A novel real-time PCR assay for quantitative detection of Campylobacter fetus based on ribosomal sequences

Background Campylobacter fetus is a pathogen of major concern for animal and human health. The species shows a great intraspecific variation, with three subspecies: C. fetus subsp. fetus, C. fetus subsp. venerealis, and C. fetus subsp. testudinum. Campylobacter fetus fetus affects a broad range of hosts and induces abortion in sheep and cows. Campylobacter fetus venerealis is restricted to cattle and causes the endemic disease bovine genital campylobacteriosis, which triggers reproductive problems and is responsible for major economic losses. Campylobacter fetus testudinum has been proposed recently based on genetically divergent strains isolated from reptiles and humans. Both C. fetus fetus and C. fetus testudinum are opportunistic pathogens for immune-compromised humans. Biochemical tests remain as the gold standard for identifying C. fetus but the fastidious growing requirements and the lack of reliability and reproducibility of some biochemical tests motivated the development of molecular diagnostic tools. These methods have been successfully tested on bovine isolates but fail to detect some genetically divergent strains isolated from other hosts. The aim of the present study was to develop a highly specific molecular assay to identify and quantify C. fetus strains. Results We developed a highly sensitive real-time PCR assay that targets a unique region of the 16S rRNA gene. This assay successfully detected all C. fetus strains, including those that were negative for the cstA gene-based assay used as a standard for molecular C. fetus identification. The assay showed high specificity and absence of cross-reactivity with other bacterial species. The analytical testing of the assay was determined using a standard curve. The assay demonstrated a wide dynamic range between 102 and 107 genome copies per reaction, and a good reproducibility with small intra- and inter-assay variability. Conclusions The possibility to characterize samples in a rapid, sensitive and reproducible way makes this assay a good option to establish a new standard in molecular identification and quantification of C. fetus species. Electronic supplementary material The online version of this article (doi:10.1186/s12917-016-0913-3) contains supplementary material, which is available to authorized users.


Background
Members of the genus Campylobacter are gram-negative epsilon-proteobacteria highly adapted to vertebrate hosts. Most species are pathogens of a wide range of livestock species and have extensive reservoirs in wildlife [1][2][3].
The species Campylobacter fetus shows a remarkable level of intraspecific variation, with three subspecies: C. fetus subsp. fetus, C. fetus subsp. venerealis, and C. fetus subsp. testudinum. Campylobacter fetus fetus and C. fetus venerealis are classified on the basis of their mechanisms of transmission, clinical presentations and two key biochemical tests (tolerance to glycine and H 2 S production) [4,5]. Campylobacter fetus fetus infects the intestinal tract of several mammalian species and induces abortion in cattle and sheep [2,5,6]. In humans, it is an opportunistic pathogen that mainly infects immunecompromised patients [7,8]. Campylobacter fetus venerealis is a cattle-restricted pathogen with tropism for genital tissues and is the etiological agent of bovine genital campylobacteriosis (BGC), a serious reproductive disease that causes infertility and abortion [9]. C. fetus venerealis includes a variant, namely C. fetus venerealis biovar intermedius that reacts differently to the H 2 S test and also causes BGC [5]. Campylobacter fetus testudinum has been proposed recently to cluster some reptilian and human strains of putative reptilian origin on the basis of notorious genetic divergence from C. fetus fetus and C. fetus venerealis [10].
Biochemical tests remain as the gold standard for identifying C. fetus and differentiating between C. fetus fetus and C. fetus venerealis, but the fastidious growth requirements and the lack of reliability and reproducibility of some assays [11], due in part to the genetic heterogeneity of some strains, motivated the development of alternative diagnostic methods.
Several studies have endeavored in determining the suitability of different genetic methods for identifying the species C. fetus using end-point PCRs. In particular, the multiplex-PCR assay designed by Hum et al. [12] has been vastly used for species identification. Detection of C. fetus in this assay is achieved using PCR primers that target signature regions of the cstA gene, and C. fetus venerealis identification is based on the parA gene. However, genetic divergence in the cstA gene could prevent their detection by this assay, as occur in reptilian strains, and thus fails as a general diagnostic tool to identify the species [10].
Other assays for C. fetus identification were later designed to target additional genes, like cpn60, which encodes the universal 60-kDa chaperonin, and nahE, which encodes a sodium/hydrogen exchanger protein [13,14]. The cpn60 and nahE gene-based methods have been updated to real-time PCR assays using different technologies [14][15][16][17]. Both real-time assays have been designed to detect C. fetus on bovine isolates, and successfully tested on this kind of samples, but may fail to detect some genetically divergent strains, particularly of reptilian origin, which have distinctive nucleotide variants in many genes. Therefore, detection of C. fetus can be improved by developing new real-time PCR assays able to detect strains from all subspecies and hosts. These assays should be designed to target highly stable genomic regions that are characteristic for the species. Ribosomal genes are one of the most common DNA regions used to design PCR assays for the identification and detection of microorganisms. The 16S rRNA gene-targeted molecular tools are widely used as its variability has been thoroughly described in all Campylobacter species [18][19][20][21][22][23]. The sequence of the 16S rRNA gene is species-specific within the genus and C. fetus has several unique nucleotide markers [24,25]. Moreover, ribosomal genes are homogeneous for C. fetus subspecies and have three identical copies per genome allowing a better detection. Despite the obvious advantages of these genes, so far, there is not a real-time PCR assay targeting ribosomal sequences for the specific detection of C. fetus.
The aim of the present study was to develop a highly sensitive real-time PCR assay, to detect and quantify C. fetus strains.

