Preparation and properties of alpha- and epsilon-amino-protected porcine relaxin derivatives

The chemical modification of the amino groups of B29 porcine relaxin resulted in pure derivatives of N alpha A1-citraconyl-B29 relaxin, N epsilon A7, N epsilon A16, N epsilon B8-tris [[[(methylsulfonyl)ethyl]oxy]carbonyl]-B29 relaxin (Msc3-relaxin), and N alpha A1, N epsilon A7, N epsilon A16, N epsilon B8-tetrakis [[[(methylsulfonyl)ethyl]oxy]carbonyl]-B29 relaxin (Msc4-relaxin). N alpha A1-Citraconyl-B29 relaxin was obtained after selective deprotection of fully acylated B29 relaxin derivatives. The quantitative reaction of N alpha A1-citraconylrelaxin with [[(methylsulfonyl)ethyl]-oxy]carbonyl succinimide ester followed by deprotection of the citraconyl group resulted in N epsilon A7, N epsilon A16, N epsilon B8-Msc3-B29 relaxin, the starting material for selective chemical modifications at the N terminus of the relaxin A chain. In mouse interpubic ligament assay both Msc3 and Msc4 derivatives of relaxin showed a bioactivity of 30%, while in the case of N alpha A1-citraconyl-B29 relaxin the bioactivity was reduced to 15%. When compared with unmodified relaxin, only the circular dichroic spectrum of N alpha A1-citraconyl-B29 relaxin revealed significant differences. Therefore, the loss in bioactivity of the N alpha A1-citraconyl-B29 relaxin seems to be related to the structural changes caused by the introduction of a negative charge at the N terminus of the A chain.     

Preparation and properties of porcine relaxin derivatives shortened at the amino terminus of the A chain

Porcine relaxins shortened at the N terminus of the A chain were produced after protection of all amino groups with the base-labile [[(methylsulfonyl)ethyl]oxy]carbonyl (Msc) protecting group. The first two amino acids were removed by cyanogen bromide digestion whereby simultaneously a free alpha-amino group was generated in position A3. The resulting des-ArgA1,MetA2-N epsilon A7,N epsilon A16,N epsilon B8-tris [[[(methylsulfonyl)ethyl]oxy]carbonyl]relaxin. was further shortened by preparative Edman degradation. The shortest derivative obtained was des-ArgA1,MetA2,ThrA3,LeuA4,SerA5,GluA6 -N epsilon A7,N epsilon A16,N epsilon B8-tris[[[(methylsulfonyl)ethyl]oxy]carbonyl]relaxin. The deprotection of the derivatives in alkaline media resulted in crude des-A(1-2)- to des-A(1-6)-relaxins, which were subsequently purified by gel filtration on Sephadex G-50 superfine followed by either ion exchange chromatography on CM-cellulose at pH 5.1 or high-performance liquid chromatography on reversed-phase columns. During the CNBr digest, a side product was isolated that was identified as the corresponding homoserine ( [HseA2]relaxin) derivative. Shortened relaxin derivatives and [HseA2]relaxin were characterized by reversed-phase chromatography, electrophoresis, end-group determination, and amino acid composition. Circular dichroism studies revealed a distinct change in the structure of relaxins that were shortened by three and more amino acid residues. In the mouse interpubic ligament assay, des-A(1-2)-relaxin and [HseA2]relaxin were fully biologically active while the bioactivity of des-A(1-3)-relaxin dropped to about 50%. Relaxins shortened by four and more amino acid residues were biologically inactive.(ABSTRACT TRUNCATED AT 250 WORDS)     

Relaxin structure. Quasi allosteric effect of the NH2-terminal A-chain helix

The NH2-terminal heptapeptide in the relaxin A-chain (Arg-Met-Thr-Leu-Ser-Glu-Lys) has been replaced by chemical means with three different helix-promoting peptides (Arg-Met-Ala-Ala-Ala-Ala-Ala-Ala-Ala-Ala-Ala-Ala, and the insulin segment Gly-Ile-Val-Glu-Gln). The partially protected NH2 terminally shortened relaxin derivative (N epsilon A16,N epsilon B8-bis(methyl-sulfonylethyloxycarbonyl)des-ArgA1,MetA2 , ThrA3,LeuA4,SerA5,GluA6,LysA7-B29-relaxin) has been prepared by a combination of cyanogen bromide digestion and Edman degradation of the epsilon-amino-protected derivative followed by mixed anhydride coupling with the synthetic peptides. All three derivatives have been isolated and purified by high performance liquid chromatography. Whole relaxins shortened at the NH2 terminus of the A-chain by 4 or more amino acid residues are biologically inactive in the mouse pubic symphysis assay (Büllesbach, E. E., and Schwabe, C. (1986) Biochemistry 25, 5998-6004). The introduction of the artificial peptides causes significant biological activity to reappear (about 30%). The loss of structural integrity of relaxins shortened by 4 or more residues of the A-chain NH2 terminus as observed by circular dichroism spectroscopy is largely reversed by the addition of the synthetic peptides. Our results suggest that no single amino acid in the NH2-terminal region of the A-chain is functionally important but that the presence of a helix is required for biological activity.     

