OUP user menu

Genotypes of multidrug-resistant Salmonella enterica serotype typhimurium from two regions of Kenya

Samuel Kariuki, Joseph O. Oundo, Jane Muyodi, Brett Lowe, E. John Threlfall, C. Anthony Hart
DOI: http://dx.doi.org/10.1111/j.1574-695X.2000.tb01498.x 9-13 First published online: 1 September 2000


A combination of phage typing and pulsed-field gel electrophoresis of XbaI-digested chromosomal DNA has been used to study the epidemiological relationships of multidrug-resistant Salmonella enterica serotype typhimurium from Nairobi (64 isolates) and Kilifi (40 isolates) collected over the period 1994–1997. Isolates from Nairobi belonged to 11 definitive phage types (DTs) encompassing eight different PFGE patterns. In contrast, isolates from Kilifi were mainly DT 56 (60%) and all fell into a single PFGE pattern. The remaining isolates did not conform to a recognisable phage type. We conclude that multidrug-resistant S. typhimurium infections from Nairobi were caused by multiple strains while those from Kilifi were likely to be from a microepidemic caused by a single clone.

  • Multidrug-resistant typhimurium
  • Kenya
  • Phage type
  • Genotype

1 Introduction

Human infections due to non-typhoidal Salmonella spp. have been increasing since the 1980s and have been shown to be related to foodborne outbreaks worldwide. As only a few phage types tend to predominate within a certain geographical region, phage typing has been used with great success to trace outbreaks [1,2]. For example, recently, multidrug-resistant (MDR) Salmonella enterica serotype typhimurium (S. typhimurium) definitive phage type (DT) 104 has been implicated in most outbreaks caused by contaminated foods of animal origin in both USA and in Europe [1,3]. However, it has not always been possible to link an outbreak to a particular phage type. For example, during the period 1988–1992, DT 193 was the most common typhimurium phage type causing infections in humans in the UK, and most infections were traced to food animals, particularly cattle and pigs. However, studies of the distribution of insertion sequence IS200 elements confirmed that DT 193 was a heterogeneous phage type containing several distinct clones [4]. In other studies [5,6], pulsed-field gel electrophoresis (PFGE) of macrorestricted genomic DNA has proved useful in providing additional information for the epidemiological analysis of outbreak-related and unrelated Salmonella infections. The objective of the present study was to use phage typing and PFGE in order to analyse epidemiological relationships of multiple drug-resistant S. typhimurium isolated from patients from two different geographical areas of Kenya.

2 Materials and methods

2.1 Patients and bacterial isolates

Patients were febrile adults seen at the main public hospital in Nairobi (Kenyatta National Hospital) and children (1–84 months of age) seen at the District Hospital in Kilifi, Coast Province some 500 km from Nairobi. Kilifi District Hospital caters for a mainly rural population and has a specialised children's ward which was selected for the study. Blood cultures and stool specimens were routinely obtained from these patients in a sub-study evaluating the aetiology of fever in hospital admissions. A total of 64 non-duplicate Salmonella enterica serotype typhimurium (S. typhimurium) (55 from blood and 9 from stools) were obtained from patients at the Kenyatta National Hospital, Nairobi and 40 isolates (28 from blood and 12 from stools) were obtained children from Kilifi District Hospital. These isolates were selected for study on the basis of being multi-drug resistant (resistant to two or more antimicrobials) between 1994–1997, and were frozen at −70°C in trypticase soy broth with glycerol until analysed. Serotyping was performed following the Kauffmann—White scheme utilising Salmonella agglutination antisera (Murex diagnostics, Dartford, UK).

2.2 Phage typing

Bacterial isolates confirmed as S. typhimurium were sent to the Laboratory of Enteric Pathogens, Central Public Health Laboratory, London, for phage typing.

2.3 Antimicrobial susceptibility

Susceptibility tests were performed on Isosensitest (Oxoid, Basingstoke, UK) agar by the disk diffusion technique. Escherichia coli ATCC 25922 was used as the sensitive control. The antibiotic disks (Oxoid) used were: ampicillin 10 µg, tetracycline 30 µg, trimethoprim 5 µg, sulfamethoxazole 100 µg, chloramphenicol 30 µg, streptomycin 10 µg, gentamicin 10 µg, co-amoxiclav 20:10 µg, ceftazidime 30 µg, ciprofloxacin, 3 µg, nalidixic acid 10 µg. Disk zone sizes were interpreted according to the NCCLS guidelines [7].

