Purification of lutropin and follitropin in high yield from horse pituitary glands.

A method has been developed for the purification of equine lutropin (eLH) and equine follitropin (eFSH) from horse pituitary glands which attains high yields of both hormones in contrast to previous methods that were devoted to one or the other with inferior recovery of the hormones. Two-pass chromatography over CM-Sephadex was used to separate eLH from eFSH. Subsequent steps employing QAE (quaternary amino-ethyl)-Sephadex chromatography and gel filtration on Sephacryl S-200 produced highly purified hormone preparations. Yields of purified eLH and eFSH were 110 and 60 mg/kg of frozen pituitaries, respectively. Subunits were prepared by dissociation in 8 M guanidine HCl followed by either gel filtration (eLH) or gel filtration followed by QAE-Sephadex chromatography (eFSH). The hormones and their subunits were characterized by sodium dodecyl sulfate-gel electrophoresis, amino acid analysis, NH2-terminal analysis, and by LH and FSH radioligand receptor assays.

* This work was supported by Grants AM 09801 and HD 8338 from the National Institutes of Health and Grants G-051 and G-147 from The Robert A. Welch Foundation. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
$ To whom correspondence should be addressed. The abbreviations used are: LH, luteinizing hormone (lutropin); FSH, follicle-stimulating hormone (follitropin); eCG, equine chorionic gonadotropin (pregnant mare serum gonadotropin); prefixes e and o in hormone abbreviations, species of origin, equine or ovine; suffixes 01 or p, the appropriate subunit derived from the hormone indicated SDS, sodium dodecyl sulfate; RLA, radioligand assay; dansyl, 5-dimethylaminonaphthalene-1-sulfonyl; PTH, phenylthiohydantoin derivative of the amino acid designated. kg), but attended by a very low yield of eFSH (7 mg/kg). They also demonstrated that the rat was an inappropriate assay animal since eLH was found to have intrinsic FSH activity in binding assays and in the Steelman-Pohley assay. These results suggested that one of the problems of resolving LH a n d FSH activities from horse pituitary extracts was the use of rat-based assays. Combarnous and Henge (4) reported a purification procedure for eFSH which produced 13.9 mg of eFSH/kg of pituitaries. Equine LH was not taken into consideration. However, they observed that eFSH readily dissociated into its subunits under slightly acidic conditions. Purification of both eLH and eFSH was later reported by Guillon and Combarnous (5). The yield of eLH was 24.2 mg/kg, and the yield of eFSH had been improved over that of the earlier report to 26 mg/kg.
We have undertaken the purification of both eLH and eFSH attempting to develop efficient methods for the purification of both of these hormones in high yield. Our approach has avoided the use of low pH steps and has employed in vitro bioassays that can distinguish eLH from eFSH. In this report we describe procedures to obtain highly purified eLH and eFSH preparations in high yield.

Materials
Frozen horse pituitary glands (Lot S1011) were obtained from Armour. Sephadex and Sephacryl chromatography products were obtained from Pharmacia. Polyhrene was obtained from Pierce Chemical Co. All other sequenator reagents were obtained from Beckman. Ovine LH and FSH reference preparations, NIH-LH-S19 and NIAMDD-oFSH-13, were obtained from the Hormone Distribution Officer, National Institute of Arthritis, Metabolism, and Digestive Diseases. All other reagents were reagent grade or the highest purity commercially available.

