NO164664B - RECOMBINANT DNA MOLECULES COVERING A DNA SEQUENCE CODING FOR A POLYPEPTIDE OF INTERFERON ALPHA (IFN-ALFA) TYPE. - Google Patents

Description

The present invention relates to recombinant DNA molecules. More particularly, the invention relates to DNA sequences expressed in suitable host organisms. The recombinant DNA molecules shown are characterized by the fact that they code for polypeptides of the interferon (IFN) cx type and comprise the DNA sequences indicated in the characterizing part of claim 1*.

As will be apparent from what follows, the DNA sequences can be used in the production of polypeptides that are useful as antiviral and antitumor or anticancer agents and methods.

In the following description, the interferon nomenclature published in Nature, 286, page 2421 (July 10, 1980) will be used. This nomenclature replaces the one used in the prioritization application, for example IF is now denoted IFN and leukocyte interferon is now denoted IFN-a.

Two classes of interferons ("IFN") are known to exist. Class I interferons are small, acid-stable (glyco)proteins that render cells resistant to viral infection (A. Isaacs and J. Lindenmann, "Virus Interference I. The Interferon", Proe. Royal Soc. Ser. B., 147, pages 258-67

(1957) and W. E. Stewart, II, The Interferon System, Springer-Verlag (1979) (hereinafter referred to as "The Interferon System")). Because to a certain extent cell-specific (The Interferon System) IFNs are not virus-specific, IFNs protect the cells against a wide spectrum of viruses.

Human interferons ("HuIFN") are classified into three groups, namely α, 3 and γ. HuIFN-ct or leukocyte interferon is produced in human leukocyte cells together with smaller amounts of HuIFN-3 (fibroblast interferon) in the lymphoblastoid cells. HuIFN-a is purified to homogeneity and characterized (eg, M. Rubenstein et al., "Human Leukocyte Interferon: Production, Purification To Homogeneity And Initial Characterization" Proe. Nati. Acad. Sei. USA, 76, pages 640-44 ( 1979)). ■ It is heterogeneous in size presumably because of its carbohydrate core. Two components are described, one with a molecular weight in the range 21,000-22,000 and the other in the range 15,000-18,000. The lower molecular weight component is reported to represent a non-glycosylated form. The minor form of HuIFN-a has also been reported to retain all or most of its HuIFN-a activity (W.E. Stewart, II et al., "Effeet Of Glycosylation Inhibitors On The Production And Properties Of Human Leukocyte Interferon", Virology, 97, pages 473-76 (1979)). A portion of the amino acid sequence of HuIFN-α from lymphoblast cells and their amino acid composition has been reported (K.C. Zoon et al., "Amino Terminal Sequence Of The Major Component Of Human Lumphoblastoid Interferon", Science, 207, pages 527-28 (1980) and M. Hunkapiller and L. Hood, personal communication (1980)).

HuIFN-a is also reported to exist in several different forms, see for example British patent application no. 2,037,296A. These forms seem to differ from each other both structurally and physiologically. No accepted nomenclature is well-known for these forms of HuIFN-a. Therefore, in the following description, each form will be denoted by a number after the general HuIFN-a designation, for example HuIFN-al or HuIFN3.

HuIFN-a can (like many other human proteins) also be polymorphic. Therefore, particular individual cells can form HuIFN-a types within the more general HuIFN-a group or forms within this group that are physiologically similar/but structurally somewhat different from the group or the form of which it is a part. Therefore, although the protein structure of HuIFN-a is generally well defined, particular individuals may produce a HuIFN-a as an insignificant variant of it, this heritable variation probably being less severe than the difference between the different forms of HuIFN-a.

HuIFN is usually not detectable in normal or healthy cells (The Interferon System, pages 55-57). Instead, the protein will be produced as a result of the cell being exposed to an IFN inducer. IFN inducers are usually viruses, but can also be non-viral in nature, such as natural or synthetic double-stranded RNA, intracellular microbes, microbial products and various chemical agents. Many efforts have been made to utilize these non-viral elicitors to render human cells resistant to viral infection (S. Baron and F. Dianzani (eds.), Texas Reports On Biology And Medicine, 35 ("Texas Reports"), pages 528-40 (1977)). These attempts have not been particularly successful. Instead, the use of exogenous HuIFN as such is now preferred.

Interferon therapy against viruses and tumors or cancer has been carried out

in different dose ranges under different routes of administration (The Interferon System, pages 305-321). For example, interferon has been effectively administered orally or by inoculation such as intravenous, intramuscular, intranasal, intradermal and subcutaneous inoculation, and in the form of eye drops, ointments and douches. It is administered one to three times daily in doses of 10 4 -10 7 units. The therapeutic treatment may depend on the patient and the condition being treated. For example, viral infections are usually treated with once or twice daily doses over several days to several weeks and tumors and cancers are usually treated with daily or multiple daily doses over several months or years. The most effective therapy for a particular patient must of course be determined by the attending physician who will take into account well-known factors with regard to the development of the disease, previous therapy and the patient's response to interferon when choosing the method of administration and dose amount.

As an antiviral agent, HuIFN is used for the treatment of

the following: respiratory infections (Texas Reports, pages 486-96), herpes simplex keratitis (Texas Reports, pages 497-500, acute hemorrhagic conjunctivitis (Texas Reports, pages 501-10), varicella zoster (Texas Reports, pages 511-15) , cytomegalovirus infection (Texas Reports, pages 523-27),

and hepatitis B (Texas Reports, pages 516-22). See also The Interferon System, pages 307-19. However, large-scale use of IFN as an antiviral agent will require larger amounts of IFN than have been available so far.

HuIFN also has other effects in addition to its antiviral effect. For example, it antagonizes the effect of colony-stimulating factor, inhibits the growth of hemopoietic-colony-forming cells and affects the normal differentiation of granulocytes and macrophage precursors (Texas Reports, pages 343-49). It also inhibits erythroid differentiation in DMSO-treated Friend leukemia cells (Texas Reports, pp.-.42Q.-28-)-.HuIFN may also play a role in regulating the immune response. Depending on the dose and timing of application in relation to an antigen, HuIFN may -a both be immunopotentiating and immunosuppressive in vivo and in vitro (Texas Reports, pages 357-69). In addition, it has been observed that especially sensitized lymphocytes will form HuIFN-a after contact with an antigen. Thus antigen-induced HuIFN-a can therefore be a regulator of the immune response and affect both the circulating antigen levels and cellular

immunity (Texas Reports, pages 370-74). HuIFN is also known to enhance the activity of killer lymphocytes and antibody-dependent cell-mediated cytotoxicity. (R. R. Her-berman et al., "Augmentation By Interferon Of Human Natural And Antibody Dependent Cell-Mediated Cytotoxicity", Nature, 277, pages 221-23 (1979), P. Beverley and D. Knight, "Kil-ling Comes Naturally", Nature, 278, pp. 119-20 (1979), Texas Reports, pp. 375-80). Both of these types are probably involved in immunological attacks on tumor cells.

Therefore, in addition to use as a human antiviral agent, HuIFN shows potential use in anti-tumour and anti-cancer therapy. (The Interferon System, pages 319-21).

It is now known that IFNs affect the growth of many classes

of tumors in many animals (The Interferon System, pages 292-304). Like many other antitumor agents, they seem most effective when used against small tumors. The antitumor effects of animal IFN depend on the dose and time, but have been demonstrated at concentrations below toxic levels. Consequently, many trials and clinical trials have been and are being conducted to clarify the antitumor and anticancer properties of IFNs.

These include the treatment of many malignancies such as osteosarcoma, acute myeloid leukemia, multiple myeloma and Hodgkin's disease (Texas Reports, pages 429-35). In addition, HuIFN has recently been reported to effect local tumor suppression when injected into subcutaneous tumor nodules in melanoma and breast carcinoma affected patients (T. Nemoto et al., "Human Interferons And Interalesional Therapy Of Melanoma And Breast Carcinoma", Amer. Assoc .For Cancer Research,

Abs. No. 994, p. 246 (1979)). Although the results of these clinical trials are encouraging, the antitumor and anticancer applications of IFN have been severely hampered due to lack of

on . sufficient supply of purified IFN. Today, HuIFN-a is produced either by growing human cells in tissue culture or from human leukocytes or collected from blood donors. It has been reported 2.6 x 10<9> IU crude HuIFN-a obtained from 800 1 cultured Namalva cells (P.J. Bridgen et al., supra). In very large blood centres, for example the Finnish Red Cross center in Helsinki, the annual production capacity is approx. 1011 IU crude HuIFN-a. Since the dose is typically 3 x IO<6> IU per patient per today, these sources are not sufficient to provide the necessary commercial quantities of HuIFN-a. Therefore, the production of HuIFN-a using other methods is desirable. As a result of the specific activity of IFN-a being

8 9

very high, of the order of 4 x 10 to 10 IU/mg, the amount of HuIFN-a for commercial use is small. For example, 100 g of pure HuIFN-a will give 3-30 million doses.

Recent advances in molecular biology have made it possible to introduce DNA coding for specific non-bacterial eukaryotic proteins into bacterial cells. In general for D.NA other than those produced via chemical synthesis^ construction of such recombinant DNA molecules will include the steps of producing a single-stranded DNA copy (cDNA) of a purified messenger RNA (mRNA) template for the desired protein, convert the cDNA into the double-stranded DNA, ligate the resulting DNA into an appropriate site in an appropriate cloning vehicle to yield a DNA molecule, and transform an appropriate host with the recombinant DNA molecule. Such a transformation can affect the host so that it produces the desired protein.\

Several non-bacterial proteins and genes have been obtained in

E.; coli using recombinant DNA technology. These include a protein displaying rat proinsulin antigenic determinants (L. Villa-Komaroff et al., "A Bacterial Clone Synthesizing Proinsulin", Proe. Nati. Acad. Sei. USA, 75, pages 3727-31 (1978)), rat growth hormone ( P.H. Seeburg et al., "Synthesis Of Growth Horraone By Bacteria", Nature, 276, pages 795-98 (1978)), mouse dihydrofolate reductase (A.C.Y. Chang et al., "Phenotypic Expression In E. coli Of A DNA Sequence Coding For Mouse Dihydrofolate Reductase", Nature, 275, pages 617-24 (1978)), human somatostatin (K. Itakura et al., "Expression In Escherichia coli Of A Chemically Synthesized Gene For The Hormone Somatostatin", Science,

198, pp. 1056-63 (1977)), European Patent Applications Nos. 0,001,929, 0,001,930 and 0,001,931, and related applications in other countries), A and B polypeptide chains in human insulin (D. V. Goeddel et al., "Expression In Escherichia coli Of Chemically Synthesized Genes For Human Insulin", Proe. Nati. Acad. Sei. USA, 76, pages 106-10 (1979), as well as the European and related patent applications mentioned above), antigens for human hepatitis B virus ( C. J. Burrell et al., "Expression In Escherichia coli: Of Hepatitis B Virus DNA Sequences Cloned In Plasmid pBR322", Nature, 279, pages 43-7 (1979) and M. Pasek et al., "Hepatitis B Virus Genes And Their Expression In E. coli", Kature, 282, pages 575-79 (1979)), human growth hormone (D. V. Goeddel et al., "Direct Expression In Escherichia coli Of A DNA Sequence Coding For Human Growth Hormone", Nature, 281, pages 544-51 (1979)), SV40 t antigen (T. M. Roberts et al., "Synthesis Of Simian Virus 40 t Antigen In Escherichia coli", Proe. Nati. Acad. Sei. USA, 76, pages 5596-600 (1979)), and human fibroblast interferon (Hu IFN-3) (T. Taniguchi et al., "Construction And Identification Of A Bacterial Plasmid Containing The Human Fibroblast Interferon Gene Sequence", Proe. Japan Acad., 55, Ser. B, pages 464-69 (1979), as well as personal communication 1980).

However, none of these recombinant DNAs target HulFN-a. This is the problem to which the present invention relates. Its solution is not made possible, such as for the recombinant DNA systems described above, by the availability of sequence information necessary to produce a synthetic gene, (such as somatostatin) or of a cell type or virus rich in a particular DNA sequence (for example hepatitis viral antigen) or mRNA types (for example rat insulin) which enable the production and identification of bacterial clones containing the desired hybrid DNA or a system which enables the selection of E. coli hosts expressing the desired protein (for example mouse dihydrofolate reductase). Nor is the problem alleviated by reporting a plasmid said to contain a DNA sequence that hybridizes to an mRNA from poly(A) RNA, which mRNA promotes HuIFN-0 activity in oocytes (for example, fibroblast interferon). Nor is the solution according to the invention directed at the apparent proposal in Research Disclosure No. 18309, pages 361-62 (1979) for the preparation of pure or substantially pure HuIFN-omRNA prior to cloning of the HuIFN-a gene.

Finally, it should be understood that selection of a DNA sequence or construction of a recombinant DNA molecule that hybridizes to mRNA from polyA RNA, which mRNA forms HuIFN activity in oocytes, is not sufficient to demonstrate that the DNA sequence or hybrid insert in the recombinant DNA molecule corresponds to HuIFN. Instead, only production of a polypeptide exhibiting an immunological and biological activity for HuIFN can truly demonstrate that the selected DNA sequence or engineered recombinant DNA molecule corresponds to HuIFN. It is further important that only after HuIFN activity has been shown via the DNA sequence, the recombinant DNA molecule or associated sequence, this can be used to select other sequences corresponding to HuIFN, according to the present invention.

It will therefore be understood from the foregoing that problems relating to the production of HuIFN-a using recombinant DNA technology differ significantly from the processes described above. In the present case, there must be a special DNA sequence of unknown structure, i.e. coding for the expression HuIFN-a in a suitable host, or this must be separated from a very complex mixture of DNA sequences in order for it to be able to. is used in the production of HuIFN-a. Furthermore, this localization and separation problem is made difficult by the extremely low concentration of the desired DNA sequence in the complex mixture and the lack of effective means to rapidly analyze the many DNA sequences in the mixture to select and separate the desired sequence .

The present invention solves the problems of locating and separating DNA sequences that encode the expression of HuIFN-oc in a suitable host and thereby provides DNA sequences and recombinant DNA molecules, by means of which a host is transformed to produce a polypeptide that exhibits a immunological or biological activity similar to that of human leukocyte interferon.

As a result of the present invention, it is possible to obtain a polypeptide that exhibits a HuIFN-a immunological or biological activity for use as antiviral, antitumor or anticancer agents or methods. The invention also allows the production of these polypeptides in quantities and by means of methods which have not been available until now.

As will be understood from the following description, the DNA sequences and the recombinant DNA molecules according to the invention will be able to direct the production, in a suitable host, towards the production of a polypeptide exhibiting a HIFN-α immunological or biological activity. Replication of these DNA sequences and recombinant DNA molecules in a suitable host also enables the production of large quantities of genes encoding these polypeptides. The molecular structure and properties of these polypeptides and genes can be easily determined. The polypeptides and genes are useful, either as produced in a host or after suitable derivatization or modification, in compositions and use in methods for detecting and improving the production of these products per se and for use as antiviral and antitumor or anticancer agents and methods in which such agents are used.

As previously mentioned, the present method differs from the known methods in that, according to the present method and in contrast to the above-mentioned methods, it involves the production and selection of a DNA sequence and recombinant DNA molecules that contain suitable DNA sequences that code for at least one polypeptide exhibiting a HuIFN-α immunological or biological activity.

It will be understood from the foregoing that the basic feature of the present invention is the provision of a DNA sequence which is characterized by the features specified in the characterizing part of claim 1. Fig. 1 schematically shows an embodiment of the present method in the production of a mixture of recombinant DNA molecules, which is characterized by inserted DNA sequences that code for the polypeptides of the IFN-oc type. Fig. 2 schematically shows the initial clone screening process. Fig. 3 schematically shows an embodiment of the clone screening process using DNA sequences according to the invention. Fig 4 shows a restriction map for one of the vectors according to the invention, the absolute position of each of the restriction sites for this clone is not shown. Figures 8-10 show more absolute positions for these restriction positions. Fig 5 schematically shows a method for determining the orientation of a DNA insert in a recombinant DNA molecule according to the invention. Fig. 6 shows the partial nucleotide sequence of certain cloning vectors useful in accordance with the invention. Fig. 7 shows results for a "Sephadex G-100" fractionation of a supernatant prepared from a bacterial culture transformed with a vector according to the invention. Figures 8-10 show the nucleotide sequence for a DNA insert in a recombinant molecule according to the invention. The sequence is numbered from the nucleotide after the polyG 5<1> tail to the nucleotide before the polyA residues and the polyC 3' tails. Nucleotides 57-125 represent the eh signal sequence and nucleotides 126-62 6 represent the fully developed interferon and the stop codon.

The amino acid sequence of the signal sequence is shown above its nucleotide sequence in lowercase letters, and the amino acid sequence of the mature interferon is shown above its nucleotide sequence in uppercase letters. Various restriction endonuclease recognition sites in this gene are also indicated in Figures 8-10. These positions are more absolutely localized than those shown in fig. 4. Fig. 11 is a schematic comparison of the restriction maps for the two DNA inserts of recombinant DNA molecules according to the invention. Fig. 12-16 show nucleotide sequences for two DNA inserts for recombinant DNA molecules according to the invention. The sequences are numbered from the nucleotide following the polyG 5' tail and to the nucleotide before the polyA residues and polyC 3' tails. The amino acid sequences of the signal sequence in each of these inserts are shown above their respective nucleotide sequences in lowercase letters and the amino acid sequence of the mature interferon is indicated above its nucleotide sequence in uppercase letters. Fig. 17 shows a partial restriction map of Z-pBR322(Pst)/HcIF-11-206 and the sequencing strategy used to determine the nucleotide sequence of the Hif-II-206 fragment shown in Figures 12-16. Fig. 18 shows a partial restriction map of a series of hybrid phages that hybridize to the Hif-2h fragment. Fig. 19 shows a partial restriction map of a hybrid insert in Z-pBR322Pst/HchrIF-35HBa and the sequencing strategy used to determine its nucleotide sequence. Figs. 20-23 show the nucleotide sequence of The HchrIF-35HBo fragment and the amino acid sequence derived therefrom. Fig. 24 shows partial linkage maps for HuIFN-α related genes. Arrows indicate regions that form R-loops that induce leukocyte poly(A) RNA. The dashed box (chr-16) indicates that this sequence was deduced from "blotting" experiments but was not apparent by R-loop mapping. Fig. 25 is a schematic diagram of the construction of plasmid C8-IFN-ctl. Fig. 26 is a schematic diagram of the construction of plasmid LAC -AUG(a2) Fig. 27 shows the construction of the AUG initiation codon and the CYS initiation codon in the construction of LAC-AUG(a2). Fig. 23 is a schematic view of the construction of the plasmid C8-IFN-cx2 and the hybrid molecules I, II, III and IV. Figs. 29-32 show the nucleotide sequence and amino acid sequence encoded thereby for IFN-α4b and its signal sequence.

The invention shall be described in more detail below and in this description the following terms are used: Nucleotide, a monomeric unit of DNA or RNA consisting of a sugar residue (pentose), a phosphate and a nitrogen-containing heterocyclic base. The base is connected to the sugar residue via the glycoside carbon (the 1<1> carbon in the pentose) and the combination of the base and the sugar is a nucleotide. The base characterizes the nucleotide. The four DNA bases are adenine ("A"), guanine ("G"), cytosine ("C"), and thymine ("T"). The four RNA bases are A, G, C and uracil ("U").

DNA sequence, a linear series of nucleotides connected to each other via phosphodiester bonds between the 3' and 5' carbons in adjacent pentoses.

Codon, a DNA sequence of three nucleotides (a triplet) which via mRNA codes for an amino acid, a translation start signal or a translation stop signal. For example, the nucleotide triplets TTA, TTG, CTT, CTC, CTA and CTG code for the amino acid leucine ("Leu"), TAG, TAA and TGA are translation stop signals and ATG is a translation start signal.

Reading frame, the grouping of codons during translation of mRNA into amino acid sequences. During transfer, the correct reading frame must be maintained. For example, the sequence GCTGGTTGT AAG can be translated into three reading frames or phases, each giving a different amino acid sequence:

Polypeptide, a linear series of amino acids connected to each other by peptide bonds between the α-amino and the carboxy groups of adjacent amino acids.

Genome, all the DNA of a cell or virus. It includes, among other things, structural genes that code for the polypeptides of the substance, as well as operator, promoter and interaction sequences such as Shine-Dalgarno sequences.

Structural gene, a DNA sequence that codes via its "messenger" RNA ("mRNA") sequence for amino acids characteristic of a specific polypeptide.

Transcription, the process of producing mRNA from a structural gene.

Translation, the process of producing a polypeptide from mRNA.

Expression on. the process that a structural gene undergoes to form a polypeptide. This is a combination of transcription and translation.

