Linkage of avian and reproductive tract tropism with sequence divergence adjacent to the 5S ribosomal subunit rrfH of Salmonella enterica


  • Editor: Reggie Lo

Correspondence: Jean Guard Bouldin, USDA-ARS, 950 College Station Road, Athens, GA 30605, USA. Tel.: +1 706 546 3446; fax: +1 706 546 3035; e-mail:


The 183 bp between the end of the 23S rrlH rRNA gene and the start of the 5S rrfH rRNA gene (ISR-1) and the 197 bp between the end of the rrfH rRNA gene and the start of the transfer RNA aspU (ISR-2) of Salmonella enterica ssp. enterica serotypes Enteritidis, Typhimurium, Pullorum, Heidelberg, Gallinarum, Typhi and Choleraesuis were compared. ISR-1s of D1 serotypes (Pullorum, Gallinarum and Enteritidis), B serotypes (Typhimurium and Heidelberg) and the C2 serotype Newport and the enteric fever pathogens serotype A Paratyphi and serotype D1 Typhi formed three clades, respectively. ISR-2 further differentiated the avian-adapted serotype Gallinarum from avian-adapted Pullorum and Salmonella bongori from S. enterica. The results suggest that serotypes Heidelberg and Choleraesuis share some evolutionary trends with egg-contaminating serotypes. In addition, ISR-1 and ISR-2 sequences that confirm serotype appear to be linked to clinically relevant host associations of the Salmonellae.


Salmonella enterica ssp. enterica serotype Enteritidis (S. Enteritidis) is currently the leading cause of salmonellosis worldwide and is the second leading cause in the United States (Fisher, 1999; Anonymous, 2005). This serotype challenges pathogen reduction strategies because it is capable of colonizing the avian reproductive tract and subsequently contaminating eggs produced by otherwise healthy hens (Lu et al., 2004; Mumma et al., 2004). Although it is known that nonhuman pathogenic Salmonella enterica serotypes Pullorum (S. Pullorum) and Gallinarum (S. Gallinarum) also contaminate the internal contents of eggs (Shivaprasad, 2003), other serotypes in eggs are sporadically reported. Of major concern is that the environmentally prevalent Salmonella enterica serotype Heidelberg (S. Heidelberg) is evolving the ability to contaminate eggs (Hennessy et al., 2004).

The long-term objective of this research is to identify small-scale genetic change within S. enterica that correlates with increased outbreak potential in humans. Salmonella Enteritidis is ideal for these investigations, because it generates clonally related subpopulations that vary in their ability to grow to high cell density, to produce cell surface structures such as capsular lipopolysaccharide (LPS) and to contaminate eggs (Guard-Bouldin et al., 2004). Reproductive tract tropism and the subsequent ability of the pathogen to contaminate eggs are the most important phenotypic characteristics for which genetic markers are sought. However, markers are also sought for the many metabolic differences between subpopulations that alter growth potential (Morales et al., 2005). Larger scale evolutionary events occur within the Salmonellae and these events provide a separate layer of virulence attributes that also influence outbreak potential (Guard-Petter et al., 1999; De Buck et al., 2004; Levy et al., 2004; Porwollik et al., 2005). However, no large-scale event such as phage acquisition has been correlated with the specific ability of S. Enteritidis to contaminate eggs as compared with its ability to be generally invasive and capable of causing gastroenteritis (Morales et al., 2005). Genetic markers that correlate with pathogenicity and growth potential are thus needed to improve epidemiological monitoring of the Salmonellae.

The Salmonellae genome is analyzed by pulsed-field gel electrophoresis (Swaminathan et al., 2001), amplified fragment length polymorphism (Liebana, 2002), multilocus sequence typing (Sukhnanand et al., 2005), multiple-locus variable-number tandem repeats (Rep-PCR) (Lindstedt et al., 2004), ribotyping (Liebana et al., 2001) and DNA–DNA microarray hybridization (Morales et al., 2005). These methods detect some heterogeneity between strains of the same serotype, but they do not reliably correlate phenotype with genotype (Morales et al., 2005). Analysis of the intergenic spacer regions (ISRs) that separate DNA-encoding ribosomal subunits (rRNA gene) achieves another layer of discrimination (Chiu et al., 2005a); however, the rRNA genes themselves are too highly conserved to be useful for discrimination between S. enterica serotypes (Jaspers & Overmann, 2004). We wanted to know whether ISRs adjacent to rRNA gene are potential candidates for indirectly linking phenotype to genotype, because they can undergo mutation without altering the critical cell functions involved in producing proteins (Woese, 1987; Laursen et al., 2005).

