Purification and characterization of porcine prorelaxin

JOURNAL

ELSEVIER

OF

biochemical and biophysical methods J. B&hem.

Biophys. Methods 31 (1996) 69-80

Research article

Purification and characterization of porcine prorelaxin Selena S. Layden *, Geoffrey W. Tregear Howard Florey Institute of Experimental Physiology and Medicine, University of Melbourne, ParkviNe, Victoria 3052, Australia

Abstract Relaxin is a two-chain 6-kDa peptide hormone. It is a member of the insulin family of peptides and is produced mainly during pregnancy to prepare the reproductive tract for birth. In the pig, relaxin is produced mainly by ovarian luteal cells. It is processed via the regulated pathway from a larger (18 kDa) precursor, prorelaxin. Protocols have been described for the purification of mature relaxin from the ovaries of pregnant gilts. Multiple forms of relaxin have been detected during isolation due to exopeptidase trimming of the peptide chains. To date, such trimming events have prevented purification of the larger relaxin precursor. Described here is a method for the isolation of milligram amounts of homogeneous and bioactive prorelaxin from porcine ovaries. Keywords: Prorelaxin; Peptide hormone; Protein purification

1. Introduction In order to Study the structure and activity of processed peptides, a source of the native material and its precursor are required. Native porcine relaxin may be isolated from the corpora lutea of pregnant gilts. To date, a purification protocol had not been designed for preparation of the native prohormone due to enzymic degradation during handling. Described here is a method for the isolation of homogeneous prorelaxin from porcine ovaries. Relaxin is a two-chain peptide hormone. It is made principally by the corpus luteum of the ovary during pregnancy [l] and its best characterized actions are on the smooth muscle and connective tissue of the female reproductive tract in preparation for parturition [2].

* Corresponding author. Fax: (61) (613) 347 1707. 0165-022X/96/$15.00 8 1996 Elsevier Science B.V. All rights reserved SSDI 0165-022X(95)00040-2

S.S. Layden, G.W. Tregear/ J. Biochem. Bivphys. Methods 31 (1996169-80

70

RLELLSKDEAENQRNLILKRFEEFNUSDKSASQQLERLSPQRESLTAKLXQPLNp

:

V F E L M

A Chain C Peptide

g

A :: T S’ LEEPQLETGPPAETMPS

Fig. 1, Primary structure and disulfide bond alignment

of porcine prorelaxin.

Newly described sources and actions of relaxin suggest it is more than just a hormone of pregnancy. Relaxin has been detected in the male [3,4] and in the brain [5] and breast tissue [6]. Binding sites for the hormone have been found in the rat heart [7] and brain

k31. There have been a number of functions suggested for the hormone as well as those associated with the reproductive tract. Relaxin affects heart rate [9,10] and blood pressure [l 11. It is also thought to regulate the release of neuropeptides such as oxytocin and vasopressin [ 121. There is increasing evidence that relaxin plays a physiological role in the regulation of water balance in the pregnant rat, at least [ 13-151. The structure and processing of native relaxin is of particular interest. Relaxin is a member of the insulin family of proteins [ 161. Like insulin, it is produced from a larger, prorelaxin precursor (Fig. 1). Prorelaxin consists of the relaxin B chain linked through its C-terminus to the N-terminus of the A chain by a connecting C peptide of N 100 residues [17]. The tertiary structure of prorelaxin is maintained by two interchain disulfide bonds between the A and B chains and one intrachain disulfide bond within the A chain. The N-terminus of prorelaxin is a pyroglytamyl residue which is formed early in the biosynthetic pathway within the cell [18]. Relaxin is transported by the regulated protein pathway to secretory granules where it is stored until secretion [19]. Like many proteins transported in this pathway, relaxin is processed by removal of the connecting C peptide. Following endopeptidase excision of the C peptide, the A and B chains are trimmed to their mature length by amino and carboxypeptidases to produce mature relaxin. See Fig. 2. The primary structure of relaxin A and B chains varies greatly between species. The disulfide bond - forming cysteines are found in the same positions of all known relaxins, except mouse relaxin which has an extra residue before the C-terminal cysteine. The only other conserved residues in the sequences of species known to date are the A chain Gly(l4) and the B chain Gly(121, Arg(l31, Arg(l7) and Gly(24) [20]. The amino acid sequence of the C peptide has been predicted from cloning and cDNA sequence studies for a number of species. The C peptide of prorelaxin is _ 100 residues [ 17,211, compared with that of proinsulin which is N 30 residues [22]. To date, functions of the relaxin C peptide have not been determined, nor is it known whether it undergoes further processing to bioactive products.

