Rapid Concentration and Molecular Enrichment Approach for Sensitive Detection of Escherichia coli and Shigella Species in Potable Water Samples

Abstract

In this work, we used a rapid, simple, and efficient concentration-and-recovery procedure combined with a DNA enrichment method (dubbed CRENAME [concentration and recovery of microbial particles, extraction of nucleic acids, and molecular enrichment]), that we coupled to an Escherichia coli/Shigella-specific real-time PCR (rtPCR) assay targeting the tuf gene, to sensitively detect E. coli/Shigella in water. This integrated method was compared to U.S. Environmental Protection Agency (EPA) culture-based Method 1604 on MI agar in terms of analytical specificity, ubiquity, detection limit, and rapidity. None of the 179 non-E. coli/Shigella strains tested was detected by both methods, with the exception of Escherichia fergusonii, which was detected by the CRENAME procedure combined with the E. coli/Shigella-specific rtPCR assay (CRENAME + E. coli rtPCR). DNA from all 90 E. coli/Shigella strains tested was amplified by the CRENAME + E. coli rtPCR, whereas the MI agar method had limited ubiquity and detected only 65 (72.2%) of the 90 strains tested. In less than 5 h, the CRENAME + E. coli rtPCR method detected 1.8 E. coli/Shigella CFU whereas the MI agar method detected 1.2 CFU/100 ml of water in 24 h (95% confidence). Consequently, the CRENAME method provides an easy and efficient approach to detect as little as one Gram-negative E. coli/Shigella cell present in a 100-ml potable water sample. Coupled with an E. coli/Shigella-specific rtPCR assay, the entire molecular procedure is comparable to U.S. EPA Method 1604 on MI agar in terms of analytical specificity and detection limit but provides significant advantages in terms of speed and ubiquity.

INTRODUCTION

The first link between enteric disease and water contaminated with fecal waste was demonstrated by Snow and Budd in 1855 (25). The recognition that safe water should be free of pathogens prompted investigators to search for means of indexing water quality. As an important member of the normal flora of the gastrointestinal tracts of humans and others mammals, Escherichia coli was found to be a suitable indicator of fecal contamination (2, 7, 13, 16) because it is specific and reliably reflects fecal contamination (12).

Since E. coli has been established as the most reliable indicator of human fecal contamination to predict the microbiological quality of potable water, many PCR assays have been proposed to complement or substitute for conventional recommended culture-based methods to monitor its presence (1, 26). The specificity of the uid chromosomal region was confirmed for E. coli and Shigella spp. by Cleuziat and Robert-Baudouy (6). Since then, the uid gene has come to be considered an ideal candidate target for DNA-based assays for detecting E. coli and Shigella species. Other housekeeping genes, such as tuf (elongation factor Tu) or clpB (heat shock protein F84.1), have also been used to design E. coli/Shigella-specific DNA-based assays (11, 23, 24).

In the field of water microbiology, there is a need for more rapid tests to improve public health protection since water is an important route of transmission for many of the most widespread and debilitating diseases that afflict humans (29). The implementation of molecular methods could represent a suitable avenue for such tests (14). However, the application of rapid molecular testing to the monitoring of the microbiological quality of potable water is lagging, mainly due to the scarcity of simple technological solutions for tackling the major task of efficiently concentrating and recovering as little as one microbial particle (indicator and/or pathogen) from a water sample (32).

Recently, Maheux et al. (21) showed the detection of 4.5 Gram-positive enterococcal CFU/100 ml of potable water in less than 5 h using a rapid and efficient concentration-and-recovery procedure coupled with a real-time PCR (rtPCR) assay, while culture-based Method 1600 on membrane-Enterococcus indoxyl-β-d-glucoside agar detected 2.3 CFU/100 ml in 24 h (21). In this study, we demonstrate that this procedure can also be used to sensitively and rapidly detect more fragile Gram-negative E. coli/Shigella cells in potable water samples.

MATERIALS AND METHODS

Ubiquity and analytical specificity of the MI agar and E. coli/Shigella-specific rtPCR assays.

The ubiquity (ability to detect all or most E. coli and Shigella strains) of the MI agar method (31) and the 2 E. coli/Shigella rtPCR primer sets tested in this study was verified by using a panel consisting of 79 E. coli strains of both clinical and environmental origins and the 11 Shigella strains (Shigella boydii, S. dysenteriae, S. flexnerii, and S. sonnei) previously used by Maheux et al. (23) (Table 1). The analytical specificity of the MI agar method and each rtPCR assay was demonstrated by testing a battery of strains consisting of 36 Gram-positive and 110 other representative Gram-negative bacterial species (Tables 2 and 3).

Table 1.

Ability of the culture-based MI agar method and the two primer-and-probe sets to detect E. coli and Shigella sp. strains

