So similar, yet so different: uncovering distinctive features in the genomes of Salmonella enterica serovars Typhimurium and Typhi


  • Editor: Ian Henderson

Correspondence: France Daigle, Department of Microbiology and Immunology, University of Montreal, C.P. 6128 Succursale Centre-Ville, Montréal, QC, Canada H3C 3J7. Tel.: +1514 343 7396; fax: +1514 343 5701; e-mail:


Salmonella enterica represents a major human and animal pathogen. Many S. enterica genomes have been completed and many more genome sequencing projects are underway, constituting an excellent resource for comparative genome analysis studies leading to a better understanding of bacterial evolution and pathogenesis. Salmonella enterica serovar Typhimurium and Typhi are the best-characterized serovars, with the first being involved in localized gastroenteritis in many hosts and the latter causing a systemic human-specific disease. Here, we summarize the major genetic differences between the two different serovars. We detail the divergent repertoires of the virulence factors responsible for the pathogenesis of the organisms and that ultimately result in the distinct clinical outcomes of infection. This comparative genomic overview highlights hypotheses for future investigations on S. enterica pathogenesis and the basis of host specificity.


Salmonella evolved as an intracellular pathogen after diverging from a common ancestor with Escherichia 100–150 million years ago (Doolittle et al., 1996). The nomenclature and taxonomy of Salmonella are complex, controversial, have changed over the years and are still evolving. The genus Salmonella is composed of two distinct species: Salmonella bongori, a commensal of cold-blooded animals, and Salmonella enterica (divided into six subspecies) (Le Minor et al., 1987; Reeves et al., 1989). The subspecies are classified into over 50 serogroups based on the O (somatic) antigen, and divided into >2400 serovars based on the H (flagellar) antigen. Some serovars are ubiquitous and generalists, while others are specifically adapted to a particular host. Only a small fraction of serovars are associated with human infections and the majority belong to S. enterica ssp. I. Salmonella enterica ssp. I is responsible for two types of disease in humans due to ingestion of contaminated food or water: gastroenteritis, a localized infection or enteric fever (typhoid), a severe systemic infection.

Gastroenteritis is caused mainly by S. enterica serovar Typhimurium (S. Typhimurium) and S. Enteritidis. Salmonella enterica serovar Typhimurium can colonize and infect a broad spectrum of warm- and cold-blooded hosts, belongs to serogroup B and is a prototroph (Fig. 1). Typhoid fever, a life-threatening illness that remains a global health problem, is caused mainly by S. enterica serovar Typhi (S. Typhi), and a clinically indistinguishable condition is caused by S. Paratyphi A. Salmonella enterica serovar Typhi is a host-restricted serovar that specifically infects humans, belongs to serogroup D and is an auxotroph (Fig. 1). As S. Typhi is restricted to humans, there are no suitable animal models. In order to study typhoid fever pathogenesis, S. Typhimurium has been used for many years in a systemic infection model using susceptible mouse strains harbouring a mutation in the Nramp1 (Slc11a1) protein (Vidal et al., 1995). Moreover, the use of S. Typhimurium with strains of mice that possess the Nramp+/+ allele, which are consequently resistant to the infection, represents a model mimicking the long-term persistence observed in S. Typhi carriers (Monack et al., 2004). These models have been crucial in understanding systemic infections by S. enterica. However, as each serovar causes a distinct type of disease in humans, conclusions regarding S. Typhi pathogenesis in humans must be interpreted carefully.

Figure 1.

 Major features distinguishing Salmonella enterica serovar Typhimurium LT2 from S. Typhi CT18.

Many recent reviews have extensively covered the literature on the pathogenesis of S. enterica (Grassl & Finlay, 2008; Haraga et al., 2008; Tsolis et al., 2008; McGhie et al., 2009). This review presents a comparative analysis of the major genetic differences between S. Typhimurium and S. Typhi and how this may contribute to our understanding of typhoid pathogenesis.

Genetic diversity

Organization of genomes allows us to gain a better understanding of the mechanisms by which species or serovars have evolved. Analysis of the chromosomal gene arrangement revealed that the genomic backbone of S. Typhimurium is very similar to the Escherichia coli genome. However, major differences in gene order have been observed in the S. Typhi chromosome. Differences in the S. Typhi genome occur mainly because of genomic rearrangements involving recombination between different rRNA operons (Liu & Sanderson, 1995; Liu & Sanderson, 1996) or IS200 elements (Alokam et al., 2002). Each serovar evolves through the acquisition of genetic elements by horizontal gene transfer or by gene degradation. The genomes of S. Typhimurium strain LT2 and S. Typhi strain CT18 are composed of 4 857 432 and 4 809 037 bp, respectively (Fig. 2) (McClelland et al., 2001; Parkhill et al., 2001). Both serovars share about 89% of genes (McClelland et al., 2001). Differences between S. Typhimurium and S. Typhi include ≈480 genes unique to S. Typhimurium and ≈600 genes unique to S. Typhi (Parkhill et al., 2001). Salmonella pathogenicity islands (SPIs), plasmids, functional prophages and phage remnants contribute significantly to the genetic diversity among S. enterica strains (Rotger & Casadesús, 1999; Boyd & Brüssow, 2002) and will be discussed below.

