Typing of Bovine Viral Diarrhea Viruses Directly from Blood of Persistently Infected Cattle by Multiplex PCR

Abstract

A nested multiplex PCR was developed for genotyping of bovine viral diarrhea viruses (BVDVs). The assay could detect as little as 3 50% tissue culture infective doses of BVDV per ml and typed 42 out of 42 cell culture isolates. BVDV was also successfully typed, with or without RNA extraction, from all 27 whole-blood samples examined from 22 carriers or probable carriers and 5 experimentally infected cattle.

Bovine viral diarrhea virus (BVDV) causes significant disease in cattle worldwide and has recently been targeted for eradication in several national programs in Europe (3). In addition to gastrointestinal disease, BVDV causes reproductive and respiratory disorders and persistent infections (for a review, see reference 1). Persistent, lifelong infections can occur when the fetus is infected in the first trimester of gestation (1). Persistently infected cattle, or carriers, usually succumb to mucosal disease, a fatal condition characterized by gastrointestinal erosion and severe diarrhea (4, 8). Because carriers are constantly viremic and continually shed and maintain the virus in the environment, their identification and removal from the herd is an essential component of programs for the control and eradication of BVDV (3, 6).

BVDV is a member of the Pestivirus genus in the family Flaviviridae (28). Recently, BVDV has been subdivided into two genotypes, BVDV1 and BVDV2 (21, 24). In addition to the above-mentioned diseases, virulent strains of BVDV2 cause severe thrombocytopenia with hemorrhage and a severe acute disease resembling mucosal disease (9, 12).

The ability to type BVDV is useful for diagnosis, for defining isolates, and for determining vaccine efficacy in herd health programs for the prevention of fetal infection. Several PCR-based assays have been developed for typing tissue culture isolates of BVDV (18, 24, 27). However, these assays were not applied to clinical samples. In this report, we describe a nested multiplex PCR that could type BVDV, with or without RNA extraction, directly from infected blood.

Primers for the PCR were designed from the NS5B gene (11). Since published sequences were limited, portions of the gene of five BVDV1 strains (Singer, New York 1, Oregon, DCP, and Hastings) and five BVDV2 strains (24301, BVD2-125c, Sl lake, Short, and MN fetus) (15, 24) were sequenced essentially as previously described (16). The external primers for primary PCR, 5′ AAGATCCACCCTTATGA(A/G)GC 3′ and 5′ AAGAAGCCATCATC(A/C)CCACA 3′, were derived from nucleotides 10385 to 10404 and 11528 to 11547, respectively (relative to BVDV-NADL [10]). The multiplex primers for secondary PCR, 5′ TGGAGATCTTTCACACAATAGC 3′ (BVDV1 specific), 5′ GGGAACCTAAGAACTAAATC 3′ (BVDV2 specific), and 5′ GCTGTTTCACCCAGTT(A/G)TACAT 3′, were derived from nucleotides 10758 to 10779, 10514 to 10533, and 11096 to 11117, respectively. Software used for primer design and synthesis of primers was as described previously (16).

RNA was extracted from 100 μl of 42 supernatants from BVDV-infected Madin-Darby bovine kidney cells and 32 infected blood or serum samples with TRIzol (Canadian Life Technologies, Burlington, Ontario, Canada) as described previously (16). Clinical samples included those from 14 carriers identified by virus isolation by the donor laboratory and 8 probable carriers (with a virus titer of ≥10⁴ 50% tissue culture infective doses [TCID₅₀]/ml) identified as viremic by PCR (17) by the donor laboratory. A carrier is defined as having virus in two blood samples obtained ∼30 days apart (6). In contrast, acutely infected cattle usually have intermittent viremia over only a few days, with lower viral titers.

