Streptococcus suis Serotype 2 Capsule In Vivo.

Many Streptococcus suis isolates from porcine endocarditis in slaughterhouses have lost their capsule and are considered avirulent. However, we retrieved capsule- and virulence-recovered S. suis after in vivo passages of a nonencapsulated strain in mice, suggesting that nonencapsulated S. suis are still potentially hazardous for persons in the swine industry.

Streptococcus suis is a gram-negative bacterium that infects pigs and causes severe economic losses to the swine industry. Moreover, it causes severe disease in persons in close contact with diseased pigs or their products (). In Japan, S. suis has been frequently isolated from pigs with endocarditis in slaughterhouses; most of the isolates were expected to be sequence types (STs) that are potentially hazardous to humans (). Many isolates from porcine endocarditis lost their capsule, and all the nonencapsulated isolates analyzed had mutations in the capsular polysaccharide synthesis (cps) genes (,). The capsule of S. suis is a major virulence factor (). Although loss of the capsule gives S. suis some benefit in causing endocarditis by enhancing the ability of bacterial cells to adhere to porcine and human platelets, a major virulence determinant for infective endocarditis (), nonencapsulated S. suis are generally considered avirulent (). However, whether nonencapsulated S. suis lurking in porcine endocarditis poses a threat to persons working in the swine industry is unknown. To investigate whether nonencapsulated S. suis can restore the ability to express the capsule and become virulent again, we repeated in vitro or in vivo passages of nonencapsulated S. suis and attempted to retrieve capsule-recovered strains.

The Study

For the in vitro passages, we used 29 S. suis strains isolated from pigs with endocarditis. These isolates had the cps gene cluster of serotype 2 but had lost their capsule because of mutations in the cps genes (Table). We subcultured them twice in liquid media and separated the cells according to the buoyant density by Percoll density gradient centrifugation (Technical Appendix). Because encapsulated cells show lower density than nonencapsulated cells (,), we investigated capsular expression of S. suis cells with low density by coagglutination tests using serotype 2 antiserum (Technical Appendix). The retrieved S. suis was also used for the next subcultures. We repeated 4 cycles of this experiment (in total 8 subcultures) but obtained no encapsulated S. suis from any of the strains tested.

Table

Nonencapsulated Streptococcus suis strains isolated from pigs with endocarditis and used for in vitro passages to investigate possible capsule recovery

Strain	Affected gene(s)	Types of mutations	Affected nucleotide(s) (affected amino acid)	Reference
NL100	cps2F	Nonsense	T696G (Tyr232TERM)	(4)
NL119	cps2F	Missense	T490C (Cys164Arg)	(4)
NL122	cps2F	Missense	G52A (Gly18Ser)	(4)
NL126	cps2F	Frameshift by insertion	TCCG	(4)
NL132	cps2E	Missense	G1199A (Arg400Lys)	(4)
	cps2H	Frameshift by deletion	TA	(4)
NL143	cps2F	Missense	G493T (Asp165Tyr)	(4)
	cps2K	Insertion	AATCATTGG	(4)
	cps2R	Missense	G496A (Gly166Arg)	(4)
NL146	cps2F	Nonsense	T482A (Leu162TERM)	(4)
NL171	cps2E	Insertion	IS element: 1,619 bp	(3)
NL174	cps2H	Frameshift by deletion	A	(4)
NL175	cps2H	Frameshift by deletion	A	(4)
NL184	cps2E	Insertion	IS element: 1,115 bp	(3)
NL194	cps2E	Insertion	IS element: 1,416 bp	(3)
NL208	cps2E	Frameshift by deletion	TAAG	(4)
NL219	cps2E	Frameshift by deletion	TAAG	(4)
NL225	cps2F	Frameshift by insertion	CCAAA	(4)
NL230	cps2F	Frameshift by insertion	A	(4)
NL240	cps2E	Nonsense	C1189T (Gln397TERM)	(4)
NL245	cps2E	Frameshift by insertion	T	(4)
NL249	cps2E	Frameshift by insertion	AGCA	(4)
NL255	cps2E	Insertion	IS element: 1,619 bp	(3)
NL257	cps2E	Frameshift by insertion	ATCT	(4)
NL266	cps2E	Frameshift by deletion	A	(4)
NL278	cps2F	Missense	T259C (Ser87Pro)	(4)
NL295	cps2F	Missense	T492G (Cys164Trp)	(4)
NL303	cps2F	Deletion	81 bp	(4)
NL322	cps2B	Missense	G469A (Asp157Asn)	(4)
	cps2G	Deletion	50 bp	(4)
NL328	cps2F	Frameshift by deletion	AG	(4)
NL342	cps2E	Frameshift by deletion	TAAG	(4)
NL345	cps2H	Deletion	23 bp	(4)
	cps2N	Missense	C706T (Pro236Ser)	(4)

