- Email: [email protected]
Partial purification of relaxin from human seminal plasma
Partial purification of relaxin from human seminal plasma Gerson Weiss, M.D., Laura T. Goldsmith, Ph.D., Cy Schoenfeld, Ph.D., and Richard D'Eletto, B.S. New York, New York Human seminal plasma contains approximately 50 ng/ml of relaxin (specific activity = 1.3 ng/mg protein). During preliminary purification, semen plasma was delipidated, rehydrated, and loaded onto small octadecyl silica columns that were eluted with a TEAF/acetonitrile gradient system. Results were monitored by radioimmunoassay. The resultant partially purified human relaxin demonstrated biologic activity in the rat uterine segment bioassay. Nine liters of semen plasma was delipidated, rehydrated in TEAF, and subjected to high-performance liquid chromatography by a step gradient followed by a linear gradient. The active eluate was further purified by ion exchange chromatography. Pooled recovery fractions provided a total of 45.8 "",g of relaxin. An aliquot flash evaporated and desalted by gel filtration chromatography provided 1.85 "",g of relaxin in 25.2 mg protein, specific activity 73.4. This material is being used as immunogen in the production of antihuman relaxin antibodies by monoclonal technique. Our procedure represents the first and only successful partial purification of human relaxin to yield sufficient quantity and purity for antibody production. (AM J OasTET GVNECOL 1986;154:749-55.)
Key words: Relaxin, semen, HPLC
Relaxin was discovered in 1926 by Hisaw. ' It was shown to be a hormone that modifies the maternal reproductive tract during pregnancy. The major source of relaxin for experimental use is porcine ovaries, structures particularly rich in relaxin and readily available from abattoirs. Since porcine relaxin is devoid of tyrosine and histidine, amino acids necessary for iodination, development of a radioimunoassay (RIA) was delayed until a technique was developed to radiolabel the molecule. 2 This allowed several laboratories to establish RIAs for relaxin. The structure of porcine relaxin was simultaneously determined by two laboratories." 4 In 1976 our laboratory clearly showed that relaxin is a human hormone of pregnancy, produced by the corpus luteum. 5 The primary structure of human relaxin, as determined by recombinant DNA technology, significantly differs from that of porcine relaxin. 6 Human relaxin has never been purified from a biologic source. This is because its concentration is relatively low, even in the corpus luteum, the richest source in humans. Starting material for purification is difficult to obtain. The hormone is a small molecular weight peptide of
From the Department of Obstetrics and Gynecology, New York University School of Medicine and Fertility Research Laboratory. Supported in part by National Institutes of Health grant HD 12395 and a grant from the Mellon Foundation. Presented by invitation at the Fourth Annual Meeting of the American Gynecological and Obstetrical Society, Hot Springs, Virginia, September 4-7, 1985. Reprint requests: Gerson Weiss, M.D., Department of Obstetrics and Gynecology, UMDNJ-New Jersey Medical School, 100 Bergen St., Newark, NJ 07103.
approximately 6000 daltons. It is lipophilic and has a high isolectric point. These characteristics result in its property of adhering to laboratory ware and other proteins. For example, we have determined that partially purified human relaxin adheres to and will not elute from certain kinds of filters. Its adherence to other substances prevents its clean precipitation by ammonium sulfate after dilipidation. Rather, it precipitates in a very broad peak with some of the hormone precipitating in all concentrations of ammonium sulfate. Similar problems occur when attempts are made to precipitate relaxin by graded ethanol concentrations after delipidation. In 1980 immunoreactive relaxin was determined to be present in human semen plasma.' We confirmed this observation and demonstrated the seminal relaxin to be biologically active in the guinea pig pubic symphysis palpation assay.s Relaxin concentration in semen plasma averages 50 ng/ml compared with approximately 1 ng/ml in serum from pregnant women. Since semen plasma is plentiful compared with other human relaxin sources, we could use this as a starting material for purification and accept low yields caused by the molecule's properties. Our strategy is to isolate and purify human semen relaxin to a level of purity sufficient to produce monoclonal antibodies to human relaxin, monitoring purification with both RIA and bioassay. The antibody can then be used to further purify human relaxin by affinity chromatography. The purified relaxin as well as the antibodies could then be used as research tools. In this report we described the partial purification of human semen relaxin to be used as an immunogen.
749
750 Weiss et al.
April,1986
Am J Obstet Gynecol
Material and methods Semen plasma. Semen plasma for starting material to be used in relaxin purification was obtained from a fertility laboratory. Semen was collected by masturbation after 3 to 5 days of abstinence. The samples were allowed to stand at room temperature for approximately 3 hours for completion of semen analysis. Samples were then centrifuged at 600 x g for 10 minutes, the pellet was discarded, and the semen plasma was pooled and stored frozen at - 20° C. Assays. Relaxin was measured by RIA with the use of rabbit antiporcine relaxin antibody R6, developed by O'Byrne and Steinetz. 9 125 1 polytyrosylated porcine relaxin was used as ligand and purified porcine relaxin (CMa') as standard. All results are expressed as nanograms per milliliter of immunoreactive equivalents of porcine relaxin. The sensitivity of this assay is 50 to 100 pg per tube. The interassay coefficient of variation is 12.9% and intraassay coefficient of variation is 6.1 %. The average precision of the assay is 13.9%. All samples were measured in triplicate. Protein was determined by the method of Lowry et al. lO All glassware was silanized with Prosil-28 (SCM Chemicals, Gainesville, Florida) before use. The rat uterine horn segment bioassay for relaxin with in vivo estrogen and progesterone pretreatment has been described in detail. 11 Immature female LongEvans rats were treated subcutaneously with 1 f.Lg of estradiol benzoate in sesame oil for 3 days. Two milligrams of progesterone in sesame oil was injected subcutaneously on days 2 and 3. The animals were killed on day 4 and their uteri isolated. Uterine horn segments were mounted in a double-jacketed organ bath attached to an isometric smooth muscle transducer. Electrodes from a direct-current stimulator were positioned so that one was free in the bath and the other touched the uterine horn segment. The muscle was bathed in 5 ml of modified Locke solution. Spontaneous contraction patterns were observed and a baseline contraction pattern was established with electric stimulation of 5 V, 60 Hz, for 5 seconds every minute. After contraction patterns were established the bath was changed and 5 ml of modified Locke solution containing the substance to be tested was added. Baseline and stimulated contractions were then monitored. At the end of the experiment the uterine segment was washed in modified Locke solution alone three times and contractions were again observed to demonstrate that inhibition of contractions was not a nonspecific toxic effect. Delipidation. The first step in all purification protocols was delipidation. Semen plasma was extracted with 10 vol of cold acetone. The supernatant was decanted and discarded. The precipitate was then diluted with 10 original vol of hexane and the supernatant
