Prevalence, Detection and Antimicrobial Resistance Pattern of Salmonella in Sudan

Infectious microbial diseases constitute a major cause of death in many parts of the world, particularly in developing countries. Salmonella has been identified as an important food and water-borne pathogen that can infect human and animals resulting in significant morbidity and mortality (Akkina et al., 1999). Salmonella is a facultative anaerobe, Gram-negative rodshaped, 2 – 3 x 0.4 – 0.6 μm in size and motile by peritrichous flagella except for S. Gallinarum and S. Pullorum which are immotile. Members of the genus have a % G+C content of 50-53. They are urease and Voges-Proskauer negative and citrate utilizing (Montville and Matthews, 2008).


Introduction
Infectious microbial diseases constitute a major cause of death in many parts of the world, particularly in developing countries. Salmonella has been identified as an important food and water-borne pathogen that can infect human and animals resulting in significant morbidity and mortality (Akkina et al., 1999). Salmonella is a facultative anaerobe, Gram-negative rodshaped, 2 -3 x 0.4 -0.6 µm in size and motile by peritrichous flagella except for S. Gallinarum and S. Pullorum which are immotile. Members of the genus have a % G+C content of 50-53. They are urease and Voges-Proskauer negative and citrate utilizing (Montville and Matthews, 2008).
Salmonellae are typically non-lactose, non-sucrose fermenting but are able to ferment glucose, maltose and mannitol with the production of acid only as in the case of S. Typhi and acid with H 2 S in the case of S. Paratyphi and for most other Salmonella serovars (Cruickshank, 1975). Optimum temperature for growth is in the range of 35 -37ºC but some can grow at temperatures as high as 54ºC and as low as 2ºC (Gray & Fedorka-Cray, 2002). Salmonella grow in a pH range of 4 -9 with the optimum being 6.5 -7.5. They require high water activity for growth (> 0.94) but can survive at a w of < 0.2 such as in dried foods. Inhibition of growth occurs at temperatures < 7ºC, pH < 3.8 or a w < 0.94 (Hanes, 2003).
Based on differences in 16S rRNA sequence data, the genus Salmonella is divided into two species: S. enterica and S. bongori. S. enterica is further divided into six subspecies: subspecies enterica, salamae, arizonae, diarizonae, houtenae and subspecies indica (Popoff & Minor, 2001). Kauffmann-White scheme classifies members of Salmonella species according to three major antigenic determinants composed of somatic (O-antigens), flagellar (H-) and virulence (K-) antigens. Agglutination by antibodies specific for the various O-antigens, groups the salmonellae into six serogroups: A, B, C 1 , C 2 , D and E. Rarely cross reactivity between Oantigens of Salmonella and other genera of Enterobacteriaceae do occur. Therefore further classification of serotypes is based on the highly specific H-antigens (Scherer & miller, 2001). H-antigens can be expressed in one of two phases: phase 1 H-antigens are serovar specific while phase 2 antigens are not. However K-antigens are produced by serovars that are characterized by extracellular polysaccharide capsules (Hu & Kopecko, 2003;Yousef & Carlstrom, 2003). Currently, Kauffmann-White scheme recognizes 2610 Salmonella serovars, the majority (2587) belongs to S. enterica while the remaining (23 serovars) are assigned to S. bongori (Guibourdenche et al., 2010).
The incidence of typhoid salmonellosis is stable, with comparatively few cases in developing countries. Cases of typhoid salmonellosis are estimated to be in the range of 16-17 million resulting in about 600,000 deaths annually (Pang et al., 1995). However, cases of non-typhoid salmonellosis are increasing worldwide. The World Health Organization (WHO) estimates 1.3 billion cases with 3 million deaths annually. Data on salmonellosis cannot be ascertained in most developing countries because many patients with acute gastroenteritis do not visit a health care provider or do not submit a specimen for laboratory testing (Portillo, 2000;Hanes, 2003;Hu & Kopecko, 2003). An unpublished data of Sudan Ministry of Health indicated that the incidence of salmonellosis has markedly increased. In 2004 these were 43,144 cases and 68 deaths in different hospitals. However, a more accurate figure of salmonellosis is difficult to determine because only large outbreaks are investigated whereas sporadic cases are under-reported.
