The active form of testosterone
Prologue: The events leading up to the discovery of the functional role of dihydrotestosterone in the prostate gland began with the observation made by Wilson and Loeb that radioactive testosterone is bound to nuclear protein in the duck preen gland. Wilson assumed that the action of testosterone on the prostate would also occur through a nuclear binding reaction involving a receptor protein possibly in a phosphorylated state. I arrived in Dallas in mid-October 1966 to begin a period of postdoctoral and residency training. In my initial experimental work I was to attempt to isolate 3H-testosterone labeled phosphoprotein from nuclei of rat prostate. The sequence of experiments leading up to the identification of dihydrotestosterone in prostatic nuclei and its binding to nuclear androgen receptor is outlined below.
Thursday, October 20, 1966 (page 5): This was the first attempt to isolate radioactive nuclear phosphoprotein from prostate. 3H-testosterone was injected into castrated, functionally hepatectomized rats and differential centrifugation was used to recover nuclei and chromatin from prostatic tissue. Chromatin was incubated with deoxyribonuclease and further centrifugation yielded a 33,000 x g pellet and supernatant. The supernatant fraction was then precipitated by dilution to reduce the concentration of non-phosphorylated protein in solution. Consistent with this, more radioactivity was recovered in the supernatant fraction after a 78,000 x g spin, but inexplicably only 10% of the radioactivity in the chromatin fraction was recovered in the 78,000 x g supernatant and pellet fractions.
Method for isolating nucleoprotein
Monday, October 24, 1966 (page 11): The experiment was repeated and again less than 10% of the radioactivity present in the 3H-labeled chromatin fraction was recovered in the final 78,000 x g fractions.
Thursday, October 27, 1966 (page 15): The experiment was repeated a third time and again the recovery of 3H-testosterone was very low. A consistent pattern of a loss of radioactivity during incubation of chromatin with deoxyribonuclease became apparent. Over 80% of the radioactivity disappeared during this step (page 14).
Undated (page 21): A control experiment was done in the wake of the unsuccessful third experiment to investigate the possibility of an adverse reaction between deoxyribonuclease and 3H-testosterone. The results showed quite clearly that there was an exponential decrease in the amount of radioactivity, such that 90% of it disappeared during four hours of incubation at 37°C (page 23).
Undated (page 27): Since deoxyribonuclease itself would not be expected to degrade 3H-testosterone, it was assumed that the disappearance of radioactivity might be due to the use of cellulose nitrate tubes in which the incubations were carried out. A comparison of the effect of incubating 3H-testosterone in glass and cellulose nitrate tubes revealed no loss of radioactivity in the glass tubes, while a striking time dependent loss was observed in the cellulose nitrate tubes (pages 27 and 29).
Counting of nuclei recovered from prostate using a hemocytometer was started at this time as a means of estimating recovery (page 31).
Undated (page 41): Conditions for optimizing the digestion of chromatin by DNase were established (page 42).
Tuesday, November 8, 1966 (page 43): In this experiment, the final 78,000 x g supernatant containing 3H-labeled nuclear protein was lyophilized, resuspended and then analyzed by the new method of gel exclusion chromatography. A 1 cm x 27 cm column of Sephadex G-25 coarse was used for this purpose. A small peak of radioactivity eluted in the fractions containing proteins (page 48).
This was the first indication that gel exclusion chromatography might be used to isolate androgen binding protein from prostatic nuclei.
Monday, November 14, 1966, (page 49): Gel exclusion chromatography was again used in an attempt to demonstrate nuclear binding of radioactivity derived from 3H-testosterone. The results of this experiment were similar to the last in that a small amount of radioactivity was bound to protein recovered in the 78,000 x g fraction. The majority of the 3H-label was retained by the column of Sephadex giving rise to a large peak of free radioactivity (page 54).
Monday, November 21, 1966 (page 55): This was another in the series of experiments using deoxyribonuclease to release nuclear protein from chromatin. The recovery of radioactivity in the 78,000 x g supernatant containing nuclear protein was markedly greater in this experiment (page 58, sample 11).
