Fig 1.
Scheme for BioH action in mycobacterial biotin biosynthetic pathway.
A. Genetic environment of bioH homologs in different microorganisms. The dots denote the conserved sequence motifs recognized by the regulatory protein (e.g., BirA and/or BioQ, shown in cyan background). The two discontinued loci are separated with the double slash. The bioC is colored green, and the bioH homolog is highlighted in purple. B. Real-time quantitative PCR (qPCR) analysis for expression of three putative BioH paralogs (bioH1 to bioH3) Namely, the three bioH isoforms included bioH1 (MSMEG_2036), bioH2 (MSMEG_1352), and bioH3 (MSMEG_6710). sigA encoding sigma A functions as an interference gene. Ct (cycle threshold) is used to measure the amplification cycles of target genes during the qPCR. C. Use of conformationally-sensitive 0.5M urea/PAGE (17.5%, pH9.5) to separate reactant M-pim-ACP and its hydrolytic product pim-ACP The minus “—” denotes no addition of either BioH or BioJ enzyme. D. The schematic representative of the “BioC-BioH” pathway of biotin synthesis. E. A scheme for multi-target biotinylation by Bpl in M. smegmatis. Unlike that in E. coli AccB (colored gray) is the only biotinylated enzyme, M. smegmatis is proposed to contain no less than five target proteins (colored pink) modified by Bpl. F. Sequence signature of the protein biotinylation by Bpl. G. Bacterial viability-based determination for minimum physiological demand for biotin in M. smegmatis MC2 155. Designations: BioA, 7,8-diaminononanoate synthase (DANS); BioB, biotin synthase; BioF, 8-amino-7-oxononanoate synthase (AONS); BioC, O-methyl transferase; BioD, dethiobiotin synthetase (DTBS); BioH, methyl pimeloyl-ACP ester carboxyl-esterase; BirA, biotin protein ligase/repressor; Bpl, biotin protein lgase; HP, hypothetical protein; ComF, a protein encoded by the late competence operon; BioQ, a TetR-type transcription factor regulating biotin operon; BioJ, an isoenzyme of BioH demethylase. FabH, β-ketoacyl-ACP synthase III; FabG, β-ketoacyl-ACP reductase; FabZ, β-hydroxyacyl-ACP dehydratase; FabI, enoyl-ACP reductase; FAS II, type II fatty acid synthesis pathway. ACP, acyl carrier protein; M-pim-ACP, methyl pimeloyl-ACP; pim-ACP, pimeloyl-ACP; DTB, dethiobiotin.
Fig 2.
Expression cloning discovered three distinct M. smegmatis genes displaying BioH-like activities.
A. M. smegmatis possesses three distinct genes whose expression consistently enables the recipient host of the E. coli ΔbioH biotin auxotrophic strain to appear on the non-permissive condition lacking biotin. A representative result of three independent experiments was given. Namely, the three BioH-like genes (designated BioH1 to BioH3) of M. smegmatis MC2 155 include MSMEG_2036, MSMEG_1352, and MSMEG_6710. B. Phylogeny of the three BioH isoenzymes (BioH1 to BioH3) from M. smegmatis. Except for BioH enzyme-including subclades, all the other four subtrees were compressed with triangles, which correspond to the previously-identified isoforms (BioJ, BioK, BioG and BioV). The majority of BioH phylogeny consists of three clusters. Namely, they include i) Subclade I, termed ‘BioH1’ (exemplified with MSMEG_2036); ii) Subclade II, labeled ‘BioH2’ (featured with MSMEG_1352); and iii) Subclade III, designated ‘BioH3’ (presented with MSMEG_6710). Number on the node denotes the bootstrap replicate. It seemed true that a large population of Mycobacteria contains both BioH1 and BioH2. In spite of its absence in the mycobacterial species other than M. smegmatis, BioH3 appears in other closely-relative cousins, like Saccharopolyspora erythraea, an erythromycin-producing actinomycete. The software of MEGA7 was applied in the generation of Maximum Likelihood (ML) tree. Jones-Taylor-Thornton (JTT) model was used, and the number of bootstrap replications is 1000. C. Use of growth curves to evaluate of the E. coli ΔbioH mutant expressing each of the three putative bioH genes (bioH1 to bioH3). It was expressed in an average ± standard deviation (SD) from three independent experiments.
Fig 3.
Integrated evidence that BioH3 is a dimer.