2.4 Conjugation experiments and plasmid extraction

Conjugation experiments were performed in broth by the method of Walia et al. [8] with E. coli K12 (nalidixic acid-resistant) as recipient. Transconjugants were selected on MacConkey Agar (Oxoid) supplemented with nalidixic acid 32 mg l−1 and ampicillin 32 mg l−1. Plasmid DNA was extracted from the transconjugants by an alkaline lysis method [9]. Plasmids were separated by electrophoresis on horizontal agarose 0.8% gels at 100 V for 2 h. Plasmid sizes were determined by co-electrophoresis with plasmids of known sizes from E. coli strains V517 (53.7, 7.2, 5.6, 3.9, 3.0, 2.7, 2.1 kb) and 39R861 (147, 63, 43.5, 6.9 kb). DNA bands were visualised with an ultraviolet transilluminator (UVP, San Gabriel, CA, USA) after staining with ethidium bromide 0.05%.

2.5 PFGE of macrorestricted chromosomal DNA

Chromosomal DNA was prepared in agarose plugs as described previously [10] from an overnight culture in Luria broth. Agarose plugs were then equilibrated for 1 h in 0.5 ml of REACT2 buffer (Life Technologies, Paisley, UK) and incubated overnight at 37°C in fresh buffer (300 µl), containing 25 units of XbaI. PFGE of agarose plug inserts was then performed in a CHEF-DR II system (Bio-Rad, Richmond, CA, USA) on a horizontal agarose 1% gel for 22 h at 120 V, pulse time of 1–40 s, at 14°C. Lambda DNA concatemers consisting of a ladder (ca. 22 fragments) of increasing size from 48.5 kb to approximately 1000 kb was included as a DNA size standard. The gel was stained with ethidium bromide and photographed on an UV transilluminator (UVP). The RE digest patterns were interpreted by considering migration distance and intensity of all visible bands, and by using guidelines described by Tenover et al. [11]. Banding patterns were compared and genetic relatedness was determined by data clustering using the unweighted pair grouping arithmetic averaging method (UPGMA) (Molecular Fingerprinting Program version 1.4.1, Bio-Rad).

3 Results

3.1 Antimicrobial susceptibility

All 64 S. typhimurium from Nairobi and the 40 isolates from Kilifi were resistant to 2 or more drugs including ampicillin, co-trimoxazole, streptomycin, tetracycline and chloramphenicol. A large proportion of isolates from Nairobi (36; 56.3%) and Kilifi (13; 33%) were multiply resistant to ampicillin, chloramphenicol, co-trimoxazole and streptomycin.

3.2 Conjugation experiments and plasmids

A plasmid of ca. 100 kb was present in each of the multidrug-resistant S. typhimurium strains from Nairobi and Kilifi, in addition to several plasmids ranging from 5–42 kb. In conjugation experiments, all 64 multidrug-resistant S. typhimurium from Nairobi, and 40 isolates from Kilifi transferred plasmids of ca. 100 kb to E. coli K12. In each case, the 100-kb plasmids transferred resistance to ampicillin; in addition resistance to co-trimoxazole, tetracycline and chloramphenicol were transferred together in 19 (29.7%) and 9 (22.5%) donor-to-E. coli K12 conjugation tests for isolates from Nairobi and Kilifi, respectively.

3.3 Phage types of S. typhimurium

The 64 S. typhimurium from Nairobi encompassed 11 DTs. Most were DT 56 (17 isolates; 26.6%), followed by DT 193 (7; 10.9%) and 21 (32.8%) isolates were in eight other smaller phage type groups; 10 (15.6%) were untypable (these strains did not react with any of the currently available typing phages at the PHLS) and 9 (14.1%) were RDNC (these strains reacted with some typing phages, but did not conform to a recognised pattern) (Table 1). In contrast, 24 isolates (60%) from Kilifi were DT 56 and 16 (40%) were RDNC.

