Combination of an Unbiased Amplification Method and a Resequencing Microarray for Detecting and Genotyping Equine Arteritis Virus

Abstract

This study shows that an unbiased amplification method applied to equine arteritis virus RNA significantly improves the sensitivity of the real-time reverse transcription-quantitative PCR (RT-qPCR) recommended by the World Organization for Animal Health. Twelve viral RNAs amplified using this method were hybridized on a high-density resequencing microarray for effective viral characterization.

TEXT

Equine arteritis virus (EAV), the causative agent of equine viral arteritis (EVA), is a member of the Arteriviridae family. Its genome is made up of a single positive-strand RNA molecule (1, 2). EAV infects Equidae and may be transmitted through the respiratory and venereal routes. EAV can persist in the reproductive tract of stallions only (3). Following the primary infection, up to 70% of stallions can be persistently infected and shed the virus in their semen, sometimes for life. These “shedder” stallions spread the virus in the horse population during breeding or when their semen is used for artificial insemination (4). The financial consequences are potentially huge, since shedder stallions lose their value. Moreover, infection may cause abortion in pregnant mares and the death of young foals (5). The World Organization for Animal Health (OIE) prescribes viral isolation (VI) on cell culture to detect EAV for international trade (6). However, a recent study showed that under certain conditions, the real-time reverse transcription-quantitative PCR (RT-qPCR) assay developed by Balasuriya et al. in 2002 (7) is as sensitive as VI for detecting EAV in semen (8). Therefore, RT-qPCR targeting open reading frame 7 (ORF7) of the EAV genome, encoding the viral nucleoprotein, is increasingly used for diagnosis.

The main challenge to EAV surveillance is detecting EAV in the semen of shedder stallions and the organs of aborted fetuses to prevent costly outbreaks. In particular, when the viral load is very low and thus the amount of viral nucleotides targeted is limited (9). The aim of our study was to increase the sensitivity of the OIE-recommended RT-qPCR method (6) by combining it with an unbiased amplification method using the Phi29 polymerase coupled to a high-density resequencing microarray (RMA) to genotype the viruses detected.

Sixty different samples were used in this study. Of the 48 EAV-positive samples, 31 were from semen (Table 1), 12 were from virus isolation cell culture supernatants, and 5 were tissue samples from the lungs, spleen, or liver of 1 aborted foetus, 3 young foals, and 1 adult. Viral RNA was extracted from semen, cell culture supernatant, or 1 g of tissue sample (previously homogenized in 5 ml of cell medium) with the QIAmp viral RNA minikit (Qiagen, Valencia, CA) following the manufacturer's instructions.

TABLE 1.

The 48 positive EAV samples tested with their origin, year of collection, and copy numbers^a

