of Equine Relaxin. Mohammed Akhter Hossain,1,2 Feng Lin,1 Soude Zhang,1 Tania Ferraro,1 Ross A. Bathgate,1. Geoffrey W. Tregear1 and John D. Wade1.
International Journal of Peptide Research and Therapeutics, Vol. 12, No. 3, September 2006 (Ó 2006), pp. 211–215 DOI: 10.1007/s10989-006-9020-9
Regioselective Disulfide Solid Phase Synthesis, Chemical Characterization and In Vitro Receptor Binding Activity of Equine Relaxin Mohammed Akhter Hossain,1,2 Feng Lin,1 Soude Zhang,1 Tania Ferraro,1 Ross A. Bathgate,1 Geoffrey W. Tregear1 and John D. Wade1 (Accepted February 7, 2006; online publication May 19, 2006)
In the equine industry, pregnancy loss during the third trimester constitutes a large percentage of fetal and neonatal mortality and represents a major financial loss and time investment for the breeder. Early identification of placental insufficiency would, in some cases, make it possible to sustain the pregnancy through medical intervention. Recent work suggests that relaxin is a valuable clinical tool for diagnosing placental insufficiency and monitoring treatment efficacy in mares. Relaxin is a polypeptide member of the insulin superfamily that consists of a two-chain structure and three disulfide bonds in a disposition identical to that of insulin. It is typically produced in the ovary during pregnancy and has primary roles in maintaining mammalian pregnancy and facilitating the delivery of the young via remodelling of the reproductive tract. The placenta is the primary source of relaxin in the mare during pregnancy. Its primary structure has been determined and shown to be the smallest of the known mammalian relaxins. It consists of a 20 residue A-chain and a 28-residue B-chain. To undertake detailed biophysical and biological characterization of the peptide, its chemical synthesis was undertaken using regioselective disulfide formation methods. The synthetic equine relaxin showed typical a-helical structure under physiological conditions. The peptide was found to bind to the relaxin receptor, LGR7, in vitro, and its binding affinity was found to be higher than that of the ‘‘gold standard’’, porcine relaxin, and similar to that of the human relaxin-2 (H2 relaxin).
KEY WORDS: Circular dichroism spectra; equine relaxin; regioselective disulfide bond formation; solid phase peptide synthesis.
immediately after birth represent a major financial loss and time investment for the breeders. Recent studies have shown that problems with the placenta are the most important factors contributing to lateterm abortion and death in foals (Ryan et al., 1998). In a different study, it has been found that the placenta is the primary source of relaxin hormone in pregnant mares (Stewart and Stabenfeldt, 1981; Stewart et al., 1982; Klonisch et al., 1997). Therefore, it is possible to have a correlation between relaxin, and placental function. Ryan et al. (1999) identified
INTRODUCTION In the equine industry, pregnancy loss during the third trimester and death of foals in the period 1
2
Howard Florey Institute of Experimental Physiology and Medicine, University of Melbourne, 3010, Parkville, Victoria, Australia. Correspondence should be addressed to: Mohammed Akhter Hossain, Howard Florey Institute of Experimental Physiology and Medicine, University of Melbourne, 3010, Parkville, Victoria, Australia. Tel.: +61-3944-7285; Fax: +61-3348-1707; e-mail: a.hossain@hfi.unimelb.edu.au.
211 1573-3149/06/0900–0211/0 Ó 2006 Springer Science+Business Media, Inc.
212 such an association between depressed systemic relaxin and placental insufficiency, which led them to conclude that circulating relaxin may be a useful biochemical means of monitoring placental function and treatment efficacy in the mare (Ryan et al., 2001). It has also been observed that there was a significant decline in serum relaxin concentration in mares with problematic pregnancies (Stewart et al., 1992). In all pregnancies when systemic serum relaxin concentrations are low, it is more likely that late-term abortion would occur, or the foals would be born hypoxic and smaller than normal foals but could survive with medical intervention. Therefore, measurement of relaxin levels is a simple but potentially important diagnostic tool for the veterinarians to predict problem pregnancies and deliveries, which may help to reduce the incidence of late-term abortions and stillbirths in mares, as well as predict whether a foal may have health problems when it is born. Since equine relaxin is not commercially available and its purification from placental tissue is labour-intensive and time-consuming, a means to produce sufficient quantities of equine relaxin for use in clinics and research is needed. Therefore, we have undertaken, for the first time ever, to synthesise equine relaxin chemically using regioselective disulfide bond formation strategy. We report here the synthetic strategy of preparation of equine relaxin along with its binding affinity for the relaxin receptor LGR7 and its biochemical activity.
MATERIALS AND METHODS Materials Fmoc-amino acid-derivatized PEG-polystyrene-resins for solid phase peptide synthesis were purchased from PerpSeptive Biosystems GmbH, Hamburg, Germany. Fmoc-amino acids were of the L-configuration and purchased from Auspep Pty Ltd., Australia. High-performance liquid chromatography (HPLC) columns were obtained from Vydac, Hesperia, USA. All other chemicals or solvents used for the synthesis or HPLC were of the highest purity available.