Since Salmonella is closely related to both human and animal health, more rapid and sensitive methods for the identification of this bacterium are required (Schrank et al., 2001). In the regions where enteric fever is common, clinical diagnosis of typhoid fever is inadequate, as the symptoms it causes are non-specific and overlap with those of other febrile illness. Serological tests, predominantly the Widal test, are available but have very low sensitivity and specificity, and no practical value in endemic areas despite their continued use (Levine et al., 1978). Isolation of the causative organism remains the most effective diagnostic method in suspected typhoid fever (Zhou & Pollard, 2010). Standard culture methods for detecting Salmonella spp. include non-selective pre-enrichment followed by selective enrichment and plating on selective and differential media (Whyte et al., 2002). These methods take approximately 4-7 days (Harvey & Price, 1979;Perales & Audicana, 1989), so they are considered laborious and time consuming.
In a clinical setting, suspected Enterobacteriaceae are often subjected to biochemical testing to investigate the ability of an isolate to grow upon certain substrates, produce various metabolic products or alter the pH. In the 1970s, efforts were being made to collect together multiple biochemical tests to allow rapid and relatively high throughput identification of clinical bacterial isolates . Numerous test kits of varying accuracy became available for the identification of the Enterobacteriaceae. The API 20E system (BioMérieux, France) was found to be the most reliable, having a 99% correlation with standard biochemical tests and a 94% identification rate . Over 60 bacterial species can currently be identified with the API20E, with identification extending to the serovar level for Typhi and Paratyphi A among others. However, results of such tests are open to interpretation and require experience and a high level of technical skill to generate reproducible results (Jamshidi et al., 2007). Therefore, several alternative faster methods for the detection of Salmonella have been suggested.
Resistance of infectious microorganisms to commonly prescribed antibiotics has emerged and spread in both developed and developing countries (Zhao et al., 2003;Ahmed et al., 2000;Grob et al., 1998). This imposes serious constraints on the options available for the treatment of many infections (Kunin et al., 1990). In the case of salmonellae, resistance to tetracyclines or chloramphenicol was first reported in 1961 (Ramsey and Edwards, 1961). Since then, reports on salmonellae resistance to one or more antibiotics have increased substantially and resistance has emerged even to newer more potent antimicrobial agents (Montville and Matthews, 2008;Piddock, 2002). In addition, multidrug resistance in Salmonella has become a public health concern (Crump and Mintz, 2010;Asai et al., 2010;Singh et al., 2010).
In the Sudan, as in most other developing countries, resistance and multiple resistance to antimicrobial agents among members of Enterobacteriaceae including some Salmonella serovars was found to increase during the last decades (Yagoub et al., 2005;Ahmed et al., 2000;Hassan, 1985, Shears et al., 1988Musa and Shears, 1998). The sensitivity of Salmonella Typhi, S. Paratyphi A and S. Paratyphi B to ten antibiotics was examined in Sudan (Ahmed et al., 2000). The examined strains were sensitive to all drugs tested except for one S. Typhi strain which was resistant to cotrimaxazole; tetracycline and sulfonamide and one S. Paratyphi A which was resistant to tetracycline. The sensitivity of Salmonella Paratyphi A and S. Paratyphi B which were isolated from Sudanese white cheese was tested against 9 antibiotics (Yagoub et al., 2006). ciprofloxacin, chloramphenicol and ofloxacin were the most effective drugs against the tested isolates. The resistance was more frequent to ampicillin, tetracycline, penicillin, gentamicin and co-trimoxazole.
Due to the indiscriminate and injudicious use of antibiotics in human and veterinary medicine as well as for the promotion of growth in food animals, Salmonella strains resistant to first line antibiotics will continue to develop at an increasing rate. Therefore updated knowledge of Salmonella serotypes resistance patterns is important for the proper selection and use of antimicrobial drugs and for the development of appropriate prescribing policies.

