The invasome of Salmonella Dublin as revealed by whole genome sequencing

Salmonellosis is one of the most common foodborne diseases worldwide. Salmonella enterica (S. enterica) causes a huge global burden of morbidity and mortality in humans. It is estimated that Salmonella serovars responsible for typhoid fever kill over 250,000 humans each year [1] while non-typhoidal Salmonella (NTS) serovars that are responsible for diarrhoeal illness cause 155,000 deaths annually [2]. Moreover, NTS might have adapted to cause invasive disease and systemic infections in humans; children, the elderly and immunocompromised and it is estimated that 680,000 people die every year as a result of infection by invasive NTS (iNTS) [3].

The most predominant iNTS serovars associated with systemic illness in humans are S. Typhimurium, S. Choleraesuis and S. Dublin [4, 5]. S. Dublin is specifically adapted to cattle [6] thus, people can be infected through contact with infected animals or consumption of contaminated food including raw milk and raw-milk cheese. Outbreaks of human S. Dublin infection have been reported in some European countries including Ireland and France [7, 8].

Several virulence factors have been identified within the human adapted serovars Typhi and Paratyphi [9, 10]. On the other hand, little is known about the molecular basis of virulence in iNTS more specifically S. Dublin. Our understanding of the genetic basis of invasiveness in S. Dublin is skewed by the fact that most studies have focused on the most common iNTS; S. Typhimurium in particular the highly invasive multidrug-resistant (MDR) S. Typhimurium of a distinct Multilocus Sequence Type (MLST), ST313 that has been associated with severe infections and deaths in humans in sub-Saharan Africa [11, 12].

The aim of this study is to characterise the invasome of S. Dublin and the virulence factors that might enable the bacteria to invade blood causing systemic illness. We therefore applied whole genome sequencing (WGS) to a set of S. Dublin isolates from different countries all over the world.

Human infection with iNTS represents a significant public health problem. A high proportion of S. Dublin cases in humans are characterized by bloodstream infection. Furthermore, the antibiotic resistance is increasing in S. Dublin as we detected two clinical isolates from blood that are resistant to multiple antibiotics.

The invasiveness of S. Dublin may be related to the expression of the Vi (virulence) antigen which is a capsular polysaccharide antigen commonly found in the human adapted S. Typhi and S. Paratyphi C. It has however also been detected in S. Dublin [21].

The Vi proteins are encoded within the viaB (Vi capsule biosynthesis) locus on SPI-7 [22, 23]. The viaB locus contains both Vi antigen biosynthetic genes (tviB, tviC, tviD and tviE) and export genes (vexA, vexB, vexC, vexD and vexE). The expression of Vi antigen is controlled by the rcsB-rcsC and ompR-envZ two-component regulator systems which lie outside the SPI-7 they however, interact with tviA; the first gene of the viaB gene cluster and regions upstream of the tviA promoter [24].

Experimental studies showed that Vi antigen is antiopsonic and antiphagocytic as it reduces the level of S. Typhi-induced tumor necrosis factor alpha (TNFα) by human macrophages and it also increases the resistance of S. Typhi to oxidative stress [25, 26].

In this study, we screened all S. Dublin isolates from France for the viaB operon and we found that Vi-coding genes were absent from all isolates except three isolates including the reference veterinary isolate; SARB13 and two clinical isolates; 04.4663 from blood and 201,208,251 from pus. We therefore conclude that it is unlikely that Vi antigen is the main virulence determinant for S. Dublin since it is absent from other invasive isolates from blood and urine.

Although Vi capsular polysaccharide antigen was not present in all invasive isolates it is possible that S. Dublin produces a unique O-antigen capsule that plays a role in increasing bacterial virulence and pathogenicity. Further in vitro and in vivo studies using O-antigen capsule deficient mutant are required to confirm this hypothesis.

Flagellum is considered as a virulence factor for motile bacteria such as Salmonella [27]. The surface domains of the flagellin protein are highly immunogenic and play an important role in triggering host innate and adaptive immune responses therefore, motile bacteria have evolved several mechanisms to overcome flagellin recognition by host receptors [28] however we detected no diversity among invasive an non-invasive S. Dublin strains in flagellar biosynthesis genes including fliC gene coding for flagellin, fliD coding for the flagellum capping protein and fliA coding for flagellum-specific sigma factor.

Although the three S. Dublin isolates; 04.4663, 201,208,251 and SARB13 are positive for the Vi antigen they lack the novel pathogenicity island ST313-GI that harbours the putative virulence gene st313-td. On the other hand, all other S. Dublin isolates harbour the st313-td gene on a degraded ST313-GI.

