Peritonitis is the major disease problem of laying hens in commercial table egg and parent stock operations. Despite its importance, the etiology and pathogenesis of this disease have not been completely clarified. Although avian pathogenic Escherichia coli (APEC) isolates have been incriminated as the causative agent of laying hen peritonitis, Gallibacterium anatis are frequently isolated from peritonitis lesions. Despite recent studies suggesting a role for G. anatis in the pathogenesis of peritonitis, little is known about the organism’s virulence mechanisms, genomic composition and population dynamics. Here, we compared the genome sequences of three G. anatis isolates in an effort to understand its virulence mechanisms and identify novel antigenic traits. A multilocus sequence typing method was also established for G. anatis and used to characterize the genotypic relatedness of 71 isolates from commercial laying hens in Iowa and 18 international reference isolates. Genomic comparisons suggest that G. anatis is a highly diverse bacterial species, with some strains possessing previously described and potential virulence factors, but with a core genome containing several antigenic candidates. Multilocus sequence typing effectively distinguished 82 sequence types and several clonal complexes of G. anatis, and some clones seemed to predominate among G. anatis populations from commercial layers in Iowa. Biofilm formation and resistance to antimicrobial agents was also observed in several clades. Overall, the genomic diversity of G. anatis suggests that multiple lineages exist with differing pathogenic potential towards birds.
Citation: Johnson TJ, Danzeisen JL, Trampel D, Nolan LK, Seemann T, Bager RJ, et al. (2013) Genome Analysis and Phylogenetic Relatedness of Gallibacterium anatis Strains from Poultry. PLoS ONE 8(1): e54844. doi:10.1371/journal.pone.0054844
Editor: Sarah K. Highlander, Baylor College of Medicine, United States of America
Received: July 17, 2012; Accepted: December 18, 2012; Published: January 24, 2013
Copyright: © 2013 Johnson et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: Funding for this study was provided by the Midwest Poultry Research Program, grant 415-49-09, Iowa Egg Council, and the Danish Research Council for Technology and Production, grant 09-065909. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Avian pathogenic Escherichia coli (APEC) has typically been incriminated as the causative agent of laying hen peritonitis . However, evidence suggests that Gallibacterium anatis may also play a role in the pathogenesis of this disease . G. anatis is a Gram-negative, non-motile, encapsulated, usually β-hemolytic coccobacillus of the Pasteurellaceae family that forms grayish, round, semi-transparent colonies. The genus was defined by Christensen et al.  and now includes four named species and three additional unnamed genomospecies . Of these, only G. anatis is regularly isolated from poultry . This organism is now considered to be an important bacterial agent responsible for decreased egg production in commercial layers, since it infects the reproductive tract and causes pathological changes , –.
Although great effort has been devoted to understanding the pathogenesis of APEC, the virulence mechanisms possessed by G. anatis, and its role as a pathogen, have not been fully elucidated. The prototypic virulent G. anatis strain 12656-12 was recently sequenced. Analysis of this sequence identified a RTX-like toxin, GtxA, which contributes to G. anatis’ virulence for chickens, and has cytotoxic activity . Furthermore, the secretion system essential for export of GtxA was identified, and the gtxA gene was found to be disrupted in non-hemolytic strains of G. anatis . Other virulence associated traits among G. anatis strains have been identified, such as protease production and hemagglutination, but the underlying genetic traits responsible for these phenotypes have not yet been determined. It has also been proposed that G. anatis isolates vary in their virulence potential , , and amplified fragment length polymorphism (AFLP) has revealed that there is substantial genetic diversity among the Gallibacterium isolates dominating among and between successive flocks . The purpose of this study was to generate and compare the genome sequences of virulent and avirulent G. anatis isolates to better understand their genetic composition, and to develop a multilocus sequence typing (MLST) procedure for assessing the genetic relatedness of G. anatis isolates and their genomic content .
Materials and Methods
Bacterial Strains and Growth Conditions
All animal experiments were performed in accordance with the Insitutional Animal Care and Use Committee at Iowa State University. Live animal were humanely euthanized using carbon dioxide gas in sealed containers. The strains sequenced in this study included G. anatis strains UMN179, 12656-12, and F149T. Strain 12656-12 is a well characterized pathogenic strain isolated from the liver of a septicemic chicken in 1981 . Strain F149T is the type strain for G. anatis and was isolated from a healthy duck in Denmark in 1979 . Strain UMN179 was isolated in 2007 from a commercial laying hen with peritonitis in Iowa, USA . Additional isolates for MLST analysis were obtained from two commercial egg laying companies in Iowa, USA in 2006 and 2007 involving laying hens from four farm systems and eleven different farms (Table 1). Healthy or diseased birds were received at the Iowa State University Veterinary Diagnostic Laboratory where they were euthanized and necropsied. Swab samples were taken from the following locations: crop, gizzard, small intestine, ceca, cloaca, trachea, lung, liver, spleen, oviduct, and peritoneum. Samples were inoculated onto Remel 5% sheep blood agar and incubated aerobically at 35°C for 24–48 hours. β-hemolytic colonies were verified as Gram-negative via Gram staining, then confirmed as follows via biochemical testing: indole(−), urease(−), trehalose(+), maltose(−), xylose(+/−), arabinose(−), mannitol(+), and sorbitol(+/−) Isolates were further confirmed as G. anatis using a PCR-based approach specific for the 16S and 23S rRNA genes, as previously described . After confirmation of the isolates as G. anatis, they were stored in BHI broth with 20% w::v glycerol at −80°C until further use. Eighteen previously characterized international Gallibacterium reference strains were also used for comparative purposes in MLST analysis, representing the defined biovars and genomospecies of Gallibacterium , , .
