Evaluation of a PCR Primer Based on the Isocitrate Dehydrogenase Gene for Detection of Helicobacter pylori in Feces

Abstract

In order to improve detection and identification of Helicobacter pylori in highly contaminated samples, we evaluated new specific primers based on the DNA base sequence within the isocitrate dehydrogenase (icd) gene to amplify a 1,200-bp DNA segment. The specificity of the icd primer was tested against DNA derived from various bacteria, including 7 Helicobacter species and a panel of 1 gram-variable, 2 gram-positive, and 16 gram-negative bacteria, as well as DNA from houseflies and feces from H. pylori-negative patients. The primers permitted the detection of all clinical H. pylori isolates tested, but no reactions were observed with negative controls. Several procedures for DNA extraction from feces were evaluated using PCR with icd primers. The lower limits of detection of H. pylori DNA from two different sources containing the same number of H. pylori organisms, a pure culture and feces spiked with H. pylori, were established for each extraction method tested. The results were 8.0 × 10³ CFU/ml for cultures of pure H. pylori, and 8.0 × 10⁶ CFU/ml for H. pylori from feces, using the phenol-chloroform method; 8.0 × 10² and 7.0 × 10³ CFU/ml, respectively, for a glass matrix and chaotropic solution protocol; 8.0 × 10² and 7.0 × 10³ CFU/ml, respectively, for the QIAamp tissue kit; and 5.0 × 10² and 5.0 × 10³ CFU/ml, respectively, for the XTRAX DNA extraction kit. We conclude that the use of the icd gene as a primer for PCR represents a specific and sensitive assay for detection of H. pylori in highly contaminated samples.

Helicobacter pylori is strongly associated with gastroduodenal disease, including chronic active gastritis, duodenal ulcers, and gastric adenocarcinoma (1). Several reliable methods for detecting H. pylori are used at present; most of them are performed on gastric biopsy samples. Since these methods are thus invasive and expensive, less-invasive methods such as serologic examination of blood and the urea breath test are becoming more popular (27). However, these noninvasive tests have drawbacks: positive results by blood serology do not necessarily indicate current infection by H. pylori (D. J. Cullen, K. J. Cullen, B. J. Collins, K. J. Christiansen, and J. Epis, Letter, Lancet 340:1161–1162, 1992), while urea breath tests require expensive, specialized equipment and reagents (2). Considering these drawbacks of such noninvasive tests, alternative methods with feces appear attractive.

In a limited number of studies, infected individuals have been shown to excrete H. pylori in feces, since PCR, enzyme immunoassay (EIA), or even culture could detect the organisms in stool specimens (7–9, 20, 23). However, many others have not yet confirmed these findings. In fact, culture of H. pylori is extraordinarily problematic due to the complexity of the fecal microflora, making direct culture of H. pylori difficult and subject to misinterpretation. Further limiting factors of isolation are technical, such as the requirement for a selective medium to culture H. pylori reliably from fecal samples. Hence, PCR has been used to detect H. pylori DNA; however, it has been shown to have its own limitations. For example, extraction of PCR-amplifiable DNA from feces has been a complicated and lengthy process (20). A variety of inhibitors present in feces can lead to erratic results (12, 24). In addition, the correct choice of highly specific and sensitive primers and the correct conformation of the amplified product are clearly essential for optimal sensitivity and specificity (10). We have recently isolated and expressed the isocitrate dehydrogenase gene (icd) of H. pylori (6). The DNA sequence of the icd gene was 96% identical to that present in the total-genome sequence of H. pylori strain 26695 (21) and only 64% homologous to the icd of Escherichia coli, 64% homologous to Vibrio sp. icd, and 63% homologous to Bacillus subtilis icd. Furthermore, the 3′ sequence was unique based on gene bank analysis. The aim of our study was to evaluate several procedures for extracting H. pylori DNA from feces and to evaluate newly designed primers based on the DNA sequence within the icd gene for PCR. An effective combination of DNA extraction and amplification may improve detection of H. pylori in the environment and in clinical settings.

MATERIALS AND METHODS

Controls for primer specificity.

Bacterial strains, flies, and fecal samples used as controls are listed in Table 1.

TABLE 1.

