The heterogeneity of human chorionic gonadotropin (hCG). II. Characteristics and origins of nicks in hCG reference standards.

0013-7227/91/1293-1551$03.00/0 Endocrinology Copyright (e) 1991 by The Endocrine Society

Vol. 129, No. 3 Printed in U.S.A.

The Heterogeneity of Human Chorionic Gonadotropin (hCG). II. Characteristics and Origins of Nicks in hCG Reference Standards* STEVEN BIRKEN, MARY ANN GAWINOWICZ, ANDREW KARDANA, AND LAURENCE A. COLE Department of Medicine, Columbia University College of Physicians and Surgeons (S.B., M.A.G.), New York, New York 10032; and the Department of Obstetrics and Gynecology, Yale University School of Medicine (A.K., L.A.C.), New Haven, Connecticut 06510

ABSTRACT. hCG, the hormone produced by the trophoblast throughout pregnancy, has peptide bond cleavages, or nicks, in the 0-subunit. We sought to compare the nature of these nicks in standard reference preparations of hCG, to determine the enzymes that may be responsible for generating the peptide bond cleavages, and to devise means of separating nicked from intact hormone. The standard reference preparations of hCG, which are purified from a commercial product made from large pools of pregnancy urine, were found to have varying concentrations of nicked hormone. The preceding report showed that 11 of 13 hCG preparations isolated from individual pregnancy urine samples were nicked at the /347-48 bond, with 2 of 13 having a second

h

CG IS a glycoprotein hormone produced by the trophoblast starting very early in pregnancy and serves to maintain steroid secretion by the corpus luteum. It is the sole placental member of the gonadotropin family of glycoprotein hormones; each has a common a-subunit and a unique 0-subunit. The other three related hormones are produced in the pituitary: FSH, LH, and TSH. Several groups, including ours, have reported peptide bond nicks in the 0-subunits of the pituitary glycoprotein hormones (1-3) and in hCG (4-9). The reports of Puisieux et al. (8) and Sakakibara et al. (9) have found that hCG isolated from the urine of pregnant individuals has a peptide bond nick in the 0-subunit between residues 47-48, while our companion report (10) shows that some preparations of hCG are also nicked at 044-45 and 04647. We have observed that the standard reference preparations of hCG, produced at Columbia University, are Received March 12,1991. Address all correspondence and requests for reprints to: Steven Birken, Department of Medicine, Columbia University, College of Physicians and Surgeons, 630 West 168th Street, New York, New York 10032. * This work was supported by NIH Grants HD-15454 (to S.B.), CA44131 (to L.A.C.), and CA-46828 (to L.A.C.).

nick at /S44-45. As shown here, all of the hCG reference standards are nicked to similar extents at both the /J47-48 bond and the j844-45 bond. The percentage of peptide bond nicking in the various hCG standard preparations ranged from 10-20% and appeared higher in the more recent preparations. We showed that human leukocyte elastase is capable of specifically cleaving the /S44-45 bond, and in extended digests it can also cleave the 048-49 and /351-52 peptide bonds. Thus, human leukocyte elastase may be the origin of some of these cleavages in the individual samples and the reference standards. Furthermore, we report that a monoclonal antibody directed to hCG a-/3 dimer binds preferentially to nonnicked hCG and much less to nicked hCG. {Endocrinology 129: 1551-1558, 1991)

cleaved at both 044-45 and 047-48 (CR series and First International Reference Preparation). These two cleavages were present in hCG reference preparation CR121, which had been used to grow crystals of hCG (7). Since the hCG reference standards differed from each other and from hCG isolated from pregnant individuals (10), we undertook this study to compare the peptide heterogeneity of all of the hCG reference standards, to infer the protease responsible for the peptide nicks, and to develop an immunological method to separate intact hormone from nicked forms. Materials and Methods Preparation of hCG standards The current purification methods for the hCG reference standards have been recently reviewed (11). The Columbia University laboratory has purified all of the CR series of hCG reference standards for the NIH and WHO (including the First International Reference Preparation for immunoassay and the Third International Standard for bioassay), which are distributed to investigators worldwide. The Organon (Diosynth division) standard clinical grade of hCG with an average specific activity of 3500 IU/mg is the starting material for the purification procedure. All reference preparations of hCG, starting

1551

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 06 July 2015. at 10:43 For personal use only. No other uses without permission. . All rights reserved.

NICKS IN hCG REFERENCE STANDARDS

1552

Endo • 1991 Voll28«No3

with CR123, were purified by gel filtration in ammonium bicarbonate, followed by anion exchange chromatography in the same buffer (11). Older preparations, such as CR119, had an additional first step, batch cation exchange at pH 5.0 before gel filtration (12). Subunits were purified by ion exchange, as described previously (13).

calculated for hCG CR127 at zero time. It was also calculated for the 2- and 21-h hLE digests. The percent decrease in the intact hCG/3 peak due to hLE digestion was calculated (we assumed that hCG CR127 was already 20% nicked, as approximated by sequence analysis). Peaks were quantitated using Nelson Analytical Xtrachrom software (PE Nelson).

