NasFED Proteins Mediate Assimilatory Nitrate and Nitrite Transport in Klebsiella oxytoca (pneumoniae) M5al

Abstract

Klebsiella oxytoca can use nitrate and nitrite as sole nitrogen sources. The enzymes required for nitrate and nitrite assimilation are encoded by the nasFEDCBA operon. We report here the complete nasFED sequence. Sequence comparisons indicate that the nasFED genes encode components of a conventional periplasmic binding protein-dependent transport system consisting of a periplasmic binding protein (NasF), a homodimeric intrinsic membrane protein (NasE), and a homodimeric ATP-binding cassette (ABC) protein (NasD). The NasF protein and the related NrtA and CmpA proteins of cyanobacteria contain leader (signal) sequences with the double-arginine motif that is hypothesized to direct prefolded proteins to an alternate protein export pathway. The NasE protein and the related NrtB and CmpB proteins of cyanobacteria contain unusual variants of the EAA loop sequence that defines membrane-intrinsic proteins of ABC transporters. To characterize nitrate and nitrite transport, we constructed in-frame nonpolar deletions of the chromosomal nasFED genes. Growth tests coupled with nitrate and nitrite uptake assays revealed that the nasFED genes are essential for nitrate transport and participate in nitrite transport as well. Interestingly, the ΔnasF strain exhibited leaky phenotypes, particularly at elevated nitrate concentrations, suggesting that the NasED proteins are not fully dependent on the NasF protein.

Klebsiella oxytoca, a member of the family Enterobacteriaceae, can assimilate nitrate (NO₃⁻) and nitrite (NO₂⁻) as sole nitrogen sources during aerobic growth. Nitrate assimilation takes place by three sequential steps: (i) nitrate transport into the cell by a specific nitrate permease; (ii) reduction to nitrite by assimilatory nitrate reductase; and (iii) further reduction to ammonium by assimilatory nitrite reductase (reviewed in reference 19). The resulting ammonium is then incorporated into central metabolism through the action of glutamine synthetase and glutamate synthase (26; reviewed in reference 35).

K. oxytoca genes involved in nitrate and nitrite assimilation are organized in a cluster, nasRFEDCBA (Fig. 1). The nasFEDCBA operon encodes the enzymes for uptake and reduction of nitrate and nitrite. Results from both genetic analysis and sequence comparisons indicate that the nasFED genes encode a nitrate and nitrite uptake system, the nasCA genes encode the two subunits of assimilatory nitrate reductase, and the nasB gene encodes assimilatory nitrite reductase (16, 17). nasF operon expression is regulated by general nitrogen control, via the NtrC transcription activator (2, 7), and by pathway-specific nitrate and nitrite induction, via the NasR transcription antiterminator (12, 18).

FIG. 1.

Physical map of the nasRFEDCBA-narLXKG region and subclones used for constructing deletions (modified from references 12 and 17). Restriction site abbreviations: B, BamHI; C, ClaI; E, EcoRI; G, BglII; H, HindIII; K, KpnI; M, SmaI; N, NsiI; S, BspEI. Subclones pVJS2502, pVJS2520, and pVJS2562 were used for constructing the nasFED, nasB, and narKG deletions, respectively. Additional subclones used for allelic exchange are also shown. See Table 1 for additional information. The inset shows Nas protein functions deduced from this and previous studies; see the text for details.

Nitrate transport is the essential first step in nitrate assimilation. The lack of both a convenient nitrogen radioisotope and appropriate mutants has impeded progress in understanding bacterial nitrate transport (46). Nevertheless, nitrate and nitrite uptake by cyanobacteria has been thoroughly studied (reviewed in references 19 and 29). In recent years, mutational analysis coupled with cloning and sequence analysis has identified nitrate transport systems in representative bacterial species (reviewed in reference 19). In the cyanobacterium Synechococcus sp. strain PCC7942, nitrate transport is mediated by an ATP-binding cassette (ABC)-type system consisting of a periplasmic binding protein (encoded by nrtA), an integral membrane protein (encoded by nrtB), and two homologous ATP-binding proteins (encoded by nrtC and nrtD, respectively) (21, 30, 32; reviewed in reference 29). The K. oxytoca nasFED genes are, through their homology to nrtABD, implicated in nitrate transport (17).

We wished to examine directly the roles of the nasFED genes in mediating nitrate and nitrite transport. We constructed large in-frame deletions in each of these genes and transplanted the deletions via allelic exchange into the K. oxytoca chromosome. The resulting mutants were used in assays for nitrate and nitrite uptake under a range of experimental conditions. The NasE and NasD proteins were essential for nitrate uptake under all conditions, whereas the NasF protein (the presumed periplasmic binding protein) was important but not essential, at least at very high nitrate concentrations. By contrast, the NasFED proteins were not essential for nitrite uptake. These results are similar to those obtained with studies of Synechococcus sp. strain PCC7942 nrt mutants (20, 30; reviewed in references 19 and 29). We conclude that the nasFED genes encode a system for the transport of both nitrate and nitrite.

MATERIALS AND METHODS

Strains and plasmids.

K. oxytoca M5al strains, the Escherichia coli K-12 strain, and plasmids used are listed in Table 1. Aerobacter aerogenes M5al was designated K. pneumoniae (see reference 44). However, its phenotypic properties (such as a positive reaction in the indole test) place this strain in the genus K. oxytoca (33). Genetic crosses were performed by bacteriophage P1 kc-mediated transduction (24). Transposon MudJ is a bacteriophage transposon which encodes kanamycin resistance and forms lacZ operon fusions (22). Standard methods were used for restriction endonuclease digestion, ligation, and transformation of DNA (22).

TABLE 1.

