Development and application of a triplex TaqMan RT-PCR assay for simultaneous detection of Feline calicivirus, Feline Herpesvirus 1 and Feline parvovirus

The feline calicivirus (FCV), feline herspesvirus 1 (FHV-1) and feline panleukemia virus (FPV) are heavily threaten the health of cats. In this study, a triplex TaqMan real-time polymerase chain reaction (RT-PCR) assay (triplex assay) was developed to detect these viruses. The optimized concentration of primers was 0.5 µM of each, probes concentration was 0.1 µM for FCV and FHV-1, 0.05 µM for FPV. The annealing temperature was set at 54 ℃ . The triplex RT-PCR assay was carefully validated. The detection limit for FPV, FCV, and FHV-1 was 5×10 1 copies/µL, which showed a 10-100-fold increase in the sensitivity compared with the conventional PCR. The coecients of variation (CV) of the intra-assay variability of the test were < 1.86%, and that of inter-assay was < 3.19%, indicating excellent repeatability and reproducibility of the triplex assay. Additionally, the assay has perfect specicity. In a pilot study, samples from 48 cats were analyzed using the triplex RT-PCR method and the commercial kits, and further conrmed by sequencing. The positive rates for the samples analyzed with these two methods were 70.83% and 62.5%, respectively, which demonstrated that the developed method was more accurate than the commercial kits in clinical diagnosis.


Introduction
Cats have been regarded as members of our family and their health has received great attention.
However, in clinical practice, infectious diseases such as feline pan-leukemia virus (FPV), feline calicivirus (FCV) and feline herpes virus 1 (FHV-1) are posing a huge threat to their health. More importantly, the mixed infection of these three viruses often occurs in the clinic and sometimes has similar clinical features [1]. As a member of Vesivirus of Calicivirdae, FCV is the most widespread feline virus, with overall prevalence ranges from approximately 15-31% [2]. The morbidity of FCV can reach to 90% in some colonies [3]. More importantly, the clinically recovered felines may become the virus carriers [4].
FHV-1 belongs to α-Herpesvirinae of Herpesviridae [1], which mainly infects the kittens aged 2-3 months [5]. When the kittens are infected, secondary infections are likely to occur due to the reduced immunity, and the nal mortality rate can reach to 70% [6]. Panleukopenia, caused by FPV, a Parvovirus of Parvoviridae, is another acute, highly contagious, and sometimes fatal feline viral disease, which is widely distributed throughout the world [7].
Although vaccines can prevent virus infection, in some less developed part of the world, the vaccines immunization is not universal, these viruses are still highly prevalent and cases of co-infection are often found, especially in multi-cat households and stray catteries [8]. In another aspect, in the early stages of these three virus infections, similar clinical symptoms often occur, such as mental illness, anorexia, sneezing, diarrhea, conjunctival congestion, eye nose secretion, dyspnea, etc. [9][10][11][12], which make it di cult to discriminate them from each other with the naked eye, and therefore miss the best treatment time. Additionally, in physical examination, this virus infection is the required items for our cats.
Therefore, it is urgent to establish a timesaving, labor-saving, sensitive, and e cient detection method that is suitable for simultaneous detection of these viruses. Current diagnostic methods for virus infection includes serological testing, virus isolation and identi cation, immunoelectron microscopy, polymerase chain reaction (PCR) and so on. Serological diagnosis needs the development of high speci c antibody, while virus isolation and culture and electron microscopy cannot be widely used in clinical diagnosis due to the high cost and time-consuming.
Nowadays, a second-generation PCR technology, RT-PCR, has been widely used in the eld of scienti c research and clinical detection due to its characteristics of high speci city, high sensitivity and shorted the detection time [13][14][15][16]. Although, several PCR methods have been reported for these virus detection [5,[17][18][19], most of the methods have low sensitivity or tedious operation, and no method can be used to detect three kinds of viruses quickly and e ciently at the same time. In this study, a triplex TaqMan RT-PCR method for simultaneously detection of FCV, FHV-1, and FPV was developed, which can differentiate these three viruses with high sensitivity, speci city, and reproducibility.

Materials And Methods
Pathogen and clinical samples An amount of 96 clinical samples from 48 cats (48 oral swabs and 48 rectal swabs) were collected in Nanjing from February 2019 to December 2019. All samples were divided into two parts, one for the triplex RT-PCR assay, and the other for commercial kits analysis. At the same time, 30 negative samples (15 oral swabs and 15 rectal swabs), con rmed to be free of FCV, FHV-1, and FPV [5,12,20], were used in the study. Methods for sample collection and storage were as described [21].