History of Salmonella research in Sudan
In Sudan, the prevalence of Salmonella serovars is not well documented, as salmonellae are not routinely isolated and identified. Only a few studies have been reported by few workers. Horgan (1947) made the first report on Salmonella infections in cattle. He investigated a food poisoning outbreak at Wad Madani town and isolated Salmonella serovar Dublin from faeces of two persons who fell sick after eating meat. Again the serovar Dublin was isolated from infected calves and from one of the apparently healthy animals (Soliman and Khan, 1959). A survey to ascertain the incidence rate of Salmonella infection in animals was made in Khartoum (Khan, 1970). During the survey, 230 Salmonella cultures were recovered from different sources belonging to 63 serotypes.  (Khan, 1970). In his attempt to assess the quality of fresh meats in Sudan, Sariy Eldin (1971) reported the occurrence of Salmonella Wein, S. Dublin, S. Havana, S. Typhimurium, S. Senegal and S. Braenderup. S. Dublin was also isolated from sheep liver (Salih and Ibrahim, 1972). Fifty-eight Salmonella strains were isolated from slaughtered chicken in Khartoum North and Omdurman (Yagoub and Mohamed, 1987). The most common serotypes reported were: S. Mons, S. Amek and S. Uganda. The incidence of S. Dublin in the mesenteric lymph nodes and faeces of sick calves in Kuku dairy cooperative farm, Omdurman and El Obeid slaughter houses was also reported (Saliem, 1987).
Forty-five Salmonella isolates (not serotyped) were isolated from carcasses, liver, spleen, intestinal contents of chickens from a poultry farm in El Obeid (unpublished data). The isolation of Salmonella enterica subspecies enterica serotype San-Diego from three goats (3.84%) at Omdurman Central Abattoir was reported (El Tom et al., 1999). Recently, Salmonella Umbadah plus 19 new serovars were reported from different sources at Khartoum (Hag Elsafi et al., 2009;El Hussein et al., 2010).

Isolation and identification of Salmonella
Salmonellae were isolated and identified according to the techniques recommended by the International Organization for Standardization described by Molla et al. (2004).
For confirmation, presumptive salmonellae were subjected to biochemical tests (Macfaddin, 1980), further identified with API 20E identification kits (Bio Merieux, Marcy, France) and a slide agglutination test was employed thereafter, using a commercially available Salmonella polyvalent O (Denkafeiken, Japan) and H antisera (Mast Diagnostic, UK). Presumptive Salmonella isolates were shipped to the Public Health Agency, Office International des´ Epizooties (OI´E) Reference Laboratory for salmonellosis, Guelph, Ontario, Canada or to the Egypt Management Central Laboratory for serotyping and phage typing.

Serotyping and phage typing
For serotyping, the somatic (O) antigens of the Salmonella isolates were determined with the slide-agglutination test as described by Ewing (1986), whereas the flagellar (H) antigens were identified by a microtechnique (Shipp and Rowe, 1980) that uses microtitre plates (Poppe et al., 2001). The antigenic formulae of Salmonella serovars as listed by Popoff (2001) were used to name the serovars.
The standard phage typing technique described by Anderson and Williams (1956) was used. Strains that did not conform to any recognized phage type were considered atypical (AT). Salmonella Enteritidis strains were phage typed according to Ward et al. (1987) with typing phages obtained from the International Centre for Enteric Phage Typing (ICEPT), Central Public Health Laboratory, Colindale, United Kingdom via the National Laboratory for Enteric Pathogens (NLEP), Health Canada, Winnipeg, Manitoba. The phagetyping-scheme and phages for S. Typhimurium, developed by Callow (1959) and further extended by Anderson (1964) and Anderson et al. (1977), were obtained from the ICEPT via the NLEP.

Antimicrobial susceptibility testing
The antimicrobial resistance of the isolates was tested against ten antimicrobial agents by the agar diffusion method with Mueller Hinton agar and antibiotic disks (Hi Media), following the guidelines of the National Committee for Clinical Laboratory Standards (NCCLS, 2000). E. coli ATCC 25922 was used for quality control. The categories susceptible or resistant were assigned on the basis of the critical points recommended by the NCCLS (2007).

Plasmid profiling analysis
For each isolate a single colony was grown overnight in 1ml LB media at 37ºC. The bacterial cells were harvested by centrifugation for 30s in a micro-centrifuge (Sanyo), the supernatant was discarded and the pellet was subjected to Plasmid DNA extraction according to alkalinedetergent method (Dillon et al., 1985). The plasmids extracted were stored at -20ºC till used for further analysis. Later, the extracted plasmid for each isolate was analyzed electrophoretically on a 0.8% agarose gel. The gel was then exposed to ultraviolet transillumination and photographed in a gel documentation system (Model GAS Uvitec. Product). One Kb DNA ladder (Invitrogen, Germany) was also used to determine the plasmids sizes.

