 mM phenazine methosulfate (PMS) were added. At various time points, the vesicles were rapidly separated from the buffer by centrifugation in a microcentrifuge. The pellets were resuspended in water and lysed by bringing the pH up to 12.5 so that the ammonia could be measured. A Corning gas-sensing ammonia combination electrode (part no. 300740.0) was used for these measurements, according to the manufacturer's instructions.
Assays of ammonium content were conducted on whole cells of the wild type and orf8, orf9, and orf8-orf9 mutants. They were grown at pH 7.5 on semidefined l-malate-containing media to the mid-logarithmic phase of growth. The cells were then harvested by centrifugation at 10,000 × g and washed by suspension in the buffer used as the base of the QA medium, at pH 7.5, followed by another centrifugation. Cells were then resuspended at 0.2 mg of cell protein/ml in 20 ml of QA medium, pH 7.5, in which the glutamine and alanine concentrations were one-half those of the normal QA growth medium. The whole suspension was harvested at zero time and 3 and 6 h by centrifugation, suspended in 2 ml of pH 7.5 buffer (as used for QA medium), and assayed for ammonium after lysis of the cells by treatment at pH 12.5. The assays of ammonium content were all conducted at pH 7.5 because use of pH 10.5 conditions raised the background ammonia levels too much.
Inasmuch as the conditions for sporulation and germination of B. pseudofirmus OF4 had never been examined, preliminary experiments were conducted to establish the following conditions for wild-type sporulation and germination. For preparation of spore suspensions, cells of B. pseudofirmus OF4 were grown overnight in the l-glutamate-containing sporulation medium described above and then 0.2 ml of this preculture was used to inoculate solid medium of the same composition. The sporulation agar plates were incubated at 30°C. After 6 days, spores were harvested from the surface and washed eight times in ice-cold deionized water by repeated suspension and centrifugation. For the wild type, washed-spore preparations were about 95% phase-bright spores. Spores were stored at −20°C in deionized water. For electron microscopic comparison of the spores of the wild type and mutants, sporulating cells were collected from the surface of plates in distilled water, centrifuged, resuspended in formaldehyde-glutaraldehyde fixative (28), and processed as described by Leatherbarrow et al. (34).
For assays of spore germination, spore suspensions were heat activated in water at 70°C for 30 min. The germination was then carried out in 125 mM 2-amino-2-methyl-1-propanol (AMP) buffer, pH 10.0, with appropriate germinants (200 mM NaCl and either 10 mM inosine, or 10 mM l-alanine with 5 μg of O-carbamyl-d-serine [an alanine racemase inhibitor]/ml). Germination was measured by the change in optical density (OD) of samples at 490 nm (initial OD was 0.5 to 0.8) in a microplate reader (Wallac Victor2 1420 multilabel counter). A 50% loss in OD corresponded to 100% germination of the B. pseudofirmus OF4 wild-type spores.
The whole orf9 locus sequence was deposited in GenBank with accession number U89914.2.
A diagrammatic representation of a gene locus containing “orf9,” encoding a CPA-2-type protein, is shown in Fig. Fig.1A.1A. The putative promoter shown in the diagram is likely to be a sigma A-type promoter rather than one associated more specifically with stress or sporulation, based on comparisons with Bacillus subtilis (21). A stem-loop structure (ΔG = −18.1 kcal/mol) is predicted between orf10 and orf11, and a second such structure (ΔG = −19.7 kcal/mol) is predicted between orf16 and orf17. Some of the proteins in the databases with the closest sequence similarity to the predicted products of genes in the region containing orf9, i.e., between the promoter and the first stem-loop, are shown in Table Table2.2. The first gene of interest, orf7, is a putative regulatory gene based on its sequence similarity to transcriptional regulators including B. subtilis YusO and Salmonella enterica serovar Typhimurium MarR. The MarR family of transcriptional regulators (Pfam PF01047) are DNA binding proteins that typically are negative regulators that function as dimers. Each monomer contains a winged-helix binding motif (1). The orf8 product is predicted to be a hydrophilic protein with a potential membrane-anchoring segment and modest overall sequence similarity to a number of K+ channel or transporter components. The domain that Orf8 has in common with these proteins corresponds to TrkA to TrkC (Pfam 02080). A C-terminal region of Orf8 exhibits 100% alignment of 85 residues that have 26% identity and 55% similarity to the TrkA-C domain consensus sequence. This domain is observed as part of a larger repeat domain structure in KTN (K+ transport nucleotide binding) proteins (7). TrkA-C has been proposed to bind a ligand, but this ligand and hence the specific function of TrkA-C are unknown (4). KTN domains are found in the C-terminal regions of the KefB and KefC K+ efflux proteins that are members of the CPA-2 protein family and are also present in the hydrophilic TrkA and KtrA subunits that assemble, respectively, with integral membrane subunits of the TrkG/H and KtrB K+ transporters (12, 45). The orf9 product is the CPA-2 transporter family member but one that does not itself contain a KTN domain. Its homology to KefB is limited to the N-terminal region, and it shows slightly more similarity to a putative NapA-like protein from Aquifex aeolicus (Table (Table2).2). The orf10 product shows sequence similarity to RelA and SpoT proteins but corresponds to a fragment that would not be expected to have any of the catalytic activities of those ppGpp-related enzymes. Summary information is also shown in Table Table22 for orf11 to orf16, because RT-PCR analyses supported the expectation that orf7 is the first gene of an orf9-containing operon but suggested that the operon extends beyond orf10. Quite a few of these genes are predicted to have roles in metabolism or transport of nitrogenous compounds. The contiguous arrangement of these genes, both from orf7 to orf10 and beyond to orf16, is not observed in the genome of either alkaliphilic Bacillus halodurans C-125 or B. subtilis.
