IE63866B1 - Molecular cloning and characterization of a further gene sequence coding for human relaxin - Google Patents

Description

Molecular cloning and characterization of a further gene sequence coding for human relaxin This invention relates to the molecular cloning and characterization of a gene sequence coding for human relaxin. The invention is also concerned with recombinant DNA techniques for the preparation of human relaxin, prorelaxin and preprorelaxin.

In our Australian Patent Application No. 17906/83 (PF 5352/81, filed 12th August, 1982), we described the molecular cloning and characterization of a gene sequence coding for human relaxin. We have now found a second gene which also codes for human relaxin.

More specifically, there is disclosed an isolated and purified (i.e., cloned) human gene coding for prorelaxin, preprorelaxin, and the A and/or B and/or C peptide chains of human relaxin, methods for isolating and purifying the genes and a method for transferring the genes to and replicating the genes in a host cell. The cloned genes are expressed by the host cell when fused with a host-expressible prokaryotic or eukaryotic gene. The genes are thus useful in the production of human relaxin for therapeutic purposes. - 2 There is further disclosed the peptides human relaxin, prorelaxin and preprorelaxin, the individual peptide chains which comprise these sequences and modified forms of these peptides.

There are disclosed modified genes coding for the individual relaxin chains and for the individual relaxin chains and for the above-mentioned modified forms.

[Note: References used in the following description are collected at the end of the description.] Pioneering work by Hisaw (1926) suggested an important role for the peptide hormone relaxin in mammals through this effects in dilating the pubic symphysis, thus facilitating the birth process. Relaxin is synthesized and stored in the corpora lutea of ovaries during pregnancy and is released into the bloodstream prior to parturition. The availability of ovaries has enabled the isolation and amino acid sequence determination of relaxin from pig (James et al, 1977; Schwabe et al, 1977) rat (John et al, 1981) and shark (Schwabe et al, 1982). The biologically active hormone consists of two peptide chains (known as the A and B chains) held together by disulphide bonds, two inter-chain and one intra-chain. The structure thus closely resembles insulin in the disposition of disulphide bonds which has led to speculation of a common ancestral gene for these hormones (James et al, 1977; Schwabe et al, 1977).

Recombinant DNA techniques have been applied to the isolation of cDNA clones for both rat and porcine relaxins (Hudson et al, 1981; Haley et al, 1982), see also Australian Patent Application No. 11834/83 (PF 2696/82). Synthetic undecamer nucleotides, prepared on the basis of amino acid sequence information, were used as primers for the synthesis of cDNA probes greatly enriched in relaxin cDNA sequences which identified relaxin cDNA clones in libraries derived from both rat and porcine ovarian tissue. The relaxin structural gene was round to code in both cases for a single chain precursor which resembles preproinsulin in the overall configuration, i.e., signal peptide/B chain/C peptide/A chain.

In our Application No. 17906/83 we described the use of probes based on the C peptide region of porcine relaxin to select out a relaxin gene from a human genomic library. This approach resulted in the successful identification of a genomic clone which we have now designated Hl from which the structure of the entire coding region of a human preprorelaxin was determined.

The present invention arises from a continuation of the work described in Application No. 17906/83 in which we sought to confirm the gene structure described in that application. We have investigated cDNA clones in libraries derived from ovarian tissue from pregnant human females using as a probe a segment of the previously identified human Hl gene corresponding to approximately 300 nucleotides of the c peptide/A-chain coding region (amino acids 64-161). Positive cDNA clones were isolated and sequencing of these revealed a cDNA sequence which was not identical with the sequence previously established and which coded for a form of preprorelaxin different to that described in our earlier application.

We have also isolated from the human genomic library described in our copending Australian Patent application No. 17906/83 (PF 5352/82) a recombinant phage containing exon 1 of the H2 gene where exon 1 comprising the coding region of the signal, B-peptide and part of the c-peptide similar to that of the Hl-gene.

It is now believed that either or both the presently described gene which we have designated H2 and the Hl gene described in our Application No. 17906/83 are expressed in human reproductive tissue, for example ovary and placenta, and/or other tissues including but not limited to gut, brain and skin, since both genes express peptides with relaxin-like activity.

The corpora lutea of the ovary as well as decidual and placental tissues are the most likely sites for expression of relaxin-related genes. However, in view of the wide distribution of many peptide hormones it is highly likely that the relaxin gene reproductive tissues, gastrointestinal tract. is also expressed in nonincluding brain and the Relaxin has the general properties of a growth factor and is capable of altering the nature of connective tissue and influencing smooth muscle contraction. We believe that one or both of the gene structures described in this specification and that of Application No. 17906/83 to be widely distributed in the body. We suggest that the relaxin peptides expressed from these genes will play an important physiological role in addition to their well documented hormonal function during reproduction..

The following abbreviations are used in this description.

Hl - the relaxin gene described in Application No. 5 17906/83, being deduced from a genomic clone.

H2 - the relaxin gene described herein, being deduced from a cNDA clone.

DNA - deoxyribonucleic acid RNA - ribonucleic acid cDNa- complementary DNA (enzymatically synthesized from an mRNA sequence) mRNA- Messenger RNA A Adenine T - Thymine G - Guanine C - Cytosine U - Uracil The coding relationships between nucleotide sequence in DNa and amino acid sequence in protein are collectively known as the genetic code, which is set out below.

First Third position (5' end) Second position position (3' end) U C A G Phe Ser Thr Cys U Phe Ser Tyr Cys C U Leu Ser Stop Stop A Leu Ser Stop Trp G Leu Pro His Arg U Leu Pro His Arg C C Leu Pro Gin Arg A Leu Pro Gin Arg G He Thr Asn Ser U lie Thr Asn Ser C A He Thr Lys Arg A Met Thr Lys Arg G Val Ala Asp Gly U Val Ala Asp Gly C G Val Ala Glu Gly A Val Ala Glu Gly G The abbreviations used for the amino acids in the table are identified as follows. Phenylalanine (Phe) Histidine (His) Leucine (Leu) Glutamine (Gin) Isoleucine (He) Asparagine (Asn) Methionine (Met) Lysine (Lys) Valine (Val) Aspartic acid(Asp) Serine (Ser) Glutamic acid(Glu) Proline (Pro) Cysteine (Cys) Threonine (Thr) Tryptophan (Try) Alanine (Ala) Arginine (Arg) Tyrosine (Tyr) Glycine (Gly) Each 3-letter codon represented in the table, e.g., AUG, CAU (otherwise known as a deoxynucleotide triplet or nucleotide triplet) corresponds to a trinucleotide of mRNA, having a 5'-end on the left and a 3'-end on the right. The letters stand for purine or pyrimidine bases forming the nucleotide sequence. All DNa sequences given herein are those of the strand whose sequence corresponds to the mRNA sequence, with thymine (T) substituted for uracil (U).

According to the present invention there is provided a human H2-relaxin analogue characterized in that the H2relaxin A-chain. 1 2 3 4 5 6 7 8 9 10 11 12 Gin Leu Tyr Ser Ala Leu Ala Asn Lys Cys Cys His 13 14 15 16 17 18 19 20 21 22 23 24 Val Gly Cys Thr Lys Arg Ser Leu Ala Arg Phe Cys and the H2 - : relaxin B-chain: -1 1 2 3 4 5 6 7 8 9 10 11 Asp Ser Trp Met Glu Glu Val He Lys Leu Cys Gly 12 13 14 15 16 17 18 19 20 21 22 Arg Glu Leu Val Arg Ala Gin lie Ala He Cys 23 24 25 26 27 28 29 30 31 32 Gly Met Ser Thr Trp Ser Lys Arg Ser Leu are modified by the deletion of up to nine amino acids from the amino terminus of the A-chain, and/or the B-chain is shortened by up to nine amino acids at the amino terminus and/or up to nine amino acids at the carboxy terminus.