Species,a origin, and strainc	Serotype	MI agar assay result	E. coli/Shigella-specificd rtPCR assay result
			uidA	tuf
E. coli
Clinical isolates
ATCC 11105	NAb	+	+	+
ATCC 11775	O1:K1:H7	+	+	+
ATCC 14763	NA	+	+	+
ATCC 23500	NA	+	−	+
ATCC 23510	O15:K14(L):H4	+	+	+
ATCC 23511	O16:K1(L):NM	+	+	+
ATCC 25922	NA	+	−	+
ATCC 29194	NA	+	−	+
ATCC 33475	NA	−	+	+
ATCC 33476	NA	−	+	+
ATCC 35401	O78:H11	+	+	+
ATCC 35218	NA	+	+	+
ATCC 39188	NA	−	+	+
ATCC 43886	O25:K98:NM	+	+	+
ATCC 43890	O157:H7	−	+	+
ATCC 43894	O157:H7	−	+	+
ATCC 43895	O157:H7	−	+	+
ATCC 43896	O78:K80:H12	+	+	+
ATCC 47076	NA	−	+	+
CCRI-1191	NA	+	+	+
CCRI-1192	NA	+	+	+
CCRI-1193	NA	+	+	+
CCRI-1213	NA	+	+	+
CCRI-2099	NA	−	+	+
CCRI-2166	NA	−	+	+
CCRI-8831	O157:H7	−	+	+
CCRI-8832	O157:H8	−	+	+
CCRI-8833	O103:H2	+	+	+
CCRI-8834	O103:H3	+	+	+
CCRI-8835	O111:H-	+	+	+
CCRI-8836	O111:H-	+	+	+
CCRI-8837	O26:NM	−	+	+
CCRI-8838	O26:NM	+	+	+
CCRI-8839	O145:NM	+	+	+
CCRI-8840	O145:NM	+	+	+
CCRI-9493	NA	+	−	+
LSPQ 2082	O4:H5	+	−	+
LSPQ 2085	O7:NM	+	+	+
LSPQ 2086	O8:H9	+	+	+
LSPQ 2089	O12:NM	−	+	+
LSPQ 2092	O18:NM	+	−	+
LSPQ 2096	O26:NM	+	+	+
LSPQ 2113	O111:NM	+	+	+
LSPQ 2115	O128:H8	+	+	+
LSPQ 2117	O113:H21	−	+	+
LSPQ 2118	O117:H4	+	+	+
LSPQ 2125	O128:NM	+	+	+
LSPQ 2127	O157:H7	−	+	+
LSPQ 3760	O157:H7	−	+	+
LSPQ 3761	O157:H7	−	+	+
LSPQ 3762	O157:H7	−	+	+
Environmental isolates
CCRI-14813	NA	+	−	+
CCRI-14858	NA	+	+	+
CCRI-14859	NA	+	+	+
CCRI-14871	NA	+	+	+
CCRI-14881	NA	+	+	+
CCRI-16465	NA	+	+	+
CCRI-16485	NA	+	+	+
CCRI-16540	NA	+	+	+
CCRI-16527	NA	+	−	+
CCRI-16528	NA	+	+	+
CCRI-16537	NA	+	−	+
CCRI-16539	NA	+	+	+
CCRI-16579	NA	+	+	+
CCRI-16580	NA	+	+	+
CCRI-17006	NA	+	+	+
CCRI-17021	NA	+	+	+
CCRI-17027	NA	+	+	+
CCRI-17042	NA	+	−	+
CCRI-17045	NA	+	+	+
CCRI-17056	NA	+	+	+
CCRI-17063	NA	+	+	+
CCRI-17065	NA	+	−	+
CCRI-17097	NA	+	+	+
CCRI-17151	NA	+	+	+
CCRI-17158	NA	+	+	+
CCRI-17161	NA	+	+	+
CCRI-17172	NA	+	+	+
CCRI-17176	NA	+	+	+
All E. coli strains		61/79 (77.2)e	68/79 (86.1)e	79/79 (100)e
S. boydii
ATCC 8700	NA	−	−	+
ATCC 9207	NA	+	+	+
ATCC 12032	13	−	−	+
S. dysenteriae
ATCC 11835	1	−	−	+
CCRI-8843	NA	−	−	+
CCRI-8844	NA	−	+	+
S. flexneri
ATCC 12022	2b	−	+	+
CCRI-2198	NA	−	+	+
S. sonnei
ATCC 29930	NA	+	+	+
ATCC 25931	NA	+	−	+
CCRI-2196	NA	+	+	+
All Shigella strains		4/11 (36.4)e	6/11 (54.5)e	11/11 (100)e

^aSeventy-nine strains were studied.

^bNA, not available.

^cCCRI, Collection of the Centre de Recherche en Infectiologie.

^dE. coli/Shigella-specific rtPCR assays: uidA (10), primers 784F and 866R and probe EC807; tuf (this study), primers TEcol553 and TEcol754 and probe TEco573-T1-B1.

^eNumber detected/total (percentage of total).

Table 2.

Gram-positive bacteria used for specificity analysis^a

Species	Strain
Abiotrophia defectiva	ATCC 49176
Clostridium lavalense	CCRI-9842
Enterococcus aquimarinus	CCRI-15963
Enterococcus avium	ATCC 14025
Enterococcus caccae	CCUG 51564
Enterococcus canis	CCUG 46666
Enterococcus casseliflavus	CCUG 37857
Enterococcus canintestini	CCRI-19376
Enterococcus cecorum	ATCC 43198
Enterococcus columbae	ATCC 51263
Enterococcus devriesei	CCUG 37865
Enterococcus dispar	ATCC 51266
Enterococcus durans	ATCC 19432
Enterococcus faecalis	ATCC 19433
Enterococcus faecalis	ATCC 29212
Enterococcus faecium	ATCC 19434
Enterococcus faecium	ATCC 700221
Enterococcus flavescens	ATCC 49996
Enterococcus gallinarum	LSPQ 3364
Enterococcus gilvus	CCUG 45553
Enterococcus haemoperoxidus	CCUG 45916
Enterococcus hirae	ATCC 8043
Enterococcus italicus	CCUG 50447
Enterococcus malodoratus	ATCC 43197
Enterococcus moraviensis	CCUG 45913
Enterococcus mundtii	ATCC 43186
Enterococcus pallens	CCUG 45554
Enterococcus phoeniculicola	CCUG 18923
Enterococcus pseudoavium	ATCC 49372
Enterococcus raffinosus	ATCC 49427
Enterococcus ratti	ATCC 700914
Enterococcus saccharolyticus	ATCC 43076
Enterococcus silesiacus	CCUG 53830
Enterococcus sulfureus	ATCC 49903
Enterococcus termitis	CCUG 53831
Enterococcus villorum	CCUG 43229
Gemella haemolysans	ATCC 10379
Granulicatella adiacens	ATCC 49175
Kocuria rhizophila	ATCC 9341
Lactobacillus acidophilus	ATCC 4356
Leifsonia aquatica	ATCC 14665
Listeria grayi	ATCC 19120
Listeria innocua	ATCC 33090
Listeria ivanovii	ATCC 19119
Listeria monocytogenes	ATCC 15313
Listeria seeligeri	ATCC 35967
Ruminococcus gauvreauii	CCRI-16110
Staphylococcus aureus	ATCC 25923
Staphylococcus capitis subsp. capitis	ATCC 27840
Staphylococcus epidermidis	ATCC 14990
Staphylococcus haemolyticus	ATCC 29970
Staphylococcus hominis subsp. hominis	ATCC 27844
Staphylococcus lugdunensis	ATCC 43809
Staphylococcus saprophyticus subsp. saprophyticus	ATCC 15305
Staphylococcus simulans	ATCC 27848
Staphylococcus warneri	ATCC 27836
Streptococcus agalactiae	ATCC 13813
Streptococcus anginosus	ATCC 33397
Streptococcus bovis	ATCC 33317
Streptococcus constellatus subsp. constellatus	ATCC 27823
Streptococcus cristatus	ATCC 51100
Streptococcus intermedius	ATCC 27335
Streptococcus gordonii	ATCC 33399
Streptococcus mutans	ATCC 25175
Streptococcus parasanguinis	ATCC 15912
Streptococcus pneumoniae	ATCC 6303
Streptococcus pyogenes	ATCC 19615
Streptococcus salivarius	ATCC 7073
Streptococcus sanguinis	ATCC 10556
Streptococcus suis	ATCC 43765

