THE JOURNAL OF BIOLOGICAL CHE~~~STRY Vol. 249, No. 18, Issue of September 25, pp. 5725-5732, 1974

Printed in U.S.A.

Isolation and Characterization of the Subunits of Highly Purified Ovine Follicle-stimulating Hormone

(Received for publication, February 11, 1974)

HERBERT J. GRIMEK AND WILLIAM H. MCSHAN

From the Department of Zoology and the Endocrinology-Reproductive Physiology Program, University of Wis- consin, Madison, Wisconsin 53706

SUMMARY

Highly purified ovine follicle-stimulating hormone (FSH) was separated into two noncovalently linked, nonidentical subunits (CX and @) by treatment with 8 M urea followed by ion exchange chromatography on DEAE-Sephadex. The in- dividual subunits were purified by gel filtration on Sephadex G-100. This step resulted in the removal of a larger molec- ular weight contaminant from the /I subunit. This contami- nant, designated as undissociated FSH because of its rela- tively high biological activity, was not dissociated by further treatment with urea, suggesting that it may consist of co- valently linked subunits. Although their electrophoretic mobilities differed greatly, both subunits exhibited micro- heterogeneity as indicated by analytical acrylamide disc gel electrophoresis. The molecular weights of native FSH and the a! and /I subunits as determined by ultracentrifugation were 33,000, 15,500, and 18,500, respectively. The subunits contained the amino acids and carbohydrates usually found in glycoprotein hormones with the exception that methionine was not present in the ,8 subunit. Although they differed markedly in amino acid and carbohydrate content, both sub- units exhibited relatively high amounts of threonine, half- cystine, glucosamine, and mannose. The amino acid com- position of the a! subunit was very similar to the amino acid contents of the LY subunits of ovine and bovine luteinizing hormone. The sum of the amino acid and carbohydrate contents of the two subunits corresponded well with the amino acid and carbohydrate composition of the native FSH. Incubation of the biologically inactive CY and p subunits fol- lowed by gel filtration on Sephadex G-100 resulted in recon- stituted FSH which possessed approximately 68% of the biological activity of the original FSH as measured by the augmentation bioassay. The reconstituted FSH was al- most as potent as the original FSH in the ability to stimulate the prostates, seminal vesicles, and testes in hypophysecto- mized male rats.

* This investigation was supported bv United States Public Health Service Training Grant 5-TOl-HD-00104-08 from the National Institute of Child Health and Human Development, Ford Foundation Grant 630-0505A, and National Science Foun- dation Research Grant GB-31112.

Since Papkoff and Samy (1) isolated the subunits of ovine lu- teinizing hormone (LH), there have been numerous papers con- cerning the isolation and characterization of the subunits of glycoprotein hormones. The amino acid sequence of ovine LHi has been determined (2-4).

Investigations on the structure of FSH have been delayed be- cause of a lack of material. Papkoff and Ekblad (5) proposed that ovine FSH is composed of two nonidentical subunits. Re- search by Saxena and Rathnam (6) suggested that human FSH is also composed of subunits.

The present paper deals with the isolation and characterization of the subunits of the highly purified ovine FSH of Sherwood et al. (7) containing approximately 133 units of National Insti- tutes of Health follicle-stimulating hormone standard (NIH- FSH-Sl) per mg, dry weight, as determined by the Steelman and Pohley augmentation bioassay. This preparation, in addition to being approximately three times as potent as other ovine FSH preparations, shows notable differences in amino acid and carbo- hydrate contents (7,8). Because of the small amounts of highly purified ovine FSH available, one of the primary objectives of the present study was to isolate the subunits in a manner that would result in a good yield of pure subunits with their primary structures intact. Analytical methods were selected for their sensitivity and specificity.

MATERIALS AND METHODS

Preparation of Ovine FSH-The ovine FSH was prepared essen- tially according to the method of Sherwood et al. (7). A minor modification was introduced into the procedure in that a column of Sephadex G-199 (1.4 X 109 cm) was substituted for the column of Sephadex G-75 (1.4 X 65 cm) originally used for the final gel filtration step.

Disc Electrophoresis-Analytical acrylamide disc gel electro- phoresis at pH 8.9 (9) on columns (G X 70 mm) was used to assess the efficacy of subunit isolation procedures.

Conductivity Measurements-Conductivity measurements were made with a Radiometer model CDM 2e conductivity meter. Where desirable the conductivity readings were converted to molarity by means of an appropriate standard curve.

Separation of Subunik-Preliminary small scale studies indi- cated that the subunits of ovine FSH could be readily separated by treating the FSH with 8 M urea followed by ion exchange chromatography on DEAE-Sephadex. In order to obtain enough subunit mat&~1 for characterization studies several experiments using 14.7 to 18.7 mg of ovine FSH per experiment were performed.

i The abbreviations used are: LH, luteinizing hormone; FSH, follicle-stimulating hormone; HCG, human chorionic gonado- tropin.

5725

by guest on April 11, 2020

http://ww

w.jbc.org/

Dow

nloaded from

5726

The FSH was incubated in 8 M urea (7 mg of FSH per ml of urea) at 37” for 12 hours. The urea was obtained from Mallinckrodt and was from a lot selected especially for purity. The urea was dissolved shortlv before use in 0.04 M Tris-HCl. nH 6.5. The nH

I I

of the solution was readjusted to 6.5 with HCl after the urea was completely dissolved.

