INTRODUCTION

Top Abstract Introduction Materials and methods Results Discussion References

The nitrogen regulatory (Ntr) system of enteric bacteria allows Klebsiella aerogenes, Escherichia coli, and other enteric bacteria to respond rapidly and effectively to changes in the quality of the nitrogen source provided (13). Under conditions of nitrogen limitation, a complex cascade of regulatory events involving uridylylation and phosphorylation reactions leads to the accumulation of the phosphorylated (active) form of the transcriptional activator NtrC. NtrC-phosphate can activate the ⁵⁴-dependent expression of a variety of genes involved in utilization of organic and inorganic nitrogen sources. NtrC-phosphate also activates transcription of the nac gene in K. aerogenes and E. coli (6).

The nitrogen assimilation control protein (NAC) is a regulatory protein responsible for activating the transcription of operons such as hutUH, putP, and ureDABCEFG, whose products supply the cell with ammonia or glutamate from histidine, proline, and urea, respectively (2). NAC is also responsible for repressing transcription of the gdhA and gltBD operons, whose products are involved in assimilating ammonia under conditions of nitrogen excess or limitation (2). NAC is also responsible for down-regulating its own transcription (3, 7). The expression of the nac gene is entirely dependent on the Ntr system such that under conditions of nitrogen-limited growth, the Ntr system is active and nac is actively transcribed by RNA polymerase charged with the minor sigma factor ⁵⁴ (6, 12). There are no known coeffectors involved in the regulation of transcription by NAC (17), so if nac is expressed, then NAC is produced and activates transcription of hut, put, and ure by RNA polymerase charged with the major sigma factor ⁷⁰. Thus, NAC serves as a coupling factor between the ⁵⁴-dependent Ntr system and ⁷⁰-dependent genes like hut (2).

NAC is a member of the large family of LysR-type transcriptional regulators (LTTRs; 15, 16). This group of over 50 proteins are identified on the basis of amino acid sequence similarity. Most of the members of this family have between 300 and 350 amino acids and show a surprisingly high degree of sequence similarity. For example, NAC from K. aerogenes shows about 40% sequence identity with OxyR from E. coli (16). Within the LTTRs, the DNA-binding domain occurs near the N terminus of the protein, where a predicted helix-turn-helix motif (approximately amino acids 20 to 40 in NAC) is located (15). In several LTTRs, the binding site for a required coeffector has been identified. These binding sites are generally in the C-terminal third of the polypeptide, suggesting that the DNA-binding and coeffector-binding domains are physically separated (15).

Very little structural information about LTTRs is available. A fragment of CysB containing the C-terminal 233 amino acids has been crystallized, and a structure was derived from X-ray diffraction data (18). However, this fragment lacks the N-terminal 87 amino acids and thus the DNA-binding domain. As a result, this structure says little about how the DNA-protein interactions occur, are altered by the presence of inducer, or result in the activation of transcription.

Comparison of NAC from E. coli (NAC_E) with NAC from K. aerogenes (NAC_K) revealed a surprising lack of sequence similarity (14). NAC_E was only about 75% identical to NAC_K, in contrast to other E. coli and K. aerogenes regulatory proteins, which are >95% identical. Most of the sequence divergence between NAC_E and NAC_K occurs in the C-terminal two-thirds of the protein, consistent with the lack of a need to conserve a coeffector-binding site in the C-terminal domain. This raised the question of whether the C-terminal domain has any role at all, leading us to seek the smallest N-terminal fragment that would retain the biological activities associated with NAC.

Two nac mutations also suggested that a C-terminal portion of NAC might be dispensable, further encouraging us to search for a small active fragment of NAC. The nac-112 allele of K. aerogenes is an insertion of Mudlac,amp at an unknown site within nac_K, which retains considerable NAC activity (12). The nac-10 allele of E. coli, an insertion of a stop codon and a drug resistance cartridge in the center of the nac_E gene, was constructed as an intermediate in constructing the nac-28 null allele (14). Strains carrying nac-10 also retained considerable NAC activity. Thus, we attempted to determine whether the C-terminal domain of NAC from E. coli and K. aerogenes plays an essential role and to define the smallest N-terminal fragment of NAC that would retain the ability to function as a transcriptional regulator.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and methods Results Discussion References

Strains and plasmids. All of the K. aerogenes strains used in this study were derived from W70 and are listed in Table 1.