Human leukocyte elastase (hLE) digestion

Immunoaffinity separations

hCG CR127 or hCG CR129 /3-subunit was dissolved (5 nmol) in 100 n\ water. hLE (E1508, Sigma, St. Louis, MO) was dissolved in elastase digestion buffer [0.1 M Tris-HCl, 0.5 M NaCl, and 0.1% Brij (wt/wt), pH 8.5] at a concentration of 1.5 U enzyme/ml. One hundred microliters of hLE were added to start the reaction, and the mixture was incubated at 37 C. Aliquots were taken at 0 min, 10 min, 30 min, 2 h, and 21 h and placed in tubes containing 2 /xi 0.15 M phenylmethylsulfonylfluoride in 2-propanol for sodium dodecyl sulfate (SDS)-gel and sequence analyses. A repeat experiment was performed with 48-h incubation using the same conditions. A second aliquot at each time point was placed in a tube that contained 10 /A of 10 mg/ml soybean trypsin inhibitor (1%) as an inhibitor of the enzyme for use in the biological and immunological assays described in the companion paper (14). A control study was run in parallel, with the omission of enzyme. Assay results were normalized to control values.

An immunoaffinity column of antibody B109 (15) was made using a Pierce SelectiSpher-10 activated Tresyl high performance affinity column (10 cm X 5 mm; Pierce Chemical Co., Rockford, IL) according to the manufacturer' s directions. Approximately 85 mg antibody were bound to this column out of 100 mg purified immunoglobulin loaded onto the column. hCG /3-subunit CR129 (5 mg) was loaded onto the B109 affinity column in 0.1 M ammonium bicarbonate. The column was washed with 5 ml water and eluted with 2 ml 0.1 M sodium citrate, pH 2.2. An additional elution using 6 M guanadine hydrochloride at pH 4 was used for both further elution of hormone and washing of the column. A second experiment was performed using 50 mg hCG /3-subunit CR129. Again, the column was loaded in 0.1 M ammonium bicarbonate, washed with 5 ml water, and eluted with 2 ml 0.1 M sodium citrate, pH 2.2, followed by 2 ml 6 M guanadine hydrochloride, pH 4. Intact hCG CR127 in 1-mg lots was injected onto the B109 column in the same fashion and eluted in a similar manner.

NH2 terminal sequence analysis Sequence analysis was performed on an Applied Biosystems model 470A gas phase sequencer (Applied Biosystems, Foster City, CA) equipped with an on-line phenylthiohydantoin analyzer. Peaks were quantitated using Nelson Analytical Xtrachrom software (PE Nelson, South Plainfield, NJ). hLE digests were loaded directly onto a precycled glass fiber support without additional purification. HPLC

Miscellaneous procedures Liquid phase RIAs were performed, as previously described (16). Assays were performed using B109 antibody and [125I] hCG as tracer, with both hCG CR127 and 100% nicked hCG preparation C5(10) as competitors, and were repeated twice. The procedures for SDS-gel electrophoresis and staining of gels were also described previously (17).

Results

Samples were incubated for 1 h at 37 C in buffer A [0.05% trifluoroacetic acid (TFA)] before injection onto a Vydac C4 reverse phase column (4.6 mm X 25 cm; Separations Group, Hesperia, CA) in order to fully dissociate the hCG. Omission of incubation resulted in an additional peak of intact hormone eluting after hCG/3. After 5 min of isocratic elution at 0% B, a gradient was run to 80% buffer B (buffer B is 0.05% TFA in 100% acetonitrile) in 40 min. Detection was accomplished by measuring absorbance at 230 and 280 nm. A Waters HPLC (Waters Associates, Milford, MA) was employed. Quantitation of percentage of nicking The percentage of bond cleavages in the hCG CR127 0subunit was estimated by sequence analysis and then corroborated by reverse phase HPLC. Reverse phase HPLC separated nicked /3 from intact /?, and the areas under the peaks were used to help calculate the percent nicking of /3 after the hLE digestion. Since hCG CR127 standard is already nicked, only the increase of nicking between the zero time and the digest time points could be calculated after the HPLC separation procedure. We divided the area of the intact hCG /3-subunit peak by the total area of all of the peaks. This fraction was

The biological activities of the hCG reference standards were determined at the NIH in the year of preparation by in vivo assays (increase in ventral prostate weight) for batch CR121 and earlier preparations (18) (see Table 1 for nomenclature). In vitro receptor assays were used for the later hCG reference standards (19, 20). Because the 95% confidence intervals inherent in biological assays are large, the differences in the biological activities of the reference standards are not significant. The 95% confidence interval of hCG CR119, for example, is 9,560-16,624 mlU/mg (21), which includes the estimated activity of every reference preparation (see Table 1). Sequence analyses indicated some quantitative differences in peptide bond heterogeneity among preparations. The ratio of each NH2-terminal sequence to residue 1 of the 0-subunit present in each sample of hormone was estimated from the recoveries of specific residues in each sequence (see Table 1). Variable recoveries of individual amino acids and some overlapping sequences make pre-



