Reinfection with Streptococcus suis analysed by whole genome sequencing

Abstract A butcher with chronic dermatitis presented with a second episode of Streptococcus suis meningitis, 8 years after the first episode. To distinguish between reinfection and persistent carriage, we compared the two S. suis isolates using whole genome sequencing. We investigated whole genome sequences of the S. suis isolates by means of substitution rates and population structure of closely related strains in addition to available clinical information. Genome‐wide analyses revealed an inserted region consisting of 12 genes in the first isolate and the calculated substitution rate between the isolates suggested infections were caused by highly similar, but unrelated strains. Continuous occupational exposure likely resulted in reinfection with S. suis in a butcher.


| INTRODUC TI ON
Streptococcus suis is a Gram-positive bacterium with zoonotic potential, causing meningitis, septicaemia and arthritis (Wertheim, Nghia, Taylor, & Schultsz, 2009). Streptococcus suis is rarely found in healthy individuals, but is a commensal of pigs and is carried in the upper respiratory and the gastrointestinal tracts of up to 100% of pigs in the Netherlands.
The S. suis serotype is determined by the antigenic properties of the polysaccharide capsule. Serotype 2 is responsible for infections in pigs and causes the majority of zoonotic infections (Goyette-Desjardins, Auger, Xu, Segura, & Gottschalk, 2014). In the Netherlands, zoonotic infections with S. suis are observed predominantly in persons who have been in close contact with pigs, such as hunters, butchers and farmers (van de Beek, Spanjaard, & Gans, 2008). Despite passive surveillance through the Netherlands Reference Laboratory of Bacterial Meningitis (NRLBM), underreporting of S. suis meningitis cases still occurs (van de Beek et al., 2008;Wertheim et al., 2009). Vaccines that protect against S. suis infection are not available for human use.
We describe a middle-aged male butcher who was known to suffer from chronic dermatitis and to carry porcine MRSA since 2007.
On 19 April 2015, the patient presented with fever, headache, vomiting and diarrhoea. Streptococcus suis was cultured from cerebrospinal fluid (CSF). The isolate was submitted to the NRLBM, where it was linked to another isolate from CSF from the same patient isolated in 2007. Using whole genome sequencing, we investigated genetic differences between the 2007 isolate, 2071319, and the 2015 isolate, 2150651, potentially explaining this recurrent infection.

| RE SULTS
Part of the accessory genome of 2071319, spanning genes 937-948, encoded an integrase, replication initiator protein, transcriptional regulator, but mostly hypothetical proteins without predicted domains. The mean GC content of these contiguous genes was 30.6% and lower than the GC content of the whole genome.
Among the remaining accessory genes is the sadP gene, encoding the Streptococcal Adhesion Protein (SadP), which had less than the 95% amino acid identity (84.8%) limit as set by Roary due to the presence of two fewer repeats. Mac family proteins also showed <95% identity due to different number of repeats, and the IS110 family transposases showed overlap, but had much lower protein identity.
Other proteins did not show similarities.
Mapping of sequencing reads against the closely related ref-  Table S1). Single Locus Variants (SLV) of ST1 were included in this analysis because single SNPs in the MLST housekeeping genes may not be representative of variation across the genome. Four main clusters could be observed.
Outliers in this tree are GZ1 as well as A7 and SC070731 (both ST7 isolates) with long diverging branches (Supporting Information Figure S1). The cluster consisting of 16 isolates including the isolates 2071319 and 2150651 resulted in a core genome of 1886 genes with an accessory genome of 129 genes, which was slightly smaller than the shared core genome between isolates 2071319 and 2150651. A maximum likelihood tree was again generated to create the highest resolution among the closest related isolates (Supporting Information Figure S2). Whilst isolates 2071319 and 2150651 cluster on the same branch, they cluster amongst other CC1 isolates from the Netherlands suggesting these two isolates are not more related to each other than to the other isolates. Using hierarchical clustering with the presence-absence matrix of the pangenome in R, we compared the accessory genome content of these 16 isolates (Supporting Information Figure S3). Isolates 2071319 and 2150651 clustered among other CC1 isolates from the Netherlands, but two main clusters of the accessory genome were separated in the dendrogram due to the presence or absence of the previously mentioned insertion in 2071319, indicative of co-evolution of two CC1 subclones circulating in the Netherlands.

| D ISCUSS I ON
Our results suggest that it is unlikely that 2150651 was a descendant from 2071319 and indicate reinfection by two unrelated isolates belonging to different ST1 subclones. The isolates differed by an inserted region, with genes which were likely inserted as a whole, but the origin of this inserted sequence is not well understood. The isolates also had different SadP genes, previously characterized as an adhesin as well as a factor H binding protein which may contribute Impacts • We report the first genomic comparison between consecutive Streptococcus suis strains, isolated from a Dutch butcher with recurring meningitis.
• We conclude this reinfection was the result of similar yet unrelated strains based on genomic analyses.
• Professionals in continuous close contact with pigs should be vigilant of reinfection even after previous S. suis related illness as natural infection may not provide protection against future infections.
Current evidence suggests that carriage of S. suis by humans in general is very rare (Nghia et al, 2011). Whilst potential carriage of S. suis due to continuous professional exposure to pigs (Bonifait, Veillette, Letourneau, Grenier, & Duchaine, 2014), combined with skin lesions related to his chronic dermatitis, cannot be ruled out in this patient, the genomic analysis does not suggest that infections occurred due to long-term carriage of a single strain as the estimated substitution rate would be too high.
Reinfection with encapsulated bacteria, such as Neisseria meningitidis and Streptococcus pneumoniae, has been associated with host complement and immunoglobulins deficiencies (Lewis & Ram, 2014). Studies in C3-and C5R-deficient mice indicated an increased susceptibility to S. suis resulting in severe infection in an intranasal mouse model (Seitz et al., 2014). A case of recurrent S. suis infections was reported in a patient after splenectomy (Francois, Gissot, Ploy, & Vignon, 1998). Whilst the patient did not have a medical history suggesting immunodeficiency, he did not consent to additional investigations to confirm or reject S. suis carriage or immunodeficiency, after his recovery.
In conclusion, we identified a patient with a S. suis reinfection on the basis of whole genome sequence analysis.

ACK N OWLED G EM ENT
We acknowledge the technicians at the Netherlands Reference Laboratory of Bacterial Meningitis for technical assistance.

CO N FLI C T O F I NTE R E S T
The authors declare no conflict of interest.