Complete genome sequences of Clostridium perfringens Del1 strain isolated from chickens affected by necrotic enteritis

A virulent CP Del1 strain (BioSample accession: SAMN06252079) was originally isolated from the intestines of a chicken on an NE-affected farm in Delaware, USA in 2009 and this strain was used to induce experimental NE in broiler chickens [3]. Bacteria were cultured anaerobically for 16 h and pellets were collected for DNA extraction using the cetyl trimethylammonium bromide method [9].

The complete genome sequence of CP Del1 was determined with the PacBio RS II platform (Pacific Biosciences, Menlo Park, CA, USA), performed and assembled at Genomic Resource Center, University of Maryland (Baltimore, MD, USA). A 20-kb DNA library was constructed using SMRTbell Template Preparation kit 1.0 according to the manufacturer’s protocol and sequenced using single-molecule real-time (SMRT) sequencing technology with P6-C4 chemistry and 1 RS II SMRT cell, and the data were assembled de novo using the hierarchical genome assembly process (HGAP). Since PacBio sequencing has a higher single-pass error rate (15%) than HiSeq platform does (1.5%) [10], the assembled whole-genome shotgun sequences were verified with Illumina HiSeq 4000 (San Diego, CA, USA) performed by Novogene Inc. (Sacramento, CA, USA). Sequencing libraries were generated using NEBNex DNA Library Prep Reagent Set (New England Biolabs Inc, Ipswich, MA, USA) and index codes were added to each sample. The clustering of the index-coded samples was performed on a cBot Cluster Generation System using HiSeq 3000/4000 PE Cluster Kit. After cluster generation, the libraries were sequenced on a HiSeq 4000 platform into paired-end 150-bp short reads.

Annotation was carried out by the NCBI Prokaryotic Genome Annotation Pipeline (released 2013). Information about this annotation pipeline can be found through the following web link: https://​www.​ncbi.​nlm.​nih.​gov/​genome/​annotation_​prok.

Genomic DNA was extracted from pure cultures of a single bacterial colony of Del1. The bacterium was identified as CP by biochemical identification kit and PCR product sequencing targeting 16S rRNA and necrotic enteritis B-like toxin (netB) gene, as described previously [3]. Potential contamination of the genomic libraries by other microorganisms was evaluated using a BLAST search against the non-redundant database. A5-miseq includes a quality-checking step that detects putative mis-assemblies by identifying clusters of read pairs that map to disjointed locations in the assembled genome. This method did not detect any putative mis-assemblies. The internal sequencing errors resulting from Pacbio sequencing were corrected and verified by Illumina sequencing.

The complete genome of CP Del1 was composed of a circular chromosome (Fig. 1a) and 4 plasmids (Fig. 1b). The genome had a chromosome of 3,559,163 bp long with a GC content of 28.3%, 3556 genes (chromosome and plasmids), 3361 coding DNA sequences (CDSs) (chromosome), 30 rRNA genes (10 5S, 10 16S, and 10 23S), 94 tRNA genes and 4 non-coding RNAs (Table 1). The plasmids, designated as pDel1_p1, pDel1_p2, pDel1_p3 and pDel1_p4, were 82,596, 69,827, 49,582 and 49,728 bp long, respectively. The genomic annotation data for CP Del1 and 3 other CP reference strains, CP4, JP55 and ATCC13124, are summarized in Table 1. One striking difference was that CP4 had only 38 RNA genes while the 3 other CP strains contained 128 genes. This discrepancy may be accounted for by incomplete gene sequences of CP4 strain.

The Clusters of Orthologous Group (COG)-associated functional genes for CP Del1 and other bacterial starins are shown in Additional file 1: Table S1. Among the various COG categories in CP Del1, 6 comprised the largest proportions (each ≥ 5% of the total COG classifications) in order from largest to least: R [general function prediction only, 319 open reading frame (ORFs), 9.76%], S (function unknown, 283 ORFs, 8.65%), G (carbohydrate transport and metabolism, 203 ORFs, 6.21%), K (transcription, 202 ORFs, 6.18%), E (amino acid transport and metabolism, 181 ORFs, 5.54%), and J (translation, ribosomal structure and biogenesis, 169 ORFs, 5.17%). Generally, all CP strains (Del1, JP55) exhibited comparable ranges in proportions of COG classifications.

To date, 52 complete genome sequences of CP strains and their annotation data are available in the GenBank database. To compare these genomes, whole genome DNA-sequence-based average nucleotide index (ANI) analysis was performed and phylogenetic trees were drawn based on genomic sequence analysis of these 52 CP genomes (Fig. 2). The dendrogram illustrates that CP Del1 strain was the most closely related to CP4 strain, and clustered with JP55, MJR7757A and 5 other JFP strains. CP4 strain, whose genomic sequence has been characterized (9), is a pathogen that causes NE [11], and contains 3 highly conserved NE-associated loci that are designated NELoc-1 (42 kb), NELoc-2 (11.2 kb) and NELoc-3 (5.6 kb), with netB residing on an ~ 42-kb plasmid-encoded pathogenicity locus (NELoc-1) [12]. In CP Del1, netB gene resides in plasmid pDel1_1.

