The widespread and indiscriminate use of antibiotics has led to the development of antibiotic resistance in pathogenic, as well as commensal, microorganisms. Resistance genes may be horizontally or vertically transferred between bacterial communities in the environment. The recipient bacterial communities may then act as a reservoir of these resistance genes. In this study, we report the incidence of antibiotic resistance in enteric bacteria isolated from the Mhlathuze River and the distribution of genetic elements that may be responsible for the observed antibiotic resistance. The resistance of the enteric bacteria isolated over a period of one year showed that resistance to the older classes of antibiotics was high (94.7% resistance to one antibiotic and 80.8% resistance to two antibiotics). Furthermore, antibiotic resistance data of the environmental isolates showed a strong correlation (r = 0.97) with data obtained from diarrhoea patients. PCR based methods demonstrated that class 1 integrons were present in >50% of the environmental bacterial isolates that were resistant to multiple antibiotics. This class of integrons is capable of transferring genes responsible for resistance to beta-lactam, aminoglycoside, sulfonamide and quaternary ammonium antimicrobial agents. Conjugate plasmids were also isolated, but from a small percentage of isolates. This study showed that the Mhlathuze River (a) is a medium for the spread of bacterial antibiotic resistance genes, (b) acts as a reservoir for these genes and (c) due to socio-economic pressures, may play a role in the development and evolution of these genes along this river system.

One hundred and thirteen enteric bacteria were isolated from the Mhlathuze River during February 2001 to January 2002. The antibiotic resistance patterns (ARPs) of these identified isolates showed that 94.7 % were resistant to at least one class of antibiotic while 75.2% were multi-resistant. All isolates were sensitive to gentamicin. The levels of resistance exhibited by isolates to specific antibiotics are as follows: penicillin, 72.6%; rifampicin, 69.2%; novobiocin, 52.1%; ampicillin, 43.6% and cephalothin, 28.2 %. The antibiotic resistance gene pool especially that for β-lactamase is likely to be widely available in the environment. The ARPs for E. coli and non-E. coli groups were very similar. The enteric bacteria isolated from downstream, which is a mainly urban and industrial area, were more resistant to several antibiotics than those from upstream which is predominantly rural. The results suggest that environmental, industrial and/or human activities impact on the level of antibiotic resistance in the environment. The strong correlation (r=0.97) between the ARPs of the clinical and the environmental isolates may suggest a link between diarrhoeal incidence and the water quality in the region. It is thus imperative that the determination of antibiotic susceptibility/resistance patterns of isolated microbes is a part of the microbial monitoring process of the water.

The Mhlathuze River has become a major reservoir for antibiotic-resistant microbes and a wide pool of antibiotic resistant genes with the environmental isolates exists in this water system. The ability of commensal organisms to carry resistant genes of clinical importance and their ability to transfer such genes to other bacteria are of greater concern than phenotypic measurements. Forty-three Enterobacteriaceae isolates, which were capable of resisting more than 4 different classes of antibiotics, were used for the molecular characterisation of antibiotic resistances. It was found that 58% of these multiple antibiotic-resistant isolates possess the Class 1 integron. Amongst these 25 isolates with positive detection of the Class 1 integron, the beta-lactamase gene (pse) was the most common, being present in 44% of these integrons. The aminoglycoside-resistant gene was detected in 16% of these integrons and 24% of Class 1 integrons contained two genes coding for sulphonamide resistance and for quaternary ammonium compounds resistance. A high degree of genotypic diversity and the lack of correlation between antimicrobial resistance patterns and molecular types of the isolates suggest convergent acquisition of resistance determinants by genetically unrelated strains rather than epidemic spread of resistant isolates in the community. Of the tested environmental isolates, 56% transferred their plasmids as well as their antibiotic resistance profiles to the recipient cells. The possibility of transmission of resistant genes between bacteria (especially pathogenic) which invade human and animal populations within this river poses a health risk to the communities who are dependent on this river for water consumption.

Fecal contamination of source waters and its associated intestinal pathogens continues to pose risks to public health although the extent and effect of microbial contamination of source waters gets very little attention in designing treatment plants in most developing countries. Coliform counts give an indication of the overall bacterial contamination of water and thus its safety for human consumption. However, their presence fails to provide information about the source of fecal contamination which is vital to managing fecal contamination problems in surface waters. This study explored the use of multiple antibiotic resistance (MAR) indexing as means of differentiating E. coli isolates from different sources. A total of 322 E. coli isolates were obtained from municipal wastewater and from fecal samples from domestic and wild animals. Conventional culture methods and standard chemical and biochemical tests were used for isolation and identification of E. coli. Isolates were assayed against 10 antibiotics using the micro-dilution technique. The results obtained generated antibiotic resistance profiles which were used to statistically group the isolates into different subsets. Correct source classification was obtained for 60% of human-derived and 95% non-human-derived E. coli respectively. These results indicate the validity of the usefulness of MAR indexing as a method of bacterial source tracking.

The microbial quality (total and faecal coliform counts) and some physico-chemical parameters of the Mhlathuze River water source were monitored during March 2001 to November 2002 and compared to the previous study conducted during 1998-1999. The results showed that most of the physical and chemical values obtained were within the recommended limits specified in the South African Water Quality Guidelines. High concentrations of metal were detected in water samples from Felixton and the Richards Bay estuary. Water samples from the Mhlathuze Pumping Station and Felixton, which contained higher concentrations of total nitrogen and phosphate, possessed higher faecal coliform contamination than other sites. The total coliform counts of the Mhlathuze River in this study period were noted to be significantly higher than those in the 1998 to 1999 period. As observed in the previous study, Felixton continues to be the site with major faecal contamination. The resuscitation results indicated that the level of faecal contamination in the Mhlathuze catchment was higher than that measured using conventional methods. Therefore the real impact of this "viable but non-culturable" state of microorganisms in this water system requires urgent attention. Larger fluctuations in the trend of total and faecal coliform counts were observed in 2001. This phenomenon coincided with the major construction of the Mhlathuze pumping station. High water surface temperatures and rainfall figures might have also contributed to this observation. Evidence from our results strongly suggests that the use of faecal coliform bacteria as indicators should be expanded and more research is indicated to identify the impact of the "viable but non-culturable" (VBNC) state of pathogens in this environment.