The Salmonellae contain seven copies of rRNA gene regions in the genome, six of which have a 16S–23S–5S arrangement (Helm et al., 2004). A number of studies focused on the 16S–23S ISR (Roth et al., 1998; Iteman et al., 2000; Gonzalez et al., 2003); however, others note a higher discriminatory power in the 23S–5S ISR (Tilsala-Timisjarvi & Alatossava, 2001; Herpers et al., 2003). Only one study has reported using the 23S–5S ISR for the purposes of characterizing Salmonella enterica serotype Typhi (S. Typhi) (Zhu et al., 1996), whereas other Salmonella ISR studies only examined the 16S–23S ISR (Bakshi et al., 2002; Chiu et al., 2005a). Investigations here examine whether the 23S–5S and 5S-aspU ISRs of rrnH rRNA gene can be used to (i) discriminate naturally occurring subpopulations of S. Enteritidis that vary in their ability to contaminate eggs, (ii) discriminate between serotypes of S. enterica ssp. enterica and (iii) identify evolutionary trends occurring within the pathogenic Salmonellae that may correlate with egg contamination or adaptation to the avian environment.

Materials and methods

Strains for analysis

Salmonella isolates and their sources used in ISR analysis are listed in Table 1. For serotype D1 Salmonellae, seven S. Pullorum, one S. Gallinarum and 20 S. Enteritidis strains were used. Salmonella Enteritidis strains included seven phage type 13a (PT13a) strains expressing either a wild-type or biofilm-forming phenotype, three phage type 4 (PT4) strains and 10 S. Enteritidis strains isolated from whole egg contents. For serotype B Salmonellae, 14 S. Heidelberg, six lipopolysaccharide O-antigen B field isolates, five Salmonella Typhimurium 5 (also known as variant Copenhagen) and two S. Typhimurium definitive type 104 (DT104) strains were used. Nine Salmonella enterica serotype Newport (S. Newport) strains representing serotype C2 were used. The generation of original sequence from these strains resulted in 14 consensus sequences, which were entered into the NCBI database as a population set in GenBank (accession numbers DQ231427DQ231440). The DNA sequences of strains NCTC 133461 (S. Gallinarum), NCTC 133492 (S. Enteritidis phage type 4), LT23 (S. Typhimurium), NCTC 133484 (S. Typhimurium definitive type 104), ATCC 91505 (Salmonella Paratyphi A), CT186 (S. Typhi), SC-B677 (Salmonella enterica serotype Choleraesuis) and ATCC 439758 (Salmonella bongori) were obtained from Internet databases.

Table 1. Salmonella strains used for characterization of ISRs adjacent to 5S rrfH rRNA gene

of bp)
% similarity
  • *

    Strains listed within a serotype heading have 100% similarity scores for both ISR-1 and -2.

  • Genus species designation for a non-pathogenic Salmonellae that is not classified Salmonella enterica; serotyping does not apply (NA).

  • ISRs, intergenic spacer regions; LPS, lipopolysaccharide.

