A Set of Activators and Repressors Control Peripheral Glucose Pathways in Pseudomonas putida To Yield a Common Central Intermediate

Abstract

Pseudomonas putida KT2440 channels glucose to the central Entner-Doudoroff intermediate 6-phosphogluconate through three convergent pathways. The genes for these convergent pathways are clustered in three independent regions on the host chromosome. A number of monocistronic units and operons coexist within each of these clusters, favoring coexpression of catabolic enzymes and transport systems. Expression of the three pathways is mediated by three transcriptional repressors, HexR, GnuR, and PtxS, and by a positive transcriptional regulator, GltR-2. In this study, we generated mutants in each of the regulators and carried out transcriptional assays using microarrays and transcriptional fusions. These studies revealed that HexR controls the genes that encode glucokinase/glucose 6-phosphate dehydrogenase that yield 6-phosphogluconate; the genes for the Entner-Doudoroff enzymes that yield glyceraldehyde-3-phosphate and pyruvate; and gap-1, which encodes glyceraldehyde-3-phosphate dehydrogenase. GltR-2 is the transcriptional regulator that controls specific porins for the entry of glucose into the periplasmic space, as well as the gtsABCD operon for glucose transport through the inner membrane. GnuR is the repressor of gluconate transport and gluconokinase responsible for the conversion of gluconate into 6-phosphogluconate. PtxS, however, controls the enzymes for oxidation of gluconate to 2-ketogluconate, its transport and metabolism, and a set of genes unrelated to glucose metabolism.

Pseudomonas putida mt-2 was isolated from soils in Japan and identified as able to grow on meta-toluate as the sole C source. Sequencing of its genome revealed that it possesses many genes encoding proteins involved in the degradation of plant-derived chemicals, including a variety of methoxylated aromatic acids, hydroxylated aromatic acids, and other compounds (6, 12, 23). The strain was also found to possess information for the degradation of fructose (39) and glucose (3, 4, 39), two of the most abundant sugars present in plant root exudates (14).

del Castillo et al. (4) showed that glucose catabolism in this strain occurred through the simultaneous operation of three peripheral pathways that converge at the level of 6-phosphogluconate (6PG) (Fig. 1). This compound is further metabolized by the Entner-Doudoroff pathway to yield glyceraldehyde-3-phosphate and pyruvate. Glucose enters the periplasmic space via the OprB outer membrane porin(s) (16, 33, 38). Once in the periplasm, glucose can be either transported to the cytoplasm via an ABC glucose transport system (2, 3, 11) or oxidized to gluconate. In the cytoplasm, glucose is phosphorylated via a glucokinase, and the resulting glucose-6-phosphate is converted to 6PG by the concerted action of the glucose-6-phosphate dehydrogenase and 6-phosphoglucolactonase. Gluconate in the periplasm can be either transported into the cytoplasm via the GntP protein and subsequently phosphorylated to 6PG or oxidized to 2-ketogluconate (2KG), which is transported to the cytoplasm and, upon phosphorylation, is reduced to 6PG by the concerted action of the kguK/kguD gene products (Fig. 1).

FIG. 1.

Glucose catabolism in P. putida as deduced from gene annotations. At the top are the events that occur in the outer membrane and the reactions that take place in the periplasmic space. Also shown is the transport of glucose, gluconate, and 2KG into the cell. The set of catabolic reactions that take place in the cytoplasm is depicted. The genes that encode the enzymes involved are indicated for all steps. OM, outer membrane; PG, periplasmic space; IM, inner membrane; G3P, glyceraldehyde 3-phosphate; PYR, pyruvate.

Genome analysis of the set of genes involved in glucose metabolism revealed that they were organized in sets of clusters scattered along the chromosome (4) (Fig. 2). Transcriptional analysis revealed a complex organization in operons and monocistronic units within each of these clusters (Fig. 2). One of the most relevant features this study disclosed was that the oprB1 gene, encoding the OprB1 porin involved in the entry of glucose into the periplasmic space, was in an operon made up of open reading frames (ORFs) PP1015 to PP1019, which encodes the glucose transport system (Fig. 2A). Another relevant finding from this study was that the zwf-1 allele encoding glucose-6-phosphate dehydrogenase (18) was in an operon within the eda gene that encodes one of the Entner-Doudoroff enzymes, whereas the gene encoding glucokinase (glk) formed an operon with the edd gene. Transcribed in the opposite direction from the zwf-1 promoter was the hexR gene, encoding a putative repressor (Fig. 2A).

FIG. 2.

