0013-7227/91/1292-0579$03.00/0 Kndocrinology Copyright ^ 1991 by The Endocrine Society
Vol. 129, No. 2 Printed in U.S.A.
Glycoprotein Hormones Were Always Composed of Subunits—We Just Had to Find Out the Hard Way fessor Albert Neuberger's laboratory at St. Mary's Hospital in London where I participated in studies on the nature of the carbohydrate-protein linkage in ovalbumin and ovomucoid. Upon my return to Berkeley it seemed sensible to perform similar studies on oLH (or ICSH, interstitial cell-stimulating hormone as it was called in Li's laboratory). I was appalled, however, at the few milligrams of oLH that Dr. Li gave me for a year's study as I had been accustomed to working with gram quantities of the egg white proteins. Using proteolytic digestions, gel filtration, and some semiquantitative carbohydrate analyses, I was able to publish a note in 1963 in Biophys Biochim Ada and prepare an abstract for the 1964 Endocrine Society Meeting which concluded that there were two carbohydrate moieties associated with oLH. It was during this period that perhaps the most important observation relating to the chemistry of oLH and subsequently the other glycoprotein hormones was made by C. H. Li and his research assistant at the time, Barbra Starman. They had been examining the molecular weights of several of the purified pituitary hormones available at that time by the technique of sedimentationequillibrium, a procedure which had become possible because of advances in ultracentrifuge technology. When they examined oLH in strongly acidic solutions (pH 1.3) they were surprised to find that the mol wt was no longer 30,000 as in neutral solutions but, instead, about 16,000, or roughly half. These results were published as a very short article that appeared in Nature in 1964 in which determination of the "monomer" mol wt of ICSH (LH) was described. There was no reason to believe at the time that this was anything but a dimer-monomer dissociation phenomenon. Darrell Ward confirmed and extended the observation the following year (1965) in an article in Analytical Biochemistry. A year later (1966) in an abstract presented at the Federation Meetings Darrell Ward presented data on the amino acid sequences associated with two glycopeptides isolated from proteolytic digests of oLH. These were different and he correctly postulated that the "monomer" chains of oLH were not identical. I had reached the same conclusion from examining tryptic digests of performic acid oxidized oLH by two dimensional chromatography-
The glycoprotein hormones of the pituitary gland (LH, FSH, and TSH) each consist of two chemically dissimilar, dissociable, glycosylated protein subunits. One of the subunits (a) is common and identical in primary structure in all three hormones; the other (jS) determines the hormone specificity of the fully active a-fi complex. The chorionic gonadotropins (human CG, equine CG) have a similar chemistry. Over the last 20 yr this basic description has appeared so many times in papers on glycoprotein hormones that the source references are no longer cited, i.e. it is now dogma. This, of course, is very aggravating to those of my generation that played a role in developing the above concepts. Since it may seem odd to our younger colleagues that the subunit nature of the glycoprotein hormones was not forever evident, and since few of us read scientific papers that are older than 5 yr other than our own masterpieces, I would like to briefly cover in this "Remembrance" those key studies that led in a matter of a few years to a radically new understanding of the structure-function relationships among the glycoprotein hormones. My recollections are highly biased, and I apologize in advance to those whose work I do not mention, especially when they are good friends. Ovine LH (oLH) was the first of the pituitary and placental glycoprotein hormones in which the subunit nature as described above was demonstrated. By 1959 two laboratories (P. G. Squire in C. H. Li's laboratory in Berkeley and D. N. Ward in Houston) had isolated oLH in a state of purity such that chemical characterization studies could be undertaken. There was reasonable agreement between the two laboratories with respect to such properties as biological potency, amino acid composition, and mol wt, the latter being 28-30,000. In general, all investigators working with gonadotropins or TSH in this period were plagued by problems of low yields, capricious losses of activity, and evidences of polymorphism such that there was difficulty if not reluctance in performing major chemical studies such as amino acid sequencing. My own work with gonadotropins and oLH in particular began after a year (1961-1962) as a Fellow in ProReceived March 18, 1991. "Remembrance" articles discuss people and events as remembered by the author. The opinion(s) expressed are solely those of the writer and do not reflect the view of the Journal or The Endocrine Society. 579
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 November 2015. at 17:24 For personal use only. No other uses without permission. . All rights reserved.
