Evidence for recombination between a sialidase (nanH) of Actinomyces naeslundii and Actinomyces oris, previously named ‘Actinomyces naeslundii genospecies 1 and 2’

Actinomyces spp., predominant members of human oral biofilms, may use extracellular sialidase to promote adhesion, deglycosylate immunoglobulins and liberation of nutrients. Partial nanH gene sequences (1077 bp) from Actinomyces oris (n =74), Actinomyces naeslundii (n =30), Actinomyces viscosus (n =1) and Actinomyces johnsonii (n =2) which included the active-site region and the bacterial neuraminidase repeats (BNRs) were compared. The sequences were aligned and each species formed a distinct cluster with five isolates having intermediate positions. These five isolates (two A. oris and three A. naeslundii) exhibited interspecies recombination. The nonsynonymous/synonymous ratio was <1 for both A. oris and A. naeslundii indicating that nanH in both species is under stabilizing selective pressure; nonsynonymous mutations are not selected. However, for A. oris significant negative values in tests for neutral selection suggested the rate of mutation in A. oris was greater than in A. naeslundii but with selection against nonsynonymous mutations. This was supported by the observation that the frequency of polymorphic sites in A. oris, which were monomorphic in A. naeslundii was significantly greater than the frequency of polymorphic sites in A. naeslundii which were monomorphic in A. oris (χ2=7.011; P =0.00081). The higher proportions of A. oris in the oral biofilm might be explained by the higher mutation rate facilitating an increased ability to respond successfully to environmental stress.

Sialidase activity also appears important in the nutrition of the oral biofilm because sialidase activity increases in the absence of host diet, and withdrawal of the normal diet increases the proportions of sialidase-producing bacteria Lucas et al., 1997;Sheehy et al., 2000). Sialic acid is utilized by various oral streptococci, Actinomyces naeslundii and by different mixed populations of subgingival plaque bacteria (ter Steeg et al., 1987;Frandsen, 1994;Byers et al., 1996Byers et al., , 1999Homer et al., 1996). Sialidase also exposes galactose residues of O-and N-linked glycans which mediate the adherence of human A. naeslundii strains described previously as genospecies 1 and 2 to human glycans including those of the salivary pellicle (Costello et al., 1979;Gibbons et al., 1990).
Two of the strains from which the nanH was sequenced were described as Actinomyces viscosus but such human strains were subsequently described as A. naeslundii genospecies 2 (Johnson et al., 1990). However, on the basis of a phylogenetic analysis of partial gene sequences of housekeeping genes, we (Henssge et al., in press) have reported that strains previously identified as A. naeslundii genospecies 2, including the sequenced strain MG1, should be reclassified as Actinomyces oris. The species A. naeslundii has been amended to include only strains previously identified as A. naeslundii genospecies 1. Therefore, we have nanH sequence data on three A. oris strains but no information of the nanH genes of A. naeslundii, Actinomyces johnsonii (previously A. naeslundii genospecies WVA 963) or animal strains of A. viscosus.
In this report, we have compared the nucleotide and amino acid sequences of the region of the nanH gene containing the active site (Crennell et al., 1993) and the five Asp boxes or bacterial neuraminidase repeats (BNRs, Henningsen et al., 1991) of these Actinomyces species. In the mouth A. naeslundii and A. oris occupy the same sites but A. oris is the predominant species (Bowden et al., 1999) while A. johnsonii was isolated from the gingival crevice (Johnson et al., 1990). We present a phylogenetic analysis of a partial sequence of nanH in these four species and present evidence of interspecies recombination between nanH genes.

Bacterial strains
The isolates have all been reported and identified in a previous taxonomic study (Henssge et al., in press) (NCTC 10951). The oral and clinical isolates were given study numbers from 1 to 94. All isolates were cultured anaerobically at 37 1C on Fastidious Anaerobe Agar (LabM) supplemented with 5% (v/v) defibrinated horse blood and stored at À 80 1C in brainheart infusion (Oxoid) containing 50% glycerol.

