Regulation of carbon utilization by sulfur availability in Escherichia coli and Salmonella typhimurium

Bacteria require an ample supply of a variety of nutrients to maintain continuous growth. Among these nutrients are sources of carbon, nitrogen, phosphorus, sulfur and iron. The use of each such nutrient is in general controlled by one or a few pleiotropic transcriptional regulatory proteins. Thus, in Escherichia coli and the closely related Salmonella typhimurium, carbon utilization is regulated by the cAMP-receptor protein (CRP) and the catabolite repressor/activator (Cra) protein (Saier & Ramseier, 1996 ; Saier et al., 1996 ), nitrogen utilization is controlled by the NtrBC sensor kinase/response regulator pair in conjunction with the GlnB and GlnD proteins (Magasanik, 1996 ; Reitzer, 1996a , b ), phosphorus utilization is controlled by the PhoR/PhoB sensor kinase/response regulator pair (Wanner, 1996 ), iron utilization is regulated by the Fur transcription factor (Earhart, 1996 ; Gralla & Collado-Vides, 1996 ) and sulfur metabolism is controlled by the CysB transcriptional activator (Kredich, 1992 ).

The cysteine regulon includes most of the genes required for synthesis of cysteine and genes for uptake of sulfur sources such as L-cystine, sulfate, thiosulfate and taurine. Transcriptional activation of these genes requires CysB, the inducer N-acetyl-L-serine and conditions of sulfur limitation (Kredich, 1992 , 1996 ). CysB is a tetrameric LysR-type regulator with an N-terminal DNA-binding domain, a central inducer-binding domain and a C-terminal oligomerization domain that is essential for stability (Lochowska et al., 2001 ). Its activity is regulated by an efflux pump specific for cysteine metabolites (Dassler et al., 2000 ). CysB is also an autorepressor, preventing expression of its own structural gene, cysB.

E. coli and S. typhimurium can utilize a number of sulfur-containing compounds as sole sulfur source, including sulfate, sulfite, thiosulfate, sulfide, glutathionine, lanthionine, taurine and L-djenkolate [S,S'-methylene bis(cysteine)] as well as cysteine and cystine (Kredich, 1996 ; van der Ploeg et al., 1997 ). Although these organisms cannot utilize methionine as a sole sulfur source, Klebsiella strains can (Seiflein & Lawrence, 2001 ). Sulfide and thiosulfate are anti-inducers, probably exerting their effects by competing with N-acetyl-L-serine for binding to the CysB activator (Hryniewicz & Kredich, 1991 ; Ostrowski & Kredich, 1990 ). Cysteine inhibits inducer synthesis, resulting in maximal repression of the sulfur regulon. Growth with poor sulfur sources such as L-djenkolate or glutathione results in maximal derepression of the sulfur regulon. Rapid growth on rich media results in moderate induction of the regulon (Antón, 2000 ).

Hartmann & Boos (1993) demonstrated that mutations in phoB, the gene encoding the pleiotropic activator of the pho regulon, affect the carbohydrate fermentation phenotype in E. coli under certain physiological conditions, particularly under conditions of phosphate starvation. CysB, normally considered to be the central pleiotropic regulator of sulfur metabolism in enteric bacteria, has been shown to be required for normal acid induction of arginine decarboxylase (Shi & Bennett, 1994 ) as well as for sensitivity to novobiocin (Rakonjac et al., 1991 ) and mecillinam (Oppezzo & Antón, 1995 ). Thus, pleiotropic effects of phoB and cysB mutations other than those expected on the basis of their involvement in phosphorus and sulfur metabolic regulation, respectively, have been reported. In this report we present the results of our preliminary studies with representative examples of cysB mutant strains and provide analyses of the basis for specific carbohydrate utilization defects observed for these mutants.

Strains and plasmids.
The strains used in this study are listed in Table 1. The cysB allele in MDA7K was transferred by P1 transduction to three different genetic backgrounds (bottom of Table 1), selecting for tetracycline resistance and screening for the cysB phenotype (non-growth on minimal medium and growth upon the addition of cysteine). Plasmids were introduced into the various E. coli strains by electroporation as described by OCallaghan & Charbit (1990) .

