Allosteric inhibition of Staphylococcus aureus MenD by 1,4-dihydroxy naphthoic acid: a feedback inhibition mechanism of the menaquinone biosynthesis pathway

Abstract

Menaquinones (MKs) are electron carriers in bacterial respiratory chains. In Staphylococcus aureus (Sau), MKs are essential for aerobic and anaerobic respiration. As MKs are redox-active, their biosynthesis likely requires tight regulation to prevent disruption of cellular redox balance. We recently found that the Mycobacterium tuberculosis MenD, the first committed enzyme of the MK biosynthesis pathway, is allosterically inhibited by the downstream metabolite 1,4-dihydroxy-2-naphthoic acid (DHNA). To understand if this is a conserved mechanism in phylogenetically distant genera that also use MK, we investigated whether the Sau-MenD is allosterically inhibited by DHNA. Our results show that DHNA binds to and inhibits the SEPHCHC synthase activity of Sau-MenD enzymes. We identified residues in the DHNA binding pocket that are important for catalysis (Arg98, Lys283, Lys309) and inhibition (Arg98, Lys283). Furthermore, we showed that exogenous DHNA inhibits the growth of Sau, an effect that can be rescued by supplementing the growth medium with MK-4. Our results demonstrate that, despite a lack of strict conservation of the DHNA binding pocket between Mtb-MenD and Sau-MenD, feedback inhibition by DHNA is a conserved mechanism in Sau-MenD and hence the Sau MK biosynthesis pathway. These findings may have implications for the development of anti-staphylococcal agents targeting MK biosynthesis.

This article is part of the theme issue ‘Reactivity and mechanism in chemical and synthetic biology’.

1. Introduction

Isoprenoid quinones such as menaquinones (MKs, vitamin K2) and ubiquinones (coenzyme Q) are redox-active, membrane-bound molecules with vital roles across all domains of life [1]. MKs, considered to be the most ancient types of isoprenoid quinones, function as electron carriers in bacterial electron transport chains, and are essential in ATP-generating redox reactions in mycobacteria, Gram-positive bacteria and anaerobically respiring Gram-negative bacteria. MKs also function as environmental sensors detecting changes in redox state and oxidative stress [2–8], and have been linked to biofilm formation in Staphylococcus aureus (Sau) [6], and latency and virulence in Mycobacterium tuberculosis (Mtb) [5,9,10].

MKs consist of a naphthoquinone head group linked to an isoprenoid tail of varying lengths and are, therefore, referred to as menaquinone-n (MK-n), where n denotes the number of isoprenyl side chain units. Side chain lengths of 6–10 are common, but isoprenyl units of 1–5 and 11–15 are also found [2,11]. Two different pathways have evolved in bacteria for the biosynthesis of MK; the classical pathway (figure 1a), used by bacteria such as Escherichia coli (Ec), Mtb and Sau, and the less common alternative (or futalosine) pathway [1,2]. Commonalities of both pathways include the separate synthesis of the naphthoquinone headgroup precursor and isoprenoid side chain. These components are then combined and further modified.

Figure 1.

The classical menaquinone biosynthesis pathway and MenD catalytic cycle. (a) The classical menaquinone biosynthesis pathway as currently known showing the reactions catalysed by MenA, B, C, D, E, F, H, I and UbiE (sometimes termed MenG) with: SEPHCHC, 2-succinyl-5-enolpyruvyl-6-hydroxy-3-cyclohexene-1-carboxylic acid; SHCHC, 2-succinyl-6-hydroxy-2,4-cyclohexadiene-1-carboxylate; OSB, o-succinylbenzoate; DHNA, 1,4-dihydroxy-2-naphthoic acid; CoA, coenzyme A; reactions from MenA onwards are known to be membrane bound. In Mtb DHNA has been shown to inhibit the MenD reaction. (b) The MenD catalytic cycle in closer detail, showing the activated ThDP ylide, the two-step reactions adding each substrate (substrate 1: α-ketoglutarate/2-oxoglutarate, and substrate 2: isochorismate) and the resulting two ThDP-bound reaction intermediates in the cycle. The aminopyrimidine ring of the ThDP is shown as the 4′-aminopyrimidine (AP) tautomer in the ThDP ylide, the APH⁺ tautomer in the intermediate I carbanion and as the 1′,4′-iminopyrimidine (IP) tautomer in the intermediate II form.

The MK biosynthesis pathway is absent in humans, and antibacterial drug discovery efforts targeting numerous enzymes from this pathway in bacterial pathogens have been reported [12–16]. Sau is an important human pathogen, and uses MK (predominantly MK-8) for both aerobic and anaerobic respiration [17–19]. Sau is associated with a variety of skin, bone, joint and blood infections with considerable morbidity and mortality [20,21]. Treatment of Sau infections can be extremely challenging owing to the formation of persister populations and small colony variants (SCVs), high tolerance of Sau biofilm to antibiotics, and the emergence of antibiotic-resistant strains [22–25]. These obstacles, in particular the appearance and spread of new drug-resistant strains, highlight the importance of identifying new inhibitors to treat Sau infections. Inhibitors of MenA, MenB and MenE enzymes in the Sau MK biosynthesis pathway exhibit promising growth inhibitory [26–30] and biofilm formation inhibitory activity [26], validating the druggability of this pathway for the development of Sau-targeting antibacterial agents.

Until recently, little was known about how the MK biosynthesis pathway is regulated. Levels of MK, and those of some upstream metabolites of the pathway, are likely to require tight regulation, as excessive concentrations of these molecules could result in cell toxicity if the redox balance is disrupted [2]. We showed evidence of feedback regulation of the MK biosynthesis pathway in Mtb [31], by identifying the negative allosteric regulation of the first committed enzyme in MK biosynthesis, MenD, by a downstream metabolite, 1,4-dihydroxy-2-naphthoic acid (DHNA).

