The nac (Nitrogen Assimilation Control) Gene from Escherichia coli

Abstract

The nitrogen assimilation control gene, nac, was detected in Escherichia coli but not in Salmonella typhimurium by Southern blotting, using a probe from the Klebsiella aerogenes nac (nac_K) gene. The E. coli nac gene (nac_E) was isolated from a cosmid clone by complementation of a nac mutation in K. aerogenes. nac_E was fully functional in this complementation assay. DNA sequence analysis showed considerable divergence between nac_E and nac_K, with a predicted amino acid sequence identity of only 79% and most of the divergence in the C-terminal half of the protein sequence. The total predicted size of NAC_E is 305 amino acids, the same as for NAC_K. A null mutation, nac-28, was generated by reverse genetics. Mutants bearing nac-28 have a variety of phenotypes related to nitrogen metabolism, including slower growth on cytosine, faster growth on arginine, and suppression of the failure of an Ntr-constitutive mutant to grow with serine as sole nitrogen source. In addition to a loss of nitrogen regulation of histidase formation, nac-28 mutants also showed a loss of a weak repression of glutamate dehydrogenase formation. This repression was unexpected because it is balanced by a NAC-independent activation of glutamate dehydrogenase formation during nitrogen-limited growth. Attempts to purify NAC_E by using methods established for NAC_K failed, and NAC_E appears to be degraded with a half-life at 30°C as short as 15 min during inhibition of protein synthesis.

The nitrogen assimilation control protein (NAC) of Klebsiella aerogenes plays an important role in regulating the nitrogen metabolism of this enteric bacterium (4, 25, 40). NAC allows the coupling of operons transcribed by RNA polymerase carrying ς⁷⁰ to the nitrogen regulatory (Ntr) system (26), which uses RNA polymerase carrying ς⁵⁴. In brief, nitrogen-limited growth leads to a starvation for glutamine. Through a complex cascade of events, glutamine starvation leads to phosphorylation (and activation) of the transcriptional regulator NtrC. Phosphorylated NtrC activates RNA polymerase carrying ς⁵⁴ to transcribe a number of genes, one of which is nac, which codes for NAC. NAC in turn activates RNA polymerase carrying ς⁷⁰ to transcribe a number of operons whose products can supply the cell with ammonium or glutamate from alternative organic sources (4). NAC also represses operons whose function is to assimilate ammonium when ammonium is present in abundance (4). The operons activated by NAC in K. aerogenes include hutUH, putP, and ureDABCEFG, which code for enzymes required for the catabolism of histidine, proline, and urea, respectively. The operons repressed by NAC include gdhA (glutamate dehydrogenase [GDH]), gltBD (glutamate synthase), and nac itself (4, 16, 25).

Using the hutUH operon as a model system, we have learned that NAC is both necessary and sufficient to activate transcription from a ς⁷⁰-dependent promoter, that no coeffector is needed for NAC’s activity, and that a NAC-binding site placed at −64 relative to the start of transcription can activate RNA polymerase even at the lacZ promoter (19, 36, 40). NAC is a member of the LysR family of transcriptional regulators (41), which includes over 50 members. NAC is a typical member of this family, with two differences: NAC does not require a coeffector for its actions, and the nac gene is not divergently transcribed from an operon that it regulates.

Although NAC plays an important role in the nitrogen regulation of K. aerogenes, Salmonella typhimurium appears to lack a functional nac gene. The hutUH operon of S. typhimurium does not respond to nitrogen starvation unless it is moved to a K. aerogenes cytoplasm (34) or unless a nac⁺ plasmid from K. aerogenes is present in S. typhimurium (6).

The existence of a nac gene in Escherichia coli K-12 has remained an open question. Many of the NAC-regulated operons of K. aerogenes either are absent from E. coli (hut and ure) or do not respond to nitrogen regulation (put and gdh) (45). When the hutUH operon from K. aerogenes (or S. typhimurium) was transferred to an E. coli cytoplasm, hutUH expression was nitrogen regulated but the degree of that regulation was much less than is seen in K. aerogenes (18). In other words, the fact that hutUH shows some nitrogen regulation in E. coli suggests the presence of an active nac gene, but the weakness of that regulation and the lack of NAC-regulated targets is suggestive of the absence of an active nac gene. The data presented here show that S. typhimurium does indeed lack a nac gene and that E. coli has a nac gene that is a fully functional analog of the nac gene from K. aerogenes.

MATERIALS AND METHODS

Strains and plasmids.

All K. aerogenes and E. coli strains used in this study are derived from W70 and K-12 respectively, and are listed in Table 1.

TABLE 1.