Methods
Extraction of Horse Pituitary Glands-Frozen horse pituitary glands were extracted by a modified version of the percolation extraction method used by Bates et al. (6) in the purification of human thyrotropin. The glands were extracted by suspension in the various solvents rather than by percolation, and a higher pH was maintained throughout.
All procedures were performed at 4 "C. Two kilograms of frozen glands were soaked overnight in 6.4 liters of alcohol. The glands were homogenized in 1-kg lots by four 15-s bursts at the high setting of a Waring Blendor. Assuming a water content of 80% for the glands, the ethanol content of the resulting homogenate was 75%. The homogenate was immediately centrifuged at 6000 X g for 20 min in a Sorvall RC-3 centrifuge. The supernatant was discarded and the pellet resuspended in 4 liters of 75% ethanol, 25% 0.5 M sodium acetate, pH 6.0, and extracted for 2 h. Following centrifugation the supernatant was discarded and the pellet extracted for 2 h in 60% ethanol, 40% 0.5 M sodium acetate, pH 6. The supernatant was saved and the pellet extracted overnight with 4 liters of 50% ethanol containing 1 M NaCl and 0.5 M Tris-acetate, pH 7. This was followed Equine Gonadotropin Purification by overnight extraction with 40% ethanol containing 1 M NaCl and 0.5 M Tris acetate, pH 7, and concluded with overnight extraction with 40% ethanol containing 1 M NaCl buffered with 0.5 M sodium acetate, pH 5. The supernatants of the 60,50, and both 40% ethanol extracts were adjusted to 80% ethanol. The resulting precipitates were allowed to settle overnight and recovered by centrifugation. The pellets were resuspended in 0.126 M ammonium bicarbonate, to which 2 mM phenylmethylsulfonyl fluoride was added, and dialyzed against distilled water. Following dialysis the insoluble fractions were removed by centrifugation, pooled, and lyophilized. The soluble fractions were lyophilized separately and assayed by rat testis radioligand assay.
Purification of Equine LH and FSH-The fractions extracted at pH 7 with 50% ethanol and with 40% ethanol from three 2-kg batches of horse pituitaries were pooled. The 26.7 g of extract were applied to a column (15 X 11 cm) of CM-Sephadex (C-50) equilibrated with 0.005 M sodium phosphate buffer, pH 6.0 (Buffer A). The column was developed with the discontinuous buffer system of Ward et al. (7)  Both CMS-1D and CMS-1E were concentrated and dialyzed by ultrafiltration over a PM-10 membrane in an Amicon ultrafiltration cell prior to lyophilization.
CMS-lA, the unabsorbed fraction from CM-Sephadex chromatography of an additional 28.6 g of extract, and the 40% ethanol, pH 5, extracts were dissolved in 0.025 M sodium phosphate buffer at pH 7.5 at a protein concentration of 20 mg/ml. A 500-p1 aliquot was taken for RLA and SDS-gel electrophoresis. Solid ammonium sulfate was added to reach 55, 60, and 80% saturation. Precipitates at each point were collected by centrifugation, dissolved in distilled water, dialyzed against distilled water, and lyophilized. The 55% saturated ammonium sulfate precipitate (18.5 g of protein) was chromatographed on a column (15 X 20 cm) of CM-Sephadex. The column was developed with Buffers A-D, described above, and with Buffer F (0.15 M sodium chloride in 0.1 M Tris-HC1, pH 9.2) at a flow rate of 2.6 ml/h/cm2. Crude eLH (CMS-2D) was eluted with Buffer D while Buffer F eluted a crude eFSH fraction (CMS-2F). CMS-2D was rechromatographed on a CM-Sephadex column (5 X 28 cm) which was developed at a flow rate of 3.1 ml/h/cm2 with Buffers A-E. The eLH fraction, CMS-4D, which was eluted with Buffer D, was desalted over a Sephadex G-25 column equilibrated with 0.126 M ammonium bicarbonate. Following lyophilization, CMS-4D was applied to a column (5 X 146 cm) of Sephadex G-100 equilibrated with 0.126 M ammonium bicarbonate buffer and the column developed at a flow rate of 4.8 ml/h/cm2. The purified eLH fraction obtained from this column was designated eLH-C.
Crude gonadotropin, CMS-lD, was rechromatographed on a column (5 X 28 cm) of CM-Sephadex developed at a rate of 3.1 ml/h/ cm2 with Buffers A-E. This produced two fractions: a crude eLH fraction, CMS-3D, and a crude eFSH fraction, CMS-3E. CMS-3D was applied to a column (15 X 1.3 cm) of QAE (quaternary aminoethyl)-Sephadex (A-50) equilibrated with 0.01 M ammonium bicarbonate (Buffer I). The column was developed at a flow rate of 0.9 ml/ h/cm2. Following removal of the unabsorbed fraction the column was eluted with 0.1 M ammonium bicarbonate (Buffer II), followed with 0.2 M ammonium bicarbonate (Buffer 111), and with 1 M ammonium bicarbonate (Buffer IV). Protein was recovered by lyophilization. The two largest fractions eluted with Buffers I1 (QAE-11-LH) and 111 (QAE-111-LH) were further purified by gel filtration on Sephacryl S-200 columns equilibrated with 0.126 M ammonium bicarbonate. QAE-11-LH was chromatographed on a S-200 column (5 X 120 cm) at a flow rate of 4 ml/h/cm2. Two eLH fractions were recovered. One was 159 mg of highly purified eLH, designated eLH-B. The other eLH fraction (S2OO-LH) was further purified by gel filtration on Sephadex G-100 to produce 239 mg of purified eLH-A. QAE-111-LH was chromatrographed on a S-200 column (2.5 X 158 cm) to yield 107 mg of eLH-C.