MenD is a thiamine diphosphate (ThDP)-dependent enzyme that catalyses the conversion of oxoglutarate and isochorismate to 2-succinyl-5-enolpyruvyl-6-hydroxy-3-cyclohexene-1-carboxylic acid (SEPHCHC) (figure 1) [32,33]. Catalysis occurs through two covalent ThDP intermediates as the substrates oxoglutarate and isochorismate are successively added to ThDP before SEPHCHC is released (figure 1b) [31,34]. The Mtb-MenD was found to bind DHNA in a pocket distinct from the active site in an arginine cage comprising Arg97, Arg277 and Arg303. All three arginine residues were identified as being crucial for full Mtb-MenD activity and for inhibition by DHNA [31]. Although based on sequence similarity, the Mtb-MenD allosteric site was found to be well-conserved among mycobacteria and closely related Rhodococcus sp.; the site was not strictly conserved in other microorganisms, including Sau. This suggested either that regulation of the MK biosynthesis pathway by DHNA may be limited to a small group of bacteria, or that DHNA (or related molecules) may still bind in this region despite the absence of some key elements, such as the full arginine cage, found in the Mtb-MenD.

In this work, we report that the Sau-MenD is allosterically inhibited by DHNA. Thus, our results demonstrate that inhibition by DHNA of the MK biosynthesis pathway is conserved between Mtb and Sau, which hints at an ancient origin of this allosteric feedback mechanism in Gram-positive bacteria and mycobacteria. Furthermore, we show that exogenous DHNA inhibits the growth of Sau bacteria, underscoring the importance of the MK pathway for growth. These findings suggest that small molecules based on the DHNA scaffold may be effective for the development of new antibacterial drugs.

2. Material and methods

(a) . Cloning, expression and purification of Sau-MenD

The MenD open reading frames (ORFs) from Sau IS1 (sequence identical to gene locus tag EKM74_RS13930 in NCBI reference sequence NZ_AP019306.1) and Sau ATCC 25923 (gene locus tag KQ76_RS04840 in NCBI reference sequence NZ_CP009361.1) were amplified by PCR using MenD_Sau_F1 and MenD_Sau_R1 primers specified in electronic supplementary material, table S1. The amplified MenD ORFs were cloned into a modified pET30a vector using the restriction enzymes NcoI and HindIII to generate rTEV (recombinant tobacco etch virus protease)-cleavable N-terminally His₆-tagged proteins.

Construction of alanine mutants (Arg98Ala, Lys283Ala and Lys309Ala) of the Sau IS1 MenD ORF was performed by site-directed mutagenesis and In-Fusion cloning using primers listed in electronic supplementary material, table S1.

His₆-tagged Sau-MenD was expressed in E. coli BL21 (DE3) cells. Seed cultures were grown for 18 ±2 h in Luria–Bertani Miller (LB) medium (Thermo Scientific) with 50 µg ml⁻¹ kanamycin at 37°C. Seed cultures were diluted 1/200 in Terrific Broth autoinduction medium (24 g l⁻¹ yeast extract, 12 g l⁻¹ tryptone, 0.8% (v/v) glycerol, 0.0162 M KH₂PO₄, 0.0528 M K₂HPO₄, 2 mM MgSO₄, 0.375% (w/v) aspartic acid, 0.015% (w/v) glucose and 0.5% w/v lactose, made up concentration pH 7.3) containing 50 µg ml⁻¹ kanamycin and cultured at 37°C under agitation (180 r.p.m.) for 3 h, followed by a further 20 h incubation at 18°C and 180 r.p.m. Cells were harvested and lysed in lysis buffer (20 mM HEPES pH 8.0, 150 mM NaCl, 5 mM CaCl₂, 5% (v/v) glycerol, 50 mM imidazole and 1 mM tris(2-carboxyethyl)phosphine (TCEP)) using a Microfluidics cell disrupter (Newton, MA). Recombinant Sau-MenD was purified at ambient temperature by immobilized metal affinity chromatography (IMAC) with 5 ml HisTrap HP columns (GE Healthcare) and an imidazole gradient of 20–500 mM over 50 ml. Peak fractions were immediately diluted 1 : 2 in 1× lysis buffer to avoid protein precipitation. To remove the N-terminal His₆-tag from the recombinant protein, TEV-cleavage [35] was performed overnight at 4°C by dialysing the protein in 1× lysis buffer without imidazole. Subtractive IMAC was then conducted at room temperature with 5 ml HisTrap HP columns to remove the cleaved His₆-tag and rTEV. Fractions containing the target protein were combined and concentrated for further purification by size-exclusion chromatography (SEC). SEC was carried out at room temperature on a Superdex 200 10/300 column with buffer containing 20 mM HEPES pH 8.0, 150 mM NaCl, 5 mM CaCl₂, 5% (v/v) glycerol and 1 mM TCEP. For long-term storage, 50% (v/v) glycerol was added to take the final concentration to 10% (v/v) and the protein solution was kept at −80°C.

(b) . Sequence alignment of MenD proteins

MAFFT sequence alignment of Sau-MenD_IS1 and Mtb-MenD was performed in Geneious [36] using MAFFT v.7.388 [37], the AUTO algorithm, Blosum45 scoring matrix, a gap open penalty of 1.53 and an offset value of 0.153.

