Multiplex detection of six swine viruses on an integrated centrifugal disk using loop-mediated isothermal amplification

Abstract

Advances in molecular testing and microfluidic technologies have opened new avenues for rapid detection of animal viruses. We used a centrifugal microfluidic disk (CMFD) to detect 6 important swine viruses, including foot-and-mouth disease virus, classical swine fever virus, porcine reproductive and respiratory swine virus–North American genotype, porcine circovirus 2, pseudorabies virus, and porcine parvovirus. Through integrating the loop-mediated isothermal amplification (LAMP) method and microfluidic chip technology, the CMFD could be successfully performed at 62℃ in 60 min. The detection limit of the CMFD was 3.2 × 10² copies per reaction, close to the sensitivity of tube-type LAMP turbidity methods (1 × 10² copies per reaction). In addition, the CMFD was highly specific in detecting the targeted viruses with no cross-reaction with other viruses, including porcine epidemic diarrhea virus, transmissible gastroenteritis virus, and porcine rotavirus. The coincidence rate of CMFD and conventional PCR was ~94%; the CMFD was more sensitive than conventional PCR for detecting mixed viral infections. The positive detection rate of 6 viruses in clinical samples by CMFD was 44.0% (102 of 232), whereas PCR was 40.1% (93 of 232). Thirty-six clinical samples were determined to be coinfected with 2 or more viruses. CMFD can be used for rapid, sensitive, and accurate detection of 6 swine viruses, offering a reliable assay for monitoring these pathogens, especially for detecting viruses in widespread mixed-infection clinical samples.

Introduction

Since the 1990s, swine diseases caused by porcine reproductive and respiratory syndrome virus–North American genotype (PRRSV-NA; species Betaarterivirus suid 2), classical swine fever virus (CSFV; species Pestivirus C), foot-and-mouth disease virus (FMDV), pseudorabies virus (PRV; Suid alphaherpesvirus 1), porcine parvovirus (PPV; species Ungulate protoparvovirus 1), and porcine circovirus 2 (PCV-2) have spread widely throughout the world, and resulted in severe economic losses in the global farming industry.^{8,10,14,17,20} FMDV, PRRSV-NA, CSFV, PRV, PCV-2, and PPV are 6 common swine viruses that can cause high morbidity and mortality among pigs.^14,15

A rapid, simple, and specific multiplex detection method is needed to detect multiple swine viruses in mixed infected clinical samples. Molecular techniques, such as PCR, real-time PCR, and loop-mediated isothermal amplification (LAMP) targeting specific genes, have been established for the detection of the above-mentioned 6 viruses.^{5–7,9,21,23} Among these methods, LAMP has proved to be a reliable, simple, and rapid detection method, able to detect targeted pathogens within 60 min.^1,2,18 However, a serious concern about employing highly sensitive methods such as LAMP is the possibility of false-positive results arising from contamination by extraneous nucleic acids (NAs).^6,12,16 This drawback has hindered the widespread use of such techniques in clinical settings.

To resolve this problem, we designed and developed a centrifugal microfluidic disk (CMFD) for multiplexed detection of 6 viruses. This technique, also called “lab-on-a-chip,” offers a number of advantages, such as smaller amounts of sample and reagents required, less energy and time required, ease of disposal, compact size, computerization, and trouble-free analysis.^15,19 Through integrating real-time fluorescence LAMP and microfluidic disk technology, our CMFD can be used for rapid, sensitive, and accurate detection of 6 swine viruses simultaneously, including FMDV, CSFV, PRRSV-NA, PCV-2, PRV, and PPV. We tested the CMFD experimentally for its ability to perform multiplexed detection of 6 swine viral pathogens in clinical samples.

Materials and methods

Sample collection and virus preparation

Swine clinical samples (n = 232) were collected for CMFD testing, including: 39 tissue samples collected from pigs with clinical signs (provided by Harbin Veterinary Research Institute, Chinese Academy of Agricultural Sciences); 193 clinical samples provided by 2 pig farms in Hebei Province and 1 pig farm in Beijing, constituting ~20% of the herds on the 3 farms. These samples included 40 blood samples, 35 serum samples, and 118 nose–throat swabs, and were collected by farms from Jan 2016 to Dec 2016 when the 3 farms implemented animal health surveillance. The period of investigation was selected randomly.