Genome Sequencing and Annotation
Sequencing of strains UMN179, F149T, and 12656-12 was performed using 454 Life Sciences pyrosequencing at the University of Minnesota Biomedical Genomics Facility, USA, and at 454 Life Sciences, Branford, CT, USA, respectively. The sequencing of strain UMN179 has been previously described (15). For 12656-12 and F149T, one shotgun library each was used and similarly assembled with finishing. Sequences of these genomes are deposited in Genbank under the following accession numbers: UMN179 (CP002667 and CP002668), F149T (PRJNA167529), and 12656-12 (PRJNA167530).
Annotation of the three G. anatis genomes was performed using publicly available tools. Putative coding regions were predicted using GeneMarkS . Gene function was assigned using HMMER3 against Pfam-A 24.0, RPS-BLASTp against CDD, and BLASTp against all microbial proteins , . tRNA genes were identified using tRNAscan-SE . rRNA genes were identified using RNAmmer . For analysis of the shared and unique proteins in G. anatis genomes sequenced, BlastP was used with a similarity cutoff of 90% identity over 90% of the protein. For genomic regions of difference, whole genome alignments were performed using MAUVE to identify regions present in strain UMN179 but absent from strains F149T and/or 12656-12 . Genomic islands for strain UMN179 were also predicted computationally using IslandViewer . All predicted proteins of UMN179 were analyzed for their predicted subcellular location using PSORT 3.0 . Linear and circular genomic maps were generated using XPlasMap and Circos, respectively .
Using the completed sequence of G. anatis strain UMN179 and the draft sequences of strains 12656-12 and F149T, primers were designed to amplify eight housekeeping genes dispersed throughout the Gallibacterium genome (Table 2). PCR was performed using an initial denaturing step at 94°C for 5 min, followed by 25 cycles of 94°C for 30 sec, 55°C for 45 sec, and 72°C for 1 min. The amplified products were electrophoresed in 2% agarose gels at 200 V for 1 hr, stained with ethidium bromide, and visualized under ultraviolet light. Products were bidirectionally sequenced using standard Sanger sequencing reactions. Resulting data were quality assessed and then aligned and trimmed to equal sizes. Concatenated sequences were generated and aligned using ClustalW . Evolutionary distances were computed using the Maximum Composite Likelihood method. A total of 4,172 nucleotide positions were used in the final dataset. Phylogenetic analyses were conducted in MEGA version 4 . Bootstrapping was performed with 500 replicates. The sequences were deposited in Genbank under the following accession numbers: atpD, JQ647935–JQ648023; fumC, JQ648024–JQ648112; gyrB, JQ648113–JQ648201; mdh, JQ648202–JQ648290; recN, JQ648291–JQ648379; infB, JQ648380–JQ648468; thdF, JQ648469–JQ648557; and adk, JQ648558–JQ648646. An MLST database for the analysis of Gallibacterium, based upon this scheme, is available for public use at http://pubmlst.org/gallibacterium/.
Isolates were tested for susceptibility to seven antimicrobial agents by an agar disk diffusion method (Kirby-Bauer) according to the CLSI MIC Interpretative Standards using Mueller–Hinton agar plates . Disks containing the following antimicrobial agents were used: ampicillin (AM, 10 µg), gentamicin (GM, 10 µg), nalidixic acid (NA, 30 µg), streptomycin (S, 10 µg), tetracycline (TE, 30 µg), sulfisoxazole (G,.25 µg) and trimethoprim (TMP, 1.2 µg) (BD, Franklin Lakes, NJ). The following E. coli strains were used as positive controls: ATCC 25922 and APEC O1.
Biofilm formation was quantified in 96-well polystyrene microtiter plates (Falcon Microtest 353072; Becton Dickinson, Franklin Lakes, NJ, USA), based on previously described methods , . Overnight BHI broth cultures of the strain to be tested were diluted 1∶100 in BHI (Difco). A volume of 200 µl aliquots of each dilution was then dispensed into a microtiter plate well. Negative control wells contained uninoculated medium, and each strain was tested in triplicate. Plates were inoculated aerobically without shaking at 37°C for 24 h. Contents of the plates were then decanted, and plates were washed with sterile double distilled water. Microplates were then stained with 200 µl of 0.1% Crystal Violet for 30 min, washed four times with ddH2O to remove excess stain, and air dried for 1 h. After drying, adherent cells were resolubilized with 200 µl of an 80∶20 solution of ethanol and acetone. A volume of 150 µl of this solution was then transferred to a new microtiter plate, and the OD600 was measured using an automated ELx808 Ultra MicroPlate Reader (BioTek Instruments, Winooski, VT, USA). All tests were carried out in triplicate, and the results were averaged. Based on the OD produced by bacterial biofilms, strains were classified into the following categories: non, weak, moderate or strong biofilm producer, based on a previously described method , . Briefly, the cutoff OD (ODc) was defined as three standard deviations above the mean OD of the negative control. Strains were classified as follows: OD<ODc = no biofilm production; ODc<OD<(2 · ODc) = weak biofilm producer (2 · ODc)<OD<(4 · ODc) = moderate biofilm producer; and (4 · ODc)<OD = strong biofilm producer.
