Inter-genomic displacement via lateral gene transfer of bacterial trp operons in an overall context of vertical genealogy
© Xie et al; licensee BioMed Central Ltd. 2004
Received: 27 January 2004
Accepted: 23 June 2004
Published: 23 June 2004
The growing conviction that lateral gene transfer plays a significant role in prokaryote genealogy opens up a need for comprehensive evaluations of gene-enzyme systems on a case-by-case basis. Genes of tryptophan biosynthesis are frequently organized as whole-pathway operons, an attribute that is expected to facilitate multi-gene transfer in a single step. We have asked whether events of lateral gene transfer are sufficient to have obscured our ability to track the vertical genealogy that underpins tryptophan biosynthesis.
In 47 complete-genome Bacteria, the genes encoding the seven catalytic domains that participate in primary tryptophan biosynthesis were distinguished from any paralogs or xenologs engaged in other specialized functions. A reliable list of orthologs with carefully ascertained functional roles has thus been assembled and should be valuable as an annotation resource. The protein domains associated with primary tryptophan biosynthesis were then concatenated, yielding single amino-acid sequence strings that represent the entire tryptophan pathway. Lateral gene transfer of several whole-pathway trp operons was demonstrated by use of phylogenetic analysis. Lateral gene transfer of partial-pathway trp operons was also shown, with newly recruited genes functioning either in primary biosynthesis (rarely) or specialized metabolism (more frequently).
(i) Concatenated tryptophan protein trees are congruent with 16S rRNA subtrees provided that the genomes represented are of sufficiently close phylogenetic spacing. There are currently seven tryptophan congruency groups in the Bacteria. Recognition of a succession of others can be expected in the near future, but ultimately these should coalesce to a single grouping that parallels the 16S rRNA tree (except for cases of lateral gene transfer). (ii) The vertical trace of evolution for tryptophan biosynthesis can be deduced. The daunting complexities engendered by paralogy, xenology, and idiosyncrasies of nomenclature at this point in time have necessitated an expert-assisted manual effort to achieve a correct analysis. Once recognized and sorted out, paralogy and xenology can be viewed as features that enrich evolutionary histories.
The process of LGT
Lateral gene transfer (LGT) undoubtedly occurs with high frequency . But what is required for any given LGT event to escape transience and actually survive as a non-vertical contributor to the evolutionary history of a species? (i) Initially the incoming gene(s) must be replicated in the original recipient cell and its immediate progeny. (ii) Selective advantages of the LGT-gene(s) must be sufficient to foster eventual domination of the species population. (iii) During this time a crucial factor influencing survival will be whether the considerable demands imposed by the recipient genome for amelioration of alien genes can be met . Obviously, there will be less amelioration pressure if the donor and recipient genomes are phylogenetically close and have similar guanine/cytosine base ratios, dinucleotide frequencies, promotor motifs, and so on. Even here a LGT insertion into a genome might be contra-selected for other reasons, for example, if the location of the insertion disrupts the symmetry and physical balance between the origin and terminus points of replication . There is a balance in that one can expect alien sources of greatest novelty to be phylogenetically distant, yet genes from such remote sources will usually confront the greatest amelioration pressures. The most obvious candidates as successfully imported alien genes will encode novel functions that confer clear selective value, such as resistance to threatening environmental agents (e.g. antibiotics) or ability to utilize a new source of carbon and energy.
Enhanced probability of LGT success can be expected if only a single gene is needed for the novel function. This does not necessarily rule out the acquisition of a new function that is complex and multi-genic. Sometimes where existing functions are complex and multi-step, a single incoming gene can create a new metabolic linkage. For example, an organism having a multi-step pathway for tyrosine catabolism could acquire a previously unavailable ability to also catabolize phenylalanine through import of a gene encoding phenylalanine hydroxylase (which converts phenylalanine to tyrosine). In cases where the total repertoire of multiple steps that define a novel function are all absent in a given organism, acquisition of that function by LGT would be highly improbable were it not for the existence of operon modules in prokaryotes. Lawrence has proposed, in fact, that gene organization as operons largely exists because of "selfish" properties that operate at the hierarchical level of genes [4, 5]. Although genes of L-tryptophan (Trp) biosynthesis, a classic operon system, are completely dispersed in some prokaryote genomes (e.g. Aquifex, Chlorobium, Wolinella and unicellular cyanobacteria such as Synechococcus, Synechocystis, and Prochlorococcus), they usually exist either as whole-pathway operons or as a combination of several partial-pathway operons .
Approaches for the detection of LGT events are either phylogenetic or parametric. Parametric approaches include the detection of nucleotide composition, dinucleotide frequencies and codon-usage biases in gene segments that are atypical of the recipient genome. Such parametric analysis is well illustrated by a study of the Escherichia coli genome . However, such atypical parametric properties will drift with time toward those of the recipient genome (amelioration). Therefore, parametric analysis is limited to detection of genes that were acquired recently. We have found only a single example where trp genes exhibited parametric features suggesting LGT. In this case a low-GC gene block in Xylella fastidiosa contains seven genes, of which two encode the subunits of anthranilate synthase and one encodes a repressor, TrpR . On the other hand, phylogenetic analysis can detect ancient events of LGT, and this approach has allowed the recognition of a number of whole-pathway and partial-pathway trp operons that were acquired by LGT. Presumably, the initial aberrant parametric properties associated with LGT have been ameliorated. Phylogenetic analysis also is a much more powerful indicator of the likely donor lineage following gene acquisition by LGT.
En blocimportation of primary pathways via LGT
Ubiquitous pathways of primary biosynthesis, having a long history of genomic optimization and metabolic integration, are generally improbable candidates for replacement by LGT . Among the challenges confronting successful LGT are that key regulatory genes may be spatially separated from the structural genes, and promoters and transcription signals vary between bacterial species, as also is the case for elements required for translation.
Exceptions can be envisioned, among them: (i) A gene of primary biosynthesis could, in fact, be synonymous with a gene of antibiotic resistance. Thus, if one or more primary-pathway enzymes is the target of an antimicrobial agent, then genes in nature that encode resistant versions of that enzyme(s) could confer strong selective pressure for replacement. The apparent displacement of an essential enzyme of isoprenoid biosynthesis (hydroxymethylglutaryl coenzyme A reductase) in Archaea by a statin-resistant enzyme of bacterial origin exemplifies this . (ii) If primary pathway genes are lost by deletion, as often occurs with pathogens or symbionts, re-acquisition of the pathway might later become advantageous. In this case, no native pathway having a sophisticated history of amelioration is present to out-compete a pathway of alien origin. In a sense, the recipient has reverted to a pristine evolutionary state that is probably subject to instability and rapid change. (iii) An alien suite of enzymes might provide extraordinary properties of catalysis and/or regulation that are of sufficient selective value to offset a lack of ameliorative history.
The pathway of tryptophan biosynthesis
Key to nomenclaturea used
New gene name
Prior gene name
Protein domain encoded
Anthranilate synthase: aminase subunit (α)
Anthranilate synthase: amidotransferase subunit (β)
Anthranilate phosphoribosyl transferase
Indoleglycerol phosphate synthase
Tryptophan synthase, α subunit
Tryptophan synthase, β subunit
Complete genomes where Trp-pathway genes were lost via reductive evolution
Chlamydophila psittaci a
Wigglesworthia glossinidia b
Yesinia pestis KIM
Trp is biochemically the most expensive of the amino acids synthesized by prokaryotes. Accordingly, it is not surprising that Trp biosynthesis is usually regulated with fine-tuned precision. Different organisms deploy completely different modes of regulation, for example, compare those of E. coli and Bacillus subtilis [15, 16]. Pseudomonas aeruginosa exhibits yet a third distinctive system of regulation, part of which involves an activator gene (trpI) . The sophisticated, complex, and highly distinctive regulation in each of the latter three organisms appears to be of recent origin based upon the relatively narrow clades possessing these particular regulatory systems. This is a very important point in support of the thesis  that modern, sophisticated genomes may be much less prone to displacement by LGT of trp genes than were ancient genomes. In this context, it will be quite important to determine whether organisms that lack known modes of regulation based upon current model organisms really have relatively unsophisticated control mechanisms, or whether unknown regulatory mechanisms are in place. For example, it has been initially surprising that trp genes of Streptomyces coelicolor are not regulated by feedback repression. However, an excellent and detailed study  has demonstrated the existence of regulation that is both growth phase-dependent and growth rate-dependent. This has been attributed to the oligotrophic lifestyle of Streptomyces. It would be interesting to know whether the clade defined by the Streptomyces mode of regulation is also narrow (and therefore of recent origin).