Plasmid, a non-chromosomal double-stranded DNA sequence comprising an intact "replicon", so that the plasmid is duplicated in the host cell. When the plasmid is placed inside a single-celled organism, the organism's properties will change or be transformed as a result of the plasmid's DNA. For example, a plasmid carrying a gene for tetracycline resistance (Tet<R>) will convert a cell that was previously sensitive to tetracycline into one that is resistant to tetracycline. A cell transformed by a plasmid is called a "transformant".

Phage or bacteriophage. bacterial virus, many of which contain DNA sequences encapsulated in a protein envelope or a protein coat ("capsid").

Clonin<g>s carrier. a plasmid, phage DNA or other DNA sequences that are capable of duplicating in a host cell and are characterized by one or a small number of endonuclease-recognizing positions, into which such DNA sequences can be inserted in a specific way without there is a loss of DNA<1>'s necessary biological function, i.e. duplication, production of corresponding proteins or loss of promoter or binding positions, and which contains a marker suitable for use in the identification of transformed cells, e.g.

eg tetracycline resistance or ampicillin resistance.

A cloning vehicle is often called a vector.

Cloning, the process of obtaining a population of organisms or DNA sequences derived from such an organism or sequence by asexual reproduction.

Recombinant DNA molecule or hybrid DNA, a molecule consisting of segments of DNA from different genomes that are joined end-to-end outside living cells and that have the ability to infect certain host cells and are retained therein.

Expression control sequence, a nucleotide sequence that controls and regulates expression of structural genes when it is operatively linked to these genes. They include the lac system, the trp system, the phage master operator and promoter regions, the fd coat protein control region, and other sequences known to control the expression of genes of prokaryolytic or eukaryolytic cells and their tissues.

With reference to fig. 1 schematically shows an embodiment of the method for producing a mixture of recombinant DNA molecules, some of which are characterized by inserted DNA sequences that code for polypeptides that display an immunological or biological activity corresponding to that of human leukocyte interferon.

PREPARATION OF POLY( A) RNA-CONTAINING HUMAN INTERFERON

mRNA (IFN-amRNA)

Human leukocytes were induced for 5 h at 37°C with Sendai virus and extracted to give a poly(A) RNA mixture containing human leukocyte interferon mRNA ("HuIFN-amRNA"). 'Induction was performed using the Cantell procedure (The Interferon System, pages 130-131 and according to the references therein). The poly(A) RNA mixture is shown in Fig. 1 without regard to its actual proportions. Induced leukocytes were collected and 10 ii cells were resuspended in 1 L of a solution containing 8 g NaCl, 0.2 g KCl, 1.15 g Na 2 HPO 2 H 2 O and 0.2 g KH 2 PO 4 dissolved in water ("PBS") and added slowly with vigorous stirring to 17 1 20 mM Tris-HCl (pH 7.5), 1 mM EDTA ("TE buffer"), 2% sodium dodecyl sulfate ("SDS") in a 50 1 separatory funnel. "Seif-digested" Pronase (Calbiochem) was added to 200 µg/ml and the solution stirred in oxygen at room temperature. 10^ counts/min. ("cpm")

125

of I-globin mRNA was added as a marker for the recovery of poly(A) RNA and to control the mRNA degradation during the subsequent steps. 2M Tris-HCl (pH 9) in an amount corresponding to 1/20 of the total volume ("1/20 vol") was added and the mixture extracted under vigorous stirring with 15 L of redistilled phenol at 10 min. 3 1 of chloroform was added and the mixture stirred for 5 min. After waiting for 3 0 min. for phase separation, the aqueous phase was recovered and again extracted with phenol and chloroform. The resulting aqueous phase totaling 19.1 L was combined with 60 g of SDS. Nucleic acids were precipitated from the aqueous phase with 1/10 vol 3M sodium acetate (pH 5.5) and 2 vol ethanol.

After overnight storage at -20°C, the fibrous nucleic acid precipitate was removed by filtration through a plastic sieve. This material was then stirred with 200 ml of TNE (50 mM Tris-HCl (pH 7.5), 100 mM NaCl, 5 mM EDTA) containing 0.5% SDS. It was finally dissolved by adding a further 350 ml of this solution. The non-fibrous precipitate was collected by centrifugation in "Sorvall" bottles in a "Sorvall RC-3" centrifuge for 15 min. at 5000 rpm. and redissolved in 350 ml TNE containing 0.5% SDS.

The two TNE solutions were combined, extracted 3 times with 1 vol phenol, 3 times with 1/2 vol ether and 3 times with 1 vol ether. The RNA recovery from the aqueous phase totaled 775 mg, determined by adsorption at 260 nm.

Isolation of the poly(A) RNA mixture was achieved by repeated batch adsorptions to oligo(dT) cellulose ("type 7", P-L Biochemicals, Inc.). 2.7 g of oligo(dT) cellulose was added to 500 ml, i.e. approx. half of the RNA-containing solution described above. After stirring for 1 h at room temperature to provide adsorption of poly(A) RNA to the oligo(dT) cellulose, the cellulose and the mixture of mRNAs were bound thereto, collected by centrifugation and washed once with 50 ml of TNE and a second time with 15 ml of TNE. The bound poly(A) RNA was then eluted with five consecutive washes with 2 ml of t₂O. The yield was 860 µg of poly(A) RNA determined by its optical density (Preparation A). The supernatant RNA solution from the first adsorption was subjected to two further adsorption cycles in the same manner as described above. The second and the third adsorption gave respectively 600 µg and 170 µg of RNA which remained. combined (Production B) .

RNA was analyzed for HuIFN-amRNA by injection into Xenopus laevis oocytes (The Interferon System), pages 93-95): RNA was dissolved in 15 mM Tris-HCl (pH 7.5), 88 mM NaCl ("TNK buffer ") to give a concentration of approx. 1 mg/ml. 50 ni of this solution was injected into each of 50 oocytes. The oocytes were incubated overnight at room temperature in a "Barth" medium (Gurdon, J. Embryol and Exper. Morph., 20, pages 401-414 (1968) and Barth, J. Embryol and Exper. Morph., 7, pages 210 -222 (1959)). The incubated oocytes were then purified and homogenized with a Pasteur pipette in a 1.5 ml "Eppendorf" centrifuge tube in 0.5 ml of 52 mM Tris glycine buffer (pH 8.9). The mixture was centrifuged

for 2 min. in an "Eppendorf" centrifuge and the supernatant was withdrawn and frozen at -20°C for analysis. IFN-α activity was determined by the plaque reduction assay described by H. Strander and K. Cantell, "Production Of Interferon By Human Leukocytes In Vitro", Ann. With. exp. Fenn., 44, pp. 265-73 (1966). One unit of IFN-a reduces viral plaques by 50%. The activity of an IFN-a preparation is expressed in

hold to human reference HuIFN-a 69/19 (International Symposium on Standardization of Interferon and Interferon Inducers, 1969). Alternatively, the determination based on reduction of the cytopathic effect, essentially as described by W. E. Stewart, II and S. E. Sulkin, "Interferon Production In Hamsters Experimentally Infected With Rabies Virus", Proe. Soc. Exp. Biol. Ned., 123, pages 650-3 (1966), is used except that human CCL-23 cells were used and infected with Mengo virus. The oocyte extracts exhibited 300 IU of IFN-α activity per injected RNA. In later determinations, the incubation time for injected oocytes was 48 hours and only the incubation medium was examined because the majority of the interferon was secreted by the oocytes (A. Colman and J. Mor-ser, "Export Of Proteins From Oocytes Of Xenopus laevis", Cell, 17, pp 517-26 (1979)). For further purification of poly(A) RNA, sufficient 0.5 M ethylenediaminetetraacetic acid ("EDTA") was added to the poly(A) RNA preparation A to bring the concentration to 5 mM EDTA. The resulting solution was extracted twice with a corresponding volume of TNE saturated with phenol and 5 times with a corresponding volume of ethers. The solution was then passed through a 0.1 ml "Chelex-100 Bio-Rad" column, heated for 90 s at 100°C and placed on a 13 ml 5-23% sucrose gradient containing 50 mM Tris-HCl (pH 7.5 ), 1 mM EDTA, 0.2 M NaCl. 10,000 cpm 5<1->terminated <32>P-labeled DNA fragments prepared by a simultaneous treatment of pBR322 with the restriction enzymes HindIII and Pstl (New England Biolabs) were added as size markers. Centrifugation was performed in a "SW40" rotor at 10°C and 35,000 rpm. for 16 h. Fractions (0.6 ml) were collected with an ISCO gradient collector at 1 ml/min. The fractions were probed for HuIFN-amRNA as described above, and their position relative to the 3 2P DNA markers was noted for future reference. In the subsequent centrifugations, HuIFN-amRNA-containing fractions were identified in relation to the markers. The fractions with HuIFN-amRNA activity contained 80 pq of poly(A) RNA. They were mixed with 2 vols of TNE containing 0.5% SDS and 0.02% polyvinyl sulfate (in later preparations polyvinyl sulfate was omitted) and added to a 50 } l oligo(dT) cellulose column. After washing the column as described above, 40 µg of the RNA mixture was eluted with 4 washes with 0.6 ml of sterile distilled water. After ethanol precipitation, RNA was dissolved to give 1 mg/ml in 0.5 mM EDTA.

A determination of HuIFN-amRNA activity was performed as described above on a portion of the poly(A) RNA precipitate. It showed a specific activity of 3600 IU interferon/µg injected RNA. The sucrose gradient had therefore enriched poly(A) RNA approx. 10-fold with respect to HuIFN-amRNA. In a subsequent similar preparation, a 40-fold enrichment was achieved. Preparation B was purified in a similar way, and when it showed a similar specific activity as preparation A, the two preparations were combined.

It should be emphasized that the actual poly(A) RNA product obtained from the sucrose gradient contained a very large number of different mRNAs. Apart from mRNA specific for IFN-α, the other mRNAs constitute unwanted contaminants (Fig. 1). Unfortunately, these contaminating RNAs will behave similarly to HuIFN-amRNA through the remaining cloning method of the invention. Therefore, their presence in poly(A) RNA will lead to the eventual production of a large number of unwanted bacterial clones containing genes encoding polypeptides other than IFN-α. This contamination leads to complex isolation problems when isolating the desired IFN-α hybrid clones. In the case of IFN-α, the isolation problem is further complicated by the lack of a sufficiently purified sample of HuIFN-amRNA or DNA or parts thereof that can act as a screening probe for identification of the desired clones. Therefore, the screening process for the IFN-a clones is very time-consuming and difficult.Furthermore, because only a very small percentage of the IFN-a clones themselves are expected to express IFN-a in a biologically active or immunologically active form/is the isolation of an active clone a "needle in a haystack" isolas ion process.

Advantageously, recombinant DNA technology can be used to provide a further purified sample of HuIFN amRNA or cDNA or part thereof. This purified mRNA or cDNA can be used to rapidly screen a large number of bacterial clones and thereby substantially increase the probability of isolating a clone that expresses IFN-α in an active form.

Poly(A) RNA enriched with respect to IFN-amRNA (Preparations A<+>B)^ was used as a template to prepare single-stranded complementary DNA (cDNA) (Fig. 1) (Cf, A. Efstratiadis et al., "Full Length And Discrete Partial Reverse Transcripts Of Globin And Chorion mRNAs", Cell, 4, pages 367-78 (1975) and references therein). 800 µl of the reaction mixture containing 40 mM Tris-HCl (pH 7.5), 30 mM NaCl, 5 mM MgCl2, 0.5 mM DTT (Cal-Biochem), 20 pg/ml oligo(dT) 12-18 (P&L Biochemicals), 5 mM dGTP (Schwarz), dCTP (Laevosan) and dTTP (Sigma), 5 mM <32>P-dATP (NEN), specific activity 100,000 cpm/nmol), 60 µg/ml poly(A) RNA and 280 units avian myeloblastosis virus (AMV) reverse transcriptase (gift from Life Sciences, Inc., St. Petersburg, Florida). After incubation for 1 hour at 37°C, 0.5 M EDTA and 20% SDS (recrystallized) was added to 10 mM EDTA and 0.1% SDS. The mixture was extracted with 1 vol of phenol (distilled). The phenol phase was washed with 200 µl of 200 mM Tris-HCl (pH 7.5), 1 mM EDTA and 0.1% SDS and the aqueous phases combined. These were extracted with an equal volume of ether ("Fluka, pro anl.") and chromatographed on a 5 ml "Sephadex G-100 column in TNE. Fractions of 0.1 ml were collected at 0.3 ml/min. Fractions which exhibits radioactivity, (measured by Cerenkov radiation) were combined and 3M sodium acetate was added to 0.3M. The nucleic acids were precipitated with 2.5 vol ethanol. After overnight storage at -20°C, the samples were centrifuged and the supernatant discarded. The precipitate

was dissolved in 180 µl of distilled water and transferred to a silicone treated "Eppendorf" tube. 20 µl of 5M MaOH was added and the mixture was kept at room temperature for 40 min. 20 µl of 5M sodium acetate, 100 µl of distilled water and 500 µl of ethanol were added. After cooling overnight at -20°C, the precipitate formed was collected by centrifugation at a force equivalent to 10,000 times gravity (10,000 G) for 20 min. at 0°C. The yield of single-stranded cDNA was 10 µg.

Again, it must be understood that the single-stranded cDNA product prepared as indicated above is in reality a complex mixture of a large number of different cDNAs transcribed from the corresponding mRNAs present in the poly(A) RNA mixture.

(Fig. 1). Only a very small number of these cDNAs are IFN-α related, i.e. H ul FN-acDNA1 is. Another factor that acts to complicate the cDNA mixture(s) is that each mRNA type in the poly(A) RNA mixture is usually not completely transcribed. Instead, for each mRNA type, the transcription process may be stopped before the end of the mRNA is reached. Therefore, a large number of cDNA types can be prepared from each mRNA type (not shown in Fig. 1). Each type will behave more or less the same in the subsequent cloning processes so that bacterial clones will be produced that contain recombinant DNA molecules that only have a fragment of the gene for a particular protein. Presence of these fragment-containing clones will further complicate the final clone screening process.

The sizes of the various single-stranded cDNAs were determined by electrophoresis of a small portion of an alkaline 2% agarose gel using 30 mM NaOH, 2 mM EDTA as electrolyte (M. W. McDonell et al., "Analysis Of Restriction Fragments Of T7 DNA And Determination Of Molecular Weights By Electrophoresis In Neutral And Alkaline Gels", J. Mol. Biol., 110, pp. 119-46. (1977)). <32>P-cDNA with a length of 600-1000 nucleotides compared to the single-chain globulin cDNA and <32>P-labeled DNA fragments were used as size markers.

The single-stranded cDNA can be made double-stranded by treatment with DNA polymerase I (T. Maniatis et al., "Amplification And Characterization Of A Ø-Globin Gene Synthesized In Vitro", Cell, 8, pp. 163-82 (1976)). The precipitated single-stranded cDNA obtained as above was dissolved in 200 µl H 2 O, heated to 100°C for 2 min. and incubated in 500 µl 0.1

M heat-denatured potassium phosphate buffer (pH 6.9), 10 mM MgCl2# 10 mM DTT (Calbiochem), 1 mM each of dATP (Merck), dGTP (Schwarz) and dCTP (Laevosan), 1 mM <3>H-dTTP (NEN , specific activity 100,000 cpm/nmol) and 150 units/ml E. coli DNA polymerase I ("Boehrincjer-Mannheirn") . After 6.5 hours at 15°C, 0.5 M EDTA and 20% SDS was added to 10 mfl EDTA and 0.1% SDS. The mixture b<1>a extracted with 500 µl of phenol and the phenol phase was re-extracted with 250 µl of 20 mM Tris-HCl (pH 7.5), 5 mM EDTA ("TE buffer"). The two aqueous phases were combined and chromatographed on a 5 ml "Sephadex G-100" column under the same conditions as previously described. Sodium acetate (3M) was added to 1.3 M and 2.5 vol of ethanol was mixed in to precipitate the DNA. A total amount of 13 µg of DNA was recovered.

The DNA obtained was treated with nuclease S^, prepared by the method according to R. C. Wiegand et al., "Specificity Of The S^ Nuclease From Aspergillus oryzal", J. Biol. Chem., 250, pp. 8848-55 (1975). The precipitated DNA was dissolved in 250 µl S1 buffer (0.2 M NaCl, 50 mM sodium acetate (pH 4.5), 10 mM zinc sulfate) and heated at 37°C for 30 min..1.5 µl S-^ enzyme (11 units/µl) was added to the mixture and this was incubated at 37°C for 30 min. SDS and EDTA were added to 0.1% SDS and 5 mM EDTA and the mixture extracted with 250 µl of phenol. The phenol phase was washed with 100 µl of TE buffer. The aqueous phases were combined and chromatographed on a "Sephadex G-100" ("Pharmacia") column in TNE. 0.1-ml fractions were collected at 0.3 ml/min. and "Cerenkov" radiation was determined for each fraction. 8 µg of double-stranded cDNA was recovered after precipitation with ethanol and sodium acetate as indicated above.

It must again be understood that the double-stranded cDNA produced as indicated above is a mixture of a large number of cDNAs and fragments thereof, while quite a few are made up of HuIFN acDNA or its fragments (Fig. 1).

A large number of host/cloning vehicle combinations can be used in cloning the double-stranded cDNA prepared as described above. E.g. suitable cloning vehicles may consist of segments of chromosomal, non-chromosomal and synthetic DNA sequences, such as various known bacterial plasmids, for example plasmids of E. coli including col E1, pCR1, pBR322 and their derivatives, broad host range plasmids, e.g. RP4, phage DNA, for example the many derivatives of phage for example NM 989, and vectors derived from combinations of plasmids and phage DNA such as plasmids modified for the use of phage DNA or other expression control sequences or yeast plasmids such as 2u plasmid or derivatives hence. Useful hosts may include bacterial hosts such as strains of E. coli, for example E. coli HB 101, E. coli X1776, E. coli X2282, E. coli MRCI and strains of Pseudomonas. Bacillus subtilis, Bacillus stearo-thermophilus and other bacteria, yeasts and fungi, animal or plant hosts such as animals (including humans)'. or plant cells in culture or other hosts. Naturally, not all host/vector combinations will be equally effective. The particular choice of host/cloning carrier combination can be made by a person skilled in the art after assessing the presented principles without deviating from the framework of the present invention.

Furthermore, within each specific cloning vehicle, different positions can be selected for the introduction of the double-stranded DNA. These positions are usually designated by the restriction endonuclease that cuts them. E.g. in pBR322, the Psti<p>osision is located in the gene for B-lactamase, between the nucleotide triplets that code for amino acids 181 and 182 in this protein. This position was used by Villa-Komaroff et al., as noted above, in their synthesis of protein displaying rat proinsulin antigen determinants. One of the two jiindll endonuclease recognition sites lies between the triplets encoding amino acids 101 and 102 and the other of several Tag positions at the triplet encoding amino acid 45 of B-lactamase in pBR322. In the same way, the EcoRI position and the Py^jII position in this plasmid lie outside any coding region, the EcoRI position being located between the genes that code for resistance to tetracycline and ampicillin, respectively. This position was used by Itakura et al. and Goeddel et al. in their recombinant synthesis schemes, mentioned above. It must of course be understood that a cloning vector useful in the practice of the present invention does not require a restriction endonuclease site for introduction of the selected DNA fragment.

The vector or cloning vehicle and the particular position selected therein for attachment of a selected DNA fragment to

yield a recombinant DNA molecule/is determined by a number of different factors, for example the number of positions sensitive to a particular restriction enzyme, the size of the protein to be expressed, the susceptibility of the desired protein to proteolytic degradation by the host cell enzymes, contamination of the protein to be expressed by the host cell proteins which are difficult to remove during purification, expression characteristics such as the position of the start and stop codons relative to the vector sequences, as well as other factors well known to a person skilled in the art. The choice of vector and insertion position for a particular gene is determined by a balance between these factors, and not all choices are equally effective for a given case.

Although a number of methods are known in the art

for the introduction of foreign DNA into a cloning vehicle or vector to give a recombinant DNA molecule, the preferred method for the first construct according to the invention is that described by Villa-Komaroff et al., supra, and shown in fig. 1. This method is characterized by treating the plasmid (especially pBR322) with the restriction enzyme specific for the chosen position for the introduction (especially Pstl) and adding dGMP tails to the end parts by terminal transferase. dGMP tails are added to the 5' ends of the split plasmid to regenerate the Pstl position and allow attachment of a cDNA fragment carrying complementary tails.

Similarly, the double-stranded cDNA is extended by adding dCMP tails to the 3* end to join the tailed plasmid. The latter and the cDNA are then joined to introduce the cDNA into the appropriate position of the plasmid and to circularize the hybrid DNA, the complementary nature of the tails enabling their cohesion (Fig. 1). The resulting recombinant DNA molecule carries a gene for the selected restriction site (Fig. 1).

