Evaluation of DNA extraction methods from mouse stomachs for the quantification of H. pylori by real-time PCR

Journal of Microbiological Methods 62 (2005) 71 – 81 www.elsevier.com/locate/jmicmeth

Evaluation of DNA extraction methods from mouse stomachs for the quantification of H. pylori by real-time PCR Yvonne Roussela, Mark Wilksa,T, Andrew Harrisb, Charles Meinc, Soad Tabaqchalia,d a

Department of Microbiology and Virology, Barts and the London NHS Trust St. Bartholomew’s Hospital, West Smithfield, London, EC1A 7BE, United Kingdom b School of Biotechnology and Biomolecular Sciences, University of New South Wales, Sydney, 2052, Australia c Genome Centre, Barts and The London, Queen Mary’s School of Medicine and Dentistry, Charterhouse Square, London, EC1M 6BQ, United Kingdom d Wolfson Institute of Preventive Medicine, Barts and The London, Queen Mary’s School of Medicine and Dentistry, Charterhouse Square, London, EC1M 6BQ, United Kingdom Received 16 July 2004; received in revised form 14 January 2005; accepted 24 January 2005 Available online 25 February 2005

Abstract Real-time PCR methods have recently been developed for the quantification of Helicobacter pylori from infected mouse stomachs. However, the extent to which results is affected by the efficiency of different methods of DNA extraction and the degree of inhibition of the subsequent PCR have largely been ignored. In this study, mouse stomachs were processed using two homogenisation methods: complete disruption using a blender and homogenisation by vortexing with glass beads. Each procedure was followed by DNA purification by three different protocols–two commercially available kits–Qiagen DNA Mini Tissue kit and Qiagen Stool Kit and a phenol–chloroform extraction method. PCR inhibition was assessed by screening for mouse DNA and for H. pylori DNA after spiking stomach extracts with H. pylori 16S rDNA. PCR inhibition was found to be lower in DNA samples prepared by vortexing and processed by column kits. Validation of procedures was performed by quantification of H. pylori DNA and mouse DNA in infected mouse stomachs. Homogenisation with glass beads followed by the Qiagen Tissue kit was found to be the most suitable protocol combining high extraction and detection efficiency of 16S rDNA in the presence of a mouse DNA background. D 2005 Elsevier B.V. All rights reserved. Keywords: Helicobacter pylori; DNA extraction; PCR inhibition; Real-time quantitative PCR; TaqMan; Mouse stomach

1. Introduction T Corresponding author. Tel.: +44 20 7601 8401; fax: +44 20 7601 8409. E-mail address: [email protected] (M. Wilks). 0167-7012/$ - see front matter D 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.mimet.2005.01.011

Helicobacter pylori is a common pathogen of the human stomach, affecting 50% of the world’s population (Nakayama and Graham, 2004). It can

72

Y. Roussel et al. / Journal of Microbiological Methods 62 (2005) 71–81

cause chronic active gastritis, peptic ulcer disease, and is a risk factor for gastric adenocarcinoma and lymphoma. For the last decade, there has been intense research into the development of vaccines. Numerous H. pylori antigens have been tested in different animal models amongst which the mouse has been the most widely used (Lachman et al., 1997; Lee et al., 1997; Marchetti et al., 1995). Reduction of Helicobacter infection following vaccine regimens has been assessed by different assays such as recovery of H. pylori by culture (Goodwin et al., 1985), urease testing (Hazell et al., 1987) and histological examination (Lee et al., 1997). Nowadays real-time quantitative PCR technology allows quantification of pathogen DNA in clinical samples with a higher level of rapidity, specificity and sensitivity. Whereas some studies have been reported on real-time PCR protocols for the detection of H. pylori (Bjorkholm et al., 2004; He et al., 2002; Huijsdens et al., 2004; Jiang et al., 2003; Mikula et al., 2003; Straubinger et al., 2003), less information is available on the impact of stomach processing and DNA purification procedures on the suitability of the extracted DNA for amplification. This issue is of critical importance in vaccine studies where accurate quantification of H. pylori strongly depends on the yield and the quality of DNA extracted from the stomach sample. It was therefore important to determine an extraction method of H. pylori DNA from gastric tissues of the greatest possible quantity, consistency and quality which would provide templates suitable for real-PCR. In our study, cells were collected from mouse stomachs by two methods: (i) by complete disruption of the stomach using a blender (HB) or (ii) by vortexing with glass beads (Vortex). Each cell extraction procedure was followed by DNA purification by three different protocols–two commercially available kits–Qiagen DNA Mini Tissue kit and Qiagen Stool Kit, and a manual phenol–chloroform extraction method. Performances of the different protocols were tested in terms of DNA yield and amplification efficiency by TaqMan PCR. Inhibition level was firstly assessed by screening for mouse DNA where amounts of mouse DNA determined by TaqMan PCR were compared with PicoGreen-based data. Secondly, inhibition level was tested by spiking the different stomach DNA extracts with H. pylori 16S rDNA and

by comparing the quantity of DNA detected by TaqMan PCR with the original input amount. Quantification of H. pylori DNA was then assessed from mouse stomachs which had been spiked with H. pylori cells and from mice which had been undergone immunisation after experimental infection with H. pylori.