Purity of the initial crude eFSH fractions determined the complexity of subsequent purification procedures. CMS-2F was applied to a column (5 x 3.9 cm) of QAE-Sephadex (A-50) equilibrated and developed with the ammonium bicarbonate Buffers I-IV used above in the purification of eLH. The fraction eluted with Buffer 111 yielded 378 mg of highly purified eFSH, designated eFSH-C. CMS-3E was purified by chromatography on a QAE-Sephadex column (2.5 X 5.6 cm). The Buffer 111 eluate (QAE-FSH) was chromatographed on the 5-200 column (2.5 X 158 cm) yielding 94 mg of purified eFSH-B.
For the crudest eFSH fraction, CMS-1E (see Fig. 1, gel inset), a special procedure was required to remove many contaminants which could not be eliminated by QAE-Sephadex chromatography alone. A column (5 X 5 cm) of the cation exchanger SE-Sephadex (C-50) equilibrated with 0.05 M ammonium bicarbonate was coupled to a column (5 X 5 cm) of QAE-Sephadex equilibrated with the same buffer. CMS-1E was applied to the SE-Sephadex column which was developed with starting buffer until all of the unabsorbed protein had been washed onto the QAE-Sephadex column. The columns were then uncoupled and developed separately. Absorbed proteins were eluted from the SE-Sephadex column with 1 M ammonium bicarbonate. The QAE-Sephadex column was eluted with 0.1, 0.2, and 1 M ammonium bicarbonate. SQ-FSH (SP-Sephadex/QAE-Sephadex-purified), the fraction eluted with 0.2 M ammonium bicarbonate, was chromatographed on S-200 as described for QAE-FSH. The purified eFSH was designated eFSH-A (150 mg of protein).
Separation of eLH into Subunits-A sample of eLH (200 mg) was incubated in 8 M guanidine HCl overnight at 37 "C. The dissociated subunits were separated on a column (2.5 X 158 cm) of Sephacryl S-200 equilibrated with 0.126 M ammonium bicarbonate. The two subunit peaks were pooled separately, lyophilized, and assessed for purity by SDS-polyacrylamide gel electrophoresis, NH2-terminal analysis, and radioligand assay.
Separation of eFSH into Subunits-Equine FSH (31.9 mg) was dissociated into subunits in 8 M guanidine HC1 and chromatographed on S-200 as described above. The subunit fraction emerged as a single peak, and protein was recovered by lyophilization. (This step proved ineffective and can be eliminated if suitable allowance for the guanidine HCl is made.) Subunits were separated by ion exchange chromatography on a column (2.5 X 6.1 cm) of QAE-Sephadex equilibrated with 0.01 M ammonium acetate, pH 5.5. The a subunit emerged unabsorbed and the p subunit was eluted with 0.4 M ammonium acetate, pH 5.5. The subunit fractions were lyophilized and assessed for purity as described above. Assessment of Purity-Purification of eLH and eFSH was followed on SDS-polyacrylamide gels using the discontinuous buffer system of Laemmli (8) with 12 or 15% slab gels as described by Moore and Ward (9) with the following modifications. Samples were boiled for 2 min rather than being incubated at 60 "C for 1 h and electrophoresed in a Bio-Rad Protean double slab electrophoresis cell. The gels were fixed in 45% methanol, 10% acetic acid with constant shaking for 2 h to remove SDS prior to staining in 0.25% Coomassie brilliant blue R-250 for 6 h or overnight. The gels were destained in 20% methanol, 10% acetic acid. Gels were scanned at 525 nm in a Helena Laboratories Quick Scan Jr instrument. The maximum for each 525-nm absorbance peak in the scan was used to determine the RF for each protein band. Molecular weights were estimated from a set of proteins of known molecular weight (10).
Gonadotropic activities were followed by LH and FSH radioligand receptor assays using 1251-labeled eLH or eFSH as radioligands and rat testis homogenate and crude horse testis membranes (11) or chicken testis homogenate (12) as receptor preparations. Relative potencies were calculated from the IDsos determined from the inhibition curves (13).
Analytical Procedures-Amino acid compositions of equine gonadotropins and their subunits were determined on samples hydrolyzed in 6 N HC1 containing 0.1% phenol for 24 and 72 h. Half-cystine values were determined on performic acid-oxidized or reduced and carboxymethylated samples. The hydrolysates were analyzed on an LKB model 4400 amino acid analyzer equipped with a Shimadzu model C-RlB recording integrator (570-nm channel) and with a Hewlett-Packard model 3309 integrator (440-nm channel). Tryptophan was determined by the method of De Traglia et al. (14). Sialic acid determinations and NH2-terminal analyses by the dansyl method were carried out as described by Moore and Ward (9). In addition, NH2-terminal amino acids were determined by automated amino acid sequencing of intact hormone samples or on isolated subunits. One mg of intact hormone or 500 pg of subunit were dissolved in 300 pl of distilled water containing 3 mg of Polybrene. The mixture was loaded into the spinning cup of an updated Beckman 890B Sequencer equipped with a liquid nitrogen trap, Sequemat P-6 auto-converter, and Sequemat SC-510 programmer. After drying the sample under Equine Gonadotropin Purification vacuum, three or more sequencer cycles were run using the 1 M Quadrol program 060275, provided by Beckman. P T H derivatives were analyzed by high performance liquid chromatography on a Glenco high performance liquid chromatograph equipped initially with a Hewlett-Packard model 3370.4 integrator and later with a Hewlett-Packard model 3309 integrator. The isocratic solvent system of Tarr (15) (38% acetonitrile/62% 0.065 M sodium acetate, pH 4.5) was employed on an Altex Ultrasphere ODS column a t 55 "C using a flow rate of 0.55 ml/min.