Strains and plasmids used

Strain	Relevant characteristics	Source or reference
E. coli
DPB271	F− λ−recD1903::Tn10dTet	Stan Maloy
EB3048	glnL302 hutC515Ka nac-10	P1 · EB3044 × YMC15
EB3044	recD1903::Tn10dTet nac-10	This study
EB3046c	hutC515K nac-10	P1 · EB3044 × YMC10
EB3095	recD1903::Tn10dTet nac-28	This study
EB3099	glnL302 hutC515K nac-28	P1 · EB3095 × YMC15
EB3364	nac-28	P1 · EB3095 × W3110
EB3365	hutC515K nac-28	P1 · EB3099 × YMC10
EB4457	YMC10/pCB1056	Transformation
EB4458	YMC10/pCB1057	Transformation
EB3846	W3110/pPC36	21
EB4459	EB3365/pCB1056	Transformation
EB4460	EB3365/pCB1057	Transformation
W3110	Wild type	R. Matthews
YMC10	hutC515K	2
YMC15	glnL302b hutC515K	9
K. aerogenes
KC 2914	KC2330/pTROY11	Transformation
KC2330	hutC515 galK6	This laboratory
KC2668	hutC515 Δ[bla]-2	This laboratory
KC2725	hutC515 nac203::Tn5-131d Δ[bla]-2	This laboratory
KC2937	KC2914 (λpgal8)	Lysogen
KC2939	KC2941 (λ347 Kmr)e	Dilysogen
KC2941	KC2937 nac-203::Tn5-131	Transduction
KC2972	KC2941/pCB511	Transformation
KC3367	KC2941 (λ345 Kmr)e	Dilysogen
Plasmids
pBGS9	High-copy-number vector, Kmr	43
pCB511	E. coli nac on a BamHI-to-HindIII fragment subcloned in pGB2	This study (see Fig. 2)
pCB524	E. coli nac on a BamHI-to-HindIII fragment subcloned in pBGS9	This study (see Fig. 2)
pCB529	Kmr cartridge from pWW97 cloned into BglII site of pCB511	This study (see Fig. 2)
pCB547	pCB529 SacI-to-PvuII deletion	This study (see Fig. 2)
pCB552	nac coding sequence in pJLA503	This study
pCB554	nac-lacZ fusion in pJLA503	This study
pCB577	5.1-kb BamHI subclone carrying E. coli nac	This study
pCB1056	E. coli nac promoter driving lacZ in pRJ800	This study
pCB1057	K. aerogenes nac promoter driving lacZ in pRJ800	This study
pCJ5	2-kb BamHI subclone carrying K. aerogenes nac	6
pGB2	Low-copy-number vector	11
pJLA503	λpL expression vector	38
pRJ800	Promoterless lacZ reporter	Reid Johnson (33)
pRS415	Promoterless lacZYA reporter	42
pTROY11	Plasmid carrying λ receptor	Robert Ludwig (12)
pWW97	Kmr	31

^aThe subscript “K” indicates that the hut operons are from K. aerogenes. The hutC515 mutation inactivates the hut-specific repressor. 

^bMutation that makes the strains constitutive for the Ntr system. 

^cP1vir-mediated generalized transduction of strain EB1926 with phage grown on strain EB3044. 

^dTn5-131 is a derivative of Tn5 with the transposase and Km^r genes replaced by a tetracycline resistance gene (13). Thus, Tn5-131 is Tnp⁻, Km^s, and Tc^r. 

^eλ347 and λ345 are members of the ordered genomic cosmid library of Kohara et al. (22). Km^r indicates that a Km^r gene has been added to the phage by the method of Henry and Cronan (20). 

Media and chemicals.

Strains were grown in W4 salts (W salts adjusted to an initial pH of 7.4 [25]) supplemented with carbon and nitrogen sources at 0.4 and 0.2% (wt/vol), respectively, or in rich LB medium (29). TB medium (44) was used for plasmid isolation. Media for plasmid-bearing strains were supplemented with ampicillin (100 μg/ml), kanamycin sulfate (50 μg/ml), or tetracycline (25 μg/ml) as indicated. Glutamine was always Calbiochem A grade. IPTG (isopropyl-β-d-thiogalactopyranoside) was from Sigma Chemical Company. Sequencing reagents (Sequenase) were from United States Biochemicals, and [α-³²P]dATP at 3,000 Ci/mmol was from ICN Pharmaceuticals.

Genetic techniques.

Recombinant DNA techniques were carried out essentially as described by Maniatis et al. (27). DNA was often purified by separating digested DNA fragments in agarose buffered with TAE (27) buffer and then recovering fragments by use of glass resin or by electroelution and precipitation with ethanol. Plasmid constructs were often passed through E. coli DH5α as a host, and in all cases, only secondary transformants were studied. Strains were made competent for transformation by treatment with calcium chloride (30). Transduction using phage P1vir was performed as described previously (17).

DNA sequence analysis.

DNA nucleotide sequence was determined by the chain-terminating method, using modified T7 DNA polymerase (Sequenase) on double-stranded templates, with the following modifications to the protocol supplied by the manufacturer. Strand denaturation by NaOH was replaced by heat denaturation. The reactions mixture, containing 1 μg of plasmid DNA, 50 pmol of primer (usually 17- to 20-bp oligomer), and water to a final volume of 10 μl, was heated to 94°C for 3 min, followed by quick cooling on ice. Reaction buffer was added after denaturation, and subsequent steps followed the Sequenase protocol. Overlapping sequences were determined for both strands with pCB511 and pCB524 as templates. The sequencing strategy involved primer walking, in which new primers were synthesized by using sequence determined by the results of the previous sequence determination.

Enzyme assays.

