Vaccine 26 (2008) 3594–3600

Contents lists available at ScienceDirect

Vaccine

journa l homepage: www.e lsev ier .com/ locate /vacc ine

Development of a porcine reproductive and respiratory syndrome virusdifferentiable (DIVA) strain through deletion of specific immunodominantepitopes

Marcelo de Limaa,b, Byungjoon Kwona, Israrul H. Ansari a, Asit K. Pattnaika,Eduardo F. Floresb, Fernando A. Osorioa,∗

a Nebraska Center for Virology and Department of Veterinary and Biomedical Sciences, University of Nebraska-Lincoln,Lincoln, NE 68583-0900, USAb Departmento de Medicina Veterinaria Preventiva, Universidade Federal de Santa Maria, Santa Maria, RS 97105-900, Brazil

(diffen of es studt serviru

he bacturae of tlong

utanttant

a r t i c l e i n f o

Article history:Received 14 March 2008Received in revised form 24 April 2008Accepted 30 April 2008Available online 19 May 2008

Keywords:PRRSVB-cell epitopesPeptidesInfectious cDNA cloneDIVA

a b s t r a c t

The availability of a DIVAthe control and eradicatiosyndrome (PRRS). Previourecognized by convalescenand respiratory syndromeepitopes can be used as tidentified on the non-strularge deletions). The choicdeletion were performed ato successfully rescue a mgrowth of the deletion mu

Marker vaccines by a commercial ELISA kit, witthe inoculated animals. In sumpotentially be developed by de

1. Introduction

Porcine reproductive and respiratory syndrome virus (PRRSV)is a small enveloped, positive-strand RNA virus associated withreproductive failure in pregnant sows and severe pneumonia inneonatal pigs [1]. Porcine reproductive and respiratory syndrome(PRRS) is currently one of the most important infectious diseasesof swine causing significant economic losses to the pig industryworldwide [2]. PRRSV, along with lactate dehydrogenase-elevatingvirus (LDEV), equine arteritis virus (EAV) and simian hemorrhagicfever virus (SHFV), is classified into the order Nidovirales, familyArteriviridae [1]. Based on significant antigenic and genetic dif-ferences reported among North American and European PRRSVstrains [3,4], two distinct genotypes of the virus have been rec-

∗ Corresponding author at: 111 K. Morrison Life Sciences Research Center, 4240Fair Street, East Campus, Lincoln, NE 68583-0900, USA. Tel.: +1 402 472 7809;fax: +1 402 472 9690.

E-mail address: fosorio@unl.edu (F.A. Osorio).

0264-410X/$ – see front matter © 2008 Elsevier Ltd. All rights reserved.doi:10.1016/j.vaccine.2008.04.078

rentiating infected from vaccinated animals) vaccine is very important forndemic infectious diseases such as porcine reproductive and respiratoryies in our laboratory identified several B-cell linear epitopes consistently

a obtained from pigs infected with a North American porcine reproductives (PRRSV) strain. To ascertain if one or more of these immunodominantsis of DIVA differentiation, we selected two epitope markers previously

l protein 2 (PRRSV NSP2, predictably the viral protein most likely to toleratehese epitopes was primarily based on their immunodominance and theirthe backbone of the wild-type cDNA infectious clone (FL12). We were ablethat fulfilled the requirements for a DIVA marker strain, such as: efficientin vitro and in vivo and induction of specific seroconversion as measured

h absence of a marker-specific peptide-ELISA response in 100% (n = 15) ofmary, our results provide proof of concept that DIVA PRRSV vaccines canletion of individual “marker” immunodominant epitopes.

© 2008 Elsevier Ltd. All rights reserved.

ognized: European (type I) and North American (type II). Althoughisolates from both genotypes can cause disease with similar clini-cal signs, their genomes exhibit divergences of approximately 40%[5].

Vaccination against PRRSV infections is being carried out since1995 in the United States. The most commonly used vaccine con-sists of a North American PRRSV strain attenuated by multiplepassages in cell cultures. The efficacy of these currently used atten-uated vaccines is somewhat controversial and is generally acceptedthat there is a need for improvements on their safety and efficacy. Inthis context, the availability of a DIVA (differentiating infected fromvaccinated animals) vaccine would be of great value for the controland eventual eradication of PRRS. Epidemiological as well as reg-ulatory considerations advise that a PRRSV DIVA vaccine shouldbe designed based on a negative marker (i.e., a marker absentfrom the vaccine strain but consistently present in wild-type (wt)strains). Classical examples of modified-live vaccines carrying dele-tions of non-essential and immunogenic structural proteins havebeen produced for large DNA viruses such as pseudorabies virus(PRV) and bovine herpesvirus-1 (BHV-1) [6–8]. While technically

cine 2

2.3. In vitro transcription, RNA electroporation, and recovery ofepitope deletion mutants

The full-length plasmid (pFL12) was digested with AclI and lin-earized DNA was used as the template to generate capped RNAusing the mMESSAGE mMACHINE Ultra T7 kit according to manu-facturer’s (Ambion) recommendations. Briefly, after in vitro RNAtranscription, the reaction mixture was treated with DNaseI todigest the DNA template, extracted with phenol/chloroform andfinally precipitated with isopropanol. Following electrophoresisthrough a glyoxal agarose gel, the integrity of the RNA transcriptswas analyzed upon ethidium bromide staining of the gel.

