IE67139B1 - Synthesis of human virus antigens by yeast - Google Patents

Description

biosynthesis of an antigen of human Hepatitis B virus (HBv) by yeast, brought about by an application of recombinant DMA techniques. Hepatitis B virus is recognised as a major, world-wide public health problem. In addition to the widespread incidence of viral hepatitis, and the persistence of asymptomatic carrier states. Hepatitis 3 virus has been implicated in the etiology of hepatocellular carcinoma. For a recent review of the molecular biology of Hepatitis B virus, see Tiollais, P., et al., Science 213, 406 (1981) .

A major effort in current research is to produce a suitable vaccine to provide protective immunity against viral infection» One line of approach to preparing a suitable vaccine has involved attempts to purify the principal antigenic component of the virus, the surface antigen» Hereinafter, the symbol HBsAg is used to identify H3V surface antigen obtained from preparations of intact virus (Dane particles) or purified from the serum of hepatitis carriers. Skelly, J. et al., Nature 290, 51 (1981) have reported the purification of water-soluble protein micelles of purified HBsAg» A significant limitation of this approach is that the amount of material which can be prepared depends upon the availability of donors» No technique is known for growinq the virus in culture; therefore, in addition to limitations in the amount of source material, there is a risk of contamination of the vaccine with active virus or other components of donor serum, and a possible heterogeneity in the products obtained form various donors.

A second approach has been the attempt to synthesize peptides eliciting antibodies against HBsAg based upon the amino acid sequence of the protein comprising the surface antigen (S-protein) and model studies predicting the most likely antigenic determinants» See, e.g. R. A. Lerner et al,, Proc. Nat. Acad» Sci. USA 78,3403 (1981). Such work is in a highly preliminary stage, and it may be difficult to assess whether the approach can produce antigens having a practical degree of immunogenicity in a cost-effective manner. 4. Ά third approach, employing recombinant DNA techniques, is the synthesis of S-protein, HBsAg or an immunologically reactive equivalent by a microorganism, by endowing a microorganism with genetic capability to produce S-protein, HBsAg or an immunologically reactive equivalent in large amounts, in the absence of other viral gene products. This approach eliminates the possibility of contamination by virus or other viral components and permits large-scale production with economies of scale. Furthermore, it is possible, through appropriate manipulations of the genetic material, to modify the sequence of the protein comprising the vaccine, in order to modify its side effects, or make the vaccine polyvalent. Toward this end, the entire genome of HBV has been cloned in Ξ. coli and its entire nucleotide sequence determined (Charnay, P., et al., Nucl. Acid Res, 7, 335 (1979) ,» Galibert, F., et al., Nature 281, 646 (1979) ; Valenzuela, P., e t a 1., Animal Virus Genetics (B. Fields, R. Jaenisch and C. F. Fox, Eds.) Academic Press, New York, N.Y. (1980) , page 57. A single region of the genome was found to code for the S-protein and also for a large pre-sequence of 163 amino acids. The structure of HBsAg is believed to consist of two S-protein chains joined by intermolecular disulfide bonds and held in a prescribed confirmation by additional intramolecular disulfide bonds. One of the two chains appears to be glycosylated. In the serum of carriers, HBsAG frequently appears in the form of spherical particles with a mean diameter of 22 nm, which are thought to aggregates of the S-protein dimers just described, and possibly contain lipids.

In the viral envelope, HBsAg is associated with the lipid-containing viral envelope, which is believed to be derived from membrane components of the host cell.

The antigenicity and immunogenicity of HBsAg depend upon several factors, not all of which are well understood. It has been observed that reduction of the disulfide bonds reduces antigenicity and immunogenicity markedly (Mishiro, S. et al., J„ Immunol. 124f 1589 (1980)). Therefore, the tertiary configuration contributed by the intramolecular and intermolecular disulfide bonds is thought to contribute to antigenicity and immunogenicity. T’he contribution of other factors, such as the extent and nature of glycosylation and association with lipid is unclear, although all are thought to contribute to some degree. Aggregation into particles such as the above-mentioned 22 nm particles is thought to contribute significantly to enhancing immunogenicity.

The S-protein has been synthesized in E. coli in the form of a fusion protein (Edman, J. C. et al., Nature 291, 503 (1981)). The product included 183 amino acids of pre-beta lactamase, 5-10 glycine residues, and 204 amino acids of S-protein lacking 22 amino acids of the amino terminal end. The fusion protein was immunoprecipitable with anti-HBsAg IgG.

Since it is known that S-protein dimers Mishiro et al», supra) and 22 nm particles incorporationg HBsAg (Cabrall, G. A. et al., J, Gen.

Virol. 38, 339 (1978)) are more antigenic than the associated S-protein, it would be highly desirable to find a biological system capable of producing HBsAg or an immunologically reactive equivalent directly, in substantial quantities.

The steps in converting S-protein to HBsAg or to 22 nm particles are not fully understood, nor is it known to what extent they are host cell-specific. Furthermore,, the S-protein gene appears to code for an unusually long pre-sequence of 163 amino acids,» whose functional significance, if any,, is unknown. In fact it is not known whether the pre-sequence is actually translated in the virus-infected cell» Yeast (Saccharomyces cerevisiae} was chosen as a host cell in which to attempt the expression of HBsAg for the following reasons: Yeast is readily grown in culture in large quantities. In fact, the technologv of yeast culture on a large scale is well understood. Also, yeast is eucaryotic, so it was hoped that some of the post-translational processing steps which are carried out in a normal host cell might be carried out in yeast. Because of the complex post-translational events that convert S-protein to HBsAg, some of which may be host-cell specific, the nomenclature adopted herein is intended to distinguish different antigenic forms recognized from the work herein disclosed» The unprocessed translation product of the structural gene for surface antigen is termed S-protein. The antigen isolated from plasma of infected donors, from Dane particles or from human hepatoma cell cultures, is termed HBsAg. The expression product of the surface antigen gene in yeast is termed γ-HBsAg. The term, immunologically reactive equivalent of HBsAg, is a general term for any immunologically cross-reactive composition comprising S-protein or a portion thereof, of which Y-HBsAg is an example.

Yeast has never previously been used for expression of the genes of a virus which normally multiplies in a different organism. Prior art attempts to express heterologous proteins in yeast have yielded mixed results. An attempt to express rabbit globin, under control of its own promoter, appears to have been unsuccessful in translation of the protein (Beggs, J. D. et al., Nature 283, 835 (1980))» A gene coding for a Drosophila gene has been reported capable of complementing a yeast ade 8 mutant, under conditions of selective pressure for genetic complementation. Isolation of a functional protein from the yeast strain was not reported. The gene for human leukocyte interferon has been expressed in yeast, under control of the yeast ADHl (alcohol dehydrogenase) promoter. In that instance, successful production of an active protein did not require post-translational processing or assembly of components.

DNA transfer vectors suitable for transfer and replication in yeast have been developed (Broach, J. R. et al». Gene 8, 121 (1979) ? Hartley, J» L. et al», Nature 285, 850 (1980) . Most yeast vectors in current use are derived from E. coli vectors, such as pBR322, into which have been inserted a yeast origin of replication. Two types of yeast replication origins are available. The first, derived from a ubiquitous naturally-occurring yeast plasmid, commonly referred to as the 2 micron circle, confers the ability to replicate independently of yeast chromosomal DNA. Another class of vectors contains a replication origin sequence termed arsl (autonomous replication sequence), derived from the yeast chromosomal replication origin, which also provides autonomous replication capability. Because both bacterial and yeast replication origins are present in the same vector, they can be used in either organism. Selection may be provided for in bacterial systems by the inclusion of antibiotic resistance genes, such as the ampicillin and tetracycline resistance genes of pBR322. Selection in yeast systems typically may be provided for by including a yeast gene complementing a mutation in a suitable auxotrophic host strain. The studies reported herein conveniently utilize yeast vectors containing a promoter isolated from the yeast gene coding for alcohol dehydrogenase (ADH1). (Bennetsen, J. L. et al., J. Biol. Chem. Vol. 257, p. 3018 (1981).