Methods Solid Phase Synthesis Both protected A- and B-chains on resins were synthesised using the continuous flow Fmoc solid-phase method (Atherton and Sheppard, 1998) as previously reported using an automatic Pioneer peptide synthesiser (PerSeptive Biosystems, USA) (Smith et al., 2001). The solid supports used were Fmoc-L-Cys(Trt)-PEG-PS and Fmoc-L-Lys(Boc)-PEG-PS for A- and B-chains respectively and 4
Hossain et al. fold molar excess of HBTU-activated Fmoc-amino acids were used throughout. All amino acid side chains were protected by trifluoroacetic acid (TFA)-labile protecting groups except for Cys7 (But) and Cys20 (Acm) in A-, and Cys22 (Acm) in B-chain. The scale of assembly was 0.2 mmol for A- chain and 0.1 mmole for B chain. Each coupling reaction was carried out for 1 h. Deprotection of the Fmoc was performed by 20% piperidine in DMF. At the end of all coupling, cleavage from the solid supports and side chain deprotection was achieved by 2 h treatment with TFA (95%) in the presence of scavengers: anisole (3%), ethandithiol (EDT, 2%) and triisopropylsilane (TIS, several drops). The crude peptides were identified by MALDI-TOFMS.
A Chain Intramolecular Disulfide Bond Formation The crude A-chain [Cys6, 11 S-thiol, Cys7 (But), Cys20 (Acm)] (200 mg, 79.20 lmoles) was dissolved in 100 ml of 6 M GnHCl and diluted with 500 ml of water, and to this was added 80 ml of 1 mM 2-pyridyl disulfide (DPDS) in MeOH (Maruyama et al., 1999). The reaction progress was monitored by analytical HPLC column, and the product was identified by MALDI-TOFMS (m/z 2518.55 [(M+H)+], calcd. 2523.08). The reaction was complete in 1 h at RT. The solution was freeze dried to give 210 mg of crude [Cys7 (But), Cys20 (Acm)] A-chain.
A Chain Conversion of Cys7 (But) to Cys7 (Pyr) [Cys7 (But), Cys20 (Acm)] A-chain (200 mg, 79.27 lmoles) and 2-DPDS (314.3 mg, 1.425 mmoles) were dissolved in 8 ml TFA and 0.88 ml anisole (Bu¨llesbach et al., 2001). The solution was chilled on ice before 8.88 ml of TFMSA/TFA (1:4 v/v) was added. The reaction was performed for 1 h at 0°C and the peptide was collected by ether precipitation, centrifuged and purified on a semipreparative RP-HPLC column. The freeze dried purified product weighed 25 mg (9.7 lmoles, 12.23%). The product [Cys7 (Pyr), Cys20 (Acm)] A-chain was identified by MALDI-TOFMS (m/z 2571.66 [(M+H)+], calcd. 2576.00).
Combination of [Cys7 (Pyr), Cys20 (Acm)] A-chain with [Cys10 (S-thiol), Cys22 (Acm)] B-chain [Cys7 (Pyr), Cys20 (Acm)] A-chain peptide (25.0 mg, 9.7 lmoles) was dissolved in 28 ml of 50 mM NH4HCO3 (pH8.5) (Lin et al., 2004). [Cys10 (S-thiol), Cys22 (Acm)] B-chain (31.41 mg, 9.7 lmoles) dissolved in water (18 ml) was added slowly to the solution of A-chain. The reaction was monitored by analytical HPLC column, and the product was identified by MALDI-TOFMS (m/z 5703.89 [(M+H)+], calcd. 5704.00). The reaction was complete by 1 h at RT in nitrogenous conditions. The reaction was stopped by adding neat TFA, and the product was purified by preparative RP-HPLC and freeze dried to give 12 mg (2.1 lmoles, 21.6%) of [Cys20A (Acm)/ Cys22B (Acm)]-A-B.
Equine Relaxin The A-B peptide [Cys20A (Acm)/Cys22B (Acm)] (8 mg, 1.4 lmoles) was dissolved in glacial acetic acid (10 ml) and to this 3.95 ml of 20 mM iodine/acetic acid and then 0.75 ml of 60 mM HCl was added. After 1 h, the reaction was stopped by addition of 4 ml of 20 mM ascorbic acid and the reaction mixture diluted with 140 ml of H2O and purified with preparative HPLC, and freeze dried to give 0.4 mg (0.072 lmoles, 5.1%) of final product. The
Regioselective Disulfide Solid Phase Synthesis peptide was identified by MALDI-TOFMS as a single species (m/z 5560.60 [(M+H)+], calcd. 5560.40), and the purity was examined by analytical RP-HPLC (Fig. 2). The amino acid composition and peptide content was determined by amino acid analysis.
Circular Dichroism Spectroscopy Circular dichroism spectra (CD) were recorded between 190 and 250 nm on AVIV 202 spectrometer (AVIV Biomedical, Lakewood NJ, USA) at 25°C using 1 mm path length cell. The peptide was dissolved in 10 mM phosphate buffer (pH7.4) with 120 mM NaCl at a concentration of 0.125 mg/ml.