(c) . Crystallization

Initial Sau-MenD low-resolution structures (incomplete and not presented in this work) were from both IS1 and ATCC wild-type (WT) enzymes in a range of MORPHEUS screen conditions [38] without and with ThDP present (1–2.5 mM). Optimized crystals for the higher-resolution structure presented in this work were grown using tag-removed recombinant Sau-MenD_IS1 (15 mg ml⁻¹ in 20 mM HEPES pH 8.0, 150 mM NaCl, 5 mM CaCl₂, 5% (v/v) glycerol, 2 mM TCEP and 2 mM ThDP) in 96-well sitting-drop format from MORPHEUS screen condition H12 (12.5% (w/v) PEG 1000, 12.5% (w/v) PEG 3350, 12.5% (v/v) 2-methyl-2,4-pentanediol (MPD), 0.02 M of amino acids (sodium l- and d-glutamate, l- and d-alanine, glycine, l- and d-lysine HCl, l- and d-serine and 0.1 M bicine/Trizma base pH 8.5)). Crystals grew within 7 days and were flash frozen in liquid nitrogen.

(d) . Data collection, structure determination and refinement

All diffraction data were collected using the macromolecular crystallography beamline MX2 at the Australian Synchrotron, equipped with a Dectris EIGER 16M detector [39]. Data were autoprocessed via the Australian synchrotron pipeline, using iterative rounds of XDS auto-indexing [40,41], and the resulting .hkl file was imported into the CCP4 program suite [42] for space group assignment, merging, truncating and generation of an R_free set of 5% using AIMLESS [43]. Analyses of merged CC½ correlations between intensity estimates from half datasets were used to influence high-resolution cut-off for data processing [44]. Matthews co-efficient analysis [45] suggested four molecules per asymmetric unit (47.9% solvent content).

The structures were solved by molecular replacement using Phaser [46], with the above-mentioned initial low-resolution partially complete Sau-MenD model used as a search model (this model was generated from lower-resolution Sau-MenD data using molecular replacement from Bacillus subtilis (Bs)-MenD Protein Data Bank (PDB) code 2X7J [33] and manual model building/refinement).

The Sau-MenD IS1 model presented here was completed with iterative rounds of manual building using Coot [47] and refinement using REFMAC5 [48] and PHENIX [49]. Additional density corresponding to ThDP, ions and glycine (from the crystallization medium) were modelled using available PDB dictionary restraints. Water molecules were identified by their spherical electron density and appropriate hydrogen-bond geometry with the surrounding structure. The final refined 2.35 Å structure was deposited in the PDB with the code 7TIN. Unless otherwise stated, all protein structure images were generated using PyMOL (PyMOL Molecular Graphics System, v.1.5, Schrödinger).

(e) . Native mass spectrometry

Protein samples for native mass spectrometry (Sau-MenD_IS1 4.7 mg ml⁻¹, and Sau-MenD_{ATCC 25923} 6.3 mg ml⁻¹) were exchanged from the SEC buffer into 2 M ammonium acetate pH 7.4 via a single-pass through a Bio-Spin P-6 column (Bio-Rad). Buffer-exchanged protein was loaded into gold-coated glass capillaries (1.0 mm outer diameter/0.75 mm inner diameter with filament), fabricated with orifice sizes of ca 5 µm, using a P-2000 puller (Sutter). Spectra were acquired using a Synapt XS (Waters) equipped with a 32k quad, with instrument settings optimized for the transmission of high-mass ions, and soft ionization and activation conditions [50,51]. The instrument was mass calibrated using CsI. Spectra were smoothed (3 × 20, mean) and mass assigned in MassLynx (Waters).

(f) . Intrinsic fluorescence quenching experiments

Intrinsic fluorescence quenching experiments were performed at 25°C with a Varian Cary Eclipse Fluorescence Spectrophotometer (Agilent Technologies) with slits set at 5 nm bandwidth. The excitation wavelength was 290 nm and emission spectra were recorded in the 300–400 nm range. Experiments were conducted in 3 ml quartz cuvettes with 1 µM Sau-MenD in an assay buffer consisting of 50 mM HEPES pH 8.0, 150 mM NaCl, 5 mM CaCl₂ and 1 mM TCEP. For Sau-MenD samples in the presence of the reaction ligands, the enzyme was preincubated with 100 µM ThDP only or 100 µM ThDP and 300 µM oxoglutarate for 45 min at ambient temperature. DHNA (Sigma-Aldrich) was titrated from stock solutions that were prepared in the assay buffer and kept on ice. Following DHNA addition, cuvette contents were mixed immediately and fluorescence intensity was measured.

The fluorescence intensities obtained at 340 nm were corrected for the dilution factor as well as for inner filter effects [52] using equation (2.1),

$$\mathrm{Fi}\hspace{0.17em}\mathrm{corr}=\mathrm{Fi}\hspace{0.17em}\mathrm{dil}\hspace{0.17em}\times {10}^{\mathrm{Aex}+\mathrm{Aem}/2},$$ 2.1

where Fi corr is the corrected value of fluorescence intensity at the titration point, Fi dil is the dilution-corrected fluorescence intensity measured, and Aex and Aem are the absorbance of DHNA at excitation (290 nm) and emission wavelengths (340 nm).

Corrected intensities were plotted against the DHNA concentration and, to determine the K_d value for DHNA, the resulting data were fitted in GraphPad Prism 9 using equation (2.2),

$$F=F_{\max }-{\displaystyle \frac{\left(X+K_{\mathrm{d}}+M-\sqrt{{(X+K_{\mathrm{d}}+M)}^2-4XM}\right)}{2M}(F_{\max }-F_{\min }),}$$ 2.2

where F is the measured fluorescence intensity of Sau-MenD, F_max is the fluorescence intensity of Sau-MenD in the absence of DHNA, F_min is the fluorescence intensity when Sau-MenD is saturated with DHNA, M is the micromolar concentration of enzyme used in the assay, X is the micromolar DHNA concentration and K_d is the dissociation constant.