Organization of the peripheral glucose catabolic genes. Gene numbering and organization are derived from the annotation of the complete genome sequence of P. putida KT2440 in the TIGR database. (A) Set of 16 genes, most of which have been assigned a function based on enzymatic assays. (B) Set of nine genes, most of which have a specific role in gluconate catabolism through 2KG. (C) Set of four genes involved in gluconate catabolism in P. putida KT2440. Numbers in italics between genes indicate the distance between the stop codon of the preceding gene and the start codon of the following gene; negative numbers indicate that contiguous genes overlap.

The gene that encodes the gluconate transporter gene (gntP) and the gene encoding the gluconokinase (gnuK) were contiguous on the chromosome, but did not form an operon (4) (Fig. 2). Transcribed divergently with respect to the gnuK gene was the gnuR gene, which encodes a repressor assigned to the LacI family (Fig. 2C). The gluconate dehydrogenase genes (PP3384 and PP3382) and the genes encoding the 2KG transporter and its subsequent metabolism are within a cluster in the chromosome that covers from PP3376 to PP3384. A regulator of the PtxS family of repressors is found within this cluster of genes, which encode enzymes for gluconate oxidation to 2KG and its subsequent metabolism (Fig. 2B).

Global transcriptomic analysis revealed that all of the catabolic genes and regulators mentioned above were up-regulated in response to glucose, except for the hexR gene (4). It was also shown that when KT2440 bearing the TOL plasmid was grown simultaneously in the presence of glucose and toluene, the glk branch and the glucose transport system were under catabolite repression control, which was not the case for the set of genes involved in gluconate metabolism. This pointed toward independent evolutionary acquisition of genes for glucose metabolism and differential gene regulation.

The present study was undertaken to further define the regulatory circuits that allow P. putida KT2440 to use glucose as the sole C source. To this end, mutants in each of the four potential regulators within the cluster of genes for glucose catabolism were generated. Growth characteristics, global transcriptomic analysis, and transcriptional fusions revealed that the GltR-2 protein is a transcriptional activator involved in the induction of the glucose transport system. HexR and GnuR are local repressors; the latter controls the expression of gluconate transport and its phosphorylation system, whereas the former controls the three operons involved in glucose metabolism via glucokinase. The regulator PtxS acted as the local repressor of the 2KG genes, although surprisingly, in a ptxS-deficient mutant, 10 genes with no relation to glucose metabolism were up- or down-regulated, indicating that this regulator exerts a more global type of control.

MATERIALS AND METHODS

Bacterial strains, plasmids, and culture conditions.

Pseudomonas putida KT2440 and a series of isogenic mutant strains with insertions in the gltR-2, hexR, ptxS, and gnuR genes are shown in Table 1. The GltR-2 and HexR mutant strains were identified by Duque et al. (6) in a general screening after random mutagenesis of P. putida with mini-Tn5-Km. The ptxS and gnuR mutants were generated in this study and are described below. Pseudomonas putida strains were grown at 30°C in M9 minimal medium with glucose (16 mM) or sodium citrate (16 mM) as a carbon source. When appropriate, antibiotics were added at the following concentration: kanamycin (Km), 25 μg/ml; and tetracycline (Tc), 10 μg/ml.

TABLE 1.

P. putida strains and plasmids used in this study

Strain or plasmid	Genotype or relevant characteristic(s)a	Source or reference
Strains
KT2440	Wild type, prototroph; Cmr Rifr	This laboratory
M2052b	gltR-2::mini-Tn5-Km; Kmr Rifr	6
M1881b	hexR::mini-Tn5-Km; Kmr, Rifr	6
PSC303c	ptxS::pCHESIΩ-Km; Rifr	This study
PSC304c	gnuR::pCHESIΩ-Km; Rifr	This study
Plasmids
pMP220	Tcr; ′lacZ promoter probe plasmid	9
pCHESIΩKm	Kmr	17
pTC1	pCHESIΩ-Km bearing ′ptxS′	This study
pTC2	pCHESIΩ-Km bearing ′gnuR′	This study
pBedd	Tcr; pM220 bearing the promoter region of the edd gene	This study
pPP1015	Tcr; pM220 bearing the region upstream of ORF PP1015	This study
pBhex	Tcr; pM220 bearing the promoter region of the hexR gene	This study
pBzwf	Tcr; pM220 bearing the promoter region of the zwf gene	This study

^aCm^r, Km^r, Tc^r, and Rif^r represent resistance to chloramphenicol, kanamycin, tetracycline, and rifampin, respectively.

^bCollection of KT2440 mutants available at the Consejo Superior de Investigaciones Científicas, Granada, Spain.

^cPseudomonas Stock Center.

DNA techniques.

Preparation of plasmid and chromosomal DNA, digestion with restriction enzymes, ligation, electrophoresis, and Southern blotting were performed as described before (28, 30).