^an = 70.

Table 3.

Gram-negative bacteria used for specificity analysis^a

Species	Strain
Acinetobacter baumannii	ATCC 19606
Acinetobacter haemolyticus	ATCC 17906
Aeromonas caviae	CCUG 44411
Aeromonas hydrophila	ATCC 7966
Burkholderia cepacia	ATCC 25416
Citrobacter amalonaticus	ATCC 25405
Citrobacter braakii	ATCC 43162
Citrobacter farmeri	ATCC 51112
Citrobacter freundii	ATCC 6879
Citrobacter gillenii	ATCC 51117
Citrobacter koseri	ATCC 27156
Citrobacter murliniae	ATCC 51641
Citrobacter sedlakii	ATCC 51115
Citrobacter werkmanii	ATCC 51114
Citrobacter youngae	ATCC 29935
Enterobacter aerogenes	ATCC 13048
Enterobacter agglomerans	ATCC 27989
Enterobacter amnigenus	ATCC 33072
Enterobacter asburiae	ATCC 35953
Enterobacter cancerogenus	ATCC 33241
Enterobacter cloacae	ATCC 7256
Enterobacter dissolvens	ATCC 23373
Enterobacter gergoviae	ATCC 33028
Enterobacter hormaechei	ATCC 49162
Enterobacter intermedius	ATCC 33110
Enterobacter nimipressuralis	ATCC 9912
Enterobacter pyrinus	ATCC 49851
Escherichia blattae	ATCC 29907
Escherichia fergusonii	ATCC 35469
Escherichia hermannii	ATCC 33650
Escherichia vulneris	ATCC 33821
Haemophilus haemolyticus	ATCC 33390
Haemophilus influenzae	ATCC 9007
Haemophilus parahaemolyticus	ATCC 10014
Haemophilus parainfluenzae	ATCC 7901
Hafnia alvei	ATCC 13337
Klebsiella oxytoca	ATCC 13182
Klebsiella pneumoniae	ATCC 27736
Leclercia adecarboxylata	ATCC 29916
Legionella pneumophila subsp. fraseri	ATCC 33156
Moraxella atlantae	ATCC 29525
Moraxella catarrhalis	ATCC 25238
Neisseria caviae	ATCC 14659
Neisseria elongata subsp. elongata	ATCC 25295
Neisseria gonorrhoeae	ATCC 35201
Neisseria meningitidis	ATCC 13077
Neisseria mucosa	ATCC 19696
Pantoea agglomerans	ATCC 27155
Pasteurella aerogenes	ATCC 27883
Photorhabdus asymbiotica	ATCC 43948
Proteus mirabilis	ATCC 25933
Proteus vulgaris	ATCC 29513
Providencia alcalifaciens	ATCC 9886
Providencia rettgeri	ATCC 9250
Providencia rustigianii	ATCC 33673
Providencia stuartii	ATCC 33672
Pseudomonas aeruginosa	ATCC 27853
Pseudomonas alcaligenes	ATCC 14909
Pseudomonas fluorescens	ATCC 2219
Pseudomonas oryzihabitans	ATCC 43272
Pseudomonas putida	ATCC 12633
Pseudomonas stutzeri	ATCC 17588
Raoultella ornithinolytica	ATCC 31898
Raoultella planticola	ATCC 33531
Raoultella terrigena	ATCC 33257
Salmonella bongori	ATCC 43975
Salmonella enterica subsp. enterica Choleraesuis	ATCC 7001
Salmonella enterica subsp. enterica Enteritidis	ATCC 13076
Salmonella enterica subsp. enterica Gallinarum	ATCC 9184
Salmonella enterica subsp. enterica Heidelberg	ATCC 8326
Salmonella enterica subsp. enterica Paratyphi A	ATCC 9150
Salmonella enterica subsp. enterica Paratyphi B	ATCC 8759
Salmonella enterica subsp. enterica Pullorum	ATCC 9120
Salmonella enterica subsp. enterica Putten	ATCC 15787
Salmonella enterica subsp. enterica Typhi	ATCC 10749
Salmonella enterica subsp. enterica Typhi	ATCC 27870
Salmonella enterica subsp. enterica Typhimurium	ATCC 14028
Salmonella enterica subsp. enterica Virchow	ATCC 51955
Salmonella enterica subsp. houtenae	ATCC 43974
Salmonella enterica subsp. indica	ATCC 43976
Salmonella enterica subsp. salamae	ATCC 43972
Serratia entomophila	ATCC 43705
Serratia ficaria	ATCC 33105
Serratia fonticola	ATCC 29844
Serratia grimesii	ATCC 14460
Serratia liquefaciens	ATCC 25641
Serratia marcescens	ATCC 8100
Serratia odorifera	ATCC 33077
Serratia plymuthica	ATCC 183
Serratia proteamaculans subsp. proteamaculans	ATCC 19323
Serratia proteamaculans	ATCC 33765
Serratia rubidaea	ATCC 27593
Stenotrophomonas maltophilia	ATCC 13637
Tetragenococcus solitarius	ATCC 49428
Vibrio alginolyticus	CCRI-14794
Vibrio cholerae	ATCC 25870
Vibrio fluvialis	CCRI-14795
Vibrio parahaemolyticus	ATCC 17802
Vibrio vulnificus	ATCC 27562
Yersinia aldovae	ATCC 35236
Yersinia bercovieri	ATCC 43970
Yersinia enterocolitica subsp. enterocolitica	ATCC 9610
Yersinia frederiksenii	ATCC 29912
Yersinia intermedia	ATCC 29909
Yersinia kristensenii	ATCC 33638
Yersinia mollaretii	ATCC 43969
Yersinia pseudotuberculosis	ATCC 29833
Yersinia rohdei	ATCC 43380
Yersinia ruckeri	ATCC 29473