Strains and plasmids

Strain or plasmid	Genotype	Reference or source
K. oxytoca M5al strains
VJSK009	hsdR1	7
VJSK573	hasdR1 Δlac-2001 nasC123::Ω-Cm nasB16::MudJ Galr Tn7	Laboratory collection
VJSK2089	hsdR1 lacZ101::Tn10d(Tc) rpsL	Laboratory collection
Derivatives of K. oxytoca VJSK2089
VJSK1973	ΔnasF132	This work
VJSK1974	ΔnasE133	This work
VJSK1975	ΔnasD134	This work
VJSK1976	Δ(nasFE)135	This work
VJSK1977	Δ(nasED)136	This work
VJSK1978	Δ(nasFED)137	This work
VJSK2216	nas+ Δ(narKG)302	This work
VJSK2207	ΔnasF132 Δ(narKG)302	This work
VJSK2208	ΔnasE133 Δ(narKG)302	This work
VJSK2209	ΔnasD134 Δ(narKG)302	This work
VJSK2210	Δ(nasFE)135 Δ(narKG)302	This work
VJSK2211	Δ(nasED)136 Δ(narKG)302	This work
VJSK2212	Δ(nasFED)137 Δ(narKG)302	This work
VJSK2214	ΔnasB138 Δ(narKG)302	This work
VJSK2217	ΔnasF132 ΔnasB138 Δ(narKG)302	This work
VJSK2218	ΔnasE133 ΔnasB138 Δ(narKG)302	This work
VJSK2219	ΔnasD134 ΔnasB138 Δ(narKG)302	This work
VJSK2220	Δ(nasFE)135 ΔnasB138 Δ(narKG)302	This work
VJSK2221	Δ(nasED)136 ΔnasB138 Δ(narKG)302	This work
VJSK2222	Δ(nasFED)137 ΔnasB138 Δ(narKG)302	This work
E. coli K-12 strain S17-1 λpir	hsdR supE44 endA thi pro recA RP4-2-Tc::Mu-Km::Tn7 λpir	42
Plasmids
pALTER-1	Tcr (vector for site-specific and loop deletion mutagenesis)	Promega
pKAS46	Apr Kmr, ori R6K rpsL+ (suicide vector)	42
pVJS2502	Tcr, nasFED; ∼5.7-kb HindIII-BamHI fragment in pALTER-1 (see Fig. 1)	This work
pVJS2549	Apr Kmr, ΔnasF132; KpnI-EcoRI fragment in pKAS46 (see Fig. 1)	This work
pVJS2550	Apr Kmr, ΔnasE133; KpnI-EcoRI fragment in pKAS46 (see Fig. 1)	This work
pVJS2515	Apr Kmr, ΔnasD134; SmaI-BamHI fragment in pKAS46 (see Fig. 1)	This work
pVJS2553	Apr Kmr, Δ(nasFE)135; KpnI-EcoRI fragment in pKAS46 (see Fig. 1)	This work
pVJS2554	Apr Kmr, Δ(nasED)136; KpnI-BamHI fragment in pKAS46 (see Fig. 1)	This work
pVJS2557	Apr Kmr, Δ(nasFED)137; KpnI-BamHI fragment in pKAS46 (see Fig. 1)	This work
pVJS2520	Tcr, nasCBA; ∼7.8-kb insert fragment in pALTER-1 (see Fig. 1)	This work
pVJS2546	Apr Kmr, ΔnasB138; EcoRI-KpnI fragment in pKAS46 (see Fig. 1)	This work
pVJS2560	Tcr, narKG; ∼6-kb BamHI fragment in pALTER-1 (see Fig. 1)	This work
pVJS2562	Tcr Apr, as pVJS2560 but narK(A115G) (see Fig. 1)	This work
pVJS2566	Apr Kmr, Δ(narKG)302; BamHI fragment in pKAS46 (see Fig. 1)	This work

Culture media.

Defined, complex, and indicator media for routine genetic manipulations were used as described previously (22). Nitrogen-free medium contained 0.2% (wt/vol) glucose, 1% (wt/vol) sodium citrate, 0.74% (wt/vol) sodium phosphate (pH 8), and 1 mM MgSO₄ (16). This medium was supplemented with additional nitrogen sources (5 mM NaNO₃, NaNO₂, or NH₄Cl, or 2.5 mM arginine) as indicated to test nitrogen utilization phenotypes. MacConkey nitrate agar (43) and LB-nitrate-formate agar (11) were used to test the phenotype of Δ(narKG) deletion mutants. 5-Bromo-4-chloro-3-indolyl-β-d-galactopyranoside (X-Gal; 40 μg/ml) was used to score the Lac phenotypes of MudJ insertion strains. E. coli transformants were selected on ampicillin (200 μg/ml). Chloramphenicol was used at 50 and 25 μg/ml for selecting K. oxytoca and E. coli transformants, respectively. Kanamycin was used at 100 and 75 μg/ml for selecting K. oxytoca and E. coli transformants, respectively. Streptomycin was used at 500 μg/ml, and tetracycline was used at 20 μg/ml.

Liquid medium to grow cultures for uptake and β-galactosidase assays was buffered with 3-[N-morpholino]propanesulfonic acid (MOPS) as previously described (43). The initial pH of this medium was adjusted with NaOH to 8.0. Glucose (40 mM) was used as the sole carbon source except as indicated. The nitrogen sources NaNO₃ (5 mM), NaNO₂ (5 mM), and l-glutamine (5 mM) were added as indicated. MOPS-, 2-[N-morpholino]ethanesulfonic acid (MES)-, and N-[2-hydroxyethyl]piperazine-N′-[3-propanesulfonic acid] (EPPS)-buffered liquid media were used as indicated. The compositions of the last two media were the same as that of MOPS-buffered medium except that 80 mM MES or EPPS was substituted for MOPS. The initial pH values of MES- and EPPS-buffered media were adjusted with NaOH to 7.0 and 9.0, respectively.

Culture conditions.

Cultures for β-galactosidase assays, growth curves, and nitrate and nitrite uptake studies were grown at 30°C to minimize deamidation of glutamine (3). Culture densities were monitored with a Klett-Summerson photoelectric colorimeter (Klett Manufacturing Co., New York, N.Y.) equipped with a no. 66 (red) filter. For dry weight (DW) determinations, cell densities were monitored with a Beckman DU-50 spectrophotometer (wavelength of 600 nm). Aerated cultures were incubated at 240 rpm in 1/10 volume of medium in 125- or 250-ml sidearm flasks. Cultures in the mid-exponential phase were harvested, chilled on ice, and washed with saline or uptake buffer prior to enzymatic or uptake assays.

DNA sequencing.

The nasFED sequence was determined from double-stranded templates by the dideoxynucleotide chain termination method (40) with modified T7 DNA polymerase (45) and [α-³⁵S]dATP labeling (5). Some sequencing was performed on an automated 373A stretch DNA sequencer by using dye terminator chemistry and Ampli Taq-FS DNA polymerase (Perkin-Elmer/Applied Biosystems Division, Foster City, Calif.). DNA templates for sequencing were prepared as described previously (15) or by using QIAprep spin plasmid kits (Qiagen Inc., Chatsworth, Calif.). DNA sequences were analyzed with programs from DNASTAR Inc. (Madison, Wis.), and database searches were performed with the BLAST programs (1).

