Dual regulation of zwf-1 by both 2-keto-3-deoxy-6-phosphogluconate and oxidative stress in Pseudomonas putida

In Pseudomonas species, the Entner–Doudoroff (ED) pathway is used to metabolize glucose. Three pathways of initial glucose breakdown to 6-phosphogluconate (6PG) have been identified in Pseudomonas spp. (Fig. 1; del Castillo et al., 2007b; Temple et al., 1994). The first of these pathways is the glucose kinase pathway, in which glucose is transported into the cytoplasm by the glucose transport system. The second is the gluconokinase pathway, in which gluconokinase mediates the direct phosphorylation of gluconate or the oxidation of glucose to gluconate. Finally, there is the ketogluconate pathway, in which the formation of 2-ketogluconate is catalysed primarily by two consecutive periplasmic oxidation reactions involving glucose dehydrogenase, encoded by the gcd gene, and gluconate dehydrogenase, encoded by the gad gene (Fig. 1). Through these pathways glucose is converted to 6PG, which is further metabolized into glyceraldehyde 3-phosphate and pyruvate by two ED enzymes, 6PG dehydratase (Edd) and 2-keto-3-deoxy-6-phosphogluconate (KDPG) aldolase (Eda) (Fig. 1; Cuskey et al., 1985; del Castillo et al., 2007b).

(26K): Fig. 1. Carbohydrate metabolism in P. putida involves the oxidation of glucose to gluconate and 2-ketogluconate, and occurs in the periplasmic space. PEP, phosphoenolpyruvate; OM, outer membrane; PG, periplasmic space; IM, inner membrane.

The genes for glucose metabolism in Pseudomonas putida are organized as several operons and are regulated by specific transcriptional regulatory proteins including HexR, GltR-2, PtxS and GnuR (del Castillo et al., 2008). GltR-2 participates in the glucose transport system, PtxS regulates gluconate oxidation and GnuR controls the transport and phosphorylation of gluconate (del Castillo et al., 2008).

HexR regulates two operons containing the zwf (glucose-6-phosphate dehydrogenase; G6PDH), pgl (6-phophogluconolactonase) and eda genes; and the edd, glk (glucokinase) and gltR-2 (DNA binding response regulator GltR) genes (del Castillo et al., 2008; Hager et al., 2000; Petruschka et al., 2002; Temple et al., 1994). HexR also regulates the expression of glyceraldehyde-3-phosphate dehydrogenase, encoded by the gap gene (del Castillo et al., 2008), and controls zwf in the ED pathway of Pseudomonas aeruginosa (Hager et al., 2000). The central intermediate of glucose metabolism, 6PG, acts as an inducer in P. aeruginosa as well as in P. putida (Petruschka et al., 2002; Temple et al., 1990). However, this study showed that KDPG, and not 6PG, induces HexR in glucose metabolism.

In many bacterial strains, the zwf gene is induced in response to oxidative stress (Lundberg et al., 1999; Park et al., 2006; Pomposiello & Demple, 2001; Sung & Lee, 2007). NADPH, produced by G6PDH activity, is needed for reductive metabolic pathways and oxidative stress-damage repair reactions (Girò et al., 2006; Lundberg et al., 1999; Singh et al., 2005). Deletion of the zwf gene increases bacterial sensitivity to oxidative stresses (Girò et al., 2006; Lundberg et al., 1999; Ma et al., 1998). The zwf gene is activated by the SoxRS regulon of Escherichia coli (Demple, 1996; Greenberg et al., 1990; Wu & Weiss, 1992). SoxR, a MerR-family transcription factor containing redox-active [2Fe–2S] centres, controls the SoxR regulon. SoxR is activated upon [2Fe–2S] oxidation and induces transcription of the divergently transcribed soxS gene (Hidalgo et al., 1997). The zwf gene is one of the targets of SoxS in E. coli (Girò et al., 2006; Pomposiello & Demple, 2001; Tsaneva & Weiss, 1990; Li & Demple, 1994) and is its direct transcriptional activator. The P. putida genome does not contain a clear SoxS homologue. We have previously shown that regulation of superoxide stress in P. putida differs from the SoxR paradigm previously described in E. coli (Park et al., 2006). The zwf-1 gene of P. putida KT2440 is strongly induced by oxidative stress reagents such as paraquat (PQ), menadione (MD) and nitric oxide, but its induction is not controlled by the SoxR system (Park et al., 2006). Here, we show that HexR is a transcriptional repressor responding to KDPG in the context of glucose metabolism, and that it might also be a key regulator of the zwf-1 gene under conditions of oxidative stress.

