mobile water kit (mwk): a smartphone compatible low-cost water monitoring system for rapid detection...

Analytical Methodswww.rsc.org/methods

ISSN 1759-9660

PAPERSushanta Mitra et al.Mobile Water Kit (MWK): a smartphone compatible low-cost water monitoring system for rapid detection of total coliform and E. coli

Volume 6 Number 16 21 August 2014 Pages 6139–6590

AnalyticalMethods

PAPER

Publ

ishe

d on

13

June

201

4. D

ownl

oade

d on

23/

08/2

014

07:4

0:24

.

View Article OnlineView Journal | View Issue

Mobile Water Kit

aDepartment of Mechanical Engineering, Uni

T6G 2G8. E-mail: sushanta.mitra@ualberta.

5017bTCS Innovation Labs, Tata Consultancy Se

IndiacDepartment of Microbiology, Bhavans Rese

058, India

† Electronic supplementary informa10.1039/c4ay01245c

Cite this: Anal. Methods, 2014, 6, 6236

Received 24th May 2014Accepted 12th June 2014

DOI: 10.1039/c4ay01245c

www.rsc.org/methods

6236 | Anal. Methods, 2014, 6, 6236–6

(MWK): a smartphone compatiblelow-cost water monitoring system for rapiddetection of total coliform and E. coli†

Naga Siva Kumar Gunda,a Selvaraj Naicker,a Sujit Shinde,b Sanjay Kimbahune,b

Sandhya Shrivastavac and Sushanta Mitra*a

In this work, we have developed and demonstrated a rapid and low-cost water monitoring sensor that can

simultaneously detect total coliform and Escherichia coli (E. coli) bacteria in contaminated drinking water

samples. The test method, called Mobile Water Kit (MWK), comprises a set of custom chemical reagents

that would serve as colorimetric or fluorometric chemosensors, syringe filter units and a smartphone

platform that would serve as the detection/analysis system. The MWK provides information about the

presence/absence of total coliform and E. coli in water samples. The MWK has preliminarily been tested

for its selectivity, sensitivity and accuracy, with samples of known concentrations of bacteria. The MWK

has also been tested with contaminated water samples collected during the two field trials conducted in

Canada and India, and the obtained results were confirmed with conventional laboratory methods. With

this MWK, we were able to detect the total coliform and E. coli bacteria in water samples within 30 min

or less, depending on the concentration of the bacteria. For one of the field samples, the MWK was able

to detect the total coliform within 35 s, which is faster than any rapid test methods available in the

market. This new technology can dramatically improve the response times for the outbreak of water-

borne diseases and will help water managers and individuals to assess the quality of water sources.

1 Introduction

There is a pressing need to access clean and safe potable waterfor the developing world and in communities with minimalresource settings. According to World Health Organization(WHO) statistics, approximately 884 million people (whichmeans 1 in 8 people worldwide) do not have access to clean andsafe potable water.1 In most of these cases, contaminated wateris consumed without any prior treatment processes. Water-borne diseases (especially diarrhoeal diseases) account forroughly 4.1% of total global burden of diseases and approxi-mately 3.4 million deaths are reported annually owing to water,sanitation, and hygiene-related causes.2–8 Almost 90% of thosedeaths, (mostly children under the age of 5) occur in thedeveloping world and in communities with minimalresources.8–10 One of the major causes for such public healthconcern is the lack of consistent and economical methods to

versity of Alberta, Edmonton, AB, Canada,

ca; Fax: +1 780 492 2200; Tel: +1 780 492

rvices, Mumbai, Maharashtra, 400 601,

arch Center, Mumbai, Maharashtra, 400

tion (ESI) available. See DOI:

246

provide safe and uncontaminated drinking water to commu-nities. Among all available methods of water treatment, aneffective chlorination process at the point of use is cost effective,which can eliminate certain pathogens in drinking water.However, chlorination results in an unpleasant taste indrinking water and is not always necessary.11–13 Moreover, thereare certain health risks associated with excessive chlorina-tion.11–13 Hence, it is important to assess the requirement forchlorination.11–13 A natural corollary to the above challenges isto develop water monitoring systems capable of rapid andaccurate determination of the pathogens or the presence orabsence of a suitable indicator (such as E. coli) in water.Therefore, there is a need for a simple, rapid, and low-cost watermonitoring system for evaluating the bacteriological quality ofdrinking water at the point of use. It will also provide answer toa critical question whether chlorination is required or not? Sucha water monitoring system can also act as an early warning forthe community by providing information on the quality andsafety of the water source. The main challenge for such a watermonitoring system (or sensor) would be to detect low concen-trations of target contaminants in a prescribed volume of watersample (�100 ml) within a reasonable time (�1 h). According toHealth Canada and United States Environmental ProtectionAgency (USEPA) standards, the target concentration value for E.coli is zero Colony-Forming Units (CFU) per 100 ml for potablewater and 126 CFU per 100 ml for recreational water.7,14–17

This journal is © The Royal Society of Chemistry 2014

http://crossmark.crossref.org/dialog/?doi=10.1039/c4ay01245c&domain=pdf&date_stamp=2014-07-23

http://dx.doi.org/10.1039/c4ay01245c

http://pubs.rsc.org/en/journals/journal/AY

http://pubs.rsc.org/en/journals/journal/AY?issueid=AY006016

Paper Analytical Methods

Publ

ishe

d on

13

June

201

4. D

ownl

oade

d on

23/

08/2

014

07:4

0:24

. View Article Online

Coliforms are a group of bacteria present in the environmentand in the intestines of all warm-blooded animals andhumans.18–20 According to the source and characteristics ofbacteria, coliform bacteria are grouped as either total coliformor faecal coliform. Total coliform comprises faecal coliformbacteria such as E. coli and other coliform bacteria that arepresent in the environment. Faecal coliform bacteria are foundin bodily waste, animal droppings, and in soil.18–20 Oen totalcoliform and E. coli counts are a good indicator of watercontamination, as their presence indicate the likelihood ofother deadly bacteria (viz. Salmonella spp., Vibrio cholerae),viruses (like Hepatitis and Rotavirus), etc. in drinking water.21,22

