Whole-Genome-Sequence of Streptococcus Suis LSM178 with a Novel ST1005 Characterized the Hyper Pathogenicity in Human Infection

Streptococcus suis ( S. suis ) has been well-recognized as a zoonotic pathogen 2 worldwide gearing up a great risk to the public health. In this study, an S. suis 3 LSM178 strain with serotype 2 and novel multi-locus sequence type of 1005, isolating 4 from a patient, was interpreted for the pathogenicity by its genetic information. 5 LSM178 was more efficiently invasive to Caco-2 cells than SC19 and P1/7. 6 Phylogenetic analysis showed that LSM178 clustered with highly virulent strains 7 including all human strains and epidemic strains. These serotype 2 S. suis from China 8 shared exclusively the typical virulence characteristics including the maximum (95/96) 9 virulent factors and type I-89 K Pathogenicity Island. Further, groups of genes were 10 identified to distinguish these highly virulent strains from other generally virulent 11 strains, emphasizing the key roles of genes modeling transcription, cell barrier, 12 replication, recombination and repair on the high pathogenicity for highly virulent 13 strains. Additionally, LSM178 contains a novel prophage conducive potentially to 14 pathogenicity . These characters would contribute to deeply studying the pathogenic 15 mechanism and virulence drift of human pathogenic S. suis . 16

Streptococcus suis (S. suis) is one of the most important swine pathogens leading to 22 severe economic losses to the porcine industry worldwide. However, S. suis has 23 emerged as an great zoonotic agent, causing fever, septicemia, meningitis, arthritis 24 and a variety of other symptoms in human. Since the first case of human S. suis 25 infection reported in 1968 [1] , it had spread in more than 30 countries and/or regions, 26 particularly the southeast Asian countries where the pathogen represents a significant 27 public health concern [2,3] . Seriously, 2 and 4 outbreaks have occurred in China 28 (Sichuan in 2005 and Jiangsu in 1998) [4] and in Thailand (Phayao in 2007, Chiang 29 Mai and Lamphoon in 2008, Phetchabul in 2010 and Uttaradit in 2019) [5] respectively. 30 Typing of S. suis strains is epidemiologically important to control the infection. The 31 most commonly used serotyping is not only used for identification and diagnosis of 32 clinical of S. suis isolates, but also its importance of on pathogenesis has been 33 suggested [5] . Of the 35 serotypes (types 1-34 and 1/2) originally identified according 34 to the antigenicity of capsular polysaccharide (CPS), six S. suis-like strains (serotypes 1 20, 22, 26, 32, 33, and 34) have been taxonomically removed from the S. suis species 2 based on phylogenetic and/or sequence analyses [6] . Additionally, new variants with 3 serotype Chz and novel cps loci were recently checked out, although their relation to 4 the virulence potential remain unclear [7,8,9,10] . The prevalence of S. suis serotypes in 5 countries and regions is different. For instance, isolates associated with pigs disease 6 were predominantly identified as serotypes 2 and 9 in Europe [11] , and serotypes 2 and 7 3 in North America [12] . However, serotype 2 is considered to be the most toxic and 8 prevalent serotype causing both pig and human infection worldwide [13,14] , although 9 other serotypes such as serotype 9 and 14 are increasingly urgent [15] . A recent study 10 reported that, upon detection of raw pork and edible pig organs collected from 88 11 sales locations in central Thailand, the positive rate of S. suis was as high as 85.23% 12 and the positive rate of serotype 2 was 17.05% [16] . 13 Besides serotyping, genetically classification by multiple sequence locus typing 14 (MLST) [17] has become increasingly important because of the higher resolution on 15 determining the strain evolution and on delineating the relationship between subtype 16 and the pathogenicity. For instance, while the serotype 2 ST1 strains present high 17 zoonotic potential worldwide, ST7 from serotype 2 and 14 is frequent to China [5,18] . 