EP0239631A1 - Recombinant apolipoproteins and methods - Google Patents

Abstract

A process for the production of a purified lipid-binding peptide which can bind to phospholipids at one or more amphipatic alpha-helical peptide regions. The method includes establishing genetic coding for the peptide and introducing the gene in expressible heterologous form into an appropriate expression system capable of synthesizing a mixture of peptides which includes the lipid-binding peptide. The addition of endogenous or exogenous lipids to the mixture of peptides forms a low density lipopeptide complex composed of the lipid and the lipid linked peptide, and this complex can be easily separated from the non lipid linked peptides in the peptide according to its size. and / or its density. The process is intended in particular for the large-scale production of purified human apolipoproteins and their alpha-helical lipid-binding regions. Related processes are also described for the production of recombinant apolipoproteins, therapeutic lipopeptide compositions, and a stabilized lipid emulsion for nutritional therapy. Also disclosed are methods of expressing apolipoproteins or their lipid-binding segments in mammalian, bacterial, and yeast cell expression systems, as well as methods of purifying lipid-binding proteins, including recombinant fused proteins. .

Description

RECOMBINANT APOLIPOPROTEINS AND METHODS

1. Field of the Invention

The present invention relates to lipid-binding peptides, methods of producing, purifying and formulating the peptides, and novel lipopeptide compositions.

2. Background of the Invention

Lipid-binding proteins and their peptide fragments are potentially important in a number of therapeutic applications. One major class of lipoproteins are the serum lipoproteins which are present in a variety of macromolecular protein/lipid complexes which transport and regulate serum levels of cholesterol, triglycerides, phospholipids, and other serum lipids. Since many cardiovascular diseases are related to lipid imbalances, it has been proposed to treat such diseases by administering one or more apolipoproteins or apolipoprotein/lipid complexes which can exert a counter effect on an undesired pattern of lipid accumulation in the tissues of the patient.

Apolipoproteins share a common design with respect to the arrangement of lipid-binding regions. In general terms, the proteins have a segment or segments which contain a series of amphipatic α-helical lipid binding regions. Some of the proteins also have a relatively polar region which may be a receptor-binding region or play some other protein-identification role.

For some of the proposed therapeutic applications, the apolipoprotein is designed to perform a protein-specific function involving the more polar region. Applications of this type are expected to require the polar regions of the complete mature apolipoproteins. In other applications -- and particularly those involving lipid-binding function only, it is likely that a number of different apolipoproteins, or the lipid-binding segments thereof, could function interchangeably.

In order to bring such therapeutic application of lipoproteins to a practical stage, it is necessary to produce the lipid-binding proteins in high-levels and usually, in relatively pure form. Although the apolipoproteins may be isolated from human serum, the purification schemes are time consuming, and the yields of some are relatively poor. In particular, difficulties arise in attempting to separate selected apolipoprotein types from one another. These difficulties are illustrated by current methods for purifying serum apolipoprotein A-4 (Weinberg, L.B, et al, J Lipid Res 26:26 (1985); and Weinberg, R.B., et al, J Biol Chem 260:4914 (1985)). Also, proteins purified by these methods can be contaminated with infectious viruses, such as hepatitis.

Recombinant DNA expression represents an alternative approach which potentially can overcome some of the limitations inherent in isolation of serum proteins. Several groups, including the inventors, have reported genomic and cDNA sequences for a number of different apolipoproteins. including apolipoprotein Al (abbreviated apoAI) (Seilhamer, J.J., et al. DNA 3:309 (1984)), apoAII (Moore, et al, Biochera Biophys Res Commun 123:1 (1984), apoAIV (Bogouski, D., et al, Proc Natl Acad Sci (USA) 81:5021 (1984)). apoCI (Knott. et al. Nucl Acids Res 12:3909 (1984). apoCII (Sharpe, C.R., et al. Nuc Acid Res 12:3917 (1984)). apoCIII (Protter. A.H., et al. DNA 3:180 (1984)), and apoE (Das, H.K., et al. J Biol Chem 260:6240 (1985)). Heretofore, however, no group has successfully achieved expression of a recombinant apolipoprotein or lipid-binding segment thereof.

Even if successful expression of apolipoproteins is achieved, it will be necessary, for practical applications, to produce the protein efficiently and at high levels on a large-scale basis. Also it will be important to provide a method for purifying large quantities of the protein by a simple and efficient method.

3. Summary of the Invention It is therefore a general object of the present invention to provide a method for producing apolipoproteins for their lipid-binding segments in purified form, in a system which is readily adaptable to large-scale production. A related object of the invention is to provide a high-producing expression system capable of producing and secreting recombinant apolipoprotein in processed, or mature, form.

Another related object of the invention is to provide a method for use in an expression system in which a heterologous lipid-binding peptide is present in a mixture of proteins produced by the system, for separating the lipid-binding peptide by simple floatation and/or size-related fractionation procedures. Yet another object of the invention is to provide a stabilized nutritional emulsion which can be formulated in large scale using methods of the invention. Producing a nascent high-density lipoprotein complex by simple recombinant DNA methods is still another object of the invention.

The invention includes a method for producing, a purified lipid-binding peptide which can bind to phospholipids at one or more amphipatic alpha-helical peptide regions. In practicing the method, a gene coding for the peptide is introduced in expressible, heterologous form in a suitable expression system capable of synthesizing a mixture of peptides which includes the lipid-binding peptide. Addition of either endogenous or exogenous lipids to the peptide mixture forms a particulate Hpopeptide complex composed of lipid and the lipid-binding peptide, and this complex can be separated easily from nonlipid-binding peptides in the peptide mixture on the basis of its size or, preferably, its density.

In one embodiment, the expression system includes Chinese hamster ovary (CHO) cells, and the gene is introduced in an expression vector having a regulatable promoter derived from the human metallothionein II gene. Synthesis of the lipid-binding protein in the system is under the control of the promoter, which is responsive to the concentration of divalent metals in the cell medium. The cells are effective to produce the apolipoprotein in mature, secreted form, and can also form, from endogenous lipid, the lipoprotein complexes which can be readily separated from nonlipid-binding proteins according to the method of the invention. High-producer CHO cells can be selected by clonal selection. The lipid which is added to the mixture of proteins produced in the expression system may be endogenous lipid. as in CHO cells, or exogenous lipid in the form of emulsion particles or phospholipid vesicles. Interaction of recombinant apolipoproteins or their lipid-binding segments with oil emulsion particles produces an emulsion which is stabilized against particle fusion, or creaming, in certain patients where inflammation is involved, and thus solves an important problem in nutritional lipid-emulsion therapy. Addition of phospholipid vesicles to the apolipoproteins, and particularly apoAI, leads to the formation of disc-like lipoprotein structures which resemble high-density lipoproteins, and which are thus expected to have therapeutic properties in reverse cholesterol transport. These and other objects and features of the present invention will become more fully apparent when the following detailed description of the invention is read in conjunction with the accompanying drawings.

Brief Description of the Drawings

Figure 1 is a flow diagram showing peptide synthesis, separation, purification, and formulation processes involved in the invention; Figure 2 shows the construction of two bacterial vectors, designated ρAp85 and pApoF, designed for expression of human apoAI C-terminal genomic fragment in E. coli;

Figure 3 is an autoradiographic analysis for the expression of cloned human apoAI C-terminal genoraic fragment in E. coli;

Figure 4 shows the construction of a bacterial vector, designated pFLAl-10. designed for expression of mature human apoAI in E. coli; Figure 5 is an autoradiographic analysis of the expression of cloned human apoAI in E. coli;

Figure 6 shows the construction of a yeast vector, designated YEp-a-aA-11, designed for expression of mature human apoAI in S. cerevisiae:

Figure 7 is an autoradiographic analysis of the expression of cloned human apoAI in S. cerevisiae;

Figure 8 shows the construction of a mammalian-cell vector, designated pMTAIR, designed for expression of mature human apoAI in Chinese hamster ovary (CHO) cells;

Figure 9 is an autoradiographic analysis of the expression of cloned human apoAI in CHO cells;

Figure 10 shows gel electrophoretic patterns of proteins expressed by CHO cells transformed with control and pMTAIR vectors, and the proteins, including apoAI, which float with lipid emulsion particles.

Figure 11 is an autoradiographic analysis of the expression of cloned human apoAII in CHO cells; Figure 12 shows the construction of a mammalian-cell vector, designed for expression of cloned human apoAI C-terminal genomic fragment in CHO cells;

Figure 13 is a negative-stain electron micrograph of lipoprotein complexes formed in accordance with the invention;

Figure 14 is an ultraviolet scan of the elution profile from an HPLC column used to purify recombinant mature human apoAI;

Figure 15 shows the gel electrophoretic pattern of mature human apoAI produced by roller bottle culture and purified in accordance with the invention;

Figure 16 is a negative-stain electron micrograph of phospholipid vesicles before lipoprotein complex formation by addition of recombinant apoAI; Figure 17 is a negative-stain electron micrograph of lipoprotein complexes formed by addition of recombinant apoAI to the lipid vesicles of Figure 16;

Figure 18 shows the DNA sequence of the cDNA insert containing the signal sequence and amino acid codons of mature apoB and the corresponding amino acid sequence; and

Figure 19 illustrates vector constructs used in isolating apoB cDNAs.

Detailed Description of the Invention

Figure 1 illustrates the general schema of the invention. The upper portion of the flow diagram indicates the steps of (1) selecting of a suitable lipid-binding peptide, (2) constructing or isolating a gene coding for the peptide, and (3) placing the gene in an expression vector. As indicated in the figure, the expression vector may be one designed for use in either a bacterial, yeast, or mammalian-cell expression system. These aspects of the invention are discussed in Section I below. In Section II, methods of expressing the selected LPB in a variety of hosts are described.

The recombinantly produced LBP can be combined with lipids to produce lipid-protein complexes which may take the form of protein-associated triglyceride emulsions (LBP/emulsions) or LBP/phospholipid vesicle/(LBP/PL). Lipid emulsions or phospholipid vesicles may be added exogenously to the LBP, to form the lipoproteins. It has also been discovered, according to one aspect of the invention, that a mammalian-cell expression system, such as Chinese hamster ovary (CHO) cells, can supply endogenous lipids which combine with the expressed LBP to form the desired lipoprotein. These aspects of the invention are considered in Section III. According to another aspect of the invention, the lipoprotein complexes can be separated readily from nonlipid binding proteins on the basis of increased size or density, to yield lipoproteins which are largely purified from nonlipid binding proteins also produced in the expression system. The lipoprotein separation provides a simple, rapid method for purifying lipid-binding peptides such as apolipoproteins in large scale. This approach will be detailed in Section IV.

The purified lipoprotein formed by LBP in association with lipid emulsion may be suitable, without further purification, as a stabilized lipid emulsion for nutritional purposes. Alternatively, the LBP in the emulsion may be further purified and recombined with emulsion lipids to form an emulsion which is generally suitable for parenteral use. The increased stability of the lipoprotein. emulsion is discussed in Section V. Similarly, the lipoprotein formed by association with phospholipids is useful directly as a nascent high-density lipoprotein (HDL), or the protein may be further purified and reassociated with phospholipid to form a nascent HDL composition. The utility of nascent HDL in reverse cholesterol uptake is also considered in Section V.

I. LBP Expression Vectors

As defined herein, lipid-binding peptide (LBP) refers to a protein or protein segment which has, or in the presence of phospholipids assumes, an amphipatic alpha-helical structure characterized by a polar side or face that can interact with the polar head groups of phospholipids, and a nonpolar side which can interact with lipid acyl chain groups. The polar face contains charged amino acid residues, typically with acidic groups toward the center of the helix paired with basic residues at the edge, and it has been postulated that these charge pairs are able to form favorable ionic interactions with zwitterionic phospholipid head groups. The nonpolar face contains largely hydrophobic residues. Compositional and sequence features of alpha-helical lipid-binding peptides are reviewed in Segrest. J.P., et al. FEBS LETT 38:247 (1974).. Human apolipoproteins are one general source of lipid-binding peptides. Apolipoproteins refer to the delipidated proteins which, when combined with lipids, such as triglycerides, cholesterol, cholesterol esters, and phospholipids (Breslow. J.L., Ann Rev Biochem 54:699 (1984)). form various types of lipoprotein complexes. Alpha-helical, amphipatic lipid-binding regions are common to all human apolipoproteins, including apoAI, apoAII, apoAIV, apoB, apoCI, apoCII, apoCIII, apoD, and apoE. The lipid-binding peptide may include the entire apolipoprotein. or one or more lipid-binding regions thereof. The amino acid sequence and composition of the LBP may precisely match that of an apolipoprotein or lipid-binding region(s), or it may contain one or more amino acid substitutions which do not interfere with the LBP's ability to form an alpha-helical structure necessary for lipid binding (neutral substitutions). The gene coding for selected LBP peptide is defined herein to include any natural or synthetic DNA sequence which codes for the selected LBP, and can be expressed in heterologous form in a suitable expression system. The gene coding for the selected LBP may be derived either from genomic DNA or cDNA (copied from an isolated mRNA). The gene is said to be derived from genomic or cDNA if it includes a partial or full-length DNA sequence coding for: (a) a human apolipoprotein, (b) one or more alpha-helical binding regions thereof, or (c) proteins or peptides containing neutral substitutions in (a) or (b). The gene may also be a synthetic polynucleotide. In the latter case, the coding sequence in the gene is constructed to correspond to the LBP amino acid sequence, or more preferably, to match the known gene sequence of a selected apolipoprotein lipid-binding region, as reported, for example, in the references given above on the full-length cDNA and/or genomic sequence of a variety of apolipoproteins and in Example XVIII. The oligonucleotide can be synthesized, typically as a series of overlapping segments, by known methods. A partial or full-length genomic DNA or cDNA for a selected apolipoprotein can be obtained according to published methods or by known cloning and selection procedures. The procedures described by the inventors (Seilhamer. J.J, et al, DNA 3:309 (1984)) for obtaining both full-length cDNA and a genomic gene fragment which includes the entire coding region of apoAI are generally applicable. In the published method, a poly A RNA (mRNA) fraction was copied, made double stranded, methylated, and cloned into a λ cloning vector by addition of convenient EcoRI (RI) linkers to the double-strand cDNA ends.

