Phenotypic and genotypic analysis of clinical isolates Salmonella serovar Typhimurium in western Kenya

Background: Salmonella spp. are recognized as some of the most common causes of enteritis worldwide. This study aimed to identify clinically isolated S. Typhimurium in western Kenya and to assess antimicrobial resistance profiles and strain inter-relatedness. Methods: The study was performed in rural Maseno, Nyanza province in Kenya, between February 2004 and June 2005. Sixty-three patients with diarrhoea and fever were recruited. S. Typhimurium isolates were confirmed using serotyping, biochemical testing, and 16S rRNA sequencing. Susceptibility to 20 antimicrobials was determined and specific resistance genes were identified by polymerase chain reaction (PCR). Strain diversity was further analyzed using pulsed-field gel electrophoresis (PFGE), fluorescence amplified fragment length polymorphism (fAFLP), and multi-locus-variable-number-tandem regions (MLVNTR). Results: Twenty S. Typhimurium strains were isolated in the course of the study and their identity was confirmed by 16S rRNA gene sequencing. All 20 S. Typhimurium strains were resistant to ampicillin, streptomycin and sulfamethoxazole; ciprofloxacin resistance and phage DT104 were not detected. PFGE, plasmid profiling, and analysis of selected VNTR loci revealed further heterogeneity among the strains in the study. Conclusion: S. Typhimurium was commonly isolated from patients with diarrhoea and fever in Maseno. Considerable phenotypic and genotypic diversity was observed among isolates, suggesting that strains belonging to multiple lineages are responsible for disease in the study region. Multiple resistance was common and mediated by a variety of resistance genes but not by phage DT104.


Introduction
Salmonella enterica are widespread in humans and animals worldwide and are of increasing public health concern as causative pathogens of food poisoning [1,2].In many countries where sanitation is poor, typhoid and paratyphoid fevers, which are transmitted by the faecal-oral route, are major causes of gastric illness [3,4].Sanitation and hygiene are difficult if not impossible to implement in many developing countries, and unfortunately, the effectiveness of antimicrobial chemotherapy is also being eroded by the emergence of antibiotic resistance [5].Non-typhoidal human Salmonella diarrhoea does not warrant antimicrobial therapy; however, there are occasions when the infections can lead to life-threatening systemic infections that require effective chemotherapy [6].Of increasing concern is the worldwide emergence of multidrug-resistant phenotypes among Salmonella serotypes, in particular S. Typhimurium [4,7], that express resistance to ampicillin, chloramphenicol, streptomycin, sulfamethoxazole and tetracycline [4,5].In Kenya, multidrug resistance to commonly available antibiotics poses a major health concern, as alternative therapeutic choices are either unavailable or too expensive to be affordable for most patients [8].Multidrug resistant S. Typhimurium has been reported to be the predominant cause of bacteraemic illness in children in Zaire and Rwanda, while in Kenya it was the main isolate in adults with Salmonellae bacteraemia [9].Phenotypic identification of Salmonella by use of Kauffman's White Scheme, a classification system that depends on cell wall O antigens and the flagellar H antigen, has not been implemented in most laboratories due to different strains [10].Therefore, the use of genotypic identification typing methods based on the characterization of either plasmid or chromosomal DNA is encouraged and can help identify strains that may be clonal.The following methods, in particular, are being encouraged: plasmid profile typing, plasmid fingerprinting, determination of variablenumber tandem repeats (VNTR), and the identification of plasmid mediated virulence genes [11].However, reliable classification of clonal bacteria into genetically related groupings can be analyzed by sequence variation of multiple genes encoding housekeeping genes and the flanking region of the VNTR locus [12,13].Currently, many laboratories use pulsed-field gel electrophoresis (PFGE) to determine strain relatedness, to confirm outbreaks of bacterial diseases, and to identify the source of a strain or outbreak.Extensive use of antimicrobial agents in human and veterinary medicine has led to the increase of resistance among food-borne pathogens.Whereas antimicrobial drug resistance in zoonotic Salmonella may be associated with adverse consequences in several ways, treatment failures have been infrequently reported until recently [14,15,16], which may indicate an epidemic spread of multiresistant clones of particular serotypes of Salmonella.Not only do these clones have the potential to spread infection, but they also have the potential to develop additional resistance to the new antibiotics, which is the case for S. Typhimurium [17].
In Africa and most other developing regions, multidrug resistance, particularly to commonly available antibiotics, remains a major challenge for the health system [4].Inadequate sanitation to prevent strain dissemination and over-the-counter distribution of antimicrobials can exacerbate Salmonella infection.In their study, Kariuki et al. documented multidrug-resistant S. Typhimurium as the predominant cause of community-acquired bacteraemic illness in both children and in adults.They observed non-typhoidal infection as well as multiple resistance to commonly available antibiotics, including ampicillin, chloramphenicol, cotrimoxazole and tetracycline [4].
Bacterial resistance to aminoglycoside antibiotics, such as kanamycin, amikacin and gentamicin, is often mediated by enzymes that modify those drugs by acetylation, adenylation, or phosphorylation [18,19].The corresponding genes are often part of plasmids or transposons [18].In 1995, there was a substantial increase in multiresistant DT 104, showing resistance to ampicillin (A), chloramphenicol (C), streptomycin (S), sulfonamide (Sul), and tetracycline (T) [(ACSSuT)] mediated by a bacteriophage with additional chromosomally encoded resistance to ciprofloxacin,the drug of choice for the treatment of invasive Salmonellosis in humans [20].A study by Gorman et al. found that 98.5% of the S. Typhimurium isolated showed a high level of drug resistance, a result which correlated with those of investigations performed in the European Union [20].Multidrug resistance was observed in 88% of S. Typhimurium strains and the resistance profile for 77.6% of these strains was ampicilin, choramphenicol, streptomycin, sulfamethoxazole, and tetracycline (ACSSuT) [21].This antimicrobial resistance phenomenon, accompanied with a dearth of data from developing countries including Kenya, prompted this study on phenotypic and genotypic identification of S. Typhimurium clinical isolates.The study also aimed to assess their antimicrobial resistance profiles and gain insight to the mechanisms of resistance acquisition and spread.