(g) . UV spectroscopy assays

The isochorismate mixture was prepared using commercially produced chorismic acid and laboratory-made Ec-MenF as previously described [31,32]. Sau-MenD activity and DHNA-inhibition assays were performed using the UV–vis spectroscopy assay measuring the decrease in isochorismate absorbance at 278 nm (ε₂₇₈ = 8300 M⁻¹ cm⁻¹) as described by Bashiri et al. [31] with some modifications. The assay buffer consisted of 50 mM HEPES pH 8.0, 150 mM NaCl and 5 mM CaCl₂. Reaction mixes consisting of 0.4 µM Sau-MenD, 100 µM ThDP and 300 µM oxoglutarate were incubated in the reaction buffer at 37°C for 30 min prior to adding 20 µM isochorismate to initiate the reaction. All assays were carried out using a Cary 400 UV–vis spectrophotometer (Agilent Technologies) and quartz cuvettes, with a final reaction volume of 150 µl. Initial rate data were fitted using the Cary WinUV software.

DHNA-inhibition assays contained 0.4 µM protein, 100 µM ThDP, 300 µM oxoglutarate and various concentrations of DHNA (0–51.2 µM). After preincubation of these components at 37°C for 30 min, 10 µM isochorismate was added to initiate the reactions. Stock solutions of DHNA were freshly prepared in assay buffer prior to each experiment and stored on ice. At 25.6 and 51.2 µM DHNA, an increase in absorbance was observed for control samples run without isochorismate, presumably owing to gradual DHNA oxidation. To account for this, the linear increase in absorbance at 278 nm of the non-isochorismate controls was deducted from the initial rate data of these samples. Initial rate data were then fitted to a four-parameter logistic Hill model with GraphPad Prism 9 using equation (2.3),

$$Y=R_{\mathrm{\infty }}+{\displaystyle \frac{(R_0-R_{\mathrm{\infty }})}{(1+{10}^{(({\mathrm{logIC}}_{50}-X)\times n)}\hspace{0.17em})},}$$ 2.3

where Y is the reaction rate, X is the logarithm of the inhibitor concentration, R₀ and R_∞ are the asymptotic values for reactions rates at very low and very high inhibitor concentrations, respectively, IC₅₀ is the half-maximal inhibitory concentration and n is the Hill slope parameter describing the steepness of the fitted curve.

(h) . Growth assays with DHNA and MK-4 rescue assays

Sau and Pseudomonas aeruginosa (electronic supplementary material, table S2) were grown for 18 ± 2 h in 5 ml of LB medium at 37°C and 180 r.p.m. Cultures were then diluted in fresh LB medium to contain 2.7 × 10⁶ CFU ml⁻¹. A Nunc Microwell 96-well flat-bottom plate (Thermo Scientific) was seeded in triplicate (from three independent starter cultures) with 75 µl of each of the diluted cultures and 75 µl of 2× working solutions of DHNA, MK-4 or DHNA + MK-4 (prepared from the following stock solutions: 100 mM DHNA dissolved in dimethyl sulfoxide (DMSO), and 50 mM MK-4 (Sigma-Aldrich) dissolved in 100% ethanol). Plate wells contained a final concentration of 10⁶ CFU well⁻¹ with 0, 50 or 100 µM DHNA for growth assays, or with 0 or 100 µM DHNA supplemented with 337 µM MK-4 for growth rescue assays. Plates were incubated at 37°C and 300 r.p.m. for 24 h. Growth was determined by measuring the OD₆₀₀ of the cultures every 30 min during the incubation period. Data from biological replicates were averaged and subtracted from blank data (sterile medium with appropriate additives).

3. Results

The Sau-MenD shares 28% identity with its well-studied homologue in Mtb (Mtb-MenD) (electronic supplementary material, figure S1) and eluted in one peak in SEC at an elution volume nearly identical to the Mtb-MenD protein (data not shown). The tetrameric conformation of this protein was confirmed by native mass spectrometry (electronic supplementary material, figure S2).

(a) . Sau-MenD binds DHNA

To determine whether the Sau-MenD can bind DHNA, the intrinsic fluorescence of MenD enzymes from Sau IS1 and Sau ATCC 25923 was monitored upon DHNA titration in the presence and absence of the cofactor ThDP, and in the presence of ThDP and first substrate oxoglutarate. The two Sau-MenD variants differed in sequence by seven amino acids and were, therefore, used to assess potential differences in their ability to bind DHNA. Although there are no tryptophan residues present in the DHNA binding sites of either Sau-MenD variant (based on the Mtb-MenD DHNA binding site), each of the enzymes contains four tryptophan residues, one of which is near the putative DHNA binding pocket (Trp293).

An emission maximum at 340 nm resulted from the excitation of the Sau-MenD proteins at 290 nm. Fluorescence quenching upon DHNA titration had saturation behaviour and was fitted to equation (2.2) to estimate the K_d values (figure 2). Our results show that the apo-enzyme, ThDP-bound and ThDP- and oxoglutarate-bound forms of Sau-MenD bind DHNA with K_d values between 58 and 110 µM for the IS1 protein and 67–85 µM for the ATCC 25923 variant. The affinity of the ThDP- and oxoglutarate-bound forms of Sau-MenD_IS1 for DHNA (K_d = 58 µM, 95% CI 38–87 µM) is slightly higher than that of the apo-enzyme (K_d = 110 µM, 95% CI 91–150 µM). While the K_d values for the apo-enzyme (K_d = 85 µM) versus the ThDP- and oxoglutarate-bound states (K_d = 67 µM) of Sau-MenD_{ATCC 25923} show a similar trend, overlapping 95% confidence intervals hinder a definitive conclusion.