All assays were performed with cells that had been washed in 1% KCl and suspended at a concentration that contained 1 to 1.5 mg of protein per ml. Most assays were performed with cells permeabilized by hexadecyltrimethylammonium bromide. Toluene was used to make cells permeable for proline oxidase assays. Assays for histidase (37), GDH (7), glutamate synthase (28), urease (25), and proline oxidase (37) have been described. Histidase, glutamate synthase, and GDH activities were measured at 37°C for historical reasons; all other enzymatic activities were measured at 30°C. Specific activities are reported as nanomoles of product formed or substrate consumed per minute per milligram of total cell protein. Cell protein was measured by the method of Lowry et al. (24) with bovine serum albumin as the standard, except that whole cells suspended in 1% KCl were added directly to the test mixture without prior disruption.

Complementation of K. aerogenes nac with genomic clones from E. coli.

Derivatives of several of the members of the Kohara E. coli genomic phage library (22) (miniset 345-348, generously provided by F. Neidhardt) that carried a kanamycin resistance (Km^r) gene were isolated as described by Henry and Cronan (20). To insert these derivatives, a K. aerogenes host strain required a preexisting λ prophage to provide sufficient homology for recombination. Since K. aerogenes lacks the λ receptor and an attλ site, plasmid pTROY11 was used to transform several stains to λ sensitivity as described previously (12). To provide a selectable marker for lysogeny, phage λpgal8 (14) (provided by K. McKenny) was used to transduce strain KC2330 to Gal⁺. The details of this lysogeny will be described elsewhere (35). The resulting λpgal8-containing strains were used as recipients for transduction by derivatized (Km^r) Kohara phage λ clones.

Southern hybridizations.

Chromosomal DNA from E. coli, S. typhimurium, and K. aerogenes was prepared from log-phase cultures by using a Puregene DNA isolation kit (Gentra Systems Inc.). Phage DNA from the Kohara miniset members was prepared as described by Chisholm (10). Restriction digests were performed with enzymes supplied by Boehringer Mannheim. DNA fragments were separated by electrophoresis on an 0.8% agarose gel buffered with Tris-borate-EDTA (TBE) and blotted as described by Maniatis et al. (27). Hybridization with probe prepared by random-primed labeling of the 920-bp AflII fragment from pCJ5 (6) containing the K. aerogenes nac gene was performed as described previously (27), varying the hybridization and wash temperatures between 50 and 65°C.

Mutagenesis of E. coli nac.

Using the cloned nac gene on plasmid pCB511, we introduced various disruptions into the coding sequence. Initially, a Km^r gene cartridge from plasmid pWW97 (31) was introduced as a BamHI fragment into the compatible BglII site internal to nac (see Fig. 2). An amber stop codon linker (SpeI; New England Biolabs) was introduced into the SmaI site present in the cartridge’s multiple cloning site (just 3′ to the now destroyed BglII site) to create plasmid pCB529. This plasmid produces a truncated NAC of 169 amino acids, 165 amino acids from NAC and 4 amino acids provided by the multiple cloning site from the BglII site to the stop codon. This construct was used to make the nac-10 allele. A further and more complete disruption was made by deleting the region between the SacI and PvuII sites of this plasmid to delete the promoter and 5′ end (N terminus) of the nac gene. This construct, on plasmid pCB547, was nac-28. Replacement of the chromosomal nac gene with the disrupted gene was achieved following electroporation of linearized plasmid DNA into a recD strain of E. coli (32). These Km^r-tagged nac alleles were immediately transferred into stable host backgrounds by P1vir-mediated transduction of various E. coli host strains.

FIG. 2.

E. coli nac region and clones.

Mobility shift retardation assays.

Mobility shift assays were performed with purified K. aerogenes NAC protein. Pure NAC protein was isolated as described by Goss and Bender (19). DNA targets were prepared by digestion and precipitation followed by resuspension in Tris-EDTA (TE). Radioactively labeled targets were prepared by filling in the overhanging 5′ ends with [α-³²P]dATP (ICN Pharmaceuticals) and the Klenow fragment of DNA polymerase I. Reaction mixtures contained 1 μl of DNA, 1 μl of poly(dI-dC) (50 ng/μl), 4 μl of double-distilled H₂O, and 1 μl NAC dilution in dilution buffer (50 mM NaPO₄ [pH 7], 125 mM NaCl, 0.5 mM MgCl₂, 0.1 mM β-mercaptoethanol, 50% glycerol, 1 mg of bovine serum albumin/μl). Reaction mixtures contained dilutions of NAC with 1 to 200 ng of protein/μl and were allowed to incubate with the DNA for 30 min before 1.5 μl of loading buffer (40 mM Tris [pH 8.4], 4 mM EDTA, 0.2% bromothymol blue, 0.2% xylene cyanol, 15% Ficoll) was added and the mixture was loaded for electrophoresis onto 4% acrylamide gels buffered with TE or 2% agarose gels with TBE.

Primer extension assays.

Primer extension analysis was carried out as described by Ausubel et al. (1), with the following modifications: RNA was purified from log-phase cultures grown in rich broth (LB) or in W4 salts medium supplemented with 0.4% glucose and 0.2% l-arginine as the sole nitrogen source (ammonia limiting) or 0.2% ammonium sulfate plus 0.2% glutamine (ammonia excess). A template of 25 μg of total RNA was used in each reaction. A modified reverse transcriptase (Superscript II) was used in the extension reaction at 52°C. Reactions were loaded alongside DNA sequencing reactions on a 0.8% sequencing gel. The same primer was used for primer extension and for sequencing.

