ANTIGEN OF TOXOPLASMA GONDII

The 54 kDa antigen identified and described in WO 92/11366 constitutes a potentially protective agent against infection by Toxoplasma gondii. the parasite responsible for the development of toxoplasmosis in certain individuals. In WO

92/11366 human T-cell clones specific for T. gondii were obtained, one of which, designated T-cell clone 32 (TCC32), was used to identify the 54KD antigen (also known as Tg34).

Whilst the identification of the 54KD antigen represents an important advance there still exists a need to present this antigen to the immune system in a way as close as possible to the natural route. In one aspect he present invention relates to expression systems for the 54KD antigen. In another aspect the invention relates to the identification of the epitope on the 54KD antigen which is recognised by TCC 32 and to the synthesis of a polypeptide corresponding to this epitope. In another aspect the invention relates to chemical conjugates or fused recombinant proteins incorporating the said epitope.

Because T lymphocytes play a major role in protective immune responses to intracellular pathogens, significant efforts have been made to identify the regions of a protein that arc recognized by the T-cell receptor. Approaches to design synthetic vaccines requires the appropriate selection and production of peptides which contain immunodominant T-cell epitopes. Several models have been elaborated to predict such antigenic segments from the primary sequence.

In making the present invention, we applied three predictive models in order to determine the region of the cloned Tg34 protein of Toxoplasma gondiiC ⁷ ) which contains the epitope recognized by the human T-cell clone 32. The details are given hereinbelow (see Example 1) from which it was found that the said epitope corresponded to amino acids 197 to 216 of the 54KD antigen of T. gondii described in WO 92/11366.

Accordingly in a first aspect the present invention provides a polypeptide corresponding to residues 197 - 216 of the 54KD antigen of T. gondii (the 'T-cell epitope').

In a related aspect the invention provides a polypeptide having the amino acid sequence TDPGDVVIEELFNRIPETSV where the letters represent standard abbreviations for amino acids. This corresponds to the sequence of amino acids 197 to 216 in the 54KD antigen of T. gondii (see WO 92/11366).

Preferably the polypeptide is in purified form, preferably greater that 60% pure, more preferably over 75% pure, advantageously over 90% pure, for example

95-100% pure.

In a preferred aspect the polypeptide (T-cell epitope) is in substantially pure

form.

The polypeptide may be synthesised by conventional techniques, for example the Merrifield solid phase method.

The epitope identified by the present inventors has potential for the formation of immunogenic molecules able to confer immunity to T. gondii infection. In one aspect such molecules may be chemical conjugates which 'present' the epitope in the correct manner to the immune system.

Accordingly in a further aspect the invention provides a compound comprising the polypeptide according to the invention chemically bound to a heterologous molecule and having the ability to cause the formation of antibodies against T. gondii following internal administration to a human or animal.

Such chemical conjugates may be formed by well known techniques, for example those described in Current Protocols in Immunology (1991; Wiley Intersciences, New York). In another aspect it is possible to form, by recombinant DNA techniques,a hybrid molecule incorporating the T-cell epitope described herein. Using techniques analogous to those already described in the literature (see Martineau et al (1991) Bio/Technology 9_, 170 - 179; Rutgers et al (1988) Bio/Technology, t_, 1065 - 1070), such molecules may be expressed in particulate form so as to present the T-gondii epitope to the immune system in an effective manner.

Accordingly the invention also provides a recombinant protein comprising the polypeptide of the invention fused to a heterologous antigen, which protein is capable of forming particles on expression in a suitable host.

Preferably the host is yeast, especially Saccharomvces cerivisae. In a particularly preferred aspect the heterologous antigen is derived from

Hepatitis B surface antigen.

By hepatitis B surface antigen is meant the 226 amino acid S protein optionally with pre-S sequence present (for a discussion of the nature of Hepatitis B surface antigen, see Ganem, D and Varmus, H. E. (1987) Ann Rev. Biochem, 5_5_, 651 - 693).

