EP0538371A1 - Streptomyces vectors for production of heterologous proteins - Google Patents

Abstract

Nucleic acid sequences and DNA vectors useful in the production of CD4 chimeric proteins, as well as other heterologous proteins, in Streptomyces. The nucleic acid sequences of the invention comprise the coding sequence for the signal peptide of the tyrosine inhibitor gene Streptomyces Longisporus (ITL) functionally linked to a propeptide sequence composed essentially of an amino acid sequence coding for 1 to about 6 amino acids, the sequence of said amino acids chosen to allow the formation in Streptomyces of a protein product having a homogeneous amino terminus after treatment in order to eliminate said signal peptide formed in the protein product during the synthesis of the latter. In another embodiment of the invention, the propeptide is omitted and the ITL signal peptide is functionally linked to a nucleic acid sequence coding for a heterologous protein having been modified in order to code for the sequence lys-ala- to l 'end 3'. The invention also relates to cells transformed with the nucleic acid sequences or vectors of the invention, as well as methods of using the nucleic acid sequences and vectors of the invention to produce heterologous proteins in Streptomyces.

Description

STREPTOMYCES VECTORS FOR PRODUCTION

OF HETEROLOGOUS PROTEINS

Field of the Invention

The present invention relates to production of heterologous proteins in microorganisms. More particularly, the present invention relates to nucleic acid sequences and expression vectors for expression of heterologous proteins, particularly soluble CD-4 proteins, in Streptomyces.

Background of the Invention

The principal route of infection of HIV viruses is mediated through attachment of the virus to the CD4 protein. CD4 (also referred to as T4) is a T cell surface glycoprotein that associates with the major histocompatibility complex class II molecules. However, CD4 is also the target for the HIV-1 envelope glycoprotein, gp120. Once the HIV-1 virus binds to CD4 on the target cell surface, the viral PNA is then introduced into the target cell by direct fusion of the apposing viral and plasma membranes (Maddon et al., Cell, 54:865-874 (1988).

The DNA sequence of the CD4 has been disclosed by Maddon et al. (Cell 42:93-104 (1985)). From this it was deduced that the mature CD4 protein consists of an extracellular region containing four immunoglobulin-like domains (V1-V4), a transmembrane domain, and a charged intracellular region of approximately 40 amino acids (Maddon et al., supra).

Recombinant soluble derivatives of CD4, such as sT4, comprise proteins in which the transmembrane and cytoplasmic domains which have been deleted, yet retain the ability to bind gp120. sT4 has also been shown to inhibit both viral infectivity and HIV mediated syncytia formation (Deen et al., Nature 331: 82-84 (1988)). This inhibition results from the association of the sT4 protein with the HIV gp120 envelope protein, thereby competing with the native CD4 protein on the cell surface. The mechanism by which CD4 and soluble CD4 inhibits viral infectivity and syncytia formation is not completely understood. In vitro, it has been shown that the addition of soluble CD4 to HIV-1 or HIV-1 infected cells appears to induce the release of gpl20 without a concomitant increase in other viral proteins. This implicates CD4 as an active rather than passive inhibitor of HIV-1.

However, in vivo, the sT4 protein and similar soluble CD4 proteins are rapidly eliminated from serum. This rapid elimination severely limits the clinical application of soluble CD4 as an inhibitor of HIV function.

To prolong the serum clearance time, several soluble CD4 chimeras have been constructed. Traunecker et al. (Nature 331:84-86 (1988)) disclosed CD4 chimeras (V1V2 and V1V2V3V4) in which the carboxy terminal portion of the protein consists of a murine immunoglobulin light chain constant region. Traunecker et al. further disclosed that these constructs were expressed in myeloma cells.

Traunecker et al. (Nature 339:68-70) (1989) subsequently disclose CD4 chimeras (V1V2) fused to murine IgM or IgG2 heavy chain constant regions. These constructs were also disclosed to form pentamers.

Seed (EP-A-325,262, published July 26, 1989) and Capon et al. (WO89/02922, published April 6, 1989) disclose CD4 chimeras fused to human IgG heavy and light chain constant regions. The proteins disclosed by Seed and Capon et al. were all expressed in mammalian systems, i.e. COS, BHK (baby hamster kidney) and CHO cells. As a result, the disclosed CD4 chimeras are expensive to produce and the overall production capacity is limited, relative to alternate expression systems.

It is thus an object of the present invention to produce CD4 chimeras in a bacterial host system to reduce the cost of production as well as to produce CD4 chimeras which have an increased serum half-life and/or potency against HIV infections relative to soluble CD4.

It is common for proteins produced by recombinant DNA techniques to be fused with a portion of a protein produced by the host cell or other protein. These extra portions typically do not add to the function of the desired protein but their presence can hinder folding or processing of the desired protein. It is thus desirable to be able to produce proteins which are as near as possible to the native protein so these effects can be avoided or reduced. Accordingly, it is also an object of the invention to produce CD4 derivatives, chimeras and other proteins in bacterial host systems that do not contain amino acids derived from host system proteins or that contain only a minimal number of amino acids added to the authentic protein sequence. It is a further object of the invention to provide vectors for producing CD4 chimeras and proteins and other proteins that result in chimeras that do not contain amino acids derived from host system proteins or that contain only a minimal number of amino acids added to the authentic protein sequence. Summary of the Invention

The present invention provides nucleic acid sequences and DNA vectors useful for the production of CD4 chimeric proteins, as well as other heterologous proteins, in Streotomyces. The nucleic acid sequences of the invention comprise the coding sequence for the signal peptide of the Streptomyces lonoisporus tyrosine inhibitor (LTI) gene operatively linked with a propeptide sequence consisting essentially of an amino acid sequence coding for from one to about 6 amino acids, the sequence of said amino acids selected to result in the formation in Streptomyces of a protein product having a homogeneous amino terminus after processing to remove said signal peptide formed on the protein product during synthesis of the protein product. In an alternative embodiment of the invention, the propeptide is omitted and the LTI signal peptide is operatively linked with a nucleic acid sequence coding for a heterologous protein which has been modified to code for the sequence lys-ala- at the 3' end.

In other aspects the present invention provides

Streptomyces cells transformed with a nucleic acid sequence or vector of the invention, and methods of using the nucleic acid sequences of the invention to produce heterologous proteins in Streptomyces.

This invention is more particularly pointed out in the appended claims and is described in its preferred embodiments in the following description.

Detailed Description of the Invention

The present invention provides nucleic acid sequences and DNA vectors useful for the production of CD4 chimeric proteins, as well as other heterologous proteins, in Streptomyces. The nucleic acid sequences of the invention comprise the coding sequence for the signal peptide of the Streptomyces lonoisporus tyrosine inhibitor (LTI) gene operatively linked with a propeptide sequence consisting essentially of an amino acid sequence coding for from one to about 6 amino acids, the sequence of said amino acids selected to result in the formation in streptomyces of a protein product having a homogeneous amino terminus after processing to remove said signal peptide formed on the protein product during synthesis of the protein product. Preferably the amino acid sequence of the propeptide is selected to cause processing of the protein product at a position between the end of the nucleic acid sequence coding for the signal peptide and the beginning of the propeptide seguence. In preferred embodiments of the invention, the amino acid seguence of the propeptide is thr-, thr-pro-ala-ala- (SEQ ID NO: 1), or thrpro-ala-ala-ala- (SEQ ID NO: 2). The propeptide sequence thris more preferred, as it is only one amino acid and provides good yield of protein product. In these embodiments of the invention, the nucleic acid sequence of the propeptide preferably comprises the sequence ACC (coding for thr); ACC CCG GCC GCT (SEQ ID NO: 3) (coding for thr-pro-ala-ala, SEQ ID NO:1) or ACC CCG GCC GCT GCT (SEQ ID NO: 4) (coding for thrpro-ala-ala-ala, SEQ ID NO:2). The nucleic acid sequences of the invention are preferably operably linked with other DNA sequences to form an expression vector which may then be inserted in Streptomyces for production of the heterologous protein. Heterologous proteins thus formed will have a homogeneous N-terminus comprising the sequence of the propeptide.

In an alternative embodiment of the invention, the propeptide is omitted and the LTI signal peptide is operatively linked with a nucleic acid sequence coding for a heterologous protein which has been modified to code for the sequence lys-ala- at the 3'end. The nucleic acid sequence is preferably operably linked with other DNA sequences to form an expression vector which may then be inserted into Streptomyces for production of the heterologous protein. In this embodiment of the invention, signal sequence is cleaved from the heterologous protein at the end of the signal sequence such that the heterologous protein is formed having an N-terminus of lys-ala-X, where X is the remainder of the heterologous protein. This embodiment of the invention is particularly advantageous for production of CD4 derivatives and chimeric proteins.

