Catabolite Repression Control of napF (Periplasmic Nitrate Reductase) Operon Expression in Escherichia coli K-12

Abstract

Escherichia coli, a facultative aerobe, expresses two distinct respiratory nitrate reductases. The periplasmic NapABC enzyme likely functions during growth in nitrate-limited environments, whereas the membrane-bound NarGHI enzyme functions during growth in nitrate-rich environments. Maximal expression of the napFDAGHBC operon encoding periplasmic nitrate reductase results from synergistic transcription activation by the Fnr and phospho-NarP proteins, acting in response to anaerobiosis and nitrate or nitrite, respectively. Here, we report that, during anaerobic growth with no added nitrate, less-preferred carbon sources stimulated napF operon expression by as much as fourfold relative to glucose. Deletion analysis identified a cyclic AMP receptor protein (Crp) binding site upstream of the NarP and Fnr sites as being required for this stimulation. The napD and nrfA operon control regions from Shewanella spp. also have apparent Crp and Fnr sites, and expression from the Shewanella oneidensis nrfA control region cloned in E. coli was subject to catabolite repression. In contrast, the carbon source had relatively little effect on expression of the narGHJI operon encoding membrane-bound nitrate reductase under any growth condition tested. Carbon source oxidation state had no influence on synthesis of either nitrate reductase. The results suggest that the Fnr and Crp proteins may act synergistically to enhance NapABC synthesis during growth with poor carbon sources to help obtain energy from low levels of nitrate.

As a facultative aerobe, Escherichia coli can use nitrate (NO₃⁻) and several other compounds as terminal oxidants for anaerobic respiration (28, 74). Nitrate respiration occurs through either of two enzymes. Membrane-bound nitrate reductase (NarGHI enzyme), encoded by the narGHJI operon, functions in respiration by coupling nitrate reduction directly to proton motive force. Periplasmic nitrate reductase (NapABC enzyme), encoded by the napFDAGHBC operon, functions indirectly in respiration and also participates in dissimilatory processes such as redox balancing (12, 46, 50, 67).

Transcription of the narG and napF operons is activated during anaerobic growth by the oxygen-responsive Fnr protein (36) and is controlled further by the nitrate- and nitrite-responsive NarX-NarL and NarQ-NarP two-component regulatory systems (69). Fnr binding to specific DNA sites requires formation of its oxygen-labile iron-sulfur cluster. The NarL and NarP response regulators bind specific DNA sites upon phosphorylation by the NarX and NarQ sensors. Phospho-NarL and -NarP activate transcription of the narG and napF operons, respectively, whereas phospho-NarL antagonizes napF operon transcription (18, 22).

In continuous cultures, napF operon transcription peaks at a relatively low nitrate concentration, 1 mM, whereas maximal narG operon transcription requires at least 8 mM nitrate (77). Likewise, Nap⁺ strains outcompete Nar⁺ strains in nitrate-limited continuous cultures (51). This suggests that NapABC and NarGHI enzymes function in nitrate-limited and -replete environments, respectively (50).

Regulation of narG operon transcription results from activation of a single promoter by both Fnr and phospho-NarL proteins (37, 62, 75). Integration host factor (IhfAB) is essential for full expression (52, 58, 76). No other regulators are known to modulate narG operon transcription directly. In contrast, regulation of napF operon transcription is relatively complex. Transcription from promoter P1 is activated by Fnr and phospho-NarP during anerobic growth with nitrate, whereas transcription from the overlapping promoter P2 occurs during aerobic growth or during anaerobic growth in the absence of nitrate (19, 22, 23, 66). Transcription from the upstream promoter P3 is repressed during normal growth conditions by the iron-sulfur cluster assembly regulator IscR (29). Finally, napF operon transcription is also controlled by the molybdate-responsive regulator ModE (43, 66) and by the nitric oxide-responsive repressor NsrR (27).

Carbon catabolite repression results from several interacting mechanisms whereby the presence of a preferred carbon source inhibits the metabolism of less-favored carbon sources (25, 49, 78). One well-studied mechanism involves transcription activation by the cyclic AMP (cAMP) receptor protein (Crp), also termed the catabolite gene activator protein (Cap) (15, 38). cAMP functions as an allosteric effector to increase the affinity of Crp for its DNA binding sites. Preferred carbon sources (such as glucose) inhibit adenylate cyclase activity by virtue of their transport through the phosphoenolpyruvate:carbohydrate phosphotransferase system (PTS) (25). Certain non-PTS sugars such as gluconate also cause catabolite repression (32). Catabolite repression is relaxed in anaerobic cultures but is imposed in nitrate-respiring cultures (21, 26).

For this study, we examined the influence of catabolite repression on nitrate reductase synthesis. We found little influence on narG operon expression but a clear effect on napF operon expression. While comparing napF operon control region sequences from different species, we noted a potential Crp binding site (Fig. 1). This site was identified independently in a genome-wide search (13). The results indicate that Crp activates napF operon expression and suggest a physiological role for stimulating NapABC enzyme synthesis under energy-limited growth conditions.

FIG. 1.

Transcription control region for enterobacterial napF operons. Sequences are for E. coli K-12, C. rodentium ICC168, and S. enterica LT2. A portion of the sequence alignment is duplicated in order to show the promoter P2 elements separately. The experimentally determined transcription initiation sites, denoted as T1, T2, and T3, are shown for the E. coli sequence (19, 29, 66). Nucleotides that match the consensus sequences are shaded black (promoter −10 and −35 elements) or gray (NarP, Fnr, and Crp) or are enclosed within boxes (NsrR and ModE). Y, C or T; M, A or C; K, G or T; and W, A or T. Binding site centers are measured from the T1 transcription initiation site. The region protected by IscR protein (29) is indicated. Deletion endpoints (22) and site-specific alterations in the promoter −10 elements (66) are shown.