discarded. The precipitate was then extracted with 10 vol of acetone and saved by suction filtration on Whatman No. 541 filter paper. The precipitate was air dried overnight at room temperature and dissolved in deionized water. The insoluble material was removed by centrifugation at 1000 x g for 10 minutes. The supernatant was lyophilized and stored at - 20° C. In the large-scale purification scheme, the acetone wash was acidified with O.lN HCl. All reagents used in purification were high-performance liquid chromatoagraphic (HPLC) grade. Octadecyl silica (ODS) purification-small scale. The powdered delipidated semen plasma from 100 ml of semen plasma was dissolved in 500 ml of 0.25N TEAF, pH 3.0. This was divided into 29 parts and loaded on disposable Sep-Pak ODS columns, 191 mg powderll7.2 ml/column. The columns had been previously activated with 10 ml of methanol and equilibrated with 10 ml of 0.25N TEAF, pH 3.0. The eluate was recirculated and reloaded for a total of 10 times to ensure complete adsorbance to the column. After loading, each Sep-Pak was washed with 10 ml of TEAF. This was followed by a wash with 10 ml of 20% acetonitrile (CH3CN)/80% TEAF. This was followed by a wash with 9.5 ml of 40% CH 3CN/60% TEAF. Most of the relaxin eluted in the last 9 ml of this wash. The corresponding fractions from ~ll Sep-Paks were combined and assayed for relaxin. HPLC was carried out on a Supelcosil semipreparatory (10 mm by 25 cm) ODS column containing 5 f.Lm of particles with 10 nm of pores (catalog No. 5-8358, Supelco Inc., Belafonte, Pennsylvania). The chromatograph used was a Beckman model 324-MP advanced gradient liquid chromatograph with two solvent delivery pumps and a CRT controller (Beckman Instruments, Inc., Fullerton, California). The column effluent was monitored at 254 nm with a Hitachi variable wavelength spectrophotometer, model 155-10, equipped with a high-pressure flow cell. The recording was made on a Kipp & Zonen dual pen strip chart recorder (Kipp & Zonen, Bohemia, New York) set on two absorbance units full scale, with a chart speed of 0.5 cm/min. Fractions were collected at I-minute intervals with a Gilson model 201 fraction collector (Gilson Medical Electronics, Inc., Middleton, Wisconsin) into borosilicate glass tubes. The column was equilibrated with 12% CH 3CN/ TEAF. The pooled Sep-Pak fractions were diluted with 2 vol of TEAF to restore CH 3CN to 12%, a concentration at which relaxin completely adsorbs to the column. Sample loading was accomplished via the aqueous solvent delivery pump. This eliminates the need for drying the Sep-Pak fraction, rehydrating, and clarifying the solution, steps that result in loss of relaxin. After loading, the pump and columns were flushed with 100 ml of 12% CH 3CN/TEAF followed by 100% TEAF.
Purification of human seminal relaxin
Volume 154 Number 4
751
Table I. Partial purification of relaxin from 100 ml of human semen plasma Procedure
Original semen plasma Acetone, hexane extracted semen plasma Sep-Pak eluate HPLC eluate
Total activity (ng)*
8860 5150
Protein (mg)t
4343 1525
1690 1450
36.5 22.5
Specific activity (nglmg)
Purification (fold)
2.0 3.4
100 58.1
(1.0) 1.7
46.3 64.4
19.1 16.4
23.2 32.2
*Total activity by RIA. tTotal protein determined by the method of Lowry et al. 10
Relaxin was eluted by means of a linear gradient of CH,CN/TEAF, increasing CH 3CN from 12% to 22%
in the first minute. In the next 21 minutes the CH 3 CN was increased to 50%. Flow rate was 4.73 mllmin. Fractions were collected in borosilicate glass test tubes (12 by 75 mm, RTU; Becton Dickinson, Parsippany, N.J.). Relaxin eluted at 77.5 ml. Relaxin was measured by RIA directly, aliquoting the mixture into 0.5 ml of RIA buffer (0.01 mollL phosphate-buffered saline solution with 1% ovalbumin, pH 7.0). Large-scale purification. Nine thousand ninety-four milliliters of pooled semen plasma containing 461.5 ILg of immunoactive relaxin and 343.1 gm of protein was extracted with acid acetone, hexane, and acid acetone. The precipitate was dried and rehydrated with 0.25N TEAF, pH 3.0. This was clarified by ultracentrification at 40,000 rpm (250,000 x g) for 1 hour. Preparatory octadecylsilyl silica high-pressure liquid chromatography. The Prep LC/System 500 (Waters Associates, Milford, Massachusetts) with ODS Prep-Pak 500 cartridge, 5.7 by 30 cm, was activated with 4 L of methanol and equilibrated with 4 L of TEAF. The delipidated semen plasma was loaded onto the Prep-Pak and recirculated for 8 hours at a flow rate of 250 mil min. Elution by a step gradient proceeded at a flow rate of 100 ml/min. Four liters ofO.25N TEAF, pH 3.0, was used as the initial wash. The second wash was 6 L of 20% CH 3 CN/80% TEAF. The next wash was 8 L of 40% CH 3 CN/60% TEAF, the first 5 L collected in 500 ml fractions and the remaining 3 L collected as one fraction. A final wash of 4.5 L of 60% CH,CN/40% TEAF was performed. Fraction 2 ofthe 40% CH 3 CN/60% TEAF step gradient, containing 53% of the load and 80% of the total recovery, was filtered on a 0.45 ILm pore size hydrophilic Millipore Durapore membrane (No. HVLP04700, Millipore/Continental Water Systems, Bedford, Massachusetts) to remove precipitated protein that formed during storage at - 20° C. The solution was diluted to 4 L in 0.25N TEAF, pH 3.0. The Prep-Pak was equilibrated with 4 L of 5% CH 3 CN!95% TEAF, loaded with the diluted fraction 2 and recirculated for 30 minutes. A linear gradient of 10% to 50% CH 3 CN/30 minutes was chosen. Fractions were collected, with fraction 1 con-
taining 970 ml, fractions 2 through 25 containing approximately 250 ml each, and fraction 26 containing 828 ml. Fractions 13 through 17 were pooled, with a total volume of 1278 ml. Ion exchange chromatography. After preliminary study, 1041 ml of the pooled fractions 13 through 17 were flash evaporated to 83 ml, ionic strength 0.2 mollL. This was diluted fivefold with water and the pH adjusted from 3.0 to 5.0 by dropwise addition of triethylamine to a volume of 415 ml, ionic concentration 0.093 mollL. During pH adjustment, a precipitate formed that was removed by ultrafiltration at 250,000 x g for 60 minutes. The supernatant had a pH of 5.0, ionic strength of 0.1 mollL, and a volume of 405 ml. There was no measurable loss of relaxin in the precipitate. Carboxymethyl Fractogel (Fractogel TSK CM-G50S Merck Chemical Division, Rathway, New Jersey) was equilibrated in 0.05 mollL ammonium acetate, pH 5.0, and packed in a Pharmacia K 50/30 glass column (5 by IS cm, bed volume 294.5 ml) fitted with Pharmacia A-50 flow adaptors at both ends (Pharmacia Fine Chemicals, Piscataway, New Jersey). Three hundred ninety-five milliliters was pumped onto the column with a Pharmacia P-4 peristaltic pump, flow rate 2 mllmin. Relaxin was not detected in the column effluent. Unadsorbed proteins were washed off with 0.05 mollL ammonium acetate, pH 5.0. Relaxin was eluted from the column with a 0 to 1 mollL NaCllinear gradient in 0.05 mollL ammonium acetate, pH 5.0. Threeminute fractions (6.1 ml) were collected. After 99 fractions were collected, baseline optical density at 280 nm was not reached. Elution was continued until 26 additional samples were collected and the baseline had reached its pregradient elution level. Relaxin was detectable in the last fraction indicating that some relaxin remained on the column. Elution was continued isocratically with I mollL N aCI in 0.1 mollL of ammonium acetate, pH 5.0, for another 75 fractions. Relaxin-containing fractions were combined into eight separate pools. Gel filtration chromatography. Size exclusion chromatography used Sephadex G-50 (fine) packed in a Pharmacia K-261100 glass column (2.6 by 100 cm; bed
