key: cord-290883-r2744fb3
authors: TORRES, JUAN M.; SÁNCHEZ, CARLOS; SUÑÉ, CARLOS; SMERDOU, CRISTIAN; PREVEC, LUDVIK; GRAHAM, FRANK; ENJUANES, LUIS
title: Induction of Antibodies Protecting against Transmissible Gastroenteritis Coronavirus (TGEV) by Recombinant Adenovirus Expressing TGEV Spike Protein
date: 1995-11-30
journal: Virology
DOI: 10.1006/viro.1995.0023
sha: 
doc_id: 290883
cord_uid: r2744fb3

Abstract Ten recombinant adenoviruses expressing either fragments of 1135, 1587, or 3329 nt or the full-length spike gene of transmissible gastroenteritis coronavirus (TGEV) have been constructed. These recombinants produce S polypeptides with apparent molecular masses of 68, 86, 135, and 200 kDa, respectively. Expression of the recombinant antigen driven by Ad5 promoters was inhibited by the insertion of an exogenous SV-40 promoter. Most of the recombinant antigens remain intracytoplasmic in infected cells. All the recombinant-directed expression products contain functional antigenic sites C and B (Gebaueret al.,1991,Virology183, 225–238). The recombinant antigen of 135 kDa and that of 200 kDa, which represents the whole spike protein, also contain antigenic sites D and A, which have previously been shown to be the major inducers of TGEV-neutralizing antibodies. Interestingly, here we show that recombinant S protein fragments expressing only sites C and B also induced TGEV-neutralizing antibodies. The chimeric Ad5–TGEV recombinants elicited lactogenic immunity in hamsters, including the production of TGEV-neutralizing antibodies. The antisera induced in swine by the Ad5 recombinants expressing the amino-terminal 26% of the spike protein (containing sites C and B) or the full-length spike protein, when mixed with a lethal dose of virus prior to administration to susceptible piglets, delayed or completely prevented the induction of symptoms of disease, respectively.

immune response to coronaviruses Enjuanes and Van der Zeijst, 1995) : the spike protein (S) Transmissible gastroenteritis coronavirus (TGEV) in- (Buchmeier et al., 1984; Cavanagh et al., 1986 ; Daniel et fects the enteric and respiratory tissues of newborn pigal., 1993; Daniel and Talbot, 1990; Koolen et al., 1990) , lets resulting in mortality of nearly 100% (Saif and Wesley, the membrane protein (Fleming et al., 1989; . Protection of newborn animals from TGEV infecal., 1992; Welch and Saif, 1988) , and the nucleoprotein tion requires the induction of secretory IgA in milk. Previ-(Buchmeier et al., 1984; ; Lecomte et ous studies have shown that precursors of mucosal IgA al., 1987; Nakanaga et al., 1986; Talbot et al., 1984;  Wesplasma cells originate in lymphoepithelial structures in seling et al., 1993) . The study of the induction of protecthe gastrointestinal and respiratory tracts. These precurtive immunity to TGEV has focused on S protein because sor cells switch to IgA production in gut-or bronchusit is the major inducer of TGEV-neutralizing antibodies associated lymphoepithelial tissues and migrate to dis- Jimé nez et al., 1986; Laude et al., seminated mucosal effector sites, including gastrointesti-1992) and it mediates binding of TGEV to its cellular nal and upper respiratory tracts, as well as to exocrine receptor Godet et al., 1994) . A correlatissues such as the mammary gland. Recombinant hution between the antigenic and the physical structure of man adenovirus 5 (Ad5) has efficiently been used to in-S protein has been established ; duce protection against viral infections (Berkner, 1988; Jimé nez et al., 1986; Suñé et al., 1990) . Site A is also Graham and Prevec, 1992) . We have reported that Ad5 involved in the induction of in vivo protection (De Diego infects mucosal tissues of swine (Torres et al., 1995 (Torres et al., ), et al., 1992 , but the precise roles of the different antiindicating that recombinant adenoviruses might be used genic sites in eliciting resistance to TGEV are unknown to induce mucosal immunity against TGEV. Helper-inde- (Enjuanes and Van der Zeijst, 1995) . pendent Ad5-based vectors with the capacity to express

In this paper we describe 10 Ad5-TGEV recombinants foreign genes of up to 4.9 kb have been developed (Bett expressing either full-length TGEV spike protein or three et al., 1993) . truncated amino-terminal fragments of this protein.

These recombinants induced immune responses in hamsters and swine which neutralized TGEV infectivity. In addition, we demonstrate that porcine serum from Ad-tion. Finally, we show that virus-neutralizing antibodies tains the 3-end of Ad5 from the XhoI site at 70 map units (m.u.) with a deletion of the XbaI D fragment from are induced in the milk of Ad-TGEV-immune hamsters.

