Demonstration of the Role of Both a Ttr and a Psr Homologue Enzymes in the Respiration of Tetrathionate by an Environmental Bacterium Shewanella sp. ANA‐3
Gwendoline Degré, Audrey Tempier, Marie Vaillant, Cécile Jourlin‐Castelli, Simon Duval, Régine Lebrun, Barbara Schoepp‐Cothenet
TL;DR
This study shows that the bacterium Shewanella sp. ANA-3 uses two enzymes to respire tetrathionate, a process previously thought to be limited to gut pathogens.
Contribution
The first evidence that a Psr homologue can catalyse tetrathionate respiration in an environmental microorganism.
Findings
Shewanella sp. ANA-3 uses Ttr and Psr homologues to respire tetrathionate.
Neither OTR nor Tsd participate in tetrathionate respiration in this species.
The Ttr's role in tetrathionate respiration is confirmed in an environmental microorganism for the first time.
Abstract
The broad metabolic capacity of Shewanella species allows them to colonise a wide range of aquatic niches. Their ability to convert sulphur compounds, including tetrathionate, has been reported. This may seem surprising, as this ability has long been considered restricted to human gut pathogens such as Enterobacteria and tetrathionate is considered unstable in the external environment. The molecular basis of Shewanella growth on tetrathionate had never been analysed. By combining the construction and metabolic characterisation of deletion mutants and complementary biochemical analyses, we determined that Shewanella sp. ANA‐3 uses two enzymes, homologous to the tetrathionate reductase Ttr and the polysulphide reductase Psr, respectively, to respire tetrathionate. This study provides the first evidence that a Psr homologue can catalyse this reaction. Neither the octahaem tetrathionate…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
FIGURE 1
FIGURE 2
FIGURE 3
FIGURE 4
FIGURE 5
FIGURE 6| Strains | Genotypes | References |
|---|---|---|
|
| ||
| C600 | F‐ tonA21 thi‐1 thr‐1 leuB6 lacY1 glnV44 rfbC1 fhuA1 λ— | (Appleyard |
| CC118 λpir | Δ( | (Herrero et al. |
| 1047/pRK2013 | Helper strain | (Figurski and Helinski |
| ANA‐3 | ||
| ANA‐3 | Wild Type | (Saltikov et al. |
|
| ANA‐3 deleted of | This work |
|
| ANA‐3 deleted of | This work |
|
| ANA‐3 deleted of | This work |
|
| ANA‐3 deleted of | This work |
|
| ANA‐3 deleted of | This work |
| Proteins identified in ANA‐3 | Accession numbers | Genes (locus tags) |
|---|---|---|
| TtrB | SHEWANA3_RS03355 | |
| TtrC | SHEWANA3_RS03350 | |
| TtrA | SHEWANA3_RS03345 | |
| PsrA | SHEWANA3_RS18910 | |
| PsrB | SHEWANA3_RS18905 | |
| PsrC | SHEWANA3_RS18900 | |
| OTR | SHEWANA3_RS19265 | |
| SirA | SHEWANA3_RS02580 | |
| SirC | SHEWANA3_RS02600 | |
| SirD | SHEWANA3_RS02605 | |
| CysN | SHEWANA3_RS04520 | |
| CysC | SHEWANA3_RS04530 | |
| CysH | SHEWANA3_RS04485 | |
| TsdA | SHEWANA3_RS18845 | |
| TsdB | SHEWANA3_RS18850 |
- —Agence Nationale de la Recherche10.13039/501100001665
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsSulfur Compounds in Biology · Nitrogen and Sulfur Effects on Brassica · Odor and Emission Control Technologies
Introduction
1
Shewanella species, around 70 isolated today, are metabolically so versatile as to be able to colonise a wide range of aquatic ecological niches, including those which are chemically polluted (Lemaire et al. 2020), which makes the genus a formidable tool for the remediation of environmental pollutants (Yin et al. 2022) and has led to a large number of studies on its species including the characterisation of their energy metabolism. These facultative anaerobic bacteria have the capacity to convert a wide range of inorganic and organic compounds using the numerous oxidases and reductases they carry. The abundance of these enzymes in certain species can complicate efforts to establish the precise enzymatic network behind the fate of a chemical compound in the cell with sometimes somehow redundant or at least overlapping enzymatic activities (Xiong et al. 2017; Liu et al. 2021; Yin et al. 2022) for examples. This is particularly true for sulphur metabolism.
Certain Shewanella species have been shown able to oxidise thiosulphate (S_2_O_3_ ^2−^) and sulphite (SO_3_ ^2−^), using the thiosulphate dehydrogenase Tsd, a cytochrome enzyme, and the sulphite dehydrogenase SDH, a molybdenum (Mo)‐enzyme, respectively (Yu et al. 2021; Sun et al. 2023). Some others have been demonstrated to anaerobically respire polysulphide (S_n_ ^2^), elemental sulphur (S^0^) and S_2_O_3_ ^2−^ using an enzyme homologous to the Mo‐enzyme polysulphide reductase Psr (Burns and DiChristina 2009) and the flavin‐enzyme N‐Psr (Warner et al. 2011), respire SO_3_ ^2−^ using the cytochrome sulphite reductase SirA (Shirodkar et al. 2011), and respire dimethyl sulphoxide (DMSO) using the Mo‐enzyme Dms (McCrindle et al. 2005; Xiong et al. 2017). Finally, the aquatic S. oneidensis MR‐1 has been shown to respire tetrathionate (S_4_O_6_ ^2−^) although this ability has been long considered as a prerogative of human gut pathogens such as the Enterobacteriaceae including Salmonella enterica (Hensel et al. 1999), since S_4_O_6_ ^2−^ is a well‐known compound produced during intestinal inflammation where this respiration could confer a competitive advantage to the pathogen. Indeed, the ability of certain environmental bacteria to reduce tetrathionate has been recognised for decades and has been published in multiple studies (Barrett and Clark 1987; Mandal et al. 2020). While the enzyme responsible for the S_4_O_6_ ^2−^ reduction, the tetrathionate reductase (Ttr), has been identified and characterised in Salmonella enterica (Hensel et al. 1999; Hinsley and Berks 2002), its functional role in environmental bacteria, including Shewanella species, remains to be established.
The soluble Octahaem Tetrathionate Reductase (OTR) from S. oneidensis MR‐1 has been overexpressed, purified and crystallised, and demonstrated to reduce S_4_O_6_ ^2−^ in vitro (Mowat et al. 2004; Atkinson et al. 2007) and proposed as the enzyme responsible for in vivo S_4_O_6_ ^2−^ reduction (Burns and DiChristina 2009). However, the physiological function of this cytochrome has never been established. The best‐known enzymatic system capable of anaerobically reducing S_4_O_6_ ^2−^ into S_2_O_3_ ^2−^ is the membrane‐bound, quinone‐reacting Mo‐enzyme TtrABC, which has been characterised in S. enterica . Since homologues of ttr genes have been identified in phyla other than Enterobacteriaceae (Radl et al. 2019; Vavourakis et al. 2019; Boughanemi et al. 2020; Calisto and Pereira 2021; Flieder et al. 2021; Hashimoto et al. 2022; Reji et al. 2022; Li et al. 2023; He et al. 2025) the Ttr‐based S_4_O_6_ ^2−^ respiration could also exist in other phyla, including Shewanella. In addition to Ttr, Tsd cytochrome is also a promising candidate for S_4_O_6_ ^2−^ reduction in Shewanella. The Tsd has been well characterised in relation to aerobic S_2_O_3_ ^2−^ oxidation in S. oneidensis (Yu et al. 2021), but has also been shown to be responsible for anaerobic reduction of S_4_O_6_ ^2−^ in Wolinella succinogenes and Campylobacter jejuni (Liu et al. 2013; Kurth et al. 2017).
