MBE Advance Access originally published online on July 31, 2008
Molecular Biology and Evolution 2008 25(10):2221-2232; doi:10.1093/molbev/msn170
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Research Articles |
A Phylogenomic Analysis of the Shikimate Dehydrogenases Reveals Broadscale Functional Diversification and Identifies One Functionally Distinct Subclass
,1
* Department of Pathology, Children's Hospital Boston, Harvard Medical School, Boston, MA
Department of Ecology and Evolutionary Biology, University of Arizona
Department of Cell and Systems Biology, University of Toronto, Toronto, Ontario, Canada
Centre for the Analysis of Genome Evolution and Function, University of Toronto, Toronto, Ontario, Canada
E-mail: sasha.singh{at}childrens.harvard.edu.
 |
Abstract |
|---|
 TOP Abstract IntroductionMaterials and MethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
The shikimate dehydrogenases (SDH) represent a widely distributed enzyme family with an essential role in secondary metabolism. This superfamily had been previously subdivided into 4 enzyme groups (AroE, YdiB, SdhL, and RifI), which show clear biochemical and functional differences ranging from amino acid biosynthesis to antibiotic production. Despite the importance of this group, little is known about how such essential enzymatic functions can evolve and diversify. We dissected the enzyme superfamily with a phylogenomic analysis of
250 fully sequenced genomes, making use of previously characterized representatives from each enzyme class, and the key substrate-binding residues known to distinguish substrate specificity. We identified 5 major evolutionary and functional SDH subgroups and several other potentially unique functional classes within this complex enzyme family and then validated the functional distinctiveness of each group by characterizing the 5 SDH homologs found in Pseudomonas putida KT2440 biochemically. We identified an entirely novel functionally distinct subgroup, which we designated Ael1 (AroE-like1) and also delineated a new group of shikimate/quinate dehydrogenases (YdiB2), which is phylogenetically distinct from the previously described Escherichia coli YdiB. The combination of biochemical, phylogenetic, and genomic approaches has revealed the broad extent to which the SDH enzyme superfamily has diversified. Five functional groups were validated with the potential for at least 5 additional subgroups. Our analysis also identified a new SDH functional group, which appears to have evolved recently from an ancestral AroE, illustrating a very prominent role of horizontal transmission and neofunctionalizaton in the evolutionary and functional diversification of this enzyme family.
Key Words: shikimate dehydrogenase • evolution • phylogeny • diversification • phylogenomics • neofunctionalization • motifs
|
Introduction |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
The emergence of highly resistant microbial pathogens has prompted widespread efforts to identify suitable novel targets for antimicrobial drug development. The recent surge in microbial genome sequencing and comparative genomic analyses has facilitated the identification of numerous essential enzymes in critical metabolic pathways in bacteria that may be ideal targets for such drugs. Shikimate dehydrogenase (SDH), the fourth enzyme in the shikimate pathway, has been studied extensively for its role in the biosynthesis of aromatic compounds in bacteria, fungi, and plants (Herrmann and Weaver 1999
). The prototypical SDH, AroE, catalyzes the reduced nicotinamide adenine dinucleotide phosphate (NADPH)–dependent reduction of dehydroshikimate to shikimate (fig. 1) and has garnered significant attention in recent years as an antimicrobial target due to its conserved and rather essential nature (Fonseca et al. 2006). Indeed, Escherichia coli aroE loss-of-function mutants are auxotrophic for the 3 aromatic amino acids, indicating the essentiality of AroE in basic metabolism (Pittard and Wallace 1966a, 1966b).
|
In recent years, however, increasing genomic data have revealed that some bacteria harbor additional homologs of SDH, including YdiB, RifI, and SdhL (Michel et al. 2003
; Ye et al. 2003; Guo and Frost 2004), each of which has distinctive biochemical activities and cellular functions. YdiB, previously reported to be paralogous to the AroE enzyme, has been predicted to function as a quinate/shikimate dehydrogenase, based on its ability to accept either quinate or shikimate as substrates (Michel et al. 2003). As such, the enzyme may play a role in partitioning shikimate toward quinate metabolism, as shown for enzymes characterized in fungi and plants (Giles et al. 1985; Kang and Scheibe 1993). The SDH homolog RifI, identified in Amycolatopsis mediterranei (Yu et al. 2001), is part of the aminoshikimate biosynthetic pathway that leads to the biosynthesis of the antibiotic rifamycin B (Yu et al. 2001; Guo and Frost 2002, 2004). Interestingly, RifI does not appear to be required for rifamycin biosynthesis in A. mediterranei (Yu et al. 2001) but has been proposed to bind and catalyze the reversible reduction of either 5-amino-dehydroshikimate or 5-amino-dehydroquinate (Yu et al. 2001; Guo and Frost 2004). Evidence of its activity was demonstrated by the ability of E. coli, which lacks the aminoshikimate pathway, to metabolize 5-amino-shikimate when recombinantly expressing the A. mediterranei rifI (Guo and Frost 2004). Likewise, the more recently described SDH-like protein from Haemophilus influenzae, SdhL (Singh et al. 2005), exhibits 1,000-fold less activity than the E. coli AroE with shikimate as a substrate; moreover, it does not catalyze the oxidation of quinate (Singh et al. 2005), suggesting that shikimate and quinate are not its preferred substrates. The presence of numerous SDH homologs among many different bacterial lineages with differing substrate specificities raises questions relating to the diversification of this enzyme family and the specific biological roles of its members. What selective pressures have driven the overall diversification of this enzyme family? Can SDH homologs substitute functionally for each other? How many independent times has SDH diversification occurred? What is the role of horizontal gene transfer in the dissemination of SDH homologs? Can the suite of SDH homologs carried by a lineage provide insight into the biology of the organism? The recent surge in the availability of completed bacterial genomes has provided an excellent opportunity to examine the distribution and evolution of SDH homologs and may inevitably have significant repercussions for the development of effective antimicrobials.