Table 1. Bacterial strains and plasmids used in this study

Growth medium.
The minimal medium contained W salts (100 mM potassium phosphate, pH 7, plus 1 mM MgSO₄) (Smith et al., 1971 ) to which was added 0·01% L-tryptophan when required (i.e. for strains LJ4508, LJ4510, LJ4511 and LJ4512) plus 0·2% (NH₄)₂SO₄ or another nitrogen source (when indicated) and the specified carbon source. For sulfur-free medium, MgSO₄ was replaced by MgCl₂ and 0·2% (NH₄)₂SO₄ was replaced by 0·4% NH₄Cl. The rich medium used was LuriaBertani (LB) broth or nutrient broth (NB). These media are limiting for cysteine, and growth of cysB mutants in these media requires addition of cysteine (an efficient sulfur source) or djenkolate (a limiting sulfur source).

Cell growth was always at 37 °C. Fermentation was tested on eosin-methylene blue (EMB) plates lacking lactose, but containing the indicated carbon source at 1%, except for rhamnose which was present at 0·5%. These plates, like the rich media described above, are sulfur-limited. Carbon oxidation was assayed using 96-well Biolog plates (Bochner, 1993 ). Cysteine, when present, was added to a final concentration of 0·2 mM. Although cysteine can be growth-inhibiting at high concentrations (Harris & Lui, 1981 ), it is not inhibitory at this concentration. For all enzyme assays, cells were harvested during exponential growth.

Enzyme assays.
ß-Galactosidase was assayed by the method described by Miller (1972) . For the data presented in Table 5, which used minimal media, IPTG (1 mM) was present during growth to eliminate any effect of the lac operon repressor. ß-Galactosidase activities are either expressed in Miller units (µmol min^-1) (mg protein)^-1 (see Table 5) or in relative activities (see Figs 2 and 3). For reporter gene transcriptional fusion assays, cells were grown overnight in 5 ml NB plus 1% glucose, glucitol or formate at 37 °C with agitation. Fresh NB (100 ml) of the same composition was inoculated from the overnight culture and cells were harvested during exponential growth at 37 °C (OD₆₀₀ approx. 0·5). The cell density was adjusted and two drops of toluene were added to permeabilize the cells during vortexing for 30 s. The samples were incubated at 30 °C and aliquots were transferred to the appropriate buffers for reporter gene product assay. The ß-galactosidase buffer consisted of 2 mM EDTA, 70 mM Na₂HPO₄, 35 mM NaH₂PO₄ (pH 7·0), 1 mM MgSO₄, 10 mM KCl and 10 mM ß-mercaptoethanol. The same conditions were used to assay native ß-galactosidase as reported in Table 5. The ß-glucuronidase buffer consisted of 50 mM sodium phosphate, pH 7·0, 0·1 % Triton X-100 and 10 mM ß-mercaptoethanol. The assay mixtures were allowed to equilibrate to room temperature for 10 min before addition of 200 µl 0·4% o-nitrophenyl galactopyranoside (gut operon expression) or 100 µl 10 mM p-nitrophenyl-ß-D-glucuronide (fdo operon expression) to give a final volume of 1·0 ml. The nitrophenol released was measured after addition of 400 µl 1 M Na₂CO₃ (gut operon expression) or 400 µl 2·5 M 2-amino-2-methylpropanediol (fdo operon expression) to stop the reaction. The samples were then read at 420 or 415 nm, respectively (Jefferson et al., 1986 ; Miller, 1972 ). Protein concentrations were measured by the Lowry method.