Bacterial strains, culture conditions, and DNA manipulation.
Bacteria (Table 1) were grown at 37 °C (E. coli) or 30 °C (P. putida) in Luria–Bertani (LB) or mineral salts medium (M9) containing different carbon sources at the following concentrations: 10 mM succinate, 10 mM glucose, 10 mM gluconate, 10 mM fructose, 10 mM pyruvate or 0.5 % (v/v) glycerol. Plasmid isolation, gel electrophoresis, transformation and PCR were performed using standard procedures (Ausubel et al., 1999). Tetracycline (15 µg ml^–1) and kanamycin (50 µg ml^–1) were added to bacterial cultures when necessary. The edd and eda mutants were purchased from Bio-Iliberis R&D. The KDPG substrate was provided by Matthew J. Walters (Duke University).

Table 1. Bacterial strains and plasmids used in this study

Northern blot analysis.
Total RNA was isolated from 3 ml of exponentially growing cells using an RNeasy kit (Qiagen) following the manufacturer's instructions. RNA concentration was determined by measuring A₂₆₀ using 4–5 µg total RNA per sample. The fractionated RNA was transferred to a nylon membrane (Schleicher & Schuell) using a Turboblotter (Schleicher & Schuell), and the amounts of zwf-1, zwf-2, zwf-3 and edd mRNA were determined by hybridizing the membrane with α-³²P-labelled probe specific for each gene (TaKaRa Bio). To generate the specific probes, the following primers were used: kt zwf-1 Pp1, kt zwf-1 Pp2, kt zwf-2 Pp1, kt zwf-2 Pp2, kt zwf-3 Pp1, kt zwf-3 Pp2, kt edd Pp1 and kt edd Pp2 (Table 2).

Table 2. Primers used in this study

Cloning procedures and construction of bacterial strains.
A broad-host-range promoter probe vector, pRK415gfp (Yin et al., 2003), was used to construct a reporter plasmid, pRKP_zwf-1gfp. Briefly, a 438 bp fragment from the zwf-1 promoter was amplified using primers kt zwf-1 pro2 and kt zwf-1 pro3. To generate pRKP_zwf-1gfp-HexR CP (a reporter plasmid containing the hexR plus the zwf-1 promoter regions for complementation), the kt zwf-1 pro1 and the kt zwf-1 pro2 primers were used. The amplicon was cloned into the EcoRI/BamHI cloning site of the pRK415gfp vector generating pRKP_zwf-1gfp. The constructed plasmid was then introduced into E. coli Top10 by electroporation. Then, the pRKP_zwf-1gfp plasmid was introduced by triparental conjugation into P. putida KT2440-R, thus creating P. putida KT2440 (pRKP_zwf-1gfp). Conjugation was performed by triparental filter mating using E. coli Top10 (pRKP_zwf-1gfp), E. coli HB101 (pRK2013) and P. putida KT2440-R as donor and recipient, respectively (Park et al., 2003).

To construct the hexR mutant, a 313 bp fragment of its internal region was amplified by PCR using the kt hexR KO1 and kt hexR KO2 primers. The fragment was cloned into the EcoRI cloning site of the pVIK112 vector (Kalogeraki & Winans, 1997), generating pVIK-HexR. The constructed plasmid was then introduced by electroporation into E. coli S17-1 λpir. Conjugation was performed by filter mating with E. coli S17-1 λpir (pVIK-HexR) and P. putida KT2440-R (Lee et al., 2006a) as the donor and recipient, respectively.