To identify and characterize the main functional SDH subgroups and provide insight into the mechanisms of SDH enzyme evolution and extent of diversification, we used a phylogenomic approach that integrated phylogenetics, genomics (Eisen 1998; Eisen and Hanawalt 1999; Eisen and Wu 2002), and biochemistry. Phylogenomics uses phylogenetic methods to infer function by identifying putative orthologs (homologous sequences directly related through speciation) among diverse species and assumes that orthologs generally maintain common functions through speciation (Eisen 1998; Eisen and Hanawalt 1999; Eisen and Wu 2002). Using this approach, we analyzed the SDH superfamily from approximately 250 completed or nearly completed bacterial genomes and identified several functionally distinct SDH subgroups. We subsequently validated the distinctiveness of these groups through biochemical characterization of the 5 evolutionarily distinct SDH homologs found in the genome of the common soil-inhabiting bacterium, Pseudomonas putida KT2440 (Nelson et al. 2002).
|
Materials and Methods |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
Phylogenetics
Shikimate homologs were identified from completed genomes by protein Blast using the H. influenzae Rd KW20 AroE. Approximately 650 homologs were identified, using a Blast cutoff value of E < 10–05. Following phylogenetic reconstruction with all 650 homologs, the data set was refined manually to eliminate genomes whose homologs clustered in unresolvable groups due to high similarity (strains within species). Perl scripts were used to extract the number of homologs present in each bacterial isolate as well as the pertinent motifs that characterize the SDH family. Multiple sequence alignment of the resulting data set was achieved using ClustalX 1.83 (Thompson et al. 1997
), with manual editing and trimming of the alignment being performed in Genedoc v2.6. Alignments were imported into MEGA3 (Kumar et al. 2004), and a Neighbor-Joining tree was constructed using the Jones–Taylor–Thornton amino acid substitution model and 500 bootstrap replicates. The Bayesian tree was generated with MrBayes, using 64,000,000 generations and 10 chains (640,000 trees total), with a burn-in of 200,000. Because of data set size, the tree was not taken to convergence.
Comparative Genomics
A 10-kb region, 5 kb immediately upstream and downstream of ael1 or aroE, was extracted from the genomes of Burkholderia ambifaria AMMD, Burkholderia cenocepacia HI2424, Polaromonas sp. JS666, Polaromonas naphthalenivorans CJ2, P. putida KT2440, and Bradyrhizobium japonicum USDA110 and the Rhodobacterales bacterium HTCC2654 and subsequently reannotated with Genemark heuristic annotation (Besemer and Borodovsky 1999). A custom Blastable database of all these genes was generated, and standalone Blast used to identify all homologs of the 11 P. putida KT2440 open reading frames (E value cutoff < 10–05). Other SDH homologs were identified by Blast against the non-redundant database at National Center for Biotechnology Information.
Cloning
Five putative SDH-encoding genes were amplified by polymerase chain reaction (PCR) from P. putida KT2440 genomic DNA using primers with incorporated restriction sites (table 1). The resulting products were cloned into either a modified pET28a vector (pET28.3, table 2) or the p15TV-LIC vector, a ligation-independent vector (table 2) due to incompatible cloning sites between the insert and pET28.3. All constructs were verified by DNA sequencing.
|
|
Protein Expression and Purification
The expression and purification steps of the P. putida SDHs were accomplished as follows: the SDH enzymes were expressed in either the E. coli strain BL21 gold (p15TV-LIC constructs) or the BL21 CodonPlus (pET28.3 constructs) in 1 l Luria-Bertani media supplemented with the appropriate antibiotics: kanamycin 50 µg/ml–1 and ampicillin 100 µg/ml–1 (BL21 gold) or kanamycin 50 µg/ml–1 (CodonPlus). Cultures were incubated at 37 °C with shaking until they reached an attenuance of 0.6–0.8 at 600 nm and then induced with 0.4 mM isopropyl-β-D-thiogalactopyranoside for 3 h at 37 °C and allowed to grow for an additional 4 h with shaking at 24 °C. Harvested cells were disrupted by sonication, and the insoluble cellular material was removed by centrifugation at 20 000 x g (30 min, 4 °C). The recombinants were purified from other contaminating proteins using nickel-nitrilotriacetic acid affinity chromatography. Binding, wash, and elution buffers consisted of 3, 30, and 300 mM imidazole, respectively, in 10 mM Tris–HCl (pH 7.5), 500 mM NaCl, and 5% glycerol. Each recombinant protein purified as a single band as determined by sodium dodecyl sulfate–polyacrylamide gel electrophoresis. Proteins were quantified at 280 nm using the extinction coefficients of 22,710 M–1 cm–1 (AroE), 18,450 M–1 cm–1 (RifI), 11,585 M–1 cm–1 (YdiB), 19,285 M–1 cm–1 (SdhL), and 14,355 M–1 cm–1 (Ael1). The amount of purified protein ranged from 10 to 30 mg/ml–1 per one liter bacterial culture. Protein samples for kinetic studies were dialyzed and stored at 4 °C in 10 mM Tris–HCl (pH7.5), 500 mM NaCl, and 5% glycerol.
