ABSTRACT
Elizabethkingia anophelis
is an emerging pathogen. Genomic analysis of strains from clinical, environmental or mosquito sources is needed to understand the epidemiological emergence of
E. anophelis
and to uncover genetic elements implicated in antimicrobial resistance, pathogenesis, or niche adaptation. Here, the genomic sequences of two nosocomial isolates that caused neonatal meningitis in Bangui, Central African Republic, were determined and compared with
Elizabethkingia
isolates from other world regions and sources. Average nucleotide identity firmly confirmed that
E. anophelis, E. meningoseptica
and
E. miricola
represent distinct genomic species and led to re-identification of several strains. Phylogenetic analysis of
E. anophelis
strains revealed several sublineages and demonstrated a single evolutionary origin of African clinical isolates, which carry unique antimicrobial resistance genes acquired by horizontal transfer. The
Elizabethkingia
genus and the species
E. anophelis
had pan-genomes comprising respectively 7,801 and 6,880 gene families, underlining their genomic heterogeneity. African isolates were capsulated and carried a distinctive capsular polysaccharide synthesis cluster. A core-genome multilocus sequence typing scheme applicable to all
Elizabethkingia
isolates was developed, made publicly available (
bigsdb.web.pasteur.fr
), and shown to provide useful insights into
E. anophelis
epidemiology. Furthermore, a clustered regularly interspaced short palindromic repeats (CRISPR) locus was uncovered in
E. meningoseptica, E. miricola
and in a few
E. anophelis
strains. CRISPR spacer variation was observed between the African isolates, illustrating the value of CRISPR for strain subtyping. This work demonstrates the dynamic evolution of
E. anophelis
genomes and provides innovative tools for
Elizabethkingia
identification, population biology and epidemiology.
IMPORTANCE
Elizabethkingia anophelis
is a recently recognized bacterial species involved in human infections and outbreaks in distinct world regions. Using whole-genome sequencing, we showed that the species comprises several sublineages, which differ markedly in their genomic features associated with antibiotic resistance and host-pathogen interactions. Further, we have devised high-resolution strain subtyping strategies and provide an open genomic sequence analysis tool, facilitating the investigation of outbreaks and tracking of strains across time and space. We illustrate the power of these tools by showing that two African healthcare-associated meningitis cases observed 5 years apart were caused by the same strain, providing evidence that
E. anophelis
can persist in the hospital environment.