580
REMEMBRANCE
electrophoresis on paper. If oLH was a monomer of 15,000 daltons it would have a lysine and arginine content such that one would expect to observe a maximum of about 12 tryptic peptides. Over twice that number was observed suggesting that the units were different but each of about the same mol wt. Additionally, I found that performic acid oxidized preparations of oLH were partially insoluble at pH 4.0 and the two fractions had very different amino acid compositions. Such materials, however, while useful for chemical studies, could not be employed for biological work, and I decided to look for a means of obtaining intact subunits. I opted for counter-current distribution (CCD) because C. H. Li's laboratory had previously made extensive use of the technique in studies on MSH, ACTH, and PRL. I had used it as well a few years earlier in the purification of porcine GH. Success would depend on finding a twophase solvent system in which the subunits would differentially partition. Most of the dozen or more organic alcohol-water systems we tried were complete failures in that none of the LH would distribute into the organic phase. One system, however, showed a glimmer of promise. This consisted of 40% ammonium sulfate solution and ethanol and had been tried unsuccessfully by Denis Gospodarowicz a couple of years previously when he worked with me in developing new purification methods for oLH and FSH. By trial and error the system was modified by the addition of dichloroacetic acid to provide an acidic dissociating environment and the introduction of n-propanol in addition to the ethanol as it increased solubility in the organic phase. A limited test of a few transfers with a few milligrams of LH showed great promise and we were ready to try a scaled up experiment with a larger quantity of LH. However, it was now late in 1966 and we had to suspend our research activity in Berkeley to prepare for our move to San Francisco. December was spent packing and January of 1967 was spent moving and unpacking. As with all such moves the new laboratory was chaos. At the beginning of February, with a minimum of functional equipment, we reinitiated our CCD studies of oLH. Our first experiment which involved 10-20 mg LH and only nine transfers was a spectacular success. The LH had distributed itself into two nearly equal fractions, one of which had a very low partition coefficient and was found mainly in the initial aqueous phases, and one which had a very high partition coefficient which had distributed into the organic phases at the end of the train. We soon had two fractions (called C-I and C-II at that time) which were chemically and electrophoretically different, each of minimal biological activity (about 7% that of intact oLH), and upon mixing and incubating together, a significant regeneration of activity (about 20%) occurred. Indeed we had separated and isolated the subunits of
Endo«1991 Vol 129 • No 2
oLH! By early May 1967, Dr. T. S. A. Samy (now at University of Miami School of Medicine) and I submitted a short paper on the preparation and characterization of the subunits of oLH to a well known weekly scientific journal in the US. Needless to say, it was rejected, but happily, was published in another journal several months later (Biochim Biophys Acta 147:175-177, 1967). Earlier in the year Marian Jutisz and his colleagues in Paris published studies in Biochem Biophys Res Commun on the dissociation of oLH in concentrated urea solutions, an approach that was later coupled with ion exchange chromatography and used extensively to dissociate and isolate the human pituitary glycoprotein hormone subunits as well as those of human CG and equine CG. Some of the elegant and now classic studies of John G. Pierce on the chemistry of bovine TSH are necessary to complete this story. While all of the above was taking place, John had decided to ignore the polymorphism of his TSH preparations and proceed with a major effort of determining its amino acid sequence. Despite the convincing evidence of subunits in oLH, similar experiments, including CCD, did not provide such evidence for TSH. On the other hand he found that the amino acid sequences of two of the three glycopeptides of TSH were identical to those of oLH reported by Darrell Ward at the Third International Congress of Endocrinology in 1968, suggesting a close similarity to oLH (J Biol Chem 1969). The big breakthrough came from a serendiptitious set of conditions which, according to John Pierce in his Eli Lilly Lecture in 1971, "allowed the subunits of TSH to more or less fall into our laps." John had found that