Smallpox subunit vaccine produced in planta confersprotection in miceMaxim Golovkin*, Sergei Spitsin*, Vyacheslav Andrianov*, Yuriy Smirnov*, Yuhong Xiao†, Natalia Pogrebnyak*,Karen Markley*, Robert Brodzik*, Yuri Gleba‡, Stuart N. Isaacs†, and Hilary Koprowski*§

*Biotechnology Foundation Laboratories, Thomas Jefferson University, Philadelphia, PA 19107-6799; †Department of Medicine, University of PennsylvaniaSchool of Medicine, Philadelphia, PA 19104; and ‡Icon Genetics, Biozentrum Halle, Weinbergweg 22, D-06120 Halle (Saale), Germany

Contributed by Hilary Koprowski, March 2, 2007 (sent for review January 18, 2007)

We report here the in planta production of the recombinantvaccinia virus B5 antigenic domain (pB5), an attractive componentof a subunit vaccine against smallpox. The antigenic domain wasexpressed by using efficient transient and constitutive plant ex-pression systems and tested by various immunization routes in twoanimal models. Whereas oral administration in mice or the minipigwith collard-derived insoluble pB5 did not generate an anti-B5immune response, intranasal administration of soluble pB5 led toa rise of B5-specific immunoglobulins, and parenteral immuniza-tion led to a strong anti-B5 immune response in both mice and theminipig. Mice immunized i.m. with pB5 generated an antibodyresponse that reduced virus spread in vitro and conferred protec-tion from challenge with a lethal dose of vaccinia virus. Theseresults indicate the feasibility of producing safe and inexpensivesubunit vaccines by using plant production systems.

plant biotechnology � transgenic plants � B5 glycoprotein �recombinant antigen

P lants have emerged as an excellent alternative to otherexpression systems for the production of complex pharma-

ceutical proteins, including recombinant subunit vaccines (1–6).It was shown that some plant-derived antigens can inducesystemic and mucosal immune responses and, in some cases,confer protection against challenge (1–3). Plants provide theadditional advantage of direct delivery through oral or othermucosal routes (1, 6). Despite some difficulties with the expres-sion of certain recombinant proteins, especially those of viralorigin, plant biotechnology holds the promise of producingmedicinal proteins to be used in vaccine formulations.

Interest in a safe smallpox vaccine has been reawakened by thethreat of bioterrorism (7, 8) and continuous outbreaks of or-thopoxvirus diseases (9, 10). A live vaccinia virus (VV)-basedvaccine has been used to eventually eradicate smallpox disease(11, 12), but does display side effects (13). Although oneapproach for developing a safer vaccine is to use a highlyattenuated live virus, recombinant protein-based vaccines arelikely to be safer. For orthopoxviruses, there are several candi-date antigens that can protect mice and nonhuman primatesfrom lethal challenge (14–20). The extracellular virus (EV)-specific membrane glycoprotein encoded by the B5R gene(21–24) is the main target of EV-neutralizing antibodies presentin human-derived vaccinia gamma globulin used to treat com-plications arising from smallpox vaccination (25).

In this study, we demonstrate that the VV B5 protein can beproduced in two plant expression systems. The use of themagnifection transient expression system (26–28) enabled rapidhigh-yield production of soluble B5 as well as selection ofoptimal subcellular targeting signals to use in stable planttransformation. Preparations of purified plant-derived B5 anti-gen (pB5) induced a strong immune response when administeredparenterally and intranasally, and mice vaccinated i.m. with pB5were protected from VV challenge.

ResultsExpression Cassette Design and Cloning Strategy. The B5 glycopro-tein (42 kDa) encoded by the VV (strain WR) B5R gene (21) was

chosen for the production in planta of the full extracellularantigenic domain (amino acids 20–275) (Fig. 1A), which containsthe major neutralization epitopes (14, 29–30). Expression cas-settes containing the c-Myc and His6 tags from pIV vectors(ImpactVector, Wageningen, the Netherlands) (Fig. 1B) weresubcloned into the magnICON Icon Genetics’ plasmidpICH115999 (Fig. 1B) and used for empirical testing to deter-mine the best expression in planta in conjunction with helperplasmid(s) supplying various plant intracellular signals in trans(26–28) [Icon Genetics, Halle (Saale), Germany]. This resultrevealed the apoplast secretion signal to be superior. Thus, theapoplast-targeting cassette from the pIV-based vector under therbcS promoter was subcloned into the binary plant transforma-tion vector pBIN-Plus (ImpactVector) (Fig. 1B), generatingpBIRub-apopB5 for stable tobacco transformation. The fusion ofthe full extracellular antigenic domain with an intracellularmembrane anchor was used to generate transgenic collard plantsproducing insoluble pB5 in abundant vegetative biomass suitablefor oral feeding (31) (Fig. 2 B and C).

