Selective acylation of epsilon-amino groups

The present invention relates to the acylation of proteins. More particularly, the invention relates to a one-step process for selectively acylating the ε-amino group of proinsulin, insulin or an insulin analog in the presence of a free α-amino group.

The acylation of amino groups is one of the most common means employed for chemically modifying proteins. General methods of acylation are set forth in Methods of Enzymology, 25:494-499 (1972) and include the use of activated esters, acid halides, or acid anhydrides. The use of activated esters, in particular N-hydroxysuccinimide esters of fatty acids is a particularly advantageous means of acylating a free amino acid with a fatty acid. Lapidot et al., J. of Lipid Res. 8:142-145 (1967). Lapidot et al. describe the preparation of N-hydroxysuccinimide esters and their use in the preparation of N-lauroyl-glycine, N-lauroyl-L-serine, and N-lauroyl-L-glutamic acid.

Early studies of selectively acylating the amino groups of insulin are described in Lindsay et al., in Biochem. J. 121:737-745 (1971). Lindsay et al., describe the reactivity of insulin with N-succinimidyl acetate at low reagent concentration and near neutral pH as producing two mono-substituted products. Phe B1 -acetyl-insulin and Gly A1 -acetyl insulin. At pH 8.5, the amount of Phe B1 -acetyl insulin produced is lowered and Lys B29 -acetyl-insulin is also produced. Thus, Lindsay et al., conclude at pH 6.9 the order of reactivity is Glycine(A1)≈Phenylalanine(B1)>>Lysine (B29) and at pH 8.5 Glycine(A1)>Phenylalanine=Lysine (B29). Id.

Lindsay et al., U.S. Pat. No. 3,869,437, disclose the acylation of the B 1 amino acid with an acyl group containing up to seven carbons and optionally blocking the A 1 - and/or B 29 -amino group with an acyl group with up to four carbons. N-hydroxysuccinimide esters are described as particularly advantageous acylating agents. In order to produce the maximum yield of insulin acylated at the B 1 -amino group, the proportion of acylating agent is relatively low (one to not more than two molar equivalents of acylating agent). In addition, the maximum yield of mono-substituted B 1 product is produced at a pH at or near about pH 7. At pH 8.5 to 9.0, the yield of the desired B 1 acylated product falls off considerably in favor of additional substitution at positions A 1 and B 29 .

D. G. Smyth, in U.S. Pat. No. 3,868,356 and Smyth et al., in U.S. Pat. No. 3,868,357 disclose N-acylated, O-substituted insulin derivatives in which at least one of the A 1 , B 1 or B 29 amino acid amino groups is converted into a blocked amino group. The acylation is carried out with a relatively small excess of acylating agent, e.g., from 2 to 3 moles per amino group at a neutral or mildly alkaline pH, e.g., 7-8. The reaction proceeds in very high yield with the formation of the di-substituted derivative resulting from the reaction of the A 1 - and B 1 -amino groups. In the presence of excess acylating agent, e.g., up to 10 molar, the reaction proceeds additionally at the B 29 -amino group to form the tri-substituted derivative.

To selectively acylate insulin, Muranishi and Kiso, in Japanese Patent Application 1-254,699, disclose a five-step synthesis for preparing fatty acid insulin derivatives. Step one, the activated fatty acid ester is prepared; Step two, the amino groups of insulin are protected with p-methoxy benzoxy carbonylazide (pMZ); Step three, the insulin-pMZ is reacted with the fatty acid ester; Step four, the acylated insulin is deprotected; and Step five, the acylated insulin is isolated and purified. Most notably, selective acylation of one amino group is only achieved by using the pMZ blocking group to protect the other amino groups. Using this methodology, Muranishi and Kiso prepare the following compounds: Lys B29 -palmitoyl insulin (the ε-amino group is acylated), Phe B1 -palmitoyl insulin (the N terminal α-amino group of the B chain is acylated), and Phe B1 , Lys B29 -dipalmitoyl insulin (both the ε-amino and the N-terminal α-amino group are acylated).

Similarly, Hashimoto et al., in Pharmaceutical Research 6:171-176 (1989), teach a four step synthesis for preparing N-palmitolyl insulin. The synthesis includes protecting and deprotecting the N-terminal. A 1 -glycine and the ε-amino group of B 29 -lysine, with pMZ. Under the conditions described in the reference, two major acylated products are prepared, B 1 -mono-palmitoyl insulin and B 1 , B 29 -dipalmitoyl insulin.

Therefore, prior to the present invention, the selective acylation of the B 29 -N ε -amino group of insulin was carried out by protecting and subsequently deprotecting the α-amino groups. The present invention provides a selective one-step synthesis for acylating the ε-amino group of proinsulin, insulin and insulin analogs. It is quite surprising that the invention is able to selectively acylate the ε-amino group in a one step process in high yield. Thus, the invention eliminates the need to protect and subsequently deprotect other amino groups of the protein. The invention provides more efficient and less expensive means of preparing ε-amino acylated insulin derivatives.