if TSH were incubated in 1 M propionic acid, lyophilized, and then gel filtered on a column of Sephadex G100 in ammonium bicarbonate, three peaks were eluted which proved to be undissociated TSH, TSH-a, and TSH-b (Federation Meetings 1970). We (Papkoff and Ekblad) later found the same procedure to be applicable to preparing subunits from ovine FSH. John Pierce's progress was now rapid. In comparing the TSH subunits with those of bovine LH prepared by CCD, he showed that one of them {a) was common and interchangeable with respect to generating biological activity and that their tryptic peptides were identical (J Biol Chem 1970,1971). Thus, all of the essential elements of the subunit nature of the glycoprotein hormones had been demonstrated in a matter of 4 yr after the separation and isolation of the oLH subunits in 1967. The reader may be interested in the origin of the a-/? nomenclature for the subunits of the glycoprotein hormones. It was clear by early 1970 that we were headed for a good deal of confusion as already a half dozen investigators were using different designations for subunits of a particular species of glycoprotein hormone. John Pierce and I must have been very concerned by this
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 November 2015. at 17:24 For personal use only. No other uses without permission. . All rights reserved.
REMEMBRANCE because we sent a joint letter dated August 13, 1970 to about 30 investigators in the field throughout the world proposing the a-fi nomenclature. In addition, John reiterated the proposal when he gave his landmark presentation on TSH a couple of weeks later at the Laurentian Hormone Conference. There was an immediate acceptance on the part of everyone, possibly because it made
581
sense, but more likely because nobody had to adopt the designations of someone else at the expense of their own nomenclature. Harold Papkoff Hormone Research Institute San Francisco, California
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 November 2015. at 17:24 For personal use only. No other uses without permission. . All rights reserved.
Vol. 129, No. 2 Printed in U.S.A.
Glycoprotein Hormones Were Always Composed of Subunits—We Just Had to Find Out the Hard Way fessor Albert Neuberger's laboratory at St. Mary's Hospital in London where I participated in studies on the nature of the carbohydrate-protein linkage in ovalbumin and ovomucoid. Upon my return to Berkeley it seemed sensible to perform similar studies on oLH (or ICSH, interstitial cell-stimulating hormone as it was called in Li's laboratory). I was appalled, however, at the few milligrams of oLH that Dr. Li gave me for a year's study as I had been accustomed to working with gram quantities of the egg white proteins. Using proteolytic digestions, gel filtration, and some semiquantitative carbohydrate analyses, I was able to publish a note in 1963 in Biophys Biochim Ada and prepare an abstract for the 1964 Endocrine Society Meeting which concluded that there were two carbohydrate moieties associated with oLH. It was during this period that perhaps the most important observation relating to the chemistry of oLH and subsequently the other glycoprotein hormones was made by C. H. Li and his research assistant at the time, Barbra Starman. They had been examining the molecular weights of several of the purified pituitary hormones available at that time by the technique of sedimentationequillibrium, a procedure which had become possible because of advances in ultracentrifuge technology. When they examined oLH in strongly acidic solutions (pH 1.3) they were surprised to find that the mol wt was no longer 30,000 as in neutral solutions but, instead, about 16,000, or roughly half. These results were published as a very short article that appeared in Nature in 1964 in which determination of the "monomer" mol wt of ICSH (LH) was described. There was no reason to believe at the time that this was anything but a dimer-monomer dissociation phenomenon. Darrell Ward confirmed and extended the observation the following year (1965) in an article in Analytical Biochemistry. A year later (1966) in an abstract presented at the Federation Meetings Darrell Ward presented data on the amino acid sequences associated with two glycopeptides isolated from proteolytic digests of oLH. These were different and he correctly postulated that the "monomer" chains of oLH were not identical. I had reached the same conclusion from examining tryptic digests of performic acid oxidized oLH by two dimensional chromatography-
