INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

A prominent class of prokaryotic enhancer-binding proteins activates transcription by the ⁵⁴ holoenzyme form of RNA polymerase (23, 26, 30, 50). One such protein is nitrogen regulatory protein C (NtrC), which activates transcription in response to limitation of combined nitrogen in the medium. The NtrC protein from enteric bacteria has been well studied as an activator for the glnA gene; this gene encodes glutamine synthetase, an enzyme with a major role in assimilation of ammonia.

NtrC can function both as an activator and as a repressor of glnA transcription, depending on the nutritional status of the cell (41, 56). Positive and negative controls of expression are achieved at two different promoters: activation at a downstream ⁵⁴-dependent promoter and repression at an upstream ⁷⁰-dependent promoter (Fig. 1A). Under nitrogen-limiting conditions, NtrC is phosphorylated by the protein kinase NtrB (17, 32). Phosphorylated NtrC forms an unusual hexamer or octamer at the two sites that constitute the glnA enhancer (61). To activate transcription, this oligomer contacts ⁵⁴ holoenzyme at the glnA promoter by means of a DNA conformational change (55). It catalyzes isomerization of closed complexes between ⁵⁴ holoenzyme and the promoter to transcriptionally productive open complexes (29, 39, 46) in a manner that depends upon hydrolysis of ATP and an energy-coupling mechanism (37, 39, 55, 57). When NtrC is bound to the enhancer, it represses transcription from a secondary ⁷⁰-dependent promoter that lies in the enhancer region (41). Phosphorylation and oligomerization, which are essential for ATP hydrolysis and therefore for transcriptional activation (1, 40, 59, 60), are not required for repression of transcription (20).

View larger version (24K): [in this window] [in a new window]   FIG. 1.   (A) Diagram of the glnA promoter-regulatory region from S. typhimurium (not to scale). The upstream binding sites for NtrC, which are centered at 140 and 108 with respect to the major 54-dependent transcriptional start site at +1, function as a transcriptional enhancer (33, 42). Conserved promoter sequences recognized by the 54 holoenzyme lie at 24 and 12, as indicated (23, 50). When phosphorylated, NtrC forms an oligomer at the enhancer (see text) and activates transcription by 54 holoenzyme. Both phosphorylated and unphosphorylated NtrC can bind to the enhancer and repress transcription from the secondary 70-dependent promoter that lies in the enhancer region (20, 41). (B) Domain structure of NtrC (not to scale) (reviewed by Kustu et al. [22], Weiss et al. [58], and Morett and Segovia [30]). An NtrC monomer (52,238 Da) contains 469 amino acid residues and is composed of three domains. The N-terminal receiver domain (~120 residues) contains the site of phosphorylation, D54. Under nitrogen-limiting conditions, this aspartate residue receives a phosphate from the phosphorylated NtrB protein, a physiological signal that is necessary for NtrC-mediated transcriptional activation (1, 32, 57). NtrB and NtrC constitute a sensory kinase-response regulator pair in a "two-component" signal transduction system (21, 34). The N-terminal domain is linked by a glutamine-rich flexible linker ("Q-linker") to the central domain (~240 residues), which appears to be directly responsible for ATP hydrolysis and transcriptional activation by 54 holoenzyme (see text). This domain is highly conserved among activators of 54 holoenzyme (Fig. 2). The C-terminal domain (~90 residues) contains a helix-turn-helix DNA-binding motif (40, 58) and the major dimerization determinants of the protein (19, 35).