Applicants have surprisingly and unexpectedly found that by altering only one amino acid at position 2 near the N-terminus of CD4 (in the VI region), a heterologous protein could be formed that is efficiently secreted and correctly processed to remove the entire LTI signal sequence, but which still retained substantial gp120 binding capacity. By altering the nucleic acid sequence coding for the CD4 derivative or chimera such that it codes for lys-ala- at the amino terminus, the product coded for contained an N-terminus amino acid sequence similar to the sheep L3T4 (CD analogue) receptor, which has an N-terminus of lys-ala-.

When it is desired to use this embodiment of the invention to prepare heterologous proteins other than CD4 derivatives or chimeras, the nucleic acid seguence coding for the heterologous protein may be modified to code for lys-ala-at the amino terminus by deleting the bases coding for the two amino acids at the N-terminus and substituting a sequence coding for lys-ala for the deleted bases, or a sequence coding for lys-ala may simply be added to the 3 'end of the coding sequence for the LTI signal peptide. Thus heterologous proteins having a modified N-terminus having the amino acid sequence lys-ala- will be formed in Streptomyces.

Because of the possible deleterious effects of additional amino acids on the function of heterologous proteins, Applicants desired to produce heterologous proteins that have the sequence of the native, protein (or portions of it) with few or no additional amino acids derived from host system proteins, as is not the case when heterologous proteins are formed as fusion proteins. Applicants have surprisingly found that the pro-peptide of the LTI gene can be deleted or modified and heterologous proteins formed that have nearly authentic N-termini and are also efficiently secreted from the Streptomyces host cell.

It became apparent the amino acid sequence surrounding the signal peptide cleavage site can have dramatic effects on. signal peptide processing. There are at least two parameters which can influence processing and consequently, secretion and production levels: the physico-chemical nature of amino acids and the net charge within the region surrounding the cleavage site.

A statistical analysis of eukaryσtic and prokaryotic species signal peptides has shown clear preferences for specific classes of amino acid at specific residues within the region surrounding the signal peptide cleavage site (VonHeijne, FEBS Letters 244:439-446 (1989)). For example, amino acids residues at positions -3 and -1 are generally small, uncharged residues. Alanine is the most frequent amino acid at these positions. Positions -3 and -1 also show a clear bias against certain classes of amino acids; for example, aromatic, charged, hydrophobic residues and proline were not found at position -1 within 78 eukaryotic signal peptides. Within the amino terminus of the mature protein, proline and glycine are seldom found at position +1.

Both experimental data and statistical analysis indicate a clear preference for a net neutral or negative charge within the region surrounding the cleavage site (Li, et al., Proc. Natl. Acad. Sci. USA 8J5:7685-7689 (1988), ϊanane and Mizushima, 263:19690-19696 (1988), Summers, et al., J. Biol. Chem. 264:20082-20088 (1989), Von Heijne, J. Mol. Biol. 192:287-290 (1986) of bacterial signal peptides. Li, et al. supra have shown that when the amino terminus of E. coli alkaline phosphate was mutagenized such that the net charge increased from 0 to +2, the total production of alkaline phosphatase dropped 50 fold. In addition, the time required for the precursor processing into the mature product changed from undetectable for the wild type protein to 30 minutes for the mutant with the net charge of +2. The severity of this defect was somewhat decreased by reducing the net charge to +1. Summers, et al. supra described a variety of hybrid proteins between the B-lactamase signal peptide and chicken muscle triσsephosphate isomerase. The native triosephosphate isomerase is not secreted into the periplasm of E. coli. This protein could be exported into the periplasm, however, when the arginine residue at position 3 of the triose phosphate isomerase was changed to a proline or serine residue. These observations lend credence to Von Heijne's statistical analysis which showed a clear preference for a net negative charge within the cleavage site region.

It has been found that the first two amino acid residues of CD4 (in the VI region) lys-lys- are absolutely required for gp 120 binding activity. In attempts to provide

CD4 derivatives and chimeras having authentic N-termini (e.g.

KK-V1V2), it was found that CD4 derivatives and chimeras could not be produced.

When the CD4 V1V2 regions are fused to the LTI signal seguence, the resulting product is processed in at least two forms. Applicants surprisingly and unexpectedly found that moving the positively charged lysine residues away from the signal peptide cleavage site by inserting up to five amino acids resulted in a CD4 heterologous protein that has a homogenous N-terminus, was efficiently transported into the culture medium, and still retained gp120 binding capacity.

Even more surprisingly, Applicants found that by changing lysine at position 2 of the CD4 molecule to alanine, a homogeneous CD4 protein having a nearly authentic N-terminus could be produced despite the presence of lysine at the N-terminus end. In addition, this heterologous protein was also efficiently secreted into the medium, and retained gpl20 binding capacity. Heterologous proteins produced using the nucleic acid sequences and vectors of the invention have homogeneous N-termini, i.e., the N-terminus of the product has the same amino acid seguence which gives them several advantages over heterologous proteins that are produced as a mixture of forms. Production costs are lower since fewer separation steps are needed in order to obtain the desired product in a purified form. A single product rather than a mixture of heterologous proteins is preferred for regulatory approval. Additionally, natural folding and function of the protein product is encouraged by the absence of amino acids derived from host system proteins.

In other embodiments the invention is directed to DNA vectors for expressing heterologous proteins in Streptomyces which comprise a coding sequence for the heterologous protein operatively linked to a promoter and a nucleic acid sequence coding for the signal sequence of the Streptomyces longisporus tyrosine inhibitor gene operatively linked with a propeptide seguence consisting essentially of an oligonucleotide coding for from one to about 6 amino acids, the sequence of said amino acids selected to result in the formation in Streptomyces of a heterologous protein having a homogeneous amino terminus after processing to remove the signal peptide formed on a heterologous protein during synthesis of the heterologous protein. Preferably the coding sequence for the heterologous protein comprises a sequence coding for a HIV gp120 binding region.

In other embodiments, the invention is directed to DNA vectors for expressing heterologous proteins in Streptomyces which comprises a coding sequence for the heterologous protein operatively linked to a promoter and the Streptomyces longisporus trypsin inhibitor gene signal sequence, wherein the sequence coding for the heterologous protein is modified at its 5' end by adding bases coding for the amino acid sequence lys-ala, or by deleting bases coding for the two amino acids at the 5' end and substituting for the deleted sequence a sequence coding for the amino acid sequence lys-ala. Preferably the coding sequence for the heterologous protein comprises a sequence coding for a HIV gp120 binding region.

Additional embodiments of the invention are directed to methods for using the DNA vectors of the invention, which comprise the steps of introducing a vector of the invention into a Streptomyces host cell, and growing the host cell in a suitable culture medium.

Further embodiments of the invention are directed to Streptomyces cells transfected with the nucleic acid sequences or DNA sequences of the invention.

The CD4 T lymphocyte receptor protein, herein referred to as "CD4", is a surface glycoprotein that interacts with the HIV envelope protein, gp120. This high affinity interaction occurs between gp120 and the extracellular domain of CD4. The extracellular domain of CD4 comprises 4 regions of limited homology with the immunoglobulin variable (V) regions. The approximate boundary domains for the CD4 variable-like regions (V1-V4) are, respectively, amino acids 100-109, amino acids 175-184, amino acids 289-298, and amino acids 360-369. Thus, as used herein, V1 refers to amino acids 1-183 (approximately) and so on.

Removal of the transmembrane and cytoplasmic domains of CD4 results in a soluble receptor protein. Such soluble CD4 proteins, consisting of all or portions of the external region of the human CD4 receptor, have been shown to inhibit HIV-1 infection and virus-induced cell fusion (see e.g., Deen et al., Nature 331: 82-84 (1988). As used herein, the term "HIV gp120 binding domain" refers to any soluble human CD4 molecule (i.e., lacking the transmembrane and cytoplasmic domains) that binds HIV gp120 with essentially the same or greater affinity than the full-length CD4 receptor protein. The binding affinity of CD4 for gp120 can be assayed as described by Arthos et al., (Cell 57:469-481 (1989).

The CD4 molecule of the present invention comprises the minimal HIV gp120 binding region found in domain V1 (amino acids 41-55). Preferably, it includes the V1V2 domains of CD4 or substantially the same sequence (i.e., differs by no more than 15 amino acids). Hence a preferred embodiment of the present invention is V1V2 joined to a human immunoglobulin constant region.

The DNA coding sequence for CD4 is disclosed by Maddon et al. (Cell 42 : 93-104 (1985) and incorporated by reference herein with the noted exception that the mature amino terminal of human CD4 is Lys-Lys and not Gln-Gly-Lys-Lys as reported by Maddon et al. (see, Littman et al., Cell 55:541 (1988). DNA encoding for CD4 is available from various sources, for example, plasmid pT4B (Maddon et al., supra). Alternatively, cDNA molecules encoding the CD4 sequence can be derived from mRNA of T lymphocytes which express CD4 using known techniques (e.g., polymerase chain reaction) or can be synthesized by standard DNA synthesis technigues.

In addition to those proteins exemplified below, one of skill in the art can readily construct DNA molecules which encode further CD4 derivatives. For example, it is within the art to construct amino acid additions, substitutions, deletions, rearrangements (e.g., V1V3 , V1V4) and chemical modifications thereof. Such derivatives, however, must retain the ability to bind HIV gp120.