Cell extracts for the assay of glucitol-6-phosphate dehydrogenase, Enzyme II^Gut and D-alanine dehydrogenase activities were prepared as follows. Cells were grown with shaking in LB broth at 37 °C, harvested by centrifugation, washed three times with minimal medium 63 and resuspended in the same medium. PMSF (0·1 mM) and ß-mercaptoethanol (0·5 mg ml^-1) were added to the samples and the cells were lysed using a French press at 10000 p.s.i. (69 MPa). After centrifugation to remove cell debris, the supernatant was used for the assays. For the assay of glucitol-6-phosphate dehydrogenase activity, aliquots of the cell extracts were added to a solution containing 0·1 M Tris/HCl (pH 8·5), 1·5 mM NAD⁺, 1 mM dithiothreitol and 2 mM glucitol 6-phosphate. The optical density at 340 nm was followed over a period of 2 min. For the assay of Enzyme II^Gut, 0, 1, 2, 5, 10 and 25 µl samples of extract were added to 100 µl twofold concentrated assay mix containing 100 mM potassium phosphate buffer, pH 6·0, 20 µM [¹⁴C]D-glucitol, 20 mM glucitol 6-phosphate, 50 mM KF, 25 mM MgCl₂ and 5 mM dithiothreitol. The samples were brought to 200 µl, final volume, vortexed and incubated at 37 °C for 40 min. One millilitre ice-cold water was then added to each assay tube to stop the reaction. The samples were applied to NaCl pre-washed ion-exchange columns (Saier et al., 1977 ), the columns were washed twice with 12 ml water and the samples were eluted with 12 ml 1 M LiCl into scintillation vials containing Biosafe II scintillation cocktail in preparation for scintillation counting. For the D-alanine dehydrogenase assay, aliquots of cell extracts were added to a solution containing 2 mM D-alanine, 10 mM potassium phosphate, pH 7·5, and 0·33 mg dinitrophenyl hydrazine ml^-1. At time intervals of 10 min, the reaction was stopped by addition of 1·67 ml 10% NaOH and the absorbance was read at 445 nm as described by Wild et al. (1974 ; see also Saier et al., 1980 ).

Transport assays.
Cells were grown with shaking at 37 °C in LB medium, harvested during exponential growth by centrifugation, washed three times with minimal medium 63 and resuspended in the same medium. The cell suspensions were pre-warmed at 37 °C for 10 min before ¹⁴C-labelled glucitol or D-alanine was added to a final concentration of 0·1 mM. At intervals, samples were transferred to pre-washed 0·2 µm (pore diameter) millipore filters. Cells were washed three times with medium 63 before the filters were dried, transferred to scintillation vials containing Biosafe-NA scintillation cocktail and counted.

cAMP assays.
Cultures were grown for net cAMP production in minimal salts medium containing 1% lactate or 1% lactate plus 0·5% glucose for 5 h before samples were removed for extraction of cAMP (Castro et al., 1976 ). The optical density of the cultures (600 nm) was approximately 0·6. Samples were placed in a boiling water bath for 5 min, chilled to 0 °C and centrifuged to remove cell debris. Total cAMP was determined using partially purified cAMP-binding protein from beef muscle (Gilman, 1970 ), essentially as described previously (Castro et al., 1976 ; Feucht & Saier, 1980 ). The amount of cAMP produced (Table 6) is expressed in nmol (mg dry wt)^-1. No background activity was subtracted from the values presented.

Effects of cysB mutations on carbon utilization in E. coli and S. typhimurium
Table 2 records the Biolog results observed for cysB mutants of both E. coli and S. typhimurium relative to their isogenic wild-type (cysB⁺) strains. The results observed for the S. typhimurium cysB mutant generally (but not always) paralleled those obtained with the E. coli mutant. Table 2 records the most pronounced effects observed. Clear-cut parallel effects were observed for D-alanine, L-fucose, D-glucitol, glucuronamide, α-hydroxyglutaric acid γ-lactone and L-rhamnose. In two cases (D-glucuronate and D-gluconate), the S. typhimurium cysB mutant exhibited decreased utilization relative to its parental strain, but the E. coli mutant did not. In two cases (glycolate and glyoxylate), the E. coli cysB mutant exhibited decreased utilization relative to its wild-type strain, but the wild-type S. typhimurium strain did not exhibit a positive response towards either compound. In one case (formate), opposite responses were observed for S. typhimurium and E. coli cysB mutants (increased utilization in the former strain; decreased utilization for the latter strain). Although the catalytic repertoires for using carbon sources are similar in E. coli and Salmonella species, significant differences are found even among the strains and substrains of each species (Lin, 1996 ).

Table 2. Carbon oxidation by wild-type and cysB mutant strains of S. typhimurium and E. coli

Fermentation properties of cysB mutants of S. typhimurium and E. coli
Table 3 summarizes the results of carbohydrate fermentation responses of wild-type and cysB mutant strains to select compounds. The cysB mutants of S. typhimurium and E. coli generally did not ferment carbohydrates as well as did the isogenic parents, and for both bacterial species, responses towards L-rhamnose and D-glucitol appeared to be most pronounced. On minimal medium, both cysB mutants exhibited negligible growth regardless of the carbon source present, as expected, since the cysB mutation gives rise to a requirement for exogenous cysteine or another sulfur source (data not shown). In all cases examined, inclusion of cysteine or djenkolate in the media partially, but not completely, reversed the growth phenotype. Even high concentrations of these compounds did not restore wild-type growth responses (data not shown).