PCR conditions and nanH sequencing
DNA was extracted from cells as described previously (Henssge et al., in press). Owing to the length of the gene fragment which was likely to contain the sites of interest, two pairs of primers were used to obtain a 1077-bp fragment of the nanH gene in all species: Sial-F1 5 0 -ACACGATCACG CAAGCCGA-3 0 and Sial-R1 5 0 -CGACCTTGTTCTCATC CA-3 0 and Sial-F2 5 0 -AACCACATCGTCCA-3 0 and Sial-R2 5 0 -GAGCCAGTTCATCGTGAA-3 0 . The PCRs were performed using Reddymix (Abgene, Epsom, Surrey, UK) and the PCR conditions for each pair of primers were an initial denaturation for 10 min at 94 1C followed by 30 cycles of 94 1C for 30 s, 49 1C for 30 s and 72 1C for 90 s. A final extension was carried out for 5 min at 72 1C. The PCR products were visualized on a 1% agarose gel stained with Gel Red (Biotium Inc.). The amplicons were cleaned using a 50 : 50 mixture of 40% polyethylene glycol and 3 M NaCl, washed twice with 70% ethanol and rehydrated in sterile water. The same primers were used for the sequencing reactions and all amplicons were sequenced in both directions using the BigDye Terminator Sequencing kit (Applied Biosystems) and reaction products were run on a 3730xl sequencer (Applied Biosystems).

nanH sequence analysis
The DNA sequences were aligned using BIOEDIT (http:// www.mbio.ncsu.edu/BioEdit/bioedit.html). The phylogenetic relationships between partial nanH nucleotide sequences of the reference strains, the human oral and clinical isolates and the two 'A. viscosus' nanH sequences in GenBank (L06898 and X62276) and the nanH of the sequenced strain MG1 (ANA2709) were analysed using MEGA 4 (Tamura et al., 2007). Distances were calculated using the Kimura two-parameter model and for clustering the neighbour-joining method of Saitou & Nei (1987) using bootstrap values based on 500 replicates was used. The amino acid sequences were clustered using CLUSTALW2 (http://www.ebi.ac.uk/). DNASP (Rozas et al., 2003) was used to investigate the nanH sequences of A. naeslundii and A. oris. The G1C content of the partial genes sequences, the number of discrete sequences, the number of polymorphic sites, average nonsynonymous/synonymous ratios (dN/dS) were calculated and nucleotide diversity was estimated by determining p (nucleotide diversity) and y (the total number of mutations). To test for neutral molecular evolution three tests were used; Tajima's D, based on the differences between the number of segregating sites and the average number of nucleotide differences, Fu and Li's D Ã , based on the differences between the number of singletons (mutations appearing only once among the sequences) and the total number of mutations and Fu and Li's F Ã , based on the differences between the number of singletons, and the average number of nucleotide differences between pairs of sequences. The extent of the DNA sequence polymorphism between A. naeslundii and A. oris was calculated and compared using a w 2 test.
Split decomposition trees were constructed with 1000 bootstrap replicates based on parsimony splits as implemented in the SPLITSTREE 4.0 (Hudson & Bryant, 2006) and the statistic phi was calculated. To identify strains with evidence of recombination the Recombination Detection Package (Martin et al., 2005) was used with the default settings.

Identification of BNR sequences and active site residues
The BNRs were identified using the InterProScan Sequence Search tool (http://www.ebi.ac.uk/Tools/InterProScan/). The amino acid polymorphisms in the BNRs were determined by visual inspection. The partial nanH amino acid sequences of the four type strains were aligned with the sequence of a Salmonella typhimurium sialidase (M55342) using Parallel Protein Information Analysis System (http:// www.cbrc.jp/papia/papia.html) and the 12 putative active site residues of the S. typhimurium sialidase (Crennell et al., 1993) were identified.

Detection of nanH transcripts
Each type strain was grown on FAA supplemented with 5% (v/v) defibrinated horse blood and total RNA was extracted using the UltraClean Microbial RNA Isolation Kit (MO BIO Laboratories Inc.). Reverse transcription was performed using the Omniscript RT kit (Qiagen Ltd) with primers, Sial-F1 and Sial-R1, for nanH and AtpA-F (CCCTGGAG TACACCACCAT) and AtpA-R (CGCCAGGGTGATCTT GAG), to amplify the housekeeping gene, atpA (ATP synthase F1, a subunit, ANA_0169). The thermal cycling was as follows: cDNA synthesis at 37 1C for 30 min, 94 1C for 1 min, followed by amplification with 40 cycles of denaturation at 94 1C for 20 s, annealing at 55 1C for 30 s, extension at 72 1C for 1 min, and then the final extension at 72 1C for 5 min. The products were run on a 1% agarose gel, with Gel Red incorporated, for 10 min at 100 V. The identity of the amplicons was confirmed by sequencing as described above.