An inherent advantage of the CD4-Ig chimeras of the present invention is the potential to increase the potency for inhibition/inactivation of an HIV infection relative to soluble CD4. This can be accomplished through the acquisition of Fc effector functions as described more fully below.

The effector functions, which lie primarily within the Fc portion of the immunoglobulin, include: prolonged serum clearance time (Nakamura et al., J Immun 100:376-383 (1968); protein A binding (Deisenhofer et al. Biochem 20:2361-2370 (1981); Fc receptor binding (and Antigen-dependent cellular cytotoxicity (ADCC): mediated through the binding of antibody to Fc receptors on cytotoxic T-cells); complement fixation (e.g., C1q binding); dimerization (Davies et al., Annu Rev Immunol 1:87 (1983); and placental transfer (Morgan et al., Adv Immunol 40:61-134 (1987).

However, when expressing portions of the immunoglobulin heavy chain, some of the effector functions will be lessened or eliminated. An advantage of the present invention, therefore, is that effector functions which are desired may be conferred back to the molecules of the present invention by the addition of amino acids which are associated with those functions. For example, crystallographic studies of Fc fragments of human IgG indicates that dimerization occurs within the hinge region and CH3 domain (see, e.g., Deisenhofer et al., J Biochem 20:2361-2370 (1981). Therefore, by adding amino acids from the hinge region or CH3 domain, one may produce IgG fragments that have the ability to dimerize. Furthermore the Fc effector functions disclosed above may be derived from any appropriate immunoglobulin constant region or isotope. For example, IgG1, IgG2, IgG3, IgG4, IgM, IgA or IgE, but preferably IgG1. Human IgG1 is preferred since this immunoglobulin subclass has been shown to be the most efficient at mediating cell killing by both complement and ADCC (Bruggeman et al., J Exp Med 166: 1351- 1361 (1987). The IgG1 constant region is comprised of several domains: the CH1, Hinge (h), CH2 and CH3 as disclosed by Ellison et al. (Nar 10: 4071-79 (1981)), and incorporated by reference herein.

The immunoglobulin constant region of the present invention may include all or a portion of a human immunoglobulin heavy chain in which the variable region has been deleted and replaced with CD4 or an HIV gp120 binding fragment thereof. In addition, the gp120 binding fragment thereof. In addition, the gp120 binding domain can be joined directly (i.e., synthesized or expressed as a single polypeptide) or indirectly (i.e., coupled after synthesis) to the immunoglobulin constant region.

One embodiment is a human immunoglobulin IgG constant region lacking most (i.e., at least 51%) or all of the CH3 domain. Preferably the immunoglobulin constant region comprises most or all of the CH2 domain, that is the CH2 domain plus or minus the hinge (h) region. Preferred embodiment include -CH1hCH2, -hCH2 and -CH2. However, the present invention need not be limited to a single constant region domain.

Another embodiment of the present invention is an aglycosylated human immunoglobulin constant region, comprising at least one. domain. Preferred embodiments include, but are not limited to, -CH1hCH2CH3, -hCH2CH3, -CH2CH3, -hCH2, -CH2 and -CH3. The most preferred embodiments include, but are not limited to, -hCH2 and -CH2. Preferably the protein of this particular embodiment is produced in a bacterial host system. It is more preferable that the host cell is from the genus Streptomyces. Methods for the preparation of DNA which encodes the heavy or light chain constant region of immunoglobulins are taught, for example, by Robinson et al., PCT Application Publication No. W087/02671, published May 7, 1987. Moreover, one skilled in the art can prepare a DNA molecule encoding the human IgG1 sequence from any cell line expressing IgGl, for example, ARH-77 (from the ATCC) using known techniques (e.g., polymerase chain reaction). Alternatively, the DNA molecule can be synthesized by standard DNA synthesis techniques.

In addition to those proteins specifically exemplified herein, one of ordinary skill in the art can readily construct DNA molecules which encode further Ig derivatives. For example, amino acid additions, substitutions, deletions, rearrangements and chemical modifications thereof. Other examples include, the substitution of deletion of cysteine residues, mutagenesis to increase Fc receptor binding and/or C1q binding, or the introduction of Fc regions from other immunoglobulin molecules. Such derivatives possess some or all of the Fc effector functions as disclosed below.

Depending on which regions of IgG are expressed, the resulting chimeric protein may bind to antibodies to human IgG, Protein A, complement (specifically C1q), and Fc receptors on appropriate cells of the immune system, e.g. macrophages. Furthermore, the binding of complement (and other effector functions) appears dependent on other factors than simply the presence of relevant sequences on the heavy chain constant region, since different subclasses of human IgG which contain a complement binding sequence differ in their complement binding capacity. Thus, other structural features may be involved.

The recombinant DNA molecule and nucleic acid sequences of this invention may comprise additional DNA sequences, for example, a vector comprising a regulatory element, one or more selectable markers, and sequences that code for replication and maintenance functions. The regulatory region typically contains a promoter found upstream from the coding sequence of this invention, which functions in the binding of RNA polymerase and in the initiation of transcription. In other words, the regulatory element or region is operatively linked to the recombinant DNA molecule of this invention. It will be appreciated by one of skill in the art that the selection of regulatory regions will depend upon the host cell employed.

The selectable marker system can be any of a number of known marker systems such that the marker gene confers a selectable new phenotype on the transformed cell. Examples include Streptomyces drug resistance genes such as thiostrepton resistance ribosomal methylase (Thompson et al., Gene 20:51 (1982)) and neomycin phosphotransferase (Thompson et al., supra).

DNA sequences coding for replication and maintenance functions include, for example, those derived from the Streptomyces. pIJ101 derivatives (see, e.g., Keiser et al., Mol Gen Genet 185:223 (1982)) or the Streptomyces SLP1 derived vectors (see, e.g., Bibb et al., Mol Gen Genet 184 : 230 (1981)). The vector of this invention may also contain a marker which permits gene amplification. Such markers which serve to amplify gene copy number in Streptomyces include the gene for spectinomycin resistance (Hornemann et al., J Bacteriol 169:2360 (1987)) and arginine auxotrophy (Altenbuchner et al., Mol Gen Genet 195:134 (1984).

The present invention also relates to a host cell transformed with the recombinant DNA molecule of this invention. Such host cell is capable of growth in a suitable culture medium and expressing the protein encoded by the recombinant DNA molecule of this invention. Such host cell is prepared by the method of this invention, i.e., by transforming a desired host cell with the plasmid of this invention. Such transformation is accomplished by utilization of conventional transformation techniques. Preferred host cells belong to the genus Streptomyces. For example such hosts include, but are not limited to, S. lividans, S. coelicolor. S. albus, and S. lonqisporus. Preferred embodiments include S. lividans and S. lonoisporus. Other host cells which may be suitable include, but are not limited to, mammalian cells, insect cells, yeast and other bacterial cells (e.g., E. coli. Salmonella. Bacillus). Thus, this invention and the products thereof need not be limited to any specific host cells.

For expression of heterologous proteins in Streptomyces, there are a variety of promoters available. Examples include the galactose-inducible promoter of the Streptomyces galactose operon (Fornwald et al., oc. Natl. Acad. Sci. USA 84: 2130 (1987)), the constitutive promoter of the S. lividans β-galactosidase gene (Eckhardt et al., J. Bacteriol 169:4249 (1987) or Brawner et al., U.S. Patent 4,717,666), the S. lonoisporus trypsin inhibitor gene (see EP-A-264, 175, published April 20, 1988), or a temporarily regulated promoter such as that reported in M. echinosporsa (Baum et al., J. Bacteriol 170:71 (1988)). Regions for transcription termination in Streptomyces are derived from the 3' end of several Streptomyces genes, for example the termination signal at the end of the Streptomyces galactose operon or that found at the end of the S. fradiae neomycin phosphotransferase gene (Thompson et al., Proc Natl Acad Sci USA 80:5190 (1983)). Sequences for protein export in Streptomyces include those isolated from the S. lividans LEP-10 gene and the S. longisporus trypsin inhibitor (LTI) gene (see EP-A-264, 175, published April 20, 1988). Preferably the export or signal sequence is derived from but not limited to the LTI gene. More preferably, the signal sequence is modified (e.g., additions, substitutions, deletions and/or rearrangements) as described in the Examples section.

For the initial expression studies, the CD4-lg chimeras as more fully described in the Examples, were fused to the LTI prepro sequence. The Streptomyces replication functions for this expression vector were provided by plasmid pIJ351, which is a derivative of PIJ101 (Keiser et al., Mol Gen Genet 185:223 (1982)). The host strain used was wild- type S. lividans 1326 (Bibb et al., Mol Gen Genet 184:230

(1981)).

The present invention also relates to a method of producing the protein encoded by the recombinant DNA molecule of this invention which comprises culturing the transformed host of the invention in an appropriate culture media and the isolation of such protein. By "appropriate culture media" is meant that media which will enable such host to express the coding sequence of the invention in recoverable quantity. It will be appreciated by one of skill in the art that the appropriate culture media to use will depend upon the host cell employed. The isolation of the protein so produced is preferably from the host's culture medium, that is, the protein of the invention is preferably exported into the culture medium.

