Bacterial Cysteine-Inducible Cysteine Resistance Systems

ABSTRACT

Cysteine donates sulfur to macromolecules and occurs naturally in many proteins. Because low concentrations of cysteine are cytotoxic, its intracellular concentration is stringently controlled. In bacteria, cysteine biosynthesis is regulated by feedback inhibition of the activities of serine acetyltransferase (SAT) and 3-phosphoglycerate dehydrogenase (3-PGDH) and is also regulated at the transcriptional level by inducing the cysteine regulon using the master regulator CysB. Here, we describe two novel cysteine-inducible systems that regulate the cysteine resistance of Pantoea ananatis, a member of the family Enterobacteriaceae that shows great potential for producing substances useful for biotechnological, medical, and industrial purposes. One locus, designated ccdA (formerly PAJ_0331), encodes a novel cysteine-inducible cysteine desulfhydrase (CD) that degrades cysteine, and its expression is controlled by the transcriptional regulator encoded by ccdR (formerly PAJ_0332 or ybaO), located just upstream of ccdA. The other locus, designated cefA (formerly PAJ_3026), encodes a novel cysteine-inducible cysteine efflux pump that is controlled by the transcriptional regulator cefR (formerly PAJ_3027), located just upstream of cefA. To our knowledge, this is the first example where the expression of CD and an efflux pump is regulated in response to cysteine and is directly involved in imparting resistance to excess levels of cysteine. We propose that ccdA and cefA function as safety valves that maintain homeostasis when the intra- or extracellular cysteine concentration fluctuates. Our findings contribute important insights into optimizing the production of cysteine and related biomaterials by P. ananatis.

IMPORTANCE Because of its toxicity, the bacterial intracellular cysteine level is stringently regulated at biosynthesis. This work describes the identification and characterization of two novel cysteine-inducible systems that regulate, through degradation and efflux, the cysteine resistance of Pantoea ananatis, a member of the family Enterobacteriaceae that shows great potential for producing substances useful for industrial purposes. We propose that this novel mechanism for sensing and regulating cysteine levels is a safety valve enabling adaptation to sudden changes in intra- or extracellular cysteine levels in bacteria. Our findings provide important insights into optimizing the production of cysteine and related biomaterials by P. ananatis and also a deep understanding of sulfur/cysteine metabolism and regulation in this plant pathogen and related bacteria.

INTRODUCTION

The cysteine residues of proteins form disulfide bonds that influence protein folding and conformation, as well as form Fe-S clusters (1). In addition, cysteine is a source of sulfur for other cellular macromolecules. However, at low concentrations, free cysteine is cytotoxic to prokaryotes and eukaryotes (2–4). Therefore, the level of intracellular cysteine is stringently controlled. Several modes of regulation at different levels have been identified, reflecting its significance as a thiol-containing molecule for cellular functions, as well as its deleterious effects, requiring careful handling. Escherichia coli regulates cysteine levels through multiple mechanisms that primarily involve biosynthesis through feedback inhibition of serine acetyltransferase (SAT) and 3-phosphoglycerate dehydrogenase (3-PGDH) by cysteine and serine, respectively (5, 6). The master regulator CysB organizes the CysB regulon by controlling the transcription of most of the genes that encode proteins involved in sulfur assimilation, including the components of sulfur transport systems and the cysteine biosynthetic pathway (7–9). Degradation of cysteine represents another regulatory mechanism. In E. coli, at least five enzymes (CysK, CysM, MetC, TnaA, and MalY) exhibit cysteine desulfhydrase (CD) activity, which decomposes cysteine into pyruvate, ammonium, and sulfide (10), and were proposed to remove excess cysteine. Because these enzymes mediate other functions (e.g., CysK and CysM are cysteine synthases [11, 12]), their physiological significance in cysteine degradation is unclear.

The efflux of cysteine may serve as a safety valve to reduce excess levels of intracellular cysteine. For example, studies on the accumulation of extracellular cysteine by strains used for fermentation show that several transporters (EamA, EamB, CydDC, and Bcr) that exhibit efflux activity mediate cysteine transport (13–16). However, the specificity of these transporters for cysteine and their significance in controlling cysteine levels as a safety valve are unclear. For example, CydDC and Bcr are primarily transporters of glutathione and multiple drugs (e.g., bicyclomycin), respectively (16, 17), and EamA and EamB may function primarily as protection against reactive oxygen species (ROS) (18).