Of course, other known methods of introducing DNA sequences into cloning vehicles to yield recombinant DNA molecules may be equally useful. This includes, for example, direct annealing, synthetic linkers, exonuclease and polymerase-linked repair reactions followed by annealing, or extension of the DNA chain with DNA polymerase and an appropriate single-stranded template followed by annealing. It must of course be understood that the nucleotide sequences or cDNA fragments introduced in the selected position in the cloning carrier may include nucleotides which are not part of the relevant structural gene for the desired polypeptide, or may include only a fragment of the complete structural gene for it. desired protein. It is only necessary that with the DNA sequence that is introduced, the transformed host will produce a polypeptide that exhibits a biological or immunological activity for HuIFN-a, or that the DNA sequence itself is used as a hybridization probe to select clones that. contains DNA sequences which are suitable for the production of polypeptides with immunological or biological activity such as HulFN-a.

The cloning vehicle or vector containing the foreign gene is used to transform a host to allow the host to express the protein or part thereof that the hybrid DNA encodes. The choice of suitable host is also controlled by a number of factors well known in the art. These include, for example, compatibility with the chosen vector, the toxicity of the proteins coded for by the hybrid plasmid, ease of recovery of the desired protein, expression characteristics, biosafety and costs. A balance between these factors must be found based on the understanding that not all hosts will be equally efficient in expressing a particular recombinant DNA molecule.

According to the present invention, the preferred initial cloning vehicle is bacterial plasmid PBR322, and the preferred initial restriction endonuclease position therein is the Pstl position (Fig. 1). The plasmid is a small (molecular weight approx. 2.6 megadaltons) plasmid that carries resistance genes for ampicillin (Amp) and tetracycline (Tet) antibiotics. The plasmid has been fully characterized (F. Bolivar et al., "Construc-tion And Characterization Of New Cloning Vehicles II. A Multi-Purpose Cloning System", Gene, pp. 95-113 (1977); J. G. Sutcliffe, "pBR322 Restriction Map Derived From the DNA Sequence: Accurate DNA Size Markers Up To 4361 Nucleotide Pairs Long", Nucleic Acids Research, 5, pp. 2721-28 (1978)). Introduction of the DNA product into this position produces a large number of bacterial clones each containing one of the DNA genes or fragments thereof present in the previously produced DNA product. Again, only a small number of these clones will contain the gene for IFN-α or fragments thereof (Fig. 1). The preferred host for the initial cloning according to the invention is E. coli HB 101. Other experiments were carried out with E. coli X1776, a host described in English Patent No. 1,516,458 and deposited in the American Type Culture Collection, Rockville, Maryland, USA, where it is designated by ATCC No. 31244.

1. Preparation of Pstl-cleaved, dGMP-extended pBR322 Plasmid pBR322 (20 µg) was treated with 21 units of Pstl endonuclease ("MRE Porton Downs" or "Nev; England Biolabs") in 150 µl of 10 mM Tris-HCl (pH 7.5), 6 mM MgCl2, 50 mM NaCl, 6 mM 2-mercaptoethanol, 200 mg/pl bovine serum albumin ("BSA")

("Calbiochem") . After 2 hours at 37°C, the mixture was extracted with a phenol-chloroform mixture (1:1) and 1 vol of ether and precipitated with ethanol.

Addition of homopolymeric dGMP tails (Fig. 1) by terminal deoxynucleotidyl transferase (TdT) (purified according to F. J. Bollum, "Deoxynucleotide Polymerizing Enzymes From Calf Thymus Gland", in Methods in Enzymology, (L. Grossman and K. Moldave , eds.), Academic Press, New York, 128, pp. 591-611 (1968)) was carried out in a 328-pl reaction volume containing 100 mM sodium cacodylate (pH 7.2), 10 mM NaH2PO4, 5 mM MgCl2 1 mM dGTP, 50 pg/pl BSA and 3-6 units of TdT (purified as indicated above) per pg DNA. The incubation was carried out at 37°C for 20 min. EDTA was added to 10 mM and the mixture extracted as indicated above and dialyzed for 2 days against TNE buffer.

"2.. Preparation of dCMP-extended DNA

Double-stranded DNA was extended with dCMP residues using standard procedures (eg, Villa-Komaroff et al., supra). 150 ng of the double-stranded cDNA described above was incubated in 8 µl of 100 mM sodium cacodylate (pH 7.2), 2.5 mM CoCl2, 50 µg/µl BSA, 0.1 mM dCTP containing 3-6 units of purified TdT per µg DNA for 8 min. at 27°C and then frozen at -20°C. As previously mentioned, the obtained dCMP-extended DNA is a mixture of different types of which only a few are IFN-related (Fig. 1).

3. Preparation of Ca++-treated E. coli X1776

A single colony of E. coli X1776 was inoculated into 100 ml of tryptone medium (C. Weissmann and W. Boll, "Reduction Of Possible Hazards In The Preparation Of Recombinant Plasmid DNA", Nature, 261, pages 428-29 (1976), supplemented with 100 µg/ml diaminopalmitic acid ("Koch-Light Laboratories"), 10 µg/ml naldixic acid ("Calbiochem") and 10 µg/ml tetracycline ("Ach-romycin", American Cyanamid). The culture was grown at 37°C until an. observed optical density of 0.6 at 650 nm (OD65Q), (measured in a "Beckman DB" spectrometer) and cooled in ice for 30 min. The culture was then sedimented at 4000 rpm in a "Sorvall H4 " swinging cuvette rotor, the cells were washed with 50 ml of 10 mM NaCl, re-pelleted by centrifugation and resuspended in 20 ml of 100 mM CaCl2. The suspension was cooled in ice for 30 min., pelleted by centrifugation and resuspended in 4 ml 100 mM CaCl2 and kept chilled on ice overnight for use E. coli HB101

was prepared for transformation according to the method of M. Mandel and A. Higa, "Calcium-Dependent Bacteriophage DNA Infection", J. Mol. Bio, L. , 53, pp. 159-62 (1970). Aliquots (0.5 ml) were kept frozen at -70°C and they retained their activity for at least 3 months.

4. Union of dGMP-extended pBR322 and dCMP-extended DNA Union of the tailed PstI-cleaved pBR3 22 and tailed cDNA was performed as described by J. Van den Berg et al., "Comparison Of Cloned Rabbit And Mouse 3 ~globin Genes Showing Strong Evolutionary Divergence Of Two

Homologous Pairs Of Introns", Nature, 276, pages 37-44

(1978). 8 ng of the dCMP-extended DNA product was mixed with 22 ng of dGMP-extended PstI-cleaved pBR322 in 50 µl TNE buffer. Incubation was carried out in 4 successive 1 hour increments at 65°C, 46°C, 37°C and 20°C. 20 µl 100 mM Tris-HCl (pH 7.5), 100 mM CaCl 2 , 100 mM MgCl 2 and 50 µl TNE buffer were added and the mixture cooled in ice for 20 min.

The product obtained is naturally a mixture of a large number of different recombinant DNA molecules and some cloning carriers without inserted DNA sequences. However, each recombinant DNA molecule contains a cDNA segment

at the Pstl position. Each such cDNA segment may comprise a gene or a fragment thereof. Only a very few cDNA segments code for IFN or a part thereof (Fig. 1). The bulk encodes one of the other proteins or parts thereof whose mRNAs were part of the poly(A) RNA used

(Fig. 1)'.

5. Transfection of E. coli X1776 with the combined hybrid plasmids

Transfection of E. coli X1776 with the mixture of recombinant DNA molecules was performed as described by J. Van den Berg et al., supra. P3 containment capabilities were used in the transfection process and all subsequent steps in which the resulting transformed bacteria were handled. The connecting pBR322 recombinant DNA molecules were added to 100 µl of Ca<++->treated E. coli X1776, prepared as above, and the mixture cooled in ice for 20 min., warmed to 20°C for 10 min. and 0.6 ml of tryptone medium added. The mixture was spread on 2 tryptone medium agar plates treated as above. The transfection efficiency was 3.3 x 10^ colonies per pg combined pBR322 transfecting DNA, native pBR322 gave 3 x 10<6> colonies per p.g.

As the plasmid pBR322 contains the gene for tetracycline resistance, E. coli hosts that have been transformed with a plasmid containing this gene will grow intact in cultures containing this antibiotic and thus exclude the growth of the bacteria that have not been transformed in this way. Therefore, growth in tetracycline-containing culture will enable the selection of hosts transformed with a recombinant DNA molecule or recycled vector. .After 48 hours at 37°, individual colonies were picked and suspended in 100 µl of tryptone medium (supplemented as above) in the wells of microtiter plates (Dynatech). After overnight incubation at 37°C, 100 µl of 40% glycerol was added to each well. The plates were stored at -20°C and a pool of 100,000 individual clones of transformed E. coli X1776 was prepared.

These 100,000 clones contained a number of recombinant DNA molecules representing complete or partial copies of the mixture of mRNAs in the poly(A) RNA preparations from IFN-induced leukocytes. (Fig. 2). The main part of these will contain only a single recombinant DNA molecule and quite a few of these recombinant DNA molecules are related to IFN. Consequently, the clones must be screened to separate IFN-related clones from the others.

Screening for a clone containing HuIFN acDNA

There are several approaches to isolating bacterial clones containing human leukocyte interferon cDNA ("HuIFN-acDNA"). These include, for example, RNA selection hybridization (Alwine et al., infra), differential1 hybridization (T. P. St. John and R. W. Davis, "Isolation Of Galactose-Inducible DNA Sequences From Saccharomyces Cerevisiae By Differential Plaque Filter Hybridization", Cell, 16, pages 443- 452 (1979) Hoeijmakers et al., infra), hybridization with a synthetic sample (probe) (B. Noyes et al., "Detection And Partial Sequence Analysis Of Gastrin mRNA By Using An Oligodeoxynucleo-tide Probe", Proe. Nati. Acad. Sci. USA, 76, pages 1770-1774 (1979)) or screening of clones producing the desired protein by immunological (L. Villa-Komaroff et al., supra) or biological (A.C.Y. Chang et al., supra) assays In the present case, RNA selection hybridization was considered to be the most suitable and promising me-.

t<p>de for a primary isolation of clones containing IFN-acDNA.

RNA selection hybridization assay

1• Overview of initial analysis

With reference to fig. 2, recombinant DNA was isolated from a culture of a mixture of 512 clones from the above-mentioned collection of clones (two mixtures of 2 clones are shown in Fig. 2) (Step A). The reason for choosing this bet size will be explained in what follows. The recombinant DNA molecules were cleaved, denatured and hybridized to leukocyte poly(A) RNA-containing IFN-amRNA, prepared as previously indicated (Step B). All recombinant DNA molecule-poly(A) RNA hybrids were separated from unhybridized poly(A) RNA (Step C). Poly(A) RNA was recovered from the hybrids and purified (Step D). Recovered RNA was assayed for its IFN-amRNA activity as indicated above (Step E). If, and only if, the mixture of recombinant DNA molecules contains a recombinant DNA molecule with an inserted nucleotide sequence capable of hybridizing to IFN mRNA in poly(A) RNA under stringent hybridization conditions, mRNA released from this hybrid will cause formation of IFN-α in oocytes because mRNA released from any other recombinant DNA molecule-poly(A) RNA hybrid will not be IFN-α related. If a group of 512 clones gave a positive response, the clones were regrouped into 8 portions of 64x and each portion was examined as indicated above. This process was repeated until a single clone responding to this assay was identified.

No assurance is given that recombinant DNA molecules and bacterial clones transformed with them, identified in this way, contain the complete IFN-acDNA sequence or IFN-α or even the DNA sequence that actually encodes IFN-α . However, the recombinant DNA molecules will certainly contain significant nucleotide sequences that are complementary to the IFN-amRNA coding sequence. Therefore, at least one recombinant DNA molecule can be used as a source of a sample to rapidly isolate other recombinant DNA molecules and clones transformed with these to identify further sets. clones that may contain an authentic and complete IFN-α nucleotide coding sequence.

3. Theoretical considerations

The conditions for hybridization (step B) are critical. The absolute concentrations and the ratio between recombinant DNA molecule and poly(A) RNA must be chosen so that the reaction rate and stoichiometry are taken into account. It is difficult to make the right choice because the share of IFN-amRNA

in poly(A) RNA is not known. To ensure controlled and correct kinetics, the hybridization was performed under conditions where the concentration of the DNA sequences from the recombinant DNA molecules was in excess compared to the assumed IFN-amRNA concentration. In a mixture of 512 possible different recombinant DNA molecules, an IFN-a related DNA sequence ("IFN-aR DNA") will either not be present (in a negative analysis result) or will constitute at least 1/512 of the recombinant DNA molecules. Thus, the concentration of the recombinant DNA molecule mixture, and consequently the concentration of IFN-αR DNA, if any, can be adjusted in the hybridization step to ensure appropriate hybridization rates. In addition, the amount of IFN-aR DNA in the reaction mixture must be sufficient to bind enough IFN-amRNA from poly

(A) RNA to enable detection of IFN-α after injection

in oocytes of mRNA recovered from the recombinant DNA molecule-poly(A) RNA hybrid.

To detect IFN-a with the available assays, the concentration should be 100 IU/ml or higher. Because 0.5 ml aliquots are required for duplicate determinations, 50 IU should be generated in the oocytes. Poly(A) RNA from induced previously used leukocytes generates approx. 500 IU IFN-a by injection of 1 pg into oocytes. Therefore, at least 0.1 pg of poly(A) RNA must be injected to generate the required 50 IU. Model experiments with rat globulin mRNA and rat B-globulin cDNA clones showed that the total recovery of I-globulin mRNA in

125

oocytes in relation to I-globulin mRNA added to the hybridization mixture was approx. 10% and recovery of mRNA

the activity was approx. 5%. Therefore, at least 0.1/0.05 = 2 /ag leukocyte poly(A) RNA should be used in the hybridization analysis. In order to achieve a sufficient safety margin, 12 µg poly(A) RNA was used per analysis. To calculate how much DNA from the recombinant DNA molecules is needed for

to bind IFN-amRNA in 12 pg of poly(A) RNA, the IFN-amRNA content in poly(A) RNA was estimated. 1 pg poly(A) RNA generates in 500 IU IF. The specific activity for IFN-a lies between 8 9 2 x 10 and 10 IU/mg protein. 500 IU IFN-a therefore corresponds to between 500/2 x IO<8> = 2.5 x 10~<6> mg (2.5 ng) and 500/10<9> =

5 x 10 ^ mg (0.5 ng) interferon.

The relationship between the amount of IFN-amRNA injected into an oocyte and the amount of IFN-a produced is unknown. In the case of B-globulin mRNA, approx. 30 protein molecules per moles of mRNA per hour. This value is approx. 6 for β-globulin (J.B. Gurdon et al., "Message Stability In Injected Frog Oocytes: Long Life Of Mammalian And B-Globin Messages", J. Mol. Biol., 80, pages 539-51 (1973)). Assuming an average value of 20 for IFN-a and a molecular weight of 18,000 for IFN-a and a molecular weight of 330,000 for IFN-amRNA, 26 mg (18,000/ 3,330,000 x 20 x 24) of IFN-a should be produced in 24 t per mg injected IFN-amRNA. If the specific activity for IFN-α is 2 x 10<8>/mg (2 x IO<2> IU/ng). then 1 ng of IFN-amRNA will give 26 x 2 x 10<2>=

3 9 5.2 x 10 IU IFN-a. If the specific activity is 10 /mg 3 4 (10 IU/ng); the amount of IFN-a produced will be 2.6 x 10 IU. Because 1 pg of leukocyte poly(A) RNA gives 500 IU of IFN-a under the above assumed conditions/the concentration of IFN-amRNA in 1 pg of poly(A) RNA should be in the range of 0.1-0.02 ng and the proportion of IFN -amRNA in leukocyte poly(A) RNA lie in the range 1:10,000 and 1:50,000. Therefore, 12 pg of poly(A) RNA contains 0.2-1.2 ng of IFN-amRNA.

Should the translation ratio of IFN-amRNA in the oocytes be one order of magnitude lower than the average for a-globulin mRNA, the IFN-amRNA content in poly(A) RNA will be approx. 10 times higher than that calculated above or be between 1:1000 to 1:5000. In this case, 12 pg of poly(A) RNA would contain 2-12 ng of IFN-amRNA. On the other hand^vis ' the translas ion ratio. for IFN-amRNA in the oocytes should be one order of magnitude lower than what is usual for globulin mRNA, the IFN-amRNA content in poly(A) RNA would be 10 times lower than that calculated above or in the range of 1:100,000 and 1:500,000. 12 pg poly(A) RNA will then contain 0.02-0.1 ng IFN-amRNA.

Plasmid pBR322 has 4361 b.p. The complete cDNA in IFN-amRNA will add approx. 800-1000 b.p. to pBR322 by generating pBR322-IFN-acDNA to a total in the range of 5200-5400 b.p. The molecular weight will thus be 12 times (2 x 5200/800) that of IFN-amRNA alone. Therefore, to ensure IFN-amRNA as above calculated to be present in the 12 pg of poly(A) RNA required for the assay^ an amount of recombinant DNA molecules equal to 12 times the amount of IFN-amRNA (stoichiometric amount ).

Because the IFN-amRNA content of the poly(A) RNA used for the preparation of recombinant DNA molecules was increased 10-40 times over that of the starting poly(A) RNA, the pool of 512 clones should contain 10-40 times more clones containing the desired IFN-amRNA than is calculated from the above. If IFN-amRNA is 1 part in 1000 parts of starting poly{A) RNA then 12 pg poly(A) RNA will contain 12 ng IFN-amRNA and the stoichiometric amount of IFN-acDNA plasmid is 144 ng. As each group of 512 clones will contain at least 5 with IFN-acDNA inserts, the total amount of hybrid plasmid DNA will require 14.8 pg (144 x 512 /5 x 10~<3>). If IFN-amRNA constitutes 1 part in 10,000/, 12 pg of poly(A) RNA will contain 1.2 ng of IFN-amRNA and the amount of IFN-acDNA plasmid required will be 14.4 ng. A pool of 512 clones will contain either 0 or 1 IFN-acDNA insert, so the total amount of hybrid plasmid DNA required is 7.4 pg (14.4 x 512 x 10 ). If IFN-amRNA is 1 part in 100,000, the total amount of hybrid plas--3

mid DNA required be 0.74 pg (1.44 x 512 x 10 ). To ensure that the hybridization reaction will proceed under DNA excess conditions (ie, excess recombinant DNA compared to poly(A) RNA), 20 µg of the mixture (approximately 1.4- to 30-fold excess) was selected for the assay.

The hybridization must be carried out under conditions that ensure (a) that the hybridized part of poly(A) RNA is recovered intact and

in a biologically active form, (b) that non-specific DNA-mRNA association is prevented and (c) that the hybridization reaction proceeds to at least 75% completeness. These conditions will most likely be satisfied by hybridization in 80% formamide, 0.4M NaCl (J. Casey and N. Davidson, "Rates Of Formation And Thermal Stability Of RNA:DNA And DNA:DNA Duplexes At High Concentrations Of Formamide" , Nucleic Acids Res., 4, pages 1539-52 (1977)). In this solution, the hybridization can be carried out at approx. 40°C instead of 60-70°C which is necessary when formamide is omitted. Lower temperatures are preferred to minimize destruction of poly(A) RNA. The chosen hybridization temperature was 56°C. This is approx. 3° lower than T1 . (J. Casey and N, Davidson, supra) and approx. 10-13 C below T^/2d (Hama9ucni & Geidushek, J. Amer. Chem. Soc., 84, page 1329). This temperature should therefore not allow hybridization of sequences with less than approx. 87%'s homology, as 1%'s mismatch lowers T-[./2d mec* l°c (T' F* Bonner et al., "Reduction In The Rate Of DNA Reassociation By Sequence Divergence", J. Mol .Biol. 81, pages 123-35

(1973)).

In the present hybridization, self-hybridization of DNA is not a major problem because the mixture of DNAs used consists of the same vector (pBR322) and a number of cDNA inserts. Therefore, most DNA sequences will be heterodup primers in which the inserts are available for hybridization to poly(A) RNA. It is unlikely that complementary cDNA inserts that form part of the different duplexes will co-react as a result of topographical constraints. In all cases, DNA:DNA reassociation will be minimized under the reaction conditions used (J. Casey and N. Davidson, supra).

