A rapid multiplex real-time PCR detection of toxigenic Clostridioides difficile directly from fecal samples

This study developed a new single-tube multiplex real-time PCR method for detecting toxigenic C. difficile directly from fecal samples using tcdA, tcdB, cdtB, and internal gene tpi as targets, which could be performed on kinds of polymerase chain reaction device including point-of-care testing (POCT), with improved detection efficiency. The specificity, sensitivity, and repeatability of each gene was evaluated using 69 C. difficile isolates and 74 fecal samples. Results were compared with established PCR, qPCR, and ELISA methods. Interspecies specificity was 100% based on six common intestinal pathogens (Escherichia coli, Enterococcus Faecium, Enterococcus faecalis, Clostridium perfringens, Bacteroides fragilis, Clostridium botulinum). The lower detection limit (LDL) for tcdA, tcdB, and cdtB with pure C. difficile DNA was 101,100, and 100 copies/μL, respectively, the coefficients of variation among different experimental batches and within each experimental batch were both less than 3%, which shows that this method has strong repeatability. And the LDL of fecal DNA was 5 × 100, 5 × 103, and 5 × 102 colony-forming units (CFU)/g, respectively. In addition, the efficiency for detection of tcdA was compared with established PCR and real-time PCR methods, demonstrating high consistency (98.4%) and similar sensitivity. ELISA was used to confirm inconsistent results, which were identical with our method. The sensitivity and specificity for detecting toxigenic C. difficile in fecal samples were 96.49% and 94.12% compared with the toxigenic culture (TC). This method effectively identified the toxigenic and non-toxigenic strains with high specificity, sensitivity, and repeatability, and could reduce the false positive rate of tcdA, and accurately identify the typical Asian strain RT017, making it potentially contribute to the surveillance of CDI in China. Supplementary Information The online version contains supplementary material available at 10.1007/s13205-022-03434-6.

. In summary, CDI poses a significant burden on both life health and social resources, which should be of great concern.
Toxins produced by C. difficile play an important role in the pathogenic process and two major toxins are produced: toxin A (enterotoxin) and toxin B (cytotoxin), which can be produced alone or simultaneously. Moreover, a third toxinbinary toxin is found in some C. difficile isolates, such as hypervirulent RT027 / NAP1 / BI (Nagy 2018) and RT078. Although the exact role of binary toxin in pathogenesis is unclear, but it is thought to be associated with high incidence, recurrence, and mortality of CDI (Bacci et al. 2011;Stewart et al. 2013). However, typical Asian strain RT017( toxin A-negative toxin B-positive) causing several outbreaks worldwide, has also been identified in recent years (Drudy et al. 2007).
There are three commonly used methods for the identification of toxigenic C. difficile: (i) TC (toxigenic culture) is a two-step method that combined with C. difficile culture and cell cytotoxicity assay (CCNA) (Nagy 2018); (ii) enzyme immunoassay (EIA) for toxin A/B and glutamate dehydrogenase(GDH) (Bagdasarian et al. 2015); (iii) nucleic acid amplification tests (NAATs) targeting toxin-encoding genes, which include PCR, quantitative PCR (QPCR), loopmediated isothermal amplification (LAMP), and helicasedependent isothermal DNA amplification (HAD). TC operation is complex and time-consuming, which is mainly used for epidemiological research and evaluation of new methods (Bagdasarian et al. 2015). The sensitivity and specificity of immunological testing can vary, and it must be combined with a high sensitivity-specific approach to make up for its shortcomings. Therefore, NAAT was recommended for diagnosing CDI worldwide in both two-steps and three-steps diagnosis procedures (Lee et al. 2021).
In the present study, our novel method can simultaneously detect tcdA, tcdB, and cdtB genes, which improves the accurate identification of toxigenic C. difficile isolates directly from fecal samples. First, it is noteworthy that the typical Asian strain RT017, which has an A-B + cdtA/B-profile, plays an important role in CDI epidemiology in China. A deletion of 1821 bp was found in the repeat region of tcdA. Although none toxin A was produced due to the presence of a stop codon in these A-B + isolates, but there still are DNA fragments remaining in tcdA-negative isolates (Rupnik et al. 2003), primers that circumvent this deleted region may lead to false-positive results for toxinA in this type strain. Second, other multiple qPCR methods only included tcdA and tcdB as target genes, but binary toxin genes were not included (Luna et al. 2011;Kubota et al. 2014;Bélanger et al. 2003). Additionally, since binary toxin genes cdtA and cdtB may be fused, choosing cdtA as the target gene may lead to false-negative results for binary toxin (Gerding et al. 2014). In other previously published multiple qPCR methods, cdtA rather than cdtB was selected as the target gene for binary toxin detection, which may lead to inaccurate detection of binary toxin (Kilic et al. 2015;Hoegh et al. 2012). Therefore, cdtB was selected as the target gene for binary toxin detection.
In a word, a multiplex real-time PCR method was developed to detect toxigenic C. difficile directly from fecal samples, involving tcdA, tcdB, cdtB, and internal gene tpi as targets, which could be performed on kinds of polymerase chain reaction devices including POCT for rapid detection and identification of clinical CDI cases. Furthermore, this method could reduce the false positive rate of tcdA, accurately identify the typical Asian strain RT017 with improved detection efficiency, making it potentially contribute to the surveillance of CDI in China.

