Characterization of Haemophilus Parasuis Serovar 2 CL120103, a Moderately Virulent Strain in China.

Haemophilus parasuis is an important bacterium affecting pigs, causing Glässer's disease. To further characterize this species, we determined the complete genomic sequence of H. parasuis CL120103, which was isolated from diseased pigs. The strain H. parasuis CL120103 was identified as serovar 2. The size of the largest scaffold is 2,326,318 bp and contains 145 large contigs, with the N50 contig being 20,573 bp in length. The complete genome of H. parasuis CL120103 is 2,305,354 bp in length with 39.97% GC content and contains 2227 protein-coding genes, 19 ribosomal rRNA operons and 60 tRNA genes. Sequence similarity of the genome of H. parasuis CL120103 to the previously sequenced genome of H. parasuis was up to 96% and query cover to 86%. Annotation of the genome of H. parasuis CL120103 identified a number of genes encoding potential virulence factors. These virulence factors are involved in metabolism, adhesion, secretion and LPS biosynthesis. These related genes pave the way to better understand mechanisms underlying metabolic capabilities. The comprehensive genetic and phylogenetic analysis shows that H. parasuis is closely related to Actinobacillus pleuropneumoniae and provides a foundation for future experimental confirmation of the virulence and pathogen-host interactions in H. parasuis.

Introduction

Haemophilus parasuis (H. parasuis), is an important disease-causing agent in pigs. It can cause Glässer’s disease and has shown multiple clinical manifestations including severe pant, pneumonia, pleurisy, peritonitis, polyserositis, arthritis, meningitis and septicemia [1, 2, 3, 4, 5, 6, 7]. In China, Glässer’s disease can cause great economic loss ascribed to half to two-thirds of the fatal cases occurring in finishing pigs who were previously healthy [3, 6, 7]. Moreover, Glässer’s disease outbreaks are seriously damaging to pigs, both on their own or when co-infections with other swine pathogens occur in pig-breeding companies, especially large-scale pig-raising enterprises [4, 8]. Previous studies have shown there are 15 serovars of H. parasuis with differences in virulence, including highly virulent serovars 1, 5, 10 and 12-14; virulent serovars 2, 4, 8 and 15; and avirulent serovars 3, 6, 7, 9 and 11 [9], and other studies have described the prevalent serovar for controlling infection [10, 11, 12, 13, 14]. However, vaccine immunity confers only limited cross-serovar protection [15]. Thus, further characterization to accurately identify serotypes is critical for epidemiological investigations or vaccine selection studies in H. parasuis infections. In this study, the field strain CL120103 of H. parasuis was identified through bacterial cultivation, morphological observation, PCR analysis of the 16S rRNA gene sequence and biochemical identification of traits. Furthermore, the genome of H. parasuis CL120103 was sequenced and compared with serovar 5, with a focus on the investigation of potential virulence factors, antibiotic-resistance genes and pathogen-host interactions in H. parasuis.

Materials and Methods

Bacterial strain isolation, identification and DNA purification

H. parasuis strain CL120103 was isolated from heart blood, lungs, ascites, articular fluid, and brain tissue samples of diseased pigs in Fujian province. The major clinical signs of the diseased pigs were observed as severe pant, pleurisy, peritonitis, arthritis, or meningitis. The bacterium was cultured on tryptic soy agar supplemented with 5 mg/ml nicotinamide adenine dinucleotide (NAD) at 37°C and inverted for 24-72 h before picking typical colonies as observed by Gram staining microscopy and pure culture. The strain was identified by determining the biological characteristics and by 16S rRNA gene sequencing according to the previous descriptions [16]. For amplification of the 16S rRNA gene, forward primer H1, 5′-GTGATGAGGAAGGGTGGTGT-3′ and reverse primer H2, 5′-GCTTCGTCACCCTCTGTA-3′ was used in this study. The reaction conditions were as follows: 5 min at 95°C, 35 cycles of 1 min at 94°C, 1 min at 58°C and 1 min at 72°C, followed by a final extension at 72°C for 10 min. Resulting reactions were subjected to agarose gel electrophoresis. Serotypes of H. parasuis strain CL120103 were identified by agar diffusion test via each type of standard serum of each separate strain according to the previous research [9].