Enzyme Kinetics
Saturation kinetics studies for the P. putida SDHs were carried out as previously described (Singh and Christendat 2006). Activity was assayed by monitoring the reduction of NADP+ or NAD+ (
= 6,220 M–1 cm–1) at 340 nm and 22 °C in the presence of shikimate or quinate. Saturation kinetics were carried out by varying the concentrations of either substrate or cofactor while keeping the other at saturation. The saturation curves were repeated 2–4 times over 2 independent purifications. The computer program GraFit (Cleland 1979) was used to calculate KM and Vmax.
|
Results |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
Phylogenetic Analysis
Approximately 650 SDH homologs were identified from completed or nearly completed bacterial genomes by protein Blast using the H. influenzae RdKW20 AroE as the query (E value < 10–05). The redundancy of the resulting data set was reduced by removing all but a single representative strain for each species, yielding a refined data set of 315 SDH homologs representing almost all major bacterial taxonomic groups. To facilitate cross-referencing throughout the analysis, all 315 homologs were given a unique reference identifier. Because a large proportion of genomes analyzed contained 2 or more SDHs (B. japonicum USDA110, e.g., has 8 distinct SDHs in its genome), each homolog was also tagged with the number of SDH enzymes present in its parent genome (supplementary table S1, Supplementary Material online), which permitted an analysis of the phylogenetic distribution of homologs within each strain. To help assign function to each phylogenetic group, we extracted from each homolog the key motifs that are putatively important for substrate recognition (Michel et al. 2003
; Padyana and Burley 2003; Singh et al. 2005; Singh and Christendat 2006). These included the equivalent positions of the Arabidopsis thaliana AroE Ser336-X-Ser338 (SXS), where X is any amino acid residue, Asn406-Thr407 (NT), and Asn421-Thr422 (of NTD) (Singh and Christendat 2006) (fig. 2A), as well as the motifs that determine NADPH or NADH specificities (Scrutton et al.1990; Michel et al. 2003; Lim et al. 2004; Singh et al. 2005).
|
To understand the evolutionary relationships among the different SDHs, we first constructed Neighbor-Joining and Bayesian phylogenetic trees using the 315 SDH homologs. The resulting gene genealogies from the 2 methods supported the major phylogenetic groups, with only minor inconsistencies in the relative placement of terminal nodes and position of specific taxa. Because of this, the Bayesian tree is presented in figure 3, with the Neighbor-Joining tree being available as supplementary figure 1 (Supplementary Material online). Subsequently, we mapped onto the tree previously characterized SDH homologs, including the AroE from E. coli (Michel et al. 2003
), the RifI enzymes from A. mediterranei and Streptomyces coelicolor (Yu et al. 2001; Guo and Frost 2002, 2004), the SdhL from H. influenzae (Singh et al. 2005), and the E. coli YdiB (Michel et al. 2003), thereby providing functional landmarks within our established evolutionary framework.
|
The SDH gene genealogy was strongly supported, yielding well-resolved clades, which suggested significant evolutionary diversification within the SDH enzyme family (fig. 3). We characterized the SDH tree by examining the clustering and evolutionary relationships among SDH homologs, the number of homologs present among the representative taxa within each major clade, and the distribution of key functional motifs. These features provide perspective into which SDH homologs are essential and which may have conditional or niche-specific functional roles.
The archetypical E. coli AroE falls within a well-supported enteric clade that contains bacteria such as Salmonella spp., Yersinia spp., Photorhabdus spp., and Buchnera spp. and is largely congruent with the evolutionary history previously established for the Proteobacteria. Congruence between the relationships among SDH sequences within a clade and the previously established species relationships as determined by 16S rRNA ribotyping (Ciccarelli et al. 2006) is a strong indication that the SDH homologs have evolved largely in a clonal or vertical manner. The 
-Proteobacteria AroE sequences form a single monophyletic group, but their overall placement within the Proteobacteria AroE clade is not well resolved. Most of the sequences in this clade contain SXS, NT, and NTD functional motifs, which are characteristic of the prototypical AroE (figs. 2B and 3). These motifs are also found in the AroE homologs in clades comprising cyanobacterial and the actinobacterial sequences, both of which are also congruent with the established 16S rRNA relationships. In contrast, AroE SDH homologs from the Firmicutes (Gram-positive bacteria) form 2 distinct clades, one being a smaller group comprising only the Bacilli (Geobacillus, Lactobacillus, Listeria, Staphylococcus) and the second forming a sister group to 1 of the 2 Bacteroidetes/Chlorobi groups. Interestingly, a well-supported clade of SDH homologs from the Gram-negative enteric bacteria (E. coli, Salmonella typhimurium, Shigella flexneri, and Yersinia. frederiksenii) are nested within the larger Firmicute AroE clade (fig. 3).