Generation of C-terminal deletions of NAC. Unidirectional exonuclease III deletions were performed to make 3' deletions in the E. coli nac gene by using the protocol described by Henikoff (9) with minor modifications. Plasmid pCB574 contains the E. coli nac gene inserted into the EcoRI site of plasmid pGB2 (4) such that the 3' end of the gene is proximal to the PstI site present in the multiple cloning site. Deletions were initiated by cutting with SalI and PstI to allow unidirectional deletions into nac. After digestion with exonuclease III and mung bean nuclease, the deleted DNA fragments were resuspended in 20 µl of ligation buffer (250 mM Tris-HCl [pH 7.6], 50 mM MgCl₂, 1 mM dithiothreitol, 0.5 mM ATP) and 100 pmol of a dephosphorylated NheI linker (containing stop codons in three reading frames) to which T4 DNA ligase was added. The ligation mixture was used to transform E. coli DH5 to streptomycin (50 µg/ml) and spectinomycin (100 µg/ml) resistance. Deletions of interest were sequenced by using the pUC reverse primer to determine the precise endpoint of each deletion. These were brought under the control of the lac promoter by PCR amplification by using the primers CGAATTCAAACTGGAGACTCATATGAAC (forward) and AGGATCCTCACACAGGAAACAGCTATGAC (reverse) and insertion into the EcoRI and BamHI sites of pGD103 (5). Specific truncations in the K. aerogenes nac gene were generated by PCR using a forward primer (CTGGAATTCCTTACAGGAGGCA) containing an EcoRI site and reverse primers complementary to the last 18 bases prior to the desired truncation point followed by an amber stop codon and GGATCC (BamHI). PCR products were introduced into the EcoRI and BamHI sites of plasmid pGD103, and the sequences were verified by dideoxy sequencing.

Purification of MBP-NAC fusions. The nac genes from E. coli and K. aerogenes, along with the C-terminal truncated derivatives of each, were inserted into vector pMalC2 (New England Biolabs) as a blunt-ended 5' end of the nac gene beginning at the initiating methionine ligated to the blunt-end XmnI site within the vector. The 3' ends were ligated into either the EcoRI or BamHI sites within the multiple cloning site. The initial construct was prepared by PCR and sequenced to verify that errors had not been introduced by PCR amplification. Later derivatives were made by replacing sequences within this clone by using internal sites (XmnI, SspI, and BglII) in the nac gene rather than reamplifying for each construct. Insertion of a blunt-ended DNA fragment into the XmnI site of the vector results in fusion of the NAC protein to the 42-kDa maltose-binding protein (MBP). The junction between the two proteins contains the recognition site for factor Xa protease (Ile Glu Gly Arg) positioned such that cleavage with this enzyme would result in separation of the intact NAC protein from the fusion.

Purification of his₆-tagged NAC. To overcome the problems associated with MBP-NAC fusions, various six-histidine (his₆)-tagged constructs were built. NAC N-terminal constructs were made by cloning nac genes into the EcoRI/HindIII sites of pET28 (Novagen). The N-terminal fusions added a leader to the NAC proteins consisting of 56 amino acids which include a stretch of six histidines in a row to facilitate nickel-nitrilotriacetic acid affinity purification (Qiagen). Constructs in which carboxy-terminal his₆ tags were added had the DNA sequence (CAT)₆ followed by an amber stop codon designed into the primers used to amplify the nac gene. The C-terminal his₆-tagged genes were inserted into a derivative of the pET28 vector in which the material between the XhoI and EcoRI sites had been removed. Histidine-tagged NAC proteins were expressed as previously described for the MBP fusions and prepared for affinity purification in a similar fashion, except that the pH of the buffer was raised to 7.8 to facilitate nickel interaction. Clarified S30 French pressate was added to a column with an internal diameter of 1.25 cm containing a 10-ml bed volume of nickel-nitrilotriacetic acid resin (Qiagen). NAC was allowed to bind the column at a flow rate of 2 ml/min. The column was washed with 150 ml of column buffer supplemented with 5 mM imidazole to eliminate nonspecific binding. NAC was eluted with elution buffer (300 mM imidazole [pH 7.5], 250 NaCl, 1 mM MgCl₂, 1 mM 2-mercaptoethanol). Fractions of 3 ml were collected and screened by Bio-Rad protein assay. NAC generally eluted in a tight peak in fraction 3. Purity was monitored by SDS-PAGE.

