Phylogeny of the genus Haemophilus as determined by comparison of partial infB sequences

Jakob Hedegaarda,1, Henrik Okkels2, Brita Bruun3, Mogens Kilian4, Kim K. Mortensen1 and Niels Nørskov-Lauritsen3,5

Departments of Molecular and Structural Biology1 and Medical Microbiology and Immunology4, University of Aarhus, DK-8000 Aarhus C, Denmark
Departments of Clinical Biochemistry2 and Clinical Microbiology5, Aalborg Hospital, DK-9000 Aalborg, Denmark
Department of Clinical Microbiology, Statens Serum Institut, DK-2300 Copenhagen S, Denmark3

Author for correspondence: Niels Nørskov-Lauritsen. Tel: +45 9932 3207. Fax: +45 9932 3216. e-mail: nnl{at}aas.nja.dk


   ABSTRACT
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
A 453 bp fragment of infB, the gene encoding translation initiation factor 2, was sequenced and compared from 66 clinical isolates and type strains of Haemophilus species and related bacteria. Analysis of the partial infB sequences obtained suggested that the human isolates dependent on X and V factor, H. influenzae, H. haemolyticus, H. aegyptius and some cryptic genospecies of H. influenzae, were closely related to each other. H. parainfluenzae constituted a heterogeneous group within the boundaries of the genus, whereas H. aphrophilus/paraphrophilus and Actinobacillus actinomycetemcomitans were only remotely related to the type species of the genus Haemophilus. H. parahaemolyticus and H. paraphrohaemolyticus took up an intermediary position and may not belong in the genus Haemophilus sensu stricto. Ambiguous results were obtained with seven isolates tentatively identified as H. segnis, which fell into two discrete clusters. The delineation of ‘Haemophilus sensu stricto’ as suggested by infB analysis supports previous results obtained by DNA hybridization, in contrast to the delineation inferred from 16S rRNA sequence comparison.

Keywords: Pasteurellaceae, taxonomy, 16S rRNA

The GenBank accession numbers for the sequences reported in this paper are AJ289629 through AJ289694, AJ290742 through AJ290767, and AJ295746.

a Present address: Danish Institute of Agricultural Sciences, Research Centre Foulum, DK-8830 Tjele, Denmark.


   INTRODUCTION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
The genus Haemophilus Winslow et al. 1917 was conceived to encompass small non-motile Gram-negative bacilli that grew only on media containing blood or other body fluids. Pfeiffer’s so-called influenza bacillus, the Koch–Weeks bacillus, and the Bordet–Gengou bacillus were the principal representatives of this group. By 1921 it had been shown conclusively that Haemophilus influenzae was dependent on two separate factors present in blood for its growth in vitro (Tjøtta & Avery, 1921 ); in 1922, bacteria with close resemblance to H. influenzae, but dependent only on the heat-labile factor, were described (Rivers, 1922 ). Growth factor requirements have remained of importance for the taxonomy of genus Haemophilus. In the 1984 edition of Bergey’s Manual of Systematic Bacteriology, the demonstrable need for X or V factor by a small Gram-negative rod would qualify that organism as a Haemophilus strain. Conversely, the absence of such a requirement would exclude the bacterium from the genus (Kilian & Biberstein, 1984 ). This definition of the genus is, however, no longer useful, as several V factor-dependent species have been transferred from the genus Haemophilus to other genera within the family Pasteurellaceae (Pohl et al., 1983 ; Mutters et al., 1985 ).

Nucleic-acid-based characterization has indicated that the genus Haemophilus should be restricted to a limited number of species for which man is the only natural host. Phylogenetic analysis by 16S rRNA sequence comparison has suggested that the genus Haemophilus should be restricted to H. influenzae, H. aegyptius, H. haemolyticus, H. segnis, H. aphrophilus, H. paraphrophilus, and possibly Actinobacillus actinomycetemcomitans (Dewhirst et al., 1992 , 1993 ), whereas DNA–DNA hybridization studies have indicated a genus limited to H. influenzae, H. aegyptius, H. haemolyticus, H. parainfluenzae and ‘H. intermedius (Mutters et al., 1989 ). DNA–rRNA hybridization experiments have placed H. aphrophilus and A. actinomycetemcomitans on separate branches in the family Pasteurellaceae, quite remote from the three authentic genera in the family, Pasteurella, Actinobacillus and Haemophilus (De Ley et al., 1990 ).

Phylogenetic analysis of DNA sequences from translated genes is a different approach with several theoretical advantages (Palys et al., 1997 ). The method has worked well when applied to a number of bacterial genera, and different housekeeping genes have been employed (Cousineau et al., 1992 ; Mollet et al., 1997 ; Swanson et al., 1997 ; Poyart et al., 1998 ). The translation initiation factor 2 (IF2) is an essential protein involved in the initiation of protein synthesis in prokaryotes. The sequence of the encoding gene, infB, is highly conserved in the segments encoding the domains of IF2 responsible for binding and hydrolysis of GTP and for the affinity to , while the 5'-segments are more variable in both length and sequence. So far, infB has only been found as a single copy in the prokaryotic genome; it encompasses 2490 bp in H. influenzae (Fleischmann et al., 1995 ). The rationale of using the partial sequence of infB, encoding the GTP-binding domain of IF2, as a phylogenetic marker is its universal distribution and its conserved, yet sufficently variable, sequence. We have successfully used the partial sequence of infB in a study of the phylogeny of Enterobacteriaceae (Hedegaard et al., 1999 ). The deduced phylogeny could, in some cases, be supported by critical biochemical characteristics, such as the clustering of the mixed acid fermenters Salmonella, Citrobacter and Escherichia (Hedegaard et al., 1999 ).