Here, we describe two novel systems of Pantoea ananatis, a member of the family Enterobacteriaceae (19), which are directly involved in resistance to excess levels of cysteine through mechanisms that mediate cysteine degradation and efflux. The systems include a CD and a cysteine efflux pump. Their levels are regulated in response to cysteine, and their functions are associated with cysteine resistance. To our knowledge, we provide here the first example of a CD and a cysteine efflux pump that are primarily involved in regulating intracellular cysteine levels. P. ananatis, a member of the bacterial family Enterobacteriaceae, has been studied as a plant pathogen (for a review, see reference 20). In recent years, nonpathogenic P. ananatis and its closely related species have attracted the interest of the fermentation industry because they have promise as tools for metabolic engineering (19, 21, 22). Recent studies have indicated the potential of P. ananatis for overproducing a wide variety of useful materials, including amino acids and related compounds (l-glutamate [23], l-aspartate [24], and l-3,4-dihydroxyphenylalanine [25]), vitamins (pyrroloquinoline quinone [22], ascorbic acid intermediates [54], and vitamin E [26]), and other chemical compounds (2,3-butanediol [X. W. Jiang, 9 May 2007, Chinese Patent Office]). Our present study provides important insights into using the bacterium for the fermentative production of cysteine and related biomaterials.

MATERIALS AND METHODS

Bacterial strains, plasmids, and growth conditions.

The strains and plasmids used in this study are listed in Table 1. All deletions and replacements of P. ananatis genes were performed using the Red-driven integration system developed by Katashkina et al. (21). Marker excision from the P. ananatis chromosome was performed using a λ Int/Xis-dependent technique described previously (21). To amplify fragments for chromosomal integration, we used pMW-(λattL-Km^r-λattR) (kanamycin marker) as a template and primers listed in Table S2 in the supplemental material. P. ananatis SC17(0)/pRSFRedTER was electroporated with PCR amplicons. The strain was designed for inducing Red-driven recombination components, and the transformants were selected for antibiotic resistance of the donor strain. Recipient cells were transformed by electroporating them with chromosomal DNA prepared from the donor strain. To induce λ Int/Xis-dependent excision of markers from the chromosome, pMW-intxis-cat was used as a helper plasmid. All strains were grown in Luria-Bertani (LB) medium, M9 minimal medium (27), or fermentation medium [15 g of (NH₄)₂SO₄, 1.5 g of KH₂PO₄, 1 g of MgSO₄ · 7H₂O, 0.1 mg of thiamine hydrochloride, 1.7 mg of FeSO₄ · 7H₂O, 0.15 mg of Na₂MoO₄ · 2H₂O, 0.7 mg of CoCl₂ · 6H₂O, 1.6 mg of MnCl₂ · 4H₂O, 0.3 mg of ZnSO₄ · 7H₂O, 0.25 mg of CuSO₄ · 5H₂O, 0.6 g of tryptone, 0.3 g of yeast extract, 0.6 g of NaCl, 20 g of CaCO₃, 135 mg of l-histidine · HCl · H₂O, 4 g of Na₂S₂O₃, 2 mg of pyridoxine hydrochloride, and 60 g of glucose per liter] supplemented with appropriate antibiotics at 37°C (E. coli) or 34°C (P. ananatis) or as otherwise indicated. The medium was supplemented with 20 μg ml⁻¹ kanamycin, 12.5 μg ml⁻¹ tetracycline, and/or 25 μg ml⁻¹ chloramphenicol as required.

TABLE 1.