Site-specific and loop deletion mutagenesis.

Oligonucleotide-directed site-specific mutagenesis used the Altered Sites system (Promega Corp., Madison, Wis.) as described previously (22). Loop deletion oligonucleotides contained an introduced NsiI site (ATGCAT) at the boundaries of the deleted sequence (Fig. 2A). The larger deletions of nasFE, nasED, and nasFD were constructed by adding two oligonucleotides to a single mutagenesis reaction. Double-deletion derivatives were isolated and subjected to NsiI reduction to excise the intervening sequence. In all cases, the newly introduced restriction site, as well as a deleted restriction site(s) where appropriate, was used as a marker(s) to screen the deletion. Deletions of nasFED, nasB, and narK-narG were constructed on plasmids pVJS2502, pVJS2520, and pVJS2562, respectively (Fig. 1). The narK-narG deletion was constructed by first introducing a BglII site within narK, followed by BglII reduction to excise the central fragment containing the 3′ portion of narK and the 5′ portion of narG (Fig. 1 and 2B).

FIG. 2.

Oligonucleotides (see Materials and Methods for details). (A) Mutagenic oligonucleotides used for constructing deletions of nasFED and nasB. The deduced numbers of aminoacyl residues in the wild-type (+) and mutant (Δ) proteins are indicated. (B) Mutagenic oligonucleotide used for introducing a BglII site into narK. (C) Primer oligonucleotides used for colony PCR analysis of chromosomal deletions.

Allelic exchange.

Deletions were transplanted into the chromosome of K. oxytoca by allelic exchange (42). Appropriate deletion-containing restriction fragments were subcloned into plasmid pKAS46, a conditionally replicating plasmid that requires the plasmid R6K π protein for maintenance. Deletion-containing subclones (Fig. 1) were conjugated from E. coli S17-1 λpir into K. oxytoca VJS2089, and kanamycin-resistant (Km^r) exconjugants were isolated and verified to be streptomycin sensitive (Sm^s) (due to the presence of rpsL⁺ on pKAS46). Segregants (Sm^r Km^s) were subsequently isolated and screened for the appropriate Nas or Nar phenotypes.

Whole-colony PCR was used to confirm the veracity of the allelic exchanges (39). A fresh colony was picked and resuspended in 50 μl of sterile distilled water. The final PCR mixture (100 μl) contained 5 μl of the colony suspension, 1× thermophilic PCR buffer, 2 mM MgCl₂, 150 μM deoxynucleoside triphosphate, a 1 mM concentration of each oligonucleotide primer, and 2.5 U of Taq DNA polymerase. Reactions were as follows: first, a single cycle of 94°C for 2 min, 59°C for 2 min, and 72°C for 3.5 min; second, 30 cycles of 94°C for 1 min, 59°C for 1 min, and 72°C for 2 min; and third, a single cycle of 94°C for 50 s, 59°C for 1 min, and 72°C for 3 min. Oligonucleotide primers are shown in Fig. 2C. Appropriate pairs of primers were chosen depending on the deletion; for example, ΔnasF was detected with primers F1 and F2, whereas Δ(nasFE) was detected with primers F1 and E2.

Nitrate and nitrite uptake.

Nitrate uptake was determined by measuring nitrite accumulation from externally added nitrate, since the ΔnasB Δ(narKG) mutant used cannot further reduce nitrite to ammonium. Assay mixtures contained cell suspension (final concentration, 0.16 mg [DW]/ml), 40 mM glucose, and 80 mM MOPS-NaOH buffer (pH 8.0) in a final volume of 1.0 ml. Time course assays used a final volume of 4.0 ml. Experiments to evaluate pH effects also used 80 mM MES-NaOH (pH 7.0) and EPPS-NaOH (pH 9.0) as indicated. Assays were initiated by adding NaNO₃ at a final concentration of 5 mM unless otherwise indicated. NH₄Cl and NaCl were added as indicated. The assay tube (10 ml) was incubated at 25°C. At defined time intervals, 0.2 ml of the assay mixture was sampled into 0.9 ml of 1% (wt/vol) sulfanilic acid–20% (vol/vol) HCl, mixed by vortexing, further mixed with 0.9 ml of Marshall’s reagent (N-naphthylethylene diamine hydrochloride, 0.129% [wt/vol]), and centrifuged. The concentration of nitrite in the supernatant was determined by measuring the A₅₄₀ (43). Uptake rates were determined during the first 10 min of the assay and are expressed as nanomoles of nitrite produced per minute per milligram (DW). Nitrite uptake was determined by measuring nitrite disappearance from the uptake mixture. The procedure was as described above except that NaNO₂ was added at a final concentration of 75 μM. All reported values are averages from at least two independent experiments.

Assimilatory nitrate reductase assay.

Reduced methyl viologen-dependent assimilatory nitrate reductase activity was determined in toluene-permeabilized cells (24) essentially as described previously (43) except that nitrate was used at 5 mM. Activity is expressed as nanomoles of nitrite produced per minute per milligram (DW) to facilitate comparison with measurements of uptake in intact cells. All reported values are averages from at least two independent experiments.

β-Galactosidase assays.

β-Galactosidase assays were done at room temperature, approximately 21°C. Cell pellets were suspended in 4 ml of Z buffer (24) and stored on ice. β-Galactosidase activity was measured in CHCl₃-sodium dodecyl sulfate-permeabilized cells by monitoring the hydrolysis of o-nitrophenyl-β-d-galactopyranoside. Activities are expressed in terms of cell density (A₆₀₀), by using the formula of Miller (24).

DW determination.

Cultures of strain VJSK2089 were grown in MOPS medium with 40 mM glucose and 5 mM glutamine. Mid-exponential-phase cultures (about 40 Klett units) were harvested by centrifugation. Pellets were washed twice, resuspended in distilled water, placed into tared aluminum weigh boats, and dried to constant weight at 80°C. The calculated cell DW was 0.456 mg/ml at an optical density at 600 nm of 1.0.

Nucleotide sequence accession number.

The DNA sequence reported in this paper has been deposited in the GenBank nucleotide sequence database under accession no. L27431.

RESULTS

nasFE sequence analysis.