Figure 2.

 Circular representation of Salmonella enterica serovar Typhimurium LT2 and S. Typhi CT18 chromosomes. The circles indicate the localization of the pathogenicity islands (red), the prophages and prophage remnants (green) and fimbrial operons (blue). The inner circle indicates the G+C content (values above the average content are in black, and those below are in grey). Asterisks indicate elements specific to each serovar. The outer scale is marked in bases. The chromosomes were generated using dnaplotter software (Carver et al., 2009).

The low level of genetic variation observed in S. Typhi genomes of distinct isolates from around the world revealed a highly conserved and clonal relation, suggesting that they emerged from a single progenitor, making S. Typhi a monomorphic organism (Baker & Dougan, 2007; Holt et al., 2008). Clonality is often encountered in human-restricted pathogens (Achtman, 2008). There is very little evidence of adaptive selection in S. Typhi genes, with the exception of a recent evolution in phenotypic traits that includes the acquisition of resistance to fluoroquinolones (Chau et al., 2007; Le et al., 2007). Examination of DNA sequences and the rate of change of single-nucleotide polymorphisms suggest that S. Typhi may be only 50 000 years old, a short time frame for bacteria to accumulate diversity (Selander et al., 1990; Kidgell et al., 2002a, b; Roumagnac et al., 2006). This situation strongly suggests that evolution in the S. Typhi strain population is mainly characterized by loss of gene function. Salmonella enterica serovar Typhi is an example of reductive evolution, where the adaptation to its human niche has led to the functional inactivation of genes, due to certain needs that have been satisfied by the host (Dagan et al., 2006). Annotation of the first completed S. Typhi genome sequence revealed that >200 genes have been disrupted or inactivated, representing approximately 5% of its genome (Parkhill et al., 2001), a characteristic that was confirmed by the sequencing of other S. Typhi strains (Deng et al., 2003; Holt et al., 2009). We will point out the different pseudogenes in each of the following sections. Surprisingly, most of the pseudogenes in S. Typhi are intact and fully functional in S. Typhimurium (McClelland et al., 2001) and could explain in part the loss of host range for serovar S. Typhi. Interestingly, many pseudogenes from S. Typhi are also conserved in Paratyphi A, a serovar that has the ability to cause enteric fever that afflicts only humans (McClelland et al., 2004; Holt et al., 2009).


Most S. Typhimurium strains contain a self-transmissible virulence plasmid (pSLT) of about 90 kb harbouring virulence genes such as the spv operon, involved in intramacrophage survival, and the plasmid-encoded fimbriae (pef) fimbrial operon (Gulig & Doyle, 1993; Ahmer et al., 1999; Rotger & Casadesús, 1999). When S. Typhimurium is cured of the plasmid, virulence in the mouse is decreased (Jones et al., 1982) and can be complemented by the sole addition of the spv operon (Gulig et al., 1992) encoding the SpvB toxin (Lesnick et al., 2001). Additionally, S. Typhimurium can also carry multidrug-resistance plasmids of high molecular weight (up to 200 kb) and much smaller plasmids (<20 kb) with unknown functions (Rychlik et al., 2006). The pSLT virulence plasmid is absent in S. Typhi strains. In S. Typhi, incHI plasmids involved in multiple-drug resistance are commonly found (Maher & Taylor, 1993; Fica et al., 1997; Wain et al., 2003). Salmonella enterica serovar Typhi strain CT18 harbours plasmid pHCM1, an incHI1 plasmid of about 218 kb with genes for resistance to antibiotics and heavy metals (Parkhill et al., 2001). Salmonella enterica serovar Typhi can also carry cryptic plasmids. Salmonella enterica serovar Typhi strain CT18 harbours the cryptic plasmid pHCM2 of about 106 kb whose function is unknown, but it is rarely present in other strains (Parkhill et al., 2001; Kidgell et al., 2002a, b). Additionally, a 27-kb linear plasmid was recently isolated in S. Typhi strains originating from Indonesia. This plasmid carries the fljBz66 gene, encoding a flagellin antigen known as H:z66 (Baker et al., 2007b). However, no plasmid has been identified yet in S. Typhi that has been associated with virulence.