Bacterial strains and plasmids

Strain or plasmid	Descriptiona	Source or referencec
Strains
E. coli
MG1655	Wild type (ATCC 47076)	ATCC
W3110	Wild type (ATCC 27325)	ATCC
W3110ΔmetC	W3110 ΔmetC	This study
P. ananatis
SC17	Low-mucus-producing mutant derived from wild-type AJ13355 (19) (NBRC BP-11091)	NITE
SC17ΔccdA	SC17 ΔccdA::Kmr	This study
SC17ΔccdR	SC17 ΔccdR::Kmr	This study
SC17ΔlacZ-PccdA-lacZ	SC17 ΔlacZ::Tetr-PccdA::lacZ-Kmr	This study
SC17ΔcefA	SC17 ΔcefA::Kmr	This study
SC17ΔcefR	SC17 ΔcefR::Kmr	This study
SC17ΔccdAΔcefA	SC17 ΔccdA ΔcefA::Kmr	This study
CYS-1	SC17 harboring pMIV-CysE5 and pACYC177	This study
CYS-2	SC17 harboring pMIV-CysE5 and pACYC-eamA	This study
CYS-3	SC17 harboring pMIV-CysE5 and pACYC-PA36ccR	This study
CYS-4	SC17 harboring pMIV-CysE5 and pACYC-cefA	This study
CYS-5	SC17 harboring pMIV-CysE5 and pACYC-PA36ccd	This study
SC17(0)	Recipient strain for λ Red-mediated integration	21
Plasmids
pSTV29	Cloning vector; Cmr	TaKaRa Bio
pCcdA	pSTV29; ccdA gene on 1.5-kb DNA fragmentb; Cmr	This study
pCcdR	pSTV29; ccdR gene on 1.0-kb DNA fragmentb; Cmr	This study
pCefA	pSTV29; cefA gene on 1.1-kb DNA fragmentb; Cmr	This study
pCefR	pSTV29; cefR gene on 2.0-kb DNA fragmentb; Cmr	This study
pACYC177	Cloning vector; Kmr	Nippon Gene
pACYC-eamA	pACYC177; eamA (E. coli) gene on 1.4-kb DNA fragmentb; Kmr	This study
pACYC-cefAF	pACYC177; cefA gene on 1.1-kb DNA fragmentb; Kmr	This study
pACYC-PA36ccd	pACYC177; cefA-cefR gene on 2.7-kb DNA fragmentb; Kmr	This study
pACYC-PA36ccR	pACYC177; cefR gene on 2.0-kb DNA fragmentb; Kmr	This study
pMIV-CysE5	cysE5 (E. coli) under control of omp promoter	53
pHSG299	Cloning vector; Kmr	TaKaRa Bio
pHSG-ccdA	pHSG299; ccdA gene on 1.5-kb DNA fragmentb; Cmr	This study
pMW-(λ attL-Kmr-λattR)	Donor attLλ-Kmr-attRλ cassette; Apr Kmr	21
pMW-intxis-cat	pSC101-ts; λ xis-int genes transcribed from λ PR promoter under cIts857 control; Cmr	21
pRSFRedTER	λ gam, bet, and exo genes under control of P element; sacB Cmr	21

^aAp^r, Km^r, Tet^r, and Cm^r, resistance to ampicillin, kanamycin, tetracycline, and chloramphenicol, respectively.

^bFragment contains 300 and 200 bp upstream and downstream, respectively, from the CDS(s).

^cATCC, American Type Culture Collection; NITE, National Institute of Technology and Education.

Screening and growth assay for cysteine resistance.

Genomic DNA of P. ananatis wild-type strain SC17 was partially digested with Sau3AI, and fragments of approximately 10 kb were cloned into the BamHI site of pSTV29 (TaKaRa Bio). E. coli MG1655 was electroporated with the genomic-DNA library, and the cells were exposed to 2 to 4 mM cysteine for 2 days. Colonies were isolated, and we determined the nucleotide sequences of the candidate genes carried by the plasmids that conferred cysteine resistance. Cells transformed by plasmids encoding cysteine resistance that harbored ccdA and cefA were grown aerobically in M9 minimal medium overnight and were diluted 1:200 to 1:200,000 and spotted onto M9 plates containing 0, 0.5, or 4 mM cysteine. The plates were incubated for 24 h and monitored for growth.

Assay of CD activity.

An overnight culture of each strain grown in LB medium was diluted 1:100 in fresh LB medium and grown aerobically to logarithmic phase (3 to 5 h) before harvesting by centrifugation. The cultures were supplemented with 30 mM cysteine as required. Native PAGE and analysis of CD activity were performed as described previously (28). Total protein concentrations were analyzed using a standard Coomassie protein assay system; 2 μg of the total protein was loaded in each lane.

Measurement of CD activity.

A 1:100 dilution of an overnight culture of each strain grown in LB medium was inoculated into fresh LB medium and cultured with or without 25 mM cysteine. Cells in logarithmic phase were collected, washed with a buffer containing 0.85% NaCl at 4°C, and suspended in 0.1 M potassium phosphate buffer (pH 8.0) containing 50 μg ml⁻¹ bovine serum albumin and 10 μM pyridoxal phosphate. The cells were disrupted by ultrasonication and centrifuged, and the supernatant (cell extract) was analyzed for CD activity as previously described (29).

β-Galactosidase lacZ reporter assay.

An overnight culture grown in M9 minimal medium was inoculated into fresh M9. After 5 h, when the cells attained logarithmic phase (optical density at 600 nm [OD₆₀₀] = 0.3), 0.1 mM cysteine or water was added to the culture medium, samples were collected hourly from 0 to 3 h, and β-galactosidase was measured using the standard Miller method (30).

Real-time quantitative PCR.