Extraction of eLH and eFSH from Frozen Horse Pituitaries
The results of the extraction of frozen horse pituitary glands are summarized in Table I. Preliminary results indicated that only fractions extracted with less than 75% ethanol contained gonadotropic activity. Therefore, only the 60, 50, and both 40% ethanol extracts were tested in rat testis LH and FSH radioligand assays. As can be seen from the table, over 90% of the LH and FSH activities were recovered in the fractions extracted with 50 or 40% ethanol at pH 7. These fractions from extractions of three batches of glands were pooled for CM-Sephadex chromatography.

Purification of eLH and eFSH
The overall scheme for the purification of eLH and eFSH is shown in Fig. 1. The procedure consisted of: 1) separation of LH and FSH activities by ion exchange chromatography on CM-Sephadex; 2) removal of contaminants by QAE-Sephadex chromatography followed with 3) final purification by gel filtration (if required).

Separation of eFSH from eLH by Ion Exchange Chromatography on CM-Sephadex
The elution profiles and activity recoveries illustrating the separation of LH and FSH activities of pituitary extracts and crude gonadotropin fractions on CM-Sephadex columns are found in Fig. 2. The relative purity of these fractions can be judged by the gel pattern insets in Fig. 1. Table I1 lists the material and activity recoveries at all stages of the purification. LH activity was determined in the horse testis RLA and FSH activity was determined in the chicken testis RLA. Fig. 2A shows the results from a chromatograph of 26.66 g of horse pituitary extract on a column (15 X 11 cm) of CM-Sephadex. Over 70% of the LH activity but only half of the FSH activity were retained. The bound activities were recovered in two fractions. CMS-1D was 2.74 g of crude gonadotropin which contained all of the bound LH activity and half of the bound FSH activity. A 5-fold purification of LH and a 2.6-fold purification of FSH were obtained at this step. CMS-1E was 741 mg of protein which contained the remaining FSH activity, enriched 8-fold over the starting material. Sim- a Values are relative percentage of the given activity per designated fraction compared to the total activity in all fractions. It was not possible to obtain reliable estimates of total activity in the glands from aliquots submitted to total homogenization. ~ ilar results were obtained when a second batch of pituitary extract was chromatographed on CM-Sephadex.
The unabsorbed fractions from these two columns had about the same specific activities in LH and FSH as the fractions extracted with 40% ethanol at pH 5 (rat testis radioligand assay). Therefore, these fractions were combined (43.1 g) and fractionated by ammonium sulfate precipitation. All of the LH and over 90% of the FSH activities were found in the 18.5 g of protein recovered in the 0.55 saturated ammonium sulfate fraction (0.55 SAS). Fig. 2B shows the results of rechromatography of 0.55 SAS on CM-Sephadex (CMS-2). In this case over 90% of the gonadotropic activities were absorbed by the resin and eLH was separated from eFSH. Buffer D eluted 1.58 g of crude eLH (CMS-2D) enriched 8.5-fold for LH activity while having the same FSH specific activity as the starting material (Table 11). Equine FSH slowly bled off the column as development continued with Buffer D. Only a small peak of material was eluted with Buffer F. These were pooled and 694 mg of protein recovered.
The fraction, CMS-2F, was enriched 15-fold in FSH activity over the starting material. Fig. 2C shows the elution profile for the rechromatography of CMS-1D on CM-Sephadex (CMS-3). Buffer D eluted 1.98 g of crude eLH (CMS-3D) which contained 96% of the LH activity. Although only a 1.3-fold purification of eLH was obtained, only 7% of the total FSH activity remained in this fraction. FSH was eluted by continued development with Buffer D as well as by Buffer E. This fraction (CMS-3E) consisted of 269 mg of crude eFSH that was virtually devoid of LH activity and contained 89% of the FSH activity. A 7fold enrichment over starting material was obtained. Fig. 2 0 shows the elution profile for the rechromatography of CMS-2D on CM-Sephadex. Virtually all of gonadotropic activity was recovered in 710 mg of CMS-4D eluted with Buffer D. A 1.4-fold purification of eLH was obtained. There was no change in the FSH specific activity in this fraction. In addition, no FSH was eluted by subsequent development with Buffer D.

Chromatography of Partially Purified eLH on QAE-Sephadex
When crude eLH (CMS-3D) was chromatographed on QAE-Sephadex (Fig. 3), LH activity was found in all fractions. About 9% was found in 136 mg of unabsorbed or weakly absorbed protein which bled off as the column was developed with Buffer I. Buffer I1 eluted 949 mg of crude eLH (QAE-II-LH) containing 70% of the LH activity. The 308 mg of crude eLH (QAE-111-LH) recovered from the peak eluted with Buffer I11 contained 12%. An additional 2% was found in 252 mg of protein eluted with Buffer IV. Equine LH from fractions I (9%) and IV (2%) was difficult to purify. They will, therefore, not be considered further.