We also found numerous important genes involved in pathogenesis and virulence in the CP Del1 chromosome and plasmids (Additional file 1: Table S2). CP Del1 and CP4 strains share similar genes, with more genes absent in the strain ATCC13124 genome. Besides alpha-toxin phospholipase C (cpa plc), CP Del1 genome also had genes encoding necrotic enteritis B-like toxin (netB), Beta 2 toxin (cpb2), mu-toxin (nagH-K), sialidase (nanH, nanI and nanJ), collagenase (peptidase_U32), and perfringolysin O (thera toxin, pfo). Alpha toxin (cpa-encoded) is the most toxic extracellular enzyme produced by CP type A. It is present in all types and essential for virulence, and its core protein, phospholipase C, hydrolyzes both important constituents of eukaryotic cell membranes: phosphatidylcholine and sphingomyelin [13, 14]. NetB is a pore-forming toxin critical for induction of NE in chickens with predisposing factors, such as preceding coccidial infection and dietary manipulation [15, 16]. Sialidase cleaves terminal sialic acid residues in the alpha configuration linked to oligosaccharide chains present on proteins and lipids and functions as a nutritional factor to support bacterial growth and as a virulence factor during bacterial pathogenesis [1].

Studies on the genetic diversity of C. perfringens in mild NE-affected chickens indicated that more than 90% of all isolates from NE-specific organ lesions carried netB gene [17]. However, immunization with vaccines containing recombinant NetB toxin only partially protected progeny chickens from necrotic enteritis challenges [18]. Other virulence factors may additively contribute to the CP pathogenesis in NE induction. Global transcriptomic analysis performed on ligated intestinal loops in chickens following infection with a netB⁺ CP strain revealed that numerous virulence factors were significantly expressed in vivo, such as cpa, netB and sialidase genes [19]. Del1 strain is a virulent strain isolated from a field disease outbreak to be used as a challenge strain. Compared to CP4 strain, this Del1 strain contains some unique genes expressing adhensin, pilin of the toxin- coregulated-pilus (tcp cluster C–J), and programmed cell death toxin (PemK) which may be involved in multitude of pathogenesis functions, including adherence to eukaryotic cells, DNA uptake, and protein/toxin secretion, and therefore enhancing the pathogenesis and favoring the colonization and infection of C. perfringens as a gastrointestinal pathogen in food animals. Sequencing analysis also indicated that Del1 strain did not contain genes expressing Tpel toxin and surface protein CPE1231 as the CP4 strain had. Previously, it was reported that Tpel protein enhanced bacterial virulence and severity of NE [16]. Identifying bacterial important genes that encode toxin and virulence factors in genomics and transcriptomics approaches is critical for recombinant vaccine development, for example, cpa, netB, sialidase B, tpel, and PemK. In general, these genes contribute to toxin generation and bacterial colonization and growth. Strain Del1 is classified as Type A due to absence of beta, iota and epsilon toxins based on the definition of the presence of toxin genes in toxinotyping [1].

By screening the genome and plasmid sequences against the GenBank library, we identified 6 antibiotic-resistance genes/clusters in CP Del1 genomes and plasmids. Specially these genes conferred resistance to tetracycline (total 3 genes: 1 gene encoding tetracycline ribosomal protection with Accession No: YP_001967743, 2 genes encoding tetracycline efflux pumps: BAB71966 and YP_001966009), 1 gene for class B beta-lactams (Subclass B1_136), as well as 1 gene for other antibiotics (YP_698526). Finally, another 1 gene encoded a major facilitator superfamily that was involved with antibiotic efflux pump (P96712), by which membrane transport proteins facilitated movement of small antibiotics across the cell membrane [20]. CP4 was also found to confer resistant to tetracycline [21]. Del1 strain was found in its susceptibility test to be resistant to tetracycline hydrochloride with minimal inhibitory concentrations of more than 256 µg/mL.

Bacteriophage DNAs serve as mobile genetic elements which confer the antibiotic-resistance genes or virulence determinants by transduction, and ultimately contribute to genome diversification of bacterial hosts, with strain to strain variations in bacteriophage repertoire [22]. However, host specificity has routine been observed relative to the bacteriophages isolated from various CP strains. This is most likely due to evolution of the receptor in the bacterial hosts and anti-receptor molecules in bacteriophages (3). Using a previously described bioinformatics platform [23], we have determined specific genes in two regions (1224604–1235574 and 2737554–2740876) in Del1 genome which have high sequence similarities to Clostridium phage vB_CpeS-CP51, Clostridium phage PhiS63, and bacteriophage phi3626. Further, these regions also have low homology between them, indicating the prophages in Del1 strain are also strain specific. Sequence annotation predicts that majority of the genes in these two prophages-like areas encode hypothetical proteins.