Usually, the conventional methods take around 24 to 48hours to detect coliform bacteria in contaminated watersamples.23–35 These conventional methods require transportingof water samples to a designated microbiology laboratory,which can be a challenge for remote locations. These methodsare also time-consuming and require trained professionals totest water samples in the laboratory. Recently, numerousadvanced methods based on interfacing micro- and nanotech-nology with molecular (immunological or genetic detection)and enzymatic methods have been developed for coliformdetection.36–54 However, these advanced techniques havesignicant challenges in terms of eld deployment, eventhough they are rapid and highly sensitive.36–54 The challengesand potential of these novel methods are reviewed in detailelsewhere.27–35 Most of the detection principles are based oninstant culturing of bacteria, enzymatic reactions, and molec-ular (immunological or genetic) methods of detection. Pres-ently, different alternative assays are commercially available todetect total coliform and E. coli. USEPA has approved ten suchmethods that include Colilert® (WP200, IDEXX, Westbrook,ME, USA), Colilert-18® (WP200-18, IDEXX, Westbrook, ME,USA), Coliscan® (250 MF, Micrology Laboratories, LLC Goshen,IN, USA), Colisure® (WCLS200, IDEXX, Westbrook, ME, USA),MI Agar (B14985, S&S Biosciences, Fisher), Readycult® Coli-forms 100 (EMD Chemicals Inc., Gibbstown, NJ, USA),Chromocult® (EMD/Merck Laboratories, Gibbstown, NJ, USA),E-Colite® (ECO-100, Charm Sciences Inc., Lawrence, MA, USA),m-Coli Blue 24® (MOOPMCB2, Hach/Millipore Billerica, MA,USA), and Colitag® (4600-0012, CPI International, Santa Rosa,CA, USA).7,55,56 Like conventional methods, most of these teststoo require certain equipment and dedicated laboratories to testthe water samples for accurate and reliable results, and may becost-intensive as well. Though, these laboratory tests are highlysensitive, they are time consuming and are not applicable in theeld for point of use testing of water quality. The main reasonfor the prolonged durations to perform these tests, is the lowconcentration of target bacteria in drinking water samples.Therefore, there is a necessity for a simple, easy to use, rapidand low-cost test method for on-eld detection of bacteria from100 ml of drinking water samples. These methods can be usedfrequently to test the water quality with minimum expenses andalso provide early warning signals about water quality. More-over, the public in general, will be more interested in qualitativeresults than quantitative results. Hence, the “positive/negative”type of detection tests at the point of use are more desirable for


limited resource communities compared to expensive analyticaltests with read-outs for actual CFU per 100 ml of water.

In the present work, we have developed a rapid and low-costeld test that can simultaneously detect total coliform and E.coli in contaminated water samples. The test method comprisesa low-cost water monitoring system, called Mobile Water Kit(MWK), coupled with a smartphone platform, for quantitativeanalysis. The MWK is simple, rapid, easy to use and is based onrapid enzymatic activity of specic enzymes produced by totalcoliform and E. coli bacteria, with specically formulatedchemical reagents. Dual enzyme substrates, the chromogen 6-chloro-3-indolyl-b-D-galactopyranoside (Red-Gal) and the uor-ogen 4-methylumbelliferyl-b-D-glucuronide (MUG), arecombined with other key ingredients to induce the productionof the enzymes galactosidase and glucuronidase, respectively.These enzymes in turn cleave the specic substrates (Red-Galand MUG) to release colored or uorescent molecules, resultingin a change in the color of the solution. Several review articlesare available on methods for detecting total coliform and E. coliby enzymatic methods.7,55,57,58 Recently, Gunda et al.59 exploredthis combination for simultaneous detection of total coliformand E. coli of contaminated water samples. The MWK has beendeveloped based on this discovery of rapid detection of patho-gens in contaminated water. This system consists of a 100 ml/60ml leur lock syringe, a 0.45 mm pore size syringe lter unit,specically formulated chemical reagents, and a smartphonewithmHealth E. coli application (App). The test method involvescollecting 100 ml of contaminated water sample in a syringeand ltering it through a 0.45 mm pore size syringe lter unit.Aer the ltration, the formulated chemical reagents are added,in a sequential manner, onto the syringe lter unit. Theappearance of the red color on the syringe lter surface indi-cates the presence of total coliform and E. coli. The change incolor can be captured using a smartphone and the images canbe analyzed using the custom built mHealth E. coli App toprovide quantitative results. The mHealth App facilitatescommunication with the end users via Short Message Service(SMS). The mHealth App can also access the in-built globalpositioning system (GPS) of the smartphone and thereby canprovide unique location identier for the contaminated watersources. This water monitoring platform is not only capable ofproviding real-time data for contaminated water source loca-tions, but can also provide an “early warning” system to detectany outbreak of water-borne diseases in the community. UsingmHealth's web based console, each eld test can be monitoredalong with its location being mapped.

In remote communities, be it in developing countries likeIndia or in First Nation communities in Canada's North, as anindividual user of water from different sources, one will beinterested to know if there is E. coli present or absent in thepotable water, rather than having quantied data in terms ofCFU per 100 ml. Such quantied data are of interest to regula-tors, municipality water board, etc., but have little signicanceto millions of people living in impoverishment. The MWKprovides an answer to this bottom billion people, who now canbe empowered to make their choice of drinking water by thissimple “yes/no” test. The present method focuses on the

Anal. Methods, 2014, 6, 6236–6246 | 6237


Analytical Methods Paper

Publ

ishe

d on

13

June

201

4. D

ownl

oade

d on

23/

08/2

014

07:4

0:24


presence or absence of the total coliform/E. coli in potable watersamples; hence visual observation of tested lter surfaces isgood enough for general public to understand whether thewater is contaminated or not. In future, our plan is to integratethe LED light source of a given excitation wavelength and areceiver to provide quantication capability for the MWK.