A 1:100 dilution of an overnight culture of each strain grown in M9 minimal medium was inoculated into the same fresh M9 medium. Cysteine (1 mM) was added after the cells were grown to logarithmic phase (approximately 6 h). Samples were taken 0, 2.5, 5, and 10 min later and mixed with RNA Protect Bacteria Reagent (Qiagen), frozen in liquid N₂, and stored at −80°C. Total RNA (2.5 μg) was prepared using an RNeasy minikit (Qiagen) and treated with DNase (Turbo DNA-free; Ambion), and then 200 ng was reverse transcribed using an ExScript RT Reagent kit (TaKaRa Bio). The cDNA was quantified using real-time PCR with the primers listed in Table S3 in the supplemental material and Power SYBR green master mix (Applied Biosystems) according to the manufacturer's instructions. Transcript levels were normalized to that of the 16S rRNA internal standard.

Fermentative production of cysteine.

Each strain was streaked onto an LB plate and grown overnight. Cells were collected using a 10-μl loop passed through 7 cm of bacteria on the culture plates and inoculated into 2 ml of the fermentation medium in test tubes (23-mm internal diameter; 200 mm long). The cells were incubated at 32°C with agitation until all the sugar was consumed (approximately 21 to 24 h). The amount of cysteine-related compounds in the culture broth was determined using a previously published method (31).

RESULTS

Screening a P. ananatis genomic library for genes conferring cysteine resistance.

To identify P. ananatis genes that when overexpressed conferred resistance to cysteine, we screened a multicopy genomic library prepared from P. ananatis genomic DNA. E. coli wild-type strain MG1655 was transformed with the library, and cells were cultured in M9 minimal medium supplemented with 2 to 4 mM cysteine (the lowest concentrations sufficient to inhibit growth for 2 days). Colonies were selected after 2 days, and the genes responsible for the phenotype were identified by sequencing the plasmid DNAs. We identified cefA (cysteine efflux; formerly PAJ_3026), which is predicted to form multiple transmembrane helixes with conserved motifs characteristic of those of members of the LysE superfamily that are common to transporters of small molecules (32, 33), suggesting a functional association with cysteine efflux. A plasmid containing cefA carried a coding sequence (CDS) designated cefR (cysteine efflux regulator; formerly PAJ_3027) that is divergently transcribed from cefA, encoding a putative helix-turn-helix (HTH)-type transcriptional regulator that is possibly involved in regulating cefA expression.

The other gene, ccdA (cysteine-inducible cysteine desulfhydrase; formerly PAJ_0331), shares motifs with members of the tryptophan synthase β superfamily, which is one of the four major superfamilies of pyridoxal 5′-phosphate (PLP)-dependent enzymes. In E. coli, five l-CDs and d-CD mediate detoxification of cysteine by converting it into pyruvate, ammonium, and hydrogen sulfide. These enzymes are PLP dependent, and CysM, CysK, and DcyD are members of the tryptophan synthase β superfamily (29, 34, 35). Based on their structural characteristics shared with CDs and their ability to confer cysteine resistance, we propose that ccdA encodes a novel CD involved in the detoxification of cysteine. Preliminary analysis of microarrays suggests that the transcription of cefA and ccdA responds to 1 mM cysteine (cysteine shock). Because evidence indicates that cefA and ccdA may mediate cysteine resistance, we conducted further biological analyses.

To investigate the effects of cefA and ccdA on growth in the presence of cysteine, we deleted these genes and assayed the growth of the mutants on solid medium containing cysteine. The growth of each mutant was significantly sensitive to cysteine, suggesting that the proteins encoded by cefA and ccdA were required for growth in the presence of toxic concentrations of cysteine. We confirmed that the sensitivity to cysteine was complemented by introduction of each gene on multicopy plasmids (Fig. 1A and B). Figure 1C shows a comparison of the cysteine sensitivity of each single and double mutant. The results indicate that ccdA and cefA affected cysteine resistance at different levels. The effect of ccdA was greater than that of cefA, and sensitivity was increased in the double-deletion mutant.

FIG 1.

Sensitivity to cysteine. The sensitivity of each strain to cysteine was determined by observing its growth on M9 plates containing 0, 0.5, or 4 mM cysteine. (A) Wild-type (WT) P. ananatis SC17 harboring the empty vector (pSTV29) and a plasmid carrying ccdA (pCcdA) and the ccdA mutant SC17ΔccdA harboring pSTV29 and pCcdA. (B) WT strain SC17 harboring the empty vector (pSTV29) and a plasmid carrying cefA (pCefA) and the cefA mutant SC17ΔcefA carrying pSTV29 and pCefA. (C) WT strain SC17, ccdA mutant SC17ΔccdA, cefA mutant SC17ΔcefA, and a ccdA cefA double mutant, SC17ΔccdAΔcefA.

A novel CD encoded by ccdA is the only major cysteine-inducible CD expressed by P. ananatis.