Site-specific homologous inactivation of gnuR and ptxS.

To construct mutant strains bearing an inactivated chromosomal version of the gnuR and ptxS genes, we generated the corresponding knockout using the appropriate derivatives of pCHESIΩKm (17). Plasmid pCHESIΩKm is based on pUC18 and bears the origin of transfer oriT of RP4 and the Ω-Km interposon of plasmid pHP45ΩKm cloned as a HindIII fragment. To generate the desired mutation, an internal fragment between 400 and 500 bp of the target gene was amplified by PCR with primers containing the EcoRI and BamHI sites to amplify an internal part of the gnuR gene and with primers containing the EcoRI and XbaI sites to amplify the ptxS gene. The amplified fragments were subsequently cloned between the EcoRI and BamHI sites of pCHESIΩKm (Table 1) in the case of gnuR and EcoRI and XbaI in the case of the ptxS gene (Table 1). In both cases, the fragments were cloned in the same transcriptional direction as the P_lac promoter. The recombinant plasmids were introduced into P. putida KT2440 by electroporation, and transformants bearing a cointegrate of the plasmid in the host chromosome were selected on M9 minimal medium with citrate as a carbon source and Km. The correct insertion of the mutant allele was confirmed by colony-screening PCR using a primer based on the Km marker gene and another primer that was annealed to the sequence complementary to the cloned gene fragment. The correctness of the construction was confirmed by Southern blotting using the target gene as a probe (not shown).

Construction of ′lacZ transcriptional fusions and determination of β-galactosidase assays.

Transcriptional fusions were constructed by cloning the putative promoters in 5′ with respect to ′lacZ in pMP220 (9). To this end, the sequence upstream of the edd, PP1015, zwf-1, and hexR genes was amplified by PCR with primers containing the EcoRI-PstI sites 5′ and 3′ of the amplified fragment. These fragments were cloned into the EcoRI-PstI sites of pMP220. All fusion constructs (pBedd, pPP1015, pBhex, and pBzwf) were confirmed by DNA sequencing (Table 1). Transcriptional fusion constructs were assayed in P. putida KT2440 and its isogenic mutant background.

The pBedd, pPP1015, pBzwf, and pBhex plasmids were introduced into Pseudomonas putida KT2440 or its isogenic mutants by electroporation. Transformants were grown overnight on minimal medium with citrate as the sole carbon source in the presence of Tc, and then the cultures were diluted to reach a turbidity (optical density at 660 nm [OD₆₆₀]) of 0.05 in the same medium. After 1 h of incubation at 30°C with shaking (200 rpm in an orbital platform), the cultures were split into five aliquots to which and 5 mM glucose, gluconate, 2KG, or fructose was added, except for the fifth, which was kept as a control. After 6 h of incubation, β-galactosidase activity was assayed in permeabilized whole cells according to Miller's method (20). Three independent assays were run in triplicate.

Pseudomonas putida microarrays.

The genome-wide DNA chip used in this work (printed by Progenika Biopharma) was described in detail previously (7, 44). It consists of an array of 5,539 oligonucleotides (50-mer) spotted in duplicate onto γ-aminosilane-treated slides and covalently linked with UV light and heat. The oligonucleotides represent 5,350 of the 5,421 predicted ORFs annotated in the P. putida KT2440 genome (23) The chips are also endowed with homogeneity controls consisting of oligonucleotides for the rpoD and rpoN genes spotted at 20 different positions, as well as duplicate negative controls at 203 predefined positions.

For RNA preparation, Pseudomonas putida KT2440 and mutant cells were grown in minimal medium with citrate as a carbon source until the early exponential phase was reached (turbidity at 660 nm was about 0.5). Cells from 12-ml culture samples were harvested by centrifugation (7,000 × g) at 4°C in tubes precooled in liquid nitrogen. Total RNA was isolated with TRI reagent (Ambion; reference no. 9738), as recommended by the manufacturer, and then subjected to DNase treatment followed by purification with RNeasy columns (Qiagen; catalog no. 74104). The RNA concentration was determined spectrophotometrically, and its integrity was assessed by agarose gel electrophoresis.

To prepare fluorescently labeled cDNA, we primed 25 μg of RNA with 7.5 μg of pd(N)₆ random hexamers (Amersham; catalog no. 27-2166-01). Probes were synthesized at 42°C for 2 h exactly as described before (7). Labeling efficiency was checked with a NanoDrop ND1000 spectrophotometer (NanoDrop Technologies). Hybridization conditions were as described before (7, 44). Images were acquired at 10-μm resolution, and the background-subtracted median spot intensities were determined with GenePix Pro 5.1 image analysis software (Axon Instruments, Inc.). Signal intensities were normalized by applying the LOWESS intensity-dependent normalization method (42) and statistically analyzed with Almazen System software (Alma Bioinformatics S.L.). For appropriate statistical analysis of the results, RNA preparations from at least four independent cultures were tested for each strain (1). P values were calculated with Student's t test. A particular ORF was considered differentially expressed if (i) the change was at least 1.8-fold and (ii) the P value was 0.05 or lower.

