Ra

Sa

b

ARRAA

KMVCRpS

1

pohdDnc

BcafCgnr

sH

h0

COREView metad

onnector

Journal of Virological Methods 233 (2016) 23–36

Contents lists available at ScienceDirect

Journal of Virological Methods

j o ur na l ho me pa ge: www.elsev ier .com/ locate / jv i romet

eal-time PCR for differential quantification of CVI988 vaccine virusnd virulent strains of Marek’s disease virus

usan J. Baigenta,∗, Venugopal K. Naira, Hervé Le Galludecb

Avian Oncogenic Virus Group, The Pirbright Institute, Woking, GU24 0NF, United KingdomZoetis International Services, 23-25 Avenue du Docteur Lannelongue, 75668 Paris Cedex 14, France

rticle history:eceived 3 December 2015eceived in revised form 1 March 2016ccepted 7 March 2016vailable online 10 March 2016

eywords:arek’s disease virus

accinationVI988/Rispens

a b s t r a c t

CVI988/Rispens vaccine, the ‘gold standard’ vaccine against Marek’s disease in poultry, is not easily distin-guishable from virulent strains of Marek’s disease herpesvirus (MDV). Accurate differential measurementof CVI988 and virulent MDV is commercially important to confirm successful vaccination, to diagnoseMarek’s disease, and to investigate causes of vaccine failure. A real-time quantitative PCR assay to dis-tinguish CVI988 and virulent MDV based on a consistent single nucleotide polymorphism in the pp38gene, was developed, optimised and validated using common primers to amplify both viruses, but dif-ferential detection of PCR products using two short probes specific for either CVI988 or virulent MDV.Both probes showed perfect specificity for three commercial preparations of CVI988 and 12 virulent MDVstrains. Validation against BAC-sequence-specific and US2-sequence-specific q-PCR, on spleen samples

brought to you by ata, citation and similar papers at core.ac.uk

provided by Elsevier - Publisher C

eal-time PCRp38ingle nucleotide polymorphism

from experimental chickens co-infected with BAC-cloned pCVI988 and wild-type virulent MDV, demon-strated that CVI988 and virulent MDV could be quantified very accurately. The assay was then used tofollow kinetics of replication of commercial CVI988 and virulent MDV in feather tips and blood of vac-cinated and challenged experimental chickens. The assay is a great improvement in enabling accuratedifferential quantification of CVI988 and virulent MDV over a biologically relevant range of virus levels.

© 2016 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license

. Introduction

Marek’s disease (MD) is a highly infectious lymphoid neo-lasm of chickens caused by oncogenic serotype-1 strains (MDV-1)f Marek’s disease virus (family Herpesviridae, subfamily Alpha-erpesvirinae, genus Mardivirus). Susceptible breeds of chickenevelop visceral and neural lymphomatous lesions (Baigent and

avison, 2004; Calnek, 2001), resulting in death or carcass condem-ation. This economically important disease has been successfullyontrolled since the 1970s by vaccination using cell-associated

Abbreviations: 40-Ct, cycle threshold value in real-time PCR subtracted from 40;AC, bacterial-artificial-chromosome; dpc, days post challenge; dpv, days post vac-ination; FAM-BHQ1, probe with carboxyfluorescein (FAM) reporter fluorochromend Black Hole Quencher; LOD, limit of detection by q-PCR; ovo, chicken ovotrans-errin gene; PBL, peripheral blood lymphocytes; pp38-CVI, PCR probe specific forVI988 pp38 gene sequence; pp38-Vir, PCR probe specific for virulent MDV pp38ene sequence; q-PCR, real-time quantitative polymerase chain reaction; SNP, singleucleotide polymorphism; YY-TAMRA, probe with Yakima Yellow reporter fluo-ochrome and tetramethylrhodamine quencher.∗ Corresponding author.

E-mail addresses:ue.baigent@outlook.com (S.J. Baigent), venugopal.nair@pirbright.ac.uk (V.K. Nair),erve.Legalludec@zoetis.com (H. Le Galludec).

ttp://dx.doi.org/10.1016/j.jviromet.2016.03.002166-0934/© 2016 The Authors. Published by Elsevier B.V. This is an open access article u

(http://creativecommons.org/licenses/by/4.0/).

live avirulent and nononcogenic vaccine strains of all threeMardivirus serotypes. Serotype 2 (MDV-2) or serotype 3 (her-pesvirus of turkeys, HVT) vaccine viruses are naturally avirulentin chickens (Okazaki et al., 1970; Schat and Calnek, 1978), whileCVI988/Rispens vaccine (de Boer et al., 1986; Rispens et al., 1972) isa naturally-attenuated MDV-1 strain. MD vaccine viruses establishpersistent infection and lifelong immunity, effectively protectingagainst tumours and mortality. However, vaccination does not pre-vent superinfection, replication and shedding of virulent challengeviruses (Davison and Nair, 2005; Gimeno, 2008), so chickens canpotentially be infected simultaneously with both vaccine and vir-ulent MDV strains.

Accurate differential measurement of vaccine and virulent virusin the same individual chicken is both commercially importantand experimentally useful. In the field, it could confirm success-ful vaccination and assist in identifying causes of vaccine failure,such as administration of a sub-optimal vaccine dose (Landmanand Verschuren, 2003), interference with vaccine virus replicationby maternal antibodies (King et al., 1981; de Boer et al., 1986), and

infection with virulent MDV field strains (Witter et al., 2005). Inthe laboratory, it could be used to study mechanisms of vaccinalprotection and differences in efficacy between vaccines.

nder the CC BY license (http://creativecommons.org/licenses/by/4.0/).

2 rologic

tqsbcRt1i(rdddtf

vtpaeceiHets

fotmovssv1brqMoePmfhu2ent

sMdiCtaRP

4 S.J. Baigent et al. / Journal of Vi

Therefore, it is important to have a reliable test for differentia-ion of MDV-infected and vaccinated animals (DIVA test). Real-timeuantitative polymerase chain reaction (q-PCR) is a sensitive andpecific molecular diagnostic method. Serotype-specific q-PCR cane used to distinguish and differentially quantify MD vaccine andhallenge virus, where vaccine is of serotype 2 (Islam et al., 2004;enz et al., 2006) or serotype 3 (Islam et al., 2004, 2006a,b), andhus of a different serotype to challenge virus (always serotype). CVI988 is the ‘gold standard’ vaccine against MD because it

s antigenically and genetically 98% identical to virulent strainsSpatz et al., 2007a), and therefore elicits a very effective immuneesponse. However, the downside of this is that CVI988 is not easilyistinguishable from virulent MDV-1, restricting the application ofifferential molecular diagnostics (Baigent et al., 2006b). To date,evelopment of a q-PCR assay which is both accurately quantita-ive and able to universally distinguish all virulent MDV strainsrom CVI988, has not been possible.

Using bacterial-artificial-chromosome (BAC) cloned CVI988irus (pCVI988), which carries the BAC vector sequences in place ofhe MDV-1 US2 gene, q-PCR assays for differential quantification ofCVI988 from virulent MDV (by targeting the BAC vector sequencend the US2 gene respectively) were developed previously (Baigentt al., 2011). Although these assays allowed investigation of repli-ation of both pCVI988 vaccine virus and virulent MDV-1 inxperimental infections (Baigent et al., 2011), they cannot be usedn the field as commercial CVI988 does not contain a BAC sequence.owever, the BAC-US2 q-PCR system, and the DNA samples gen-rated in that study, represents an ideal system against whicho validate any prospective CVI988-specific and virulent MDV-1-pecific q-PCR assays.

Spatz and Silva (2007) identified several types of sequence dif-erence in the repeat long (RL) genomic regions of 13 MDV-1 strainsf varying virulence: polynucleotide sequence expansions, shortandem repeat polymorphisms, sequence deletions, frameshift

utations and single nucleotide polymorphisms (SNPs). The effortsf other researchers to develop DIVA PCR assays for CVI988 andirulent MDV have focused on the polynucleotide sequence expan-ion of the 132-bp repeat region, the 177-bp insertion of duplicatedequence in the CVI988 meq gene, and SNPs in the pp38 gene. Con-entional PCR to amplify the 132-bp repeat region (Becker et al.,992; Silva, 1992) could distinguish CVI988 from virulent MDVsased on a difference in the number of 132-bp repeats, but hadelatively low sensitivity (Davidson et al., 2002). A highly sensitiveuantitative meq gene q-PCR, that detects both CVI988 and virulentDV (Baigent et al., 2005a), can be used to quantify either CVI988

r virulent MDV if only one of those viruses is present (Baigentt al., 2005a,b, 2006a,b), or can be paired with the 132-bp repeatCR to assist with distinction between CVI988 and virulent MDV inixed infections. However, 132-bp repeat PCR is not very sensitive

or CVI988 (since the multiple bands of the CVI988 PCR productsave much lower intensity than the single/double bands of the vir-lent MDV PCR products), not absolutely conclusive (Baigent et al.,005b; van Iddekinge et al., 1999; Niikura et al., 2006; Petherbridget al., 2003; Silva and Gimeno, 2007), laborious, and the repetitiveature of the 132-bp repeat region precludes its use as a quantita-ive assay.

The meq gene is polymorphic, with variation in the number ofequence repeats in the proline-rich region (Chang et al., 2002).urata et al. (2007) developed a non-quantitative nested PCR to

istinguish between standard meq and large (L-)meq (with a 177-bpnsertion). However, while virulent MDV strains have standard meq,VI988 may have either L-meq or standard meq (Lee et al., 2000), so

he 177-bp insertion is not a reliable differential marker for CVI988nd virulent MDV. Based on a stable polymorphism in the meq gene,enz et al. (2013) developed two sensitive and specific real-timeCRs to differentiate and accurately quantify 20 Australian viru-

al Methods 233 (2016) 23–36

lent MDV isolates and three commercial CVI988 vaccines used inAustralia. However, virulent isolates from USA, Asia and Europe donot differ from CVI988 at this polymorphism so, while this assaycould be very successfully used in Australia, it could not be usedglobally (Renz et al., 2013).