78.5 to 84.3 m.u. within the Ad5 E3 coding region. Plasmid pAB14 also contains the 3-end of Ad5 from map unit 70 MATERIALS AND METHODS to 100 with a 2685-nucleotide deletion in the E3 coding Eukaryotic cells and viruses region. Plasmid pFG173 contains a deletion of essential sequences to the left of E3 in the Ad genome that renders The epithelial swine testicle (ST) cell line (McClurkin it unable to produce infectious Ad5 (Bett et al., 1993; and Norman, 1966) and human 293 cells which constitu-Hanke et al., 1990; Mittal et al., 1993) . tively express the 5-end 11% of the Ad5 genome (Graham et al., 1977) were used to grow the recombinant Construction of recombinant vectors adenoviruses. PUR46-MAD strain of TGEV (Sá nchez et al., 1990) was cloned, sequenced, and used as a source

The general procedure followed to construct recombiof the S gene . Neutralization of nant Ad5 viruses expressing TGEV S gene fragments TGEV was performed by incubating serial 10-fold dilu-(Ad-TS) is summarized in Fig. 1 . S gene sequences were tions of the virus with a 1/20 dilution of the antibody at flanked by SV-40 Pr and polyadenylation sequences 37Њ for 30 min, and the virus-antibody mixture was plated when indicated (Fig. 2) , by subcloning them into plasmid on ST cells as previously described (Correa et al., 1988) . pSV2X3 or pSV2X4. Cassettes with S gene sequences The neutralization index (NI) was defined as the log 10 of were inserted into the unique XbaI site of the partially the ratio of the PFU after incubating the virus in the deleted E3 gene on plasmid pFG144K3 or pAB14, both presence of medium or the indicated antiserum. NI indiof which include the 3-end of Ad5. Alternatively, S gene ces are determined rather than titers since in the first fragments were removed from the original plasmid or procedure virus-antibody mixtures are evaluated in the from pSV2X3-TS vectors without SV-40 Pr signal, or withplaque assay without further dilution of the antibody, proout both Pr and polyadenylation sequences, using the viding highly reproducible results and information about restriction endonucleases indicated in Fig. 1 . In this case, the potency of the antibody (the titer reduction expressed fragment ends were blunted with Klenow and T4 DNA in logarithmic units rather than the ability of the serum polymerase and cloned into the XbaI site of pFG144K3 to neutralize a few PFU).

or pAB14 plasmids that were blunted and dephosphory-Ad5 strain dl309 contains a small deletion from 83 to lated according to standard procedures (Maniatis et al., 85 map units and an unknown substitution in the E3 1989) . Each of these plasmids is noninfectious by itself, region (Jones and Shenk, 1979) . pFG140 is an infectious but can generate infectious virus following cotransfection circularized form of Ad5 dl309 carrying a 2.2-kb DNA of 293 cells along with a plasmid, pFG173, which coninsert (pMX2) encoding ampicillin resistance (Apr) and a tains the 5-end of Ad5 ( Fig. 1 ) (Graham and Prevec, bacterial origin of replication. Plasmid pFG140 was used 1992; Hitt et al., 1995 Hitt et al., , 1994 . This results in the rescue as positive control for infectious Ad5 DNA (Graham et of genes cloned into the E3 region of viral vectors. Coal., 1988) .

transfection was performed essentially as described using the calcium phosphate precipitate technique (Gra-Plasmids and bacteria ham and van der Eb, 1973) . After 8 to 15 days, plaques were isolated and expanded, and viral DNA was ana-The TGEV S gene was cloned into Bluescript (Stralyzed by HindIII restriction enzyme digestion. Viruses tagene) or pYA plasmids (Smerdou et al., 1995) as prewith the expected DNA pattern were plaque purified viously described . Escherichia coli three times and the junction of the constructs was se-DH5 or XL1-blue cells (Stratagene) were transformed quenced to verify the expected primary structure. Recomwith newly constructed plasmids by electroporation binants Ad-TS01 and Ad-TS02 are identical to recombi- (Dower et al., 1988) . Plasmid DNA was prepared by the nants Ad-TS5 and Ad-TS6, respectively, except that the alkaline lysis method (Birnboim and Doly, 1979) and purifirst two were constructed using cloning vector pAB14 fied by CsCl-ethidium bromide density gradient centrifuwith the large deletion on E3 gene, while in the construcgation. S gene fragments or the full-length S gene were tion of the second pair of recombinants plasmid flanked either by SV-40 promoter (Pr) alone or by both pFG144K3, with the smaller deletion on E3, was used. Pr and polyadenylation sequences, as indicated. S gene fragments were first subcloned into pSV2X3 or pSV2X4

Immunoprecipitation of S antigens expressed by plasmids . The structures of the three recombinant Ad-TS key plasmids (pFG144K3, pAB14, and pFG173) used in the construction of Ad5-TGEV recombinants have been Subconfluent 293 cells grown in Dulbecco's modified Eagle medium with 5% horse serum (Gibco Europe) were reported previously (Bett et al., 1993; Mittal et al., 1993) . Plasmid pFG144K3 was derived from pFG144 (Ghosh-infected with Ad-TS recombinants at a multiplicity of infection (m.o.i.) of 30 PFU per cell. After 1 hr of virus Choudhury et al., 1986) and as essential features con- (Maniatis et al., 1989) . S gene sequences previously cloned into Bluescript(SK 0 ) (Promega) or pYA (Smerdou et al., 1995) plasmids were excised using the indicated restriction endonucleases and subcloned into pSV2X3 or pSV2X4, in which the S gene sequences were flanked by SV-40 Pr, polyadenylation sequences, or both. To generate recombinants Ad-TS07, Ad-TS05, Ad-TS9, and Ad-TS06 S gene sequences were cloned directly into plasmid pFG144K3 or pAB14. S gene sequences either alone or flanked by SV-40 sequences were subcloned into the XbaI site of pFG144K3 or pAB14, or excised with the indicated restriction endonucleases, blunted using the Klenow polymerase fragment, and cloned into blunted XbaI unique site of these vectors. Infectious Ad-TS recombinants expressing S protein fragments were generated by cotransfecting 293 cells with pFG144K3-TS or pAB14-TS (which carry S gene sequences from TGEV and pFG173 plasmids). Diagrams are not to scale. The origins of DNA fragments flanking the S gene are indicated with squares filled with different motifs. Numbers below the bar representing the Ad5 genome (bottom) indicate map units. mcs, multicloning site; Pr, promoter; An, polyadenylation signal; DE3, deletion in E3 gene; R.E., restriction endonuclease; TS refers to sequences derived from TGEV spike gene. adsorption at 37Њ, fresh medium was added and cells dium, and refed with fresh medium containing 50 mCi/ml of Pro-Mix: L-[ 35 S] in vitro methionine/cysteine labeling were incubated for 22 hr at 37Њ. Medium was then replaced by methionine-and cysteine-free medium con-mix (1 Ci/mmol, Cod. No. SJQ0079, Amersham Ibé rica). Cell monolayers were incubated 1.5 hr, detached with a taining 2% dialyzed serum. Cells were incubated for 1 hr at 37Њ, washed with methionine-and cysteine-free me-rubber policeman, washed with cold phosphate-buffered saline, pH 7.2 (PBS), collected by centrifugation at 3000 standard) and Ad-TS recombinants grown under the same conditions were immunoprecipitated in parallel. rpm for 15 min at 4Њ in a microfuge, and lysed in RIPA buffer (50 mM Tris-HCl buffer, pH 7.5, 150 mM NaCl, 1%

