Mycobacterial cell envelope components have been a major focus of research due to their unique features that confer intrinsic resistance to antibiotics and chemicals apart from serving as a low-permeability barrier. The complex lipids secreted by Mycobacteria are known to evoke/repress host-immune response and thus contribute to its pathogenicity. This study focuses on the comparative genomics of the biosynthetic machinery of cell wall components across 21-mycobacterial genomes available in GenBank release 179.0. An insight into survival in varied environments could be attributed to its variation in the biosynthetic machinery. Gene-specific motifs like ‘DLLAQPTPAW’ of ufaA1 gene, novel functional linkages such as involvement of Rv0227c in mycolate biosynthesis; Rv2613c in LAM biosynthesis and Rv1209 in arabinogalactan peptidoglycan biosynthesis were detected in this study. These predictions correlate well with the available mutant and coexpression data from TBDB. It also helped to arrive at a minimal functional gene set for these biosynthetic pathways that complements findings using TraSH.
Citation: Banerjee R, Vats P, Dahale S, Kasibhatla SM, Joshi R (2011) Comparative Genomics of Cell Envelope Components in Mycobacteria. PLoS ONE 6(5): e19280. https://doi.org/10.1371/journal.pone.0019280
Editor: Olivier Neyrolles, Institut de Pharmacologie et de Biologie Structurale, France
Received: January 27, 2011; Accepted: March 25, 2011; Published: May 6, 2011
Copyright: © 2011 Banerjee et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: Bioinformatics Division, Depatment of Information Technology, Government of India. url: http://www.mit.gov.in/content/bio-informatics. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
The genus Mycobacterium includes bacteria known to cause dreadful diseases like tuberculosis, leprosy and skin ulcers . One of the characteristic features of the members of this genus is the presence of a low-permeability cell envelope with high proportion of complex lipids that is organized into three superposed compartments viz., the plasma membrane, the cell wall skeleton and the capsule. Cell envelope components include mycolic acids, arabinogalactan (AG), lipoproteins, lipomannan (LM) and lipoarabinomannan (LAM). The pathogenicity and survival of Mycobacterium species in varied environments has been attributed to the variation in structural components of cell wall complex  and hence the variation in the biosynthetic machinery. The mycolic acid-arabinogalactan-peptidoglycan polymer is arranged to form a hydrophobic layer with other lipids , . A variety of unique lipids, such as LAM, trehalose dimycolate and phthiocerol dimycocerate, anchor non-covalently with the cell membrane and appear to play an important role in virulence . The enzymatic pathways that synthesize M. tuberculosis cell envelope lipids are the target of presently available antituberculosis antimicrobials and may be candidates for future antibiotic development.
In this study we have carried out comparative genomics of the biosynthetic machinery of cell wall components amongst 21-mycobacterial genomes using metabolic pathway context, sequence similarity tools and phylogenetic profiling. Phylogenetic profiling predicts functionally linked genes i.e., genes that are a part of a same biological process or cellular system based on the presence or absence of a protein in a set of reference genomes , . Prediction of functional linkages helps in the annotation of uncharacterized genes thereby reducing the gap between rate of genome sequencing and annotation. Previous studies have established the fact that the conserved co-evolution patterns of gene-pairs across different genomes as suitable indicators of functionally linked genes , . The predicted functional linked genes obtained by our studies were mapped with the co-expression data wherever available.
Comparative genomics of the mycobacterial genomes with a focus on genes involved in biosynthesis of cell envelope components was carried out using methodologies viz., sequence similarity, metabolic pathway reconstruction and phylogenetic profiling. Similarity search was carried out using MPI version of SSEARCH program available in FASTA3 package version 34  with parameters E-value: e−20, %identity: 50 and %query coverage: 80. The query dataset used for the analysis is the proteome of M. tuberculosis H37Rv.
Multiple sequence alignment of proteins was carried out using parallel implementation of ClustalW-MPI version 0.13 . Phylogenetic trees were reconstructed using parsimony (protpars) available in Phylip package version 3.67 . Prior to phylogenetic reconstruction, to assess the statistical significance of the topology obtained, the data was bootstrapped to generate 100 data sets using seqboot program of Phylip. The trees were visualized using MEGA version 4 .
Metabolic pathway reconstruction was carried out using Pathway Tools version 14.0  with MetaCyc version 14.1  as a reference database. As the curated genbank formatted file is the input for Pathway Tools, the genomes were curated with respect to functional information like ‘Enzyme Commission Number’ using the tool EFICAz2 version 13  and UniProt database . Metabolic pathways like LAM biosynthesis are unavailable in MetaCyc and were incorporated manually by mining of literature and using tools such as MarvinSketch [http://www.chemaxon.com/products/marvin/marvinsketch/] and ChemSpider [http://www.chemspider.com/] to add chemical structures of the metabolites involved. Pathway holes were filled using the ‘Power User’ mode with a probability cutoff of 0.9. Functionally linked genes were predicted using the methodology of phylogenetic profiling. The strength of this methodology lies in the fact that it takes into account the entire proteome and aids in the annotation of hypothetical genes that in-turn are capable of filling the pathway holes. Similarity profile matrix of all pairwise combination of genes was determined using SSEARCH, which implements the highly accurate and sensitive approach of dynamic programming for database similarity searching. The E-values obtained via SSEARCH were normalized  such that it enables capture of sequence divergence as well as generates continuous variables which are amenable to rigorous statistical treatment such as estimating mutual information content using B-spline function. Profile matrices were generated for both real as well as random datasets. Randomization aids in the calculation of ‘relationship-strength’ between gene-pairs that has occurred by chance and was implemented using rand function available in MATLAB [http://www.mathworks.com/]. Functionally linked genes were inferred based on calculation of mutual information content using mis_calc program , Pearson correlation coefficient and Hamming distance (using an in-house developed perl script). A plot of the above 3 parameters using both real and random data helped to generate the cutoff criteria for analyses of significant linkages (Figure 1 & Figure 2). Functionally linked genes in this study are defined as the genes that satisfy thecutoff criteria of 0.9 for mutual information content and 0.8 for Pearson correlation coefficient as these values differentiated the real from random data. Such stringent criteria were used as the comparison is between species belonging to the same genus viz., Mycobacterium. Hypothetical genes belonging to MTBH37Rv that exhibit conserved coevolution patterns in terms of similar profile to a well-characterized gene belonging to cell envelope biogenesis are analyzed in the present study. Genes displaying > = 50 functional linkages are predicted as ‘network-hubs’. The predicted functional linkages obtained were further substantiated by mapping withe co-expression data retrieved from TBDB  along with knockout mutants obtained via TraSH analysis – wherever available. For data management and efficient retrieval of ∼7700000 records, MySQL was used as DBMS. Anvaya, an in-house developed workflow environment that includes pre-defined workflows for ortholog identification, motif detection, phylogenetic reconstruction and phylogenetic profiling was used to perform all the above analysis.