The ADHl promoter region was isolated from the 5’-flanking region of the yeast ADHl gene. A fragment containing approximately 1600 base pairs of the ADHl sequence extending from position -1550 to 4-17 within the coding region was fused to the yeast CYCl coding sequence. Studies on transcription of the attached CYC1 coding sequence demonstrated that transcript starting specificity could be transferred from one yeast gene to another. Smaller fragments, lacking all of the ADH coding region, have subsequently been constructed, and shown to be functional in the expression of human interferon.

One such fragment, designated 921, is terminated after position -9, and was employed in the present studies.

Because substances reactive with anti-HBsAg . antibody exist in several forms, a nomenclature has' been adopted herein to distinguish these forms. The translation product of the HBV surface antigen gene is termed the S-orotein. S-protein has 226 amino acids whose sequence has been inferred from the nucleotide sequence of its gene and by partial sequence analysis. HBsAg as used herein includes the major surface antigenic component of HBV found in infected patients' serum and in Alexander cells, a hepatocellular carcinoma cell line which synthesizes and excretes 22 nm HBsAg particles (Alexander, J. J. et al., S-.Afr. Med, J, 50, 1124 (1976). Both S-protein and HBsAg are antigenic, however, the latter is more reactive against anti-HBV antibody and is considered more immunogenic. Since the structure of HBsAg is not fully characterized, and the contributions to antigenicity and immunogenicity of various modifying steps not fully understood, the term HBsAg is used herein to include any modified form of the S-protein which contributes to its antigenic and immunogenic properties, including, but not limited to, dimerization, glycosylation, and particle assembly.

General information on the gene coding for hepatitis 3 virus antigen and its expression in a host cell is given in the following documents; EP-A-73 657; Proc.Natl.Acad.Sci.USA, 77 (1980) p.45494553; EP-A-38 765; and Nature, 280 (1979) p.815-819.

The present invention relates fo synthesis of HBsAg in yeast. Yeast expression vectors comprising a yeast promoter, ADH1, have been constructed. The region of the HBV genome coding for the S-protein, excluding a possible 163 amino acid presequence, has been transferred to the yeast expression vector.

Using the described veast vector, the successful synthesis of HBsAg by yeast has been achieved. The product is antigenic (reactive with anti-HBsAg), and a substantial portion is found associated with particles identical in electron microscopic appearance to those found in the serum of HBV-infected patients and in Alexander cells but having a smaller particle size diameter. The HBsAg synthesised by yeast has identical sedimentation behavior to purified, naturally-occurring HBsAg particles purified from Alexander cells as measured by sucrose gradient sedimentation. The present invention demonstrates synthesis and assembly of a higher ordered multi-component structure resulting from expression of a heterologous DNA coding segment in a microorganism.

The present invention is believed to be the first instance of biosynthesis and particle assembly of a virus protein in a heterologous host, where the heterologous host (in this instance yeast) is far removed on an evolutionary scale from the normal host (man). The invention was made possible by the development of autonomously replicating DNA transfer vectors for yeast and also by the cloning and characterization of the HBV genome in bacteria. In the present invention, the promoter for the yeast 1 ADHl gene was used to provide a high level of transcription of the inserted S-protein coding region. In principle, any yeast promoter could be employed instead, preferably an active promoter which provides a high level of transcription. Other suitable active promoters of yeast include those for glyceraldehyde 3-phosphate dehydrogenase, aldolase, pyruvate kinase and phosphoglycerafce kinase. It may be that heterologous promoters, such as the HBV S-protein promoter, may also be employed. However, at present, the use of yeast promoters is preferred.

Other proteins of HBV, such as the core antigen, should also be synthesizable by employing the principles and techniques of the present invention. Beyond that, the invention is applicable and will be particularly advantageous in any system where posttranslational processes are desired for making a biologically functional end product, including glycosylation, particle assembly, and possibly specific protein cleavage reactions.

The S-protein gene has three potential points for initiation of translation. The first two are AUG codons located approximately 70 and 90 base pairs from the putative HBV S-protein promoter. The third, which begins the known coding sequence of mature S-protein, is located 522 and 489 base pairs from the first and second, respectively. The third potential start point is therefore much further away from the HBV S-protein promoter. At present, it is not known which of the AUG codons is the actual starting point for translation. If translation is initiated at either the first or second potential 2 start codons, the transcript either comprises the coding sequence for an unusually long leader of 163 amino acids which must be removed by posttranslational processing, or it constitutes an unusual intron which is removed by posttranscriptional processing. If the third AUG is the actual initiation point, then there is an unusually large spacing between the promoter and the start codon.

The data presented herein demonstrate that Υ-HBsAg is a particulate, immunologically cross-reactive equivalent of HBsAg which differs from the latter in several ways, although its morphological appearance in the electron microscope is similar to HBsAg. Furthermore, the data demonstrates that Υ-HBsAg may be at least as antigenic, per unit weight, as HBsAg, and Y-HBsAg may be at least as effective as HBsAg in eliciting antibody reactive to HBsAg in rodents and in primates.

The rate of expression of the S-protein coding segment may be enhanced by a variety of means. These include modifying the expression vectors to optimise the spacing between the promoter and the start codon of the coding segment, in order to optimize the rate of translation initiation. The addition of a terminator sequence, which directs termination of transcription at a point in the 3s untranslated region following the stop condon of the coding segment, enhances expression, presumably by stabilizing the mRNA transcripts. In the absence of a termination signal, it has been observed that 3 optimising the length of the 3s untranslated region itself enhanced expression» The adoption of means to enhance vector stability also increases the yield of the expression product from a culture. Many vectors adapted for cloning in yeast include genetic markers to insure growth of transformed yeast cells under selection pressure, for example, by including a ^'RPl gene to permit the growth of a trpl~ host in medium lacking tryptophan. Host cell cultures containing such vectors may contain large numbers of untransformed segregants when grown under nonselective conditions, especially when grown to high cell densities. Therefore, it is advantageous to employ expression vectors which do not require growth under selection conditions, in order to permit growth to high densities and to minimise the proportion of untransformed segregants. Vectors which contain a substantial portion of the naturally occuring two circle plasmid are able to replicate stably with minimal segregation of untransformed cells, even at high cell densities, when transformed into host strains previously lacking two micron circles. Such host strains are termed circle zero (cir°) strains. Additionally, the rate of cell growth at low cell densities may be enhanced by incorporating regulatory control over the promoter such that the expression of the S- protein coding region is minimized in dilute cultures such as early to middle log phase, then turned on for maximum expression at high cell densities. Such a control strategy increases the efficiency of cell growth in the fermentation process and further reduces the frequency of segregation of untransformed cells.

In the examples that follow, many of the techniques, reactions and separating procedures are already well known in the art. All enzymes, unless otherwise stated,, are available from one or more commercial sources, such as New England BioLabs, Beverly, Massachusetts; Collaborative Research, Waltham, Massachusetts; Miles Laboratories, Elkhart, Indiana; Boehringer Biochemicals Inc., Indianapolis, Indiana and Bethesda Research Laboratory, Rockville, Maryland, to mention a representative few. Buffers and reaction conditions for restriction enzyme digestion were used according to recommendations supplied by the manufacturer for each enzyme.

Partial digestions with restriction enzymes were carried out using a reduced enzyme concentration which was predetermined from preliminary experiments for each enzyme batch. Standard methodology for ✓ other enzyme reactions, gel electrophoresis separations and E. coli transformation may be found in Methods in Enzymology, Vol. 58, Ray Wu, Ed., Academic Press (1379). Transformation of yeast protoplasts was carried out essentially as described by Beggs, (Nature 275,, 104-109 (1978) „ E. coli strains useful for transformation include X1776; K12 strain 294 (ATCC No. 31445); RRl and ΚΒ101» Yeast strains XvSlO-Sc having the genotype (a ade2 adeo leu2 lysl trpl canl) and GM-3C-2, Paye, G. et al., Proc. Nat. Acad. Sci. USA 78, 2258 (1981) Genotype: (a Leu2 Trpl His4 CYC1-1 CYP3-1), were used for yeast transformations.