Although these results suggested that mutations in cps genes are not repaired easily, the conditions faced by S. suis in vivo could influence capsular expression. To investigate this possibility, we selected strain NL119 as a representative. NL119 is an ST1 strain, one of the types hazardous to humans, but one that has lost the capsule because of a point mutation that occurred at nt 490 (T490C, Cys164Arg) of a glycosyltransferase gene (cps2F) (Table; Figure 1, panel A) (). We inoculated groups of 5 mice with 5 × 108 CFU of NL119 (online Technical Appendix). Bacteria persistent in mice were retrieved 36 h after infection from the blood, in which capsular expression works favorably for survival. We investigated capsular expression of the retrieved NL119 by coagglutination tests and used the colony giving the strongest reaction within 30 s for the subsequent in vivo passage.

Figure 1

Capsule recovery of Streptococcus suis strain NL119 in vivo. A) The genetic organization of the S. suis serotype 2 capsular polysaccharide synthesis (cps) gene cluster and mutations observed in isolate NL119 and strains retrieved from NL119-infected mice after each in vivo passage (NL119 P1–P4; DDBJ/EMBL/GenBank accession nos. LC147077, LC147078, LC147079, LC147080, and LC077855, respectively). Gray arrows indicate genes putatively involved in capsule synthesis; open arrows indicate genes with unknown functions; numbers indicate nucleotide positions in cps2F. NL119 lost the ability to synthesize the capsule because of a missense mutation at nt 490 (T490C, Cys164Arg) of cps2F (). B, C) NL119 P1 and P2 retrieved from mice after the first and second in vivo passages, remained nonencapsulated, and their cps2F sequences were identical to that of NL119. In NL119 P3 and P4 retrieved after the third and fourth passages, a further missense mutation at nt 491 (G491C, Arg164Pro) of cps2F restored the function of the gene, resulting in capsule recovery of the strains. Dot-ELISA of NL119 and strains retrieved from NL119-infected mice after each in vivo passage (NL119 P1–P4) using monoclonal antibody Z3 (B) and polyclonal anti–S. suis serotype 2 serum adsorbed with NL119 (C). Monoclonal antibody Z3 specifically recognizes the sialic acid moiety of the S. suis serotype 2 capsule.

As expected, the coagglutination test of the parental strain NL119 showed a negative result. Similarly, NL119 after the first and second passages (NL119 P1 and P2, respectively) reacted weakly, comparable to those of the parental strain, suggesting poor encapsulation. Meanwhile, NL119 after the third and fourth passages (NL119 P3 and P4, respectively) reacted strongly, suggesting recovery of the capsule. To confirm this finding, we further analyzed formalin-killed bacteria by dot-ELISA using monoclonal antibody Z3, which reacts with the sialic acid moiety of the serotype 2 capsule (), and an anti–S. suis serotype 2 serum adsorbed with parental strain NL119 to select the capsule-specific antibodies (Technical Appendix). In accordance with the coagglutination test, NL119 P1 and P2 gave weak reactions similar to those of NL119, whereas strong signals were detected in NL119 P3 and P4 with both the monoclonal antibody and serum (Figure 1, panels B, C). Because NL119 P1–P4 were also ST1 as determined by multilocus sequence typing, these results suggested that NL119 had recovered the capsule during passages in animals.