752 Weiss et al.
April, 1986 Am J Obstet Gynecol
1.8
40 1" ELUTION
2nd ELUTION
3,d ELUTION
E 1>
"
m
...... o z
......
20
F m
800
c:
o
~1.4
......
!;:t
.5 600
0'
I.JJ
z
~ 1.0
(lJ
~ x
12
;:;:l
6
.,
X
I
80...!..
GJ400
cr
o(f)
cr
~ 0.6
0:
E
60 ~ ('Tl
3579135791 3 5 7 9 ELUTION VOLUME (ml)
200
o 0.2
r-
40
I I I
~--_---..r-'V/A~
20
d z :ti r
('Tl
1 2 3 4 5 6 7 8 9 10
Fig. 1. Triple-step gradient of CH,CN/TEAF used in the elution of human semen relaxin from a Sep-Pak.
FRACTION NUMBER
Fig. 3. Multiple-step gradient elution of human semen relaxin from a Prep-Pak with CH,CN/TEAF. The Prep-Pak load was solubilized delipidated powder from 9094 ml of human semen plasma.
A
8.0
E
7.0
~ 6.0
.5
z 5.0
B
50
b 4.0
g:
....
-
40 ~ ..... 30 0~
3.0
T 2.0
20
1.0
10
..... ~
r ,..,
8
FRACTION NUMBER
Fig. 2. Effect of relaxin on electrically stimulated rat uterine horn segments. In A, 2 ng of porcine relaxin (NIH-R-PI) is added at arrows 1-4. Arrow 5 indicates a wash. In B, 8.8 ng of partially purified human semen relaxin is added at arrows 1 and 2. Arrow 3 indicates a wash.
volume 530 ml) fitted on top with a Pharmacia A-26 flow adaptor. The column was equilibrated in 0.01 mollL ammonium acetate, pH 5.0, at a flow rate of 2 ml/min. All fractions were collected at 4-minute intervals in 13 by 100 mm borosilicate glass RTU test tubes. The column was pre saturated with 0.1 % bovine serum albumin. The load solution, 20.85 ml, flash evaporated from 468 ml taken from a preliminary recovery from ion exchange chromatography, contained 5.06 f.Lg of relaxin, with 221.8 mg of protein, specific activity 22.82 ng/mg.
Results Small-scale purification. The results of the smallscale purification of 100 ml of semen plasma can be seen in Table I. The results of the step-gradient elution of relaxin from a Sep-Pak can be seen in Fig. 1. No relaxin is eluted by the 20% CH 3 CN/80% TEAF wash. Most of the relaxin is eluted in the 40% CH 3 CN/60% TEAF wash.
Fig. 4. Rechromatography by a linear gradient of CH,CNI TEAF on the Prep-Pak of fraction 2 from the step gradient (Fig. 3).
The pooled mobile phase from the relaxin eluate from HPLC was evaporated under vacuum and lyophilized four times to completely eliminate all solvent. This material was rehydrated and applied to the rat uterine horn segment relaxin bioassay. As can be seen in Fig. 2, human seminal relaxin purified by these methods is active in this assay. The high sensitivity of this assay compared with other bioassays of relaxin allows its use in the monitoring of purification of human semen relaxin. Large-scale purification. After delipidation, the pooled semen plasma contained 437.9 f.Lg of relaxin. This was chromatographed by Prep-Pak HPLC by a step gradient. The results are seen in Fig. 3. Fraction 2 contained 232.3 f.Lg of relaxinlll.l gm protein, specific activity 20.93 ng/mg. The results of Prep-Pak HPLC with a linear gradient are seen in Fig. 4. The load of relaxin was 204.6 f.Lg. Fractions 13 through 17 contained 198.8 f.Lg of relaxin and 8.95 gm of protein, specific activity 22.21 ng/mg.
Purification of human seminal relaxin
Volume 154 Number 4
753
I §
!;! 0.3
1.200 ~
:il
800~
~ 0.2
ifl
«
~
0.1
400.[
c------'
I 10
20 FRACTION
NUMBER
Fig. 5. Chromatography of Prep-Pak purified human semen plasma by ion exchange chromatog raphy on carboxymethyl Fractogel TSK with the use of an NaCI gradient. The vertical arrow indicates where the second session of column elution started. Roman numerals signify pooled fractions.
The results of ion exchange chromatography of 1041 ml of pooled fractions 13 through 17 from HPLC are shown in Fig. 5. Recovery of fractions 70 through 124 contain 37.82 f.Lg of relaxin and 0.737 gm of protein, specific activity 51.32 ng/mg. The purest pool was VIII, containing 1.79 f.Lg of relaxin/5.95 mg of protein, specific activity 301 ng/mg. This is a 13.5-fold increase in purity over that of the column load. If pools III through VIII were arithmetically combined, the relaxin content of this pool would be 29.1 f.Lg of relaxinll38.74 mg of protein, specific activity 210.9 ng/mg. The estimated total recovery, if the entire quantity of starting material had been pumped through all of the steps, would be 10.67%. Gel filtration chromatography. The desalting procedure was run with the eluate of 468 ml from pooled recovery fractions 57 through 135 of a preliminary ion exchange chromatography run. These results can be seen in Fig. 6. Pool II contains 1.85 f.Lg of relaxin and 25.2 mg of protein, specific activity 73.4 nglmg, with an ionic strength of 0.01 moilL. Pool III contained 0.458 ng of relaxin I 1.04 f.Lg of protein, specific activity 440.4 ng/mg. However, this pool had an ionic strength of 0.43 moilL, indicating that it was not salt free. Pool II is being used for monoclonal antibody production.