Identical enzyme steps operating in different pathways
We refer to trp genes that produce Trp in general support of protein synthesis as genes of primary metabolism. Those trp genes responsible for the production of intermediates (or Trp) for any other purpose (e.g. as precursors of antibiotics or pigments) are referred to as genes of specialized (or secondary) metabolism. There is ample precedent for maintenance in a given genome of co-existing structural genes whose gene products catalyze the same reaction, but which function in differing temporal or spatial modes in different pathways. Such genes may be differentially regulated by distinctive control mechanisms or mechanisms that accomplish spatial separation. For example, S. coelicolor possesses trp paralog genes of identical catalytic function that exist in separate operons dedicated to primary biosynthesis, on the one hand, or to calcium-dependent antibiotic (CDA) production, on the other hand [6, 19]. It is interesting that, in such cases, one or more of the pathway genes sometimes exist as a single copy and, therefore, must be shared between both functions.
Phylogenetic trees for proteins require a continuum of close relatives
This study is primarily a phylogenetic analysis and depends upon protein trees. However, sequences of proteins such as the Trp enzymes are not nearly as conserved as 16S rRNA, and it is well known that they are of limited value for making phylogenetic inferences over wide phylogenetic distances . On the other hand, for a group of very closely related organisms, the 16S rRNA sequences can be so similar that there is a limited basis for discriminating order of branching. Here some protein trees may yield more refined branching relationships over short phylogenetic distances. Until recently, there have been relatively few sequenced genomes available that would provide a critical mass of closely related organisms as a source of Trp-protein sequences. The sequencing of new genomes is now beginning to provide phylogenetic groups of ever more dense genome representation. Health-related organisms heavily influence priorities, and wide gaps in the currently available tree exist. Phylogenetic regions where genome representation is sparse will undoubtedly persist for many years, and it would be helpful if new genomes for sequencing were selected specifically to fill phylogenetic gaps. One can anticipate that a given protein tree will progressively increase its useful phylogenetic span of discrimination as the gaps between regions of sufficient genome representation are filled.
Additional daunting problems (which are minimal for 16S rRNA trees) assail the phylogenetic validity of protein trees. These are: (i) unrecognized paralogy and (ii) xenology. Ancient paralogs that arose in a common cell will usually have diverged greatly in the contemporary organisms that house them. If one or the other paralog has been lost in various lineages in an erratic fashion and the genealogical history of these losses is not recognized, this situation has been termed "unrecognized paralogy". Differential loss of ancient paralogs in different lineages has the potential to place surviving homolog proteins of distantly related organisms closer to one another than to the surviving proteins of closely related organisms. In this case, xenology can be erroneously inferred. Although LGT has been increasingly recognized for genes in general, genes encoding 16S rRNA are thought to be recalcitrant to LGT (but see Gogarten et al.  and Doolittle .
Congruency and implied cases of LGT
Since particular lineages have optimized Trp biosynthesis to overall demands of primary and secondary metabolism that are both qualitatively and quantitatively individualistic, it seems that abandonment of native trp genes in favor of imported alien genes would not occur very often. If correct, then one would expect that the vertical trace of evolutionary descent for trp genes engaged in primary Trp biosynthesis would be demonstrable, as has indeed been shown following a very comprehensive analysis . In the latter study, individual Trp-protein trees were found to be generally congruent with 16S rRNA trees in subtree regions that were supported by sufficiently dense phylogenetic representation.
Against this background of congruency (and, indeed, enabled because of the overall congruency), several clear examples were found of LGT of whole-pathway trp operons in which all of the native trp genes were displaced, occasionally leaving a surviving remnant behind. In addition, examples were found of LGT of partial-pathway trp operons that either displaced native genes of primary biosynthesis (one case) or became associated with a secondary function. In the literature relevant to LGT there is frequently little evidence about likely donor genomes, and even less indications of direction of transfer . This paper provides detailed information and analysis not given in , and it provides an especially thorough documentation in that the occasional LGT transposition of the entire suite of seven genes responsible for Trp biosynthesis are evaluated. Importantly, we were able to identify donor lineages, to identify a gene remnant of the pre-existing Trp system in three divergent descendants of one of the recipient genomes, and to identify in one genome several evolutionary steps of specific operon perturbation that resumed in the vertical genealogy following LGT.
Seven tryptophan congruency groups (TCGs)
Membership composition of Tryptophan Congruency Groupsa
[Corynebacterium glutamicum] b
[Helicobacter pylori ] b
Buchnera aphidicola c
Blochmannia floridanus c
Anabaena (Nostoc) sp. PCC 7120
Prochlorococcus marinus CCMP 1986 (MED4)
Synechococcus sp. WH8102
Synechocystis sp. PCC 6803
Anabaena variabilis ATCC 29413
Crocosphaera watsonii WH 8501
Gloeobacter violaceus PCC 2471
Prochlorococcus marinus CCMP1375 (SS120)
Prochlorococcus marinus MIT9313
Synechococcus elongatus PCC 7942
Thermosynechococcus elongatus BP-1
The few exceptions that violate the otherwise good congruency between TCGs and 16S rRNA subtree regions point strongly to instances of whole-pathway trp-operon displacement. These include the incongruent presence of Trp-protein concatenates from both Helicobacter pylori and the coryneform bacteria in TCG-1. This indicates that inter-genomic transfer of whole-pathway trp operons occurred, with the donor identifiable as a member of TCG-1.
Orange-highlighted dashed lines in Figures 1 and 2 mark lineages where closely related genomes have not yet been sequenced, that is, where there is sparse genome representation. There is no congruency of the branching positions of these ten orphan organisms when the 16S rRNA tree (Figure 1) is compared with the Trp-protein concatenate tree (Figure 2). It is generally believed that the position of these lineages on the 16S rRNA tree is reliable because 16S rRNA comparisons can discriminate very well over wide phylogenetic distances. In contrast, support for the position of branching obtained for these lineages with respect to the concatenated Trp-protein tree is not significant, that is, low bootstrap values. One can reasonably expect that, as more genomes are sequenced, new TCGs will emerge that include these current orphan organisms. For example, our preliminary data indicate that the Trp proteins of Bacteroides thetaiotaomicron will reside within a new TCG group with Cytophaga hutchinsonii.
Individual Trp-protein trees
Xie et al.  can be consulted for a detailed overview of the Trp pathway reactions (summarized here in Table 1) and for a perspective on the nomenclature issues. Table 4 is a comprehensive listing [see Additional File 1] that contains gi (gene identification) numbers (in bold) that correspond to each of the seven domains asserted to function in primary Trp biosynthesis. Paralog and xenolog gi numbers are presented in regular type. Each gi number is hyperlinked to the corresponding record at NCBI (National Center for Biotechnology Information).
TrpAa is the aminase subunit of anthranilate synthase. Figure 3 shows an individual protein tree for the TrpAa catalytic domain. Forty-seven of these are domain segments of the 47 concatenates making up the tree of Figure 2. An additional three sequences are putative specialized paralogs or xenologs that have been excluded from Figure 2. These include an apparent TrpAa xenolog from X. fastidiosa (Xfa_2) . The origin of Xfa_2 TrpAa is unknown. It is part of an apparent six-gene operon that is divergently oriented next to trpR . It is perhaps suggestive of epsilon proteobacteria origin that Xfa_2 TrpAa exhibits a 100% bootstrap score with the TrpAa of Campylobacter, although the branch distances are great. Except for Xfa TrpAb (which also exhibits a suggestively close relationship with Campylobacter TrpAb; see Figure 5), the Campylobacter genome possesses neither the other four genes of the Xylella operon, nor trpR.