Although the exact role of st313-td in increasing bacterial virulence is not known experimental studies showed that st313-td is associated with the virulence and systemic infection in invasive S. Typhimurium of a distinct MLST, ST313, that has emerged recently in Africa [12]. It has been shown that virulence of S. Typhimurium in the mouse model was reduced in the absence of st313-td therefore, it is hypothesised that st313-td might help in bacterial evasion from host’s immune system through decreasing the binding of the bacteria to specific antibodies causing less uptake by macrophages or through decreasing the susceptibility to complement mediated lysis [11, 28].

We hypothesise that it is likely that st313-td plays a role in the virulence of S. Dublin hence, we detected it in all isolates (except the three isolates that are positive for Vi antigen; 04.4663, 201,208,251 and SARB13) however further experimental work is required to test its role in the virulence of S. Dublin.

There are other mobile genetic elements (MGEs) (excluding SPI-7 harbouring Vi antigen and genomic island ST313-GI encoding the putative virulence gene st313-td) that might also contribute to the virulence of S. Dublin in humans including SPI-6 and SPI-19 which harbour two different types of T6SS including T6SS_SPI-6 and T6SS_SPI-19 respectively.

It has been shown that T6SSs play a significant role in bacterial pathogenesis as they function as contractile nanomachines to puncture human cells and deliver virulence and lethal factors [29]. Experimental studies showed that T6SS_SPI-6 plays a crucial role in the invasiveness and the systemic spread of S. Typhimurium [30] while T6SS_SPI-19 has been reported to contribute to the survival of the poultry adapted serovar S. Gallinarum within infected macrophages [31].

All S. Dublin isolates harbour SPI-6 and SPI-19 that encode T6SS_SPI-6 and T6SS_SPI-19 respectively. Interestingly, SPI-6 and SPI-19 harbouring the two different T6SS have been also detected in clinical human isolates from Ireland including invasive and gastroenteritis isolates [32] suggesting that they contribute to the ability of S. Dublin to cause invasive disease in humans.

Another MGE that might contribute to the virulence and pathogenesis of Salmonella serovars including S. Dublin is the lambdoid prophage Gifsy-2 [33] as it encodes several virulence genes including sodCI, sseI and gtgE [34].

In this study, we found that all S. Dublin isolates are lysogenic for Gifsy-2 prophage and they all harbor the gene encoding Gifsy-2 prophage attachment and invasion protein suggesting that Gifsy-2 might be associated with the predisposition of S. Dublin to cause systemic illness in humans.

Plasmids are other MGEs that have been shown to play a significant role in the acquisition of virulence and antimicrobial resistance genes in several Salmonella serovars [35]. Experimental studies showed that S. Dublin plasmid encoding the virulence spv locus is involved in the dissemination and spread of bacteria to the blood [36, 37].

Interestingly, all S. Dublin isolates sequenced in this study except three clinical isolates including 93.3170 and 201,005,507 from blood and 93.1557 from stool harbour the virulence S. Dublin plasmid pCT02021853 that encodes the spv genes. Although the cattle isolate (93.1086) harbours a closely related plasmid to the virulence plasmid pCT02021853 the plasmid does not harbour the virulence spv locus. We therefore conclude that it is unlikely that the virulence plasmid is the main virulence determinant for S. Dublin and it is not critical for bacterial invasiveness since it is absent from two invasive clinical isolates from blood; 93.3170 and 05.1078.

Among the virulence factors that might contribute to the virulence of S. Dublin is γ-glutamyl transpeptidase (GGT) as it has been reported to contribute to the virulence of Helicobacter pylori [38] furthermore, it plays a significant role in inhibiting T-cell proliferation [39]. We found that the gene ggt is harboured by all S. Dublin isolates including human and animal isolates.

Another virulence gene, PagN, encodes for an outer membrane protein (PagN) that has been reported to contribute to the virulence of S. Typhimurium through mediating bacterial adhesion and invasion of mammalian cells [40, 41]. Experimental studies showed that PagN is a strong immuogen in mice and it can therefore be a potential vaccine candidate for salmonellosis [42]. Interestingly, we found that PagN is harboured by all S. Dublin isolates including invasive and non-invasive isolates suggesting its relation to bacterial virulence and invasiveness.

Although S. Dublin is adapted to cattle as a result of extensive genome decay and pseudogenes accumulation [6] S. Dublin can infect other animals [43, 44]. The phylogenetic SNP analysis showed the close relation among the veterinary isolates and human isolates as SNPs were randomly distributed around the chromosome of S. Dublin. All S. Dublin isolates were intermixed and there were no genomic differences among clinical invasive and non-invasive isolates.

We identified several virulence factors in S. Dublin isolates that form the bacterial invasome however, we could not detect any genomic markers that differentiate invasive from non-invasive disease suggesting that host factors and immune response play a significant role in the disease outcome.