MATERIALS AND METHODS

Strains.

Table 1 lists the strains used in this study. Strains were constructed by generalized transduction with bacteriophage P1 (45). The Δcya-854 allele (11, 30) was introduced in two steps: (i) an ilv::Tn10 insertion was brought in by selection for resistance to tetracycline, and (ii) the cya allele was brought in by cotransduction with ilv⁺. The Δcrp-45 allele (55) was introduced by cotransduction with rpsL.

TABLE 1.

E. coli K-12 strains used in this study

Strain	Genotype	Source or reference
CA8306	HfrH thi-1 Δcya-854	11
EC8445	HfrH thi-1 Δcya-854 Δcrp-45 rpsL136	55
VJS632	F- λ- prototroph	68
VJS676	As VJS632 but Δ(argF-lacIZYA)U169	68
VJS711	As VJS632 but Δcya-854	This study
Derivatives of strain VJS676
VJS3861	λΦ(nrfA-lacZ)	67
VJS6906	λΦ(napFHi-lacZ) [Δ260]	65
VJS4800	λΦ(napF-lacZ) [Δ123] narL215::Tn10	22
VJS5101	λΦ(napF-lacZ) [Δ85] narL215::Tn10	22
VJS5509	λΦ(napF-lacZ) [Δ146-G69A] narL215::Tn10	22
VJS5516	λΦ(napF-lacZ) [Δ146] narL215::Tn10	22
VJS7443	λΦ(napF-lacZ) [Δ146 P1-] narL215::Tn10	66
VJS7444	λΦ(napF-lacZ) [Δ146 P2-] narL215::Tn10	66
VJS9439	λΦ(napF-lacZ) [Δ123] Δ(narXL)265::Km	This study
VJS9440	λΦ(napF-lacZ) [Δ85] Δ(narXL)265::Km	This study
VJS9447	λΦ(nrfA-lacZ) Δ(narXL)265::Km	This study
VJS9448	λΦ(napFHi-lacZ) [Δ260] Δ(narXL)265::Km	This study
VJS9451	λΦ(napF-lacZ) [Δ123] Δ(narXL)265::Km Δcya-854	This study
VJS9452	λΦ(napF-lacZ) [Δ85] Δ(narXL)265::Km Δcya-854	This study
VJS9453	λ- Δ(attλ-lom)::bla [Φ(nrfASo-lacZ)] Δ(narXL)265::Km	This study
VJS11436	λΦ(napF-lacZ) [Δ146] narL215::Tn10 Δcya-854	This study
VJS11437	λΦ(napF-lacZ) [Δ146 P1-] narL215::Tn10 Δcya-854	This study
VJS11438	λΦ(napF-lacZ) [Δ146; P2-] narL215::Tn10 Δcya-854	This study
VJS11439	λΦ(napF-lacZ) [Δ146] narL215::Tn10 fnr-275::Km	This study
VJS11440	λΦ(napF-lacZ) [Δ146 P1-] narL215::Tn10 fnr-275::Km	This study
VJS11441	λΦ(napF-lacZ) [Δ146 P2-] narL215::Tn10 fnr-275::Km	This study
VJS11445	λΦ(napF-lacZ) [Δ146] narL215::Tn10 Δcrp-45 rpsL136	This study
VJS11446	λΦ(napF-lacZ) [Δ146 P1-] narL215::Tn10 Δcrp-45 rpsL136	This study
VJS11447	λΦ(napF-lacZ) [Δ146 P2-] narL215::Tn10 Δcrp-45 rpsL136	This study

Experiments to monitor napF and nrfA expression employed strains with either an narXL deletion or a narL insertion, because phospho-NarL inhibits expression of these operons during batch culture with excess nitrate (22, 73). The narL215::Tn10 insertion has been described (62). The Δ(narXL)265::Km insertion was constructed by λ Red-mediated recombineering (24), using PCR primers LLC1285 and LLC1289 (5′-ATGGCGATGCTTGGAACTGCGTTGAACAATATGTCTattccggggatccgtcgacc and 5′-CATTTTCTTCAGCATGTGCTTGACGTGCACTTTTACgtgtaggctggagctgcttc) with plasmid pKD4 (Km) (24) as the template. (Sequence complementary to pKD4 is shown in lowercase.) This results in deletion from narX codon Ala-221 through narL codon Thr-186.

The Δfnr-275::Km insertion was constructed by using primers RS2203 and RS2204 (5′-AATTATACGGCGCATTCAGTCTGGCGGTTGTGCTATgtgtaggctggagctgcttc and 5′-GATGTATTTACCTTTGACTGCCAGCATGCCGCTTTTattccggggatccgtcgacc). This resuts in deletion of fnr codons His-19 through Gln-219.

Gene and operon fusion constructs (Table 1) are carried on monocopy bacteriophage λ specialized transducing phage (61) integrated at attλ. We used PCR amplification to sequence the control regions from our permanent stock cultures of the Φ(napF-lacZ) [Δ123] and Φ(napF-lacZ) [Δ85] strains (22) to confirm that they have the indicated deletion endpoints.