The MDV-1 pp38 gene has a SNP at base #320, which is con-sistent between CVI988 (which has ‘G’ at this position) and allsequenced virulent strains (which have ‘A’) (Cui et al., 1991; Endohet al., 1994; Spatz et al., 2007a,b). The current study therefore inves-tigated the possibility of using SNP #320 as a reliable biomarkerfor differentiating and quantifying CVI988 and virulent strains byreal-time PCR (pp38 SNP q-PCR assay). Subsequent to developmentof this assay, Gimeno et al. (2014) published a real-time PCR assayalso based on SNP #320. However, their assay differs from the assayreported in the current paper in terms of the molecular methodsused to amplify and detect the polymorphism, their ability to fullyvalidate the assay, and its quantitative accuracy.

The aims of the current work were (1) to design and optimise aq-PCR assay based on SNP #320 in the pp38 gene to specificallydistinguish CVI988 from virulent MDV-1 strains; (2) to investi-gate the feasibility of making this assay accurately quantitativeover a biologically relevant range of virus levels; (3) to validatethe assay against the BAC/US2 q-PCR assay; and (4) to use the assayto investigate replication of CVI988 and virulent MDV in tissues ofvaccinated, challenged experimental chickens.

2. Materials and methods

2.1. Design of q-PCR primers and probes

Primers and probes were designed by AlleLogic Biosciences Corp(Hayward, CA). Primer and TaqMan® probe sequences, and theirlocation within the pp38 gene, are shown in Fig. 1. The forward andreverse primers (pp38-FP and pp38-RP) are common to both theCVI988 and virulent MDV-1 pp38 sequences, and the 99-bp ampli-con covers the polymorphic region within which the probes weredesigned. One 15-mer probe (pp38-Generic) was designed to targeta region common to all MDV-1 strains. Two 15-mer probes weredesigned to incorporate the consistent G/A SNP #320, one specificfor the CVI988 sequence (pp38-CVI), and one specific for the vir-ulent MDV-1 sequence (pp38-Vir(1)). Subsequently, an additionalprobe (pp38-Vir(3)) was designed for detection of virulent MDV-1, to have a melting temperature similar to that of the pp38-CVIprobe and the pp38-Generic probe. Probe pp38-Vir(3) also incor-porates the ‘non-consistent’ SNP #326, so it is possible that affinityfor some field strains may be reduced. However, in this study, thespecificity of pp38-Vir(1) and pp38-Vir(3) was identical, but pp38-Vir(3) was chosen as the probe for routine detection of virulentMDV-1 because it was more sensitive. Each probe was labelledwith 5′FAM reporter, and 3′BHQ1 quencher. All probes were usedin combination with the same primer pair. All primers and probeswere manufactured by Sigma Genosys (UK).

2.2. DNA samples from virus stocks and chicken tissues

HVT strain FC126, MDV-1 strain HPRS-B14, and USA MDV-1strains 675A, 584A, 648A, 660A, 595, Md5, 549, 571, JM102/W(Witter, 1997; Witter et al., 2005) were obtained from Dr. A.M.Fadly (Avian Disease and Oncology Laboratory, USA) all as 7thduck embryo fibroblast passage stocks. MDV-1 strain RB-1B wasprepared as previously described (Baigent et al., 2007). Euro-

pean MDV-1 isolate C12/130 (Barrow and Venugopal, 1999) wasprepared as previously described (Smith et al., 2011). Three com-mercial CVI988 vaccine stocks were used: Pfizer Poulvac Marek CVI,Merial BioMarekR, and Intervet Nobilis Rismavac. MDV-2 strain

S.J. Baigent et al. / Journal of Virological Methods 233 (2016) 23–36 25

ATGGAATTCGAAGCAGAACACGAAGGGCTGACGGCGTCTTGGGTCGCCCCCGCTCCCCAGGGTGGAAAAGGGGCGGAGGGCCGCGCAGGGGTCGCCGACGAGGCAGGGCATGGGAAAACAGAAGCGGAATGCGCCGAGGACGGCGAGAAATGCGGGGACGCCGAGATGAGCGCTTTGGATCGGGTCCAGAGGGACCGGTGGAGATTCAGTTCTCCGCCCCCTCACTCTGGAGTCACGGGGAAGGGGGCTATTCCAATAAAGGGTGATGGGAAGGCGATAGAATGCCAGGAGCTAACCGGAGAGGGAGAGTGGCTGTCACXGTGGGYGGAGCTACCGCCTGAGCCCCGGAGGTCAGGGAATGAACATCTTGACGAAAGTCGGTATGCGAAACAAACCGAAAGGGGTAGCTCTACGGGGAAAGAAGAGGGAGATGGTATGAAGCAGATGGGGGAGCTTGCCCAGCAGTGCGAAGGAGGAACATATGCGGACTTGCTTGTCGAAGCAGAGCAAGCTGTTGTACATTCCGTTCGCGCATTAATGCTGGCCGAAAGACAAAACCCAAATATATTGGGGGAGCATTTGAATAAAAAACGGGTTCTTGTACAACGACCCCGTACTATTCTATCCGTGGAGTCAGAGAATGCAACAATGCGTTCTTATATGCTGGTTACATTGATCTGTTCTGCAAAATCATTATTACTAGGATCGTGCATGTCATTTTTCGCTGGTATGTTAGTCGGTAGAACGGCAGACGTAAAAACACCATTATGGGATACTGTATGTTTGTTAATGGCTTTCTGTGCAGGCATTGTCGTTGGGGGAGTGGATTCTGGGGAGGTGGAATCTGGAGAAACAAAATCTGAATCAAATTAA

Fig. 1. Location of primers and probes in MDV-1 pp38 gene sequences.The pp38 gene sequence shows two single nucleotide differences (G/A SNPs) between CVI988 and Md5 (a representative virulent strain). The bold SNP X (#320) rep-resents a consistent difference between CVI988 and virulent strains. The bold italicised SNP Y (#326) is not consistent in all virulent strains. Sequences of the primers(underlined text) were: forward primer 5′-GAGCTAACCGGAGAGGGAGA-3′ , and reverse primer 5′-CGCATACCGACTTTCGTCAA-3′ , and the amplicon size was 99 bp. Two 15-m conta( ed bo( hich t

Svppbaascs

2

drceoLbo5wmwRatvqapCCCap2(pbsce

er probes (light grey shading), designed to the minus strand, covered the region5′-CCCACTGTGACAGCC-3′ , Tm = 54.1). A third (17-mer) probe, pp38-Vir(3) includpp38-Generic, 5′-GCTACCGCCTGAGCC-3′ , designed to the plus strand, Tm = 57.4) w

B-1 was a commercial vaccine stock (SB-vacc, Intervet). pCVI988irus was prepared from the BAC clone of Poulvac CVI988 DNA, asreviously described (Petherbridge et al., 2003). All viruses wererepared as cell-associated stocks grown in chicken embryo fibro-last cells (CEF). For each virus stock, DNA was prepared frompproximately 5 × 106 cells using the DNeasy-96 kit (Qiagen),ccording to the manufacturer’s instructions. Chicken tissue DNAamples were available from a previous study from experimentalhickens vaccinated with pCVI988 and/or challenged with virulenttrain RB-1B (Baigent et al., 2011).

.3. Real-time quantitative PCR

Reactions were set up on ice, in 96-well plates in a PCR-edicated cabinet, using PCR-dedicated pipettes and aerosolesistant autoclaved tips. A master-mix was prepared diluting theomponents in water, and 21 �l aliquotted per reaction, so thatach reaction contained: primers pp38-FP and pp38-RP (0.4 �M),ne of the above pp38 probes (0.2 �M) and ABsolute Blue® q-PCRow Rox master-mix (Thermofisher UK). Duplex reactions to detectoth pp38 and the chicken ovotransferrin (ovo) gene also containedvo forward and reverse primers (0.4 �M), and ovo probe (0.2 �M,′Yakima Yellow-3′TAMRA, Eurogentec), the sequences of whichere reported previously (Baigent et al., 2005a). All reactions wereade up to 25 �l by the addition of 4 �l water (‘no template control’ells), test DNA, standard DNA or control DNA. Every run includedB-1B DNA control (positive control for virulent MDV detectionnd negative control for CVI988 detection) and CVI988 DNA con-rol (positive control for CVI988 detection and negative control forirulent MDV detection) to ensure correct setting of the threshold.-PCRs targeting the BAC and US2 gene sequences were performeds previously described (Baigent et al., 2011). Standard curves wererepared using 10-fold serial dilutions of DNA from RB-1B-infectedEF (for pp38-Vir reaction and for US2 reaction), pCVI988-infectedEF (for pp38-CVI reaction and for BAC reaction) and non-infectedEF (for ovo reaction) all of which had been accurately calibratedgainst plasmid constructs of known target gene copy number:RB-1B (Petherbridge et al., 2004), pCVI988 (Petherbridge et al.,003), and pGEMT-ovo (Baigent et al., 2005a). An ABI7500® systemApplied Biosystems) was used to amplify and detect the reactionroducts, under the following conditions: 95 ◦C for 15 min, followed

y 40 cycles of 95 ◦C (15 s) and 60 ◦C (1 min). For chicken tissue DNAamples, MDV genomes were normalised to 2 × 104 copies of thehicken ovo gene, so the standard curves enabled quantification ofach virus as genomes per 104 cells (Baigent et al., 2005a).

ining SNP #320: pp38-CVI (5′-CCCACCGTGACAGCC-3′ , Tm = 57.7) and pp38-Vir(1)th SNPs #320 and #326 (5′-CTCCCACTGTGACAGCC-3′ , Tm = 57.1). A fourth probeargets a region common to pp38 of all MDV-1 strains, is shaded dark grey.

2.4. Quantitative accuracy of detection in virus mixtures.

Testing of sensitivity and quantitative accuracy of the pp38q-PCRs, in mixed samples over a wide range of CVI988:virulentMDV ratios, was performed and validated against BAC-specificand US2-specific q-PCRs, as described below. DNA prepared fromCEF infected with pCVI988 virus, and DNA prepared from CEFinfected with RB-1B was used to prepare dilution series. The start-ing concentration of each DNA sample was approximately 5 × 105

genomes per 4 �l. DNA of pCVI988 was diluted in DNA of RB-1B orin DNA from non-infected CEF, in two parallel dilution series; andDNA of RB-1B was diluted in DNA of pCVI988 or in DNA from non-infected CEF, in two parallel dilution series. Dilutions were 1:2, 1:5,1:10, 1:30, 1:102, 1:103 and 1:104, each in duplicate. The pp38 SNPq-PCR was validated by two means: (1) Comparing measurementsfrom samples diluted in DNA from the other virus with measure-ments from samples diluted in chicken DNA from non-infectedCEF; (2) Comparing measurements made by the pp38-CVI q-PCRand the pp38-Vir q-PCR, with those made using BAC-specific andUS2-specific q-PCRs respectively.