The present invention provides a process of selectively acylating proinsulin, insulin, or an insulin analog having a free ε-amino group and a free α-amino group with a fatty acid, which comprises reacting the ε-amino group with a soluble activated fatty acid ester under basic conditions in a polar solvent.

All amino acid abbreviations used in this disclosure are those accepted by the United States Patent and Trademark Office as set forth in 37 C.F.R. § 1,822(B)(2).

As noted above, the present invention provides a highly selective, one step acylation of the ε-amino group of proinsulin, insulin or an insulin analog. The invention specifies conditions that preferentially acylate the ε-amino group over the α-amino groups. Generally, the mono-acylated α-amino group is produced in less than 5% yield.

The term “insulin” as used herein means human insulin, pork insulin, or beef insulin. Insulin possesses three free amino groups: B 1 -Phenylalanine, A 1 -Glycine, and B 29 -Lysine. The free amino groups at positions A 1 and B 1 are α-amino groups. The free amino group at position B29 is an ε-amino group.

The term “proinsulin” as used herein is a properly cross-linked protein of the formula:

wherein:

A is the A chain of insulin or a functional derivative thereof;

B is the B chain of insulin or a functional derivative thereof having an ε-amino group; and

C is the connecting peptide of proinsulin. Preferably, proinsulin is the A chain of human insulin, the B chain of human insulin, and C is the natural connecting peptide. When proinsulin is the natural sequence, proinsulin possesses two free amino groups: B 1 -Phenylalanine (α-amino group) and B 29 -Lysine (ε-amino group).

The term “insulin analog” as used herein is a properly cross-linked protein of the formula:

A—B

wherein:

A is a functional analog of the insulin A chain; and

B is a functional analog of the insulin B chain having an ε-amino group.

Preferred insulin analogs include insulin wherein:

the amino acid residue at position B 28 is Asp, Lys, Leu, Val, or Ala;

the amino acid residue at position B 29 is Lys or Pro;

the amino acid residue at position B 10 is His or Asp;

the amino acid residue at position B 1 is Phe, Asp, or deleted alone or in combination with a deletion of the residue at position B 2 ;

the amino acid residue at position B 30 is Thr, Ala, or deleted; and

the amino acid residue at position B 9 is Ser or Asp; provided that either position B 28 or B 29 is Lys.

In standard biochemical terms known to the ordinarily skilled artisan the preferred insulin analogs are LysB 28 Pro B29 -human insulin (B 28 is Lys; B 29 is Pro); Asp B28 -human insulin (B 28 is Asp); Asp B1 -human insulin, Arg B31 , B32 -human insulin, Asp B10 -human insulin, Arg A0 -human insulin, Asp B1, Glu B13 -human insulin, Ala B26 -human insulin, and Gly A21 -human insulin.

The term “acylating” means the introduction of one or more acyl groups covalently bonded to the free amino groups of the protein.

The term “selective acylation” means the preferential acylation of the ε-amino group(s) over the α-amino groups. Generally, selective acylation results in a ratio of the amount of mono-acylated ε-amino group product to mono-acylated α-amino group product greater than about 5. Preferably, the ratio is greater than about 10, and most preferably greater than about 50.

The term “fatty acid” means a saturated or unsaturated C 6 -C 21 fatty acid. The term “activated fatty acid ester” means a fatty acid which has been activated using general techniques described in Methods of Enzymology 25, 494-499 (1972) and Lapidot et al., in J. of Lipid Res. 8:142-145 (1967). The preferred fatty acids are saturated and include myristic acid (C 14 ), pentadecylic acid (C 15 ), palmitic acid (C 16 ), heptadecylic acid (C 17 ) and stearic acid (C 18 ). Most preferably, the fatty acid is palmitic acid. Activated fatty acid ester includes derivatives of agents such as hydroxybenzotriazide (HOBT), N-hydroxysuccinimide and derivatives thereof. The preferred activated ester is N-succinimidyl palmitate.

The term “soluble” indicates that a sufficient amount of ester is present in the liquid phase to acylate the insulin, insulin analog or proinsulin. Preferably, about 1 to 4 molar equivalents of activated ester per mole of insulin are in the liquid phase.

The term “basic conditions” as used herein refers to the basicity of the reaction. The reaction must be carried out with all the free amino groups substantially deprotonated. In an aqueous solvent or semi-aqueous solvent mixture, basic conditions means the reaction is carried out at a pH greater than 9.0. In a non-aqueous organic solvent, the reaction is carried out in the presence of a base with basicity equivalent to a pK a greater than or equal to 10.75 in water.

