The kinetics of formation and biological activity of native and hybrid molecules of human follicle-stimulating hormone.

Abstract A rat testes tubule tissue receptor assay was utilized to study the kinetics of combination of the hormone-specific β subunit of human follicle-stimulating hormone (hFSH) with the α subunits of hFSH, human luteinizing hormone (hLH), bovine luteinizing hormone (bLH), and human chorionic gonadotropin (hCG). In this competitive protein-binding assay, 125I-hFSH is the radioligand and FSH activity is measured by the ability of the test substance to inhibit uptake of the radioligand by the rat testes tubule receptors. Rate data for combination of subunits were analyzed for quality of fit to first or second order integrated rate equations by nonlinear regression analysis. The time course for the appearance of FSH activity upon incubation of the various subunit combinations was neither clearly first or second order, although each simple mechanism could be fitted to the experimental data. We conclude that at least one kinetically distinguishable state follows formation of the initial α/β subunit complex and that this complex has a binding affinity for its receptors in rat testes tubules which is much less than those of subsequent states. Hybrid molecules of bLH-α/hFSH-β, hLH-α/hFSH-β, and hCG-α/hFSH-β were formed at differing rates, but each was more rapid than the combination always gous hFSH-α/hFSH-β pair. Maximum combination always occurred by 1440 min of incubation. Hybrid molecules formed after this time were tested for FSH activity by the in vitro rat testes tubule receptor assay and results compared with FSH activity as measured by a classical in vivo bioassay. The in vitro activity of hFSH-α/hFSH-β, hLH-α/hFSH-β and hCG-α/hFSH-β hybrid molecules was from 2 to 4 times greater than their activity as measured in vivo. The in vitro activity of these hybrids was similar to that of intact hFSH, but the activity of the hybrid molecules in vivo was less than that of intact hFSH. Of particular interest were the properties of the hybrid molecule formed between bLH-α/hFSH-β. This combinant had an in vitro activity which was 2.8 times greater than that of purified hFSH. The activity of the bLH-α/hFSH-β hybrid as measured by the in vivo bioassay, however, was only 0.14 times the activity of purified hFSH. The ratio of activity in vitro to in vivo for this artificial FSH molecule was 19.8. It is suggested that the bLH-α/hFSH-β hybrid has a greater affinity than native hFSH for receptors in rat testes tubules, but that expression of in vivo biological activity is depressed probably because of factors related to its more rapid clearance from the circulation.

ever, was only 0.14 times the activity of purified hFSH. The ratio of activity in vitro to in vivo for this artificial FSH molecule was 19.8. It is suggested that the bLH-a/hFSH-fl hybrid has a greater affinity than native hFSH for receptors in rat testes tubules, but that expression of in uiuo biological activity is depressed probably because of factors related to its more rapid clearance from the circulation. Fituitary and placental glycoprotein hormones of different types and from several sources have been shown to be composed of two dissimilar subunits, designated as the (Y (common) and fl (hormone-specific) subunits (1). It has been possible to form hybrid molecules from various combinations of these subunits, with the qualitative nature of the Q + /3 hybrid molecule being dependent upon the source of the /3 subunit.
For example, a molecule formed from the cx subunit of LH1 and the /3 subunit of TSH will have TSH biological activity (1). We have previously reported initial studies on the formation and biological activity of native and hybrid molecules of LH and hCG (2, 3). The rate of combination of the (Y and /3 subunits of LH from various species was followed through use of a carefully validated rat testes Leydig cell receptor assay for LH (a), and the properties of the hormonereceptor interaction have also been studied in detail (3,4). We have demonstrated, on the basis of biological studies, that human FSH also consists of two dissimilar subunits (5,6) and have succeeded in achieving a separation and purification of the hFSH a! and fi subunits (7). The details of the interaction of hFSH with specific receptors in rat testes tubules have been described earlier in this journal (8), and application of the tissue receptor system to quantitation of human and animal pituit.ary FSH has also been reported (9,10).