Primer extension analysis.

For primer extension analyses, we isolated RNA as described above, and the process was carried out as described by Marqués et al. (19). The amount of total RNA template used in each reaction varied between 90 and 100 μg. About 10⁵ cpm of ³²P-labeled 5′-end-specific oligonucleotides was used as a primer in extension reactions. To improve electrophoresis quality, the cDNA products obtained after reverse transcriptase were treated with DNase-free RNase (Roche; reference no. 1119915). The cDNA products were separated and analyzed in urea-polyacrylamide sequencing gels which were exposed to a phosphor screen (Fuji Photo Film Co., Ltd.) for 5 to 12 h. Phosphor screens were scanned with a phosphorimaging instrument (Molecular Imager FX; Bio-Rad). Data were quantified with Quantity One software (Bio-Rad).

RESULTS

Construction of mutants in genes putatively involved in the regulation of glucose metabolism and their characterization.

Pseudomonas putida KT2440 and the four isogenic knockout mutants in gltR-2, hexR, ptxS, and gnuR were grown overnight on M9 minimal medium with glucose as the sole C source. Then they were washed and suspended in the same medium to reach an initial cell density of 0.05 U at OD₆₆₀, and growth was monitored. We found that after an initial 2-h lag phase, the wild-type strain grew exponentially with a μ_max of 0.59 ± 0.02 h⁻¹ and consumed glucose at a rate of 7.60 ± 0.19 mmol glucose/g cell biomass·h⁻¹. The mutants deficient in gnuR, hexR, and ptxS grew promptly, with growth rates (0.57 ± 0.01 h⁻¹ to 0.60 ± 0.02 h⁻¹) and glucose consumption rates (6.84 ± 0.8 to 9.7 ± 0.5 mmol glucose/g cell biomass·h⁻¹) similar to those of the parental strain. In the gltR-2 mutant, the lag phase was much longer (about 5 h), but once the culture began to grow the growth and glucose consumption rates were almost identical to those of the parental strain.

These results suggested that in general, no major effect on growth rate and glucose uptake rate was associated with deficiency in any of the potential regulatory genes for glucose catabolism. However, the above results provided no indication of the specific effects of the mutations on the different catabolic segments involved in glucose degradation. To gain insights into local and global effects, the parental strain and each of the mutant strains were grown on M9 minimal medium with citrate and total mRNA was prepared and labeled as indicated in Materials and Methods. The expression levels were then compared between the mutant strain and the parental one (Tables 2 to 5).

TABLE 2.

Genes differentially up-regulated in the GnuR mutant background with respect to the genes expressed in the parental strain

ORF	Gene	Function	Fold induction	P value	Gene organizationa
PP1023	pgl	6-Phosphogluconolactonase	1.86	0.01	O
PP1516		RND efflux pump	1.91	0.05	O
PP2333		GntR family of regulators	1.82	0.03	O
PP2741		Hypothetical protein	1.96	0.04	O
PP3415	gnuR	Transcriptional repressor	2.85	0.02	M
PP3416	gnuK	Gluconate kinase	2.60	0.05	M
PP3417	gntP	Gluconate transporter	1.82	0.05	O
PP3418		Hypothetical protein	1.80	0.01	O
PP3642		Hypothetical protein	1.93	0.5	M

^aM, monocistronic; O, operon.

TABLE 5.

Genes differentially up-regulated in the GltR2 mutant background with respect to the genes expressed in the parental strain

ORF	Family	Fold change	P value	Gene organizationa
PP0299	Conserved hypothetical protein	1.90	0.004	M
PP1516	RND efflux pump	1.95	0.004	O
PP1545	Hypothetical protein	1.81	0.01	O
PP2059	Conserved hypothetical protein	1.86	0.008	M
PP2722	Conserved hypothetical protein	1.97	0.05	M

^aM, monocistronic; O, operon.

