Anaerobic Respiration of Elemental Sulfur and Thiosulfate by Shewanella oneidensis MR-1 Requires psrA, a Homolog of the phsA Gene of Salmonella enterica Serovar Typhimurium LT2

Abstract

Shewanella oneidensis MR-1, a facultatively anaerobic gammaproteobacterium, respires a variety of anaerobic terminal electron acceptors, including the inorganic sulfur compounds sulfite (SO₃²⁻), thiosulfate (S₂O₃²⁻), tetrathionate (S₄O₆²⁻), and elemental sulfur (S⁰). The molecular mechanism of anaerobic respiration of inorganic sulfur compounds by S. oneidensis, however, is poorly understood. In the present study, we identified a three-gene cluster in the S. oneidensis genome whose translated products displayed 59 to 73% amino acid similarity to the products of phsABC, a gene cluster required for S⁰ and S₂O₃²⁻ respiration by Salmonella enterica serovar Typhimurium LT2. Homologs of phsA (annotated as psrA) were identified in the genomes of Shewanella strains that reduce S⁰ and S₂O₃²⁻ yet were missing from the genomes of Shewanella strains unable to reduce these electron acceptors. A new suicide vector was constructed and used to generate a markerless, in-frame deletion of psrA, the gene encoding the putative thiosulfate reductase. The psrA deletion mutant (PSRA1) retained expression of downstream genes psrB and psrC but was unable to respire S⁰ or S₂O₃²⁻ as the terminal electron acceptor. Based on these results, we postulate that PsrA functions as the main subunit of the S. oneidensis S₂O₃²⁻ terminal reductase whose end products (sulfide [HS⁻] or SO₃²⁻) participate in an intraspecies sulfur cycle that drives S⁰ respiration.

Microbial reduction of inorganic sulfur compounds is central to the biogeochemical cycling of sulfur and other elements such as carbon and metals (29). The ability to reduce elemental sulfur (S⁰) is found in members of both prokaryotic domains (20), including mesophilic deltaproteobacteria (Desulfovibrio vulgaris, Pelobacter carbinolicus, Geobacter sulfurreducens) (6, 9, 36, 51), thermophilic deltaproteobacteria (Desulfurella acetivorans) (39), gammaproteobacteria (Shewanella putrefaciens) (41), epsilonproteobacteria (Wolinella succinogenes) (49), cyanobacteria (“Oscillatoria limnetica”) (45), and hyperthermophilic archaea (1, 53). Partially reduced inorganic sulfur compounds such as tetrathionate (S₄O₆²⁻), thiosulfate (S₂O₃²⁻), and sulfite (SO₃²⁻) are also important electron acceptors in the biogeochemical cycling of sulfur (29, 51). S₄O₆²⁻-reducing bacteria, for example, may produce S₂O₃²⁻ as a metabolic end product of S₄O₆²⁻ reduction, while S₂O₃²⁻ disproportionation is a key reaction catalyzed by sulfate-reducing bacteria, resulting in the formation of sulfate (SO₄²⁻) and sulfide (S²⁻) (26).

Shewanella oneidensis MR-1, a facultatively anaerobic gammaproteobacterium, respires a variety of compounds as an anaerobic electron acceptor, including the inorganic sulfur compounds S⁰, SO₃²⁻, S₂O₃²⁻, and S₄O₆²⁻; transition metals [e.g., Fe(III) and Mn(IV)]; and radionuclides [e.g., U(VI) and Tc(VII)] (8, 21, 41, 44, 50, 55, 56). The majority of studies of anaerobic respiration by S. oneidensis have focused on the mechanism of electron transport to transition metals and radionuclides (11, 14, 34, 46, 58, 59), while the mechanism of electron transport to inorganic sulfur compounds has not been thoroughly examined.

Microbial S⁰ respiration is postulated to occur via two pathways, both of which are based on an intraspecies sulfur cycle. In the first pathway (catalyzed by members of the genus Salmonella [20]), S₂O₃²⁻ is reduced, yielding HS⁻ and SO₃²⁻ (24). SO₃²⁻ diffuses from the cell and reacts chemically with extracellular S⁰ to form S₂O₃²⁻, which reenters the periplasm and is rereduced, thereby sustaining an intraspecies sulfur cycle. In the second pathway (catalyzed by W. succinogenes [24]), water-soluble polysulfides (S_n²⁻; n > 2), formed by chemical interactions of S⁰ at pHs >7 (52), are reduced stepwise in the periplasm to S_{n − 1}²⁻ and HS⁻. Similarly to what occurs with the first pathway, microbially produced HS⁻ diffuses from the cell and reacts chemically with S⁰ to produce additional S_n²⁻, which reenters the periplasm and is rereduced to sustain an analogous intraspecies sulfur cycle (24).