The objective of the present work was to decipher the S_4_O_6_ ^2−^ reduction process in Shewanella at a molecular level, using Shewanella sp. ANA‐3 (abbreviated as ANA‐3 hereafter). This strain has been shown to be able to reduce a number of compounds, including arsenate, antimonate, iodate and iron (Saltikov and Newman 2003; Reyes et al. 2012; Wang et al. 2020; Jiang et al. 2023) but its ability to reduce sulphur compounds has never been examined. Following a detailed analysis of the genome to identify all the enzymes likely to be directly or indirectly involved in S_4_O_6_ ^2−^ metabolism, we analysed the phenotypes of deletion mutants obtained using the suicide cloning vector pKNG101 for in‐frame gene deletion mutagenesis (Kaniga et al. 1991) in ANA‐3. We deleted each gene cluster that was potentially involved in the S_4_O_6_ ^2−^ metabolism or deleted a combination of them, and analysed their respiratory capacity in relation to S_2_O_3_ ^2−^ and S_4_O_6_ ^2−^. We also examined the in vitro enzymatic activities of the putative S_4_O_6_ ^2−^ respiratory enzymes. We discussed the function of the widespread Ttr‐like enzymes in the light of sequence analyses.
Experimental Procedures
2
Tetrathionate Sources
2.1
Three sources of S_4_O_6_ ^2−^ have been used for this study: the product currently sold by Sigma Aldrich‐Merck (reference P2926), the product currently sold by Genaxxon Bioscience (reference M6007.0100), and a 20‐year‐old product from Merck (reference 5169) no longer available on the market. While the K_2_S_4_O_6_ powder from Genaxxon is almost fully converted instantly into S_3_O_6_ ^2−^ when dissolved in aqueous solutions, the K_2_S_4_O_6_ currently sold by Sigma Aldrich‐Merck or the one sold in the past by Merck showed significantly better stability in solution, with the best chemical stability for the old product no longer commercially available. This one has been consequently used in our study.
Strains, Plasmids and Primers
2.2
Bacterial strains of Escherichia coli and ANA‐3 as well as the plasmids used or constructed in this study are listed in Table 1. Primers used for all constructions are listed in Table SI‐1.
Growth Conditions
2.3
The different strains of ANA‐3 were pre‐cultured overnight in Luria‐Bertani (LB) medium at 28°C under aerobic conditions. For anaerobic growth, a minimal medium supplemented with vitamins and a SL‐10 trace element solution was used as described by Saltikov et al. (2003). Depending on the assays, 30 mM formate was added as electron donor, along with either 5 mM S_4_O_6_ ^2−^ or 10 mM S_2_O_3_ ^2−^ as electron acceptors. The medium was inoculated with 1% of an overnight aerobic culture. The cultures were then incubated at 30°C, which was determined in this study to be the optimal growth temperature, and continuously flushed with nitrogen. For strains harbouring plasmids, the appropriate antibiotic was added prior to inoculation. E. coli strains were cultured in LB medium containing the relevant antibiotics and incubated overnight at 37°C under aerobic conditions. The antibiotic concentrations used were: 100 μg/mL of streptomycin, 25 μg/mL of kanamycin, 25 μg/mL of chloramphenicol and 50 μg/mL of ampicillin. Shewanella is naturally resistant to ampicillin. Thus, the use of this antibiotic completes the colicin selection effect in deletion mutagenesis (see below).
Metabolic Monitoring
2.4
A 1 mL sample of the culture was taken at different growth times. To stabilise sulphur (S) compounds, a final concentration of 1% isopropanol was added. The sample was centrifuged for 10 min at 13,000 g to remove the cells, after which the supernatant was stored at −20°C. The concentrations of S compounds, including SO_3_ ^2−^, SO_4_ ^2−^, S_2_O_3_ ^2−^, S_3_O_6_ ^2−^ and S_4_O_6_ ^2−^ were quantified by ion‐exchange chromatography (ECO‐IC Metrohm). The thawed supernatants were diluted 1:500 in a 1% aqueous isopropanol solution, and filtered through a 0.2 μm membrane. The elution rate was set to 1 mL/min. An isocratic elution method was used with an eluent of 1.8 mM Na_2_CO_3_/1.7 mM NaHCO_3_/20% acetone, which facilitated the separation of the compounds based on their size and charge. The individual peaks were identified based on their retention times using a calibration curve, and the concentration of each compound was determined by measuring the area under the peak. To determine the concentration of trithionate (S_3_O_6_ ^2−^), a specific calibration was performed during this study. To this end, solutions of varying concentration of the K_2_S_4_O_6_ product (reference M6007.0100) from Genaxxon Bioscience, were injected; this product converts completely into S_3_O_6_ ^2−^ once in solution.
Mutagenesis
2.5
Deletion mutant strains of ANA‐3 were generated to investigate the role of each enzyme in sulphur metabolism. Gene deletions were performed using a suicide vector, pKNG101, which carries the sacB cassette and the strAB genes that encode the streptomycin phosphotransferase (Kaniga et al. 1991) as previously described in detail (Baraquet et al. 2009). Briefly, the 500 bp‐regions upstream and downstream flanking the genes to be deleted were amplified using ANA‐3 genomic DNA. A third amplicon was obtained by fusing the upstream and downstream fragments via a third PCR. The amplicon was digested with ApaI and BamHI and cloned into pKNG101. The ligation product was introduced into chemically competent E. coli CC118λpir cells. Transformants were selected on agar plates containing streptomycin and screened by PCR. The resulting recombinant plasmid was then introduced into the relevant ANA‐3 strain via conjugation using the E. coli helper strain 1047/pRK2013. Shewanella integrant cells were selected on a medium containing colicin and ampicillin (to prevent the growth of E. coli donor cells) as well as streptomycin. Colonies were selected for the second homologous recombination, which produced the deletion, on LB agar supplemented with 6.25% sucrose. The deletants were screened for their sensitivity to streptomycin, then screened by PCR and finally sequenced.
Complementation
2.6
The ttrBCA and the psrABC gene clusters, which were amplified from ANA‐3 genomic DNA, were introduced between the restriction sites BamHI and XhoI. They were then cloned into pBBR1MCS2 and introduced into chemically competent E. coli C600 or DH5α via heat shock treatment. Transformants were selected using kanamycin, screened by PCR, and finally sequenced. The resulting plasmids were introduced by conjugation into ANA‐3 ΔttrΔpsr and ANA‐3 Δpsr.
In Gel Activity Assay
2.7
ANA‐3 Δttr and ANA‐3 Δpsr were grown on 10 mM S_2_O_3_ ^2−^ or 5 mM S_4_O_6_ ^2−^, respectively, for 24 h. Cells were then centrifuged at 4000 g for 30 min and washed in a 50 mM HEPES buffer (pH 7). The cells were then resuspended in a 50 mM HEPES buffer (pH 7) containing protease inhibitors (Sigma). Cells were lysed by passing three times through a French press at 1000 psi. Unbroken cells were removed by centrifugation at 6000 g for 5 min at 4°C. The “soluble fraction” and “membrane fraction” were separated using ultracentrifugation for 1 h at 180,000 g at 4°C. The resulting supernatant, representing the soluble fraction, was concentrated using a 30 kDa cut‐off ultrafiltration unit (Millipore). The pellet, corresponding to the membrane fraction, was resuspended in 50 mM HEPES buffer (pH 7) containing 1% Triton X‐100 and incubated for 30 min at 4°C. After incubation, this sample was centrifuged at 10,000 g for 10 min and the pellet discarded. Both the soluble and membrane fractions were loaded onto a native polyacrylamide gel (7% running, 5% stacking) as described previously (Guiral et al. 2009). The gel was stained in a MES 30 mM/HEPES 30 mM/Tricine 30 mM/AMPSO 30 mM/NaCl 100 mM buffer (pH 7) under anoxic conditions, in the presence of 2 mM reduced methyl viologen (as reduced by titanium citrate). Finally, electron acceptors, 2.5 mM S_2_O_3_ ^2−^ or S_4_O_6_ ^2−^ were added. De‐stained bands were identified by proteomic analysis.