Bacteria were grown and selected according to procedures described by Miller, J. H., Experiments in Molecular Genetics, Cold Spring Harbor Laboratory, Cold Sprig Harbor, N.Y. (1972). Yeast were grown on the following media: YEPD contained 1% (w/v) yeast extract, 2% (w/v) peptone, and 2% (w/v) glucose; and, in the case of plating media, 3% (w/v) agar. YNB plus CAA cotained S.7 grams of yeast nitrogen base (Difco Laboratories, Minneapolis, Minnesota) , 10 mg of adenine, 10 m of uracil, 5 g casamino acids (CAA) (Difco), 20 g glucose; and, in the case of plating media, 30 g agar per liter. Selection for tryptophan prototrophy was made on plates containing 6,7 g yeast nitrogen base (lacking amino acids), and supplemented for all growth requirements of the strain to be transformed except tryptophan.

EXAMPLE 1 Construction of yeast vectors. Two yeast vectors have been constructed, one having an arsl replication origin, the other comprising a 2 μ circle replication origin.

The plasmid pFRP-921 has been previously described by Hifczeman et al., Nature 293, 717 (1981) . The vector contains the ampicillin and tetracycline resistance genes and replication origin of bacterial plasmid pBR322, the yeast arsl replicaton origin and trpl gene together with the ADHl promoter fragment, designated 921, terminated after position -9 in the nucleotide sequence. A map of pFRP-921 is shown in Figure 1.

Plasmid ρΜΑ5β contained the sequence of bacterial plasmid pBR322, a yeast trpl gene for selection in yeast, the yeast 2μ circle replication origin, and an ADH1 promoter fragment designated 90S, terminated after nucleotide -15 at the 33-end. Steps and construction of pMA56 are outlined as follows and diagrammed in Figure 2.

Plasmid YRp79 (Stinchcomb, D. T. et al., Nature 232, 39 (1979)) containing the yeast trpl and arsl sequences inserted at the EcoRI site of pBR322 was used as starting material. As a result of inserting the yeast sequences, the plasmid contained two BcoRI sites (figure 2). One of these was deleted by partial digestion with EcoRI endonuclease, to digest, on the average, only one of the two sites per molecule. The resulting unpaired ends of the linear molecules were filled in by the reaction catalyzed by DNA polymerase I (Klenow fragment), and the resulting blunt ends were rejoined in a DNA ligasecatalyzed reaction to re-establish a closed circular DNA molecule. One of the resulting plasmids, designated pFRT, was chosen since if had retained the BcoRI site adjacent to the arsl region.

As shown in Figure 2, the arsl replication origin was bounded by a Pstl site and an EcoRI site. At the same time, the plasmid YEpl3 (Broach et al.., supra) contained the 2μ circle replicaton origin similarly bounded by a Pstl site and an EcoRI site. Therefore, cleavage of pFRT and YEpl3 by Pstl and EcoRI endonucleases yielded, respectively, a large linear fragment lacking a yeast replication origin from pFRT and a small DNA fragment comprising the 2μ circle replication origin from YEpl3. Plasmid pFRT was digested with Pstl endonuclease under partial digestion conditions to reduce the frequency of cleavage at the Pstl site within the ampicillinresistance gene. The desired fragments were purified by preparative gel electrophoresis, mixed together and covalently joined in a DNA ligase-catalyzed reaction. The resulting plasmid, designated pMW5, was selected for ability to confer ampicillin resistance.

The ADHl fragment 906 was inserted into pBR322 between the BamHI and EcoRI sites., The promoter fragment was released by digestion with BamHI and EcoRI endonucleases. A fragment of 1.5 kilobases (kb), the ADH1-906 fragment, was isolated by preparative gel electrophoresis., Plasmid MW5 was simiarly digested with EcoRI and BamHI endonucleases. The large fragment, having an EcoRI-specific end and a BamHI-specific end, was isolated by preparative gel electrophoresis, mixed with the ADH1-90S fragment, and covalently joined by a DNA ligase-catalyzed reaction. The resulting plasmid, designated pMA56 and diagrammed in Figure 2, was selected by ampicillin resistance in E. coli.

Plasmids pFRP-921 and pMA5o are structurally similar, differing primarily in having an arsl replication origin (pFRP-921) or a 2μ circle replication origin (pMA56), respectively. In addition, the ADHl promoter fragments differ slightly, as described. Both are similar in having bacterial replication origins and selection markers for growth in E- coli. Both contain a yeast TRP1 gene to permit selection in yeast trpl host strains. i 8 EXAMPLE 2 Construction of a yeast plasmid containing the S-protein coding region. Analysis of the nucleotide sequence of HBV DNA reported by Valenzuela, P. et al. in Animal Virus Genetics, Academic Press, New York, N.Y (1980), pp. 57-70, showed the location of the S-protein coding region. The region is contained within the Tad-HpaI fragment of 835 base pairs length. This fragment includes 26 base pairs preceding the AUG codon for the N-terminal methionine of the S-protein. (Most of the region coding for the putative presequence described supra, as well as the first two AUG codons, are missing from the Tael-HpaI fragment). The fragment also contains the entire S-protein coding region (678 bp), a TAA stop codon, and 128 bp following the stop codon. 675 bp Approximately 500 ug of DNA from plasmid pHBV-3300 (Valenzuela, P. et al., Nature 280, 815 (1969)) were digested to completion with a combination of the restriction enzymes EcoRI and Hpal. Approximately 60 ug of the fragment EcoRI-Hpal of 965 base pairs were isolated by preparative gel electrophoresis in agarose. This fragment was then digested to completion with the restriction enzyme Tad. Approximately 30 ug of the 835 base pair Tael-Hpal fragment were isolated by preparative gel electrophosresis in agarose. 9 The 835 bp fragment was treated to provide EcoRI specific ends by addition of EcoRI linker oligonucleotides (obtained commercially from Collaborative Research, Waltham, Massachusetts) . The linker oligonucleotides were joined to approximately 3-5 ug of the fragment by blunt-end ligation catalyzed by T4 DMA ligase. The fragment was then digested with EcoRI endonuclease to cleave unreacted and self-ligated linkers and to produce EcoRIspecific unpaired (’’sticky) ends.

Plasmid pFRP-921 was digested with EcoRI endonuclease and treated with alkaline phosphatase to prevent self-ligation (Shine, J., U.S. Patent No. 4,264,731). A mixture of the Tad-HpaI fragment with EcoRi ends and EcoRI-digested pFRP-921 was incubated with DNA ligase to form covalently closed circular DNA having the S-protein coding fragment inserted in the yeast vector. The resulting plasmids were used to transform E. coli, selecting for ampicillin resistance. Both possible orientations of the S-protein coding sequence with respect to the yeast promoter were isolated and characterized by the cleavage products resulting from treatment with a restriction enzyme acting on an asymmetrically located site within the HBV insert. Plasmids pHBS-11 (in correct orientation) and pHBS-12 (opposite orientation) were selected and amplified in E. coli.

The foregoing procedure was employed as described, using pMA56 instead of pFRP921 for inserting the Tacl-Hpal S-protein coding fragment into pMASS at the EcoRi site. Two plasmids were isolated and characterized, pHBS-16 containing the .2 Ο viral gene in the correct orientation with respect to the ADH1 promoter, and pHBS-20 having the S-protein gene in the opposite orientation.

EXAMPLE 3 5 Synthesis of HBsAg in yeast. Protoplasts of the yeast-recipient strain XVS10-8C or GM3 C-2, were separately incubated with DMA from each of the four plasmids, pHBS-11, pHBS-12, pHBS-lS and pHBS-20, under the transformation conditions described, and lu plated on agar plates in medium lacking tryptophan. Surviving colonies, transformed to tryptophan prototrophy, were isolated. To test for HBsAg synthesis, yeast strains tranformed with each of the four plasmids were separately grown in liquid cultures, in medium lacking tryptophan, and harvested in mid-log phase. The cells were collected by centrifugation, and cell extracts were prepared by grinding the cells with glass beads in a buffer of 0.01M sodium phosphate (pH 7.4) containing 0.01M beta20 mercaptoethanol, and 0.1% (v/v) NP-40 detergent [polyoxyethylene (9)octaphenol]. The presence of KBV surface antigen was assayed using a radioimmunoassay kit commercially available from Abbot Laboratories., North Chicago, Illinois. Qualitatively, the plasmids containing the surface antigen coding fragment in correct orientation, pHBS-11 and pHBS-lS, produced readily detectible amounts of surface antigen, whereas no detectible surface antigen was found in extracts of cells transformed with pHBS-12 or 3U pHBS-20. Quantitatively, a 200 ml culture of XV-610-8C containing pHBS-16 produced 1-2 ug of surface antigen protein. Cells containing pHBS-ll produced 1/2 to 1/3 as much surface antigen, possibly attributable to a lower copy number per cell of plasmids having the arsl replication origin.

