S

BhT

NV

a

ARRA

KMCBCT

1

El1par1taswito

c

v

h0b

y COREView me

er Connector

Veterinary Immunology and Immunopathology 167 (2015) 80–85

Contents lists available at ScienceDirect

Veterinary Immunology and Immunopathology

jo ur nal ho me pag e: www.elsev ier .com/ locate /vet imm

hort communication

oLA-6*01301 and BoLA-6*01302, two allelic variants of the A18aplotype, present the same epitope from the Tp1 antigen ofheileria parva

. Svitek, E. Awino, V. Nene, L. Steinaa ∗

accine Biosciences, International Livestock Research Institute, P.O. Box 30709, Nairobi 00100, Kenya

r t i c l e i n f o

rticle history:eceived 2 April 2015eceived in revised form 28 May 2015ccepted 12 June 2015

eywords:HC polymorphism

ytotoxic T cells

a b s t r a c t

We have recently shown that the BoLA-A18 variant haplotype (BoLA-6*01302) is more prevalent than theBoLA-A18 haplotype (BoLA-6*01301) in a sample of Holstein/Friesian cattle in Kenya. These MHC classI allelic variants differ by a single amino acid polymorphism (Glu97 to Leu97) in the peptide-bindinggroove. We have previously mapped an 11-mer peptide epitope from the Theileria parva antigen Tp1(Tp1214–224) that is presented by BoLA-6*01301. Crystal structure data indicates that Glu97 in the MHCmolecule plays a role in epitope binding through electro-static interaction with a lysine residue in position5 of the epitope, which also functions as an additional anchor residue. In contrast to expectations, we

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

provided by Elsevier - Publish

oLATL epitopeheileria parva

demonstrate that the amino acid substitution in BoLA-6*01302 does not divert the CTL response awayfrom Tp1214–224. The two MHC molecules exhibit similar affinity for the Tp1 epitope and can present theepitope to parasite-specific CTLs derived from either BoLA allelic variants. These data confirm that thisBoLA polymorphism does not alter Tp1 epitope specificity and that both allelic variants can be used forTp1 vaccine studies.

© 2015 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND

. Introduction

Theileria parva, a protozoan parasite that causes the diseaseast Coast fever (ECF) in cattle is responsible for major economicosses in eastern, central and southern Africa (De Leeuw et al.,995; Mukhebi, 1992). CD8 T cells that are primed by a live T.arva infection and treatment method (ITM) are believed to medi-te protection to disease and several T. parva CD8 T-cell epitopesecognized by immune CTL have been identified (McKeever et al.,994; Morrison et al., 1995). Unlike in malaria, the specificity ofhe CTL response to this vaccine in individual cattle is skewed to

few epitopes and most likely contributes to the observation oftrain specific immunity in ECF (MacHugh et al., 2009). Vaccinationith a cocktail of T. parva parasite isolates provides broad-spectrum

mmunity to ECF (Di Giulio et al., 2009). Armed with this informa-ion, our research is focused on the identification and development

f a subunit vaccine for the control of ECF.

It has been shown that ITM vaccination of Holstein/Friesianattle homozygous for the BoLA-A18 haplotype with the Muguga

∗ Corresponding author. Tel.: +254 717 666 290.E-mail addresses: n.svitek@cgiar.org (N. Svitek), e.awino@cgiar.org (E. Awino),

.nene@cgiar.org (V. Nene), l.steinaa@cgiar.org (L. Steinaa).

ttp://dx.doi.org/10.1016/j.vetimm.2015.06.007165-2427/© 2015 The Authors. Published by Elsevier B.V. This is an open accessy-nc-nd/4.0/).