S.S. Layden. G.W. Tregear/J.

Biochem. Biophys.

Methods31 (1996) 69-80

Preprorelaxin

Prorelaxin

1

Exopeptidase trimming

Relaxin

Fig. 2. Postulated porcine prorelaxin processing events. Following exopeptidase removal of the C peptide, the A and B chains are trimmed to their mature lengths. Non-specific trimming of relaxins may occur during purification and result in product heterogeneity.

It has been suggested that the C peptide assists the folding of the prorelaxin molecule and aligns the A and B chains for correct disulfide bond formation. The X-ray structure of relaxin shows the C-terminus of the B chain to be spatially near the N-terminus of the A chain [23]. This information and also that from computer modelling of prorelaxin and relaxin [24] imply a C peptide of one hundred residues is not necessary to align the A and B chains correctly. A modified human prorelaxin molecule with the two chains linked by a short peptide of sequence RREFKR was correctly processed and secreted by the yeast S. cereuiseae [25] and was active in the mouse pubic symphysis bioassay. In another study a relaxin fusion protein was produced in E. coli with a C peptide of 13 residues. Upon isolation, refolding and proteolytic treatment, mature relaxin was obtained [26]. This information further suggests that a long C peptide is not required. The sequences of the known C peptides are well conserved between species compared with the A and B chain sequences. The conserved regions include potential processing sites consisting of single and dibasic amino acids [27,28]. It is possible the C peptide has biological functions of its own or is further processed to bioactive fragments. Unprocessed prorelaxin may also be biologically active. Prorelaxin is the major form of the hormone in the ovary of the rat late in pregnancy 1291. It has also been described in the serum of rats [30] and pigs [3 11.,The. porcine prorelaxin sequence was expressed in Chinese hamster ovary cells to test the, ability of furin to correctly process it to relaxin

72

S.S. Layden, G. W. Tregear / J. Biochem. Biophys. Methods

31 Cl 996169-80

[32]. Unprocessed prorelaxin was secreted and this, also, was found to be active in the human endometrial cell bioassay. Further study of prorelaxin structure, activity and processing is important for our understanding of the hormone. In order that such investigations be carried out, a source of homogeneous native prorelaxin is required. Purification of mature relaxin from the ovaries of the pregnant pig has been described [33]. Acid-acetone extraction was employed as described, [34], and the extract separated by gel filtration. Peaks containing relaxin bioactivity were then adsorbed to carboxymethyl cellulose (CMC) and upon elution with a linear salt gradient, three peaks were obtained. These peaks were very similar structurally, with only microheterogeneity detectable, and all were of high and similar potency in the mouse interpubic ligament relaxin bioassay [33]. The three peaks from these preparations corresponded to three different forms of relaxin, differing by up to four residues in the length of the B chain. It was suggested these different forms were produced by proteolytic trimming of the C-terminus of the B chain during purification. Purification of relaxin species by other investigators has also demonstrated microheterogeneity of product, again due to C-terminal trimming of the relaxin B chain [35-371. A method for extraction of small, acid-resistant peptides [38] was modified for use in isolation of relaxin [39]. Acidic buffers were used and purification carried out on octadecylsilica columns. Using this protocol, the proportion of relaxins with trimmed B chains was much decreased, and improved yields of homogeneous relaxin were obtained. Higher molecular mass forms of relaxin have been detected during isolation of the native hormone [4O]. Attempts to purify prorelaxin have shown it to be highly heterogeneous in nature due to exopeptidase trimming. Prorelaxin for research use has been made by recombinant methods [41] and to date there has not been a protocol published describing purification of the native protein. This paper describes a method for isolation of milligram quantities of homogeneous, nondegraded and nonprocessed porcine prorelaxin. The method, based on that for relaxin isolation [39], allows preparation of a product which is biologically active in the isolated rat atria assay.