The glycoprotein hormones of the pituitary gland (LH, FSH, and TSH) each consist of two chemically dissimilar, dissociable, glycosylated protein subunits. One of the subunits (a) is common and identical in primary structure in all three hormones; the other (jS) determines the hormone specificity of the fully active a-fi complex. The chorionic gonadotropins (human CG, equine CG) have a similar chemistry. Over the last 20 yr this basic description has appeared so many times in papers on glycoprotein hormones that the source references are no longer cited, i.e. it is now dogma. This, of course, is very aggravating to those of my generation that played a role in developing the above concepts. Since it may seem odd to our younger colleagues that the subunit nature of the glycoprotein hormones was not forever evident, and since few of us read scientific papers that are older than 5 yr other than our own masterpieces, I would like to briefly cover in this "Remembrance" those key studies that led in a matter of a few years to a radically new understanding of the structure-function relationships among the glycoprotein hormones. My recollections are highly biased, and I apologize in advance to those whose work I do not mention, especially when they are good friends. Ovine LH (oLH) was the first of the pituitary and placental glycoprotein hormones in which the subunit nature as described above was demonstrated. By 1959 two laboratories (P. G. Squire in C. H. Li's laboratory in Berkeley and D. N. Ward in Houston) had isolated oLH in a state of purity such that chemical characterization studies could be undertaken. There was reasonable agreement between the two laboratories with respect to such properties as biological potency, amino acid composition, and mol wt, the latter being 28-30,000. In general, all investigators working with gonadotropins or TSH in this period were plagued by problems of low yields, capricious losses of activity, and evidences of polymorphism such that there was difficulty if not reluctance in performing major chemical studies such as amino acid sequencing. My own work with gonadotropins and oLH in particular began after a year (1961-1962) as a Fellow in ProReceived March 18, 1991. "Remembrance" articles discuss people and events as remembered by the author. The opinion(s) expressed are solely those of the writer and do not reflect the view of the Journal or The Endocrine Society. 579
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 November 2015. at 17:24 For personal use only. No other uses without permission. . All rights reserved.
580
REMEMBRANCE
electrophoresis on paper. If oLH was a monomer of 15,000 daltons it would have a lysine and arginine content such that one would expect to observe a maximum of about 12 tryptic peptides. Over twice that number was observed suggesting that the units were different but each of about the same mol wt. Additionally, I found that performic acid oxidized preparations of oLH were partially insoluble at pH 4.0 and the two fractions had very different amino acid compositions. Such materials, however, while useful for chemical studies, could not be employed for biological work, and I decided to look for a means of obtaining intact subunits. I opted for counter-current distribution (CCD) because C. H. Li's laboratory had previously made extensive use of the technique in studies on MSH, ACTH, and PRL. I had used it as well a few years earlier in the purification of porcine GH. Success would depend on finding a twophase solvent system in which the subunits would differentially partition. Most of the dozen or more organic alcohol-water systems we tried were complete failures in that none of the LH would distribute into the organic phase. One system, however, showed a glimmer of promise. This consisted of 40% ammonium sulfate solution and ethanol and had been tried unsuccessfully by Denis Gospodarowicz a couple of years previously when he worked with me in developing new purification methods for oLH and FSH. By trial and error the system was modified by the addition of dichloroacetic acid to provide an acidic dissociating environment and the introduction of n-propanol in addition to the ethanol as it increased solubility in the organic phase. A limited test of a few transfers with a few milligrams of LH showed great promise and we were ready to try a scaled up experiment with a larger quantity of LH. However, it was now late in 1966 and we had to suspend our research activity in Berkeley to prepare for our move to San Francisco. December was spent packing and January of 1967 was spent moving and unpacking. As with all such moves the new laboratory was chaos. At the beginning of