Single base extension reaction and its applications in environmental microbiological studies

Chapter 5: A High-throughput and Quantitative Hierarchical Oligonucleotide Primer Extension HOPE-based Approach to Identify Sources of Fecal Contamination in Water Bodies 5.1 Abstract

SINGLE BASE EXTENSION REACTIONS AND THEIR

APPLICATIONS IN ENVIRONMENTAL MICROBIOLOGICAL STUDIES

HONG PEI YING

NATIONAL UNIVERSITY OF SINGAPORE

2009

SINGLE BASE EXTENSION REACTIONS AND THEIR

APPLICATIONS IN ENVIRONMENTAL MICROBIOLOGICAL STUDIES

HONG PEI YING

(B Eng (Hons), NUS)

A THESIS SUBMITTED FOR THE DEGREE OF DOCTOR OF PHILOSOPHY DIVISION OF ENVIRONMENTAL SCIENCE AND

ENGINEERING NATIONAL UNIVERSITY OF SINGAPORE

2009

Acknowledgements

I am grateful for the presence of many special people throughout these five years They encourage, inspire and provide me with ample opportunities to discover my passion for research A big thank you to the following persons:

Professor Liu Wen-Tso, thank you for providing all the right formulations to groom me

Professor F Michael Saunders He is the one who encouraged me to ask questions and is always there to render ready assistance whenever needed

Dr Wu Jer-Horng He is a great mentor who tolerated with my ignorance when I first started my candidature

Dr Zhang Rui for his sincere advice

Dr Zang Kaisai for the FISH images displayed at the last chapter of this dissertation

The great seniors: Ezrein Shah Bin Selamat is my Yoda in microarray fabrication; Dr Johnson Ng Kian Kok provides his excellent guidance in writing; Dr Pang Chee Meng gives blunt criticisms and thought-provoking scientific discussions to improve myself; Dr Chen Chia Lung shows me how patience is the key to eventual success

The ESE administrative and lab staff They shoulder the administrative headaches for me

All my lab friends and the cat Thank you for your great company!

My family My attitude towards research has made me a stranger to home Thank you for showering me with your never-ending understanding and concern

For every ending lies a new beginning

Table of Contents

Acknowledgements… ……….……….…….…….…i

Table of Contents……….…….……… ……… …….….……… iii

Summary……… …… ix

List of Tables……… ….…… xii

List of Figures……… ………xiv

Abbreviations……….….… xvii

Publications……… … xix

Chapter 1: Introduction 1.1 Background and problem statement 1

1.2 Objectives and aims 5

1.3 Organization of thesis 11

Chapter 2: Literature Review 2.1 In the Land of the One-Eyed King, do we see the whole picture? 13