M. de Lima et al. / Vac

straightforward in the case of some double-stranded DNA viruses,deleting antigen-coding sequences from the genome of a small RNAvirus like PRRSV, which encode only a few proteins with essen-tial functions, seems to be a more difficult task [9–11]. Thus, thegeneration of a mutant virus carrying a deletion of an immunodom-inant and conserved protein segment (or a combination of deletionswithin the same protein or even in different proteins) would be anattractive alternative to generate a live-attenuated marker vaccinestrain.

Using a systematic and detailed approach, we previouslydemonstrated by Pepscan analysis the presence of B-cell linearepitopes in the non-structural protein 2 (NSP2) and structural pro-teins of a North American strain of PRRSV, consistently recognizedby the humoral immune response elicited by PRRSV-infected pigs[12]. In this context, the selection of immunodominant epitopesand deletion of these regions in a full-length infectious cDNA clonewould be an alternative approach for the development of markerlive-attenuated PRRSV vaccines since the genome of small RNAviruses, like PRRSV, generally does not tolerate less subtle changessuch as deletions of entire genes [9–11]. Based primarily on theimmunodominance and level of amino acid conservation observedfor some of the peptides distributed along the different proteins,we selected target epitopes (serological markers candidates) to bedeleted in the wt infectious cDNA clone (FL12) by reverse genetics.

The approach of epitope deletion has proved to be fea-sible for arteriviruses through deletion of a 46 amino acidimmunodominant region from the ectodomain of the glycopro-tein L (gL) of EAV without deleterious effects on the replicationand immunogenicity of the virus [13]. Furthermore, a peptideELISA based on this particular domain enabled serological dis-crimination between vaccinated and wt virus-infected animals[13].

The present study describes the generation of a PRRSV dele-tion mutant, named FLdNsp2/44, which lacks amino acid residues431–445 within the sequence of the NSP2 derived from a US-typestrain of PRRSV. In order to explore the potential of this approach togenerate a live-attenuated marker vaccine against PRRSV, we eval-uated the replication efficacy in vitro of the epitope deletion mutantand its biological properties in vivo, such as replication efficiencyand immunogenicity in pigs as well as virulence in a pregnant sowmodel.

2. Material and methods

2.1. Cells, viruses and antibodies

The virus used for animal inoculation was recovered fromMARC-145 cells transfected with RNA transcripts produced invitro from the full-length infectious cDNA clone (FL12) derivedfrom PRRSV NVSL 97-7895 type II strain (GenBank accession no.AY545985) [14]. MARC-145 cells were propagated in Dulbecco’sModified Eagle’s Medium (DMEM) containing 10% fetal bovineserum (FBS) and antibiotics (100 units/ml of penicillin, 20 �g/mlof streptomycin, and 20 �g/ml of kanamycin) and used for RNAelectroporation, viral infection and amplification, virus titrationand experiments of growth kinetics. Indirect immunofluores-cence assays (IFA) were performed as previously described [15],using N protein-specific monoclonal antibodies (MAbs): SDOW17(National Veterinary Services Laboratories – NVSL, Ames, IA) andSR30 [kindly provided by Dr. Eric Nelson (South Dakota StateUniversity, SD, USA)]. The secondary antibody used was a goat anti-mouse IgG antibody (Alexa Fluor – 488, Molecular Probes, Eugene,OR).

6 (2008) 3594–3600 3595

2.2. Introduction of deletions into the full-length PRRSV cDNAinfectious clone

In order to explore deletion of possible markers, two regionswere previously selected for deletion in the NSP2. Each regionspans 15 amino acids residues which were found to be highlyimmunogenic, relatively conserved among US-type PRRSV strainsand most importantly, consistently recognized by antibodiesof PRRSV-infected pigs [12]. The regions selected as sero-logical marker candidates were P431PPPPRVQPRKTKSV445 andK441TKSVKSLPGNKPVP455 within NSP2 amino acid sequence. Alldeletions were introduced into the pFL12 plasmid which containsthe full-length cDNA of NVSL 97-7895 PRRSV strain [14], by over-lap extension method as previously described [16]. Sequences fromeither side of the point of deletion were amplified by using spe-cific primers designed such that their 3′ ends hybridize to templatesequence on one side of deletion and the 5′ ends are comple-mentary to template sequence on the other side of the deletion(Table 1). Using this approach the products generated from thePCR reaction using the reverse and forward primers with overhangregions are therefore overlapping at the deletion point. After ampli-fication of the flanking regions, both amplicons were recoveredby phenol/chloroform extraction and precipitated with ethanolusing standard protocol as described elsewhere [17]. The equimo-lar amounts of the two amplicons were mixed and subjected to 4–5cycles of PCR followed by the additional rounds of PCR amplifica-tion with external primers. After gel purification and precipitationthe DNA was digested with SpeI and SphI or XhoI and the frag-ment was cloned directionally into the pFL12. Confirmation of thedeletion and absence of any other mutations within the region wasconfirmed by nucleotide sequencing.