2. Materials and methods 2.1. Mice and H. pylori C57BL/6 female specific pathogen-free mice were purchased from Harlan (Oxon, UK). All procedures were performed in compliance with the Guidance on the Operation of the Animals (Scientific Procedures) Act 1986. Stomachs from 15 uninfected mice were used for determination of total DNA yield and PCR inhibition assessment. Two stomachs were used for the artificial infection model where freshly collected stomachs were opened and spiked each with 106 CFU equivalents of H. pylori cells. H. pylori strain SS1 (Lee et al., 1997) was cultured by previously described methods (Harris et al., 2003) and resuspended in BHI. The optical density of cell suspensions was read at 600 nm and converted to units of CFU/ml according to Blom’s formula (Blom et al., 2002). Four mice were used for the experimentally low-infected mouse model. Each mouse received 100 Ag of purified recombinant urease (OraVax, Cambridge MA) plus 10 Ag of cholera toxin (C8052, Sigma) via orogastric route. Immunisation was repeated three times with equivalent doses at 2-week intervals. Three weeks after the last immunisation dose, mice were challenged twice at 2-day intervals by orogastric route with 108 CFU of H. pylori SS1. Three weeks after the H. pylori challenge, animals were sacrificed and their stomachs were collected for quantification of H. pylori infection. For the highinfected mouse model, four mice were challenged twice with 108 CFU of H. pylori as described above. For all mice (infected or not), stomachs were processed immediately after killing and divided longitudinally into two halves. Each half of the stomach was rinsed twice by rapid submersion in normal saline (0.9%), allowed to drain briefly and weighed.

Y. Roussel et al. / Journal of Microbiological Methods 62 (2005) 71–81

2.2. Stomach processing Two different methods were used as follows (i) Homogenisation using a blender (HB). Each half stomach was resuspended into 2 ml of saline and disrupted at full speed for 20–30 s with a blade-blender homogeniser (UltraTurraxR T8, Janki and Kunkel, Staufen, Germany) producing a coarse homogenous suspension. To minimise sample to sample contamination, the probe of the homogeniser was cleaned between each stomach by operating it in 10% SDS, in 70% (v/v) alcohol and finally in three different PBS solutions. (ii) Vortexing with glass beads (Vortex/VGB). The other half stomach was resuspended in 1 ml of PBS with 0.5 g of glass beads (1 mm diameter, Zirconia/silica, Biospec, Bartlesville, OK) and vortexed at full speed five times for 30 s. This process left the half-stomach unbroken and produced a clear cell homogenate which was used for DNA extraction. Homogenates produced by either method were divided into 100-Al aliquots and stored at 20 8C until DNA purification. 2.3. DNA purification from the stomach preparations Three protocols of DNA purification were performed on the stomach homogenate preparations: (i) Qiagen DNA Tissue Mini Kit (Tissue kit): The manufacturer’s protocol (09/2001, Qiagen, Crawley, UK) was used with the following modifications. To 200 Al of stomach homogenate, 500 Al of Buffer ATL was added and tissue digestion was carried out by incubating for 2 h at 56 8C in presence of Proteinase K (20 mg/ml, Qiagen). Both Buffer AL and ethanol were added (600 Al each) to the cell lysate before loading the QIAamp columns. The columns were then washed according to the recommended protocol and DNA eluted twice with 200 Al of the supplied AE elution buffer. (ii) Qiagen DNA Stool Mini Kit (Stool kit). The protocol (08/2001, Qiagen) was used with the following modifications. 500 Al of Buffer ASL

73

was added to 200 Al of stomach homogenate and incubated for 5 min at 70 8C. After centrifugation, half an InhibitEX pellet was added to the supernatant and processed until step 9 of the recommended protocol. To 600 Al of the resulting supernatant from step 8, 20 Al of Proteinase K (Qiagen) and 600 Al of buffer AL were added. After incubation at 70 8C for 10 min, 600 Al of ethanol was added to the mixture and loaded into a QIAamp spin column. Column washes were performed as recommended in the kit and DNA was finally eluted twice with 200 Al of the supplied AE elution buffer. (iii) Phenol–chloroform extraction method (Phenol protocol). 100 Al of stomach homogenate was digested in the presence of 1% SDS (Promega, Southampton, UK) and 12 mAU of ProteinaseK (Qiagen) in 400 Al of TE at 56 8C for 1 h. DNA was then extracted with an equal volume of phenol (pH 8.5, Sigma), followed by extraction with an equal volume of phenol/ chloroform/isoamyl alcohol (25:24:1 ratio) and finally with an equal volume of chloroform/ isoamyl alcohol (24:1 ratio) (Sigma). DNA was then precipitated with 2.5 vol of ethanol and 1/ 10 vol of 3 M Na acetate pH5.2 (Sigma) overnight at 20 8C. After centrifugation at 13,000 rpm for 10 min, pellets were washed with 70% ethanol and air-dried for 15 min, after which they were dissolved in 200 Al of AE buffer (Qiagen) and stored at 20 8C until further use. 2.4. Measurement of DNA concentration (i) Fluorescence with PicoGreen. DNA concentrations were determined by fluorescence with PicoGreen (Molecular Probes, Leiden, Netherlands) using DNA calibration standards made of Calf Thymus DNA (Sigma) serially diluted from 2.4 to 0.0375 ng/Al. DNA preparations were diluted 100- to 500-fold to a concentration up to 0.1 ng/Al (according to spectrophotometric values). Fluorescence emission was measured at 530 nm on a Wallac Victor2 (Perkin Elmer Life Sciences, UK) after excitation at 480 nm.

