Correspondence: Eiji Masai, Department of Bioengineering, Nagaoka University of Technology, Nagaoka, Niigata 940-2188, Japan. Tel.: +81 258 47 9428; fax: +81 258 47 9450; e-mail: firstname.lastname@example.org
Sphingobium sp. strain SYK-6 is able to degrade various lignin-derived aromatic compounds including ferulate, vanillate, and syringate. In the SYK-6 cells, ferulate is converted to vanillin and acetyl-coenzyme A (acetyl-CoA) through the reactions catalyzed by feruloyl-CoA synthetase and feruloyl-CoA hydratase/lyase encoded by ferA and ferB, respectively. Here, we characterized the transcriptional regulation of ferBA controlled by a MarR-type transcriptional regulator, FerC. The ferC gene is located upstream of ferB. Reverse transcription (RT)-PCR analysis suggested that the ferBA genes form an operon. Quantitative RT-PCR analyses of SYK-6 and its mutant cells revealed that the transcription of the ferBA operon is negatively regulated by FerC, and feruloyl-CoA was identified as an inducer. The transcription start site of ferB was mapped at 30 nucleotides upstream from the ferB initiation codon. Purified His-tagged FerC bound to the ferC–ferB intergenic region. This region contains an inverted repeat sequence, which overlaps with a part of the −10 sequence and the transcriptional start site of ferB. The binding of FerC to the operator sequence was inhibited by the addition of feruloyl-CoA, indicating that FerC interacts with feruloyl-CoA as an effector molecule. Furthermore, hydroxycinnamoyl-CoAs, including p-coumaroyl-CoA, caffeoyl-CoA, and sinapoyl-CoA also acted as effector.
Lignin is the most abundant aromatic compound in nature, and its mineralization is a fundamental step in the terrestrial carbon cycle. In nature, it is considered that white rot fungi, which secrete extracellular phenol oxidases, initiate the degradation of native lignin (Higuchi, 1971; ten Have & Teunissen, 2001), and the resulting lignin-derived aromatic compounds are mineralized by bacteria (Vicuña, 1988). Sphingobium sp. strain SYK-6, one of the best characterized degraders of lignin-derived aromatics, is capable of utilizing a wide variety of lignin-derived biaryls, including β-aryl ether (Sato et al., 2009), biphenyl (Peng et al., 2005), phenylcoumaran, and diarylpropane, as well as various lignin-derived monoaryls, including ferulate (Masai et al., 2002), vanillin, and syringaldehyde (Masai et al., 2007b) as the sole source of carbon and energy. These lignin-derived compounds are converted to vanillate or syringate, which are then further degraded via aromatic-ring cleavage pathways (Masai et al., 2007a).
In the SYK-6 cells, ferulate is transformed to feruloyl-coenzyme A (feruloyl-CoA) by feruloyl-CoA synthetase encoded by ferA in the presence of CoA, ATP, and Mg2+ (Masai et al., 2002). The resultant feruloyl-CoA is hydrated to 4-hydroxy-3-methoxyphenyl-β-hydroxypropionyl-CoA and then further degraded to produce vanillin and acetyl-CoA by feruloyl-CoA hydratase/lyase encoded by ferB (Fig. 1a). Vanillin is oxidized by the reaction of vanillin dehydrogenase encoded by ligV, which is located at a different locus from ferBA (Masai et al., 2007b). The resultant vanillate is further metabolized by the protocatechuate (PCA) 4,5-cleavage pathway after the conversion of vanillate to PCA by O demethylation catalyzed by vanillate/3-O-methylgallate O-demethylase, LigM (Abe et al., 2005; Masai et al., 2007a). Among the SYK-6 genes involved in the ferulate catabolism, we characterized transcriptional regulation of the PCA 4,5-cleavage pathway genes controlled by a LysR-type transcriptional regulator, LigR (Kamimura et al., 2010). However, the regulation of other genes including ferBA remains unknown.