Pullorum99113D1178/53Canadian Food Inspection AgencyChicken house
99114D1178/53100/100Canadian Food Inspection AgencyChicken house
99116D1178/53100/100Canadian Food Inspection AgencyChicken house
99118D1178/53100/100Canadian Food Inspection AgencyChicken house
99119D1178/53100/100Canadian Food Inspection AgencyChicken house
23022D1178/53100/100ESQRU, USABackyard flocks
23023D1178/53100/100ESQRU, USABackyard flocks
GallinarumNCTC 13346D1178/190Sanger Institute, UKGenome database
99117D1178/190Canadian Food Inspection AgencyChicken house
Enteritidis PT13a21026D1179/190ESQRU, USAMouse intestine
21027D1179/190100/100ESQRUMouse spleen
21028D1179/190100/100ESQRUMouse spleen
21029D1179/190100/100ESQRUMouse spleen
21042D1179/190100/100ESQRUMouse spleen
21046D1179/190100/100ESQRUMouse spleen
21050D1179/190100/100ESQRUMouse intestine
Enteritidis PT4NCTC 13349D1179/190Sanger Institute, UKGenome database
22076D1179/190100/100University of California, DavisFeline gut
22075D1179/190100/100University of CaliforniaChicken liver
22079D1179/190100/100University of CaliforniaCreek
Heidelberg23010B178/190100/100Centers for Disease ControlEgg
23011B178/190100/100Centers for Disease ControlEgg
23012B178/190100/100Centers for Disease ControlEgg
23013B178/190100/100Centers for Disease ControlEgg
23014B178/190100/100Centers for Disease ControlEgg
23015B178/190100/100Centers for Disease ControlEgg
23016B178/190100/100Centers for Disease ControlEgg
23017B178/190100/100Centers for Disease ControlEgg
23018B178/190100/100Centers for Disease ControlEgg
23019B178/190100/100Centers for Disease ControlEgg
23020B178/190100/100Centers for Disease ControlEgg
23021B178/190100/100Centers for Disease ControlEgg
24001B178/190100/100National Veterinary Services LaboratoryUnknown
24002B178/190100/100National Veterinary Services LaboratoryUnknown
Newport24010C2178/190100/100National Veterinary Services LaboratoryUnknown
24011C2178/190100/100National Veterinary Services LaboratoryUnknown
24012C2178/190100/100National Veterinary Services LaboratoryUnknown
24013C2178/190100/100National Veterinary Services LaboratoryUnknown
24014C2178/190100/100National Veterinary Services LaboratoryUnknown
24015C2178/190100/100National Veterinary Services LaboratoryUnknown
24016C2178/190100/100National Veterinary Services LaboratoryUnknown
24017C2178/190100/100National Veterinary Services LaboratoryUnknown
24018C2178/190100/100National Veterinary Services LaboratoryUnknown
TyphimuriumLT2B178/190Washington University GSC, MOGenome database
Serogroup B Group 199187B178/19099/100ESQRUMouse spleen
99188B178/19099/100ESQRUMouse spleen
20143B178/19099/100ESQRUMouse spleen
Serogroup B Group 220144B178/19099/94ESQRUMouse spleen
20146B178/19099/94ESQRUMouse spleen
20147B178/19099/94ESQRUMouse spleen
Typhimurium var Copenhagen99163B179/190100/100USDA ARS, College Station, TXPigeon
99167B179/190100/100USDA ARSPigeon
99168B179/190100/100USDA ARSPigeon
99170B179/190100/100USDA ARSPigeon
99172B179/190100/100USDA ARSPigeon
TyphiCT18D184/53Sanger Institute, UKGenome database
ParatyphiATCC 9150A178/190Washington University GSC, MOGenome database
CholeraesuisSC-B67C1179/190Chang Gung Genomic Medical CenterGenome database
Salmonella bongori66NA179/191Sanger Institute, UKGenome database

DNA isolation

DNA was isolated from bacteria grown in 10 mL of brain heart infusion broth (Difco, Becton, Dickinson and Company) at 37°C for 16 h. Bacterial cells were pelleted in a Sorvall RC5B Plus centrifuge at 5000 g for 15 min in a Sorvall Super-lite SLA 600TC rotor. Total DNA was extracted using a Qiagen Genomic-tip 100/G kit following the protocol designated for bacteria (Qiagen Inc.). Precipitated DNA was dissolved in 200 μL of Tris-EDTA buffer (10 mM Tris-HCl, 1 mM EDTA, pH 8) and stored at −20°C.

PCR and sequencing

Primers ISRH-1 (5′-GATGCGTTGAGCTAACCGGTACTA-3′) and ISRH-2 (5′-ATTCTTCGACAGACACGGCATCAC-3′) were designed to amplify the two ISRs between the 23S rrlH ORF and the yafB ORF, yielding a product size of 1118 bp in S. Enteritidis (Fig. 1). The cycling conditions for an Applied Biosystems 2400 Gene Amp PCR system were as follows: initial denaturation of 95°C for 1 min, then 30 cycles of 95°C for 30 s, 64°C for 30 s and 72°C for 1 min. Each reaction contained 400 nM of each primer, 200 μM ACGU deoxynucleotide triphosphates, 1.5 mM Mg++, 2.5 U Taq enzyme (Fisher, Pittsburg, PA) and 1 μL template DNA. Single amplicon products were confirmed by agarose gel electrophoresis. PCR products were purified using a Qiagen QIAquick PCR Purification kit and submitted for sequencing to the South Atlantic Area DNA Sequencing Facility (Athens, GA). Sequencing primers ISRHfs (5′-GTGGAGCGGTAG TTCAGTTGGTTA-3′) and ISRHrs (5′-TAACCAACTGAACTACCGCTCCAC-3′) were submitted in addition to primers ISRH-1 and ISRH-2 in order to achieve full nucleotide coverage of the 1119 bp region. PCR amplicons were sequenced using an Applied Biosystems BigDye Terminator 1.1 reaction mix on an Applied Biosystems 3730 DNA Analyzer.