74

Y. Roussel et al. / Journal of Microbiological Methods 62 (2005) 71–81

(ii) TaqMan PCR. The concentration of mouse DNA was determined by quantitative PCR using the GAPDH TaqMan system described below (Section 2.5). The assay was calibrated by measuring the signal obtained from known amounts of mouse liver DNA over the range 80 ng to 0.8 pg per reaction. Mouse liver DNA was extracted from C57BL/6 mouse liver tissue using the WizardR Genomic DNA Purification Kit (Promega) and DNA concentration was determined by both PicoGreen fluorescence and spectrophotometry at 260 nm. To determine the concentration of mouse DNA from stomach extracts, 200-fold dilution of HB samples and 20-fold dilution of Vortex samples were used in the TaqMan reactions. 2.5. TaqMan PCR assays TaqMan PCR assays were carried out in ABI PRISMk Optical 384-Well reaction plates (Applied Biosystems, Warrington, UK) on an ABI PRISMR 7900HT Sequence Detection using all default program settings. Each 10-Al TaqMan reaction consisted of 5 Al of the TaqMan Universal PCR Master Mix (Applied Biosystems), 4 Al of DNA sample to be tested and 1 Al of oligonucleotide mixture containing each of the two primers and the dually labelled probe. H. pylori specific primers HP5 (5V-TTTGTTAGAGAAGATAATGACGGTATCTAAC-3V) and HP6 (5VCATAGGATTTCACACCTGACTGACTATC-3V) were used at the concentration of 900 nM each with the Taqman probe 16S1 (5V-CGTGCCAGCAGCCGCGGT-3V) at the concentration of 125 nM. Mouse-specific primers GAPDHsense (5V-TGCACCACCAACTGCTTAG-3V) and GAPDHantisense (5V-GGATGCAGGGATGATGTTC-3V) were used at the concentration of 600 nM each and GAPDH probe (5V-CAGAAGACTGTGGATGGCCCT-3V) at the concentration of 200 nM. Unlabelled primers were cartridge-purified and obtained from InvitroGen Life Technologies (InvitroGen, Glasgow, UK). TaqMan probes, labelled at the 5V-end with the 6-carboxyfluorescein (FAM) reporter dye and at the 3V-end with the 6carboxy-tetramethylrhodamine (TAMRA) quencher dye were purchased from Sigma Genosys (Sigma, Haverhill, UK).

All samples were measured at least three times in each assay, and negative controls without template were included in each PCR run. Amplification and data analysis were performed using the SDS2.1 software (Applied Biosystems). Gene copies per PCR for both assays were calculated using a standard curve generated at each assay. 2.6. Statistics Statistics tests were performed with SPSS version 11.5 (SPSS, Chicago, IL). To test reproducibility of assays and compare variability of DNA detection between methods, the coefficient of variation (CV) was determined as the standard deviation expressed as a percentage of the mean. Ttests and Mann–Whitney tests were used to compare means, mean differences were considered significant at pb0.05. For PCR inhibition experiments, DNA was extracted from five mouse gastric samples using the same protocol and pooled. The pooled specimen was then used to assess PCR inhibition by quantifying the amount of mouse DNA with the GAPDH Q-PCR system or by quantifying the amount of H. pylori 16S gene after spiking the pooled specimens with p16S rDNA.

3. Results 3.1. Total DNA extraction yield Stomachs from 15 mice were extracted using the six different protocols for the determination of total DNA yields and variability (Fig. 1). Yields ranged from 83 Ag DNA per gram of stomach with the Vortex-Stool protocol to 1655 Ag per gram for the Blender-Tissue protocol. About seven times more DNA was extracted after homogenisation with the Blender compared to Vortex method, irrespective of the DNA purification method used. In terms of DNA purification procedures, the Tissue kit extracted slightly more DNA than the Phenol protocol and approximately three times more than the Stool kit. Variability in the DNA yield (CV) was between 32% and 44% for all protocols except for Vortex-Stool which reached 54%.

Y. Roussel et al. / Journal of Microbiological Methods 62 (2005) 71–81

75

Fig. 1. Yield of total DNA extracted from mouse stomachs. Values for each preparation method are an average of 15 samples and error bars show 95% confidence intervals (CI). Yield was expressed in micrograms of DNA (determined by PicoGreen) per gram of stomach (wet weight).

3.2. Assessment of PCR inhibition in DNA extracts 3.2.1. Screening of mouse DNA with the GAPDH TaqMan system To assess PCR inhibition in DNA samples produced by each of the six protocols, mouse DNA concentrations were estimated by TaqMan PCR with

the GAPDH primer system on serially diluted samples and compared with concentrations determined by PicoGreen (Fig. 2). There was complete inhibition of PCR in undiluted samples when stomachs had been homogenised with the blender regardless of the DNA purification method used. The highest amplification rates of mouse DNA were obtained when DNA

Fig. 2. Assessment of PCR inhibition by screening of mouse DNA in serially diluted DNA extracts. Serial dilutions were neat, 1/2, 1/10 and 1/ 20 for the Vortex samples and 1/20, 1/100 and 1/200 for the Blender samples. Concentrations of mouse DNA were determined by Q-PCR with the GAPDH primer system (empty bars) by PicoGreen fluorescence (dotted bars). Q-PCR values are average of three independent experiments and error bars indicate standard deviations of the mean.

