Most bacteria get their purines and pyrimidines by de novo synthesis or by salvaging them from exogenous sources. Once obtained by either route, the nitrogen bases react with phosphoribosyl pyrophosphate to form ribonucleotide monophosphates, which are then converted into all the ribo- and deoxyribo- nucleotides needed for nucleic acid biosynthesis and other cellular activities (Zalkin and Nygaard, 1996; Neuhard and Klein, 1996). Chlamydiae can neither synthesize purines and pyrimidines de novo nor, with limited exception, acquire them by salvage. They lack almost all the genes for both pathways, including the absolutely essential gene for phosphoribosyl pyrophosphate synthetase [See: Genome comparisons for sequence references; also McClarty, 1994; 1999]. They are also uniquely limited in their ability to metabolize nucleotides. C. trachomatis serovar D entirely depends on the importation of three of the four ribonucleotide triphosphates from its host (Tipples and McClarty, 1993; McClarty, 1994; 1999). Other chlamydiae may be less dependent (Read et al., 2000). Loss of genes for de novo synthesis of purines and pyrimidines is common among bacteria with small genomes, but of all such bacteria with sequenced genomes, only the obligately intracellular chlamydiae and rickettsiae have lost so many of the genes of the salvage pathways as well (Table 3). 

Table 3. Nucleotide metabolism in intracellular and epicellular bacteria with small genomes

Interconversion of nucleotides is even more restricted in chlamydiae than in the rickettsiae, which convert nucleotide monophosphates into all the nucleotides they require (Andersson et al., 1998; Ogata et al., 2001).

In addition to theATP/ADP transferase gene described in the previous section, Chlamydiaceae have a second ribonucleotide triphosphate transferase gene that is found in all chlamydial genome sequences so far completed [see Genome comparisons for references]. The two genes are almost identical in sequence, but in C. trachomatis serovar L2 their gene products exhibit very different substrate specificities (Tjaden et al., 1999). One, the ATP importer, is specific for ATP and ADP. It does not transport AMP, ribonucleoside triphosphates other than ATP, or any deoxyribonucleotide triphosphate. It supplies chlamydiae with energy from their hosts. The other is a ribonucleoside triphosphate importer that transports all four ribonucleotide triphosphates, but not ribonucleotide mono- and diphosphates or deoxyribonucleoside triphosphates. It supplies the ribonucleoside triphosphates needed by C. trachomatis. Because this transporter is energized by a H⁺ [proton] pump, it is a ribonucleotide / H⁺ transporter that provides for the net uptake of ribonucleotide triphosphates. The ribonucleoside triphosphate / H⁺ transporters of other Chlamydia and Chlamydophila spp. may be of different specificity.

Although chlamydiae, rickettsiae and Arabidopsis chloroplasts all have transporters specific for ATP and ADP, only chlamydiae have transporters with broad ribonucleotide triphosphate specificity, and only chlamydiae are auxotrophic for ribonucleotide triphosphates. Considering the great similarity between the two chlamydial genes, the ribonucleotide triphosphate / H⁺ gene probably originated by duplication of the ATP / ADP transferase gene It does not seem to have been received from Arabidopsis or transferred to rickettsiae. It would be nice to know if one or both of the ribonucleoside triphosphate transferases are present in the other families of Chlamydiales.

Why have chlamydiae, despite their capacity for using host ATP and the other ribonucleotide triphosphates for energy sources and biosynthetic intermediates, still retained the ability to generate ATP on their own? In the ante - genomic era, it was believed that chlamydiae could not make their own ATP [see McClarty (1999) for summary of the evidence that led to this conclusion], but the coming of complete genome sequences revealed that Chlamydiaceae has the genes for self-generation of ATP, probably by both the glycolytic pathway and a truncated tricarboxylic acid cycle [see: Genome comparisons and McClarty (1999) and Wyrick (2000)]. Rickettsiae, the only other intracellular bacteria known to import ATP, have retained complete tricarboxylic acid cycles but have lost their glycolytic pathways (Andersson et al., 1998, Ogata et al., 2001).