PhnA from P. aeruginosa , together with PhnB (see Figure 5), encode a partial-pathway operon that originated from the enteric lineage. A TrpAa paralog (Sco_CDA) in S. coelicolor functions in the calcium-dependent antibiotic (CDA) pathway .
The •TrpAb domain of TrpAa•TrpAb or TrpAa•TrpAb_phzE fusion proteins showed similar tight clustering when comprehensive TrpAb trees were constructed (data not shown). We have suggested that a number of TrpAa•TrpAb fusions have occurred as independent fusion events, but that the sequences have then converged due to rigid constraints imposed for proper protein-protein interactions of these subunits .
TrpAb is the glutamine-binding subunit of anthranilate synthase responsible for function of the TrpAa/TrpAb complex as an amidotransferase. It is the smallest and least conserved of the seven Trp domains. Sequence features of TrpAb do not distinguish it from from PabAb, an amidotransferase subunit of p-aminobenzoate (PABA) synthase. Usually, however, the functional role of the two members of this homolog group can be deduced by the presence of at least one homolog in a trp operon or a pab operon. In some cases, as with B. subtilis and other members of TCG-6, a single subunit participates in both anthranilate and PABA synthesis. In this case it has dual-pathway function and has the ability to complex with TrpAa, on the one hand, or with PabAa, on the other hand. Most of the TCGs seen in Figure 2 are recognizable on the TrpAb tree in faint outline (supported with only marginal bootstrap values), with TCG-2 and TCG-7 being especially fragmented (Figure 5). The position of the •TrpAb domain from T. fusca is not shown. (The discussion given immediately above for TrpAa• from T. fusca applies here.) Note that TrpAb from Thermotoga, Campylobacter, and a small subclade of TCG-1 represented in Figure 5 by E. coli are fused with TrpB (TrpAb•TrpB). These seem to have arisen as independent fusions, and there is no indication here of convergent evolution (in contrast to the TrpAa•TrpAb fusions discussed above).
PhnB from P. aeruginosa is encoded by a gene within an operon that also contains phnA (Figure 3). The phnA and phnB xenologs in P. aeruginosa are both divergent from the P. aeruginosa trpAa and trpAb homologs that are engaged in primary biosynthesis. The phnA/phnB operon originated by LGT from within TCG-1.
The gene encoding the TrpAb paralog (Sco_CDA) is a member of a large operon (along with TrpAa, TrpB, and TrpD paralogs) that is engaged in "calcium-dependent antibiotic" biosynthesis  and presumably arose intragenomically by gene duplication.
TrpB catalyzes the anthranilate phosphoribosyl transferase step in which anthranilate and PRPP combine to yield phosphoribosyl-anthranilate. In this paper the usage of "TrpB" refers to TrpB_1, the major and ubiquitous subtype species. The distinctly divergent TrpB_2 subtype is narrowly distributed (being cyanobacteria-specific) and usually co-exists with TrpB_1 in cyanobacteria (Table 4 [see Additional File 1]). TrpB_2 might prove to be an ancient paralog in cyanobacteria that has diverged to perform some other phosphoribosyl-transfer function. However, at present the closest homolog of TrpB_2 proteins are the TrpB_1 enzymes that catalyse the second step of Trp biosynthesis.
Figure 6 includes another S. coelicolor paralog dedicated to antibiotic synthesis (Sco_CDA) that can be distinguished from the primary-biosynthesis paralog. Dehalococcoides ethanogenes has two highly divergent paralogs, but Det_2 can be reasonably excluded from the primary-biosynthesis pathway (and, therefore, from the Det concatenated sequence of Figure 2) because it alone is not located within the whole-pathway operon trpAaAbBDC(aroA Iβ)EbEa . Det_2 could be a very ancient paralog, or it may have come from a LGT donor that happens to be relatively distant from any of the homologs included in Figure 6.
In the case of the unicellular cyanobacteria, a single set of completely dispersed trp genes exists in every genome. The Nostoc/Anabaena subclade is more complex, having additional operon-organized trp genes. Since these species also possess the set of dispersed trp genes, it has been concluded that the dispersed trp genes constitute a basic set of conserved genes whose functional role is primary Trp biosynthesis in all cyanobacteria . Accordingly, the trp genes that are present in operon organization (including trpB) have been excluded from the Trp concatenates constructed for Anabaena sp. and Nostoc punctiforme (Figure 2).
As observed above (Figure 5) with the TrpAb• domain of three phylogentically separated TrpAb•TrpB fusions, the corresponding •TrpB domains (marked with asterisks in Figure 6) also show no indication of evolutionary convergence (suggesting that for this particular domain combination, fusion does not impose a set of rigid constraints that must be met in each independent fusion).
TrpC catalyzes the phosphoribosyl-anthranilate isomerase step, which yields 1-(o-carboxyphenylamino)-1-deoxyribulose 5-phosphate (CDRP). Figure 7 shows the TrpC-protein tree on which the seven TCG groups are marked, as well as TrpC from the ten lineages (orange highlight) whose branching positions are not supported by significant bootstrap values. The •TrpC domain of H. pylori is the most divergent sequence within TCG-1. It clusters with TrpC from Campylobacter jejuni with a bootstrap value of 81%. All members of TCG-1 in the TrpC tree are •TrpC domains of TrpD•TrpC fusions, except for C. jejuni TrpC (whose location in TCG-1 is probably artifactual).
TCG-5 organisms (actinomycete bacteria) lack a trpC gene (except, of course, for the horizontally transferred genes of coryneform bacteria). hisA (priA) from these organisms presumably encodes a dual-pathway enzyme that performs the isomerase function in both the histidine and Trp biosynthetic pathways . These dual-function priA genes are listed in the hisA column of Table 4 [see Additional File 1]. Since •trpC in the coryneform bacteria is coordinated with the other trp-operon genes that originated by LGT, their hisA (priA) gene can be considered a remnant in the context of Trp biosynthesis. However, an essential role in histidine biosynthesis has undoubtedly selected for maintenance of hisA. It would be interesting to know whether the trpC function of hisA in the coryneform species of bacteria has deteriorated or not.
HisA catalyses an Amadori rearrangement that is analogous to the TrpC rearrangement. This HisA isomerase catalyses the rearrangement of N'– [(5'-phosphoribosyl) forimino]-5-aminoimidazole-4-carboxamide ribonucleotide to N'– [(5'-phosphoribulosyl) forimino]-5-aminoimidazole-4-carboxamide ribonucleotide.
TrpD catalyzes the reaction of indoleglycerol phosphate synthase, whereby CDRP is converted to indole 3-glycerol phosphate with the release of carbon dioxide (CO2) and water (H2O). The TCG groupings seen in Figure 2 are well articulated in the TrpD-protein tree of Figure 9. Within the cyanobacterial TCG-4, both Anabaena sp. and N. punctiforme possess multiple paralogs that are speculated  to have arisen by two gene duplications that occurred in a common ancester of Anabaena and Nostoc. The TrpD• domains of H. pylori and the coryneform bacteria cluster within TCG-1. S. coelicolor possesses a divergent paralog, Sco_CDA, that is dedicated to antibiotic biosynthesis.
All of the actinomycete bacteria (bottom of Figure 9) possess monofunctional TrpD proteins that cluster within TCG-5. The trpD genes from coryneform bacteria in TCG-5 are remnants that have been functionally displaced by the trpD• domain (located in TCG-1). The latter exists as part of the trpD•trpC fusion within the trp operon imported via LGT. The long branches of the TrpD sequences (Cgl_r and Cdi_r) from the coryneform bacteria (especially from C. glutamicum) suggest strongly that these remnants may have lost catalytic function.