The cDNA inserts were subcloned into a bacterial plasmid, and the resulting plasmid library screened with a 15-mer representing the anti-codons for the last five amino acids of apoAI. One of the clones, containing a 300 bp insert encompassing the 3'-terminal end of apoAI mRNA was selected, and from this an Alul fragment was obtained for use as a probe in screening for full-length cDNA. The largest cDNA probe identified contained a 5' untranslated region 39 nucleotides long, and an open reading frame of 801 nucleotides coding for the complete 267 amino acid prepro-apoAI. Following the TGA termination codon was a 3' untranslated region of 55 nucleotides containing the polyadenylation signal AATAAA and a short polyA tail.

A complete-sequence genomic apoAI gene was obtained by screening the cloned genomic fragment with the above Alul cDNA probe. One 2.2 kb PstI fragment was identified as containing the entire apoAI gene, including 5' and 3' untranslated regions. A comparison of the genomic and cDNA sequences revealed the presence in the genomic gene of three introns which interrupt the coding sequence in (a) the 5' untranslated region, (b) an 18 amino acid signal sequence, and (c) the region coding for the mature peptide.

Example XI describes similar methods for obtaining a full-length genomic clone coding for apoAII. References noted above also give details for obtaining genomic and/or cDNA genes for apo AIV, CII, CHI, and E. Example XVIII describes similar methods for obtaining a cDNA that encodes the signal sequence and first 430 amino acids of apoB, which includes the receptor-binding domain and two amphipatic helices. This cDNA is described in detail in co-owned patent application for "Novel Lipoprotein-Based Drug-Delivery Systems", Serial No. 783,787, filed 3 October 1985.

The LBP gene from above is placed into a suitable expression vector for expression in a bacterial, yeast, or mammalian-cell system. Typically, the vector used in the construction contains the necessary control elements for gene transcription and translation in a selected host system, and one or more restrictions sites at which the heterologous gene can be inserted in expressible form.

Vectors suitable for expressing protein in bacteria, typically E. coli, and yeast, typically S. cerevisiae. are available, and methods for introducing genes in expressible forms into the vectors are known. The Methods section below outlines general procedure for inserting heterologous gene sequences into expression vectors, selecting successful recombinants, and verifying the constructions. The methods described in Example I for construction of a bacterial vector having an apoAI genomic fragment insert, and in Examples III and V, for construction of a bacterial vector and yeast vector, respectively, with a full-length apoAI cDNA insert are generally applicable. Full-length cDNA was used in the vectors designed for production of complete apolipoproteins, to avoid limitations in bacterial and yeast systems in transcribing intron-containing genomic genes. The gene in the yeast vector contained codons for the mature apoAI minus the first seven N-terminal codons.

A preferred mammalian-cell expression vector has been described in co-owned patent application for "A Superior Mammalian Expression System". Serial No. 701,296, filed 13 February 1985. The vector contains the human metallothionein-II (hMT-II) promoter and is designed to transform Chinese hamster ovary (CHO) cells. The vector/cell system is capable of high levels of expression of the heterologous gene, and gene expression is inducible, under the control of the hMT-II promoter, by addition of divalent metals, such as divalent zinc and iron. Studies performed in support of the present application and other co-owned applications indicate that genomic (intron-containing) genes, and cDNA copies thereof, encoding a variety of protein and protein segments, including apolipoproteins and apolipoprotein α-helical segments, are expressed efficiently and at high levels in the CHO system. As will be seen below, and according to an important feature of the system, the CHO cells are also capable of processing expressed apolipoproteins to mature proteins, and of forming nascent lipoprotein complexes with endogenous producing lipid. The vector construction methods detailed below for construction of a hMT-II vector containing a full-length genomic apoAI gene (Example VII) and a full-length genomic apoAII gene (Example XII), are generally applicable to other apoproteins and lipid-binding segments thereof. An exemplary CHO expression vector containing the hMT-II promoter and a full-length genomic apoAI coding region under the control of the promoter in a high-producing CHO cell line was deposited with the American Type Culture Collection ATCC Patent Depository on 3 October 1985 and has deposit # CRL 8911. The vector in the deposited CHO cells can be readily manipulated, by known techniques, to substitute other apolipoprotein or protein fragment coding regions for the apoAI region. In another embodiment of the invention, a gene coding for an LBP is joined to the gene of a second protein or peptide in the expression vector, to produce a fused-protein gene composed of the LBP gene and the coding sequence for a second, typically hydrophilic, protein. The LBP moiety in the expressed fused protein can be used to facilitate protein purification, particularly in a scaled-up production method, according to procedures detailed in Section IV below. Methods for fusing an LBP gene to a second coding sequence would follow known cloning procedures. Preferably, the LBP gene would be attached at the C-terminus of the second-protein's coding region. If neither peptide moiety contains a methionine residue, the two genes can be fused at a methionine codon, to allow the fused protein to be cleaved into its unfused components, after purification, by cyanogen bromide cleavage, according to known methods.

II. LBP Expression

Expression of the selected LBP gene is achieved by transforming a suitable bacterial, yeast, or mammalian-cell host with the expression vector constructed as above. Methods for transforming bacterial and yeast cells with suitable expression vectors are known, and generally follow the methods outlined in Examples II and IV, for transforming E. coli. and Example VI, for transforming S. cerevisiae. Also described in these examples are typical reaction conditions under which protein synthesis in the expression system takes place.

To confirm that the expression system has indeed expressed the heterologous LBP, the proteins produced in the system are analyzed for the presence of a peptide that can be distinguished on the basis of its unique size or reactivity with anti-LBP-antibody. For proteins formed in a bacterial expression system, which are often produced as intracellular forms, the cells must be ruptured or lysed, such as by detergent treatment, and the cell debris removed, such as by centrifugation or filtration. The released cellular proteins, when fractionated by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE), show a series of fractionated proteins which migrate on the gel according to their molecular weight. Comparison of the fractionated proteins from cells transformed either with the LBP-gene vector, or a control vector may show the presence in the LBP cell proteins of a unique molecular weight species corresponding to the expected molecular weight of the LBP. The identity of the LBP can be further confirmed by precipitating the protein mixture with anti-LBP antibody, and fractionating the precipitated material by SDS-PAGE, as described generally in Example II.

Using the above protein identification methods, it is shown in Examples II and IV that a bacterial system (E. coli) is capable of producing both a apoAI fragment containing three amphipatic, alpha-helical protein regions and full-sequence apoAI itself. As noted above, the apolipoproteins have strong sequence similarities in their alpha-helical binding regions. The finding herein that the bacterial system is capable of expressing both apoAI and a lipid-binding region thereof indicates that the system is capable of expressing genes which encode apolipoprotein alpha-helical lipid binding regions generally, whether the lipid binding regions are portions of an apolipoprotein, or the full-length apolipoprotein. It is noted that the bacterial system of Examples II and IV does not carry out post-translation processing of the expressed LBP because the gene was tailored to exclude the N-terminus signal sequence and propeptide codon(s). Similar methods were used to show that a yeast system (S. cerevisiae) is capable of expressing apoAI (Example VI). In the yeast system, the LBP are secreted into the extracellular medium, and therefore can be harvested simply by removing cells, such as by centrifugation. from the cell medium.

A preferred system for expression of LBP is the mammalian-cell system composed of CHO cells transformed with the hMT-II promoter vector. For a variety of reasons, this system is well suited to scaled-up production and purification of apolipoproteins and other LBPs. One advantage of the system is its inducibility for LBP expression, by addition of divalent metals to the cell medium. Example X below demonstrates the increasing levels of apoAI production which are produced at increasingly greater zinc concentrations. Expression levels may be increased substantially in clonal strains of CHO cells which are selected for high production of LBP. Example IX illustrates the selection of clonal variants of the CHO cells which are high producers of mature apoAI. The selection procedure involves culturing the cells under low cell concentration conditions which lead to individual-cell (clonal) colonies in the culture. These colonies are picked and screened for high levels of LBP expression, for example, by dot-blot Western blotting (Jahn et al, Proc Natl Acad Sci (USA) 81:1684-1687 (1984)). In the selection procedure described in Example IX, about 200 clonal cell colonies were screened. Of these, two high-producing strains produced apoAI at a level about 30 times that of the original CHO cells as a whole. In the high-producer cells, the proportion of apoAI in the secreted protein mixture is between about 30-50%. This high percentage of specific expression contributes, in part, to the ability to obtain substantially purified LBP by a simple lipoprotein separation step (Section IV). LBP production by the CHO cells can be readily scaled up by culturing the cells in large-volume roller bottles, as described in Example XIX. The CHO strain used here was a high-producing strain, and liquid volume in each bottle was 250 ml. As seen in Table I of the example. apoAI levels of up to 30 mg/liter/day can be achieved.

According to another advantage of the CHO cell system, the apolipoproteins are processed and secreted in mature form. Cellular processing of full-length (prepro) apoAI involves cleavage of an 18 amino acid leader sequence, to form the proapoAI, and further cleavage of 6 amino N-terminal residues to form the mature protein. Although earlier studies reported in the above-cited patent application for "A Superior Mammalian Expression System" showed that the system effectively secretes proteins containing a leader sequence, with cleavage of the sequence, it was not known heretofore that the system could further process proapolipoproteins to produce mature proteins. Studies confirming the processing of apoAI to mature protein are detailed in Example XV.

It is known that both apoAI and apoAII have preproapo, proapo. and mature apo forms, and that in the body apparently at least a portion of the processing from the proapo to mature apo form occurs in the serum (Bojanovski, D., et al, J Lipid Res 26:185 (1985)). This suggests that the serum contains special peptidases capable of cleaving the small N-terminal pro peptide from proapo proteins. The finding that the CHO system is capable of processing the proapo form to mature apo form is therefore especially surprising, since the CHO cells are cultured in serum free medium. Apparently the CHO cells contain their own enzyme for carrying out this final processing step. The general procedures for expressing other apolipoproteins in the CHO system have been verified for expression and secretion of apoAII (Example XIX).

III. Forming Lipoprotein Complexes

The bacterial, yeast, and mammalian-cell expression systems described above produce a mixture of proteins (peptides) which include a relatively small proportion of the desired heterologous LBP. For bacterial systems -- where the heterologous LBP is not secreted -- the LBP-containing mixture is in the intracellular protein fraction and produced by lysing the cells, and removing cellular debris. For yeast and mammalian-cell systems, where the LBP is secreted, the protein mixture is in the extracellular protein fraction remaining after removing the cells from the reaction medium. The number of heterogeneous sized proteins contained in the protein mixtures from the various expression systems can be seen from the SDS gel electrophoretic patterns of expressed proteins in Figures 3 and 5 (for E. coli). Figure 7 (for S. cerevisiae) and Figures 9, 10, and 11 (for CHO cells). Typically the heterologously produced LBP constitutes only about 0.5 to 2 percent of the total protein mixture. In the most favorable system -- clonal high-producer CHO cells under favorable metal ion induction conditions -- approximately 30-50% of the protein is the desired LBP. According to one aspect of the invention, the heterologous LBP gene product produced in an expression system can be purified conveniently and in large volume from the protein mixture by (a) adding lipid to the mixture, to form particulate lipoprotein complexes composed of lipid particles and associated LBP, and (b) separating the complex from nonlipid binding peptides in the mixture on the basis of the lower density or greater size of the complex. The present section examines methods for forming such lipoprotein complexes in mixtures of peptides.

The lipid added to an LBP-containing protein mixture may be an oil emulsion, and typically a triglyceride oil emulsion containing a phospholipid emulsifier. Such emulsions are commercially available for use in parenteral nutrition or may be prepared by emulsifying oil/phospholipid mixtures by homogenization or sonication, according to known procedures. One standard oil emulsion is INTRALIPID, a commercially available triglyceride/phosphatidylcholine emulsion used for parenteral nutrition. The emulsion typically has particle sizes in the 0.1 to 0.5 micron range.