Naturally occurring porcine relaxins and large-scale preparation of the B29 hormone

Porcine ovaries were collected from pregnant sows under conditions designed to keep autolysis to an absolute minimum. During the extraction the tissues were never allowed to warm up to 0 degree C until submerged in 1.6 N HCl. Isolation and fractionation of the various relaxin forms became possible by application of CM-cellulose chromatography at pH 5.5 and 7.8, gel filtration, and high-performance liquid chromatography. The new isolation procedure has made it possible to isolate and identify LeuB32 relaxin. Also, [Leu-PheA0]B29 relaxin was identified and the existence of a [Leu-PheA0]B32 relaxin may be deduced from our data. Controlled digestion of B-chain-extended relaxins with carboxypeptidase A led to the large-scale production of homogeneous B29 relaxin, a suitable starting material for controlled chemical modification of porcine relaxin.     

A0-phenylalanyl] relaxin (porcine): an active intermediate

A [phenylalanylA0] relaxin has been isolated as a byproduct during large scale porcine relaxin preparations, using ion exchange chromatography on CM-cellulose at pH 7.8 followed by high performance liquid chromatography on reversed phase columns. The elongation at the N terminus of the A-chain has been demonstrated by amino acid and sequence analyses of the isolated and carboxymethylated relaxin-A-chain. The phenylalanyl relaxin and B29 relaxin are indistinguishable by circular dichroism spectroscopy, in mouse pubic ligament assay, and radioimmunoassay. The occurrence of phenylalanyl relaxin may be caused by an incomplete conversion of prorelaxin to relaxin.     

Purification and sequence determination of canine relaxin

Relaxin immunological activity has been observed in the plasma of pregnant bitches, and preliminary studies in our laboratory indicated that the highest relaxin concentrations were found in placentas. Therefore, canine placentas were collected at term and also from spay and relaxin was purified by methods developed for equine relaxin. Tissue was prepared by homogenization and purification on a C₁₈ column. The preparation was further purified by stepwise elution ion-exchange chromatography, gel filtration, and gradient elution ion-exchange chromatography. One predominant peak in relaxin immunoactivity was collected. Canine relaxin was found to be larger than either porcine or equine relaxin as determined by SDS-PAGE. It migrated faster under reducing conditions, indicating a subunit structure. Purified canine relaxin was used for tracer and standard in a canine radioimmunoassay (RIA) using an antiporcine relaxin antibody. Concentrations of relaxin immunoactivity using the canine assay were up to 300-fold higher in placental preparations than those measured in the porcine relaxin assay. Sequence analysis of canine relaxin revealed a structure similar to other relaxins in the presence and placement of cystine residues.     

Total synthesis of human relaxin and human relaxin derivatives by solid-phase peptide synthesis and site-directed chain combination

Human relaxin, a two-chain protein hormone, was synthesized by solid-phase peptide synthesis in combination with a novel thiol-protecting group strategy whereby the three disulfide bonds could be synthesized sequentially and without error. The final product was shown to be homogeneous by reversed-phase high performance liquid chromatography and electrophoresis and had the correct amino acid composition and sequence. Tryptic digestion and peptide mapping of the synthetic relaxin by reversed-phase high performance liquid chromatography resulted in a pattern identical with that produced by standard tryptic relaxin fragments synthetized by different methods. Three human relaxin derivatives containing oxidized methionine, formyltryptophan, and bis[B13,B17-citrulline]-relaxin, were produced and their biological activity and structural similarity to human relaxin was assessed. All derivatives, except those containing modified tryptophan residues, showed indistinguishable circular dichroic spectra, indicating that the modifications did not cause significant structural changes. However, only human relaxin and the tryptophan- and methionine-protected relaxin derivatives showed bioactivity. The derivative in which the two arginines in positions B13 and B17 had been replaced by the uncharged isosteric amino acid citrulline were biologically inactive. This observation confirms preliminary studies (Büllesbach, E. E. and Schwabe, C. (1988) Int. J. Pept. Protein Res. 32, 361-367) that suggested that these two conserved arginines located in the midregion of the relaxin B chain are essential for the function of the hormone.     