The protein(s) of the present invention may be isolated and purified in accordance with standard techniques. For example, extraction, selective precipitation, column chromatography, affinity chromatography or electrophoresis. For example, the IgG chimeric proteins may be purified by passing a solution containing said chimeric protein through a column which contains immobilized protein A or protein G which selectively binds the Fc portion of the fusion protein. See for example, Reis et al., J Immunol 132:3098-3102 (1984). The chimeric protein may then be eluted by treatment with a chaotropic salt or by a change in pH (e.g., 0.3M acetic acid). Alternatively, the protein of the invention may be purified on anti-CD4 antibody columns, or anti-immunoglobulin antibody columns.

The protein and protein product of the present invention may be used for the treatment of HIV viral infections. As a prophylactic, the CD4-Ig chimeras are administered to individuals at high risk to AIDS or individuals who show exposure to HIV by the presence of antibodies to HIV. Administration of an effective amount of the chimeric protein at an early stage of the disease or prior to its onset would act to inhibit infection of CD4⁺ lymphocytes. As a therapeutic, administration of CD4-Ig chimeras to individuals infected with HIV would inhibit extracellular spread of the virus.

The dosage ranges for the administration of the chimeric protein of the invention are those to produce the desired effect whereby symptoms of HIV or HIV infection are ameliorated. This amount administered is selected so as to maintain an amount which suppresses or inhibits secondary infection by syncytia formation or by circulating virus throughout the period during which HIV infection is evidenced such as by presence of anti-HIV antibodies, presence of culturable virus and presence of -24 antigen in patent sera. The presence of anti-HIV antibodies can be determined through use of standard ELISA or western assays for example, antigpl20, anti-gp41, anti-tat, anti-p55, anti-p17 antibodies, etc. The dosage will generally vary with age, extent of the infection, and counterindications, if any, for example, immune tolerance. The dosage can vary from 0.01 mg/kg/day to 50mg/kg/day, but preferably 0.01 to 1.0 mg/kg/day. The chimeric molecule can be administered intravenously, intraperitoneally, intramuscularly, or subcutaneously. if administered parenterally, it can be by single injunction (e.g., bolus) or by gradual perfusion over time.

The proteins of the invention may be used in combination with other agents, for example, in association with agents directed against other HIV proteins, such as reverse transcriptase, protease or tat. An effective therapeutic agent against HIV should prevent virus medicated as well as cell to cell transmission of infection. The proteins may also be used in combination with other antiviral agents, for example, azidothymide (AZT).

The proteins of this invention can also be used as reagents to identify natural, synthetic or recombinant molecules which act as therapeutic agents or inhibitors of CD4⁺ cell interactions. For example, the proteins can be used in screening assays, such as assays for protein interaction measured by ELISA based methodologies to assay for competitors of the CD4 receptor surface domain.

Based on in vitro data showing that soluble CD4 proteins bind to cells expressing HIV env proteins, the proteins of the present invention can also serve as selective targeting molecules for HIV infected cells in vivo. As a target specific carrier protein, CD4-Ig proteins can serve, for example, as the carrier protein for delivery of cytotoxic agents to the infected cells, including the delivery of liposome formulations.

The examples which follow are illustrative, and not to be construed as limiting of the present invention.

EXAMPLES

Enzymes used in the genetic manipulations were obtained from commercial sources and were used substantially in accordance with the vendor's instructions. Except where otherwise indicated, procedures were carried out substantially as described by Maniatis et al.. Molecular Cloning. Cold Spring Harbor Laboratory, 1989.

Example 1 Preparation of CD4-IgG Chimeras

Using recombinant DNA manipulations, the coding sequence for the V1V2 domains of CD4 are followed by the Hinge-CH2 or Hinge-CH2-CH3 regions of human in various plasmids, which function in various, host cells.

The V1V2 region contains amino acids 1-183 (see

Maddon et al., Cell 42: 93-104 (1985) and Littman et al., Cell, 55:541 (1988)). The Hinge-CH2 and Hinge-CH2-CH3 regions comprise IgG1 amino acids. 97- 228 and 97-330, respectively (see Ellison et al., NAR 10::4071-79 (1982). A) V1V2-hCH2DHFR and V1V2-hCH2CH3DHFR (CHO cells):

The expression vector ST4184.DHFR, disclosed in EP-A-331,356 (published September 6, 1989), was modified to create plasmid V1V2183DHFR: Plasmid ST4184.DHFR was cut with EcoRI and Nhel, and a 682 base paid fragment, containing nucleotide 1-682 of the CD4 DNA was isolated. This fragment was ligated to a synthetic linker encoding CD4 amino acids 177-183 followed by a TAA termination codon. In addition, the synthetic linker created a HindIII site by changing nucleotides 693 (G to A) and 696 (C to T) of the CD4 sequence. These nucleotide changes do not alter the amino acid sequence. The resulting fragment, which is flanked by EcoRI and Xbal ends, was ligated into the EcoRI and Xbal site of another sT4184.DHFR plasmid. The sequence of the synthetic linker is substantially as follows:

5' CTAGCTTTCCAGAAAGCTTCCTAAT 3'

3' GAAAGGTCTTTCGAAGGATTAGATC 5'

The resulting plasmid, V1V2183DHFR, contains the mouse β-globin promoter, DHFR gene, SV40 poly A region and the CD4 coding sequence for amino acids 1-183.

V1V2183DLHFR was subsequently linearized by HindIIl. which cuts after the V1V2 coding sequence.

Next, a BanII-AvaI fragment which encodes the HingeCH2 region (nucleotides 299-677, Ellison et al., supra) was isolated from human IgGl cDNA. (The Hindlll site was introduced by PCR mutagenesis.) Each fragment was ligated into the HindIII site of a different V1V2183DHFR plasmid with synthetic linkers:

i) A Hindlll-BanII linker; to ligate the 3' end of the V1V2 gene with the 5' end of Hinge CH2 or Hinge-CH2CH3.

5' AGCTTCCΛAGGTGGΛGCC 3 '

3' AGGTTCCACC 5' ii) An Aval-HindIII linker; used to introduce a stop codon after the CH2 coding sequence and to join the 3' end of the Hinge-CH2 fragement with the 5' end of the V1V2 sequence.

5' CCGAGAGTAGTGACTGCAGA 3'

3' CTCATCACTGACGTCTTCGA 5'

These constructs maintain the correct reading frame without the addition of non-CD4 or non-IgG1 amino acid residues at the HindIII junction and are herein referred to as V1V2-hCH2DHFR and V1V2-hCH2CH3COS, respectfully.

B) V1V2-hCH2COS and V1V2-hCH2CH3COS (COS Cells):

The vectors V1V2-hCH2DHFR and V1V2-hCH2CH3DLHFR were digested with NheI. Fragments of 1,817 and 2,147bp, respectfully, were isolated which comprises the carboxy end of V2 the complete Hinge-CH2 or Hinge-CH2CH3 coding region, the bovine polyA region, the mouse β-globin promoter, and part of the mouse DHFR coding region.

The vector V1V2183COS2 (described below) contains the Rouse sarcoma virus LTR, the SV40 early promoter, VIV2, the SV40 poly A early region, and the mouse β-globin promoter operatively linked to the mouse DHFR coding region. When V1V2183COS2 was digested with NheI, a 1,723 bp fragment containing the carboxy end of V2 to the mouse DHFR coding region was removed. The NheI fragments from V1V2-hCH2DHFR and VlV2-hCH2CH3DHFR were then ligated into the NheI sites of vector V1V2183COS2, which restores, in the correct reading frame, the carboxy end of V2 and the DHFR gene, respectively. The resultant vectors are herein referred to as V1V2-hCH2COS and V1V2-hCH2CH3COS, which was used to transect COS cells.

Construction of V1V2183COS: To create plasmid V1V2183COS2, plasmid V1V2183COS2, plasmid V1V2183DHFR was digested with EcoRI and Xbal. The EcoRI site was filled in, and a 707 bp fragment consisting of the V1V2 coding region was isolated. This fragment was ligated to plasmid Rst4COS2 (described below), which had been cut with Smal and Xbal to delete the sT4 coding region. The resulting plasmid was V1V2183VOS2.

Construction of Rst4COS2: To create plasmid

Rst4COS2, plasmid st4DHFR (Maddon et al., PCT/WO88/01304, published February 25, 1988) was digested with SmaI and EcoRI. The EcoRI site was filled in and a 338 bp fragment containing the SV40 early promoter was isolated. This fragment was ligated into Rst4DHFR (see below), which had been cut with BamHI and the sticky ends filled in. The resulting plasmids were screened for orientation, and one in which the SV40 early promoter is in the opposite orientation to RSV LTR was selected as plasmid Rst4COS2.

