Transcriptional regulation of the sulfate-starvation-induced gene sfnA by a {sigma}54-dependent activator of Pseudomonas putida

Sulfur is an essential element for bacterial growth and is normally acquired through the assimilation of inorganic sulfate. However, in soil environments sulfate may not be freely available because inorganic sulfate constitutes less than 5 % of the total sulfur in these environments (Autry & Fitzgerald, 1990). In addition, most of the residual sulfur is present in the soil as peptides/amino acids, sulfate esters and/or sulfonates (Autry & Fitzgerald, 1990). Thus, several soil bacterial species have evolved the ability to use a wide variety of organosulfur compounds, and both the genes and enzymes involved in the use of such compounds (especially the alkanesulfonates, sulfate esters and taurine) have been studied in detail (see reviews by Kertesz, 2000, 2004). When the preferred inorganic sulfur compounds are not available, these bacteria produce a set of proteins to adapt to sulfur limitation (Kertesz et al., 1993; Coppée et al., 2001). These proteins are known as sulfate-starvation-induced proteins and include metabolic enzymes, regulatory proteins, and transport systems for scavenging sulfur from organosulfur compounds. This sulfate-starvation response plays an important role in the survival of bacteria in soil environments (Kahnert et al., 2002). Although the emission of dimethylsulfide (DMS) from soils has been reported (Minami, 1982), very little is known about the mechanism of the sulfate-starvation response in the use of organosulfides.

We isolated Pseudomonas putida strain DS1, which uses DMS as a sulfur source and desulfurizes it via DMSO, dimethyl sulfone (DMSO₂), and methanesulfonate (MSA; Fig. 1a; Endoh et al., 2003a). As reported for other Pseudomonas and Escherichia coli strains (Eichhorn et al., 1999; Kahnert et al., 2000; van der Ploeg et al., 1999), desulfonation of MSA to provide inorganic sulfur for growth is catalysed by SsuD (FMNH₂-dependent monooxygenase) with an NAD(P)H-dependent FMN reductase (Endoh et al., 2003a). The ssuD gene of DS1 is located within the ssu operon (ssuEADCBF; Endoh et al., 2003a). An FMNH₂-dependent monooxygenase, SfnG, which is involved in the conversion of DMSO₂ to MSA, has also been identified, and the sfnG gene is located within the sfnFG operon (Endoh et al., 2005). Interestingly, the expression of the sfnFG operon is controlled by a novel σ⁵⁴-dependent transcriptional regulator, SfnR, that is similar to NtrC-type regulators, but lacks the N-terminal phospho-receiver domain (Endoh et al., 2003b, 2005).

(14K): Fig. 1. (a) Proposed dimethyl sulfide (DMS) desulfurization pathway in P. putida strain DS1. DMS is thought to be oxidized in two steps by unidentified monooxygenases via the conversion of dimethyl sulfoxide (DMSO) to dimethyl sulfone (DMSO2). DMSO2 is then converted to methanesulfonate (MSA) by the FMNH2-dependent monooxygenase, SfnG. Sulfite is finally released from MSA by the FMNH2-dependent monooxygenase, SsuD. The reduced FMN required for the activity of SfnG and SsuD could be provided by the FMN reductase, SfnE. (b) Organization of the sfnAB–sfnFG gene cluster of P. putida strain DS1. Pentagonal boxes indicate the length and direction of the sfn genes. Dots with numbers (1, 2 and 3) represent the location of the three SfnR-binding sites (sites 1, 2 and 3, respectively), which were previously identified by DNase I footprinting (Endoh et al., 2005). Double-headed arrows indicate the locations of the DNA fragments amplified by RT-PCR (see text and Fig. 2).