Production of B5 Protein in Plants. The magnifection system de-veloped by Icon Genetics provided a robust and rapid method(26–28) to express recombinant pB5 in wild-type plants (Fig.3A). The pB5 protein was readily detected at 6–8 days postin-fection in the leaf tissues transfected with the B5 and apoplast-targeting module (Fig. 3B). The pB5 antigen remained stable inlyophilized plant tissues for several months at �100 mg/kg levelwhen stored in air-tight containers at ambient temperatures or4°C. Powdered material was used in a two-step affinity purifi-cation (Ni- and c-Myc Mab affinity columns) to obtain stan-dardized pB5 samples at �50% purity (Fig. 3 B and C) suitablefor intranasal and parenteral immunization.

Stable transgenic tobacco plants expressing soluble B5 proteinwere generated (Fig. 2A) by using the pBIRub-apopB5. Severalindependent transgenic lines revealed the presence of pB5 byWestern blotting (Fig. 2 A Right Lower) and showed no morpho-logical abnormalities (Fig. 2 A Left).

Transgenic collard plants expressing insoluble pB5 were gen-erated by our method (31) by using the pB002-based binaryvector (32) carrying the pB5-anchor fusion expression cassette(Fig. 1B). Several transgenic lines revealed an antigen-specificband in leaf tissues by Western analysis (Fig. 2B Right). Theaccumulation of pB5 in collard was higher when compared withtransgenic tobacco expressing soluble pB5. The presence ofadditional amino acids at the C-terminal end (anchor) is prob-ably responsible for the differences in migration of soluble and

Author contributions: M.G., S.S., R.B., S.N.I., and H.K. designed research; M.G., S.S., V.A.,Y.S., Y.X., N.P., K.M., and S.N.I. performed research; M.G., Y.S., R.B., and Y.G. contributednew reagents/analytic tools; M.G., S.S., V.A., Y.S., Y.X., N.P., K.M., and S.N.I. analyzed data;and M.G., S.S., N.P., and S.N.I. wrote the paper.

The authors declare no conflict of interest.

Abbreviations: VV, vaccinia virus; CT, cholera toxin; TSP, total soluble protein.

§To whom correspondence should be addressed. E-mail: hilary.koprowski@jefferson.edu.

© 2007 by The National Academy of Sciences of the USA

6864–6869 � PNAS � April 17, 2007 � vol. 104 � no. 16 www.pnas.org�cgi�doi�10.1073�pnas.0701451104

Dow

nloa

ded

by g

uest

on

Mar

ch 6

, 202

1

insoluble forms of pB5 on SDS/PAGE and the B5-specificdouble-banding (Fig. 2B Right Lower). Transgenic collard plantsgrew large rosettes that were �30 cm in diameter (Fig. 2B Left)and weighed �1.5 kg. Transgenic collard leaves were directlyused for oral immunization in either fresh form or standardized1-g pellets (Fig. 2C).

Immunogenicity of pB5 Protein in Mice. Sera from mice immunizedorally, intranasally, or parenterally with pB5 were tested for thepresence of B5-specific antibodies by ELISA (Fig. 4). No im-mune response was detected in serum of mice fed with transgenic

collard plant material (fresh leaves, pellets) as well as withpurified pB5 administered by gavage (Fig. 4A). Supplementationof plant material with cholera toxin (CT) did not invoke a

A

B

Fig. 1. Vaccinia virus EV B5 glycoprotein: expression cassettes and cloning strategy. (A) Schematic diagram of the full-length B5 protein comprised of a signalpeptide (SP; amino acids 1–19), four modular short consensus repeat (SCR) domains (amino acids 20–237), stalk region (amino acids 238–275), putativetransmembrane domain (TM), and cytoplasmic tail (CT). Open triangles indicate neutralizing antibody epitopes; arrowhead indicates the B5-specific Mab site(B5 Mab) within the stalk region; pin heads indicate the three N-linked glycosylation sites within SCR2. (B) The B5 antigenic ectodomain (amino acids 20–275)was cloned into various expression cassettes with C-terminus-specific tags (c-Myc and/or His6), an intracellular targeting signal, such as apoplast (Apo), or thesubcellular outer membrane anchor. The pICH115999 provector was used for in planta transient transfection. For stable transformation, the pBIN-Plus (solublepB5) and pB002-based (insoluble pB5) vectors under the rbcS (black arrow) and CaMV-35S (gray arrow) promoters, respectively, were used. For bacterialexpression, the B5 cassettes were subcloned into the modified pGEX3 vector (pGEX3�GST).

A

B

C

Fig. 2. Constitutive expression of B5 in tobacco and collard. (A) Transgenictobacco expressing soluble pB5 grown to maturity (Left) and used for proteingel analysis (Right). Equal amounts of total soluble proteins (TSPs) wereloaded onto the SDS/PAGE (Right Upper) and analyzed by Western blotting byusing c-Myc Mab (Right Lower). Bands of expected size were detected intransgenic lines. Bacterial-expressed B5 was used as a positive control (�);nontransgenic TSP extract was used as a negative control (wt). (B) Transgeniccollard was grown to maturity (Left), and leaf tissue was collected and used forprotein analysis (Right). Equal amounts of TSP extracts were loaded onto anSDS/PAGE and stained (Right Upper) or used for Western blotting with B5-specific MAb (Right Lower). Band(s) of expected size (pB5; Right Lower) weredetected in several transgenic lines. (C) For oral feeding, plant material wasused in fresh or lyophilized form (Left); 1-g tablets (Right) containing astandardized quantity of recombinant (intracellular, membrane-bound, in-soluble) pB5 antigen were prepared in plastic molds (Center).

