Cell surface lipid composition and hydrophobicity governs tuberculosis evolution and pathogenicity

The evolution of tubercle bacilli correlates closely with changes in cell envelope surface lipid composition (Donoghue et al. Diversity 2017, 9:46; Jankute et al. Scientific Reports 2017, 7:1315). Smooth, hydrophilic “Mycobacterium canettii” is the first recognisable member of the Mycobacterium tuberculosis complex, but it has reduced pathogenicity and poor aerosol transmission. In contrast, rough M. tuberculosis is very hydrophobic and readily spread in aerosols. Starting from hydrophilic surface lipids in environmental Mycobacterium kansasii, intermediate “M. canettii” adds hydrophobic lipids but retains overall cell hydrophilicity. Eliminating hydrophilic lipooligosaccharides (LOSs) and phenolic glycolipids (PGLs) from “M. canettii” leads to M. tuberculosis with a refined selection of hydrophobic lipids, namely phthiocerol dimycocerosates (PDIMs), pentaacyl trehaloses (PATs) and sulfoglycolipids (SGLs). The relative hydrophobicity of M. tuberculosis is double that of representatives of M. kansasii and “M. canettii”.

The above changes have implications both for the onset of tuberculosis and pinpointing evolutionary hosts. Tuberculosis has not been found in Homo sapiens during the Late Pleistocene, but megafauna are the most likely hosts; characteristic bone lesions have been validated by TB DNA amplification and lipid biomarkers in bison metacarpals up to 17,000 years old. Late Pleistocene enhanced TB hydrophobicity and aerosolisation may have produced megafaunal pandemics, with extinction of bison, mastodons and contemporary taxa. The oldest H. sapiens tuberculosis is from the “Fertile Crescent” back to 9-11ka BP at the start of the Holocene. Naïve humans arriving “Out of Africa” may have encountered newly virulent tubercle bacilli of megafaunal origin, recently refined through a distinct “bottleneck”.