Title of Invention	ACYLATED INSULIN
Abstract	ABSTRACT The present invention relates to protracted human Insulin derivatives in which the A21 and the B3 amino acid residues are, independently, any amino acid residue which can be coded for by the genetic code except Lys, Arg and Cys; PheB1 may be deleted; the B30 amino acid residue is a) a non-codable, lipophilic amino acid having from LO to 24 carbon atoms, in which case an acyl group of a carboxylic acid with up to 5 carbon atoms is bound to the e-amino group of Lys829; or b) the B30 amino acid residue is deleted or is any amino acid residue which can be coded for by the genetic code except Lys, Arg and Cys, in any of which cases the e-amino group of LysBZ9 has a lipophilic substituent; and any Zn2* complexes thereof with the proviso that when B3 0 is Thr or Ala and A21 and B3 are both Asn, and PheB1 is present, then the insulin derivative is always present as a Zn2* complex.

Title of Invention

ACYLATED INSULIN

Abstract

ABSTRACT The present invention relates to protracted human Insulin derivatives in which the A21 and the B3 amino acid residues are, independently, any amino acid residue which can be coded for by the genetic code except Lys, Arg and Cys; PheB1 may be deleted; the B30 amino acid residue is a) a non-codable, lipophilic amino acid having from LO to 24 carbon atoms, in which case an acyl group of a carboxylic acid with up to 5 carbon atoms is bound to the e-amino group of Lys829; or b) the B30 amino acid residue is deleted or is any amino acid residue which can be coded for by the genetic code except Lys, Arg and Cys, in any of which cases the e-amino group of LysBZ9 has a lipophilic substituent; and any Zn2* complexes thereof with the proviso that when B3 0 is Thr or Ala and A21 and B3 are both Asn, and PheB1 is present, then the insulin derivative is always present as a Zn2* complex.

