Characterization of the genome composition of Bartonella koehlerae by microarray comparative genomic hybridization profiling

Abstract

Bartonella henselae is present in a wide range of wild and domestic feline hosts and causes cat-scratch disease and bacillary angiomatosis in humans. We have estimated here the gene content of Bartonella koehlerae, a novel species isolated from cats that was recently identified as an agent of human endocarditis. The investigation was accomplished by comparative genomic hybridization (CGH) to a microarray constructed from the sequenced 1.93-Mb genome of B. henselae. Control hybridizations of labeled DNA from the human pathogen Bartonella quintana with a reduced genome of 1.58 Mb were performed to evaluate the accuracy of the array for genes with known levels of sequence divergence. Genome size estimates of B. koehlerae by pulsed-field gel electrophoresis matched that calculated by the CGH, indicating a genome of 1.7 to 1.8 Mb with few unique genes. As in B. quintana, sequences in the prophage and the genomic islands were reported absent in B. koehlerae. In addition, sequence variability was recorded in the chromosome II-like region, where B. koehlerae showed an intermediate retention pattern of both coding and noncoding sequences. Although most of the genes missing in B. koehlerae are also absent from B. quintana, its phylogenetic placement near B. henselae suggests independent deletion events, indicating that host specificity is not solely attributed to genes in the genomic islands. Rather, the results underscore the instability of the genomic islands even within bacterial populations adapted to the same host-vector system, as in the case of B. henselae and B. koehlerae.

FIG. 1.

Relationship between percent sequence identity in a global alignment of the B. henselae probe sequence to the B. quintana genome and the relative signal (M = log₂BQ/log₂BH) from control hybridizations of B. quintana versus B. henselae.

FIG. 2.

PFGE-RFLP analysis of isolates of Bartonella. Lanes 1 and 8, DNA size standard (lambda ladder). Lanes 2 to 4 (digestion with AscI): lane 2, B. koehlerae; lane 3, B. quintana strain Toulouse; lane 4, B. henselae Houston 1-980517. Lanes 5 to 7 (digestion with NotI): lane 5, B. koehlerae; lane 6, B. quintana strain Toulouse; lane 7, B. henselae strain Houston 1-980517.

FIG. 3.

Regions missing from B. koehlerae in the B. henselae prophage, the genomic islands HGIa, HGIb, and HGIc (left panel) and the chromosome II-like region (chrII) (right panel). Both panels show, from the top: repeat number of B. henselae sequences (the number at which each sequence occurs in the genome with more than 80% sequence identity over 100 bp); the B. henselae genes on “+” and “−” strands (green, annotated gene; pink, hypothetical conserved; yellow, Bartonella specific); expected array results for B. quintana (red, absent [<75% identity]; blue, present [>75% identity]); obtained array results for B. quintana (red, absent [M < −2]; blue, present [M > −2]); and obtained array results for B. koehlerae (red, absent [M < −2.7]; blue, present [M > −2.7]; pink, uncertain [M between −2.7 and −2]). In the right panel, the gene order in B. quintana is shown above the B. henselae genes. Orthologous sequences are highlighted by vertical lines between genes in B. henselae and B. quintana. Triangles show the position of tRNA genes and rhombi the position of integrase genes.

FIG. 3.

Regions missing from B. koehlerae in the B. henselae prophage, the genomic islands HGIa, HGIb, and HGIc (left panel) and the chromosome II-like region (chrII) (right panel). Both panels show, from the top: repeat number of B. henselae sequences (the number at which each sequence occurs in the genome with more than 80% sequence identity over 100 bp); the B. henselae genes on “+” and “−” strands (green, annotated gene; pink, hypothetical conserved; yellow, Bartonella specific); expected array results for B. quintana (red, absent [<75% identity]; blue, present [>75% identity]); obtained array results for B. quintana (red, absent [M < −2]; blue, present [M > −2]); and obtained array results for B. koehlerae (red, absent [M < −2.7]; blue, present [M > −2.7]; pink, uncertain [M between −2.7 and −2]). In the right panel, the gene order in B. quintana is shown above the B. henselae genes. Orthologous sequences are highlighted by vertical lines between genes in B. henselae and B. quintana. Triangles show the position of tRNA genes and rhombi the position of integrase genes.

FIG. 4.

Content of coding (a) and noncoding (b) DNA in the chromosome II-like region of B. henselae (BH), B. koehlerae (BK), and B. quintana (BQ). Upper and lower levels for B. koehlerae show the estimated number sequences classified as present based on M cutoff values of −2.7 and −2, respectively.