The 39 tissue samples provided by the Harbin Veterinary Research Institute were collected from pigs with clinical signs on a farm in Heilongjiang. Clinical signs included fever, the appearance of vesicles, hemorrhage, lameness, diarrhea, apathy, coughing, increased mortality, decreased feed intake, bruising, eyelid edema, abortions, wasting, weight loss, respiratory distress, dyspnea, mummifications, dark red papules and macules on skin, ataxic gait, tremor, and motor incoordination.

We used the following virus isolates in our study: FMDV O/HB/HK/99 strain, CSFV C strain, PRRSV-NA CH-1R strain, PCV2 ZJ/C strain, PRV Bartha-K61Z strain, and PPV PPVS-1 strain. In addition, isolates of the following viruses were used to validate the methods: transmissible gastroenteritis virus (TGEV; species Alphacoronavirus 1) attenuated HUA strain, porcine epidemic diarrhea virus (PEDV) attenuated CV777 strain, and porcine rotavirus (RV) NX strain. All of the virus isolates used were received from the Harbin Veterinary Research Institute and were stored at the Institute of Animal Quarantine, Chinese Academy of Inspection and Quarantine (CAIQ).

Extractions of nucleic acids

Viral RNA was extracted from CSFV, PRRSV-NA, FMDV, RV, TGEV, and PEDV (RNeasy mini kit; Qiagen, Germantown, MD). Viral DNA was extracted from PCV-2, PRV, and PPV (DNeasy blood & tissue kit; Qiagen).

For clinical samples, genomic DNA and total RNA were simultaneously extracted from the same samples. Serum and nose–throat swabs (PureLink viral RNA/DNA mini kit; Invitrogen, Carlsbad, CA) and blood and tissues (TRIzol reagent; Invitrogen) were extracted, according to the manufacturer’s instructions. Extracted NA mixtures obtained from clinical samples were stored at −80°C until use.

The quantity and quality of the purified genomic DNA and total RNA were examined (NanoDrop 2000 spectrophotometer; Thermo Fisher Scientific, Waltham, MA) and diluted with water to 100 ng/µL. For high-quality RNA, the optical density (OD)_260/280 ratios should be 1.9–2.1, and the corresponding OD_260/230 ratio values should be >2. For high-quality DNA, the OD_260/280 ratio values should be 1.6–1.8, and the corresponding OD_260/230 ratio values obtained should be >2.

Primers and reagents

PCR primers were designed with Primer Premier v.5.0 based on conserved sequences of the 3D gene of FMDV (GenBank accession KR401154.1), 5’–noncoding region of CSFV (AF531433.1), open reading frame 7 (ORF7) of PRRSV-NA (EU807840.1), ORF1 of PCV-2 (EF524518.1), VP1 gene of PPV (AY684864.1), and gG gene of PRV (JF797217.1). LAMP primers were designed with Primer Explorer v.4.0. The primers for PCR (Table 1) and LAMP (Table 2) were synthesized by Sangon Biotech (Shanghai, China).

Table 1.

Primers used for PCR detection of 6 swine viruses.

Primer	Sequence (5’–3’)	Fragment (bp)
FMDV-P1	GGATCCTAATACGACTCACTATAGGGAGGCGAGGCTATCCTCTCCTT	235
FMDV-P2	CCCAAGCTTCTGAGGGATTATGCGTCAC
CSFV-P1	GGATCCTAATACGACTCACTATAGGGAGGCCCAAGCTTCGGAGGGACTAGCCGTAGTG	269
CSFV-P2	CGGGGTACCGTGCCATGTACAGCAGAG
PRRSV-NA-P1	GGATCCTAATACGACTCACTATAGGGAGGCCCAAGCTTAGCCTCGTGTTGGGTGGCAGA	414
PRRSV-NA-P2	CGGGGTACCCAAACTAAACTCCACAGTGTAAC
PCV-2-P1	CTAGATCTCAAGGACAAC	1,768
PCV-2-P2	CACACTCCATCAGTAAGT
PRV-P1	CCTCTCGACGTATCGTCC	283
PRV-P2	TAGGCGTCGTCGCCAAAG
PPV-P1	TTAAGGGGGTGGTGGGCATA	1,051
PPV-P2	GTGCTGCTGGTGTGTATGGA

See text for explanation of virus abbreviations.

Table 2.

Primers used for LAMP detection of 6 swine viruses.