View this table:
Table 1

Phage types of Salmonella enterica serotype typhimurium from Nairobi and Kilifi, Kenya

Phage typeNumber of isolates
Nairobi (n=64)Kilifi (n=40)
  • S. typhimurium isolates not specific to any of the standard phages used for typing.

  • UNT, S. typhimurium isolates untypable.

3.4 Pulsed-field gel electrophoresis groups

The 64 isolates from Nairobi could be divided into 8 PFGE clusters (Table 2). Isolates within each cluster did not differ by more than four bands. Ten (58.8%) isolates of the most common phage type, DT 56, were in PFGE cluster 2 and the other seven were found to be distributed among isolates in four other PFGE clusters. More than a third of the isolates from Nairobi (24; 37.5%) were in PFGE cluster 2. This cluster contained ten different definitive phage types. Apart from PFGE cluster 7, which contained one untypable isolate and cluster 8 which contained isolates of a single phage type, all other PFGE clusters contained isolates from more than one phage type. When digested with XbaI, the 100-kb plasmids were shown to produce fragments of less than 20 kb which were not scored during the analysis. PFGE analyses of the 40 isolates from Kilifi showed that the banding patterns were similar for both DT 56 and RDNC as they fell into the same PFGE cluster 2. Thus, S. typhimurium from Kilifi were closely related, but not identical, to 10 out of 17 DT 56 isolates from Nairobi. In total, therefore, 64 (61.5%) of S. typhimurium from the present study were closely related by PFGE. The fragment patterns of XbaI-digested chromosomal DNA of MDR S. typhimurium isolates from Kilifi and Nairobi are shown in Fig. 1A,B.

View this table:
Table 2

Pulsed field gel electrophoresis cluster groups and phage types of S. typhimurium from Nairobi and Kilifi

PFGE cluster and (no. of isolates)Phage type and no. of isolates from
1856 (2), 135 (1), 204a (1), 193 (1), RDNC (1), UNT (2)
2274056 (10), 135 (1), 49 (2), 12 (1), 193 (1), 1 (2) 204a (1), 52 (1), 2 (1), RDNC (5), UNT (2)56 (24), RDNC (16)
3756 (2), 29 (1), 1 (2), RDNC (1), UNT (1)
4256 (1), UNT (1)
5956 (2), 204a (2), 193 (2), 135 (1), 12 (1), UNT (1)
68193 (2), 135 (2), 204a (1), RDNC (2), UNT (1)
71UNT (1)
82193 (1), UNT (1)
Figure 1

(A) Restriction endonuclease fragment patterns of XbaI-digested DNA from representative S. typhimurium isolates from Nairobi. Tracks 1 and 16 contained the 48.5-kb DNA molecular mass standard. Tracks 2–15 contained: 2, N150 (DT 56, cluster 2); 3, N3916 (DT 56, cluster 2); 4, N3344 (DT 56, cluster 2); 5, N815 (DT 135, cluster 3); 6, N7830 (DT 135, cluster 4); 7, N338 (DT 193, cluster 4); 8, N1487 (DT 135, cluster 4); 9, N592 (DT 1, cluster 5); 10, N710 (DT 1, cluster 5); 11, N617 (DT 12, cluster 3); 12, N844 (RDNC, cluster 4); 13, N2981 (RDNC, cluster 4); 14, N7118 (untypable, cluster 3); 15, N235 (untypable, cluster 2). (B) Restriction endonuclease fragment patterns of XbaI-digested DNA from representative S. typhimurium isolates from Kilifi. Track 1 contained the 48.5-kb DNA molecular mass standard. Tracks 2–21 contained: 2, K1 (DT RDNC); 3, K2 (DT 56); 4, K3 (DT 56); 5, K4 (DT RDNC); 6, K5 (DT RDNC); 7, K6 (DT 56); 8, K7 (DT 56); 9, K8 (DT 56); 10, K9 (DT 56); 11, K10 (DT 56); 12, K11 (DT 56); 13, K12 (DT RDNC); 14, K13 (DT RDNC); 15, K14 (DT 56); 16, K15 (DT RDNC); 17, K16 (DT RDNC); 18, K17 (DT 56); 19, K18 (DT 56); 20, K19 (DT 56); and 21, K20 (DT 56).