Figure 1.

Salmonella ISRs 1 and 2 locations in the rrnH operon. ISR-1 is defined as the 183 bp between the end of the 23S rrlH rRNA gene and the start of the 5S rrfH rRNA gene and ISR-2 is the 197 bp between the end of the rrfH rRNA gene and the start of the aspU, although exceptions between Salmonella serotypes are noted in the text. The numbers shown on the top reference the nucleotide locations within the genome of Salmonella Enteritidis phage type 4 (NCTC 13349).

Sequence alignment and similarity scores

Sequences were aligned using MegAlign of the Lasergene software suite for sequence analysis version 6.0 (DNASTAR) by the Clustal W method (Thompson et al., 1994). Default parameter settings were used, with the exception of the transition weight set to 0.0 for the purposes of equally evaluating any type of nucleotide substitution in a noncoding region of DNA. Sequences were trimmed to include only the ISR between rrlH (23S) and rrfH (5S) for ISR-1 and only the ISR between rrfH and aspU for ISR-2. Similarity scores were calculated by MegAlign using direct pair-wise comparisons of the sequences. Intraserotype similarity scores were based on an arbitrarily designated strain chosen to be representative of each grouping in Table 1. Some manual adjustment to alignment was made in those cases where strains with deletions had some sequence that was obviously misplaced by program parameters. For this reason, phylogenetic trees by themselves were not appropriate for showing relationships, because the small genetic distance between groups necessitated showing the final sequence alignment assigned per group (Tables 2 and 3).

Table 2.   Grouping of Salmonella by alignment of intergenic spacer region-1*
  • *

    Base pairs that differ from the group consensus sequence have a single underline.

  • Serotyping is only done within Salmonella enterica and does not apply to Salmonella bongori.

  • Also known as Salmonella Typhimurium var Copenhagen.

  • NA, not applicable.

Group I
Group II
Group III
GAACGAAAGAT---------------------------------------S. TyphiD1
-------------------------------------------------S. TyphiD1
Table 3.   Grouping of Salmonella by alignment of intergenic spacer region-2
  • *

    Single and double underlines show polymorphisms occurring between and within groups, respectively; bold type, 19 bp element also at end of 5S rrfH rRNA gene.

  • Salmonella enterica serotypes Typhimurium, Newport, Paratyphi A and Group B-1.

  • Salmonella enterica serotypes Gallinarum, Enteritidis, Heidelberg, Group B-2, Typhimurium 5- and Choleraesuis.

  • §

    Salmonella bongori only.

  • Salmonella enterica serotypes Typhi and Pullorum.



Table 2 shows alignment of ISR-1 sequences into three genetic groups. Twelve of the 13 Salmonellae examined had ISR-1 nucleotide sequence lengths of 182 or 183 bp. However, ISR-1 of S. Typhi was 84 bp, which was due to a 94 bp internal deletion (Table 2). Overall, there were 10 unique ISR-1 sequences. Percent identity of ISR-1 for S. bongori as compared with S. enterica ranged from 71.5 (S. Choleraesuis) to 98.3 (S. Heidelberg). Percent identity within S. enterica serotypes ranged from 73.6 to 100 (Table 4). Alignments for determination of percent identity in ISR-1 excluded S. Typhi, because it had a large deletion. These results indicate that heterogeneity within serotypes of S. enterica is similar to that which occurs between S. enterica and S. bongori.

Table 4.   Interserotype percent similarity scores for intergenic spacer region-1*
 Serotype no. and serotypeSequence percent similarity score for the following Salmonella serotypes
  • *

    Alignment excluded Salmonella Typhi, which had a large deletion in comparison with all other sequences.

  • The serotype numbers correspond to the serotype numbers identified on the left.

