A NOVEL TRIPLEX QPCR STRATEGY FOR QUANTIFICATION OF TOXIGENIC AND NON-TOXIGENIC VIBRIO CHOLERAE IN AQUATIC ENVIRONMENTS

A NOVEL TRIPLEX QPCR STRATEGY FOR QUANTIFICATION 1

OF TOXIGENIC AND NON-TOXIGENIC VIBRIO CHOLERAE 2

IN AQUATIC ENVIRONMENTS 3

4

Rupert Bliem1,4, Sonja Schauer1, Helga Plicka4, Adelheid Obwaller5, Regina Sommer1,3, Adolf 5

Steinrigl6, Munirul Alam7, Georg H. Reischer2,3, Andreas H. Farnleitner2,3 & Alexander 6

Kirschner1,3* 7

1 Institute for Hygiene and Applied Immunology, Water Hygiene, Medical University of Vienna, 8

Kinderspitalgasse 15, A-1090 Vienna, Austria 9

2 Institute of Chemical Engineering, Research Group Environmental Microbiology and Molecular 10

Ecology, Vienna University of Technology, Gumpendorferstraße 1a, A-1060 Vienna, Austria 11

3 Interuniversity Cooperation Centre Water and Health (ICC Water & Health), Vienna, Austria. 12

www.waterandhealth.at 13

4Armament and Defence Technology Agency, NBC & Environmental Protection Technology 14

Division, Rossauer Lände 1, A-1090 Vienna, Austria 15

5 Federal Ministry of Defence and Sports, Division of Science, Research and Development, 16

Rossauer Laende 1, 1090 Vienna, Austria 17

6 Institute for Veterinary Disease Control Mödling, Dept. for Molecular Biology, Austrian Agency 18

for Health and Food Safety (AGES), Robert Koch Gasse 17, 2340 Mödling, Austria 19

7 Enteric & Food Microbiology Laboratory, Centre for Food and Waterborne Diseases (CFWD) 20

International Centre for Diarrhoeal Disease Research, Bangladesh (ICDDR, B), Dhaka 1000 21

Bangladesh 22

23

*Corresponding author. Mailing address: Institute for Hygiene and Applied Immunology; Water 24

Hygiene; Medical University Vienna, Kinderspitalgasse 15, 1090 Vienna, Austria. Phone: +43-1-25

40160-33060. Fax: +43-1-40160-933000. E-mail: [email protected] 26

AEM Accepted Manuscript Posted Online 27 February 2015Appl. Environ. Microbiol. doi:10.1128/AEM.03516-14Copyright © 2015, American Society for Microbiology. All Rights Reserved.

Running title: Novel QPCR strategy for quantification of V.cholerae 27

Keywords: qPCR, quantification, Vibrio cholerae, environment 28

29

30

Abstract 31

Vibrio cholerae is a severe human pathogen and a frequent member of aquatic ecosystems. 32

Quantification of V.cholerae in environmental water samples is therefore fundamental for 33

ecological studies and health risk assessment. Beside time-consuming cultivation techniques, qPCR 34

has the potential to provide reliable quantitative data and offers the opportunity to quantify multiple 35

targets simultaneously. A novel triplex qPCR strategy was developed in order to simultaneously 36

quantify toxigenic and non-toxigenic V.cholerae in environmental water samples. To obtain quality-37

controlled PCR results, an internal amplification control was included. The qPCR assay was 38

specific, highly sensitive and quantitative across the tested 5-log dynamic range down to a method 39

detection limit of 5 copies per reaction. Repeatability and reproducibility were high for all three 40

tested target genes. For environmental application, global DNA recovery (GR) rates were assessed 41

for drinking water, river water and water from different lakes. GR rates ranged from 1.6% to 76.4% 42

and were dependent on the environmental background. Uncorrected and GR-corrected V.cholerae 43

abundances were determined in two lakes with extremely high turbidity. Uncorrected abundances 44

ranged from 4.6×102 to 2.3×104 cell equivalents L-1, whereas GR-corrected abundances ranged from 45

4.7×103 to 1.6×106 cell equivalents L-1. GR-corrected qPCR results were in good agreement with an 46

independent cell-based direct detection method, but were up to 1.6 log higher than cultivation based 47

abundances. We recommend the newly developed triplex qPCR strategy as a powerful tool to 48

simultaneously quantify toxigenic and non-toxigenic V.cholerae in various aquatic environments 49

for ecological studies as well as for risk assessment programs. 50

51

INTRODUCTION 52

Vibrio cholerae is a waterborne bacterium found worldwide in brackish water, coastal areas and 53

estuarine environments (1). Although the species V. cholerae comprises more than 200 serotypes, 54

only serotypes O1 and O139 are currently able to cause epidemic and pandemic cholera outbreaks 55

with more than 100.000 reported death cases per year (2). All other non-O1/non-O139 serotypes are 56

usually associated with less-severe gastrointestinal, blood, wound and ear infections (3). V. cholerae 57

has been classified as a potential category B terrorism agent by the Centers for Disease Control and 58

Prevention, USA (4). Detection and quantification of V. cholerae, especially of serogroup O1/O139 59

strains in environmental samples is still a difficult task and no international standard is available. 60

Currently accepted cultivation based methods take at least 48h to obtain a final result (5) and 61

severely underestimate the presence of O1/O139 serogroups, due to the fact that these strains 62

predominantly enter a viable but non-culturable (VBNC) state once released from the human 63

intestine into the environment (6, 7). Alternatively, fluorescence in situ hybridization (FISH) and 64

the direct fluorescent antibody assay (DFA) are direct cell-based detection methods that include 65

