Lyophilization Prior to Direct DNA Extraction from Bovine Feces Improves the Quantification of Escherichia coli O157:H7 and Campylobacter jejuni

Lyophilization was used to concentrate bovine feces prior to DNA extraction and analysis using real-time PCR. Lyophilization significantly improved the sensitivity of detection compared to that in fresh feces and was associated with reliable quantification of both Escherichia coli O157:H7 and Campylobacter jejuni bacteria present in feces at concentrations ranging between 2 log₁₀ and 6 log₁₀ CFU g⁻¹.Bovines are a reservoir for verotoxigenic Escherichia coli O157:H7 and Campylobacter jejuni, pathogenic microorganisms responsible for severe human gastrointestinal disease (5, 12). Qualitative and quantitative detection of these organisms in bovine feces is essential for evaluating risk to human health. Real-time PCR (quantitative PCR qPCR]) assays have been developed to detect and quantify both E. coli O157:H7 and C. jejuni bacteria by using DNA directly extracted from animal feces (20, 22). Analysis of DNA extracted from bovine feces can generate a high level of correlation between the actual target cell density and the PCR signal (7, 8). However, the detection of E. coli O157:H7 and C. jejuni by direct DNA extraction is less sensitive and more variable than detection by procedures based on a preliminary enrichment step (e.g., laboratory culture) (7, 9, 16, 20). We explored the potential of lyophilization for improving overall detection by qPCR through increasing the amount of bovine fecal material available for DNA extraction.Four sets of five fresh bovine fecal samples were collected, and each sample was divided into four equal portions. Samples were seeded with either (i) E. coli O157:H7 (strain NZRM 3614) grown for 18 h at 37°C in tryptic soy broth (BD, Sparks, MD) or (ii) C. jejuni (strain NZRM 1958) grown for 48 h at 42°C in Exeter broth (11) to obtain the following concentrations: set 1, 0 CFU g⁻¹ (unseeded) and 3.5 log₁₀, 4.5 log₁₀, and 5.5 log₁₀ CFU of E. coli O157:H7 g⁻¹, and set 2, 0 CFU g⁻¹ (unseeded) to 5.2 log₁₀ CFU of E. coli O157:H7 g⁻¹. Set 3 and 4 concentrations varied from 0 CFU g⁻¹ (unseeded) to 6.4 log₁₀ C. jejuni CFU g⁻¹. DNA was either extracted directly from fresh samples or extracted from samples after lyophilization. Lyophilization involved mixing of prepared fecal samples in phosphate-buffered saline (145 mM NaCl, 59 mM Na₂HPO₄, 8 mM KH₂PO₄, pH 7.5) at a ratio of 1:10 (wt/vol), homogenization with a lab blender model 400 (Seward Medical, London, United Kingdom), cooling to −35°C, and concentration using a 1015GP lyophilizer (Cuddon Ltd., Blenheim, New Zealand). Total DNA was extracted from 0.2 g of a fresh or lyophilized fecal sample by using a QIAamp DNA stool minikit (Qiagen Inc., Mississauga, Canada). DNA was amplified using either a TaqMan E. coli O157:H7 detection kit (Applied Biosystems, Foster City, CA) or mapA primers and a corresponding probe (1). Amplification and fluorescence data were collected with optical-grade 96-well plates by using a TaqMan 7300 PCR system (Applied Biosystems). For each DNA sample, a mean threshold cycle (C_T) value for triplicate qPCR runs was calculated. When no C_T value was obtained, an arbitrary C_T value of 40 was assigned. All data were reported as equivalent concentrations in fresh feces. Significance levels were determined by one-way analysis of variance. The relationship between the log₁₀ numbers of CFU g⁻¹ fresh feces (viable-cell counts) and C_T values was analyzed using GenStat software (version 10.2.0.175; VSN International, Oxford, United Kingdom). Confidence intervals were obtained using the software program Flexi (21).Lyophilized samples were associated with significantly improved sensitivity (P < 0.001) at seeding levels of 4.5 and 5.5 log₁₀ E. coli O157:H7 CFU g⁻¹ (Table ​(Table1).1). At 3.5 log₁₀ CFU g⁻¹, the rate of E. coli O157:H7 detection was also higher, with all lyophilized samples producing a C_T value of <40 (Table ​(Table1).1). Individual C_T values for the three qPCR amplification runs were sufficiently similar to allow averaging (P > 0.05). Regression analysis of the averaged set 2 and 3 data (Fig. ​(Fig.1)1) demonstrated that the detection of both E. coli O157:H7 and C. jejuni was linear for seeding levels ranging from ca. 2 log₁₀ to 6 log₁₀ CFU g⁻¹ fresh feces. The range of concentrations used reflects the reported range of concentrations of these bacteria in feces (i.e., 0 to 6 log₁₀ CFU g⁻¹) as determined by conventional culture (3, 4, 18, 19). The high coefficients of correlation for the relationships between the log₁₀ numbers of CFU g⁻¹ feces and the C_T values indicated the specific amplification of the target DNA. The reproducibility of detection of E. coli O157:H7 was reduced at the lowest seeding concentration (i.e., 2.2 log₁₀ CFU g⁻¹ feces), with 75% of the samples giving a C_T value of <40. The limit for 100% successful detection after lyophilization was 2.9 log₁₀ E. coli O157:H7 CFU g⁻¹. The detection of C. jejuni by qPCR varied between sets. For set 3, 100% reproducibility occurred at 2.2 log₁₀ C. jejuni CFU g⁻¹. For set 4, satisfactory detection was obtained only after dilution of the DNA extract prior to qPCR. Despite this requirement for dilution, C. jejuni was still detected in 80% of the samples of set 4 seeded at a density of 2.2 log₁₀ C. jejuni CFU g⁻¹.Open in a separate windowFIG. 1.Ranges of quantification of E. coli O157:H7 (A) and C. jejuni (B) bacteria obtained from lyophilized fecal samples by real-time PCR. Each point represents the average C_T value for triplicate runs of one fecal sample at one seeding concentration. The hatched areas represent the 95% confidence intervals.