All cell extracts were analyzed by sucrose gradient sedimentation. Pre-formed 5% (w/v) to 30% (w/v) sucrose gradients were layered with an extract of XV610-8C/pHBS-16 cells prepared as described, and control gradients were layered with a preparation of HBsAg purified from an Alexander cell culture. The gradients were centrifuged in a swinging bucket rotor for 8 hours at 27,000 rpm. After centrifugation, fractions were collected and assayed by the above-described radioimmunoassay. Surprisinaly, HBsAg synthesized by yeast was found to have exactly the same sedimentation properties as HBsAg isolated from Alexander cells. A sedimentation value of approximately SOS was calculated for both HBsAg preparations.

HBsAg synthesized by yeast was purified by a combination of equilibrium centrifugation in cesium chloride and sedimentation in a sucrose gradient. 100 Ml of cells grown to an O.D. of 2.0 at 650 nm were harvested by centrifugation to yield 0.150 ml packed cells. The cell extracts were prepared by grinding the cells with glass beads (as described above) such that, after centrifugation at 6,000 rom for 15 minutes to remove cell debris, a total volume of 0.5 ml of extract resulted. The extract contained about 30 mg/ml protein and a total of about 1 pg 2 HBsAg. The extract was layered on a discontinuous cesium chloride gradient from 1.1 g/cm3 to 1.4 g/cm3, and centrifuged in a swinging bucket rotor (SW41, Beckman instruments, Fullerton, California) at ,000 rpm for 24 hours. After centrifugation, fractions were collected and assayed as before. A control tube containing Alexander cell HBsAg was identically treated, as a marker. Yeast HBsAg comigrated with the Alexander cell HBsAg peak, with a 3 bouyant density of 1.1S g/cm . Fractions containing yeast HBsAg were pooled, dialyzed and loaded on a 5-30% (w/v) preformed sucrose gradient, and centrifuged at 30,000 rpm for 36 hours. Again, as previously observed, the peak of yeast HBsAg coincided exactly with HBsAg from Alexander cells. Pooled peak fractions had a total protein concentration of 0.01 mg/ml and an overall yield of HBsAg of 15%.

From the sedimentation data, it was apparent that HBsAg was synthesised in yeast in the form of particles or aggregates. The nature of these particles was further characterized by electron microscopy. HBsAg particles synthesized from yeast and purified as described were adsorbed onto carbon film grids and stained with uranyl acetate stain (2% (w/v) for 7 minutes). Under the electron microscope, particles of HBsAg synthesized in yeast were observed, had an identical appearance but smaller diameter compared with HBsAg from Alexander cells. In these studies Alexander cell HBsAg particles had a diameter of about 20 nm, whereas Y-HBsAg particle diameter was from about IS to about 17 nm (see figure 3). These 3 results are believed to be the first demonstration of assembly into a higher order structure of a heterologous protein in a microorganism host.

Higher yields, up to five-fold, have been obtained using as host the yeast strain GM3C-2. The strain is a petite strain whose increased dependence upon carbohydrate metabolism may result in a higher activity for the ADH promoter. The high expression level observed in GM-3C-2 may also be the result of using a modified vector, pHBS-25, containing the S-protein coding fragment flanked by an ADH1 promoter fragment and an ADH termination fragment.. The Tael-HpaI HBV coding segment was fitted with Hindlll oligonucleotide linkers and joined at the 5s-end to the ADHl promoter fragment ADH1-90S terminated in a Hindlll linker sequence at its 3" end.., The 38-end of the HBV segment was joined to a 450 bp HindIII-BamHI fragment of the ADHl gene containing the coding region for the 43 C-terminal amino acids of ADH, the stop codon TAA and part of the 3’ untranslated region (see Bennetzen et al-, J· Biol., Chem., Vol. 257, p. 3018 (1982), such that both fragments were oriented in the same direction of transcription. The resulting composite segment, ADH-H-5-S-proteinADH-terminator, was flanked by BamHI sites, permitting insertion at the BamHI site of pMA56.

When inserted such that the ADH terminator section was adjacent to the ADH1-906 promoter fragment, the resulting vector was designated pHBS-25.

The construction of vectors of the 25 series analagous to pHBS56-3 and pHBS56-5 (Example 6) is readily accomplished using the composite genes for 4 S-protein flanked by the ADH promoter and terminator segments described in Example 6. The composite gene derived from pHBSXS-3 is inserted at the Sphl site of pt'lASo, linearised by Sphl digestion and treated with alkaline phosphatase to prevent reconstitution of pMA56 in the absence of the inserted composite gene, as described by Shine, U.S, Patent 4,264,371. In this construct ion, pHBS25-3 and pHBS25-5 differ from pHBS25 in that the composite gene is inserted at the Sohl site of pMA56 rather than the nearby BamHl site of pMA56.

The transfer vector pHBS-16 and a yeast strain comprising the strain XV610-8C transformed by plasmid pHBS-16 have been placed on deposit in the American Type Culture Collection, 12301 Parklawn Drive, Rockville, Maryland.

EXAMPLE 4 This example demonstrates removal of a 5s~ untranslated segment of HBV-DNA. The DNA segment comprising the S-protein region isolated as described in Example 2 included an untranslated 26 base pair segment at the 58-end of the coding region, lying between the promoter and the ATG start codon. The following procedure was developed to remove all, or all but one, of the bases of the 51-untranslated region of the HBV-DNA preceding the S-protein coding region.

The plasmid pHBS-5 was digested with EcoRI endonuclease generating a fragment of approximately 850 base pairs including the S-protein coding region and flanking 39-and 5’-untranslated regions, terminated by EcoRI linker oligonucleotide segments. The HBV-DNA segment was reisolated by preparative gel electrophoresis, electroeluted and divided into samples which were digested with the exonuclease Bal-31 for varying times from 0.5 to 30 minutes at 37°C. The extent of exonuclease digestion was characterized qualitatively by digesting a portion of each sample with Xbal endonuclease. The S-protein coding region contains an Xbal site beginning 92 base pairs from the first base of the start condon. Therefore, samples in which Ba1-31 digestion had proceeded beyond the Xbal site would yield only one fragment upon qel electrophoresis after Xbal endonuclease incubation while samples with fewer bases removed would vield two classes of fragment: a homogeneous large fragment and a heterogeneously sized small fragment. Samples yielding only one Xbal fragment were discarded. Samples yielding two size classes of fragments were blunt-ended by incubation with DNA polymerase I (Klenow fragment, see Klenow, Η., et al., Proc. Nat. Acad. Sci. 65., 168 (1970) in the presence of all four deoxynucleotide triphosphates. Linker oligonucleotides containing the EcoRI recognition site were added by blunt-end ligation using T4 DNA ligase. EcoRI specific cohesive ends were generated by digestion with EcoRI endonuclease. The modified DNA was isolated by gel electrophoresis, electroeluted and joined to EcoRI-digested, alkaline phosphatase treated pBR322, in a DNA ligase catalyzed reaction. The recombinant plasmids were then used to transform E. coli HB-101. 6 Two strategies were employed for screening and characterizing clones with shortened 5’untranslated segments. In the first, individual colonies were screened for the presence of S-protein coding region by in situ colony hybridization using labeled S-protein coding DNA as a probe. Colonies screening positively for the presence of the S-protein coding region were used to start cultures from which vector DNA was prepared. The vector DNA was incubated by EcoRi endonuclease to excise the S-protein coding region. The S-protein coding DNA thus prepared was analyzed either by determining the size of fragments generated by Xbal endonuclease digestion or by DNA sequence analysis of the 5’-terminal sequences (Maxum, A., et al., Proc. Nat. Acad. Sci. USA 74, 560 (1977)).