Reverse transcription (RT) and PCR were combined in a single step. One microliter of the extracted RNA or sample was added to a reaction mixture (total volume of 50 μl) containing 2 mM MgCl₂, PCR buffer (20 mM Tris-HCl [pH 8.4], 50 mM KCl), 0.2 mM deoxynucleoside triphosphates (Pharmacia, Baie D’Urfe, Quebec, Canada), 0.25 μg of external primers, 5 U of RNAguard RNase inhibitor (Pharmacia), 50 U of Moloney murine leukemia virus reverse transcriptase (Canadian Life Technologies), and 1.25 U of Taq DNA polymerase (Canadian Life Technologies). RT was carried out at 37°C for 30 min, followed by denaturation at 94°C for 3 min. The reactions were cycled 25 times at 94°C for 20 s, 50°C for 30 s, and 72°C for 30 s, with a final extension step of 72°C for 15 min. The product (1 μl) was used in secondary PCR for 40 cycles. This was performed in the same manner as the primary PCR but with multiplex primers and without reverse transcriptase, RNase inhibitor, and external primers. Products were electrophoresed on a 2% agarose gel and stained with ethidium bromide.

Amplification products of 604 and 360 bp were predicted for BVDV2 and BVDV1, respectively. By using RNA extracted from the medium of infected cell cultures for RT-PCR, products consistent with those predicted were obtained (Fig. 1A, lanes 2 to 6). The products derived from the reference strains BVDV2-890 (23) and BVDV1-Singer were sequenced and confirmed to be BVDV specific. A total of 42 BVDV isolates were typed by PCR and tested against type-specific monoclonal antibodies (14, 15) in an immunoperoxidase assay (13). The typing results correlated perfectly except for one isolate not recognized by either antibody (Table 1). Cell culture virus was also tested directly, without RNA extraction, by PCR. Surprisingly, PCR products of the appropriate sizes were obtained from cultures of BVDV2 and BVDV1 (Fig. 1A, lanes 8 to 12).

FIG. 1.

Typing of BVDV by PCR. (A) BVDV from medium of infected cell cultures. PCR was carried out with RNA extraction (lanes 2 to 7) and without RNA extraction (lanes 8 to 13) with BVDV2 strains 890 (lanes 2 and 8) and BVD2-125c (lanes 3 and 9), BVDV1 strains Singer (lanes 4 and 10), New York 1 (lanes 5 and 11), and Hastings (lanes 6 and 12), and medium from an uninfected cell culture (lanes 7 and 13). (B) BVDV from whole blood. PCR was carried out with RNA extraction (lanes 2 to 7) and without RNA extraction (lanes 8 to 13) from the blood of five carriers (samples B1 to B5, listed in Table 2) (lanes 2 to 6 and 8 to 12, respectively) and BVDV-negative blood (lanes 7 and 13). Lanes 1, 123-bp ladder.

TABLE 1.

Multiplex PCR of BVDV

Virus(es) (no.)	Response to monoclonal antibody:		PCR product sizea (type)
	157	BA29
BVDV1 reference strains (6)b	+	−	360 (1)
U.S. and Canadian isolates (8)	+	−	360 (1)
Fetal bovine serum isolates (4)	+	−	360 (1)
Alberta bison isolate (1)	+	−	360 (1)
Alberta bison isolate (1)	−	−	360 (1)
BVDV2 reference strains (2)c	−	+	604 (2)
U.S. and Ontario isolates (19)d	−	+	604 (2)
Fetal bovine serum isolate (1)	−	+	604 (2)

^aEstimated product size in base pairs. 

^bBVDV1 strains NADL, Singer, New York 1, Oregon, DCP (15), and Hastings (15). 

^cBVDV2 strains 890 and BVD2-125c. 

^dIsolates associated with hemorrhagic syndrome or severe acute disease. 

To examine the specificity of the primers, other bovine viruses were tested. No PCR products were obtained for stocks of bovine herpesvirus 1, coronavirus, rotavirus, adenovirus 3, and parainfluenzavirus 3. A stock of respiratory syncytial virus gave a product of ∼604 bp. However, sequence analysis determined that the stock was contaminated with BVDV2.

Because the PCR produced prominent DNA products for different BVDVs propagated in cell culture and since carriers commonly have high BVDV titers (10³ to 10⁵ TCID₅₀/ml) (25), we examined the utility of the assay for clinical samples by testing paired samples of whole blood and serum from five carriers and five BVDV-negative cattle. The PCR detected and typed BVDV from all carrier samples when RNA was extracted and, remarkably, from all blood samples used directly without RNA extraction (Fig. 1B). Sera from carriers, however, gave inconsistent results when used directly (data not shown). No PCR products were obtained from paired samples from the 5 negative animals or from the blood of an additional 20 negative cattle subsequently tested.