Primer	Sequence
CSFV-F3	AGGGACTAGCCGTAGTGG
CSFV-B3	GTGCCATGTACAGCAGAG
CSFV-FIP	TGCTCACGTCGAACTACTGACGAGCTCCCTGGGTGGTCT
CSFV-BIP	AGCCCACCTCGAGATGCTATGTGATTTCACCCTAGCGA
CSFV-FLP	GACTGTCCTGTACTCAGGACTT
CSFV-BLP	CCCAAGACACACCTTAACCCT
PRRSV-NA-F3	CAGGGAGTGGTAAACCTTGT
PRRSV-NA-B3	GTCTGGATCGACGACAGA
PRRSV-NA-FIP	CCTTGCCTCTGGACTGGTTTCAGTCAATCAGCTGTGCC
PRRSV-NA-BIP	AGCCCCATTTYCCTCTAGCGACCACAATTGCCGCTCACTA
PRRSV-NA-FLP	GCGATGATCTTACCCAGCA
PRRSV-NA-BLP	GATGACGTCAGGCATCACTT
FMDV-F3	CGAGGCTATCCTCTCCTT
FMDV-B3	CTGAGGGATTATGCGTCAC
FMDV-FIP	TCCTGCCACGGAGATCAA
FMDV-BIP	TGAGATTCCAAGCTACAGATCA
FMDV-FLP	AGAGACGCCGGTACTCGTTGGGACCATACAGGAGAAG
FMDV-BLP	GAGCCCTTCCAGGGYCTCTCACCCAACGCAGGTAAAG
PCV-2-F3	CTCGATCTCAAGGACAACGG
PCV-2-B3	GCCCCACAATGACGTGTAC
PCV-2-FIP	CTCTGCAACGGTCACCAGACTCGTGACCTGTCTACTGCTGTG
PCV-2-BIP	TTGTCAGAAATTTCCGCGGGCTTTGGTCTTCCAATCACGCTT
PCV-2-FLP	CCGCTCTCCAACAAGGTACT
PCV-2-BLP	GGCTGAACTTTTGAAAGTGAGC
PPV-F3	CCAGGATACAAATACCTTGGT
PPV-B3	TGTCCAATTTTACCTCCGTA
PPV-FIP	GTCGTGTTCTTTTGCTGCGGCCAGGAAACTCACTAGACCA
PPV-BIP	TCCATACATCTACTTCTCAGCAGCGTCTTTTGCGTGTTCAGTT
PPV-FLP	TCTGATGGATTAGTTGGTTCTCCT
PPV-BLP	TGATGAGAAATTCATA
PRV-F3	CGGACCCCGTCAACGT
PRV-B3	CGAGGGCGTACTCCATCAG
PRV-FIP	ACTCGCGGTGGACGATGGGGACCGTCGCCTGGTTCT
PRV-BIP	GCCATGCCCTCCGTCGAGATGTCGATGCGCGTGCG
PRV-FLP	TGCCGCCGGTGACGTT
PRV-BLP	ACACCGGCGGGTACTCGT
PC-F3-P	CTTTAAAACTTGCTCTGCT
PC-B3-P	AGATCGGCTTGTTTCCTT
PC-FIP-P	CGGACAACTTTGGGACTCTAGGATTTAGTCTTAAATCGCTTCCTTCATG
PC-BIP-P	GCAAGGTGGGGAAACTGAACGATTTCAAAAAGGTACAACTCTCA
PC-LF-P	TCTAATCGGCGTGTTTGC
PC-LB-P	ATGTTTGCTTGCCAGTGCATTTTAG

PC = positive control. See text for explanation of virus abbreviations.

The reverse-transcription, loop-mediated isothermal amplification (RT-LAMP) kit was purchased from Eiken China (Beijing, China). Calcein and manganese chloride were purchased from Sigma-Aldrich (Shanghai, China). The DL 2000 DNA marker was purchased from TaKaRa (Dalian, China). The gel and disk images were obtained and processed by gel imager (FluorChem M; ProteinSimple, San Jose, CA).