The divergent organization of gnuR and the gnuK gntP genes (Fig. 2) suggested that GnuR might regulate these genes. Indeed, we found that in the gnuR mutant background the gnuK gene was expressed at a level 2.6-fold higher than in the parental strain, as was also the case for gnuR, indicating that GnuR controls its own expression. Reverse transcription-PCR confirmed that the ORFs encoding PP3417 and PP3418 formed an operon. In agreement with this observation was the fact that expression of both genes increased, and that of the ORF encoding the PP3418 hypothetical protein was 1.8-fold times higher in the mutant strain than in the parental one. We found that expression of the pgl gene was induced about 1.9-fold, whereas expression of the zwf-1 gene upstream of pgl was not induced. This was somewhat surprising since our previous reverse transcription-PCR assays showed that they were part of the hex regulon (see below). The results also revealed that the level of four other genes varied ≥1.8-fold. This was found for a set of genes that encode proteins of unknown function (Table 2).

In the hexR mutant background, some of the genes in the PP1009 to PP1024 segment of the chromosome changed their expression, whereas another set did not do so (Table 3). In the hexR mutant background, expression of PP1009, which encodes gap-1, and the operon PP1010 to PP1012 increased between 4.89- and 1.94-fold. PP1022 through PP1024 also increased its expression between 2.04- and 6.07-fold. Only one gene outside the cluster also exhibited significant increase in expression: the ORF encoding the PP4488 hypothetical protein. Surprisingly, the expression level of the segment PP1015 to PP1018, which encodes the glucose ABC transport systems, did not change. This result showed that its regulation is not under the control of HexR. The level of the hexR gene did not change either, which suggests that the protein does not regulate its own expression. Therefore, HexR seems to control the expression of the glucokinase and Entner-Doudoroff pathway enzymes, together with glyceraldehyde-3-phosphate dehydrogenase, the enzyme that acts on the final product of glucose metabolism and helps to channel glucose to Krebs cycle intermediates (Fig. 1).

TABLE 3.

Genes differentially up-regulated in the HexR mutant background with respect to the genes expressed in the parental strain

ORF	Gene	Function	Fold induction	P value	Gene organizationa
PP1009	gap-1	Glyceraldehyde-3-P-dehydrogenase	4.89	0.01	M
PP1010	edd	6-Phosphogluconate dehydratase	4.06	0.05	O
PP1011	glk	Glucokinase	2.21	0.05	O
PP1012	gltR-2	DNA-binding response regulator GltR	1.94	0.033	O
PP1022	zwf-1	Glucose-6-P-dehydrogenase	2.04	0.032	O
PP1023	pgl	6-Phosphogluconolactonase	6.07	0.017	O
PP1024	eda	2-Dehydro-3-deoxyphosphogluconate aldolase	4.77	0.023	O
PP4488		Conserved hypothetical protein	3.38	0.05	M

^aM, monocistronic; O, operon.

The ptxS gene is located within a set of genes that includes the gluconate dehydrogenase operon (PP3384 to PP3382) and the 2-ketogluconate metabolism operon (kguE through kguD). We found that in the ptxS mutant background, expression of the first gene in these two operons and that of ptxS itself increased between 2.06- and 3.23-fold (Table 4). We also found that the genes for two hypothetical proteins (PP2741 and PP2984) also increased in this mutant background (Table 4). We found that eight genes, unrelated to glucose catabolism, were repressed in this mutant background 1.81- to 3.67-fold (Table 4).

TABLE 4.

Genes differentially regulated in the PtxS mutant background with respect to the genes expressed in the parental strain

ORF	Gene	Family	Fold change	P value	Gene organizationa
PP3064		gpX-pyocin R2 tail component protein	−2.37	0.000	O
PP3087		Exonuclease ABC A subunit, putative	−1.81	0.003	M
PP3088		Conserved hypothetical protein	−3.06	0.002	O
PP3089		Conserved hypothetical protein	−3.67	0.002	O
PP3095		Chaperone-associated ATPase, putative	−2.21	0.005	O
PP3099		Conserved hypothetical protein	−1.94	0.002	O
PP3100		Conserved hypothetical protein	−1.93	0.008	O
PP3252		Modification methylase HemK family	−2.80	0.003	O
PP2741		Conserved hypothetical protein	1.92	0.009	O
PP2984		Hypothetical protein	1.97	0.005	O
PP3379	kguE	Epimerase KguE, putative	2.20	0.007	O
PP3380	ptxS	PtxS family	3.23	0.004	O
PP3384		Conserved hypothetical protein	2.06	0.008	O

^aM, monocistronic; O, operon.

A gltR-2 mutant was generated as described in Materials and Methods. When the transcriptional pattern in cells grown on citrate was compared to that of the parental strain, we found a significant increase in expression (almost twofold) for a set of five ORFs that encoded proteins of unknown function, which were scattered along the chromosome and had no links with the glucose catabolism gene cluster (Table 5). These results suggested that GltR-2 does not function as a repressor for glucose metabolism.