Previous sequence analysis showed that the deduced K. oxytoca NasF, NasE, and NasD proteins are homologous to the deduced Synechococcus NrtA, NrtB, and NrtD proteins, respectively. However, four regions of the nasFE sequence were determined on only one strand, and the stop codon of the nasF gene and the start codon of the nasE gene were uncertain (17). Therefore, we completed and refined the entire nasFED sequence. A few errors, including some double frameshifts, were corrected.

The presumed translational start site for the nasF gene is 5′-TTTCTGGAGCGGTTATGGGC-3′, where the Shine-Dalgarno region and presumed start codon for the nasF gene are underlined. Conceptual translation of the nasF sequence yields a protein of 418 amino acids with molecular mass of 46,176 Da. The presumed translational start site for the nasE gene is 5′-CAACGTAAGGGGGCATGAGATGAAA-3′, where the presumed stop codon for the nasF gene and the Shine-Dalgarno region and presumed start codon for the nasE gene are underlined. Conceptual translation of the nasE sequence yields a protein of 294 amino acids with molecular mass of 32,302 Da. The presumed stop codon for the nasE gene is within the previously reported sequence 5′-AAATAA GGAGTCGCAGATGAAA-3′ (17), where the presumed stop codon for the nasE gene and the Shine-Dalgarno region and presumed start codon for the nasD gene are underlined. Conceptual translation of the nasD sequence yields a protein of 262 amino acids with molecular mass of 28,999 Da, slight differences from the values previously reported (17).

In-frame deletions in the nasFED genes.

Previous mutational and complementation analysis showed that the nasD gene is required for nitrate but not nitrite assimilation by K. oxytoca (17). To further examine the role(s) of the nasFED genes, we constructed six in-frame deletions within this region: ΔnasF, ΔnasE, ΔnasD, Δ(nasFE), Δ(nasED), and Δ(nasFED). Large in-frame deletions of the nasF, nasE, and nasD genes were created in plasmid pVJS2502 by oligonucleotide-mediated loop formation (see Materials and Methods and Fig. 1 and 2A). The loop deletion strategy was guided by the following principles: (i) the numbers of deleted nucleotides were 3n (n, number of amino acid residues); (ii) translational initiation and termination signals were not changed; and (iii) a unique restriction site (NsiI) was introduced at the deletion junction (Fig. 1 and 2A). The Δ(nasFE), Δ(nasED), and Δ(nasFED) deletions were created from NsiI reductions of the appropriate single deletions.

We transplanted the resulting deletions into the K. oxytoca chromosome by allelic exchange (42). After plasmid integration and segregation, potential deletion mutants were screened for nitrogen utilization phenotypes (Table 2). The mutants were unable to use nitrate as the sole nitrogen source (the ΔnasF mutant showed a leaky phenotype). We verified the authenticity of the chromosomal deletions by colony PCR coupled with NsiI restriction analysis. The sizes of PCR products from the wild type and mutants, and of NsiI-restricted fragments from mutants, were as expected (data not shown).

TABLE 2.

Phenotypes of the wild-type strain and nas deletion mutants of K. oxytoca

Strain	Genotype	Growth on nitrogen-free mediuma
		No supplement	+NO3−	+NO2−	+NH4+
VJSK2089	nas+	−	++	++	++
VJSK1973	ΔnasF132	−	±	++	++
VJSK1974	ΔnasE133	−	−	++	++
VJSK1975	ΔnasD134	−	−	++	++
VJSK1976	Δ(nasFE)135	−	−	++	++
VJSK1977	Δ(nasED)136	−	−	++	++
VJSK1978	Δ(nasFED)137	−	−	++	++

^a−, poor growth; ++, good growth; ±, very weak growth after 24 h incubation but substantial growth after 48 h. 

Polarity analysis of nasFED deletions.

We expected the in-frame deletions to be nonpolar on downstream gene expression. To determine this directly, we used generalized transduction to construct a series of strains, each carrying one of the nas deletions in cis with a chromosomal Φ(nasB-lacZ) operon fusion. The deletions are closely linked to the nasB::MudJ insertion, so we developed screens to identify the desired recombinant class, i.e., progeny that inherit the donor nasB::MudJ insertion but retain the recipient deletion allele. In an initial control experiment, we transduced nasB::MudJ from VJSK573 (nasC::Ω-Cm, nasB::MudJ) into the wild type (VJSK2089). We screened 148 Km^r transductants for their Lac phenotypes on X-Gal medium containing arginine as the sole nitrogen source (permissive for nasF operon expression). Four Lac⁺ colonies were identified, each of which proved to be chloramphenicol sensitive (Cm^s). By contrast, the remaining 144 Lac⁻ colonies were Cm^r. Thus, nasB::MudJ transductants that inherited the nasC::Ω-Cm insertion were Lac⁻ as expected, because of the strong polar effect on Φ(nasB-lacZ) expression.

We then performed a series of crosses, again using strain VJSK573 as the donor, with each of the six nas deletion strains as recipients. We reasoned that Cm^s transductants, which did not inherit the donor nasC::Ω-Cm allele, were likely to have retained the recipient nas deletion allele. Three to seven Km^r Cm^s transductants were identified from among 150 to 300 Km^r colonies in each cross. We used colony PCR coupled with NsiI restriction to analyze two Km^r Cm^s Lac⁺ strains from each cross to verify that these strains did retain the recipient’s nas deletion. Subsequent measurements of β-galactosidase activity revealed that the deletions had no significant effect on downstream Φ(nasB-lacZ) expression (the various strains synthesized 483 ± 69 Miller units).

Deletion of the narKG genes.

K. oxytoca can also use nitrate as an electron acceptor for anaerobic respiration. The narK gene, encoding a nitrate and/or nitrite transport protein (9, 38), and the narGHJI operon, encoding respiratory nitrate reductase, are located downstream of the nasFEDCBA operon (Fig. 1) (17). We therefore constructed a Δ(narKG) chromosomal deletion to eliminate potential complications from these activities. We determined the DNA sequence for much of the nar region, including the amino-terminal and carboxyl-terminal coding regions of narK and narG, respectively (data not shown). These sequences in K. oxytoca are homologous to those of E. coli. We used site-specific mutagenesis to introduce a BglII site at position 115 in the narK sequence. A subsequent BglII reduction deleted 95% of narK and 60% of narG. This deletion was transplanted into the K. oxytoca chromosome of the wild type and the nas deletion mutants by allelic exchange (see Materials and Methods). Segregants (Sm^r) were screened for the NarG⁻ phenotype on both MacConkey nitrate medium (43) and formate-nitrate medium (11). We used colony PCR coupled with BglII restriction to analyze two strains from each allelic replacement to verify that these strains contained the narKG deletion. These constructions yielded a set of seven strains, each carrying Δ(narKG); collectively, the strains carried the wild-type nas gene and all combinations of deletions of the nasFED genes (Table 1).