According to the present invention there is further provided a human H2-relaxin analogue characterized in that the A-chain: 1 2 3 4 5 6 7 8 9 10 11 12 Gin Leu Tyr Ser Ala Leu Ala Asn Lys Cys Cys His 13 14 15 16 17 18 19 20 21 22 23 24 Val Gly Cys Thr Lys Arg Ser Leu Ala Arg Phe Cys and the H2-relaxin B-chain: -1 1 2 3 4 5 6 7 8 9 10 11 Asp Ser Trp Met Glu Glu Val He Lys Leu Cys Gly 12 13 14 15 16 17 18 19 20 21 22 Arg Glu Leu Val Arg Ala Gin lie Ala He Cys 23 24 25 26 27 28 29 30 31 32 Gly Met Ser Thr Trp Ser Lys Arg Ser Leu are modified by the replacement of at least one of the natural amino acids with a different amino acid with the proviso that said analogue processes relaxin activity.

The invention will be further described and elucidated in the following discussion. Reference will be made to the accompanying drawings in which: Figure 1: shows an abbreviated restriction map and sequencing strategy for the cDNA clone in pBR322, genomic clone Hll, and GT10 cDNA clones a-f. Arrows indicate the direction of sequencing on end-labelled fragments (see methods). GT10 clones a-f were sequences by subcloning into an M13 vector as described hereinafter. Nucleotides are numbered from the AUG initiation codon, position 1-3, to the termination codon, position 554-556.

Figure 2: compares the amino acid and MRNA sequence of human preprorelaxin H2 (upper) with the corresponding Hl (lower) sequence. The sequences have been aligned to maximize homology with nucleotide homolugues by boxed-inareas. Amino acids are numbered from the start of the Bchain (H2 gene sequence starting at -1 and Hl sequence at +1) although this position represents only the hypothetical start of the B chain sequence and has been deduced simply from the homoloy. to the related porcine and rat preprorelaxin structures. The asterisk beneath Ala 45 in the C peptide denotes the position of an intron in the G/CA codon in both genes.

Figure 3: autoradiographs cf identical nitrocellulose strips taken from either a) Northern gel transfer of human ovarian RNA or b) λplaques corresponding tc the Hl gene (λΗ7) or H2 gene (xGT10-a) using as hybridization probes A: a random primed 600 bp H2 relaxin cDNA fragment (72-660), B: H2-specific 25 mer (483-507), C: El-specific 25 mer (483-507), D: Hl-specific 25 mer (248-272).

Figure 4 : autoradiographs of identical nitrocellulose strips following Northern gel transfer of human ovarian RNA using as hybridization probes fragments of the H2 cDNA clone in pBR322 (see Figure 1). A: 600 bp fragment (72-660) corresponding to most of the coding region B: 5' untranslated region (tc Hinf I site at nucleotide 30) C: 3’ untranslated region (from Hinf I site at nucleotide 660) D: 3’ untranslated region (from Hpa I site at nucleotide 850) Figure 5: comparison of the amino acid sequences of the B and A chains between the two human relaxin genes, human insulin, and other members cf the relaxin family. Boxed areas highlight residue which are conserved between the two human relaxin genes and the relaxin family. Arrows indicate probable sites of proteolytic cleavage with confirmation by protein sequencing data of the amino terminal residue of the B and A chains of porcine (Schwabe et al, 1977; James et al, 1977), rat (John et al, 1981), shark (Schwabe et al, 1982) and dogfish relaxins (Schwabe et al, 1983) .

The H2 mRNA sequence shown in Figure 2 was determined by the methods described hereinafter. For ease of comparison, the numbering of the amino acids previously used for the peptide derived from the Hl sequence has been maintained in the present description of the H2-derived peptide. The structure of Hlpreprorelaxin was deduced frcm the genomic sequence by comparison with the homologous structures of pig and rat relaxin. The H2-preprorelaxin structure was deduced by comparison with the Hl structure as well as the pig and rat structures. Confirmation of the A and B peptide chain structures has been provided by synthesis and chain recombination in vitro which produces a material which is biologically active in the uterine contraction assay.

It will be seen from Figure 2 that the present and previous sequences shew significant: differences as well as similarities. Notable are: (1) Significant amino acid differences in three main (a) the N-terminus of the B-chain (b) the N-terminus of the A-chain (c) the middle of the C-peptide. (2) Regions of strong homology in the B-chain and C-peptide: 47 (a) 120 identical bases rrom Val to lie , (b) 88 - 90 identical bases from Phe0* to Ser132.

The two genes are therefore very similar but the differences are sufficient to indicate that the H2- gene is indeed a second gene and not simply a polymorph of Hl.

The mode of in vitro processing of the Hl preprorelaxin is not yet fully known but by analogy with pig relaxin, cleavage of the signal peptide would be expected to occur at the Ala~^-Lys bond. Similarly excision of the Hl-C peptide is predicted to occur at Leu3^ - Ser33 and Arg^3^ - Arg^-3, thus giving the Hl-3 and Hl-A chains of respectively 32 and 24 residues (Figure 2) .

In H2 preprorelaxin Ala 1 has been replaced by Asp and so we would predict cleavage of the signal peptide after the alanine corresponding to position -2 in Hl.

Cleavage at the H2-B chain/C peptide junction is 32 expected after Leu by anlogy to alx other prorelaxms, thus leaving the H2-B chain with 33 residues. Cleavage at the H2-C peptide/A chain junction would occur after Arg136 by analogy to rat preprorelaxin, thus leaving the H2-A chain with 24 residues.

As noted in our studies on pig relaxin, there are' core sequences in the pig relaxin 3 and A chains which contain all the essential elements for biological activity. Our synthetic studies on the human relaxin -12chain show similar results, as set out in more detail hereinafter.

There is disclosed a gene for the expression of human H2-preprorelaxin.

Except where otherwise specified, all following references to the gene sequences for preprorelaxin, prorelaxin, relaxin and the signal, A, B and C peptides, and to the peptides themselves will be understood to refer to the H2 variants and to exclude the H1 variants.

More specifically, there is disclosed a double-stranded DNA fragment for the expression of human preprorelaxin, which comprises a coding strand and a complementary strand corresponding to the complete mRNA (codons -25 to 160) sequence shown in Figure 2 of the accompanying drawings.

The disclosure includes any subunit of the H2preprorelaxin gene sequence described herein, or any equivalent of the said sequence or sub-unit. Among any subunits to be included by this statement are the individual structural genes coding for the signal peptide chain and the separate H2-A and H2-B peptides and the H2-C chain of human preprorelaxin (see Figure 2) and any combinations of these chains, e.g., the genes for expressing the H2-A and H2-B peptides, separately or as prorelaxin (with the H2-C chain). The sub-units also include fragments and combinations of fragments of any of said gene sequences. -13There is disclosed a gene for the expression of human prorelaxin.

There is further disclosed a double-stranded DNA fragment for the expression of human prorelaxin, which comprises a coding strand and a complementary strand corresponding to the codons numbered as 1 to 160 of the mRNA sequence shown in Figure 2 of the accompanying drawings.

