Tracing phylogenomic events leading to diversity of Haemophilus influenzae and the emergence of Brazilian Purpuric Fever (BPF)-associated clones

Abstract

Here we report the use of a multi-genome DNA microarray to elucidate the genomic events associated with the emergence of the clonal variants of Haemophilus influenzae biogroup aegyptius causing Brazilian Purpuric Fever (BPF), an important pediatric disease with a high mortality rate. We performed directed genome sequencing of strain HK1212 unique loci to construct a species DNA microarray. Comparative genome hybridization using this microarray enabled us to determine and compare gene complements, and infer reliable phylogenomic relationships among members of the species. The higher genomic variability observed in the genomes of BPF-related strains (clones) and their close relatives may be characterized by significant gene flux related to a subset of functional role categories. We found that the acquisition of a large number of virulence determinants featuring numerous cell membrane proteins coupled to the loss of genes involved in transport, central biosynthetic pathways and in particular, energy production pathways to be characteristics of the BPF genomic variants.

Figure 1

MLST-based phylogenetic tree using H. influenzae sequences available in the MLST database (www.mlst.net). Red dots denote strains used in this study. Relationships are drawn using the minimum evolution algorithm applied to concatenated sequences of the seven MLST housekeeping genes and a fragment of hap [7]. The MLST-based calculated mean genetic distance within our collection is 0.026 ± 0.002, which is in agreement with the one calculated from all recognized sequence types (STs), i.e., 0.025 ± 0.002.

Figure 2

Putative sources of the novel genomic content of HK1212 based on blast analysis. A. Summary of HK1212 sequence database searches using blastx. Species-associated sequences with the best blastx scores are shown with (top) and without (bottom) those with no match (hypotheticals). B. Same as A, except the summary represents HK1212 features (partial ORF) analysis using blastp. Significant blast scores to sequences from non-typable H. influenzae and H. aegyptius strains constitute the majority of “Haemophilus spp.” fraction.

Figure 3

Comparative sequence analysis of shared, presume absent and gained cellular role categories in HK1212 relative to the H. influenzae KW20/Rd reference genome. Absent and shared role categories were determined by CGH using a single genome array based on H. influenzae KW20/Rd. Gained functions were identified based on the Gene Discovery (GD) approach.

Figure 4

Total gene content predictions derived from the CGH analysis using the species microarray and PFGE-based genome size estimates for the strains characterized in this study.

Figure 5

Phylogenomic relationships among Haemophilus strains based on the CGH patterns using different markers sets. A. CGH clustering based on the global gene hybridization patterns of all 4578 70-mers using the trinary designations “0”, “0.5” and “1” representing those coding sequences determined to be “absent”, “divergent” and “present” respectively. Three major lineages, i.e. Clades 1, 2 and 3 are color-coded in brown, green and blue respectively. B. Dendrogram is based on the trinary designations from the 329 markers representing putative virulence associated genes. C. Clustering approach was based on the ratio percentile values of the conserved markers. D. Dendrogram is based on the trinary designations from the markers representing variable genes.

Figure 6

Summary of functions characteristic for the BPF/aegyptoids (Clade 1) group members based on the marker association analysis using Fisher’s Exact Test. Findings are summarized according to the putative cellular functional role categories of the coding DNA sequences. Numbers in brackets represent the number of characteristic genes for each group. The group of strains designate as “Others*” refer to Clade 2, Clade 3, HK369 and HK1141 genomes.

Figure 7

Summary of functions characteristic for strains of Clades 2 and 3 (Figure 3) based on the marker association analysis using Fisher’s Exact Test. Findings are summarized according to the putative cellular functional role categories of the coding DNA sequences. Numbers in brackets represent the number of characteristic genes for each group. The group of strains designated as “Others*” refer to Clade 1, Clade 2, HK369 and HK1141 genome, whereas “Others**” refer to Clade 1, Clade 3, HK369 and HK1141 genomes.