Growth tests.

To begin examining the function of the NasFED proteins in nitrate and nitrite assimilation, we determined the growth rates for Δ(narKG) derivatives of the nasFED⁺, ΔnasF, ΔnasE, and ΔnasD strains in defined media with glucose as the sole carbon source. Different buffers (MES, MOPS, and EPPS) were used to provide different pH values (7, 8, and 9, respectively). With glutamine (5 mM) as the sole nitrogen source (3), the wild type and mutants exhibited similar growth rates irrespective of pH (Fig. 3). With nitrate (5 mM) as the sole nitrogen source, only the nasFED⁺ strain grew, again at a rate that was indifferent to pH (Fig. 3). However, with nitrite (5 mM) as the sole nitrogen source, pH was an important variable. At pH 7 and 8, the wild type and the deletion mutants were indistinguishable. By contrast, at pH 9, the mutants displayed longer lag phases and slower growth rates, although the final cell yields approached those of the wild type. The ΔnasF, ΔnasE (Fig. 3), and ΔnasD (data not shown) mutants were indistinguishable in all of these tests.

FIG. 3.

Growth of Δnas deletion strains with nitrate and nitrite. The nas⁺ (VJSK2216), ΔnasF (VKSK2207), and ΔnasE (VJSK2208) strains were cultured in defined glucose media buffered with MES (pH 7.0), MOPS (pH 8.0), or EPPS (pH 9.0) as indicated. Nitrogen sources were 5 mM glutamine (▪), 5 mM nitrate (•), and 5 mM nitrite (▾). ○, no nitrogen source. All strains were nasB⁺ Δ(narKG).

We used the wild type and the ΔnasE mutant to further explore the pH effects. With glucose as the carbon source in EPPS-buffered medium, culture pH decreased from about 8.8 initially to about 8.2 at culture saturation (data not shown). With glucitol as the carbon source, culture pH decreased from about 8.9 to about 8.5. In both cases, significant growth occurred at pH values near 8.8, in which the calculated nitrous acid concentration is approximately 18 nM. Furthermore, we measured nitrite uptake at pH 7, 8, and 9 in the wild type and the ΔnasE mutant (see below).

We draw three conclusions from these results. First, consistent with results of genetic tests (above), the deletions were not significantly polar on downstream gene expression, because each mutant grew at the wild-type rate on nitrite (which requires nasB function). Second, each of the mutants was fully blocked in nitrate assimilation (at the 5 mM concentration tested; see below). Third, each of the mutants exhibited a NasFED-independent, pH-dependent ability to assimilate nitrite.

We also examined the effect of nitrate concentration on doubling time (data not shown). The nasFED⁺ strain reached its maximal doubling time with 20 mM nitrate. The ΔnasE and ΔnasD mutants failed to grow even with 100 mM nitrate. However, the ΔnasF mutant grew slightly with 10 mM nitrate and reached a maximal doubling time of about 25% of the wild-type value with 80 mM nitrate. The ΔnasF mutant’s leaky phenotype was also evident in plate tests (Table 2). This deletion removes more than half of the nasF coding region, so it is unlikely that the mutant protein retains function. Thus, a high external nitrate concentration was able to partially bypass the requirement for NasF, the presumed periplasmic binding protein, but not NasE and NasD, the presumed membrane and cytoplasmic components.

Nitrate uptake.

The term transport defines the movement of a substrate from one side of the cytoplasmic membrane to the other. An assay for transport therefore (usually) requires a radiolabeled substrate, to monitor intracellular accumulation. However, the radioisotope ¹³N has a half-life of approximately 10 min, requiring special facilities for its use (46). Uptake is an operational term, used when the transported substrate is further metabolized within the cell. Our experiments did not monitor transport directly but rather used indirect assays to measure nitrate and nitrite uptake.

We studied nitrate uptake by using a sensitive colorimetric assay to measure accumulation of its reduction product, nitrite. To block in vivo nitrite reduction, we constructed strains carrying a large in-frame deletion of nasB, the structural gene for assimilatory nitrite reductase (Fig. 1 and 2A). This deletion was constructed by loop mutagenesis as described above for the other nas deletions. The ΔnasB deletion was transplanted into the chromosome of the nas⁺ Δ(narKG) and the Δnas Δ(narKG) double mutants by allelic exchange (see Materials and Methods). Segregants (Sm^r) were screened for the NasB⁻ phenotype on defined medium containing nitrite as the sole nitrogen source. We used colony PCR coupled with NsiI restriction to analyze two strains from each allelic replacement to verify that these strains contained the nasB deletion. These constructions yielded a set of seven strains, each carrying ΔnasB and Δ(narKG); collectively, the strains carried the wild-type nas gene and all combinations of deletions of the nasFED genes (Table 1).

To measure nitrate uptake, ΔnasB Δ(narKG) strains were cultured to the mid-exponential phase in defined medium containing glucose, glutamine, and nitrate. Uptake assays were performed as described in Materials and Methods by measuring the time-dependent accumulation of nitrite resulting from transport and reduction of nitrate. In preliminary experiments, the Δ(nasFED), Δ(nasFE), and Δ(nasED) strains failed to accumulate detectable levels of nitrite. Subsequent experiments focused on the single-gene deletion strains. The nasFED⁺ strain accumulated nitrite at a significant rate, whereas accumulation by the ΔnasE and ΔnasD mutants was negligible (Fig. 4 and Table 3). Strikingly, the ΔnasF mutant accumulated nitrite at about one-half of the wild-type rate (Fig. 4 and Table 3). This observation is congruent with results of growth tests (above), which indicated that the ΔnasF strain retains significant capacity for nitrate assimilation.

FIG. 4.

Nitrate uptake by Δnas deletion strains. The nasFED⁺ (VJSK2214; •) ΔnasF (VJSK2217; ▾), ΔnasE (VJSK2218; ▪), and ΔnasD (VJSK2219; ▴) strains were cultured in defined glucose medium buffered with MOPS (pH 8.0) and supplemented with 5 mM glutamine as the nitrogen source and 5 mM nitrate to induce nasF operon expression. Nitrate uptake, estimated from the accumulation of nitrite, was determined in 80 mM MOPS-NaOH buffer (pH 8.0) as described in Materials and Methods. All strains were ΔnasB Δ(narKG).

