Multiplexed Detection of Foodborne Pathogens from Contaminated

Subscriber access provided by READING UNIV

Article

Multiplexed Detection of Foodborne Pathogens from Contaminated Lettuces Using a Handheld Multistep Lateral Flow Assay Device Joong Ho Shin, Jisoo Hong, Hyeyun Go, Juhwan Park, Minsuk Kong, Sangryeol Ryu, Kwangpyo Kim, Eunjung Roh, and Je-Kyun Park J. Agric. Food Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jafc.7b03582 • Publication Date (Web): 02 Dec 2017 Downloaded from http://pubs.acs.org on December 4, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Journal of Agricultural and Food Chemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 30

Journal of Agricultural and Food Chemistry

1

Multiplexed Detection of Foodborne Pathogens from

2

Contaminated Lettuces Using a Handheld Multistep

3

Lateral Flow Assay Device

4

Joong Ho Shin,†,⊥ Jisoo Hong,‡, ⊥ Hyeyun Go,† Juhwan Park,† Minsuk Kong,§ Sangryeol Ryu,§

5

Kwang-Pyo Kim,ǁ Eunjung Roh*,‡ and Je-Kyun Park*,†

6 7

†

8

Technology (KAIST), Daejeon 34141, Republic of Korea

9

‡

Department of Bio and Brain Engineering, Korea Advanced Institute of Science and

Microbial Safety Team, National Institute of Agricultural Sciences, Rural Development

10

Administration, Wanju-gun 55365, Republic of Korea

11

§

12

of Korea

13

ǁ

14

Republic of Korea

Department of Agricultural Biotechnology, Seoul National University, Seoul 08826, Republic

Department of Food Science and Technology, Chonbuk National University, Jeonju 54896,

15 16

Corresponding Authors:

17

*Fax: +82-63-238-3840. Phone: +82-63-238-3406. E-mail: [email protected] (E.R).

18

*Fax: +82-42-350-4310. Phone: +82-42-350-4315. E-mail: [email protected] (J.-K.P).

1 ACS Paragon Plus Environment


Page 2 of 30

19

ABSTRACT: This paper presents a handheld device that is capable of simplifying multistep

20

assays to perform sensitive detection of foodborne pathogens. The device is capable of

21

multiplexed detection of Escherichia coli (E. coli) O157:H7, Salmonella typhimurium (S.

22

typhimurium), Staphylococcus aureus and Bacillus cereus. The limit of detection for each

23

bacterium was characterized and then the detection of bacteria from contaminated fresh lettuces

24

was demonstrated for two representative foodborne pathogens. We employed a sample

25

pretreatment protocol to recover and concentrate target bacteria from contaminated lettuces,

26

which can detect 1.87 × 104 CFU of E. coli O157:H7 and 1.47 × 104 CFU of S. typhimurium per

27

gram of lettuce without an enrichment process. Lastly, we demonstrated that the limit of

28

detection can be reduced to 1 CFU of E. coli O157:H7 and 1 CFU of S. typhimurium per gram of

29

lettuce by including a 6-h enrichment of contaminated lettuces in growth media before

30

pretreatment.

31 32

Keywords: contaminated lettuces; foodborne pathogens; handheld device; lateral flow assay;

33

multiplexed detection

34 35


Page 3 of 30


36

INTRODUCTION

37

Foodborne illness is a serious problem that causes a number of problems such as nausea,

38

diarrhea, hospitalization and even death. In the United States, estimates of more than 9 million

39

foodborne illnesses are reported each year.1 Although it is known that the use of proper heat in

40

meat and poultry cooking prevents infection, such a heating process is skipped if fresh products

41

such as fruits and vegetables are to be consumed fresh. With the increased consumption of fresh

42

products for the prevention of chronic diseases, the number of foodborne outbreaks from the

43

consumption of fresh products has also increased.2,3 Furthermore, globalization and increasing

44

trade, along with a faster distribution of food products also increase the risk of outbreaks.4 Thus,

45

a rapid detection process needs to be developed that can be applied to the food industry in order

46

to prevent future outbreaks.

47

The current detection method for fresh products, which is considered as “gold standard,” is a

48

culture-based method.5,6 The method requires a series of processes that involves stomaching,

49

enrichment, bacteria culture, and identification of target bacteria. First, to remove bacteria

50

present on the surface of and also deep within the food sample, the sample is put in a sterile

51

plastic bag with bacteria growth medium and put into a stomacher machine. The machine

52

vigorously pounds the outer surface of the plastic bag, repeatedly and violently compressing and

53

shearing the food sample.7 Then the extracts are collected into sterile tubes and are put in a 37 °C

54

incubator over a long period of time, usually between 18 and 24 h to enrich and increase the

55

number of bacteria. This process usually ends up containing non-target bacteria and it is

56

impossible to identify the presence of target bacteria in this state. The enriched sample is then

57

streaked and incubated for another 24 h on an agar plate for colony analysis. Then, a colony has

58

to be isolated to ensure purity and an additional incubation is required. The purified colonies can



Page 4 of 30

59

be identified visually in this state, however, a combination of assays such as agglutination assay,

60

additional culture on special media, enzyme/biochemical tests and target-specific selective gram

61

staining are required to confirm the identity of the colony. This whole process is labor intensive,

62

requires proper lab equipment and takes several days (3–5 days) for the analysis result.