There are disclosed genes for the separate expression of the A, B and C chains of human relaxin or any combination of two or more of the said chains and any fragment or combination of fragments of the said chains.

More specifically, this disclosure provides doublestranded DNA fragments for the separate expression of the A and/or B and/or C chains of human relaxin (or fragments as described above) which comprise a coding strand and a complementary strand corresponding to the codons numbered -1 to 32, 33 to 136 and 137 to 160 of the mRNA sequence shown in Figure 2 of the accompanying drawings.

The genes described above in addition to the codons specified may also include the appropriate start and stop codons, e.g., AUG and UGA respectively (codons -25 and 161 in Figure 2). — 1 αThose skilled in the art will appreciate that polymorphic forms of the genes may exist. Such forms are included in the present invention.

The disclosure further includes the complements of the above sequences, sub-units or equivalents, and the Corresponding RNA sequences, sub-units or equivalents.

According to another aspect of the present invention there is provided a DNA transfer vector comprising the deoxynucleotide sequences corresponding to the genes of the present invention.

As shown above, the genetic code contains redundancies, that is certain amino acids are coded for by more than one codon. Thus the invention includes deoxynucleotide sequences in which the codons depicted in the drawings or their cDNA equivalents are replaced by other codons which code for the same amino-acid.

Furthermore, as already indicated above, peptides with relaxin activity may be produced which differ from the B and/or A chain structures of natural relaxin. Such differences may involve deletion of one or more amino acids and/or addition of further amino acids and/or substitution of different amino acids in the natural chains. thus the invention also includes genes and DNA transfer vectors as described above wherein one or more of the natural codons are deleted and/or are replaced by codons which code for amino acids other than that coded by the natural codon, and/or further codons are added to the natural sequence.

The transfer vectors of the invention may also include inter alia, genetic information which ensures their replication when transferred to a host cell. Such cells may include, for example, the cells of prokaryotic microorganisms, such as bacteria, yeasts and moulds, and also eukaryotic cells, including mammalian cells and cell lines.

Examples of transfer vectors commonly used in bacterial genetics are plasmids and the DNA of certain bacteriophages. Both phage DNA and bacterial plasmids have been used as the transfer vectors in the present work. It will be understood however, that other types of transfer vectors may be employed. The general techniques of forming such transfer vectors and transforming them into microorganisms are well known in the art.

The invention also includes a prokaryotic or eukaryotic cell transformed by any of the transfer vectors described above.

One preferred microorganism is the very familiar Escherichia coli. but any other suitable microorganism may be used.

According to a still further aspect of the present invention, there is provided a process for making a DNA transfer vector for use in maintaining and replicating a deoxynucleotide sequence coding for human preprorelaxin according to the present invention, characterized by ligating a deoxynucleotide sequence coding for human preprorelaxin with a DNA molecule prepared by cleaving a transfer vector with a restriction enzyme.

DNA transfer vectors for use in maintaining and replicating deoxynucleotide sequences coding for human prorelaxin and for the A and B chains of human relaxin may be similarly prepared from the appropriate deoxynucleotides.

The A and B peptide chains, and also prorelaxin and preprorelaxin may be prepared by the usual process of gene expression, that is by growing cells containing the appropriate transformed transfer vectors and isolating and purifying the required peptide(s) produced by the cells.

Thus, the invention further includes a process for making a fusion protein comprising the amino acid sequence of human preprorelaxin according to the present invention as its C-terminal sequence and a protein of a prokaryotic or eukaryotic protein as its N-terminal sequence, characterized by incubating a cell culture transformed by an expression transfer vector comprising a deoxynucleotide sequence coding for human preprorelaxin, prepared in accordance with the process described above.

Fusion proteins comprising the amino acid sequences for human prorelaxin and/or the A and/or B and/or C chains of human relaxin may be similarly prepared.

The fusion peptide products thus produced will be in the form of a fusion protein in which the desired peptide is linked with a protein of a prokaryotic or eukaryotic protein characteristic of the host cell. Such fusion proteins also form a part of this invention. -17The invention also includes a process for synthesizing human prorelaxin according to the invention comprising the A and B peptides separated from each other by a C peptide, characterized by incubating a culture of cells, transformed by an expression transfer vector comprising a deoxynucleotide sequence coding for said human prorelaxin, prepared as described above, under conditions suitable for expression of said sequence coding for human prorelaxin, and purifying human prorelaxin from the lysate or culture medium of said cells.

The peptide of interest can be recovered from the fusion product by any suitable known cleavage procedure.

As already indicated above the transfer vector may be modified by codon substitution/deletion/addition and such modifications will give rise to modified fusion peptides.

In this way appropriate modifications may be made to facilitate the cleavage of the fusion peptides, for example, at the junction of B/C or C/A chains or to modify the peptide chain behaviour during subsequent chemical or biological processing.

As indicated above, the invention also provides analogues of human H2-relaxin, prorelaxin and preprorelaxin.

Relaxin may be prepared by direct combination of the separate A and B chains by any of the procedures currently known and used for the preparation of insulin.

Also in a similar manner to insulin, relaxin may be prepared from prorelaxin by oxidising or otherwise converting the sulfhydryl groups on the A and B peptides -18of relaxin, prepared as described herein, to form disulfide crosslinks between said A and B peptides, and then excising the C peptides, for example, by an enzyme-catalyzed hydrolysis specific for the bonds joining the C peptide to the A and B peptides.

Accordingly, the present invention further provides a method for the synthesis of human H2-relaxin analogues which comprises combining the A and B chains of relaxin (in their full-length, shortened or modified forms) by methods known per se for combination of A and B chains of human insulin.

One such method comprises reducing a mixture of the Ssulphonated A and B chains and then allowing the mixture to oxidize in air.

We have also found that the efficiency of the above procedure is improved when one or both of the A and B chains is in the form of an S-thioethyl-cys derivative rather than the S-sulpho form.

In our Australian Patent Application No. 15413/83 (PF 4385/82) we also showed that one or both of the A and B chains of relaxin can be shortened at the amino and/or carboxy terminii without significant loss of biological activity and with improved combination yields. These techniques apply equally to the preparation of human relaxin.

The invention also provides a method for producing a human relaxin ana 1 ague which comprises the step of forming the shortened and/or modified B and/or A peptide chains and combining them by any of the methods described above.

Our investigations with both pig and human relaxin (Hl) show that relaxin activity may be present with human A chains as short as A(10-24) and B chains as short as B(1022) although the expected practical minima of pig relaxin are respectively A(4-24) and B(4-23). The peptide A(4-24)1q B(l-25) is already known to have relaxin activity.

In general, for the present relaxin structure (H2) the A chain can be varied from A(l-24) to A(10-24) and B chain from B(-l-32) to B(10-22).

The preferred combinations are derived from: A B (1-24) (-1-23) any of (2-24) with any of (up to) (3-24) (-1-32) Modifications of the B and/or A chains, in accordance with the present invention may involve either genetic modification, as described above, or chemical modification of the B and/or A chains (in either full-length or shortened form) prior to combination by the method of the invention. Two types of modification may be employed, either singly or in combination.

The first type involves the modification cf one or more of the amino-acids which occur in the natural or shortened E and/or A chains. Such modification will generally involve protection of active groups on one or more of the amino-acids by methods known per se, and the protecting groups may, if desired, be removed after combination of the (modified) A and B chains.

Examples of this type of modification include the acetylation, formylation or similar protection of free amino groups, including the N-terminal, amidation of C-terminal groups, or the formation of esters of hydroxyl or carboxylic groups. The formyl group is a typical example of a readily-removable protecting group.