TABLE 3.

Nitrate uptake and nitrate reductase activity in K. oxytoca^a

Strain	Genotype	Nitrate uptakeb (nmol of NO2− min−1 mg [DW]−1)	Nitrate reductasec (nmol of NO2− min−1 mg [DW]−1)
VJSK2214	nasFED+	9.2	10.2
VJSK2217	ΔnasF132	4.6	12.4
VJSK2218	ΔnasE133	<0.1	11.3
VJSK2219	ΔnasD134	<0.1	10.5

^aAll strains are ΔnasB Δ(narKG). 

^bGlucose-dependent nitrate uptake rates in intact cells were determined over a 10-min time course as described in Materials and Methods. 

^cReduced methyl viologen-dependent nitrate reductase enzyme activities in toluene-permeabilized cells were determined over a 10-min time course as described in Materials and Methods. 

Our assay for nitrate uptake measures two consecutive steps: transport, and reduction to nitrite. We therefore wished to measure nitrate reduction independent of transport, to determine whether nitrate reductase activity was limiting in the uptake assays. In preliminary experiments to evaluate a variety of permeabilizing treatments, we found that toluene (24) gave the most satisfactory results. In toluene-treated cells, the rates of reduced viologen-dependent nitrate reduction were indistinguishable in the wild-type, ΔnasF, ΔnasE, and ΔnasD strains (Table 3). This result stands in contrast to the rates of glucose-dependent nitrate uptake in whole cells (Table 3). We therefore conclude that nitrate reductase activity was not limiting in the uptake assays. These results further demonstrate that the Δnas deletions were not polar on downstream (nasA) gene expression.

Nitrate uptake rates were constant for the first 10 min of the assays (Fig. 4). Therefore, in most of the experiments described below, uptake rates were estimated from a single measurement at 10 min. Uptake was absolutely dependent on a source of energy (glucose) in the assay medium. To further characterize the factors affecting nitrate uptake, we determined the effects of growth medium, temperature, pH, ammonium, and sodium on nitrate uptake by the wild type.

(i) Growth medium.

Cultures were routinely grown in MOPS-buffered minimal medium with glucose as the carbon source, glutamine as the nitrogen source, and nitrate to induce nasF operon expression. The presence of nitrate in the growth medium caused a twofold increase in the measured rate of nitrate uptake: 9.0 nmol of nitrite min⁻¹ mg (DW)⁻¹ with nitrate versus 4.1 nmol of nitrite min⁻¹ mg (DW)⁻¹ with glutamine only.

(ii) Temperature.

Nitrate uptake rates at 25, 30, and 37°C were 8.8, 5.6, and 0.8 nmol of nitrite min⁻¹ mg (DW)⁻¹, respectively. Thus, the rate of nitrate uptake decreased 10-fold over a 12°C increase in assay temperature. This result indicates that nitrate uptake in K. oxytoca is temperature sensitive. All other uptake assays reported herein were performed at 25°C.

(iii) pH.

Nitrate uptake rates at pH 7 (MES), 8 (MOPS), and 9 (EPPS) were 7.3, 8.8, and 10.0 nmol of nitrite min⁻¹ mg (DW)⁻¹, respectively. Thus, the rate of nitrate uptake increased only slightly over a two-unit increase in assay pH. All other uptake assays reported herein were performed at pH 8.

(iv) Ammonium.

To determine if ammonium inhibits nitrate uptake, we added ammonium (final concentration, 0.5 mM) at 0 or 5 min in a standard time course experiment. Nitrate uptake was immediately inhibited upon the addition of ammonium (Fig. 5).

FIG. 5.

Effect of added ammonium on nitrate uptake. The nasFED⁺ ΔnasB Δ(narKG) strain (VJSK2214) was cultured in defined glucose medium buffered with MOPS (pH 8.0) and supplemented with 5 mM glutamine as the nitrogen source and 5 mM nitrate to induce nasF operon expression. Nitrate uptake, estimated from the accumulation of nitrite, was determined in 80 mM MOPS-NaOH buffer (pH 8.0) as described in Materials and Methods. Curve A shows uptake in the absence of added ammonium. Ammonium (0.5 mM) was added to the uptake assay at 0 min (curve B) or 5 min (curve C) after the addition of nitrate.

(v) Sodium.

It has been reported that sodium is required for Nrt-dependent nitrate uptake in Synechococcus sp. strain PCC7942 (36). We therefore determined whether nitrate uptake by K. oxytoca is influenced by external sodium. We prepared MOPS-glucose defined medium in which all sodium salts were replaced by their potassium equivalents. Nitrate uptake assays were performed in MOPS-KOH buffer (pH 8.0), and potassium nitrate was used as the substrate. The rate of nitrate uptake was determined in reaction mixtures containing 0 to 100 mM added NaCl. In these experiments, sodium had no effect on nitrate uptake by either the nasFED⁺ or the ΔnasF mutant. No nitrate uptake was observed in the Δ(nasED) and Δ(nasFED) mutants even in the presence of 100 mM sodium (data not shown). These results indicate that nitrate uptake in K. oxytoca is indifferent to the extracellular sodium concentration.

Nitrite uptake.

We studied nitrite uptake by measuring nitrite consumption from an initial concentration of 75 μM. Assays were performed at 25°C as described above. These experiments used nasB⁺ Δ(narKG) strains. Again, initial uptake rates were constant for approximately 10 min (Fig. 6). Strikingly, the rates of nitrite uptake (transport and reduction to ammonium in a nasB⁺ strain) were approximately fourfold greater than the rates of nitrate uptake (transport and reduction to nitrite in a ΔnasB strain). We do not know if this difference in rates reflects slower transport of nitrate relative to nitrite, slower reduction of nitrate relative to nitrite, or a combination of the two. We have been unable to measure reliably nitrite reductase activity in permeabilized cells.

FIG. 6.

Effect of pH on nitrite uptake. Note the different time scales. The nas⁺ (VJSK2216; filled symbols) and ΔnasE (VJSK2208; open symbols) strains were cultured in defined glucose medium buffered with MOPS (pH 8.0) and supplemented with 5 mM glutamine as the nitrogen source and 5 mM nitrate to induce nasF operon expression. Nitrite uptake was determined at pH 7.0 (80 mM MES; circles), 8.0 (80 mM MOPS; triangles), and 9.0 (80 mM EPPS; squares) as described in Materials and Methods. Both strains were nasB⁺ Δ(narKG).