Comment To our knowledge, this article reports the first successful partial purification of human relaxin from a biologic source in sufficient quantity and purity for antibody production. Our attempts at purification of human seminal relaxin have provided several useful techniques to decrease procedural losses. These include keeping the relaxin in acidic solution and silanization to decrease adherence to labware. Avoidance of drying and filtering partially purified relaxin, but rather maintaining it in a dissolved state, also helps decrease procedurallosses. For example, in an earlier purification protocol we filtered relaxin on a Millex GV filter, a polyvinylidene difluoride hydrophilic membrane contained in a disposable polystyrene holder. We first tested the filter with porcine relaxin and it allowed passage without loss. However, this filter unit retained 75%
§
Vo
VI
0
I
I
gj 0.3 "5
240
~
w
200
~ 0.2 m n::
160 ~
U
:
l>
120 ~
0
Ul
m
« 0.1
80 -;-
> ::::>
40
I 5
""
10 15 20 25 30 35 40 45 50 55 60 65 FRACTION NUMBER
Fig. 6. Desalting of CM Fractogel TSK-purified human semen plasma on a Sephadex G-50 (fine) column. Roman numerals indicate pooled fractions. Vo, void volume; Vt, total volume.
of the human relaxin. We were unable to elute this relaxin by any technique. Our procedures will provide experience in devising techniques for the purification of relaxin-like peptides from other sources, some with much lower concentration or availability. It is likely that there will be considerable microheterogeneity in our relaxin preparation. Semen plasma contains many proteolytic enzymes such as amino and carboxypeptidases, which may partially degrade the native hormone but may not affect its biologic activity. Porcine relaxin can accept signifigant loss of the C-terminus of the B chain without loss of biologic activity.12 Hudson et al. 6 have shown that there are two genes for human relaxin as identified from a human genomic library. These two genes encode markedly different relaxin peptides. Little is known of the expression of these genes except that one is preferentially expressed by human ovaries during pregnancy. Characterization of seminal relaxin is necessary to determine which gene is expressed in the male. It is tempting to postulate that there is divergent gene expression in the male and female. This possibility can be evaluated by purification ofthe seminal hormone and comparison of its structure to that of ovarian relaxin. The studies may pTovide models for the evaluation of the hormonal regulation of gene expression. The partial purification of relaxin described herein
754
Weiss et al.
will provide material for the production of antihuman relaxin antibodies, which, in turn, can be used for final purification of seminal relaxin as well as establishment of human relaxin immunoassays. Antirelaxin antibodies will also be useful tools in determination of the biologic roles of relaxin. Because of significant primary structural differences, administration of nonhuman (porcine) relaxin to humans has the potential danger of being immunogenic. Human relaxin in sufficient quantities for clinical use will be available only as material generated by recombinant DNA methodology. However, use of the recombinant DNA-derived relaxin will require verification of the structure of the native human hormone. REFERENCES 1. Hisaw FL. Experimental relaxation of the pubic ligament of the guinea pig. Proc Soc Exp Bioi Med 1926;23:661. 2. Sherwood OD, Rosentreter KR, Birkhimer ML. Development of a radioimmunoassay for porcine relaxin using 125I-labelled polytyrosyl-relaxin. Endocrinology 1975; 96:1106. 3. Schwabe C, McDonaldjK, Steinetz BG. Primary structure of the B-chain of porcine relaxin. Biochem Biophys Res Commun 1977;75:503. 4. james R, Niall H, Kwok S, Bryant-Greenwood G. Primary structure of porcine relaxin: homology with insulin and related growth factors. Nature 1977;267:544. 5. Weiss G, O'Byrne EM, Steinetz BG. Relaxin: a product of the human corpus luteum of pregnancy. Science 1976; 194:948. 6. Hudson P,john M, Crawford R, Haralambidisj, Scanlon D, Gorman j, Tregear G, Shine j, Niall H. Relaxin gene expression in human ovaries and the predicted structure of a human pre pro relaxin by analysis of cDNA clones. EMBO j 1984;3:2333. 7. Loumaye E, DeCooman S, Thomas K. Immunoreactive relaxin-like substance in human seminal plasma. j Clin Endocrinol Metab 1980;50: 1142. 8. Essig M, Schoenfeld C, D'Eletto R, Amelar R, Dubin L, Steinetz BG, O'Byrne EM, Weiss G. Relaxin in human seminal plasma. Ann NY Acad Sci 1982;380:224. 9. O'Byrne EM, Steinetz BG. RIA of relaxin in serum of various species using antiserum to porcine relaxin. Proc Soc Exp Bioi Med 1976;152:272. 10. Lowry OH, Rosebrough Nj, Farr AL. Protein measurement with folin phenol reagent. j Bioi Chern 1951; 193:265. II. Sarosi P, Schmidt CL, Essig M, Steinetz BG, Weiss G. The effect of relaxin and progesterone on rat uterine contractions. AM j OBSTET GYNECOL 1983; 145:402. 12. john Mj, Walsh jR, james Rj, Kwok S, Byrant-Greenwood GD, Bradshaw RA, Niall HD. Heterogenecity of porcine relaxin. In: Byrant-Greenwood GD, Niall HD, Greenwood FC, eds. Relaxin. Amsterdam: Elsevier Biomedical Press, 1981: 17.
Editors' note: This manuscript was revised after these discussions were presented.
Discussion DR. KAMRAN S. MOGHISSI, Detroit, Michigan. Relaxin was discovered almost 60 years ago. It is a polypeptide hormone with a similar structural identity to that of
April, 1986 Am J Obstet Gynecol
insulin and nerve growth factor. Its main source appears to be the corpus luteum of pregnancy, but it is also produced by the male prostate gland and is found in many target tissues such as cervix, myometrium, decidua, and breast connective tissue. Its principal mechanism of action appears to be the facilitation of remodeling of connective tissue in target organs to allow the necessary changes in organ structure during pregnancy and parturition. Additionally, in some mammals, relaxin inhibits myometrial contractility until near the end of pregnancy. Possible roles for relaxin in the human include enhancement of sperm motility and penetration into cervical mucus, ovum penetration, implantation of blastocyst, uterine stromal remodeling during pregnancy, the inhibition of premature labor, cervical ripening, and collagen biosynthesis in disorders of collagen metabolism such as scleroderma and arthritis. 1.2 Porcine relaxin has been well studied and purified. This preparation has been used in most animal experiments and some clinical studies. The amino acid sequence of a human relaxin derived from the recent identification of a genomic clone for relaxin is now known. s Approximately 50% homology between porcine relaxin and this human relaxin exists. Porcine and human relaxins share the same tertiary structure and some of the same amino acid sequences at biologically active sites. Human relaxin has not, as yet, been purified and the development of a specific assay must await the availability of purified samples of human relaxin for antibody development. The stated goals of Dr. Weiss and his associates are as follows: (I) purification ofrelaxin to a sufficient degree so that workable amounts of relaxin-specific antibodies can be obtained, (2) use of the antirelaxin antibodies to prepare affinity immunosorbants that can be used to complete the relaxin purification, (3) sequence of the purified relaxin to determine molecular structure, and (4) use of recombinant DNA methods to batch up adequate amounts of relaxin for clinical studies. The work presented here deals with only the first item, namely partial purification. The purification of relaxin from seminal plasma is a difficult problem and Dr. Weiss and his colleagues should be complimented for their persistance and their ability to collect "9 L" of semen as their starting material. However, there are severe limitations in the experimental approach used by the authors. First, the levels of immunoreactive relaxin in human seminal plasma are exceedingly low, approximately 9 fLg per 9 L (representing 4300 fLg of proteins). Even assuming purification to homogeneity with 100% yield, which would require a 500,000-fold enrichment, the amount of relaxin recovered would be barely sufficient for structural analysis. Dr. Weiss reports a 32-fold purification with only 16% yield after three purification steps. This is certainly not an encouraging result.
