In Partial Fulfillment of The Requirements

For The Degree of

Master of Science

December 2014

Assessing local water quality in Saudi Arabia and its impact on food safety

Thesis by

Dhafer Alsalah

1

EXAMINATION COMMITTEE APPROVAL FORM

This thesis by Dhafer Alsalah is approved by the examination committee.

Committee Chairperson: Dr. Peiying Hong

Committee Member: Dr. Uli Stingl

Committee Member: Dr. Pascal Saikaly

2

© December 2014

Dhafer Alsalah

All Rights Reserved

3

ABSTRACT

Assessing local water quality in Saudi Arabia and its impact on food safety

Thesis by

Dhafer Alsalah

Saudi Arabia produces a majority of its fruits and vegetables locally in small-scale

production farms. These farms utilize groundwater as the main source of irrigation water. The

water-regulating authorities in Saudi Arabia rely on traditional culturing methods to monitor

coliforms as indicators of microbial contamination. These methods are time-consuming, do not

address the sources of contamination, and do not permit assessment on the associated health risk.

To address these knowledge gaps, the study investigates the sources of contamination in eight

wells northeast of Mecca, Saudi Arabia. The study focuses on the potential impact on

groundwater quality due to a nearby chicken farm and urban runoffs from human residential

areas. Besides performing conventional methods to determine nutrient content and to enumerate

coliforms, quantitative PCR using four host-associated primer sets were used to distinguish

microbial contamination from humans and livestock. High-throughput sequencing was also

performed to determine the relative abundance of several genera associated with opportunistic

pathogens. Bacterial isolates were cultivated from the vegetable samples harvested from these

farms, and were characterized for their phylogenetic identities. Lastly, the study collates the

information to perform quantitative microbial risk assessment due to ingesting antibiotic-

resistant Listeria monocytogenes, Pseudomonas aeruginosa and Enterococcus faecalis in these

vegetable samples.

4

ACKNOWLEDGEMENTS

I would start by thanking Dr. Hong for her consistent support and guidance throughout the

journey of this thesis. Moreover, I would like to thank her for teaching me more about academia

in general.

Also, I would like to distinguish the farmers from Wadi Yamaniyah and from Badalah for

opening their farms to a stranger like me, and for allowing access to their water and food

samples. These samples serve to form an essential part of this thesis.

Last but not least, I would like to thank my friends and my colleagues Moustapha Harb, David

Mantilla, Nada Aljassim, and Dr. Mohammed Ikram Ansari in Dr. Hong’s group. Also, I would

like to send many thanks to my friends from outside the group for their endless encouragement.

5

TABLE OF CONTENTS

Page

EXAMINATION COMMITTEE APPROVAL FORM

1

COPYRIGHT PAGE

2

ABSTRACT

4

ACKNOWLEDGEMENTS

4

TABLE OF CONTENTS

5

LIST OF ABBREVIATIONS

7

LIST OF FIGURES

8

LIST OF TABLES

9

1. Introduction

11

2. Literature Review

15

2.1. Irrigation water in Saudi Arabia 15

2.2. Water standards and regulations in Saudi Arabia 16

2.3. Methods of detecting microbial contamination 19

2.3.1. Multiple-tube fermentation as conventional method to enumerate coliforms 19

2.3.2. Molecular-based approach to monitor anthropogenic contamination 21

3. Materials and Methods

23

3.1. Sampling site 23

3.2. Sampling procedure 24

3.3 Water quality analysis

25

3.4. Vegetable sampling, processing, and cultivation

26

3.5. Bacteria counting and isolation

26

3.6. DNA extraction 26

6

3.7. Phylogenetic identification of bacterial isolates 27

3.8. Quantitative PCR

28

3.9. High-throughput sequencing and data analysis

29

3.10. Risk assessment

30

4. Results and discussion

32

4.1. TOC and TN analysis

32

4.2. Microbiological quality

33

4.3. Fecal source tracking

35

4.4. Microbial community analysis

37

4.5. Bacterial counts

39

4.6. Risk assessment

42

5. Conclusion

43

REFERENCES

46

APPENDIX AND SUPPLEMENTARY DATA 50

7

LIST OF ABBREVIATIONS

ARB

Antibiotic Resistant Bacteria

B.G. Brilliant Green Agar

Ceft. Ceftazidime

FAO Food and Agriculture Organization of the United Nation

FC Fecal Coliforms

Mac MacConkey Agar

MMRA Ministry of Municipal and Rural Affairs

MoWE Ministry of Water and Electricity

MPN Most Probable Number

OTU Operation Taxonomical Units

P Peppers

PCR Polymerase chain reaction

PME Presidency of Meteorology and Environment

qPCR Quantitative PCR

T Tomatoes

TC Total Coliforms

TN Total Nitrogen

USEPA United States Environmental Protection Agency

WHO World Health Organization

8

LIST OF FIGURES

Page

Figure 1. Foodborne illness cases in Saudi Arabia from 1989 to 2003 12

Figure 2. Sampling site 23

Figure 3. Heterotrophic plate count per 50 g of peeled skin 39

Figure 4. ARB count per 50 g of peeled skin 41

Supplementary Figures

Figure S1. Microbial richness in water samples up to 3250 sequences deep 51

Figure S2. Cells/L of the water samples using flow cytometry analysis 51

Figure S3. Rarefaction curves of water samples 52

9

LIST OF TABLES

Page

Table 1: Total irrigated area per province per irrigation method 16

Table 2: Ministry of Municipal and Rural Affairs (MMRA) and Ministry of Water and Electricity (MoWE) irrigation water standards for 2003 and 2006.

17

Table 3: Saudi Arabian PME regulation for groundwater 19

Table 4: Detailed information on sampling site 24

Table 5: List of the host-specific markers 29

Table 6: TOC and TN analysis 33

Table 7: Total coliforms and fecal coliforms count 34

Table 8: Copy numbers per liter of the host-specific markers 36

Table 9: Summary of the sources of fecal contamination 36

Table 10: Percentage of the genera associated with pathogens in the microbial community

38

Table 11: The relative abundance of ARB 41

Table 12: The daily and annual risks from pathogenic ARB 44

10

Supplementary Tables

Table S1: Salinity of the water samples 52

Table S2: Heterotrophic plate count per 50 g of peeled skin 53

Table S3: ARB count per 50 g of peeled skin 53

Table S4: Total microbial richness of the water samples including all sequences 53

11

1. Introduction

The production of food is intricately linked to water consumption. Agricultural

production remains to be the thirstiest sector relative to the needs for domestic and industrial

use. 70-80% of the freshwater supplies are withdrawn to feed the never-ending thirst for water

by the agriculture sector (Dabbagh and Abderrahman, 1997; Zaharani et al., 2011). While each

person drinks on average 2 to 4 L of water per day, they can consume up to 2000 to 5000 L of

virtual water from the three meals that they have on a daily basis (http://virtualwater.eu/). In

many water-stressed countries like Saudi Arabia, the production of food is conducted at an

unsustainable rate. Groundwater supplies are pumped up from underground at an alarming rate,

and stored in well waters for agricultural irrigation.

The rate of consuming non-renewable groundwater supplies has meant that underlying

aquifers can be vulnerable to surface infiltration from urban runoffs. This problem is further

compounded by the rapid rate of urbanization and other industrial activities in Saudi Arabia

(Ellis, 1999), which can result in anthropogenic contamination of the groundwater resources.

Groundwater with a compromised quality because of human and livestock activities can impact

the food safety during the pre-harvest stage. For example, human activities and the lack of

proper sewage systems in Mpraeso, Ghana are the main reason of the degradation of the

groundwater quality (Omari and Yeboah-Manu, 2012). According to the Ministry of Health in

Saudi Arabia, the number of reported foodborne illnesses has been increasing steadily for more

than a decade (Figure 1) (Al-Mazrou, 2004). While a portion of these cases arise due to poor

hygienic practices during food preparation, some cases can also arise because the groundwater

quality has been impacted and compromised food were sold to the public consumers.

12

Figure 1. Foodborne illness cases in Saudi Arabia from 1989 to 2003 adapted from (Al-Mazrou, 2004)

There are a limited number of papers that detail the nutrient contents and coliform

numbers of coliforms in groundwater of Saudi Arabia (El-Din et al., 1993; Alaa el-Din et al.,

1994; Al-Salamah and Nassar, 2009; AlOtaibi, 2009; Al-Salamah et al., 2011). In those studies,

it was found that in some of the groundwater wells, nitrate contents can be greater than 45 mg/L

which exceeded the level recommended by the local regulating agencies. Furthermore, fecal

coliforms were present in 21.4% of the well waters that were tested throughout Saudi Arabia

over a one-year period in 1989 (El-Din et al., 1993; Alaa el-Din et al., 1994). There are

numerous factors that can account for the high nutrient content in the groundwater. Agricultural

soils in Saudi Arabia are impoverished and farmers have to rely on heavy usage of fertilizers and

pesticides to achieve a good crop yield. The soil and sand conditions are highly permeable, and

0

100

200

300

400

500

600

1989

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

Food

born

e ill

ness

cas

es

Year

13

nutrients can infiltrate easily into the vadose zones of groundwater aquifers (Al-Salamah and

Nassar, 2009; Al-Salamah et al., 2011; Zaharani et al., 2011). On the same note, only 40% of

Saudi Arabia land mass is connected by sewage pipelines (O'Sullivan, 2010). Human settlement

areas that are not served by sanitary networks have to rely on local septic tanks to store their

municipal wastes. Lack of maintenance on these septic tanks can lead to leakage of the municipal

wastes, which in turn contaminate the groundwater supplies.