Transcriptional regulation of the sulfate-starvation-induced genes has been extensively investigated in Pseudomonas aeruginosa, as well as in E. coli, and the LysR-family regulator CysB plays a key role in the expression of genes involved in both the cysteine biosynthetic pathway and organosulfur assimilation (e.g. the ssu operon; Delic-Attree et al., 1997; Hummerjohann et al., 2000). Four other LysR-family regulators of organosulfur metabolism, SdsB (Davison et al., 1992), AsfR (Vermeij et al., 1999), SftR (Kahnert et al., 2002) and SsuR (Iwanicka-Nowicka et al., 2007), have also been discovered in Pseudomonas (or Burkholderia) strains. However, before our report on SfnR (Endoh et al., 2003b), there had been no reports that this σ⁵⁴-dependent transcriptional regulator was involved in sulfur assimilation (Cases et al., 2003). Thus, this was apparently the first study to suggest that at least two different sulfate-starvation response mechanisms function in P. putida for organosulfide or organosulfone assimilation; i.e. (1) conversion of DMSO₂ to MSA, controlled by a σ⁵⁴-dependent transcriptional regulator, and (2) conversion of MSA to sulfite, controlled by a σ⁷⁰-dependent LysR-family regulator.

SfnR binds three DNA regions (sites 1, 2 and 3) upstream of the sfnFG operon (Fig. 1b; Endoh et al., 2005). Only the site 1 region, which is proximal to the sfnF gene and contains the two overlapping imperfect inverted repeat sequences (IIRSs), is necessary for the SfnR-dependent expression of the sfnFG operon (Fig. 1b; Endoh et al., 2005). However, the functions of the other SfnR-binding sites (sites 2 and 3) remain unknown. Because these SfnR-binding sites are located in the sfnF–sfnA intergenic region (Fig. 1b), it is possible that sites 2 and 3 may be involved in the expression of the sfnAB operon. If the sfnA and sfnB genes, both of which exhibit sequence identity with the acyl-CoA dehydrogenase family protein, are indeed targets of a σ⁵⁴-dependent transcriptional regulator, the function of these genes in the assimilation of organosulfur compounds warrants further research. Therefore, we conducted a transcriptional analysis of the sfnAB operon to identify other σ⁵⁴-dependent organosulfur-assimilating genes in addition to the sfnFG operon.

Bacterial strains, plasmids and culture conditions.
The bacterial strains and plasmids used are listed in Table 1. Pseudomonas putida DS1 was cultivated with various sulfur sources using established methods (Endoh et al., 2005). E. coli strains for cloning were cultivated at 37 °C using LB broth, 2xYT medium, or Terrific broth, as described by Sambrook & Russell (2001). Ampicillin (Ap), tetracycline (Tet), gentamicin (Gm), kanamycin (Km), chloramphenicol (Cm), and IPTG were added to the media (when necessary) at final concentrations of 50, 50, 15, 50 and 30 µg ml^–1, and 0.1 mM, respectively. In some cases, Km and Gm chloride were prepared from sulfate salts using ion exchange (van der Ploeg et al., 1996). Briefly, the sulfate ion of commercially available Km and Gm was replaced by chloride by passage over the Cl^– form of Amberlite IRA400 J Cl resin, elution with water and lyophilization.

Table 1. Bacterial strains and plasmids

DNA manipulation.
Total DNA preparation, plasmid isolation, restriction enzyme digestion and transformation of E. coli were performed as described elsewhere (Sambrook & Russell, 2001). The DNA ligation kit version 2 (Takara) and EZNA gel extraction kit (Omega Bio-tek) were used according to the manufacturer's instructions. Other commercially available enzymes and kits were used as indicated by the respective manufacturers. P. putida was transformed by electroporation in a 0.1 cm cuvette (25 µF, 200 Ω, 17 kV cm^–1) using a GenePulserII apparatus (Bio-Rad). DNA sequencing was conducted using an ABI PRISM 310 Genetic Analyzer (Applied Biosystems).