The prototypical H. influenzae SdhL sequence falls within a well defined and diverse lineage containing representatives from the 
-Proteobacteria, Actinomycetes, and Deinococcus groups (figs. 3 and 4B). Interestingly, the placement of the species within this clade is not congruent with their accepted 16S rRNA relationships, indicating an important role for horizontal gene transfer. All SdhL homologs retain the NT and NTD motifs characteristic of the AroE (fig. 2B) but instead of the conserved serine motif (SXS), they possess a (N/S/T/)XG motif (fig. 2B). All but 3 taxa in this monophyletic group have more than 1 SDH homolog (supplementary table S1, Supplementary Material online). The 3 basal taxa within this group (Syntrophus aciditrophicus, Desulfotalea psychrophila, and Syntrophobacter fumaroxidans) are all 
-Proteobacteria that carry only one homolog having motifs consistent with AroE enzymes. Furthermore, their position at the base of this clade in this tree is not well supported, prompting exclusion from the SdhL group.
|
The prototypical RifI enzymes of A. mediterranei and S. coelicolor are the only enzymes included in this analysis that were not obtained from completely sequenced genomes. The RifI phylogenomic group is highly structured, comprising primarily Actinomycetes and Proteobacteria. The 3 key motifs vary considerably within each subclade, possibly suggesting further functional diversification (fig. 2B). We attempted to subdivide the RifI supergroup into subgroups using both phylogenetic clustering and the distinctiveness of the substrate-binding motifs, yielding at least 4 additional subgroups within this clade (named RifI2-RifI5). Some taxa, like B. japonicum USDA110, P. naphthalenivorans CJ2, Azotobacter vinelandii AvOP, and Frankia sp. CCI3, have multiple homologs within the RifI supergroup, with B. japonicum USDA110 having one divergent homolog that is sufficiently distinct (both phylogenetically and with respect to its motifs) to warrant assignment to its own subgroup (RifI4). As with the SdhL group, all taxa represented in each subgroup have more than one homolog outside of the RifI supergroup, one of which is always a predicted AroE
One of the very surprising findings of the phylogenetic analysis was the observation that the prototypical E. coli YdiB sequence clusters with homologs from several other enteric bacteria in a clade nested within the Gram-positive AroE lineage (figs. 3 and 4A). Because of this unexpected phylogenetic placement, we believe that the E. coli YdiB sequence, and its very close relatives, is a xenolog (acquired through horizontal gene transfer) rather than a paralog as previously reported (Michel et al. 2003). Given this, we did not use this sequence for the functional analysis of the putative shikimate–quinate dehydrogenase group. We did note, however, that one P. putida homolog fell into a separate and distinct and well-supported clade that formed a sister lineage to the RifI supergroup (fig. 3). Although there were no biochemically characterized SDHs within this group to serve as a reference, numerous characteristics suggested that this clade represented quinate/shikimate dehydrogenases. All species represented in this group had at least one other SDH homolog elsewhere in the phylogeny (supplementary table S1, Supplementary Material online), and the relationships among the bacteria were not congruent with the 16S rRNA species tree, strongly suggesting that these sequences are frequently transferred among strains via horizontal gene transfer. In addition, much like the AroE class, this enzyme group contains the NT and NTD motifs, but the second serine of the SXS motif can also be a threonine, SX(T/S) (fig. 2B). Because of the high similarity in the primary motifs, we extended our analysis to include the cofactor-binding residues, which are found approximately 20 residues downstream from the G-motif (GXGGXX) (Scrutton et al. 1990; Lesk 1995; Carugo and Argos 1997a, 1997b). In contrast to the AroE enzymes discussed above, whose cofactor-binding motif possesses the arginine typified of NADPH-binding enzymes, this group has a conserved aspartate instead, indicating a general preference for NADH. This is consistent with the kinetic profiles of the quinate/shikimate dehydrogenase (YdiB) enzymes from fungi and plants, which also exhibit a preference of YdiB for NADH over NADPH (Giles et al. 1985; Kang and Scheibe 1993). Based on this, we designated the group YdiB2.
With the exception of the Gram-positive groups, we have subdivided each group within the gene genealogy into a putative function, using motif profiling and phylogenomic approaches. The consensus motifs generated for each of the main functional subclasses will provide strong hypotheses for the development of activity models and for establishing putative function of new SDH homologs. A prominent motif emerges for all the established subclasses, except for the RifI supergroup, which likely represents several distinct functional subgroups (fig. 2B). To confirm the functional categorization of our phylogenetically defined groups, we characterized biochemically the 5 evolutionarily distinct SDH homologs found in P. putida KT2440.
Functional Validation through Kinetics
We measured saturation kinetic parameters from purified recombinant proteins for each of the 5 P. putida homologs, considering the ability of each to catalyze the oxidation of shikimate or quinate in the presence of either NAD+ or NADP+. The P. putida SDH homolog #1 falls in the putative AroE group on the phylogenetic tree (figs. 3 and 4B) and was therefore predicted to exhibit biochemical characteristics attributed to other AroE enzymes. As expected, this homolog catalyzes the oxidation of shikimate but not quinate and utilizes either NAD+ or NADP+ (table 3). It binds NADP+ preferentially (Km = 112 µM) over NAD+ (Km = 5 mM), indicating that NADP+ is the biological cofactor. A similar dual specificity for the cofactor was reported for the AroE from the thermophile, Archaeoglobus fulgidus, which also had a catalytic efficiency (with NADP+ as a cofactor) comparable to all others reported (Michel et al. 2003; Lim et al. 2004; Fonseca et al. 2006; Singh and Christendat 2006). These data serve to corroborate our previous designation of this homolog and other members within this group as the AroE class.
|
The P. putida homolog #460 can catalyze the NAD+-dependent oxidation of either shikimate or quinate (table 3), consistent with the properties of quinate dehydrogenases from fungi and plants (Giles et al. 1985
; Kang et al. 1994). We therefore designated this homolog YdiB2Ppu. The turnover rate of the E. coli YdiB (Michel et al. 2003), which clusters with the Gram-positive SDH homologs, is 100-fold less than that of YdiB2Ppu. This is likely attributable to the Gram-positive origin of the E. coli homolog (YdiB), whose SDH ancestor may not have been optimized for quinate metabolism. The 2 SDH homologs, #453 and #504, which fall within taxonomically diverse groups, exhibit no measurable dehydrogenase activity when all 4 combinations of available ligand and substrates were tested (table 3). We have designated these RifI2Ppu (#453) and SdhLPpu (#504) due to their position near their respective marker homologs in our phylogenetic tree (fig. 3).
