Citrate synthase gene comparison, a new tool for phylogenetic analysis, and its application for the rickettsiae.

Using PCR and an automated laser fluorescent DNA sequencer, we amplified and sequenced a 1,234-bp fragment of the citrate synthase-encoding gene (gltA) of 28 bacteria belonging to the genus Rickettsia. Comparative sequence analysis showed that most of the spotted fever group (SFG) rickettsiae belonged to one of two subgroups. The first subgroup included Rickettsia massiliae, strain Bar 29, Rickettsia rhipicephali, "Rickettsia aeschlimanni," and Rickettsia montana, which have been isolated only from ticks. The second subgroup was larger and included the majority of the human pathogens and also rickettsiae isolated only from ticks; the members of this subgroup were strain S, Rickettsia africae, "Rickettsia monglotimonae," Rickettsia sibirica, Rickettsia parkeri, Rickettsia conorii, Rickettsia rickettsii, the Thai tick typhus rickettsia, the Israeli tick typhus rickettsia, the Astrakhan fever rickettsia, "Rickettsia slovaca," and Rickettsia japonica. The sequence analysis also showed that the tick-borne organisms Rickettsia helvetica and Rickettsia australis and the mite-borne organism Rickettsia akari were associated with the SFG cluster, that Rickettsia prowazekii and Rickettsia typhi, two representatives of the typhus group, clustered together, and that Rickettsia canada; Rickettsia bellii, and the AB bacterium probably represent three new groups. We compared the phylogenetic trees inferred from citrate synthase gene sequences and from 16S ribosomal DNA (rDNA) sequences. For rickettsial phylogeny, the citrate synthase approach was more suitable, as demonstrated by significant bootstrap values for all of the nodes except those in the larger subgroup defined above. We also compared phylogenetic analysis results obtained in a comparison of the sequences of both genes for all of the representatives of the domain Bacteria for which the gltA sequence was determined. We believe that comparison of gltA sequences could be a complementary approach to 16S rDNA sequencing for inferring bacterial evolution, especially when unstable phylogenetic models are obtained from ribosomal sequences because of high levels of sequence similarity between the bacteria studied.