TrpEa is the alpha subunit of tryptophan synthase. It converts indole 3-glycerol phosphate to indole with the release of glyceraldehyde 3-P. It forms a tight complex with TrpEb . The TCG clusters are well defined, as shown in Figure 10. Within cyanobacterial TCG-4, trpEa appears to have been duplicated in a common ancestor of Anabaena and Nostoc, as discussed previously . The close relatives, P. aeruginosa and Pseudomonas syringae, possess a single trpEa gene, which clearly falls into TCG-3, rather than TCG-2. Thus, the native trpEa in these Pseudomonas species must have been displaced by a gene originating within the lineage of TCG-3 organisms.
TrpEb is the beta subunit of tryptophan synthase and condenses indole with L-serine to produce L-tryptophan. Indole is not a free intermediate and passes through a tunnel that is created within the TrpEa/TrpEb complex . A small number of Archaea and Bacteria possess a homolog of TrpEb that falls into a distinct subcluster termed TrpEb_2 . (In this manuscript the major enzyme species, TrpEb_1, may simply be referred to as TrpEb in any context where there would be no confusion.) It has been suggested that the TrpEb_2 species does not form a complex with TrpEa and might have some other stand-alone function, such as that of serine deaminase . Figure 11 shows four bacterial genomes that possess trpEb_2 in addition to trpEb_1, namely Thermotoga, Geobacter, Aquifex, and Chlorobium. These organisms are widely spaced on the 16S rRNA tree (Figure 1). Geobacter has the partial-pathway operon trpAa/trpAb/trpB/trpEb_2/trpC. trpEa and trpEb_1 are located outside the operon and are unlinked to one another. Even though TrpEb_2 is located in the operon, it was excluded from the Gsu concatenate string (Figure 2) because trpEb_1 encodes all of the critical contact residues known to be important for physical association with TrpEa .
The TrpEb_1 sequences in Figure 11 form TCG clusters that are relatively well defined, judging from the excellent conformation with the tree of concatenated Trp sequences (Figure 2). The paralogs present in Anabaena and Nostoc within TCG-4 presumably originated following gene duplication in a common ancestor of these genera . The sequences of TrpEb from Thermomonospora, Streptomyces, and Mycobacteria cluster together as expected for actinomycetes (TCG-5). Exceptions are the TrpEb proteins of the coryneform bacteria, which originated by LGT from a source within TCG-1. Interestingly, after LGT, gene duplication appears to have resulted in a second TrpEb species (Cdi_2) in C. diptheriae. The branching position shown in Figure 11 suggests that the duplication preceded speciation. If so, C. glutamicum has since lost the paralog. Alternatively, rapid divergence associated with probable status of Cdi_2 as a pseudogene might account for the position on the distance tree. It is interesting that it has recently been asserted that duplication occurs more often among laterally transferred genes than for native genes .
Aromatic Congruency Groups implicate two deteriorated TCGs
The chlamydiae have an intact common pathway of aromatic biosynthesis, but none have a complete trp pathway. In one case, Chlamydophila pneumoniae, no trp genes are present. At the other extreme, C. psittaci possesses all trp genes except for trpAa and trpAb. Other species possess an intermediate number of trp genes. Various rationales have been advanced [10, 11] in support of novel functions for these incomplete pathways that have some interesting implications for pathogenesis. Those few trp proteins remaining in the chlamydiae are always very tightly clustered with one another on individual protein trees. We expect that when sufficient related genomes are sequenced, Trp congruency will be established for the chlamydiae. We mainly base this on the fact that trees corresponding to the seven enzymes of the common aromatic pathway yield a very cohesive Aromatic Congruency Group (Xie and Jensen, in progress).
The epsilon proteobacteria, Helicobacter pylori and C. jejuni, were among the early complete and published genomes. Recently, the genomes of H. hepaticus and Wolinella succinogenes have also become available. These four organisms all possess common-aromatic pathway enzymes whose sequences cluster tightly together (Xie and Jensen, in progress). Since the H. pylori trp genes had been replaced via LGT after divergence from their closest neighbor (H. hepaticus), we expected that the Trp proteins of the three remaining epsilon proteobacteria would define a TCG grouping. However, Trp proteins from H. hepaticus, W. succinogenes and C. jejuni are all divergent from one another. They are also divergent from sequences of other organisms. There is, so far, no indication of an LGT history. C. jejuni has a trpAa/trpAb•trpB/trpC/trpEb/trpEa operon (trpD is isolated in an unlinked position). H. hepaticus has spaced the trp genes in three locations, where trpAa/trpAb•trpB, trpD•trpC and trpEb/trpEa reside. It is anomalous and quite possibly symptomatic of ongoing reductive evolution that trpEa and trpEb are divergently oriented. H. hepaticus not only has the trpAb•trpB fusion, but also has a trpD•trpC fusion. The trpD•trpC fusion of H. hepaticus appears to be of independent origin with respect to all of the others in TCG-1, based upon phylogenetic analysis and inter-domain linker analysis. In W. succinogenes the trp genes are all dispersed and none of them are fused. In the future, a core of epsilon proteobacteria may become available as a source of sequences to build a Tryptophan Congruency Group. The organisms currently available have experienced dynamic (and perhaps disruptive) evolutionary events that currently prevent definitive conclusions about the common ancestor.
For comparison, we took a preliminary look at the histidine pathway in these four epsilon proteobacteria. H. pylori has completely lost the histidine pathway, and C. jejuni appears to have displaced its pathway with the histidine operon from the enteric lineage. Genes of histidine biosynthesis are completely dispersed in H. hepaticus and W. succinogenes, and all yield mutually best BLAST hits. Thus, H. hepaticus and W. succinogenes may be core genomes that represent a congruency group with respect to histidine biosynthesis.
Some underlying complexities of Trp-protein concatenates
Convergence of TrpAa•TrpAb fusions
Fusions of general aromatic biosynthetic pathway genes, including those of the Trp pathway, have occurred frequently. It is not uncommon for the same genes in different organisms to have undergone fusion independently. In cases where the catalytic domains function separately, there may be great latitude for successful fusion orientations, for example, aroQ•pheA fusions . In contrast, TrpAb delivers ammonia to the active site of TrpAa in a way that may impose particular demands upon the spatial orientation of the protein-protein interaction operating for the anthranilate synthase complex. Although we suggest that TrpAa•TrpAb fusions have occurred independently at least five times in widely spaced lineages, they all group tightly together in both TrpAa trees (Figure 4) and TrpAb trees (not shown). We believe that this is due to strong selective pressure for convergence. Thus, the individual TrpAa and TrpAb trees place the T. fusca TrpAa and TrpAb proteins in an anomalous group that includes proteins from Legionella, Azospirillum, Brucella (and close relatives) and the Nostoc/Anabaena group of cyanobacteria. This phenomenon of convergence disrupts the congruence of the individual TrpAa tree when these organisms are included, but this is not enough to undermine the proper placement of the T. fusca Trp-protein concatenate within TCG-5.
The bifunctional priAgene of TCG-5 actinomycetes
The enzyme encoded by priA is competent for the isomerase step in both the histidine and Trp pathways of actinomycete bacteria . Thus, PriA sequences would be concatenate components defining not only TCG-5, but also a Histidine Congruency Group (HCG-5) (Figure 8). Presumably the acquisition of the trp operon in coryneform species leaves priA unnecessary for a role in Trp biosynthesis. The dual substrate capabilities of PriA might reflect an ancient state of broad-specificity competence, with hisA and trpC being derived from specialized offspring of gene duplication . On the other hand, because PriA is clearly a homolog of HisA, but an uncertain homolog of TrpC, it is possible that an ancestral trpC was lost and subsequently replaced by a divergent duplicate of hisA. The overall effect of PriA being uniquely present in Trp-protein concatenates of actinomycete bacteria (i.e. being forced to align with TrpC in all other concatenates) is to compress slightly the actinomycete TCG grouping.