The last P. putida SDH homolog, #132, clusters in a clade of 6 proteins within the AroE group of the β-Proteobacteria (marked with an asterisk in fig. 3), which includes sequences obtained from -, β-, and -Proteobacteria. All members of this clade have a consensus motif of (S/T)XS, N(A/C), and NFD, which differs substantially from the established AroE motif (figs. 2B and 3). It exhibits the broadest range of substrate specificity as it is capable of catalyzing the oxidation of shikimate with either cofactor (table 3). Although the enzyme preferentially binds NADP+ (KM = 7.3 µM) over NAD+ (KM > 5 mM), its turnover rate is 10-fold greater with NAD+ (table 3). It also has a higher affinity for shikimate (23-fold greater, using NADP+), but a lower turnover rate (45-fold lower), compared with AroEPpu (table 3). This lower turnover is compensated by the increased affinity for shikimate, making it as efficient as AroEPpu. Because of its unique enzymatic profile, binding site motif, and the fact that this homolog is contained within a small monophyletic group within the AroE lineage of the β-Proteobacteria, we have designated this homolog and the members of this group Ael1 for AroE-like.
Comparative Genomic Analysis of the Ael1 Subgroup
In order to understand the evolution of this unique homolog and further support the contention of distinct functionality, we took a comparative genomic approach to evaluate whether the ael1 homologs share similar genomic context and synteny. Orthology of both ael1 as well as flanking loci in diverse bacterial species would further support the common function of this novel SDH group. The P. putida Ael1 SDH homolog is nested within the β-Proteobacteria AroE lineage in a clade that includes representatives from only the -, β-, and -Proteobacteria. The high similarity among the Ael1 proteins within this group suggests that it has evolved quite recently and has been acquired by each of these organisms independently through horizontal gene transfer. Using pairwise BlastP, we compared the ael1 locus and the 5 genes immediately upstream and downstream of the P. putida ael1 with 5 other species within the Ael1 subgroup (fig. 4B). All 5 taxa had syntenic blocks consistent with the P. putida ael1 region. Our genomic comparisons also revealed several surprises. First, the ael1 homologs of both Polaromonas strains and B. japonicum are in an operon with a second SDH homolog (fig. 5). In the Polaromonas strains, the second SDH clusters in the main RifI group (homologs #454 and #546), whereas the B. japonicum homolog clusters in the YdiB2 group (homolog #463). Second, a type II dehydroquinase is found in the same operon as ael1 in P. putida and Polaromonas. This is highly suggestive of coregulated SDHs, which may be coordinately activated during accumulation of shikimate resulting in its redirection to alternate anabolic or catabolic pathways.
|
To further validate the distinct function of the Ael1 group, we expanded our analysis to include the genomic region encompassing the gene encoding the putative AroE of the Rhodobacterales bacterium HTCC2654. This strain clusters just basal to the ael1 group. The position of the HTCC2654 AroE in the SDH phylogenetic tree relative to the Ael1 group is well supported, and because the other HTCC2654 homolog clusters within the YdiB, it suggests that this homolog functions as the primary SDH (AroE). For this reason, we did not included it as a member of the new Ael1 subgroup. Our comparative genomic analysis shows that unlike ael1, the HTCC2654 aroE appears to be coregulated with only an upstream type II dehydroquinase and possibly a hypothetical protein. In addition, the genetic composition and organization within this region is substantially different from the genomic islands containing ael1 (fig. 5), strongly supporting the unique evolutionary derivation of Ael1.
|
Discussion |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
We identified and characterized several functional subgroups of the SDH protein family using a combined biochemical and phylogenetic approach. Biochemical features (substrate-binding site motifs) and kinetic properties of characterized representatives from each functional class were mapped onto a phylogenetic treatment of all SDH homologs to identify and differentiate evolutionarily and functionally distinct groups. This approach led to the phylogenetic characterization of the 4 previously identified SDHs (AroE, YdiB, SdhL, and RifI), the discovery of a novel SDH homolog, Ael1, the identification of a phylogenetically distinct shikimate/quinate dehydrogenase group, YdiB2, and the extent of diversification in the RifI supergroup.
All bacteria analyzed in this study encoded an AroE SDH homolog, and by and large, the phylogenetic relationships of these sequences are consistent with the accepted relationships among species (Ciccarelli et al. 2006). This trend is a strong indicator of vertical evolutionary descent during speciation, reflecting the essential and ancestral nature of aroE and its related pathway. In contrast, the phylogenetic structure obtained for the other SDH homologs (YdiB2, SdhL, RifI, RifI2, RifI3, RifI4, RifI5, and Ael1) is not congruent with the bacterial species tree, suggesting a combination of independent evolution and acquisition through horizontal gene transfer. The relatively restricted phylogenetic distribution of these homologs, in contrast to AroE, is consistent with their involvement in highly specialized or even niche-specific metabolic processes, perhaps indicating nonessential or facultative roles. This is further reinforced by the fact that all taxa represented in these groups possess the ancestral AroE homolog. Interestingly, we were unable to identify the AroE using motif profiling and phylogenetic analyses for 7 taxa that had multiple SDH homologs either because all homologs within the genome had highly similar AroE-specific motifs or their motifs did not conform to what had been previously described for the vast majority of AroE enzymes. These homologs are highlighted in the phylogeny shown in figure 3. Five of these 7 taxa were Gram-positive bacteria, suggesting an even more limited understanding of SDH function and diversity within this diverse bacterial group. We anticipate a more comprehensive phylogenetic analysis will be possible for this group of bacteria as more Gram-positive genome sequences become available.