A

B

C

Fig. 3. Transient expression and purification of soluble pB5. (A) Six-week-oldN. benthamiana (Left) or Swiss chard (Right) plants were used for magnICON-based expression. (B) Major purification steps of soluble pB5 visualized bySDS/PAGE staining (Upper) and Western blot analysis with c-Myc Mab (Lower).pB5 in TSP extract detected as a single band (lane 1). Lanes 2–5 representconsecutive Ni-column elutions (Step I), and lanes 6–9 represent fractionseluted from the c-Myc-Mab-affinity column (Step II). Lane 1, TSP extract; lane2, extract washed with 5 mM imidazole; lane 3, extract washed with 20 mMimidazole; lane 4, elution in 80 mM imidazole; lane 5, consequent elution in120 mM imidazole; lane 6, dialyses of eluate for c-Myc-Mab-affinity columnapplication (c-Myc-column); lane 7, flow-through in PBS buffer; lane 8, elutionwith c-Myc peptide; lane 9, consequent elution in acid glycine buffer (pH 2.5);lane 10, marker. (C) Visual evaluation of concentrated recombinant pB5 fromseparate purified batches against a BSA titration curve.

Golovkin et al. PNAS � April 17, 2007 � vol. 104 � no. 16 � 6865

PLA

NT

BIO

LOG

Y

Dow

nloa

ded

by g

uest

on

Mar

ch 6

, 202

1

response, although it did result in detection of anti-CT responsespresent in sera (IgG) and fecal pellet extracts (IgA) (data notshown). Intranasal immunization with purified pB5 without anyadjuvant did not produce antibodies in mice, although whensupplemented with CT led to a distinct anti-B5 immune response(Fig. 4B). The strongest anti-B5 immune response was detectedin sera of mice immunized parenterally (Fig. 4C). Westernanalysis of protein extracts from plant- and bacterially expressedB5 (Fig. 4D) confirmed the B5 specificity of the mouse serumantibodies.

Immunogenicity of pB5 Protein in the Minipig. Analysis of theimmune response in the minipig revealed a response patternsimilar to that observed in mice (Fig. 5). After feeding, noanti-B5-specific antibodies were detected in the serum (Fig. 5A),saliva, or vaginal secretions (data not shown). However, afterintranasal immunization, together with CT the serum showed anincrease in B5-specific IgG (Fig. 5B). No B5-specific IgA waspresent in either the saliva or vaginal secretions, whereas a strong

anti-CT response was detected in both the serum (IgG) andsaliva (IgA). Furthermore, no anti-CT response was detected inthe fecal matter or vaginal washes (data not shown). As shownin Fig. 5C, a sustained increase in B5-specific serum IgG wasdetected after i.m. immunization with purified pB5 in combi-nation with CpG-ODN and alum (CpG/alum).

Protection of pB5-Vaccinated Mice Against Lethal Challenge with VV.The level of protection in immunized mice was analyzed for thepresence of B5-specific antibodies in sera by measuring the invitro functional anti-VV activity and challenging the mice with alethal dose of VV (Fig. 6). Mice were inoculated with VV onceby tail scarification or three times with plant-derived B5 (inCpG/alum) i.m.; control mice were injected with extracts of totalsoluble protein (in CpG/alum) or left uninoculated. Three weeksafter the third vaccination, sera were tested by ELISA (Fig. 6A)and in the comet inhibition assay (Fig. 6B) as a qualitativemeasure of anti-EV antibody activity. We note that the serumimmune response in mice was equally pronounced when usingeither CpG/alum or Freund’s as an adjuvant (compare Figs. 6Aand 4C). As shown in Fig. 6B, sera from pB5-vaccinated micealtered comet formation, whereas sera from VV-vaccinated mice

A

B

C

D

Fig. 4. Serum antibody response to pB5 in mice. IgG titers in sera of miceimmunized with pB5 by three different routes. Each group was immunized asdescribed in Materials and Methods. (A) Anti-B5 response after oral immuni-zation of mice fed three times with fresh leaves or pellets supplemented withCT. (B) B5-specific immune response after intranasal application of purifiedpB5 alone (1st and 2nd) and together with CT adjuvant (3rd and 4th). (C)Anti-B5 response after parenteral immunization with purified pB5 appliedwith Freund’s adjuvant. (D) Western blot analysis of sera from mice immu-nized parenterally (3rd i.p. samples in C). Sequential dilutions of TSP extractsfrom transgenic plants (left) and E.coli (right) expressing B5 probed with c-MycMab or sera from mice that received parenteral immunization with pB5.Mobility difference of B5 from plant versus E.coli, indicated by arrowheads, islikely due to glycosylation of pB5.