^an = 109.

The identification of all strains used in this study was reconfirmed with either an automated MicroScan Autoscan-4 system (Siemens Healthcare Diagnostics Inc., Newark, DE) or with Vitek 32 (bioMérieux SA, Marcy l'Étoile, France). The bacterial strains listed in Tables 1, 2, and 3 were first grown from frozen stocks, stored at −80°C in brain heart infusion (BHI) medium (BD, Mississauga, Ontario, Canada) containing 10% glycerol, and cultured on sheep blood agar (BD), chocolate agar (BD), or buffered charcoal yeast extract agar (BD), depending upon the specific growth requirement of each species. Bacterial strains were then grown to logarithmic phase (optical density at 600 nm [OD₆₀₀], 0.5 to 0.6) in BHI medium, and cultures were adjusted to a 0.5 McFarland standard (Thermo Fisher Scientific Company, Ottawa, Ontario, Canada) by dilution with phosphate-buffered saline (PBS).

To determine the ubiquity and analytical specificity of the MI agar method, an aliquot of a 10⁻⁵ dilution was used to spike spring water (Labrador; Anjou, Québec, Canada) to produce a 100-ml suspension containing between 20 and 80 CFU/100 ml that was filtered through a GN-6 membrane filter (47-mm diameter, 0.45-μm pore size; PALL Corporation, Mississauga, Ontario, Canada) on a 3-place standard manifold (Millipore Corporation, Billerica, MA). The filter was then incubated on MI agar for 24 ± 2 h at 35 ± 0.5°C to evaluate colony growth and color. Tests to confirm the sterility of the filter membranes and buffer used to rinse the filtration apparatus were also performed.

To determine the ubiquity and analytical specificity of rtPCR assays, standardized cell suspensions were lysed using the BD Diagnostics GeneOhm Rapid Lysis kit as recommended by the manufacturer (BD Diagnostics GeneOhm, Québec City, Québec, Canada). One microliter of a standardized lysed bacterial suspension was transferred directly to a 24-μl PCR mixture containing 50 mM KCl, 10 mM Tris-HCl (pH 9.0), 0.1% Triton X-100, 2.5 mM MgCl₂, 0.4 μM each primer, 0.2 μM probe, 200 μM each deoxyribonucleoside triphosphate (GE Healthcare Bio-Sciences Inc., Baie d'Urfé, Québec, Canada), 3.3 μg/μl bovine serum albumin (BSA; Sigma-Aldrich Canada Ltd., Oakville, Ontario, Canada), 0.06 μg/μl methoxsalen (Sigma-Aldrich Canada Ltd.), 0.025 U of Taq DNA polymerase (Promega, Madison, WI), and TaqStart antibody (Clontech Laboratories, Mountain View, CA). Decontamination of the PCR mixtures prior to rtPCR was achieved as described by Maheux et al. (23). In each experiment, 1 μl of sterile water was added to PCR mixtures as a negative control. The PCR mixtures were subjected to thermal cycling for 1 min at 95°C and then 35 cycles of 2 s at 95°C, 10 s at 58°C, and 20 s at 72°C for the E. coli/Shigella tuf primer-and-probe set and for 3 min at 95°C and then 35 cycles of 15 s at 95°C and 60 s at 60°C for the E. coli/Shigella uidA primer-and-probe set (10).

The sequences of the rtPCR primers and probes used in this study to evaluate the ubiquity and analytical specificity of the E. coli/Shigella-specific rtPCR assays are listed in Table 4. The uidA rtPCR primer (784F and 866R)-and-probe (EC807) set used for E. coli/Shigella detection was previously described by Frahm and Obst (10). The rtPCR primer set (TEcol553 and TEcol754) used to detect the E. coli/Shigella tuf gene was described by Maheux et al. (23), while the TEco573-T1-B1-specific dually labeled (TaqMan) rtPCR probe was developed by building an alignment of multiple tuf sequences retrieved from public databases with GCG programs (version 8.0; Accelrys, Madison, WI) (this work). The Oligo primer analysis software (version 5.0; National Biosciences, Plymouth, MN) was used to select candidate primer and probe sequences from the alignment. Oligonucleotide primers and dually labeled probes were synthesized by Integrated DNA Technologies (Coralville, IA).

