<i>E. coli</i> O157 is shed into the environment from the faeces of infected animals. Once in the soil, the availability of hydrological pathways enables cells to disperse through the soil and enter the groundwater.  This thesis investigates the transmission and survival of <i>E. coli</i> O157 in the farming environment with particular emphasis on the role of hydrological pathways.  Findings from fieldwork included the observation that <i>E. coli</i> O157 population densities in faeces varied due to animal shedding rates and increasing temperature.  Rainfall events were found to increase transport of <i>E. coli</i> O157 to a local stream.  Strain typing using MLVA demonstrated that one strain had been transported from faeces into overland flow. Microcosm based experiments found that soil water content, temperature and soil type affected the recoverability of <i>E. coli</i> O157 from soil.  The transport of <i>E. coli</i> O157 was influenced by the presence of preferential flow pathways under heavy rainfall conditions.  The days after rainfall event caused changes in both the population density and the cellular activity of <i>E. coli </i>O157. The population dynamics and activity of <i>E. coli </i>O157 was monitored in different environmental matrices at different temperatures.  These variables were found to influence the population and cellular activity of <i>E. coli</i> O157.  During this experiment, cells appeared to enter a viable but non-culturable state. The study showed that the dispersal of <i>E. coli</i> O157 is mediated by soil-related hydrological pathways.  By using the thesis findings to improve current risk assessment strategies, the number of <i>E. coli</i> O157 cases from environmental sources may be reduced.

Potassium is the major intracellular cation of the bacterium Escherichia coli and is accumulated at high levels in order to maintain turgor pressure. Regulation of potassium is achieved by a number of uptake and efflux systems. The efflux of potassium is regulated by five different systems, two of which are KefB and KefC. The KefB and KefC systems play a major role in projection of cells against the toxicity of electrophiles (e.g. N-ethylmaleimide and methylglyoxal) and these proteins exhibit strong similarity at most levels. Both systems consist of a channel protein (KefB and KefC) and an ancillary protein (KefG and KefF) that is required for full activity. The two systems are maintained in a closed state by glutathione (GSH). Both systems are activated by GSH adducts, which are formed by the reaction between GSH and an electrophile. They differ in their response to methylglyoxal, with only KefB being strongly activated by this electrophile. The KefC protein comprises of two major domains (each of which have sub-domains)- an N-terminal membrane domain and a C-terminal “soluble” domain, conducted by a Q-linker.  The linker sequence is a predominantly hydrophilic with a large number of acidic residues. This thesis presents mutational studies on the linker sequence that show the linker length and residue position are important for full activity of the KefC system. Within the N-terminal domain of KefC a conserved sequence (HALESDIEP) has been identified that is important for channel regulation and gating. Further experiments show that this region interacts with specific residues within the two sub-domains of the carboxy-terminal domain. The C- terminal domain consists of a KTN and a SAM domain, residues in both these sub-domains are confirmed to play a role in channel gating and regulation via interactions with GSH and the HALESDIEP sequence. Cross-linking studies show the linker region to be highly flexible also the KefC channel protein is shown to form oligomers.