The urea solution containing the dissociated FSH was applied directly to a column (1.1 X u) cm) of DEAE-Sephadex A-50 equilibrated with 0.04 M, Tris-HCl, pH 6.5. The column was operated with the equilibrating buffer until the unadsorbed pro- tein (a subunit) was eluted. The adsorbed protein (j3 subunit and undissociated FSH) was eluted using a linear salt gradient to 0.5 M NaCl in the equilibrating buffer, with 125 ml of each solution. The eluate was monitored at 280 nm by means of a Beckman DB spectrophotometer. The fraction containing the a subunit was freeze-dried. Because of its high salt content the fraction consisting of the 0 subunit and undissociated FSH was concentrated in an Amicon ultrafiltration cell with a UM2 mem- brane prior to freeze-drying.

Gel Filtration on Sephadex G-loo-The subunits were desalted and purified by gel filtration on columns (1.4 X 100 cm) of Sepha- dex G-100 equilibrated with 0.05 M ammonium acetate. These columns (one for each subunit) were operated at 4”. The eluate was monitored at 280 nm by means of a LKB Uvicord II. The respective fractions comprising the a subunit, undissociated FSH, and p subunit were pooled and freeze-dried several times to re- move the ammonium acetate. The 0 subunit preparations used for chemical and biological studies were checked for complete removal of undissociated FSH by rechromatography on the same Sephadex G-100 column or a longer column (1.4 X 124 cm) of Sephadex G-100.

Sedimentation Equilibrium Studies-The molecular weights of ovine FSH and its subunits were determined according to the meniscus depletion method of Yphantis (10) by Dr. James Blair of the Enzyme Institute, Madison, Wis., using a Beckman Spinco model E ultracentrifuge at 4”. The samples were dissolved in 0.1 ionic strength sodium phosphate buffer, pH 7.5 (200 rg of glyco- protein per ml of buffer), and a O.l-ml aliquot was placed on a special layering fluid in a 12-mm double sector cell with sapphire windows. Ultracentrifugation was conducted at 28,090 rpm for the FSH and 44,006 rpm for the subunits. Photographs were taken after 20 hours using Rayleigh interference optics and ana- lyzed with a Gaertner microcomparator.

Amino Acid Analyses-Amino acids were quantitated on a Beckman model 120C amino acid analyzer equipped with an expanded range bridge in the recorder. The glycoprotein sam- ples, which had been dried to constant weight in a vacuum desic- cator over anhydrous calcium sulfate (Drierite) at room tempera- ture, were hydrolyzed in 6 N HCl (1 mg per ml of acid) for 22, 48, and 72 hours in sealed tubes under nitrogen. The hydrolyzed samples were dried in a vacuum desiccator containing KOH and HISO, at 0”. The desiccator was connected to a VirTis freeze- drier throughout the drying procedure. The residue was redis- solved in a small amount of water and freeze-dried again to remove any remaining HCl. The samples were analyzed accord- ing to the method of Benson and Patterson (11). Duplicate runs were made at each of the three hydrolysis times. The values for serine and threonine were extrapolated to zero time (12).

Tryptophan was determined according to t,he procedure of Bencze and Schmid (13).

Carbohydrate Analyses-The samples of ovine FSH and its subunits used for the carbohydrate determinations were dried as described previously under “Amino Acid Analyses.” Gluco- samine and galactosamine were quantitated on the Beckman model 12OC amino acid analyzer. The conditions for analysis were the same as those used for amino acid analysis except that the samples were hydrolyzed in 4 N HCl for 4 hours and the buffer change time was reduced from 95 to 75 min to allow better sepa- ration of the hexosamines from the amino acids.

The neutral sugars were determined as the alditol acetate derivatives by gas-liquid chromatography on a glass column (0.64 X 183 cm) packed with Gas-chrom Q precoated with 1% ECNSS-M (14). The column was used in a Nuclear Chicago Selectra System 5000 gas chromatograph equipped with a temper- ature programmer and flame ionization detector. The conditions for chromatography were as follows: column temperature (initial 135”, final 175”), program rate (2” per min), detector cell temper-

ature (300”), injector cell temperature (285”), hydrogen to detector (35 ml per, min). The detector settings were as follows: attenu- ation 16, range 10, balance 10.

One-milligram samples of ovine FSH and its subunits were prepared for gas chromatography essentially according to the procedure of Lehnhardt and Winzler (15) except that resin hydrol- ysis was carried out for 24 hours in an oven at 109” instead of 40 hours in a steam bath. Following ion removal and reduction the samples were stored in a desiccator containing PBOS until the day before they were scheduled to be analyzed at which time they were acetylated. Five-microliter aliquots equivalent to 50 rg of hormone or subunit were injected in triplicate into the gas chromatograph. 2-Deoxyglucose was used as an internal stan- dard.

Sialic acid was determined by the thiobarbituric acid method of Warren (16). Hydrolysis of the samples was done in 0.1 N sulfuric acid (1 mg of glycoprotein per ml of acid) at 80” for 1 hour in sealed tubes under nitrogen.

Reassociation of Subunits-Two reassociation experiments were performed. In the first experiment approximately 1106 rg of 01 subunit were combined with 1400 pg of @ subunit. In the second experiment 1509 pg of each subunit were used. Both experiments were conducted in an identical manner. Aliquots containing the above mentioned amounts of 01 and fl subunits were mixed in a 15-ml conical centrifuge tube. The contents of the tubes were freeze-dried and dissolved in 0.5 ml of 0.01 M phos- phate buffer, pH 7.0, and incubated at 37” for 16 hours. After incubation the subunit mixture was cooled to 4” and placed on a column (1.4 X 100 cm) of Sephadex G-100 equilibrated with 0.05 M ammonium acetate.

