Structural remodelling of the 2OG oxygenase Rv3406 enables sulfur-scavenging in Mycobacterium tuberculosis

Abstract

Rv3406 evolved from the ubiquitous taurine-catabolising enzyme TauD and functions as a sulfur-scavenging protein in Mycobacterium tuberculosis. Structural and biochemical analyses reveal specific changes that shape its chemical environment for ligand interaction and explain its broad substrate range. These findings show how amino acid substitutions redefine protein function and drive adaptation to the unique metabolic context of Mycobacteria.

---

Mycobacterium tuberculosis (Mtb), the pathogen that causes tuberculosis, assimilates nutrients efficiently from many different sources.^1,2 This survival mechanism allows the bacterium to persist in the nutrient-deprived environment of host macrophages and persist long-term within the host.^1,2 Sulfur is one of the essential nutrients that must be scavenged by Mtb.^3–5 It is incorporated into amino acids, cofactors, and other essential biomolecules, sustaining core biochemical function, structural integrity, and energy production.^3–5 Recent studies have shown that Mtb not only acquires sulfur from inorganic sulfate,^3–5 the preferred sulfur source of many bacteria, but also exploits other host-derived sulfur sources such as alkyl sulfate esters.⁶

In Mtb, Rv3406 functions as an alkyl sulfatase that hydroxylates alkyl sulfate esters, liberating inorganic sulfate for metabolic use (Fig. 1a).⁶ Rv3406 is the only alkyl sulfatase characterised in Mtb and is a member of the Fe(II)- and 2-oxoglutarate (2OG)-dependent oxygenase superfamily.⁷ Sequence alignment shows homology to other bacterial sulfur-scavenging oxygenases, the closest matches being to AtsK, an alkyl sulfatase from Pseudomonas putida (54% sequence identity),⁸ and TauD, a taurine dioxygenase from Escherichia coli (36% sequence identity).⁹ These homologues are functionally equivalent to Rv3406, and also allow P. putida and E. coli to scavenge sulfur from alternative, non-inorganic sulfate, sources.^8,9

Fig. 1 Rv3406 is an Fe(II) and 2OG-dependent oxygenase in Mtb. (a) Rv3406 catalyses the hydroxylation of alkyl sulfate esters to release inorganic sulfate and an aldehyde. (b) NMR monitoring of Rv3406 activity in real time: signals for the substrates, 2OG and (R)-2EHS, and products, succinate and 2-ethylhexanal, can be observed directly. Spectra are at 8 min intervals, and the first spectrum was recorded 4 min after sample mixing. (c) ¹H NMR time course allows the concurrent measurement of 2OG oxidation and substrate hydroxylation. The reaction mixture contained 2 μM Rv3406, 50 μM Fe(II), 1 mM ascorbate, 500 μM 2OG, 500 μM (R)-2EHS, 50 mM Tris-D11 (pH 8.0) in 10% D₂O/90% H₂O. We were unable to accurately integrate the resonances for 2-ethylhexanal due to signal overlap.

Despite their homology and related biological functions, Rv3406, AtsK, and TauD, have different substrate specificity. While TauD accepts taurine, a small, non-proteinogenic amino acid,⁹ Rv3406 and AtsK prefer alkyl sulfates that contain large, hydrophobic sidechains, e.g., (R)-2-ethylhexyl sulfate ((R)-2EHS).^6,8 The substrate preferences of AtsK and TauD likely reflect the sulfur sources available to their respective hosts.^6,8,9 Understanding the structural determinants of substrate specificity in these closely related enzymes is key to clarifying the role of Rv3406 in Mtb sulfur metabolism and the nutrient conditions within macrophages. Here, we provide evidence that structural divergence of Rv3406 from its homologue TauD underlies its broader substrate specificity and hints at its refined role as a sulfur-scavenging enzyme in Mtb.

Previous biochemical studies of Rv3406 suggested broad substrate specificity but were limited to testing only medium-chain alkyl sulfates.⁶ In the standard 2OG oxygenase mechanism, oxidative decarboxylation of 2OG produces a reactive Fe(IV)=O intermediate that oxidizes the substrate.^10,11 The coupling ratio and relative rates of 2OG and substrate oxidation are key indicators of substrate preference, as a weak substrate may not bind close enough to the metal centre or remain bound long enough for oxidation. To further explore the substrate preference of Rv3406, we produced recombinant Rv3406 and tested a wider range of alkyl sulfates (methyl, hexyl, octyl and dodecyl sulfate) (Table 1 and Fig. S1, SI). The reaction was characterised using a nuclear magnetic resonance (NMR)-based assay, which allows concurrent monitoring of substrate consumption and product formation (Fig. 1b and Fig. S2–S5, SI). This method is superior to other spectrophotometry-based assays as it allows the coupling ratio between substrate and co-substrate oxidation to be measured.