63

Many methods of rapid detection have been attempted as an alternative to the time consuming

64

culture-based method. Some of the methods include the use of enzyme-linked immunosorbent

65

assay (ELISA), polymerase chain reaction (PCR),8 bioluminescence signal,9 flow cytometry,10

66

and latex bead agglutination test.11 ELISA is reported to detect bacteria in the order of 104

67

colony-forming unit (CFU) per mL,12 and PCR provides highly sensitive and highly specific

68

detection results due to the amplification of the target fragment. Although both ELISA and PCR

69

can provide sensitive detection results, they are both complicated, require a trained technician, as

70

well as expensive and bulky equipment, which hinders them from being applied for on-site

71

detection. For on-site analysis purposes, a lateral flow test (LFT) is one of the most promising

72

methods due to its convenience, quick readout (10–20 min),13 and the signal that can be

73

interpreted with the naked eye. The operation starts by loading a liquid sample into the sample

74

pad. The sample flows forward and rehydrates pre-dried labeling agents and forms

75

immunocomplexes. The immunocomplexes continue to flow through the strip by capillary force,

76

and they are captured by antibodies on the test line. The flow is sustained by the absorbent pad

77

and the result appears as a visible line with the naked eye. LFT kits for various pathogens exist

78

on the market today, but there is still room for improvement in terms of the limit of detection.

79

One way to simplify and improve the detection limit of LFT is to add an additional reagent

80

loading step and perform a multistep assay on LFT. By increasing the size of the gold

81

nanoparticles (AuNPs) as signaling probes with enhancers after completion of an assay, the


Page 5 of 30


82

signal intensity can be increased and a 100-fold improvement of the detection limit was

83

demonstrated.14−16 Enzyme-linked immunoreactions were also performed on LFTs by first

84

applying a sample flow along the length of the strip and then flowing a substrate across the width

85

of the strip for enzyme-substrate-based signal generation.17,18 The processes however require

86

manual reconfiguration of multiple pads and wait for the primary reaction to finish before

87

performing a secondary reaction. Although such requirements are not suitable for on-site

88

detection purposes, the multistep assays have demonstrated sensitive analyte detection, and in

89

order to simplify the assay process, customized packaging technologies are being developed to

90

perform the sequential tasks with ease.19−21

91

With the recent advancement in paper-based fluidics and by utilizing fluid delay phenomena,

92

an engineered paper network with multiple channels of different fluidic path lengths was

93

developed.22,23 Fluids travel longer time in the longer channel, thus arrives at the test zone much

94

later as compared to fluids that flow through shorter channel. The device exploits this

95

mechanism to perform the sequential flow of samples, buffers and enhancing solution to increase

96

the sensitivity of an assay. However, such a format requires up to 1 h due to the fluidic resistance

97

that increases with increasing channel length; and the use of single wicking pad, whose

98

absorption rate saturates over time.24

99

Our group recently developed a rotary type device that offers solution to the abovementioned

100

problem. It was used to simplify the process required for ELISA on LFT and was able to detect

101

Escherichia coli (E. coli) O157:H7 in 22 min.25 In this paper, we modified the design of the

102

handheld device and expanded the number of test strips for multiplexed detection. In order to

103

reduce and simplify the assay steps, we utilized AuNP and enhancer instead of enzyme-substrate

104

pair for colorimetric signal. The device consists of four test lines that indicate the presence of E.



Page 6 of 30

105

coli O157:H7, Staphylococcus aureus (S. aureus), Salmonella typhimurium (S. typhimurium) and

106

Bacillus cereus (B. cereus), thereby allowing multiplexed detection. We used the device to detect

107

bacteria from contaminated lettuce, and then demonstrated increased detection sensitivity by

108

incorporating a bacteria enrichment step.

109 110

MATERIALS AND METHODS

111

Design and Fabrication. The proposed device is designed to simplify the complicated and

112

laborious multistep assays while simultaneously detecting up to four pathogenic bacteria with

113

high sensitivity. As shown in Figure 1A, it is equipped with two test strips: the left strip is used

114

to detect E. coli O157:H7 (ATCC 35150) and S. aureus (ATCC 25923); the right strip is used to

115

detect S. typhimurium (ATCC 14028) and B. cereus (ATCC 10987). EC, SA, ST and BC stands

116

for E. coli, S. aureus, S. typhimurium and B. cereus, respectively. The device consists of a

117

stationary part (bottom piece) and a rotating part (top piece). Sample pads (S1–S3) and their

118

corresponding absorbent pads are contained in the top piece and nitrocellulose (NC) test strip is

119

contained in the bottom piece. S21 contains a mixture of AuNPs that label E. coli O157 and S.

120

aureus antibodies, while S22 contains a mixture of AuNPs that label S. typhimurium, as well as

121

AuNPs for B. cereus.

122

A patterned OHP transparency film, which has through-holes, is attached to the bottom of the

123

top piece (Figure 1B). This enables a direct contact between the sample pad and the test strip as

124

well as between the test strip and the absorbent pad. The top and bottom pieces were made of 3

125

mm-thick poly(methyl methacrylate) (PMMA) plates, which were patterned and etched with a

126

laser cutter (C40-60W; Coryart, Anyang, Korea). The diameter of the handheld device was 8 cm.