Full Text	FIELD OF TEE INVENTION The present invention relates to novel human insulin derivatives which are soluble and have a protracted profile of action, to a method of providing such derivatives, to pharmaceutical compositions containing them, and to the use of such insulin derivatives in the treatment of diabetes. BACKGROUND OF THE INVENTION Many diabetic patients are treated with multiple daily insulin injections in a regimen comprising one or two daily injections of a protracted insulin to cover the basal reguirement supplemented by bolus injections of a rapid acting insulin to cover the reguirement related to meals. Protracted insulin compositions are well Known in the art. Thus, one main type of protracted insulin compositions comprises injectable aqueous suspensions of insulin crystals or amorphous insulin. In these compositions, the insulin compounds utilized typically are protamine insulin, zinc insulin or protamine zinc insulin. Certain drawbacks are associated with the use of insulin suspensions. Thus, in order to secure an accurate dosing, the insulin particles must be suspended homogeneously by gentle shaking before a defined volume of the suspension is withdrawn from a vial or expelled from a cartridge. Also, for the storage of insulin suspensions, the temperature must be kept within more narrow limits than for insulin solutions in order to avoid lump formation or coagulation. While it was earlier believed that protamines were non- immunogenic, it has now turned out that protamines can be immunogenic in man and that their use for medical purposes may lead to formation of antibodies (Samuel et al., Studies on the immunogenecity of protamines in humans and experimental animals by means of a micro-complement fixation test, Clin. Exp. Immunol. 11, pp. 252-260 (1978)). Also, evidence has been found that the protamine-insulin complex is itself immunogenic (Kurtz et al., Circulating IgG antibody to protamine in patients treated with protamine-insulins. Diabetologica 25., pp. 322-324 (1983)). Therefore, with some patients the use of protracted insulin compositions containing protamines must be avoided. Another type of protracted insulin compositions are solutions having a pH value below physiological pH from which the insulin will precipitate because of the rise in the pH value when the solution is injected. A drawback with these solutions is that the particle size distribution of the precipitate formed in the tissue on injection, and thus the timing of the medication, depends on the blood flow at the injection site and other parameters in a somewhat unpredictable manner. A further irawback is that the solid particles of the insulin may act as a local irritant causing inflammation of the tissue at the site of injection. Wo 91/12817 (Novo Nordisk A/S) discloses protracted, soluble insulin compositions comprising insulin complexes of zobalt(III). The protraction of these complexes is only intermediate and the bioavailability is reduced. Human insulin has three primary amino groups; the N-terminal group of the A-chain and of the B-chain and the e-amino group of LysB29. Several insulin derivatives which are substituted in one or more of these groups are known in the prior art. Thus, JS Patent No. 3,528,960 (Eli Lilly) relates to N-carboxyaroyl insulins in which one, two or three primary amino groups of the insulin molecule has a carboxyaroyl group. No specifically NB29-substituted insulins are disclosed. According to GB Patent No. 1,492.997 (Nat. Res. Dev. Corp.), it has been found that insulin with a carbamyl substitution at N'^ has an improved profile of hypoglycaemic effect. JP laid-open patent application No. 1-254699 (Kodama Co., Ltd.) discloses insulin wherein a fatty acid is bound to the amino group of Phe^' or to the e-amino group of Lys^^ or to both of these. The stated purpose of the derivatisation is to obtain a pharmacologically acceptable, stable insulin preparation. Insulins, which in the B30 position have an amino acid having (1 at least five carbon atoms which cannot necessarily be coded for by a triplet of nucleotides, are described in JP laid-open patent application No. 57-067548 (Shionogi). The insulin analogues are claimed to be useful in the treatment of diabetes mellitus, particularly in patients who are insulin resistant due to generation of bovine or swine insulin antibodies. By "insulin derivative" as used herein is meant a compound having a molecular structure similar to that of human insulin including the disulfide bridges between CysA7 and CysA7 and between CysA20 and CysB19 and an internal disulfide bridge between cysA6 and CysA11, and which have insulin activity. However, there still is a need for protracted injectable insulin compositions which are solutions and contain insulins which stay in solution after injection and possess minimal inflammatory and immunogenic properties. One object of the present invention is to provide human insulin derivatives, with a protracted profile of action, which are soluble at physiological pH values. Another object of the present invention is to provide a pharmaceutical composition comprising the human insulin derivatives according to the invention. It is a further object of the invention to provide a method of making the human insulin derivatives of the invention. SUMMMIY OF THE IliVBllTIOM Surprisingly, it has turned out that certain human insulin derivatives, wherein the e-amino group of Lys"^ has a lipophilic substituent, have a protracted profile of action and are soluble at physiological pH values. Accordingly, in its broadest aspect, the present invention relates to an insulin derivative having the following sequence: wherein Xaa at positions A21 and B3 are, independently, any amino acid residue which can be coded for by the genetic code except Lys, Arg and Cys; Xaa at position Bl is Phe or is deleted; Xaa at position B30 is (a) a non-codable, lipophilic amino acid having from 10 to 24 carbon atoms, in which case an acyl group of a carboxylic acid with up to 5 carbon atoms is bound to the e-amino group of LygB29^ (b) any amino acid residue which can be coded for by the genetic code except Lys, Arg and Cys, in which case the e-amino group of Lys^^ has a lipophilic substituent or (c) deleted, in which case the c-amino group of Lys^^ has a lipophilic substituent; and any Zn^ complexes thereof,provided that when Xaa at position B30 is Thr or Ala, Xaa at positions A21 and B3 are both Asn, and Xaa at position Bl is Phe, then the insulin derivative is a Zn^ complex. In one preferred embodiment, the invention relates to a human insulin derivative in which the B3 0 amino acid residue is deleted or is any amino acid residue which can be coded for by the genetic code except Lys, Arg and Cys; the A21 and the B3 amino acid residues are, independently, any amino acid residues which can be coded for by the genetic code except Lys, Arg and Cys; Phe^^ may be deleted; the e-amino group of Lys^ has a lipophilic substituent which comprises at least 6 carbon atoms; and 2-4 Zn^ ions may be bound to each insulin hexamer with the proviso that when B30 is Thr or Ala and A21 and B3 are both Asn, and Phe^' is not deleted, then 2-4 Zn^* ions are bound to each hexamer of the insulin derivative. In another preferred embodiment, the invention relates to a human insulin derivative in which the B30 amino acid residue is deleted or is any amino acid residue which can be coded for by the genetic code except Lys, Arg and Cys; the A21 and the B3 amino acid residues are, independently, any amino acid residues Biiiii r. i"'i. III.I 11.1" «a3.3W'tii- Cys, with the proviso that if the B30 amino acid residue is Ala or Thr, then at least one of the residues A21 and B3 is different from Asn; Phe^^ may be deleted; and the c-amino group of Lys^^ has a lipophilic substituent which comprises at least 6 carbon atoms. In another preferred embodiment, the invention relates to a human insulin derivative in which the B30 amino acid residue is deleted or is any amino acid residue which can be coded for by the genetic code except Lys, Arg and Cys; the A21 and the B3 amino acid residues are, independently, any amino acid residues which can be coded for by the genetic code except Lys, Arg and Cys; Phe'^ may be deleted; the e-amino group of Lys^ has a lipophilic substituent which comprises at least 6 carbon atoms; i and 2-4 Zn^'' ions are bound to each insulin hexamer. In another preferred embodiment, the invention relates to a human insulin derivative in which the B30 amino acid residue is deleted. In another preferred embodiment, the invention relates to a 1 human insulin derivative in which the 830 amino acid residue is Asp. In another preferred embodiment, the invention relates to a human insulin derivative in which the B30 amino acid residue is Glu. I In another preferred embodiment, the invention relates to a human insulin derivative in which the B30 amino acid residue is Thr. In another preferred embodiment, the invention relates to a human insulin derivative in which the 830 amino acid is a I lipophilic amino acid having at least 10 carbon atoms. In another preferred embodiment, the invention relates to a human insulin derivative in which the B30 amino acid is a lipophilic a-amino acid having from 10 to 24 carbon atoms. In another preferred embodiment, the invention relates to a human insulin derivative in which the B30 amino acid is a straight chain, saturated, aliphatic a-amino acid having from 10 to 24 carbon atoms. In another preferred embodiment, the invention relates to a human insulin derivative in which the B30 amino acid is D- or L-N'-dodecanoyllysine. In another preferred embodiment, the invention relates to a human insulin derivative in which the B30 amino acid is o-amino decanoic acid. In another preferred embodiment, the invention relates to a human insulin derivative in which the B30 amino acid is a-aroino undecanoic acid. In another preferred embodiment, the invention relates to a human insulin derivative in which the B30 amino acid is a-amino dodecanoic acid. In another preferred embodiment, the invention relates to a human insulin derivative in which the 830 amino acid is a-amino tridecanoic acid. In another preferred embodiment, the invention relates to a human insulin derivative in which the B30 amino acid is o-amino tetradecanoic acid. In another preferred embodiment, the invention relates to a human insulin derivative in which the B30 amino acid is a-amino pentadecanoic acid. In another preferred embodiment, the invention relates to a human insulin derivative in which the B30 amino acid is a-araino hexadecanoic acid. In another preferred embodiment, the invention relates to a human insulin derivative in which the B30 amino acid is an a-amino acid. In another preferred embodiment, the invention relates to a human insulin derivative in which the A21 amino acid residue is Ala. In another preferred embodiment, the invention relates to a human insulin derivative in which the A21 amino acid residue is Gin. In another preferred embodiment, the invention relates to a human insulin derivative in which the A21 amino acid residue is Gly. In another preferred embodiment, the invention relates to a human insulin derivative in which the A21 amino acid residue is Ser. In another preferred embodiment, the invention relates to a human insulin derivative in which the B3 amino acid residue is Asp. In another preferred embodiment, the invention relates to a human insulin derivative in which the B3 amino acid residue is Gin. In another preferred embodiment, the invention relates to a human insulin derivative in which the B3 amino acid residue is Thr. In another preferred embodiment, the invention relates to a human insulin derivative in which the e-amino group of Lys^^ has a lipophilic substituent which is an acyl group corresponding to a carboxylic acid having at least 6 carbon atoms. In another preferred embodiment, the invention relates to a human insulin derivative in which the e-amino group of Lys"^ has a lipophilic substituent which is an acyl group, branched or unbranched, which corresponds to a carboxylic acid having a chain of carbon atoms 8 to 24 atoms long. In another preferred embodiment, the invention relates to a human insulin derivative in which the c-amino group of Lys"^ has a lipophilic substituent which is an acyl group corresponding to a fatty acid having at least 6 carbon atoms. In another preferred embodiment, the invention relates to a human insulin derivative in which the c-amino group of Lys^^ has a lipophilic substituent which is an acyl group corresponding to a linear, saturated carboxylic acid having from 6 to 24 carbon atoms. In another preferred embodiment, the invention relates to a human insulin derivative in which the c-amino group of Lys^^ has a lipophilic substituent which is an acyl group corresponding to a linear, saturated carboxylic acid having from 8 to 12 carbon atoms. In another preferred embodiment, the invention relates to a human insulin derivative in which the e-amino group of Lys^^ has a lipophilic substituent which is an acyl group corresponding to a linear, saturated carboxylic acid having from 10 to 16 carbon atoms. In another preferred embodiment, the invention relates to a human insulin derivative in which the e-amino group of Lys^^ has a lipophilic substituent which is an oligo oxyethylene group comprising up to 10, preferably up to 5, oxyethylene units. In another preferred embodiment, the invention relates to a human insulin derivative in which the e-amino group of LysB29 has a lipophilic substituent which is an oligo oxypropylene group comprising up to 10, preferably up to 5, oxypropylene units. In another preferred embodiment, the invention relates to a human insulin derivative in which each insulin hexamer binds 2 Zn^* ions. In another preferred embodiment, the invention relates to a human insulin derivative in which each insulin hexamer binds 3 Zn^* ions. In another preferred embodiment, the invention relates to a human insulin derivative in which each insulin hexamer binds 4 Zn^* ions. In another preferred embodiment, the invention relates to the use of a human insulin derivative according to the invention for the preparation of a medicament for treating diabetes. In another preferred embodiment, the invention relates to a pharmaceutical composition for the treatment of diabetes in a patient in need of such a treatment comprising a therapeutically effective amount of a human insulin derivative according to the invention together with a pharmaceutically acceptable carrier. In another preferred embodiment, the invention relates to a pharmaceutical composition for the treatment of diabetes in a patient in need of such a treatment comprising a therapeutically effective amount of a human insulin derivative according to the invention, in mixture with an insulin or an insulin analogue which has a rapid onset of action, together with a pharmaceutically acceptable carrier. In another preferred embodiment, the invention relates to a pharmaceutical composition comprising a human insulin derivative according to the invention which is solvible at physiological pH values. In another preferred embodiment, the invention relates to a pharmaceutical composition comprising a human insulin derivative according to the invention which is soluble at pH values in the interval from about 6.5 to about 8.5. In another preferred embodiment, the invention relates to a protracted pharmaceutical composition comprising a human insulin derivative according to the invention. In another preferred embodiment, the invention relates to a pharmaceutical composition which is a solution containing from about 120 nmol/ml to about 1200 nmol/ml, preferably about 600 nmol/ml of a human insulin derivative according to the invention. In another preferred embodiment, the invention relates to a method of treating diabetes in a patient in need of such a treatment comprising administering to the patient a therapeutically effective amount of an insulin derivative according to this invention together with a pharmaceutically acceptable carrier. In another preferred embodiment, the invention relates to a method of treating diabetes in a patient in need of such a treatment comprising administering to the patient a therapeutically effective amount of an insulin derivative according to this invention, in mixture with an insulin or an insulin analogue which has a rapid onset of action, together with a pharmaceutically acceptable carrier. Examples of preferred human insulin derivatives according to the present invention in which no Zn2+ ions are bound are the following: Examples of preferred human insulin derivatives according to the present invention in which two zn2+ ions are bound per insulin hexamer are the following: (NB29_tridecanoyl des(B30) human insulin)6, 2Zn2+, (NB29-tetradecanoyl des(B30) human insulin)6, 2Zn2+, Examples of preferred human insulin derivatives according to the present invention in which three Zn2+ ions are bound per insulin hexamer are the following: (NB29-tridecanoyl des{B30) human insulin)6, 3Zn2+, Examples of preferred human insulin derivatives according to the present Invention in which four Zn2+ ions are bound per insulin hexamer are the following: BRIEF DBSCRZPTIOH OF THE DRAWINGS The present invention is further illustrated with reference to the appended drawings wherein Fig. 1 shows the construction of the plasmid pEA5.3.2; Fig. 2 shows the construction of the plasmid pEAlOS; and Fig. 3 shows the construction of the plasmid pEAli3. DETAILED DESCRIPTION OF THE INVENTION Teminology The three letter codes and one letter codes for the amino acid residues used herein are those stated in J. Biol. chem. 243. p. 3558 (1968). In the DNA seguences, A is adenine, C is cytosine, G is guanine, and T is thymine. The following acronyms are used: DMSO for dimethyl sulphoxide, DMF for dimethylformamide, Boc for tert-butoxycarbonyl, RP-HPLC for reversed phase high performance liquid chromatography, X-OSu is an N-hydroxysuccinimid ester, X is an acyl group, and TFA for trifluoroacetic acid. Preparation of lipophilic insulin derivatives The insulin derivatives according to the present invention can be prepared i.a. as described in the following: 1. Insulin derivatives featuring in position B30 an amino acid residue which can be coded for by the genetic code. e.g. threonine (human insulin) or alanine (porcine insulin). 1.1 starting from human insulin. Human insulin is treated with a Boc-reagent (e.g. di-tfirt-butyl dicarbonate) to form (A1,B1)-diBoc human insulin, i.e., human insulin in which the N-terminal end of both chains are protected by a Boc-group. After an optional purification, e.g. by HPLC, an acyl group is introduced in the c-amino group of LysB29 by allowing the product to react with a N-hydroxysuccinimide ester of the formula X-OSu wherein X is the acyl group to be introduced. In the final step, TFA is used to remove the Hoc-groups and the product^ NB29-X human insulin, is isolated. 1.2 starting from a single chain insulin precursor. A single chain insulin precursor, extended in position Bl with an extension (Ext) which is connected to Bl via an arginine residue and in which the bridge from B30 to Al is an arginine residue, i.e. a compound of the general formula Ext-Arg-B(1-30)-Arg-A(l-21) , can be used as starting material. Acylation of this starting material with a N-hydroxysuccinimide ester of the general formula X-OSu wherein X is an acyl group, introduces the acyl group X in the e-amino group of Lys^" and in the N-terminal amino group of the precursor. On treating this acylated precursor of the formula (NB29-X),X-Ext-Arg-B(l-30)-Arg-A(l-21) with trypsin in a mixture of water and a suitable water-miscible organic solvent, e.g. DMF, DMSO or a lower alcohol, an intermediate of the formula (NB29-X),ArgB31 insulin is obtained. Treating this intermediate with carboxypeptidase B yields the desired product, (NB29-X) insulin. 2. Insulin derivatives with no amino acid residue in position B30. i.e. des(B30^ insulins. 30 2.1 starting from human insulin or porcine insulin. On treatment with carboxypeptidase A in anunonium buffer, human insulin and porcine insulin both yield des(B30) insulin. After an optional purification, the des(B30) insulin is treated with a Boc-reagent (e.g. di-tert-butyl dicarbonate) to form (Al,Bl)- rdiBoc des(B30) insulin, i.e., des(B30) insulin in which the N- terminal end of both chains are protected by a Boc-group. After an optional purification, e.g. by HPLC, an acyl group is introduced in the e-amino group of LysB29 by allowing the product to react with a N-hydroxysuccinimide ester of the formula X-OSu wherein X is the acyl group to be introduced. In the final step, TFA is used to remove the Boc-groups and the product, (NB29-X) des(B30) insulin, is isolated. 2.2 Starting from a single chain human insulin precursor. A single chain human insulin precursor, which is extended in position Bl with an extension (Ext) which is connected to Bl via an arginine residue and which has a bridge from B30 to Al can be a useful starting material. Preferably, the bridge is a peptide of the formula Y^-Arg, where Y is a codable amino acid except lysine and arginine, and n is zero or an integer between 1 and 35. When n>l, the Y's may designate different amino acids. Preferred examples of the bridge from B3 0 to Al are: AlaAlaArg, SerArg, serAspAspAlaArg and Arg (European Patent No. 163529). Treatment of such a precursor of the general formula Ext-Arg-B( 1-30)-Y„-Arg-A( 1-21) with a lysyl endopeptidase, e.g. ActroTnobacter lyticus protease, yields Ext-Arg-B(l-29) Thr-Y„- Arg-A(l-21) des(B30) insulin. Acylation of this intermediate with a N-hydroxysuccinimide ester of the general formula X-OSu wherein X is an acyl group, introduces the acyl group X in the e-amino group of LysB29, and in the N-terminal amino group of the chain and the B-chain to give (NB29-X) X-Ext-Arg-B(l-29) X- Thr-Y^-Arg-A(l-21) des(B30) insulin. This intermediate on treatment with trypsin in mixture of water and a suitable organic solvent, e.g. DMF, DMSO or a lower alcohol, gives the desired derivative, CNB29-X) des(B30) human insulin. Data on MB29 Kodified insulins. Certain experimental data on NB29 modified insulins are given in Table 1. The lipophilicity of an insulin derivative relative to human insulin, krel, was measured on a LiChrosorb RP18 (um, 250x4 man) HPLC column by isocratic elution at 40"C using mixtures of A) 0.1 M sodium phosphate buffer, pH 7.3, containing 10% acetonitrile, and B) 50% acetonitrile in water as eluents. The elution was monitored by following the UV absorption of the eluate at 214 nm. Void time, tg, was found by injecting 0.1 mM sodium nitrate. Retention time for human insul in, t^^^f^, was adjusted to at least 2to by varying the ratio between the A and B solutions. The degree of prolongation of the blood glucose lowering effect was studied in rabbits. Each insulin derivative was tested by subcutaneous injection of 12 nmol thereof in each of six rabbits in the single day retardation test. Blood sampling for glucose analysis was performed before injection and at 1, 2, 4 and 6 hours after injection. The glucose values found are of expressed as percent of initial values. The Index of Protraction, which was calculated from the blood glucose values, is the scaled Index of Protraction (prolongation), see p. 211 in Markussen et al.. Protein Engineering 1 (1987) 205- 213. The formula has been scaled to render a value of 100 with bovine ultralente insulin and a value of 0 with Actrapid* insulin (Novo Nordisk A/S, 2880 Bagsvaerd, Denmark). The insulin derivatives listed in Table 1 were administered in solutions containing 3 Zn2+ per insulin hexamer, except those specifically indicated to be Zn-free. For the very protracted analogues the rabbit model is inadequate because the decrease in blood glucose from initial is too small to estimate the index of protraction. The prolongation of such analogues is better characterized by the disappearance rate in pigs. T50% is the time when 50% of the A14 Tyr(125I) analogue has disappeared from the site of injection as measured with an external Tf-counter (Ribel, U et al.. The Pig as a Model for Subcutaneous Absorption in Man. In: M. serrano-Rios and P.J. Lefebre (Eds): Diabetes 1985; Proceedings of the l2th Congress of the International Diabetes Federation, Madrid, Spain, 1985 (Excerpta Medica, Amsterdam, (1986) 891-96). In Table 2 are given the T50% values of a series of very protracted insulin analogues. The analogues were administered in solutions containing 3 Zn2+ per insulin hexamer. solubility The solubility of all the NB29 modified insulins mentioned in Table 1, which contain 3 Zn2+ ions per insulin hexamer, exceeds 600 nmol/ml in a neutral (pH 7.5), aqueous, pharmaceutical formulation which further comprises 0.3% phenol as preservative, and 1.6% glycerol to achieve isotonicity. 