Expression of the E. coli NAC protein.

Initial expression of the E. coli NAC protein was carried out by putting the nac gene under the control of a temperature-inducible promoter in the expression vector pJLA503 (38). This construct, pCB552, was made by PCR amplification of nac from pCB511. Plasmid pCB554 was made by cloning the β-galactosidase gene from pRS415 into the EcoRI and SalI sites of pCB552 such that lacZ would be transcriptionally fused to nac and thus coexpressed when the promoter was induced by raising the temperature above 35°C.

SDS-PAGE analysis of expressed protein.

Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) analysis of whole-cell extracts was carried out as described by Ausubel et al. (1), using a polyacrylamide concentration of either 10, 12.5, or 15%.

To determine the half-life of overexpressed NAC protein relative to β-galactosidase, E. coli DH5α bearing plasmid pCB554 was induced by shifting growth from 30 to 45°C for 40 min when the culture reached 50 Klett units (green filter). Cells were allowed to grow at this temperature for 1 h, at which time chloramphenicol and rifampin were added to 100 and 50 μg/ml, respectively, along with reduction of the growth temperature to 30°C. Samples (0.2 ml) were taken every 5 min from the culture flask and placed immediately on ice. Cells were collected by centrifugation at 4°C and resuspended in 20 μl of TE; 10 μl of each sample was loaded on an SDS-gel for PAGE.

Nucleotide sequence accession number.

The DNA sequence has been deposited in GenBank with accession no. U56736.

RESULTS

E. coli carries a functional nac gene.

Interspecies complementation has suggested that E. coli expresses a functional NAC but S. typhimurium does not (6, 18). To test whether E. coli and S. typhimurium possess a homolog of the K. aerogenes nac gene, we looked for hybridization between the K. aerogenes nac gene and DNA from these organisms. DNA from K. aerogenes W70 (strain KC1043), E. coli K-12 (strain W3110), and S. typhimurium 15-59 (strain NE7) was digested with EcoRI or with BamHI. After separation of the resulting fragments by agarose gel electrophoresis, a hybridization analysis by the method of Southern was carried out with an AflII fragment from within the K. aerogenes nac gene as probe. Under stringent conditions (hybridization and washes at 65°C), this 920-bp probe hybridized to a single fragment of K. aerogenes DNA but did not hybridize to any DNA fragment from E. coli or S. typhimurium. When the stringency of the hybridization and washes was reduced (58°C), we detected a single fragment from E. coli DNA but no hybridization with S. typhimurium DNA (Fig. 1). The positions of the bands suggested that the EcoRI and BamHI fragments from E. coli were about 11.6 and 5.1 kb in size. When the stringency of the hybridization and washes was reduced still further (50°C), hybridization to S. typhimurium DNA was detected but there were many fragments showing hybridization in the lanes containing DNA from each of the three organisms under these conditions (not shown). Thus, E. coli appears to have a locus with strong sequence similarity to nac from K. aerogenes; S. typhimurium appears to lack such a sequence.

FIG. 1.

Southern blot of digested chromosomal DNA probed with a 920-bp AflII fragment (carrying the K. aerogenes nac gene). Lanes 1 to 3 and 4 to 6, K. aerogenes, E. coli, and S. typhimurium, respectively, digested with EcoRI (lanes 1 to 3) and BamHI (lanes 4 to 6); lanes 7 and 8, K. aerogenes and E. coli plasmid clones (pCJ5 and pCB577, respectively) digested with BamHI.

The genetic maps of E. coli and K. aerogenes are quite similar in all regions that have been compared (5), so it seemed likely that the nac gene of E. coli would lie in a region analogous to that of the nac gene of K. aerogenes. Several λ clones from the Kohara ordered set of genomic clones (22) were tested for complementation of a nac mutant of K. aerogenes. Four clones (345 through 348) believed to cover the E. coli chromosomal region analogous to that containing the nac gene in K. aerogenes were chosen and modified to encode kanamycin resistance as described by Henry and Cronan (20). These derivatized phage were then used to transduce a nac-deficient λ lysogen of K. aerogenes (strain KC2941) to kanamycin resistance. Transductants were assayed for histidase and GDH activities. Only the transductants derived from clones 346 and 347 showed NAC-dependent regulation of histidase and glutamate dehydrogenase formation. As shown in Table 2, the strain carrying the prophage derived from clone 347 showed the same activation of histidase, urease, and proline oxidase in response to nitrogen starvation as the wild type. That strain also showed the same NAC-dependent repression of GDH and glutamate synthase as the wild type (compare KC2939 with KC2668). In contrast, a strain carrying the prophage derived from clone 345 was as defective as the nac mutant in activation and repression of enzyme formation in response to nitrogen starvation (compare KC3367 with KC2941). Thus, the nac gene from E. coli seems fully able to complement a K. aerogenes nac mutant.

TABLE 2.