In a particular embodiment of the invention there is provided a recombinant protein which consists of an initiation methionine followed by the sequence TDPGDVVIEELFNRIPETSV, the 13 amino acids from the preS2 region of HBsAg and the 226 amino acids from the major HBsAg coding region (S protein). There is also provided a vaccine composition comprising an immunologically effective amount of any compound according to the invention admixed with a pharmaceutically acceptable carrier.

The term 'immunoprotective' refers to the amount necessary to elicit an immune response against a T. gondii challenge such that disease is averted or

mitigated and transmission of the disease is blocked or delayed.

Vaccine preparation is generally described in New Trends and Developments in vaccines, Voller _____] (eds), University park Press, Baltimore, Maryland, 1978.

The amount of protein of the present invention present in each vaccine dose is selected as an amount which induces an immunoprotective response without significant adverse side effects in typical vaccines.

Such amount will vary depending on which specific immunogen is employed and whether or not the vaccine is adjuvanted. Generally it is expected that each dose will comprise 1-lOOOug of protein, preferably 1 - 200ug. An optimal amount for a particular vaccine can be ascertained by standard studies involving observation of antibody titres and other responses in subjects. Following an initial vaccination, subjects will preferably receive a boost in about 4 weeks, followed by repeated boosts every six months for as long as a risk of infection exists.

The invention also provides a method of preventing T. gondii infection in humans or animals by administering to a human or animal in need thereof an immunologically effective amount of a vaccine composition according to the invention.

As described further in the examples below the present inventors have surprisingly found that it is advantageous to express the 54KD antigen of T. gondii or the T-cell epitope thereof as herein described in hosts which were not mentioned in WO 92/11366. Particular hosts include vaccinia virus and M. bovis - BCG. These have the advantage that they are live vaccination vehicles. CHO cells may also be advantageously used as host cells. These have the advantage of providing stable transformants. Accordingly in yet a further aspect the invention provides a metiiod of producing the 54 KD antigen of T. gondii or the T-cell epitope thereof (amino acids 197 - 216) by causing the product to be expressed from a host cell comprising a vaccinia virus or M. bovis - BCG, followed by purification of the product by conventional techniques. The antigen produced may be engineered to be cleared of the transmembrane domain. This has the advantage of the protein being secreted into the culture supernatant. The full length protein or a truncated derivative may be produced in yeast. Typically removal of the transmembrane domain will involve removing the majority of residues 464 to 485. In a related aspect the invention provides a vaccinia infected cell or M. bovis -

BCG host cell transfected or transformed with a vector comprising DNA encoding the 54KD antigen of T. gondii or the T-cell epitope thereof (amino acids 197 - 216). Particular vectors for this purpose,described below, also form part of the invention. The following examples and Figures (explained below) illustrate the

invention. Literature references follow the examples.

Example 1: Identification of the Tg34 epitope recognized by the human T-cell clone 32 The method of Berzofsky et al. (-), called the amphipathic helix model, is based on the strong correlation found between helper T-cell antigenicity and α-helical amphipathicity. In this model, antigenic sites are postulated to be helices with one face predominantly polar (which is recognized by the T-cell receptor) and the opposite face predominantly apolar (which is bound to class I or class II molecules of the major histocompatibility complex on the surface of the presenting cell).

The method of Rothbard et al. ( ² ) is based on the observation that a large percentage of helper and cytotoxic determinants contain a linear pattern composed of a charged residue or a glycine followed by two or three hydrophobic residues, and terminated by a polar residue or a glycine. The method of Kourilsky et al. ( ³ ), called the self peptidic model, is based on the statistically significant tendency of T immunogenic peptides to be constituted of clusters of rare tetrapeptides, as evaluated from the available sequence data banks.

Methods and Results 1. Prediction of T cell epitopes

The algorithm of Berzofsky has been reproduced from the publication of Spouge et al. 0) with some adaptative changes to our computer. The general scheme of the program is, first, convert the amino acid sequence into a sequence of hydrophobicity values. Second, divide the hydrophobicity sequence into overlapping blocks (7 or 11 residues long). Third, in each of the blocks, search for periodicity in hydrophobicity consistent with a regular amphipathic helical structure. Fourth, distinguish between stable and unstable amphipathic helical segments. In practice, the original program offered the possibility to perform the analysis by choosing a window respectively of 7 or 11 residues. We have added the possibility to use both windows consecutively and to represent the amphipathic segments aligned under the primary sequence by using different symbols in order to distinguish regions predicted by both windows, from regions predicted by only one window and from regions predicted by the window of 11 residues but with a weak amphipathic score (<8).