Page 7 of 30


127

The through-holes were patterned on the OHP film with a laser cutter, and then adhered to the

128

top piece with a double-sided tape.

129

1 mg/ml anti-S. aureus antibody (ab20002; Abcam) and 1 mg/ml anti-E. coli O157 antibody

130

(ab20976; Abcam) was immobilized parallel to each other on 15 µm pore NC strip as two test

131

lines (SA, EC), respectively, and 1 mg/ml anti-mouse IgG antibody (M4155; Sigma Aldrich,

132

MO, USA) was dispensed as a control line (CL 1) for the left strip. After the antibodies were

133

dried, the NC strip was cut into 4 × 25 mm strips. 10 µL blocking buffer (2% skim milk and 0.15

134

% Tween-20 in phosphate buffered saline (PBS)) was dropped on the center of each strip and

135

dried at room temperature. This strip was used as a left strip for S. aureus and E. coli O157:H7

136

detection. A cell wall binding domain (CBD) of an endolysins produced by a B. cereus phage

137

was prepared for B. cereus detection as previously described.26 1 mg/ml anti-S. typhimurium

138

antibody (ab8274; Abcam) and 1.5 mg/ml CBD against B. cereus were immobilized on 15 µm

139

pore NC strip as two test lines (ST, BC) for the right strip, respectively. CBD was diluted with

140

PBS containing 5% BSA (Santa Cruz Biotechnology Inc., Santa Cruz, CA, USA) prior to

141

immobilization to minimize the nonspecific signal and increase the detection signal-to-noise

142

ratio.27 1 mg/ml anti-mouse IgG antibody were dispensed as the first control line (CL 1) and 1

143

mg/ml anti-glutathione-S-transferase (GST) antibody (A190-122A; Bethyl) was dispensed as the

144

second control line (CL 2).

145

Gold Nanoparticle Preparation. AuNPs for the labeling of E. coli O157:H7, S. aureus and S.

146

typhmurium are conjugated with antibodies. For the AuNP-antibody conjugation, the pH of the

147

AuNP (20 nm diameter) solution was adjusted to 9 by adding 3 µL of K2CO3 (0.2 M) per 150 µL

148

of AuNP (753610; Sigma-Aldrich). The stock concentration (1×) was 6.54 × 1011 particles/ml.

149

Then, 1 µg antibody per 150 µL was added to the AuNP and incubated at room temperature for



Page 8 of 30

150

20 min. Thereafter, to block the surfaces of antibody conjugated AuNPs, 100 µL of PBS

151

containing 3% BSA was added and incubated for 20 min. The AuNP was then centrifuged at

152

7,000 × g for 30 min to remove unbound antibodies. The supernatant was removed and the

153

AuNP was resuspended in a PBS solution containing 3% BSA and 0.05% Tween-20.

154

AuNPs for the labeling of B. cereus are conjugated with CBD proteins. For the AuNP-CBD

155

conjugation, 1 M Tris-Cl buffer (pH 9.0) was added to 1 mL of an AuNP (20 nm diameter)

156

solution suspended in citrate buffer (741965; Sigma-Adrich). 5 µL of Cys-GST-tagged CBD

157

protein solution was added to the AuNP solution and incubated for 1 h at room temperature. To

158

block the surfaces of the CBD-conjugated AuNP, 100 µL of 2 mM PEG (dissolved in 10 mM

159

Tris-Cl, pH 9) (SunBio, Anyang, Korea) was added and incubated for 1 h at room temperature.

160

The solution was then centrifuged at 8,000 × g for 30 min to remove unbound proteins. The

161

supernatant was removed and the AuNP pellet was resuspended in a PBS solution containing 3%

162

BSA and 0.05% Tween-20.

163

Device Operation. Prior to starting the detection process, 40 µL AuNP mixtures were loaded

164

into their respective AuNP pads: AuNP-E. coli O157 antibody mixed with AuNP-S. aureus

165

antibody was loaded into S21 and AuNP-S. typhimurium antibody mixed with AuNP-CBD is

166

loaded into S22. Then, 80 µL enhancer was loaded into S3. To begin the assay, 260 µL of

167

bacteria sample suspended in 2% BSA containing PBS was loaded into S1. As shown in the

168

scheme in Figure 1C, the assay first requires delivery of the sample to the NC strip for the

169

bacteria capture step. After 10 min of sample flow, the top piece was rotated counterclockwise to

170

deliver AuNPs to label the captured bacteria. After 6 min of AuNP labeling, an enhancer solution

171

(GoldEnhance EM; Nanoprobes, Yaphank, NY, USA) was delivered to amplify the AuNP signal

172

intensity. The gold enhancer solution consists of an enhancer, activator, initiator, and buffer,


Page 9 of 30


173

which are mixed prior to use. As mentioned in previous publications,28,29 the gold enhancer has

174

the ability to decrease the limit of detection and help distinguish unclear signals, which can

175

prevent false negative detection results. After 3 min of enhancement, the top layer was detached

176

from the device and pictures of the test lines were taken and analyzed. The sequential delivery

177

for the aforementioned multistep assay was easily performed simply by incrementally rotating

178

the top piece of the device with the hands.

