1 Supporting Information Adapting Enzyme-Based Microbial Water Quality Analysis to Remote Areas in Low-Income Countries Adam Abramson * , Maya Ben Ami, Noam Weisbrod Department of Environmental Hydrology & Microbiology, Zuckerberg Institute for Water Research, The Jacob Blaustein Institutes for Desert Research, Ben-Gurion University of the Negev, Sede Boqer Campus, Midreshet Ben-Gurion, 84990 Israel * Corresponding author. Tel.: +972 8 6563433; fax: +972 8 6596909 E-mail address: [email protected]Number of Pages: 11 Number of Figures: 3 Number of Tables: 7
Transcript
1
Supporting Information
Adapting Enzyme-Based Microbial Water Quality
Analysis to Remote Areas in Low-Income Countries
Adam Abramson*, Maya Ben Ami, Noam Weisbrod
Department of Environmental Hydrology & Microbiology, Zuckerberg Institute for Water Research,
The Jacob Blaustein Institutes for Desert Research, Ben-Gurion University of the Negev, Sede
Ten enzyme-based methods were approved by the US Environmental Protection Agency
(EPA) between 1999 and 2004, and their relevance to remote water monitoring, as well as
comparative performance as investigated by Olstadt et al.1, are summarized in Table S1. We applied
these findings to remote areas according to the following criteria: (1) ability to simplify aseptic
procedure; (2) suppress non-fecal bacteria, and (3) reliably detect E. coli. Accordingly, 4 of the 10
techniques hold potential for remote applications.
Table S1: Comparative analysis of 10 EPA-approved enzyme-based methodologies.2-14 Potential for
remote areas is positive if all three preceding criteria are fulfilled.
Method
Simplify
aseptic
procedure?
Suppress non-
fecal bacteria
reliably?
Observed Failure Rate
Quantification
possible?
Potential for
remote areas? % Acceptable?
Colilert + + 0 + + +
Colilert-18 + + 3.3 + + +
Colisure – 24 h + + 20 +
Colisure – 48 h + + 0 + + +
mColiBlue 24 23 +
Readycult + + 20
Chromocult 0 + +
Coliscan 0 + +
E*Colite + 0 (E. coli) +
MI Agar 0 + +
Colitag + + 0 + +
S2. Colisure Methodology
Colisure’s Defined Substrate Technology is based on utilization of the specific enzymatic
activity of β-D galactosidase and β-D glucuronidase for the detection of total coliforms and
Escherichia coli, respectively. Identification of these enzymes is accomplished first by resuscitation
and enumeration of the target organisms through a combination of nutrients and inorganic salts, and
then by hydrolysis of chromogenic or fluorogenic enzyme substrates. At the same time, the medium
selectively inhibits the growth of non-coliforms through a combination of detergents and antibiotics.
3
When total coliforms metabolize Colisure’s nutrient-indicator chlorophenol red
ß-D-galactopyranoside (CPRG), the sample turns from yellow to red/magenta. When E. coli
metabolizes Colisure’s nutrient-indicator 4-methyl-umbelliferyl-ß-D-glucoronide (MUG), the
sample also fluoresces under a longwave fluorescent UV lamp.5 Colisure can simultaneously detect
these bacteria at 1 CFU/100 mL within 24 h, even in the presence of as many as 2 million
heterotrophic bacteria per 100 mL.5 The on-site requirements for this process are an incubator and a
portable 6 W, longwave (365 nm) UV lamp source. A supply of 100-mL sterile plastic, non-
fluorescing sampling bottles, precoated with non-coliform inhibitors, are included with a snap-pack
of reagent medium, which is added to the sample within 6 h of sampling, and immediately before
incubation. After incubation, the samples are analyzed for total coliforms with the naked eye, and for
E. coli with the fluorescent light, according to the simple color scheme outlined in Table S2.
Table S2. Colisure Interpretation Scheme 5
Appearance Result
Yellow/gold Negative for total coliforms and E. coli
Red or magenta Positive for total coliforms
Red/magenta and fluorescent Positive for E. coli
Pink or orange Incubate sample 4 additional h (up to 48 h maximum). Before 48 h, shake the bottle and check for red/magenta color change and fluorescence, indicating positive for total coliforms and/or positive for E. coli.