The same number of infected cells was analyzed for each recombinant. Similar relative expression levels were ob-Triton X-100, 1% sodium dodecyl sulfate (SDS), and 0.2 mM PMSF). Viscosity was reduced by mixing the tubes tained in many (ú5) experiments. After protein resolution in polyacrylamide gel electrophoresis and autoradiogra-in a Vortex mixer and passing the samples through a 0.6-mm needle 10 times. Extracts were centrifuged at phy, the intensity of the immunoprecipitated bands from Ad-TS extracts was compared with that of the reference 30,000 g for 30 min at 4Њ in a microfuge. Labeled proteins were immunoprecipitated with TGEV-specific porcine se-[ 35 S]TGEV with known protein concentration (determined using BCA Protein Assay Reagent, Pierce) to estimate rum which had been preadsorbed several times with 293 cells infected with adenovirus Ad5 dl309. Further the amount of S antigen. absorption of the antiserum did not eliminate the nonspecific bands. Antigen-antibody complexes were bound Immunofluorescence to protein A-Sepharose by overnight incubation at 4Њ. Sepharose beads were washed three times with RIPA ST cells at a density of approximately 1.5 1 10 5 cells/ cm 2 in microslide culture chambers (Miles Scientific) buffer containing 0.2% SDS, and the final pellet was resuspended in electrophoresis sample buffer containing were infected with adenovirus Ad140 which contains no S gene insert, or with Ad-TS recombinants, at a m.o.i. 2.5% SDS and 5% 2-mercaptoethanol (Laemmli, 1970) . Samples were boiled for 3 min, the beads were sedi-of 3 PFU/cell. At 24 hr postinfection, cell monolayers were washed and fixed either with methanol:acetone (1:1) at mented by low-speed centrifugation, and supernatants were analyzed by polyacrylamide gel electrophoresis 020Њ for 15 min or with 4% paraformaldehyde in PBS for 20 min at room temperature. Cells were washed three and autoradiography. To estimate the amount of protein expressed by each recombinant different dilutions of su-times with PBS and once with 0.3% bovine serum albumin (BSA) in PBS for 10 min at room temperature. The cells crose gradient-purified 35 S-labeled TGEV (used as an were incubated with hybridoma supernatants containing highest dilution giving a binding threefold higher than background. a mixture of MAbs 1D.B12, 5B.H1, and 1D.G3 (specific for S protein sites B, C, and D, respectively) or with MAb Detection of the different antigenic sites in the S protein fragments, encoded by recombinant Ad-TS, was H2-19 specific for a 70K Ad5 antigen. After three additional washings with PBS, cells were covered with a carried out by cRIA using the antiserum elicited in hamsters by the different recombinants. The binding of 125 I-1:200 dilution of fluoresceinated goat anti-mouse immunoglobulins (Cappel Laboratories) in 0.3% BSA in PBS, labeled MAbs to purified TGEV bound to microplates was performed as previously reported (Correa et al., 1988) incubated for 40 min at room temperature, washed five times for 10 min each with PBS, and mounted on glyc-with some modifications. Briefly, purified TGEV (0.1 mg/ well) was plated, remaining binding sites were saturated erol-PBS (9:1).

with 5% BSA in PBS, and 125 I-labeled MAbs (sp act 1.5 Binding of 125 I-labeled MAbs to 293 cells infected 1 10 7 cpm/mg; 4 1 10 5 cpm/well) were added and incuwith recombinant Ad-TS bated for 2 hr at 37Њ in the presence of fivefold dilutions of the competitor antiserum prepared in PBS with 0.1% Confluent ST cell monolayers plated on 24-well mi-BSA. Microplates were washed six times with 0.1% BSA croplates were infected (m.o.i. 40 PFU/cell) with recombiand 0.1% Tween-20 in PBS. Well bottoms were cut and nant Ad-TS viruses. At 24 hr postinfection, cells were bound radioactivity was determined in a gamma counter. washed with PBS and fixed in methanol:acetone (1:1) for

The percentage of radioactivity bound was determined 15 min at 020Њ or in 4% paraformaldehyde in PBS for 20 in relation to the radioactivity bound in the absence of min at room temperature. Cells were washed three times competitor MAb. Purified homologous MAbs were used with PBS and for 2 hr with 0.5% BSA in PBS. Aliquots of as positive controls in the cRIA. 0.25 ml of 125 I-labeled purified MAbs (1 1 10 6 cpm/well; 1.5 1 10 7 cpm/mg) (Greenwood et al., 1963) 