A. Gel filtration profile suggests that solution structure of the purified BioH3 is a dimer. B. Chemical cross-linking analysis of the BioH3 protein. In addition to the monomeric band, dimeric band is given in the chemical cross-linking assay with the EGS cross-linker. C. Overall structure of dimeric BioH3. Architectural analysis of BioH3 suggests two binding-interfaces between protomer I and protomer II, namely Interface I and Interface II. D. Enlarged views of Interface I and Interface II. The critical residues involved in interface formation are labeled. The residues from protomer I are shown in magenta, and those from protomer II are given in cyan. Designations: mAU, milli absorbance units; kDa, kilo Dalton; EGS, ethylene glycol bis (succinimidyl succinate); FakA, fatty acid kinase A; FakB2, fatty acid kinase subunit B2; PA0502, a putative BioH enzyme from Pseudomonas aeruginosa; N, N-terminus; C, C-terminus.
Fig 4.
Functional and structural analysis of the two biochemically-amenable demethylases BioH2 and BioH3.
Ribbon structures of three known biotin gatekeeper enzymes BioH (A), BioG (B), and BioJ (C). D. Ribbon presentation of the monomeric structure of BioH3 (MSMEG_6710). BioH3 seemed unusual in that its lid domain displays a unique folding mode compared with the other three known enzymes. E. BioH2 (MSMEG_1352) is enzymatically active, albeit of its aggregation in solution. F. BioH3 (MSMEG_6710) displays enzymatic activity of removing the methyl (ethyl) moiety from M-pim-ACP (E-pim-ACP) to give pim-ACP product. The enzymatic mechanism of BioH2 (and/or BioH3) action was examined in vitro, using its physiological substrate M-pim-ACP, as well as an alternative one E-pim-ACP. The reactant and product from BioH2 (BioH3) reaction were separated using conformationally-sensitive gel [0.5M urea/17.5% PAGE (pH9.5)]. In general, the reactant of M-pim-ACP (and/or E-pim-ACP) migrates faster than the product pim-ACP in such PAGE gel containing 0.5M urea. Designations: α, α-helices; β, β-sheet; N, N-terminus; C, C-terminus; ACP, Acyl carrier protein; M-pim-ACP, Methyl-pimeloyl-ACP; E-pim-ACP, Ethyl-pimeloyl-ACP; Pim-ACP, pimeloyl-ACP. The minus symbol “—” denotes no addition of any enzyme.
Fig 5.
MS evidence for the enzymatic cleavage of the physiological substrate of pimeloyl ACP methyl ester by BioH2 and BioH3.
A. MALDI-TOF profile for the substrate of pimeloyl-ACP methyl ester. B. MALDI-TOF identification of the pimeloyl-ACP product arising from BioJ hydrolysis of pimeloyl-ACP methyl ester. MALDI-TOF evidence that BioH2 (C) and BioH3 (D) both demethylate the reactant pimeloyl-ACP methyl thioester to give pimeloyl-ACP product. E. Use of high-resolution mass spectrometry to detect a pool of ACP peptide fragments of which the serine 36 (S36) features with a single pimeloyl modification. Designations: ACP, Acyl carrier protein; Pim-ACP, pimeloyl-ACP; M-pim-ACP, Methyl-pimeloyl-ACP; BioJ, Fransicella pimeloyl-ACP methyl ester carboxyl-esterase; NA, not available; Ppan, Phosphopantetheine.
Table 1.
Data collection and refinement statistics.
Fig 6.
Structure-guided functional analysis for the conserved catalytic triad amongst the three BioH enzymes (BioH1 to BioH3).
Structural snapshots of catalytic triads from M. smegmatis BioH1 (A) and BioH2 (B). Structures of both BioH1 and BioH2 were predicted with Alpha-fold. C. X-ray crystal structure-based visualization for the BioH3 (MSMEG_6710) catalytic triad. The three conserved residues from the catalytic triad correspond to i) S110, D251 & H279 (BioH1); ii) S127, D255 & H283 (BioH2); and iii) S103, D232 & H265 (BioH3). D. Use of enzymatic assays to evaluate the roles of catalytic triad (S103, D232, and H265) in BioH3 (MSMEG_6710) activity. Three alanine-substituted mutants of BioH3 created here, included S103A, D232A, and H265A, respectively. The in vitro enzymatic actions of BioH3 mutants were evaluated using conformationally-sensitive gel as described in Fig 4. E. The BioH1 mutant defective in its catalytic triad cannot restore bacterial growth of the E. coli ΔbioH biotin auxotrophic strain on the non-permissive condition lacking biotin. F. None of BioH2 mutants with the defection of catalytic triad enables the E. coli ΔbioH strain to appear on the biotin-deficient growth condition. G. The certain mutation of catalytic triad causes functional loss of BioH3. In total, 9 mutants of M. smegmatis BioH (BioH1 to BioH3) were engineered into the ΔbioH biotin auxotrophic strain. On the basis of biotin-lacking cultivation condition, growth curves were plotted to address the in vivo role of BioH (BioH1 to BioH3) catalytic triad. Three independent experiments were conducted, and final output was given in an average ± SD.