If negative after 24 h, samples are incubated for an additional 24 h to avoid false negatives as
recommended by Chuang et al.15 and Olstadt et al.1. Samples are considered negative if after 48 h,
no red/magenta color and/or fluorescence can be seen. Red/magenta or red/magenta with
fluorescence observed before the 24 h time point is a valid positive. No sample is considered valid
after 48 h because heterotrophic bacteria can overwhelm Colisure’s inhibition system and turn the
sample red/magenta or red/magenta with fluorescence. A sample that remains pink or orange at 48 h
is considered negative for both coliforms and E. coli.
4
Table S3. False Positive or Negative Readings of Colisure at Various Incubation Temperatures
in a Standard Laboratory Incubator
Incubation
temperature (°C)
Staphylo-
coccus
aureus
Enterococcus
faecalis
Klebsiella
pneumonia
Escherichia
coli
Toilet water
25
27
30
32
35
37
40
42
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
100
0
0
0
0
0
0
0
0
0
0
100
66.7
0
0
0
0
33.3
100
0
0
0
0
0
0
5
S3. Field Incubator Performance
Figure S1. (A) In-situ water sample temperatures at three points in the field incubator over 7 days'
incubation during the second (rainy season) sampling. (B) The field incubator layout as used in the
field study in Zambia, with position numbers labeled, and specified width and height dimensions.
While there exists a visible spatial variability in relation to distance from the heat source,
these differences are such that incubation ranges at all locations are within an acceptable temperature
range. For locations at, near and far from the heat source, incubation ranges are 36.36 ± 0.90, 35.90
6
± 1.26 and 34.07 ± 1.88, respectively. The greatest point of diversion from mean temperatures
corresponded to the peak of incubation volume, when initial in-situ temperatures were lower.
Variability decreases significantly after peak volume. The calculated performance for the first 3 days
is 35.26 ± 1.64, whereas for the last 4 days, it is 35.56 ± 0.40.
S4. Water Sample Results by Containers
In addition to the samples of water sources, water samples were taken from containers used to
transport, store and consume water. Figure S2 demonstrates the distribution of contamination results
by container type.
Figure S2. Distribution of Colisure results by container type. Samples of transport containers were
taken only in cases where separate containers were used to store and transport water.
S5. Regression Analysis
Nominal logistic and standard least squares regression were conducted to investigate the impact of
various factors on the presence of fecal contamination in the transport and storage containers and the
user perception of water quality, respectively. The results are presented in Tables S4 and S5.
7
Table S4. Results of Nominal Logistics Model of E. coli in Container (for Log Odds of 0/1)
A.
Model -LogLikelihood DF ChiSquare Prob>ChiSq Difference 17.055287 13 34.11057 0.0012* Full 50.137519 Reduced 67.192806 RSquare (U) 0.2538 AICc 131.635 BIC 169.458 Observations (or sum wts)
140
B.
Term Estimate Std Error ChiSquare Prob>ChiSq Intercept -4.0676536 1.0172219 15.99 <.0001* Fecal at source? [no] 1.71569169 0.5886745 8.49 0.0036* Hours in storage 0.01097043 0.0106303 1.07 0.3021 Last time container cleaned 0.02106018 0.0566306 0.14 0.7100 Sand to wash container? [no] 0.44899771 0.3269666 1.89 0.1697 Soap to wash container? [no] -0.1284145 0.3070173 0.17 0.6758 Number of children 0.18882906 0.146441 1.66 0.1972 Treat water? [always] -0.4220301 0.7992461 0.28 0.5975 Treat water? [never] 0.70308283 0.4946333 2.02 0.1552 Use soap to wash hands? [always]
0.07032226 0.3973925 0.03 0.8595
Use soap to wash hands? [never] 0.23320406 0.3998865 0.34 0.5598 Use a latrine? [no] -0.0498316 0.2895127 0.03 0.8633 Dip cup? [no] 0.14152394 0.3199168 0.20 0.6582 Container closed [no] 0.34818967 0.2718608 1.64 0.2003
Table S5. Least Squares Regression of User Perception of Water Quality (1 = “I Fear Very
Much that Someone May Get Sick from It” and 5 = “I Am Sure No One Will Get Sick from It”)
A.
Source DF Sum of Squares Mean Square F Ratio Model 8 108.15424 13.5193 11.6783 Error 343 397.07019 1.1576 Prob > F C. total 351 505.22443 <.0001*
B.