Preparation of virus DNA and RNA standards

For 3 DNA viruses, ORF1 of PCV-2, VP1 gene of PPV, and gG gene of PRV were amplified with primers (Table 1). The amplicons were purified (E.Z.N.A. gel extraction kit; Omega Bio-Tek, Norcross, GA), and cloned into the pGEM-T Easy vector (Promega, Madison, WI). The synthetic plasmids were purified (E.Z.N.A. plasmid mini kit I; Omega Bio-Tek). The copy numbers of plasmids were calculated as described previously.¹¹

The 3D gene of FMDV, 5’–noncoding region of CSFV, and ORF7 of PRRSV-NA were amplified with primers containing the T7 promoter sequence in the 5’-end of the reverse primer (Table 1) and were in vitro transcribed with T7 RNA polymerase (TaKaRa) according to the manufacturer’s instructions. The synthesized complementary RNA (cRNA) was treated with RQ1 RNase-free DNase (Promega) to remove DNA vectors. Then cRNA was purified by ethanol precipitation, and the cRNA pellet was dissolved in diethylpyrocarbonate (DEPC)-treated water. The copy numbers of cRNA standards were calculated as described previously.²²

The viral synthetic plasmids and viral cRNA transcripts were quantified, and 10-fold dilutions of 10⁷–10⁰ copies/µL were prepared.

To prepare pre-mixed template standards, 5 μL of each of 6 viral templates at a concentration of 10⁶ copies/µL were mixed together, including 3 viral synthetic plasmids and 3 viral cRNA transcripts. Next, 20 µL of RNase-free water was added to the mixed templates. After mixing, a 50-µL pre-mixed template standard was obtained, containing 3 synthetic viral plasmids and 3 viral cRNA transcripts at a concentration of 10⁵ copies/µL. Subsequently, the pre-mixed template standard was serially 10-fold diluted in RNase-free water to yield mixed template standards of 10⁴–10⁰ copies/µL. All of the template standards were stored in a –80°C freezer.

Tu-LAMP reaction

Tube-type LAMP turbidity (Tu-LAMP) reaction systems were optimized for detection of 6 swine viruses, and RT-LAMP was carried out in a final volume of 25 µL, containing 12.5 µL of 2 × RM (reaction buffer mix), 1.5 µL of enzyme mix (mixture of 8 U of Bst DNA polymerase and 10 U of AMV reverse transcriptase), 1 µL of MnCl₂ (12 mM), 1 µL of calcein (300 µM), 0.5 µL of each of the outer primers F3 and B3 (10 µM), 1.25 µL of each of the primers FIP and BIP (40 µM), 0.5 µL of each of the primers LF and LB (40 µM), 1 µL of DNA/RNA templates, and 3.5 µL of DEPC-treated H₂O. Tu-LAMP reactions were carried out at 62°C for 60 min and then terminated by heating the mixture at 80°C for 5 min using a real-time turbidimeter (LA320C; Teramecs, Tokyo, Japan). The sensitivity of the Tu-LAMP was established, and standard curves were derived, using 10-fold serial dilutions of the synthetic plasmids and RNA transcripts with concentrations of 10⁵–10⁰ copies/µL. One microliter of each dilution was utilized as template. Three independent reactions were repeated.

Standard curves of Tu-LAMP

The crossing points (the threshold values of time) were calculated directly as the coordinates of points at which the threshold line actually crossed the lines representing the turbidity plots obtained. The threshold was chosen at a turbidity of 0.1. Standard curves were generated by plotting the threshold values of time against log copy numbers, and linear regressions were calculated (Excel; Microsoft, Redmond, WA).

Fabrication and architecture of the disk

Polymethyl methacrylate (PMMA) was used to make disks. Every disk was divided into 2 equal units with 1 injection port, 1 overflow port, 1 main channel, and 12 reaction pools (Fig. 1). Each unit could be connected by the main channel and detect up to 12 genes simultaneously. The entire disk could analyze 2 samples and up to 24 genes simultaneously.

Figure 1.

The architecture of the disk. Specific LAMP primer pairs were preloaded in reaction pools. Reaction pool 1 = porcine circovirus 2 (PCV-2)-1; 2 = PCV-2-2; 3 = foot-and-mouth disease virus (FMDV)-1; 4 = FMDV-2; 5 = porcine parvovirus (PPV)-1; 6 = PPV-2; 7 = pseudorabies virus (PRV)-1; 8 = PRV-2; 9 = porcine reproductive and respiratory syndrome virus–North American genotype (PRRSV-NA)-1; 10 = PRRSV-NA-2; 11 = classical swine fever virus (CSFV)-1; 12 = CSFV-2; 13 = negative-PCV2; 14 = negative-FMDV; 15 = negative-PPV; 16 = negative-PRV; 17 = negative-PRRSV-NA; 18 = negative-CSFV; 19 = positive-1; 20 = positive-2; 21 = positive-3; 22 = positive-4; 23 = positive-5; 24 = positive-6.

