Clonal Diversification and Changes in Lipid Traits and Colony Morphology in Mycobacterium abscessus Clinical Isolates

Abstract

The smooth-to-rough colony morphology shift in Mycobacterium abscessus has been implicated in loss of glycopeptidolipid (GPL), increased pathogenicity, and clinical decline in cystic fibrosis (CF) patients. However, the evolutionary phenotypic and genetic changes remain obscure. Serial isolates from nine non-CF patients with persistent M. abscessus infection were characterized by colony morphology, lipid profile via thin-layer chromatography and matrix-assisted laser desorption ionization-time of flight mass spectrometry (MALDI-TOF MS), sequencing of eight genes in the GPL locus, and expression level of fadD23, a key gene involved in the biosynthesis of complex lipids. All 50 isolates were typed as M. abscessus subspecies abscessus and were clonally related within each patient. Rough isolates, all lacking GPL, predominated at later disease stages, some showing variation within rough morphology. While most (77%) rough isolates harbored detrimental mutations in mps1 and mps2, 13% displayed previously unreported mutations in mmpL4a and mmpS4, the latter yielding a putative GPL precursor. Two isolates showed no deleterious mutations in any of the eight genes sequenced. Mixed populations harboring different GPL locus mutations were detected in 5 patients, demonstrating clonal diversification, which was likely overlooked by conventional acid-fast bacillus (AFB) culture methods. Our work highlights applications of MALDI-TOF MS beyond identification, focusing on mycobacterial lipids relevant in virulence and adaptation. Later isolates displayed accumulation of triacylglycerol and reduced expression of fadD23, sometimes preceding rough colony onset. Our results indicate that clonal diversification and a shift in lipid metabolism, including the loss of GPL, occur during chronic lung infection with M. abscessus. GPL loss alone may not account for all traits associated with rough morphology.

FIG 1

Phylogenetic tree based on MLST (concatenated hsp65-secA-rpoB sequences). The clinical isolates in this study clustered with other M. abscessus genomes available in NCBI, apart from M. abscessus subsp. massiliense or M. abscessus subsp. bolletii. The serial isolates from each patient showed shared evolutionary lineage, demonstrating their close genetic relatedness.

FIG 2

Representative single-colony pictures of the serial isolates from patients 1, 2, 5, and 7 at a magnification of ×100, displayed chronologically from left to right. The isolates obtained on the same date are shown in one column (single-colony pictures of the isolates from patients 3, 4, 6, 8, and 9 are shown in Fig. S1 in the supplemental material). WT, wild type; S, smooth; R, rough; SNP, single nucleotide polymorphism.

FIG 3

A segment of the GPL locus at the 5′ end in the M. abscessus genome, as described by Ripoll et al. (9). The genes in bold at the top are the 8 genes that underwent targeted sequencing in this study. We found a detrimental mutation in 4 of the 8 genes (mmpS4, mmpL4a, mps1, and mps2 [*]) in 35 of the 39 rough clinical isolates included in the study. Most of these mutations were present in either in mps1 or mps2, but the 3 rough isolates from patient 7 had detrimental mutations in mmpL4a, while the 2 rough isolates from patient 2 had them in mmpS4.

FIG 4

(A) MALDI-TOF MS of the lipid extracts from patient 1 isolates. The MALDI-TOF MS spectra corresponding to GPL are present only in the smooth isolates. This indicates complete absence of GPL in the rough isolates. (B) The TLC results also agreed with the findings of MALDI-TOF MS. (C) MALDI-TOF MS and TLC of the isolates from patient 2. The rough isolates from this patient were found to have mutations in mmpS4. Note that mmpS4 mutants showed spectra in the molecular mass range where GPL is detected in MALDI-TOF MS, but the major peaks have atomic mass values higher than those of GPL (whose spectrum is shown together for comparison). The TLC of the mmps4 mutants did not show GPL bands. MALDI-TOF MS and TLC of the first and last isolates from patients 3, 4, 5, 6, 7, 8, and 9 are shown in Fig. S2 in the supplemental material. R, rough; S, smooth.

FIG 5

(A) The band corresponding to TAG becomes increasingly darker in the later clinical isolates, suggesting accumulation of TAG in M. abscessus as it gains persistence in the human lung. The solvent system used was hexane-diethyl ether-acetic acid at 70:30:1. Bands were visualized by spraying with 10% (wt/vol) CuSO₄ in 10% phosphoric acid, followed by charring at 110°C (20, 21). (B) MALDI-TOF MS of the lipid extract from the silica scraped off the band corresponding to TAG in the TLC run of 1-8(R). The peaks had molecular masses matching those of TAG in the Lipidomics Gateway database (http://www.lipidmaps.org) and were consistent with what has been reported in the literature on TAG in NTM species (30). R, rough; S, smooth.

FIG 6

fadD23 cDNA copy numbers converted from the RNA extracts of isolates from patients 1, 3, 5, and 7. The graphs display the qPCR results from two separate RNA extraction and cDNA conversion experiments for each isolate. A significant reduction in the copy number of fadD23 mRNA was observed in the later serial isolates. In patients 3 and 7, the reduced expression of fadD23 occurred even before the appearance of rough morphology. S, smooth; R, rough.