The sulfonamide functional group is widely used in medicinal chemistry and is found in many approved and experimental drugs1. Sulfonamide-containing compounds exhibit diverse biological activities through their interactions with metal ions, proteins, and nucleic acids2. Despite its synthetic and medicinal chemistry importance, the sulfonamide group is a rare functional group in natural products and little is known about its biosynthesis3.
Previously our group reported the biosynthesis of a sulfonamide antibiotic natural product altemicidin4,5. In the biosynthesis of altemicidin we found that one cupin superfamily enzyme SbzM is responsible for the formation of sulfonamide moiety of altemicidin: SbzM catalyzed the conversion of L-cysteine (L-Cys) into 2-sulfamoylacetic aldehyde (1), which is further oxidized to 2-sulfamoylacetic acid by the aldehyde dehydrogenase SbzJ (Figure 1). However, the structural basis and the catalytic mechanism of SbzM remains unknown. In this study, we further investigated SbzM via biochemical, spectroscopic and computational studies.
Figure 1. Sulfonamide moiety formation catalyzed by SbzM.
Since cupin superfamily enzymes require a metal ion (generally iron) for their catalytic activities6, firstly we want to know the metal dependency of SbzM. We used AmSbzM in this study, a SbzM homolog from Actinokineospora mzabensis, sharing 43% sequence similarity with SbzM. By cutting the his tag and treated with EDTA, AmSbzM was incubated with L-Cys and various metal ions. AmSbzM catalyzed the generation of 1 only in the presence of nickel, suggesting that AmSbzM is a nickel-dependent enzyme. By inductively coupled plasma–mass spectrometry (ICP-MS) and isothermal titration calorimetry (ITC) we further confirmed that AmSbzM can only bind to nickel but not any other metal ions including iron. By X-ray absorption spectroscopy (XAS) we found that during the catalytic reaction Ni(II) bound to AmSbzM is oxidized to Ni(III), suggesting a Ni(II)/Ni(III) redox cycle during catalysis. Only two Ni-dependent cupin enzymes, acireductone dioxygenase (ARD) and quercetin 2,4-dioxygenase (QueD) has been reported so far7, and Ni retains its divalent state during their catalysis. Compared with them, AmSbzM is a special Ni-dependent cupin enzyme having Ni(II)/Ni(III) redox cycle during catalysis.
To understand the structural basis of the sulfonamide formation reaction, the O2-bound structure of AmSbzM was solved at 1.94 Å resolution. AmSbzM has a classical cupin-fold domain, and an additional C-terminal domain comprising three α-helices and two β-sheets, which is not observed in the canonical cupin enzymes. We found three metal binding pockets: one 3x His motif in cupin-fold domain and two 1x His+3x Cys zinc-finger like motifs in the C-terminal domain (Figure 2). Mutagenesis of any residue of three metal sites led to the total abolishment of enzyme activity.
Next we docked L-Cys into the crystal structure of AmSbzM and verified the docking model by molecular dynamics (MD) simulation. Based on the docking model, we performed mutagenesis for the possible pocket residues around the 3x His motif. All variants showed largely decreased 1-forming activity, suggesting the importance of the pocket residues for the catalytic reaction. Interestingly among the variants, H107A variant still maintained the ability to consume L-Cys but can hardly generate 1, suggesting that the variant generates other products. By using different derivatization reagents, we successfully identified mercaptoacetaldehyde (MA) and ammonium as the products of H107A variant. Considering MA and ammonium may be the products of an imine hydrolysis reaction, we assumed that a mercaptoimine (4) may be the real product of H107A variant and then 4 was hydrolyzed into MA and ammonium during the detection. We successfully detected 4 in the reaction of H107A variant, suggesting 4 is the product of H107A variant. Interestingly, small amount of 4 was also detected in the reaction of AmSbzM WT, suggesting that 4 is not only a new product from variant but also possibly an intermediate of wild type.