Results
The construction of a phylogenetic tree using the nanH sequences indicated that the four species were separated from each other except that two of A. oris strains (strains 60 and 61) were separate from the major A. oris cluster and three A. naeslundii strains (strains 25, 51 and CCUG 34725) were distinct from the major A. naeslundii cluster (Fig. 1). Essentially, the same tree topography was found for the analysis of the derived amino acid sequences (data not shown) with the same isolates being found on the periphery of the clusters composed of the majority of the A. oris and A. naeslundii isolates and clearly distinct from the A. johnsonii and A. viscosus sequences.
The dN/dS values for A. oris and A. naeslundii were 0.1466 and 0.2134, respectively. Tajima's D, Fu and Li's D Ã and F Ã were all significantly negative for A. oris but did not achieve significance for A. naeslundii (Table 1). The test for DNA divergence between populations implemented in DNASP indicated that the number of polymorphic sites in A. naeslundii and monomorphic in A. oris was 59 but the number of polymorphic sites in A. oris and monomorphic in A. naeslundii was 287 (w 2 = 7.011; P = 0.00081).
Analysis of the amino acid sequences of the BNRs of the four species is shown in supporting Table S1. There was overall great consistency in the pattern SXDXGXTW within each of the BNRs for all of the species with only single exceptions in the first for BNRs. In the fifth BNR the pattern  was SXDXGXSW in all of the isolates except for one A. johnsonii isolate which had the sequence SCGNGASW.
The alignment of the amino acid sequences inferred from the partial nanH sequences from the four species with that of S. typhimurium LT2 demonstrated that the active site identified in the S. typhimurium sialidase was recognized in each of the Actinomyces sequences. Nine of the 12 amino acids in the S. typhimurium active site were identical in all the Actinomyces sequences (supporting Fig.  S1). The exceptions among the sequences were the met-99 and trp-121 in S. typhimurium which were both replaced by serine and leu-175 which was replaced by phenylalanine.
To test for the statistical evidence of recombination in the first instance the Splitstrees method was used and the phi test provided evidence of significant recombination when all strains, except the A. viscosus and A. johnsonii strains, were included in the analysis (P = o 10 À20 ). Consideration of A. oris or A. naeslundii strains alone yielded phi values with P = 3.43 Â 10 À6 and P = 1.37 Â 10 À5 , respectively, indicating statistically significant evidence of recombination in each of the species. To identify the strains with evidence of recombination we analysed the data of these two species together using the seven programs within the RDP suite and found that only A. naeslundii strains 25, 51 and CCUG 34725 and A. oris strains 60 and 61 gave significant evidence of recombination. The recombination events are summarized in Fig. 2. In the A. naeslundii strains 25, 51 and CCUG 34725 and A. oris strain 60 the recombination event extended beyond the available partial sequences.
The type strains of each of the four species examined expressed the nanH gene when grown on a complex medium containing defibrinated horse blood and 1% (w/v) glucose (Fig. 3) and sequence analysis confirmed that the individual products were derived from the nanH genes of the appropriate species.