The Φ(nrfA_So-lacZ) gene fusion was constructed essentially as described previously (39). Briefly, whole-colony PCR was used to clone the nrfA control region from S. oneidensis (nrfA_So) MR-1 by using primers PJB2024 and PJB2025 (5′-GAGGTAAGACTGCCTgaattcATATGTCGAATACCTTGTGG and 5′-CACTTAATGCAAAAGTCggatccGTCATCTTCTTCATCATC). These introduce an EcoRI site overlapping codons 1 to 3 of the divergently transcribed narQ gene and a BamHI site overlapping codons 6 to 8 of the nrfA gene. (Restriction sites are shown in lowercase.) The resulting gene fusion construct was integrated into the chromosome as described previously (8, 64).

Culture media and conditions.

Defined, complex, and indicator media for genetic manipulations were used as described previously (42). Defined medium (47) to grow cultures for enzyme assays was buffered with 3-[N-morpholino] propanesulfonic acid (MOPS) as described previously (68). The initial pH of this medium was set at 8.0 to ameliorate nitrite toxicity. Because the pK_a of MOPS is 7.2, the buffering capacity of this medium continually increases as acidic fermentation products accumulate; at harvest, cultures typically had a pH value of about 7.5. Carbon sources, acid-hydrolyzed casein, tryptone, and yeast extract were added as indicated. The respiratory oxidants NaNO₃ and sodium fumarate were added at 40 mM, and trimethylamine N-oxide was added at 20 mM, as indicated. 3′,5′-cAMP and isopropyl-β-d-thiogalactopyranoside (IPTG) were added at 5 mM and 1 mM, respectively, as indicated.

Cultures were grown at 37°C to the early exponential phase, about 25 to 35 Klett units. Culture densities were monitored with a Klett-Summerson photoelectric colorimeter (Klett Manufacturing Co., New York, NY) equipped with a no. 66 (red) filter. Aerated and anaerobic cultures were grown and harvested as described previously (68).

Carbon sources.

We reasoned that different carbon sources might influence respiratory enzyme synthesis through catabolite repression, catabolic pathway, oxidation state, or combinations thereof. Accordingly, we examined a variety of carbon sources.

Glucose and mannitol are both class A carbon sources that elicit effective catabolite repression, whereas mannose and glucitol (sorbitol) are both class B carbon sources that do not (49). Although not a PTS substrate, gluconate also elicits effective catabolite repression (32). Additionally, catabolite repression is enhanced in medium supplemented with a source of amino acids (26, 78), so some experiments included either acid-hydrolyzed casein or tryptone plus yeast extract.

Hexitols and hexoses are metabolized by the Embden-Meyerhof-Parnas pathway, whereas sugar acids are metabolized by the Entner-Doudoroff pathway (48). Glycerol catabolism requires either aerobic or anaerobic glycerol-3-phosphate dehydrogenase, both of which are coupled to respiration, whereas pyruvate, the ultimate product of glycerol catabolism, is fermented. Thus, growth with glycerol as primary carbon source compels the organism to deploy its respiratory metabolism (28, 74).

The hexitols (glucitol and mannitol) are more reduced than the corresponding hexoses (glucose and mannose), whereas the sugar acids gluconate and glucuronate are progressively more oxidized. During anaerobic growth, the mixed-acid fermentation maintains redox balance by generating relatively higher concentrations of reduced end products from more reduced carbon sources and relatively higher concentrations of oxidized end products from more oxidized carbon sources (1, 20). Growth with respiratory oxidants influences redox balance by providing alternate routes for NADH oxidation.

Enzyme assays.

Enzyme specific activities were determined at room temperature (approximately 21°C). Washed cell pellets were stored overnight at −20°.

Reduced methyl viologen-nitrate reductase activity was measured in cell extracts by monitoring production of nitrite from nitrate (40). The β-galactosidase activities reported in Table 2 were measured in cell extracts by monitoring production of o-nitrophenol from o-nitrophenyl-β-d-galactopyranoside (45). Cells were suspended in 4 ml of cold potassium phosphate (0.32 M; pH 7.1) and ruptured by passage through a French pressure cell (SLM Aminco, Inc., Urbana, IL) at 20,000 lb/in². Unbroken cells were removed by sedimentation at 3,000 × g for 5 min. Protein concentrations were estimated by the procedure of Bradford (9). Units are expressed as μmol product formed min⁻¹ mg⁻¹. Cultures were assayed in duplicate, and reported values are averaged from three to five independent experiments.

TABLE 2.

Effects of carbon source, culture aeration, and nitrate on NarG and LacZ synthesis

Carbon sourcea	Casein additionb	Sp act
		NarGc			LacZd
		+O2-NO3−	-O2 +NO3−	-O2-NO3−	+O2-NO3−	-O2 +NO3−	-O2-NO3−
Glucose	-	0.005	1.4	0.034	2.8	2.6	4.0
Mannitol	-	0.015	1.7	0.035	5.7	4.4	7.0
Glucitol	-	0.016	1.8	0.037	7.0	7.9	11
Mannose	-	0.017	1.8	0.030	9.1	8.4	10
Glucose	+	0.005	1.3	0.021	1.9	2.0	3.2
Glucitol	+	0.014	1.1	0.026	4.5	4.7	5.4
Pyruvate	+	0.012	1.0	0.028	5.2	6.0	7.1
Glycerole	+	0.013	1.2	0.027	5.2	5.9	7.5

^aStrain VJS632 (nar⁺ lac⁺) was cultured areobically or anaerobically to the early exponential phase in MOPS defined medium. The indicated carbon source was added (40 mM).

^bAcid-hydrolyzed casein (0.5%) was added as indicated.

^cViologen-dependent activity in cell extracts, expressed as μmol nitrite formed min⁻¹ mg⁻¹.

^dActivity in cell extracts, expressed as μmol o-nitrophenol formed min⁻¹ mg⁻¹.

^eFumarate added as electron acceptor.