Each monomer of the dimeric NtrC protein is comprised of three domains (Fig. 1B). The N-terminal domain contains the site of phosphorylation, D54 (17, 32). The C-terminal domain is responsible for binding to the enhancer (7, 40). The central domain of NtrC, amino acid residues 141 to 376 (Fig. 2), is directly responsible for transcriptional activation (2, 10, 14, 15, 37), which appears to entail oligomerization, nucleotide binding and hydrolysis, and coupling of energy to a thermodynamically unfavorable change in the conformation of ⁵⁴ holoenzyme. The functional significance of the central domain of NtrC is underscored by the high degree of sequence similarity shared by all members of the ⁵⁴ activator family (4, 6, 44). Sequence comparisons among these homologous proteins have revealed seven highly conserved amino acid segments (designated C1 to C7) along the central domain (30). When such comparisons were combined with genetic and biochemical analyses and were expanded to include a broader range of purine nucleotide-binding proteins (38, 47), interesting parallels that provided insights into the distribution of the subfunctions of positive control within the central domain emerged. The C1 and C4 regions, respectively, were identified as likely Walker A and Walker B motifs, which are involved in ATP binding and hydrolysis (37, 38, 43, 44, 52, 57). The C3 region appears to be analogous to the "switch I" region of other purine nucleotide-binding proteins (37, 38), a region that lies between the Walker A and B motifs and plays an important role in coupling nucleotide hydrolysis to biological output (e.g., see references 3, 13, 18, and 48). The C7 region is required for nucleotide binding and is probably involved in binding the nucleotide base (38, 43). Finally, recent secondary structure predictions coupled with the use of recognition algorithms for protein folds indicated that the central domain of activators of ⁵⁴ holoenzyme adopts a mononucleotide-binding fold similar to those of the eukaryotic signaling protein p21^ras and the G domain of the bacterial polypeptide elongation factor Tu (38).

View larger version (84K): [in this window] [in a new window]   FIG. 2.   Alignment of the central domains of nine activators of 54 holoenzyme and the locations of amino acid substitutions in NtrC mutant proteins. The activators are chosen from members of the , , and  subgroups of the purple bacteria. Respectively, the sequences belong to S. typhimurium NTRC, Klebsiella pneumoniae NIFA, R. leguminosarum DCTD, Alcaligenes eutrophus HOXA, Pseudomonas putida XYLR, Caulobacter crescentus FLBD, E. coli FHLA, A. eutrophus ACOR, and Pseudomonas aeruginosa PILR (original references given in reference 35). Residues that are identical in at least 6 activators are boxed. The common boundaries of their central domains are indicated by arrows, and their exact positions are according to Morett and Segovia (30). Another pair of arrows identify the locations of the AgeI and AscI restriction sites that correspond to the start and the end of the PCR mutagenesis target region (amino acids 173 to 321 in S. typhimurium NtrC). Four functional determinants are highlighted by symbols directly beneath their positions in the sequence alignment: the heavy black line denotes the glycine-rich segment that has been proposed to correspond to the Walker A motif of purine nucleotide-binding proteins (37, 38, 43, 44, 52, 57); the asterisk denotes the highly conserved aspartate at the end of the Walker B motif; the plain line denotes the putative switch I (effector) region (18, 37, 38); and the dashed line marks four residues believed to be involved in binding of the nucleotide base (38, 43). Letters above the sequence alignment designate all single amino acid substitutions for which in vitro measurements were made in this study and known substitutions resulting in NtrC(Rep) proteins. In cases where more than one such substitution was found for a given residue, the letters are arranged according to the severity of the effect, such that those resulting in the largest defects are placed closest to the alignment. Plain letters denote known NtrC(Rep) substitutions that had been described previously (37, 43, 56-58). Boldface letters indicate amino acid substitutions found in this study to yield NtrC(Rep) proteins. Like known NtrC(Rep) proteins, these new proteins were characterized by the absence of detectable transcriptional activities in vivo or in vitro. Italic letters represent other mutational substitutions generated in this study which resulted in little transcriptional activation and glutamine auxotrophy in vivo but readily detectable and different amounts of transcriptional activation in vitro (see Results).

Genetic studies of the NtrC protein of Salmonella typhimurium were facilitated by the isolation of a class of mutant proteins known as NtrC repressor [NtrC(Rep)] forms (37, 43, 56, 57). NtrC(Rep) proteins, which have been called "positive control" or "PC" forms in other systems, were defined as those that had lost the ability to activate transcription by ⁵⁴ holoenzyme but retained the ability to repress transcription by ⁷⁰ holoenzyme (37). Prior to this study, 13 NtrC(Rep) proteins carrying single residue changes in five of the seven conserved regions of the central domain (30) had been characterized. Their properties in vitro allowed us to distinguish between those regions involved specifically in Mg-ATP binding and those responsible for ATP hydrolysis per se (43) and to identify one region (C3 or switch I) required for transcriptional activation but not for nucleotide hydrolysis (37); by inference, the C3 region appears to be involved in interaction with ⁵⁴ holoenzyme, that is, in contact with the polymerase, and/or coupling of energy to a change in its conformation that allows it to form open complexes. Further advances in functional mapping were limited by the size of our mutant pool.