Figure 2.

Intrinsic fluorescence quenching of Sau-MenD enzymes upon DHNA titration. Intrinsic fluorescence quenching experiments were performed with 1 µM enzyme. For Sau-MenD samples complexed with 100 µM ThDP only, or 100 µM ThDP and 300 µM oxoglutarate (oxo), enzymes were preincubated with the respective ligands for 45 min at ambient temperature. DHNA was then titrated, cuvette contents were mixed immediately and fluorescence intensity was measured twice for each titration. The fluorescence intensities obtained at 340 nm were corrected for the dilution factor and inner filter effects. To obtain dissociation constants for DHNA, the resulting data were fitted in GraphPad Prism with equation (2.2). K_d values are shown with respective 95% confidence intervals in parentheses.

(b) . DHNA inhibits Sau-MenD SEPHCHC synthase activity

We previously showed that DHNA inhibits the SEPHCHC activity of 0.6 µM Mtb-MenD with an IC₅₀ of 53 nM [31]. To determine whether this inhibitory mechanism is conserved in Sau-MenD, the effect of DHNA on the activity of MenD enzymes from Sau IS1 and Sau ATCC 25923 was investigated in a UV–vis spectroscopy-based assay by measuring isochorismate consumption at 278 nm in the presence of various concentrations of DHNA (0–51.2 µM). Our assay results (figure 3 and table 1) show that DHNA inhibited 0.4 µM Sau-MenD_IS1 and Sau-MenD_{ATCC 25923} with IC₅₀ values of 3.7 and 2.3 µM, respectively, suggesting this last soluble metabolite of the pathway is also an allosteric regulator of the Sau-MenD.

Figure 3.

Inhibition of Sau-MenD enzymes by DHNA. Using a UV spectroscopy-based assay for isochorismate consumption, inhibition assays were performed with 0.4 µM enzyme, 100 µM ThDP, 300 µM oxoglutarate and various concentrations of DHNA (0.1–51.2 µM). Following a 30 min preincubation of enzyme with ThDP, oxoglutarate and DHNA, 10 µM isochorismate was added to initiate the reaction. Initial rates were measured and fitted to the four-parameter logistic Hill equation. Data shown for each enzyme are from two independent experiments.

Table 1.

Activity and DHNA inhibition of Sau-MenD variants.

	activity relative to WT (%)	95% CI for activity relative to WT	activity reduction in the presence of 12.5 µM DHNA (%)	IC50 of DHNA (µM)	95% CI for DHNA IC50	Hill slope	95% CI for Hill slope
WT (IS1)			82	3.7	2.5–6.7	−1.0	−1.4 to −0.70
Arg98Ala	150	130–170	21	>10	—	−2.3	n.d.a
Lys283Ala	220	210–230	28	>10	—	−1.2	−1.8 to n.d.a
Lys309Ala	160	120–200	80	4.6	3.3–7.5	−1.0	−1.4 to −0.74
ATCC 25923	99	88–110	87	2.3	1.7–3.2	−0.99	−1.3 to −0.75

^an.d. indicates that a complete confidence interval (CI) was unable to be determined and therefore the respective best-fit values should be interpreted with caution.

(c) . Structure of Sau-MenD

The Sau-MenD_IS1 structure was solved (processing and refinement statistics—electronic supplementary material, table S2) to a resolution of 2.35 Å in space group P2₁2₁2₁ with four chains (A–D, with interpretable density across their length for residues 2–554/557) in the asymmetric unit comprising the biological ‘dimer of dimers’ tetrameric unit (figure 4a), consistent with our native mass spectroscopy analysis (electronic supplementary material, figure S2). The Sau-MenD_IS1 tetramer structure is typical of all known MenD structures to date and, like the Bs-MenD [33] and Ec-MenD [53] structures, is symmetrical, with nearly identical conformations for each monomer (r.m.s.d. using SSM [54] of 0.29–0.44 Å over the length of the chain), and well-defined ThDP and Ca²⁺ cofactors bound to all four active sites.

Figure 4.

Structure of Sau-MenD. (a) Cartoon depiction of the tetrameric ‘dimer of dimers’ present in the asymmetric unit. Dimer 1 comprises the green and orange chains, dimer 2 the teal and pink chains, and the four ThDP in the active sites are shown as yellow sticks. (b) Cartoon depiction of the three-domain monomer fold. Domain I (PYR) is shown in light blue, domain II (TH3) in teal, and domain III (PP) in green, and ThDP is shown as yellow sticks with the associated Ca²⁺ ion as a green sphere. The location of the DHNA binding site from Mtb-MenD is depicted by a star and the Mtb-MenD DHNA from the overlay in (d) is shown in blue sticks. (c) A close-up of the Sau-MenD active site. ThDP is shown as yellow sticks with the associated Ca²⁺ ion as a green sphere. Active site residues are shown as sticks (domain III (PP) is green with green labels and domain I (PYR) is light blue with blue labels) and selected hydrogen bonds are shown as yellow dashes. (d) Overlay of the Mtb-MenD DHNA binding site (grey-blue sticks with grey labels) with DHNA bound (blue sticks) and the Sau-MenD putative allosteric site (green sticks with green labels).