2.2 Limitations in the SSU rRNA-based approach 13

2.2.1 Sampling and storage bias 14

2.2.1.1 Sampling design 14

2.2.1.2 Sample storage 16

2.2.2 DNA extraction bias 17

2.2.3 PCR bias 19

2.2.3.1 False-negatives and false-positives 19

2.2.3.2 Microbial variability 21

2.2.3.3 Oligonucleotide primers 22

2.2.4 Whole genome amplification (WGA) bias 24

2.3 Molecular tools 26

2.3.1 Identifying and profiling the change in microbial community 27

2.3.1.1 Microarray 27

2.3.1.2 16S rRNA gene libraries and metagenomics 30

2.3.1.3 Terminal restriction fragment length polymorphism 32

2.3.2 Quantifying the abundances of bacterial targets 33

2.3.2.1 Whole cell hybridization 33

2.3.2.2 Quantitative PCR (Q-PCR) 35

2.3.2.3 Hierarchical oligonucleotide primer extension (HOPE) 36

2.4 Environmental microbiological studies 40

2.4.1 Human gut and fecal microbiota 40

2.4.2 Water quality and human health 41

2.5 Concluding remarks 44

Chapter 3: Materials and Methods 3.1 Fecal samples 45

3.1.1 Human feces 45

3.1.2 Pig feces 46

3.1.3 Cow feces 46

3.1.4 Dog feces 46

3.2 Contaminated aqueous samples 47

3.2.1 Municipal wastewater 47

3.2.2 Swine wastewater 47

3.2.3 Aqueous medium contaminated with cow and dog feces 47

3.2.4 Blind samples 47

3.2.5 Environmental water samples 48

3.3 Bacterial reference strains 49

3.3.1 Bacteroides spp 50

3.3.2 Bifidobacterium spp 50

3.3.3 Animal-specific uncultivated Bacteroidales 51

3.3.4 Non-Bacteroides and Non-Bifidobacterium spp 51

3.4 DNA extraction 52

3.5 Oligonucleotide primers 52

3.5.1 Oligonucleotide primers for PCR 52

3.5.2 Oligonucleotide primers for HOPE 53

3.5.3 Oligonucleotide primers for Q-PCR 57

3.5.4 Oligonucleotide probes for FISH-flow cytometry 59

3.6 Polymerase chain reaction 59

3.7 Hierarchical Oligonucleotide Primer Extension (HOPE) 60

3.7.1 Applied Biosystems PRISM 3130 60

3.7.2 Beckman Coulter CEQ 8000 61

3.7.3 Calculation of relative abundances of bacterial targets 62

3.8 Quantitative PCR (Q-PCR) 63

3.9 Fluorescent in-situ hybridization- flow cytometry (FISH-FC) 63

3.9.1 Cell fixation 63

3.9.2 Cell permeabilization 64

3.9.3 Hybridization 64

3.9.4 Flow cytometry 64

3.10 Statistical and in silico analyses 65

3.10.1 Primer design 65

3.10.2 In silico analysis of probe specificity 65

3.10.3 Multi Dimensional Scaling 66

3.10.4 Correlation analysis 66

3.11 Nucleotide sequence ascension number 66

Chapter 4: Relative Abundance of Bacteroides spp in Stools and Wastewaters as Determined by Hierarchical Oligonucleotide Primer Extension (HOPE) 4.1 Abstract 67

4.2 Introduction 68

4.3 Results 74

4.3.1 PCR amplification 74

4.3.2 Specificity of primer extension 76

4.3.3 Calculation of calibration factor (CF) values 77

4.3.4 Electropherograms of tested samples 78

4.3.5 Relative abundance of Bacteroides spp as determined by HOPE 80

4.3.6 Abundances of Bacteroides spp as quantified by Q-PCR 86

4.3.7 Relative abundance of Bacteroides spp in wastewaters 86

4.4 Discussion……… 89

Chapter 5: A High-throughput and Quantitative Hierarchical Oligonucleotide

Primer Extension (HOPE)-based Approach to Identify Sources of Fecal

Contamination in Water Bodies

5.1 Abstract 94

5.2 Introduction 95

5.3 Results 99

5.3.1 Primer specificities 99

5.3.2 Detection limits of PCR assays and HOPE 102

5.3.3 HOPE-based approach on feces 102

5.3.4 HOPE-based approach on aqueous samples 106

5.3.5 HOPE-based approach on blind samples 107

5.3.6 HOPE-based approach on environmental waters 110

5.3.7 Validation by Q-PCR 111

5.3.8 Validation by fecal coliform test 113

5.4 Discussion 114

Chapter 6: Hierarchical Oligonucleotide Primer Extension (HOPE) as a Time- and Cost-effective Approach for the Quantitative Determination of Bifidobacterium spp in Infant Feces 6.1 Abstract 118

6.2 Introduction 119

6.3 Results 120

6.3.1 Primer and probe specificities 120

6.3.2 Detection limits of newly designed HOPE primers 122

6.3.3 Relative abundance of Bifidobacterium spp 123

6.3.4 Statistical correlation analysis 126

6.4 Discussion 127

Chapter 7: Conclusion and Recommendations 7.1 Conclusion 131

7.2 Recommendations 135

7.2.1 Method development 135

7.2.2 Environmental study 136

7.2.2.1 Human gut and fecal microbiota 136

7.2.2.2 Water quality and fecal source tracking 136

References………138

Summary

The study of microbial ecology involves identification and quantification of

functionally important microorganisms, which in turn are important steps necessary to achieve homeostasis in both natural and engineered ecosystems A small subunit (SSU) rRNA-based approach can bring new insights to our understanding of the microbial

community, but is fraught with technical shortcomings at almost every step A large suite

of molecular tools would allow one to complementarily utilize the most appropriate set of experimental tools to address specific questions, therefore minimizing limitations and biases associated with individual SSU rRNA-based method Till today, quantitative-PCR (Q-PCR) remains to be the most powerful tool to quantify abundances of bacterial targets

in environmental sample, but is limited in its throughput capabilities This drives the need

to develop hierarchical oligonucleotide primer extension (HOPE) as an alternative method that can complement Q-PCR and meet the needs for high-throughput quantitative

measurement of bacterial targets

To achieve this overall objective, this dissertation first demonstrated in chapter 4 the applicability of HOPE to examine microbiota in feces and fecal-contaminated

environment Nineteen oligonucleotide primers that target 14 functionally important

Bacteroides spp are designed and applied in HOPE to detect and quantify their

abundances in feces and contaminated wastewaters The findings show that the 14

Bacteroides spp are human-associated and are present in human feces at abundances that are five and two-fold higher than in non-human feces and wastewaters, respectively The

relative abundances of Bacteroides spp as quantified by HOPE are comparable to those

reported in published literature and those obtained using conventional methods like

Q-PCR Compared to Q-PCR which can only achieve up to five-plexing, HOPE is

demonstrated to achieve a higher throughput per reaction, and is able to detect bacterial targets that comprise at least 0.1% of the total amplified 16S rRNA These strengths are therefore potentially useful to address hypotheses related to microbial ecosystems