TABLE 1.

Difference in C_T values obtained for real-time PCR detection of E. coli O157:H7 in seeded fecal samples (n = 5) with and without lyophilization

Seeding level (log10 CFU g−1 fresh feces) or status	Average CT value (range)
	Without lyophilization	With lyophilization
5.5	31.50 (31.02-32.18)	28.34 (28.04-29.03)
4.5	34.79 (33.43-35.75)	31.33 (31.01-31.89)
3.5	35.45a	33.52 (33.21-33.87)
Unseeded	>40	>40

Open in a separate window^aOnly one fecal sample gave a C_T value of <40.Overall, the removal of water by lyophilization provided an approximately 10-fold increase in the amount of fecal material used. Consequently, the test sensitivity was 10-fold greater than that reported previously (17, 7). Lyophilization of feces has been reported to be useful for PCR-based studies of pigs (14), and our results indicate a useful role for the quantification of E. coli O157:H7 bacteria in cattle feces. Indeed, the slopes and the linear regression coefficients for the qPCR signal (C_T values) and the known concentrations of microbial pathogen cells in the feces are in agreement with published values (2). Our methodology shows a lower limit of C. jejuni quantification by qPCR (ca. 2 log₁₀ CFU g⁻¹ in seeded fresh feces) than that reported previously (8), demonstrating the usefulness of lyophilization to improve detection and quantification of bacteria in feces.In our study, the accurate detection of C. jejuni after DNA extraction from lyophilized feces was adversely affected for some samples. Interference due to partial removal of PCR inhibitors after DNA extraction using the QIAamp DNA stool minikit has been reported by other workers (10, 15). For lyophilized samples, the inhibition was successfully overcome by dilution of DNA. Recent reports confirmed the importance of diluting DNA (up to 3 log) to increase the accuracy of detection by real-time PCR (6, 13). Lyophilization presents the advantage that lyophilized material can be stored for long periods at room temperature, is easy to transport, and can also be used for complementary chemical analysis.