Gel Filtration of eLH on Sephacryl S-200 and
Sephadex G-1 00 Fig. 4, top, shows the elution profile for 947 mg of QAE-II-LH on Sephacryl S-200. The cross-hatched area indicates fractions that were pooled to obtain 159 mg of purified eLH (eLH-B). The 255 mg of S2OO-LH recovered from the shaded area of the chromatogram was 89% eLH. Contaminants from the lower M , peak were readily removed from S2OO-LH by gel filtration on Sephadex G-100, as can be seen in Fig. 4, bottom. This additional gel filtration step produced 228 mg of eLH-A.
The potencies of eLH-A and eLH-B in horse testis LH radioligand assay were 6.3 and 5.62 times NIH-LH-S19, respectively. Gel filtration of QAE-111-LH on s-200 yielded a  CMS-4D were applied directly to a Sephadex G-100 column. The fraction recovered from the single major peak emerging from this column was designated eLH-D. It consisted of 347 mg of purified eLH, enriched 1.7-fold over CMS-4D. It contained 25% of the recovered eLH and was 4.9 x NIH-LH-S19 in horse testis LH RLA.   Fig. 6. Most of the material (93.5 mg) was recovered from the single major peak that eluted with a VJVO of 1.48. This fraction, designated eFSH-B, was found to be 156 x NIAMDD-oFSH-13 in chicken testis FSH radioligand assay, a 1.1-fold increase in activity over QAE-FSH. Gel filtration removed a broad M , = 30,000 band and most of the contaminants which appeared as sharp M, = 11,000 and 12,000 bands (compare gel insets in Fig. 1). Similar results were obtained when SQ-FSH was chromatographed on the same column to obtain eFSH-A. The preparation eFSH-C, derived from QAE-Sephadex chromatography of CMS-2F, was of sufficient purity (compare FSH gel C with A and B, Fig. 1) that gel filtration was deemed unnecessary.

Isolation of eLH Subunits
The separation of eLH into its subunits by gel filtration on Sephacryl S-200 following dissociation in guanidine HC1 is shown in Fig. 7. Two peaks emerged from the column. SDSpolyacrylamide gel electrophoresis (see Fig. 7, inset) identified the first peak that eluted with a V,/Vo of 1.48 as the / 3 subunit and the second peak, which eluted with a VJVn of 1.79 as the CY subunit. Equine LH and its subunits were compared in both LH and FSH rat testis radioligand receptor assays. The a subunit was found to have 1.3% the LH activity and 2.7% the FSH activity of intact eLH. The p subunit was slightly more active, having 2.3% the LH and 3.7% the FSH activity of eLH.

Isolation of eFSH Subunits
Gel filtration of eFSH after overnight incubation in guanidine HC1 resulted in a single peak eluting with a VJV0 of 1.67, midway between the positions of the subunits of eLH (see Fig. 7, dashed profile). Fig. 8 shows the elution profile when the subunit fraction was applied to a column of QAE-Sephadex equilibrated with 0.01 M ammonium acetate, pH 5.5. The unabsorbed fraction was identified as the CY subunit by SDS-gel electrophoresis (Fig. 8, inset) while the fraction eluted with 0.4 M ammonium acetate was the /3 subunit. When eFSH and its subunits were tested in rat testis radioligand assays, the subunits were found to have no detectable LH activity. In the FSH assay, eFSHa was 0.7% as active and eFSHP was 0.2% as active as intact eFSH.

Activities of Purified Equine Gonadotropins in Radioligand Assays
The four purified eLH preparations were found to have potencies ranging from 3.84 to 6.3 times NIH-LH-S19 in horse testis radioligand assays (Table 11). In the chicken testis radioligand asasy using '251-eFSH as radioligand, the eLH preparations' potencies ranged from 1.5 to 6.7 times NIAMDD -oFSH -13. The eFSH preparations' potencies ranged from 116 to 156 times NIAMDD-oFSH-13 in this same assay. In the horse testis LH radioligand assays the eFSH preparations were found to have potencies ranging from <0.004 to 0.05 times NIH-LH-S19.

Amino Acid Composition
Preliminary results comparing 24-h hydrolysates of samples from each gonadotropin preparation indicated that the amino acid compositions of the eLH preparations were identical and that those of the eFSH preparations were also identical to each other. Only one preparation of eLH, eFSH, and their subunits was used for more detailed amino acid analysis. These results along with those reported by other workers can be found in Tables 111-VII. Our results are very similar to those of other workers. The most frequent differences are seen in the values for threonine, serine, proline, and halfcystine which are difficult amino acids to quantitate. Tryptophan determinations were performed on eFSH, eFSHa, eFSHP, and eLHP. The absence of tryptophan in eLHp and   FIG . 4 (right). Gel filtration of eLH. Top, elution profile for QAE-11-LH on Sephacryl S-200. The cross-hatched area contained highly purified eLH-B while the shuded area yielded a largely eLH fraction, S2OO-LH, which was subsequently purified on G-100. Bottom, elution profile for S2OO-LH on Sephadex G-100. The bar indicates the portion of the chromatogram pooled to yield eLH-A. Column dimensions and elution conditions can be found under "Methods."