Genetic analyses of S₂O₃²⁻ reduction-deficient mutants of Salmonella enterica serovar Typhimurium have demonstrated that phsA (denoting production of hydrogen sulfide) is required for HS⁻ production during S₂O₃²⁻ respiration (10, 17, 22). In addition, phsA-deficient mutants are unable to reduce S⁰ as an electron acceptor (24). The phsA homolog of W. succinogenes (annotated as psrA, for polysulfide reduction) is required for S⁰ respiration (32, 37). W. succinogenes psrA is the first gene of a three-gene cluster (including psrA, psrB, and psrC) whose products encode a polysulfide reductase, a quinol oxidase, and a membrane anchor, respectively (15). In addition, the structure of the polysulfide reductase complex (PsrABC) from Thermus thermophilus has recently been solved, and results indicate that PsrC acts as a quinol oxidase that transfers electrons stepwise via PsrB and PsrA to S_n²⁻ during anaerobic S⁰ respiration (27). The main objectives of the present study were to (i) identify the S. Typhimurium phsA homolog in the S. oneidensis genome, (ii) employ a newly constructed suicide cloning vector for in-frame gene deletion mutagenesis in S. oneidensis to delete the S. Typhimurium phsA homolog of S. oneidensis, and (iii) test the S. oneidensis psrA deletion mutant for respiratory activity on a combination of two electron donors and 11 electron acceptors, including the inorganic sulfur compounds S₄O₆²⁻, S₂O₃²⁻, and S⁰.

MATERIALS AND METHODS

Growth media and cultivation conditions.

S. oneidensis MR-1 was cultured at 30°C in Luria-Bertani (LB) medium (10 g liter⁻¹ NaCl, 5 g liter⁻¹ yeast extract, 10 g liter⁻¹ tryptone) for genetic manipulations. For anaerobic growth experiments, cells were cultured in a defined salts medium (44) supplemented with lactate (18 mM) or formate (30 mM) as the carbon/energy source. Anaerobic growth experiments were carried out with 13-ml Hungate tubes (Bellco Glass, Inc.) filled with 10 ml of salts medium and sealed with black butyl rubber stoppers under an N₂ atmosphere. Growth experiments were performed in two parallel yet independent incubations. In a second set of anaerobic growth experiments with S⁰ and S₂O₃²⁻, S. oneidensis cultures were incubated with continual flushing of the headspace with N₂ to remove bacterially produced HS⁻ (37). Electron acceptors were added from filter-sterilized stocks (synthesized as described previously [13]), except where indicated, at the following final concentrations: O₂ (atmospheric); NO₃⁻, 10 mM; Fe(III) citrate, 50 mM; hydrous ferric oxide and Mn(III)-pyrophosphate, 40 mM (30) and 10 mM, respectively; trimethylamine-N-oxide (TMAO), 25 mM; S₂O₃²⁻, 10 mM; S₄O₆²⁻, 2 mM; fumarate, 30 mM; dimethyl sulfoxide (DMSO), 25 mM; and S⁰, 20 mM (41). When required, antibiotics were supplemented at the following final concentrations: gentamicin (Gm), 15 μg ml⁻¹, and chloramphenicol, 25 μg ml⁻¹. For growth of Escherichia coli β2155 λ pir (12), diaminopimelate was supplemented at a final concentration of 100 μg ml⁻¹.

Analytical techniques.

Cell growth was monitored by direct cell counts via epifluorescence microscopy and by measuring terminal electron acceptor depletion or end product accumulation. Acridine orange-stained cells were counted (Carl Zeiss AxioImager Z1 microscope) according to previously described procedures (35). Cell numbers at each time point were calculated as the average of 10 counts from two parallel yet independent anaerobic incubations. NO₂⁻ was measured spectrophotometrically with sulfanilic acid-N-1-naphthyl-ethylene-diamine dihydrochloride solution (40). Fe(III) reduction was monitored by measuring Fe(II) production with the ferrozine technique (54). Mn(III)-pyrophosphate concentration was measured colorimetrically as previously described (30). S₂O₃²⁻ and S₄O₆²⁻ concentrations were measured by cyanolysis as previously described (28). Growth on O₂, TMAO, DMSO, and fumarate was detected by monitoring cell growth only. Control experiments consisted of incubations with cells that were heat killed at 80°C for 30 min prior to inoculation or by omission of the terminal electron acceptor.

Nucleotide and amino acid sequence analyses.