The present study was undertaken to investigate the delineation and speciation of the genus Haemophilus as evaluated by partial infB sequences. The principal scope was the species that have qualified as true members of the genus by 16S rRNA sequencing or DNA hybridization, i.e. H. influenzae, H. haemolyticus, H. parainfluenzae, H. segnis, H. aphrophilus/paraphrophilus and A. actinomycetemcomitans, as well as the phenotypically related human parasites H. parahaemolyticus and H. paraphrohaemolyticus. Attempts were made to include enough clinical isolates to render possible a distinction of genetic clusters. A few single isolates or type strains of non-human origin from the non-Haemophilus genera of the family Pasteurellaceae were included to provide an outline of the family.


   METHODS
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Bacterial strains, characterization, and DNA preparation.
The strains sequenced in this work are listed in Table 1. Isolates from the collections at the Statens Serum Institut were kept at -80 °C in broth containing 10% (v/v) glycerol. Bacterial strains from other sources were obtained lyophilized. Upon inclusion in the study, cells were thawed or rehydrated and the identification was checked by characterization of X and V factor requirements, haemolysis on 5% horse blood agar, and production of tryptophanase (indole test), urease, ornithine decarboxylase and ß-galactosidase (ONPG test). In selected cases, carbohydrate fermentation and production of H2S were also investigated. Methods for characterization were performed as described by Kilian & Frederiksen (1981) , and biotypes were defined as described by Kilian (1991) . Genomic DNA was released using the Chelex method (De Lamballerie et al., 1992 ). In brief, a small loopful of cells was suspended in H2O, pelleted by centrifugation, and resuspended in 5% Chelex-100 (Bio-Rad). The cells were incubated at 56 °C for 15 min, Vortex-mixed for 10 s, and incubated at 100 °C for 8 min. After centrifugation at 10000 g for 2 min in a benchtop centrifuge, the supernatant was withdrawn and diluted tenfold in H2O. One microlitre was used for a 100 µl PCR reaction.


View this table:
[in this window]
[in a new window]
 
Table 1. Strains of Haemophilus, Actinobacillus and Pasteurella included in the study

 
PCR amplification and sequencing of infB.
Fragments of infB encoding the GTP-binding domain of IF2 were amplified by PCR using degenerate oligonucleotide primers and purified as previously described (Hedegaard et al., 1999 ). The purified PCR products were sequenced and analysed on an ALF DNA Sequencer (Amersham Pharmacia Biotech) using the 5'-fluorescein labelled primers 1748 Reverse (GTAGCAACCGGACCACGACCTTTAT) or 1745 Reverse (GCAACCGRICCICTICCTTTRTC; R=A or G, I=dInosine) (Hedegaard et al., 1999 ). With H. ducreyi CCUG 4438, a 3' mismatch prevented the use of the internal primers, and the sequencing was performed using the Thermo Sequenase dye terminator sequencing kit (Amersham Pharmacia Biotech) and the ABI377 DNA Sequencer (PE Biosystems) in combination with the primers used for the PCR amplification. A 453 bp DNA sequence and the deduced amino acid sequence were obtained and compared for all isolates, corresponding to the amino acid positions 424–562 with reference to Escherichia coli IF2-1 (Sacerdot et al., 1984 ; accession no. X00513). Sequences were checked by visual examination of the electropherograms, by translation into amino acid sequence, and finally by re-examination of all polymorphic positions. Ambiguities were resolved by repeating the PCR and sequencing reactions.

16S rRNA gene sequencing.
Partial 16S rRNA gene sequencing was performed using a commercial kit, Microseq 16S rDNA (PE Biosystems), according to the manufacturer’s instructions. Amplified fragments were separated and detected on a ABI310 Genetic analyser (PE Biosystems). For amplification and partial sequencing of the 16S rRNA gene of H. parahaemolyticus NCTC 8479T the following two primers were employed: 5'-TATTACCGCGGCTGCTGGCA-3' and 5'-TCAGATTGAACGCTGGCGGC-3' (Brosius et al., 1978 ). Sequencing in both directions was performed with a Thermo Sequanase dye terminator sequencing kit (Pharmacia Biotech) and analysed with the ABI377 DNA sequencer. Fragments of 423–426 bp (corresponding to nt 62–485 of H. influenzaeT 16S rRNA; accession no. M35019) were obtained from all isolates and used for analysis.

Sequence analysis.
Multiple alignments were performed and edited using the PILEUP and LineUp programs of the GCG package (the Genetic Computer Group, University of Wisconsin, Madison). Phylogenetic trees were constructed by the minimum-evolution (Kidd & Cavalli-Sforza, 1971 ; Rzhetshy & Nei, 1992 ) and the maximum-parsimony methods. The distances were corrected by the Kimura two-parameter method (Kimura, 1980 ) and initial trees were generated by the neighbour-joining algorithm or by stepwise addition with simple or random addition. For both minimum-evolution and maximum-parsimony analyses, trees were generated by a heuristic search with tree bisection-reconnection branch swapping. The robustness of the trees was examined by resampling with 100 bootstrap replication. The calculation of similarity between clusters given in Tables 2 and 3 was performed with even-weighted characters ignoring gaps and ambiguous positions. All sequence analyses were performed with PAUP 4.0.0d55 for UNIX (Smithsonian Institution).