license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

isolate of T. parva results in approximately 70% of the CTL beingdirected to a single epitope on an antigen called Tp1 (MacHughet al., 2009). The epitope maps to residues 214–224 (Tp1214–224) andis presented by BoLA-6*01301 (Graham et al., 2006, 2008). A recom-binant version of this MHC class I molecule has been crystalizedtogether with the Tp1 peptide epitope, VGYPKVKEEML (Macdonaldet al., 2010). According to this model the 11-mer peptide bindsBoLA-6*01301 in a raised conformation and Glu97, an amino acidresidue located in the BoLA peptide binding groove bonds withlysine at position 5 (P5) in the Tp1 peptide. An alanine substitu-tion scan of the Tp1 epitope sequence revealed the C-terminusand the P5 position as anchor residues for binding to the MHCmolecule. The BoLA-6*01301 molecule also binds nonameric pep-tides and an alanine scan of a high affinity self-peptide, TIMPKDIQL,revealed position P2 as an additional anchor residue to P5 and theC-terminus. This peptide binding groove appears to be quite flex-ible as a truncated 10-mer Tp1 peptide, GYPKVKEEML, also bindsto BoLA-6*01301 but it is not recognized by Tp1-specific CTL sug-gesting the 10-mer peptide adopts a different conformation to the11-mer (Macdonald et al., 2010).

In a recent survey of more than 200 Holstein/Friesian cat-tle in Kenya we found an A18 haplotype with a variant of theBoLA-6*01301 allele, BoLA-6*01302, to be more prevalent thanBoLA-6*01301 (Svitek et al., 2015) and more recent BoLA-A18

article under the CC BY-NC-ND license (http://creativecommons.org/licenses/

N. Svitek et al. / Veterinary Immunology and Immunopathology 167 (2015) 80–85 81

Fig. 1. Characterization of CTL recognizing Tp1214–224 in cattle of BoLA-6*01302 and BoLA 6*01301 type. CTL from BA219 (A), BF092 (B and D), BV115 (C) were tested for lysisof Tp1214–224 pulsed PBMC and of T. parva (Muguga) infected cell lines (TpM). (A and B): CTL from BA219 and BF092 (BoLA-6*01302) were tested for lysis of autologous PBMCp nd D)C pM anp

tctotwncC

iaBp

ulsed with 1 �M Tp1 peptide, PBMC alone, autologous TpM and control TpM. (C aTL can kill T. parva infected target cells expressing either BoLA-6*01301 (BV115 Terformed twice with similar result.

yping indicate that the BoLA-6*01302 is more than 90% prevalentompared to the BoLA-6*01301 allele (data not shown). There arewo single nucleotide differences between these alleles but onlyne translates into an amino acid substitution, Glu → Leu at posi-ion 97 (Immuno Polymorphism Database (part: IPD-MHC): http://ww.ebi.ac.uk/ipd/mhc/bola/align.htm). This is a change from aegatively charged amino acid to a hydrophobic one raising con-erns that this polymorphism could lead to a change in the primaryTL specificity in BoLA-6*01302 cattle vaccinated by ITM.

As part of a newly established East Coast fever consortium, it

s planned to use the Tp1 antigen as a model antigen for gener-ting CTL in cattle with various delivery systems and adjuvants.ecause the BoLA-6*01302 allele is much more prevalent in Euro-ean cattle in Kenya and possibly also in other African countries,

: BV115 CTL (BoLA 6*01301) and BF092 CTL (BoLA 6*01302) demonstrate that thed BH054 TpM) or BoLA-6*01302 allele (BF092 and BA219). The experiments were

it is very difficult to source sufficient numbers of animals of theBoLA-6*01301 type for experiment. Hence, we decided to eluci-date if BoLA-6*01302 animals also generated CTL toward the Tp1epitope and if there were any major differences in peptide–BoLAbinding assays.

2. Material and methods

2.1. Cell lines

Cell lines infected with T. parva were established by infection ofPBMC in vitro with sporozoites as described previously (Goddeerisand Morrison, 1988) Cryopreserved sporozoites stabilates of the T.parva Muguga 3308 were used for the production of infected cell

82 N. Svitek et al. / Veterinary Immunology and Immunopathology 167 (2015) 80–85

F ) BV1c oLA-66

lia

Tpaa

2

ueiLcrTw(P

2c

fpB(2(eto

ig. 2. Strength of CTL responses using BoLA-6*01301 and BoLA-6*01302 targets. (Aoncentrations on PBMC from BF091, BH054 (BoLA-6*01301) and BF092, BA219 (B:1 for BA219. The experiment was performed twice with similar result.

ines. CTL lines were generated by re-stimulation of PBMC frommmunized cattle with autologous infected cell lines (Goddeerisnd Morrison, 1988; Steinaa et al., 2012).