4 Discussion

A combination of phage typing and PFGE of XbaI-digested chromosomal DNA has been used to study S. typhimurium isolates from two geographically unrelated regions in Kenya. PFGE confirmed that MDR isolates from Nairobi were multiclonal, while those from Kilifi were essentially of the same clone. Although many reports of infection due to S. typhimurium in developed countries, such as in Europe [2,6], USA [3] and Canada [12] have been attributed to a single phage type often of an epidemic nature, our results indicate that this may not always be the case in developing countries.

Forty S. typhimurium isolates from Kilifi gave PFGE fragment patterns that were indistinguishable, suggesting that all isolates represented strains within the same clone. The present study has also shown that S. typhimurium DT 56 with a prevalence of 26.6% in Nairobi and 60% in Kilifi, was the most common phage type in both areas. In addition, S. typhimurium isolates from Kilifi that did not conform to a recognisable phage type (RDNC) produced PFGE fragment patterns that were indistinguishable from those of DT 56 strains suggesting that these RDNC strains were closely related to DT 56. Previous studies have shown that PFGE is superior to phage typing and antibiogram typing for Salmonella spp. [13]. For example, the acquisition or loss of lysogenic phages might result in a change in phage type. Thus, PFGE was able to further discriminate between strains of similar phage type. Using PFGE, our results also indicate that a number of S. typhimurium phage types are closely related, and that phage typing is useful in analysing strains that are only epidemiologically closely linked [3]. Combining PFGE patterns and phage typing for S. typhimurium isolates from Kilifi indicates the existence of a MDR clone of micro-epidemic nature that has been previously unrecognised in this fairly stable rural community at the coastal region of Kenya. In contrast, S. typhimurium isolates from Nairobi yielded a total of 11 phage types distributed within eight PFGE patterns indicating the existence of S. typhimurium strains of a multiclonal nature. It appears from our study that S. typhimurium from Nairobi are more diverse, even within the same phage type. Due to the fairly high population turnover and movement of the general population of Nairobi it is possible that S. typhimurium isolates come from different parts of the country and that they cause sporadic infections of a multiclonal nature rather than epidemic outbreaks.

In developed countries MDR S. typhimurium DT 104 with a common resistance phenotype of ampicillin, chloramphenicol, streptomycin, sulfonamides and tetracycline have been implicated in many cases of epidemic foodborne salmonellosis [2,3,11]. This phage type has remained fairly sensitive to newer generation cephalosporins and quinolones, which have been used as alternative drugs of choice for treatment of MDR S. typhimurium. However, a significant proportion of isolates of MDR DT 104 from the UK now shows decreased sensitivity to ciprofloxacin [14]. In the present study, S. typhimurium were resistant to all drugs commonly available in Kenya including ampicillin, cotrimoxazole, tetracycline, streptomycin and chloramphenicol. In addition these resistance phenotypes were self-transmissible on large 100-kb plasmids that was maintained across all the different phage types and PFGE clusters. The high prevalence of MDR S. typhimurium complicates the treatment and management of these infections in developing countries including Kenya where effective alternative drugs for treatment of salmonellosis are often unaffordable or limited in availability, thus leading to treatment failure. In addition, unlike in developed countries where S. typhimurium infections have often been associated with contaminated foods of animal origin, the natural reservoirs for non-typhoidal Salmonella spp. in developing countries including Kenya are not fully understood. Although there has been speculation that person-to-person transmission is important, case control studies need to be carried out in order to assess the risk factors for salmonellae infections in Kenya.

In conclusion, using phage typing and PFGE the present study has demonstrated that Salmonella infections due to MDR S. typhimurium in Kilifi were likely to be clonal in nature arising from a microepidemic in the region. In contrast, those from Nairobi were likely to be from sporadic outbreaks caused by a variety of strain types.


The Director of the Kenya Medical Research Institute for permission to publish this work. S.K. is supported by The Wellcome Trust Research Development Award in Tropical Medicine. We are grateful to Mrs. L.R. Ward from the Laboratory of Enteric Pathogens, CPHL for phage typing the isolates of S. typhimurium referred to in this study.


  1. [1].
  2. [2].
  3. [3].
  4. [4].
  5. [5].
  6. [6].
  7. [7].
  8. [8].
  9. [9].
  10. [10].
  11. [11].
  12. [12].
  13. [13].
  14. [14].
View Abstract