While there are a number of reports on the catabolic genes of ferulate (Rahouti et al., 1989; Gasson et al., 1998; Venturi et al., 1998; Overhage et al., 1999; Achterholt et al., 2000; Civolani et al., 2000; Masai et al., 2002), the information regarding the transcriptional regulation of the ferulate catabolic genes is very scarce. However, Calisti et al. (2008) reported on regulation of the ferulate catabolic genes of Pseudomonas fluorescens BF13. In this strain, the ech-vdh-fcs genes, which respectively encode feruloyl-CoA hydratase/lyase, vanillin dehydrogenase, and feruloyl-CoA synthetase, formed an operon, and the transcription of this operon was negatively regulated by a MarR-type transcriptional regulator, FerR. Based on the ech promoter assay using BF13 mutants, feruloyl-CoA was identified as an inducer molecule of the ech-vdh-fcs operon. Similar observation had been described in the regulation of the hydroxycinnamate catabolic genes (hca) of Acinetobacter baylyi ADP1 (Parke & Ornston, 2003). The hca genes were shown to be negatively regulated by a MarR-type transcriptional regulator, HcaR. Furthermore, p-coumaroyl-CoA was identified as a true inducer. However, the biochemical analysis of the interaction of the regulator protein with the operator sequence has not been documented, and there is no direct proof that hydroxycinnamoyl-CoAs are the effector molecules of these MarR-type transcriptional regulator proteins.
In this study, we characterized the transcriptional regulation of the ferBA genes of SYK-6 controlled by a MarR-type transcriptional regulator, FerC. In vitro assay demonstrated the interaction between FerC and the operator sequence, and actual effector molecules of FerC were identified.
Materials and methods
Bacterial strains, plasmids, and culture conditions
The bacterial strains and plasmids used in this study are listed in Supporting Information, Table S1. Sphingobium sp. strain SYK-6 and its mutant derivatives were routinely grown at 30 °C in Luria-Bertani (LB) medium or Wx minimal salt medium (Table S2) containing 5 mM ferulate, 5 mM vanillin, 10 mM vanillate, or SEMP (10 mM sucrose, 10 mM glutamate, 0.34 mM methionine, and 10 mM proline). The ferC mutant, SME043 was obtained by introduction of the ferC disruption plasmid, pFESIBI into the cells of SYK-6, and disruption of the gene was examined by Southern hybridization analysis as described previously (Sato et al., 2009). Escherichia coli strains were grown in LB medium at 37 °C. For cultures of cells carrying antibiotic resistance markers, the media for E. coli transformants were supplemented with 100 mg of ampicillin per liter or 25 mg of kanamycin per liter.
Reverse transcription (RT)-PCR and quantitative RT-PCR (qRT-PCR)
For RT-PCR analysis of the fer genes, SYK-6 cells grown in LB medium were washed with Wx medium and resuspended in the same medium containing 5 mM ferulate and SEMP to an absorbance at 600 nm (A600nm) of 0.2. The cultures were then incubated at 30 °C for 12 h. For qRT-PCR analysis of SYK-6 and ferC mutant, the cells prepared, as described earlier, were used as induced cells, while the cells incubated in LB medium for 12 h, were employed as uninduced cells. For identification of an inducer, cells of SYK-6, FAK, and FBK were incubated in Wx-SEMP medium at 30 °C for 4 h. After the addition of 5 mM ferulate, 5 mM vanillin, or 10 mM vanillate, the cells were further incubated for 6 h. Total RNAs were isolated as described previously (Kamimura et al., 2010) and then treated with RNase-free DNase I (Takara Bio Inc.) to remove contaminating DNA. RT-PCR and qRT-PCR analyses were carried out according to the previous reports (Kamimura et al., 2010; Kasai et al., 2010).
A Beckman dye D4 (D4)-labeled primer, PEferB, complementary to the ferB mRNA from 91 to 111 nucleotides downstream from the ferB start codon, was used to detect the start site of the ferB mRNA (Table S3). Primer extension reactions were performed as described previously (Kasai et al., 2010).
LacZ reporter assays
The reporter plasmids carrying the ferB promoter-lacZ fusion with or without ferC were introduced into SME043 cells. The resulting transformants were grown in Wx medium containing 5 mM ferulate or grown in LB medium at 30 °C for 12 h. Preparation of cell extracts and β-galactosidase assays were performed as described previously (Kasai et al., 2010).