To determine the necessary hybridization time to ensure at least 75% conversion, a second-order reaction equation was used:

in which:

R = the molar nucleotide concentration in hybridized

RNA

Co= the molar nucleotide concentration of the initial

DNA to be hybridized

Ro= the molar nucleotide concentration of the initial

RNA to be hybridized

kR= turnover rate constant for RNA-DNA hybridization

t time (f)

and :

0.7 5 (75% turnover)

kR 472 (kR = 1/12 k^ (J. Casey and N. Davidson, supra)

where: k^ = second-order reaction constant for DNA under the selected hybridization conditions

and: kd = 1.7 x IO<5> x L1//2 x N_1

(J. R. Hutton and J. G. Wetmur, "Renaturation Of Bacteriophage 8 X17 4 DNA-RNA Hybrid: RNA Length

Effect And Nucleation Rate Constant", J. Mol. Biol., 77, pages 495-500 (1973))

L 900 (chain length in b.p., approx. 900

is present in the complete IFN-acDNA insert)

N 900 (complexity in b.p. of the hybrid chain. Here the complexity is 900 because the 900 nucleotides of IFN-amRNA are connected with complementary 900 nucleotides in the IFN-acDNA insert)

-7

Co = 2.5 x 10 (Based on 40 pl solution containing the previously determined

35

agreed 20 µg of recombinant DNA molecules to be used in the analysis, again under the assumption that the IFN-acDNA insert will be 1/12 of the recombinant DNA molecule and will occur at least in 1 of 512 clones and indicate 662 as the average molecular weight of a DNA base pair)

Ro = 8.7 x 10 — 8 (Based on 40 µl of solution containing the previously determined 12 µg of poly(A) RNA used in the assay, again assuming that poly(A) RNA contains 1:10,000 parts of IFN-amRNA ( assuming that in the large excess of DNA, a different proportion will have little effect on the hybridization rate) and assuming 3 43 as the average molecular weight of a ribonucleotide in RNA)

3. Performing the initial analysis

Step A-Preparation and Cleavage of the Mixture of Recombinant DNA Molecules

The desired number of bacterial clones was inoculated onto tryptophan medium agar plates supplemented as indicated above by transferring an aliquot from each microtiter well using a mechanical device. After incubation at 37°C, each clone had given rise to a colony several millimeters in diameter. All colonies were washed off the plate(s) and pooled to give an inoculum used to inoculate 1 L of tryptophan medium, supplemented as indicated above, into a 2 L Erlenmyer flask. The culture was shaken at 37°C to an observed 0D,C- of approx. 0.8 (determined visually). A volume of supplemented tryptophan medium and chloramphenicol to 170 µg/ml was added to the culture which was further shaken at 37°C for 16 hours.

20 ml of chloroform was added and the culture again shaken for 10 min. at 37°C to kill the bacteria (C. Weissmann and W. Boll, supra). The culture was decanted from the chloroform and the cells collected by centrifugation ("Sorvall GS3" rotor)

for 15 min. at 6000 rpm. at 4°C. About. 1-2 g of cells were obtained with each liter of the preparation. The cells were suspended in 30 ml of 20 mM Tris-HCl (pH 7.5), centrifuged for 20 min. at 5000 rpm. and 4°C ("Sorvall SW" rotor) and resuspended in 30 ml of 50 mM Tris-HCl (pH 7.5). 0.25 vol of lysozyme solution (10 mg/ml in 50 mM Tris-HCl (pH 7.5)) was added, and after cooling for 10 min at 0°C, 0.33 vol (calculated on the volume of the original 50 mM Tris-HCl culture suspension) 0.5 M EDTA (pH 8.0) and gently mixed in without shaking. After another 10 min. at 0°C, 1/16 vol (again calculated on the original volume) of 2% "Triton X-100" was added. After 60 min. the sample was centrifuged for 60 min. at 10,000 rpm. at 0°C in a "Sorvall SW" rotor) The supernatant liquid was transferred to a beaker containing a magnetic stirrer and 3M NaOH was added with stirring until a pH of 12.5 was reached, measured at 20°C using a glass electrode and an "Orion Research" model 601 pH meter, standardized with Beckman pH 10 Carbonate Buffer Standard (No. 3505). After stirring for 10 min. at 20 oC the pH was adjusted to 8.5. After 3 min. further stirring, 1/9 vol 5 M NaCl and 1 vol phenol (distilled and brought into equilibrium with 9.5 M NaCl) were added and vigorously stirred for 5 min. The phases were separated by centrifugation ("GSA Sorvall" rotor) at 10,000 rpm. at 0°C for 10 min. The supernatant containing Form I DNA (circular double-stranded DNA) was carefully removed from the intermediate phase (containing single-stranded DNA) and extracted 3 times with chloroform. (Phenol must be largely removed at this step). The Form I DNA fraction will contain the recombinant DNA molecules (pBR322 cDNA insert) that were originally used to transform the host cells that formed part of the 512 clones selected for investigation.

Pancreatic RNAase A (5 mg/ml, preheated 10 min. at 85°C) was added to Form I DNA to a concentration of 20 µg/ml, and the mixture incubated at 60 min. at 37°C. 1/5 vol 5 M

NaCl was added and the mixture adjusted with 30% polyethylene

glycol 6000 ("Union Carbide", autoclaved for 20 min. at 120 oC) up to a final concentration of 7.5% PEG. After 2-16 h at -10°C, the precipitate was collected in a "Sorvall SW"

rotor for 20 min. at 8,000 rpm. at 0°C, dissolved in 0.075 M

NaCl, 0.0075 M Na citrate to an absorbance of 20 at 260 nm

and adjusted to 0.5% SDS. The solution was incubated for 30 min.

at 37°C with 0.5 mg/ml pronase (self-"digested" at 20 mg/ml for 2 d at 37 C) and extracted 3 times with 1 vol of distilled phenol and 2 times with 1 vol of chloroform. The sample (up to 2 ml of a 1 mg/ml DNA solution) was centrifuged through a 5% to 23% sucrose gradient in 50 mM Tris-HCl (pH 7.5)

1 mM EDTA for 15 h at 21,000 rpm. at 15°C during use

else of a "SW 27 Beckman Rotor". The fractions were collected and the OD260 was observed. DNA-containing fractions were pooled and DNA precipitated with sodium acetate and ethanol. 20-100 pg of the Form I DNA mixture was recovered by the centrifugation.

20 µg of purified Form I DNA was treated in 150 µl of 10 mM Tris-

HCl (pH 7.5), 6 mM MgCl, 50 mM NaCl, 6 mM 2-mercaptoethanol,

200 pg/ml BSA or gelatin and 20 units HindIII (New England Biolabs). The Hindlll restriction enzyme cleaves Form

1 DNA at the position inside the pBR322 unit (it is unlikely

equal to that the cDNA unit is also cleaved, but if it is cleaved, the determination will not be affected to a significant extent). After 2 h at 37°C, an aliquot (1%) was analyzed by electrophoresis through a 1% agarose gel in 50 mM Tris-acetate (pH 7.8),

2 mM EDTA oxygen at 50 mA to determine if cleavage was complete. If this was not the case, more HindIII was added and the incubation continued for 2 hours. When

Form I DNA was completely converted into linear molecules,

pronase (Calbiochem), EDTA and SDS were added to 0.5 mg/ml respectively,

10 mM and 0.5%. After 30 min. at 37°C the solution was extracted with 30 µl phenol-chloroform (1:1). The organic phase was washed with 50 µl of 20 mM Tris-HCl (pH 7.5) 1 mM EDTA and the combined aqueous phases extracted 3 times with ether,

filtered through a 0.1-ml "Chelex" column, collected in an EDTA-boiled "Pyrex" tube and precipitated with 1/10 vol 3M sodium

acetate and 2.5 vol ethanol. After a reprieve overnight at o

-20 C, DNA was collected by centrifugation.

Step B - Hybridization of DNA with poly(A)

RNA

Two hybridization mixtures were prepared.

Mixture I contained 4 µl of 10x concentrated hybridization buffer (4M NaCl, 0.1 PIPES (pH 6.4, 1.4 piperazine-dietane-125 sulfonic acid, "Sigma"), 50 mM EDTA, 0.5 µl (ca 5 ng of I-globin mRNA (5000 cpm) and 6 µl of induced leukocyte poly(A) RNA (2 pg/µl) in an amount sufficient to generate 6000 IU of IFN upon injection of oocytes.

Mixture II contained 10 µg of the HindIII-cleaved Form I DNA, as prepared above, and 0.1 µg of the PstI-cleaved Z-pBR322-(H3)/RcBG-4,13 (a pBR322 derivative containing B-globulin- the sequence in the HindIII position) (Mantei et al., "Rabbit 3-globin mRNA Production In Mouse L Cells Transformed With Cloned Rabbit 3-globin Chromosomal DNA", Nature 281, pages 40-46

125

(1979)). I-globulin mRNA in mixture I and B-globulin

DNA in mixture II serves as internal positive controls for the hybridization assay. Both mixtures were dried in a vapor of nitrogen gas. 40 µl of 80% formamide was added to the residue of mixture II and the solution was denatured for 10 min. at 100°C and cooled rapidly in ice. The denatured solution was used to dissolve the residue from mixture I and the resulting solution incubated at 56°C for 4 hours.

Step C - Separation of Hybridized Poly(A)

RNA-DNA from unhybridized poly(A) RNA

After dilution to 1 ml with cold 0.9 M NaCl, 0.09 M Na citrate and formamide (100%) to 4% v/v, the solution was filtered at 0.5 ml/min. through a "Millipore" filter (pore size 0.45 pm), the filter having previously been examined with respect to its ability to retain RNA-DNA hybrids, because not all filters obtained from the manufacturer were equally effective.

Step D - Purification of Hybridized Poly(A)

RNA

The above-mentioned filter with associated poly(A) RNA hybrids was immersed in 1 ml of 0.15 mM NaCl, 0.015 M Na citrate, 0.5% SDS for 10 min. at 37°C, cleared with 50 mM Tris-HCl (pH 7.5), 10 mM MgCl2, 2 mM CaCl-, and placed in 0.6 ml fresh buffer.

After addition of 5 µl iodoacetate-treated DNAase (5mg/ml)

(S. B. Zimmermann and G. Sandeen, Anal. Biochem., 14, p. 269

(1966); P. A. Price et al., "Alkylation Of A Histidine Residue At The Active Site Of Bovine Pancreativ Deoxyribonuclease", J. Biol. Chem., 244, pages 924-32 (1969)), the filter was incubated at 37°C for 10 min.

The filter was removed and the solution extracted with 1 vol of phenol and 1 vol of ether and passed through a 0.1 ml "Chelex" column. 5 µg of carrier RNA (purified yeast RNA) was then added to the solution and the RNA precipitated with sodium acetate and ethanol. The precipitate was collected by centrifugation at 10,000 G, dissolved in 100 µl 1 mM EDTA, heated for 90 s at 100°C and TNE and SDS were added accordingly

add 2 times TNE and 0.5% SDS. RNA was absorbed onto a 100 µl oligo(dT) cellulose column, eluted with 4 washes of 0.3

ml of distilled water and precipitated with sodium acetate and ethanol. After 16 h at -20°C, the precipitated RNA was separated by centrifugation and dissolved in 2 µl TNK buffer.

Step E - Determination of IFN-amRNA activity

The poly(A) RNA solution obtained from the previous step was injected into 40 oocytes (approximately 50 ni per oocyte). The oocytes were incubated at 23°C for 24-48 h, homogenized and centrifuged (or the incubation medium was recovered) and analyzed as described previously for IFN-α.

4. Subsequent determination - hybridization to filter-bound DNA

Most subsequent analyzes of a recombinant DNA molecule from a single clone were performed with DBM or DPT paper-bound DNA because the analysis conditions were no longer critical and the determination more suitable. DPT paper produced lower background noise and was preferably used. D3M paper was prepared as described by J. C. Alwine et al., "Method For Detection Of Specific RNAs In Agarose Gels By Transfer To Diazobenzyl Oxymethyl-Paper And Hybridization With DNA Probes", Proe. Nati. Aeaad. Pollock. USA, 14, pages 5350-54 (1977) APT paper was prepared by the method of B. Seed (personal communication): Sheets of "Whatman 540" paper (20 g) were stirred for 16 t at 20 oC with a mixture of 70 ml of 0.5 M NaOH, 2 mg/ml NaBH^ and 30 ml of 1,4 butanediol diglycidyl ether. The paper was then transferred to a solution of 10 mL of 2-aminothiophenol in 40 mL of acetone and stirred for 10 h. The paper was thoroughly washed with acetone, 0.1 N HCl, H 2 O, 0.1 N HCl, H 2 O and dried. APT paper was diazotized to DPT paper as described for conversion of ABM to DBM paper (Alwine et al., supra).

2

DNA (up to 15 pg) was bound to 50 mm diazotized ABM (DBM) or diazotized APT (DPT) paper as described by J.H.J. Hoeijmakers et al. "The Isolation Of Plasmids Containing DNA Complementary To Messenger RNA For Variant Surface Glycoproteins Of Trypanosoma Brucei", Gene, in press 1980), as noted hereinafter.

Hybrid plasmid DNA was treated with endonuclease Pstl, treated with 500 pg pronase per ml, 0.5% SDS and 10 mM EDTA for 30 min. at 37°C, extracted with phenol and ether, passed through a 0.1 ml "Chelex" column and precipitated with ethanol. The heat-denatured DNA (up to 5 pg with a small amount

3 2

labeled P-DNA added as marker) was incubated overnight at 0°C with 1 cm 2 DBM or DPT paper in 200 µl 25 mM potassium phosphate buffer (pH 6.5). The filters were washed three times for 5 min. at room temperature in 50 mM potassium phosphate buffer (pH 6.5) 1% glycine and three times with 99% recrystallized formamide.

A further incubation with 99% formamide for 2 min. at 68°C was followed by three washes in 50 mM potassium phosphate buffer (pH 6.5) at 20°C and two washes in 0.4 M NaOH at 37°C for 10 min. About. 40-60% of the radioactivity was retained in the filters. These filters were incubated for 3 hours at 38°C in the prehybridization medium A, which had been supplemented with 1% glycine, using a\, 330 µl per filter. Medium A contains 50% formamide, 5 x SSC, 0.04% polyvinylpyrrolidone, 0.04% "Ficoll"

(Pharmacia), 0.1% SDS, 25 pg poly(A) (P & L) and 100 pg yeast RNA (BDH, extracted six times with phenol and precipitated with ethanol). The filters were washed twice in medium A and then hybridized for 16 h at 38°C with poly(A) RNA as indicated (typically 5-8 pg) in medium A under paraffin oil. RNA was added as follows: a wet DNA filter was pressed and introduced into a sterile Petri dish, 20-40 µl of the RNA solution was pipetted onto the filter and a second DNA filter (either a duplicate or a control) was placed on top of the first, and the two overlying filters were covered with sterile paraffin oil. After hybridization, the filters were successively washed in medium A (2 times) in a solution containing 1 x SSC, 0.2% SDS, 1 mM EDTA (3 times 10 min. at 20°C each time), medium A ( 2 h at 38°C) and in 50% formamide, 5 x SSC, 0.1% SDS (3 times, 10 min. at 20°C). Hybridized RNA was eluted by heating for 1 min. at 100°C in 200 µl 10 mM Tris-HCl (pH 7.4), 1 mM EDTA and 0.1% SDS. The elution step was repeated twice and the eluates combined and the RNA precipitated with ethanol after addition of 2 pg of yeast RNA (purified as indicated above). The washed pellet was vacuum dried, dissolved in 3 µl H 2 O and injected into oocytes. The IFN-α activity was determined as indicated above.

5. Results of RNA Selection Hybridization Assay The determinations from 8 groups of 512 clones (i.e. groups T, Y, j, K, • , 0, <£ and rz were negative. The determinations from 4 groups of 512 clones (i.e. groups 1, 6, N and \) were positive but not consistent. The positive results are reported as follows: IU/ml IFN-α produced by RNA released from poly(A) RNA-DNA hybrid (determination from control hybridization using Z -pBR322(H3)/Rc3 G-4,13, supra), the determinations where the experimental results were higher than the background control are underlined.

Group A was divided into 8 groups of 6 4 clones and the hybridization analyzed as before. The subgroups produced the following results shown in the same format as above:

Subgroups X-III were divided into 8 sets of 8 clones and hybridized and analyzed:.

Because the first positive results were obtained with the set A-III-4, the individual colonies in this set (designated A-H) were hybridized and analyzed:

Therefore, clone X-III-4-C contains a recombinant molecule capable of hybridizing IFN-amRNA.

The recombinant DNA molecule in this clone is designated: Z-pBR322(Pst)/HcIF-4C ("Hif-4C"), and the bacterial strain containing it: E. coli X1776 (Z-pBR322 (Pst)/HcIF-4C) ("E. coli Hif-4C"). This nomenclature indicates that the recombinant DNA molecule originates from Zurich (Z) and the plasmid pBR322 contains at the Pstl position a HIFN acDNA ("HcIF"). The particular recombinant DNA molecule is derived from clone X- -III-4-C ( ("4C").

CLONING AND CHARACTERIZATION OF Z- pBR322( Pst)/ HcIF- 4C

As primary clones of transformed cells often contain more than one type of recombinant DNA molecule (Efstratiadis et al., "The Primary Structure Of Rabbit Ø-globin mRNA As Determined From Cloned DNA", Cell, 10, pages 571-85 (1977)) , Hif-4C was isolated from E. coli X1776 (Hif-4C) clones and purified as the DBM paper method was used in this assay. described above. Samples of Hit-4C and pBR322 were treated and digested with Pst I and analyzed by electrophoresis on a 1% agarose gel. Hif-4C gave two bands, one with the mobility for the Pst-cleaved pBR322 and the other with the mobility corresponding to approx. 320 b.p.

E. coli HB101 was transformed with the isolated Hif-4C as described above. Six clones of tetracycline-resistant transformed bacteria were picked and small cultures prepared and form I DNA purified and analyzed by Pstl digestion and agarose gel electrophoresed as indicated. All samples showed cleavage patterns identical to Hif-4C. One of these re-cloned recombinant DNA molecules was designated Z-pBR322 (Pst)/HcIF-4c ("Hif-4c") and used for further experiments. Small "c" denotes a cloned DNA molecule.

To determine the capacity of Hif-4c and its cDNA insert to hybridize to IFN-amRNA, Hif-4c (115 µg) was completely digested with 125 units of Pstl, extracted with phenol and chloroform, and precipitated with ethanol as described above. An aliquot (10 pg) was 5' terminally labeled

(to serve as marker in subsequent steps) by dissolving it in 100 µl of 50 mM Tris-HCl (pH 7.5), passing the solution through a 0.1 ml "Chelex 100" column and treating it with 0.6 units of bacterial alkaline phosphatase in 1 tifie at6 5°C. Ten times concentrated TNE (40 µl) was added and the solution extracted 3 times with 1 vol of phenol and 3 times with 1 vol of chloroform. DNA was precipitated with 2 vol ethanol at -20°C overnight and collected by centrifugation. For further purification, a sample of 0.5 ml of TNA was adsorbed on 0.25 ml of DEAE cellulose ("Whatman DE52", prewashed with 2 ml of 150 mM NaCl, 50 mM Tris-HCl (pH 7.5), 2 mM EDTA) ("NET buffer"), washed with 2 ml NET buffer, eluted with 0.4 ml 1.5 M NaCl,

20 mM Tris-HCl (pH 7.5), 2 mM EDTA and precipitated with ethanol as indicated previously. The DNA obtained was incubated with γ-<32>P-ATP (specific activity about 5000 Ci/mmol) and polynucleotide kinase, (A.M. Maxam and W. Gilbert, "A New Method For Sequencing DNA", Proe. Nat. Acad. Sci. USA, 74, pages 560-564 (1977)) and purified by chromatography on 3 ml "Sephadex-

G50" column in TNE. The eluted fractions were pooled

32

and P-DNA pr7 precipitated with ethanol as previously indicated. yield approx. 10 ppm.

The unlabeled PstI-digested Hif-4c DNA (90 µg) was mixed

5 32

with 6 x 10 dpm P-labeled, Pstl-cleaved Hif-4c DNA as obtained above and electrophoresed through a 10 x 20 x 0.7 cm, 2% horizontal agarose gel in 50 mM Tris-acetate buffer (pH 7.8) using a gap of 2.5 cm. An x-ray film was exposed to the gel and the position of the 320-bp fragment determined. The gel strip containing the radioactive band (1.3 x 10 5 dpm) was excised, crushed by pressing through a plastic 2 ml injection syringe and extracted overnight at 4°C by stirring with 10 times the volume of the gel with NET buffer. DNA was adsorbed on a 0.1 ml hydroxyapatite column (prewashed with 1 ml NET buffer). The column was washed with 1 ml of 0.1 M K-phosphate buffer (pH 7.5) and DNA eluted with 0.2 ml of 1 M K-phosphate buffer (pH 7.5). The eluate was diluted 10-fold with sterile, distilled H 2 O and DNA adsorbed to and eluted from DEAE and precipitated with ethanol as described previously. This DNA was designated "Hif-4c fragment".