Table 3. Fermentation properties of cysB mutants of S. typhimurium and E. coli on solid medium

Glucitol uptake in S. typhimurium
Fig. 1 shows time courses for the uptake of [¹⁴C]D-glucitol by S. typhimurium strain LT-2 and by the isogenic cysB mutant following growth in LB broth (moderately sulfur-limiting conditions comparable to our fermentation conditions). Uptake was reduced nearly tenfold by loss of CysB. When the cysB mutant was grown in the presence of glucitol plus djenkolate, uptake was partially restored. These results are in full agreement with the fermentation results recorded in Table 3.

(16K): Fig. 1. Uptake of [14C]glucitol by wild-type cells () and by cysB mutant cells grown in LB broth plus 0·5% glucitol with () or without () djenkolate at a concentration of 1 mg ml-1. Cells were harvested 3 h after inoculation during exponential growth as described under Methods.

Effects of cysB mutations on the specific activities of enzymes and transport systems concerned with glucitol and D-alanine utilization
Table 4 summarizes the results of studies aimed at quantifying the specific activities of transport systems and catabolic enzymes concerned with glucitol and D-alanine utilization in S. typhimurium. Glucitol is utilized via a pathway that involves concomitant transport and phosphorylation of the sugar, yielding cytoplasmic glucitol 6-phosphate, followed by oxidation of glucitol 6-phosphate to fructose 6-phosphate (Lengeler, 1975a , b ). The Enzyme II^Gut (glucitol PTS permease) and glucitol-6-phosphate dehydrogenase were therefore assayed in crude extracts derived from wild-type and cysB mutant cells of S. typhimurium. As summarized in Table 4, both activities were depressed to about 50% of the wild-type level by the cysB mutation, and the inclusion of djenkolate in the growth medium partially restored these activities. Inclusion of cysteine had an effect similar to that of djenkolate (data not shown). These results are in agreement with the fermentation and transport results cited above.

Table 4. Effects of a cysB mutation on the specific activities of transport systems and enzymes concerned with glucitol and alanine utilization in S. typhimurium

When D-alanine uptake and D-alanine dehydrogenase were similarly assayed, qualitatively similar results were obtained. Thus, D-alanine transport activity was reduced to 70% whilst D-alanine dehydrogenase activity decreased to 9%. The dramatic effect of the cysB mutation on D-alanine dehydrogenase activity can account for the poor response to alanine as measured with Biolog plates.

Use of reporter gene transcriptional fusions to study the effects of a cysB mutation on gene expression
A plasmid bearing a gutBlacZ transcriptional fusion was transformed into wild-type and cysB mutant strains of E. coli. Cells bearing this multicopy plasmid were then grown in nutrient medium under non-inducing, repressing and inducing conditions as described in Methods. The lack of induction by glucitol presumably reflects the multiple copies of the gut operon control region since the chromosomally encoded regulatory proteins, GutM and GutR (Yamada & Saier, 1988 ), may not be present in sufficient amounts to regulate all plasmid operators. The results depicted in Fig. 2(a) clearly show that when cells were grown under either non-inducing conditions (NB alone) or inducing conditions (NB plus 1% glucitol), the presence of the cysB mutation drastically reduced glucitol operon expression (Yamada & Saier, 1987 , 1988 ; Yamada et al., 1990 ). CRP is present in E. coli cells in large amounts. Consequently, catabolite repression would not be expected to be altered by expression of a target gene on a multicopy plasmid. As can be seen, strong catabolite repression is observed. Unexpectedly, however, there was little effect of the cysB mutation under glucose-repressing conditions.

(23K): Fig. 2. Transcription of the glucitol (gut, a) and formate dehydrogenase (fdo, b) operons in wild-type (WT) and cysB mutant strains of E. coli. (a) A plasmid-encoded gutBlacZ fusion was introduced into the two strains by electroporation as described in Methods. Cells were grown in NB with or without glucose (Glc; 1%) or glucitol (Gut; 1%) at 37 °C for 3 h before harvesting, cell rupture and assay of ß-galactosidase. (b) A plasmid-encoded fdoGuidA fusion was similarly introduced into the two strains by electroporation. The transformed cells were grown in nutrient broth with or without formate (Form) or glucose (Glc) (1% each) for 3 h before harvesting, cell rupture and assay of ß-glucuronidase as described in Methods. White boxes at the top of the dark bars denote the error expressed in standard deviations with the measured mean values denoted by the dark bars. The lack of induction reflects the use of high-copy-number plasmids.