Bacterial genomic DNA was extracted and purified with the QIAamp DNA Mini Kit (Qiagen, Germany). The concentration of genomic DNA was measured using a Qubit 2.0 Fluorometer (Thermo Scientific, USA). Purity of the DNA samples (UV A260/A280) was assessed using a NanoDrop 2000 Spectrophotometer (Thermo Scientific, USA). The Quant-iT Picogreen dsDNA kit (Invitrogen, Shanghai, China), Nano-2000 (Thermo Scientific, Waltham, US), and gel electrophoresis were used to evaluate the quality and quantity of genomic DNA.

Ethical approval

The research related to animals use has been complied with all the relevant national regulations and institutional policies for the care and use of animals.

Genome sequencing, assembly and annotation of H. parasuis CL120103

The genome of strain CL120103 was sequenced using the PacBio RSII platform. A 20-kb DNA library was constructed according to the manufacturer’s instructions and PacBio using single-molecule real-time sequencing technology (Pacific Biosciences, Gene Denovo, Guangzhou, China) [17, 18]. The rDNA sequences were publicly accessible and those which have high similarity to various serovars of H. parasuis by the BLASTn analysis were chosen [19]. The complete genome of H. parasuis CL120103 was added to the NCBI GenBank. The NCBI non-redundant (nr) database was applied to align the amino acid sequences by BLASTp [19].

Phylogenetic analysis of H. parasuis CL120103

The complete genomes of eight members of the genus Haemophilus and four closely related bacteria from other genera were used in the phylogenetic analysis. The accession no. for all species was shown in the front of the description (CP015099.1, CP007715.1, CP011226.1, CP001091.1, CP005384.1, CP006957.1, NC_011852.1, CP009237.1, CP007471.1, CP006955.1, CP000947.1, CL120103, and CP009158.1). The sequenced draft genome of H. parasuis serovar 5 strain 29755 (GenBank accession no. NZ_ABKM00000000) was downloaded from NCBI and aligned to the complete genome of H. parasuis strain SH03 by using the BLASTN (expected threshold of 1e-5 and minimum alignment length of 91%). Orthologous genes were identified by BLASTn suite [20]. A Bayesian phylogenetic tree was reconstructed in the software MEGA 6 [21]. For comparison within the species of H. parasuis, reciprocal BLAST was performed according to the previous description and numbers of orthologs shared between them were calculated by in-house Perl scripts [19].

COG and Pfam analysis of H. parasuis CL120103

The COG annotations were verified by comparing them to the annotations of the COG members in RefSeq databases [22]. The protein domain names in the Pfam database were used to predict protein-coding sequences and protein structure domains (ftp://ftp.sanger.ac.uk/pub/databases/Pfam/Tools/) [23, 24]. Alignment length over 90% of amino acid sequences and over 20% match identity were chosen and the description of the best hit was assigned as the annotation of the predicted gene. All annotated genes were then classified based on the COG database [25] and COG classes. COG-annotated genes and Pfam-annotation of H. parasuis strain CL120103 were compared to that of H. parasuis strain SH03.

Virulence factors of pathogenic bacteria (VFDB) analysis of H. parasuis CL120103

Virulence factors database (VFDB) of pathogenic bacteria was used to analyse H. parasuis CL120103 using chlamydia and mycoplasma pathogenic factors (http://www.mgc.ac.cn/VFs/main.htm). Also, the VFDB database was applied to align the amino acid sequences by BLASTp, and amino acid sequences with alignment length over 90% and match over 20% were identified as predicted genes [19].

Virulence factors involved in pathway analysis of H. parasuis CL120103 and pathogenesis

All sequences of H. parasuis CL120103 were translated into amino acids and submitted to the KEGG database for pathway annotation (http://www.genome.jp/kaas-bin/kaas_main) [26]. All VFDB annotated and involved in the pathway were manually downloaded.