The three-domain α/β monomeric fold (figure 4b) is typical of enzymes from the decarboxylase superfamily of ThDP-dependent enzymes [55], with domain I (PYR domain, residues 1–200) pairing with domain III (PP domain, residues 364–557) from the other monomer in the dimer to form each of the four paired MenD active sites. Similar to Bs-MenD [33], the extended Sau-MenD_IS1 C-terminal helix at the end of domain III extends down, packing against a domain III active site lid (residues 469–489) before ending in domain II (TH3, residues approx. 220–364) (figure 4b). Domain II (TH3), the most sequence-diverse of the MenD domains, is connected to domain I via a long linker (approx. residues 201–220) and to domain III via a long helix (residues 341–364). Domain II, which has no known function in many decarboxylase superfamily enzymes, has been associated with the formation of the allosteric regulatory DHNA binding site in Mtb-MenD [31].

Despite the overall modest sequence conservation across MenD enzymes, the Sau-MenD_IS1 active site shows high conservation with other well-studied MenD enzymes for the residues involved in ThDP, substrate/intermediate binding and catalysis (figure 4c). The ThDP diphosphate and associated Ca²⁺ are anchored to the enzyme via domain III (Ser390, Met391, Ser440, Asp438, Gly's 467–469, Asn465, Ile471), as is the ThDP thiazole ring, which is sandwiched between the side chains of Phe471 and Ile416, facilitating the catalytically active V-shaped ThDP conformation (figure 4c). The ThDP aminopyrimidine ring inserts into a pocket between domain I and domain III, forming hydrogen bonds with Asn414, Ile416 and Glu56—the latter playing an important role in proton transfer and stabilization of the various catalytically important aminopyrimidine ring tautomers [32,34].

Substrate/intermediate binding residues come from both domain I and III: intermediate I (Arg394, Arg411) and intermediate II (Ile470, Phe471 and Leu474) interacting residues from domain III; carboxylate/decarboxylation binding pocket residues Ser33, Arg34 and intermediate II interacting residues Arg108 and Glu119 from domain I (figure 4c) [31,32]. Invariant in MenD enzymes, Glu119 has been ascribed several important roles during catalysis, including product release [31,32]. Both Arg108 and Glu119 reside on a mobile loop shown in Mtb-MenD to connect the active site and allosteric sites (loop residues 114–116 cap the allosteric site (figure 4d)) [31]. This loop normally contains another key intermediate I/II binding residue (equivalent to Mtb-MenD Asn117), but in Sau-MenD this is replaced by a proline, and the side chain of Asn414 from domain III takes on a structurally analogous role, similar to Bs-MenD [32,33].

By contrast to the high conservation at the active site, conservation of the Mtb-MenD DHNA binding allosteric site across MenD enzymes appears limited. Sequence comparisons with Sau-MenD indicated low conservation of the allosteric region, with the potential for conservation of one of the arginine cage residues and substitution of another for a Lys (Sau-MenD Arg98 and Lys283). Attempts to co-crystallize DHNA with Sau-MenD resulted in precipitation of the enzyme, while soaking greatly impaired diffraction. Therefore, structural comparisons of the Sau-MenD structure and DHNA-bound Mtb-MenD were used to gain insight into the structural conservation at the putative allosteric site. These overlays reveal that the likely structural equivalents for the arginine cage are Arg98, Lys283 and Lys309 (figure 4d). None of these residues matches the side chain conformers that would be needed for DHNA binding, suggesting that for DHNA to bind, conformational rearrangements would be needed along with backbone movements around the 305–312 region. In the Sau-MenD structures, Arg98 is extensively hydrogen bonded to the backbone of Asn308, Lys309 and the Asn308 side chain. Interaction and stacking with DHNA would interfere with these interactions and could facilitate the movement of the region to accommodate DHNA.

(d) . Residues in the DHNA binding pocket are important for activity and inhibition

To validate our structural findings and investigate whether the equivalent arginine cage candidate residues in Sau-MenD (Arg98, Lys283 and Lys309) are important for activity and DHNA inhibition, we performed alanine mutagenesis experiments of the WT Sau-MenD_IS1. Interestingly, all three mutants showed enhanced activity compared with the WT Sau-MenD_IS1 when measured under the same conditions and in the absence of DHNA (figure 5a, table 1). Mean activity compared with the WT was 150, 220 and 160% for the Arg98Ala, Lys283Ala and Lys309Ala mutants, respectively, suggesting all three residues play an important role in catalytic activity, as was shown for the equivalent residues in Mtb-MenD. However, unlike Mtb-MenD, where these mutations negatively impacted activity, the mutations in Sau-MenD consistently enhanced activity in the conditions tested. This implies that some feature associated with the WT allosteric site in its unoccupied conformation that limits activity may be altered by the changes that occur when this region is mutated.

Figure 5.

Activity and DHNA inhibition of Sau-MenD mutants. Using a UV spectroscopy-based assay for isochorismate consumption, activity (in the absence of DHNA) and DHNA inhibition of Sau-MenD alanine mutants were investigated. (a) Activity of Sau-MenD mutants compared with WT Sau-MenD_IS1 and Sau-MenD_{ATCC 25923}. Three independent experiments were performed in technical triplicate and initial rate data were determined. (b) IC₅₀ data for DHNA against Sau-MenD mutants. Data shown for each enzyme are from two independent experiments. Initial rates were measured and fitted to the four-parameter logistic Hill equation. Assays were performed with 0.4 µM enzyme, 100 µM ThDP and 300 µM oxoglutarate. DHNA (0.1–51.2 µM) was present for inhibition assays only. Following a 30 min preincubation of enzyme with ThDP, oxoglutarate and DHNA (inhibition assays only), reactions were initiated by the addition of 20 µM (or 10 µM for inhibition assays) isochorismate.