Proteomic Analysis
2.8
In gel proteins were digested with the Trypsin/LysC mix from Pseudomonas aeruginosa (Promega), following reduction and alkylation steps (Santin et al. 2018). The tryptic peptides were injected into an EasySpray PepMapNeo column (1500 bars, 75 μm × 500 mm, C18, 2 μm, 100 A) (Thermo Fisher Scientific) using a Vanquish Neo UHPLC system (Thermo Fisher Scientific, Germany), and separated using a two step‐gradient 2%–25% in 90 min, then 25%–50% in 20 min, of mobile phase B (0.1% (v/v) formic acid/80% (v/v) acetonitrile) in mobile phase A (0.1% (v/v) formic acid). The eluting peptides were detected and analysed using an ESI‐Q‐Exactive Plus mass spectrometer (Thermo Fisher Scientific), with the same parameters as previously described (Santin et al. 2018). The data were searched using the Proteome Discoverer 3.0.1.27 software (Thermo Fisher Scientific) and the Sequest algorithm for protein identification. The following global settings were used: ANA‐3 database (TxID 22), downloaded from NCBI last modified on 04/02/2021 (4736 entries), amino acid modifications; and fixed carbamidomethylation on cysteine and variable oxidation on methionine and variable acetylation on protein N‐term. Proteins were identified if a minimum of two unique peptide sequences, each consisting of more than six amino acids, passed the high confidence filter.
Phylogenetic Analysis
2.9
For the phylogenetic analysis, the open reading frames that code for the Ttr and Psr subunits homologues were retrieved from the National Centre for Biotechnology Information (http://www.ncbi.nlm.nih.gov) via a BLAST search. The Soe, Aio, Arr and Arx sequences originate from previous studies (Duval et al. 2008; Boughanemi et al. 2020; Szyttenholm et al. 2020). The TtrB/C/A sequences from Salmonella enterica , Desulfolithobacter dissulfuricans GF1T, Pyrococcus aerophilum and Anaeromyxobacter sp. strain PSR‐1 (Hensel et al. 1999; Haja et al. 2020; Muramatsu et al. 2020; Hashimoto et al. 2022) were used as query templates for Ttr‐homologues. The PsrA/B/C sequences from W. succinogenes and Thermus thermophilus as well as the thiosulphate reductase PhsA/B/C from S. enterica were used as query templates for Phs (Krafft et al. 1992; Heinzinger et al. 1995; Jormakka et al. 2008). The sulphur reductase SreA/B/C sequences from Acidianus ambivalens and HSR2 (Laska et al. 2003; Sorokin et al. 2016) were used as query templates for Sre‐homologue sequences.
A multiple sequence alignment was produced using Expresso, a special mode of the T‐Coffee server that produces structure‐guided sequence alignments (Di Tommaso et al. 2011). This alignment was refined using Seaview (Gouy et al. 2010) based on structural alignments obtained using PYMOL. Structures from Arr [6CZ7], Aio [4AAY], and Psr [2VPZ] were obtained from the pdb database (http://www.rcsb.org/pdb/welcome.do). A phylogenetic tree was reconstructed from this alignment using the Maximum Likelihood method implemented in MEGAX.
Results
3
Genomic Analyses of Shewanella sp. ANA‐3 Reveals Several Potential Enzymes Involved in Tetrathionate Reduction
3.1
To elucidate the metabolisms of S_4_O_6_ ^2−^ and S_2_O_3_ ^2−^ (the latter's reduction product) in ANA‐3, it was essential to first establish the bacterium's genomic content in genes encoding enzymes likely to be involved in the conversion of these compounds and their products. Although several sulphur metabolism enzymes have been identified in Shewanella species, only a SO_3_ ^2−^ reductase Sir had been identified in ANA‐3 (Shirodkar et al. 2011). Using a BLAST analysis approach with sequences from already characterised enzymes, a polysulphide reductase Psr‐homologue and a tetrathionate reductase Ttr‐homologue were identified as well as the Octahaem Tetrathionate Reductase OTR (Table 2) (Krafft et al. 1992; Hensel et al. 1999; Atkinson et al. 2007). A TsdBA system has been identified that could be responsible for both the reduction of S_4_O_6_ ^2−^ and oxidation of S_2_O_3_ ^2−^ as exposed in the Introduction section (Yu et al. 2021). Since SO_3_ ^2−^ is a marker for the combined functions of Ttr, OTR, and Psr, to be monitored, we looked for enzymes that could interfere with its evolution over time by consuming or producing SO_3_ ^2−^ in addition to the Sir. Part of the cytoplasmic assimilatory reduction pathway of sulphate (SO_4_ ^2−^) was revealed, with the ATP sulphurylase CysN, the APS kinase CysC, and the PAPS reductase CysH, allowing the production of SO_3_ ^2−^ from SO_4_ ^2−^ (Frigaard and Dahl 2009).
The Identified Ttr‐Homologue Is Involved in Tetrathionate Metabolism
3.2
The triad SHEWANA3_RS03345‐ SHEWANA3_RS03350‐ SHEWANA3_RS03355 was identified as a potential ttrBCA gene cluster, suggesting that ANA‐3 might be able to respire S_4_O_6_ ^2−^ as was previously observed with S. oneidensis MR‐1. To decipher the molecular basis of S_4_O_6_ ^2−^ respiration, K_2_S_4_O_6_ was tested as an anaerobic respiratory substrate in a minimum medium. Even under anaerobic conditions, S_4_O_6_ ^2−^ has been shown to undergo abiotic conversion over time into S_3_O_6_ ^2−^ and S_2_O_3_ ^2−^ (Hinsley and Berks 2002). The kinetics of its abiotic conversion at 30°C in a growth medium free of bacteria resulted in about 17% decomposition within 9 h (Figure 1A). In line with previous studies, the products of this abiotic conversion are S_3_O_6_ ^2−^ and S_2_O_3_ ^2−^ (Figure 1B–D).
Phenotypes of wild‐type and single deletion mutants of ANA‐3 when grown on S4O6 2− 5 mM. Evolution over time of (A) Tetrathionate (S4O6 2−), (B) Trithionate (S3O6 2−), (C) Thiosulphate (S2O3 2−) and (D) Sulphite (SO3 2−) concentrations in wild‐type (orange circles), Δttr (yellow triangles), Δpsr (purple inverted triangles) cultures and chemical controls (grey squares).