Reverse transcription (RT) PCR analysis.
Strain DS1 and Dfi74 J (sfnR disruptant with Tn5) were grown to early exponential phase in sulfur-free mineral medium (SFMM) (Omori et al., 1995) supplemented with either sulfate or MSA (to final substrate concentrations of 1 mM) as a sulfur source. A 2 ml volume of the respective cultures was then centrifuged, and total RNA from the harvested cells was extracted using an RNeasy Mini kit (Qiagen) or NucleoSpin RNA II (Macherey-Nagel), combined with RQ1 RNase-Free DNase (Promega) according to the manufacturer's instructions. RT-PCR was performed with a TaKaRa One Step RNA PCR kit (AMV) (Takara Shuzo). The gene-specific reverse primers used to synthesize cDNA were ssuE-RV for ssuE, sfnA-RV for sfnA, and sfnB-RT-r for the sfnA–sfnB intergenic region (Table 2). The reaction mixture (50 µl) contained 5 µl 10x One Step RNA PCR buffer, 5 µl MgCl₂, 5 µl dNTP, 40 U RNase inhibitor, 5 U AMV RTase XL, 5 U AMV-optimized Taq, 1 µl each forward and reverse primer, and 1 µg total RNA prepared as described above. PCR was performed using a PCR Thermal Cycler Dice (Takara Shuzo) as follows: 50 °C for 30 min and 94 °C for 2 min, followed by 25 cycles of 94 °C for 30 s, 60 °C for 30 s, and 72 °C for 1.5 min.

Table 2. Primer sequences used

Primer extension analysis.
For the construction of pMEsfnA-PE, the DNA region containing the 5'-terminal end of sfnA and upstream sequences was amplified with a primer set (sfnA-FW-1 and sfnA-RV-PE; Table 2), using pEN18 (Endoh et al., 2005) as a template. Each amplified fragment was cloned into a pT7Blue (R) vector (Novagene), and the nucleotide sequences of each reporter plasmid were confirmed. Clones were digested with both HindIII and EcoRI (sites incorporated in the primers; Table 2), and the fragments were then cloned between the HindIII and EcoRI sites of pMElacZ (Endoh et al., 2003b) to produce pMEsfnA-PE. Total RNA (5 µg) from DS1 (pMEsfnA-PE) grown on MSA as a sulfur source was subjected to an RT reaction using SuperScript III reverse transcriptase (Invitrogen) and a primer, sfnA-PE1 (Table 2), whose 5' end was labelled with IRD₈₀₀ (Aloka). The primer extension products were purified using phenol/chloroform extraction and ethanol precipitation. The products were subjected to electrophoresis together with a sequencing reaction using the same primer and a Li-Cor model 4200I-2 Auto DNA sequencer running Base ImaglR data collection software 4.0 (LI-COR), according to the manufacturer's instructions.

β-Galactosidase reporter assay.
For the construction of a series of sfnA'–lacZ transcriptional fusion plasmids (pMEPsfnA-328, pMEPsfnA-188, pMEPsfnA-138, pMEPsfnA-108, pMEPsfnA-58, pMEPsfnA-138SM1, pMEPsfnA-138SM2, and pMEPsfnA-138DM3), the regions upstream of sfnA were amplified using a reverse primer (sfnA-RVrep) and a series of forward primers (sfnA-FW-1 to sfnA-FW-5; sfnA-FW-3M1 to sfnA-FW-3M12; Table 2), using pEN18 (Endoh et al., 2005) as a template. Each amplified fragment was cloned into a pT7Blue (R) vector (Novagen), and the nucleotide sequences of each reporter plasmid were confirmed. Clones were digested with both HindIII and EcoRI (sites incorporated in the primers; Table 2), and the fragments were then cloned between the HindIII and EcoRI sites of pMElacZ (Endoh et al., 2003b). The strains of P. putida transformed using reporter plasmids were grown to exponential phase on SFMM containing 1 mM sulfate or 1 mM organosulfur compound at 30 °C. A lacZ assay was conducted on the cultures according to Miller (1972), with some modifications.