IC₅₀ experiments revealed that DHNA does not inhibit the Arg98Ala and Lys283Ala mutants as effectively as it does the WT protein (figures 3 and 5b, and table 1). DHNA has strong absorbance at 278 nm; thus it was not possible to test DHNA concentrations above 51.2 µM. This led to our inability to accurately determine IC₅₀ values greater than 10 µM and truncated inhibition curves for the Arg98Ala and Lys283Ala mutants (figure 5b). As such, we were unable to determine IC₅₀ values for these mutants. The truncated inhibition curves of the Arg98Ala and Lys283Ala mutants are an indication that these residues play an important role in DHNA inhibition. By contrast, the Lys309Ala mutant remained sensitive to DHNA inhibition with an IC₅₀ of 4.6 µM, similar to that of the WT.

To illustrate the differences in DHNA sensitivity between the mutant and WT enzymes, the percentage inhibition in activity at 12.5 µM DHNA was also determined (table 1). This was the concentration at which the greatest differences in sensitivity were observed between the WT and mutant enzymes and it was the highest concentration of DHNA tested that did not require correction of the initial rate data owing to gradual DHNA oxidation. At 12.5 µM DHNA, the catalytic activities of the WT enzyme and Lys309Ala mutant were reduced by 82 and 80%, respectively, whereas the activities of the Arg98Ala and Lys283Ala mutants were notably less affected, with reductions of 21 and 28%, respectively.

(e) . DHNA supplementation inhibits the growth of Sau in media

As the first metabolite in the MK biosynthesis pathway with a complete naphthoquinol ring, DHNA levels in the cells are likely to require tight regulation, as excessive amounts could result in cell toxicity if the redox balance is disrupted. To test this premise, four different Sau strains, including methicillin-resistant and methicillin-sensitive clinical isolates, were grown in media supplemented with various concentrations of DHNA (0–150 µM). Pseudomonas aeruginosa was used as a control as this bacterium does not produce MK, only ubiquinone, and therefore does not contain classical MK biosynthesis genes [56]. We found that exogenous DHNA at 50 µM clearly impaired the growth of all four Sau strains (figure 6a–d). The growth impact was more severe at 100 µM DHNA, and at 150 µM DHNA the growth of all four Sau was completely inhibited for the duration of the assay (24 h). By contrast, the growth of P. aeruginosa was unaffected at DHNA concentrations up to 150 µM (figure 6e).

Figure 6.

Growth assays with DHNA and MK-4 rescue assays. (a–e) Exogenous DHNA inhibits the growth of Sau. Bacteria were grown at 37°C under agitation in LB medium containing various concentrations of DHNA (0, 50, 100 and 150 µM). (f–i) MK-4 rescues growth of DHNA-treated Sau. Sau strains were grown at 37°C under agitation in LB medium with DHNA (100 µM) and without DHNA, and supplemented with 337 µM MK-4. Growth for both assays was monitored over 24 h by OD₆₀₀ measurements taken every 30 min. Data represent the mean ± s.d. of three biological replicates.

(f) . MK-4 rescues growth defects caused by exogenous DHNA

To ascertain whether growth inhibition caused by DHNA is due to inhibition of the MK biosynthetic pathway, rescue experiments were performed by adding 337 µM MK-4 (150 µg ml⁻¹ MK-4) to culture medium with (100 µM) and without DHNA present, and monitoring growth of the Sau bacteria over 24 h. In the absence of DHNA, the addition of MK-4 slightly impacted the growth of the bacteria (figure 6f–i). This may be due to oxidative stress, as MK is a redox-active metabolite. Despite this, MK-4 addition to culture medium containing 100 µM DHNA resulted in a significant growth improvement of the strains compared with growth in the presence of 100 µM DHNA only. Notably, growth was not completely restored to levels observed in the absence of DHNA, which is possibly because MK-4 is not the preferred MK for Sau (predominant MK in Sau is MK-8).

4. Discussion

We recently discovered feedback regulation of the MK biosynthesis pathway, with the last cytosolic metabolite of the pathway, DHNA, found to allosterically inhibit Mtb-MenD [31]. Sequence alignments of MenD homologues suggested a limited conservation of this allosteric site in other bacteria, including the pathogen Sau. However, data presented here demonstrate that DHNA is also able to bind to and inhibit the SEPHCHC synthase activity of Sau-MenD enzymes.

To better understand the basis for the conservation of allosteric inhibition of Sau-MenD by DHNA, we solved the Sau-MenD structure and investigated conservation at a structural level. Like Mtb-MenD, the Sau-MenD structure revealed a tetrameric protein, consisting of a dimer of dimers with paired active sites. Consistent with the full occupancy of the symmetrical Ec- and Bs-MenD structures, and in contrast to the half-sites occupancy of the asymmetrical structures of Mtb-MenD, we did not observe asymmetry in the Sau-MenD cofactor-bound structure, and all four active sites were occupied with ThDP. There is limited conservation of the allosteric region; we identified only one of three key arginines of the Mtb-MenD arginine cage (Arg97) conserved in Sau-MenD (Arg98), with the other two (Mtb-MenD Arg277 and Arg303) being conservatively replaced by Lys283 and Lys309, respectively.

These differences in the Sau-MenD allosteric site cage were reflected in notably higher IC₅₀ values for DHNA, in the low micromolar range, compared with the low nanomolar range for the Mtb-MenD [31]. The presence of lysine residues in Sau-MenD at the equivalent positions to two key arginine residues in Mtb-MenD is likely to impact the strength of the stacking interactions with the aryl ring of the DHNA. Mutation of the three Sau-MenD allosteric cage candidates validated a significant role for two (Arg98 and Lys283) in the inhibition of SEPHCHC synthase activity by DHNA. The inhibitory potency of DHNA remained relatively unaffected by the mutation of Lys309, suggesting this residue plays a less significant role in DHNA binding and inhibition.