Quantification of GFP fluorescence.
Bacterial cells at the exponential growth phase (OD₆₀₀ ∼0.3) grown in M9 minimal medium supplemented with glucose, gluconate, glycerol, fructose, succinate or pyruvate, were collected using a microcentrifuge (15 800 g) and washed twice with PBS (137 mM NaCl, 10 mM phosphate, 2.7 mM KCl, pH 7.4). Then, both the OD₆₀₀ and GFP fluorescence intensity of the resuspended cultures were quantified using a microtitre plate reader (Victor3, Bio-Rad). This reporter strain expresses a stable GFP variant that absorbs light at 488 nm.

Overexpression of His–HexR.
HexR was cloned from the genomic DNA of P. putida KT2440 by PCR (Kt HexR-OE1/OE2 primer pair used; Table 2) and inserted into pET28a(+). The recombinant plasmid was introduced into E. coli BL21 to obtain expression of the recombinant plasmid pET28a(+)-HexR. The transformed E. coli BL21, harbouring pET28a(+)-HexR, was cultured overnight in 5 ml LB medium supplemented with kanamycin at 37 °C with shaking at 220 r.p.m. The cells were then diluted 100-fold in 20 ml of the same medium and incubated for 3 h. IPTG (0.5 mM) was added, and the cells were incubated for an additional 4 h. The cells were harvested, washed twice with PBS, resuspended in Tris-Cl buffer (pH 7.5, containing 1 mM DTT), and lysed by sonication. Crude extracts were clarified by centrifugation at 10 000 g for 10 min at 4 °C, and the protein concentration was determined by the Bradford method using BSA as the standard.

Electrophoretic mobility shift assays (EMSAs).
The P_zwf-1 DNA probe was generated by PCR amplification using the zwf-1 sp. pro1/zwf-1 sp. pro2 primer pair (for the 136 region) and the zwf-1 sp. pro2/zwf-2 sp. pro3 primer pair (for the 221 region). The PCR product was dephosphorylated and labelled with [γ-³²P]ATP and T4 polynucleotide kinase. The reaction mixture (20 µl final volume), containing the P_zwf-1 probe, crude extract and loading buffer in 5x binding buffer (50 mM Tris, pH 7.5, 10 mM MgCl₂, 50 %, v/v, glycerol, 10 mM DTT, and 375 mM KCl), was incubated for 20 min at room temperature. The resulting complexes were analysed by electrophoresis on 4 % polyacrylamide gels in 0.5x Tris-borate/EDTA (TBE) buffer (5xTBE=1.1 M Tris, 900 mM borate, 25 mM EDTA, pH 8.3).

Oxidative stress sensitivity assay.
Cells were grown in liquid LB medium overnight and diluted 100-fold in the same medium. After 3 h further incubation, serially diluted cells were spotted on LB agar with or without PQ (7 µM), MD (1 mM) or arsenic (As) (10 p.p.m. arsenic oxide, Sigma).

Measurement of G6PDH activity.
Cells grown under each set of conditions were lysed by sonication in 50 mM Tris buffer (pH 7.5). G6PDH activity was monitored using a spectrophotometer (340 nm) to measure the production of NADPH in 1 ml reaction mixtures containing 2 mM glucose 6-phosphate, 0.3 mM NADP⁺ and the cell extract.

Induction of the zwf genes in the presence of various carbon sources
There are three putative homologues of zwf (zwf-1, zwf-2 and zwf-3) in the P. putida KT2440 genome (Fig. 2a, b, c). In the presence of various sources of carbon, we investigated their expression using Northern blot analysis. Total RNAs were isolated from 3 ml exponentially growing cells. The level of zwf-1 transcription, as determined by Northern blotting, increased dramatically when P. putida KT2440 was cultured in minimal medium containing either glucose or gluconate as the sole carbon source. However, in minimal medium supplemented with either pyruvate or succinate, the level of zwf-1 gene induction was very low and likely represented a basal level (data not shown). The expression levels of zwf-2 and zwf-3 were constitutive, and they remained low regardless of the carbon source (data not shown). These results indicate that zwf-2 and zwf-3 play minor roles in glucose metabolism, whereas the zwf-1 gene product is the main enzyme required for converting glucose 6-phosphate to 6-phosphoglucolactone. Induction of the zwf-1 gene occurred when glucose or gluconate metabolism was required, whereas the organic acids pyruvate and succinate had no apparent effect on zwf-1 induction.