Results and Discussion
Comparative genomics of pathogenic and non-pathogenic Mycobacteria has played an instrumental role in unraveling many underlying factors responsible for virulence and host-specificity . Detailed analysis of the biosynthetic machinery of mycolic acid, arabinogalactan, lipomannan/lipoarabinomannan, phthiocerol dimycocerate and lipoproteins is provided.
Mycolic acids are alpha-alkyl, beta-hydroxy fatty acids and are the signature lipids of the hydrophobic mycobacterial cell wall . The composition and quantities of mycolic acids are known to affect the virulence, growth rate, colony morphology and permeability of Mycobacteria –. Cells with reduced mycolate content show a higher permeability for substance uptake into the cells or excreted into the culture medium . The biosynthesis pathway involves 5 different steps viz., production of malonyl CoA, fatty acid initiation (I, II and III) and fatty acid elongation (FAS I and FAS II) followed by the actual biosynthesis of mycolates. Altogether, 46 genes are known to participate in a coordinated manner to generate mycolates in Mycobacteriaceae members. Comparative analysis across the 21-mycobacterial genomes revealed that the biosynthetic pathway is conserved across all the species with differences arising due to redundant genes in different species.
fabH gene is a pivotal link between Fatty acid biosynthesis I and II pathways . It elongates acyl-CoA primers derived from FAS-I to form thioesters through condensation with malonyl-ACP. Current work revealed that fabH is absent in M. leprae strains and the domains are truncated either at C/N termini in MMCS, MJLS, MKMS, MGlPYR and MSgMC2 and hence its functionality may be affected in these organisms. However, few studies report that the mycolate biosynthesis is not hindered, hence suggesting that there may exist alternate genes or pathways, which circumvent this step .
cmaA2, mmaA1 and umaA
Cyclopropanation of mycolic acids is one of the distinguishing features of pathogenic Mycobacteria suggesting that this modification may be associated with an increase in oxidative stress experienced by the slow-growing species, and is catalysed by a family of S-adenosyl methionine dependent methyltransferases . Even though the enzymes of this family share a conserved fold, they display high specificity for cis/trans cyclopropane ring formation in proximal/distal ends. These genes share a 50 to 70% identity between them; hence a lot of ambiguity arises in distinguishing these genes using sequence alignment tools alone. In order to detect sequence-motifs responsible for the observed specificity, we retrieved all the methyl transferase genes (totaling 170) from the 21 Mycobacteria and subjected them to multiple sequence alignment (MSA).
The MSA (Figure 3) clearly depicted that ‘ADGAGDA’ motif is unique to cmaA2 (RefSeq ID: NP_215017), which encodes the enzyme responsible for trans-cyclopropanation at the proximal end of meromycolate chain as reported earlier . This hydrophilic stretch is surface accessible inspite of being adjacent to binding site residues (which are buried in the hydrophobic core) and forms a loop away from the active site. Recent reports suggest that cmaA2 carries out both cis and trans cyclopropanation at the proximal position of the oxygenated mycolates .
A protein of MLpTN with NCBI accession id NP_302570.1 is annotated as a cyclopropane synthase. This protein may be re-annotated as cmaA2 as it contains the cmaA2-specific ‘ADGAGDA’ motif. The study also revealed that cmaA2 gene is absent in fast growing species, MAbATCC, MGlPYR, MVaPYR, MSgMC2, MJLS, MKMS and MMCS as the corresponding orthologues do not contain the ‘ADGAGDA’ motif (Figure 3). Hence in the above listed organisms, trans cyclopropanation at proximal end may not take place, as this reaction is specific to cmaA2. This further reiterates the rationale for the implication of cmaA2 as one of the major factors contributing to the pathogenicity of Mycobacteria as mutants lacking this gene are known to evoke 5-fold increase in host – immune response .
In order to gain further insight into the relationship amongst the several cyclopropane synthases, a phylogenetic tree was reconstructed using parsimony (Figure 4). An in-depth analysis of the phylogenetic tree revealed that mmaA1 responsible for methylation at the proximal end of mycolic acid  is present in MTbH37Rv, MBoAF, MLpTN and MAv104 strains and forms a distinct cluster. Methylation and cis-trans isomerization by mmaA1 is succeeded by trans cyclopropanation by cmaA2. It is interesting to observe that all organisms containing cmaA2 also contain mmaA1 except MUlAg. However, it needs to be mentioned here that the mmaA1 cluster is shared with another node that contain umaA genes with significant bootstrap value (97 times out of 100) and MUlAg contains one such umaA gene (Locus tag: MUL_4538). The mmaA1 cluster also includes a node with a small cluster of genes without any specific functionality but contain the cyclopropane synthase domain. These genes belong to MSgMC2, MJLS, MKMS, MMCS, MGlPYR, MVaPYR and MAbATCC.
mmaA2 has two distinct roles viz., cis cyclopropanation at distal end in alpha mycolates and cis cyclopropanation at proximal end in oxygenated mycolates . Previous experiments have shown that distal cyclopropanation helps in increasing the resistance of Mycobacteria to hydrogen peroxide, a major factor contributing to the oxidative stress experienced by the bacteria . mmaA2 is absent in non-pathogenic strains such as, MJLS, MKMS, MMCS, MGlPYR, MSgMC2 and MVaPYR. In the phylogenetic tree, these genes form a distinct cluster.
mmaA3 catalyses the addition of methyl moiety at the hydroxyl group which is newly formed by mmaA4 at the distal end in oxygenated mycolates , , . It is interesting to note that mmaA3 is absent in MLpTN, MLpBr, MAvK-10 and MAv104, hence may affect the methoxy mycolate production in these organisms. In the phylogenetic tree mmaA3 cluster share the same node with mmaA4 cluster.
mmaA4 catalyses the addition of methyl and hydroxyl branch at distal end in oxygenated mycolates , . Deletion of mmaA4 abolishes synthesis of both methoxy- and ketomycolates in Mycobacteria , . The node containing mmaA4 also includes a set of genes which do not have well defined function but contain the methyltransferase domain. These ‘undefined methyltransferase genes’ belong to non-pathogenic strains such as MJLS, MKMS, MMCS, MGlPY and MVaPYR. It is interesting to note that these organisms do not contain mmaA4. Hence it can hypothesized that these undefined methyltransferase genes may have function similar to mmaA4 as absence of mmaA4 leads to reduction in ketomycolate production, causing increased permeability of the cell wall, with the hypersensitivity to both ampicillin and RIF and impaired growth at low temperature .
ufaA1 and other cyclopropane synthase containing proteins cluster separately. MSA of ufaA1 genes helped to delineate the motif ‘DLLAQPTPAW’ (Figure 5). The orthologues of this gene are absent in MAbATCC, MVaPYR, MSgMC2, MJLS, MKMS, MMCS, MGlPYR, MLpTN, MLpBr, MAvK-10 and MAv104 as the motif ‘DLLAQPTPAW’ is absent. However, the affect of the absence of ufaA1 in mycolate biosynthesis needs to be probed further. This motif is however not conserved in ufaA1 orthologs of MMrM and MUlAg. This observation can be attributed to the fact that both these organisms have taxonomically distinct relationship in comparison to other mycobacteria and are known to produce stereochemically different mycolates .