2 Materials and methods2.1 Materials

An enzymatic chromogen substrate, Red-Gal (6-chloro-3-indolyl-b-D-galactoside), was obtained from Research Organics,Cleveland, OH, USA and a uorogen substrate, 4-methyl-lumbelliferyl-b-D-glucuronide (MUG), was purchased from Bio-world, Dublin, OH, USA. Bacteria protein extraction reagent (B-PER), Lauryl Tryptose Broth (LTB) nutrient broth medium andthe LTB medium with MUG were obtained from Fisher Scien-tic, Canada. N,N-Dimethylformamide (DMF) and anhydrousFerric Chloride (FeCl3) were obtained from Sigma-Aldrich, USA.

E. coli Castellani and Chalmers (American Type CultureCollection (ATCC) 11229), E. coli O157 (National Collection ofType Cultures (NCTC) 12900 and NCTC 13125) were obtainedfrom Stream Technologies, Inc., (Edmonton, Canada). Thecultured E. coli Castellani and Chalmers (ATCC 8739) andEnterobacter spp. were obtained from Bhavan's Research Center(BRC), Mumbai, India. E. coli DH5a (ATCC 67878) strain wasacquired from Biological and Medicinal Chemistry Laboratory(BMCL), University of Alberta, Edmonton, Canada. ListeriaMonocytogenes (ATCC 43251) and Pseudomonas uorescens werepurchased from Cedarlane, Burlington, ON, Canada. All thesebacteria samples were cultured in nutrient broth mediumovernight at 37 �C in a microbiological incubator (Model IMC18120V, Heratherm, Thermo Scientic, Canada). Unless otherwisestated, serial dilution was made in LTB to produce bacteriaconcentrations in the range of 2–2 � 108 CFU ml�1. Here, inorder to generate contaminated water samples for laboratorytesting purpose, we have further diluted above preparedconcentrated samples (1 ml) with DI Water (99 ml) to generate100 ml of laboratory test water sample with nal bacteriaconcentrations in the range of 2–2 � 108 CFU per 100 ml. Theseknown bacteria concentration water samples were used to checkthe performance of the MWK in the laboratory. The LTB(without MUG) is used to culture the bacteria and the LTB withMUG is used for formulating chemical reagents.

Sterile Millex®Syringe lter units with mixed celluloseesters, 33 mm diameter, 0.45 mm pore size, 150 mm thick, wereobtained from EMD Millipore Corporation, Billerica, MA, USA.100 ml/60 ml capacity plastic syringes (sterile, leur lock,disposable) were obtained from BD (Becton, Dickinson andCompany), Canada. Microcentrifuge tubes of 1.5 ml, sterile,snap-t, plastic were purchased from Fisherbrand premium,Fisher Scientic, Canada. Pipette (10 ml to 100 ml) and therespective pipette tips were obtained from Eppendorf Canada,Mississauga, ON, Canada. Deionized (DI) water was used toprepare most of the solutions. Materials were sterilized when-ever needed in an autoclave (Tuttnauer 3850M Autoclave, Hei-dolph North America, Elk Grove Village, IL, USA).

6238 | Anal. Methods, 2014, 6, 6236–6246

2.2 Formulation of chemical reagents

Four chemical reagents (A, B, C and D) were specicallyformulated using the procured chemicals and stored in sepa-rate sterile microcentrifuge tubes. Reagent A was prepared bydissolving a mixture of LTB with MUG (35.7 mg) and Red-Gal(0.30 mg) in deionized (DI) water (1 ml). Reagent B was B-PER (1ml). Reagent C was prepared by dissolving FeCl3 (20 mg) in DIwater (1 ml). Reagent D was prepared by dissolving Red-Gal (30mg) in a mixture of DMF (0.5 ml) and DI water (0.5 ml). For asingle test, we require approximately 100 ml of reagent A andapproximately 20 ml of each other reagents B, C and D.59,60 All thechemical reagents were maintained at pH 7. The combinationof Red-Gal and MUG is used to detect E. coli pathogens thatsecrete b-galactosidase and b-glucuronidase enzymes.

2.3 Components of the MWK

The MWK can be used for simultaneous detection of totalcoliform and E. coli. It allows users to simultaneously test threecontaminated water samples. The MWK consists of a box withthree components: the main unit allows placement of threesyringe lter units, four specically formulated chemicalreagents (A, B, C and D), 12 pipette tips/droppers to dispensethe chemical reagents on the lter unit. The bottom reservoir isused to store the ltered water, which can be periodicallydrained as required. The top lid is used for covering the box toavoid exposure to dust and other environmental contamina-tions. In addition, the kit contains three sterile 100 ml/60 mlleur lock syringes, a sterile container to collect water samplesand a smartphone operating on android (which is equippedwith a custom built mobile application called mHealth E. coliApp). Fig. 1 shows the main components of a eld deployableMWK.

2.4 Water monitoring using the MWK

The contaminated water sample was collected in a sterilecontainer, followed by transferring the sample to a 100ml/60 mlsterile syringe (as per USEPA standard). The syringe wasattached to one of the three syringe lter units. Any air bubblesin the syringe should be removed before attaching it to thesyringe lter unit. Aer attachment, the syringe plunger rod ispushed to lter the water sample until no water residue ispresent in the syringe (complete ltration of water sample). Anybacterium present in the water sample is retained or concen-trated onto the surface of the syringe lter unit. The syringe wasremoved from the syringe lter unit aer ltration. A fewmicroliters of specically formulated chemical reagents A, B, Cand D were then added into the syringe lter unit in asequential manner. The syringe lter unit was incubated for acertain period at 37 �C. The lters were then monitored for overa period of one hour, followed by every 15 min for a period of 2 hto observe any changes in the color. The presence of total coli-form bacteria in the contaminated water sample was detectedby the appearance of the red color on the surface of the syringelter unit. The observation of color change is for rapid



Fig. 1 Mobile water kit (MWK) used in the field trials for simultaneous detection of total coliform and E. coli.


Publ

ishe

d on

13

June

201

4. D

ownl

oade

d on

23/

08/2

014

07:4

0:24


qualitative results only. The entire procedure to use the MWKfor water monitoring is schematically illustrated in Fig. 2.