2. Materials and methods All reagents used were analytical reagent grade. All of the extraction steps were carried out on ice or at 4°C and all analysis was carried out at room temperature unless stated otherwise. Whole ovaries from late-pregnant gilts were collected into liquid nitrogen and stored at - 80°C. Corpora lutea of pregnancy were decapsulated then homogenized in ice-cold extraction buffer consisting of 15% (v/v) trifluoroacetic acid (TFA), 5% (v/v) formic acid, 1% (v/v) NaCl, 1 M HCl and a cocktail of enzyme inhibitors consisting of 5 mM EDTA, 10 mM benzamidine, 1 mM 1, IO-phenanthroline and 0.5 Mm pepstatin. 75 ml of extraction buffer was used for each 10 g of decapsulated corpus luteum tissue.

S.S. Layden, G.W. Tregear/J.

0.6

A.

Biochem. Biophys. Methods 31 (1996) 69-80

-l

73

B.

0.6.

0.5.

go.4.

d 74 tjo.3.

0.1.

0.

0.7

0.7

0.6

0.6

0.5

0.5

g 0.4

so.4

s N

s c(

;0.3-

io.3

i 8 0.2-

4 $0.2

0.1-

0% 5

r

D.

0.1

0,

Fig. 3. Analytical RP-HPLC resolution of Cl8 Sep-pak column eluates. Material was eiuted from sep-pak columns using 0.1% TFA in acetonitrile/water at A, lO:W, B, 20: 80, C, 30:70; D, 40:60; E, 50:50, F, 80: 20. RP-HLPC separation was carried out using 0.1% TFA in acetonitrile (.5-65% over 60 min) and material detected at 214 nm.

74

S.S. Luyden,

G.W. Tregear/J.