To characterize new functional determinants of NtrC, we performed random mutagenesis to scan most of the central domain of the S. typhimurium NtrC protein (residues 172 to 322) for new amino acid residues and regions important for positive control. We determined transcriptional activities of mutant proteins both in vivo and in vitro. As expected, these studies yielded additional NtrC(Rep) proteins. Unexpectedly, they also yielded a new and interesting class of mutant proteins that were more active than wild-type NtrC (NtrC^WT) at low concentrations but had greatly decreased transcriptional activation capacities at high concentrations.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Media. Complex media were nutrient broth (NB), Luria broth (LB), and green indicator plates (5), which were supplemented with glutamine (2 mM) and/or ampicillin (100 µg/ml) when appropriate. The minimal medium, NC (12), was supplemented with glucose (0.4%) as the carbon source, glutamine (5 mM) as the nitrogen source, and L-histidine (0.2 mM) to satisfy an auxotrophic requirement. Ampicillin (50 µg/ml) was added for plasmid-carrying strains to prevent loss of plasmids.

Random mutagenesis with PCR. Most of the DNA sequence encoding the central domain of NtrC (amino acids 141 to 376 or nucleotides 421 to 1128, where +1 designates the A of the ATG start codon for NtrC; Fig. 2) was subjected to random mutagenesis with PCR. The two oligonucleotide primers used to amplify this segment were 5'-ATTGGTCGGCTGTCGCGTTC-3' (nucleotides 460 to 479; upstream primer) and 5'-ATTTGGCTTCCACGCCTAAT-3' (nucleotides 1015 to 996; downstream primer). PCR conditions were chosen to exploit the inherent infidelity of DNA synthesis by Thermus aquaticus (Taq) DNA polymerase (24, 62). The template was pJES559 (Table 1), which carries malE-ntrC in an overexpression vector. The Mg²⁺ concentration was elevated to provide a large excess over the total deoxynucleoside triphosphates (dNTPs) present, a condition that further reduces the copying accuracy of the Taq polymerase (8). Reaction mixtures contained 5 ng of template, 100 ng of each primer, 50 µM (each) dNTP, 10 mM Tris-Cl (pH 8.3), 50 mM KCl, 4 mM MgCl₂ and 5 units of Taq DNA polymerase (AmpliTaq) in a total volume of 50 µl. PCRs were carried out on a Perkin-Elmer Cetus DNA Thermal Cycler with the following cycles programmed: 30 cycles of 1 min at 94°C, 1 min at 48°C, and 1 min at 70°C.

Genetic screen. A newly developed plate test was employed to discriminate between transformants carrying ntrC(Rep) alleles and those carrying the wild-type allele or ntrC null alleles. The screening medium was NB, which lacks sufficient glutamine to allow growth of glutamine auxotrophs. Because DH5 has an intact chromosomal ntrC gene, the introduction of an ntrC mutant allele on a multicopy expression plasmid (strain NCM1237 [Table 1]) did not affect glutamine-independent growth of the host strain. By contrast, when an ntrC(Rep) allele was introduced (strain NCM1946), there was no growth on NB because both glnA promoters on the host chromosome were inactivated. Interestingly, the control strain NCM1099, which expressed NtrC^WT from pJES311, did not grow as well as the host strain on NB (phenotype intermediate between those of NCM1237 and NCM1946). This "semirepressed" phenotype is probably explained by the fact that there is only a single chromosomal copy of ntrB, and hence NtrC is not phosphorylated well enough to drive adequate expression from the downstream promoter.

DNA sequencing. Strains were streaked out fresh from glycerol stocks, and plasmids were extracted with QIAGEN Miniprep kits. DNA sequencing of the entire mutagenized region was performed on an ABI373 automated sequencer using the upstream primer 5'-TATCGAGGTTAACGGCCCGA-3' (nucleotides 390 to 409). For each sample, a second sequencing run was also performed to confirm the mutation from the reverse direction. The downstream primer was the same as that used previously for PCR mutagenesis.

Strain constructions. Preparation of lysates of phage P22, transductional crosses, and the use of green indicator plates to purify transductants free of phage were as described previously (5).