NHz-terminal Amino Acid Determination
For eLH-Phenylalanine, serine, and traces of glycine were consistently observed in NH2-terminal analyses by the dansyl method. Automated sequence analysis of intact eLH showed phenylalanine and serine as the two major PTH derivatives in Cycle 1 (56 and 35% of the total, respectively). Small amounts of glycine (5%) and aspartic acid (4%) were also observed. For eLHp, only serine was found in Cycle 1. However, for eLHa, three PTH derivatives were observed after the first Edman cycle. The principal amino acid was phenylalanine which comprised 80% of the amino acids detected while aspartic acid (9%) and glycine (11%) comprised the remainder. The same pattern was followed on each successive cycle (see Table VIII); one principal PTH derivative appeared which corresponded to the NHz-terminal amino acid sequence for the a subunit of equine glycoprotein hormones (16), along with two minor PTH derivatives consistent with two subpopulations of a subunits, one which began with the aspartic acid at position 3 and the other beginning with the glycine at position 4. This was particularly evident at Cycle 5 where both minor sequences should be threonine. After allowing for carry over from the preceding step, the only minor PTH derivative detected was threonine.
For eFSH-Phenylalanine and aspartic acid (indicating either aspartic acid or asparagine) were the most prominent amino acids detected in the NHz-terminal analysis by the sequence began with the lysine a t position 15, the NH, terminus proposed for eFSHa by Rathnam et al. (2). We have previously noted our inability to confirm their reported sequence (18).

On the Extraction
initial results could not be easily reproduced. Eventually it was found that efficiency of extraction could be increased by stirring the pituitary homogenate in the extraction buffers. This allowed greater activity recoveries but also extracted more than twice as much protein as the original method. The additional inactive protein complicated the purification procedure, requiring the introduction of additional chromatographic steps as in the present report. The highest yield of eFSH reported was obtained after extraction at pH 6 (2) and because of the extreme acid lability of eFSH (4) which we had also observed, the original 0.5 M sodium acetate, pH 5 , buffer was replaced with 0.5 M Tris acetate buffer, pH 7. Since more total protein would be extracted at pH 7, the alcohol content was increased from 40 to 50% to compensate. It has also been found necessary to introduce a preliminary extraction with 75% ethanol, 25% 0.5 M sodium acetate at pH 6 to reduce the amount of inactive protein in subsequent extractions. Bates et al. (6) suggested that inert proteins tended to be denatured by this treatment while gonadotropins were not. Our current extraction procedure extracts about 40% of the amount of   (14).
Not determined.
protein extracted with the Koenig and King procedure (20) applied to similar glands yet provides as good or better solubilization of activity. Consideration of the Separation of eLH from eFSH on CM-Sephadex-In our initial purification method (19) virtually all of the gonadotropic activities were absorbed to CM-Sephadex and all but a few per cent were eluted with Buffers D and E. The results of chromatography of the pituitary extracts prepared in this study differ in that significant amounts of LH and especially of FSH were not retained by the resin. In addition the LH fraction, CMS-lD, contained half of the bound eFSH. The crude FSH fraction, CMS-lE, was essentially devoid of LH activity as before. However, it was heavily contaminated with other proteins. All three problems were eventually solved. When the unabsorbed CMS-1A fraction was applied to a second CM-Sephadex column over 90% of the gonadotropic activities were absorbed. We presume absorption to impurities precluded binding on the first column passage. In addition the LH and FSH activities were well resolved. The LH was in the CMS-2D fraction eluted by Buffer D. The FSH (CMS-2F) began to emerge in the tail of  Gln (9) 8 a Since initial yields tend to vary from experiment to experiment, the total nanomoles of PTH-derivatives detected in Cycle 1 were set equal to 100% and the relative amounts of each PTH-derivative found in that cycle and all subsequent cycles were expressed as percentages of that initial total.
Numbers in parentheses indicate positions of the amino acids in the sequence determined for equine chorionic gonadotropin a subunit (16), Phe-Pro-Asp-Gly-Glu-Phe-Thr-Thr-Gln-. . . .

