Title Page

ABSTRACT

Contents

Chapter 1. General Introduction 11

1. Characterizing fecal sources using Escherichia coli 12

Strain level phylogenetic diversity 12

Antibiotic resistance 15

Pathogenicity 16

2. Microbial source tracking 17

Known MST methods 17

New approach (Next generation sequencing technology) 20

3. Scope of this study 21

Chapter 2. Absence of Escherichia coli Phylogenetic Group B2 Strains in Humans and Domesticated Animals from Jeonnam Province, Korea 23

Introduction 24

Materials and Methods 26

Isolation of E. coli from humans and domesticated animals. 26

Isolation of E. coli from Yeongsan River, Jeonnam Province, Korea 28

Horizontal fluorophore-enhanced rep-PCR DNA fingerprinting 28

Phylogenetic grouping and virulence gene identification 29

Results and Discussion 30

Phylogenetic grouping patterns 30

Virulence gene distribution 36

Conclusions 47

Chapter 3. High Diversity and Abundance of Antibiotic Resistant Esherichia coli Isolated from Humans and Farm Animal Hosts in Jeonnam Province, South Korea 50

Introduction 51

Materials and Methods 51

E. coli isolation 51

Horizontal fluorophore-enhanced rep-PCR DNA Fingerprinting 52

Antibiotic resistance analyses 53

PCR detection for antibiotic resistance genes and integrase 53

Bacterial conjugation 54

Plasmid profiling and Southern blotting 54

Statistical analysis 55

Results and Discussion 55

Genotypically unique strain determination with HFERP DNA fingerprints 55

Distribution and types of antibiotic resistance genes 68

Chapter 4. Genotypic and Phenotypic Trends in Antibiotic Resistant Pathogenic Escherichia coli Isolated from Humans and Farm Animals in South Korea 75

Introduction 76

Material and Methods 78

E. coli isolation 78

Phylogenetic grouping and virulence gene identification 80

Antimicrobial resistance test 80

Statistical analyses 81

Results and discussion 81

Occurrence of genomically unique E. coli isolates 81

Virulence gene distribution 82

Genomic similarity and population structure of ExPEC strains 87

Occurrence of antibiotic resistant pathogenic E. coli 89

Chapter 5. Use of Barcoded Pyrosequencing and Shared OTUs to Determine Sources of Fecal Bacteria in Watersheds 95

Introduction 96

Materials and Methods 98

Calculation of colony forming unit (CFU) 98

Fecal and Environmental DNA extraction 98

Pyrosequencing 102

Analyses of β diversity of fecal bacteria community 103

Determination of pyrosequencing reads shared between fecal and environmental samples. 104

Results and Discussion 104

Species richness and diversity in fecal bacteria community characterized by barcoded pyrosequencing 104

Comparison of fecal bacterial diversity by relative abundance of each taxon 107

Application of barcoded pyrosequencing derived shared OTUs for microbial source tracking 108

Chapter 6. Summary 125

REFERENCES 129

List of Tables

Table 1. Sources and numbers of E. coli isolates used in this study 27

Table 2. The occurrence of E. coli strains with virulence genes and phylogenetic groups 37

Table 3. Comparison of the virulence gene patterns in unique E. coli strains obtained from different human and domesticated animal sources. 42

Table 4. Assignment of unique strains to phylogenetic groups by using HFERP DNA fingerprint and Jackknife analyses 46

Table 5. Number of feces, E. coli isolates and unique E. coli isolates used in this study 57

Table 6. Occurrence of antibiotic-resistant Escherichia coli isolates obtained from various hosts. 58

Table 7. Antibiotic resistance and integron distribution patterns in E. coli isolates that were resistant to more than 10 antibiotics from various hosts. 64

Table 8. Occurrence of multiple antibiotic resistance (MAR) and integrons in E. coli isolates obtained from various hosts 66

Table 9. Occurrence of antibiotic resistance genes in E. coli isolates carrying plasmids 69