Bioassays-Specific FSH activity was determined by the aug- mentation method of Steelman and Pohley (17) using 50 i.u. of HCG. Three doses of the NIH-FSH reference standard and each unknown were administered to immature female rats ob- tained from the Holtzman Co., Madison, Wis. Four animals were used for each dose. The 50- to 60-g rats were received at 21 days of age (day 21) and were injected twice daily for 3 days beginning on day 22. The rats were killed on the morning of day 25. The results were evaluated by the parallel line method of Bliss (18). The relative potencies were adjusted to NIH- FSH-Sl. Because preliminary bioassays had suggested that the subunits were too inactive to allow a valid assessment of relative potency, they were bioassayed at two dose levels only. The results were compared to the HCG controls.

Native and reconstituted FSH, as well as the subunits were also bioassayed using hypophysectomized male rats. Two dose levels of the NIH-LH reference standard and each unknown were administered to immature rats which had been hypophysecto- mized on day 22 by Altech Laboratories, Madison, Wis. The rats were injected twice daily for 4 days beginning on the morning of day 25 and were killed on the morning of day 29. The ventral prostates, seminal vesicles, and testes were removed and weighed. The LH activity was estimated by the ventral prostate bioassay (19)

RESULTS

Isolation of Subunits-As shown in Fig. 1 DEAE-Sephadex A-50 resolved urea-treated FSH into two protein fractions. Ovine FSH not treated with urea when chromatographed on DEAE-Sephadex under similar conditions was eluted as one peak at approximately the same sodium chloride molarity as the peak designated ,f3 subunit and undiisociated FSH in Fig. 1. The “spike” peak on the side of the peak designated as the (Y subunit was an artifact caused by urea. The sharp decrease in the per cent transmission probably resulted from light scattering caused by the mixing of the 8 M urea with the buffer in the flow cell of the Beckman DB spectrophotometer. When 2 ml of 8 M urea were chromatographed on an identical column of DEAE-Sepha- dex a similar spike peak was eluted at the same position in the chromatogram.

Typical results of further purification and desalting of the subunits on Sephadex G-100 are shown in Figs. 2 and 3. The (Y subunit was eluted as one symmetrical peak with a V,/VO value

by guest on April 11, 2020

http://ww

w.jbc.org/

Dow

nloaded from

of 1.79 (Fig. 2). The /3 subunit as obtained from DEAE-Sepha- dex was resolved into two protein components (Fig. 3). The larger molecular weight component presumably consisting of un- dissociated FSH was eluted with a V,/VO value of 1.45. In com- parison native ovine FSH passed through a similar Sephadex G-100 column was eluted with a V,/VO value of 1.56. The @ subunit was eluted with a Vs/VO value of 1.84. The undissoci- ated FSH was bioassayed by the augmentation method of Steel- man and Pohley (17) and was found to contain 52.1 f 4.2 units of NIH-FSH-Sl per mg based on two bioassays. The sizable low molecular weight fraction (V,/V, 2.71) contained the buffer salts as indicated by the conductivity measurements. As this fraction was not tested for amino acids or small peptides the reason for the relatively high absorbance is not clear. In the subunit isolation experiment shown in Figs. 1, 2, and 3 in which 17.7 mg of ovine FSH were dissociated, the amounts recovered in the various fractions after Sephadex G-100 chromatography were as follows: cr subunit 5.7 mg, fi subunit 7.1 mg, and undis- sociated FSH 1.7 mg.

The results of rechromatographing the /3 subunit on Sephadex G-100 are shown in Fig. 4. Essentially all of the undissociated FSH was removed from the @ subunit by pooling the contents of the tubes as shown in Fig. 3.

The results of analytical disc gel electrophoresis conducted on the fractions obtained from the Sephadex G-100 columns are compared with native ovine FSH and urea-treated FSH in Fig. 5. The native ovine FSH exhibited one diffuse band (Fig. 5a). Treatment with 8 M urea (Fig. 56) dissociated the FSH into three

100 0 IO 20 30 40 50 60 70 60

TUBE NUMBER

FIG. 1. Ion exchange chromatography on DEAE-Sephadex of 17.7 mg of ovine FSH treated with urea. Fractions were col- lected at 12-min intervals at a flow rate of 16 ml per hour. The contents of the tubes were pooled as indicated by the stippled and hatched areas.

slowly migrating zones, the lowermost of which was very faint, and two diffuse rapidly migrating zones, the uppermost of which showed an electrophoretic mobility similar to that of native FSH. The isolated subunits and undissociated FSH (Fig. 5, c, d, and e) clearly account for the zones observed after urea treatment (Fig. 5b). Further treatment of the unclissociated FSH with 8 M urea at 37” for 12 hours failed to dissociate it into subunits as evidenced by analytical disc electrophoresis (Fig. 5f).

Sedimenlation Equilibrium Studies--The plots of log fringe dis- placement uersus the square of the radii were linear, indicating that the FSH and subunit preparations were homogeneous. The calculated molecular weights were as follows: ovine FSH, 33,000 using a partial specific volume of 0.70; (Y subunit, 15,500 using a partial specific volume of 0.70; /3 subunit, 18,500 using a partial specific volume of 0.69.

The partial specific volumes of ovine FSH and its subunits were calculated from the partial specific volumes of the individual amino acid and carbohydrate residues. The partial specific vol- umes of the amino acid residues were taken from Schachman (20) and Cohn and Edsall (21). The partial specific volumes of the carbohydrate residues were obtained from Gibbons (22).