Specific primer sets (Table 2) were preloaded into each reaction pool and dried at room temperature before sealing the disk. For the negative control (NC), no template was added in the reaction pools preloaded with 6 LAMP primer sets targeting 6 viruses (Table 2). For the positive control (PC), a human genome fragment of neuropilin 2 (NRP2) and the specific LAMP primers to detect NRP2 were preloaded in the reaction pools. The PC and NC were run in every CMFD test.

LAMP master mix (26 µL, including DNA/RNA templates; Table 3) was introduced into a unit of the disk by pipettes through an injection port. The disk was centrifuged (Mini G centrifuge; IKA, Staufen, Germany) for 5–10 s at 2,000 × g, and the reaction mixture was centrifuged into the reaction pools. Each reaction pool contained 1.4 µL of master mix.

Table 3.

Reaction system of LAMP in CMFD.

Reaction components	Final concentration
2× reaction buffer mix	13.00 µL
Enzyme mix	1.56 µL
F3	0.20 µM each virus
B3	0.20 µM each virus
FIP	2.00 µM each virus
BIP	2.00 µM each virus
LF	0.80 µM each virus
LB	0.80 µM each virus
Calcein	12.50 µM
MnCl2	461 µM
RNA/DNA templates	6.00 µL
Add DEPC-treated H2O to volume	26.00 µL

B3 = backward outer prime; BIP = backward inner prime; F3 = forward outer prime; FIP = forward inner prime; LB = backward loop prime; LF = forward loop primer.

Detection of 6 swine viruses through CMFD

After the reaction mixture was centrifuged into the reaction pools, the disk was put into the real-time fluorescence LAMP machine (RTisochip-B; CapitalBio Technology, Beijing, China) to start the LAMP assay. The CMFD assay was run at 62°C for 60 min. The LAMP amplification in each reaction pool was independent, and there was no interference from adjacent pools. Fluorescence intensities were monitored by real-time fluorescence scanning, and results were analyzed based on amplification curves.

Specificity of CMFD

To determine the specificity of CMFD, the RNA genomes of 3 common swine viruses TGEV, PEDV, and RV were utilized as templates. Six microliters of each viral genome (100 ng/µL) were added to the 6 sets of LAMP primers for the CMFD assay. The final amount of viral genomes template per each reaction pool was:

$$6\mu \mathrm{L}\times 100\mathrm{ng}/\mu \mathrm{L}\times 1.4\mu \mathrm{L}/26\mu \mathrm{L}=3.2\times {10}^1\mathrm{ng}.$$

Three independent reactions were repeated for each virus.

Sensitivity of CMFD

The sensitivity of CMFD was tested using 10-fold serial dilutions of synthetic viral plasmids or viral cRNA transcripts of 6 targeted viruses with concentrations of 10⁵–10⁰ copies /µL. Six microliters of each dilution was then utilized as template by CMFD. Three independent reactions were repeated. The final amount of template per reaction pool was (at 10³ copies/µL for example):

$$6\mu {\mathrm{L}\times 10}^3\mathrm{copies}/\mu \mathrm{L}\times 1.4\mu \mathrm{L}/26\mu \mathrm{L}=3{.2\times 10}^2\mathrm{copies}.$$

6µL × 103copies/µL × 1.4µL/26µL = 3.2 × 102copies

The sensitivity of CMFD was further tested using 10-fold serial dilutions of pre-mixed nucleotide template standards of 6 targeted viruses at concentrations of 10⁵–10⁰ copies/µL. Six microliters of each dilution was then utilized as template by CMFD. The final amount of template of each virus per reaction pool was 3.2 × 10⁴ to 3.2 × 10⁰ copies. Three independent reactions were repeated.

Detection of 6 swine viruses in clinical samples by using CMFD

To further evaluate the performance of CMFD, the NA mixtures of 232 clinical samples were extracted, and these samples were screened for 6 targeted viruses by CMFD. The results were compared with detection results of conventional PCR.

For the RNA viruses (FMDV, CSFV, and PRRSV-NA), first-strand cDNAs were synthesized (GoScript reverse transcriptase system; Promega) according to the manufacturer’s recommendation. Final volume of reverse-transcription reaction mix was 20 µL; 5 µL experimental RNA (up to 1 µg) was utilized in each reaction.