Fig 7.
Structure-aided functional insights into recognition of the M-pim-ACP substrate by BioH3.
A. Electrostatic surface illustration for BioH3 structure. A PEG-bound tunnel is detected, and highlighted with a dashed line rectangle. The 2Fo-Fc electron density maps of PEG were contoured at 1.2δ level. It probably mimics the entry of physiological substrate M-pim-ACP into BioH3 enzyme. B. Structural presentation of BioH3 docked with the methyl pimeloyl moiety of M-C7-ACP. Presumably, the interplay of BioH3 with ACP comprises two interaction interfaces (highlighted with dashed line square), namely Interface I and Interface II. Both BioH3 and ACP are shown with cylinders. BioH3 is colored cyan, and ACP is displayed orange. Ppan-linked methyl pimeloyl moiety is given in sticks. C. Enlarged views of Interface I and Interface II engaged in BioH3-ACP interaction. As for Interface I, two electrostatic interactions are proposed. In brief, the two negatively-charged residues (D36 and D39) of ACP α2-helices pair with the two positively-charged amino acids (R136 and R134). Four positively-charged residues are suggested to involve in Interface II formation, namely R217, R218, R221, and R249. D. Use of site-directed mutagenesis to determine the contribution of interface I & II to BioH3 activity in vivo1 E. Growth curves of the E. coli ΔbioH derivatives carrying an array of plasmid-borne bioH3 mutants on the biotin-deficient, non-permissive condition. A panel of bioH3 mutants are cloned into pBAD322, transformed into the E. coli ΔbioH biotin auxotrophic strain, and functionally assayed on the basis of bacterial growth on the non-permissive M9 medium lacking biotin. 0.1% arabinose (0.1% ara) was supplemented to induce expression of bioH3 (its mutants). Namely, six single mutants of bioH3 included R134A, R136A, R217, R218, R221, and R249. Two of six residues (R134 and R136) are found to play major roles in BioH3 activity. A representative result from three independent experiments was given. Designations: α, α-helices; N, N-terminus; C, C-terminus; ACP, Acyl carrier protein; M-C7-ACP, Methyl-pimeloyl-ACP.
Fig 8.
Genetic and biochemical evidence that three BioH isoenzymes are dedicated to biotin synthesis.
A. Schematic illustration for genetic context of the triple bioH mutant (ΔbioH1/2/3) of M. smegmatis, a biotin auxotrophic strain. B. Chemical reaction for the three BioH isoforms that cleave the M-pim-ACP to produce pim-ACP. C. An individual bioH gene (bioH1 to bioH3) can significantly restore the viability of the triple mutant of ΔbioH1/2/3 on the biotin-deficient growth condition. The MC2 155 strain (WT) is the positive control, and the biotin auxotrophic strain of M. smegmatis (ΔbioAFD) lacking the late step of biotin synthesis, is the negative control. The triple mutant of bioH is denoted with ΔbioH1/2/3, and its complementary strain carrying plasmid-borne bioH is labeled with ΔbioH1/2/3+pbioH (H1 to H3). D. Schematic representation of the in vitro reconstitution of DTB synthesis. E. The purified form of BioH2 (and/or BioH3) protein reconstitutes DTB/biotin synthesis in vitro. The biotin auxotrophic strain FYJ283 of A. tumefaciens NTL4 (ΔbioBFDA) we earlier developed [47], acted as an indicator bacterium as described recently by Zhang et al. [37]. Designations: M-pim-ACP, Methyl pimeloyl-ACP ester; Bpl, Biotin protein ligase; M-C7-ACP, Pimeloyl-ACP methyl ester.