Respiration of S_4_O_6_ ^2−^ by the wild‐type ANA‐3 strain resulted in its total reduction within 9 h (Figure 1A). The sequential reduction products of this respiration are S_3_O_6_ ^2−^, S_2_O_3_ ^2−^, SO_3_ ^2−^ (Figure 1B–D; a representation of the global metabolism by strain is also available in Figure SI‐1). The S_3_O_6_ ^2−^ concentration as a function of time during the culture of the wild‐type strain is biphasic. The initial phase (0–6 h) corresponds to the transient accumulation of 0.8 mM S_3_O_6_ ^2−^ from the 5 mM S_4_O_6_ ^2−^, while the second phase (6–9 h), which was not observed in the chemical controls, corresponds to the predominant biological respiration of S_3_O_6_ ^2−^ in parallel with direct respiration of S_4_O_6_ ^2−^ (Figure 1B) generating S_2_O_3_ ^2−^ (Figure 1C). This suggests the involvement of a Ttr‐type enzyme. Interestingly, the initial production phase of S_3_O_6_ ^2−^ was faster in the wild‐type ANA‐3 strain than in the chemical controls. This could be explained by the biological production of SO_3_ ^2−^ (Figure 1D), since SO_3_ ^2−^ is known to react with S_4_O_6_ ^2−^ to produce S_3_O_6_ ^2−^ (Pan et al. 2019). During the final phase of the S_4_O_6_ ^2−^ metabolism (9–58 h) in the ANA‐3 wild‐type strain, accumulated S_2_O_3_ ^2−^ is consumed (Figure 1C) to produce SO_3_ ^2−^ (Figure 1D) and a black precipitate (followed by O.D; not shown) in a strictly correlated manner. This black precipitate is the typical FeS formed from the reaction of HS^−^ with Fe^2+^ in the medium (see (Xu et al. 2025) for recent review). The generation of both SO_3_ ^2−^ and HS^−^ from S_2_O_3_ ^2−^ suggests the involvement of the Psr in the S_2_O_3_ ^2−^ metabolism.
To verify the role of the identified ttrBCA homologue genes in the respiration of S_4_O_6_ ^2−^, as suggested by the kinetics of production/reduction of S_3_O_6_ ^2−^, S_2_O_3_ ^2−^ and SO_3_ ^2−^, the ttrBCA homologue genes in ANA‐3 were deleted by using the suicide cloning vector pKNG101 for in‐frame gene deletion mutagenesis as described in the “Experimental procedures” Section. The reduction of S_4_O_6_ ^2−^ by the resulting deletion mutant (ANA‐3 Δttr) was significantly slower than that of the wild‐type, though not completely abolished (Figure 1A). The kinetics of S_3_O_6_ ^2−^ production and reduction, as well as S_2_O_3_ ^2−^ production, were also significantly slowed down, while S_2_O_3_ ^2−^ reduction remained unaffected (Figure 1B,C). These results suggest that the Ttr‐homologue plays a key role in the reduction of S_4_O_6_ ^2−^ and S_3_O_6_ ^2−^, as previously demonstrated for the Ttr in S. enterica (Hinsley and Berks 2002). The delay in the production of SO_3_ ^2−^ in the Δttr mutant is perfectly consistent with the deletion of a typical Ttr enzyme, which converts S_3_O_6_ ^2−^ to both S_2_O_3_ ^2−^ and SO_3_ ^2−^. However, these results imply that the Ttr‐homologue is not the sole enzyme responsible for reducing S_4_O_6_ ^2−^ in ANA‐3. Furthermore, they suggest that this enzyme is not involved in the S_2_O_3_ ^2−^ reduction.
To identify additional systems involved in S_4_O_6_ ^2−^ respiration, the psrABC homologue genes were deleted from ANA‐3 using the suicide cloning vector pKNG101 for in‐frame gene deletion mutagenesis. The fates of the resulting Δpsr mutant for S_4_O_6_ ^2−^, S_3_O_6_ ^2−^ and S_2_O_3_ ^2−^ remained almost identical compared with those of the wild‐type. However, the mutant exhibited a complete abolition of both consumption of S_2_O_3_ ^2−^and production of SO_3_ ^2−^ (Figure 1C,D). These results suggest that the psr homologue genes are the only ones involved in the respiration of S_2_O_3_ ^2−^ in ANA‐3 and that they are not involved in the reduction of S_4_O_6_ ^2−^.
The Identified Ttr‐Homologue Clusters With the
S. enterica Ttr‐Enzyme
3.3
The observation of a S_4_O_6_ ^2−^ reduction based only partially on ttr homologue genes identified in the genome from ANA‐3 led us to analyse further the identified sequences. The typical Ttr (as exemplified by that characterised in S. enterica ) is a trimeric enzyme, member of the DMSO reductase family. Named “the workhorse of the bioenergetics”, this enzyme family contains enzymes with very diverse substrate specificity: arsenic, nitrogen, carbon, selenium, sulphur compounds (Rothery et al. 2008; Grimaldi et al. 2013). Ttr is anchored in the cytoplasmic membrane by a subunit, TtrC, without any cofactors. The catalytic subunit, TtrA, carries both a Mo cofactor and an iron–sulphur (Fe‐S) cluster, while the electron transfer subunit, TtrB, carries four Fe‐S clusters. Although the three Ttr subunits are homologue to their counterparts in other Mo‐enzymes, they show a specific gene organisation in the ttr gene cluster (ttrBCA) and primary sequence differences that enable specific recognition. Notably these include the presence of the TAT leader sequence on both the catalytic and the electron transfer subunits and the nine helices in the membrane subunit (Hensel et al. 1999; Rothery et al. 2008). The Ttr sequences identified in ANA‐3 meet all these criteria. However, the literature reports such Ttr homologues that actually are As(V) reductases, phylogenetically distinct from both true Ttrs and true Arrs (Haja et al. 2020; Muramatsu et al. 2020). A Ttr‐homologue has furthermore been proposed as functioning in S_2_O_3_ ^2−^ disproportionation (Hashimoto et al. 2022). We reconstructed a phylogenetic tree that included homologues from S_4_O_6_ ^2−^ respiring S. enterica , S_2_O_3_ ^2−^ disproportionating Desulfolithobacter dissulfuricans GF1T, the As(V) respiring Pyrococcus aerophilum and Anaeromyxobacter sp. Strain PSR‐1, as well as the protein WP_011715816.1 identified in ANA‐3 (Table 2). The reconstructed tree reveals a great phylogenetic diversity among Ttr‐like sequences suggesting diverse functions. The Ttr‐like As(V) reductases form two distinct phylogenetic clades, as previously observed (Gavrilov et al. 2017; Muramatsu et al. 2020). The sequences identified in bacteria that disproportionate sulphur compounds (highlighted in Figure 2, panel A with red arrows) appear to represent a variety of cases, including potential As^V^ reductases, true S_4_O_6_ ^2−^ reductases and two new clades of unknown homologues (highlighted in Figure 2, panel A with question marks). Within the Ttr‐subtree, the WP_011715816.1 sequence from ANA‐3 clusters with the bona fide Ttr (see Figure 2, panel A). Preliminary proteomic analysis results (data not shown) also suggest the specific over‐production of the Ttr‐homologue under S_4_O_6_ ^2−^ and not under As^V^ or S_2_O_3_ ^2−^. These results are consistent with the S_4_O_6_ ^2−^ reductase activity attributed to the SHEWANA3_RS03355‐ SHEWANA3_RS03350‐ SHEWANA3_RS03345 gene triad (Table 2), as determined by the mutagenesis experiments.