Although these studies provide insights to the groundwater quality in Saudi Arabia, many

of these studies date back to before the year 2000, and groundwater quality may have changed

over the years. Monitoring efforts on the groundwater quality should be renewed to determine

the quality of these waters and their impact on food safety if these waters were to be used for

agricultural irrigation.

This current study serves to address the knowledge gaps relating to the quality of

irrigation waters used by farmers in Saudi Arabia, and the consequential effect on food safety.

The aims of this study are to:

i. investigate the quality of well irrigation water,

ii. track sources of fecal contamination

iii. explore the consequences on food safety due to poor water quality

To achieve these aims, conventional methods to determine the nutrient contents and coliform

numbers were complemented by molecular-based approaches. This include the use of qPCR to

perform microbial source tracking and the use of Ion PGM sequencing to determine the relative

abundance of genera associated with opportunistic pathogens. The food samples were analyzed

through enriching and cultivating for bacterial isolates that were resistant to antibiotics. These

bacterial isolates were then phylogenetically identified based on their 16S rRNA genes. The

14

data were then utilized for quantitative microbial risk assessment to estimate the associated risk

imposed on public consumers.

15

2. Literature Review

2.1. Irrigation Water in Saudi Arabia

Freshwater resources are very scarce in Saudi Arabia. The country has no surface water

bodies of any kind excluding the Red Sea and the Persian Gulf, which cannot be utilized in

agriculture except after going through an expensive desalination process. Most of the

agricultural farms are located in the mid-eastern region of Saudi Arabia where the only sources

of water in those locations have been rainfall and non-renewable groundwater. Together the

agriculture industry utilizes up to 80% of the freshwater supplies in Saudi Arabia to produce a

small variety of food products (Al-Jasser, 2011; Zaharani et al., 2011). The food products

traditionally include cucumber, tomato, peppers, lettuce and alfalfa.

According to the Food and Agriculture Organization of United Nations (FAO), 32% of

the agriculture in Saudi Arabia uses traditional irrigation methods rather than modern irrigation

methods (Table 1) (Food and Agriculture Organization of the United Nations, 2009). A typical

form of traditional irrigation method is surface flooding, which requires excessive amount of

irrigation water and can result in surface runoff or infiltration of nutrient-rich waters into the

underlying aquifers. Modern irrigation methods include drip irrigation, otherwise also referred

to as trickle irrigation, and overhead sprinkler irrigation. In drip irrigation, water is delivered

drop by drop near to the roots of plants (Goyal, 2012). For overhead spray systems, irrigation

water is sprayed through pressurized nozzles into the open air and the water particulates then

settle down onto the leaves and roots of the crops. Among these three mentioned irrigation

methods, flooding and overhead sprinklers would introduce more water to come in contact with

the vegetables (Fox et al., 2012). This would mean that if irrigation water is compromised with

poor microbial quality, there is a higher likelihood that bacterial particulates would adhere onto

vegetables and crops. This can affect the quality of crops and raise concerns on food safety.

16

Table 1: Total irrigated area per province and per irrigation method.

Region Traditional irrigation Modern irrigation Area (ha) % Area (ha) %

Riyadh 43010 15% 243275 85% Makkah 43924 98% 1032 2% Madinah 26618 93% 2020 7% Qassim 15541 7% 208712 93% Eastern Province 16081 15% 92987 85%

Asir 22232 99% 296 1% Tabuk 5113 11% 42057 89% Hail 12368 10% 116139 90% Northern Boarders 19 14% 114 86%

Jazan 177375 99% 1995 1% Najran 8811 69% 4008 31% Baha 2658 98% 55 2% Jouf 11688 11% 93224 89% Total 385438 32% 805914 68%

2.2. Water quality standards and regulations in Saudi Arabia

As such, irrigation water quality should be monitored and standards should be in place to

ensure that water quality is well-regulated. There are three water-regulating parties in Saudi

Arabia tasked to ensure the quality of irrigation water. These three government agencies include

Ministry of Water and Electricity (MoWE), Ministry of Municipal and Rural Affairs (MMRA),

and the Presidency of Meteorology and Environment (PME). Guidelines and Standards by these

organizations differ in some aspects and comply with each other in other aspects of the water

quality (Tables 2 and 3).

However, these standards are not well-enforced and regulated despite the number of

regulating parties in Saudi Arabia. Nor are monitoring efforts on the groundwater quality

routinely conducted. Furthermore, the current standards relating to microbial pollution are

17

limited to the TOC, TN and the abundance of fecal coliforms, and stop short on determining the

sources of contamination. Determining the sources of contamination would allow proper

management strategies to be put in place to deter subsequent contamination events. The lapse in

monitoring and enforcement arise mainly because most of the agriculture in Saudi Arabia are

small-scale production and conducted on private properties. Poor literacy rates among farmers

would also mean that best management practices are not conveyed effectively to them.

18

Table 2: Ministry of Municipal and Rural Affairs (MMRA) and Ministry of Water and Electricity (MoWE) irrigation water standards for 2003 and 2006.

Parameter

Unit Unrestricted irrigation

Restricted irrigation

2003-MMRA

2006-MoWE

2003-MMRA

2006-MoWE

Physical parameters Floatable materials Absent Absent Absent

Total suspended solids (TSS) mg/L 10 10 40 40 pH 6–8.4 6–8.4 6–8.4

Turbidity NTU 5 5 Chemical parameters Organic chemicals parameters

Biochemical oxygen demand (BOD5) mg/L 10 10 40 40 Chemical oxygen demand (COD) mg/L 50

Total organic carbon mg/L 40 Oil and grease mg/L Absent Absent Absent

Phenol mg/L 0.002 0.002 0.002 Inorganic chemicals parameters Chemical compounds

Total dissolved solids (TDS) mg/L 2000 2500 2000 2500 Chloride (Cl2) mg/L 100 Sulfate (SO4) mg/L 600

Ammonia (NH3–N) mg/L 5 5 5 Nitrate (NO3–N) mg/L 10 10 10

Free residual chlorine mg/L 0.2 0.5 0.5 Fluoride (F) mg/L 1 1

Biological parameters Fecal coliforms per 100 mL 2.2 2.2 1000 1000 Intestinal nematodes per litre No./L 1 1 1

Adapted from MoWE and MMRA

19

Table 3: Saudi Arabian Presidency of Meteorology and Environment regulations for groundwater quality.

Parameter Unit Groundwater

Biochemical oxygen demand (BOD5)

mg/L n/a

Chemical oxygen demand (COD)

mg/L n/a

Chemical parameters

Ammonia (free, as NH3)

mg/L 0.3

Inorganic nitrogen

(as nitrate and nitrite)

mg/L 30

Biological parameters

Fecal coliforms per 100 mL

CFU/100mL 10

2.3. Methods of detecting microbial contamination

2.3.1. Multiple-tube fermentation as a conventional method to enumerate coliforms

The standard method 9221B is a standard total coliform multiple-tube fermentation

technique that is recommended by most regulatory bodies to be used for enumerating total

coliforms in a sample (EPA, 2010).This method relies on the use of selective media to enrich for

total coliforms through a three phases process. In the first phase, lauryl tryptose broth is used in

the presumptive test, and the cell suspension incubated at 35 oC for 24 h. Tubes that produced

Adapted from PME 24/03/2012

20

gas (i.e., carbon dioxide) and acid and exhibited cell growth were further tested for total

coliforms in the second phase (i.e., confirmation test). The brilliant green lactose bile broth is

used in the confirmation test, and the cell suspension is again incubated at 35 oC for 48 h.

Formation of gas in any amount during this test would constitute a positive result. Further

confirmation of the presence of fecal coliforms can be performed using EC broth, and the cell

culture is then incubated at 44.5 oC for 24 h. For each of the three phases, enumeration of the

bacterial density can be performed using the multiple tube method to generate the Most Probable

Number (MPN). For example, a range of different sample volumes (e.g. 10 ml, 1 ml and 0.1 ml)

can be inoculated into the lauryl tryptose broth, with five replicates for each of the different

tested sample volume. At the end of the incubation, if all five replicates for the tested sample

volumes generate gas production, the combination of positives would be recorded as 5-5-5,

signifying > 1600 MPN/100 ml of the sample volume.