Primers and probes design
The published VP2 gene of FPV virus, ORF2 gene of FCV virus, and TK gene of FHV-1 virus were obtained from GenBank and aligned by DNAMAN (LynnonBiosoft, USA) to nd the conservative regions. Then six pairs of speci c primers and three speci c probes were designed, and the speci city was con rmed using BLAST in NCBI. Three pairs of longer fragments were used to construct plasmids, and three pairs of shorter fragments and probes were used for uorescence detection of three viruses. The three probes were labelled with FAM/BHQ1, VIC/BHQ1, and Texas Red /BHQ2 at its 5' and 3' terminals, respectively. The probes were purchased from Sangon Biotech (Shanghai, China) and the primers were purchased from Nanjing Kingsley Biotechnology Co. Ltd, the details of these fragments are shown in Table 1.  Nucleic acids extraction and standard plasmid preparation The nucleic acids of FCV, FHV-1 and FPV were extracted by DNA/RNA Extraction kit (Sangon Biotech, China). Reverse transcription was performed to synthesize viral cDNA following the manufacturer's instructions (Thermo Scienti c, USA). The concentration and purity of the nucleic acid were determined by measuring the absorbance at 260/280 nm with a NanoDrop2000c spectrophotometer (Thermo Scienti c, USA). All products were stored at -80 ℃ until use.
FPV, FHV-1, and FCV gene fragments were ampli ed by PCR. The PCR products were puri ed and recovered using a DNA Gel Extraction Kit (Axygen, China). Then, these recovered fragments were cloned to pMD18T vector (TaKaRa, China) to obtain the pMD18T-FCV, pMD18T-FHV, and pMD18T-FPV, respectively. The positive plasmids were used as the standard and establish the standard curves. The concentration of plasmids was calculated according to the absorbance measurement and the conversion of the copy number of the plasmid has been described in the previous study [22].

Experimental design and RT-PCR
To obtain a more sensitive, stable, and easy PCR method, the annealing temperature, primer concentration, and probes concentration for each target gene were carefully optimized. D-optimal design (MODDE 12.1 software) was carried out to comprehensively analyze the in uence of these factors [23]. The experimental conditions with the highest uorescence signal and the lowest Ct value were used as the optimal reaction conditions.
The triplex RT-PCR assay was carried out in a nal volume of 20.0 µL on a LightCycler 96 RT

Validation of the RT-PCR method
The speci city of the established RT-PCR method was con rmed using RV, PRV, FCoV, FIV, and FeLV. To ascertain the detection limit and the linear range of the method, the plasmids containing the target genes were diluted by 10 times gradient (from 5×10 7 copies/µL to 5×10 0 copies/µL) and subjected to the RT-PCR assay. The ampli cation e ciency (AE) and correlation coe cient (R 2 ) were used as parameters to evaluate the sensitivity of the triplex assay [24]. The repeatability of the method was tested using plasmid as the templates at the concentrations of 5×10 7 , 5×10 5 , and 5×10 3 copies/µL. Each independent experiment was carried out in triplicate for the intra-assay repeatability test, and triplicate runs over three days were performed by different operators for the inter-assay repeatability test.
A pilot study of the RT-PCR method To ensure the reliability of the experiment, we conducted a co-infection experiment, in which three viral plasmids were mixed in different combinations and proportions and then used for the triplex assay. Then the method was used to analyze 96 collected clinical samples. In addition, we also compared the results with the commercial kits. All positive samples were con rmed by sequencing.

Statistical analysis
Data generation and collection were carried out with LightCycler SW 1.1. Data management, analysis, and graphics generation were performed using Microsoft Excel 2007 (Microsoft, USA) and MODDE 12.1 software (Umetrics, Sweden). Results are presented as mean value (average) ± standard deviation (SD). The intra-and inter-assay variations were calculated from the mean Ct values and expressed as coe cients of variation (CV).