Sub-confluent monolayers of MARC-145 cells were used forelectroporation with approximately 5 �g of in vitro produced tran-scripts along with 5 �g of carrier RNA isolated from uninfectedMARC-145 cells. About 2 × 106 cells in 400 �l of DMEM contain-ing 1.25% DMSO were pulsed once using Bio-Rad Gene Pulser Xcellat 250 V, 950 �F in a 4.0-mm cuvette. After this treatment, thecells were diluted in normal growth media containing 10% of fetalbovine serum and placed into a 60-mm cell culture plate. The

Table 1Primers used for amplification of specific fragments in each of the selected targetsfor deletion

Protein Primer Nucleotide sequence (5′–3′)

NSP2 d44Rev GTTCCCTGGCAAGCTCTTCGGTGTCCACCGTGGNSP2 d44For CCCACGGTGGACACCGAAGAGCTTGCCAGGGAACNSP2 d45Rev CTGACCTTCCTGCGTGGAGCTCGAGGCTGAACTCTTGGNSP2 d45For CCAAGAGTTCAGCCTCGAGCTCCACGCAGGAAGGTCAG

Primers were designed such that their 3′ ends hybridize to template sequence onone side of deletion and the 5′ ends are complementary to template sequence on theother side of the deletion. The selected protein and primers with its 5′–3′ nucleotidesequence are shown.

cine 2

3596 M. de Lima et al. / Vac

expression of N protein at 24 h post RNA electroporation was anindicator of genome replication and transcription. After confirmingexpression of N protein using indirect immunofluorescence assay,the culture supernatant from electroporated cells was collected at72 h post-electroporation, clarified and inoculated onto a confluentmonolayer of uninfected MARC-145 cells. The cells were then mon-itored on a daily basis for characteristic cytopathic effect (CPE) andalso examined for expression of N protein. Culture supernatantsfrom infected cells showing both CPE and positive fluorescencewere assessed as containing infectious virus. The rescued virus wasamplified and small aliquots were stored at −80 ◦C for further stud-ies. In all the experiments, FL12 (containing wt PRRSV genome) andFL12 pol− (containing polymerase-defective PRRSV genome) wereused as positive and negative controls respectively, as describedelsewhere [14].

2.4. Kinetics of virus growth

MARC-145 cells were infected either with FL12 or FLdNsp2/44at an MOI of 0.1 TCID50 per cell and incubated at 37 ◦C in 5% CO2atmosphere. Aliquots of culture supernatants from infected cellswere collected at different time points (0, 6, 12, 24, 48 and 72 h post-infection, pi) and the virus titer was determined by 50% end-pointanalysis and the titers were expressed as tissue culture infectiousdose 50 per ml (TCID50/ml). The viral growth kinetics assays wereperformed in triplicate.

2.5. Animal experiments

Two groups of fifteen mixed-breed (Landrace × Large White)piglets averaging 3 weeks of age obtained from a PRRSV-free farmwere allocated in three BL-2 isolation rooms. At the beginningof the experiment, all animals were tested negative for PRRSV-specific antibodies as measured by a commercially available ELISAkit (IDEXX Labs, Inc.). The animals from the experimental groupswere inoculated with a total dose of 105.0 TCID50/3 ml of PRRSV FL12or the epitope deletion mutant (FLdNsp2/44) by intranasal (1 mlin each nostril) and intramuscular (1 ml) routes. The inoculatedanimals were clinically monitored on a daily basis and their rec-tal temperatures were recorded from day 3 pre-inoculation to day15 post-infection. Sequential blood samples were collected from allanimals at days 0 (zero), 7, 15, 30, 45 and 60 pi. In order to assess thevirulence of the epitope deletion mutant (FLdNsp2/44), we inocu-lated two pregnant sows at day 90 of gestation by intranasal route

with 105.0 TCID50/ml of the mutant virus to evaluate the viabil-ity scores of the offspring at birth and weaning. The sows wereacquired from a specific pathogen-free herd with a certified recordof absence of PRRSV infection. Inoculated sows were clinically mon-itored from 3 days prior to inoculation until 15 days post-farrowing.Blood samples were collected at days 0 (zero), 7, 15, 30, 45 and 60pi for serological and virological analyses.

2.6. Peptide ELISA

Serum samples collected until day 60 pi from all piglets exper-imentally infected with FL12 (wt strain) and FLdNSP2/44 (epitopedeletion mutant) and from the sows infected with the PRRSV dele-tion mutant were tested using a peptide-based ELISA for screeningof the peptide-specific antibody response, as previously described[12]. Briefly, Immulon 2HB flat bottom microtiter 96 well plates(Thermo Electron, Milford, MA) were coated with 100 �l of apeptide (P431PPPPRVQPRKTKSV445) solution (10 �g/ml) in 0.1 Mcarbonate buffer (pH 9.6), and incubated overnight at 4 ◦C. Afterblocking with 250 �l of a 10 wt.% nonfat dry milk solution for 4 hat room temperature on a plate shaker, the plates were washed