Purification and characterization of recombinant porcine prorelaxin expressed in Escherichia coli.

In this report we describe the purification and characterization of recombinant porcine prorelaxin expressed in Escherichia coli. Nucleotide sequence encoding porcine prorelaxin was inserted into an E. coli expression vector, pOTS, and the recombinant plasmid was transformed into the E. coli host (AR120). Upon induction with nalidixic acid, the 19-kDa recombinant porcine prorelaxin was produced at a level of approximately 8% of the total accumulated cell protein. The recombinant prorelaxin was purified to homogeneity by CM-cellulose chromatography and reversed-phase HPLC, after refolding in the presence of reduced and oxidized glutathione and a low concentration of guanidine-HCl. The identity of the recombinant prorelaxin was confirmed by the correct size, immunoreactivity with antibodies against native porcine relaxin, and direct amino-terminal sequence analysis. Furthermore, the purified recombinant prorelaxin could be converted to the 6-kDa relaxin by limited digestion with trypsin. Trypsin was shown to cleave at the carboxyl side of Arg29 and Arg137 residues of the recombinant prorelaxin, producing the des-ArgA1-B29-relaxin, and degrade the 13-kDa connecting peptide into small peptides. Both the recombinant prorelaxin and converted relaxin were found to be biologically active in an in vitro bioassay for relaxin.     

Abnormal relaxin secretion during pregnancy in women with type 1 diabetes

To test the hypothesis that relaxin may play a role in the fetal abnormalities associated with pregnancy in type 1 diabetic women, we previously compared gestational relaxin concentrations in diabetic and clinically normal women using a porcine relaxin radioimmunoassay (RIA): Serum immunoactive relaxin was significantly (P < 0.001) elevated in the diabetic women. To confirm and extend this work in a larger group of subjects, we have now used an enzyme-linked immunosorbent assay (ELISA) specific for human H2 relaxin (the normal human gene product) to determine immunoactive serum relaxin concentrations in serial samples from 61 Type 1 diabetic and 21 normal pregnant women. Samples from 22 of the diabetic and nine of the normal women were also directly compared in the porcine relaxin RIA. ELISA-determined serum relaxin was higher (P < 0.001) at 24 and 36 weeks of pregnancy in type 1 diabetic women than in controls, confirming previous findings. However, the geometric mean increase in immunoactive relaxin concentration in identical samples from pregnant diabetic women over that of controls was significantly greater with the RIA than with the ELISA (271% vs 44%; P < 0.001). To investigate this discrepancy, the specificity and epitope selectivity of the RIA and the ELISA were compared using several synthetic polypeptides, including human relaxins H1 and H2, and relaxin and insulin derivatives. Both assays showed great specificity, but the porcine RIA selectively identified the epitopes of the receptor-binding domain of the relaxin B chain and cross-reacted strongly with H1 and H2 relaxins. In contrast, only the H2 peptide was detected by the ELISA antiserum. Therefore, the marked discrepancy between the RIA and the ELISA could be due to the presence in the diabetic samples of another relaxin-like molecule in addition to the normal H2 relaxin. The biological consequences of elevated serum relaxin in diabetic pregnancy remain to be elucidated.     

Antigenically important regions of the relaxin molecule: studies with monoclonal antibodies

A panel of 22 monoclonal antibodies was produced against porcine relaxin. Antibodies were shown to be specific for porcine relaxin and appeared to recognize antigenic determinants associated with at least two different areas of the molecular surface, one of which was the N-terminal region of the A chain.     

B-chain sequence and in situ hybridization of the rabbit placental relaxin-like gene product.

We reported that the nucleotide sequence of a cDNA generated from rabbit placental poly(A)(+) RNA using porcine preprorelaxin primers was identical to SQ10, a product of squamous differentiated tracheal epithelial cells. However, these results did not confirm that SQ10 was the biologically active rabbit relaxin that had been isolated previously yet not sequenced. In this study, a 7-kDa protein isolated from rabbit placentas exhibited relaxin bioactivity and cross-reacted with a porcine relaxin antiserum. A partial amino acid sequence of this protein revealed a sequence identical to that of SQ10. Although the amino acid sequence of the putative relaxin receptor-binding domain found in the B chain of relaxin was modified in SQ10 from CGRDYVR to CRNDFVR, the placental protein was bioactive. These results suggest that SQ10 is the rabbit relaxin. In situ hybridization, using an SQ10 riboprobe, indicated radiolabeling in the syncytiotrophoblast cells of the rabbit placenta. The pattern of labeling corresponded with the immunohistochemical staining for relaxin observed with use of a porcine relaxin antiserum. These results indicate that the syncytiotrophoblast cells are a site of synthesis for SQ10 and that the immunostaining is not solely the result of sequestering SQ10 through receptor-mediated endocytosis. A potential role for relaxin in implantation is discussed.     

