Mycobacterial phenolic glycolipid virulence factor biosynthesis: mechanism and small-molecule inhibition of polyketide chain initiation

Abstract

Phenolic glycolipids (PGLs) are polyketide-derived virulence factors produced by Mycobacterium tuberculosis, M. leprae, and other mycobacterial pathogens. We have combined bioinformatic, genetic, biochemical, and chemical biology approaches to illuminate the mechanism of chain initiation required for assembly of the p-hydroxyphenyl-polyketide moiety of PGLs. Our studies have led to the identification of a stand-alone, didomain initiation module, FadD22, comprised of a p-hydroxybenzoic acid adenylation domain and an aroyl carrier protein domain. FadD22 forms an acyl-S-enzyme covalent intermediate in the p-hydroxyphenyl-polyketide chain assembly line. We also used this information to develop a small-molecule inhibitor of PGL biosynthesis. Overall, these studies provide insights into the biosynthesis of an important group of small-molecule mycobacterial virulence factors and support the feasibility of targeting PGL biosynthesis to develop new drugs to treat mycobacterial infections.

Figure 1. Representative structures of M. tuberculosis dimycocerosate esters (DIMs)

The p-hydroxybenzoic acid-derived phenol moiety of the PGLs is highlighted in red.

Figure 2. Proposed CoA-independent initiation of phenolphthiocerol biosynthesis parallels the mechanism of rifamycin B and (carboxy)mycobactin biosynthesis initiation

(A) Proposed CoA-independent biosynthesis initiation process required for assembly of the p-hydroxyphenyl-polyketide moiety of PGLs. (B) Mechanism for biosynthesis initiation of rifamycin B. (C) Mechanism for biosynthesis initiation of mycobactins and carboxymycobactins. A, adenylation domain; ArCP, aroyl carrier protein domain; pHBA, p-hydroxybenzoic acid; Sal, salicylic acid; AHBA, 3-amino-5-hydroxybenzoic acid; pHB-/Sal-/AHB-S-ArCP, p-hydroxybenzoyl-/salicyl/3-amino-5-hydroxybenzoyl-ArCP domain thioester. Non-covalent binding of acyl-AMP intermediates to the adenylation domains is indicated with a vertical dotted line. The phosphopantetheinyl group in the carrier domains is represented by a wavy line.

Figure 3. FadD22 is required for PGL production but not for PDIM or PNDIM production

(A) TLC analysis of DIMs from M. bovis, M. marinum, and M. kansasii showing incorporation of the label of [¹⁴C]-p-hydroxybenzoic acid (pHBA) in PGLs and [¹⁴C]-propionate (Prop) in PGLs, PDIMs, and PNDIMs. The profile of triplicate labeling experiments (lanes 1, 2, and 3) are shown in each panel. M. kansasii lacks PDIMs due to a (phenol)phthiodiolone ketoreductase deficiency. (B) Radio-TLC analysis demonstrating that fadD22 is required for PGL production but not for PDIM/PNDIM production. [¹⁴C]-Labeled apolar lipids of M. bovis strains wild-type (lane 1), ΔfadD22 (lane 2), ΔfadD22 with vector pJAM2 (lane 3), and ΔfadD22 with pJAM2-fadD22tb (pJAM2 expressing FadD22) (lane 4). Indicated TLC eluents were used for PGL and PDIM/PNDIM visualization as reported (see Experimental Procedures). PE, petroleum ether; Et₂O, diethyl ether. PGLs remain at the origin with PE/Et₂O (9:1 v/v) eluent. PDIMs/PNDIMs run with the eluent front in CHCl₃/MeOH (95:5 v/v).