February, with a minimum of functional equipment, we reinitiated our CCD studies of oLH. Our first experiment which involved 10-20 mg LH and only nine transfers was a spectacular success. The LH had distributed itself into two nearly equal fractions, one of which had a very low partition coefficient and was found mainly in the initial aqueous phases, and one which had a very high partition coefficient which had distributed into the organic phases at the end of the train. We soon had two fractions (called C-I and C-II at that time) which were chemically and electrophoretically different, each of minimal biological activity (about 7% that of intact oLH), and upon mixing and incubating together, a significant regeneration of activity (about 20%) occurred. Indeed we had separated and isolated the subunits of
Endo«1991 Vol 129 • No 2
oLH! By early May 1967, Dr. T. S. A. Samy (now at University of Miami School of Medicine) and I submitted a short paper on the preparation and characterization of the subunits of oLH to a well known weekly scientific journal in the US. Needless to say, it was rejected, but happily, was published in another journal several months later (Biochim Biophys Acta 147:175-177, 1967). Earlier in the year Marian Jutisz and his colleagues in Paris published studies in Biochem Biophys Res Commun on the dissociation of oLH in concentrated urea solutions, an approach that was later coupled with ion exchange chromatography and used extensively to dissociate and isolate the human pituitary glycoprotein hormone subunits as well as those of human CG and equine CG. Some of the elegant and now classic studies of John G. Pierce on the chemistry of bovine TSH are necessary to complete this story. While all of the above was taking place, John had decided to ignore the polymorphism of his TSH preparations and proceed with a major effort of determining its amino acid sequence. Despite the convincing evidence of subunits in oLH, similar experiments, including CCD, did not provide such evidence for TSH. On the other hand he found that the amino acid sequences of two of the three glycopeptides of TSH were identical to those of oLH reported by Darrell Ward at the Third International Congress of Endocrinology in 1968, suggesting a close similarity to oLH (J Biol Chem 1969). The big breakthrough came from a serendiptitious set of conditions which, according to John Pierce in his Eli Lilly Lecture in 1971, "allowed the subunits of TSH to more or less fall into our laps." John had found that if TSH were incubated in 1 M propionic acid, lyophilized, and then gel filtered on a column of Sephadex G100 in ammonium bicarbonate, three peaks were eluted which proved to be undissociated TSH, TSH-a, and TSH-b (Federation Meetings 1970). We (Papkoff and Ekblad) later found the same procedure to be applicable to preparing subunits from ovine FSH. John Pierce's progress was now rapid. In comparing the TSH subunits with those of bovine LH prepared by CCD, he showed that one of them {a) was common and interchangeable with respect to generating biological activity and that their tryptic peptides were identical (J Biol Chem 1970,1971). Thus, all of the essential elements of the subunit nature of the glycoprotein hormones had been demonstrated in a matter of 4 yr after the separation and isolation of the oLH subunits in 1967. The reader may be interested in the origin of the a-/? nomenclature for the subunits of the glycoprotein hormones. It was clear by early 1970 that we were headed for a good deal of confusion as already a half dozen investigators were using different designations for subunits of a particular species of glycoprotein hormone. John Pierce and I must have been very concerned by this
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 November 2015. at 17:24 For personal use only. No other uses without permission. . All rights reserved.
REMEMBRANCE because we sent a joint letter dated August 13, 1970 to about 30 investigators in the field throughout the world proposing the a-fi nomenclature. In addition, John reiterated the proposal when he gave his landmark presentation on TSH a couple of weeks later at the Laurentian Hormone Conference. There was an immediate acceptance on the part of everyone, possibly because it made
581
sense, but more likely because nobody had to adopt the designations of someone else at the expense of their own nomenclature. Harold Papkoff Hormone Research Institute San Francisco, California
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 November 2015. at 17:24 For personal use only. No other uses without permission. . All rights reserved.