We next performed native PAGE and enzymatic assays to detect CD activity. A crude cell extract prepared from exponentially grown cells in LB medium at 37°C was prepared and subjected to native PAGE, and the gel was analyzed for CD activity. As a control, we assayed the activity of the CD encoded by metC in extracts prepared from wild-type E. coli (W3110) and an E. coli metC deletion mutant (W3110 ΔmetC) (Fig. 2A). CD activity was detected in the lanes loaded with the extract of wild-type P. ananatis (SC17) but not in those loaded with the extract of the ccdA deletion mutant. CD activity was detected in the lanes loaded with extracts prepared from the ccdA deletion mutant after it was transformed with the multicopy plasmid (Fig. 2A). These data suggest that ccdA encodes a CD. Moreover, CD activity was induced when cysteine was added during culture (Fig. 2A). The levels of CD activity were determined using crude cell extracts prepared from the wild-type strain and the ccdA deletion mutant, and the results show that CD activity was induced 10-fold by cysteine (Table 2). The results further suggest that ccdA is the only gene that encodes an inducible CD activity, because the deletion mutant failed to respond to cysteine. Moreover, ccdA must encode a major CD, because CD activity was drastically decreased in extracts prepared from the ccdA deletion mutant.

FIG 2.

CD activity of CcdA and its induction by cysteine. (A) Assays of CD activity of the metC product of E. coli (lanes 1 and 2) and the ccdA product of P. ananatis (lanes 3 to 8). The strains and conditions were as follows: E. coli wild type (W3110) (lane 1), metC mutant (W3110ΔmetC) (lane 2), P. ananatis wild type (SC17) (lane 3), ccdA mutant (SC17ΔccdA) (lane 4), wild type (SC17) cultivated with 30 mM cysteine (lane 5), ccdA mutant (SC17ΔccdA) cultivated with 30 mM cysteine (lane 6), wild type (SC17) harboring pHSG299 (vector control) (lane 7), and wild type (SC17) harboring pHSG-ccdA (expressing plasmid-borne ccdA) (lane 8). (B) lacZ reporter assay of the ccdA promoter in the presence or absence of cysteine, using the P_ccdA-lacZ fusion. The strain SC17ΔlacZ-P_ccdA-lacZ was grown in M9 medium to logarithmic phase, and the time course of β-galactosidase activity was measured hourly from 0 to 3 h after addition of 100 μM cysteine (solid circles) and a cysteine-free solution (open circles). The values represent the averages of the results of two independent experiments. (C) Time course of relative ccdA mRNA levels after cysteine addition. SC17 (wild type) grown in M9 medium to logarithmic phase was treated with 1 mM cysteine (cysteine shock [solid circles]) or without cysteine (open circles). Shown are the levels of ccdA mRNA compared with those before addition of cysteine. The values represent the averages of the results of three independent experiments. (D) Relative levels of ccdA mRNA after the addition of cysteine in the presence or absence of ccdR. (Top) ccdR deletion mutant construct. (Bottom) Level of ccdA mRNA 5 min after cysteine shock (1 mM) using the ccdR mutant (SC17ΔccdR) harboring the empty vector (pSTV29) and a plasmid carrying ccdR (pCcdR). The values were normalized to the levels of mRNAs before stimulation (0 min = 1). The values represent the averages for three independent experiments, and the error bars represent 1 standard deviation. nt, nucleotides.

TABLE 2.

Total cellular CD activity in the absence or presence of cysteine

Strain	Cysteinea	CD activity (mU)b
SC17	−	11.2 ± 0.3
	+	121.9 ± 3.7
SC17ΔccdA	−	1.1 ± 0.1
	+	1.1 ± 0.1

^a−, absent; +, present.

^bThe values are averages ± 1 standard deviation for four independent experiments; 1 U = 1 μmol/min/mg of protein.

Cysteine-mediated induction of ccdA is rapid and elicits a strong response regulated at the transcriptional level.

To test whether cysteine-inducible CD activity was due to activation of the ccdA promoter, we transformed cells with a construct comprising the promoter region of ccdA fused to a lacZ reporter sequence and measured β-galactosidase activity after adding cysteine to the culture. The results show that β-galactosidase activity was induced in the presence of cysteine (Fig. 2B). Real-time PCR analysis (Fig. 2C) showed the extremely rapid (≤2.5 min) and robust induction (approximately 400-fold compared with the zero-time control) and indicated that ccdA transcription responded to cysteine. Although we did not determine the minimum concentration required for induction, maximum induction was achieved using ≤100 μM cysteine (data not shown). The E. coli genome contains genes that encode at least five CDs, and among them, the transcription of tnaA is upregulated when cells are grown in the presence of a high concentration of cysteine (28). In contrast, we found that transcription of E. coli tnaA was not induced by cysteine shock, raising the possibility that the regulation was not direct (data not shown). We propose that ccdA is the first CD gene to be discovered that rapidly and robustly responds directly to cysteine and that the primary physiological role is closely associated with the detoxification of cysteine.