(22K): Fig. 2. Genes involved in glucose catabolism in P. putida KT2440. The zwf-1 (a) gene homologues are zwf-2 (b) and zwf-3 (c). The complete genome sequence of P. putida KT2440 in the NCBI database was the source for the gene numbers and organization.

HexR represses zwf-1 induction
The mechanism of HexR regulation in P. putida KT2440 was investigated using a reporter plasmid that contained the gfp gene under the control of P_zwf-1. The reporter plasmid was introduced into either the wild-type or the HexR knockout mutant strain (HexR–) to generate P. putida KT2440 (pRKP_zwf-1gfp) and P. putida KT2440 (HexR–/ pRKP_zwf-1gfp). Consistent with the results of the Northern blot assay, the reporter assay indicated that P_zwf-1 activity increased significantly in minimal medium supplemented with either glucose or gluconate compared with its activity in minimal medium containing either pyruvate or succinate (Fig. 3). The GFP expression levels in minimal medium containing glucose or gluconate were 25–50-fold higher than those observed in pyruvate- or succinate-containing minimal medium. The P_zwf-1 activity in either glycerol- or fructose-containing minimal medium was much higher than in either the pyruvate- or succinate-containing medium, but it remained at only half of the activity level observed in the glucose-containing medium (Fig. 3). When these results are taken together, the conclusion appears to be that the zwf-1 gene is inducible, and that this occurs when metabolism of glucose, gluconate, glycerol or fructose is required. However, in the HexR– strain the GFP activity was constitutively high under all conditions (Fig. 3, black bars). The levels of P_zwf-1 activity were similar in wild-type and HexR-complementation strains (Fig. 3, white bars) under all conditions tested. Collectively, it appears that HexR regulates zwf-1 induction as a repressor.

(21K): Fig. 3. Quantification of GFP expression in wild-type and HexR– reporter strains grown in the presence of various carbon sources. GFP was measured as described in Methods. Grey bars, P. putida KT2440 (pRKPzwf-1gfp); black bars, HexR– (pRKPzwf-1gfp); white bars, HexR– (pRKPzwf-1gfp-HexR CP).

To investigate whether HexR binds to the zwf-1 promoter region, EMSAs were conducted using a P_zwf-1 DNA probe (–136 to +34 of the zwf-1 promoter region) with cell extracts from E. coli BL21 harbouring a HexR-overexpression vector. The EMSA results demonstrated that a crude extract of E. coli BL21 [pET28a(+)-HexR] cells retards the migration of the P_zwf-1 DNA probe (Fig. 4a). The EMSA data indicated that a HexR operator site might lie in the zwf-1 promoter-specific region.

(22K): Fig. 4. (a) EMSA of the zwf-1 promoter region probe with cell extracts from BL21 (HexR). Lane 1, free zwf-1 promoter region probe; lanes 2–6, 0.25, 0.5, 1, 5 and 10 µg BL21 (HexR) cell extract, respectively; lane 7, BL21 cell extracts harbouring pET28. (b) Schematic representation of the 5'-deleted zwf-1 promoter region fused to the gfp reporter plasmid. The transcription start point is taken from del Castillo et al. (2008). (c) Quantification of GFP expression in wild-type and HexR-derivative reporter strains grown with gluconate or succinate.