Regulation of enzyme formation in a K. aerogenes nac mutant carrying the nac gene from E. coli

Strain	Relevant genotype	Growth mediumb	Sp activity (U/mg of protein)a
			Histidase	GDH	Urease	Proline oxidase	Glutamate synthase
KC2668	Wild type	+N	33	361	11	9	193
		−N	357	52	1,380	47	70
KC2941	nac-203::Tn5-131	+N	33	379	10	12	260
		−N	51	401	242	17	245
KC2939	nac-203(λ347 Kmr)c	+N	39	369	15	11	216
		−N	319	94	1,212	32	125
KC3367	nac-203(λ345 Kmr)	+N	29	423	21	11	NDe
		−N	54	411	207	13	ND
KC2972	nac-203/pCB511d	+N	31	292	13	8	232
		−N	460	41	1,275	39	98

^aAverage of three independent cultures. Units of urease activity are calculated based on urea cleaved. 

^bNitrogen-limiting medium (−N) is W4 salts supplemented with glucose (0.4%) and l-glutamine (0.2%) as sole nitrogen sources. Nitrogen-excess medium (+N) is further supplemented with ammonium sulfate (0.2%). Cultures were grown at 30°C with vigorous shaking to mid-log phase (50 Klett units). Media for cells used in urease assays were supplemented with 1 μM nickel sulfate. Media for cells used in proline oxidase assays were supplemented with 0.01% l-proline. 

^cλ347 carries the E. coli nac gene; λ345 does not. 

^dPlasmid pCB511 carries the intact E. coli nac gene. 

^eND, not done. 

The region of overlap between the two complementing clones (346 and 347) includes an EcoRI fragment of about 12 kb which in turn contains a BamHI fragment of about 5.4 kb, in close agreement with the sizes predicted from the Southern blot in Fig. 1.

Isolation and characterization of the E. coli nac gene.

The E. coli nac gene was isolated as a 5.4-kb BamHI fragment from the Kohara miniset clone 347 in both a low-copy-number vector (pGB2) and a high-copy-number vector (pBGS9). High-copy-number plasmids carrying nac cause the cells that carry them to grow slowly, perhaps because of the high-level expression of the adjacent asnV gene (tRNA^Asn). A restriction map of this fragment is shown in Fig. 2.

The sequence of the nac gene and those of the two adjacent genes (asnV and cbl) were determined by the dideoxy sequencing method on both strands. The sequence of the E. coli nac promoter region is very different from the corresponding region from the K. aerogenes nac gene, except in those regions known to specify the binding of important regulatory proteins, including NAC, NtrC, and RNA polymerase (Fig. 3). The deduced protein sequence of E. coli NAC differs considerably from that of K. aerogenes, in contrast to the situation with most regulatory proteins, where strong similarity (often approaching 100%) exists. The N-terminal one-third is quite similar (greater than 90% identity), but the C-terminal two-thirds differs much more than expected (only about 75% identity). Thus, the overall deduced amino acid sequence identity is about 79%.

FIG. 3.

Comparison of the promoter regions from the nac genes of E. coli (top) and K. aerogenes (bottom). |, identity; ∧, space added for better alignment; *, start of transcription as determined by primer extension.

Like the K. aerogenes nac gene (41), the E. coli nac gene is preceded by a sequence resembling a ς⁵⁴-dependent promoter. The start of transcription was determined by primer extension analysis (Fig. 4) and confirms that the putative ς⁵⁴-dependent promoter is in fact the promoter of the E. coli nac gene.

FIG. 4.

Primer extension analysis of the E. coli nac promoter. RNA was extracted from YMC10 cells grown in nitrogen-excess medium containing ammonium sulfate (0.2%) and glutamine (0.2%) as nitrogen sources (lane 1) and from YMC10 cells grown in nitrogen-limiting medium containing arginine (0.2%) as the nitrogen source (lane 2). A sequencing ladder was generated by using the nac containing plasmid pCB524 as template. The same oligonucleotide (5′ GCTGGTTGTGCGATATGCA 3′) was used as a primer for primer extension and for the sequencing reactions; thus, the sequencing ladder represents the complement of the sequence shown. The sequence at the bottom is the E. coli nac promoter, with the start of transcription indicated by the bold A.

Isolation of nac mutations.

Two nac mutations were isolated by reverse genetics, taking advantage of a BglII site at the 165th codon of the nac coding sequence and a PvuII site upstream of the nac promoter. The nac-10 mutation is an insertion of a Km^r cassette from plasmid pWW97 into the BglII site of nac. Such a mutation leads to a truncated NAC with only the N-terminal 165 amino acids of NAC plus 4 amino acids encoded in the linker used to provide a stop codon. The nac-28 mutation was derived from nac-10 by deletion of the DNA from the PvuII site upstream of the nac promoter to a SacI site just inside the Km^r cassette. Thus, nac-28 is a null allele for nac that deletes the nac promoter and the entire N-terminal half of the nac coding sequence. The nac-28 mutation was crossed onto the E. coli chromosome as described in Materials and Methods, and replacement of the wild-type nac gene was confirmed by Southern blot analysis (not shown). The resulting mutant strains were tested for regulation of the K. aerogenes hutUH operon in these E. coli strains (Table 3).

TABLE 3.

Nitrogen regulation of enzyme formation in E. coli nac mutants

Strain	Relevant genotype	Growth mediumb	Sp activity (U/mg of protein)a
			Histidase	GDH
YMC10	hutC515K	+N	98	175
		−N	320	145
EB3365	YMC10 nac-28	+N	96	118
		−N	108	257
YMC15	glnL302 hutC515K	+N	375	86
EB3099	YMC15 nac-28	+N	145	130
W3110	Wild type	+N	NAc	96
		−N	NA	84
EB3364	W3110 nac-28	+N	NA	91
		−N	NA	146

^aAverage of four independent cultures. 