Integrated bacteriophages represent major loci of genetic diversity in bacterial genomes (Brüssow et al., 2004). Salmonella genomes contain several prophages or prophage remnants with similarity to the lambda, Mu, P2 and P4 families (Thomson et al., 2004; Bossi & Figueroa-Bossi, 2005). The contribution of prophages to S. enterica virulence has been recognized only recently. Some prophages carry nonessential ‘cargo’ genes involved in fitness and/or virulence, including several type three secreted effectors (Ehrbar & Hardt, 2005). Each strain of S. Typhimurium seems to have a distinct set of prophage elements within its genome. Salmonella enterica serovar Typhimurium strain LT2 harbours four prophages, including Gifsy-1, Gifsy-2, Fels-1 and Fels-2 (McClelland et al., 2001; Brüssow et al., 2004). Both the Gifsy-3 and the SopE prophages, found in S. Typhimurium strains 14028 and SL1344, respectively, are absent in S. Typhimurium strain LT2 (Figueroa-Bossi et al., 2001; Brüssow et al., 2004; Thomson et al., 2004). Salmonella enterica serovar Typhimurium strains SL1344 and 14028 both contain Gifsy-1 and Gifsy-2, but not Fels-1 and Fels-2 (Figueroa-Bossi et al., 2001). Salmonella enterica serovar Typhi harbours seven distinct prophage-like elements, spanning >180 kb, that are generally conserved between strains (Fig. 2) (Thomson et al., 2004). The modular nature of prophage genomes makes a significant contribution to serovar variation and comprises most of the variation in gene content among strains of the same serovar (Boyd et al., 2003; Vernikos et al., 2007).


Salmonella has many virulence-associated genes found within clusters in its genome, which are known as SPIs (Mills et al., 1995). Virulence factors encoded by SPI genes tamper with host cellular mechanisms and are thought to dictate the host specificity of the different S. enterica serovars (Eswarappa et al., 2008). Many of the SPIs are found next to a tRNA gene (Supporting Information, Fig. S1) and their G+C content differs from the rest of the genome (Fig. 2). Hence, such genomic islands were most likely inserted into the DNA of Salmonella by horizontal transfer events, although this explanation remains uncertain (Amavisit et al., 2003). Twenty-one SPIs are known to date in Salmonella (McClelland et al., 2001; Parkhill et al., 2001; Chiu et al., 2005; Shah et al., 2005; Vernikos & Parkhill, 2006; Fuentes et al., 2008; Blondel et al., 2009). The S. Typhimurium and S. Typhi genomes contain 11 common SPIs (SPIs-1 to 6, 9, 11, 12, 13 and 16) (Fig. 2). SPIs-8 and 10 were initially found in S. Typhi, and considered as absent in S. Typhimurium. However, at both locations in S. Typhimurium, there is a completely different set of genes. There is only one SPI specific to S. Typhimurium, SPI-14 (Shah et al., 2005), and four SPIs are specific to S. Typhi (SPIs-7, 15, 17 and 18) (Fig. 2). SPIs-19, 20 and 21 are absent in both of these serovars and will not be discussed further (Blondel et al., 2009). Even if many of these islands are found in both serovars, differences emerge when comparing equivalent SPIs. In the following section, the genomic differences between S. Typhimurium and S. Typhi are described for each SPI using S. Typhimurium strain LT2 and S. Typhi strain CT18 as the genomic references. Amino acid alignments of SPIs between these strains were performed using the xbase software (Chaudhuri & Pallen, 2006) and can be seen in Fig. S1.

SPI-1 is a 40 kb locus located at centisome 63 encoding a type three secretion system (T3SS) (Mills et al., 1995) and the sit metal transport system (Janakiraman & Slauch, 2000). The T3SS is involved in the invasion of nonphagocytic cells and proinflammatory responses (Galán & Curtiss, 1989; Mills et al., 1995; Galán & Collmer, 1999). T3SS are used by the bacteria to inject proteins, called effectors, directly inside the host cells that will act as mediators of cell invasion and modifications contributing to intracellular growth. Effectors can be encoded by genes located inside or outside SPI-1. Genomic comparison confirmed a high degree of identity between the two serovars and revealed the presence of four additional ORFs in S. Typhimurium, including the bacterial effector avrA (Hardt & Galán, 1997) and three distal ORFs (STM2901, STM2902 and STM2903) encoding putative cytoplasmic proteins (Fig. S1a) (Parkhill et al., 2001). In S. Typhi, a partial insertion sequence and transposase are present at the end of the locus. Therefore, the major difference in SPI-1 between both serovars may be at the functional level, as some genes coding effectors located outside SPI-1 are missing (sspH1, steB) or are pseudogenes (sopA, sopE2 and slrP) in S. Typhi. All known SPI-1 and SPI-2 effectors of the two serovars are listed in Table S1. Amino acid substitutions in the SipD translocon and the SptP effector were identified between these serovars and may reflect a potential functionality difference (Eswarappa et al., 2008).