The ccdR gene, located immediately upstream and transcribed divergently from ccdA, encodes a transcriptional regulator responsible for cysteine-mediated induction of ccdA.

We tested whether ccdR (cysteine-inducible cysteine desulfhydrase regulator; formerly PAJ_0332 or ybaO), which is located immediately upstream and transcribed divergently from ccdA, was responsible for the cysteine-mediated induction of ccdA. ccdR encodes a putative LRP-like transcriptional regulator that is considered an ortholog of E. coli YbaO (it contains 74% identical amino acids) (36). We constructed a deletion mutant of ccdR to determine the effects on the induction of ccdA in cells grown in the presence of cysteine. To maintain the integrity of the putative regulatory region of ccdA, ccdR was partially deleted, leaving 300 bp immediately upstream of ccdA (Fig. 2D). Using real-time PCR analysis, we compared the levels of ccdA mRNA after addition of cysteine with those of cells lacking ccdR, as well as those of cells transformed by a plasmid encoding ccdR. The results show that deletion of ccdR abolished the induction of ccdA, which was complemented by a plasmid-encoded ccdR (Fig. 2D). These data support the idea that ccdR is a transcriptional regulator that positively controls the cysteine-mediated induction of ccdA.

CefA functions as an efflux pump for cysteine.

To determine whether cefA gene product-mediated efflux of cysteine affected the enhancement of cefA expression on the overproduction of cysteine, we employed cysteine-producing strains. In E. coli, cysteine synthesis is regulated by SAT that is encoded by cysE and is subject to feedback inhibition by cysteine (5, 6). Therefore, cells must express a feedback-insensitive SAT to overproduce cysteine (37–39). Moreover, enhancing the activity of efflux pumps exerts additional effects on the overproduction of cysteine (14, 16, 40). For these reasons, we employed the P. ananatis wild-type strain (SC17) harboring the plasmid pMIV-cysE5 (40) carrying a cysE5 gene (6) that encodes a feedback-insensitive mutant of SAT to determine whether cefA encodes a cysteine efflux pump. As a control, we overexpressed the cysteine efflux pump EamA (8) in the cysteine-producing strain SC17 harboring the plasmid pMIV-cysE5. We were unable to detect the production of cysteine by the wild-type strain SC17, and cells transformed by cysE5 produced cysteine, which was accelerated by the expression of EamA (Table 3). The increased production of cysteine when cefA is overexpressed supports the conclusion that the product of cefA mediates the efflux of cysteine (Table 3). Further, an additional effect on the production of cysteine was observed when cells were cotransformed with cefR and cefA, suggesting that the transcriptional regulator encoded by cefR modulated the expression of cefA.

TABLE 3.

Production of cysteine

Expt and strain	Enhanced gene(s) on plasmids	Production of cysteinea (μg ml−1)	OD600	Cultivation timeb (h)
Expt 1
CYS-1	cysE5, vector	169 ± 4	27.8 ± 0.5	23
CYS-2	cysE5, eamA	545 ± 28	21.8 ± 0.1	<19
Expt 2
CYS-1	cysE5, vector	226 ± 40	32.6 ± 0.8	20
CYS-3	cysE5, cefR	188 ± 50	32.9 ± 0.9	17.5
CYS-4	cysE5, cefA	496 ± 75	30.1 ± 1.3	19.5
CYS-5	cysE5, cefA, cefR	900 ± 96	29.9 ± 0.2	<16

^aProduction of cysteine and related products in the medium was measured as described by Gaitonde (31). The values are averages ± 1 standard deviation for four independent experiments.

^bCultivation was terminated after the initial glucose (60 g liter⁻¹) was consumed.

Cysteine induces the transcription of cefA.

Because our preliminary microarray data indicated that cysteine shock induced the transcription of cefA, using real-time PCR, we determined the kinetics of cefA expression. The design of the experiment was the same as that for ccdA described above. The results confirm that cysteine rapidly induced cefA mRNA (Fig. 3A).

FIG 3.