To find mutations that had contributed to the capsule recovery, we sequenced the cps2F gene of NL119 P1–P4. Although the cytosine residue at nt 490 was not changed in comparison with the parental strain, we found a further missense mutation at nt 491 (G491C, Arg164Pro) of the cps2F gene in NL119 P3 and P4 (Figure 1, panel A). To investigate whether this mutation was involved in the capsule recovery, we cloned cps2F of NL119 P4 into a gene expression vector pMX1 () and introduced it into the parental strain NL119 (Technical Appendix). A coagglutination test using serotype 2 antiserum showed positive reactions in all transformants tested, demonstrating that the further missense mutation restored the function of cps2F, resulting in capsule recovery of NL119 P3 and P4, although how the Cps2F function was recovered by the amino acid substitution is unknown. Isolation of the capsule-recovered strains in vivo could have been the consequence of selection of encapsulated cells, which were already present as a subpopulation in the original nonencapsulated NL119 population, by resisting host immunity including phagocytosis. However, because NL119 was a nonencapsulated strain originally recovered from a single colony and well-isolated by repeated passages in vitro, and no encapsulated subpopulation was ever retrieved in vitro by the selection using Percoll density gradient centrifugation, the most plausible hypothesis would be that the capsule-recovered S. suis was generated in vivo.

To evaluate whether the capsule-recovered S. suis isolate also recovered its virulence, we infected mice with either NL119 or NL119 P4 (online Technical Appendix). Rates of death differed significantly (p<0.05): 50% death in the NL119 P4-infected mice 14 days after infection, compared with 0% for the nonencapsulated NL119 (Figure 2, panel A). Recovery of the capsule also significantly increased its survival in blood 24 h after infection (p<0.05). All but 1 surviving NL119 P4-infected mice had significant blood bacterial titers (>5 × 103 CFU/mL; geometric mean 104 CFU/mL). In contrast, except for 1 mouse, all mice infected with NL119 had blood bacterial titers <104 CFU/mL (geometric mean 102 CFU/mL) (Figure 2, panel B).

Figure 2

Virulence of nonencapsulated Streptococcus suis strain NL119 and capsule-recovered NL119 P4 in mice. A) Survival of C57BL/6 mice (n = 10 mice per strain; until 14 days after infection) inoculated intraperitoneally with 5 × 107 CFU of either NL119 or NL119 P4. B) Blood bacterial burden at 24 h after infection. Data of individual mice are presented as log10 CFU/mL with the geometric mean. Asterisks indicate a significant difference between NL119 and NL119 P4 (p<0.05).

Conclusions

Although capsule loss might contribute to S. suis infection by enhancing bacterial adherence to host cells and biofilm formation (,–), capsule loss makes S. suis cells susceptible to phagocytosis; therefore, the virulence of nonencapsulated mutants was attenuated when evaluated in animal models (). In accordance with previous studies, nonencapsulated NL119 was avirulent. However, NL119 P4, which recovered its capsule in vivo, also recovered virulence. Because various mutations in cps genes, including large deletions and insertions, cause capsule loss in S. suis (,), not all mutations will be repaired like NL119. However, our results demonstrated the presence of a nonencapsulated mutant, which can recover the capsule and virulence in vivo. Hence, nonencapsulated S. suis strains can cause severe diseases to the next hosts by recovering the capsule, which indicates that some nonencapsulated S. suis lurking in pigs with endocarditis are still potentially hazardous to persons handling such pigs and their products. Further investigations using a variety of naturally occurring and laboratory-derived mutants are needed for a comprehensive understanding of the biological significance and mechanisms of this phenomenon.

Technical Appendix

Expanded methods for study of possible capsule recovery of nonencapsulated Streptococcus suis.