The cluster of pcaA that carries out the proximal cyclopropanation of alpha-mycolates  is shared with MT-11. pcaA is uniquely present only in pathogenic strains viz., MTbH37Rv, MTbH37Ra, MtbCDC, MBoAF MboTokyo, MBoBCG, MKZN, MTbF11, MAv104, MAvK-10, MMrM, MulAg, MLpTN and MLpBr. The phylogenetic tree depicts that MAv104 (Locus tag: MAV_4679) and MTbCDC (Locus tag: MT0486) are pcaA genes even though the same is not annotated accordingly in public domain databases.
Methyltransferase type-11 (MT-11) is involved in DNA regulation. These genes have been picked up during the search for ‘methyltransferase containing domains’. They however do not have a direct evident role in mycolate biosynthesis and hence were not analysed further. It is however interesting to note that MT-11 is present only in MVaPYR, MSgMC2, MJLS, MKMS, MMCS, MGlPYR, MMrM, MAvK-10 and MAv104.
otsB2 encodes trehalose-6-phosphate phosphatase which dephosphorylates trehalose-6-phosphate to yield trehalose-mono-phosphate. The truncation of otsB2 at N' in MSgMC2 was observed during the comparative analysis, and may have affect on its functionality. However, studies by Woodruff et al.,  revealed that this gene is functional.
Predicted functional linkages
Novel functional linkages could be identified for genes involved in mycolate biosynthesis using phylogenetic profiling. These linkages are said to be novel, as the present annotation does not suggest any definite role of these genes in mycolic acid biosynthesis. A total of 1661 unique protein pairs satisfied the criteria for MI and CC (Table S1). Of these, 91 proteins have well-defined function in public domain databases like Tuberculist [http://tuberculist.epfl.ch/] and TBDB. Analysis of the proteins with known function revealed that Rv0503c (cmaA2 involved in trans cyclopropanation at proximal end of mycolate), Rv0470c (pcaA involved in cis cyclopropanation at proximal end of mycolate), Rv1273c (ABC transporter) and Rv3804c (fbpA involved in cell wall biosynthesis via its mycolyltransferase activity) display >230 functional linkages (Table 2). These genes can thus serve as network hubs  and may be probable chokepoints. This hypothesis is further substantiated by the crucial role played by these genes in mycolate biosynthesis as deduced by several experimental studies pertaining to knockout mutants–, , – and co-expression[TBDB] – For e.g., Rv0227c (a conserved hypothetical protein) shares significant MI and PCC with mycolate genes like Rv3801c (fadD32) and Rv3799c (accD4), and hence, Rv0227c may be functionally linked to mycolic acid biosynthesis (Figure 6). It has also been proved to be essential for growth by Sassetti et al, 2003 – and co-expression values with fadD32 and accD4 is 0.43 and 0.61 respectively. Hence, it can be hypothesized that this gene seems to have a vital role in mycolate synthesis.
Nodes in red colour are genes with known function in mycolate biosynthesis while those in blue colour are hypothetical genes. Black line shows direct linkage of Rv0227c.
The hypothetical proteins which share significant MI (>0.9) and PCC (>0.8) with genes involved in mycolate biosynthesis, and their role is further substantiated by recent literature, were summarized in Table 3. When the functional linkages displayed by the hypothetical proteins predicted to be involved in mycolate biosynthesis were analysed (Table 3), the following were the observations. Rv3802c recently annotated as phospholipase thioesterase , displayed maximum linkages (total: 121). Based on these observations it can be hence hypothesized that this enzyme may play a crucial role in mycolate biosynthesis and hints at the possibility that the functionality encoded by this gene is unique. Rv3722c a conserved hypothetical gene, shown to be essential by Sasetti et al, 2003, , displays significant co-expression with mycolate biosynthesis genes, hence providing a possible role of this gene in this pathway.
Please refer to the Table S1 for the complete list of hypothetical proteins that share significant functional linkages with mycolate biosynthesis genes, and hence may play a role either in transport of the protein products or in the regulation of the genes.
Arabinogalactan tethers the mycolic acid layer to the peptidoglycan , . The biosynthesis of AG has been reviewed extensively and the same is included as a component of ‘mycolyl-arabinogalactan-peptidoglycan complex’ biosynthesis in MetaCyc. Comparative analyses revealed that all the genes involved in AG biosynthesis are conserved in the 21-mycobacterial genomes with minor differences arising in genes with redundant functionality. rmlC gene encodes dTDP-4-keto-6-deoxyglucose epimerase, the third enzyme in the M. tuberculosis dTDP-L-rhamnose pathway. rmlC is conserved in all organisms except MGlPYR wherein it is present as a bifunctional protein (Locus tag: Mflv_3297) possessing truncated reductase and epimerase domains. Rv3779 is a glycosyltransferase responsible for direct synthesis of polyprenyl-phospho-mannopyranose, an intermediate in AG biosynthesis . The orthologs of this gene are absent in fast-growing species like MGlPYR, MVaPYR, MSgMC2, MJLS, MKMS and MMCS and hence may be responsible for the observed variations in cell growth and shape in these organisms.
The number of predicted functionally linked protein pairs for AG biosynthesis is 2086. These include, Rv1302 (rfe, involved in AG biosynthesis), Rv1086 (short-chain isoprenyl diphosphate synthase), Rv2361c (long-chain isoprenyl diphosphate synthase; essential gene), Rv3265c (wbbL1, rhamnosyl transferase; essential gene), Rv3464 (rmlB, dTDP-GLUCOSE 4,6-DEHYDRATASE; essential gene), Rv3794 (embA, arabinosyl transferase; essential gene) and Rv3809c (glf, galactose pyrannose mutase) and display >230 functional linkages (Table 2). Mutant studies for the above proteins provide evidence for them to be “essential” for survival –. The large number of functional linkages displayed by these proteins when correlated with mutant studies suggest their role as ‘hubs’ in protein interaction networks, which may further translate into them being “choke points” in metabolic networks. Thus, the high proportion of functional linkages correlate directly with the mutant studies thus providing a rationale for the hypothesis that number of linkages is directly proportional to the essentiality of the gene.