Results and Discussion
Overview of the G. anatis Genomes
The completed sequence of strain UMN179 was previously generated and used here for genomic comparisons (15). A total of 390,760 reads were used to draft assemble F149, resulting in 128 contigs yielding a draft assembled size of 2,403,410 bp at an average of 31-fold genome coverage. A total of 606,921 reads were used to draft assemble 12656-12, resulting in 115 contigs yielding a draft assembled size of 2,540,739 bp at an average of 26-fold genome coverage.
The complete sequence of UMN179 was compared to the draft sequence of F149T in an effort to identify genomic regions present in strain UMN179, greater then 2,000 bp, that were not present in F149T, excluding insertion sequence element insertions not involving additional genetic material (Fig. 1 and Table 3). In total, 47 genomic regions of difference were identified, 18 of which were also present in 12656-12. These regions ranged from 2.1 to 55.6 kb in size. Four of the regions contained prophage-like elements. Analysis of the predicted proteins of the three sequenced strains using a 95% amino acid similarity cutoff revealed that 69% of the predicted proteins of the three sequenced genomes were shared, 15% of the proteins of UMN179 were unique compared to F149T and 12656-12, and 8% of the proteins were shared with UMN179 and F149T or 12656-12, respectively (Fig. 2). This is roughly similar to what has been previously observed for E. coli , suggesting that the proportional pangenome size of G. anatis might be similar to that of E. coli and other Gammaproteobacteria.
The outer circle depicts scale in kilobase pairs. The next circle designates genomic regions present in UMN179 but absent from F149T. Circles 3–4 depict coding regions (blue = positive strand; green = minus strand; yellow = island genes). Circle 5 depicts tRNA and rRNA loci. Circles 6–7 depict the presence of coding regions in UMN179 compared to 12656-12 and F149, respectively, based upon amino acid similarity heatmap ranging from red (high similarity) to blue (no similarity). Circle 8 depicts G+C content relative to the genome average of 40%, with salmon spikes above 40% and light blue spikes below 40%. Circle 9 depicts single nucleotide polymorphism abundance within the core genes.
Percentages are relative to UMN179 total predicted proteins.
Strain UMN179 contained a 6,804-bp plasmid, pUMN179. The replication gene of this plasmid exhibited closest similarity to small plasmids from Pasteurella multocida strain BB1034 and Mannheimia haemolytica strain OVINE. The plasmid also contained genes encoding a TraG-like protein and a prophage-like integrase. No antibiotic resistance genes were identified on pUMN179 and no plasmid-like contigs were identified in strains 12656-12 or F149.
The G+C content of UMN179 was approximately 40%. The G+C content in UMN179 was analyzed using a 1,000 bp sliding scale, which demonstrated that numerous genomic islands contained a higher or lower G+C content than the average, further suggesting horizontal transfer of these regions (Fig. 1). Single nucleotide polymorphisms were also examined between UMN179, F149T, and 12656-12 within the core genome (Fig. 1). Regions of difference not present in all three sequenced strains were excluded from this analysis. SNPs were mostly evenly distributed throughout the core genome and coding regions with a higher density of SNPs did not exhibit any discernable patterns, such as similarity in function or subcellular location.
Whole genome alignments were performed between the completed sequence of strain UMN179 and seven other completed bacterial genomes. Inferred phylogeny was then constructed based upon SNPs within conserved regions of the genomes, and compared to inferred phylogeny using their 16S rRNA genes (Fig. 3). In both comparisons, G. anatis strain UMN179 was the most divergent genome compared to its closest available completed sequences in the database. However, this was accentuated in the whole genome alignments compared to 16S rRNA comparisons, and whole genome alignments painted a slightly different picture on the relatedness of these genomes. Specifically, whole genome alignment suggests early divergence of Gallibacterium with Haemophilus parasuis, while based on 16S rRNA analysis Gallibacterium clusters with Histophilus somni and Actinobacillus succinogenes.
Panel A: Inferred phylogeny of eight completed bacterial genomes using their 16S rRNA sequences. Data was analyzed using the Maximum Likelihood method based on the JTT matrix-based model conducted in MEGA5. There were a total of 1,515 positions in the final dataset and 100 bootstrap replicates were included. Panel B: Inferred phylogeny of eight completed bacterial genomes using whole genome alignment. Alignments were conducted in MAUVE  and SNPs within conserved regions were extracted. Data was analyzed using the Maximum Likelihood method based on the JTT matrix-based model conducted in MEGA5. There were a total of 98,659 positions in the final dataset and 100 bootstrap replicates were included.