We determined the lower limits of detection of the PCR with 10-fold serial dilutions of cell culture virus in duplicate and parallel assays. For virus diluted in medium, the limits were ∼30 and ∼50 TCID₅₀/ml for strains 890 and Singer, respectively, both when RNA was extracted and when the sample was used directly without RNA extraction (data not shown). For virus diluted in blood, the limits were ∼3 and ∼30 TCID₅₀/ml for strain 890 and ∼5 and ∼50 TCID₅₀/ml for strain Singer when RNA was extracted and by the direct method, respectively (Fig. 2). These results suggested that little, if any, inhibition of PCR by blood constituents occurred.

FIG. 2.

Sensitivity of typing from blood. (A) Sensitivity of PCR with RNA extraction. PCR was done with serial 10-fold dilutions in whole blood of strain 890 containing 3,000 to 0.3 TCID₅₀/ml (lanes 3 to 7, respectively) and strain Singer containing 5,000 to 0.5 TCID₅₀/ml (lanes 8 to 12, respectively) and with BVDV-negative blood and control without RNA (water) (lanes 2 and 13, respectively). PCR was positive to 3 and 5 TCID₅₀/ml for strains 890 (lane 6) and Singer (lane 11), respectively. (B) Sensitivity of PCR without RNA extraction. PCR was done with serial dilutions of strains 890 and Singer and controls (lanes are as described for panel A). PCR was positive to 30 and 50 TCID₅₀/ml for strains 890 (lane 5) and Singer (lane 10), respectively. Lanes 1, 123-bp ladder.

We increased our test panel of whole blood to a total of 27 samples from carriers and probable carriers and from five cattle experimentally infected with BVDV2 (Table 2). Some of these samples had undergone several freeze-thaw cycles before a virus titer was determined; thus, the titer expressed as “≥” must be considered a minimum of the original titer. Both RNA extraction and direct methods typed BVDV from all samples. One sample (no. 875) gave inconsistent results with the direct method. This was one of only a few samples that we were able to test as fresh, unfrozen blood. Although other samples generated PCR products without freeze-thaw cycles, sample 875 gave negative results with fresh blood but subsequently gave positive results upon freeze-thawing. Whether samples which are freeze-thawed give results superior to those of fresh, unfrozen samples when the direct method is used remains to be further evaluated.

TABLE 2.

Typing BVDV from whole blood samples by multiplex PCR

Sample, origin,a anticoagulantc	Clinical status of donor animalb	Virus titer (TCID50/ml)	PCR (typing result)
			With RNA extraction	Without RNA extraction
B1, Ohio	C	≥1 × 104	2	2
B2, Ohio	C	≥5 × 102	1	1
B3, Ohio	C	NDd	1	1
B4, Ohio	C	≥1 × 104	1	1
B5, Ohio	C	≥1 × 104	1	1
105, Ontario	C	≥3 × 104	1	1
106, Ontario	C	≥6 × 104	1	1
108, Ontario	C	≥2 × 104	1	1
109, Ontario	C	≥1 × 105	1	1
8, Iowae	C	≥4 × 104	1	1
262, Iowa	C	≥4 × 103	2	2
794, Iowa, C	C	3 × 104	1	1
854, Iowa, C	C	2 × 104	2	2
859, Iowa, C	C	1 × 104	2	2
358, Manitobae	PC	≥3 × 104	1	1
625, Manitoba	PC	≥1 × 104	1	1
287, Manitoba, H	PC	1 × 104	1	1f
291, Manitoba, H	PC	1 × 105	1	1f
292, Manitoba, H	PC	3 × 104	1	1f
296, Manitoba, H	PC	1 × 105	1	1f
875, Manitoba, H	PC	1 × 105	1	1fg
1086, Manitoba, C	PC	6 × 104	1	1f
E1-515p, C	EI	2 × 104	2	2f
E2-515p, C	EI	1 × 102	2	2f
E3-515p, C	EI	1 × 102	2	2f
E4-515u, C	EI	4 × 103	2	2
E5-p11Q, C	EI	5 × 101	2	2f

^aBlood tested was previously frozen unless stated otherwise (see footnote f). Origin, location of donating laboratory. 