This study further illustrates the application of HOPE to address hypotheses

related to water quality and human health in two separate chapters In chapter 5, it is

hypothesized that the abundances of certain Bacteroidales are significantly different in

various host populations like humans, pigs, cows and dogs By quantifying for their abundances and performing statistical evaluation of these data, it will be possible to identify the source of fecal contamination A total of 12 primers are optimized for HOPE reactions, and are used to target and quantify for the abundances of human-associated,

pig-, cow- and dog-specific Bacteroidales When statistically evaluated in a

multi-dimensional scaling plot, it is observed that samples contaminated with feces from the same host population clustered well together and are distanced apart from those of other origins The method is further applied to blind samples and environmental waters, and is able to correctly identify fecal contamination that originates from humans, pigs and cows Unlike conventional fecal coliform test which can only indicate impaired water quality, this study has demonstrated that a HOPE-based approach is able to indicate both impaired water quality and the fecal contamination sources The HOPE findings can therefore facilitate relevant authorities to make a more effective compliance decision

In chapter 6, it is hypothesized that the abundances of genus Bifidobacterium may

be markedly different in the fecal microbiota of infants with and without eczema

However, inconsistent findings in the abundances of Bifidobacterium spp have prevented

precise conclusions to their role in modulating the immune response of infants towards eczema HOPE is demonstrated as a time- and cost-effective method to determine the

abundances of nine Bifidobacterium spp in 30 fecal samples It is observed that the

microbial compositions in infants are highly dynamic Most notably, a large portion of the

genus Bifidobacterium remains undetected by the current set of HOPE-based primers and

should be identified by other conventional methods prior to quantification by HOPE Therefore, by complementarily utilizing HOPE with other molecular tools (e.g., T-RFLP and gene libraries), a better elucidation of the roles of microbiota in modulating the

immune response can be achieved

In summary, this study has demonstrated HOPE as a rapid and high-throughput method that can be used to address hypotheses related to microbial ecosystems in a time- and cost-effective manner The high adaptability of this method also means that primers developed to target other bacterial targets can easily be added to new HOPE reaction tubes The method can also be complementarily utilized with other molecular tools to provide a better understanding of the microbial ecosystem

List of Tables

Table 2.1: 47F primer and the primer modifications that can be made to increase the

primer coverage……… 23

Table 2.2: List of waterborne bacterial, protozoal and viral pathogens……….… 43

Table 3.1: Medical information of the 10 infants……… 45

Table 3.2: List of oligonucleotide primers used for PCR……… 53

Table 3.3: HOPE primers that were designed to target human-specific Bacteroides spp at various hierarchical levels (species, group, order and domain level)……… 54

Table 3.4: Host-specific primers used in HOPE……… 55

Table 3.5: HOPE primers that were designed to target human-specific Bifidobacterium spp at various hierarchical levels (species, group, and domain level) 56

Table 3.6: Primers targeting Bacteroides spp at various taxonomical levels were used in Q-PCR to validate against HOPE results……… 57

Table 3.7: Primers used in Q-PCR to validate HOPE on environmental waters……… 58

Table 3.8: Oligonucleotide probes used in FISH-flow cytometry to hybridize to the 16S rRNA genes of Bifidobacterium spp……… 59

Table 4.1: Relative abundance of bacterial targets in the human, pig and cow feces, together with those present in the influent of a municipal and swine wastewater treatment plant, were determined individually using multiplexing HOPE……… 83

Table 4.2: Comparison of the abundance of Bacteroides at different hierarchical levels, obtained from various published literature……… 88

Table 5.1: Animal-specific uncultivated Bacteroidales clones and their nucleotide sequences at the primer binding region……… 101

Table 5.2: Results of blind samples validated by HOPE……… 108

Table 5.3: Results of environmental water samples as evaluated with HOPE, Q-PCR and

conventional fecal coliform (FC) test……… 112

Table 6.1: List of Bifidobacterium primers……… 121

Table 6.2: The coverage of oligonucleotide primers and probes targeting Bifidobacterium

spp……… 122 Table 6.3: Non-parametric correlation analysis for the relative abundance of

List of Figures

Figure 1.1: The schematics of this dissertation……… 5

Figure 2.1: A typical schematic of the SSU rRNA-based approach……… 14 Figure 2.2: A rarefraction curve……… 16

Figure 2.3: A plot of the PCR amplicon mass against the number of cycles………… 20

Figure 2.4: Whole genome amplification……… 25 Figure 2.5: The schematics of a multiplexing single base extension reaction………… 37

Figure 2.6: The schematics of hierarchical oligonucleotide primer extension……… 39