To test whether GltR-2 could act as an activator of the glucose transport system as occurs in P. aeruginosa (10), we fused the putative promoter region of ORF PP1015 to ′lacZ in pMP220 to yield P_gstA and assayed the fusion both in the wild type and in the set of mutant regulators of the glucose degradation pathway. We found that in the parental strain, in the hexR, gnuR, and ptxS mutant backgrounds, β-galactosidase activity increased from negligible levels in the absence of glucose to 630 to 790 Miller units in the presence of this sugar. In the gltR-2 mutant background, expression was null even in the presence of glucose. This set of results indicated that GltR-2 is the positive transcriptional regulator of the glucose transport system.

Fusion of the glucose catabolic operon promoter regions to ′lacZ to validate the arrays.

To verify the set of results presented above, we also generated fusions of the promoter of the edd-glk operon and the zwf-1-pgl-eda operon (P_zwf) to ′lacZ (P_edd) to yield plasmids pBedd and pBzwf, respectively (Table 1). These plasmids were transformed in the wild-type strain and the hexR mutant, and β-galactosidase activity was determined in the absence and in the presence of glucose (Table 6). Expression of P_edd and P_zwf in the absence of glucose was lower in the parental strain than in the hexR mutant, which suggested that HexR represses expression of these two promoters. However, in the presence of glucose the levels of expression from these promoters were equally high in both backgrounds (Table 6). We also fused the hexR promoter P_hexR to ′lacZ and tested its expression. As shown in Table 6, the levels of expression were equally high in both backgrounds regardless of the presence of glucose. Hence, these results are in consonance with microarray data.

TABLE 6.

Expression from P_edd and P_zwf in KT2440 and the hexR mutant strain

Promoter	Background	β-Galactosidase activity (Miller units):
		− Glucose	+ Glucose
Pedd	Wild type	60 ± 3	150 ± 5
	hexR mutant	120 ± 5	200 ± 5
Pzwf	Wild type	100 ± 20	1,210 ± 10
	hexR mutant	1,260 ± 10	1,340 ± 40
PhexR	Wild type	220 ± 10	295 ± 10
	hexR mutant	205 ± 10	255 ± 20

We also used fusions of P_gnuK and P_orfPP3384 to ′lacZ and measured their expression in GnuR and PtxS mutant backgrounds. In the gnuR mutant background, expression from P_gnuK was about fourfold higher than in the parental strain. Expression from P_orfPP3384 was threefold higher in the PtxS-deficient background than in the parental strain (not shown). Therefore, this set of results also confirmed our transcriptomic results.

Identification of the tsp of the glucose catabolism operon.

To learn more about expression of the operons regulated by the three transcriptional repressors (PtxS, GnuR, HexR) and the activator (GltR-2) in glucose metabolism, we decided to determine the transcription start points (tsp) of a number of well-established operons.

(i) HexR-regulated genes.

For the HexR-regulated genes, we determined the tsp of hexR, gap-1, the edd-glk operon and the zwf-1-eda operon (see Fig. 3 for gap-1). In all cases, a single tsp was found. For hexR, the transcription start point was a C located 67 bp upstream of the first G of the first potential GTG codon. We found −10 (−12 TACGAT −7) and −35 (−35 TGGTAC −30) hexamers similar to those recognized by sigma-70 in Pseudomonas (5).

FIG. 3.

Transcription start point of the promoters in front of the ptxS, kguE, gnuK, gnuR, and gap-1 genes. Preparation of RNA from cells growing exponentially on glucose is described in Materials and Methods. Sequences of the primers used for primer extension assays will be supplied on request. In all panels, M refers to single-strand markers and ATCG is a sequencing ladder. The primer extension products for ptxS (S1), kguE (E2), gnuK (K2), gnuR (R1), and gap1 (1 and 1.1, since two different primers were used) are shown. Other conditions are as described in Materials and Methods.

For gap-1, edd, and zwf, a single tsp was also identified, and −10 and −35 hexamers similar to those recognized by RNA polymerase with sigma-70 were found. These three promoters are under HexR repressor control. Repressors have often been found to exert their role by binding to inverted repeats that partly or fully overlap the RNA-polymerase binding site (31). For this reason, we searched for potential inverted repeats between −50 and +50. We found a GntTtTaN₁₂TAAAAnC inverted motif that was located between −21 and +8 in the zwf-1 promoter, between −16 and +13 in the edd promoter, and between −17 and +15 in the gap-1 promoter. The potential role of this sequence as the true site of recognition by HexR awaits further experimental studies. In accordance with HexR recognizing the above inverted motif is the fact it was also found in the promoter region of the ORF PP4488, whose level also increased in the hexR mutant background.

