Thesis (MSc Food Sc)--Stellenbosch University, 2013. / ENGLISH ABSTRACT: Food-borne disease outbreaks caused by Escherichia coli have been linked to the use of faecally-polluted irrigation waters. Thus the overall aim of this research was to evaluate irrigation water and agricultural effluents as sources of antibiotic resistant E. coli in the Western Cape. The aim of the first study was to enumerate and characterise E. coli present in irrigation water and in potential contamination sources. Maximum total coliform and E. coli counts for irrigation sites was log 7.862 and log 5.364 MPN.100 mL-1, respectively. Five out of seven irrigation sites had E. coli counts exceeding national and international guidelines for ‘safe’ irrigation water (<1 000 counts.100 mL-1), making it unsafe for the irrigation of fresh produce. In this study, 46.6% of the E. coli strains were characterised in phylogenetic group B1. It has been shown that E. coli in group B1 have the ability to survive and persist in the external environment. Group B1 was also the most common group among isolates from irrigation sites (79.4%), while isolates from environmental sites grouped mainly in group A0 (54.1%). It was concluded that the wide variation of E. coli types present in irrigation water is a concern that should be further investigated. This raises human health implications since the increased exposure to faecal organisms increases the risk of food-borne outbreaks. The E. coli isolates (n = 120) and the marker (n = 37) and reference strains (n = 6), were evaluated for antibiotic resistance to seven medically-important antibiotics from different classes using the Kirby-Bauer disc diffusion method. Thirty-five strains (35/163 = 21.5%) exhibited resistance to one or more antibiotics. Piggery effluent was found to harbour the most antibiotic resistant E. coli isolates (9/35 = 25.7%). Among the resistant E. coli strains, the highest occurrence of antibiotic resistance was to trimethoprim (2.5 μg) (68.6%), tetracycline (30 μg) (57.1%), ampicillin (10 μg) (45.7%) and chloramphenicol (30 μg) (34.3%). Seventy-four percent (26/35) exhibited multiple antibiotic resistances to two or more antibiotics. The antibiotic resistant E. coli strains were evaluated for the presence of pathotypes using Polymerase Chain Reaction analysis to detect Intestinal Pathogenic E. coli (InPEC) and Extra-intestinal Pathogenic E. coli (ExPEC). Five InPEC strains were characterised as four Entero-Pathogenic E. coli (EPEC) strains resistant to three or four antibiotics and one Entero-Aggregative E. coli (EAEC) strain resistant to trimethoprim. The antibiotic resistant EAEC strain also possessed the ExPEC-related gene iutA. Two E. coli isolated from the Mosselbank River were both resistant to chloramphenicol and trimethoprim and also possessed the ExPEC-related gene iutA. It was concluded that the diverse antibiotic resistances of E. coli pathotypes present in irrigation water is a concern that should be further investigated. / AFRIKAANSE OPSOMMING: Voedselverwante siekte uitbrake wat deur Escherichia coli veroorsaak word, is gekoppel aan die gebruik van fekale besoedelde besproeiingswater. Dus was die hoof doel van die navorsing om besproeiingswater en landbou-afvalwater te evalueer as bronne van antibiotika-weerstandbiedende E. coli in die Wes-kaap. Die doel van die eerste studie was om die getalle en eienskappe van E. coli te bepaal wat in besproeiingswater en in ander potensiële besmettingsbronne teenwoordig is. Maksimum totale koliforme en E. coli-tellings vir besproeiingspunte was onderskeidelik log 7.862 en log 5.364 MPN.100 