Amino Acid and Carbohydrate Analyses-Ovine FSH and its subunits consisted of the amino acids and carbohydrates com- monly found in glycoprotein hormones (Table I). The FSH was high in lysine, aspartic acid, threonine, glutamic acid, and half- cystine but low in methionine and tryptophan. Lysine, threo- nine, glutamic acid, and half-cystine were the most abundant amino acids in the (Y subunit. The (11 subunit was low in arginine, isoleucine, leucine, and tryptophan, and its amino acid composi- tion was similar to the amino acid compositions of the (Y subunits

a-" 100 IO 20 30 40 50 60

TUBE NUMBER

FIG. 4. &chromatography on Sephadex G-106 of 5.9 mg of /3 subunit. Fractions were collected at 12-min intervals at a flow rate of 12 ml per hour. The contents of the tubes were pooled as indicated by the hatched area.

: s 2oc c! 2 z 60 r 0 v, m IO z ii 5 03 a0 F :, 0

c" 02

s 100 0 IO 20 30 40 50 60 70 TUBE NUMBER TUBE NUMBER

FIG. 2 (left). Purification and desalting on Sephadex G-100 of FIG. 3 (right). Purification and desalting on Sephadex G-100 the o subunit from 17.7 mg of ovine FSH. Fractions were col- of the 6 subunit from 17.7 mg of ovine FSH. Fractions were lected at 12-min intervals at a flow rate of 10.5 ml per hour. The collected at 12-min intervals at a flow rate of 10.5 ml per hour. contents of the tubes comprising the (I subunit were pooled as The contents of the tubes comprising the fl subunit and undisso- indicated by the stippled area. The void volume of the column ciated FSH were pooled as indicated by the hatched and stippled was 51 ml. areas, respectively. The void volume of the column was 48 ml.

by guest on April 11, 2020

http://ww

w.jbc.org/

Dow

nloaded from

5728

Comparison of the amino acid and carbohydrate compositio?zs of ovine FSH and ils subunits

ic - a subunii

M. w. = 15,500

Amino acid or

carbohydrate 3 subunit M. w. =

L8,500

Lysine 8.8 5.3 Histidine 3.0 2.8 Arginine 2.9 4.7 Aspartic acid 5.5 9.8 Threonine 8.2 11.0 Serine 5.3 7.3 Glutamic acid 7.6 8.9 Proline 6.5 5.3 Glycine 3.5 5.2 Alanine 7.1 6.0 Half-cystine 8.4 10.4 Valine 5.0 6.1 Methionine 3.7 0.0 Isoleucine 2.2 5.7 Leucine 2.3 5.0 Tyrosine 4.6 6.5 Phenylalanine 3.8 Tryptophanb

3.0 1.6 2.6

Total amino acid residues 90.0 1 06.2

Glucosamine Galactosamine F”COSEZ M.S”“OSC! Galactose Sialic acid

6.7 8.7 1.2 1.1 0.4 1.6 5.4 4.7 1.7 3.6 2.0 4.8

Total carbohydrate residues

aThe data for the amln 737 values from 22, 48, a nd

17.4 uzlds rep

72 hour it h;

24.5 sent the drolyses.

1 a e-i- I+ B I.W. = 34,000

FSH M. w. = 33,000

: c b

1

a

14.1 5.8 7.6

15.3 19.2 12.6 16.5 11.8

8.7 13.1 18.8 11.7

::9' 7.3

11.1 6.8 4.2

14.6 5.7 7.6

15.5 18.9 11.8 17.4 11.2

1Z 16.4 11.3

3.6 7.7 8.7

10.1 7.2 4.0

96.2 193.4

15.4 14.2 2.3 2.4 2.0 1.6

10.1 9.9 5.3 5.1 6.8 6.3

41.9 39.5

"elIage f extrapolated

b Tryptophan was determined by the procedure of Bencze and Schmid (13).

TABLE II Comparison of the amino acid compositions of the subunits

of bovine LH, ovine LH, and ovine FSH FIG. 5. Analytical acrylamide disc gel electrophoresis of a, ovine FSH; b, urea-treated FSH; c, FSH-CX; d, FSH-8; e, undis- sociated FSH; f, undissociated FSH treated with urea; g, FSH reconstituted from the subunits. Electrophoresis was conducted at pH 8.9 (8) at a current of 4 ma per gel. The origin was at the top and the anode at the bottom.

B 1 P G

a Subunit B Subunit Amino acid ovine ovine ovine

LH LH FSH apkoff Lamkin

and I I

et al. an(23)

-- (24)

B 6oc Residues per mole -

8.7 2.7 3.4 5.8 8.6 5.5 8.0 8.4 4.0 7.0 7.2 4.7 3.1 2.1 3.2 3.9 4.7 not iven <

9.1 3.1 3.8 5.4 8.6 6.0 7.7 9.0 4.1 7.6 8.2 5.3 3.4 2.3 3.0 4.0 3.5 not

given

8.8 3.0 2.9 5.5 8.2 5.3 7.6 6.5 3.5 7.1 8.4 5.0 3.7 2.2 2.3 4.6 3.8 1.6

Residues per mole

1 60

L

4

Lysine Histidlne Arginine Aspartx acid Threonuie Serine Glutamic acid Proline Glycine Alanine Half-cystine Valine Methionine Isoleucine Leucine Tyrosine Phenylalanine Tryptophan

2.5 2.2 2.3 2.8 5.2 7.6 4.1 5.1 5.6 6.0 6.4 6.8 7.1 6.0

16.6 19.9 6.2 6.6 6.3 7.4 6.9 9.6 5.1 7.0 1.9 1.8 3.3 3.4 8.4 10.6 1.6 1.5 2.1 2.7 not not

given given

5.3 2.8 4.7 9.8

11.0 7.3 8.9 5.3 5.2 6.0

10.4 6.7 0.0 5.7 5.0 6.5 3.0 2.6

f5 Yil t REASSOCIATED FSH

TUBE NUMBER

FIG. 6. Purification on Sephadex G-100 of FSH reconstituted from the subunits. Fractions were collected at 12.min intervals at a flow rate of 10 ml per hour. The contents of the tubes were pooled as indicated by the stippled and hatched areas.