The RT-PCR data are summarized in Fig. Fig.1B.1B. The primers used in the analyses are shown under a diagrammatic representation of the DNA fragment encompassing orf6 to orf17 (Fig. (Fig.1A).1A). First, the R1 primer, corresponding to part of orf9, was used for RT and as the reverse primer for the PCR. Use of F2 as the forward primer resulted in an 0.98-kb product from RNA from both the wild type and orf9 deletion mutant (lanes 1 and 3) as well as in control reactions with chromosomal DNA (lane 5). The primer pair of F1-R1 produced no PCR product from RT samples from either the wild type (lane 2) or the orf9 deletion mutant (lane 4), while the genomic DNA control template gave a product of approximately 1.4 kb (lane 6). The F1 primer is located in a sequence between the orf6 and orf7 genes, 12 nt after orf6 and before the predicted promoter, and the F2 primer is within orf7. Thus, orf6 and orf7 are presumed not to be part of the same transcriptional unit, whereas orf7 to orf9 are. Although we initially expected orf10 to complete the operon, further RT-PCR linked both orf10 and orf11 to orf8 and orf9. When genomic DNA was used as PCR template, the F3 primer, which corresponds to part of orf8, and R2, R3, and R4 primers that respectively correspond to parts of orf9, orf10, and orf11 gave the expected bands at 1.9, 2.2, and 2.5 kb (lanes 13, 14, and 15). However, the wild-type RNA templates prepared with the same R2, R3, and R4 primers gave negative results in the PCRs primed with F3 (in orf8) and R2 (in orf10) (lanes 7, 8, and 9). While orf10 might not be transcriptionally linked to the three upstream genes, it seemed possible that the expected products were too big for effective RT. A product of 0.7 kb was indeed generated from F3-R2 PCRs with RNA prepared using primers R2, R3, and R4 and a template from the orf9 deletion mutant (lanes 10, 11, and 12), i.e., from which the expected RNA products from RT would be significantly smaller than those from wild type. The product formed from F3-R2-primed PCR when R2 and R3 had been used to prime the RT (lanes 10 and 11) indicated that orf10 was part of the same transcriptional unit as orf8. Since a product of identical size, albeit a weaker band, was formed from F3-R2-primed PCR using R4 (corresponding to part of orf11) (lane 12), some transcripts contain both orf8 and orf11. Perhaps only some of the transcripts extend that far, or perhaps the longer transcript is less stable. As shown in the right panel of Fig. Fig.1B,1B, F4-R3 (corresponding to parts of orf10 that should produce a small product) were used to prime PCRs with R3- or R4-primed RT products as template. The identical product was formed from these and control reactions (lanes 16 to 20), supporting the conclusion that orf11 can be found on the same RNA as orf10. Attempts to link any of the more downstream genes to orf10 or orf11 were made using a series of primers in orf12 and orf13. These attempts are not depicted in the figure because they were unsuccessful, suggesting that the longer RNA transcripts might not be sufficiently stable. It is likely, but remains undocumented by experimental results, that the operon extends beyond orf11 and through the end of orf16, given the overlap between orf11 and orf12 and the stem-loop that follows orf16. The next orf gene, orf17, is transcribed in the opposite direction. Perhaps the stem-loop after orf10 has some relevance to expression of the genes upstream of that first stem-loop. The study focused on orf7-orf10, which most closely surrounded the orf9 gene of particular interest.