Genome sequence data for S. oneidensis MR-1 (21), S. Typhimurium LT2 (38), and W. succinogenes (5) were obtained from the comprehensive microbial resource (J. Craig Venter Institute; http://cmr.jcvi.org). S. oneidensis proteins that displayed significant similarity to S. Typhimurium PhsA (designated PsrA in the S. oneidensis genome), tetrathionate reductase (TtrA), and anaerobic sulfite reductase (AsrA) were identified via BLAST analysis (3). Functional motifs were analyzed via Pfam (http://pfam.sanger.ac.uk) and ClustalW (33). The S. oneidensis PsrA sequence was used as the query sequence for BLAST analysis of several other recently sequenced Shewanella genomes, including S. putrefaciens strain 200, S. putrefaciens CN32, S. putrefaciens W3-18-1, S. amazonensis SB2B, S. denitrificans OS217, S. baltica OS155, S. frigidimarina NCIMB400, S. pealeana ATCC 700345, S. woodyi ATCC 51908, Shewanella sp. strain ANA-3, Shewanella sp. strain MR-4, Shewanella sp. strain MR-7, S. loihica PV-4, S. halifaxensis HAW-EB4, S. piezotolerans WP3, S. sediminis HAW-EB3, and S. benthica KT99. Preliminary genome sequence data for these strains were obtained from the Department of Energy, Joint Genome Institute (http://www.jgi.doe.gov).

Construction of suicide vector pKO2.0.

The strategy for construction of the suicide vector pKO2.0 (Table 1) is outlined in Fig. S1 in the supplemental material. The R6K oriV and the RP4 origin of transfer (oriT) genes of pKNOCK-Gm (2) were PCR amplified with NheI restriction sites engineered onto the ends of the oriV-oriT fragment. The chloramphenicol acetyltransferase (cat) gene and the LacZ-containing multiple-cloning site (MCS) were PCR amplified from pBBR1MCS (31) with NheI sites also engineered onto the ends (see Table S1 in the supplemental material for corresponding primers). The resulting PCR products were ligated to form pKOCat-MCS. The open reading frames encoding SacB and Gm acetyltransferase (conferring Gm resistance [Gm^r]) were PCR amplified from pJQ200 (48) as a contiguous fragment with ApaLI and MunI restriction sites engineered onto the ends. The cat gene was removed from pKOCat-MCS by inverse PCR, resulting in a fragment containing R6K oriV, RP4 oriT, and the LacZ-MCS with ApaLI and MunI sites on the ends. These two fragments were ligated to form pKO2.0 (R6K oriT LacZ-MCS Gm^r sacB).

TABLE 1.

Strains and plasmids used in this study

Strain or plasmid	Feature(s)	Source or reference
Strains
Shewanella oneidensis
MR-1	Wild-type strain	ATCC
PSRA1	In-frame deletion mutant	This study
PSRA+	“Knock-in” complementation mutant	This study
Escherichia coli
EC100D pir-116	F−mcrA Δ(mrr-hsdRMS-mcrBC)	Epicentre
	φ80dlacZΔM15 ΔlacX74 recA1 endA1
	araD139 Δ(ara leu)7697galUgalK λ−
β2155 λ pir	thrB1004 pro thi strA hsdS lacZΔM15	12
	(F9 lacZΔM15 lacIqtraD36 proA1 proB1)
	ΔdapA::erm pir::RP4 Kmr
Plasmids
pKO2.0	4.5-kb γR6K, mob RP4 sacB GmrlacZ	This study
pKNOCK-Gm	1.6-kb γR6K, mob RP4 Gmr	2
pBBR1MCS	CmrlacZ	31
pJQ200	ColE1 sacB Gmr	48
pKOPSRA	pKO2.0 containing in-frame deletion of psrA	This study
pKOPSRA+	pKO2.0 containing wild-type psrA	This study

In-frame gene deletion mutagenesis.

psrA was deleted from the S. oneidensis genome by the following method. Regions corresponding to ∼750 bp upstream and downstream of the psrA open reading frame were PCR amplified using iProof ultrahigh-fidelity polymerase (Bio-Rad, Hercules, CA). Primers used for construction of the psrA deletion mutant are listed in Table 2. PCR cycling routines consisted of 98°C for 30 s, 35 cycles of 98°C for 15 s, 60°C for 30 s, and 72°C for 30 s, concluding with a final extension step at 72°C for 7 min. PCRs (50-μl total mixture volume) were performed on a Bio-Rad iCycler (Bio-Rad, Hercules, CA), and each mixture contained 20 ng of S. oneidensis genomic DNA (DNAzol; Invitrogen), 250 μM of the deoxynucleoside triphosphates, and 50 ng of each oligonucleotide. The resulting fragments were separated by agarose gel electrophoresis, isolated from the gel (QIAquick gel extraction kit; Qiagen), and subsequently joined using overlap extension PCR (25) to generate a DNA fragment containing regions homologous to the regions flanking psrA. This region was cloned into the newly constructed knockout vector pKO2.0, and the resulting plasmid (pKOPSRA) was electroporated into E. coli EC100D pir-116 (Epicentre) (Bio-Rad Xcell electroporation system; Bio-Rad, Hercules, CA). Recipients were detected on LB agar medium containing Gm and 40 μg ml⁻¹ X-Gal (5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside). The recombinant plasmid was confirmed by detection of an ∼1.5-kb insertion following digestion with BamHI. Following confirmation, pKOPSRA was electroporated into E. coli β2155 λ pir (12) (Bio-Rad Xcell electroporation system; Bio-Rad, Hercules, CA), and the resulting donor strain was mated biparentally with S. oneidensis on LB agar medium supplemented with diaminopimelate. Following an 18-h incubation, the mating mix was resuspended in LB liquid medium and subsequently plated onto LB agar medium containing Gm.