This Hif-4c fragment (120 ng) was bound to DPT paper

(0.5 x 0.5 cm) as previously described. As a control, 120 ng of B-globulin cDNA fragment was excited with HindIII from the hybrid plasmid Z-pBR322(H3)RcBG-4,13 (F. Meyer et al., "Transposition Of AT-linked, Cloned DNA From One Vector. To Another ", Experimentia, 35, page 972 (1979), N. Mantel et al., supra) and treated in the same way. The hybridization of the duplicate filter to poly(A) RNA (in 20 µl), washing of the filters, and recovery of RNA from the filters were performed as previously described. After injection into oocytes, the following IFN-α activities were detected:

Thus, it can be seen that Hif-4c contains an insert capable of hybridizing to IFN-amRNA.

IDENTIFICATION OF CLONES OF E. COLI CONTAINING RECOMBINANT DNA MOLECULES CROSS-HYBRIDIZED TO THE HIF-4C INSERTION

When the cDNA insert in the recombinant DNA molecule Hif-4c was only approx. 320 b.p., or one third of the estimated size for IFN-amRNA, the purified Hif-4c fragment described above was used as a sample to isolate bacterial clones containing recombinant DNA molecules containing related hybrid DNA inserts (Fig. 3 ).

64 bacterial clones constituting subgroup Å.-III described above, were pressed onto a "Millipore" membrane (diameter 8 cm), placed on an agar plate (supplemented with diaminopimelic acid, nalidixic acid and tetracycline, as previously indicated) and incubated for 24 hours at 37°C. The filter was placed on a 0.75 ml drop of 0.5 M NaOH and after 2-3 min was transferred to a paper towel to remove excess liquid, after which the step was repeated. The filter was neutralized using 1 M Tris-HCl (pH 7.5) and washed with 1.5 M NaCl - 0.5 M Tris-HCl (pH 7.4) in the same manner as above and air-dried. The filter was dipped in 0.3 M NaCl, air dried and

heated to 80 C for 2 h under vacuum. The Hif-4c Pst fragment

32

(30 ng) was P-tagged by "nick" translation (A.J. Jeffreys and R.A. Flavell, "The Rabbit 3-Globin Gene Contains A Large Insert In The Coding Sequence", Cell, 12, pages 1097-1108

32 32

(1977)) using β-β dATP and α-β dCTP (specific activity, 40 Ci/mmol each). The filter carrying the A-III colonies was prehybridized in 4 x SETT (SETTer 0.15 M NaCl, 30 mM Tris-HCl (pH 8.0), 1 mM EDTA), 0.1% (w/v) "Ficoll" , 0.1% polyvinylpyrrolidine, 0.1% (w/v) BSA, 0.5% SDS and 200 µg/ml denatured, fragmented salmon sperm DNA for 7 h at 68°C and

5 32

hybridized with 2 x 10 cpm P-labeled Hif-4c in 4 x SET 0.02% (w/v) "Ficoll", 0.02% polyvinylpyrrolidine, 0.02% w/v BSA, 0.5% SDS and 200 pg/ml denatured salmon sperm DNA at 68°C for 16 h. The filter was washed with SETT-0.5% SDS at room temperature, washed 2 x SETT-0.5% SDS for 5 h at 68°C, the solution yielded a once, and with 3 mM "Trizma" base at room temperature for 4 thaws, replacing the solution once. After drying the filter, an x-ray film was exposed to the filter for 8 tirres using one layer. Three colonies gave a strong positive response, namely A.-III-7D, \-III-2H and A-III-4C and two gave a weak response, namely A. -IH-IE, X-III-3D.

Small cultures were prepared from Hif-4c related clones, form I DNA was purified, digested with PstI and analyzed by agarose gel electrophoresis as previously described. All form I DNAs gave a large fragment (plasmid pBR322 core) and a small fragment (hybrid insert). The recombinant DNA molecule from X.-III-2H released the largest deposit, namely approx. 900 b.p. This recombinant DNA molecule was designated Z-pBR322(Pst)/ HcIF-2H ("Hif-2H") and its insert "Hif-2H fragment".

Hif-2H was examined for its ability to bind IFN-amRNA by binding it to DPT paper (4 pgAm 2 ) and then hybridizing it to poly(A) RNA (0.3 µg/µl), performed as previously described, for 16 h and then determine the IFN-amRNA activity.

In a further experiment, a further set of E. coli clones containing recombinant DNA molecules was prepared and the colonies hybridised until labeled Hif-4c fragments were identified. To ensure a high yield of plasmids with long

32 cDNA inserts were part of the double-stranded P-labeled leukocyte cDNA prepared enzymatically from leukocyte poly(A) RNA (the same cDNA preparation prepared as previously indicated) was size fractionated by centrifugation through a sucrose density gradient using the same procedure as described for centrifugation of poly(A) RNA. The fractions containing cDNA with a sedimentation rate corresponding to a 600 b.p. DNA fragment or larger was pooled and cDNA recovered after ethanol precipitation. The cDNA was extended with dCMP residues, hybridized to dGMP-extended Pstl cleaved pBR322 and the resulting hybrid DNA used to transform E. coli as previously indicated, except that E. coli HB101 was used. The bacteria were distributed on 8 cm diameter "Millipore" filters placed on "Tryptone" medium agar plates (containing 10 pg/ml tetracycline) and grown until small colonies appeared. A duplicate filter was prepared by pressing a new, moist "Millipore" filter onto the colony-bearing filter, after which the filter was pulled off, placed face up on an agar plate containing 4.4%

glycerol and incubated until small colonies appeared. This colony-bearing filter was covered with an additional "Millipore" filter, frozen at -55°C and stored (D. Hanahan and M. Meselson, "A Protocol For High Density Plasmid Screening", Sept. 1978, personal communication. A total of eighteen filters with a total of approximately 5,000 colonies were prepared.A replica of each filter was

used for hybridization to P-labeled, Psjtl-treated Hif-4c DNA fragment, in the same manner as described above.

About. 185 positive colonies were identified on an autoradiography, recloned on the "Millipore" filter and identified once again by hybridization. - 95 clones that gave the strongest hybridization response were designated Z-pBR322(Pst)/HcIF-SN1 to SN95 and used for further investigations. It is of course obvious that when Hif-2h has the ability to produce a polypeptide that exhibits an immunological or biological effect of HuIFN (infra), then this Hif-2h and other DNA sequences related to it, for example Hif-4c, will be able is used in this method for clone screening equally well on other clones containing DNA sequences obtained by recombinant DNA technology, syntheses, natural sources or a combination thereof or clones containing DNA sequences related to any of the above-mentioned DNA sequences by mutation including single or multiple base substitutions, insertions, inversions or deletions to select other DNA sequences and clones that also encode HuIFN. Therefore, such DNA sequences and their identification will fall within the scope of the present invention. (For example infra). It must also be understood that DNA sequences which are not isolated by means of the above-mentioned DNA sequences, but which nevertheless result from their rearrangement of the nucleotide code for the expression of polypeptides which code by means of expression of the above-mentioned DNA sequences , will fall within the scope of the invention.

Further characterization of the Hif-2h DNA insert

As described above, the recombinant DNA molecule Hif-2h contains a deposit of approx. 900 b.p. and hybridizes to human leukocyte interferon mRNA. The following additional characteristics were determined.

1. Hybrid arrested translation

If an mRNA hybridizes to a cloned, complementary cDNA, the translation of the mRNA is prevented. However, heat-denaturation of the hybrid releases translatable mRNA (B. M. Paterson et al., "Structural Gene Identification And Mapping By DNA-mRNA Hybrid-Arrcstcd Col1-Free■ Translation", Proe. Nati. Acad. Sei. USA, 74, pages 4370 -74 (1977)). 2.2 µg _s_fcl-cleaved Hif-2h (and as a control 2 µg HindIII-cleaved Z-pBR322(H3)/RcSG-4.13 ("Rc3G")) were denatured in 10 µl 80% (vol/vol ) deionized formamide - 20 mM "PIPES" buffer (pH 6.4) for 10 min at 80° C. The solution was added to an Eppendorf tube into which leukocyte poly(A) RNA (5 pg), NaCl (4 pmol) oq LDTA (10 nmol) was dried. The mixture was heated for 7 h at 4 8°C under a layer of paraffin oil, cooled and diluted with 200 pl H 2 O. The two samples were each divided into two equal parts, one of h<y> of these were heated to 100°C for 30 s. The nucleic acids were precipitated with ethanol, dissolved in 3 µl H^O and examined for IFN-amRNA activity in oocytes as indicated above:

Therefore, Hif-2h, when hybridized with poly(A) RNA, inhibited the translation of IFN-amRNA in poly(A) RNA, and after denaturation of the hybrid, IFN-amRNA became translatable again. This experiment further confirms that Hif-2h contains sequences complementary to IFN-amRNA.

2. Analysis using restriction enzyme digestion and determination of nucleotide/amino acid sequences and restriction maps

Digestion of Hif-2h with various restriction enzymes (New England Biolab) was performed and the resulting products analyzed by agarose gel electrophoresis. The underlined fragments are not common to pBR322 and Hif-2h:

32

In addition, the 5' terminated P-tagged Pstl cleaved Hif-2h was additionally cleaved with several restriction enzymes and the size of the radioactive fragments derived from the cDNA insert as the recombinant DNA molecule was determined.

On the basis of these data, a restriction map for Hif-2h was deduced (Fig. 4). The map is not definitive with respect to the absolute location with respect to the positions and may be incomplete with respect to the MbolI positions within the deposit. Only the positions closest to the insert are given inside the pBR322 core. The arrow indicates the orientation of the IFN-acDNA coding chain.

Although the current structure of the Hif-2h fragment or other insertions in the clones according to the invention or for the amino acid sequence or the structure of the polypeptides encoded therefrom is not necessary for a person skilled in the art to practice the invention, the above-mentioned data and the restriction map were included in the original description as the best available information regarding the structure of the fragment at the time of filing the original application. Since that time is as expected (supra, p 9, lines 27-29) these data and the restriction map of the Hif-2h fragment improved using well-known techniques, for nucleotide sequencing and restriction analysis. See, for example, A. M. Maxam and W. Gilbert, "A New Method For Sequencing DNA", Proe. Nati. Acad. Pollock. USA, 74, pp. 560-64 (1977). Plasmid DNA was prepared by Method B (N. M. Wilkie, et al., "Hybrid Plasmids Containing An Active Thymidine Kinase Gene Of Herpes Simplex Virus 1", Nucleic Acids Research, 7, pages 859-77

(1979)) and digested by various restriction enzymes essentially according to the supplier, except that 200 µg/ml gelatin replaced the C-serum albumin in the enzyme buffers.

(EcoRI was a gift from W. Boll, Bspl a gift from A. Kiss and other enzymes were obtained from New England Biolabs).

Digested DNA (20 pg) was extracted with phenol, precipitated with ethanol, dissolved in 0.05 M Tris-HCl (pH 8) and passed through a small column "Chelex-100". Fragments with straight or 5<* >"sticky" ends were dephosphorylated by treatment with 0.2 units of calf digestive alkaline phosphatase (Boehringer) per pmol DNA 5' ends in 200 µl 0.05 M Tris-HCl (pH 8) for 60 min at 37°C. The enzyme was inactivated by heating for 60 min at 65°C. For DNA fragments with 3' "sticky " ends, bacterial alkaline phosphatase (Worthington) was used as described by (A. M. Maxairt and W. Gilbert, supra), except that the incubation time was 30 min at 65° C. The dephosphorylated DNA was purified by adsorption on and elution from DEAE -cellulose (W. Muller et al., "Site-Directed Mutagenesis In DNA: Generation of Point Mutations Corresponding To Amino Acids 121-123", J. Mol. Biol., 124, pages 343-58 (1978)) or subjected polyacrylamide gel electrophoresis when necessary (see below) Fragments recovered from the polyacrylamide (or agarose) gel in 0.15 M NaCl, 0.05 M Tris-HCl (ppH 8) were absorbed on a 0.1 ml hydroxyapatite ("Biorad HTP") column, washed 4 times with 1 ml of 0.1 M potassium phosphate buffer (pH 7) and eluted with 0.3 ml of 1 M potassium phosphate buffer (pH 7). The solution was diluted 10 times and DNA absorb deposited on DEAE cellulose and recovered 10 times as described by W. Muller et al., supra.

After ethanol precipitation, DNA was 5<1->terminated labeled with

32

[Y-P] ATP (12-34 pCi per pmol DNA 5' ends) and polynucleotid kinase (New England Biolabs or P-L Biochemicals Inc.)

substantially as described by A. M. Maxam and W. Gilbert, supra, except that the DNA was not denatured prior to the kinase reaction. Specific activities of 1-1.5 pCi-

32

[ P] phosphate per pmol DNA 5' ends were obtained.

For sequencing, labeled fragments were digested with a second restriction enzyme and the products separated by electrophoresis through a 5% polyacrylamide gel in tris-borate-EDTA buffer. The desired fragments were extracted from the gel and purified (Muller et al., supra). The different fragments for sequencing were prepared as follows (the number indicates the nominal fragment chain length in base pairs, the indicated positions are marked with an asterisk): (1) cleavage of Hif-2h with BspI, isolation of BspI-BspI-232 and BspI-BspI -949 by 5% polyacrylamide gel electrophoresis in <M>Loening's" buffer (U. E. Loening "The Fraction Of High Molecular Weight Ribonucleic Acid By Polyacrylamide Gel Elec-trophoreses", J. Biochem. page 102 (1967)); (2) cleavage of Hif-2h with BspI, labeling, cleavage with PstI, isolation of BspI<*->PstI-83 and BspI*- PstI-827; (3) cleavage of Hif-2h with Bglll, labeling, cleavage by Pstl, isolation of BglII*- PstI-336 and BglII*- PstI-570; (4) digestion of Hif-2h with MboII, labelling, digestion with PstI and HindIII (to cleave an interfering 350 bp pBR322 fragment), isolation of MboII - PstI -519 and MboII - PstI-351; (5) cleavage of Hif-2h with EcoRI, labelling, cleavage with PstI, isolation of EcoRI*- PstI-708 and EcoRI*- PstI-198; (6) cleavage ning of Hif-2h with Pstl, labeling, cleavage with Bglll, isolation of PstI*- BglII-570 and Pstl -BglII-336; (7) cleavage of Hif-2h with AvalI, labelling, cleavage with Pstl and Bglll, isolation of AvaII - PstI-186 and Avall - Bglll-147; (8) cleavage of Hif-2h with PvuII, labeling, cleavage with Pstl and Bglll, isolation of PvuII - PstI-486.

The fragments were degraded according to the procedure of A. M. Maxam and W. Gilbert, supra, with the modifications described in the protocols provided by the same authors in September 1978. The products were fractionated at 0.1 x 25 x 36 cm 12% 'ig polyacrylamide gels (acrylamide/bisacrylamide = 18/1) in 50 mM tris-borate, 1 mM EDTA (pH 8.3) with experiments performed at 2, 8, 18 and 26 h at 900 V after a 6 h pre-electrophoresis at 700 V. Best results were obtained when the gels were kept at room temperature 2-3 days before use.

Each stretch of the cDNA insert was sequenced from both bands and each restriction site, which served as a labeled terminus, was sequenced using a fragment comprising it. The nucleotide sequences thus obtained are reproduced in figures 8-10. As expected, the positions of the various restriction sites in this insert are located more absolutely than those determined by restriction enzyme digestion alone and reproduced in fig. 4.

With reference to fig. 8-10 it can be seen that the heteropolymeric part of the insert is flanked by 23G residues at the 5' end and by 7A residues (possibly reflecting the poly(A) termination of the mRNA) followed by 15C residues at 3 ' end part. For reference, the insert is numbered from the first nucleotide following the dG tail to the last nucleotide before the poly(A) residues. An ATG initiating triplet at positions 57-59 and a TAA terminating triplet at positions 624-626 define a reading frame not interrupted by nonsense codons. Both of the other two reading frames in this region of the insert contain 18 and 12 nonsense codons, respectively. Additional,

the only other sequences, i.e. in different reading frames flanked by ATG and GTG and a terminating signal,

which codes for a polypeptide of 25 amino acids or more (

are located respectively between nucleotides 226 and 304, 640 and 778 and 683 and 743. The area between nucleotides 57 and 626 will therefore most likely include the nucleotide sequences for the Hif-2h fragment which codes for a polypeptide that exhibits a biological and immunological activity of IFN-a.

It must of course be understood that cloned cDNA from poly(A) RNA by the usual methods (A. Efstratiadis et al., supra) lacks 5' terminal nucleotides and may also contain artificial sequences (R. I. Richards et al., "Molecular Cloning . And Sequence Analysis Of Adult Chicken B-Globin cDNA", Nucleic Acids Research, 7, pages 1137-46 (1979)). It is therefore not certain that the ATG located in nucleotides 57-59 is in reality the first ATG in authentic mRNA. However, for the purposes of the following description, it is assumed that the ATG at nucleotides 57-59 is the first ATG in authentic mRNA.

By comparing the polypeptide encoded by this region of the insert to the 35 amino-terminal amino acid sequence of authentic human lymphoblastoid interferon—SerAspLeuProGlnThrHis-SerLeuGlyAsnArgArgAlaLeuIleLeuLeuAlaGlnfletGlyArglleSerLeu-PheSerCysLeuLysAspArgHisAsp—determined by K. C. Zoon et al., supra and M. Hunkapiller and L. Hood, supra, it seems that the selected reading frame is correct and that nucleotides 57-124 may code for a signal sequence that precedes the nucleotide sequence that codes for the fully developed polypeptide, because a comparison of the published sequence with the determined sequence (from the 24 amino acid and beyond) show an extensive homology (ie for 26 out of 35 amino acids).

In eukaryotic mRNAs, the first AUG triplet from the 5' end is usually the initiating position for protein synthesis (M. Kozak, "How Do Eukaryotic Ribosomes Select Initiation Regions In Messenger RNA", Cell, 15, pages 1109 -25

(1978)). The codon in the Hif-2h fragment corresponding to the first amino acid of lymphoblastoid interferon^ is 22 codons from the first AUG (and 14 codons from the second) indicating that the sequence that codes for interferon may precede the sequence that determines a signal peptide of 23 (or less likely 15) amino acids. The longest of the putative signal sequences contains an uninterrupted series of 11 hydrophobic amino acids (and the shortest one of 6 hydrophobic amino acids). This accumulation of hydrophobic residues is characteristic for a signal sequence (cf-, B. D. Davis and P. C. Tai, "The Mechanism Of Protein Secretion Across Membranes", Nature, 283, pages 433-38 (1980)).

The sequence which apparently corresponds to fully developed IFN-α polypeptide comprises 498 nucleotides which code for 166 amino acids. Assuming that no carboxy-terminal truncation (processing) occurs, the molecular weight of the interferon polypeptide is 19,388. The base composition of the coding sequence is 50% GC. The codon usage in the interferon coding sequence is in reasonable agreement with that suggested for mammalian mRNAs in general. (R. Grantham , et al., "Codon Catalog Usage And The Genome Hypothesis", Nucleic Acids Research, 8, pages 49-62 (1980)).Any discrepancies observed may be attributed to the small numbers involved.

The structure of the polypeptide shown in fig. 8-10 for the Hif-2h fragment, naturally does not take into account possible modifications of the polypeptide caused by the interaction with the in vivo enzymes (eg glycosylation). It must therefore be understood that this structure does not have to be identical to the in vivo-formed IFN-α, but has very similar, if not identical, biological and immunological properties. Nor will this structure exclude the possibility that other modifications such as mutations, including single or multiple, base substitutions, deletions, insertions, or inversions or chemical derivatives from this structure will form compounds that also exhibit IFN-a activity.

3. Determination of the plus strand in the inserted IFN acDNA

The DNA strand which has the same sequence as the mRNA is referred to as the plus strand and its complement as the minus strand. The plus strand of the IFN-acDNA insert was identified as indicated in fig. 5. Hif-2h DNA was digested with the restriction enzyme Bglll. the ends labeled with 3 2P-phosphate, as described above for Pstl-cleaved ends) and the DNA was treated with Pstl to give longer (545 b.p. (570 b.p. determined by the more refined method of analysis reported above)) and shorter 340 bp (336 bp as determined by the more refined method of analysis indicated above)) radioactive fragments. These fragments were denatured and hybridized to poly(A) RNA from induced leukocytes in 80% formamide, 0.4 M NaCl, i.e. under conditions where DNA-DNA reassociation does not take place (supra). The nucleic acids were treated with nuclease Sl, which degrades all single-stranded nucleic acids, especially non-hybrid-32 ■

d.ized P-DNA (and the products were separated on a polyacrylamide gel (R. F. Weaver and C. Weissmann, "Mapping Of RNA By A Modification • Of The Berk-Sharp Procedure", Nucleic Acid Research, 7, pages 1175-93 (1979 . page).