Biochem. Biophys. Methods 31 (1996) 69-80

The homogenate was centrifuged at 15 000 rpm at 4°C for 20 min (Sorvall RC-5B refrigerated superspeed centrifuge). The supematant was collected and filtered successively through filter paper (Whatman, UK), a glass fibre prefilter (Sartorius, Germany) and a Millipore (Bedford, MA) 0.45 pm filter. Disposable octadecylsilica (Cl81 columns - Sep-pak classics - (Waters, Milford, MA) were joined in series (three for every 20 g of corpora lutea tissue) and connected to a peristaltic pump. The columns were washed with 80% acetonin-ile, 0.1% TFA (15 ml for each 10 g of tissue = 1 vol.) then equilibrated with 2 ~01s. of distilled water. The filtered ovarian extract was pumped onto the columns, with the effluent from the first application collected and reapplied. As is observed in relaxin preparations [39], the Sep-paks became yellow-brown in colour due to the adsorbed protein. They were washed with 1 vol. of 10% acetonitrile, 0.1% TFA then the material on the columns eluted with successive washes of 1 vol. each 20, 30, 40, 50 and 80% acetonitrile, 0.1% TFA. The effluent of each step was collected and analysed by reversed-phase high-performance liquid chromatography (HPLC). Analytical HPLC was carried out using an Aquapore BP-300 C8 column (Applied Biosystems, San Jose, CA) connected to a Waters 600 Multisolvent Delivery System. Column dimensions were 22 cm X 4.6 mm internal diameter, packed with C8 solid support of particle size 7 pm. A gradient of 0.1% TFA, 5% acetonitrile to 0.1% TFA, 65% acetonitrile over 30 or 60 min was used at a flow rate of 1.5 ml/min. Eluate from the column was detected at 214 nm using a Waters 484 Tunable Absorbance Detector. Porcine prorelaxin and relaxin in the column eluates were identified by coelution with standards previously purified and characterized in our laboratory. Those samples containing the target material were then subjected to preparative HPLC. Up to 20 ml of Sep-pak eluate was separated at a time using a Cl8 preparative column (Vydac, Hesperia, CA) and the gradient described above. The preparative column was of dimensions 25 cm X 22 mm internal diameter and packed with a solid support of lo-pm spheroidal particles. A flow rate of 10 ml/min was used and the eluate monitored at 230 nm. Eluted material was collected manually with the material from each peak pooled and tested for purity as well as coelution with the relaxin and prorelaxin standards. Homogeneity of the material was assessed using analytical HPLC as described. Buffers containing 0.1% TFA, as described above, and also buffers composed of 10 mM triethylamine acetate (TEAA) in 10% and 80% acetonitrile were used over a range of gradients to test that a discrete product had been obtained. Capillary electrophoresis was carried out using a Beckman System 2100 with System Gold software. Separation was carried out at 30 kV (30 PA) at 25°C using a capillary of 72 cm (50 cm to detector) X 50 pm. The sample was prepared in 20 mM sodium citrate buffer, pH 2.5, and the components from a 3-s injection were detected at wavelength 200 nm. The purified material was subjected to sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis using 15% gels and stained with Coomassie brilliant blue. Samples were run in a nonreduced form and after reduction with @mercaptoethanol. The size and homogeneity of the prorelaxin was confirmed via mass spectrometry, using a Finnegan Mat laser mat mass spectrometer in linear mode. Amino acid analysis of the whole molecule and of the fragments produced upon tryptic cleavage was carried out using a Beckman 6300 ion-exchange amino acid

S.S. Layden, G.W. Tregear/ 0.05

0.04

I ;

0.03

P

P

0.02

Iii -t

0.01

0

i-

J. Biochem. Biophys. Methods 31 (1996) 69-80

75

0.12

A.

11 ;

5

Elutim

time(mins)

I5

25

0.10

g 0.08 ,8 f 0.06

1 0.04 < 0.02

o- 35

45

55

6

I

I

,

4

5

6

,

I

7

8

I-

9

Eludon mne (mins)

Fig. 4. Purified native porcine prorelaxin assessed by analytical RP-HPLC and CE. (A) Analytical RP-HPLC: buffers 0.1% TFA in water/acetonihile, pH 2.5. (B) Analytical RP-HPLC: buffers 10 mM TEAA in water/acetonitrile, pH 7.5. (C) Analytical CE: separated at 30 kV, 30 PA in 20 mM sodium citrate, pH 2.5.

analyser fitted with a Hewlett Packard 3390A integrator. Acid hydrolysis of the samples was carried out in vacua at 110°C for 24 h. Isolated porcine prorelaxin was subjected to the isolated rat atria bioassay. Four separate samples were tested as described [lo], and compared with standard porcine relaxin.

3. Results HPLC analysis of eluates from the Sep-pak columns (Fig. 3) demonstrates most protein was eluted at concentrations of 30 to 40% acetonitrile. The two major peaks in these traces coeluted with previously purified and characterized porcine prorelaxin and relaxin. These peaks were separated using preparative HPLC and each tested for homogeneity. The peak that coeluted with porcine prorelaxin was assessed by analytical HPLC and found to be a discrete product. A single peak was detected using different HPLC buffers and also upon separation by capillary electrophoresis under various gradient and buffer conditions (Fig. 4). Six preparations of prorelaxin have been carried out using this method. From these experiments, the yield of HPLC-purified prorelaxin averaged 6 mg (range 4.2-7.1 mg) from an ovary of average wet weight 16.5 g. The weight of decapsulated corpus luteum tissue from such an ovary averaged 8.3 g. Thus, the yield was 0.72 mg of prorelaxin per gram wet weight of decapsulated corpus luteum tissue. Upon SDS-polyacrylamide gel electrophoresis, under both reducing and nonreducing conditions, the prorelaxin sample ran as a single band of apparent molecular mass 18 kDa (Fig. 5). This is the expected size of intact prorelaxin ,with no hydrolysis of the peptide bonds or processing having occurred. Mature relaxin isolated from the same

76

S.S. Layden, G.W. Tregear/ J. B&hem.