Induction of cefA transcription by cysteine. (A) Time course of relative cefA mRNA levels after the addition of cysteine. P. ananatis SC17 (wild type) was grown in M9 medium to logarithmic phase and treated with 1 mM cysteine (solid diamonds). The mRNA levels of cells not treated with cysteine are indicated by the open diamonds. The relative levels of cefA mRNA compared with the level of mRNA before cysteine was added are shown. The values represent the averages of the results of three independent experiments. (B) Relative levels of cefA mRNA after the addition of cysteine in the presence or absence of cefR. (Top) Structure of the cefR deletion. (Bottom) Level of cefA mRNA 5 min after cysteine shock (1 mM) using the cefR mutant (SC17ΔcefR) harboring an empty vector (pSTV29) and cefR carried by a plasmid (pCefR). The values were normalized to each amount of mRNA before the addition of cysteine (0 min = 1) and represent the averages of the results of three independent experiments, and the error bars represent 1 standard deviation.

The cefR gene, located immediately upstream and transcribed divergently from cefA, encodes a transcriptional regulator responsible for cysteine-mediated induction of cefA.

We tested whether cefR was responsible for the cysteine-mediated induction of cefA. The level of cefA mRNA after cysteine shock revealed that deletion of cefR abolished the induction of cefA, which was complemented by the plasmid carrying cefR (Fig. 3B). These data support the idea that cefR is a transcriptional regulator that positively controls the cysteine-mediated induction of cefA, which represents a modular regulatory system analogous to ccdA-ccdR. Further experiments at the molecular level that confirm cysteine-mediated promoter binding or activation of the transcription factors (CcdR and CefR) are necessary to fully demonstrate these points.

DISCUSSION

In the present study, we identified a cysteine-inducible CD and a cysteine efflux pump in P. ananatis that may directly mediate resistance to excess levels of cysteine. We previously reported some of these findings in our patent application in 2009 (55), which discloses the functional characterization of ccdA (described as d0191) as a cysteine-inducible CD involved in cysteine resistance and ccdR (described as c0263) as a transcriptional regulator that controls cysteine-mediated induction. This report, together with our patent application, describes the first example, to our knowledge, of a modular cysteine resistance catalyst/regulator system in bacteria. Our findings are supported by those of Oguri et al., who characterized the cdsH-cutR system, which is the Salmonella ortholog of ccdA-ccdR (41). These systems, along with well-characterized feedback inhibition loops involving SAT and 3-PGDH, as well as the transcriptional regulator CysB and its regulon, provide a finely tuned mechanism to counter the deleterious effects of cysteine.

To explore the conservation and variation of these systems among organisms, we queried the STRINGS database (42) and identified six groups associated with the genome context near ccdA-ccdR (Fig. 4). The sequences of ccdA, ccdR, or both are conserved in most gammaproteobacteria and less well conserved in alphaproteobacteria, suggesting that the system is most common in Enterobacteriaceae. Further, we identified the well-conserved gene cluster mdlAB (encoding a predicted multidrug ABC transporter) immediately downstream of ccdR, which may represent an operon. Group A comprises the full set of genes, whereas all other groups lack at least one of the components. We found it interesting that some groups contain the regulator ccdR that is missing its counterpart ccdA (groups C and D) and vice versa (group E).

FIG 4.

Comparison of the genomic contexts of ccdR-ccdA among bacterial genomes. Representative examples of bacterial genomes are shown below.

An informative example is provided by E. coli, which harbors ybaO (ccdR ortholog) and mdlAB but not ccdA (group D). The transcription factor encoded by ybaO may have been conserved fortuitously independently of ccdA during evolution, or it may possess alternative functions independent of CcdA, such as that of a regulator of mdlAB or other target genes on a different genomic locus. In contrast, cefA-cefR is conserved in only a few genera (e.g., Serratia and Erwinia of the Enterobacteriaceae). This limited conservation may reflect the diminished contribution of CefA to the regulation of cysteine levels, because its magnitude of induction (Fig. 2D and 3B) and resistance to the levels of cysteine (Fig. 1) are significantly lower than those of CcdA. Alternatively, the cefA-cefR system may function under specific environmental conditions and is therefore functionally complementary to the ccdA-ccdR system.

Because most CDs in E. coli are involved in other physiological functions, the significance of their roles in cysteine degradation is unclear. For example, the O-acetylserine sulfhydrylases CysK and CysM catalyze the reaction of O-acetyl-l-serine with sulfide to produce cysteine (11, 12). The cystathionine β-lyase family members MetC and MalY cleave the Cβ-S bond of diverse substrates, for example, to convert cystathionine to homocysteine (43, 44). The tryptophanase TnaA catalyzes the hydrolysis of l-tryptophan to indole, pyruvate, and ammonia (45). E. coli expresses several less specialized CDs, whereas Corynebacterium glutamicum utilizes a single major CD that may provide a more specific function in cysteine metabolism. However, the only known C. glutamicum CD is encoded by aecD, which is not characterized regarding its regulation associated with cysteine levels (46).