Design of primers and probes
Complete sequences of tcdA (NZ_FUUL01000004.1, N C _ 0 1 7 1 7 8 . 1 , N Z _ C D N D 0 1 0 0 0 0 0 1 . 1 , N C _ 0 1 7 1 7 4 . 1 , N C _ 0 1 3 3 1 6 . 1 ) , t c d B ( N Z _ FUUL01000004.1, NC_017178.1, NZ_CDND01000001.1, NC_017174.1, NC_013316.1), and cdtB (AF271719.1, HQ639679.1, FN538970.1, NC_017178.1, NZ_ CDND01000001.1, NC_017174.1)were download from NCBI GenBank entries and aligned to determine the conserved regions. According to the consensus sequences, primers and probes were designed by Primer Express software v.3.0 (Applied Biosystems, Foster City, CA), and then were BLAST in GenBank to test the specificity. There was no crossover with other pathogens using BLAST searches in the NCBI database. All primers and probes were included in patent application CN202010825309. X. RT017 is a typical Asian strain characterized by a partial deletion in the repeat region of the tcdA gene. Due to the presence of a termination codon, it does not produce toxin A, but most of the tcdA negative isolates still possess remaining DNA fragments. Therefore, it is necessary to design primers targeting the absent region of the tcdA gene for the accurate identification of toxin A.

Construction of standard plasmids and standard curves for the three genes
The C. difficile isolate ATCC-BBA1803 (RT027) was used to amplify the three target toxin genes. PCR products were purified using an EasyPure Quick Gel Extraction Kit (Trans, China), and then ligated into the PMD18-T vector (TaKaRa, Japan), which were subsequently transformed into JM109 competent cells (TaKaRa). Positive plasmids were identified by Blue-White Screening of positive colonies and sequencing of PCR products from plasmids. The standard plasmid concentration was obtained according to the copy number conversion formula (6.02 × 10 23 copies/M × (concentration)/ (MWg/mol) = copies/ml), and ten-fold dilutions of the standard into ten gradients were used as templates to prepare the standard curve.

Evaluation of the specificity, sensitivity, and repeatability of qPCR for pure bacterial DNA
Standard strains of another six common intestinal pathogens (Escherichia coli, Enterococcus Faecium, Enterococcus faecalis, Clostridium perfringens, Bacteroides fragilis, Clostridium botulinum) were used to test interspecies specificity (Table S1). A total of 69 C. difficile isolates with different toxigenic types, including eight ATCC standard strains and 61 clinical isolates, were used to verify intraspecies specificity (Table S1). Through standard curve analysis, the pure bacterial lower detection limit (LDL), also known as the minimum concentration of detected plasmid, was determined. To assess the repeatability of this method, different concentrations (high, medium, and low copies) of three standard plasmids harboring tcdA, tcdB, and cdtB genes were tested in triplicate every other day. Then three parallel DNA samples were tested to obtain the coefficient of variation between batches and the coefficient of variation within groups (Table S2).

Preparation and detection of C. difficile-simulated fecal samples
A pure clone of the C. difficile ATCC1803 toxigenic strain was picked and mixed with 1 ml BHI (Brain Heart Infusion Broth) to achieve an McF 0.5 suspension of ~ 10 5 colonyforming units (CFU)/ml by the standard plate counting method. Bacterial suspensions at 10 0 to 10 5 CFU/ml were obtained by serial dilution, and 0.1 ml of each suspension was transferred into 0.2 mg C. difficile-negative fecal samples from healthy individuals at a final concentration of 5.0 × 10 − 1 to 5.0 × 10 4 CFU/g. Total DNA was extracted from these simulated fecal samples for qPCR evaluation of LDLs for the three target genes (tcdA, tcdB, and cdtB; Table 1).

Detection of human fecal samples using multiplex real-time PCR
A total of 74 frozen fecal samples in our laboratory were used to evaluate our qPCR detection method. A Fecal DNA Extraction Kit (TIANGEN) was used to extract fecal DNA according to the manufacturer's instructions. The results were compared with those of the gold standard (TC) method using tpi gene as internal control to evaluate the performance (Table 2).