Putative Virulence Factors
Few virulence factors of G. anatis have been described. One key virulence factor of G. anatis is its RTX-like toxin named GtxA , and its associated secretion system . All three genomes sequenced were found to contain gtxA in similar genomic locations (UMN179_1781). However, as previously described, in strain F149T this protein is disrupted by an insertion sequence, likely explaining its lack of hemolytic activity. However, it has been shown that isolates lacking hemolytic ability are not phylogenetically related . Thus, it appears that such insertional inactivations of GtxA have occurred on separate occasions resulting in convergent evolution of non-hemolytic G. anatis. The secretion system for GtxA is gtxEBD . This system is also found within the core genome of G. anatis at similar locations in all sequenced strains (UMN179_1226–1228).
Previous work has identified secreted metalloproteases in G. anatis with proteolytic activity capable of degrading chicken IgG . We identified one predicted protein (UMN179_900) that contained a metal-dependent endonuclease domain, was predicted to be extracellular, and was present in all sequenced G. anatis. Additionally UMN179_1874 was identified as a zinc metalloprotease that was present in all sequenced genomes, and UMN179_572 was identified as an ATP-dependent metalloprotease present in all sequenced genomes. These proteins could be responsible for the proteolytic capability of G. anatis.
The genome of strain UMN179 was found to contain a number of predicted adhesins. Three products, UMN179_416, UMN179_2244, and UMN179_2445, were similar to one another and possessed hemagglutinin activity and pre-toxin domains. Each of these also appears to be a filamentous hemagglutinin-like protein linked to a two-partner secretion system, with adjacent proteins containing domains related to secretion and activation of hemolysins/hemagglutinins. These three predicted proteins were unique to UMN179, as compared to 12656-12 and F149T. Strain UMN179 also contained six predicted hemagglutinin/hemolysin/calcium-binding-like proteins of varying lengths and similarities with one another. All were unique and shared low similarity with the other predicted proteins in UMN179, but all were more similar to one another than to other predicted proteins in the NCBI database. Interestingly, many of these genes were adjacent to truncated versions of predicted hemagglutinins. In strains 12656-12 and F149T, similar predicted proteins were identified in some cases. For example, UMN179_1030 and UMN179_1346 shared 45% and 80% identity, respectively, with predicted proteins from 12656-12 in the same genomic locations. UMN179 also possessed several additional predicted proteins with YadA domains, one putative autotransporter/adhesin (UMN179_1565), and one putative serine protease (UMN179_1925).
Strain UMN179 contained three predicted chaperone/usher fimbrial loci. All were four-gene F17-like systems encoding apparently intact chaperone and usher proteins. Two fimbrial systems (UMN179_293–295 and UMN179_809–812) were present in all three sequenced genomes, whereas an intact version of the remaining system (UMN179_750–753) could only be identified in the genomes of UMN179 and 12656-12. A truncated version could, though, be identified in F149T. This system lacked the genes encoding the usher and adhesin protein, and moreover, the chaperone-encoding gene was only partly represented. We compared the fimbrial ushers for each of these systems with existing fimbrial systems from the NCBI database (Fig. 4). From amino acid alignments of these proteins, each protein was most similar to those from the same locus (in different isolates) and same species (in different loci), but all predicted proteins from G. anatis were most similar to one another. This is suggestive that the F17-like fimbrial systems in G. anatis have arisen through duplication of these systems resulting in multiple paralogs following speciation. Interestingly, a transposon was located just upstream the truncated version of the fimbrial locus in F149T. A similar transposon was also found in the region upstream the locus-homologs in strain UMN179 and 12656-12, respectively, further suggesting horizontal gene transfer of paralogs within G. anatis.
Capsular Polysaccharide Components
The genomes of UMN179 and F149T did not contain detectable capsular polysaccharide biosynthesis genes. However, 12656-12 contained a putative capsular biosynthesis gene cluster inserted within genes similar to UMN179_44 and UMN179_45. This gene cluster contained 13 coding regions designated lipA2-lipA1, gcbG-A, and gexD-A. Capsular components of G. anatis have not yet been described, but some G. anatis do possess a capsular structure (A.M. Bojesen, unpublished results).