Term Estimate Std Error t Ratio Prob>|t| Intercept 2.626577 0.083962 31.28 <.0001* Sick from water? [0] 0.3604545 0.079794 4.52 <.0001* Water path point [storage] 0.0488118 0.109571 0.45 0.6563 Water path point [transport] 0.068981 0.13997 0.49 0.6225 E. coli? [0] 0.3623515 0.0769 4.71 <.0001* Water path point[storage]*E. coli? [0] -0.277748 0.098114 -2.83 0.0049* Water path point[transport]*E. coli? [0] 0.1173343 0.131155 0.89 0.3716 Water path point[storage]*Sick from water? [0] -0.037835 0.099262 -0.38 0.7033 Water path point[transport]*Sick from water? [0] -0.119844 0.134695 -0.89 0.3742
Table S6 presents the results of chi-square test of independence between reported diarrhea and
presence of fecal contamination, for source and point-of-use water samples.
8
Table S6. Results of Chi-Square Test of Independence: Reported Diarrhea by Fecal
Contamination in Source and Point-of-Use Water Samples
Water Point Test N DF -LogLikelihood R Square ChiSquare Prob > ChiSquare
Source Likelihood Ratio 164 1 1.489 0.023 2.978 0.0844
Pearson 3.019 0.0823
Point-of-use Likelihood Ratio 163 1 1.362 0.0211 2.724 0.0988
Pearson 2.312 0.1284
As indicated by Pearson’s chi-square for both source (p = 0.0823) and point-of-use (p = 0.128) water
samples, although the evidence raises doubt as to the independence of experienced diarrhea and
presence of fecal contamination, it is insufficient to reject the null hypothesis. The relationship is
stronger at the source than at the point-of-use.
S6. Comparative Analysis of Microbial Water Test Feasibility in Remote Areas
While there are numerous feasible approaches for on-site monitoring of microbial water
quality in remote areas, only a small number meet the requirement of being approved or accepted as
standard practice. Other techniques, such as the H2S strip test16,17 , are technically feasible in low
resource settings, but have not been approved by the U.S. EPA or listed in the Standard Methods for
the Examination of Water and Wastewater. Portable field kits (PFKs) that modify conventional
membrane-filtration tests into a miniature laboratory format remove the need for most of the
conventional requirements except for highly trained staff, since aseptic procedure is still required for
analysis. As a result, analysis becomes feasible in very remote areas, but the costs remain high due
to the small size of the kits needed for portability (i.e. number of samples that can be analyzed per
day) and the requirement of a paid technician. We investigate Potatest and Potaflex (Wagtech, UK),
which offer a wide range of physicochemical and microbiological water analyses in a modular,
portable kit.18
To enumerate total and fecal coliforms, 18 and 40 reusable aluminum petri dishes are
included in each kit, respectively. This means that these kits can only process 18 and 40 petri dishes
per 14-h round, and these dishes must be resterilized between incubations.
The Colilert-Petrifilm package known as the EC-Kit removes the need for lab-grade
incubation through use of a specialized waistbelt to incubate the samples with body heat.15
A summary of the technical complexity of each on-site approach investigated is presented in
Figure S3.
9
Figure S3. A comparison of on-site water-analysis approaches across four categories of technical
complexity. Complexity is investigated across a qualitative array of indicators including incubation,
aseptic procedure, equipment and electricity requirements. Across all indicators, the portable field
kits are more complex than the alternatives.
S7. Cost Analysis of Microbial Tests
The costs of each analysis ($ per sample) were calculated according to Eq. 1:
Cost = Equipment Cost + Test Material Cost + Sampling Cost +
�
Operation Cost
Samples Analyzed (1)
where equipment refers to any non-consumable component of the test and test material is the
consumable component. Table S7 provides the cost figures used in these calculations.
Table S7. Overview of Costs Associated with Four Feasible On-Site Microbial Water Tests
Method Equipment cost
Test material cost (per sample)
Sampling cost (per sample)
Samples analyzed (per
person per day)
Operation cost range
investigated ($/day)
Potatest $1,900/unit $0.40
$0.50 18/unit* 50-200
Potaflex $3,000/ unit $0.40 40/unit* 50-200
EC-Kit $40 $4.66 100 0-100
Colisure $65 $2.00 100 0-100
*We assume a limit of 200 samples per person per day.