An apparent limitation of this conventional method is that it requires up to 48 h before the

presence of total coliforms can be determined. An additional 24 h is required to determine for the

presence of fecal coliforms. This time-consuming process would mean that a fecal contamination

event cannot be communicated to the public in a timely manner (Csuros, 1999; Field et al.,

2003). Furthermore, the presence of the fecal coliforms does not allow regulators to distinguish if

the fecal contamination event originates from anthropogenic sources or from animals. The

multiple-tube fermentation technique depends on detecting the gas produced from the lactose

fermentation and the acidic activity. The general assumption in this technique is that the use of

selective media would permit only the growth of coliforms, and that the gas produced would be

exclusively associated with the total coliforms and fecal coliforms. However, non-coliform and

lactose-fermenting bacteria like Aeromonas can generate false-positives and does not

conclusively indicate that fecal contamination events had happened (Hendricks and United,

21

1979). On the other hand, certain strains of E. coli do not produce gas when fermenting lactose

and can generate false negative results during the multiple fermentation tube method (Corry et

al., 1996).

Summing up, these limitations of multiple fermentation tube method suggest that a sole

reliance on this conventional method may not protect the public against potential health hazards

arising from contamination of the irrigation water.

2.3.2. Molecular-based approaches to monitor anthropogenic contamination

Recent advances in molecular-based approaches allow better detection sensitivity and

quantitative analysis of the microbial contaminants. Molecular methods like high-throughput

sequencing and, quantitative PCR (qPCR) provide a higher detection sensitivity and more

precise quantitation than traditional culturing methods. To illustrate, host-specific primer assays

were developed to detect for fecal contamination originating from humans and chickens (Allsop

and Stickler, 1985; Bernhard and Field, 2000; Dick et al., 2005; Lu et al., 2007; Weidhaas et al.,

2010; Toledo-Hernandez et al., 2013). Most of the developed host-specific primers are designed

to target for the functional genes of host-specific Bacteroidales. Thus far, Bacteroides spp. and

Prevotella spp. that are associated to humans or are specific to dogs, cows, elks or pigs have

been identified. In contrast, Bacteroidales are usually present in low abundances in chickens and

do not enable effective detection of fecal contamination originating from chickens. Therefore,

chicken-specific primers tend to be designed to target non-Bacteroidales group (e.g.

Brevibacterium) that are uniquely associated with chicken feces.

A positive detection of these microbial targets, coupled with their copy numbers would

enable one to determine the source of the contamination. This is illustrated by field studies that

were conducted in various geographical locations and in different water matrices (Allsop and

22

Stickler, 1985; Bernhard and Field, 2000; Dick et al., 2005; Lu et al., 2007; Silkie and Nelson,

2009; Weidhaas et al., 2010; Hagedorn et al., 2011; Toledo-Hernandez et al., 2013). For

example, Schriewer et al. detected the presence of Bacteroidales in surface waters in the

California coast of USA, and correlated the presence of Bacteroidales to that of pathogens

(Schriewer et al., 2010). Zhang and coworkers performed microbial source tracking on karst

groundwater samples and detected high abundances of human-associated Bacteroides in some of

these waters (Zhang et al., 2014).Unlike the fecal coliform tests, microbial source tracking

permits identification and differentiation of the source of contamination is identified. This allows

concerned parties to strategize and devise appropriate mitigation measures. Moreover, compared

to the traditional culturing methods, molecular methods like qPCR can be completed in 2-4

hours. This required amount of time is much shorter than the 48 h required in conducting the

presumptive, confirmed, and completed tests of multiple-tube fermentation technique.

Besides qPCR, high-throughput sequencing using platforms like 454 FLX, Illumina

HiSeq or MiSeq and Ion PGM provides insights to the microbial communities of environmental

samples. Depending on the number of sequencing reads achieved, low abundances of rare taxa or

genera associated with opportunistic pathogens can be detected. Comparative analyses can also

be performed on the data generated from high-throughput sequencing to understand the

microbial ecology of the environmental systems. Previously, this approach has been

demonstrated to determine the planktonic communities in the Red Sea (Lee et al., 2011).

However, both qPCR-based microbial source tracking and high-throughput sequencing have not

been performed on monitoring for the irrigation water quality in Saudi Arabia.

23

3. Materials and Methods

3.1. Sampling site

The study area is a narrow valley north of Mecca with an approximate coordinates of 21.7 ˚

N, 40 ˚ E.

Figure 2: The sampling site in Badalah and Wadi Yamaniyah. Sampling locations are classified into two groups, namely group 1, which are wells within the 20 km distance from the chicken farm (i.e., wells 2,3,4,5,6,) and group 2, which are wells beyond the 20 km distance from the chicken farm (i.e., wells 1,7, 8).

Eight different wells were collected for their water samples (Figure 2). Three wells in

Wadi Yamaniyah were located east of a chicken farm with an up and running wastewater

treatment facility in Az Zemah labeled 6, 7, and 8, and three wells in Badalah were located

northwest of the chicken farm labeled 1, 2, and 3. Two sampling trips were made in March 2014.

For the first sampling trip, water samples were collected from the three wells in Wadi

Yamaniyah. For the second trip, another set of water samples were collected from wadi

24

Yamaniyah from the same wells. Water samples were also collected form the three wells in

Badalah, and two wells that were 1 km and 6 km downstream from the chicken farm in Az

Zemah labeled 4 and 5. Fresh vegetables were sampled from farms that utilize the well waters to

irrigate these crops Table 4 clarifies more details about the distances and the layout of the wells.

Table 4: Detailed information on sampling site and the type of samples collected from the various wells.

Site Well Name (Well number)

Veg. Samples

Upstream or Downstream

from the Chicken

farm

Distance from the Chicken

Farm (km)

Sample group

number

Wadi Yamaniyah

(Well 8) No Upstream 25.5 2

(Well 7)

Yes

(Tomatoes & peppers)

Upstream 24 2

(Well 6) No Upstream 17 1

Badalah

(Well 3)

Yes

(Tomatoes & peppers)

Downstream 5

1

(Well 2) No Downstream 19 1

(Well 1) No Downstream 32.5 2

1 km

Secondary Branch

(Well 4) No Downstream 1

1

6 km Main

Branch (Well 5) No Downstream 6

1

25

3.2. Sampling procedure

Pumps were activated to pump up the water samples from the wells. Water was flushed

for approximately 3 min prior collection into sanitized plastic bottles. Vegetable samples were

plucked by farmers and immediately placed into a plastic bag. All samples were placed into a

cooler and kept cool during transport. Once at the laboratory, samples were placed in 4 oC fridge

until further analyses.

3.3. Water quality analysis

All the water samples were analyzed for non-particulate organic carbon (NPOC) and total

nitrogen (TN) using the TOC Analyzer Wet Oxidation/NDIR Method Model from Shimadzu.

The water samples were filtered using a 0.45 µm membrane to eliminate the particulate organic

carbon (POC). A number of dilutions ranging from 1/8 to 1/20 were prepared for each water

sample to insure that NPOC and the TN fall within the valid range of the machine.

The microbial water quality of the samples was assessed using the USEPA method 9131,

which is the multiple-tube fermentation technique (EPA, 2010). A total of 15 tubes were used

with a volume of 10 ml of the broth. The first 5 tubes contained broth at 2X concentration so as

to permit adding 10 ml of the water samples to form the correct final concentration of broth.

Then, inverted vial were placed within the tubes to collect produced gas. After that, metal caps

were placed to cover the tubes and the tubes were autoclaved. Finally, 10 ml of the water sample

is added to the two folds tubes and 1 ml to the next 5 tubes and 0.1 ml to the final 5 tubes. First,

lauryl trypose broth was used as a medium for the presumptive test and tubes were incubated for

24 hours at 35 oC. Second, brilliant green lactose bile broth is used as a medium and incubated

at 35 oC. Similarly EC broth was used for the confirmed test and tubes were incubated at 44.5 oC.

26

Tubes showing gas production are positives. Positive tubes from the presumptive test were

moved to the confirmed test and the completed test stage to determine the MPN of the total

coliforms and the fecal coliforms.

3.4. Vegetable sampling, processing and cultivation

The skins of the tomatoes and peppers were peeled to collect 50 g from each. Then, the

50 g of each of the vegetables was blended with a 250 ml of deionized water until a relatively

homogenous mixture was obtained. 50 ml of the blend was mixed with 50 ml of Luria Broth and

incubated for 24 h at 37 oC to enrich the bacterial community. The enrichment was diluted by

104, 106, and 108, and 100 µl of each dilution was spread on the media plates. The media plates

include MacConkey (Weidhaas et al.) and Brilliant Green (B.G.) agar with and without

antibiotics. Two different types of antibiotics, namely meropenem and ceftazidime with a

concentration of 8 µg / ml, were used to test for the presence of antibiotic resistant bacteria

(ARB) in the vegetable samples.