β-Galactosidase activities reported in Tables 3 to 8 were measured in CHCl₃-sodium dodecyl sulfate-permeabilized cells. Specific activities are expressed in arbitrary (Miller) units (45). Cultures were assayed in duplicate, and reported values are averaged from at least two independent experiments.

TABLE 3.

Effects of carbon source and nitrate on Φ(napF-lacZ) expression

Carbon sourcea	Redoxb	LacZ sp act (Miller units)c		Expression (fold) relative to glucose
		-NO3−	+NO3−	-NO3−	+NO3−
Glucose	+4	260	6,860	1.0	1.0
Gluconate	+2	410	6,180	1.6	0.9
Glucuronate	0	470	10,750	1.8	1.6
Mannitol	+6	500	9,880	1.9	1.4
Glucitol	+6	630	11,030	2.4	1.6
Mannose	+4	1,110	14,220	4.3	2.1

^aThe indicated carbon source was added (40 mM).

^bRedox state as H produced per C₆ (20).

^cStrain VJS5516 was cultured anaerobically to the early exponential phase in MOPS defined medium. The genotype was narQ⁺ narP⁺ narX⁺ narL215::Tn10 Φ(napF-lacZ) [Δ146] (Table 1).

TABLE 8.

Effects of carbon source, complex medium, and nitrate on reporter expression

Straina	Fusionb	TY additionc	LacZ sp act (Miller units)d				Catabolite repression (fold)
			-NO3−		+NO3−
			Glycerol	Glucose	Glycerol	Glucose	-NO3−	+NO3−
VJS9447	nrfAEc	-	29	4	480	86	7.2	5.6
VJS9447	nrfAEc	+	21	3	250	50	7.0	5.0
VJS9439	Δ123	-	2,510	520	14,030	8,730	4.8	1.6
VJS9439	Δ123	+	1,380	620	12,700	3,980	2.2	3.2
VJS9440	Δ85	-	330	350	12,020	8,010	0.9	1.5
VJS9440	Δ85	+	540	370	9,450	2,830	1.5	3.3
VJS9448	napFHi	-	55	29	370	260	1.9	1.4
VJS9448	napFHi	+	110	64	850	420	1.7	2.0
VJS9453	nrfASo	-	70	29	740	170	2.4	4.4
VJS9453	nrfASo	+	83	28	870	180	3.0	4.8

^aGenotype: narQ⁺ narP⁺ Δ(narXL)265::Km (Table 1).

^bIndicated lacZ fusion at attλ.

^cTY (1% tryptone plus 0.5% yeast extract) was added as indicated.

^dStrains were cultured anaerobically to the early exponential phase in MOPS defined medium plus TMAO. Glycerol (80 mM) or glucose (40 mM) was added as indicated.

Genome database searches.

Analyses employed the BLAST programs (2) at the National Center for Biotechnology Information (http://www.ncbi.nlm.nih.gov). Draft genome sequence data for Citrobacter rodentium ICC168 were produced by the Bacterial Genomes Sequencing Group at the Sanger Institute (http://www.sanger.ac.uk). Completed genome sequence data were accessed through GenBank for E. coli K-12 (GenBank accession no. NC_000913), Salmonella enterica LT2 (NC_009137), Shewanella oneidensis MR-1 (NC_004347), Shewanella sp. strain MR-4 (NC_008321), Shewanella amazonensis SB2B (NC_008700), and Shewanella loihica PV-4 (NC_009092).

RESULTS

Experiments to examine catabolite repression were motivated by two goals: (i) to investigate the influence of carbon source oxidation state on anaerobic respiratory enzyme synthesis (20) and (ii) to characterize further the influence of nitrate respiration on catabolite repression in anaerobic cultures (21, 26). Initial experiments examined NarG and β-galactosidase (LacZ) synthesis in order to establish relative responses to different culture conditions. We then focused attention on napF operon expression.

Expression of the narG and lacZ operons.

During mid-exponential-phase growth in defined medium, NarG synthesis was induced efficiently by nitrate irrespective of carbon source (Table 2). However, there were subtle differences in the induced level of expression, with the pattern glucose < mannitol < glucitol ≅ mannose. These results are not consistent with the idea that more reduced carbon sources lead to higher levels of NarG synthesis, but they may indicate weak catabolite repression.

IPTG-induced lacZ operon expression followed the same general pattern (Table 2). However, LacZ activities varied by approximately 3-fold in the different media, whereas NarG activities varied by less than 1.3-fold. In some experiments, we also measured induced levels of tryptophanase (TnaA), an enzyme whose synthesis is quite sensitive to catabolite repression (7). The relative patterns of TnaA synthesis paralleled those for LacZ synthesis (data not shown).

Table 2 also shows that LacZ activities were comparable in aerated and nitrate-respiring cultures, whereas they were higher in anaerobic cultures, reflecting the anaerobic relaxation of catabolite repression (21, 26).

Finally, enzyme synthesis was also measured from cultures supplemented with acid-hydrolyzed casein (Table 2). LacZ activities relative to those in defined media were lower, reflecting enhanced catabolite repression (26, 78). The relatively subtle differences in NarG activities did not follow a pattern consistent with either catabolite repression or the oxidation state of the principal carbon source.

Effects of cAMP on narG and lacZ operon expression.

cAMP is synthesized by adenylate cyclase, the product of the cya gene (11). Strains carrying a cya-null allele exhibit a conditional phenotype: Crp-dependent gene expression is activated by added cAMP, even during growth on glucose. Therefore, we tested the effects of added cAMP on NarG and LacZ synthesis in both cya⁺ and cya-null strains cultured in glucose medium supplemented with acid-hydrolyzed casein. As expected, cAMP had little effect on NarG synthesis; indeed, expression was slightly elevated in the cya-null strain VJS711 (data not shown). In contrast, LacZ synthesis was strongly dependent upon either the cya⁺ genotype or added cAMP.