10 hisF645 putPA

glnAp381 glnAp131 hisF645

Growth conditions and -galactosidase assay. Cultures of strains carrying putative ntrC(Rep) alleles in the SK3424 background were adapted to growth on glutamine as the sole nitrogen source by overnight growth in minimal medium containing 5 mM glutamine and 2 mM NH₄Cl. These precultures were subjected to centrifugation at room temperature. Cells were washed once with fresh medium containing 5 mM glutamine as the sole nitrogen source and were inoculated into 10 ml of this medium at an initial optical density at 650 nm (OD₆₅₀) of 0.05. Cultures were incubated in baffled culture tubes on a New Brunswick model G76 water bath shaker at 37°C and 240 rpm. Samples in the exponential phase of growth were harvested in duplicate between OD₆₅₀s of 0.2 and 0.6 and were immediately frozen on dry ice. Cultures of control strains SK3424 and SK3425 were grown along with each set of strains carrying putative ntrC(Rep) alleles in the overexpression vector. -Galactosidase assays were carried out, and units of activity were calculated according to Miller (28) (see Fig. 3).

Western blotting. Cultures grown under the same conditions as those used for -galactosidase assays (500 µl at an OD₆₅₀ of 0.5) were subjected to centrifugation, and cell pellets were suspended in 20 µl of sodium dodecyl sulfate (SDS) gel-loading buffer (~25-fold concentration). In each case, the entire sample was subjected to electrophoresis in an SDS-10% polyacrylamide gel and proteins were transferred to nitrocellulose membranes as described previously (45). Membranes were exposed to polyclonal mouse antiserum directed against the carboxy-terminal domain of the NtrC protein from S. typhimurium, and bands were detected with secondary antibodies directed against mouse immunoglobulin G and coupled to alkaline phosphatase (Bio-Rad). Purified NtrC and prestained markers (Broad Range; New England Biolabs) were used as standards.

Overexpression and small-scale purification of NtrC proteins. Plasmids for overproduction of NtrC proteins could not be maintained in Escherichia coli strains that contained T7 RNA polymerase (57); therefore, all of the 55 plasmids carrying putative ntrC(Rep) alleles were transformed into an Hfr strain of E. coli (NCM724), and induction was accomplished by infection with an M13 phage that carries T7 RNA polymerase under control of the lac promoter (M13mGP1-2). Cultures were grown in 50 ml of LB medium containing 2 mM glutamine and 100 µg of ampicillin per ml at 37°C with vigorous shaking. At an OD₆₀₀ of ~0.5, M13mGP1-2 was added to a multiplicity of infection of about 10 (49). IPTG (isopropyl--D-thiogalactoside) was also added, to 0.5 mM, to allow induction of T7 RNA polymerase. After 2 h, cultures were harvested by centrifugation. Cells were suspended in 5 ml of breakage buffer (50 mM Tris-acetate [pH 8.2], 200 mM KCl, 1 mM EDTA, 1 mM dithiothreitol [DTT]) and were lysed by two passages through a French pressure cell (SLM-Aminco) at 8,000 lb/in². The resulting lysate was subjected to centrifugation at 27,000 × g for 30 min at 4°C, and ammonium sulfate (35% final concentration) was added to the supernatant at room temperature. The ammonium sulfate pellet was dissolved in 10 ml of B-50 buffer (10 mM Tris-acetate [pH 8.2], 50 mM KCl, 0.1 mM EDTA, 5% glycerol, 1 mM DTT) and loaded on DEAE Sephacel (Pharmacia Biotech) (1 ml of resin packed into Poly-Prep chromatography columns [Bio-Rad]) at room temperature. Columns were washed twice with 5 ml of B-50 buffer, and then the bottom of each column was closed and 2 ml of B-500 buffer (10 mM Tris-acetate [pH 8.2], 500 mM KCl, 0.1 mM EDTA, 5% glycerol, 1 mM DTT) was added. After 10 min at room temperature, the proteins were eluted by centrifugation at 2,500 × g for 5 min at 4°C. To each sample (2 ml), 18 ml of B-0 buffer (10 mM Tris-acetate [pH 8.2], 0.1 mM EDTA, 5% glycerol, 1 mM DTT) was added and the diluted samples were loaded on heparin Sepharose CL-6B (Pharmacia Biotech) (0.5 ml of resin packed in the columns described above) at room temperature. The columns were washed with 5 ml of B-50 buffer, and then the bottoms were closed and 2 ml of B-500 buffer was added. After 10 min of incubation at room temperature, the proteins were eluted and dialyzed into B-50 buffer overnight at 4°C. Proteins were assessed to be at least 90% pure by visual inspection of SDS-10% polyacrylamide gels stained with Coomassie blue (Sigma). NtrC concentrations were determined from absorption at 280 nm in the presence of 6 M guanidine hydrochloride (Pierce) by using an extinction coefficient of 44,907.8 M¹ cm¹ for NtrC (9). Because none of the amino acid substitutions in mutant proteins involved W residues, all mutant proteins had approximately the same extinction coefficient as NtrC^WT.