Conventional PCR amplifications were carried out (KOD FX Neo polymerase KFX-201; ToYoBo, Tokyo, Japan). For each PCR assay, a 50-µL reaction volume was employed. The standard reaction solution contained 25 µL of 2× PCR buffer for KOD FX Neo, 10 µL of 2 mM dNTPs, 3 µL of PCR primer mix (0.3 µM of each primer in the final reaction solution), 2 µL of DNA/RNA templates, 1 µL of KOD FX Neo polymerase (1 unit), then distilled water to a final reaction of 50 µL. PCR assays were conducted (9700 thermal cycler; Applied Biosystems, Foster City, CA; Table 4).

Table 4.

Reaction conditions of PCR for detection of 6 swine viruses.

Procedure	Virus
	FMDV	CSFV	PRRSV-NA	PCV-2	PRV	PPV
Initial denaturation	94°C, 5 min
Denaturation	94°C, 30 s
Annealing	62°C, 30 s	62°C, 30 s	62°C, 30 s	54°C, 30 s	55°C, 30 s	58°C, 30 s
Extension	72°C, 40 s	72°C, 40 s	72°C, 50 s	72°C, 2 min	72°C, 40 s	72°C, 80 s
Cycle	40 cycles
Final extension	72°C, 5 min

See text for explanation of virus abbreviations.

The amplicons of LAMP, PCR, and RT-PCR were analyzed through gel electrophoresis, and the positive samples were sequenced (Sangon Biotech). The 36 sequencing results from confirmed samples have been deposited in GenBank as accessions MH487827–MH487862.

Results

Analytic performance

The detection limits of CMFD for FMDV, PRRSV-NA, PCV-2, and PRV were 3.2 × 10² copies per reaction (Fig. 2A, 2E, 2G, 2K), which were very close to the detection limits obtained by Tu-LAMP for these viruses (Fig. 2B, 2F, 2H, 2L). Detection limits of CMFD for CSFV and PPV were also 3.2 × 10² copies per reaction (Fig. 2C, 2I), which were 30 times less sensitive than that of Tu-LAMP (10¹ copies per reaction; Fig. 2D, 2J). The standard curves generated by Tu-LAMP are provided in Supplementary Fig 1. The R² values were calculated as 0.995 (FMDV), 0.985 (CSFV), 0.998 (PRRSV), 0.993 (PCV-2), 0.980 (PRV), and 0.968 (PPV). CMFD could not detect the 6 targeted viruses at the concentration of 3.2 × 10¹ copies per reaction (Supplementary Fig. 2).

Figure 2.

Sensitivity tests of CMFD for detection of 6 swine viruses compared with tube-type, LAMP turbidity methods (Tu-LAMP) using 10-fold serially diluted nucleic acid templates. A. Amplification plots of CMFD were based on FMDV cRNA at the concentration of 3.2 × 10² copies per reaction. B. Amplification plots of Tu-LAMP using 10-fold serially diluted FMDV cRNAs of 10⁵–10⁰ copies per reaction. C. Amplification plots of CMFD were based on CSFV cRNA at the concentration of 3.2 × 10² copies per reaction. D. Amplification plots of Tu-LAMP using 10-fold serially diluted CSFV cRNAs of 10⁵–10⁰ copies per reaction. E. Amplification plots of CMFD were based on PRRSV-NA cRNA at the concentration of 3.2 × 10² copies per reaction. F. Amplification plots of Tu-LAMP using 10-fold serially diluted PRRSV-NA cRNAs of 10⁵–10⁰ copies per reaction. G. Amplification plots of CMFD were based on PCV-2 synthesized plasmid at the concentration of 3.2 × 10² copies per reaction. H. Amplification plots of Tu-LAMP using 10-fold serially diluted PCV-2 synthesized plasmids of 10⁵–10⁰ copies per reaction. I. Amplification plots of CMFD were based on PPV synthesized plasmid at the concentration of 3.2 × 10² copies per reaction. J. Amplification plots of Tu-LAMP using 10-fold serially diluted PPV synthesized plasmids of 10⁵–10⁰ copies per reaction. K. Amplification plots of CMFD were based on PRV synthesized plasmid at the concentration of 3.2 × 10² copies per reaction. L. Amplification plots of Tu-LAMP using 10-fold serially diluted PRV synthesized plasmids ranging from 10⁵–10⁰ copies per reaction. NC = negative control; PC = positive control.