Specificity of PCR primer sets
Ten specific primer sets (Invitrogen, Germany) each targeting a different gene were evaluated for their specificity and sensitivity to detect locally isolated Salmonella serovars. DNA from each Salmonella serovar and non-Salmonella strain was extracted according to the boiling -centrifugation method (Soumet et al., 1994). A single colony of a pure nutrient agar culture was grown overnight at 37ºC in 1ml Luria -Bertani broth. Bacterial cells were precipitated by centrifugation at 13,000 rpm for 5min in a micro-centrifuge (MSE, MSBo1o.cx2.5, Sanyo, UK). The supernatant was discarded and the pellet was re-suspended in 500µl deionized distilled water. The suspension was boiled for 10min in a water bath then immediately cooled on ice. Extracted DNA was then stored refrigerated at 4oC until used as a template for PCR amplification.
The extracted chromosomal DNA was amplified by an established PCR technique (Sambrook et al., 1989). PCR amplification reactions were carried out in 25 µl total volume of PCR mixture containing 5 µl of template DNA, 12.5µl of the PCR master mix (Promega) (50 unit/ml Taq DNA polymerase in an appropriate reaction buffer {pH 8.5}, 400 µM each dNTPs and 3mM MgCl2) and 0.1 µM of each of primer pair. DNA was amplified according to reaction conditions published for each primer pair in a thermal cycler (Techne/ Flexigene -biotech).
Appearance of the target band specified for each primer set on the 1.2% agarose gel under specified gel electrophoresis conditions is considered as a positive amplification product.

Sensitivity of PCR primers
To determine the sensitivity of each PCR primer set, a single colony of a pure culture of Salmonella Typhi was grown overnight at 37ºC in 10ml Luria -Bertani broth. After incubation for 24 hours ten-fold dilutions (10 to 10-9) of the broth culture were made. The viable cell count in each dilution was determined using plate count media. For each mixture 100µl was cultured immediately in a plate count media (Somasegaran and Hoben, 1985;Vincent, 1970). The cultures were incubated overnight at 37ºC. The numbers of formed colonies were then counted (each colony is considered to be formed by a single cell). The DNA from each dilution was extracted as previously described and was used as a PCR template for each primer set.

Prevalence
Out of 1921 collected and examined samples; 833 (43.4%) belonged to poultry, 680 (35.4%) to food items, 224 (11.7%) to human faeces, 107 (5.6%) to chlorinated drinking water and 77 (4%) to food animal faeces (Table 1). In Total, 213 (11.09%) Salmonella strains belonging to 54 different serovars were isolated. Of these, 210 were members of S. enterica subspecies enterica,   . The remaining 38 serovars were represented each by less than six isolates. S. Typhimurium isolates were phagetype 2 while five of S. Enteritidis isolates were 21a phenotype and one was an Atypical phenotype. Table 2 shows that the serovar Typhimurium was isolated from four of five sources examined whereas Stanleyville, Livingstone and Molade were isolated from three of the Sources. The remaining serovars were isolated each from two or only one source. The most common serovar of the poultry isolates (n=70) was S. Kentucky (24.3%) followed by Stanleyville (17.1%), Virchow (12.9%) and Blockley, Hadar and Alachua (8.6% each). S. Typhi (9 isolates) was the most common (9.7%) serovar among human faeces isolates (n=93) followed by Paratyphi B (6.5%), Agona and Kalina (5.4 % each). Among the food isolates (n=32) S. Stanleyville and Livingstone were the most common (15.6%) with 5 isolates each. The most common serovars among water isolates (n=10) and animal faeces isolates (n=8) were S. Molade (50%) and Uganda (25%), respectively. To the best of our knowledge, 21 of the serovars reported here were isolated for the first time in Sudan and were all from human source only. These included: S. Stanley, Okerara, Sarajane, Limete, Massenya, Edinburg, Isangi, Inganda, Java, Alabany, Kalina, Tudu, Ottershaw, Saugus, Amoundrenes, Lexington, Kisii, Ituri, Lagos, salamae