Although not shown, expression of orf8, orf9, or orf8-orf9 together failed to complement the Na+-sensitive phenotype of Na+/H+ antiporter-deficient E. coli KNabc (ΔchaA ΔnhaA ΔnhaB) or E. coli EP432 (ΔnhaA ΔnhaB) under conditions in which expression of E. coli nhaA or Bacillus cereus gerN did complement. Consistent with that, we were unable to detect Na+/H+ antiport activity in vesicle assays or Na+/(H+)K+ antiport when K+ was present on the trans side of the membrane; such activities were conferred by gerN. Expression of orf9, but not orf8, did have a pronounced effect on the K+-uptake-deficient strain E. coli TK2420 (kdp kup trk). As shown in Fig. Fig.2,2, the transformant expressing orf9 required a much higher [K+] than did the control transformant in order to grow optimally. When orf8 and orf9 were expressed together, the transformant showed none of the detrimental effect of orf9 on growth. Measurements of cellular K+ concentrations by flame photometry confirmed that the LBK-grown transformant expressing orf9 had reduced levels of K+ compared to the control transformant when sampled 30 min after dilution into defined growth medium containing 25 mM added K+. In the presence of valinomycin, which is expected to move K+ rapidly down its electrochemical gradient to achieve equilibrium, the cytoplasmic K+ concentration of both transformants was 37 mM. In the absence of valinomycin, the cytoplasmic K+ concentration of the vector control transformant was 100 mM whereas that of the orf9-expressing transformant was 50 mM. The results suggested that Orf9 supports K+ efflux that almost completely outpaces the residual K+ uptake activity of E. coli TK2420. No K+/H+ antiport activity was detected in fluorescence-based assays of membrane vesicles.
We were unable to detect Orf9-dependent K+ uptake in membrane vesicles by flame photometry or with 86Rb+. However, 86Rb+ is not used by all K+-translocating systems and flame photometry-based transport assays generally have been reported for whole cells. We therefore turned to experiments using K+-depleted cells to further explore the possibility that Orf9 might function in K+ uptake. Transformants of the K+-uptake-deficient E. coli TK2420 were depleted of K+ by Tris-EDTA treatment (38) and assayed using the approach of Elmore et al. (13). The experiments were conducted in the defined medium for this strain with glycerol instead of the standard glucose that was used in the growth experiment shown in Fig. Fig.2.2. This medium contains abundant Na+, but inasmuch as there was no indication of a Na+-related phenotype or interference by Na+ with the K+-related phenotype, Na+ is unlikely to be an Orf9 substrate or inhibitor. As shown in Fig. Fig.33 (left panel), the pattern of K+ association with cells of the vector control transformant of K+-depleted E. coli TK2420 was identical to that of transformants expressing orf8 or orf8-orf9; each of these showed a significant initial increase in cell-associated K+ followed by a slower phase. The transformant expressing orf9 alone exhibited a much greater and more sustained increase in cell-associated K+ than did the other transformants. As shown in the middle panel, addition of ammonium at the same time as the K+ abolished the effect of Orf9. Although not shown, addition of the same concentration of methylamine did not inhibit the Orf9-dependent increase in cell-associated K+. A significant amount of the cell-associated K+ measured in this assay may be bound to the outside of the cell, although some residual K+ uptake capacity presumably exists in the K+-uptake-deficient E. coli mutant since it is viable. To conservatively correct for binding, the K+ associated with the control transformant (empty vector) was subtracted from that found for the others at each time point. This is shown in the right panel of Fig. Fig.3,3, where the patterns of transport for the transformants with orf8, orf8-orf9, and orf9 with and without added ammonium are shown after subtraction of the uptake observed for the vector control transformant.
The well-buffered semidefined medium that is routinely used in our laboratory is supplemented by addition of malate, trace salts, and yeast extract and contains 1 mM added K+. Growth experiments next compared growth rates of the wild-type alkaliphile and the panel of deletion strains on this medium and the same medium with either 100 or 200 mM added K+. Under the conditions used, which involved growth in conical tubes, the growth of the wild type was slightly lower at pH 10.5 than at pH 7.5, differing in that respect from results in more highly aerated continuous culture or flask cultures (18, 51). Under these conditions, only the strain with a single deletion in orf8 alone exhibited a phenotype (Fig. (Fig.4).4). This effect of orf8 deletion was negative and was not affected by pH. It was, however, abolished by deletion of both orf8 and orf9 instead of just orf8 alone. Reasoning from the effects of expression of orf9 versus expression of orf8-orf9 in E. coli TK2420 (Fig. (Fig.3),3), we hypothesized that Orf8 negatively modulated a K+ efflux capacity and that the deletion in orf8 led to adverse loss of K+ from the alkaliphile at both pH values. However, addition of high external [K+] did not improve the growth of the orf8 deletion mutant (Fig. (Fig.4).4). This raised the possibility that Orf9 catalyzes fluxes of something other than K+ that was the physiologically important substrate in the alkaliphile. Growth experiments were therefore initiated in a defined medium, with a view towards exploring variables that might shed light on such an alternate flux. Given the inhibition of K+ uptake by NH4+ in the E. coli experiments (Fig. (Fig.3),3), a plausible alternate substrate was NH4+. Loss of NH4+ from the orf8 deletion strain or uncoupling by ammonium cycling might account for its growth defect in the semidefined alkaliphile medium in which a low concentration of yeast extract was the sole nitrogen source.