Size (kD)

Biophys. Methods 31 Cl996) 69-80

M(r) Marker.

A.

Fig. 5. SDS-polyacryamide gel electrophoresis of HPLC-putified sample. (B) Sample reduced using IS-mercaptoethanol.

native porcine prorelaxin.

(A)

8995.3

46.

42.

26s

22. I

4ooo

6mil

8ooo

loo00

12wo

16000

2m

Mass(m/z)

Fig. 6. Mass speceometer profile of purified native porcine prorelaxin. 18,200 Da (theoretical mass of porcine prorelaxin is 18,107 Da).

Main component

of detected

sample

S.S. Layden, G.W. Tregear/J.

Biochem. Biophys. Methods 31 (1996) 69-80

77

ovary preparation ran with apparent size of 6.5 kDa, as expected, and upon reduction ran as a smaller, broad band corresponding in size to the separate A and B chains, at - 3 kDa (data not shown). Mass spectrometry confirmed the prorelaxin was homogeneous and of expected molecular mass - 18 kDa (Fig. 6). This suggests exopeptidase trimming of the purified prohormone was minimal. Edman sequencing of the prorelaxin was not undertaken as it was confirmed to be N-terminally blocked. Porcine prorelaxin has a glutamine at its N-terminus which is cyclised to pyroglutamic acid in nature [42]. Amino acid analysis of tryptic fragments of the purified prorelaxin demonstrated the expected amino acid ratios in fragments including those at the N- and C-termini of the peptide (results not shown). The purified porcine prorelaxin was tested for activity in the isolated rat atria bioassay [43]. It was found to be as active as mature porcine relaxin both inotropically and chronotropically. Maximal activity was lower for prorelaxin in both tests but the dose required for half maximal response; pD, (- log iD of molar concentration) was not statistically different. For inotropic activity mature relaxin pD, was 9.10 f 0.07 and prorelaxin pD, was 9.01 + 0.06. For chronotropic activity mature relaxin pD, was 8.90 f 0.03 and prorelaxin pD, was 8.88 + 0.05.

4. Discussion A method has been devised to isolate milligram amounts of porcine prorelaxin. The protocol is based on that used by Walsh and Niall [39] for purification of mature mammalian relaxin. To improve yield and homogeneity of the target proteins modifications were made to this protocol as follows: The decapsulated corpora lutea of porcine ovaries were used rather than whole ovaries. Decapsulating served to remove the bulk of vasculature and connective tissue. This reduced the amount of extraneous material in the preparation and made homogenization easier, thus improving yield of protein. Removing vascular material eliminates many serum-borne proteases and reduces proteolysis of the target material. It is important to minimise the degradation of prorelaxin by enzymes that specifically process the precursor to mature relaxin and also by nonspecific enzymes. While the enzymes that process prorelaxin have not been elucidated, it is thought they may be members of the prohormone convertase family. Prohormone convertases are calcium dependent serine proteases [44]. For this reason EDTA and benzamidine were added to the extraction buffer, along with a cocktail of inhibitors for other proteases. All purification steps were carried out at or below 4°C to further retard proteolysis. Previous methods for relaxin purification using octadecylsilica columns involved applying the extract to the columns then eluting all proteins with 80% acetonitrile. In order to better fractionate the material, a stepwise gradient of increasing acetonitrile concentration was used in this protocol. Eluate from the 30 and 40% acetonitrile washes were enriched for relaxin and prorelaxin (Fig. 3). One of the most important modifications in this method is that material is taken straight from the octadecylsilica columns and purified by reversed phase HPLC. The

78

S.S. Layden. G.W. Tregear/J.