and arizonae. Table 3 show that the highest frequency of resistance observed was to streptomycin (41.3%) followed by tetracycline (31.9%), gentamycin (28.2%), ampicillin (25.4%), nalidixic acid (22.1%), co-trimoxazole (17.4%), ciprofloxacin (8.9%), chloramphenicol (8%), norfloxacin (7.5%) and apramycin (5.6%). All isolates of the serovars Tudu, Montevideo, Meleagridis, Johannesburg, Umbadah, Poona, Kisii and one Rough isolate which were found infrequently (n ranging from 1-3) were susceptible to all of the antibiotics tested. Out of the 37 isolates which showed resistance against cotrimoxazole, 27 (73%) belonged to the serovars Sanleyville (11/18 isolates), Virchow (10/10 isolates) and Typhi (6/9 isolates). Similarly, out of the 68 tetracycline-resistant isolates, 40 (58.8%) belonged to the serovars Stanleyville, Virchow, Blockley and Kentucky, each contributing by 10 isolates. Isolates of S. Blockley (n=10) were all resistant to tetracycline but were fully sensitive to co-trimoxazole, chloramphenicol, ciprofloxacin, ampicillin, norfloxacin and apramycin. All S. Kentucky isolates (n=19) were resistant to nalidixic acid but fully sensitive to chloramphenicol. Of the 22 isolates belonging to S. Typhi (9), Paratyphi B (6) and Typhimurium (7), 14 were resistant to ampicillin, 6 to gentamycin, 4 to chloramphenicol and 3 to each of ciprofloxacin and nalidixic acid but none to norfloxacin.  Twenty-one isolates belonging to ten serovars showed resistance to four antibiotics and exhibited ten resistance patterns. Of these, three patterns (TeCnNaSt, NaCipNorAmp and NaCipNorApr) were shown by eight isolates of S. Kentucky. One of the patterns, TeCnNaSt, was displayed by five of the 19 S. Kentucky isolates while the three isolates of S. Molade and two of S. Typhi shared the pattern TeStCotAmp. The other six patterns were shown by eight isolates belonging to the serovars Edinburg (2), Paratyphi B (1), Rissen (1), Java (1), Livingstone (1), Enteritidis (1) and Albany (1) Kentucky and one isolate of S. Isangi. They were all resistant to ampicillin and ciprofloxacin and showed three different resistance patterns. The smallest plasmid (1.2 Kb) detected in three isolates of each of S. Molade and S. Stainleyville. These isolates were resistant to tetracycline, co-trimoxazole and ampicillin. It is also clear that, all of the ciprofloxacin resistant isolates were found to contain at least one of the largest plasmids (12.2, 12.5 and 15.6 Kb). The 12.2 Kb plasmid was present, either alone or in combination with plasmids of other sizes, in 15 (40.5%) of the different resistance patterns.

Detection by PCR
DNA extracted from 213 Salmonella strains and 12 closely related non-Salmonella strains were used to evaluate the specificity and sensitivity of ten primer sets to detect Salmonella sp. using the polymerase chain reaction technique ( Table 5). The primer pairs targeting invA, hilA, iroB, oriC, fimA, hitJ and stn genes successfully amplified the DNA extracted from all Salmonella isolates generating the specific amplicon for each primer, and no amplification products were detected with DNA from non-Salmonella strains. These primers were, therefore, recommended as reliable means for simple and rapid PCR -based detection of the locally isolated Salmonella serovars.

Primer
Salmonella isolates (n=213) The ompC primer set produced the expected amplicon with the DNA extracted from most of the Salmonella isolates, but failed to do so with DNA of S. Molade and S. Meleagridis. Negative results were also obtained with all non-Salmonella strains. The primer set 16S rDNA generated target size amplicons with all Salmonella isolates. However, similar amplicons were produced from the DNA of some non-Salmonella strains including Proteus sp., Shigella dysenteriae, Citrobacter freundii and Pseudomonas aeruginosa. The repetitive fragment of DNA primer set failed to amplify the DNA of 2 isolates of each of S.