VBNC bacteria. DFA is a monoclonal antibody staining application that can only be used for 66

serotype O1 (8). FISH based on 16S rRNA-targeted oligonucleotide probes cannot discriminate 67

between toxigenic and non-toxigenic V. cholerae genotypes and a separation between V. mimicus 68

and V. cholerae is not possible (9). New molecular biological methods based on quantitative real-69

time PCR (qPCR) technology have the ability to provide bacterial genome equivalent estimates in 70

as little as 3 hours (10, 11). For V. cholerae several qPCR protocols have been developed in the past 71

decade, but all of them have limitations when applied for the quantification of toxigenic and non-72

toxigenic V. cholerae in environmental water samples (4, 12-16). Either the published qPCR 73

methods did not perform as a multiplex qPCR (which is necessary to distinguish between toxigenic 74

and non-toxigenic V. cholerae in one assay), they lacked an internal amplification control, no 75

standards for quantification were used, or the method could not separate between toxigenic and 76

non-toxigenic V. cholerae (4, 12-16). Moreover, the application of qPCR to environmental samples 77

has often been limited by high concentrations of substances with negative effects on DNA 78

extraction or PCR efficiency, such as salts or humic substances (17) (18). 79

The aim of the current work was to establish a two-level multiplex qPCR strategy to quantify 80

toxigenic and non-toxigenic V. cholerae in environmental water samples, even with high 81

concentrations of PCR-inhibiting substances. At the first level, a quality-controlled triplex TaqMan 82

qPCR was developed based on the three targets ompW for the presence of total V. cholerae, ctxA 83

for toxigenic V. cholerae and egfp (enhanced green fluorescent protein) as an internal amplification 84

control. At the second level, global DNA recovery (GR) rates were determined for water samples 85

from a range of ecologically different aquatic environments. By comparing the qPCR results to 86

those of two independent methods, a cell-based direct detection and a cultivation-based approach, it 87

was possible to assess the accuracy of our newly developed triplex qPCR strategy. 88

89

MATERIALS AND METHODS 90

Development of the qPCR assay 91 92 Signature genes and internal amplification control. Selection of the promising signature genes 93

ompW (outer membrane protein W) and ctxA (cholera toxin A) for the specific detection of non-94

toxigenic and toxigenic V. cholerae was based on published data (14). New primers, ompW-out and 95

ctx-out (Thermo Scientific, Ulm, Germany) were designed to provide additional sequence 96

information (Table 1). They were positioned > 150 bp up- and downstream, respectively, of the 97

published primer binding sites for these target genes (14, 16) to increase the number of possible 98

primer and probe binding positions for the hydrolysis assay. SYBR-Green based qPCR was 99

performed with a LightCycler 480II System (Roche Applied Science, Meylan, France) under 100

conditions described in the Supplemental Information. Each PCR product was sequenced in both 101

directions utilizing capillary electrophoresis on the 3130xl Genetic Analyser (Applied Biosystems) 102

as described in the Supplemental Information. Sequences were deposited at GenBank under 103

accession numbers KJ558403-KJ558422. Resulting sequences were added to the sequences 104

retrieved from the public database NCBI/EMBL, aligned and compared using Bionumerics 105

(Applied Maths, Ghent, Belgium). Primer Express software (Applied Biosystems) was used to 106

design hydrolysis minor groove binding (MGB) probes and primers targeting regions identified to 107

be specific for ompW and ctxA genes (Table 1). In silico amplicons were checked with nucleic acid 108

folding software mfold (RNA Institute, College of Arts and Sciences, State University of New York 109

at Albany [http://mfold.rna.albany.edu/?q=mfold]). 110

The egfp gene was selected as internal amplification control (IC), as proposed by Hoffmann et al 111

(19). Using the primers egfp-12-F and egfp-10-R (19) and intype IC-DNA (Labordiagnostik GmbH, 112

Leipzig, Germany) as template, a 374bp fragment of the egfp gene was amplified by PCR. The 113

specific primers and probe for amplification of the IC used here are listed in Table 1. The PCR 114

products were analysed by electrophoresis on a 1.5% agarose gel and DNA was purified with the 115

QIAquick® Gel Extraction kit (Qiagen). Amplicons were then inserted into the pCR® II standard 116

vector (Invitrogen, Lofer, Austria) and cloned into E. coli One Shot® Top10 competent cells 117

(Invitrogen), according to the manufacturer´s protocol. The insert sequences were checked by 118

sequencing with M13 standard primers. Sequencing was performed as described in the 119

Supplemental Information. Plasmids were purified with the plasmid mini kit (Qiagen) according to 120

the manufacturer´s instructions. Plasmid concentrations were determined by QubitTM Fluorometer 121

(Invitrogen). Finally, plasmids were serially diluted in nuclease-free water containing yeast carrier 122

tRNA (30 ng µl-1), aliquoted and stored at -20°C. 123

Triplex qPCR amplification. All qPCR reactions were carried out in a final reaction volume of 20 124

µl containing 10 µl 2× QuantiTect Multiplex PCR NoROX Master Mix (Qiagen), 2 µl of each 125

primer-probe mixture, 2 µl of DNA template including 250 copies of the IC plasmid, and PCR-126

grade water. The primer-probe mixture contained 200 nM primer concentrations for ctxA and IC, 127

150 nM of ompW-F1 and 50 nM of ompW-F2 (degenerate forward primer set), 200 nM of ompW-128

R, as well as the hydrolysis probes for IC at a concentration of 100 nM (Thermo Scientific, Ulm, 129

Germany), and for ompW and ctxA at a concentration of 200 nM (Applied Biosystems). The triplex 130

qPCR assay was designed for an optimal annealing temperature of 60°C and the cycling conditions 131

were as follows; hot start enzyme activation for 15 min, followed by 45 cycles at 94°C for 15 sec 132

and 60°C for 30 sec. Each triplex qPCR experiment included a standard dilution series of the ctxA, 133

ompW and egfp-plasmid, a positive control containing all three targets and a negative (no template) 134

control. Measurements and analyses were performed on a LightCycler 480 II software release 1.5 135

(Roche). Quantification cycle (Cq) values were calculated based on the curve fit method and a Cq 136

cut-off level of 39 was defined for both ctxA and ompW. Colour compensation was performed 137

according to the manufacturer’s guidelines. 138

Performance characteristics of the triplex qPCR. Specificity and sensitivity of the multiplex 139

qPCR assay were evaluated by using plasmids containing the desired gene targets and genomic 140

DNA from bacterial strains listed in the Supplemental Information (Table S1). In total, 19 141

V.cholerae O1/O139, 23 V. cholerae nonO1/nonO139, 29 non-target Vibrio spp. and 38 non-target 142

bacteria were tested. All V. cholerae were positive for ompW, only V.cholerae O1/O139 strains 143

were positive for ctxA. All other bacteria were negative for both targets. The method detection limit 144

(MDL) was calculated by using known concentrations of purified genomic DNA from V. cholerae 145

O395 (measured by using a QubitTM Fluorometer, Invitrogen), and the IC plasmid DNA. Ten-fold 146

to two-fold serial dilutions of the template were performed to calculate the MDLs from the 147

proportion of positive qPCRs at each dilution. Five independent experiments, each with four 148

replicates per dilution step were measured by qPCR. According to the MIQE guidelines (20) the 149