Sample no.	Sample type	Country of origin	Yr of collection	Assay 1, purified RNA		Assay 2, cDNA		Assay 3, WTA 1/100		Ratio of copy no. (assay 3/assay 1) × 100
				Copy no.	SD	Copy no.	SD	Copy no.	SD
1b	VIc	France	2008	209 × 103	19 × 103	282 × 104	84 × 104	504 × 106	120 × 106	24 × 104
2	Semen	France	2008	642	72	3 × 104	1.3 × 104	3 × 106	0.4 × 106	46 × 104
3b	Semen	France	2008	16 × 103	4 × 103	21 × 104	7.2 × 104	7.9 × 106	0.3 × 106	4.75 × 104
4b	Semen	France	2008	745	135	16.3 × 104	2.1 × 104	8.8 × 106	0.8 × 106	118 × 104
5b	VI	France	2008	4.7 × 103	1.8 × 103	88.2 × 104	26.8 × 104	537 × 106	44 .5 × 106	1140 × 104
6b	VI	France	2008	47.5 × 103	6.5 × 103	43 × 104	10 × 104	386 × 106	31.6 × 106	81.5 × 104
7	VI	Netherlands	2012	377	205	2,271	334	32 × 106	7.7 × 106	848 × 104
8	VI	Netherlands	2012	10.8 × 103	2.7 × 103	43.3 × 104	6.2 × 104	231.5 × 106	24.5 × 106	215 × 104
9	VI	Netherlands	2012	62 × 103	18 × 103	210.8 × 104	45.7 × 104	116.5 × 106	12.8 × 106	18.7 × 104
10	VI	Netherlands	2012	33.7 × 103	2.5 × 103	45.9 × 104	14.3 × 104	248 × 106	29 × 106	73.5 × 104
11	VI	Netherlands	2012	5 × 103	1 × 103	5.9 × 104	0.6 × 104	3.5 × 106	0.1 × 106	6.74 × 104
12	Semen	Germany	2010	6.4 × 103	1 × 103	2.7 × 104	0.35 × 104	0.85 × 106	19,218	1.33 × 104
13	Semen	Germany	2011	3 × 103	2.5 × 103	3,929	617	1.8 × 106	0.25 × 106	5.91 × 104
14	Semen	Germany	2010	15.3 × 103	4.3 × 103	11.7 × 104	1.7 × 104	8.5 × 106	0.4 × 106	5.52 × 104
15	Semen	Qatar	2011	1 × 103	121	1.26 × 104	0.35 × 104	1.8 × 106	0.28 × 106	17.1 × 104
16	Semen	Germany	2010	8 × 103	7.7 × 103	4 × 104	1.3 × 104	2 .6 × 106	0.45 × 106	3 × 104
17	Semen	Netherlands	2010	820	390	3.7 × 104	1.2 × 104	0.73 × 106	71,203	8.99 × 104
18	Semen	Germany	2012	33 × 103	13 × 103	23.3 × 104	0.5 × 104	10.6 × 106	2 × 106	3.21 × 104
19b	Lung	France	2007	539 × 103	159 × 103	141.3 × 104	14.6 × 104	87 × 106	4.2 × 106	1.62 × 104
20	Liver	France	2007	140 × 103	77 × 103	5.75 × 104	10.5 × 104	61.5 × 106	12.9 × 106	4.38 × 104
21	Lung	France	2007	113 × 103	15.7 × 103	44.9 × 104	10.3 × 104	31 × 106	2.8 × 106	2.74 × 104
22	Lung	France	2007	84	26	1,445	221	7.4 × 106	1.2 × 106	873 × 104
23	Spleen	France	2007	942	150	2,215	612	94,033	10,357	1 × 104
24	VI	France	2007	174	48	517	133	51,357	11,365	2.96 × 104
25b	Semen	France	2007	133 × 103	7.5 × 103	49.9 × 104	3.3 × 104	46.9 × 106	4 × 106	3.52 × 104
26b	Semen	France	2007	80.5 × 103	11.2 × 103	25.9 × 104	2 × 104	73.3 × 106	3.7 × 106	9.07 × 104
27	Semen	France	2007	46.6 × 103	33 × 103	12.3 × 104	2.5 × 104	13.25 × 106	1 × 106	2.84 × 104
28	Semen	France	2007	9.6 × 103	2.3 × 103	1.65 × 104	1,694	1.4 × 106	219,702	1.47 × 104
29	Semen	France	2007	6.5 × 103	701	9,615	1,997	4.1 × 106	615,044	6.46 × 104
30	Semen	France	2008	11 × 103	3 × 103	3.3 × 104	0.8 × 104	12.5 × 106	1.8 × 106	11.3 × 104
31	Semen	France	2008	48.5 × 103	17.7 × 103	14.5 × 104	6 × 104	26.8 × 106	2.9 × 106	5.53 × 104
32b	Semen	France	2008	32.5 × 103	6 × 103	4 × 104	5.9 × 104	21.1 × 106	10.6 × 106	6.50 × 104
33	Semen	France	2007	46.5 × 103	2.3 × 103	10.6 × 104	0.9 × 104	5.8 × 106	0.82 × 106	1.25 × 104
34	Semen	France	2007	182	23	461	31	6,770	2,514	3,720
35	Semen	France	2008	11 × 103	3 × 103	5.5 × 104	2.6 × 104	0.950 × 106	0.5 × 106	8,670
36	Semen	France	2008	15.2 × 103	1.5 × 103	7.9 × 104	0.45 × 104	1.9 × 106	32,714	1.25 × 104
37	Semen	France	2008	21.2 × 103	778	6 × 104	0.4 × 104	9.3 × 106	0.8 × 106	4.36 × 104
38b	VI	France	2008	200 × 103	11.5 × 103	119 × 104	22.5 × 104	196 × 106	20.5 × 106	9.78 × 104
39	VI	France	2008	155 × 103	18.2 × 103	9.8 × 104	5.4 × 104	107.8 × 106	16.5 × 106	6.96 × 104
40	VI	Netherlands	2012	184 × 103	31 × 103	84 × 104	8.7 × 104	156.5 × 106	5 × 106	8.49 × 104
41	Semen	Germany	2010	15.8 × 103	1.2 × 103	4 × 104	1 × 104	4.4 × 106	0.4 × 106	2.78 × 104
42	Semen	Germany	2012	61.3 × 103	36.3 × 103	3.75 × 104	1.3 × 104	1.499 × 106	2.12 × 106	245 × 104
43b	VI	France	2008	184 × 103	7.6 × 103	51.6 × 104	13.7 × 104	916 × 106	161 × 106	49.8 × 104
44	Semen	France	2008	2.3 × 103	1.8 × 103	3,248	1,056	7,458	521	326
45	Semen	France	2008	3.7 × 103	1 × 103	5,418	673	0.65 × 106	18,091	1.77 × 104
46	Semen	France	2008	320	95	586	146	375	90	117
47	Semen	France	2008	167	49	1,502	865	5,566	180	3,320
48b	VI	USA (Bucyrus strain)	1953	57.8 × 103	11.8 × 103	12.5 × 104	3 × 104	9.9 × 106	0.28 × 106	1.72 × 104
Mean copy no.				53.75 × 103d,e	11.4 × 103	31.104d	6.6 × 104	112.5 × 106e	55 × 106