Relaxin concentrations in endometrial, placental, and ovarian tissues and in sera from ewes during middle and late pregnancy

These studies were designed to determine the tissue source of ovine relaxin and to determine the feasibility of using the pregnant ewe for study of relaxin production and secretion. On Day 4 of gestation, ewes were laparotomized, the nonpregnant uterine horn was ligated, and the ovary not containing the corpus luteum was removed. During a second surgery at Day 45 (n = 8) or 140 (n = 9) of gestation, 10-ml blood samples were drawn from a uterine artery, the ovarian vein, and veins draining the pregnant and nonpregnant uterine horns. Endometrial, placental, and luteal tissues were obtained for immunocytochemistry and extraction. Relaxin was detected by a heterologous porcine radioimmunoassay (RIA) in 3 of 54 serum samples (701.3 +/- 25.4 pg/ml, mean +/- SEM). Relaxin was not detected in crude tissue extracts, but low quantities were detected by RIA following Sephadex G-50 column chromatography of tissue extracts. Total relaxin activity for all tissues was equivalent to 0.57 +/- 0.13 ng of porcine relaxin/g tissue (w.w.). Relaxin was not detected immunocytochemically by light or electron microscopy. These data indicate that low quantities of relaxin are present in tissues and sera of pregnant ewes.     

Determination of leukocyte DNA 6-thioguanine nucleotide levels by high-performance liquid chromatography with fluorescence detection

A HPLC method for determination of 6-thioguanine nucleotide in DNA was developed. Leukocyte DNA was isolated from peripheral blood, derivatized with chloroacetaldehyde and the formed etheno derivatives N(2),3-etheno 6-thioguanine (epsilon6TG), 1,N(6)-etheno adenine (epsilonA) and N(2),3-etheno guanine (epsilonG) were released from the DNA backbone by hydrolysis at pH 6.0 and 80 degrees C for 60 min. After extraction of epsilon6TG by immobilized metal ion affinity chromatography (IMAC) the sample was analysed by ion-pair reversed-phase HPLC with fluorescence detection. The limit of quantification was 9.0 nM and the intra- and interday precision ranged from 2.8 to 15.5%. In a small cohort of eight children with acute lymphoblastic leukaemia (ALL), a median of one 6-thioguanine base was found for each 3000 normal bases (range 1:2000-1:11000).     

A novel, simple bioactivity assay for relaxin based on inhibition of platelet aggregation

In humans, the relaxin hormone family includes H1, H2 and H3 isoforms and insulin-like peptides 3 to 6. The ever-increasing interest in relaxin as potential new drug requires reliable methods to compare bioactivity of different relaxins. The existing bioassays include in vivo or ex vivo methods evaluating the organ-specific responses to relaxin and in vitro methods based on measurement of cAMP increase in relaxin receptor-bearing cells. We previously demonstrated that relaxin dose-dependently inhibits platelet aggregation. On this basis, we have developed a simple, reliable bioassay for relaxin used to compare purified porcine relaxin, assumed as reference standard, with two recombinant human H2 relaxins, H3 relaxin, insulin-like peptides 3 and 5. Pre-incubation of platelets with relaxins (3, 10, 30,100, 300 ng/ml; 10 min.) caused the inhibition of ADP-induced platelet aggregation. Within the 10-100 ng/ml range, porcine relaxin showed the highest effects and a nearly linear dose-response correlation. Lower peptide concentrations were ineffective, as were insulin-like peptides 3 and 5 at any concentration assayed. Platelet inhibition was mediated by specific RXFP1 relaxin receptor and cGMP, whose intracellular levels dose-dependently increased upon relaxin. For comparison, we stimulated THP-1 cells, a relaxin receptor-bearing cell line, with porcine relaxin, human H2 and H3 relaxins at the above concentrations (15 min.). We observed a dose-related increase of intracellular cAMP similar to the trend of platelet inhibition. Insulin like peptide 5 was ineffective. In conclusion, this study shows that inhibition of platelet aggregation may be used to assess bioactivity of relaxin preparations for experimental and clinical purposes.     