The algorithm of Rothbard has been deduced from the publication mentioned above (2) where the four sets of amino acids composing the proposed patterns are defined. We have just added tyrosine to the fourth set containing polar residues because we observed, in the table reporting the known epitopes containing the motifs, that Tyr could be present in the last position of a motif of either 4 or 5 residues.

To implement the algorithm of Kourilsky, we first generated a subset of the

protein sequence databank SWISSPROT containing only human sequences from which all immune system-related polymorphic protein sequences were excluded (i.e. immunoglobulins, T-cell receptors, MHC molecules). In a second step, this databank subset, representing the somatic self catalog of the human specie, was read dirough a window of 4 residues and the number of occurrence for all the possible tetrapeptide sequences were scored using a four dimensional matrix. Finally, submission of a proteic sequence to this program gives for each tetrapeptide found in this sequence the corresponding value reported by the matrix.

However, because different regions of a given protein may be selected by each model, we have written a program which performs under the primary sequence of interest an alignment of the results generated from each predictive method. This approach has been used to increase the likelihood of identification of T-cell determinants allowing to focus upon those antigenic segments that are congruently predicted by different algorithms. All the computer programs described above have been written in FORTRAN, a powerful scientific programming language.

The primary sequence of Tg34 protein has been submitted to this set of three programs and on the basis of the resulting analysis we have chosen to synthesize chemically three peptides out of ten predicted segments:

The segment 197-216 contains an amphipatic α-helix predicted by both windows with amphipathic scores of 23.9 for the window of 7 residues and 28.9 for the window of 11 residues. It contains also 2 contiguous motifs of Rothbard (underlined), the first one being a pentamer and the second one a tetramer. 197 TDPGDVVIEELFNRIPETSV 216 20 amino acids

The segment 303-410 contains two overlapping predicted amphipathic α- helices predicted by the window of 11 residues with amphipathic scores of 8.6 and 4.3). It contains also two non contiguous motifs of Rothbard (underlined), and three consecutives tetrapeptides with nul occurrence in the self peptidic human set of tetrapeptides.

393 LQLIRLΔΔSLQHYGLVHA 410 18 amino acids The segment 501-524 contains an amphipathic α-helix predicted by both windows with amphipathic scores of 47.4 for the window of 7 residues and 54.7 for the window of 11 residues. It contains also three motifs of Rothbard (underlined), two of them are contiguous, the first one being a pentamer and the second one a tetramer. In addition, the amino terminal part of this segment presents tetrapeptides with low occurrence values in the human subset databank.

501 IEWIYRRCKNIPOPVRALLEGFLR 524 24 amino acids

2. Synthesis of Peptides and Conjugates

Peptides were synthesized by the Merrifield solid-phase method (4) _Q n a fully automated peptide synthesizer (AB1 model 430 A, Foster City, CA), according to the strategy tert-butyloxycarbonyl trifluoroacetic acid (tBoc TFA). The tBoc-N-a protected amino-acids were coupled sequentially to a tBoc-Ala-OCH2 - phenylacetamidomethyl (Pa ) resin, a tBoc-VAl-OCH2*-Pam resin and a tBoc-tosyl- Arg-OCH2-Pam resin. The trifunctional amino-acids were protected as follows: Arg (tosyl), Asp and Glu (benzyl ester), Ser and Thr (benzyl ether), Lys (2-chloro- benzyloxycarbonyl), Tyr (di-chloro-benzyl), Tip (for yl), Cys (methylbenzyl), and His (dinitrophenyl). At the end of the synthesis, the protective group of His residue was removed by thiolysis before cleavage and deprotection by "low-high" hydrogen fluoride procedure. The crude peptides were purified by gel-filtration on TSK HW 40s (Merck, Rahway, NJ) and reverse phase HPLC on Nucleosil C\ g. Then peptides were checked for homogeneity by RP-HPLC on Vydac Cjg and thin-layer chromatography, and for identity by amino-acid analysis after total acid hydrolysis.

