To characterize the function of Rv2869c in mycobacteria, we deleted this gene from the chromosomes of M. bovis BCG and M. tuberculosis Erdman by specialized transduction ¹⁸. Figure 1b shows the genomic location of Rv2869c between dxr and gcpE, two genes in the non-mevalonate pathway of isoprenoid biosynthesis ^19,20. Southern blotting (Figure 1c) confirmed successful replacement of Rv2869c with a Hygromycin resistance cassette in both BCG and M. tuberculosis, demonstrating that, in contrast to YaeL in E. coli, Rv2869c is not an essential gene for M. tuberculosis viability.

When grown on solid media, the Rv2869c null mutant displayed altered colony morphology. In pathogenic mycobacteria, the colonial and microscopic morphology of cording has long been associated with virulence and is dependent on multiple cell envelope lipids ^2,5. Both the BCG and Mtb ΔRv2869c strains lacked cording, as measured by colonial morphology (Figure 2A) and by microscopic examination of Auramine-Rhodamine stained bacilli (Figure 2B). Wild type cording was restored to the ΔRv2869c strain by a wild type copy of Rv2869c, confirming that loss of Rv2869c function was responsible for the noncording phenotype (Figure 2). Alanine substitution mutations in the conserved HExxH (H21A) and FDG (D341A) motifs, which are required for proteolytic activity of RIP proteases^10,13,16, did not restore wild type colony morphology (Fig 2C). These results demonstrate that the proteolytic activity of Rv2869c regulates cell envelope composition in pathogenic mycobacteria. These results suggested that RIP is a conserved mechanism of membrane composition control in prokaryotes.

The SREBP pathway regulates multiple pathways of lipid biosynthesis, including cholesterol and fatty acids²¹. To directly ask whether Rv2869c regulates lipid composition in Mtb, we analyzed the extractable and esterified mycolic acids of the cell envelope of wild type cells, ΔRv2869c cells, and complemented mutant cells by quantitative thin layer chromatography (TLC) of ¹⁴C acetate labeled cells in the presence or absence of detergent in the culture medium. In the esterified mycolic acids, we observed similar quantities of the three major mycolic acids in all strains regardless of culture conditions (figure 3A, top panel). In contrast, while wild type cells maintained extractable mycolic acid synthesis after shift to detergent free media (Figure 3A, bottom panel), ΔRv2869c cells downregulated alpha mycolate synthesis by 4.6fold, methoxymycolates by 3.5 fold, and ketomycolates by 2.3 fold. In addition, ΔRv2869 cells upregulated synthesis of a slow migrating lipid near the origin of the TLC plate by 6 fold compared to no Tween conditions. This lipid was absent from wild type or complemented mutant extractable lipids (Figure 3A, bottom panel). These results indicated that Rv2869c regulates the composition of extractable mycolic acids in the cell envelope in response to changes in membrane fluidity but has no role in the covalently esterified mycolates of the cell wall. To examine whether Rv2869c also regulates other classes of cell envelope lipids, we examined the composition of phosphatidylinositol mannoside (PIM). Two dimensional TLC of chloroform-methanol extracts revealed typical pattern of PIM2 and PIM6 with differing degrees of acylation ^22,23. We found no difference between mutant and wild type cells in PIM2 species (Figure S1, spots 1 and 2) but observed a two fold decrease in the abundance of a PIM6 subspecies (spot 4) in mutant but not wild type or complemented mutant cells.

To ask whether Rv2869c controls cell envelope composition through transcriptional regulation, we compared the transcriptional profiles of the wild type and ΔRv2869c strains in the presence or absence of detergent using an M. tuberculosis whole genome oligonucleotide microarray. We found that Rv2869c was both a positive and negative transcriptional regulator of multiple lipid biosynthetic and lipid degrading genes. Consistent with the lipid analysis, Rv2869c mediated complex transcriptional regulation of mycolic acid biosynthetic genes in response to detergent, including pks13, kasB, accD6, fatty acid synthase, fabG1, fabG2, and two lipid desaturases (figure3B and supplementary data). Cluster analysis grouped the mycolic acid biosynthetic genes kasB, pks13, and accD6 together but indicated that other components of the pathway such as fabG1/fabG2 were divergently regulated by Rv2869c (Figure 3, S2, supplementary tables). Other lipid biosynthetic genes were positively regulated by Rv2869c, including multiple genes involved in PDIM synthesis (mas, ppsA, drrB). Absence of Rv2869c resulted in strong overexpression of a lipase (lipP) and an epoxide hydrolase (ephC) and strong underexpression of rpfC (Tables S1,S2), three genes putatively involved in lipid or cell wall degradation ²⁴. Taken together, these data demonstrate that the lipid perturbations in the cell envelope of the Rv2869c mutant resulted from altered transcriptional control of diverse lipid anabolic and catabolic pathways.

An emerging model in Mtb pathogenesis is that the extractable lipids of the cell envelope act as direct effectors of pathogenesis either to modulate host immune responses or alter intracellular trafficking ^2,3,7,25-29. Some of these lipids are regulated by Rv2869c, suggesting that RIP might be a proximal regulator of critical cell envelope pathogenesis determinants. To test this idea, we examined the Rv2869c mutant in the mouse model of aerosol Mtb infection. Despite identical inocula one day after infection, the Rv2869c mutant was impaired for bacterial growth during the acute phase of infection such that mutant bacterial titers in the lung after three weeks of infection were approximately 100 fold lower than wild type (Figure 4a). In mice, wild type Mtb persists at constant titer in the lung for the life of the animal. Surprisingly, we found that Rv2869c was also required for this persistence phase of the infection. ΔRv2869c organisms were progressively eliminated from the lung such that by 22 weeks after infection, the number of bacilli in the lung infected with Rv2869c mutant was 10,000 fold lower than wild type. This in vivo phenotype was due to loss of Rv2869c function and not a polar effect on neighboring genes because the wild type growth and persistence phenotypes were restored by a plasmid expressing Rv2869c (Figure 4b). In the liver, Rv2869c organisms were completely cleared by 22 weeks of infection (Figure S3). Gross and microscopic examination of infected lungs revealed a dramatic attenuation of granulomatous histopathology in the mutant infected lungs (Figure 4c-d).

While S2P mediated RIP of SREBPs is a well described mechanism controlling eukaryotic membrane composition, S2P family members in prokaryotes characterized to date regulate stress responses^15,16,30, sporulation¹³, and cell polarity¹⁷. The present study implicates RIP in controlling the lipid composition of the complex mycobacterial cell envelope. Consistent with the role of other RIP proteases in controlling transcription, the observed changes in cell envelope lipids were associated with transcriptional dysregulation of lipid biosynthetic and degradative genes. Further mechanistic studies will determine whether Rv2869c directly cleaves membrane bound transcriptional regulators or controls membrane composition through cleavage of other membrane proteins that control membrane composition.

SI Guide

Supp Figs S1-S

Supp tables S1