The term “cross-link” means the formation of disulfide bonds between cysteine residues. A properly cross-linked proinsulin, insulin or insulin analog contains three disulfide bridges. The first disulfide bridge is formed between the cysteine residues at positions 6 and 11 of the A-chain. The second disulfide bridge links the cysteine residues at position 7 of the A-chain to the cysteine at position 7 of the B-chain. The third disulfide bridge links the cysteine at position 20 of the A-chain to the cysteine at position 19 of the B-chain.

Before the present invention, one skilled in the art selectively acylated the ε-amino group by the use of a protecting group in a multi-step synthesis. Muranishi and Kiso, Japanese Patent Application 1-254,699, disclose a five-step synthesis for preparing acylated insulin derivatives. Likewise, Hashimoto et al., in Pharmaceutical Research 6:171-176 (1989), teach a four step synthesis for preparing N-palmitoyl insulin. TO selectively acylate the insulin, both references teach the use of the pMZ protecting group.

The present invention produces an N ε -acylated proinsulin, insulin, or insulin analog in a high yield, one step synthesis. The reaction permits the preparation of N ε -acylated proteins without the use of amino-protecting groups. The acylation is carried out by reacting an activated fatty acid ester with the ε-amino group of the protein under basic conditions in a polar solvent. Under weakly basic conditions, all the free amino groups are not deprotonated and significant acylation of the N-terminal amino groups results. In an aqueous solvent or semi-aqueous solvent mixture, basic conditions means the reaction is carried out at a pH greater than 9.0. Because protein degradation results at a pH range exceeding 12.0, the pH of the reaction mixture is preferably pH 9.5 to 11.5, and most preferably 10.5. The pH measurement of the reaction mixture in a mixed organic and aqueous solvent is the pH of the aqueous phase prior to mixing. The data in Table 1 demonstrates the effect of the basicity of the reaction on the selectivity of the reaction. The data presented in Table 1 was generated with human insulin acylated with two molar equivalents N-succinimidyl palmitate in 50% CH 3 CN/water.

Table 1 demonstrates that the acylation of the ε-amino group is dependent on the basicity of the reaction. At a pH greater than 9.0, the reaction selectively acylates the ε-amino group of B29-lysine.

In a non-aqueous solvent, the reaction is carried out in the presence of a base with basicity equivalent to a pK a greater than or equal to 10.75 in water in order to sufficiently deprotonate the ε-amino group(s). That is, the base must be at least as strong as triethylamine. Preferably, the base is tetramethylguanidine (TMG), diisopropylethylamine, or tetrabutylammonium hydroxide.

The choice of polar solvent is dependent largely on the solubility of the proinsulin, insulin, or insulin analog and the fatty acid ester. Most significantly, the solvent may be wholly organic. Generally acceptable organic solvents include DMSO, DMF and the like. Aqueous solvent and mixtures of aqueous and organic solvents are also operable. The selection of the polar solvents is limited only by the solubility of the reagents. Preferred solvents and solvent systems are DMSO; DMF; acetonitrile and water; acetone and water; ethanol and water; isopropyl alcohol and water; isopropyl alcohol, ethanol and water; and ethanol, propanol and water. Preferably, the solvent is acetonitrile and water; most preferably 50% acetonitrile. One skilled in the art would recognize that other polar solvents are also operable.

The ratio of the reactants is not critical. Generally it is preferred that the activated fatty acid ester be in molar excess. Preferably the reaction is carried out with 1 to 4 molar equivalents, most preferably 1 to 2 molar equivalents, of the ester. However, one skilled in the art would recognize that at very high levels of activated ester, bis or tri-acylated product will be produced in significant quantity.

The temperature of the reaction is not critical. The reaction is carried out at between 0 to 40 degrees Celsius and is generally complete in 15 minutes to 24 hours.

After acylation, the reaction is quenched, and the product is purified by standard methods such as reverse phase or hydrophobic chromatography. Thereafter, the product is recovered by standard methods such as freeze drying or crystallization.

Proinsulin, insulin and insulin analogs can be prepared by any of a variety of recognized peptide synthesis techniques including classical (solution) methods, solid phase methods, semi-synthetic methods, and more recent recombinant DNA methods. For example, Chance et al., U.S. patent application Ser. No. 07/388,201, EPO publication number 383 472, Brange et al., EPO 214 826, and Belagaje et al., U.S. Pat. No. 5,304,473 disclose the preparation of various proinsulin and insulin analogs and are herein incorporated by reference. The A and B chains of the insulin analogs of the present invention may also be prepared via a proinsulin-like precursor molecule using recombinant DNA techniques. See Frank et al., Peptides: Synthesis-Structure-Function Proc. Seventh Am. Pept. Symp., Eds. D. Rich and E. Gross (1981) which is incorporated herein by reference.

The following examples are provided merely to further illustrate the invention. The scope of the invention is not construed as merely consisting of the following examples.