600 nmol/ml is the concentration of human insulin found in the 100 lU/ml compositions usually employed in the clinic. The e-B29 amino group can be a component of an amide bond, a sulphonamide bond, a carbamide, a thiocarbamide, or a carbamate. The lipophilic substituent carried by the C-B29 amino group can also be an alkyl group. Pharmaceutical compositions containing a human insulin derivative according to the present invention may be administered parenterally to patients in need of such a treatment. Parenteral administration may be performed by subcutaneous, intramuscular or intravenous injection by means of a syringe, optionally a pen-like syringe. Alternatively, parenteral administration can be performed by means of an infusion pump. A further option is a composition which may be a powder or a liquid for the administration of the human insulin derivative in the form of a nasal spray. The injectable human insulin compositions of the invention can be prepared using the conventional techniques of the pharmaceutical industry which involves dissolving and mixing the ingredients as appropriate to give the desired end product. Thus, according to one procedure, the human insulin derivative is dissolved in an amount of water which is somewhat less than the final volume of the composition to be prepared. An isotonic agent, a preservative and a buffer is added as required and the pH value of the solution is adjusted - if necessary - using an acid, e.g. hydrochloric acid, or a base, e.g. aqueous sodium hydroxide as needed- Finally, the volume of the solution is adjusted with water to give the desired concentration of the ingredients. Examples of isotonic agents are sodium chloride, mannitol and glycerol. Examples of preservatives are phenol, m-cresol, methyl p- hydroxybenzoate and benzyl alcohol. Examples of suitable buffers are sodium acetate and sodium phosphate. composition for nasal administration of an insulin derivative according to the present invention may, for example, be prepared as described in European Patent No. 272097 (to Novo Nordisk A/S). The insulin compositions of this invention can be used in the treatment of diabetes. The optimal dose level for any patient will depend on a variety of factors including the efficacy of the specific human insulin derivative employed, the age, body weight, physical activity, and diet of the patient, on a possible combination with other drugs, and on the severity of the case of diabetes. It is recommended that the daily dosage of the human insulin derivative of this invention be determined for each individual patient by those skilled in the art in a similar way as for known insulin compositions. Where expedient, the human insulin derivatives of this invention may be used in mixture with other types of insulin, e.g. human insulin or porcine insulin or insulin analogues with a more rapid onset of action. Examples of such insulin analogues are described e.g. in the European patent applications having the publication Nos. EP 214826 (Novo Nordisk A/S), EP 375437 (Novo Nordisk A/S) and EP 383472 (Eli Lilly & Co,). The present invention is further illustrated by the following examples which, however, are not to be construed as limiting the scope of protection. The features disclosed in the foregoing description and in the following examples may, both separately and in any combination thereof, be material for realizing the invention in diverse forms thereof. EXAMPLES Flasmids and DMA naterial All expression plasmids are of the cPOT type. Such plasmids are described in EP patent application No. 171 142 and are characterized in containing the Schizosaccharomyces pombe triose phosphate isomerase gene (POT) for the purpose of plasmid selection and stabilization. A plasmid containing the POT-gene is available from a deposited E. coli strain {ATCC 39685). The plasmids furthermore contain the S* ggrevisiae triose phosphate isomerase promoter and terminator (P^p, and T1p1). They are identical to pMT742 (Egel-Mitanl, M. et al., Gene 73 (1988) 113-120) (see Fig. 1) except for the region defined by the ECoRI-Xbal restriction sites encompassing the coding region for signal/leader/product. Synthetic DNA fragments were synthesized on an automatic DNA synthesizer (Applied Biosystems model 3BOA) using phosphoramidite chemistry and commercially available reagents (Beaucage, S.L. and Caruthers, M.H., Tetrahedron Letters 22 (1981) 1859-1869) . All other methods and materials used are common state of the art knowledge (see, e.g. Sambrook, J., Fritsch, E.F. and Maniatis, T., Molecular Cloning; A Laboratory Napual. Cold Spring Harbor Laboratory Press, New York, 1989). Analytical Molecular masses of the insulins prepared were obtained by MS (mass spectroscopy), either by PDMS (plasma desorption mass spectrometry) using a Bio-Ion 20 instrument (Bio-Ion Nordic AB, Uppsala, Sweden) or by ESMS (electrospray mass spectrometry) using an API III Biomolecular Mass Analyzer (Perkin-Elner Sciex Instruments, Thornhill, Canada), EXAMPLE 1 Synthesis of AlaA21 AspB3 human insulin precursor from Yeast strain yEA002 using the LaC212spx3 signal/leader. 58.5 ul of water One cycle was performed: 94°c for 45 sec, 49°c for 1 min, 72'C for 2 min. Subsequently, 5/xl of oligonucleotides #16 and #126 was added and 15 cycles were performed: 94"C for 45 sec, 45°c for 1 min, 72'C for 1.5 min. The PCR mixture was loaded onto a 2.5 % agarose gel and subjected to electrophoresis using standard techniques (Sambrook et al.. Molecular cloning, Cold Spring Harbour Laboratory Press, 1989). The resulting DNA fragment was cut out of the agarose gel and isolated using the Gene Clean Kit (Bio 101 Inc., PO BOX 2284, La Jolla, CA 92038, USA) according to the manufacturer's instructions. The purified PCR DNA fragment was dissolved in 10 fxl of water and restriction endonuclease buffer and cut with the restriction endonucleases N col and Xba I according to standard techniques, run on a 2.5% agarose gel and purified using the Gene Clean Kit as described. The plasmid pAK188 consists of a DNA sequence of 412 bp composed of a EcoRI/NcoI fragment encoding the synthetic yeast signal/leader gene LaC212spx3 (described in Example 3 of HO 89/02463) followed by a synthetic Ncol/Xbal fragment encoding the insulin precursor MI5, which has a SerAspAspAlaLys bridge connecting the B29 and the Al amino acid residues (see SEQ ID NOS. 14, 15 and 16), inserted into the EcoRI/Xbal fragment of the vector (phagemid) pBLUESCRIPT IIsk(+/-) (Stratagene, USA). The plasmid pAK188 is shown in Fig. 1. The plasmid pAK188 was also cut with the restriction endonucleases Ncol and Xbal and the vector fragment of 3139 bp isolated. The two DNA fragments were ligated together using T4 DNA ligase and standard conditions (Sambrook et al.. Molecular Cloning, Cold Spring Harbour Laboratory Press, 1989). The ligation mixture was transformed into a competent E. coli strain (R-, M+) followed by selection for ampicillin resistance. Plasmids were isolated from the resulting E. coli colonies using standard DNA miniprep technique (Sambrook et al., Molecular Cloning, Cold Spring Harbour Laboratory Press, 1989), checked with appropriate restrictions endonucleases i.e. EcoRI, Xba I, Ncol and Hpal. The selected plasmid was shown by DNA sequencing analyses (Sequenase, U.S. Biochemical Corp.) to contain the correct sequence for the Ala^^, Asp^^ human insulin precursor and named pEA5.3. The plasmid pKFN1627 is an E. coli - S. cerevisiae shuttle vector, identical to plasmid pKFN1003 described in EP patent No, 375718, except for a short DNA sequence upstream from the unique Xbal site. In pKFNloOB, this sequence is a 178 bp fragment encoding a synthetic aprotinin gene fused in-frame to the yeast mating factor alpha l signal-leader sequence. In pKFN1627, the corresponding 184 bp sequence encodes the insulin precursor MIS (Glu^^, Glu^^Sj (i.e. B(l-29, G1U^\G1U"^)- SerAspAspAlaLys-A(l-21) fused in-frame to the mating factor alpha 1 sequence (see SEQ ID NOS. 17, 18 and 19). The vector pKFN1627 is shown in Fig. 1. pEA5.3 was cut with the restriction endonucleases EcoRI and Xbal and the resulting DNA fragment of 412 bp was isolated. The yeast expression vector pKFN1627 was cut with the restriction endonucleases Ncol and Xbal and with Ncol and EcoRI and the DNA fragment of 9273 bp was isolated from the first digestion and the DNA fragment of 1644 bp was isolated from the second. The 412 bp EcoRI/Xbal fragment was then ligated to the two other fragments, that is the 9273 bp Ncol I/Xbal fragment and the 1644 bp NcoI/EcoRI fragment using standard techniques. The ligation mixture was transformed into E. coli as described above. Plasmid from the resulting E. coli was isolated using standard techniques, and checked with appropriate restriction endonucleases i.e. EcoRI, Xbal, Ncol, Hpa I. The selected plasmid was shown by DNA sequence analysis (using the Sequenase kit as described by the manufacturer, U.S. Biochemical) to contain the correct sequence for the Ala21 AspB3 human insulin precursor DNA and to be inserted after the DNA encoding the liaC212spx3 signal/leader DNA. The plasmid was named pEA5.3.2 and is shown in Fig. 1. The DNA sequence encoding the LaC2l2spx3 signal/leader/Ala^-'- Asp^^ human insulin precursor complex and the amino acid sequence thereof are SEQ ID NOS. 20, 21 and 22. The plasmid pEA5.3.2 was transformed into S^. h cerevisiae strain MT663 as described in European patent ' application having the publication No. 214826 and the resulting strain was named yEA002. The DNA encoding GlyA21 AspB3 human insulin precursor was constructed in the same manner as described for the DNA encoding Ala^^ Asp^^ human insulin precursor in Example 1. The DNA sequence encoding the LaC212spx3 signal/leader/Gly^-- Asp^-' human insulin precursor complex and the amino acid sequence thereof are SEQ ID NOS. 26, 27 and 28. The plasmid pEAl.5.6 was shown to contain the desired sequence, transformed into £j. cerevisiae strain MT663 as described in Example 1 and the resulting strain was named yEA007. The DNA encoding Gly^ ^ Thr^-' human insul in precursor was constructed in the same manner as described for the DNA encoding Ala^^ Asp^-^ human insulin precursor in Example l. The DNA sequence encoding the LaC212spx3 signal/leader/Gly^^^ Thr®"' human insulin precursor complex and the amino acid seguence thereof are SEQ ID NOS. 29, 30 and 31. The plasmid pEA4.4.11 was shown to contain the desired DNA seguence, transformed into S. cerevisiae strain MT663 as described in Example l and the resulting strain was named yEA006, a DM sequence of 412 bp encoding the synthetic yeast signal'leader LaC2l2spx3 (described in Example 3 of WO 89/02463) followed by the insulin precursor MIS (see SEQ ID NOS. 14, 15 and 16) inserted into the vector (phagemld) pBLUESCRIPT IIsk(+/-) (Stratagene, USA). A total of 16 cycles were performed, each cycle comprising 1 minute at 95C; 1 minute at 40'C; and 2 minutes at 72'C, The PCR mixture was then loaded onto a 2% agarose gel and subjected to electrophoresis using standard techniques. The resulting DNA fragment was cut out of the agarose gel and isolated using the Gene Clean kit (Bio 101 Inc., PO BOX 2284, La Jolla, CA 92038, USA) according to the manufacture's instructions. The purified PCR DNA fragment was dissolved in 10 jxl of water and restriction endonuclease buffer and cut with the restriction endonucleases Hindlll and Xbal according to standard techniques. The Hindlll/Xbal DNA fragment was purified using The Gene Clean Kit as described. The plasmid pAK406 consists of a DNA sequence of 520 bp comprising an EcoRI/Hindlll fragment derived from pMT636 (described in WO 90/10075) encoding the yeast alpha factor l eader and part of the insulin precursor ligated to the Hindlll/Xbal fragment from pAK188 encoding the rest of the insulin precursor MIS (see SEQ ID NOS. 32, 33 and 34) inserted into the vector cPOT. The vector pAK406 is shown in Fig. 2. The plasmid pAK233 consists of a DNA sequence o£ 412 bp encoding the synthetic yeast signal/leader LaC212spx3 (described in Example 3 of WO 89/02463) followed by the gene for the insulin precursor B(l-29)-GluLysArg-A(l-2l) (A21-Gly) (see SEQ ID NOS, 35, 36 and 37) inserted into the vector cPOT. The plasmid pAK233 is shown in Fig. 2. The plasmid pAK233 was cut with the restriction endonucleases Ncol and Xbal and the vector fragment of 9273 bp Isolated. The plasmid pAK406 was cut with the restriction endonucleases Ncol and Hindlll and the vector fragment of 2012 bp isolated. These two DNA fragments were ligated together with the Hindlll/Xbal PCR fragment using T4 DNA ligase and standard conditions. The ligation mixture was then transformed into a competent E. coli strain (R-, M+) followed by selection for ampicillin resistance. Plasmids were isolated from the resulting E. coli colonies using a standard DNA miniprep technique and checked with appropriate restriction endonucleases i.e. EcoRI, Xbal, Ncol, Hindlll. The selected plasmid was shown by DNA sequencing analyses to contain the correct sequence for the Arg° single chain human insulin precursor DNA and to be inserted after the DNA encoding the S. cerevisiae alpha factor DNA. The plasmid was named pEAlOS and is shown in Fig. 2. The DNA sequence encoding the alpha factor leader/ArgB31 single chain human insulin precursor complex and the amino acid sequence thereof are SEQ ID NOS. 38, 39 and 40. The plasmid pEA 108 was transformed into S. cerevisiae strain MT663 as described in Example l and the resulting strain was named yEAlOS. B) The following Polymerase Chain Reaction (PCR) was performed using the Gene Amp PCR reagent kit (Perkin Elmer, 761 Main Avewalk, CT 06859, USA) according to the manufacturer's instructions. In all cases, the PCR mixture was overlayed with 100 ul of mineral oil (Sigma Chemical Co., St. Louis, MO, USA) 5 ul of oligonucleotide #220 (100 pmol) 5 ul of oligonucleotide #307 (100 pmol) 10 ul of lOX PCR buffer 16 ul of dNTP mix 0.5 ul of Taq enzyme 0.2 ul of pEAlOS plasmid as template (0.1 ug DNA) 63 ul of water A total of 16 cycles were performed, each cycle comprising 1 minute at 95'C; 1 minute at 40'C; and 2 minutes at 72'C. The PCR mixture was then loaded onto an 2% agarose gel and subjected to electrophoresis using standard techniques. The resulting DNA fragment was cut out of the agarose gel and isolated using the Gene Clean kit (Bio 101 Inc., PO BOX 2284, La Jolla, CA 92038, USA) according to the manufacture * s instructions. The purified PCR DNA fragment was dissolved in 10 ul of water and restriction endonuclease buffer and cut with the restriction endonucleases Ncol and Xbal according to standard techniques. The Ncol/Xbal DNA fragment was purified using The Gene Clean Kit as described. The plasmid pAK401 consists of a DNA sequence of 523 bp composed of an EcoRI/NcoI fragment derived from pMT636 (described in WO 90/10075) (constructed by by introducing a Ncol site in the 3'-end of the alpha leader by site directed mutagenesis) encoding the alpha factor leader followed by a Ncol/Xbal fragment from pAK188 encoding the insulin precursor MI5 (see SEQ ID NOS. 41, 42 and 43) inserted into the vector (phagemid) pBLUESCRIPT IIsk(+/-) (Stratagene, USA). The plasmid pAK40l is shown in Fig. 3. The plasmid pAK40l was cut with the restriction endonucleases Ncol and Xbal and the vector fragment of 3254 bp isolated and ligated together with the Ncol/Xbal PCR fragment. The ligation mixture was then transformed into a competent E. coli strain and plasmids were isolated from the resulting E. coli colonies using a standard DNA miniprep technique and checked with appropriate restriction endonucleases i.e. EcoRI, Xbal, Ncol. The selected plasmid, named pll3A (shown in Fig. 3), was cut with EcoRI and Xbal and the fragment of 535 bp Isolated. The plasmid pAK233 was cut with the restriction endonucleases Ncol and Xbal, and with EcoRl/Ncol and the fragments of 9273 and 1644 bp isolated. These two DNA fragments were ligated together with the EcoRI/Xbal fragment from pll3A using T4 DKA ligase and standard conditions. The ligation mixture was then transformed into a competent E. coli strain (R-, M+) followed by selection for ampicillin resistance. Plasmids were isolated from the resulting E. coli colonies using a standard DNA miniprep technique and checked with appropriate restriction endonucleases i.e. EcoRI, Xbal, Ncol, Hindlll. The selected plasmid was shown by DNA sequencing analyses to contain the correct sequence for the ArgB31 single chain human insulin precursor DNA with the N-terminal extension GluGluAlaGluAlaGluAlaArg and to be inserted after the DNA encoding the 5. cerevisiae alpha factor DNA. The plasmid was named pEA113 and is shown in Fig. 3. The DNA sequence encoding the alpha factor leader/ArgB-1 ArgB31 single chain human insulin precursor having an N-terminal extension (GluGluAlaGluAlaGluAlaArg) and the amino acid sequence thereof are SEQ ID NOS. 44, 45 and 46. The plasmid pEA113 was transformed into S. cerevisiae strain MT663 as described in Example 1 and the resulting strain was named yEA113. EXAMPLE 6 Synthesis of Arg Arg single chain human insulin precursor having an N-terminal extension (GluGluAlaGluAlaGluAlaGluArg) from Yeast strain yEA136 using the alpha factor leader. The following oligonucleotide was synthesized: #389 5'-GCTAACGTCGCCATGGCTAAGAGAGAAGAAGCTGAAGCGAAG CTGAAAGATTCGTTAACCAACAC-3' (SEQ ID NO:13) The following PCR was performed using the Gene Amp PCR reagent kit 5 u1 of oligonucleotide #220 (100 pmol) 5 Ml of oligonucleotide #389 (100 pmol) 10 Ml of lOX PCR buffer 16 Ml of dNTP mix 3.5 Ml of Tag enzyme 2 Ml of pEA113 plasmid as template (0.5 ug DNA) S3 Ml of water \ total of 12 cycles were performed, each cycle comprising 1 ninute at 95°C; 1 minute at 37°C; and 2 minutes at 72°C. The DNA encoding alpha factor leader/ArgB-1 ArgB31 single chain luman insulin precursor having an N-terminal extension (GluGluAlaGluAlaGluAlaGluArg) was constructed in the same nanner as described for the DNA encoding alpha factor Leader/Arg Arg single chain human insulin precursor laving an N-terminal extension (GluGluAlaGluAlaGluAlaArg) in Example 5. The plasmid was named pEA13 6. The DNA sequence ancoding the alpha factor leader/ArgB-1 ArgB31 single chain luman insulin precursor having an N-terminal extension (GluGluAlaGluAlaGluAlaGluArg) and the amino acid sequence thereof are SEQ ID NOS. 47, 48 and 49. The plasmid pEA136 was bransformed into S. cerevisiae strain MT663 as described in Example I and the resulting strain was named yEA136. EXAMPLE 7 Synthesis of (A1,B1)-diBoc human insulin. 5 g of zinc-free human insulin was dissolved in 41.3 ml of DMSO. To the solution was added 3.090 ml of acetic acid. The reaction was conducted at room temperature and initiated by addition of 565 mg of di-tert-butyl pyrocarbonate dissolved in 5.650 ml of DMSO. The reaction was allowed to proceed for 5k hour and then stopped by addition of 250 u1 of ethanolamine. The product was precipitated by addition of 1500 ml of acetone. The precipitate was isolated by centrifugation and dried in vacuum. A yield of 6.85 g material was obtained. (Al,Bl)-diBoc insulin was purified by reversed phase HPLC as follows: The crude product was dissolved in 100 ml of 25% ethanol in water, adjusted to pH 3,0 with HCl and applied to a column (5 cm diameter, 30 cm high) packed with octadecyJdimethylsilyl-substituted silica particles (mean particle size 15 um, pore size 100 A) and equilibrated with elution buffer. The elution was performed using mixtures of ethanol and 1 mM aqueous HCl, 0.3 M KCl at a flow of 2 1/h. The insulin was eluted by increasing the ethanol content from 30% to 45%. The appropriate fraction was diluted to 20% ethanol and precipitated at pH 4.8. The precipitated material was isolated by centrifugation and dried in vacuum. Thus 1.701 g of (A1,B1)-diBoc human insulin was obtained at a purity of 94.5%. EXAMPLE 0 Synthesis of (N°"-benzoyl human insulin)^, 3Zn^. 400 mg of (A1,B1)-diBoc human insulin was dissolved in 2 ml of DMSO. To the solution was added 748 ^1 of a mixture of N-methylmorpholine and DMSO (1:9, v/v). The reaction was conducted at 15'C and initiated by addition of 14.6 mg of benzoic acid N-hydroxysuccinimide ester dissolved in 132 ^1 DMF. The reaction was stopped after 2 hours by addition of 100 ml of acetone. The precipitated material was isolated by centrifugation and dried in vacuum. 343 mg of material was collected. The Boc protecting groups were eliminated by addition of 4 ml of TFA. The dissolved material was incubated for 30 minutes and then precipitated by addition of 50 ml of acetone. The precipitate was isolated by centrifugation and dried in vacuum. NB29-benzoyl human insulin was purified by reversed phase HPLC as described in Example 7. A yield of 230 mg was obtained. Recrystallization from 15% aqueous ethanol containing 6 mM 2n^* and 50 mM citrate at pH 5.5 gave crystals of the title compound which were isolated by centrifugation and dried in vacuum. The yield was 190 mg. Molecular mass, found by MS: 5911, theory: 5911. EXAMPLE 9 Synthesis of (N'^^-lithocholoyl human insulin)^, SZn^. 400 mg of (Al,Bl)-diBoc human insulin was dissolved in 2 ml of DMSO. To the solution was added 748 ^il of a mixture of N-methylmorpholine and DMSO (1:9, v/v). The reaction was conducted at 15' C and initiated by addition of 31.94 mg of lithocholic acid N-hydroxysuccinimide ester dissolved in 300 nl of DMF. The reaction was stopped after 2 hours by addition of 100 ml of acetone. The precipitated material was isolated by centrifugation and dried in vacuum. 331 mg of material was obtained, The Boc protecting groups were eliminated by addition of 4 ml of TFA. The dissolved material was incubated for 30 minutes and then precipitated by addition of 50 ml of acetone. The precipitate was isolated by centrifugation and dried in vacuum. The yield was 376 mg. B29-lithocholoyl insulin was purified by reversed phase HPLC as described in Example 7. A final yield of 67 mg was obtained at a purity of 94%. Recrystallization from 15% aqueous ethanol containing 6 mM Zn2+ and 50 mM citrate at pH 5.5 gave crystals of the title compound which were isolated by centrifugation and dried in vacuum. The yield was 49 mg. Molecular mass, found by MS: 6160, theory: 6166. EXAMPLE 10 Synthesis of (NB29-decanoyl human insulin)6, 3Zn2+. 