In this report, we describe application of the rat testes tubule receptor assay to the study of the kinetics of combination of the hormone specific /3 subunit of hFSH with the OL subunits of hFSH, hLH, bLH, and hCG. The newly formed FSH molecules are compared with respect to their FSH activity in the rat testes tubule receptor assay and in a hormone-specific in tivo bioassay.
In this function, X represents the amount of FSH formed after the time, t, in terms of the adjustable parameters, X,, the limiting amount of FSH formed after t = m , and k,, the first order rate constant expressed in hours-l.

RESULTS
Kinetic Analysis-Seven data sets were obtained, four of them for the combination of the a and p subunits of hFSH and one each for the combination of the hFSH-/3 subunit with the (Y subunits of hLH, bLH, and hCG. Each data set was tested for fit The results depicted in the lower panel can be related to Fig. 1B. The curve depicted in the upper pane2 for combination of hCG-a! + hFSH-@ subunits were obtained in a similar manner, to each of the trial functions as given in Equations 1, 2, and 3. In the case of Equation 3, trials with p fixed at values corresponding to rxperimental conditions and with p free were also made. Figs. 2 and 3 summarize the data obtained in the combination experiments.
The curves through the experimental points were calculated from Equation 2, the second order rate function.
Our fitting trials show that t.he first and second order equations gave fits of similar quality.
In all cases, whether first or second order trials, small systematic deviations between predicted and &- a Range of the 95% confidence limit)s. b Est,imated from per cent-l hr-l by assuming an average molecular weight for each subunit of 15,000.
c Rate constant for combination of hLH subunits calculated as described in (2). served values were seen during t'he first 2 hours of the reaction. Because the question of the apparent rate equation cannot be regarded as settled, Table II presents bot'h first and second order rate constants obtained from the fitting trials.
The variat,ion of the cyo $0 ratio in the hFSH-cr/hFSH$ combination reaction did not significantly affect the quality of fit, and t'hese observations also indicate that simple first or second order kinetics do not apply.
In Vitro and in Vivo Activity of Native and Hybrid Molecules of FSH-The various combinations of subunits studied above were also examined with respect to relative activity in the in vitro rat testes tubule receptor assay (9) and in the FSH-specific in vivo hCG-augmented ovarian weight gain assay (14) ( Table 1). Incubation of different subunit combinations was allowed to proceed for 24 hours, since kinetic studies had indicated maximal combination of hybrid molecules had occurred by this time (Figs.  I, 2, and 3). In the receptor assay, relative potencies were determined by comparison of the 1251-hFSH uptake inhibition activity of the hybrid combinants with that of an accepted human pituitary FSII reference preparation, LER-907, a.ssigned a potency of 20 i.u. per mg (25). The hybrid combinants of Q? subunits with hFSH-P subunit gave response curves parallel to that of intact hFSH (Fig. 1) so that standard parallel line assay statistics could be utilized for est,imation of biological activity, Parallelism was also observed when the hybrid molecules were compared with intact hFSH in the in vivo bioassay.
The hybrid molecules resulting from combination of hLH-a + hFSH-P and hCG-a + hFSH-p had in vitro activities (receptor assay) at least as great as that of undissociated hFSH, while in each instance, the irz vitro activity of the hybrid molecules was significantly less than that of undissociated hFSH (Table I). This discrepancy was even more marked with the combination bLH-a + BFSH-p.
Here the in vitro act.ivity of the hybrid was 7081 i.u. per mg or 278y0 that of undissociated human FSH.
In contrast, the in viva activity of the hybrid was 510 i.u. per mg. The ratio of activity in vitro t'o in vivo ranged from 2 to 4 for all comparisons, with the exception of that for the bLH-a + hFSH-P hybrid, where the ratio was 19.8. The significance of these results will be considered under YXscussion." The activity of recombined hFSH-a + hFSH-P was relatively low in each system (Table I). In terms of the in vivo response this combinant had an activity of 510 ix. per mg, which was 14y0 of the activity of intact hFSH.
