MmpL8MAB controls Mycobacterium abscessus virulence and production of a previously unknown glycolipid family

Abstract

Mycobacterium abscessus is a peculiar rapid-growing Mycobacterium (RGM) capable of surviving within eukaryotic cells thanks to an arsenal of virulence genes also found in slow-growing mycobacteria (SGM), such as Mycobacterium tuberculosis A screen based on the intracellular survival in amoebae and macrophages (MΦ) of an M. abscessus transposon mutant library revealed the important role of MAB_0855, a yet uncharacterized Mycobacterial membrane protein Large (MmpL). Large-scale comparisons with SGM and RGM genomes uncovered MmpL12 proteins as putative orthologs of MAB_0855 and a locus-scale synteny between the MAB_0855 and Mycobacterium chelonae mmpL8 loci. A KO mutant of the MAB_0855 gene, designated herein as mmpL8_MAB , had impaired adhesion to MΦ and displayed a decreased intracellular viability. Despite retaining the ability to block phagosomal acidification, like the WT strain, the mmpL8_MAB mutant was delayed in damaging the phagosomal membrane and in making contact with the cytosol. Virulence attenuation of the mutant was confirmed in vivo by impaired zebrafish killing and a diminished propensity to induce granuloma formation. The previously shown role of MmpL in lipid transport prompted us to investigate the potential lipid substrates of MmpL8_MAB Systematic lipid analysis revealed that MmpL8_MAB was required for the proper expression of a glycolipid entity, a glycosyl diacylated nonadecyl diol (GDND) alcohol comprising different combinations of oleic and stearic acids. This study shows the importance of MmpL8_MAB in modifying interactions between the bacteria and phagocytic cells and in the production of a previously unknown glycolipid family.

Keywords: GDND; MMPL; Mycobacterium; abscessus; glycolipid.

Fig. 1.

Identification of a putative MmpL required for intracellular replication in both amoebae and MΦ from a Tn library screen. Tn mutants impaired in their intracellular survival compared with the WT strain GI in amoebae and MΦ. Amoebae (A. castellanii) and murine MΦ were infected at an MOI of 10 and lysed at different dpi (2 and 3 dpi, respectively) before cfu determination. The Tn insertion site was cloned to identify the disrupted genes. Identified genes were categorized with the COG classification. Mutants for which the Tn insertion occurred in a gene implicated in either lipid metabolism/transport or related processes (COG M and I, respectively) and 55F6 uncharacterized (COG R) mutant putatively implicated in lipid transport and metabolism are depicted.

Fig. 2.

MAB_0855 transporter is specific to the MCC. (A) MAB_0855 phylogenetic position among M. tuberculosis MmpL transporters. In this study, BH scores were calculated with the Opscan program. A BBH between two proteins corresponds to a putative ortholog. M. abscessus and M. tuberculosis MmpL phylogeny was inferred by Maximum Likelihood using the PHYML software and the aLRT statistic from the Multiple Sequence Alignment of M. abscessus and M. tuberculosis performed with Clustal. The aLRT statistical values over 0.9 are indicated. When a BBH was found between two proteins, the BBH score is indicated in brackets. (Scale bar: 0.5 estimated amino acid substitutions.) (B) Conservation of MAB_0855 in mycobacteria. MAB_0855 and MAB_0937c BBHs with MmpL proteins outside M. abscessus strains are represented by check boxes (Left). BBHs of these foreign MmpLs with M. tuberculosis MmpLs, if any, are indicated (Center). MAB_0855 and MAB_0937c hit scores, which give a clue on MAB_0855 and MAB_0937c level of conservation among mycobacteria, are also depicted by a range of colors (Right). The lowest scores are depicted in blue; the highest scores are depicted in orange. (C) Conservation of MAB_0855 locus among pathogenic and nonpathogenic mycobacteria. The chromosomal loci encompassing the relevant genes depicted in the figure were built with the GenoPlotR package. Genes for which no BBHs were found are depicted in gray. Genes exhibiting BBH toward MAB_0855 locus genes and the two fad genes upstream of the locus are depicted in color. A gray band links other BBH genes (in white).

Fig. 3.

Elucidation of the mmpL8_MAB role in M. abscessus virulence. Intracellular growth was assessed by cfu counting throughout the infection. Murine MΦ were infected at 10 MOI with the SΔmmpL8_MAB or with the complemented strain carrying the MYCMA_0439 or the mmpL8_MAB gene (A), the M. tuberculosis-closest homologs to M. abscessus mmpL8 (mmpL8_MTB, mmpL10_MTB, mmpL12_MTB; B), or mmpL10_MTB (C); cfu histograms with error bars represent means ± SD calculated using data from three independent experiments. Differences between means were analyzed by two-way ANOVA and the Tukey posttest, allowing multiple comparisons. ns, Nonsignificant. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.