Peptides 197-216, 393-410 and 501-524 have been covalently linked to bovine serum albumin (BSA) respectively with dimethylsuberimidate, l-(3- dimethylaminopropyl)-3-ethylcarbodiimide and gutaraldehyde (5). After dialysis to remove any uncoupled peptide, the conjugates were characterized to determine the ratio peptide/carrier. The quantity of protein was evaluated with the Bio-Rad Protein Assay. During peptide synthesis, we used ^H-labelled Boc-Leu (Dupont, NEN Research Products, Boston, MA), in order to determine the number of linked peptide by measuring the total radioactivity of the purified conjugate.

3. T Cell Proliferation Assay

The human T-cell clone 32 was maintained in culture by stimulation every 2-3 weeks in the presence of allogeneic irradiated PBMC, phytohaemagglutinin (PHA), and recombinant IL-2 as described ("). For proliferation assays, T cells were incubated with the appropriate concentration of antigen in the presence of cryopreserved cells of the autologous B lymphoblastoid cell line as antigen presenting cellsW. Proliferation was assessed three days later by [-1H]thymidine incorporation as described previously^).

Fig. 1 shows the dose-response curves of the T-cell clone 32 obtained with the three peptides. As can be observed a significant proliferation was detected with the peptide corresponding to residues 197-216, with the maximum response obtained between 10 ^" ^ and 3 x 10 ^" ^ μg/ml. No response was observed with the two other peptides. Significant proliferation was observed in the presence of the fraction F3 or an extract of the Tg34 lysogen (positive controls); no significant proliferation was observed with the extract of the wild lambda gtl 1 lysogen (negative control). The

conjugates were also tested in the proliferation assay; we observed a response in the presence of the conjugate corresponding to the peptide 197-216 only (data not shown).

It can be concluded that the amino acid sequence 197-216 of the Tg34 protein contains the epitope recognized by the T-cell clone 32.

EXAMPLE 2

Construction of recombinant vectors carrying the 54 kDa antigen (Tg34), either complete or restricted to the T-cell epitope (aa 197-216).

A. Vector Construction for Transfer into Vaccinia Virus

The pBluescript KS ⁺ vector carrying a ± 1800 bp EcoRI DNA piece encoding the 54 kDa antigen (pNIV3401) has been described in EXAMPLE 5 of WO 92/11366. From there, a 1556 bp Nhel-Ncol DNA fragment, encompassing most of the coding sequence for the antigen, was recovered and ligated into the BgHI-NcoI sites of the multipurpose cloning vehicle pJRD184 (Heuterspreute £t al, Gene 59. 299-304, 1980) together with a 59 bp BgHI-Nhel synthetic double-stranded oligonucleotide (OL1) coding for the 5' end of the Tg34 DNA sequence (Fig. 2A). The resulting intermediate construct was then cut with EcoRV and Ncol to insert a second 33 bp synthetic double-stranded oligonucleotide (OL2), flanked by EcoRV (5'-end) and Ncol (3'-end), coding for the 3'-end of the Tg34 DNA sequence (Fig. 2A). The final construct, pNIV3404, thus carries the coding sequence for the 54 kDa antigen, including a 5' initiation consensus sequence (Kozak, M.. Cell 44. 283-292, 1986), an ATG initiation codon and a stop codon (Fig. 2B). pNIV3404 was then cut with BglH and EcoRV to recover a 1648 bp DNA module encoding the 54 kDa antigen which was finally cloned into the vaccinia transfer plasmid pULB5213 (a derivative of the standard vaccinia vector pSCl 1; Chakrabati s___ , Mol. Cell. Biology 5. 3404-3409, 1985). The resulting plasmid, pNIV3402, is illustrated in Fig. 2C; it carries the sequence for the Tg34 antigen under the control of the P7 5 vaccinia promoter.