Table 4.

rtPCR primers and probes used in this study

Assay, genetic target, and primer or probe	Sequence (5′→3′)	Reference
uidA
784F	GTGTGATATCTACCCGCTTCGC	10
866R	GAGAACGGTTTGTGGTTAATCAGGA
EC807	FAMa-TCGGCATCCGGTCAGTGGCAGT-BHQ-1b
tuf
TEcol553	TGGGAAGCGAAAATCCTG	23
TEcol754	CAGTACAGGTAGACTTCTG
TEco573-T1-B1	TETc-AACTGGCTGGCTTCCTGG-BHQ-1	This study
atpD
ABgl158	CACTTCATTTAGGCGACGATACT	27
ABgl345a	TTGTCTGTGAATCGGATCTTTCTC
Abgl-T1-A1	FAM-CGTCCCAATGTTACATTACCAA-CCGGCACT-(BHQ-1)-GAAATAGG

^aFAM, 6-carboxyfluorescein (fluorescence reporter dye).

^bBHQ-1, Black Hole Quencher 1 (fluorescence quencher dye).

^cTET, tetrachlorofluorescein (fluorescence reporter dye).

Comparison studies of the MI agar method and the CRENAME (concentration and recovery of microbial particles, extraction of nucleic acids, and molecular enrichment) procedure combined with an E. coli/Shigella-specific rtPCR assay. (i) Water sample preparation.

The bacterial strain used for spiking experiments was E. coli ATCC 11775. E. coli cells were grown to logarithmic phase (OD₆₀₀, 0.5 to 0.6) in BHI medium, and the culture was adjusted to a 0.5 McFarland standard (Thermo Fisher Scientific Company) before being serially diluted 10-fold in PBS (137 mM NaCl, 6.4 mM Na₂HPO₄, 2.7 mM KCl, 0.88 mM KH₂PO₄, pH 7.4). An aliquot of the 10⁻⁵ dilution was used to spike spring water (Labrador, Anjou, Québec, Canada) to produce suspensions containing approximately 100, 50, 25, 16, 8, 4, 2, and 1 CFU/100 ml. Bacterial counts were confirmed by filtering 100 ml of each spiked water sample through a GN-6 membrane filter (47-mm diameter, 0.45-μm pore size; PALL Corporation) on a 3-place standard manifold (Millipore Corporation, Billerica, MA). The filter was then incubated on sheep blood agar plates for 24 ± 2 h at 35.0 ± 0.5°C prior to the determination of colony counts. Tests to confirm the sterility of filter membranes and buffer used for rinsing the filtration apparatus were also performed.

To determine the ability of the CRENAME procedure coupled with the E. coli/Shigella-specific rtPCR assay (CRENAME + E. coli rtPCR assay) to detect E. coli cells in different potable water samples, 10 different well water samples were collected in the Québec City area during fall 2008 were spiked with diluted sewage (in PBS) to produce suspensions having approximately 100 CFU of E. coli/100 ml of water. Untreated sewage was harvested just before the entrance of the municipal treatment plant of St-Nicolas (Québec, Canada) and held at 4°C for a maximum of 2 days. Finally, to compare the limit of detection (LOD) of the MI agar method (31) and that of the CRENAME + E. coli rtPCR assay with real water samples, another well water sample was spiked with diluted sewage to produce suspensions having 100, 50 10, 5, and 1 of E. coli CFU/100 ml of water.

A process control consisting of approximately 60 Bacillus atrophaeus subsp. globigii (CCRI-9827 [equivalent to strain NRS 1221A]) spores/100 ml was added to all water samples prior to filtration. Spores were prepared as described by Picard et al. (27). B. atrophaeus subsp. globigii detection serves to monitor the integrity of the sample preparation, nucleic acid extraction, and molecular enrichment methods and to verify the absence of whole-genome amplification (WGA) and/or rtPCR inhibition.

(ii) Membrane filtration.

The membrane filtration step is used for both the MI agar and CRENAME + E. coli rtPCR assay procedures. For each spiked sample, two 100-ml aliquots of a 200-ml spiked sample were filtered on GN-6 membrane filters (47-mm diameter, 0.45-μm pore size; PALL Corporation) using a standard manifold (Millipore Corporation). One filter was used for the MI agar method, and the other was used for the CRENAME + E. coli rtPCR assay procedure.

For the MI agar method, the filter was placed onto solid MI medium (BD Diagnostic Systems, Sparks, MD) supplemented with 5 μg/ml cefsulodin (Sigma-Aldrich, St. Louis, MO) and incubated for 24 ± 2 h at 35 ± 0.5°C. After incubation, colony counts and color were recorded (31). Quality control of each batch of MI agar plates was conducted as recommended by the U.S. Environmental Protection Agency (EPA). In addition, filter, buffer, and rinse water blanks were included as sterility controls.

For the CRENAME + E. coli rtPCR assay procedure, the filter was aseptically transferred to a 15-ml polypropylene tube (Sarstedt) and treated by the CRENAME procedure as described below.

(iii) CRENAME procedure.

The CRENAME method is summarized in Fig. 1. Briefly, the CRENAME method is composed of (i) a method for the concentration and recovery of microbial particles, (ii) a nucleic acid extraction procedure, and (iii) a molecular enrichment by WGA. WGA is an isothermal procedure that amplifies the genomic DNA of microbial cells recovered during the previous concentration step (3, 18, 20).

Fig. 1.

Performance of U.S. EPA Method 1604 and that of the CRENAME + E. coli-specific rtPCR procedure for the detection of E. coli in potable water.

(iv) Concentration and recovery of E. coli cells followed by the nucleic acid extraction procedure.

The membrane aseptically transferred to the 15-ml polypropylene tube was exposed for 10 s to 8.5 ml of high-performance liquid chromatography grade methanol (Sigma-Aldrich) and vigorously agitated on a vortex mixer for 10 s. The reaction tube and its contents were then centrifuged for 3 min at 2,100 × g. The supernatant was removed and discarded, 1 ml of histological-grade acetone (EMD Chemicals, San Diego, CA) was added to the pellet, and complete dissolution was achieved by vigorous agitation on a vortex mixer. The resulting clear acetone solution was transferred to a 2-ml tube containing a mixture of sterile, acid-washed glass beads (150 to 212 μm and 710 to 1,180 μm; Sigma-Aldrich) and centrifuged for 3 min at 15,800 × g, and the supernatant was removed.