76

Y. Roussel et al. / Journal of Microbiological Methods 62 (2005) 71–81

samples were diluted at 1:20 (Vortex) or 1:200 (Blender) where the column purification protocols led to detection rates of above 90% and the Phenol protocol resulted in 77% detection. 3.2.2. Assessment of PCR inhibition with 16S rDNA TaqMan system To examine how PCR inhibition could affect quantification of H. pylori 16S genes in the DNA samples, dilutions of the DNA samples were spiked with a small quantity of p16S plasmid DNA and analysed by TaqMan PCR with the H. pylori 16S primer system. Detection rates of 16S rDNA in the samples produced by each of the six protocols are shown in Fig. 3. As previously observed with the mouse TaqMan system, there was no amplification of DNA in the undiluted DNA preparations obtained from blender-homogenised stomach whereas 16S rDNA could be detected in undiluted DNA samples produced from vortexed stomachs (62% for Stool, 60% for Tissue and 24% for Phenol). Detection of above 90% of the 16S rDNA input could be attained in the Vortex samples (VGB) when 10-fold diluted and in Blender samples (HB) when 100-fold diluted. 3.2.3. Assessment of non-target DNA involvement in PCR inhibition To determine if PCR inhibition detected in DNA samples extracted from stomach tissues could be due to an excess of mouse DNA, p16S plasmid DNA

(2300 copies) was assayed in the presence of stomach mouse DNA (Blender-Phenol DNA at 20 ng Al 1) and/or non-stomach mouse DNA. Inhibition was found to be significantly higher in stomach DNA (30%) than in non-stomach DNA (18%) when present at the same concentration of 20 ng Al 1 ( p=0.012). In the presence of mouse stomach DNA, addition of the same amount of non-stomach mouse DNA did not change the inhibition rate (30% vs. 32%, p=0.63). When p16S DNA was assayed in non-stomach DNA present at 80 ng Al 1, high inhibition was observed (79%) and was significantly increased to 91% in presence of stomach mouse DNA ( pV0.001). 3.3. Quantification of H. pylori DNA extracted from mouse stomachs 3.3.1. Variability of DNA extraction yield for each protocol To assess reproducibility of the entire assay (extraction and PCR), open stomachs were spiked with equal volumes of the same suspension of H. pylori cells before processing by the six different protocols. DNA samples were then produced in quadruplicate from each of the homogenates, and TaqMan PCR assays were used to quantify H. pylori 16S and GAPDH DNA in each extracts. Yields of H. pylori DNA and mouse DNA and the variability of DNA extraction yields are shown in Table 1. As previously found with total DNA yields, approxi-

100

Detection rate (%)

80 60 40 20 0 Neat

VGB Stool 1/10 1/100

VGB Tissue Neat 1/10 1/100

VGB Phenol Neat

1/10 1/100 Neat

HB Stool 1/10 1/100 Neat

HB Tissue 1/10

HB Phenol 1/100 Neat

1/10

1/100

Fig. 3. Assessment of PCR inhibition in spiked DNA extracts obtained by six different methods. Each of the diluted DNA extracts (undiluted, diluted 1:10 and diluted 1:100) was spiked with 500 copies of p16S1 per PCR reaction and analysed by the 16S H. pylori Q-PCR assay. Detection rates were calculated by dividing the CT values obtained by the spiked DNA extracts by the CT value determined when 500 copies of p16S1 copies was resuspended in the same quantity of water. Results are the average of four independent experiments and error bars indicate standard error of the means.

Y. Roussel et al. / Journal of Microbiological Methods 62 (2005) 71–81

77

Table 1 Yield of H. pylori DNA and mouse DNA in extracts from stomach spiked with H. pylori cells Vortex a

H. pylori DNA (ng/g) Mouse DNAb (Ag/g) H. pylori/mouse DNA ratio (10 3)

Blender

Stool

Tissue

Phenol

Stool

Tissue

Phenol

44 (F28%) 41 (F34%) 1.09 (F15%)

182 (F4%) 146 (F10%) 1.26 (F11%)

132 (F10%) 124 (F12%) 1.07 (F13%)

89 (F18%) 311 (F22%) 0.29 (F8%)

335 (F4%) 1037 (F4%) 0.32 (F5%)

242 (F5%) 930 (F6%) 0.26 (F4%)

Variability is shown in parentheses as coefficients of variation based on the mean from four independent extractions per sample. a Mean quantities (ng) of H. pylori DNA per gram of stomach, calculated on the basis of 1.78 fg per H. pylori genome (1.65 Mb). b Mean quantities (Ag) of mouse DNA per gram of stomach.