CMFD also achieved a sensitivity of 3.2 × 10² copies per reaction when pre-mixed templates (containing 6 viral templates) were utilized (Fig. 3), which was equivalent to the detection limits obtained by CMFD using single virus templates.

Figure 3.

Sensitivity test of CMFD for the simultaneous detection of 6 swine viruses using pre-mixed templates containing 3 recombinant plasmids and 3 RNA transcripts. CMFD could achieve a sensitivity of 3.2 × 10² copies per reaction for each virus when all pre-mixed recombinant plasmids and RNA transcript templates for 6 targeted viruses were present. NC = negative control; PC = positive control.

The specificity of CMFD was evaluated by detecting viral genomes extracted from 6 targeted viruses (FMDV, PRRSV-NA, CSFV, PPV, PRV, and PCV-2) and 3 other closely related viruses (TGEV, PEDV, and RV). Each set of LAMP primers specifically detected the targeted virus, and no cross-reaction was observed with other viruses (Fig. 4).

Figure 4.

Specificity test of CMFD using 9 viral genomes of 6 viruses (CSFV, FMDV, PRRSV-NA, PCV-2, PPV, and PRV) and 3 other important swine viruses (PEDV, TGEV, RV). Amplification plots of CMFD (at 32 ng per reaction) were based on A. PEDV genomes; B. TGEV genomes; C. RV genomes; D. CSFV genomes; E. FMDV genomes; F. PCV-2 genomes; G. PRRSV-NA genomes; H. PRV genomes; and I. PPV genomes. NC = negative control; PC = positive control.

Diagnostic performance

The diagnostic performance of CMFD was further evaluated through testing 232 clinical samples and comparing results with those of conventional PCR (Table 5). The positive detection rates of CMFD for FMDV, PRRSV-NA, CSFV, PPV, PCV-2, and PRV were 2.6% (6 of 232), 9.5% (22 of 232), 12.1% (28 of 232), 12.1% (28 of 232), 23.3% (54 of 232), and 11.6% (27 of 232), respectively. The positive detection rates of PCR for FMDV, PRRSV-NA, CSFV, PPV, PCV-2 and PRV by testing the same clinical samples were 2.6% (6 of 232), 8.6% (20 of 232), 11.2% (26 of 232), 11.6% (27 of 232), 21.1% (49 of 232), and 10.3% (24 of 232), respectively. All of the above positive detection results were further validated through gene sequencing and BLASTs.

Table 5.

Performance of CMFD for detection of 6 swine viruses in 232 clinical samples.

Clinical sample/Virus	No. of positive clinical samples
	CMFD	PCR
Positive for 1 virus
FMDV	6	6
PRRSV-NA	8	6
CSFV	10	8
PPV	13	14
PCV-2	22	19
PRV	7	8
Positive for 2 viruses
PCV-2 + PPV	4	4
PCV-2 + PRV	6	4
PCV-2 + CSFV	4	4
PPV + PRV	2	1
Positive for 3 viruses
PCV-2 + PPV + CSFV	3	3
PCV-2 + PPV + PRV	1	0
PCV-2 + PRV + CSFV	2	2
PCV-2 + PRV + PRRSV-NA	5	5
PCV-2 + CSFV + PRRSV-NA	2	2
PPV + CSFV + PRRSV-NA	1	1
Positive for 4 viruses
PCV-2 + PPV + CSFV + PRRSV-NA	2	2
PCV-2 + PRV + CSFV + PRRSV-NA	2	2
Positive for 5 viruses
PCV-2 + PPV + PRV + CSFV + PRRSV-NA	2	2
Total	102	93

See text for explanation of virus abbreviations.

The coincidence rate of the 2 detection methods was 94.0% (218 of 232); the detection results for 14 clinical samples were different between the 2 methods. All of the 14 clinical samples were determined to be positive for PRRSV-NA, CSFV, PCV-2, PPV, or PRV by CMFD, whereas these samples were determined to be negative for specific virus by PCR. Furthermore, the 14 clinical samples were validated to be positive through gene sequencing. The sensitivity of CMFD for 6 swine viruses was higher than the sensitivity of the conventional PCR.