Nitrite uptake in the nasFED⁺ strain was not markedly influenced by the pH of the assay buffer, exhibiting initial rates of approximately 44, 41, and 34 nmol of nitrite consumed min⁻¹ mg (DW)⁻¹ at pH 7, 8, and 9, respectively (Fig. 6). By contrast, uptake in the ΔnasE mutant was strongly dependent on pH, exhibiting initial rates of approximately 16, 12, and 7 nmol of nitrite consumed min⁻¹ mg (DW)⁻¹ at pH 7, 8, and 9, respectively (Fig. 6; note different time scale). This observation is congruent with results of growth tests (see above), in which nitrite-dependent growth of the Δnas mutants was strongly affected by culture medium pH.

DISCUSSION

Bacterial assimilatory nitrate and nitrite reductases are soluble cytoplasmic proteins (reviewed in reference 19). Thus, nitrate and nitrite transport is the obligatory first step in assimilation. Previous work revealed high-affinity active transport of nitrate by K. oxytoca M5al (46), but the physiology and genetics of nitrate and nitrite transport had not been further studied. Genetic analysis and sequence comparisons in our laboratory indicated that the nasFED genes appear to encode components of a typical ABC transporter (see below and reference 17). To further characterize the roles of the NasFED proteins, we constructed nonpolar nasFED deletion mutants and studied their growth and uptake phenotypes with respect to nitrate and nitrite. We conclude that the NasFED proteins are essential for high-affinity nitrate transport (Fig. 1) and participate in nitrite transport as well. (Conclusions regarding nitrite transport are complicated by the fact that high-affinity nitrite uptake at neutral pH was partially NasFED independent [see below].) Our results are largely in accord with studies of Synechococcus sp. strain PCC7942, which expresses a similar nitrate uptake system, NrtABCD (reviewed in reference 29).

Nitrate uptake.

We studied nitrate uptake by measuring the accumulation of nitrite, which requires both nitrate transport and reduction. Activity measurements in permeabilized cells indicated that nitrate reduction was not a limiting factor in our uptake assays. Strains for these experiments carried a deletion of narK and narG, encoding components involved in respiratory (anaerobic) nitrate transport and reduction, and a nonpolar deletion of nasB, encoding assimilatory nitrite reductase. In this strain background, nonpolar deletions of nasE (encoding the presumed membrane-spanning component) and nasD (encoding the presumed ATP-binding component [see below]) blocked assimilation even at 100 mM nitrate. This observation apparently differs from results with a Synechococcus nrtABCD deletion strain, which reportedly grew in medium with 60 mM nitrate (data not shown in reference 21). The ΔnasE and ΔnasD alleles were judged to be nonpolar on downstream nasB gene expression by three criteria: expression of a Φ(nasB-lacZ) fusion, growth on nitrite, and measurement of assimilatory nitrate reductase activity. Thus, K. oxytoca nitrate uptake appears to be absolutely dependent on the NasE and NasD proteins.

By contrast, the strain carrying a nonpolar deletion of nasF (encoding the presumed periplasmic binding protein) exhibited a leaky phenotype with respect to assimilation and uptake at moderate to high nitrate concentrations. A nrtA insertion mutant of Synechococcus also exhibited significant growth at 20 mM nitrate (32). Although the periplasmic binding protein is considered to be essential for the function of most ABC-type transporters (reviewed in reference 6), it has been possible to isolate binding protein-independent mutations in malF and malG, encoding the cytoplasmic membrane components of the maltose transporter (47). Thus, results with nrtA and nasF mutants may suggest that the NrtB and NasE proteins, unlike similar components (such as MalFG) in most ABC-type transporters, can interact with substrate (nitrate) to a significant degree in the absence of the binding protein. Alternatively, some other periplasmic binding protein(s) might partially substitute for NasF and NrtA.

Nitrite uptake.

The involvement of Synechococcus NrtABCD proteins in nitrite uptake has been more difficult to assess, because of the membrane permeability of nitrous acid (HNO₂; pK_a′ 3.35), which can bypass the need for active transport at neutral pH (reviewed in references 19 and 29). Nitrite is reportedly a competitive inhibitor for nitrate uptake in Synechococcus (37), and the NrtA protein binds both nitrate and nitrite with nearly equal affinities (21). Indeed, a Synechococcus nrtD insertion mutant failed to take up nitrite at elevated pH (20). This result was interpreted as demonstrating that nitrate transport and active nitrite transport are both mediated by the NrtABCD permease.

By contrast, recent results with Synechococcus nrt deletion mutants have revealed significant residual nitrite uptake, about 30% of the wild-type activity, even at pH 9.6 (21). This observation suggests that Synechococcus may also express a nitrite-specific transport system. Our results with K. oxytoca nasFED deletion mutants support this view. At pH 8.8, these mutants both took up and assimilated nitrite at about 20% of the wild-type rate (Fig. 3 and 6).

We previously observed that the nasFED genes (along with nasB) are required to bestow nitrite assimilation upon E. coli (17). A conservative interpretation of this observation is that K. oxytoca has a separate nitrite transport system that is not present in E. coli. Results reported in this paper are consistent with that conclusion.

Characteristics of NasFED and related proteins.

ABC transporters utilize the energy from ATP hydrolysis to catalyze the uptake of a wide variety of compounds (reviewed in reference 6). Cyanobacterial nitrate transporters evidently contain four distinct polypeptide components (reviewed in references 19 and 29): a monomeric periplasmic binding protein (NrtA), a homodimeric membrane-spanning protein (NrtB), and a heterodimeric ATP-binding protein (NrtC and NrtD). The homologous CmpABCD proteins, involved in carotenoid binding, show considerable sequence similarity (28, 34).

(i) Periplasmic binding proteins.

In cyanobacteria and gram-positive bacteria, periplasmic proteins are anchored to the external face of the cytoplasmic membrane through a covalently attached lipid moiety (see reference 21). Recent biochemical analysis has established the Synechococcus NrtA protein as the periplasmic nitrate- and nitrite-binding protein (21).