Infected cell lines, BA219 TpM, BF092 TpM, BH054 TpM, BV115pM and CTL lines BA219, BF092, BH054 and BV115 were fromrevious studies (Graham et al., 2006; Wendoh et al., 2014). BA219nd BF092 expressed the BoLA-6*01302 A18 allele whereas BH054nd BV115 expressed the BoLA-6*01301 A18 allele.

.2. Cytotoxicity assay

Standard 4 h release assay using 51Cr-labeled target cells weresed to measure cytotoxicity (Goddeeris & Morrison, 1988; Steinaat al., 2012). 51Cr was obtained from American Radiolabeled Chem-cals, Inc., St. Louis, MO, USA. Supernatants were counted usingumaplates in a TopCounter (PerkinElmer, Waltham, MA, USA). Theytotoxicity was calculated as: (experimental release-spontaneouselease/total release-spontaneous release). Target cells were either. parva infected PBMC or peptide pulsed PBMC. Autologous PBMCere pulsed with 10-fold dilutions of the Tp1214–224 CTL epitope

Mimotopes, Clayton, Victoria, Australia), and labeled with 51Cr.BMC were washed twice in PBS before use.

.3. Production of peptide–MHC class I tetramers and flowytometry

Generation of peptide (p)–MHC class I tetramers was per-ormed as described in (Svitek et al., 2014). Tp1-BoLA-6*01301–MHC tetramers were labeled with allophycyanin (APC) and Tp1-oLA-6*01302 p–MHC tetramers were labeled with phycoerythrinPE). Flow cytometry was performed as described in (Svitek et al.,014). Briefly, BB007 Tp1-specific CTL line was stained with FVS450

BD Biosciences), anti-CD8 ILRI monoclonal antibody ILA51 andither Tp1–BoLA-6*01301–APC or Tp1–BoLA6*01302-PE p–MHCetramers or both together. Anti-IgG1-FITC was used as the sec-ndary antibody. Flow cytometry data were acquired on a BD

15 CTL, (B) BA219 CTL. Target cells: Tp1214–224 epitope was titrated in the indicated*01302). CTL assays were performed at a effector:target ratio of 3:1 for BV115 and

FACSCantoTM II instrument (BD Biosciences) and analyzed withFlowJo.

2.4. Peptide–MHC class I binding assay

Tp1 binding affinity assay was performed as described in (Sviteket al., 2014) except that human �2m was used instead of the bovine�2m at the concentration of 150 nM and heavy chain concentrationof 25 nM. BoLA-6*01301 and BoLA-6*01302 heavy chain moleculeswere incubated with the Tp1 peptide at concentrations of 3.2 nM,16 nM, 80 nM, 400 nM, 2 �M, 10 �M, and 50 �M. Kd values werecalculated by Graph Pad Prism by generating a fit curve using theOne Site-Specific binding equation.

2.5. Peptide prediction and logo generation

Twenty-six protein sequences from Theileria parva, includingthe Tp1 sequence, were used to predict top peptide binders usingNetMHCpan 2.8 (http://www.cbs.dtu.dk/services/NetMHCpan) ofBoLA-6*01301 and BoLA-6*01302 molecules using the 11-mer pep-tide length and the top 2% binders were used to generate the logoswith Seq2logo (http://www.cbs.dtu.dk/biotools/Seq2Logo/).

3. Results and discussion

Cattle of the BoLA-A18 serological specificity were typed usingsequence specific primer PCR, which amplify both of the BoLA-A18 haplotype alleles, BoLA-6*01301 and BoLA-6*01302. The PCRproducts were discriminated by restriction fragment length poly-morphism, as the BoLA-6*01301 allele contains a site for PvuII notpresent in BoLA-6*01302 (Svitek et al., 2015). The allele assign-

ments were further confirmed by sequencing (data not shown).

First, we tested two cattle of the BoLA-6*01302 type (BA219and BF092), not selected to be homozygous, for generation ofCTL toward Tp1214–224 following ITM with the Muguga isolate of

N. Svitek et al. / Veterinary Immunology and Immunopathology 167 (2015) 80–85 83

Fig. 3. Multiparametric flow cytometry analysis of Tp1–BoLA-6*01301/01302 p–MHC class I tetramer TCR binding specificities. (A) lymphocytes population, (B) single cellsp unstaip ation.