MDL was defined as the lowest DNA target concentration that could be detected with a probability 150

of ≥ 95%. The amplification efficiency was calculated from the log-linear portion of the standard 151

curve, covering 6 orders of magnitude. Repeatability (intra-assay variance, IAV) and reproducibility 152

(inter-assay variance, IEV) were calculated from the standard deviations (SD) and coefficients of 153

variation (CV) of the Cq-values and calculated concentrations from four replicate serial dilutions 154

consisting of ten dilution steps, each measured in triplicate. The triplex qPCR assay was also tested 155

for potential limiting effects (i.e. preferential amplification of one template over the others) under 156

various template mixing ratios (template matrix assay). Eight different mixing ratios with plasmid 157

concentrations of 1 × 102 and 1 × 104 copies per reaction diluted with yeast carrier tRNA (30 ng µl-158

1) were carried out in triplicate (Supplemental Information; Table S2). To check whether the qPCR 159

also provides reliable results for ctx positive V. cholerae when high abundances of non-toxic 160

variants are present, an additional matrix assay was run (Supplemental Information; Table S3). 161

162

Spiking experiments and recovery rates 163

Known numbers of V. cholerae O395 (source and growth conditions listed in Table S1, 164

Supplemental Information) were seeded into environmental water samples, collected from a 165

drinking water distribution system, the River Danube and the Lake Neusiedler See (Austria). The 166

samples had been collected with sterile 2 L glass-bottles and transported to the laboratory at in situ 167

temperature (±2 °C) in a cooling box within 3 h. The exact numbers of the added V. cholerae cell 168

suspensions, prepared in 10% (final conc.) of ethanol, were determined via epifluorescence 169

microscopy and solid phase cytometry as described elsewhere (9). Two-ml aliquots (cell numbers of 170

5.4 × 106 ml-1) were stored at -80°C to act as standard for cell recovery estimations. Environmental 171

water samples (see above) were aliquoted into two 200 ml centrifuge tubes. Into one tube, the V. 172

cholerae O395 standard was added. 173

DNA extraction. DNA extraction was performed using a modified version (preventing the loss of 174

sample material during the extraction process) of the procedure supplied with the MO BIO 175

PowerSoil® DNA Extraction Kit (MO BIO, Carlsbad, CA, USA) (21, 22). Environmental water 176

samples were centrifuged at 3,220 × g for 30 min, the supernatant was carefully decanted and the 177

pellet was resuspended in sterile tap water and transferred into 2 ml tubes (21). After a further 178

centrifugation step at 4,000 × g for 15 min, the resulting pellet was resuspended with 550 µl 179

PowerLyzerTM PowerSoil® bead solution and 0.3 g (150-212 µm) acid-washed glass beads (Sigma, 180

St. Louis, MO, USA) were added. The complete tube contents were processed in a Retsch MM301 181

mixer-mill (Retsch, Düsseldorf, Germany) at 30 Hz for four min. Eluted DNA was stored at -20°C 182

until multiplex qPCR analysis was performed. To determine the DNA binding capacity of the used 183

DNA extraction kit a range of ~2 × 102 to ~2 × 106 V. cholerae cells (exact numbers determined by 184

microscopic counting, see above) was tested. The stock suspension (5.4 × 106 cells) of type strain V. 185

cholerae O395 was ten-fold serially diluted in sterile autoclaved tap water and extracted with 186

MOBIO PowerSoil® DNA Extraction Kit in triplicate. Multiplex qPCR was performed as 187

described above, applying triplicate measurements for each sample. 188

Global recovery rate. Global recovery (GR) was defined as the efficiency of the cell concentration 189

and DNA extraction procedure to recover bacteria spiked into environmental water samples (23, 190

24). To test the reproducibility of the GR rates three independent samples from Lake Neusiedler See 191

were spiked with 5.4 × 105 cells into 200 ml of lake water. From the obtained Cq values mean 192

abundances, standard deviations and recovery rates (as percentages of the spiked cell number) were 193

calculated. Background abundances in non-spiked parallel samples were subtracted. In further 194

experiments, the new triplex qPCR method was also tested with unfiltered drinking water, river 195

water from the Danube, lake water from the Neusiedler See and some surrounding shallow saline 196

lakes to check for an influence of environmental conditions on GR. Seven sampling sites with 197

different physico-chemical conditions were selected. Measured parameters were the concentration 198

of dissolved organic carbon (DOC), the photometric ratio of the DOC measured at 365 nm and 254 199

nm as an indicator for the concentration of humic substances, the electrical conductivity and the 200

concentration of total suspended solids. These parameters were determined with methods used in an 201

earlier publication (9). 202

Environmental application 203

In order to determine V. cholerae abundances in environmental samples, the measured qPCR results 204

(in genomic units per sample volume) were used with and without correction with the respective 205

GR rate, determined for each sample. The obtained values were compared to two other independent 206

methods (solid phase cytometry and cultivation, see below) applied to the same samples. Water 207

samples were taken as described above from the Austrian lake Neusiedler See and one shallow 208

saline lake situated along the eastern shore of the lake (Zicklacke). Permanent autochthonous 209

occurrence of non-O1/non-O139 V. cholerae strains had been reported in these lakes that are known 210

for their high turbidity, high concentrations of DOC and humic substances (25, 26). For solid phase 211

cytometry, 3 to 8 replicate subsamples of appropriate volume (between 10 and 100 µl, depending on 212

the turbidity of the water) were taken and filled up with 1×PBS to reach a final volume of 1 ml and 213

processed as described in Schauer et al. (9). Cultivation-based quantification of V. cholerae was 214

done by membrane filtration as described in Schauer et al. (9). Briefly, appropriate volumes (10 ml 215

down to 10 µl, filled up to 10 ml with sterile 1×PBS for an even distribution on the filter) were 216

filtered through 0.45µm-pore-size cellulose nitrate filters and the filter directly placed on TCBS agar 217

plates (Merck, Darmstadt, Germany). Plates were incubated for 18 h at 37°C. Yellow, flat, 1- to 3-218

mm diameter colonies were counted and streaked onto plates with nutrient agar without NaCl. After 219

incubation overnight at 37°C, colonies growing on agar without NaCl were considered presumptive 220

V. cholerae. Representative isolates were confirmed by species-specific ompW-based PCR (5). 221

222

RESULTS 223

Performance characteristics of the triplex qPCR 224

Accuracy and method detection limit. The newly developed triplex qPCR exhibits a high 225

accuracy and low method detection limit for all three targets (Table 2). The mean slope of all 226

standard curves was between -3.15 and -3.35 with coefficients of determination (r2) of > 0.999 227