A second screening strategy was used to detect clones in which the 5-untranslated region had been completely removed or terminated one base short of the start codon. The method exploited the observation that the first four bases of the S-protexn coding sequence, ATGG, when joined to an EcoRi linker oligonucleotide (GGAATTCC) generated a recognition site for the restriction endoculease NeoI; CCATGG. The NeoI site thus generated would be unique in the vector, since neither pBR322 nor the S-protein coding region contains an Ncol site. Therefore, any S-protein coding segment in which the Bal-31 digestion was terminated precisely at the ATG start codon would be characterized by the generation of a new Ncol site when joined to an EcoRi linker oligonucleotide. As it happens, in HBv-DMA there is 7 a C residue adjacent to the ATG start codon of the S-protein region, in the 5s-untranslated region. Therefore, Bal-31 digests that retain only the last C of the 5’-untranslated region will also generate an Ncol recognition site when jointed to an EcoRI linker oligonucleotide. These two specific constructions were screened for by incubating the clones with a combination of Ncol and XbaI endonucleases followed by gel electrophoresis of the fragments generated, if any. Those clones yielding a 96 base pair fragment were therefore selected, as these had all or all but one of the 58-untranslated base removed, but retained the ATG start codon. The exact sequence was then confirmed by DNA sequence analysis using the Maxam et al. procedure. The resulting plasmid, combining pBR322 with a modified HBV segment with the entire 5s> untranslated region of the S-protein gene deleted, inserted at the EcoRI site, was designated pHBS5~3. (See Figure 4)„ The HBv-DNA segment in pHBS5-3 was also incidentally modified by the removal of about 40 base pairs of the 3s untranslated region, due to concomitant action of Bal-31 exonuclease at the 38 end.

Expression vector construction analogous to pHBSlS, described in Example 2, was carried out by insertion of the modified HBV-DNA segment of pHBS5-3 in place of the corresponding segment in pHBS16. For this purpose, a 16-type" vector was prepared by EcoRI endonuclease digestion and religation, followed by selection for a vector in which the HBV-DNA was deleted. Expression vectors constructed by the insertion, of modified HBV-DNA segments at the EcoRI 8 site of the 16-type vector were characterized by the designation pHBS16-X, where X is a number characterizing the modification of HBV-DNA inserted at the EcoRI site. Thus, the HBV-DNA segment transferred from pHBS5-3 to the lo vector generated an expression plasmid designated phHBS16-3. (See Figure 5). All constructions were screened for correct orientation of the S-protein codinq region with respect to the ADH1 promoter by combined digestion with BamHI and Xbal endonuclease digestion. Correct orientation yielded a Bam- Xba fragment approximately 1600 base pairs in length whereas incorrect orientation yielded a longer fragment.

The host strain for the 15 series expression vectors was Saccharomyces cerevisiae AB 35-D3-D a., leu2-3, leu2-112, ura3~52, trpl-289, his4-580, ade2 or Saccharomyces cerevisiae AB 35-14-D. Sample cultures of the host strain transformed with either pHBS16 or pHBSlS-3 were grown under equivalent conditions and Y~HBsAg was qualitatively assayed by radioimmune assay as described in Example 3. Cells transformed with pH3S15-3 produced approximately 2.2 times as much Υ-HBsAg per cell as those transformed by pHBSlo.

EXAMPLE 5 A further modification was made of the vector construction described in Example 4, in which the 3'-untranslated region removed during the Bal-31 digestion was restored. The strategy of this construction was to combine, at the Xbal site 9 internal to the S-protein coding region, a fragment of the coding region derived from pHBS5-3, modified as described in Example 4, together with an unmodified fragment from pHBS5 having an intact 3’untranslated region.

The HBV-DNA segment of pHBSS containing the 58-untranslated region and promoter proximal part of the S-protein coding region was removed by the sequential action of Clal endonuclease and Xbal endonuclease. The plasmid was first cleaved with Clal endonuclease. The resulting unpaired ends were filled in using DNA polymerase I Klenow fragment in the presence of the four deoxynucleotide triphosphates to provide a linear vector with blunt ends. The DNA was then digested with Xbal endonuclease and alkaline phosphatase. The latter treatment was intended to insure that the ends generated by the foregoing series of steps could not rejoin to one another in the presence of DNA ligase, (Shine, J., supra).

The modified HBV-DNA of pH3S5-3 containing the promoter proximal portion of the coding region for S-protein was also prepared by sequetial endonuclease digestion. Plasmid pHBS5-3 was first cleaved with EcoRI endonuclease and blunt ended with DNA polymerase I Klenow fragment in the presence of the four deoxynucleotide triphosphates. The DNA was then cleaved with Xba endonuclease. The small fragment resulting from Xbal cleavage, approximately 100 base pairs having a blunt EcoRI end and an Xbal end was isolated by gel electrophoresis and electroelution. The purpose of sequential Ο endonuclease treatment in both instances was to insure that fche 100 base pair fragment would be joined in correct orientation with the cleaved vector DNA. The 100 base pair fragment derived from pHBS5-3 was mixed with modified vector DNA derived from pHBS5, in the presence of DNA ligase under conditions permitting blunt end ligation as well as the joining of paired ends derived from the Xbal cuts. Transformants were selected and identifed by the existance of an Ncol site, derived from the small fragment from pHBS5~3 (See Example 4).

It was anticipated that the EcoRI site adjacent to the Ncol site would be regenerated by the construction strategy employed. However, one base pair in the filled in EcoRI site was not regenerated by the DNA polymerase treatment. Consequently, the EcoRI site was not regenerated as expected.

Fortuitously however, the rejoined sequences did regenerate the Clal site.

DNA nucleotide sequence analysis of the resulting vector, designated pHBS6, diagrammed in Figure 4, confirmed that the HBV-DNA region of the vector contained the entire S-protein coding region and 39 untranslated region together with a deletion of the 5’ untranslated region as decribed for pHBS5-3. The HBV-DNA segment modified as in pHBS6 was transfered to an expression vector of the 16 series as follows: The HBV-DNA region of pHBSS was isolated by combined action of Ncol and EcoRI endonucleases and blunt ended by incubation with DNA polymerase I Klenow fragment in the presence of the four deoxynueleotide triphosphates. The resulting 820 base pair fragment was isolated by gel electrophoresis and electroelution. The expression vector pHBSlo, or the ,916-vector described in Example 4, was cleaved by EcoRI endonuclease action, blunt ended using DNA polymerase 1 Klenow fragment in the presence of the four deoxynucleotide triphosphates, and treated with alkaline phosphatase.

The HBV-DNA fragment was then joined to the treated IS vector by blunt end ligation using T4 DNA. ligase. Correct orientation of the fragments regenerated an EcoRI site between the ADH promoter and the start codon of the S-protein coding region. DNA nucleotide sequence analysis was carried out confirming the structure of the resulting construction, designated pHBSlo-5, diagrammed in Figure 5.

The relative rate of expression of Y-HBsAg for yeast cells transformed by pHBS16-5 was measured under the same conditions as for pHBS!6-3, described in Example 4. Expression of Υ-HBsAg by Saccharomyces cerevisiae AB-35-D3-D transformed by pHBS16-5 was approximately 2.8 times greater per cell as measured by radioimmumo assay, than expression by cells transformed with pHBSlS.

A related construction using the HBV-DNA segment of pHBSS was carried out using an identical procedure with the exception that the HBV-DNA fragment was excised by an EcoRI endonuclease preparation having some EcoRI* activity. After ligation with the 16 vector, prepared as previously described, the expression vector was found to have lost the EcoRI site as well as the Ncol site adjacent to the ATG start codon of the S-protein coding 2 region. The resulting expression plasmid was designated pHBS16-4. The nucleotide sequence adjacent to the S-protein start codon was 59»..ACTATCTGGCATGG»..3. The rate of expression in yeast cells transformed with pHBS16-4 was comparable to that of pHBS16-5 transformed cells, within experimental error. The structure of pHBS16-4 is diagrammed in Figure 5.