1553

TABLE 1. Characteristics of hCG reference standards Preparation

Lot no.

yr 1972 1973 1974 1977 1978 1988 1989

CR117 hCG CR119 hCG CR121 hCG CR123 hCG CR125 hCG CR127 hCG CR129 hCG CR119 jS-subunit CR127 0-subunit CR129 j8-subunit

Biological activity (IU/mg)

N-Terminal° al"

a3

«4

j81 e

/345

048

10,600 13,500 14,000 12,780 11,900 14,900 11,300

0.95 0.77 0.74 1.08 1.16 1.12 1.47

0.11 0.08 0.0 0.10 0.16 0.03 0.05

0.54 0.29 0.31 0.22 0.36 0.27 0.26

1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00

0.06 0.05 0.05 0.10 0.11 0.17 0.16 0.06 0.11 0.05

0.05 0.04 0.04 0.16 0.15 NDd 0.20 0.02c 0.13e 0.05"

0

Ratio of sequences contained in each sample determined by setting the picomole amount of/31 sequence at 1.00. * Picomole amount of a sequences determined by Ala in cycle 1, Asp in cycle 1, and Val in cycle 1. c Picomole amount of/3 sequences determined by Lys in cycle 2, Glu in cycle 3, Leu in cycle 1, and Leu in cycle 2. Wherever appropriate, values are corrected for lag and background. d Not determined. Leu in cycle 2 obscured by large Lys peak. e Based on the presence of Val in cycle 1.

cise quantitation by sequence analysis of hCG a-0 dimer difficult, since the percentage of bond nicking in the standards is low, but the values presented are reasonable estimates, calculated as described above. All standard preparations exhibited a amino-terminal heterogeneity, as previously described (22). In all standard hCG reference preparations, the presence of a nick at the 044-45 bond was clear, while the 047-48 bond cleavage was more difficult to detect at low levels because of the heterogeneity of the or-chain. The 047-48 bond cleavage was more easily identified in the isolated 0-subunits (see Table I). A recent study of nicked hCG employed a reverse phase HPLC system to separate nicked 0-subunit from intact hormone (9). Using a modification of this method, we separated the reference preparation hCG CR129 into four main peaks (Fig. 1, upper panel). The 0-subunit of CR129, prepared by dissociation of hCG a-0 dimer followed by ion exchange chromatography (13), was separated into three peaks (Fig. 1, lower panel). Drawing on our experience with the fingerprint of hCG on reverse phase HPLC (17), we were able to identify all the peaks shown in Fig. 1. The first peak to elute, as shown by sequence analysis and SDS-gel analysis (23.6 min), was nicked hCG0. The next peak, eluting at 24.7-24.8 min, was a mixture of a- and 0-chains and was at the position identified in an earlier study (17) as that of a-subunit containing methionines oxidized to sulfoxides and 0subunit with its single methionine likewise oxidized. The next two major peaks were the intact a-subunit (25.3 min) and the intact 0-subunit (26.1 min), as reported by Sakakibara et al. (9). By separating the reference standard hCG CR127 on HPLC in this manner (not shown), we confirmed that this standard contained equal molar ratios of the 044-45 bond cleavage and the 047-48 bond cleavage in the peak representing nicked 0-subunit (see

Table 1). Supporting this finding was the observation that purified 0-subunit preparations hCG CR119 and hCG CR129, which were made by ion exchange procedures (13) from the standard hCG reference preparations, CR119 and CR129, respectively, contained both peptide bond nicks. In addition, it was shown that the peak containing the oxidized hCG displayed low concentrations of several cleaved chains, including a-subunit cleaved at the 70-71 bond position and 0 cleaved at 4445. hLE cleavage of hCG and hCG (3-subunit

We sought to identify the enzyme responsible for the peptide bond cleavages. As shown in Table 2, hLE specifically cleaves hCG at 044-45. A time-course study showed that although the hCG CR127 starting material had nicks at 047-48 and 044-45 (Table 1), after a 2-h digestion with hLE further cleavage occurred only at the 044-45 bond (30%). The free 0-subunit proved more susceptible to cleavage, and in 2 h was 75% cleaved at the 044-45 bond and 54% cleaved at the 051-52 position, while prolonged enzymatic treatment of the 0-subunit (21 h) produced significant cleavage at the 05-6 position. Extended digestion of hCG with hLE (48 h) demonstrated that the enzyme would cleave at 05-6 as well as 048-49 with the loss of the 045 amino-terminus, presumably as a result of cleavage and loss of peptide 45-48, and also resulted in some dissociation of hCG, observed upon gel filtration (data not shown). Reverse phase HPLC separation techniques were also used to help interpret the results of hLE digestion of hCG CR127. Figure 2 shows a comparison of 2- and 21h hLE digests with standard hCG CR127 starting material separated on the reverse phase HPLC system. Amino acid sequence analysis determined the relative ratio of



1554

Endo • 1991 Vol 129 • No 3

TABLE 2. N-Terminal analysis of hLE digest Time point (h)

