Evolution of the Translocation and Assembly Module (TAM).
Bottom Line: We report that TamA and a closely related protein TamL are confined almost exclusively to Proteobacteria and Bacteroidetes/Chlorobi respectively, whereas TamB is widely distributed across the majority of Gram-negative bacterial lineages.Several sequence characteristics were discovered to define the TamB protein family: A signal-anchor linkage to the inner membrane, beta-helical structure, conserved domain architecture and a C-terminal region that mimics outer membrane protein beta-strands.Taken together, the structural and phylogenetic analyses suggest that the TAM likely evolved from an original combination of BamA and TamB, with a later gene duplication event of BamA, giving rise to an additional Omp85 sequence that evolved to be TamA in Proteobacteria and TamL in Bacteroidetes/Chlorobi.
Affiliation: Department of Microbiology, Monash University, Melbourne, Victoria, Australia.Show MeSH
Mentions: Initial hidden Markov model (HMM) searches (HMMER) of TamB highlighted a similarity to AsmA. Given the low overall similarity of TamB sequences we observed, and the unclear assignment of several sequences as either TamB or AsmA, we chose a very lenient cutoff to include as many divergent TamB-like sequences as possible for in-depth analysis. Domain analysis of the sequences identified in five iterative HMM searches (jackHMMER) using the E. coli TamB protein (UniProt: P39321) as the initial search input, showed that the number of proteins containing the DUF490 domain characteristic for TamB (Selkrig et al. 2012), as well as the AsmA domain, each plateaued at the fifth iteration (supplementary fig. S1, Supplementary Material online). Markov clustering (MCL) was then applied to the full set of 16,908 sequences after the fifth iteration (fig. 2). We note here a similarity between members of the TamB and AsmA protein families and the mitochondrial inner membrane proteins Mdm31 and Mdm32 (Dimmer et al. 2005), whether this is of functional significance awaits further investigation. The resulting subclusters were grouped following manual annotation, as well as removing fragments (<200 amino acids) and eukaryotic proteins resulting in 16,107 sequences (supplementary table S1, Supplementary Material online). Clustering and manual analyses also revealed the similarity of TamB not only to AsmA (including the two E. coli AsmA paralogs, YicH and YhjG) but also to a group of protein sequences with DUF748 and DUF3971 domains of unknown function (fig. 2 and supplementary fig. S2, Supplementary Material online). This relationship was confirmed in a comparison of the respective Pfam motifs using a pairwise comparison of profile HMMs (HHPred; http://toolkit.tuebingen.mpg.de/hhpred (last accessed October 11, 2014); Soding et al. 2005; supplementary table S2, Supplementary Material online): There is an underlying similarity of the Pfam domains between DUF490 (found in TamB) and AsmA, Mdm31/32, DUF3971, and DUF748. The protein clusters were manually joined into groups; despite several attempts at clustering and phylogenetic analyses, some sequence groups could not unambiguously be grouped with TamB or AsmA, and are therefore labeled as AsmA-TamB hereafter. The proteins grouped as TamB predominantly showed distinct sequence relationships according to their taxonomic species of origin (fig. 2 and supplementary fig. S2, Supplementary Material online).Fig. 2.—
Affiliation: Department of Microbiology, Monash University, Melbourne, Victoria, Australia.