TABLE 2.

Primers used in this study

Primer	Sequence
PSRAD1	GACTGGATCCCACAGCTTATTTGGTCG
PSRAD2	ATATCTTTTCGTCATTTGAGCCTCCAAGGTTACCTCCATCACACTCA
PSRAD3	TGAGTGTGATGGAGGTAACCTTGGAGGCTCAAATGACGAAAAGATAT
PSRAD4	GACTGTCGACACAAGCCAAGCCTAAGCTGATGG
PSRADTF	GCGTGCATTTTGAACGACAG
PSRADTR	GCTTTCTAAAGGGCATAAGCAGC
RT-PSRAF	GTCGCCGATAAAGCCGATGAATGGTA
RT-PSRAR	GGCATATCTTTTACGCCGGGCTTAC
RT-PSRBF	TATGGCGAAATGCCCAATCTGC
RT-PSRBR	AAGCGGGTGTCCTTACAGAAGT
RT-PSRCF	AGTCACGACAAGACGTTAGCCA
RT-PSRCR	GCTTTGGCCCGCATACACAATA
RT-RPOAF	GTTCAACACGAGCTGCTTCTA
RT-RPOAR	GGCTTAGCAGTGACTATCGAG

S. oneidensis MR-1 colonies with pKOPSRA integrated into the genome were screened by PCR using primers flanking the recombination region. The resulting strain (MR-1::pKOPSRA) was grown in LB medium with NaCl omitted and subsequently transferred to LB agar medium with NaCl omitted and containing 10% (wt/vol) sucrose. Colonies were patched to LB agar medium containing Gm to confirm the loss of pKO2.0, and the deletion in psrA was confirmed in the Gm-sensitive colonies via PCR using primers flanking the recombination region. One mutant strain (PSRA1) was also confirmed by direct DNA sequencing of the targeted region (University of Nevada, Reno Genomics Center).

Confirmation of in-frame deletions.

Liquid cultures of S. oneidensis wild-type and deletion mutant strains were grown to mid-log phase (corresponding to 1.8 × 10⁹ cells ml⁻¹) in LB medium, and total RNA was extracted with PureZol reagent (Bio-Rad, Hercules, CA) according to the manufacturer's instructions. RNA (∼150 ng μl⁻¹) was further purified (RNeasy Kit; Qiagen) with on-column DNase I treatment according to the manufacturer's instructions. RNA was reversely transcribed with a high-capacity cDNA reverse transcription (RT) kit (Applied Biosystems). RT reaction mixtures (20 μl) consisted of 150 ng RNA, 50 U MultiScribe reverse transcriptase, 4 mM deoxynucleoside triphosphates, and random hexamers. Cycling routines were 25°C for 10 min, 37°C for 120 min, and 85°C for 5 min. The resulting cDNA was PCR amplified with primers specific for internal regions of psrA, psrB, and psrC (Table 2). RNA polymerase factor rpoA served as a positive control. Negative controls consisted of identical reactions with reverse transcriptase omitted.

Knock-in complementation analysis of PSRA1.

PSRA1 was complemented via a strategy analogous to that followed in the gene deletion protocol. Wild-type psrA was PCR amplified from MR-1 genomic DNA with psrA-specific primers PSRAD1 and PSRAD4. The resulting amplicon contained the entire open reading frame and ∼750 bp of upstream and downstream DNA for subsequent recombination into PSRA1. The amplicon was cloned into pKO2.0 using identical restriction sites, and the resulting construct was subsequently transformed into E. coli strains as described above. “Knock-in” complementation was performed as described for in-frame deletion above, except that PSRA1 was used as the recipient strain. Insertion into the proper location of the genome was confirmed via PCR amplification with flanking primers PSRADTF and PSRADTR (Table 2) followed by DNA sequencing (University of Nevada, Reno Genomics Center). The resulting “knock-in” complementation strain was designated PSRA+.

RESULTS

Construction of new suicide vector pKO2.0.

Suicide vector pKO2.0 was constructed in a stepwise fashion, integrating the hallmark components of previously constructed vectors. The R6K oriV and RP4 oriT origins from pKNOCK-Gm and the chloramphenicol resistance marker (cat) and the LacZ-containing polylinker (MCS) from the broad-host-range vector pBBR1MCS were independently PCR amplified and joined to form pKOCat-MCS (See Fig. S1 in the supplemental material). The cat gene was subsequently removed via inverse PCR, and the resulting product (containing R6Kori, oriT, and MCS) was ligated with the PCR product containing the Gm^r gene and sacB, encoding levansucrase (see Fig. S1 in the supplemental material).