1S. bongori71.574.373.298.375.474.373.797.274.397.898.9
2S. Choleraesuis 95.59572.67675.496.173.295.573.673
3S. Enteritidis  98.974.375.475.498.974.310074.774.7
4S. Gallinarum   73.675.375.398.973.699.473.673.6
5S. Heidelberg    75.874.774.298.974.798.9100
6S. Newport     97.875.876.475.876.475.8
7S. Paratyphi A      75.875.375.875.374.7
8S. Pullorum       74.299.474.274.2
9S. Typhimurium        74.310098.9
10S. Typhimurium 5-         74.774.7
11Serogroup B group 1          98.9
12Serogroup B group 2           

Salmonella bongori ISR-1 was highly similar to that of S. Typhimurium, which is the prototype strain for ISR-1 Group I. There was no variation in ISR-1 between three Typhimurium genomic databases and the field isolates examined here. Classification of S. bongori within this group suggests that ISR-1 is highly conserved across a certain lineage of Salmonella species. However, ISR-1 was able to discriminate between serotypes within this group that have different food and host attributions. For example, S. Typhimurium and S. Heidelberg had two single nucleotide polymorphisms (SNPs). In addition, these two O-antigen B serotypes were identical to serogroup B-1 and B-2 serotypes, respectively (Table 2).

The other O-antigen B serotype, namely, S. Typhimurium 5-, clustered with avian-adapted O-antigen D1 S. Pullorum, S. Gallinarum and S. Enteritidis in ISR-1 Group II. Salmonella Choleraesuis, which is a common pathogen of pigs that can cause food-borne illness, also clustered within ISR-1 Group II. The finding that ISR-1 differentiates between O-antigen B serotypes strongly suggests that evolution independent of serotype is taking place within the Salmonellae.

ISR-1 Group III included strains that cause enteric fever in humans, namely, Salmonella Paratyphi and S. Typhi. The third member of this group is S. Newport, which could not be intuited from current knowledge of its association with dairy products. In summary, the results from analysis of ISR-1 suggest that SNPs in this region correlate with evolutionary divergence between the pathogenic Salmonellae that impact pathogen–host associations regardless of O-antigen serotype.

ISR-2, which begins immediately 3′ to the end of rrfH and ends at aspU, produced five unique sequences that clustered into three groups for S. enterica (Table 3). Differentiation of S. bongori from S. enterica suggests that ISR-2 is a marker for a branch in evolution that occurred after speciation because the 191 bp ISR-2 of S. bongori did not cluster with the S. enterica sequences. Heterogeneity was pronounced in the last 50 bp of ISR-2. The percent identity of ISR-2 of S. bongori to the other S. enterica serotypes ranged between 88.5 and 89.0, whereas percent identity ranged from 91.6 to 100 within S. enterica (Table 5). Alignments for determination of percent identity in ISR-2 excluded S. Typhi and S. Pullorum because each had a similar large deletion.

Table 5.   Interserotype percent similarity scores for intergenic spacer region-2*
 Serotype no. and serotypeSequence percent similarity score for the following Salmonella serotypes
  • *

    Alignment excluded Salmonella Typhi and Salmonella Pullorum, which had large deletions.

  • The serotype numbers correspond to the serotype numbers identified on the left.

1S. bongori888989.58988.588.588.58988.989.5
2S. Choleraesuis 98.998.498.991.691.691.698.991.698.9
3S. Enteritidis  99.510092.692.692.610092.6100
4S. Gallinarum   99.592.
5S. Heidelberg    92.692.692.610092.6100
6S. Newport     10010092.610092.6
7S. Paratyphi A      10092.610092.6
8S. Typhimurium       92.610092.6
9S. Typhimurium 5-        92.6100
10Serogroup B group 1         92.6
11Serogroup B group 2          

For S. enterica, ISR-2 was 197 bp in length, except for S. Typhi and S. Pullorum, which each had similar deletions. These two S. enterica serotypes are somewhat unusual, because they are host specific and cause systemic disease in humans and birds, respectively. They resulted in a 54 bp ISR-2 that was shortened due to a deletion beginning 4 bp after rrfH rRNA gene that was 137 bp in length. There were two 19 bp repeating elements with the sequence 5′-AACTGCCAGGCATCAAATT-3′ that defined the site of deletion. The first 19 bp element included the last 15 bp of rrfH and 4 bp of ISR-2, which were present in strains without a deletion. The second 19 bp element was at the end of the deletion itself. The sequence of the 19 bp element is found in many plasmid shuttle vectors, such as pBAD322 (Cronan, 2006). Unlike S. Pullorum, the other avian-restricted pathogenic serotype S. Gallinarum did not have the 137 bp deletion. Salmonella Pullorum and S. Gallinarum have otherwise highly similar genetic content but differ in their ability to cause disease in chickens of different ages (Shivaprasad, 2003); thus, the 137 bp element may correlate with evolution in S. enterica that is associated with subtle rather than overt shifts in pathophysiology.