The zwf-1 promoter region contains a HexR operator site
To examine the HexR binding regions, the 221 P_zwf-1, 136 P_zwf-1, 67 P_zwf-1 and 19 P_zwf-1 variants of the zwf-1 promoter were constructed (kt zwf-1 sp. pro3, kt Pzwf-1 pt1, pt2, pt3 as upper primer, respectively, and kt zwf-1 pro2 as a lower primer; Table 2, Fig. 4b). Potential binding regions of differing sizes were cloned into the gfp reporter vector (pRK415gfp) and were introduced into the wild-type and HexR– strains. The 221 P_zwf-1 yielded the highest levels of GFP expression in gluconate-containing minimal medium, whereas GFP did not accumulate in the succinate-containing minimal medium (Fig. 4c). The 136 P_zwf-1 behaved similarly to 221 P_zwf-1 (Fig. 4c). However, the 136 P_zwf-1 activity was twofold lower than that of 221 P_zwf-1 in gluconate or succinate, regardless of the presence of HexR (Fig. 4c). Due to the absence of the RNA polymerase binding site, there was no 19 P_zwf-1 activity (Fig. 4c). Interestingly, we found that the reporter strain containing 67 P_zwf-1 expressed GFP in the succinate medium (Fig. 4c), and similar levels of GFP were observed in the HexR– reporter strain carrying 67 P_zwf-1 in minimal medium supplemented with either succinate or gluconate (Fig. 4c). Together, these data provide evidence that: (1) the region between –67 and –136 is important for HexR binding, (2) 67 P_zwf-1 is not involved in the regulation of HexR, and (3) the region far upstream of the HexR binding site might be essential for full expression of the zwf-1 gene.

KDPG is a real inducer of Hex operons
It has been suggested that 6PG is the inducer of the Hex operon (Petruschka et al., 2002; Wu & Weiss, 1992). However, our data show that 6PG does not inhibit the binding of HexR to the zwf-1 promoter region. Instead, KDPG influences the binding of HexR to its cognate promoter region (Fig. 5a). The EMSA assay showed that binding of HexR to the zwf-1 promoter region is blocked by KDPG but not by 6PG.

(33K): Fig. 5. (a) EMSA of the Pzwf-1 probe using BL21 (HexR) crude extracts (5 µg), KDPG or 6PG. The first lane contains the free zwf-1 promoter region probe. (b) Pzwf-1 activity in wild-type, Edd and Eda mutant strains. Each strain was precultured in 10-fold-diluted LB medium until growth reached the exponential phase (OD600 0.3–0.4), and cells were harvested, washed with PBS, and divided into gluconate or succinate medium. GFP intensity was determined after 1 h.

The edd and eda deletion strains were used to confirm that KDPG is a real inducer of HexR in vivo. The edd mutant cannot make KDPG, and the eda mutant accumulates KDPG inside cells. Wild-type cells and mutants bearing the pRKP_zwf-1gfp reporter plasmid were precultured in 10-fold diluted LB medium until they reached exponential growth phase and were transferred to minimal medium containing either gluconate or succinate. After a 1 h incubation, we measured the level of GFP induction. The P_zwf-1 activity of the eda mutant increased dramatically even in the succinate-amended conditions (Fig. 5b). However, the GFP activity of the edd mutant remained low in the gluconate-amended medium (Fig. 5b). As such, it appears that KDPG is the inducer of zwf-1 expression.

The zwf-1 gene is also induced by an oxidative stress condition
Bacteria may protect themselves against oxidative stress by using enzymes that include superoxide dismutase, catalase, glutathione/thioredoxin peroxidase and/or G6PDH to scavenge reactive oxygen species (ROS) (Halliwell & Gutteridge, 1999). We have previously shown that the zwf-1 gene in P. putida KT2440 is induced by oxidative stress in response to reagents such as PQ or MD (Park et al., 2006). To investigate the response of each of the zwf gene homologues to oxidative stress, Northern blot analysis was performed in the presence of PQ, MD, As or cumene hydroperoxide (CHP). Each of these reagents is able to produce superoxide or hydroxyl radicals under aerobic conditions (Gant et al., 1998; Parvatiyar et al., 2005). Under these conditions, the zwf-2 and zwf-3 induction levels were negligible (Fig. 6a); however, the zwf-1 gene was highly expressed under all conditions except in the presence of PQ (Fig. 6a, b). The zwf-1 gene was the only oxidative stress-related gene that responded to these reagents. Similar levels of induction were observed when the G6PDH activities were measured after a 30 min exposure to PQ (1 mM), MD (1 mM) or As (10 p.p.m.). With As and MD, the activity (U mg^–1) was 0.6±0.05 and 0.517±0.02, respectively. The level of G6PDH activity was similar with PQ (0.375±0.04) and under control (0.375±0.01) conditions, probably because of the low level of zwf-1 expression.