The mHealth E. coli App operated through a smartphone canbe used to analyze the image of the tested syringe lter unit. Thecolor change of the lter surface is then processed usingmHealth and then over the mobile phone network. ThemHealth App analyzes the captured image by scanning forpixels of a specic color (in this case red) and also simulta-neously cancels any background color due to the presence ofany other particles on the lter surface. The step by stepprocedure for the use of mHealth E. coli App to capture the

Fig. 2 Step by step procedure to use the MWK for water monitoring: (Ssyringe; (Step 3) fix syringe to the syringe filter; (Step 4) push syringe plungadd to syringe filter unit; (Step 7) repeat steps 5 and 6 to add other chem(Step 9) put support for controlling smartphone location; (Step 10) take pic(Step 11) transmit the picture to a server and initiate other App operation


image of the tested syringe lter unit, thereby providing an earlywarning system through short messaging service (SMS), isillustrated in Fig. 3.

In summary, the test method with the MWK involves: (a)ltering of the contaminated water through a syringe lter unitto concentrate the contaminants on the surface of the lter, (b)adding the specically formulated chemical reagents on thelter surface to interact with contaminants, (c) allowing thechemical reagents to interact with contaminants for a certainperiod, (d) detecting the presence of total coliform and E. coli bythe appearance of red color on the lter surface, (e) capturing

tep 1) collect water in a syringe; (Step 2) remove any air bubble in theer to filter water; (Step 5) take chemical reagent A using pipette; (Step 6)ical reagents B, C and D; (Step 8) observe the appearance of red color;ture of the tested filter surfacewith a smartphone camera after 1 h; ands (details provided in Fig. 3) using the cellular network.

Anal. Methods, 2014, 6, 6236–6246 | 6239


Fig. 3 Step by step procedure to capture the image of the tested syringe filter unit usingmHealth E. coli App. (a) Select themHealth E. coli App inthe smartphone; (b) select language English; (c) accept disclaimer statement; (d) click take E. coli photo; (e) select the water source type (well,tap, river, etc.); (f) take the picture of the tested filter; (g) click the submit button; (h) wait to upload the picture and receive the result; (i) view theresult (if result is Yes: E. coli is present or if result is No: E. coli is absent) and (j) received text message on subscribed users' mobile regarding thequality of water.


Publ

ishe

d on

13

June

201

4. D

ownl

oade

d on

23/

08/2

014

07:4

0:24


the image of the tested lter surface (refer Fig. 4) for furtherimage processing and conrming the presence of total coliformand E. coli through custom made mHealth E. coli App in asmartphone, (f) determining the risk level for drinking watersources according to total coliform and E. coli levels, (g) creatingan interactive, searchable map of water quality results throughin-built GPS, and (h) can transmit water quality alerts to thecommunity via SMS.

2.5 Conrmation of the MWK with standard methods

Classical microbiological tests, including estimation of numberof the coliform groups by the multiple tube dilution tests

Fig. 4 Comparison of the filter surface between an unused (clean) anda used syringe filter unit. It is to be noted that the appearance of coloron the used filter indicates the presence of total coliform and E. coli.

6240 | Anal. Methods, 2014, 6, 6236–6246

(Presumptive test, conrmed test, or completed test) (IS1622:1981) were performed on eld water samples to cross-check theresults, which were obtained by using the MWK. The watersamples for these conrmatory tests were collected directlyfrom different sources, for which MWKs were deployed. Foreach test, three sterile polystyrene bottles (capacity 100 ml) wereused to collect three samples for the same water source. Well-established conrmatory laboratory methods (as per IS1622:1981) such as MacConkey's Broth,61,62 brilliant green lactose bile(BGLB) broth63,64 and Indole test65,66 were used for identifyingand enumerating total coliform, faecal coliform and E. coli,respectively. These standard water quality tests were conductedat the BRC, Mumbai, India. A detailed description of theconrmatory tests is provided in the ESI.†

3 Results and discussions

In the present study, we focused our efforts only to observe thered color due to enzymatic reaction of the b-galactosidaseenzyme with Red-Gal and in future studies, we will incorporatea system to observe the blue uorescence generated by theenzymatic reaction of b-glucuronidase with MUG. The inclusionof B-PER in the MWK is to accelerate the extraction of b-galac-tosidase and b-glucuronidase enzymes by lysing the bacteriacells without denaturing the bacteria. FeCl3 solution is used toaccelerate the dimerization of the released indoxyls to indigo,the color producing molecule on the syringe lter surface,which is visible by naked eye. The combination of these



Fig. 6 Variation in the response time of the MWK with the appearanceof red color on the syringe filter with different known concentrationsof E. coli in laboratory water samples.


Publ

ishe

d on

13

June

201

4. D

ownl

oade

d on

23/

08/2

014

07:4

0:24


chemical reagents enhances the rapid detection process, whichwas not explored previously in any commercially available testkits.

The MWKmethod was rst tested with known concentrationof bacteria before deploying in the eld. E. coli Castellani andChalmers (ATCC 11229) were used as standards for the valida-tion of the kit. E. coli Castellani and Chalmers act as indicatorsfor both total coliform and E. coli since it has both b-galacto-sidase and b-glucuronidase enzymes.

3.1 Sensitivity

The MWK has been repeatedly used for 10 to 12 times withdifferent concentrations of bacteria to determine the sensitivityand limit of detection. Fig. 5 shows the appearance of red coloron the syringe lter unit for different known concentrations ofE. coli (ATCC 11229) mixed with DI water. It was observed thatthe intensity of color increases with the increase in theconcentration of E. coli. Fig. 6 depicts the variation in theresponse time of the system for different concentrations of E.coli samples. It was found that the appearance of the red colorfor samples (2 � 108 CFU per 100 ml and 2 � 107 CFU per 100ml) occurs within 30 s to 60 s. The present method was able todetect up to 2 CFU per 100 ml (detection time between 30 minand 65min) and this can be considered as the limit of detection.Water samples with more than 200 CFU per 100 ml of E. coli canbe easily detected within 15 min to 20 min, which is very rapidcompared to other existing methods.7,55,56

3.2 Specicity

The reported method was veried for its specicity by usingdifferent positive and negative controls. E. coli Castellani andChalmers (ATCC 11229 and ATCC 8739), E. coli O157 (NCTC12900 and NCTC 13125), Enterobacter and E. coli DH5a (ATCC67878) were used as positive controls for both total coliform andE. coli. Listeria Monocytogenes (ATCC 43251), Pseudomonas

Fig. 5 Appearance of red color on the syringe filter unit for different knocolor intensity depends on the incubation time and the concentration o


uorescens and deionized water without any E. coli were used asnegative controls. Fig. 7 shows the tested syringe lter unit withpositive and negative controls. Fig. 7(a) shows the appearance ofred color on the surface of the syringe lter unit due to differentpositive controls. As expected, negative controls have notproduced any color on the syringe lter unit (see Fig. 7(b))indicating the specicity of the present method. These lterswith controlled samples were kept for 2 h.