Figure 4. Formation of pHB-AMP by FadD22 and FadD22(S576A)

(A) TLC analysis of radiolabeled pHB-AMP formation. Enzymes were included at 5 μM in the reactions. No product was detected in reaction without enzyme (not shown). The mass spectrum of the pHB-AMP product is shown. (B) Influence of CoA on pHB-AMP formation. Substrates and enzymes (5 μM) included in the reactions are indicated. BZLRp, R. palustris benzoyl-CoA ligase. (C) Time course for pHB-AMP formation by FadD22 (2 μM) and FadD22(S576A) (2 μM). (D) FadD22-pHB-AMP association determined by size exclusion chromatography. FadD22 was included at 15 μM in the reaction. Graphs show means of triplicate reactions with SEM. Only one of triplicate reactions is shown in the representative TLC image below each graph in A and B. pHB-AMP, black bars; pHB-CoA, white bars.

Figure 5. FadD22 is phosphopantetheinylated and autoacylated with the pHBA, whereas FadD22(S576A) is not

(A) Time course for in vitro phosphopantetheinylation of wild-type FadD22 (1 μM) and FadD22(S576A) (1 μM). (B) Auto-loading competence of FadD22 and FadD22(S576A) expressed alone or together with Sfp. Proteins (10 μM) in acylation reactions with the indicated compositions were bound to the membrane under denaturing conditions for analysis of covalently-bound radiolabel in the proteins. Only holo-FadD22 was competent for autoacylation (upper, right quadrant). Triplicate reactions were spotted on the membrane. (C) pHB-AMP formation competence of FadD22 (10 μM) and FadD22(S576A) (10 μM) expressed alone or together with Sfp. TLC excerpt image shows formed pHB-AMP (or lack thereof) in reactions of the indicated composition. (D) Time course for holo FadD22 autoacylation. The concentration of the pHB-S-FadD22 formed was normalized to the concentration of FadD22 (5 μM) in the reaction ([pHB-S-FadD22]/[FadD22]). Means of triplicate reactions with SEM are shown.

Figure 6. pHB-AMS inhibits FadD22-catalyzed pHBA adenylation and autoacylation

(A) p-Hydroxybenzoyl-AMP (1) (pHB-AMP) intermediate and its mimic, 5′-O-(N-[4-hydroxybenzoyl]sulfamoyl)adenosine (2) (pHB-AMS). (B) Dose-response for inhibition of pHB-AMP formation plotted with fractional velocity as a function of pHB-AMS concentration. (C) Dose-response for inhibition of pHB-S-FadD22 formation plotted with fractional velocity as a function of pHB-AMS concentration. v_i and v_c, are velocities in inhibitor- and DMSO-containing reactions, respectively. Data points are means of triplicates reactions.

Figure 7. pHB-AMS inhibits PGL production

(A) Inhibition of PGL production by pHB-AMS. The dose-response plots show the effect of pHB-AMS on production of PGLs. PGL_T/PGL_C, ratio of PGL production in inhibitor-treated (PGL_T) to DMSO-treated (PGL_C) cultures. Values represent means from triplicate cultures with SEM. The maximum fold changes induced by pHB-AMS on PGL, PDIM, and PNDIM production are depicted in the bar graphs. No PDIM fold-change is shown for M. kansasii, which is PDIM deficient. (B) Reduction of the PGL-production inhibitory efficacy of pHB-AMS by multi-copy suppressor effect. pHB-AMS was tested at the concentration shown against M. kansasii (Mk) and Mk transformed with the indicated plasmids. pCP0, vector; pCP0-FadD22tb, pCP0 expressing M. tuberculosis FadD22; pCP0-FadD22tbAd, pCP0 expressing M. tuberculosis FadD22 adenylation domain; pCP0-FadD22ml, pCP0 expressing M. leprae FadD22. Data represent means of triplicate cultures with SEM. (C) Effect of pHB-AMS on growth of M. tuberculosis (Mt, red circles), M. bovis (Mb, blue squares), M. marinum (Mm, green circles), and M. kansasii (Mk, orange squares). The effect was evaluated on cultures started with the indicated inoculum levels. OD_T/OD_C, ratio of OD_580nm of inhibitor-treated (OD_T) to OD_580nm of DMSO-treated (OD_C) cultures. Means of triplicate cultures with SEM are shown.