(64K): Fig. 6. Induction of the zwf homologue genes (zwf-1, zwf-2 and zwf-3) exposed to oxidative stress. Total mRNA was extracted 1, 5 or 10 min after treatment of P. putida KT2440 with (a) 1 mM PQ, 0.5 mM MD, 10 p.p.m. As or (b) 5 mM CHP. C, control.

HexR may directly detect oxidative stress
To demonstrate the existence of an additional activation system, EMSA was conducted using crude extracts of wild-type and HexR– cells. Proteins in the HexR– cell crude extracts (25 µg) did not bind to the zwf-1 promoter region, and gel-shift retardation occurred in the crude extracts from wild-type (25 µg) and BL21 (HexR) (and HexR– cells) (Fig. 7a). When the full-length promoter regions were used to rule out the possibility of an upstream activator protein, we observed the same result (Fig. 7a). If there was an additional system for zwf-1 gene induction, then during exposure to oxidative stress the level of zwf-1 transcription in the wild-type strains would exceed that of the HexR– strain. However, Northern blot analyses indicated that the levels of zwf-1 induction in the HexR– strain were almost identical in the presence or absence of oxidative stress reagents (Fig. 7b). Thus, it is highly probable that the only operative oxidative stress system involves HexR-regulated zwf-1 activation. These results suggest that HexR is the only protein able to bind to the zwf-1 promoter region, and that HexR is a good candidate for the oxidative stress-sensing regulator of the zwf-1 promoter.

(46K): Fig. 7. Investigation of the zwf-1 gene activation system. (a) EMSA of the Pzwf-1 probe (221 and 136 bases upstream of the zwf-1 transcription start point) using cell extracts from wild-type (25 µg), HexR– (25 µg) and BL21 (HexR) (10 µg) strains. (b) Northern blot analysis of the zwf-1 gene in HexR– bacteria exposed to 1 mM PQ, 0.5 mM MD and 10 p.p.m. As. Cells were exposed to each substrate for 1 min. (d) Northern blot analysis of the zwf-1 gene in wild-type cells grown in minimal medium with succinate then exposed to As (1, 5 and 10 p.p.m.) for 1 and 10 min. (e) Northern blot analysis of the edd gene in wild-type cells that were exposed to 0.5 mM MD, 10 p.p.m. As or 5 mM CHP. Cells were exposed to each substrate for 1 or 10 min (CHP). (c, f) Northern blot analysis of the zwf-1 (c) or edd (f) genes in different mutant strains exposed to As (10 p.p.m.) for 1 min. C, control.

We also investigated the indirect induction of the zwf-1 gene by HexR, under the conditions of oxidative stress. If KDPG were overproduced, then high levels of zwf-1 expression could occur, through an unknown mechanism, under conditions of oxidative stress. To confirm this possibility, Northern blot analysis was performed using the ΔEdd and ΔEda strains. The ΔEdd strain expressed the zwf-1 gene in response to As exposure in the absence of KDPG (Fig. 7c). Consistent with the results of the GFP-based reporter assay, a high level of zwf-1 expression was observed in the ΔEda strain cultured in LB medium with or without As exposure (Fig. 7c). Thus, it seems unlikely that KDPG is required for zwf-1 gene induction under oxidative stress conditions. This observation also confirmed that there was no other activation system for zwf-1 induction in the presence of oxidative stress reagents (Fig. 7c). This observation is further supported by the fact that wild-type cells exhibited higher zwf-1 expression in response to oxidative stress in the succinate-amended minimal medium (Fig. 7d).