Construction of RST4DHFR: Plasmid TND (Connors et al., DNA. 7 : 651-661 (1988) was digested with Bg1II and HindIII. A 600 bp fragment containing the Rous Sarcoma Virus (RSV) LTR was isolated. Using a commercially available linker (New England Biolabs, Beverly, MA) consisting of a HindIII "sticky end", a SmaI site, an a EcoRI "sticky end", the RHV LTR fragment was ligated to plasmid ST4DHFR (Maddon et al., supa) which had ben digested with BgIII and EcoRI to delete the SV40 early promoter.

C) OmpAy1V2-hCH2 and OmpAV1V2-hCH2CH3 (E. coli)

The vectors V1V2-hCH2DHFR nd V1V2-hCH2CH3DHFR were digested with Af1III and XbaI. Isolated were fragments of 746 and 1,041bp; respectfully which contains part of VI (approximately 73 amino acids) and the complete Hinge-CH2 or Hinge-CH2CH3 coding regions. These Af1II-XbaI fragments replaced the Af1II-XbaI fragments of plasmid OmpAV1V2, disclosed in EP-A-331,356 (published September 6, 1989), to create a vector containing the lambda p_L promoter and the ompA signal sequence fused to either V1V2-hCH2 or V1V1-hCH2CH3. The vectors, which function in E. coli are here in referred to as OmpAv1V2-hCH2 and OmpAV1V2-hCH2CH3.

D) V1V1-hCH2-Tkk, V1V2-hCH2-KA, V1V2-hCHLStrept and

V1V2-hCH2CH3strent (Streptomyces)

The vectors V1V2-hCHL2DHFR and V1V2-hCH2CH3DHFR were digested with Af1III and XbaI. Isolated were fragments of 746 and 1,041bp, respectfully which contains part of VI and the complete Hinge-CH2 or Hinge-CH2CH3 coding regions. These Af1III and XbaI. fragments replaced the AF1II -XbaI fragments of plasmid 12B1. Plasmid 12B1 contains V1V2 operatively linked to the Streptomyces lividans longisporus trypsin inhibitor (LTI) promoter and signal sequence (see, EP-A-264,175, published April 20, 1988), as well as Streptomyces replication functions as found on plasmid pIJ351 (Keisert et al., Mon Gen Genet 185:223 (1982)). The resultant vectors herein referred to as V1V1-hCH2strept and V1V2-hCH2CH3strept, comprise the LTI signal sequence fused or linked to V1V2-Hinge-CH2 or V1V2-Hinge-CH2CH3. These vectors are designed to function in Streptomyces and export V1V2-hCH2 or V1V2-hCH2CH# into the culture medium. However, the naturally occurring LTI protein is expressed as a prepro protein, that is, the signal sequence is cleaved upon secretion and the pro-sequence is cleaved extracellularly. As a result, the LTI protein isolated can have heterologous amino terminal ends. The expression of VlV2-hCH2 or VlV2-hCH2CH3 from V1V2-hCH2strept and V1V2-hCH2CH3strept, respectively, produce this anomaly; i.e proteins with heterologous amino terminal ends. The ratio of pro-protein to mature protein i.e. lacking all signal sequence amino acid residues can be shifted by variations in culture conditions, e.g. pH, dissolved O₂ etc.

As an alternative, the wild-type LTI prepro sequence i.e. signal sequence plus pro-peptide) has been modified to produce proteins with homogeneous amino terminal ends. The wild-type prepro sequence comprises the following amino acids:

Met-Arg-Asn-Thr-Ala-Arg-Trp-Ala-Ala-Thr-Leu-Ala-Leu-Thr-Ala- Thr-Ala-Val-Cys-Gly-Pro-Leu-Thr-Gly-Ala-Ala-Leu-Ala-↓-Thr- Pro-Ala-Ala-Ala-Pro-Ala-Ser; wherein ↓ denotes the cleavage site between the signal sequence and pro-protein. In one construct, the pro-protein sequence was deleted and replaced with Thr (T). (The amino terminus of the mature CD4 protein is KK). Hence the amino terminus of this chimeric CD4-Immunoglobulin protein is Thr-Lys-Lys (TKK). In another construct, the pro-peptide sequence and the first two amino acids of the mature chimeric protein were deleted and replaced with Lys-Ala (KA). The vectors that encode VlV2-hCH2 are hereby referred to as VlV2-hCH2-TKK and VlV2-hCH2-KA, respectively. These vectors encode VlV2-hCH2 proteins having a homogeneous amino terminus of Thr-Lys-Lys or Lys-Ala, respectively.

The complete nucleotide sequence for vector V1V2-hCH2-KA and the corresponding amino acid sequence for the chimeric protein V1V2-hCH2-KA is disclosed as follows:

Wherein the signal seguence begins at nucleotide 648 and ends at nucleotide 731; VI begins at nucleotide 732 and ends at nucleotide 1286, the Hinge region begins at nucleotide 1287 and ends at nucletotide 1331, CH2 begins at 1332 and continues to the end of the coding seguence.

Example 2 Expression of CD4-IgG Chimeras

A) V1V1- CH2DHFR and V1V2hCH2CI,H3DHFR in CHO cells:

V1V2-hCH2 was transfected into CHO cells via electroporation. 15μg ofVlV2-hCH2 DNA suspended in 15 μl phosphate buffered sucrose (PDSucrose) (272mM sucroe, 7 mM sodium phospate pH7.4, 1 mM MgCl2, sterile filtered) was mixe with 1.0 × 107 CHO cells (0.8ml) and incubated on ice for 15 min in a Gene Pulser cuvette (Dio-Rad, Richmond, Calif.). The cuvette was then placed in the electrode chamber of the Gene Pulser (Dio-Rad) and one pulse of 700 volts/3_/xFd (time constant=0.7) was delivered to the cells. The cells were placed on ice for 10 minutes and then diluted to 10 ml with Growth medium (containing lxITS [insulin, transferrin, selenium; Collaborative Research, Bedford, MAJ, 1x Lipides [Gibo, Grand Island, NY]). Cells were then plated onto 96 well microtiter plates at 3 × 10³ cells/well on a total of 5 plates. After 48 hours the medium was changed to Selection medium (i.e., nucleoside-free Growth medium) for DHFR selection.

Cells were maintained on Selective medium for approximately 4 weeks. Western blot analysis was performed using 15% SDS-PAGE gel electrophoresis and anti-sT4 polyclonal antibodies. The V1V2-hCH2 gene product was identified by a single positive band at a molecular weight of approximately 40,000 daltons. Cells from positive wells were moved to 24 well plates and maintained on Selective medium, until ready for scale up. The two clones (3F8 and 2F6) were chosen for scale up. Clone 3F8 was amplified in Selection medium containing 50 nM methotrexate. Cells were replated onto 96 well microtiter plates at 3 × 10³ cells/well. Amplification was started at methotrexate levels of 50 nM and were to be increased on each successful round of amplification according to standard protocols.

Similarly, the V1V2-hCH2CH3 vector was transfected into CHO cells by electroporation. Positive clones were selected by growing the cells on nucleoside-free medium and screening the supernatant of healthy wells by Western blot analysis using anti-sT4 antibodies. The resultant positive clones were then amplified using increasing amounts of methotrexate.

B) V1V2-hCH2COS and V1V2-hCH2CH3COS in COS cells: COS cells were obtained from the ATCC (Rockville, MD) and grown according to ATCC recommended protocols. The. cells were then transfected with 10 μg of plasmid DNA in 2.5% Nu serum (Collaborative Research, Cambridge, MA) in DMEM (Dulbecco's minimal essential medium) )GIBCO Grand Island, NY with 400 g/ml DEAE-dextran (Pharmacia) and 100uM chloroquine (Seed et al., Proc Natl Acad Sci 84:3365-3369 (1987)). Following a 4 hour incubation at 37°C, the medium was removed and the cells were treated with 2ml of 10% dimethylsufoxide in PBS (phosphate buffered saline) for 3 min at 25°, the medium was removed and replaced with DMEM.

Both V1V2-hCH2 and V1V2-hCH2CH3 were expressed and exported into the culture media. In a standard transfection in a 6 mm dish (about 10⁶ cells), 20μg of VlV2-hCH2 had accumulated int he medium at 70 hrs post-transfection. Approximately 0.4μg of material was present in the nucleifree cell lystate. For comparison, the corresponding values for sT4, a soluble CD4 protein which had been expressed at high levels in several mammalian cell types were 20μg in the supernatant and 1μg in the cell lysate. The V1V2-hCH2 protein in the medium migrated as two closely spaced fragments. Under nonreducing conditions, the V1V2-hCH2 migrated principally a monomer although approximately 10% of the sample appeared to migrate as a dimer. The V1V2-hCH2 protein in the medium bound to gP120 coupled to a sepharose resin.

V1V2-hCH2CH3 was analyzed in the same manner as described above. This protein also expressed well as a secreted product, accumulating to 20-25μg in the medium and 1-2μg in the cell lysate. The protein in the medium migrated as a single fragment of approximately 50 kD by SDS/PAGE under reducing conditions, while the cell lysate sample migrated as two equivalent bands of approximately 40 and 50 kD. Under nonreducing conditions the V1V2-hCH2CH3 migrated principally as a single fragment with a molecular weight consistent with dimer formation. This protein bound to gpl20 coupled to sepharose.

