Advertisement
Research Article

The Pan-Genome of the Animal Pathogen Corynebacterium pseudotuberculosis Reveals Differences in Genome Plasticity between the Biovar ovis and equi Strains

  • Siomar C. Soares,

    Affiliations: Department of General Biology, Federal University of Minas Gerais, Belo Horizonte, Minas Gerais, Brazil, Center for Biotechnology, Bielefeld University, Bielefeld, Nordrhein-Westfalen, Germany, CLIB Graduate Cluster Industrial Biotechnology, Center for Biotechnology, Bielefeld University, Bielefeld, Nordrhein-Westfalen, Germany

    X
  • Artur Silva,

    Affiliation: Department of Genetics, Federal University of Pará, Belém, Pará, Brazil

    X
  • Eva Trost,

    Affiliations: Center for Biotechnology, Bielefeld University, Bielefeld, Nordrhein-Westfalen, Germany, CLIB Graduate Cluster Industrial Biotechnology, Center for Biotechnology, Bielefeld University, Bielefeld, Nordrhein-Westfalen, Germany

    X
  • Jochen Blom,

    Affiliation: Center for Biotechnology, Bielefeld University, Bielefeld, Nordrhein-Westfalen, Germany

    X
  • Rommel Ramos,

    Affiliation: Department of Genetics, Federal University of Pará, Belém, Pará, Brazil

    X
  • Adriana Carneiro,

    Affiliation: Department of Genetics, Federal University of Pará, Belém, Pará, Brazil

    X
  • Amjad Ali,

    Affiliation: Department of General Biology, Federal University of Minas Gerais, Belo Horizonte, Minas Gerais, Brazil

    X
  • Anderson R. Santos,

    Affiliation: Department of General Biology, Federal University of Minas Gerais, Belo Horizonte, Minas Gerais, Brazil

    X
  • Anne C. Pinto,

    Affiliation: Department of General Biology, Federal University of Minas Gerais, Belo Horizonte, Minas Gerais, Brazil

    X
  • Carlos Diniz,

    Affiliation: Department of General Biology, Federal University of Minas Gerais, Belo Horizonte, Minas Gerais, Brazil

    X
  • Eudes G. V. Barbosa,

    Affiliation: Department of General Biology, Federal University of Minas Gerais, Belo Horizonte, Minas Gerais, Brazil

    X
  • Fernanda A. Dorella,

    Affiliation: Department of General Biology, Federal University of Minas Gerais, Belo Horizonte, Minas Gerais, Brazil

    X
  • Flávia Aburjaile,

    Affiliation: Department of General Biology, Federal University of Minas Gerais, Belo Horizonte, Minas Gerais, Brazil

    X
  • Flávia S. Rocha,

    Affiliation: Department of General Biology, Federal University of Minas Gerais, Belo Horizonte, Minas Gerais, Brazil

    X
  • Karina K. F. Nascimento,

    Affiliation: Department of General Biology, Federal University of Minas Gerais, Belo Horizonte, Minas Gerais, Brazil

    X
  • Luís C. Guimarães,

    Affiliations: Department of General Biology, Federal University of Minas Gerais, Belo Horizonte, Minas Gerais, Brazil, Center for Biotechnology, Bielefeld University, Bielefeld, Nordrhein-Westfalen, Germany, CLIB Graduate Cluster Industrial Biotechnology, Center for Biotechnology, Bielefeld University, Bielefeld, Nordrhein-Westfalen, Germany

    X
  • Sintia Almeida,

    Affiliation: Department of General Biology, Federal University of Minas Gerais, Belo Horizonte, Minas Gerais, Brazil

    X
  • Syed S. Hassan,

    Affiliation: Department of General Biology, Federal University of Minas Gerais, Belo Horizonte, Minas Gerais, Brazil

    X
  • Syeda M. Bakhtiar,

    Affiliation: Department of General Biology, Federal University of Minas Gerais, Belo Horizonte, Minas Gerais, Brazil

    X
  • Ulisses P. Pereira,

    Affiliation: Department of Veterinary Medicine, Federal University of Lavras, Lavras, Brazil

    X
  • Vinicius A. C. Abreu,

    Affiliation: Department of General Biology, Federal University of Minas Gerais, Belo Horizonte, Minas Gerais, Brazil

    X
  • Maria P. C. Schneider,

    Affiliation: Department of Genetics, Federal University of Pará, Belém, Pará, Brazil

    X
  • Anderson Miyoshi,

    Affiliation: Department of General Biology, Federal University of Minas Gerais, Belo Horizonte, Minas Gerais, Brazil

    X
  • Andreas Tauch equal contributor,

    equal contributor Contributed equally to this work with: Andreas Tauch, Vasco Azevedo

    Affiliation: Center for Biotechnology, Bielefeld University, Bielefeld, Nordrhein-Westfalen, Germany

    X
  • Vasco Azevedo equal contributor mail

    equal contributor Contributed equally to this work with: Andreas Tauch, Vasco Azevedo

    vasco@icb.ufmg.br

    Affiliation: Department of General Biology, Federal University of Minas Gerais, Belo Horizonte, Minas Gerais, Brazil

    X
  • Published: January 14, 2013
  • DOI: 10.1371/journal.pone.0053818

Abstract

Corynebacterium pseudotuberculosis is a facultative intracellular pathogen and the causative agent of several infectious and contagious chronic diseases, including caseous lymphadenitis, ulcerative lymphangitis, mastitis, and edematous skin disease, in a broad spectrum of hosts. In addition, Corynebacterium pseudotuberculosis infections pose a rising worldwide economic problem in ruminants. The complete genome sequences of 15 C. pseudotuberculosis strains isolated from different hosts and countries were comparatively analyzed using a pan-genomic strategy. Phylogenomic, pan-genomic, core genomic, and singleton analyses revealed close relationships among pathogenic corynebacteria, the clonal-like behavior of C. pseudotuberculosis and slow increases in the sizes of pan-genomes. According to extrapolations based on the pan-genomes, core genomes and singletons, the C. pseudotuberculosis biovar ovis shows a more clonal-like behavior than the C. pseudotuberculosis biovar equi. Most of the variable genes of the biovar ovis strains were acquired in a block through horizontal gene transfer and are highly conserved, whereas the biovar equi strains contain great variability, both intra- and inter-biovar, in the 16 detected pathogenicity islands (PAIs). With respect to the gene content of the PAIs, the most interesting finding is the high similarity of the pilus genes in the biovar ovis strains compared with the great variability of these genes in the biovar equi strains. Concluding, the polymerization of complete pilus structures in biovar ovis could be responsible for a remarkable ability of these strains to spread throughout host tissues and penetrate cells to live intracellularly, in contrast with the biovar equi, which rarely attacks visceral organs. Intracellularly, the biovar ovis strains are expected to have less contact with other organisms than the biovar equi strains, thereby explaining the significant clonal-like behavior of the biovar ovis strains.

Introduction

The genus Corynebacterium belongs to the CMNR group from the supra-generic group of Actinomycetes, which includes genera of great medical, veterinary, and biotechnological importance, such as Corynebacterium, Mycobacterium, Nocardia, and Rhodococcus. These genera have specific features in common, such as a high DNA G+C content and a specific organization of the cell wall, which is mainly composed of peptidoglycans, arabinogalactans, and mycolic acids [1]. The genus Corynebacterium was originally created to include Corynebacterium diphtheriae and other pathogenic species [2]. Several other bacteria that differed in shape, pathogenicity and sporulation were later added to this group [3]. Currently, the genus is composed of pathogenic species such as Corynebacterium diphtheriae, the causative agent of diphtheria [4]; opportunistic pathogens such as Corynebacterium jeikeium, which is responsible for some nosocomial infections in humans [5]; and non-pathogenic species such as Corynebacterium glutamicum, which is highly utilized in industrial amino acid production [6].

Corynebacterium pseudotuberculosis is a facultative intracellular and pleomorphic member of the genus Corynebacterium. This bacterium is non-motile, although it does possess fimbriae, and it is the causative agent of caseous lymphadenitis (CLA) in sheep and goats [7]. A close taxonomic relationship between C. pseudotuberculosis and Corynebacterium ulcerans has been suggested because these organisms are the only corynebacteria that produce the exotoxin phospholipase D [8], [9]. Moreover, some strains of C. pseudotuberculosis and C. ulcerans express the diphtheria toxin, which indicates a relationship between both species and C. diphtheriae [10]. This relationship has also been demonstrated by a phylogenetic analysis of the rpoB gene [1]. The initial classification of C. pseudotuberculosis was based on morphological and biochemical characteristics [7], [11]: the results of the nitrate reduction test play an important role in distinguishing the biovar ovis (isolated from sheep and goats; negative nitrate reduction) from the biovar equi (isolated from horses and bovines; positive nitrate reduction) [12].

In sheep and goats, C. pseudotuberculosis biovar ovis strains are responsible for causing the aforementioned infectious, contagious, chronic disease CLA, which is mainly characterized by the presence of caseous necrosis on the lymphatic glands or abscess formation in superficial lymph nodes and subcutaneous tissues [13]. CLA is a widespread disease that has been reported in several countries, including Australia, Brazil, Canada, New Zealand, South Africa, and the United States, where sheep and goat farming are prevalent [1], [14][18]. CLA produces economic losses for sheep and goat farmers by causing skin deterioration and reducing yields of milk and wool. In addition to these effects, the visceral form of the disease can affect internal organs, resulting in weight loss, carcass condemnation and death [19]. The disease is transmitted through direct contact with superficial wounds, which can be the result of common procedures such as castration and shearing [20]. The transmission and dissemination of C. pseudotuberculosis are also associated with the following: a high resistance to environmental conditions [21][23]; a low detection rate, with the visceral form of the disease usually being detected in the later stages or in the slaughterhouse [24]; the inefficacy of antibiotic therapies due to abscess formation and an intra-macrophagic lifestyle [25]; high variability in the severity of the disease in vaccinated animals and in the protection levels of the vaccines [26]; and the variable efficacy of licensed vaccines, which are intended for use in sheep, in goat immunizations [27].

Although C. pseudotuberculosis was initially identified as causing CLA in sheep and goats, this bacterium has also been isolated from other species that exhibit different symptoms, including horses, cows, camels, buffalo, and even humans [1], [28][30]. Despite the broad host spectrum, natural cross-species transmission of C. pseudotuberculosis between small ruminants and cattle does not appear to occur [12], although infections of cattle with both biovars have been previously reported [31].