Figure 3.1: Map of the sampling site……… 49

Figure 4.1: Phylogenetic representation of different cultivated and uncultivated

Figure 4.2: Schematic illustration of HOPE PCR amplicons, fluorophore-labeled

ddNTPs, DNA polymerase and oligonucleotide primers were mixed together

to form a HOPE reaction mixture……… 72 Figure 4.3: A graphical representation of PCR amplicon mass obtained at different

number of thermal cycles……… 75 Figure 4.4: The actual electropherograms of HOPE products obtained in the presence of

total PCR-amplified 16S rRNA genes from Bacteroides thetaiotaomicron

BCRC10624 clone……… 78 Figure 4.5: The actual electropherograms of HOPE products obtained in the presence of

total PCR-amplified 16S rRNA genes from one of the human fecal

microbiota……… 79

Figure 4.6: Statistical interpretation of the HOPE results (a) Abundances of B fragilis

cluster, B fragilis subcluster, and the P distasonis cluster relative to the total

amplified 16S rRNA genes were averaged for the individual sample types

(b) B massiliensis, B vulgatus, B eggerthii, B uniformis, B intestinalis, B

caccae and B fragilis were detected, and their average relative abundances within the B fragilis cluster are illustrated A predominant portion within the B

Figure 4.7: Statistical interpretation of the HOPE results A multidimensional scaling

(MDS) plot obtained by first performing a Euclidean distance square-root transformation on the raw data to generate a similarity matrix The matrix then undergoes a 100-iterations MDS analysis……… … 85 Figure 5.1: Phylogenetic tree constructed based on the partial 16S rRNA gene sequences

that were available in NCBI and the sequences of the uncultivated

Figure 5.2: Box-whiskers plots of the relative abundances of human-associated and

animals-specific bacterial targets……… 104 Figure 5.3: Multidimensional scaling (MDS) plot obtained based on a Euclidean distance

square-root transformation of the abundances of bacterial targets in the feces, aqueous waters and blind samples……… … 105

Figure 6.1: The relative abundances of genus Bifidobacterium in the feces sampled from

all individuals at one month, three months and 12 months after birth… 124

Figure 6.2: The relative abundances of Bifidobacterium spp against the total

Bifidobacterium The abundances were averaged from the infants in the

respective health group and time point, and were quantified by both HOPE (○) and FISH-flow cytometry (∆)……… 125

Figure 7.1: A possible schematic that complementarily utilizes various molecular tools to

address hypotheses related to health and engineering system………… 134

Abbreviations

AOB Ammonia Oxidizing Bacteria

ATP Adenosine triphosphate

CARD-FISH Catalyzed Reporter Deposition Fluorescence In Situ Hybridization

DGGE Denaturing Gradient Gel Electrophoresis

DNA Deoxyribonucleic Acid

dNTP Deoxynucleotiside Triphosphate

ddNTP Dideoxynucleoside Triphosphate

EDTA Ethylenediaminetetraacetic Acid

FISH Fluorescent In Situ Hybridization

FISH-FC Fluorescent In Situ Hybridization combined with Flow Cytometry

HOPE Hierarchical Oligonucleotide Primer Extension

NOB Nitrite Oxidizing Bacteria

MAR-FISH Microautoradiography Fluorescent In Situ Hybridization

MDS Multi-Dimensional Scaling

OTU Operational Taxonomic Unit

PBS Phosphate-buffered Saline

PCR Polymerase Chain Reaction

Q-PCR Quantitative Polymerase Chain Reaction

RDP Ribosomal Database Project

RISA Ribosomal Inter-genic Spacing Analysis

SBE Single Base Extension

SDS Sodium Dodecyl Sulfate

spp Plural form for species

sp Singular form for species

SNP Single Nucleotide Polymorphism

T-RFLP Terminal Restriction Fragment Length Polymorphism

UV-VIS Ultraviolet-visible spectrum

WGA Whole Genome Amplification

Publications

Journal articles

1 Hong, P.-Y., J.-H Wu, and W.-T Liu (2008) Relative abundance of Bacteroides

spp in stools and wastewaters as determined by hierarchical oligonucleotide primer extension Applied and Environmental Microbiology 74: 2882-2893

2 Hong, P.-Y., J.-H Wu, and W.-T Liu (2009) A high-throughput and quantitative

hierarchical oligonucleotide primer extension (HOPE)-based approach to identify sources of fecal contamination in water bodies Environmental Microbiology 11: 1672-1681

3 Hong, P.-Y., G.C Yap, B.W Lee, K.Y Chua, and W.-T Liu (2009) Hierarchical

oligonucleotide primer extension (HOPE) as a time- and cost-effective approach

for the quantitative determination of Bifidobacterium spp in infant feces Applied

and Environmental Microbiology 75: 2573-2576

4 Wu, J.-H., P.-Y Hong, and W.-T Liu (2009) Quantitative effects of position and

type of single mismatch on single base extension Journal of Microbiological Methods 77: 267-275

Conference presentations

1 Liu, W.-T., and P.-Y Hong Hierarchical oligonucleotide primer extension

(HOPE) for quantitative and qualitative analyses of biotic contaminants The 235thAmerican Chemical Society National Meeting, April 6 to 10, 2008, New Orleans, USA (accepted for oral)