179

Detection of Bacteria from Contaminated Lettuce. To determine whether the developed

180

device can be applied for bacteria detection from contaminated fresh vegetables samples, fresh

181

lettuces were inoculated with different amounts of E. coli O157:H7 and S. typhimurium. Bacteria

182

were grown in tryptic soy broth (TSB) (BD Difco, Sparks, MD, USA), centrifuged at 3,000 × g

183

for 20 min and resuspended in PBS, then diluted to the target inoculation number by measuring

184

the optical density (OD) of the bacteria suspension. The target inoculation amounts were 104,

185

105, 106 CFU per 10 g of lettuce. However, it is technically difficult to measure the exact target

186

CFU of the bacteria sample and dilute it to the exact target concentrations based on the OD

187

measurement alone, a portion of the diluted samples were dotted (10 µL of samples were spotted

188

on the agar plate after further log-dilution) on agar plate. The actual number of inoculated

189

bacteria was calculated by counting the number of bacteria from the agar plate the next day when

190

the colonies were visible. 100 µL of the diluted bacteria solutions were dropped on different

191

spots of the lettuces, which were then dried for 5 min, and put into a plastic bag. 90 mL of PBS

192

were poured into each plastic bag, and then the bags were put into a stomacher machine

193

(BagMixer 400; Interscience, Saint Nom, France) for 1 min to homogenize the solution and

194

retrieve the bacteria from the lettuce. The solution from each bag was poured into two 50 mL

195

tubes and centrifuged at 3,000 × g for 20 min. The supernatants were decanted and the sample



Page 10 of 30

196

were resuspended in 150 µL PBS. From the suspension, 10 µL was log-diluted and dotted on

197

agar plate to measure the retrieved number of bacteria, which was counted the next day. From

198

the remaining solution, 130 µL was taken and diluted to a final volume of 260 µL sample

199

containing 2% BSA in PBS, which was loaded into the device for detection.

200

Bacteria Enrichment. Fewer number of bacteria (in the order of 101, 102, 103 CFU per 10 g of

201

lettuce) were inoculated on the lettuce. After the contaminated lettuces were put in the bag, 90

202

mL growth media were poured into the bag instead of pouring PBS. The bag was then incubated

203

in an incubator at 37 °C for 6 h. After enrichment, the sample was put into a stomacher machine,

204

centrifuged and concentrated as described previously.

205 206 207

RESULTS AND DISCUSSION

208

Optimization of Gold Nanoparticle Concentration and Duration. The concentration of

209

AuNP-Ab used to label the captured bacteria is related to the total amount of AuNP and antibody

210

used, which is directly related to the total cost of an assay. Moreover, the duration of AuNP flow

211

is related to the total time required to perform each assay so that both the duration of AuNP flow

212

and the concentration of AuNP need to be optimized in order to reduce the total cost and time.

213

Antigen–antibody binding is a reversible interaction with the association and dissociation of

214

antigen–antibody complexes. The total number of antigen–antibody complex, which determines

215

the signal intensity, depends on both the concentration of bacteria and the concentration of

216

AuNP. Theoretically, using a higher concentration of AuNP would result in a stronger signal.

217

However, to develop an economically efficient rapid detection system, the device was optimized

218

to use the lowest AuNP concentration that allows the signal to appear at the lowest bacteria

219

concentration as possible.


Page 11 of 30


220

The test line’s intensity using several concentrations of AuNP were screened by captured S.

221

typhimurium (107 CFU/mL flowed for 10 min). The concentrations of antibody-conjugated

222

AuNP were 2×, 1×, 0.50×, and 0.25× with 1× being 6.54 × 1011 particles/mL. As shown in

223

Figure 2A, the intensity of detection signal increases over time, and the detection signal

224

generated with a higher concentration of AuNPs saturates at a higher intensity (Figure 2B).

225

According to the graph, the use of 2× AuNP and the flow duration for 14 min would result in a

226

highest signal intensity. Before reducing the AuNP concentration and the flow duration, the limit

227

of detection of S. typhimurium with 2× AuNP and 14 min flow duration was determined to be

228

105 CFU/mL with an enhancement step (Figure S1 of the Supporting Information). Thus,

229

additional experiments were performed to reduce the AuNP concentration and the flow duration

230

while maintaining the limit of detection.

231

To find an optimum condition, the AuNP flow duration was varied with a fixed AuNP

232

concentration (2×). 6, 10, and 14 min of AuNP flow durations were tested and the pictures show

233

that 6 min is sufficient to generate a detection signal of 105 CFU/mL S. typhimurium (Figure 2C).

234

Although the intensity of the detection signal is stronger after flowing for 10 min, a prolonged

235

AuNP flow is not necessary to generate a visible signal. It is interesting to note that the flowing

236

for 14 min results in a reduced signal intensity. The reason behind the decreased signal intensity

237

is unclear. It is assumed that a long duration causes dissociation of bacteria from the capture

238

antibodies, thereby resulting in a reduction in the total number of captured bacteria, which would

239

also reduce the signal intensity.

240

Lastly, the AuNP flow duration was fixed at 6 min while the AuNP concentration was varied.

241

The detection signals of 105 CFU/mL S. typhimurium show that no signal is visible below 1×

242

AuNP concentration and 1× AuNP is sufficient to generate a visible signal (Figure 2D). This data



Page 12 of 30

243

indicates that the minimum AuNP concentration required to generate a visible signal of lowest

244

bacteria concentration is 1×. Thus, the optimal labeling condition for the rest of the experiments

245

in this study was determined to be 6 min AuNP flow at 6.54 × 1011 particles/mL.