B. Construction of transfer plasmids for expression into Mycobacterium bovis-BCG

1. Preparation of starting materials

a) Plasmid pNIV2229

A MluI-BamHI fragment of 763 bp was isolated from pRIBlOOO (Thole et al..

Infection and Immunity 50. 800-806, 1985). It contains the putative promoter of the P64 (HSP60) mycobacterial antigen. It was cloned into the Hindi site of pUC19 plasmid to generate pNIV2229.

b) Plasmid pNIV3403

The complete coding sequence of Tg34 was transferred as a Bglll-EcoRV fragment of 1600 bp from pNIV3404 between the BamHI and Hindi sites of pUC19, generating pNIV3403.

2. Construction of pNIV3406

A Pstl-MscI DNA fragment was isolated from pNIV2229. It contains the P64 promoter up to the first base (G) following the translation initiation ATG. It was ligated between the EcoRI and Nhel sites of pNIV3403 with a synthetic DNA adaptor (NATBCG1/NATBCG2, Fig. 3 A) coding for the first amino acids of Tg34. A 1300 bp DNA containing the Kan- gene from Tn903 (Pharmacia) was then inserted into the unique PstI site of the resulting plasmid (pNIV3405) to generate pNTV3406. This construction is described in the following figure (Fig. 3B).

3. Construction of pNIV3407 Plasmid pNIV3407 results from the insertion into the unique Smal site of pNIV3406 of a 1500 bp blunted BamHI fragment derived from a temperate mycobacteriophage FRAT1 (isolated previously in our laboratory, Haeseleer, Timm and Jacobs, 1989, Acta Leprologica 7. (suppl.l) 252-253).

C. Vector construction for expression in yeast as particles made of a fusion between the Tg34 T-cell epitope (aa 197-216) and the preS2-S antigen of the hepatitis B virus (HBV)

In order to use HBsAg particles as carriers for the presentation of the Tg34 T- cell epitope to the immune system, the sequence coding for the Tg34 T-cell epitope (aa 196-216) was fused upstream to the DNA encoding the preS2 gene (aa 133-145) and the complete sequence for the S protein of HBV (aa 175-400) (Fig. 4A). Plasmid TCR50 (pRIT13220) was digested by BamHI and EcoRI to recover a 722 bp DNA fragment corresponding to amino acid residues 133-145 of the preS2 protein and 175- 400 of the S protein of HBV (Cabezon et al, Vaccines 90, 199-203, 1990). Synthetic oligonucleotides (64 mers) were used in order to reconstitute the DNA sequence encoding the sequence for the Tg34 T-cell epitope (aa 196-216) and to provide the junction between the fragment obtained above and the plasmid FF6 (pRIT13717) cut by Ncol and EcoRI. The pRIT13717 vector contains a yeast glyceraldehyde-3P- dehydrogenase (TDH3) promoter fragment followed by a yeast DNA fragment

carrying the ARG3 transcription terminator (Fig. 4B). The 2980 bp expression cassette was excised as a BgUI-Sall fragment and inserted into the yeast shuttle vector OJ10 (pRIT12741) to obtain pNIV3409 (Fig. 4C). The pRIT12741 vector is composed of the yeast 2-micron DNA and the yeast LEU2 as marker of selection in yeast (Harford et al, Postgrad. Med. J. 63, 65-70, 1987). The cassette inserted in pNIV3409 is predicted to express a 260 amino acids protein. The hybrid protein consists of an initiation methionine followed by the 20 amino acids of the Tg34 T-cell epitope, the 13 amino acids from the preS2 region and the 226 amino acids from the major HBsAg coding region.

D. Vector construction for expression in CHO cells

a) The 54 KDa antigen

Starting from the plasmid pNIV 3404 (see Example 2A), a 1648 bp Hind III - EcoRV DNA fragment was recovered by digestion and inserted into the pRIT 14073 eukaryotic expression vector derived from pEE 14 (Cockett et al, 1990, Bio/Technology, 8, 662 - 667), cut with Hind III and Sma I. The resulting expression plasmid, pNIV 3412 (Fig. 5) contains, under the control of the major immediate early promoter of the human cytomegalovirus (hCMV - MIE), the sequence encoding the 54KDa antigen described in Example 2 A.