SPI-2 is a 40 kb locus inserted next to the valV tRNA gene at centisome 30 and encodes a second T3SS, which is involved in intracellular survival (Shea et al., 1996; Hensel et al., 1998). Using comparative genomics, no major differences in SPI-2 were observed between both serovars (Fig. S1b). Three ORFs (STY1735, STY1739 and STY1742) are pseudogenes in S. Typhi. These ORFs, however, are not part of the T3SS, but part of a tetrathionate reductase complex. As with SPI-1, some genes encoding effectors in S. Typhimurium that are located outside SPI-2 are missing (sseI, sseK1, sseK2 and sseK3) or are pseudogenes (sopD2, sseJ) in S. Typhi (Table S1). Molecular differences were observed in translocon genes sseC and sseD, and effectors sseF and sifA (Eswarappa et al., 2008), reflecting a probable difference in functionality between these serovars.

SPI-3 is a 36 kb locus inserted next to the selC tRNA gene located at centisome 82, is involved in intracellular survival and encodes a magnesium transporter (Blanc-Potard & Groisman, 1997). SPI-3 shows extensive variations in its structure in various S. enterica serovars and can be divided into three regions (Fig. S1c) (Blanc-Potard et al., 1999; Amavisit et al., 2003). The region found next to the selC tRNA gene is where variations between S. Typhimurium and S. Typhi are the highest, including deletions and insertions. This region contains many pseudogenes in S. Typhi: STY4024 (cigR), STY4027 (marT), STY4030 (misL), STY4034, STY4035 and STY4037. A few more pseudogenes in S. Typhi are found in the second and third portions of SPI-3, including STY4012, STY4007 and STY4003 (Fig. S1c). In brief, the autotransporter MisL involved in intestinal colonization (Dorsey et al., 2005), its regulator MarT (Tükel et al., 2007) and an unknown putative transcriptional regulator (STY4012) are inactivated in S. Typhi.

SPI-4 is a 24 kb fragment located next to a potential tRNA-like gene at centisome 92 (Fig. S1d) and involved in adhesion to epithelial cells (Wong et al., 1998). SPI-4 harbours the siiABCDEF gene cluster encoding a type one secretion system (T1SS) for SiiE, a giant nonfimbrial adhesin of 595 kDa (Morgan et al., 2004; Gerlach et al., 2007; Morgan et al., 2007). SiiE mediates a close interaction with microvilli found on the apical side of epithelial cells, thereby aiding efficient translocation of SPI-1 effectors required for apical membrane ruffling (Gerlach et al., 2008). SiiE is encoded by one ORF in S. Typhimurium (STM4261), but is segmented into two ORFs in S. Typhi (STY4458 and STY4459) because of a stop codon, also present in S. Typhi strain Ty2 (Fig. S1d) (Deng et al., 2003). This suggests that siiE is a pseudogene in S. Typhi (Parkhill et al., 2001; Morgan et al., 2004), which correlates with a loss of function for an adhesin that contributes to intestinal colonization by S. Typhimurium (Morgan et al., 2007).

SPI-5 is an island <8 kb in size, inserted next to the serT tRNA gene at centisome 25, and is required for enteropathogenicity (Wood et al., 1998). SPI-5 encodes effectors of both SPI-1 and SPI-2. No difference is observed between the two serovars, except that an additional ORF (STY1114) is predicted to encode a transposase in S. Typhi (Fig. S1e).

SPI-6 is located next to the aspV tRNA gene at centisome 7 and is a 47 kb island in S. Typhimurium (Folkesson et al., 1999; Folkesson et al., 2002), whereas it is rather 59 kb in S. Typhi (Parkhill et al., 2001). It was previously shown that the complete deletion of this island reduced the entry of S. Typhimurium in Hep2 cells (Folkesson et al., 2002). Located on this island are a type six secretion system (T6SS), the safABCD fimbrial gene cluster and the invasin pagN (Lambert & Smith, 2008), all present in both serovars (Folkesson et al., 1999; Townsend et al., 2001; Porwollik & McClelland, 2003). A 10 kb fragment downstream of the saf operon is found only in S. Typhi, and includes probable transposase remnants (STY0343 and STY0344, both pseudogenes), the fimbrial operon tcfABCD and genes tinR (STY0349) and tioA (STY0350) (Fig. S1f) (Folkesson et al., 1999; Townsend et al., 2001; Porwollik & McClelland, 2003). The T6SS of S. Typhi contains two pseudogenes, sciI (STY0298) and sciS (STY0308), and some ORFs are missing or divergent, probably rendering its T6SS nonfunctional. Interestingly, sciS was shown to limit the intracellular growth of S. Typhimurium in macrophages at a late stage of infection and to decrease virulence in mice (Parsons & Heffron, 2005).