400 mg of (Al,Bl}-diBoc human Insulin was dissolved in 2 ml of DMSO. To the solution was added 748 ul of a mixture of N- methylraorpholine and DMSO (1:9/ v/v). The reaction was conducted at 15C and initiated by addition of 18.0 mg of decanoic acid N-hydroxysuccinimide ester dissolved in 132 fil of DMF. The reaction was stopped after 60 minutes and the product precipitated by addition of 100 ml of acetone. The precipitated material was isolated by centrifugation and dried in vacuum. 420 mg of intermediate product was collected. The Boc protecting groups were eliminated by addition of 4 ml of TFA. The dissolved material was incubated for 30 minutes and the product was then precipitated by addition of 50 ml of acetone. The precipitate was isolated by centrifugation and dried in vacuum. The yield of crude product was 420 mg. The crude product was purified by reversed phase HPLC as described in Example 7. A final yield of 254 mg of the title product was obtained. The purity was 96.1%. Recrystallization from 15% aqueous ethanol containing 6 mM Zn^* and 50 mM citrate at pH 5.5 gave crystals of the title compound which were isolated by centrifugation and dried in vacuum. The yield was 217 mg. Molecular mass, found by MS: 5962, theory: 5962. EXAMPLE 11 Synthesis of des(B30) human insulin. Synthesis of des(B30) human insulin was carried out as described by Markussen (Methods in diabetes research, Vol. I, Laboratory methods, part B, 404-410. Ed: J. Lamer and S. Phol, John Wiley & Sons, 1984). 5 g of human insulin was dissolved in 500 ml of water while the pH value of the solution was kept at 2.6 by addition of 0.5 M sulphuric acid. Subsequently, the insulin was salted out by addition of 100 g of ammonium sulphate and the precipitate was isolated by centrifugation. The pellet was dissolved in 800 ml of 0.1 M ammonium hydrogen carbonate and the pH value of the solution was adjusted to 8.4 with 1 H ammonia. 50 mg of bovine carboxypeptidase A was suspended in 25 ml of water and isolated by centrifugation. The crystals were suspended in 25 ml of water and 1 M ammonia was added until a clear solution was obtained at a final pH of 10. The carboxypeptidase solution was added to the insulin solution and the reaction was allowed to proceed for 24 hours. A few drops of toluene were added to act as preservative during the reaction. After 24 hours the des(B30) human insulin was crystallized by successive addition of 80 g of sodium chloride while the solution was stirred. The pH value was then adjusted to 8.3 and the crystallization was allowed to proceed for 20 hours with gentle stirring. The crystals were isolated on a 1.2 ^m filter, washed with 250 ml of ice cold 2-propanol and finally dried in vacuum. EXAMPLE 12 Synthesis of (A1,B1)-diBoc des(B30) human insulin. The title compound was synthesized by a method similar to that described in Example 7, using des(B30) porcine insulin as the starting material. The crude product was precipitated by acetone and dried in vacuum. The (A1,B1)-diBoc (ies(B30) human insulin was purified by reversed phase HPLC as described in Example 7. EXAMPLE 13 Synthesis of NB29-decanoyl des(B30) human insulin. 400 ng of (Al,Bl)-diBoc des(B30) human insulin was used as starting material for the synthesis of N*^-decanoyl des(B30) human insulin, following the procedure described in Example 10. The crude product was precipitated by acetone, dried in vacuum and deprotected using TFA. The resulting product was precipitated by acetone and dried in vacuum. N^^-decanoyl des(B30} human insulin was then purified by reversed phase HPLC as described in Example 10. Molecular mass, found by MS: 5856, theory: 5861. EXAMPLE 14 Synthesis of NB29-dodecanoyl des(B30) human insulin, a. IBUDObiliBatioD of A. Ivticus prctease 13 mg of A. lytipus protease, dissolved in 5 ml of aqueous 0.2 M NaHCO3 buffer, pH 9.4, was mixed with 4 ml of settled MiniLeak Medium gel, which had been washed with the same buffer (MiniLeak is a divinylsulfone activated Sepharose CL 6B, obtained from KemEnTec, Copenhagen) . The gel was kept in suspension by gentle stirring for 24 hours at room temperature. Then, the gel was isolated by filtration, washed with water, and suspended in 20 ml of 1 M ethanolamine buffer, pH 9.4, and kept in suspension for 24 hours at room temperature. Finally, the gel was washed with water followed by 0.1 M acetic acid and stored at 4°C. The enzyme activity in the filtrate was 13% of that in the initial solution, indicating a yield in the immobilization reaction of about 87%. b. Imnobilizatlon of porcine trypsin Porcine trypsin was immobilized to MiniLeak* Low to a degree of substitution of 1 mg per ml of gel, using the conditions described above for immobilization of A. lyticus. To 200 mg of Glu(GluAla)3Arg-B(l-29)-ThrArg-A(l-21) single-chain human insulin precursor, dissolved in 20 ml of 0.1 M NaHCO, buffer, pH 9.0, was added 4 ml of the gel carrying the immobilized A. lyticus protease. After the gel had been kept in suspension in the reaction mixture for 6 hours at room temperature the hydrolysis was complete, rendering GluCGluAla),-Arg-B(l-29), ThrArg-A(l-21) human insulin (the reaction was followed by reversed phase HPLC). After the hydrolysis, the gel was removed by filtration. To the filtrate was added 5 ml of ethanol and 15 ^L of 1 ZnClj and the pH was adjusted to 5.0 using HCl. The precipitation of the product was completed on standing overnight at 4°C with gentle stirring. The product was isolated by centrifugation. After one washing with 1 ml of ice cold 20% ethanol and drying in vacuo the yield was 190 mg. 190 mg (30 umol) of Glu(GluAla) 3Arg-B(l-29) , ThrArg-A(l-21) insulin was dissolved in 1 ml of DMSO and 1.05 ml of a 0.572 M solution of N,N-diisopropylethylamine in DMF. The solution was cooled to 15°C and 36 mg (120 umol) of dodecanoic acid N- hydroxysuccinimide ester dissolved in 0.6 ml of DMSO was added. The reaction was completed within 24 hours. The lipophilic title compound was not isolated. The product from the previous step, d., contained in approximately 2,65 mX of DMSO/DMF/N,N-diisopropylethylamine was diluted with 10.6 ml of a 50 mM glycine buffer comprising 20% ethanol and the pH adjusted to 10 with NaOH. After standing for 1 hour at room temperature 1 ml of MiniLeak gel, carrying 1 mg of iroroobilized trypsin per ml of gel, was added. The reaction mixture was stirred gently for 48 hours at room temperature. In order to isolate the desired product, the reaction mixture was applied to a reversed phase HPLC column (5 cm in diameter, 30 cm high) , packed with octadecyldimethylsilyl-substituted silica particles (mean particle size 15 nm, pore size 100 A). For the elution was used 20 fliM Tris/HCl buffers, adjusted to pH 7.7 and comprising an increasing concentration of ethanol, from 40% to 44% (v/v), at a rate of 2000 ml/h. The major peak eluting at about 43-44% of ethanol contained the title compound. The fractions containing the major peak were pooled, water was added to reduce the ethanol concentration to 20% (v/v), and the pH was adjusted to 5.5. The solution was left overnight at -20"C, whereby the product precipitated. The precipitate was isolated by centrifugation at -8 ° C and dried in vacuo. The yield of the title compound was 90 mg. Molecular mass, found by MS: 5892, theory: 5890. EXAMPLE 15 Synthesis of NB29-(N'myristoyl-Q(-glutamyl) human insulin. 500 mg of (Al,Bl)-diBoc human insulin was dissolved in 2.5 ml of DMSO and 428 ul of ethyl diisopropylamine, diluted with 2.5 ml of DMSO/DMF 1/1 (v/v), was added. The temperature was adjusted to 15'C and 85 mg of N-myristoyl-Glu(OBut) N-hydroxysuccinimide ester, dissolved in 2.5 ml of DMSO/DMF 1/1 (v/v), was added. After 30 min the reaction mixture was poured into 60 ml of water, the pH adjusted to 5 and the precipitate isolated by centrifugation. The precipitate was dried in vacuo. The dried reaction mixture was dissolved in 25 ml of TFA, and the solution was left for 30 min at room temperature. The TFA was removed by evaporation in vacuo. The gelatinous residue was dissolved in 60 ml of water and the pH was adjusted to 11,2 using concentrated ammonia. The title compound was crystallized from this solution by adjustment of the pH to 8.5 using 6 N HCl. The product was isolated by centrifugation, washed once by 10 ml of water, and dried in vacuo. Yield 356 ng. Purity by HPLC 94%. The product of this example is thus human insulin wherein the e-amino group of LysB29 has a substituent of the following structure: CH3 (CH2)12CONHCH (CH2CH2COOH) CO-. Molecular mass, found by MS: 6146, theory: 6148. EXMPLE 16 Synthesis of NB29-undecanoyl des(B30) human insulin. The title compound was synthesized analogously to NB29-dodecanoyl des(B30) human Insulin as described in Example 14, by using undecanoic acid N-hydroxysuccinimide ester instead of dodecanoic acid N-hydroxysuccinimide ester. Molecular mass of the product found by MS: 5876, theory: 5876. EXAMPLE 17 Synthesis of NB29-tridecanoyl des(B30) human insulin. The title compound was synthesized analogously to NB29-dodecanoyl des{B30) human insulin as described in Example 14, by using tridecanoic acid N-hydroxysuccinimide ester instead of dodecanoic acid N-hydroxysuccinimide ester. Molecular mass of the product found by MS: 5899, theory: 5904. EXAMPLE 18 Synthesis of NB29-inyristoyl des(B30) human insulin. The title compound was synthesized analogously to N^-dodecanoyl des(B30) human insulin as described in Example 14, by using myristic acid N-hydroxysuccinimide ester instead of dodecanoic acid N-hydroxysuccinimide ester. Molecular mass of the product found by MS: 5923, theory: 5918. EXAMPLE 19 Synthesis of NB29-palmitoyl des(B30) human insulin. The title compound was synthesized analogously to NB29-dodecanoyl des(B30) human insulin as described in Example 14, by using palmitic acid N-hydroxysuccinimide ester instead of dodecanoic acid N-hydroxysuccinimide ester. Molecular mass of the product found by MS: 5944, theory: 5946. EXAMPLE 20 Synthesis of NB29-suberoyl-D-thyroxine human insulin. a. Preparation of N-(succinimidYlsuberoyl)-D-thyroxine. Disuccinimidyl suberate (1.0 g, Pierce) was dissolved in DMF (50 ml), and D-thyroxine (2.0 g, Aldrich) was added with stirring at 20C. The thyroxine slowly dissolved, and after 20 hours the solvent was removed by evaporation in vacuo. The oily residue was crystallized from 2-propanol to yield 0.6 g of N-(succinijnidylsuberoyl)-D-thyroxine, m.p. 128-133'C. b. Rgagtj.pn o£ lAl. Bl) -diBoc human insul in with H- fBuccinimidvlsuberovll-D-thyroxine. (A1,B1)-diBoc human insulin (200 mg) was dissolved in dry DMF (10 ml) by addition of tri ethyl amine (20 M1) at room temperature- Then, N-(succinimidylsuberoyl)-D-thyroxine (80 rag) was added. The reaction was monitored by reversed phase HPLC and when the reaction was about 90% complete, the solvent was removed in vacuo. To the evaporation residue, anhydrous trifluoroacetic acid (5 ml) was added, and the solution was kept for 1 hour at room temperature. After removal of the trifluoroacetic acid in vacuo, the residue was dissolved in a mixture of IM acetic acid (5 ml) and acetonitrile (1.5 ml), purified by preparative reversed phase HPLC and desalted on a PD-IO column. The yield of N°^-suberoyl-D-thyroxine human insulin was 50 mg. The product of this example is thus human insulin wherein the €-amino group of Lys^^ has a substituent of the following structure: Thyrox-CO(CHg)^co-, wherein Thyrox is thyroxine which is bound to the octanedioic acid moiety via an amide bond to its a-amino group. Molecular mass of the product found by MS: 6724, theory: 6723. SZ&MFLE 21 Synthesis of N^^-(2-succinylamido)myristic acid human insulin. a. Preparation of a-aminomvristic acid methyl ester.HCl. To methanol (5 ml, Merck) at -10"C, thionyl chloride (0.2 ml, Aldrich) was added dropwise while stirring vigorously. Then, a- aminomyristic acid (0.7 g, prepared from the a-bromo acid by reaction with ammonia) was added. The reaction mixture was stirred at room temperature overnight, and then evaporated to dryness. The crude product (0.7 g) was used directly in step b. b. Preparation of N-succinoyl-a-aminomyristic acid methyl ester. a-Aminomyristic acid methyl ester,HCl (0.7 g) was dissolved in chloroform (25 ml, Merck). Triethylamine (0.35 ml, FluJca) was added, foilowed by succinic anhydride (0.3 g, Fluka). The reaction mixture was stirred at room temperature for 2 hours, concentrated to dryness, and the residue recrystallized from ethyl acetate/petroleum ether (1/1). Yield: 0,8 g. c. Preparation of N-(succiniinidylsuccinoyl) -a-aminomyristic acid methyl ester. N-succinoyl-a-aminomyristic acid methyl ester (0.8 g) was dissolved in dry DMF (10 ml, Merck, dried over 4A molecular sieve). Dry pyridine (80 ^1* Merck), and di(N-suc- cinimidyl)carbonate {1.8 g, Fluka) were added, and the reaction mixture was stirred overnight at room temperature. The evaporation residue was purified by flash chromatography on silica gel 60 (Merck), and recrystallized from 2- propanol/petr oleum ether (1/1). Yield of N- (succinimidylsuccinoyl)-a-aminomyristic acid methyl ester: 0.13 g, m.p. 64-66'C. d. Reaction of (Al, Bl) -dj.Boc human insul in with N- fsuccinimidylsuccinoyl)-a-aminomyristic acid methyl ester. The reaction was carried out as in Example 20 b., but using N- (succinimidylsuccinoyl)-a-aminomyristic acid methyl ester (16 mg) instead of N-(succinimidylsuberoyl)-D-thyroxine. After removal of the trifluoroacetic acid in vacuo, the evaporation residue was treated with O.IM sodium hydroxide at O'C to saponify the methyl ester. When the saponification was judged to be complete by reversed phase HPLC, the pH value in the solution was adjusted to 3, and the solution was lyophilized. After purification by preparative reversed phase HPLC and desalting on a PD-10 column, the yield of NB29-(2- succinylamido)myristic acid human insulin was 39 mg. The product of this example is thus human insulin wherein the €-amino group of Lys"^ has a substituent of the following structure: CH3 (CHj) „CH (COOH) NHCOCHjCHjCO-. Molecular mass of the product found by MS: 6130, theory: 6133. EXAHPLE 22 Synthesis of N'^^-octyloxycarbonyl human insulin. The synthesis was carried out as in Example 20 b., but using n- octyloxycarbonyl N-hydroxysuccinimide (9 mg, prepared from n- octyl chloroformate (Aldrich) and N-hydroxysuccinimide), instead of N-(succiniraidylsuberoyl)-D-thyroxine. The yield of NB29-octyloxycarbonyl human insulin was 86 mg. The product of this example is thus human insulin wherein the e-amino group of LysB29 has a substituent of the following structure: CH2(CH2)7OCo Molecular mass of the product found by MS: 5960, theory: 5964. EXAMPLE 23 Synthesis of N^"-(2-succinylamido)palmitic acid human insulin. a. Preparation of N-fsuccinimidylsuccinoyl)-a-amino palmitic acid methyl ester. This compound was prepared as described in Example 21 a.-c, using a-amino palmitic acid instead of a-amino myristic acid. Reaction of (Ai.BD-diBoc human insulin with N=. succinimidvlsuccinovli-a-aminopalmitictic acid methvl ester. The reaction was carried out as in Example 21 d., but using N-{succinimidylsuccinoyl)-a-aminopalmitic acid methyl ester instead of N-(succinimidylsuccinoyl)-a-aminopalmitic acid methyl ester to give NB29-(2-succinylamido)palinitic acid human insulin. The product of this example is thus human insulin wherein the e-amino group of LysB29 has a substituent of the following structure: CHjCCHj) i5CH(COOH)NHCOCH2CH2CO-. EXAMPLE 24 Synthesis of NB29-(2-succinylamidoethyloxy)palmitic acid human insulin. a. Preparation of N-f3UCcinimidyl5uccinovlW2-amimethvloxy palmitic acid methyl ester. This compound was prepared as described in Example 21 a.-c. but using 2-aminoethyloxy palmitic acid (synthesized by the general procedure described by R. TenBrink, J^ Org. Chem. 52 (1987) 418-422 instead of a-amino myristlc acid. b. Reaction of fAl.BlWdiBoc Imfflan insulin with N- (succiniroidvlsuccinovll-2-aminoethvloxvpalmitictic acid methvl ester. The reaction was carried out as in Example 21 d., but using N- (succinimidylsuccinoyl)-2-aminoethyloxypalmitic acid methyl ester instead of N-(succinimidylsuccinoyl) -a-aminomyristic acid methyl ester to give NB29-(2-succinylamidoethyloxy)palmitic acid human insulin. The product of this example is thus human insulin wherein the e-amino group of LysB29 has a substituent of the following structure: CH3(CH2)i3CH(COOH)NHCH2CH20COCH2CH2CO-. EXAMPLE 25 Synthesis of NB29-lithocholoyl-a-glutamyl des(B30) human insulin. The synthesis was carried out as in Example 13 using N-lithocholoyl-L-glutaraic acid a-N-hydroxysuccinimide ester, 7-tert-butyl ester instead of decanoic acid N-hydroxysuccinimide ester. The product of this example is thus des(B30) human insulin wherein the e-amino group of LysB29 has a substituent of the following structure: lithocholoyl-NHCH(CH2CH2COOH)CO-, Molecular mass of the product found by MS: 6194, theory: 6193. EZ&MPLE 26 Synthesis of NB29-3,3 ,5,5'-tetralodothyroacetyl human insulin. The synthesis was carried out as in Example 10 using 3,3',5,5-tetraiodothyroacetlc acid N-hydroxysuccinimlde ester, instead of decanoic acid N-hydroxysuccinlmlde ester. Molecular mass of the product found by MS: 6536, theory: 6538. EXAMPLE 27 Synthesis of NB29-L-thyroxyl human insulin. The synthesis was carried out as in Example 10 using Boc-L-thyroxine N-hydroxysuccinimide ester, instead of decanoic acid N-hydroxysuccinimide ester. Molecular mass of the product found by MS: 6572, theory: 6567, EXAMPLE 28 A pharmaceutical composition comprising 600 nmol/ml of NB29-decanoyl des(B30) human insulin, l/3Zn^ in solution. NB29-decanoyl des(B30) human insulin (1.2 ^mol) was dissolved in water (0.8 ml) and the pH value was adjusted to 7.5 by addition of 0.2 H sodium hydroxide. O.Ol M zinc acetate (60 ^1) and a solution containing 0.75% of phenol and 4% of glycerol (0.8 ml) was added. The pH value of the solution was adjusted to 7.5 using 0.2 M sodium hydroxide and the volume of the solution was adjusted to 2 ml with water. The resulting solution was sterilized by filtration and transferred aseptically to a cartridge or a vial. EXAMPLE 29 A pharmaceutical composition comprising 600 nmol/ml of NB29- decanoyl human insulin, 1/2Zn2+ in solution. 1.2 umol of the title compound was dissolved in water (0,8 ml) and the pH value was adjusted to 7.5 by addition of 0.2 M sodium hydroxide. A solution containing 0.75% of phenol and 1.75% of sodium chloride (0.8 ml) was added. The pH value of the solution was adjusted to 7.5 using 0.2 M sodium hydroxide and the volume of the solution was adjusted to 2 ml with water. The resulting solution was sterilized by filtration and transferred aseptically to a cartridge or a vial. EXAMPLE 30 A pharmaceutical composition comprising 600 nmol/ml of NB29-lithocholoyl human insulin in solution. 1.2 umol of the title compound was suspended in water (0.8 ml) and dissolved by adjusting the pH value of the solution to 8.5 using 0.2 M sodium hydroxide. To the solution was then added 0.8 ml of a stock solution containing 0.75 % cresol and 4% glycerol in water. Finally, the pH value was again adjusted to 8.5 and the volume of the solution was adjusted to 2 ml with water. The resulting solution was sterilized by filtration and transferred aseptically to a cartridge or a vial. WE CLAIM: 1. An insulin derivative having the following sequence: wherein Xaa at positions A21 and B3 are, independently, any amino acid residue which can be coded for by the genetic code except Lys, Arg and Cys; Xaa at position Bl is Phe or is deleted; Xaa at position B30 is (a) a non-codable, lipophilic amino acid having from 10 to 24 carbon atoms, in which case an acyl group of a carboxylic acid with up to 5 carbon atoms is bound to the e-amino group of LysB29, (b) any amino acid residue which can be coded for by the genetic code except Lys, Arg and Cys, in which case the e-amino group of LysB29 has a lipophilic substituent or (c) deleted, in which case the e-amino group of LysB29 has a lipophilic substituent ; and any Zn2* complexes thereof, provided that when Xaa at position B30 is Thr or Ala, Xaa at positions A21 and B3 are both Asn, and Xaa at position Bl is Phe, then the insulin derivative is a Zn2* complex. 2. The insulin derivative according to claim 1, wherein Xaa at positions A21 and B3 are, independently, any amino acid residue which can be coded for by the genetic code except Lys, Arg and Cys; Xaa at position Bl is Phe or is deleted; Xaa at position B30 is a non-codable, lipophilic amino acid having from 10 to 24 carbon atoms and an acyl group is bound to the e-amino group of Lys829, wherein the acyl group is an acyl group of a monocarboxylic acid with up to 4 carbon atoms or of a dicarboxylic acid with up to 5 carbon atoms. 3. The insulin derivative according to claim 1, wherein Xaa at positions A21 and B3 are, independently, any amino acid residue which can be coded for by the genetic code except Lys, Arg and Cys; Xaa at position Bl is Phe or is deleted; Xaa at position B30 is deleted or is any amino acid residue which can be coded for by the genetic code except Lys, Arg and Cys and the e-amino group of Lys829 has a lipophilic substituent which comprises at least 6 carbon atoms. 4. The insulin derivative according to claim 2, wherein Xaa at position B30 is selected from the group consisting of a-amino decanoic acid, a-amino dodecanoic acid, a-amino tetradecanoic acid and a-amino hexadecanoic acid. 5. The insulin derivative according to claim 2, wherein the acyl group bound to the e-amino group of Lys829 is selected from the group consisting of formyl, acetyl, propionyl and n-butyryl. 6. The insulin derivative according to claim 2, wherein the acyl group bound to the e-amino group of Lys829 is an acyl group of succinic acid. 7. The insulin derivative according to claim 3, wherein Xaa at position B30 is deleted. 8. The insulin derivative according to claim 3, wherein Xaa at position B30 is Asp,Glu,or Thr. 9. The insulin derivative according to claim 3, wherein the lipophilic substituent bound to the e-amino group of LysB29 is an acyl group derived from a carboxylic acid having at least 6 carbon atoms. 10. The insulin derivative according to claim 9, wherein the acyl group, which may be branched, comprises a main chain of carbon atoms 8-24 atoms long. 11. The insulin derivative according to claim 9, wherein the acyl group is an acyl group of a fatty acid having at least 6 carbon atoms. 12. The insulin derivative according to claim 9, wherein the acyl group is an acyl group of a linear, saturated carboxylic acid having from 6 to 24 carbon atoms. 13. The insulin derivative according to claim 9, wherein the acyl group is selected from the group comprising dodecanoic acid, tridecanoic acid and tetradecanoic acid. 14. The insulin derivative according to claim 1, wherein Xaa at position A21 is Ala, Gin, Gly or Ser. 15. The insulin derivative according to claim 1, wherein Xaa at position B3 is Asp, Gin or Thr. 16. The insulin derivative according to claim 1, wherein Xaa at position Bl is deleted. 17. The insulin derivative according to claim 3 which is NEB29-tetradecanoyl des(B30) human insulin. 18. An insulin derivative according to claim 3 which is any Zn2+ complex of NsB29- tetradecanoyl des(B30) human insulin. 19. An insulin derivative according to claim 3 which is a Zn2+ complex of NeB29 - tetradecanoyl des(B30) human insulin containing 2, 3, or 4 Zn2+ ions per insulin hexamer. 20. An insulin derivative according to claim 3 which is NEB29-(lithocholoyl-glutamy) des(B30) human insulin. 21. An insulin derivative according to claim 3 which is any Zn2+complex of NEB29 (lithocholoyl-glutamyl) des(B30) human insulin. 22. An insulin derivative according to claim 3 which is a Zn2+ complex of NeB29 (lithocholoyl-glutamyl) des(B30) human insulin containing 2, 3, or 4 Zn2+ ions per insulin hexamer.