mL-1. Vyf uit sewe besproeiingspunte het E. coli-tellings gehad wat hoër is as die nasionale en internasionale riglyne vir ‘veilige’ besproeiingswater (<1 000 tellings.100 mL-1). Dit maak dit onveilig vir die besproeiing van vars produkte. In hierdie studie was 46.6% van die E. coli-stamme in filogenetiese groep B1 gegroepeer. Dit is reeds bewys dat E. coli in groep B1 oor die vermoë beskik om in die eksterne omgewing te oorleef en voort te bestaan. Groep B1 was ook die mees algemene groep onder die isolate van besproeiingspunte (79.4%), terwyl isolate van omgewingspunte meestal in groep A0 (54.1%) gegroepeer is. Die breë variasie E. coli tipes in die besproeiingswater is bekommerniswaardig en sal gevolglik verder ondersoek moet word. Dit bring gesondsheidsimplikasies mee vir mense aangesien die verhoogde blootstelling aan fekale organismes die risiko van voedselverwante uitbrake verhoog. Die E. coli isolate (n = 120) en die merker (n = 37) en verwysingsstamme (n = 6), is teen sewe medies belangrike antibiotikas uit verskillende klasse getoets vir antibiotika-weerstandbiedendheid. Die Kirby-Bauer skyfie diffusie metode is gebruik. Vyf-en-dertig stamme (35/163 = 21.5%) het weerstand teen een of meer antibiotika getoon. Dit is gevind dat vark-afvalwater die meeste antibiotika-weerstandbiedende E. coli-isolate (9/35 = 25.7%) bevat. Die weerstandbiedende E. coli-stamme het die hoogste antibiotika-weerstandheid getoon teen "trimethoprim" (2.5 μg) (68.6%), tetrasiklien (30 μg) (57.1%), ampisillien (10 μg) (45.7%) en chloramfenikol (30 μg) (34.3%). Vier-en-sewentig persent (26/35) het meervoudige weerstandbiedheid teen twee of meer antibiotikas getoon. Die antibiotika-weerstandbiedende E. Coli stamme is getoets vir die teenwoordigheid van patogene deur van Polimerase Ketting Reaksie analise gebruik te maak om ‘Intestinal Pathogenic’ E. coli (InPEC) en ‘Extra-intestinal Pathogenic’ E. coli (ExPEC) op te spoor. Vyf InPEC-stamme is geklassifiseer as vier ‘Entero-Pathogenic’ E. coli (EPEC)-stamme wat weerstandbiedend teen drie of vier antibiotika getoon het en een ‘Entero-Aggregative’ E. coli (EAEC)-stam wat weerstandbiedendheid getoon het teen "trimethoprim". Die antibiotika-weerstandbiedende EAEC-stam het ook die ExPEC-verwante geen, iutA, besit. Twee E. coli isolate van die Mosselbankrivier het weerstand teen beide chloramfenikol en "trimethoprim" getoon en het ook die ExPEC-verwante geen, iutA, besit. Daar is tot die gevolgtrekking gekom dat die diverse antibiotika-weerstandbiedenheid van E. coli patogene teenwoordig in besproeiingswaters bekommerniswaardig is en verder ondersoek behoort te word.

Thesis (MSc Food Sc)--Stellenbosch University, 2015. / ENGLISH ABSTRACT: Fresh produce irrigation water from Western Cape Rivers carries faecal coliforms (FC) (Escherichia coli) at concentrations which often exceed the suggested limit of 1 000 FC per 100 mL and presents a health risk to consumers. On-farm ultraviolet (UV) irradiation presents several advantages for water disinfection but is an uncommon practice in South Africa. The aim of this study was to investigate the use of UV irradiation for river water disinfection prior to irrigation. Escherichia coli (E. coli) strains were exposed to low-pressure (LP) UV (4 mJ.cm-2) and UV/Hydrogen peroxide (H2O2) (4 mJ.cm-2/20 mg.L-1) treatments in Sterile Saline Solution (SSS). Strain variation in reductions was observed and ranged from 1.58 to 3.68 and 1.34 to 3.60 log for the UV and UV/H2O2 treatments, respectively. The UV/H2O2 treatment (4 mJ.cm-2/20 mg.L-1) was more effective, compared to UV singly, against some of the E. coli