Table 1 Rv3406 has a broad substrate specificity. Rv3406 substrate analogues that were screened for activity, coupling ratio and relative rate (based on 2OG turnover) compared to the previously identified substrate (R)-2EHS. The reaction mixture contained 2 μM Rv3406, 50 μM Fe(II), 1 mM ascorbate, 500 μM 2OG, 500 μM potential substrate, 50 mM Tris-D11 (pH 8.0) in 10% D₂O/90% H₂O. For structures of the compounds tested, see Fig. S1, SI. ‘No activity’ indicates no evidence of substrate consumption. 2OG-to-succinate turnover was monitored by tracking the 2OG resonance at 2.35 ppm and the succinate resonance at 2.1 ppm. Substrate turnover was monitored by the decrease in NMR signal(s) of the substrate that do not overlap with other signal resonances in the spectra

Substrate	Activity	2OG-to-substrate coupling	Relative rate to (R)-2EHS (%)
(R)-2EHS	Yes	1 : 1	100
Methyl sulfate	No	—	—
Hexyl sulfate	Yes	90 : 1	0.3
Octyl sulfate	Yes	3.5 : 1	15.0
Dodecyl sulfate	Yes	1 : 1	31.7
Methanesulfonate	No	—	—
Propane-1-sulfonate	No	—	—
Heptane-1-sulfonate	Yes	8 : 1	3.8
Octane-1-sulfonate	Yes	5 : 1	4.5
Dodecane-1-sulfonate	Yes	1 : 1	39.5
4-Dodecylbenzenesulfonate	No	—	—
Taurine	No	—	—
Isethionate	No	—	—
2-Mercaptoethane-1-sulfonate	No	—	—

---

Mtb Rv3406 did not catalyse the turnover of methyl sulfate but displayed activity towards longer alkyl sulfate analogues. Among these, the highest activity was observed with (R)-2EHS (Table 1), for which 2OG oxidation was tightly coupled to hydroxylation of the substrate (Fig. 1c). Interestingly, although the alkyl moiety in hexyl sulfate has the same length as the principal chain in (R)-2EHS, without the additional ethyl branch, the relative turnover rate of hexyl sulfate was dramatically decreased, from 100% to ∼0.3% (based on 2OG turnover) (Table 1). The 2OG : substrate coupling ratio switched from 1 : 1 to 90 : 1 (Table 1), indicating that, for every 90 rounds of 2OG oxidation, only a single molecule of hexyl sulfate is hydroxylated. The trend in the sulfate series of substrates is for the relative turnover rate to increase and the 2OG : substrate coupling ratio to decrease, as the alkyl chain length increases (Table 1). Alkyl chain length and branching are clearly critical factors governing substrate turnover in Rv3406. Our results are also consistent with those reported by Sogi et al., who found that Rv3406 could catalyse the turnover of (R)-2EHS, hexyl sulfate, heptyl sulfate, and, to a smaller extent, pentyl sulfate.⁶

Dodecyl sulfate, the longest analogue tested, showed a coupling ratio of 1 : 1, the same as for (R)-2EHS, but the relative rate does not approach the 100% control, instead only reaching 31.7%. Interestingly, no aldehyde peak was detected in the reaction (Fig. S7, SI). This indicates that hydroxylation occurs at a different position, possibly away from the carbon directly bonded to the sulfate group. We propose that the longer alkyl chain of dodecyl sulfate forces the molecule to bend within the enzyme's binding pocket and shifts the site of hydroxylation. Frustratingly, we were also unable to observe a clear set of signals that correspond to the hydroxylated product in the NMR spectra. This could be due to the limited solubility and low yield of the product. Nonetheless, our findings clearly show that hydroxylation does not necessarily happen immediately adjacent to the ester oxygen on the substrate.

We then evaluated alkyl sulfonates which lack the bridging oxygen atom present in alkyl sulfates. Sulfonate substrates followed a similar trend to their sulfate analogues (Table 1). Sulfonates with less than seven carbons in their alkyl chain remain unmodified (Table 1). Heptane-1-sulfonate has slow turnover and an 8 : 1 coupling ratio; maximal turnover of the twelve carbon molecule affords only 39.5% conversion but a 1 : 1 coupling ratio (Table 1). The 4-dodecylbenzenesulfonate substrate shows that size is not the only determining factor. This largest sulfonate analogue, is not turned over and suggests the enzyme only recognises linear or simple branched chains, and not aromatic groups as substrate (we note that this chain is equivalent in length to a 15 carbon alkyl chain).