Receptor Binding Assays The ability of the synthetic equine relaxin to bind to the relaxin receptor, LGR7, was tested as previously described (Sudo et al., 2003) in comparison to H2 relaxin (courtesy of BAS Medical, San Mateo, CA) and porcine relaxin. All data were analysed using GraphPad PRISM (GraphPad Software, San Diego, CA) with a non-linear regression one site binding model being used to obtain pKi (competition binding).
RESULTS AND DISCUSSION Equine relaxin was previously isolated and purified by acetone extraction, gel filtration, and ion exchange chromatography (Stewart and Papkoff, 1986). In an attempt to develop a more rapid and efficient method for relaxin purification, the same group (Stewart and Papkoff, 1991) used affinity chromatography coupled with RP-HPLC. An extract of term equine placentas was passed through the affinity column and washed, and relaxin was eluted by a change in pH. Equine relaxin was sequenced by Edman degradation, and its sequence confirmed by fast atom bombardment mass spectrometry. Equine relaxin appears to be the smallest mammalian relaxin sequenced having two chains A (20 amino acids) and B (28 amino acids) linked by two intermolecular disulfide bonds, and having a third intramolecular disulfide bond in the A-chain (Fig. 1). Relaxin or other insulin-like peptides are commonly prepared in our laboratory by random combination of the individual A- and B-chains (Wade et al., 1997; Dawson et al., 1999; Tang et al., 2003). Following their chemical synthesis, the purified S-reduced chains were combined by aeration in solution to produce the target native peptide in
213 modest to good yield. Surprisingly, to date, we have been unable to successfully chain combine insulin-like peptide (INSL) 4 and H3 relaxin by the random combination method. For this reason, a ‘‘forced’’ technique called regioselective disulfide synthesis (RDS) has been applied to synthesise insulin 4 (Lin et al., 2004) and H3 relaxin (Bathgate et al., 2005). A variety of insulins, relaxin and their analogues have been synthesised successfully using RDS approach (Bu¨llesbach and Schwabe, 1991, 1995; Akaji et al., 1993). Therefore, to avoid the uncertainty of the random combination method, we chose this RDS approach to synthesise equine relaxin. Differential cysteine S-protecting groups were used to allow the directed formation of three disulfide bonds. Solid phase synthesis of the separate, selectively S-protected A- and B-chains followed by their purification and subsequent stepwise formation of each of the three disulfides via oxidation (Maruyama et al. 1992a, b), thioloysis (Bu¨llesbach and Schwabe, 2001) and iodolysis (Lin et al., 2004) led to the successful acquisition of equine relaxin. The RP-HPLC profile (Fig. 2) shows the high purity of synthetic equine relaxin. The conformation of the synthetic equine relaxin and the effect of TFE on its secondary structure were studied by CD spectroscopy (Fig. 3). The CD for H2 relaxin was also measured because of its sequence similarity with equine relaxin. The studies revealed that both equine and H2 relaxins form a significant degree of helical conformation along with b-sheet and/random coiled structure. The helix content of equine relaxin, calculated from the mean residual weight ellipticity at 222 nm, [h]222 (Scholtz et al., 1991), was found to be 26% which is lower than that of H2 relaxin (43.5%). It might be due to the shorter A as well as B chain relative to H2 relaxin. It is a common observation that peptides in aqueous TFE exhibit increased helix content probably because of strengthened peptide H-bonds (Luo and Baldwin, 1997). The CD spectra of both equine and H2 relaxins were measured in 20% TFE, and as expected, their helix contents were found to be significantly increased (Fig. 3). Equine relaxin demonstrated a high affinity for the human relaxin receptor, LGR7. Indeed the
pyroGlu-Leu-Ser-His-Lys-Cys-Cys-Tyr-Trp-Gly-Cys-Thr-Arg-Lys-Glu-Leu-Ala-Arg-Gln-Cys pyroGlu-Lys-Pro-Asp-Asp-Val-Ile-Lys-Ala-Cys-Gly-Arg-Glu-Leu-Ala-Arg-Leu-Arg-Ile-Glu-Ile-Cys-Gly-Ser-Leu-Ser-Trp-Lys
Fig. 1. Primary structure of equine relaxin.
214
Hossain et al.
18.797
% Specific binding
Human LGR7 [33P]-relaxin binding H2 relaxin Equine relaxin Porcine relaxin
100 80 60 40 20 0 0
0
10
20
30
Time (min) Fig. 2. RP-HPLC profile of purified equine relaxin. (Eluent A: 0.1% aq. TFA; eluent B; 0.1% TFA in Acetonitrile, [Gradient]=15–45% B for 30 min, Vydac analytical C4 column).
affinity for LGR7 (pKi=10.34±0.2, n=3) was similar to the native ligand H2 relaxin (pKi=10.21±0.36, n=3) and higher than porcine relaxin (pKi=9.32± 0.11, n=3; P