Key words: Relaxin, semen, HPLC
Relaxin was discovered in 1926 by Hisaw. ' It was shown to be a hormone that modifies the maternal reproductive tract during pregnancy. The major source of relaxin for experimental use is porcine ovaries, structures particularly rich in relaxin and readily available from abattoirs. Since porcine relaxin is devoid of tyrosine and histidine, amino acids necessary for iodination, development of a radioimunoassay (RIA) was delayed until a technique was developed to radiolabel the molecule. 2 This allowed several laboratories to establish RIAs for relaxin. The structure of porcine relaxin was simultaneously determined by two laboratories." 4 In 1976 our laboratory clearly showed that relaxin is a human hormone of pregnancy, produced by the corpus luteum. 5 The primary structure of human relaxin, as determined by recombinant DNA technology, significantly differs from that of porcine relaxin. 6 Human relaxin has never been purified from a biologic source. This is because its concentration is relatively low, even in the corpus luteum, the richest source in humans. Starting material for purification is difficult to obtain. The hormone is a small molecular weight peptide of
From the Department of Obstetrics and Gynecology, New York University School of Medicine and Fertility Research Laboratory. Supported in part by National Institutes of Health grant HD 12395 and a grant from the Mellon Foundation. Presented by invitation at the Fourth Annual Meeting of the American Gynecological and Obstetrical Society, Hot Springs, Virginia, September 4-7, 1985. Reprint requests: Gerson Weiss, M.D., Department of Obstetrics and Gynecology, UMDNJ-New Jersey Medical School, 100 Bergen St., Newark, NJ 07103.
approximately 6000 daltons. It is lipophilic and has a high isolectric point. These characteristics result in its property of adhering to laboratory ware and other proteins. For example, we have determined that partially purified human relaxin adheres to and will not elute from certain kinds of filters. Its adherence to other substances prevents its clean precipitation by ammonium sulfate after dilipidation. Rather, it precipitates in a very broad peak with some of the hormone precipitating in all concentrations of ammonium sulfate. Similar problems occur when attempts are made to precipitate relaxin by graded ethanol concentrations after delipidation. In 1980 immunoreactive relaxin was determined to be present in human semen plasma.' We confirmed this observation and demonstrated the seminal relaxin to be biologically active in the guinea pig pubic symphysis palpation assay.s Relaxin concentration in semen plasma averages 50 ng/ml compared with approximately 1 ng/ml in serum from pregnant women. Since semen plasma is plentiful compared with other human relaxin sources, we could use this as a starting material for purification and accept low yields caused by the molecule's properties. Our strategy is to isolate and purify human semen relaxin to a level of purity sufficient to produce monoclonal antibodies to human relaxin, monitoring purification with both RIA and bioassay. The antibody can then be used to further purify human relaxin by affinity chromatography. The purified relaxin as well as the antibodies could then be used as research tools. In this report we described the partial purification of human semen relaxin to be used as an immunogen.
749
750 Weiss et al.
April,1986
Am J Obstet Gynecol
Material and methods Semen plasma. Semen plasma for starting material to be used in relaxin purification was obtained from a fertility laboratory. Semen was collected by masturbation after 3 to 5 days of abstinence. The samples were allowed to stand at room temperature for approximately 3 hours for completion of semen analysis. Samples were then centrifuged at 600 x g for 10 minutes, the pellet was discarded, and the semen plasma was pooled and stored frozen at - 20° C. Assays. Relaxin was measured by RIA with the use of rabbit antiporcine relaxin antibody R6, developed by O'Byrne and Steinetz. 9 125 1 polytyrosylated porcine relaxin was used as ligand and purified porcine relaxin (CMa') as standard. All results are expressed as nanograms per milliliter of immunoreactive equivalents of porcine relaxin. The sensitivity of this assay is 50 to 100 pg per tube. The interassay coefficient of variation is 12.9% and intraassay coefficient of variation is 6.1 %. The average precision of the assay is 13.9%. All samples were measured in triplicate. Protein was determined by the method of Lowry et al. lO All glassware was silanized with Prosil-28 (SCM Chemicals, Gainesville, Florida) before use. The rat uterine horn segment bioassay for relaxin with in vivo estrogen and progesterone pretreatment has been described in detail. 11 Immature female LongEvans rats were treated subcutaneously with 1 f.Lg of estradiol benzoate in sesame oil for 3 days. Two milligrams of progesterone in sesame oil was injected subcutaneously on days 2 and 3. The animals were killed on day 4 and their uteri isolated. Uterine horn segments were mounted in a double-jacketed organ bath attached to an isometric smooth muscle transducer. Electrodes from a direct-current stimulator were positioned so that one was free in the bath and the other touched the uterine horn segment. The muscle was bathed in 5 ml of modified Locke solution. Spontaneous contraction patterns were observed and a baseline contraction pattern was established with electric stimulation of 5 V, 60 Hz, for 5 seconds every minute. After contraction patterns were established the bath was changed and 5 ml of modified Locke solution containing the substance to be tested was added. Baseline and stimulated contractions were then monitored. At the end of the experiment the uterine segment was washed in modified Locke solution alone three times and contractions were again observed to demonstrate that inhibition of contractions was not a nonspecific toxic effect. Delipidation. The first step in all purification protocols was delipidation. Semen plasma was extracted with 10 vol of cold acetone. The supernatant was decanted and discarded. The precipitate was then diluted with 10 original vol of hexane and the supernatant
discarded. The precipitate was then extracted with 10 vol of acetone and saved by suction filtration on Whatman No. 541 filter paper. The precipitate was air dried overnight at room temperature and dissolved in deionized water. The insoluble material was removed by centrifugation at 1000 x g for 10 minutes. The supernatant was lyophilized and stored at - 20° C. In the large-scale purification scheme, the acetone wash was acidified with O.lN HCl. All reagents used in purification were high-performance liquid chromatoagraphic (HPLC) grade. Octadecyl silica (ODS) purification-small scale. The powdered delipidated semen plasma from 100 ml of semen plasma was dissolved in 500 ml of 0.25N TEAF, pH 3.0. This was divided into 29 parts and loaded on disposable Sep-Pak ODS columns, 191 mg powderll7.2 ml/column. The columns had been previously activated with 10 ml of methanol and equilibrated with 10 ml of 0.25N TEAF, pH 3.0. The eluate was recirculated and reloaded for a total of 10 times to ensure complete adsorbance to the column. After loading, each Sep-Pak was washed with 10 ml of TEAF. This was followed by a wash with 10 ml of 20% acetonitrile (CH3CN)/80% TEAF. This was followed by a wash with 9.5 ml of 40% CH 3CN/60% TEAF. Most of the relaxin eluted in the last 9 ml of this wash. The corresponding fractions from ~ll Sep-Paks were combined and assayed for relaxin. HPLC was carried out on a Supelcosil semipreparatory (10 mm by 25 cm) ODS column containing 5 f.Lm of particles with 10 nm of pores (catalog No. 5-8358, Supelco Inc., Belafonte, Pennsylvania). The chromatograph used was a Beckman model 324-MP advanced gradient liquid chromatograph with two solvent delivery pumps and a CRT controller (Beckman Instruments, Inc., Fullerton, California). The column effluent was monitored at 254 nm with a Hitachi variable wavelength spectrophotometer, model 155-10, equipped with a high-pressure flow cell. The recording was made on a Kipp & Zonen dual pen strip chart recorder (Kipp & Zonen, Bohemia, New York) set on two absorbance units full scale, with a chart speed of 0.5 cm/min. Fractions were collected at I-minute intervals with a Gilson model 201 fraction collector (Gilson Medical Electronics, Inc., Middleton, Wisconsin) into borosilicate glass tubes. The column was equilibrated with 12% CH 3CN/ TEAF. The pooled Sep-Pak fractions were diluted with 2 vol of TEAF to restore CH 3CN to 12%, a concentration at which relaxin completely adsorbs to the column. Sample loading was accomplished via the aqueous solvent delivery pump. This eliminates the need for drying the Sep-Pak fraction, rehydrating, and clarifying the solution, steps that result in loss of relaxin. After loading, the pump and columns were flushed with 100 ml of 12% CH 3CN/TEAF followed by 100% TEAF.