Identification of S. oneidensis gene products that display similarity to the PhsA homolog of S. Typhimurium.

BLAST analysis revealed that the S. oneidensis MR-1 genome contained a three-gene cluster (annotated as psrABC) whose translated products displayed 59 to 73% and 65 to 80% amino acid similarity to the phsABC (psrABC) clusters of S. Typhimurium LT2 and W. succinogenes, respectively (Table 3; see also Fig. S3 in the supplemental material). The three-gene psr cluster in S. oneidensis MR-1 has the same genomic organization as the three-gene psr clusters in both S. Typhimurium and W. succinogenes (data not shown). ClustalW multiple alignments of PsrA functional domains with those of other proteobacteria indicated that S. oneidensis PsrA contains the conserved 4Fe-4S cluster and molybdopterin guanine-dinucleotide (MGD) binding motifs postulated to contribute to PsrA activity (Fig. 1). PsrA contained the predicted Pfam domains, consistent with other members of the polysulfide (thiosulfate) reductase family of proteins, including the 4Fe-4S cluster (Pfam accession no. PF04879) and MGD binding motifs (accession no. PF01568) (Fig. 1). Proteins homologous to S. Typhimurium TtrA (tetrathionate reductase) or AsrA (anaerobic sulfite reductase) were not detected in the S. oneidensis genome (data not shown).

TABLE 3.

Identities and similarities between PsrA, PsrB, and PsrC homologs

Protein	Shewanella spp.a			S. Typhimurium LT2			GenBankb
	Sim	%ID	E value	Sim	%ID	E value	Best hit	Sim	%ID	E value
PsrA	85-98	75-96	0.0	73	56	0.0	W. succinogenes	65	47	0.0
PsrB	90-99	79-98	10−87-10−107	71	54	10−57	W. succinogenes	80	65	10−63
PsrC	66-99	51-96	10−57-10−137	59	40	10−38	W. succinogenes	67	45	10−62

^aPercentages of sequence similarity (Sim), percentages of identity (%ID), and expect values (E values) between S. oneidensis Psr predicted amino acid sequences obtained from TIGR. Ranges were determined by pairwise comparison with translated sequence data from draft genome sequences of 17 recently sequenced Shewanella strains, including S. putrefaciens 200, S. putrefaciens CN32, S. putrefaciens W3-18-1, S. amazonensis SB2B, S. denitrificans OS217, S. baltica OS195, S. frigidimarina NCIMB400, S. pealeana ATCC 700345, S. woodyi ATCC 51908, Shewanella strain ANA-3, Shewanella strain MR-4, Shewanella strain MR-7, S. loihica PV-4, S. halifaxensis, S. piezotolerans, S. benthica, and S. sediminis.

^bOrganism outside the genus Shewanella or Salmonella, with the ortholog of highest similarity (best hit) determined by BLASTP analysis of the nonredundant GenBank database.

FIG. 1.

Identification of conserved Pfam domains in the PsrA family of proteins. ClustalW multiple alignment of PsrA family proteins containing conserved Pfam domains. (A) Multiple alignment of Pfam domain PF04879 (4Fe-4S binding domain) with S. oneidensis MR-1 PsrA (PsrA_MR1), S. Typhimurium LT2 PhsA (PhsA_LT2), and W. succinogenes PsrA (PsrA_Ws). (B) Multiple alignments of Pfam domain PF01568 (bis-MGD binding domain) with PsrA and PhsA proteins. Identical residues are shaded.

S₂O₃²⁻-reducing members of the genus Shewanella contain Psr homologs with high amino acid sequence similarity.

The psrABC cluster was also identified in the recently sequenced genomes of Shewanella strains that respire S⁰ and S₂O₃²⁻, including S. putrefaciens CN32, S. putrefaciens 200, S. putrefaciens W3-18-1, S. amazonensis SB2B, S. frigidimarina NCIMB400, S. baltica OS155, S. pealeana, Shewanella sp. strain ANA-3, Shewanella loihica PV-4, Shewanella sp. strain MR-4, Shewanella sp. strain MR-7, S. benthica, S. halifaxensis, S. piezotolerans, and S. sediminis. The psr gene cluster was missing from the genomes of S. denitrificans OS217 and S. woodyi, a finding that correlated with the inability of these two strains to respire S₂O₃²⁻ (57). PsrA (85 to 98%), PsrB (90 to 99%), and PsrC (66 to 99%) amino acid sequence similarities were highly conserved among S₂O₃²⁻-respiring Shewanella species (Table 3).

Construction and confirmation of in-frame deletion mutant strain PSRA1 and knock-in complementation strain PSRA+.