strains. Selected strains showed increased sensitivity at higher UV doses (8, 10 and 13 mJ.cm-2) and H2O2 concentrations (100 and 200 mg.L-1 with 4 mJ.cm-2) but a 3 log target reduction was not always reached. For all UV and UV/H2O2 treatments maximum resistance was shown by an environmental strain. Reference strains should, therefore, not be used for the optimisation of UV based disinfection parameters. At 10 mJ.cm-2 an American Type Culture Collection (ATCC) reference strain and an environmental strain (ATCC 25922 and F11.2) were both significantly less inactivated in sterilised river water compared to SSS. Enhanced water quality allowed for improved inactivation of the ATCC strain. Also, the efficiency of LP UV (5, 7 and 10 mJ.cm-2) and medium-pressure (MP) UV (13, 17 and 24 mJ.cm-2) radiation was investigated using water from the Plankenburg River. Water was sampled and treated on three respective days (Trials 1 to 3). Physico-chemical and microbiological water quality was always poor. The FC concentration reached a maximum of 6.41 log cfu.100 mL-1 while UV transmission was always below 38%. For LP and MP UV irradiation increased doses resulted in increased disinfection but a 3 log reduction of FC was only attained when MP UV light was used in Trial 1. Disinfection efficiency was dependent on water quality and on the characteristics of the microbial population in the water. Since FC were never reduced to below 3 log cfu.100 mL-1, the water did not adhere to guidelines for produce irrigation. Photo-repair following irradiation was investigated in river water using MP UV doses of 13 and 24 mJ.cm-2 and 3.5 kLux reactivating light, initially. Ultraviolet transmission was close to 50% and total coliform (TC) reduction exceeded 3 log, even at 13 mJ.cm-2. However, TC were reactivated from below 1 000 cfu.100.mL-1 to 3.93 and 4.41 log cfu.100 mL-1 for the 13 and 24 mJ.cm-2 treatments, respectively. A higher MP dose (40 mJ.cm-2) and a different treatment regime (2 x 20 mJ.cm-2) inhibited photo-repair (compared to 13 and 24 mJ.cm-2) but TC were always recovered to a final concentration surpassing 3 log cfu.100 mL-1, even under lower light intensities (1.0 to 2.0 kLux). In the current study UV irradiation did not produce water of acceptable standards for produce irrigation, mainly as a result of extremely poor water quality. However, on farm-scale, UV efficiency could be enhanced by improving water quality before irradiation. Also, stronger lamps that deliver higher UV doses may result in adequate disinfection, irrespective of water quality. Higher UV doses and the use of combination treatments (such as UV/Chlorine and UV/Peracetic acid) should be further investigated also to determine its disinfection efficiency and possible capability to inhibit post-disinfection repair. / AFRIKAANSE OPSOMMING: Varsproduk besproeiingswater vanuit Wes-Kaapse riviere bevat fekale kolivorme (FK) (Escherichia coli) in konsentrasies wat dikwels die voorgestelde limiet van 1 000 FK per 100 mL oorskry en hou `n gesondheidsrisiko vir verbruikers in. Plaasvlak ultraviolet (UV) bestraling bied verskeie voordele met verwysing na water dekontaminering, maar word selde aangewend in Suid-Afrika. Die doel van hierdie studie was om die gebruik van UV bestraling vir die dekontaminering van rivierwater, voor besproeiing, te ondersoek. Escherichia coli (E. coli) isolate is blootgestel aan lae-druk (LD) UV (4 mJ.cm-2) en UV/Waterstofperoksied (H2O2) (4 mJ.cm-2/20 mg.L-1) behandelings in Steriele Sout Oplossing (SSO). Isolaat variasie in reduksies is waargeneem en het gewissel tussen 1.58 tot 3.68 en 1.34 tot 3.60 