6 (2008) 3594–3600

three times with PBS containing 0.1% Tween 20 (PBST-20). Unboundreagents were further removed by striking the plates repeatedly,bottom up, on a stack of absorbent paper towel. Then, 100 �l ofpig sera (1:20) diluted in 5 wt.% nonfat dry milk in PBST-20 wasadded per well and plates were incubated in the shaker for 1 h atroom temperature. After washing five times with PBST-20, eachwell received 100 �l of the affinity purified antibody peroxidaselabeled goat anti-swine IgG (KPL, Gaithersburg, MD) diluted 1:2000in PBST-20 with 5 wt.% nonfat dry milk for and the plate was incu-bated for 30 min at room temperature. Following a final wash,100 �l of ABTS (KPL) peroxidase substrate was added for 15 minat 37 ◦C and the reaction was stopped by adding 100 �l of 1% SDS.Sera were considered positive when the OD value was above thecut-off point (the mean OD absorbance at 405 nm of the negativesera plus three standard deviations).

2.7. Sequence analysis

Multiple alignment of nucleotide and amino acid sequencesobtained either from Gene Bank data base or from the sequenc-ing facilities that analyzed our mutants were made using ClustalW[18]. Alignments were retrieved and analyzed by Bio-Edit sequencealignment editor v.7.0.5.

3. Results

3.1. Recovery of a mutant PRRSV lacking an immunodominantB-cell epitope of NSP2

In order to investigate whether an immunodominant regionwithin NSP2 is dispensable for viral replication, we introduceda deletion of 45 nucleotides into the full-length cDNA infectiousclone of PRRSV (FL12) [14]. Two mutant PRRS viruses namedFLdNsp2/44 and FLdNsp2/45 carrying deletions of amino acidresidues P431PPPPRVQPRKTKSV445 and K441TKSVKSLPGNKPVP455,respectively were successfully rescued after RNA electroporationinto MARC-145 cells. Such segments correspond to immunodom-inant epitopes located at a predicted hydrophilic domain of NSP2,and they were found to be consistently recognized by antibod-ies from pigs experimentally infected with a North Americanstrain of PRRSV (NVSL97-7895) [12]. Based on a slightly higherdegree of conservation of the deleted epitope (data not shown),we then selected one of the mutant strains (FLdNsp2/44) asthe mutant virus for in vitro and in vivo experiments. Although

the presence of the 45 nucleotide deletion introduced into theFLdNsp2/44 mutant had been previously verified by sequencing ineach of the cloning steps, it was further confirmed after sequenc-ing of one fragment amplified by RT-PCR directly from the culturesupernatant collected from infected MARC-145 cells (data notshown).

3.2. In vitro growth kinetics of the epitope deletion mutant(FLdNsp2/44) and its parental virus (FL12)

In order to investigate the possible effects of the deletion of theNSP2 epitope on virus growth, MARC-145 cells were infected inparallel (at an MOI of 0.1) with either FL12 or FLdNsp2/44 strains.Aliquots of culture supernatants from infected cells were collectedat different time points after infection and the virus titer wasdetermined. The viral growth kinetics assays were performed intriplicate and the mean titer values obtained in each time point canbe observed in Fig. 1. The multiple-step growth curves revealedthat the mutant virus exhibited a delayed onset of extracellularvirus accumulation, reaching its maximum titer approximately 24 hlater than the parental virus. Although the titer reached by the

cine 2

capacity to grow in vitro when compared to the wt virus despitethe lower titers reached in the first passages in cell culture (Fig. 1,

M. de Lima et al. / Vac

Fig. 1. Growth kinetics of the epitope deletion mutant (FLdNsp2/44) and its parentalvirus strain (FL12). MARC-145 cells were infected at an MOI of 0.1 and culture super-natant was collected at different time points after infection. Virus infectivity wascalculated by 50% end-point titration according to the Reed and Muench method.Results represent the mean values obtained in three independent experiments.

mutant virus was lower than that of the wt, similar virus yieldswere obtained for both viruses in subsequent passages in cell cul-ture (data not shown). These results suggest that the removal of theNSP2 B-cell epitope does not adversely affect the efficiency of viralreplication in vitro.

3.3. Serological response to the mutant virus assessed in weanedpiglets

The serological response to the mutant virus was assessedby inoculation of two groups of 15 piglets with a total dose of105.0 TCID50 of either FL12 or FLdNsp2/44 by intranasal and intra-muscular routes. Piglets inoculated with FL12 were served aspositive post-infection serologic control group. No significant clin-ical alteration was observed in either of both groups of animalsafter virus inoculation, except for a slight increase in rectal tem-perature (≤1.5 ◦C) observed between days 2 and 5 post-infection inthe group infected with the wt virus (FL12). In both experimentalgroups, viremia was detected from days 4 to 15 pi with a peak inviral titers on day 7 pi (103.32 ± 0.22 TCID50/ml for FLdNsp2/44 and104.28 ± 0.49 TCID50/ml for FL12) in the serum samples indicatingactive viral replication (Table 2). Sera collected from the experi-mentally infected animals in both groups, through days 15–60 pi,were positive for PRRSV-specific antibodies as assayed by a com-mercially available ELISA kit (IDEXX Labs, Inc.) (Fig. 2). A 1076 bpgenomic fragment was amplified by RT-PCR from virus isolated

Fig. 2. Serum antibody titers in piglets experimentally inoculated with PRRSV FL12strain (wild-type virus) or FLdNsp2/44 (epitope deletion mutant). S/P ratios wereexpressed according to a commercial IDEXX ELISA kit and the mean values obtainedfrom the serum samples collected at 0, 7, 15, 30, 45 and 60 days post-infectionfrom the 15 piglets infected in each group are shown. Horizontal bars represent thestandard deviation. A dashed line at 0.4 S/P ratio corresponds to the threshold valueabove which samples are considered positive for PRRSV-specific antibodies.