(calculated by the ROCHE LC480 software); Intra-assay variance was less than 0.3 Cq, inter-assay 228

variance was less than 11 %. The experimentally determined MDL was 5 copies for all targets, the 229

dynamic range span from 5 to 1 × 105 copies. The results from the template matrix assay showed 230

that a mutual influence of target amplification leading to false results occurred only at unrealistic 231

scenarios (see Supplemental Information, Table S2). In addition, toxigenic V. cholerae can be 232

reliably quantified even in the presence of much higher numbers of non-toxigenic variants 233

(Supplemental Information, Table S3). 234

Strain specificity and sensitivity. From the whole panel of 67 tested non-target organisms only 235

one showed a weak ctxA false-positive signal (V. fluvialis, Cq 42.93). In addition, one tested non-236

O1/non-O139 target strain showed a weak false-positive ctxA signal (V. cholerae, Cq 40.3). 237

However, both signals were beyond the ctxA qPCR cut-off level of Cq = 39, and can thus be 238

considered as negative. All other tested target and non-target strains gave the expected signals 239

(Supplemental Information, Table S1). 240

Efficacy of the DNA extraction procedure. In order to evaluate the range of concentrations in 241

which the qPCR results are truly quantitative, DNA binding capacity tests were done. These tests 242

showed a very high concordance between the number of spiked cells and the qPCR results for both 243

the ompW and the ctxA target over the whole tested concentration range (~2 × 102 to ~2 × 106) (Fig. 244

1). This shows that the used DNA extraction procedure can be reliably used for quantifying V. 245

cholerae numbers from 2 × 102 to at least 2 × 106 cells per sample volume. 246

In order to detect possible influences of high ctxA and ompW target concentrations on detection of 247

the IC-target, the detection rate of the latter was also evaluated: the detection rate of the IC target 248

was close to 100%, with the exception of the highest ctxA and ompW concentration. This case of 249

reduced amplification of the IC may be caused by the very high ratio between the ompW and ctxA 250

targets to the IC plasmid DNA that was added to the DNA extract at a concentration of only 250 251

plasmids per reaction. 252

Determination of global recovery rates (GR). Global recovery rates were evaluated for a variety 253

of different environmental conditions (Fig. 2). Spiking experiments with non-chlorinated drinking 254

water, river water and lake water were performed to check for the influence of different 255

environmental parameters like DOC, concentration of humic substances (calculated from the 365 256

nm/254 nm photometric ratio of DOC and the DOC concentration), electrical conductivity, and total 257

suspended solids) on the GR. The DOC concentration had the most negative influence on the GR 258

(drinking water 76.4% - shallow soda lake Oberer Stinker 1.6%; Spearman Rank correlation 259

coefficient rho = -0.93; p < 0.01; n = 7) as shown in the main chart of Fig. 2. For all other 260

parameters, a decreasing GR was observed with increasing values of the tested parameter, too, but 261

statistical relationships were partly insignificant (see Supplemental Information, Fig. S1). For the 262

lake Neusiedler See, the main target of our future investigations, GR rates were determined for 263

three representative sampling sites in more detail (Fig. 2, small insert). Mean GR rates for the 264

ompW target ranged from 15.8 to 26.7% with low standard deviations of ≤ 4.1 %. GR rates for the 265

ctxA target ranged from 17.2% to 28.9% with SD ≤ 1.9%). 266

267

Environmental testing 268

Finally, the V. cholerae abundances in selected environmental samples were quantified by the 269

newly developed qPCR in comparison to two independent non-qPCR based methods. The global 270

recovery rate was assessed for each sample to correct for the extraction efficiency of the DNA. 271

Environmental water samples were collected from two stations of the lake Neusiedler See as well as 272

from one shallow saline lake at various time points and processed as described above. To detect 273

potential inhibition, the original DNA extract and a 2-fold dilution were measured. None of the 274

diluted samples showed an increased IC value in comparison to the undiluted sample and thus 275

inhibition was ruled out. Measured qPCR concentrations (ompW target only, all samples showed 276

negative ctxA-qPCR results) are reported uncorrected (without correction for the GR rate) and 277

corrected for the GR rate, individually determined for each sample from a spiked sample aliquot. 278

Uncorrected and GR-corrected results were compared to cultivation and SPC/CARD-FISH data 279

(Fig. 3). V. cholerae numbers determined by cultivation ranged from 3.7×103 to 1.4×104 CFU L-1 in 280

the lake Neusiedler See and from 2.0×104 to 3.8×104 CFU L-1 in the Zicklacke. The numbers of V. 281

cholerae cells enumerated by SPC/CARD-FISH were consistently higher than cultivation-based 282

results and ranged from 1.2×104 to 8.3×104 cells L-1 for the lake Neusiedler See and from 4.8×105 283

to 9.0×105 cells L-1 for the Zicklacke. GR-uncorrected qPCR results ranged from 4.6×102 to 284

7.4×103 cell equivalents L-1 in the Neusiedler See and from 5.6×103 to 2.3×104 cell equivalents L-1 285

in the Zicklacke. With the exception of sample 5C, uncorrected qPCR results were always lower 286

than results obtained by cultivation or by the SPC/CARD-FISH procedure. Nevertheless, a 287

significant correlation between log-transformed uncorrected qPCR data and SPC/CARD-FISH data 288

was observed (r = 0.72, p < 0.05). GR-corrected qPCR results ranged from 4.7×103 to 2.0×105 cell 289

equivalents L-1 in the Neusiedler See and from 1.8×105 to 1.6×106 cell equivalents L-1 in the 290

Zicklacke. With the exception of samples 5A and 5B, where similar values were determined, GR-291

corrected qPCR results were significantly higher (p < 0.05) than results obtained by cultivation. For 292

these samples, the ratio of qPCR to cultivation ranged from 3.2 (sample 36C) to 43 (sample ZLB). 293

In contrast, the ratio of the GR-corrected qPCR results to the SPC/CARD-FISH results ranged from 294