Methylmalonyl-CoA mutase (MCM, E.C. 5.4.99.2), a coenzyme B12-dependent enzyme, catalyses the inter conversion of succinyl-CoA and methylmalonyl- CoA. The gene (sbm) encoding this enzyme is found in Escherichia coli (E. coli) at 62.3min on the E. coli chromosome. However, the metabolic role of this enzyme in the organism is not known. This project involves an investigation into this metabolic obscurity. The sbm gene is part of a four gene operon which also includes argK (or ygfD) that codes for a protein kinase catalysing the phosphorylation of two periplasmic binding proteins involved in cationic amino acid transport, ygfG that codes for methylmalonyl-CoA decarboxylase and ygfH that codes for propionyl-CoA: succinyl-CoA transferase. From existing literature we suspect that this operon, including the sbm gene, could be involved in the utilisation of unusual carbon sources such as succinate and propionate. An insertion mutant of the sbm gene created by transposon mediated mutagenesis was used for investigating the role of this gene. The wild type E. coli K12 strain, E. coli TR6524 and the mutant E. coli K12 (sbm::MudJ) were used in this study. Growth of the two strains (E. coli TR6524 and FA1P1) in minimal media with three different concentrations (0.05, 0.5, 5.0μg/mL) of vitamin B12 and in the presence succinate, propionate or glucose as the sole source of carbon, was studied. Growth was typical in media with glucose with no major differences in the growth pattern of the wild type and mutant strain. However, the two strains exhibited a differential growth pattern in media containing succinate, with the wild type growing faster than the mutant, indicating the role of the sbm gene in the utilisation of this carbon source. Growth in media containing propionate as the sole carbon source indicated only marginal differences in the growth pattern of the wild type and mutant strain. This result possibly suggests that the other pathways for propionate utilisation in E. coli compensate for the lack of a functional Sbm protein in the mutant strain. Promoter analysis indicated the presence of a promoter induced by σS, a transcription factor involved in the expression of proteins under stress or stationary phase growth conditions. Reverse transcription polymerase chain reaction (RT-PCR) studies of the genes of the sbm operon (sbm-argK-ygfGygfH) under the same growth conditions were carried out. Densitometric analysis of the PCR products suggested that the transcription level of sbm was higher in E. coli grown in succinate as compared to when grown in glucose and not as much when grown in propionate indicating a transcriptional level control of the sbm gene expression during the utilisation of succinate. RT-PCR studies also indicated a higher level of transcription of the gene in the stationary phase of the culture during the utilisation of succinate. Real time reverse transcription PCR (QPCR) analysis was used for the absolute quantification of the transcription of the genes of the sbm operon. An increase in the mRNA levels corresponding to the sbm, argK and ygfG genes was observed as E. coli TR6524 growth reached stationary phase, in the presence of succinate or propionate as the sole source of carbon as compared to glucose, In contrast, the highest mRNA levels corresponding to the ygfH gene were observed in the early log-phase of growth. This indicated a differential transcriptional level control of the genes within the operon. This study further established the possible role of this operon in the utilisation of succinate and propionate. The MCM enzyme activity measurement in the whole cell extracts of the wild type E. coli K12, grown under the above mentioned conditions, led to the first ever measurement of MCM activity in wild type E. coli. These measurements also revealed a four fold increase of the MCM specific activity in the case of growth in succinate (4.76x10-3U/mg) and a two fold increase for growth in propionate (2.79x10-3U/mg) compared to that observed with growth in glucose (1.37x10-3U/mg), indicating a significant level of involvement of the enzyme in succinate utilisation, and to a lesser extent in propionate utilisation. The proteomic analysis to understand the gene expression pattern of E. coli TR6524 was carried out using cells harvested at the stationary phase. The results showed that growth conditions induced the expression of transport related (HisJ, DppA) and energy generating proteins (PckA, AceF) required by E. coli to cope with the stressful growth conditions. However, Sbm was not identified among the limited protein spots that were analysed. Finally, E. coli K12 sbm gene was successfully cloned into B. cereus SPV leading to the development of a metabolically engineered polyhydroxyalkanoate producing strain of B. cereus. The intention was to provide the bacteria with a natural intracellular source of propionyl-CoA, leading to the production of the P(3HB-co-3HV) copolymer from structurally non related carbon sources like glucose. Hence, this work has initiated investigation into the metabolic role of the sbm gene product in E. coli. In addition, it has also led to the use of this gene product in metabolic engineering applications.