b) The truncated 54 KDa antigen

The amino acid sequence of the 54KDa antigen contains a putative membrane - anchorage domain spanning residues 464 to 485, which may prevent secretion of the recombinant protein by host cells. We have constructed a DNA encoding a truncated form of the 54 KDa antigen wherein the sequence corresponding to residues 464 to 485 has been deleted. The construction proceeded as follows. A 1356 bp Hind III - _____ I DNA fragment was recovered from plasmid pNIV 3402 (see Example 2A) and introduced, together with a 49 bp ≤fς I - Sma I synthetic DNA fragment (Fig. 6a), into the pRIT 14073 expression vector cut with Hind III and Sma I, yielding an intermediate construct. This construct was then digested with Sma I to accomodate, by ligation, into this site, a 151 bp Hpa I - EcoRI DNA fragment obtained by the polymerase chain reaction (primers 1 and 2, DNA of pNIV 3402 as template, Fig 6b). The final construct, pNIV 3414, thus encodes, under the control of the hCMV promoter, a truncated form of the 54 KDa antigen lacking residues 464 to 485 (Fig. 6c).

EXAMPLE 3

Expression of the 54 kDa antigen, complete or restricted to the T-cell epitope (aa

197-216), in host cells

A. Expression in eukaryotic cells via vaccinia virus recombinants

The recombinant transfer plasmid pNIV3402 was transfected into vaccinia- infected CV-1 cells and recombinant viruses were isolated after Bromo-uridine selection and plaque purification on the basis of their blue colour in the presence of X-gal. They will be referred to as VV3402. The human HI 43 fibroblast TK strain was used preferably for plaque assays. The vaccinia virus used to infect cells was of the WR type (origin Borysiewicz, U.K.). The procedure follows that one previously described for the obtention of vaccinia virus recombinants (Mackett M. and Smith G.L., J.Gen. Virology 67. 2067-2082, 1986; Mackett M., Smith G.L., and Moss B., J. Virology 49. 857-864, 1984).

The recombinant vaccinia virus V V3402 was used to infect CV- 1 cells in culture at a multiplicity of infection of 1. Infected cells (about 3.10*-' per assay) and spent culture medium (about 2 ml) were collected between 18 and 48 hours post infection. The presence of the 54 kDa antigen was identified following separation of proteins by polyacrylamide gel electrophoresis in the presence of 12% sodium dodecyl sulphate, transfer onto nitrocellulose membranes and immunodetection. Details of these procedures are given in WO 92/11366.

B. Expression in M.bovis-BCG

Plasmids pNIV3406 and pNIV3407 were electroporated into M.bovis-BCG: the first plasmid was used in its linear form (via Seal digestion) and the second one was used as a circular molecule. M.bovis-BCG recombinants recovered from die transformation procedure were analysed for expression by the immunodetection methods described under subtitles A 13 and A 15 of WO 92/11366.

C. Expression in Yeast

1. Plasmid pNIV3409 was used to transform S.cerevisiae strain DC5 (his3- 11 , his3-15, leu2-3, leu2-112, canl-11 cir° , Harford and Peeters, Curr.Genet. 11:315- 319, 1987) to the Leu ⁺ phenotype. The transformants were named Y420. One of them was grown in 40ml of yeast nitrogen base minimal medium with 2% glucose as carbon source up to mid exponential phase. Cells were washed, collected by centrifugation and broken with glass beads in 300ml of Tris-HCl 25mM pH 8, 4 mM EDTA, 1% Tween 20, 4mM phenylmethylsulfonyl fluoride, 10% propanol-2. The homogenates were centrifuged at 30,000 x g for 30 min to obtain crude extracts. The supernatant (± 50μg of total protein) was analyzed on 12.5% SDS-polyacrylamide gel. Separated protein were then blotted on nitrocellulose (Towbin et al., Proc. Natl. Acad. Sci. USA 76:4350-4354, 1979) and the sheet was incubated with monoclonal antibodies directed against the HBsAg (mAb HBS1, SmithKline Beecham) or the