2.5. Validation of pp38 q-PCR assays on chicken tissue samples

Spleen DNA samples from experimental chickens which hadbeen vaccinated with BAC-cloned CVI988 virus (pCVI988) and/orchallenged with RB-1B, were available from a previous study(Baigent et al., 2011). In that study, pCVI988 vaccine virus and RB-1B challenge virus were distinguished and specifically quantifiedusing TaqMan® q-PCR to measure a BAC-specific target sequence,and the US2 gene target sequence, respectively. In the current work,the spleen DNA samples from this study were subjected to pp38SNP/ovo duplex q-PCR assays, using either pp38-CVI, pp38-Vir(3),or pp38-Generic as the pp38 detection probe. The appropriate stan-dard curves were used to calculate genome copy no. per 104 cellsfor CVI988, RB-1B or total MDV-1, respectively.

2.6. Time-course of replication of commercial CVI988 vaccine andRB-1B challenge virus in individual chickens

Thirty-two one-day-old specified pathogen-free, maternal-antibody-free Rhode Island Red chickens were randomly allocated

to three experimental groups. All procedures were performedaccording to the guidelines of the UK Home Office. The studywas performed under a Project Licence issued by the UK HomeOffice and approved by the Ethical Review Committee at The Pir-

2 rologic

bbp

ooas(BbeTicwanelpwfPtjodfaccw

2

dRiuvgqccvgmstecntesc

3

3p

V1

6 S.J. Baigent et al. / Journal of Vi

right Institute. Vaccination and challenge were performed onlyy trained persons in possession of a Personal Home Office Licenceermitting them to perform these procedures.

Chickens were vaccinated by subcutaneous administration ofne commercial dose of Poulvac CVI988 (Pfizer Animal Health) atne day of age. Chickens were challenged at 8 days of age, by intra-bdominal administration of 100 �l dose containing 1000 pfu MDVtrain RB-1B (Schat et al., 1982) prepared as previously describedBaigent et al., 2007). Group A (n = 8) received vaccine only, Group

(n = 14) received challenge virus only, and Group C (n = 10) wasoth vaccinated and challenged. Group A initially had 10 chick-ns, but two were ‘non-starters’ and were culled prior to challenge.he greater number of chickens in Group B was intentional, sincet was expected that some would die in the first two weeks posthallenge. The three groups were housed in separate animal roomshich were high efficiency particulate air-filtered and maintained

t negative pressure. Every chicken was sampled by trained person-el at ten time-points between -4 and 38 days post challenge (dpc),quivalent to 3 and 45 days post vaccination (dpv). Peripheral bloodymphocyte (PBL) and feather tip samples were collected and pre-ared as previously described (Baigent et al., 2005a,b). Feather tipsere collected from 10 dpv (3 dpc), the first time-point at which

eathers were present in sufficient size. DNA was prepared fromBL and feather tips using the DNeasy-96 kit (Qiagen), accordingo the manufacturer’s instructions, and each DNA sample was sub-ected to pp38 SNP/ovo duplex q-PCR assays, using either pp38-CVIr pp38-Vir(3) as the pp38 detection probe. The appropriate stan-ard curves were used to calculate genome copy no. per 104 cellsor CVI988 or RB-1B. Chickens were monitored daily for morbiditynd mortality, and any chicken reaching the humane end-point wasulled using a Schedule 1 method by trained persons. All survivinghickens were culled at 46 dpv (39 dpc). Post mortem examinationas performed on every chicken.

.7. Statistical analyses

Statistical analyses were performed using Minitab v15. To vali-ate the efficacy of the pp38 SNP q-PCRs in quantifying CVI988 andB-1B in virus mixtures, mean 40-Ct values for virus DNA diluted

n non-infected CEF DNA were compared with mean 40-Ct val-es for virus DNA diluted in DNA of CEF infected with the otherirus, using analysis of variance. To investigate correlation betweenenome copies calculated using the two q-PCR systems (BAC-US2-PCR vs pp38 SNP q-PCRs), samples were compared in pairwiseombinations using regression analysis. To examine the effect ofo-infection on replication of CVI988 vaccine and RB-1B challengeiruses in PBL and feather tips, arithmetic mean values for virusenome copy number and 95% confidence intervals were deter-ined using the log10 transformed copy number for each individual

ample, then back-transformed to obtain actual values for presen-ation of results. Analysis of variance was performed to examine theffect of challenge with RB-1B on the level of CVI988 per 104 cells inhickens from the two vaccinated groups, and the effect of vacci-ation on the level of RB-1B virus per 104 cells in chickens fromhe two challenged groups. This analysis was performed withinach sampling time-point (not as a repeated measures analysis),o sampling time was excluded as a factor in the model. Pairwiseomparisons were made using Tukey simultaneous tests.

. Results

.1. Specificity of CVI988-specific and virulent MDV-1-specificp38 probes in singleplex reactions

Fig. 2 shows the change in fluorescence when pp38-CVI, pp38-ir(3) and pp38-Generic probes were used in 40-cycle q-PCR using00 ng DNA from either CVI988-infected CEF or RB-1B-infected CEF

al Methods 233 (2016) 23–36

as a template. The pp38-CVI probe gave a large increase in fluores-cence with CVI988 DNA template, while increase in fluorescencewith RB-1B DNA template was minimal (fluorescence did not crossthe threshold line at the default setting of 0.2, thus the sampleis scored negative) (Fig. 2a). The pp38-Vir(3) probe gave a largeincrease in fluorescence with RB-1B DNA, and some increase withCVI988 DNA (Fig. 2b). Raising the threshold to 0.6 ensured thatCVI988 was always scored negative in this assay and preventedfalse positive results. The pp38-Generic probe gave a large increasein fluorescence with both RB-1B and CVI988, with a threshold set-ting of 0.2 (Fig. 2c). None of the three probes gave any increase influorescence in no-template-control wells.

The specificity of the two probes was then tested against tar-get DNA samples from a range of Mardiviruses: HVT (serotype 3),SB-1 (serotype 2), CVI988 vaccine from three different commercialsources, and 12 virulent MDV-1 strains covering the spectrum ofvirulence from virulent (v), very virulent (vv) to very virulent plus(vv+) (Witter et al., 2005). Fig. 2d shows that, using a detectionthreshold of 0.2 for pp38-CVI, this probe detected all three com-mercial CVI988 preparations, but did not detect any of the virulentMDV-1 strains, or HVT, or SB-1. Using a detection threshold of 0.6for pp38-Vir(3), this probe detected all virulent MDV-1 strains, butdid not detect CVI988, HVT or SB-1. Using a detection threshold of0.2, pp38-Generic probe detected all CVI988 and all virulent strainsof MDV-1, but did not detect HVT or SB-1.

3.2. Assay sensitivity, efficiency, linearity and reproducibility

The probes were tested in q-PCR assays using 10-fold serialdilutions of DNA prepared from CEF infected with either RB-1Bor CVI988. Mean amplification efficiencies, R2 values and limits ofdetection (LOD) are summarised in Table 1. LOD is the value forx (gene copy no.) when y (40-Ct value) = 0. Fig. 3 shows the repro-ducibility of standard curves, across three independent q-PCR runs,using pp38-CVI probe with serial dilutions of CVI988 DNA (Fig. 3a)and pp38-Vir(3) probe with serial dilutions of RB-1B DNA (Fig. 3b).The assays show good reproducibility, sensitivity, linearity, effi-ciency, and a wide dynamic range.

3.3. Duplex q-PCR for absolute quantification of CVI988 andvirulent MDV-1

In order to use the pp38 SNP q-PCR assays for absolute quantifi-cation of vaccine and challenge virus in tissue samples (expressedas virus genomes per 104 cells), it was necessary to use duplex reac-tions to simultaneously measure either the CVI988 pp38 gene andthe ovo gene, or the virulent MDV-1 pp38 gene and the ovo gene.The ovo probe was labelled with YY-TAMRA so as to be detectedin a different channel to the pp38 probe fluorescence. The ovo andpp38 reactions were shown to proceed with the same efficiencywhen used in duplex reactions as in singleplex reactions (data notshown). The mean amplification efficiency, R2 value and LOD forthe ovo reaction are shown in Table 1.

3.4. Quantitative accuracy of detection in virus mixtures

Since the pp38 SNP q-PCR assay uses a universal primer pair,complementary to a conserved region of the pp38 gene, bothCVI988 and virulent strains of MDV will be amplified from a mixedpopulation, and thus there is no reaction specificity at this level(specificity is determined solely by probe binding). Therefore, whenboth CVI988 and virulent MDV target sequences are present, they

will compete for the same primers. If one virus is present at a highlevel, and the other at only a low level, it is possible that competi-tion for primers would prevent amplification (and thus detection)of the less abundant virus. The ability of the reactions to accurately

S.J. Baigent et al. / Journal of Virological Methods 233 (2016) 23–36 27

Fig. 2. Specificity of probes for CVI988 and virulent MDV-1.Change in fluorescence (delta Rn) when pp38-CVI, pp38-Vir(3) and pp38-Generic probes were used in 40-cycle q-PCR using 100 ng DNA from either CVI988-infected CEF orRB-1B-infected CEF as a template. The thick horizontal line with an arrowhead at each end is the threshold (default value 0.2). (a) pp38-CVI probe detects CVI988 DNA butnot RB-1B DNA; NTC is no-template control. (b) pp38-Vir(3) probe detects RB-1B DNA, while the signal from the CVI988 PCR product is very low. By raising the threshold lineto 0.6, CVI988 is scored negative in this reaction. (c) pp38-Generic probe detects both CVI988 DNA and RB-1B DNA. (d) The specificities of pp38-CVI, pp38Vir-A(3) and pp38-Generic probes were tested against a range of Mardivirus strains in a 40-cycle q-PCR. The 40-Ct value (the cycle threshold value, Ct, subtracted from 40) is proportional to theamount of target DNA present. A 40-Ct value of 0 indicates that fluorescence did not cross the threshold, i.e. the probe did not bind to the PCR product. ‘att MDV1′ = attenuatedMDV-1, vMDV1 = virulent MDV-1, vvMDV1 = very virulent MDV-1, vv+MDV1 = very virulent plus MDV-1.