Clustered regularly interspaced short palindromic repeats (CRISPRs) are arrays or repeats common among bacteria and archaeae that provide protection against foreign DNA . These loci are identified by their possession of cas gene clusters, and the presence of non-contiguous direct repeats separated by variable spacer regions. Several CRISPR loci were identified within the three G. anatis genomes sequenced . The first CRISPR locus was found between UMN179_2281 and UMN179_2300 and was present in UMN179 and 12656-12 on GI44 (Table 3). This region contained the genes csm1-csm2-csm3-csm4-csm5-cas6-csx16-csm6-csx1-cas2-cas1-cas2, with 31 direct repeats in UMN179 and 15 direct repeats in 12656-12. A second CRISPR locus was identified between UMN179_2337 and UMN179_2343 and was present in all three genomes. This locus contained cas3-cys4-cys3-cys2-cys1 and had 8 direct repeats in UMN179, and 7 direct repeats each in F149T and 12656-12. A third locus was present in strains F149T and 12656-12, but absent in UMN179, situated between UMN179_173 and UMN179_174. This region contained cas3-cas5-csd1-csd2-cas4-cas1-cas2 and had 6 direct repeats in 12656-12 and 48 direct repeats in F149T. None of the spacers within any of the three strains shared significant nucleotide similarity with spacers of other strains. Similarly, none of the three CRISPR loci genes shared significant amino acid similarity with genes from other loci. The presence of all CRISPR loci in multiple strains of the three sequenced genomes suggest that they were acquired by ancestral strains and have rapidly evolved to acquire different subsets of spacer elements. Since none of the strains sequenced here are very closely phylogenetically related (see below), it is not surprising that they differ in their spacer compositions. It remains to be determined if CRISPR loci in Gallibacterium can be used as phylogenetic tools and as a means of predicting the propensity to acquire foreign DNA.
Other loci of importance for DNA uptake were encoded in the competence regulon, which was present in all three strains, as previously described . Interestingly, F149T exhibited a transformation frequency that was three orders of magnitude lower than 12656-12 . Thus, although most G. anatis strains studied appear to be naturally competent, this ability seems differentially regulated between strains .
Integrative Conjugative Elements (ICE) and Resistance Elements of Gallibacterium
Integrative conjugative elements (ICEs) are well-described entities that are able to integrate and excise from bacterial genomes . ICEs have some trademark signatures, including a site-specific integrase, transfer genes, and genes regulating excision and transfer. ICEs also often possess accessory elements conferring antimicrobial resistance. All three sequenced G. anatis strains contained single but distinct ICEs within their genomes. These ICE was integrated within the tRNALeu in strains UMN179 (UMN179_316 through UMN179_382), similar to ICEPmu1 found within the genome of Pasteurella multocida strain 36950 . In strain 12656-12, its ICE element was integrated into tRNAMet. We were unable to discern the genomic location of the ICE of strain F149. These ICEs did not appear to harbor any antimicrobial resistance genes but did contain YadA domain-containing coding regions with similarity to a hemagglutinin/invasin (Fig. S1). The presence of distinct ICEs in multiple locales within the three G. anatis genomes suggests that these may play an important and active role in their evolution. Based upon phylogenetic analysis of the topB gene of the ICEs in G. anatis, it seems that the ICEs of G. anatis are more closely related to one another than those of closely related species (Fig. 5).
Data was analyzed using the Maximum Likelihood method based on the JTT matrix-based model conducted in MEGA5. There were a total of 596 positions in the final dataset and 500 bootstrap replicates were included.
It was previously reported that the tet(31) resistance determinant is common among G. anatis . In strain UMN179, this region is absent but another antimicrobial resistance island is present. This island is flanked by IS10 transposase elements and encodes an AraC family transcriptional regulator annotated in some cases as TetD, a TetR family transcriptional regulator annotated in some cases as TetC, a major facilitator superfamily protein annotated in some cases as TetA (56% amino acid similarity to TetA(31)), a second TetR transcriptional regulator (64% amino acid similarity to TetR(31)), an ArsR family transcriptional regulator, an amino acid binding ACT domain protein, and a truncated sodium/glutamate symporter. Since strain UMN179 is resistant to tetracycline, it is likely that this region plays a role in tetracycline resistance and represents a mechanism in addition to tet(31) by which G. anatis is able to resist tetracycline.
Prediction of Antigenic Candidates of G. anatis
Predicted subcellular locations of the proteins of UMN179 were assessed using PSORTb version 3.0 in an effort to determine possible antigenic candidates that are surface exposed proteins (Table 4). In total 54 (2.2%) outer membrane proteins and 21 (0.8%) extracellular proteins were predicted. Predicted outer membrane proteins included putative autotransporter/adhesins, fimbrial usher proteins, hemolysin secretion proteins, the secretion system for the RTX-like toxin GtxA, iron regulated proteins, and other predicted outer membrane proteins of unknown function (Tables S1Tables S1 and S2). Many of these proteins were conserved among the G. anatis strains sequenced, suggesting their potential as future antigenic candidates for vaccine production.