*Analyses of the Ttr, Psr and Arr sequences from ANA‐3. (A) Phylogenetic analysis of the proteins WP_011715816.1 (), WP_041412810.1 () and ArrA () relative to known Mo‐enzymes subfamilies and characterised members within the Mo‐bisPGD enzymes family. Maximum likelihood phylogenetic tree was reconstructed using MEGAX. AioA, arsenite oxidase; ArrA, arsenate reductase (from Chrysiogenes arsenatis ), ArxA, alternative arsenite oxidase; PhsA, thiosulfate reductase (from Salmonella enterica ); PsrA, polysulfide reductase (from Thermus thermophilus and from Wolinella succinogenes ); Sre, sulphur reductase; TtrA, tetrathionate reductase (from Salmonella enterica , from Desulfolithobacter dissulfuricans GF1T) and Ttr‐like arsenate reductases (from Anaeromyxobacter sp. PSR‐1 and
Pyrobaculum aerophilum ). Selected Ttr‐like sequences identified in sulphur compounds disproportionating bacteria are indicated by a red arrow (B) Sequence alignment analysis of the binding site from bona fide Arr and Ttr compared with that from Ttr‐like arsenate reductases. The absence of the GR motif in the bona fide Ttr and in the WP_011715816.1 protein from ANA‐3 appears to distinguish these enzymes from all the enzymes able to bind arsenic.*
The ΔttrΔpsr
Mutant is Unable to Respire Tetrathionate
3.4
To further investigate other(s) enzyme(s) potentially involved in the reduction of S_4_O_6_ ^2−^ in addition to the Ttr, several double deletion mutant strains of ANA‐3 were constructed, including ΔttrΔtsdA, ΔttrΔotr, and ΔttrΔpsr. While Tsd has been established as a S_2_O_3_ ^2−^ oxidase in the presence of oxygen in S. oneidensis MR‐1 (Yu et al. 2021), it has been proposed as a S_4_O_6_ ^2−^ reductase under oxygen‐limited conditions (Liu et al. 2013). The Tsd system is a soluble monomeric or dimeric enzyme consisting of one (TsdA) or two (TsdA‐TsdB) dihaem cytochromes (Kurth et al. 2016). In Shewanella species, including ANA‐3, the Tsd exists in the dimeric form (Yu et al. 2021). In the ANA‐3 ΔttrΔtsdA strain, the global metabolism of S_4_O_6_ ^2−^, including the fate of S_3_O_6_ ^2−^, S_2_O_3_ ^2−^ and SO_3_ ^2−^ was almost identical to that observed in the ANA‐3 Δttr strain (see Figure 3), suggesting that Tsd does not play a role in S_4_O_6_ ^2−^ metabolism under anoxic conditions. This is consistent with the presence of a tsdB gene alongside the tsdA gene in ANA‐3, whereas strains that use TsdA for the respiration of S_4_O_6_ ^2−^ carry only the tsdA gene (Liu et al. 2013; Kurth et al. 2016).
Phenotypes of double deletion mutants of ANA‐3 when grown on S4O6 2− 5 mM. Evolution over time of (A) Tetrathionate (S4O6 2−), (B) Trithionate (S3O6 2−), (C) Thiosulphate (S2O3 2−) and (D) Sulphite (SO3 2−) in ΔttrΔtsdA (green lozenges), ΔttrΔotr (red triangles), ΔttrΔpsr (blue triangles), ΔttrΔpsr/pBBR1MCS2‐ttrBCA (cyan hexagons) and ΔttrΔpsr/pBBR1MCS2‐psrABC (magenta stars) cultures.
As OTR has been suggested to act as a S_4_O_6_ ^2−^ reductase (Mowat et al. 2004; Atkinson et al. 2007), we investigated the capacity of an ANA‐3 ΔttrΔotr strain to convert S_4_O_6_ ^2−^. The consumption of S_2_O_6_ ^2−^ and the production of SO_3_ ^2−^of this double mutant strain were increased, which could be interpreted as an implication of OTR in the metabolism of sulphur compounds. However, the fates of S_4_O_6_ ^2−^ and S_3_O_6_ ^2−^ in this double mutant strain were almost identical to those observed in the strain lacking genes coding for the Tt‐r‐homologue (compare Figure 3 to Figure 1). These results suggest the absence of any role of OTR in the metabolism of S_4_O_6_ ^2−^ and S_3_O_6_ ^2−^, and are consistent with those obtained by Wu in her thesis (Wu 2010). Additionally, preliminary proteomic analyses (not shown) suggested that OTR production is not significantly induced by S_4_O_6_ ^2−^.
Considering that the Psr‐homologue could be indirectly involved in the metabolism of S_4_O_6_ ^2−^, through respiration of S_2_O_3_ ^2−^ formed either abiotically or biologically from S_4_O_6_ ^2−^, thus shifting the chemical equilibrium between S_4_O_6_ ^2−^/ S_2_O_3_ ^2−^, we finally analysed the conversion of S_4_O_6_ ^2−^ in an ANA‐3 ΔttrΔpsr strain. This revealed a significant slowdown of the overall S_4_O_6_ ^2−^ metabolism (Figure 3A), including slowdown of S_4_O_6_ ^2−^ and S_2_O_3_ ^2−^ reduction, as well as the production of SO_3_ ^2−^, down to the level of the chemical controls (see Figure 1). Complementation in trans of this double deletion strain with a plasmid carrying the ttr gene cluster (pBBR1MCS2‐ttrBCA) restored the S_4_O_6_ ^2−^ reduction, and the one with a plasmid carrying the psr gene cluster (pBBR1MCS2‐psrABC) restored both the S_4_O_6_ ^2−^ and S_2_O_3_ ^2−^ reductions (Figure 3). These results suggest that the identified Psr plays a role in S_4_O_6_ ^2−^ metabolism alongside the Ttr‐homologue.
The Psr‐Homologue Is the Only Enzyme Involved in Thiosulfate Respiration
3.5
To confirm that the psr homologue genes identified in the ANA‐3 genome encode the unique enzyme responsible for S_2_O_3_ ^2−^ respiration, we analysed the direct conversion of S_2_O_3_ ^2−^ by the Δpsr single mutant, comparing it to the wild‐type strain and the abiotic control. The ANA‐3 wild‐type strain exhibited S_2_O_3_ ^2−^ consumption and SO_3_ ^2−^ production consistent with the ability of the identified Psr‐homologue to reduce S_2_O_3_ ^2−^ as demonstrated in S. oneidensis MR‐1 (Figure 4). However, it should be noted that the kinetics of S_2_O_3_ ^2−^ reduction in our study are slower than previously observed (Burns and DiChristina 2009). This discrepancy can easily be explained by the use of a different strain and growth conditions, such as the volume of culture, the N_2_ flushing method, and the concentration of substrates. Deletion of the psr genes resulted in the stabilisation of S_2_O_3_ ^2−^ concentrations with no more reduction. In the ANA‐3 Δpsr strain, only trace amounts of SO_3_ ^2−^ were detected, suggesting that Psr‐homolgue is the primary source of SO_3_ ^2−^ production in the wild‐type strain, with minimal contribution from the Ttr‐homologue (through the reduction of S_3_O_6_ ^2−^). The absence of FeS (black precipitates) in the ANA‐3 Δpsr culture, an indirect indicator of HS^−^ production, further supports the idea that HS^−^ is mainly produced by the Psr‐homologue. Complementation in trans of the ANA‐3 Δpsr strain with a plasmid carrying the psr gene cluster (pBBR1MCS2‐psrABC) restored the wild‐type phenotype (Figure 4). These results demonstrate firstly that Psr is indeed the enzyme responsible for the S_2_O_3_ ^2−^ respiration and secondly that it is the only one.
Phenotypes of wild‐type and deletion mutants of ANA‐3 when grown on S2O3 2− 10 mM. (A) Thiosulphate (S2O3 2−) consumption and (B) Sulphite (SO3 2−) production of wild‐type strain (orange circles), Δpsr mutant (purple triangles), Δpsr/pBBR1MCS2‐psrABC (magenta stars) and chemical controls (grey squares). Neither tetrathionate nor trithionate was observed in these growth conditions.
