An in silico analysis of T-box regulated genes and T-box evolution in prokaryotes, with emphasis on prediction of substrate specificity of transporters

Abstract

Background: T-box anti-termination is an elegant and sensitive mechanism by which many bacteria maintain constant levels of amino acid-charged tRNAs. The amino acid specificity of the regulatory element is related to a so-called specifier codon and can in principle be used to guide the functional annotation of the genes controlled via the T-box anti-termination mechanism.

Results: Hidden Markov Models were defined to search the T-box regulatory element and were applied to all completed prokaryotic genomes. The vast majority of the genes found downstream of the retrieved elements encoded functionalities related to transport and synthesis of amino acids and the charging of tRNA. This is completely in line with findings reported in literature and with the proposed biological role of the regulatory element. For several species, the functional annotation of a large number of genes encoding proteins involved in amino acid transport could be improved significantly on basis of the amino acid specificity of the identified T-boxes. In addition, these annotations could be extrapolated to a larger number of orthologous systems in other species. Analysis of T-box distribution confirmed that the element is restricted predominantly to species of the phylum Firmicutes. Furthermore, it appeared that the distribution was highly species specific and that in the case of amino acid transport some boxes seemed to "pop-up" only recently.

Conclusion: We have demonstrated that the identification of the molecular specificity of a regulatory element can be of great help in solving notoriously difficult annotation issues, e.g. by defining the substrate specificity of genes encoding amino acid transporters on basis of the amino acid specificity of the regulatory T-box. Furthermore, our analysis of the species-dependency of the occurrence of specific T-boxes indicated that these regulatory elements propagate in a semi-independent way from the genes that they control.

Figure 1

Sequence logo[64]visualization of the 5 different T-box motifs. Both the consensus sequence and relative conservation of individual residues is displayed. Motif 1 (information content: 19.2 bits) displays the motif used to perform the T-box identification. Validation was performed by checking the presence of motif 2 (29.7 bits) or 3 (20.9 bits) together with motifs 4 (31.4 bits) and 5 (25.8 bits). Motif 1 includes the (a-specific) tRNA interaction site (T-box sequence, consensus GGTGG) located in the antiterminator loop. The other motifs include different parts of the specifier loop [23,26]: GNTG- and AG-box in motif 2 and 4, TGA-, AGGA- and AGTA-box in motif 3 and a conserved part of the specifier loop in motif 5 (GAG). The specifier codon is to be found within 1 – 5 nucleotides upstream of this conserved GAG.

Figure 2

Overview of T-box-regulated methionine biosynthesis in L. plantarum. Reactions coloured in blue are catalyzed by proteins encoded by genes regulated by a T-box. The figure was generared using the the Simpheny software tool (Genomatica, San Diego, USA). Reactions were based on the metabolic model of L plantarum published by [65].

Figure 3

Occurrence of T-boxes in relation with different transporter families. The figure displays a NJ-tree of the LIVCS-family transporters of the Firmicutes (left) and partial trees related to the APC-, MFS- and NSS-family transporters. It appears T-boxes are only associated with very few proteins of these families and the association appears to be very species-specific. Bootstrap values are given for those clusters that contain T-box regulated systems (indicated in black). Those systems that are controlled by a T-box are colored. For the LIVCS-family the sequences of the experimentally studied transporters from Pseudomonas aeruginosa (BraZ [49]), Corynebacterium glutamicum (BrnQ [48]) and Lactobacillus delbrueckii (BrnQ [43]) were included in the analysis. For the APC-family all B. subtilis sequences were included together with the orthologous clusters containing T-box regulated systems. For the MFS- and NSS-family only the orthologous clusters containing T-box regulated systems are shown. In the case of the MFS-family the asterisk indicates that the upstream sequence of these systems contains a box that seems to be degenerated.