To form the LBP/emulsion complex, the lipid emulsion is added to the LBP-containing protein mixture, and the suspension allowed to incubate for up to one or more hours with shaking. The protein mixture is preferably concentrated to about 5-10 mg/ml before addition of lipids. Lipid is added in a weight ratio of at least about 10:1 lipid/protein and preferably at a ratio of 50:1 or more.

In a second general method, the added lipid is phospholipid, usually taking the form of an aqueous suspension of uni- or multilamellar bilayer structures commonly known as liposomes. Methods for preparing liposome suspensions from phospholipid or phospholipid/cholesterol mixtures are well known. A preferred liposome composition includes phosphatidylcholine (PC), such as egg PC, or a PC/cholesterol mixture. Typically the vesicle forming lipids are dried to a thin film, and hydrated slowly with an aqueous medium, forming a suspension of lipids. The suspension can be further processed, for example, by homogenization or sonication, to reduce the size heterogeneity of the particles. Example XVI below describes the preparation of a PC liposome suspension whose uni- and multilamellar structures are seen in the negative stain photomicrograph of Figure 16. Liposome particle sizes typically range from about 0.01 to 0.1 microns.

Addition of the lipid suspension to the protein mixture from the expression system, and incubation of the protein/lipid suspension are substantially the same as for lipid emulsion addition to the protein mixture. That is, the liposome suspension is added at a weight ratio of preferably about 10:1 or more, and incubated for up to one or more hours. The morphology of the lipoprotein. (LBP/phospholipid) complexes which form will depend on the nature of the apolipoprotein. Where the apolipoprotein is apoAI, the lipoprotein complexes include many disc-like structures, such as those shown in the photomicrograph of Figure 17. These structure strongly resemble the disc-like nascent high-density lipoprotein (HDL) particles found in liver perfusate and composed predominantly of phospholipid and apoAI and apoE.

As indicated in Figure 1, a third lipid source is endogenous lipids which are supplied by the expression system itself. Heretofore, it has not been realized that an expression system can supply lipid in an amount and form that is useful in producing particulate lipoprotein complexes with heterologous expressed lipid-binding proteins. The lipids supplied by the system, in the case of CHO cells, are presumably phospholipids which can combine with the expressed LBP to form characteristic LBP/phospholipid structures. A comparison of the photomicrographs of Figures 13 and 17 indicates that the lipoprotein particles formed in the CHO system reported in Example XIII, involving apoAI expression, resemble the disc-like structures seen when purified apoAI is mixed with PC liposomes.

IV. Lipoprotein Purification The particulate lipoprotein complex from above can be separated readily from nonlipid binding proteins in the expression system on the basis of either size or density.

As discussed in Section III. the LBP/lipid complex is a particulate complex composed of lipid particles -- either emulsion, liposomal. or lipid-disc particles -- with associated LBP. The lipoprotein particles are in the general size range of between about 0.01 to 0.55μ, and as such, can be readily separated from smaller, soluble proteins by particle exclusion in molecular sieve chromatography. The molecular-sieve material is preferably selected to exclude material in the range 500,000 daltons or greater, insuring that essentially all of noncomplexed proteins will be included (retarded) by passage through the chromatography column bed. The larger lipoprotein complex, which is excluded, will elute in the void volume of the column. The eluted fractions can be monitored, conventionally, by UV spectroscopy or the like. Preferred molecular-sieve material includes Agarose 10%, whose exclusion size is about 1,000,000 daltons.

For large-scale separation, the lipoprotein complex is preferably separated from nonlipid binding peptides by flotation. Typically this is done by adjusting the density of the peptide-mixture medium to between about 1.06-1.21 by addition of a mono- or disaccharide, such as sucrose, or a salt, such as KCl. The medium is then placed in a centrifuge tube, overlayered with a lower density medium, and centrifuged until the complex has concentrated at the top of the tube. The lipoprotein material removed from the gradient may be resuspended in the same density medium and separated a second time, to further remove nonlipid-binding proteins. This procedure is illustrated in Example XIV for the purification of apoAI. As seen in Figure 15, and discussed in Example XIV, the purification method produced about a 95% purification of apoAI, as judged by a gel density scan of the fractionated protein.

Alternatively, the lipoprotein complex will separate on its own by floatation if allowed to stand undisturbed in a medium of sufficient density. This floatation procedure is, of course, advantageous in large-scale preparation, since both chromatography and centrifugation steps are eliminated.

The LBP associated with the separated lipoprotein complex may be further purified, as necessary, by delipidating the complex and purifying the peptide by conventional purification methods, such as described in Example XV.

The purification method just described is part of a general LBP-producing system in which recombinantly produced LBP is associated to particulate lipid and isolated from nonbinding lipids by simple floatation or molecular sizing procedures. The method is more broadly applicable to purifying other proteins which have lipid-binding regions that promote protein association with particulate lipids. Such proteins include lung surfactant protein, C-reactive protein, and fused LBP proteins of the type described above. In the more general method, particulate lipid is added to a mixture of proteins produced in the expression system, to produce lipid-particle complexes with the heterologous lipophilic protein. The particulate lipoprotein complexes are then separated on the basis of size or density as described.

V. Utility of Lipoprotein Compositions

A. Nascent HDL Composition

It is known that the apoprotein components of the high-density lipoproteins (HDL) -- apoAI, apoAII, apoCI, apoCII, and apoCIII -- when reconstituted with phospholipid and then incubated in the presence of tissue culture cells, remove cholesterol from the cell membrane in vitro (Jackson, et al, J Biol Chem 250:7204 (1975)) and in vivo (Miller. N.E., et al. Nature 314:109 (1985)). The complex formed by apoAI and phospholipid appears to be particularly effective in reverse cholesterol transport.

As described in Section III, the lipoprotein complex formed by recombinant apoAI and PC liposomes resembles the disc-like structures of nascent HDL. This finding indicates that apolipoproteins formed in accordance with the invention retain their lipid-binding properties and that lipoproteins formed with the recombinant proteins would therefore have the same therapeutic uses, for reverse cholesterol transport, as HDL particles or nascent HDL particles formed with isolated serum apolipoproteins. B. Other Therapeutic Uses

It is known that either apoAI or apoCI activate lecithin cholesterol acyl transferase (LCAT) in vitro (Soutaw, A.K., et al. Biochem 14:3057 (1975)). LCAT acetylates cholesterol and this is thought to be the mechanism that prevents cholesterol exchange out of HDL. Thus, infusion of apoAI or apoCI or a mixture of both in various Hpopeptide mixtures could be effective to enhance LCAT activity and thereby lower the rate of, or prevent, plaque formation and consequent atherosclerotic disease. ApoCII is known to enhance lipoprotein lipase activity in vitro (Shelburne. F. et al. J Clin Invest 65:652 (1980)) which promotes the metabolism of triglyceride-carrying chylomicrons, lowering the serum levels of triglycerides. In fact, deficiency of apoCII has been shown to cause one form of hypertriglyceridemia (Breckenridge, W.C., et al, New Engl J Med 298:1265 (1978)) and apoCΪI therapy may be effective in controlling this disease. ApoCII and apoCIII are both known to facilitate the removal of triglycerides in vivo (Erkelens. D.W., et al, Metabolism 34:222 (1985)) and they could be important in therapy to lower serum triglyceride levels for the prevention of hypertriglyceridemia. ApoE have been associated with type III lipoproteinemia, (Zannis, V.I. et al, J Biol Chem 255:1759 (1980)). Supplying these patients with the normal protein could be a useful treatment for this disease. ApoB deficiency has been associated with a-ß-lipoproteinemia (Barson. F.A. et al, Blood 5:301 (1959)). and preparations with this apolipoprotein could be useful in treating these patients. C. Stabilized Emulsion

Although fat emulsions are widely used for parenteral nutritional therapy, in many acutely ill patients and in infants, problems relating to instability of emulsion particle size have been encountered. In infants, particle size instability is evidenced by fat accumulation in the lungs, leading to fat embolism and death (Barson. et al. Archives of Disease in Childhood 53:218 (1978)). A similar problem has been observed in acutely ill adults, and it has been shown that the sera of acutely ill patients agglutinate, or cream, lipid emulsions such as INTRALIPID. The conditions most commonly associated with creaming are neoplasia; severe infections, e.g., pneumonia, septicaemia; major surgery; trauma; myocardial infarction and cerebrovascular accident; inflammatory disorders including rheumatoid arthritis; Crohn's disease and polyarteritis nodosa; and lymphoma. This finding has led to the suggestion that C-reactive protein, synthesized at high levels in people stressed by inflammation, tissue injury, or infection (this group includes premature infants), may be enhancing the fusion of INTRALIPID lipid particles to particles of a larger size that are filtered out of the bloodstream by the lungs, sometimes leading to fat embolism and death. Studies conducted in support of the present invention, and reported in Example XVII show that serum from stressed animals causes a significant size increase in a major portion of INTRALIPID particles, whereas emulsion particle size is substantially stable in serum from unstressed animals (Table IV).

According to an important discovery herein, it was found that the lipoprotein complex formed between apoAI and INTRALIPID is stable in size when exposed to serum from either stressed or unstressed animals. That is. apoAI stabilized the emulsion particles against fusion and size growth in the presence of serum factor (s) (presumably including C-reactive protein) associated with stress.

The invention thus includes a stabilized lipid emulsion for parenteral nutritional use, composed of lipid emulsion particles and associated recombinant apolipoprotein, such as apoAI. From the foregoing it can be appreciated how various objects and features of the invention are met. The invention provides a method for producing recombinant apolipoproteins, and lipid-binding peptides thereof, in a variety of expression systems, including E. coli, S. cerevisiae, and CHO cells. These recombinant apolipoproteins retain their native lipid-binding properties, as judged both by (a) their ability to bind to added lipid, either in the form of an oil/phospholipid emulsion or phospholipid vesicle suspension, (b) their ability to associate with and stabilize oil/phospholipid emulsion particles against fusion in the presence of fusogenic serum factor(s) and (c) their ability to interact with phospholipid vesicles, producing disc-like structures similar in appearance to nascent HDL particles.

The CHO cell expression system described herein in particular provides several advantages for large-scale production of apolipoproteins and their peptide fragments. The system is inducible for heterologous gene expression, and high-producer clonal strains which are selected readily can increase heterologous gene expression by more than an order of magnitude and to a level in which the recombinant LBP constitutes 30%-50% of total secreted protein. This cell system also has the ability to process full-sequence apolipoproteins to mature proteins, an ability that may require two separate peptide cleavages for apoAI and apoAII. The lipid-binding capacity of the recombinant proteins is exploited in a simple, efficient method for producing purified apolipoprotein by recombinant methods. The purity of the purified Al may be as high as 95% using only centrifugation steps. The method is especially useful for producing apolipoproteins in large scale, in combination with the high-level CHO expression system.

In broader scope the method of producing pure or nearly pure recombinant protein by lipid addition to recombinantly produced proteins is applicable to any lipid-binding recombinant protein which can associate stably with lipid particles.

Methods

A. Vector Construction

Construction of suitable vectors containing the desired coding and control sequences employs standard ligation and restriction techniques which are well understood in the art. Isolated plasmids, DNA sequences, or synthesized oligonucleotides are cleaved, tailored, and religated in the form desired.

Site specific DNA cleavage is performed by treating with the suitable restriction enzyme (or enzymes) under conditions specified by the manufacturer of these commercially available restriction enzymes.

See, e.g., New England Biolabs, Product Catalog. In general, about 1 μg of plasmid or DNA sequence is cleaved by one unit of enzyme in about 20 μl of buffer solution; in the examples herein, typically, an excess of restriction enzyme is used to insure complete digestion of the DNA substrate. Incubation times of about one hour to two hours at about 37°C are workable, although variations can be tolerated. After each incubation, protein is removed by extraction with phenol/chloroform, and may be followed by ether extraction, and the nucleic acid recovered from aqueous fractions by precipitation with ethanol. If desired, size separation of the cleaved fragments may be performed by polyacrylamide gel or agarose gel electrophoresis using standard techniques. A general description of size separations is found in Methods in Enzymology 65:499 (1980). Restriction cleaved fragments may be blunt ended by treating with the large fragment of E. coli DNA polymerase I (Klenow) in the presence of the four deoxynucleotide triphosphates (dNTPs) using incubation times of about 15 to 25 min at 20 to 25°C in 50 mM Tris pH 7.6, 50 mM NaCl, 6 mM MgCl₂, 6 mM DTT and 5-10 μM dNTPs. The Klenow fragment fills in at 5' sticky ends but chews back protruding 3' single strands, even though the four dNTPs are present. If desired, selective repair can be performed by supplying only one of the, or selected, dNTPs within the limitations dictated by the nature of the sticky ends. After treatment with Klenow, the mixture is extracted with phenol/chloroform and ethanol precipitated. Treatment under appropriate conditions with S1 nuclease or exonuclease Bal-31 results in hydrolysis of any single-stranded portion. Ligations are performed in 15-50 μl volumes under the following standard conditions and temperatures: 20 mM Tris-HCl pH 7.5, 10 mM MgCl₂, 10 mM DTT, 33 μg/ml BSA, 10 mM-50 mM NaCl, and either 40 μM ATP, 0.01-0.02 (Weiss) units T4 DNA ligase at 0°C (for "sticky end" ligation) or 1 mM ATP, 0.3-0.6 (Weiss) units T4 DNA ligase at 14°C (for "blunt end" ligation). Intermolecular "sticky end" ligations are usually performed at 33-100 μg/ml total DNA concentrations (5-100 nM total end concentration). Intermolecular blunt end ligations (usually employing a 10-30 fold molar excess of linkers) are performed at 1 μM total ends concentration. In vector construction employing "vector fragments", the vector fragment is commonly treated with bacterial alkaline phosphatase (BAP) or calf intestinal alkaline phosphatase (CIP) in order to remove the 5' phosphate and prevent religation of the vector. Digestions are conducted at pH 8 in approximately 150 mM Tris, in the presence of Na and Mg using about 1 unit of BAP or CIP per μg of vector at 60° for about one hour. In order to recover the nucleic acid fragments, the preparation is extracted with phenol/chloroform and ethanol precipitated.