The second type of modification includes replacement of one or more of the natural amino-acids in the B and/or A chains with a different amino acid (including the D-form of a natural amino-acid). This general type of modification may also involve the deletion of a natural amino-acid from the chain or the addition of one or more extra amino-acids to the chain.

The purpose of such modifications is to enhance the combination yields of the A and B chains, while maintaining the activity of the product, i.e., relaxin or an analogue thereof, or to enhance or modify the activity of the product for a given combination yield. Such modification may extend to the production of synthetic analogues which have relaxin-blocking or -antagonistic effects.

A specific example of the first type of modification is the modification of the tryptophan (Trp) residue at B2 by addition ot a formyl group.

Examples of the second type of modification are replacement of the Met moiety at B24 with norleucine (Nle), valine (Val), alanine (Ala), glycine (Gly), serine (Ser) or homoserine (HomoSer).

The invention in this aspect also includes human relaxin analogues formed from natural or shortened B and/or A chains modified in accordance with the invention as described above.

The A and B peptide chains, and also prorelaxin and preprorelaxin may be prepared by the usual process of gene expression, that is by growing a microorganism containing the appropriate transformed transfer vector and isolating and purifying the required peptide(s) produced by the microorganism.

The peptide products thus produced may be in the form of a fusion protein in which the desired peptide is linked with a portion of a procaryotic protein The invention is further described and illustrated by the following description of the experimental procedures used and the results obtained thereby.

Methods and Materials Messenger RKA isolation and cDNA cloning Human ovarian tissue obtained during surgery for the treatment of an ectopic pregnancy was quickly frozen on dry ice and the RNA isolated in 5M guanidinium thiocyanate (Merck) according to the method of Chirgwin et al. , 1979. Pcly-A+ RNA was converted into double stranded DNA (Wickers et al, 1978) and cloned either by the homopolymeric G/C tailing method into a pBR322 plasmid vector (Chang et al., 1978) or by the lambda packaging method using the XGT10 vector (Huynh et al., 1983). In our experience the efficiency 4 of transformation with the pBR322 method (10 reccmbinants/ug of cDNA) was far less efficient than the lambda technique (up to 10 θ recombinants/tg of cDNA).

Preparation of hybridization probes Radiolabelled probes were prepared by primed synthesis on various DNA fragments using denatured random primers of calf thymus DNA (Hudson et al., 1983, Taylor et al., 1976) . The DNA template (100-200 ng) was denatured with the random primers (1 ug) by boiling in 20 μΐ of water for 2 minutes. Synthesis was initiated by the addition of a 30 ul reaction mixture containing 50mM Tris-HCl pH 8.0, 50mM NaCl, ImM DTT, lOmil MgCl2, 5 units of E. coli DNA Polymerase 1 (Klenow fragment), 500 uM each of dCTP, dGTP, dTTP and 0.3 uM 32 α-[ P]-dATP (Approx. 3000 Ci/mmol, Amersham). After incubation at 37°C for 30 minutes the reaction was terminated by dilution inro 300 ul of a buffer containing 0.3M NaCl, lOml-l Tris-HCl, pH 8.0, ImM EDTA and passed through a Sephadex-G50 column, (1cm x 5cm) in the same buffer. The radiolabelled probe was collected from the peak fractions at void volume and precipitated with 2 volumes of ethanol at -20°C for 2 hours using tRNA (10 ug) as carrier.

Selection of specific cDNA clones To screen the human ovarian cDNA clone bank for relaxin specific sequences we used as a probe a segment of the previously identified human Hl gene corresponding to a 400 nucleotide segment coding for the C peptide and A-chain from amino acid 64, through the termination codon and including 80 bases of the 3' untranslated region. A single positive cDNA clone from the pBR322 library was isolated and sequenced. 23 unique recombinants were isolated from the xGTIO libraries, but of these only 6 were subjected to complete nucleotide sequence analysis.

DNA sequence analysis The sequencing strategy and an abbreviated restriction map of the cDNA clones are summarized in Figure 1. The recombinant plasmid in pBR322 was digested with restriction enzymes Hpa II (P), Hinf I (F) or Taq I (T) and end-labelled using reverse transcriptase and the appropriate a-labelled deoxynucleotide triphosphate (dCTP for Hpa II, and Taq I, dATP for Hinf I). Fragments were cleaved internally with a second restriction endonuclease and then separated by electrophoresis on 8% polyacrylamide gels prior to sequencing by the chemical degradation method of Maxam and Gilbert et al, 1977. cDNA clones in AGTIQ were sequenced by subcloning Eco Rl restriction fragments into Ml3mp9 and employing the techniques described by Sanger et al, (1977).

Southern and Northern gel analyses Performed on purified genomic DNA after restriction endonuclease cleavage by the method of Southern (1975) or on purified RNA. The DNA fragments which were used as probes were found to be specific for either exon I or exon II of the Hl genomic clone despite having a small amount of flanking sequences. These fragments were generated by subcloning into M13mp8 a 500 bp Alu I fragment of the λH7 clone in the case of the exon I probe, or a 400bp Eco RI-Ava II fragment for the exon II probe. A probe from the H2 cDNA clone was generated by digesting with Hint I and isolating a 300 bp doublet corresponding to the coding region from Asp 1 to the termination codon and including 110 bases of the 3' untranslated region (Figure 1). Oligonucleotide probes were synthesized by the phosphite chemistry method of Beaucage and Caruthers (1981) and were end-labelled with γ- P-ATP using T4 polynucleotide kinase. Hybridization conditions were calculated on the basis of the G+C content.

Isolation and nucleotide sequence analysis of the H2 genomic clone The human genomic lambda library of Lawn et al (1978) was screened by method described earlier (Hudson et al, 1983) except that a mixture of DNA fragments corresponding to exons I and II of the Hl genomic clone was used for the probe as described above. Positive phage were grown in litre scale liquid cultures, the DNA isolated and digested with restricticn endonucleases prior to mapping with the exon I and II probes. A 4 kilobase EcoRI fragment was found to contain the entire exon I coding region which differentiated this clone from the homologous Hl gene structure. This fragment was subcloned into M13mp8 and sequenced by the technique of Maxam and Gilbert (1977). After digesting with Ava I, fragments spanning the coding region were end-labelled and cleaved internally by a second restriction enzyme (Hpa II of Hinf I) to generate fragments suitable for sequence analysis.

Isolation of a cDNA clone Samples of human corpus luteum were made available to us as a result of surgical intervention in ectopic pregnancies or from lutectomy at the time of Caesarian section. From the KNA isolated from a single corpus luteum a cDNA library was constructed in pBR322 providing about 300 unique recombinants. Screening this library with an Hl-cDNA probe revealed a single recombinant with sequence homology to human relaxin I. To increase the total number of recombinants from such small amounts of ovarian tissue we constructed cDNA libraries using the xGTIO cloning system (Huynh et al, 1983). Screening with a relaxin-specific probe identified 23 unique cDNA clones of which six were characterized as shown in Figure 1. Nucleotide sequence analysis revealed that all 6 cDNA recombinants encoded fragments of the same relaxin structural gene (Figure 2), yet this sequence was different to the genomic clone reported earlier (Hudson et al., 1983).

We expected that this novel sequence corresponded to the second human relaxin gene (H2) which had been observed in genomic DNA.