The dual-pathway pabAbof TCG-6 bacilli
Very much like the above situation, TCG-6 Bacillus species have lost trpAb from an otherwise complete trp operon. pabAb, located in the pab operon, is competent to function for both Trp and folate biosynthesis (see Xie et al. ). Thus, within the narrow Bacillus clade that includes B. subtilis, B. halodurans and B. stearothermophilus, PabAb sequences help to define not only TCG-6, but also a pending Folate Congruency Group.
Evolutionary scenario for LGT of whole-pathway Trp operons
Figure 12 depicts the evolutionary events proposed to account for the location in both H. pylori and coryneform bacteria of all seven domains that are associated with primary Trp biosynthesis within TCG-1, which otherwise is populated by members of the enteric lineage. Members of TCG-1 all possess whole-pathway operons with a trpD•trpC fusion. As an isolated observation, we cannot absolutely rule out independent fusions of trpD with trpC in Helicobacter and coryneform bacteria, followed by convergent evolution with respect to the remaining trpD•trpC fusions, to explain their joint locations in TCG-1 for TrpD and TrpC trees. However, the identical overall gene order of the 7-gene operon, the phylogenetic positioning of the remaining five protein domains that are not fused, and the existence of congruent trpD remnants in coryneform bacteria are observations that support the conclusion of LGT with conviction. Although the phylogenetic analysis pinpoints the LGT donor of the whole-pathway trp operon to a member within the enteric lineage, the subclade that includes E. coli, Salmonella typhimurium, Klebsiella pneumoniae, and Shigella flexneri can be excluded. This is because the latter four organisms all possess a recent gene fusion (trpAb•trpB) that is absent elsewhere in the enteric lineage (absent also, of course, in H. pylori and coryneform bacteria).
The ancestor of coryneform bacteria also imported a whole-pathway trp operon, but in this case including a coding region for a leader peptide. It is significant that three species of coryneform bacteria possess an unlinked second copy of trpD (denoted TrpD_r), which is a remnant of the native displaced operon that still persists in other actinomycete Bacteria. This conclusion is supported (Figure 9) by the clustering of the stand-alone TrpD_r from both C. glutamicum and C. diptheriae with TrpD proteins from other actinomycete bacteria (species of Bifidobacteria, Mycobacteria, Thermomonospora and Streptomyces) in TCG-5. The latter organisms possess trpD within a partial-pathway trp operon having the order trpAaDEbEa . The elongated branches seen for TrpD_r in Cgl and Cdi in Figure 9 are consistent with drift of the remnant genes from their previous functional role, due to the lack of selective pressure that followed functional replacement by TrpD• of the LGT operon.
It is interesting that C. diptheriae exhibits evidence of changes to its trp operon that are both relatively recent and disruptive (Figure 12). A duplicate of trpEb has been inserted at the beginning of the operon between trpL (encoding the leader peptide) and trpAa. This TrpEb protein may not be functional, considering that invariant residue 162G (S. typhimurium numbering) has been changed to 162E and invariant residue 167K has been changed to 167S, as pointed out earlier . Residue 167K is crucial to the formation of an intersubunit salt bridge with 56D of TrpEa .
Another post-LGT event in C. diptheriae has been the insertion of two genes of D-pantothenate biosynthesis between trpD•C and trpEb. In C. glutamicum the panB/panC operon is remote from the trp operon and has been well characterized . In C. diptheriae the insertion between trpD•C and trpEb has been in a scrambled orientation such that panC and panB are transcribed convergently, hence no longer being an operon. The transcription of panB in the opposite direction as the flanking genes of trp biosynthesis should prevent the formation of a single transcript from the trp genes. It would appear that the original xenologous trp operon of C. diptheriae has been split such that trpAaAbBD•C and trpEbEa would be separately transcribed. This may illustrate the phenomenon of reductive evolution for a pathogen in the process of discarding un-needed genes.
All elements of operon regulation were not transferred
The trp operon of E. coli (and presumably other members of the enteric clade) is subject to control at several levels [15, 16, 27, 33]. These include (i) allosteric control of anthranilate synthase, whereby Trp acts as a potent feedback inhibitor; and (ii) an attenuator mechanism that is mediated by a Trp-rich leader peptide (encoded by trpL), and repression control mediated by trpR, which binds Trp as a corepressor moiety. Because sensitivity to feedback inhibition is built into the allosteric domain of TrpAa, and because trpL is located immediately upstream of the E. coli trp operon, it is not surprising that recipient organisms such as the coryneform bacteria possess both of the latter regulatory features. However, trpR is distant from the operon and was not co-transferred with the operon.
Perhaps the contemporary trp operon acquired by LGT offered selective advantages to the ancestor of coryneform bacteria, but Trp regulation in coryneform bacteria would appear to be relatively unsophisticated without trpR. In E. coli the impact of attenuation is relatively weak compared with the impact of repression [16, 27]. Repression detects free Trp whereas attenuation detects uncharged tRNATrp. Since free Trp concentrations can be fairly low and still maintain highly charged tRNATrp, trpR-mediated repression acts over a large range of expression. Relief from attenuation ensues after maximal derepression has occurred. In E. coli, trpR also functions beyond the Trp pathway in that it binds to an operator for an initial gene of aromatic biosynthesis.
LGT of trpLin coryneform bacteria
In the case of H. pylori, the trp operon was either acquired without trpL, or trpL was soon lost. Thus, the trp pathway of H. pylori appears to lack regulation by both attenuation and repression. Assuming that the displaced pathway of H. pylori was similar to that of the contemporary C. jejuni or other epsilon-proteobacteria, it would be interesting to compare the relative efficiencies of the modern Trp pathways in these fairly close relatives.
LGT of partial-pathway Trp operons
P. aeruginosa possesses a partial-pathway operon consisting of genes encoding the two subunits of tryptophan synthase that are of xenologous origin. These trpEb/trpEa genes are engaged in primary biosynthesis. The donor is clearly a member of the TCG-3 clade. Thus, in the scenario considered  for gene separation of the ancestral trpC/trpEb/trpEa operon to yield a stand-alone trpC gene and a trpEb/trpEa operon, the native trpEb/trpEa must have been displaced by LGT, after the intragenomic translocation. Because the regulatory gene, trpI, is adjacent in divergent orientation to trpEb in the P. aeruginosa/P. syringae clade (and three other close relatives; see asterisks in Table 3), it will be interesting if any genomic members of TCG-3 to be sequenced in the future prove to have trpI adjacent to a trpEb/trpEa operon. (However, TrpI is a member of the ubiquitous and large family of LysR transcriptional activators, and could easily have emerged recently in the P. aeruginosa clade following gene duplication.)
P. aeruginosa also possesses a partial-pathway trp operon encoding the two subunits of anthranilate synthase. Phylogenetic analysis indicates that these are trpAa trpAb xenologs that originated from within the enteric lineage. As with the whole-pathway operons just discussed, the E. coli/S. typhimurium/K. pneumoniae/S. flexneri clade can be excluded as the specific donor because this clade possesses a trpAb•trpB fusion. This xenologous two-gene operon of P. aeruginosa (originally denoted phnA and phnB by Crawford et al. [25, 37]) has not displaced the corresponding trp genes that are engaged in primary biosynthesis. Although phnA and phnB were originally thought to function in phenazine production, as the naming implies, phenazine pigments are not derived from anthranilate . Instead, the xenolog pair appear to have an unknown specialized function that is geared to stationary-phase physiology . The LGT acquisition of the phnA/phnB xenolog operon by P. aeruginosa occurred after its divergence from close relatives such as P. putida, P. fluorescens, P. syringae and Azotobacter vinelandii because none of these relatives have the xenologous phnA/phnB operon. Hence, this acquisition appears to be quite recent.