Discussion
Salmonella was isolated from different sources with an overall prevalence of 11.09%, which is comparable to the isolation frequency (9.2 and 11.2%) reported by El Hussein et al. (2010) in Khartoum State, Sudan and by Mohammad et al. (2006) in Zahedan, Iran respectively. The prevalence percentage reported demonstrates the widespread occurrence and distribution of Salmonella in Sudan. This prevalence was clearly higher than the range of 3.34 -4.0%, reported in the limited studies conducted previously (Soliman and Khan, 1959;Khan, 1962Khan, , 1970Yagoub and Mohamed, 1987;El Tom et al., 1999;Yagoub et al., 2006 andHag Elsafi et al., 2009). Several studies in other developing countries have reported a higher overall prevalence of Salmonella (human, food, and animal) such as 68.2% in Ethiopia, 51.2% in Argentina , 25.9% in Korea, and 72% in Thailand (Cardinale et al., 2003). It is important to recognize that the prevalence and distribution of Salmonella serovars varies from region to region (Dominguez et al., 2002;Uyttendaele et al., 1998) and isolation rates depend upon the country where the study was carried out, the sampling plan and the detection limit of the methodology (Roberts, 1982;Uyttendaele et al., 1998). Consequently, it is difficult to make comparisons between Salmonella surveillance conducted in different countries. However, the serovars isolated from the various sample types in our survey in Sudan were comparable to the results reported by various investigators in other countries. For example, Baudart et al. (2000) reported the prevalence of S. Agona, S. Enteritidis, S. Infantis, S. Mbandaka, S. Muenster, S. Rissen, S. Typhimurium, S. Montevideo and S. Virchow in different aquatic environments. The same serovars plus S. Senftenberg were isolated by Saha et al. (2001) from faecal samples taken from hospitalized diarrhoeal children in India. Similar to this study, Liebana et al. (2001Liebana et al. ( , 2002and Tamada et al. (2001) reported the isolation of S. Mbandaka, S. Montevideo and S. Livingstone from animal sources. The findings of Delicato et al. (2004) and of Fernandez et al. (2003) were also comparable to our results. They reported the isolation of S. Enteritidis, S. Infantis and S. Typhimurium from human faeces. S. Typhimurium and S. Enteritidis, in particular, are regarded worldwide as significant pathogenic serovars, with certain phagetypes being associated with serious illness in humans, chickens and animals (Dominguez et al., 2002;Jorgensen et al., 2002;Roy et al., 2002& Mohammad et al., 2006. The predominant phagetype of S. Typhimurium was PT2, while those for S. Enteritidis were PT21a and Atypical. In Western Europe, phagetype 4 was generally reported as a dominant phage for S. Enteritidis (Humphrey et al., 1991), while in the United States, PT8 was generally dominant (Hickman-Brenner et al., 1991).
The most predominant serovar in this study was S. Kentucky constituting 8.9% of the recovered isolates. The ranking of this serovar among other serovars increased substantially during the past decade in most European countries Gill and Threlfall, 2007;Bonalli et al., 2011). However, this serovar was considered an unsuccessful pathogen because it was rarely associated with human illness (Collard et al., 2007). S. Kentucky and S. Stanleyville were mostly isolated from poultry samples in comparison with other serotypes (24.3 and 17.1%, respectively). In chickens, it is well established that S. Enteritidis is the most predominant Salmonella serovar, followed by S. Typhimurium (Mohammad et al., 2006;Suresh et al., 2011). However, no S. Enteritidis and only one isolate of S. Typhimurium was isolated from poultry in this study.
Many animal species harbour Salmonella and can act as potential reservoirs for human infections. For example, S. Menden, S. Enteritidis, S. Montevideo and S. Senftenberg, which were recovered from humans in this study, were previously isolated from different animals in Sudan (Khan, 1970). Salmonella may enter the food chain through carcass contamination with animal faeces at slaughter and during processing, or through food or food handlers. However, human infection may also occur through contaminated water, pets, and exotic animals. Measures taken to control these routes of transmission are an effective way of preventing salmonellosis. The collection of prevalence data of Salmonella serovars is an important component of a successful epidemiological surveillance for public health management in any country.
Since all human isolates were recovered from stool samples of food handlers, it is likely that food can act as potential reservoirs for salmonellosis epidemics. More strict measures should be implemented in the food industry to curtail the spread of salmonellosis in Sudan. For example, monitoring of restaurant workers should be performed at more frequent intervals rather than the current mandatory annual check-up for renewal of work permits. This should be accompanied by organized training programs involving suppliers of food items and corporate administrators as well as front line restaurant employees. More involvement by public health authorities in surveillance programs is needed to ensure that public safety regulations are properly implemented.