Catabolite repression of napF operon expression.

We compared several carbon sources that differ in their oxidation states and in their abilities to elicit catabolite repression (see Materials and Methods). These experiments employed an narL-null strain carrying the Φ(napF-lacZ) [Δ146] construct cultured in defined medium.

During growth in the absence of nitrate, Φ(napF-lacZ) expression varied over a 4.3-fold range from glucose to mannose (Table 3). In the presence of nitrate, expression varied by only 2.1-fold. Thus, relative levels of expression were controlled by catabolite repression and not by carbon source oxidation state. We previously described, but did not show, preliminary data that led to the same conclusion (67). We decided to use glucose and mannose for most subsequent experiments, since these exhibited the greatest differences in expression levels and have identical oxidation states.

Catabolite repression influenced IPTG-induced lacZ operon transcription by about threefold (Table 2). In IPTG-induced cultures, inducer exclusion does not operate, so catabolite repression of lacZ operon transcription reflects only Crp activation (25, 49). Therefore, the catabolite repression effects on napF and lacZ operon transcription activation were similar in magnitude.

A Crp binding site in the napF operon control region.

Previous studies with the napF operon control region have characterized three promoters as well as sites for binding the phospho-NarP, Fnr, and ModE regulatory proteins (Fig. 1). A candidate site for binding the NsrR repressor is also evident. The region bound by IscR protein is shown, but the sequence specificity determinants for IscR binding to this site have not been defined.

While examining the napF control region, we noted a sequence with similarity to the Crp protein consensus binding site (Fig. 1), which consists of inverted hendacamer sequences (consensus aaaTGTGAtct, where uppercase letters represent more highly conserved positions) (31). This potential Crp protein binding site is centered at position −105.5 relative to the T1 transcription initiation site.

Figure 1 shows comparisons between napF operon control region sequences from three close relatives: Escherichia coli K-12, Citrobacter rodentium ICC168, and Salmonella enterica LT2. In these sequences, the P1 and P2 promoter elements and binding sites for the above-noted regulatory proteins are well conserved. This phylogenetic footprint (70) supports the idea that these conserved sequences are likely to be functionally important for regulated gene expression (6, 17).

Previously, Darwin and Stewart isolated a series of 5′ deletions in the napF operon control region (22). Deletions are denoted by their position relative to the T1 transcription initiation site. Our laboratory has used the Δ146 construct for most subsequent experiments. Two other deletion constructs, Δ123 and Δ85, differ only in the presence and absence, respectively, of the Crp site (Fig. 1).

Crp site and cAMP influence napF operon expression.

We examined Φ(napF-lacZ) expression from constructs that retain the Crp site (Δ146 and Δ123) or have had it deleted (Δ85). In the absence of nitrate, glucose inhibited expression from the former constructs by 4.5-fold but inhibited expression from the latter by only 1.3-fold (Table 4). This indicates that Crp-cAMP, bound to the site centered at position −105.5, mediates strong catabolite repression under this growth condition. In the presence of nitrate, glucose inhibited expression from all three constructs equally, by about twofold.

TABLE 4.

Effects of Crp site, carbon source, cAMP, and nitrate on Φ(napF-lacZ) expression

Fusiona	NO3−b	LacZ sp act (Miller units)c				Sp act ratiod
		Mannose	Glucose	+cAMP	-cAMP	Carbon	cAMP
Δ146	−	1,000	220	—e	—	4.5	—
Δ123	−	1,440	320	990	290	4.5	3.4
Δ85	−	240	180	460	360	1.3	1.3
Δ146	+	14,800	6,830	—	—	2.2	—
Δ123	+	16,050	8,500	9,190	4,380	1.9	2.1
Δ85	+	12,690	5,960	7,280	4,620	2.1	1.6

^aControl region deletion in the Φ(napF-lacZ) construct (Fig. 1). The genotype for strains to compare mannose versus glucose is narQ⁺ narP⁺ narX⁺ narL215::Tn10 λΦ(napF-lacZ). The genotype for strains to determine the effect of cAMP is narQ⁺ narP⁺ Δ(narXL)265::Km Δcya-854 λΦ(napF-lacZ) (Table 1).

^bPresence (+) or absence (-) of 40 mM nitrate during growth.

^cStrains were cultured anaerobically to the early exponential phase in MOPS defined medium. The carbon source was added (40 mM) as indicated, and cAMP was added (5 mM) as indicated. Cultures to test cAMP effects contained glucose as the sole carbon source.

^dRatio of LacZ specific activities between cultures with mannose and glucose or with and without added cAMP.

^e—, not determined.

We introduced the cya-null allele into strains carrying the Δ123 and Δ85 versions of the Φ(napF-lacZ) construct. During growth in the absence of nitrate, cAMP induced expression by 3.4-fold in the Δ123 strain but only by 1.3-fold in the Δ85 strain (Table 4). During growth in the presence of nitrate, the difference in induction level was less pronounced (2.1-fold versus 1.6-fold). Thus, the presence and absence of cAMP for the Δcya cultures mimicked growth on mannose and glucose, respectively, for the wild type. These results support the hypothesis that the influence of the presumed Crp site on Φ(napF-lacZ) expression results from binding by Crp-cAMP.

Catabolite repression controls promoter P2.