^bC, carrier; PC, probable carrier; EI, experimentally infected with BVDV2 (blood collected on day 6 or 7 postinfection). 

^cH, heparinized; C, citrated; otherwise not recorded. 

^dND, not determined. This sample was negative at a 1/100 dilution in an immunoperoxidase assay but was positive in an antigen capture enzyme-linked immunosorbent assay. It was obtained from animal BJ, which had previously been shown to have very low serum titers (7). 

^eBVDV in all Iowa and Manitoba samples was previously typed by PCR-based methods with identical results as above (4a). 

^fSamples were tested fresh without previous freezing. 

^gProduced a type 1 product only after freeze-thawing. 

To determine the sensitivity of the PCR for blood from naturally infected animals, we tested blood from carrier animals, diluted in citrated BVDV-negative blood. When extracted RNA was used, the assay could detect ∼4 to 60 TCID₅₀/ml in samples 794, 854, 859, and 1086 (Table 2). The sensitivity of the direct method was equal to that observed with RNA extraction for two samples (no. 794 and 859) but less by factors of 5 and 10, respectively, for the other two samples. Based solely on volumes used in the two methods, the conventional method would be expected to be ∼4-fold more sensitive than the direct method.

In this report, we describe a PCR which can be used to type BVDV from infected cell cultures and blood. Remarkably, the PCR could be used on both types of samples without RNA extraction. This is an important advance, since RNA extraction procedures can be very laborious. Our positive PCR results without RNA extraction suggest that a portion of BVDV RNA in the blood of infected cattle may be virus free and resistant to degradation and thus was readily available for RT. Alternatively, RNA may have been released from virus and/or cells during sample handling or during RT-PCR. Although definitive experiments are planned to distinguish between these possibilities, initial experiments show that incubation in the RT-PCR mixture has a dramatic negative effect on the titer of infective virus, which lends some support to the second hypothesis.

Infected blood that was known to be either citrated or heparinized generated products by PCR (Table 2). However, citrate and EDTA are considered more suitable as anticoagulants than heparin, since the latter has been shown to inhibit PCR (19).

Previously, several PCR-based assays for typing BVDV were reported (18, 24, 27). Typing with one of these assays, however, was indirect and involved restriction endonuclease digestion of PCR products for typing (18). A second assay used the specific amplification of BVDV2 for typing, with a negative result indicating BVDV1 (24). The advantage of the multiplex PCR is that a specific product is produced for both genotypes. The multiplex PCR of Sullivan and Akkina (27) could type border disease virus (BDV), another pestivirus, in addition to BVDV1 and BVDV2. This is an advantage for typing pestiviruses from sheep, which can be infected with all three viruses, but in cattle BDV does not appear to be readily infectious. BDV has not been isolated from North American or European cattle (20, 22, 27), and only one bovine BDV isolate has ever been reported; its original isolation is thought to have been made in the 1960s (2).

The sequences from which the PCR primers were derived appear to be highly conserved, as all cell culture isolates and BVDV-positive samples examined generated a PCR product. The sensitivity of the assay, with and without RNA extraction, also appears to be sufficient for the identification of carriers.

Further studies are planned to evaluate these methods. Specifically, because colostral antibodies may interfere with BVDV detection by virus isolation and antigen capture enzyme-linked immunosorbent assay (5, 25, 26), studies to evaluate the PCR for screening the carrier status of young calves in herds are planned. PCR may be especially useful in screening beef herds, since early and accurate identification of carriers avoids excessive rounding up of cattle from pasture for testing. Our results indicate that the PCR may have a role in screening herds for carriers and that it is a valuable tool for typing BVDV from cell culture and directly from blood.

Nucleotide sequence accession numbers.

Sequences of the entire BVDV-Singer NS5B gene and portions of the NS5B gene of other BVDVs mentioned in this report were deposited in GenBank (accession no. AF078534 to AF078543).