(ii) GnuR-regulated genes.

We determined the +1 residue of the gnuR and gnuK promoters (Fig. 3). The +1 residue in both cases was a G located 37 and 151 bp upstream of the A of the first potential ATG. As above, in front of these genes we found potential −10 and −35 hexamers similar to those sigma-70 recognizes in Pseudomonas (5). Overlapping them was an inverted repeat, GTCCnTACN₃GTAnGGAC, located between −78 and −96 in the gnuR promoter and +59 and +77 in the gnuK promoter. This motif, however, was not found in front of the other five genes whose level of expression increased in the absence of GnuR, suggesting they may not be under the direct control by this regulator.

(iii) PtxS-regulated promoters.

In PtxS, we determined the +1 of the ptxS gene (PP3380) and the operon promoter of the 2KG operon metabolism (kguE) (Fig. 3). Both promoters had a single tsp with AT-rich −10 hexamers and relatively well-conserved −35 hexamers. The potential PtxS recognition site was a 5′-TGAAACCGGTTTCA-3′ inverted repeat that overlapped the −10 region in kguE and covered the +1 region in ptxS. This motif is also present in the ptxS gene in P. aeruginosa and has been shown to be the target of PtxS in this human pathogen (35). This motif also overlaps the −10 region of the promoter of PP3384. We searched for this motif in other regulated promoters in the ptxS mutant background, although it was not found in the putative promoter regions of these genes.

(iv) GltR-2 controls expression from the gtsA gene in response to glucose.

In cells growing on glucose, we mapped the +1 residue of gtsA, which corresponded to A located 96 bp upstream of the start ATG codon. The promoter exhibited relatively well-conserved −10 hexamers, but lacked a defined −35 hexamer, as is typical of promoters mediated by activators that respond to chemical signals (9, 13, 34).

DISCUSSION

When Pseudomonas putida is exposed to glucose, the sugar—upon reaching the periplasm—can either be internalized and subsequently metabolized via glucokinase/glucose-6-phosphate dehydrogenase to 6PG or converted into 2KG in the periplasmic space via glucose dehydrogenase and gluconate dehydrogenase. Gluconate and 2KG are transported to the cytoplasm and converted into 6PG (Fig. 1). The scattered distribution of the cluster of genes for each of the three pathways along the chromosome of P. putida suggests that each of the metabolic pathways has been acquired independently, and the presence of genes that encode catabolic enzymes and regulators was indicative of the potentially independent regulation of each segment. In agreement with this proposal are a series of previous studies that indicated that the glucokinase pathway was under catabolite repression control when cells were exposed to alternative C sources such as toluene or succinate, according to a process mediated by the global Crc regulator (5, 41), whereas neither of the other two pathways was under catabolic repression (5).

The set of genes for the glucokinase pathway cover a cluster consisting of PP1009 to PP1024. There are two potential transcriptional regulators within this cluster: GltR-2 and HexR. The set of catabolic genes not only includes the glucokinase pathway, but also includes the Entner-Doudoroff edd and eda genes in a peculiar organization: the edd gene is cotranscribed with glk, and eda is cotranscribed with zwf-1. Divergently with respect to edd is gap-1, which encodes glyceraldehyde-3-phosphate dehydrogenase. The cluster also includes an operon that encodes an ABC transport system mediating the ATP-dependent transport of glucose into the periplasmic space. All of these genes except for hexR are inducible by glucose (4).

In Pseudomonas aeruginosa, Sage et al. (32) reported that the gltR gene encoded a product homologous to the response element of two-component systems, whose disruption caused the loss of glucose transport activity. Inactivation of the gltR-2 gene in P. putida had an effect on growth in the presence of glucose that resulted in a prolonged lag when cells were transferred, for example, from M9 minimal medium with citrate to glucose-containing medium as the sole C source. This seems to be a consequence of (i) the lack of induction of the gtsABCD genes that encode the active transport system in P. putida, as shown by our transcriptional fusion assays of a gtsA::′lacZ construct in the gltR-2 mutant background; and (ii) the lack of induction of pgl encoding 6-phosphogluconolactonase, an enzyme required for full operation at the highest rate of the glucokinase pathway (10, 26). It should be noted that pgl is under the control of HexR (Table 3) and GnuR (Table 2), which indicates that the gene can be controlled from the zwf-1 promoter under HexR control and from its own (as yet unidentified) promoter under GnuR control. This is somewhat surprising, but it may be related to the need to induce pgl and eda for the efficient catabolism of gluconate/2KG when these chemicals are used as the sole C source, as is also the case in P. fluorescens (27) and in Escherichia coli (8, 22).