However, the isolated hFSH-a subunit had an in tivo activity of less than 10 i.u. per mg (0.3a/, the activity of pure hFSH) and the isolated hFSH-P subunit had an in vivo activity of 39 i.u. per mg (1.1 y0 the activity of pure hFSB) (Ta-ble I), Assuming no increase in FSH activity resulting from combination of the subunits to form active a + /3 dimers, the in vivo biological activity of a 1: 1 mixture of the subunits would be the mean value of the activity of the individual subunits, or 26 i.u. per mg. The observed potency of 510 i.u. per mg, therefore, represented a highly significant al-fold augmentation of FSH activity.
By similar reasoning, the activity of hFSH-a! + hFSH-fl as measured by the in vitro assay, 862 i.u. per mg, represents a highly significant 25.fold increase in activity over what would be expected if no combination of subunits occurred. It should be noted that the in vivo activity of the hCG-cr + hFSH-/3 hybrid relative to intact hFSH, was relatively high, about 60%. Therefore, the low recovery of activity in the hFSH-Lu + hFSH-fl combinant would seem to reside in a defect in the hFSH-cr, rather than with the hFSH-/3 subunit. DISCUSSION We have recently reported studies on the kinetics of formation and biological activity of native and hybrid molecules of LH and hCG, utilizing as a tool for assessment of recombination a rat testes Leydig cell receptor assay for LH (2,3).
The rates of dissociation and combination of subunits of hLH and hCG have also been studied by application of a fluorescence probe, anilinonaphthalene sulfonate, to follow the course of the reaction (22). There is no information, however, on similar studies with FSH. We have reported a procedure for separation and isolation of subunits of hFSH (7) and have developed a rat t&es tubule receptor assay suitable for quantitation of FSH biological activity (9). It has been previously reported that the fluorescent probe anilinonaphthalene sulfonic acid does not bind to hFSH (26) but even if binding were observed, it would remain to be demonstrated that enhanced anilinonaphthalenesulfonic acid fluorescence is correlated with biological activity.
Thus, the receptor assay for FSH (9) seems ideally suited for studies of kinetics of formation of native and hybrid molecules of FSH, as well as for comparisons of receptor binding activity with in Z&JO biological activity.
The general mechanism suggested in our previous communication (2) and represented by Equation 4 provides a basis for discussion of the FSH subunit combination reaction. a+P &x,&x*%...& X" (4) The model states that a bimolecular reaction between the subunits leads to formation of an initial complex, X1, which is formed from 1 molecule each of (Y and /3 subunit. Yroceeding to the right, the formation of X1 is followed by formation of Xz, X8, and so forth, until the "final" species, X,, results. The species Xl. . . X, are shown as reversibly related, with the'implication that equilibrium among them exists. The degree to which this can occur depends upon the relative magnitudes of the various rate constants.
The apparent kinetics to be observed in such a model will also depend upon the relative magnitudes of the rate constants, but at a sufficiently low subunit concentration, the reaction would necessarily appear as second order because of the first step in the sequence.
In the present study, the analysis for extent of reaction is based upon the ability of FSH formed in the reaction mixture after various periods of incubation to compete with Y-hFSH for FSHspecific binding sites in the rat testes tubule homogenate.
In terms of the general mechanism (Equation 4), it is not known at the outset whether any of the species beyond X1 are able to bind to the hormone receptor sites, or how many such species there may be. If second order kinetics were clearly established, it could indicate that (a) there is no species beyond X1, (b) limiting low concentration conditions for a and /3 existed during the rate measurements, or (c) active species may exist beyond X1 but the bimolecular step is rate-limiting. If first order kinetics were seen, then the existence of at least one species beyond X1 would be established.
Our kinetic measurements show that although the combination of LY and /3 subunits of hFSH gives a satisfactory fit to the second order integrated rate equation (Equation 2)) a fit of similar quality is obtained with the first order equation (Table II). To provide a more stringent test, the ratio of CY to /3 subunit was varied; the rate data however proved insensitive to these variations in reactant concentration ratios (p in Equation 3). The quality of fit to the first order law was also substantially unaffected, keeping in mind that small systematic deviations are seen in our fits to both kinetic laws. It does not seem likely that more data would alter either the systematic deviations in fit or the insensitivity to initial subunit concentration ratios. The evidence then, appears to rule out a single step bimolecular formation of active hormone, that is, only species X1 being formed.