Study site, design and patient inclusion criteria
The study was performed in a rural setting in Maseno, Nyanza province, Kenya.Consent to collect samples from the patients in the hospital was obtained from relevant authorities, namely, from Maseno University as well as from the Provincial Medical Officer and respective hospital superintendents.Between February 2004 and June 2005, patients with both fever (≥ 38°C) and diarrhoea who presented at Maseno and St. Elizabeth Mukumu Mission Hospitals and consented to participate in the study were recruited.In order to isolate as many Salmonella as possible, no age limit was set in the study.Blood, pus, cerebral spinal fluid, and stool samples were collected from a total of 63 patients who met the inclusion criteria.

Phenotypic isolation and identification of Salmonella species in blood, cerebral spinal fluid (CSF) and stool
Out of the 63 Salmonella samples collected, only 59 were viable when subcultured in Germany.Forty of the 59 samples were collected from stool and 12 were from blood.Four pus samples from peritoneum tissue were also collected from four of the patients whose blood had been collected but who were experiencing swollen abdomen and flatulence, and three from patients who were anaemic and convulsing.The collected samples were inoculated in Selenite-F medium (HIMedia Laboratories Pvt.Ltd Mumbai, India) and incubated at 37°C for 18 hours for maximum recovery of the isolates.To analyze for the presence of Salmonella, inocula were obtained from Selenite-F medium using a sterile cotton swab onto MacConkey agar (Oxoid No.3 CM 115 Basingstoke, England) media and incubated into Deoxycholate Citrate Agar (DCA) (HIMedia Laboratories Pvt.Ltd Mumbai, India).The isolates were then subcultured in Kliger Iron Agar (KIA) (HIMedia Laboratories Pvt.Ltd Mumbai, India) and Simmon Iron Medium (SIM) (Oxoid Basingstoke, England), Peptone water (Oxoid Basingstoke, England), phenylalanine media, and lysine agar slants.These steps were followed by biochemical and serotyping identification.For all the culture media used, a Salmonella-positive control obtained from KEMRI/CDC (ATTC 14028-American Type Culture Collection) was used to compare the isolates.The Salmonella-positive specimens were then subcultured in nutrient broth and stored in the refrigerator at 8°C for antibiotic susceptibility testing.