N-Terminal" al

all

/3-Subunit (CR129) 0.17

1.00 1.00 1.00 1.00

0.5 2 21

Nicked

f

P

Oxidized

Nicked _J

P

FIG. 1. Reverse phase HPLC chromatogram developed on a Vydac C4 protein column using a TFA-acetonitrile eluant system. Buffer A was 0.05% TFA, while buffer B was 0.05% TFA in 100% acetonitrile. The flow rate was 1 ml/min, and the gradient was 5 min at 0% B, 45 min to 80% B. The abscissa represents time at a chart speed rate of 0.5 cm/ min. The times (minutes) of the eluted peaks appear above each peak. The ordinate is absorbance at 230 nm [1.0 absorbance units full scale (AUFS)]. The upper panel is the separation of hCG CR129 (400 ng), while the lower panel is the separation of hCG 0-subunit CR129 (1000 jig), which was purified from the hCG shown in the upper panel by ion exchange techniques (13). The (a + 0)' in the upper panel denotes forms of hCGa and hCG0 with oxidized methionines.

amino-termini in each of the peaks as follows. The nicked 0 region of CR127 hCG at time zero (peaks 1-2) contained 01 (1.0), 045 (0.57), and 048 (0.53). The corresponding region of the 2-h hLE digest contained 01 (1.0), 045 (1.0), and 048 (0.37). The same region of the 21-h hLE digest contained 01 (1.0), 045 (0.54), and 049 (0.36). We corroborated the estimates of the percentage of nicking derived from sequence analysis using the integrated areas of the HPLC peaks (see Materials and Methods). The order of peak elution in Fig. 2 is identical to that in

hCG (CR127) 0.17 1.22 0.5 1.14 2 0.70 21 1.25 48 0.77

01

0.21 0.55

1.00 1.00 1.00 1.00 1.00

06

0.05 0.66

0.22

045

048

049

0.10 0.31 0.54 0.95

0.58 0.54 0.75 0.18 0.25 0.30 ND6 ND

052

0.16 0.13 0.15 ND ND

0.44 0.41

° The amount of each sequence was estimated from the concentration of phenylthiohydantoin amino acids (picomoles) and was normalized to the sequence starting at 01. 6 Not determined. Overlapping residues from other sequences and contaminating peaks from digest buffer obscure these sequences, if present.

Fig. 1, although the abscissa is expanded. Peaks 1 and 2 are nicked 0-subunit, peak 3 contains oxidized forms of hCG, peak 4 is intact a, and peak 5 is intact 0. If the standard hCG CR127 intact 0-subunit peak 5 area is taken as a reference, this peak declines by 32% in 2 h and by 86% in 21 h. This means that if hCG CR127 contains approximately 80% intact hCG, then after 2 h of hLE digestion it is only 54% intact and after 21 h it is only 11% intact. Methods were explored to separate intact hormone from nicked hCG, avoiding the dissociation that takes place during reverse phase chromatography. In the course of an attempt to remove hCG a-0 dimer from a preparation of 0-subunit (CR129), the 0-subunit was passed through a HPLC monoclonal antibody affinity column (monoclonal antibody B109 bound by Tresyl link to the column matrix). B109 is a monoclonal antibody directed to hCG a-0 dimer (15). The reverse phase separations shown in Fig. 1 were used to evaluate the passthrough and bound products of the affinity column. Nicked 0-subunit was easily separated on such a reverse phase column. This HPLC column method used in conjunction with SDS-gel electrophoresis and amino acid sequence analysis permitted complete evaluation of the products of the HPLC affinity column. Two HPLC immunoaffinity experiments with different column loads of 0-subunit were performed. When 5 mg hCG0 were injected into the column, the pass-through consisted of a small quantity of nicked hCG a-0 dimer. Elution of the column with citrate buffer released about half of the bound 0-subunit, which proved by sequence analysis to be free of all nicks. In a second experiment using 50 mg of the subunit preparation, the pass-through contained about 40 mg 0-subunit, with nicks (see Fig. 3, upper



1555

21H ELRSTflSE DIGEST

FIG. 2. Reverse phase HPLC chromatograms of reference standard hCG CR127 and the products of 2- and 21-h hLE digestions of the same batch of hCG CR127 (see Materials and Methods). The gradient was run as described in Fig. 1, except the absorbance data were converted to digital data and integrated and plotted using Nelson Analytical Xtrachrom software. The abscissa has been expanded to permit easier visualization of the pattern of eluted peaks.