SPI-7 remains the largest island identified to date and is absent in S. Typhimurium, but present in S. Typhi (Parkhill et al., 2001; Pickard et al., 2003; Bueno et al., 2004). In S. Typhi, it is 134 kb in size, corresponding to approximately 150 genes inserted between duplicated pheU tRNA sequences (Hansen-Wester & Hensel, 2002; Pickard et al., 2003). This island contains the Vi capsule biosynthesis genes (Hashimoto et al., 1993), whose production is associated with virulence (see section below), a type IVB pilus operon (Zhang et al., 2000) and the SopE prophage (ST44) encoding the SPI-1 effector SopE (Mirold et al., 1999). SopE is also encoded in S. Typhimurium's genome, but within the temperate SopE prophage (Hardt et al., 1998) located at a different location (sopE is absent in most S. Typhimurium strains, including S. Typhimurium strain LT2, but present and located on a prophage in S. Typhimurium strains SL1344 and 14028) (Hardt et al., 1998; Mirold et al., 1999; Pelludat et al., 2003). At the SPI-7 location in S. Typhimurium LT2, we find a single complete pheU tRNA sequence and STM4320 (a putative merR family bacterial regulatory protein) (Fig. S1g).

SPI-8 is an 8 kb DNA fragment found next to the pheV tRNA gene that is part of SPI-13 and will be discussed in that section (Fig. S1l) (Parkhill et al., 2001; Hensel, 2004).

SPI-9 is a 16 kb locus present in both serovars (Fig. S1h). This island contains three genes encoding for a T1SS and one for a large protein, sharing an overall 40% nucleotide identity to siiCDEF genes from SPI-4 (Morgan et al., 2004, 2007). The large protein-coding ORF (STM2689) in S. Typhimurium strain LT2 was first suggested to be a pseudogene (McClelland et al., 2001; Morgan et al., 2004). However, a subsequent study showed an undisrupted gene coding a putative 386 kDa product renamed BapA (Latasa et al., 2005).

SPI-10 is an island found next to the leuX tRNA gene at centisome 93. This locus is completely different in each serovar and has been termed SPI-10 only in S. Typhi. In S. Typhimurium, it is substituted by a 20 kb uncharacterized island without any SPI annotation (Fig. S1i), comprising functionally unrelated genes that share little homology to sequences from the genomic databases (Edwards et al., 2001; Bishop et al., 2005). However, a possible relationship of these genes with DNA repair has been proposed (Porwollik & McClelland, 2003). Deletion of this island in S. Typhimurium strain 14028 caused attenuation of virulence in mice (Haneda et al., 2009). In S. Typhi's genome, this island corresponds to a 33 kb fragment (Parkhill et al., 2001) carrying a full P4-related prophage, termed ST46 (Edwards et al., 2001; Thomson et al., 2004; Bishop et al., 2005). ST46 harbours the prpZ cluster as cargo genes encoding eukaryotic-type Ser/Thr protein kinases and phosphatases involved in S. Typhi survival in macrophages (Faucher et al., 2008). There is also a complete, but inactivated sefABCDR (S. Enteritidis fimbriae) fimbrial operon (Fig S1i). Many pseudogenes are found in S. Typhi: STY4835 (IS1230), STY4836 (sefA), STY4839 (sefD), STY4841 (sefR), STY4845 (a thiol : disulphide interchange protein) and STY4848 (putative transposase) (Fig. S1i). Interestingly, ORFs STY4842–4846 of S. Typhi are homologues to S. Typhimurium genes located on the virulence plasmid, including srgA (Rodríguez-Peña et al., 1997). srgA encodes a functional disulphide oxidoreductase in S. Typhimurium and is a pseudogene in S. Typhi (STY4845) (Bouwman et al., 2003). It was shown that SrgA acts in concert with DsbA, another disulphide oxidoreductase, to target SipA (a SPI-2 effector), and that an srgA dsbA double mutant had a stronger attenuation than either single mutants, with a level of attenuation similar to a SPI-2 mutant (Miki et al., 2004).

SPI-11 was initially identified in the genome sequencing of serovar Choleraesuis as a 14 kb fragment inserted next to the Gifsy-1 prophage (Chiu et al., 2005). This SPI is shorter in S. Typhimurium (6.7 kb) and in S. Typhi (10 kb) (Fig. S1j). SPI-11 includes the phoP-activated genes pagD and pagC involved in intramacrophage survival (Miller et al., 1989; Gunn et al., 1995). The putative envelope lipoprotein envF is absent in S. Typhi, while six additional ORFs (STY1884–1891), including the typhoid toxin cdtB, are present in S. Typhi (Fig. S1j) (Spanòet al., 2008).