Surprisingly, none of the cDNA clones contained a polyadenosine sequence at the 3’ end, although the size of cDNA clones in PBR322 and X3T10 (1800 bp and 1900 bp respectively) indicate that large transcription products were being synthesized during the cloning procedure. These two cDNA clones had overlapping sequence identity at the 3' terminus confirming that they were derived from the same mRNA structure. We attributed the loss of the poly-A tail either to premature termination of the double-stranding transcription reaction or to excessive SI nuclease degradation during the cloning procedure.

Isolation of a genomic clone corresponding to the second gene g A thorough screen of 10 recombinant phage from the human genomic library of Lawn et. al. , (1978) using mixed probes specific for exon I or II of the \H7 relaxin clone revealed 16 positive phage. Small scale restriction mapping analysis revealed that 14 of these recombinant phage corresponded to the Hl relaxin gene reported earlier (11 were identical to the λΗ7 genomic clone; 3 were identical to λΗ5 a different genomic clone of the Hl gene as previously reported by Hudson et. al. , 1983) . However, the other 2 recombinant phage were identical and had a unique restriction pattern characteristic of the H2 relaxin gene whose structure is given in Figure 1. The unusual ratio of recombinants reflects either their proportion in the original genomic library or results from selective growth during amplification. Southern blot analyses of this new recombinant phage (λΗΙΙ) using separate probes corresponding to either exon I or II of the λΗ7 clone, revealed that λΗΙΙ contained only the exon I coding region. Attempts to find a full length genomic clone corresponding to the H2 relaxin gene either in the library of Lawn et. al. (1978) or in another library (Dr. R. Crawford, unpublished) have so far been unsuccessful.

The nucleotide sequence of the relaxin coding region of λΗΙΙ was found to be identical to that observed in the cDNA clone shown in Figure 2. An intron interrupts the coding region in exactly the same position as in the λΗ7 genomic clone (Hudson et. al., 1983) suggesting that these genes arose by a gene duplication event at some point in evolution.

Northern gel analysis RNA was isolated from several samples of human corpora lutea taken from different individuals during surgical intervention for ectopic pregnancy or during Caesarian section operations. Northern gel analysis using probes made from the coding region of either relaxin gene revealed that two major mRNA species of approximate sizes 1000 bp and 2000 bp were present in five human ovarian RNA samples tested (Figure 3). The smaller mRNA species were 2-3 fold more abundant in the RNA samples tested and this result was independent of whether the probe used in the analysis corresponded to Hl or H2 relaxin indicating that high crosshydridization rates occur under our experimental conditions. To differentiate whether these two mRNA species represent the separate products of the Hl and H2 genes, oligonucleotide probes were synthesized over a region of minimum homology (60%) between the two relaxin genes (residues 137-144 in Figure 2). These synthetic 25 mers were radiolabelled by kinasing with 32 γ- P-ATP and used as hybridization probes under conditions shown to provide specificity for either the Hl or H2 gene (Figure 3). Northern gel analysis using these radiolabelled probes revealed that both mRNA species corresponded to products of the H2 gene. We could not detect any transcription products from the Hl gene using the specific probes, although low level expression (less than 5% of the H2 level) would have been difficult to identify.

To analyse the different mRNA transcripts from the H2 gene, we made specific probes from segments of the two large H2 cDNA clones corresponding to the coding region and 5' and 3’ untranslated regions (Figure 4). The larger mRNA transcript (approximately 2kb in length) selectively hybridized to segments of the 3 ’ untranslated region from both cDNA clones, from a position approximately 100 bases from the termination codon. A potential polyadenvlation signal exists in the nucleotide sequence of the cDNA clones, 140 bases from the termination codon, and this region does have homology to the porcine relaxin polyadenylation site. However, the question of whether the shorter mRNA product is polyadenylated near this position cannot be resolved until full length cDNA clones corresponding to both mRNA forms have been isolated and characterised.

In the absence of the genomic sequence of the H2 gene it is impossible to define the mechanisms leading to the formation of the two mRNA transcripts. It is possible, like the collagen and β-microglobulin genes, that cleavage of the primary RNA transcript could occur at alternative polyadenylation sites. On the other hand we cannot rule out the possibility of alternative splicing mechanisms such as occurs in the calcitonin, growth hormone and α-crystallin genes.

The primary structure of preprorelaxin encoded by the H2 gene The mode of in vivo processing of the human preprorelaxin genes is not yet fully understood and has to be deduced by analogy to the processing of porcine and rat preprorelaxins (Figure 5) . The predicted B and A chain structures for the Hl and H2 genes have been aligned to other members of the relaxin family and human insulin in Figure 5.

Cleavage of the signal peptide in Hl has been predicted (Hudson et al., 1983) to occur after a short side chain residue such as Ala -2 or -4 or after Ser-6. Cleavage after Ala-1 is consistent with the homology to porcine preprorelaxin and human preproinsulin. Similarly, cleavage of the H2 signal peptide probably occurs after Ala-2 by such analogy, although cleavage after Ala-4 or Ser-6 are other possibilities.

By analogy to rat and pig prorelaxins, cleavage at the B chain/C peptide junction would occur after Leu 32 in both Hl and H2 precursors. However, both human relaxin B chains possess at positions 29-30 the conserved dibasic sequence Lys-Arg, which is a known processing site in other prohormones such as proinsulin, ana cleavage here cannot be excluded.

Direct amino acid sequence analysis of relaxin isolated from corpora lutea of pregnancy will be required to · settle this point. In the meantime it seems that the most likely structure of the Hl B chain would be 32 residues in length (Lys 1 to Leu 32) and the H2 B chain would be 33 residues (Asp-1 to Leu 32).

Cleavage at the C peptide/A chain junction of Hl prorelaxin has been predicted (Hudson et al., 1983) to occur after Arg 136 within a group of 4 basic residues because the Arg-Fro imide bond at 137-138 would be resistant to proteolysis. H2 prorelaxin has the same sequence of 4 basic residues and a similar processing step after Arg 136 would result in both the Hl and H2 relaxin A chains being 24 residues in length.

Biological Activity of the H2 gene As noted in earlier studies on synthetic pig relaxin peptides, there are core sequences in the pig relaxin B and A chains which contain all the essential elements for biological activity. Our synthetic studies on the Hl relaxin peptides has shown that combination of the complete Hl A chain (Arg 137-CYS 160) to a shortened form the Hl B chain (Lys 1-Ser 25) produced material which possessed biological activity (Hudson et al. , 1983). Further studies on both the Hl and H2 gene structures using peptide synthesis reveals that both genes code for forms of relaxin which are biologically active in the rat uterine contractility assay.

Chemical Synthesis of a modified human relaxin H2 (hRLX) A(l-24) - 3(-1-24) (i) Synthesis of human relaxin A-chain, H2 hRLX A(l-24) The amino acid sequence corresponding to residues 1 to 24 of the human relaxin A-chain, deduced as described above from the nucleotide sequence of the cDNA clone, was synthesized by the solid-phase procedure according to the general principles described by Merrifield (e.g. Barany, G. and Merrifield, R.B. In The Peptides. Ξά. E. Gross & J. Meienhofer, Academic Press, N.Y., pp. 1-284, 1980). * Ν-α-tertiarybutyloxycarbony1 -4-methy1benzyl-L-cysteine (*hereinafter BOC) was coupled to a 1% crosslinked polystyrene resin via the phenylacetamidomethyi (PAM) linkage to a level of 0.30 mmole/gm using the method of Tam et al., (Synthesis 12, 955-957, 1979). The BOC-L-CYS-PAM resin (8.0 gm) was transferred to the reaction vessel of a Beckman Model 990 Peptide Synthesizer and the amino acid sequence from residues 23 through to 1 was assembled by the stepwise addition of each suitably protected amino acid. The amino terminal BOC protecting group of each amino acid was removed by treatment of the resin with 35% trifluoroacetic acid in methylene chloride for 30 minutes followed by neutralization with 5% diisopropylethylamine in methylene chloride for 15 minutes. After each treatment the resin was washed thoroughly with methylene chloride. The next amino acid in the sequence (suitably protected at the a-araino with the BOC group and where necessary with the 7side-chain functional group appropriately protected) was coupled to the resin using dicyclohexylcarbodiimide (DCC) . The resin was stirred with the amino acid in methylene chloride for 10 minutes prior to the introduction of the DCC which was also dissolved in methylene chloride. A 2.5 molar excess (6.0 mmole) of amino acid and DCC was used for each coupling. After stirring for 1 hour a sample of the resin was removed frcm the reaction mixture and tested for the presence of free amino groups using the ninhydrin procedure of Kaiser et al.