The hypothetical proteins which share significant MI (>0.9) and PCC (>0.9) with genes involved in AG biosynthesis, and their role is further substantiated by recent literature, were summarized in Table 4. Genes Rv1209c and Rv3031 are annotated as ‘hypothetical proteins’ and share significant mutual information content and Pearson correlation coefficient with genes involved in AG biosynthesis, hence suggesting a role of these genes in AG metabolism (Table 4). Recent findings by Jackson et al., 2009  and Kaur et al., 2009  support the same. Please refer to the Table S2 for the complete list of hypothetical proteins that share significant functional linkages with AG biosynthesis genes.
Lipomannan and lipoarabinomannan biosynthesis
Lipoarabinomannan and lipomannan phosphatidylinositol mannosides (PIMs) are major phosphatidylinositol (PI)-based lipoglycans/glycolipids of Mycobacterium. They play a major role in phagocytosis, persistence of bacilli in phagocytic cells, CD-1-restricted antigen presentation, initiation of innate immunity, and in antibody-mediated immunity . Recent work of Kaur D et al., 2009  has provided a thorough understanding of LM and LAM biosynthesis. However the same is not included in public-domain databases.
For comparative analysis, an exclusive list of genes catalyzing the LM/LAM biosynthesis process is mandatory. In order to facilitate ease-of-mapping across 21 genomes, we have reconstructed the pathway manually via literature curation and Pathway Tools software. Some of the interesting observations apart from the variations reported earlier in Kaur D et al., 2009  in genes responsible for capping includes Rv2181, an integral membrane protein. It is the alpha (1→2) ManT responsible for the synthesis of the alpha (1→2) ManP-linked branches, characteristic of the mannan backbone of LM and LAM; orthologous gene of which is absent in MAbATCC as it may contain unusual alpha (1→3) mannosyl side chains as in M. chelonae,  instead of alpha (1→2) which is commonly found in all other mycobacterial species , , . Rv1635c is a transmembrane protein that carries out mannose capping of LAM moieties in the periplasmic side of plasmamembrane in MTbH37Rv . The orthologs of this gene are absent in MSgMC2, MKMS, MMCS, MJLS and MAbATCC, as these species are devoid of ManLAM and contain alternate capping of LAM moieties viz., PILAM and AraLAM –, .
473 genes were predicted to be functionally linked (via phylogenetic profiling) with genes known to be involved in LAM biosynthesis (Table 2). Rv2174 (mptA), a polyprenol-P- mannose alpha 1→6 mannosyltransferase displayed the maximum linkages (>230). The large number of linkages can be attributed to the fact that disruption of this gene affects the optimal growth of the mycobacterium ,  as revealed by the TraSH analysis of this gene. Rv2188c, pimB, an alpha-D-Mannose-alpha 1→6-phosphatidyl-myoinositol mannosyltransferase and Rv2610c (pimA) display 112 linkages each, thus providing a rationale for their inclusion as potential candidates for therapeutic development.
The hypothetical proteins which share significant MI (>0.9) and PCC (>0.9) with genes involved in LAM biosynthesis, and their role is further substantiated by recent literature, were summarized in Table 5. Few studies have found evidence that Rv2613c encodes GT4 family glycosyl transferase  and Rv2257c encodes a homologue of pbpX of MSgMC2, which has a role in antibiotic resistance . Table S3 catalogs the complete list of hypothetical genes that may have a possible role in LAM biosynthesis.
Phthiocerol dimycocerosate (PDIM) biosynthesis
PDIM has a prominent role in evoking adaptive immune response as well as in combating oxidative stress by scavenging the oxygen free radicals . PDIM biosynthesis involves four main steps viz., priming of long-fatty acids and synthesis of diol component of phthiocerol, biosynthesis of phthiocerol ppsE protein, enzymatic synthesis of mycocerosic acid and transesterification of mycocerosic acid onto the diol component of phthiocerol . It is now understood that the complete PDIM molecules are synthesized in the cytoplasm of M. tuberculosis before being translocated into the cell wall . Previous studies have reported the presence of PDIM in pathogenic Mycobacteria with the exception of M.gastri . PDIM biosynthetic pathway has been reconstructed in-house using Pathway Tools. The biosynthesis of PDIM is carried out by genes ppsA-E, which encode a type I modular polyketide synthase responsible for the synthesis of phthiocerol and phenolphthiocerol by elongation of a C20–C22 fatty acyl chain or an acyl chain containing a phenol moiety with three malonyl-CoA and two methylmalonyl-CoA units . mas encodes an iterative type I polyketide synthase that produces mycocerosic acids after two to four rounds of extension of C18–C20 fatty acids with methylmalonyl-CoA units . papA5 catalyzes diesterification of phthiocerol and phthiodiolone with mycocerosate  along with fadD26, a fatty acyl- AMP ligase involved in the activation and transfer of long-chain fatty acids . drrC and mmpL7 are necessary for the proper localization of DIM .
Comparative studies revealed that FadD26 is present in all 21 mycobacterial species. ppsA-E genes are absent in MAv104, MAvK-10, MSgMC2, MAbATCC, MKMS, MJLS and MMCS. ppsC-E genes are absent in MVaPYR and papA5 is absent in MAv104, MAvK-10 and MSgMC2. The gene mas is absent in MVaPYR and MGlPYR. These findings hints at the absence of PDIM production in the organisms listed above. The transport proteins of PDIM viz., mmpl7 is absent in MAv104, MAvK-10, MSgMC2, MKMS, MJLS, MMCS, MGlPYR and MVaPYR; drrC is however present in all the 21 Mycobacteria. This suggests that the daunorubicin resistance, which is one of the activities of drrC apart from its role in PDIM translocation, is present is these organisms inspite the absence PDIM production.
Phylogenetic profiling for PDIM biosynthesis
A total of 677 unique protein pairs satisfied the criteria for MI and CC and were predicted to be functionally linked (via phylogenetic profiling) with genes known to be involved in PDIM biosynthesis (Table S4). Rv1528c (polyketide synthase associated protein PapA4) displayed the maximum number of linkages (99). Analysis of the proteins with known function revealed that Rv2933 (rfe, involved in PDIM biosynthesis), Rv2930 (acyl-CoA synthetase), Rv2941 (acyl-CoA synthetase), Rv2942 (transmembrane transport protein MmpL7), Rv3820c (polyketide synthase associated protein PapA2) and Rv3824c (polyketide synthase associated protein PapA1) displayed >90 functional linkages (Table 2).