Specificity of icd primers in detecting H. pylori DNA

Control specimen	Amplification with icd primersa
Arcobacter butzleri	−
Bacteroides fragilis	−
Campylobacter coli	−
Campylobacter jejunii	−
Campylobacter lari	−
Campylobacter coli	−
Clostridium perfringens	−
Escherichia coli	−
Enterococcus faecalis	−
Gardnerella vaginalis	−
Haemophilus influenzae	−
Haemophilus parahaemolyticus	−
Helicobacter spp.
H. bilis	−
H. canis	−
H. felis	−
H. hepaticus	−
H. mustelae	−
H. pullorum	−
H. pylori (10 isolates)	+
H. rodentium	−
Providencia rettgeri	−
Salmonella sp.	−
Serratia marcescens	−
Shigella flexneri	−
Streptococcus pneumoniae	−
Yersinia enterocolitica	−
Vibrio parahaemolyticus	−
Control flies
Spiked with H. pylori	+
Nonspiked	−
Feces
Spiked with H. pylori	+
Nonspiked	−

^a+, target amplification; −, no target amplification. 

Clinical isolates of H. pylori from 10 patients with duodenal ulcers were grown on Brucella agar (Difco, Detroit, Mich.) with lysed sheep blood (5% [vol/vol]; Remel, Lenexa, Kans.) and were incubated under a microaerophilic atmosphere at 37°C for 72 h. Campylobacter spp. were also grown on this medium at 42°C for 2 days. All other bacteria were grown on blood agar plates at 37°C for 24 h; for Enterococcus faecalis, a 5% CO₂ atmosphere was used. Stock cultures were maintained at −80°C in 20% (vol/vol) glycerol-Brucella broth.

Seven different Helicobacter species were cultured like the clinical isolates above in order to assess the specificity of our icd primers. Furthermore, we tested the icd primers against houseflies (Musca domestica) and human feces. Fecal samples were obtained from patients proven H. pylori negative by histology of endoscopically obtained gastric biopsy specimens and were tested individually. In this study, feces from H. pylori-positive patients was not used for the evaluation of the sensitivity and specificity of our PCR methods, since the number of H. pylori organisms in these samples is unknown. The purpose of our work was to evaluate our primers and DNA extraction methods; therefore, samples with defined H. pylori loads were used.

Flies were chosen as controls because they contain a large variety of viral, bacterial, and parasite DNA, as well as food and host DNA (5). In addition, we recently reported that flies are able to carry and excrete viable H. pylori and may therefore represent a vector for the transmission of H. pylori (4). Several hundred housefly pupae (Carolina Biological Supply Company, Burlington, N.C.) were placed in well-ventilated autoclaved buckets maintained at 25°C. Emerging adult flies were removed, and pools of five flies were transferred into sterile test containers and frozen at −80°C until subsequent preparation for PCR. For DNA extraction, flies were thawed. The abdomen of each fly was separated aseptically from the chest under a dissecting microscope. The bodies of each fly pool were homogenized in 0.5 ml of sterile water by using a sterile glass grinder. All bacterial and fecal samples, as well as fly homogenates, were processed by using the XTRAX kit (see below) to purify genomic DNA for PCR.

PCR.

PCR was used to identify the icd gene of H. pylori in our samples. Chromosomal DNA was prepared as described under “DNA extraction and method sensitivity” below. Oligonucleotide primers (Tufts University Protein and Nucleic Acid Analysis Unit, Boston, Mass.), were specific for the 5′ and 3′ termini of the icd gene respectively, with the sequences 5′-ATGGCTTACAACCCTAAAATTTTACAAAAGCC-3′ and 5′-TCACATGTTTTCAATCATCACGC-3′. The reaction was performed in a volume of 100 μl, comprising 50 pM each primer, 10 μl of chromosomal DNA, 2 U of Taq DNA polymerase (Boehringer Mannheim Biochemicals), 0.2 mM deoxynucleoside triphosphates, and 1.5 mM MgCl₂, in a single block thermal cycler (Ericomp, San Diego, Calif.). Cycle profiles were as follows: DNA denaturation at 94°C for 5 min, followed by 30 cycles consisting of 1-min denaturation at 94°C, 2-min annealing at 55°C, and 2-min extension at 72°C. Samples were run simultaneously with a clinical H. pylori isolate as a positive control. Negative-control reactions with distilled water were performed with each batch of amplification to exclude the possibility of contamination. The specificity of icd primers was compared with that of species-specific protein antigen primers, which have been reported to be highly specific for H. pylori (14). Species-specific protein antigen and icd primers were tested against control flies, H. pylori, and the non-Helicobacter control bacteria listed in Table 1. In addition, icd primers were studied with seven Helicobacter species. PCR products were analyzed by electrophoresis of a 10-μl aliquot using a 1.3% (wt/vol) agarose gel. Bands were visualized by excitation under UV light after staining with 1 μg of ethidium bromide/ml.