Measurement of growth characteristics.
P. putida strains were grown in 1 ml SFMM with 1 mM DMSO₂ as a sulfur source at 30 °C. Cells were harvested by centrifugation and washed three times by suspension in 1 ml SFMM. The washed cells were resuspended into 1 ml SFMM containing 1 mM of the appropriate sulfur source. The sulfur sources used were benzylmethyl sulfide, benzylmethyl sulfone, methanethiol, ethanesulfonate, propanesulfonate, butanesulfonate, pentanesulfonate, taurine and hexylsulfate. After appropriate intervals, OD₅₅₀ was measured using a spectrophotometer (model DU-7400, Beckman).

Chemicals.
All chemicals used were of the highest purity commercially available (i.e. 98–100 %; Merck, Sigma-Aldrich, Kanto Chemical, Wako Pure Chemical, Nacalai Tesque).

The sfnAB operon is regulated by SfnR under sulfate starvation
To investigate the co-transcription of the sfnA and sfnB genes, RT-PCR was performed with total RNA from sulfate- or MSA-grown DS1 cells and a set of primers (sfnA-RT-f and sfnB-RT-r) based on the 3'-terminus of sfnA and the 5'-terminus of sfnB (Table 2). The expected 760 bp amplified product was only observed in the reaction mixture containing total RNA from MSA-grown cells (data not shown), indicating that the sfnAB gene cluster was organized as a sulfate-starvation-induced operon. Subsequently, RT-PCR analysis was performed to investigate whether the expression of the sfnA gene was controlled by a novel σ⁵⁴-dependent regulator (SfnR) under sulfate starvation. Total RNA prepared from both MSA-grown DS1 and Dfi74 J (sfnR : : Tn5; DMSO₂-use deficiency) was subjected to RT-PCR with the primer sets ssuE-FW/RV for ssuE and sfnA-FW/RV for sfnA (Table 2). The amplified product of the ssuE gene, which is expressed under sulfate starvation, but is not regulated by SfnR (Endoh et al., 2003a, b), was observed in both cells. In contrast, the amplified product of the sfnA gene was not found in the reaction mixture containing total RNA from the MSA-grown Dfi74 J cells (Fig. 2). These results indicate that SfnR is involved in the upregulation of the expression of the sfnAB operon under sulfate limitation.

Table 2). Total RNA was prepared from cultures of strains DS1 and Dfi74 J (sfnR disruptant) on sulfate or methanesulfonate (MSA); 1 µg total RNA was subjected to RT-PCR. In controls total DNA was used as template.

Determination of the transcription start point of the sfnA gene
The results of the RT-PCR analysis pointed to the existence of an σ⁵⁴-dependent promoter activated by SfnR in the region upstream of the sfnA gene. To determine the transcriptional start point of sfnA, total RNA was extracted from DS1(pMEsfnA-PE) cells grown on sulfate or MSA, and a primer extension analysis was then performed. These analyses revealed one transcriptional start point (a single G base) 32 bp upstream from the translation start point of sfnA, but only in MSA-grown cells (Fig. 3, lane 2). Through comparison with the σ⁵⁴-dependent consensus promoter (5'-TGGCACN₅TTGCW-3') of E. coli (Reitzer & Schneider, 2001), the promoter of the sfnAB operon was determined to be 5'-GGGCACGGAGTTTGCG-3' (underlined sequences are highly conserved; Fig. 3).

Table 2) annealed to a specific site within the sfnA gene. Lanes G, A, T and C correspond to the sequence ladder generated using the same primer, and the sequence pattern is shown on the right. In the sequence of the sfnF–sfnA intergenic region, the position of the transcriptional start point is shown. The promoter of the sfnAB operon containing a consensus σ54-binding sequence is shown in bold. The boxed sequences (sites 1, 2 and 3) indicate the SfnR-binding regions (Endoh et al., 2005).