^aCopy numbers obtained with purified RNA (assay 1), purified RNA retrotranscribed into cDNA (assay 2), and purified RNA retrotranscribed and then amplified with the Phi29 polymerase (assay 3) are shown. The gains in sensitivity between assays 1 and 3 are also shown (ratio of copy numbers).

^bSamples tested on the high-density resequencing DNA microarray.

^cVI, culture fluid from virus isolation performed on RK-13 cells (ATCC CCL-37), as described in reference 6.

^dAmong these two values, P < 0.05 (using Student's t test).

^eAmong these two values, P < 0.05 (using Student's t test).

In order to evaluate the gain in sensitivity of the newly developed method, extracted and purified viral RNA was amplified in triplicate using three different protocols (Fig. 1). First, RNA samples were amplified using the one-step RT-qPCR (assay 1) method recommended by the OIE (7). In parallel, the extracted RNA was treated with Turbo DNase (Life Technologies, Inc., Carlsbad, CA, USA) and then retrotranscribed into cDNA using the SuperScript III first-strand synthesis kit (Life Technologies, Inc.) in the presence of random hexamers. These cDNA samples were then either amplified using qPCR (assay 2) with the same program and the same primers and probe sequences as those used in assay 1 or subjected to random unbiased isothermal Phi29 amplification using the QuantiTect whole transcriptome kit (Qiagen, Valencia, CA), as described by Berthet et al. (10), prior to qPCR amplification (assay 3).

FIG 1.

Schematic view of EAV detection by qPCR and RT-qPCR using three different assays (assays 1, 2, and 3).

Table 1 presents the results obtained with assays 1,2, and 3. To ensure that the comparisons between the assays were normalized and standardized, the RNA samples used in each of the three assays were from the same extraction process. Interestingly, all positive samples were detected by all three assays, indicating that isothermal amplification with Phi29 polymerase did not change the specificity of the OIE-recommended RT-qPCR (6). Moreover, no PCR signals were detected from the 12 negative samples, indicating that the assay was highly specific for EAV detection. Furthermore, the analytical sensitivity of each assay was evaluated and varied markedly depending on the assay performed. The mean copy numbers of viral genome for the 48 positive samples obtained with assays 1, 2, and 3 were 53.75 × 10³ ± 11.4 × 10³, 31 × 10⁴ ± 6.6 × 10⁴, and 112.5 × 10⁶ ± 55 × 10⁶, respectively. Our results show that isothermal amplification with Phi29 polymerase significantly increased the ratio of amplification compared to assay 1. Indeed, the detected amount of DNA with assay 3 was 10² to 10⁷ times the amount detected using assay 1 (Table 1). Therefore, compared to the OIE-recommended RT-qPCR, the analytical sensitivity of assay 3 is significantly enhanced (Student's t test, P < 0.05) (Table 1). These findings were obtained regardless of the sample's origin (culture fluid, raw semen, or homogenized organs).