4. Demonstration that poly(A) RNA from non-induced human leukocytes does not hybridize to Hif-2h DNA

An experiment identical to that described in the previous section was performed, however poly(A) RNA from non-induced human leukocytes was used, prepared by the same procedure as described for Sendai virus-induced leukocytes. No detectable amounts of labeled DNA were protected. By comparison with the results from the previous section, poly(A) RNA from non-induced cells contained less than 1/20 the amount of mRNA hybridized to Hif-2h than that of poly(A) RNA induced cells.

SYNTHESIS OF A POLYPEPTIDE WITH INTERFERON ACTIVITY USING E. COLI CONTAINING RECOMBINANT DNA MOLECULES RELATED

WITH Z- pBR322( Pst)/ HcIF~ 4c

The pstl position for pBR322 is within the 3-lactamase (penicillinase) gene. Therefore, when a coding DNA segment (for example a cDNA comprising all or part of a gene) is inoculated

in the position in the correct orientation and correct reading frame, a fused protein can occur. The protein will consist of an amino-terminal part of 3-lactamase followed by the amino acid sequence for which the inserted DNA sequence codes

(L. Villa-Komaroff et al., supra). If the inserted DNA segment comprises a DNA sequence containing its own initiation signal and has a sequence preceding this with a terminating signal in phase with the 3-lactamase sequence, termination and reinitiation can take place at the second initiation signal and a non- -fused, active protein may occur (A.C.Y. Chang et al., supra). To ensure that the DNA

the insert related to Ii.if-.4c was inserted into the correct reading frame for expression within the 3-lactamase body, a set of derivatives of pBR322, namely pKT279, pKT280 and pKT287 (constructed by K. Talmadge, personal communication 1979) became. Each of these derivatives has a Pstl position located so that a DNA insert grafted into this position will be in a different reading frame compared to an insert at the Pstl position for the other derivatives of the set (Fig. 6). Therefore, the kit enables the introduction of DNA into the 3-lactamase gene in all three reading frames. The Pstl excised insert from Hif-2h was prepared as described for the fragment Hif-4c. The Hif-2h Pst fragment (10 ng) was mixed with PstI-cleaved pBR322, pKT279, pKT280 or pKT287 (10 ng in each case) in 20 µl 10 mM Tris-HCl (pH 7.5), 6 mM MgCl2, 100 mM NaCl, 6 mM 3-mercaptoethanol, 200 pg/ml gelatin and 0.1 mM ATP and incubated with 0.1 units of T4 DNA ligase (New England Biolabs) for 16 h at 10°C. The resulting recombinant DNA molecules are designated Z-pBR322 (Pst)/HcIF-2h, Z-pKT279(Pst)/HcIF-2h, Z-pKT280(Pst)/HcIF-2h and Z-pKT287(Pst)/HcIF- 2h. E. coli HB101 was transformed with each of these recombinant DNA molecules, and transformed colonies were selected on tetracycline-containing agar plates, as previously described. Since tetracycline-resistant clones of transformed bacteria may also contain the recycled vector, the bacterial colonies from each set were grown on "Millipore" filters and colonies hybridized to <32>P-labeled Hif-4c fragments were identified and selected, as described above. These strains were termed as follows,

E. coli HB101 (Z-pBR322(Pst)/HcIF-2h-AH1) to (-AH3),

E. coli HB101 (Z-pKT279(Pst)/HcIF-2h-AH1). to (-AH8),

E. coli HB101 (Z-pKT280(Pst)/HcIF-2h-AH1) to (-AH8),

E. coli HB101 (Z-pKT287(Pst)/HcIF-2h-AH1) to (-AH8),

Extracts of some of the above-mentioned strains as well as some of the strains Z-pBR322 (Pst)/HcIF-SNl to 95; were examined with respect to IFN-α activity. The bacteria were grown in a tryptone medium to stationary phase, collected, washed with 1/20 vol (calculated on the volume of the culture) 50 mM Tris-HCl (pH 8), 30 mM NaCl and frozen. After thawing, the cells were resuspended in the volume indicated below of the previous buffer and lysozyme was added to 1 mg/ml. After 60 min. at 0°C, the suspensions were frozen (in an ethanol dry ice bath) and thawed (at 37°C) 5 times and centrifuged for 10 min. at 12,000 rpm. in a "GSA Sorvall" rotor.

In certain cases, a portion of the supernatant (S30) was further centrifuged at 100,000 x g in a "Type 65 Spinco" rotor and the supernatants (S100) recovered. Such supernatants were examined with respect to IFN-ct activity determined by reduction of the cytopathic effect (Experiment 1). The colonies that showed a positive response in experiment 1 were again examined, as well as 49 clones from the set Z-pBR322/ HcIF-SN-1 to SN-95 as described previously, at a smaller dilution (experiment 2).

Source of the extract:

Source of the extract:

Some of the most active manufacturers from those mentioned above were examined in more detail. Cultures were grown to late "log-phase" (observed OD650 ca" °'9' and cells lysed as indicated above in 1/50 of the culture volume. The following activities were found using Z-pBR322(Pst)/HcIF-SN32 as negative control:

It must be understood that the protein in question produced by these strains has not been structurally analyzed to determine whether it is produced fused to amino acids unrelated to IFN or to all or part of IFN's signal sequence. However, whatever protein is produced, it exhibits an immunological and biological activity like IFN. Therefore, the protein as expressed is useful. More important

is that the protein's activity demonstrates that the DNA sequence that codes for it is a DNA sequence related to HuIFN-a. It is therefore within the state of the art to use this DNA sequence, as shown, to select other similar HuIFN-a related DNA sequences and to provide a basis for other constructs which will express fully developed interferon or other variants thereof, or which will improve the yield of the particular protein expressed.

The production level of a protein is controlled by two main factors: the number of copies of its gene inside the cell and the efficiency with which these gene copies are transcribed and translated. The efficiency of transcription and translation (which together constitute expression) is in turn dependent on the nucleotide sequences that normally occur in front of the desired coding sequence.

These nucleotide sequences or expression control sequences define, among other things, the location at which RNA polymerase acts to initiate transcription (the promoter sequence) and to which ribosomes bind and interact with mRNA (the product of the transcription) to initiate translation. Not all such expression control sequences operate with the same efficiency. It is thus advantageous to separate the specific coding sequences for the desired protein from their adjacent nucleotide sequences and instead fuse them to other known expression control sequences, thus promoting high levels of expression. When this has been achieved, the newly constructed DNA fragment can be inserted into a multicopy plasmid or a bacteriophage derivative to increase the number of gene copies inside the cell and thereby further improve the yield of the expressed protein.

Several expression control sequences can be used as described above. These include the operator, promoter and ribosome binding and effector sequences (including sequences such as the Shine-Dalgarno sequences) for the lactose operon of

E. coli ("the lac system"), the corresponding sequences for the tryptophan synthetase system of E. coli ("the trp system"), the main operator and promoter regions of phage X (0_P_ and OrPr), the control region of phage fd coat protein , or other sequences that control the expression of genes of prokaryotic or eukaryotic cells and their viruses. Therefore, to enhance the production of particular polypeptides in a suitable host, the gene encoding this polypeptide can be produced as before and removed from a recombinant DNA molecule containing it, and again introduced into a recombinant DNA molecule closer to its former expression control sequence or under control of one of the above-mentioned expression<on~ troll sequences. Such methods are known from the prior art.

Further increase in cellular yield of the desired products is dependent on the increase in the number of genes that can be utilized in the cell. This is achieved, for example, by introducing recombinant DNA molecules constructed as previously described into the appropriate ("template") bacteriophage A. (NM989), most simply by treating the plasmid with a restriction enzyme to give a linear molecule which is then mixed with a restriction-treated phage A cloning vehicle (for example, the type described by N. E. Murray et al., "Lambdoid Phages That Simplify The Recovery Of In Vitro Recombinants", Molec. gen. Genet. 150, pages 53-61 (1977) and N. E. Murray et al., "Molecular Cloning Of The DNA Ligase Gene From Bacteriophage T4", J. Mol. Biol., 132, pages 493-505 (1979)) and the recombinant DNA molecule prepared by incubation with DNA ligase. The desired recombinant phage is selected as before and used to lysogenize a host of the E. coli strain.

Particularly useful A. clone carriers contain a temperature tolerance-like mutant of the repression gene cl and suppressible mutations in the gene S, the product of which is required for lysis of the host cell, and the gene E, the product of which is the major capsid protein of the virus. With this sister, the lysogenic cells are grown at 32°C and then heated to 45°C to induce truncation of the prophage. Further growth at 37°C leads to high production levels of the protein which is retained inside the cells as these are not lysed by the phage products in the normal way and as the phage deposits are not encapsidated it will remain available for further transcription. Artificial lysis of the cells will thereby release the product in high yield.

In a first attempt to increase the yield of the polypeptide exhibiting a biological or immunological activity as human leukocyte interferon, prepared from hosts transformed with Z-pBR322(Pst)/HcIF-SN35 by the method described above, a restriction ion map of the DNA insert for the hybrid determined. This mapping showed that compared to Hif-2h, Hif-SN35 lacked a Pstl position flanking the 3' end of the sequence, part of the signal sequence was missing (up to and including codon 7) and the Avall position in the signal sequence was replaced with a Bspl position . Therefore, Hif-SN35 trol:g is a polymeric or heritable variant of Hif-2h.

The plasmid Hif-SN35 was opened with Pstl and the resulting DNA band chewed back at both ends using standard procedure and the LAC-Alu fragment (infra) inserted therein and the plasmid resealed. The current structure of the modified plasmid, identified as Z-pBR322(Pst)/HcIF-SN3 5-AHL6, and the amino acid sequence at the amino terminus of the protein produced in E. coli have been determined. The nucleotide sequence of this construct shows that the LAC-Alu fragment was linked to an amino acid away from the first amino acid of IFN-α1 (SN35). This amino sequence of the amino terminal portion of the protein expressed in E. coli showed that a fusion protein was prepared with six amino acids fused to the IFN-α1 (SN35) sequence. However, hosts transformed with the modified plasmid produced approx. 100 times more polypeptide that exhibited the biological or immunological activity for human leukocyte interferon, compared to hosts transformed with unmodified Z-pBR322(Pst)/HcIF-SN35.

Referring to Figure 25, another attempt to improve the yield of the polypeptide exhibiting an immunological or biological activity for human leukocyte interferon, produced by hosts transformed with Z-pBR322-(Pst)/HcIF-SN35 (SN35 in Figure 25) is shown. . The hybrid plasmid was cleaved using standard procedures with BspI (a gift from Dr. Kiss). After heat inactivation (65°C, 30 min.) of the restriction enzyme, the mixture was adjusted to 50 mM Tris-HCl (pH 8) and heated (37°C, 30 min.). After extraction with phenol and ether, the largest, namely the alcDNA fragment, was isolated on low-temperature gelling agarose (0.8%)

and HindIII linkers attached. The modified fragment was ligated with HindIII-cleaved plasmid HS-pBR322 (Eco)/lacUV5-150 ("LAC-150")<*> (a gift from H. Schaller) by melting fragment-containing gel pieces (approximately 20 µl each) at 65°C, cool to 37°C and add 20 units/µl T4 DNA ligase. After 16 h at 15°C, ligation took place in the solidified gel (H. Lehrach, (personal communication 1980). One-tenth volume of 100 mM Tris-HCl (pH 7.5), 100 mM CaCl2, 100 mM MgCl2 was added to the sample, and it was heated for 5 min at 65° C and cooled to 37° C. The samples were then used to transform a Ca 2 -treated E. coli HB101, incubated at

OE OE

0 C for 20 min., heated at 42 C for 1 min. and for 10 min. at 20 oC. After adding 1 mol of tryptone medium, the samples were incubated for 60 min. at 37°C and placed on agar plates containing ampicillin. Plasmid DNA was isolated from these cultures as previously indicated, and the hybrid plasmid containing the IFN-α1 insert at its 5' end adjacent LAC fragments identified by restriction analysis. The plasmid was then digested with EcoRI using conventional methods and treated with exonuclease BAL-31 (0.06

+ This plasmid contains the lac promoter HaeIII-202 bp fragment (W. Gilbert et al., "Lactose Operator Sequences And The Action Of Lac Repressor" in Protein Ligand Interactions, H. Sund and G. Blauer, eds. (Berlin, Walter de Gruyter), pages 193-206 and. K. Backman et al., "Maximizing Gene Expression On A Plasmid Using Recombination In Vitro", Cell. 13. pages 65-71 (1978) flanked by an EcoRI linker at its 3' end.

units/ml, 2-4 min. at 30°C) to remove the overhanging EcoRI tail of the LAC fragment and to truncate the B-galactosidase coding segment.

To ensure that the treated plasmid contained a complete IFN-α1 coding sequence, the plasmid was cleaved with BglII, using standard procedures, prepared as before, and the largest fragment was separated on agarose gel (0.8%). This fragment was then combined with a BspI-BglII fragment from Z-pBR322 (Pst)/HcIF-SN35, and the resulting hybrid plasmid used to transform E. coli HB101 as previously indicated. The transformed colonies were screened for IFN activity and a clone with a high level of IFN activity was selected. This clone was designated E. coli HB101 (C8-IFN-al) and its hybrid plasmid C8-IFN-al.

DNA sequence analysis of C8-IFN-α1 showed that the coding sequence following the initiator triplet determines the first 7 amino acids of B-galactosidase, a Pro residue generated by fusion, amino acids 16-23 of the IFN-α1 signal sequence and IFN-α1 (SN35 ) sequence. E. coli minicell strains (DS410) transformed with the hybrid plasmid C8-IFN-a1 produce approx. 50 million units of IFN per liter or approx. 2500 times more polypeptide exhibiting an immunological or biological activity similar to that of HuIFN, compared to minicells transformed with unmodified Z-pBR322(Pst)/Hif-SN35. The amino acid sequence of the polypeptide produced by the plasmid C8-IFN-a1 confirms that the product is a fused protein of 7 amino acids from 3-galactosidase, an amino acid generated by fusing amino acids 16-23 of the IFN-a1 signal sequence fused to IFN-a1.

Further examples of different constructions to improve the protein yield according to the present invention are discussed in connection with other forms of IFN-a (infra).

CHARACTERISTICS OF INTERFERON ACTIVITY PRODUCED BY E. COLI TRANSFORMED WITH HYBRID PLASMIDS

1. The sensitivity of IFN-α activity to trypsin

50 µl samples of authentic HuIFN-α (specific activity 1.2 x 10^ units/mg, 50U) and the SlOO extracts described above of E. coli HB101 (Z-pKT287(Pst)/HcIF-2h-AG6) ("Hif -287-6 extracts") (200 units/ml, 10 units) and of E. coli HB101 (Z-pBR322(Pst)/HcIF-SN35) ("Hif-35 extracts") (1000 units/ml, 50 units) were incubated with different amounts of trypsin, as indicated below, for 30 min. at 37°C.

As the S100 extracts have a higher protein content, which HuIFN-a does not, a mixture of HuIFN-a and control SlOO extracts Hif-32 was examined in parallel:

Accordingly, IFN-α from the extracts is sensitive to trypsin and is therefore a protein.

2. Behavior by chromatography on "Sephadex G-100" Extract Hif-35 (1 ml) and the S100 extract of E. coli HB101 (Z-pBR322(Pst)/HcIF-SN32) ("Hif-32 extracts") was chromatographed on a 32-ml Sephadex G-100 column at 4 o'C in 50 mM K-phosphate buffer (pHi 7.4). Cytochrome c (0.2 mg) was added as an internal marker. The flow rate was 2 ml/ 5 and 1 ml fractions were collected. The absorbance at 280 nm and 405 nm (cytochrome c) and the IFN-α activity were determined. As shown in Fig. 7, the IFN-α activity of the Hif-35 extracts eluted before cytochrome c, with a k^ value of about 0.45. Therefore, the observed molecular weight for the substance is between 20,000 and 30,000 (The molecular weight determined by nucleotide sequencing and assuming no carboxy-terminal "processing" is 19,388). No activity was detected for fractions of the control extract Hif-32 3. Inhibition of interferon activity for Hif-35 and Hif-287-6 by means of antibody against human leukocyte interferon.

HuIFN-a (specific activity 1.2 x 10<6> IU/mg), and the Hif-35/ Hif-287-6 S100 extracts were incubated with various dilutions of sheep antiserum against HuIFN-a (prepared by K. Cantell, 24 February 1976, specific activity 450,000 units/ml) in 100 µl "Modified Eagles Medium" (MEM) with 10% calf serum for 30 min. at 37°C after which 45 µl were examined with regard to IFN-α activity in reducing the cytophatic effect. (The antibody itself does not cause any cytophatic effect).

To show that the effect of the antibody was not due to an unspecified effect, such as a proteolytic degradation, similar experiments were performed with the mouse interferon system:

With antibodies active against HuIFN-α, the IFN-α activity is inhibited in particular for polypeptides produced in E. coli transformed with certain recombinant DNA molecules containing the HcIF-2h DNA sequence. The apparently low affinity of the antibody towards IFN-α produced by E. coli may possibly reflect structural differences between the latter and natural HuIFN-α, for example the absence of a carbohydrate unit, the presence of a signal sequence or the fusion of two parts of B- the lactamase sequence.

4. Reduced activity of Hif-35 and Hif-287-6 extracts against mouse cells

Human CCL23 cells or Mouse L929 cells were treated with

E. coli extracts, HuIFN-a (prepared by K. Cantell, specific activity 1.2 x 106 units/mg) or mouse IF (N.I.H. standard) were infected with virus (Mengo virus when the human cells and VSV when gjalt the mouse cells) and the IFN-α activity determined by reduction of the cytopathic effect determination.

The results show that Hif-35 and Hif-287-6 extracts have a protective effect on human cells and only a small effect (~10%) on mouse cells, which is typical for human interferon.

5. Effect on certain cell functions

Extracts from E. coli HB101 (Z-pBR322(Pst)/Hif-SN35-AHL6) were compared with authentic IFN for its effect on certain cell functions. IFN-a produced by E. coli exhibits the following properties as for natural IFN-a: (1) it enhances the natural killing activity of human lymphocytes, (2) it enhances antibody-dependent cell-associated cytotoxicity, (3) it suppresses antigen- and mitogen-induced leukocyte migration inhibition and (4) it inhibits growth of IFN-sensitive Burkitt lymphocytes. These properties are indicative of E. coli synthesized IFN-α1's activity against human tumors and cancer. 6. Activity of IFN-α without amino-terminal sequences IFN-α without amino-terminal sequences has also been produced in E. coli and has been shown to exhibit activity consistent with IFN-activity.

To construct the appropriate recombinant DNA molecule, plasmid Hif-2H (supra) was partially digested with EcoRI and BamHI and the fragment containing the IFN-α1 coding sequence separated on agarose gel and combined with the non-IFN-α1 coding sequence obtained from an EcoRI/ BamHI digestion of the plasmid Hif-SN35 lacking a Pstl position adjacent to the 3' end of the hybrid insert, compared to Hif-2h (supra). The resulting plasmid was then digested with PstI/BgII to remove part of the amino-terminal part of the IFN-α1 coding sequence. In its place was inserted a series of IFN-α1 fragments prepared by cleavage of the plasmid Hif-2h with PvuII, treatment with Bal exonuclease, attachment of Pstl linkers and cleavage

ing with Bglll. The resulting plasmids thus contained a series of IFN-α1 coding sequences lacking various parts of their amino-terminal sequences. One of these (plasmid 2H-M8) was digested with Pstl and its nucleotide sequence determined.

The sequence showed that plasmid 2H-M8 contained several nucleotides between the Pst position and the first codon (CYS<N> of IFN-α1. Therefore, PstI-cleaved plasmid 2H-M8 was treated with T4-polynuclease/dATP, Sl exonuclease and cleaved with Sali. This procedure generated a series of fragments whose IFN-α1 coding sequences were missing parts of their amino terminus. These fragments were then placed under LAC control by operatively linking them to the GUA-LAC fragment prepared from plasmid Lac3V8 by digestion with EcoRI, treatment with exonuclease S1 and digestion with SalI Thus, the resulting fragmenting series of plasmids had IFN-α1 coding sequences lacking various parts of their amino termini linked in a "counter-clockwise" direction via an AUG to a fragment containing the LAC promoter.

Some of these plasmids were sequenced. One began at the fifth amino acid in IFN-α and one at the tenth amino acid i

IFN-al. In E. coli minicells (DS410), both of these plasmids produced polypeptides exhibiting IFN activity. Therefore, not all of the IFN-α1 protein is required for IFN activity.