Biochem. Biophys. Methods 31 (1996) 69-80

method of Walsh and Niall [39] involves rotary evaporation for 10 to 15 min at 38°C to concentrate the extract before applying it to a gel filtration column. It is likely the target material would undergo degradation during treatment at 38”C, even with the addition of enzyme inhibitors. For this reason, the rotary evaporation step has been removed. Omission of the gel filtration step in this protocol allows volumes to be kept small and losses of material minimised. This method affords N 6 mg of prorelaxin and 4 mg of relaxin from one ovary of a late pregnant gilt. This compares with 1 to 2 mg of relaxin obtained per ovary as described by Walsh and Niall [39]. Our protocol affords improved yields of homogeneous, full-length prorelaxin than previously described. Purified prorelaxin is an essential requirement for experiments designed to study prorelaxin processing and the fate of the connecting C peptide.

Acknowledgements The authors thank Mr. Mick Petrovski for collection of ovary material and for carrying out amino acid analyses, Dr. Get? Talbo for performing the mass spectrometry, and Mr. Yan Yeow Tan for the rat atria bioassay. Mass spectrometry facilities were made available by the Department of Biochemistry, University of Melbourne. The Howard Florey Institute is supported by an Institute block grant from the National Health & Medical Research Council of Australia.

References [l] Sherwood, O.D., Relaxin, in The Physiology of Reproduction, Knobil, E. and Neill, J. @.dsl, 1994, Raven Press: New York. pp. 861-1009. [2] Downing, S.J. and Sherwood, O.D., The physiological role of relaxin in the pregnant rat III. The influence of relaxin on cervical extensibility. Endocrinology, 116 (1985) 1215-1220. [3] Hansell, D.J., Bryant-Greenwood, G.D. and Greenwood, F.C., Expression of the human relaxin HI gene in the decidua, trophoblast and prostate. J. Clin. Endocrin. Metab., 72 (1991) 899-904. [4] Winslow, J.W., Shih, A., Bourell, J.H., Weiss, G., Reed, B., Stults, J.T. and Goldsmith, L.T., Human seminal relaxin is a product of the same gene as human luteal relaxin. Endocrinology, 130 (1992) 2660-2668. [5] Osheroff, P.L. and Ho, W.-H., Expression of relaxin mRNA and relaxin receptors in postnatal and adult rat brains and hearts. J. Biol. Chem., 268 (1993) 15193-15199. [6] Mazoujian, G. and Bryant-Greenwood, G.D., Relaxin in breast tissue. Lancet, 335 (1990) 298-299. [7] Osheroff, P.L. and Phillips, H.S.. Autoradiographic localisation of relaxin binding sites in rat brain. hoc. Natl. Acad. Sci. USA, 88 (1991) 6413-6417. [8] Osheroff, P.L., Ling, V.T., Vandlen, R.L., Cronin, M.J. and Lofgren, J.A., Preparation of biologically active 32P-labelled human relaxin. J. Biol. Chem., 265 (1990) 9396-9401. 191 Parry, L.J., Poterski, R.S., Summerlee, A.J.S. and Jones, S.A., Mechanism of the haemotensive action of porcine relaxin in anaesthetized rats. J. Neurol., 2 (1990) 53-58. [lo] Kakouris, H.. Eddie, L.W. and Summers, R.J., Cardiac effects of relaxin in rats. Lancet, 339 (1992) 1076-1078. [ill St-Louis, J. and Massicotte. G.. Chronic decrease of blood pressure by rat relaxin in spontaneously hypertensive rats. Life Sci., 37 (1985) 1351-1357.

S.S. Luyden, G.W. Tregear/J.