were obtained by infecting mice which had undergone a prophylactic immunisation with purified urease combined with cholera toxin, high H. pylori content mice were infected but not immunised. H. pylori infection rates and variability obtained for the different protocols are shown in Table 2. When mice had been infected but not immunised, high infection rates were obtained, with the highest value achieved when stomach homogenates were processed with the Vortex-Tissue protocol (5.66 ng of H. pylori DNA per mg mouse DNA). Whatever the protocol used, analysis of the samples with high content of H. pylori DNA resulted in low variability, with intra-assay CV ranging from 8% to 16% and inter-assay CV ranging from 14% to 22%. When mice had undergone immunisation, H. pylori infection rates were found to be very low (below 2.8410 2ng of H. pylori DNA per mg of mouse DNA) showing a reduction of 2.6 log10 units compared to unimmunised mice. Discrepancies were found in data obtained from samples with low levels of H. pylori DNA especially in samples that had been produced with the blender.

mately seven times more mouse DNA was measured from the Blender samples than from those produced by Vortex. Twice as much H. pylori DNA was measured in the Blender samples compared to the Vortex samples. Variation in DNA extraction yields was higher with the Stool kit (CV: 18–34%) than with the two other purification methods Tissue and Phenol (CV ranging from 4% to 23%). However, when quantification was performed by normalising H. pylori DNA with mouse DNA (ratio), variation measured between the quadriplate extracts was reduced and found similar among the three different purifications methods (CV from 11% to 13% for Vortex homogenates and CV ranging from 4% to 8% for the Blender samples). 3.3.2. Quantification of H. pylori DNA extracted from stomachs of mice infected with high and low level of H. pylori Applicability of the six different protocols was tested on experimentally infected mice with low and high content of H. pylori. Low H. pylori content mice

Table 2 H. pylori infection rates determined by TaqMan PCR from mouse samples produced by different extraction protocols Vortex

Blender

Stool

Tissue

Phenol

Stool

Tissue

Phenol

H. pylori/mouse ratio Intra-assay CV Inter-assay CV

5.35 8 20

5.66 8 22

5.25 11 20

1.63 16 14

2.13 11 20

1.30 12 18

high H. pylori content

H. pylori/mouse ratio Intra-assay CV Inter-assay CV

2.7510 23 30

ND 162 75

ND 143 43

ND 124 51

low H. pylori content

2

2.8410 31 29

2

2.3110 32 27

2

H. pylori/mouse ratio was calculated by dividing mean quantity of H. pylori (ng/g of stomach) by mean quantity of mouse DNA (Ag/g of stomach) (data not shown). Mean coefficients of variations (CV) are the mean variation between triplicate measurements within the same TaqMan PCR assay (intra-assay) and between independent measurements (inter-assay). ND: not determined as row data were very variable and did not allow calculation of means.

78

Y. Roussel et al. / Journal of Microbiological Methods 62 (2005) 71–81

Variability values were so large (intra-assay CVN124% and inter-assay CV ranging from 43% to 75%) that no infection rates could be determined with accuracy and precision. When the GAPDH mouse assay was applied to these DNA samples, we obtained low intra-assay and inter-assay CV (below 10% and values ranging from 16% to 25%, respectively) indicating that high variability values observed with the 16S assay was likely to be attributable to quantification of low-copynumber target in the real-time PCR.

4. Discussion Real-time quantitative PCR is a fast and sensitive method used to assess H. pylori load in infected mice, but there is a need to standardise the preparation of DNA from gastric specimens. To the best of our knowledge, this study, using six different preparation protocols is the most comprehensive study undertaken to determine yields of DNA obtained from mouse gastric samples and its suitability for real-time PCR. When H. pylori stomach infection studies are carried out in mice, H. pylori cells are often recovered manually by scraping the stomach mucosa with a glass slide or a scalpel for the quantification by colony counts or real-time PCR. However, H. pylori cells are not uniformly distributed in the stomach but show specific gastric tissue tropism with some strains colonising the corpus and some others (like SS1) located preferentially in the antrum (Akada et al., 2003; Lee et al., 1997). The disparate location of H. pylori cells in the stomach as well as the small size of mouse stomachs render manual cell-recovery methods inherently unreliable. Standardised mechanical protocols of cell recovery as presented in this study produce a more consistent and representative sampling of the mouse gastric tissue. Vortexing the stomach with glass beads facilitates the release of cells from the mucosa of the gastric tissue and therefore allows a more selective recovery of bacterial cells. Homogenisation with a blender (HB) induces a thorough disruption of the gastric tissue and a complete recovery of all cells. As a result, yields of total and mouse DNA were found to be seven times higher from homogenisation by blender than from vortexing (Fig. 1 and Table 1). Among the three DNA purification methods, the Tissue kit resulted in the highest DNA yield (up to 1.7 Ag of