6 (2008) 3594–3600 3597

from sera of five FLdNsp2/44-infected piglets at day 10 pi andsequenced, confirming the identity of the epitope deletion mutant.The results of virus isolation from blood and subsequent positiveserology demonstrated that a productive virus replication occurredin all experimentally infected piglets. This data indicates that theepitope deletion does not negatively affect the integral humoralimmune response elicited by the mutant virus as detected by theroutine commercial ELISA that detects a polyclonal response to theN antigen (IDEXX Laboratories kit, Portland Maine). In contrast,when the peptide 44-specific ELISA was used to assess serologyof these pigs, we observed that pigs inoculated with the epitopedeletion mutant (FLdNsp2/44) did not develop antibodies againstthe peptide (ep#44) corresponding to the deleted epitope, whereasa strong reactivity was observed in the sera obtained from animalsinfected with the wt virus (FL12) using the same peptide-basedELISA (Fig. 3).

3.4. The mutant virus is stable in cell cultures and in vivo

An important requirement for the candidate marker deletionis that the referred deletion be stable throughout the in vivo infec-tion. In this context, the rescued mutant virus also exhibited similar

also discussed in Section 3.1). In addition, the duration of viremiain piglets upon experimental inoculation was comparable to theparental virus, although lower titers have been detected in thesera of the infected animals through days 4–15 post-inoculation(Table 2). A fragment of 1076 bp, encompassing the deleted regionwithin NSP2 was amplified by RT-PCR from virus isolated from seraof five infected piglets at day 10 pi. Nucleotide sequencing of thefragment confirmed the consistency of the epitope deletion mutantin all five samples indicating that the deletion was stable after sev-eral rounds of replication in vivo.

3.5. Assessment of virulence of the PRRSV epitope deletionmutant in a pregnant sow model

In order to verify whether the presence of the deletion in theNSP2 gene affected the overall biological properties of the mutantstrain (FLdNsp2/44) we focused on the measurement of virulence.To that end, a pregnant sow model which has been extensivelyused in our lab was used to assess virulence of wt and mutant

Fig. 3. Serological response of pigs experimentally infected with 105 TCID50/ml ofFL12 (wild-type) or FLdNsp2/44 (epitope deletion mutant) from day 0 to day 60 post-infection. The values represent the mean optical density obtained in a peptide-basedELISA where plates were coated with a 15 mer peptide corresponding to the deletedregion in FLdNsp2/44. Horizontal bars indicate the standard deviation and a dashedline corresponds to the cut-off point.

cine 2

r FLd

perim

3598 M. de Lima et al. / Vac

Table 2Viremia in 15 pigs experimentally infected with 105 TCID50/ml of FL12 (wild-type) o

Group 4 dpi 7 dpi

FLdNsp2/44 2.88 (±0.38) 3.32 (±0.22)FL12 3.86 (±0.43) 4.28 (±0.49)

Infectivity is expressed as mean log10 PRRSV titer TCID50/ml in the sera of the 15 exa Thirteen out of 15 pigs showed detectable levels of virus in sera.b ND: not detectable (<101.7 TCID50/ml).

viruses [15,17]. We evaluated the phenotype of the mutant virususing the reproductive failure model assessed by PRRSV challengeof pregnant sow at 90th of gestation and measuring the viabilityof the offspring at birth and upon weaning at 15 days of age. Wepostulated that this deletion would not affect the overall biology ofthe virus, as assessed by the parameters of in vivo replication andtransplacental pathogenesis that characterize the reproductivefailure model. The virulent phenotype of FL12 was unequivocallydemonstrated by the high rate of mortality at birth (Table 3). Like-wise, the epitope deletion mutant (FLdNsp2/44) was identicallyvirulent to FL12 when inoculated in pregnant sows, inducing abor-tion rates comparable to its parental virus (Table 3). In one sow,abortion was observed 10 days after infection with FLdNsp2/44 andpresence of the authentic mutant virus was confirmed by real-timePCR from RNA extracted from thoracic fluid and tissues collectedfrom aborted fetuses (data not shown). As shown in Table 3, twopiglets born from one sow infected with the mutant virus werealive at 15 days post-farrowing; however, they were euthanizeddue to weakness and poor health condition. These results showthat deletion of 15-mer NSP2 marker epitope did not contribute

for attenuation of the mutant virus in the pregnant sow model.