During growth in the absence and presence of nitrate, Φ(napF-lacZ) expression is mainly attributable to promoters P2 and P1, respectively. Previously, we constructed site-specific alterations in the Φ(napF-lacZ) Δ146 construct to ablate either of the two overlapping promoters (Fig. 1) (66). We used narL-null derivatives of these strains to examine catabolite repression.

Expression from the construct in which only promoter P2 is active was inhibited roughly fivefold during growth with glucose compared to mannose, even in the presence of nitrate (Table 5). In contrast, expression from the construct in which only promoter P1 is active was inhibited only about 1.5-fold during growth with glucose, even in the absence of nitrate. Analogous results were obtained for crp⁺ strains cultured with acid-hydrolyzed casein plus fructose or glucose (see below; Table 6).

TABLE 5.

Effects of control region, carbon source, cAMP, and nitrate on Φ(napF-lacZ) expression

Control region alteration		NO3−b	LacZ sp act (Miller units)c				Sp act ratiod
P1	P2		Mannose	Glucose	+cAMP	-cAMP	Carbon	cAMP
+	+	-	1,090	280	430	170	3.9	2.5
+	-	-	320	210	540	120	1.5	4.5
-	+	-	1,260	240	740	200	5.2	3.7
+	+	+	16,980	9,360	7,970	4,360	1.8	1.8
+	-	+	21,550	15,350	12,090	9,390	1.4	1.3
-	+	+	490	100	120	70	4.9	1.7

^a Control region alteration in Φ(napF-lacZ) construct (Fig. 1). The genotype for strains to compare mannose versus glucose is narQ⁺ narP⁺ narX⁺ narL215::Tn10 λΦ(napF-lacZ) [Δ146]. The genotype for strains to determine the effect of cAMP is narQ⁺ narP⁺ narX⁺ narL215::Tn10 λΦ(napF-lacZ) [Δ146] Δcya-854 (Table 1).

^bPresence (+) or absence (-) of 40 mM nitrate during growth.

^cStrains were cultured anaerobically to the early exponential phase in MOPS defined medium. The carbon source was added (40 mM) as indicated, and cAMP was added (5 mM) as indicated. Cultures to test cAMP effects contained glucose as the sole carbon source.

^dRatio of LacZ specific activities between cultures with mannose and glucose or with and without added cAMP.

TABLE 6.

Effects of control region, carbon source, and Crp on Φ(napF-lacZ) expression in the absence of nitrate

Straina	Control region alterationb		crp allelec	LacZ sp act (Miller units)d		Catabolite repression (fold)	Crp activation (fold)
	P1	P2		Fructose	Glucose		Fructose	Glucose
VJS5516	+	+	+	1,120	600	1.9	7.0	3.3
VJS11445	+	+	-	160	180	0.9
VJS7444	+	-	+	680	560	1.2	4.0	2.2
VJS11447	+	-	-	170	250	0.7
VJS7443	-	+	+	540	150	3.6	6.0	1.9
VJS11446	-	+	-	90	80	1.1

^aGenotype: narQ⁺ narP⁺ narX⁺ narL215::Tn10 λΦ(napF-lacZ) [Δ146] (Table 1).

^bControl region alteration in Φ(napF-lacZ) construct (Fig. 1).

^c+, crp⁺; -, Δcrp-45.

^dStrains were cultured anaerobically to the early exponential phase in MOPS defined medium supplemented with 0.2% acid-hydrolyzed casein. The carbon source was added (40 mM) as indicated.

We also examined the effect of cAMP on expression from these mutant promoters. Strains carrying the cya-null allele were cultured in defined medium with glucose as the carbon source. During growth in the absence of nitrate, added cAMP stimulated expression from the P2 and P1 promoters by 3.7- and 4.5-fold, respectively (Table 5). In contrast, added cAMP stimulated Φ(napF-lacZ) expression by less than twofold during growth in the presence of nitrate.

Crp controls promoters P1 and P2.

We next tested the influence of Crp by examining expression in crp-null strains cultured in medium with either fructose or glucose as the predominant carbon source. Although fructose can be catabolized by crp- and cya-null mutants (57), it nevertheless was necessary to add acid-hydrolyzed casein for growth (33). We only examined cultures grown in the absence of nitrate.

In the fructose-grown cultures, Φ(napF-lacZ) expression was increased by four- to sevenfold in the crp⁺ strains, whereas in glucose-grown cultures, expression was increased by about two- to threefold (Table 6).

Fnr influences catabolite repression of promoter P2.

We also tested Φ(napF-lacZ) expression in fnr-null strains cultured in defined medium with no added nitrate. Either mannose or glucose was present as the sole carbon source. Expression from the P1 promoter was decreased more than fivefold in the fnr-null strain regardless of carbon source (Table 7). In contrast, expression from the P2 promoter was decreased nearly sevenfold in the fnr-null strain grown with mannose but only about twofold with glucose.

TABLE 7.

Effects of control region, carbon source, and Fnr on Φ(napF-lacZ) expression in the absence of nitrate

Straina	Control region alterationb		fnr allelec	LacZ sp act (Miller units)d		Catabolite repression (fold)	Fnr activation (fold)
	P1	P2		Mannose	Glucose		Mannose	Glucose
VJS5516	+	+	+	1,240	300	4.0	6.5	4.9
VJS11439	+	+	-	190	61	3.1
VJS7444	+	-	+	320	290	1.1	5.5	5.6
VJS11441	+	-	-	49	52	0.9
VJS7443	-	+	+	1,170	210	5.6	6.9	2.2
VJS11440	-	+	-	170	95	1.8

^aGenotype: narQ⁺ narP⁺ narX⁺ narL215::Tn10 λΦ(napF-lacZ) [Δ146] (Table 1).