S. oneidensis psrA was deleted in frame via application of the newly developed pKO2.0-based gene deletion system. Regions flanking psrA were PCR amplified and joined to construct a region used for gene deletion. This region was cloned into pKO2.0 (creating plasmid pKOPSRA) and subsequently mobilized into S. oneidensis via conjugal transfer. Single integrants were selected on the basis of their resistance to Gm, and their sequences were confirmed via PCR using test primers flanking the recombination region. The second step involved resolution of the single integration by counterselection on medium containing sucrose. Several sucrose-resistant colonies were screened by PCR and selected for further study. A single strain (designated PSRA1) displaying a PCR product corresponding to an in-frame psrA deletion was chosen and tested for expression of downstream genes psrB and psrC via RT-PCR. Transcripts for psrB and psrC were detected in PSRA1, while transcripts for psrA were not detected (Fig. 2).

FIG. 2.

RT-PCR of genes in the psr operon. PCR products generated from cDNA from the wild type (lane 1), PSRA1 (lane 2), or PSRA+ (lane 3) were separated electrophoretically and visualized via staining with ethidium bromide. Transcripts for psr or housekeeping (rpoA) genes are indicated above the gel lanes. Negative-control lanes 4, 5, and 6 include cDNA from the wild type, PSRA1, and PSRA+, respectively, with reverse transcriptase omitted from the cDNA synthesis reaction. Lanes M, molecular size markers.

To avoid problems associated with in trans complementation (e.g., inadvertent overexpression), the newly constructed suicide vector pKO2.0 was used to generate a “knock-in” genetic complementation strain, designated PSRA+. psrA and regions upstream and downstream (identical to those used to construct PSRA1) were PCR amplified and cloned into pKO2.0 as described above. pKOPSRA+ was used to reinsert the psrA gene into PSRA1 as described for construction of the in-frame deletion, with the exception that PSRA1 was used as the recipient strain. Insertion of wild-type psrA back into PSRA1 was confirmed via DNA sequencing and RT-PCR analyses as described above (Fig. 2).

Anaerobic growth capabilities of S. oneidensis strains PSRA1 and PSRA+.

Wild-type S. oneidensis, deletion mutant PSRA1, and complemented mutant PSRA+ were tested for anaerobic respiration on a combination of two electron donors (lactate and formate) and a set of alternate electron acceptors. PSRA1 was severely impaired in its ability to respire anaerobically on S₂O₃²⁻ and S⁰ with lactate or formate (Fig. 3 and 4) yet retained the ability to grow on all other combinations of electron donors and electron acceptors, including O₂, DMSO, Fe(III)-citrate, hydrous ferric oxide, NO₃⁻, fumarate, Mn(III)-pyrophosphate, and TMAO (see Fig. S2 in the supplemental material). Anaerobic growth of wild-type S. oneidensis and PSRA+ on S⁰-containing solid growth medium was accompanied by the production of a clearing zone in the colony periphery. The clearing zone, indicative of S⁰ reduction, was more pronounced with formate than with lactate as an electron donor (Fig. 4). Neither cell mass nor a clearing zone was observed in anaerobic incubations of PSRA1 on S⁰-containing solid growth medium.

FIG. 3.

S₂O₃²⁻ reduction phenotypes of the wild-type (MR-1) (filled circles), psrA deletion mutant (PSRA1) (open triangles), and “knock-in” complementation (PSRA+) (filled squares) strains. (A) S₂O₃²⁻ concentration as a function of time for the wild-type, PSRA1, and PSRA+ strains with lactate (18 mM) as the electron donor. (B) Cell growth as a function of time with lactate as the electron donor. (C) S₂O₃²⁻ concentration as a function of time with formate (30 mM) as the electron donor. (D) Cell growth as a function of time with formate as the electron donor.

FIG. 4.

S⁰-respiratory phenotypes of the wild-type (MR-1), psrA deletion mutant (PSRA1), and “knock-in” complementation (PSRA+) strains. Anaerobic cell growth was measured on solid medium containing S⁰ (40 mM) as the electron acceptor with either lactate (18 mM) (A) or formate (30 mM) (B) as the electron donor (90% N₂, 10% CO₂ atmosphere). Following a 48-h anaerobic incubation, plates were visualized against a black background to observe the clearing zone associated with S⁰ reduction by MR-1 and PSRA+ (but not by PSRA1).

PSRA1 retained S₄O₆²⁻-respiratory activity and consequently produced S₂O₃²⁻, although not at wild-type rates (Fig. 5). With formate as an electron donor, the wild-type strain and PSRA+ depleted S₄O₆²⁻ to low levels and switched to anaerobic respiration of the produced S₂O₃²⁻ (noted by a depletion of S₂O₃²⁻ after 24 h). With lactate as an electron donor, the wild-type strain and PSRA+ reduced S₄O₆²⁻ to S₂O₃²⁻; however, neither strain was able to respire the produced S₂O₃²⁻. PSRA1, on the other hand, reduced S₄O₆²⁻ and produced S₂O₃²⁻ but was unable to respire the produced S₂O₃²⁻ with either lactate or formate as an electron donor. As a result, PSRA1 displayed a marked decrease in growth rate and S₄O₆²⁻ reduction activity (Fig. 5).