Alternatively, religation can be prevented in vectors which have been double digested by additional restriction enzyme digestion of the unwanted fragments. The desired sequences are thus recovered from colonies responding to probe.

B. Verification of Construction

In the constructions set forth below, correct ligations for plasmid construction are confirmed by first transforming E. coli strain MC1061 obtained from Dr. M. Casadaban (Casadaban, M., et al, J Mol Biol 138:179 (1980)) or other suitable host with the ligation mixture. Successful transformants are selected by ampicillin, tetracycline or other antibiotic resistance or using other markers depending on the mode of plasmid construction, as is understood in the art. Plasmids from the transformants are then prepared according to the method of Clewell, D. B., et al, Proc Natl Acad Sci (USA) 62:1159 (1969). optionally following chloramphenicol amplification (Clewell, D.B., J Bacteriol 110:667 (1972)). The isolated DNA is analyzed by restriction and/or sequenced by the dideoxy method of Sanger. F., et al, Proc Natl Acad Sci (USA) 74:5463 (1977) as further described by Messing, et al, Nucleie Acids Res 9:309 (1981), or by the method of Maxam, et al. Methods in Enzymology 65:499 (1980).

Example I Bacterial Vectors with Al Genomic Fragment

The recombinant phage. λAI-12, containing a 2.2 kb PstI fragment that encodes most of the apoAI gene, including all of the coding sequence, has been described (see Seilhamer, J.J., et al, DNA 3:309 (1984)). Plasmid pPSAI.2 was constructed by isolating the 2.2 kb PstI fragment and inserting it into the PstI site of pBR322 (Maniatis, T., et al. Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory, NY (1982)). The pPSAI.2 insert was removed by PstI digestion and isolated by gel electrophoresis. The isolated insert (10μg) was digested with 5 units of StuI, and the 1065 base pair StuI fragment, which includes the part of the apoAI gene region that encodes amino acids 110-187 was isolated and digested with Narl, as shown in Figure 2. The 5' overhang resulting from Narl digestion was blunt-ended by incubating the DNA with the Klenow fragment of DNA polymerase I in K.L. Buffer (10 mM Tris-HCl, pH 8.0. 5 mM MgCl₂ ), and 0.8 mM each dATP, dTTP, dCTP, and dGTP for 5 hours at 17°C. The self-complementary oligonucleotide #2A-14 (5'-dCCTGAAGCTTCAGG) which contains translation stop signal (TGA) and Hindlll site (AAGCTT) was ligated to both ends of the blunt-ended fragment by incubating with 50 units T4 DNA ligase in K.L. Buffer with 10 mM ATP for 12 hours at 12°C. The linkers were cut back with Hindi II and the fragment was subcloned into the HindIII site of pKT19 (Talmadge. K., et al, Gene 12:235 (1980)) to give the plasmid pPSAI.4 (Figure 2). The pPSAI.4 plasmid (10μg) was cut with 2 units of MboII for 2 hours at 37°C, and the resulting 3' overhang was removed by treatment with T4 DNA polymerase (Maniatis, et al, supra). The self-complementary oligonucleotide #1A--10 (CATGGCCATG) which contains a translation initiation signal (TGA) was ligated to the MboII cut, blunt-ended DNA fragments, cut with Ncol and Hindlll and ligated into pBR329 (catalogued in the National Institutes of Health, Bethesda, MD data base of vectors and gene sequences) which had been cut with Ncol and Hindlll (see Figure 2). The ligation mixtures were used to transform E. coli HB101 cells as described above. Plasmids with the correct insert were identified by restriction mapping. The resulting plasmid, pPSAI.6, includes an Ncol/Hindlll fragment with the apoAI gene fragment encoding amino acids 110-187 in frame with an ATG translation initiation signal at the 5' end and a TGA translation stop signal at the 3' end. The Ncol/Hindlll insert of pPSAI.6 was isolated and ligated into (a) the expression vector pKK233-2 (Brosius. J., et al. Proc Natl Acad Sci (USA) 81:6929 (1984)) which had been cut with Ncol and Hindlll, and (b) the expression vector pKT52, which had been cut with the same enzymes. pKK233-2 and pKT52 were constructed as follows. The construction of these vectors is described in U.S. serial nos. 616,488 and 622,639, assigned to the same assignee, and incorporated herein by reference. With reference to Figure 2, the "trc" promoter contains the upstream portions of the trp promoter and the downstream, operator-containing, regions of the lac promoter and was originally prepared from two readily available plasmids containing these promoters. To construct the trc promoter as a BamHI/HindlII cassette, an intermediate plasmid pKK10-0 was prepared containing the hybrid promoter.

To prepare pKK10-0. pEA300 (Amman. E., et al, Gene (1983) 25 :167-178) was digested with PvuII and Clal, filled in using dCTP only in the presence of DNA polymerase (Klenow), followed by digestion with mungbean nuclease, and the large vector fragment isolated. This vector fragment contains the upstream portions of the trp promoter. The fragment was ligated with a 55 bp blunt-ended Hpall/PvuII digest excised from pGL101 (Lauer. G., et al, J Mol Appl Genet (1981) 1:139-147), which was prepared by digesting pGL101 with PvuII and Hpall followed by repair in the presence of dGTP and labeled dCTP. This fragment contains the lac operator region. The ligation product of these two blunt-end fragments was pKK10-0. A BamHI site was inserted into pKK10-0 upstream of the trp/lac (trc) promoter/operator by digestion with EcoRI, filling in with Klenow, and insertion of the BamHI linker 5'-CCGGATCCGG-3'. The resulting plasmid, pKK10-1 was digested with PvuII, and ligated to the Ncol linker, 5'-ACCATGGT-3'. digested with Ncol, filled in, and then ligated to a double-stranded linker containing PstI and Hindlll sites provided as two complementary oligonucleotides, 5'-GCTGCAGCCAAGCTTGG-3' and its complement. The ligation mixture was used to transform E. coli to Amp^R. The isolated plasmid DNA was digested with BamHI and Hindlll, and the small BamHI/HindlII fragment obtained on electrophoresis contains the trc promoter. To complete pKT52. the BamHI/HindlII fragment containing the trc promoter was ligated into the large fragment obtained from BamHI/HindlH digestion of pKK10-2 (Brosius. J., Gene (1984) 27:161-172) which contains the Amp^R gene and the origin of replication. The resulting plasmid, pKK233-1 was digested to completion with Pvul and then partially with Bgll and ligated with the 360 bp Pvul/Bgll fragment containing the corresponding portion of the ampicillin resistance gene but lacking a PstI site from pUC8. The ligation mixture was used to transform E. coli and transformants were screened for the presence of only one PstI site next to the trc promoter. The correct construction, pKK233-2. was digested with EcoRI and PvuII, filled in with dATP and dTTP, and religated to obtain the correct construction pKT52. pKT52 contains the desired trc promoter, a downstream ATG start codon. and downstream Ncol, PstI and Hindlll sites.

After ligation of the Ncol/Hindlll fragment of pPSAI.6 into the two digested vectors, sequence analysis confirmed the structure of the Figure 2 plasmid constructs, designated pAP85 (the pKK233-2 expression vector) and pApoF (the pKT52 expression vector).

Example II

Expression of Al Genomic Fragment in E. coli E. coli JA221 (Nakamura. K., et al, J Mol Appl Genet 1:289 (1982)) (lpp-/F'lacI^Q) containing plasmid pKK233-2, pAP85 or pApoF were grown in M9 medium (Miller, J., Experiments in Molecular Genetics, Cold Spring Harbor Laboratory, Cold Spring Harbor, New York, 1972) supplemented with glucose (2 mg/ml). thiamine (10 μg/ml). MgSO₄-7H₂O (250 μg/ml). CaCl₂ (11 μg/ml), tryptophan (20 μg/ml), leucine (20 μg/ml). methionine (2 μg/ml), ampicillin (100 μg/ral), and isopropyl-1-thio-D-galactopyranoside (IPTG, 1 mM).

At a cell density of approximately 5 x 10⁸ cells/ml (OD₅₉₀nm = 0.5), 10 μCi/ml of L-[³²S]-methionine (>1000 Ci/mmole) was added.

Following 30 sec of incubation, 1 ml of culture was removed, added to 0.34 ml of ice-cold 20% (w/v) trichloracetic acid (TCA) in a 1.5 ml Eppendorf centrifuge tube, vortexed, and incubated at 0°C (on ice) for 30 minutes. The mixture was then centrifuged at 4°C for 10 min in an Eppendorf centrifuge at 15,000 x g. The supernatant was discarded and the pellet was washed once with 1 ml of ice-cold acetone, followed by centrifugation for 2 min. The acetone was aspirated off and the pellet allowed to air dry.

The dried TCA pellet was resuspended in 40 μl of 50 mM Tris-HCl. pH 6.8, 1 mM EDTA, and 1% (w/v) SDS. Ten μl of this resuspension was added to 10 μl of 2 x SDS gel sample buffer (125 mM Tris-HCl, pH 6.8. 20% glycerol. 2% SDS, 2% ß-mercaptoethanol. and 0.1% bromophenol blue), and incubated at 100°C for 2 minutes.

The remaining 30 μl of the resuspension was brought to a 1 ml total volume by the. addition of 970 μl of 50 mM Tris-HCl, pH 6.8, 0.15 M NaCl, 0.1 mM EDTA. and 2% (v/v) Triton X-100. Two μl of sheep anti-h-apolipoprotein-AI-antiserum (obtained from Boehring Mannheim) was added to the suspension. The suspension was incubated at 0°C on ice overnight. Following incubation, 50 μl of Pansorbin (10% (w/v) cell suspension, Staphylococcus aureus cells obtained from Calbiochem-Behring. La Jolla, CA) was added to the suspension, gently mixed, and incubation continued at 0°C for 30 min. The suspension was centrifuged in the Eppendorf centrifuge for 2 min at 4°C. The supernatant was aspirated and the cell pellet resuspended in 0.5 ml of 50 mM Tris-HCl, pH 7.5. 5 mM EDTA, 0.5 M NaCl, 0.5% (v/v) NP-40, and 1 mg/ml ovalbumin. The pellet was washed with vigorous vortexing, followed by centrifugation and aspiration of the supernatant. This washing procedure was repeated three additional times. Two further washes followed with 0.5 ml of 50 mM Tris-HCl, pH 7.5, 5 mM EDTA, 0.15 M NaCl, and 0.5% (v/v) NP-40. The final cell pellet was resuspended in 50 μl of SDS gel sample buffer (75 mM Tris-HCl, pH 6.8, 10% (w/v) glycerol, 1% (w/v) SDS, 1% (w/v) ß-mercaptoethanol. 0.05% (w/v) bromophenol blue), followed by incubation at 100°C for 5 min. Cell debris was removed by centrifugation in an Eppendorf centrifuge for 5 min at room temperature.

The total and immunoadsorbed samples were subjected to SDS-polyacrylamide gel electrophoresis according to the method of Laemmli (Laemmli, U.K., Nature (London) 227:680 (1970)) in the presence of 6 M urea. Sixteen cm long slab, gels (1.5 mm thick) of 16% acrylamide were electrophoresed at 30 mArap constant current until the bromophenol blue dye reached the bottom of the gel. Gels were fixed for 30 min in 30%

(v/v) methanol, 10% (w/v) TCA, and 10% (v/v) acetic acid and treated with En³Hance obtained from New England

Nuclear (Boston, Massachusetts) according to the manufacturer's directions, followed by drying and fluorography at -70°C using Kodak XAR-5 X-ray film. A comparison of polypeptides extracted from cells containing plasmids pKK233-2, pAP85, and pApoF, labeled with L-[³⁵S]-methionine as shown in Figure 3.