Purification of human seminal relaxin
Volume 154 Number 4
751
Table I. Partial purification of relaxin from 100 ml of human semen plasma Procedure
Original semen plasma Acetone, hexane extracted semen plasma Sep-Pak eluate HPLC eluate
Total activity (ng)*
8860 5150
Protein (mg)t
4343 1525
1690 1450
36.5 22.5
Specific activity (nglmg)
Purification (fold)
2.0 3.4
100 58.1
(1.0) 1.7
46.3 64.4
19.1 16.4
23.2 32.2
*Total activity by RIA. tTotal protein determined by the method of Lowry et al. 10
Relaxin was eluted by means of a linear gradient of CH,CN/TEAF, increasing CH 3CN from 12% to 22%
in the first minute. In the next 21 minutes the CH 3 CN was increased to 50%. Flow rate was 4.73 mllmin. Fractions were collected in borosilicate glass test tubes (12 by 75 mm, RTU; Becton Dickinson, Parsippany, N.J.). Relaxin eluted at 77.5 ml. Relaxin was measured by RIA directly, aliquoting the mixture into 0.5 ml of RIA buffer (0.01 mollL phosphate-buffered saline solution with 1% ovalbumin, pH 7.0). Large-scale purification. Nine thousand ninety-four milliliters of pooled semen plasma containing 461.5 ILg of immunoactive relaxin and 343.1 gm of protein was extracted with acid acetone, hexane, and acid acetone. The precipitate was dried and rehydrated with 0.25N TEAF, pH 3.0. This was clarified by ultracentrification at 40,000 rpm (250,000 x g) for 1 hour. Preparatory octadecylsilyl silica high-pressure liquid chromatography. The Prep LC/System 500 (Waters Associates, Milford, Massachusetts) with ODS Prep-Pak 500 cartridge, 5.7 by 30 cm, was activated with 4 L of methanol and equilibrated with 4 L of TEAF. The delipidated semen plasma was loaded onto the Prep-Pak and recirculated for 8 hours at a flow rate of 250 mil min. Elution by a step gradient proceeded at a flow rate of 100 ml/min. Four liters ofO.25N TEAF, pH 3.0, was used as the initial wash. The second wash was 6 L of 20% CH 3 CN/80% TEAF. The next wash was 8 L of 40% CH 3 CN/60% TEAF, the first 5 L collected in 500 ml fractions and the remaining 3 L collected as one fraction. A final wash of 4.5 L of 60% CH,CN/40% TEAF was performed. Fraction 2 ofthe 40% CH 3 CN/60% TEAF step gradient, containing 53% of the load and 80% of the total recovery, was filtered on a 0.45 ILm pore size hydrophilic Millipore Durapore membrane (No. HVLP04700, Millipore/Continental Water Systems, Bedford, Massachusetts) to remove precipitated protein that formed during storage at - 20° C. The solution was diluted to 4 L in 0.25N TEAF, pH 3.0. The Prep-Pak was equilibrated with 4 L of 5% CH 3 CN!95% TEAF, loaded with the diluted fraction 2 and recirculated for 30 minutes. A linear gradient of 10% to 50% CH 3 CN/30 minutes was chosen. Fractions were collected, with fraction 1 con-
taining 970 ml, fractions 2 through 25 containing approximately 250 ml each, and fraction 26 containing 828 ml. Fractions 13 through 17 were pooled, with a total volume of 1278 ml. Ion exchange chromatography. After preliminary study, 1041 ml of the pooled fractions 13 through 17 were flash evaporated to 83 ml, ionic strength 0.2 mollL. This was diluted fivefold with water and the pH adjusted from 3.0 to 5.0 by dropwise addition of triethylamine to a volume of 415 ml, ionic concentration 0.093 mollL. During pH adjustment, a precipitate formed that was removed by ultrafiltration at 250,000 x g for 60 minutes. The supernatant had a pH of 5.0, ionic strength of 0.1 mollL, and a volume of 405 ml. There was no measurable loss of relaxin in the precipitate. Carboxymethyl Fractogel (Fractogel TSK CM-G50S Merck Chemical Division, Rathway, New Jersey) was equilibrated in 0.05 mollL ammonium acetate, pH 5.0, and packed in a Pharmacia K 50/30 glass column (5 by IS cm, bed volume 294.5 ml) fitted with Pharmacia A-50 flow adaptors at both ends (Pharmacia Fine Chemicals, Piscataway, New Jersey). Three hundred ninety-five milliliters was pumped onto the column with a Pharmacia P-4 peristaltic pump, flow rate 2 mllmin. Relaxin was not detected in the column effluent. Unadsorbed proteins were washed off with 0.05 mollL ammonium acetate, pH 5.0. Relaxin was eluted from the column with a 0 to 1 mollL NaCllinear gradient in 0.05 mollL ammonium acetate, pH 5.0. Threeminute fractions (6.1 ml) were collected. After 99 fractions were collected, baseline optical density at 280 nm was not reached. Elution was continued until 26 additional samples were collected and the baseline had reached its pregradient elution level. Relaxin was detectable in the last fraction indicating that some relaxin remained on the column. Elution was continued isocratically with I mollL N aCI in 0.1 mollL of ammonium acetate, pH 5.0, for another 75 fractions. Relaxin-containing fractions were combined into eight separate pools. Gel filtration chromatography. Size exclusion chromatography used Sephadex G-50 (fine) packed in a Pharmacia K-261100 glass column (2.6 by 100 cm; bed
752 Weiss et al.
April, 1986 Am J Obstet Gynecol
1.8
40 1" ELUTION
2nd ELUTION
3,d ELUTION
E 1>
"
m
...... o z
......