The 15-ml polypropylene tube was briefly rinsed with 1.0 ml of histological-grade acetone, and the resulting mixture was transferred to the glass bead tube previously used. The tube was then centrifuged for 3 min at 15,800 × g, and the supernatant was removed and discarded. The resulting pellet was washed with 1.0 ml of TE (Tris-HCl at 100 mM, EDTA at 1 mM, pH 8.0) and centrifuged for 3 min at 15,800 × g. After centrifugation of the washed filtrate-glass bead suspension in the presence of TE buffer, the supernatant was removed and discarded. The dead volume in the glass beads is approximately 25 μl. At this point, the tube containing the concentrated E. coli cells was treated to evaluate the recovery rate and efficiency of the concentration-and-recovery step only (see next paragraph) or submitted to molecular enrichment by WGA for the sensitive detection of E. coli cells contained in the 100-ml water sample (see WGA section below).

To evaluate the recovery rate and efficiency of the concentration-and-recovery step only, 15 μl of TE (100 mM Tris-HCl, 1 mM EDTA, pH 8.0) was added to the tube and lysis of the cells contained in the pellet was achieved by vigorous mixing at maximum speed on a vortex mixer for 5 min. The reaction tube containing the cell lysate was then incubated for 2 min at 95°C, briefly spun in a microcentrifuge, and kept at −20°C until rtPCR amplification. A 1-μl sample of the 40-μl final volume obtained after DNA extraction was then directly used to perform an E. coli-specific rtPCR.

(v) WGA.

Forty microliters of Illustra GenomiPhi V2 sample buffer (part of the Illustra GenomiPhi DNA amplification kit; GE Healthcare, Montréal, Québec, Canada) was added to the 25-μl reaction mixture. The cells contained in the pellet were mechanically lysed by vigorous mixing at maximum speed on a vortex mixer for 5 min at room temperature. The reaction tube containing the crude cell extract was then incubated for 3 min at 95°C and kept on ice for a minimum of 3 min. A mixture of 45 μl of GenomiPhi reaction buffer and 4 μl of φ29 DNA polymerase (GenomiPhi DNA amplification kit) was added to the extract and gently mixed by finger tapping before being briefly spun in a microcentrifuge. The WGA reaction mixture was incubated for 3 h at 30°C. The enzymatic reaction was then arrested by a 10-min incubation at 65°C. One microliter of WGA-amplified products was then used as the template for E. coli and B. atrophaeus subsp. globigii rtPCR amplification. To ensure that the tested water samples were free of E. coli, CRENAME + E. coli rtPCR negative controls were also performed using unspiked water.

(vi) rtPCR conditions used to compare U.S. EPA Method 1604 on MI agar and the CRENAME procedure combined with the E. coli/Shigella-specific rtPCR assay.

The sequences of the rtPCR primers and probes used to compare U.S. EPA Method 1604 on MI agar and the CRENAME + E. coli rtPCR assay are shown in Table 4. The rtPCR primer set (TEcol553 and TEcol754) and the TEco573-T1-B1 E. coli/Shigella-specific rtPCR dually labeled (TaqMan) probe were developed as described in the section on the ubiquity and analytical specificity of the MI agar and E. coli/Shigella-specific rtPCR assays. The rtPCR primers (ABgl158 and ABgl345a) and probe (ABgl-T1-B1) used for the B. atrophaeus subsp. globigii assay are described elsewhere (27).

One microliter of a standardized lysed bacterial suspension or of WGA amplification products was transferred directly to a 24-μl PCR mixture containing 50 mM KCl, 10 mM Tris-HCl (pH 9.0), 0.1% Triton X-100, 2.5 mM MgCl₂, 0.4 μM tuf E. coli/Shigella or B. atrophaeus subsp. globigii primers, 0.2 μM tuf E. coli/Shigella or B. atrophaeus subsp. globigii probe, 200 μM each deoxyribonucleoside triphosphate (GE Healthcare Bio-Sciences Inc.), 3.3 μg/μl BSA (Sigma-Aldrich Canada, Ltd.), 0.06 μg/μl methoxsalen (Sigma-Aldrich Canada, Ltd.), 0.025 U of Taq DNA polymerase (Promega), and TaqStart antibody (Clontech Laboratories). Decontamination of the PCR mixtures prior to rtPCR was achieved as described by Maheux et al. (23). In each experiment, 1 μl of sterile water was added to PCR mixtures as a negative control.

The PCR mixtures were subjected to thermal cycling for 1 min at 95°C and then 45 cycles of 2 s at 95°C, 10 s at 58°C, and 20 s at 72°C for the E. coli/Shigella tuf primer-and-probe set (this study); thermal cycling for 3 min at 95°C and then 45 cycles of 15 s at 95°C and 60 s at 60°C for the E. coli/Shigella uidA primer-and-probe set (10); or thermal cycling for 1 min at 95°C and then 45 cycles of 15 s at 95°C and 60 s at 60°C for the B. atrophaeus subsp. globigii primer-and-probe set in a Rotor-Gene thermal cycler (Qiagen Inc.).

Statistical analysis.

Logistic regression statistical analysis was done using the software JMP v8.0 (29a) and R (28).

RESULTS AND DISCUSSION

Ubiquity of the MI agar and E. coli/Shigella rtPCR assays.