View this table:
[in this window]
[in a new window]
 
Table 2. Similarity between infB sequence clusters of human haemophili

 

View this table:
[in this window]
[in a new window]
 
Table 3. 16S rRNA similarity between selected infB clusters

 

   RESULTS
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Amplification and partial sequencing of infB
The expected amplicon of approximately 650 bp was obtained with all bacterial strains investigated, and a 453 bp segment of infB was unambiguously determined and used for analysis. A total of 174 polymorphic positions were found in the investigated segment, with the majority of them being located at the third-base position in codons. In the deduced amino acid sequence, a total of 34 polymorphic positions were found. Compared to the sequence of IF2 from E. coli, all strains of Pasteurellaceae were found to possess an inserted lysine between glycine438 and methionine439 (E. coli IF2–1 numbering). Within the Pasteurellaceae, a single amino acid deletion was found in P. multocidaT at position 375 (H. influenzae IF2 numbering; Fleischmann et al., 1995 , accession no. AAC22933). The consensus sequence of five regions essential for the GTP-binding activity has previously been described (Bourne et al., 1991 ). Four of the regions are located within the studied segment of IF2 and were found to be highly conserved.

Phylogenetic analysis based on infB
The phylogenetic tree constructed by the minimum-evolution method using the partial sequences of infB is shown in Fig. 1. Analysis by the maximum-parsimony method resulted in more than 100 trees, and a 50% majority rule consensus tree was constructed and compared with the tree obtained with the minimum-evolution method. Branches in Fig. 1 marked with an asterisk are supported by the consensus tree from maximum-parsimony analysis. Fifty-nine strains clustered in nine sequence groups (Fig. 1), and these will be addressed below. The grouping from Fig. 1 is used in the calculation of similarity given in Tables 2 and 3.



View larger version (30K):
[in this window]
[in a new window]
 
Fig. 1. Minimum-evolution tree based on partial sequences of infB with evolutionary distance in terms of number of base substitutions weighted by the Kimura two-parameter model. The type strains (T) are emphasized in bold. The value above or to the left of the branches is the percentage of bootstrap replications supporting the branch. Only percentages above 50 are shown. Branches marked with an asterisk (*) are also found in the 50% majority rule consensus tree from the maximum-parsimony analysis. The infB sequence from E. coli (Sacerdot et al., 1984 ; accession no. X00513) is used as outgroup.

 
The H. influenzae branch (1). All strains with requirements for X and V factor were placed on a common branch by minimum-evolution and maximum-parsimony analyses. Typical clinical isolates of H. influenzae, as well as the type strains of H. influenzae and H. aegyptius, formed a very homogeneous subcluster (1A) with a sequence homology in excess of 98% (Fig. 1, Table 2). Six different biotypes were represented in the cluster, but no correlation with the phenotypic diversity was found in the distribution of sequences. Identical sequences were obtained from P1442 and P1491 (biotypes IV and III, respectively). Subcluster 1A may represent ‘H. influenzae sensu stricto’.

Seven other isolates (subcluster 1B) were clearly located on the H. influenzae branch, but were more distantly related to the ‘H. influenzae sensu strictu’ cluster. A sequence homology at the 94–97% level was found between subclusters 1A and 1B (Table 2). These isolates included two representatives of the so-called ‘H. influenzae cryptic genospecies’ (biotype IV, CCUG 31339 and 31340), which has been found predominantly in urogenital and neonatal infections (Quentin et al., 1996 ). Four strains phenotypically classified as H. haemolyticus were placed on a common branch in subcluster 1B, HK385T and U175 (indole-negative), and HK676 and HK680 (indole-positive). Two strains (HK855 and 856) were found to be haemolytic on primary isolation but expressed IgA1 protease activity, a property of H. influenzae rather than H. haemolyticus. Upon re-examination for the present study, both strains had lost the ability to haemolyse. HK856 was phenotypically classified as H. influenzae biotype II. HK855 produced gas from glucose and emitted H2S, while other phenotypic tests were indicative of H. influenzae biotype VII. With respect to sequence analysis, HK856 clustered with ‘H. influenzae sensu strictu’, whereas HK855 was more unique (Fig. 1).

The H. parainfluenzae branch (2). A complex pattern was obtained with 27 isolates representing H. parainfluenzae and strains tentatively identified as H. segnis. The cluster is depicted in Fig. 2 together with selected phenotypic characteristics of the individual isolates. By infB analysis, four separate groups were generated by the minimum-evolution method, but one of these was only represented by a single strain (HK23). The remaining 26 strains fell into three subclusters (2A, B, and C), as follows.



View larger version (36K):
[in this window]
[in a new window]
 
Fig. 2. Selected phenotypic characteristics of bacterial strains located on the H. parainfluenzae branch. The branching is based on infB data and is adapted from Fig. 1. nd, not determined.