Reassociation of Subunits-The results of gel filtration of the incubation mixture of the second reassociation experiment where 1500 pg of o( subunit were recombined with 1500 pg of /3 subunit are shown in Fig. 6. The results of the first reassociation experi- ment where 1100 pg of a! subunit were combined with 1400 pg of @ subunit were similar. The reconstituted FSH was eluted with a V,/Vo value of 1.58 which is essentially identical with the V,/VO value of 1.56 obtained for native ovine FSH passed through a similar column. Approximately 1.8 mg of reconstituted FSH were obtained from 3.0 mg of subunit material.

As shown in Fig. 5g the electrophoretic properties of the FSH reconstituted from the subunits were similar to those of the orig-

of bovine and ovine LH (Table II). The p subunit showed rela- tively high amounts of threonine and half-cystine and low amounts of histidine and tryptophan. The @ subunit contained no methionine and its amino acid composition differed greatly from the amino acid compositions of the p subunits of bovine and ovine LH (Table II).

In general, glucosamine and mannose were the most abundant carbohydrates found in the ovine FSH and its subunits. Galac- tosamine and fucose were the least abundant. The ovine FSH and the p subunit also contained substantial amounts of sialic acid.

by guest on April 11, 2020

http://ww

w.jbc.org/

Dow

nloaded from

5729

inal FSH. Disc electrophoresis of the “shoulder” observed in a 90% restoration of the ability to stimulate these organs. Al- Fig. 6 showed it to consist of the expected mixture of subunits and though the relative potencies given in the second footnote to reassociated FSH. Table V are useful fo.r comparing the native FSH to the recon-

Bioassays-The individual subunits were inactive at dose stituted FSH, they probably should not be accepted as absolute levelsashigh as 200pginthe augmentationbioassay (Table III). values because of the nonparallelism at the 0.05 level of proba- When the subunits were recombined, however, 67 to 68% of the bility duetothe relatively poor dose response observed with the original biological activity was recovered (Table IV). NIH-LH standard.

Theindividualsubunitsdidnotstimulatetheventralprostates, seminalvesicles, ortestesinhypophysectomized male rats (Table TABLE IV

V). Recombination of the subunits resulted in approximately Comparison of the results of Steelman-Pohley augmen bioassays of native ovine FSH and FSH

reconstituted from the subunits

tation

TABLE III Steelman-Pohley augmentation bioassays of the

subunits of ovine FSH

Experi- menP Preparation

Preparation FSH (FSH-o( + FSH-PY’

HCG control (50 i.u.) NIH-FSH-S9

wz

50 100 200

50 200

50 200

Ovarian weighta 1

mg 58.5 zk 3.0 2 69.5 z!z 4.1 96.2 f 9.3 ~

FSH (FSH-or + FSH-fi)*

FSH-a

FSH-fl

a Results are mean f standard error.

89.24 119.90 60.27 80.62

103.43 141.87 72.08 97.18

166.65 115.42

192.66 0.1526 130.82 0.1444

Index of precision

174.0 * 10.4 59.8 zt 5.3 55.0 * 1.5 56.5 z!z 3.7 66.0 f 5.7

a In Experiment 11100pgofor subunit were combined with1400 pg of p subunit. In Experiment 2 1500 rg of a subunit were com- bined with 1500rg of p subunit (see “Materials andMethods”).

* The values for the reconstituted FSH represent the average for two groups of rats compared with the same standard. Each group consisted of the usual three-dose levels, using four animals per dose.

TABLE V Biological activity of ovine FSH, the subunits, and FSH reconstituted from the subunits in hypophysectomized male rats

Preparation

Saline controls

NIH-LH-S17

FSH

FSH-a

FSH-B

(FSH-cx + FsH-i3)’

Dose ug

9.3 + 0.9(4) 5.8 + 0.3 200.3 + 10.4 -

5 13.5 + 1.2(4) 7.0 + 0.0 256.0 + 29.0 20 19.3 E 3.7(3) 7.3 T 0.9 281.3 + 64.0 - -

15 17.0 + 1.8(4) 6.8 + 0.5 507.5 + 16.2 60 32.7 + 4.4(3) 16.0 T 1.0 521.3 T 33.2 - - -

50 11.0 + l-7(4) 6.8 + 1.5 225.8 + 12.8 200 9.3 T 0.9(3) 7.3 T 0.3 225.0 T 20.2 - - -

50 11.3 + 2.0(3) 7.7 + 0.3 226.3 + 22.5 200 9.0 + 0.0(3) 7.0 + 0.6 204.0 r 17.5 - - -

25 20.0 + 1.7(3) 9.3 + 0.9 482.7 + 22.8 100 35.3 T 2.0(3) 21.0 + 0.0 564.7 T 23.4 - - -

Ventral Prostateb Seminal Vesicles Testes

Organ weightsa m9

aMean organ weight + the standard error. The number of rats are given in parentheses following the ventral prostate weights. Where only three rats are indicated, the fourth rat either died or was rejected due to incomplete hypophysectomy.

b The statistical evaluation of the ventral prostate data was as follows: (relative potency ovine FSH = 1.00 x NIH-LH-S17; 95% confidence limits = 0.51 - 3.58; index of precision = 0.2568; significant nonparallelism at p = 0.05 but not at p = O.Ol), (relative potency of reconstituted FSH = 0.88 x NIH-LH-S17; 95% confidence limits = 0.49 - 2.43: index of precision = 0.1961; significant nonparallelism at p = 0.05 but not at p = 0.01).