18 And, in Thai for human infections, ST104 are almost exclusively predominant in 19 serotype 2 and the main serotype 14 isolates was ST105 [5] . So far as to April 8, 2021, 20 2,808 STs have been recorded in the S. suis MLST database, showing that S. suis is 21 constantly evolving as the environment changes. Though only several STs have been 22 found to be mainly responsible for human infections, the increasing diversity brings 23 new risks and challenges, such as ST658 isolated in China [19] . 24 It is no doubt that virulence arsenal play more roles in human infection, since S. suis 25 was suggested to be a cause of community-acquired pathogen [20] . It is difficult to 26 make any clear distinction about the virulence factors belonging exclusively to pigs 27 versus humans. In a recent study, no any defined genomic differences between human 28 strains and pig strains were suggested, although human disease isolates are limited to 29 a single virulent population whose origin nevertheless coincided with the first 30 intensification of pig production [21] . Even, it couldn't definitely determine whether a 31 strain is a virulent one only by the presence of proposed virulence factors, making the 32 ambiguous definition for virulence factors [22] . Virulence factors play key roles in 33 many aspects. One reason why serotype 2 strain showed a high zoonotic potential and 34 virulence in human was attributed to its better adherence to a human intestinal 1 epithelial cell [5,23,24] . Deletion of these factors greatly attenuated virulence [25,26,27,28,29] . 2 As well, the ability to escape immune clearance is necessary for strain survival, 3 dissemination and pathogenesis. For example, CPS has been shown to enhance 4 bacterial resistance against the killing by host phagocytes [30] . Many virulence factors 5 have potential to be vaccine candidates [31] . It should be noted that a universal 6 cross-protective vaccine is highly challenging due to the diversity of S. suis. The 7 control of infection mainly depends on antibiotics therapy. And, the emergence of 8 antibiotic-resistant strains, especially multidrug-resistant strains, poses an intricate 9 problem [32,33] . 10 In this study, a human S. suis LSM178 with serotype 2 and novel ST1005 causing 11 fever, nausea and general malaise, was isolated and comprehensively assessed as a 12 hyper virulent strain based on toxicity tests and genomic analysis. and 36-40 h) in each of 3 independent repeated tests ( Fig. 1 A, as one representative). 29 The main symptoms of the death caused by LSM178 infection are seriously 30 abdominal swelling, blood spots and ecchymosis ( Fig. 1 B), which were similar with 31 those caused by SC19 and P1/7.

27
Phylogenetic tree analysis 28 A phylogenetic tree was generated based on 51,520 core-genome SNP sites using the 29 genome sequence of LSM178 and 52 published S. suis complete genomes (Fig. 4). 30 Within the 18 strains clustered with LSM178 (LSM178-branch), 7 out of 12 pig 31 isolates and 5 out of 6 human strains were from China. However, the closest 2 strains 32 (NCTC10234 and S735) were neither from China nor from human. Among the 53 33 strains, most ST1 (5/9) and ST7 (9/12) were clustered in LSM178-branch. It was 1 interesting that in LSM178-branch, the STs of human isolates (LSM178, LSM102, 2 05ZYH33, SC84, BM407 and 98HAH33 with ST1005, ST658, ST945, ST7, ST1 and 3 ST890 respectively) are more diverse than those of pig strains (just including ST1 and 4 ST7).   Table 1). 9 Since LSM178 shares these important virulence markers with epidemic strains and 10 almost all human strains, these strains were referred here as the highly virulent strains 11 (HVS) at least for serotype 2 strains from China, those are LSM178, ZY05719, 12 98HAH33, SC84, 05ZYH33, SS2-1, SC19 and LSM102. 