246

Double-Strip Device Detection Specificity. The specificity of the device was tested while being

247

equipped with two strips. First, a PBS sample containing only 2% BSA (negative control) was

248

tested and showed no signal from the test lines (Figure 3A). This is very important for the

249

negative control to avoid any background noise because a higher background noise results in a

250

low signal-to-noise ratio and can potentially increase the limit of detection. Then, a sample

251

containing all four species of bacteria (each species was 107 CFU/mL) was loaded into the

252

device. As shown in Figure 3B, detection signals from all test lines indicate no significant

253

problem caused by clogging or interference among the bacteria. One of the most crucial features

254

of the multiple detection sensor is that signals must appear only when the specific target exists.

255

Thus, a sample containing 107 CFU/mL of each species was loaded into the device to test its

256

specificity. Figure 3C–F shows that test lines only appear when the target bacterium exists. This

257

confirms that the device is capable of performing a specific detection of bacteria and that there is

258

no cross-reactivity or background noise.

259 260

Assessment of Limit of Detection. Individual bacterial species were serially diluted (log-

261

dilution in each step) in a concentration ranging from 103 to 107 CFU/mL. As shown in Figure

262

4A–D, E. coli O157:H7, S. aureus, S. typhimurium and B. cereus’ limits of detection were

263

determined to be 105 CFU/mL, 106 CFU/mL, 105 CFU/mL, and 105 CFU/mL, respectively. In

264

general, the detection signal increases as the bacteria concentration increases.

265


Page 13 of 30


266

Detection of Bacteria Recovered from Contaminated Lettuce. Experiments were performed

267

to validate that the device can be actually used to detect bacteria from fresh vegetables. As a

268

demonstration, lettuce was chosen and contaminated with varying concentration of E. coli

269

O157:H7 and S. typhimurium separately. Bacteria were recovered from the contaminated lettuces

270

and concentrated into a small volume so that it can be loaded into the device. The amount of

271

recovered bacteria has a significant impact on the detection result. If the majority of bacteria are

272

lost during recovery, the detection result may show false negative despite the sample

273

contaminated with a sufficient amount of pathogens for detection. Thus, the bacteria count

274

before and after the recovery process was measured. Figure 5A shows the relationship between

275

the recovered bacteria from 10 g of lettuce versus the initial inoculation amount. The trend of

276

both E. coli O157:H7 and S. typhimurium shows linear relationships between the inoculated and

277

the recovered bacteria (R2 values were 0.99 for both). The average bacteria recovery rate from

278

contaminated lettuce was 41.80 ± 3.61% for E. coli O157:H7 and 24.47 ± 4.10% for S.

279

typhimurium.

280

The recovered samples were loaded into the sample pad and the detection was performed with

281

lettuces contaminated with three different concentrations of bacteria. The amount of bacteria

282

used for contamination, recovered bacteria count, loaded sample concentration, and signal

283

appearance are shown in Table S1 of the Supporting Information. As shown in Figures 5B–E, the

284

rotary device is able to detect lettuces contaminated with E. coli O157:H7 above 1.87 × 105

285

CFU/10 g and S. typhimurium above 1.47 × 105 CFU/10 g. When the bacteria from these

286

samples are concentrated, their concentrations are 2.88 × 105 and 1.13 × 105 CFU/mL,

287

respectively. These concentrations are within the same order of magnitude of the previously

288

characterized limit of detection from spiked samples in PBS (Figure 4). The device, however,



Page 14 of 30

289

cannot detect bacteria if the lettuces are contaminated with fewer number of bacteria. It is

290

difficult to conclude whether the detection limits are meaningful for practical uses for all target

291

bacteria because the infective dose are different for each species. For example, for B. cereus, the

292

infective dose is about 105–108 cells per gram of food in order to produce sufficient toxins,30

293

which means that the current system has no problem to detect the bacteria. On the other hand, as

294

few as 10 CFU of E. coli O157:H7 can cause infection in human,31 which means any E. coli

295

O157:H7 that exists below the current limit of detection can cause harm even if the detection

296

result is negative. For bacteria whose infective dose is below the limit of detection in the rotary

297

device, the bacteria need to be enriched in order to increase their number and also to increase the

298

concentration for successful detection.

299

Detection of Bacteria after Enrichment. In order to detect bacteria that exist below the limit

300

of detection, the contaminated lettuce sample needs to be enriched with growth media and the

301

bacteria count must be increased. According to the previous data, the recovered and concentrated

302

sample needs to be at least 105 CFU/mL before loading into the device. Considering the final

303

volume of the concentrated sample that is loaded into the device (260 µL), the total number of

304

recovered bacteria from 10 g of lettuce needs to be at least 2.6 × 104 CFU, or 2.6 × 103 CFU/g.

305

This means that, regardless of the initial contaminated amount, as long as at least 2.6 × 104 CFU

306

bacteria can be recovered from 10 g of contaminated lettuce after enrichment, the device can

307

theoretically detect the contamination. For practical applications, it is important to determine the

308

minimum time required to enrich the bacteria to reach the sufficient amount required for

309

detection.