'Fifteen hundred ug of cx subunit were combined with 1500 ug of $ subunit as described under "MATERIALS AND METHODS".

by guest on April 11, 2020

http://ww

w.jbc.org/

Dow

nloaded from

5730

DISCUSSION

Urea has been used successfully to separate the subunits of a number of glycoprotein hormones including HCG (25), human FSH (6), and human LH (26). Urea is preferable as a dissociat- ing agent because of its relative mildness and lack of charge, allowing it to be applied directly to an ion exchange column.

One molar propionic acid which has a pH of approximately 2.4, as used by Papkoff and Ekblad (5) to dissociate ovine FSH, was not considered desirable because it may cause hydrolysis of labile bonds. Dissociation at low pH values also may cause charge effects which can complicate physical characterization (27). The apparent aggregation of the subunits of ovine FSH observed by Papkoff and Ekblad (5) may be an example of this.

The use of urea as a dissociating agent is of course not com- pletely without problems. Cyanate ions will slowly accumulate in concentrated urea solutions (28, 29). An 8 M urea solution will eventually become 0.02 M with respect to cyanate ion. This concentration of cyanate may be great enough to cause carbam- ylation of various functional groups on the protein especially the e-amino groups of lysine (30).

As a precaution against carbamylation the 8 M urea solutions used in these studies were freshly prepared from high purity urea shortly before each dissociation experiment, and the dissociation was conducted at pH 6.5 using a Tris-containing buffer. Ac- cording to Peterson (31) carbamylation of the e-amino groups of lysine does not occur at acidic pH if the urea is low in cyanate. Peterson (31), however, did not specify the pH or cyanate con- centration. A buffer containing Tris should serve to keep the cyanate concentration low because Tris competes with the protein for cyanate ions formed from the urea (32).

Dissociation at a pH slightly below neutrality may also have a retarding effect on disulfide interchange reactions. Huggins et al. (33) found that a 5.6% solution of plasma albumin in 8 M urea gelled in 30 min at pH 11. At pH 7.0, however, 16 hours were required for gelation. By the use of various sulfhydryl reagents they concluded that the gelation was due to the formation of in- termolecular disulfide bonds.

The ovine FSH and its subunits displayed a degree of micro- heterogeneity as indicated by analytical disc gel electrophoresis (Fig. 5, a, b, c, and d). Microheterogeneity appears to be com- mon among the gonadotropins and their subunits. Saxena and Rathnam (6) noted that both subunits of human FSH exhibited multiple zones during disc electrophoresis. The electrophoretic patterns of the subunits of ovine FSH obtained by Papkoff and Ekblad (5) also suggest microheterogeneity. Microheterogeneity of the subunits has also been noted for HCG (25, 34) and human LH (26).

The microheterogeneity could have its origin in either the pep- tide or carbohydrate moiety. Grimek el al. (8) using the 5-di- methylaminonaphthalene-1-sulfonyl chloride procedure of Gros and Labouesse (35) recovered serine, glycine, phenylalanine, pro- line, and isoleucine from the N termini of ovine FSH. Hetero- geneity of the N terminus has also been noted for the a! subunit of ovine LH (3). The origin of this heterogeneity is not clear, but it is possible that labile bonds at the terminal positions may be cleaved by proteolytic enzymes early in the purification pro- cedure.

Microheterogeneity of oligosaccharide units has been reported for a variety of glycoproteins (36). In addition to the variable stages of completion of the carbohydrate moieties of glycopro- teins, there is also the possibility of positional isomerism. In (~1 acid glycoprotein, for example, sialic acid can be linked to

either carbon 3,4, or 6 of the adjoining galactose residue result- ing in polymorphism on starch gel electrophoresis (36).

The results of amino acid and carbohydrate analyses presented in Table I indicate that ovine FSH is composed of two nonidenti- cal subunits in a 1: 1 ratio. The similarity of the amino acid composition of the (Y subunit of ovine FSH to the amino acid compositions of the LY subunits of ovine and bovine LH (Table 11) suggests the possibility that their ammo acid sequences may also be similar, if not identical. It has recently been shown that the (Y subunits of ovine LH and bovine thyroid-stimulating hormone share identical amino acid sequences (2,3, 37).

The results of the amino acid and carbohydrate analyses of the ovine FSH and its subunits obtained in this study are not in ac- cord with those of an earlier study by Papkoff and Ekblad (5). The original FSH preparations with the exception of threonine, half-cystine, leucine, and tyrosine are somewhat similar in amino acid content. The differences among the cy subunits are pro- nounced. The previous study (5) suggested that the a! subunit is high in aspartic acid, glutamic acid, and leucine but low in half- cystine and extremely low in methionine. The present study indicates, however, that the (Y subunit is high in half-cystine and contains a substantial amount of methionine. In addition the current results indicate that the cr subunit is very low in leucine and possesses significantly lower amounts of aspartic and glu- tamic acids.

The amino acid compositions of the fi subunits obtained in the two studies are somewhat similar. Although the values for lysine, glutamic acid, alanine, and tyrosine differ by almost 3 residues per mole, there are no outstanding differences as in the case of the (Y subunits. With the exception of the hexose content of the @ subunit the values obtained for the carbohydrate com- position of ovine FSH and its subunits in the previous study are much lower than those given in this report. The differences ob- served between the present study and that of Papkoff and Ekblad (5) may reflect differences in the purity of the original FSH prep- arations and/or the methods used to isolate the subunits.