To define the structural determinants governing substrate binding to Rv3406 and turnover, we carried out extensive crystallography experiments. While crystals complexed with (R)-2EHS could not be obtained, we succeeded in crystallising Rv3406 using Ni(II) as a substitute for Fe(II) together with N-oxalylglycine (NOG) as an unreactive analogue/substitute of 2OG (Table S1 and S2, SI). The enzyme structure (Fig. 2a) displays the canonical structure displayed by earlier apo and metal bound structures (PDB: 4FFA and 4CVY), including tetrameric assembly and the conserved metal binding active site architecture (Fig. S8, SI). The Ni(II) is coordinated by the triad HXD/E…H motif (His97, Asp99, and His251) (Fig. 2b), conserved in all 2OG-dependent oxygenases,¹⁰ while NOG coordinates the Ni(II) in a bidentate manner (Fig. 2b) with the final metal coordination site opposite His251, occupied by a water molecule as a proxy for the substrate hydroxylation site (Fig. 2b). Comparison with the crystal structures of E. coli TauD (PDB: 1GY9)¹² and P. putida AtsK (PDB: 1OIK),¹³ both protein–Fe(II)–2OG–substrate complexes, shows that the Ni(II) in our structure occupies the same position and has similar ligands as the Fe(II) in TauD and AtsK. NOG binds in a manner analogous to 2OG in these enzymes (Fig. S7 and S8, SI).

Fig. 2 Structure of Rv3406 with co-substrate mimic NOG. (a) Rv3406-Ni(II)-NOG structure (PDB: 8EVN). The molecular structure of NOG is depicted in the inset. The overall backbone and metal binding residues of our Rv3406-Ni(II)-NOG complex are the same as those of the published Rv3406-Fe(II) structure.⁶ (b) NOG binds to the active site metal ion in a bidentate manner. The final coordination site on Ni(II) opposite His251, is occupied by a water molecule as a proxy for the hydroxylation site for substrate oxidation. (c) The active site of TauD-Fe(II)-2OG with taurine binding (PDB: 1GY9) defines Asn95_TauD and Tyr73_TauD as critical residues stabilising the ligand. These two amino acids are substituted by alanine residues in Rv3406 and explain the inability of RV3406 to turnover taurine as substrate. Docking of (R)-2EHS to Rv3406-Ni(II)-NOG shows a binding model similar to that seen in the AtsK-Fe(II)-2OG-(R)-2EHS structure (PDB: 1OIK).

Using our Rv3406–Ni(II)–NOG structure, we docked (R)-2EHS into the active site and found it positioned in the same location and orientation as in the AtsK–Fe(II)–2OG–(R)-2EHS structure (PDB: 1OIK). Our model places the sulfate of (R)-2EHS close to the water-coordination site of Ni(II) and held in place by two hydrogen bond interactions between sulfate oxygens and Val100 and Arg266 (Fig. 2c). The non-polar alkyl chain putatively extends into a non-polar pocket of the enzyme although a loop in this section of the X-ray crystal structure (residues 82–102) is disordered and not modelled. The location of disordered loops is similar to that observed in the AtsK-Fe(II)-2OG-(R)-2EHS structure (PDB: 1OIK) (S9, SI).¹³ In TauD (PDB: 1GY9), the equivalent sequence is an ordered β-strand structure (Fig. S10, SI).¹² The flexibility within the disordered loop(s) may allow Rv3406 to accommodate alkyl chains of different length in the active site.

The amine group of taurine in the TauD-Fe(II)-2OG-taurine structure (PDB: 1GY9),¹² hydrogen bonds with Tyr73_TauD and Asn95_TauD (Fig. 2c), residues that are substituted with alanine in Rv3406 (Fig S11, SI). These substitutions likely increase the size and hydrophobicity of the substrate binding pocket and could explain the observed preference for alkyl sulfates over taurine by Rv3406. To further understand the roles of Tyr73_TauD and Asn95_TauD in governing the substrate preference between TauD and Rv3406, we produced TauD Y73A, TauD N95A, and TauD Y73A/N95A mutants and compared them with WT TauD and Rv3406.