2 Hong, P.-Y., and W.-T Liu Hierarchical oligonucleotide primer extension

(HOPE) for quantitative and qualitative analyses in environmental microbiological studies The 108th American Society of Microorganisms General meeting, June 1

to 5, 2008, Boston, USA (accepted for poster)

3 Hong, P.-Y., J.-H Wu, and W.-T Liu Hierarchical oligonucleotide primer

extension (HOPE) for quantitative multiplexing analyses of PCR-amplified 16S ribosomal RNA genes The 12th International Symposium of Microbial Ecology Bi-annual meeting, August 17 to 22, 2008 Cairns, Australia (accepted for poster)

4 Hong, P.-Y., J.-H Wu, and W.-T Liu Development of hierarchical

oligonucleotide primer extension (HOPE) to identify sources of fecal

contamination in water bodies The 12th International Symposium of Microbial Ecology Bi-annual meeting, August 17 to 22, 2008 Cairns, Australia (accepted for poster)

5 Hong, P.-Y., and W.-T Liu Development of hierarchical oligonucleotide primer

extension (HOPE) to rapidly profile the predominant gut flora The 12th

International Symposium of Microbial Ecology Bi-annual meeting, August 17 to

22, 2008 Cairns, Australia (accepted for poster)

“The elegant and powerful tools of molecular biology continue to bring new dimensions

to our experimental approaches, and we seem to be at a stage in our science at which the rate-limiting step to important discoveries is not the availability of adequate technology, but rather our ability to ask the right questions; to design the right experiments.”

John Breznak, PhD Michigan State University

Chapter 1 Introduction

1.1 Background and problem statement

The study of microbial ecology is important to achieve homeostasis in both natural and engineered ecosystems To illustrate, by identifying new microorganisms like

anammox bacteria (i.e., Candidatus Brocadia, Kuenenia, Scalindula and

content from wastewaters in a cost-effective manner (Ye and Thomas, 2001) Similarly, in engineered nitrification processes, factors such as abundance ratio of ammonia oxidizing bacteria (AOB) to nitrite oxidizing bacteria (NOB) are hypothesized to affect the fragile mutualism between these two bacterial groups, and could upset the nitrification process

(Graham et al., 2007) A study which involves the identification and quantification of

their abundances would thus be an essential step towards achieving a better nitrification efficacy

To identify microorganisms, past studies had relied on providing favorable growth conditions to cultivate bacteria However, culture-based identification is selective for a

particular group of microorganisms (Wagner et al., 1993), does not provide quantitative

determination of the microorganisms that are not cultivated, and often requires a long culturing time In contrast, a molecular-based analysis, in particular a 16S rRNA gene-based approach, would allow a systematic identification and quantification of both

cultured and yet-to-culture microorganisms

In the 16S rRNA gene-based approach, the first step usually includes extraction of genomic DNA, followed by PCR amplification of the 16S rRNA genes The total

amplified 16S rRNA genes is then analyzed based on a suite of molecular tools like the 16S rRNA gene clone libraries, terminal restriction fragment length polymorphism (T-

RFLP), DNA microarray, denaturing gradient gel electrophoresis (DGGE) and so on In other instances, whole-cell hybridization like fluorescent in situ hybridization (FISH) can

be used to provide a semi-quantitative measurement of the abundances of bacterial targets

in the community without any prior DNA extraction and PCR

While these technological advances continue to progress at an alarmingly fast pace

to bring new insights to our understanding of the microbial community, a paradox also lies beneath The molecular-based approach is fraught with technical shortcomings at almost every step For example, certain DNA extraction procedures may favor gram-negative bacteria to gram-positive ones (Madigan, 2009), and PCR may

disproportionately amplify selective bacterial targets that are present in the microbial

community (von Wintzingerode et al., 1997) Furthermore, taxonomical resolution and

detection limits of molecular-based analytical tools vary and can result in differences in

the detected microbial diversity (Forney et al., 2004)

Although molecular-based methods have their individual shortcomings, a large suite of them would allow one to choose the most appropriate set of experimental tools to address specific questions Molecular-based methods like 16S rRNA gene clone libraries and DNA microarrays can be utilized to identify the bacterial species present, and

fingerprinting tools like T-RFLP and DGGE can be used to rapidly monitor the spatial and temporal changes in the microbial community However, to quantify for the

percentage abundances of bacterial species, these molecular methods only provide a quantitative measurement of abundances Till today, quantitative-PCR (Q-PCR) remains

semi-to be the most powerful semi-tool semi-to quantify abundances of bacterial targets in the

environmental sample Q-PCR can provide excellent quantitative limit of detection (down

to a few copies of target fragments) but requires standard curves to be established for

individual targets (Jung et al., 2000) Furthermore, given the limited number of spectrally

distinct fluorophores that are available in UV-VIS spectrum (Levsky and Singer, 2003),