Protection of swine by immune serum 0.2% BSA were added to each well and incubated for 1 hr at room temperature, and the cell monolayers were

The virulent TGEV strain PUR46-SW11-ST2 (1 1 10 7 washed six times with PBS. MAb binding was deter-PFU/swine) was mixed with 3 ml of the porcine antiserum mined by collecting the cells in 0.25 ml of 0.5 N NaOH induced by recombinants Ad-TS8 or Ad-TS06, incuand counting the radioactivity in a gamma counter. bated at 37Њ for 60 min, and administered using a gastric tube to 2-day-old miniswine born from TGEV-seronega-Immunization of hamsters and swine tive sows. Inoculated animals were fed three times per day with milk formula for newborns (Nidina 1, Nestlé ) Eight-week-old golden Syrian hamsters were immucontaining 3 ml of the antiserum. Control animals were nized with infectious Ad-TS recombinants by three treated following the same procedure but using serum routes: oral (4 1 10 8 PFU in 0.2 ml of PBS), nasal (2 1 induced by wt Ad5. Virus titers after 1, 2, and 3 days in 10 8 PFU/0.1 ml), and intraperitoneal (1 1 10 9 PFU/0.5 ml). animals challenged with virus treated with control serum The virus was administered at Days 0, 32, 60, and 90, and 1, 2, or 5 days postinoculation in animals challenged and orbital plexus puncture bleedings were performed with TGEV immune serum-treated virus were determined at Days 0, 32, 47, 87, 105, and 115. Females with highest in tissue extracts from jejunum and ileum, lungs, mesentiters of TGEV-specific antibodies were crossed with nonteric, and mediastinal lymph nodes. Tissue homogenizaimmune males, and 8 days later another dose of the tion was performed at 4Њ using an OMNI 2000 homogehomologous Ad-TS recombinant was administered.

nizer (Omni International). Twenty-four hours after delivery, hamsters were subcutaneously administered 10 IU of oxytocin. The milk was collected 1 hr later by applying vacuum with a syringe.

RESULTS Milk was diluted fourfold in PBS and stored at 020Њ.

Ad5-TGEV recombinants One-month-old swine, from crossing Large White and Belgium Landrace, were immunized three times at 0, 28, Ten Ad5-TGEV recombinants expressing TGEV S and 56 days, each time by three routes: oral (1 1 10 9 gene fragments were constructed using vectors with dif-PFU), nasal (1 1 10 9 PFU), and intraperitoneal (1 1 10 9 ferent deletions on E3 gene or combinations of SV-40 PFU per dose). Serum was collected 14 days after the promoter and polyadenylation signals. Using these relast immunization.

combinants S protein fragments of four different sizes were expressed. The recombinants were obtained by Radioimmunoassay (RIA) and competitive RIA (cRIA) replacing the E3 gene of the Ad5 genome with S gene with 125 I-labeled MAbs sequences starting from nt 08 and the first 5-end 1135, 1587, 3329, or 4341 nt of the S gene. These recombinants RIA was performed using purified TGEV as antigen (0.1 mg/well) as previously described (Jimé nez et al.,

code for fragments of 378, 529, 1109, and 1447 amino acids (aa) extended from the amino-terminus (Fig. 2) . The 1986). Titers in RIA were defined as the inverse of the faint band (results not shown). Recombinant products with apparent molecular masses of 68 and 135 kDa (Fig.  3, lanes c and e, respectively) were obtained for recombinant S protein fragments of 378 and 1109 aa, respectively. Recombinants Ad-TS07 and Ad-TS5, both coding for polypeptides of 529 aa, gave a main band of 86 kDa and a minor band of 80 kDa (lane d), which probably corresponds to an underglycosylated form of the antigen or to a degradation product. The difference between the expected and the apparent molecular mass of the recombinant products suggests that these are heavily glycosylated, as occurs during S protein synthesis after TGEV serum using extracts from 293 cells infected with the chimeric Ad-TS viruses (Fig. 3) . The amount of S protein was based on the comparison of band intensity after last product represents the full-length spike protein. The immunoprecipitation and autoradiography of 35 S-labeled constructs were obtained using either plasmid recombinant antigens and reference sucrose gradient-pFG144K3 or plasmid pAB14 (Fig. 1) , with deletions of purified 35 S-labeled TGEV with known protein concentra-1.88 or 2.69 kb, respectively, in E3 (Bett et al., 1993) .

tion. Both reference virus and recombinant antigens were Recombinant plasmids were constructed as summarized labeled and analyzed in parallel using the same experi- (Fig. 1) . When indicated, the S gene fragments were mental conditions. Since the distribution of the methioflanked by Pr and polyadenylation signals (Fig. 2 ) by clonnine and cysteine in the different fragments was similar, ing them into vector pSV2X3 or pSV2X4. Inserts were no significant correction of band intensity was necessary subcloned into plasmid pFG144K3 or pAB14 containing in the analysis. The expression levels ranged from 0.1 the 3-end half of Ad5. Human 293 cells were cotransto 10 mg of S protein per 10 6 infected cells. Maximum fected with one of these plasmids and pFG173, which expression levels (5 to 10 mg/10 6 cells) were obtained contains almost the entire Ad5 genome with a lethal for recombinants Ad-TS5, Ad-TS8, and Ad-TS07, interdeletion across the E3 region. Fully infectious Ad-TS mediate levels (1 to 3 mg/10 6 cells) for Ad-TS9, Ad-TS2, viruses were recovered following recombination in co-Ad-TS05, and Ad-TS6, and minimum (around 0.1 mg/ transfected 293 cells. Recombinant viruses were plaque 10 6 cells) for recombinants Ad-TS01, Ad-TS02, and Adpurified. The DNA from all the recombinants gave the TS06. Relative expression levels were highly reproducpattern and sequence expected for each insert by HindIII ible in different experiments. All the recombinants, inrestriction endonuclease analysis and sequencing of cluding those expressing minimum amounts of antigen, DNA junctions (results not shown).