3 ,

2H ELflSTflSE DIGEST

2 ,

CR127 HCG

20.00

panel), while the citrate eluate contained intact ,8-subunit (see Fig. 3, lower panel) as did the guanadine hydrochloride wash (see Materials and Methods). It appeared that the B109 column could bind approximately 10 mg intact /3-subunit. The SDS-gel electrophoresis analysis of the pass-through of the B109 column showed that a band below the /?-subunit was removed by the B109 column (see gel inset labeled sub-,8, lower panel of Fig. 3). This band has not been identified. The band labeled /?' is ]9 45/48-145 derived from nicked 0-subunit. In an attempt to purify nicked hCG from intact hormone, the reference preparation of hCG CR127 was loaded onto the B109 affinity column. It was found that although the pass-through was enriched in nicked hCG or-/3 dimer over the column load, a significant quantity of nicked hCG a-/3 dimer bound and was eluted from the B109 column with citrate buffer, pH 2.2, along with intact hCG a-,8 dimer. Liquid phase dose-response curves were performed to define the binding properties of the antibody attached to this column. The assay system was composed of antibody B109, iodinated hCG CR127, and three competitors: hCG CR127, preparation C5 (10) (which is 100% nicked hCG), and hCG 0-subunit CR129. It was found that 100% nicked hCG bound to B109 less than 10% as well as hCG CR127 at the 50% displacement point, while the free /3-subunit bound less than 0.1% as well as hCG to B109.

Discussion There have been a number of reports concerning the presence of peptide bond cleavages within the ,8-subunit

30.00

of the glycoprotein hormones. As early as 1973, Shome and Parlow (1) described cleavages in the 43-49 region of hLHjS isolated from pituitary tissue. Ward et al. (3) reported that hLH molecules with cleavages in the 4349 region had reduced biological activity. It was not unexpected that pituitary hormones would be damaged by proteases, since that tissue is very rich in such enzymes, and human pituitaries are generally not removed until 24-72 h after death. However, many of the peptide bond cleavages may occur during purification procedures, in addition to the cleavages that are already present when the hormones are extracted from the pituitary tissue itself (2). Nishimura et al. (4), Bidart et al. (5), Cole et al. (6), and Lustbader et al. (7) reported cleavages in a similar region of the hCG ,8-subunit. Bidart et al. (5), Puisieux et al. (8), and Sakakibara et al (9) reported hCG cleavage only at the ,847-48 bond. The companion publication shows that 11 of 13 hCG preparations isolated from the urine of individual women were nicked at the ,847-48 bond, 2 of 13 at ,844-45, and 1 of 13 at 04647 (10). We have observed that all hCG reference preparations are nicked at the first 2 of these sites. Various standard reference preparations of hCG (CR series) have been purified at Columbia University during the past 2 decades and have been supplied to investigators around the world through the Contraceptive Research Development program of the NICHHD and the National Hormone and Pituitary Program of the NIDDK. Table 1 details the characteristics of each of these reference preparations of hCG. Preparation CR119 was donated to the WHO and was designated the First International Reference Preparation for Immunoassay



1556

Oxidized

Nicked

CM

Oxidized

6K*

M fi fi load pom

FIG. 3. Reverse phase HPLC chromatograms of samples of the passthrough (upper panel) and acid eluate (lower panel) of hCG /3-subunit CR129 from a B109 immunoaffinity column. The abscissa is time, with the elution times in minutes marked above each peak. The ordinate is absorbance at 230 nm, 1.0 AUFS for the upper panel and 0.2 AUFS for the lower panel. The Vydac C4 column was run as described in Fig. 1. The reduced SDS-gel electrophoresis inset in the lower panel shows the silver-stained band pattern of the hCG /3-subunit preparation loaded onto the B109 column compared to the sample of the pass-through of the B109 column that is chromatographed in the upper panel. The B109 column removed the sub-/S band, which allowed the j8 band, composed of 045-145 and 048-145, to pass through the column.

(21, 23) and later as the Third International Standard for hCG, established in 1986 as code no. 75/537 at the 37th Meeting of the WHO Expert Committee on Biological Standardization (24). Although the method of preparation of hCG by Organon is proprietary, this starting material was assumed not to have changed during the past 15 yr, and thus, the

Endo«1991 Vol 129 • No 3

reference preparations should be very similar to each other. Although the Columbia purification scheme was abbreviated starting with hCG CR123 by elimination of a first batch cation exchange step, biological activity assays of all of the reference preparations appear to confirm their similarity, since the activity values do not vary to a significant extent. Microsequence analysis of each of the reference preparations indicated somewhat increased peptide nicking in more recent preparations. Although it is difficult to precisely estimate the percentages of nicking of the whole hormone when the nicks are present at low levels, the reference standard preparations appear to range from approximately 10% nicking for CR119 to as much as 20% in CR127. The preceding report shows that hCG purified from individual urine samples, whether from pregnancy, hydatidiform mole, or choriocarcinoma, displayed a wide range of /3-subunit nicking, from 0-100% (10). None of the normal pregnancy samples had the |344-45 bond nick, and none had the a-subunit NH2-terminal heterogeneity that was observed in the reference standards. The reference standards are derived from very large pools of pregnancy urine and should represent an average composition. In the light of these observations, we question whether the reference standards are appropriate for measurement of hCG in certain individual urines, since significant individual variation was shown to exist (10). We sought to answer three basic questions. 1) Could we infer which enzyme was causing the nicking in urinary hCG? 2) Why were the standard hCG preparations made from the Organon starting material nicked at 044-45 and heterogeneous in sequence at the NH2-terminus of the hCG a-subunit while the individual pregnancy samples were not? 3) Could we devise a means to separate nicked hormone from intact? The human protease hLE cleaves after glycine, valine, and leucine within exposed hydrophobic regions (25). Since the /338-57 loop of hCG is amphipathic, that is it can be viewed as being a helix, with one hydrophobic face and one hydrophilic face, this region appeared to be a candidate for cleavage by this enzyme within its hydrophobic face (26). Furthermore, leukocytes are present in high concentrations around trophoblastic tissue, which itself may secrete additional leukocyte proteases, and at the trophoblast-myometrium interface (27). We found that this elastase enzyme did produce the /544—45 bond cleavage in hCG within a 2-h digestion period. In the case of free hCG /3-subunit, which was cleaved much more rapidly than intact hormone, a second cleavage at /351-52 appeared simultaneously with the 044-45 cleavage. Prolonged digestion of free j8 or intact hCG generated additional cleavages, including /35-6 and /S48-49. The a-subunit was also cleaved at the 70-71 bond after 21 h of digestion. Since hLE cleaves at multiple sites in