Results

Characterization and complete genome sequencing and assembly of H. parasuis CL120103

The isolated strain was cultured for 24h on an agar plate containing 0.005% NAD and 5% horse blood [27] and results showed that the bacterial colonies were translucent, moist, smooth and a single small colony was the size of the sample tip (Figure 1A). The strain did not grow on MacConkey agar plates and was Gramnegative (Figure 1B). The PCR results [16] showed that the 16S rRNA gene fragment length was 821bp, equal to the designed and expected fragment size (Figure 1C). The agar diffusion test showed that the isolated H. parasuis CL120103 was identified as serovar 2 (Figure 1D). The Haemophilus parasuis CL120103 genome was sequenced and its complete de novo assembly was achieved by way of overlap using Single Molecule Real Time [18]. A total of 150, 292 reads (481,933,423 bases) and 30,653 paired-end reads (429,578,421 bases) were generated by PacBio RS II sequencing, in which read quality was 99.796% and 99.143%, respectively. The size of the largest scaffold was 2,326,318 bp, which contained 145 large contigs and the N50 contig was 20,573 bp in length, suggesting that this raw assembly is highly continuous [7]. The complete genome of H. parasuis CL120103 was 2,305,354 bp in length with GC content of 39.97% (Figure 1E).

Figure 1

Identification, complete genome sequencing and assembly and circular representation of H. parasuis strain CL120103 genome. (A) Colony of H. parasuis; (B) Gram staining analysis of H. parasuis; (C) PCR appraisal of H. parasuis; (D) serotype classification of H. parasuis determined by agar gel immunodifusion (AGID; the red arrow means standard serotype 2); (E) circular representation of H. parasuis strain CL120103 genome. Circles range from the outermost circle to the innermost circle. The outer two circles show protein-coding genes on the forward and reverse strands in CL120103, colored according to COG categories. All genes are colored based on biological functions and different colors in the COG collection; The third circle shows the coordinates of BLAST hits of the H. parasuis CL120103 complete genome; Fourth circle, insertion sequence elements; Fifth circle, tRNA genes; Sixth circle, rRNA operons; Seventh circle, G+C content; Eighth circle, GC skew plot [(G2C)/(G+C)].

Genome annotation of H. parasuis CL120103

Based on the 2,227 predicted genes of H. parasuis CL120103, the 2,133 CDS were annotated by BLAST search from the NCBI non-redundant database (File S1). Sixty tRNA genes and 19 rRNA genes were found in the genome of H. parasuis CL120103. The same number of rRNA genes was found in the genome of strain CL120103, strain SH03, strain 19392 and strain SC1401, and the full annotation of repetitive sequences is attached as File S2a, S2b and S2c. General features of the whole genomes of H. parasuis strains are shown in Table 1.

Table 1

General features of whole genomes of H. parasuis (CL120103)

Genbank accession No.	None	CP009158.1	CP011226.1	CP007715.1	CP009237.1	CP015099.1
Strain	CL120103	SH03	GU5	19392	KL0138	SC1401
Total length (bp)	2305354 bp	2265927 bp	1624874 bp	2431533 bp	2280878 bp	2277540 bp
Genes	2227	2148	1574	2264	2175	2180
CDS	2133	1979	1422	2117	1997	2098
Pseudo Genes	54	90	101	11	101	101
Ribosome RNA	19	19	0	19	20	19
Number of tRNA	60	59	50	62	56	59
Frameshifted Gene	35	78	69	7	78	0

Phylogenetic analysis showed that H. parasuis CL120103 shares the closest evolutionary origin to strain SH03, as expected (Figure 2). Interestingly, with numbers of orthologs (Table 2), Actinobacillus pleuropneumoniae serovar 7 strain AP76 had a similar evolutionary relationship to H. parasuis strain SC1401. H. parasuis CL120103 was illustrated by COG-annotated class distribution and the top COG classes are shown in Figure 3. The majority of the genes were involved in basic cellular functions, such as general function prediction only, replication, recombination and repair, amino acid transport and metabolism, and energy production and conversion. However, 8.46% of the genes have unknown functions in the COG database. For the full COG functional annotation, refer to File S3.

Figure 2

Bayesian phylogenetic tree of H. parasuis strain CL120103 and other closely related bacteria. Phylogenetic tree was reconstructed using MEGA 6 (Tamura et al., 2013).