28 S.J. Baigent et al. / Journal of Virological Methods 233 (2016) 23–36

Table 1q-PCR assay efficiency and reproducibility.

Primers Probe DNA template Slopea R2 valueb LODc

pp38-FP +pp38-RP pp38-CVI CVI988-infected CEF DNA 3.52 0.995 5.7pp38-FP +pp38-RP pp38-Vir(1) RB-1B-infected CEF DNA 3.53 0.995 29.6pp38-FP +pp38-RP pp38-Vir(3) RB-1B-infected CEF DNA 3.48 0.994 9.6pp38-FP +pp38-RP pp38-Generic CVI988-infected CEF DNA 3.27 0.992 8.9pp38-FP +pp38-RP pp38-Generic RB-1B-infected CEF DNA 3.59 0.995 7.4Ovo-FP + Ovo-RP Ovo Non-infected CEF 3.58 0.982 1.1

a measure of amplification efficiency (good efficiency is considered to be 90% [slope = 3.59] − 105% [slope = 3.21]).b R2 coefficient of determination, a measure of linearity.c Limit of detection (copies of target gene).

R² = 0.99 37

0

5

10

15

20

25

1 10 10 0 10 00 10 000 10000 0 10 0000 0 10 0000 00

40-C

t val

ue ( p

p38

gene

-vi

rule

nt)

Copy no. pp38 gene - RB-1B target DNA

b)

R² = 0.9945

0

5

10

15

20

25

1 10 100 10 00 10 000 10 000 0 10000 00 10 0000 00

40-C

t val

ue (p

p38

gene

-CV

I988

)

Copy no. pp 38 gene - CVI988 target DNA

a)

Fig. 3. Assay sensitivity, efficiency, linearity and reproducibility.The probes were tested in q-PCR assays using serial dilutions of DNA extracted from CVI988-infected CEF or RB-1B-infected CEF. Mean 40-Ct values from triplicate assays,w p38-Cs cate as

qePa

wnsR

ith 95% confidence limits for the regression and the R2 value, are shown for (a) perial dilutions of RB-1B CEF DNA. The underlying points for 40-Ct values from tripliquares, black diamonds and grey triangles).

uantify both CVI988 and virulent MDV-1 in mixed samples wasxamined and validated against BAC-specific and US2-specific q-CRs (which will distinguish cloned pCVI988 and RB-1B withoutny competition for reagents).

The BAC-specific q-PCR detected the same level of pCVI988 DNA,hether the pCVI988 DNA was diluted into RB-1B CEF DNA or into

on-infected CEF DNA at each dilution (data not shown). The US2-pecific q-PCR detected the same level of RB-1B DNA, whether theB-1B DNA was diluted into pCVI988CEF DNA or into non-infected

VI probe with serial dilutions of CVI988CEF DNA, and (b) pp38-Vir(3) probe withssays are shown using three different symbols to distinguish the three assays (open

CEF DNA at each dilution (data not shown). These assays confirmedthat the dilution series were accurately prepared and directly com-parable, and that there was no competition for reagents in theseq-PCR assays. For the pp38-CVI q-PCR, 40-Ct values for pCVI988DNA diluted in RB-1B CEF DNA were only slightly (≤1 Ct unit) lowerthan those for pCVI988 DNA diluted in non-infected CEF DNA, at

dilutions down to 1:5 (Fig. 4a). At 1:10 and 1:30 the differencewas greater (2Ct units) and statistically significant and, at dilutionsof 1:102, 1:103, the difference was 4–7 Ct units (p < 0.001). Thus,

S.J. Baigent et al. / Journal of Virological Methods 233 (2016) 23–36 29

0

2

4

6

8

10

12

14

16

18

20

0.00 01 0.00 1 0.01 0.1 1

40-C

t val

ue fo

r pp3

8 ge

ne o

f RB-

1B

Di l u�on of RB-1B DNA in DNA from pCVI988 infected CEF or in DNA from non-infected CEF

b)

0

2

4

6

8

10

12

14

16

18

20

0.0001 0.001 0.01 0.1 1

40-C

t val

ue fo

r pp3

8 ge

ne o

f CVI

988

Di lu�on of pCVI988 DNA in DNA from RB-1B infected CEF or in DNA from non-infected CEF

a)

Fig. 4. Quantitative accuracy of detection in virus mixtures using pp38 SNP q-PCR assays.DNA was prepared from CEF cells infected with pCVI988, and from CEF cells infected with RB-1B, and each was diluted to approximately 5 × 105 virus genomes per 4 �l. (a)DNA of pCVI988-infected CEF was diluted in DNA of RB-1B-infected CEF (open triangles) or in DNA from non-infected CEF (black squares). (b) DNA of RB-1B-infected CEFw nfectep r dupc

abpliRofDa

3s

aw(cpTP

as diluted in DNA of pCVI988 infected CEF (open triangles) or in DNA from non-ip38-Vir(3) probe. The figures show mean 40-Ct values with standard error bars folarity).

t the greater dilutions, pCVI988 DNA could clearly be detectedut not accurately quantified. At 1:104 dilution in RB-1B DNA, theCVI988 DNA was barely detectable, despite still being present at

evels detectable by BAC q-PCR, showing that reagent competitionn the pp38 SNP q-PCRs reduces detection of pCVI988 DNA whenB-1B DNA is in >5-fold excess. For the pp38-Vir(3) q-PCR, the effectf reagent competition on measured level of RB-1B was significantrom 1:5 dilution in pCVI988 DNA (p < 0.01) and, although RB-1BNA was still clearly detected at further dilutions, it would not beccurately quantifiable (Fig. 4b).

.5. Validation of pp38 SNP q-PCR assays on chicken tissueamples

Spleen DNA samples from experimental chickens were avail-ble from a previous study (Baigent et al., 2011) where the chickensere vaccinated at one day of age with BAC-cloned CVI988 virus

pCVI988) and/or challenged at 15 days of age with RB-1B, and four

hickens per group sacrificed to collect spleen at each of eight time-oints between 4 and 34 dpv (equivalent to −10 dpc to 20 dpc).hese DNA samples were now subjected to pp38 SNP/ovo duplex q-CR assays to determine level of vaccine virus and challenge virus

d CEF (black squares). Samples were run in q-PCR using either pp38-CVI probe orlicate dilution series and duplicate reactions (only upper error bars are shown, for

as genome copies per 104 cells. The calculated level of pCVI988was compared with that previously determined using the BAC-specific q-PCR assay (Baigent et al., 2011), and the level of RB-1Bwas compared with that previously determined using the US2-specific q-PCR assay (Baigent et al., 2011). The specificity of eachof the reactions was as expected for the four groups of chickens:pCVI988 was detected only in the two vaccinated groups and RB-1Bonly in the two challenged groups. For each individual spleen DNAsample, total MDV-1 genome copy no., determined using the pp38-Generic probe, was compared with that obtained from the additiveresults from US2 and BAC reactions, and from the additive resultsfrom pp38-CVI and pp38-Vir(3) reactions. Fig. 5 shows correlationbetween levels of pCVI988 vaccine virus, RB-1B challenge virus,and total MDV-1 measured by the two q-PCR systems. There was ahighly significant relationship between the level of pCVI988 mea-sured by the two q-PCR systems (p < 0.0001, Fig. 5a), and betweenthe level of RB-1B measured by the two q-PCR systems (p < 0.0001,Fig. 5b). There was a highly significant relationship between the

level of total MDV measured by pp38-Generic, BAC + US2 and pp38-CVI + pp38-Vir(3) (p < 0.0001 for each of the comparisons; Fig. 5cand d). These data confirm that pp38-CVI and pp38-Vir(3) reactionsshow excellent specificity, and can be used to accurately quantify

30 S.J. Baigent et al. / Journal of Virological Methods 233 (2016) 23–36

R² = 0.9413

-0.5

0.5

1.5

2.5

3.5

-0.5 0.5 1.5 2.5 3.5

Leve

l of p

CVI9

88 b

y BA

C-sp

ecifi

c qP

CR(lo

g10

geno

mes

/ 1

0^4

cells

)

Level of pCVI988 by pp38 -CVI qPCR (log10 genome s / 10^4 cell s)

a)

R² = 0.9749

0

1

2

3

4

5

0 1 2 3 4 5

Leve

l of R

B -1B

by

US2

- spe

cific

qPC

R(lo

g10

geno

mes

/ 1

0^4

cells

)

Level of RB-1B by pp38 -Vir(3) qPCR(log10 genome s / 10^4 cells)

b)

R² = 0.9717

-1

0

1

2

3

4

5

-1 0 1 2 3 4 5

Tota

l MDV

-1 b

y pp

38-G

ener

ic q

PCR

(log

10 g

enom

es /

10^

4 ce

lls)

Total MDV -1 by addi �ve data from pp38 -CVI & pp38 -Vir(3) qPCRs(log10 genomes / 10^4 cell s)

c)

R² = 0.9793

-1

0

1

2

3

4

5

-1 0 1 2 3 4 5

Tota

l MDV

-1 b

y ad

di�v

e da

ta fr

om p

p38-

CVI

& p

p38 -

Vir(

3) q

PCRs

(log

10 g

enom

es /

10^4

ce

lls)

Total MDV -1 by addi �ve data from BAC & US2 qPCR s (log10 genomes / 10^4 cells)

d)

Fig. 5. Validation of pp38 SNP q-PCR assays against BAC and US2 q-PCR assays.Each data point shows values for an individual chicken (log10 genomes per 104 cells). (a) Comparison of BAC q-PCR and pp38-CVI q-PCR to measure pCVI988 invaccinated/non-challenged and vaccinated/challenged birds (n = 74). (b) Comparison of US2 q-PCR and pp38-Vir(3) q-PCR to measure RB-1B in challenged/non-vaccinatedand challenged/vaccinated birds (n = 79). (c) Comparison of pp38-Generic q-PCR with additive results from pp38-CVI and pp38-Vir(3) q-PCRs to measure total MDV-1 in allvaccinated and/or challenged birds (n = 102). (d) Comparison of additive results from pp38-CVI and pp38-Vir(3) q-PCRs, with additive results from BAC and US2 q-PCRs tomeasure total MDV-1 in all vaccinated and/or challenged birds (n = 102).

rologic

Cb

3a

qlrivirdFvic

cphioci1MtlwfR

rntioPelr

fpcoIgbCaf

4

qsndbv

S.J. Baigent et al. / Journal of Vi

VI988 vaccine and virulent MDV-1 in tissue samples containingoth viruses in biologically relevant amounts.