Analysis of other ISRs in S. enterica ssp. I

The 511 bp ISR between 16S rrsH and 23S rrlH ribosomal subunits was evaluated for its ability to discriminate between serotypes. Perusal of public databases for 13 Salmonella genomes, namely Typhimurium LT2, Typhimurium DT104, Typhimurium SL1344, Typhimurium DT2, Typhimurium D23580, Choleraesuis, Paratyphi A, Typhi CT18, Typhi Ty2, Infantis, Hadar, Gallinarum and Enteritidis PT4, indicated that this region did not discriminate between S. Typhimurium, S. Paratyphi and Salmonella Infantis. There was some discrimination between these three serotypes and S. Enteritidis (93% similarity), but there was no discrimination between S. Enteritidis and S. Gallinarum. These results indicate that the 16S–23S ISR in the rrnH operon does not adequately discriminate between serotypes.

The ISRs flanking the other six 5S rrf genes were similarly evaluated for their ability to discriminate between serotypes. Flanking sequences for rrfB and rrfA were similar in size and arrangement to those of rrfH. The ISRs of S. Typhimurium rrfB between rrlB-rrfB and rrfB-murB were 192 (rrfB ISR-1) and 178 bp (rrfB ISR-2), respectively, and the ISRs of rrfA between rrlA-rrfA and rrfA-mobB were 191 (rrfA ISR-1) and 260 bp (rrfA ISR-2), respectively. The S. Typhimurium rrfB ISR-1 was 96% similar to that of S. Choleraesuis, S. Enteritidis and S. Gallinarum, but it did not discriminate adequately between the latter 3 serotypes. Salmonella. Typhimurium rrfB ISR-2 was 95% and 96% similar to S. Cholerasuis and S. Enteritidis, respectively; however, S. Gallinarum lacked identity. Both ISRs flanking rrfA provided some discrimination between serotypes, but rrfA ISR-1 (rrlA–rrfA) was similar between S. Enteritidis and S. Gallinarum.

The ISRs flanking rrfG, rrfF, rrfD and rrfC were in some cases substantially different in size and arrangement from those of rrfH. For example, rrfG–rrlG (rrfG ISR-2) was 193 bp, but the rrfG–kgtP region (rrfG ISR-1) was 763 bp. The rrlC–rrfC (rrfC ISR-1) and rrfC–aspT (rrfC ISR-2) regions were 90 and 53 bp, respectively. Finally, the rrfF and rrfD genes were adjacent to each other, which resulted in naming three ISRs for these 5S genes. Namely, yhdV–rrfF (778 bp) was rrfF ISR-1, rrfF–rrfD (123 bp) was rrfF ISR-2 and rrfD–rrlD (89 bp) was rrfD ISR-1. There was no rrfD ISR-2. The 193 bp rrfG ISR-2 of S. Typhimurium was examined further because it was similar in size to the ISRs flanking rrfH. Similarity scores were 98% to S. Paratyphi and 96% to S. Infantis, whereas S. Choleraesuis, S. Enteritidis, S. Gallinarum and Salmonella Hadar had scores of 85% or lower. These results suggest that 5S genes other than rrfH could be used to improve the resolution of the serotype. However, the juxtaposition of two discriminatory flanking sequences around rrfH makes this region especially useful for determining the serotype within a single sequencing reaction.