The /3 subunit of ovine FSH as obtained from DEAESephadex contained approximately 19% of a larger molecular weight sub- stance tentatively designated as undissociated FSH. Rathnam and Saxena (38) found that the /3 subunit of human FSH as pro- cured from DEAE-Sephadex (6) contained more than 50% of a higher molecular weight fraction. Similar results were noted for human LH by Rathnam and Saxena (26). Although they did not report the results of any bioassays or other studies conducted on this contaminant of the fi subunit, they postulated that it prob- ably consisted of undissociated or reassociated hormone or a polymerized subunit.

The observation that the undissociated FSH removed from the /3 subunit of ovine FSH by gel filtration on Sephadex G-100 (Fig. 3) apparently could not be dissociated into subunits by further treatment with 8 M urea (Fig. 5f) suggests some interesting possi- bilities. In consideration of its biological activity (approxi- mately 52 X NIH-FSH-Sl) it is unlikely that this fraction simply represents an aggregate of @ subunits.

The undissociated FSH may consist of a! and /? subunits pos- sessing the same electrophoretic mobility and hence will appear as essentially one zone during electrophoresis. It is difficult, however, to conjure up a reasonable explanation why the undis- sociated FSH should be composed only of subunits with identical electrophoretic mobilities.

Another possibility is that the two subunits of undissociated FSH are joined by disulfide bonds. Disulfide interchange reac- tions can occur in proteins especially during treatment with de-

by guest on April 11, 2020

http://ww

w.jbc.org/

Dow

nloaded from

5731

naturing agents such as urea (33,39). Disulfide interchange pre- Acknowledgments--We are grateful to Mr. Lyman L. Lyons for sumably could result in the formation of disulfide bridges between his assistance in the preparation of ovine FSH and Dr. Jack E. the subunits. Lilien and Mr. Donald J. O’Brien for help with the gas-liquid

An intriguing speculation is that the undissociated FSH repre- chromatography of the neutral sugars. sents a prohormone form of FSH in which the two subunits are joined by a connecting peptide. Vaitukaitis (40) has recently proposed that HCG is synthesized as a prohormone.

Although other investigators (26, 38) have noted incomplete dissociation of gonadotropins following urea treatment, the exact nature of the undissociated portion remains unknown. Studies involving cleavage of the disulfide bonds, amino acid analysis, and determination of the NHt- and COOH-terminal amino acids should help to clarify this matter.

1.

2.

3.

4. Although a previous study (5) suggested that the subunits of

ovine FSH have slight biological activity, the results presented in this report indicate that they are completely inactive. The low activity noted in the previous study may have been due to contamination with undissociated or reassociated FSH. Bio- logical activity of the subunits, especially the /3 subunit, has also been noted in the case of human FSH (6). These results, how- ever, were based on the B subunit as obtained from DEAE-Sepha- dex. The /3 subunit of human FSH was later found to be con- taminated with large amounts of a higher molecular weight material, possibly undissociated FSH (38).

5.

6.

7.

8.

9. 10. 11.

The FSH reconstituted from the subunits possessed 67 to 68% of the activity of the original FSH (Table IV). These figures are somewhat higher than those previously report.ed (5). The ob- served difference in recovery of activity after recombination may be partially due to the fact that in the present study the subunit mixture was subjected to gel filtration after incubation. The bioassay was therefore based only on the reconstituted FSH rather than on a mixture of reconstituted FSH and subunits as apparently was the case in the previous study (5).

12.

13.

14. 15.

16. 17.

The recovery of biological activity after recombination prob- ably could be increased. Hartree et al. (41) recovered 100% of the original activity of human LH upon recombination of the subunits. The complete recovery of activity reported by these workers may be due to the fact that they incubated relatively large quantities of subunits (about 8 mg of each subunit) in a small volume (0.5 ml) at 4”. They also subjected the subunit mixtuie to gel filtration after incubation. It should be pointed out, however, that one or both of the subunits may never com- pletely regain the original tertiary structure after treatment with 8 M urea.

18.

19.

20. 21.

22.

23.

24.

cept that a degree of LH-like activity is inherent in the ovine FSH molecule.

Previous studies (7, 42) have suggested that ovine FSH possesses intrinsic LH-like activity. The fact that in the present study recombination of the subunits, which were by themselves inactive, resulted in an almost complete restoration of the ability to stimulate the ventral prostates and seminal vesicles in hy- pophysectomized male rats (Table V), strongly supports the con-

25.

26.

27.

28. DIRNHUBER, P., AND SCHUTZ, F. (1948) Biochem. J. 42. 6% 632

VAN HOLDE, K. E. (1970) in Protein Sequence Determination (NEEDLEMAN, S. B., ed) pp. 4-24, Springer-Verlag, New York

In conclusion, the subunits of highly purified ovine FSH ap- proximately three times as potent as previous preparations have been isolated and characterized. The two dissimilar subunits which were biologically inactive individually, can be recombined with a substantial recovery of the biological activity of the orig- inal hormone. The amino acid composition of the (Y subunit was similar to those of the a! subunits of ovine and bovine LH. These studies should pave the way for sequence analysis of ovine FSH. Admittedly, sequence analysis of ovine FSH will be diffi- cult because of the small amounts of material available and will require the most modern techniques in microanalysis.