Q-PCR can currently obtain a maximum of five-plexing in each reaction (Vet et al., 1999;

http://www.biorad.com) This drives the need to develop alternative method that can complement Q-PCR, and to meet the needs for high-throughput quantitative measurement

of bacterial targets

Single base extension (SBE) reaction, which is primarily used in molecular

biology studies for detection of single nucleotide polymorphisms (SNP), has displayed

good multiplexing capabilities in previous studies (Gwee et al., 2003; Wang et al., 2003)

However, the method has not been applied to most environmental microbiology studies, and has not been demonstrated for quantitative measurements

Recently, Wu and Liu have demonstrated the feasibility of applying SBE-based reactions to achieve quantitative detection of bacterial targets at various taxonomical levels (Wu and Liu, 2007) The method, termed as hierarchical oligonucleotide primer extension (HOPE), can be done by first arranging multiple oligonucleotide primers

targeting 16S rRNA genes at different phylogenetic specificities into a single reaction tube, and allowing SBE reaction to take place The primers modified with different

lengths of poly-A at the 5’ end can anneal to their complementary bacterial targets and be extended with a single fluorescently labeled dideoxynucleoside triphosphate (i.e ddATP, ddTTP, ddCTP and ddGTP) The extended primers are separated and identified in a genetic analyzer based on fragment size and dye color Based on fluorescence intensity detected, the peak areas of a given extended specific primer is determined, and the ratio

of the area to that for a common reference primer is calculated and used to determine the relative abundance of the bacterial targets of interest (Wu and Liu, 2007)

Wu and Liu further developed a 10-plexing reaction targeting Bacteroides spp at

domain, group and species levels and applied the reaction to detect the relative

abundances of Bacteroides spp in influent and effluent of domestic wastewater treatment

plant (Wu and Liu, 2007) They successfully showed that the method could achieve single mismatch discrimination and had a detection limit of as little as 0.01 to 0.05% of the total set of PCR-amplified 16S rRNA genes HOPE has therefore demonstrated

promising capabilities, and this further encourages the development and application of HOPE to ecological systems like feces and fecal-contaminated waters as illustrated in this doctoral study

1.2 Objectives and aims

The primary objective of this doctoral study is to develop HOPE as a rapid and high-throughput molecular method that can provide quantitative measures of the relative abundances of bacterial targets among a pool of total amplified 16S rRNA genes In microbial ecosystems that require quantitative measurements of the abundances of

bacterial targets, this method can complement the available suite of tools, and facilitate better understanding of microbial community in different ecological systems

Figure 1.1 The schematics of this dissertation

Feces

Fecal contaminated environment

Ecosystem Implications

Health status

of host Water quality

Current molecular methods

to study the ecosystem

1 Q-PCR

2 FISH

3 16S rRNA gene libraries

4 T-RFLP And so on…

Primary objective

Develop HOPE to add on to the suite of methods

A1

Further study these implications through

quantification of selected bacterial targets

Case study 1 : Water quality

Hypothesis Quantifying the

abundance of host-specific

Bacteroidales and statistically

evaluating the dataset will provide an

alternative means to identify the fecal

correlate to the occurrence of allergy

However, current molecular tools do not provide a time- and cost-effective means to examine this hypothesis Direct implication

Indirect implication A1: Specific aim in Chapter 4

A2: Specific aim in Chapter 5

A3: Specific aim in Chapter 6

Feces

Fecal contaminated environment

Ecosystem Implications

Health status

of host Water quality

Current molecular methods

to study the ecosystem

1 Q-PCR

2 FISH

3 16S rRNA gene libraries

4 T-RFLP And so on…

Primary objective

Develop HOPE to add on to the suite of methods

A1

Further study these implications through

quantification of selected bacterial targets

Case study 1 : Water quality

Hypothesis Quantifying the

abundance of host-specific

Bacteroidales and statistically

evaluating the dataset will provide an

alternative means to identify the fecal

correlate to the occurrence of allergy

However, current molecular tools do not provide a time- and cost-effective means to examine this hypothesis Direct implication

Indirect implication A1: Specific aim in Chapter 4

A2: Specific aim in Chapter 5

A3: Specific aim in Chapter 6

To demonstrate the primary objective, this study has identified feces and

fecal-contaminated environment as important ecosystems to be studied As illustrated in Figure 1.1, the use of HOPE in fecal and fecal-contaminated ecosystems will first be validated Subsequently, HOPE will be applied to examine two areas related to the fecal

environment, namely, fecal contamination in water bodies (i.e., water quality

microbiology) and health-related fecal microbiota

1.2.1 Validate the use of HOPE in feces and fecal-contaminated environment

Specific aim 1 To demonstrate HOPE as a robust method that is applicable in feces and

fecal-contaminated environment, and that the method is able to rapidly detect and quantify the relative abundances of bacterial targets in such ecosystems

