The incidences of diarrhoeal episodes in infants and children have mostly been associated with the consumption of contaminated weaning foods. This is especially true in developing countries where factors such as the lack of sanitation systems and electricity have been found to contribute to an increase in the incidence of microbiologically contaminated weaning foods. The process of fermentation has been found to reduce the amount of microbiological contamination in such foods as a result of the production of antimicrobial compounds such as organic acids, peroxides, carbon dioxide and bacteriocins. In this study, microbiological surveys were conducted on sorghum powder samples and their corresponding fermented and cooked fermented porridge samples collected from an informal settlement of the Gauteng Province of South Africa. The process of fermentation was found to result in significant decreases (P>0.05) in Gram-negative counts and spore counts, while aerobic plate counts decreased slightly. Lactic acid bacteria counts, however, increased significantly (P>0.05). The cooking process was found to result in further significant decreases (P>0.05) in all counts. Sorghum powder samples and fermented porridge samples were found to be contaminated with potential foodborne pathogens, including Bacillus cereus, Clostridium perfringens and Escherichia coli, however, none of the pathogens tested for were detected in any of the cooked fermented porridge samples. SDS-PAGE and phenotypic analysis of 180 lactic acid bacteria isolated from sorghum powder samples and their corresponding fermented and cooked fermented porridge samples showed that a majority of the isolates were lactobacilli and leuconostocs, however, some isolates were identified as pediococci and lactococci. These results demonstrated the heterogeneity of the lactic acid bacteria isolates that were associated with fermentation processes in this study. Of the lactic acid bacteria identified, Lactobacillus plantarum and Leuconostoc mesenteroides strains were found to have the highest distribution frequencies, being distributed in 87% and 73% of the households, respectively. Analysis of Lactobacillus plantarum (58) and Leuconostoc mesenteroides (46) strains isolated from sorghum powder samples and corresponding fermented and cooked fermented porridge samples by AFLP fingerprinting showed that they originated from a common source, which was sorghum powder. There was, however, evidence of strains that may have been introduced at household level. Antimicrobial activity of selected lactic acid bacteria was found to be mainly due to a decrease in pH in fermented and cooked fermented porridge samples. None of the lactic acid bacteria tested seemed to produce bacteriocins.

Thesis (MSc)--University of Stellenbosch, 2007. / ENGLISH ABSTRACT: The World Health Organisation (WHO) estimates that at least 75% of infants receive infant formula milk (IFM) either entirely or in conjunction with breast milk during the first four months after birth. The presence of the emerging pathogen Enterobacter sakazakii in IFM has been associated with rare but fatal cases of neonatal infections and deaths. There is thus a need for accurate methods for the rapid detection of E. sakazakii in foods. At present, the methods used to detect and identify this micro-organism are inadequate, controversial and contradictory. The aim of this study was to determine the most suitable method for E. sakazakii detection after evaluation of the currently available methods. A further aim was to optimise a polymerase chain reaction (PCR) method for the detection of only viable E. sakazakii cells utilising the DNA-intercalating dyes ethidium monoazide (EMA) and propidium monoazide (PMA). The Food and Drug Administration (FDA) method for E. sakazakii detection was utilised to select 50 isolates from IFM and 14 from the environment, regardless of colony appearance. These isolates were identified by sequencing a 1.5 kilobase (kb) fragment of the 16S ribosomal DNA (rDNA) and by using the National Centre for Biotechnological Information (NCBI) database to confirm the closet known relatives. Seven of the 50 (14%) IFM isolates and six of the 14 (43%) environmental isolates were identified as E. sakazakii. The methods that were evaluated for accuracy in detecting and identifying these E. sakazakii isolates included yellow pigment production on tryptone soy agar (TSA), chromogenic Druggan-Forsythe-Iversen (DFI) and Enterobacter sakazakii (ES) agars and PCR using six different species-specific primer pairs described in the literature. The suitability of the FDA method was lowered by the low sensitivity, specificity and accuracy (87%, 71% and 74%, respectively) of using yellow pigment production for E. sakazakii identification. DFI and ES agars were shown to be sensitive, specific and accurate (100%, 98% and 98%, respectively) for the detection of E. sakazakii. The specificity of the PCR amplifications was found to vary between 8% and 92%, with Esakf and Esakr being the most accurate of the primer pairs evaluated. The current FDA method for E. sakazakii detection requires revision in the light of the availability of more sensitive, specific and accurate detection methods. Based on the results obtained in this study, a new method is proposed for the detection of E. sakazakii in food and environmental samples. This proposed method replaces the culturing steps on violet red bile glucose agar (VRBGA) and TSA with culturing on chromogenic DFI or ES agar. For identification and confirmation of presumptive E. sakazakii isolates, the oxidase test, yellow