The ubiquity of the MI agar assay and both E. coli/Shigella-specific rtPCR assays was assessed by testing genomic DNA isolated from 79 E. coli and 11 Shigella sp. strains of different serotypes and geographic origins (Table 1). Sixty-one (77.2%) of the 79 E. coli strains tested and 4 (36.4%) of the 11 Shigella strains tested yielded a β-glucuronidase-positive signal by the MI agar method. The uidA rtPCR assay result was positive for 68 (86.1%) of the 79 E. coli strains, while the tuf rtPCR assay yielded positive signals for all of the E. coli strains, leading to a ubiquity of 100%. Only the primer-and-probe set targeting tuf was able to detect all of the E. coli and Shigella strains tested. Indeed, 5 (45%) of the 11 Shigella strains used were not detected by the uidA primer-and-probe set.

Historically, E. coli and Shigella are considered a single genetic species but are classified in two different genera on the basis of biochemical and pathogenicity tests (4, 19, 26). Since no distinction between pathogenic and nonpathogenic strains is required to assess water quality in environmental microbiology, E. coli and Shigella can be considered the same genetic species and detection of their presence in water should indicate a high probability of fecal contamination. Consequently, the rtPCR assay targeting the tuf gene detects more E. coli/Shigella strains than does the uidA rtPCR assay tested in this study.

Analytical specificity of the MI agar and rtPCR assays.

The analytical specificity of the MI agar assay and both E. coli/Shigella-specific rtPCR assays was verified by testing a panel composed of 179 nontarget strains representing 70 species of Gram-positive and 109 species of Gram-negative bacteria that are frequently encountered in either clinical or environmental settings. The panel also includes bacteria phylogenetically closely related to E. coli/Shigella (Tables 2 and 3). All of these bacterial species tested negative with both the MI agar and uidA rtPCR assays. However, the tuf rtPCR assay amplified only DNA from Escherichia fergusonii (1/179), for an analytical specificity of 99.4%. E. fergusonii is genetically closely related to E. coli based on DNA-DNA hybridization (8) and phylogenetic analysis (26). Indeed, their respective tuf genes are 100% identical. To date, E. fergusonii was identified in human and bird feces (33), but its prevalence in water remains unknown.

LOD of the E. coli/Shigella rtPCR assays.

The LOD of the rtPCR assays was verified by using purified E. coli genomic DNA. Both rtPCR assays were able to detect as little as one purified genome copy (∼6 to 7 fg of genomic DNA) of E. coli per rtPCR (not shown).

Recovery rate and efficiency of the CRENAME concentration-and-recovery step.

The major advantage of the CRENAME concentration-and-recovery step lies in its efficiency in robustly recovering as little as 1 Gram-positive enterococcal cell from a 100-ml water sample (21). In this study, the efficiency of the CRENAME concentration-and-recovery step in recovering more fragile Gram-negative E. coli/Shigella cells was tested. First, tests were performed to determine the LOD and repeatability of the CRENAME concentration-and-recovery step without molecular enrichment. A preliminary experiment was performed to determine the concentrations at which E. coli cells were always detected. This experiment showed that E. coli cells were always detected at concentrations as low as 10 CFU/rtPCR. The same experiment was also performed with 10 different well water samples randomly collected from the region of Québec City during fall 2008. Recovery levels were similar to those obtained during the preliminary experiment. Following these experiments, tests were performed to determine both the LOD and the repeatability of the CRENAME concentration-and-recovery step. This was done by testing replicates at low levels (between 0.5 and 10 CFU/rtPCR; Table 5). Positive and negative controls performed as expected. The number of CFU incorporated into each rtPCR was estimated by dividing the number of CFU obtained on MI agar (in a membrane filtration experiment done with a paired sample) by 40 since only 1 μl of the extracted sample is tested by rtPCR. The LOD at 95% for E. coli calculated by logistic regression was 2.98 CFU/rtPCR (29 CFU/100 ml; P value of 0.0328).

Table 5.

Comparative recovery of E. coli cells by counting procedures and the membrane dissolution step (without WGA)

Target microbial counta	Avg E. coli CFU counta on MI agar ± SD	Estimated avg no. of E. coli CFU/rtPCR ± SDb	tuf rtPCR resultsc
80	65.0 ± 5.5	1.63 ± 0.16	−, +, +
80	63.0 ± 5.0	1.58 ± 0.16	+, +, +
40	31.0 ± 2.2	0.78 ± 0.06	−, −, +
40	29.0 ± 1.4	0.73 ± 0.04	−, +, +
20	14.3 ± 1.7	0.36 ± 0.04	−, −, −
20	14.0 ± 0.8	0.35 ± 0.02	−, −, −
0d	0	0	−, −, −

^aNumber of CFU/100 ml.

^bThe number of E. coli CFU/rtPCR is estimated to 1/40 of the average E. coli count on MI agar.

^cResults of three experiments are shown.

^dNegative control.

When an rtPCR assay is optimized, the LOD can be as little as a single copy of DNA. However, it is impossible, based on Poisson probability, to guarantee that a single copy can be delivered into a particular reaction tube (5). Bustin et al. (5) stated that the most sensitive LOD theoretically possible is 3 copies per PCR, assuming a Poisson distribution, a 95% chance of including at least 1 copy in the rtPCR, and single-copy detection. In this study, we always detected at least 3 microbial particles per μl after membrane dissolution. Thus, the loss of E. coli cells during this part of the procedure was negligible. This capacity may be explained by a combination of factors. First, the initial methanol step produces small membrane pieces that apparently create conditions that favor the confinement of microbial particles during centrifugation, limiting losses. Furthermore, the methanol step reduces the amount of acetone required to completely dissolve filtration membrane fragments. Thus, the remaining 1 ml of acetone can be easily transferred in a 2-ml microtube, where glass beads contribute to the efficiency of microbial particle recovery, acting as a secondary confinement matrix.

LOD of the CRENAME + E. coli rtPCR procedure.

Testing of 1/40 of the original water sample is insufficient for drinking water quality monitoring, and this is attributable mainly to the limitations imposed by the final volume obtained after DNA extraction, since only 1 or 2 μl is used for rtPCR. Since common DNA purification procedures are not highly efficient at recovering DNA at low concentrations, WGA was used to increase the amount of E. coli DNA to a level detectable by rtPCR.