All three TauD mutants were highly active and were able to catalyse uncoupled 2OG turnover. Y73A/N95A TauD shows ∼2.5 times more 2OG turnover compared to the other mutants and both WT enzymes (Fig. 3a). The reason for this is not known. A slower turnover rate for uncoupled 2OG turnover may be due to 2OG binding in multiple orientations at the active-site metal, as there is no substrate to anchor 2OG in the “productive” orientation. A reduction in polar amino acids at the active site could prevent binding in non-productive orientations, since fewer hydrogen bond donors or acceptors would be available to stabilise 2OG in these positions. We then tested the mutants and WT enzymes with taurine as substrate (Fig. 3b). As expected, WT TauD turns over the most taurine, with equimolar amounts of succinate and aminoacetaldehyde produced at the reaction end point. WT Rv3406 is unable to utilise taurine as a substrate (Fig. 3b). All three TauD mutants accept taurine as a substrate; N95A TauD shows the highest activity, with an almost 1 : 1 2OG : substrate coupling ratio, whereas Y73A and Y73A/N95A TauD exhibit reduced activity compared to N95A (Fig. 3b). These results suggest that the tyrosine residue is important for taurine catalysis and likely acts synergistically with asparagine, as indicated by the cumulative loss of activity in the Y73A/N95A double mutant. In contrast, it is striking that with a more hydrophobic and spacious environment, TauD mutants were able to utilise (R)-2EHS as a substrate (Fig. 3c). This demonstrates the influence of the alanine substitutions in diversifying the sulfatase behaviour of the Rv3406 protein from that of its predecessor in taurine catabolism.

Fig. 3 Mutagenesis reveals residues that govern substrate specificity. (a) Uncoupled turnover of 2OG for WT and mutant proteins expressed as a percentage of starting substrate concentration. The reaction mixture contained 2 μM enzyme, 50 μM Fe(II), 1 mM ascorbate, 500 μM 2OG, 5 μM catalase, 50 mM Tris-D11 (pH 8.0) in 10% D₂O : 90% H₂O. Reaction time was 1 hour. (b) Catalytic activity of TauD WT, Rv3406 WT and mutant TauD towards taurine. Coupling ratios are shown above the product formation plots. The reaction conditions are the same as above in the 2OG only reaction in (a), but with the addition of 500 μM taurine. (c) Catalytic activity of TauD WT, Rv3406 WT and mutant TauD protein towards (R)-2EHS. Coupling ratios are shown above the product formation plots. The reaction conditions are the same as above for 2OG but with the addition of 500 μM (R)-2EHS.

We examined Rv3406's substrate preference and expanded our experimental range to include medium- and long-chain alkyl sulfates and sulfonates. Since many microorganisms use these compounds as alternative sulfur sources,¹⁴ our findings suggest that Mtb may similarly exploit them through the extended substrate profile of Rv3406. Through cross-species mutagenesis, we probed the structural basis for this promiscuous substrate preference and identified specific amino acid residues driving sulfatase binding and catalysis in Rv3406. Sequence similarity and structural analysis suggest that TauD is ancestral to Rv3406 (Fig. S12, SI). Comparing the two is intriguing because, while they perform similar biological roles – utilising alternative sulfur sources – they rely on entirely different organic substrates. The intracellular niche of Mtb within macrophages contrasts sharply with the commensal existence of E. coli in the gastrointestinal tract, exposing each organism to distinct nutritional landscapes, particularly in relation to sulfur availability. These environmental differences likely drive selection for enzymes capable of exploiting diverse substrates. Although the availability of different sulfate and sulfonate substrates inside host macrophage is unclear, it is possible that the broad substrate specificity of Rv3406 may represent an adaptive strategy to access varied sulfur sources within the macrophage. This warrants further investigation as little is known about the abundance, composition, and physiological relevance of these compounds. In addition to its role in sulfur scavenging, Rv3406 has been implicated in the oxidation and inactivation of several antitubercular agents.^15–17 This work therefore provides valuable structure–activity relationship insights to guide drug development, either by avoiding susceptible functional groups or by designing inhibitors as potential adjuvants.

We acknowledge use of the Nuclear Magnetic Resonance Centre and Mass Spectrometry Hub at the University of Auckland, and the Melbourne Magnetic Resonance Platform, Melbourne Protein Characterisation Platform, and Mass Spectrometry and Proteomics Facility at the University of Melbourne for this work. We thank Dr Nicholas Demarais and Martin Middleditch of the University of Auckland for their technical input for mass spectrometry analyses. We thank Jingsong Zhou for his help with NMR experiments. This research was undertaken in part using the MX2 beamline at the Australian Synchrotron, part of ANSTO, and made use of the Australian Cancer Research Foundation (ACRF) detector. EYWH was supported by the Melbourne Research Scholarship, Rowden White Scholarship, Norma Hilda Scholarship, Dame Margaret Blackwood Soroptimist Scholarship, and Dr Albert Shimmins Postgraduate Writing-Up Award. IKHL acknowledge the University of Auckland and the University of Melbourne for financial support of this project.