Genotypic isolation and analysis of Salmonella
Discreet S. Typhimurium colonies were picked from the saturated solid agar-disk plates using a sterile plastic inoculation loop, and DNA extracted using a QIAmp DNA mini kit (Qiagen, west Sussex, UK).For the presence of malic dehydrogenase (mdh) and definitive Type 104 (DT 104), all amplifications were performed in a final volume of 50 µl containing 0.5 µl of both primers (tables 1 and 2) as well as 2 µl Qiagen dNTP, 5 µl of 10 × Buffer, 1 µl MgCl 2 , 0.25 µl Qiagen Taq polymerase, 5 µg DNA containing sample and 36 µl water to make up the volume.The cycling conditions were as follows: 94 C for 5 minutes, followed by 35 cycles of 94 C for 25 seconds, 54 C for 45 seconds, 72 C for 45 seconds, and a final extension step at 72 C for 7 minutes.The amplicons were then loaded onto a casted 1.5% agarose gel (Eurobio, Les Ulis, France) run in TBE buffer (89 mmol l -1 Tris pH 8.3, 89 mmol l -1 borate and 2 mmol l -1 EDTA).The gel was stained with ethidium bromide solution (0.5 µg ml -1 ) and run at 135 V for 25 minutes then photographed under ultraviolet light (Gel Logic 100 Imaging System, Kodak).A negative and a positive control were included in each PCR reaction.
To compare the results, analysis of VNTR was also performed according to methods described by Lindstedt et al.Gel analysis was done as previously mentioned.

Pulsed-field gel electrophoresis of microrestricted chromosomal DNA
Twenty isolated Salmonella species were prepared for PFGE by individual suspension of the bacterial cells, grown on trypic soy agar at 37ºC.The cells were then transferred into a cell suspension buffer (100mM Tris-HCl, 100mM EDTA, pH 8.0) and measurements taken using spectrophotometer at an absorbance of 0.7 ± 0.05 at 612nm.Protease K (20 µl) was then added to 400 µl of the suspension along with 400 µl of molten (54ºC) 1% Seakem Gold Agar.These were mixed quickly, and approximately 300 µl were dispensed into prepared plug molds.Once solidified, the plugs were placed into 1.5 ml cell lysis buffer (50 mM Tris / HCl, 50 mM EDTA, pH 8.0, 1% Sarcosyl) and 40 µl of proteinase K and incubated for 1.5 hours at 54°C in a shaking water bath.The plugs were washed twice in ultrapure water for 15 minutes in a 50°C water bath followed by four washings in Tris -EDTA (TE) buffer (10 mM Tris -HCl, 1 mM EDTA, pH 8.0).The washed plugs were cut into 3mm × 9mm pieces and then digested in 173 µl of sterile water, 2 µl of bovine serum albumin, 20 µl of 10 × ReAct II buffer, and 5 µl of XbaI [New England Biolabs, Beverly, and Mass] (5' -TCTAGA -3') (10U/µl) at 37°C in a shaking water bath for 1.5 hours.The plugs were run in a 1% agarose gel using a CHEF III Pulsed-Field System (Bio-Rad) in 0.5% Trisborate -EDTA buffer (Sigma) at 10°C.The parameters were set with the initial switch time at 2.2 seconds, the final switch time at 64 seconds, a voltage of 6 V/cm, and a duration of 21 hours.Included on the gel were Xba I-digested plugs of S. Typhimurium to be used as size standards.The gels were stained with ethidium bromide and recorded on a Gel Doc System (Bio-Rad Laboratories, Inc., Hercules, CA).The file images were processed by BioNumerics software (Applied Maths BVBA, Kortrijk, Belgium).All the isolates within a PFGE subtype had identical bands; therefore, one isolate from each PFGE subtype was randomly selected as a representative and cluster analysis was completed on the subtypes by Dice similarity coefficient and 0.8% band position tolerance as recommended by BioNumerics [31].