^bNitrogen-limiting (−N) and nitrogen-excess (+N) indicate growth at 37°C in W4 salts supplemented with glucose (0.4%) and either l-arginine (0.2%; −N) or ammonium sulfate (0.2%; +N). 

^cNA, not applicable. 

Strains carrying nac-28 were unable to derepress histidase synthesis whether nac expression was induced physiologically (by nitrogen-limited growth of strain EB3365) or genetically (by the glnL302 mutation in strain EB3099, which makes the Ntr system constitutively active).

The gdhA genes (encoding GDH) from E. coli and K. aerogenes are quite similar. In particular, a sequence resembling a NAC-binding site is found at about −80 in the promoter regions of the gdhA genes from both organisms. However, expression of GDH activity in E. coli does not vary appreciably in response to nitrogen limitation, in contrast to the strong repression seen in K. aerogenes. Analysis of GDH expression in strains with the nac-28 mutation reveals a subtle regulatory complexity. The nac-28 mutants show about a twofold increase in GDH expression when the Ntr system is activated either by nitrogen limitation or by the glnL302 mutation (strains EB3365 and EB3099 in Table 3). In contrast, the nac⁺ strain YMC10 showed a slight but reproducible repression of GDH expression. Thus, there appears to be an approximately twofold repression of GDH expression in E. coli by NAC, and this is almost exactly balanced by a nearly twofold increase in GDH expression during nitrogen-limited growth by an unknown mechanism. There is considerable variation in the expression of GDH among different E. coli strain backgrounds; therefore, we compared a nac-28 mutant of strain W3110 with its nac⁺ parent and found essentially the same effect as seen with YMC10 (Table 3). Although the regulatory effects of NAC on GDH expression are small, they are physiologically significant. gltD mutants lack glutamate synthase activity, and their assimilation of ammonium is completely dependent on GDH. Strain EB3135 (gltD nac-28) grows faster than EB3134 (gltD) in glucose medium with ammonium as the sole nitrogen source (doubling times of 49 ± 8 and 65 ± 10 min for EB3135 and EB3134, respectively).

Effect of nac mutations on growth rate of E. coli.

In K. aerogenes, nac mutations have no phenotype except slower growth on substrates like histidine, proline, or urea that are catabolized by products of NAC-dependent operons (3). This was also seen in E. coli in that nac mutants grew significantly more slowly in glucose minimal medium with cytosine as the sole nitrogen source, with doubling times of 180 ± 5 and 235 ± 10 min for the nac⁺ and nac strains, respectively. The nac mutant also grew somewhat more slowly when either serine or threonine was the sole nitrogen source. This was readily seen as a difference in colony size, but the doubling times of nac mutants on serine or threonine were only about 5 min (about 3%) longer than the nac⁺ parent.

In contrast, when nac mutants of E. coli were grown on glucose-arginine medium (i.e., glucose minimal medium with l-arginine as the sole nitrogen source [Table 4]) the growth rate actually increased, with doubling times of 147 ± 3 and 134 ± 4 for the nac⁺ and nac strains, respectively. The increase was always small but was clearly seen in all five of the paired growth experiments attempted. This difference was particularly large in strains which carry the glnL302 allele, which leads to constitutive expression of the Ntr system and thus constitutive NAC expression. In addition to an increased growth rate on glucose-arginine medium, there is also a decrease in the growth lag when cells are transferred from nitrogen-rich medium (glucose minimal medium supplemented with 0.2% ammonium sulfate) to glucose-arginine. The growth rate advantage of the nac mutants was minimized when even small amounts of glutamate (0.02%) or aspartate (0.01%) were added to the glucose-arginine medium (Table 4). Another phenotype involved the ability of a glnL302 strain to grow in glucose minimal medium with serine as the sole nitrogen source. Strains carrying the glnL302 mutation failed to grow on glucose-serine medium. The reason for this failure is not known. Nonetheless, nac mutations allow the glnL strain to grow in glucose with serine as the sole nitrogen source (as do glnE mutations).

TABLE 4.

Growth rates of E. coli strains containing nac and glnL mutations^a

Medium	Strain	Relevant genotype	Lag time (min)	Doubling time (min)
Glucose (0.4%)– l-arginine (0.2%)
	YMC10	Wild type	245	147
	YMC15	glnL302	225	140
	EB3365	nac-28	196	134
	EB3099	glnL302 nac-28	152	109
Glucose (0.4%)– l-arginine (0.2%)– glutamate (0.02%)
	YMC10	Wild type	198	135
	YMC15	glnL302	159	110
	EB3365	nac-28	193	133
	EB3099	glnL302 nac-28	110	106

^aGrowth rates were measured at 37°C in W4 medium supplemented as indicated. Lag times were calculated upon shift from log-phase growth on glucose (0.4%)-ammonia (0.2%) cultures to the indicated medium. Cells were washed twice in cold 1% KCl before inoculation into new medium. Growth rate was determined by monitoring the cell density of a vigorously shaking culture in sidearm flasks. 

Effect of NAC on nac expression in E. coli.