Table 2

Orthologs of predicted CDSs of H. parasuis CL120103 compared with complete genomes of related organisms.

	Name	No. of orthologs	% of CDS in H. parasuis
1	Homologous to H. parasuis strain SH03	1462	68.5%
2	Homologous to H. somnus strain 2336	1236	57.9%
3	Homologous to B. trehalosi USDA-ARS-USMARC-189	1226	57.5%
4	Homologous to H. influenzae strain C486	1205	56.5%
5	Homologous to H. parasuis strain KL0318	1143	53.5%
6	Homologous to H. parasuis strain SH0165	1133	53.1%
7	Homologous to M. haemolytica USDA-ARS-USMARC-184	1097	51.4%
8	Homologous to H. parasuis strain ZJ0906	1065	49.9%
9	Homologous to A. pleuropneumoniae serovar 7 AP76	1061	49.7%
10	Homologous to H. ducreyi strain GU5	1054	49.4%
11	Homologous to A. equuli subsp. equuli strain 19392	1022	47.9%
12	Homologous to H. parasuis strain SC1401	1007	47.2%

Figure 3

COG class distribution of H. parasuis CL120103 genome. The COG-annotated genes are grouped under their respective COG classes. Percentages of the top ten classes are labeled for easy reference and presentation of the gene functional-categories.

As shown in Table 3, virulence-factor annotated genes in strain CL120103 were identified, and a list of potential virulence factors was compiled, which included gene clusters such as the peptidoglycan-binding protein LysM, 3-ketoacyl-ACP reductase, opacity-associated protein (OapA), opacity-associated A LysM-like domain protein, GlcNAc transferase, capsular polysaccharide biosynthesis protein, protein WbjB, and Fnl, the iron(III) ABC transporter ATP-binding protein, glycosyl transferase 2 protein, FbpC, ADP-L-glycero-D-manno-heptose-6-epimerase (ADP-LgDmh6e), UDP-glucose 4-epimerase (UDP-g4e), polysaccharide biosynthesis family protein (pbf), and lacto-N-neotetraose biosynthesis glycosyl transferase (LgtA) (File S4). These genes are involved in bacterial metabolism, adhesion, secretion and surface lipopolysaccharide (LPS) biosynthesis (File S5). Despite the increased complexity of virulence in the mature bacterium, this work has shown that the list of virulence factors found in the genome of H. parasuis CL120103 was compared with the other serovar strains SH03, KL0138 and SC1401. A list of genes coding for enzymes involved in the pentose phosphate pathway (PPP) was identified, including heptose 1-phosphate adenyltransferase, RfaEI, prsA, rbsK, gntK, rpe-3, rpiA, talB, pgi, fba, pfkA, gnd, devB, zwf, tktA, deoC, and fbp (Figure 4A, File S6). In this study, we also identified the sedoheptulose-7-phosphate isomerase (GmhA, H. parasuis_orf001260 and H. parasuis_orf001739), which could increase the sedoheptulose-7-phosphate content to synthesize LPS (Figure 4B and File S1). As shown in Figure 4C, ATP-dependent Clp protease subunit B (ClpB) is an important component in the complex cellular process of CtrA degradation entry into S phase. Furthermore, the identified genes were also involved in energy metabolism, glucose metabolism, DNA damage repair, autophagy and protection of proteins by chaperones according to the results of the KEGG analysis (File S7). The chaperones included GroEL (molecular chaperone belonging to heat-shock protein system), MutS (DNA mismatch repair) and lysine-tRNA ligase, and might increase the virulence of H. parasuis CL120103, which should be identified in future research.