Fig. 5. Serum antibody response to pB5 protein in minipig. The minipig wasused for all immunizations as described in Materials and Methods. IgG titersin serum are given as the mean (� SD) of triplicate determinations. (A) Oralimmunization with transgenic plant material. (B) Intranasal immunizationwith purified pB5. (C) Intramuscular immunization with pB5 in CpG/alum. (D)Western blot analysis of sera from the minipig injected with pB5 (3rd injec-tion). Mobility difference of B5 from plant vs. E.coli is indicated by arrowheads.

6866 � www.pnas.org�cgi�doi�10.1073�pnas.0701451104 Golovkin et al.

Dow

nloa

ded

by g

uest

on

Mar

ch 6

, 202

1

inhibited comet formation completely (Left); control mouse seradid not alter comet formation (Right).

Vaccinated mice were then challenged intranasally with alethal dose of VV and monitored daily for weight loss andmorbidity (Fig. 6C). Only the mice vaccinated with either VV orpB5 survived the challenge (Fig. 6C Lower). Control naı̈ve-,CpG/alum-, and nontransgenic plant material-vaccinated miceall died, indicating that survival is B5-dependent. Although100% of the pB5-vaccinated mice survived, they did experiencea greater weight loss than the VV-vaccinated mice (Fig. 6CUpper). Maximal weight loss of �5% was observed on day-3postchallenge, whereas pB5-vaccinated mice underwent a lossclose to 30% on day-8 postchallenge.

DiscussionRecombinant protein production is a well established technologythat utilizes efficient strategies for the generation of subunitvaccines (39, 40, 41). Recently developed plant biotechnologyoffers additional advantages in production scale, economy,product safety, and ways of delivery (1–6). At present, the maingoal is to increase the overall expression yield of functionalplant-derived proteins, especially those of viral origin, whichsometimes seriously impair transgenic plant growth and devel-opment (42–44).

To overcome the impediment in the case of VV B5, we usedtwo plant expression systems. The transient system (magnifec-tion) was used to optimize expression cassette arrangements for

the antigen and expedite the production of B5 protein in asoluble form, which is more amenable for extraction and puri-fication (45). We note that, from several plant-specific targetingsignals used to produce a soluble form of pB5, the apoplast signalled to the highest expression of recombinant protein. Thisinformation aided in efficient production of soluble pB5 intransgenic plants. The use of C-terminal tags facilitated theprocess of protein purification of �50% purity. Although theexpression levels were lower than those obtained by magnifec-tion, our transgenic plants are readily available for inexpensiveup-scaled production.

An obvious advantage of producing recombinant medicinalproteins in plants is the possibility of oral administration (1, 6).Considering the inevitable protein degradation in the gastroin-testinal tract, the antigen must be present at higher levels so asto induce an adequate immune response. In addition, planttissues might serve to protect subunit vaccines against digestionin the gut. To test this hypothesis, we constitutively expressed B5in transgenic collard displaying a large vegetative biomass (31).We confirmed the plant retaining a normal morphology,whereas insoluble antigen was expressed at high levels suitablefor direct feeding experiments. We previously implemented thisstrategy to express rabies viral G protein in Arabidposis as afusion protein rendering it membrane-bound and facing thecytosol (M.G., N.P., S.S., K.M., H.K, unpublished data), thusensuring correct posttranslational modifications (especially gly-cosylation) and proper folding of the viral glycoproteins. Thegoal was to increase the likelihood of increased antigen expres-sion of immunologically functional quality.

Mice and the minipig fed with transgenic collard and CTexhibited no detectable pB5 immune response, although a clearCT-specific IgG and IgA response was observed. This responsewas induced in both animal models to a lower dose of CT thanthe overall amount of pB5 in the feed. Moreover, no B5-specificantibody response was detected in mice immunized with solubleand purified pB5 plus CT by gavage. Because there was noresponse to orally administered pB5, whereas there was aresponse to CT, it is plausible that an antigenic protein able toinduce an immune response after oral administration mustnaturally be taken up through the oral route (6). Intranasaladministration of a soluble plant-derived antigen, together withCT, led to a steady increase in antibody titers after eachimmunization in both mice and the minipig. The titers in theminipig were lower than in mice. Perhaps this outcome is due toa suboptimal dose of pB5 and/or the use of nonoptimal adju-vants. Intranasal administration of antigen leads to the appear-ance of IgA in stool, saliva, and vaginal secretions (46); however,we did not detect any. Nevertheless, a CT-specific IgA waspresent in the saliva of the immunized minipig and not in stoolor vaginal samples.