Part of this supernatant was immediately analyzed by gel exclusion chromatography as before. The remainder was divided into aliquots which were then lyophilized prior to further analysis. The results were again encouraging showing a small amount of radioactivity appearing in the same fractions as the excluded protein followed by a large peak of free radioactivity which was retained by the column of Sephadex G-25 and presumed to be unmetabolized isotope (pages 63 and 67).
A repeat chromatography of the excluded fractions reproduced the peak of bound radioactivity and suggested that the dissociation of 3H-testosterone was relatively slow (page 71).
Thursday, December 1, 1966 (page 73): Nuclei were again labeled in vivo by injecting 3H-testosterone. The nuclear fraction was divided into three equal parts, each of which was frozen and then thawed. This time, no deoxyribonuclease was used and instead the thawed nuclei were sonicated and then extracted with buffer containing either 0.014 molar NaCl or a 10-fold higher concentration of 0.14 molar NaCl. The resultant solution of chromatin was next centrifuged at 78,000 x g and the supernatant, containing about 60% of the radioactivity, was applied to a column of Sephadex-G25. Although the radioactivity eluted from the column was resolved into peaks corresponding to bound and free fractions, the overall recovery of radioactivity was poor (page 77).
This initial experiment suggested that improved extraction of the sonicated nuclear fraction was needed since almost 40% of the radioactivity remained associated with the 78,000 x g pellet (page 75).
In the second part of the experiment, sonicated chromatin was dissolved in buffer containing 0.14 molar NaCl, and the resulting solution was centrifuged at 78,000 x g. Gel-exclusion chromatography of the supernatant yielded protein bound and free peaks of radioactivity with a much improved 70% recovery of input radioactivity (pages 80 and 82).
The 78,000 x g pellet which retained about one-third of the total radioactivity in the original nuclear fraction was dissolved in 1 molar NaCl and the solution centrifuged at 33,000 x g. Chromatography of the supernatant resulted in clearly defined peaks of bound and free radioactivity with a 75% recovery of radioactivity from the column. (pages 84 and 85).
This was a critical experiment because it suggested that the nuclear androgen-binding protein could be solubilized more completely in the presence of a high concentration of NaCl.
In the third part of the experiment, sonicated chromatin was dissolved directly in buffer containing 0.5 molar NaCl and the solution was centrifuged at low speed to remove membranous debris. The supernatant fraction was chromatographed on a column of Sephadex G-25 leading to the recovery of two large peaks of radioactivity representing the bound and free fractions. (page 93).
This was the best result obtained up to that point and indicated that nuclear androgen binding protein could be isolated without the use of deoxyribonuclease.
A control experiment done as part of this series suggested that the incubation of dissolved chromatin with exogenous 3H-testosterone for a prolonged period of time would result in labeling of the chromatin associated androgen-binding protein (page 91).
Tuesday, December 20, 1966 (page 99): Taking advantage of the greater solubility of nuclear androgen binding protein in high salt, a simplified extraction procedure was tried. The basic steps were freezing and thawing of 3H-labeled nuclei, sonication of the nuclear fraction and then titration of the NaCl concentration in the suspension of chromatin up to 0.5 molar. Centrifugation of the solution at 17,000 x g yielded a clear supernatant of high specific activity (page 102, aliquot 7).
When this was chromatographed on a column of Sephadex G25, a perfect separation of bound and free radioactivity was obtained. The recovery of radioactivity applied to the column was almost 70% and of that 41% was in the bound form (page 104).
Column chromatography with polyacrylamide-10 produced a less satisfactory result (page 107) as did the omission of the sonication step (page 109).
Thursday, January 12, 1967 (page 111): With reproducible separation of nuclear radioactivity into bound and free fractions, the first of several experiments was undertaken to confirm the chemical integrity of the radioactive isotope which was presumed to be 3H-testosterone. Fractions eluted from the Sephadex G-25 gel were pooled and extracted with chloroform: methanol (Folch method) and the resultant sample was analyzed by thin-layer chromatography on a plate coated with silica gel H. After chromatography, the coating of silica was divided into 10 horizontal bands which were scraped individually and transferred to scintillation counting vials. A puzzling result was obtained in that the majority of the radioactivity was recovered between 4.5 cm and 6.5 cm from the origin well ahead of the testosterone standard which migrated at a distance of only 3.5 cm from the origin (pages 115 and 116).