The hypothetical proteins which share significant MI (>0.9) and PCC (>0.8) with genes involved in PDIM biosynthesis, and their role is further substantiated by literature, are summarized in Table 6. Rv0748, Rv1301 and Rv2681, annotated as ‘conserved hypothetical proteins’ may have a role to play in in this pathway as they share significant MI and PCC with PDIM genes and have also been shown to be essential by mutant studies. Rv1461, a conserved hypothetical protein, shown to be essential by Sasseti et al., 2003, displays significant co-expression with Rv2930 (acyl- CoA synthetase). Previous studies report that Rv1461 (ppS1) to be an ortholog of SufB, a highly conserved component of the [Fe-S] cluster, assembly and repair SUF (mobilization of sulfur) machinery, crucial for survival . This clearly shows the crucial role of this enzyme in PDIM biosynthesis and hints at the possibility that the functionality encoded by this gene is unique. Rv2681, a conserved hypothetical protein, shows significant co-expression with Rv2933 and Rv2942, hence providing a possible role of this gene in this pathway . Please refer to the Table S4 for the complete list of hypothetical proteins that share significant functional linkages with PDIM biosynthesis genes.
Identification of putative drug targets
The worldwide increase in multi-drug resistant Mycobacterium tuberculosis strains poses a great threat to human health and highlights the need to identify new anti-tubercular agents. The construction and analysis of molecular interaction networks provides a powerful means to understand the complexity of biological systems and to reveal hidden relationships between drugs, genes, proteins, and diseases . We have used the knowledge gained from the above analyses for rational identification of putative drug targets and estimated their appropriateness by sequence analysis. Many currently unexploited MTbH37Rv receptors may be chemically druggable and could serve as novel anti-tubercular targets. Those genes in the above analysis that were classified as essential, automatically form a first list of putative targets for anti-tubercular drugs, since their total inactivation may result in loss of production of cell wall components and hence the viability or the pathogenicity of the bacteria. However, it is reasoned that an ideal target should be essential not only in terms of the reaction it can catalyse, but also as the only protein coded by the genome that can perform the same task. Moreover, an ideal target should also have no recognisable homologue in the host system, which can in principle compete with the same drug, leading to adverse effects in the host system , . Sequence analysis of MTbH37Rv with human proteomes was therefore carried out for each of the identified targets and the results are summarised in Table 7.
Of the six proteins classified to be essential as well as functionally linked with cell wall components, (Table 7), no close homologues were observed in human proteome. Literature survey revealed that studies have listed all the genes except Rv0227 to be plausible targets for drug design , . Hence this is the first study to report Rv0227c as a novel target for MtbH37Rv.
Comparative genomics of genes involved in cell envelope biosynthesis amongst the 21 mycobacterial species at different levels of biocomplexity viz., sequence similarity, metabolic pathway context and phylogenetic profiling provide a rationale for the observed variation in components of cell wall according to their niche occupancy. Our findings suggest that the genes involved in mycolate biosynthesis are highly conserved with variations observed in genes, which form cylcopropane rings. AG biosynthesis is conserved in all the 21 Mycobacteria. LM/LAM biosynthetic machinery is conserved with known-variations in capping. PDIM-specific polyketide synthases are present only in pathogenic strains. The predicted functional linkages augment the search space responsible for the biosynthesis of the crucial components of cell wall apart from providing a rationale for the analyzing of network hubs and understanding the subtle relationships between various pathways. Experimental data can further validate the specific function encoded by proteins predicted through phylogenetic profiling studies in different metabolic pathways. Moreover, the shortlisted probable drug targets provide a hypothesis for use in tuberculosis drug design and needs to be tested experimentally. The methodology addresses several issues related to annotation discrepancies amongst closely related organisms apart from providing broader network of genes involved in any metabolic process. The conserved genes complement the TraSH data to arrive at a catalogue of ‘minimal gene set’ that Mycobacteria require for their survival and hence pathogenicity. The variant gene-set suggest the existence of alternate routes for biosynthesis of the cell envelope components. The methodology used is robust and is applicable for analyses of hundreds of prokaryotic genomes that are being sequenced due to the advent of NGS technologies.
Complete list of predicted functionally linked genes involved in mycolate biosynthesis.
Complete list of predicted functionally linked genes involved in Arabinogalactan biosynthesis.
Complete list of predicted functionally linked genes involved in Lipomannan and lipoarabinomannan biosynthesis.
The authors thank Ms. Bhakti Limaye for generating the input dataset for phylogenetic profiling.
Conceived and designed the experiments: RJ SMK. Performed the experiments: RB PV SD. Analyzed the data: RB PV SD SMK. Wrote the paper: RB PV SD SMK. Comparative genomics: RB. Phylogenetic profiling: PV. Metabolic pathway reconstruction: SD.
- 1. Vissa V, Sakamuri RM, Li Wei, Brennan PJ (2009) Defining mycobacteria: Shared and specific genome features for different lifestyles. Indian J of Microbiol 49: 11–47.
- 2. Guenin-Macé L, Siméone R, Demangel C (2009) Lipids of pathogenic Mycobacteria: contributions to virulence and host immune suppression. Transbound Emerg Dis 56: 255–268.
- 3. Nikaido H, Kim SH, Rosenberg EY (1993) Physical organization of lipids in the cell wall of Mycobacterium chelonae. Mol Microbiol 8: 1025–1030.
- 4. Minnikin DE, Kremer L, Dover LG, Besra GS (2002) The methyl-branched fortifications of Mycobacterium tuberculosis. Chem Biol 9: 545–553.
- 5. Meena LS, Rajni (2010) Survival mechanisms of pathogenic Mycobacterium tuberculosis H37Rv. FEBS J 277: 2416–2427.
- 6. Pellegrini M, Marcotte EM, Thompson MJ, Eisenberg D, Yeates TO (1999) Assigning protein functions by comparative genome analysis: protein phylogenetic profiles Proc Natl Acad Sci U S A 96: 4285–4288.
- 7. Date SV, Peregrín-Alvarez JM (2008) Phylogenetic profiling. In: Keith JonathanM, editor. Methods Mol Biol. New Jersey: Humana Press. pp. 201–216.
- 8. Date SV, Marcotte EM (2003) Discovery of uncharacterized cellular systems by genome-wide analysis of functional linkages. Nat Biotechnol 221: 1055–1062.