TeaacomBcat

opulation, (C) live cells population, (D) CD8 positive (+) population, (E) negative

opulation, and (H) double Tp1–BoLA-6*01301-APC/Tp1–BoLA-6*01302-PE+ popul

. parva (Radley et al., 1975). Antigen specific CTL lines were gen-rated as previously described by re-stimulation of PBMC withutologous T. parva infected lymphoblastoid cell lines (Goddeerisnd Morrison, 1988; Steinaa et al., 2012). These were tested inytolytic assays using Tp1214–224 pulsed autologous PBMC, PBMCnly, the autologous Muguga 3308 infected cell lines and a BoLAismatched T. parva infected cell line as target cells (Fig. 1A and

). Both animals generated robust CTL toward autologous infectedells and the Tp1 epitope. This clearly shows that the BoLA-6*01302nimals generate CTL to the same Tp1 epitope as BoLA-6*01301 cat-le. Apart from the CTL assays as shown here, this was confirmed

ned control, (F) Tp1–BoLA-6*01301-APC+ population, (G) Tp1–BoLA-6*01302-PE+

Arrows in orange indicate origin of each gated populations.

using interferon-� ELISPOT assays using peptide pulsed targets(data not shown).

Next, we performed a CTL experiment where CTL generatedfrom cattle possessing one of the two MHC alleles were testedon targets expressing the other MHC allele. If there was no dif-ference in presenting the peptide between the two MHC types,CTL from each type should kill infected cells of both types (Fig. 1C

and D). As seen in Fig. 1C and D, this was the case. For simplic-ity only one animal of each of the two MHC types is shown. InFig. 1C, BV115 CTL (BoLA-6*01301) killed infected cell lines gen-erated from both BoLA-6*01302 animals (BA219 and BF092) to

84 N. Svitek et al. / Veterinary Immunology and Immunopathology 167 (2015) 80–85

Fig. 4. (A) BoLA-6*01301 (blue) and BoLA-6*01302 (red) binding specificity for Tp1. Graph is depicting the curve of Tp1 concentrations up to 2 �M. The solid lines representthe fit curves and the dashed lines indicate the concentration at which we observe 50% folding of the BoLA molecules (Kd values). (B) Sequence logo of the amino acidp s.

tBBltatpeebaBairTda

(atAstpatcFt

references of BoLA-6*01301 (left panel) and BoLA-6*01302 (right panel) molecule

he same extent as infected BoLA-6*01301 cell lines (BV115 andH054). Likewise, in Fig. 1D, BF092 CTL (BoLA-6*01302) killed theoLA-6*01301 target cells equally well as the BoLA-6*01302 cell

ines. Since the target cells were parasite infected cells it leaveshe possibility that CTL specificities besides those to Tp1 couldccount for the observed cross killing. Therefore, we decided toest the CTL on PBMC pulsed with different concentrations of Tp1eptide from 10−11 M to 10−6 M (Fig. 2), which in addition to thepitope specific response also will show the cytotoxicity at limitingpitope concentrations, a condition which can reveal differentialinding to the MHC/peptide complex. BV115 CTL (BoLA-6*01301)nd BA219 CTL (BoLA-6*01302) were tested on PBMC from the twooLA-6*01301 animals (BF091, BH054), and the two BoLA-6*01302nimals (BF092, BA219). As seen, there were no major differencesn the ability of the CTL lines to lyse the different target cells. Similaresults were obtained from two other CTL lines (data not shown).he slight differences in killing profile seen in Fig. 2 may be due toifferences in expression levels of the MHC molecules in differentnimals.