 
Subcluster 2A, ‘H. parainfluenzae sensu stricto’. Eighteen strains constituted one heterogeneous cluster (2A) with a sequence similarity above the 92% level. All typical isolates of H. parainfluenzae, including the type strain, were found in this cluster. Moreover, the cluster encompassed haemolytic strains (of biotypes II and IV), and isolates which were difficult to differentiate from H. segnis (negative in the indole, urease, and ornithine decarboxylase tests). Six separate biotypes of H. parainfluenzae were included in the study, but the isolates were dispersed throughout subcluster 2A regardless of biotype or other variable phenotypic traits of H. parainfluenzae, haemolysis or production of ß-galactosidase (Fig. 2). Three strains, HK755, HK758 and HK759, were negative in the reactions used for biotyping of H. parainfluenzae, i.e. indole, urease and ornithine decarboxylase. However, they produced H2S and were more saccharolytic than typical representatives of H. segnis (Fig. 2; results from carbohydrate fermentation not shown). By infB analysis, these isolates were located in subcluster 2A and conceivably represent H. parainfluenzae biotype V (Kilian, 1991 ).

Subcluster 2B. The type strain and three clinical isolates of H. segnis were located in this cluster. Relatively large differences were observed, with sequence similarities between 92 and 95%. Strains complying phenotypically with H. segnis were also found in subcluster 4C (see below).

Subcluster 2C. Four clinical isolates constituted one homogeneous cluster clearly separated from all other strains of the H. parainfluenzae branch. These isolates expressed a distinct and peculiar phenotype, with haemolysis in combination with biochemical reactions representative of H. parainfluenzae biotype V (or H. segnis) (Fig. 2). They may be representatives of a hitherto undescribed species. The isolates were cultivated from various human specimens from Denmark and Brazil during a 20 year period (Table 1).

The H. parahaemolyticus branch (3). H. parahaemolyticus and H. paraphrohaemolyticus are species of close resemblance that have qualified as separate species by 16S rRNA sequence comparison (Dewhirst et al., 1992 , 1993 ), but not by phenotypical analysis (Kilian, 1976 ; Broom & Sneath, 1981 ). The two type strains and one additional clinical isolate of each species were included in this study. Phylogenetic analysis by partial infB sequence placed these four isolates on a common branch (cluster 3), distinct from all other strains investigated (Fig. 1). The cluster was relatively heterogeneous, with a sequence similarity in the range of 94–98% (Table 2).

The A. actinomycetemcomitans branch (4). A. actinomycetemcomitans (subcluster 4A), H. aphrophilus/paraphrophilus (4B), and representatives of H. segnis (4C) constituted three separate subclusters on a common branch only remotely related to the type species of the genus Haemophilus. Only small differences were observed between the five isolates of H. aphrophilus/paraphrophilus, or between the four A. actinomycetemcomitans, and the homogeneity of these species was comparable to that of ‘H. influenzae sensu strictu’ (Table 2). The interspecies similarity of these two subclusters was 88–90%, equivalent to the similarity between H. influenzae and H. parainfluenzae. Three clinical isolates phenotypically identified as H. segnis constituted subcluster 4C. They were positive in the ONPG reaction, in contrast to the four H. segnis isolates of cluster 2B (Fig. 2). No other phenotypic tests employed could discriminate between strains of clusters 2B and 4C (data not shown).

Unclustered strains. Six strains were found to be unclustered (the type strains of A. pleuropneumoniae, P. multocida, P. avium and H. ducreyi, plus CCUG 18944 and HK22) (Fig. 1).

Comparison of partial 16S rRNA gene sequences
16S rRNA gene sequences were available for the type strains investigated, and a fragment of approximately 424 bp was sequenced from 27 other study strains (Table 1). Despite repeated sequencing, eight bases were ambiguously determined, probably due to differences in sequence of the multiple rrn operons. The fragment studied encompassed 423 bp (H. paraphrohaemolyticus, accession no. M75076), 426 bp (three representatives of subcluster 2C, accession nos AJ290753 and AJ290755–6) and 424 bp (other strains). The previously published 16S rRNA gene sequences (used in Table 3) of the type species of H. influenzae, H. segnis, H. parainfluenzae and H. paraphrohaemolyticus contain 1, 10, 10, and 6 undetermined bases in the fragment under study.

An unexpectedly low sequence homology (92%) of the available 16S rRNA gene sequence (accession no. M75073) of the type strain of H. parahaemolyicus and the 16S rRNA gene sequence of strain HIM570-6 prompted us to re-examine the former. The 424 bp sequence of H. parahaemolyticus NCTC 8479T determined by us showed only 90% identity to the sequence of Dewhirst et al. (1992) for the same strain, but the expected high homology (98%) to H. parahaemolyticus strain HIM570-6. This indicates that the 16S rRNA sequence of the type strain of H. parahaemolyticus available in databases as M75073 is based on an incorrect strain.

The partial 16S rRNA sequence from the unclustered HK22 (accession no. AJ290746) was compared with sequences in the EMBL database and found to be most closely related to H. aphrophilus/paraphrophilus and A. actinomycetemcomitans (search results not shown).