13 14 Genomic comparative analysis 15 Checking virulence factors between strains in GZ1-branch and closely related 16 avirulent T15, T15 have 87 virulence factors which were shared by GZ1-branch 17 strains except A7, P1/7, S10 and GZ1 (Supplementary Table 1 Generally, S. suis infections in humans were restricted to workers in close contact 5 with pigs or swine byproducts. However, in southeast Asia, the bacterium has been 6 reported to affect the general population [20] . The sporadic human S. suis, particular 7 those with serotype 2, should be get attention in epidemiological monitoring because 8 of their unpredictable adaptive potential, for instance of LSM178 here, a serotype 2 9 strain with novel ST1005. LSM178 was more efficiently invasive to Caco-2 cells 10 compared with P1/7 and SC84. However, no significant virulence were detected in 11 cytotoxic activity and challenging zebrafish which has been used as a model to 12 evaluate the virulence of S. suis [35] . In deed, virulence of S. suis could not be 13 intensively evaluated and compared in other models [36] . Probably, virulent strains 14 have their own outstanding virulence aspects, which can balance the weak ones and 15 eventually lead to a similar pathogenicity at least under the inoculation. For instance, 16 P1/7 and SC84 were prominent in biofilm formation ( Supplementary Fig. 2) and 17 adhesion respectively. 18 Recent studies suggested that quinolones, beta-lactams, florfenicol and 19 trimethoprim/sulfamethoxazole could still treat well the clinic S. suis infection [37,38,39] . 20 However, the strains resistant to widely used effective beta-lactams have been 21 increasely reported [40,41] . LSM178 showed the sensitivity at least to beta-lactams and  Fig. 4). Additionally, it is interesting that the P1/7 30 without ANT(6)-Ia confered resistance against streptomycin, whereas ANT(6)-Ia 31 containing LSM178 could not [38] . However, ANT(6)-Ia of LSM178 is only 49.5% 32 identical to the functional homologue (UniProtKB -P12055 (AADK_STAAU)) from 33 Staphylococcus aureus [43] .
Several factors are deemed to be important for the pathogenesis, such as CPS, Fbps, 1 enolases, dipeptidylpeptidase DppIV and SrtA [22] , which were all found in LSM178. It 2 has been demonstrated that S. suis serotype 2 virulent strains are able to exacerbate 3 inflammatory activation scavenging bacteria. All the 15 known anti-immunity 4 factors [19] exist in LSM178 and would modulate the immune responses improving its 5 survival at the inflammation area. Although it is hardly to define a virulent strain only 6 by proposed virulence genes, many avirulent strains isolated from healthy animal are 7 found with less virulence factors (less than 87), such as WUSS351, 0061, 1081,  17 It is very clear that intermediately pathogenic strain could evolve to highly pathogenic 18 strain and then epidemic strain [3] . In LSM178-branch, all 18 strains possessed the 19 serotype 2. However, while 6 human strains have their individual STs, 12 pig strains 20 were just ST1 (4 strains) or ST7 (8 strains). It may imply that while S. suis has been 21 purified to relatively stable genotype in pig, multiple evolutionary directions are in 22 progress to be epidemic during their adaption to human, at least for serotype 2 23 strains in China. 24 Type I-89K-PAI is specific to highly pathogenic S. suis linked to Chinese epidemics [19] 25 and could horizontally transfer among strains [44] . Humans can carry avirulent S. suis  6 In published S. suis genomes, many prophages were held as remnants [46,47] . A few 7 intact prophages was described and one of them was reported to be induced to lyse S. 8 suis [48] . The two prophages in LSM178 could not be induced to form plaques on 9 several S. suis strains. The reason might be that improper induction method was 10 adopted or that the phages produced are defective on infection even for the almost 11 intact pha17802 [48] . Alternatively, the lysis spectrum of the phages is very narrow, or 12 no phages were induced at all, which could be supported by the facts that pha17801 increase the fitness of the bacteria by modulating host metabolism [48] . 