- 9. Jothi R, Przytycka TM, Aravind L (2007) Discovering functional linkages and uncharacterized cellular pathways using phylogenetic profile comparisons: a comprehensive assessment. BMC Bioinformatics 8: 173.
- 10. Benson DA, Karsch-Mizrachi I, Lipman DJ, Ostell J, Keith WheelerDL (2008) NucleicAcids Res 36: D25–D30.
- 11. Ripoll F, Pasek S, Schenowitz C, Dossat C, Barbe V, et al. (2009) Non mycobacterial virulence genes in the genome of the emerging pathogen Mycobacterium abscessus. PLoS One 4:
- 12. Li L, Bannantine JP, Zhang Q, Amonsin A, May BJ, et al. (2005) The complete genome sequence of Mycobacterium avium subspecies paratuberculosis. Proc Natl Acad Sci U S A 102: 12344–12349.
- 13. Garnier T, Eiglmeier K, Camus JC, Medina N, Mansoor H, et al. (2003) The complete genome sequence of Mycobacterium bovis. Proc Natl Acad Sci U S A 100: 7877–7882.
- 14. Brosch R, Gordon SV, Garnier T, Eiglmeier K, Frigui W, et al. (2007) Genome plasticity of BCG and impact on vaccine efficacy. Proc Natl Acad Sci U S A 104: 5596–5601.
- 15. Seki M, Honda I, Fujita I, Yano I, Yamamoto S, et al. (2009) Whole genome sequence analysis of Mycobacterium bovis bacillus Calmette-Guérin (BCG) Tokyo 172: a comparative study of BCG vaccine substrains. Vaccine 27: 1710–1716.
- 16. Monot M, Honoré N, Garnier T, Zidane N, Sherafi D, et al. (2009) Comparative genomic and phylogeographic analysis of Mycobacterium leprae. Nat Genet 41: 1282–1289.
- 17. Cole ST, Eiglmeier K, Parkhill J, James KD, Thomson NR, et al. (2001) Massive gene decay in the leprosy bacillus. Nature 409: 1007–1011.
- 18. Stinear TP, Seemann T, Harrison PF, Jenkin GA, Davies JK, et al. (2008) Insights from the complete genome sequence of Mycobacterium marinum on the evolution of Mycobacterium tuberculosis. Genome Res 18: 729–741.
- 19. Fleischmann RD, Alland D, Eisen JA, Carpenter L, White O, et al. (2002) Whole-genome comparison of Mycobacterium tuberculosis clinical and laboratory strains. J Bacteriol 184: 5479–5490.
- 20. Zheng H, Lu L, Wang B, Pu S, Zhang X, et al. (2008) Genetic basis of virulence attenuation revealed by comparative genomic analysis of Mycobacterium tuberculosis strain H37Ra versus H37Rv. PLoS One 3:
- 21. Camus JC, Pryor MJ, Médigue C, Cole ST (2002) Re-annotation of the genome sequence of Mycobacterium tuberculosis H37Rv. Microbiology 148: 2967–2973.
- 22. Stinear TP, Seemann T, Pidot S, Frigui W, Reysset G, et al. (2007) Reductive evolution and niche adaptation inferred from the genome of Mycobacterium ulcerans, the causative agent of Buruli ulcer. Genome Res 17: 192–200.
- 23. Pearson WR, Lipman DJ (1988) Improved tools for biological sequence comparison. Proc Natl Acad Sci U S A 85: 2444–2448.
- 24. Kuo-Bin Li (2003) ClustalW-MPI: ClustalW Analysis Using Distributed and Parallel Computing. Bioinformatics 19: 1585–1586.
- 25. Felsenstein J (1989) PHYLIP-Phylogeny Inference Package (Version 3.2). Cladistics 5: 164–166.
- 26. Tamura K, Dudley J, Nei M, Kumar S (2007) MEGA4: Molecular Evolutionary Genetics Analysis (MEGA) software version 4.0. Mol Biol and Evol 24: 1596–1599.
- 27. Karp PD, Paley SM, Krummenacker M, Latendresse M, Dale JM, et al. (2010) Pathway Tools version 13.0: integrated software for pathway/genome informatics and systems biology. Brief Bioinform. 11: 40–79.
- 28. Caspi R, Altman T, Dale JM, Dreher K, Fulcher CA, et al. (2010) The MetaCyc database of metabolic pathways and enzymes and the BioCyc collection of pathway/genome databases. Nucleic Acids Res 38: D473–D479.
- 29. Arakaki AK, Huang Y, Skolnick J (2009) EFICAz2: enzyme function inference by a combined approach enhanced by machine learning. BMC Bioinformatics 10: 107.
- 30. UniProt Consortium. (2010) The Universal Protein Resource (UniProt) in 2010. Nucleic Acids Res 38: D142–D148.
- 31. Daub CO, Steuer R, Selbig J, Kloska S (2004) Estimating mutual information using B-spline functions-an improved similarity measure for analysing gene expression data. BMC Bioinformatics 5: 118.
- 32. Reddy TB, Riley R, Wymore F, Montgomery P, DeCaprio D, et al. (2009) TB database: an integrated platform for tuberculosis research. Nucleic Acids Res 37: D499–D508.
- 33. Beste DJ, Espasa M, Bonde B, Kierzek AM, Stewart GR, et al. (2009) The genetic requirements for fast and slow growth in mycobacteria. PLoS One 4:
- 34. Sassetti CM, Rubin EJ (2003) Genetic requirements for mycobacterial survival during infection. Proc Natl Acad Sci U S A100: 12989–12994.
- 35. Sassetti CM (2003) Genes required for mycobacterial growth defined by high density mutagenesis. Mol Microbiol 48: 77–84.
- 36. Marri PR, Bannantine JP, Golding GB (2006) Comparative genomics of metabolic pathways in Mycobacterium species: gene duplication, gene decay and lateral gene transfer. FEMS Microbiol Rev 30: 906–925.
- 37. Takayama K, Wang C, Besra GS (2005) Pathway to synthesis and processing of mycolic acids in Mycobacterium tuberculosis. Clin Microbiol Rev 18: 81–101.
- 38. Dubnau E, Chan J, Raynaud C, Mohan VP, Lanéelle MA, et al. (2000) Oxygenated mycolic acids are necessary for virulence of Mycobacterium tuberculosis in mice. Mol Microbiol 36: 630–637.
- 39. Glickman MS, Cox JS Jacobs WR Jr<1/americanism> (2000) A novel mycolic acid cyclopropane synthetase is required for cording, persistence, and virulence of Mycobacterium tuberculosis. Mol Cell 5: 717–727.