IDENTIFICATION OF E. COLI CLONES CONTAINING RECOMBINANT DNA MOLECULES, WHOSE DNA INSERTS CROSS-HYBRIDIZE WEAKLY TO THE INSERTS IN Hif-4c AND WHICH HAVE A DIFFERENT RESTRICTION MAP THAN THE Hif-2h FRAGMENT

Comparison of the first 35 amino acids of authentic lymphoblastoid interferon (Zoon et al., supra, and M. Hunkapiller and L. Hood, supra) and the sequence deduced from the Hif-2h fragment revealed 9 differences. In all cases, the codons for the separating amino acids should be related by "one base" forward changes. The amino acid compositions determined directly for authentic lymphoblastoid interferon on the one hand and deduced from the sequence of the Hif-2h fragment on the other hand also show remarkable differences with respect to their content of Gly, Pro, Cys and Met. These differences are too large to be explained by polymorphism. Instead, they probably reflect the existence of at least two non-inherited genes, because the degree of deviation for the two proteins (26% mismatch) corresponds, for example, to that between human and sheep P-globulin (23% mismatch)• Consequently, the clones that exhibited weak hybridization to fragment Hif-4c, previously identified (supra), was examined, and a clone E. coli HB101 (Z-pBR322(Pst)/HcIF-II-206) was identified.

The hybrid plasmid Z-pBR322(Pst)/HcIF-II-206 ("HcIF-II-206") in this clone and its DNA insert "Hif-II-206 fragment" hybridize weakly to Hif-4c and Hif-2h- the fragment. E. coli transformed with plasmid Hif-II-206 produces polypeptides that exhibit a biological or immunological activity similar to that of HuIFN-α. The Hif-11-206 fragment has a restriction map

which is different from that determined for the Hif-2h fragment.

A comparison of the two restriction maps is shown in fig. 11. Again, the absolute localization of the section positions for the Hif-II-206 fragment is not determined by restriction mapping alone. However, sequencing the nucleotide sequences of this insert, using the procedures described above, allows a more absolute determination

of these locations. However, due to the difference in the restriction map of the Hif-II-206 fragment compared to that of the Hif-2h fragment it is clear that the interferon genes for these two inserts have different nucleotide sequences.

With reference to figures 12-16, nucleotide sequences are shown for the inserted DNA sequence -- the Hif-II-206 fragment -- of culture HcIF-G and for the inserted DNA sequence -- the Hif-2h fragment -- of culture HcIF -E (infra), previously determined and corresponding amino acid sequences for proteins encoded by these nucleotide sequences. The nucleotide sequences of the Hif-II-206 fragment, namely the Pst I fragment (790bp) of plasmid DNA, isolated using the method of method B described by N.M. Wilkie et al., Nucl. Acids Res., 7, pp. 859-77 (1979) from culture HcIF-G was determined using the standard procedure of A. M. Maxam and W. Gilbert, supra. The sequencing strategy used is shown in figure 17.

The restriction digested DNA (usually about 10 µg) was 5'-terminally sequenced

as described by N. Mantei et al., Gene, 10, pages 1-

10 (1980). Labeled fragments were cleaved with a second restriction enzyme, the products separated by electrophoresis through a 5% polyacrylamide gel in Tris-borate-EDTA buffer (A.

C. Peacock and C.W. Dingman. Bioch., 6, pages 1818-27

(1967)), extracted from the gel and purified as described by

W. Muller et al., J. Mol. Biol., 124, pp. 343-58 (1978).

With reference to fig. 17, the various fragments for sequencing were prepared as follows: 25 and 26 — cleavage of Hif-II-206 with Pstl, labelling, cleavage with Bglll, isolation of a Pstl*-Bglll fragment (257bp) ("25") and a PstI<*->BglII fragment (279bp) ("26"),

21, 22 and 23 — cleavage of Hif-II-206 with PvuII, labelling, cleavage with Bglll, isolation of a PvuII*- BglII fragment (88bp) ("21"), a PvuII<*->BglII fragment ( 176bp) ("22") and a PvuI<*->BglII fragment (214bp) ("23"),

11, 12, 13 and 14 — cleavage of Hif-II-206 with Bglll, labelling, cleavage with Pstl, isolation of a Bglll -Pstl fragment (279bp) ("14") and a co-migrating mixture of a Bglll - Psti- fragment and a Bglll - Bglll fragment. Cleavage of the mixture with PvuII and isolation of a BglII - Pstl fragment (257bp) ("13") and a BglII*- PvuII fragment (176bp)

("12") and a BglII<*->PvuIl fragment (88bp) ("11").

27L, 27U, 41, 43, 44 and 45 — cleavage of Hif-II-206 with HinfI, labelling, isolation of the precursor fragments: HinfI tit-Hinfl<*> (113bp) ("27P"), HinfI*- HinfI* ( 146bp) ("28P"), HinfI*- Hinfl* (159bp) ("30P"), HinfI*- HinfI* (397bp) ("31P") and HinfI*- HinfI* (1522bp) ("32P"). Cleavage of 28P with MboII and isolation of a fragment HinfI - MboII (112bp)

("41"). Cleavage of 30P with MboII and isolation of a fragment HinfI*- MboII (126bp) ("43"). Cleavage of 31P with Pstl and isolation of a fragment HinfI - Pstl (151bp) ("44").

Hf Cleavage of 32P with Pstl and isolation of a fragment HinfI - Pstl (139bp) ("45"). Chain separation of 27P to give two chains

:..("27U" and "27L" ) .

All segments, except 27U and 27L, were sequenced on both strands and above the réstriks jonsposis jonefté that served as starting sites for sequencing, except for the Bglll alignment at position 185.

From a comparison of the amino acid sequences encoded by the two inserts, it is apparent that the Hif-II-206 fragment encodes an interferon-like protein containing one amino acid less than the Hif-2h fragment (amino acid 44 (Asp) present in Hif-2h is missing in Hif-II-206j A further 10% of the nucleotide positions and 17% of the derived amino acid residues in the two fragments are different.

Additionally, when comparing the 35 amino acids determined for the amino terminus of lymphoblastoid interferon (K. C. Zoon et al., Science, 207, pages 527-28 (1980)), the Hif-II-206 insert encodes a protein that differs by 5 residues from the 35 amino acid residues determined by Zoon et al., supra. Therefore, at least 3 different leukocyte-type (a-type) IFN genes must exist, namely the Hif-2h fragment, the Hif-II-206 fragment and the gene for Zoon's IFN. According to the newly proposed nomenclature for interferon, in the following the proteins coded for by these genes will be identified as follows:

The difference between IFN-α1 and IFN-α2 is also reflected in their varying activities on human CCL23 and bovine fetal kidney (BEK) cells:

Therefore, IFN-al is approx. 30 times less active on human cells than IFN-a2. They are still approx. equally active on bovine cells. Therefore, in addition to their usual use as antiviral and antitumor or anticancer agents for humans, IFNs may also be useful in the treatment of these conditions in cattle. For example production of HuIFN-a could be used in a standard way (supra) in the treatment of FMDV and other well-known virus infections in cattle. This is more particularly the case for IFN-al as its activity on bovine cells is approx. 20 times greater than its activity on human cells.

As a result of the improved yield obtained in the case of IFN-α1 with construction C8 (supra), a similar construction was made for IFN-α2. Z-pBR322(Pst)/Hif-II-206 was digested completely with BspI and partially with PvuII (at P1) (Fig. 28) and 867 bp fragments were isolated from a 6% polyacrylamide gel. This fragment was then cleaved into a 2590 bp PvuII fragment by C8-IFN-α1. (C8-IFN-al has three

PvuII restriction sites (Fig. 28).^2590 bp fragments lie between P. and P^.) The resulting hybrid plasmid was used to transform E. coli HB101 and the clones screened foc AFN activity. One clone showed a high activity and was selected and designated E. coli HB101 (C0-IFN-α2).

DNA sequencing of the hybrid plasmid C8-IFN-a2 showed that, like C8-IFN-a1, it had a coding sequence following the initiator triplet that determined the first 7 amino acids of 6-galactosidase, a Pro residue generated by melting the amino acids 16 to 23 of the IFN.-α2 signal sequence. Therefore, it is again likely that a fused protein containing IFN-α2 is expressed.

Mini cells transformed with this plasmid gave 100-200 million units per 1 of IFN or 20000-40000 times higher yields for IFN-α2 than minicells transformed with unmodified Z-pBR322(Pst)/Hif-II-206.

A comparison of the relative yields for C8-IFN-al

(~50 x 10 /r units/liter) and C8-IFN-a2 (~100-200 million units per liter) are surprising at first sight because IFN-a2 has been shown to be approx. 30 times more active than IFN-al on human cells (supra). However, quantitative analysis of the amount of the two proteins produced from the minicells, transformed with the two C8 plasmids shows that in C8-IFN-a1 approx. 5-6 times more protein than in C8-IFN-a2. Therefore, the yields measured on the basis of IFN activity are shifted due to the much larger amount of protein produced for the case of C8-IFN-α1.

With reference to fig. 26 another construction is shown in an attempt to improve the yield of IFN-α2. First, an expression plasmid was prepared containing the LAC-Alu fragment with restriction of the known lac promoter with Alul and extension of the fragment as shown (for 1 end part) in fig. 27 with EcoRI linkers. The extended fragment was introduced into pBR322 at the EcoRI position and a small EcoRI-EcoRI fragment was removed from the construct. The resulting plasmid, designated 404 in FIG. 2(}/ was digested with a HindIII and PvuII to introduce the IFN-α2-containing fragment. This IFN-α2 fragment was prepared by partial Sau3A digestion of Z-pBR322(Pst)/HcIF-II-206 (" 206" in Fig. 26), extension of the Sau3A fragment with HindIII linkers (Fig. 27) and cleavage with Bsp_I. After introduction of this fragment into the HindiII-PvuII-cleaved plasmid 404, the resulting plasmid was cleaved with HindIII and EcoRI, treated with S1 nuclease to bring the LAC promoter closer to the IFN-α2 gene and religated This construct^identified as plasmid LAC-AUG(α2)/ has the IFN-α2 DNA sequence under the control of the LAC promoter.Additionally The IFN-α2 sequence immediately follows the initiating AUG codon of this promoter (see Fig. 27). Therefore, at least a portion of the IFN produced by these plasmids will be fully developed IFN, e.g., IFN without any amino acids from the signal sequence)' .

In mini cells, the yield of IFN-cc2 obtained with plasmid LAC-AUG(a2) was 5-10 million units per litres.

Another IFN-ct2 construction based on similar attachment principles has also been produced. Here, the penicillinase expression control sequence for pBR322 was connected via an AUG initiator codon to the IFN-ct2 gene from HIF-II-206.<*>

This construction could most efficiently be done by a partial dissolution of pBR322 with MboII, treatment with Sl, attachment of EcoRI linkers as previously indicated and re-introducing the fragment into the EcoRI position of pBR322 and omitting an EcoRI position. The resulting plasmid ("B-lac-AUG plasmid") could then be combined with a Sl treated (mild) HindIII linker-BspI fragment of 206, as previously described. After cleavage with EcoRI and treatment with Sl and phosphatase, the expression plasmid B-lac-AUG(a2) is isolated.

The construction with other genes could be carried out in the same way by utilizing the constructed B-lac-AUG plasmid for the introduction of other genes or constructions or more preferably by using the Sau3A position in the 3-lac-AUG(a2) plasmid itself.

This plasmid, designated as B-lac-AUG(a2) f enables, when used to transform host cells, the production of IFN-a2 without fusion to other protein sequences. In mini cells, yields are 50-100 million units per liters observed. .This plasmid is the most preferred plasmid according to the present invention. It is also preferred for use with other IFN-a genes shown in the present description. The preferred host according to the invention is E. coli DS410 (ara azide TonA lac y Tsx min a min b galXxyl step ). This strain has been deposited with a sample of the preferred plasmid B-lac-AUG(a2) as HcIF-K.

Other constructions in which different promoter sequences, ribosome binding positions, Sh.ine-Dalgarnose sequences and DNA sequences between the promoter and the AUG initiator codon are used and using different IFN-α genes shown herein can also be constructed using similar method principles.

HYBRID MOLECULES OF IFN-α1 AND IFN-α2

A number of hybrid molecules of IFN-α1 and IFN-α2 have been constructed. Surprisingly, these hybrid constructs exhibit quantitatively different properties and activities compared to each of their parents, namely IFN-α1 or IFN-α2.

Referring to Fig. 28, in which schematic construction of four of these hybrid molecules is shown. For convenience, these hybrid molecules are designated as plasmids I, II, III and IV. In these constructions, digestion with restriction enzymes (obtained from Biolabs, except e.g., a gift from Dr. Kiss) performed essentially as recommended by the suppliers. Partial DNA digestion was performed with reduced enzyme amounts. After heat inactivation (65°C, 30 min) of the restriction enzyme, samples were adjusted to 50 mM Tris-HCl (pH 8) and, if necessary, calf intestinal alkaline phosphatase (Boehringen) (1 vol per pg DNA) added. After 30 min at 37°C, the samples were extracted with phenol and ether. In most cases, The DNA fragments separated on low-temperature gelled agarose (0.8%). For ligation, fragment-containing gel pieces (each about 20 µl) were melted at 65°C, cooled to 37°C and 20 units per µl of T4 DNA were added -ligase The mixture was kept at 15°C for 16 hours and ligation took place in the solidified gel (H. Lehrach, personal communication 1980). One-tenth volume of 100 mM Tris-HCl (pH 7.5), 100 mM CaCl2, 100 mM MgCl2 was added and the sample heated for 5 min. at 65°C and then cooled to 37°C.

+2

The samples were then added to Ca - treated minicells, incubated at 0°C for 20 min, heated for 1 min. at 42°C and 10 min. at 20°C and 1 m in tryptone medium added. After incubation for 60 min. at 37°c the cultures were plated on agar plates containing suitable antibiotics. All plasmids were characterized by nucleotide sequence analysis of the junction in the IFN sequence.

Hybrid molecule I, a g-1 (PvuII)a-2 hybrid, was constructed by partial cleavage of C8-IFN-a1 (supra) (C8-a1 in Fig. 28) with PvuII, dephosphorylated, cleaved with Pstl and Pstl-PvuII (P2) 13 4 6 bp fragment isolated. This fragment was ligated to

a 2135 bp Ps_tl (a) - PvuII (P-,) fragment prepared by total cleavage of C8-IFN-ri2 (supra) (C8-a2 in Fig. 28) with PvuII and partially cleaving this with Pstl.

Hybrid molecule II, the a-1(Bglll)a-2 hybrid, was constructed by cleaving hybrid molecule I with Bglll. After dephosphorylation, the larger Bglll fragment was isolated and ligated to the smaller Bglll fragment of C8-IFN-α2. After cloning, the hybrid plasmid with the smaller BglII fragment in the correct orientation was identified by restriction analysis.

Hybrid molecule III, an α-2(PvuII)α-1 hybrid, was constructed by partial cleavage of C8-IFN-α1 with PvuII, dephosphorylation, cleavage with Aval and isolation of 1686 bp PvuII(P2)-Aval and 3233 bp PvuII(P^)- The Aval fragments. These fragments were then ligated to the 300 bp PvuII (P.^ - PvuII (P2 ) fragment in HcIF-II-206 (supra) (SN206 in Fig. 28) and the plasmid containing the smaller PvuII fragment was identified by examining transformed E. coli strains for IFN-α activity.

The hybrid molecule IV, an α-2(Bglll)α-1 hybrid, was constructed by cleaving C8-IFN-α1 with Bglll and Aval and the 1776 bp fragment isolated. This fragment was then ligated to the 3543 bp Bglll-Aval fragment in hybrid molecule III.

The biological activities of the different interferon types in relation to each other were also determined. Cultures of minicells (DS410), transformed with the different plasmids, were grown and the bacteria collected by centrifugation, washed with PBS, suspended in PBS (about 1/20 of the original volume), incubated for 60 min. at 0°C with 1 mg/ml lysozyme, 10 mM EDTA, freeze-thaw 4 times, sheared by passing 5 times through a syringe and separated by centrifugation.

The activities against human, mouse, guinea pig and bovine cells were as follows:

Surprisingly, all the interferons show approximately the same activity against cattle cells, while IFN-α1 and the two hybrid IFNs (I and II) which have the same amino terminal groups of IFN-α1 show 10 to 1000 times lower activity on human cells than IFN-α2 and the two hybrid IFNs (III and IV) containing the amino terminal groups of IFN-α2. It is even more surprising that the two hybrid IFNs (I and II) with the amino terminal part of IIN-al exhibit more than 10 times lower activity on human cells than IFN-al itself. Nevertheless, one of the hybrids (III), with the amino-terminal part of IFN-α2, exhibits approximately the same activity on human cells as IFN-α2.

IDENTIFICATION OF CHROMOSOMAL GENES FOR IFN-g

A collection of hybrid phages derived from human fetal chromosomal DNA generated by partial digestion with Haelll and Alul and joined with EcoRI linkers A.Charon 4A arms has been prepared by R. M. Lawn et al., Cell, 15, pages 1157-74

(1978). This gene bank was screened by an "in situ" method (W.D. Benton and R.W. Davis, Science, 196, pp. 180-82 (1977); T. Maniatis et al., Cell, 15, pp. 687-701

(1978)) using as a "probe" the 3 2P-labelled IFN-g1 cDNA insert removed from pBR322(Pst)/Hif-2h. Sixteen hybridization-positive phage clones were isolated from 240,000 plates by repeated plate purifications (T. Maniatis et al., supra). Ten of the hybrid phage DNA preparations were digested with HindIII, Tad, Hhal, BamHI, EcoRI and BglII, respectively, and the fragments separated by electrophoresis on an agarose gel, transferred to a "Millipore" membrane (E. M. Southern, J. Mol. Biol., 98 , pages 503-17 (1975)) and hybridized with <32>P-labeled Hif-2h cDNA insert. Fig. 18 summarizes the results in the form of partial restriction maps and various tables. As shown, for each hybrid phage DNA preparation, at least a few characteristic restriction sites and region(s) were established that hybridize to the separated IFN-gl gene "probe"

(black arrow).

The shaded area on chr-16 in Figures 18 and 2 4 represents a sequence that hybridizes

weak to Hif-2h cDNA but which did not show large R-"looping").

Referring to Figures 18 and 24, it can be seen that clones chr-3 and chr-26 may represent DNA segments that overlap greatly over their length because they have several EcoRI and HindIII fragments in common. In addition, the hybridizing part of chr-1 and one of the hybridizing parts of chr-10 may be the same because the Kindlll-Hindlll and EcoRI-EcoRI fragments that hybridize with the Hif-2h "probe" have the same length (3.2kb respectively and 0.95kb). It also appears that the "right hand" hybridizing part of chr-16 (marked "r" in Fig.

18) may be identical to the hybridizing part of chr-35, although it is oriented in the opposite direction, as each of the two clones gives a 1.4 kb Bg.l.II-Bglll fragment and a 2 kb EcoRI-EcoRI fragment which hybridizes with the Hif-2h cDNA "probe".

It is therefore very likely that chr-16 and chr-35 deposits overlap.

Consequently, it seems that the 13 hybridizing parts of the

11 hybrid chromosomal DNAs fall into no less than 10 distinct classes — chr-1, chr-3, chr-12, chr-13, chr-16 (left hand, labeled "1" in Fig. 18) chr-26 , chr-30, chr-35, chr-19 and chr-27.

With reference to fig. 24 is shown the overlap of chr-1, chr-3, chr-10 and chr-26 and that of chr-16 and chr-35.

The data given above indicate that the genome of a single human contains no less than 10 different DNA sequences that cross-hybridize to Hif-2h. This conclusion is reinforced by the fact that the proportion of Hif-2h related sequences detected in the clone bank is approx. 1 in 16000. Assuming a value of 3 x 10 9 bp for the haploid human, the expected value for a single gene copy with an average DNA fragment size of 16kb (the average value for the examined clones) approx. 1:190000. Therefore, the frequency of Hif-2h related fragments is 12 times higher than expected for a single gene.

When comparing these data/when Lawn et al. (supra) screened 300,000 plates from the same genebank with a 3-globulin cDNA probe, only 2 positive clones were identified, the expected value being 1.6:300,000.

Therefore, there may be 10-15 distinct chromosomal DNA segments in the human gnome that cross-hybridize to the Hif-2h fragment or IFN-alcDNA.

FURTHER CHARACTERIZATION OF Hif-chr3 5

By way of illustration only, the hybridization sequence of chr-35 ("Hif-chr 35") was further characterized.

The hybridizing portion of chr-35 ("Hif-chr35") (and the right segment of Hif-chrl6 to which it is most likely identical, supra) is the only hybridizing chromosomal DNA portion of a Bglll position. As IFN-a1 and IFN-a2 cDNAs have respectively 1 and 2 Bglll positions within their coding sequences, it seems likely that Hif-chr35 is a counterpart to one of the two previously cloned interferon genes. Hif-chr 35's strong hybridization to the 3' terminal Hif-2h cDNA fragment (containing only the 3' non-coding region) compared to the weaker hybridization for the other chromosomal DNAs of this "probe" supports the likelihood of correspondence between Hif -chr35 and Hif-2h (f. Kafatos et al., Proe. Nati. Acad. Sei USA, 74, pages 5618-22 (1977)).