~_ _
this peak as well as in the following fraction eluted with Buffer F. Since most of the contaminants which plagued the purification of the gonadotropin fraction eluted from the first CM-Sephadex column were already removed, subsequent purification of eLH and eFSH from fractions from CMS-2 was much easier. Rechromatography of CMS-2D may not have been necessary since separation of eLH from eFSH had already occurred. The FSH fraction CMS-2E was found to be relatively pure so that QAE-Sephadex chromatography was sufficient to complete the purification. In contrast, purification of the gonadotropin fractions eluted from CMS-1 was more difficult because of the contaminating proteins. Rechromatography of crude gonadotropin CMS-1D on a smaller column of CM-Sephadex achieved separation of eLH from eFSH. The LH was eluted as a sharp peak with Buffer D. When the peak began to tail off, FSH activity began to protein/ml of packed resin compared to the ratio of 5 mg/ml used in the initial purification method and in subsequent CM-Sephadex chromatography steps during this purification procedure (CMS-2-CMS-4). However, when crude pituitary extract was applied to a CM-Sephadex column at the ratio of protein to resin of 5 mg/ml or 40 mg/ml activity breakthrough was still observed. Thus, this seems to be dependent on interactions between the gonadotropin and contaminating proteins in the extract.
On the QAE-Sephadex Chromatography of eLH-It is difficult to evaluate the utility of QAE-Sephadex in the purification of eLH. It may be useful to remove some contaminants of intermediate size that interfere with the gel filtration step.
On the other hand, if crude gonadotropin is not rechromatographed on CM-Sephadex, but rather, is subjected to QAE-Sephadex chromatography directly, the LH fraction (eluted with 0.125 M ammonium bicarbonate) fails to resolve adequately on an S-200 column. The variable results with anion exchange resins appear to be due to sialic acid heterogeneity. When purified preparations of eLH are fractionated on anion exchangers the fractions eluting with the lowest ionic strength buffers have the lowest sialic acid content while those eluted with higher ionic strength buffers have correspondingly higher sialic acid contents. Therefore, one disadvantage in using QAE-Sephadex has been the separation of eLH into several fractions which must be further purified separately. Another disadvantage is the fact that samples bleed slowly off the column rather than eluting as sharp peaks. For this reason and because better separation of eLH and eFSH could be obtained, volatile buffers were employed in order to simplify recovery of protein from the large volumes of column effluent that were generated.
Use of QAE-Sephadex in the Purification of eFSH-QAE-Sephadex proved to be very useful in the purification of eFSH, possibly because less sialic acid heterogeneity was observed in eFSH preparations. In only one of three columns was a significant amount of FSH activity found in any fraction other than that eluted with 0.2 M ammonium bicarbonate. Moreover, the M , = 28,000 protein found in partially purified eFSH fractions (Fig. 1) which overlaps with eFSH on S-200 columns can be removed by QAE-Sephadex.
Final Purification of eLH by Gel Filtration-The final step in the purification of eLH was gel filtration. The basic characteristics of most chromatograms are seen in Fig. 4, top. The eLH fraction emerged first followed by a second larger peak of inactive contaminants, which in some earlier preparations (extracted at pH 5 ) included eFSH subunits. The shape of the eLH peak depends on the sialic acid heterogeneity of the preparation. Equine LH-B which was recovered from the leading shoulder contained 6.3% sialic acid. S200-LH, recovered from the main peak, contained 5.4% sialic acid. Sialic acid has been demonstrated to alter the apparent molecular weights of glycoproteins as determined by gel filtration experiments (21). S200-LH contained 11% of a low M , contaminant which was readily removed by gel filtration on Sephadex G-100 (Fig. 4, bottom). This contaminant may be free LY subunit since its amino acid composition is very similar to those of purified eLH and eFSH LY subunits, and one of the principal amino acids observed in NH2-terminal analysis was phenylalanine. Characterization at this point is inadequate to make a firm conclusion.
Gel Filtration of eFSH on Sephacryl S-200"Following Equine Gonadotropin Purification QAE-Sephadex chromatography the eFSH preparations were highly purified as indicated by SDS-gel electrophoresis (Fig.  1). The purity of these preparations was confirmed by gel filtration since the eFSH emerged as a single major peak, having only small amounts of contaminants emerging immediately before and after the eFSH fraction.
Biological Assessment of Gonadotropin Purity-We have used radioligand receptor assays to evaluate cross-contamination of our eLH and eFSH preparations. The potencies of eFSH compared with eLH in the horse testis LH/RLA were less than 1% as active as the eLH preparations. When the FSH activities of eLH preparations were determined with the chicken testis RLA they were found to be 3-4.6% as active as eFSH. This is not due to eFSH contamination but is intrinsic activity of eLH in this FSH assay.' Radioimmunoassay of some eLH preparations (19) indicated less than 1% contamination by eFSH. The chicken testis assay overestimates FSH contamination but is more specific than the rat testis assay' and provides a convenient system to follow purification of equine gonadotropins. Chicken testes have several advantages over horse testes. They are easier to obtain, the receptor preparation is easier to prepare, and they have not been as variable in their ability to bind lZ5I-labeled eFSH as the horse testes have been. There is, however, one problem in direct comparisons of potency with reports from other laboratories. That is the low activity of ovine FSH preparations in the chicken assay. For example, with the horse testis preparations using NIAMDD-oFSH-13 as a reference preparation, two eFSH preparations were determined to be 169 and 228 x NIH-FSH-S1 (4). Using the same reference preparation in the chicken testis RLA our eFSH preparations have potencies Licht et al. (3) reported that their eLH preparation was 4.24 x oLH, using a purified oLH preparation that was 2.7 X NIH-LH-S1 by ovarian ascorbic acid depletion assay (e.g. a calculated 11.4 X NIH standard). This is somewhat higher than our preparations that were up to 6.3 X NIH-LH-S19.