were also consistently positive in the immunofluores-After infection of 293 cells with Ad-TS recombinants, cence and 125 I binding assays and in the induction of S protein antigens remained cell associated. Tris buffer TGEV-specific antibodies (see below). containing 1% SDS was used to solubilize them. The When indicated, the S gene fragment cloned into Ad5 estimated size of recombinant S antigen expressed by was flanked by Pr and polyadenylation signals (Fig. 2) .

Comparison of the expression levels in constructs with and representative results are shown (Fig. 3) . S polypep-S gene fragments of the same size indicated that Ad5 tides were detected with a polyclonal TGEV-specific porrecombinants made using pFG144K3 plasmids excine serum. Good specific immunoprecipitation bands pressed higher levels of antigen than those based on were systematically obtained with all recombinants explasmid pAB14, although in some cases (i.e., recombinant Ad-TS07 compared with Ad-TS5) the level of ex-cept Ad-TS01, Ad-TS02, and Ad-TS06, which gave a pression was similar (results not shown). In recombinants with the same E3 deletion it was also observed that removal of SV-40 Pr yielded Ad-TS recombinants with higher expression levels (results not shown).

To study the cellular location of recombinant S antigen, we used immunofluorescence analysis of ST cells infected with four selected recombinants each coding for S fragments of different size: 387, 529, 1109, and 1447 (full-length S protein) amino acids. A bright fluorescent signal was observed in the cytoplasm of methanol-acetone-fixed cells infected with recombinants Ad-TS8, Ad-TS5, and Ad-TS9 (results not shown). Highest fluorescence intensity was seen with TGEV-infected cells and TS recombinants. When immunofluorescence was performed with a human Ad5-specific MAb (which binds 72K protein) bright fluorescence was observed on discrete though amino acids 380 to 387 of S protein site D are areas of the nucleus, but not in the cytoplasm (results coded by recombinant Ad-TS5, this site was poorly recnot shown), in contrast to the cytoplasmic fluorescence ognized by MAb 1D.G3 specific for D site on Ad-TS5observed with TGEV-specific MAbs.

infected cells (Fig. 4) . The four antigenic sites were An estimation of the relative amount of S antigen loweakly detected in cells infected by recombinant Adcated in the cytoplasm or accessible on the surface of TS06, probably due to the low replication level of this Ad5-infected ST cells was determined by studying the recombinant. binding of 125 I-labeled MAb 1D.B12 (site B-specific) to methanol-or paraformaldehyde-fixed cells (results not Immunogenicity of the recombinants shown). This MAb was selected because it recognizes an epitope present in all Ad-TS recombinants. Cells in-Immune responses elicited by the different recombinants were studied by inoculating hamsters both orona-fected with recombinants Ad-TS8, Ad-TS5, and Ad-TS9 permeabilized with methanol-acetone expressed the sally and intraperitoneally (Fig. 5) . Seven of the ten recombinants summarized in Fig. 2 elicited titers in RIA highest amount of S antigen, which ranged between 60 and 66% of the amount expressed on ST cells infected higher than 2500 and NI between 1 and 3. The best inducers of TGEV-neutralizing antibodies were recombi-with TGEV. In cells infected with these recombinants the binding of site B-specific MAb to exposed antigen was nants Ad-TS8, Ad-TS2, and Ad-TS06, expressing either the smallest protein fragment or the full-length protein around 10% of the binding to cytoplasmic S antigen of TGEV-infected cells. That is, the amount of S antigen (Fig. 2) . Four recombinants (Ad-TS8, Ad-TS5, Ad-TS9, and detected on the surface of the infected cells was at least sixfold lower than that seen in the cytoplasm. The recom-Ad-TS06), each expressing S gene fragments of different lengths (Fig. 2) were selected to study the induction of binant products were not detected in the supernatants of infected cells, although the media were not concen-an immune response to sites A, B, and D by cRIA (Fig.  6 ). Site C was not included in the study because the trated to detect small antigen amounts.

Proper folding of the S protein fragments expressed amino acid sequence PNSD recognized by MAbs specific for this site is present in pro-by the four selected recombinants was evaluated by determining the amount of 125 I-labeled MAb specific for anti-teins of the immunoglobulin superfamily and other serum proteins (Correa et al., 1988;  I. Correa and L. Enjuanes, genic sites A, B, C, and D bound to infected ST cells (Fig.  4) . All recombinants expressed sites C and B. Recombi-unpublished results). Recombinant Ad-TS9 induced an immune response to antigenic sites B, D, and A (Fig. 6) . nant Ad-TS9, in addition, expressed sites D and A. Al- FIG. 5 . Immune response induced by Ad-TS recombinants in hamsters. Groups of four golden Syrian hamsters were immunized at Time 0 and at times indicated by arrows (see Materials and Methods) with the indicated recombinants. Sera collected at 0, 32, 47, 87, and, in some cases, at 105 and 115 days postinfection were evaluated by RIA and neutralization against TGEV. Mean serum titers and standard deviation errors are represented for each time point. The titer by RIA was defined as the inverse of the highest antibody dilution giving a binding three times higher than the background in the RIA assay. The NI was defined as the log 10 of the ratio of the PFU after incubating the virus in the presence of medium or the indicated antiserum.