.6. Time-course of replication of commercial CVI988 vaccine,nd RB-1B challenge virus in individual chicks

Having demonstrated, in the above experiment, that pp38 SNP-PCR can be used to accurately quantify CVI988 vaccine and viru-

ent MDV-1 in tissue samples containing both viruses in biologicallyelevant amounts, a further experiment was performed to exam-ne replication of RB-1B challenge virus in experimental chickensaccinated with one dose of commercial CVI988 vaccine. For eachndividual chicken, mortality and presence of MD lesions wereecorded, and the level of CVI988 or RB-1B at each time-point wasetermined by pp38-CVI q-PCR and pp38-Vir q-PCR respectively.or each chicken, the detection kinetics of vaccine or challengeirus in PBL (Table 2) broadly correlated with detection kineticsn feather tips (data not shown). CVI988 was detected only in vac-inated chickens and RB-1B only in challenged chickens.

No death and no macroscopic MD lesions were observed in anyhicken in Group A (vaccinated only). Each of these chickens wasositive in pp38-CVI q-PCR and showed a uniform response, withigh early replication of CVI988. RB-1B replicated to high levels

n the PBL and feathers of non-vaccinated chickens (Group B). Tenut of 14 chickens (72%) developed the disease following RB-1Bhallenge, each of these having become RB-1B-positive by 13 dpcn PBL. Mortality occurred in two phases: ‘Early mortality’, from 8 to1 dpc, was associated with rapid onset paralysis (6 chickens); andD tumour phase, 20–33 dpc (4 chickens). At the end of the trial,

hree surviving chickens (#2421, #2436 and #2440) had high RB-1Bevels and were observed to have small MD tumours; these chickens

ould likely have succumbed to MD had the trial continued for aurther two weeks. The fourth survivor (#2342) did not becomeB-1B-positive until 20 dpc, and had no visible lesions.

In Group C (vaccinated and challenged), one chicken (10%)eached humane end-point and was culled at 10 dpc. This bird wasegative for CVI988 but had high early levels of RB-1B, indicatinghat it had not been successfully vaccinated. In the nine surviv-ng chickens CVI988 was readily detected and no MD lesions werebserved. In five of these chickens, RB-1B was not detected in eitherBL or feather tips at any time-point. In the remaining four chick-ns, replication of RB-1B was considerably delayed, and was at aow level. Thus, vaccination had a major inhibitory effect on theeplication of RB-1B in PBL and feathers of most chickens.

The mean level of virus in feather tips and PBL was calculatedor CVI988 (Fig. 6a) and RB-1B (Fig. 6b) for each group at each time-oint. In feathers the mean level of CVI988 was lower in the RB-1Bhallenged group than in the non-challenged group from 33 dpvnly, but the effect of RB-1B was not significant at any time-point.n PBL the mean level of CVI988 was lower in the RB-1B challengedroup throughout the course of the experiment, this differenceeing significant from 17 dpv onwards (p < 0.05). Vaccination withVI988 significantly lowered the mean level of measured RB-1Bt 6 dpc and all time-points thereafter, in both PBL (p < 0.001) andeathers (p < 0.001).

. Discussion

It has been a long-term goal to have a reliable test for differentialuantification of CVI988 vaccine and virulent MDV-1 in chickens,ince this is commercially important to confirm successful vacci-

ation, to investigate causes of vaccine failure and to confirm aiagnosis of MD. To date, development of a q-PCR assay which isoth accurately quantitative, and able to universally distinguish allirulent MDV strains from CVI988, has not been possible.

al Methods 233 (2016) 23–36 31

The current work resulted in development, optimisation andvalidation of an SNP-based real-time PCR assay which was specific,sensitive, efficient, reproducible, free from false-positive results,and enabled accurate differential quantification of CVI988 andAmerican and European virulent MDV-1 strains, over a biologicallyrelevant range of virus levels. The assay therefore offers severaladvantages over previously published differential assays.

It was necessary to identify both a suitable genetic difference onwhich to base the assays, and a suitable technology to enable quan-titative differential detection of the difference. The selected geneticdifference was a consistent SNP at position #320 in the MDV-1 pp38gene, and the selected technology involved use of unconventionallyshort (15–17 nucleotide) TaqMan® probes.

The pp38 gene sequences have now been determined for over 40virulent isolates of MDV (NCBI database). The pp38 gene sequenceis identical in all MDV-1 strains with the exception of bases #320and #326 (Cui et al., 1991; Endoh et al., 1994; Spatz et al., 2007a,b),which differ between CVI988 and virulent strains. The polymor-phisms result in a TspI restriction enzyme site in the pp38 geneof all MDV-1s except CVI988, and different antibody epitopes (Cuiet al., 1999; Lee et al., 2005; Zhizhong et al., 2004), both of whichmay be used to differentiate CVI988 from virulent strains. At SNP#326, CVI988 has ‘G’ and the majority of virulent strains have ‘A’;however, some ‘atypical’ virulent strains (including GA strain, 571strain, three Chinese strains and six recent Pennsylvanian fieldisolates) have ‘G’ and so SNP #326 cannot be used to reliablydistinguish virulent strains from CVI988 (Gimeno et al., 2014). Con-versely, SNP #320 is consistent between CVI988 (which has ‘G’at this position) and all sequenced virulent strains (which have‘A’) (Gimeno et al., 2014; Spatz et al., 2007a). Based on this poly-morphism, Dunn et al. (2010, 2012) developed pyrosequencingassays to distinguish and semi-quantify recombinant viruses (thathad either the virulent-specific or CVI988-specific pp38 gene) invirus super-infection and competition experiments. Although assayspecificity was 100%, sensitivity was relatively poor.

The current study therefore sought to investigate the possibilityof using SNP #320 as a reliable biomarker for differential quantifi-cation of CVI988 and virulent MDV by real-time PCR. Real-time PCRtechnology for differential detection of SNPs is now widely used, butthis only allows confirmation that the virus in question is presentor absent, without quantification. The goal was achieved by usingshort TaqMan® probes to differentially detect PCR products follow-ing amplification with common primers. Short probes can improveSNP detection by increasing quenching efficiency and specificityof PCR product detection. Probe pp38-Vir(3) also incorporates the‘non-consistent’ SNP #326, in order to have a melting temperaturesimilar to that of the pp38-CVI and pp38-Generic probes. This probewas chosen for use in the current work because (for all of the viru-lent MDVs tested here) it showed identical specificity, but greatersensitivity than the probe which contains only SNP #320. Whenthese probes were first designed, atypical MDV field strains having‘G’ at #326 had been observed only very rarely (Spatz et al., 2007a).However, since completion of these studies, several atypical strainshave been identified in some regions of the USA (Gimeno et al.,2014) raising concerns that such strains are becoming increasinglyprevalent. Using probe pp38-Vir(3) could therefore pose problemsfor the analysis of atypical field isolates: there is a possibility thatsuch strains will not be efficiently detected by probe pp38-Vir(3)containing SNP #326 (whereas they would be efficiently detectedusing probe pp38-Vir(1) containing only SNP #320). The currentwork included one atypical MDV (strain 571) which was detectedby both probes pp38-Vir(3) and pp38-Vir(1) with equal sensitiv-

ity, specificity and efficiency, which is reassuring. However, it willbe important to test more atypical strains with both probes and,should issues with sensitivity, specificity and efficiency of probe

32 S.J. Baigent et al. / Journal of Virological Methods 233 (2016) 23–36

Table 2Measurement of CVI988 vaccine and RB-1B challenge virus in PBL of experimental chickens, using pp38 SNP q-PCR assay.

Group Bird # Virus Sampling time-point: PBL (virus level: genome copies per 10,000 cells) Survived S/Died D(time of death dpc)

3 dpv 6 dpv 10 dpv 13 dpv 17 dpv 24 dpv 27 dpv 33 dpv 40 dpv 45 dpv

−4 dpc −1 dpc 3 dpc 6 dpc 10 dpc 17 dpc 20 dpc 26 dpc 33 dpc 38 dpc

AVacc only (level of CVI) a

2428 CVI —c 66 81 43 101 85 79 135 144 NDd S2433 CVI – 7 36 31 40 21 15 23 22 18 S2434 CVI – 74 75 53 75 57 56 42 10 31 S2437 CVI 12 84 110 78 69 93 105 115 137 134 S2442 CVI – 12 77 48 45 54 42 38 54 47 S2444 CVI – 23 61 29 29 44 31 37 40 35 S2446 CVI – 42 9 ND 24 34 15 16 15 9 S2447 CVI – 17 192 52 151 129 91 89 92 113 S

BChall only (level ofRB-1B) b

2401 RB1B – – 3.1 913 13231 15924 30017 De D D D 20 dpc2403 RB1B – – – 133 3138 D D D D D D 11 dpc2405 RB1B – – – 249 D D D D D D D 8 dpc2421 RB1B – – – 336 6586 2968 3783 7878 16290 32241 S2424 RB1B – – – 309 D D D D D D D 10 dpc2429 RB1B – – 57 1421 D D D D D D D 8 dpc2432 RB1B – – – – – – 4 17 1270 – S2435 RB1B – – – 35 1528 1739 2632 3491 9352 D D 33 dpc2436 RB1B – – – 73 2171 2881 5375 5391 4592 4902 S2439 RB1B – 1.1 2.8 190 3631 4170 8707 D D D D 24dpc2440 RB1B – – – 99 4400 2532 3893 6842 7983 14220 S2443 RB1B – – 4.0 225 2493 8329 16123 D D D D 21 dpc2445 RB1B – – 23 697 D D D D D D D 8 dpc2448 RB1B – – – 419 D D D D D D D 8 dpc

CVacc & Chall (level ofCVI and level of RB-1B)

2404 CVI – 10 98 31 18 9 12 12 9 7 SRB1B – – – – – – – – – 87

2407 CVI – 12 44 22 7 12 11 10 11 21 SRB1B – – – – – – – – – –

2414 CVI – 12 53 26 34 32 28 41 50 108 SRB1B – – – – – – – – – –

2417 CVI – 6 25 4 5 2 2 0.3 0.3 1 SRB1B – – – – – – – – – –

2418 CVI – 3 28 33 24 29 25 21 15 24 SRB1B – – – – – – – – – –

2419 CVI – 6 7 10 11 8 16 9 7 12 SRB1B – – – – – – – – – –

2420 CVI – – – – D D D D D D D 10 dpcRB1B – 0.9 0.9 460 D D D D D D

2422 CVI – 23 37 23 10 10 11 9 4 2 SRB1B – – – – – 2 2 10 50 134

2423 CVI – 6 20 31 17 10 12 7 4 0.3 SRB1B – – – – – – – – – 171

2425 CVI – 9 51 65 22 25 8 4 3 4 SRB1B – – – – – – – – – 3

a All Group A birds were negative for RB-1B at each time-point (data not shown).b All Group B birds were negative for CVI988 at each time-point (data not shown).

