Multi-omics Investigation into the Mechanism of Action of an Anti-tubercular Fatty Acid Analogue

Abstract

The mechanism of action (MoA) of a clickable fatty acid analogue 8-(2-cyclobuten-1-yl)octanoic acid (DA-CB) has been investigated for the first time. Proteomics, metabolomics, and lipidomics were combined with a network analysis to investigate the MoA of DA-CB against Mycobacterium smegmatis (Msm). The metabolomics results showed that DA-CB has a general MoA related to that of ethionamide (ETH), a mycolic acid inhibitor that targets enoyl-ACP reductase (InhA), but DA-CB likely inhibits a step downstream from InhA. Our combined multi-omics approach showed that DA-CB appears to disrupt the pathway leading to the biosynthesis of mycolic acids, an essential mycobacterial fatty acid for both Msm and Mycobacterium tuberculosis (Mtb). DA-CB decreased keto-meromycolic acid biosynthesis. This intermediate is essential in the formation of mature mycolic acid, which is a key component of the mycobacterial cell wall in a process that is catalyzed by the essential polyketide synthase Pks13 and the associated ligase FadD32. The multi-omics analysis revealed further collateral alterations in bacterial metabolism, including the overproduction of shorter carbon chain hydroxy fatty acids and branched chain fatty acids, alterations in pyrimidine metabolism, and a predominate downregulation of proteins involved in fatty acid biosynthesis. Overall, the results with DA-CB suggest the exploration of this and related compounds as a new class of tuberculosis (TB) therapeutics. Furthermore, the clickable nature of DA-CB may be leveraged to trace the cellular fate of the modified fatty acid or any derived metabolite or biosynthetic intermediate.

Figure 1.

(A) Principal component analysis (PCA) scores plot generated from 1D ¹H NMR data set using MVAPACK with 3 principal components. The PCA plot exhibits group differentiation with R²=0.755 and Q²=0.640. Untargeted metabolomics of Msm cells (purple), cells treated with isoniazid (INH, red), D-cycloserine (DCS, dark green), ethionamide (ETH, green), ciprofloxacin (magenta), streptomycin (blue) and DA-CB (yellow). Ellipsoids represent a 95% confidence limit of the normal distribution of each cluster. (B) Dendrogram generated from the PCA score plots with each Mahalanobis distance p-value represented between the nodes shows DA-CB is closely clustered with ethionamide.

Figure 2.

(A) Principal component analysis (R²=0.535, Q²=0.391) with 5 principal components and (B) Orthogonal projection to latent structures discriminant analysis (OPLS-DA, R²=0.990, Q²=0.959, p-value < 0.03) models generated from the LC-MS metabolomics data set. Msm cells were treated with either 400 μM of DA-CB (green) or 10 μL of DMSO (Control, red). Quality control (QC) pooled samples combine 45 μL from each DA-CB and Control sample. Ellipses represent a 95% confidence limit of the normal distribution of each cluster.

Figure 3.

(A) An enrichment analysis using the 26 metabolites with statistically significant changes in Msm cells following a DA-CB treatment. Representative box plots for significantly altered metabolites corresponding to the following enriched pathways: (B) amino acids, (C) dipeptides, (D) purine nucleosides, (E) purine nucleotides, (F) acyl-CoA and (G) pyrimidine nucleosides. DA-CB indicates DA-CB treated Msm cells and Control indicates only DMSO treated Msm cells.

Figure 4.

(A) Principal component analysis (PCA) (R²=0.745, Q²=0.644) with 5 principal components and (B) Orthogonal projection to latent structures discriminant analysis (OPLS-DA, R²=0.992, Q²= 0.980, p-value < 0.01) models generated from the positive ionization mode LC-MS lipidomics data set. (C) PCA (R²=0.726, Q²=0.613) and (D) OPLS-DA (R²=0.996, Q²= 0.984, p-value < 0.01) models generated from the negative ionization mode LC-MS lipidomics data set. Msm cells were treated with either 400 μM of DA-CB (green) or 10 μL of DMSO (Control, red). Quality control (QC) pooled samples combine 45 μL from each DA-CB and Control sample. Ellipses represent a 95% confidence limit of the normal distribution of each cluster.

Figure 5.

(A) An enrichment analysis using the 48 lipids with statistically significant changes in Msm cells following a DA-CB treatment. (B-G) Representative box plots for significantly altered lipids corresponding to the following enriched pathways: (B) Wax monoesters (C) branched chain fatty acids, (D) mycolic acids, and (E) hydroxy fatty acids, (F) saturated fatty acids, (G) unsaturated fatty acids. DA-CB indicates DA-CB treated Msm cells and Control indicates only DMSO treated Msm cells.

Figure 6.

(A) A Cytoscape (https://cytoscape.org/) network generated from the 123 differentially expressed proteins in the proteomics data set and using the Msm protein database. Nodes using a square symbol indicate altered pathways from the Kyoto Encyclopedia of Genes and Genomes (KEGG) database, while nodes with a hexagon symbol indicate altered biological processes (BP). A summary of the (B) KEGG and (C) BP terms that were significantly altered according to the network analysis of the proteomics data set.

Figure 7.

A schematic of an integrated network of the multi-omics data sets summarizing the consensus changes in the Msm lipidome (green), metabolome (blue), and proteome (red) resulting from a DA-CB treatment. An up arrow indicates an increase in DA-CB treated cells and a down arrow indicates a decrease. The relative thickness of the arrow indicates the range of the fold change for the lipid, metabolite, or protein (insert). A molecule colored black was not detected or altered by DA-CB but was included to highlight important nodes or to connect to other observed nodes.

Scheme 1.

Workflow of multi-omics sample preparation and data acquisition using Reversed phase liquid chromatography electrospray ionization high resolution-mass spectrometry (RPLC ESI HR-MS^E ) from DA-CB treated Msm. Created with BioRender.com.