In the ANA‐3 Δpsr strain culture under S_4_O_6_ ^2−^, S_2_O_3_ ^2−^ produced from S_4_O_6_ ^2−^ (tentatively attributed to Ttr) was no longer consumed (Figure 1B). This further supports the hypothesis that only the Psr is involved in S_2_O_3_ ^2−^ reduction. It should be noted that while the double deletion mutant ANA‐3 ΔttrΔpsr strain showed a significantly slower S_4_O_6_ ^2−^ respiration than the single deletion mutant ANA‐3 Δttr, the single deletion mutant ANA‐3 Δpsr showed no difference in S_4_O_6_ ^2−^respiration compared with the wild‐type. This could be interpreted as suggesting that the activity from the Ttr towards S_4_O_6_ ^2−^ is so high that the substrate availability is limiting.
The Psr‐Homologue Closely Resembles a Bona Fide Psr
3.6
As studies on Shewanella strains ((Burns and DiChristina 2009); present study) are the only ones to examine the ability of Psr‐enzymes (which physiologically convert S^0^ and S_n_) to convert S_2_O_3_ ^2−^, the physiological substrate of the thiosulphate reductase Phs (Hinsley and Berks 2002), we conducted a detailed analysis of Psr‐homologue protein sequences. The Psr is a trimeric enzyme belonging to the DMSO reductase family (Rothery et al. 2008; Grimaldi et al. 2013). To date, only one representative has been characterised in detail, in Wolinella succinogenes (Krafft et al. 1992; Dietrich and Klimmek 2002). The PsrA subunit carries the Mo‐cofactor and an Fe‐S cluster, the PsrB subunit carries four Fe‐S clusters and the PsrC subunit, which lacks any cofactor, is responsible for anchoring the enzyme to the membrane and, more importantly, for reacting with the quinones, which are its electron donors. The polysulphide reductase Psr, thiosulphate reductase Phs, selenite reductase Srr, and sulphur reductase Sre, share a conserved trimeric structure and highly similar amino acid sequences. Yet the specific sequence differences between the SreA and PsrA/PhsA/SrrA subunits enable their classification into two distinct phylogenetic clades, as discussed in (Sorokin et al. 2016; Wells et al. 2019). However, the homology between the PsrA, the PhsA and the SrrA is such that they cannot be distinguished and are mixed up in a phylogenetic tree reconstructed using the catalytic subunit sequences (Wells et al. 2019). The distinction between psr and srr is based solely on the operon composition: the srr operon carries additional genes (srrE, srrD and srrF; (Wells et al. 2019)). These genes are not present in the ANA‐3 psr gene cluster. The distinction between Psr and Phs is based solely on the nature of its membrane subunit. The PsrC is a subunit with eight transmembrane helices and is devoid of any cofactor (Dietrich and Klimmek 2002; Jormakka et al. 2008), whereas the PhsC is predicted to have five transmembrane helices and two b hemes (Rothery et al. 2008; Stoffels et al. 2012). The predicted topology of the protein WP_011718395.1 (Table 2) identified in ANA‐3 unambiguously shows that it is a PsrC (Figure SI‐2). Therefore, we conclude that the enzyme identified in S. oneidensis MR‐1 and ANA‐3 is most likely a Psr; however, this requires experimental validation by testing its ability to reduce polysulphide.
The Psr‐Homologue Is Able to Directly Convert the Tetrathionate
3.7
The results presented above suggested that the role of the Psr‐homologue in reduction of S_4_O_6_ ^2−^ was indirect, involving the shifting of the chemical equilibrium between S_4_O_6_ ^2−^ and S_2_O_3_ ^2−^ by depleting the medium in S_2_O_3_ ^2−^. To test this hypothesis, we monitored the S_4_O_6_ ^2−^ and S_2_O_3_ ^2−^ reductive activities of cellular fractions (membranous and soluble ones) from both the Δttr and Δpsr strains, which were cultivated in the presence of S_2_O_3_ ^2−^ and S_4_O_6_ ^2−^ to induce the production of Psr and Ttr, respectively. After staining the gels with reduced (by titanium citrate) methyl viologen as electron donor, and then revealing the enzymatic activities using S_2_O_3_ ^2−^ or S_4_O_6_ ^2−^ as the electron acceptor (Figure 5), we observed that no enzymatic activity was detected in the presence of S_2_O_3_ ^2−^ in the Δpsr mutant samples, even when grown in the presence of S_4_O_6_ ^2−^ to induce Ttr production. This indicates that the Ttr‐homologue was unable to reduce S_2_O_3_ ^2−^ in ANA‐3. This finding supports the results obtained in S. enterica (Hinsley and Berks 2002) and our results from the mutant phenotype analysis. In the gels revealed using S_4_O_6_ ^2−^ as the electron acceptor, de‐stained bands appeared in both the membrane and soluble fractions of the Δpsr strain grown under S_4_O_6_ ^2−^. Proteomic analysis (Table SI‐2) indeed identified TtrA in the destained band, thereby confirming the analysis of mutant phenotype analysis conducted using S_4_O_6_ ^2−^. The presence of the Ttr in the soluble fraction may be surprising, but membrane‐bound members of this Mo‐bisPGD enzyme superfamily have been shown to detach from the membranes (see (Haja et al. 2020)). Additionally, destained bands also appeared in the Δttr strain grown under S_2_O_3_ ^2−^ to induce the expression of the psr genes. Proteomic analysis of the band specifically revealed by S_4_O_6_ ^2−^ in both the membrane and soluble samples identified both PsrA and PsrB, with PsrA being the most abundant protein in the band (Table SI‐3). This result suggests that the Psr‐homologue, induced under S_2_O_3_ ^2−^ growth conditions, is responsible for the reduction of S_4_O_6_ ^2−^ in the gel. This finding supports the hypothesis that Psr directly reduces tetrathionate in vivo and suggests that Psr homologues play a broader role in sulphur metabolism than previously recognised, capable of reducing not only S_n_ ^2−^, S_2_O_3_ ^2−^ and S (Krafft et al. 1992; Burns and DiChristina 2009) but also S_4_O_6_ ^2−^.
Enzymatic activity of cellular fractions from Δttr and Δpsr mutants of ANA‐3. Membranous fraction (M) or Soluble fraction (S) running in a native gel. Electron donor: Methyl viologen reduced by titanium citrate. Electron acceptor: Thiosulphate (S2O3 2−) or Tetrathionate (S4O6 2−). TtrA is identified in the band labelled Ttr while PsrA and PsrB are identified in the band labelled Psr (see Tables SI‐2 and 3).