FIG. 5.

S₄O₆²⁻-respiratory phenotypes of the wild-type (MR-1), psrA deletion mutant (PSRA1), and “knock-in” complementation (PSRA+) strains. (A and C) S₄O₆²⁻ reduction by the wild-type (filled circles), PSRA1 (filled triangles), or PSRA+ (filled squares) strain with lactate (18 mM) (A) or formate (30 mM) (C) as the electron donor. Respiration of S₄O₆²⁻ results in an accumulation of S₂O₃²⁻ by the wild-type (open circles), PSRA1 (open triangles), or PSRA+ (open squares) strain. (B and D) Cell growth as a function of time for the wild-type (filled circles), PSRA1 (open triangles), or PSRA+ (filled squares) strain with lactate (B) or formate (D) as the electron donor. In some cases, error bars are smaller than the symbol.

DISCUSSION

Metal-respiring members of the genus Shewanella provide attractive models for determining the molecular mechanism of anaerobic respiration of metals, metalloids, and radionuclides. Genetic manipulations in Shewanella can be performed aerobically prior to subsequent phenotypic tests under anaerobic conditions. With the recent release of the genomic sequences of 17 Shewanella strains (Joint Genome Institute; http://www.jgi.doe.gov), targeted, in-frame gene deletion mutagenesis may now be carried out to examine the molecular mechanism of anaerobic respiration. Earlier genetic manipulation techniques, such as transposon and chemical mutagenesis, have been problematic for a variety of reasons, including polar effects on downstream genes, inadvertent genomic rearrangements and deletions, and altered growth rates due to the energetic requirements for maintaining antibiotic resistance markers (7, 13, 43). The suicide vector constructed in the present study (pKO2.0) contains features ideal for generating in-frame, markerless gene deletions in Shewanella: pKO2.0 replicates efficiently in E. coli donor strains but not in S. oneidensis, contains an appropriate broad-host-range origin of transfer (oriT), and encodes a functional counterselection marker (SacB) (16, 18, 19). In addition, pKO2.0 is ∼5 kb smaller than previously constructed vectors (18) and contains a LacZ polylinker for more-efficient ligation of DNA constructs. Shewanella deletion mutants may therefore be constructed more efficiently due to increased transformation efficiency and elimination of problems with blunt-end cloning (19). pKO2.0 was used to construct in-frame deletions of the gene encoding the putative PsrA homolog identified in the S. oneidensis genome.

The majority of studies of anaerobic respiration by S. oneidensis have focused on the mechanism of electron transport to transition metals and radionuclides, while the mechanism of electron transport to metalloids such as inorganic sulfur compounds remains poorly understood (8, 11, 14, 46, 56). Previous studies demonstrated that several members of the Shewanella genus couple the oxidation of organic matter (or hydrogen) to the dissimilatory reduction of S⁰, a process that does not require direct contact between the cell and the S⁰ particle (41, 47). S⁰ respiration in S. Typhimurium and W. succinogenes involves a PhsA (PsrA)-dependent intraspecies sulfur cycle. In S. Typhimurium, PhsA reduces S₂O₃²⁻ to HS⁻ and SO₃²⁻. The produced SO₃²⁻ diffuses from the cell and reacts chemically with S⁰ to form additional S₂O₃²⁻, which reenters the periplasm and is reduced by PhsA to HS⁻ and SO₃²⁻, thereby sustaining an intraspecies sulfur cycle. In W. succinogenes, PsrA reduces S_n²⁻ stepwise to S_{n − 1}²⁻ and HS⁻. The HS⁻ diffuses from the cell and reacts with S⁰ to form additional S_n²⁻, which reenters the periplasm and is rereduced (24). In this manner, the terminal reduction step for S⁰ respiration by S. Typhimurium and W. succinogenes is driven by the abiotic interaction of S⁰ with either SO₃²⁻ or HS⁻ and not by direct enzymatic reduction of S⁰.

In the present study, all predicted PsrA homologs identified in the recently sequenced Shewanella genomes contain Pfam domains indicative of the Phs (Psr) family of thiosulfate reductases, including 4Fe-4S cluster and MGD binding motifs. psrA-like genes are not found in the genomes of S. denitrificans or S. woodyi, two Shewanella species unable to respire S⁰ or S₂O₃²⁻ (57). Interestingly, S. denitrificans and S. woodyi are also unable to respire Fe(III) oxides as the terminal electron acceptor (57), a pattern of electron acceptor utilization reflecting the close association of the S⁰ and Fe(III) reduction phenotypes in metal-respiring bacteria (57).