SPI-12, located next to the proL tRNA gene at centisome 48, is 15.8 kb long in S. Typhimurium and 6.3 kb long in S. Typhi (Fig. S1k) (Hansen-Wester & Hensel, 2002). It contains the effector SspH2 (Miao et al., 1999). The additional 9.5 kb fragment in S. Typhimurium contains 11 ORFs, which include some putative and phage-associated genes as well as oafA, encoding a Salmonella-specific gene for O-antigen acetylase (Fig. S1k) (Slauch et al., 1996; Hansen-Wester & Hensel, 2002). SPI-12 was shown to be required for systemic infection of mice in S. Typhimurium strain 14028 (Haneda et al., 2009). In S. Typhi, three ORFs are pseudogenes (STY2466a, STY2468 and STY2469), leaving only the sspH2 gene as functional on this island.

SPI-13 was initially identified in serovar Gallinarum (Shah et al., 2005). This 25 kb gene cluster is found next to the pheV tRNA gene at centisome 67 in S. Typhimurium and in S. Typhi. However, an 8 kb portion is different in each serovar and corresponds to SPI-8 only in S. Typhi (Fig. S1l). In S. Typhimurium, this region contains the ORFs STM3117 to STM3123, a cluster unique to S. Typhimurium, coding genes for a putative lyase, hydrolase, oxidase, arylsulphatase and arylsulphatase regulator as well as two putative LysR family transcriptional regulators (Fig. S1l). In strain S. Typhimurium 14028, STM3117–STM3121 are novel virulence-associated genes, as they were shown to be involved in systemic infection of mice (Haneda et al., 2009) and replication inside murine macrophages (Shi et al., 2006). In S. Typhi, the virulence function of SPI-8 is unknown and it harbours two bacteriocin immunity proteins (STY3281 and STY3283) and four pseudogenes (Fig. S1l) (Parkhill et al., 2001). The 17 kb conserved portion of SPI-13 has not been shown to contribute to virulence (Haneda et al., 2009).

SPI-14 corresponds to 9 kb present in S. Typhimurium at centisome 19 and is absent in S. Typhi (Shah et al., 2005; Morgan, 2007). It harbours seven ORFs encoding putative cytoplasmic proteins (Fig. S1m). The function of genes on this island is unknown, but gene upregulation was observed in macrophages infected by S. Typhimurium strain SL1344 (Eriksson et al., 2003).

SPI-15 is a 6.5 kb island of five ORFs encoding hypothetical proteins, is inserted near the glyU tRNA gene in S. Typhi and is absent in S. Typhimurium (Fig. S1n) (Vernikos & Parkhill, 2006). Different genes are found at the same location in S. Typhi strain Ty2 (Fig. S1n) (Vernikos & Parkhill, 2006). SPI-15, as well as SPI-16 and 17, were identified by bioinformatic work (Vernikos & Parkhill, 2006).

SPI-16 is found in S. Typhimurium and S. Typhi as a 4.5 kb fragment inserted next to an argU tRNA site, and encodes five or seven ORFs, respectively, four of which are pseudogenes in S. Typhi (Fig. S1o). The three remaining ORFs show a high level of identity with P22 phage genes involved in seroconversion (Vernikos & Parkhill, 2006) and were suggested to mediate O-antigen glycosylation (Mavris et al., 1997; Guan et al., 1999) and cell surface variation (Allison & Verma, 2000; Bogomolnaya et al., 2008). These ORFs (genes yfdH, rfbI and STM0557) were required for the intestinal persistence of S. Typhimurium in mice (Bogomolnaya et al., 2008).

SPI-17 is a 5 kb island encoding six ORFs inserted next to an argW tRNA site and is absent in S. Typhimurium, but present in S. Typhi (Fig. S1p) (Vernikos & Parkhill, 2006). Seroconversion genes homologous to P22 phage are present and showed high homology to genes of SPI-16, including a putative lipopolysaccharide modification acyltransferase. Most of these genes (four) are pseudogenes in S. Typhi (Fig. S1p).

SPI-18 was recently identified in S. Typhi as a 2.3 kb fragment harbouring only two ORFs: STY1498 and STY1499 (Fig. S1q) (Fuentes et al., 2008). clyA (STY1498), also known as hlyE or sheA, encodes a 34 kDa pore-forming secreted cytolysin found in E. coli and S. enterica serovars Typhi and Paratyphi A (del Castillo et al., 1997; Green & Baldwin, 1997; Oscarsson et al., 1999, 2002). clyA is important for invasion of human epithelial cells in vitro, with its heterologous expression in S. Typhimurium leading to colonization of deep organs in a murine model (Fuentes et al., 2008). taiA (STY1499) is a secreted 27 kDa invasin that increases bacterial uptake by human macrophages (Faucher et al., 2009). Both genes are part of a common operon and are controlled by the virulence-related regulator PhoP (Faucher et al., 2009).