MLST of G. anatis
In an effort to better appreciate the phylogenetic diversity of G. anatis, 71 isolates belonging to four farm systems and collected at ten timepoints during 2006–2007 were examined for their genetic relatedness using a MLST scheme developed here. Some isolates came from differing body sites within the same bird, while others came from different birds within the same flock (Table 1). Eighteen reference strains belonging to various subgroups of G. anatis were also included for comparison purposes, including the isolates sequenced here. Assignment of sequence type (ST) was performed and used to develop an MLST database (http://pubmlst.org/gallibacterium/). In total, 82 STs were identified from the 89 isolates examined, underscoring the genetic diversity of the isolates examined. The dendrogram constructed based upon concatenated gene sequences revealed significant diversity among all of the isolates examined, with the greatest similarities between isolates belonging to the same farm systems and/or timepoints examined (Figs. 6 and 7). The location of isolation within the bird did not correlate with the dendrogram, as isolates from multiple bird sites were scattered throughout the dendrogram. Isolates were grouped by farm source. Group ‘A’ isolates (n = 14) were found mostly in one predominant lineage (n = 7) with the remainder distributed throughout the dendrogram. Group ‘B’ isolates (n = 9) were mostly found within two closely related lineages (n = 8). Similarly, group ‘C’ isolates (n = 9) also clustered into two distinct and unrelated groups. Nearly all group ‘D’ isolates (n = 20) were found in several closely related clusters that were distinct from the rest of the isolates examined. In contrast, group ‘E’ isolates (n = 13) were more diverse and spread throughout the dendrogram. Notably, most reference isolates examined clustered closely, while additional clinical isolates examined expanded the diversity of the dendrogram considerably.
Data was analyzed using the Maximum Likelihood method based on the Hasegawa-Kishino-Yano model in MEGA5. A total of 4,172 positions were used in the final dataset and 100 bootstrap replicates were included.
Farms are colored by source. Biofilm data are classified by strong (S), intermediate (I), or weak (W) biofilm formers. Antimicrobial susceptibilities are depicted as resistant (R), intermediate (I), or susceptible (S) using disk diffusion. Abbreviations for antibiotics used are S10 = streptomycin, NA30 = nalidixic acid, TE30 = tetracycline, GM10 = gentamicin, TMP5 = trimethoprim, G25 = sulfisoxazole, C30 = chloramphenicol, AM10 = ampicillin, RA5 = rifampin, and K30 = kanamycin. The dendrogram was constructed as described in Fig. 6 and is identical order to Fig. 6.
Previously, several approaches have been used to examine the genetic diversity of Gallibacterium isolates. Bojesen et al used AFLP to examine fifty Gallibacterium isolates belonging to 29 flocks in California, USA . They demonstrated substantial genetic diversity within the Gallibacterium genus, differentiation of isolates belonging to Gallibacterium genomospecies 1 and 2 from formal G. anatis isolates, and the identification of dominant clones that persisted in the same operation throughout a 6-year period. Other typing approaches have been used to characterize Gallibacterium, including DNA:DNA hybridization,16S rRNA analysis, and MLST, that also clearly separate G. anatis from Gallibacterium genomospecies 1 and 2 , , . However, correlations between phenotypic biovar and phylogeny within G. anatis are less clear. We can conclude from the development of the current MLST approach that the previously identified G. anatis biovars are representative of much of the genetic diversity within G. anatis, but diversity within this genus is even greater than that of currently defined biovars. Also, we conclude that dominant clones of G. anatis exist in some farm operations, while in other farms the populations are more diverse. The publicly available MLST database for Gallibacterium will provide a worldwide opportunity to further study such strain diversity.
Isolates assessed by MLST were also compared for their ability to form biofilms on a plastic substrate and their susceptibility to ten antimicrobials using disk diffusion (Fig. 7). Biofilm formation was classified as ‘strong,’ ‘intermediate,’ or ‘weak’ based upon a previously used crystal violet assay . Patterns were observed when relating biofilm formation to inferred phylogeny. For example, several clusters that contained previously characterized reference strains were classified as strong or intermediate biofilm formers, as were most isolates belonging to the cluster of isolates containing UMN179. Strains F149T, CCM5974 and 12158/5 salp. have previously been shown to be strong at biofilm formation, which to some degree validates the current results . Many isolates from the Iowa collection, though, were weak biofilm formers that belonged outside of the aforementioned clusters. We are unable at this point to determine if certain genomic regions are associated with a stronger ability to form biofilm, however there does appear to be a phylogenetic correlation with strong versus weak biofilm formers.
Most isolates examined were resistant to tetracycline, sulfisoxazole, and rifampin, while some were also resistant to kanamycin, and these isolates tended to cluster together on the dendrogram. The genetic basis for tetracycline resistance has been described above, but mechanisms of resistance to these other antimicrobials are currently unknown. The presence of entire lineages resistant to certain antimicrobials suggests that some of these traits are well established in such lineages and are likely encoded within their core chromosomes rather than within accessory genomic islands or transmissible plasmids. The presence of resistance traits on islands was previously suggested by Bojesen et al.,  who showed that distinct resistance phenotypes could be identified in genotypically different isolates from unrelated flocks. Therefore, horizontal gene transfer likely plays an additional role in the acquisition of resistance to antimicrobial agents by Gallibacterium.
In conclusion, this study defined the initial core and accessory genomes of G. anatis and identified a number of loci with the potential to play a role in the pathogenesis of this bacterium. Substantial genomic differences among the strains examined suggest that G. anatis, a commensal organism in avian species, comprise a diverse group of isolates that differ in their pathogenic potential. Future work is necessary to fully understand the role of the genomic repertoire of G. anatis relative to its disease-causing ability.