C. pseudotuberculosis infections in horses can display three different disease patterns: external abscesses (pigeon fever), ulcerative lymphangitis of the limbs, and a visceral form that affects the internal organs [32], [33]. Additionally, several clinical symptons of the diseases caused by C. pseudotuberculosis have been described in cattle: pyogranulomatous reactions, abscess formation, mastitis, visceral commitment, and necrotic and ulcerative dermatitis on the heel of the foot, which is accompanied by edematous swelling and lameness [24]. In bulls and buffalo, there is evidence of the mechanical transmission of C. pseudotuberculosis by houseflies or other diptera, in addition to transmission via skin contact between animals [23], [24], [34][37]. Moreover, all reported outbreaks of CLA in horses in the United States have been preceded by large populations of houseflies and other diptera during the summer, a phenomenon promoted by high environmental temperatures and drought conditions [38] that may also be related to a rise in the number of affected herds in Israel [24].

Although the pathogenic mechanism of CLA is well understood, there remains a lack of information about the virulence factors of C. pseudotuberculosis and the pathogenic mechanisms of the other diseases caused by this bacterium [1], [39], [40]. Virulence factors play an important role in the adhesion, invasion, colonization, spread inside the host, and immune system evasion of pathogenic bacteria; they also allow contact, penetration and survival inside the host [41]. Billington et al. [42] reported four C. pseudotuberculosis genetic factors, the fagABC operon and the fagD gene, that play an important role in virulence; they are involved in iron acquirement and, therefore, enable the bacterium to survive in environments where iron is scarce. The fagABC operon and the fagD gene are found in a pathogenicity island along with the pld gene, which encodes phospholipase D (PLD) [43]. PLD is the primary virulence factor of C. pseudotuberculosis; it promotes the hydrolysis and degradation of sphingomyelin in endothelial cell membranes, which increases vascular permeability and contributes to the spread and persistence of the bacterium in the host [27], [44], [45]. More recently, Trost et al. [46] reported the presence of two pilus gene clusters in the C. pseudotuberculosis FRC41 strain, which is in agreement with the previously reported visualization of pilin structures in other strains of C. pseudotuberculosis [47]. Pili are helical, cylinder-shaped structures, which are observed attached to and protruding from the bacterial cell surface. Pili play an important role in virulence as they enable pathogens to bind to molecules on various host tissues. After attaching to the host cell surface, the pathogen is able to initiate specific biochemical processes, such as extracellular and intracellular invasion, that will result in its proliferation in and dissemination among the host tissues [48].

To better understand the different symptoms of C. pseudotuberculosis infections in the broad spectrum of hosts and how genome plasticity is related to the symptom patterns, we performed pan-genomic comparative analyses of 15 C. pseudotuberculosis strains. In the following sections, we present the phylogenomic correlations between C. pseudotuberculosis and other corynebacteria. Furthermore, we describe the content and extrapolations of the following gene subsets from C. pseudotuberculosis: the “pan-genome”, which is the complete inventory of genes found in any member of the species; the “core genome”, which is composed of the genes that are present in all the species strains and that are thus important for basic life processes; and the “singletons”, which represent genes found only in a given strain. Finally, we provide insights into the specific subsets (singletons and the pan- and core genomes) of both biovars of C. pseudotuberculosis, ovis and equi, and we correlate these subsets with the plasticity of pathogenicity islands, virulence genes, and biovar-specific diseases.

Materials and Methods

Genome Sequences

The genome sequences of 15 C. pseudotuberculosis strains were retrieved from the NCBI database (http://www.ncbi.nlm.nih.gov/genbank/): 9 biovar ovis strains, which were isolated from sheep, goats, humans, llamas, antelopes, and cows, and 6 biovar equi strains, which were isolated from horses, camels, and buffalo (Table 1). The strains were isolated in Oceania (Australia), South America (Brazil and Argentina), North America (United States), Africa (South Africa, Egypt and Kenya), southwestern Asia (Israel), and Europe (the United Kingdom, Belgium, France and Scotland). The clinical symptoms of infections with these strains vary broadly and include abscesses, mastitis, lymphangitis, necrogranuloma, and edematous skin disease (Table 1).

thumbnail

Table 1. General information about the 15 C. pseudotuberculosis strains used in this work.

doi:10.1371/journal.pone.0053818.t001

Corynebacterium Genus Phylogenomic Analyses

The Gegenees (version 1.1.4) software was used to perform the phylogenomic analyses at the genus level and to retrieve the GenBank sequences of all the complete Corynebacterium genomes from the NCBI ftp site. Briefly, Gegenees was used to divide the genomes into small sequences and to perform an all-versus-all similarity search to determine the minimum content shared by all the genomes. Next, the minimum shared content was subtracted from all the genomes, resulting in the variable content, which was compared with all the other strains to generate the percentages of similarity. Finally, these percentages were plotted in a heatmap chart with a spectrum ranging from red (low similarity) to green (high similarity) [60]. The Gegenees data can also be exported as a distance matrix file in nexus format. Here, we used the distance matrix as an input file for the SplitsTree (version 4.12.6) software to generate a phylogenomic tree using the UPGMA method [61], [62].

Pan-genome, Core Genome and Singleton Analyses

This section describes the analyses that were performed for all of the following three datasets: A) all strains, using C. pseudotuberculosis strain 1002 as a reference; B) the biovar ovis strains, using C. pseudotuberculosis strain 1002 as a reference; and C) the biovar equi strains, using C. pseudotuberculosis strain CIP52.97 as a reference. To calculate the pan-genome, core genome and singletons of the C. pseudotuberculosis species, we used EDGAR (version 1.2), multiple-strain genome comparison software that performs homology analyses based on a specific cutoff that is automatically adjusted to the query dataset [63]. Initially, the genome sequences of C. pseudotuberculosis were retrieved from GenBank, and a new project was created on the annotation platform GenDB (version 2.4) to homogenize the genome annotations [64]. Subsequently, an EDGAR project was created based on the GenDB annotations, and homology calculations based on BLAST Score Ratio Values (SRVs) were performed. According to the SRV method, instead of using raw BLAST scores or E-values, a normalization of each BLAST bit score is calculated by considering the maximum possible bit score (i.e., the bit score of the subject gene against itself). This results in a value ranging from 0 to 1 [65], which is multiplied by 100 and rounded in a percentage value of homology. Finally, a sliding window on the SRV distribution pattern was used to automatically calculate the SRV cutoff with EDGAR [63]. For this work, a SRV cuttof of 59 was estimated. Pairs of genes exhibiting a Bidirectional Best Hit where both single hits have a SRV higher than the specific cutoff were considered to be orthologous genes.

The core genome was calculated as the subset of genes presenting orthologs in all the selected strains. The gene set of subject strain A was compared with the gene set of query strain B, and only genes with orthologs in both strains were members of core AB. The resulting subset was then compared with the gene set of query strain C, and the comparisons continued in a reductive manner. The pan-genome was calculated in the same way, but in an additive manner: the initial pan-genome was composed of strain A, and the non-orthologous genes of strain B were added to pan-genome A to create the pan-genome AB. The resulting set of genes was then compared with strain C, and the comparisons continued in the same manner. Finally, the singletons were calculated as genes that were present in only one strain and thus did not present orthologs in any other C. pseudotuberculosis sequenced strain.

The developments of the core genome, pan-genome and singletons of C. pseudotuberculosis were calculated based on permutations of all the sequenced genomes. The developments of the core genome and singletons were calculated using the least-squares fit of the exponential regression decay to the mean values. In contrast, the statistical computing language R was used to calculate the pan-genome extrapolation using Heaps’ Law by estimating the parameters κ and γ using the nonlinear least-squares curve fit to the mean values [66], [67].

The core genes of all the strains, including the biovar ovis strains and the biovar equi strains, were classified by their Cluster of Orthologous Genes (COG) functional category as the following: 1. information storage and processing; 2. cellular processes and signaling; 3. metabolism; and 4. poorly characterized. To perform this analysis, the query sets of core genes were submitted to BLAST protein (blastp) similarity searches against the COG database, the proteins with E-values higher than 10−6 were discarded, and the best BLAST results for each protein were considered for the COG functional category information retrieval. Finally, the whole-genome comparison maps were visualized using the software CGView Comparison Tool (CCT) [68]. All the strains were plotted against C. pseudotuberculosis strains 1002 and CIP52.97 to generate two genome comparison maps.

Pathogenicity Island Prediction

The plasticity of the 15 genomes was assessed using PIPS: Pathogenicity Island Prediction Software (version 1.1.2). PIPS is a multi-pronged approach that predicts pathogenicity islands (PAIs) based on common features, such as G+C content, codon usage deviation, high concentrations of virulence factors and hypothetical proteins, the presence of transposases and tRNA flanking sequences, and the absence of the query region in non-pathogenic organisms of the same genus or related species [69]. C. glutamicum strain ATCC 13032 was selected as the non-pathogenic organism of the same genus [6], and separate predictions were performed for each strain. The sizes of the islands were compared with those of all the other strains via ACT: Artemis Comparison Tool (version 10.2.0) and CCT [68], [70]. Following the curation of the PAIs, the genes of all the islands in each strain were assessed for their presence/absence in all the other strains using the pan-genome data generated by EDGAR. The overall number of genes in the PAIs of the subject strain that were shared by the query strains was expressed as a percentage and plotted in a heatmap. The percentages were also converted into a nexus file, which was used in SplitsTree (version 4.12.6) to create a phylogenomic tree using the UPGMA method [61], [62]. Finally, zoomed PAI figures were created using a script from CCT (create_zoomed_maps.sh) with the zoom option selected as 30×.

Results

Phylogenomics of the Genus Corynebacterium and C. pseudotuberculosis Biovars

To evaluate the phylogenomic relationships between C. pseudotuberculosis strains and other species of the genus Corynebacterium, the Corynebacterium shared gene content was automatically determined using Gegenees. Then, the shared gene content was subtracted from all genomes and the resulting variable content of each genome sequence was cross-compared to generate a phylogenomic tree and to plot a heatmap (Figure 1). According to the generated phylogenomic tree, the pathogenic species C. diphtheriae, C. pseudotuberculosis, and C. ulcerans formed three closely related clusters. Moreover, C. glutamicum and Corynebacterium efficiens, two non-pathogenic bacteria of great industrial importance as amino acid producers [6], [71], appeared closely related in a different cluster. Additionally, Corynebacterium kroppenstedtii, another pathogenic bacterium of the Corynebacterium genus, was positioned between the clusters of pathogens (C. pseudotuberculosis, C. diphtheriae and C. ulcerans) and non-pathogens (C. glutamicum and C. efficiens). Finally, the opportunistic bacteria C. jeikeium, Corynebacterium urealyticum and Corynebacterium resistens [5], [72], [73] clustered together with the non-pathogenic Corynebacterium variabile [74], whereas Corynebacterium aurimucosum formed a new branch [75].

thumbnail

Figure 1. Phylogenomic tree and heatmap analyses of the genus Corynebacterium.