Phage typing, plasmid detection and sizing
Bacteria isolates confirmed as S. Typhimurium were sent to the Robert Koch's Institute, Enteric Pathogen Unit, Germany, for phage typing and processed according to Threlfall et al. [32] and Scalzo et al. [33].Plasmid DNA extraction was performed using a plasmid Mini Prep Kit (Qiagen, west Sussex, UK) according to the manufacturer's instructions.
Plasmids were separated by electrophoresis on horizontal 0.8% agarose gels at 100 V for two hours.Plasmid sizes were determined by co-electrophoresis with plasmids of known sizes from E. coli strains V157 (NCTC 50193) (53.7,7.2,5.6,3.9,3.0,2.7,2.1kb) and 39R861 (NCTC 50192) (147,63,43.5,6.9kb).DNA bands were visualized with an ultraviolet transilluminator (UVP Inc) after staining with 0. Phenotypic and genotypic antimicrobial screening of the S. Typhimurium isolates Antibiotic susceptibility was evaluated according to the National Committee for Clinical Laboratory Standards, 2000 and 2001.Commercial antibiotic disks (HIMedia Laboratories Pvt.Ltd Mumbai India) of different antibiotic concentrations [(ampicilin, (10 µg/ml), chloramphenicol (30 µg/ml), tetracycline (30 µg/ml), streptomycin (10 µg/ml), ciprofloxacin (5 µg/ml), nalidixic acid (30 µg/ml) and cefotaxime (30 µg/ml)] were used (table 1).The zone of inhibition diameter of any growth observed was measured by use of Vanier callipers and compared with that of the positive control organism obtained from the Kenya Medical Research Institute (KEMRI), CDC Microbiology Unit (CDC 6516-60), and the American Type Culture Collection (ATCC 14028).MIC determination by use of VITEK R 2 [bioMeriux], a semi-automated system for 11 antimicrobials that works on MIC principles and can also distinguish between S. Typhi and non-Typhi isolates, was performed by obtaining the XDL bacteria growth with a sterile plastic loop and emulsified in 0.9% sodium chloride solution.DNA was extracted using a QIAmp DNA mini kit (Qiagen, west Sussex, UK) and amplified for various antimicrobial resistance genes using respective primers.Identification of class 1 integron gene cassettes was performed in 50µl final volumes containing 5 µl, 10 × Qiagen buffer, 10 µl, Q-buffer, 0.5 µl primer, 2 µl dNTPs ,0.5 µl Hot start Taq poly, 5 µl DNA template and topped with 26.5 µl water.All the amplification reactions were performed under the following conditions: hot start temperature at 95 C for 15 minutes, followed by 40 cycles of 94 C for one minute, 55 C for 45 seconds, 72 C for 30 seconds, and a final extension step at 72 C for 5 minutes.The same procedure was applied to bla PSE-1 , bla TEM , aphA1-lab, aac (6)-1, aadA, aadB, strB, sul A, and int genes.The gel was read as described earlier.
Molecular confirmation of S. Typhimurium fliC gene analysis specific for phase 1 gene (flagellin phase -1) and 2 flagellar (rfb cluster genephase 2) proteins (the H 1 and H 2 antigens respectively) observed during polyvalent "H" serotyping confirmed Salmonella species.Positive PCR results using the primer pair Fli 15 and Tym (ref.35 table 1) confirmed the molecular identification of the 24 strains as S. Typhimurium (table 2).However, four of the 24 strains produced a 312 bp band suggesting S. Enteritidis.Amplification of Salmonella housekeeping gene, malic acid dehydrogenase (mdh) (GenBank accession no.X61029) from 20 of the Salmonella isolates gave a specific band of 261 bp for all 20 S. Typhimurium isolates (table 2).Four out of the 20 strains were then randomly selected and sequenced for 16S rRNA.The 16S rRNA of the strains was found to be 99% homologous to the S. Typhimurium LT2 sequence since they had ≥ 99% sequence homology to a sequence entry at the

Bacteria Strains
Number

Discussion
These strains were identified as S. Typhimurium based on their O and H antigens, fliC, mdh, ST 11-ST 15, Fli 15 and Tym, Fli 15 and Ty04 genes (table 2).Thirty-nine other strains that did not conform to this test set were of the S. Typhi and S. Enteritidis groups [36,37].Molecular analysis by PFGE, MLVNTR, 16S rRNA and fAFLP demonstrated that the S. Typhimurium isolates were considerably heterogeneous (table 3).PFGE analysis showed that genetic variations are present among different phage types, suggesting diversification of serovar Typhimurium in Maseno and its environs.Similarly, the strains, which were multiply resistant, showed a broad range of antimicrobial resistance profiles.The presence of bla PSE-1 and bla TEM gene was not strongly associated with class 1 integrons.In isolates where bla PSE-1 was negative, bla TEM was positive and the isolates were resistant to trimethoprim.Class 1