Brief background Feces contain approximately 1014 bacterial cells per gram (Franks

et al , 1998), and several phyla like Firmicutes, Bacteroidetes, Proteobacteria and

Actinobacteria make up predominant of the fecal microbiota (Eckburg et al., 2005) The fecal microbiota is representative of the gut microbiota (Palmer et al., 2006; Palmer et al.,

2007), and plays important roles in supplying to the catabolic and metabolic needs of the host (Guarner and Malagelada, 2003) Based on niche exclusion theory, endemic gut/fecal microbiota colonizes the gut and inhibits occurrence of enteroinvasive microorganisms, therefore modulating a stable health condition in its host (Walter, 2008) When discharged

to the environment (e.g., water bodies), pathogens in feces can be transmitted through oral route and impose health concerns Therefore, feces and fecal-contaminated environment are identified as important ecosystems that will be studied in this dissertation

Experimental approach This study serves to first validate the use of HOPE in feces

and fecal-contaminated environment through the use of genus Bacteroides as a model group of bacterial targets The genus Bacteroides is highly predominant in human feces

and can play a significant role in maintaining the host health (Guarner and Malagelada,

2003) For example, species like B vulgatus and B caccae can colonize the surface area

of intestinal mucosa and inhibit adherence of enteroinvasive pathogens (Finegold and

Jousimies-Somer, 1997) The numerical dominance of different Bacteroides spp is

therefore hypothesized to correlate to human health Furthermore, B fragilis, B

thetaiotaomicron and B vulgatus are reported to be ubiquitous in feces, and detection of

these species in environmental waters would indicate impaired water quality (Kreader,

1995) By further quantifying for their abundances, persistence of Bacteroides in the

environment can be better understood It is therefore important to utilize a molecular

method that quantifies the abundance of each individual Bacteroides sp that is present in

feces and fecal-contaminated environments It is assumed that HOPE is a suitable

molecular method that can be used to quantify the abundances of Bacteroides in feces and

fecal-contaminated samples It is further assumed that the abundances as quantified by HOPE will be comparable to those obtained from conventional molecular methods To

examine these assumptions, 14 Bacteroides spp are chosen and their abundances in

human and animal feces, as well as in wastewaters, are quantified by HOPE To further evaluate for comparability, the abundances are compared against those obtained by Q-PCR

If the abovementioned assumptions are valid, HOPE is suitable for quantifying bacterial targets in feces and fecal-contaminated environment, and can be utilized to look into other studies related to this ecosystem, namely water quality microbiology and health-related microbiota (Figure 1.1)

1.2.2 Water quality microbiology (case study one)

Specific aim 2 To utilize HOPE as a rapid and high-throughout method for the

identification of fecal contamination sources

Brief background Besides fecal coliforms, certain clusters of order Bacteroidales are host-specific and associate with fecal contamination (Dick et al., 2005a), and detection of

them in environmental waters would indicate impaired water quality Previous studies

utilized PCR assays to denote for the occurrence of host-specific Bacteroidales in the environment (Kreader, 1998; Dick et al., 2005a) However, PCR is prone to the

occurrence of false-positives (Field et al., 2003), and may not provide accurate

identification of fecal contamination sources

Experimental approach As shown in Figure 1.1, it is hypothesized that the

abundances of these host-specific Bacteroidales are markedly different in various host

populations, and a statistical evaluation of these dataset can provide an alternative means

to identify the sources of fecal contamination A HOPE-based approach is utilized to look

into this hypothesis The abundances of host-specific Bacteroidales in feces and

contaminated aqueous samples are first quantified, and used to derive a database that can aid in subsequent identification of the fecal contamination sources in unknown samples

The method is then applied to blind samples and environmental waters, and the findings

are validated against Q-PCR and conventional fecal coliform test

1.2.3 Health-related fecal microbiota (case study two)

Specific aim 3 To utilize HOPE as a time- and cost-effective method for the

examination of health-related bacterial populations

Brief background Compared to other bacterial groups, genus Bifidobacterium is highly predominant in the infants’ gut and fecal microbiota (Harmsen et al., 2000) The numerical dominance of Bifidobacterium spp leads to the assumption that the genus plays

an important role in modulating the immune response of infants towards atopic eczema

(Mah et al., 2006; Mah et al., 2007) In particular, certain species in the genus may

correlate closely to occurrence of eczema in infants at different stages of infancy For

example, Bifidobacterium pseudocatenulatum is more commonly found in infants with eczema than in healthy infants (Gore et al., 2008) To draw conclusive findings to support

this hypothesis, studies have to be performed on a statistically representative large sample

size and on numerous Bifidobacterium spp that may potentially correlate to the

occurrence of eczema in infants Current studies have obtained inconsistent conclusion to support this hypothesis, primarily because of the lack of appropriate molecular tools to

examine the abundances of numerous Bifidobacterium spp in a time- and cost-effective

manner

Experimental approach To circumvent the abovementioned problem, HOPE is used

to examine the relative abundance of Bifidobacterium spp in a high-throughput and effective manner A total of eight Bifidobacterium-targeting primers are used to examine

cost-the abundances of Bifidobacterium spp in 30 infant feces, and cost-the abundances as

quantified by both HOPE and FISH-flow cytometry are statistically compared

1.3 Organization of thesis

The thesis is subdivided into the following chapters

• Chapter 2: Literature review

This chapter aims to provide a comprehensive review of the shortcomings in the 16S rRNA gene-based approach, and the various molecular methods that can be utilized to address specific questions related to microbial ecology