Enteropathogenic Escherichia coli (EPEC) strains are diarrheagenic pathogens that colonize the gut mucosa via attaching-and-effacing (A/E) lesion formation. EPEC utilize a type III secretion system (T3SS) to translocate effector proteins, including Tir and EspH that subvert host cell signalling to sustain colonization and multiplication. In vitro, EPEC induces the formation of an actin-rich ‘pseudopod’-like structure, termed pedestals. My research focussed on EspH, which is implicated in the elongation of actin pedestals, and inactivation of Rho GTPase signalling. In the first study, I showed that EspH promotes Tir-dependent actin assembly at the bacterial attachment sites independently of the Tir tyrosine residues Y454 and Y474, which are implicated in actin pedestal formation. Moreover, EspH promotes the recruitment of N-WASP and the Arp2/3 complex to the bacterial attachment sites. EspH-mediated N-WASP recruitment is dependent on the N-WASP binding protein WIP. Expression of WIP and EspH during EPEC infection induces extensive actin cytoskeletal rearrangements at the bacterial attachment site, forming a protrusive actin patch that is associated with EPEC microcolonies, termed “macro-pedestals”. This study implicates the critical role of WIP in EPEC-mediated actin polymerization, and represents the first instance whereby N-WASP is efficiently activated at the EPEC attachment sites independently of the canonical Tir:Nck and Tir:IRTKS/IRSp53 pathways. In my second study, I investigated the interplay of EspH, a eukaryotic RhoGEF inhibitor, and T3SS bacterial RhoGEFs during EPEC infection. EspH does not inhibit the E. coli RhoGEFs Map, EspT, and EspM, as well as the Salmonella RhoGEF SopE. Expression of EspH induces cell rounding, detachment, caspase-3 activation and death. Importantly, EPEC EspT and EspM2, and Salmonella SopE are able to block EspH-induced cell detachment and caspase-3 activation. In addition, the bacterial RhoGEFs blocked staurosporine-induced caspase-3 activation, and can functionally compensate for the anti-apoptotic T3SS effector NleH. Therefore, this study expands the role of bacterial RhoGEFs to include cell adhesion and survival, and demonstrates that A/E pathogens silence eukaryotic DH-PH RhoGEFs while translocating their own RhoGEFs to hijack Rho GTPase signalling for the exclusive benefit of the pathogen. Highlighted in the two studies is also the concept of effector interplay, where T3SS effectors function in concert with other effectors during the subversion of host cell signaling pathways. Understanding the functional interactions between effectors enables us to gain further insight into the sophisticated nature of their biological relevance beyond that of their biochemical functions.

Despite the HACCP systems which have been introduced to the pork industry, cross-contamination which occurs within pork slaughterlines remains an important concern for food safety of the final carcass. The aim of this work was to understand the dissemination and cross-contamination of enteric bacteria during slaughter processing by investigating Escherichia coli populations. E. coli is widely used as an indicator of faecal or enteric pathogen contamination, and a strong correlation between the presence of Salmonella and E. coli levels was seen in this study. With microbiological counts and molecular typing of E. coli, changes of contamination levels as well as of the bacterial communities was observed during processing. The results demonstrated that temperature variation at different carcass sites during the singeing process allowed strains to survive on the cooler sites of the carcass and be present in the subsequent processing stages. The polishing process was recognised as an important site of cross-contamination not only because of an increase in contamination levels but also because a high variety of sources contributed contamination at this site, including strains surviving through singeing, strains that persist in the polisher overnight and strains from faecal leakage during the polishing process. A high percentage of virulence factor-carrying E. coli were present on the slaughtered carcasses and recombination between virulence genes from different pathovars was observed. These findings suggest carcasses slaughtered from a healthy pig herd may still be a potential source for E. coli pathovars in the food chain.

- / -

The bacterial mechanosensitive channels MscS and MscL have been shown to protect cells from hypo-osmotic shock-induced lysis. Bacterial strains deficient for MscS and MscL are severely compromised and fail to survive a hypo-osmotic shock. Both channels exhibit redundant function such that re-introduction of either of these proteins is sufficient to restore cell survival. Several proteins paralogous to MscS have been identified in E. coli, but their function remains unknown. Mechanosensitive channel homologues are also being discovered in a variety of organisms including Archaea, plants and fungi and their function is starting to emerge.