Table 3 shows the 16S rRNA similarity between selected infB clusters, based on partial sequences from 27 strains. The clusters are listed by decreasing similarity to the type species of the genus. By partial 16S rRNA gene comparison, the seven isolates of H. segnis (infB subclusters 2B and 4C) did not split into separate 16S rRNA gene clusters (Table 3). Eleven representatives of subcluster 2A, ‘H. parainfluenzae sensu stricto were relatively homogeneous in their 16S rRNA gene sequences, and a distinction from the related subcluster 2C (‘haemolytic biotype V’) was indicated. The new 16S rRNA sequence of H. parahaemolyticusT confirmed the existence of one separate cluster encompassing strains representative of H. parahaemolyticus and H. paraphrohaemolyticus (Table 3).


   DISCUSSION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Phylogenetic analysis based on single gene sequences is linked to the concept of bacterial populations as arrays of stable lineages (clones) that maintain their chromosomal gene composition with little or no rearrangement for extended periods of time (Ørskov & Ørskov, 1983 ). Recombination events involving the segment of infB studied would segregate the phylogeny of the gene from the phylogeny of the individual bacteria. Strains of most Haemophilus species are competent for transformation (Albritton et al., 1984 ). Accordingly, comparative sequence studies of selected genes encoding putative virulence factors have revealed mosaic structures, which is clear evidence of recombination (Proulx et al., 1991 ; Lomholt et al., 1995 ). Nevertheless, population-genetic studies based on large collections of capsulated and non-capsulated isolates of H. influenzae suggest that the extent of recombination is not sufficient to disrupt a basically clonal population structure (Porras et al., 1986 ; Musser et al., 1990 ). Less is known of the other members of the genus, but a single recent study of the 16S/23S ribosomal spacer from H. parainfluenzae revealed a mosaic-like organization of sequence blocks (Privitera et al., 1998 ). Recombination per se does not necessarily exclude the possibility of obtaining a valid picture of the phylogeny of bacterial species and genera based on sequence analyses of single housekeeping genes, provided that recombination is restricted to occur exclusively, or almost exclusively, within individual taxa.

In general, the phylogenetic picture obtained in this study by sequence analyses of infB was in good agreement with the delineation of the genus Haemophilus suggested by DNA hybridization (Mutters et al., 1989 ). An exception was observed for seven presumed representatives of H. segnis, which fell into two separate clusters by infB analysis, but not by 16S rRNA sequence comparison. The distance between subcluster 2B on the H. parainfluenzae branch and subcluster 4C on the A. actinomycetemcomitans branch (Fig. 1) is so wide that it makes recombination a plausible explanation. DNA–DNA hybridization, as well as 16S rRNA sequencing, has clearly demonstrated a close affiliation of H. segnis to H. aphrophilus/paraphrophilus (Potts et al., 1986 ; Dewhirst et al., 1992 ). It is conceivable that the recombinational transfer of the infB locus took place from H. parainfluenzae to H. segnis, with subcluster 4C as a likely candidate for the correct phylogenetic position of the overall H. segnis chromosome. Based on only seven strains, subclusters 2B and 4C were discriminated from one another by results of the ONPG test. DNA hybridization experiments involving strains of subclusters 2B and 4C will be needed to finally clarify their relationship.

Based on infB sequence comparison, the genus Haemophilus is composed of two major divisions. The first includes H. influenzae, H. aegyptius, H. haemolyticus, and possibly one or more cryptic genospecies. The designation ‘H. intermedius’ has been suggested for strains of a particular genotype found by DNA hybridization (Mutters et al., 1989 ), and Quentin et al. (1996) characterized a subset of H. influenzae, biotype IV, as a cryptic genospecies. Only representatives of the latter were included in the present study (CCUG 31339 and 31340), and they occupied a unique position in the infB dendrogram (subcluster 1B, Fig. 1), in agreement with 16S rRNA sequences published by Quentin et al. (1996) . Sequences obtained with various haemolytic isolates showed relatively wide differences, and the sequence of HK855 was as unique as the ones obtained with the cryptic genospecies of Quentin and co-workers. H. haemolyticus and H. aegyptius are not very different from H. influenzae phenotypically, by infB analysis (Table 2), or by 16S rRNA sequence comparison (Dewhirst et al., 1993 ). Indeed, DNA hybridization, which is the currently accepted reference methodology, suggests that H. aegyptius and H. influenzae do not merit separate species rank (Casin et al., 1986 ). How far speciation should be pursued in the H. influenzae branch is not clear at present.

Bacteria classified as H. parainfluenzae constitute the other major division of the genus Haemophilus. By infB sequence analysis, large differences were encountered between isolates of H. parainfluenzae; in fact, the dissimilarity between 18 isolates of ‘H. parainfluenzae sensu strictu’ (subcluster 2A) exceeded differences between subclusters 1A and 1B (Table 2). Broad differences have previously been noticed within H. parainfluenzae (Pohl, 1981 ). No correlation was found between the phylogenic relationships suggested by partial infB sequences and the phenotypic traits within subcluster 2A (Fig. 2). It is possible that the distribution of infB sequences in cluster 2A is a product of intraspecies recombination. The phylogeny of other genes located at a different position on the bacterial chromosome would clarify this hypothesis.