310

Experiments were performed to measure the bacteria growth over time. The lettuces were

311

contaminated with a varying number of E. coli O157:H7 and S. typhimurium that range from 101


Page 15 of 30


312

to 103 CFU per 1 gram of lettuce, and the number of recovered bacteria was measured

313

immediately after, 3 h after, and 6 h after enrichment in growth media. As expected, the graphs

314

in Figure 6A,B show that recovering E. coli O157:H7 and S. typhimurium from the lettuce

315

immediately after the contamination results in a loss, which is also below the limit of detection

316

(indicated by red dotted line). After 3 h, the lettuces contaminated with 101 and 102 CFU per

317

gram do not reach the minimum number of bacteria required for detection, while the initial

318

contamination of 103 CFU/g does reach the minimum number of bacteria for both species. 6 h

319

enrichment is enough to enrich all levels of contamination to reach the minimum threshold for

320

detection. On the basis of these results, it is concluded that as few as 101 CFU/g of contamination

321

can be theoretically detected after 6 h of enrichment.

322

To verify the enrichment process allowing for more sensitive detection of bacteria, bacteria

323

contaminations were also tested in the following ranges; 0, 1–101, 101–102, 102–103 CFU/g for

324

E. coli O157:H7 and 0, 1–101, 101–102 CFU/g for S. typhimurium (Table S2 of the Supporting

325

Information). As shown in Figure 6C,D the detection signals were visible for all tested

326

concentrations except for the negative control. It is worth noting that even the lettuces

327

contaminated with as low as 1 CFU/1 g shows a positive detection result for both E. coli

328

O157:H7 and S. typhimurium after 6 h of enrichment. This can be explained by the fact that the

329

concentrations of the concentrated samples loaded into the device for detection were all above

330

105 CFU/mL, which was above the characterized the limit of detection for both species. This

331

result indicates that 6 h of enrichment is enough for detection of as few as 1 bacterium in 1 gram

332

of lettuce, which is equivalent to about five bacteria per single lettuce leaf.

333

In this paper, a handheld device capable of simplifying multistep assays was used to perform

334

sensitive detection of foodborne pathogens. The result shows that contamination can be detected



Page 16 of 30

335

as long as the concentrated sample contains bacteria above 105 CFU/mL, which translates to the

336

initial contamination amount of 2.6 × 103 CFU per 1 g of lettuce. To further improve the limit of

337

detection, the contaminated sample were incubated in growth media and as few as 1 CFU of E.

338

coli O157:H7 or S. typhimurium per 1 g of lettuce was able to be detected after 6 h of

339

enrichment. Such sensitivity is comparable to that of more expensive and complicated molecular

340

detection methods such as PCR, which often involves a 24-h enrichment.8

341

The recovery and enrichment experiments with lettuce confirm that the device can actually be

342

applied to detect pathogens from contaminated fresh vegetable samples. One of the biggest

343

problems in the detection of food products are the debris and chunks that exist within the sample

344

solution after the pretreatment process, which are known to interfere with assays. In our study,

345

although food dye and some debris were observed between the sample pad and the NC strip (at

346

the contact zone), they had no detrimental effect on the assay results.

347

The combination of the sample enrichment process along with the use of a rotary device can

348

determine the presence of target bacteria much more quickly than previous methods. In general,

349

pathogen identification by selective growth media, PCR or LFT requires 18 to 24 h of

350

enrichment and sample pretreatment prior to testing. After enrichment, identification by selective

351

growth media along with biochemical tests requires additional 2 to 3 days; PCR needs 2 – 3 h for

352

amplification; and LFT takes about 20 min to complete the test. After enrichment and sample

353

treatment, our device takes about 19 min to perform a multiplexed detection, which is much

354

shorter than the time required for PCR amplification and biochemical tests. Furthermore, we

355

were able to demonstrate that 6 h enrichment is enough for detection of as low as 1 CFU of

356

pathogen per gram of sample. Considering the time required for sample enrichment, pretreatment

357

and detection, the entire test can be performed within 7 h using our protocol, which is a realistic


Page 17 of 30


358

time frame that can be applied for fresh product distribution. The proposed detection protocol

359

can possibly be used in food processing facilities where raw materials need to be screened prior

360

to processing and packaging. Such processes may include the packaging for ready-to-eat salads

361

and assorted fresh vegetables. Furthermore, it is expected to be used for regular screening of

362

fresh vegetables before distribution and help prevent food poisoning and foodborne pathogen

363

outbreaks before they occur, potentially saving many lives.

364 365

ASSOCIATED CONTENT

366

Supporting Information

367

The Supporting Information is available free of charge on the ACS Publications website at

368

DOI: 10.1021/acs.jafc.#######.

369

Assessment of limit of detection (Figure S1), the average number of contaminated and

370

recovered bacteria, and loaded sample concentration for the recovery experiment (Table S1), and

371

the number of contaminated and recovered bacteria after enrichment, and loaded sample

372

concentration for the enrichment experiment (Table S2) (PDF).

373 374

AUTHOR INFORMATION

375 376

Corresponding Authors

377

*Fax: +82-63-238-3840. Phone: +82-63-238-3406. E-mail: [email protected] (E.R).

378

*Fax: +82-42-350-4310. Phone: +82-42-350-4315. E-mail: [email protected] (J.-K.P).