We also tested the binding of two Tp1–tetramers to a CTL lineBB007) expressing BoLA-6*01301, one based on BoLA-6*01301nd the other based on BoLA-6*01302. The Tp1214–224 epitopeetramers were labeled with different fluorochromes (PE or APC).s seen in Fig. 3, gating was sequentially performed on forwardcatter (FSC)/side scatter (SSC) as a rough dead/live discrimina-ion, then singlets followed by gating on viable cells. From thisopulation of live cells, CD8 positive/tetramer double positive cellsre shown in Fig. 3H. Clearly, the TCR expressed by BB007 bound

he two tetramers equally well. This was also the case when aell line expressing BoLA 6*01302 was used (data not shown).inally, we also assayed the binding of the peptide epitope to thewo MHC alleles (Fig. 4A), using an ELISA binding assay previously

described (Svitek et al., 2014; Sylvester-Hvid et al., 2002). Theseresults showed that the Kd of the binding of the peptide to the MHCmolecules were similar, 7.47 nM for BoLA-6*01301 versus 8.63 nMfor BoLA-6*01302 with no statistical significant difference betweenthem.

This study shows that the polymorphism found in amino acidposition 97 of BoLA-6*01302 does not affect the priming andgeneration of CTL to the known Tp1 epitope restricted by BoLA-6*01301. It appears that the BoLA Glu97–LysP5 peptide interactionas described by Macdonald et al. (2010) is not critical and that otherinteractions between the Tp1 peptide and BoLA-6*01302 moleculecompensate for the Glu97 to Leu97 sequence polymorphism in thepeptide binding groove. Interestingly, our experimental data verifyNetMHCpan predictions that Tp1214–224 remains a strong peptidebinder of BoLA-6*01302 (data not shown).

In conclusion, animals possessing either BoLA A18 alleles, BoLA-6*01301 or BoLA-6*01302, can be used/interchanged for studyingthe CTL response to Tp1214–224 in subunit vaccine studies. Whetherthis can be extrapolated to other epitopes restricted by thesealleles would have to be experimentally verified on a case-to-casebasis since it remains a possibility that the BoLA-A18 amino acidpolymorphism may affect binding of other epitopes. We used theNetMHCpan server to predict the preferred amino acid at the dif-ferent positions in an 11mer binding peptide condition and therelative importance of each positions in the binding is illustrated inFig. 4B. The Logos and amino acids at anchor positions are indeedvery similar, which suggest that the two BoLA molecules will bindsimilar epitopes.

Conflict of interests

The authors declare that they have no conflict of interests.

y and

A

tFT

R

D

D

G

G

G

M

N. Svitek et al. / Veterinary Immunolog

cknowledgements

This study was funded by the Norman Borlaug Commemora-ive Research Initiative (#58-5348-2-117F), Bill & Melinda Gatesoundation (#OPP107879) and National Council for Science andechnology (Kenya) (#NSC/5/003/W/2ndCALL/102).

eferences

e Leeuw, P.N., McDermott, J.J., Lebbie, S.H.B., 1995. Monitoring of livestock healthand production in sub-Saharan Africa. Prev. Vet. Med. 25, 195–212.

i Giulio, G., Lynen, G., Morzaria, S., Oura, C., Bishop, R., 2009. Live immunizationagainst East Coast fever – current status. Trends Parasitol. 25, 85–92.

oddeeris, B.M., Morrison, W.I., 1988. Techniques for generation, cloning, andcharachterization of bovine cytotoxic T cells specific for the protozoan Theileriaparva. J. Tiss. Cult. Methods 11, 101.

raham, S.P., Pelle, R., Honda, Y., Mwangi, D.M., Tonukari, N.J., Yamage, M., Glew,E.J., de Villiers, E.P., Shah, T., Bishop, R., Abuya, E., Awino, E., Gachanja, J., Luyai,A.E., Mbwika, F., Muthiani, A.M., Ndegwa, D.M., Njahira, M., Nyanjui, J.K.,Onono, F.O., Osaso, J., Saya, R.M., Wildmann, C., Fraser, C.M., Maudlin, I.,Gardner, M.J., Morzaria, S.P., Loosmore, S., Gilbert, S.C., Audonnet, J.C., van derBruggen, P., Nene, V., Taracha, E.L., 2006. Theileria parva candidate vaccineantigens recognized by immune bovine cytotoxic T lymphocytes. Proc. Natl.Acad. Sci. U. S. A. 103, 3286–3291.