The combination of negative reactions in the three biochemical tests used for biotyping of H. parainfluenzae (indole production, urease and ornithine decarboxylase) is characteristic of both H. paraphrophilus and H. segnis. These two species are differentiated by the stronger acid production of the former and its ability to ferment lactose. However, it has been unclear if strains of H. parainfluenzae negative in the three biotyping tests exist. The clustering of non-haemolytic strains with these characteristics according to both infB and 16S rRNA sequences (Fig. 1, Table 3) clearly demonstrates two different positions of such strains. While some strains clustered together with the type strain of H. segnis, other strains (HK755, HK758, HK759) clustered among strains of H. parainfluenzae. In contrast to strains of H. segnis, the latter strains produced H2S like most strains of H. parainfluenzae (Fig. 2). Thus, these strains represent H. parainfluenzae biovar V, a designation reserved for this combination of characteristics (Kilian, 1991 ).

Subcluster 2C on the H. parainfluenzae branch may qualify for status as a separate species, with four isolates being unique with respect to phenotype, 16S rRNA, and infB sequences (Figs 1 and 2, Table 3). Three of these strains were positive in the H2S reaction (Fig. 2) and their studied fragment of the 16S rRNA gene encompassed 426 bp, in contrast to 423–424 bp found in other strains with available 16S rRNA sequences (Table 1, alignment not shown).

The type strains of H. parahaemolyticus and H. paraphrohaemolyticus, plus two additional isolates (P1300 and HIM570-6), had related infB sequences clearly separate from all other sequences generated in the study. The distinction of these species from each other by phenotypic traits is difficult (Kilian, 1976 ). Strain P1300 was positive in the ONPG test and negative for IgA1 protease activity, which is why identification of H. paraphrohaemolyticus was preferred to identification as H. parahaemolyticus. By infB sequence, a division into two separate species is not found (Fig. 1), in contrast to other molecular data (Mutters et al., 1989 ; Dewhirst et al., 1993 ). However, re-examination of the 16S rRNA gene sequence of the type strain of H. parahaemolyticus clearly indicates that the sequence deposited by Dewhirst et al. (1993) , which often has been part of the arguments for reorganization of the genus Haemophilus, is not based on the correct strain. The sequence obtained in the present study for this type strain showed the expected high homology to another verified strain of H. parahaemolyticus (HIM570-6) and to the type strain of H. paraphrohaemolyticus, in full agreement with the observed infB homologies. The interrelationship and affiliation to genus of these species should, however, be considered unresolved, as a limited number of strains has been included in this and previous investigations.

The close relationship of A. actinomycetemcomitans, H. aphrophilus/paraphrophilus and H. segnis has been demonstrated by 16S rRNA sequence (Dewhirst et al., 1992 , 1993 ), by DNA hybridization (Potts et al., 1986 ), and by infB sequence (Fig. 1, subclusters 4A–C). It is the affiliation to genus of these species that is a matter of uncertainty. The results from 16S rRNA comparison showed them to be close relatives of H. influenzae, the type species of the genus Haemophilus (Dewhirst et al., 1992 ). This is in conflict with the results obtained by DNA–rRNA hybridization (deLey et al., 1990 ) and by infB sequence comparison, which have found these species to be only remotely related to the genus Haemophilus.

Five strains, the type strains of H. ducreyi, P. multocida, P. avium and A. pleuropneumoniae, plus CCUG 18944, were included in order to give an outline of the family Pasteurellaceae. The last four, which were clearly separate from the genus Haemophilus by infB analysis, were of non-human origin. Of these four strains, three have a requirement for V factor. CCUG 18944 complies phenotypically with H. parainfluenzae biotype I, although fermentation of sucrose was not detected (data not shown). The human origin thus appears to be more predictive of affiliation to the genus than the V factor requirement. The taxonomic position of species represented by only a single strain should naturally await analysis of a greater number of representative isolates. The only H. ducreyi examined in the study was the type strain, as this species is known to be only remotely related to the type species of the genus Haemophilus. Speciation of the sixth unclustered strain, HK22, presents a particular problem. This human isolate is phenotypically a H. parainfluenzae biotype III (Kilian, 1976 ), it is related to H. aphrophilus by partial 16S rRNA gene sequence (accession no. AJ290746), and it is related to H. paraphrohaemolyticus by DNA hybridization [HK22 was included in the DNA hybridization experiments done by Mutters et al. (1989) . The matrix of individual DNA-binding values obtained between strains is given in the PhD thesis of Burbach (1987) .]

In conclusion, the same delineation of the genus ‘Haemophilus sensu stricto’ is suggested by infB sequence comparison and DNA–DNA hybridization. Quite a different phylogeny is inferred from 16S rRNA sequence analysis, where H. aphrophilus/paraphrophilus, H. segnis, and possibly A. actinomycetemcomitans, are included, whereas H. parainfluenzae is excluded from the genus. More data on the relations of the genus Haemophilus and family Pasteurelleceae are needed, and the discrepancies above emphasize the potential problems of basing conclusions on studies based solely on 16S rRNA sequences or other single gene loci.


   ACKNOWLEDGEMENTS
 
The study was financially supported by Aalborg Stiftstidendes Julelotteri and by grants from the Danish Natural Science Research Council to the Department of Molecular and Structural Biology (28807-9502036, 9602401).