Comparable experiments were conducted with a transcriptional reporter gene fusion to the fdoGHI (aerobic formate dehydrogenase) operon. In this fdoGuidA fusion, the reporter gene encodes ß-glucuronidase (Abaibou et al., 1995 ) (Fig. 2b). Results paralleled those observed for the glucitol operon gene fusion. Thus, induced and uninduced activities were comparable, and unrepressed activity (i.e. in NB-grown cells) was reduced about twofold by the cysB mutation. This activity was largely restored by inclusion of cysteine in the growth medium. By contrast, glucose-repressed activity was only weakly affected by the cysB mutation. These results led to the possibility that the cysB mutation prevented the normal cAMPCRP-mediated activation of catabolite-sensitive operons under non-repressing conditions. Activity under non-repressing conditions was thus affected by the cysB mutation much more than that under repressing conditions.

Effect of exogenous cAMP on gene expression
The results reported in Fig. 2 suggested that the cysB defect might interfere with the function of the cAMPCRP complex that normally activates catabolite-sensitive genes. To test this possibility, expression of the two reporter gene fusions (gutBlacZ and fdoGuidA) was studied as a function of the concentration of exogenous cAMP. The results are summarized in Fig. 3. As can be seen, cAMP increased expression of both fusions in the cysB genetic background, but decreased expression of both fusions in the wild-type background. The mild repressive effect in the wild-type background suggests that, in the absence of exogenous cAMP, these cells synthesize sufficient cAMP for maximal transcription. The results suggest that a primary deficiency, giving rise to defective expression of carbon catabolic enzyme-encoding genes, is related to depressed levels of cytoplasmic cAMP in the cysB mutant relative to the wild-type strain.

(14K): Fig. 3. Effects of cAMP on expression of the genes encoding Enzyme IIA of the glucitol PTS (a) and formate dehydrogenase (b) of E. coli. Cells (WT, ; cysB, ) were grown at 37 °C in nutrient broth for 3 h in the presence of the concentration of exogenous cAMP indicated. ß-Galactosidase (a) or ß-glucuronidase (b) was assayed as described in Methods. Five millimolar cAMP depressed growth rates of the wild-type strain, but not of the cysB mutant.

Control of lac operon expression in response to sulfur limitation
Table 5 presents the results of experiments designed to test the effects of sulfur limitation on native ß-galactosidase synthesis when lacZ is expressed under the control of the lac operon promoter and control region. These experiments were conducted in the genetic background of Crookes strain (SB2249; cpd; Castro et al., 1976 ). Since this strain lacks cAMP phosphodiesterase, cAMP degradation cannot occur. This leads to higher cytoplasmic levels of this nucleotide. In the wild-type strain, ß-galactosidase activity was high when cells were grown in 1% lactate-containing minimal medium and substantially lower when grown in the presence of 0·5% glucose. Under these conditions, the presence of djenkolate or cysteine had little effect on lacZ expression, although expression was slightly higher when cysteine (0·2 mM) was present as compared with djenkolate (1 mM). These results are consistent with the conclusion that in cysB⁺ cells, cAMP is present at high cytoplasmic concentrations and glucose only moderately reduces these concentrations. However, in the isogenic cysB mutant, lactate-grown cells showed substantially less ß-galactosidase activity than the wild-type strain and the difference between cysteine- and djenkolate-grown cells was somewhat greater. Growth in the presence of djenkolate is known to promote sulfur limitation in a cysB mutant to a greater extent than growth in the presence of cysteine (Kredich, 1996 ). Most striking, however, was the effect of sulfur limitation on the cysB mutant when grown in the presence of glucose plus djenkolate. ß-Galactosidase activity dropped to about 20% of the value obtained when the wild-type strain was grown with glucose with or without either cysteine or djenkolate. Activity was also high when the cysB mutant was grown with glucose plus cysteine. These results show that lacZ expression is greatly reduced only in the cysB mutant under conditions of sulfur starvation and substantiate the suggestion that sulfur limitation lowers cellular cAMP levels.