Sampling costs are the same for all methods, since they all require the sterile collection of
100-mL samples. We assume that a local resident with minimal training can sample 10 water points
per day at a daily wage of $5. Equipment costs vary considerably by method. Potatest and Potaflex
cost $1,900 and $3,000, respectively, for one kit including all equipment as well as consumables for
performing 200 tests. Every additional order of 200 samples costs $80. For these methods, sampling
efforts with 1, 3, 5 and 10 portable field kits were investigated, with only the least-cost configuration
10
presented. We assume that no dilutions are performed, in order to detect, not enumerate, fecal
contamination. For the other methods, equipment consists of the portable UV light ($40) and, for
Colisure, the space heater ($25).
The number of samples analyzed also differs by methodology. Only 18 and 40 samples can
be analyzed per field kit per day, whereas we assume that 100 samples can be analyzed per person
per day with the enzyme-based methods. This is because of the technical restrictions associated with
field-sized incubation with reusable petri dishes.
Operational costs differ between methodologies requiring local and external staff due to the
additional costs of transportation, accommodation and wages for external technicians. Cost of
analysis is determined primarily by the amount of training required by the practitioner. If a trained
microbiologist or laboratory technician must be hired, the cost is higher than if the analysis can be
performed primarily by, for example, community health workers. Rural health clinicians could
feasibly be paid $50 per day or less, whereas the operation costs of bringing external technicians is
not likely to be less than $50 per day. Transport costs depend primarily on the distance between the
water sampling and the analysis.
11
References
(1) Olstadt, J.; Schauer, J.; Standridge, J. A comparison of ten US EPA approved total coliform/E. coli tests. Journal of water and health 2007 2(5), 267-282.
(2) Charm Sciences, Inc. 2004 E*Colite Manufacturer Literature, 659 Andover St., Lawrence, MA 01843
(3) Colitag manufacturer literature, CPI International, 5580 Skylane Blvd, Santa Rosa, CA 95403
(4) Hach Company 2004 m-ColiBlue24 Manufacturer Literature. PO Box 389, Loveland, CO 80539
(5) IDEXX Laboratories, Inc. 2004 Colilert , Colilert-18 , Colisure. One IDEXX Dr, Westbrook, ME 04092
(6) Micrology Laboratories 2004 Coliscan MF Manufacturer Literature. 1303 Eisenhower Dr South, Goshen, IN 46526.
(7) US EPA - Federal Register 40 CFR Part 141, June 29, 1989a, Vol. 54, No. 124 - Colilert (total coliform).
(8) US EPA - Federal Register 40 CFR Part 141, June 17, 1989b, Vol. 54, No. 135. - MPN Method.
(9) US EPA - Federal Register 40 CFR Part 141, June 10, 1992, Vol. 57, No. 112 - Colilert, Colilert-18 Final Approval.
(10) US EPA - Federal Register 40 CFR Part 141, Dec. 5, 1994 - Colisure. (11) US EPA - Federal Register 40 CFR Part 141, July 31, 1998, Vol. 63. - MI Agar. (12) US EPA - Federal Register 40 CFR Part 141, December 1, 1999, Vol. 64, No. 230. -
Colisure -24 hour, E * Colite and m-Coli Blue 24. (13) US EPA - Federal Register 40 CFR Part 141, October 29, 2002, Vol. 67, No. 209. -
Chromocult and Readycult Coliforms 100. (14) US EPA - Federal Register 40 CFR Part 141, February 13, 2004, Vol. 69, No. 30. - Colitag. (15) Chuang, P.; Trottier, S.; Murcott, S. Comparison and verification of four field-based
microbiological tests: H 2S test, Easygel ®, Colilert ®, Petrifilm™. Journal of Water,
Sanitation and Hygiene for Development 2011, 1, 68-85. (16) WHO (World Health Organization). Evaluation of the H2S Method for Detection of Fecal
Contamination of Drinking Water. 2002, 1–44. (17) Pathak, S. P.; Gopal, K. Efficiency of Modified H2S Test for Detection of Faecal
Contamination in Water. Environ Monit Assess 2005, 108, 59–65. (18) Wagtech, Inc. Potatest WE-10005 Instructions, available online at www.wagtech.co.uk