3.5. Bacteria counting and isolation

Media plates that did not contain antibiotics were incubated for 24 h at 37 ˚C while media

plates that were spiked with antibiotics were incubated for 65 h as bacterial colonies grew slower

in the presence of antibiotics. The total number of bacterial colonies per 50 g of vegetable skin

was calculated based on:

CFU/50 g= (Dilution factor)(Plate count) X 100 ml of LB mixture X 250 ml of food blend 0.1 ml 50 ml of food blend

27

ARB colonies growing on the media plates were randomly picked and re-streaked twice to

acquire a pure culture.

3.6. DNA extraction

Water samples were filtered through 0.45 µm membrane. Volumes of the water samples

ranged from 2 L to 5 L. The membranes were cut into thin strips with sterile razor blades, and

extracted for the microbial genomic DNA using Ultra Clean soil kit (MoBio, US) based on

manufacturer’s protocol. Minor adjustments were applied to the protocol. Briefly, 12 μl of 100

mg/ml lysozyme (Bimboim and Doly, 1979) and 12 μl of 1mg/ml achromopeptidase (Simonet et

al., 1984) were added to the lysis solution and samples so as to provide a further enzymatic lysis

of the cell walls. The suspensions were incubated for 1 h before starting the protocol. Bacterial

isolates were extracted for their genomic DNA using the DNeasy Blood & Tissue kit (Qiagen,

US) based on manufacturer’s protocol. All genomic DNA concentrations were quantified by

Qubit Broad Range DNA assay or with Nanodrop.

3.7. Phylogenetic identification of bacterial isolates

The 16S rRNA genes of bacterial isolates were amplified by primers 11F (5`-

GTTYGATYCTGGCTCAG-3`) and 1492R (5`-GGYTACCTTGTTACGACTT-3`). PCR

reaction mixture was prepared using 1 µl of each primer, 25 µl Premix F (Epicentre

Biotechnologies, US), 0.3 µl Taq Polymerase (Takara, Japan), 22 µl of molecular biology grade

water and 1 µl of the DNA template. A negative control was used in each PCR run. The negative

control contained all the substances mentioned above in the PCR reaction mixture except the

DNA template. The thermal cycler program consisted of an initial denaturation temperature of 95

˚C for 3 minutes. Then, 30 cycles of denaturation for 30 seconds at 95 ˚C, annealing temperature

28

of 50 ˚C for 45 seconds, and an extension temperature of 72 ˚C follow the denaturation period

for one minute. Finally, a final stage of extension at 72 ˚C follows the 30 cycles for 10 minutes.

The PCR product was subsequently purified using the commercially available kit from Promega

which is Wizard® Genomic DNA Purification Kit. PCR amplicons were sent in for Sanger

sequencing at the KAUST Genomics core lab, and were sequenced using the 338F primer (5′-

ACTYCTACGGRAGGCWGC-3′). The sequencing results were identified by using BLASTN

(Basic Local Alignment Tool) against the 16S rRNA genes database.

3.8. Quantitative PCR

Microbial source tracking was conducted using a qPCR approach. Primer assays that

target the 16S rRNA genes of human-associated Bacteroides spp. (i.e., B. vulgatus, B. fragilis

and B. uniformis) and cow-specific Bacteroidales, as well as primer assays that target the

inorganic ion metabolism gene of B. fragilis unique in chicken feces (Allsop and Stickler, 1984;

Allsop and Stickler, 1985; Lu et al., 2007; Toledo-Hernandez et al., 2013) are listed in (Table 5).

The qPCR standards were prepared from B. vulgatus BCRC12903, B. uniformis JCM5828, B.

fragilis BCRC10619 and CP2-9 amplicon product from genomic DNA of chicken feces. The

gene amplicons were cloned into pGEM-T easy vector (Promega, US) and transformed into One

Shot ® TOP10 E. coli competent cells. Blue/white screening was performed to pick out clones

with gene inserts, and PureYieldTM Plasmid Miniprep System (Promega, US) was used to purify

the. Plasmids were sequenced for the gene insert region to ensure perfectly matched sequences to

the primer assays. The concentration of the plasmids with the insert was measured using The

Qubit Broad Range dsDNA assay, and copy numbers of plasmids per µl were determined. Then,

108,106,105,104,103 series of dilution were prepared to generate standard curves for each of the

primer assays. Amplifications to obtain standard curves were performed in triplicate, while test

29

amplifications and negative blanks were run in duplicates. Each reaction volume of 20 μl

contained 10 µl of FAST SYBR Green master mix, 0.4 µl of each primer (10 µM), 1 µl of DNA

template and 8.2 µl molecular biology grade water. The 7900 HT Fast protocol was used for

thermal cycling. The protocol includes 40 cycles of 1 s denaturation at 95oC and 60s of annealing

and extension. Dissociation curve analysis was included to detect non-specific amplification.

Table 5: List of the host-specific markers.

Primers Target Sequence (5’-3’) Amplicon size

927F Human-associated Bacteroides

vulgatus

GGG CCC GCA CAA GCG G 110

Bvg1016R ATG CCT TGC GGC TTA CGG C

927F Human-associated Bacteroides fragilis

GGG CCC GCA CAA GCG G 118

Bfrg1024R TCA CAG CGG TGA TTG CTC A

927F Human-associated Bacteroides uniformis

GGG CCC GCA CAA GCG G 110

Bufm1018R CTG CCT TGC GGC TGA CA

C367F Cow-specific Bacteroidales

GGA AGA CTG AAC CAG CCA AGT A 116

C467R GCT TAT TCA TAC GGT ACA TAC AAG

CP2-9F Chicken-associated Bacteroides fragilis

GTA AGA CAG CAA CCC CAT GTA 245

CP2-9R ACC TAT GGT TCA ACA CGC TTT A

3.9. High-throughput sequencing and data analysis

The universal barcoded primers 515F (5`-GTGYCAGCMGCCGCGGTA-3`) and 909R

(5`- CCCCGYCA-ATTCMTTTRAGT-3`) were used to amplify the gDNA of bacterial

community in the water samples. 50 μl of PCR mixture was prepared for each water sample. The

30

PCR mixture was composed of 1 μl of 10 μM 515F and 1 μl of 10 μM of 909R barcoded

primers, 25 μl of 2-fold premix F, 0.4 μl Ex Taq polymerase from Takara, and 1 μl of extracted

gDNA from the water samples described in section 3.6 with 55 °C for annealing temperature. A

C1000 Touch thermocycler (BioRad Labs Inc.) was used with the following conditions: an initial

denaturation stage at 95 °C for 3 minutes, followed by 30 cycles of denaturation at 95 °C for 30

seconds, annealing at 55 °C for 45 seconds and extension at 72 °C for 1 minute, and then a final

extension stage at 72 °C for 10 minutes.

Raw sequence reads were first trimmed for their barcode, adaptor and primer sequences.

Trimmed sequences were then checked for their quality by removing reads that are < 250 nt in

length and with Phred score < 20. Chimeras were identified on UCHIME (Edgar et al., 2011) by

referencing a core reference set that was downloaded from Greengenes (i.e., gold strains gg16 –

aligned.fasta). RDP Classifier was used for taxonomical assignments of the 16S rRNA gene

sequences at 95% confidence level (Cole et al., 2009). Sequences were aligned using the RDP

Infernal Aligner, and aligned sequences were binned for unique operational taxonomic units

(OTUs) identified at 97% 16S rRNA gene similarity. The cluster matrix generated from RDP

pipeline was then used in rarefaction analysis. Rarefaction curves were generated and shown in

Supplementary Figure S3. Microbial richness for each sample was denoted from the rarefaction

curves based on a defined sequencing depth of 3250 sequences.

3.10. Risk Assessment

Quantitative microbial risk assessment (QMRAwiki) was performed for genera

Pseudomonas, Listeria, and Enterococcus as they accounted for a higher relative abundance

among the isolates recovered from the vegetables (Table 11). Phylogenetic identification of

isolates denoted the presence of Listeria monocytogenes, Pseudomonas aeruginosa, and

31

Enterococcus faecalis, at an individual fraction of 0.23, 0.075, and 0.075 respectively relative to

the total counts of antibiotic resistant isolates. The risks arising from these bacterium species and

from consuming the vegetables were calculated based on an assumption of 70 kg average body

weight per person and that each individual consumed an average 2.9 g/kg/d (USEPA, 1997).

The risk from the opportunistic pathogenic species was characterized using the

exponential distribution model for daily and annual risks (Haas and Eisenberg, 2001):

P (response) = 1- e (-k x dose)

k The probability of an organism to survive to reach and infect

P The probability of infection or death

The dose was calculated by multiplying the fraction of Listeria monocytogenes or

Pseudomonas aeruginosa to the total counts of antibiotic resistant bacteria, followed by an

assumed probability of transmission of 2 x 10-6. The k constant of Listeria monocytogenes is

5.71 x 10-7 (Golnazarian et al., 1989; Czuprynski et al., 2003), that of P. aeruginosa is 6.93 x 10-

8 based on the ID50 (Hazlett et al., 1978; Lawin-Brussel et al., 1993), and that of E. faecalis is

8.77 x 10-9 based on the LD50 and assuming exponential model (Bourgogne et al., 2008).