4. Discussion

These experiments resulted in the generation of a PRRSV car-rying a deletion of an immunodominant B-cell linear epitopecontained in the NSP2 gene. This epitope had been previ-ously identified as being consistently recognized by the antibodyresponse in PRRSV-infected animals [12]. The successful recov-ery of the epitope deletion mutant (FLdNsp2/44) and its activereplication in pigs, demonstrates that this specific region isdispensable for virus replication in vitro and in vivo. Previousreports have shown that the NSP2 replicase protein of PRRSVtolerates insertions [19], and deletions [20], thus representingan ideal target for the development of marker vaccines [21].Most importantly, our mutant exhibited an efficient growth invitro and in vivo and the induction of specific seroconversion asmeasured by a commercial ELISA kit, with absence of a marker-specific peptide-ELISA response in 100% (n = 15) of the inoculatedanimals, thus fulfilling the requirement for DIVA marker differen-tiation.

Table 3In vivo phenotype of the wild-type virus (FL12) and epitope deletion mutant(FLdNsp2/44) assessed in a reproductive failure model in pregnant sows (90 days ofgestation)

Group Sow # Offspring Viability at birth Viability at 15 days

Dead Live Live

FLdNsp2/44 1 18 18 0 02 14 11 3 2

FL12a 1 16 13 3 02 14 13 1 0

The viability scores of offspring at birth and 15 days after farrowing are indicated.a Data obtained from a previous experiment [15].

6 (2008) 3594–3600

Nsp2/44 (epitope deletion mutant)

10 dpi 15 dpi 30 dpi

2.55 (±0.46) 2.29 (±0.42)a NDb

4.25 (±0.38) 3.41 (±0.20) ND

entally infected pigs from days 4 to 30 post-infection ± the standard deviation.

To test whether the epitope deletion made on the mutant hadaltered the general biology of the strain, we examined the in vivophenotype of the epitope deletion mutant by inoculating two preg-nant sows at day 90 of gestation. Infection of pregnant sows withvirulent PRRSV strains, specifically with our wt infectious cloneFL12, invariably results in abortion, mummified or stillbirth pigletsbeing a reproductive failure model to study virulence [15]. Infec-tion of the pregnant sows with FLdNsp2/44 resulted in abortionin similar levels when compared to the wt virus indicating thatthe removal of the 45 nucleotides from NSP2 did not result inattenuation. Although the absence of virulence will be an essentialrequirement for the final live virus vaccine candidate, we did notexpect that any significant level of attenuation could be attributedto the 15 amino acid deletion. Current research in our laboratoryindicates that the main determinants of virulence appear to belocated in several regions of the PRRSV genome other than NSP2,such as NSP3-8, ORF 2 and ORF5 (Kwon et al., April 2008, submittedfor publication).

The use of marker vaccines and their companion differentialdiagnostic tests have become popular and even mandatory incampaigns aiming towards the control and eradication of animaldiseases with economic impact in national and international trade[8,22]. The best examples of eradication programs successfullyachieved by the use of marker vaccines are those of pseudora-bies and infectious bovine rhinotracheitis [8]. The developmentof the marker vaccine for these dsDNA viruses consisted of thedeletion of a gene encoding an immunogenic, conserved proteinand dispensable for viral replication, without impairing the wholeimmunogenicity of the vaccine virus [8]. Other examples of markervaccines include licensed subunit vaccines against classical swinefever (CSF) based on the recombinant expression of the E2 of CSFvirus in a baculovirus system [23], and its differential serologi-cal test [24]. However, it is generally accepted that there is stillneed of improvement regarding the efficacy of these new sub-

unit vaccines when compared to the classical live-attenuated CSFVvaccine [25,26]. Promising results have also been described foranother flavivirus, the bovine viral diarrhea virus, by using a similarmethodology as described for CSFV despite the incomplete pro-tection conferred against a wide-range of BVDV isolates [27,28].Further approaches aiming the development of a marker vaccinehave also been reported for Newcastle disease virus (NDV) eitherby construction of recombinant chimeric viruses [29], or deletionof immunodominant epitopes [30]. Recently, vaccine candidatesof Rift Valley fever (RVF) virus were generated by reverse genet-ics through deletion of non-structural genes, allowing differentialidentification of vaccinated and infected animals [31].

Vaccination against PRRSV infections is being carried out inthe United States since 1995. The most common vaccine cur-rently in use contains a North American PRRSV strain attenuatedby multiple passages in cell culture. Modified-live vaccines areoften preferred than inactivated-virus formulations due to itshigher efficacy and long-lasting immunity induced. Despite theavailability and extensive use of live-attenuated PRRSV vaccinesin the United States, an important drawback of these vaccinesconsists of the lack of a marker feature. Control and eventual

cine 2

[

[

[

[

[

[

[

[

[

M. de Lima et al. / Vac

eradication of PRRSV infections could be more easily achieved bya systematic vaccination program with a new generation of DIVAvaccines accompanied with a differential diagnostic test whichallows serological discrimination between vaccinated and naturallyinfected animals. This technical improvement on the current atten-uated PRRSV vaccines would be a significant advance for PRRSVvaccinology and consequently highly desirable in eradication pro-grams.