^bControl region alteration in Φ(napF-lacZ) construct (Fig. 1).

^c+, fnr⁺; -, fnr-275::Km.

^dStrains were cultured anaerobically to the early exponential phase in MOPS defined medium. The carbon source was added (40 mM) as indicated.

Catabolite repression of nrfA and napF expression.

Finally, we examined catabolite repression of napF operon expression in comparison to that of similarly regulated operons. These experiments used glycerol and glucose, for comparison with previous analysis of nrfA operon expression (72). Cultures were grown in defined medium and also in complex medium made by adding tryptone plus yeast extract.

The E. coli nrfABCDEFG operon, encoding periplasmic nitrite reductase, is subject to strong catabolite repression that is enhanced during growth in complex medium (72). During growth in defined medium, glucose strongly inhibited Φ(nrfA-lacZ) expression in both the absence and presence of nitrate (Table 8; 7.2-fold and 5.6-fold inhibition, respectively). Similar levels of catabolite repression occurred also in complex medium, in which overall expression was reduced nearly twofold in the nitrate-supplemented cultures.

During growth in defined medium without nitrate, glucose inhibited expression from the Φ(napF-lacZ) [Δ123] and Φ(napF-lacZ) [Δ85] constructs by 4.8-fold and 0.9-fold, respectively (Table 8). In contrast, catabolite repression of expression from both constructs was only about 1.5-fold during growth with nitrate. Thus, expression during growth with glycerol was similar to that during growth with mannose (compare Tables 3 and 8), further indicating that metabolic route per se does not influence napF operon expression (see Materials and Methods).

During growth in complex medium without nitrate, glucose inhibited expression from the Δ123 construct to a slightly greater degree (2.2-fold) than that from the Δ85 construct (Table 8; 1.5-fold). In contrast, catabolite repression of expression from both constructs was stronger (about 3.2-fold) during growth with nitrate. Thus, complex medium enhanced catabolite repression during growth with nitrate independently of the Crp site.

Genomes from many strains of Shewanella spp. carry the nrfA gene along with napDAGHB and napEDABC operons. We constructed a Φ(nrfA_So-lacZ) operon fusion in E. coli, using control region DNA from S. oneidensis MR-1. Catabolite repression was observed during growth in either the absence or presence of nitrate (Table 8). Strikingly, complex medium had virtually no effect on catabolite repression.

Transcription from the Haemophilus influenzae napF (napF_Hi) operon control region is activated by Fnr and phospho-NarP or phospho-NarL in E. coli (65). Here, glucose inhibited Φ(napF_Hi-lacZ) expression by less than twofold under all conditions studied (Table 8). Additionally, overall expression was enhanced nearly twofold during growth in complex medium. These results reinforce the conclusion from examining narG operon transcription (Table 2) that strong catabolite repression by glucose and by complex medium is not an essential property of Fnr- and Nar-regulated transcription.

DISCUSSION

The carbon source had little effect on narG operon transcription under all growth conditions examined (Table 2). In contrast, during growth in defined medium with no added nitrate, catabolite repression modulated napF operon transcription by at least fourfold (Tables 3, 4, and 8). NapABC enzyme likely functions to scavenge nitrate at low concentrations (51, 77). Thus, induction during growth with less-preferred carbon sources may serve to enhance energy capture in environments where nitrate is present transiently or in trace amounts.

The Crp binding site.

We used visual inspection and phylogenetic footprinting to identify a Crp binding site in the napF operon control region (Fig. 1). The upstream half-site matches the core consensus sequence (TGTGA), whereas the downstream half-site has two to four matches, depending on the species. The site also exhibits several matches to the less-conserved positions that flank or separate the half-site core sequences (31).

The napF operon Crp site was identified independently by Brown and Callan, who used a computer-enabled whole-genome approach to phylogenetic footprinting (13). Their calculated binding energy for the napF operon site is similar to those for some documented sites, including that for the well-studied galE operon (79).

For cultures grown without nitrate, glucose inhibited expression from the Δ123 and Δ146 constructs by fourfold or greater (Table 4). In contrast, expression from the Δ85 construct, in which the Crp site is deleted, was inhibited by 1.3-fold at most. Moreover, in a cya-null strain, cAMP stimulated expression by 3.4-fold from the Δ123 construct, but by only 1.3-fold from the Δ85 construct (Table 4). Thus, the Crp site is required for carbon source regulation of napF operon transcription in the absence of nitrate.

Crp-dependent promoter P2 activation.

E. coli napF operon transcription is initiated from two overlapping promoters. (A third, upstream promoter is active under iron-sulfur cluster stress conditions [29].) Studies with glucose-grown cultures concluded that transcription from promoter P1 is stimulated in response to anaerobiosis and nitrate and that promoter P2 accounts for the majority of napF operon transcription during growth without added nitrate (19, 23, 66). It was speculated that transcription from promoter P2 might be regulated under growth conditions not yet explored (66).

Expression from a mutant construct in which promoter P1 is inactivated was inhibited fivefold by glucose (Table 5), stimulated nearly fourfold by cAMP in a cya-null strain (Table 5), and reduced sixfold in a crp-null strain cultured with fructose as the major carbon source (Table 6). This suggests that a major function of Crp-cAMP is to activate transcription from promoter P2 during growth with less-preferred carbon sources and no added nitrate.

