Approximately 1.5 mL K-EDTA blood received from Novartis was frozen (–80°C) and then thawed 1 week later after lysis of the erythrocytes. After centrifugation of the sample at 3,000 × g for 30 min, the pellet was resuspended in M199 (Cellgro, Mediatech, Inc., Herndon, VA, USA) containing 20% (vol/vol) fetal bovine serum, 22.5% (vol/vol) sodium bicarbonate, 100 mmol/L sodium pyruvate and GlutaMAX-1 (Gibco Life Technologies, Grand Island, NY, USA) and spread onto trypticase soy agar containing 5% (vol/vol) rabbit blood and chocolate agar (Becton Dickinson, Cockeysville, MD, USA), respectively. Plates were incubated at 35°C under 5% CO2 and monitored for up to 6 weeks.
Five hundred microliters phosphate-buffered saline (PBS) was added to 200 μL blood (previously frozen at –80°C) and centrifuged at 20,817 × g for 6 min. The supernatant was removed, and the pellets were resuspended in 500 μL 1× PBS followed by centrifugation for 6 min. After removing the supernatant, and resuspending the samples in 200 μL PBS, we extracted DNA by using a QIAamp Blood Kit (Qiagen, Chatsworth, CA, USA) DNA from culture-grown B. henselae strain Houston-1, B. vinsonii subspecies berkhoffii (93CO-1), B. elizabethae, B. clarridgeiae (NCSU 94-F40), and B. quintana (ATCC VR-358) were used for all PCRs as control templates.
Amplification of the 16S rDNA and the 16S–23S intergenic spacer (ITS) regions was performed as described earlier (2,3
). Amplification conditions for the citrate synthase gene (gltA
) were the same as for the 16S–23S ITS region except that primers BhCS 1137n1 (5´ AATGCAAAAAGA ACAGTAAACA 3´) and CS443f 2 (5´ GCTATGTCTGCATTCTATCA 3´) were used (4
). Selective PCR amplifications for the 16S rDNA, 23S rDNA, and rnpB
were performed as described (2
After cloning, recombinant plasmid DNA for gltA
and the 16S–23S ITS region was sequenced bidirectionally with the infrared fluorescently labeled primers M13Reverse (5´ CAGGAAACAGCTATGACCATG) and T7 (5´ TAATACGACTCACTATAGGGCGA). The recombinant DNA carrying the genes for 23S rDNA, 16S rDNA, and rnpB
was sequenced as described elsewhere (5
). All sequences were aligned by using the multiple sequence alignment editor ALIGN-IR (LI-COR), and consensus sequences for every gene sequenced were determined. Consensus sequences were then used to identify the closest match within GenBank. To determine the exact phylogenetic relationship of the new isolate within the genus Bartonella
, we analyzed an alignment that contained the sequences of 3 important phylogenetic markers, ribonuclease P RNA (RNase P RNA), 16S rDNA, and 23S rDNA, merged by catenation and organized by secondary structure elements, as described (5
). Our dataset comprises 14 Bartonella
strains (), including the 7 strains known to be human pathogens. We have also used the sequence information for the gltA
as well as the 16S–23S rDNA ITS for sequence similarity analysis. Sequences have been deposited in GenBank with accession numbers AY484592 (16S rDNA), AY484593 (23S rDNA), AY484594 (RNase P RNA), and AY484595 (gltA
Strains used for the phylogenetic analysis within genus Bartonella*
On day 14 after blood plating, growth typical for members of the genus Bartonella was obtained. Sixty-two small to medium-sized, white, shiny, smooth, nonadherent colonies were detected on chocolate agar. By day 16 after plating, 43 colonies of similar appearance were evident on blood agar. The strain was designated Bartonella strain CMO-01-1.
DNA could be successfully extracted, and subsequent PCR reactions resulted in PCR products representing 23S rRNA, 16S rRNA, RNase P RNA, 16S-23S rDNA ITS sequence, and the citrate synthase gene. All products were successfully cloned and sequenced. Sequencing of multiple clones for each gene resulted in sequences that were >99% identical to existing sequences derived from B. quintana, with the exception of the 16S–23S rDNA ITS sequence (>98.4%) and gltA (98%). Initial BLAST search results showed that the sequences for the 23S rDNA, the 16S rDNA, RNase P RNA, 16S–23S rDNA ITS sequence, and gltA derived from strain CMO-01-1 best matched B. quintana sequences that have been reported to GenBank. The data could be reproduced by using DNA extracted from the K-EDTA blood sample or from pure colonies grown on both chocolate and blood agar.
Subsequent comprehensive phylogenetic analysis clearly identified the isolate CMO-01-1 as a close relative of B. quintana type strain "Fuller." The statistical support for this relationship is 100%, as indicated by the bootstrap values for the phylogenetic tree ().
Figure 3 Phylogenetic tree of Bartonella species () based on the combined RNase P RNA, 16S, and 23S rRNA sequence alignment. Agrobacterium tumefaciens serves as the outgroup in this tree. The tree shown was generated by using the neighbor-joining method. (more ...)