Table 10. Number of samples and isolates used in this study 78

Table 11. Virulence genes detected among E. coli isolates obtained from humans and animals. 83

Table 12. Distribution patterns of virulence genes among E. coli used in this study. 86

Table 13. Antibiotic resistance patterns among potential pathogenic E. coli isolates used in this study. 91

Table 14. Occurrence of multiple antibiotic resistant pathogenic diarrheagenic E. coli and ExPEC isolates. 92

Table 15. Virulence gene profiles among potentially pathogenic E. coli resistant to multiple antibiotics. 93

Table 16. Detailed information of the samples used in this study 100

Table 17. Comparison of microbial diversity richness estimation of the 16S rRNA gene. 105

Table 18. Number of shared OTUs identified with Mothur program 110

Table 19. Ten most frequently observed genus in Proteobacteria within shared OTUs between the goose fecal sample and the environmental samples. 113

Table 20. Bacterial community distance matrix between fecal and environmental samples calculated with UniFrac service. (H: Human; C: Chicken; D: Duck; G: Goose; BC: Beef cattle; DC: Dairy cattle; and S: Swine) 114

Table 21. Suggestion of fecal contamination parameter using results obtained from the barcoded pyrosequencing 117

Table 22. Density of pyrosequencing reads in each fecal sample of the ten most 120

List of Figures

Figure 1. HFERP image separation by dual wave-length scanning 14

Figure 2. Distribution of phylogenetic groups among E. coli isolates obtained from humans and domesticated animals 31

Figure 3. Seasonal variations in phylogenetic group distribution among E. coli obtained from the Yeongsan River, Jeonnam province, Korea 35

Figure 4. Distribution of virulence genes among phylogenetic groups of E. coli obtained from humans and domesticated animals 39

Figure 5. Genetic relatedness of E. coli strains possessing virulence genes. 44

Figure 6. Phylogenetic grouping analysis of HFERP DNA fingerprints using MANOVA 46

Figure 7. Cluster analysis of E. coli isolates from various hosts based on antibiotic resistance profiles 60

Figure 8. Occurrence of E. coli isolates resistant to different number of antibiotics 61

Figure 9. Occurrence of integron-carrying E. coli isolates compared with the total number of antibiotics to which the strains were resistant. 67

Figure 10. Plasmid profiles of E. coli strains resistant to more than 9 antibiotics. 71

Figure 11. Plasmid mediated antibiotic gene transfer 72

Figure 12. Genomic similarities among extraintestinal pathogenic E. coli (ExPEC) isolates obtained from humans and animals in Gwangju, JangSeong, Naju and Damyang. 87

Figure 13. Genomic relatedness E. coli isolates obtained from poultry sources. 89

Figure 14. Rarefaction analysis of OTUs obtained from fecal samples used in this study. 106

Figure 15. Cluster profiling based on the relative abundance in each taxon with all reads(A), reads with 100% query coverage and a similarity !97% (B), and cluster profiling based on sequence divergence with all reads (C). Legend: H: Humans; C: Chickens; D: Ducks; G: Geese; BC: Beef cattle; DC: Dairy cattle; and S: Swine. 108

Figure 16. The relative abundance of phyla among OTUs from humans, animals, and environment samples 109

Figure 17. Cluster analysis of shared species between fecal and environmental samples using UniFrac service for samples obtained at sites Y1(A), Y2 (B), and Y3(C). Legend: H: Humans; C: Chickens; D: Ducks; G: Geese; BC: Beef cattle; DC: Dairy cattle; and S: Swine. 111

Figure 18. Cluster analysis of shared species between fecal and environmental samples at the phylum level for samples obtained at sitesY1 (A), Y2 (B) and Y3(C). Legend: H: Humans; C: Chickens; D: Ducks; G: Geese; BC: Beef cattle; DC: Dairy cattle; and S: Swine. 112

Figure 19. Percentage of phylum (A), shared OTUs (B), and total density ratio (C) within shared OTUs between fecal and environmental sample. 116

According to the World Health Organization (WHO), diarrhea kills 2.2 million people globally each year, therefore, and fecal contamination of water is often cited as being of major concern. Microbial source tracking (MST) was developed to identify sources of fecal contamination by using microorganisms, such as bacteria and virus. The hypothesis behind mainstream MST methods is that gut microorganisms are different (genetically, phenotypically, or ecologically) among various animal species, including humans, and is likely due to gut environmental differences. Therefore, comparing significant differences among gut microorganisms in each animal species may lead us to identify the source of fecal microorganisms in environments.