A very different growth pattern from that observed in the semidefined yeast extract-containing medium (Fig. (Fig.4)4) was seen in the defined QA medium (Fig. (Fig.5).5). The defined medium contained malate as a carbon source and high concentrations of glutamine and alanine as a nitrogen source and as an additional carbon source. This medium was developed empirically. The high amine content, i.e., 18 mM, of the QA was optimal for growth of the wild-type strain. However, growth on the QA medium reduced the ability of the alkaliphile to grow at pH 10.5 relative to pH 7.5 beyond that noted in semidefined medium with the tube culture experiments (Fig. (Fig.44 versus Fig. Fig.5A).5A). Ammonium cannot be used as the sole nitrogen source at very high pH, where much of it is lost as ammonia, but nonetheless, the cytosolic accumulation of ammonium interferes with pH homeostasis (44). As shown in Fig. Fig.5A,5A, the only mutant strain that grew in the QA medium was the orf8 deletion strain. Strikingly, its growth was more modest than that of the wild type at pH 7.5 but much greater than that of the wild type at pH 10.5. Addition of 100 mM KCl to the medium (with added NaCl as a control for the added ionic strength) did not “rescue” any of the other mutants and appeared to particularly stress the wild type at highly alkaline pH. The only positive effect of KCl or NaCl addition was a modest improvement of the growth of both the wild type and mutant by addition of NaCl at pH 7.5.
The growth patterns of the wild-type and mutant alkaliphile strains on QA medium were consistent with the hypothesis that NH4+ is a physiologically relevant substrate of Orf9 in the alkaliphile and might be required to release excess ammonium in a medium with high levels of amine-nitrogen; upon deletion of orf8, Orf9-dependent ammonium release could occur. Such NH4+ efflux would be enhanced by high external pH at which much of the effluxed substrate would be converted to ammonia gas. This would reduce the risk of reuptake if Orf9 catalyzes bidirectional NH4+ translocation, as inferred from the E. coli experiments for its translocation of K+. To assess whether Orf9 could facilitate NH4+ acquisition when the amine-nitrogen level was suboptimal, growth of the wild type and that of the orf8 mutant were compared at high pH in QA medium containing one-quarter of the usual concentration of QA. As shown in Fig. Fig.5B,5B, the mutant grew better than the wild type even without added NH4+ and responded better than the wild type to addition of subinhibitory concentrations of NH4+. The results were consistent with a capacity of Orf9 to enhance NH4+ uptake. The better growth of the orf8 mutant even in the absence of added NH4+ was likely due to the presence of some free NH4 that is released from the amino acids in the medium (as was observed in experiments described below).
Experiments monitoring cytosolic ammonium contents of wild-type and mutant cells were then undertaken to further test the idea that the orf9 locus is involved in ammonium homeostasis, i.e., minimizing cytosolic NH4+ accumulation under amine-nitrogen-replete conditions. The experiments were conducted on the wild type and orf8, orf9, and orf8-orf9 mutants that were grown on semidefined medium at pH 7.5 and then washed and resuspended in QA medium containing one-half of the usual glutamine and alanine content. This was found optimal for the assays in pilot experiments. Malate was present as a major carbon source. The pH used for the medium was pH 7.5, because the background ammonia-ammonium produced at higher pH values made it impossible to assess Orf9-dependent effects. Immediately after the shift (“zero time”) and after 3 and 6 h of further incubation, the ammonium content of the cells was determined using an ammonium electrode, as described under Materials and Methods. The time frame was chosen empirically to allow sufficient catabolism of amino acids to produce cytoplasmic ammonium and mimic the condition that is nonpermissive for growth of orf9 mutants. As shown in Table Table3,3, the ammonium concentrations inside the cells of the wild-type and mutant strains were comparable right after the shift. After 3 h of incubation in QA medium, the mutant cells all had much higher concentrations of cytosolic ammonium than did the wild-type cells, even the orf8 mutant strain. After incubation for an additional 3 h, the orf8 mutant exhibited a significant reduction in cytoplasmic ammonium content, bringing it much closer to the lower wild-type levels, whereas the two mutants with a deletion in orf9 showed a further, substantial increase in cytoplasmic ammonium. The capacity of the orf8 deletion mutant to reduce its cytosolic ammonium relative to the mutant strains without orf9 probably accounts for its ability to grow on QA medium. The fact that the orf8 mutant first accumulates cytosolic ammonium far above levels seen in the wild type may indicate that the Orf9 functioning in the absence of Orf8 is not precisely equivalent to the wild-type Orf9 that functions in its presence (see Discussion).