We show here that P. ananatis expressed a single major CD that was synthesized at levels and with kinetics apparently different from those of other CDs that impart cysteine resistance. For example, the threshold concentration of cysteine that affects cell growth is 10 times higher in P. ananatis than in E. coli (see Fig. S1 in the supplemental material). Because deletion of ccdA abolished this effect, ccdA may make a major contribution to the distinctive ability of P. ananatis to tolerate levels of cysteine that are toxic to other bacterial species. Thus, the present study is the first to identify a CD with a distinct physiological role in imparting resistance to cysteine through a mechanism involving the degradation and efflux of cysteine.

E. coli and its relatives utilize several pumps that efflux cysteine; however, we are unaware of direct evidence that implicates them in the efflux of cysteine that serves as a safety valve to mitigate toxicity. EamA and EamB, the representative cysteine exporters of E. coli (13, 14), may serve primarily as components of the cysteine/cystine shuttle system that protects cells from ROS. Moreover, these exporters are highly induced by hydrogen peroxide compared with cysteine, suggesting that their physiological function is to protect cells against ROS (18). CydDC transports cysteine and glutathione (15, 17) and mediates cytochrome assembly (47). Bcr and other multidrug transporters export cysteine; however, they were originally identified as drug exporters (48). These findings support our claim that cefA, together with its regulator cefR, represents the first example of a bacterial system that responds directly to excess cysteine and mediates cysteine resistance. Efflux of toxic substances out of cells provides one of the simplest countermeasures; however, it might result in futile influx/efflux cycles (49). Cysteine is generally present as its reduced form in a reductive environment, such as the cytoplasm, and is present as the oxidized form (i.e., cystine) in an oxidative environment, such as the periplasm (18). In E. coli and related bacteria (including P. ananatis), import of cystine is mediated by two transporter systems, YdjN and FliY-YecSC (50). Both systems are controlled by the transcriptional regulator CysB (51), a master regulator of sulfur metabolism that activates expression of most of the genes involved in assimilative production of cysteine when cysteine is in short supply (7, 8). Thus, to our knowledge, the conditions under which the safety valve is activated with transient excess levels of cysteine are not those under which cystine transport is activated through CysB, thereby creating a futile cycle. Alternatively, outer membrane porins may play an important role in preventing the futile cycle between cytoplasm and periplasm. Indeed, the outer membrane channel TolC may be involved in cysteine tolerance through export of cysteine species from the periplasm (38).

In general, bacterial amino acid efflux pumps exhibit rather low substrate specificity [14; K. Takumi and G. Nonaka, U.S. patent application EP2218729(B1)]. Our preliminary experiments suggest that CefA exhibits broad amino acid recognition, which may provide opportunities for its application to a wide variety of fermentative amino acid production processes (40).

Figure 5 summarizes our present findings, together with previously characterized bacterial systems that regulate cysteine metabolism. The known regulatory systems control the biosynthesis of cysteine through feedback inhibition of SAT and 3-PGDH (5, 6), as well as the transcriptional regulation of the cysteine regulator CysB, which controls the output of its biosynthetic pathway (7–9). Our present findings expand the regulatory mechanisms employed by bacteria to rapidly and effectively protect them against cysteine cytotoxicity. We propose that CcdA and CefA function as safety valves that respond to a sudden change in the concentration of intra- or extracellular cysteine. Because of the importance of cysteine to bacterial physiology and its potentially deleterious effects, the survival of bacteria throughout evolution likely depended on the acquisition of sophisticated and exquisitely sensitive mechanisms to protect them from cysteine toxicity.

FIG 5.

Mechanisms employed by P. ananatis to regulate cysteine levels. The dashed lines indicate regulation at the enzymatic and transcriptional levels. NAS, N-acetylserine; OAS, O-acetylserine; PAPS, 3-phosphoadenosine 5-phosphosulfate; 3-PG, d-3-phosphoglycerate; TF, transcription factor.

Because of its advantageous phenotype, P. ananatis attracts the interest of those involved in the fermentation industry (23, 52; O. N. Mokhova, T. M. Kuvaeva, L. I. Golubeva, A. V. Kolokolova, and J. Y. Katashkina, 8 April 2010, International Patent Cooperation Treaty). The data shown here provide important insights into the use of P. ananatis for the fermentative production of cysteine and related biomaterials and show that inactivation of major CDs and enhancement of efflux pumps are key factors that make the bacterium an ideal host cell for the fermentative production of cysteine.

Funding Statement

The funder had no role in study design, data collection and interpretation, or the decision to submit the work for publication.