Identification of the DNA region required for expression of the sfnAB operon
To identify the DNA region required for the sfnR-dependent expression of the sfnAB operon, deletion analyses of the region upstream of the sfnA gene were performed using lacZ as a reporter. Strain DS1 was transformed by each sfnA'–lacZ reporter plasmid (Fig. 4), and the β-galactosidase activities of the resultant transformants were measured in the presence of sulfate or DMSO₂. In the reaction mixtures with DMSO₂, the deletion of the region containing only site 1 (located upstream of position –138 from the sfnA transcription start point) had no effect on the induction of β-galactosidase activity. However, deletion of the region containing site 2 (located upstream of position –108 from the sfnA transcription start point) exhibited very low β-galactosidase activity. In contrast, the β-galactosidase activities of the transformants were very low in reaction mixtures with sulfate (Fig. 4), and no activity was observed in DS1 harbouring pMElacZ (control vector). These results indicate that the DNA region downstream of position –138 with respect to the sfnA transcription start point was necessary for SfnR-dependent expression of the sfnAB operon.

(9K): Fig. 4. Deletion analysis of the sfnA upstream region. The sfnA regions fused with lacZ are represented by the lines on the left. β-Galactosidase activities of the DS1 transformants grown on sulfate (white bars) or DMSO2 (black bars) are shown on the right; results are presented as the mean±SD of three independent experiments.

Mutational analysis of SfnR-binding sites
We investigated whether the 18 bp imperfect inverted repeat sequence (IIRS; 5'-CTGTN₁₀ACAG-3') in site 2 (previously expected to be a candidate for the consensus sequence bound by SfnR; Endoh et al., 2005) plays an important role in the expression of the sfnAB operon. Three sfnA'–lacZ reporter plasmids with mutations in one or two of the IIRSs were constructed (Fig. 5a), and β-galactosidase activities were compared with those of the sfnA'–lacZ reporter plasmids without mutations. All mutations within the IIRS of site 2 led to the loss of the induction of β-galactosidase activity under sulfate starvation (Fig. 5b), suggesting that this IIRS in site 2 is a potential upstream activating sequence (UAS) for SfnR binding.

(21K): Fig. 5. Mutational analysis of SfnR-binding sites. (a) The mutated sequences (bold italics) in the reporter plasmids. (b) β-Galactosidase activities of the DS1 transformants grown on sulfate (white bars) or DMSO2 (black bars). Results are presented as the mean±SD of three independent experiments.

Growth characteristics of SAK1 and SBK1
According to the results of the transcriptional analyses, the expression of the sfnA and sfnB genes was apparently sulfate-starvation-induced and regulated by SfnR. However, strains SAK1 (sfnA disruptant) and SBK1 (sfnB disruptant) are able to use DMS, DMSO, DMSO₂, MSA, diethyl sulfone, hexanesulfonate, methionine, methionine sulfoxide and methionine sulfone as sulfur sources (Endoh et al., 2005). Therefore, to determine the function of these genes, the growth characteristics of SAK1 and SBK1 were investigated further using other organosulfur compounds as sulfur sources. Among the organosulfur compounds tested, differences in growth were only reliably observed when methanethiol was used as a sulfur source. Whereas SBK1 had nearly identical growth capabilities to wild-type DS1, SAK1 exhibited a reduced growth rate compared to DS1 (Fig. 6). Similar results were obtained when SAK1 (pEN10) carrying the sfnB expression plasmid was used to avoid a polar effect (data not shown). This observation suggests that SfnA functions in the assimilation of methanethiol. Because UK1 (ssuD, encoding MSA-sulfonatase, disruptant with kan) lost the ability to grow on methanethiol (Fig. 6), we hypothesize that SfnA may catalyse an enzymic reaction (oxygenation) involved in the conversion of methanethiol to MSA.

(14K): Fig. 6. Growth of strain DS1 and its derivatives with methanethiol as a sulfur source. Symbols: , DS1 (wild-type); , SAK1 (sfnA disruptant); •, SBK1 (sfnB disruptant); , UK1 (ssuD disruptant). Results are presented as the mean of three independent experiments; error bars (SD) are smaller than the symbols.