the 41-54 region of the 0-subunit and specifically cleaves 05-6, it may be involved in the production of 0-fragment, a form of 0-subunit produced in pregnancy which is missing residues 1-5 and 41-54 (16). Why are the hCG reference standards nicked differently from hCG purified from single pregnant individuals? It was possible that the reason why individual urinary hCG preparations did not exhibit the 044-45 bond cleavage was because the 045-47 region had been removed by sequential protease digestion. The report by Kardana et al. (10) showed that this was not the case and that nicked hCG from a single individual contained peptide 01-47 and not 01-44. Therefore, it was unlikely that hLE alone had caused the 047-48 bond cleavage. In addition, as shown in the companion paper (10), none of six individual pregnancy hCG samples were nicked at 044-45, and none had a-subunit heterogeneity. In contrast, two of seven trophoblast disease individual samples were nicked at 044-45, and six of seven had a-subunit NH2-terminal heterogeneity. These findings suggest that nicking at 044-45 and a-subunit heterogeneity in the Organon preparations may originate from samples from individuals with molar or choriocarcinoma or other abnormal pregnancies. Another possible explanation for both types of bond cleavages may be the presence of pus in some urine specimens, which would result in the release of many leukocyte proteases into the urine. It is also possible that small quantities of proteases acted on hCG during purification from the Organon material to produce additional nicks. Kardana et al. (10) used acetone and ethanol extractions as initial steps for purification of hCG from urine samples, while the procedure that Organon employs for purification has not been published. Some of these structural differences may be related to differences in isolation techniques. In an attempt to further purify hCG CR129 0-subunit free of traces of intact biologically active hCG, the 0subunit was applied to a B109 immunoaffinity column. Antibody B109 was selected to bind a-0 dimer hCG and was the basis of a highly sensitive and specific assay for hCG (15, 28). It was observed that free 0-subunit bound to the column when high concentrations of subunit were loaded. The 0-subunit, which had negligible cross-reaction with B109 in the liquid phase, was binding to the immunoabsorbant by mass action effects of the high free 0 and antibody concentrations. The striking observation was that the hCG 0-subunit bound to the B109 column was free of nicked material and that nicked material preferentially passed through the immunoaffinity column. This was interpreted to mean that it might be possible to use B109 to separate nicked from intact hCG. When this experiment was carried out, it was found that while nicked 0 did not bind, some nicked hCG did, in fact, bind to the B109 column, although to a lesser extent

1557

than intact hCG. The affinity of nicked free 0-subunit for B109 must be much lower than that of intact 0, since no nicked 0 bound to the column at all. The different results found upon loading 5 mg 0-subunit on the column compared to 50 mg 0-subunit were now explained. Although free 0-subunit reacts less than l/1000th as well as nicked hCG, at very high concentrations 0 competes with nicked hCG for the B109 sites. This was the case for the 5-mg load when all 5 mg 0 bound and some nicked hCG passed through the column. At 50 mg, the capacity of the B109 immunoaffinity column for 0-subunit was exceeded, and 0-subunit appeared in the pass-through along with nicked hCG. Therefore, B109 cannot be used in immunoaffinity format to separate nicked hCG from intact hCG, but proved to be a very useful measurement tool in an immunoradiometric assay format, as described in the following article by Cole et al. (14). The reference preparations of hCG have served the scientific community for nearly 2 decades and have significantly furthered the understanding of the structure and function of the glycoprotein hormones. Preparation hCG CR119 has also served as the First International Immunoassay Standard and now the Third International Reference Standard for bioassay. The presence of peptide bond nicks in these hormones has been noted for many years by our laboratory. Careful examination of the reference preparations indicates a possible trend toward more bond nicking in more recent preparations, perhaps the result of a change in urine collection procedures or in the contributing population. The discoveries that hLE is a likely contributor to these bond cleavages and that certain antibodies widely used to measure intact hormone are sensitive to such nicking raise important issues in the quantitation of hCG (14). The differences in structure between the reference preparations of hCG and individual urinary hCG samples reinforce such concerns (10). Acknowledgments We wish to acknowledge Dr. Gabe Bialy and the Contraceptive Research Development Branch for support in developing the reference preparations of hCG. We also wish to thank Dr. Bruce Nisula for information and thoughtful discussions about the biological activities of these reference preparations. All biological assays were performed at the NIH, with most supervised by Dr. Nisula. In addition, we acknowledge the expert technical assistance of Ms. Gladys Agosto in the performance of immunoassays. References 1. Shome B, Parlow AF 1973 The primary structure of the hormonespecific, (8 subunit of human pituitary luteinizing hormone (hLH). J Clin Endocrinol Metab 36:618-624 2. Hartree AS, Lester JB, Shownkeen RC 1985 Studies of the heterogeneity of human pituitary LH by fast protein liquid chromatog-