Other pathogenicity islands are found in the S. Typhimurium and S. Typhi genomes and have not been identified as SPIs, but encode genes responsible for virulence in the host, such as CS54. The CS54 island is a 25 kb region found between xseA and yfgJ at centisome 54 in S. Typhimurium (Kingsley et al., 2003) and S. Typhi (Fig. S1r). Five genes are found within this island, which are shdA, ratB, ratA, sinI and sinH (sivH). In S. Typhimurium, ShdA was shown to be an outer membrane protein of the autotransporter family that binds fibronectin, RatB is a predicted secreted protein of unknown function and SinH is a putative outer membrane protein (Kingsley & Bäumler, 2002; Kingsley et al., 2003; Abd El Ghany et al., 2007). shdA, ratB and sinH (sivH) are all implicated in intestinal colonization of BALB/c mice by S. Typhimurium, but are all pseudogenes in S. Typhi (Kingsley et al., 2003).

Other virulence factors


Fimbriae or pili are proteinous structures found on bacteria that can mediate interaction with cells. Fimbriae are normally specific to a receptor and can be used at different critical times during the infection. Each serovar harbours a unique combination of fimbrial operons (Fig. 2). Whole-genome sequence analysis revealed eight putative fimbrial operons shared by both S. Typhimurium and S. Typhi [bcf, csg (agf), fim, saf, stb, stc, std, sth] (McClelland et al., 2001; Parkhill et al., 2001). Salmonella enterica serovar Typhimurium possesses five unique fimbrial sequences known as lpf, stf, pef, sti and stj (McClelland et al., 2001). These unique fimbriae were not involved in systemic colonization of the spleen, and Lpf was shown to be involved in intestinal colonization of mice (Weening et al., 2005). A type IVB pilus located on SPI-7 is only found within the S. Typhi genome, along with five other unique fimbrial operons (sef, sta, ste, stg and tcf) (Parkhill et al., 2001). For the majority of these fimbriae, little is known about their true function, expression conditions or their importance for virulence during infection. Type IV pili are used by Typhi for adhesion to human monocytes and epithelial cells by interaction with the cystic fibrosis transmembrane conductance regulator receptor (Pier et al., 1998; Zhang et al., 2000; Tsui et al., 2003; Pan et al., 2005). Tcf was recognized by human sera from patients with typhoid (Harris et al., 2006) and Stg mediates adherence to epithelial cells and reduces phagocytosis (Forest et al., 2007). All fimbrial operons are intact in S. Typhimurium strain LT2, although pseudogenes are found within six fimbrial operons of S. Typhi strain CT18 (fimI, safE, sefA, sefD, bcfC, steA, stgC, sthC, sthE) (Townsend et al., 2001) ( The unique repertoire of fimbrial adhesin systems may explain in part some differences observed between the host tropism colonization niches of these two serovars.


In Salmonella, the major subunit of flagella is usually encoded by fliC or fljB, which correspond to the H1 and H2 variants of the H antigen, respectively (Silverman & Simon, 1980). Only one type of flagellin can be expressed at a specific time by a mechanism known as phase variation (Lederberg & Iino, 1956; Simon et al., 1980). This antigenic variation can be observed in S. Typhimurium, but most S. Typhi strains are considered monophasic, as they lack a corresponding fljB locus (Frankel et al., 1989). However, some S. Typhi isolates from Indonesia contain a linear plasmid encoding a novel flagellin, fljBz66, but reversion to fliC is considered irreversible due to a deletion (Baker et al., 2007a). fliB, involved in methylation of the flagellin in S. Typhimurium, is a pseudogene in S. Typhi (Parkhill et al., 2001).