While the development of this mutant FLdNsp2/44 supports theconcept of development of live PRRSV strains that can be differ-entiated serologically by a specific peptide ELISA, the particularmarker herein reported still falls short of being the ideal DIVAmarker candidate for PRRSV vaccines. While the NSP2 epitopeused for deletion in these experiments to generate FLdNsp2/44was definitely recognized as immunodominant in our previousresearch [12], evaluation of field sera with the peptide specificELISA indicates that the level of conservation of this particular NSP2epitope is low, with just about 25% of field sera from wt PRRSV-infected animals exhibiting positive reactivity against this epitope(unpublished research). Therefore, even if this epitope is immun-odominant, thus inducing antibody response in the majority ofanimals infected with the same strain, its level of conservationamong PRRSV strains seems relatively low or suboptimal. Effortsare therefore now concentrated on obtaining deletions or substitu-tions of the most conserved markers previously reported, which arebased mainly in regions located at the endodomain of the M protein[12]. In this case a particular challenge is posed by the fact that thiswell-conserved epitopes are not amenable to deletion, thus oblig-ing us to attempt the elimination of the antigenicity of the markersequence without losing viability of the strain using single pointmutations (Osorio et al., unpublished ongoing research).

In summary, in vitro and in vivo characterization of theFLdNsp2/44 showed that removal of a 15-mer NSP2 epitope hadno effect on immunogenicity, growth properties or virulence ofthe mutant virus. In addition, pigs inoculated with FLdNsp2/44 didnot develop antibodies to the selected epitope as measured by apeptide-based ELISA, whereas a strong reactivity was observed inthe sera derived from animals infected with the wt virus (FL12).Taken together, our results provide proof of concept demonstrat-ing the feasibility of constructing a PRRSV live-attenuated markervaccine by deleting an immunodominant B-cell linear epitope fromthe vaccine strain. In addition, the combination of a mutant viruscarrying an epitope deletion and its corresponding peptide-basedELISA represents an attractive approach for the development of

PRRS differential vaccines. As this NSP2 epitope maker would besuboptimal for widespread detection of PRRSV wt strains due to thelow conservation and high variability of NSP2, efforts should nowfocus on identifying an optimal marker such as those in PRRSV Mendodomain [12], currently under investigation in our laboratory.

Acknowledgments

We gratefully thank Dr. Debasis Nayak and Dr. Subash Dasfor technical support. This research has been supported primar-ily by a grant from the National Pork Board (NPB#05-159) andalso partially by a USDA NRICGP (project 2004-01576). The animalexperiments described in this paper were reviewed and approvedby the Institutional Animal Care Committee of the University ofNebraska-Lincoln under protocol IACUC no. 04-08-046.

References

[1] Snijder EJ, Meulenberg JM. Arteriviruses. In: Kniper D, et al, editors. Fieldsvirology, vol. 1, 4th ed. Philadelphia: Lippincott Williams & Wilkins; 2001. p.1205–20.

[

[

[

[

[

[

[

[

[

[

[

6 (2008) 3594–3600 3599

[2] Ropp SL, Wees CE, Fang Y, Nelson EA, Rossow KD, Bien M, et al. Characteriza-tion of emerging European-like porcine reproductive and respiratory syndromevirus isolates in the United States. J Virol 2004;78(7):3684–703.

[3] Allende R, Lewis TL, Lu Z, Rock DL, Kutish GF, Ali A, et al. North American andEuropean porcine reproductive and respiratory syndrome viruses differ in non-structural protein coding regions. J Gen Virol 1999;80(Pt 2):307–15.

[4] Wensvoort G, de Kluyver EP, Luijtze EA, den Besten A, Harris L, Collins JE, etal. Antigenic comparison of Lelystad virus and swine infertility and respiratorysyndrome (SIRS) virus. J Vet Diagn Invest 1992;4(2):134–8.

[5] Nelsen CJ, Murtaugh MP, Faaberg KS. Porcine reproductive and respiratorysyndrome virus comparison: divergent evolution on two continents. J Virol1999;73(1):270–80.

[6] Moormann RJ, de Rover T, Briaire J, Peeters BP, Gielkens AL, van Oirschot JT. Inac-tivation of the thymidine kinase gene of a gI deletion mutant of pseudorabiesvirus generates a safe but still highly immunogenic vaccine strain. J Gen Virol1990;71(Pt 7):1591–5.

[7] Kaashoek MJ, Moerman A, Madic J, Rijsewijk FA, Quak J, Gielkens AL, et al. Aconventionally attenuated glycoprotein E-negative strain of bovine herpesvirustype 1 is an efficacious and safe vaccine. Vaccine 1994;12(5):439–44.

[8] van Oirschot JT. DIVA vaccines that reduce virus transmission. J Biotechnol1999;73(2–3):195–205.

[9] Yoo D, Welch SK, Lee C, Calvert JG. Infectious cDNA clones of porcine reproduc-tive and respiratory syndrome virus and their potential as vaccine vectors. VetImmunol Immunopathol 2004;102(3):143–54.

10] Wissink EH, Kroese MV, van Wijk HA, Rijsewijk FA, Meulenberg JJ, Rottier PJ.Envelope protein requirements for the assembly of infectious virions of porcinereproductive and respiratory syndrome virus. J Virol 2005;79(19):12495–506.

11] Welch SK, Jolie R, Pearce DS, Koertje WD, Fuog E, Shields SL, et al. Construc-tion and evaluation of genetically engineered replication-defective porcinereproductive and respiratory syndrome virus vaccine candidates. Vet ImmunolImmunopathol 2004;102(3):277–90.