Results
Strains were assigned to C. fetus and its subspecies using standard bacteriological methods (Table 1). Additionally, we performed the molecular characterization in the same collection of strains ( Table 1). The results of bacteriological and molecular classification do not always match, particularly at the subspecies level One bovine (INTA 89/222) and the reptilian isolate (RA8/Italy/2011) were phenotypically identified as C. fetus but were negative for the cstA gene amplicon that is currently used as a marker for C. fetus. The bovine isolate was positive for the subspecies (C. fetus venerealis) markers of both tests and the reptilian isolate was negative. The assignment of these isolates to the species C. fetus was confirmed by sequencing a fragment of the 16S rRNA gene, which unequivocally discriminates between Campylobacter species and from other bacterial species [21,24].
The 16SPb probe is species specific and has a minimum of one mismatch with a single sequence from C. hyointestinalis, and a maximum of nine differences with other Campylobacter species (e.g. C. rectus and C. showae). The forward primer's sequence is species specific and has a minimum of one and a maximum of four mismatches with other Campylobacter species (Figs. 1 and 2, Additional file 1). The reverse primer's sequence is identical in some Campylobacter species but has one or two differences with others. The combination of primers and probe only matches perfectly with the 16S rRNA gene of C. fetus.
All PCR reactions using template DNA from C. fetus fetus, C. fetus venerealis, C. fetus venerealis bv. intermedius, and C. fetus testudinum yielded a VIC signal corresponding to the C. fetus-specific probe. This result indicates a 100% clinical sensitivity and 95% confidence interval of 94-100% (Clopper-Pearson interval).
The analytical performance of the assay was determined using a standard curve (Fig. 3). The linear dynamic range of the assay was established between 10 2 and 10 7 genome copies per reaction. The amplification efficiency and the coefficient of determination (R 2 ) were 93% and 0.9973, respectively. Intra-and inter-assay reproducibility was calculated using the coefficient of variation (CV), which showed considerable low values, being the highest 2.19% ( Table 2).
No fluorescent signal was observed using template DNA from non-C. fetus bacterial species used as negative controls (i.e. C. geochelonis, C. hyointestinalis, C. jejuni, C. coli and C. sputorum). This result corresponds to a clinical