Figure 4

T-box regulation of tRNA ligase encoding genes in the Firmicutes. The color coding relates to the presence or absence of a T-box upstream of the genes encoding the amino acid-specific tRNA ligases in the various species and strains. Green indicates the tRNA ligase(s) is (are) regulated by a T-box and red that the tRNA ligase(s) is (are) not regulated by a T-box. Although most tRNA ligases are present in one copy on the genome, several organisms contain two, or in some cases three copies of specific ligases (indicated by a number in the box). Orange indicates that 1 of the 2 tRNA ligases is regulated by a T-box or 1 out of 3 in the case of the argS genes in B. cereus ATCC 10987 and the aspS genes in C. acetobutylicum. Light green indicates that the tRNA ligase is not the first in the operon, but is regulated by a T-box with the same specificity. Yellow color coding indicates that the regulated tRNA ligase is the second gene in an operon in combination with another tRNA ligase gene regulated by a T-box with different specificity. White indicates that no tRNA ligase of this type is present in the organism. In principle, a species needs at least one specific tRNA ligase for each amino acid. Nevertheless, there are exceptions. For instance, all but one (Clostridium perfringens) of the analyzed genomes lack the gene that encodes a Gln-tRNA ligase and the genomes of the Chloroflexi, Actinobacteria and Thermoanaerobacter tencongens also lack an Asn-tRNA ligase. In these cases, the biological role of the Gln-tRNA ligase is taken over by the Glu tRNA ligase, which couples a Glu residue to the tRNA^Gln. The residue is subsequently transformed into a Gln by a tRNA specific amidotransferase [66]. Similarly, an Asn-tRNA^Asn is formed via transamidation of an Asp residue (Asp-tRNA^Asn to Asn-tRNA^Asn) in bacteria that lack an Asn tRNA ligase [67]. Consequently, we found that all species lacking either the Gln-tRNA ligase or the Asn-tRNA ligase have an orthologous gene coding for the corresponding amidotransferase. No T-boxes were identified upstream of those genes.

Figure 5

T-boxes preceding the genes related to amino acid biosynthesis in Firmicutes. Color coding identifies the presence of the biosynthesis pathway and whether it is regulated by a T-box: Green; T-box regulated; red; not T-box regulated; no color; pathway absent. ⁺TRAP protein is present. ^M Pathway genes organized in multiple operons. BCA indicates the branched chain amino acids valine, leucine and isoleucine.

Figure 6

The evolutionary relationship between some T-boxes. (A) shows a putative phylogeny of the branched-chain amino acid tRNA ligases of B. anthracis Ames, B. subtilis 168, C. acetobutylicum ATCC824D, L. acidophilus NCFM, L. plantarum WCFS1, L. mesenteroides ATCC8293 and S. aureus Mu50. (B) shows the Neighbor-Joining tree for the related T-boxes. The underlying alignments were made with the complete 200 – 300 nucleotides of the identified T-boxes. These alignments were homogeneous in the sense that the fully conserved motifs aligned perfectly and that between those conserved elements there were little gaps. Nevertheless, the low bootstrap support for the various branches indicates that this tree is unlikely to reflect the true phylogeny of the regulatory elements. (C) shows the Neighbor-Joining tree for various T-boxes found in B. anthracis Ames. Next to the tree, the part of the corresponding multiple sequence alignment containing the specifier codon (indicated in white letters) is depicted. The amino acid specificity of the specifier codon is color-coded: Red and orange relate to Ile, green to Leu, light blue to Phe, beige to Pro or Gln, pink to Ser, brown to Thr, turquoise to Trp, purple to Tyr and dark blue to Val. The family of the protein encoded by the regulated gene is indicated by the letters that follow the amino acid code. These protein families included the APC, LIVCS, MFS, NhaC and NSS transporter protein families and various tRNA-ligase families (S or Smr for mupirocin resistant tRNA ligase). The NSS-family transport proteins regulated by a Leu, Phe and Trp T-box are in-paralogs characteristic for the species of the Bacillus cereus group. The purple numbers between brackets indicate the bootstrap support for the displayed clusters (out of 1000).