The highest serum antibody response was found to be inparenterally immunized animals together with various adjuvants.Intramuscular vaccination of mice with purified pB5 in CpG/alum generated an antibody response that showed in vitro and invivo activity against VV. Sera from vaccinated animals, but notfrom control groups, were able to alter virus spread in thecomet-inhibition assay. The greater comet-inhibition activity ofsera from vaccinia-vaccinated mice, compared with that frompB5-vaccinated mice, likely reflects the reaction to multiple EVtargets after VV vaccination. The pB5-immunized mice werefound to be protected from lethal challenge with VV, and allmice survived the challenge. However, these mice experienced agreater weight loss compared with the VV-vaccinated mice. Thisresult was expected given that optimal protection to the chal-lenge is provided by a multicomponent vaccine (17, 18). There-fore, several potent antigens are under further investigation forthe development of a highly efficient, multicomponent subunitsmallpox vaccine made entirely in planta.

A

B

C

Fig. 6. pB5-specific antibody response protects from lethal challenge. (A) IgGtiters detected in sera from mice immunized with live VV or with pB5 inCpg/alum (Left). Control groups were injected with adjuvant alone (CpG/alum) or plant TSP extracts, or left unvaccinated (Right). (B) Characteristiccomets that form with VV are altered by the B5 antibodies present in sera frommice immunized with pB5 (Left); comet formation was not altered in negativecontrol groups: adjuvant only (CpG/alum), TSP vaccinated, unvaccinated(Right). Comet formation was completely inhibited by sera vaccinated with VV(far left). (C) Survival (Lower) and weight loss (Upper) after lethal intranasalchallenge with VV 3 weeks after the last i.m. vaccination (or 7 weeks after thesingle vaccinia vaccination by tail scarification). Mice were challenged intra-nasally with live VV and monitored for weight loss. Those with a �30% initialweight loss (or morbid) were killed. Symbols and colors for groups are thesame as in A.

Golovkin et al. PNAS � April 17, 2007 � vol. 104 � no. 16 � 6867

PLA

NT

BIO

LOG

Y

Dow

nloa

ded

by g

uest

on

Mar

ch 6

, 202

1

In conclusion, the efficient production of the B5 antigenicdomain in planta was analyzed for its ability to induce protectiveimmunity. The antigen was produced in soluble and insolubleforms upon transient and stable plant transformation. Thesoluble pB5 was purified to be used for immunization by gavage,intranasal, and parenteral routes. Transgenic collard with largeedible biomass was primarily used for direct oral administration.A B5-specific response was detected after intranasal immuniza-tion, whereas the strongest response was observed after paren-teral administrations. Moreover, i.m.-immunized mice gener-ated neutralizing antibodies that inhibited the spread of VV invitro and provided 100% protection from death after lethalchallenge.

Our study presents a major step toward the efficient produc-tion of viral proteins in plants that are functionally potent andready for use in subunit vaccine formulations to counter infec-tious diseases such as smallpox.

Materials and MethodsGeneration of Expression Cassettes and Transformation Vectors. Plas-mid pSI-80–10 (24) carrying a fragment of the VV (strain WR)genome was used to amplify the full antigenic ectodomainportion of B5 (amino acids 20–275) by using the followingprimers: 5�-CCA TGG ATT GTA CTG TAC CCA CTA TGAATA ACG-3� and 5�-GCG GCC GCA TGA TAA GTT GCTTCT AAC GAT TCT A-3�. The PCR fragment was cloned(NcoI-NotI) into a group of five intermediate pIV1 (1–5) vectors,thus fusing the gene to targeting signals and c-Myc tags and6x-His6 at C terminus (ImpactVector). As determined by transientin planta expression, the best-expression apoplast-targeting(33) cassette was subcloned (XbaI-SacI) into the stable planttransformation binary vector pBIN-Plus (ImpactVector) withkanamycin selection to generate pBIRub-apopB5.

The pB5 NcoI-SacI fragments from the corresponding pIV-based plasmids were subcloned into the carrier provector plas-mid pICH11599 (magnICON) and used for vacuum infiltrationof wild-type plants.

An insoluble (membrane-bound) form of pB5 protein wasproduced in transgenic collard plants by fusing the B5 antigenicregion amplified by two consecutive rounds of PCR with prim-ers: F, 5�-CTT TCA AAT ACT TCC ACC ATG GGA TGT ACTGTA CCC ACT ATG AAT AAC G-3�; R, 5�-AAG ATC CTCCTC GCT AAT AAG CTT TTG ATG ATA AGT TGC TTCTAA CGA TTC TAT TTC-3�; F, 5�-GCT CTA GAC GTT TTTATT TTT AAT TTT CTT TCA AAT ACT TCC ACC ATGGGA-3�; and R, 5�-ATG ATA AGT TGC TTC TAA CGA TTCTAT TTC-3�. The fragment was cloned into a binary vectorpB002 (32) expression cassette driven by CaMV-35S promoterand harboring a subcellular anchor.

For bacterial expression, the XbaI-SacI DNA fragments ex-cised from the pIV-based constructs were subcloned into thecorresponding sites of the modified pGEX3 vector (AmershamPharmacia Biotech, Piscataway, NJ) in which the GST tag wasdeleted (pGEX3��GST). Resulting constructs were transformedinto the Escherichia coli Rosetta-2 (DE3) strain (Novagen,Madison, WI) for expression.