- 40. Liu J, Barry CE III, Besra GS Nikaido H (1996) Mycolic acid structure determines the fluidity of the mycobacterial cell wall. J Biol Chem 271: 29545–29551.
- 41. Yuan Y, Zhu Y, Crane DD, Barry CE III (1998) The effect of oxygenated mycolic acid composition on cell wall function and macrophage growth in Mycobacterium tuberculosis. Mol Microbiol 29: 1449–1458.
- 42. Gebhardt H, Meniche X, Tropis M, Krämer R, Daffé M, et al. (2007) The key role of the mycolic acid content in the functionality of the cell wall permeability barrier in Corynebacterineae. Microbiology 153: 1424–1434.
- 43. Vissa VD, Brennan PJ (2001) The genome of Mycobacterium leprae: a minimal mycobacterial gene set. Genome Biol 2:
- 44. Barkan D, Liu Z, Sacchettini JC, Glickman MS (2009) Mycolic acid cyclopropanation is essential for viability, drug resistance, and cell wall integrity of Mycobacterium tuberculosis. Chem Biol 16: 499–509.
- 45. Huang CC, Smith CV, Glickman MS, Jacobs WR Jr, Sacchettini JC (2002) Crystal structures of mycolic acid cyclopropane synthases from Mycobacterium tuberculosis. J Biol Chem 277: 11559–11569.
- 46. Barkan D, Rao V, Sukenick GD, Glickman MS (2010) Redundant function of cmaA2 and mmaA2 in Mycobacterium tuberculosis cis cyclopropanation of oxygenated mycolates. J Bacteriol 192: 3661–3668.
- 47. Rao V, Gao F, Chen B, Jacobs WR Jr, Glickman MS (2006) Trans-cyclopropanation of mycolic acids on trehalose dimycolate suppresses Mycobacterium tuberculosis -induced inflammation and virulence. J Clin Invest 116: 1660–1667.
- 48. Glickman MS (2003) The mmaA2 gene of Mycobacterium tuberculosis encodes the distal cyclopropane synthase of the alpha-mycolic acid. J Biol Chem 278: 7844–7849.
- 49. Yuan Y, Lee RE, Besra GS, Belisle JT, Barry CE (1995) Identification of a gene involved in the biosynthesis of cyclopropanated mycolic acids in Mycobacterium tuberculosis. Proc Natl Acad Sci U S A 92: 6630–6634.
- 50. Behr MA, Schroeder BG, Brinkman JN, Slayden RA, Barry CE (2000) A point mutation in the mma3 gene is responsible for impaired methoxymycolic acid production in Mycobacterium bovis BCG strains obtained after 1927. J Bacteriol 2000 182: 3394–3399.
- 51. Alahari A, Alibaud L, Trivelli X, Gupta R, Lamichhane G, et al. (2009) Mycolic acid methyltransferase, MmaA4, is necessary for thiacetazone susceptibility in Mycobacterium tuberculosis. Mol Microbiol 71: 1263–1277.
- 52. Yuan Y, Barry CE (1996) A common mechanism for the biosynthesis of methoxy and cyclopropyl mycolic acids in Mycobacterium tuberculosis. Proc Natl Acad Sci U S A 93: 12828–12833.
- 53. Daffe M, Laneelle MA, Lacave C (1991) Structure and stereochemistry of mycolic acids of Mycobacterium marinum and Mycobacterium ulcerans. Res Microbiol 142: 397–403.
- 54. Woodruff PJ, Carlson BL, Siridechadilok B, Pratt MR, Senaratne RH, et al. (2004) Trehalose is required for growth of Mycobacterium smegmatis. J Biol Chem 279: 28835–28843.
- 55. Agarwal S, Deane CM, Porter MA, Jones NS (2010) Revisiting date and party hubs: novel approaches to role assignment in protein interaction networks. PLoS Comput Biol 6:
- 56. Glickman MS, Cahill SM, Jacobs WR Jr (2001) The Mycobacterium tuberculosis cmaA2 gene encodes a mycolic acid trans-cyclopropane synthetase. J Biol Chem 276: 2228–2233.
- 57. Målen H, De Souza GA, Pathak S, Søfteland T, Wiker HG (2011) Comparison of membrane proteins of Mycobacterium tuberculosis H37Rv and H37Ra strains. BMC Microbiol 11: 18.
- 58. Parker SK, Barkley RM, Rino JG, Vasil ML (2009) Mycobacterium tuberculosis Rv3802c encodes a phospholipase/thioesterase and is inhibited by the antimycobacterial agent tetrahydrolipstatin. PLoS One 4:
- 59. Kaur D, Pham H, Larrouy-Maumus G, Rivière M, Vissa V, et al. (2009) Initiation of methylglucose lipopolysaccharide biosynthesis in mycobacteria. PLoS One 4:
- 60. Bhowruth V, Alderwick LJ, Brown AK, Bhatt A, Besra GS (2008) Tuberculosis: a balanced diet of lipids and carbohydrates. Biochem Soc Trans 36: 555–565.
- 61. Scherman H, Kaur D, Pham H, Skovierová H, Jackson M, et al. (2009) Identification of a polyprenylphosphomannosyl synthase involved in the synthesis of mycobacterial mannosides. J Bacteriol 191: 6769–6772.
- 62. Jackson M, Brennan PJ (2009) Polymethylated polysaccharides from Mycobacterium species revisited. J Biol Chem 284: 1949–1953.
- 63. Guerardel Y, Maes E, Elass E, Leroy Y, Timmerman P, et al. (2002) Structural study of lipomannan and lipoarabinomannan from Mycobacterium chelonae. Presence of unusual components with alpha 1,3-mannopyranose side chains. J Biol Chem 277: 30635–30648.
- 64. Dinadayala P, Kaur D, Berg S, Amin AG, Vissa VD, et al. (2006) Genetic basis for the synthesis of the immunomodulatory mannose caps of lipoarabinomannan in Mycobacterium tuberculosis. J Biol Chem 281: 20027–20035.
- 65. Vignal C, Guérardel Y, Kremer L, Masson M, Legrand D, et al. (2003) Lipomannans, but not lipoarabinomannans, purified from Mycobacterium chelonae and Mycobacterium kansasii induce TNF-alpha and IL-8 secretion by a CD14-toll-like receptor 2-dependent mechanism. J Immunol 171: 2014–2023.
- 66. Delmas C, Gilleron M, Brando T, Vercellone A, Gheorghui M, et al. (1997) Comparative structural study of the mannosylated-lipoarabinomannans from Mycobacterium bovis BCG vaccine strains: characterization and localization of succinates. Glycobiology 7: 811–817.