20
F m
800
c:
o
~1.4
......
!;:t
.5 600
0'
I.JJ
z
~ 1.0
(lJ
~ x
12
;:;:l
6
.,
X
I
80...!..
GJ400
cr
o(f)
cr
~ 0.6
0:
E
60 ~ ('Tl
3579135791 3 5 7 9 ELUTION VOLUME (ml)
200
o 0.2
r-
40
I I I
~--_---..r-'V/A~
20
d z :ti r
('Tl
1 2 3 4 5 6 7 8 9 10
Fig. 1. Triple-step gradient of CH,CN/TEAF used in the elution of human semen relaxin from a Sep-Pak.
FRACTION NUMBER
Fig. 3. Multiple-step gradient elution of human semen relaxin from a Prep-Pak with CH,CN/TEAF. The Prep-Pak load was solubilized delipidated powder from 9094 ml of human semen plasma.
A
8.0
E
7.0
~ 6.0
.5
z 5.0
B
50
b 4.0
g:
....
-
40 ~ ..... 30 0~
3.0
T 2.0
20
1.0
10
..... ~
r ,..,
8
FRACTION NUMBER
Fig. 2. Effect of relaxin on electrically stimulated rat uterine horn segments. In A, 2 ng of porcine relaxin (NIH-R-PI) is added at arrows 1-4. Arrow 5 indicates a wash. In B, 8.8 ng of partially purified human semen relaxin is added at arrows 1 and 2. Arrow 3 indicates a wash.
volume 530 ml) fitted on top with a Pharmacia A-26 flow adaptor. The column was equilibrated in 0.01 mollL ammonium acetate, pH 5.0, at a flow rate of 2 ml/min. All fractions were collected at 4-minute intervals in 13 by 100 mm borosilicate glass RTU test tubes. The column was pre saturated with 0.1 % bovine serum albumin. The load solution, 20.85 ml, flash evaporated from 468 ml taken from a preliminary recovery from ion exchange chromatography, contained 5.06 f.Lg of relaxin, with 221.8 mg of protein, specific activity 22.82 ng/mg.
Results Small-scale purification. The results of the smallscale purification of 100 ml of semen plasma can be seen in Table I. The results of the step-gradient elution of relaxin from a Sep-Pak can be seen in Fig. 1. No relaxin is eluted by the 20% CH 3 CN/80% TEAF wash. Most of the relaxin is eluted in the 40% CH 3 CN/60% TEAF wash.
Fig. 4. Rechromatography by a linear gradient of CH,CNI TEAF on the Prep-Pak of fraction 2 from the step gradient (Fig. 3).
The pooled mobile phase from the relaxin eluate from HPLC was evaporated under vacuum and lyophilized four times to completely eliminate all solvent. This material was rehydrated and applied to the rat uterine horn segment relaxin bioassay. As can be seen in Fig. 2, human seminal relaxin purified by these methods is active in this assay. The high sensitivity of this assay compared with other bioassays of relaxin allows its use in the monitoring of purification of human semen relaxin. Large-scale purification. After delipidation, the pooled semen plasma contained 437.9 f.Lg of relaxin. This was chromatographed by Prep-Pak HPLC by a step gradient. The results are seen in Fig. 3. Fraction 2 contained 232.3 f.Lg of relaxinlll.l gm protein, specific activity 20.93 ng/mg. The results of Prep-Pak HPLC with a linear gradient are seen in Fig. 4. The load of relaxin was 204.6 f.Lg. Fractions 13 through 17 contained 198.8 f.Lg of relaxin and 8.95 gm of protein, specific activity 22.21 ng/mg.
Purification of human seminal relaxin
Volume 154 Number 4
753
I §
!;! 0.3
1.200 ~
:il
800~
~ 0.2
ifl
«
~
0.1
400.[
c------'
I 10
20 FRACTION
NUMBER
Fig. 5. Chromatography of Prep-Pak purified human semen plasma by ion exchange chromatog raphy on carboxymethyl Fractogel TSK with the use of an NaCI gradient. The vertical arrow indicates where the second session of column elution started. Roman numerals signify pooled fractions.
The results of ion exchange chromatography of 1041 ml of pooled fractions 13 through 17 from HPLC are shown in Fig. 5. Recovery of fractions 70 through 124 contain 37.82 f.Lg of relaxin and 0.737 gm of protein, specific activity 51.32 ng/mg. The purest pool was VIII, containing 1.79 f.Lg of relaxin/5.95 mg of protein, specific activity 301 ng/mg. This is a 13.5-fold increase in purity over that of the column load. If pools III through VIII were arithmetically combined, the relaxin content of this pool would be 29.1 f.Lg of relaxinll38.74 mg of protein, specific activity 210.9 ng/mg. The estimated total recovery, if the entire quantity of starting material had been pumped through all of the steps, would be 10.67%. Gel filtration chromatography. The desalting procedure was run with the eluate of 468 ml from pooled recovery fractions 57 through 135 of a preliminary ion exchange chromatography run. These results can be seen in Fig. 6. Pool II contains 1.85 f.Lg of relaxin and 25.2 mg of protein, specific activity 73.4 nglmg, with an ionic strength of 0.01 moilL. Pool III contained 0.458 ng of relaxin I 1.04 f.Lg of protein, specific activity 440.4 ng/mg. However, this pool had an ionic strength of 0.43 moilL, indicating that it was not salt free. Pool II is being used for monoclonal antibody production.