Discussion
Campylobacter fetus is a pathogen of great relevance for the cattle industry and public health. It is mandatory to report the presence C. fetus venerealis to the World Organization for Animal Health (OIE). In humans it is necessary to detect this opportunistic pathogen to achieve a better treatment and for epidemiological surveys. Detection of C. fetus in humans is difficult because both C. fetus fetus and C. fetus testudinum are potential pathogens and well-established methods would fail to detect strains of reptilian origin [10]. Therefore, cost-effective, automated and straightforward tools for the unambiguous identification of C. fetus are of paramount importance.
Bacteriological analysis, like culture isolation and biochemical tests, are well standardized and extensively used but challenging by the slow growing and few differential phenotypic properties of C. fetus [26]. These methods are also laborious and time-consuming, a disadvantage when processing samples at large-scale or delivering a fast diagnosis. To improve the quality and complement the gold-standard bacteriological methods for C. fetus detection, some end-point PCR methods have been designed based on the presence of speciesspecific amplicons [12,[27][28][29]; these assays fulfill various criteria such as accuracy, high detection probability and well-standardized protocols for its application and interpretation. Real-time PCR methods have been also designed with the same purpose [14][15][16][17] and have provided additional technical improvements to C. fetus As described in Hum et al. [12] c As described in Iraola et al. [41] d These strains were assayed both starting from a resuspended culture and directly from bovine samples of placenta or vaginal mucus, without a previous isolation step detection protocols, like the prevention of cross contamination and the minimization of manipulation and running times. However, both end-point and real-time PCR methods described to date are designed to identify C. fetus in bovine samples and do not deal with the intra-specific genetic variability of the bacteria that is found in diverse hosts. In comparison to conventional PCR methods, real-time PCR assays provide increased sensitivity and an accurate quantification of target DNA to study the dynamics of the bacteria in different hosts and tissues. To the best of our knowledge, there is not a real-time PCR method that uses ribosomal sequences for the identification and quantification of C. fetus.
Here, we have improved the current molecular methods for C. fetus detection by designing a new real-time PCR assay that targets the multi-copy 16S rRNA gene. The variability of these sequences within Campylobacter species supports its suitability as a target for amplificationbased methods using fluorescent probes. The inclusion in the assay of a TaqMan-MGB probe provides higher specificity, sensitivity and accuracy than traditional TaqMan probes and discriminates between sequences that differ in just one nucleotide [30][31][32].
Our assay was compared to the cstA gene end-point PCR proposed by Hum et al. [12] and currently used as standard for molecular diagnosis of C. fetus. The bovine  (Table 1), revealing that the sensitivity of this method for bovine isolates is not complete as previously reported [12,17,[33][34][35][36]. These isolates were confirmed as belonging to C. fetus by sequencing a fragment of the 16SrRNA gene; therefore the lack of amplification of the cstA gene could be due to the absence of the target cstA gene in these strains, or the presence of sequence variations that prevent the correct annealing of primers. Our attempt to amplify a larger region including Hum's PCR target region also failed, indicating the absence of this gene in these strains or an even greater sequence divergence within the cstA gene (data not shown). To test this hypothesis, it would be necessary to conduct the whole genome analysis of these strains. This notion is supported by the presence of several differences in Hum's primers binding sites in the complete genome of the reptilian strain C. fetus subsp. testudinum 03-427 (GenBank Acc. number NC_022759). This explains why the 13 isolates used for the description of this subspecies, and the RA8/Italy/2011 strain analyzed here, were negative for Hum's method based on the cstA gene [10]. Given the importance of this gene in the metabolism of nitrogen, and in the interaction with the host in C. jejuni [37], it is necessary to continue investigating its variations and possible roles in C. fetus.
Our novel real-time PCR assay detected all C. fetus tested in this study, but was negative for other Campylobacter species. The complete identity of primer and probe targets in all C. fetus strains deposited in the GenBank database (including reptilian isolates) supports that our  assay is expected to detect the currently described subspecies from diverse hosts (Fig. 1). These results indicate the excellent sensitivity and specificity of the assay. In addition, the primers and probe sequences are conserved in the 16SrRNA gene of the three subspecies (Fig. 1), in contrast with what happens with primers that amplify the cstA gene.
The assay here described has some advantages over other real-time PCR methods described in the literature. The nahE assay reported by Van der Graaf-van Bloois et al. [17] uses a TaqMan probe that provides high sensitivity and detection capability, but its quantification capability has not been ascertained using a standard curve. It is also uncertain whether this assay would detect reptilian C. fetus testudinum isolates, for which it was not designed, as the probe and the forward PCR primers have two mismatches each with respect to the C. fetus testudinum reference strain 03-427. The hybridization of primers and probes to the nahE gene could be also affected because it is embedded in a region that shows genomic rearrangements in most of the complete genome sequences available in the databases (not shown). The methodology to detect the cpn60 gene described by Chaban et al. [14] uses specific primers and SYBR green chemistry to identify C. fetus species, but its performance is sub-optimal in samples with low bacterial concentrations [15], such as the uncultured samples that were successfully tested in the present assay (Table 1).

Conclusions
The 16S rRNA gene-targeted assay here developed is excellent for the accurate detection and quantification of C. fetus in clinical samples and pure cultures. The possibility to characterize a large number of samples in a rapid, sensitive and reproducible way makes this assays a suitable tool for routine testing and research. For these reasons, this method has the potential to become a new standard in molecular identification of C. fetus species.

Real-time PCR design
The assay is based on a set of primers that amplifies a 78-bp sequence of the 16S rRNA gene (16SFw: 5′-GC ACCTGTCTCAACTTTC-3′and 16SRv: 5′-CCTTACCT GGGCTTGAT-3′) and a TaqMan-MGB probe (16SPb: 5′-VIC-ATCTCTAAGAGATTAGTTG-MGB/NFQ-3′), which targets a 19-bp polymorphic region that discriminates strains of C. fetus from the remaining Campylobacter species and other bacteria. This polymorphic region ( Fig. 1) was detected by visual inspection of over 3859 partial and complete 16S rRNA gene sequences aligned with T-Coffee [38]. The constructed alignment comprised sequences from all recognized Campylobacter species and from unassigned strains belonging to the genus, which were obtained from the SILVA database [39]. An alignment of 1907 representative sequences (removing identical sequences) is shown in Additional file 1. BLAST algorithm [40] was used to check in silico the specificity of primers and probe sequences, and to evaluate the occurrence of non-specific matches within the genomes of C. fetus and other bacterial species.