The nucleotide sequence adjacent to the S-protein start codon of PHBS-16-3 was 5' ...ACTATCTGGAATTCCCATGG...3’ . The sequence for pHBS-16-5 was 5’...ACTATCTGGAATTCATGG.. . 3 ’ . The sequence difference between 16-3 and 16-5 was a consequence of blunt-ending the DNA after EcoRI digestion of pHBS-6., EXAMPLE 6 This example describes details of the construction of a series of vectors for expression characterised by having the entire two micron circle plasmid DNA sequence within their sequence together with DNA segments comprising the promoter and transcription terminator sequences of the yeast ADH gene, with the S-protein coding region sandwiched between the ADH promoter and the ADH terminator regions. These vectors were designated ’’So63 series vectors and their nomenclature is consistent with the nomenclature of the 16 series of expression vectors. Thus, the expression vector pHBS16-3 contains the S-protein gene modified as described for pHBS16-3, while pHBS56-5 contains HBV-DNA modified as described for pHBSlS-5, as described in Examples 4 and 5 respectively. The full length two micron circle DNA ’S (d 4-» provides for stable replication in a circle zero host strain in the absence of metabolic selection pressure. The ADH terminator was provided to enhance the stability of S-protein mRNA transcripts.

The parent plasmid for the construction of 56 type vectors was pCI/Ι which was a hybrid plasmid between pBR-322 and a two micron circle plasmid joined at their EcoRI sites. The two micron circle portion was previously modified to contain an inserted LEU2 gene of yeast and obtained from the plasmid PJBD219 described by Beggs, J. et al., Nature 275, 104 (1978) . The restriction map of pCI/Ι is shown in Figure 6.

It can be seen from Figure 6 that digestion of pCl/1 with endonuclease Sphl deleted a portion of the plasmid spanning the two micron-pBR-322 joint.

It was observed that the active portion of the ADHl promoter region was contained within an Sphl HindHI fragment of approximately 300 base pairs lenqth (See sequence of the ADHl gne, by Bennetzen, J. L. and Hall, B. D., J. Biol. Chem. 257, 301 (1982)). The recognition sequence for Sphl is GCATGC and such a sequence exists in the ADH promoter beginning at position -413. Similarly, the yeast terminator sequence was contained within a HindiII Sphl fragment of about 330 base pairs. In both cases the Sphl site was distal to the coding region so that the HBV S-protein coding region could be inserted between them if provided with HindHI sites at its termini. 4 The precursor source for the ADH promoter and terminator segments was plasmid ΡΆΑΗ5 containing a 1500 base pair ADHl promoter fragment terminated at position "9 in the nucleotide sequence (Hitzeman, R„ A. et al., supra, and an approximately 450 base pair terminator unit from nucleotides 913 to 1368 in the ADH bene nucleotide sequence, joined by a Hindlll site between the fragments and cloned into the BamHI site of the vector ΥΞρ13, (Broach, J. and Hicks, J., Gene 8, 121 (1979)). The HBv-DMA segment of pHBS5 was excised by EcoRi digestion. The protruding ends were filled in using DNA polymerase I Klenow fragment and joined at both ends with Hindlll linker oligonucleotides having the sequence CAAGCTTG. After Hindlll endonuclease digestion to expose unpaired, Hindlll specific ends on the HBV-DNA segment, the segment was joined to Hindlll cut plasmid pAAH5, thereby placing the HBV S-protein coding sequence between the ADH promoter and terminator fragments. A plasmid with the S-protein gene in correct orientation with respect to the promoter and terminator fragments, as determined by restriction analysis, was designated pHBS-22. The ADH promoter and terminator sequences were each found to contain an SphI site (recognition sequence GCATGC) making it possible to excise the entire composite gene comprising about 400 base pairs of ADHl promoter, HBV S-protein region and about 330 base pairs of ADHl terminator by digestion with SphI endonuclease. Digestion of pHBS22 with SphI endonuclease yielded the intact composite gene in a fragment of approximately 1500 base pairs. The fragment was •3 joined with SphI-cut vector pCI/Ι. E. coli. HB101 transformants were screened for ampicillin resistance and sensitivity to tetracycline, since the segment excised by SphI endonuclease digestion of pCl/1 deleted a portion of the tetracycline resistance gene of the pBR322 segment. The structure of the resulting vector, designated pHBS56 was further confirmed by restriction analysis. E. coli. HB101 transformants obtained from the products of the ligase reaction were cloned on plates containing ampicillin. Plasmid DNA from single colony isolates grown in culture was screened by restriction endonuclease analysis for the insertion and correct orientation of the S-protein coding region, ^he plasmid selected, pHBS-56 containing the HBV S-protein coding region in correct orientation with respect to the ADH promoter and terminator segments, is shown in Figure 6.

Two additional 56-type vectors were constructed using a promoter fragment and promoter proximal region of the S-protein coding segment obtained from pHBSlo-3 and from pHBS16-5. These constructions were designated pHBS56-3 and pHBS56-5, respectively. In both cases, an SphI - Xbal fragment was joined in a DNA ligase catalyzed reaction with the larger of the two SphI - Xbal fragments obtained by digestion of pHBS55. The larger fragment, approximately 1080 base pairs, extends from the Xbal site within the S-protein coding region to the SphI site of the ADH terminator region. ^his fragment was isolated by gel electrophoresis and electroelution prior to joining to the SphI - Xbal fragment of 6 either pKBSlS-3 or pHBS16-5, in a DNA ligase catalyzed reaction. The two composite genes thus constructed were identical except for the sequence of the 5s-untranslated region between the ADH oromoter and the S-protein start codon, these differences arising from differences in the respective source vectors, pHBSlS-3 and pHBS16-5 respectively. Both composite genes were sub-cloned, in separate reactions, at the Sphl site of PBR322, situated between the BamHI site and the Sail site of PBR322, for the purpose of obtaining amplified amounts of composite gene DNA. The sub-cloning vectors were selected by their ability to confer ampicillin resistance and tetracycline sensitivity phenotype to E. coli. HB101 transformants.

In the final step, the large fragment produced by Sphl cleavage of pHBSSo was treated with alkaline phosphatase, and isolated by gel electrophoresis and electroelution» Similarly, the composite genes were excised from their sub-cloning vectors by Sphl cleavage and isolated by gel electrophoresis and electroelution, but without phosphatase treatment. These were combined, in separate reactions, with the large Sphl fragment of pHBS56 and joined in DNA ligase catalyzed reactions yielding PHBS56-3 and pHBS5©~5, respectively. E. coli. HB101 transformants were selected by ampicillin resistance and tetracycline sensitivity and further characterized by restriction analysis. A diagram of the construction steps and maps of the relevant vectors are shown in Figure 7. Although the nomenclature of the 56 type vectors is parallel with 7 that of the 16 type vectors, it will be understood that the distinctions in the case of the 56 series vectors refer only to the 5s untranslated region in each instance, and that the 3’ untranslated region is the same in each member of the series. Specifically, pHBS56-3 lacks the 40 base pair deletion in the 38 untranslated region that occurs in pHBS-16-3.

After cloning selection and characterization, vectors of the 56 series were used to transform a circle zero yeast strain designated 2150-2-3. The strain was derived from a genetic cross between strain Y379-5-D cyh2 nibl (rho ) Livingston, D. Genetics 86, 73 (1977) and DC 04 a Adel AdeX leu2~04 (cir°) (Broach, J., Cell, 21, 501 (1980). The diploid strain resulting from the cross was permitted to sporulate and tetrads were disected by a standard procedure. One of the haploid spores gave rise to strain 2150-2-3 a Adel Leu2-04 (cir^) . Yeast transformants were selected for Leu' phenotype conferred by the presence of plasmids of the pHBS-56 series.

The relative rates of Y-HBsAg expression in various plasmid-host combinations were compared in the following manner: A one liter volume of cell culture was grown to its limiting cell density, the cells harvested in crude lysates analyzed for total soluble protein and for Υ-HBsAg, using radioimmunoassay, as described, supra. The results were expressed both as micrograms Υ-HBsAg per liter of culture and Υ-HBsAg as percent by weight of total yeast soluble protein. The latter provides a measure of the amount of cell metabolism devoted to Y-HBsAg production while the former provides a measure of the 8 overall yield of Υ-HBsAg obtainable upon growth of cultures to limiting density. Differences arise particularly in the case of the 56 series vectors, because the cells carrying these vectors can be grown in rich media without selection pressure whereas the 16 and 25 series vectors require growth in a defined medium lacking tryptophan to prevent the accumulation of untransformed segregants. The results are shown in the accompanying table. The weights of Y-HBsAg given in the table were determined by a commercial radioimmunoassay using antibody raised against HBsAg. Therefore, the amounts of Υ-HBsAg reported may not be an absolute measure of mass but may be considered to be internally consistent for purposes of comparison.