Discussion

We confirmed several advantages to the CMFD. First, CMFD could simultaneously detect 6 swine viruses within 1 h, whereas conventional PCR and Tu-LAMP only detect a single virus at a time. Furthermore, detection of 6 targeted viruses in one sample through CMFD only needs the addition of sample once, whereas other methods, such as Tu-LAMP and PCR, need multiple additions. Multiple additions of samples could lead to cross-contamination. Second, the reaction volume of CMFD per reaction pool is 1.4 µL, whereas reaction volumes of Tu-LAMP and PCR are 25 µL. The consumption of reagents for CMFD is much less than traditional methods. Finally, CMFD is a sensitive and specific assay for targeted viruses. Sensitivity of the CMFD was examined using 10-fold serially diluted viral DNA/RNA templates. The detection limits of the CMFD for the viral pathogens were 3.2 × 10² copies per reaction. The detection limits of CMFD for 6 targeted viruses were lower than Tu-LAMP. The possible reason is that the reaction volume of CMFD is only 1.4 µL, whereas that of Tu-LAMP is 25 µL. The reduced reaction volume will decrease the absolute quantity of virus DNA/RNA templates and the sensitivity of CMFD. But the CMFD method proved highly specific for targeted viruses, and no cross-reaction was observed with the nontargeted viruses. The ability of CMFD to detect viruses was evaluated by testing 232 clinical samples and compared with that of PCR. In our study, 102 clinical samples were determined to be positive for targeted viruses, and the positive detection rate of CMFD is 44.0% (102 of 232). Meanwhile, the positive detection rate of PCR was 40.1% (93 of 232). The coincidence rate of the 2 detection methods was 94.0% (218 of 232). We could deduce that there exist subclinical infections of targeted viruses in the sampled herds. The detection sensitivity of CMFD was better than conventional PCR, given that CMFD could detect viruses in 14 clinical samples, whereas PCR could not.

To guarantee the success of CMFD tests, several issues should be noted. First, the quality of extracted NAs is significant for the CMFD tests and the presence of inhibitors such as protein or phenol would lead to the failure of CMFD. To ensure the purity of extracted NAs, we utilized high-quality NA extraction reagents, such as PureLink Viral RNA/DNA mini kits (Invitrogen) and TRIzol reagent (Invitrogen) according to the manufacturer’s instructions. Then the quality of DNAs and RNAs was examined with a NanoDrop 2000 spectrophotometer. The OD_260/280 and OD_260/230 ratio values indicated that the quality of extracted NA templates met the requirements and should not have great implications for the PCR/LAMP amplifications. Our aim was to establish a high-sensitivity method utilized for detection of swine viruses in the field, and our results demonstrated that CMFD can meet the above need. When the same NA templates were utilized for PCR and CMFD simultaneously, CMFD was more sensitive than conventional PCR. Second, the high amplification efficiency of LAMP may lead to false-positive results in the CMFD assay. One target gene could be amplified to 10⁹ copies within 20 min by LAMP, which is very sensitive to aerosol contamination.¹³ To prevent aerosol contamination, multiple functional zones of laboratories, and specific operation and cleaning of the test platform, must be guaranteed. Third, the quality of the disks utilized in CMFD is important; unstable CMFD products would lead to false-negative results. Intra-assay and inter-assay variability was assessed when the CMFD was established. The coefficients of variation (CVs) of the intra-assay and inter-assay variations were <10% in all cases. This indicated that the repeatability and stability of CMFD are acceptable.^3,4 CMFD would be utilized in rapid, primary screening of possible infection by any of 6 swine viruses in swine with clinical signs. CMFD established in our study could be used as a method for field screening in outbreak situations. Further validation is necessary and pathogenic virus isolation, enzyme-linked immunosorbent assay, or indirect immunofluorescence assay could be employed to confirm CMFD-positive suspicious samples.

The lack of an internal control for inhibition monitoring is a limitation of the CMFD method. No matter what methods were utilized for NA extraction, the inhibitors could not be completely avoided; the intent of CMFD is for deployment in the field, where the NAs might not be pristine. Moreover, NanoDrop could not be utilized to monitor the quality of NAs in the field. Therefore, the possibility of failure of CMFD induced by inhibitors within NA extractions exists, and an internal control is necessary for CMFD if it is utilized in the field. To address this issue, further research is needed to improve the CMFD method. The CMFD disk would be redesigned, and the disk should contain more than 12 reaction pools per unit to include the duplicate reaction pools for internal control together with reactions pools for 6 targeted viruses.