• Chapter 3: Materials and methods

This chapter lists the experimental materials and methodologies used in the entire doctoral study

• Chapter 4: Relative abundances of Bacteroides spp in stools and wastewaters

as determined by hierarchical oligonucleotide primer extension

This chapter details how HOPE can be used to determine the relative abundance of

predominant Bacteroides spp present in fecal microbiota and wastewaters

• Chapter 5: A high-throughput and quantitative hierarchical oligonucleotide

primer extension-based approach to identify sources of fecal contamination in water bodies

This chapter demonstrates the use of HOPE to identify the origins of

contamination in feces, contaminated aqueous samples, blind samples and

environmental waters

• Chapter 6: Hierarchical oligonucleotide primer extension as a time- and

cost-effective approach for the quantitative determination of Bifidobacterium spp

in infant stools

This chapter demonstrates the use of HOPE to examine Bifidobacterium spp in

the fecal microbiota of infants with and without eczema

• Chapter 7: Conclusion and recommendations

The overall conclusion and recommendations for future studies are presented here

Chapter 2 Literature Review

2.1 In the Land of the One-Eyed King, do we see the whole picture?

Ever since the SSU rRNA (i.e., 30S ribosomal subunit containing 16S rRNA gene)

was proposed as an evolutionary marker to study Bacteria and Archaea (Woese and Fox,

1977), a SSU rRNA-based approach is utilized to identify microorganisms that are present

in the ecosystem Changes in the profiles and abundances of bacterial targets are

correlated to environmental factors, and the findings are hypothesized to provide insights

to the functional roles of microorganisms in maintaining the homeostasis of a microbial ecosystem

Although we continue to accrue new insights to our understanding of the microbial community using the molecular approach, it is also recognized to be fraught with

technical shortcomings that blind our understanding of the microbial community (Forney

et al., 2004) In this chapter, we illustrate the shortcomings in a typical schematic of a SSU rRNA-based approach Upon recognizing that the paradox is inevitable, this chapter aims to discuss the essential steps that can be adopted to minimize the error

In the following subsections, problems associated with each step of a SSU based approach will be systematically discussed (Figure 2.1)

rRNA-Figure 2.1 A typical schematic of the SSU rRNA-based approach

2.2.1 Sampling and storage bias

2.2.1.1 Sampling design

A proper sampling design involves making decisions on the sampling mode and the number of samples to collect (Green, 1979) Depending on the prior information of sampling area, random or systematic sampling can be performed Random sampling is useful when the target-of-interest is relatively homogenous within the sampling area In contrast, systematic sampling is particularly useful for locating target-of-interest at

concentrations or abundances that are significantly different from the background

Objectives, hypotheses

1 Microarray e.g Phylochip

2 16S rRNA gene libraries

3 Terminal restriction fragment

Objectives, hypotheses

1 Microarray e.g Phylochip

2 16S rRNA gene libraries

3 Terminal restriction fragment

length polymorphism (T-RFLP)

1 Whole cell hybridization

2 Quantitative PCR

(USEPA, 2002) The sampling number can also be decided based on statistical

information derived in prior samplings or after a cost-benefit evaluation (Baú and Mayer, 2007) Alternatively, one can collect more samples at sites with high level of uncertainty, therefore improving the confidence level

Although a good sampling design is essential, it is recognized that the total

microbial community can never be fully sampled as the relationship between the number

of species and the area (i.e., sampling area) is described by the Power-law (Garcia Martin and Goldenfeld, 2006)

Power-Law: S = cA z

Where, S is the number of species, c is the intercept in the log-log space, A is the area, and z is the measure of the rate of species turnover across space (approximately 0.0475 to 0.0959 in microorganisms) Given that c and z are numerical constants, this relationship marks that the number of species would increase with the sampling area, and the total microbial diversity cannot be fully sampled with finite resources

Similarly, the inability to study the total microbial diversity is also shown in a rarefraction curve, correlating the number of new operationally taxonomic units (defined based on 97% sequence similarity in 16S rRNA gene) against the number of clones

sequenced As the number of sequences increases, the slope of rarefraction curves

changes from linear to quasi-linear and never reaches plateau (Figure 2.2) These

observations clearly illustrate that a good sampling design, although essential, is never able to fully capture the total microbial community in a complex ecosystem