La contamination des zones littorales par des bactéries entériques potentiellement pathogènes pour l’Homme constitue un problème majeur pour la pérennité de certains usages tel que la conchyliculture. Ces bactéries provenant de rejets agricoles ou urbains peuvent atteindre les zones conchylicoles et être impliquées dans des toxi-infections alimentaires collectives (TIAC). Actuellement, très peu de données sont disponibles sur la présence et la diversité des bactéries entériques telles que les Escherichia coli (E. coli) pathogènes de type E. coli producteurs de Shiga-toxines (STEC) et E.coli entéropathogènes (EPEC) dans les coquillages en France.La présence de ces E. coli pathogènes a été recherchée pendant deux ans, dans trois zones conchylicoles françaises et leurs bassins versants. Un total de 28 souches STEC et 89 souches EPEC différentes ont été isolées dans des coquillages, des eaux aux exutoires, le sédiment et l’eau de mer, représentant 1% de la totalité des souches E. coli isolées (n=12016). Ces souches isolées présentaient néanmoins une grande diversité. Elles étaient réparties au sein de 73 profils de virulence différents dont une souche STEC de sérotype O26 :H11 présentant 47 gènes de virulence isolée dans un lot de moule. Soixante-quinze pourcents des souches EPEC présentaient des gènes de virulence associés à des îlots de pathogénicités caractéristiques de souches pathogènes responsables d’infection grave chez l’Homme, révélant le potentiel pouvoir pathogène de certaines souches. Enfin, l’étude de la cinétique de contamination, décontamination d’huîtres au contact de souches STEC, n’a pas montré de différence comparée à un E. coli non STEC.Les travaux réalisés au cours de cette thèse sont à notre connaissance les premiers de ce genre. Ils ont permis de mettre en évidence la faible présence de STEC et de EPEC au niveau des zones conchylicoles françaises étudiées ainsi que le potentiel pouvoir pathogène de certaines souches. La faible prévalence des souches sur ces sites de catégorie B ou C (purification des coquillages avant commercialisation) reste néanmoins plutôt en faveur d’un risque faible de contamination dans les zones sélectionnées. Les résultats acquis au cours de cette thèse sont des éléments importants pour mieux appréhender le risque sanitaire potentiel lié aux STEC et aux EPEC en zone littorale. / The contamination of coastal areas by potentially pathogenic enteric bacteria is of concern for the sustainability of some uses, such as shellfish farming, recreational shellfish harvesting and bathing. The contamination of these environments may occur through the land-spreading of livestock wastes, animal feces deposited on pastures, wastewaters from slaughterhouses. The presence of these bacteria in coastal environment may present a potential risk to human health. In fact, shellfish-borne outbreaks may occur by the consumption of shellfish from contaminated areas. To date, few studies focusing on the presence and the diversity of enteric bacteria, such as pathogenic Escherichia coli (E. coli) more precisely, Shiga-toxin-producing E. coli (STEC) and enteropathogenic E. coli (EPEC) in coastal environments and shellfish have been reported.For this purpose, during a 2-year study, shellfish batches, freshwater, seawater, and surface sediment samples from three French selected shellfish-harvesting areas and their upstream catchments, were analyzed to evaluate the presence of STEC and EPEC strains. Twenty-eight STEC and 89 EPEC strains were isolated representing 1% of the total E. coli(n=12 016). The isolated STEC and EPEC strains belonged to a high diversity. One STEC strain isolated from a mussel batch, belonging to the serotype O26:H11 displayed 47 additional virulence genes. Seventy-five percent of EPEC strains displayed several virulence genes associated with pathogenicity islands specific to pathogenic E. coli involved in human infections.No difference in the kinetics of the contamination and depuration of oysters by STEC and non-STEC E. coli was found.To our knowledge, this study is the first to focus on the diversity of STEC and EPEC strains isolated from coastal environments. This study highlights the weak presence of STECs and EPECs in the French shellfish-harvesting areas studied and a potential pathogenicity of some strains. The low prevalence of STEC and EPEC strains in shellfish fromB- and C-categories (depuration of shellfish before commercialization), is rather in favor of a limited risk of contamination of shellfish in the studied areas. The results obtained during this study are important to better understand the health risk associated with STEC and EPEC in coastal areas.