Assay for open complex formation. The ability of NtrC proteins to catalyze the formation of open complexes by ⁵⁴ holoenzyme was assessed on a supercoiled plasmid template by using a single-cycle transcription assay, as described previously (10, 39, 55, 57). The template was plasmid pJES534 (1 nM) (40), which carries a "strong enhancer" situated ~460 bp from the glnA promoter and directs synthesis of a 155-nucleotide transcript that contains no uracil (57). Final concentrations of reagents added to the buffer were as follows: core RNA polymerase, 30 nM; ⁵⁴, 50 nM; carbamyl phosphate, 10 mM; ATP, 4 mM; GTP, 400 µM; CTP, 100 µM containing 5 µCi of [-³²P]CTP. The NtrC concentration was varied. After allowing phosphorylation of NtrC by carbamyl phosphate for 10 min at 37°C, open complex formation was initiated by adding ATP to the reaction mixture. After 10 min, synthesis of transcripts was initiated by adding a mixture of heparin (100 µg/ml) and the remaining two nucleotides. After an additional 10 min, transcripts were precipitated, electrophoresed on 6% sequencing gels, and quantified with a Molecular Dynamics PhosphorImager. To test the transcriptional ability of the NtrC mutant proteins in the absence of phosphorylation, carbamyl phosphate was omitted from the reaction mixture. The detection limit with this assay was 0.05 fmol.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Random PCR mutagenesis yielded 55 different single amino acid substitutions in the central domain of NtrC. To generate new mutant forms of NtrC defective in positive control, we performed PCR mutagenesis of the DNA encoding most of its central domain (see Materials and Methods). PCR-amplified DNA fragments were then substituted for the corresponding fragment of pJES311, which expresses wild-type NtrC from a strong T7 promoter and translational start. The resulting plasmids were transformed into E. coli DH5 to generate a random mutant library.

Assays of glnA'-`lacZ expression indicated that all 55 new mutant forms of NtrC were deficient in transcriptional activation in vivo. We measured levels of expression of a (glnA'-`lacZ) fusion at the put locus of S. typhimurium for all 55 mutant NtrC proteins containing single residue changes to assess their residual ability to activate transcription (Table 2 and Fig. 3). The -galactosidase activity of the host strain (SK3424 ntrC) was ~850 U/ml/OD₆₅₀ and is attributable to transcription from the upstream ⁷⁰-dependent promoter. When a multicopy expression vector encoding a known NtrC(Rep) protein, NtrC^S207F (that is, NtrC with a change from serine to phenylalanine at position 207), was introduced into SK3424, the -galactosidase activity dropped to 16 U/ml/OD₆₅₀ because NtrC^S207F represses transcription from the upstream promoter and is unable to activate transcription from the ⁵⁴-dependent promoter. By contrast, when the multicopy expression vector encoding the NtrC^WT protein was introduced, expression increased to ~1,350 U/ml/OD₆₅₀ under the derepressing growth conditions employed (see Materials and Methods). This increase in glnA'-`lacZ expression was attributed to NtrC-dependent transcriptional activation at the downstream ⁵⁴-dependent promoter because transcription from the upstream ⁷⁰-dependent promoter was presumably repressed. When similar experiments were done with plasmids encoding all 55 new mutant forms of NtrC, they allowed us to divide these forms into two classes. One class (40 of 55) yielded -galactosidase values comparable to those of known NtrC(Rep) proteins, i.e., 70 U/ml/OD₆₅₀. The remainder yielded values ranging from ~90 to ~500 U/ml/OD₆₅₀, less than that of the ntrC mutant host strain.