Fig. 4.

MmpL8_MAB plays a crucial role in M. abscessus interactions with cells by modulating M. abscessus entry and its survival inside the cells. (A) Adhesion to murine MΦ assessment by flow cytometry. MΦ were cooled down to avoid bacterial internalization and then infected with CFDA-SE–stained bacteria at an MOI of 100 for 1 h at 4 °C with gentle agitation to favor bacteria-to-cell contact. Nonadherent bacteria were well separated from intact cells (Left), and a substantial shift was observed for the cell population with adherent bacteria (Right; infected cells). The bacterial adhesion index is given that represents the mean fluorescence intensity (MFI) of each population with adherent bacteria reported to the MFI of the cells infected with the WT strain. FSC represents forward scatter and SSC represents side scatter. NI represents noninfected cells. (B) Phagosome acidification test. MΦ intraphagosomal pH after infection at an MOI of 10 was assessed by the mean FITC fluorescence measurement (485 nm) combined with mCherry fluorescence (565 nm) of double-stained bacteria. The pH along the infection is extrapolated from a pH standard curve. HK corresponds to heat killed bacteria. (C) Intracellular behavior in human THP-1. Cells were infected at an MOI of 10. The ability of bacteria to establish a contact with the cell cytosol was assessed by FRET at 20 and 30 hpi. (D) Kinetics of IL-1β production during THP-1 MΦ infection. The cellular response to infection with M. abscessus WT, ΔmmpL8_MAB, and complemented strains was assessed by quantifying IL-1β secretion. Histograms with error bars represent means ± SD calculated using data from three independent experiments. Differences between means were analyzed by ANOVA and the Tukey posttest, allowing multiple comparisons to be performed. ns, Nonsignificant. *P < 0.05; ***P < 0.001; ****P < 0.0001.

Fig. 5.

The mmpL8 gene plays a role in M. abscessus in vivo survival and virulence. (A) Survival curves of embryos. Embryos were injected into the bloodstream with 200 cfu of tdTomato-expressing M. abscessus R (parental strain), RΔmmpL8_MAB mutant, or the complemented strain R∆mmpL8::mmpL8_MAB (n = 20). Data are representative of two independent experiments. (B) Kinetics of granuloma formation. Embryos were i.v. infected with 200 cfu of tdTomato-expressing M. abscessus R, M. abscessus S, RΔmmpL8_MAB, SΔmmpL8_MAB mutants, S∆mmpL8::mmpL8_MAB, or R∆mmpL8::mmpL8_MAB. Granulomas were then counted at 5 dpi. Histograms represent means ± SEM values calculated from two independent experiments. The statistical test used was the Fisher’s exact test (Left), and means were analyzed by ANOVA and the Tukey posttest (Right). *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.

Fig. 6.

Production of a polar lipid is altered by deletion of mmpL8_MAB. (A and B) TLC showing a diminished amount of a set of polar lipids in strains lacking mmpL8_MAB. Polar lipids extracted from freeze-dried bacteria (A, S strain; B, R strain) were separated by TLC using chloroform:methanol:water [90:10:1 (vol/vol/vol)] as the solvent system. TLCs were subsequently stained with orcinol [0.1% (wt/vol) in a 20% (vol/vol) H₂SO₄ solution] and charred. The presence of a particular lipid (the lipid with structure that was determined in this study; indicated with brackets) is decreased in the S Tn MYCMA_0439 mutant and KO ΔmmpL8_MAB mutant, which is particularly evident in the M. abscessus R background lacking GPL. (C) TLC of lipids (named Pa3 lipid fraction) from the polar lipid fraction of the R M. abscessus strain. The presence of the particular lipid is again indicated with the bracket. (D) MALDI-MS and MSⁿ analyses of purified lipids by successive fragmentations of relevant fragment ions. (Row 1) MALDI-MS spectrum of purified lipid. (Row 2) MS² fragmentation spectrum of the parent ion at m/z 987. (Row 3) MS³ fragmentation spectrum of the fragment ion [M-Hex + H]⁺ at m/z 825. (Row 4) MS⁴ fragmentation spectrum of the fragment ion [M-Hex-C18 + H]⁺ at m/z 541. On top of the MALDI-MS spectra, the structure of a major component at m/z 987 based on MSⁿ, GC-MS, and NMR data is shown. (E) Multidimension NMR spectroscopy analysis of the purified lipid. (Top) ¹H/¹H TOCSY, (Middle) ¹H/¹³C HSQC, and (Bottom) ¹H/¹³C HSQC-TOCSY NMR spectra showing the spin system of the glycan moiety of the purified lipid.