0.4 (sample 5B) to 5.1 (sample 36B), with a median of 0.96, indicating that these two methods 295

provide on average comparable results. Moreover, log-transformed data obtained by the two 296

methods were significantly inter-correlated (r = 0.80, p < 0.02). 297

298

DISCUSSION 299

Targets of the qPCR. The key objective of this study was to develop a qPCR with an in-built 300

quality control for the simultaneous detection and quantification of toxigenic and non-toxigenic V. 301

cholerae in water samples from various environmental conditions. Two V. cholerae specific genes, 302

ompW for total V. cholerae and ctxA for toxigenic V.cholerae were chosen for the multiplex qPCR 303

assay. In contrast to other published assays (4, 15, 27), an internal amplification control (a plasmid 304

harbouring a partial sequence of the egfp gene) was additionally included in our assay to avoid 305

false-negative results due to PCR-inhibition. A huge amount of different targets are currently used 306

for the detection of V. cholerae by various qPCR assays, like hlyA, zot, ompU, ompW, ctxA, rtxA 307

and toxR (14, 28). It has been revealed that a variety of closely related bacteria possess sequences 308

similar to V. cholerae, invalidating some of the used targets for specific detection (29). In addition, 309

naturally occurring gene transfer based on transduction and transformation leads to an on-going 310

horizontal shift of gene variants (30, 31). To ensure the specificity and sensitivity of the selected 311

target gene ompW, 29 closely related Vibrio strains belonging to 6 different species and 22 312

additional V. cholerae isolates from the lake Neusiedler See were analysed (Supplemental 313

Information, Table S1) and the newly developed primers and probes were targeted against 314

appropriate sequence sections. Goel et al. proposed the target gene ctxAB as an indicator for 315

toxigenic V. cholerae (14); however, other groups found the presence of ctxAB in the closely related 316

V. mimicus (31-33). Also, a positive detection of the ctxAB gene may originate from the presence of 317

the CTXΦ phages in the water column (34). In any case, the detection and quantification of the ctxA 318

gene in other “hosts” than V. cholerae is relevant for risk assessment and environmental monitoring 319

because all of them may cause cholera-like diseases or act as an environmental reservoir (5, 31). 320

Quality parameters of the qPCR. For many published V. cholerae qPCR assays the authors did 321

not describe the respective detection limits (35-37); others stated that it ranged from 2.6 CFU up to 322

1000 CFU (12, 15, 38). In most publications the reported detection limit is simply referring to the 323

method detection limit (MDL), i.e. the lowest number of DNA targets that can be detected at a 324

probability level of 95% (20). However, for any environmental application, the sample limit of 325

detection (SLOD) should be reported, which describes the lowest quantity of the target DNA that 326

can be reliably detected in a certain volume of sample (39) The SLOD includes sources of bias 327

originating from sample processing and from the distribution of the target in the sample (i.e. 328

Poisson distribution at low abundances). It is also necessary to determine the recovery rates of the 329

used method to be able to reliably quantify absolute V. cholerae numbers in different environmental 330

matrices. The MDL of the novel triplex qPCR was 5 plasmid copies per reaction for all 3 target 331

genes; results lower than 3 copies are not reproducible because of the Poisson distribution of the 332

targets in solution (20). The SLOD (in cell equivalents) was calculated considering the MDL, the 333

sample volume (SV), the volume of the eluate of the DNA extraction kit (100 µl), the volume of the 334

DNA extract added to the PCR reaction (2 µl) and the measured GR rate according to the formula: 335

SLOD = ((MDL * (100µl / 2µl) * GR-1) * SV)-1 336

SLOD values found in this study thus differed largely between environments, ranging from 330 cell 337

equivalents per 200 ml in drinking water (GR = 0.76) to 1,000 per 200 ml cell equivalents in lake 338

water (GR = 0.25) and 15,500 cell equivalents per 200 ml in soda lakes with a highly interfering 339

background matrix (GR = 0.016). If a sample volume less than 200 ml is processed for turbid 340

surface waters (e.g. when filtration is used as concentration step instead of centrifugation) higher 341

SLOD values would be achieved; for drinking water lower SLOD values would be possible, when 342

larger sample volumes can be concentrated. 343

A primer probe matrix assay was performed to determine the optimal primer and probe 344

concentrations for the triplex qPCR (data not shown). No interference was observed. Intra-assay 345

variance (IAV) and inter-assay variance (IEV) were also tested, as recommended by the MIQE 346

guidelines (20). The low variability of IAV (< 0.29 Cq) strongly underlines the robustness and 347

precision of the assay, an IEV of < 11% clearly demonstrates the possible application for other 348

laboratories with the potential of comparing their own results with results of other groups. Our 349

qPCR was tested against a panel of various target concentrations (template matrix assay) to 350

visualize problematic cases, which may lead to a false negative result. A moderate to strong 351

influence on the Cq at low ompW concentrations (1 × 102 copies per reaction) was recognized only 352

in the presence of high concentrations of ctxA and IC resulting in delayed Cq values. This shows 353

that the newly developed qPCR is only biased in the rather unrealistic/rare situation of a higher ctxA 354

than ompW concentration (i.e. when the ctxA gene is present at high concentrations in a “host” other 355

than V. cholerae). 356

Applicability for environmental samples. When developing a triplex qPCR assay for 357

quantification of specific targets in various environmental matrices, one has to be aware of (i) 358

potential inhibitory effects in the PCR and (ii) how much sample material gets lost through the 359

sample preparation process. First, with the implementation of an internal amplification control we 360

could prove the reliability of the PCR in a wide range of environmental samples. Second, the DNA 361

recovery efficiency of the used extraction method must be checked. The DNA extraction procedure 362

used in this study was shown to provide correct quantitative results over a large range of cell 363

numbers (1×106 to 1×102 cells). However, a loss of sample DNA may occur during the 364

concentration and extraction procedure. Concentration of water samples by filtration may be more 365

effective and less damaging to V. cholerae cells. Recovery rates obtained via filtration ranged from 366

51% to more than 90% (23, 24), compared to the rates obtained by centrifugation (this study), with 367

a wide range from 1.6% to 76.4%. However, because of the high amount of suspended solids in 368

some of the investigated environmental waters the use of a filter was impossible, because it was 369

blocked after a few millilitres of water. Assuming constant global recovery rates for all samples 370

from a specific environment is probably an invalid approach to extrapolate qPCR results to 371

corrected cell numbers because of the great variability of GR rates that can be observed in 372

environmental samples even from the same body of water, as shown in this study. In contrast, by 373

determining global recovery rates for each sample from spiked parallel samples, a more realistic 374

calculation of the actual V. cholerae abundance in the environment should be possible. In our 375

experiments, special attention was paid to a variety of parameters responsible for potential influence 376

on DNA recovery or PCR inhibition (40). Along a turbidity gradient ranging from drinking water to 377

extremely turbid soda lake water, global recovery was most strongly impacted by the amount of 378

DOC. Most likely, co-extracted charged organic compounds such as humic acids present in high 379

concentrations in the investigated lakes compete with nucleic acids for silica-binding sites, causing 380

many of the nucleic acids to pass (41). Another possible mechanism for apparent losses of target 381