The lack of response to variations in the (;~:/3 ratio supports the view that at least one species beyond X1 does esist, and that X1 itself does not compete for receptor sites. The last conclusion follows since if X1 were active in the assay, second order kinetics for its appearance would be expected. Conclusions of a related nature were reached by Ingham et al. (22) for the recombination of hLH a and /3 subunits, as followed by use of fluorescent probe techniques.
We have included in Table II data on the kinetics of combination of the isolated subunits of hLH as determined using a radioligand receptor assay for LH (3). As can be seen, the rate constant for the combination of the hLH subunits is an order of magnitude higher than observed for any of the combinations involving the 0 subunit of hFSH.
The rate equations (Equations 1,2, and 3) were utilized because they represent mechanisms easily accommodated within the framework of the general reaction model (Equation 4). The failure of the hFSH-or + hFSH-/3 combination data to respond as expected to variations in the CUO:@O ratio, together with the systematic deviations observed at times early in the course of the reaction, suggest that neither simple expression will suffice, and that further analysis will be needed to permit a more precise description of the kinetic events. Nevertheless, it is clearly shown in the present study that the subunit combination reactions do lead to formation of a-0 dimer molecules capable of competing with labeled hFSH for specific receptor sites in the testes tubule homogenate, as well as stimulating ovarian growth in viva (Table II. Taken together with the in tivo biological activity of the hybrid molecules in the specific Steelman-Yohley bioassay this would seem to confirm the hormone-specific character of the hFSH-@ subunit fraction, as well as the efficacy of the subunit dissociation and isolation procedure previously reported (7). The high activity of the bLH-a + hFSH-@ hybrid molecule in w&o was somewhat surprising.
However, there is a precedent for hybrid molecules having receptor binding activity greater than the most active intact hormone.
A similar situation was encountered in our earlier studies with LH and hCG (2) where it was observed that a hybrid molecule formed between the (Y subunit of porcine LH and the 0 subunit of hCG had a receptor binding activity somewhat greater than that of intact hCG, which was the most active LH molecule examined in that study.
Presumably, the configuration of the hybrid dimer is such as to confer a greater affinity of the molecule for the receptor than that of the native hormone.
by guest on March 24, 2020 http://www.jbc.org/ It is of interest that the bovine LH-a/hFSH-/? fraction discussed above had such a low activity in wivo. A possible explanation could be related to the sialic acid content of hFSH (27), which is considerably greater than that of bovine LH (28). Morel1 et al. (29) have shown that sialic acid content is related to sustained levels of glycoproteins in plasma, and Vaitukaitis and Ross (30) have shown a relationship between sialic acid content of urinary FSH and biological activity It is possible that the discrepancy between in vitro and in vivo activity of the bLH-cr/ hFSHj3 dimer is due to differences in the sialic acid content of the bovine LH-a! compared to that of hFSH-cu.2 This might result in an accelerated clearance of the hybrid molecule from the circulation.
Such an explanation has been invoked to explain certain pharmacological differences in the biological properties of bovine, porcine, and human FSH (15,31). The greater in vitro activity of bLH-a/hFSH# hybrid compared to that of the homologous human subunit pair could also be due to damage to the hFSH-a subunit during its dissociation and separation. bLH-a subunit is prepared by acid dissociation and separated from bLH-/3 subunit by counter current distribution, while hFSH-a was dissociated from hFSH-/3 by incubation with urea. The latter dissociating agent may be responsible for a structural perturbation in the hFSH-cr subunit which prevents efficient recombination with the hFSH-/3 subunit. For example, variations in the degree of denaturation of the a! subunits could explain the differences in measured activity of the respective hybrid FSH molecules. Also, the possible presence of impurities in the subunit preparations could account for failure to recover greater biological activity after incubation under conditions favoring formation of CX//~ dimers.
To our knowledge, this is the first report dealing with the kinetics of formation of subunits of hFSH, as well as kinetics of formation and biological activity, in vitro and in tivo, of hybrid FSH molecules.
We are currently extending these studies to probe the structural requirements for FSH subunit combination and receptor binding affinity.