Bacterial strains: species and subspecies identification
The real-time PCR assay was tested with a collection of C. fetus strains isolated from cattle, humans and reptiles. Two of the strains (INTA 97/C1N3 and INTA 97/608) were assayed also directly from bovine samples of placenta or vaginal mucus, without a previous isolation step. Ten additional strains from four non-fetus Campylobacter species that occasionally occur in bovine samples were used to verify the specificity of the assays (Table 1).
Strains were previously typed using bacteriological methods to test the assay specificity. Samples were grown in Brucella semi-solid Broth and Campylobacter selective medium under microaerophillic conditions (85% H 2 , 5% O 2 , 10% CO 2 ) for 48 h at 37°C. The presumptive Campylobacter colonies were tested by catalase and oxidase tests, and grown in Brucella broth (Sigma-Aldrich, St. Louis, USA) with 1, 1.3, 1.5 and 1.9% glycine (Sigma-Aldrich), without glycine and in Brucella broth with NaCl and cysteine (Sigma-Aldrich) to detect H 2 S production with a lead acetate paper (Sigma-Aldrich). Sodium selenite reduction test was also performed. Colonies that grew in 1% glycine were classified as C. fetus fetus or C. fetus testudinum by their positive or negative H 2 S production, respectively. Glycine-sensitive colonies were assigned to the subspecies C. fetus venerealis (H 2 S negative) or C. fetus venerealis bv intermedius (H 2 S positive) ( Table 1).
Out of a total of 60 strains, 25 were C. fetus fetus, 20 C. fetus venerealis, 10 C. fetus venerealis bv intermedius, one was C. fetus testudinum, and four were not analyzed. Strains were further characterized using the multiplex-PCR assays designed by Hum et al. [12] and Iraola et al. [41]. Both assays use the same species-specific primers to detect the cstA gene and different genes to identify the subspecies. The first method includes a fragment of the parA gene as a C. fetus venerealis marker, and the second uses a fragment of the virB11 gene (Table 1) [42].
In cases where multiplex-PCR based methods failed to identify the isolates, molecular identification of species was confirmed by sequencing a fragment of the 16S rRNA gene, which was amplified using the C412F and C1288R primers described by Linton et al [21].

Real-time PCR assays
DNA was extracted from 500 μL of a suspension of live bacteria in a phosphate-buffered saline pH 7.4 solution (1 × 10 8 CFU/mL), or from 1 mL of preputial washing or vaginal mucus. The QIAamp DNA Mini Kit (Qiagen, Hilden, Germany) was used for all DNA extractions and the DNA purity was measured as the ratio of absorbance at 260 and 280 nm (A260/280) using a Nanodrop 2000 (Thermo Scientific, Waltham, USA).
Real-time PCR was carried out in a 25-μL reaction containing 1 × TaqMan Genotyping Master Mix (Applied Biosystems, Foster City, USA), 1 × Custom TaqMan SNP Genotyping Assay (0.9 μM each primer and 0.2 μM probe), and 1 μL genomic DNA. Thermocycling was performed on an ABIPrism 7500 (Applied Biosystems) and consisted of a 5 min incubation step at 50°C, denaturation for 10 min at 95°C, followed by 40 cycles of 15 s at 95°C and 1 min at 60°C, and a final step of 5 min at 70°C. Fluorescence measurements from VIC fluorophore was collected at the 5 min initial incubation stage, at the 60°C step of each cycle, and at the end of the run.

Standard curve generation for analytical testing
To construct the standard curve for the ribosomal probe we generated 10-fold serial dilutions containing 10 0 -10 7 genome copies/μL. Number of genome copies was determined by the following formula: Y (genome copies/μL) = [X (g/μL) DNA/ (nt genome length × 660)] × (6.022 × 10 23 ) using the DNA concentration of the dilution (X) and the genome size of the strain Cff 82-40 (1.77 Mb; GenBank accession number NC008599). The log dilution series of C. fetus genomes and negative controls containing nuclease-free water were tested with real-time PCR in triplicate and in three independent runs. Standard curve was generated by plotting threshold cycle (Ct) values per three replicates per standard dilution versus the logarithm of the bacterial genome copies to determine analytical sensitivity and efficiency of the assay. The amplification efficiency was calculated with the equation E = (10 (−1/k) ) − 1, where (k) is the slope of the linear regression line [43,44]. A value of 1 corresponds to 100% amplification efficiency. The coefficient of determination (R 2 ) was also assessed and was considered to be suitable when it was higher than 0.980 in a single run [45,46]. The coefficients of variation (CVs) of Ct values were assessed separately for each standard bacterial dilution by analyzing the replicates of the same analytical run (intraassay) and the repeated analyses from different analytical runs (inter-assay).

Additional files
Additional file 1: Alignment of 16S rRNA gene sequences from Campylobacter species obtained from the SILVA database, in FASTA format. (TXT 4882 kb)