raham, S.P., Pelle, R., Yamage, M., Mwangi, D.M., Honda, Y., Mwakubambanya,R.S., de Villiers, E.P., Abuya, E., Awino, E., Gachanja, J., Mbwika, F., Muthiani,A.M., Muriuki, C., Nyanjui, J.K., Onono, F.O., Osaso, J., Riitho, V., Saya, R.M., Ellis,S.A., McKeever, D.J., MacHugh, N.D., Gilbert, S.C., Audonnet, J.C., Morrison, W.I.,van der Bruggen, P., Taracha, E.L., 2008. Characterization of the fine specificityof bovine CD8 T-cell responses to defined antigens from the protozoan parasite

Theileria parva. Infect. Immun. 76, 685–694.

acdonald, I.K., Harkiolaki, M., Hunt, L., Connelley, T., Carroll, A.V., MacHugh, N.D.,Graham, S.P., Jones, E.Y., Morrison, W.I., Flower, D.R., Ellis, S.A., 2010. MHC classI bound to an immunodominant Theileria parva epitope demonstratesunconventional presentation to T cell receptors. PLoS Pathog. 6, e1001149.

Immunopathology 167 (2015) 80–85 85

MacHugh, N.D., Connelley, T., Graham, S.P., Pelle, R., Formisano, P., Taracha, E.L.,Ellis, S.A., McKeever, D.J., Burrells, A., Morrison, W.I., 2009. CD8+ T-cellresponses to Theileria parva are preferentially directed to a single dominantantigen: Implications for parasite strain-specific immunity. Eur. J. Immunol.39, 2459–2469.

McKeever, D.J., Taracha, E.L., Innes, E.L., MacHugh, N.D., Awino, E., Goddeeris, B.M.,Morrison, W.I., 1994. Adoptive transfer of immunity to Theileria parva in theCD8+ fraction of responding efferent lymph. Proc. Natl. Acad. Sci. U. S. A. 91,1959–1963.

Morrison, W.I., Taracha, E.L., McKeever, D.J., 1995. Theileriosis: progress towardsvaccine development through understanding immune responses to theparasite. Vet. Parasitol. 57, 177–187.

Mukhebi, A.W., 1992. Economic impact of theileriosis and its control in Africa. In:Mukhebi, A.W., Norval, R.A., Perry, B.D., Young, A.S. (Eds.), The Epidemiology ofTheileriosis in Africa. Academic Press, London, pp. 379–403.

Radley, D.E., Brown, C.G., Cunnigham, M.P., Kimber, C.D., Musisi, F.L., Purnell, R.E.,Stagg, S.M., 1975. East Coast fever: challenge of immunised cattle by prolongedexposure to infected ticks. Vet. Rec. 96, 525–527.

Steinaa, L., Saya, R., Awino, E., Toye, P., 2012. Cytotoxic T lymphocytes from cattleimmunized against Theileria parva exhibit pronounced cross-reactivity amongdifferent strain-specific epitopes of the Tp1 antigen. Vet. Immunol.Immunopathol. 145, 571–581.

Svitek, N., Hansen, A.M., Steinaa, L., Saya, R., Awino, E., Nielsen, M., Buus, S., Nene,V., 2014. Use of “one-pot, mix-and-read” peptide-MHC class I tetramers andpredictive algorithms to improve detection of cytotoxic T lymphocyteresponses in cattle. Vet. Res. 45, 50.

Svitek, N., Nzau, B., Steinaa, L., Nene, V., 2015. A method to discriminate betweenclosely related bovine major histocompatibility complex class I alleles bycombining established PCR–SSP assays with RFLPs. Tissue Antigens 4,278–282.

Sylvester-Hvid, C., Kristensen, N., Blicher, T., Ferre, H., Lauemoller, S.L., Wolf, X.A.,Lamberth, K., Nissen, M.H., Pedersen, L.O., Buus, S., 2002. Establishment of a

quantitative ELISA capable of determining peptide–MHC class I interaction.Tissue Antigens 59, 251–258.

Wendoh, J., Waihenya, R., Saya, R., Awino, E., Steinaa, L., 2014. Perforin expressionin Theileria parva specific T cells correlates with cytotoxicity. Open J. Immunol.4, 117–127.