REFERENCES

PAPKOFF, H., AND SAMY, T. S. A. (1967) B&him. Biophys. Acta 147, 175-177

PAPKOFF, H., SAIRAM, M. R., AND LI, C. H. (1971) J. Amer. Chem. Sot. 93, 1531-1532

LIU, W.-K., NAHM, H. S., SWEENEY, C. M., LAMKIN, W. M., BAKER, H. N., AND WARD, D. N. (1972) J. Biol. Chem. 247, 43514364

LIU, W.-K., NAHM, H. S., SWEENEY, C. M., HOL~OMB, G. N., AND WARD, D. N. (1972) J. Biol. Chem. 24’7.4365-4381

PAPKOFF, H., AND EKBLAD, M. (1970) Biochem. Biophys. Res. Commun. 40, 614-621

SAXENA, B. B., AND RATHNAM, P. (1971) J. Biol. Chem. 246, 3549-3554

SHERWOOD, 0. D., GRIMEK, H. J., AND MCSHAN, W. H. (1970) J. Biol. Chem. 246, 2328-2336

GRIMEK, H. J., NUTI, L. C., .~ND MCSHAN, W. H. (1971) Bio- them. Biophys. Res. Commun. 46, 879-886

DAVIS, B. J. (1964) Ann. N. Y. Acad. Sci. 121, 404-427 YPHANTIS, D. A. (1964) Biochemistry 3, 297-317 BENSON, J. V., JR., AND PATTERSON, J. A. (1965) Anal. Bio-

them. 13, 265-280 SMITH, E. L., AND STOCKELL, A. (1954) J. Biol. Chem. 207,

501-514 BENCZF,, W. L., AND SCHMID, K. (1957) Anal. Chem. 29, 1193-

1196 NIEDERMEIER, W. (1971) Anal. Biochem. 40, 465-475 LEHNHARDT, W. F., AND WINZLER, R. J. (1968) J. ChTomatogr.

34, 471479 WARREN, L. (1959) J. Biol. Chem. 234, 1971-1975 STEELM.4N, S. I,., AND POHLEY, F. M. (1953) Endocrinology

63, 604616 BLISS, C. I. (1952) The Statistics of Bioassay, Academic Press,

New York GREEP, R. O., VAN DYKE, H. B., AND CHOW, B. F. (1941)

Proc. Sot. Exp. Biol. Med. 46, 644-649 SCHACHMAN, H. K. (1957) Methods Enzymol. 4.32-103 COHN, E. J., AND EDSALL, J. T. (1943) Proteins, Amino Acids,

and Peptides, pp. 37&381, Reinhold, New York GIBBONS, R. A. (1966) in Glycoproteins (GOTTSCHALK, A., ed)

pp. 29-95, American Elsevier, New York PAPKOFF, H., AND GAN, J. (1970) Arch. Biochem. Biophys.

136, 522-528 LAMKIN, W. M., FUJINO, M., MAYFIELD, J. D., HOLCOMB,

G. N., AND WARD, D. N. (1970) Biochim. Biophys. Acta 214. 290-298

SWAMINATHAN, N., AND BAHL, 0. P. (1970) Biochem. Biophys. Res. Commun. 40, 422-427

RATHNAM, P., AND SAXENA, B. B. (1971) J. Riol. Chem. 246, 7087-7094

29. MARIER, J. R., AND ROSE, D. (1964) Anal. Biochem. 7, 304- 314

30. STARK, G. R., STEIN, W. H., AND MOORE, S. (1960) J. Biol. Chem. 236, 31773181

31. PETERSON, R. F. (1972) Methods Enzymol. 26, 178-182 32. HIRS, C. H. W. (1967) Methods Enzymol. 11, 199-203 33. HUGGINS, C., TAPLEY, D. F., AND JENSEN, E. V. (1951) Nature

167, 592-593 34. MORGAN, F. J., AND CANFIELD, R. E. (1971) Endocrinology

36, 10451053 35. GROS, C., AND LABOUESSE, B. (1969) Eur. J. Biochem. 7,

463470 36. SPIRO, R. G. (1969) New Eng. J. Med. 281, 991-1001

by guest on April 11, 2020

http://ww

w.jbc.org/

Dow

nloaded from

5732

37. LIAO, T.-H., AND PIERCE, J. G. (1971) J. Biol. Chem. 246, 85&865

40. VAITUKAITIS, J. L. (1973) Abstracts of the 66th Meeting of the

38. RATHNAM, P., AND SAXENA, B. B. (1972) in Gonadotropins Endocrine Society, Chicago, Ill.

(SAXENA, B. B., BELING, C. G., AND GANDY, H. M., eds) 41. HARTREE, A. S., THOMAS, M., BRAIKEVITCH, M., BELL, E. T.,

pp. 120-131, Wiley-Interscience, New York CHRISTIE, D. W., STAULL, G. V., TAYLOR, R., AND PIERCE,

39. HAUROWITZ, F. (1963) The Chemistry and Function of Pro- J. G. (1971) J. Endocrinol. 61, 169-180 teins, 2nd Ed, pp. 132-183, Academic Press, New York 42. PAPKOFF, H. (1965) Acta Endocrinol. 48, 439-445

by guest on April 11, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Herbert J. Grimek and William H. McShanFollicle-stimulating Hormone

Isolation and Characterization of the Subunits of Highly Purified Ovine

1974, 249:5725-5732.J. Biol. Chem. 

  http://www.jbc.org/content/249/18/5725Access the most updated version of this article at

 Alerts:

  When a correction for this article is posted• 

When this article is cited• 

to choose from all of JBC's e-mail alertsClick here

  http://www.jbc.org/content/249/18/5725.full.html#ref-list-1

This article cites 0 references, 0 of which can be accessed free at

by guest on April 11, 2020

http://ww

w.jbc.org/

Dow

nloaded from