Annual risk was calculated based on the equation below:

P annual = 1 – (1 – Pdaily) number of exposure days per year

32

4. Results and Discussion

4.1. TOC and TN Analysis

The total nitrogen and the total organic carbon of the wells water samples ranged between

15.21 – 61.33 mg/l and 10.63-70.60 mg/l, respectively (Table 6). Also, the TN and the TOC of

the water samples from Wadi Yamaniyah (Well 6, 7, and 8) show a very significant fluctuation.

The TN concentrations of water from all wells were > 15 mg/L (Table 6), which exceeded the

MoWE standards for total ammonia and nitrate (Table 2). Also, some of the water samples TN

exceeded the Presidency of Meteorology and Environment’s groundwater standards for TN of 30

mg/l (Table 3) (Environment, 2012). Several of the water samples, particularly those collected

from all three wells in the Badalah (Well 1, 2, and 3) site had an exceedance of > 40 mg/L of

TOC content which exceeded the MMRA standards for TOC (Table 2). The TN concentrations

in the water samples collected from wells within 20 km distance of the chicken farm were

significantly higher than those further away from the chicken farm (One-way ANOVA, P-value

= 0.00). The TOC in water samples collected from both sets of wells were however not

significantly different (One-way ANOVA, P-value = 0.89).

33

Table 6: Total organic carbon (Bourgogne et al.) and total nitrogen (TN) analysis.

4.2. Microbiological Quality

Only water samples collected from the wells at Badalah vegetable site ( Well 3) and the two

Well number Sample group

number

Average Standard deviation

TN TOC TN TOC

(mg/L) (mg/L) (mg/L) (mg/L)

1st Trip

Well 8 2 19.63 69.11 0.2 1.98

Well 7 2 23.02 16.09 0.06 0.20

Well 6 1 53.66 70.56 0.31 2.77

2nd Trip

Well 8 2 23.14 33.64 0.62 0.32

Well 7 2 37.60 34.91 0.98 14.17

Well 6 1 61.33 22.73 5.61 2.22

Well 3 1 49.52 68.75 0.62 0.46

Well 2 1 42.21 69.86 2.24 2.90

Well 1 2 15.21 70.60 0.13 0.26

Well 4 1 55.08 10.63 2.88 0.91

Well 5 1 40.81 12.14 0.45 0.01

34

wells that are 1 km and 6 km (Well 4 and Well 5) downstream from the chicken farm produced

gas during the presumptive test. All of these wells were located within the 20 km distance from

the chicken farm. The confirmed tests showed that the MPN of total coliforms were > 1600 per

100 ml. In particular, water samples from wells at Badalah veg site and at 1km downstream of

chicken farm had an exceedingly high 1600 MPN fecal coliforms per 100 ml (Table 7). This

number of fecal coliforms exceeded the current regulations of 2.2 MPN/100 ml for unrestricted

irrigation and 1000 MPN/100ml for restricted irrigation. The water collected from the 6 km well

had a lower MPN of fecal coliforms (i.e., 12 MPN/100 ml). In this study, the wells are located in

farms that produce vegetables through furrow irrigation technique. Therefore, irrigation water

comes into contact with the farmers quite often, and the microbial quality of these waters should

be compared against the regulations stipulated for unrestricted irrigation.

Well Presumptive (10ml-1ml-

0.1ml)

Total coliforms

(10ml-1ml-0.1ml)

Fecal coliforms

(10ml-1ml-0.1ml)

Total MPN /100 ml

Fecal MPN / 100 ml

Well 8 0-0-0 - - - -

Well 7 0-0-0 - - - -

Well 6 0-0-0 - - - -

Well 3 5-5-1 5-5-5 5-5-3 >1600 920

Well 2 0-0-0 - - - -

Well 1 0-0-0 - - - -

Well 4 5-5-5 5-5-5 2-1-2 >1600 12

Well 5 5-5-3 5-5-5 5-5-4 >1600 1600

Table 7: Total coliform and fecal coliforms analysis.

35

4.3. Fecal Source Tracking

Water samples collected from the upstream well of Wadi Yamaniyah (Well 8) during the

first sampling trip was positive for human fecal contamination, while samples collected from the

other two wells during the second trip to Wadi Yamaniyah were positive for human fecal

contamination (Well 6 and 7) (Table 8). The copy numbers of the B. fragilis ranged from 7.1 x

101 to 7.34 x 103 copies/L and were lower than that from B. vulgatus and B. uniformis. Water

samples collected from wells within the 20 km distance from chicken farm were positively

detected for the chicken-specific Bacteroidales marker. The abundance of the chicken-specific

Bacteroidales was lower than that of human-associated Bacteroides, and ranged from 2.0 x101 to

4.51x102 copies/L (Table 8). Table 10 summarizes the sources of fecal contamination.

The water samples collected from the 1 km and 6 km wells (Well 4 and 5) were positive

for fecal coliforms based on the MPN methods and qPCR-based microbial source tracking

further complemented the results by suggesting that the fecal coliforms likely originate from

human feces. In contrast, the fecal coliforms present in water samples collected from the Badalah

vegetable site (Well 3) likely originated from chicken feces since it turned negative for all the

human markers. However, there was no apparent correlation between the qPCR-based microbial

source tracking results with that obtained from MPN method for all remaining samples.

36

Table 8: Copy numbers per liter of the host-specific markers.

Table 9: Summary of fecal contamination sources.

Well Copies/L Normalized against amplification efficiency

1st Trip B. fragilis B. vulgatus B. uniformis CP2-9 1st CP2-9 2nd

Well 8 - 4.33x105

2.92x107 - -

Well 7 - - - 4.51x102 4.14x102

Well 6 - - - 1.99x102 2.13x102

2nd Trip - - - - -

Well 8 7.10x101 - - - -

Well 7 8.30x101 8.14x104 - - - Well 6 3.62 x102 2.72x105 - - -

Well 3 - - - 2.55 x102 3.10 x102 Well 2 - - - 2.0 x101 1.17 x102

Well 1 - - - 1.79 x102 1.09 x102

Well 4 7.33 x103

3.55 x105

3.53 x107 - -

Well 5 3.37 x103 - 5.40 x106

4.5 x101

2.30 x102

Well Source of Contamination

1st Trip Wadi Y 2nd Trip Wadi Y

Well 8 Human - Well 7 Chicken Human Well 6 Chicken Human Well 3 Chicken Well 2 Chicken Well 1 Chicken Well 4 Human Well 5 Both

37

4.4 Microbial community analysis

To serve as a rough estimate of contamination in the water samples, the microbial

communities were evaluated using high-throughput sequencing. Table 10 illustrates the

significant genera associated with pathogens with high abundance among the samples. Notably,

Acinetobacter was the most abundant among all the other genera and ranged from 0.03% to 48%

of the total microbial community in the water samples. The main concern from Acinetobacter

genera is A. baumannii. Also, Pseudomonas genera ranged from 0.055% to 2.96 % of the

microbial community within water samples and the main pathogenic species is P. aeruginosa.

Also, the presence of Legionella, Bacillus, and Mycobacterium presence is a concern as these

pathogen-associated genera were the most common after the Acinetobacter and Pseudomonas.

Also, the presence of Enterococcus and Streptococcus indicates a contamination of human or

animal wastes and fecal pathogens, as they are major genera in the human and animal

microbiota.

38

Table 10: Percentage of the genera associated with pathogens in the microbial community.

Well Percentage Of the Total Microbial Community Acinetobacter Pseudomonas Streptococcus Enterococcus Legionella Bacillus Mycobacterium

Well 8 1st 3.50 2.960 0.024 0.104 0.009 0.007 0.096 Well 7 1st 2.04 2.138 0.246 0.114 0.029 0.034 0.069 Well 6 1st 48.76 0.720 0.074 0.009 0.025 0.003 0.015 Well 8 2nd 12.30 0.830 0.067 0.013 0.018 0.080 0.009 Well 7 2nd 11.13 1.559 0.232 0.031 0.006 0.000 0.019 Well 6 2nd 1.52 2.346 0.617 0.060 0.048 0.000 0.230

Well 3 2.24 2.426 0.768 0.105 0.075 0.000 1.778 Well 2 1.54 0.056 0.013 0.045 0.043 0.054 0.022 Well 1 5.81 0.512 0.022 0.128 0.014 0.025 1.483 Well 4 0.04 0.055 0.000 0.014 0.000 0.000 0.000 Well 5 0.03 0.092 0.000 0.000 0.000 0.000 0.031

39

4.5. Bacterial counts

Figure 3: Heterotrophic plate count per 50g of peeled skin.