379

ORCID

380

Je-Kyun Park: 0000-0003-4522-2574



Page 18 of 30

381

Author Contributions

382

⊥

383

Funding

384

This research was supported by the Rural Development Administration of Korea (Grant No.

Joong Ho Shin and Jisoo Hong contributed equally to this work.

385

PJ009842)

and

the

National

Research

Foundation

386

2016R1A2B3015986) funded by the Ministry of Science and ICT.

387

Notes

388

The authors declare no competing financial interest.

of

Korea

(Grant

No.

NRF-

389


Page 19 of 30

390


REFERENCES

391 392 393 394 395

(1) Painter, J. A.; Hoekstra, R. M.; Ayers, T.; Tauxe, R. V.; Braden, C. R.; Angulo, F. J.; Griffin, P. M., Emerging Infect. Dis. 2013, 19, 407–415. (2) Callejón, R. M.; Rodríguez-Naranjo, M. I.; Ubeda, C.; Hornedo-Ortega, R.; Garcia-Parrilla, M. C.; Troncoso, A. M., Foodborne Pathog. Dis. 2015, 12, 32–38.

396

(3) Olaimat, A. N.; Holley, R. A., Food Microbiol. 2012, 32, 1–19.

397

(4) Newell, D. G.; Koopmans, M.; Verhoef, L.; Duizer, E.; Aidara-Kane, A.; Sprong, H.;

398

Opsteegh, M.; Langelaar, M.; Threfall, J.; Scheutz, F., Int. J. Food Microbiol. 2010, 139,

399

S3–S15.

400 401 402 403

(5) Blackburn, C. W.; McClure, P. J., Foodborne pathogens: Hazards, risk analysis and control. 2nd Ed., CRC Press: 2009. (6) García-Cañas, V.; Simó, C.; Herrero, M.; Ibáñez, E.; Cifuentes, A., Anal. Chem. 2012, 84, 10150–10159.

404

(7) Sharpe, A.; Jackson, A., Appl. Environ. Microbiol. 1972, 24, 175–178.

405

(8) O'Grady, J.; Ruttledge, M.; Sedano-Balbás, S.; Smith, T. J.; Barry, T.; Maher, M., Food

406

Microbiol. 2009, 26, 4–7.

407

(9) Samkutty, P. J.; Gough, R. H.; Adkinson, R.; McGrew, P., J. Food Prot. 2001, 64, 208–212.

408

(10) Veal, D.; Deere, D.; Ferrari, B.; Piper, J.; Attfield, P., J. Immunol. Methods 2000, 243, 191–

409

210. 19 ACS Paragon Plus Environment


410 411 412 413 414 415 416 417 418 419 420 421 422 423

Page 20 of 30

(11) van Griethuysen, A.; Bes, M.; Etienne, J.; Zbinden, R.; Kluytmans, J., J. Clin. Microbiol. 2001, 39, 86–89. (12) Magliulo, M.; Simoni, P.; Guardigli, M.; Michelini, E.; Luciani, M.; Lelli, R.; Roda, A., J. Agric. Food Chem. 2007, 55, 4933–4939. (13) Posthuma-Trumpie, G. A.; Korf, J.; van Amerongen, A., Anal. Bioanal. Chem. 2009, 393, 569–582. (14) Anfossi, L.; Di Nardo, F.; Giovannoli, C.; Passini, C.; Baggiani, C., Anal. Bioanal. Chem. 2013, 405, 9859–9867. (15) Yang, W.; Li, X.-b.; Liu, G.-w.; Zhang, B.-b.; Zhang, Y.; Kong, T.; Tang, J.-j.; Li, D.-n.; Wang, Z., Biosens. Bioelectron. 2011, 26, 3710–3713. (16) Linares, E. M.; Kubota, L. T.; Michaelis, J.; Thalhammer, S., J. Immunol. Methods 2012, 375, 264–270. (17) Park, S.; Kim, H.; Paek, S.-H.; Hong, J. W.; Kim, Y.-K., Ultramicroscopy 2008, 108, 1348– 1351.

424

(18) Cho, J.-H.; Paek, E.-H.; Cho, I.-H.; Paek, S.-H., Anal. Chem. 2005, 77, 4091–4097.

425

(19) Cho, J.-H.; Han, S.-M.; Paek, E.-H.; Cho, I.-H.; Paek, S.-H., Anal. Chem. 2006, 78, 793–

426 427 428

800. (20) Fu, E.; Liang, T.; Houghtaling, J.; Ramachandran, S.; Ramsey, S. A.; Lutz, B.; Yager, P., Anal. Chem. 2011, 83, 7941–7946.


Page 21 of 30

429 430 431 432


(21) Kim, S.-W.; Cho, I.-H.; Lim, G.-S.; Park, G.-N.; Paek, S.-H., Biosens. Bioelectron. 2017, 98, 7–14. (22) Fu, E.; Liang, T.; Spicar-Mihalic, P.; Houghtaling, J.; Ramachandran, S.; Yager, P., Anal. Chem. 2012, 84, 4574–4579.

433

(23) Fridley, G. E.; Le, H.; Yager, P., Anal. Chem. 2014, 86, 6447–6453.

434

(24) Grant, B. D.; Smith, C. A.; Karvonen, K.; Richards-Kortum, R., Anal. Chem. 2016, 88,

435

2553–2557.