A unique polypeptide migrating at a molecular weight of approximately 4500 daltons appears in lanes 2 and 3 which are protein patterns from cells containing plasmids pApoF and pAP85. respectively. Cells containing plasmid pKK233-2 (lane 1) do not have this unique peptide. The estimated molecular weight of the apoAI fragment peptide based on amino acid composition is 8,700 daltons. The discrepancy between measured and expected molecular weight may be due either to the lipophilic nature of the protein, which may effect its binding to SDS, or to a processing event which is taking place in the cell. This unique polypeptide is specifically iramunoreactive to serum raised against apoAI (Figure 3, lanes 5 and 6). ApoAI antiserum does not react with any unique polypeptides extracted from cells containing plasmid pKK233-2 (lane 4).

Example III Bacterial Vector with Al cDNA Plasmid pBL13AI containing a full-length Al cDNA was prepared as described in Seilhamer, J.J., et al, supra. The plasmid (100 μg) was digested to completion with EcoRI for 3 hr at 37°C. The 965 base pair EcoRI fragment containing the apoAI cDNA was isolated on a 4% (w/v) nondenaturing polyacrylamide gel (Maniatis, et al, supra). The apoAI cDNA fragment was excised from the gel, electroeluted, and concentrated by ethanol precipitation. The resulting DNA pellet was dried in vacu and resuspended in H₂O. This EcoRI fragment was further digested with Sau3A for 30 min at 37°C with a DNA-to-enzyme ratio of 1 μg : 2 units. The resulting Sau3A fragments were separated as described above. The 783 base pair partial fragment was isolated and concentrated as described above.

Two complementary oligonucleotides (Figure 4) containing the coding sequence of the first 8 amino acids of mature apoAI protein were synthesized on a Biosearch SAM I DNA Synthesizer (Biosearch, San Rafael,

CA) according to the manufacturer's instructions. The oligonucleotides were kinased with ³²P-ATP using T4 polynucleotide kinase (Maniatis, et al, supra). The two oligonucleotides (1 ng each) were mixed, boiled for 2 min, and allowed to hybridize at RT for 60 min. The hybridized oligonucleotides were ligated to 25 ng of Sau3A partial apoAI cDNA fragment with the addition of T4 DNA ligase by the method of Maniatis. et al, supra. After ligation. the mixture was digested to completion with EcoRI and Ncol. Products were separated on a 5% nondenaturing polyacrylamide gel. The 811 base pair fragment corresponding to full-length coding sequence of mature apoAI protein was excised. The DNA was eluted from the gel slice in 10 mM Tris-HCl, pH 8, 1 mM EDTA, 0.4 M NaCl with agitation at 37°C overnight and concentrated by ethanol precipitation. The NcoI-EcoRI fragment was then ligated into pBR329 vector. The resulting plasmid. shown in Figure 4, is designated pFLAI-2.

In order to confirm the correct DNA sequence through the constructed oligonucleotide site, an Rsal fragment of pFLAI-2 encompassing the first 56 base pairs of the coding region for mature apoAI, as well as 112 base pairs downstream into pBR329 was sequenced by the dideoxy method (Messing, J., et a,. Gene 19:259 (1982) and Sanger. et al. Proc Natl Acad Sci (USA) 74:5463 (1977)). The NcoI-EcoRI construct was the ligated into the bacterial expression vector pKK233-2 of Figure 2. Plasmid pFLAI-2 (10 μg) was digested to completion with EcoRI and the ends filled in the method of Maniatis et al, supra, with E. coli DNA Polymerase I, Klenow fragment, and the addition of 0.5 mM dATP and dTTP. The blunted DNA was then phenol/chloroform extracted, and concentrated by ethanol precipitation. The blunted DNA was further digested to competition with Ncol. The 811 base pair fragment containing the coding region of mature apoAI was purified on a 5% nondenaturing acrylamide gel and concentrated as described above. The Ncol blunt fragment was then ligated using T4 DNA ligase (Maniatis et al, supra) into pKK233-2 which was cut with Hindlll, blunted, phenol extracted, and cut with Ncol. The resulting plasmid is designated pFLAI-10 (Figure 4).

Example IV Expression of Full-Lenqth ApoAI in E. coli E. coli JA221 (lpp-/F'lacI^Q) containing plasmids pKK233-2 or pFLAI-10 (Example III) were grown and labelled in M9 medium (Miller. J., Experiments in Molecular Genetics, Cold Spring Harbor Laboratory, Cold

Spring Harbor, New York, 1972), as in Example II. Following 30 sec of incubation with [³⁵S]-methionine,

1 ml of culture was removed and washed TCA precipitable material was obtained, as in Example II.

The pellet was resuspended in 50 μl of 5 mM

Tris-HCl, pH 6.8, 1 mM EDTA, and 1% (w/v) SDS, and incubated at 100°C for 3 min. An aliquot equal to 2 x 10⁵ cpm was added to the SDS gel sample buffer of

Example II, and incubated 100°C for 2 min. A second aliquot equal to 2 x 10⁶ cpm was brought up to 1 ml total volume by the addition of 5 mM Tris-HCl, pH 6.8, 0.15 M NaCl. 0.1 mM EDTA, and 2% (v/v) Triton X-100. 2 μl of rabbit antiserum raised against human apoAI was added to the suspension. The suspension was incubated at room temperature for 30 min and further incubated on ice overnight. Following incubation, 50 μl of a 10% (w/v) Protein A Sepharose CL-4B (obtained from Sigma, St. Louis. MO) suspension in 50 mM Tris-HCl, pH 7.5, 5 mM EDTA. 0.15 M NaCl, 0.5% (v/v) NP-40. and 1 mg/ml ovalbumin was added. Incubation continued with gentle agitation for 60 min at 4°C. The suspension was centrifuged 15,000 x g for 3 min at 4°C. The supernatant was aspirated, and the Sepharose beads were washed 5 times with 0.5 ml 50 mM Tris-HCl, pH 7.5, 5 mM EDTA. 0.5 M NaCl. 0.5% (v/v) NP-40, 1 mg/ml ovalbumin. 2 times with 50 mM Tris HCl. pH 7.5, 5 mM EDTA, 0.15 M

NaCl. 0.5% (v/v) NP-40. and once with 10 mM Tris-HCl. pH 7.5. The pellet was then resuspended in 30 μl of SDS gel sample buffer and incubated at 100°C for 5 min. The total and immunoadsorbed samples were subjected to SDS-polyacrylamide gel electrophoresis according to the method of Laemmli (Laemmli, U.K., supra). Sixteen cm slab gels (0.7 mm thick), 10-20% acrylamide gradients, were electrophoresed at 15 mAmps until the BPB dye reached the bottom of the gel. Gels were fixed in 25% (v/v) methanol, 10% (w/v) TCA and 10% (v/v) acetic acid for 30 min. Gel autoradiography was performed as in Example II.

A comparison of polypeptides extracted from cells containing plasmids pKK233-2 and pFLAI-10 is shown in Figure 5. A unique polypeptide migrating at a molecular weight of 25,000 daltons appears in cells harboring the plasmid pFLAI-10 (lane 4). This unique protein is not present in cells harboring plasmid PKK233-2 (lane 2). Commercially prepared mature apoAI protein (Calbiochem-Behring. La Jolla. CA) also migrates at 25.000 daltons, identically to the bacterial product. It should be noted that the calculated molecular weight of mature apoAI protein is 27,000 daltons. but the hydrophobic nature of this protein causes it to migrate at an apparent molecular weight of 25,000 daltons in this gel system. This unique polypeptide is specifi- cally immunoreactive to serum raised against apoAI (lane 5). ApoAI antiserum does not react with any unique proteins from cells harboring plasmid pKK233-2 (lane 3).

Example V Yeast Vector with ApoAI cDNA A yeast library in the E. coli-yeast shuttle vector YEp13 (Nasmyth. K., et al, Cell 19:753 (1980)) was screened using a 5'-³²P end-labeled oligodeoxynucleotide 5'-dCCTGGCCAACCAATG-3' (Maniatis et al, at pp. 324-325). Plasmids containing inserts of yeast DNA hybridizing to this oligonucleotide were subsequently isolated. One of these plasmids contained an insert of approximately 15 kb of yeast DNA, and was shown to contain the 1.7 kb EcoRI fragment containing the alpha-factor gene as described by Kurjan and Herskowitz (Kurjan. J., et al. Cell 30:933 (1982)). The end of the 1.7 kb EcoRI fragment were made blunt by incubation with DNA Polymerase I (Klenow fragment) and BamHI linkers using T4-DNA ligase (Maniatis, et al, at pp. 113-114, 116, 392-394). The BamHI ends were made cohesive by digestions with Bam HI restriction endonuclease, subsequently ligated into the BamHI site of the yeast-E. coli shuttle plasmid pCV7-Hin2 (Broach, J.R., et al. Cell 21:503 (1980)). A deletion around the Hindlll site of the plasmid CV7 was made by Hindlll digestion. treatment with exonuclease III. treatment with SI nuclease, and religations with T4-DNA ligase to generate the plasmid pCV7-Hin2, all using the method described in Maniatis et al, supra. This plasmid containing the yeast alpha-factor gene is diagramed in Figure 6, and designated YEp-α-8.

A fragment of DNA from pBL13AI (Example III) encoding the entire mature apoAI sequence minus the first eight amino acids was inserted into the unique Hindlll cloning site of YEp-α-8 (Figure 6). With reference to the figure, digestion of pBL13AI (10 μg) with EcoRI endonuclease followed by partial digestion with Sau3A endonuclease produced a 0.78 kb Sau3A-EcoRI fragment the ends of which were made blunt with DNA Polymerase I (Klenow fragment) and synthetic Hindlll linkers (pCCAAGCTTGG) using T4 DNA ligase (Maniatis, et al. supra, pp. 113-114, 392-394). The Hindlll ends were made cohesive by digestion with Hindlll endonuclease and ligated into Hindlll-cleaved YEp-α-8 which had been treated with alkaline phosphatase to remove its 5' phosphate moieties (Maniatis, et al, supra, pp. 366-369). With insertion into the yeast vector, amino acid number eight of mature apoAI is restored by the Hindlll linker, allowing for coding of amino acids 8-243 of mature apoAI.

Expression vector YEp-α-aA-11 (Figure 6) with the apoAI insert in the correct orientation encodes a chimeric protein consisting of factor signal/leader peptide and almost the entire mature apoAI sequence (amino acids 8-243). Expression vector YEp-α-aA-5 with the insert in the reverse orientation encodes a chimeric protein with the same signal/leader peptide and a protein unrelated to apoAI. Example VI Expression of ApoAI Full-Length cDNA in S. cerevisiae DNA was prepared from E. coli cultures containing the plasmids YEp-α-8. YEp-α-aA-11, and YEp-α-aA-5 (Example V) and was used to transform yeast strain W301-18A (α ade 2-1. trp 1-1, leu 2-3, -112, can 1-100, ura 3-1, his 3-11, -15) (Kramer, et al, Proc Natl Acad Sci (USA) 81:367 (1984)) to Leu 2 prototrophy. Yeast strains were grown on standard media (Sherman et al. Methods in Yeast Genetics. Cold Spring Harbor Press. Cold Spring Harbor, New York). Plasmid DNA from E. coli was also recloned into M13 for sequencing and confirmation of the α-factor apoAI DNA constructions (Messing. J., et al, Gene 19:259 (1982) and Sanger, et al, Proc Natl Acad Sci (USA) 74:5463 (1977)).

Yeast cultures Y9-1 (YEp-α-8/W301), Y9-2 (YEp-α-aA-11/W301), and Y9-3 (YEρ-α-aA-5/W301) were maintained in synthetic medium lacking leucine. Saturated overnights were diluted back in fresh media with the addition of 100 μg/ml BSA (bovine serum albumin) to OD600 nm = 1.5. Cultures were shaken at

30°C for 30 min at which time 200 μCi/ml of L-[³⁵S]-methionine (>1000 Ci/mmole, Amersham Corp., Chicago, ILL) was added. Growth was allowed to continue for 4 hr, at which time PMSF (phenylmethysulfonylfluoride) was added to 1 mM. Cultures were centrifuged at 5,000 rpm for 5 min, and the cell pellet discarded. Medium supernatant proteins were concentrated by the addition of ice cold TCA to a final concentration of 10%. The supernatant was incubated on ice for 30 min, centrifuged for 15 min at 4°C in an Eppendorf centrifuge at 15,000 x g, and the pellet was washed with 1 ml of ice cold acetone. The final is dried in vacu. resuspended in 50 mM Tris-HCl, pH 6.8. 1 mM EDTA, 1% (w/v) SDS, and incubated at 100°C for 3 min. 10 μl of this resuspension was added to 10 μl of 2x SDS gel sample buffer. The remaining 40 μl was brought up to 1 ml by the addition of 960 μl of 50 mM Tris-HCl, pH 6.8, 1 mM EDTA. and 2% (v/v) Triton X-100. Two μl of rabbit antiserum raised against human apoAI was added to the suspension. The suspension was then incubated, reacted with Protein A Sepharose. and washed as described in Example IV. The total and immunoadsorbed samples were subjected to SDS-polyacrylamide gel electrophoresis as described in Example IV.