Phylogenetic analysis of the two ISRs adjacent to 5S rrfH rRNA gene of S. enterica indicates that sequencing of a single genomic region of less than 500 bp provides a useful approach for clinicians to gain information about the serotype of any one strain of S. enterica, with the additional benefit that it resolves questions about how to classify some atypical serotypes. For example, S. Typhimurium 5, which is historically associated with pigeons (Kiessling et al., 1990), clustered with the avian-restricted serotypes S. Pullorum and S. Gallinarum, as well as with the broader host range S. Enteritidis that contaminates eggs. The genetic linkage of these serotypes may be with reproductive tract tropism in general, because S. Choleraesuis also clustered with this group. Although S. Choleraesuis is not an avian pathogen, it is strongly associated with porcine reproductive and respiratory syndrome (Wills et al., 2000; Chiu et al., 2005b). Finding that ISR-2 aligns S. Heidelberg with S. Enteritidis is a concern because S. Heidelberg is a more prevalent serotype in the on-farm environment of the chicken than is S. Enteritidis and there are some data to suggest that it too can contaminate eggs (Nayak et al., 2003; Gast et al., 2004; Johnson et al., 2005). The clustering of S. Newport with S. Paratyphi A and S. Typhi may have some biological consequence because S. Newport is sometimes associated with symptoms in humans that are more similar to those of enteric fever than to those of simple gastroenteritis (Saxena et al., 1990; Weir et al., 2004; Sukhnanand et al., 2005). Salmonella Newport has, in recent years, become a more prominent and persistent pathogen that is attributed to dairy products (Gupta et al., 2003).

ISR-2 helped to resolve the identity of those strains that were difficult to serotype. The Salmonellae are classified within subspecies enterica using the Kauffman–White scheme, which classifies the Salmonellae based on development of an immuno-precipitate formed between monospecific antisera and lipopolysaccharide O-antigens and flagellar H-antigens that are produced by any one strain (Popoff et al., 1997). Atypical serotypes that lack H-antigens and that otherwise produce lipopolysaccharide O-antigen B have been recently associated with increased human illness (Echeita et al., 1999; Tavechio et al., 2004). The ability of ISR sequence to resolve serotype definitively according to genetic lineage will aid classification of O-antigen B serotypes that are atypical. The inability to type reliably within S. enterica by H-antigens suggests that (i) current methods of producing typing antisera are not adequate, (ii) Salmonella O-antigen B serotypes have evolved that suppress at least one of two flagella structural genes and/or (iii) there is some masking of flagellar H-antigens that selectively spares O-antigen. Of these possibilities, there is increasing concern about the commercial availability of well-defined O- and H-antisera and about the propensity of some strains of egg-contaminating S. enterica to suppress flagellation or to produce antigen-masking capsules (Herrera-Leon et al., 2004; McQuiston et al., 2004). The availability of more definitive nucleic acid-based typing methods to resolve issues of genetic lineage and serotype should thus be a benefit when identity is in doubt or when appropriate antisera are not available.

In summary, the 23S–5S rrfH (ISR-1) and 5S-aspU (ISR-2) sequences located adjacent to rrfH of 72 Salmonella strains that included 10 known and two unknown serotypes of S. enterica as well as S. bongori were analyzed. This region provided valuable information about genetic lineages within S. enterica that have known associations with specific hosts and foods. However, inclusion of other regions of the Salmonellae genome will provide additional discrimination. Maximum discrimination from a minimal number of sequences is needed to obtain the most information from typing approaches that follow changes in the epidemiology of the Salmonellae. For example, neither rrfH ISR examined in this study distinguished between phenotypically variant subpopulations of S. Enteritidis that were derived from a single parent (Guard-Bouldin et al., 2004; Morales et al., 2005). By including other regions that are linked to virulence phenotype in combination with the ISR sequences analyzed herein, it may be possible to follow evolutionary trends within the Salmonellae that substantially impact epidemiological patterns of illness in humans.


  1. 1. These sequence data were produced by the Beowulf Genomics Sequencing Group at the Sanger Institute and can be obtained from

  2. 2. These sequence data were produced by the Beowulf Genomics Sequencing Group at the Sanger Institute and can be obtained from

  3. 3. GenBank accession number AE006468

  4. 4. These sequence data were produced by the Beowulf Genomics Sequencing Group at the Sanger Institute and can be obtained from

  5. 5. GenBank accession number AL513382

  6. 6. GenBank accession number CP000026

  7. 7. GenBank accession number AE017220

  8. 8. These sequence data were produced by the Beowulf Genomics Sequencing Group at the Sanger Institute and can be obtained from


The authors involved in this research have no commercial or other association that might pose a conflict of interest. Funding for this research was provided by ARS Project 6612-32000-042, ‘Molecular Pathobiology and Epidemiology of Egg-Contaminating Salmonella Enteritidis’. This research was previously presented in abstract form at the 2005 General Meeting of the American Society for Microbiology that met June 5–9, 2005, in Atlanta, GA (abstract number B-199).

This work would not have been possible without the policy of the Sanger Institute (Pathogen Sequencing Unit) and the National Center for Biotechnology Information to allow public access to genomic sequence pre- and postpublication.