All the complete genomes from the genus Corynebacterium were retrieved from the NCBI ftp site. Comparisons between the variable content of all the strains were plotted as percentages of similarity on the heatmap using Gegenees (version 1.1.4). The percentage of similarity was used to generate a phylogenomic tree with SplitsTree (version 4.12.6). Numbers from 1 to 39 (upper-left to upper-right corner) represent species from Corynebacterium aurimucosum ATCC 70097 to Corynebacterium variable DSM 44702 (upper-left to lower-left corner). Percentages were plotted with a spectrum ranging from red (low similarity) to green (high similarity). On the heatmap, the upper portion is not symmetrical to the lower portion because the variable contents of all genomes present different sizes. Therefore, considering a scenario where the variable content from genomes A and B are composed of 100 and 80 genes, respectively, with a common repertoire of 40 genes, genome A will present 40% of similarity to genome B and genome B will present 50% of similarity to genome A.

doi:10.1371/journal.pone.0053818.g001

At the species level, the C. pseudotuberculosis genomes clustered in two separate groups representing the two biovars of the species: biovar ovis, with more than 99% similarity according to the heatmap; and biovar equi, with a similarity ranging from 95% to 100%. Moreover, the heatmap indicated an almost clonal-like behavior of C. pseudotuberculosis compared with the C. diphtheriae species, which presented similarities raging from 82% to 100%.

An alternative to assess the clonal-like behavior of species is the use of a circular genome comparison, which was performed with the software CCT. The results reveal regions of plasticity based on a chosen reference and, interestingly, plot the genomes from outer to inner circles by order of decreasing similarity. As shown in Figure 2, we plotted all the genomes using C. pseudotuberculosis strain 1002 (bv. ovis) and C. pseudotuberculosis strain CIP52.97 (bv. equi) as references. Figure 2A shows specific patterns of deletions in all the biovar equi strains compared with C. pseudotuberculosis 1002. In Figure 2B, however, the deletions in the comparison with C. pseudotuberculosis CIP52.97 are not specific to particular biovars, but rather are generalized. In both cases, the genomes that were classified as having the same biovar as the reference strain were clustered together in the outer circles, whereas the other strains were clustered in the inner circles.

thumbnail

Figure 2. Comparative genomic maps of the C. pseudotuberculosis biovar equi and ovis strains.

A, all the C. pseudotuberculosis strains were aligned using C. pseudotuberculosis strain 1002 as a reference. From the inner to outer circle on A: the biovar equi strains Cp31, Cp1/06-A, CpCp162, Cp258, Cp316 and CpCIP52.97; and, the biovar ovis strains CpC231, CpP54B96, Cp267, CpPAT10, CpI19, Cp42/02-A, Cp3/99-5, CpFRC41 and Cp1002. B, all the C. pseudotuberculosis strains were aligned using C. pseudotuberculosis strain CIP52.97 as a reference. From the inner to outer circle on B: the biovar ovis strains CpC231, Cp1002, CpPAT10, Cp267, CpP54B96, CpI19, Cp42/02-A, CpFRC41, Cp3/99-5; and, the biovar equi strains Cp1/06-A Cp31, CpCp162, Cp316, Cp258 and CpCIP52.97. CDS, coding sequences; tRNA, transfer RNA; rRNA, ribosomal RNA; and PAI, pathogenicity island.

doi:10.1371/journal.pone.0053818.g002

The Pan-genome of the Species C. pseudotuberculosis

To achieve a global view of the genome repertoire of C. pseudotuberculosis, the pan-genome (i.e., the total number of non-redundant genes) was calculated using the abovementioned SRV method with the software EDGAR (Figure 3). The resulting pan-genome of C. pseudotuberculosis contained a total of 2,782 genes, which is 1.3-fold the average total number of genes in each of the 15 strains (2,078). However, when the pan-genomes of the biovars were calculated separately, a slightly different scenario emerged, in which the biovar ovis had a pan-genome of 2,403 genes, 1.14-fold the average total number of genes in each biovar ovis strain (2,098), and the biovar equi had a pan-genome with 2,521 genes, 1.23-fold the average total number of genes in each biovar equi strain (2,047).

thumbnail

Figure 3. Pan-genome development of C. pseudotuberculosis.

Center chart, the pan-genome development using permutations of all 15 strains of C. pseudotuberculosis; upper-right chart, the pan-genome development of the C. pseudotuberculosis biovar ovis strains; lower-right chart, the pan-genome development of the C. pseudotuberculosis biovar equi strains.

doi:10.1371/journal.pone.0053818.g003

Additionally, the extrapolation of the C. pseudotuberculosis pan-genome was calculated by curve fitting based on Heaps’ Law, as represented by the formula n = κ * N−α, where n is the expected number of genes for a given number of genomes, N is the number of genomes, and the other terms are constants defined to fit the specific curve [67]. The variables κ and γ were determined to be 2,043.06 and 0.11, respectively, by using the statistical computing language R. According to Heaps’ Law,1) an α≤1 is representative of an open pan-genome, meaning that each added genome will contribute some new genes and the pan-genome will increase, and 2) an α>1 represents a closed pan-genome, in which the addition of new genomes will not significantly affect the pan-genome. Using the formula α = 1−γ, we inferred that the pan-genome of C. pseudotuberculosis is increasing with an α of 0.89, indicating that it has an open pan-genome. The extrapolation of the pan-genome was also separately calculated for both biovars, ovis and equi. Although the biovar equi had the same α as the entire pan-genome (0.89), the biovar ovis had a much-higher α of 0.94.

Core Genome of the Species C. pseudotuberculosis

The core genome of a species is defined as the subset of genes from the pan-genome that are shared by all strains. Here, the core genome of C. pseudotuberculosis was calculated with the software EDGAR by defining the subset of genes that presented orthologs in all the strains using the SRV method. The subset of core genes of C. pseudotuberculosis contained 1,504 genes, which represented 54% of the entire pan-genome of the species (2,782 genes). This subset may decrease with the addition of new genomes, as shown by the tendency of the core genes in the blue curve (Figure 4). However, although this subset may slightly decrease, the extrapolation of the curve can be calculated by the least-squares fit of the exponential regression decay to the mean values, as represented by the formula n = κ * exp[−x/τ]+tg(θ), where n is the expected subset of genes for a given number of genomes, x is the number of genomes, exp is Euler’s number, and the other terms are constants defined to fit the specific curve. Interestingly, that formula can be used to predict that with a high number of genomes (x), the κ * exp[−x/τ] term will tend toward 0, where tg(θ) represents the convergence of the genome subset. Based on this observation, the core genome of C. pseudotuberculosis tended to converge to 1,347 genes, which represented 48% of the pan-genome of the species (2,782 genes).

thumbnail

Figure 4. Core genome and singleton development of C. pseudotuberculosis.

Upper-left, the core genome development using permutations of all 15 strains of C. pseudotuberculosis; upper-center, the core genome development of the C. pseudotuberculosis biovar ovis strains; upper-right, the core genome development of the C. pseudotuberculosis biovar equi strains; lower-left, the singleton development using permutations of all 15 strains of C. pseudotuberculosis; lower-center, the singleton development of the C. pseudotuberculosis biovar ovis strains; lower-right, the singleton development of the C. pseudotuberculosis biovar equi strains.

doi:10.1371/journal.pone.0053818.g004

The separate analyses of the core genomes of biovars ovis and equi (Figure 4) presented different scenarios. The core genome of the C. pseudotuberculosis biovar ovis strains contained 1,818 genes, and it tended to stabilize at approximately 1,719 genes, according to the exponential regression decay. The C. pseudotuberculosis biovar equi strains, however, presented a more compact core genome of 1,599 genes and tended to stabilize at 1,404 genes. Altogether, with a total C. pseudotuberculosis core genome of 1,504 genes and a biovar ovis core genome of 1,818 genes, the core genome of biovar ovis is predicted to contain 314 orthologous genes that are shared by all strains from this biovar and are absent from one or more strains of biovar equi (Figure 5). Additionally, using the same strategy, the biovar equi, with 1,599 genes, contained 95 core genes that were absent from one or more strains of biovar ovis (Figure 5).

thumbnail

Figure 5. Venn diagram representing the core genomes of the C. pseudotuberculosis strains.

All genomes, the number of genes composing the core genome of all the strains; equi, the number of genes of the core genome of the C. pseudotuberculosis biovar equi strains, which were absent in one or more of the C. pseudotuberculosis biovar ovis strains; ovis, the number of genes of the core genome of the C. pseudotuberculosis biovar ovis strains, which were absent in one or more of the C. pseudotuberculosis biovar equi strains.

doi:10.1371/journal.pone.0053818.g005

The core genome of all the strains and the differential core genome of the biovar ovis and equi strains were classified by COG functional category. According to the chart in Figure 6, the core genome of all the strains had a large number of genes related to the categories “Metabolism” and “Information storage and processing”. Moreover, a high proportion of the core genome of all the strains was classified as “Poorly characterized”. However, when analyzing the differential core genes of the biovar ovis and equi strains separately, a higher proportion of “Poorly characterized” genes was clearly detected in the differential core genes when compared with the core genome of all the strains (Figure 6). Finally, the biovar equi had a larger number of genes classified under the functional category “Cellular processes and signaling” than biovar ovis strains.

thumbnail

Figure 6. Core genes of the C. pseudotuberculosis strains classified by COG functional category.

Core all, the genes composing the core genome of all the strains; core ovis, the genes of the core genome of the C. pseudotuberculosis biovar ovis strains, which were absent in one or more of the C. pseudotuberculosis biovar equi strains; core equi, the genes of the core genome of the C. pseudotuberculosis biovar equi strains, which were absent in one or more of the C. pseudotuberculosis biovar ovis strains.

doi:10.1371/journal.pone.0053818.g006

Singletons: Strain-specific Genes Detected in the Species C. pseudotuberculosis

The singletons of a strain are defined as the subset of genes that are absent from all the other strains and are thus responsible for increases in the number of genes in the pan-genome. We used the SRV method and EDGAR to calculate the subset of C. pseudotuberculosis singletons as the genes that did not present orthologs in any other strain. Moreover, by the least-squares fit of the exponential regression decay to the mean values, as previously described by the formula n = κ * exp[−x/τ]+tg(θ), we calculated the tg(θ) (Figure 4) for the three datasets: A) all the genomes, B) the biovar ovis genomes, and C) the biovar equi genomes. The tg(θ) for all the genomes was 18.805, meaning that each sequenced genome added approximately 19 genes to the total gene pool of the species C. pseudotuberculosis, i.e., the pan-genome. However, the individual analysis of each biovar revealed a scenario in which each sequenced biovar ovis strain contributed ~16 genes, but each sequenced biovar equi strain contributed ~34 genes.

Detection of PAIs in the C. pseudotuberculosis Genomes

Intraspecies genome plasticity may result from several events, of which horizontal gene transfer is particularly important because it can cause the acquisition of blocks of genes (genomic islands, or GEIs), producing evolution by quantum leaps [76]. PAIs are important in this context because they represent a class of GEIs that carry virulence genes, i.e., factors that enable or enhance the parasitic growth of an organism inside a host [77]. Therefore, high concentrations of the two following subsets of genes would be expected inside PAIs: 1) shared genes, which are shared by two or more, but not all, strains; and 2) singletons.