View larger version (15K): [in this window] [in a new window]   FIG. 3.   In vivo expression from the glnA promoter in three strains carrying mutant NtrC proteins. Expression of a (glnA'-`lacZ) fusion at the put locus of S. typhimurium was measured as a function of the OD650 for cultures grown under nitrogen-limiting conditions (see Materials and Methods). Curves labeled A220V (open diamonds), M244V (open circles), and G291R (open squares) give data for strains carrying plasmids that overexpress NtrC proteins with these three amino acid substitutions (strains SK3468, SK3458, and SK3447, respectively). In each case, two data points were used to determine the slope of the line, which reflects the differential rate of glnA'-`lacZ expression. All three mutant NtrC proteins activated transcription better than NtrCWT at low concentrations in vitro (Fig. 4B through D). Curves labeled WT (closed squares) and S207F (open triangles) give data for control strains carrying plasmids that overexpress NtrCWT (strain SK3425) and NtrCS207F (strain SK3426), a known NtrC(Rep) protein. The curve labeled host (closed circles) gives data for the ntrC mutant host strain SK3424. For control strains, three data points were taken to allow reliable comparisons between experiments.

Residual glnA'-`lacZ expression in strains expressing mutant NtrC proteins is <sup>54</sup> dependent. For mutant NtrC proteins that yielded the highest residual glnA'-`lacZ expression (90 to 500 U/ml/OD₆₅₀), we wanted to determine whether this activity was due to residual transcriptional activation from the downstream ⁵⁴-dependent promoter or to loss of repression at the upstream ⁷⁰-dependent promoter. To do so we performed two sorts of experiments: first, we ascertained whether residual expression was dependent on function of the ntrA gene product, ⁵⁴; second, we assessed immunologically whether any of the mutant proteins were significantly degraded.

Forty-one of 55 mutant forms of NtrC were soluble and could be purified. To characterize their transcriptional activation capacities in vitro, we attempted to purify all 55 NtrC mutant proteins by using the rapid protocol described in Materials and Methods. Fourteen of the proteins could not be purified by this protocol because they were in the pellet after cells were broken with the French pressure cell (Table 2). The 41 proteins that were purified yielded from 2.6 to 18 nmol of dimer per 50 ml of culture (0.85 to 6 µM), with the average yield around 6 nmol (2 µM). The variation in yield could be due to inefficient induction during overexpression or to differences in behavior during purification. All proteins were at least 90% pure based on visual inspection of SDS-polyacrylamide gels stained with Coomassie blue, and none showed evidence of degradation (data not shown).

Seventeen of 41 soluble mutant proteins failed to activate transcription by ⁵⁴ holoenzyme in vitro. The 41 NtrC mutant proteins purified were tested for the ability to catalyze the formation of open complexes by ⁵⁴ holoenzyme at the glnA promoter (Table 2; see also Materials and Methods). Phosphorylated NtrC^WT catalyzed the formation of open complexes at concentrations ranging from 10 nM dimer (the lowest concentration tested) up to 500 nM dimer (the highest concentration tested) (Fig. 4 and Table 3). In contrast, 17 of the 41 phosphorylated NtrC mutant proteins, which had amino acid substitutions in six of the seven highly conserved regions in the central domain, failed to catalyze open complex formation at 10, 20, 50, and 100 nM dimer (Table 2). Four of these 17 (N196H, F217L, L253V, and V303A) were also tested at 200 and 500 nM dimer and, in all cases, yielded <1% as many open complexes as the wild-type protein at the same concentrations (data not shown). The 17 mutant proteins that failed to activate transcription from the ⁵⁴-dependent glnA promoter in vitro also showed little glnA transcription in vivo (<70 U/ml/OD₆₅₀ of -galactosidase activity from the glnA'-`lacZ fusion). Hence, they behaved like the NtrC(Rep) proteins characterized previously.

View larger version (23K): [in this window] [in a new window]   FIG. 4.   Transcriptional activation by selected mutant NtrC proteins in vitro. Formation of open complexes by 54 holoenzyme was assessed by a single-round transcription assay as described in Materials and Methods. NtrC proteins were present at the concentrations indicated and were phosphorylated with carbamyl phosphate (10 mM). Effects of single amino acid substitutions were assessed by comparison to NtrCWT (open circles) in each panel. (A) V177A and I200V. (B) A220V. (C) Q230R and M244V. (D) Y262C and G291R. Proteins were chosen to illustrate their hyperactivity at low concentrations and the widespread locations of their amino acid substitutions within the central domain of NtrC: V177A between conserved regions 1 and 2 (C1-C2), I200V (C2), A220V (C3), Q230R (C3-C4), M244V (C4-C5), Y262C (C5), and G291R (C6). Data for NtrCWT is included in each plot to illustrate the variability between experiments.