log vir die UV en UV/H2O2 behandelings, onderskeidelik. In vergelyking met UV bestraling alleen was die UV/H2O2 behandeling (4 mJ.cm-2/20 mg.L-1) meer effektief teen sommige E. coli isolate. Geselekteerde isolate was meer sensitief tot hoër UV dosisse (8, 10 en 13 mJ.cm-2) en H2O2 konsentrasies (100 en 200 mg.L-1 met 4 mJ.cm-2), maar `n 3 log teikenreduksie was nie altyd haalbaar nie. Vir alle UV en UV/H2O2 behandlinge was die meeste weerstand deur `n omgewingsisolaat gebied. Verwysingsisolate behoort daarom nie aangewend te word vir die optimisering van UV-gebaseerde behandelingsparameters nie. By 10 mJ.cm-2 was beide `n ATCC verwysingsisolaat en `n omgewingsisolaat (ATCC 25922 en F11.2) betekenisvol minder gedeaktiveer in rivierwater as in SSO. Verbeterde waterkwaliteit het verhoogde inaktivering van die ATCC isolaat toegelaat. Die doeltreffendheid van LD UV (5, 7 en 10 mJ.cm-2) en medium-druk (MD) UV (13, 17 en 24 mJ.cm-2) bestraling is ook ondersoek deur watermonsters vanuit die Plankenburg Rivier te gebruik. Watermonsters was getrek en behandel op drie verskillende dae (Proewe 1 tot 3). Fisies-chemiese en mikrobiologiese kwaliteit van die water was deurentyd swak. Die FK konsentrasie het `n maksimum van 6.41 log kve.100 mL-1 bereik terwyl UV transmissie altyd laer as 38% was. Vir beide LD en MD UV bestraling het verhoogde dosisse gelei tot verbeterde dekontaminering, maar `n 3 log reduksie is slegs bereik toe MD UV lig gebruik is in Proef 1. Die effektiwiteit van die behandelings was afhanklik van waterkwaliteit en die eienskappe van die mikrobiese populasie in die water. Aangesien FK nooit tot onder 3 log kve.100 mL-1 verminder is nie het die water nie voldoen aan riglyne vir varsproduk-besproeiing nie. Fotoherstel na bestraling was ondersoek in rivierwater deur aanvanklik gebruik te maak van MD UV dosisse van 13 en 24 mJ.cm-2 en 3.5 kLux heraktiverende lig. Ultraviolettransmissie het byna 50% bereik en reduksie van totale kolivorme (TK) het 3 log oorskry, selfs by 13 mJ.cm-2. Totale kolivorme was egter geheraktiveer van onder 1 000 kve.100.mL-1 tot 3.93 en 4.41 log kve.100 mL-1 vir die 13 en 24 mJ.cm-2 behandelings, onderskeidelik. In vergelyking met 13 en 24 mJ.cm-2 het `n hoër MD dosis (40 mJ.cm-2) en `n veranderde bestralingstegniek (2 x 20 mJ.cm-2) fotoherstel onderdruk, maar TK was in elke geval geheraktiveer tot `n finale konsentrasie hoër as 3 log kve.100 mL-1, selfs onder laer intensiteit lig (1.0 tot 2.0 kLux). In hierdie ondersoek het UV bestraling nie water van aanvaarbare standaarde vir varsproduk besproeiing gelewer nie, hoofsaaklik as gevolg van swak waterkwaliteit. Nietemin, op plaasvlak mag die effektiwiteit van UV bestraling verhoog word deur waterkwaliteit voor bestraling te verbeter. Die gebruik van sterker lampe, om hoër UV dosisse te produseer, mag verder bydra tot voldoende dekontaminasie, ongeag van waterkwaliteit. Hoër UV dosisse en die gebruik van kombinasie behandelinge (soos UV/Chloor en UV/Perasynsuur) moet ook verder ondersoek word om die dekontaminasie effektiwiteit, en vermoë daarvan om heraktivering na dekontaminering te onderdruk, vas te stel.