In previous studies, seven PAIs were identified in C. pseudotuberculosis biovar ovis strains 1002 and C231 (PiCps 1–7) [43], and four additional PAIs have been identified in C. pseudotuberculosis strain 1002 by further comparisons with C. pseudotuberculosis strains 316 and 258 (PiCps 8–11) [54][56]. The latter subset of PAIs was identified due to a better view of the two biovars and their specific patterns of plasticity. Here, we applied the same methodology used in the previous studies, using the software PIPS to achieve a global view of the PAIs in 15 C. pseudotuberculosis strains. Briefly, in addition to the previously identified 11 PAIs, we found 5 new PAIs, identified as PiCps 12–16. Although the 16 PAIs are present in all strains, they have different patterns of deletions, especially in the biovar equi strains (Figure 2). PiCp1, as previously described [43], harbors the pld gene and the fag operon and is present in all the strains. PiCp3 harbors the diphtheria toxin gene (tox) in C. pseudotuberculosis strain 31, and PiCps 7 and 15 harbor the spaD and spaA pilus gene clusters, respectively.

To assess the level of plasticity in the PAIs, we used the orthologous data predicted by EDGAR to calculate the percentage of PAIs (from each strain) present in each of the other strains. Using these data, we generated a phylogenomic tree of the strains with SplitsTree (Figure 7). The phylogenomic tree produced a clear separation of the ovis and equi biovar strains, similar to the phylogenomic tree created using Gegenees (Figure 1). A further comparison of the Gegenees and PAI phylogenomic trees revealed that the latter strategy did not cluster C. pseudotuberculosis strains 42/02-A and C231 in the same branch as did the former. However, two other branches were in agreement with the phylogenomic tree created by Gegenees: C. pseudotuberculosis strains 258 and 316 clustered together in a biovar equi group, and C. pseudotuberculosis strains 3/99-5 and FRC41 clustered in a biovar ovis group.

thumbnail

Figure 7. Phylogenomic tree and heatmap analyses of the Corynebacterium pseudotuberculosis strains based on pathogenicity island plasticity.

Comparisons between the PAI contents of all the strains were plotted as percentages of similarity on the heatmap using Gegenees (version 1.1.4). The percentages of similarity were used to generate a phylogenomic tree with SplitsTree (version 4.12.6). Numbers from 1 to 15 (upper-left to upper-right corner) represent the strains from Cp1002 to Cp1/06-A (upper-left to lower-left corner). On the heatmap, the upper portion is not symmetrical to the lower portion because the pathogenicity islands contents of all genomes present different sizes. Therefore, considering a scenario where the pathogenicity islands content from genomes A and B are composed of 100 and 80 genes, respectively, with a common repertoire of 40 genes, genome A will present 40% of similarity to genome B and genome B will present 50% of similarity to genome A.

doi:10.1371/journal.pone.0053818.g007

Additionally, we used the comparison data generated by the PAI analyses to create a new heatmap (Figure 7), from which we deduced a high level of intra-biovar similarity in the ovis strains with respect to the PAI content (82–100%). Although biovar ovis showed a lower level of similarity to biovar equi with respect to the PAI content (78–91%), the former tended to present a similar deletion pattern in the same PAIs, independent of the strain. The biovar equi strains, however, contained large deletions and a lower level of similarity intra-biovar (77–88%) and also compared with the biovar ovis PAIs (62–74%) (Figure 2A).

Variations in Pathogenicity Islands Encoding Exotoxin Virulence Factors

As described previously, the major toxin of C. pseudotuberculosis is phospholipase D (PLD), which is encoded by the pld gene and is strongly associated with the spread of bacteria throughout the host cells [1]. In a previous study, this toxin was shown to be harbored by a PAI (PiCp1) close to the fag operon, which also encodes important virulence factors that are responsible for iron acquisition in environments where this element is scarce [43]. Here, we found that the pld gene was present in 14 of 15 strains, with similarities ranging from 98–100%. This finding was expected due to the important role of PLD during the disease course; pld mutants present a diminished ability to spread throughout the host [1].

Although the pld gene plays a pivotal role in pathogenesis, C. pseudotuberculosis strain 31 contains a frameshift mutation near the 3′-end of this gene that could decrease the ability of this strain to spread throughout the host. However, C. pseudotuberculosis strain 31 was the only strain in our dataset to present another important virulence factor, the diphtheria toxin gene (tox) (Cp31_0135). The diphtheria toxin (DT) is an important virulence factor in C. diphtheriae, in which the gene was acquired through lysogenization by corynephages, meaning that the tox gene is also present in a PAI in this species and can be horizontally transferred to other organisms. Briefly, the tox gene is regulated by the chromosomal iron-dependent repressor DtxR [78], which blocks the transcription process by binding to the tox operator [79]. When gene transcription is activated, the toxin precursor is exported and cleaved into two fragments (A and B), which are joined by a disulfide bond [80]; fragment B binds the membrane of the host cell, mediating the internalization of fragment A, which exhibits ADP-ribosyltransferase activity [79], [81].

The exotoxin catalizes the transfer of adenosine diphosphate ribose (ADP-ribosylation) from nicotinamide adenine dinucleotide (NAD) to a histidine residue of elongation factor 2 (EF-2), called diphthamide. This process leads to inactivation of EF-2 and inhibits chain elongation during protein synthesis [82]. This toxin has also been identified in C. ulcerans strains, where it causes diphtheria-like illness [83], [84], and, interestingly, in two C. pseudotuberculosis strains isolated from buffalo in Egypt [10], [85]. The tox gene from C. pseudotuberculosis 31 has 560 amino acids in length, does not present any frameshift and has ~96–97% similarity to the tox genes from several C. diphtheriae strains and from corynephage β, as well as ~94–95% similarity to the tox gene from C. ulcerans 0102 (data not shown). Given the absence of the pld gene, the similarity of the tox gene from C. pseudotuberculosis to those from the C. diphtheriae strains, the conservation of all the domains and the presence of the gene in other strains isolated from buffalo in Egypt, the following question can be raised: is DT required for C. pseudotuberculosis to infect buffalo or is this feature more closely related to the geographical location (Egypt) than to the host?

Variations and Deletions Detected in PiCps 4, 5 and 9

Specific patterns of deletions in PiCps 4, 5 and 9 of C. pseudotuberculosis CIP52.97, 316 and 258 (biovar equi strains) have been demonstrated [54][56]. Here, we detected the same deletions in all the biovar equi strains, which indicates that these deletion events were specific to the mentioned biovar (Figure S1). Although most of the deleted CDSs encoded hypothetical or phage proteins (integrases and phage-associated proteins), one gene of PiCp5 encoded a putative sigma 70 factor (Cp1002_1452) and deserves attention because it is most likely involved in the correct assembly of the transcription machinery at specific promoters and is therefore associated with the general transcription process [43].

Differences between Pilus Gene Clusters Located on PiCp15 and PiCp7

According to work performed by Yanagawa and Honda in 1976 [47], C. pseudotuberculosis cells possess pilus structures, although the number of pili per bacterial cell is small, and at times, a long bundle measuring more than several micrometers in length was the only pilus observed. In a more recent genomic study, two clusters of pilus genes were described in C. pseudotuberculosis FRC41 and were named according to their major pilin gene: the spaA (srtB-spaA-srtA-spaB-spaX-spaC) and spaD (srtC-spaD-spaY-spaE-spaF) clusters, where srtA and srtB are the specific sortases of the spaA cluster; spaA, spaB and spaC encode the major, base and tip pilin proteins, respectively, of the spaA cluster; srtC is the specific sortase of the spaD cluster; spaD, spaE and spaF encode the major, base and tip pilin proteins, respectively, of the spaD cluster; and spaX and spaY have currently unknown functions. Additionally, a housekeeping sortase (srtD) is likely responsible for anchoring the pili to the cellwall [46].

Interestingly, the spaA and spaD gene clusters were located in PAIs (PiCps 15 and 7, respectively) (Figure 8), which is in agreement with the presence of pilin genes in horizontally acquired regions of Gram-negative and Gram-positive bacteria, such as Vibrio cholerae and C. diphtheriae, respectively [86], [87]. Moreover, although the biovar ovis strains had a complete spaA cluster, the biovar equi strains contained a large deletion at the position where the spaA and srtB genes should be located (PiCp15). Furthermore, the entire srtA-spaB-spaX-spaC region presented a low similarity to the same region in the biovar ovis strains, which was caused by small deletions, frameshift mutations and nucleotide substitutions (Figure 8).

thumbnail

Figure 8. Plasticity of the pilus gene clusters spaA and spaD in C. pseudotuberculosis.

A1 and B1, PiCp15 harboring the spaA cluster of genes; A2 and B2, PiCp7 harboring the spaD cluster of genes. A, all the C. pseudotuberculosis strains were aligned using C. pseudotuberculosis strain 1002 as a reference. From the inner to outer circle on A1 and A2: the biovar equi strains Cp31, Cp1/06-A, CpCp162, Cp258, Cp316, CpCIP52.97; and, the biovar ovis strains CpC231, CpP54B96, Cp267, CpPAT10, CpI19, Cp42/02-A, Cp3/99-5, CpFRC41 and Cp1002. B, all the C. pseudotuberculosis strains were aligned using C. pseudotuberculosis strain CIP52.97 as a reference. From the inner to outer circle on B1 and B2: the biovar ovis strains CpC231, Cp1002, CpPAT10, Cp267, CpP54B96, CpI19, Cp42/02-A, CpFRC41, Cp3/99-5, Cp1/06-A; and, the biovar equi strains Cp31, CpCp162, Cp316, Cp258 and CpCIP52.97. CDS, coding sequences; tRNA, transfer RNA; rRNA, ribosomal RNA; and PAI, pathogenicity island.

doi:10.1371/journal.pone.0053818.g008

With respect to the spaD cluster of the biovar ovis strains, the major pilin gene spaD contains a frameshift in C. pseudotuberculosis P54B96 and PAT10; and in C. pseudotuberculosis 267, the tip pilin gene spaF also contains a frameshift. In biovar equi strains, the spaD gene of all the strains had 99% similarity to the spaD gene of the biovar ovis strains. However, C. pseudotuberculosis CIP52.97 contains a frameshift mutation in the specific sortase gene srtC. Furthermore, the base and tip pilin genes, spaE and spaF, respectively, of C. pseudotuberculosis strains 258, 316, 1/06-A and Cp162 are merged into the same reading frame.