Figure 3 illustrates the heterotrophic plate count per 50g in each of the peppers and the

tomatoes. Badalah food samples (Well 3) contained 2.45-fold more of total bacteria compared to

the Wadi Yamaniyah samples (Well 7).

Seventeen ceftazidime-resistant bacterial colonies were isolated from Wadi Yamaniyah

vegetable samples (Well 7). No meropenem-resistant bacterial isolates were obtained from Wadi

Yamaniyah vegetable samples. In contrast, there were up to 3.5 x 106 CFU of meropenem-

resistant bacterial isolates per 50 g of tomato skin sampled from Badalah site (Well 3). Similar to

the Wadi Yamaniyah samples, 23 ARB colonies from the Badalah samples were randomly

selected (Table 11). Badalah food samples contained 2-fold more of total ARB than Wadi

Yamaniyah food samples (Figure 4).

Upon phylogenetic identification based on the 16S rRNA genes, it was determined that

0.00E+00

1.00E+11

2.00E+11

3.00E+11

4.00E+11

5.00E+11

6.00E+11

7.00E+11

8.00E+11

T Mac T B.G. P Mac P B.G.

CFU

/ 50

g

Wadi Y

Badalah

40

Enterococcus spp. account as the predominant group of bacterial species isolated (Table 11). As

mentioned above, the Enterococci inhibit the guts of warm-blooded animals and humans. Also,

they exist in high abundances in feces e.g. often range between 104 and 106 cells per gram of

human feces (Slanetz and Bartley, 1957; Layton et al., 2010). Even though Enterococci occur at

lower abundance than E. coli and Bacteroidales, they are still used as fecal indicators because

they are ubiquitous among all feces and persist better in water than E. coli and Bacteroidales

(Zubrzycki and Spaulding, 1962; Anderson et al., 1997; Tendolkar et al., 2003; Byappanahalli et

al., 2012).

Listeria monocytogenes accounted as the second predominant bacterial species on the

vegeTable samples. L. monocytogenes is a foodborne pathogen associated with a number of

sporadic outbreaks throughout the world (Gray and Killinger, 1966; Ciesielski et al., 1987). The

foodborne pathogen mainly causes listeroisis infections in patients with compromised immune

system. For instance, listeroisis cases among AIDS patients occur at a rate of 150 times

compared to healthy individuals (Gellin et al., 1991). Also, the elderly, pregnant women and

infants are more susceptible to listeroisis than any other group except the immunocompromised

individuals (Tauxe, 2001). Moreover, meningitis is a common manifestation among the patients.

Fatality rate among patients can reach up to 25% and hospitalization rate of around 91% (Goulet

and Marchetti, 1996).

In addition, Pseudomonas aeruginosa, which is an opportunistic pathogen, was also recovered

from the vegetable samples.

41

Figure 4. ARB Count per 50 g of peeled skin.

Table 11: The relative abundance of ARB.

Place Antibiotic Resistance ID % in total ARB

(n=40)

Wadi Yamaniyah

Well 7

Ceft. L. monocytogenes 17.5

Ceft. E. faecalis 20

Ceft. Listonella

anguillarum 2.5

Badalah

Well 3

Ceft. E. mundtii 32.5

Meropenem P. entomophila 10

Meropenem P. monteilii 7.5

Meropenem P. aeruginosa 7.5

0 0 0 0

3.50E+06

1.50E+06

0.00E+00 5.50E+08

3.50E+10

0.00E+00

1.00E+10

2.00E+10

3.00E+10

4.00E+10

5.00E+10

6.00E+10

7.00E+10

0.00E+00

5.00E+05

1.00E+06

1.50E+06

2.00E+06

2.50E+06

3.00E+06

3.50E+06

4.00E+06

T Mac T B.G. P Mac P B.G.

CFU

/ 50

g

Ceft

.

CFU

/ 50

g

Mer

open

em

Meropenem Wadi Y Meropenem Badalah Ceft. Badalah Ceft. Wadi Y

42

4.6. Risk assessment

Given that Enterococcus faecalis, Listeria monocytogenes and Pseudomonas aeruginosa

were the top three predominant bacteria species isolated from the vegetable samples, quantitative

microbial risk assessment was performed using these three bacterial species. The daily and the

annual risk for P. aeruginosa, E. faecalis v583 and L. monocytogenes are shown in Table 12,

where in all cases, the annual risk from accidental ingestion of these bacterial species in the

Wadi Yamaniyah and Badalah vegetable samples were higher than the agreeable limit of 10-4,

which is based on the microbial risk level acceptable in drinking water in Netherlands (De Roda

Husman and Medema, 2005).

Table 12: Daily and annual risk from pathogenic ARB.

L. monocytogenes Tomato Pepper

Point estimate of risk 8.57 x10-3 5.12 x10-3 Annual risk 8.09 x10-1 6.26 x10-1

P. aeruginosa Point estimate of risk 3.41 x10-4 2.03 x10-4 Annual risk 6.33 x10-2 3.82 x10-2

E. faecilas v583 Point estimate of risk 4.31 x10-5 2.57 x10-5 Annual risk 8.25 x10-3 4.92 x10-3

43

5. Conclusion

This study carries out a recent monitoring effort on the irrigation water quality at a site

that is heavily populated by agricultural farms and livestock production (Figure 2). The findings

agree with past surveys performed on groundwater in Saudi Arabia (El-Din et al., 1993; Alaa el-

Din et al., 1994; Dabbagh and Abderrahman, 1997; Al-Salamah and Nassar, 2009; AlOtaibi,

2009; Al-Salamah et al., 2011), reiterating a high nutrient content in the irrigation water. The

main concern of high-nitrate water is usually associated with drinking water because the

denitrifiers in the human gut microbiota will reduce the nitrate to nitrite, which causes

methemoglobinemia by binding with the hemoglobin. According to Volk and his colleagues, the

concentration of nutrients is directly proportional to the growth rate of bacteria, which will not

only lead to the persistence of pathogens and fecal bacteria but also will aid them to thrive (Volk

and LeChevallier, 1999; Ward et al., 2005; Yamahara et al., 2012; Amato et al., 2014).

The high nutrient contents in water sampled from this site may be due to the fertilizers

applied by farmers, anthropogenic contamination from the nearby human settlements and/or

from wastewater streams of the poultry livestock production farm. Water sampled from some of

the wells that were within 20 km distance from the chicken production farm and near to the

Badalah area (i.e., Wells 3, 4 and 5) were positive for fecal coliform counts that exceeded the 2.2

CFU/100 ml permitted for unrestricted irrigation. Although the traditional multiple fermentation

tube method provides coliform counts that can be compared against the regulations and

standards, the method does not provide any indication on the origin of fecal contamination. This

knowledge gap can be addressed by performing qPCR-based microbial source tracking. Given

that the sampling sites were located in close proximity to human settlement and livestock

production farms, primer assays that targeted human-associated and chicken-specific indicators

were used. All the Badalah wells and all the wells within 20 km from the chicken farm were

44

positive for chicken fecal contamination. On the other hand, all the wells except the Badalah

wells showed high copy number per liter of human-associated assays, although the occurrences

of these targets were sporadic.

Antibiotics are commonly utilized at the sub-therapeutic concentration in most livestock

production farms. Over 50% of the antibiotics worldwide are fed to livestock (Novick, 1981;

Kümmerer, 2003; Zhao et al., 2010), which eventually end up in the wastewater treatment

facility and ultimately leaking to the neighboring environments. Human sewage is also likely to

contain low concentrations of antibiotics (Kümmerer et al., 2000; Karthikeyan and Meyer,

2006). However, it is likely that these treatment processes did not achieve a total removal of

antibiotic residues from the wastewater streams. This study did not attempt to measure

antibiotics residues that may be present in the water samples. However, our findings suggested

that crops irrigated with the well water from Badalah had a higher abundance of antibiotic

resistance bacteria compared to those from Wadi Yamaniyah. It is likely that water samples from

wells located in Badalah may be contaminated by wastewater from the chicken farm upstream

since it turned negative for all of the three human-associated markers, and that this

contamination may have attributed to the ARB contamination in the food samples from Badalah.

(Fox et al., 2012)

Among the isolated ARBs that were enriched from the vegetable samples, Listeria

monoctyogenes and Pseudomonas aeruginosa account as the two predominant bacterial species.

Further resolution at the strain level should be performed based on multi-locus strain typing to

determine if the isolated strains would belong to clonal lineages that are of concern to public

health. Given that these isolated bacterial species were resistant to at least one type of antibiotic,

our study attempted to further explore the associated risks from ingesting tomatoes or peppers

from the site through QMRA. The annual risks were higher than the agreeable limit of 10-4.