An assay of ammonium uptake was then developed in RSO vesicles from the wild-type and mutant alkaliphile strains. Fifteen-minute data points, but not earlier points, were reproducibly obtained by a protocol in which separation of the vesicles from the supernatant was achieved by a high-speed centrifugation step (described under Materials and Methods). As shown in Table Table4,4, all of the vesicle preparations exhibited ammonium uptake in the presence of the electron donor ascorbate-PMS but little or no uptake in the absence of the electron donor. In the presence of ascorbate-PMS, the ammonium uptake activity of vesicles from the orf8 mutant was more than double that of the wild type at 15 min, whereas the activity of vesicles from the orf9 and orf8-orf9 mutant strains was less than half of the wild-type activity. The ammonium uptake by all the vesicle preparations, assayed in the presence of 1 mM NH4+, was inhibited by the addition of 1 mM K+ at the same time, with the inhibition of the wild-type and orf8 mutant preparations (80 to 86%) exceeding that observed in the vesicles without orf9 (60 to 70%). Although not shown, experiments were conducted to test the effect of preloading a concentrated suspension of RSO vesicles with K+ (10 mM) followed by dilution into buffer to which 1 mM NH4+ was added. The intravesicular K+ did not enhance ammonium uptake, either in the presence or in the absence of electron donor, as would be likely if Orf9 were capable of catalyzing K+/NH4+ antiport. The presence of K+/NH4+ antiport in bacteria has been proposed (26) but not yet associated with a particular gene product.
Because of the reported activities of other members of the CPA-2 transporter family and the sequence similarity of Orf7 to MarR, the effects of a variety of other conditions were examined for their effects upon the orf9 and orf7-orf10 deletion strains in comparison with the wild type. These experiments are not shown because they revealed no differences between the wild-type and mutant strains. They included sensitivity to sublethal concentrations of antibiotics including chloramphenicol, tetracycline, acriflavine, rifampin, and benzalkonium; sensitivity to methylglyoxal; and growth at reduced or elevated levels of added Na+ or Fe2+.
There was no specific indication for an involvement of the orf9 gene locus in sporulation or germination, e.g., by the predicted promoter. Indeed, the RT-PCR work and the effect of mutations in the orf9 locus on vegetative phenomena indicated that expression was probably not restricted to sporulation. Nevertheless, the precedent of involvement of several monovalent cation transporters from Bacillus, including CPA-2 transporter members (50, 52, 53) as well as the mrp antiporter (30), in sporulation or germination made it of interest to examine this possibility. Further, amino acids and peptides have important roles in various aspects of the sporulation-germination program and at least one global regulator, CodY, regulates aspects of both nitrogen utilization and sporulation (16). Even if Orf9 were an ammonium transporter with housekeeping functions, it might also have an indirect but significant role in ammonium homeostasis during sporulation.
After 6 days on sporulation agar, wild-type cultures contained approximately 95% phase-bright spores with approximately 5% remaining vegetative cells. The orf9 mutant had a much-reduced spore yield, 5 to 25% phase-bright spores, and the remaining cells were predominantly lysed cells. The orf7-orf10 mutant was similar in sporulation frequency and lysis to the orf9 mutant but exhibited a somewhat higher sporulation frequency at 10 to 30%. Although the sporulation agar does not give full synchrony, electron microscopy of wild-type cultures showed approximately 50% of cells with prespores around state IV after 24 h and about 80% of cells at around stage IV to V after 48 h (Fig. (Fig.6A6A and B). By contrast, cultures of the orf9 deletion strain contained only 10 to 20% of cells with prespores at 24 h, and at 48 h, 80% of the cells had lysed. Many of the lysed mutant cells contained multiple septa. It was also noted that at 24 h a large number of the mutant cells contained lightly staining granules of unknown nature, which were still present in the cell ghosts after 48 h (Fig. (Fig.6C6C and D). Essentially identical results were obtained for the orf7-orf10 mutant.