pigment production and API biochemical profiling is replaced by DNA sequencing and/or species-specific PCR with the most accurate primer pair (Esakf and Esakr). The amendments to the current FDA method will reduce the time to detect E. sakazakii from approximately 7 days to 4 days and should prove to be more sensitive, specific and accurate for E. sakazakii detection. In this study, a novel PCR-based method was developed which was shown to be capable of discriminating between viable and dead E. sakazakii cells. This was achieved utilising the irreversible binding of bacterial DNA to photo-activated PMA or EMA in order to prevent PCR amplification from the dead cells. At concentrations of 50 and 100 μg.ml-1, PMA completely inhibited PCR amplification from dead cells, while causing no significant inhibition of the PCR amplification from viable cells. EMA was equally effective in preventing PCR amplification from dead cells, however, it also inhibited PCR amplification from viable cells. PMA-PCR in particular, will be useful for assessing the efficacy of processing techniques, as well as for monitoring the resistance, survival strategies and stress responses of E. sakazakii. This will be an important step in the efforts to eliminate E. sakazakii from food and food production environments. / AFRIKAANSE OPSOMMING: Die Wêreld Gesondheidsorganisasie (WGO) beraam dat ten minste 75% van alle babas net baba formule melk (BFM) of BFM in kombinasie met moedersmelk in die eerste vier maande na geboorte kry. Die teenwoordigheid van die voortkomende patogeen Enterobacter sakazakii in BFM is al geassosieer met skaars maar noodlottige gevalle van neonatale infeksies en sterftes. Akkurate metodes word dus benodig vir die vinnige deteksie van E. sakazakii in voedsel. Die metodes wat huidiglik gebruik word vir die deteksie en identifikasie van hierdie mikroörganisme is onvoldoende, kontroversieël en teenstrydig. Die doel van hierdie studie was om die beste metode vir die deteksie van E. sakazakii te bepaal, na 'n evaluasie van die metodes wat huidiglik beskikbaar is. 'n Verdere doel was om 'n polimerase ketting reaksie (PKR) metode vir die deteksie van slegs lewensvatbare E. sakazakii selle te optimiseer deur gebruik te maak van die DNSbindende kleurstowwe, etidium mono-asied (EMA) en propidium mono-asied (PMA). Die Voedsel en Medisyne Administrasie (VMA) se metode vir E. sakazakii deteksie is gebruik om, ongeag van die kolonie kleur, 50 isolate vanuit BFM en 14 isolate vanuit die omgewing te kies. Hierdie isolate is geïdentifiseer deur die DNS volgorde van 'n 1.5 kilo-basis (kb) fragment van die 16S ribosomale DNS (rDNS) te bepaal en die Nationale Sentrum vir Biotegnologiese Informasie (NSBI) databasis te gebruik om die mees verwante spesie te bevestig. Sewe van die 50 (14%) BFM isolate en ses van die 14 (43%) omgewings isolate is geïdentifiseer as E. sakazakii. Die metodes wat geëvalueer is in terme van akkuraatheid vir deteksie en identifikasie van hierdie E. sakazakii isolate het PKR met ses verskillende spesie-spesifieke peiler pare soos beskryf in die literatuur, geel-pigment produksie op triptoon soja agar (TSA) en chromogeniese Druggan-Forsythe-Iversen (DFI) en Enterobacter sakazakii (ES) agars ingesluit. Die geskiktheid van die VMA metode is verlaag deur die lae sensitiwiteit, spesifisiteit en akkuraatheid (87%, 71% en 74% onderskeidelik) van geel pigment produksie vir E. sakazakii identifikasie. Chromogeniese DFI en ES agars was sensitief, spesifiek en akkuraat (100%, 98% en 98% onderskeidelik) vir die identifikasie van E. sakazakii. Die spesifisiteit van die PKR produkte het gewissel tussen 8% en 92%, en Esakf en Esakr is as die akkuraatste geëvalueerde peiler paar geidentifiseer. Die huidige VMA metode vir E. sakazakii deteksie vereis hersiening aangesien meer sensitiewe, spesifieke en akkurate deteksiemetodes voortdurend beskikbaar word. 'n Nuwe metode, gebaseer op die resultate van hierdie studie, word voorgestel vir die deteksie van E. sakazakii in voedsel- en omgewingsmonsters. Die voorgestelde metode vervang die kwekingsstap op violet rooi gal glukose agar (VRGGA) en TSA deur kweking op chromogeniese DFI of ES agars. Verder word die oksidase toets, geel pigment produksie en API biochemiese profiele van vermoeidelike E. sakazakii isolate vervang deur DNS volgorde bepaling en/of spesie-spesifieke PKR met die mees spesifieke peiler paar (Esakf and Esakf) vir die identifikasie en bevestiging van E. sakazakii. Die voorgestelde wysigings van die VMA metode sal die tydsduur van E. sakazakii identifikasie van 7 dae na 4 dae verminder, en behoort ook meer sensitief, spesifiek en akkuraat te wees vir die deteksie van E. sakazakii. 'n Nuwe PKR-gebaseerde metode wat tussen lewensvatbare en dooie E. sakazakii selle kan onderskei is in hierdie studie ontwikkel. Dit is bereik deur die onomkeerbare binding van bakteriële DNS aan lig-geaktiveerde EMA of PMA om die PKR amplifisering van dooie selle te voorkom. Konsentrasies van 50 en 100 μg.ml-1 PMA het PKR amplifikasie heeltemal geïnhibeer, terwyl geen inhibisie van lewensvatbare selle bespeur kon word nie. EMA was ook suksesvol in die voorkoming van die PKR amplifikasie van dooie selle, alhoewel daar ook 'n mate van DNS inhibisie was tydens die amplifikasie van lewensvatbare selle. PMA-PKR kan ook van nut wees vir die assessering van die doeltreffendheid van prosesseringstegnieke, en ook vir die waarneming van die weerstandigheid, oorlewingsstrategieë en stresresponse van E. sakazakii. Dit sal 'n belangrike stap wees in pogings om E. sakazakii van voedsel en voedsel produksieomgewings te elimineer.