Downstream of the concentration-and-recovery step, the CRENAME + E. coli-specific rtPCR assay was used to specifically detect E. coli cells in 100-ml potable water samples spiked with different concentrations of target bacteria. E. coli cells were always detected by molecular amplification at concentrations as low as 3.3 ± 1.3 CFU/100 ml, as confirmed by microbiological counts on MI agar. The LOD of the whole new procedure was estimated to be 1.8 CFU/100 ml, whereas for the same water samples, the LOD of the MI agar method was 1.2 CFU/100 ml (95% confidence; Table 6). Culture enrichment steps, requiring 8 to 16 h, are often used in molecular environmental microbiology to reach the LOD required to assess drinking water quality (9, 10, 30). The CRENAME + E. coli rtPCR procedure provides an alternate means to detect the presence of bacteria in drinking water samples and identify them in only 5 h, without prior culture enrichment. Finally, the chemical decontamination procedure performed on the reagents before adding DNA samples ensured that trace amounts of E. coli DNA commonly found in reagents (10, 15) were not amplified and do not contribute to false-positive results.

Table 6.

Comparative recovery and detection of E. coli by the MI agar and CRENAME + E. coli rtPCR procedures

Target E. coli count (CFU/100 ml)	Avg bacterial count ± SD (CFU/100 ml)	MI agar method resultsa	CRENAME + E. coli rtPCR assay resultsa
100	89.3 ± 7.2	+, +, +, ND,b ND, ND	+, +, +, ND, ND, ND
50	47.4 ± 5.8	+, +, +, ND, ND, ND	+, +, +, ND, ND, ND
25	22.3 ± 4.2	+, +, +, ND, ND, ND	+, +, +, ND, ND, ND
16	10.3 ± 5.0	+, +, +, ND, ND, ND	+, +, +, ND, ND, ND
8	5.0 ± 1.3	+, +, +, ND, ND, ND	+, +, +, ND, ND, ND
4	3.3 ± 1.3	+, +, +, +, +, +	+, +, +, +, +, +
2	1.5 ± 0.5	+, +, +, +, +, +	+, +, +, +, −, −
1	1.3 ± 0.9	+, +, +, +, +, −,	+, +, +, +, +, −
1	0.8 ± 0.4	+, +, +, +, +, −,	+, +, +, +, −, −
0c	0	−, ND, ND, ND, ND, ND	−, ND, ND, ND, ND, ND

^aResults (presence or absence for each replicate) of six experiments are shown.

^bND, not done.

^cNegative control.

To determine the ability of the CRENAME + E. coli rtPCR assay to detect E. coli cells in different natural potable water samples, 10 different well water samples harvested in the Québec City area during fall 2008 were spiked with diluted sewage to produce suspensions having approximately 100 CFU of E. coli/100 ml of water. All well water samples were subjected to CRENAME + E. coli rtPCR before and after spiking. All 10 well water samples tested negative before spiking, whereas they tested positive after spiking with sewage. As a process control, B. atrophaeus subsp. globigii was detected in all cases and the cycling thresholds for all E. coli/Shigella detection curves were quite similar (data not shown), indicating that the inhibitors present in the well water samples were (bio)chemically equivalent and/or not present at a sufficiently high concentration to inhibit WGA or rtPCR processes. Furthermore, the LODs of the MI agar and CRENAME + E. coli rtPCR methods were similar with both sewage- and E. coli-spiked samples. Indeed, the CRENAME + E. coli rtPCR assay result was positive when E. coli colonies grew on MI agar while it was negative when there was no growth on MI agar (Table 7). These preliminary results suggest that the overall LOD of the whole molecular microbiology procedure is equivalent to that of the MI agar method.

Table 7.

LOD of the MI agar method compared to that of the CRENAME + E. coli-rtPCR assay

Target E. coli counta	E. coli counta by MI agar assay	CRENAME + E. coli rtPCR resultb
100	112	+
50	46	+
10	6	+
5	3	+
1	0	−
0.5	0	−
0.1	0	−
0c	0	−

^aNumber of CFU/100 ml.

^bPresence or absence is shown for each bacterial titer.

^cUnspiked.

Since a PCR assay could also detect injured or dead cells, further comparison studies will help to determine the usefulness of a nucleic acid-based assay in drinking water analysis. However, a preliminary study in our laboratory showed that there is no rtPCR amplification in drinking water samples spiked with up to 0.05 ng (equivalent to 10⁴ genome copies) of purified genomic DNA of E. coli (data not shown). These data suggest that free DNA found in a drinking water sample flows through the filter during the filtration step, thus confirming that no free E. coli/Shigella DNA in water can be detected by the CRENAME + E. coli rtPCR procedure.

In this report, we demonstrate a good level of correlation between the presence/absence determination of E. coli/Shigella by WGA rtPCR and the microbiological counts obtained by the culture-based MI agar method. This preliminary study suggests that our molecular microbiology approach could be equivalent to conventional microbiology in terms of LOD and specificity but offers the advantage of decreasing the time to detection by approximately 19 h. Validation studies comprising a large number of natural samples are, however, needed to confirm our results and the water quality assessment procedure's acceptability to regulatory authorities.

Conclusion.

We coupled a rapid, simple, and efficient sample preparation method to a WGA procedure, rendering feasible the highly specific detection by rtPCR of 1.8 E. coli/Shigella CFU/100 ml (95% confidence) in less than 5 h without culture enrichment. Consequently, this procedure provides an easy way to concentrate and detect very low numbers of Gram-negative E. coli/Shigella cells present in 100-ml potable water samples. The entire molecular procedure (CRENAME + E. coli rtPCR assay) is comparable to EPA Method 1604 on MI agar in terms of analytical specificity and LOD but provides significant advantages in terms of speed and ubiquity.