After extensive washing, spore preparations of the mutants were obtained with approximately 50% phase-bright spores and 50% lysed debris. Electron microscopic examination did not reveal detectable structural differences between the wild-type and mutant spores (data not shown). Some modest differences in germination properties were observed, however, which might still reflect undetected differences between the wild-type and mutant spores rather than a direct relationship between the mutations and germination per se. Incubation of wild-type spores in AMP buffer at 30°C with either NaCl or KCl, at 200 mM, but not in their absence, resulted in slow germination after a significant lag, as measured by the reduction in OD at 490 nm. NaCl was significantly more efficacious than KCl in supporting germination and was used for subsequent studies. pH 10 was used routinely after it was found that germination of wild-type spores was comparable at pH 10 to 11.9 but was barely detectable at pH 9 and below. As shown in Fig. Fig.7A,7A, germination of wild-type spores in the presence of 200 mM NaCl was significantly enhanced by addition of either l-alanine or inosine. Although not shown, neither of the organic germinants resulted in any germination when added to AMP buffer in the absence of NaCl or KCl. As shown in Fig. Fig.7B7B and C, the germination response of the orf9 and orf7-orf10 mutant strains was less than that of the wild type with NaCl and either l-alanine or inosine.
The focus of the present study was Orf9, its activities and its modulation by the presence or absence of Orf8. In the K+-uptake-deficient mutant of E. coli, TK2420, expression of this CPA-2 transporter exacerbates the K+-dependent phenotype (Fig. (Fig.2)2) and reduces cytosolic K+ accumulation when the transporter is not modulated by the presence of Orf8. Orf9 is thus Kef (K+ efflux)-like in this respect. Evidence for an Orf9-dependent K+ uptake capacity was shown with K+-depleted cells (Fig. (Fig.3).3). Together, the results suggest that the K+ translocation capacity of Orf9 is bidirectional. No evidence for an antiport mechanism was found in the E. coli experiments. Rb+ probably does not serve as a substrate for Orf9 since no Orf9-dependent activities could be monitored via 86Rb+. We were thus far unsuccessful in establishing a vesicle assay of ammonium uptake in E. coli vesicles, comparable to that developed in alkaliphile vesicles. However, an inhibitory effect of ammonium, but not methylamine, was demonstrated in the uptake experiments conducted in K+-depleted E. coli cells.
The likelihood that NH4+ is actually a substrate for Orf9 emerged from the inability of the alkaliphile mutants, except for the orf8 mutant, to grow in QA medium (Fig. (Fig.5A).5A). Supporting this idea was the pattern of orf9 mutant versus wild-type ammonium accumulation during a prolonged incubation under conditions in which QA catabolism could produce cytosolic ammonium (Table (Table3).3). The “rescue” of the alkaliphile by deletion of orf8 correlated with the ability of this mutant to lower its cytosolic ammonium accumulation to only about twice that observed in the wild type by the end of a 6-h incubation period in QA medium (with one-half the normal amine-nitrogen and pH 7.5 used for technical reasons). Under the same conditions, the orf9 mutants progressively accumulated about 10 times the wild-type level of ammonium. The orf8 deletion more effectively supported growth of the orf8 mutant on QA medium at pH 10.5 than at pH 7.5, such that the mutant growth significantly exceeded that of the wild type only at the higher pH (Fig. (Fig.5A).5A). This is consistent with the hypothesis that the reduced alkaliphily of B. pseudofirmus OF4 in the QA medium is a result of the severe challenge that cytosolic ammonium accumulation poses to pH homeostasis at the higher pH. At pH 7.5, there is a significant risk that the enhanced flux of ammonium would be accompanied by reuptake, especially if Orf9-mediated transport is bidirectional; by contrast, the ammonium effluxed at pH 10.5 would substantially be lost as ammonia gas. Supporting an ammonium uptake capacity of Orf9 is the further observation that the orf8 mutation enhances growth, relative to the wild type, in QA media with suboptimal QA concentrations at pH 10.5 (Fig. (Fig.5B5B).
The strongest evidence for an ammonium transport function for Orf9 was the observation of greater electron donor-dependent, K+-inhibitable ammonium uptake in wild-type and orf8 mutant RSO vesicles than in vesicles from strains lacking an intact orf9 (Table (Table4).4). The dependence upon an added electron donor, ascorbate-PMS, indicates that the proton motive force is a driving force. This in turn suggests that the transport is an electrogenic transport of the charged ammonium ion. The inhibitory effect of K+ supports electrogenicity. K+ inhibition would not be anticipated if the Orf9 substrate were ammonia gas, as has been suggested by some investigators (48, 49), but not all (35, 36), for Amt-MEP-type ammonium transporters. The present studies do not establish that K+ is also a substrate for Orf9 in the alkaliphile. Given the results in E. coli, this is possible, but since no K+-related phenotype was evident in the orf8 mutant, it is likely that any Orf9-dependent K+ fluxes in the alkaliphile setting are minor. The assays of Orf9-dependent activities in this study were all conducted in an orf8 mutant of B. pseudofirmus OF4 that has a full complement of K+ transport systems as well as at least one alternate ammonium transport system (see below). This makes it impossible to use a whole-cell assay of K+ uptake comparable to that used in a strain of E. coli with multiple mutations.