DNA could be the formation of complexes between DNA and water sample constituents during 382

extraction that render the DNA targets unavailable or non-amplifiable and thus non-detectable by 383

qPCR analysis (24). 384

Despite these restrictions, the newly developed triplex qPCR allowed us to quantify V. cholerae in 385

eight environmental water samples from the lake Neusiedler See and a selected shallow soda lake. 386

Both, uncorrected and GR-corrected qPCR results were significantly correlated with abundances 387

determined by an independent direct detection assay based on CARD-FISH in combination with 388

solid phase cytometry (9) in the same samples. However, GR-corrected qPCR values showed higher 389

correlation coefficients and were of comparable magnitude as the CARD-FISH/SPC approach. This 390

strongly supports our assumption that using the GR rate as a multiplication factor to estimate the 391

true V. cholerae cell abundance is acceptable over the whole range of GR results. Even at observed 392

low GR values ranging from ~2% to ~14%, the corrected V. cholerae numbers were on average in 393

good agreement with the CARD-FISH/SPC results. Significantly lower numbers were mostly 394

obtained with the cultivation method, which is in accordance with previous results (presence of 395

VBNC or dead cells, one CFU does not originate from one cell) as discussed in detail in Schauer et 396

al (36). No ctxA positive results were obtained, i.e. no toxigenic V. cholerae were detected in these 397

samples. However, as the qPCR performance characteristics for the ctxA target were the same as for 398

the ompW target, a successful quantification of toxigenic V. cholerae should be simultaneously 399

possible in water samples from respective geographic regions. 400

We can thus recommend the presented strategy for ecological and quantitative microbial risk 401

assessment studies. Both the quality-controlled uncorrected qPCR data and the GR-corrected data 402

may be used to feed a recently developed software tool for quantitative microbial risk assessment 403

from surface water to potable water (42). 404

405

CONCLUSIONS 406

Our newly designed qPCR strategy is a powerful tool for rapid and specific simultaneous 407

quantification of toxigenic and non-toxigenic V. cholerae cells in environmental samples. By the 408

integration of (1) an internal amplification control in the multiplex assay to verify the absence of 409

PCR-inhibition and (2) the calculation of a global recovery rate based on a standard addition to each 410

sample to correct for DNA extraction efficiency, the V. cholerae abundance can be determined even 411

in water samples with complex environmental matrix. This approach can be used for ecological and 412

microbial risk assessment studies. 413

414

ACKNOWLEDGMENTS 415

We thank Christopher Grim (University of Maryland) and Pavol Bakoss (Comenius University 416

Bratislava) for providing reference strains. The study was financed by the Austrian Science Fund 417

(FWF, project nr: P21625-B20). Dr. Alam of icddr,b has significantly contributed to this study. 418

icddr,b is thankful to the governments of Australia, Bangladesh, Canada, Sweden and the UK for 419

providing core/unrestricted support. This study is a joint publication of the Interuniversity 420

Cooperation Centre “ICC Water and Health” (www.waterandhealth.at) funded by the Austrian 421

Federal Ministery for Science, Research and Economy. Opinions presented in this work are those of 422

the authors and do not represent the official policy of the Austrian armed forces. 423

424

REFERENCES 425

1. Wang D, Xu X, Deng X, Chen C, Li B, Tan H, Wang H, Tang S, Qiu H, Chen J, Ke B, 426 Ke C, Kan B. 2010. Detection of Vibrio cholerae O1 and O139 in environmental water 427 samples by an immunofluorescent-aggregation assay. Appl Environ Microbiol 76:5520-428 5525. 429

2. Ali M, Lopez AL, You YA, Kim YE, Sah B, Maskery B, Clemens J. 2012. The global 430 burden of cholera. Bull World Health Organ 90:209-218A. 431

3. Cheasty T, Said B, Threlfall EJ. 1999. V cholerae non-O1: implications for man? Lancet 432 354:89-90. 433

4. Koskela KA, Matero P, Blatny JM, Fykse EM, Olsen JS, Nuotio LO, Nikkari S. 2009. 434 A multiplatform real-time polymerase chain reaction detection assay for Vibrio cholerae. 435 Diagnostic Microbiology and Infectious Disease 65:339-344. 436

5. Baron S, Chevalier S, Lesne J. 2007. Vibrio cholerae in the environment: a simple method 437 for reliable identification of the species. J Health Popul Nutr 25:312-318. 438

6. Reidl J, Klose KE. 2002. Vibrio cholerae and cholera: out of the water and into the host. 439 FEMS Microbiology Reviews 26:125-139. 440

7. Das M, Bhowmick TS, Nandy RK, Nair GB, Sarkar BL. 2009. Surveillance of 441 vibriophages reveals their role as biomonitoring agents in Kolkata. FEMS Microbiology 442 Ecology 67:502-510. 443

8. Hasan JA, Bernstein D, Huq A, Loomis L, Tamplin ML, Colwell RR. 1994. Cholera 444 DFA: an improved direct fluorescent monoclonal antibody staining kit for rapid detection 445 and enumeration of Vibrio cholerae O1. FEMS Microbiol Lett 120:143-148. 446

9. Schauer S, Sommer R, Farnleitner AH, Kirschner AK. 2012. Rapid and sensitive 447 quantification of Vibrio cholerae and Vibrio mimicus cells in water samples by use of 448 catalyzed reporter deposition fluorescence in situ hybridization combined with solid-phase 449 cytometry. Appl Environ Microbiol 78:7369-7375. 450

10. Haugland RA, Siefring SC, Wymer LJ, Brenner KP, Dufour AP. 2005. Comparison of 451 Enterococcus measurements in freshwater at two recreational beaches by quantitative 452 polymerase chain reaction and membrane filter culture analysis. Water Res 39:559-568. 453

11. Griffith JF, Weisberg SB. 2011. Challenges in Implementing New Technology for Beach 454 Water Quality Monitoring: Lessons From a California Demonstration Project. Marine 455 Technology Society Journal 45:65-73. 456

12. Blackstone GM, Nordstrom JL, Bowen MD, Meyer RF, Imbro P, DePaola A. 2007. Use 457 of a real time PCR assay for detection of the ctxA gene of Vibrio cholerae in an 458 environmental survey of Mobile Bay. Journal of Microbiological Methods 68:254-259. 459

13. Chua AL, Elina HT, Lim BH, Yean CY, Ravichandran M, Lalitha P. 2011. 460 Development of a dry reagent-based triplex PCR for the detection of toxigenic and non-461 toxigenic Vibrio cholerae. J Med Microbiol 60:481-485. 462