) and

pp

Gapatit‘tbpatl

c —indicates samples that were considered negative in the q-PCR (40-Ct value ≤1d ND = Not done (chicken inadvertently not sampled at this time-point).e D = bird dead, therefore no sample at this time-point.

p38-Vir(3) be identified, it would be prudent to use only probep38-Vir(1) to test field samples in the future.

Simultaneous to the current work on developing this assay,imeno et al. (2014) sought to use the same polymorphism tochieve the same goal. However, their assay differs from the assayresented in this paper in terms of the molecular methods used tomplify and detect the polymorphism, their ability to fully validatehe assay, and its quantitative accuracy. Specificity in real-time PCRs ideally required at two levels: PCR product specificity (differen-ial amplification) and probe specificity (differential detection). TheGimeno assay’ is based solely on differential amplification, whilehe current assay is based solely on differential detection. Whileoth the current assay and the Gimeno assay have some com-lementary advantages and disadvantages, the former represents

n improvement in that accurate differentiation and quantifica-ion has been validated over a biologically-relevant range of virusevels. Gimeno et al. (2014) developed a Mismatch Amplification

resulted in a baseline value of 0.6 genome copies per 104 cells.

Mutation Assay, which incorporates an intentional mismatch at the3′ penultimate position in the PCR primers, a recognized methodof increasing assay specificity. They used forward primers whichincorporated the SNP at the last base (a CVI988-specific primer anda virulent MDV-specific primer) and a generic reverse primer, in anon-probe (SYBR-green) real-time PCR. PCR assays using primersdesigned to distinguish between a single nucleotide are frequentlyvery inefficient. For this reason, the current work sought to developprobe-based TaqMan® q-PCR using primers that would amplifypp38 of both CVI988 and virulent strains, and to incorporate theSNP into the short probes (a CVI988-specific probe and a virulentMDV-specific probe).

When either CVI988 only or virulent MDV only was present ina DNA sample, both the Gimeno assay and the current assay were

highly specific and either virus could be accurately quantified overa wide dynamic range using standard curves. Both assays werereproducible and very sensitive but, while the limit of detection

S.J. Baigent et al. / Journal of Virological Methods 233 (2016) 23–36 33

0.1

10

1000

100000

10000000

0 5 10 15 20 25 30 35 40 45

Mea

n C

VI9

88 le

vel (

geno

mes

per

10^

4 ce

lls)

Time post vac cination (days)

CVI988-vaccine , no challen ge (Feathers) CV I988 v accine, RB- 1B ch allen ge (Fe ather s)

CVI988-vaccine , no challen ge (PBL) CV I988 v accine, RB- 1B ch allen ge (P BL)

a)

0.1

10

1000

100000

10000000

-5 0 5 10 15 20 25 30 35 40

Mea

n R

B-1

B le

vel (

geno

mes

per

10^

4 ce

lls)

Time post challen ge (days)

No vacci ne, R B-1B ch allen ge (Feathers) CV I988 vaccine, RB- 1B ch allenge (Fe athers)

No vacci ne, R B-1B ch allen ge (PBL) CV I988 vaccine, RB- 1B ch allenge (PBL)

b)

Fig. 6. Time-course of replication of commercial CVI988 vaccine and RB-1B challenge virus in individual chickens.One-day-old Rhode Island Red chickens were divided between three experimental groups. Group A (n = 8) received CVI988 vaccine only, Group B (n = 14) received RB-1Bchallenge virus only, and Group C (n = 10) was both vaccinated and challenged. Feather and blood samples were collected from each chicken at regular intervals. DNA wass he pp3M ype at( terva

itoC

lt(gw

ubjected to pp38 SNP/ovo duplex q-PCR, using either pp38-CVI or pp38-Vir(3) as tean virus levels (with 95% confidence intervals) for each group and each sample t

b). Values of 0.6 (baseline value) are considered negative. Only upper confidence in

n the current assay was approximately 10 copies of virus genome,he Gimeno assay showed higher sensitivity detecting as little asne copy. Thus, both assays could successfully differentially detectVI988 and virulent MDV.

However, both the current assay and the Gimeno assay had someimitations when attempting to specifically and accurately quan-

ify CVI988 or virulent MDV in mixtures of virus-infected cell DNAcurrent assay) or DNA of plasmid carrying an insert of the pp38ene of either CVI988 or virulent MDV (Gimeno et al., 2014) over aide range of genome copies. The Gimeno assay suffered from ‘false

8 detection probe, and the level of each virus quantified as genomes per 104 cells. each time-point are shown for CVI988 vaccine virus (a) and RB-1B challenge virusls are shown for clarity.

positive’ results (loss of specificity if miss-matched sequence is inexcess), and was only accurately quantitative if the homologoussequence was present above a certain level, and the heterologoussequence below a certain level. Conversely, the assay reported inthis paper had issues with ‘false negative’ results (always specific,but loss of sensitivity if miss-matched sequence is in excess). This

assay has an additional level of specificity through use of probeswhich each bound minimally to the miss-matched PCR product.However, use of universal primers to amplify both CVI988 and vir-ulent MDV can yield false negatives since, in mixtures of CVI988

3 rologic

apMptlq

ivivcCCoswtpsCmoaaCto

dRanvsReet(vv(bsiittprbcltRRe

(CmteC

4 S.J. Baigent et al. / Journal of Vi

nd virulent MDV DNA, these two target sequences will compete forrimers. Consistent with this, when DNAs from CVI988 and virulentDV were mixed over a wide range of ratios, competition for PCR

rimers by the predominant virus reduced the amplification (andhus detection) of the less abundant virus. Therefore, although theess abundant virus could still be detected, it could not be accuratelyuantified when the other virus was predominant.

Field samples will contain CVI988 and virulent MDV-1 in vary-ng ratios. Both the current study, and Gimeno et al. (2014),alidated the respective assays using tissue samples from exper-mental chickens which were either CVI988-vaccinated only,vMDV-challenged only, or both CVI988-vaccinated and vvMDV-hallenged. However, while Gimeno et al. (2014) used commercialVI988 vaccine, the current study initially used BAC-clonedVI988 vaccine which gave a significant advantage in confirmationf reliable and confident validation, since the BAC-sequence-pecific/US2-sequence-specific q-PCR system (Baigent et al., 2011),here there is no competition for primers, was used for valida-

ion of the pp38 SNP q-PCR assay. In the current study, use of thep38-Generic reaction to validate the additive results of the virus-pecific assays showed that, over a biologically-relevant range ofVI988 and virulent MDV levels generated in the in vivo challengeodels used here, either virus could be quantified with a high level

f accuracy. When CVI988-vaccination was performed correctlynd effectively, the immune response kept the challenge virus at

level which was too low to impede accurate measurement ofVI988. Conversely, when CVI988 vaccination was not successful,he level of CVI988 was too low to impede accurate measurementf challenge virus.

The pp38 SNP q-PCR assay could therefore be used with confi-ence to follow kinetics of replication of commercial CVI988 andB-1B in individual chickens. Mean kinetics of replication in PBLnd feathers were consistent with previous reports for CVI988 inon-challenged chickens (Baigent et al., 2005b) and RB-1B in non-accinated chickens (Singh et al., 2010). Vaccination with CVI988ignificantly reduced or almost completely inhibited detection ofB-1B in PBL and feathers, consistent with the observations of Islamt al. (2013, 2014) following challenge of CVI988-vaccinated chick-ns with an Australian vvMDV strain and distinction between thewo viruses using the Australian strain-specific meq q-PCR assaysRenz et al., 2013). The reduction in virulent MDV load in CVI988-accinated chickens is greater than that observed in either SB-1accinated chickens (Singh et al., 2010), HVT-vaccinated chickensIslam et al., 2006b; Walkden-Brown et al., 2013a), or MDV-2/HVTivalent-vaccinated chickens (Walkden-Brown et al., 2013a) con-istent with the fact that CVI988 is the most effective vaccine. It wasnteresting that the mean level of CVI988 measured in PBL (but notn feathers) was significantly lower in the RB-1B challenged grouphan in the non-challenged group, whereas Islam et al. (2014) foundhat challenge on or after 5 dpv had little effect on CVI988 load. Theossibility that this result in the current study could be a false result,eflecting competition for q-PCR reagents in samples containingoth CVI988 and RB-1B reducing efficacy of CVI988 detection, wasonsidered. However, given that the measured level of RB-1B wasow in this group, it seems unlikely that such reagent competi-ion had any marked effect on measurement of either CVI988 orB-1B in these samples. Therefore, it is likely that challenge withB-1B had a true effect on reducing replication of CVI988 under thexperimental conditions used in the current study.

Prior to publication of validation of their assay by Gimeno et al.2014), Haq et al. (2012) used this assay to examine replication ofVI988 vaccine and virulent challenge virus in feathers of experi-

ental chickens. Consistent with the results of the current study,

hey observed that challenge with virulent MDV had no significantffect on replication of CVI988 in feathers, but that vaccination withVI988 reduced the level of RB-1B measured in feathers. This reduc-

al Methods 233 (2016) 23–36

tion was more dramatic in the current study, but this may reflectdifferences in the vaccine used (Poulvac CVI988 vs Merial CVI988),the age at which vaccination was performed (day-old vs in ovo), orthe fact that the Gimeno assay may yield false positive results.