Recently, del Castillo and colleagues have reported that HexR controls edd as well as gap-1 in P. putida (Fig. 1; del Castillo et al., 2008). This implies that the edd and gap-1 promoter regions contain HexR operator sites, and that the edd and gap-1 genes are regulated by the HexR repressor. If HexR is a direct sensor of oxidative stress, then there should be an increase in edd expression with oxidative stress. The results revealed a high level of edd gene expression under oxidative stress conditions (Fig. 7e). As in HexR- and KDPG-related zwf-1 gene induction, the edd gene was also induced in the HexR– and ΔEda strains due to the absence of the HexR repressor and the presence of the HexR inducer, respectively (Fig. 7f). It seems that the induction of oxidative stress via the edd promoter region occurs as a result of more complex mechanisms.

In P. aeruginosa PAO1, a locus designated hexC (129 bp, cis-element) exists between the gap and edd genes (Temple et al., 1994). This region reportedly contains binding sites not only for HexR but also for the integration-host factor (IHF; Proctor et al., 1997). We conducted an in vitro EMSA analysis to examine whether HexR is a direct sensor of oxidative stress. The data showed that binding of HexR to its cognate promoter was inhibited by CHP (50 mM) and MD (20 mM; data not shown). Although these concentrations may not be physiologically relevant, we chose them because high concentrations of other non-specific proteins and reagents such as hydroperoxide (500 mM) and tert-butyl hydroperoxide (500 mM) did not inhibit HexR binding in crude extracts (data not shown). Although the precise mechanism remains unknown, we have provided clear evidence that HexR may directly detect oxidative stress.

HexR– deletion improves tolerance to oxidative stress
The tolerance to oxidative stress was compared in wild-type and HexR cells. In the presence of MD or PQ, the HexR– strain was less sensitive than the wild-type cells (Fig. 8). The HexR– strain might be more resistant to oxidative stress due to the constitutive activity of G6PDH. However, exposure to As did not result in differential sensitivity of the wild-type and HexR– strains to stress (Fig. 8). It is possible that zwf-1 expression in response to As exposure might be sufficiently rapid and strong to defend against oxidative stress (Fig. 6).

(56K): Fig. 8. Sensitivity of wild-type and HexR– strains to oxidative stress. The strains were cultured until they reached the exponential growth phase, serially diluted and spotted onto LB agar supplemented with various inducers of oxidative stress.

G6PDH is required in the ED pathway of Pseudomonas species. NADPH, produced by G6PDH, is essential for defence against oxidative stress. Interestingly, the P. putida KT2440 genome contains three homologues of zwf. The amino acid identities of zwf-1 to zwf-2 and of zwf-1 to zwf-3 are 36.8 % and 48.2 %, respectively. Northern blot analysis showed that zwf-2 and zwf-3 do not play major roles in the ED pathway, but that the zwf-1 gene is a key enzyme in this pathway and is also important for defence against oxidative stress (Fig. 6a, b). Our results showed that, under conditions of oxidative stress, the zwf-1 gene is induced, whereas the levels of zwf-2 and zwf-3 expression do not change (Fig. 6a). Although the NCBI database shows that various micro-organisms contain more than one zwf gene, the functions of the additional genes and of G6PDH remain unclear. Nevertheless, only zwf-1 gene products responded to glucose metabolism in P. putida. Deinococcus radiophilus carries the G6PDH-1 and G6PDH-2 isoforms of G6PDH (Sung & Lee, 2007). When D. radiophilus was treated with potassium superoxide, the activity of G6PDH-1 increases markedly, but the G6PDH-2 activity does not increase (Sung & Lee, 2007). The functions of zwf-2 and zwf-3 in P. putida remain to be investigated.