Species of Xylella also possess a xenologous two-gene trpAa/trpAb partial-pathway trp operon that appears to have a specialized function that is distinct from homologs engaged in primary Trp biosynthesis. Xie et al.  have suggested that the gene denoted acl probably encodes an aryl-CoA ligase. Since acl appears to exist within the Xylella trpAa/trpAb operon, its gene product may have the specificity of anthranilate-CoA ligase. Activated anthranilate may then function as a key precursor for production of antibiotic, siderophore, and so on. Note that, if this is correct, the primary trp-pathway genes trpB/trpC/trpD/trpEa/trpEb are irrelevant to the specialized pathway. Thus, reference to the xenologous trpAa/trpAb pair as a "partial-pathway operon" could be somewhat a misnomer, and perhaps "hybrid operon" might prove to be more apt. The origin of the Xylella trpAa/trpAb genes might have been from a close relative of C. jejuni because the best matches of both genes are with C. jejuni. It is also suggestive that these genes have a low-GC ratio (about 38%), compared with a genomic GC ratio of 30.2% for C. jejuni. C. jejuni cannot have been the direct donor, however, because it possesses a trpAb•trpB fusion.
The overall impact of LGT in Bacteria and Archaea is currently a highly contentious issue. Critical reviews written from completely different viewpoints are recommended for a sense of the status of the ongoing debate, as well as for a substantial listing of key references [39, 40]. Gogarten et al.  have summarized a contemporary rationale to describe the evolutionary process as a phylogenetic "synthesis" that could integrate a traditional tree-like behavior (vertical descent of genes) and web-like, reticulate behavior (LGT). The latter paper follows up on a balanced and highly insightful review produced by Doolittle in 1999 . In an essay by Martin  about the extent to which bacterial chromosomes might be mosaic, it was emphasized that "careful gene-by-gene phylogenetic comparisons in addition to genome-by-genome comparisons.....are needed." We consider the Trp system to be a particularly apt choice for gaining detailed insight into the relative contributions of lateral and vertical events in the evolutionary history of a major segment of primary metabolism. On the one hand, genes of Trp biosynthesis are expected to represent a level and type of fundamental metabolism that is least prone to LGT if it is correct that a "core" of genes exists that is relatively recalcitrant to LGT [23, 42]. On the other hand, the wide distribution of intact whole-pathway trp operons should facilitate the probability of LGT (at least the initial acquisition event). For the Trp pathway, we conclude that events of LGT and paralogy do not obliterate the vertical trace of evolutionary history. This view is reinforced by the contention that whole-genome trees (mean pairwise similarities between shared genomic proteins) have largely converged on the rRNA-sequence tree . The latter work was preceded by the gene-content "trees" introduced by Snel et al. . Also, see Eisen .
Gogarten et al.  have discussed the radical possibility that rRNA genes are highly mosaic and are so useful for prokaryotic taxonomy "precisely because they are mosaics and reflect the mosaic character of the genome as a whole". According to this view, a vertical trace of genealogical history will not be found, because it only exists in short jumps. (If so, the two whole-pathway LGT events described here reflect LGT events that stand out against the overall mosaic trends.)
The above caveat aside, 16S rRNA trees do appear to provide a reasonable guide to the vertical trace of evolutionary descent . If so, the mapping of the Trp-pathway system upon the 16S rRNA tree exemplifies a case where the evolutionary history can indeed be tracked as a vertical genealogy that features some intriguing reticulate relationships. Proteins of the Trp pathway generally exhibit a genealogy that is parallel with the 16S rRNA tree. In the case of Helicobacter and Corynebacterium species, the genomes are a mosaic with respect to Trp genes. This means that the history of the trp genes in the latter organisms, instead of being that shared with their closest relatives, is the same as that of enteric bacteria (the donor lineage) up to the time of LGT. Following this time a new vertical genealogical progression has begun. Each organism (lineage) can be envisioned to possess an individualistic repertoire of chimeric features that have been assimilated into a recognizable skeleton of vertical events. It would seem that the evolutionary history of the primary pathway of Trp biosynthesis parallels the organismal phylogeny in most organisms. This presently includes all of the organisms in Figure 2 that belong to TCG groupings, except for H. pylori and the coryneform bacteria. Mosaicism can occasionally exist at a level of individual trp genes, as illustrated by P. aeruginosa and its closest relatives. Here the history of trpAa/trpAb/trpB/trpC/trpD (primary biosynthesis) parallels the organismal phylogeny, but trpEb/trpEa exhibit a reticulate genealogy.
It will be important to obtain other analyses of similar detail for other pathways to assess to what extent the evolutionary process that underpins Trp biosynthesis reflects the general process in other pathways. We have already seen in preliminary work that construction of similar congruency groups pertinent to common aromatic-pathway biosynthesis, folate biosynthesis, and histidine biosynthesis has good potential to expand the analysis. It seems probable that there are no prokaryote species whose evolutionary history exclusively involved a vertical line of descent for all of its genes. Lawrence  has written a cogent essay in support of the view that the Linnean paradigm of hierarchical descent fails to describe the evolution of prokaryotes because LGT occurs "at all levels of taxonomic inclusiveness". However, Woese et al.  argue that the dynamic of LGT has become progressively diminished through time as simple, primitive early cells became complex and refined. This is in accord with our thesis  that early and simple trp-gene assemblages may have been unstable until their progression to complexities of regulation and to establishment of individualistic metabolic ties that conferred increasing operon stability.
Refer to the Gold Genomes Online Database  for a list of published complete genomes and links to the corresponding references. Synonymous names include Rhizobium meliloti = Sinorhizobium meliloti; Corynebacterium glutamicum = Brevibacterium lactofermentum = Brevibacterium flavum = Brevibacterium divaricatum; Rhizobium meliloti = Sinorhizobium meliloti; Rhizobium loti = Mesorhizobium loti; Thermomonospora fusca = Thermobifida fusca; Chlamydophila psittaci = Chlamydophila caviae; Thiobacillus ferrooxidans = Acidithiobacillus ferrooxidans; Shewanella putrefaciens = Shewanella oneidensis; Sphingomonas aromaticivorans = Novosphingobium aromativorans.
16S rRNA phylogenetic trees
Protein phylogenetic trees
Unrooted phylogenetic trees were derived from ClustalW alignment [52, 53]. The neighbor-joining and Fitch  programs were employed to obtain distance-based trees. The distance matrix was obtained using Protdist with a Dayhoff Pam matrix. The Seqboot and Consense programs were then used to assess the statistical strength of the tree using bootstrap resampling. Neighbor-joining and Fitch trees yielded similar clusters and arrangement of taxa within them. Bootstrap values indicate the number of times a node was supported in 100 resampling replications.
Concatenated sequences of Trp-pathway proteins
Multiple sequence alignments were derived by input of the indicated homolog amino acid sequences into the ClustalW program (Version 1.4) [52, 53]. Manual alignment adjustments were made as needed with the assistance of the BioEdit multiple alignment tool of Hall .
After each of the individual Trp-protein domain alignments were generated, both N-terminal and C-terminal unaligned regions were trimmed manually through visual inspection. Then the seven protein domains responsible for the Trp pathway of biosynthesis were concatenated in the order: TrpAa/TrpAb/TrpB/TrpC/TrpD/TrpEa/TrpEb. The resulting concatenated multiple alignment was used as input for generation of a phylogenetic tree using the program package PHYLIP .
Analysis of raw DNA sequence data
Raw DNA contig sequences available from NCBI  and the TIGR Unfinished Microbial Genomes database  were screened using the built-in BLAST service. The protein sequences from GenBank were used as query entries. The BLAST 2.0 and the ORF Finder (Open Reading Frame Finder) offered by NCBI  were used to locate open reading frames and to confirm the similarity search result of the raw sequence.