The in vivo challenge model used in the current work (usingchickens free of maternal antibody against MDV, a vaccination-challenge interval of 7 days, and vvMDV strain RB-1B as thechallenge virus) did not result in any instance of the presence of rel-atively high levels of both vaccine and challenge virus together inan individual bird, and so there was no opportunity to demonstratethat the assay could accurately differentially quantify relativelyhigh levels of co-infection in a chicken tissue sample. Use of avv+ challenge strain and shorter vaccination-challenge intervalmight have resulted in higher levels of both viruses in the co-infected group. However, the challenge model was carefully chosento aim for a balance between sufficient replication of challengevirus (to ensure that it still replicated in the CVI988-vaccinatedgroup) and sufficient survival of chicks (to ensure multiple sam-ples at multiple time-points for analysis). Inclusion of a challengedbut non-vaccinated control group was necessary in order to observethe effect of vaccination on reduction of challenge virus load, butalmost 50% of non-vaccinated chicks had died by 10 dpc. This earlymortality would have been exacerbated had a shorter vaccination-challenge interval, and/or a vv+ challenge strain (Read et al., 2015)been used, so sufficient data would not have obtained from thiscontrol group, although the co-infected group would likely havehad higher levels of both viruses. Use of chicks having maternal-antibody against MDV (to reduce the issues with early mortality),combined with a higher dose of RB-1B challenge virus, might pro-vide higher levels of CVI988/RB-1B co-infection, and this would beinteresting to examine.

The technology used in these MDV q-PCRs could be applied toother herpesvirus-induced diseases of man and animals, where liveattenuated vaccines are used but there are minimal genetic dif-ferences between vaccine and virulent strains. Similar technologyis available for human Varicella Zoster Virus, where SNPs in fivegenes identify four major subtypes and additional SNPs can dif-ferentiate the live attenuated Oka vaccine strain from wild-typestrains (Quinlivan et al., 2005; Parker et al., 2006), but this was notquantitative.

Live attenuated vaccines, with high genetic similarity to vir-ulent strains, are used against infectious laryngotracheitis virus(ILTV), another herpesvirus of chickens. Generation of new viru-lent ILTV strains by recombination between two attenuated vaccinestrains was recently demonstrated in Australia (Lee et al., 2012).This potentially raises concerns for the long-term reliability of theassay presented in the current work for distinguishing CVI988 fromvirulent MDV. Given the now well-documented instability of thepolymorphism at position 326 (Gimeno et al., 2014; Spatz et al.,2007a), how stable is the ‘consistent’ polymorphism at position320 on which both this assay, and the Gimeno assay rely? Couldmutation, or recombination between CVI988 and virulent strains,result in generation of new virulent viruses which have ‘G’ at posi-tion 320 and are therefore detected by the CVI988-specific assayand not by the virulent-specific assay? Although such viruses havebeen generated in the laboratory by genetic engineering (Cui et al.,1999; Lee et al., 2005) they have never been reported in the field:all virulent isolates sequenced to date, which have been isolatedover a period of 40 years and cover the whole virulence spec-trum, have ‘A’ at position 320. Therefore, this can be considered ahighly stable polymorphism, so the pp38-Vir(1) probe should reli-ably differentiate virulent strains from CVI988 in the long-term.

Nevertheless, should clinical signs of MD be observed in the absenceof detection of any virus by the ‘virulent-specific’ assay, it would beprudent to sequence the whole virus genome and to perform in vivopathogenicity studies in SPF chicks to further characterise the virus.

rologic

aCaaeospe2Vt

A

LaPctwRHob

R

B

B

B

B

B

B

B

B

B

B

C

C

C

C

S.J. Baigent et al. / Journal of Vi

In summary, the pp38 SNP q-PCR assay allows fast, reli-ble, sensitive, specific and accurate differential quantification ofVI988 vaccine and virulent MDV in chicken tissue samples over

biologically-relevant range of virus levels, and thus representsn advance over currently available methods. In the commercialnvironment, the assays could be exploited for confirming successf CVI988 vaccination, for monitoring circulation of virulent fieldtrains and for diagnosis of MD, using feather, blood or organ sam-les; and for monitoring MDV contamination of the poultry housenvironment in samples of poultry dust (Walkden-Brown et al.,013b; Baigent et al., 2013). Ultimately, use of the pp38-CVI, pp38-ir(1) and ovo probes in a single triplex reaction could be optimised

o enable simultaneous quantification of CVI988 and virulent MDV.

cknowledgements

We gratefully acknowledge the help of Dr. Shane Xin (AlleLogictd) for her invaluable input into the design of the pp38 primersnd probes. We thank the staff of The Pirbright Institute Poultryroduction Unit and the Experimental Animal House for everydayare of the chickens, Mrs. L. Kgosana and Mrs. L. Smith for assis-ance with collection of samples from experimental chickens. Thisork was supported by the Biotechnology and Biological Sciencesesearch Council and by a grant from Zoetis (formerly Pfizer Animalealth). The study funders had no input into the design, executionr data analysis of the study, or into the writing of the manuscript,ut gave approval for submission of the manuscript.

eferences

aigent, S., Davison, F., 2004. Marek’s disease virus: biology and life cycle. In:Davison, F., Nair, V. (Eds.), Marek’s Disease, an Evolving Problem. , 1st ed.Elsevier Academic Press, Oxford, pp. 62–77.

aigent, S.J., Petherbridge, L.J., Howes, K., Smith, L.P., Currie, R.J., Nair, V.K., 2005a.Absolute quantitation of Marek’s disease virus genome copy number inchicken feather and lymphocyte samples using real-time PCR. J. Virol. Methods23, 53–64.

aigent, S.J., Smith, L.P., Currie, R.J., Nair, V.K., 2005b. Replication kinetics ofMarek’s disease vaccine virus in feathers and lymphoid tissues using PCR andvirus isolation. J. Gen. Virol. 86, 2989–2998.

aigent, S.J., Smith, L.P., Nair, V.K., Currie, R.J., 2006a. Vaccinal control of Marek’sdisease: current challenges, and future strategies to maximize protection. Vet.Immunol. Immunopathol. 112, 78–86.

aigent, S., Nair, V., Currie, R., 2006b. Real-time quantitative PCR for Marek’sdisease vaccine virus in feather samples: applications and opportunities. In:Vannier, P., Espeseth, D. (Eds.), New Diagnostic Technology: Applications inAnimal Health and Biologics Controls, vol. 126. Basel, Karger, pp. 271–281.

aigent, S.J., Smith, L.P., Currie, R.J., Nair, V.K., 2007. Correlation of Marek’s diseaseherpesvirus vaccine virus genome load in feather tips with protection, using anexperimental challenge model. Avian Pathol. 36, 467–474.

aigent, S.J., Smith, L.P., Petherbridge, L.J., Nair, V.K., 2011. Differentialquantification of cloned CVI988 vaccine strain and virulent RB-1B strain ofMarek’s disease viruses in chicken tissues using real-time PCR. Res. Vet. Sci. 91,167–174.

aigent, S.J., Kgosana, L.B., Gamawa, A.A., Smith, L.P., Read, A.F., Nair, V.K., 2013.Relationship between levels of very virulent MDV in poultry dust and infeather tips from vaccinated chickens. Avian Dis. 57, 440–447.

arrow, A., Venugopal, K., 1999. Molecular characteristics of very virulentEuropean MDV isolates. Acta Virol. 43, 90–93.

ecker, Y., Asher, Y., Tabor, E., Davidson, I., Malkinson, M., Weisman, Y., 1992.Polymerase chain reaction for differentiation between pathogenic andnon-pathogenic serotype 1 Marek’s disease viruses (MDV) and vaccine virusesof MDV-serotypes 2 and 3. J. Virol. Methods 40, 307–322.

alnek, B.W., 2001. Pathogenesis of Marek’s disease virus infection. In: Hirai, K.(Ed.), Current Topics in Microbiology and Immunology, vol. 255. Springer,Berlin, pp. 25–55.

hang, K.-S., Ohashi, K., Onuma, M., 2002. Diversity (polymorphism) of the meqgene in the attenuated Marek’s disease virus (MDV) serotype 1 andMDV-transformed cell lines. J. Vet. Med. Sci. 64, 1097–1101.

ui, Z.Z., Lee, L.F., Liu, J.L., Kung, H.J., 1991. Structural analysis and transcriptional

mapping of the Marek’s disease virus gene encoding pp38, an antigenassociated with transformed cells. J. Virol. 65, 6509–6515.

ui, Z.Z., Qin, A., Lee, L.F., Wu, P., Kung, H.J., 1999. Construction andcharacterisation of a H19 epitope point mutant of MDV CVI988/Rispens strain.Acta Virol. 43, 169–173.

al Methods 233 (2016) 23–36 35

Davidson, I., Borenshtain, R., Weisman, Y., 2002. Molecular identification of theMarek’s disease virus vaccine strain CVI988 in vaccinated chickens. J. Vet. Med.49, 83–87.

Davison, F., Nair, V., 2005. Use of Marek’s disease vaccines: could they be drivingthe virus to increasing virulence? Expert Rev. Vaccines 4, 77–88.

de Boer, G.F., Groenendal, J.E., Boerrigter, H.M., Kok, G.L., Pol, J.M., 1986. Protectiveefficacy of Marek’s disease virus (MDV) CVI-988CEF65 clone C againstchallenge infection with three very virulent MDV strains. Avian Dis. 30,276–283.

Dunn, J.R., Witter, R.L., Silva, R.F., Lee, L.F., Finlay, J., Marker, B.A., Kaneene, J.B.,Fulton, R.M., Fitzgerald, S.D., 2010. The effect of the time interval betweenexposures on the susceptibility of chickens to superinfection with Marek’sdisease virus. Avian Dis. 54, 1038–1049.

Dunn, J.R., Silva, R.F., Lee, L.F., Witter, R.L., 2012. Competition between two virulentMarek’s disease virus strains in vivo. Avian Pathol. 41, 267–275.

Endoh, D., Niikura, M., Hirai, K., Inagaki, N., Hayashi, M., Nakajima, K., Sato, F., 1994.Expression and purification of recombinant Marek’s disease virus serotype 1specific phosphorylated protein pp38 in E. coli. J. Vet. Med. Sci 56, 823–826.

Gimeno, I.M., Dunn, J.R., Cortes, A.L., El-Gohary, A.E.-G., Silva, R.F., 2014. Detectionand differentiation of CVI988 (Rispens vaccine) from other serotype 1 Marek’sdisease viruses. Avian Dis. 58, 232–243.