C) OmpAV1V2-hCH2 and OmpAV1V2-hCH2CH3 in E. coli

The E. coli lambda lysogens, AR58 (Debouck et al., EP-A-0,216, 747, published Aril '1, 1987) and Ar120 (Mott et al., Proc Natl Acad Sci 82: 88-92 (1985)) were transformed with OPMAVlV2-hCH2 and OmpAv1V2-hCH2CH3, respectively, using standard procedures. Expression of V1V2-hCH2 and V1V2-hCH2CH3 in strain AR58 was accomplished by raising the temperature of the culture media from 32ºC to 42ºC (see for example, Rosenberg et al., Meth Enzymtol 101:123 (1983)).

The expression of V1V2-hCH2 and V1V2-hCH2CH3 in strain AR120 was accomplished by the addition of nalidixic acid (Nal) to the culture media (see for example, Mott et al., Proc Natl Acad Sci 82:88-92 (1985) as follows. A culture of AR120 was grown at an optical density (at 650nm) of 0.4 absorbance units at 37ºC whereupon nalidixic acid was added to a final concentration of 50μg/ml. The culture was maintained at 37ºC in a shaker incubator for approximately 5 hours at which point the cells were centifuged and subsequently chilled to stop the induction.

For both the heat and Nal inductions, the cell pellets and the clarified culture media (i.e. centrifuged media) were assayed for V1V2-hCH2 and V1V2-hCH2CH3 expression. Both chimeric proteins expressed well, however only a small percentage of V1V2-hCH2CH3 was detected in the culture media (i.e. exported).

D) V1V2-hCH2-TKK, V1V1-hCH2-KA, V1V2-hCH2 strept and

V1V2-hCH2CH3strept in Streptomyces

Plasmids V1V2-hCH2-TKK, V1V2-hC2strept and V1V2-hC3strept were transformed into S. lividans strain 1326 (Bibb et al., Mol Gen Genet 184:230 (1981)) using standard procedures (see Hopwood et al., Genetic Manipulation of Streptomyces - A Laboratory Manual, F. Crowe & Sons, Ltd., Norwich, England (1985)). Transformants were selected by overlaying the transformation plates with agar (0.4%) containing 100μg/ml thiostrepton. Colonies which express the protein (s) of interest are then grown in trypticase soy broth media supplemented with 5 μg/ml thiostrepton. All of the chimeric proteins were secreted into the culture media, however the V1V2-hCH2 constructs were expressed at a higher level than the V1V2-hCH2CH3 construct. V1V2-hCH2-KA was expressed at approximately 14 mg/L.

Example 3 Characterization of CD4-IqG Chimeras

A) Subunit Structure:

The native molecular form of secreted V1V2-hCH2- KA and V1V2-hCH2CH3 were analyzed by electrophoresis on a 15% SDS-polyacrylamide gel under non-reducing conditions. The protein bands were identified by Western Blot analysis and the results are shown in Table I.

B) Antibody Recognition:

All of the chimeric proteins made by the present invention are recognized by both anti-CD4 antibodies (Deen et al., Nature, 331:82-84 (1988)) and anti-human IgG Fc receptor antibodies (Cappel, Malvern, PA.) See Table I. C) gp120 Binding:

The chimeric proteins made in S. lividans. COOS an CHO cells all exhibited gp120 binding (refer to Table I) either by the immunoprecipation assay as described by Arthoos et al. (Cell, 57:469 (1989)) or by binding to gp120 immobilized to sepharose as described below:

Samples are diluted to a total volume of 300μl with precipitation (ppt) buffer (phosphate buffered saline containing 0.5% dry milk and 0.1% NP40). A 50% slurry of gp120 sepharose beads (100μl, described below) of wa added to the diluted sample and incubated at 4ºC for 60 min. The sample was spun (30 sec) in a microfuge to precipitate the chimeric protein/gp120-Sepharose complex. The complex was washed 5 times with 400μl of ppt buffer and once with ice cold phosphate buffered saline. The complex was pun dow again and resuspended in 60μl loading buffer (Laemmli, Nature, 227: 680 (1970)), boiled for 5 minutes and loaded onto a 15% polyacrylamide gel followed by Western blot analysis. The chimeric proteins expressed in E. coli were not tested for gpl20 binding.

gp120 was obtained from in-house sources. It is also available commercially, e.g., American BioTechnology, Cambridge, MA. The sepharose beads were prepared as follows: 1.0 liter of Sepharose C1-6B (Pharmacia Fine Chemicals, Piscataway, NJ) was reacted with 52.5 ml epibromohydrin in a basic solution of 0.5N NaOH/30% tetrahydrofuran (THF) at 40ºC for four hours. The activated sephrose was collected by filtration on a sintered glass funnel and was washed extensively with 30% THF to remove unreacted epibromohydrin. The product was further washed with water until the washing was neutral. This gel was then resuspended with ethylenediamine (50ml) overnight at room temperature. The gel was filtered and unreacted ethylenediamine was removed by washing the gel with 0.1M acetic acid followed by water. The gel was resuspended in 1.0 liter of water, then was reacted with succinic anhydride (25 gm) at pH 6.0. The gel was further washed with 1.0 liter sodium carbonate solution (0.2M) followed by water until the washing was found to be neutral. This gel was finally washed with isopropyl alcohol and stored at 4ºC for further use as a moist powder.

D) Protein A & Protein G Binding:

The V1V2-hCH2 chimeric proteins expressed in S. lividans and COS cells do not bind to protein A or protein _G. In contrast, the V1V2-hCH2CH3 chimeric protein expressed in S. lividans and COS cells do exhibit binding affinity to both proteins A and G. albeit at different affinities. The V1V2-h CH2CH3 produced in Streptomyces has a lower affinity for proteins A and G compared to the V1V2-hCH2CH3 produced in COS ce411s. The results are summarized in Table I.

i

Example 4 - Construction of fusion plasmids TKK-VIV2, TPAA-AA-VIV2 AND and KA-VIV2.

Fusion plasmids TKK-VIV2, TPAA-VIV2, TPAAA-VIV2 AND KA-VIV2 were constructed as follows from plasmid 12B1.

Construction of plasmid 12B1. Plasmid 12B1 contains VIV2 operatively linked to the Streptomyces longisporus trypsin inhibitor (LTI) promoter and signal sequence (see EP-A-264, 175, published April 20, 1988) as well as Streptomyces replication functions as found on plasmid pIJ351 (Keiser et al. Mol Gen Genet 185:223 (1982).

The parental plasmid 12B1 was constructed as follows. The BbvI cleavage site within the coding sequence for the CD4 signal peptide (between nucleotides 148 and 149 of the CD4 DNA sequence; Maddon, et al. Cell 42: 93-104 (1985)) was moved by site directed mutagenesis (Kunkel, Proc. Natl. Acad. Sci. USA 82: 488-492)) such that the BbvI cleavage site was placed between nucleotides 150 and 151. This mutation (named 1478) was inserted into a sCD4 minigene which contained the coding sequence for amino acid residues 1 - 129. An EcoRI + HindIII fragment containing the 1478 mutation was transferred from M13mpl8 into pUC18 to generate pUCV1pV2(1478). Following BbvI digestion of pUCV1pV2 (1478), it was treated with the Klenow fragment of DNA polymerase I to fill-in the 5' single-stranded sequence, then digested with HindIII. The HindIII-blunt end fragment resulting from these manipulations was cloned into pLTI450 which had been digested with AccI, treated with the DNA polymerase I Klenow fragment and digested with HindIII. The resulting plasmid, 12B1/1477, contains a sCD4 minigene (amino acid residues 1-129) fused to the coding sequence of the LTI signal sequence such that the expressed V1V2 protein will contain at its amino terminus the 6 amino acid LTI pro peptide plus residues 1 and 2 of the mature LTI protein. A Streptomyces replicon and selectable marker were cloned into 12B1/1477 by inserting pIJ351 (Kieser, et al., Mol. gen. Genet. 185:223-238 (1982) using the unique PstI site within both plasmids. To create a complete V1V2 minigene (amino acid residues 1-183) in the 12B1/1477 plasmid an Af1II + XbaI fragment from DHFR V1V2 183#7 in Example 1 was inserted into 12B1/1477 which had been digested with Af1II and XbaI. The resulting plasmid was 12B1.

Construction of plasmid PLTI450: A 0.92 kb SacI-KpnI fragment containing the LTI gene was inserted into pUC18 which had been digested with Sad and KpnI. This plasmid, pLTI520, was partially digested with EagI and totally digested with Sail then ligated to a synthetic linker (2x stranded) which had EagI and Sa1I ends:

5'-GGCCGCCGCCCCCGCG (SEQ ID NO: 5)

CGGCGGCGGGGGCGCAGCT-5 ' (SEQ ID NO: 6)

Plasmids obtained from this ligation where screened for insertion of the synthetic linker into the EagI site located approximately 0.5 kb from the Sad site. This EagI site is located at base pair 86 with respect ot the 5' end of the LTI gene. The resulting plasmid contains the LTI promoter and the coding seguence for the signal peptide and signal peptide cleavage site.