Linear map of the integrative conjugative element of Gallibacterium anatis strain UMN179.
Predicted proteins localized to the outer membrane of Gallibacterium anatis strain UMN179.
Predicted proteins localized as extracellular in Gallibacterium anatis UMN179.
The authors are grateful to Dr. Keith Jolley (University of Oxford) for assistance in the development of the Gallibacterium MLST database. This work was carried out in part using computing resources at the University of Minnesota Supercomputing Institute.
Conceived and designed the experiments: TJJ DT LKN AMB. Performed the experiments: TJJ JLD DT. Analyzed the data: TJJ TS RJB AMB. Contributed reagents/materials/analysis tools: TJJ AMB. Wrote the paper: TJJ AMB.
- 1. Saif YM, Fadly AM (2008) Diseases of Poultry. Ames, Iowa: Blackwell Pub. 1324 p.
- 2. Pattison M (2008) Poultry Diseases. Edinburgh; New York: Elsevier/Butterworth-Heinemann. 611 p.
- 3. Christensen H, Bisgaard M, Bojesen AM, Mutters R, Olsen JE (2003) Genetic relationships among avian isolates classified as Pasteurella haemolytica, ‘Actinobacillus salpingitidis’ or Pasteurella anatis with proposal of Gallibacterium anatis gen. nov., comb. nov. and description of additional genomospecies within Gallibacterium gen. nov. Int J Syst Evol Microbiol 53: 275–287.
- 4. Bisgaard M, Korczak BM, Busse HJ, Kuhnert P, Bojesen AM, et al. (2009) Classification of the taxon 2 and taxon 3 complex of Bisgaard within Gallibacterium and description of Gallibacterium melopsittaci sp. nov., Gallibacterium trehalosifermentans sp. nov. and Gallibacterium salpingitidis sp. nov. Int J Syst Evol Microbiol 59: 735–744.
- 5. Bojesen AM, Torpdahl M, Christensen H, Olsen JE, Bisgaard M (2003) Genetic diversity of Gallibacterium anatis isolates from different chicken flocks. J Clin Microbiol 41: 2737–2740.
- 6. Lin MY, Lin KJ, Lan YC, Liaw MF, Tung MC (2001) Pathogenicity and drug susceptibility of the Pasteurella anatis isolated in chickens in Taiwan. Avian Dis 45: 655–658.
- 7. Bojesen AM, Nielsen OL, Christensen JP, Bisgaard M (2004) In vivo studies of Gallibacterium anatis infection in chickens. Avian Pathol 33: 145–152.
- 8. Neubauer C, De Souza-Pilz M, Bojesen AM, Bisgaard M, Hess M (2009) Tissue distribution of haemolytic Gallibacterium anatis isolates in laying birds with reproductive disorders. Avian Pathol 38: 1–7.
- 9. Kristensen BM, Frees D, Bojesen AM (2010) GtxA from Gallibacterium anatis, a cytolytic RTX-toxin with a novel domain organisation. Vet Res 41: 25.
- 10. Kristensen BM, Frees D, Bojesen AM (2011) Expression and secretion of the RTX-toxin GtxA among members of the genus Gallibacterium. Vet Microbiol.
- 11. Zepeda VA, Calderon-Apodaca NL, Paasch ML, Martin PG, Paredes DA, et al. (2010) Histopathologic findings in chickens experimentally infected with Gallibacterium anatis by nasal instillation. Avian Dis 54: 1306–1309.
- 12. Zepeda A, Ramirez S, Vega V, Morales V, Talavera M, et al. (2009) Hemagglutinating activity of Gallibacterium strains. Avian Dis 53: 115–118.
- 13. Bojesen AM, Shivaprasad HL (2007) Genetic diversity of Gallibacterium isolates from California turkeys. Avian Pathol 36: 227–230.
- 14. Johnson TJ, Fernandez-Alarcon C, Bojesen AM, Nolan LK, Trampel DW, et al. (2011) Complete genome sequence of Gallibacterium anatis strain UMN179, isolated from a laying hen with peritonitis. J Bacteriol 193: 3676–3677.
- 15. Bojesen AM, Vazquez ME, Robles F, Gonzalez C, Soriano EV, et al. (2007) Specific identification of Gallibacterium by a PCR using primers targeting the 16S rRNA and 23S rRNA genes. Vet Microbiol 123: 262–268.
- 16. Besemer J, Lomsadze A, Borodovsky M (2001) GeneMarkS: a self-training method for prediction of gene starts in microbial genomes. Implications for finding sequence motifs in regulatory regions. Nucleic Acids Res 29: 2607–2618.
- 17. McClure MA, Smith C, Elton P (1996) Parameterization studies for the SAM and HMMER methods of hidden Markov model generation. Proc Int Conf Intell Syst Mol Biol 4: 155–164.
- 18. Altschul SF, Madden TL, Schaffer AA, Zhang J, Zhang Z, et al. (1997) Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res 25: 3389–3402.
- 19. Lowe TM, Eddy SR (1997) tRNAscan-SE: a program for improved detection of transfer RNA genes in genomic sequence. Nucleic Acids Res 25: 955–964.