Full Text

FIELD OF TEE INVENTION
The present invention relates to novel human insulin derivatives which are soluble and have a protracted profile of action, to a method of providing such derivatives, to pharmaceutical compositions containing them, and to the use of such insulin derivatives in the treatment of diabetes.
BACKGROUND OF THE INVENTION
Many diabetic patients are treated with multiple daily insulin injections in a regimen comprising one or two daily injections of a protracted insulin to cover the basal reguirement
supplemented by bolus injections of a rapid acting insulin to
cover the reguirement related to meals.
Protracted insulin compositions are well Known in the art. Thus, one main type of protracted insulin compositions comprises injectable aqueous suspensions of insulin crystals or
amorphous insulin. In these compositions, the insulin compounds
utilized typically are protamine insulin, zinc insulin or
protamine zinc insulin.

Certain drawbacks are associated with the use of insulin suspensions. Thus, in order to secure an accurate dosing, the insulin particles must be suspended homogeneously by gentle shaking before a defined volume of the suspension is withdrawn from a vial or expelled from a cartridge. Also, for the storage of insulin suspensions, the temperature must be kept within more narrow limits than for insulin solutions in order to avoid lump formation or coagulation.
While it was earlier believed that protamines were non-
immunogenic, it has now turned out that protamines can be
immunogenic in man and that their use for medical purposes may

lead to formation of antibodies (Samuel et al., Studies on the immunogenecity of protamines in humans and experimental animals by means of a micro-complement fixation test, Clin. Exp. Immunol. 11, pp. 252-260 (1978)).
Also, evidence has been found that the protamine-insulin complex is itself immunogenic (Kurtz et al., Circulating IgG antibody to protamine in patients treated with protamine-insulins. Diabetologica 25., pp. 322-324 (1983)). Therefore, with some patients the use of protracted insulin compositions containing protamines must be avoided.
Another type of protracted insulin compositions are solutions having a pH value below physiological pH from which the insulin will precipitate because of the rise in the pH value when the solution is injected. A drawback with these solutions is that the particle size distribution of the precipitate formed in the tissue on injection, and thus the timing of the medication, depends on the blood flow at the injection site and other parameters in a somewhat unpredictable manner. A further irawback is that the solid particles of the insulin may act as a local irritant causing inflammation of the tissue at the site of injection.
Wo 91/12817 (Novo Nordisk A/S) discloses protracted, soluble insulin compositions comprising insulin complexes of zobalt(III). The protraction of these complexes is only intermediate and the bioavailability is reduced.
Human insulin has three primary amino groups; the N-terminal group of the A-chain and of the B-chain and the e-amino group of LysB29. Several insulin derivatives which are substituted in one or more of these groups are known in the prior art. Thus, JS Patent No. 3,528,960 (Eli Lilly) relates to N-carboxyaroyl insulins in which one, two or three primary amino groups of the insulin molecule has a carboxyaroyl group. No specifically NB29-substituted insulins are disclosed.

According to GB Patent No. 1,492.997 (Nat. Res. Dev. Corp.), it has been found that insulin with a carbamyl substitution at N*'^ has an improved profile of hypoglycaemic effect.
JP laid-open patent application No. 1-254699 (Kodama Co., Ltd.) discloses insulin wherein a fatty acid is bound to the amino group of Phe^' or to the e-amino group of Lys^^ or to both of these. The stated purpose of the derivatisation is to obtain a pharmacologically acceptable, stable insulin preparation.
Insulins, which in the B30 position have an amino acid having
(1 at least five carbon atoms which cannot necessarily be coded
for by a triplet of nucleotides, are described in JP laid-open
patent application No. 57-067548 (Shionogi). The insulin
analogues are claimed to be useful in the treatment of diabetes
mellitus, particularly in patients who are insulin resistant
due to generation of bovine or swine insulin antibodies.
By "insulin derivative" as used herein is meant a compound having a molecular structure similar to that of human insulin including the disulfide bridges between CysA7 and CysA7 and between CysA20 and CysB19 and an internal disulfide bridge between cysA6 and CysA11, and which have insulin activity.
However, there still is a need for protracted injectable insulin compositions which are solutions and contain insulins which stay in solution after injection and possess minimal inflammatory and immunogenic properties.
One object of the present invention is to provide human insulin derivatives, with a protracted profile of action, which are soluble at physiological pH values.
Another object of the present invention is to provide a pharmaceutical composition comprising the human insulin derivatives according to the invention.

It is a further object of the invention to provide a method of making the human insulin derivatives of the invention.
SUMMMIY OF THE IliVBllTIOM
Surprisingly, it has turned out that certain human insulin derivatives, wherein the e-amino group of Lys"^ has a lipophilic substituent, have a protracted profile of action and are soluble at physiological pH values.
Accordingly, in its broadest aspect, the present invention relates to an insulin derivative having the following sequence:

wherein
Xaa at positions A21 and B3 are, independently, any amino acid residue which can be coded for by the genetic code except Lys, Arg and Cys; Xaa at position Bl is Phe or is deleted; Xaa at position B30 is (a) a non-codable, lipophilic amino acid having from 10 to 24 carbon atoms, in which case an acyl group of a carboxylic acid with up to 5 carbon atoms is bound to the e-amino group of LygB29^ (b) any amino acid residue which can be coded for by the genetic code except Lys, Arg and Cys, in which case the e-amino group of Lys^^ has a lipophilic substituent or (c) deleted, in which case the c-amino group of Lys^^ has a lipophilic substituent; and any Zn^ complexes thereof,provided that when Xaa at position B30 is Thr or Ala, Xaa at positions A21 and B3 are both Asn, and Xaa at position Bl is Phe, then the insulin derivative is a Zn^* complex.
In one preferred embodiment, the invention relates to a human insulin derivative in which the B3 0 amino acid residue is deleted or is any amino acid residue which can be coded for by the genetic code except Lys, Arg and Cys; the A21 and the B3 amino acid residues are, independently, any amino acid residues which can be coded for by the genetic code except Lys, Arg and Cys; Phe^^ may be deleted; the e-amino group of Lys*^ has a lipophilic substituent which comprises at least 6 carbon atoms; and 2-4 Zn^* ions may be bound to each insulin hexamer with the proviso that when B30 is Thr or Ala and A21 and B3 are both Asn, and Phe^' is not deleted, then 2-4 Zn^* ions are bound to each hexamer of the insulin derivative.
In another preferred embodiment, the invention relates to a human insulin derivative in which the B30 amino acid residue is deleted or is any amino acid residue which can be coded for by the genetic code except Lys, Arg and Cys; the A21 and the B3 amino acid residues are, independently, any amino acid residues
Biiiii r. i"'i. III.I 11.1" «a3.3W'tii-

Cys, with the proviso that if the B30 amino acid residue is Ala or Thr, then at least one of the residues A21 and B3 is different from Asn; Phe^^ may be deleted; and the c-amino group of Lys^^ has a lipophilic substituent which comprises at least 6 carbon atoms.
In another preferred embodiment, the invention relates to a human insulin derivative in which the B30 amino acid residue is deleted or is any amino acid residue which can be coded for by the genetic code except Lys, Arg and Cys; the A21 and the B3 amino acid residues are, independently, any amino acid residues which can be coded for by the genetic code except Lys, Arg and Cys; Phe'^ may be deleted; the e-amino group of Lys*^ has a lipophilic substituent which comprises at least 6 carbon atoms; i and 2-4 Zn^'*' ions are bound to each insulin hexamer.
In another preferred embodiment, the invention relates to a human insulin derivative in which the B30 amino acid residue is deleted.
In another preferred embodiment, the invention relates to a 1 human insulin derivative in which the 830 amino acid residue is Asp.
In another preferred embodiment, the invention relates to a human insulin derivative in which the B30 amino acid residue is Glu.
I In another preferred embodiment, the invention relates to a human insulin derivative in which the B30 amino acid residue is Thr.
In another preferred embodiment, the invention relates to a human insulin derivative in which the 830 amino acid is a I lipophilic amino acid having at least 10 carbon atoms.

In another preferred embodiment, the invention relates to a human insulin derivative in which the B30 amino acid is a lipophilic a-amino acid having from 10 to 24 carbon atoms.
In another preferred embodiment, the invention relates to a human insulin derivative in which the B30 amino acid is a straight chain, saturated, aliphatic a-amino acid having from 10 to 24 carbon atoms.
In another preferred embodiment, the invention relates to a human insulin derivative in which the B30 amino acid is D- or L-N'-dodecanoyllysine.
In another preferred embodiment, the invention relates to a human insulin derivative in which the B30 amino acid is o-amino decanoic acid.
In another preferred embodiment, the invention relates to a human insulin derivative in which the B30 amino acid is a-aroino undecanoic acid.
In another preferred embodiment, the invention relates to a human insulin derivative in which the B30 amino acid is a-amino dodecanoic acid.
In another preferred embodiment, the invention relates to a human insulin derivative in which the 830 amino acid is a-amino tridecanoic acid.
In another preferred embodiment, the invention relates to a human insulin derivative in which the B30 amino acid is o-amino tetradecanoic acid.
In another preferred embodiment, the invention relates to a human insulin derivative in which the B30 amino acid is a-amino pentadecanoic acid.

In another preferred embodiment, the invention relates to a human insulin derivative in which the B30 amino acid is a-araino hexadecanoic acid.
In another preferred embodiment, the invention relates to a human insulin derivative in which the B30 amino acid is an a-amino acid.
In another preferred embodiment, the invention relates to a human insulin derivative in which the A21 amino acid residue is Ala.
In another preferred embodiment, the invention relates to a human insulin derivative in which the A21 amino acid residue is Gin.
In another preferred embodiment, the invention relates to a human insulin derivative in which the A21 amino acid residue is Gly.
In another preferred embodiment, the invention relates to a human insulin derivative in which the A21 amino acid residue is Ser.
In another preferred embodiment, the invention relates to a human insulin derivative in which the B3 amino acid residue is Asp.
In another preferred embodiment, the invention relates to a human insulin derivative in which the B3 amino acid residue is Gin.
In another preferred embodiment, the invention relates to a human insulin derivative in which the B3 amino acid residue is Thr.
In another preferred embodiment, the invention relates to a human insulin derivative in which the e-amino group of Lys^^ has

a lipophilic substituent which is an acyl group corresponding to a carboxylic acid having at least 6 carbon atoms.
In another preferred embodiment, the invention relates to a human insulin derivative in which the e-amino group of Lys"^ has a lipophilic substituent which is an acyl group, branched or unbranched, which corresponds to a carboxylic acid having a chain of carbon atoms 8 to 24 atoms long.
In another preferred embodiment, the invention relates to a human insulin derivative in which the c-amino group of Lys"^ has a lipophilic substituent which is an acyl group corresponding to a fatty acid having at least 6 carbon atoms.
In another preferred embodiment, the invention relates to a human insulin derivative in which the c-amino group of Lys^^ has a lipophilic substituent which is an acyl group corresponding to a linear, saturated carboxylic acid having from 6 to 24 carbon atoms.
In another preferred embodiment, the invention relates to a human insulin derivative in which the c-amino group of Lys^^ has a lipophilic substituent which is an acyl group corresponding to a linear, saturated carboxylic acid having from 8 to 12 carbon atoms.
In another preferred embodiment, the invention relates to a human insulin derivative in which the e-amino group of Lys^^ has a lipophilic substituent which is an acyl group corresponding to a linear, saturated carboxylic acid having from 10 to 16 carbon atoms.
In another preferred embodiment, the invention relates to a human insulin derivative in which the e-amino group of Lys^^ has a lipophilic substituent which is an oligo oxyethylene group comprising up to 10, preferably up to 5, oxyethylene units.

In another preferred embodiment, the invention relates to a human insulin derivative in which the e-amino group of LysB29 has a lipophilic substituent which is an oligo oxypropylene group comprising up to 10, preferably up to 5, oxypropylene units.
In another preferred embodiment, the invention relates to a human insulin derivative in which each insulin hexamer binds 2 Zn^* ions.
In another preferred embodiment, the invention relates to a human insulin derivative in which each insulin hexamer binds 3 Zn^* ions.
In another preferred embodiment, the invention relates to a human insulin derivative in which each insulin hexamer binds 4 Zn^* ions.
In another preferred embodiment, the invention relates to the use of a human insulin derivative according to the invention for the preparation of a medicament for treating diabetes.
In another preferred embodiment, the invention relates to a pharmaceutical composition for the treatment of diabetes in a patient in need of such a treatment comprising a therapeutically effective amount of a human insulin derivative according to the invention together with a pharmaceutically acceptable carrier.
In another preferred embodiment, the invention relates to a pharmaceutical composition for the treatment of diabetes in a patient in need of such a treatment comprising a therapeutically effective amount of a human insulin derivative according to the invention, in mixture with an insulin or an insulin analogue which has a rapid onset of action, together with a pharmaceutically acceptable carrier.
In another preferred embodiment, the invention relates to a pharmaceutical composition comprising a human insulin

derivative according to the invention which is solvible at physiological pH values.
In another preferred embodiment, the invention relates to a pharmaceutical composition comprising a human insulin derivative according to the invention which is soluble at pH values in the interval from about 6.5 to about 8.5.
In another preferred embodiment, the invention relates to a protracted pharmaceutical composition comprising a human insulin derivative according to the invention.
In another preferred embodiment, the invention relates to a pharmaceutical composition which is a solution containing from about 120 nmol/ml to about 1200 nmol/ml, preferably about 600 nmol/ml of a human insulin derivative according to the invention.
In another preferred embodiment, the invention relates to a method of treating diabetes in a patient in need of such a treatment comprising administering to the patient a therapeutically effective amount of an insulin derivative according to this invention together with a pharmaceutically acceptable carrier.
In another preferred embodiment, the invention relates to a method of treating diabetes in a patient in need of such a treatment comprising administering to the patient a therapeutically effective amount of an insulin derivative according to this invention, in mixture with an insulin or an insulin analogue which has a rapid onset of action, together with a pharmaceutically acceptable carrier.
Examples of preferred human insulin derivatives according to the present invention in which no Zn2+ ions are bound are the following:

Examples of preferred human insulin derivatives according to the present invention in which two zn2+ ions are bound per insulin hexamer are the following:
(NB29_tridecanoyl des(B30) human insulin)6, 2Zn2+, (NB29-tetradecanoyl des(B30) human insulin)6, 2Zn2+,

Examples of preferred human insulin derivatives according to the present invention in which three Zn2+ ions are bound per insulin hexamer are the following:
(NB29-tridecanoyl des{B30) human insulin)6, 3Zn2+,

Examples of preferred human insulin derivatives according to the present Invention in which four Zn2+ ions are bound per insulin hexamer are the following:

BRIEF DBSCRZPTIOH OF THE DRAWINGS
The present invention is further illustrated with reference to the appended drawings wherein
Fig. 1 shows the construction of the plasmid pEA5.3.2;
Fig. 2 shows the construction of the plasmid pEAlOS; and
Fig. 3 shows the construction of the plasmid pEAli3.
DETAILED DESCRIPTION OF THE INVENTION
Teminology
The three letter codes and one letter codes for the amino acid residues used herein are those stated in J. Biol. chem. 243. p.
3558 (1968).
In the DNA seguences, A is adenine, C is cytosine, G is guanine, and T is thymine. The following acronyms are used: DMSO for dimethyl sulphoxide, DMF for dimethylformamide, Boc for tert-butoxycarbonyl, RP-HPLC for reversed phase high performance liquid chromatography, X-OSu is an N-hydroxysuccinimid ester, X is an acyl group, and TFA for trifluoroacetic acid.
Preparation of lipophilic insulin derivatives
The insulin derivatives according to the present invention can be prepared i.a. as described in the following:
1. Insulin derivatives featuring in position B30 an amino acid residue which can be coded for by the genetic code. e.g. threonine (human insulin) or alanine (porcine insulin).

1.1 starting from human insulin.
Human insulin is treated with a Boc-reagent (e.g. di-tfirt-butyl
dicarbonate) to form (A1,B1)-diBoc human insulin, i.e., human
insulin in which the N-terminal end of both chains are
protected by a Boc-group. After an optional purification, e.g.
by HPLC, an acyl group is introduced in the c-amino group of LysB29 by allowing the product to react with a N-hydroxysuccinimide ester of the formula X-OSu wherein X is the acyl group to be introduced. In the final step, TFA is used to
remove the Hoc-groups and the product^ NB29-X human insulin, is
isolated.
1.2 starting from a single chain insulin precursor.
A single chain insulin precursor, extended in position Bl with an extension (Ext) which is connected to Bl via an arginine
residue and in which the bridge from B30 to Al is an arginine residue, i.e. a compound of the general formula Ext-Arg-B(1-30)-Arg-A(l-21) , can be used as starting material. Acylation of this starting material with a N-hydroxysuccinimide ester of the general formula X-OSu wherein X is an acyl group, introduces
the acyl group X in the e-amino group of Lys^" and in the N-terminal amino group of the precursor. On treating this acylated precursor of the formula (NB29-X),X-Ext-Arg-B(l-30)-Arg-A(l-21) with trypsin in a mixture of water and a suitable water-miscible organic solvent, e.g. DMF, DMSO or a lower
alcohol, an intermediate of the formula (NB29-X),ArgB31 insulin
is obtained. Treating this intermediate with carboxypeptidase B yields the desired product, (NB29-X) insulin.
2. Insulin derivatives with no amino acid residue in position B30. i.e. des(B30^ insulins.
30 2.1 starting from human insulin or porcine insulin.

On treatment with carboxypeptidase A in anunonium buffer, human
insulin and porcine insulin both yield des(B30) insulin. After
an optional purification, the des(B30) insulin is treated with
a Boc-reagent (e.g. di-tert-butyl dicarbonate) to form (Al,Bl)-
rdiBoc des(B30) insulin, i.e., des(B30) insulin in which the N-
terminal end of both chains are protected by a Boc-group. After
an optional purification, e.g. by HPLC, an acyl group is
introduced in the e-amino group of LysB29 by allowing the product
to react with a N-hydroxysuccinimide ester of the formula X-OSu
wherein X is the acyl group to be introduced. In the final
step, TFA is used to remove the Boc-groups and the product,
(NB29-X) des(B30) insulin, is isolated.
2.2 Starting from a single chain human insulin precursor.
A single chain human insulin precursor, which is extended in
position Bl with an extension (Ext) which is connected to Bl
via an arginine residue and which has a bridge from B30 to Al
can be a useful starting material. Preferably, the bridge is a
peptide of the formula Y^-Arg, where Y is a codable amino acid
except lysine and arginine, and n is zero or an integer between
1 and 35. When n>l, the Y's may designate different amino
acids. Preferred examples of the bridge from B3 0 to Al are:
AlaAlaArg, SerArg, serAspAspAlaArg and Arg (European Patent No.
163529). Treatment of such a precursor of the general formula
Ext-Arg-B( 1-30)-Y„-Arg-A( 1-21) with a lysyl endopeptidase, e.g.
ActroTnobacter lyticus protease, yields Ext-Arg-B(l-29) Thr-Y„-
Arg-A(l-21) des(B30) insulin. Acylation of this intermediate
with a N-hydroxysuccinimide ester of the general formula X-OSu
wherein X is an acyl group, introduces the acyl group X in the
e-amino group of LysB29, and in the N-terminal amino group of the
chain and the B-chain to give (NB29-X) X-Ext-Arg-B(l-29) X-
Thr-Y^-Arg-A(l-21) des(B30) insulin. This intermediate on
treatment with trypsin in mixture of water and a suitable
organic solvent, e.g. DMF, DMSO or a lower alcohol, gives the
desired derivative, CNB29-X) des(B30) human insulin.