This curious finding made it necessary to confirm the identity of the radioactive steroid in the nucleus in later experiments.
Thursday, January 12, 1967 (page 117) The second experiment started on this day was an examination of the effects of long periods of freezing on the stability of the testosterone-nucleoprotein complex. About 5 x 108 labeled nuclei were recovered from prostatic tissue and divided into 5 equal parts. One part was taken for immediate analysis while the remaining ones were frozen in super-cooled alcohol for periods of 1, 4 and 8 days. After thawing, all nuclear preparations were subjected to the same treatment as the unfrozen zero-time control. Each sample was sonicated and chromatin dissolved in a buffer containing NaCl at a concentration of 0.5 molar. The resultant solution was centrifuged at 17,000 x g yielding a supernatant which was then passed through a column of Sephadex G-25. The experimental procedure worked well giving rise to a high 72-98% recovery of radioactivity with each chromatography. No effect of freezing was observed on the distribution of radioactivity between the bound and the free peaks (pages 122, 128, 134 & 140).
A mean 55% of the radioactivity was associated with the eluted fractions containing nuclear protein (page 139).
This experiment defined the apparent ideal conditions for the reproducible recovery of 3H-labeled nuclear protein and indicated that such protein could be stored at low temperature without any loss of binding activity (page 142).
Monday, January 23, 1967 (page 143): In returning to the matter of the anomalous migration of radioactive steroid extracted from prostatic nuclei (page 116), the next group of experiments was undertaken “to study the composition of radioactive peaks with regards to steroids using thin-layer chromatography”. In the first experiment of this series, a solution of 3H-labeled chromatin was chromatographed on a column of Sephadex-G25 to compare the elution profiles of radioactivity, protein, DNA and RNA more precisely. Most of the radioactivity was resolved into two peaks with the first one coinciding with the excluded peak of nuclear protein (page 149).
In the second experiment, the pellet and supernatant resulting from the centrifugation of sonicated 3H-labeled nuclei at 17,000 x g were extracted with chloroform-methanol and analyzed by thin-layer chromatography. In the third experiment, the 3H-labeled supernatant fraction was passed through a column of Sephadex-G25 to separate bound and free radioactivity. Representative fractions from each peak (page 157)
were pooled into two samples which were then extracted and analyzed by thin-layer chromatography. Most of the radioactivity recovered from the nuclear pellet and supernatant, and the Sephadex-derived peaks, migrated ahead of the testosterone standard on the thin-layer plates (page 158).
This discrepancy between the migration of the testosterone standard and the 3H-steroid recovered from nuclei was the first indication of the presence of a metabolic derivative of testosterone.
Tuesday, January 31, 1967 (page 159): Another effort was made to determine the optimum conditions of further recovery of nucleoprotein in a soluble form. In this experiment the concentration of sodium chloride used in the extraction of buffer varied between .05M and 2.50M. A 17,000 x g spin yielded supernatant and pellet fractions which were analyzed for radioactivity, protein, RNA and DNA. Maximum solubility of radioactivity and protein occurred at a concentration of NaCl of 0.6M or greater (page 168).
After these results were obtained, the extraction of protein from prostatic nuclei was routinely carried out using buffer containing 0.6M NaCl.
Monday, February 13, 1967 (page 171): To further refine the recovery of 3H-labeled nuclear protein, the uptake of radioactivity into the cytoplasmic and nuclear compartments of prostatic tissue was studied in more detail. Identification of the radioactive steroids in the cytoplasmic and nuclear fractions was undertaken with both thin-layer chromatography and gas-liquid chromatography. Prostatic tissue was recovered from castrated, functionally hepatectomized rats at 60 and 120 minutes after injection of testosterone-3H. At each time point, nuclei were prepared from the tissue and sonicated. Part of the resultant suspension was extracted directly with chloroform:methanol while the remainder was dissolved in buffer containing 0.5M NaCl and centrifuged at 17,000 x g. The supernatant and pellet fractions were then extracted with chloroform:methanol. Longer thin-layer chromatography plates were used this time to obtain better separation of the extracted 3H-steroids. On this occasion, a list of the possible metabolites of testosterone was drawn up showing dihydrotestosterone as one of the compounds (page 181).