To determine if S. oneidensis psrA is required for S⁰ and S₂O₃²⁻ respiration, the S. Typhimurium phsA homolog identified in the S. oneidensis genome was deleted in frame with the new pKO2.0-based in-frame gene deletion system, and the resulting mutant was tested for anaerobic respiration on a combination of two electron donors and a set of 11 alternate electron acceptors, including S⁰ and S₂O₃²⁻. S. oneidensis deletion mutant PSRA1 is unable to respire anaerobically on S₂O₃²⁻ or S⁰ as an electron acceptor, yet it retains respiratory activity on nine other electron acceptors tested, including S₄O₆²⁻. Following depletion of S₄O₆²⁻, wild-type S. oneidensis switches to anaerobic respiration of the produced S₂O₃²⁻. PSRA1, on the other hand, is unable to respire the produced S₂O₃²⁻ and, as a result, displays decreased cell growth and S₄O₆²⁻ depletion rates (Fig. 5). This finding indicates that S₄O₆²⁻ is reduced to HS⁻ stepwise with S₂O₃²⁻ as an intermediate. Cells respiring S₄O₆²⁻ with lactate as an electron donor were unable to reduce the produced S₂O₃²⁻ (Fig. 5A). The reasons for this finding are not clear, but it may be due to repression of the Psr system by residual (1 mM) S₄O₆²⁻ remaining in the system. A similar growth pattern was recently reported for cytochrome c maturation (ccmB) mutants of S. putrefaciens 200 that are unable to reduce NO₂⁻ but retain NO₃⁻ reduction activity (11). The S. putrefaciens CCMB1 mutant (with an in-frame deletion of ccmB) stoichiometrically converts NO₃⁻ to NO₂⁻ but is unable to respire the produced NO₂⁻, and growth halts.

S₄O₆²⁻-respiring bacteria generally contain the three-gene cluster ttrBAC, in which ttrA encodes the major subunit (TtrA) of typical S₄O₆²⁻ terminal reductases. TtrA contains ferredoxin and MGD binding motifs, similarly to the polysulfide (thiosulfate) reductases PhsA and PsrA (23). The ability of PSRA1 to respire S₄O₆²⁻ indicates that PsrA does not function as the S₄O₆²⁻ reductase in S. oneidensis. Interestingly, the genomes of other S₄O₆²⁻-respiring members of the genus Shewanella contain the traditional ttrBAC gene cluster, while S. oneidensis does not (21). A possible candidate for S₄O₆²⁻ reductase in S. oneidensis is an atypical octaheme c-type cytochrome (SO4144) that displays S₄O₆²⁻, NO₂⁻, and NH₂OH reduction activities in vitro (4, 42). The physiological role of SO4144 in vivo remains to be examined.

The results of the present study indicate that pKO2.0 is an effective suicide vector for generating markerless, in-frame gene deletions in S. oneidensis MR-1. The vector is also suitable for generating “knock-in” complements of the deletion mutants, thereby avoiding problems associated with genetic complementation in trans. The results also indicate that PsrA is required for S. oneidensis to respire anaerobically on either S⁰ or S₂O₃²⁻ as an electron acceptor, most likely functioning as the S₂O₃²⁻ terminal reductase. The large clearing zone observed in the periphery of wild-type S. oneidensis colonies incubated anaerobically on solid medium amended with S⁰ corroborates previous findings that direct contact is not required for S⁰ respiration by S. oneidensis (41). S. oneidensis may therefore respire S⁰ via an intraspecies sulfur cycle catalyzed by extracellular, abiotic (purely chemical) interactions between S⁰ and either HS⁻ or SO₃²⁻ produced during anaerobic S⁰ respiration. The results of the present study are unable to differentiate between the HS⁻- and SO₃²⁻-catalyzed pathways. One potential strategy for differentiating between the HS⁻ and SO₃²⁻ cycling pathways is to construct SO₃²⁻ reductase deletion mutants. An increase in the diameters of clearing zones in the periphery of SO₃²⁻ reductase deletion mutants incubated anaerobically on S⁰-amended solid growth medium is an indication that SO₃²⁻ contributes to S⁰ mobilization in S. Typhimurium (24) (produced SO₃²⁻ is not reduced and is therefore more readily available to interact with extracellular S⁰, resulting in a larger clearing zone). This experimental strategy is not currently possible for S. oneidensis, however, since a genome-wide scan indicated that genes encoding traditional dissimilatory SO₃²⁻ terminal reductases (DsrA or AsrA) are missing from the S. oneidensis genome. Current work is focused on identification of the novel SO₃²⁻ reductase in S. oneidensis and examination of the S⁰ reduction activity of the corresponding SO₃²⁻ reductase deletion mutants.