To create the fusion plasmids TKK-VIV2, TPAA-VIV2, TPAAA-VIV2 and KA-VIV2 oligonucleotide mutagenesis was used to delete the LTI pro peptide coding sequence and create the amino acid coding seguence of choice following the signal peptide cleavage site. The M13 single stranded DNA used for these mutagenesis was mp18V1V2/12B1 which was created by inserting a 1.1 kb EcoRI-Xbal fragment from 12B1 into M13mp18 digested with EcoRI and Xbal. Summarized in Table 3 are the oligonucleotides used for site directed mutagenesis.

Table 3 - Oliqonucleotides used to create V1V2 derivatives

Derivative Oligo- Oligonucleotide

Nucleotide Sequence

TKK-V1V2 2214 5'GCCCAGCACCACTTTCTTGGTGGCGAGCGCGGCTCC-3'

TPAA-V1V2 2253 5'-GCCCAGCACCACTTTCTTAGCGGCCGGGGTGGC-3'

TPAAA-V1V2 2254 5'-GCCCAGCACCACTTTCTTAGCAGCGGCCGGGG-3'

KA-V1V2 2216 5'-GCCCAGCACCACGGCCTTGGCGAGCGCGGCTCC-3'

In Table 3, nucleotide sequence 2214 is SEQ ID NO: 7; nucleotide sequence 2253 is SEQ ID NO: 8; nucleotide seuqence

2254 is SEQ ID NO: 9; nucleotide sequence 2216 is SEQ ID

NO:10. A 0.7kb EcoRI-Af1II fragment isolated from the RF form of the mutagenized mp18V1V2 was exchanged with a 0.75kb EcoRI-Af1II fragment from 12B1 to generate plasmids pVlV2-2214 (or PTKK-V1V2), pVIV2-2253 (or pTPAA-VIV2), pV1V2-2254 (or pTPAAA- VIV2) and pV1V2-2216 (or pKA-V1V2). pV1V2-2214, also named PTKK-V1V2, was created using oligonucleotide 2214 as shown in Table 3 and contains the LTI signal sequence linked to the modified propeptide sequence threonine, and the modified propeptide sequence threonine is linked to V1V2. pVlV2-2253, also named pTPAA-V1V1 was constructed using oligonucleotide 2253 as shown in Table 3 and contains the LTI signal sequence linked to the modified propeptide sequence thr-pro-ala-ala. PV1V2-2254, also named pTPAAA-V1V2 was constructed using oligonucleotide 2254 as shown in Table 3 and contains the LTI signal peptide sequence linked with the modified propeptide sequence thr-pro-ala-ala-ala, which is in turn linked with V1V2. pV1V2-2216, also named pKA-V1V2 was constructed using oligonucleotide 2216 as shown in Table 3 and contains the LTI signal peptide sequence which is in turn linked with V1V2. Example 5 - Construction of βgal *KK-V1V2 fusion plasmid

βgal *KK-VIV2 was constructed by mutagenesis from plasmid pøgal sT4/7. Pβgal sT4/7 was constructed as follows. Construction of pβgalsT4/7: Plasmid p*galsT4/7 was constructed from plasmids pUCst4, pIJ702 and p3SSXMCP. p3SSXMCP is a pUC9 (Viera and Messing, Gene 19:259(1982) derivative which contains the β-galactosidase promoter and signal sequence. Twenty-six base pairs downstream of the coding sequence for the signal peptide cleavage site is an XmnI site (Eckhardt, et al., J. Bacteriol. 169 : 4249(1987)). Inserted into this site was a synthetic linker which contains a BamHI recognition sequence (New England Biolabs, Beverly, Massachusetts) to generate plasmid p3SSX10. This vector was treated with BamHI and reverse transcriptase followed by XhoI digestion. Ligated to this vector was a fragment which had an NcoI end treated with reverse transcriptase to generate a blunt end and a SalI end. Ligation of the filled-in BamHI site with the filled-in Ncol site recreated both the BamHI and NcoI sites. The resulting plasmid was p3SSXMCP. pUCsT4 was treated with XbaI and reverse transcriptase followed by NcoI digestion. This 1.1 kb NcoI-Xbal (RT) fragment was then inserted into p3SSXMCP which had been treated with SacI and T4 DNA polymerase I followed by NcoI digestion. The Streptomyces replication functions were provided by pIJ702 which was inserted into p3SSXsT4 via the unique Bg1II site on both plasmids. The resulting plasmid was pβgalsT4/7.

A 1.6kb EcoRI-Af1III fragment from pβgal sT4/7 was exchanged with the 0.75kb EcoRI-Aflll fragment from mpl8VlV22214 to generate mp18BgalV1V2. Site directed mutagenesis was used to delete the coding sequence for amino acid residues -9 to -1 of the CD4 signal peptide, change the arginine residues at positions -3 and -4 of the βgal signal peptide to aspartamate and glutamate, respectively, and delete the coding sequence for residues 1 to 8 of the mature β-galactosidase protein. These mutations were made using oligonuclotide 2256 whose sequence is

5'-GCCCAGCACCACTTTCTTCGCCGCGTCCTCTACGGCGCCTG-3'

(SEQ ID NO: 11).

A 1.6kb (EcoRI-Af1II fragment from the RF form of mp18βga1V1V2-2256 was exchanged with the 0.75kb EcoRI-Af1II fragment from plasmid 12B1. The resulting plasmid, pV1V2βgal-2256 (βgal *KK-V1V2) has the following characteristics: (1) V1V2 expression is directed by the Bgal promoter, (2) the coding sequence of the βgal signal peptide has been altered such that amino acid residue -4 is a glutamate and -3 is an aspartamate, and (3) V1V2 is fused to the βgal signal sequence such that the predicted N-terminal amino acid sequence of V1V2 is Lys-Lys.

Example 6 - V1V2 Expression in S. lividans

Plasmids pV1V2-2214, pV1V2-2253, pV1V2-2254, pV1V2-2216, and pV1V2Bgal-2256 from Example 4 were transformed into S. lividans 1326 (Bibb et al., Mol. Gen. Genet. 184: 230-240

(1981) using standard procedures (Hopwood, et al., Genetic Manipulation of Streptomyces - A Laboratory Manual, F. Crowe & Sons Ltd., Norwich, England (1985). Transformants were selected by overlaying the R2YE transformation plates with 3ml of 0.4% agar + 100μg/ml thiostrepton. Transformants which express the protein of interest were grown in trypticase soy broth or ME1 + 5% HycaseSF (40g/liter glucose, 50g/liter HycaseSF, 50g/liter Hysoy, 1g/liter yeast extract, 1g/liter CaCO4, and 0.001g/liter CoC12 supplemented with 5μg/ml thiostrepton. Cell-free supernatants used for the gp120 binding assay and the gp120 affinity chromatography were harvested from cultures grown in ME1 + 5% Hycase.

The V1V2 derivatives when expressed in S. lividans were initially characterized by immunoblotting.

Immunoblot Procedure - Cell free supernatants were separated on a 15% polyacrylamide (30:0.8 acrylamide:bis)- sodium dodecy] sulfate gel (Laemmli (1970) Nature 227: 680-685), then transferred to nitrocellulose (Towbin, et al. (1979) Proc. Natl Acad. Sci. USA 76: 4350-4354). The nitrocellulose filter was processed to detect V1V2 (Brawner et al. (1985) Gene 40: 191-201) using rabbit anti-serum prepared against denatured sCD4 protein. The bound antibody was detected with ¹²⁵I-protein A. The immunoreacting proteins were visualized by autoradiography.

The TKK-VIV2, TPAA-VIV2, TPAAA-VIV2 , KA-VIV2 and βgal*KK derivatives were produced as a single, immunoreacting band. A doublet was produced from the KK-V1V2 derivative. One of the KK-V1V2 proteins co-migrated with V1V2 produced from CHO cells. The other band, which migrated with a slower mobility than the V1V2 reference protein, may be the result of incomplete processing of the LTI signal peptide or processing at an alternative cleavage site within the LTI signal peptide. Because of the heterogenous nature of the product expressed by the KK-V1V2 derivative no additional studies were undertaken.

The biological activity of these V1V2 derivatives was then determined by gp120 binding. As judged by quantitative gp120 immunoprecipitation (Arthos, et al., Cell 57:469-481 (1989), the Streptomyces produced V1V2 proteins were fully active. Results of the immunopreciptation assay are shown in Table 2. All of the VIV2 derivatives tested produce gp120 binding of greater than 90%.