Expression of His tag-fused ferC in E. coli and purification of His-tagged FerC
The coding region of ferC was amplified by PCR using Ex Taq DNA polymerase (Takara Bio Inc.) together with NdeferC-F and R primer pair (Table S3). The 0.6-kb NdeI-XhoI fragment of the PCR product was inserted into pET-16b to generate pETRR1. E. coli BL21(DE3) cells harboring pETRR1 were grown in 100 mL of LB medium at 30 °C. When A600nm of the culture reached 0.5, expression of ferC with an N-terminal His tag was induced for 6 h by adding 1 mM isopropyl-β-d-thiogalactopyranoside. After the incubation, cells were resuspended in 50 mM Tris–HCl buffer (pH 7.5) and broken by an ultrasonic disintegrator (UD-201; Tomy Seiko Co.). The supernatant obtained by centrifugation (19 000 g, 20 min) was applied to a His Spin Trap (GE Healthcare) previously equilibrated with buffer A, consisting of 50 mM Tris–HCl (pH 7.5), 500 mM NaCl, and 100 mM imidazole. After the centrifugation at 100 g for 30 s, samples were washed twice with 500 μL of buffer A. His-tagged FerC (ht-FerC) was eluted with 400 μL of buffer B, consisting of 50 mM Tris–HCl (pH 7.5), 500 mM NaCl, and 500 mM imidazole, and resultant fractions were subjected to desalting and concentration by centrifugal filtration with an Amicon Ultra-0.5 10k filter unit (Millipore). The purity of the enzyme preparation was examined by sodium dodecyl sulfate-12% polyacrylamide gel electrophoresis (SDS-PAGE). In vitro cross-linking of ht-FerC was performed as descried in a previous study (Kamimura et al., 2012).
Electrophoretic mobility shift assays (EMSAs)
EMSAs for ht-FerC were performed with a DIG gel shift kit 2nd generation (Roche). To prepare FER-102, FER-66, FER-50, and FER-48 probes, DNA fragments containing the ferC-ferB intergenic region were synthesized by PCR with specific primer pairs (Table S3), and the 3′ ends of the fragments were labeled with digoxigenin (DIG)-11-ddUTP. DNA binding assays were performed at 20 °C in a total volume of 10 μL mixture containing 1–32 ng of purified ht-FerC (0.025–0.80 pmol dimer), a DIG-labeled probe (0.5 nM of FER-102 or FER-66 probe; or 1.0 nM of FER-50 or FER-48 probe), 1.0 μg of poly[d(I-C)], 0.1 μg of poly-l-lysine, and a reaction buffer [20 mM HEPES, 10 mM (NH4)2SO4, 1 mM dithiothreitol, 0.2% (w/v) Tween 20, 30 mM KCl, and 1 mM EDTA, pH 7.6] for 20 min, following the same procedure described earlier (Kamimura et al., 2010). To test the association of FerC with effector molecules, ht-FerC (5 ng, 0.13 pmol) was previously incubated with 100 μM of feruloyl-CoA or other hydroxycinnamoyl-CoAs at 20 °C for 10 min. A FER-102 probe (1.0 nM) was then added to the mixture and incubated for 10 min. Gel electrophoresis and the detection of signals were performed according to a previous description (Kamimura et al., 2010).
Preparation of hydroxycinnamoyl-CoAs
The ferA coding sequence was amplified using Prime STAR GXL DNA polymerase (Takara Bio Inc.) and the primer pair of ferA-Nde-F and ferA-Bam-R (Table S3). This fragment was inserted into pET-16b to yield pE16FA. FerA with an N-terminal His tag (ht-FerA) was produced in E. coli BL21(DE3) and purified by His Spin Trap column, and the purity of ht-FerA was examined by SDS-PAGE. To prepare feruloyl-CoA, p-coumaroyl-CoA, caffeoyl-CoA, and sinapoyl-CoA, 2 mM of corresponding hydroxycinnamates were incubated with 20 μg of purified ht-FerA at 25 °C for 6 h in the presence of 2.5 mM CoA, 3 mM MgSO4, and 3 mM ATP. Degradation of each hydroxycinnamate was examined by high-performance liquid chromatography (ACQUITY ultraperformance liquid chromatography system; Waters). The change in absorbance of each reaction mixture was monitored by a V-630 spectrophotometer (Jasco Corp.) at the wavelengths of 345, 333, 346, and 346 nm derived from feruloyl-CoA, p-coumaroyl-CoA, caffeoyl-CoA, and sinapoyl-CoA, respectively (Beuerle & Pichersky, 2002). The reaction mixtures were filtered by an Amicon ultra spin filter unit (3-kDa cutoff, Millipore), and then the filtrates were used as preparations of 2 mM hydroxycinnamoyl-CoAs.