The radioactivity taken up by nuclei reached a maximum level about 1 hour after injection of 3H-testosterone (page 184)
and was resolved into two distinct peaks by thin-layer chromatography (page 185).
In the cytoplasm, the majority of the radioactivity migrated with the testosterone standard. However, the reverse was true in the nuclear fractions with the major peak of radioactivity migrating ahead of the testosterone standard.
When the same samples were analyzed by gas-liquid chromatography, the retention times for the major peaks of radioactivity corresponded to testosterone and dihydrotestosterone (page 187).
The results of this analysis generated a great deal of excitement owing to the identification of dihydrotestosterone as the principle metabolite of testosterone in prostatic tissue, and especially in the nuclear compartment. At this point there was some debate over whether to pursue further studies on the metabolism of testosterone or to concentrate on the isolation of the nuclear androgen-binding protein. It was decided to repeat the time course study of the uptake of testosterone-3H into nuclei and the formation of metabolites.
Monday, February 27, 1967 (page 211): When the experiment was repeated, the uptake results (page 214), and the results of the analyses by thin-layer chromatography (page 221) and gas-liquid chromatography (page 225) were virtually identical to the previous findings. Thus, the presence of dihydrotestosterone in nuclei of prostatic cells was confirmed.
Friday, March 3, 1967 (page 231): With the discovery of the high concentration of dihydrotestosterone in prostatic nuclei, it was decided to test for the presence of a nuclear form 5a-reductase, the enzyme that converts testosterone to dihydrotestosterone. Prostatic nuclei were incubated with a NADPH2-generating system together with 3H- testosterone (page233).
There was no apparent metabolism of radioactive testosterone in the control incubation mixture of nuclei and 3H-steroid. When examined by thin-layer chromatography, most of the radioactivity recovered from this sample corresponded to the testosterone standard (page 236, experiment 9-0, TLC bands 9 and 10).
In contrast, in the incubation mixture supplemented with co-factors, a small percentage of the 3H-testosterone was metabolized to dihydrotestosterone (page 36, experiment 10-0, bands 12 and 13). These preliminary results were consistent with the idea that prostatic nuclei are capable of forming dihydrotestosterone, thus accounting for its presence in the nuclear compartment. This study became the prototype for a number of similar experiments that were to follow in which the in vitro conversion of testosterone to dihydrotestosterone was examined in prostate and other tissues in much more detail.
Monday, March 13, 1967 (page 241): The next experiment in this series was a logical extension of the studies initiated on Monday, February 13, 1967 (page 171) and Monday, February 27, 1967 (page 211) in which the nuclear localization of dihydrotestosterone was demonstrated and then confirmed. The next step was to establish the actual composition of the radioactive peaks obtained when 3H-labeled nucleoprotein was chromatographed on Sephadex G-25. Prostatic nuclei were recovered from functionally hepatectomized rats injected with 3H-testosterone. After sonication, the nuclear fraction was divided into three equal parts each of which was extracted with a different concentration of NaCl, either 0.15M, 0.60M, or 1.00M (page 243). Centrifugation of the solutions yielded three supernatant fractions which were analyzed by gel-exclusion chromatography (page 243).
The radioactivity associated with the soluble nucleoprotein was resolved into two peaks, one excluded from the column, the other retained (page 248).
The aqueous fractions containing the radioactivity were extracted with chloroform-methanol and then analyzed by thin-layer chromatography. The results indicated that the principle 3H-androgen associated with soluble nucleoprotein was dihydrotestosterone whereas the unbound 3H-androgen was chiefly testosterone (page 253).