V1V2 production of the TKK-V1V2, TPAA-V1V2, TPAAA-V1V2, and β-gal*KK derivatives was compared to that obtained from plasmid 12B1 which produces V1V2 at levels equal to or greater than 100 mg/l. V1V2 expression levels of the TPAA-V1V2 and TPAAA-V1V2 derivatives were comparable to that obtained from 12B1; KA-V1V2 expression was 70% of that obtained from 12B1; TKK-V1V2 expression was 30% of that from 12B1; βgal*KK expression was approximately 1 mg/l.

Example 7 - N-terminal amino acid analysis of heterologous proteins produced bv plasmids PVIV2-2214. PVIV2-2253. PVIV2- 2254. pVIV2-2216 and PVIV2 βgal-2256.

A. Purification of VIV2 from Streptomyces media for N-terminal sequencing by using GP-120 sepharose affinity column

1. Preparation of gp120 Sepharose

Purified gp120 was immobilized on Sepharose CL-6B (Pharmacia) through amino group using active ester chemistry. Approximately 0.25 mg of gp120 was coupled to 1ml of resin and binding capacity for sT4 was determined as 20 μg/ml.

2. Sample preparation from media.

Fermentation media containing VIV2 constructs was diluted 5-fold in column eguilibration buffer (see below), centrifuged for 15 minutes at 20,000 × g and filtered through 0.2μ Acrodisc low protein binding filter (Gelmen).

3. Gp120 affinity chromatography

Gp120 Sepharose column (1.6 cm × 6 cm) was equilibrated with 50mM Hepes pH 7.5, 150mM NaCl. The prepared media sample was applied to the column, washed with equilibration buffer until the absorbance reached to baseline, washed again with 50mM Hepes pH7.5, 500mM NaCl to remove any nonspecifically bound impurities. V1V2 was eluted with 0.1 M acetic acid. 4. Removal of peptide impurities from affinity purified V1V2 by using reverse phase HPLC V1V2 eluted from affinity column was made to 0.05% TFA solution and applied to C3 RP-HPLC column (DuPont Pro 10/300, 4.6cm × 250 cm) equilibrated in 0.05% TFA. The column was eluted at 1ml/min for 60 minutes with linear gradient of o60% acetonitrile in 0.05% TFA. The V1V2 product was eluted as a single peak at 50 minutes. Pooled fractions were concentrated in Centricon 3 (Amicon) and used for N-terminal seguencing and amino acid analysis.

B. N-Terminal Sequence

The N-terminal amino acid sequence was identified to determine if signal peptides processing had occurred at the predicted cleavage site. The V1V2 proteins were purified from culture supernatant using gpl20 sepharose affinity chromatography and reverse phase HPLC. The partially purified V1V2 preparations were further purified by polyacrylamideSDS gel electrophoresis followed by electroelution into a polyvinylidene difluoride (PVDF) membrane in preparation for N-terminal amino acid sequencing (Matsudaria, J. Biol. Chem. 262:10035-10038 (1987)). The N-terminal amino acid sequence was determined using a gas-phase protein seguenator. This analysis showed that the V1V2 proteins TKK-V1V2, TPAAKK-V1V2 , TPAAAKK-V1V2 and KA-V1V2 fused to the carboxy terminus of the LTI signal peptide were correctly processed at the LTI signal peptide cleavage site. The N-terminal amino acid sequence is summarized in Table 2. The N-terminal end of the V1V2 peptide is LYS-LYS. The Bgal*KK-V1V2 derivative, however, was not processed at the natural signal sequence cleavage site. In this derivative, the amino acids preceeding LYS-LYS are derived from the coding sequence at the 3' end of the β-gal * signal peptide. Signal sequence cleavage for the R(-4)E/R(-3)D signal peptide mutation occured within the βgal signal sequence between positions -8 and -7. Table 2 - Characteristics of the VIV2 derivatives

gp120

Derivative N-terminal Amino Acid Sequence Binding TKK-V1V2 Thr-LYS-LYS >90%

TPAAKK-V1V2 Thr-Pro-Ala-Ala-LYS-LYS >90%

TPAAAKK-V1V2 Thr-Pro-Ala-Ala-Ala-LYS-LYS >90%

TPAAAKK-V1V2 LYS-LYS ND (1)

KA-V1V2 LYS-Ala--- >90%

Bgal*KK-V1V2 Ala-Ala-Val-GLu-Asp-Ala-Ala-LYS-LYS >90% (1) Not Determined.

The above description and examples fully disclose the invention including preferred embodiments thereof. However, it is appreciated that the invention is not limited to the particular embodiments described above. Modifications of the methods described above that are obvious to those of ordinary skill in the art are intended to be within the scope of the following claims.

Claims

Claims

1. A nucleic acid sequence comprising a nucleic acid sequence coding for the signal sequence of the Streptomyces lσnqisporus tyrosine inhibitor gene operatively linked with a propeptide sequence consisting essentially of an oligonucleotide coding for from one to about 6 amino acids, the sequence of said amino acids selected to result in the formation in Streptomyces of a heterologous protein having a homogeneous amino terminus after processing to remove the signal peptide formed on said heterologous protein during synthesis of said heterologous protein.

2. The nucleic acid sequence of claim 1 wherein said propeptide sequence causes processing of said heterologous protein at a position between the end of the nucleic acid sequence coding for said signal sequence and the beginning of said propeptide sequence.

3. The nucleic acid sequence of claim 2 wherein said propeptide codes for the amino acid threonine.

4. The nucleic acid sequence of claim 2 wherein said propeptide codes for the amino acid sequence thr-proala-ala (SEQ ID NO:1).

5. The nucleic acid sequence of claim 2 wherein said propeptide codes for the amino acid sequence thr-proala-ala-ala (SEQ ID NO: 2)

6. The nucleic acid sequence of claim 1 wherein said nucleic acid sequence coding for the signal sequence of the Streptomyces longisporus tyrosine inhibitor gene comprises the sequence

TGC GGA AGG ATG CAC ACA ATG CGG AAC ACC GCG CGC TGG GCA GCC ACC CTC GCC CTC ACG

GCC ACC GCC GTC TGC GGA CCC CTC ACC GGA GCC GCG CTC GCC, or derivative thereof capable of acting as a signal peptide in Streptomyces.

7. The nucleic acid sequence of claim 6 wherein said propeptide sequence comprises the sequence ACC.

8. The nucleic acid sequence of claim 6 wherein said propeptide sequence comprises the sequence ACC CCG GCC GCT (SEQ ID NO: 3).

9. The nucleic acid sequence of claim 6 wherein said propeptide sequence comprises the sequence ACC CCG GCC GCT GCT (SEQ ID NO:4).

10. A nucleic acid sequence comprising a nucleic acid sequence coding for the signal sequence of the Streptomyces longjsporus tyrosine inhibitor gene operatively linked with a nucleic acid sequence coding for a polypeptide modified at the 3'end to code for lys-ala-.

11. A DNA vector for expressing heterologous proteins in Streptomyces which comprises a coding seguence for the heterologous protein operatively linked to a promoter and a nucleic acid sequence of claim 1.

12. The DNA vector of claim 11 wherein said coding sequence for the heterologous protein is a sequence coding for a HIV gp120 binding region.

13. The DNA vector of claim 11 wherein said coding seguence for the heterologous protein is a sequence coding for a HIV gp120 binding region joined to a portion of a human immunoglobulin constant region encoding a polypeptide lacking most or all of the CH3 domain.

14. A DNA vector for expressing heterologous proteins in Streptomyces which comprises a coding sequence for the heterologous protein operatively linked to a promoter and the Streptomyces longisporus trypsin inhibitor gene signal seguence, wherein said seguence coding for the heterologous protein is modified at its 3' end by adding bases coding for the amino acid sequence lys-ala, or by deleting bases coding for the two amino acids at the 3' end and substituting for the deleted sequence a sequence coding for the the amino acid sequence lys-ala.

15. The DNA vector of claim 14 wherein said coding sequence for the heterologous protein is a sequence coding for a HIV gp120 binding region and wherein the bases coding for the two amino acids at the 3' end have been deleted and a sequence coding for the amino acids lys-ala have been substituted for the deleted deguence.

16. The DNA vector of claim 14 wherein said coding sequence for the heterologous protein is a sequence coding for a HIV gp120 binding region joined to a portion of a human immunoglobulin constant region encoding a polypeptide lacking most or all of the CH3 domain and wherein the bases coding for the two amino acids at the 3' end have been deleted and a sequence coding for the amino acids lys-ala have been substituted for the deleted sequence.

17. A method for using the DNA vector of claim 12 comprising the steps of introducing the vector of claim 12 into a Streptomyces host cell, and growing said host cell in a suitable culture medium.

18. A method for using the DNA vector of claim 14 comprising the steps of introducing the vector of claim 14 into a Streptomyces host cell, and growing said host cell in a suitable culture medium.

19. A Streptomyces cell transfected with the nucleic acid sequence of claim 1.

20. A Streptomyces cell transfected with the nucleic acid seguence of claim 6.

21. A Streptomyces cell transfected with the DNA vector of claim 7.

22. A Streptomyces cell transfected with the DNA vector of claim 10.