Comment To our knowledge, this article reports the first successful partial purification of human relaxin from a biologic source in sufficient quantity and purity for antibody production. Our attempts at purification of human seminal relaxin have provided several useful techniques to decrease procedural losses. These include keeping the relaxin in acidic solution and silanization to decrease adherence to labware. Avoidance of drying and filtering partially purified relaxin, but rather maintaining it in a dissolved state, also helps decrease procedurallosses. For example, in an earlier purification protocol we filtered relaxin on a Millex GV filter, a polyvinylidene difluoride hydrophilic membrane contained in a disposable polystyrene holder. We first tested the filter with porcine relaxin and it allowed passage without loss. However, this filter unit retained 75%
§
Vo
VI
0
I
I
gj 0.3 "5
240
~
w
200
~ 0.2 m n::
160 ~
U
:
l>
120 ~
0
Ul
m
« 0.1
80 -;-
> ::::>
40
I 5
""
10 15 20 25 30 35 40 45 50 55 60 65 FRACTION NUMBER
Fig. 6. Desalting of CM Fractogel TSK-purified human semen plasma on a Sephadex G-50 (fine) column. Roman numerals indicate pooled fractions. Vo, void volume; Vt, total volume.
of the human relaxin. We were unable to elute this relaxin by any technique. Our procedures will provide experience in devising techniques for the purification of relaxin-like peptides from other sources, some with much lower concentration or availability. It is likely that there will be considerable microheterogeneity in our relaxin preparation. Semen plasma contains many proteolytic enzymes such as amino and carboxypeptidases, which may partially degrade the native hormone but may not affect its biologic activity. Porcine relaxin can accept signifigant loss of the C-terminus of the B chain without loss of biologic activity.12 Hudson et al. 6 have shown that there are two genes for human relaxin as identified from a human genomic library. These two genes encode markedly different relaxin peptides. Little is known of the expression of these genes except that one is preferentially expressed by human ovaries during pregnancy. Characterization of seminal relaxin is necessary to determine which gene is expressed in the male. It is tempting to postulate that there is divergent gene expression in the male and female. This possibility can be evaluated by purification ofthe seminal hormone and comparison of its structure to that of ovarian relaxin. The studies may pTovide models for the evaluation of the hormonal regulation of gene expression. The partial purification of relaxin described herein
754
Weiss et al.
will provide material for the production of antihuman relaxin antibodies, which, in turn, can be used for final purification of seminal relaxin as well as establishment of human relaxin immunoassays. Antirelaxin antibodies will also be useful tools in determination of the biologic roles of relaxin. Because of significant primary structural differences, administration of nonhuman (porcine) relaxin to humans has the potential danger of being immunogenic. Human relaxin in sufficient quantities for clinical use will be available only as material generated by recombinant DNA methodology. However, use of the recombinant DNA-derived relaxin will require verification of the structure of the native human hormone. REFERENCES 1. Hisaw FL. Experimental relaxation of the pubic ligament of the guinea pig. Proc Soc Exp Bioi Med 1926;23:661. 2. Sherwood OD, Rosentreter KR, Birkhimer ML. Development of a radioimmunoassay for porcine relaxin using 125I-labelled polytyrosyl-relaxin. Endocrinology 1975; 96:1106. 3. Schwabe C, McDonaldjK, Steinetz BG. Primary structure of the B-chain of porcine relaxin. Biochem Biophys Res Commun 1977;75:503. 4. james R, Niall H, Kwok S, Bryant-Greenwood G. Primary structure of porcine relaxin: homology with insulin and related growth factors. Nature 1977;267:544. 5. Weiss G, O'Byrne EM, Steinetz BG. Relaxin: a product of the human corpus luteum of pregnancy. Science 1976; 194:948. 6. Hudson P,john M, Crawford R, Haralambidisj, Scanlon D, Gorman j, Tregear G, Shine j, Niall H. Relaxin gene expression in human ovaries and the predicted structure of a human pre pro relaxin by analysis of cDNA clones. EMBO j 1984;3:2333. 7. Loumaye E, DeCooman S, Thomas K. Immunoreactive relaxin-like substance in human seminal plasma. j Clin Endocrinol Metab 1980;50: 1142. 8. Essig M, Schoenfeld C, D'Eletto R, Amelar R, Dubin L, Steinetz BG, O'Byrne EM, Weiss G. Relaxin in human seminal plasma. Ann NY Acad Sci 1982;380:224. 9. O'Byrne EM, Steinetz BG. RIA of relaxin in serum of various species using antiserum to porcine relaxin. Proc Soc Exp Bioi Med 1976;152:272. 10. Lowry OH, Rosebrough Nj, Farr AL. Protein measurement with folin phenol reagent. j Bioi Chern 1951; 193:265. II. Sarosi P, Schmidt CL, Essig M, Steinetz BG, Weiss G. The effect of relaxin and progesterone on rat uterine contractions. AM j OBSTET GYNECOL 1983; 145:402. 12. john Mj, Walsh jR, james Rj, Kwok S, Byrant-Greenwood GD, Bradshaw RA, Niall HD. Heterogenecity of porcine relaxin. In: Byrant-Greenwood GD, Niall HD, Greenwood FC, eds. Relaxin. Amsterdam: Elsevier Biomedical Press, 1981: 17.
Editors' note: This manuscript was revised after these discussions were presented.
Discussion DR. KAMRAN S. MOGHISSI, Detroit, Michigan. Relaxin was discovered almost 60 years ago. It is a polypeptide hormone with a similar structural identity to that of
April, 1986 Am J Obstet Gynecol
insulin and nerve growth factor. Its main source appears to be the corpus luteum of pregnancy, but it is also produced by the male prostate gland and is found in many target tissues such as cervix, myometrium, decidua, and breast connective tissue. Its principal mechanism of action appears to be the facilitation of remodeling of connective tissue in target organs to allow the necessary changes in organ structure during pregnancy and parturition. Additionally, in some mammals, relaxin inhibits myometrial contractility until near the end of pregnancy. Possible roles for relaxin in the human include enhancement of sperm motility and penetration into cervical mucus, ovum penetration, implantation of blastocyst, uterine stromal remodeling during pregnancy, the inhibition of premature labor, cervical ripening, and collagen biosynthesis in disorders of collagen metabolism such as scleroderma and arthritis. 1.2 Porcine relaxin has been well studied and purified. This preparation has been used in most animal experiments and some clinical studies. The amino acid sequence of a human relaxin derived from the recent identification of a genomic clone for relaxin is now known. s Approximately 50% homology between porcine relaxin and this human relaxin exists. Porcine and human relaxins share the same tertiary structure and some of the same amino acid sequences at biologically active sites. Human relaxin has not, as yet, been purified and the development of a specific assay must await the availability of purified samples of human relaxin for antibody development. The stated goals of Dr. Weiss and his associates are as follows: (I) purification ofrelaxin to a sufficient degree so that workable amounts of relaxin-specific antibodies can be obtained, (2) use of the antirelaxin antibodies to prepare affinity immunosorbants that can be used to complete the relaxin purification, (3) sequence of the purified relaxin to determine molecular structure, and (4) use of recombinant DNA methods to batch up adequate amounts of relaxin for clinical studies. The work presented here deals with only the first item, namely partial purification. The purification of relaxin from seminal plasma is a difficult problem and Dr. Weiss and his colleagues should be complimented for their persistance and their ability to collect "9 L" of semen as their starting material. However, there are severe limitations in the experimental approach used by the authors. First, the levels of immunoreactive relaxin in human seminal plasma are exceedingly low, approximately 9 fLg per 9 L (representing 4300 fLg of proteins). Even assuming purification to homogeneity with 100% yield, which would require a 500,000-fold enrichment, the amount of relaxin recovered would be barely sufficient for structural analysis. Dr. Weiss reports a 32-fold purification with only 16% yield after three purification steps. This is certainly not an encouraging result.