The data are incomplete in that direct evidence for Orf9-dependent K+ fluxes derives only from the work in E. coli and evidence for ammonium fluxes derives only from the work in the alkaliphile. Nonetheless, the transport and phenotypic data, together, support the conclusion that Orf9 has a capacity to translocate ammonium, that this flux is inhibited by K+, and that K+ is an alternate substrate at least in some contexts. Ammonium transport is a new function for the widely distributed CPA-2 transporter family, one that is likely to be found in other prokaryotes since there are other operons in the databases that have orf8-orf9 homologues (e.g., YhaT-YhaU in B. subtilis). Interestingly, the yhaU gene that encodes the closest homologue of Orf9 has been reported to be part of the σW regulon of B. subtilis that is induced upon alkaline shock (57). Juxtaposition of a cross-inhibition between K+ and NH4+ transport and proposals for transport proteins that use both of these cations as substrates are not themselves new. Inhibition by NH4+ has been observed in some K+ channels (5, 36, 47), and fluxes of both cations have been detected in studies of the plant KAT1 channel (36). A capacity for both K+ and NH4+ translocation has also been suggested for the Kdp transporter of E. coli on the basis of indications of Kdp-dependent futile cycling of ammonium under conditions of high ammonium and limiting K+ (9).
The conjoining of a capacity for K+ and ammonium ion flux in Orf9 raises the possibility that, even under energy depletion, the activation of Orf9 could mediate active ammonium uptake. Were a low intracellular nitrogen signal, e.g., low glutamine levels, to cause Orf9 activation, initial K+ efflux could set up a potential that could drive subsequent active ammonium ion accumulation via the same channel or uniporter. Such a sequence of activities might be considered in connection with the acidification of the cytoplasm that follows electrophile-activated efflux of K+ via KefB/C in E. coli (13) if those CPA-2 proteins turn out to also have the additional capacity to transport ammonium. Orf9-mediated ammonium uptake, energized by a proton motive force, could occur even though cytosolic K+ concentrations are substantial. In contrast to this proton motive force-energized uptake flux, ammonium efflux during growth on media with high amine-nitrogen levels would have to overcome the potential inhibition by high cytosolic K+ as well as the presence of a counterproductive chemiosmotic driving force. Perhaps high cytosolic ammonium concentrations reduce the proton motive force by uncoupling effects. Solutions to the specific problem of inhibition of ammonium efflux by cytosolic K+ could include one or more of the following: (i) that ammonium is not effluxed by Orf9 until it sufficiently exceeds the K+ concentration; (ii) that because of high binding of K+ within the cytosol, the active K+ concentration is too low to compete effectively with accumulating ammonium; or (iii) that accumulating ammonium replaces cytosolic K+ significantly, as has been suggested to occur in E. coli (10).
The capacity for ammonium uptake by RSO vesicles from orf9 and orf8-orf9 mutants of the alkaliphile was lower than the capacity of those from the wild-type strain or orf8 mutant, but the vesicles without Orf9 still exhibited significant ammonium uptake that was dependent upon an electron donor. This transport was inhibited significantly by K+, although that inhibition was slightly less than that exhibited by the Orf9-mediated flux (Table (Table4).4). The clear inference is that B. pseudofirmus OF4 possesses at least one other ammonium transporter. It is likely that this alkaliphile possesses a gene encoding a member of the widely found Amt-MEP family (35, 46, 48, 49). There is only one gene (NP244701.1) that is predicted to encode an ammonium transporter in the annotation of the alkaliphilic B. halodurans C-125 genome. That protein is an Amt-MEP family member, albeit not closely related to the AmtB protein of E. coli or the putative family member from B. subtilis. If indeed an Amt type of protein is responsible for the residual ammonium uptake detected in the RSO vesicles of B. pseudofirmus OF4, the strong dependence of that uptake upon the proton motive force would suggest that this alkaliphile transporter also catalyzes transport of the ionized substrate.