- 20. Lagesen K, Hallin P, Rodland EA, Staerfeldt HH, Rognes T, et al. (2007) RNAmmer: consistent and rapid annotation of ribosomal RNA genes. Nucleic Acids Res 35: 3100–3108.
- 21. Darling AC, Mau B, Blattner FR, Perna NT (2004) Mauve: multiple alignment of conserved genomic sequence with rearrangements. Genome Res 14: 1394–1403.
- 22. Langille MG, Brinkman FS (2009) IslandViewer: an integrated interface for computational identification and visualization of genomic islands. Bioinformatics 25: 664–665.
- 23. Nakai K, Horton P (1999) PSORT: a program for detecting sorting signals in proteins and predicting their subcellular localization. Trends Biochem Sci 24: 34–36.
- 24. Darzentas N (2010) Circoletto: visualizing sequence similarity with Circos. Bioinformatics 26: 2620–2621.
- 25. Thompson JD, Gibson TJ, Higgins DG (2002) Multiple sequence alignment using ClustalW and ClustalX. Curr Protoc Bioinformatics Chapter 2: Unit 2 3.
- 26. Tamura K, Dudley J, Nei M, Kumar S (2007) MEGA4: Molecular Evolutionary Genetics Analysis (MEGA) software version 4.0. Molecular Biol Evol 24: 1596–1599.
- 27. Clinical and Laboratory Standards Institute (2005) Performance standards for antimicrobial susceptibility testing.
- 28. O’Toole GA, Pratt LA, Watnick PI, Newman DK, Weaver VB, et al. (1999) Genetic approaches to study of biofilms. Methods Enzymol 310: 91–109.
- 29. Stepanovic S, Cirkovic I, Ranin L, Svabic-Vlahovic M (2004) Biofilm formation by Salmonella spp. and Listeria monocytogenes on plastic surface. Lett Appl Microbiol 38: 428–432.
- 30. Skyberg JA, Siek KE, Doetkott C, Nolan LK (2007) Biofilm formation by avian Escherichia coli in relation to media, source and phylogeny. J Appl Microbiol 102: 548–554.
- 31. Welch RA, Burland V, Plunkett G 3rd, Redford P, Roesch P, et al (2002) Extensive mosaic structure revealed by the complete genome sequence of uropathogenic Escherichia coli. Proc Natl Acad Sci USA 99: 17020–17024.
- 32. Kristensen BM, Frees D, Bojesen AM (2011) Expression and secretion of the RTX-toxin GtxA among members of the genus Gallibacterium. Vet Microbiol 153: 116–123.
- 33. Garcia-Gomez E, Vaca S, Perez-Mendez A, Ibarra-Caballero J, Perez-Marquez V, et al. (2005) Gallibacterium anatis-secreted metalloproteases degrade chicken IgG. Avian Pathol 34: 426–429.
- 34. Fricke WF, Mammel MK, McDermott PF, Tartera C, White DG, et al. (2011) Comparative genomics of 28 Salmonella enterica isolates: evidence for CRISPR-mediated adaptive sublineage evolution. J Bacteriol 193: 3556–3568.
- 35. Grissa I, Vergnaud G, Pourcel C (2007) CRISPRFinder: a web tool to identify clustered regularly interspaced short palindromic repeats. Nucleic Acids Res 35: W52–57.
- 36. Kristensen BM, Sinha S, Boyce JD, Bojesen AM, Mell JC, et al. (2012) Natural transformation of Gallibacterium anatis. Appl Environ Microbiol 78: 4914–4922.
- 37. Wozniak RA, Fouts DE, Spagnoletti M, Colombo MM, Ceccarelli D, et al. (2009) Comparative ICE genomics: insights into the evolution of the SXT/R391 family of ICEs. PLoS Genet 5: e1000786.
- 38. Michael GB, Kadlec K, Sweeney MT, Brzuszkiewicz E, Liesegang H, et al. (2012) ICEPmu1, an integrative conjugative element (ICE) of Pasteurella multocida: structure and transfer. J Antimicrob Chemother 67: 91–100.
- 39. Bojesen AM, Bager RJ, Ifrah D, Aarestrup FM (2011) The rarely reported tet(31) tetracycline resistance determinant is common in Gallibacterium anatis. Vet Microbiol 149: 497–499.
- 40. Christensen H, Kuhnert P, Olsen JE, Bisgaard M (2004) Comparative phylogenies of the housekeeping genes atpD, infB and rpoB and the 16S rRNA gene within the Pasteurellaceae. Int J Syst Evol Microbiol 54: 1601–1609.
- 41. Vaca S, Monroy E, Rojas L, Vazquez C, Sanchez P, et al. (2011) Adherence of Gallibacterium anatis to inert surfaces. J Anim Vet Adv 10: 1688–1693.
- 42. Bojesen AM, Vazquez ME, Bager RJ, Ifrah D, Gonzalez C, et al. (2011) Antimicrobial susceptibility and tetracycline resistance determinant genotyping of Gallibacterium anatis. Vet Microbiol 148: 105–110.