Transcription of the K. aerogenes nac gene is strongly regulated by the Ntr system in response to the nitrogen supply (25). It is also repressed by the binding of NAC to a site centered about 77 bp upstream from the start of transcription (6, 16). When the K. aerogenes nac promoter was cloned on a high-copy-number plasmid driving β-galactosidase expression, the E. coli Ntr system was fully functional in regulating transcription from the K. aerogenes nac promoter (Table 5, lines 1 and 2 or 3 and 4). The chromosomally encoded NAC exerted a weak repression on nac-driven β-galactosidase expression from this multicopy plasmid (Table 5, lines 2 and 4). When the E. coli nac promoter replaced that from K. aerogenes, a similar regulation in response to the nitrogen supply (Table 5, lines 5 and 6 or 7 and 8), as well as a similar repression by NAC (Table 5, lines 6 and 8), was seen. The E. coli nac promoter contains a sequence resembling the NAC-binding site from the K. aerogenes nac promoter, and this site is also in a very similar position (centered at −76). A gel mobility shift assay (Fig. 5) confirms that NAC binds to the E. coli nac promoter region, consistent with the argument that the E. coli nac gene is autogenously regulated, just like that from K. aerogenes.

TABLE 5.

Regulation of nac promoter expression by E. coli nac

Strain	Relevant genotypea	Growth mediumb	β-Galactosidase sp act (U/mg)c
EB4458	Wild type/pCB1057	+N	487
		−N	13,074
EB4460	nac-28/pCB1057	+N	502
		−N	25,011
EB4457	Wild type/pCB1056	+N	449
		−N	16,901
EB4459	nac-28/pCB1056	+N	507
		−N	28,290

^aPlasmid pCB1056 carried the nac promoter from E. coli fused to the promoterless lacZ gene of pRJ800; pCB1057 carries the nac promoter from K. aerogenes fused to the promoterless lacZ gene of pRJ800. 

^bNitrogen-excess growth medium (+N) is W4 salts supplemented with glucose (0.4%), ammonium sulfate (0.2%), and l-arginine (0.2%). Nitrogen-limiting (−N) medium was supplemented with glucose (0.4%) and l-arginine (0.2%). Cultures were grown at 37°C assayed as described in Materials and Methods. 

^cAverage of five independent determinations. 

FIG. 5.

Gel mobility shift assay showing the interaction of NAC with the E. coli nac region. Lanes: 1 to 4, 330-bp EcoRI-HindIII fragment containing the K. aerogenes ureD promoter incubated with 0, 0.35, 0.7, 1.7 pmol of purified K. aerogenes NAC, respectively; 5 to 8, 378-bp EcoRI-to-BamHI fragment containing the E. coli nac promoter incubated with 0, 0.35, 0.7, and 1.7 pmol of K. aerogenes NAC, respectively.

Expression of E. coli nac coding sequence.

The coding sequence of E. coli nac was expressed from a plasmid expression vector with a temperature-inducible phage λ p_L promoter (38). SDS-PAGE of expressed NAC showed a protein of 33 kDa as expected from the sequence. Curiously, no NAC protein was seen when strains bearing this construct were grown at 42°C. To visualize protein, the strains had to be grown at 46°C. Control experiments using this vector (38) suggested that an ample amount of protein should have been made at 42°C. To reconcile this discrepancy, a transcriptional fusion was made by placing the lacZ coding sequence immediately downstream of the nac coding sequence, creating an artificial operon with nac and lacZ cotranscribed. Expressed β-galactosidase was readily detected at the lower temperatures (37 and 42°C), suggesting that both genes were being adequately transcribed. It seemed likely that either the NAC transcript was being inefficiently translated or the NAC protein was being degraded. Therefore, we compared the stabilities of the NAC and β-galactosidase polypeptides. NAC and β-galactosidase were produced by growing strains at 46°C for 1 h and then shifting the temperature to 30°C along with the addition of chloramphenicol and rifampin to prevent further translation and transcription. The disappearance of the NAC and β-galactosidase polypeptides was monitored by SDS-PAGE (Fig. 6). The NAC bands exhibited a half-life of about 15 min following overexpression and growth at 30°C, in contrast to the more stable β-galactosidase protein (half-life of >35 min).

FIG. 6.

Relative stabilities of E. coli NAC and β-galactosidase. E. coli NAC and β-galactosidase were cotranscribed from the temperature-inducible phage λp_L promoter in plasmid pCB554. Cells were grown to a density of approximately 50 Klett units, and expression was induced by growth at 45°C for 40 min. At time zero, the culture was shifted to 30°C and chloramphenicol and rifampin were added. Samples (0.2 ml) were removed at 5-min intervals and concentrated 10-fold, and 10 μl was applied to an SDS–12.5% polyacrylamide gel. Lanes: 1, 5 μg of K. aerogenes NAC as a standard; 2, sample at time zero; 3 to 9, samples at 5-min intervals; 10 and 11, E. coli NAC expressed from plasmid pCB552 (which transcribes NAC alone).