In addition to enhancing the sensitivity of the current diagnostic method, the second objective of our study was to rapidly genotype the viruses detected. We therefore decided to combine the unbiased amplification of nucleic acids with a high-density resequencing microarray. The two EAV sequences tiled on the microarray cover a region of 267 bp (Table 2) located in the ORF 1 region encoding the nonstructural protein 9 (nsp9) that is involved in RNA synthesis (11). Amplified DNA samples were fragmented and labeled using the GeneChip resequencing assay kit (Affymetrix, Inc., Santa Clara, CA). After overnight hybridization at 45°C, the RMA was washed, stained, and scanned according to the manufacturer's instructions. The raw image file (.DAT) obtained after scanning was transformed into a fluorescence intensity file (.CEL) using GeneChip operating software (GCOS) (Affymetrix, Inc.). Bases were called using a derivative of the ABACUS base-calling algorithm (12). Sequences were output in FASTA format, and for each sequence obtained, the call rate value was calculated as the ratio of the number of determined bases (A, T, C, or G) to a sequence length of 243 bp (Table 3). Sequences retrieved from the microarray were used to perform an evolutionary study. The phylogenetic trees were constructed using the maximum-likelihood method, and the tree's statistical robustness was assessed by bootstrap resampling (100 data sets) of the multiple alignments. Phylogenetic reconstructions were performed using SeaView software, version 4 (13). In the literature, EAV isolates have been classified into two distinct groups: the North American group (NA) and the European group (EU) (14). The EU group can be subdivided into two subgroups known as European subgroup 1 (EU-1) and European subgroup 2 (EU-2) (14). This classification was obtained by aligning and comparing either the 3-kb sequences covering the ORF2-to-ORF7 region of the EAV genome, which encodes the structural protein, or a 518-nucleotide portion of ORF5, from positions 11,296 to 11,813, encoding viral glycoprotein 5 (14). A phylogenetic analysis with EAV sequences, retrieved from either GenBank or the 12 samples in our study (samples 1, 3, 4, 5, 6, 19, 25, 26, 32, 38, 43, and 48), was performed according to the internationally recommended rules described above. A phylogenetic tree constructed with the 518-nucleotide sequence obtained following ORF5 sequencing (Fig. 2a) confirmed the classification of known EAV strains. Moreover, samples 1 to 6, 19 to 38, and 43 to 48 were classified as being in the EU-1, EU-2, and NA groups, respectively (Fig. 2b). The products obtained after isothermal amplification from these 12 samples were hybridized on the RMA (Table 1). Surprisingly, the phylogenetic tree obtained with the nsp9 nucleotide sequences retrieved from the microarray separated the strains into the NA and EU groups and divided the EU group into subgroups EU-1 and EU-2 (Fig. 2b). Thus, this result indicates that the 243-nucleotide sequence from the ORF1 region amplified in our study could also be used for genotyping EAV.

TABLE 2.

EAV partial ORF1 sequences of two 267-nucleotide strains from the Arteriviridae family tiled on the resequencing microarray

Viral strain	Location along ORF 1	GenBank Accession no.
CW 96	6579–6845	AY349167
Bucyrus (EAV ATCC VR-796)	6575–6841	DQ846750

TABLE 3.

Call rate value for each of the 12 strains tested on the RMA^a

Sample no.	Call rate value (%)	Phylogenetic group
1	34.69	European subgroup 1
3	34.69	European subgroup 1
4	34.69	European subgroup 1
5	51.43	European subgroup 1
6	50.61	European subgroup 1
19	56.33	European subgroup 2
25	52.65	European subgroup 2
26	45.97	European subgroup 2
32	60	European subgroup 2
38	71.43	European subgroup 2
43	73.47	North American group
48	90.2	North American group

^aCall rate value is the ratio of the number of determined bases (A, T, C, or G) to a sequence length of 243 bp.

FIG 2.

Phylogenetic analysis of the partial ORF5 nucleotide sequences (518 nucleotides in length) (a) and partial ORF1 nucleotide sequences (243 nucleotides in length) (b) of 27 EAV strains. Twelve isolates are specific to this study, and 15 are from previously published EAV strains. Neighbor-joining and maximum-likelihood trees were constructed using SeaView version 4, and equivalent results were obtained. Only the maximum-likelihood trees are shown. The percentages of trees, >50 only, in which the associated taxa clustered together are shown next to the branches. The North American group and European subgroups 1 and 2 are indicated.

This study is the first in the scientific literature to use an unbiased isothermal amplification method using the Phi29 polymerase in combination with a high-density resequencing microarray for diagnosing nidoviruses. Furthermore, amplifying EAV RNA by whole transcription amplification, which requires an additional time of 10 h, provides much greater sensitivity than the OIE-recommended RT-qPCR method developed in 2002 by Balasuriya et al. (6, 7). Its association with a resequencing microarray has greatly reduced the time taken for genotyping even if it remains more expensive than reference RT-qPCR and Sanger sequencing of PCR amplicons. However, the resequencing microarray is less expensive than next-generation sequencing for EAV genotyping. This method can be recommended for the detection of EAV in semen and aborted fetuses, especially when the viral load is very low. In addition, this study validated the usefulness of the high-density resequencing DNA microarray for the diagnosis of equine viral diseases and the genotyping of RNA viruses such as equine arteritis virus.