Discussion

Corynebacterium pseudotuberculosis – all Strains

According to the rpoB gene tree generated by Khamis et al. [88], C. jeikeium, C. urealyticum, C. kroppenstedtii and C. variabile cluster together in group 3, and C. aurimucosum appears in group 1. Moreover, C. glutamicum and C. efficiens cluster together in one branch, whereas C. pseudotuberculosis, C. diphtheriae and C. ulcerans appear closely related in another branch. Furthermore, C. ulcerans appears closer to C. pseudotuberculosis than to C. diphtheriae. Based on our results, we can deduce that although many variable regions exist between the pathogenic members of the genus Corynebacterium, these species tend to cluster together because they most likely share some core virulence determinants. Finally, although C. kroppenstedtii did not cluster with group 3, the other species were in perfect agreement with the rpoB analysis of Khamis et al. [88].

Two striking characteristics of C. kroppenstedtii are the absence of mycolic acids in the cell wall (due to the losses of a condensase gene cluster and a mycolate reductase gene) and a lipophilic phenotype (due to the absence of a microbial type I fatty acid synthase gene) [89]. Therefore, the transitional phylogenomic position of C. kroppenstedtii between the pathogenic and non-pathogenic species was in agreement with the lack of important virulence genes and the low pathogenic potential characteristic of C. kroppenstedtii [89][91].

At the species level, the heatmap indicated a clonal-like behavior of C. pseudotuberculosis compared with the C. diphtheriae species. Trost et al. [87] have highlighted the high plasticity of the C. diphtheriae genome, which is mainly related to the 57 genomic islands identified in this species. With respect to the clonal-like behavior of C. pseudotuberculosis, Bolt [92] have identified 10 STs among 73 strains of C. pseudotuberculosis typed by MLST, where 7 and 4 STs were associated with 64 and 9 strains of biovar ovis and equi, respectively. The few number of STs identified by MLST was in agreement with previous typing studies [17], [93], [94] in that the strains of C. pseudotuberculosis are clonally related. Moreover, although there were 7 STs identified for biovar ovis strains, 6 and 7 of them were clustered in one sole eBURST group when considering single locus variants (SLVs) and double locus variants (DLVs), respectively; and, all the STs identified for biovar equi shared two alleles with the biovar ovis strains [92]. Finally, the MLST findings indicate that: 1) biovar ovis and equi strains share a common evolutionary origin, although they are now relatively distinct genotypic clusters; and, 2) biovar ovis is a clonal-like organism. Our results with respect to this clonal-like behavior of C. pseudotuberculosis are also in agreement with PFGE data from Connor et al. [17] and can also be inferred from the extrapolation of the pan-genome data, in which C. pseudotuberculosis had a slightly higher α value of 0.89 compared with the C. diphtheriae α value of 0.69; and, from the total number of genes in the pan-genome of C. pseudotuberculosis (2,782 genes), which is compact compared with that of the closely related species C. diphtheriae, which contains 4,786 genes [87].

Although C. pseudotuberculosis displays some clonal-like behavior, the resulting α of 0.89 from the extrapolation of the pan-genome indicates that it has an open pan-genome. Moreover, considering that α is inversely proportional to the pan-genome increasing rate, in contrast to the C. diphtheriae α of 0.69, the α of 0.89 of the C. pseudotuberculosis pan-genome indicates that the latter is increasing at a slower rate. This slow increase is related to the low number of singletons (~19) added to the pan-genome of C. pseudotuberculosis by each newly sequenced strain, whereas each strain of C. diphtheriae added ~65 genes to the entire pan-genome [87]. Moreover, the slow increase and higher α value are in agreement with the intracellular facultative behavior of this species. Because strictly intracellular organisms tend to have closed pan-genomes due to their limited contact with potential gene donors, an intracellular facultative organism such as C. pseudotuberculosis, even when it has different hosts, can be expected to have an α that is closer to 1 than that of C. diphtheriae [95], [96].

With respect to the core genome of all the strains, a large number of genes are related to the categories “Metabolism” and “Information storage and processing”. The “Information storage and processing” category contains genes involved in translation, ribosomal structure and biogenesis, RNA processing and modification, transcription, replication, recombination and repair, and other important functions; the “Metabolism” category contains genes involved in the production and conversion of energy, as well as the transport and metabolism of carbohydrates, amino acids, nucleotides, coenzymes, lipids, inorganic ions and secondary metabolites. Given the importance of the core genome, these two functional categories are expected to be highly represented in the analyzed subset. Finally, although a large number of “Poorly characterized” genes were identified in the core gene subset, this result is in agreement with previous core genome analyses of Aggregatibacter actinomycetemcomitans, in which one-third of the genes were categorized as “Poorly characterized” and approximately one-third were classified under “Metabolism” [97].

Corynebacterium pseudotuberculosis – Biovars Ovis and Equi

Connor et al. [17] and Bolt [92] have investigated the clonal aspect of C. pseudotuberculosis using PFGE and MLST, respectively, which enabled them to differentiate the equi and ovis biovars. On the phylogenomic tree, the C. pseudotuberculosis genomes also clustered in two separate groups representing the two biovars of the species: biovar ovis, with more than 99% similarity according to the heatmap; and biovar equi, with a similarity ranging from 95% to almost 100%. This result highlights the higher plasticity of C. pseudotuberculosis biovar equi compared with the biovar ovis strains, although this plasticity is not as high as that described for C. diphtheriae strains. Moreover, the same conclusion (regarding the relative plasticity of the two biovars) may be drawn from the number of singletons, in which the biovar equi strains presented higher levels of variability in the number of singletons, compared with the biovar ovis strains (Table 1). The circular genome comparison generated by CCT also revealed the clonal-like behavior of biovar ovis, with all the ovis strains containing minor deletions compared with C. pseudotuberculosis strain 1002 (Figure 2A); and the presence of a higher number of singletons in biovar equi, with all the strains from both biovars presenting similar deletion patterns when compared with C. pseudotuberculosis strain CIP52.97 (Figure 2B). Finally, the majority of the genomic variations on the circular genome comparison were found in PAI regions, which are very important for virulence potential and host adaptation and are known as mosaic and unstable [69].

Interestingly, the analysis of the pan-genome subsets revealed that the ovis and equi biovar strains contain major variations of the data found in the entire pan-genome. Although the pan-genome of biovar equi had an invariable α value of 0.89, the pan-genome of the biovar ovis had a higher α value of 0.94, which was strictly correlated to the higher clonal-like behavior of this biovar compared with biovar equi [92]. Moreover, its high α value and the pan-genome curve suggest that the pan-genome of biovar ovis is increasing at a slower rate than that of biovar equi.

The same conclusion may be drawn from the development of singletons: each biovar ovis strain added ~16 singletons to the pan-genome, but each biovar equi strain added ~34 singletons to the gene pool. Moreover, although the core genome subset of the biovar ovis strains (1,818 CDS) was slightly higher than that of the biovar equi strains (1,599 CDS), most of the variable genes of the biovar ovis strains were acquired in blocks through horizontal gene transfer and are highly conserved throughout the entire biovar, as shown in Figure 2A. In contrast, the biovar equi strains presented great variability, both intra- and inter-biovar, in the content of the detected pathogenicity islands (Figure 2B). Finally, a comparison of the similarity levels on the two heatmaps, generated by Gegenees (93–100%, Figure 1) and from PAI contents (62–100%, Figure 7), also revealed that most of the variability defining the biovars ovis and equi arose from the gene content of the PAIs.

In view of this, one possible explanation for the large number of “Poorly characterized” genes in the differential core subsets of both biovars ovis and equi is the abovementioned acquisition of these subsets by horizontal gene transfer, which tends to involve a large number of hypothetical proteins [98], and the maintenance of these acquired regions in different biovars because they enabled the biovars to colonize specific hosts. Finally, the higher proportion of the functional category “Cellular processes and signaling” in biovar equi is most likely related to host adaptation because many genes in this cluster had functions such as defense mechanisms, signal transduction mechanisms, cell wall/membrane/envelope biogenesis, cell motility, and extracellular structures.

Variations in Pilus Gene Clusters

With respect to the gene content of the PAIs, the most interesting finding is the high similarity of the pilus genes in the biovar ovis strains, which is in contrast to the large variability of these genes in the biovar equi strains. Pilus gene clusters are normally acquired in a block through horizontal gene transfer and are composed of a specific sortase gene and the major, base and tip pilin genes. Briefly, the specific sortase protein of each cluster is responsible for cleaving the LPxTG motif of the major, base and tip pilin proteins of that cluster between the threonine (T) and glycine (G) amino acids, capturing the cleaved polypeptides, polymerizing the monomers, and transferring the final product to the housekeeping sortase of the bacterium for its final incorporation into the cell wall [99], [100]. In the absence of a housekeeping sortase, the pilus-specific sortase can mediate the incorporation of the polymer into the cell wall. However, the presence of both housekeeping and specific sortases is necessary to efficiently anchor the pilus to the cell wall [101]. Moreover, although the expression of the major pilin is absolutely required for the specific pilus polymerization, the base and tip pilin monomers may still attach to the cell wall in its absence [100][103].

Although the biovar ovis strains present a complete spaA cluster, the biovar equi were shown to present large deletions in this cluster. Because of the deletion of the major pilin SpaA in the biovar equi, the base and tip pilin monomers would be expected to be the only pilin structures that could attach to the cell wall in a non-polymerized manner. Moreover, the deletion of one of the specific sortase genes in biovar equi, srtB, could also interfere in the efficient cell wall-anchoring of these monomers, causing them to be secreted [101]. Finally, even the production and sizes of these proteins may vary among the biovar equi strains because these proteins contain small deletions and frameshift mutations. Altogether, the differences in the spaA cluster of the biovar equi strains could account for the different levels of host cell attachment compared with the biovar ovis strains and even among the biovar equi strains, as found in the C. diphtheriae species 87,104,105.

In contrast to the high similarity found between the spaA clusters of the biovar ovis strains, the spaD clusters presented differences in three strains of this biovar. In C. pseudotuberculosis P54B96 and PAT10, a frameshit in the major pilin gene spaD impairs the coding of the entire protein and, thus, the polymerization of the pilin structure; and, in C. pseudotuberculosis 267, the tip pilin gene spaF also contains a frameshift. Although the tip pilin is not required for the polymerization of the pilin structure and adhesion to the host cell wall, its absence can slightly decrease the degree of adherence, which could reduce the spread of C. pseudotuberculosis strain 267 [106]. With respect to the spaD cluster of the biovar equi strains, a frameshift mutation in the specific sortase gene srtC of C. pseudotuberculosis CIP52.97 prevents the polymerization of the pilin structure. Moreover, the base and tip pilin genes, spaE and spaF, respectively, of C. pseudotuberculosis strains 258, 316, 1/06-A and Cp162 are merged into the same reading frame. Overall, these results suggest that although C. pseudotuberculosis 258, 316, 1/06-A and Cp162 can polymerize the major pilin, C. pseudotuberculosis strain 31 is most likely the only biovar equi strain able to polymerize an entire pilin structure from the spaD cluster, whereas all the biovar ovis strains are likely capable of producing one or two types of pilin structures (spaA and spaD).