29 While the core virulence genes were identified through the comparation between 30 avirulent T15 and virulent strains in GZ1-branch, they should be responsible for the 31 enhanced pathogenicity. Moreover, some genes could be used to make a distinction 32 between HVSs and other general virulent strains. Particularly, it should emphasize the 33 roles of genes with functions of transcription, cell barrier, replication, recombination 34 and repair in virulence enhancement, since their number increased obviously in HVSs 1 core genes. Thus, the importance of the mobilizable elements is beyond all doubt, 2 because it is the carrier responsible for genetic differences [50] . In fact, GIs make up a 3 considerable part of the genome for LSM178 (10.4%). And, higher proportion of 4 virulence-enhancing genes are dispersed in GIs, for example that GIs of LSM178 are 5 colonized with 3.7% of core general virulence genes but with 23.5% of core high 6 virulence genes. However, there are less GIs in LSM178 (number of 23 and total 7 length of 21, 5087 bp) than T15 (29 and 31, 4245 bp) ( Supplementary Fig. 5 B). 8 These suggested that fusion of specific PAIs increased the virulence. The Biofilm formation assay 15 Strains (20 μL, 5.0×10 7 CFU/ml) were inoculated into 2 mL TSB medium 16 (containing 10% fresh FBS) and cultured in a 24-well cell plate at 37℃ for 3 days, 17 and the un-inoculated medium was used as the control. After that, the strains were were inoculated with 500 μL bacterial suspension (1 × 10 6 CFU). After washing to 3 remove unadhesive strain, cells were then treated by trypsin digestion for 2 hours. In 4 invasion, extracellular bacteria were treated with gentamicin (100 μg/ml) and 5 penicillin G (5 μg/ml) before trypsin treatment. The digested cells were lysed using 6 1% saponin and the lysis was inoculated on THB plate. The rate of adhesion (Ra) and 7 invasion (Ri) was expressed as (CFU determined from plate / CFU original inoculum) × 100%. The 8 relative invasion rate was expressed as Ri strain / Ri P1/7 × 100%.  Table 1) and used to detect the potential virulence genes 26 in genome. Antibiotic resistance genes were predicted with the comprehensive 27 antibiotic resistance database (CARD) with default settings [54] . The ST was 28 determined using MLST typing scheme (https://pubmlst.org/ssuis/). Serotyping 29 strategy was executed based on the homology and coverage (80%, evalue≤1e-10) 30 between WZY amino acid sequence of the LSM178 and 33 standard strains with 31 known serotypes. The sequence of CpsK was used to discriminate the serotype 2 32 (W161) from 1/2 (C161) for all strains with serotype 2 or 1/2 [55,56] . ST complexes 33 were analysed by goeBURST [54] program (http://goeburst.phyloviz.net). Prediction of 1 gene islands (GI) and prophages were performed using IslandViewer 4 [57] and 2 PHAST [58] respectively. Clustered Regularly Interspaced Short Palindromic Repeats 3 (CRISPRs) were predicted by CRISPR recognition tool (CRT) [59] . Open reading 4 frames (ORFs), tRNA and rRNA were predicted with Prokka. The proteins in 5 prophage and PAI were integrally annotated by databases of NR, eggNOG, KEGG, 6 Swiss-Prot and GO. Genome maping with informations was generated by CGView [60] . 7 To identify the unique regions in the genome of LSM178, the whole-genome 8 sequence was used as a reference to compared with that of the other 52 strains using 9 BLAST Ring Image Generator (BRIG) [ Table   20 2). The chromosomal sequences were aligned using Parsnp program generating   15 The authors declare no conflicts of interest and no competing financial interests.       The circular diagram of the LSM178 genome. From inside to outside, the rst circle, the scale of genome;

Conflict of Interest
the second circle, GC skew; the third circle, GC content; the fourth and seventh circles, the COG category of coding sequence (CDS) on two strands; the fth and sixth circles, the position of CDS, tRNA and rRNA on two strands. Phylogenetic tree of 53 S. suis based on core genome SNP.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.