The data presented demonstrated that HexR binds to the promoter regions of the zwf-1 gene and functions as a repressor. Because the full-length promoter expresses zwf-1 at levels approximately twofold higher than those observed with promoters of various lengths, there may be more complex mechanisms of zwf-1 gene regulation (Fig. 4c). It may be that the upstream region (–221 to –136) contains cis elements that are important for the full induction of zwf-1 transcription. Many research groups have proposed that 6PG is an inducer of HexR; however, our results demonstrated that HexR binding is inhibited by KDPG rather than by 6PG (Fig. 5a). In P. aeruginosa, KDPG may physiologically induce repressed genes, including edd, gap and zwf (Temple et al., 1998). However, to our knowledge, there are no data to support this idea. Interestingly, it has recently been reported that KDPG, generated during glucose utilization, is also a chemical signal involved in toluene catabolic repression in P. putida KT2440 (del Castillo & Ramos, 2007a). Previous data, along with our data, have demonstrated that KDPG might be a very important regulatory intermediate for controlling efficient energy metabolism.

Because it was very difficult to purify the HexR proteins, we used crude extracts harbouring overexpressed HexR throughout our in vitro experiments. The difficulty of purifying HexR is caused by its complex polymer structure. Although we used crude extracts containing His-tagged HexR, the same results were obtained with crude extracts containing wild-type HexR (data not shown). We strongly believe that there is no difference between His-tagged HexR and wild-type HexR in terms of both their binding to the cognate DNA region and their sensing KDPG and oxidative stress. Although the structure of HexR has not been determined, the NCBI database shows that HexR proteins feature a helix–turn–helix domain in their N-terminal region and a sugar isomerase (SIS) domain in their C-terminal region. The SIS domain functions in phospho-sugar binding, and KDPG might bind to the HexR-SIS domain, alter its conformation, and inhibit its binding to DNA.

ROS, such as the superoxide anion radical and its derivatives hydrogen peroxide and the hydroxyl radical, can damage bacterial cells (Girò et al., 2006). To protect itself against superoxide and nitric oxide stresses, E. coli has the SoxR regulon (Demple, 1996; Park et al., 2006). Under normal conditions, the [2Fe–2S] centre in SoxR exists in a reduced state. If this iron–sulfur cluster is oxidized by ROS, the SoxR is activated, and SoxS expression is induced. In E. coli, SoxS is known to be a transcriptional activator for zwf expression. The defence role of G6PDH against oxidative stress has been described before (Ceccarelli et al., 2004; Lundberg et al., 1999). Although a SoxR homologue is present, SoxS is absent from P. putida, and little is known about the scavenging mechanism in Pseudomonas species that is induced by oxidative stress (Park et al., 2006). The fpr gene, which encodes ferredoxin-NADP⁺ reductase in P. putida, is involved in oxidative stress defence and its expression is regulated by FinR (Lee et al., 2006b). However, the expression of fpr in E. coli occurs under the control of SoxS. The system regulating zwf-1 induction under conditions of oxidative stress in P. putida remains unknown.

To our knowledge, this is the first report to demonstrate that HexR detects ROS and controls zwf-1 expression under the conditions of oxidative stress. Our conclusions are strongly supported by the findings that oxidative stress (1) induces the zwf-1 gene, despite the absence of intracellular KDPG; (2) inhibits HexR binding to its cognate promoter regions; and (3) strongly induces the edd gene, which does not appear to play a significant role in oxidative stress defence. Because some oxidative stress-inducing reagents strongly induced zwf-1 expression, but did not inhibit HexR binding to its cognate promoter, there may be a more complex mechanism through which HexR detects oxidative stress (data not shown).

This work was supported by a NCRC (National Core Research Center) grant (R15-2003-002-01002-0) and a Korea Science and Engineering Foundation (KOSEF) grant (R01-2008-000-10697-0) to W. P. This work was also supported by a Korea University Grant (2008). We thank Matthew J. Walters (Dr Eric J. Toone's laboratory, Department of Chemistry, Duke University) for providing the KDPG substrate.

Edited by: M. A. Kertesz