Comparison with previously reported detection methods for tcdA and confirmation by ELISA for TcdA
Two pairs of classic PCR primers for gene tcdA (Kato et al. 1999;Lemee et al. 2004) were compared with our method using 33 C. difficile RT017 isolates and 30 other types of clinical isolates preserved in our laboratory (Table S3). For inconsistent results, ELISA was employed for clarification. Briefly, negative control (ATCC43593), positive control (ATCC-BBA1803), and tested C. difficile (N20) isolates (Table S3) were anaerobically cultured in BHI liquid medium at 37 °C for 24 h. After centrifugation, the supernatant was taken and filtered through a 0.45 μm filter, then incubated overnight at 4 °C in a 96-well plate.
The plate was washed with PBST(Phosphate Buffered Saline + Tween) and blocked with 5% bovine serum albumin at 37 °C for 2 h. Polyclonal antibody (List Biological Laboratories, Toxin A IgY, 1:1000 dilution) was used for the first hybridization, and horseradish peroxidase (HRP-conjugated goat anti-chicken IgY (Invitrogen, 1:5000 dilution) served as the secondary antibody. Finally, the results were determined using an EL-TMB Chromogenic Reagent Kit (SANGON, China) at an absorbance of 450 nm. Standard curves were prepared using Toxin A (List Biological Laboratories) as a positive control at an initial concentration of 1600 ng/ml that was diluted twice to 0.05 ng/ml. The absorbance ratio (P/N) of the tested sample (N20) and negative control (ATCC43593) was calculated, and results were considered negative at P/N < 1.5, suspicious at P/N ≥ 1.5 and < 2.1, and positive at P/N ≥ 2.1. In addition, our results were also compared with two previously reported qPCR methods (Luna et al. 2011;Kubota et al. 2014), in which our primers for gene tcdA were located in the missing region to avoid potential false positives (Fig. 1a). The qPCR conditions and mixture were previously described. The lower limits of these three pairs of primers for tcdA were tested and compared by testing the C. difficile-simulated fecal samples (Table 3) prepared as described above.
Letter Y refers to positive results and N means negative results. The red letters refer to inconsistent results  b Complete sequences of tcdA extracted from whole genomes of 33 RT017 isolates and the locations of primers were synthesized and tested in this study. M68 (NC_017175.1), CF5 (NC_017173.1) and CD630 (NC_009089.1) were used as reference strains

Flow chart
A flow chart for better understanding of the manuscript (Fig. 2).

The real-time PCR method is specific, sensitive, and repeatable
Standard strains of the other six common intestinal pathogens were tested using our primers and probes, and no specific products were amplified (Table S1). ATCC Standard strains of different toxigenic types of C. difficile were used to assess intraspecies specificity, and the results were consistent with the defined toxin profiles (Table S1). According to the above results, the specificity of this method was 100%, confirming its ability to detect toxigenic C. difficile strains accurately and selectively (Fig. 3a). Standard curves of the three target genes (tcdA, tcdB, and cdtB) were constructed using plasmids diluted tenfold as templates for qPCR detection (Fig. 3b). The three plasmids were linearly correlated with their corresponding CT (Cycle threshold) values at 10 0 − 10 9 copies/μl. Using the lowest concentration of each plasmid that could be detected as the LDL, we verified that the detection limits of DNA from C. difficile isolates for tcdA, tcdB, and cdtB were 10 1 , 10 0 , and 10 0 copies/ μl, respectively. PCR amplification efficiency for each gene was 102%, 103%, and 102%, respectively, and correlation coefficients reached 0.998 (Fig. 3b).
To evaluate the repeatability of this method, coefficients of variation among different experimental batches were calculated. Three replicate samples for each gene within each experiment batch were also tested. The coefficients of variation among different experimental batches and within each experimental batch were both less than 3%, which shows that this method has strong repeatability ( Table S2).

Detection of toxigenic C. difficile in simulated fecal samples
When 0.2 g of feces was mixed with 10 0 CFU bacteria, the method could stably detect tcdA, and the bacteria concentration in feces was 5.0 × 10 0 CFU/g. Gene tcdB could be stably detected when 0.2 g of feces was mixed with 10 3 CFU bacteria, and the bacterium content in feces was 5.0 × 10 3 CFU/g. And when 0.2 g of feces were mixed with 10 2 CFU bacteria, the method could stably detect cdtB, and the bacteria content in feces was 5.0 × 10 2 CFU/g (Table 1). Therefore, the LDL of simulated fecal samples for the three genes was 5 × 10 0 , 5 × 10 3 , and 5 × 10 2 colony-forming units (CFU)/g, respectively.

The real-time PCR method detects toxigenic C. difficile in human fecal samples
A total of 74 fecal samples, including 57 samples with toxigenic C. difficile and 17 samples with non-toxigenic C. difficile, all confirmed by TC, were randomly selected to evaluate our method (Table 2). Using our method, we detected 56 toxin-positive samples, including one false-positive, and 18 toxin-negative samples, including two false-negative (Table 2). Therefore, the true-positive and true-negative values were 96.49% (55/57) and 94.12% (16/17), respectively.