45

However, there are several limitations to our QMRA approach. Firstly, the enrichment of

microbial community from the food samples might have led to a slight overestimation of the

microbial risk from these food samples. Secondly, the assumed consumption rate of 4 days per

week and at 200 g per day was based on reported statistics in a US cohort. These assumptions

may not be valid for the majority of the Saudi population. Lastly, the assumed weight of 200 g of

vegetables should be based on the total weight of the consumed food. However, in order to

simplify our risk assessment approach in this study, we assumed that the entire 200 g was the

peeled skins of either tomatoes or peppers. This again might have resulted in an overestimation

of the actual microbial risk.

Future studies should aim to increase the number of samples collected over a longer

period to account for seasonal variations. Also, the bacterial isolates should be tested for their

housekeeping genes (i.e., multi-locus strain typing), as well as their resistance or susceptibility to

a wider spectrum of antibiotics. Q-PCR can also be performed to target the different types of

antibiotic resistance genes, and to determine the copy numbers of these genes in the water

samples.

46

REFERENCES

Al-Jasser, A. O. (2011). Saudi Wastewater Reuse Standards for Agricultural Irrigation: Riyadh Treatment Plants Effluent Compliance. Journal of King Saud University - Engineering Sciences 23: 1-8.

Al-Mazrou, Y. Y. (2004). Food Poisoning in Saudi Arabia. Potential for Prevention? Saudi Medical Journal 25: 11-14.

Al-Salamah, I. and I. Nassar (2009). Trends in Drinking Water Quality for Some Wells in Qassim, Saudi Arabia, 1997-2009. Journal of Applied Sciences 9: 3555-3561.

Al-Salamah, I. S., Y. M. Ghazaw and A. R. Ghumman (2011). Groundwater Modeling of Saq Aquifer Buraydah Al Qassim for Better Water Management Strategies. Environ Monit Assess 173: 851-860.

Alaa el-Din, M. N., I. M. Madany, A. al-Tayaran, A. H. al-Jubair and A. Gomaa (1994). Trends in Water Quality of Some Wells in Saudi Arabia, 1984-1989. Sci Total Environ 143: 173-181.

Allsop, K. and D. J. Stickler (1984). The Enumeration of Bacteroides Fragilis Group Organisms from Sewage and Natural Waters. J Appl Bacteriol 56: 15-24.

Allsop, K. and D. J. Stickler (1985). An Assessment of Bacteroides Fragilis Group Organisms as Indicators of Human Faecal Pollution. J Appl Bacteriol 58: 95-99.

AlOtaibi, E. L. (2009). Bacteriological Assessment of Urban Water Sources in Khamis Mushait Governorate, Southwestern Saudi Arabia. Int J Health Geogr 8: 16.

Amato, S. M., C. H. Fazen, T. C. Henry, W. W. K. Mok, M. A. Orman, E. L. Sandvik, et al. (2014). The Role of Metabolism in Bacterial Persistence. Frontiers in Microbiology 5: 70.

Anderson, S., S. Turner and G. Lewis (1997). Enterococci in the New Zealand Environment: Implications for Water Quality Monitoring. Water Science and Technology 35: 325-331.

Bernhard, A. E. and K. G. Field (2000). Identification of Nonpoint Sources of Fecal Pollution in Coastal Waters by Using Host-Specific 16s Ribosomal DNA Genetic Markers from Fecal Anaerobes. Appl Environ Microbiol 66: 1587-1594.

Bimboim, H. and J. Doly (1979). A Rapid Alkaline Extraction Procedure for Screening Recombinant Plasmid DNA. Nucleic acids research 7: 1513-1523.

Bourgogne, A., D. A. Garsin, X. Qin, K. V. Singh, J. Sillanpaa, S. Yerrapragada, et al. (2008). Large Scale Variation in Enterococcus Faecalis Illustrated by the Genome Analysis of Strain Og1rf. Genome Biol 9: R110.

Byappanahalli, M. N., M. B. Nevers, A. Korajkic, Z. R. Staley and V. J. Harwood (2012). Enterococci in the Environment. Microbiology and Molecular Biology Reviews 76: 685-706.

Ciesielski, C. A., B. Swaminathan and C. V. Broome (1987). Listeria Monocytogenes—a Foodborne Pathogen. Clinical Microbiology Newsletter 9: 149-152.

Cole, J. R., Q. Wang, E. Cardenas, J. Fish, B. Chai, R. J. Farris, et al. (2009). The Ribosomal Database Project: Improved Alignments and New Tools for Rrna Analysis. Nucleic Acids Res 37: D141-145.

Corry, J. E., G. Curtis and R. M. Baird (1996). Culture Media for Food Microbiology, Elsevier. Csuros, M. (1999). Microbiological Examination of Water and Wastewater, CRC Press. Czuprynski, C. J., N. G. Faith and H. Steinberg (2003). A/J Mice Are Susceptible and C57bl/6

Mice Are Resistant to Listeria Monocytogenes Infection by Intragastric Inoculation. Infect Immun 71: 682-689.

47

Dabbagh, A. E. and W. A. Abderrahman (1997). Management of Groundwater Resources under Various Irrigation Water Use Scenarios in Saudi Arabia. Arabian Journal for Science and Engineering 22: 47-64.

De Roda Husman, A. and G. Medema (2005). Inspectorate Guideline–Assessment of the Microbial Safety of Drinking Water. Inspectorate of the Ministry of Housing, Physical Planning and the Environment, Art 5318.

Dick, L. K., A. E. Bernhard, T. J. Brodeur, J. W. Santo Domingo, J. M. Simpson, S. P. Walters, et al. (2005). Host Distributions of Uncultivated Fecal Bacteroidales Bacteria Reveal Genetic Markers for Fecal Source Identification. Appl Environ Microbiol 71: 3184-3191.

Edgar, R. C., B. J. Haas, J. C. Clemente, C. Quince and R. Knight (2011). Uchime Improves Sensitivity and Speed of Chimera Detection. Bioinformatics 27: 2194-2200.

El-Din, M. N., I. Madany, A. Al-Tayaran, A. H. Al-Jubair and A. Gomaa (1993). Quality of Water from Some Wells in Saudi Arabia. Water, Air, and Soil Pollution 66: 135-143.

Ellis, J. B. (1999). Impacts of Urban Growth on Surface Water and Groundwater Quality: Proceedings of an International Symposium Held During Iugg 99, the Xxii General Assembly of the International Union of Geodesy and Geophysics, at Birmingham, Uk 18-30 July 1999, IAHS Press.

Environment, P. o. M. a. (2012). The National Ambient Water Quality Standard for Ksa, PME. EPA, U. S. (2010). Method 1680: Fecal Coliforms in Sewage Sludge (Biosolids) by Multiple-

Tube Fermentation Using Lauryl Tryptose Broth (Ltb) and Ec Medium. Washington, DC, U.S. EPA.

Field, K. G., E. C. Chern, L. K. Dick, J. Fuhrman, J. Griffith, P. A. Holden, et al. (2003). A Comparative Study of Culture-Independent, Library-Independent Genotypic Methods of Fecal Source Tracking. J Water Health 1: 181-194.

Food and Agriculture Organization of the United Nations. (2009). Saudi Arabia. Retrieved November 24, 2014, from http://www.fao.org/nr/water/aquastat/countries_regions/saudi_arabia/index.stm.

Fox, B. K., C. S. Tamaru, J. Hollyer, L. F. Castro, J. M. Fonseca, M. Jay-Russell, et al. (2012). A Preliminary Study of Microbial Water Quality Related to Food Safety in Recirculating Aquaponic Fish and Vegetable Production Systems. College of Tropical Agriculture and Human Resources, University of Hawaii at Manoa. Food Safety and Technology.

Gellin, B. G., C. V. Broome, W. F. Bibb, R. E. Weaver, S. Gaventa and L. Mascola (1991). The Epidemiology of Listeriosis in the United States--1986. Listeriosis Study Group. Am J Epidemiol 133: 392-401.

Golnazarian, C. A., C. W. Donnelly, S. J. Pintauro and D. B. Howard (1989). Comparison of Infectious Dose of Listeria Monocytogenes F5817 as Determined for Normal Versus Compromised C57b1/6j Mice. Journal of Food Protection 52: 696-701.

Goulet, V. and P. Marchetti (1996). Listeriosis in 225 Non-Pregnant Patients in 1992: Clinical Aspects and Outcome in Relation to Predisposing Conditions. Scand J Infect Dis 28: 367-374.

Goyal, M. R. (2012). Management of Drip/Trickle or Micro Irrigation, CRC Press. Gray, M. L. and A. H. Killinger (1966). Listeria Monocytogenes and Listeric Infections.

Bacteriol Rev 30: 309-382. Haas, C. and J. N. Eisenberg (2001). Risk Assessment. Water Quality: Guidelines, Standards and

Health: 161-183. Hagedorn, C., A. R. Blanch and V. J. Harwood (2011). Microbial Source Tracking: Methods,

Applications, and Case Studies, Springer.

48

Hazlett, L. D., D. D. Rosen and R. S. Berk (1978). Age-Related Susceptibility to Pseudomonas Aeruginosa Ocular Infections in Mice. Infect Immun 20: 25-29.