total DNA per mg of stomach). The significantly lower yield obtained by the Stool kit (0.6 Ag DNA per mg stomach) could be due to an extra-step involving adsorption of potential PCR inhibitors to a solid matrix, which could contribute to a simultaneous loss of DNA. Another explanation might be that there is a less complete digestion of the gastric tissue by the Stool protocol as this kit was designed to preferentially lyse bacteria present in tissue. A previous study (He et al., 2002) comparing a phenol-based extraction method to a column-based method (InvitroGen Easy-DNA kit) for purification of DNA from gastric mucous found a much higher yield with the phenol protocol (up to 7 Ag of DNA per mg of stomach). This yield was approximately six-fold higher than our phenol yield DNA (1.2 Ag DNA per mg of stomach). This important difference could result from the use of spectrometry as a method of DNA quantification. Indeed, we observed in our study that spectrophotometric DNA concentrations were four to five times higher than concentrations determined by PicoGreen fluorescence or TaqMan (data not shown). This overestimation of spectrophotometric DNA concentrations have already been reported (Dionisi et al., 2003; Zipper et al., 2003) as this method is unable to distinguish between DNA and RNA, has a much lower sensitivity and is more prone to interference by co-purified products. More important than the high recovery of total DNA in the purification protocol is the quality of the extracted sample which should allow a sensitive and accurate quantification of H. pylori DNA by PCR. In our first real-time PCR assays, reactions performed with samples extracted from mouse stomach resulted in inconsistent amplification levels indicating that PCR inhibition was an important issue to consider and assess. Although the importance of PCR inhibition has been widely recognised in the detection of pathogens from clinical samples (Lantz et al., 2000), genital samples (Coutlee et al., 2000; Wilcox et al., 2000), respiratory samples (Scott et al., 2003; Ursi et al., 2003) and clinical swabs (Bezold et al., 2000; Cloud et al., 2003), only a few groups have reported the occurrence of PCR inhibition for the detection of H. pylori from gastric specimens (Chisholm et al., 2001; Furuta et al., 1993; Mikula et al., 2003). Although ideally DNA extracts should be free of potential PCR inhibitors, in practice the extent of PCR

Y. Roussel et al. / Journal of Microbiological Methods 62 (2005) 71–81

inhibition should at least be known. In our study, we assessed the extent of PCR inhibition by mouse DNA screening as suggested by Coutlee (Coutlee et al., 2000) but the real impact of inhibitors in the detection of H. pylori DNA was performed by measuring the quantity of spiked 16S rDNA in the DNA samples produced by the different protocols. Quantity of p16S DNA spiked into DNA samples was deliberately low as we and other investigators observed that PCR inhibitors affected the detection of low-concentration target more than high-concentration target (personal data; Cogswell et al., 1996; Fischer-Romero et al., 2000). In many reports, PCR inhibition might have unwittingly been disregarded because high levels of target DNA were used in their real-time PCR. In our study, extensive PCR inhibition was observed in DNA samples obtained by homogenisation using the blender (HB) whereas DNA samples prepared by vortex showed lower inhibition rates. It is common practice to assess DNA quality with spectrophotometric absorbance ratios; however, we did not find any correlation between A 260 /A 280 and A 260 /A 230 ratios and PCR inhibition (data not shown). Nevertheless, a strong correlation was found between PCR inhibition and DNA concentrations (r=0.99, pb0.01, Spearman’s Test) indicating that total DNA concentration was a good marker of PCR inhibition. An excess of non-target DNA has been described as a significant PCR inhibitor when present at critical concentrations (Mikula et al., 2003; Tebbe and Vahjen, 1993). In our study, detection of 16S rDNA was found not to be affected by presence of extra mouse DNA at the concentration normally found in DNA extracts. This indicates that PCR inhibition would rather be due to substances co-purified with DNA. Highest inhibition rates were observed with samples purified with the Phenol protocol, suggesting that inhibition might possibly result from carry-over phenol which then would inhibit the polymerase (Katcher and Schwartz, 1994). Lowest PCR inhibition rate was achieved with the Stool protocol; however, this is probably due to the lower quantity of total DNA extracted and suggests that the matrix specially designed to remove inhibitory compounds from faeces may have little effects on gastric tissue. Inhibition experiments indicated that all DNA extracts required dilution in order to overcome PCR inhibition. To achieve less than 10% inhibition in the detection of the small

79

amount of 16S rDNA, DNA samples have to be diluted 100-fold when produced from the blender homogenates and 10-fold diluted when prepared from vortexed stomachs. In order to evaluate overall the detection efficiency (extraction and PCR) of the different extraction procedures, pre-extraction spike controls were performed in experiments where H. pylori cells were added to stomachs before DNA extraction. As already observed with total DNA yield, quantification of both 16S rDNA and mouse DNA showed the greatest variability with the Stool protocol. However, the high variability shown by the Stool kit was significantly reduced when 16S quantities of H. pylori 16S rDNA was normalised with respect to the quantity of mouse DNA extracted. This shows that variability in DNA extraction caused by a particular protocol can be easily overcome by correcting the amount of H. pylori DNA with the amount of total or mouse DNA. For higher accuracy, it is therefore better to carry out a single extraction, quantify 16S bacterial DNA, and correct for the mouse DNA concentration rather than performing a 16S assay on repeated separate DNA extractions. 16S rDNA quantification performed on DNA samples obtained from mice with high and low H. pylori content allowed us to test assay sensitivities and showed that quantification was of particular concern when the 16S target DNA is present at very low concentration relative to non-target mouse DNA. Intra- and inter-assay variability was relatively low (from 8% to 16% and from 14% to 22%, respectively) when the infection rate was above 10 3 (as measured as H. pylori–mouse DNA ratio). However when the ratio was below 210 5 as in the case of successfully immunised mice, variability was significantly higher and did not permit the determination of accurate infection rates in the cases of blender samples. In these samples, the initial low-content of 16S target DNA was further reduced by the 100-fold dilution to alleviate PCR inhibition. This led to a pre-PCR DNA concentration below the detection limit of the assay because DNA is subject to the inherent randomness of PCR and may therefore result in very disparate quantities of amplicon. This shows a limitation of the real-time PCR technology, detection sensitivity should be increased by refining DNA extraction techniques and improving quantitative PCR reactions. Specific procedures of DNA extraction from stomach