Target DNA sequence determination.

DNA sequencing was performed on the amplified products in order to eliminate false-positive results. DNA was extracted from 90 μl of the reaction mixture by using a PCR purification kit (QIAGEN, Hilden, Germany). Double-stranded DNA PCR was carried out by the dideoxy chain termination method (15) with Taq DNA polymerase (GIBCO BRL) by using the same oligonucleotide primers specific for the icd gene as in the PCR.

DNA extraction and method sensitivity.

In the present study, four methods to purify DNA from one pure clinical H. pylori isolate in suspension as well as from spiked fecal samples were compared, as follows.

(i) The classical phenol-chloroform method (11).

Protein was denatured using STE buffer with 1% sodium dodecyl sulfate and 25 μg proteinase of K, and DNA was concentrated in the aqueous layer using phenol-chloroform. Ethanol was then used to remove residual phenol-chloroform and salts and to precipitate the DNA.

(ii) Glass matrix and chaotropic solution (18).

Guanidine-thiocyanate was used to lyse cells and denature proteins. DNA was bound to glass beads (GlasPac; National Scientific Supply Company, San Rafael, Calif.), and the suspension was washed in wash buffer (50% ethanol, 10 mM Tris-HCl [pH 7.5], 100 mM NaCl).

(iii) QIAamp tissue kit (Qiagen).

Proteins were denatured using proteinase K. DNA was bound to the silica membrane of spin columns, leading to the removal of other contaminants by washing with buffers provided in the QIAamp kit.

(iv) XTRAX DNA extraction kit (Gull Laboratories, Salt Lake City, Utah) (26).

Samples were microwaved in the presence of an extraction buffer containing detergents and salts, followed by centrifugation and DNA precipitation with isopropyl alcohol.

DNA was extracted from an H. pylori dilution series (10⁰ to 10⁻⁶, beginning with an optical density at 600 nm [OD₆₀₀] of 1.0), using 100 μl of a pure H. pylori suspension in one series and 50 μl of the creamy layer of a stool sample spiked with 100 μl of bacterial suspension in the other. The creamy layer contains fecal bacteria in a more concentrated form and was prepared as described by Thomas (19). In brief, stool samples were centrifuged (7,000 × g for 10 min) to separate three identifiable layers: a liquid supernatant on top, a creamy layer in the middle, and a solid layer on the bottom.

Quantification of H. pylori present in bacterial suspensions and in aliquots used to spike stool samples was performed by plate counts on Skirrow's agar (Difco, San Jose, Calif.).

PCR using the icd primers was performed to determine the sensitivities of the extraction methods (see the protocols above).

RESULTS

Our icd primers proved to be very specific for H. pylori only and did not result in false-positive results with any of the other bacterial species, including seven Helicobacter species (Table 1). In addition, these primers specifically amplified H. pylori DNA from feces and fly abdomens, which represent a pool of heterogeneous DNA. This underlines the usefulness of these primers in the detection of H. pylori in highly contaminated samples, such as flies and stool. In contrast, the species-specific antigen primers (14) produced false-positive signals in our initial screening tests and reacted with negative-control flies, Campylobacter jejuni, and E. faecalis (data not shown).

The DNA sequence of the amplified target DNA of our controls confirmed the PCR products with more than 96% homology to the icd gene. The control clinical isolates of H. pylori showed 95% to 97% homology to the icd gene of reference strain 26695 (21). The fact that the sequences are not identical minimizes the chance that contamination occurred during sample preparation.

In comparison with two other DNA extraction techniques, the QIAamp kit and the XTRAX kit proved to be the preferred methods due to their sensitivity and the fact that complete kits reduce the length and number of preparatory steps (see Table 2). Extraction using guanidine-thiocyanate, however, was as sensitive as the QIAamp and XTRAX techniques for pure H. pylori as well as for bacteria from feces but was less convenient, since it required more preparation. Phenol-chloroform extraction recovered the smallest amount of DNA from pure and fecal H. pylori, and exposes one to harmful fumes (Table 2).

TABLE 2.