Results and discussion
Operon structure of the fer genes
Nucleotide sequence of the SYK-6 genome (Masai et al., 2012) revealed that SLG_25040 (ferC), which is located 87 bp upstream of ferB (Fig. 1b), showed 20–27% identity at amino acid level with ferR of P. fluorescens BF13 (Calisti et al., 2008), badR of Rhodopseudomonas palustris (Egland & Harwood, 1999), and mobR of Comamonas testosteroni KH122-3a (Hiromoto et al., 2006; Yoshida et al., 2007). These gene products involved in the catabolism of ferulate, benzoate, and 3-hydroxybenzoate, respectively, belong to the family of MarR-type transcriptional regulator; therefore, ferC appears to encode a MarR-type transcriptional regulator.
To determine the operon structure of ferBA, RT-PCR analyses were carried out with the total RNA extracted from the cells of SYK-6 grown with ferulate and primers complementary to the neighboring genes (Fig. 1c). PCR-amplified products of the expected sizes were detected for the internal region of ferB and the intergenic region between ferB and ferA; however, no products for ferC–ferB and ferA-SLG_25010 intergenic regions were obtained. These results suggested that ferB and ferA are organized in the same transcriptional unit.
Transcriptional regulation of the ferBA operon controlled by FerC
qRT-PCR analyses were performed to determine the transcriptional regulation of the ferBA operon. As shown in Fig. 2a, the transcription of ferB was induced 6.5-fold in the SYK-6 cells grown on ferulate. However, no induction was observed in the cells grown in the presence of the metabolites of ferulate, vanillin or vanillate, suggesting that the inducer molecule of the ferBA operon is ferulate or its first metabolite, feruloyl-CoA (Fig. 2b).
To examine the role of ferC in the transcriptional regulation of the ferBA operon, ferC mutant (SME043) was created. qRT-PCR analyses showed that ferB was constitutively expressed at a high level in the SME043 cells, indicating that the ferBA operon is negatively regulated by the ferC gene product (Fig. 2a).
The ferA mutant (FAK), which is unable to transform ferulate, and the ferB mutant (FBK), which is scarcely able to transform feruloyl-CoA were employed for the qRT-PCR analysis to determine the inducer of the ferBA operon. SYK-6 has two feruloyl-CoA hydratase/lyase genes, ferB and ferB2 (Masai et al., 2002), but the level of ferB2 transcription was < 10% of that of ferB (data not shown). In the FAK cells, the transcriptional induction of ferB was not observed in the presence of ferulate (Fig. 2b). On the other hand, the transcription of ferB was significantly induced in the FBK cells when ferulate was supplemented (Fig. 2b). These results indicated that feruloyl-CoA is the actual inducer of the ferBA operon. This fact corresponded to the observation that CoA-thioester intermediates act as inducers for the regulation by FerR, HcaR, and BadR (Egland & Harwood, 1999; Parke & Ornston, 2003; Calisti et al., 2008).
Promoter analysis of the ferBA operon
To determine the promoter region of the ferBA operon, a DNA fragment containing ferC and the ferC-ferB intergenic region was cloned into a promoter-probe vector pPR9TZ (Kamimura et al., 2010), generating a transcriptional fusion to the promoterless lacZ reporter gene (pPR85). The levels of expression of the lacZ fusion in SME043 cells harboring pPR85 were examined. The β-galactosidase activity was increased 16-fold in the cells grown in the presence of ferulate (Fig. S1). Therefore, the cis-acting region necessary for the transcriptional regulation in response to an inducer seemed to be in the ferC–ferB intergenic region. On the other hand, SME043 cells harboring pPR05, which contains the ferC–ferB intergenic region but not ferC, showed constitutive expression (Fig. S1), suggesting that FerC regulates the ferBA transcription through the binding to the ferC–ferB intergenic region.