Data on MB29 Kodified insulins.
Certain experimental data on NB29 modified insulins are given in Table 1.
The lipophilicity of an insulin derivative relative to human
insulin, krel, was measured on a LiChrosorb RP18 (um, 250x4 man)
HPLC column by isocratic elution at 40"C using mixtures of A)
0.1 M sodium phosphate buffer, pH 7.3, containing 10%
acetonitrile, and B) 50% acetonitrile in water as eluents. The
elution was monitored by following the UV absorption of the
eluate at 214 nm. Void time, tg, was found by injecting 0.1 mM
sodium nitrate. Retention time for human insul in, t^^^f^, was
adjusted to at least 2to by varying the ratio between the A and
B solutions.
The degree of prolongation of the blood glucose lowering effect
was studied in rabbits. Each insulin derivative was tested by
subcutaneous injection of 12 nmol thereof in each of six
rabbits in the single day retardation test. Blood sampling for
glucose analysis was performed before injection and at 1, 2, 4
and 6 hours after injection. The glucose values found are
of expressed as percent of initial values. The Index of
Protraction, which was calculated from the blood glucose
values, is the scaled Index of Protraction (prolongation), see
p. 211 in Markussen et al.. Protein Engineering 1 (1987) 205-
213. The formula has been scaled to render a value of 100 with
bovine ultralente insulin and a value of 0 with Actrapid*
insulin (Novo Nordisk A/S, 2880 Bagsvaerd, Denmark).
The insulin derivatives listed in Table 1 were administered in solutions containing 3 Zn2+ per insulin hexamer, except those specifically indicated to be Zn-free.
For the very protracted analogues the rabbit model is inadequate because the decrease in blood glucose from initial is too small to estimate the index of protraction. The prolongation of such analogues is better characterized by the

disappearance rate in pigs. T50% is the time when 50% of the A14 Tyr(125I) analogue has disappeared from the site of injection as measured with an external Tf-counter (Ribel, U et al.. The Pig as a Model for Subcutaneous Absorption in Man. In:
M. serrano-Rios and P.J. Lefebre (Eds): Diabetes 1985;
Proceedings of the l2th Congress of the International Diabetes Federation, Madrid, Spain, 1985 (Excerpta Medica, Amsterdam,
(1986) 891-96).
In Table 2 are given the T50% values of a series of very protracted insulin analogues. The analogues were administered in solutions containing 3 Zn2+ per insulin hexamer.

solubility
The solubility of all the NB29 modified insulins mentioned in Table 1, which contain 3 Zn2+ ions per insulin hexamer, exceeds 600 nmol/ml in a neutral (pH 7.5), aqueous, pharmaceutical formulation which further comprises 0.3% phenol as preservative, and 1.6% glycerol to achieve isotonicity. 600 nmol/ml is the concentration of human insulin found in the 100 lU/ml compositions usually employed in the clinic.
The e-B29 amino group can be a component of an amide bond, a sulphonamide bond, a carbamide, a thiocarbamide, or a carbamate. The lipophilic substituent carried by the C-B29 amino group can also be an alkyl group.
Pharmaceutical compositions containing a human insulin derivative according to the present invention may be administered parenterally to patients in need of such a treatment. Parenteral administration may be performed by subcutaneous, intramuscular or intravenous injection by means of a syringe, optionally a pen-like syringe. Alternatively, parenteral administration can be performed by means of an infusion pump. A further option is a composition which may be a powder or a liquid for the administration of the human insulin derivative in the form of a nasal spray.
The injectable human insulin compositions of the invention can be prepared using the conventional techniques of the pharmaceutical industry which involves dissolving and mixing the ingredients as appropriate to give the desired end product.
Thus, according to one procedure, the human insulin derivative is dissolved in an amount of water which is somewhat less than the final volume of the composition to be prepared. An isotonic agent, a preservative and a buffer is added as required and the pH value of the solution is adjusted - if necessary - using an acid, e.g. hydrochloric acid, or a base, e.g. aqueous sodium hydroxide as needed- Finally, the volume of the solution is

adjusted with water to give the desired concentration of the ingredients.
Examples of isotonic agents are sodium chloride, mannitol and glycerol.
Examples of preservatives are phenol, m-cresol, methyl p- hydroxybenzoate and benzyl alcohol.
Examples of suitable buffers are sodium acetate and sodium phosphate.
composition for nasal administration of an insulin derivative according to the present invention may, for example, be prepared as described in European Patent No. 272097 (to Novo Nordisk A/S).
The insulin compositions of this invention can be used in the
treatment of diabetes. The optimal dose level for any patient
will depend on a variety of factors including the efficacy of
the specific human insulin derivative employed, the age, body
weight, physical activity, and diet of the patient, on a
possible combination with other drugs, and on the severity of
the case of diabetes. It is recommended that the daily dosage
of the human insulin derivative of this invention be determined
for each individual patient by those skilled in the art in a
similar way as for known insulin compositions.
Where expedient, the human insulin derivatives of this
invention may be used in mixture with other types of insulin,
e.g. human insulin or porcine insulin or insulin analogues with
a more rapid onset of action. Examples of such insulin
analogues are described e.g. in the European patent
applications having the publication Nos. EP 214826 (Novo
Nordisk A/S), EP 375437 (Novo Nordisk A/S) and EP 383472 (Eli
Lilly & Co,).

The present invention is further illustrated by the following examples which, however, are not to be construed as limiting the scope of protection. The features disclosed in the foregoing description and in the following examples may, both separately and in any combination thereof, be material for realizing the invention in diverse forms thereof.
EXAMPLES
Flasmids and DMA naterial
All expression plasmids are of the cPOT type. Such plasmids are
described in EP patent application No. 171 142 and are
characterized in containing the Schizosaccharomyces pombe
triose phosphate isomerase gene (POT) for the purpose of
plasmid selection and stabilization. A plasmid containing the
POT-gene is available from a deposited E. coli strain {ATCC
39685). The plasmids furthermore contain the S* ggrevisiae
triose phosphate isomerase promoter and terminator (P^p, and
T1p1). They are identical to pMT742 (Egel-Mitanl, M. et al., Gene
73 (1988) 113-120) (see Fig. 1) except for the region defined
by the ECoRI-Xbal restriction sites encompassing the coding
region for signal/leader/product.
Synthetic DNA fragments were synthesized on an automatic DNA synthesizer (Applied Biosystems model 3BOA) using phosphoramidite chemistry and commercially available reagents (Beaucage, S.L. and Caruthers, M.H., Tetrahedron Letters 22 (1981) 1859-1869) .
All other methods and materials used are common state of the art knowledge (see, e.g. Sambrook, J., Fritsch, E.F. and Maniatis, T., Molecular Cloning; A Laboratory Napual. Cold Spring Harbor Laboratory Press, New York, 1989).

Analytical
Molecular masses of the insulins prepared were obtained by MS (mass spectroscopy), either by PDMS (plasma desorption mass spectrometry) using a Bio-Ion 20 instrument (Bio-Ion Nordic AB, Uppsala, Sweden) or by ESMS (electrospray mass spectrometry) using an API III Biomolecular Mass Analyzer (Perkin-Elner Sciex Instruments, Thornhill, Canada),
EXAMPLE 1
Synthesis of AlaA21 AspB3 human insulin precursor from Yeast strain yEA002 using the LaC212spx3 signal/leader.

58.5 ul of water
One cycle was performed: 94°c for 45 sec, 49°c for 1 min, 72'C for 2 min.
Subsequently, 5/xl of oligonucleotides #16 and #126 was added
and 15 cycles were performed: 94"C for 45 sec, 45°c for 1 min,
72'C for 1.5 min. The PCR mixture was loaded onto a 2.5 %
agarose gel and subjected to electrophoresis using standard
techniques (Sambrook et al.. Molecular cloning, Cold Spring
Harbour Laboratory Press, 1989). The resulting DNA fragment was
cut out of the agarose gel and isolated using the Gene Clean
Kit (Bio 101 Inc., PO BOX 2284, La Jolla, CA 92038, USA)
according to the manufacturer's instructions. The purified PCR
DNA fragment was dissolved in 10 fxl of water and restriction
endonuclease buffer and cut with the restriction endonucleases
N
col and Xba I according to standard techniques, run on a 2.5% agarose gel and purified using the Gene Clean Kit as described.
The plasmid pAK188 consists of a DNA sequence of 412 bp composed of a EcoRI/NcoI fragment encoding the synthetic yeast
signal/leader gene LaC212spx3 (described in Example 3 of HO 89/02463) followed by a synthetic Ncol/Xbal fragment encoding the insulin precursor MI5, which has a SerAspAspAlaLys bridge connecting the B29 and the Al amino acid residues (see SEQ ID NOS. 14, 15 and 16), inserted into the EcoRI/Xbal fragment of the vector (phagemid) pBLUESCRIPT IIsk(+/-) (Stratagene, USA).
The plasmid pAK188 is shown in Fig. 1.
The plasmid pAK188 was also cut with the restriction endonucleases Ncol and Xbal and the vector fragment of 3139 bp isolated. The two DNA fragments were ligated together using T4 DNA ligase and standard conditions (Sambrook et al.. Molecular Cloning, Cold Spring Harbour Laboratory Press, 1989). The ligation mixture was transformed into a competent E. coli strain (R-, M+) followed by selection for ampicillin resistance. Plasmids were isolated from the resulting E. coli colonies using standard DNA miniprep technique (Sambrook et

al., Molecular Cloning, Cold Spring Harbour Laboratory Press, 1989), checked with appropriate restrictions endonucleases i.e. EcoRI, Xba I, Ncol and Hpal. The selected plasmid was shown by DNA sequencing analyses (Sequenase, U.S. Biochemical Corp.) to contain the correct sequence for the Ala*^^, Asp^^ human insulin precursor and named pEA5.3.
The plasmid pKFN1627 is an E. coli - S. cerevisiae shuttle vector, identical to plasmid pKFN1003 described in EP patent No, 375718, except for a short DNA sequence upstream from the
unique Xbal site. In pKFNloOB, this sequence is a 178 bp fragment encoding a synthetic aprotinin gene fused in-frame to the yeast mating factor alpha l signal-leader sequence. In pKFN1627, the corresponding 184 bp sequence encodes the insulin precursor MIS (Glu^^, Glu^^Sj (i.e. B(l-29, G1U^\G1U"^)-
SerAspAspAlaLys-A(l-21) fused in-frame to the mating factor alpha 1 sequence (see SEQ ID NOS. 17, 18 and 19). The vector pKFN1627 is shown in Fig. 1.
pEA5.3 was cut with the restriction endonucleases EcoRI and
Xbal and the resulting DNA fragment of 412 bp was isolated. The
yeast expression vector pKFN1627 was cut with the restriction
endonucleases Ncol and Xbal and with Ncol and EcoRI and the DNA
fragment of 9273 bp was isolated from the first digestion and
the DNA fragment of 1644 bp was isolated from the second. The
412 bp EcoRI/Xbal fragment was then ligated to the two other
fragments, that is the 9273 bp Ncol I/Xbal fragment and the
1644 bp NcoI/EcoRI fragment using standard techniques.
The ligation mixture was transformed into E. coli as described above. Plasmid from the resulting E. coli was isolated using standard techniques, and checked with appropriate restriction
endonucleases i.e. EcoRI, Xbal, Ncol, Hpa I. The selected plasmid was shown by DNA sequence analysis (using the Sequenase kit as described by the manufacturer, U.S. Biochemical) to contain the correct sequence for the Ala21 AspB3 human insulin precursor DNA and to be inserted after the DNA encoding the
liaC212spx3 signal/leader DNA. The plasmid was named pEA5.3.2

and is shown in Fig. 1. The DNA sequence encoding the LaC2l2spx3 signal/leader/Ala*^-'- Asp^^ human insulin precursor complex and the amino acid sequence thereof are SEQ ID NOS. 20, 21 and 22. The plasmid pEA5.3.2 was transformed into S^. h cerevisiae strain MT663 as described in European patent ' application having the publication No. 214826 and the resulting strain was named yEA002.

The DNA encoding GlyA21 AspB3 human insulin precursor was constructed in the same manner as described for the DNA encoding Ala*^^ Asp^^ human insulin precursor in Example 1. The DNA sequence encoding the LaC212spx3 signal/leader/Gly*^-*- Asp^-' human insulin precursor complex and the amino acid sequence thereof are SEQ ID NOS. 26, 27 and 28. The plasmid pEAl.5.6 was shown to contain the desired sequence, transformed into £j. cerevisiae strain MT663 as described in Example 1 and the resulting strain was named yEA007.

The DNA encoding Gly*^ ^ Thr^-' human insul in precursor was
constructed in the same manner as described for the DNA
encoding Ala*^^ Asp^-^ human insulin precursor in Example l. The
DNA sequence encoding the LaC212spx3 signal/leader/Gly^^^ Thr®"'
human insulin precursor complex and the amino acid seguence
thereof are SEQ ID NOS. 29, 30 and 31. The plasmid pEA4.4.11
was shown to contain the desired DNA seguence, transformed into
S. cerevisiae strain MT663 as described in Example l and the
resulting strain was named yEA006,

a DM sequence of 412 bp encoding the synthetic yeast signal'leader LaC2l2spx3 (described in Example 3 of WO 89/02463) followed by the insulin precursor MIS (see SEQ ID NOS. 14, 15 and 16) inserted into the vector (phagemld) pBLUESCRIPT IIsk(+/-) (Stratagene, USA).

A total of 16 cycles were performed, each cycle comprising 1
minute at 95*C; 1 minute at 40'C; and 2 minutes at 72'C, The
PCR mixture was then loaded onto a 2% agarose gel and subjected
to electrophoresis using standard techniques. The resulting DNA
fragment was cut out of the agarose gel and isolated using the
Gene Clean kit (Bio 101 Inc., PO BOX 2284, La Jolla, CA 92038,
USA) according to the manufacture's instructions. The purified
PCR DNA fragment was dissolved in 10 jxl of water and
restriction endonuclease buffer and cut with the restriction
endonucleases Hindlll and Xbal according to standard
techniques. The Hindlll/Xbal DNA fragment was purified using
The Gene Clean Kit as described.
The plasmid pAK406 consists of a DNA sequence of 520 bp comprising an EcoRI/Hindlll fragment derived from pMT636 (described in WO 90/10075) encoding the yeast alpha factor
l
eader and part of the insulin precursor ligated to the Hindlll/Xbal fragment from pAK188 encoding the rest of the insulin precursor MIS (see SEQ ID NOS. 32, 33 and 34) inserted into the vector cPOT. The vector pAK406 is shown in Fig. 2.

The plasmid pAK233 consists of a DNA sequence o£ 412 bp encoding the synthetic yeast signal/leader LaC212spx3 (described in Example 3 of WO 89/02463) followed by the gene for the insulin precursor B(l-29)-GluLysArg-A(l-2l) (A21-Gly)
(see SEQ ID NOS, 35, 36 and 37) inserted into the vector cPOT.
The plasmid pAK233 is shown in Fig. 2.
The plasmid pAK233 was cut with the restriction endonucleases
Ncol and Xbal and the vector fragment of 9273 bp Isolated. The
plasmid pAK406 was cut with the restriction endonucleases Ncol
and Hindlll and the vector fragment of 2012 bp isolated. These
two DNA fragments were ligated together with the Hindlll/Xbal
PCR fragment using T4 DNA ligase and standard conditions. The
ligation mixture was then transformed into a competent E. coli
strain (R-, M+) followed by selection for ampicillin
resistance. Plasmids were isolated from the resulting E. coli
colonies using a standard DNA miniprep technique and checked
with appropriate restriction endonucleases i.e. EcoRI, Xbal,
Ncol, Hindlll. The selected plasmid was shown by DNA sequencing
analyses to contain the correct sequence for the Arg° single
chain human insulin precursor DNA and to be inserted after the
DNA encoding the S. cerevisiae alpha factor DNA. The plasmid
was named pEAlOS and is shown in Fig. 2. The DNA sequence
encoding the alpha factor leader/ArgB31 single chain human
insulin precursor complex and the amino acid sequence thereof
are SEQ ID NOS. 38, 39 and 40. The plasmid pEA 108 was
transformed into S. cerevisiae strain MT663 as described in
Example l and the resulting strain was named yEAlOS.
B)
The following Polymerase Chain Reaction (PCR) was performed
using the Gene Amp PCR reagent kit (Perkin Elmer, 761 Main
Avewalk, CT 06859, USA) according to the manufacturer's
instructions. In all cases, the PCR mixture was overlayed with
100 ul of mineral oil (Sigma Chemical Co., St. Louis, MO, USA)
5 ul of oligonucleotide #220 (100 pmol)
5 ul of oligonucleotide #307 (100 pmol)
10 ul of lOX PCR buffer
16 ul of dNTP mix 0.5 ul of Taq enzyme 0.2 ul of pEAlOS plasmid as template (0.1 ug DNA)
63 ul of water
A total of 16 cycles were performed, each cycle comprising 1
minute at 95'C; 1 minute at 40'C; and 2 minutes at 72'C. The
PCR mixture was then loaded onto an 2% agarose gel and
subjected to electrophoresis using standard techniques. The
resulting DNA fragment was cut out of the agarose gel and
isolated using the Gene Clean kit (Bio 101 Inc., PO BOX 2284,
La Jolla, CA 92038, USA) according to the manufacture * s
instructions. The purified PCR DNA fragment was dissolved in 10
ul of water and restriction endonuclease buffer and cut with
the restriction endonucleases Ncol and Xbal according to
standard techniques. The Ncol/Xbal DNA fragment was purified
using The Gene Clean Kit as described.
The plasmid pAK401 consists of a DNA sequence of 523 bp
composed of an EcoRI/NcoI fragment derived from pMT636
(described in WO 90/10075) (constructed by by introducing a
Ncol site in the 3'-end of the alpha leader by site directed
mutagenesis) encoding the alpha factor leader followed by a
Ncol/Xbal fragment from pAK188 encoding the insulin precursor MI5 (see SEQ ID NOS. 41, 42 and 43) inserted into the vector (phagemid) pBLUESCRIPT IIsk(+/-) (Stratagene, USA). The plasmid pAK40l is shown in Fig. 3.
The plasmid pAK40l was cut with the restriction endonucleases Ncol and Xbal and the vector fragment of 3254 bp isolated and ligated together with the Ncol/Xbal PCR fragment. The ligation mixture was then transformed into a competent E. coli strain and plasmids were isolated from the resulting E. coli colonies using a standard DNA miniprep technique and checked with appropriate restriction endonucleases i.e. EcoRI, Xbal, Ncol.