Discussion
The complete functions of sialidase in the oral cavity are certainly not fully understood but dietary restriction in macaque monkeys resulted in the increase in levels of sialidase activity in the oral biofilm suggesting a role in bacterial nutrition . The levels of sialidase-producing bacteria also increase in the absence of host's diet, again supporting the hypothesis that it plays a role in providing nutrient, both C and N, for microbial growth and proliferation Lucas et al., 1997;Sheehy et al., 2000). Bacterial sialidases, including those of the oral Actinomyces, also expose galactose residues of glycoprotein glycans located on mucosal surfaces or adsorbed to the tooth as part of the enamel pellicle. The exposed galactose residues may then be used as attachment sites by members of the oral biofilm promoting dental plaque accumulation (Costello et al., 1979;Gibbons et al., 1990). All sialidase enzymes have a highly conserved array of residues, which in the tertiary structure of the enzyme, form the active and binding sites of the molecule (Crennell et al., 1993). In addition to these residues all sialidases also have a number of BNRs structural entities for which a function cannot be definitely ascribed because Asp boxes are found in many protein families (Henningsen et al., 1991;Copley et al., 2001). In determining the sequence diversity of nanH in these Actinomyces species, we selected a portion of the gene which contained the 12 residues involved in substrate interactions with, and stabilization of, the active site and the BNRs originally described in 'A. viscosus' strain DSM 43798, now identified as A. oris. The gene fragment of 1077 bp represents c. 30% of the entire nanH sequence in A. oris. The determination of the partial sequences of the nanH genes from multiple isolates of A. oris and A. naeslundii has enabled the amino acid conservation within each species and nucleotide sequence variation between the multiple independent strains of A. oris and A. naeslundii to be determined. The phylogenetic analysis clearly demonstrated that these two species have nanH genes with different nucleotide sequences. It was also apparent that the nucleotide sequences of A. johnsonii and A. viscosus were characteristic and that, despite considerable site-specific conservation of amino acid residues, the amino acid sequences of all four species were distinct and characteristic. The BNRs of the four Actinomyces species showed a very high degree of conservation with respect to the amino acid sequence as would be expected (Vimr, 1994). However, in the fifth BNR the threonine residue was replaced by a serine in all the 107 sequences determined; such a structurally synonymous substitution would not be expected to significantly modify the activity of the enzyme. The same variations observed here in the conserved amino acids involved in the active-site region have also been reported in other high G1C organisms (Sakurada et al., 1992;Jost et al., 2001).
The dN/dS values were o 1 for both A. oris and A. naeslundii indicating that nanH in both species is under stabilizing selective pressure; nonsynonymous mutations are not selected. However, for A. oris significant negative values Tajima's D and Fu and Li's D Ã and F Ã were calculated suggesting that, as all nucleotides of any codon are assumed to be equally mutatable, the rate of mutation in A. oris must be greater than in A. naeslundii but with selection against nonsynonymous mutations. The proposed higher rate of mutation in the nanH of A. oris was supported by the observation that the frequency of polymorphic sites in A. oris which were monomorphic in A. naeslundii was significantly greater than the frequency of polymorphic sites in A. naeslundii which were monomorphic in A. oris. Consideration of our previously published sequence data on six housekeeping genes (Henssge et al., in press) indicated that the frequency of polymorphic sites in A. oris which were monomorphic in A.naeslundii was also significantly (P o 0.05) greater in gltA, metG, pgi and rpoB, while in gyrA and atpA no significant difference was found. The reasons for the apparent greater frequency of mutations in A. oris is not known but may suggest that A. oris is subject to greater stress than A. naeslundii in the oral biofilm because environmental stress may increase the rate of mutation of organisms in biofilms (Bjedov et al., 2003;Tenaillon et al., 2004). However, given the structural requirements of a functioning sialidase, nonsynonymous mutations may not be beneficial and will, therefore, not be selected. Actinomyces oris (as A. naeslundii genospecies 2) is numerically more successful in the oral biofilm than A. naeslundii (as A. naeslundii genospecies 1) (Bowden et al., 1999) which might in part be explained by its higher mutation rate facilitating genome plasticity and an increased ability to respond successfully to environmental stress.
The methods used to interrogate data for the presence of recombinational events provided evidence for significant but limited recombination between the A. oris and A. naeslundii. This is the first evidence of recombination between these two species although recombination between sialidase genes of other oro-pharyngeal organisms, Streptococcus oralis and Streptococcus pneumoniae, has been reported (King et al., 2005). We found evidence of recombination in three of 30 A. naeslundii strains and in two of 74 A. oris isolates suggesting that recombination between these species is either not uncommon or these may be rare events and the recombinants may be more successful than the parent strains. However, the growth or survival benefit acquired by these recombinant strains is not known but because they have persisted and proliferated, at least in the mouths of the individuals from whom they were isolated, this suggests benefit has been accrued. The survival and growth benefits may not be due to recombinational events in nanH but to other unknown events, congruent with possession of a recombinant nanH gene, which might be responsible for the proliferation of these particular strains.

Supporting Information
Additional Supporting Information may be found in the online version of this article:  Fig. S1. Alignment of partial sequences of the nanH gene of the type strains of A. naeslundii (accession number EU805601), A. oris (EU805602), A. johnsonii (EU805600) and A. viscosus (EU805603) with the nanH gene sequence of S. typhimurium (M55342). Please note: Wiley-Blackwell is not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.