[I21

Biochem. Biophys. Methods 31 (1996) 69-80

79

Summerlee, A.J.S., G’Byme, K.T., Paisley, A.C., Breeze, M.F. and Porter, D.G., Relaxin affects the central controlof oxytocin release. Nature, 309 (1984) 372-374. [I31Weisenger, R.S., Bums, P., Eddie, L.W. and Wintour, E.M., Relaxin alters the plasma osmolality arginine vasopressin relationship in the rat. J. Endocrinol., 137 (1993) 505-510. [141Zhao, S., Malmgren, C.H., Shanks,R.D. and Sherwood, O.D., Monoclonal antibodies specific for rat relaxin VIII. Passive immunization with monoclonal antibodies throughout the second half of pregnancy reduces water consumption in rats. J. Endocrinol., 136 (1995) 1892-1897. [I51Thornton, S.N. and Fitzsimons, J.T., The effects of centrally administered porcine relaxin on drinking behavoior in male and female rats. J. Neuroendocrinol., 7 (1995) 165-170. b51 Blundell, T.L. and Humbel, R.E., Hormone families: pancreatic hormones and homologous growth factors. Nature, 287 (1980) 781-787. [171Haley, J., Hudson, P., Scanlon, D., John, M., Cronk, M., Shine, J., Tregear, G. and Niall, H., Porcine relaxin: molecular cloning and cDNA structure. DNA, 1 (1982) 155-161. [I81 Gast, M.J., Studies on luteal generation and processing of the high molecular weight relaxin precursor. Ann. NY. Acad. Sci., 380 (1982) 111-121. G.D., Ultrastructural immunoperoxidase demonstra1191Kendall, J.Z., Plopper, C.G. and Bryant-Greenwood, tion of relaxin in corpora lutea from a pregnant sow. Biol. Rep., 18 (1978) 94-98. ml Schwabe, C. and Bullesbach, E.E., Relaxin: structures, functions, promises and nonevolution. FASEB J., 8 (1994) 1152-1160. El1 Hudson, P., John, M., Crawford, R., Haralambidis, J., Scanlon, D., Gorman, J., Tregear, G., Shine, J. and Niall, H., Relaxin gene expression in human ovaries and the predicted structure of a human preprorelaxin by analysis of cDNA clones. EMBO J., 3 (1984) 2333-2339. ml Tager, H.S., Emdin, S.O., Clark, J.L. and Steiner, D.F., Studies on the conversion of proinsulin to insulin. J. Biol. Chem., 248 (1973) 3476-3482. [23] Eigenbrot, C., Randal, M., Quan, C., Bumier, J., O’Connell, L., Rinderknecht, E. and Kossiakoff, A.A., X-Ray structure of human relaxin at 1.5A - comparison to insulin and implications for receptor binding determinants. J. Mol. Biol., 221 (1991) 15-22. [24] Bedarkar, S., Blundell, T., Gowan, L.K., McDonald, J.K. and Schwabe, C., On the three dimensional structure of relaxin, in Steinetz, B.G., Schwabe, C. and Weiss, G. &is.), Relaxin: Structure, Function and Evolution Conference Proceedings, 1982, New York Academy of Sciences, pp. 22-33. [25] Yang, S., Heyn, H., Zhang, Y.Z., Bullesbach, E.E. and Schwabe, C., The expression of human relaxin in yeast. Arch. Biochem. Biophys., 300 (1993) 734-737. [26] Vandlen, R., Winslow, J., Moffat, B. and Rinderknecht, E., Human relaxin: purification, characterization and production of recombinant relaxins for structure function studies, in MacLennan, A.H., Tregear, G.W. and Bryant-Greenwood, G.D. @Is.). Progress in Relaxin Research. Proceedings of the Second International Congress on the Hormone Relaxin, 1994, Global Publication Services, Singapore: Adelaide, South Australia. pp. 59-72. [27] Devi, L., Consensus sequence for processing of peptide precursors at monobasic sites. FEBS Lett., 280 (1991) 189-194. [28] Gluschankof, P., Gomez, S., Lepage, A., Creminon, C., Nyberg, F., Terenius, L. and Cohen, P., Role of peptide substrate structure in the selective processing of peptide prohormones at basic amino acid pairs by endoproteases. FEBS Len., 234 (1988) 149-152. [291Soloff, M.S., Shaw, A.R., Gentry, L.E., Marquardt, H. and Vasilenko, P., Demonstration of relaxin precursors in pregnant rat ovaries with antisera against bacterially expressed prorelaxin. Endocrinology, 130 (1992) 1844-1851. [301Sherwood, O.D., Key, R.H., Tarbell, M.K. and Downing, S.J., Dynamic changes of multiple forms of serum immunoactive relaxin during pregnancy in the rat. Endocrinology, 114 (1984) 806-813. G.D., Jeffrey, R., Ralph, M.M. and Seamark, R.F., Relaxin production by the porcine [311Bryant-Greenwood, ovarian Graafian follicle in vitro. Biol. Reprod., 23 (1980) 792-800. [321Vu, A.L., Green, C.B., Roby, K.F., Soares, M.J., Fei, D.T.W., Chen, A.B. and Kwok, C.M., Recombinant porcine prorelaxin produced in Chinese hamster ovary cells is biologically active. Life Sciences, 52 (1993) 1055-1061. of porcine relaxin. Arch. B&hem. [331Shetwood, C.D. and O’Byme, E.M., Purification and characterization Biophys., 160 (1974) 185-196.