- 67. Guérardel Y, Maes E, Briken V, Chirat F, Leroy Y, et al. (2003) Lipomannan and lipoarabinomannan from a clinical isolate of Mycobacterium kansasii: novel structural features and apoptosis-inducing properties. J Biol Chem 278: 36637–36651.
- 68. Joe M, Sun D, Taha H, Completo GC, Croudace JE, et al. (2006) The 5-deoxy-5-methylthio-xylofuranose residue in mycobacterial lipoarabinomannan. absolute stereochemistry, linkage position, conformation, and immunomodulatory activity. J Am Chem Soc 128: 5059–5072.
- 69. Treumann A, Xidong F, McDonnell L, Derrick PJ, Ashcroft AE, et al. (2002) 5-Methylthiopentose: a new substituent on lipoarabinomannan in Mycobacterium tuberculosis. J Mol Biol 316: 89–100.
- 70. Turnbull WB, Shimizu KH, Chatterjee D, Homans SW, Treumann A (2004) Identification of the 5-methylthiopentosyl substituent in Mycobacterium tuberculosis lipoarabinomannan. Angew Chem Int Ed Engl 43: 3918–3922.
- 71. Kaur D, Obregón-Henao A, Pham H, Chatterjee D, Brennan PJ, et al. (2008) Lipoarabinomannan of Mycobacterium: mannose capping by a multifunctional terminal mannosyltransferase. Proc Natl Acad Sci U S A 105: 17973–17977.
- 72. Jackson M, Crick DC, Brennan PJ (2000) Phosphatidylinositol is an essential phospholipid of mycobacteria. J Biol Chem 275: 30092–30099.
- 73. Flores AR, Parsons LM, Pavelka MS Jr (2005) Characterization of novel Mycobacterium tuberculosis and Mycobacterium smegmatis mutants hypersusceptible to beta-lactam antibiotics. J Bacteriol 187: 1892–1900.
- 74. Onwueme KC, Vos CJ, Zurita J, Ferreras JA, Quadri LE (2005) The dimycocerosate ester polyketide virulence factors of mycobacteria. Prog Lipid Res 44: 259–302.
- 75. Trivedi OA, Arora P, Vats A, Ansari MZ, Tickoo R, et al. (2005) Dissecting the mechanism and assembly of a complex virulence mycobacterial lipid. Mol Cell 17: 631–643.
- 76. Camacho LR, Constant P, Raynaud C, Laneelle MA, Triccas JA, et al. (2001) Analysis of the phthiocerol dimycocerosate locus of Mycobacterium tuberculosis. Evidence that this lipid is involved in the cell wall permeability barrier. J Biol Chem 276: 19845–19854.
- 77. Azad AK, Sirakova TD, Fernandes ND, Kolattukudy PE (1997) Gene knockout reveals a novel gene cluster for the synthesis of a class of cell wall lipids unique to pathogenic mycobacteria. J Biol Chem 272: 16741–16745.
- 78. Azad AK, Sirakova TD, Rogers LM, Kolattukudy PE (1996) Targeted replacement of the mycocerosic acid synthase gene in Mycobacterium bovis BCG produces a mutant that lacks mycosides. Proc Natl Acad Sci U S A 93: 4787–4792.
- 79. Buglino J, Onwueme KC, Ferreras JA, Quadri LE, Lima CD (2004) Crystal structure of PapA5, a phthiocerol dimycocerosyl transferase from Mycobacterium tuberculosis. J Biol Chem 279: 30634–30642.
- 80. Trivedi OA, Arora P, Sridharan V, Tickoo R, Mohanty D, et al. (2004) Enzymic activation and transfer of fatty acids as acyl-adenylates in mycobacteria. Nature 428: 441–445.
- 81. Huet G, Castaing JP, Fournier D, Daffé M, Saves I (2006) Protein splicing of SufB is crucial for the functionality of the Mycobacterium tuberculosis SUF machinery. J Bacteriol 188: 3412–3414.
- 82. Zuo Y, Wang Y, Malhotra A (2005) Crystal structure of Escherichia coli RNase D, an exoribonuclease involved in structured RNA processing. Structure 13: 973–984.
- 83. Kinnings SL, Xie L, Fung KH, Jackson RM, Xie L, et al. (2010) The Mycobacterium tuberculosis drugome and its polypharmacological implications. PLoS Comput Biol 6:
- 84. Hasan S, Daugelat S, Rao PS, Schreiber M (2006) Prioritizing genomic drug targets in pathogens: application to Mycobacterium tuberculosis. PLoS Comput Biol 2:
- 85. Raman K, Rajagopalan P, Chandra N (2005) Flux balance analysis of mycolic acid pathway: targets for anti-tubercular drugs. PLoS Comput Biol 1:
- 86. Raman K, Yeturu K, Chandra N (2008) targetTB: a target identification pipeline for Mycobacterium tuberculosis through an interactome, reactome and genome-scale structural analysis. BMC Syst Biol 2: 109.
- 87. Arnvig KB, Young DB (2009) Identification of small RNAs in Mycobacterium tuberculosis. Mol Microbiol 73: 397–408.
- 88. Dos Vultos T, Blázquez J, Rauzier J, Matic I, Gicquel B (2006) Identification of Nudix hydrolase family members with an antimutator role in Mycobacterium tuberculosis and Mycobacterium smegmatis. J Bacteriol 188: 3159–3161.
- 89. Erb TJ, Brecht V, Fuchs G, Müller M, Alber BE (2009) Carboxylation mechanism and stereochemistry of crotonyl-CoA carboxylase/reductase, a carboxylating enoyl-thioester reductase. Proc Natl Acad Sci U S A 106: 8871–8876.
- 90. Goldstone RM, Moreland NJ, Bashiri G, Baker EN, Shaun Lott J (2008) A new Gateway vector and expression protocol for fast and efficient recombinant protein expression in Mycobacterium smegmatis. Protein Expr Purif 57: 81–87.
- 91. Morris RP, Nguyen L, Gatfield J, Visconti K, Nguyen K, et al. (2005) Ancestral antibiotic resistance in Mycobacterium tuberculosis. Proc Natl Acad Sci U S A 102: 12200–12205.
- 92. Rachman H, Strong M, Ulrichs T, Grode L, Schuchhardt J, et al. (2006) Unique transcriptome signature of Mycobacterium tuberculosis in pulmonary tuberculosis. Infect Immun 74: 1233–1242.
- 93. Siméone R, Constant P, Guilhot C, Daffé M, Chalut C (2007) Identification of the missing trans-acting enoyl reductase required for phthiocerol dimycocerosate and phenolglycolipid biosynthesis in Mycobacterium tuberculosis. J Bacteriol 189: 4597–4602.