Gel filtration analysis of purified proteins. Gel filtration chromatography was used to determine the size and oligomeric state of purified NAC proteins. Size determinations were made by one of two methods: fast protein liquid chromatography (FPLC)-mediated gel filtration or gravity-fed chromatography. For determinations by FPLC, a Pharmacia LCC-501 FPLC apparatus was used in conjunction with a Superdex 75 HR 10/30 prepacked column. The column was equilibrated with buffer 4 (described above) and loaded with 30 µl (75 µg) of NAC. Separation was allowed to continue at a flow rate of 1 ml/min for 40 min. Detection of eluting protein was done by monitoring A₂₈₀. Standardization of the column was done with gel filtration protein standards purchased from Sigma Chemical Co. resuspended in buffer 4. Generally, dimeric NAC (66-kDa dimer) eluted at a 15-ml volume at the same fraction as bovine serum albumin (66 kDa). Gravity flow gel filtration was performed by using an SR 25/100 column (25-mm internal diameter) packed with a 150-ml bed volume of Sephacryl S-100. Standardization of this column was done with the same buffers, globular weight standards, and concentrations as the FPLC, but the loading volume was increased to 1 ml. The flow rate was adjusted to 0.5 ml/min, and 150 ml was allowed to flow. Generally, dimeric NAC (66 kDa) eluted at a volume of 16 ml from this column.

In vitro transcription assays. In vitro transcriptions were carried out essentially as described previously (8). The template used was plasmid pCB695, which carries the K. aerogenes hutUH promoter inserted into vector pTE103, similar to pAM1202, which was described previously (8). Transcription mixtures contained 0.16 pmol of supercoiled plasmid as the template and 2 to 6 pmol of RNA polymerase in a total volume of 25 µl.

Enzyme assays. Histidase and glutamate dehydrogenase (GDH) activities were measured as described previously (12). Specific activities are reported as nanomoles of urocanate formed (histidase) or NADPH₂ oxidized (GDH) per minute per milligram of cell protein. Cell protein was determined by the method of Lowry et al. (11).

Gel mobility shift assays. Gel mobility shift assays were performed by using purified NAC_E or NAC_K or fragments thereof. The ability of NAC to bind DNA was determined by using the 330-bp EcoRI-to-HindIII fragment of the ureD promoter, which contains a strong NAC-binding site. Binding reaction mixtures consisted of 1 µl of DNA (0.05 pmol), 1 µl of poly[d(I·C) (50 ng/ml), 4 µl of deionized distilled H₂O, and 1 µl of a NAC dilution (0.35 to 1.7 pmol) in 50 mM NaH₂PO₄ (pH 7)-125 mM NaCl-0.5 mM MgCl₂-0.1 mM -mercaptoethanol-50% glycerol, 1-mg/ml bovine serum albumin. NAC was incubated with the DNA for 30 min before addition of 1.5 µl of loading buffer (40 mM Tris [pH 8.4], 4 mM EDTA, 0.2% bromthymol blue, 0.2% xylene cyanol, 15% Ficoll) and applied to a 2% Tris-borate-EDTA agarose gel. Electrophoresis was carried out at room temperature at 15 V/cm for approximately 1 h. Gels were stained with ethidium bromide, and the fragments were visualized by UV fluorescence.

	RESULTS

Top Abstract Introduction Materials and methods Results Discussion References

Residual NAC activity in nac-10 mutants. The nac-10 allele of the E. coli nac gene has a stop codon and a kanamycin resistance gene inserted into the middle of nac_E after codon 165. The resulting NAC polypeptide has four amino acids encoded by the polylinker fused to the 165th amino acid of NAC; the C-terminal 140 amino acids of NAC are not present. The nac gene of E. coli was replaced with nac-10 as described elsewhere (14), and the replacement was confirmed by Southern blotting (data not shown). To test for NAC activity, we measured the ability of a strain carrying nac-10 to activate the NAC-dependent expression of the hut operon in response to nitrogen-limited growth. The nac-10 mutant showed about half as much histidase as the wild type under NAC-dependent conditions (data not shown), suggesting that the NAC protein encoded by nac-10 retained considerable activity. However, NAC-dependent activation of histidase expression in E. coli is weak, making it dangerous to draw strong conclusions based on small differences.