The σ⁵⁴-dependent transcriptional regulator SfnR binds the sfnA–sfnF intergenic region and upregulates the expression of the sfnFG operon, which is essential for the DMS and DMSO₂ metabolism of strain DS1 (Endoh et al., 2005). Of the three SfnR-binding sites within the intergenic region, the site 1 region is indispensable for the expression of these catabolic genes (Endoh et al., 2005). To find other target genes controlled by this novel transcriptional regulator, we investigated whether SfnR-binding sites other than site 1 are involved in the expression of the sfnAB operon. In addition to the DNase I footprinting previously conducted (Endoh et al., 2005), the RT-PCR analysis, primer extension and lacZ reporter assays demonstrated that the expression of the sfnAB operon was activated by the binding of SfnR to sites 2 and 3.

We demonstrated that two divergently oriented transcriptional units (sfnFG and sfnAB) were regulated by SfnR, both of which were strictly regulated in response to sulfate limitation. However, the features of their UASs bound by SfnR were quite different. Normally, two adjacent UASs for a σ⁵⁴-dependent activator are placed in the correct configuration, and their sequences are formed by two IIRSs (approx. 16 bp) with a spacing of 29–42 bp between the centres of the palindromes (Pérez-Martín & de Lorenzo, 1996; Tropel & van der Meer, 2002). For transcriptional activation of the sfnAB operon, two adjacent UASs (i.e. IIRS in site 2 and IIRS in site 3) were placed in the usual position with a space of 30 bp between the centres of the palindromes. In contrast, the two UASs (IIRSs in site 1) for the expression of sfnFG operon were partially overlapping (Endoh et al., 2005). Therefore, the relationship between the promoter geometry bound by large protein complexes of SfnR and the transcriptional activation levels of the respective operons warrants further investigation.

Both sfnA and sfnB are not required for DMS and DMSO₂ metabolism (Endoh et al., 2005). Thus, we investigated the function of the SfnR-dependent sulfate-starvation-induced gene, sfnA, in organosulfur assimilation. SAK1 (sfnA : : kan) repeatedly exhibited reduced growth compared to the wild-type strain on methanethiol. This suggests that the sfnA gene may function in methanethiol metabolism; however, the detailed oxidative pathway for degradation of methanethiol as a sulfur source has not yet been identified. Vermeij & Kertesz (1999) reported that P. putida converted methanethiol to MSA through an unknown pathway and subsequently desulfonated MSA to yield sulfite. Because UK1 (ssuD : : kan) failed to grow on methanethiol (Fig. 6), DS1 also metabolized methanethiol to sulfite via MSA. Considering that SAK1 can grow on MSA as a sulfur source, the sfnA gene may be involved in the metabolic step from methanethiol to MSA. To elucidate the methanethiol metabolic pathway at the molecular level, strain DS1 must be transposon-mutagenized to obtain methanethiol-deficient mutants. Such an investigation may help explain why SAK1 growth was restored to nearly the same level as in DS1 or SBK1 (Fig. 6) because possible methanethiol metabolite candidates (e.g. methanesulfinate) are non-enzymically oxidized to MSA in the liquid phase by hydroxyl radicals (Arsene et al., 2002; Bardouki et al., 2002). SfnA may be involved in the oxidation of such metabolic intermediates of methanethiol.

Currently, genome sequences are available for several Pseudomonas species (Pseudomonas Genome Project; ). According to a homology search, their genomes contain many FMNH₂-dependent monooxygenase homologues (acyl-CoA dehydrogenase-family homologues) such as the sfnA and sfnB genes. However, many of their functions remain unknown. In strain DS1, we also found at least five sulfate-starvation-induced genes encoding FMNH₂-dependent monooxygenase homologues [i.e. ssuD (Endoh et al., 2003a), sfnC (Endoh et al., 2003b), sfnG, sfnA and sfnB (Endoh et al., 2005)], but the enzymic functions of SfnC, SfnA and SfnB are not known. Further study of SfnA may provide insights into the functions of such unknown sulfate-starvation-induced genes in organosulfur assimilation.

A part of this study was supported by a Grant-in-Aid for Scientific Research (no. 17780057) to H. H. from the Ministry of Education, Culture, Sports, Science, and Technology of Japan.

Edited by: M. A. Kertesz