The selected plasmid, named pll3A (shown in Fig. 3), was cut with EcoRI and Xbal and the fragment of 535 bp Isolated.
The plasmid pAK233 was cut with the restriction endonucleases
Ncol and Xbal, and with EcoRl/Ncol and the fragments of 9273
and 1644 bp isolated. These two DNA fragments were ligated
together with the EcoRI/Xbal fragment from pll3A using T4 DKA
ligase and standard conditions. The ligation mixture was then
transformed into a competent E. coli strain (R-, M+) followed
by selection for ampicillin resistance. Plasmids were isolated
from the resulting E. coli colonies using a standard DNA
miniprep technique and checked with appropriate restriction
endonucleases i.e. EcoRI, Xbal, Ncol, Hindlll. The selected
plasmid was shown by DNA sequencing analyses to contain the
correct sequence for the ArgB31 single chain human insulin
precursor DNA with the N-terminal extension
GluGluAlaGluAlaGluAlaArg and to be inserted after the DNA
encoding the 5. cerevisiae alpha factor DNA. The plasmid was
named pEA113 and is shown in Fig. 3. The DNA sequence encoding
the alpha factor leader/ArgB-1 ArgB31 single chain human
insulin precursor having an N-terminal extension
(GluGluAlaGluAlaGluAlaArg) and the amino acid sequence thereof
are SEQ ID NOS. 44, 45 and 46. The plasmid pEA113 was
transformed into S. cerevisiae strain MT663 as described in
Example 1 and the resulting strain was named yEA113.
EXAMPLE 6
Synthesis of Arg Arg single chain human insulin precursor having an N-terminal extension (GluGluAlaGluAlaGluAlaGluArg) from Yeast strain yEA136 using the alpha factor leader.
The following oligonucleotide was synthesized:
#389 5'-GCTAACGTCGCCATGGCTAAGAGAGAAGAAGCTGAAGCGAAG CTGAAAGATTCGTTAACCAACAC-3' (SEQ ID NO:13)

The following PCR was performed using the Gene Amp PCR reagent kit
5 u1 of oligonucleotide #220 (100 pmol) 5 Ml of oligonucleotide #389 (100 pmol) 10 Ml of lOX PCR buffer 16 Ml of dNTP mix 3.5 Ml of Tag enzyme
2 Ml of pEA113 plasmid as template (0.5 ug DNA)
S3 Ml of water
\ total of 12 cycles were performed, each cycle comprising 1 ninute at 95°C; 1 minute at 37°C; and 2 minutes at 72°C.
The DNA encoding alpha factor leader/ArgB-1 ArgB31 single chain luman insulin precursor having an N-terminal extension (GluGluAlaGluAlaGluAlaGluArg) was constructed in the same nanner as described for the DNA encoding alpha factor Leader/Arg Arg single chain human insulin precursor laving an N-terminal extension (GluGluAlaGluAlaGluAlaArg) in Example 5. The plasmid was named pEA13 6. The DNA sequence ancoding the alpha factor leader/ArgB-1 ArgB31 single chain luman insulin precursor having an N-terminal extension (GluGluAlaGluAlaGluAlaGluArg) and the amino acid sequence thereof are SEQ ID NOS. 47, 48 and 49. The plasmid pEA136 was bransformed into S. cerevisiae strain MT663 as described in Example I and the resulting strain was named yEA136.
EXAMPLE 7
Synthesis of (A1,B1)-diBoc human insulin.
5 g of zinc-free human insulin was dissolved in 41.3 ml of
DMSO. To the solution was added 3.090 ml of acetic acid. The
reaction was conducted at room temperature and initiated by
addition of 565 mg of di-tert-butyl pyrocarbonate dissolved in
5.650 ml of DMSO. The reaction was allowed to proceed for 5k

hour and then stopped by addition of 250 u1 of ethanolamine. The product was precipitated by addition of 1500 ml of acetone. The precipitate was isolated by centrifugation and dried in vacuum. A yield of 6.85 g material was obtained.
(Al,Bl)-diBoc insulin was purified by reversed phase HPLC as follows: The crude product was dissolved in 100 ml of 25% ethanol in water, adjusted to pH 3,0 with HCl and applied to a column (5 cm diameter, 30 cm high) packed with octadecyJdimethylsilyl-substituted silica particles (mean
particle size 15 um, pore size 100 A) and equilibrated with elution buffer. The elution was performed using mixtures of ethanol and 1 mM aqueous HCl, 0.3 M KCl at a flow of 2 1/h. The insulin was eluted by increasing the ethanol content from 30% to 45%. The appropriate fraction was diluted to 20% ethanol and
precipitated at pH 4.8. The precipitated material was isolated by centrifugation and dried in vacuum. Thus 1.701 g of (A1,B1)-diBoc human insulin was obtained at a purity of 94.5%.
EXAMPLE 0
Synthesis of (N*°"-benzoyl human insulin)^, 3Zn^*.

400 mg of (A1,B1)-diBoc human insulin was dissolved in 2 ml of DMSO. To the solution was added 748 ^1 of a mixture of N-methylmorpholine and DMSO (1:9, v/v). The reaction was conducted at 15'C and initiated by addition of 14.6 mg of benzoic acid N-hydroxysuccinimide ester dissolved in 132 ^1 DMF. The reaction was stopped after 2 hours by addition of 100 ml of acetone. The precipitated material was isolated by centrifugation and dried in vacuum. 343 mg of material was collected.
The Boc protecting groups were eliminated by addition of 4 ml of TFA. The dissolved material was incubated for 30 minutes and then precipitated by addition of 50 ml of acetone. The precipitate was isolated by centrifugation and dried in vacuum.

NB29-benzoyl human insulin was purified by reversed phase HPLC as described in Example 7. A yield of 230 mg was obtained. Recrystallization from 15% aqueous ethanol containing 6 mM 2n^* and 50 mM citrate at pH 5.5 gave crystals of the title compound which were isolated by centrifugation and dried in vacuum. The yield was 190 mg.
Molecular mass, found by MS: 5911, theory: 5911.
EXAMPLE 9
Synthesis of (N'^^-lithocholoyl human insulin)^, SZn^*.

400 mg of (Al,Bl)-diBoc human insulin was dissolved in 2 ml of DMSO. To the solution was added 748 ^il of a mixture of N-methylmorpholine and DMSO (1:9, v/v). The reaction was conducted at 15' C and initiated by addition of 31.94 mg of lithocholic acid N-hydroxysuccinimide ester dissolved in 300 nl of DMF. The reaction was stopped after 2 hours by addition of 100 ml of acetone. The precipitated material was isolated by centrifugation and dried in vacuum. 331 mg of material was obtained,
The Boc protecting groups were eliminated by addition of 4 ml of TFA. The dissolved material was incubated for 30 minutes and then precipitated by addition of 50 ml of acetone. The precipitate was isolated by centrifugation and dried in vacuum. The yield was 376 mg.
B29-lithocholoyl insulin was purified by reversed phase HPLC as described in Example 7. A final yield of 67 mg was obtained at a purity of 94%. Recrystallization from 15% aqueous ethanol containing 6 mM Zn2+ and 50 mM citrate at pH 5.5 gave crystals of the title compound which were isolated by centrifugation and dried in vacuum. The yield was 49 mg.
Molecular mass, found by MS: 6160, theory: 6166.

EXAMPLE 10
Synthesis of (NB29-decanoyl human insulin)6, 3Zn2+.
400 mg of (Al,Bl}-diBoc human Insulin was dissolved in 2 ml of
DMSO. To the solution was added 748 ul of a mixture of N-
methylraorpholine and DMSO (1:9/ v/v). The reaction was
conducted at 15*C and initiated by addition of 18.0 mg of
decanoic acid N-hydroxysuccinimide ester dissolved in 132 fil of
DMF. The reaction was stopped after 60 minutes and the product
precipitated by addition of 100 ml of acetone. The precipitated
material was isolated by centrifugation and dried in vacuum.
420 mg of intermediate product was collected.
The Boc protecting groups were eliminated by addition of 4 ml
of TFA. The dissolved material was incubated for 30 minutes and
the product was then precipitated by addition of 50 ml of
acetone. The precipitate was isolated by centrifugation and
dried in vacuum. The yield of crude product was 420 mg.
The crude product was purified by reversed phase HPLC as described in Example 7. A final yield of 254 mg of the title product was obtained. The purity was 96.1%. Recrystallization from 15% aqueous ethanol containing 6 mM Zn^* and 50 mM citrate at pH 5.5 gave crystals of the title compound which were isolated by centrifugation and dried in vacuum. The yield was 217 mg.
Molecular mass, found by MS: 5962, theory: 5962.
EXAMPLE 11
Synthesis of des(B30) human insulin.
Synthesis of des(B30) human insulin was carried out as described by Markussen (Methods in diabetes research, Vol. I,

Laboratory methods, part B, 404-410. Ed: J. Lamer and S. Phol, John Wiley & Sons, 1984). 5 g of human insulin was dissolved in 500 ml of water while the pH value of the solution was kept at 2.6 by addition of 0.5 M sulphuric acid. Subsequently, the insulin was salted out by addition of 100 g of ammonium sulphate and the precipitate was isolated by centrifugation. The pellet was dissolved in 800 ml of 0.1 M ammonium hydrogen carbonate and the pH value of the solution was adjusted to 8.4 with 1 H ammonia.
50 mg of bovine carboxypeptidase A was suspended in 25 ml of water and isolated by centrifugation. The crystals were suspended in 25 ml of water and 1 M ammonia was added until a clear solution was obtained at a final pH of 10. The carboxypeptidase solution was added to the insulin solution and the reaction was allowed to proceed for 24 hours. A few drops of toluene were added to act as preservative during the reaction.
After 24 hours the des(B30) human insulin was crystallized by successive addition of 80 g of sodium chloride while the solution was stirred. The pH value was then adjusted to 8.3 and the crystallization was allowed to proceed for 20 hours with gentle stirring. The crystals were isolated on a 1.2 ^m filter, washed with 250 ml of ice cold 2-propanol and finally dried in vacuum.
EXAMPLE 12
Synthesis of (A1,B1)-diBoc des(B30) human insulin.
The title compound was synthesized by a method similar to that described in Example 7, using des(B30) porcine insulin as the starting material. The crude product was precipitated by acetone and dried in vacuum. The (A1,B1)-diBoc (ies(B30) human insulin was purified by reversed phase HPLC as described in Example 7.

EXAMPLE 13
Synthesis of NB29-decanoyl des(B30) human insulin.
400 ng of (Al,Bl)-diBoc des(B30) human insulin was used as starting material for the synthesis of N**^-decanoyl des(B30)
human insulin, following the procedure described in Example 10. The crude product was precipitated by acetone, dried in vacuum and deprotected using TFA. The resulting product was precipitated by acetone and dried in vacuum. N*^^-decanoyl
des(B30} human insulin was then purified by reversed phase HPLC
as described in Example 10.
Molecular mass, found by MS: 5856, theory: 5861.
EXAMPLE 14
Synthesis of NB29-dodecanoyl des(B30) human insulin,
a. IBUDObiliBatioD of A. Ivticus prctease
13 mg of A. lytipus protease, dissolved in 5 ml of aqueous 0.2
M NaHCO3 buffer, pH 9.4, was mixed with 4 ml of settled
MiniLeak Medium gel, which had been washed with the same buffer
(MiniLeak is a divinylsulfone activated Sepharose CL 6B,
obtained from KemEnTec, Copenhagen) . The gel was kept in
suspension by gentle stirring for 24 hours at room temperature.
Then, the gel was isolated by filtration, washed with water,
and suspended in 20 ml of 1 M ethanolamine buffer, pH 9.4, and
kept in suspension for 24 hours at room temperature. Finally,
the gel was washed with water followed by 0.1 M acetic acid and
stored at 4°C. The enzyme activity in the filtrate was 13% of
that in the initial solution, indicating a yield in the
immobilization reaction of about 87%.

b. Imnobilizatlon of porcine trypsin
Porcine trypsin was immobilized to MiniLeak* Low to a degree of substitution of 1 mg per ml of gel, using the conditions described above for immobilization of A. lyticus.

To 200 mg of Glu(GluAla)3Arg-B(l-29)-ThrArg-A(l-21) single-chain human insulin precursor, dissolved in 20 ml of 0.1 M NaHCO, buffer, pH 9.0, was added 4 ml of the gel carrying the
immobilized A. lyticus protease. After the gel had been kept in suspension in the reaction mixture for 6 hours at room temperature the hydrolysis was complete, rendering GluCGluAla),-Arg-B(l-29), ThrArg-A(l-21) human insulin (the reaction was followed by reversed phase HPLC). After the hydrolysis, the gel
was removed by filtration. To the filtrate was added 5 ml of
ethanol and 15 ^L of 1 ** ZnClj and the pH was adjusted to 5.0 using HCl. The precipitation of the product was completed on standing overnight at 4°C with gentle stirring. The product was isolated by centrifugation. After one washing with 1 ml of ice
cold 20% ethanol and drying in vacuo the yield was 190 mg.

190 mg (30 umol) of Glu(GluAla) 3Arg-B(l-29) , ThrArg-A(l-21) insulin was dissolved in 1 ml of DMSO and 1.05 ml of a 0.572 M solution of N,N-diisopropylethylamine in DMF. The solution was
cooled to 15°C and 36 mg (120 umol) of dodecanoic acid N-
hydroxysuccinimide ester dissolved in 0.6 ml of DMSO was added.
The reaction was completed within 24 hours. The lipophilic title compound was not isolated.

The product from the previous step, d., contained in
approximately 2,65 mX of DMSO/DMF/N,N-diisopropylethylamine was
diluted with 10.6 ml of a 50 mM glycine buffer comprising 20%
ethanol and the pH adjusted to 10 with NaOH. After standing for
1 hour at room temperature 1 ml of MiniLeak gel, carrying 1 mg
of iroroobilized trypsin per ml of gel, was added. The reaction
mixture was stirred gently for 48 hours at room temperature. In
order to isolate the desired product, the reaction mixture was
applied to a reversed phase HPLC column (5 cm in diameter, 30
cm high) , packed with octadecyldimethylsilyl-substituted silica
particles (mean particle size 15 nm, pore size 100 A). For the
elution was used 20 fliM Tris/HCl buffers, adjusted to pH 7.7 and
comprising an increasing concentration of ethanol, from 40% to
44% (v/v), at a rate of 2000 ml/h. The major peak eluting at
about 43-44% of ethanol contained the title compound. The
fractions containing the major peak were pooled, water was
added to reduce the ethanol concentration to 20% (v/v), and the
pH was adjusted to 5.5. The solution was left overnight at
-20"C, whereby the product precipitated. The precipitate was
isolated by centrifugation at -8 ° C and dried in vacuo. The
yield of the title compound was 90 mg.
Molecular mass, found by MS: 5892, theory: 5890.
EXAMPLE 15
Synthesis of NB29-(N'myristoyl-Q(-glutamyl) human insulin.
500 mg of (Al,Bl)-diBoc human insulin was dissolved in 2.5 ml of DMSO and 428 ul of ethyl diisopropylamine, diluted with 2.5 ml of DMSO/DMF 1/1 (v/v), was added. The temperature was adjusted to 15'C and 85 mg of N-myristoyl-Glu(OBut) N-hydroxysuccinimide ester, dissolved in 2.5 ml of DMSO/DMF 1/1 (v/v), was added. After 30 min the reaction mixture was poured into 60 ml of water, the pH adjusted to 5 and the precipitate

isolated by centrifugation. The precipitate was dried in vacuo.
The dried reaction mixture was dissolved in 25 ml of TFA, and
the solution was left for 30 min at room temperature. The TFA
was removed by evaporation in vacuo. The gelatinous residue was
dissolved in 60 ml of water and the pH was adjusted to 11,2
using concentrated ammonia. The title compound was crystallized
from this solution by adjustment of the pH to 8.5 using 6 N
HCl. The product was isolated by centrifugation, washed once by
10 ml of water, and dried in vacuo. Yield 356 ng. Purity by
HPLC 94%.
The product of this example is thus human insulin wherein the e-amino group of LysB29 has a substituent of the following structure: CH3 (CH2)12CONHCH (CH2CH2COOH) CO-.
Molecular mass, found by MS: 6146, theory: 6148.
EXMPLE 16
Synthesis of NB29-undecanoyl des(B30) human insulin.
The title compound was synthesized analogously to NB29-dodecanoyl des(B30) human Insulin as described in Example 14,
by using undecanoic acid N-hydroxysuccinimide ester instead of
dodecanoic acid N-hydroxysuccinimide ester.
Molecular mass of the product found by MS: 5876, theory: 5876.
EXAMPLE 17
Synthesis of NB29-tridecanoyl des(B30) human insulin.