The remaining 24 of 41 soluble mutant proteins did activate transcription by ⁵⁴ holoenzyme in vitro. Unexpectedly, in vitro transcription assays revealed a large class of 24 mutant proteins that did catalyze formation of open complexes by ⁵⁴ holoenzyme in vitro. Among these proteins, 14 yielded essentially no glnA transcription in vivo (<70 U/ml/OD₆₅₀ of -galactosidase activity from the glnA'-`lacZ fusion). Ten of these proteins (A180T, N196S, M197V, E208V, F210L, K214E, Y262C, Y262N, F298C, and I304F) failed to reach the maximum activity of phosphorylated NtrC^WT in vitro (achieved at 100 nM dimer) at any concentration, and two of these (E208V and K214E) had very low levels of activity (Table 3 and Fig. 4D). The remaining four proteins of this subclass (A194T, A220V, Q230R, and M244V) did reach the maximum activity of NtrC^WT (Table 3; Fig. 3, 4B, and 4C).

Nineteen of the 24 NtrC mutant proteins that activated transcription by ⁵⁴ holoenzyme in vitro showed hyperactivity at low concentrations. Perhaps our most interesting result was the observation that 19 of the 24 mutant proteins that catalyze the formation of open complexes by ⁵⁴ holoenzyme in vitro (discussed above) did so better than NtrC^WT at low concentrations. They yielded 3 to 20 times as many open complexes as NtrC^WT at 10 nM dimer (Table 3; Fig. 4A through D). Figures 4A through D show some particularly striking examples of hyperactivity (V177A, I200V, A220V, Q230R, M244V, and G291R) and a single example (Y262C) in which hyperactivity at low concentrations is not accompanied by the ability to reach the maximum activity for NtrC^WT at any concentration.

All 24 mutant proteins that are active in vitro must be phosphorylated. Some NtrC mutant proteins, called NtrC constitutive [NtrC(Con)], can hydrolyze ATP and activate transcription in vitro without being phosphorylated. Amino acid substitutions in NtrC(Con) proteins have been localized to both the N-terminal regulatory domain and the central domain (10). To determine whether any of the 24 NtrC mutant proteins that were active in vitro were active without being phosphorylated, i.e., had the NtrC(Con) phenotype, all were tested for transcriptional activation in the absence of carbamyl phosphate, the in vitro phosphate donor. All 24 mutant proteins were tested at concentrations of 50 and 100 nM dimer, concentrations at which all of the phosphorylated proteins gave rise to clearly visible transcripts, as did the unphosphorylated NtrC(Con) protein NtrC^S160F (10), which was used as a positive control. No visible transcripts were observed for the 24 new mutant proteins, showing that all must be phosphorylated to catalyze formation of open complexes by ⁵⁴ holoenzyme at the glnA promoter.

Levels of NtrC protein from the expression plasmid in the absence of T7 polymerase are hundreds of times higher than the chromosomal level. The unexpected activities of many mutant NtrC proteins in vitro, together with the observation that their activities dropped considerably with an increase in protein concentration in many instances, led us to quantitate the overexpression of these proteins in vivo. Preliminary Western blot experiments performed with cells carrying plasmids that encoded either NtrC^WT (pJES311) or the mutant proteins had already attracted our attention to the high levels of overexpression even in the absence of T7 RNA polymerase.

13

2

2

7

13

7

View larger version (48K): [in this window] [in a new window]   FIG. 5.   Amounts of NtrC protein expressed from a single chromosomal ntrC+ allele or from plasmid pJES311, which overexpresses NtrCWT. Cells were cultivated under nitrogen-limiting conditions as for assays of glnA'-`lacZ expression (see Materials and Methods), and 0.5 ml of culture at an OD650 of 0.5 was concentrated and loaded on an SDS-10% polyacrylamide gel. Amounts of NtrC protein in crude lysates were determined by Western blotting with mouse polyclonal antiserum raised against the carboxy-terminal domain of NtrC. (A) Crude lysates of SK3424 (ntrC) (lane 1) and SK1490 (ntrC+) (lane 2) were compared with dilutions of lysates of SK3425(pJES311): lane 3, 1:1,250; lane 4, 1:625; lane 5, 1:125; lane 6, 1:25; lane 7, undiluted. (B) Crude lysates of strain SK1490 (lanes 2, 4, 6, and 8) were compared with different amounts of purified NtrCWT added to crude lysates of SK3424: lane 1, 0 ng; lane 3, 12.5 ng; lane 5, 25 ng; lane 7, 50 ng. MW, BenchMark molecular weight standards (Gibco BRL).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