Comparison of DNA extraction methods for detection of H. pylori in pure solution and in human feces

Method	Sensitivity (CFU/ml) for:		Duration (min)	Cost/sample ($)	Convenience
	Pure H. pylori	H. pylori and feces
Phenol-chloroform	8.0 × 103	8.0 × 106	600	0.50	Low
Guanidine	8.0 × 102	7.0 × 103	180	1.90	Moderate
QIAamp	8.0 × 102	7.0 × 103	330	1.80	High
XTRAX	5.0 × 102	5.0 × 103	60	1.60	High

DISCUSSION

This study shows that the icd gene PCR assay is a sensitive and specific method for the detection of H. pylori in feces. Comparison of four DNA extraction protocols that included columns, centrifugation, glass matrices, or patented extraction matrices showed that the QIAamp method and XTRAX kit were the best procedures due to their sensitivity and convenience of execution.

Isolation of H. pylori from feces as a diagnostic tool or for research purposes for detecting colonization is attractive because it is noninvasive. However, culture of H. pylori from patients' fecal samples has been problematic because currently available selective culture media for isolation of H. pylori are frequently overgrown by other, faster-growing gram-negative bacteria. Similarly, detection of H. pylori in feces by using PCR and standard extraction methods has proved difficult, producing erratic results due to a variety of fecal inhibitors, such as acidic polysaccharides, metabolic products, and large amounts of irrelevant DNA (12). Furthermore, previous PCR analysis for H. pylori in feces has shown low sensitivity (24) due to by multiple factors such as variations in the sensitivity and amplification performance of the polymerase and in the annealing affinity of the chosen primers to their target sequence (17). For this reason, DNA extraction procedures that remove inhibitors of PCR and concentrate the pathogen or total DNA were developed. In concert with these effective techniques, our highly specific icd primer represents a specific and sensitive method for detection of H. pylori.

Makristathis et al. (8) described species-specific protein antigen primers as specific for H. pylori, using only one other Helicobacter species to assess specificity. Our screening experiments, which included all the bacterial species listed in Table 1, as well as flies, showed that the use of these primers resulted in false-positive amplifications with control flies, C. jejuni, and E. faecalis. It is therefore possible that Makristathis' assay may have amplified non-H. pylori DNA, resulting in false positives. We demonstrated that the icd primers were more specific for H. pylori than the species-specific protein antigen primers when tested against a variety of DNA samples. Others have used primers based on the 16S rRNA, which have also displayed good specificity when tested against many microorganisms, including several Helicobacter species (16, 25), although no studies have been performed to assess the detection limit of these primers for H. pylori in feces.

The minimum number of organisms in feces required for a positive PCR result was 3.3 × 10⁴ CFU/ml of stool in our most sensitive assay, which is comparable to other reported sensitivities (24). Despite concentrating H. pylori organisms by immunomagnetic-bead separation techniques, Osaki et al. showed comparable detection limits of 2 × 10⁴ CFU for fecal samples obtained from gnotobiotic mice (13). Similar numbers, 1.8 × 10⁴ CFU, are required for the recently developed stool antigen test (HpSA-EIA) (3). The HpSA-EIA is another attractive, noninvasive fecal test that seems to predict H. pylori status reliably (23), but it has been less suitable for evaluating the outcome of eradication therapy (22).

Another variable that may limit routine clinical use of fecal assays is the fact that H. pylori proteins or DNA, to be tested, have to pass unharmed through the entire gastrointestinal tract despite exposure to numerous digestive enzymes and bacteria. Furthermore, it is unknown how many H. pylori organisms an infected individual excretes, whether there is patchy distribution in a bowel movement, and how long H. pylori fragments remain in the intestinal tract after eradication therapy. Two recent studies showed that there was still significant detection of H. pylori debris by PCR as well as EIA at a 1-month follow-up for patients from whom H. pylori had been eradicated (8, 22). Hence, more data on the intestinal elimination kinetics of H. pylori are necessary in order to use molecular or immunoassay tests for evaluation of eradication treatment.

In conclusion, the icd gene assay has been shown to allow differentiation of H. pylori from seven other Helicobacter species. This newly designed primer pair may be useful for PCR used in epidemiological studies (P. Grübel, L. Huang, N. Masubuchi, F. J. Stutzenberger, and D. R. Cave, Letter, Lancet 352:788–789, 1998). In a next step, we will test the icd primers on feces from H. pylori-positive patients to evaluate their potential for clinical application.