Vector pHBS- Host Υ-HBsAg Yields Culture Density O.D. % by weight of total yeast soluble protein pg Y-HBsAg per liter culture 16-3 AB-35-14D 2.0 0.1% 20 16-4 A3-35-14D 2.0 0.1% 20 16-5 AB-35-14D 2.0 0.1% 20 25 GM-3C-2 4.0 0.5% 200 56 2150-2-3 12.0 0.3% 300 & EXAMPLE 7 Preparation of a vaccine comprising HBsAg synthesized by yeast. Υ-HBsAg particles are purified from cell extracts by the method of Example 3, or by suitable methods known in the art. for example, as described in U.S patents 4,088,748 or 4,181,713. Purified HBsAg particles are dialyzed against physiological saline or phosphate-buffer saline and adjusted to 100 ug protein/ml final contentration. Guinea pigs are subjected subcutaneously at 9, 14 and 56 day intervals with 1 ml of the HBsAg preparation. The serum of the test animals is sampled at 0, 28, 56 and 84 days and assayed for antibody titre agaist Dane particles or HBsAg purified from Alexander cells. The radioimmune assay described in Example 3 is employed, or in the alternative, the radioimmunoassay of Hollinger, F. et al., J. Immunol. 107, 1099 (1971) is employed. The majority of animals exhibit antibodies cross-reactive with HBsAg 84 days after administration of the particles.

Similar results are obtained upon injection of monkeys. Accordingly, HBsAg synthesized by yeast is immunogenic and is capable of eliciting antibodies cross-reactive with naturally-occurring HBsAg.

HBsAg synthesized by yeast has the advantage of being available in significantly larger quantities than that obtained from Dane particles or carrier serum. A more uniform product is obtainable at an advantageous cost per unit, which may be expected to decrease with increasing production volume.

Furthermore, there is no danger of accidental infection, since there is no intact HBV, and can be Ο no intact HBV, in the surface antigen prepared from yeast» By contrast, viral proteins purified from serum or other natural sources always pose the danger of viral contamination.

EXAMPLE 8 As shown in Exmple 7, HBsAg synthesized by yeast is capable of eliciting antibodies cross-reactive with naturally-occurring HBsAg. It therefore follows that such antigens and antigen aggregates, when purified as described and administered in a physiologically acceptable medium, constitute a vaccine for protection agaist infection by hepatitis B virus.

Sixteen chimpanzees are divided into three groups. Group A (six animals) is inoculated intravenously with 1 ml of a standard Bureau of Biologies Hepatitis B virus preparation; Group B (four animals) is inoculated intravenously with 1 ml containing 200 ug of HBsAg synthesized in yeast and purified as described in Example 3, in physiological saline; Group C (six animals) is the control group and receives no inoculation. All chimps in Group A have evidence of clinical Hepatitis B (either antigenemia, enzyme elevations and/or antibody response) within 40 weeks. None of the animals in Groups B or C shows evidence of clinical Hepatitis B infection over the same 40-week period. The chimps of Group 3 are rendered immune to subsequent challenge when inoculated intravenously with 1.0 ml of BOB Hepatitis 3 virus. 1 EXAMPLE 9 Υ-HBsAg differed in several respects from plasma-derived HBsAg. The diameters of HBsAg and Υ-HBsAg particles were measured from negatively stained electron micrographs. The yeast-derived antigen had a diameter range of from about 14 to about 18 nm while plasma-derived antigen had a diameter range of from about 20 to about 24 nm. Υ-HBsAg was unstable at pH 2, and to pepsin at pH 2, whereas plasma-"derived HBsAg was stable under the same conditions. To a 1 ml suspension of purified Y-HBsAg was added 0.03 ml of IN HCl to lower the pH to 2.0. The sample was divided in halves and to one-half was added 1 ug of pepsin while no enzyme was added to the other half. Both samples were held at 37°C for sixteen hours and then 0.03 ml of IM NaOH was added to each to raise the pH to 7.0. The two samples were measured for antigen binding activity in a quantitative radioimmunoassay (RIA). Over 95% of the RIA activity was lost in each sample. Under the same conditions the plasma derived HBsAg retained all of its antigen binding activity.

A sample of purified Υ-HBsAg was heated in sodium dodecylsulfate (SDS) and 2-mercaptoethanol at 90°C for five minutes. It was then electrophoresed through a 10% polyacrylamide gel containing 0.1% SDS. Subsequent staining of the gel with protein stains revealed a single band at a molecular weight equivalence location of about 25,000» A plasma purified HBsAg sample treated identically showed two bands after staining: one band at about 25,000 daltons and a second, band at about 28,000 daltons. 2 Unlike plasma-derived HBsAg, Υ-HBsAg did not bind to a monoclonal antibody (HBsAb) selected against HBsAg. A crude extract of yeast cells containing surface antigen was passed through an affinity adsorbent column prepared by chemically coupling monoclonal HBsAb to an agarose gel. Measurement of the column effluent revealed 90% of the Υ-HBsAg charged to the column was present in the effluent. A plasma-derived HBsAg passed through the same column of monoclonal HBsAb revealed less than 10% of the charged antigen in the column effluent. Υ-HBsAg exhibited higher activity in mouse potency tests. Υ-HBsAg was adsorbed to an aluminum hydroxide gel prior to administration and was diluted to contain 10, 2.5, 0.62, 0.15 and 0.0375 pg/ml. 1 Ml quantities of the foregoing concentrations were injected intraperitoneally into each of five groups of five-week old female mice. Each of these concentrations was injected into one of the five groups, each group containing ten mice. The yeast-produced antigen had an ΕΏ^θ (the concentration of antigen needed to produce antibody in one-half of the mice) of about 0.05 ug/ml while the plasma-derived antigen in the same procedure had an ED^q of about 0.5 ug/ml. Υ-HBsAg purified as described was essentially free of contaminating chemicals. Measurement of the Lowry protein of Y-HBsAg showed 53 ug/ml while measurement of the RIA antigen binding ability of this antigen indicated a concentration of 12 pg/ml. Measurement of the Lowry protein and RIA antigen binding ability of a pure preparation of plasma-derived HBsAg revealed a protein concentration of 44 ug/ml and a binding activity of 50 ug/ml.

No differences were observed in the physical, chemical or antigenic properties of Y-HBsAg produced from cells tranformed with pHBs-16, -25 or -56.

EXAMPLE 10 Comparative study of Potency in Mice of Yeast-Derived lu Antigen with Plasma-derived Antigen_ A total of 80 five-week old female mice were divided into two groups of 40 and each group was further subdivided into four sub-groups of 10 mice. The 10 mice from each sub-group were injected intraperitoneally with either the antigen prepared from Example 3 using vector pHBS-25 or plasma-derived antigen at a concentration, respectively, of 10, 2.5, 0.625 or 0.156 ug/ml. Saline-alum placebo was used as diluent to dilute the antigen concentration where necessary to obtain the foregoing concentrations.