(Anal. Biochem., 34 , 595-558, 1970). If the ninhydrin test was negative indicating complete coupling the reaction cycle was continued with BOC deprotection, neutralization and coupling of the next amino acid.

For a positive ninhydrin test the coupling reaction was repeated with further amino acid and DCC.

Amino acids with side-chain functional groups were used as the following protected derivatives: N-a-BOC2,6-dichlorobenzyl-L-tyrosine, N-a-BOC- ς-chlcrobenzyloxycarbonyl-L-lysine; N-a-BOC-Lserine O-benzyl ether; N-a-amyloxycarbonyl Q N -tosyl-L-arginine; Ν-α-BOC-L-threonine O-benzyi ether; N-a-BOC-S-ethyl mercapto-L-cysteine (for CYS at A-chain sequence position 15, 11 and 10).

Following the assembly of the 1-24 peptide sequence, the final BOC group on the amino terminal arginine was removed using the deprotectdion neutralization cycle and the peptide-resin dried in vacuo (wt of peptide resin 13.0 gm) . A portion of the peptide-resin (2 gm) was treated with anhydrous hydrogen fluoride in the presence of anisole (2 ml) at 0°C for 30 minutes. The total time for contact of the resin-peptide with hydrogen fluoride (HF) was kepr to a minimum (not more than 70 minutes) by rapid removal of the HF under oil-pump vacuum. The resin-peptide was then washed several times with ethyl acetate to remove excess anisole, the peptide extracted into 1M acetic acid and the solution lyophilized. The yield of crude peptide, (with the cysteines at positions 10, 11 and 15 still protected as the S-thioethyl derivative) was 392 mg. Initial purification of the crude peptide was by gel-filtration cn Biogel P10 in 0.1M acetic acid. The fractions representing the major peak frcm this column, which eluted at a position corresponding to a molecular weight of approximately 3000, were collected and lyophilized. Amino acid analysis of a sample of this peptide indicated that all the amino acids of the 1-24 seguence were present in the correct ratio.

Further purification cf the [S-thioethyl Cys^-hRLX A(l-24) peptide was effected by preparative reverse-phase HPLC on a Waters C-1S Bondapak column using a 0.1% TFA-water/acetonitrile solvent system.

A sample (80 mg) of the peptide purified by gel-filtration was S-sulfonated with a mixture of sodium sulfite and sodium tetrathionate (total reaction time of 3 hours) according to the method described by Du et al., (Scienria Sinica, 101, 84-104 (1961)). The precipitate which formed during the S-sulfonation reaction was removed by filtration and both the precipitate and the supernatant solution dialyzed against distilled water at 4°C for 48 hours. The contents of the dialysis bags were lyophilized to yield 39.5 mg of peptide from the supernatant solution and 20.3 mg of peptide from the precipitate which occurred during the S-sulfonation reaction. A sample of the 'soluble* [S-sulfo Cys^'^RLX A(l-24) peptide was purified by preparative reverse-phase HPLC on a Waters C-18 Bondapak column using a 0.1% TFA-water/acetonitrile solvent system. (ii) Synthesis of shortened human relaxin B-chain, E2 hRLX B(-l-24) The amino acid sequence corresponding to residues -1 to 24 of the H2 human relaxin B-chain was synthesized using the procedures described above and commencing with 6.0 gm N-a-tertiarybutyloxycarbonylL-methionine-O-benzyl-L-serine-phenylacetamido-methyl polystyrene resin wirh a loading of 0.5 mmole Met per gm. The side-chain protecting groups used in the A-chain synthesis were also employed for the B-chain including the S-ethyl mercapto derivative for both cysteines at positions 10 and 22. The glutamic acid residues at positions 4 and 5 and the aspartic acid residue at -1 were added as the N-ct-BOC-benzyl ester derivative. The glutamine at position 18 was coupled by the active ester procedure using Ν-α-BOC-L-glutamine-p-nitrophenyl ester in DMF. Following coupling of the tryptophan at position 2, 0.1% indole was added to the trifluoroacetic acid deprotecting reagent and to the subsequent methylene chloride washes.

The final weight of peptide-resin after removal of the BOC group from the amino terminal aspartic acid residue and vacuum-drying was 8.5 gm. A portion of the peptide resin (3.5 gm) was treated with anhydrous hydrogen fluoride in the presence of anisole (2 ml) at 0°C for 30 minutes and the B-chain peptide isolated using the procedure described above for the A-chain. 22 The crude [S-thioethyl Cys ' ] hRLX E(-l-24) (0.97 gm) was purified by gel filtration on EicGel P10 in IM acetic acid followed by preparative HPLC.

A sample (100 mg) of the peptide purified by gel filtration was S-sulfonated at pH 8.3 for 3 hours, the reaction mixture filtered and the precipitate and supernatant solutions dialyzed against distilled water. The 'soluble* peptide recovered after lyophilization was 42.4 mg; the 'insoluble' peptide was 59.5 mg. The S-sulfonated B-chain peptides were further purified by preparative HPLC using a C-18 reverse-phase column and 0.1% TFA-water-acetonitrile solvent system. (iii) Chain Combination The synthetic H2 hRLX A(1-24, and H2 hRLX B(-l-24) peptides were combined using the procedure described by Chance and Hoffmann (Australian Patent Application No. 68844/81) for insulin chains wherein the S-sulfonated peptides were mixed in a ratio of A : B of 2.6: 1 at a peptide concentration of 10 mg/ml in glycine buffer pH 10.5. Dithiothreitol in glycine buffer was then added in an amount to give a total of 1.0 sulfhydryl groups for each S-sulfo group. The reaction mixture was thenstirred in an open vessel for 24 hours.

As a further modification to this procedure we have found that the chain combination reaction to form biologically active relaxin proceeded efficiently when one or preferably both of the peptice chains are used as their S-thioethyl-Cys derivatives rather than in the S-sulfo form specified by Chance and Hoffmann (op.cit.) in the case of insulin. The use of S-thioethyl Cys peptides eliminates a reaction ana purification step required to convert the peptides to the S-sulfc derivatives. In our experience the S-sulfonation reaction of relaxin peptides is accompanied by side reactions which render the S-sulfo peptides difficult to purify resulting in low yields.

Using the above conditions chain combination yields from 1.5 to 6.0% have been achieved as measured by biological activity in the rat uterine contractility assay of Wiqvist & Paul (Acta Endocrinol., 29, 135-136, 1958).