This initial survey has provided useful information about the status of salmonellosis in Khartoum, Sudan. We have demonstrated that Salmonella was isolated mostly from humans followed by chickens suggesting that chicken and chicken-food products could be a potential source of salmonellosis in the food chain. The S. Kentucky serovar, which has a lesser potential for infection, was common to human, chicken and food items. Furthermore, S. Kentucky isolates are known for their resistance to ciprofloxacin, the antibiotic of choice for the treatment of typhoid in Sudan (Bonalli et al., 2011). Thus this phenotype may spread to other serovars which might have greater potential for infection.
The increased level of drug resistance in Salmonella has become a public health concern (Pui et al., 2011). As antibiotic usage varies among countries, different resistant phenotypes and genotypes can be expected. Thus monitoring of antimicrobial resistance patterns from different sources and regions is an important issue.
In this study, the rate of resistance to ampicillin (25.4%) and chloramphenicol (8.0%) was comparable to those detected for the same antibiotics in the year 2000. At that time, Ahmed et al. (2000) reported 25.0% and 13.0% resistance to ampicillin and chloramphenicol, respectively. However, the percentage of Salmonella isolates resistant to nalidixic acid and ciprofloxacin in Sudan has increased from zero percent (Ahmed et al., 2000) to 22.0 and 8.9%, respectively. The wide resistance to Nalidixic acid is a matter of concern, since nalidixic acid resistance has been associated with a decrease in susceptibility to fluoroquinolones, including ciprofloxacin, which are used to treat salmonellosis in humans (Gorman and Adley, 2004). The human isolates were more resistant to ciprofloxacin (16.1%) than the poultry isolates of which only 5.7% showed resistance to the antibiotic. Our results are consistent with what has been reported by Tassios et al. (1997), but disagree with the findings of Al-Bahry et al. (2007) who found that isolates of human origin were less resistant than those of chicken. In this study, resistance was 41.3% to streptomycin and 31.9% to tetracycline which is inconsistent with a report of complete susceptibility, to both antibiotics, by Singh et al. (2010) in India. Our results are in line with other investigators who observed high rates of resistance to streptomycin and tetracycline by Salmonella isolates (Zhao et al., 2005;Stevens et al., 2006;Dogru et al., 2009;Iseri and Erol, 2010). De Oliveira et al. (2010) attributed the prevalence of resistance to streptomycin and tetracycline to their frequent administration in veterinary medicine.
A lower percentage of resistance was observed for norfloxacin (7.5%), chloramphenicol (8.0%) and ciprofloxacin (8.9%). Lower resistance rates to these drugs were also reported by other workers (Gulsen et al., 2004;Zhao et al., 2006;Lestari et al., 2009). Therefore, these drugs may continue to be the drugs of choice for treatment of human salmonellosis in Sudan. apramycin, to which only 5.6% of Salmonella isolates were resistant, can replace the currently used antibiotics (mostly streptomycin and tetracycline) in the treatment of poultry.
Reduced susceptibility to fluoroquinolones and chloramphenicol has also been reported particularly for S. Kentucky isolates in many countries of the region such as Ethiopia (Molla et al., 2006), Morocco (Bouchrif et al., 2009) and Tunisia (Turki et al., 2011). Weill and Le Hello (2008) indicated that the acquisition and spread of distinct antibiotic resistance, especially resistance to ciprofloxacin, the drug of choice in Sudan, is associated with the emergence of S. Kentucky. Turki et al. (2011) reported a correlation between resistance to nalidixic acid and ciprofloxacin. This was not the case in our study; instead, nalidixic acid resistance closely correlates with tetracycline resistance. Strong evidence of cross resistance between chloramphenicol and ampicillin was also observed. Of 17 chloramphenicol resistant isolates, 12 (70.6%) were also resistant to ampicillin. Values of 100%, 85% and 21% cross resistance between these two antibiotics was reported by different investigators (Goldstein et al., 1983;Rowe et al., 1992;Gupta et al., 1990).