14. Goel AK, Ponmariappan S, Kamboj DV, Singh L. 2007. Single multiplex polymerase 463 chain reaction for environmental surveillance of toxigenic-pathogenic O1 and non-O1 464 Vibrio cholerae. Folia Microbiol (Praha) 52:81-85. 465

15. Gubala AJ. 2006. Multiplex real-time PCR detection of Vibrio cholerae. J Microbiol 466 Methods 65:278-293. 467

16. Nandi B, Nandy RK, Mukhopadhyay S, Nair GB, Shimada T, Ghose AC. 2000. Rapid 468 method for species-specific identification of Vibrio cholerae using primers targeted to the 469 gene of outer membrane protein OmpW. J Clin Microbiol 38:4145-4151. 470

17. Wilson IG. 1997. Inhibition and facilitation of nucleic acid amplification. Appl Environ 471 Microbiol 63:3741-3751. 472

18. Matheson CD, Gurney C, Esau N, Lahto R. 2010. Assessing PCR inhibition from humic 473 substances. The Open Enzyme Inhibition Journal 3:38-45. 474

19. Hoffmann B, Depner K, Schirrmeier H, Beer M. 2006. A universal heterologous internal 475 control system for duplex real-time RT-PCR assays used in a detection system for 476 pestiviruses. Journal of Virological Methods 136:200-209. 477

20. Bustin SA, Benes V, Garson JA, Hellemans J, Huggett J, Kubista M, Mueller R, Nolan 478 T, Pfaffl MW, Shipley GL, Vandesompele J, Wittwer CT. 2009. The MIQE Guidelines: 479 Minimum Information for Publication of Quantitative Real-Time PCR Experiments. 480 Clinical Chemistry 55:611-622. 481

21. Wellinghausen N, Frost C, Marre R. 2001. Detection of legionellae in hospital water 482 samples by quantitative real-time LightCycler PCR. Appl Environ Microbiol 67:3985-3993. 483

22. Bushon RN, Kephart CM, Koltun GF, Francy DS, Schaefer FW, 3rd, Alan Lindquist 484 HD. 2010. Statistical assessment of DNA extraction reagent lot variability in real-time 485 quantitative PCR. Lett Appl Microbiol 50:276-282. 486

23. Tissier A, Denis M, Hartemann P, Gassilloud B. 2012. Development of a rapid and 487 sensitive method combining a cellulose ester microfilter and a real-time quantitative PCR 488 assay to detect Campylobacter jejuni and Campylobacter coli in 20 liters of drinking water 489 or low-turbidity waters. Appl Environ Microbiol 78:839-845. 490

24. Haugland RA, Siefring S, Lavender J, Varma M. 2012. Influences of sample interference 491 and interference controls on quantification of enterococci fecal indicator bacteria in surface 492 water samples by the qPCR method. Water Research 46:5989-6001. 493

25. Kirschner AK, Schlesinger J, Farnleitner AH, Hornek R, Suss B, Golda B, Herzig A, 494 Reitner B. 2008. Rapid growth of planktonic Vibrio cholerae non-O1/non-O139 strains in a 495 large alkaline lake in Austria: dependence on temperature and dissolved organic carbon 496 quality. Appl Environ Microbiol 74:2004-2015. 497

26. Eiler A, Farnleitner AH, Zechmeister TC, Herzig A, Hurban C, Wesner W, Krachler 498 R, Velimirov B, Kirschner AK. 2003. Factors controlling extremely productive 499 heterotrophic bacterial communities in shallow soda pools. Microb Ecol 46:43-54. 500

27. Dickinson G, Lim KY, Jiang SC. 2013. Quantitative microbial risk assessment of 501 pathogenic vibrios in marine recreational waters of southern california. Appl Environ 502 Microbiol 79:294-302. 503

28. Singh DV, Isac SR, Colwell RR. 2002. Development of a hexaplex PCR assay for rapid 504 detection of virulence and regulatory genes in Vibrio cholerae and Vibrio mimicus. J Clin 505 Microbiol 40:4321-4324. 506

29. Angela Di P, Giuseppina C, Giuseppina T, Domenico C, Vito Tony F. 2005. A 507 Collagenase-Targeted Multiplex PCR Assay for Identification of Vibrio alginolyticus, 508 Vibrio cholerae, and Vibrio parahaemolyticus. Journal of Food Protection 68:150-153. 509

30. Maiti D, Das B, Saha A, Nandy RK, Nair GB, Bhadra RK. 2006. Genetic organization of 510 pre-CTX and CTX prophages in the genome of an environmental Vibrio cholerae non-O1, 511 non-O139 strain. Microbiology 152:3633-3641. 512

31. Campos E, Bolanos H, Acuna MT, Diaz G, Matamoros MC, Raventos H, Sanchez LM, 513 Sanchez O, Barquero C. 1996. Vibrio mimicus diarrhea following ingestion of raw turtle 514 eggs. Appl Environ Microbiol 62:1141-1144. 515

32. Faruque SM, Rahman MM, Asadulghani, Nasirul Islam KM, Mekalanos JJ. 1999. 516 Lysogenic conversion of environmental Vibrio mimicus strains by CTXPhi. Infect Immun 517 67:5723-5729. 518

33. Acuna MT, Diaz G, Bolanos H, Barquero C, Sanchez O, Sanchez LM, Mora G, Chaves 519 A, Campos E. 1999. Sources of Vibrio mimicus contamination of turtle eggs. Appl Environ 520 Microbiol 65:336-338. 521

34. Faruque SM, Asadulghani, Alim AR, Albert MJ, Islam KM, Mekalanos JJ. 1998. 522 Induction of the lysogenic phage encoding cholera toxin in naturally occurring strains of 523 toxigenic Vibrio cholerae O1 and O139. Infect Immun 66:3752-3757. 524

35. Lyon WJ. 2001. TaqMan PCR for detection of Vibrio cholerae O1, O139, non-O1, and 525 non-O139 in pure cultures, raw oysters, and synthetic seawater. Appl Environ Microbiol 526 67:4685-4693. 527

36. Rivera IN, Chun J, Huq A, Sack RB, Colwell RR. 2001. Genotypes associated with 528 virulence in environmental isolates of Vibrio cholerae. Appl Environ Microbiol 67:2421-529 2429. 530

37. Albert MJ, Islam D, Nahar S, Qadri F, Falklind S, Weintraub A. 1997. Rapid detection 531 of Vibrio cholerae O139 Bengal from stool specimens by PCR. J Clin Microbiol 35:1633-532 1635. 533