Hendricks, C. W. and S. United (1979). Evaluation of the Microbiology Standards for Drinking Water. Washington, U.S. Environmental Protection Agency, Office of Drinking Water.

Karthikeyan, K. and M. T. Meyer (2006). Occurrence of Antibiotics in Wastewater Treatment Facilities in Wisconsin, USA. Science of the Total Environment 361: 196-207.

Kümmerer, K. (2003). Significance of Antibiotics in the Environment. Journal of Antimicrobial Chemotherapy 52: 5-7.

Kümmerer, K., A. Al-Ahmad and V. Mersch-Sundermann (2000). Biodegradability of Some Antibiotics, Elimination of the Genotoxicity and Affection of Wastewater Bacteria in a Simple Test. Chemosphere 40: 701-710.

Lawin-Brussel, C. A., M. F. Refojo, F. L. Leong, L. Hanninen and K. R. Kenyon (1993). Effect of Pseudomonas Aeruginosa Concentration in Experimental Contact Lens-Related Microbial Keratitis. Cornea 12: 10-18.

Layton, B. A., S. P. Walters, L. H. Lam and A. B. Boehm (2010). Enterococcus Species Distribution among Human and Animal Hosts Using Multiplex Pcr. J Appl Microbiol 109: 539-547.

Lee, O. O., Y. Wang, J. Yang, F. F. Lafi, A. Al-Suwailem and P. Y. Qian (2011). Pyrosequencing Reveals Highly Diverse and Species-Specific Microbial Communities in Sponges from the Red Sea. Isme j 5: 650-664.

Lu, J., J. Santo Domingo and O. C. Shanks (2007). Identification of Chicken-Specific Fecal Microbial Sequences Using a Metagenomic Approach. Water Res 41: 3561-3574.

Novick, R. P. (1981). The Development and Spread of Antibiotic-Resistant Bacteria as a Consequence of Feeding Antibiotics to Livestock. Annals of the New York Academy of Sciences 368: 23-60.

O'Sullivan, E. (2010). Saudi Arabia Presses Ahead with a Wastewater Revolution. MEED. Omari, S. and D. Yeboah-Manu (2012). The Study of Bacterial Contamination of Drinking

Water Sources: A Case Study of Mpraeso, Ghana. The Internet Journal of Microbiology 10.

QMRAwiki. (2013). Dose Response Assessment. Retrieved November 24, 2014, from http://qmrawiki.msu.edu/index.php?title=Dose_response_assessment.

Schriewer, A., W. A. Miller, B. A. Byrne, M. A. Miller, S. Oates, P. A. Conrad, et al. (2010). Presence of Bacteroidales as a Predictor of Pathogens in Surface Waters of the Central California Coast. Appl Environ Microbiol 76: 5802-5814.

Silkie, S. S. and K. L. Nelson (2009). Concentrations of Host-Specific and Generic Fecal Markers Measured by Quantitative PCR in Raw Sewage and Fresh Animal Feces. Water Research 43: 4860-4871.

Simonet, P., A. Capellano, E. Navarro, R. Bardin and A. Moiroud (1984). An Improved Method for Lysis of Frankia with Achromopeptidase Allows Detection of New Plasmids. Canadian journal of microbiology 30: 1292-1295.

Slanetz, L. W. and C. H. Bartley (1957). Numbers of Enterococci in Water, Sewage, and Feces Determined by the Membrane Filter Technique with an Improved Medium. J Bacteriol 74: 591-595.

Tauxe, R. V. (2001). Food Safety and Irradiation: Protecting the Public from Foodborne Infections. Emerg Infect Dis 7: 516-521.

Tendolkar, P. M., A. S. Baghdayan and N. Shankar (2003). Pathogenic Enterococci: New Developments in the 21st Century. Cell Mol Life Sci 60: 2622-2636.

49

Toledo-Hernandez, C., H. Ryu, J. Gonzalez-Nieves, E. Huertas, G. A. Toranzos and J. W. Santo Domingo (2013). Tracking the Primary Sources of Fecal Pollution in a Tropical Watershed in a One-Year Study. Appl Environ Microbiol 79: 1689-1696.

USEPA, E. (1997). Exposure Factors Handbook. Office of Research and Development, Washington.

Volk, C. J. and M. W. LeChevallier (1999). Impacts of the Reduction of Nutrient Levels on Bacterial Water Quality in Distribution Systems. Appl Environ Microbiol 65: 4957-4966.

Ward, M. H., T. M. DeKok, P. Levallois, J. Brender, G. Gulis, B. T. Nolan, et al. (2005). Workgroup Report: Drinking-Water Nitrate and Health-Recent Findings and Research Needs. Environmental health perspectives: 1607-1614.

Weidhaas, J. L., T. W. Macbeth, R. L. Olsen, M. J. Sadowsky, D. Norat and V. J. Harwood (2010). Identification of a Brevibacterium Marker Gene Specific to Poultry Litter and Development of a Quantitative Pcr Assay. J Appl Microbiol 109: 334-347.

Yamahara, K. M., L. M. Sassoubre, K. D. Goodwin and A. B. Boehm (2012). Occurrence and Persistence of Bacterial Pathogens and Indicator Organisms in Beach Sand Along the California Coast. Applied and environmental microbiology 78: 1733-1745.

Zaharani, K., M. S. Al-SHAyAA and M. BAIg (2011). Water Conservation in the Kingdom of Saudi Arabia for Better Environment: Implications for Extension and Education. Bulgarian Journal of Agricultural Science 17: 389-395.

Zhang, Y., W. R. Kelly, S. V. Panno and W. T. Liu (2014). Tracing Fecal Pollution Sources in Karst Groundwater by Bacteroidales Genetic Biomarkers, Bacterial Indicators, and Environmental Variables. Sci Total Environ 490: 1082-1090.

Zhao, L., Y. H. Dong and H. Wang (2010). Residues of Veterinary Antibiotics in Manures from Feedlot Livestock in Eight Provinces of China. Science of the Total Environment 408: 1069-1075.

Zubrzycki, L. and E. H. Spaulding (1962). Studies on the Stability of the Normal Human Fecal Flora. J Bacteriol 83: 968-974.

50

APPENDIX AND SUPPLEMENTARY DATA

Figure S1: Microbial Richness of the Water Samples at 3250 Sequences Depth

0

200

400

600

800

1000

1200

1400

1600

Wel

l 8 1

stW

ell 7

1st

Wel

l 6 1

st

Wel

l 8 2

ndW

ell 7

2nd

Wel

l 6 2

nd

Wel

l 3W

ell 2

Wel

l 1

Wel

l 4W

ell 5

OTU

's

51

Figure S2: Cells/L of the water samples using flow cytometry.

Table S1: Salinity analysis of water samples.

Well Salinity μS/cm

Well 8 935

Well 7 1057

Well 6 2.22

Well 3 2.16

Well 2 1766

Well 1 860

Well 4 2.83

Well 5 2.81

0.00E+00

2.00E+08

4.00E+08

6.00E+08

8.00E+08

1.00E+09

1.20E+09

1.40E+09

1.60E+09

1.80E+09

Well8 1st

Well7 1st

Well6 1st

Well8 2nd

Well7 2nd

Well6 2nd

Well3

Well2

Well1

Well4

Well3

Cells

/ L

52

Figure S3: Rarefaction curves.

Table S2: Heterotrophic plate count per 50 g of peeled skin.

Site T Mac T B.G. P Mac P B.G. Wadi Y 2.00E+10 2.50E+10 3.05E+11 2.65E+11 Badalah 3.10E+11 2.45E+11 1.95E+11 7.45E+11

Table S3. ARB Count per 50 g of peeled skin.

Antibiotics Site T Mac T B.G. P Mac P B.G. Meropenem Wadi Y 0.00E+00 0.00E+00 0.00E+00 0.00E+00 Meropenem Badalah 3.50E+06 1.50E+06 0.00E+00 0.00E+00 Ceftazidime Wadi Y 0.00E+00 5.50E+08 0.00E+00 3.50E+10 Ceftazidime Badalah 0.00E+00 6.50E+10 0.00E+00 4.05E+09

0

200

400

600

800

1000

1200

1400

1600

0 1000 2000 3000 4000

OTU

s

Number of sequences

Well 5

Well 4

Well 8 1st

Well 7 1st

Well 6 1st

Well 8 2nd

Well 7 2nd

Well 6 2nd

Well 3

Well 2

53

Table S4: Total microbial richness of the water samples.

Well Number of Sequences

Number of OUT's

Well 8 1st 76230 11719.87 Well 7 1st 17478 3938.98 Well 6 1st 32637 3243.76 Well 8 2nd 22402 2558.95 Well 7 2nd 15968 2178.41 Well 6 2nd 8262 2481.63 Well 3 6636 1631.84 Well 2 125622 8475.75 Well 1 138722 9715.68 Well 4 7216 434.91 Well 5 3256 540.62