Summarizing, all the C. pseudotuberculosis biovar ovis strains likely contain a functional spaA cluster of pilus genes; only three strains (267, P54B96 and PAT10) are unable to polymerize an entire spaD pilin structure (most likely, they instead attach monomers or incompletely polymerized pilin structures). In contrast, all the biovar equi strains contain deletions, which render them unable to polymerize any spaA pilin structures; within this biovar, only C. pseudotuberculosis 31 appears to be able to polymerize an entire spaD pilin structure. Given the pivotal role played by pili in the processes of adhesion and internalization, the polymerization of complete pilin structures in the biovar ovis strains could be responsible for the great ability of these strains to spread throughout host tissues and penetrate cells to grow intracellularly [48], [101], [106], [107]. Based on this observation, the biovar ovis strains are expected to have less contact with other organisms than the biovar equi strains and to therefore show more clonal-like behavior. Finally, these results could also explain the distinct pattern of the diseases caused by C. pseudotuberculosis in horses, which involves ulcerative lymphangitis that rarely evolves to a visceral form [108]. However, more studies are needed to assess whether the C. pseudotuberculosis biovars equi and ovis truly present different patterns of pilin formation and, thus, variable degrees of host tissue adhesion, spreading and cell internalization.

Supporting Information

Figure S1.

Plasticity of PiCps 4, 5 and 9. A1 and B1, PiCp9; A2 and B2, PiCp4; A3 and B3, PiCp5. A, all the C. pseudotuberculosis strains were aligned using C. pseudotuberculosis strain 1002 as a reference. From the inner to outer circle on A1, A2 and A3: the biovar equi strains Cp31, Cp1/06-A, CpCp162, Cp258, Cp316, CpCIP52.97; and, the biovar ovis strains CpC231, CpP54B96, Cp267, CpPAT10, CpI19, Cp42/02-A, Cp3/99-5, CpFRC41 and Cp1002. B, all the C. pseudotuberculosis strains were aligned using C. pseudotuberculosis strain CIP52.97 as a reference. From the inner to outer circle on B1, B2 and B3: the biovar ovis strains CpC231, Cp1002, CpPAT10, Cp267, CpP54B96, CpI19, Cp42/02-A, CpFRC41, Cp3/99-5, Cp1/06-A; and, the biovar equi strains Cp31, CpCp162, Cp316, Cp258 and CpCIP52.97. CDS, coding sequences; tRNA, transfer RNA; rRNA, ribosomal RNA; and PAI, pathogenicity island.

doi:10.1371/journal.pone.0053818.s001

(TIFF)

Acknowledgments

The authors thank the CAPES/DAAD international cooperation for financing an exchange scholarship (http://www.capes.gov.br/cooperacao-inter​nacional) (grant number 5117119) and CLIB - Graduate Cluster Industrial Biotechnology (http://www.graduatecluster.net/), where SCS and LCG are currently participating as guest students.

Author Contributions

Read and gave insights about the manuscript: SCS AS ET JB RR AC AA ARS ACP CD EGVB FAD FA FSR KKFN LCG SA SSH SMB UPP VACA MPCS AM AT VA. Conceived and designed the experiments: AT VA. Performed the experiments: SCS ET JB RR AC AA ARS ACP CD EGVB FAD FA FSR KKFN LCG SA SSH SMB UPP VACA. Analyzed the data: SCS ET JB. Contributed reagents/materials/analysis tools: SCS AS ET JB AT VA. Wrote the paper: SCS AS MPCS AM AT VA.