1558


raphy. J Endocrinol 105:405-414 3. Ward DN, Glenn SD, Nahm HS, Wen T 1986 Characterization of cleavage products in selected human lutropin preparations. Int J Peptide Protein Res 27:70-78 4. Nishimura R, Ide K, Utsunomiya T, Kitajima T, Yuki Y, Mochizuki M 1988 Fragmentation of the /3 subunit of human chorionic gonadotropin produced by choriocarcinoma. Endocrinology 123:420-425 5. Bidart J-M, Puisieux A, Troalen F, Foglietti MJ, Bohuon C, Bellet D 1988 Characterization of the cleavage product in the human choriogonadotropin /3 subunit. Biochem Biophys Res Commun 154:626-632 6. Cole LA, Kardana A, Birken S 1989 The isomers, subunits and fragments of hCG. Serono Symp 65:59-78 7. Lustbader JW, Birken S, Pileggi NF, Kolks MAG, Pollak S, Cuff ME, Yang W, Hendrickson WA, Canfield RE 1989 Crystallization and characterization of human chorionic gonadotropin in chemically deglycosylated and enzymatically desialylated states. Biochemistry 28:9239-9243 8. Puisieux A, Bellet D, Troalen F, Razafindratsita A, Lhomme C, Bohoun C, Bidart J-M 1990 Occurrence of fragmentation of free and combined forms of the /? subunit of human chorionic gonadotropin. Endocrinology 126:687-694 9. Sakakibara R, Miyazaki S, Ishiguro M 1990 A nicked /3 subunit of human chorionic gonadotropin purified from pregnancy urine. J Biochem 107:858-862 10. Kardana A, Elliott MM, Gawinowicz MA, Birken S, Cole LA 1991 The heterogeneity of human chorionic gonadotropin (hCG). I. Characterization of peptide heterogeneity in 13 individual preparations of hCG. Endocrinology 129:1541-1550 11. Birken S, Krichevsky A, O'Connor J, Lustbader J, Canfield RE 1990 Chemistry and immunochemistry of hCG, its subunits and its fragments. In: Chin WW, Boime I (eds) Glycoprotein Hormones, Serono Symposia, USA. Raven Press, New York, pp 45-61 12. Canfield RE, Morgan FJ 1973 Human chorionic gonadotropin (hCG). I. Purification and biochemical characterization. In: Berson SA, Yalow RS (eds) Methods in Investigative and Diagnostic Endocrinology. North Holland, Amsterdam, pp 727-733 13. Morgan FJ, Canfield RE, Vaitukaitis JL, Ross GT 1974 Properties of the subunits of human chorionic gonadotropin. Endocrinology 94:1601-1606 14. Cole LA, Kardana A, Andrade-Gordon P, Gawinowicz MA, Morris JC, Bergert ER, O'Connor J, Birken S 1991 The heterogeneity of human chorionic gonadotropin (hCG). III. The occurrence and biological and immunological activities of nicked hCG. Endocrinology 129:1559-1567 15. Krichevsky A, Armstrong EG, Schlatterer J, Birken S, O'Connor J, Bikel K, Silverberg S, Lustbader JW, Canfield RE 1988 Prepa-

16.

17. 18. 19.

20.

21. 22.

23.

24. 25. 26.

27. 28.