692
integrons are known to capture and express mobile genes, known as cassettes, which in most cases are antibiotic resistant genes; this group of genes were commonly observed in S. Typhimurium isolates [38] (table 4).The chromosome-located apha-1 gene responsible for kanamycin resistance [39] was not detected in this study, though investigations done elsewhere have reported that genes coding for bla PSE-1 and bla TEM are clearly predominant and present as the only beta-lactamase gene in the absence of Apha-1 gene.Initially bla PSE-1 accorded bacteria resistance to ampicillin / kanamicin, but lately bla TEM has also been observed to play the same role.
Phage DT 104 was not identified in this study.However, DT104 was characterized by chromosomal resistance to ampicillin (A), chloramphenicol (C), streptomycin (S), sulfonamides (Sul), and tetracycline (T) and is commonly referred to as having resistance (R) type ACSSuT [40].It is thought that independent acquisition of transposons before the development/acquisition of the chromosomal multiresistance gene cluster ACSSuT might provide an explanation for the presence of more than one gene coding for the same resistance property in these strains, as documented by Frech et al. [37].It is also possible that for class 1 integron-positive strains, one gene was associated with integron, but the strain also harboured a plasmid containing the other resistance gene.The sul1 gene was not detected in the study even though most strains were sulphonamide resistant, an indication that sul1 gene is not a consistent maker for the presence of class 1 integrons.The acquisition of a multiple antibiotic resistance plasmid observed by PFGE may have been favoured by the selective pressure of antibiotics in the environment.Asymptomatic carriers, who were unnoticed within the community set-up, could provide an explanation for the existence and dissemination of these heterogeneous strains in our geographical area.However, in some instances, strains negative for class 1 integrons also contained two different resistance genes for the same antibiotic.
Chloramphenicol resistance was observed in most strains by agar disk diffusion, an indication of the presence of chloramphenicol acetyl-transferase activity.Though the presence of chloramphenical acetyl-transferase activity was not determined in this study, resistance to chloramphenical may be with the expression of the chloramphenicol transferase cat gene as documented by Faldynova et al.The isolates were resistant to sulfamethoxazole conferred by genes other than sul1, potentially sul2 or sul3, which were not tested in this study.The former is a likely possibility since, according to Guerra et al., the sul 2 gene often appears to be associated with genes that confer resistance to streptomycin (strA, strB) and all the isolates were positive for str B gene [42,43].
Our data reveals that the tested isolates in this study did not show the presence of phage DT 104 that is responsible to drug resistance.We therefore conclude that S. Typhimurium strains from western Kenya show considerable diversity and are also different from strains which originate from other parts of Kenya (Nairobi, Kiambu and Rift Valley) and around the world [4][5][6][7][8][9][10][11].

Table 1 .
[34]ml -1 ethidium bromide.Salmonella primers used were according to Soumet et al.[34](table1).Amplifications were conducted in a total volume of 50 µl, containing 0.5 µl of Hot start Primer sequences of various oligonucleotides used in the study.

Table 2 .
Number of Salmonella species using five primer pairs on different Salmonella strains.The 216bp fragment was amplified with the primer mdh d The 559bp fragment was obtained with primers Fli 15 and Tym b The 312bp fragment by the primer sef 167 and sef 478 (sef A) .e The 620bp fragment was amplified with the primers fliC 15 and Ty04 c The 429bp fragment was obtained with primer ST 11 -ST 15 BLAST search function located in the NCBI data base (accession AE008893 version AE 008893.1),but were not identical to each other.This acted as confirmation of the 16S rRNA (BLAST LT2), and phenotypic results obtained by serology, biochemical test and VITEK R 2. a

Table 4 .
Phenotypic and genotypic antimicrobial resistance profile of clinically isolated S. Typhimurium.