References

  1. 1. Dorella FA, Pacheco LGC, Oliveira SC, Miyoshi A, Azevedo V (2006) Corynebacterium pseudotuberculosis: microbiology, biochemical properties, pathogenesis and molecular studies of virulence. Vet Res 37: 201–218. doi: 10.1051/vetres:2005056
  2. 2. Lehman KB, Neumann R (1896) Atlas und grundriss der bakeriologie und lehrbuch der speziellen bakteriologischen diagnositk. 1st ed. J.F. Lehmann, Munchen.
  3. 3. Pascual C, Lawson PA, Farrow JA, Gimenez MN, Collins MD (1995) Phylogenetic analysis of the genus Corynebacterium based on 16S rRNA gene sequences. Int J Syst Bacteriol 45: 724–728. doi: 10.1099/00207713-45-4-724
  4. 4. Cerdeño-Tárraga AM, Efstratiou A, Dover LG, Holden MTG, Pallen M, et al. (2003) The complete genome sequence and analysis of Corynebacterium diphtheriae NCTC13129. Nucleic Acids Res 31: 6516–6523. doi: 10.1093/nar/gkg874
  5. 5. Tauch A, Kaiser O, Hain T, Goesmann A, Weisshaar B, et al. (2005) Complete genome sequence and analysis of the multiresistant nosocomial pathogen Corynebacterium jeikeium K411, a lipid-requiring bacterium of the human skin flora. J Bacteriol 187: 4671–4682. doi: 10.1128/jb.187.13.4671-4682.2005
  6. 6. Kalinowski J, Bathe B, Bartels D, Bischoff N, Bott M, et al. (2003) The complete Corynebacterium glutamicum ATCC 13032 genome sequence and its impact on the production of L-aspartate-derived amino acids and vitamins. J Biotechnol 104: 5–25. doi: 10.1016/s0168-1656(03)00154-8
  7. 7. Jones D, Collins MD (1986) Irregular, nonsporing gram-positive rods, section 15. pages 1261–1579 in bergey’s manual of systematic bacteriology. Williams & Wilkins, Co., Baltimore, MD.
  8. 8. Buck GA, Cross RE, Wong TP, Loera J, Groman N (1985) DNA relationships among some tox-bearing corynebacteriophages. Infect Immun 49: 679–684.
  9. 9. Groman N, Schiller J, Russell J (1984) Corynebacterium ulcerans and Corynebacterium pseudotuberculosis responses to DNA probes derived from corynephage beta and Corynebacterium diphtheriae. Infect Immun 45: 511–517.
  10. 10. Wong TP, Groman N (1984) Production of diphtheria toxin by selected isolates of Corynebacterium ulcerans and Corynebacterium pseudotuberculosis. Infect Immun 43: 1114–1116.
  11. 11. Muckle CA, Gyles CL (1982) Characterization of strains of Corynebacterium pseudotuberculosis. Can J Comp Med 46: 206–208.
  12. 12. Biberstein EL, Knight HD, Jang S (1971) Two biotypes of Corynebacterium pseudotuberculosis. Vet Rec 89: 691–692. doi: 10.1136/vr.89.26.691
  13. 13. Ayers JL (1977) Caseous lymphadenitis in goat and sheep: Review of diagnosis, pathogenesis, and immunity. JAVMA n. 171: 1251–1254.
  14. 14. Ben Saïd MS, Ben Maitigue H, Benzarti M, Messadi L, Rejeb A, et al. (2002) Epidemiological and clinical studies of ovine caseous lymphadenitis. Arch Inst Pasteur Tunis 79: 51–57.
  15. 15. Arsenault J, Girard C, Dubreuil P, Daignault D, Galarneau JR, et al. (2003) Prevalence of and carcass condemnation from maedi-visna, paratuberculosis and caseous lymphadenitis in culled sheep from Quebec, Canada. Prev Vet Med 59: 67–81. doi: 10.1016/s0167-5877(03)00060-6
  16. 16. Binns SH, Bailey M, Green LE (2002) Postal survey of ovine caseous lymphadenitis in the United Kingdom between 1990 and 1999. Vet Rec 150: 263–268. doi: 10.1136/vr.150.9.263
  17. 17. Connor KM, Quirie MM, Baird G, Donachie W (2000) Characterization of United Kingdom isolates of Corynebacterium pseudotuberculosis using pulsed-field gel electrophoresis. J Clin Microbiol 38: 2633–2637. doi: 10.1051/vetres:2007013
  18. 18. Paton MW, Walker SB, Rose IR, Watt GF (2003) Prevalence of caseous lymphadenitis and usage of caseous lymphadenitis vaccines in sheep flocks. Aust Vet J 81: 91–95. doi: 10.1111/j.1751-0813.2003.tb11443.x
  19. 19. Hodgson AL, Carter K, Tachedjian M, Krywult J, Corner LA, et al. (1999) Efficacy of an ovine caseous lymphadenitis vaccine formulated using a genetically inactive form of the Corynebacterium pseudotuberculosis phospholipase D. Vaccine. 17: 802–808. doi: 10.1016/s0264-410x(98)00264-3
  20. 20. Pugh DG (2002) Caseous Lymphadenitis. In: Sheep & Goat Medicine Saunders 207–208.
  21. 21. Radostits OM, Gay CC, Blood DC, Hinchcliff KW (2002) Clínica veterinária. um tratado de doenças dos bovinos, ovinos, suínos, caprinos e eqüinos. Ed. Guanabara, Koogan, 9a edição.
  22. 22. Augustine JL, Renshaw HW (1986) Survival of Corynebacterium pseudotuberculosis in axenic purulent exudate on common barnyard fomites. Am J Vet Res 47: 713–715.
  23. 23. Yeruham I, Friedman S, Perl S, Elad D, Berkovich Y, et al. (2004) A herd level analysis of a Corynebacterium pseudotuberculosis outbreak in a dairy cattle herd. Vet Dermatol 15: 315–320. doi: 10.1111/j.1365-3164.2004.00388.x
  24. 24. Yeruham I, Elad D, Friedman S, Perl S (2003) Corynebacterium pseudotuberculosis infection in Israeli dairy cattle. Epidemiol Infect 131: 947–955. doi: 10.1017/s095026880300894x
  25. 25. Collett MG, Bath GF, Cameron CM (1994) Corynebacterium pseudotuberculosis infections. In: Infections diseases of livestock with special reference to Southern Africa. Oxford University Press 2: 1387–1395.
  26. 26. Dorella FA, Pacheco LG, Seyffert N, Portela RW, Meyer R, et al. (2009) Antigens of Corynebacterium pseudotuberculosis and prospects for vaccine development. Expert Rev Vaccines 8: 205–213. doi: 10.1586/14760584.8.2.205
  27. 27. Williamson LH (2001) Caseous lymphadenitis in small ruminants. Vet. Clin. North Am. Food Anim. Pract 17: 359–371.
  28. 28. Liu DTL, Chan W, Fan DSP, Lam DSC (2005) An infected hydrogel buckle with Corynebacterium pseudotuberculosis. Br J Ophthalmol 89: 245–246. doi: 10.1136/bjo.2004.051698
  29. 29. Mills AE, Mitchell RD, Lim EK (1997) Corynebacterium pseudotuberculosis is a cause of human necrotising granulomatous lymphadenitis. Pathology 29: 231–233. doi: 10.1080/00313029700169944
  30. 30. Peel MM, Palmer GG, Stacpoole AM, Kerr TG (1997) Human lymphadenitis due to Corynebacterium pseudotuberculosis: report of ten cases from Australia and review. Clin Infect Dis 24: 185–191. doi: 10.1093/clinids/24.2.185
  31. 31. Barakat AA, Selim SA, Atef A, Saber MS, Nafie EK, et al. (1984) Two serotypes of Corynebacterium pseudotuberculosis isolated from different animal species. Revue Scientifique et Technique de l’OIE 3(1): 151–163.
  32. 32. Aleman M, Spier SJ, Wilson WD, Doherr M (1996) Corynebacterium pseudotuberculosis infection in horses: 538 cases (1982–1993). J Am Vet Med Assoc 209: 804–809.
  33. 33. Pratt SM, Spier SJ, Carroll SP, Vaughan B, Whitcomb MB, et al. (2005) Evaluation of clinical characteristics, diagnostic test results, and outcome in horses with internal infection caused by Corynebacterium pseudotuberculosis: 30 cases (1995–2003). J Am Vet Med Assoc 227: 441–448. doi: 10.2460/javma.2005.227.441
  34. 34. Braverman Y, Chizov-Ginzburg A, Saran A, Winkler M (1999) The role of houseflies (Musca domestica) in harbouring Corynebacterium pseudotuberculosis in dairy herds in Israel. Revue Scientifique et Technique de l’OIE 18 n° 3: 681–690.
  35. 35. Addo P (1983) Role of the common house fly (Musca domestica) in the spread of ulcerative lymphangitis. Vet Rec 113(21): 496–497. doi: 10.1136/vr.113.21.496
  36. 36. Selim SA (2001) Oedematous skin disease of buffalo in Egypt. J Vet Med B Infect Dis Vet Public Health 48: 241–258. doi: 10.1046/j.1439-0450.2001.00451.x
  37. 37. Yeruham I, Braverman Y, Shpigel NY, Chizov-Ginzburg A, Saran A, et al. (1996) Mastitis in dairy cattle caused by Corynebacterium pseudotuberculosis and the feasibility of transmission by houseflies. I. Vet Q 18: 87–89. doi: 10.1080/01652176.1996.9694623
  38. 38. Spier S (2008) Corynebacterium pseudotuberculosis infection in horses: An emerging disease associated with climate change? Equine Veterinary Education 20: 37–39. doi: 10.2746/095777307x260106
  39. 39. McKean S, Davies J, Moore R (2005) Identification of macrophage induced genes of Corynebacterium pseudotuberculosis by differential fluorescence induction. Microbes Infect 7: 1352–1363. doi: 10.1016/j.micinf.2005.05.002
  40. 40. McKean SC, Davies JK, Moore RJ (2007) Expression of phospholipase D, the major virulence factor of Corynebacterium pseudotuberculosis, is regulated by multiple environmental factors and plays a role in macrophage death. Microbiology 153: 2203–2211. doi: 10.1099/mic.0.2007/005926-0
  41. 41. Schumann W (2007) Thermosensors in eubacteria: role and evolution. J Biosci 32: 549–557. doi: 10.1007/s12038-007-0054-8
  42. 42. Billington SJ, Esmay PA, Songer JG, Jost BH (2002) Identification and role in virulence of putative iron acquisition genes from Corynebacterium pseudotuberculosis. FEMS Microbiol : Lett. 208, 41–45.
  43. 43. Ruiz JC, D’Afonseca V, Silva A, Ali A, Pinto AC, et al. (2011) Evidence for reductive genome evolution and lateral acquisition of virulence functions in two Corynebacterium pseudotuberculosis strains. PLoS One 6: e18551. doi: 10.1371/journal.pone.0018551
  44. 44. Alves FSF, Olander H (1999) Uso de vacina toxóide no controle da linfadenite caseosa em caprinos. Veterinária Notícias, Uberlândia n° 5: 69–75.
  45. 45. Songer JG, Libby SJ, Iandolo JJ, Cuevas WA (1990) Cloning and expression of the phospholipase D gene from Corynebacterium pseudotuberculosis in Escherichia coli. Infect Immun 58: 131–136.
  46. 46. Trost E, Ott L, Schneider J, Schröder J, Jaenicke S, et al. (2010) The complete genome sequence of Corynebacterium pseudotuberculosis FRC41 isolated from a 12-year-old girl with necrotizing lymphadenitis reveals insights into gene-regulatory networks contributing to virulence. BMC Genomics 11: 728. doi: 10.1186/1471-2164-11-728
  47. 47. Yanagawa R, Honda E (1976) Presence of pili in species of human and animal parasites and pathogens of the genus Corynebacterium. Infect Immun 13: 1293–1295.
  48. 48. Wilson JW, Schurr MJ, LeBlanc CL, Ramamurthy R, Buchanan KL, et al. (2002) Mechanisms of bacterial pathogenicity. Postgrad Med J 78: 216–224.
  49. 49. Pethick FE, Lainson AF, Yaga R, Flockhart A, Smith DGE, et al. (2012) Complete Genome Sequences of Corynebacterium pseudotuberculosis Strains 3/99–5 and 42/02-A, Isolated from Sheep in Scotland and Australia, Respectively. J Bacteriol 194: 4736–4737. doi: 10.1128/jb.00918-12
  50. 50. Cerdeira LT, Pinto AC, Schneider MPC, de Almeida SS, dos Santos AR, et al. (2011) Whole-genome sequence of Corynebacterium pseudotuberculosis PAT10 strain isolated from sheep in Patagonia, Argentina. J Bacteriol 193: 6420–6421. doi: 10.1128/jb.06044-11
  51. 51. Lopes T, Silva A, Thiago R, Carneiro A, Dorella FA, et al. (2012) Complete Genome Sequence of Corynebacterium pseudotuberculosis Strain Cp267, Isolated from a Llama. J Bacteriol 194: 3567–3568. doi: 10.1128/jb.00461-12
  52. 52. Silva A, Schneider MPC, Cerdeira L, Barbosa MS, Ramos RTJ, et al. (2011) Complete genome sequence of Corynebacterium pseudotuberculosis I19, a strain isolated from a cow in Israel with bovine mastitis. J Bacteriol 193: 323–324. doi: 10.1128/jb.01211-10
  53. 53. Cerdeira LT, Schneider MPC, Pinto AC, de Almeida SS, dos Santos AR, et al. (2011) Complete genome sequence of Corynebacterium pseudotuberculosis strain CIP 52.97, isolated from a horse in Kenya. J Bacteriol 193: 7025–7026. doi: 10.1128/jb.06293-11
  54. 54. Ramos RTJ, Silva A, Carneiro AR, Pinto AC, Soares SDC, et al. (2012) Genome Sequence of the Corynebacterium pseudotuberculosis Cp316 Strain, Isolated from the Abscess of a Californian Horse. J Bacteriol 194: 6620–6621. doi: 10.1128/jb.01616-12
  55. 55. Ramos RTJ, Carneiro AR, Soares SC, Santos AR, Almeida SS, et al.. (2013) Tips and tricks for the assembly a Corynebacterium pseudotuberculosis genome using a semiconductor sequencer. Microbial Biotechnology in press.