Gimeno, I.M., 2008. Marek’s disease vaccines: a solution for today but a worry fortomorrow? Vaccine 26 (Suppl. (3)), C31–41.

Haq, K., Fear, T., Ibraheem, A., Abdul-Careem, M.F., Sharif, S., 2012. Influence ofvaccination with CVI988/Rispens on load and replication of a very virulentMarek’s disease virus strain in feathers of chickens. Avian Pathol. 41, 69–75.

Islam, A., Harrison, B., Cheetham, B.F., Mahony, T.J., Young, P.L., Walkden-Brown,S.W., 2004. Differential amplification and quantitation of Marek’s diseaseviruses using real-time polymerase chain reaction. J. Virol. Methods 119,103–113.

Islam, A., Cheetham, B.F., Mahony, T.J., Young, P.L., Walkden-Brown, S.W., 2006a.Absolute quantitation of Marek’s disease virus and Herpesvirus of turkeys inchicken lymphocyte, feather tip and dust samples using real-time PCR. J. Virol.Methods 132, 127–134.

Islam, A.F., Walkden-Brown, S.W., Islam, A., Underwood, G.J., Groves, P.J., 2006b.Relationship between Marek’s disease virus load in peripheral bloodlymphocytes at various stages of infection and clinical Marek’s disease inbroiler chickens. Avian Pathol. 35, 42–48.

Islam, T., Walkden-Brown, S.W., Renz, K.G., Fakhrul Islam, A.F.M., Ralapanawe, S.,2013. Vaccination-challenge interval markedly influences protection providedby Rispens CVI988 vaccine against very virulent Marek’s disease viruschallenge. Avian Pathol. 42, 516–526.

Islam, T., Walkden-Brown, S.W., Renz, K.G., Fakhrul Islam, A.F.M., Ralapanawe, S.,2014. Replication kinetics and shedding of very virulent Marek’s disease virusand vaccinal Rispens/CVI988 virus during single and mixed infections varyingin order and interval between infections. Vet. Microbiol. 173, 208–223.

King, D., Page, D., Schat, K.A., Calnek, B.W., 1981. Difference between influences ofhomologous and heterologous maternal antibodies on response to serotype-2and serotype-3 Marek’s disease vaccines. Avian Dis. 25, 74–81.

Landman, W.J., Verschuren, S.B., 2003. Titration of Marek’s disease cell-associatedvaccine virus (CVI 988) of reconstituted vaccine and vaccine ampoules fromDutch hatcheries. Avian Dis. 47, 1458–1465.

Lee, S.I., Takagi, M., Ohashi, K., Sugimoto, C., Onuma, M., 2000. Difference in themeq gene between oncogenic and attenuated strains Marek’s disease virusserotype 1. J. Vet. Med. Sci. 62, 287–292.

Lee, L.F., Cui, X., Cui, Z., Gimeno, I., Lupiani, B., Reddy, S.M., 2005. Characterizationof a very virulent Marek’s disease virus mutant expressing the pp38 proteinfrom the serotype 1 vaccine strain CVI988/Rispens. Virus Genes 31, 73–80.

Lee, S.-W., Markham, P.F., Coppo, M.J.C., Legione, A.R., Markham, J.F.,Noormohammadi, A.H., Browning, G.F., Ficorilli, N., Hartley, C.A., Devlin, J.M.,2012. Attenuated vaccines can recombine to form virulent field viruses.Science 337, 188.

Murata, S., Chang, K.S., Lee, S.I., Konnai, S., Onuma, M., Ohashi, K., 2007.Development of a nested polymerase chain reaction method to detectoncogenic Marek’s disease virus from feather tips. J. Vet. Diagn. Invest. 19,471–478.

Niikura, M., Dodgson, J.B., Cheng, H.H., 2006. Stability of Marek’s disease virus132-bp repeats during serial in vitro passages. Arch. Virol. 151, 1431–1438.

Okazaki, W., Purchase, H.G., Burmester, B.R., 1970. Protection against Marek’sdisease by vaccination with a herpesvirus of turkeys. Avian Dis. 14, 413–429.

Parker, S.P., Quinlivan, M., Taha, Y., Breuer, J., 2006. Genotyping of varicella-zostervirus and the discrimination of Oka vaccine strains by TaqMan® real-time PCR.J. Clin. Microbiol. 44, 3911–3914.

Petherbridge, L., Howes, K., Baigent, S.J., Sacco, M.A., Evans, S., Osterrieder, N., Nair,V., 2003. Replication-competent bacterial artificial chromosomes of Marek’sdisease virus: novel tools for generation of molecularly defined herpesvirusvaccines. J. Virol. 77, 8712–8718.

Petherbridge, L., Brown, A.C., Baigent, S.J., Howes, K., Sacco, M., Osterrieder, N.,Nair, V.K., 2004. Oncogenicity of virulent Marek’s disease virus cloned asbacterial artificial chromosomes. J. Virol. 78, 13376–13380.

Quinlivan, M., Gershon, A.A., Steinberg, S.P., Breuer, J., 2005. An evaluation of single

nucleotide polymorphisms used to differentiate vaccine and wild-type strainsof varicella-zoster virus. J. Med. Virol. 75, 174–180.

Read, A.F., Baigent, S.J., Powers, C., Kgosana, L.B., Blackwell, L., Smith, L.P., Kennedy,D., Walkden-Brown, S.W., Nair, V.K., 2015. Imperfect vaccination can enhancethe transmission of highly virulent pathogens. PLoS Biol. 13 (7).

3 rologic

R

R

R

S

S

S

S

S

S

comparing chicken immunological reactions to viruses point-mutated in the

6 S.J. Baigent et al. / Journal of Vi

enz, K.G., Islam, A., Cheetham, B.F., Walkden-Brown, S.W., 2006. Absolutequantification using real-time polymerase chain reaction of Marek’s diseasevirus serotype 2 in field dust samples, feather tips and spleens. J. Virol.Methods 135, 186–191.

enz, K.G., Cheetham, B.F., Walkden-Brown, S.W., 2013. Differentiation betweenpathogenic serotype 1 isolates of Marek’s disease virus and the Rispens CVI988vaccine in Australia using real-time PCR and high resolution melt curveanalysis. J. Virol. Methods 187, 144–152.

ispens, B.H., Van Vloten, J., Mastenbroek, N., Maas, H.J.L., Schat, K.A., 1972. Controlof marek’s disease in the Netherlands. I. Isolation of an avirulent Marek’sdisease virus (strain CVI 988) and its use in laboratory vaccination trials. AvianDis. 16, 108–125.

chat, K.A., Calnek, B.W., 1978. Characterization of an apparently nononcogenicMarek’s disease virus. J. Natl. Cancer Inst. 60, 1075–1082.

chat, K.A., Calnek, B.W., Fabricant, J., 1982. Characterisation of two highlyoncogenic strains of Marek’s disease virus. Avian Pathol. 11, 593–605.

ilva, R.F., Gimeno, I., 2007. Oncogenic Marek’s disease viruses lacking the 132 basepair repeats can still be attenuated by serial in vitro cell culture passages. VirusGenes 34, 87–90.

ilva, R.F., 1992. Differentiation of pathogenic and non-pathogenic serotype 1Marek’s disease viruses (MDVs) by the polymerase chain reaction amplificationof the tandem direct repeats within the MDV genome. Avian Dis. 36, 521–528.

ingh, S.M., Baigent, S.J., Petherbridge, L.J., Smith, L.P., Nair, V.K., 2010. Comparativeefficacy of BAC-derived recombinant SB-1 vaccine and the parent wild typestrain in preventing replication, shedding and disease induced by virulent

Marek’s disease virus. Res. Vet. Sci. 89, 140–145.

mith, L.P., Petherbridge, L.J., Baigent, S.J., Simpson, J., Nair, V., 2011. Pathogenicityof a very virulent strain of Marek’s disease herpesvirus cloned as infectiousbacterial artificial chromosomes. J. Biomed. Biotechnol., 412829, http://dx.doi.org/10.1155/2011/412829.

al Methods 233 (2016) 23–36

Spatz, S.J., Silva, R.F., 2007. Polymorphisms in the repeat long regions of oncogenicand attenuated pathotypes of Marek’s disease virus 1. Virus Genes 35, 41–53.

Spatz, S.J., Petherbridge, L., Zhao, Y., Nair, V., 2007a. Comparative full-lengthsequence analysis of oncogenic and vaccine (Rispens) strains of Marek’sdisease virus. J. Gen. Virol. 88, 1080–1096.

Spatz, S.J., Zhao, Y., Petherbridge, L., Smith, L.P., Baigent, S.J., Nair, V., 2007b.Comparative sequence analysis of a highly oncogenic but horizontalspread-defective clone of Marek’s disease virus. Virus Genes 35, 753–766.

Walkden-Brown, S.W., Islam, A., Islam, A.F., Burgess, S.K., Groves, P.J., Cooke, J.,2013a. Pathotyping of Australian isolates of Marek’s disease virus incommercial broiler chickens vaccinated with herpesvirus of turkeys (HVT) orbivalent (HVT/SB1) vaccine and association with viral load in the spleen andfeather dander. Aust. Vet. J 91, 341–350.

Walkden-Brown, S.W., Islam, A., Groves, P.J., Rubite, A., Sharpe, S.M., Burgess, S.K.,2013b. Development, application and results of routine monitoring of Marek’sdisease virus in broiler house dust using real time quantitative PCR. Avian Dis.57, 544–554.

Witter, R.L., 1997. Increased virulence of Marek’s disease virus field isolates. AvianDis. 41, 149–163.

Witter, R.L., Calnek, B.W., Buscaglia, C., Gimeno, I.M., Schat, K.A., 2005.Classification of Marek’s disease viruses according to pathotype: philosophyand methodology. Avian Pathol. 34, 75–90.

Zhizhong, C., Zhi, Z., Aijian, Q., Lucy, L.F., 2004. Analyzing the H19- andT65-epitopes in 38 kd phosphorylated protein of Marek’s disease viruses and

epitopes. Sci. China C Life Sci. 47, 82–91.van Iddekinge, B.J., Stenzler, L., Schat, K.A., Boerrigter, H., Koch, G., 1999. Genome

analysis of Marek’s disease virus strain CVI-988: effect of cell culture passageon the inverted repeat regions. Avian Dis. 43, 182–188.