Agrobacterium Infiltration of Plants Using the Magnifection System.For rapid transient expression of recombinant B5 in planta, weused the so-called ‘‘magnifection’’ procedure [Icon Genetics,Halle (Saale), Germany] as described (26–28), with minormodifications. GV3101 agrocultures carrying the B5 cassettes inpIV plasmids were each combined with equal volumes of culturescarrying the premanufactured helper plasmids supplying thesubcellular targeting signals and recombinase–integrase com-plex. The mixture was diluted (1:50) in buffer [10 mM MgSO4,10 mM Mes (pH 5.6), 5% sucrose, 200 �M acetosyringon, and0.02% Silwet L-77] and applied to mature Nicotiana benthami-

ana (Tobacco) and Beta vulgaris var. cicla (Swiss chard) plants(6–8 weeks old). Plant tissues were harvested within 7–10 daysand analyzed by Western blot and ELISA.

Generation of Transgenic Plants. Tobacco (Nicotiana tabacum cvWisconsin) leaf discs were transformed (34) with pBIRub-apopB5, and Collard (Brassica oleracea cv. Morris Heading)explants were transformed by our method (31) using the pB002-based construct. Kanymycin (tobacco)- and phosphinotricin(collard)-resistant transgenic lines were selected and tested byPCR and Western blotting for the presence of pB5.

pB5 Protein Expression Analysis. Harvested and lyophilized planttissues were homogenized and ground to a fine powder in thepresence of liquid nitrogen and/or dry ice in extraction buffer [50mM sodium phosphate (pH 8.0), 0.3M NaCl, 0.2% Tween-20, 2.5mM �-mercaptoethanol] supplemented with complete proteaseinhibitor (Sigma–Aldrich, St. Louis, MO). For SDS/PAGE,extracts of total soluble protein were loaded at �50 �g per lane.In Western analysis, proteins were detected by using MYC1–9E10.2 (IgG1) monoclonal antibody (c-Myc mab) (ATCC,Manassas, VA) and/or B5-specific Mab 206C5-F12 (35), ormouse/minipig pB5-specific serum antibodies.

Isolation and Purification of pB5 Recombinant Protein for Immuniza-tion. Lyophilized and powdered leaf tissue from magnifection-transfected N. benthamiana, Swiss chard, and transgenic N.tabacum plants was homogenized in extraction buffer. Aftercentrifugation, the supernatant was filtered through Miracloth(IMD Biosciences, San Diego, CA) and loaded on a Ni-Sepharose column (GE Healthcare, Piscataway, NJ). Eluate wasdialyzed against PBS with 0.1% Tween 20 and applied on a c-Mycmab-Sepharose column to be eluted with 0.1M glycine-HCl (pH2.5). Batches were concentrated, dialyzed against PBS � 0.1%Tween-20, and desalted on Sephadex-G50 spin columns (Roche,Nutley, NJ). Samples were brought to a concentration of 0.1�g/�l of pB5 with sterile PBS buffer and frozen in liquid nitrogenas 10-�g aliquots to be stored at �80°C.

Preparation of Plant Material for Oral Immunization. Fresh collardleaves were harvested from healthy transgenic plants. Collard-based pellets were prepared by mixing lyophilized and powderedplant tissue in PBS with 0.25% (wt/vol) of cornstarch. Pelletswere formed by using plastic moulds, lyophilized, and stored at4°C in air-tight containers.

Immunological Assessment of pB5 in Mice. Six- to 8-week-old femaleBALB/c mice (5 or 10 mice per group) were used in allexperiments. For oral immunization, each mouse was fed with2–3 g of fresh tissue or 1 g of pellets (�100 �g of antigen) overa period of 6–8 h. Control mice received wild-type plantmaterial. In some groups plant material was supplemented with10 �g of CT (IMD Biosciences) as an adjuvant. An additionalgroup of mice received purified pB5 (2 �g) with CT by gavage.For intranasal immunization, 2 �g of pB5 was administered in 10�l of saline into both nostrils; in some groups, pB5 was supple-mented with CT (1 �g). Mice were immunized three times at2-week intervals. For parenteral immunization, mice were in-jected three times at 2-week intervals with 2 �g of purified pB5.First and second immunizations were given s.c. with completeand incomplete Freund’s adjuvant, respectively (Difco, Detroit,MI). The third dose was administered i.p. in saline. Blood andfecal matter were collected 10 days after each immunization.Proteins from fecal pellets were extracted in PBS (10 vol/wt)supplemented with 1% BSA and protease inhibitors as described(36). Mice were killed 10 days after the last immunization andbled by cardiac puncture, and sera were analyzed by ELISA andWestern blotting.

6868 � www.pnas.org�cgi�doi�10.1073�pnas.0701451104 Golovkin et al.