80

Y. Roussel et al. / Journal of Microbiological Methods 62 (2005) 71–81

samples have been designed to selectively eliminate PCR contaminants, such as the use of Chelex 100 which removes multivalent cations and lyses cells (Walsh et al., 1991). However, it may not be sufficient for DNA purification from gastric mucosal samples as PCR inhibition has been observed in these conditions (Mikula et al., 2003). The addition of bovine serum albumin (BSA), as was suggested for the analysis of DNA obtained from clinical samples (Abu Al-Soud and Radstrom, 2000), did not improve amplification efficiency in our DNA samples and neither did heat inactivation (data not shown). In conclusion, the Vortex/Tissue protocol produced DNA from mouse stomachs with the highest extraction and amplification efficiencies. Moreover, the vortex procedure is more practical than the blender homogenisation as a large number of samples can be handled at the same time by using a device such as the MiniBead Beater (Biospec). To minimise PCR inhibition in Vortex Tissue samples, a 10-fold dilution should be applied before the 16S PCR assay. This allows accurate determination of H. pylori DNA and provides a good sensitivity even when mice have low H. pylori content. When infection rates are to be determined, it is important to correct the 16S H. pylori DNA quantities with the host DNA quantities determined as total DNA or as mouse GAPDH DNA. Acknowledgments Research support was provided by the European Commission (Framework Program 5; project QLK12000-00146: DEPROHEALTH bProbiotic Strains With Designed Health Properties.Q We acknowledge Simon Warwick for his comments regarding the methodologies used. References Abu Al-Soud, W., Radstrom, P., 2000. Effects of amplification facilitators on diagnostic PCR in the presence of blood, feces, and meat. J. Clin. Microbiol. 38, 4463 – 4470. Akada, J.K., Ogura, K., Dailidiene, D., Dailide, G., Cheverud, J.M., Berg, D.E., 2003. Helicobacter pylori tissue tropism: mousecolonizing strains can target different gastric niches. Microbiology 149, 1901 – 1909. Bezold, G., Volkenandt, M., Gottlober, P., Peter, R.U., 2000. Detection of herpes simplex virus and varicella zoster virus in

clinical swabs: frequent inhibition of PCR as determined by internal controls. Mol. Diagn. 5, 279 – 284. Bjorkholm, B., Guruge, J., Karlsson, M., O’Donnell, D., Engstrand, L., Falk, P., Gordon, J., 2004. Gnotobiotic transgenic mice reveal that transmission of Helicobacter pylori is facilitated by loss of acid-producing parietal cells in donors and recipients. Microbes Infect. 6, 213 – 220. Blom, K., Svennerholm, A.M., Bolin, I., 2002. The expression of the Helicobacter pylori genes ureA and nap is higher in vivo than in vitro as measured by quantitative competitive reverse transcriptase-PCR. FEMS Immunol. Med. Microbiol. 32, 219 – 226. Chisholm, S.A., Owen, R.J., Teare, E.L., Saverymuttu, S., 2001. PCR-based diagnosis of Helicobacter pylori infection and real-time determination of clarithromycin resistance directly from human gastric biopsy samples. J. Clin. Microbiol. 39, 1217 – 1220. Cloud, J.L., Hymas, W.C., Turlak, A., Croft, A., Reischl, U., Daly, J.A., Carroll, K.C., 2003. Description of a multiplex Bordetella pertussis and Bordetella parapertussis LightCycler PCR assay with inhibition control. Diagn. Microbiol. Infect. Dis. 46, 189 – 195. Cogswell, F.B., Bantar, C.E., Hughes, T.G., Gu, Y., Philipp, M.T., 1996. Host DNA can interfere with detection of Borrelia burgdorferi in skin biopsy specimens by PCR. J. Clin. Microbiol. 34, 980 – 982. Coutlee, F., de Ladurantaye, M., Tremblay, C., Vincelette, J., Labrecque, L., Roger, M., 2000. An important proportion of genital samples submitted for Chlamydia trachomatis detection by PCR contain small amounts of cellular DNA as measured by beta-globin gene amplification. J. Clin. Microbiol. 38, 2512 – 2515. Dionisi, H.M., Harms, G., Layton, A.C., Gregory, I.R., Parker, J., Hawkins, S.A., Robinson, K.G., Sayler, G.S., 2003. Power analysis for real-time PCR quantification of genes in activated sludge and analysis of the variability introduced by DNA extraction. Appl. Environ. Microbiol. 69, 6597 – 6604. Fischer-Romero, C., Luthy-Hottenstein, J., Altwegg, M., 2000. Development and evaluation of a broad-range PCR-ELISA assay with Borrelia burgdorferi and Streptococcus pneumoniae as model organisms for reactive arthritis and bacterial meningitis. J. Microbiol. Methods 40, 79 – 88. Furuta, T., Arai, H., Honda, S., Suzuki, M., Kaneko, E., 1993. Quantification of Helicobacter pylori in the gastric mucus by competitive PCR using synthetic DNA fragments. Nippon Shokakibyo Gakkai Zasshi 90, 2962. Goodwin, C.S., Blincow, E.D., Warren, J.R., Waters, T.E., Sanderson, C.R., Easton, L., 1985. Evaluation of cultural techniques for isolating Campylobacter pyloridis from endoscopic biopsies of gastric mucosa. J. Clin. Pathol. 38, 1127 – 1131. Harris, A.G., Wilson, J.E., Danon, S.J., Dixon, M.F., Donegan, K., Hazell, S.L., 2003. Catalase (KatA) and KatA-associated protein (KapA) are essential to persistent colonization in the Helicobacter pylori SS1 mouse model. Microbiology 149, 665 – 672. Hazell, S.L., Borody, T.J., Gal, A., Lee, A., 1987. Campylobacter pyloridis gastritis: I. Detection of urease as a marker of bacterial colonization and gastritis. Am. J. Gastroenterol. 82, 292 – 296.

