Lineage-specific Virulence Determinants of Haemophilus influenzae Biogroup aegyptius

Fiona R. Strouts¹, Peter Power, Nicholas J. Croucher, Nicola Corton, Andries van Tonder, Michael A. Quail, Paul R. Langford, Michael J. Hudson, Julian Parkhill, J. Simon Kroll

, and Stephen D. Bentley

Author affiliations: Imperial College London, London, UK (F.R. Strouts, P.R. Langford, J.S. Kroll); University of Oxford, Oxford, UK (P. Power); Wellcome Trust Sanger Institute, Cambridge, UK (N.J. Croucher, N. Corton, A. van Tonder, M.A. Quail, J. Parkhill, S.D. Bentley); The Health Protection Agency, Salisbury, UK (M.J. Hudson)

Abstract

An emergent clone of Haemophilus influenzae biogroup aegyptius (Hae) is responsible for outbreaks of Brazilian purpuric fever (BPF). First recorded in Brazil in 1984, the so-called BPF clone of Hae caused a fulminant disease that started with conjunctivitis but developed into septicemic shock; mortality rates were as high as 70%. To identify virulence determinants, we conducted a pan-genomic analysis. Sequencing of the genomes of the BPF clone strain F3031 and a noninvasive conjunctivitis strain, F3047, and comparison of these sequences with 5 other complete H. influenzae genomes showed that >77% of the F3031 genome is shared among all H. influenzae strains. Delineation of the Hae accessory genome enabled characterization of 163 predicted protein-coding genes; identified differences in established autotransporter adhesins; and revealed a suite of novel adhesins unique to Hae, including novel trimeric autotransporter adhesins and 4 new fimbrial operons. These novel adhesins might play a critical role in host–pathogen interactions.

For more than a century, Haemophilus influenzae biogroup aegyptius (Hae) has caused worldwide seasonal epidemics of acute, purulent conjunctivitis (1,2). In 1984, an entirely new syndrome, Brazilian purpuric fever (BPF), emerged in the town of Promissão, São Paulo State, Brazil. Caused by an emergent clone of Hae, the virulence of BPF in children was unprecedented and fatal. Invasive infection was preceded by purulent conjunctivitis that resolved before the onset of an acute bacteremic illness, which rapidly evolved into septic shock complicated by purpura fulminans (3). In the 11 years to 1995, several hundred cases of BPF were reported, of which all but 3 were in Brazil (4,5); overall mortality rate was 40%. Cases occurred sporadically and in outbreaks, mainly in small towns, although some were in the state capital, where an epidemic was feared because of crowding and deprivation. A collaborative task force by the Brazilian Health Authorities and the US Centers for Disease Control and Prevention was created to investigate this emergent infection and identified the cause as the BPF clone of Hae (HaeBPF) (6).

After 1995, no more cases were reported for more than a decade, although cases may have been missed, submerged in periodic surges of clinically indistinguishable hyperendemic or epidemic meningococcal disease. The potential of the disease to reappear with devastating effect is, however, underscored by the recent report of a suspected outbreak (7 cases, 5 fatal within 24 hours) in 2007 in the town of Anajás in the previously unaffected Brazilian Amazon region (7); thus, it cannot be assumed that this emergent infection has gone away.

The emergence of new pathogens causing human and animal diseases represents a constant threat. Distinguishing invasive strains from their noninvasive relatives is relevant for diagnosis, treatment, and prevention of the spread of emerging infectious diseases. HaeBPF constitutes a unique H. influenzae clade separate from the usual conjunctivitis-causing Hae strains (8); in experimental infections, it has caused sustained septicemia (9) and endothelial cytotoxicity (10).

However, despite intensive research spanning 2 decades, these phenotypes remain unexplained. HaeBPF, a strain of nontypeable H. influenzae (NTHI), lacks genes encoding the polysaccharide capsule, a major virulence determinant of invasive H. influenzae. Although 1 animal study has indicated that a phase-variable lipopolysaccharide structure might play a part in the serum resistance of HaeBPF (11), in other respects, a novel lipopolysaccharide has not convincingly explained its virulence (12). With regard to adhesins, Farley et al. (13) identified duplication of fimbrial (haf) genes, with sequences differing from H. influenzae type b pilin (hif), but could find no systematic difference in binding of HaeBPF and conventional Hae strains to human epithelial cells and could not conclusively implicate this locus in virulence. Various other BPF-specific outer membrane proteins potentially involved in host–pathogen interactions have been identified, including a partially characterized hemagglutinin (14) and an ≈145-kDa phase-variable protein eliciting protective immunity (15), but none have been fully characterized, and their role in disease has not been established. HaeBPF (but not other Hae strains) has a copy of the Haemophilus insertion element IS1016 (16), which has been implicated in acquisition of capsulation genes and other unspecified virulence factors in other H. influenzae strains (17,18), but its role has not been defined.

To better define the role of HaeBPF, we conducted a pan-genomic analysis. This comparison with 5 other complete H. influenzae genomes available in public databases has enabled delineation of the accessory genome for Hae and HaeBPF, characterizing all Hae-specific features that might contribute to the differences in the biology of this lineage of H. influenzae. This study goes beyond other H. influenzae pan-genome studies (19) by comparing only complete genomes and provides an absolute genomic comparison among the strains. Analysis of differences in genome content between the Hae strains and other H. influenzae revealed a plethora of novel adhesins that might play a critical role in host–pathogen interactions.