Our phylogenomic analysis led to the identification of a novel subclass of SDH, Ael1. By incorporating substrate-binding motifs into the SDH phylogenetic framework, we identified a unique binding/catalytic motif among a single restricted clade, suggesting a novel evolutionary and functional class of SDH enzyme. Subsequent biochemical analysis of the P. putida Ael1 homolog showed that this enzyme has enzymatic properties distinct from any other SDH subclass (table 3). Moreover, the observation of strong synteny in the flanking genomic regions of closely related ael1 orthologs indicates that they likely share similar function, are involved in a pathway separate from the other SDH subclasses, and evolved from a common ancestor. Lastly, our observation that Ael1 retains significant SDH activity indicates that the role of this enzyme may include the redirection of shikimate to an alternate pathway, as observed for the YdiB subclass (Knop et al. 2001) (fig. 1). The similarity between Ael1 and AroE in terms of a high affinity for shikimate is consistent with its phylogenetic placement within the AroE enzymes.
An important finding that emerged from our analysis was that the second SDH (originally designated YdiB) encoded by Gram-negative enteric bacteria (including E. coli, S. typhimurium, S. flexneri, and Y. frederiksenii) is of Gram-positive (Firmicute) origin. This homolog clearly clusters with SDH enzymes from other Firmicutes with significant support, and the representation of several enteric bacteria in this group strongly supports its acquisition through horizontal gene transfer by the ancestor of the enterics. Consistent with the xenologous nature of this homolog is previous reports of this second E. coli SDH catalyzing the oxidation of shikimate and quinate but with lower than expected substrate conversion rates for a quinate/shikimate dehydrogenase (YdiB) (Michel et al. 2003). The ancestor of the enteric bacteria may have been devoid of this enzyme, and the acquisition of the Firmicute-originating homolog via horizontal gene transfer may have permitted better exploitation of existing hosts or the expansion of their host range. In support of this hypothesis, the fungus Rhizoctonia solani has been shown to alter virulence against its plant host by modulation the quinate and shikimate pathways, suggesting that subtle changes in these pathways may have drastic effects on fitness (Liu et al. 2003). An alternative hypothesis is that the enteric ancestor may have lost its bona fide Gram-negative YdiB in exchange for this Firmicute-derived homolog. This may have been selected if the Firmicute YdiB-like enzyme has additional substrates other than shikimate and quinate. Evaluating the contribution of YdiB to the virulence of these bacterial pathogens may be achieved through comparative pathogenicity assays between wild-type strains and ydiB mutants, although a thorough understanding of substrate specificity of the Firmicute YdiB is essential to evaluating its contribution to fitness. It is notable that there is no E. coli representative in the YdiB2 subgroup.
Although most of this study has focused on the 5 SDH homologs identified in P. putida, remarkably, B. japonicum USDA110 has 8 SDH homologs (Kaneko et al. 2002). Three of these enzymes cluster within or near RifI supergroup, 2 within or near the newly designated YdiB2 group, 2 within or near the AroE group, and 1 within the newly designated Ael1 group. The second YdiB and AroE fall just outside the main groups established here and may represent horizontally acquired or paralogous SDHs. It is unclear what selective pressures may have driven this diversification or how the genetically diverse SDH complement within this particular strain enhances its fitness. Given that B. japonicum is believed to be a soil-dwelling, rhizosphere-associating bacterium, it is plausible that 2 of the homologs that cluster near or within the RifI group may be genetic components of antibiotic biosynthetic clusters, providing an advantage to the bacterium during its free-living stage. The other SDH homologs may enhance nutritional versatility, enabling the bacterium to exploit alternate substrates as energy sources during periods of privation.
The SDH family represents an excellent model for the study of enzyme evolution. Our analysis indicates that SDH activity, exemplified by AroE, is the most likely ancestral function of the SDH group, a notion consistent with its presence in all characterized bacteria and the congruence of the gene genealogy to the species tree (Ciccarelli et al. 2006). This permits evolutionary polarization of enzyme diversification, with aroE functioning as the ancestral state, and new functional classes evolving from aroE via gene duplication (Jensen 1976; O'Brien and Herschlag 1999). Functional divergence of these duplicates may result in subfunctionalization (the subdivision of the ancestral function among the duplicated genes) or neofunctionalization (the evolution of a new and distinct function for one of the duplicated genes) (Taylor and Raes 2004; Ciccarelli et al. 2006). The varying extent to which the other SDH homologs have retained their ancestral SDH activity (table 3) supports the model of neofunctionalization in a stepwise manner. During such an evolutionary process, the gradual accumulation of changes in critical regions of a gene duplicate results in the eventual reduction of affinity for the ancestral substrate and the gain of binding capacity for an alternate substrate. Catalytic promiscuity is a particularly strong indicator of the propensity for enzyme functional diversification (Aharoni et al. 2005).