In Mtb-MenD, Arg97 (the residue with hydrogen bond and electrostatic interactions with the DHNA carboxylate) is the arginine cage residue whose mutation results in the largest change in DHNA response. It is also the most conserved residue in sequence alignments [31]. Our results validate the importance of this arginine (Arg98 in Sau-MenD) for allosteric regulation. Furthermore, they show that regulation, albeit at different potencies, can be conserved across bacterial species despite significant sequence divergence, likely reflecting their metabolic needs. However, it is not yet known how concentrations of DHNA vary across different bacterial species, or are modulated over time within a bacterial cell, or how DHNA is partitioned between the cytosol (where it is produced by MenI) and the membrane (where it is a substrate for MenA). By contrast to Mtb, which relies on oxidative phosphorylation for growth [57,58], substrate-level phosphorylation of fermentable sources allows Sau to grow slowly. In fact, respiration-deficient Sau, known as SCVs, have a slow-growth phenotype as a result of resorting to energy generation via fermentation. Therefore, Mtb and Sau may have different needs for the regulation of this pathway, requiring adaptations of the allosteric site.

Assuming DHNA adopts a similar binding pose in Sau-MenD, our overlays with the Mtb-MenD structure suggest that several residues in the Sau-MenD allosteric site (all three cage residues and residues 305–314) would need to undergo movement to accommodate DHNA. It is plausible that the initiation of DHNA binding facilitates a series of structural changes that help to achieve this. In the Sau-MenD structure, the Arg98 side chain forms hydrogen bonds to residues 308–309. Movement of the Arg98 side chain to interact with DHNA would free the region around residues 305–312 and enable the adoption of a more open conformation for these residues. There is precedence for the movement of this region in Mtb-MenD; the equivalent region (residues 300–308) has conformational differences between the DHNA-bound and DHNA-free forms. Rearrangement of this part of the allosteric site upon ligand binding is not unprecedented across other ThDP-dependent decarboxylase superfamily enzymes either. In fact, it can be important for the allosteric mechanism; pyruvate decarboxylase binds the positive allosteric regulator pyruvate in a similar structural location to the DHNA binding site in MenD, causing significant rearrangement of the allosteric site region and propagation of allosteric signal to the active site [59].

The exploitation of allosteric sites is an emerging strategy for inhibitor development, but it was not known whether compounds targeting the MenD allosteric site would inhibit the growth of whole cells in vivo. One advantage of allosteric sites is that they tend to bind relatively hydrophobic compounds, which are associated with cell permeability [31]. This is also the case for MenD, which binds hydrophilic molecules in its active site and the relatively hydrophobic DHNA in the allosteric site. Our results show that treatment of Sau bacteria with exogenous DHNA inhibits their growth. Furthermore, the addition of MK-4 to the culture medium of DHNA-treated cells results in growth rescue, suggesting that growth inhibition of the bacteria is caused by a lack of MK. While we cannot rule out that DHNA may also target other enzymes within the pathway, these data suggest that inhibition of cell growth may be attributable to the inhibition of the Sau-MenD enzyme. This is supported by recent findings that the addition of DHNA to culture medium increased the susceptibility of Sau to the antibiotic adjuvant cannabidiol, while MK-4 addition had the opposite effect and prevented cannabidiol-mediated growth impairment [60].

This work demonstrates that despite a lack of strict sequence conservation of the DHNA binding pocket between Sau and Mtb, feedback inhibition by DHNA is observed in Sau-MenD and the Sau MK biosynthesis pathway. Our results suggest MenD may be a suitable target for the development of anti-staphylococcal agents.

Data accessibility

Crystallographic data and structure files are available from the Protein Data Bank under the PDB code 7TIN. The other data from this work are given in the graphs and tables presented in the main manuscript and electronic supplementary material [61], with raw data available on request to the authors.

Authors' contributions

T.S.: conceptualization, data curation, formal analysis, investigation, methodology, project administration, supervision, validation, visualization, writing—original draft, writing—review and editing; N.A.T.H.: data curation, formal analysis, investigation, methodology, writing—review and editing; E.M.M.B.: data curation, validation, writing—review and editing; G.B.: funding acquisition, writing—review and editing; S.S.D.: data curation, validation, writing—review and editing; E.W.A.: investigation, writing—review and editing; J.T.: investigation; T.M.A.: data curation, methodology, visualization, writing—review and editing; W.J.: data curation, formal analysis, investigation, writing—review and editing; J.M.J.: conceptualization, data curation, formal analysis, funding acquisition, investigation, project administration, resources, supervision, validation, visualization, writing—original draft, writing—review and editing.

All authors gave final approval for publication and agreed to be held accountable for the work performed herein.

Conflict of interest declaration

We declare we have no competing interests.

Funding

This work was primarily supported by the Canterbury Medical Research Foundation (T.S. and J.M.J.), with additional funding support from a Marsden Fund from the Royal Society Te Apārangi (N.A.T.H., W.J., E.M.M.B., G.B., J.M.J.), the Maurice Wilkins Centre (MWC) for Molecular Biodiscovery (E.W.A., G.B., S.S.D., J.M.J., E.M.M.B.) and the Biomolecular Interaction Centre (J.T., T.M.A., J.M.J.). G.B. is supported by a Sir Charles Hercus Fellowship through the Health Research Council of New Zealand. This research required the use of the MX2 beamline at the Australian Synchrotron, part of the Australian Nuclear Science and Technology Organisation and made use of the Australian Cancer Research Foundation detector. Access to the Australian Synchrotron was supported by the New Zealand Synchrotron Group Ltd.