Stimulatory effect of phorbol diester on relaxin release by porcine luteal cells in culture

The role of protein kinase C (PKC) activation in the modulation of secretion of the peptide hormone, relaxin, by porcine luteal cells was examined by use of a reverse-hemolytic plaque assay. In this assay, luteal cells were cocultured in monolayers with protein-A-coupled ovine erythrocytes. In the presence of porcine relaxin anti-serum and complement, a zone of hemolysis--a plaque--developed around relaxin-releasing luteal cells. The rate of development of plaques in timecourse studies was then used as an index of the rate of relaxin release. The tumor-promoting agent, phorbol 12-myristate 13-acetate (PMA) activates a phospholipid- and calcium-dependent kinase, PKC. This enzyme is present in high concentrations in porcine luteal tissue, although its physiological role(s) is unknown. We report here that PMA exerted a time- and dose-dependent stimulatory effect on relaxin release by enzyme-dispersed porcine luteal cells in culture. Maximum stimulation was achieved by 50nM PMA. In contrast, the non-PKC-activating phorbol ester, 4 alpha-phorbol-12,13-didecanoate, exerted no significant effect on the rate of relaxin in doses up to 1 microM. We further observed that a synthetic 1,2-diacyl-glycerol (1-oleoyl-2-acetyl-rac-glycerol; 125 microM) mimicked the action of PMA in stimulating relaxin secretion. These results are consistent with the view that activation of PKC provides at least one intracellular mechanism that regulates relaxin secretion by porcine luteal cells.     

Evidence for immunoreactive relaxin in boar seminal vesicles using combined light and electron microscope immunocytochemistry.

Light-microscope immunocytochemistry using the peroxidase-antiperoxidase technique and a polyclonal rabbit antiserum raised against purified porcine relaxin showed that cytoplasmic immunostaining for relaxin could be visualized in the epithelial cells of the seminal vesicle. No relaxin immunoreactivity was seen in the testis, epididymis, ductus deferens, prostate or bulbo-urethral gland. A ten times higher concentration of porcine relaxin antiserum was necessary to achieve immunostaining in the seminal vesicle comparable to that in the corpora lutea of pregnant sows. Ultrastructural examination showed that the epithelial cells of the boar seminal vesicle resembled typical protein-secreting cells with prominent rough endoplasmic reticulum and well-developed Golgi apparatus. The most striking feature of these cells was the accumulation of granules with a limiting membrane, which ranged from 200 to 600 nm in diameter and contained flocculent material of moderate electron density. Electron-microscope immunocytochemistry using the protein A-gold technique and relaxin antiserum demonstrated that the granules were the only intracellular organelles that showed immunoreactivity for relaxin. These results indicate that a relaxin-like substance is present in boar seminal vesicles and that the subcellular site of its localization is the granules, suggesting that the seminal vesicle produces and stores a relaxin-like substance, but that it is present at much lower concentrations than in the corpora lutea of pregnant sows.     

Self-association of human and porcine relaxin as assessed by analytical ultracentrifugation and circular dichroism.

The self-association properties of recombinant DNA derived human relaxin, and porcine relaxin isolated from porcine ovaries, have been studied by sedimentation equilibrium analytical ultracentrifugation and circular dichroism (CD). The human relaxin ultracentrifuge data were adequately defined by a monomer-dimer self-association model with an association constant of approximately 6 x 10(5) M-1, whereas porcine relaxin was essentially monomeric in solution. An approximate 5-fold increase in weight fraction of human relaxin monomer elicited by dilution of the protein resulted in no change in the far-UV CD spectrum at 220 nm. In contrast, after the same increase in weight fraction of monomer, the near-UV circular dichroism spectra for human relaxin exhibited a significant decrease in the amplitude for the CD bands near 277 and 284 nm. These CD bands, which may be assigned to the lone tyrosine in human relaxin, are superimposed on a broad envelope that is probably due to the three disulfide chromophores. Although both the human and porcine proteins contain two tryptophan residues, the near-UV CD spectra exhibit only a broad shoulder near 295 nm rather than the strong CD bands often found for tryptophan. Moreover, there is little change in this broad band after dilution of human relaxin to concentrations that resulted in a 4-fold increase in monomer weight fraction. These data suggest that dissociation of the human relaxin dimer to monomer is not accompanied by large overall changes in secondary structure or alteration in the average tryptophan environment, whereas there is a significant change in the tyrosine environment.(ABSTRACT TRUNCATED AT 250 WORDS)