436

(25) Shin, J. H.; Park, J.-K., Anal. Chem. 2016, 88, 10374–10378.

437

(26) Kong, M.; Shin, J. H.; Heu, S.; Park, J.-K.; Ryu, S., Biosens. Bioelectron. 2017, 96, 173–

438 439 440

177. (27) Preechakasedkit, P.; Pinwattana, K.; Dungchai, W.; Siangproh, W.; Chaicumpa, W.; Tongtawe, P.; Chailapakul, O., Biosens. Bioelectron. 2012, 31, 562–566.

441

(28) Shin, J. H.; Park, J.; Kim, S. H.; Park, J.-K., Biomicrofluidics 2014, 8, 054121.

442

(29) Park, J.; Shin, J. H.; Park, J.-K., Anal. Chem. 2016, 88, 3781–3788.

443

(30) Tewari, A.; Abdullah, S., J. Food Sci. Technol. 2015, 52, 2500–2511.

444

(31) Xu, T.; Marr, E.; Lam, H.; Ripp, S.; Sayler, G.; Close, D., Ecotoxicology 2015, 24, 2133–

445

2140.

446 447



Page 22 of 30

448 449

FIGURE CAPTIONS

450

Figure 1. Working principle of a handheld multistep lateral flow assay device for the detection

451

of pathogens. (A) Photo of the double-strip rotary detection device with two test strips and a

452

schematic showing the test line layout for multiple detection of pathogens. The right test strip has

453

two control lines because AuNP conjugated with antibody and AuNP conjugated with CBD both

454

have to be captured. (B) An exploded view of the rotary device showing its components and its

455

layers. (C) Schematic showing each step of AuNP-based detection involving bacteria capture,

456

AuNP labeling, and AuNP signal enhancement steps.

457 458

Figure 2. Screening result of AuNP concentrations and flow duration. (A) Graph showing the

459

intensity of detection signal of 107 CFU/mL S. typhimurium over time for different

460

concentrations of AuNP used (n = 3). (B) Pictures showing the saturated intensity of the

461

detection signal after 16 min of AuNP flow. (C) Pictures and quantified intensities of 105

462

CFU/mL S. typhimurium with varying duration of 2× AuNP flow (n = 3). (D) The enhanced

463

detection signal generated with varying concentrations of AuNP after 6 min flow. Red arrows

464

indicate the visible signals. The data points and error bars represent the mean and standard

465

deviation of the mean.

466 467

Figure 3. Photos showing the detection specificity of multiple detection of all four species of

468

bacteria. (A) Negative control result shows that there is no background signal. (B) All signals are

469

present with a sample containing all four samples (the concentration of each species in the

470

sample is 107 CFU/mL). (C–F) Sample containing 107 CFU/mL of each species results in


Page 23 of 30


471

appearance of their respective test line, indicating detection specificity. Red arrows indicate

472

positive test lines.

473 474

Figure 4. Test signal intensity versus bacteria concentration for each species. Photos of detection

475

results and graphs showing the quantified signal intensity of (A) E. coli O157:H7 (n = 3), (B) S.

476

aureus (n = 3), (C) S. typhimurium (n = 3) and (D) B. cereus (n = 3). The limits of detection of E.

477

coli O157:H7, S. aureus, S. typhimurium and B. cereus were determined to be 105 CFU/mL, 106

478

CFU/mL, 105 CFU/mL, and 105 CFU/mL, respectively. The data points and error bars represent

479

the mean and standard deviation of the mean.

480 481

Figure 5. A bacteria recovery rate and corresponding signal intensity from contaminated lettuce.

482

(A) Graph showing the recovered bacteria counts versus the number of inoculated bacteria per 10

483

g of lettuce (n = 3). (B) Photos and (C) quantified signal intensity of recovered E. coli O157:H7

484

from contaminated lettuce (n = 3). (D) Photos and (E) quantified signal intensity of recovered S.

485

typhimurium from contaminated lettuce (n = 3). The data points and error bars represent the

486

mean and standard deviation of the mean.

487 488

Figure 6. Graph showing the number of bacteria after varying enrichment time for (A) E. coli

489

O157:H7 (n = 3) and (B) S. typhimurium (n = 3) with varying bacteria count of initial

490

contamination. Red dotted line indicates the minimum amount of recovered bacteria required for

491

detection (2.6 × 103 CFU/g). Pictures showing the detection results of (C) E. coli O157:H7 and

492

(D) S. typhimurium after enrichment with different amounts of initial bacteria inoculation. The

493

data points and error bars represent the mean and standard deviation of the mean.



Page 24 of 30

494

Table of Contents artwork

495 496

(1st: Food Safety and Toxicology; 2nd: Analytical Methods; 3rd: Agricultural and Environmental Chemistry)

497

498 499


Page 25 of 30

500 501 502


Figure 1

503 504 505



506 507 508

Page 26 of 30

Figure 2

509 510 511


Page 27 of 30

512 513 514


Figure 3

515 516 517



518 519 520

Page 28 of 30

Figure 4

521 522 523


Page 29 of 30

524 525 526


Figure 5

527 528 529



530 531 532

Page 30 of 30

Figure 6

533


Multiplexed Detection of Foodborne Pathogens from Contaminated

Recommend Documents