Determination of the transcription start site of the ferBA operon
The transcription start site of the ferBA operon was determined by primer extension analysis using a specific primer that anneals to the ferB sequence. Due to the fact that no significant extension product was observed when total RNA from SYK-6 cells was used as a template, we prepared total RNA from SYK-6 cells harboring pKTBIEI85, which contains the ferC–ferB intergenic region and 5′ terminal of ferB. Based on the size of the major product obtained using RNA isolated from the cells grown in the presence of ferulate (Fig. 3a), the transcription start site of the ferBA operon was mapped at T residue located at 30 nucleotides upstream from the initiation codon of ferB (Fig. 3b). Upstream of the transcription start site, putative −35 and −10 sequences (ATGGCT-N17-AATGCT) that are similar to the conserved sequence of σ70-dependent promoter were found. In addition, we found two inverted repeat sequences, IR1 (CGATGGCTTGCTCCCATCG) and IR2 (ATGCTATGGCTTATAGCAT), which overlap with −35 element and the region containing a part of −10 element and the transcriptional start site, respectively. One of these inverted repeat sequences, or both, appeared to be involved in the binding of FerC.
Binding of ht-FerC to the ferC–ferB intergenic region
The His tag-fused ferC gene was expressed in E. coli cells harboring pETRR1, and the production of a 16-kDa protein was observed by SDS-PAGE (Fig. S2). ht-FerC purified to near homogeneity by Ni affinity chromatography was subjected to an in vitro cross-linking experiment to estimate the molecular mass of native ht-FerC (Fig. S2). SDS-PAGE of the cross-linked sample showed a major shifted band at ca. 34 kDa. This result suggested that ht-FerC forms a homodimer.
EMSAs were performed using purified ht-FerC and four different DNA fragments carrying the ferC–ferB intergenic region (Fig. 4a). The binding experiments showed that the mobility of FER-102 and FER-66 probes, which contain the region from positions −56 to +46 and −20 to +46 relative to the transcriptional start site of the ferBA operon, respectively, were retarded upon the addition of ht-FerC (Fig. 4b). The intensities of the shifted bands of FER-102 and FER-66 probes were enhanced through an increase in the amount of protein, suggesting that ht-FerC directly bound to the region from positions −20 to +46, which contains IR2. By contrast, no retardation was observed when FER-48 and FER-50 probes were employed (Fig. 4b). Because FER-48 and FER-50 probes do not include the upstream half site and downstream half site of IR2, respectively, it is strongly suggested that the palindromic motif of IR2 is essential for the binding of ht-FerC. In light of the position of IR2, FerC appeared to inhibit the binding of RNA polymerase to the promoter to repress the transcription. MarR-type transcriptional regulators are reported to bind to operator sites containing a perfect or imperfect inverted repeat sequence as a dimer (Tropel & van der Meer, 2004). FerC required both half sites of IR2 for its binding, and each subunit of the dimeric structure of FerC seemed to bind to one of the two half sites of IR2, as a similar binding mechanism proposed in MarR of E. coli (Alekshun et al., 2001).
Identification of effector molecules of FerC
To examine whether the effector molecule of FerC is truly feruloyl-CoA, feruloyl-CoA was prepared from ferulate using a purified FerA enzyme (Fig. S3). When prepared feruloyl-CoA was added to the EMSA reaction mixture, the ht-FerC binding to the FER-102 fragment was inhibited (Fig. 4c). To test the ability of other hydroxycinnamoyl-CoAs to inhibit the binding of FerC to the fer operator sequence, caffeoyl-CoA, p-coumaroyl-CoA, and sinapoyl-CoA were also enzymatically prepared from caffeate, p-coumarate and sinapate, respectively. Interestingly, all the above hydroxycinnamoyl-CoAs inhibited the binding of FerC to the FER-102 probe. These results clearly indicated that FerC is able to interact with hydroxycinnamoyl-CoAs, and these interactions seemed to enable SYK-6 to metabolize not only ferulate but also other hydroxycinnamates by the relief of the repression of the ferBA operon.
At the time this manuscript was being written, the transcriptional regulation of the p-coumarate catabolic genes, couAB, of R. palustris was reported (Hirakawa et al., 2012). This reported research found that the transcription of couAB is negatively regulated by a MarR-type transcriptional repressor, CouR, and it was demonstrated that the binding of CouR to the operator sequence was antagonized by p-coumaroyl-CoA. Although the amino acid sequence identity between FerC and CouR is only ca. 23%, both regulators appeared to regulate the target genes in a similar manner. Our results in this study provide a definite proof that feruloyl-CoA is the actual effector of FerC in the catabolism of ferulate, and hydroxycinnamoyl-CoAs also act as effector molecules of FerC.
D.K. and N.K. contributed equally to this work.
Conflict of interest
This study has no conflict of interest between authors.