All recombinants induced a strong response to site B and milk was determined between Days 1 and 2 during lactation (Fig. 7) . The three recombinants induced anti- (Fig. 6A) which is conformation and glycosylation depenbodies in serum with titers in RIA ranging from 5 1 10 3 dent . As expected, site A was only to 1.5 1 10 4 and in milk from 2 1 10 3 to 3 1 10 3 (Fig. reconstituted by recombinants Ad-TS06 and Ad-TS9, 7A). Serum and milk antibodies neutralized TGEV with expressing the full-length S protein or the 135-kDa S NIs ranging from 2 to 4 and around 1, respectively (Fig. antigen, but not by recombinants which do not include 7B). As expected, recombinants with no insert did not the residues implicated in this site (Fig. 6C) .

elicit TGEV-specific antibodies. While antibody titers in Induction of lactogenic immunity by Ad-TS sera decreased with insert size, the NI increased, sugrecombinants gesting that antibodies to site A contributed significantly to the neutralization of TGEV.

Induction of immune response in swine by Ad-TS8, Ad-TS9, and Ad-TS06 were crossed with nonrecombinants Ad-TS8 and Ad-TS06 immune males and administered a third dose of the homologous Ad-TS recombinant 10 days before delivery.

The Ad-TS8 and Ad-TS06 recombinants expressing the smallest insert and the full-length spike protein, re-The presence of TGEV-specific antibodies in the sera and 5, respectively). To study the potential of these antisera for protection against TGEV, sera induced by these recombinants were examined for the ability to prevent TGEV infection. Virulent TGEV (PUR46-SW11-ST2 strain, 1 1 10 7 PFU/dose) was mixed with the antibody induced by each recombinant, incubated at 37Њ for 60 min, and administered to highly susceptible 2-day-old miniswine. Virus titers were determined in jejunum and ileum, lungs, mesenteric, and mediastinal lymph nodes at 1, 2, 3, and 5 days postinoculation. The results (Fig. 8) indicated that virus titers found in the enteric tissues were between 10 2 and 10 3 -fold lower when virus was premixed with antiserum induced by recombinant Ad-TS8 (Fig. 8D) , and very low titers (õ5 1 10 2 PFU/g of tissue) of infectious virus were detected in the small intestine of newborn pigs that were administered the antibody elicited by recombinant Ad-TS06 (Fig. 8F) . In contrast, titers ranging between 7 1 10 3 and 1 1 10 7 PFU/g of tissue were detected in the tissues of control animals to which serum induced by wt Ad5, used as a control, was administered (Fig. 8B ). In addition, neither mortality nor clinical symptoms were observed in animals treated with serum induced by recombinant Ad-TS06 (Fig. 8E) , while control animals presented diarrhea 24-30 hr postinfection and died around Day 3 postinfection (Fig. 8A) .

Ten Ad5-TGEV recombinants have been constructed and screened for their ability to express spike protein fragments of TGEV. Four recombinants expressing the full-length spike protein or truncated fragments spanning different lengths of S protein from the amino-terminus have been selected, and their ability to induce virusneutralizing antibodies was determined. These Ad-TS viruses induced lactogenic immunity in hamsters, and the recombinant expressing the full-length S protein elicited antisera that, when mixed with a lethal dose of virus prior to administration to susceptible piglets, prevented the induction of disease symptoms.

Helper-independent Ad5 viruses with a deletion in the E3 gene have been constructed, and the S gene was inserted into the E3 gene. Two types of Ad5 recombinants interest to determine the comparative levels of expression in these two plasmids. Expression levels were always higher using Ad5 viruses with the smaller deletion spectively, were selected to study the induction of TGEVneutralizing antibodies in swine. Although the level of in E3, independent of the insert size, suggesting that removal of the splicing acceptor site after the L4 gene recombinant antigen produced in ST cells was high for recombinant Ad-TS8 and low for Ad-TS06 (Figs. 3 and might have reduced E3 gene expression. Sequences inserted without an exogenous polyadenylation signal 4), both recombinants induced high titers of TGEV-specific antibodies in swine as determined by RIA (1 1 10 4 were successfully expressed, indicating that the polyadenylation signal of the E3 gene has probably been used. and 5 1 10 4 , respectively) and by neutralization (NI of 2 In general, recombinants with relatively small inserts low levels of S antigen and, accordingly, of all antigenic sites (A, B, C, and D), probably due to low replication (1135, 1587, and 3329 nt) expressed larger amounts of S polypeptide than those with larger (4470-nt) inserts.

levels. Nevertheless, antigenic sites A and B were properly folded after infection with Ad-TS06 virus since high The recombinants with smaller inserts gave Ad5 titers in cell culture between 3 1 10 8 and 1 1 10 9 PFU/ml, while antibody levels against these sites were elicited in hamsters, as detected by cRIA (Fig. 6 ). S protein trimer forma-Ad5-TS virus with an insert of 4470 nt consistently gave titers lower than 10 7 PFU/ml. Thus, the level of expression tion easily explains the dichotomy between low expression levels and high efficiency in eliciting a high immune in these recombinants correlates well with their level of replication. The three recombinants (Ad-TS8, Ad-TS5, response. S protein trimers (the native form of the glycoprotein in the virus) probably are more stable and better and Ad-TS9) with genome sizes lower than 104% of wt Ad5 were stable after 10 passages, while the recombi-represent the peplomer in the native virion. Although recombinant Ad-TS5 contains the sequences coding for nant with a genome size close to 105% of wt Ad5 (Ad-TS06) was unstable (results not shown). These results site D core (located in S protein from aa 377 to 390) Lenstra et al., 1991;  Posthumus et are in line with previous work suggesting that the Ad5 virion has the ability to package approximately 105% of al., 1990), it was very weakly detected by site D-specific MAbs, while sites C and B, also encoded in this recombi-the wt genome length. This value is generally considered to be the maximum working capacity of the system nant, were well represented. Site D may have been hidden by incorrect folding of the S protein in this area. (Ghosh-Choudhury et al., 1986; Berkner, 1988; Bett et al., 1993) .