Evaluation of the detection of the tcdA gene
The developed method demonstrated high consistency (98.4%) with two sets of standard PCR primers reported previously for detecting tcdA in 63 C. difficile isolates, and only one strain (N20) yielded different results (Table S3). Therefore, the gold standard ELISA method for detecting toxin A in isolate N20 was used for further confirmation, from which the standard curve shows that the concentration of TcdA was linear at eight concentrations (0.4 ng/ml, 0.8 ng/ml, 1.6 ng/ ml, 3.2 ng/ml, 6.4 ng/ml, 12.8 ng/ml, 25 ng/ml, and 50 ng/ ml). The ratio of the absorbance between N20 and the negative control, P(0.363)/N (0.15), was 2.42, confirming that isolate N20 is a TcdA-positive strain, with a corresponding toxin A concentration was 3.53 ng/ml (Fig. 4). The ELISA results support this study's detection results for the tcdA gene.

Discussion
Toxigenic culture and CCNA (cell cytotoxicity assay) are considered the gold standards for CDI detection (Crobach et al. 2009). However, the CCNA procedure is complicated, time-consuming, and limited by cell types and equipment (Buss et al. 2015). Nowadays, immunological techniques are used as the preliminary screening method for CDI, but the sensitivity and specificity can be unstable and the results must be interpreted together with those of PCR approaches (Guery et al. 2019). In recent years, NAAT has been recommended for auxiliary diagnosis of CDI cases due to its high sensitivity, specificity, and time-saving. To date, at least 15 kinds of C. difficile nucleic acid detection methods have been approved by the United States Food and Drug Administration (FDA), of which 12 methods rely on specific instruments, 6 methods only detect the tcdB gene, and no methods detect tcdA, tcdB, and cdtB simultaneously (https:// www. fda. gov/ medic al-devic es/ vitro-diagn ostics/ nucle ic-acid-based-tests).
Our novel method can simultaneously detect tcdA, tcdB, and cdtB genes, which improves the accurate identification of toxigenic C. difficile isolates directly from fecal samples. In four previously reported qPCR methods, primers for tcdA were not located in the deletion region (Fig. 1a) (Kilic et al. 2015;Hoegh et al. 2012;Avbersek et al. 2011), which might cause the inaccurate identification of tcdA-negative Fig. 3 Specificity and sensitivity of the developed qPCR method for detecting tcdA, tcdB, and cdtB. a The specificity for detecting the three target genes was determined by comparing them with another six intestinal bacteria. b Standard curves and amplification curves for the three target genes isolates including RT017/ST37. In the present study, primers for tcdA were situated in the deletion region (7692-7965; Fig. 1a), and results with these primers were highly consistent (98.4%) with those of standard PCR primers reported previously for detecting tcdA in 63 C. difficile isolates, except for only one strain (N20) showing different results (Lemee et al. 2004;Kato et al. 1999) (Table S3). ELISA of the inconsistent N20 strain confirmed the results for our method (Fig. 4).
According to previously reported qPCR methods for detecting toxigenic C. difficile isolates, two pairs of tcdA primers from Luna and Kubota are located in the missing region, which were selected and compared with our method. The LDLs for these three qPCR methods were compared using simulated fecal samples. The LDL for tcdA in fecal samples was 2.5 × 10 2 CFU/ml according to Luna et al., while the LDL reported by Kubota et al. was 10 3 cells/g (Luna et al. 2011;Kubota et al. 2014). The LDLs for the methods of Kubota and Luna were 5.0 × 10 0 CFU/g and 5.0 × 10 1 CFU/g, respectively (Table 3), and our method yielded an LDL comparable with that of the Kubota team (Kubota et al. 2014).
In conclusion, we established and validated a real-time PCR method for the detection of toxigenic C. difficile simultaneously targeting the tcdA, tcdB, and cdtB genes. The method displayed good performance for the specificity, sensitivity, and repeatability which has been submitted for a Chinese invention patent (CN202010825309.X). Despite the great advantages mentioned above there are some limitations: the reaction parameters and system should be further optimized to improve the overall sensitivity of the method and to optimize the amplification efficiency, in addition, the sample size should also be increased to obtain more accurate evaluation data.
It is worth mentioning that the method improved the identification of the A-B + straind, including typical Asian strain RT017. It could also be combined with the method in Chinese invention patent CN202010821402.3 for early warning testing of the highly toxigenic strain BI/ NAPI/027, and be applicable to CDI epidemiological surveillance and potential outbreak control. In addition, our method is applicable to a variety of fluorescence quantitative PCR instruments, and can also be combined with POCT for rapid detection and identification of clinical CDI cases. otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.