These findings were clearly significant for two reasons. Not only did they validate the use of gel-exclusion chromatography for fractionating prostatic nucleoprotein, but also they were the first to demonstrate the presence of a nuclear androgen receptor with a high affinity for dihydrotestosterone.
Tuesday, April 11, 1967 (page 261): The capacity of prostatic nuclei to metabolize testosterone was again studied with more careful attention given to the co-factor requirements (page 262).
Only in the incubation mixture containing prostatic nuclei, 3H-testosterone, glucose-6-phosphate, glucose-6-phosphate dehydrogenase and NADP+ was there any appreciable metabolism of the radioactive testosterone. When the metabolic products were analyzed by thin-layer chromatography, it was found that more than 50% of the 3H-androgen recovered was in the form of dihydrotestosterone while 34% remained as unmetabolized testosterone. Almost identical figures of 58% and 26% were obtained when the analysis was carried out by gas-liquid chromatography (supplement to page 262).
Supplement to page 262
Monday, May 1, 1967 (book II, page 21): In another of the series of similar experiments to characterize the optimal conditions for the in vitro conversion of testosterone to dihydrotestosterone, this time-course experiment conducted at the beginning of May proved very helpful in an unexpected way. The yield of radioactive dihydrotestosterone was unusually high and some highly purified material was set aside for recrystalization analysis (book II, page 26).
Book II, page 26
Friday, May 12, 1967 (book II, page 71): A series of recrystalizations were started at this time using the solvents methanol-H2O, acetone-H20, ether-hexane, ethylacetate-cyclohexane and benzene-heptane. Crystals of constant specific activity were obtained with each of the five solvents, findings that were consistent with the previous identification of dihydrotestosterone by thin-layer and gas-liquid chromatography. A similar recrystalization procedure confirmed the identity of 3H-androstanediol as the metabolic derivative 3H-dihydrotestosterone (book II, page 9).
Book II, page 9
The results of the final recrystalizations became available towards the end of July and left little doubt that the prostate contained enzymes very active in converting testosterone to dihydrotestosterone, and the latter to androstanediol.
Epilogue: By August enough data had accumulated to support the writing of a report describing the enzymatic conversion of testosterone to dihydrotestosterone in vivo and
in vitro. The third draft of the paper was finished by the end of October and distributed locally for comment. On the suggestion of Dr. M.D. Siperstein, more emphasis was given in the discussion to the possible role of dihydrotestosterone as the active androgen. Although the ability of prostatic tissue to form dihydrotestosterone enzymatically from testosterone had been demonstrated previously by Pearlman, Farnsworth and Shimazaki, the experimental data now in hand seemed strong enough to justify the conclusion that dihydrotestosterone is the active form of testosterone in peripheral tissues. Nuclear association of testosterone 5a-reductase, intranuclear abundance of dihydrotestosterone and retention of dihydrotestosterone exclusively by target tissues for testosterone were used as supporting evidence. The report was submitted for publication in November and notice of its acceptance was received two months later in mid-January. With completion of this paper, the writing of a second one was started almost immediately. The next dealt with the intranuclear binding of testosterone and dihydrotestosterone and contributed unique information about the nuclear hormone receptor site for dihydrotestosterone. The papers subsequently appeared in the 1968 April and November issues of the Journal of Biological Chemistry.
Owing to the preservation of the original laboratory notebooks, it is possible to produce a historical account of the experiments performed and the motives behind them. What is more difficult to recapture and put into words is the interplay of all the creative forces that made for a very dynamic working environment. To draw attention to a few of these only, the Wilson Laboratory was superbly equipped and the professional technical staff, Mr. G. Crowley and Mrs. J. Walker, were outstanding. Progress on experiments was reviewed in stimulating weekly discussions with Wilson and from these, plans for new experiments were formulated. The weekly meeting of the Division of Endocrinology provided another vibrant setting for the presentation and discussion of experimental data. Many people regularly contributed expertise and knowledge through such formal and informal exchanges. In fact, it is hard to imagine a more impressive concentration of scientific and academic resources available for medical research at that time. This was a reflection of the growing strength of the Southwestern Medical School in the mid 1960’s.