Table 3

Identification of the potential virulence factors involved in adhesion, secretion and lipopolysaccharide (LPS) biosynthesis in the H. parasuis CL120103 genome

Sequenced No.	Gene	Functional description
LPS biosynthesis
H.parasuis_org000375	GmhD	ADP-L-glycero-D-manno-heptose-6-epimerase
H.parasuis_org001728	GmhC	heptose 1-phosphate adenyltransferase; bifunctional protein RfaE, domain I; bifunctional heptose 7-phosphate kinase/heptose 1-phosphate adenyltransferase
Adhesion and secretion
H.parasuis_orf000018	aidA	pertactin family Virulence factor, outer membrane autotransporter/type V secretion pathway, adhesin AidA; IgA1 protease, Type V secretory pathway, adhesin AidA/ outer membrane autotransporter barrel domain protein/
H.parasuis_orf000080
H.parasuis_orf001940
H.parasuis_orf001745
H.parasuis_orf001750
H.parasuis_orf000066	Rlp	rare lipoA family protein
H.parasuis_orf001352	secA	preprotein translocase subunit SecA
H.parasuis_orf002200
H.parasuis_orf000468	secB	preprotein translocase subunit SecB
H.parasuis_orf000285	secE	preprotein translocase subunit SecE
H.parasuis_orf000603	secY	preprotein translocase subunit SecY
H.parasuis_orf001367	lolA	outer membrane lipocarrier protein LolA
H.parasuis_orf000143	lolB	membrane protein/ outer membrane lipoprotein LolB
H.parasuis_orf001118
H.parasuis_orf000637	IspA	signal peptidase II/ lipoprotein signal peptidase
H.parasuis_orf000663	ftsY	cell division protein FtsY
H.parasuis_orf001041	pulG	Type II secretory pathway, pseudopilin PulG
H.parasuis_orf001355	nlpE	lipoprotein copper homeostasis and adhesion, NlpE
H.parasuis_orf001452	yajC	preprotein translocase subunit YajC
H.parasuis_orf001453	sirA	sirA-like family protein
H.parasuis_orf001539	secG	preprotein translocase subunit SecG
H.parasuis_orf001857	yajC	preprotein translocase YajC subunit
H.parasuis_orf001858	SecD	preprotein translocase subunit SecD
H.parasuis_orf001859	SecF	protein-export membrane protein SecF
H.parasuis_orf001928	pilB/hofB	Tfp pilus assembly pathway, ATPase PilB/protein transporter HofB
H.parasuis_orf000937	pilW	type IV pilus biogenesis/stability protein PilW
H.parasuis_orf001929	fimB	fimbrial protein/type II secretion system protein F
H.parasuis_orf001930	pilD	type IV leader peptidase family protein/Tfp pilus assembly pathway, fimbrial leader peptidase PilD
H.parasuis_orf002180	yidC	inner-membrane protein insertion factor/membrane protein insertase/ protein translocase component YidC
H.parasuis_orf002181	tatA	protein translocase TatA
H.parasuis_orf002182	tatB	preprotein translocase subunit TatB/twin arginine-targeting protein translocase TatB
H.parasuis_orf002183	tatC	preprotein translocase subunit TatC

Figure 4

Virulence factors of pathogenic bacteria (VFDB) analysis of H. parasuis strain CL120103 genome. (A) VFDB analysis involved in the pentose phosphate pathway; (B) VFDB analysis involved in lipopolysaccharides biosynthesis; (C) VFDB analysis involved in H. parasuis maturation. In the figure, red color box shows the identified gene based on the H. parasuis strain CL120103 genome.

Discussion

In the present study, genomic analysis showed that the genome sequence is 2,305,354 bp in length with GC content of 39.97%, and contains 145 large contigs and the N50 contig. Sequence similarity to the SH0165 has been identified and results showed that the sequence similarity was up to 96% and query cover to 86% [6]. In the H. parasuis CL120103 genome, 60 tRNA genes and 19 rRNA genes were found and associated with prophage and elements targeted to tRNA-Arg-4 (File S2b). According to previous studies, phage elements as a strategy of horizontal gene transfer might play a potentially important role in genetic diversity and virulence variations in many bacteria [10, 28, 29]. However, the inheritance differences and virulence were closely related to the evolution of the bacteria. In the present study, evolutionary relationships showed that H. parasuis CL120103 displays an evolutionary relationship to A. pleuropneumoniae serovar 7 strain AP76, which was in accordance with previous findings for H. parasuis by Xu’s group [6]. Similarly, A. pleuropneumoniae is an important etiological agent in pig disease, pleuropneumonia, which is transmitted via direct contact or airborne route. This characteristic of H. parasuis and A. pleuropneumoniae may be partially explained by their common habitat in the upper respiratory tract of pigs.