Discussion
4
On the Role of OTR
4.1
Since the initial proposal by Mowat et al. (2004) that the OTR from S. oneidensis MR‐1 is a tetrathionate reductase, no study has provided experimental confirmation of this physiological function, either in vivo or in vitro. The OTR is a soluble octahaem cytochrome encoded by the otr gene, and homologues of this gene have been identified in environmental bacteria, predominantly gamma‐proteobacteria (Buckley et al. 2019). Contradictory results from transcriptomic and proteomic experiments have suggested that S_2_O_3_ ^2−^, S_4_O_6_ ^2−^ or S could be the physiological substrate of OTR, respectively (Wu 2010; Buckley et al. 2019; Flieder et al. 2021). However, in vitro activities of OTR enzymes revealed nitrite and hydroxylamine reductase activities to be the most prevalent, with very low S_4_O_6_ ^2−^ reductase or S_2_O_3_ ^2−^ oxidase activities (Atkinson et al. 2007; Buckley et al. 2019). This is consistent with the structural similarity between OTR (Mowat et al. 2004) and nitrite reductases such as ccNIR/NrfA (Einsle et al. 1999, 2000; Bamford et al. 2002; Cunha et al. 2003; Rodrigues et al. 2006; Ali et al. 2019; Campeciño et al. 2020; Denkhaus et al. 2023) as well as Hydroxylamine Oxidoreductase HAO (Igarashi et al. 1997). The results obtained with the Δttr Δotr mutant of ANA‐3 confirm the absence of the Δotr deletion mutant phenotype in S. oneidensis MR‐1 on S_4_O_6_ ^2−^ (Wu 2010). However, the results show a slight increase in the accumulation of SO_3_ ^2−^, in the absence of the otr gene (see Figure 3). This observation could be interpreted as evidence for a SO₃^2−^ reduction function for OTR, consistent with its close structural homology to MccA/SirA. The eight haems of these enzymes are superimposable (Hermann et al. 2015). However, the presence of a copper atom in the structure of MccA/SirA appears to distinguish it from OTR. OTR/MccA/SirA are distinct from the extensively studied dissimilatory sulphite reductases found in sulfate‐reducing prokaryotes such as Desulfovibrio vulgaris and Archeoglobus fulgibus which contain sirohaem (see (Fritz et al. 2005) for review). The reaction performed by MccA/SirA involves the direct reduction of SO_3_ ^2−^ to HS^−^ through a six‐electron process. This function would not only explain the higher accumulation of SO_3_ ^2−^, but also the higher respiration of S_2_O_3_ ^2−^ in the absence of OTR in the Δttr Δotr mutant. Without OTR, SO_3_ ^2−^ would no longer be consumed and would therefore accumulate. Regarding S_2_O_3_ ^2−^ respiration by Psr, it has long been known to be inhibited by H_2_S/HS^−^/S^2−^ accumulation (Macy et al. 1986). The absence of OTR activity in the Δttr Δotr mutant would reduce the HS^−^/S^2−^ generation and restrict it to that of Psr. This would partially remove the HS^−^/S^2−^ inhibition of S_2_O_3_ ^2−^ respiration by Psr, resulting in accelerated S_2_O_3_ ^2−^ consumption. Studies of the interactions between OTR and a series of periplasmic proteins in S. oneidensis suggest that OTR could ultimately receive the electrons from the CymA membrane protein (Alves et al. 2015). However, this hypothesis should be tempered by the observation that the Psr turnover is highly sensitive towards the H_2_S/HS^−^/S^2−^ accumulation. The hypothesis regarding the physiological function of OTR, proposed in Figure 6, still needs to be confirmed through dedicated work. This work should include precise monitoring of H_2_S/HS^−^/S^2−^ accumulation, and clarify the respective roles of MccA/SirA and OTR in bacteria where there are both present, as is the case for ANA‐3.
Anaerobic tetrathionate metabolism in ANA‐3, reconstructed from genomic, deletion mutant phenotypes and biochemical analyses. The Psr‐homologue would be responsible for both thiosulphate (S2O3 2−) and tetrathionate (S4O6 2−) respiration, while the Ttr‐homologue would be responsible for S4O6 2− and S3O6 2− respiration. The OTR seems to be responsible for sulphite (SO3 2−) reduction as it is known for the Sir system. CymA has been proposed as electron donor to OTR elsewhere. The Tsd system appears not to be involved in anaerobic S4O6 2− respiration but rather in aerobic S2O3 2− oxidation. Arrows in grey indicate reactions not tested in the present study but proposed in the literature (see the main text). It is to note that despite the chemical equilibrium of HSO3 2− with SO3 2− and H2S with HS−, at pH 7.3, the pH value of ANA‐3 growth, only one form of the respective chemical species is represented in the scheme (scheme created with BioRender.com).
On the Diversity of Ttr‐Type Enzymes, Their Function and Their Molecular Differences
4.2
The Ttr characterised in ANA‐3 represents only the second tetrathionate reductase to be experimentally confirmed, following the enzyme from S. enterica, as represented in the model (Figure 6). Notably, Ttr‐type enzymes have also been proposed to be involved in other metabolic processes, including As(V) respiration and S_2_O_3_ ^2−^ disproportionation. As discussed above, phylogenetic analysis suggests that the Ttr from Desulfolithobacter dissulfuricans GF1T, previously proposed to participate in S_2_O_3_ ^2−^ disproportionation (Hashimoto et al. 2022), clusters with authentic Ttr. In the proposed S_2_O_3_ ^2−^ disproportionation metabolism, S_2_O_3_ ^2−^undergoes both oxidation and reduction, suggesting that Ttr could be involved in either reaction. However, our results and those of Hensel et al. (1999) show that Ttr is unable to reduce S_2_O_3_ ^2−^. Furthermore, the thermodynamics of the oxidation of S_2_O_3_ ^2−^ to S_4_O_6_ ^2−^ (E^0′^ = +198 mV (Kurth et al. 2015)) are difficult to reconcile with the use of quinone (which is the co‐substrate of Ttr (Hensel et al. 1999)), even a UQ (E^0′^ = +100 mV). Therefore, the disproportionation metabolism of S_2_O_3_ ^2−^ by a Ttr is not straightforward to explain. Furthermore, phylogenetic analysis (Figure 2, Panel A, red arrows) shows that many of the bacteria demonstrated to carry out S_2_O_3_ ^2−^ disproportionation do not actually carry true Ttr. The role of Ttr‐like enzymes in S_2_O_3_ ^2−^ disproportionation therefore requires further investigation.
Despite their homology, Ttr‐like enzymes involved in sulphur or arsenic metabolism exhibit remarkable substrate specificity. True Ttr from ANA‐3 is not able to reduce As(V), while Ttr‐like arsenate reductase cannot reduce S_4_O_6_ ^2−^ (Saltikov and Newman 2003; Muramatsu et al. 2020). This raises a critical question: what features distinguish a real Ttr from a Ttr‐like arsenate reductase? To address this, we hypothesised that specific amino acid residues within the substrate binding site distinguish them. We conducted an in‐depth comparative analysis of the substrate binding sites of the bona fide Ttr sequences, Ttr‐like arsenate reductases, and three phylogenetically related arsenic‐metabolising enzymes: bona fide arsenate reductases Arr, as well as the alternative arsenite oxidases Arx (Saltikov and Newman 2003; Zargar et al. 2010; Muramatsu et al. 2020). Our analysis revealed that the substrate binding sites of bona fide TtrA enzymes exhibit a distinct pattern of amino acid residues, clearly differentiating them from the three As‐metabolising enzymes. Two strictly conserved residues, Gly‐Arg, are present in all As‐metabolising enzymes, but are absent in Ttr (Figure 2, Panel B). In Arr, the Arg (R165 in ANA‐3 numbering) has been demonstrated to participate in As(V) binding, alongside other key residues highlighted in cyan (Figure 2B) (Glasser et al. 2018). To determine whether these residues are critical for As binding and functionally distinguish bona fide Ttr from As‐metabolising enzymes, site‐directed mutagenesis is needed.
On the Capacity of the Psr‐Homologue to Reduce Tetrathionate
4.3
While the Δpsr single mutant shows no phenotype towards S_4_O_6_ ^2−^ reduction, combined in gel activity assays and double deletion mutant metabolism analysis clearly suggest that the Psr‐homologue is involved in S_4_O_6_ ^2−^ reduction (as proposed in Figure 6). psrA gene deletion experiments previously performed on S. oneidensis MR‐1 were interpreted in a contradictory way (Burns and DiChristina 2009). However, the examination of the published results (see Figure 5 of the publication) is in agreement with ours. The S. oneidensis MR‐1 psrA deletion mutant (equivalent to the ANA‐3 ΔttrΔpsr mutant since S. oneidensis doesn't carry ttr genes) no longer respires S_4_O_6_ ^2−^. Growth of the deletion mutant is correspondingly affected. The very slow remaining conversion of S_4_O_6_ ^2−^ to S_2_O_3_ ^2−^ could easily be explained by chemical decomposition, followed by bacterial growth on S_2_O_3_ ^2−^.