Our analyses suggest the existence of many additional functional subgroups aside from the five functional classes discussed in this study, a conclusion that has two important implications for the development of effective antimicrobials. First, proposals for the development of drugs targeting the ubiquitous and highly conserved bacterial AroE enzyme have been put forth recently (Fonseca et al. 2006); however, given the diversity of SDHs and the evolutionary versatility of bacteria, the precise extent to which other SDHs may compensate for AroE loss is not entirely clear. The ability of one SDH to substitute for the function of AroE would render drugs targeting this essential enzyme largely ineffective and impractical, requiring a thorough understanding of existing substrate specificities and affinities, as well as the number and type of molecular changes required for conversion of one SDH into another. Our analysis of the SDH complement of P. putida has revealed that Ael1 and YdiB2, in addition to AroE, are capable of utilizing shikimate as a substrate. This potential for functional compensation during AroE loss by these two homologous enzymes could render antimicrobials targeting AroE useless. The retention of ancestral function by homologous enzymes may be selectively favored, such that substrate diversification may evolve more by subfunctionalization or neofunctionalization with partial retention of ancestral function.
The prospect of the RifI supergroup representing a diversity of novel antibiotic biosynthetic mechanisms is also especially exciting and timely in the wake of increasingly resistant bacterial pathogens. If the members of this SDH subgroup function to modify precursors during antibiotic biosynthesis, a thorough exploration of their substrate specificity and biochemistry would have significant repercussions for the generation antibiotic diversity through side group modification or "decoration" (Fjaervik and Zotchev 2005). The serial or combinatorial application of multiple SDH enzymes to a given substrate precursor may facilitate the development of entirely new antibiotics to combat highly aggressive and resistant superpathogens.
|
Supplementary Material |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
Supplementary figure 1, table S1, and bayesian and neighbour joining trees in Newick format are available at Molecular Biology and Evolution online (http://www.mbe.oxfordjournals.org/).
|
Acknowledgements |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
We would like to extend our sincerest thanks to Dr Bruce Ramsay at Queen's University for providing Pseudomonas putida KT2440. D.S.G. is supported by grants from the Canada Research Chairs program, Natural Sciences and Engineering Research Council of Canada (NSERC), and the Canadian Institutes of Health Research. J.S. is currently funded by an NSERC Postdoctoral Research Fellowship. D.C. is funded by an NSERC Discovery Grant and by an Early Research Award.
|
Footnotes |
|---|
1 These authors contributed equally to this work.
Claudia Schmidt-Dannert, Associate Editor
|
References |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
Aharoni A, Gaidukov L, Khersonsky O, Mc QGS, Roodveldt C, Tawfik DS. The ‘evolvability’ of promiscuous protein functions. Nat Genet (2005) 37:73–76.[Web of Science][Medline]
Besemer J, Borodovsky M. Heuristic approach to deriving models for gene finding. Nucleic Acids Res (1999) 27:3911–3920.
Carugo O, Argos P. NADP-dependent enzymes. II: evolution of the mono- and dinucleotide binding domains. Proteins (1997a) 28:29–40.[CrossRef][Web of Science][Medline]
Carugo O, Argos P. NADP-dependent enzymes. I: conserved stereochemistry of cofactor binding. Proteins (1997b) 28:10–28.[CrossRef][Web of Science][Medline]
Ciccarelli FD, Doerks T, von Mering C, Creevey CJ, Snel B, Bork P. Toward automatic reconstruction of a highly resolved tree of life. Science (2006) 311:1283–1287.
Cleland WW. Statistical analysis of enzyme kinetic data. Methods Enzymol (1979) 63:103–138.[Medline]
Crooks GE, Hon G, Chandonia JM, Brenner SE. WebLogo: a sequence logo generator. Genome Res (2004) 14:1188–1190.
Eisen JA. Phylogenomics: improving functional predictions for uncharacterized genes by evolutionary analysis. Genome Res (1998) 8:163–167.
Eisen JA, Hanawalt PC. A phylogenomic study of DNA repair genes, proteins, and processes. Mutat Res (1999) 435:171–213.[Web of Science][Medline]
Eisen JA, Wu M. Phylogenetic analysis and gene functional predictions: phylogenomics in action. Theor Popul Biol (2002) 61:481–487.[CrossRef][Web of Science][Medline]
Fjaervik E, Zotchev SB. Biosynthesis of the polyene macrolide antibiotic nystatin in Streptomyces noursei. Appl Microbiol Biotechnol (2005) 67:436–443.[CrossRef][Web of Science][Medline]
Fonseca IO, Magalhaes MLB, Oliveira JS, Silva RG, Mendes MA, Palma MS, Santos DS, Basso LA. Functional shikimate dehydrogenase from Mycobacterium tuberculosis H37Rv: purification and characterization. Protein Expr Purif (2006) 46:429–437.[CrossRef][Web of Science][Medline]
Giles NH, Case ME, Baum J, Geever R, Huiet L, Patel V, Tyler B. Gene organization and regulation in the qa (quinic acid) gene cluster of Neurospora crassa. Microbiol Rev (1985) 49:338–358.