Table 5. ß-Galactosidase levels in Crookes strain and the isogenic cysB mutant strain

Direct measurement of cAMP production
Total cAMP production (cells plus medium) was quantified in the same isogenic cpd and cpd cysB strains used in the experiments described in the previous section (Table 6). In this experiment, cells were grown in the same minimal medium supplemented with lactate or glucose plus djenkolate. Because Crookes strain (SB2249) lacks the cAMP-degradative enzyme, cAMP phosphodiesterase, it is clear that the effects are due to the regulation of cAMP synthesis, not cAMP degradation. The cysB mutant showed slightly lower cAMP production than the isogenic cysB⁺ strain when lactate was the sole carbon source present. In the presence of glucose, however, cAMP production was greatly reduced and the cysB mutation reduced production even further. The presence of the ptsI crr double mutations reduced cAMP production during growth on lactate, but increased production in the presence of glucose. In this case, glucose may have increased cAMP production by increasing energy availability and thereby increasing the level of ATP in the cell. crr mutants are insensitive to PTS-mediated regulation of adenylate cyclase (Castro et al., 1976 ; Feucht & Saier, 1980 ). In the crr genetic background, in the presence of glucose, the cysB mutation did not reduce cAMP production levels to those observed for glucose-grown LJ4508 (cysB). The results are therefore in qualitative agreement with the results presented in Table 5, assuming that cellular cAMP levels explain the results reported therein. The data suggest, but are insufficient to establish, that CysB exerts its effect on adenylate cyclase by influencing the PTS-regulatory apparatus rather than directly affecting adenylate cyclase itself.

Table 6. cAMP levels in strains grown on lactate and djenkolate

Mutations affecting pleiotropic transcription factors controlling nitrogen, sulfur, phosphorus or iron also have a negative effect on carbon utilization (Hantke, 1987 ; Hartmann & Boos, 1993 ; Shi & Bennett, 1994 ; and our unpublished results). Thus, for example, in E. coli, using Biolog plates, a phoB mutant showed decreased glucose, maltose, melibiose and trehalose utilization, a glnB mutant exhibited overall weaker responses than did the parental strain and a fur mutant showed poor utilization of proline, L-fucose, D,L-α-glycerol phosphate, formate, L-rhamnose, fumarate and glycyl-L-aspartate. In some cases, these effects might be indirect, due to a deficiency for one or more specific compound(s) synthesized or secured by the cell. Such a compound might be required for carbon metabolic processes to proceed normally. Such indirect effects may be of great physiological significance, allowing the bacterium to coordinate the utilization of one nutrient class of compounds with another. We anticipate that such regulatory effects have been selected for during evolution. However, very few studies have addressed the mechanisms by which these regulatory interactions occur.

Recently, evidence for a connection between the cAMPCRP-regulated modulon and the stringent response in E. coli has been reported (Johansson et al., 2000 ). Moreover, the homologous nucleoid proteins, H-NS and StpA, were shown to play a positive role in expression of stringently regulated genes as well as in cell growth. The results reported suggest that catabolite repression, the stringent response and control of gene expression by structural nucleoid proteins are all interconnected (Johansson et al., 2000 ). The relationship between these observations and those reported here, if any, have yet to be investigated, but the results of Johansson et al. (2000) corroborate our postulate that the different regulons in the bacterial cell are interrelated.

In this communication, we not only document the effects of cysB mutations in E. coli and S. typhimurium on carbon oxidation and carbohydrate fermentation, but we also pursue the effect of the cysB mutation on carbon utilization to the mechanistic level. The results reported revealed that the effects were at the transcriptional level for both the gut and the fdo operons. Both effects appear to be partially reversed by exogenous cAMP or a sulfur source such as cysteine or djenkolate. Furthermore, we provide evidence that the effect is on regulation of the cAMP biosynthetic enzyme, adenylate cyclase, by the IIA^Glc protein of the PTS, and not on the inherent activity of adenylate cyclase itself. Our current postulate is that these effects are indirect, that they affect the regulation of adenylate cyclase as a primary target and that the level and/or activity of this enzyme will prove to be hypersensitive to the metabolic state of the cell. Such sensitivity would allow the bacterium to coordinate carbon/energy metabolism with the availability of other macro- and micronutrients. The mechanistic details of these interactions are currently under study in our laboratory.

We thank Jackie Richardson and Mary Beth Hiller for assistance in the preparation of this manuscript. This work was supported by USPHS grant AI14176 from the National Institute of Allergy and Infectious Diseases.