Conflicts of interest

There are no conflicts to declare.

Data availability

The data supporting this article have been included as part of the supplementary information (SI). Supplementary information is available. See DOI: https://doi.org/10.1039/d5cc05573c.

Crystallographic data for Rv3406-Ni(II)-NOG has been deposited at the Protein Data Bank (PBD) under identification code 8EVN and can be obtained from https://doi.org/10.2210/pdb8EVN/pdb.

Notes and references

1. M. Berney and L. Berney-Meyer, Microbiol. Spectrum, 2017, 5, 1128.
2. S. Ehrt, D. Schnappinger and K. Y. Rhee, Nat. Rev. Microbiol., 2018, 16, 496–507.
3. S. K. Hatzios and C. R. Bertozzi, PLoS Pathog., 2011, 7, e1002036.
4. J. M. Lensmire and N. D. Hammer, Curr. Opin. Microbiol., 2019, 47, 52–58.
5. M. W. Schelle and C. R. Bertozzi, ChemBioChem, 2006, 7, 1465–1626.
6. K. M. Sogi, Z. J. Gartner, M. A. Breidenbach, M. J. Appel, M. W. Schelle and C. R. Bertozzi, PLoS One, 2013, 8, e65080.
7. J. D. Mougous, R. E. Green, S. J. Williams, S. E. Brenner and C. R. Bertozzi, Chem. Biol., 2002, 9, 767–776.
8. A. Kahnert and M. A. Kertesz, J. Biol. Chem., 2000, 275, 31661–31667.
9. E. Eichhorn, J. R. van der Ploeg, M. A. Kertesz and T. Leisinger, J. Biol. Chem., 1997, 272, 23031–23036.
10. C. J. Schofield and Z. Zhang, Curr. Opin. Struct. Biol., 1999, 9, 722–731.
11. S. Martinez and R. P. Hausinger, J. Biol. Chem., 2015, 290, 20702–20711.
12. J. M. Elkins, M. J. Ryle, I. J. Clifton, J. C. Dunning Hotopp, J. S. Lloyd, N. I. Burzlaff, J. E. Baldwin, R. P. Hausinger and P. L. Roach, Biochemistry, 2002, 41, 5185–5192.
13. I. Müller, A. Kahnert, T. Pape, G. M. Sheldrick, W. Meyer-Klaucke, T. Dierks, M. Kertesz and I. Usón, Biochemistry, 2004, 43, 3075–3088.
14. M. A. Kertesz, FEMS Microbiol. Rev., 2000, 24, 135–175.
15. J. Neres, R. C. Hartkoorn, L. R. Chiarelli, R. Gadupudi, M. R. Pasca, G. Mori, A. Venturelli, S. Savina, V. Makarov, G. S. Kolly, E. Molteni, C. Binda, N. Dhar, S. Ferrari, P. Brodin, V. Delorme, V. Landry, A. L. de Jesus Lopes, R. D. Farina, P. Saxena, F. Pojer, A. Carta, R. Luciani, A. Porta, G. Zanoni, E. De Rossi, M. P. Costi, G. Riccardi and S. T. Cole, ACS Chem. Biol., 2015, 10, 705–714.
16. M. R. Bockman, C. A. Engelhart, S. Dawadi, P. Larson, D. Tiwari, D. M. Ferguson, D. Schnappinger and C. C. Aldrich, ACS Infect. Dis., 2018, 4, 1102–1113.
17. D. Recchia, G. Stelitano, A. Egorova, G. B. Biffignandi, K. Savková, R. Kafková, S. Huszár, A. M. Cerrato, R. A. Slayden, J. E. Cummings, N. Whittel, A. A. Bauman, G. T. Robertson, L. Rank, F. Urbina, T. R. Lane, S. Ekins, O. Riabova, E. Kazakova, K. Mikušová, D. Sassera, G. Degiacomi, L. R. Chiarelli, V. Makarov and M. R. Pasca, ACS Infect. Dis., 2025, 11, 986–997.

---

Footnote

† Vicky Tzu-Jung Juan and Patrick Bajan contributed equally to this study.

---