   REFERENCES
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Albritton, W. L., Setlow, J. K., Thomas, M., Sottnek, F. & Steigerwalt, A. G. (1984). Heterospecific transformation in the genus Haemophilus. Mol Gen Genet 193, 358-363.[Medline]

Bourne, H. R., Sanders, D. A. & McCormick, F. (1991). The GTPase superfamily: conserved structure and molecular mechanism. Nature 349, 117-127.[Medline]

Broom, A. & Sneath, P. H. (1981). Numerical taxonomy of Haemophilus. J Gen Microbiol 126, 123-149.[Medline]

Brosius, J., Palmer, M. L., Kennedy, P. J. & Noller, H. F. (1978). Complete nucleotide sequence of a 16S ribosomal RNA gene from Escherichia coli. Proc Natl Acad Sci USA 75, 4801-4805.[Abstract]

Burbach, S. (1987). Reklassifizierung der Gattung Haemophilus Winslow et al., 1917 auf Grund der DNA-Basensequenzhomologie. Inaugural Dissertation, Philipps-Universität Marburg, Marburg, Germany.

Casin, I., Grimont, F. & Grimont, P. A. (1986). Deoxyribonucleic acid relatedness between Haemophilus aegyptius and Haemophilus influenzae. Ann Inst Pasteur Microbiol 137B, 155-163.

Cousineau, B., Cerpa, C., Lefebvre, J. & Cedergren, R. (1992). The sequence of the gene encoding elongation factor Tu from Chlamydia trachomatis compared with those of other organisms. Gene 120, 33-41.[Medline]

De Lamballerie, X., Zandotti, C., Vignoli, C., Bollet, C. & de Micco, P. (1992). A one-step microbial DNA extraction method using ‘‘Chelex 100’’ suitable for gene amplification. Res Microbiol 143, 785-790.[Medline]

De Ley, J., Mannheim, W., Mutters, R. & 7 other authors (1990). Inter- and intrafamilial similarities of rRNA cistrons of the Pasteurellaceae. Int J Syst Bacteriol 40, 126–137.[Abstract]

Dewhirst, F. E., Paster, B. J., Olsen, I. & Fraser, G. J. (1992). Phylogeny of 54 representative strains of species in the family Pasteurellaceae as determined by comparison of 16S rRNA sequences. J Bacteriol 174, 2002-2013.[Abstract]

Dewhirst, F. E., Paster, B. J., Olsen, I. & Fraser, G. J. (1993). Phylogeny of the Pasteurellaceae as determined by comparison of 16S ribosomal ribonucleic acid sequences. Zentbl Bakteriol 279, 35-44.

Fleischmann, R. D., Adams, M. D., White, O. & 37 other authors (1995). Whole-genome random sequencing and assembly of Haemophilus influenzae Rd. Science 269, 496–512.[Medline]

Haubek, D., Dirienzo, J. M., Tinoco, E. M., Westergaard, J., Lopez, N. J., Chung, C. P., Poulsen, K. & Kilian, M. (1997). Racial tropism of a highly toxic clone of Actinobacillus actinomycetemcomitans associated with juvenile periodontitis. J Clin Microbiol 35, 3037-3042.[Abstract]

Hedegaard, J., Steffensen, S. A. de A., Nørskov-Lauritsen, N., Mortensen, K. K. & Sperling-Petersen, H. U. (1999). Identification of Enterobacteriaceae by partial sequencing of the gene encoding translation initiation factor 2. Int J Syst Bacteriol 49, 1531-1538.[Abstract]

Kidd, K. K. & Cavalli-Sforza, K. (1971) Number of characters examined and error in reconstruction of evolutionary trees. In Mathematics in the Archaeological and Historical Sciences, pp. 335–346. Edited by F. R. Hodson & P. Tautu. Edinburgh: Edinburgh University Press.

Kilian, M. (1976). A taxonomic study of the genus Haemophilus, with the proposal of a new species. J Gen Microbiol 93, 9-62.[Medline]

Kilian, M. (1991). Haemophilus. In Manual of Clinical Microbiology , pp. 463-470. Edited by A. Balows, W. J. Hausler, K. L. Herrmann, H. D. Isenberg & H. J. Shadomy. Washington, DC:American Society for Microbiology.

Kilian, M. & Biberstein, E. L. (1984). Haemophilus Winslow, Broadhurst, Buchanan, Krumwiede, Rogers and Smith 1917, 561AL. In Bergey’s Manual of Systematic Bacteriology , pp. 558-569. Edited by N. R. Krieg & J. G. Holt. Baltimore:Williams & Wilkins.

Kilian, M. & Frederiksen, W. (1981). Identification tables for the Haemophilus–Pasteurella–Actinobacillus group. In Haemophilus, Pasteurella and Actinobacillus , pp. 281-290. Edited by M. Kilian, W. Frederiksen & E. L. Biberstein. London:Academic Press.

Kimura, M. (1980). A simple method for estimating evolutionary rates of base substitutions through comparative studies of nucleotide sequences. J Mol Evol 16, 111-120.[Medline]

Lomholt, H., Poulsen, K. & Kilian, M. (1995). Comparative characterization of the iga gene encoding IgA1 protease in Neisseria meningitidis, Neisseria gonorrhoeae and Haemophilus influenzae. Mol Microbiol 15, 495-506.[Medline]

Mollet, C., Drancourt, M. & Raoult, D. (1997). rpoB sequence analysis as a novel basis for bacterial identification. Mol Microbiol 26, 1005-1011.[Medline]

Musser, J. M., Kroll, J. S., Granoff, D. M. & 24 other authors (1990). Global genetic structure and molecular epidemiology of encapsulated Haemophilus influenzae. Rev Infect Dis 12, 75–111.[Medline]

Mutters, R., Piechulla, K., Hinz, K. H. & Mannheim, W. (1985). Pasteuella avium (Hinz and Kunjara 1977) comb. nov. and Pasteurella volantium sp. nov. Int J Syst Bacteriol 35, 5-9.