Example of Chain Combination Reaction Human relaxin H2 [S-thioethyl Cys^ ' ' *^] A(l-24) (4.2 mg dry wt. , 2.4 mg peptide by amino acid analysis, 0.84 nmole) was dissolved in 500 μΐ of 0.1M glycine buffer pH 10.5 in a 3 ml stoppered plastic centrifuge tube. Human relaxin H2 [S-sulfo Cys10,11] B(-l-24) (1.60mg, 1.60 mg peptide by amino acid analysis, 0.33 nmole) dissolved in 200 nl of O.iM glycine buffer pH .5 was added and the mixture agitated. An aliquot (23.0 μΐ, 2.21 nmole DTT) of a stock solution of dithithreitol (DTT) made up in 0.1 M glycine buffer pH .5 (0.96 nmole DTT in 10 nl) was added to the peptide solution and following a brief agitation the reaction mixture was allowed to stand at 4°C for 24 hours open to the air. The mixture was then centrifuged and aliquots of the supernatant solution tested for relaxin biological activity in the rat uterine contractility assay. Alicuots of the reaction mixture inhibited the spontaneous contractions of the rat uterus in a dose-related manner. A 75 μΐ aliquot completely inhibited uterine contractions equivalent to a chain combination yield of 5.3% as compared to a native pig relaxin A22 B31 standard.

Synthesis of authentic human relaxin H2 : hRLX A(1-24) - B(-l-32) (i) Synthesis of full length H2 human relaxin B-chain ;hRLX B(-l-32) The amino acid sequence corresponding to residues -1 to +32 of the full length H2 human relaxin B-chain was synthesised using the procedures described above and commencing with 6.4 gm N-a-tertiarybutyloxycarbonyl-L-Ieucine phenyl acetamido methyl polystyrene resin with a loading of 0.23 mmol Leu per gm. The side-chain protecting groups used for the A(l-24) and B(-l - 24) peptides were also employed for the full length B-chain including the S-etnyl mercapto derivative for both cysteines at positions 10 and 22.

A modification of this strategy was the use of the N-formyi derivative of BOC-L-tryptophan for coupling at sequence positions 27 and 2.

The final weight of the peptide-resin following chain assembly was 8.2gm. A portion of the peptide resin (4.0gm) was treated with anhydrous hydrogen fluoride-anisole as described in previous examples to 10 yield 1.50gm of crude [S-thioethyl Cys ' Η-formyl Trp2'27j hRLX B(-l - 32). The crude pepuide was purified by gel filtration on BioGel P6 in 0.1ΙΊ acetic acid. The major peaks eluting from the gel filtration column were characterised by amino acid analysis. The fractions with analyses consistent with the -1 to +32 peptide sequence were collected and lyophilised. Deformylation of the tryptophan residues was effected by treating the peptide (lOOmg) with sodium hydroxide solution (5mi) pH 11.5 for 5 min. during which time the peptide precipitated from solution. The reaction mixture was neutralised to dissolve the peptide and applied directly to a BioGel P6 column in 0.1M acetic acid. Removal of the formyl groups from tryprophan was monitored by UV spectroscopy by following the disappearance of the N-formyl absorption at 300 nm and the appearance of the characteristic tryptophan spectra with an absorption maximum at 280 nm. Peptide fractions eluting from the column with the correct amino acid anlysis were collected and lyophilised.

Attempts to further purity the [S-thioethyl in ?? Cys ' ] hRLX B(-l - 32) peptide by preparative HPLC were not successful because of loss of peptide by adsorption to the column media. Peptide purified by gel chromatography was used directly in chain combination experiments. (ii) Chain combination of A(l-24) with B(-l - 32) : preparation of human relaxin H2 The synthetic S-sulfonated and S-thioethyl H2 human relaxin A(1-24) peptides were coupled to S-thioethyl H2 human relaxin B(-l - 32) using the same chain combination procedures described previously for the shortened B-chain (-1 - 24). Samples of the recombination mixture were tested for relaxin biological activity in the rat uterine contractility assay. Aliquots of the reaction mixture inhibited the spontaneous contractions of the rat uterus in a dose-related manner. A 100ul aliquot completely inhibited uterine contractions equivalent to a chain combination yield of 3.0% as compared to a native pig relaxin A22 B31 standard.

References Anderson, M.L., Long, J.A. and Hayashida, T.

Immunofluorescence studues on the localisation of relaxin in the corpus iuteum of the pregnant rat.

Biol. Reprod. 1.3, 499-504 (1975).

Beaucage, S.L. and Caruthers, M.H. Tetrahedron Lett. 22, 1859-1862 (1981) .

Chang, A.C.Y. Nature 275 , 617-624 (1978).

Chirgwin, J.M., Przybyla, A.E., MacDonald, R.J. and Rutter, W.J., Isolation of biologically active ribonucleic acid from sources enriched in ribonuclease. Biochem. 18 , 5294-5299, (1979).

Haley, J., Hudson, P., Scanlon, D., John, M., Cronk, M. , Shine, J., Trecrear, G. and Niall, H. DNA 1, 155-162 (1982) ' Hisaw, F.L. Proc. Soc. Exp. Biol. Med. 23, 661-663 (1926).

Hudson, P., Haley, J., Cronk, M., Shine, J. and Niall, H. Nature, 291, 127-131 (1981).

Hudson, P., Haley, J., John, M. Cronk, M., Crawford, R., Haralambidis, J., Tregear, G., Shine, J. and Niall H. Structure of a genomic clone encoding biologically active human relaxin. Nature 301, 628-631 (1983). Huynh, T., Saint, R. and Davis, R. (1983) personal communication.

James, R., Niall, H., Kwok, S. and Bryant-Greenwood, G. Nature, 267, 544-546 (1977).

John, M.J., Walsh, J.R., Borjesson, B.W. and Niall, H.D. Endocrinology 108, 726-729 (1981).

Lawn, R.M., Fritsch, E.F., Parker, R.C., Blake, G. and Maniatis, T. all 15, 1157-1174 (1978).

Maxam, A.M. and Gilbert, W. (1977) A new method for sequencing DNA. Proc. Natl. Acad. Sci. USA 74, 560-564.

Morrison, D.A., In: Methods in Enzymology, R. Wu, ed. (New York : Academic Press) pp. 326-331 (1979) .

Roychoudbury, R. Jay, E. and Wu, R. (1976) Terminal labelling and addition of homopolymer tracts to duplex DNA fragments by terminal decxynuclectidyl transferase. Nucleic Acid Res. 3, 863-877 (1976).

Sanger, F., Coulson, A.R., Barrell, B.G., Smith, A.J.H. and Roe, B.A., J. Mol. Biol. 143 , 161-178 (1989).

Schwabe, C., Gowan, L.K. and Reinig, J.W., Ann. N.Y. Acad. Sci. 380, 6-12 (1982).

Schwabe, C., McDonald, J.K. and Steinetz, B.C. Biochem. Biophys. Res. Comrnun. 75 , 503-510 (1977) . Southern, E.M., J. Mol. Biol. 98., 503-517 (1975).

Taylor, J.M., Illmersee, R., and Summers, J.Bicchim Biophys. Acta 442, 324-330 (1976).

Ullrich, A., Shine, J., Chirgvin, J., Picket, R., Tischer, E., Rutter, W.J. and Goodman, H.M. Rat insulin genes: construction of plasmids containing the coding sequences. Science 196, 1313-1319 (1977).

Vogt, V.M. Purification and further properties of single-strand-specific nuclease from Aspergillus oryzae. Eur. J. Biochem. 33., 192-200 (1973) .

