Comparative genomics of bacterial and plant folate synthesis and salvage: predictions and validations

Abstract

Background: Folate synthesis and salvage pathways are relatively well known from classical biochemistry and genetics but they have not been subjected to comparative genomic analysis. The availability of genome sequences from hundreds of diverse bacteria, and from Arabidopsis thaliana, enabled such an analysis using the SEED database and its tools. This study reports the results of the analysis and integrates them with new and existing experimental data.

Results: Based on sequence similarity and the clustering, fusion, and phylogenetic distribution of genes, several functional predictions emerged from this analysis. For bacteria, these included the existence of novel GTP cyclohydrolase I and folylpolyglutamate synthase gene families, and of a trifunctional p-aminobenzoate synthesis gene. For plants and bacteria, the predictions comprised the identities of a 'missing' folate synthesis gene (folQ) and of a folate transporter, and the absence from plants of a folate salvage enzyme. Genetic and biochemical tests bore out these predictions.

Conclusion: For bacteria, these results demonstrate that much can be learnt from comparative genomics, even for well-explored primary metabolic pathways. For plants, the findings particularly illustrate the potential for rapid functional assignment of unknown genes that have prokaryotic homologs, by analyzing which genes are associated with the latter. More generally, our data indicate how combined genomic analysis of both plants and prokaryotes can be more powerful than isolated examination of either group alone.

Figure 1

The structure of tetrahydrofolate. In natural folates, the pterin ring exists in tetrahydro form (as shown) or in 7,8-dihydro form (as in DHF). The ring is fully oxidized in folic acid, which is not a natural folate. Folates usually have a γ-linked polyglutamyl tail of up to about eight residues attached to the first glutamate. One-carbon units (formyl, methyl, etc.) can be coupled to the N5 and/or N10 positions.

Figure 2

Major folate-dependent reactions of one-carbon metabolism. The gene names are for E. coli (except for sarcosine dehydrogenase). Note that the formation of 5-formyl-THF from 5,10-methenyl-THF occurs via a second catalytic activity of serine hydroxymethyltransferase (glyA), and that 5-formyl-THF is reconverted to 5,10-methenyl-THF by 5-formyl-THF cycloligase (ygfA). For simplicity, THF is not shown as a participant in most reactions in which it is consumed or released. GCV, glycine cleavage complex, comprising the products of the gcvT, gcvH, gcvP, and lpd genes; SDH, sarcosine dehydrogenase (not present in E. coli).

Figure 3

Folate synthesis and salvage pathways. Gene names are white-on-gray; all except folQ and PTR1 are from E. coli. The folQ gene has been identified only in Lactococcus lactis and plants, and PTR1 only in Leishmania and other trypanosomatids. Note that DHN aldolase also mediates epimerization of DHN to 7,8-dihydromonapterin and aldol cleavage of 7,8-dihydromonapterin. ADC, aminodeoxychorismate; DHFR, dihydrofolate reductase; DHFS, dihydrofolate synthase; DHNA, dihydroneopterin aldolase; DHNTPase, dihydroneopterin triphosphate pyrophosphatase; DHPS, dihydropteroate synthase; FPGS, folylpolyglutamyl synthase; HPPK, hydroxymethyldihydropterin pyrophosphokinase; NP, nonspecific phosphatase; PTR1, pteridine reductase 1; the subscript _ox denotes the fully oxidized forms of pterins.

Figure 4

Compartmentation of the folate synthesis pathway in plants. The steps in the pterin branch of the pathway are in blue, those in the pABA branch are in green, and the others (condensation, glutamylation, and reduction) are in red. Note that the compartmentation of the pathway and of its folate end products implies the existence of pterin or folate carriers in the mitochondria, chloroplast, and vacuolar membranes. Pathway intermediates are designated by the symbols used in Figure 3. THF-Glu_n, THF polyglutamates.

Figure 5

Clustering of predicted folate-related genes with known folate synthesis genes. Gene names are as described in the text or given below. [For full gene and genome names, see Additional File 1.] Matching colors correspond to orthologous genes. Pale grey arrows are non-folate related genes. A. Clustering of folC2 and pqqC-like genes. 5-fcl, 5-formyl-THF cycloligase. B. Clustering of folQ2 genes. fhs, formate-tetrahydrofolate ligase; dhfr, dihydrofolate reductase. C. Clustering of folB2 (fructose-6-phosphate aldolase-like) genes. D. Clustering of folM genes.

Figure 6

Complementation of folC function by Chlamydia trachomatis CT611. Complementation of E. coli folC mutant SF4 by a pBAD24 plasmid harboring CT611 from Chlamydia trachomatis on MS minimal medium with or without glycine plus methionine. E. coli folC was included as a positive control. 1, pCA24N::EcfolC; 2, pBAD24::CT611; 3, pBAD24 alone. SF4 shows slower growth even in the presence of the added amino acids. The appropriate antibiotics and inducers were included in the media as indicated. Amp, ampicillin; Ara, arabinose; Cm, chloramphenicol.

Figure 7

Complementation of pabAa pabAb function by Corynebacterium diphtheriae DIP1790. Complementation of E. coli pabAa pabAb double mutant BN1163 by a pLOI707HE plasmid harboring DIP1790 from Corynebacterium diphtheriae on M9 minimal medium with or without pABA. Arabidopsis ADCS was included as a positive control. 1, pLOI707HE::ADCS; 2, pLOI707HE::DIP1790; 3, pLOI707HE alone. The medium contained IPTG and appropriate antibiotics.

Figure 8

Pterin utilization by an E. coli folE deletant harboring Xylella fastidiosa PD0677 or Leishmania major PTR1. Deletant cells transformed with pBluescript (pBS) alone, or pBS harboring Xylella fastidiosa Temecula1 PD0677 or Leishmania major PTR1, were streaked on LB medium containing IPTG and appropriate antibiotics, without or with 300 μM thymidine (dT) or 19 μM 6-hydroxymethylpterin (HMPt). For PD0677, neopterin, monapterin, and pterin-6-aldehyde were also tested and found not to support growth (not shown). 1, pBS::PD0677; 2, pBS; 3, pBS::PTR1.