Experiments performed on S. enterica, which carries both Ttr and Phs, suggested that Phs is unable to reduce S_4_O_6_ ^2−^ (Hensel et al. 1999; Hinsley and Berks 2002). This suggests that, despite their structural similarity, the PsrA and PhsA differ in their catalytic capacities: the Psr can react with S_n_ ^2−^, S_4_O_6_ ^2−^, S_2_O_3_ ^2−^and S^0^, whereas Phs has only been demonstrated to react with S_2_O_3_ ^2−^ and S^0^. The thermodynamics of the reaction does not easily explain the difference in the reactivity of Psr and Phs towards S_4_O_6_ ^2−^. The redox potential of the S_4_O_6_ ^2−^/S_2_O_3_ ^2−^ couple is E^0′^ = +198 mV (Kurth et al. 2015), which is well above the values of the S_2_O_3_ ^2−^/HS^−^ + SO_3_ ^2−^ (E^0′^ = −402 mV) and S^0^/HS^−^ (E^0′^ = −275 mV) couples. Therefore, the in vitro reaction using methyl viologen (E^0′^ = −450 mV) as electron donor and S_4_O_6_ ^2−^ as electron acceptor is much more thermodynamically favourable than the reaction involving the natural substrates for both Psr and Phs, that is, menaquinones (E^0′^ = −125/−70 mV) as electron donor (Stoffels et al. 2012; Hein et al. 2017) and S^0^ or S_2_O_3_ ^2−^ as electron acceptor, respectively.
To rationalise the distinct substrate reactivity, the most plausible hypotheses involve either differences in the substrate coordination network within the active site or variations in substrate accessibility to the active site. Unfortunately, structural insights remain limited: while the structure of Phs has yet to be resolved, the structure of Psr has been determined only in the absence of its substrate (Jormakka et al. 2008).
On the Metabolism of Tetrathionate in Environment
4.4
Genes that encode homologues of Ttr have been identified in environmental prokaryotes (Radl et al. 2019; Vavourakis et al. 2019; Boughanemi et al. 2020; Calisto and Pereira 2021; Flieder et al. 2021; Li et al. 2023; He et al. 2025). Previous publications have also reported the expression of ttr‐homologues in environmental communities to be correlated with reduction of S_4_O_6_ ^2−^ to S_2_O_3_ ^2−^ (Vavourakis et al. 2019). However, the present work is the first study to unambiguously demonstrate that a Ttr‐homologue is involved in respiration of S_4_O_6_ ^2−^ in an environmental prokaryote. Furthermore, our work demonstrates that another enzyme, a Psr‐homologue, is also partly responsible for this respiration.
The observation of bacterial S_4_O_6_ ^2−^ respiration in the environment has traditionally been attributed to its transient availability, arising from microbial transformations of sulphur compounds, particularly oxidation of S_2_O_3_ ^2−^, under syntrophic life conditions (Barrett and Clark 1987; Mandal et al. 2020; He et al. 2025). However, the present results challenge this view by demonstrating the relative stability of S_4_O_6_ ^2−^ over time: at neutral pH, only 55% of S_4_O_6_ ^2−^ was abiotically reduced to S_2_O_3_ ^2−^ after 60 h. This stability suggests that S_4_O_6_ ^2−^ may serve as an energetically favourable electron acceptor in anaerobic environments. At a pH of 7, the redox couple S_4_O_6_ ^2−^/S_2_O_3_ ^2−^ exhibits a redox potential of +198 mV (Kurth et al. 2015), the highest among sulphur compounds (see (D'Ermo et al. 2024) for a review). Consequently, the Gibbs free energy available for microbial growth through the oxidation of organic compounds coupled to the reduction of S_4_O_6_ ^2−^ (a) exceeds that of any other sulphur‐based respiration (see (b) for the case of polysulphide). In summary, our results suggest that S_4_O_6_ ^2−^ respiration, by Ttr‐ and by Psr‐homologues, may be far more environmentally widespread than previously acknowledged.
Author Contributions
Gwendoline Degré: writing – original draft, investigation, writing – review and editing. Audrey Tempier: investigation, writing – review and editing. Marie Vaillant: investigation, writing – review and editing. Cécile Jourlin‐Castelli: writing – review and editing, methodology. Simon Duval: writing – review and editing, investigation. Régine Lebrun: investigation, writing‐review and editing. Barbara Schoepp‐Cothenet: conceptualization, funding acquisition, writing – original draft, methodology, writing – review and editing, project administration, supervision, data curation, validation.
Funding
This work was supported by Agence Nationale de la Recherche, ANR‐23‐CE44‐0012‐01.
Conflicts of Interest
The authors declare no conflicts of interest.
Supporting information
Data S1: Supplementary Information.
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1Ali, M. , N. Stein , Y. Mao , et al. 2019. “Trapping of a Putative Intermediate in the Cytochrome c Nitrite Reductase (cc Ni R)‐Catalyzed Reduction of Nitrite: Implications for the cc Ni R Reaction Mechanism.” Journal of the American Chemical Society 141: 13358–13371.31381304 10.1021/jacs.9b 03036 · doi ↗ · pubmed ↗
- 2Alves, M. N. , S. E. Neto , A. S. Alves , et al. 2015. “Characterization of the Periplasmic Redox Network That Sustains the Versatile Anaerobic Metabolism of Shewanella oneidensis MR‐1.” Frontiers in Microbiology 6: 665.26175726 10.3389/fmicb.2015.00665 PMC 4484225 · doi ↗ · pubmed ↗
- 3Appleyard, R. K. 1954. “Segregation of New Lysogenic Types During Growth of a Doubly Lysogenic Strain Derived From Escherichia Coli K 12.” Genetics 39: 440–452.17247495 10.1093/genetics/39.4.440PMC 1209664 · doi ↗ · pubmed ↗
- 4Atkinson, S. J. , C. G. Mowat , G. A. Reid , and S. K. Chapman . 2007. “An Octaheme c‐Type Cytochrome From Shewanella oneidensis Can Reduce Nitrite and Hydroxylamine.” FEBS Letters 581: 3805–3808.17659281 10.1016/j.febslet.2007.07.005 · doi ↗ · pubmed ↗
- 5Bamford, V. A. , H. C. Angove , H. E. Seward , et al. 2002. “Structure and Spectroscopy of the Periplasmic Cytochrome c Nitrite Reductase From Escherichia coli .” Biochemistry 41: 2921–2931.11863430 10.1021/bi 015765 d · doi ↗ · pubmed ↗
- 6Baraquet, C. , L. Théraulaz , C. Iobbi‐Nivol , V. Méjean , and C. Jourlin‐Castelli . 2009. “Unexpected Chemoreceptors Mediate Energy Taxis Towards Electron Acceptors in Shewanella oneidensis .” Molecular Microbiology 73: 278–290.19555457 10.1111/j.1365-2958.2009.06770.x · doi ↗ · pubmed ↗
- 7Barrett, E. L. , and M. A. Clark . 1987. “Tetrathionate Reduction and Production of Hydrogen Sulfide From Thiosulfate.” Microbiological Reviews 51: 192–205.3299028 10.1128/mr.51.2.192-205.1987 PMC 373103 · doi ↗ · pubmed ↗
- 8Boughanemi, S. , P. Infossi , M. T. Giudici‐Orticoni , B. Schoepp‐Cothenet , and M. Guiral . 2020. “Sulfite Oxidation by the Quinone‐Reducing Molybdenum Sulfite Dehydrogenase Soe ABC From the Bacterium Aquifex aeolicus .” Biochimica et Biophysica Acta, Bioenergetics 1861: 148279.32735861 10.1016/j.bbabio.2020.148279 · doi ↗ · pubmed ↗