3.3 Field trials in Canada and India

The rst set of eld trials was conducted by testing 21 differentwater samples collected from diverse sources such asrivers, wells, lakes, and ponds in two disparate geographical

wn concentrations of E. coli in laboratory water samples. Note that thef E. coli.

Anal. Methods, 2014, 6, 6236–6246 | 6241


Fig. 7 (a) Positive control tests using different strains (as labeled) of coliform bacteria including E. coli. (b) Negative control tests with DI water (noE. coli cells) and other pathogens (as labeled) which do not belong to the coliform bacteria strain.


Publ

ishe

d on

13

June

201

4. D

ownl

oade

d on

23/

08/2

014

07:4

0:24


regions – India and Canada. Fig. 8 depicts the test results for thecollected water samples. Sample # 1 is a negative control thathas DI water without any bacteria. Samples # 2 to 7 are fromnearby ponds and river creeks in undisclosed locations inEdmonton, Canada. Samples # 8 to 21 are from different wells,river, and hand pump from undisclosed locations in Mumbai,India. Most of the water samples (Sample # 2, 3, 6, 9, 11, 13, 15,16, 17, 19 and 21) conrmed the presence of total coliform andE. coli within 30 s to 5 min with the appearance of pinkish red,salmon or brownish red on the syringe lter units. A fewsamples produced less intensity of the red color (Sample # 4, 5,7, 12 and 20) within 8 min to 15 min aer completing thesequential dispensing of reagents A, B, C and D. The yellowcolor observed in these lter unit images will be automaticallycanceled by the mHealth App and only the red pixels will betaken into consideration during image analysis. Samples # 1, 8,10, 14 and 18 do not show any color development even aer one

Fig. 8 Appearance of red color on the syringe filter unit for 21 differenSample # 1 acts as a negative control. Note that the color intensity depe

6242 | Anal. Methods, 2014, 6, 6236–6246

hour. As predicted, the negative control sample # 1 does notproduce any red color. Further, preliminary conrmatory testswere conducted by selecting two random water sources inMumbai, India, and sending the water samples from thosesources for conventional laboratory tests at BRC, Mumbai. Thedetailed results (see Table S1†) for such conrmatory tests aredocumented in the ESI.† The conrmatory tests suggest that therst eld trial was overall successful in detecting total coliformand E. coli in contaminated water sources.

During the second eld trial, to demonstrate the reproduc-ibility, we used three sterile syringes of 100 ml capacity andlled them with water from the same source and tested themindividually in the MWK. As shown in Fig. 9, these tests arereproducible across the three lter surfaces, where similar colorchange (visible to naked eyes) is observed for the same sourcewater. The sample # in Fig. 9 indicates different water sources,located at different geographical locations within the

t water samples collected in first field trial (in Canada and India). Herends on the concentration of E. coli.




Publ

ishe

d on

13

June

201

4. D

ownl

oade

d on

23/

08/2

014

07:4

0:24


community, that were tested with the kit and each time we haveobtained reproducible results without performing any adjust-ments to the MWK. All samples, except Sample # A4 (thissample was from a deep tube well where chances of any faecalcontamination was minimal), indicate the presence of totalcoliform in the collected water samples. The obtained presence/absence test results were conrmed by testing the same watersamples with conventional laboratory methods at BRC, Mum-bai, the details (see Table S2†) of which are provided in the ESI.†Such conrmatory tests again demonstrated that the secondeld trial was equally successful in properly detecting the water-borne pathogens. So far the MWK has been cross-checked withrespect to model uids containing known concentration of E.coli and a few actual eld samples collected in Canada andIndia. Furthermore, detailed validation exercise is required withdifferent water source samples correspondingly tested withdifferent known laboratory methods to make the MWK acommercially viable product.

4 Future outlook: towards acontinuous mobile water monitoringsystem

Typical potable water sources for rural communities, particu-larly in countries like India, are wells, rivers, and hand-pumps.

Fig. 9 Appearance of red color on the syringe filter units for 5 differentwater samples in the second field trial in India.


Oen the water from these sources are polluted by sewagesludge, agricultural wastes, fertilizers, chemicals, organic,inorganic materials, dead animals and other contaminantsleading to the growth of water-borne pathogens includingbacteria (e.g. Salmonella, Campylobacter, Listeria and variousstrains of E. coli), viruses (e.g. Rota and Hepatitis), protozoa (e.g.Cryptosporidium, Giardia) and other intestinal parasites (e.g.Helminths). Therefore, the deployment of the Mobile Water Kit(MWK) has been focused for developing countries, like India,for continuous monitoring of water quality through the detec-tion of the presence of total coliform and E. coli in water.

Fig. 10 shows an integrated approach of deployment of theMWK in limited resource communities, where not only onedoes the water quality monitoring, but also provides watertreatment solution to the compromised water sources througheffective chlorination treatment. A typical community, asdepicted in Fig. 10, would have a number of wells, water taps,and hand pumps, oen distributed across a wide geographicallocation. The water quality testing using the MWK can becarried out in such communities by the accredited social healthactivist (ASHA) workers (community health female worker), whoare typically employed through the primary health center (PHC)of the community. These ASHA workers, empowered withMWKs and smartphones will upload the test results (image ofthe tested syringe lter unit) to the server through the mHealthE. coli App on their smartphones. The image of the lter surfaceis then processed using mHealth and then over the mobilephone network via SMS. The soware can also access the in-built GPS of the smartphone and thereby identify the testlocation and create a searchable water quality map for thecommunity. The soware can also send messages to all the endusers including villagers and the PHC doctors about the qualityof the water sources. This “early warning” system will help thehealth official (at the PHC) to take appropriate measures forchlorine treatment of the infected wells. Aer effective chlori-nation, the wells will be tested again to assess the quality oftreated water. The mHealth platform has uniquely tagged ID(unique identiers through GPS location tracker in-built withthe mobile phone) for all the wells in that region (includingthose which need chlorination). In parallel, ASHA workers cansend the water samples to a conventional microbiology labo-ratory (like the Bhavan's Research Center (BRC), Mumbai,India) for conrmatory testing of results obtained by the MWK.