Analysis of fusion proteins
Fusion protein sequences from GenBank and NCBI Microbial Genomes Blast Databases  were screened using the BLAST  program. Multiple alignments were obtained by input of single-domain and fusion-protein sequences into the ClustalW [52, 53] program (version 1.4).
This is Florida Agricultural Experiment Station Journal Series No. R-09802. R Jensen thanks the National Library of Medicine (Grant G13 LM008297) for partial support. Some of the preliminary sequence data were obtained from the Institute for Genomic Research , National Center for Biotechnology Information , and Integrated Genomics, Inc. . We thank Charles Yanofsky and Ross Overbeek for their critiques of an early version of this manuscript.
- Lawrence JG, Ochman H: Molecular archaeology of the Escherichia coli genome. Proc Natl Acad Sci USA. 1998, 95: 9413-9417. 10.1073/pnas.95.16.9413.PubMed CentralView ArticlePubMedGoogle Scholar
- Lawrence JG, Ochman H: Amelioration of bacterial genomes: rate of change and exchange. J Mol Evol. 1997, 44: 383-397.View ArticlePubMedGoogle Scholar
- Song J, Ware A, Liu S-L: Wavelet to predict bacterial ori and ter: a tendency towards a physical balance. BMC Genomics. 2003, 4: 17-10.1186/1471-2164-4-17.PubMed CentralView ArticlePubMedGoogle Scholar
- Lawrence J: Selfish operons: the evolutionary impact of gene clustering in prokaryotes and eukaryotes. Curr Opin Genet Dev. 1999, 9: 642-648. 10.1016/S0959-437X(99)00025-8.View ArticlePubMedGoogle Scholar
- Lawrence JG: Selfish operons and speciation. Trends Microbiol. 1997, 5: 355-359. 10.1016/S0966-842X(97)01110-4.View ArticlePubMedGoogle Scholar
- Xie G, Keyhani NO, Bonner CA, Jensen RA: The ancient origin of the tryptophan operon and tracking the subsequent dynamics of evolutionary change. Microbiol Mol Biol Rev. 2003, 67: 303-342. 10.1128/MMBR.67.3.303-342.2003.PubMed CentralView ArticlePubMedGoogle Scholar
- Xie G, Bonner CA, Brettin T, Gottardo R, Keyhani NO, Jensen RA: Lateral gene transfer and ancient paralogy of operons containing redundant copies of tryptophan-pathway genes in Xylella species and in heterocystous cyanobacteria. Genome Biol. 2003, 4: R-14. 10.1186/gb-2003-4-2-r14.View ArticleGoogle Scholar
- Lawrence JG: Gene transfer, speciation, and the evolution of bacterial genomes. Curr Opin Microbiol. 1999, 2: 519-523. 10.1016/S1369-5274(99)00010-7.View ArticlePubMedGoogle Scholar
- Boucher Y, Huber H, L'Haridon S, Stetter KO, Doolittle WF: Bacterial origin for the isoprenoid biosynthesis enzyme HMG-CoA reductase of the archaeal orders Thermoplasmatales and Archaeoglobales. Mol Biol Evol. 2001, 18: 1378-1388.View ArticlePubMedGoogle Scholar
- Xie G, Bonner CA, Jensen RA: Dynamic diversity of the tryptophan pathway in chlamydiae: reductive evolution and a novel operon for tryptophan recapture. Genome Biol. 2002, 3: research0051.1-0051.17. 10.1186/gb-2002-3-9-research0051.View ArticleGoogle Scholar
- Fehlner-Gardniner C, Roshick C, Carlson JH, Hughes S, Belland RJ, Caldwell HD, McClarty G: Molecular basis defining human Chlamydia trachomatis tissue tropism: a possible role for tryptophan synthase. J Biol Chem. 2002, 277: 26893-26903. 10.1074/jbc.M203937200.View ArticleGoogle Scholar
- Wood H, Roshick C, McClarty G: Tryptophan recycling is responsible for the interferon-gamma resistance of Chlamydia psittaci GPIC in indoleamine dioxygenase-expressing host cells. Mol Microbiol. 2004, 52: 903-916. 10.1111/j.1365-2958.2004.04029.x.View ArticlePubMedGoogle Scholar
- Barona-Gómez F, Hodgson DA: Occurrence of a putative ancient-like isomerase involved in histidine and tryptophan biosynthesis. EMBO Reports. 2003, 4: 296-300. 10.1038/sj.embor.embor771.PubMed CentralView ArticlePubMedGoogle Scholar
- Osterman A, Overbeek R: Missing genes in metabolic pathways: a comparative genomics approach. Curr Opin Chem Biol. 2003, 7: 238-251. 10.1016/S1367-5931(03)00027-9.View ArticlePubMedGoogle Scholar
- Yanofsky C: Advancing our knowledge in biochemistry, genetics, and microbiology through studies on tryptophan metabolism. Ann Rev Biochem. 2001, 70: 1-37. 10.1146/annurev.biochem.70.1.1.View ArticlePubMedGoogle Scholar
- Yanofsky C: Reflections: using studies on tryptophan metabolism to answer basic biological questions. J Biol Chem. 2003, 278: 10859-10878. 10.1074/jbc.X200012200.View ArticlePubMedGoogle Scholar
- Auerbach S, Gao J, Gussin GN: Nucleotide sequences of the trpI, trpB, and trpA genes of Pseudomonas syringae: positive control unique to fluorescent pseudomonads. Gene. 1993, 123: 25-32. 10.1016/0378-1119(93)90534-A.View ArticlePubMedGoogle Scholar
- Hu DS-J, Hood DW, Heidstra R, Hodgson D: The expression of the trpD, trpC and trpBA genes of Streptomyces coelicolor A3(2) is regulated by growth rate and growth phase but not by feedback repression. Mol Microbiol. 1999, 32: 869-880. 10.1046/j.1365-2958.1999.01407.x.View ArticlePubMedGoogle Scholar
- Ryding NJ, Anderson TB, Champness WC: Regulation of the Streptomyces coelicolor calcium-dependent antibiotic by absA, encoding a cluster-linked two-component system. J Bacteriol. 2002, 184: 794-805.PubMed CentralView ArticlePubMedGoogle Scholar
- Brown JR, Doolittle WF: Archaea and the prokaryote-to-eukaryote transition. Microbiol Mol Biol Rev. 1997, 61: 456-502.PubMed CentralPubMedGoogle Scholar
- Gogarten JP, Doolittle WF, Lawrence JG: Prokaryotic evolution in light of gene transfer. Mol Biol Evol. 2002, 19: 2226-2238.View ArticlePubMedGoogle Scholar
- Doolittle WF: Phylogenetic classification and the universal tree. Science. 1999, 284: 2124-2128. 10.1126/science.284.5423.2124.View ArticlePubMedGoogle Scholar
- Eisen JA: Horizontal gene transfer among microbial genomes: new insights from complete genome analysis. Curr Opin Genet Dev. 2000, 10: 606-611. 10.1016/S0959-437X(00)00143-X.View ArticlePubMedGoogle Scholar
- Eisen JA: Phylogenomics: improving functional predictions for uncharacterized genes by evolutionary analysis. Genome Res. 1998, 8: 163-167.View ArticlePubMedGoogle Scholar
- Crawford IP, Milkman R: Orthologous and paralogous divergence, reticulate evolution, and lateral gene transfer in bacterial trp genes. In: Evolution at the Molecular Level. Edited by: Selander RK, Clark AG, Whittam TS. 1990, Sunderland, Mass.: Sinauer & Assoc, 77-95.Google Scholar
- McDonald MD, Mavrodi DV, Thomashow LS, Floss HG: Phenazine biosynthesis in Pseudomonas fluorescens : branchpoint from the primary shikimate biosynthetic pathway and role of phenazine-1,6-dicarboxylic acid. J Am Chem Soc. 2001, 123: 9459-9460. 10.1021/ja011243+.View ArticlePubMedGoogle Scholar