Η διατήρηση της μικροβιολογικής ποιότητας του υδάτινου περιβάλλοντος είναι υψίστης σημασίας δεδομένων των κινδύνων που ενέχονται για τη δημόσια υγεία. Η αξιολόγηση της μικροβιολογικής ποιότητας των υδάτων πραγματοποιείται με την ανίχνευση της κοπρανώδους μόλυνσης και με τον έλεγχο της παρουσίας και συγκέντρωσης συγκεκριμένων μικροοργανισμών – δεικτών, όπως είναι η Escherichia coli. Ωστόσο, η απλή ανίχνευση κοπρανώδους μόλυνσης δεν επαρκεί για την υπόδειξη τρόπων εξυγίανσης και αντιμετώπισης του εκάστοτε προβλήματος. Οι δύο κύριες ομάδες στις οποίες διακρίνεται η κοπρανώδης μόλυνση είναι η ανθρώπινη και η ζωική, οι οποίες υποδηλώνουν πιθανή παρουσία διαφορετικών κάθε φορά παθογόνων μικροοργανισμών για τον άνθρωπο. Έτσι, προκειμένου να οριοθετηθεί ο κίνδυνος για τη δημόσια υγεία και να καθοριστούν μέτρα αντιμετώπισης της μόλυνσης ενδείκνυται ο προσδιορισμός της ανθρώπινης ή ζωικής προέλευσης της κοπρανώδους μόλυνσης. Στην παρούσα μελέτη αναπτύχθηκαν, εφαρμόστηκαν και αξιολογήθηκαν οι μέθοδοι: α)Έλεγχος πολλαπλής ανθεκτικότητας σε αντιβιοτικά (Multiple Antibiotic Resistance – MAR – φαινοτυπική μέθοδος) και β) PCR με τυχαία ενισχυμένα τμήματα πολυμορφικού DNA - Random Amplified Polymorphic DNA-PCR (RAPD-PCR – γονοτυπική μέθοδος), ως τεχνικές προσδιορισμού και διάκρισης προέλευσης μικροοργανισμών. Κατά το πρώτο στάδιο καθορίστηκαν οι παράμετροι των μεθόδων για το διαχωρισμό στελεχών E. coli γνωστής προέλευσης (60 στελέχη απομονωμένα από ζωικά κόπρανα και 68 στελέχη από ανθρώπινα). Για το διαχωρισμό και κατηγοριοποίηση των στελεχών εφαρμόστηκαν η Ιεραρχική Ανάλυση Κατά Συστάδες και η Διαχωριστική Ανάλυση. Με τη MAR ανάλυση τα στελέχη E. coli εμφάνισαν διαφορετικούς συνδυασμούς ανθεκτικότητας και διαχωρίστηκαν βάσει της προέλευσής τους με μέσο ποσοστό σωστής ταξινόμησης (ARCC) 99,2%. Με την RAPD-PCR χρησιμοποιήθηκαν δύο εκκινητές ξεχωριστά (1254 & 1290) και τα 128 στελέχη E. coli γνωστής προέλευσης διαχωρίστηκαν σε ανθρώπινης και ζωικής πηγής με ARCC 98,4% και με τους δύο εκκινητές. Η διακριτική ικανότητα της RAPD-PCR με τους δύο εκκινητές ήταν D1254=0,97 & D1290=0,90. Επιπλέον, η αξιολόγηση της επαναληψιμότητας της RAPD-PCR και με τους δύο εκκινητές έδωσε ικανοποιητικά αποτελέσματα με την εμφάνιση ίδιων ηλεκτροφορητικών εικόνων για τα ίδια βακτηριακά στελέχη. Στη συνέχεια οι επιλεγμένες τεχνικές εφαρμόστηκαν για την ταξινόμηση και κατηγοριοποίηση στελεχών E. coli άγνωστης προέλευσης εκτιμώντας την ανθρώπινη ή ζωική πηγή τους βάσει του μοντέλου διαχωρισμού των E. coli γνωστής προέλευσης. Οι E. coli άγνωστης προέλευσης (234 στελέχη) απομονώθηκαν από δείγματα πόσιμου νερού δικτύου από 11 περιοχές και δείγματα μη επεξεργασμένων λυμάτων από τις εισόδους τεσσάρων σταθμών βιολογικού καθαρισμού (ΚΕΡΕΦΥΤ – Νομός Αττικής, ΨΥΤΤΑΛΕΙΑ – Νομός Αττικής, ΡΙΟ – Νομός Αχαΐας και ΠΑΤΡΑ - Νομός Αχαΐας). Τα 234 στελέχη με τη MAR ανάλυση ταξινομήθηκαν ως ανθρώπινα και ζωικά σε ποσοστά 46,6% και 53,4% αντίστοιχα. Τα αποτελέσματα ταξινόμησης ήταν διαφορετικά με τη μέθοδο RAPD-PCR. Με τον εκκινητή 1254 τα άγνωστα στελέχη προσδιορίστηκαν ως ανθρώπινα κατά το 64,9% και ως ζωικά κατά το 35,1%. Αντίστοιχα, με τον εκκινητή 1290 τα ποσοστά ήταν 60,3% ανθρώπινα και 39,7% ζωικά. Τα στελέχη του