Figure 2. Protein structure of AmSbzM. Ni and Zn ions are shown as green and white spheres, respectively.
Next we want to figure out how oxygen gets involved into the catalytic reaction of AmSbzM. Firstly we found that no activity was observed for the reaction in anaerobic condition, suggesting that AmSbzM is an oxygen dependent enzyme. By measuring the O2 stoichiometry of the enzyme reaction we found that AmSbzM wild type consume two molecular oxygens to generate one molecule of 1, while AmSbzM H107A variant consume one molecular oxygen to generate one molecule of 4. This leads us to a hypothesis that 4 is an intermediate when the enzyme only takes one O2 molecule into the catalytic reaction, and will be consumed when a second O2 molecule goes into the pocket. To check this hypothesis, we performed the AmSbzM wild type reaction under different oxygen concentration. In the oxygen limited condition 4 significantly accumulated compared to the reaction performed under air condition. This elegantly matched our hypothesis that one O2 molecule is required for AmSbzM wild type to catalyze the reaction from L-Cys to 4 and then another O2 molecule is required for the reaction from 4 to 1. What’s more, by using 18O2, we confirmed that two oxygen atoms of sulfonamide moiety are derived from two different O2 molecules, indicating that although two O2 molecules come to the pocket in order, they equally contribute one of their oxygen atoms to construct the sulfonamide group together.
Figure 3. Brief catalytic mechanism of AmSbzM.
Based on all the experimental evidence we have found, we proposed a brief catalytic mechanism of AmSbzM (Figure 3): the first molecular oxygen binds to Ni(II), generating a Ni(III)-superoxo species. This Ni(III)-superoxo species induces a radical decarboxylation of L-Cys, leading to the formation of 4 and Ni(II)-peroxo. In the reaction of AmSbzM wild type under general condition, the second oxygen goes into the pocket, catalyzing the formation of final product 1, while in the oxygen limited condition or in the H107A variant’s reaction, the reaction stop because the second oxygen is hard to bind to the pocket and 4 releases to the solvent. With detailed QM/MM calculations we finally proposed a detailed catalytic mechanism of AmSbzM.
Our work provides mechanistic insights into the biosynthesis of rare sulfonamides moiety in nature and offers a foundation for exploring the largely uncharacterized cupin-type sulfonamide synthase superfamily enzymes. This study opens avenues for developing biocatalytic strategies to synthesize sulfonamide-containing compounds, contributing to advancements in pharmaceutical and chemical applications.
More details of this work can be found here: Structure–function and mechanistic analyses of nickel-dependent sulfonamide synthase | Nature Catalysis
- Das, P., Delost, M. D., Qureshi, M. H., Smith, D. T. & Njardarson, J. T. A Survey of the Structures of US FDA Approved Combination Drugs. J Med Chem 62, 4265-4311(2019).
- Kumar Verma, S. et al. Antibacterial activities of sulfonyl or sulfonamide containing heterocyclic derivatives and its structure-activity relationships (SAR) studies: A critical review. Bioorg Chem 105, 104400 (2020).
- Mujumdar, P. & Poulsen, S.-A. Natural Product Primary Sulfonamides and Primary Sulfamates. J Nat Prod 78, 1470-1477 (2015).
- Hu, Z., Awakawa, T., Ma, Z. & Abe, I. Aminoacyl sulfonamide assembly in SB-203208 biosynthesis. Nat Commun 10, 184 (2019).
- Barra, L. et al. β-NAD as a building block in natural product biosynthesis. Nature 600, 754-758 (2021).
- Joseph, C. A. & Maroney, M. J. Cysteine dioxygenase: structure and mechanism. Chem Commun 3338-3349 (2007).
- Boer, J. L., Mulrooney, S. B. & Hausinger, R. P. Nickel-dependent metalloenzymes. Arch Biochem Biophys 544, 142-152 (2014).