Mutters, R., Mannheim, W. & Bisgaard, M. (1989). Taxonomy of the group. In Pasteurella and Pasteurellosis , pp. 3-34. Edited by C. Adlam & J. M. Rutter. London:Academic Press.

Ørskov, F. & Ørskov, I. (1983). From the national institutes of health. Summary of a workshop on the clone concept in the epidemiology, taxonomy, and evolution of the Enterobacteriaceae and other bacteria. J Infect Dis 148, 346-357.[Medline]

Palys, T., Nakamura, L. K. & Cohan, F. M. (1997). Discovery and classification of ecological diversity in the bacterial world: the role of DNA sequence data. Int J Syst Bacteriol 47, 1145-1156.[Abstract/Free Full Text]

Pohl, S. (1981). DNA relatedness in HPA. In Haemophilus, Pasteurella and Actinobacillus , pp. 245-253. Edited by M. Kilian, W. Frederiksen & E. L. Biberstein. London:Academic Press.

Pohl, S., Bertchinger, W., Frederiksen, W. & Mannheim, W. (1983). Transfer of Haemophilus pleuropneumoniae and the Pasteurella haemolytica-like organism causing porcine necrotic pleuropneumonia to the genus Actinobacillus (Actinobacillus pleuropneumoniae comb. nov.) on the basis of phenotypic and deoxyribonucleic acid relatedness. J Syst Bacteriol 33, 510-514.

Porras, O., Caugant, D. A., Gray, B., Lagergard, T., Levin, B. R. & Svanborg-Eden, C. (1986). Difference in structure between type b and nontypable Haemophilus influenzae populations. Infect Immun 53, 79-89.[Medline]

Potts, T. V., Mitra, T., O’Keefe, T., Zambon, J. J. & Genco, R. J. (1986). Relationships among isolates of oral haemophili as determined by DNA–DNA hybridization. Arch Microbiol 145, 136-141.[Medline]

Poyart, C., Quesne, G., Coulon, S., Berche, P. & Trieu-Cuot, P. (1998). Identification of streptococci to species level by sequencing the gene encoding the manganese-dependent superoxide dismutase. J Clin Microbiol 36, 41-47.[Abstract/Free Full Text]

Privitera, A., Rappazzo, G., Sangari, P., Giannino, V., Licciardello, L. & Stefani, S. (1998). Cloning and sequencing of a 16S/23S ribosomal spacer from Haemophilus parainfluenzae reveals an invariant, mosaic-like organisation of sequence blocks. FEMS Microbiol Lett 164, 289-294.[Medline]

Proulx, C., Munson, R. S. J., Grass, S., Hamel, J., Martin, D. & Brodeur, B. R. (1991). Identification of a surface-exposed immunodominant epitope on outer membrane protein P1 of Haemophilus influenzae type b. Infect Immun 59, 963-970.[Medline]

Quentin, R., Ruimy, R., Rosenau, A., Musser, J. M. & Christen, R. (1996). Genetic identification of cryptic genospecies of Haemophilus causing urogenital and neonatal infections by PCR using specific primers targeting genes coding for 16S rRNA. J Clin Microbiol 34, 1380-1385.[Abstract]

Rivers, T. M. (1922). Influenzae-like bacilli: growth of influenzae-like bacilli on media containing only an autoclave-labile substance as an accessory food factor. Johns Hopkins Hosp Bull 33, 429-431.

Rzhetshy, N. & Nei, M. (1992). A simple method for estimating and testing minimum-evolution trees. Mol Biol Evol 9, 945-967.[Free Full Text]

Sacerdot, C., Dessen, P., Hershey, J. W., Plumbridge, J. A. & Grunberg-Manago, M. (1984). Sequence of the initiation factor IF2 gene: unusual protein features and homologies with elongation factors. Proc Natl Acad Sci USA 81, 7787-7791.[Abstract]

Sedlacek, I., Gerner-Smidt, P., Schmidt, J. & Frederiksen, W. (1993). Genetic relationship of strains of Haemophilus aphrophilus, H. paraphrophilus, and Actinobacillus actinomycetemcomitans studied by ribotyping. Zentbl Bakteriol 279, 51-59.

Swanson, D. S., Kapur, V., Stockbauer, K., Pan, X., Frothingham, R. & Musser, J. M. (1997). Subspecific differentiation of Mycobacterium avium complex strains by automated sequencing of a region of the gene (hsp65) encoding a 65-kilodalton heat shock protein. Int J Syst Bacteriol 47, 414-419.[Abstract/Free Full Text]

Tjøtta, T. & Avery, O. T. (1921). Studies on bacterial nutrition. II. Growth accessory substances in the cultivation of haemophilic bacilli. J Exp Med 34, 97-114.

Received 22 January 2001; revised 3 May 2001; accepted 9 May 2001.