Isolation of eLH and eFSH Subunits-Subunits were prepared by methods adapted from Moore and Ward (9) for the isolation of subunits from eCG. The eLH gel filtration on Sephacryl S-200 produced well resolved subunits, suitable for chemical studies. Equine FSH subunits, however, are about the same size (Ref. 22 and Fig. 6). Therefore, QAE-Sephadex was used to separate eFSH subunits after gel filtration to remove guanidine-HCl. The Sephacryl S-200 treatment could possibly be omitted if the dissociation were done with urea. However, we have found some preparations to contain about 10% non-dissociable eFSH. The alpha subunit as obtained here is highly purified. Since intact hormone as well as subunit are retained by the QAE-Sephadex the /3 subunit fraction can be contaminated with a subunit. In our studies this happened with one of three subunit preparations. In this case cross-contamination was readily apparent by SDS gel electrophoresis as well as by RLA.
The results of gel filtration of dissociated eLH and eFSH (Fig. 7) and of SDS-gel electrophoresis (Fig. 8) confirm the observations of Landefeld and McShan (22, 23) that eLHP is larger than eLHa whereas the subunits of eFSH are of similar size. The differences in molecular size of these subunits probably reflect differences in the carbohydrate moieties rather than in the polypeptide chain. Equine LHP is shown to have more carbohydrate than eLHa (23). Equine FSHa has a G. R. Bousfield and D. N. Ward, manuscript in preparation. This manuscript provides an extensive evaluation of eLH and eFSH responses, together with their subunits, in the several test systems employed.
of 1740 to 2340 X NIH-FSH-S1. ~~~ molecular weight slightly higher than that of eLHa on SDS gels (Fig. 8) and elutes from S-200 columns between eLHP and eLHa. All our data to date show that eLHa and eFSHa are identical in amino acid sequence (18). Thus, the size differences must reflect differences in carbohydrate. It is interesting that both a subunits appear as relatively compact dark staining bands on SDS gels whereas both P subunit bands are broader and more diffuse, suggesting greater carbohydrate heterogeneity for the latter. The nature of the carbohydrate moieties of equine gonadotropins present an interesting topic for further study.
Rat testis radioligand receptor assay was used to measure biological activities of the subunits. Equine LH subunits were less than 3% as active as intact eLH. The low activity of eLHa in the FSH/RLA is contrary to the results reported by Agganval et al. (24). According to these authors eLHa retains significant activity (12% that of intact eLH) in the porcine granulosa cell FSH-binding asasy and is as active as eLH in inhibiting cyclic AMP production by seminiferous tubule cells from 18-day-old rats. However, the method by which eLH subunits were obtained was not reported and no chemical data on the purity of the subunits were presented. The residual binding activities of LH subunits have been demonstrated by immunological techniques to be due to the presence of intact hormone in the subunit preparation rather than to intrinsic activity of the subunits themselves (25). We have found that eLH remains intact under conditions of low pH that completely dissociate eFSH or oLH.' It is entirely possible that the high biological activity of eLHa reported (24) was due to undissociated eLH in their subunit preparation.
Studies on the NH2-terminal Amino Acids of Equine Gonadotropins-Our results for the NHz-terminal amino acids of eLH are in agreement with the results of earlier studies (3,26). We have extended these by determining phenylalanine to be the NH, terminus for the a-subunit and serine to be the NH2 terminus of the P subunit, as is the case for several other lutropin species as well as for eCG (27). Our results for the NHz-terminal amino acids of eFSH agree with those of Nuti et al. (28). Licht et al. (3) reported phenylalanine as one NH, terminus for their eFSH preparation, but also reported serine as the other, the same as they had found for eLH. Fujiki et al. (17) reported either aspartic acid or asparagine as the NHz terminus of eFSHP. We determined that asparagine is the NHz-terminal amino acid of eFSHP. The same NHz terminus has been reported for the P-subunit of human, porcine, and ovine FSH (27,29).
For eFSHa we found phenylalanine as the NH2 terminus; Rathnam et al. (2) reported lysine. Furthermore, their sequence lacks the first 14 amino acids, including two halfcystines, that have been reported in the amino acid sequence determination of eCGa (16). Their bioassay results indicate that they prepared highly active hormone. They interpreted this to mean that the NH2-terminal portion of the a subunit is not important for biological activity. However, their NHZterminal determinations were performed on isolated subunits, not on intact hormone. It would have been informative to compare results of an NHz-terminal determination on their intact biologically active eFSH. Except for trace amounts of lysine observed in NHz-terminal analysis by the dansyl method of some eFSH preparations, we have no evidence for a small subpopulation of a-subunit molecules having an NHzterminal lysine. At cycle one during automated sequence analysis of the a-subunits of both eLH and eFSH no lysine was detected, nor at cycle 2 was there leucine detectable as would be required by their proposed sequence. In our equine gonadotropin a-subunit preparations we conclude there is no Equine Gonadot detectable population of a-subunit molecules having an NH2terminal sequence as proposed by Rathnam et al. (2).
In conclusion, we have been able to extract and purify both LH and FSH from horse pituitary glands in high yield. Our method is more cumbersome than that we originally reported (19); however, the yield of highly purified eFSH compares very favorably with those of other laboratories, and the yield of eLH is the highest yet reported. After a two-pass chromatography on CM-Sephadex, separation of eLH from eFSH was reproducibly obtained on CM-Sephadex using the same elution conditions each pass. QAE-Sephadex chromatography as we used it may be unnecessary for the purification of eLH. The use of anion exchangers in the purification of eLH should be restricted as it tends more to separate LH into subfractions than to remove contaminants. In contrast, QAE-Sephadex was very useful in the purification of eFSH. This was probably due to less heterogeneity in the sialic acid content of eFSH.
The hormones produced are highly purified according to SDS-polyacrylamide gel electrophoresis and NHz-terminal amino acid sequence determination. The subunits prepared by methods described herein had little cross-contamination and are suitable for chemical analysis including amino acid sequence determination. Equine LH subunits had greater residual activity, which may reflect their resistance to dissociation compared with that of eFSH.