Thesis (MSc Food Sc)--Stellenbosch University, 2013. / ENGLISH ABSTRACT: Over the past two decades, there has been an increase in the use of water sources for irrigation, as well as an increase in Escherichia coli outbreaks linked to fresh produce. The full extent and type of E. coli contamination present in natural water sources is unknown and the contamination sources have also not been confirmed. The aim of this study was to enumerate and characterise E. coli from both irrigation water and potential contamination source sites. Total coliform and E. coli counts found in contamination source sites were as high as log 7.114 and log 6.912 MPN.100 mL-1, respectively. Total coliform and E. coli counts for irrigation sites were lower, with maximum counts of log 5.788 and log 5.768 MPN.100 mL-1, respectively. It was found that more than one third (5/14 = 35.71%) of the irrigation sites had E. coli counts exceeding the guidelines (<1 000 counts.100 mL-1) for ‘safe’ irrigation water for fresh produce (<1 000 counts.100 mL-1) as set by the Department of Water Affairs (DWA) and World Health Organisation (WHO), making the water unsuitable for the irrigation of fresh produce. Phylogenetic subgroups (A0, A1, B1, B22, B23, D1 and D2) and the MALDI Biotyper system (PCA dendrogram) were used to create a fingerprint of each E. coli isolated from the environment. These were then used to link E. coli strains from irrigation water to their most probable contamination origin. Escherichia coli population structure was found in this study, to be better suited for linking E. coli strains from irrigation water to their most likely source, than just applying the phylogenetic grouping. The MALDI Biotyper data in combination with the phylogenetic subgroup assignment was then used to group similar strains and link E. coli from irrigation water to their contamination sources by comparing population structures. Strains isolated from surface and groundwater showed similar distribution patterns, but groundwater strains showed a population structure more indicative of porcine and bovine origin, while surface water showed population characteristics which could not be used to make conclusive links between the irrigation water and suspected contamination sources. When investigating the population structures of individual sample sites, it was found that phylogenetic subgroups A0, A1 and B1 frequently made up the bulk of the E. coli population. It was also found that linking individual irrigation sites to contamination sources was successful, as irrigation site Berg-2 was found to have a similar population structure to contamination source site Plank-1 which represents human pollution from an informal settlement. This led to the conclusion that Berg-2 was being contaminated by human pollution, most probably from an informal settlement. Upon further investigation it was found that Berg-2 is downstream of an informal settlement, proving that E. coli population structure is a successful means of microbial source tracking (MST). Virulence factors of the 153 E. coli isolated during the study were identified and the potential risk associated with using the investigated irrigation water for irrigation of fresh produce, was determined. Two enteropathogenic E. coli (EPEC) strains were isolated from the irrigation water, one from the Plankenburg River water, and the other from a borehole in the Drakenstein area. The latter indicates that borehole water is not as safe as was once thought, and that there are bacterial contaminants finding their way into groundwater. The occurrence of an EPEC strain in river water shows that neither ground nor surface water is guaranteed to be safe, and that treatment of water being used for the irrigation of fresh produce should be implemented. / AFRIKAANSE OPSOMMING: Oor die afgelope twee dekades was daar nie net 'n toename in die gebruik van alternatiewe waterbronne vir besproeiing nie, maar ook 'n toename in uitbrake van Escherichia coli uitbrake wat aan vars produkte gekoppel kan word. Die tipe E. coli-besmetting wat in natuurlike waterbronne teenwoordig