  56. 56. Soares SC, Trost E, Ramos RTJ, Carneiro AR, Santos AR, et al.. (2012) Genome sequence of Corynebacterium pseudotuberculosis biovar equi strain 258 and prediction of antigenic targets to improve biotechnological vaccine production. J Biotechnol in press.
  57. 57. Pethick FE, Lainson AF, Yaga R, Flockhart A, Smith DGE, et al. (2012) Complete Genome Sequence of Corynebacterium pseudotuberculosis Strain 1/06-A, Isolated from a Horse in North America. J Bacteriol 194: 4476. doi: 10.1128/jb.00922-12
  58. 58. Hassan SS, Schneider MPC, Ramos RTJ, Carneiro AR, Ranieri A, et al. (2012) Whole-Genome Sequence of Corynebacterium pseudotuberculosis Strain Cp162, Isolated from Camel. J Bacteriol 194: 5718–5719. doi: 10.1128/jb.01373-12
  59. 59. Silva A, Ramos RTJ, Ribeiro Carneiro A, Cybelle Pinto A, de Castro Soares S, et al. (2012) Complete Genome Sequence of Corynebacterium pseudotuberculosis Cp31, Isolated from an Egyptian Buffalo. J Bacteriol 194: 6663–6664. doi: 10.1128/jb.01782-12
  60. 60. Agren J, Sundström A, Håfström T, Segerman B (2012) Gegenees: fragmented alignment of multiple genomes for determining phylogenomic distances and genetic signatures unique for specified target groups. PLoS One 7: e39107. doi: 10.1371/journal.pone.0039107
  61. 61. Huson DH, Bryant D (2006) Application of phylogenetic networks in evolutionary studies. Mol Biol Evol 23: 254–267. doi: 10.1093/molbev/msj030
  62. 62. Kloepper TH, Huson DH (2008) Drawing explicit phylogenetic networks and their integration into SplitsTree. BMC Evol Biol 8: 22. doi: 10.1186/1471-2148-8-22
  63. 63. Blom J, Albaum SP, Doppmeier D, Pühler A, Vorhölter F, et al. (2009) EDGAR: a software framework for the comparative analysis of prokaryotic genomes. BMC Bioinformatics 10: 154. doi: 10.1186/1471-2105-10-154
  64. 64. Meyer F, Goesmann A, McHardy AC, Bartels D, Bekel T, et al. (2003) GenDB–an open source genome annotation system for prokaryote genomes. Nucleic Acids Res 31: 2187–2195. doi: 10.1093/nar/gkg312
  65. 65. Lerat E, Daubin V, Moran NA (2003) From gene trees to organismal phylogeny in prokaryotes: the case of the gamma-Proteobacteria. PLoS Biol 1: E19. doi: 10.1371/journal.pbio.0000019
  66. 66. Tettelin H, Masignani V, Cieslewicz MJ, Donati C, Medini D, et al. (2005) Genome analysis of multiple pathogenic isolates of Streptococcus agalactiae: implications for the microbial pan-genome. Proc Natl Acad Sci U S A 102: 13950–13955. doi: 10.1073/pnas.0506758102
  67. 67. Tettelin H, Riley D, Cattuto C, Medini D (2008) Comparative genomics: the bacterial pan-genome. Curr Opin Microbiol 11: 472–477. doi: 10.1016/j.mib.2008.09.006
  68. 68. Grant JR, Arantes AS, Stothard P (2012) Comparing thousands of circular genomes using the CGView Comparison Tool. BMC Genomics 13: 202. doi: 10.1186/1471-2164-13-202
  69. 69. Soares SC, Abreu VAC, Ramos RTJ, Cerdeira L, Silva A, et al. (2012) PIPS: pathogenicity island prediction software. PLoS One 7: e30848. doi: 10.1371/journal.pone.0030848
  70. 70. Carver TJ, Rutherford KM, Berriman M, Rajandream M, Barrell BG, et al. (2005) ACT: the Artemis Comparison Tool. Bioinformatics 21: 3422–3423. doi: 10.1093/bioinformatics/bti553
  71. 71. Nishio Y, Nakamura Y, Kawarabayasi Y, Usuda Y, Kimura E, et al. (2003) Comparative complete genome sequence analysis of the amino acid replacements responsible for the thermostability of Corynebacterium efficiens. Genome Res 13: 1572–1579. doi: 10.1101/gr.1285603
  72. 72. Schröder J, Maus I, Meyer K, Wördemann S, Blom J, et al. (2012) Complete genome sequence, lifestyle, and multi-drug resistance of the human pathogen Corynebacterium resistens DSM 45100 isolated from blood samples of a leukemia patient. BMC Genomics 13: 141. doi: 10.1186/1471-2164-13-141
  73. 73. Tauch A, Trost E, Tilker A, Ludewig U, Schneiker S, et al. (2008) The lifestyle of Corynebacterium urealyticum derived from its complete genome sequence established by pyrosequencing. J Biotechnol 136: 11–21. doi: 10.1016/j.jbiotec.2008.02.009
  74. 74. Schröder J, Maus I, Trost E, Tauch A (2011) Complete genome sequence of Corynebacterium variabile DSM 44702 isolated from the surface of smear-ripened cheeses and insights into cheese ripening and flavor generation. BMC Genomics 12: 545. doi: 10.1186/1471-2164-12-545
  75. 75. Trost E, Götker S, Schneider J, Schneiker-Bekel S, Szczepanowski R, et al. (2010) Complete genome sequence and lifestyle of black-pigmented Corynebacterium aurimucosum ATCC 700975 (formerly C. nigricans CN-1) isolated from a vaginal swab of a woman with spontaneous abortion. BMC Genomics 11: 91. doi: 10.1186/1471-2164-11-91
  76. 76. Schmidt H, Hensel M (2004) Pathogenicity islands in bacterial pathogenesis. Clin Microbiol Rev 17: 14–56. doi: 10.1128/cmr.17.1.14-56.2004
  77. 77. Karaolis DK, Johnson JA, Bailey CC, Boedeker EC, Kaper JB, et al. (1998) A Vibrio cholerae pathogenicity island associated with epidemic and pandemic strains. Proc Natl Acad Sci U S A 95: 3134–3139. doi: 10.1073/pnas.95.6.3134
  78. 78. Oram DM, Avdalovic A, Holmes RK (2002) Construction and characterization of transposon insertion mutations in Corynebacterium diphtheriae that affect expression of the diphtheria toxin repressor (DtxR). J Bacteriol 184: 5723–5732. doi: 10.1128/jb.184.20.5723-5732.2002
  79. 79. Nakao H, Pruckler JM, Mazurova IK, Narvskaia OV, Glushkevich T, et al. (1996) Heterogeneity of diphtheria toxin gene, tox, and its regulatory element, dtxR, in Corynebacterium diphtheriae strains causing epidemic diphtheria in Russia and Ukraine. J Clin Microbiol 34: 1711–1716. doi: 10.1016/s0923-2508(97)81899-2
  80. 80. Hadfield TL, McEvoy P, Polotsky Y, Tzinserling VA, Yakovlev AA (2000) The pathology of diphtheria. J Infect Dis (Suppl 1): S116–20.
  81. 81. Murphy JR (2011) Mechanism of Diphtheria Toxin Catalytic Domain Delivery to the Eukaryotic Cell Cytosol and the Cellular Factors that Directly Participate in the Process. Toxins (Basel) 3: 294–308. doi: 10.3390/toxins3030294
  82. 82. Holmes RK (2000) Biology and molecular epidemiology of diphtheria toxin and the tox gene. J Infect Dis 181 Suppl 1S156–67. doi: 10.1086/315554
  83. 83. Sekizuka T, Yamamoto A, Komiya T, Kenri T, Takeuchi F, et al. (2012) Corynebacterium ulcerans 0102 carries the gene encoding diphtheria toxin on a prophage different from the C. diphtheriae NCTC 13129 prophage. BMC Microbiol 12: 72. doi: 10.1186/1471-2180-12-72
  84. 84. Sing A, Bierschenk S, Heesemann J (2005) Classical diphtheria caused by Corynebacterium ulcerans in Germany: amino acid sequence differences between diphtheria toxins from Corynebacterium diphtheriae and C. ulcerans. Clin Infect Dis 40: 325–326. doi: 10.1086/426687
  85. 85. Maximescu P, Oprişan A, Pop A, Potorac E (1974) Further studies on Corynebacterium species capable of producing diphtheria toxin (C. diphtheriae, C. ulcerans, C. ovis). J Gen Microbiol 82: 49–56. doi: 10.1099/00221287-82-1-49
  86. 86. LeMieux J, Hava DL, Basset A, Camilli A (2006) RrgA and RrgB are components of a multisubunit pilus encoded by the Streptococcus pneumoniae rlrA pathogenicity islet. Infect Immun 74: 2453–2456. doi: 10.1128/iai.74.4.2453-2456.2006
  87. 87. Trost E, Blom J, Soares SDC, Huang I, Al-Dilaimi A, et al. (2012) Pangenomic study of Corynebacterium diphtheriae that provides insights into the genomic diversity of pathogenic isolates from cases of classical diphtheria, endocarditis, and pneumonia. J Bacteriol 194: 3199–3215. doi: 10.1128/jb.00183-12
  88. 88. Khamis A, Raoult D, La Scola B (2004) rpoB gene sequencing for identification of Corynebacterium species. J Clin Microbiol 42: 3925–3931. doi: 10.1128/jcm.42.9.3925-3931.2004
  89. 89. Tauch A, Schneider J, Szczepanowski R, Tilker A, Viehoever P, et al. (2008) Ultrafast pyrosequencing of Corynebacterium kroppenstedtii DSM44385 revealed insights into the physiology of a lipophilic corynebacterium that lacks mycolic acids. J Biotechnol 136: 22–30. doi: 10.1016/j.jbiotec.2008.03.004
  90. 90. Collins MD, Falsen E, Akervall E, Sjöden B, Alvarez A (1998) Corynebacterium kroppenstedtii sp. nov., a novel Corynebacterium that does not contain mycolic acids. Int J Syst Bacteriol 48 Pt 4: 1449–1454. doi: 10.1099/00207713-48-4-1449
  91. 91. Paviour S, Musaad S, Roberts S, Taylor G, Taylor S, et al. (2002) Corynebacterium species isolated from patients with mastitis. Clin Infect Dis 35: 1434–1440. doi: 10.1086/344463
  92. 92. Bolt F (2009) The population structure of the Corynebacterium diphtheriae group. University of Warwick. PhD thesis. Available: http://wrap.warwick.ac.uk/1759/. Accessed 26 November 2012.
  93. 93. Songer JG, Beckenbach K, Marshall MM, Olson GB, Kelley L (1988) Biochemical and genetic characterization of Corynebacterium pseudotuberculosis. Am J Vet Res 49: 223–226.
  94. 94. Sutherland SS, Hart RA, Buller NB (1993) Ribotype analysis of Corynebacterium pseudotuberculosis isolates from sheep and goats. Aust Vet J 70: 454–456. doi: 10.1111/j.1751-0813.1993.tb00851.x
  95. 95. Halachev MR, Loman NJ, Pallen MJ (2011) Calculating orthologs in bacteria and Archaea: a divide and conquer approach. PLoS One 6: e28388. doi: 10.1371/journal.pone.0028388
  96. 96. Medini D, Donati C, Tettelin H, Masignani V, Rappuoli R (2005) The microbial pan-genome. Curr Opin Genet Dev 15: 589–594. doi: 10.1016/j.gde.2005.09.006
  97. 97. Kittichotirat W, Bumgarner RE, Asikainen S, Chen C (2011) Identification of the pangenome and its components in 14 distinct Aggregatibacter actinomycetemcomitans strains by comparative genomic analysis. PLoS One 6: e22420. doi: 10.1371/journal.pone.0022420
  98. 98. Hsiao WWL, Ung K, Aeschliman D, Bryan J, Finlay BB, et al. (2005) Evidence of a large novel gene pool associated with prokaryotic genomic islands. PLoS Genet 1: e62. doi: 10.1371/journal.pgen.0010062.eor
  99. 99. Ton-That H, Schneewind O (2004) Assembly of pili in Gram-positive bacteria. Trends Microbiol 12: 228–234. doi: 10.1016/j.tim.2004.03.004
  100. 100. Ton-That H, Marraffini LA, Schneewind O (2004) Sortases and pilin elements involved in pilus assembly of Corynebacterium diphtheriae. Mol Microbiol 53: 251–261. doi: 10.1111/j.1365-2958.2004.04117.x
  101. 101. Mandlik A, Swierczynski A, Das A, Ton-That H (2008) Pili in Gram-positive bacteria: assembly, involvement in colonization and biofilm development. Trends Microbiol 16: 33–40. doi: 10.1016/j.tim.2007.10.010
  102. 102. Ton-That H, Marraffini LA, Schneewind O (2004) Protein sorting to the cell wall envelope of Gram-positive bacteria. Biochim Biophys Acta 1694: 269–278. doi: 10.1016/j.bbamcr.2004.04.014
  103. 103. Ton-That H, Schneewind O (2003) Assembly of pili on the surface of Corynebacterium diphtheriae. Mol Microbiol 50: 1429–1438. doi: 10.1046/j.1365-2958.2003.03782.x
  104. 104. Hirata Jr R, Pereira GA, Filardy AA, Gomes DLR, Damasco PV, et al. (2008) Potential pathogenic role of aggregative-adhering Corynebacterium diphtheriae of different clonal groups in endocarditis. Braz J Med Biol Res 41: 986–991. doi: 10.1590/s0100-879x2008001100007
  105. 105. Hirata RJ, Souza SMS, Rocha-de-Souza CM, Andrade AFB, Monteiro-Leal LH, et al. (2004) Patterns of adherence to HEp-2 cells and actin polymerisation by toxigenic Corynebacterium diphtheriae strains. Microb Pathog 36: 125–130. doi: 10.1016/j.micpath.2003.10.002
  106. 106. Mandlik A, Swierczynski A, Das A, Ton-That H (2007) Corynebacterium diphtheriae employs specific minor pilins to target human pharyngeal epithelial cells. Mol Microbiol 64: 111–124. doi: 10.1111/j.1365-2958.2007.05630.x
  107. 107. Zasada AA, Formińska K, Rzeczkowska M (2012) Occurence of pili genes in Corynebacterium diphtheriae strains. Med Dosw Mikrobiol 64(1): 19–27.
  108. 108. Hall K, McCluskey BJ, Cunningham W (2001) Corynebacterium pseudotuberculosis infections (Pigeon Fever) in horses in Western Colorado: An epidemiological investigation. Journal of Equine Veterinary Science 21(6): 284–286. doi: 10.1016/s0737-0806(01)70058-5