The mice were individually bled and sacrificed at 28 days,. Antibody determinations were performed by the Ausab (Abbott) radioimmune assay. The serological results are summarized in the following table wherein titers are expressed as Estimated Ausab Units®. 4 Anti-HBs Titer Groun Material Cone. (u/ml) Sero- conversion Estimated Ausab Units 5 I YeastDer ived Antigen 10 (RIA) 10/10 5400, 7200, 800, 13,500, 7200, 183, 800, 136,000, 15,800, 23,500 ιυ II 2.5 10/10 18,300, 5400, 800, 7200, 5400, 800, 5400, 1600, 135,000, 72 15 III 0.625 8/10 800, 8, 1600, 16,000, 158,000, 1600, 3600, 8, 8, 8000 IV 0.125 8/10 8, 800, 800, 32,200, 5400, 8, 112,000, 18,300, 4 / /, 412 20 V Plasma- Derived Antigen 10 (Lowry) 9/10 292,000, 20,800, 8, 3 8,2 0 0, 16, 16 , 13,500, 36, 54, 512 25 VI 2.5 10/10 36, 38,200, 15,800, 800, 800, 800, 7200 18,300, 208, 15,800 VII 0.625 4/9 8, 8, 8, 800, 512, 512, 8, 158, 8 30 VIII 0.125 0/10 8, 8, 8, 8, 8 8, 8, 8, 8, 8 EXAMPLE 11 Comparative Study of Potency in African Green Monkeys of Yeast-Derived Antigen with Placebo Derived Antigen A total of 32 African Green Monkeys were divided into two groups of 16, and each group was further subdivided into four sub-groups of 4 monkeys. Each sub-group was injected intramuscularly at day 0 and day 28 with yeast-derived antigen or plasmaderived antigen at a concentration, respectively, of , 2.5, 0.625 or 0.156 pg/ml using saline-alum placebo as diluent to dilute the antigen concentration where necessary to obtain the foregoing concentrations. Bleedings were collected at weekly intervals for 14 weeks and antibody determinations were performed by the Ausab (Abbott) radioimmune assay with titers expressed as ^Estimated Ausab Units. The results are summarized in the following tables. ιυ (Yeai.t-AJ.uBi) Aiiissl f Cone. (kca/hI) (RIA) Tlsss/Week/Zg Sleeted Auoab Units -1 c* 2 4* 6 0 10 12 • 14 V-SQ015 10 <8 ¢0 54 36 322 5400 3500 1600 2350 73144 10 «8 <8 320 112 135 292 5120 20,800 20,200 73147 10 «8 <3 72 72 135 4120 13,500 29,200 26,200 S0Q12 10 «8 512 203 135 206 3320 2620 47,200 V-80006 2.5 ' «8 «8 512 352 3220 3&9000 54,000 72,000 16,000 79148 2.5 «§ <8 512 112 92 112 92 135 512 soocs 2.5 <0 <3 302 112 158 183 5120 23,200 32,200 80016 2.5 «8 «3 720 262 5120 800 20,800 18,300 23,500 Y-80105 0.625 «0 sS 183 412 3120 6000 16,000 20,800 20,800 00132 0.625 ΐ 8 <3 7200 158 92 158 5120 18,300 11,200 80133 0.625 ¢0 3400 352 5120 16,000 13,500 15,300 26,200 80201 0.625 < 8 ¢,0 103 112 112 92 112 235 47,2 V-S01Q1 0.156 <8 « 8 512 352 2920 3400 1600 1350 1350 80173 0.156 «0 18 677 262 158 92 32 412 5120 80100 0.156 < g < g <8 <9 5600 1600 360 350 720 80175 0.156 «3 < 8 c 3 382 3220 320 350 360 360 * Inoculate - 1.0 ml l.a. at day 0, 4 weeks S3 AnlM.1 / Cons. (eft/sd) (Lowry) Tiaa/Wgsk/Es stiaaied Ausab Unitsra,l 0* 2 4* 6 3 20 12 14 V-80176 10 c 3 <0 <£ © ig ig <8 3 ig 8 80102{A) 10 ig 512 512 5120 36,000 18,000' 7200 3600 son? 10 <3 «Ε 0 208 5120 3600 3220 7200 11,200· 30177 10 ig <3 <8 <0 16 ig ig <8 V-79142 2.3 <0 512 292 512 38,200 8000 20,800 23,500 30035 2.5 «3 512 352 92 135 5120· 3400' 5400 78089 2.5 <8 i8 ig 262 5120 5400 5400 7200 7200 80120 2.3 «8 512 312 5120 54,000 16,000 11,200 7200 V-78151 0.625 ig ¢3 ¢0 512 7200 5600 5400 3600 79149 0.623 «3 <0 ig ft 77 1830 1120 1830 3220 7200 80182(S) 0.825 ig «3 <8 8 5120' 1600 720 330 540 79143 0.825 cB «8 235 512 5120 36,000 16,000 16,000 32,200 ¥-79140 0.158 10 •iQ ig 16 183 72 133 352 S120 79146 0.156 ig <0 <0 1600 360 36 36 72 77024 0.156 <0 <£@ 512 208 1600 2820 SOO 520 340 00155 0.156 ig <0 ig «0 ig i0 0 Λ α * Incculits - 1.0 ml i.». a£ d*y 0, 4 weeks D " Died 8 The present invention represents a substantial advance in applying recombinant DNA technology. The practical goal of svnthesizing HBsAg in a microorganism host has been achieved.

Modifications to increase HBsAg production and improve the yield of HBsAg upon purification, which fall within the scope of ordinary skill in the art, are deemed equivalent variants within the scope of the claimed invention. Examples of such IU modifications could include improved promoter systems, more productive host cell strains, improvements in purification technique and modifications to improve the antigenicity of the product or its immunogenicity.

The following plasmids and transformed yeast strains were deposited in the American Type Culture Collection, 12301 Parklawn Drive, Rockville, Md. 20852, U.S.A.

Descr ipt ion Deposit Date Accession No. 1. Plasmid pHBS-16 August 4, 1981 40043 2. S. Cerevisiae XV610~8C/pHBS~16 Aucj ·» 4 p 1981 20619 3. Plasmid pHBS56 July 7, 1982 40047 4. Plasmid pKBSl6-3 July 7, 1982 40048 5. S. Cerevisiae AB35-14D/pHBSl6~3 July 7, 1982 20647 6. S. Cerevisiae AB35-14D/pHBS16-4 July 7, 1982 20646 7. S. Cerevisiae 2150-2-3/pHBS56 July 7, 1982 20648 8, Plasmid pHBS16-4 July 7, 1982 40046 9. Plasmid pHBS16-5 July 7, 1982 40045 10. S. Cerevisiae AB35-14D/pHBS16-5 July 7, 1982 20645 11. Plasmid pHBS56-3 July 14 , 1982 40051 12. Plasmid pHBS56-5 July 14 , 1982 40052 13. S, Cerevisiae 2150-2-3/pHBS56-3 July 14 , 1982 20649 14. S. Cerevisiae 2150~2-3/pHBS56~5 July 14 , 1982 20650 The depository was requested to handle the above-described deposits in accordance with the terms and conditions of the Budapest Treaty on the International Recognition of the Deposit of Microorganisms for the Purposes/Patent Procedure»

Claims (14)

CLAIMS :

1. A DNA expression vector capable of replication and phenotypic selection in yeast host strain comprising a promoter compatible with a yeast host strain and a DNA sequence encoding hepatitis B surface antigen, said sequence being positioned together with translational start and stop signals in said vector under control of said promoter such that in a transformant yeast strain it is expressed to produce hepatitis B surface antigen in particle form having a sedimentation rate which is virtually identical to that of authentic 22 nm hepatitis surface antigen particles.

2. A yeast strain transformed with the DNA expression vector according to claim 1.

3. The yeast strain according to claim 2 obtained by transforming a leu2 auxotrophic yeast strain.

4. The yeast strain according to claim 2 obtained by transforming strain XV610-8C.

5. A fermentation culture comprising a transformed yeast according to claim 2, 3, or 4.

6. The vector of claim 1 wherein the DNA sequence encoding hepatitis 3 surface antigen encodes only the mature hepatitis B surface antigen structural gene. 5 2

7. A method of producing hepatitis B surface antigen in particle form suitable for use in conferring immunogenicity to hepatitis B virus in a susceptible human, comprising: a) providing a transformed yeast strain according to claims 2 through 4; b) allowing the transformant strain to grow under fermentation conditions until said hepatitis B surface antigen is produced therein; and c) recovering said hepatitis B surface antigen in particle form having a sedimentation rate that is identical to that of authentic 22 nm surface antigen particles .

8. The method according to claim 7 wherein the segment encoding hepatitis B surface antigen comprises, in order from the 5'-end of its coding strand a translational start codon, nucleotides encoding hepatitis B surface antigen of the hepatitis B genome, and one or more translational stop signals.

9. The method according to claim 7 wherein the yeast strain of step (a) is XV610-8C.

10. The method according to any one of claims 7, 8, or 9 wherein the promoter of step (a) is derived from the yeast PGK promoter region.

11. The method of claim 8 wherein the start codon is the start codon for mature hepatitis B surface antigen structural gene. 5 3

12. A DNA expression vector according to claim 1, substantially as hereinbefore described and exemplified

13. A fermentation culture according to claim 5, 5 substantially as hereinbefore described.

14. A method according to claim 7, substantially as hereinbefore described and exemplified.