The title compound was synthesized analogously to NB29-dodecanoyl des{B30) human insulin as described in Example 14, by using tridecanoic acid N-hydroxysuccinimide ester instead of dodecanoic acid N-hydroxysuccinimide ester.

Molecular mass of the product found by MS: 5899, theory: 5904.
EXAMPLE 18
Synthesis of NB29-inyristoyl des(B30) human insulin.
The title compound was synthesized analogously to N**^-dodecanoyl des(B30) human insulin as described in Example 14, by using myristic acid N-hydroxysuccinimide ester instead of dodecanoic acid N-hydroxysuccinimide ester.
Molecular mass of the product found by MS: 5923, theory: 5918.
EXAMPLE 19
Synthesis of NB29-palmitoyl des(B30) human insulin.
The title compound was synthesized analogously to NB29-dodecanoyl des(B30) human insulin as described in Example 14, by using palmitic acid N-hydroxysuccinimide ester instead of dodecanoic acid N-hydroxysuccinimide ester.
Molecular mass of the product found by MS: 5944, theory: 5946.
EXAMPLE 20
Synthesis of NB29-suberoyl-D-thyroxine human insulin.

a. Preparation of N-(succinimidYlsuberoyl)-D-thyroxine. Disuccinimidyl suberate (1.0 g, Pierce) was dissolved in DMF (50 ml), and D-thyroxine (2.0 g, Aldrich) was added with stirring at 20C. The thyroxine slowly dissolved, and after 20 hours the solvent was removed by evaporation in vacuo. The oily

residue was crystallized from 2-propanol to yield 0.6 g of N-(succinijnidylsuberoyl)-D-thyroxine, m.p. 128-133'C.
b. Rgagtj.pn o£ lAl. Bl) -diBoc human insul in with H-
fBuccinimidvlsuberovll-D-thyroxine.
(A1,B1)-diBoc human insulin (200 mg) was dissolved in dry DMF
(10 ml) by addition of tri ethyl amine (20 M1) at room
temperature- Then, N-(succinimidylsuberoyl)-D-thyroxine (80 rag)
was added. The reaction was monitored by reversed phase HPLC
and when the reaction was about 90% complete, the solvent was

removed in vacuo. To the evaporation residue, anhydrous trifluoroacetic acid (5 ml) was added, and the solution was kept for 1 hour at room temperature. After removal of the trifluoroacetic acid in vacuo, the residue was dissolved in a mixture of IM acetic acid (5 ml) and acetonitrile (1.5 ml), purified by preparative reversed phase HPLC and desalted on a PD-IO column. The yield of N*°^-suberoyl-D-thyroxine human insulin was 50 mg.
The product of this example is thus human insulin wherein the
€-amino group of Lys^^ has a substituent of the following
structure: Thyrox-CO(CHg)^co-, wherein Thyrox is thyroxine which
is bound to the octanedioic acid moiety via an amide bond to
its a-amino group.
Molecular mass of the product found by MS: 6724, theory: 6723.
SZ&MFLE 21
Synthesis of N*^^-(2-succinylamido)myristic acid human insulin.
a. Preparation of a-aminomvristic acid methyl ester.HCl. To methanol (5 ml, Merck) at -10"C, thionyl chloride (0.2 ml, Aldrich) was added dropwise while stirring vigorously. Then, a- aminomyristic acid (0.7 g, prepared from the a-bromo acid by reaction with ammonia) was added. The reaction mixture was

stirred at room temperature overnight, and then evaporated to dryness. The crude product (0.7 g) was used directly in step b.
b. Preparation of N-succinoyl-a-aminomyristic acid methyl
ester.
a-Aminomyristic acid methyl ester,HCl (0.7 g) was dissolved in chloroform (25 ml, Merck). Triethylamine (0.35 ml, FluJca) was added, foilowed by succinic anhydride (0.3 g, Fluka). The reaction mixture was stirred at room temperature for 2 hours, concentrated to dryness, and the residue recrystallized from
ethyl acetate/petroleum ether (1/1). Yield: 0,8 g.
c. Preparation of N-(succiniinidylsuccinoyl) -a-aminomyristic
acid methyl ester.
N-succinoyl-a-aminomyristic acid methyl ester (0.8 g) was
dissolved in dry DMF (10 ml, Merck, dried over 4A molecular
sieve). Dry pyridine (80 ^1* Merck), and di(N-suc-
cinimidyl)carbonate {1.8 g, Fluka) were added, and the reaction
mixture was stirred overnight at room temperature. The
evaporation residue was purified by flash chromatography on
silica gel 60 (Merck), and recrystallized from 2-
propanol/petr oleum ether (1/1). Yield of N-
(succinimidylsuccinoyl)-a-aminomyristic acid methyl ester: 0.13
g, m.p. 64-66'C.
d. Reaction of (Al, Bl) -dj.Boc human insul in with N-
fsuccinimidylsuccinoyl)-a-aminomyristic acid methyl ester.
The reaction was carried out as in Example 20 b., but using N-
(succinimidylsuccinoyl)-a-aminomyristic acid methyl ester (16
mg) instead of N-(succinimidylsuberoyl)-D-thyroxine. After
removal of the trifluoroacetic acid in vacuo, the evaporation
residue was treated with O.IM sodium hydroxide at O'C to
saponify the methyl ester. When the saponification was judged
to be complete by reversed phase HPLC, the pH value in the
solution was adjusted to 3, and the solution was lyophilized.
After purification by preparative reversed phase HPLC and
desalting on a PD-10 column, the yield of NB29-(2-
succinylamido)myristic acid human insulin was 39 mg.

The product of this example is thus human insulin wherein the €-amino group of Lys"^ has a substituent of the following structure: CH3 (CHj) „CH (COOH) NHCOCHjCHjCO-.
Molecular mass of the product found by MS: 6130, theory: 6133.
EXAHPLE 22
Synthesis of N'^^-octyloxycarbonyl human insulin.
The synthesis was carried out as in Example 20 b., but using n-
octyloxycarbonyl N-hydroxysuccinimide (9 mg, prepared from n-
octyl chloroformate (Aldrich) and N-hydroxysuccinimide),
instead of N-(succiniraidylsuberoyl)-D-thyroxine. The yield of
NB29-octyloxycarbonyl human insulin was 86 mg.
The product of this example is thus human insulin wherein the e-amino group of LysB29 has a substituent of the following structure: CH2(CH2)7OCo
Molecular mass of the product found by MS: 5960, theory: 5964.
EXAMPLE 23
Synthesis of N*^"-(2-succinylamido)palmitic acid human insulin.
a. Preparation of N-fsuccinimidylsuccinoyl)-a-amino palmitic
acid methyl ester.
This compound was prepared as described in Example 21 a.-c, using a-amino palmitic acid instead of a-amino myristic acid.
Reaction of (Ai.BD-diBoc human insulin with N=.
succinimidvlsuccinovli-a-aminopalmitictic acid methvl ester. The reaction was carried out as in Example 21 d., but using N-{succinimidylsuccinoyl)-a-aminopalmitic acid methyl ester instead of N-(succinimidylsuccinoyl)-a-aminopalmitic acid

methyl ester to give NB29-(2-succinylamido)palinitic acid human insulin.
The product of this example is thus human insulin wherein the e-amino group of LysB29 has a substituent of the following structure: CHjCCHj) i5CH(COOH)NHCOCH2CH2CO-.
EXAMPLE 24
Synthesis of NB29-(2-succinylamidoethyloxy)palmitic acid human insulin.
a. Preparation of N-f3UCcinimidyl5uccinovlW2-amimethvloxy
palmitic acid methyl ester.
This compound was prepared as described in Example 21 a.-c. but using 2-aminoethyloxy palmitic acid (synthesized by the general procedure described by R. TenBrink, J^ Org. Chem. 52 (1987)
418-422 instead of a-amino myristlc acid.
b. Reaction of fAl.BlWdiBoc Imfflan insulin with N-
(succiniroidvlsuccinovll-2-aminoethvloxvpalmitictic acid methvl
ester.
The reaction was carried out as in Example 21 d., but using N- (succinimidylsuccinoyl)-2-aminoethyloxypalmitic acid methyl ester instead of N-(succinimidylsuccinoyl) -a-aminomyristic acid
methyl ester to give NB29-(2-succinylamidoethyloxy)palmitic acid
human insulin.
The product of this example is thus human insulin wherein the e-amino group of LysB29 has a substituent of the following
structure: CH3(CH2)i3CH(COOH)NHCH2CH20COCH2CH2CO-.

EXAMPLE 25
Synthesis of NB29-lithocholoyl-a-glutamyl des(B30) human insulin.
The synthesis was carried out as in Example 13 using N-lithocholoyl-L-glutaraic acid a-N-hydroxysuccinimide ester, 7-tert-butyl ester instead of decanoic acid N-hydroxysuccinimide ester.
The product of this example is thus des(B30) human insulin wherein the e-amino group of LysB29 has a substituent of the following structure: lithocholoyl-NHCH(CH2CH2COOH)CO-,
Molecular mass of the product found by MS: 6194, theory: 6193. EZ&MPLE 26
Synthesis of NB29-3,3 *,5,5'-tetralodothyroacetyl human insulin.
The synthesis was carried out as in Example 10 using 3,3',5,5*-tetraiodothyroacetlc acid N-hydroxysuccinimlde ester, instead of decanoic acid N-hydroxysuccinlmlde ester.
Molecular mass of the product found by MS: 6536, theory: 6538.
EXAMPLE 27
Synthesis of NB29-L-thyroxyl human insulin.
The synthesis was carried out as in Example 10 using Boc-L-thyroxine N-hydroxysuccinimide ester, instead of decanoic acid N-hydroxysuccinimide ester.
Molecular mass of the product found by MS: 6572, theory: 6567,

EXAMPLE 28
A pharmaceutical composition comprising 600 nmol/ml of NB29-decanoyl des(B30) human insulin, l/3Zn^* in solution.
NB29-decanoyl des(B30) human insulin (1.2 ^mol) was dissolved in
water (0.8 ml) and the pH value was adjusted to 7.5 by addition
of 0.2 H sodium hydroxide. O.Ol M zinc acetate (60 ^1) and a
solution containing 0.75% of phenol and 4% of glycerol (0.8 ml)
was added. The pH value of the solution was adjusted to 7.5
using 0.2 M sodium hydroxide and the volume of the solution was
adjusted to 2 ml with water.
The resulting solution was sterilized by filtration and transferred aseptically to a cartridge or a vial.
EXAMPLE 29
A pharmaceutical composition comprising 600 nmol/ml of NB29- decanoyl human insulin, 1/2Zn2+ in solution.
1.2 umol of the title compound was dissolved in water (0,8 ml) and the pH value was adjusted to 7.5 by addition of 0.2 M sodium hydroxide. A solution containing 0.75% of phenol and 1.75% of sodium chloride (0.8 ml) was added. The pH value of the solution was adjusted to 7.5 using 0.2 M sodium hydroxide and the volume of the solution was adjusted to 2 ml with water.
The resulting solution was sterilized by filtration and transferred aseptically to a cartridge or a vial.
EXAMPLE 30
A pharmaceutical composition comprising 600 nmol/ml of NB29-lithocholoyl human insulin in solution.

1.2 umol of the title compound was suspended in water (0.8 ml) and dissolved by adjusting the pH value of the solution to 8.5 using 0.2 M sodium hydroxide. To the solution was then added 0.8 ml of a stock solution containing 0.75 % cresol and 4% glycerol in water. Finally, the pH value was again adjusted to 8.5 and the volume of the solution was adjusted to 2 ml with water.
The resulting solution was sterilized by filtration and transferred aseptically to a cartridge or a vial.

WE CLAIM:
1. An insulin derivative having the following sequence:

wherein
Xaa at positions A21 and B3 are, independently, any
amino acid residue which can be coded for by the genetic code
except Lys, Arg and Cys;
Xaa at position Bl is Phe or is deleted;
Xaa at position B30 is (a) a non-codable, lipophilic amino acid having from 10 to 24 carbon atoms, in which case an acyl group of a carboxylic acid with up to 5 carbon atoms is bound to the e-amino group of LysB29, (b) any amino acid residue which can be coded for by the genetic code except Lys, Arg and Cys, in which case the e-amino group of LysB29 has a lipophilic substituent or (c) deleted, in which case the e-amino group of LysB29 has a lipophilic substituent ; and any Zn2* complexes thereof,

provided that when Xaa at position B30 is Thr or Ala, Xaa at positions A21 and B3 are both Asn, and Xaa at position Bl is Phe, then the insulin derivative is a Zn2* complex.
2. The insulin derivative according to claim 1, wherein
Xaa at positions A21 and B3 are, independently, any amino acid residue which can be coded for by the genetic code except Lys, Arg and Cys;
Xaa at position Bl is Phe or is deleted;
Xaa at position B30 is a non-codable, lipophilic amino acid having from 10 to 24 carbon atoms and an acyl group is bound to the e-amino group of Lys829, wherein the acyl group is an acyl group of a monocarboxylic acid with up to 4 carbon atoms or of a dicarboxylic acid with up to 5 carbon atoms.
3. The insulin derivative according to claim 1, wherein
Xaa at positions A21 and B3 are, independently, any amino acid residue which can be coded for by the genetic code except Lys, Arg and Cys;
Xaa at position Bl is Phe or is deleted;
Xaa at position B30 is deleted or is any amino acid residue which can be coded for by the genetic code except Lys, Arg and Cys and the e-amino group of Lys829 has a lipophilic substituent which comprises at least 6 carbon atoms.
4. The insulin derivative according to claim 2, wherein Xaa at position B30 is selected from the group consisting of a-amino decanoic acid, a-amino dodecanoic acid, a-amino tetradecanoic acid and a-amino hexadecanoic acid.
5. The insulin derivative according to claim 2, wherein the acyl group bound to the e-amino group of Lys829 is selected from the group consisting of formyl, acetyl, propionyl and n-butyryl.
6. The insulin derivative according to claim 2, wherein the acyl group bound to the e-amino group of Lys829 is an acyl group of succinic acid.

7. The insulin derivative according to claim 3, wherein Xaa at position B30 is deleted.
8. The insulin derivative according to claim 3, wherein Xaa at position B30 is Asp,Glu,or Thr.
9. The insulin derivative according to claim 3, wherein the lipophilic substituent bound to the e-amino group of LysB29 is an acyl group derived from a carboxylic acid having at least 6 carbon atoms.
10. The insulin derivative according to claim 9, wherein the acyl group, which may be branched, comprises a main chain of carbon atoms 8-24 atoms long.
11. The insulin derivative according to claim 9, wherein the acyl group is an acyl group of a fatty acid having at least 6 carbon atoms.
12. The insulin derivative according to claim 9, wherein the acyl group is an acyl group of a linear, saturated carboxylic acid having from 6 to 24 carbon atoms.
13. The insulin derivative according to claim 9, wherein the acyl group is selected from the group comprising dodecanoic acid, tridecanoic acid and tetradecanoic acid.
14. The insulin derivative according to claim 1, wherein Xaa at position A21 is Ala, Gin, Gly or Ser.
15. The insulin derivative according to claim 1, wherein Xaa at position B3 is Asp, Gin or Thr.

16. The insulin derivative according to claim 1, wherein Xaa at position Bl is deleted.
17. The insulin derivative according to claim 3 which is NEB29-tetradecanoyl des(B30) human insulin.
18. An insulin derivative according to claim 3 which is any Zn2+ complex of NsB29-
tetradecanoyl des(B30) human insulin.
19. An insulin derivative according to claim 3 which is a Zn2+ complex of NeB29 -
tetradecanoyl des(B30) human insulin containing 2, 3, or 4 Zn2+ ions per insulin
hexamer.
20. An insulin derivative according to claim 3 which is NEB29-(lithocholoyl-glutamy) des(B30) human insulin.
21. An insulin derivative according to claim 3 which is any Zn2+complex of NEB29 (lithocholoyl-glutamyl) des(B30) human insulin.
22. An insulin derivative according to claim 3 which is a Zn2+ complex of NeB29 (lithocholoyl-glutamyl) des(B30) human insulin containing 2, 3, or 4 Zn2+ ions per insulin hexamer.

Documents:

0318-mas-1995 abstract-duplicate.pdf

0318-mas-1995 abstract.pdf

0318-mas-1995 claims-duplicate.pdf

0318-mas-1995 claims.pdf

0318-mas-1995 correspondence-others.pdf

0318-mas-1995 correspondence-po.pdf

0318-mas-1995 description (complete)-1.pdf

0318-mas-1995 description (complete)-duplicate 1.pdf

0318-mas-1995 description (complete)-duplicate 2.pdf

0318-mas-1995 description (complete).pdf

0318-mas-1995 drawings.pdf

0318-mas-1995 form-1.pdf

0318-mas-1995 form-19.pdf

0318-mas-1995 form-26.pdf

0318-mas-1995 form-29.pdf

0318-mas-1995 form-4.pdf

0318-mas-1995 others.pdf

0318-mas-1995 petition.pdf

« Previous Patent

Next Patent »

Patent Number

220879

Indian Patent Application Number

318/MAS/1995

PG Journal Number

29/2008

Publication Date

18-Jul-2008

Grant Date

10-Jun-2008

Date of Filing

16-Mar-1995

Name of Patentee

NOVO NORDISK A/S

Applicant Address

Inventors:

#	Inventor's Name	Inventor's Address
1	IB JOHASSEN
2	JOHN BROBERG HALSTROM
3	ASSER SLOTH ANDERSEN
4	SVEND HAVELUND
5	JAN MARKUSSEN

PCT International Classification Number

C07K 014/62

PCT International Application Number

N/A

PCT International Filing date

PCT Conventions:

#	PCT Application Number	Date of Convention	Priority Country
1			NA