Wickers, M.P., Buell, G.N. and Schimke, R.T. Synthesis of double-stranded Dna complementary to lysozyme, ovomucoid, and ovalbumin mRNAs. J.Biol. Chem. 253, 2483-2495 (1978).

Claims (29)

1. A human H2-relaxin analogue characterized in that » H2-relaxin A-chain: 1 2 3 4 5 6 7 8 9 10 11 12 Gin Leu Tyr Ser Ala Leu Ala Asn Lys Cys Cys His 13 14 15 16 17 18 19 20 21 22 23 24 Val Gly Cys Thr Lys Arg Ser Leu Ala Arg Phe Cys and the H2-relaxin B-chain: -1 1 2 3 4 5 6 7 8 9 10 11 Asp Ser Trp Met Glu Glu Val He Lys : Leu Cys Gly 12 13 14 15 16 17 18 19 20 21 22 Arg Glu Leu Val Arg Ala Gin lie Ala He Cys 23 24 25 26 27 28 29 30 31 32 Gly Met Ser Thr Trp Ser Lys Arg Ser Leu are modified by the deletion of up to nine amino acids from 15 the amino terminus of the A-chain, and/or the B-chain is shortened by up to nine amino acids at the amino terminus and/or up to nine amino acids at the carboxy terminus.

2. A human H2-relaxin analogue according to Claim l characterized in that it consists of any of the A-chains 20 A(l-24), A(2-24), A(3-24) in combination with any one of the B chains B(-l-23) to B(-l-32), wherein, if the A chain is A(l-24), the B chain is not B(-l-32). -433. A human H2-relaxin analogue according to Claim 1 characterized in that the relaxin A and/or B chain are modified by the addition of a protective group to a free amino, carboxy or hydroxy group.

3. 5 4. A human H2-relaxin analogue characterized in that the A-chain: 1 2 3 4 5 6 7 8 9 10 11 12 GlnLeu Tyr Ser Ala Leu Ala Asn Lys Cys Cys His 13 14 15 16 17 18 19 20 21 22 23 24 Val Gly Cys Thr Lys Arg Ser Leu Ala Arg Phe Cys and the H2-relaxin B-chain: -1 1 2 3 4 5 6 7 8 9 10 11 Asp SerTrp Met Glu i Slu Val lie Lys Leu Cys Gly 12 13 14 15 16 17 18 19 20 21 22 15 Arg Glu Leu Val Arg Ala Gin lie Ala lie Cys 23 24 25 26 27 28 29 30 31 32 Gly Met Ser Thr Trp Ser Lys Arg Ser Leu are modified by the replacement of at least one of the natural amino acids vith a different amino acid vith the 20 proviso that said analogue possesses relaxin activity. 5. An analogue as claimed in Claim 4 characterized in that the A-chain is shortened by up to 9 aminoacids at the amino terminus and/or the B-chain is shortened by up to 9 amino-acids at the amino terminus and up to 9 amino-acids at the carboxyl terminus.

4. 6. A human H2-relaxin analogue according to Claim 4 or 5 wherein Met of B-24 is replaced with Val, Ala, Gly, or Ser.

5. 7. A human H2-relaxin analogue according to Claim 4 or 5 wherein Met of B-24 is replaced with Norleucine or Homoserine.

6. 8. A process for preparing a human H2-relaxin analogue according to any one of Claims 1 to 6, characterized in that the A and B claims are combined by a method known per se for combination of the A and B chains of insulin.

7. 9. A process for preparing a human H2-relaxin analogue according to any one of Claims 1 to 6, characterized in that the A and B chains are combined by reducing a mixture of the S-sulfonated and/or S-thioalkylated A and B chains, allowing the mixture to oxidize in air and recovering the human H2 relaxin analogue thus produced.

8. 10. A process for preparing a human H2-relaxin analogue to any one of Claims 1 to 6 from human H2-prorelaxin or an analogue, characterized in that the sulfhydryl groups on the A and B peptide chains are converted to disulfide crosslinks between said A and B peptides and the C-peptide is then excised.

9. 11. A human H2-preprorelaxin analogue which analogue comprises a shortened and/or modified form of the amino acid sequence set out in Figure 2, and from which analogue is derivable a human H2-relaxin analogue according to any one of Claims 1 to 6.

10. 12. A human H2-prorelaxin analogue which analogue comprises a shortened and/or modified form of the amino acid sequence set out in Figure 2 and from which analogue is derivable a human H2-relaxin analogue according to any one of Claims 1 to 6.

11. 13. A DNA fragment encoding a human H2-relaxin analogue according to any one of Claims 1 to 6.

12. 14. A DNA fragment encoding a human H2-preproelaxin analogue according to Claim 11.

13. 15. A DNA fragment encoding a human H2-prorelaxin analogue according to Claim 12.

14. 16. A DNA transfer vector characterized in that it contains a DNA sequence corresponding to a DNA fragment according to Claim 13.

15. 17. A DNA transfer vector characterized in that it contains a DNA sequence corresponding to a DNA fragment according to Claim 14.

16. 18. A DNA transfer vector characterized in that it contains a DNA sequence corresponding to a DNa fragment according to Claim 15.

17. 19. A DNA transfer vector as claimed in any one of Claims 15 to 18 characterized in that it is a bacterial plasmid.

18. 20. A DNa transfer vector as claimed in any one of Claims 15 to 18 characterized in that it is a bacteriophage DNA.

19. 21. A cell transformed by a transfer vector as claimed in any one of Claims 15 to 20.

20. 22. A process for making a DNA transfer vector according to any one of Claim 16 to 20 which process comprises reacting a DNA sequence according to any one of Claims 13 or 14 molecule prepared by cleaving a transfer vector with a restriction enzyme.

21. 23. A process for making a fusion protein comprising an amino acid sequence consisting of all or part of the amino acid sequence of a human H2 relaxin, prorelaxin or preprorelaxin analogue according to any one of Claims 1 to 6, 11 or 12 as its C terminal sequence and a portion of a prokaryotic protein as its N terminal sequence, characterized in that it comprises incubating a microorganism transformed by an expression transfer vector comprising the appropriate deoxynucleotide sequence.

22. 24. A process for synthesizing a human H2-prorelaxin analogue according to Claim 12, comprising the A and B peptide separated from each other by a C peptide, which process comprises incubating a micro-organism transformed by an expression transfer vector comprising a deoxynucleotide sequence coding for said human prorelaxin analogue under conditions suitable for expression of said sequence coding for human prorelaxin analogue and purifying human prorelaxin analogue from the lysate or culture medium of said micro-organism.

23. 25. A fusion protein comprising an amino acid sequence characterized in that it consists all or part of the amino acid sequence of human H2-preprorelaxin, prorelaxin or relaxin analogue according to any one of Claims 1 to 7, 11 or 12 as its C terminal sequence and a portion of a procaryotic protein as its N terminal sequence.

24. 26. A gene encoding one or more of the modified A, B or C peptide chains of a human H2-relaxin, prorelaxinor preprorelaxin analogue according to any one of Claims 1 to 7, 11 or 12.

25. 27. A human H2-relaxin analogue substantially as hereinbefore described with reference to the drawings and examples.

26. 28. A process substantially as hereinbefore described with reference to the drawings and examples.

27. 29. A DNA fragment substantially as hereinbefore described with reference to the drawings and examples.

28. 30. A DNA transfer vector substantially as hereinbefore described with reference to the drawings and examples.

29. 31. A fusion protein substantially as hereinbefore described with reference to the drawings and examples.