Site A, the major inducer of TGEV-neutralizing antibodies, was detected in larger amounts after infection by recom-Viruses in which the inserted gene was flanked by an SV-40 Pr always showed lower expression levels than binant Ad-TS9 (expressing S protein without the membrane anchor domain) than by recombinant Ad-TS06 those not flanked by this Pr (Fig. 3) . This suggests that the SV-40 Pr, in the context that has been used in this (which expresses the full-length S protein). This may be a consequence of the higher expression levels provided work, is inhibiting and transcription is probably driven from the nearby Ad5 E3 Pr. The transcription could also by Ad-TS9, since it has been previously shown (Godet et al., 1991) that the full-length spike forms trimers and be driven from the major late protein Pr that is located far to the left at m.u. 16. Similar observations have been reconstitutes site A better than truncated S proteins missing the membrane anchor domain. In fact, one of the two made with other Ad5-based vectors containing analogous E3 substitutions (Schneider et al., 1989; major inducers of TGEV-neutralizing antibodies was Ad-TS06 virus, in spite of the low amount of S protein pro-and Both et al., 1993) .

Antigenic sites C, B, D, and A (starting from the amino-duced by this recombinant. Seven of ten Ad-TS recombinants expressing S frag-terminal end) have been defined on S protein (Correa et al., 1988; Gebauer et al., 1991) . Sequences coding for ments induced TGEV-neutralizing antibodies in hamsters. Recombinant Ad-TS8, expressing a truncated form sites C and B were included in all recombinants and, in fact, S polypeptides with these two sites were detected of S protein spanning 378 aa from the amino-terminus (which includes sites C and B but not site A), induced after infection with all Ad-TS viruses. The recombinant coding for the full-length S protein (Ad-TS06) expressed virus-neutralizing antibodies. Since site C does not in- FIG. 8 . Protection of swine with porcine sera elicited by Ad-TS recombinants. TGEV-specific swine antiserum was elicited by administration of wt Ad5 virus, Ad-TS8, or Ad-TS06 recombinants (see Materials and Methods). The number of swine surviving after the oral administration of 1 1 10 7 PFU of the virulent strain PUR46-SW11-ST2 of TGEV mixed with antisera induced by (A) wt Ad5 or by the recombinants (C) Ad-TS8 or (E) Ad-TS06 expressing the 1135 amino-terminal nt or the full-length spike protein, respectively, is shown. The recovery of infectious virus was determined 1, 2, and 3 or 5 days postinfection (when the animals either died or were sacrificed) in the indicated tissue homogenates, in animals administered the virulent virus with serum from (B) Ad5, (D) Ad-TS8, or (F) Ad-TS06 immune swine. Three groups of five swine were used to follow the survival rate. The infectious virus was followed in three groups of three animals each. Mean values have been represented. Standard deviations were lower than 25% in all cases and are not shown. duce virus-neutralizing antibodies, site B, or neighboring and Van der Zeijst, 1995) or, alternatively, other factors similarly to those described in mouse hepatitis virus sys-antigenic domains involved in virus neutralization, have been reconstituted in a functional form. It has been pro-tem (Fazakerley et al., 1992; Yokomori et al., 1993) . Recombinant adenoviruses expressing only the 378 amino-posed that factors mapping in the S segment which has been deleted in the porcine respiratory coronavirus terminal residues of the S polypeptide (which are mostly deleted in PRCV strains) provide partial protection (PRCV) (from aa 21 to 241) (Callebaut et al., 1988; Sá nchez et al., 1990) , and more precisely alterations in amino against TGEV. These data indicate that the amino-terminal S protein fragment might be relevant to confer enteric acid 219 (or residues close to it) might be involved in the loss of enteric tropism (Sá nchez et al., 1992) . These fac-tropism by complementing the binding of N-aminopeptidase (identified as a major TGEV receptor) to an S protein tors might be the presence of a second receptor binding site recognized by a putative second receptor (Enjuanes domain mapping close to antigenic site A (Delmas et al., 1992; Godet et al., 1994) . Another observation supports with higher titers in swine than in hamsters, although both species were permissive to virus infection. S protein this hypothesis. Ad5 vectors have been used to express the amino-terminal 564 aa of PRCV S protein, resulting has been previously expressed using E. coli (Hu et al., 1984 (Hu et al., , 1987 or poxviruses (Pulford and Britton, 1991) , but in production of TGEV-neutralizing antibodies (Callebaut et al., 1994) which did not protect against challenge with TGEV-neutralizing antibodies were only elicited with recombinant poxviruses. Expression of S antigenic site D, virulent TGEV (Callebaut and Pensaert, 1995). By contrast, the Ad-TS recombinant eliciting antiserum provid-as a fusion protein on the surface of E. coli led to induction of TGEV-neutralizing antibodies when purified re-ing passive protection against challenge with virulent TGEV carries S sequences derived from TGEV instead combinant antigen was used as immunogen, but not when live vector was administered (Bousquet et al., of PRCV. The presence of 224 aa (from residue 21 to 244) in recombinant Ad-TS06, which are deleted in the 1994). Using Salmonella typhimurium, site D has been expressed and TGEV-neutralizing antibodies have been PRCV, might have been critical to achieve the observed protection. This interpretation is in agreement with the elicited in serum and in mucosal areas using live recombinant bacteria (Smerdou et al., 1995) but protection ex-partial protection seeing with the antiserum elicited in swine by recombinant Ad-TS8, which includes the se-periments using these systems have not been reported. Ad5 vectors have a high probability of inducing effective quences deleted in PRCV, but at the same time indicates that larger spike protein fragments (as those including mucosal immunity against TGEV, since this virus showed tropism for mucosal tissues in pigs, and the animals site A) are needed to elicite full protection.