38. Fields PI, Popovic T, Wachsmuth K, Olsvik O. 1992. Use of polymerase chain reaction 534 for detection of toxigenic Vibrio cholerae O1 strains from the Latin American cholera 535 epidemic. J Clin Microbiol 30:2118-2121. 536

39. International Organisation for Standardization. 2006. Foodstuffs -- Methods of analysis 537 for the detection of genetically modified organisms and derived products -- General 538 requirements and definitions. ISO/TC 34/SC 16:16. 539

40. Rock C, Alum A, Abbaszadegan M. 2010. PCR inhibitor levels in concentrates of biosolid 540 samples predicted by a new method based on excitation-emission matrix spectroscopy. Appl 541 Environ Microbiol 76:8102-8109. 542

41. Lloyd KG, MacGregor BJ, Teske A. 2010. Quantitative PCR methods for RNA and DNA 543 in marine sediments: maximizing yield while overcoming inhibition. FEMS Microbiology 544 Ecology 72:143-151. 545

42. Schijven JF, Teunis PF, Rutjes SA, Bouwknegt M, de Roda Husman AM. 2011. 546 QMRAspot: a tool for Quantitative Microbial Risk Assessment from surface water to 547 potable water. Water Res 45:5564-5576. 548

549 550

551

Table 1: Gene targets, primers, probes and references used in this study. 552

gene primer/ probe a primer sequence (5´- 3´)b Amplicon sizes source

ompW ompW-F1

ompW-F2

ompW-R

ompW-P

AAG CTC CGC TCC TGT ATT TGC

ACT AGC CGC TCC TGT ATT TGC

GCT ATT AAC TGC CAA CTC ACT TTG AG

6-FAM-CAC CAA GAA GGT GAC TTT-MGB-NFQ

127

this study

ctxA ctxA-F

ctxA-R

ctxA-P

GCA TAG AGC TTG GAG GGA AGA G

CAT CGA TGA TCT TGG AGC ATT C

VIC-CAT CAT GCA CCG CCG-MGB-NFQ

76 this study

egfp

egfp

IC-F

IC-R

IC-P

12-F

10-R

GAC CAC TAC CAG CAG AAC AC

GAA CTC CAG CAG GAC CAT G

Cy5-AGC ACC CAG TCC GCC CTG AGC A-BHQ2

TCG AGG GCG ACA CCC TG

CTT GTA CAG CTC GTC CAT GC

132

374

(19)

(19)

ompW ompW-out-F

ompW-out-R

AAA CAC ATT ATT AAT GTG CCT AAA TGA

CGT TAG CAG CAA GTC CC

613 this study

ctxA ctxA-out-F

ctxA-out-R

CAG ATG GTT ATG GAT TGG CAG

GGC AAA ACG GTT GCT TC

684 this study

a F, forward; R, reverse; P, probe 553 b MGB-NFQ, minor groove binding non-fluorescent quencher, BHQ2, non-fluorescent black hole 554 quencher 2 555 556

Table 2: Accuracy parameters and analytical sensitivity of the triplex qPCR. 557

Targeta Slope

(SD)

r2

(SD)

Efficiency

[%]

(SD)

IAV [Ct]b

(range)

IEV [%]c

(range)

MDLd

[copies per reaction]

ompW - 3.32

(0.22)

0.999

(0.000)

96.3

(4.67)

0.28

(0.15-0.37)

11

(2-28)

5

ctxA - 3.35

(0.059)

0.999

(0.000)

97.8

(2.08)

0.21

(0.1-0.3)

8

(2-15)

5

egfp - 3.15

(0.163)

0.999

(0.000)

105.2

(3.73)

0.29

(0.13-0.4)

9

(4-14)

5

a all values based on 5 replicate measurements, bintra-assay variance, cinter-assay variance, dmethod detection limit 558

559

FIGURE LEGENDS 560

FIG. 1. Results from DNA binding capacity test: Quantification of the ompW and ctxA targets used 561

in the qPCR assay after spiking of ten-fold serial dilutions of V. cholerae O395 type strain cells into 562

the DNA extraction kit. Two microliters of each extraction eluate (~ 106 to ~ 102 copies per 100µl 563

eluate) were used as a template in the presence of 250 copies of IC plasmid per reaction. The bars 564

represent means and SD of triplicate extractions, each dilution step was also run in triplicate. 565

Diamonds show the detected IC targets as percentage of added IC targets. X-axis: number of V. 566

cholerae spiked into the extraction kit 567

FIG. 2. Global recovery rates of the triplex qPCR obtained from spiking a known number of V. 568

cholerae O395 cells into different environmental water samples (non-chlorinated drinking water; 569

surface water from the River Danube and lake Neusiedler See as well as from Zicklacke (ZL) and 570

Oberer Stinker (OS). Main graph: GR rates (in percent) along a gradient of dissolved organic 571

carbon (DOC; white diamonds). Bars are the mean and SD of triplicate qPCR results obtained from 572

one spiked sample. Background concentrations of ompW in non-spiked parallel samples were 573

subtracted. V. cholerae was present in Neusiedler See, ZL and OS, but not in the Danube and 574

drinking water samples. Only ompW results are shown; results from parallel qPCR for ctxA were 575

not significantly different (paired T-test, T = 0.82, p > 0.4; data not shown). Small insert: GR (in 576

absolute cell numbers and as percent values) after spiking 5.4 × 105 V. cholerae O395 cells into 200 577

ml Lake Neusiedler See water at three different stations. Horizontal lines represent 5.4 × 105 cells. 578

Bars represent mean and SD of triplicate qPCR results for ompW and ctxA. 579

FIG 3. Quantification of autochthonous V. cholerae by the triplex qPCR (only ompW target 580

shown), CARD-FISH/SPC and cultivation in water samples taken from the lake Neusiedler See 581

(5A-C, 36A-C) and the shallow soda lake Zicklacke (ZLA,B). For qPCR, concentrations were 582

obtained without correction (black lines) and with correction of triplicate qPCR results (grey 583

columns) with the respective global recovery rate (white diamonds) obtained from parallel spiked 584

samples. All values represent the mean and 1 standard deviation of triplicate samples. Standard 585

deviations of uncorrected q-PCR results are too small to be visible. In addition, no positive ctxA 586

signals were obtained. Results were sorted along decreasing global recovery rates. CARD-587

FISH/SPC results and cultivation results were determined according to the protocols described in 588

Schauer et al. (9). 589