A comparison of polypeptides extracted from the media of S. cerevisiae cultures carrying the expression constructs is shown in Figure 7. A unique polypeptide migrating at a molecular weight of approximately 23,000 daltons appears in cultures of yeast carrying the alpha factor expression vector with the apoAI cDNA insert in the correct orientation (Y9-2, lane 3). This unique protein is not present in yeast carrying the α factor expression vector only (Y9-1) or the apoAI cDNA insert in the wrong orientation (Y9-3) (lanes 1 and 5, respectively). This unique protein is specifically immunoreactive with rabbit antiserum raised against human apoAI (lane 4). Proteins from the other yeast cultures Y9-1 and Y9-3 have no specific reaction with apoAI antiserum (lanes 2 and 6. respectively). The recombinant α factor-apoAI cDNA expression vector (YEp-α-aA-11) encodes for an apoAI protein 7 amino acids shorter from the N-terminus than mature apoAI (only amino acids 8-243 are encoded). This accounts for the fact that the apoAI product synthesized in yeast cultures migrates ahead of mature apoAI produced by bacterial expression (Example IV and Figure 5) and by mammalian-cell expression (Example VIII and Figure 9 below).

Example VII

Construction of Mammalian-Cell Vector for Al Expression To facilitate the expression of apoAI in mammalian cells, a hybrid gene was constructed in which the coding segment for apoAI was fused to a powerful regulated promoter derived from the human metallothionein II (hMT-II) gene. This was performed in two steps, which are illustrated in Figure 8. First, an expression vector was prepared. The expression vector, pHSI, carried 840 nucleotide base pairs of hMT-II sequence (Karin, M., et al. Nature 299:797 (1982)) from a naturally occurring Hindlll restriction site at base -765 from the start of transcription to base +70, located in the 5' untranslated region adjacent to the coding region. pHSI also carries a region into which coding sequences may be inserted . To construct pHSI, the plasmid p84H (Karin, et al, supra), which carries the hMT-II gene, was digested to completion with restriction endonuclease BamHI followed by treatment with exonuclease Bal-31 to remove terminal nucleotides. Following digestion with Hindlll. the products of this reaction were ligated into plasmid pUC8 (Messing, J., et al, supra) which had been opened with Hindlll and Hindi digestion. One of the resulting plasmid recombinants had the composition of pHSI as determined by nucleotide sequencing.

To complete the construction of the hybrid gene, the PstI fragment spanning the apoAI gene was isolated from plasmid pPSAI.2 (Example I) by digestion with PstI followed by polyacrylamide gel purification. The PstI fragment extends from a point in the 5' untranslated region through the entire coding sequence (with introns) and terminates beyond the poly A addition site. The purified fragment was blunt ended by digesting away the single-stranded 3' termini with T4 DNA polymerase in the presence of dCTP. The flush-ended molecules were then ligated to the expression vector plasmid pHSI, which had been opened by restriction with Sraal. followed by treatment with bacterial alkaline phosphatase. The products of the reaction were introduced into E. coli MC1061, and the recombinant. pMTAIR was identified. In the resulting plasmid. pMTAIR, the apolipoprotein Al coding sequences are in position to be expressed by hMT-II promoter. The plasmid is contained in the deposited CHO cells, CRL #8911.

Example VIII Al Synthesis in CHO Cells Plasmid pMTAIR from Example VII was introduced into the Chinese hamster ovary (CHO:KI) line of cultured cells (growing in McCoy's 5A medium with 10% fetal bovine serum) by co-transformation with pSV2:NEO (Southern, P., et al, J Mol Appl Genet 1:327 (1982)). a plasmid carrying a functional gene conferring resistance to the neomycin analog G418. Five hundred ng of pSV2:NEO and 5 μg of pMTAIR were applied to a 60 mm dish of cells in a calcium phosphate-DNA co-precipitate according to standard protocols (Wigler. M., et al. Cell 16:777 (1979)) with the inclusion of a two min "shock" with 15% (w/v) glycerol after 4 hr exposure to the DNA. A day later the cells were subjected to exposure to G418 at 1 mg/ml. This procedure yielded a pool of G418-resistant colonies most of which had also acquired stable inheritance of pMTAIR.

CHO:KI cells transformed with either plasmid pMT401 (control, hMT-II plasmid with no foreign insert) or pMTAIR were grown to 70% confluency in 9.6 cm² wells in standard medium (RPMI plus 10% dialyzed fetal bovine serum). Cells were preinduced with 1.5 x 10^-4

M ZnCl₂ for 7 hr, at which time 0.15 μg/ml L-[³⁵S]-methionine was added. Cells were incubated in the presence of label overnight, at which time medium was harvested, and cell debris removed by low speed centrifugation. One portion of the medium (1 x 10⁵ cpm) was added to SDS gel sample buffer and incubated at 100°C for 2 min. A second portion of medium (5 x 10⁵ cpm) was brought up to 1 ml final volume by the addition of

50 mM Tris-HCl. pH 6.8, 0.15 M NaCl, 0.1 mM EDTA, and 2%

(v/v) Triton X-100. Two μl of rabbit antiserum raised against human apoAI was added to the suspension. The suspension was then incubated, reacted with Protein A Sepharose, and washed as described in Example IV. The total and immunoabsorbed samples were subjected to SDS-PAGE as described in Example IV.

A comparison of polypeptides extracted from medium of CHO cells transformed with the recombinant plasmids pMT401 and pMTAIR is shown in Figure 9. A unique polypeptide migrating at molecular weight of 25,000 daltons appears in the medium from cells transformed with pMTAIR (lane 4). This unique protein is specifically immunoreactive to serum raised against apoAI (lane 5). There are no unique polypeptides present in lane 2, proteins extracted from cells transformed with pMT401, nor do any proteins immunoreact with apoAI antiserum, lane 3. Commercially prepared mature apoAI co-migrates with the mammalian expression product, and both migrate an an apparent molecular weight of 25,000 daltons due to the lipophilic nature of apoAI and the gel system used. Protein in Figure 9 represents apoAI being made from a transformed pool of CHO:KI cells (CHO/pMTAIR).

Example IX High-Producing CHO Cells for Al Production The pool of transformed CHO:KI cells from

Example VIII was plated at low density (100-200 cells/ml) in standard medium (DMEM 21/Coon's F-12 plus 10% FBS), producing individual clonal colonies after 4-7 days growth at 37°C. The colonies were separately picked and grown in the above medium to a cell density of about 10⁶ cell/ml.

The clonal cultures were individually assayed for Al expression by dot-blot Western, using the method described in Jahn, et al, Proc Natl Acad Sci (USA) 81:1684 (1984). Individual clones were seeded at 25% confluency in 12 well dishes in 1.5 ml DMEM21/Coons F12 plus 10% FBS. After 24 hr, the cells were washed once with 1 ml PBS and fresh medium containing 1 x 10^-4 M

Zn sulfate was added to begin preinduction. At 16 hr after preinduction, the cells were washed twice with 1 ml PBS and refed with 0.65 ml serum-free medium containing 3 x 10^-5 M Zn sulfate and 3 x 10^-5 M Fe sulfate. After 48 hr of serum-free induction, the media was harvested and centrifuged at 1000 rpm for 5 min to remove cell debris. 0.5 ml serum-free conditioned medium was applied to each well onto a nitrocellulose filter, (Schleicher & Schuell, Keene, NH) using a dot-blot microfiltration apparatus (Bio-Rad

Laboratories, Richmond, CA). Blots were probed with sheep antiserum raised against human apoAI

(Boehring-Mannhein) at 1:50 dilution. Antigen-antibody complex was detected with I¹²⁵ labeled Protein A

(Amersham). About 200 clonal colonies were examined. One of the Al high-producing cells, designated clone 104. was examined for Al expression, substantially according to methods described in Example VIII. Figure

10 shows SDS gel electrophoresis patterns of the proteins produced in the system. Lanes 1 and 3 are total S³⁵-methionine-labeled secreted proteins from

CHO-pMT401 (control) cells, and lanes 2 and 4, from

CHO: pMTAIR cells of clone 104. As seen. apoAI secreted from clone 104 is easily identifiable in the total secreted medium proteins without immunoprecipi- tation. This clone is producing apoAI at approximately a 30-fold higher level than the pool cells.

Example X Large-Scale Al Production A second high-producing CHO: pMTAIR clone from

Example IX, producing the same level of apoAI as clone 104 and designated clone 143, was used for large-scale Al production in roller bottles. The clone 104 is the high-producing CHO cell line identified as CRL 8911. A 10 cm plate of clone 143 was grown to confluence in 10 ml DMEM-21/Coon's F-12 (Gibco, Irvine, CA) plus 10% fetal bovine serum (Gibco) medium. The cells in the plate were washed once with 2 ml phosphate buffered saline (PBS), trypsinized. and suspended in 250 ml DMEM-21/Coon's F-12 medium containing 10% FBS and 15 mM HEPES buffer in an 850 cm² roller bottle. After 2 days of incubation at 37°C, zinc sulfate was added to a final concentration of 1 x 10^-4 M. At day 3, the cells were switched to serum-free DMEM-21/Coon's F-12 medium containing 6 x 10^-5 M zinc sulfate and 3 x 10^-5 M iron sulfate, marking the beginning of the production period.

The cells were cultured for 2 days following the medium switch (day 2 in Table I), and the conditioned medium harvested by low-speed centrifugation to separate cell debris from spent medium. The cell medium was fractionated by SDS gel electrophoresis, as in Example VIII. and the gel. after staining with Coomassie Brilliant Blue, was scanned optically to quantitate the amount of apoAI protein in the cell-free medium, using a purified Al standard obtained from

Calbiochem (La Jolla, CA) to produce a gel-staining calibration curve. Fresh serum-free medium containing 7 x 10^-5 M zinc (and 3 x 10^-5 M iron sulfate) was added to the cultured roller bottle and cultured for an additional 2 days (day 4 in Table I). Cell-free conditioned medium was again harvested at day 6 and examined for Al concentration, as above, and fresh medium containing 7.5 x 10^-5 M Zn was replaced in the roller bottle. The procedure was repeated at days 6-11, as indicated in Table I, with 8 x 10^-5 Zn being added at day 6 and thereafter. The concentration of zinc shown in the middle column in the table is that which the cells were exposed to from the period between harvests. Thus, the cells were exposed to 6 x 10^-5 Zn between production days 0-2, and so forth. The amount of Al produced, expressed in μg/ml/day, is shown at the right in the table. As seen, Al expression increased over the first

8 days, then plateaued at about 30 μg/ml/day after day

8.

Example XI

Cloning and Expression of AH in CHO Cells A human fetal liver cDNA library in λgt10 prepared as in Huynh. V.T., et al, DNA Cloning Techniques: A Practical Approach (IRL Press, Oxford, 1984) was probed with a 45-base pair oligonucleotide coding for amino acid residues 140-164 of human apolipoprotein AH (AH) (see Sharpe. C.R., et al, Nucl Acids Res 12:3917 (1984)) of 750,000 recombinants which were screened, and 10 positive were obtained. One of these, designated λAII, had a 440 base EcoRI insert corresponding to the full-sequence AH cDNA (Sharpe. et al, supra) plus ~ 20 bases of 5' untranslated region. The EcoRI insert was isolated from the phage and cloned into pHSI.

Two versions, one with the AH insert in the correct orientation (pHSI/AII) and one in the incorrect orientation (pHSI/incorrect). along with pHSI alone. were each transfected into CHO cells and selected with G418. The cells were labelled with ³⁵S-methionine and 0.5 ml of medium immunoprecipitated with 5 μl of rabbit anti-human apoAII antiserum obtained from Boeringer-Mannheim (Indianapolis. IN).

The secreted proteins were analyzed by SDS gel electrophoresis. as in Example II, with the results shown in Figure 13. Lanes 1, 2, and 3 are cell-free proteins from CHO cells transformed with pHSI (control), pHSI/AII, and pHSI/incorrect, respectively. As seen, the pHSI/AII transformed cells produce two low-molecular peptides of about 6,000-8,000 dalton molecular weight which are not seen in the two controls. The larger of the two pHSI/AII bands may be pro AH (78 amino acids), the other being mature AH (73 amino acids) or the different molecular weights may represent different degrees of glycosylation.

Example XII Cloning and Expression of Al Fragment in CHO Cells An expression system capable of producing apoAI (110-187) in CHO cells is constructed as described here. The plasmid pAP85 containing the apoAI (110-187) fragment is cut with Ncol and Hindlll and the insert containing the DNA coding for Met-apoAI (110-187)-Stop is isolated. The insert is blunted and ligated to the BamHI-NcoI adapter 5'-CATGGGATCC and the BamHI-Hindlll adapter 5'-AGCTGGATCC at the 5' Ncol site and the 3' Hindlll site, respectively. The DNA adapters are cut back with BamHI and the modified insert is purified by gel electrophoresis. The BamHI fragment is then ligated to the BamHI site of pMT apoAI(BS), the metallothionein expression plasmid which contains the 3' untranslated region of apoAI which includes a transcription termination signal, as described in detail in co-owned patent application Serial No. 680,358, and incorporated herein by reference. As seen in Figure 12, the plasmid is formed by digesting the pMTAIR vector from Example VII (on deposit) first with BamHI. and after blunt-ending with Klenow fragment, further digesting with StuI and treating with CIP. The ligation produces the desired pMT apoAI (BS) vector. The plasmid containing the insert in the correct orientation can be determined by restriction mapping. The plasmid containing the insert in the correct orientation is used to transfect CHO cells and high-producing clones selected as described.