Expression from a mutant construct in which promoter P2 is inactivated was controlled only weakly by catabolite repression (1.5-fold or less) during growth in either the absence or presence of nitrate (Tables 5 to 7). However, during growth with no added nitrate, expression was stimulated 4.5-fold by cAMP in a cya-null strain (Table 5) and reduced fourfold in a crp-null strain cultured with fructose as the major carbon source (Table 6). Whether this reflects synergy between Crp and Fnr or an indirect effect of Crp (see below) currently is unknown. The physiological relevance of these results is uncertain, however, since transcription initiation from promoter P1 in wild-type constructs is only detected during growth with nitrate (19, 66) or in the presence of phospho-NarP protein (23).

Synergy between Crp and Fnr?

The Crp site is centered at position −102.5 with respect to the promoter P2 transcription initiation site (Fig. 1). This location is in accordance with established spacing constraints for Crp action (15) and is similar to the uhpT control region for which the Crp site is centered at position −103.5 (44). The Fnr site is centered at position −61.5 relative to the promoter P2 initiation site (Fig. 1).

Synergistic activation can occur with tandem binding by two dimers of Fnr or Crp or one dimer of each (3, 16, 35, 59). For one model control region, optimal activation by tandem Crp dimers occurs when one binding site is centered at position −61.5 and the other is near position −93.5 or −103.5 (71). Thus, the locations of the Fnr and Crp sites suggest that the two proteins may act in synergy to control aspects of napF operon transcription.

In glucose-grown cultures, expression from a mutant construct in which promoter P1 is inactivated is reduced only about twofold in an fnr-null strain (66) (Table 7). However, in mannose-grown cultures, expression was reduced by nearly sevenfold in an fnr-null strain (Table 7). Therefore, maximal transcription from promoter 2 requires both Crp and Fnr (Tables 6 and 7), as found previously for the E. coli ansB gene (34, 59).

Crp site-independent catabolite repression in nitrate-grown cultures.

In nitrate-grown cultures with defined medium, glucose inhibited Φ(napF-lacZ) expression between 1.5- and 2-fold, even from the Δ85 construct in which the Crp site is deleted (Tables 4 and 8). Furthermore, in a cya-null strain, cAMP stimulated expression by 2.1-fold and 1.6-fold from the Δ123 and Δ85 constructs, respectively (Table 4). Thus, the relatively weak catabolite repression of nitrate-activated napF operon transcription appears to be cAMP dependent. This could reflect direct action by Crp-cAMP from a weak binding site not apparent from sequence inspection. It is conceivable, albeit unlikely, that Crp binds weakly to the Fnr site (5, 80).

Alternatively, if the effects of Crp and cAMP were indirect, a different regulator might influence carbon source or growth rate control of nitrate-activated napF operon expression. Previous work revealed strong catabolite repression of nrfA operon expression (Table 8) (72). Detailed analysis of the nrfA operon control region has failed to uncover a role for Crp-cAMP. Instead, catabolite repression is mediated by the Fis and Cra proteins (14, 72). We note a recent global analysis which found that expression of the napF, nrfA, and narG operons is affected by a fis-null allele (10).

Catabolite repression of nrfA operon expression is enhanced during growth in complex medium (72) (Table 8). Complex medium similarly enhanced catabolite repression of napF operon expression in nitrate-grown cultures (Table 8). Thus, synthesis of the periplasmic nitrate ammonification pathway appears to be regulated in response to both quality and quantity of available carbon.

nap and nrf regulation in other species.

Variants of the nap and nrf operons are present in members of the Gammaproteobacteria that also have the narQ and narP genes (53, 63). Upstream control sequences from representative species contain NarP and Fnr sites in all cases examined (53; data not shown). However, these sequences from members of the Vibrionaceae and Pasteurellaceae generally do not contain apparent Crp sites. (One exception is the napF operon from Mannheimia succiniciproducens MBEL55E [not shown].) Consistent with this finding, expression of the H. influenzae napF operon cloned in E. coli varied by only twofold or less in response to carbon source (Table 8).

Genomes from many species of Shewanella contain narQP, napDAGHB, nrfA, and napEDABC operons (63; data not shown). In all such cases examined, the napD and nrfA operon upstream control regions contain NarP, Fnr, and Crp sites. In contrast, napE operons from the same genomes contain NarP and Fnr sites but no apparent sites for Crp. Representative examples are shown in Fig. 2. Indeed, expression of the S. oneidensis nrfA operon cloned in E. coli varied as much as fivefold in response to carbon source (Table 8), indicating that the Crp site is functional.

FIG. 2.

Upstream regions for Shewanella nrfA (A), napD (B), and napE (C) operons. The sequences are for S. oneidensis MR-1, Shewanella sp. strain MR-4, S. amazonensis SB2B, and S. loihica PV-4. Nucleotides that match the consensus sequences are shaded black (promoter −10 and −35 elements) or gray (NarP, Fnr, and Crp). Y, C or T; M, A or C; K, G or T; and W, A or T. Sequence removed between the NarP and −10 elements is indicated by numbers.

Studies with S. oneidensis MR-1 have implicated Fnr (termed EtrA) and Crp separately in controlling anaerobic respiratory gene expression (4, 41, 56). Work presented here suggests a hypothesis to reconcile these results: Fnr and Crp may act synergistically, so that loss of either leads to strongly decreased gene expression.

Carbon source and redox balance.

Paracoccus pantotrophus NapABC enzyme synthesis is induced during growth on reduced carbon substrates to dissipate excess reducing power through nitrate dissimilation (54, 60). E. coli NapABC enzyme similarly plays a role in maintaining redox balance during anaerobic growth (12). However, we observed no influence of carbon source redox potential on the control of NarGHI or NapABC enzyme synthesis (Tables 2 and 3). Presumably, the mixed-acid fermentation provides sufficient flexibility to accommodate carbon sources of different redox potentials (20).