New NtrC(Rep) proteins (positive control forms) have lesions in regions of the central domain that are not well studied. In vitro assays of transcriptional activation at the ⁵⁴-dependent glnA promoter by 41 NtrC mutant proteins with decreased activity in vivo yielded 16 new forms that failed in transcriptional activation. Together with the 13 previously studied forms, lesions in these NtrC(Rep) proteins affect residues in or, in one case, directly adjacent to all seven of the conserved motifs in the central domain (30, 38). These include two regionsC2 and C5in which such lesions had not been found previously and the unusually long conserved region C6, in which only a single lesion had been characterized (37, 43, 57). Lesions affecting C2 and C5 have not yet been characterized for other ⁵⁴-dependent activators, and few lesions in C6 have been characterized (54). In only 1 of 29 cases (T280A in the C6 region) was the amino acid residue in the inactive mutant form of NtrC found at the corresponding position of a homologous activator of ⁵⁴ holoenzyme (DctD of Rhizobium leguminosarum). Several of the residues that were altered in the inactive mutant forms of NtrC (N196, T218, D296, R358) had been predicted to be "functional" residues (38).

NtrC mutant forms that are hyperactive at low concentrations have lesions throughout the central domain. Unexpectedly, 24 new mutant forms of NtrC that failed to activate transcription in vivo, or activated poorly, did activate in vitro. Because the activation by many of these forms decreased at high concentrations in vitro, their apparent failure to activate in vivo is probably accounted for by high levels of overexpression (~500-fold [see Results]). Nineteen such forms were more active at low concentrations (10 to 50 nM) than NtrC^WT, and many of these showed decreases in activity at higher concentrations (200 to 500 nM). These mutant forms have lesions that are widespread in the central domain of NtrC. They occur in the region between C1, the Walker A motif, and C2 (e.g., V177A and L181I); in or immediately adjacent to the C2 region (e.g., A194T and I200V); in the region between C4 and C5 (e.g., M244V and T250I); and in region C6, where they affect the three adjacent residues E290 to K292 and residue E298. There is one such lesion (A220V) in C3, the switch I region, and one (Q230R) between C3 and the Walker B motif, C4. The significance of the widespread occurrence of these forms is not understood, nor is the functional basis for their unexpected phenotype in vitro. Hyperactivity at low concentrations may be due to increased oligomerization and ATPase activity, increased biological output, or both, and the basis may be different for different lesions. Lesions to hyperactivity that occurred between, rather than within, conserved regions of NtrC tended to affect residues that were tree determining for the NtrC subfamily of ⁵⁴-dependent activators (e.g., V177, L181, Q230, and T250) (38). These residues were postulated to participate in unique aspects of oligomerization or of the response to signals from the regulatory domain.

Comments on new lesions in the Walker A (C1), Walker B (C4), and switch I (C3) regions. The Walker A motif for activators of ⁵⁴ holoenzyme (C1) differs from that of other members of the purine nucleotide-binding protein family in ending with a conserved E or D, rather than the usual S or T, which is often involved in coordination of the divalent cation (38). NtrC^E175G, in which the conserved glutamate is altered, was inactive in vivo and in vitro. The same was true when the corresponding glutamate in the homologous activator DctD was changed to either T or A (11). The mutant DctD proteins were shown to be greatly defective in ATP hydrolysis (11). Lesions in NtrC that altered the two hydrophobic residues preceding D239, the conserved aspartate of the Walker B motif, resulted in precipitation of the corresponding proteins (NtrC^F237S and NtrC^L238P). The hydrophobic residues normally present at these positions of purine nucleotide-binding proteins are part of a -strand (38). Like previously characterized lesions at position 239 (D239A, N, and C [43]), each of which resulted in loss of ATPase activity but not ATP binding, an additional lesion at this position, D239G, resulted in loss of transcriptional activation. Interestingly, although NtrC has a glycine at position 242, as do other "DEXG proteins," the E247G substitution at the corresponding position of Rhizobium meliloti DctD caused loss of transcriptional activation both in vivo and in vitro (54). This appears to be the only example in which changing a residue of a homologous activator to that found at the corresponding position of NtrC results in loss of activity.