Example XIII ApoAI Separation/Endogenous Lipid Culture medium from the CHO:pMTAIR clone 143 cells in Example X was used. The medium was adjusted to a density of 1.125 g/ml by addition of solid potassium bromide, then centrifuged at 38K rpm for 18 hours in a swinging-bucket rotor. The top fraction was then removed by a conventional slicing method, and dialyzed extensively against saline. The material, when examined by SDS gel electrophoresis as in Example VIII, gave a prominent band in the 25,000 molecular weight range, corresponding to apoAI. Between about 10%-20% of the total apoAI produced by the CHO:pMTAIR cells was in the upper fraction, as judged by the relative staining intensities of the 25,000 molecular weight bands on SDS gels from the upper and lower centrifugation bands.

Electron micrograph of the top-fraction material shows the presence of numerous disc-shape structures characteristic serum of apoAI/phospholipid complexes, as shown in Fig. 13 (described in Hamilton, R.L., et al, J Clin Invest 58:667 (1976)). Example XIV ApoAI Separation/Added Lipid Emulsion INTRALIPID, obtained from Cutter Labs (Berkeley, CA) was used as the substrate to bind the apolipoprotein ligand. INTRALIPID is an artificial lipid emulsion composed of soybean triacylglycerol and egg lecithins. The mesophase, or phospholipid-rich portion of the emulsion was removed by ultracentrifugal floatation in a discontinuous sucrose gradient. The bottom layer contained 2 ml of emulsion, 0.6 g sucrose and saline to give a final volume of 4 ml and a density of 1.06 g/ml. This layer, in a polyallomer tube of a Beckman SW41 rotor, was overlayed with 6 ml of an NaCl solution of d = 1.02 g/ml. Finally a third layer was made of 2 ml of saline solution of d = 1.006 g/ml. Centrifugation was at 28,000 rev/min at 10°C for 60 minutes.

After centrifugation, the triglyceride rich emulsion on the top of the gradient was separated from the infranatant solution by the tube slicing technique using a Beckman slicer.

High-producing CHO:pMTAIR clone 143 from Example IX and control CHO:pMT401 cells were cultured in roller bottles, as described in Example X. The culture medium was concentrated 100 times by ultrafiltration using an Amicon YM 10 membrane. The concentrated medium was incubated with purified INTRALIPID and centrifuged as described for INTRALIPID except that only one step centrifugation was used. Lipid-to-protein ratios of either 0.1, 0.3, or 0.6 ml of concentrated medium per ml INTRALIPID were used.

A sample from each top fraction was delipidated and fractionated by SDS-PAGE. The Coomassie-stained gels are shown in Figure 15. Bands 1, 2, and 3 are the gel patterns for the INTRALIPID fraction incubated with 0.1, 0.3, or 0.6 ml concentrated medium, respectively. Lane 4 is the non-lipid-bound protein fraction. As seen, (in the INTRALIPID bound fractions). apoAI is the only major protein that is separated with the less dense INTRALIPID fraction, and increasing quantities of medium yield increasing amounts of purified apoAI. The purity of apoAI was about 95%, as determined by staining density on the gels.

The unbound fraction contains some apoAI and other contaminating proteins secreted from the CHO cells.

Example XV Al Purification The Al enriched, delipidated protein precipitate from Example XIV was dissolved in 0.01 M Tris-HCl, pH 8.3, buffer made in 6 M urea. A 50 μl aliquot was injected in a high-performance liquid chromatography system (HPLC) with a C_{1 8} column equilibrated in a 20% acetonitrile, 0.1% trifluoroacetic acid (TFA) solution. After injection the sample was eluted with a gradient of acetonitrile from 20 to 70% containing 0.1% TFA at 2 ml/min flow rate. The elution profile, seen in Figure 14, shows a single major peak. The fractions corresponding to this peak were pooled and acetonitrile was removed by vacuum. The dried pellet was dissolved in phosphate buffer and analyzed by SDS polyacrylamide gel, as in Example VIII. Only a single band, corresponding in molecular weight to about 25,000 daltons, was observed.

The first 39 amino acids of the purified protein was sequenced by conventional methods. Table III below gives the results to the first 8 amino acids of the mature protein. The major sequence, about 95% of the protein, is that of mature apoAI. The minor (5%) sequence corresponds to the proapoAI having six additional N-terminal amino acids.

Table HI

Major Sequence

Asp Glu Pro Pro Gin Ser Pro Trp

Minor Sequence

Arg His Phe Trp --- --- Asp Glu Pro Pro Gin Ser Pro Trp

The sequence shows that the apoAI secreted into the medium is secreted as a processed mature apoAI and not as a pro Al. The first 39 amino acids were sequenced and found to be all correct for mature native apoAI.

Example XVI Nascent HDL-Like Particle Formation with Purified Al Repurified egg phosphatidylcholine (PC) obtained from Sigma Chemicals Co. (St. Louis, MO) was dissolved in ethanol (4 mg/ml). The lipid solution (4 ml) was dried under vacuum to a thin film and hydrated with 2 ml of PBS, pH 7.4. The cloudy suspension was sonicated at 15°C for 1 hour under a stream of nitrogen. The sonicated suspension was centrifuged at 38 K rpm for 1 hour, and the supernatant carefully removed. Figure 16, which is a negative-stain electron micrograph of the liposome preparation shows a mixture of unilamellar and multilamellar structures.

Pure apoAI was incubated with the PC liposomes in a weight ratio of 1:5 at 37°C for 1 hour. After incubation, apoAI bound to liposomes was separated from free apoAI by gel filtration on 10% Agarose column. This material was then analyzed by negative-stain electron microscopy.

Figure 17 shows (a) the many disc-like structures which have formed in the presence of Al, and (b) the disappearance of more of the clearly identifiable lamellar vesicle structure seen in Figure 16. The disc-like structures are similar in appearance to nascent HDL-like particles isolated from liver perfusate (see Hamilton. R.L., et al, supra).

Example XVII Stabilized Lipid Emulsion The emulsion was mixed with purified Al protein from Example XVI, at a wt ratio of 100:2 mg emulsion lipid/mg Al. The mixture was incubated with shaking at 37°C for 1 hour, to bind apoAI to the emulsion. Unbound apoAI was removed by centrifugation. The stability in serum of the lipid emulsion with and without apoAI was compared. Serum from rats treated with turpentine (5 ml/kg) or from control rats treated the same way but with saline were used. In each test, 180 μl of serum was incubated with 20 μl of lipid emulsion for 2 hours at 37°C with gentle shaking. After incubation, lipid-particle diameters were determined by laser light scattering using a sub-micron particle analyzer with optional size distribution processor analysis and multiple scattering angle detection (Coulter Model N4, Hialeah. FL).

The results are shown in Table IV below. The lipid emulsion without apoAI was unstable in serum from rats treated with turpentine. The emulsion being characterized by a bimodal distribution of sizes centered around 200 and 500 nm. By contrast, the emulsion containing apoAI showed no significant size change on exposure to serum obtained from turpentine-treated animals. Both emulsions were size stable in serum from control rats.

The results indicate that (a) serum from animals with inflammation contains a factor which promotes fusion of emulsion particles and (b) the fusion event is largely eliminated in emulsion stabilized with apolipoprotein.

Example XVIII

Cloning of Receptor Binding and Adjacent

Lipid Binding Regions of ApoB

A. First 5097 Nucleotides An approximately 5 x 10 ⁵member human adult liver cDNA library (where the insert size averaged 1 kb and the inserts were ligated into the EcoRI site of λgt10) was prepared by the method of Huynh, T., et al,

DNA Cloning Techniques: A Practical Approach (1984), Grover, D., ed., IRL Press, Oxford. For screening, 9 x 10 ⁵plaques propagated in C600 (HFL) cells were transferred to replica nitrocellulose filters and processed as described by Seilhamer, J.J., et al, DNA

3:309 (1984). The filters were prewashed for 2 hr in 3 x NaCl/Cit (1 x NaCl/Cit is 150 mM NaCl/15 mM sodium citrate, pH 7.0), 0.1% SDS at 55°C. and then prehybridized in 6 x NaCl/Cit, 200 μg/ml denatured salmon sperm DNA, 5 x Denhardt's, 0.05% sodium pyrophosphate for 1 hr at 50°C.

A 192-fold degenerate 23 base oligonucleotide probe which encodes, taking account of codon redundancy, the first 8 amino acids of the previously determined sequence of apoB-26 was used as a probe. The probe was

5' end labelled with T polynucleotide kinase (PL Biochemicals) and γ-[³²P]-ATP, added to the filters and incubated for 14 hr at 50°C. The filters were washed twice at room temperature in 5 x NaCl/Cit, 0.1%

SDS, 0.05% sodium pyrophosphate for 15 min and once at

50°C for 20 min, dried and autoradiographed with intensifying screens.

One positive plaque, designated LB25-1, was purified and the cDNA insert was subcloned in both orientations into M13/mp8 for sequencing. The nucleotide sequence of this 970 bp insert is shown in

Figure 18, along with the deduced amino acid sequence of that reading frame which agrees with the amino terminal sequence determined from the B-26 peptide. The sequence of LB25-1 contains an open reading frame extending 800 nucleotides downstream encoding 294 amino acids, and analysis of the predicted protein sequence directly upstream suggests the presence of a hydrophobic signal sequence preceded by a methionine residue. The EcoRI insert was subcloned into pBR322 to obtain pLB25-1 for amplification. pLB25-1 thus contains some 5' untranslated region, the 28 amino acids of the signal sequence, and the first 266 amino acids of the mature protein. Additional portions of the apoB encoding sequence were obtained by preparing a 2 x 10 ⁵member human adult intestine cDNA library in λgt10, as described above. The approximately 1 kb insert of pLB25 was denatured and used as probe to isolate a cDNA fragment designated IB7, containing an approximately 1.3 kb insert, about 800 bp of which extended beyond the 3' end of clone pLB25. Isolated, denatured IB7 insert was subcloned into pBR322 for amplification, creating pIB7. The purified pIB7 insert was denatured and used to screen the intestine library. One positive cDNA fragment designated I10 contained an approximately 3 kb insert, about 2.5 kb of which extended beyond the 3' end of IB7. The cDNA insert was subcloned into the EcoRI site of pBR322, creating pB10. Linearized, denatured pB10 insert was used as a probe to obtain a fourth cDNA fragment designated IB-(2)1, containing an approximately 2 kb insert, about 1 kb of which extends beyond the IB-10 sequence. The EcoRI cDNA insert was also subcloned into the EcoRI site of pBR322, creating pB(2)1. To produce an N-terminal portion of apoB in a suitable host cell, the 4 overlapping clones are digested to produce a single cDNA, as follows. A 479 bp EcoRI-SacI fragment from the pBL25 insert, and a 1162 bp SacI-EcoRI fragment from the pBI7 insert, are ligated into the EcoRI site of pBR322. The resulting plasmid is called pB25-7. When digested with EcoRI and Aval, the larger fragment contains 1342 bp encoding the signal sequence and first 427 amino acids of apoB. TWO synthetic oligonucleoties, oligo 1 with the sequence 5'-AATTCTGAATGATTGAG-3' and oligo 2 with the sequence 5'-TCGACTCAATCATTCAG-3', are synthesized by standard methods using commercially available reagents, and hybridized by the method of Rossi et al, J Biol Chem 257:9226-9229 (1982). The resulting fragment has the sequence: oligo 1 5'-AATTCTGAATGATTGAG oligo 2 GACTTACTAACTCAGCT-5'.

This fragment has an EcoRI overhang, stop codons in all 3 frames, and Sail overhangs. The EcoRI insert, isolated from pBI10, is ligated to the synthetic fragment. The insert now has stop codons and Sail overhangs at both ends. This insert is cloned into the Sail site of pBR322 for amplification, creating p10/stop. When p10/stop is digested with Sail and Aval, the larger Sall-Aval fragment contains 2733 bp encoding the carboxy-terminus of the cloned apoB protein portion. The expressed protein terminates (with one extra amino acid due to the insertion of the synthetic fragment) in frame after amino acid 1299.

The amino-terminal EcoRI-AvaI fragment isolated from pB25-7, described above, and the carboxy-terminal Aval-Sail fragment from p10/stop, also described above, are ligated into EcoRI-Sall digested pBR322. The resulting vector encodes the signal sequence and the first 1299 amino acids of the mature apoB. The amino acid sequence is identical to the first 1299 apoB amino acids shown in Figure 18 and includes an additional C-terminal aspartic acid residue due to the synthetic fragment. The EcoRI/Sall fragment is then excised and placed in a suitable expression vector, like those described above, for expression of the apoB N-terminal portion.