Endo • 1991 Vol 129 • No 3

ration and characterization of antibodies to the urinary fragment of the human chorionic gonadotropin /J-subunit. Endocrinology 123:584-593 Birken S, Armstrong EG, Kolks MAG, Cole LA, Agosto GM, Krichevsky A, Vaitukaitis JL, Canfield RE 1988 Structure of the human chorionic gonadotropin /3-subunit fragment from pregnancy urine. Endocrinology 123:572-583 Pollak S, Halpine S, Chait B, Birken S 1990 High resolution HPLC fingerprinting of purified hCG demonstrates that oxidation is a cause of hormone heterogeneity. Endocrinology 126:199-208 McArthur JW 1952 The identification of pituitary interstitial cell stimulating hormone in human urine. Endocrinology 50:304-310 VanDamme MP, Robertson DM, Dicsfalusy E 1974 An improved in vitro bioassay method for measuring luteinizing hormone (LH) activity using mouse leydig cell preparations. Acta Endocrinol (Copenh) 77:655-671 Taliadourous GS, Amr S, Louvet JP, Birken S, Canfield RE, Nisula BC 1982 Biological and immunological characterization of crude commercial human choriogonadotropin. J Clin Endocrinol Metab 54:1002-1009 Canfield RE, Ross GT 1976 A new reference preparation of human chorionic gonadotropin and its subunits. Bull WHO T-54:463-470 Birken S, Canfield RE 1980 Chemistry and immunochemistry of human chorionic gonadotropin. In: Segal S (ed) The Human Chorionic Gonadotropin Molecule. Plenum Press, New York, pp 6588 Storring PL, Gaines Das RE, Bangham DR 1980 International reference preparation of human chorionic gonadotrophin for immunoassay: potency estimates in various bioassay and protein binding assay systems; and international reference preparations of the a and /3 subunits of human chorionic gonadotrophin for immunoassay. J Endocrinol 84:295-310 Insert with vials of the 3rd International Standard for chorionic gonadotrophin, code 751537, National Institute for Biological Standards, Control 1987, Hertfordshire, England Barret AM 1981 Leukocyte elastase. Methods Enzymol 80:581-587 Keutmann HT, Charlesworth MC, Mason KA, Ostrea T, Johnson L, Ryan RJ 1987 A receptor-binding region in human choriogonadotropin/lutropin 0 subunit. Proc Natl Acad Sci USA 84:20382042 Main EK, Stirzki J, Schochet P 1987 Placental production of immunoregulatory factors: trophoblast is a source of interleukin I. Trophoblast Res 2:149-160 O'Connor JG, Schlatterer JP, Birken S, Krichevsky A, Armstrong EG, McMahon D, Canfield RE 1988 Development of highly sensitive immunoassays to measure human chorionic gonadotropin, its j8 subunit and /3 core fragment in the urine: application to malignancies. Cancer Res 48:1361-1366


The heterogeneity of human chorionic gonadotropin (hCG). I. Characterization of peptide heterogeneity in 13 individual preparations of hCG.

The heterogeneity of human chorionic gonadotropin (hCG). III. The occurrence and biological and immunological activities of nicked hCG.

Evidence for the presence of human chorionic gonadotropin (hCG) and free beta-subunit of hCG in the human pituitary.

Early serum human chorionic gonadotropin (hCG) trends after medication abortion.

The carbohydrate moiety of human chorionic gonadotropin: lack of competition with HCG for testicular receptors and anti-HCG-serum.

Number and topography of epitopes of human chorionic gonadotropin (hCG) are shared by desialylated and deglycosylated hCG.

Chemical synthesis of the β-subunit of human luteinizing (hLH) and chorionic gonadotropin (hCG) glycoprotein hormones.

Human chorionic gonadotropin (HCG) in the treatment of obesity: a critical assessment of the Simeons method.

Combined human chorionic gonadotropin (HCG) and human menopausal gonadotropin (HMG) treatment in gonadotropin-deficient males with pituitary dwarfism.

Neoadjuvant human Chorionic Gonadotropin (hCG) therapy may improve the position of undescended testis: a preliminary report.

Assessment of human chorionic gonadotropin (HCG) levels during luteal phase in women using intrauterine contraception.

Hook effect in Abbott i-STAT β-human chorionic gonadotropin (β-hCG) point of care assay.

Human chorionic gonadotropin (hCG) interacts directly with recombinant human TSH receptors.

[Destructive bladder mole--follow-up by human chorionic gonadotropin (HCG), ultrasound and digital subtraction angiography (DSA)].

Inhibitory effect of human chorionic gonadotropin (hCG) on HIV-1 transmission from lymphocytes to trophoblasts.

CG receptor.

Evidence for a modulation of human chorionic gonadotropin (hCG) subunit messenger ribonucleic acid levels and hCG secretion by gamma-aminobutyric acid in human first trimester placenta in vitro.

First-trimester aneuploidy screening using serum human chorionic gonadotropin (hCG), free ahCG, and progesterone.

Twelve of fourteen surface epitopes of receptor-bound human chorionic gonadotropin (hCG) being antibody-inaccessible suggest an extensive involvement of the long extracellular domain of the hCG receptor.

Evaluation of the CLINITEST® human chorionic gonadotropin (hCG) pregnancy test for susceptibility to the hook effect by the hCG β core fragment.

Quantitation of human chorionic gonadotrophin (hCG) by radioreceptor assay.

Screening method to evaluate point-of-care human chorionic gonadotropin (hCG) devices for susceptibility to the hook effect by hCG β core fragment: evaluation of 11 devices.

Laser light scattering spectroscopic immunoassay in the agglutination-inhibition mode for human chorionic gonadotropin (hCG) and human luteinizing hormone (hLH).

Effect of human chorionic gonadotropin (hCG) on Neisseria gonorrhoeae invasion of and IgA secretion by human fallopian tube mucosa.