πόσιμου νερού που προέρχονταν από τους σταθμούς δειγματοληψίας που ήταν αστικά κέντρα χαρακτηρίστηκαν εξ ολοκλήρου ως ανθρώπινης προέλευσης. Αντίθετα, στις περιοχές δειγματοληψίας με ανεπτυγμένη κτηνοτροφία βρέθηκαν και στελέχη ζωικής προέλευσης, γεγονός που υποδηλώνει την είσοδο στο δίκτυο κοπρανώδους υλικού προερχόμενου από ζώα των συγκεκριμένων περιοχών, τα οποία ενδεχομένως να έχουν άμεση πρόσβαση στις πηγές και γεωτρήσεις. Όσον αφορά στο χαρακτηρισμό των E. coli που καταλήγουν στους αναφερόμενους βιολογικούς καθαρισμούς, η πλειοψηφία ανίχνευσης ανθρωπίνων στελεχών δηλώνει την πιθανή παρουσία στα ακατέργαστα λύματα πολλών ανθρωπίνων εντερικών παθογόνων σημαντικών για τη δημόσια υγεία. Δεδομένου ότι τα τελευταία χρόνια οι ερευνητές έχουν αποδυθεί σε μια προσπάθεια επαναχρησιμοποίησης επεξεργασμένων λυμάτων επισημαίνεται η ανάγκη επεξεργασίας τους σε διάφορα στάδια για τη διασφάλιση της δημόσιας υγείας. Παρατηρήθηκε συμφωνία αποτελεσμάτων με τη χρήση των δύο εκκινητών καθώς η διαφορά στα ποσοστά δεν ήταν στατιστικά σημαντική (P>0,05). Συγκρίνοντας τα αποτελέσματα που ελήφθησαν με τις δύο μεθόδους, τη φαινοτυπική (MAR ανάλυση) και τη γονοτυπική (RAPD-PCR), υπήρξε στατιστικά σημαντική διαφορά (P<0,05), με συνέπεια να τίθεται θέμα επιλογής της πιο ενδεδειγμένης μεθόδου τυποποίησης και διάκρισης περιβαλλοντικών μικροοργανισμών. H παρούσα μελέτη αναδεικνύει την RAPD-PCR ως μια γονοτυπική μέθοδο με ικανοποιητική διακριτική ικανότητα, ευαισθησία, επαναληψιμότητα υπό αυστηρά καθορισμένες συνθήκες και χαμηλού κόστους. Η ευκολία εφαρμογής για την τυποποίηση μεγάλου αριθμού βακτηριακών στελεχών, χωρίς την απαίτηση γνώσης της νουκλεοτιδικής αλληλουχίας του γενετικού υλικού την καθιστούν ιδιαίτερα προσιτή σε εργαστήρια μοριακής μικροβιολογίας, ως τεχνική διάκρισης προέλευσης της κοπρανώδους μόλυνσης στο υδάτινο περιβάλλον. / Maintenance of the microbiological quality and safety of water systems is imperative, as their faecal contamination may exact high risks to human health as well as result in significant economic losses. The microbiological quality of water systems is evaluated by detecting their faecal pollution and especially specific faecal indicators such as Escherichia coli. Simple detection of faecal pollution is not sufficient in order to apply appropriate management plans to remedy the problem and to prevent any further contamination. Human faecal material is generally perceived as constituting a grater human health risk than animal faecal material, considering that it is more likely to contain human-specific enteric pathogens. Thus, it would be desirable to determine the source of the faecal material, especially for the assessment of risk for public health and for the development of monitoring plans. In the present study the development and assessment of Multiple Antibiotic Resistance Analysis (MAR – phenotypic method) and Randomly Amplified Polymorphic DNA-PCR Analysis (RAPD-PCR – genotypic method) were established as microbial source tracking methods. Firstly, parameters of the two selected methods were determined for the discrimination of E. coli isolates of known source (60 isolates from animal faecal material & 68 isolates from human faecal material). Hierarchical