B. Nucleotides 5097-8247

The clones encoding the first 5097 nucleotides have been described above and in Protter, et al. (Proc. Nat. Acad. Sci. USA 83, 1467-1471 (1986) and Protter, et al. (Proc. Nat. Acad. Sci. USA 83, 5678-5682 (1986). This portion of the example describes the isolation of additional human intestine cDNA clones and their nucleotide and inferred amino acid sequences are described here.

The insert for the intestine cDNA clone 1(2)1 was made radioactive and used as a probe to screen the human intestine library as described previously (Protter, et al. PNAS 83, 5678-5682 (1986)). A clone 1(3)13 was plaque purified and the nucleotide sequence of the insert DNA obtained as described previously. A clone 1(7)1 was plaque purified and the sequence of the insert determined.

Expression of the ApoB sequences.

The construction described here is shown in Figure 19. The insert for the intestine cDNA clone 1(2)1 (Figure 19A) was isolated by agarose gel electrophoresis. cut with BstXI and the large fragment purified. The inser for the intestine cDNA clone 1(3)13 (Figure 19B) was isolat4d. cut with BstXI. and the large fragment purified. The two fragments were incubated, with EcoRI cut pBR322, ligated. and transfected into HB101. The resultant plasmid pBI (2)1/(3)13 has the restriction map shown in Figure 19C.

The plasmid pBI (2)1/(3)13 was partially cut with EcoRI and the insert of the intestine cDNA clone 1(7)1 was isolated by cutting with EcoRI and preparative agarose gel electrophoresis. The two pieces of DNA were ligated and used to transfect competent HB101 bacteria. The plasmid with the correct orientation of EcoRI fragment can be identified by restriction mapping. The resultant clone pBI(2)1/(3)13/(7)1 is shown in Figure 19E.

In order to add a stop codon to the 3' end of the apoB sequence the following oligos was used: oligo 3 (5'-AATTCTGAATG-3') and Oligo 4 (5'-TCGACATTCAG-3') in the construct. The plasmid pBI (2)1/(3)13/(7)1 was partially digested with EcoRI, partially digested with Sail, and then ligated to the oligos oligo 3 and oligo 4. The reaction was was then transfected into competent HB101 bacteria and the resultant colonies were screened for the correct construct with the oligos, oligo 5 (5'-ATTCCAGAATTCTGAATGTCGA-3') and oligo 6

(5'-CTGAATGTCGACCGATGCCCTTG-3') which have been kinased with (gamma-³²p)ATP. The selected plasmid is called pBI(2)1/(3)13/(7)1/stop (Figure 19F).

The plasmid above contains the EcoRI-Aval fragment isolated from pB25-7 and the carboxy-terminal Aval-Sail fragment from p10/stop in the EcoRI-Sall site of the pBR322. This plasmid was digested with EcoRI and Xhol and the large DNA fragment containing the amino-terminal region was isolated by agarose gel electrophoresis. The insert of the plasmid- pBI(2)1/(3)13/(7)1/stop was purified by complete digestion with Xhol, partial digestion with Sail, and preparative gel electrophoresis. These two DNA fragments were ligated into the EcoRI/Sall fragment of pBR322. The resulting correct construct can be described by restriction fragment mapping. The EcoRI/Sall insert can be excised and placed in a suitable expression vector. The nucleotide sequence and the corresponding amino acid sequence is shown in Figure 18.

The coding sequence, or suitable LBP coding portions thereof, can be placed in a suitable expression vector, by techniques like those described above, to achieve apoB expression in the corresponding expression system, such the high-producer CHO cells. The recombinantly produced apoBl can be purified according to the method of the invention.

Although the invention has been described with respect to specific embodiments and examples, it will be apparent that various changes and modifications can be made without departing from the invention.

Claims

IT IS CLAIMED:

1. A method of producing a purified lipid-binding peptide adapted to bind to phospholipids at one or more amphipatic alpha-helical peptide regions, said method comprising: providing a gene coding for the peptide. introducing the gene in expressible, heterologous form in a suitable expression system capable of synthesizing a mixture of peptides which includes the lipid-binding peptide, adding lipid to the peptide mixture to form a particulate, Hpopeptide complex composed of lipid particles and associated lipid-binding peptide, separating the Hpopeptide complex from nonlipid-binding peptides in the peptide mixture on the basis of the size or density of the complex.

2. The method of claim 1, wherein the gene coding for the lipid-binding peptide is derived from a human apolipoprotein gene.

3. The method of claim 2, wherein the gene coding for the lipid-binding peptide is a human apolipoprotein Al gene, or a portion thereof coding for an amphipatic, alpha-helical region of apoAI.

4. The method of claim 2, wherein the gene coding for the lipid-binding peptide is the human apolipoprotein AII gene, or a portion thereof coding for an amphipatic, alpha-helical region of apoAII.

5. The method of claim 2, wherein the gene coding for the lipid-binding peptide encodes at least a portion of a peptide selected from the group consisting of amino acid residues:

(a) 41-67 from apoCIII,

(b) 13-22, 28-39, and 42-51 from apoCII, (c) 7-14, 18-30, and 32-53 from apoCI,

(d) 17-30 and 51-60 from apoAII,

(e) 141-206 from apoAI,

(f) 203-221, 226-243, 245-266 from apoE,

(g) 320-352, 378-394 from apoB, (h) 33-330 in apoAIV.

6. The method of claim 1. wherein said introducing includes placing the gene in expressible form in a bacterial expression vector, and transforming the bacteria with the vector.

7. The method of claim 6, which further includes lysing the transformed bacteria to release such a mixture of peptides prior to said adding.

8. The method of claim 6, wherein the gene coding for the lipid-binding peptide is a human apoAI gene, or portion thereof coding for at least one lipid-binding region of apoAI. and the gene is introduced and expressed in E. coli.

9. The method of claim 1, wherein said introducing includes placing the gene in expressible form in a yeast expression vector, and transforming yeast cells with the vector, the peptide mixture consists of extracellular proteins secreted by the cells, and said contacting includes adding exogenous lipid particles to the mixture of secreted peptides.

10. The method of claim 1, wherein the gene coding for the lipid-binding peptide is a human apoAI gene, or portion thereof coding for at least one lipid-binding region of apoAI.

11. The method of claim 1, wherein said introducing includes placing the gene in expressible form in a mammalian-cell expression vector, and transforming cultured mammalian cells with the vector, the peptide mixture consists of extracellular proteins secreted by the cells, and said contacting includes adding exogenous lipid particles to the mixture of secreted peptides.

12. The method of claim 11, wherein the culture cells are Chinese hamster ovary cells, the expression vector includes a regulatable promoter derived from the human metallothionein II gene, and synthesis of the lipid-binding protein in Chinese hamster ovary (CHO) cells infected with the vector is under the control of the promoter, which is responsive to the concentration of divalent metals in the cell medium.

13. The method of claim 14, wherein the gene coding for the lipid-binding peptide is a human apoAI, apoAII or apoB gene.

14. The method of claim 12, wherein the CHO cells are selected from clonal variants for high-level expression of the peptide-coding gene.

15. The method of claim 14, wherein the gene coding for the lipid-binding protein is a human apoAI gene.

16. The method of claim 12, wherein the peptides are synthesized in the presence of a divalent metal ion selected from the group consisting of Zn, Mg, and Mn, at a metal ion concentration of at least about 3 x 10^-5 M.

17. The method of claim 1, wherein the lipid added is a triglyceride/phospholipid emulsion.

18. The method of claim 1, wherein the lipid added is a phospholipid vesicle suspension.

19. The method of claim 1, wherein said separating includes floating the lipopeptide complex in an aqueous medium whose density is greater than that of the complex.

20. The method of claim 19, wherein said floating includes centrifuging the complex-containing medium.

21. The method of claim 1, wherein said separating includes fractionating the lipid-containing peptide mixture by molecular-exclusion chromatography.

22. A method of producing purified human apoAI comprising: providing a gene coding for full-length apoAI, introducing the gene in expressible, heterologous form in an expression vector containing a regulatable promoter derived from the human metallothionein II gene. transforming Chinese hamster ovary (CHO) cells with the vector, by culturing the cells under promoter-induction conditions, producing a mixture of extracellular proteins containing mature apoAI, adding lipid to the extracellular protein mixture, to form a lipoprotein complex composed of lipid particles and associated apoAI. and separating the lipoprotein complex from nonlipid-binding proteins in the extracellular medium by floatation.

23. The method of claim 22. wherein the lipid added to the peptide mixture is supplied as endogenous lipid by the CHO cells.

24. The method of claim 22. wherein the lipid added is a triglyceride/phospholipid emulsion.

25. The method of claim 22, wherein the lipid added is a phospholipid vesicle suspension.

26. A method of producing a therapeutic lipopeptide composition composed of phospholipid-containing lipid and a lipid-binding peptide derived from human apoAI, apoAII, apoCI, apoCII, apoCIII, apoB, apoE, or apoAIV, providing a gene coding for the peptide, introducing the gene in expressible, heterologous form in a suitable expression system capable of synthesizing a mixture of peptides which includes the lipid-binding peptide. adding such lipid to the peptide mixture to form the lipoprotein composition, and separating such composition from nonlipid-binding peptides in the peptide mixture by floatation.

27. The method of claim 26, for producing a composition useful use in promoting reverse cholesterol transport, wherein the selected gene codes for mature apoAI or an alpha-helical, lipid-binding segment thereof, and the lipid includes phosphatidylcholine.

28. The method of claim 26, wherein said introducing includes placing the gene in expressible form in a mammalian-cell expression vector, and transforming Chinese hamster ovary (CHO) cells with the vector, the peptide mixture consists of extracellular peptides secreted by the cells, and the lipid added to the peptide mixture is supplied as endogenous lipid by the CHO cells.

29. The method of claim 26, for producing a stabilized lipid emulsion useful in parenteral nutrition therapy, wherein the lipid added is a triglyceride/phospholipid emulsion.

30. The method of claim 25 which further includes delipidizing the lipopeptide complex, purifying the delipidized apopeptide from contaminant peptides, and adding lipid to the resulting purified apopeptide.

31. An emulsion for parenteral nutrition therapy comprising lipid emulsion particles stabilized by a peptide containing one or more amphipatic alpha helical lipid-binding regions.

32. The emulsion of claim 31, wherein the lipid emulsion particles include triglyceride/phospholipid particles, and the peptide is derived from a human apolipoprotein.

33. The emulsion of claim 32, wherein the peptide is human apoAI.

34. The emulsion of claim 32, wherein the lipid emulsion particles are substantially free of phospholipid vesicle contaminants.

35. A method of producing a human apolipoprotein or lipid-binding peptide segment thereof, comprising: providing a gene coding for the protein or peptide. and introducing the gene in expressible, heterologous form in a suitable expression system capable of synthesizing a protein mixture which includes the expressed gene product.

36. The method of claim 35, wherein the expression system is E. coli. and the gene codes for apoAI or a lipid-binding peptide thereof.

37. The method of claim 36, wherein the gene codes for mature apoAI or a lipid-binding segment thereof.

38. The method of claim 35, wherein the expression system is S. cerevisiae. and the gene codes for apoAI or a portion thereof which codes for at least one lipid-binding region of apoAI.

39. The method of claim 38, wherein the gene codes for mature apoAI.

40. The method of claim 35, wherein the expression system is mammalian cells, and the gene is introduced in a compatible expression vector.

41. The method of claim 35. wherein the expression system is CHO cells, and the gene is introduced in an expression vector under the control therein of a regulatable promoter derived from the human metallothionein-II gene.

42. The method of claim 41, wherein the CHO cells include a clonal strain of cells which are selected for high-level expression of the gene.

43. The method of claim 41, wherein the gene codes for a full-length human apolipoprotein, and the CHO cells are effective to process the expressed protein to a mature apolipoprotein.

44. The method of claim 41, wherein the gene codes for apoAI, apoAII, apoB, or a lipid binding peptide thereof.

45. A method of purifying a recombinant lipid binding peptide produced in a bacterial, yeast, or mammalian-cell expression system comprising obtaining a mixture of peptides, including the lipid-binding peptide from the expression system, adding lipid to the peptide mixture to form a particulate. lipopeptide complex composed of lipid particles and associated lipid-binding peptide, separating the lipopeptide complex from nonlipid-binding peptides in the peptide mixture on the basis of the size or density of the complex.

46. The method of claim 45, wherein the peptide is selected from the group consisting of apolipoproteins, apolipoprotein lipid-binding segments, C-reactive protein, lung-surfactant proteins and lipid-binding serum complement factors.

47. The method of claim 45. wherein the peptide is a fused protein containing a lipid-binding peptide moiety fused to a hydrophilic protein moiety.

48. The method of claim 47, wherein the lipid-binding peptide moiety is attached at the C-terminus of the hydrophilic protein moiety.

49. The method of claim 47, wherein the two moieties are joined at a unique methionine residue in the fused peptide.

50. An N-terminal apolipoprotein B peptide and its gene having the following amino acid and nucleotide sequences, respectively: