Abstract: Different Salmonella species (159 strains) were isolated from human and non-human sources were exposed to eighteen different antibiotics. 48 strains showed resistance to one or more antibiotics. Resistant strains were examined for their R- plasmids contents by glass fines, boiling miniprep and agarose electrophoresis procedures. All but three of the tested strains from chicken contained a plasmid with similar molecular weight (m.w.) even though the strains were resistant to different antibiotics. The remaining three strains contained an additional plasmid with a different electrophoretic migration pattern. Many of the plasmids when digested with endonucleases were found to have similar fragments. Out of four plasmids extracted from Salmonella strains isolated from sheep, three contained one type of a plasmid with similar m.w. and fragmentation. Broad diversity was noted in plasmids extracted from Salmonella strains isolated from humans. Endonuclease digestion of plasmids revealed that some plasmids have a common fingerprint pattern.
Introduction
The Abuse of antibiotics around the globe for preservation of food materials, treating animals with different kinds of diseases, or for preventive purposes have selected far antibiotic resistant strains. These strains, which exhibit drug resistance, continue to multiply when challenged with an antibiotic (Ryder et al., 1980). Resistance to antibiotics may be acquired by mutation of chromosomal genes or by conjugation which leads to the selection of multiple resistant bacteria. Certain constitutive mutations cause a bacterial strain to lose some important parts of its cell surface proteins. For example, some S. typhimurium which has lost surface Ompc porin increased several fold resistance to cephalosporins (Medeiros et al., 1987). Other resistant strain to ferrichrome analogue antibiotics, were shown to be defective in their iron transport system (Neilands, 1982).
Unlike chromosomal mutations, conjugative plasmids (i.e. R-factors) can transfer genes encoding for multiple resistance and to mobilize other non-conjugative plasmids and chromosomal genes to sensitive host cells (Saxena et al., 1984). Accordingly, plasmid profiling can be used as an epidemiological tool to follow up the source of bacterial strains that cause outbreaks. Little information from Oman is available on Salmonella-drug susceptibility and nothing is known about their plasmid content. Therefore, results obtained in this investigation will provide an important baseline epidemiological data for Salmonella of Oman.
Materials and Methods
Out of 159 salmonellae strains isolated previously from human and non-human sources. 48 strains were resistant to antibiotics, mawy of which with similar antibiotic resistant patterns and therefore were classified into four (I-IV) resistant groups (Al-Bahry, 1999). Three strains from each of the four groups were analyzed for their plasmid contents. All other resistant strains that differed in their resistance patterns were not among those four groups. These strains were also analyzed for their plasmid content, forming a total of 26 strains. The strains were grown in Luria Bertani broth (LB), containing a challenging antibiotic (25 mg/ml) which the isolates were resistant to. Plasmids were isolated according to glass fine boiling method (Davis et al., 1986), The E. coil K-12 containing PVW3 (12.5 Kb) plasmid was used as a control to check plasmids isolation procedure used for Salmonella. The plasmids pRL149 (5.4 Kb) and pUC13 (2.7 Kb) were kindly provided by Dr. Kailas of University of Wisconsin, Oshkosh, USA. All of the three standard plasmids were used to estimate the size of the unknown extracted plasmids of the resistant strains of Salmonella. Gel electrophoresis was carried out on 0.7% low melting point agarose grade gel. The plasmids isolated from Salmonella were digested with Eco RI and Hind lII (Parker and Seed, 1980).
Results
Since there are 4 groups (group I, II, III and IV) from non-human isolates (Table 1) with a similar antibiotic resistance pattern, three strains from each group were analyzed for their plasmid content. All other resistant strains from human and non-human sources that differ in their resistance pattern from the four groups were also analyzed for their plasmid content (Table 2, Fig. 1). Although salmonellae that were isolated from non-human sources were resistant to 3-8 antibiotics (Table 1), most of the strains contained closely related plasmids with sizes ranging from 28-32 kb (Table 2). Only one plasmid was detected from each strain of Group I, II, IV, and antibiotic resistant of ungrouped strains. However, group ill, strains 7, 8 and 9, contained an additional small plasmid with a (molecular weight) mol wt of 9 kb. The number of plasmids from resistant strains isolated from humans varied from (1-6) and their sizes ranged from 2.2-30 kb (Table 2, Fig. 1).
All plasmids from different salmonellae were listed as plasmid types after the strains from which they were extracted. Some strains contained more than one plasmid. These plasmids were numbered alphabetically as shown in Table 2. For example, strain 47 consisted of 2 plasmid types that were named 47a and 47b.
Although plasmids were extracted from three non-human strains from each of the four groups (Group I-IV) based on the antibiotic patterns and because of the m.w. of the extracted plasmids within a group were identical, only two plasmid types were digested with endonuclease. All the human strains were also digested with Eco RI or Hind III because of their dissimilarities. A total of 38 plasmids were extracted from 26 strains of Salmonella. The plasmid type 46,, plasmid size 2.2 Kb could neither be digested with Eco RI, nor with Hind Ill. Most endonuclease fragments of the plasmids that were extracted from salmonellae isolated from non-human strains were found to be closely related in terms of their mol wt fragmentation, and the total mol wt of the fragments (Table 2).
| Table 1: | Antibiotic Resistance Pattern of Salmonella Strains |
| *Antibiotic resistant of ungrouped strains | |
| Fig. 1: | Plasmid profiling of Salmonella resistant strains to antibiotics |
| Table 2: | Plasmid DNA Bands of Antibiotic Resistant Salmonella Strains |
* = All plasmids were cleaved by Eco RI unless mentioned otherwise, b= Cleaved with Hind III, * = Not cleaved by Eco RI or Hind III |
|
Of the strains isolated from chicken, that were digested with endonuclease, strains 1, 3, 4 and 5 had exactly the same number of plasmids, with similar mol wt of fragments. Whereas strain 6-8 each had a similar plasmid with a mol wt of 31 kb. Additionally Strain 7 and 8 had another plasmid identical to each other with a mol wt of 9 kb. The strains 32 and 33 isolated from sheep (Group IV) also had similar antibiotic pattern, and plasmids with identical mol wt and fragmentation (Table 1 and 2). The sheep strain 34 (ungrouped) showed different antibiotic resistant pattern and plasmid of different mol wt and fragmentation.
The human strains varied considerably in their antibiotic pattern and their plasmids also varied in mol wt and fragmentation (Table 1 and 2), However, the digestion fragments of plasmid types 40c, 40d of strain 40 were found to be closely related as were plasmid types 45a and 45b of Salmonella strain 45. These four plasmids with somewhat similar fragment sizes had identical total mol wts of their fragments.
The strain number 40 showed resistance to 10 antibiotics and number 45 to only 4 (Table 1) and yet the two strains had common antibiotic resistance to CKSmax, possibly indicating a common ancestry.
Although 38 plasmid types were digested with endonucleases, for simplicity the gel photograph shows plasmid for 11 strains only (Fig. 1). The 11 strains were selected based on similarities and dissimilarities in fragmentation (Table 2). Endonuclease fragments of the plasmids that had similar plasmid mol wt, revealed similar size of each fragment and similar migration in the gel (Fig. 1, strain 7a and 8a). Resistant salmonellae isolated from chicken were found to be closely related in terms of their fragmentation and mol wts, so were the sheep strains. The digested plasmids from human Salmonella strains showed broad diversity in terms of mol wts and fragmentation represented by plasmid type 37, 38, 40c and 46c,d,a, 47a, 47b and 48a (Fig. 1).
Discussion
Of the 159 different strains of Salmonella, 48 strains (30%) were resistant to at least two antibiotics. Interestingly, Salmonella strains that were isolated from sheep gave resistance pattern of KMinTe which is very similar to the resistance pattern of many Salmonella strains isolated from chicken. This indicates the possibility of cross infection between sheep and chicken and of common origin of the resistant strains (Al-Bahry, 1999).
Although no clear similarities were found in this study between the antibiotic resistant patterns and the plasmids isolated from salmonellae from animals and humans, it does not exclude the possibility that resistant strains from animals could not be transferred to humans. Both pathogenic and non-pathogenic strains resistant to drugs may be transported from animals to humans via food (Linton et al., 1977). Such strains act as an important source for in vivo transmission of R-plasmids to drug sensitive strains in the human intestine (Platt et al., 1986). Undoubtedly, the improper and unnecessary use of antimicrobial drugs in man also promotes development of resistant strains with R-plasmids.
The human and animal isolates did not exhibit a common antibiotic resistant pattern or common plasmid profile. This is probably because of the time frame of isolation of salmonellae and the geographical distance that separated the two sources at time of isolation (human and non-human sources). The dissimilarity in resistance pattern and plasmid profiles between human and non-human isolates does not disapprove that R-plasmids from bacteria in animals are not transferred to humans. The common antibiotic resistance pattern, and transfer of R-plasmids from animals to human strains may be demonstrated by proper sample collection during outbreaks of antibiotic resistant salmonellae. In such outbreaks, immediate action must be carried out to isolate resistant strains in order to analyze their antibiotic pattern, and their R-factors. Any delay in analyzing strains that are responsible for outbreaks can obscure the actual source of an outbreak. This is because many resistant strains tend to lose their resistance to antibiotics after longer storage periods and maintenance of cultures, or due to genetic changes during cell division or transmission from one host to another as was observed in some strains (Al-Bahry, 1999).
The plasmids DNA analysis of 28 different strains showed that the size of the DNA in the various strains ranged from a molecular of 3.1 kb to 32 kb. Most non-human strains had larger plasmid and most human strains have relatively much smaller plasmids. The human strains 38, 41, 43, and 44 each contained a single plasmid ranging from 10.8, 9, 4.7 and 6.2 kb, respectively, yet they were resistant to a large number of antibiotics. This observation of multiple drug resistance with only one plasmid needs further investigation.
Most of the extracted plasmids from salmonellae isolated from chickens were found to be closely related. Most of these plasmids were found to have mw wts of 28-32 kb, and they revealed similar fragments when cleaved with Eco RI. This suggests that these plasmids may have evolved from a common ancestor. Similarities were also observed between plasmid types, 40c, 40d, 45a, 45b, 46c, and 47b. These plasmids were extracted from human strains 40, 45, 46, and 47 respectively. Surprisingly, strain 40, which contained 4 plasmids, and 45, which contained 2 plasmids had a common antibiotic pattern of CKSmx. The strain 46 which contained 6 plasmids, and 47, which contained 2 plasmids also shared similar antibiotic pattern of AmSmxTeTmp.
Acknowledgments
I thank the staff members of the Central Bacteriology Laboratory, Public Health, Ministry of Health for their helping in sample collection. Dr. T. Kalas. University of Wisconsin-Oshkosh for the plasmid standards. Dr. M. A. Rouf, University of Wisconsin-Oshkosh for his suggestions and directions.