Cluster Analysis and Discriminant Analysis were applied for the classification of the isolates. With MAR analysis E. coli isolates developed different resistance profiles and were discriminated according to their source with an average rate of correct classification (ARCC) of 85.2%. With RAPD-PCR analysis two different 10-nt primers of arbitrary sequence were used (1254 & 1290) and the 128 E. coli isolates of known origin were classified as human and animal with the following ARCC: ARCC1254= 87.5% & ARCC1290= 81.3%. The discriminatory power of RAPD-PCR with the two selected primers was D1254=0.97 & D1290=0.90. Furthermore, the assessment of reproducibility of RAPD-PCR analysis provided satisfactory results with both primers, as RAPD profiles were identical for the same bacterial isolates. The assessment of specificity of the method resulted in the discrimination among RAPD profiles of E. coli isolates and other reference bacteria. The selected methods were applied for the classification and the source tracking of E. coli isolates, derived from tap water and raw sewage samples. In total 234 E. coli strains were isolated from tap water from 11 areas and raw sewage samples from four treatment plants (KEREFYT – prefecture of Attiki, PSITALIA - prefecture of Attiki, RIO - prefecture of Achaia and PATRA - prefecture of Achaia). With MAR analysis the 234 isolates were classified as human and animal in percentages of 46.6% & 53.4%, respectively. Classification results were different with RAPD-PCR analysis. With primer 1254 the classification was: 64.9% of human origin and 35.1% of animal origin and with primer 1290 the classification was: 60.3% of human origin and 39.7% of animal origin. Isolates derived from tap water of urban areas were classified in total as of human origin. On the contrary, in areas with many farm breeders many isolates were classified as of animal origin, indicating presence of faecal material in the water systems derived animal activities. As far as E. coli isolates from raw sewage samples are concerned, the majority of them were classified as of human source, indicating the possible presence of other human enteric pathogens as well. Taking into account the fact that there has been an effort in order to reuse treated sewage, it seems necessary a multi-stage process to renovate wastewater before it re-enters a body of water. There was an agreement of results of classification obtained form the use of the two different primers as the percentages did vary statistically (P>0.05). Comparing results obtained from the two selected methods, the difference was statistically significant (P<0.05), raising a question of the appropriate method for the typing and discrimination of environmental microorganisms. The present study demonstrates RAPD analysis as a simple, cost effective genotypic method with satisfactory discriminatory power, sensitivity and reproducibility. It can be applied for the analysis of a large number of bacterial isolates without the prior knowledge of nucleotide sequence of DNA to be necessary. Finally, it may fulfil environmental for the determination of origin of faecal pollution protecting water resources and public health.