preS2 region (mAb S2.5, SmithKline Beecham). Detection of die primary antibody was by the alkaline phosphatase method. The anti-IgG second antibodies conjugated with alkaline phosphatase were detected by NBT (nitroblue tetrazolium) and BCIP (5- bromo-4-chloro-3-indolyl phosphate). A protein of about 28 kDa not present in the negative control (Y1017) was revealed with the mAb's HBS 1 and S2.5. These results indicate that the yeast strain Y420 expresses a fusion protein of MW 28 kDa that consists of an initiation methionine followed by the 20 amino acids of die Tg34 T-cell epitope, the amino acids 133-145 from the preS2 region and the 226 amino acids from the major HBsAg coding region.

2) Formation of particles

Strain Y420 was grown in selective medium (800ml + 80μg/ml histidine) to late log phase. Cells were collected by centrifugation, washed with cold PBS and resuspended in 5ml of 50mM sodium phosphate buffer (pH 8.1), 4mM EDTA (ethylenediaminetetraacetic and disodium salt), 1% Tween-20 (polyoxyethylene sorbitan monolaurate), 4mM PMSF (phenylmethylsulfonylfluoride) and 10% isopropanol. Cells were disrupted by passage twice through a French pressure cell at 20,000 psi. The suspension was centrifuged for 30 min at 30,000 xg. The total protein cencentration of the supernatant liquid was measured by the Bio-Rad protein assay (n° 500-0001) based on the Bradford dye-binding procedure (Bradford, M. (1976) Anal. Biochem. 72, 248). The presence of HBsAg was assayed using an ELISA kit (Enzygnost-HBsAg micro) commercially available from Behringwerke, Marburg, Germany).

7ml of crude cells extracts were centrifuged to equilibrium in 1.5M CsCl, 25mM sodium phosphate buffer, pH 7.4, for 68 hours at 40,000 xg in a Beckman 60 Ti rotor. Collected fractions were assayed for the presence of HBsAg with the ELISA method described above. HBsAg was found to equilibrate to 1.22 g/cm ³ . Immunoblot analysis of the gradient fraction confirmed the ELISA results (Fig.7A). 0.5ml of the material from the CsCl peak fraction dialysed against PBS was subjected to a velocity sucrose gradient for 13 hours at 38,OOOg in a SW40 rotor.

Each fraction was analyzed for HBsAg antigenicity and for the presence of the hybrid protein respectively by ELISA (Enzygnost-HBsAg micro) and by immunoblot with the mAb HBS1 (Fig.7B). The results show that the hybrid protein assembles into lipoproteic particles.

C. Expression in yeast

3) Plasmid.pNI V3417 , ρNIV3421 and pNIV3423 were used to transform S. cerevisiae strain DC5 to the Leu ⁺ phenotype. The transformants were named

respectively YN479, YN478 and YN481. Cultures and analysis of crude extracts were released as described above. Proteins in the culture medium was precipitated by

TCA (trichloroacetic acid).

D. Expression in CHO cells (stable transformants)

a) The 54 KDa antigen

Plasmid pNIV 3412 was transfected by calcium phosphate coprecipitation into the CHO - Kl cells, using 20 ug DNA per 1.25 10 ⁶ cells. The CHO - Kl cells were grown in GMEM - S medium (Gibco). Transfectants were selected by addition of 25 uM methionine sulfoximine two days after transfection. After 10 to 14 days, resistant colonies were picked up, transferred into 96 well plates, and later into 24 well plates and subsequently to 80 cm^ flasks. GS transfectants were assayed for die 54KDa antigen when cells reached about 80% confluency. The procedure follows the one described in Cockett & al, Bio/Technology 8, 662 - 667 (1990).

b) The truncated 54 KDa antigen

Procedures for the expression of the truncated 54 KDa antigen followed tiiose described above.

E. Vector construction for expression of the 54KD antigen in yeast

A 1648 bp Hind III-Bgiπ DNA fragment was recovered from pNIV3402 and introduced blund-ended in the correct orientation for expression into the blunt ended BamHI site of the expression vector ρRTT13145 (TCM97) (FIG.7). The resulting plasmid pNIV3417, contains, under die control of the ARG3 promoter, the sequence encoding the 54KD antigen described in example 2A.