80

S.S. Layden, G.W. Tregear/J.

Biochem. Biophys.

Methods 31 (1996) 69-80

[34] Doczi, J., Process for the Extraction and Purification of Relaxin. 1963, U.S. Patent No. 3.096.246 1351 Ftieden, E.H., Rawitch, A.B., Wu, L.-H. and Chen, S.-W., The isolation of two proline-containing relaxin species from a porcine relaxin concentrate. Proc. Sot. Exp. Biol. Med., 163 (1980) 521-527. [36] Frieden, E.H. and Rawitch, A.B., Isolation and Chemical Properties of Relaxin Subspecies, in BryantGreenwood, G.D., Niall, H.D. and Greenwood, F.C. (Eds.), Relaxin, 1981, Elsevier, New York, NY, pp. 2 l-26. [37] Frieden, E.H.. Pollock, H.G., Steinetz, B.G. and Rawitch, A.B., Structure of a porcine relaxin inactive on the mouse pubic ligament. Arch. B&hem. Biophys., 266 (1988) 334-342. [38] Bennett, H.P.J., Hudson, A.M., Kelly, L.. McMattin. C. and Purdon, G.E., A rapid method, using octadecasilyl-silica, for the extraction of certain peptides from tissues. B&hem. J., 175 (1978) 1139-I 141. [39] Walsh, J.R. and Niall, H.D., Use of an octadecylsilica purification method minimizes proteolysis during isolation of porcine and rat relaxins. Endocrinology, 107 (1980) 1258-1260. [40] Kwok. S.C.M., Charnley, W.A. and Bryant-Greenwood, G.D., High molecular weight forms of relaxin in pregnant sow ovaries. Bioch. Biophys. Res. Commun., 82 (1978) 997-1005. [41] Reddy, G.K., Gunwar. S., Green, C.B., Fei, D.T.W., Chen, A.B. and Kwok, S.C.M., Purification and charactetisation of recombinant porcine prorelaxin expressed in Escherichia coli. Arch. B&hem. Biophys., 294 (1992) 579-585. [42] Schwabe, C. and McDonald, K., Demonstration of a pyroglutamyl residue at the N terminus of the B-chain of porcine relaxin. Bioch. Biophys. Res. Commun., 74 (1977) 1501-1504. [43] Kakouris, H., Eddie, L.W. and Summers, R.J., Relaxin: more than just a hormone of pregnancy. Trends Pharmacol. Sci., 14 (1993) 4-6. [44] van de Ven, W.J.M. and Roebroek, A.J.M., Structure and function of eukaryotic proprotein processing enzymes of the subtilisin family of serine proteases. Crit. Rev. Oncogen., 4 (1993) 115-136.