Deficiency in HexR resulted in the constitutive expression of the gap-1 gene and the edd-glk and eda-zwf-1 operons. This suggests that HexR is a transcriptional repressor of these operons. This observation is in agreement with findings by Phibbs and colleagues in P. aeruginosa (10). In the hexR mutant background, expression of the ABC glucose transport system comprising gtsABCD was under the control of the GltR-2 protein.

Our β-galactosidase assays with a P_edd::lacZ fusion revealed that HexR exhibits a relatively broad spectrum of effectors responding not only to glucose, but also to gluconate, 2KG, and fructose. Expression was equally high with all of these carbohydrates (T. del Castillo, unpublished observation). This is of physiological significance since the edd/eda gene products are required for P. putida to grow on gluconate, 2KG, and fructose. Examples of transcriptional regulators that recognize a broad range of effectors are not common in the HexR/LacI family, but some repressors of the TetR family (29) and the IclR family of regulators (21) have been reported to exhibit broad spectra. HexR also controls the expression of glyceraldehyde-3-phosphate dehydrogenase, the enzyme that acts on one of the final metabolites of the Entner-Doudoroff pathway.

The tsp of the three glucose catabolism transcriptional units regulated by HexR were mapped. This allowed us to identify the corresponding promoters, which exhibited −10 and −35 hexamers by RNA polymerase with sigma-70. In P. aeruginosa, hexR controls the expression of gap-1 (24, 36) like it happens in P. putida, in which an almost identical dyad to that recognized by P. aeruginosa in hexR was found covering the −10 region of the three promoters under HexR control. This dyad element was also found upstream from the first ATG of the ORF encoding PP4488.

In E. coli, the edd and eda genes are cotranscribed and regulated by a member of the GntR family of repressors that responds primarily to gluconate (37). del Castillo et al. (4) suggested that the glucokinase pathway might contribute up to 40% of the glucose income in P. putida, with the 2KG loop representing the metabolism of 40% to 45% of the glucose. The set of genes involved in the initial metabolism of gluconate to 2KG forms an operon (PP3384 to PP3382), whereas the genes for 2KG transport (PP3377) and its metabolism (PP3379, PP3378, and PP3376) form another operon that is transcribed from its own promoter. The ptxS gene is located downstream of the kguE gene (Fig. 2). Inactivation of the ptxS gene led to specific up-regulation of the promoters upstream from PP3384 and PP3379, as well as up-regulation of the ptxS gene (PP3380) itself. In P. aeruginosa, ptxS has been identified as involved in the control of the toxA gene and of its own synthesis (35, 40). In P. aeruginosa, the target of PtxS is a 14-bp dyad sequence whose disruption in front of this gene resulted in its overexpression. We identified the same 14-bp dyad element (5′-TGAAACCGGTTTCA-3′) within the 35 nucleotides 5′ upstream from each of the promoter start sites and suggest that this motif is also the target of PtxS in P. putida.

In Pseudomonas putida, the direct phosphorylation of gluconate to 6PG is of minor importance in metabolic terms, which contrasts with the situation in other gram-negative (8, 22, 25, 27, 37) and gram-positive (Corynebacterium glutamicum [15] or Bacillus subtilis [43]) bacteria. However, the basic mode of regulation of gluconate metabolism in P. putida is similar to that in other bacteria. In P. putida, the GnuR repressor is highly homologous to the GntR repressor of E. coli, B. subtilis, and Corynebacterium and controls the expression of the gluconate transporter gene and glucokinase gene. In E. coli and B. subtilis, expression of the gluconokinase is under global regulation, which is overimposed on local specific regulation (37, 43). This level of complexity is not found in P. putida, probably due to the minor role of gluconokinase in the assimilation of glucose in this bacterium. The GnuR protein seems to recognize a dyad element, as is the case for the two other regulators of the glucose operons.

In summary, our results indicate that we have identified four transcriptional regulators involved in glucose catabolism. HexR controls the flux of glucose to 6PG and further down to pyruvate and glyceraldehyde-3-phosphate, due to the link between the edd and eda genes of the Entner-Doudoroff pathway and the glk and zwf-1 genes, as discussed above. Concerning this pathway, it is worth noting that the positive transcriptional regulator GltR-2 controls the expression of the glucose transport system that allows the internalization of glucose into the cytoplasm. PtxS controls the expression of genes that encode gluconate dehydrogenase and which convert gluconate into 2KG and the set of genes whose products convert 2KG into 6PG. GnuR is another repressor that specifically responds to gluconate and controls its transport and phosphorylation. This intricate set of regulators guarantees the simultaneous channeling of glucose via three pathways that converge at the level of 6PG.