The idea of using the MWK is to empower individualcommunity households in limited resource settings with theirchoice of clean water. Typically, each household sends amember of the family (oen women) to fetch water from wellsand other sources in early morning hours. It is envisaged, thateach ASHA worker will be provided with a MWK with necessarysupplies and a smartphone, who will be responsible for 3–4wells in a given community. As explained earlier, these ASHAworkers will perform the water quality test by the procedureoutlined in Fig. 2 (or Movie S1†). At the end of sequentialoperation of dispensing all the chemicals (A, B, C, and D), theperson has to wait for 1 h before capturing the lter surfaceimage by the smartphone camera (Step 10 in Fig. 2). Eventhough the person can observe the presence of red color on the

Anal. Methods, 2014, 6, 6236–6246 | 6243


Fig. 10 Schematic showing an integrated plan for the deployment of the mobile water monitoring system in India.


Publ

ishe

d on

13

June

201

4. D

ownl

oade

d on

23/

08/2

014

07:4

0:24


lter surface much earlier than 1 h and can report it via the Appsystem, shown in Fig. 3, it is important to provide a standard-ized protocol to the ASHA worker. It may so happen that thecontamination level in the wells can be very low (�3 to 5 CFUper 100 ml) and our controlled laboratory experiments, asshown in Fig. 6, suggest that under such low concentrations, thecolor appearance only happens aer around 50 min. Hence, toavoid any possibility of misinterpretation of the MWK result, wefelt that a standard protocol of taking image aer 1 h would be agood practice for the ASHA workers. It is to be noted thatwhile performing eld trials in the designated community inMumbai, the level of contamination in these wells were sosignicant, that we always found the color change to occurwithin 10 min.

It is to be noted that the present version of theMWK does nothave the capability of capturing blue uorescence emitted bythe syringe lter surface in the presence of E. coli in water. Thecurrent MWK only provides qualitative indication of the pres-ence of E. coli/total coliform through the visual observation ofred color on the syringe lter surface. There could be twomechanisms by which one can quantify the number of E. colicells in contaminated water through the use of the MWK (to beincorporated in its future version). Firstly, one can correlate theintensity of the red color on the lter surface, captured by thecell phone camera, with the concentration of E. coli cellsthrough controlled laboratory experiments and then create alook-up table, which can be integrated with the App. Hence, inthe eld, once an image of the ltered surface is captured by thecell phone camera, the end-user can immediately know thepossible range of E. coli concentration in 100 ml of actual eldsamples. Secondly, one can observe the appearance of blue

6244 | Anal. Methods, 2014, 6, 6236–6246

uorescence under exposure to long wavelength ultraviolet light(350–360 nm). The blue uorescence results conrm the pres-ence of E. coli due to the enzymatic reaction of b-glucuronidasewith MUG. One can use a battery operated handheld uores-cence reader to quantify the blue uorescence signal emitted onthe syringe lter surface. However, as mentioned earlier, suchquantication may be of interest to regulators, but have veryless relevance to public at large. The chemicals used in theMWK allow us to adopt both the methods for quanticationpurpose, which will be taken up in near future as one of thegoals for product development.

In this work, the MWK is deployed and tested for drinkingwater only. We are not using the present kit for any industrialwaste water or processed industrial water. Therefore, chances ofany background interference on the color producing reagentsare minimal, which was substantiated by successful eld trialswhere different sources of potable water were tested andmatched with independent laboratory tests for the same watersample. Future scope of present work will be to study the effectof other environmental parameters such as turbidity, salinity,etc.

Further enhancements of the present MWK system areunderway: (a) analysis of different mobile camera models andits effect on the quality of pictures taken by its in-built camera;(b) improving the image processing algorithm to quantify thebacterial count; (c) offline storage and upload support. In thecase of no network coverage, an application will store the dataand will auto upload to a server on availability of the network;(d) once sufficient data related to contamination is gathered,their trend analysis can be done, which can further ne tune thechlorination process.




Publ

ishe

d on

13

June

201

4. D

ownl

oade

d on

23/

08/2

014

07:4

0:24


5 Conclusions

We have demonstrated a simple, easy to use, inexpensive, rapidand eld deployable test method, called Mobile Water Kit(MWK), for detecting total coliform and E. coli in contaminatedwater samples. The present method has two distinct noveltiescompared to conventional methods as well as recently devel-oped advanced methods for bacteria detection. These aresummarized here. (a) Rapid pre-concentration of E. coli from100ml water samples: the method presented here is muchmorerapid compared to current commercial methods due to theinclusion of the pre-concentration step using syringe lterunits. The use of 0.45 mm syringe lter units simultaneouslyenables the pre-concentration of E. coli from 100 ml of water onthe lter surface and also provides a larger surface area in orderto enhance the interaction between the bacteria cells and thechemical reagents. Such a system would result in a faster colordevelopment on the lter surface thereby making the detectionof E. coli simpler and quicker; (b) mHealth E. coli App: thechange in color on the syringe lter unit can be captured using asmartphone and the images can be analyzed using the custombuilt mHealth E. coli App to provide quantitative results. ThemHealth App facilitates communication with the end users viaSMS. The mHealth App can also access the in-built GPS of thesmartphone and thereby can provide a unique location identi-er for the contaminated water sources. This water monitoringplatform is not only capable of providing real-time data forcontaminated water source locations, but can also provide an“early warning” system to detect any outbreak of water-bornediseases in the community.