DISCUSSION

The data presented here show clearly that E. coli has a functional homolog of the K. aerogenes nac gene. DNA sequence analysis of E. coli nac (hereafter referred to as nac_E) predicts a protein with about 80% identity to the K. aerogenes NAC (hereafter referred to as NAC_K). Moreover, nac_E is fully able to complement a K. aerogenes nac mutant for activation of hut, put, and ure operon expression as well as for repression of gdhA expression and probably for autogenous repression of nac_K expression. On one hand, the existence of nac_E was not surprising given the close evolutionary relationship between E. coli and K. aerogenes. On the other hand, we have long known that the operons that NAC regulates in K. aerogenes are absent from E. coli (hut and ure) or are either unregulated or only slightly regulated by nitrogen in E. coli (gdhA and put). That raised the question of whether E. coli had a nac gene and, if so, what that gene did. We continue to find operons in K. aerogenes with a NAC-dependent regulation, most recently the dadAB operon which is involved in alanine catabolism (21). Again, dad is NAC regulated in K. aerogenes but not in E. coli. Nevertheless, it seems likely that as more NAC-regulated operons are discovered, some of them will be NAC regulated in E. coli as well.

Although NAC_E is a functional homolog of NAC_K, the degree of sequence divergence between NAC_E and NAC_K is surprising. Most regulatory proteins are highly conserved between these two bacteria. NtrC and Lrp are identical in the two strains, and catabolite gene activator protein and the α subunit of RNA polymerase differ by only one and two amino acids in these bacteria. Homologous LysR family members also tend to be highly conserved within the enteric bacteria. For example, the CysB proteins from E. coli and S. typhimurium are 95% identical in amino acid sequence (39). In fact, the entire LysR family shows surprising sequence conservation. NAC from K. aerogenes is 50% identical to OxyR (a regulator that senses oxidative stress) from E. coli (41). Thus, the >20% nonidentity between NAC_E and NAC_K is striking. It should be noted that there is strong sequence conservation in the amino-terminal one-third of the protein and that most of the sequence divergence is found in the carboxy-terminal two-thirds. The carboxy-terminal portion is generally thought to be important for interaction of LysR family proteins with their regulatory coeffectors (39). NAC_K appears to lack any coeffectors and to be a constitutively active transcriptional regulator (25). This may explain the lack of sequence conservation in this region. However, it must be noted that NAC_E and NAC_K both have exactly 305 amino acids and that their length is quite typical of all other LysR family members. Therefore, the carboxy-terminal domain(s) may play some role (such as stability) other than response to a coeffector signal. In support of this argument, we noted that the hydropathy plots of NAC_E and NAC_K are remarkably similar, despite the 25% divergence of the amino acid sequence in the carboxy-terminal region.

The physiological role for NAC_E remains unproved, though the data presented here suggest that NAC_E, like NAC_K, is involved in regulating the nitrogen metabolism of the cell. Several lines of evidence support this hypothesis. First, the nac-28 mutation affects the growth rate of E. coli on a variety of nitrogen sources, including arginine (faster), cytosine (slower), and serine (allowing growth of Ntr-constitutive strains). Second, even though the strong repression of gdhA by NAC seen in K. aerogenes does not occur in E. coli, a small but significant effect is detectable. In the nac-28 mutant, GDH expression increases about twofold in response to nitrogen limitation. In wild-type strains, this increase does not occur. In a gltB mutant, where glutamate formation from ammonia is completely dependent on GDH, a strain carrying nac-28 grows faster with ammonium as nitrogen source than does a nac⁺ strain, suggesting that this twofold effect on GDH expression is important. Third, the key regulatory sites in the nac_K promoter region (15, 16) are well conserved, both in sequence and in position, in the nac_E promoter region. These include the ς⁵⁴-dependent promoter and the NtrC-binding enhancer sequence. Thus, it is likely that nac_E, like nac_K, is transcribed as part of the Ntr response during nitrogen limitation.

The absence of any nac-like sequences in S. typhimurium is consistent with older observations that NAC-dependent operons cannot be expressed in response to nitrogen limitation (6, 8). The absence of a nac gene in S. typhimurium also raises questions about the evolutionary plasticity of this chromosomal region in the Enterobacteriaceae. The nac_K gene is preceded by asnV, the gene that codes for tRNA^Asn, and it is followed by an apparent operon composed of cblA (orf2), itself a member of the LysR family, and orf3. The function of this operon is unknown but may be related to sulfur metabolism. In E. coli, nac is again preceded by asnV and followed by cblA, but orf3 is located upstream from nac, on the other side of asnV. In S. typhimurium, nac is not present. There is evidence that the asnV region may be where the large cluster of cob genes, encoding the enzymes of cobalamine synthesis, lies in S. typhimurium (23). (E. coli lacks this cob cluster.) Similarly, the large cluster of nif genes (encoding the enzymes of nitrogen fixation) in K. pneumoniae appear to be linked to nac. (The nif genes are not present in K. aerogenes.) In short, this region has been subject to considerable insertion or deletion during the evolutionary divergence of the enteric bacteria. It is tempting to note that asnV and asnU flank nac_E and present an 86-bp direct repeat. If no essential genes lie between these (as is seen for E. coli), simple deletion of the intervening region by asnV-asnU recombination could lead to loss of nac, but such a speculation remains untested.

In summary, E. coli has a nac gene that is a functional homolog of the nac gene from K. aerogenes, and this nac gene is probably involved in the nitrogen regulation of E. coli. The strong sequence conservation within the first 100 or so amino acids suggests an important role for the domain(s) in this region, whereas the large number of differences in the last 200 or so amino acids suggests that the domain(s) in this region are quite plastic and can withstand considerable amino acid substitution without loss of function.