Y. Roussel et al. / Journal of Microbiological Methods 62 (2005) 71–81 He, Q., Wang, J.P., Osato, M., Lachman, L.B., 2002. Real-time quantitative PCR for detection of Helicobacter pylori. J. Clin. Microbiol. 40, 3720 – 3728. Huijsdens, X.W., Linskens, R.K., Koppes, J., Tang, Y.L., Meuwissen, S.G., Vandenbroucke-Grauls, C.M., Savelkoul, P.H., 2004. Detection of Helicobacter species DNA by quantitative PCR in the gastrointestinal tract of healthy individuals and of patients with inflammatory bowel disease. FEMS Immunol. Med. Microbiol. 41, 79 – 84. Jiang, W., Baker, H.J., Smith, B.F., 2003. Mucosal immunization with helicobacter, CpG DNA, and cholera toxin is protective. Infect. Immun. 71, 40 – 46. Katcher, H.L., Schwartz, I., 1994. A distinctive property of Tth DNA polymerase: enzymatic amplification in the presence of phenol. BioTechniques 16, 84 – 92. Lachman, L.B., Ozpolat, B., Rao, X.M., Graham, D.Y., Osato, M., 1997. Development of a murine model of Helicobacter pylori infection. Helicobacter 2, 78 – 81. Lantz, P.G., Abu al-Soud, W., Knutsson, R., Hahn-Hagerdal, B., Radstrom, P., 2000. Biotechnical use of polymerase chain reaction for microbiological analysis of biological samples. Biotechnol. Annu. Rev. 5, 87 – 130. Lee, A., O’Rourke, J., De Ungria, M.C., Robertson, B., Daskalopoulos, G., Dixon, M.F., 1997. A standardized mouse model of Helicobacter pylori infection: introducing the Sydney strain. Gastroenterology 112, 1386 – 1397. Marchetti, M., Arico, B., Burroni, D., Figura, N., Rappuoli, R., Ghiara, P., 1995. Development of a mouse model of Helicobacter pylori infection that mimics human disease. Science 267, 1655 – 1658. Mikula, M., Dzwonek, A., Jagusztyn-Krynicka, K., Ostrowski, J., 2003. Quantitative detection for low levels of Helicobacter pylori infection in experimentally infected mice by real-time PCR. J. Microbiol. Methods 55, 351 – 359.

81

Nakayama, Y., Graham, D.Y., 2004. Helicobacter pylori infection: diagnosis and treatment. Expert Rev. Anti. Infect. Ther. 2, 599 – 610. Scott, J.A., Marston, E.L., Hall, A.J., Marsh, K., 2003. Diagnosis of pneumococcal pneumonia by psaA PCR analysis of lung aspirates from adult patients in Kenya. J. Clin. Microbiol. 41, 2554 – 2559. Straubinger, R.K., Greiter, A., McDonough, S.P., Gerold, A., Scanziani, E., Soldati, S., Dailidiene, D., Dailide, G., Berg, D.E., Simpson, K.W., 2003. Quantitative evaluation of inflammatory and immune responses in the early stages of chronic Helicobacter pylori infection. Infect. Immun. 71, 2693 – 2703. Tebbe, C.C., Vahjen, W., 1993. Interference of humic acids and DNA extracted directly from soil in detection and transformation of recombinant DNA from bacteria and a yeast. Appl. Environ. Microbiol. 59, 2657 – 2665. Ursi, D., Dirven, K., Loens, K., Ieven, M., Goossens, H., 2003. Detection of Mycoplasma pneumoniae in respiratory samples by real-time PCR using an inhibition control. J. Microbiol. Methods 55, 149 – 153. Walsh, P.S., Metzger, D.A., Higuchi, R., 1991. Chelex 100 as a medium for simple extraction of DNA for PCR-based typing from forensic material. BioTechniques 10, 506 – 513. Wilcox, M.H., Reynolds, M.T., Hoy, C.M., Brayson, J., 2000. Combined cervical swab and urine specimens for PCR diagnosis of genital Chlamydia trachomatis infection. Sex Transm. Infect. 76, 177 – 178. Zipper, H., Buta, C., Lammle, K., Brunner, H., Bernhagen, J., Vitzthum, F., 2003. Mechanisms underlying the impact of humic acids on DNA quantification by SYBR Green I and consequences for the analysis of soils and aquatic sediments. Nucleic Acids Res. 31, e39.