Vi antigen

The Vi antigen is a polysaccharidic capsule absent in S. Typhimurium and present in S. Typhi. Vi is important for virulence and is controlled by two loci: viaA and viaB (Kolyva et al., 1992). The viaB locus located on SPI-7 is composed of two operons: tviABCDE and vexABCDE. The Vi capsule causes several differences between S. Typhimurium and S. Typhi at the level of the host's response to infection. The Vi capsule is associated with inhibition of complement activation, resistance to serum and to phagocytosis and is involved in survival inside phagocytes (Looney & Steigbigel, 1986; Hirose et al., 1997; Miyake et al., 1998). The viaB locus lowers the invasiveness of the bacteria towards epithelial cells, as viaB mutants are superinvasive (Arricau et al., 1998; Zhao et al., 2001), and S. Typhimurium harbouring the viaB locus is less invasive (Haneda et al., 2009). TviA avoids interleukin-8 production in the intestinal mucosa by repressing flagellin secretion, which reduces the recognition and activation of Toll-like receptor (TLR)-5 (Raffatellu et al., 2005; Winter et al., 2008). Vi also prevents the recognition of lipopolysaccharide by TLR-4 and reduces inflammation in the intestinal mucosa (Sharma & Qadri, 2004; Wilson et al., 2008). Salmonella enterica serovar Typhimurium sets off an immune response, which causes inflammation characterized by an important neutrophil influx that may be the result of its lack of capsule. Thus, Vi allows S. Typhi to disseminate systemically in its human host by crossing intestinal cells without activating the immune response, promotes resistance to killing by serum and contributes to survival inside phagocytes (Raffatellu et al., 2006). Vi is a protective antigen and the actual constituent of the parenteral typhoid fever vaccine.


Acquisition and loss of genetic material play an important role in bacterial evolution. Here, we have described the major genetic differences between S. Typhimurium and S. Typhi, two important S. enterica serovars associated with distinct diseases in humans (Fig. 1). Gene degradation in S. Typhi may be responsible for its human host restriction, but factors contributing to its systemic dispersion and survival during typhoid may be multiple and scattered, which complicates the identification of genomic regions that reflect differences in habitat and lifestyle.

Each serovar is distinguished by its own repertoire of virulence factors, whether surface expressed or secreted, that leads to specific diseases or hosts. The bacterial cell surface is different between the two serovars, as represented by various O:H:K antigens. Lipopolysaccharide differences (O antigen) allowed the classification of S. Typhimurium in serogroup B, while S. Typhi belongs to serogroup D. Only S. Typhimurium is capable of phase variation of its H antigen, by differential expression of two flagella subunits. The most important feature is undoubtedly the presence of a polysaccharidic capsule (K antigen) specific to S. Typhi, the Vi antigen. However, it is also interesting to note that some S. Typhi strains and S. Paratyphi A lack the Vi antigen, but both cause a disease very similar to S. Typhi in the human host (McClelland et al., 2004; Baker et al., 2005).

Virulence factors can be secreted using the general secretion machinery of the bacteria or by specific systems, such as the T3SS used to inject proteins directly into the host. No major differences were observed in both T3SS (Fig. S1a,b), but some of the effectors were missing in S. Typhi (Table S1). However, the T6SS included on SPI-6 is probably inactivated in S. Typhi by the presence of pseudogenes. Some toxins were specific to S. Typhimurium, such as SpvB present on the virulence plasmid. On the other hand, the CdtB and ClyA toxins are only produced by S. Typhi.

Interestingly, most of the genes involved in intestinal colonization identified in S. Typhimurium are inactivated in S. Typhi. These genes encode autotransporters MisL and ShdA, adhesin SiiE, secreted protein RatB, putative outer membrane protein SinH and Lpf fimbriae (Fig. 1), suggesting that they are not needed by S. Typhi in the human host. This particular divergence is further acknowledged when looking at some work involving vaccine development (DiPetrillo et al., 1999; Angelakopoulos & Hohmann, 2000; Hindle et al., 2002). Salmonella enterica serovar Typhimurium and S. Typhi live-attenuated vaccine strains, both modified with the same genetic deletions, did not show the same level of intestinal colonization when administered orally to human volunteers. Prolonged bacterial shedding by the host was observed over time with S. Typhimurium, but not with S. Typhi. Thus, precautions must be taken when extrapolating the S. Typhimurium data to S. Typhi. Many clinical trials investigating novel S. Typhi vaccine strains harbouring mutations that render S. Typhimurium avirulent and immunogenic in mice led to disappointing results at the attenuation level when administered to humans (Hone et al., 1988; Tacket et al., 1992a, b).

The completion of additional genome sequencing projects of other Salmonella serovars or strains will contribute considerably to our understanding of niche adaptation and bacterial evolution in general, as well as conceiving the molecular basis of epidemics and how new virulent strains emerge. However, the availability of whole-genome sequences of several strains of different S. enterica serovars has not revealed any explanation that correlates with their specific niches and fitness, which suggests that the answer is probably already under our noses ….


The authors would like to thank Charles M. Dozois and Frédéric Douesnard-Malo for critical comments concerning this manuscript. This work was supported by the Natural Sciences and Engineering Research Council (NSERC) grant number 251114-06. S.C.S. was supported by a scholarship from the Fonds de la Recherche en Santé du Québec (FRSQ). C.G.F. and J.M.L. were supported by scholarships from NSERC. C.G.F. was also supported by a scholarship from the Centre de Recherche en Infectiologie Porcine (CRIP).