is, is onbekend en die besmettingsbron is ook nog nie bevestig nie. Daarom was die doel van hierdie studie om die voorkomssyfer van E. coli van beide besproeiingswater en potensiële kontaminasiebronne te bepaal, asook om die E. coli te karakteriseer. Totale kolivorme en E. coli-tellings wat in kontaminasiebronne gevind is, het ‘n maksimum van log 7,114 en log 6,912 MPN.100 mL-1 onderskeidelik bereik, terwyl die totale kolivorme en E. coli-tellings vir besproeiingswater laer was, met 'n maksimum van log 5,788 en 5,768 MPN.100 mL-1, onderskeidelik. Dit is bevind dat meer as 1/3 (5/14 = 35,71%) van die besproeiingswaterbronne meer E. coli bevat as wat toegelaat word in die riglyne vir "veilige" besproeiingswater vir vars produkte (<1 000 fekale koliforme.100 mL-1) wat deur die Departement Waterwese (DWA) en die Organisasie vir Wêreldgesondheid (WHO) aanbeveel word. Filogenetiese subgroepe (A0, A1, B1, B22, B23, D1 en D2) en die ‘MALDI Biotyper’-stelsel (PKA dendrogram) is gebruik om unieke profiele vir elke geïsoleerde E. coli te skep. Dié profiele is daarna gebruik om E. coli-stamme van besproeiingswater te koppel aan die mees waarskynlike oorsprong van kontaminasie. Daar is in hierdie studie bevind dat die E. coli-populasiestruktuur beter geskik was vir die koppeling van E. coli-stamme van besproeiingswater na hul mees waarskynlikste bron, as net die toepassing van die filogenetiese groepering. Dit was toe gevind dat E. coli wat uit oppervlak- en grondwater geïsoleer is, soortgelyke verspreidingspatrone het, maar grondwaterstamme se bevolkingstruktuur is meer aanduidend van fekale besmetting deur varke en beeste, terwyl oppervlakwater se bevolkingseienskappe nie duidelik genoeg was om ‘n gevolgtrekking oor moontlike bronne van besmetting te maak nie. Toe die populasiestruktuur van die individuele bemonsteringspunte ondersoek is, is daar bevind dat die filogenetiese subgroepe A0, A1 en B1 dikwels die meeste tot die E. coli bevolking bydra. Daar is ook bevind dat die koppeling van isolate in individuele besproeiingswaterbronne met hul mees waarskynlike bronne van kontaminasie suksesvol was. Besproeiingswater van Berg-2 het 'n soortgelyke populasiestruktuur as Plank-1 wat beskou is as ‘n kontaminasiebron. Dit het gelei tot die gevolgtrekking dat Berg-2 heel waarskynlik deur menslike besoedeling beïnvloed word, soos Plank-1, en dat daar moontlik ook ‘n informele nedersetting by Berg-2 betrokke is. Na verdere ondersoek is gevind dat Berg-2 inderdaad ook stroomaf van 'n ander informele nedersetting geleë is, wat bewys dat die E. coli-populasiestruktuur 'n suksesvolle manier is om E. coli kontaminasie te verbind met besproeiingswater. Patogeniese faktore, wat in E. coli voorkom en maagkieme veroorsaak, is vooraf getoets in elkeen van die 153 E. coli-isolate wat tydens die studie geïdentifiseer is. Twee ‘enteropathogenic’ E. coli (EPEC)-stamme is uit die besproeiingswater geïsoleer: een uit die Plankenburgrivier (Plank-3), en die ander uit 'n boorgat in die Drakenstein-gebied (Boorgat A1). Hierdie inligting dui aan dat boorgatwater nie so veilig is as wat voorheen vermoed is nie, en dat bakteriese kontaminasie wel vookom wat nie alleen die grondwater besmet nie, maar ook daarin oorleef. Die voorkoms van die EPEC-stamme in hierdie studie is ‘n aanduiding dat beide grond- en opppervlakwater ewe gevaarlik kan wees, en dat daar dus geen waarborg vir die veiligheid van die water is nie. Die behandeling van grond- en oppervlakwater, wat vir die besproeiing van vars produkte gebruik word, moet daarom ernstig oorweeg word om moontlike uitbrake van E. coli op vars produkte te verhoed.