The N-terminal 100 amino acids of NAC are sufficient to activate and repress transcription. A collection of mutants with deletions of portions of the C-terminal part of the K. aerogenes and E. coli nac genes (nac_K and nac_E, respectively) was generated by exonuclease III digestion as described in Materials and Methods. These truncated nac genes were cloned into pGD103, where nac expression was under the control of the lacZ promoter. K. aerogenes nac-203 mutant KC2725 was transformed with the resulting plasmids, and histidase and glutamate dehydrogenase activities were determined as a measure of NAC-dependent activation and repression. For convenience, the truncated proteins will be named by the number of the last remaining amino acid. Thus, wild-type E. coli NAC is NAC_E305 and a K. aerogenes NAC lacking the last five amino acids is NAC_K300.

View larger version (28K): [in this window] [in a new window]   FIG. 1.   hut activation (A) and gdh repression (B) by N-terminal fragments of NAC. K. aerogenes KC2725 (nac-203::Tn5-131) was transformed with derivatives of pGD103 containing an insert that coded for either full-length NAC or truncated versions of NAC. These transformants were assayed for NAC-dependent activation of histidase formation and repression of GDH formation. The values on the x axis represent the number of amino acids remaining in the NAC fragment. Thus, 305 represents the wild type and 100 represents the N-terminal 100 amino acids. Histidase and GDH values are reported as specific activities.

Activity of N-terminal fragments of NAC purified using MBP tags. Fragments of NAC containing only the N-terminal 100, 120, or 129 amino acids were nearly as active as wild-type NAC in vivo. In order to prove that these fragments were directly responsible for the activity, we attempted to demonstrate activation of transcription in vitro by using purified components. To purify these fragments, we fused the MBP to the N terminus via a peptide linker that could be cleaved with factor Xa protease, leaving an N-terminal methionine, as is seen in wild-type NAC. MBP-tagged variants of wild-type NAC (NAC_K305 and NAC_E305) and two N-terminal fragments (NAC_K129 and NAC_E100) were constructed. All of these fusion proteins were inactive in vivo.

View larger version (47K): [in this window] [in a new window]   FIG. 2.   Activation of hutUH transcription in vitro by NAC and N-terminal fragments of NAC. In vitro transcription was carried out by using purified components as described in Materials and Methods. The supercoiled template, pCB695, was similar to pAM1202 (8) and contained the hutUH promoter region cloned upstream of a strong transcriptional terminator such that transcription from this promoter would yield a transcript with the mobility on SDS-PAGE indicated at the right. All reactions contained RNA polymerase, nucleotides (including radioactive UTP), and buffers. NAC and NAC fragments were added as indicated. (A) Addition of full-length NAC. Lanes: 1, no NAC; 2 and 3, 0.04 and 0.4 µg of NAC (0.06 and 0.6 pmol of dimers) purified from K. aerogenes by salt precipitation (8); 4 and 5, 0.1 and 1 µg of MBP-NACK (1.3 and 13.3 pmol of MBP-NACK monomers) treated with factor Xa for 30 min immediately before addition to the transcription mixture; 6 and 7, 0.1 and 1 µg of MBP-NACE (1.3 and 13.3 pmol of monomers) treated with factor Xa for 30 min immediately before addition to the transcription mixture; 8, 3 µg of MBP-NACE (40 pmol of monomers) treated with factor Xa for 16 h before addition to the transcription mixture. (B) Addition of NAC fragments. Lanes: 1, no NAC; 2 and 3, 0.07 and 0.7 µg of NAC (1.1 and 10.5 pmol of dimers) purified from K. aerogenes by salt precipitation; 4 and 5, 0.1 and 1 µg of MBP-NACE100 (1.9 and 19 pmol of monomers) treated with factor Xa for 30 min immediately before addition to the transcription mixture; 6 and 7, 0.13 and 1.3 µg of MBP-NACK129 (1.9 and 19 pmol of monomers) treated with factor Xa for 30 min immediately before addition to the transcription mixture; 8 and 9, 0.14 and 1.4 µg of MBP-NACK305 (1.9 and 19 pmol of monomers) treated with factor Xa for 30 min immediately before addition to the transcription mixture.

Activity of NAC fragments with his₆ tags. Full-length NAC_E305 and NAC_K305 with the nickel-binding his₆ tag at the C terminus were active in vivo. In fact, the activation of histidase and repression of GDH formation were even stronger with NAC305-his₆ that with the corresponding unmodified NAC305 (Table 3). This increased activation and repression appears not to be specific to the nature of the C-terminal tag, since a cloning artifact that fused an 11-amino-acid peptide (DFGRSGHTDSL) to the C terminus of NAC_K305 also resulted in more activation and repression than NAC_K305 (Table 3). Although the modified E. coli NAC (NAC_E305-his₆) was active in vivo, the modification did not prove useful as a means of purifying NAC_E, since NAC_E305-his₆ was insoluble in all of the buffers tested, just like NAC_E305.