- Yanofsky C, Miles E, Bauerle R, Kirschner K: Trp Operon. 1999, New York: John Wiley & Sons, Inc, 4:Google Scholar
- Xie G, Forst C, Bonner CA, Jensen RA: Significance of two distinct types of tryptophan synthase beta chain in Bacteria, Archaea and higher plants. Genome Biology. 2001, 3: research005.1-005.13. 10.1186/gb-2001-3-1-research0004.View ArticleGoogle Scholar
- Hooper SD, Berg OG: Duplication is more common among laterally transferred genes than among indigenous genes. Genome Biol. 2003, 4: R48-10.1186/gb-2003-4-8-r48.PubMed CentralView ArticlePubMedGoogle Scholar
- Calhoun DH, Bonner CA, Gu W, Xie G, Jensen RA: The emerging periplasm-localized subclass of AroQ chorismate mutases, exemplified by those from Salmonella typhimurium and Pseudomonas aeruginosa. Genome Biol. 2001, 2: research0030.1-0030.16. 10.1186/gb-2001-2-8-research0030.View ArticleGoogle Scholar
- Miles EW, Bauerle R, Ahmed SA: Tryptophan synthase from Escherichia coli and Salmonella typhimurium. Methods Enzymol. 1987, 142: 398-414. 10.1016/S0076-6879(87)42051-X.View ArticlePubMedGoogle Scholar
- Sahm H, Eggeling L: D-Pantothenate synthesis in Corynebacterium glutamicum and use of panBC and genes encoding L-valine synthesis for D-pantothenate overproduction. Appl Environ Microbiol. 1999, 65: 1973-1979.PubMed CentralPubMedGoogle Scholar
- Xiu ZI, Chang ZY, Zeng AP: Nonlinear dynamics of regulation of bacterial trp operon: model analysis of integrated effects of repression, feedback inhibition, and attenuation. Biotech. 2002, 18: 686-693. 10.1021/bp020052n.Google Scholar
- Matsui K, Sano K, Ohtsubo E: Sequence analysis of the Brevibacterium lactofermentum trp operon. Mol Gen Genet. 1987, 209: 299-305.View ArticlePubMedGoogle Scholar
- Heery DM, Dunican LK: Cloning of the trp gene cluster from a tryptophan-hyperproducing strain of Corynebacterium glutamicum: identification of a mutation in the trp leader sequence. Appl Envir Microbiol. 1993, 59: 791-799.Google Scholar
- Su YC, Chen SL: Cloning of the tryptophan operon of Brevibacterium divaricatum and its expression in E. coli. Proc Natl Acad Sci USA. 1996, 20: 87-91.Google Scholar
- Essar DW, Eberly L, Hadero A, Crawford IP: Identification and characterization of genes for a second anthranilate synthase in Pseudomonas aeruginosa : interchangeability of the two anthranilate synthases and evolutionary implications. J Bacteriol. 1990, 172: 884-900.PubMed CentralPubMedGoogle Scholar
- Mavrodi DV, Bonsall RF, Delaney SM, Soule MJ, Phillips G, Thomashow LS: Functional analysis of genes for biosynthesis of pyocyamin and phenazine-1-carboxamide from Pseudomonas aeruginosa PAO1. J Bacteriol. 2001, 183: 6454-6465. 10.1128/JB.183.21.6454-6465.2001.PubMed CentralView ArticlePubMedGoogle Scholar
- Kurland CG, Canback B, Berg OG: Horizontal gene transfer: A critical view. Proc Natl Acad Sci USA. 2003, 100: 9658-9662. 10.1073/pnas.1632870100.PubMed CentralView ArticlePubMedGoogle Scholar
- Lawrence JG, Henderickson H: Lateral gene transfer: when will adolescence end?. Mol Microbiol. 2003, 50: 739-749. 10.1046/j.1365-2958.2003.03778.x.View ArticlePubMedGoogle Scholar
- Martin W: Mosaic bacterial chromosomes: a challenge en route to a tree of genomes. Bioessays. 1999, 21: 99-104. 10.1002/(SICI)1521-1878(199902)21:2<99::AID-BIES3>3.3.CO;2-2.View ArticlePubMedGoogle Scholar
- Lerat E, Daubin V, Moran NA: From gene trees to organismal phylogeny in prokaryotes: the case of the alpha-proteobacteria. PLoS Biology. 2003, 1: E1-10.1371/journal.pbio.0000019.View ArticleGoogle Scholar
- Clarke AR, Beiko RG, Ragan MA, Charlebois RL: Inferring genome trees by using a filter to eliminate phylogenetically discordant sequences and a distance matrix based on mean normalized BLAST scores. J Bacteriol. 2002, 184: 2072-2080. 10.1128/JB.184.8.2072-2080.2002.PubMed CentralView ArticlePubMedGoogle Scholar
- Snel B, Bork P, Huynen MA: Genome phylogeny based on gene content. Nat Genet. 1999, 21: 108-110. 10.1038/5052.View ArticlePubMedGoogle Scholar
- Eisen JA: Assessing evolutionary relationships among microbes from whole-genome analysis. Curr Opin Microbiol. 2000, 3: 475-480. 10.1016/S1369-5274(00)00125-9.View ArticlePubMedGoogle Scholar
- Woese CR: Interpreting the universal phylogenetic tree. Proc Natl Acad Sci USA. 2000, 97: 8392-8396. 10.1073/pnas.97.15.8392.PubMed CentralView ArticlePubMedGoogle Scholar
- Lawrence J: Gene transfer in Bacteria : speciation without species?. Theoret Populat Biol. 2002, 61: 449-460. 10.1006/tpbi.2002.1587.View ArticleGoogle Scholar
- Woese CR, Olsen GJ, Ibba M, Soll D: Aminoacyl-tRNA synthetases, the genetic code, and the evolutionary process. Microbiol Mol Biol Rev. 2000, 64: 202-236. 10.1128/MMBR.64.1.202-236.2000.PubMed CentralView ArticlePubMedGoogle Scholar
- Gold Genomes OnLine Database. [http://www.genomesonline.org/]
- Maidak BL, Cole JR, Lilburn TG, Parker CT, Saxman PR, Farris RJ, Garrity GM, Olsen GJ, Schmidt TM, Tiedje JM: The RDP-II (Ribosomal Database Project). Nucl Acids Res. 2001, 29: 173-17. 10.1093/nar/29.1.173.PubMed CentralView ArticlePubMedGoogle Scholar
- Ribosomal Database Project II. [http://rdp.cme.msu.edu/html]
- Thompson JD, Higgins DG, Gibson TJ: Clustal W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucl Acids Res. 1994, 22: 4673-4680.PubMed CentralView ArticlePubMedGoogle Scholar
- ClustalW, version 1.4. [http://www.ebi.ac.uk/clustalw]
- Fitch WM: Toward defining the course of evolution: minimum change for a specific tree topology. Syst Zool. 1971, 20: 406-416.View ArticleGoogle Scholar
- Hall TA: Biological sequence alignment editor for Windows 95/98/NT, 4.8.6. 2000, [http://www.mbio.ncsu.edu/BioEdit/bioedit.html]Google Scholar
- Felsenstein J: PHYLIP-Phylogeny Inference Package (version 3.2). Cladistics. 1989, 5: 164-166.Google Scholar
- National Center for Biotechnology Information. [http://www.ncbi.nlm.nih.gov/Microb_blast/unfinishedgemome.html.old/]
- The Institute for Genomic Research. [http://www.tigr.org/tdb/ufmg/]
- National Center for Biotechnology Information. [http://www.ncbi.nlm.nih.gov/]
- BLAST. [http://www.ncbi.nlm.nih.gov/blast/]
- Integrated Genomics, Inc. [http://wit.integratedgenomics.com/ERGO]
This article is published under license to BioMed Central Ltd. This is an Open Access article: verbatim copying and redistribution of this article are permitted in all media for any purpose, provided this notice is preserved along with the article's original URL.