It is unlikely that the many genes of the pathways leading to endogenous synthesis of ATP survived the drastic degradation of chlamydial genomes without being of adaptive value in the chlamydial way of life. But if the capacity for importing ribonucleotide triphosphates was often accompanied by the loss of biosynthetic capacity for these nucleotides, why was there not a corresponding loss of ATP - synthesizing ability? Perhaps ATP is not transferred from host to chlamydiae fast enough to provide for maximum growth, so that exogenous ATP must be supplemented by ATP of endogenous origin. The relationship between ATP concentrations in host cytosol, intra - inclusion fluid, and chlamydiae is not known. Levels of ATP in the host cytosol have been reported to either decrease (Tipples and McClarty, 1993) or increase (Ojcius et al., 1998) as infection proceeds. It may also be that chlamydia-made ATP is unique, in that it meets a need that imported host ATP cannot satisfy. The peptidoglycan pathway has persisted in the Chlamydiaceae genome in the absence of synthesis of structural peptidoglycan, probably because in some poorly understood way peptidoglycan is needed for division of RBs [see: Chlamydial evolution: Ftsz, peptidoglycan, cell division]. Could there be a comparable hidden function for endogenous ATP? The requirement for exogenous ATP is established, but an absolute need for endogenous ATP has not been demonstrated. Sensitivity to penicillin has been taken to mean that chlamydiae must make some peptidoglycan, but there is unfortunately no specific inhibitor for the generation of ATP by chlamydiae. The idea of a unique function for chlamydia-made ATP implies the existence of segregated, non-mixing intracellular pools of ATP, something that to my knowledge has never been demonstrated in any organism.

dTTP for bacterial DNA is usually made by de novo synthesis from duMP via thymidylate synthetase or by the salvage of thymine and thymidine by thymine phosphorylase and thymidine kinase (Neuhard and Klein, 1996). Since chlamydiae have been known for a long time to be without thymidine kinase and not to incorporate thymidine into their DNA, it was concluded that they must use thymidylate synthetase to make dTTP (McClarty, 1994, 1999). Again the complete sequences provided a surprize. Not only are the chlamydial genomes without a gene for thymidine kinase, they also have none for thymidylate synthetase (see Genome comparisons for references)! As one of the many salvage pathway genes lost during genome reduction, the thymidine kinase gene is frequently not found in the small genomes of epicellular and intracellular bacteria (Table 3). However, the absence of thymidylate synthetase is rarer and not readily explicable. The other small genomes without the thymidylate synthetase gene are those of Helicobacter pylori and Treponema pallidum, both epicellular, not extracellular, organisms (Table 3). Making interpretation even more difficult is the absence of the gene in the cyanobacterium Synechocystis, a free-living bacterium of ancient lineage ( Kaneko and Tabata, 1997). Chlamydiae could have lost the thymidylate synthetase gene as part of their progression toward greater and greater host dependency, but the Chlamydiales lineage may never have had the gene at all. How chlamydiae and other thymidylate synthetase-less bacteria get their dTTP is another unanswered question.

The Chlamydiales ancestor probably brought into its first host cell a more or less complete set of genes for salvage, synthesis, and interconversion of nucleotides. However, by the time the LCA of Chlamydiaceae arrived on the evolutionary landscape, many of these genes must have already been eliminated by genome reduction. With the accumulation of complete genome sequences for more and more Chlamydiaceae species, it may in time be possible to visualize how the LCA transported, synthesized, and interconverted nucleotides. If it is assumed that variations in nucleotide metabolism stem from different patterns of gene loss and gene modification as the Chlamydiaceae radiated into different niches and evolved there along independent pathways, then it may eventually be possible to work backward to the hypothetical LCA. Even within the limited number of genome sequences now available, differences are apparent. On the basis of genome comparison, Read et al., (2000) concluded that C. muridarum and C. pneuomoniae are less dependent on host ribonucleotide triphosphates than C. trachomatis, possible because the ribonucleoside /H⁺ transporter has evolved different substrate specificities in different hosts. They visualize a simple ATP /H+ transporter in C. muridarum, a transporter with ATP and GTP specificity in C. pneumoniae, and a transporter with broad ribonucleotide triphosphate specificity in C. trachomatis.
[JWM]