Results And Discussion
Optimization of the triplex assay Annealing temperature is an important factor affecting PCR speci city and ampli cation e ciency (14).
High annealing temperature will reduce the binding e ciency of primers to templates, and low annealing temperature will lead to non-speci c ampli cation. In addition, the concentration of primers and probes also affect the PCR ampli cation reaction, with low concentration leading incomplete reaction while high concentration inhibiting the reaction. Therefore, nding the optimal reaction conditions is important to establish a PCR method. In this study, the D-optimal design consisting of 16 runs (in triplicate measurements) was rst adopted to explore the in uence of probe concentrations, primer concentrations and annealing temperature on RT-PCR. Taking FCV as an example, the three-dimension response surface curves were shown in Fig. 1A. Red areas represented lower Ct value and blue areas represented higher Ct value. The abscissa and ordinate of the lowest point were the optimal conditions. The 4D plots further illustrated the interaction between the three factors (Fig. 1B). We found that when the primer concentration was lower than 0.35 µM, the Ct value remained high regardless of the change in probe concentration and annealing temperature. However, when the primer concentration was in the range of 0.4-0.6 µM, the Ct value was lower with a lower probe concentration and annealing temperature.
In addition, we optimized the uorescence signal of the method (Fig. 2). The results showed the uorescence signal of FCV, FHV-1 and FPV was higher when the primer concentration was in the range of 0.5-0.6 µM ( Fig. 2A, B, C). For FPV (Fig. 2D), the uorescence signal of high concentration probes (0.15 µM and 0.2 µM) is stronger than that of low concentration probes (0.05 µM). However, for FCV ( Fig. 2E) and FHV-1 (Fig. 2F), a low concentration probe (0.05 µM) can obtain a stronger uorescence signal. Annealing temperature had an obvious affection on uorescence signal, that is, as the temperature decreased, the uorescence signal increased (Fig. 2G, H, I). Finally, as a compromise, the optimized experimental conditions was set as follows: primer concentration at 0.5 µM for each virus, probe concentration at 0.15 µM for FPV, 0.05 µM for FCV or FHV-1, and annealing temperature at 54 °C.
The cut-off for positivity was determined based on the Ct values of the negative samples, which exceeded 36. Once the Ct value of the sample exceeded 36, it was treated as a negative result.
Speci city and sensitivity FPV, FCV, FHV-1, and other pathogens (including FIV, FeLV, FCoV, RV, PRV) were used for the speci city test. The results showed that only FCV, FHV-1, and FPV showed speci c ampli cation curves, and the Ct values were all less than 36 (Fig. 3), while the uorescence signals of other pathogens and negative controls were below the baseline detection levels, indicating the triplex assay has high speci city.
The sensitivity test of RT-PCR and conventional PCR was performed with 10-fold serial dilution of plasmids as the templates (from 5×10 7 copies/µL to 5×10 1 copies/µL). Combined with the number of cycles and the brightness of the gel electrophoresis, the sensitivity test of real-time PCR and conventional PCR was performed. As shown in Table 2, the Ct value of real-time PCR remained positive and conformed to the linear trend at 5×10 1 copies/µL for three viruses. But the sensitivity of conventional PCR was between 5×10 2 -5×10 3 copies/µL (Fig. 4), which demonstrated a 10-100-fold decrease in the sensitivity than that of the triplex assay. Besides, the sensitivity of the triplex mixed system was the same as that of the uniplex system, which was due to the excellent design of the primers and probes and the optimization of reaction conditions (6). b Each reaction was performed in triplicate and the results were shown as the mean ± SD.
The standard curves of the triplex assay for these viruses were generated (Fig. 5). As shown in the gure, the triplex assay was linear over a 5×10 7 -5×10 1 copies/µL range with an R 2 value above 0.9966 for all three viruses. Besides, the AE was 90.37% for FPV, 93.88% for FCV, and 104.19% for FHV-1. The above results indicated that the assay has good sensitivity while maintaining good ampli cation e ciency.
The limit of detection (LOD) was determined by the serial dilutions of the recombinant plasmids that corresponded to the lowest copy number that gave a probability of at least 95% of detecting a PCR positive test result [25]. When the plasmids were diluted below 5×10 1 copies/µL, apparently randomly distributed Ct values in the range of 36 to 38 were observed by the triplex assay [26,27]. We speculated that it might be caused by non-speci c ampli cation. Therefore, we chose 5×10 1 copies/µL as the LOD of the developed method.
Repeatability and reproducibility of the triplex assay To assess the stability of the triplex assay, three different concentrations of plasmid were ampli ed for the repeatability and reproducibility test of the established real-time PCR method. As shown in Table 3, the intra-and inter-assay CVs for Ct values between 0.68-1.86%, and 0.85-3.19%, respectively, indicating triplex assay is highly reliable and accurate. Co-infection models and clinical sample detection n clinical cases, viruses often infect felines with different combinations and concentrations, especially in stray cats and kittens with weakened immunity. To simulate this situation, we created the co-infection models for testing. As shown in Table 4, the method could detect three viruses at the combinations of different concentrations, regardless of triplex infection or only duplex infections. Furthermore, the Ct value of the co-infection experiment also satis ed the linear standard, indicating its applicability in virus quanti cation during the co-infection. Finally, 96 clinical samples from 48 cats were examined using the developed method (Table 5). The results demonstrated that 34 cats were infected with FPV, FCV, or FHV-1. The positive rate for FPV, FCV, and FHV-1 was 29.17% (14/48), 50% (24/48), and 33.33% (16/48), respectively. Simultaneously, these samples were also detected with commercial kits, which showed that only 30 cats were infected with these pathogen, with a relatively low positive rate of 41.67%(20/48)for FCV and 29.17% 14/48 for FHV-1. It is worth mentioning that all positive samples detected by commercial kits were also tested positive by triplex assay. However, 4 cats that were recognized as FCV positive with the triplex assay were negative detected by the commercial kits. More importantly, all positive samples were con rmed by sequencing. Thus, the established RT-PCR method showed high accuracy than commercial kits in clinical diagnosis.  Furthermore, we found that the positive rate of FPV, FCV, and FHV-1 in this study was higher than previous surveys, which might be due to clinical samples were obtained from diseased cats with clinical symptoms. In addition, the mixed infection was common, most of which were co-infected with FCV and FHV-1, with a ratio of 20.83% (10/48). It worth noting that, rstly, most of the co-infection cases were stray cats or kittens within 3 months, which was consistent with the previous survey. The reason lies that, rstly, cats aged 2-3 months are trapped in weaning and their immune system is not fully developed, and therefore they are susceptible to viruses. Secondly, most of the stray cats are not vaccinated, which are prone to carry the virus and spread it to domestic cats. Therefore, our research highlights the great need for virus surveillance in stray cats.