This present work mainly focused on developing a MST method which is also suitable to monitor fecal pollution in the Yeongsan River, Jeonnam Province, South Korea. The present work also included surveillance studies of antibiotic resistance and virulence gene distribution among a fecal indicator bacterium, E. coli, isolated from humans and animals in Jeonnam Province, South Korea.

The following experiments were conducted.

1. Comparison of phylogenetic diversity of Escherichia coli isolates obtained from humans and animals

2. Occurrence of antibiotic resistant E. coli isolates among humans and domesticated animals

3. Genotypic characterization of antibiotic resistant potential pathogenic E. coli

4. Comparison of microbial community in human feces, animal feces, and environments

E. coli was first genotypically characterized using phylogenetic grouping based on a using multiplex PCR (Clermont phylogenetic type: A, B1, B2, and D). In addition, horizontal fluorophore-enhanced rep-PCR (HFERP) DNA fingerprint analysis was performed with all E. coli isolates obtained from humans and animals in order to characterize E. coli genetic diversity further. Our major finding in this study was that the Clermont type E. coli phylo-group, B2, was found to be very rare. E. coli belonging to phylo-group B2 was generally known to be potential pathogens, although they can be found in almost all humans and animals. The occurrence of each phylogenetic group was similar between chicken and duck isolates, beef and dairy cattle isolates, and healthy human and patient isolates, although only about 70% phylogenetic grouping was correctly classified by HFERP DNA fingerprint analysis. Secondly, we have conducted a surveillance study of antibiotic resistant E. coli in Jeonnam Province. The E. coli isolates used for the study were obtained from humans and animals raised in the Yeongsan River watershed. Consequently, the occurrence of antibiotic resistant E. coli observed in this study may indirectly characterize those found in the Yeongsan River. We hypothesized that the excessive use of antibiotics in South Korea may have triggered the high incidence of antibiotic resistant E. coli. In fact, many antibiotic resistant pathogenic bacteria, such as antibiotic resistant Campylobacter, Salmonella and E. coli O157:H7 have been isolated in South Korea. In this study, we have also demonstrated how easy antibiotic resistance determinants were transferred from one to the other E. coli isolates. Thirdly, we have conducted another surveillance study for virulence genes among the E. coli isolates by using several known primers designated to amplify known virulence genes of diarrheagenic E .coli and extraintestinal pathogenic E. coli (ExPEC) strains. By comparing virulence profiles against HFERP DNA fingerprints, we found that some unique genotypes among poultry isolates are likely ExPEC. In addition, many of them were found to be highly antibiotic resistant. In conclusion, E. coli genetic diversity in this region appeared to be different from other countries. We have conducted HFERP DNA fingerprint analysis to differentiate sources of E. coli isolates, but the results were not congruent among four locations (Damyang, Naju, Gwangju, Jangseong) where E. coli isolates were obtained.

Lastly, we developed a novel MST method by applying recently developed new DNA sequencing technology, so called “next generation sequencing (NGS)”. Using universal 16S rRNA primers targeting V1 to V3 regions, fecal bacteria communities were explored with NGS, and operational taxonomic units (OTUs) obtained for environmental bacteria communities were compared to those obtained from fecal bacteria communities. Our results suggest that the use of shared OTUs between environmental and fecal bacteria communities indicate the source of fecal contamination. This method has also showed that the percentage of shared OTUs between feces and environments likely correlates with the degree of fecal contamination. However, further study is required to develop this concept to take part in MST practices.