With regard to H. parasuis, the facultative anaerobe possessed metabolic pathways of both fermentation and respiration for energy generation, and carbon source utilization was important to produce energy [5, 30, 31]. In this study, virulence factors genes coding for adhesins or invasins may be located on transmissible genetic elements such as transposons [32], outer membrane protein P5 and outer membrane protein A, etc. [11, 33, 34]. However, beyond that, we have identified one of the two sugar transport systems in the H. parasuis CL120103 genome encoding ATP-binding cassette (ABC) transport complexes involved in the utilization of sugars [35]. ABC transport complexes comprise the largest protein transporter super-family in all organisms. This family of genes codes for different proteins that transport molecules such as amino acids, proteins, ions, sugars, cholesterol, peptides, metabolites and toxins across extra- and intracellular membranes [35, 36, 37, 38, 39]. A previous study indicated that all ABC transporters contain two domains, the nucleotidebinding domain (NBD) and the transmembrane domain (TMDs) [40, 41]. The two domains are roughly divided into two functional areas, which appear to specialize in handling various tasks, for example, the NBD catalyzes ATP hydrolysis and the TMDs translocates substances to use energy through the membrane by conformational changes [42]. These confirmed genotypes support the previously observed biochemical patterns of carbon source utilization in H. parasuis [5, 30]. As with previous studies, fructose as an alternative start point to glycolysis was controlled by Rbsk2, which can catalyze the phosphorylation of fructose to fructose-6-phosphate [43]. Furthermore, based on PPP, sedoheptulose was produced and used in the crucial components of the biosynthesis process of the LPS, lipooligosaccharides (LOS), capsules, O-antigens, and glycan moieties of bacterial cell surface (S-layer) glycoproteins [44]. In this study, we also identified the sedoheptulose-7-phosphate isomerase (GmhA, H. parasuis_orf001260 and _orf001739), which is the first biosynthesis step of the L, D-heptose component of the LPS and responsible for catalyzing isomerization of the D-sedoheptulose 7-phosphate into D-glycero-α, β-D-manno-heptose-7-phosphate, and leading to generation of GDP-D-glycero-α-D-manno-heptose and ADP-L-glycero-β-D-manno-heptose [44]. Thus, we speculated that the PPP plays a key role in the virulence of H. parasuis.

Furthermore, the identified gene was involved in the biosynthesis process of surface lipopolysaccharide, which is generally the initiation step on the bacterial surface for bacterial infection or bacterial adhesion on cells or substrates. Previous studies also reported that LPS was an important functional component of the Gram-negative bacterial outer membrane that can mediate bacterial adhesion on substrates/cells [45, 46]. As previously reported, LPSs includes three covalently linked biochemical moieties: the core oligosaccharide, the O-polysaccharide and the lipid A [6, 11, 33, 46]. The O-polysaccharide plays a vital role in bacterial adherence, invasion and immune evasion [46]. Furthermore, a gene cluster pilABCDW coding for type IV leader peptidase/fimbrial family protein has been identified in a number of Gram-negative pathogens of the genera Haemophilus, Vibrio, Actinobacillus and others [47]. Also identified in the genome of H. parasuis CL120103 was the type IV fimbrial genes encoding the major structural unit pilB (H. parasuis_orf001928) and biogenesis/stability protein pilD (H. parasuis_orf001930) and pilDW (H. parasuis_orf001930) for mediating bacterial adherence.

In conclusion, understanding the function of the complete genome of H. parasuis strain CL120103 will facilitate the development of safe and effective vaccines via approaches focused on genomic analysis to prevent and control swine disease. In particular, our work demonstrated the crucial function of virulence factors in metabolism, adhesion, secretion and LPS synthesis. These findings underscore the significance of VFDB as a target for therapeutics. A putative communication with the pentose phosphate pathway and virulence involved in LPS synthesis is proposed and requires further experimental confirmation in H. parasuis.