The deduced NasF protein of K. oxytoca is homologous to the deduced NrtA and CmpA proteins from Synechococcus sp. strain PCC7924 (37 and 30% identical residues, respectively [Fig. 7B]) and other cyanobacteria (alignments not shown). Our previously reported amino-terminal sequence for nasF contained a double frameshift; the corrected sequence is shown here. Like NrtA, NasF contains an apparent leader (signal) peptide. Indeed, the signal peptide sequences of all available NrtA-like sequences contain the double-arginine motif (consensus S/T-R-R-X-F-L-K [Fig. 7A]) characteristic of periplasmic redox enzymes that are likely assembled in the cytoplasm prior to export (reviewed in reference 4). This suggests that NrtA-like proteins may likewise be folded prior to export. However, sequence inspection has not revealed motifs indicative of redox cofactor binding sites.

FIG. 7.

Sequence of the deduced K. oxytoca NasF (KoNasF) protein and alignment with representative related proteins. (A) Amino-terminal leader regions of NrtA-like proteins. The double-arginine consensus is indicated below the sequences (4). The amino-terminal Cys residue of mature (processed) NrtA is indicated by a filled circle (21). The leader peptide cleavage site for the KoNasF protein is unknown. The Ser-Thr-Gly-Ala-rich linker region is indicated with an overline. Amino acid residues are indicated in the standard single-letter code; identical residues are boxed. SuCmpA and SuNrtA are the periplasmic binding proteins for nitrate and carotenoids, respectively, in Synechococcus sp. strain PCC7924 (27, 31, 34). PlNrtA is from Phormidium laminosum (23), and SiNrtA (sll1450) and SiCmpA (slr0040) are from Synechocystis sp. strain PCC6803 (14). AnNrtA is from Anabaena sp. strain PCC7120 (10); a double-frameshift of the deposited DNA sequence, influencing amino acid residues 18 to 46, has been introduced for this alignment. (B) Alignment of representative sequences. The remaining full-length sequences of SuCmpA, SuNrtA, and KoNasF are aligned for comparison; see panel A for the amino-terminal leader and linker sequences. These sequences were also aligned with nine other similar sequences (not shown). Filled circles denote residues that are identical in all 12 sequences aligned, whereas open circles denote residues that are identical in the 8 aligned sequences excluding the CmpA and CmpC sequences from each of the two unicellular cyanobacteria. Aligned sequences include those listed above plus SuNrtC (30), SuCmpC (28), SiNrtC (sll1452) and SiCmpC (slr0043 [14]). Azotobacter vinelandii NasS (13) was adjusted with a double-frameshift of the deposited DNA sequence, influencing amino acid residues 191 to 200, for this alignment.

Deduced NrtA sequences from a variety of cyanobacteria have been reported (reviewed in reference 19). Each sequence includes an apparent amino-terminal signal sequence, and each includes a motif resembling LXGC, required for signal cleavage (Fig. 7A). The resulting amino-terminal Cys residue serves as the site for covalent lipid attachment (21). Although the NasF leader sequence shares many features with the NrtA sequences, the conserved Cys residue is not present in NasF, as expected for a proteobacterial periplasmic binding protein.

The cyanobacterial NrtA and CmpA sequences also share an approximately 20-residue-long region, immediately adjacent to the conserved amino-terminal Cys residue, that contains a preponderance of Ser, Thr, Gly, and Ala residues (Fig. 7A and alignments not shown). This region may serve as a flexible linker between the membrane-buried lipid moiety and the nitrate-binding domain.

(ii) Membrane-spanning proteins.

The NasE protein presumably comprises a membrane-spanning homodimer. The deduced NasE protein of K. oxytoca is homologous to the deduced NrtB and CmpB proteins from Synechococcus sp. strain PCC7924 (42 and 44% identical residues, respectively [Fig. 8A]) and other cyanobacteria (alignments not shown). The E. coli MalG protein is a well-characterized transmembrane protein component of an ABC transporter required for maltose uptake (reviewed in reference 6). Membrane topology analysis indicates that MalG, like many (but not all) such proteins, contains six transmembrane helices (8). Although CmpB, NrtB, and NasE have very few residues in common with MalG, alignment of these sequences suggests that all four proteins have similar overall membrane topologies, with similarly deployed transmembrane helices (Fig. 8A).

FIG. 8.

Sequence of the deduced K. oxytoca NasE (KoNasE) protein and alignment with representative related proteins. (A) Alignment of representative sequences. The six transmembrane segments (TM-1 through TM-6) of E. coli MalG (EcMalG) are indicated with overlines (8); “p” and “c” denote the periplasmic and cytoplasmic ends of each segment, respectively. The EAA loop region is indicated with a thick underline. Amino acid residues are indicated in the standard single-letter code; identical residues are boxed. SuNrtB and SuCmpB are the integral membrane proteins involved in nitrate and carotenoid transport, respectively, in Synechococcus sp. strain PCC7924 (28, 30). (B) EAA loop regions of NrtB-like sequences. The EAA loop general consensus and the EAA loop subconsensus for ion transporters (41) are both indicated below the sequence, as is the presumed helix-loop-helix structure. Aligned sequences include the four listed above plus PlNrtB from Phormidium laminosum (23) and SiNrtB (sll1451) and SiCmpB (slr0041) from Synechocystis sp. strain PCC6803 (14).

Recently, membrane-intrinsic components of ABC transporters have been characterized as containing a conserved sequence motif termed the EAA loop (41). This sequence, thought to adopt a helix-loop-helix structure, is hypothesized to be involved in mediating interaction with the ATP-binding component of the transporter (25). Sequence analysis of proteins with different substrate specificities revealed variations on the overall EAA loop motif. However, all such sequences studied are characterized by an invariant Gly residue located four residues downstream of the Glu-Ala-Ala (EAA) triad (41).

The NasE, NrtB, and CmpB sequences do not contain this invariant Gly residue, although a region showing similarity to the ion transporter subclass of EAA loop motifs is readily apparent (Fig. 8B). Indeed, the position of this Gly residue is occupied by large basic residues (Arg or Lys) in all sequences except NasE, which has Gln. The Glu-Ala-Ala triad in these sequences is also replaced with an Asn-Val-Ala triad (Fig. 8B). These variations on the EAA loop motif therefore constitute a characteristic feature of NrtB-CmpB-NasE proteins. The structural and functional consequences of these differences are currently unknown.

(iii) ATP-binding proteins.

The defining characteristic of the ABC transporter superfamily is the highly conserved ABC motif. The deduced NasD protein, which contains the ABC motif, shares 47% identical residues with the Synechococcus NrtD protein (17). The NasD and NrtD sequences contain all the characteristic sequence motifs for this protein family, including both of the Walker-type ATP-binding motifs (17, 30).