12] de Lima M, Pattnaik AK, Flores EF, Osorio FA. Serologic marker candidates iden-tified among B-cell linear epitopes of Nsp2 and structural proteins of a NorthAmerican strain of porcine reproductive and respiratory syndrome virus. Virol-ogy 2006;353(2):410–21.

13] Castillo-Olivares J, Wieringa R, Bakonyi T, de Vries AA, Davis-Poynter NJ, Rot-tier PJ. Generation of a candidate live marker vaccine for equine arteritis virusby deletion of the major virus neutralization domain. J Virol 2003;77(15):8470–80.

14] Truong HM, Lu Z, Kutish GF, Galeota J, Osorio FA, Pattnaik AK. A highlypathogenic porcine reproductive and respiratory syndrome virus generatedfrom an infectious cDNA clone retains the in vivo virulence and transmissibilityproperties of the parental virus. Virology 2004;325(2):308–19.

15] Kwon B, Ansari IH, Osorio FA, Pattnaik AK. Infectious clone-derived viruses fromvirulent and vaccine strains of porcine reproductive and respiratory syndromevirus mimic biological properties of their parental viruses in a pregnant sowmodel. Vaccine 2006;24(49–50):7071–80.

16] Ho SN, Hunt HD, Horton RM, Pullen JK, Pease LR. Site-directed mutagenesis byoverlap extension using the polymerase chain reaction. Gene 1989;77(1):51–9.

17] Ansari IH, Kwon B, Osorio FA, Pattnaik AK. Influence of N-linked glycosyla-tion of porcine reproductive and respiratory syndrome virus GP5 on virusinfectivity, antigenicity, and ability to induce neutralizing antibodies. J Virol2006;80(8):3994–4004.

18] Thompson JD, Higgins DG, Gibson TJ. CLUSTALW: improving the sensitiv-ity of progressive multiple sequence alignment through sequence weighting,position-specific gap penalties and weight matrix choice. Nucl Acids Res1994;22(22):4673–80.

19] Kim DY, Calvert JG, Chang KO, Horlen K, Kerrigan M, Rowland RR. Expressionand stability of foreign tags inserted into nsp2 of porcine reproductive andrespiratory syndrome virus (PRRSV). Virus Res 2007;128(1–2):106–14.

20] Han J, Liu G, Wang Y, Faaberg KS. Identification of nonessential regions of thensp2 replicase protein of porcine reproductive and respiratory syndrome virusstrain VR-2332 for replication in cell culture. J Virol 2007;81(18):9878–90.

21] Fang Y, Kim DY, Ropp S, Steen P, Christopher-Hennings J, Nelson EA, et al.Heterogeneity in Nsp2 of European-like porcine reproductive and respiratorysyndrome viruses isolated in the United States. Virus Res 2004;100(2):229–35.

22] Babiuk LA. Broadening the approaches to developing more effective vaccines.Vaccine 1999;17(13–14):1587–95.

23] Moormann RJ, Bouma A, Kramps JA, Terpstra C, De Smit HJ. Development of aclassical swine fever subunit marker vaccine and companion diagnostic test.Vet Microbiol 2000;73(2–3):209–19.

24] Floegel-Niesmann G. Classical swine fever (CSF) marker vaccine. Trial III. Eval-uation of discriminatory ELISAs. Vet Microbiol 2001;83(2):121–36.

25] van Oirschot JT. Vaccinology of classical swine fever: from lab to field. VetMicrobiol 2003;96(4):367–84.

26] van Rijn PA, van Gennip HG, Moormann RJ. An experimental marker vaccine andaccompanying serological diagnostic test both based on envelope glycoproteinE2 of classical swine fever virus (CSFV). Vaccine 1999;17(5):433–40.

27] Kalaycioglu AT. Bovine viral diarrhoea virus (BVDV) diversity and vaccination:a review. Vet Q 2007;29(2):60–7.

28] Bolin SR, Ridpath JF. Glycoprotein E2 of bovine viral diarrhea virus expressedin insect cells provides calves limited protection from systemic infection anddisease. Arch Virol 1996;141(8):1463–77.

29] Peeters BP, de Leeuw OS, Verstegen I, Koch G, Gielkens AL. Generationof a recombinant chimeric Newcastle disease virus vaccine that allows

[

3600 M. de Lima et al. / Vaccine 2

serological differentiation between vaccinated and infected animals. Vaccine2001;19(13–14):1616–27.

30] Mebatsion T, Koolen MJ, de Vaan LT, de Haas N, Braber M, Romer-OberdorferA, et al. Newcastle disease virus (NDV) marker vaccine: an immunodominantepitope on the nucleoprotein gene of NDV can be deleted or replaced by aforeign epitope. J Virol 2002;76(20):10138–46.

[

6 (2008) 3594–3600

31] Bird BH, Albarino CG, Hartman AL, Erickson BR, Ksiazek TG, Nichol ST. Rift Val-ley fever virus lacking the NSs and NSm genes is highly attenuated, confersprotective immunity from virulent virus challenge, and allows for differen-tial identification of infected and vaccinated animals. J Virol 2008;82(6):2681–91.