Protection by recombinant adenoviruses expressing S infected with this virus experienced neither respiratory nor intestinal disorders (Callebaut and Pensaert, 1995; protein fragments lacking site A extends the results recently reported (Tulboly et al., 1994) on S protein expres- Callebaut et al., 1994; Torres et al., 1995) . sion using baculoviruses. These authors showed induction of TGEV-neutralizing antibodies only with recombi-ACKNOWLEDGMENTS nants expressing S protein fragments spanning 745 aa We thank Granja Cantoblanco de Animales de Laboratorio (Hospital or more from the amino-terminus, that is, with S protein General G. Marañó n, Comunidad de Madrid) and Laboratorios Sobrino Cyanamid (Olot, Girona) for providing inbred and outbred swine, re-fragments including site A, but not with S protein frag-

Antigenic differentiation between transmissible gastroenteri-TGEV expressed by recombinant baculovirus

TGEV coronavirus ORF4 encodes a membrane protein that is incorporated Callebaut

Major antigen of porcine respiratory coronavirus

Virus retaining spike glycopolypeptide S2 but not S1

Adenovirus-based expression is unable to induce virus-neutralizing or haemagglutination-inhibiting vectors and recombinant vaccines

A new technique for the Antigenic structure of the E2 glycoprotein from transmissible gastroassay of infectivity of human adenovirus 5 DNA

McDerinfection by affinity-purified spike glycoprotein of murine hepatitis mott

Charactermurine coronavirus spike glycoprotein and evidence that it forms istics of a human cell line transformed by DNA from human adenovipart of a complex tridimensional structure

The prepara-Epitope specificity of protective lactogenic immunity against swine tion of 131 I-labelled human growth hormone of high specific radioactransmissible gastroenteritis virus

5367-Herpes simplex virus IgG Fc receptor induced using recombinant 5375. adenovirus vectors expressing glycoproteins E and I

Techniques for human adenovirus vector construction and character

Cell Biology: A Four major antigenic sites of the coronavirus transmissible gastroen-Laboratory Handbook

High efficiency expression of the surface glycoprotein gp195 of porcine transmissitransformation of E. coli by high voltage electroporation. Nucleic ble gastroenteritis virus

Studies of Fazakerley

TGEV S protein gp195 expressed in E. coli and by a TGE-vaccinia The V5A13.1 envelope glycoprotein deletion mutant of mouse hepativirus recombinant

Monoclonal antibodies to the matrix (E1) glycoprotein (1986). Critical epitopes in transmissible gastroenteritis virus neutralof mouse hepatitis virus protect mice from encephalitis. Virology 168, ization

Isolation of adenovirus type 5 host

range deletion mutants defective for transformation of rat embryo S

missible gastroenteritis virus

Immunogenic peptide comprising a mouse hepatitis virus

A59 B-cell epitope and an influenza virus T-cell epitope protects Residues involved in the formation of the antigenic sites of the S against lethal infection

Human adenovirus cloning vectors based on infectious acid changes in the viral glycoprotein M affect induction of alpha bacterial plasmids

Processing and antigenicity of entire and anchor-free spike glycoprotein-S of coronavirus Lecomte

Antigenic homology among coronaviruses retype 3-induced acute disease by an anti-nucleoprotein monoclonal antibody

glycoprotein of VSV by infectious adenovirus vectors

Selection of mimotopes from a random sequence expression library by monoclonal antibodies against transmissible gastroentritis coro

Molecular Cloning: tis coronavirus S protein fused to E. coli heat-labile toxin B subunit in attenuated Salmonella for oral immunization. Submitted for publi-A Laboratory Manual

Studies on transmissible gastroenteritis of swine. II. Selected characteristics of a cytopatho-In ''Immunochemistry of Viruses. II. The Basis for Serodiagnosis and Vaccines

Mechanisms of transmissible gastroenteritis coronavirus neutralization

rus based vector using the firefly luciferase as a reporter gene

Protective effect of monoclonal antibodies on lethal mouse hepatitis virus infection hepatitis virus-4 (strain JHM): Correlation with biological activities

276, and their use in protection against transmissible gastroenteritis virus

Immunogenicity of the S protein of transmissible gastroenteritis virus ex-F. L. (1990). A recombinant human adenovirus vaccine against rabies

Intracellular processing of the porcine coronavirus transmissible gastroenteritis virus spike protein strain of transmissible gastroenteritis virus: Comparison of reactivity with virulent and attenuated virus

Mouse hepatitis ''Diseases of Swine

Iowa State Univ. Press, virus spike and nucleocapsid proteins expressed by adenovirus vector protect mice against a lethal infection

A spike protein-dependent cellular factor other than the viral receptor is required for mouse hepatitis virus entry