To further analyze the Hif-chr35 fragment, a HindIII-BamHI fragment was cleaved from chr-35. This fragment (3.4kb) contained the hybridizing part ("Hif-chr35") of chr-35. This fragment was subcloned into the Pstl position of pBR322 using the well-known dC-dG ligation procedures (L. Villa-Komaroff et al., supra) and E. coli HB101 was transformed with the resulting recombinant DNA molecule, under using well-known methods (for example, S. Nagata et al., Nature, 284, pages 316-20 (1980)).

Clones of these transformants were screened by in situ colony hybridization (D. Hanahan and M. Meselson, Gene, 10, pages 63-67 (1980)) with the 3 2P-labeled Hif-2h fragment (supra), and plasmid DNA, Z-pBR322(Pst)/HchrIF-35HB "HchrIF-35HB" was separated from positive clones (N.M. Wilkie et al., Nucleic Acids Res., 7, pages 859-77 (1979); Method B). The orientation of the hybrid insert "HchrIF-35HB fragment" in the plasmid in relation to the 3-lactamase gene of pBR322 was determined by EcoRI cleavage and size determination of the fragments obtained. The insert oriented so that it coincided with 3-lactamase was denoted by a and the opposite orientation by 3 -

Cultures of these positive clones were grown to a freezing OD65Q = 0.8 and the bacteria collected and lysed by the lysozyme freeze-thaw method described by S. Nagata et al., supra. Seven of the 10 clones examined showed IFN-α activity of 75-500 units/g cells determined by reduction of the cytopathic effect.

The DNA insert in one of these 7 IFN-producing clones, namely E. coli HB 101 (Z-DBR322(Pst)/HchrIF-35HBa), was further characterized by restriction analysis and determination of the nucleotide sequence. Plasmid DNA (HchrIF-35HBa) was prepared from the clone as previously described and restriction positions determined by Smith-Birnstiel mapping (H. 0. Smith and M. L. Birnstiel, Nucl. Acids Res., 3, pages 2387-98 (1976)); where HchrIF-35HBa was digested with EcoRI, labeled at the 5' terminus and digested with BglII (and PstI to cleave unwanted fragments of about 1kb).

1.04kb EcoRI-Bglll (3' proximal) and 0.96kb EcoRI-Bglll (5' proximal) fragments were isolated by agarose gel electrophoresis as described by A. C. Peacock and C. W. Dingman, Biochemistry, 6, pages 1818-27 (1967) ). Both fragments were partially cleaved with Hinf, BspI and MboII, respectively, and the products separated on a 1% agarose gel in triacetate buffer (pH 7.8) containing 1 pg/ml ethidium bromide. After staining, the radioactive bands were visualized by autoradiography. The Bst.NI and HgiAI positions were determined similarly for the 1.04kb (3' proximal) fragment. The results of this analysis are shown in fig. 19.

For nucleotide sequence determination, HchrIF-35HBa was cleaved with various restriction enzymes and the products separated by electrophoresis through a 5% polyacrylamide gel in Tris-borate-EDTA buffer (A.C. Peacock and C.W. Dingman, supra) and extracted from the gel and purified as described by W. Muller et al., J. Mol. Biol., 124, pages 343-58

(1978).

The sequencing strategy used is shown in fig. 19 and can

is described as follows:

I and 2, — cleavage of HchrIF-35HBa with BglII, labelling, digestion with EcoRI and Pstl and isolation of a BglII - EcoRI fragment (940bp) ("1") and a BglII<*->EcoRI (360bp) (" 2")

3 and 4, — cleavage of HchrIF-35HBa with EcoRI, labelling,

fairy

digestion with BspI and isolation of an EcoRI - BspI fragment (680bp) ("3") and an EcoRI<*->BspI fragment (880bp) ("4"),

5, 6, 7 and 8, — cleavage of HchrIF-35HBa with PvuII, labelling, cleavage with Bglll and EcoRI and isolation of PvuII - EcoRI fragment (780bp) ("5"), a PvuII*- BglII (215bp)

("6"), a P_£UII<*->BgJ.II fragment (90bp) ("7"), and a PvuII<*->EcoRI fragment (290bp) ("8"), 9 and 10 , cleavage of HchrIF-35HBa with EcoRI, isolation by 1% agarose gel electrophoresis in Tris-borate EDTA buffer of a 1300 bp EcoRI-EcoRI fragment, further cleavage with HinfI and isolation of a HinfI-HinfI fragment (450bp), and a HinfI- HinfI fragment (180bp).

Labeling of the larger HinfI-HinfI fragment and cleavage with MboII enabled the isolation of a HinfI-MboII (190bp) ("9"). Labeling of the shorter HinfI- HinfI fragment and cleavage

fairy

with AvaII enabled the isolation of a HinfI - AvaII fragment (150bp) ("10"),

II — cleavage of HchrIF-35HBa with MboII, labeling, cleavage with Bglll, and isolation of an MboII - Bglll fragment (465bp) ("11")

12, 13 and 14 — digestion of HchrIF-35HBa with BspI and Bglll, isolation by agarose electrophoresis as above of a 1200bp BspI-BglII fragment and (a) digestion with HgiAI, labelling, digestion with MboII and isolation of a HgiAI - MboII- fragment (300bp) ("12") and a H_jiAI<*->MboII fragment (360bp) ("13"), or (b) digestion with BstNI, labeling, digestion with EcoRI and isolation of a BstNI*- EcoRI- fragment (380bp) ("14").

The various fragments were sequenced using the Maxam-Gilbert procedure (supra). All fragments were sequenced on both strands and above the restriction sites that served as starting points for sequencing.

A comparison of the nucleotide sequences of the coding region of HchrIF-35HBa and that of Hif-2h (coding region) (Figs. 8-10, compared with Figures 20-23) shows that they are identical. In particular, it is surprising that there is no indication of the presence of introns within the coding sequence of the HchrIF-35HBa fragment, i.e. between the HinfI position and in the 5' non-coding region and the EcoRI position in the 3' non-coding region - the coding region. Thus, no intron could be detected in the chromosomal sequence corresponding to fully developed IFN-amRNA.

FURTHER CHARACTERIZATION OF Hif-chr 2 6 AND Hif-chr 3

The gene containing segments of chr-3 and chr-26

identical by heteroduplex analysis, but differ by at least one BglII restriction site when examined by nucleotide sequencing. Five nucleotide differences in 725 base pairs were found. Only two of these are present in the coding sequences. Then not only the genes but also at least 3.5 Kbp preceding and 6.0 Kbp following them, formed a perfect heteroduplex, and due to the relatively low sequence divergence that includes only two amino acid changes it seems that Hif-chr3 and Hif -chr26 are inherited forms of the same gene. These are designated IFN-α4a (Hif-chr3) and IFN-α4b (Hif-chr26). The nucleotide sequence and corresponding amino acid sequence for IFN-α4b determined by the previously described conventional sequencing techniques are shown in Figures 29-32.

A comparison of Figures 29-32 with Figures 8-10, 12-16 and 20-23 shows that proteins encoded by each of these sequences differ from each other by approx. 15% of their residuals. This divergence is typical for products of non-heritable genes that have diverged 20 - 90 million years ago.

Expression of Hifchr35 in mouse cells

Plasmid Z-pBR322(pst)/HchrIF-35HDa (supra) was used as a source of a Hif-chr35 fragment for expression in mouse cells. The plasmid was digested with PstI and treated with 5' exonuclease to remove the 5<1->dG tails. This fragment was introduced into a 5<1->dG-capped KpnI fragment of a plasmid prepared by joining BamHI-BamHI fragments from pBR322 and polynomial DNA. The resulting vector was used to transform mouse 3'f3 cells using the calcium phosphate technique (N. Mantei et al., Nature, 281 pages 40-46

(1979)). These transformed cells were designated mouse-3T3 (polynoma-Hif-chr35) for convenience. After 20-40 h, determinations showed an IFN-α activity of 300 units/ml of IFN-α on human cells and approx. 3000 units/ml IFN-a on bovine cells.

It should of course be understood that in the nucleotide sequence described in figures 8-10, 12-16, 20-23 and 29-32, possible modifications in the nucleotide sequences such as mutation, single or multiple, base substitutions, insertions, inversions or deletions which has already taken place, or which can be used in the future. Furthermore, in the sequence, a possible substitution of other codons coding for the same amino acid as a codon depicted in these figures is not taken into account. It should therefore be understood that such modified sequences, as a code for polypeptides exhibiting an immunological or biological activity for IFN-a, also fall within the scope of the present invention.

In addition, it must be understood that in the amino acid sequences depicted

in Figures 8-10, 12-16, 20-23 and 29-32, possible modifications in the polypeptides caused by their exposure to in vivo or in vitro agents, e.g. in vivo glycosylation enzymes.

PRODUCTION IN BACTERIAL HOSTS POLYPEPTIDES EXHIBITING AN IMMUNOLOGICAL OR BIOLOGICAL ACTIVITY LIKE INTERFERON

Then the determination of the reduction of the cytopathic effect

(W. E. Stewart II and S. E. Sulkin, S. E. Proe. Soc. Exp.

Biol. Med., 123, pages 650-53 (1966)) can detect extremely small amounts of IFN, namely less than one active molecule per bacteria cell, lysates of E. coli HB101 were infected with ten hybrid A. phages, as previously described and examined for the presence of IFN. Seven of the eleven phages (all except chr-10, chr-12, chr-19 and chr-27) gave lysates containing IFN activity in the range of 3 to 50 units/ml. For the case of chr-10 and chr-12, hybridization (to Hif-2h) HindIII-HindiII or EcoRi-EcoRI fragments, subcloned into the Pstl position of pBR322, as previously described, expressed iFN-α activity in E. coli .

Since E. coli is believed to be unable to splice mRNA (O. Mercereau-Puijalon and P. Kourilsky, Nature, 279, pages 647-49 (1979)), these IFN-α chromosomal genes probably do not contain introns in its coding region.

FINAL CONCLUSIONS

A set of recombinant DNA molecules containing cDNA prepared from poly(A) RNA from Sendai virus-treated (induced) human leukocytes has been isolated, representative of which exhibit the following properties: (1) They hybridize to poly(A) RNA from induced, but not from uninduced human leukocytes. (2) They hybridize to IFN-amRNA, as shown by their ability to select this RNA from a mixture of RNAs as well as by their ability to inhibit (reversible) translation of interferon mRNA by hybrid-arrested translation assay. (3) E. coli containing certain members of the set produces a compound with the following properties.

(a) It is sensitive to trypsin

(b) It exhibits IFN-α activity in the human cell system and only

little activity in the mouse cell sister network

(c) It has a molecular weight in the range of 20,000 - 30,000 (19,388

based on the nucleotide sequencing in Figures 8 - 10)

(d) IFN-α activity is specifically inhibited by antibody

to human leukocyte interferon.

(4) The DNA inserts of hybrid plasmids of the invention are available, in addition to their availability for selecting IFN-amRNA from a mixture of RNA<*>s for selecting IFN--aDNA from mixtures of different sources including cDNAs and from hybrid phage gene banks of chromosomal DNA. (5) A number of different chromosomal genes for IFN-α exist. Unexpectedly, these genes lack interons and allow direct expression of interferon and interferon-like polypeptides in appropriate hosts. (6) At least three of the nucleotide sequences of these recombinant DNA molecules are different and indicate the existence of at least 3 non-hereditary genes for IFN-α. (7) The proteins encoded by these three nucleotide sequences differ from the 35 amino acids determined from authentic lymphoblastoid interferon. (8) Hybrid proteins prepared for different combinations of IFN-α gene segments exhibit quantitatively different properties from each other, or their parents and the proteins have additional amino acids linked to IFN-α or proteins comprising IFN-α without part of its amino-terminal sequence exhibiting IFN-α Activity.

These properties show that recombinant DNA molecules described according to the present invention contain at least part of the coding sequence for human leukocyte interferon and that some of these plasmids lead to expression in E. coli of a polypeptide that exhibits an immunological or biological activity corresponding to that of human leukocyte interferon. It should also be obvious that the shown polypeptides can be fragmented, modified or converted into derivatives, as is well known in the protein art. Microorganisms and recombinant DNA molecules produced by the present method are exemplified by cultures deposited in the culture collection of the Deutsche Sammlung von Mikroorganismen in Gottingen, West Germany on January 7, 1980 and identified as HcIF-A to E:

A: E. coli HB101 (Z-pBR322(Pst)/HcIF-4c

B: E. coli HB101 (Z-pBR322(Pst)/HcIF-2h

C: E. coli HB101 (Z-pBR322(Pst)/HcIF-SN35

0: E. coli HB101 (Z-pBR322(Pst)/HcIF-SN42

E: E. coli HB101 (Z-pKT322(Pst)/HcIF-2h-AH6

These cultures were respectively given the registration numbers DSM 1699-1703.

In addition, the microorganisms and recombinant DNA molecules produced by the present method are exemplified by cultures deposited in the American Type Culture Collection, Rockville, Maryland on March 27

1980 and identified as HcIF-G to H and given ATCC registration numbers 31633 and 3i634 respectively:

G: E.coli HB101 (Z-pBR322(Pst)/HcIF-II-206)

H: E.coli HB101 (Z-pBR322(Pst)/HcIF-SN35-AHL6)

Other microorganisms produced by the present method are exemplified by cultures deposited in the culture collection of the Deutsche Sammlung von Mikroorganismen, in Gottingen, West Germany on October 1, 1980 and identified as HchrIF-A to J and given the registration numbers DSM 1914-1923:

A. subcloned HindIII fragment of chr 3 in E. coli HB101,

B. subcloned EcoRI fragment of chr 12 in E. coli HB101,

C. cloned HindIII fragment of chr 12 in E. coli HB101,

D. subcloned EcoRI fragment of chr 13 in E. coli HB101,

E. subcloned EcoRI fragment of chr 23 in E. coli HB101,

F. the subcloned HindIII fragment of chr 23 in E. coli HB101,

G. subcloned EcoRI fragment of chr 26 in E. coli HB101,

H. subcloned HindIII fragment of chr 26 in E. coli HB101,

I. subcloned HindIII/BamHI fragment of chr 35 in E. coli

HB101,

J. subcloned BamHI fragment of chr 35 in E. coli HB101.

In conclusion, the microorganisms produced by the present method are exemplified by cultures deposited in the American Type Gul ture Collection, Rockville, Maryland, December 15, 1980 and identified as HchrIF-K to HchrIF-Q and HcIF-I to HcIF-K and given the ATTC registration numbers:

31760 HchrIF-K

31761 HchrIF-L

31762 HchrlF-M

31763 HchrIF-N

31764 HchrIF-0

31765 HchrlF-P

31766 HchrIF-Q

31767 HcIF-I

31768 HcIF-J

31769 HcIF-K,

respectively:

K. subcloned Tac-Tac fragment of chr 23 in E. coli

HB101. L. subcloned Bglll-Bglll fragment of chr 10 1 in E. coli HB101. M. subcloned HindIII- HindIII fragment of chr 10r in E.

coli HB101.

N. subcloned Bglll-Bglll fragment of chr 26 in E. coli

HB101.

O. subcloned HindIII- HindIII fragment of chr 30 in E.

coli HB101.

P. subcloned the Bglll-Tac fragment of chr 13 in E. coli

HB101.

Q. subcloned the Bglll-Tac fragment of chr 16 1 in E. coli

HB101.'

HcIF-I: E. coli DS410 (C8-IFN-α2)

HcIF-J: E. coli DS410 (LAC-AUG(a2))

HcIF-K: E. coli DS410 (B-Lac-AUG(_2))

Claims (9)

1. Recombinant DNA molecule, characterized in that it comprises a DNA sequence that codes for a polypeptide of the IFN-α type selected from: (a) the DNA inserts Z-pBR322(Pst)/HcIF-4c, Z-pBR322(Pst)/HcIF-2h, Z-pBR322(Pst)/HcIF-SN35, Z-pBR322(Pst)/HcIF-SN42 and Z-pKT287(Pst) /HcIF-2h-AH6, where these DNA inserts are DNA inserts of the recombinant DNA molecules in the microorganisms identified by accession numbers DSM (Deutsche Sammlung von Mikroorganismen) 1699-1703, (b) DNA sequences that hybridize to any of the preceding DNA inserts and which encode for an IFN-α type polypeptide, and (c) DNA sequences which are degenerate as a result of the genetic code relative to the DNA sequences and inserts indicated under (a) and (b), and which encode a polypeptide of IFN-cx type, where these DNA sequences are operatively linked to an expression control sequence in the recombinant DNA molecule. 2. Recombinant DNA molecule according to claim 1, characterized in that the DNA sequence (b) that hybridizes to the DNA insert (a) is selected from: (d) the DNA inserts Z-pBR322(Pst)/HcIF-II-206 and Z-pBR322(Pst)/HcIF-SN35-AHL6, which DNA inserts are DNA inserts of the recombinant DNA molecules in the microorganisms identified by ATCC (American Type Culture Collection) accession numbers 31633-31634), (e) DNA- sequences which hybridize to any of the foregoing DNA inserts and which encode an IFN-α type polypeptide, and (f) DNA sequences which are degenerate as a result of the genetic code of the DNA sequences and inserts set forth under (d ) and (e) and which encodes an IFN-a type polypeptide. 3. Recombinant DNA molecule according to claim 1 or 2, characterized in that the DNA sequence (b) or (e), which hybridizes to the DNA insert (a) or (d), is selected from: (g) the HindIII fragment of Hif-chr3, EcoRI fragment of Hif-chrl2, HindIII fragment of Hif-chrl2, EcoRI fragment of Hif-chrl3, EcoRI fragment of Hif-chr23, HindIII fragment of Hif-chr23, EcoRI fragment of Hif-chr26, HindIII fragment of Hif -chr26, Hindlll-BamHI fragment of Hif-chr35, BamHI fragment of Hif-chr35, Tac-Tac fragment of Hif-chr23, Billl-Bglll fragment of Hif-chrlOÆ, Hindlll-Hindlll fragment of Hif-chrlOr, Bglll-Bglll fragment of Hif-chr26, the Hindlll-Hindlll fragment of Hif-chr30, the Bglll-Tac fragment of Hif-chrl3, and the Bglll-Tac fragment of Hif-chrl6Æ, which hybridizing parts are the DNA inserts of the recombinant DNA molecules in the microorganisms identified by accession numbers DSM (Deutsche Sammlung von Mikroorganismen) 1914-1923 and ATCC (American Type Culture Collection) 31760-31766 , (h) DNA sequences which hybridize to any of the foregoing DNA sequences and which encode an IFN-α type polypeptide, and (i) DNA sequences which are degenerate as a result of the genetic code of the DNA sequences and the inserts indicated under points (g) and (h) and which code for a polypeptide of the IFN-α type. 4. Recombinant DNA molecule according to any one of claims 1 to 3, characterized in that the DNA sequence codes for the precursor or fully developed form of an IFN-al type polypeptide with formula: 5. Recombinant DNA molecule according to each of claims 1 to 3, characterized in that the DNA sequence codes for the precursor or fully developed form of an IFN-ot2 polypeptide of formula: 6. Recombinant DNA molecule according to any one of claims 1 to 3, characterized in that the DNA sequence codes for precursor or fully developed form of an IFN-cx4b type polypeptide of formula: 7. Recombinant DNA molecule according to claim 1, characterized in that the expression control sequence is selected from a lac system, a p-lac system, a trp system, main operator and promoter regions of phageX and the control region of fd coat protein, and other sequences that control the expression of genes in bacteria, yeast or animal cells. 8. Recombinant DNA molecule according to claim 1, characterized in that it is selected from C8-IFN-a1, C8-IFN-CX2, LAC-AUG(oc2) and e-lac-AUG(a2), which recombinant DNA molecules are the recombinant DNA molecules in the microorganisms identified by the accession numbers ATCC (American Type Culture Collection) 31767-31769. 9. Recombinant DNA molecule according to claim 1, characterized in that it is selected from Z-pBR322(Pst)/HcIF-2h, Z-pBR322(Pst)/HcIF-SN35, Z-pBR322-(Pst)/HcIF-SN42 , Z-pKT287(Pst)/HcIF-2h-AH6, Z-pBR322(Pst)/- HcIF-II-206, and Z-pBR322(Pst)/HcIF-SN35-AHL6, which recombinant DNA molecules are the recombinant DNA molecules in the microorganisms identified by accession numbers DSM (Deutsche Sammlung von Mikroorganismen) 1699-1703 and ATCC (American Type Culture Collection) 31633-31634.