This method has been successfully tested with knownconcentrations of bacteria as well as with contaminated watersamples in the eld. The signicant feature of this technologyis the ability to sense the presence of coliform bacteria within30 min or less. Compared to conventional standard labora-tory methods, which can take between 24 and 48 hours, thisnew technology can drastically reduce the response time ofcoliform and E. coli detection in water. It was found that theresults from the two eld trials, conducted in Canada andIndia, corroborate well with the laboratory test results fortotal coliform. The MWK is a promising test method for waterquality monitoring in limited resource setting communities.It needs to be further validated for its functionality andapplicability towards broader implementation in communi-ties around the world. The present method is focused on thepresence or absence of the total coliform/E. coli in potablewater samples; hence visual observation of tested ltersurfaces is a sufficient mechanism for the end-user tounderstand whether the water is contaminated or not. Futurescope of the present work is to integrate the LED light sourceof a given excitation wavelength and a receiver to providequantication capability for the MWK. Presently, the MWK isdeployed and tested for drinking water only. Future scope ofthe present work will also be to study the effect of otherenvironmental parameters such as turbidity, salinity andinterferences from other water-borne microbes.


Acknowledgements

Financial support from Enterprise and Advanced Educationthrough research grant is acknowledged here. The authorswould like to thank Stream Technologies, Inc. (Edmonton, AB,Canada) and Biological and Medicinal Chemistry Laboratory(BMCL), University of Alberta, Edmonton, Canada for providingdifferent strains of E. coli. The authors would also like to thankTata Consultancy Services (TCS), Innovation Lab, Mumbai,India, Robonik India Pvt. Ltd., Mumbai, India, Bhavan'sResearch Center, Mumbai, India and King Edward Memorial(KEM) Hospital, Mumbai, India, for their help in collectingwater samples in India, constructive feedback on the developedmethods and conducting the laboratory testing of collectedwater samples for verication. Special thanks to Dr RatendraShinde, Dr Sunita Shanbhag, Dr Mridula Solanki and Dr PawanSable, KEM Hospital, Mumbai, India and Syed Ghouse, TCS,Mumbai, India. The authors gratefully acknowledge RaviShankar Chavali, Lab Technician, Micro and Nano-scaleTransport Laboratory, Department of Mechanical Engineering,University of Alberta, for his help in culturing the bacteriasamples. The authors gratefully thank Aleksey Baldygi, Grad-uate student, Micro and Nano-scale Transport Laboratory,Department of Mechanical Engineering, University of Alberta,for his help in preparation of schematics (Fig. 2) for describingthe MWK procedure.

References

1 E. Kallman, J. A. Smith and V. Oyanedel-Craver, Proceedingsof the Water Environment Federation, 2009, 2009, 19–30.

2 M. D. Sobsey, C. E. Stauber, L. M. Casanova, J. M. Brown andM. A. Elliott, Environ. Sci. Technol., 2008, 42, 4261–4267.

3 J. Bartram and S. Cairncross, PLoS Medicine, 2010, 7,e1000367.

4 A. Pruss-Ustun, R. Bos, F. Gore and J. Bartram, et al., Saferwater, better health: costs, benets and sustainability ofinterventions to protect and promote health, World HealthOrganization, 2008.

5 L. Fewtrell and J. Bartram, Water quality: Guidelines,standards, and health: assessment of risk and riskmanagement for water-related infectious disease, IWAPublishing, 2001.

6 K. Onda, J. LoBuglio and J. Bartram, Int. J. Environ. Res. PublicHealth, 2012, 9, 880–894.

7 J. Olstadt, J. Schauer, J. Standridge and S. Kluender, J. WaterHealth, 2007, 5, 267–282.

8 E. Machdar, N. van der Steen, L. Raschid-Sally and P. Lens,Sci. Total Environ., 2013, 449, 134–142.

9 M. Du Preez, K. G. McGuigan and R. M. Conroy, Environ. Sci.Technol., 2010, 44, 8744–8749.

10 C. Boschi-Pinto, C. F. Lanata and R. E. Black, Maternal andChild Health, Springer, 2009, pp. 225–243.

11 U. Von Gunten, Water Res., 2003, 37, 1469–1487.12 M. I. Cedergren, A. J. Selbing, O. Lofman and B. A. Kallen,

Environ. Res., 2002, 89, 124–130.

Anal. Methods, 2014, 6, 6236–6246 | 6245



Publ

ishe

d on

13

June

201

4. D

ownl

oade

d on

23/

08/2

014

07:4

0:24


13 A. Adin, J. Katzhendler, D. Alkaslassy and C. Rav-Acha,WaterRes., 1991, 25, 797–805.

14 M. H. Cassin, A. M. Lammerding, E. C. Todd, W. Ross andR. S. McColl, Int. J. Food Microbiol., 1998, 41, 21–44.

15 K. Dhriti, G. Mathur, R. Prem, M. Singh, V. Rawat,G. Seshadri, A. Manjeet and R. Khandal, Water and Health,Springer, 2014, pp. 305–353.

16 P. Payment, J. Siemiatycki, L. Richardson, G. Renaud,E. Franco and M. Prevost, Int. J. Environ. Health Res., 1997,7, 5–31.

17 J. Bartram, Heterotrophic plate counts and drinking-watersafety: the signicance of HPCs for water quality and humanhealth, IWA Publishing, 2003.

18 I. Deportes, J.-L. Benoit-Guyod and D. Zmirou, Sci. TotalEnviron., 1995, 172, 197–222.

19 R. Girones, Microbe-American Society for Microbiology, 2006,vol. 1, p. 19.

20 M. Steele and J. Odumeru, J. Food Prot., 2004, 67, 2839–2849.

21 E. M. Smith and C. P. Gerba, Appl. Environ. Microbiol., 1982,43, 1440–1450.

22 V. Richardson, J. Hernandez-Pichardo, M. Quintanar-Solares, M. Esparza-Aguilar, B. Johnson, C. M. Gomez-Altamirano, U. Parashar and M. Patel, N. Engl. J. Med.,2010, 362, 299–305.