Subunit structures. We next used the tagged NAC polypeptides to search for the dimerization site on NAC. These data (along with a summary of the in vitro activation data and DNA-binding data) are presented in Table 4. NAC_E and NAC_K with an N-terminal MBP tag eluted as monomers. If factor Xa was added, the elution profile was complex and included a peak of the size expected for MBP-NAC (uncleaved, monomer), MBP (monomer), NAC (dimer), and an unknown peak with an apparent size of 8.5 kDa (presumed to be a degradation product of the 33-kDa NAC monomer).

	DISCUSSION

Top Abstract Introduction Materials and methods Results Discussion References

Perhaps the most surprising finding in these experiments is that the N-terminal 100 amino acids of NAC contain the determinants for all of the known functions of NAC. NAC100 was able to activate histidase formation in vivo and hutUp transcription in vitro. NAC100 was able to repress GDH formation in vivo. NAC_K100-his₆ was able to bind a DNA fragment with a NAC-binding site, and NAC_K100-his₆ was present as a dimer in solution. In short, NAC_K100 (or NAC_K100-his₆) displayed all four known properties of NAC: activation of transcription, repression of transcription, site-specific DNA binding, and dimerization.

All of the truncations of NAC isolated in these experiments, except NAC_K300, showed considerable NAC activity. However, several tight nac mutants have been isolated before by using selection and screening. One of these, nac-203::Tn5, has been sequenced and is an insertion of Tn5 after the 199th codon of nac_K. The data in Fig. 1 suggest that this mutation should leave NAC partially active, but nac-203 is one of the tightest nac mutants in our collection (although even nac-203 retains some activity, particularly on the ure operon promoter). Thus, it seems likely that NAC_K300 and NAC_K199 are inactive because of instability or misfolding rather than because of any lack of informative sequence.

The data presented here also provide some information about the function of the N and C termini of NAC. Additions to the N terminus block the ability of NAC to activate transcription, and small additions to the C terminus increase this activity, at least in vivo. The effect of the N-terminal additions appears to be intrinsic to the protein, since it is also seen in vitro with purified protein. Moreover, the ability to activate transcription can be recovered (in the case of MBP-NAC) by removing the N-terminal addition. The effect of the C-terminal additions may likewise be intrinsic to the protein, but it may also reflect an increased stability of the C-terminally modified NAC. We have been unable to raise a high-titer antibody with sufficient specificity for NAC to address this question directly, but we have previously suggested that wild-type NAC is rapidly turned over in vivo (14).

Two lines of evidence led us to suspect an interaction between the N and C termini of NAC and perhaps other LTTRs. First, MBP-NAC-his₆ was completely resistant to cleavage by factor Xa, while MBP-NAC was readily cleaved. Second, other work in this laboratory (1) has shown that replacement of the amino-terminal 20 amino acids of NAC with the corresponding amino acids from OxyR results in an inactive NAC unless the C-terminal domain of this NAC-OxyR chimera is also replaced with the C-terminal domain from OxyR.

The helix-turn-helix region thought to be the DNA-binding site is located very close to the N terminus of LTTRs (about amino acids 20 to 40 for NAC). The binding site for coeffectors of LTTRs has generally been found in the C-terminal domain, between amino acids 100 and 200 (15). Thus, an effect of the C-terminal domain on the N-terminal domain must exist. The specificity requirement in the OxyR-NAC chimeras (1) and the ability of the C-terminal his₆ tag to prevent cleavage of MBP from the N terminus of MBP-NAC-his₆ seem more consistent with a hypothesis of direct interaction than with transmission of a signal through the middle part of the protein. However, in the absence of experiments to test a direct interaction, such arguments remain speculative at best.

It was exciting to discover that the four known regulatory properties of NAC are encoded within a NAC fragment containing only the N-terminal 100 amino acids, NAC100. It was disappointing to discover that a fifth property of NAC, the insolubility of NAC_E, was also encoded within this fragment. Most of the amino acid differences between NAC_E and NAC_K lie in the C-terminal domain; the N-terminal domains are more similar (14). Nevertheless, even NAC_E100-his₆ is much more insoluble than NAC_K100-his₆. Thus, future studies will continue to focus on NAC_K for in vitro studies.