Dow

nloa

ded

by g

uest

on

Mar

ch 6

, 202

1

Immunizations for the challenge experiments were done byusing CpG oligodeoxynucleotides (ODN) (sequence no. 1826:TCCATGACGTTCCTGACGTT; Coley PharmaceuticalGroup, Wellesley, MA) and alum (Alhydrogel; Accurate Chem-ical, Westbury, NY) as an adjuvant (CpG/alum) at 50 �g permouse (38). Mice received three i.m. vaccinations at 2-weekintervals with 8 �g of purified pB5, 8 �g of wild-type plant totalsoluble protein, or CpG/alum alone. Additional control groupsincluded naive, nonvaccinated mice and mice that were vacci-nated with VV by tail scarification. Mice were bled before thethird boost and 3 weeks later immediately before intranasalchallenge with VV.

Immunological Assessment of pB5 in Minipig. A single 8-month-oldminipig was immunized consecutively by oral, intranasal, andi.m. routes. Oral immunization consisted of two feedings with300 g of transgenic collard leaves supplemented with 15 �g of CTwithin 14 days. After 2 weeks, the minipig received threeintranasal doses of pB5 (15 �g) together with 2 �g of CT (75 �lper nostril in saline) at 2-week intervals. After an additional 3weeks, the minipig received three i.m. doses of 10 �g of purifiedpB5 supplemented with 50 �g of CpG/alum. Blood, fecal matter,saliva, and vaginal secretion samples were obtained 10 days aftereach immunization and analyzed by ELISA and Westernblotting.

Solid-Phase ELISA. The assay (37) was performed in 96-wellMaxiSorp plates (Nalgen Nunc, Rochester, NY) coated over-night at 4°C with E. coli-purified B5 at 1 �g/ml in PBS.Antigen-specific antibodies were detected by using the followingantibodies: goat anti-mouse IgG (BD Biosciences, San Jose, CA)

and anti-mouse IgA (Sigma–Aldrich) or goat anti-pig IgG/IgA(Bethyl Laboratory, Montgomery, TX). Results are given asmean � SD.

Comet Inhibition Assay. The in vitro assay was essentially per-formed as described in ref. 30. Briefly, a confluent monolayer ofBSC-1 cells grown in 12-well plates was infected with VV strainIHDJ at �5 pfu per well for 2 h at 37°C. The inoculum wasremoved and 1 ml of fresh MEM medium containing 2.5% FCSand 25 �l (1:40 dilution) of pooled sera obtained from mice3 weeks after the last boost vaccination (just before chal-lenge) were added. Plates were incubated for �36 h at 37°C andstained with 0.2% crystal violet in 4% ethanol, and wells werephotographed.

VV Challenge. Three weeks after the last protein vaccination, micewere intranasally challenged with 1.2 � 106 pfu of VV strain WRin 20 �l of sterile PBS. Mice were weighed and monitored eachday. Animals that appeared morbid or had lost �30% of theirinitial body weight were killed in accordance with the institu-tional guidelines for animal welfare (38).

We thank G. Golovin and R. Egolf for technical help and greenhousework; Dr. P. Kozlowski and Dr. M. Neutra for advice on mucosaladjuvants; Dr. T. Kohl for helpful comments on the manuscript; andThomas Jefferson University and Kimmel Cancer Center research andanimal facilities for their support. This work was supported by a grantfrom Commonwealth of Pennsylvania Department of Health to Bio-technology Foundation Laboratories (to H.K.) and National Institutes ofHealth/National Institute of Allergy and Infection Diseases MiddleAtlantic Regional Center of Excellence Grant U54 AI057168 (to Y.X.and S.N.I.).

1. Goldstein DA, Thomas JA (2004) Q J Med 97:705–716.2. Fox JL (2006) Nat Biotechnol 24:1191–1193.3. Ma JK-C, Chikwamba R, Sparrow P, Fischer R, Mahoney R, Twyman RM

(2005) Trends Plant Sci 10:580–585.4. Gleba Y, Marillonnet S, Klimyuk V (2004) Biotechnol Genet Eng Rev 21:325–

367.5. Koprowski H (2005) Vaccine 23:1757–1763.6. Streatfield SJ (2006) Methods 38:150–157.7. Henderson DA (1999) Science 283:1279–1282.8. Enserink M (2004) Science 304:809.9. Reed KD, Melski JW, Graham MB, Regnery RL, Sotir MJ, Wegner MV,

Kazmierczak JJ, Stratman EJ, Li Y, Fairley JA, et al. (2004) N Engl J Med350:342–350.

10. Nalca A, Rimoin AW, Bavari S, Whitehouse CA (2005) Clin Infect Dis41:1765–1771.

11. Fenner F, Henderson DA, Arita I, Jezek Z, Ladnyi ID (1988) World HealthOrganization (WHO, Geneva).

12. Hammarlund E, Lewis MW, Carter SV, Amanna I, Hansen SG, Strelow LI,Wong SW, Yoshihara P, Hanifin JM, Slifka MK (2005) Nat Med 11:1005–1008.