Secretion from the yeast S.cerevisiae of heterologous proteins with their own signal sequence has observed in some but not all cases. However, their secretion has mostly achieved by fusing signal sequences of yeast secretory proteins to the coding region of the foreign protein (see Romanos et al (1992) Yeast 8: 423-488). The signal sequence for the yeast pheromone MFα-1 has proved particularly productive in this respect (see Brake) (1989) in Yeast Genetic Engineering 269-280).

Thus, we have constructed a DNA encoding for the signal sequence of MFα-1 followed by die amino acid residues 28 to 538 of Tg54 protein. The construction proceeded as follows. Plasmid pNIV3402 was digested by Rsal and Sail to recover a 470 bp DNA fragment. Synthetic oligonucleotides (65 mers) were used in order to reconstitute the DNA sequence encoding the signal sequence of MFα-1 (19 amino acids) and to provide the juction between the fragment obtained above and die plasmid pUC19 cut by Hind Hl-Sall (FIG.δa). The resulting intermediate construct

was then cut the Sail and BamHI to insert a Sall-BglH DNA fragment from pNIV3402 to give the pNIV3416 (FIG.8b). This one was cut by Hind III and ECΩRI to recover a 1626 bp DNA fragment, blunt-ended and introduced in the blunt-ended

BamHI site of the pRIT 13145 (TCM97). The final construct, pNIV3421, thus encodes, under the control of the ARG3 promoter, the signal sequence of MFα-1 fused to residues 28 to 538 of the 54KD antigen (FIG. 8c).

As die putative membrane-anchorage domain of the 54KD antigen (residues

464 to 485) may prevent secretion of the recombinant protein, we have constructed a

DNA encoding a truncated form of the MFα-l-Tg54 antigen wherein the sequence corresponding to residues 464 to 485 of the Tg54 antigen has been deleted. Thus, the Sall-BgHI DNA fragment from pNIV3414 was introduced in the pNIV3415 cut by Sail and BamHI to give the pNIV3422 (FIG.9a) The resulting plasmid was tiien digested by Hind III and EcoRI to recover a 1400 bp DNA fragment, blunt-ended and introduced in the blunt-ended BamHI site of the pRIT13145 (TCM97). The resulting expression plasmid, pNIV3423, contains, under the control of the ARG3 promoter, the sequence encoding the signal sequence of MFα-1 followed by residues 28 to 465 and 486 to 538 of the 54KD antigen (FIG.9b).

References 1. Margalit H„ Spouge J.L., Cornette J.L., Cease K.B., Delisi C, and Berzofsky J. A. (1987). Prediction of immunodominant helper T cell antigenic sites from the primary sequence. J. Immunol., 138, pp 2213-2229.

2. Rothbard J.B., and Taylor W.R. (1988). A sequence pattern common to T cell epitopes. EMBO J., 7, pp 93-100. 3. Claverie J-M., Kourilsky P., Langlade-Demoyen P., Chalufour-Prochnicka A., Dadaglio G., Tekaia F., Plata F., and Bouguerlet L. (1988). T-immunogenic peptides are constituted of rare sequence patterns. Use in the identification of T epitopes in die human imunodeficiency virus gag protein. Eur. J. Immunol., 18, pp 1547-1553.

4. Kent S.B. (1988). Chemical synthesis of peptides and proteins. Annu. Rev. Biochem., 57, pp 957-985.

5. Van Regenmortel M.H.V., Briand J.P., Muller S., and Plaue S. (1988) Syndietic polypeptides as antigens, ed. Elsevier, pp 95-130.

6. Saavedra R. and Herion P. (1991). Human T-cell clones against Toxoplasma gondii: production of interferon-gamma, interleukin 2, and strain cross-reactivity. Parasitol. Res., 77, 379 - 385.

7. Saavedra R., De Meuter F., Decourt J-L., and Herion P. Human T-cell clone identifies a potentially protective 54-kilodalton protein antigen of Toxoplasma gondii cloned and expressed in E. coli. J. Immunol. \__1, 1975 - 1982 (1991).