A channel mechanism has been suggested for the E. coli KefB and KefC proteins to which Orf9 shows sequence similarity (8). A bidirectional channel or uniport mechanism is hypothesized for Orf9, since attempts to demonstrate antiport mechanisms were all negative. The role of Orf8 could be some combination of a critical gating, deactivating, or modulatory element (42, 45), similar to the KTN proteins that are found in association with diverse K+ channels and transporters (12, 45). In the E. coli setting, in which orf9 was expressed either alone or together with orf8, the presence of Orf8 appeared to largely abolish the Orf9-mediated effects upon K+ content and fluxes (Fig. (Fig.22 and and3).3). In the alkaliphile too, a suppressive effect of Orf8 on Orf9-mediated transport activities was inferred from the detection of a phenotype in the orf8 deletion mutant that was abolished when orf9 was also deleted (Fig. (Fig.44 and and5;5; Table Table4).4). However, there are also indications that the properties of Orf9 are different in the wild-type alkaliphile than in the orf8 deletion mutant. In the assays of cytosolic ammonium (Table (Table3),3), the wild type did not accumulate significant cytosolic ammonium after 3 or 6 h of incubation on QA medium whereas the orf8 mutant reduced its cytosolic ammonium levels only after 3 h of incubation and initial high accumulation. This suggests that the presence of Orf8 modulates the activity of Orf9 rather than simply determining whether it is “on” or “off” with respect to transport. For example, it might affect the affinity of the Orf9-mediated transport as well as its activation-gating state. The active form that functions in the total absence of Orf8 might be a higher-flux-lower-affinity form or have a higher capacity for uptake relative to efflux than the form that is modulated by Orf8. Such modulatory effects of Orf8 could also account for the higher ammonium uptake activity of the vesicles from the orf8 mutant than of those from the wild type (Table (Table4).4). Since the TrkA-C domain found in Orf8 participates in a variety of effects of other KTN-containing, K+-translocating proteins (12, 45), the possibility that Orf8 has roles in changing both the on-off state and the transport properties of Orf9 is worthy of further exploration.
It was notable that individual nonpolar disruptions in orf7, orf9, and orf10 as well as the orf7-orf10 deletion all had similar growth phenotypes. Orf7 is likely to be a regulatory protein. Since a deletion in orf7 results in the same phenotype as does the orf9 deletion, Orf7 is hypothesized to have a regulatory activity required for expression of other genes in the operon. It will be of interest to determine whether regulation is mediated by cytosolic ammonium levels. Designation of orf7 to orf9, respectively, as amhR, amhM, and amhT is suggested, to represent their proposed regulatory, modulatory, and transport roles in support of ammonium homeostasis. Possibilities for Orf10 involvement are not as easily suggested by proteins to which it exhibits sequence similarity. However, it could turn out to have some chaperone or assembly function necessary for Orf9 activity. The downstream group of genes also do not offer clear-cut functional clues, but the closest homologues of several of the predicted gene products are associated with aspects of nitrogen metabolism.
A robust capacity for minimizing rises in cytoplasmic ammonium is particularly important for the alkaliphile growing at high pH, since the extraordinary capacity of the alkaliphile to acidify its cytoplasm relative to the medium is central to extreme alkaliphily (32, 51). Ammonia produced in the cytoplasm from amino acid catabolism would be protonated at the expense of that pH homeostatic mechanism. Similarly, the damaging effect of mutations in the orf9 locus on formation of normal endospores is likely to reflect the need for a robust capacity for handling the amino groups released from the substantial amino acid complement of the sporulation medium, especially at highly alkaline pH. The defect in organic germinant-dependent germination of mutant spores may reflect a role for Orf9 function during the germination process. However, given the gross abnormalities in those mutant sporangia that do not give mature spores, it is more likely that the spores formed successfully have subtle structural or physiological defects that reduce the rate of response to germinants. Clearly Orf9 is not serving a very specific role in germination comparable to that of monovalent cation/proton antiporter members of the CPA-2 family such as GrmA and GerN. In view of the particular importance of nitrogen-ammonium balance in an extreme alkaliphile, it is notable that thus far the operon described here appears to be unique. Among the Bacillus species, the nonalkaliphilic B. subtilis has a smaller putative operon, yhaS to yhaU, containing orf7 to orf9 homologues and alkaliphilic B. halodurans C-125 has several genes that may encode CPA-2 proteins but are not arranged in a comparable operon. Since B. pseudofirmus OF4 grows to higher pH values on more defined, nonfermentative media than does B. halodurans C-125 (22), it may well be that it has developed special strategies for handling nitrogen metabolism in a manner that does not jeopardize its pH homeostatic capacity.
The work described was supported by a BBSRC studentship to T.W.S.; a Grant-in-Aid for Scientific Research on Priority Areas (C) “Genomic Biology” from the Ministry of Education, Culture, Sports, Science and Technology of Japan to M.I.; and research grant GM28454 from the National Institutes of Health to T.A.K.