This study demonstrated that 35.7% of the Salmonella isolates tested (n=213) were multidrug resistant. This level was comparable to that reported by Bouchrif et al. (2009) in Morocco (44%) and Van et al. (2007) in Vietnam (34%) but much lower than the level reported by Thong and Modarressi (2011) in Kuala Lumpur (67%). Among all Salmonella isolates from different sources, 37 antibiotic resistance patterns were detected, indicating wide spread multidrug resistance. This result is comparable with the findings of Singh et al. (2010) who identified 24 different resistance patterns among their Salmonella isolates. The authors attributed the observed high number in multidrug resistance patterns to the frequent use of antibiotics in the environments from which the isolates originated. It should be emphasized that most of the MDR isolates (n=76) in this study originated from human (42) and poultry sources (26). Nowroozi et al. (2004) demonstrated that indiscriminate use of antibiotics in poultry production has increased the emergence and maintenance of MDR bacteria in the environment. Similarly, Ahmed et al. (2000) attributed the high level of multiple resistance salmonellae among human isolates to the inordinate and irrational use of antimicrobial agents in the Sudan. The authors mentioned a number of factors that have led to the prevalence of antibiotic resistance. The most important among which are the deliberate self-administration of antibiotics by patients themselves, the wide use of antibiotics due to the high prevalence of infectious diseases, lack of laboratory support in rural areas and selective prescribing due to cost constrains.
Plasmid DNA was demonstrated in 67.1% of the multiple resistant isolates. Our results indicated the association of large integrons (12.2 -15.6 Kb) with the resistance to ampicillin and/or ciprofloxacin. Certain plasmid sizes may be responsible for resistance to particular antibiotics (White et al., 2001). Resistance to ampicillin in Salmonella was mediated by a βlactamase gene carried on a plasmid (Thong & Modarressi, 2011). Since some of the isolates were resistant to one or more antibiotics and yet did not harbour any plasmid, the antibiotic resistance might be chromosomally mediated or mediated by other mobile elements such as transposons (Yah & Eghafona, 2008). The finding of both plasmid and non-plasmid mediated antibiotic resistance is consistent with other studies (Rodrigue et al., 1992;Ansary et al., 2006). This implies that there is no consistent relationship between the detected plasmid profile and antibiotic resistance patterns. Similar conclusions were previously drawn by Shears et al. (1988) in Sudan and Mamun et al. (1993) in Bangladesh.
In summary, this study confirmed that food, water, poultry and food handlers might act as reservoirs for antimicrobial resistant Salmonella. To diminish contamination rates by Salmonella, risk reduction strategies such as training of food handlers regarding food safety, institutional periodic focused medical checkups for food handlers, more restrictions on the irrational use of antibiotics, and establishment of standardized monitoring systems for on-going drug resistance surveillance are required. Although the current study addressed a local problem in Sudan where it appears that there is a high incident of antibiotic resistant salmonellae, many other developing countries are faced with a similar situation (Al-Bahry et al., 2007;Turki et al., 2011;Thong and Modarressi, 2011). In addition, dissemination of information on antibiotic resistance is important for development of prescribing polices and for general medical practitioners in a remote area who may not have access to microbiology laboratory back-up and, hence, must depend on prevailing knowledge of antibiotic-resistant Salmonella.
Results of the primers pairs targeting invA, hilA, iroB, oriC, fimA, hitJ and stn genes indicated their reliability, sensitivity and accuracy for the detection of Salmonella enterica. These findings affirm results of the previous studies (Moganedi et al., 2007;Murphy et al., 2007;Narvaneni and Jamil, 2005;Moore and Feist, 2006). It has been shown that invA and hilA genes are associated with pathogenicity of Salmonella spp. The former gene was reported to be essential for the invasion of epithelial cells by Salmonella (Moganedi et al., 2007); consequently, all Salmonella isolates studied are capable of invading the epithelial cells of the host (Bajaj et al., 1995). The hilA gene proved to be highly conserved among our Salmonella serovars. Being unspecific, the repetitive fragment of DNA, 16S rDNA and opmC primers were considered as unsuitable for PCR detection of locally isolated Salmonella serovars. The positive results obtained by the primer set 16S rDNA with some non-Salmonella strains could be explained by the fact that this primer set is constructed from 16SFl and 16SIII derived from the two regions of 16S rRNA gene. 16SF1 is the reverse and complementary strand of 16SI, which has been found to hybridize with Salmonella as well as with Citrobacter sp., while the 16SIII sequence was able to hybridize with Klebsiella and Serratia spp., in addition to Salmonella (Lin and Tsen, 1996). Similarly, the repetitive DNA fragment PCR assay was found to be unspecific for Salmonella as it failed to amplify the DNA of a number of Salmonella isolates and positive results were obtained for most of the non-Salmonella strains. Comparably, Ziemer and Steadham (2003) reported the ability of this primer set to amplify the DNA of bacteria commonly associated with intestinal samples.