Gourley DG, Shrive AK, Polikarpov I, Krell T, Coggins JR, Hawkins AR, Isaacs NW, Sawyer L. The two types of 3-dehydroquinase have distinct structures but catalyze the same overall reaction. Nat Struct Biol (1999) 6:521–525.[CrossRef][Web of Science][Medline]
Guo J, Frost JW. Kanosamine biosynthesis: a likely source of the aminoshikimate pathway's nitrogen atom. J Am Chem Soc (2002) 124:10642–10643.[CrossRef][Web of Science][Medline]
Guo J, Frost JW. Synthesis of aminoshikimic acid. Org Lett (2004) 6:1585–1588.[CrossRef][Web of Science][Medline]
Herrmann KM, Weaver LM. The shikimate pathway. Annu Rev Plant Physiol Plant Mol Biol (1999) 50:473–503.[CrossRef][Web of Science]
Jensen RA. Enzyme recruitment in evolution of new function. Annu Rev Microbiol (1976) 30:409–425.[CrossRef][Web of Science][Medline]
Kaneko T, Nakamura Y, Sato S, et al, (17 co-authors). Complete genomic sequence of nitrogen-fixing symbiotic bacterium Bradyrhizobium japonicum USDA110. DNA Res (2002) 9:189–197.[Abstract]
Kang X, Neuhaus HE, Scheibe R. Subcellular localization of quinate:oxidoreductase from Phaseolus mungo L. sprouts. Zeitschrift fur Naturforschung (1994) 49:415–420.
Kang X, Scheibe R. Purification and characterization of the quinate:oxidoreductase from Phaseolus mungo sprouts. Phytochemistry (1993) 33:769–773.[CrossRef][Web of Science]
Knop DR, Draths KM, Chandran SS, Barker JL, von Daeniken R, Weber W, Frost JW. Hydroaromatic equilibration during biosynthesis of shikimic acid. J Am Chem Soc (2001) 123:10173–10182.[CrossRef][Web of Science][Medline]
Kumar S, Tamura K, Nei M. MEGA3: integrated software for molecular evolutionary genetics analysis and sequence alignment. Brief Bioinform (2004) 5:150–163.
Lesk AM. NAD-binding domains of dehydrogenases. Curr Opin Struct Biol (1995) 5:775–783.[CrossRef][Web of Science][Medline]
Leuschner C, Herrmann KM, Schultz G. The metabolism of quinate in pea roots (purification and partial characterization of a quinate hydrolyase). Plant Physiol (1995) 108:319–325.[Abstract]
Lim S, Schroder I, Monbouquette HG. A thermostable shikimate 5-dehydrogenase from the archaeon Archaeoglobus fulgidus. FEMS Microbiol Lett (2004) 238:101–106.[Web of Science][Medline]
Liu CY, Lakshman DK, Tavantzis SM. Quinic acid induces hypovirulence and expression of a hypovirulence-associated double-stranded RNA in Rhizoctonia solani. Curr Genet (2003) 43:103–111.[Web of Science][Medline]
Michel G, Roszak AW, Sauve V, Maclean J, Matte A, Coggins JR, Cygler M, Lapthorn AJ. Structures of shikimate dehydrogenase AroE and its paralog YdiB. A common structural framework for different activities. J Biol Chem (2003) 278:19463–19472.
Nelson KE, Weinel C, Paulsen IT, et al, (43 co-authors). Complete genome sequence and comparative analysis of the metabolically versatile Pseudomonas putida KT2440. Environ Microbiol (2002) 4:799–808.[CrossRef][Medline]
O'Brien PJ, Herschlag D. Catalytic promiscuity and the evolution of new enzymatic activities. Chem Biol (1999) 6:R91–R105.[Web of Science][Medline]
Padyana AK, Burley SK. Crystal structure of shikimate 5-dehydrogenase (SDH) bound to NADP: insights into function and evolution. Structure (2003) 11:1005–1013.[Medline]
Pittard J, Wallace BJ. Gene controlling the uptake of shikimic acid by Escherichia coli. J Bacteriol (1966a) 92:1070–1075.
Pittard J, Wallace BJ. Distribution and function of genes concerned with aromatic biosynthesis in Escherichia coli. J Bacteriol (1966b) 91:1494–1508.
Scrutton NS, Berry A, Perham RN. Redesign of the coenzyme specificity of a dehydrogenase by protein engineering. Nature (1990) 343:38–43.[CrossRef][Web of Science][Medline]
Singh S, Korolev S, Koroleva O, Zarembinski T, Collart F, Joachimiak A, Christendat D. Crystal structure of a novel shikimate dehydrogenase from Haemophilus influenzae. J Biol Chem (2005) 280:17101–17108.
Singh SA, Christendat D. Structure of Arabidopsis dehydroquinate dehydratase-shikimate dehydrogenase and implications for metabolic channeling in the shikimate pathway. Biochemistry (2006) 45:7787–7796.[CrossRef][Web of Science][Medline]
Taylor JS, Raes J. Duplication and divergence: the evolution of new genes and old ideas. Annu Rev Genet (2004) 38:615–643.[CrossRef][Web of Science][Medline]
Thompson JD, Gibson TJ, Plewniak F, Jeanmougin F, Higgins DG. The CLUSTAL_X windows interface: flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acids Res (1997) 25:4876–4882.
Ye S, Von Delft F, Brooun A, Knuth MW, Swanson RV, McRee DE. The crystal structure of shikimate dehydrogenase (AroE) reveals a unique NADPH binding mode. J Bacteriol (2003) 185:4144–4151.
Yu TW, Muller R, Muller M, Zhang X, Draeger G, Kim CG, Leistner E, Floss HG. Mutational analysis and reconstituted expression of the biosynthetic genes involved in the formation of 3-amino-5-hydroxybenzoic acid, the starter unit of rifamycin biosynthesis in Amycolatopsis mediterranei S699. J Biol Chem (2001) 276:12546–12555.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  H. Teramoto, M. Inui, and H. Yukawa Regulation of Expression of Genes Involved in Quinate and Shikimate Utilization in Corynebacterium glutamicum Appl. Envir. Microbiol., June 1, 2009; 75(11): 3461 - 3468. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||