23 M. Grant, Appl. Environ. Microbiol., 1997, 63, 3526–3530.24 R. Lifshitz and R. Joshi, Environ. Toxicol. Water Qual., 1998,

13, 157–164.25 R. T. Noble, S. B. Weisberg, M. K. Leecaster, C. D. McGee,

K. Ritter, K. O. Walker and P. M. Vainik, Environ. Monit.Assess., 2003, 81, 301–312.

26 R. T. Noble, M. K. Leecaster, C. D. McGee, S. B. Weisberg andK. Ritter, Water Res., 2004, 38, 1183–1188.

27 S. B. Shinde, C. B. Fernandes and V. B. Patravale, J.Controlled Release, 2012, 159, 164–180.

28 G. Lopez-Campos, J. V. Martınez-Suarez, M. Aguado-Urdaand V. Lopez-Alonso, Microarray Detection andCharacterization of Bacterial Foodborne Pathogens, Springer,2012, pp. 13–32.

29 P. Arora, A. Sindhu, N. Dilbaghi and A. Chaudhury, Biosens.Bioelectron., 2011, 28, 1–12.

30 A. K. Bhunia, V. Nanduri, E. Bae and E. D. Hirleman,Encyclopedia of Industrial Biotechnology: Bioprocess,Bioseparation, and Cell Technology, 2010.

31 V. Velusamy, K. Arshak, O. Korostynska, K. Oliwa andC. Adley, Biotechnol. Adv., 2010, 28, 232–254.

32 A. K. Bhunia, Adv. Food Nutr. Res., 2008, 54, 1–44.33 O. Tokarskyy and D. L. Marshall, Food Microbiol., 2008, 25, 1.34 O. Lazcka, F. Campo and F. X. Munoz, Biosens. Bioelectron.,

2007, 22, 1205–1217.35 A. K. Deisingh and M. Thompson, Can. J. Microbiol., 2004,

50, 69–77.36 R. Kwapiszewski, M. Skolimowski, K. Ziołkowska, E. Jedrych,

M. Chudy, A. Dybko and Z. Brzozka, Biomed. Microdevices,2011, 13, 431–440.

37 A. J. Baeumner, R. N. Cohen, V. Miksic and J. Min, Biosens.Bioelectron., 2003, 18, 405–413.

6246 | Anal. Methods, 2014, 6, 6236–6246

38 H. Zhu, U. Sikora and A. Ozcan, Analyst, 2012, 137, 2541–2544.

39 J. Comas-Riu and N. Rius, J. Ind. Microbiol. Biotechnol., 2009,36, 999–1011.

40 R. Kong, S. Lee, T. Law, S. Law and R. Wu, Water Res., 2002,36, 2802–2812.

41 N. Wery, C. Lhoutellier, F. Ducray, J.-P. Delgenes andJ.-J. Godon, Water Res., 2008, 42, 53–62.

42 L. Heijnen and G. Medema,Water Res., 2009, 43, 3124–3132.43 J. E. Afset, G. Bruant, R. Brousseau, J. Harel, E. Anderssen,

L. Bevanger and K. Bergh, J. Clin. Microbiol., 2006, 44,3703–3711.

44 S. Bekal, R. Brousseau, L. Masson, G. Prefontaine,J. Fairbrother and J. Harel, J. Clin. Microbiol., 2003, 41,2113–2125.

45 G. Bruant, C. Maynard, S. Bekal, I. Gaucher, L. Masson,R. Brousseau and J. Harel, Appl. Environ. Microbiol., 2006,72, 3780–3784.

46 S. Chen, S. Zhao, P. F. McDermott, C. M. Schroeder,D. G. White and J. Meng,Mol. Cell. Probes, 2005, 19, 195–201.

47 J. Waswa, J. Irudayaraj and C. DebRoy, LWT–Food Sci.Technol., 2007, 40, 187–192.

48 C.-J. Huang, J. Dostalek, A. Sessitsch and W. Knoll, Anal.Chem., 2011, 83, 674–677.

49 V. Jasson, L. Jacxsens, P. Luning, A. Rajkovic andM. Uyttendaele, Food Microbiol., 2010, 27, 710–730.

50 P. Kiesel, J. Martini, M. Huck, M. Bern and N. Johnson,CLEO: Science and Innovations, 2011.

51 L. Yang, Talanta, 2008, 74, 1621–1629.52 L. Yang and R. Bashir, Biotechnol. Adv., 2008, 26, 135–150.53 X. Guo, A. Kulkarni, A. Doepke, H. B. Halsall, S. Iyer and

W. R. Heineman, Anal. Chem., 2011, 84, 241–246.54 B. Bottari, D. Ercolini, M. Gatti and E. Neviani, Appl.

Microbiol. Biotechnol., 2006, 73, 485–494.55 M. Mana, Int. J. Food Microbiol., 2000, 60, 205–218.56 W. P. Hamilton, M. Kim and E. L. Thackston, Water Res.,

2005, 39, 4869–4878.57 M. Mana, W. Kneifel and S. Bascomb,Microbiol. Rev., 1991,

55, 335–348.58 M. Mana, Int. J. Food Microbiol., 1996, 31, 45–58.59 S. K. Mitra, N. S. K. Gunda, S. Naicker and P. Banerji, US

patent application, 62007133, Coliform Detection andReporting, 2014.

60 S. K. Mitra, N. S. K. Gunda, S. Naicker, S. Bhattacharjee,T. Thundat and M. Ghoraishi, US patent application,62010821, Microuidic Detection of Coliform Bacteria, 2014.

61 W. Cuthbert, J. Panes and E. Hill, J. Appl. Microbiol., 1955, 18,408–414.

62 B. Eddy and A. Kitchell, J. Appl. Microbiol., 1959, 22, 57–63.63 J. E. Sworobuk, C. B. Law and G. K. Bissonnette, Water, Air,

Soil Pollut., 1987, 36, 163–170.64 K. Venkateswaran, A. Murakoshi and M. Satake, Appl.

Environ. Microbiol., 1996, 62, 2236–2243.65 S. C. Edberg and C. M. Kontnick, J. Clin. Microbiol., 1986, 24,

368–371.66 R. M. NIEMI, J. Mentu, A. Siitonen and S. Niemela, J. Appl.

Microbiol., 2003, 95, 1242–1249.



mobile water kit (mwk): a smartphone compatible low-cost water monitoring system for rapid detection...

Documents