13. Fulginiti VA, Papier A, Lane M, Neff JM, Henderson DA (2003) Clin InfectDis 37:241–250.

14. Galmiche MC, Goenada J, Rindisbacher L (1999) Virology 254:71–80.15. Hooper JW, Custer DM, Schmaljohn CS, Schmaljohn AL (2000) Virology

266:329–339.16. Hooper JW, Custer DM, Thompson E (2003) Virology 306:181–195.17. Hooper JW, Thompson E, Wilhelmsen C, Zimmerman M, Ichou MA, Steffen

SE, Schmaljohn CS, Schmaljohn AL, Jahrling PB (2004) J Virol 78:4433–4443.18. Fogg C, Lustig S, Whitbeck JC, Eisenberg RJ, Cohen GH, Moss B (2004) J Virol

78:10230–10237.19. Heraud JM, Smith YE, Ayala V, Kalisz I, Parrino J, Kalyanaraman VS,

Manischewitz J, King LR, Hryniewicz A, Trindade CJ, et al. (2006) J Immunol177:2552–2564.

20. Fang M, Cheng H, Dai Z, Bu Z, Sigal LJ (2006) Virology 345:231–243.21. Isaacs SN, Wolffe EJ, Payne LG, Moss B (1992) J Virol 66:7217–7224.22. Engelstad M, Howard ST, Smith GL (1992) Virology 188:801–810.23. Engelstad M, Smith GL (1993) Virology 194:627–637.24. Wolffe EJ, Isaacs SN, Moss BJ (1993) J Virol 67:4732–4741.

25. Bell E, Shamim M, Whitbeck JC, Sfyroera G, Lambris JD, Isaacs SN (2004)Virology 325:425–431.

26. Marillonnet S, Giritch A, Gils M, Kandzia R, Klimyuk V, Gleba Y (2004) ProcNatl Acad Sci USA 101:6852–6857.

27. Marillonnet S, Thoeringer C, Kandzia R, Klimyuk V, Gleba Y (2005) NatBiotechnol 23:718–723.

28. Gleba Y, Klimyuk V, Marillonet S (2005) Vaccine 23:2042–2048.29. Law M, Smith G (2001) Virology 280:132–142.30. Aldaz-Carroll L, Whitbeck JC, Ponce de Leon M, Lou H, Hirao L, Isaacs SN,

Moss B, Eisenberg RJ, Cohen GH (2005) J Virol 79:6260–6271.31. Pogrebnyak N, Markley K, Smirnov Y, Brodzik Y, Bandurska K, Koprowski H,

Golovkin M (2006) Plant Sci 171:677–685.32. Golovkin M, Reddy ASN (2003) Proc Natl Acad Sci USA 100:10558–10663.33. Outchkourov NS, Peters J, Jong J, Rademakers W, Jongsma MA (2003) Planta

216:1003–1012.34. Horsch RB, Klee HJ, Stachel S (1986) Proc Natl Acad Sci USA 83:2571–2575.35. Izmailyan RA, Huang CY, Mohammad S, Isaacs SN, Chang W (2006) J Virol

80:8402–8410.36. Mantis NJ, Kozlowski PA, Mielcarz DW, Weissenhorn W, Neutra MR (2001)

Vaccine 19:3990–4001.37. Pogrebnyak N, Golovkin M, Andrianov V, Spitsin S, Smirnov Y, Egolf R,

Koprowski H (2005) Proc Natl Acad Sci USA 102:9062–9067.38. Xiao Y, Aldaz-Carroll L, Ortiz AM, Whitbeck JC, Alexander E, Lou H, Davis

HL, Braciale TJ, Eisenberg RJ, Cohen GH, Isaacs SN (2006) Vaccine 25:1214–1224.

39. Ulmer JB, Valley U, Rappuoli R (2006) Nat Biotechnol 24:1377–1383.40. Hansson M, Nygren, P-A, Stahl S (2000) Biotechnol Appl Biochem 32:95–107.41. Mitragotri S (2005) Nat Rev Immunol 5:905–916.42. Bohmert K, Balbo I, Steinbuchel A, Tishendorf G, Willmitzer L (2002) Plant

Physiol 128:1282–1290.43. Wu H, Singh NK, Locy RD, Scissum-Gunn K, Giambrone JJ (2004) Biotechnol

Lett 26:787–792.44. Cheon BY, Kim HJ, Oh KH, Bahn SC, Ahn JH, Choi JW, Ok SH, Bae JM, Shin

JS (2004) Transgenic Res 13:541–549.45. Santi L, Giritch A, Roy CJ, Marillonnet S, Klimyuk V, Gleba Y, Webb R,

Arntzen CJ, Mason HS (2006) Proc Natl Acad Sci USA 103:861–866.46. Koslowski PA, Williams SB, Lynch RM, Flangigan TP, Patterson RR, Cuvin

S, Neutra MR (2002) J Immunol 169:566–574.

Golovkin et al. PNAS � April 17, 2007 � vol. 104 � no. 16 � 6869

PLA

NT

BIO

LOG

Y

Dow

nloa

ded

by g

uest

on

Mar

ch 6

, 202

1