Molecular and Cellular Endocrinology, 69 (1990) 145-156 Elsevier Scientific Publishers Ireland, Ltd.

MOLCEL

145

02237

Synthesis and secretion of human chorionic gonadotropin and its subunits in choriocarcinoma cells: a comparative study with normal placental cells Y. Takeuchi,

R. Sakakibara

and M. Ishiguro

Department of Biochemistry, School of Clinical Pharmaceutical Sciences, Nagasaki University, l-14 Bunkyo-machi, Nagasaki 852, Japan (Received

Key words: Human

chorionic

gonadotropin:

14 August

Biosynthesis;

1989; accepted

Secretion;

4 December

1989)

BeWo cell; Placental

cell

Summary The human choriocarcinoma cell line, BeWo, synthesizes the glycoprotein hormone, human chorionic gonadotropin (hCG). We have undertaken this study to compare the synthesized and secreted forms of hCG and their II- and &subunits in cell cultures of BeWo cells to those forms of normal placental cells by immunobinding techniques. BeWo cells appeared to synthesize and secrete one species of the respective hCG subunit. The immature (Y- and /3-subunits, synthesized in BeWo cells as well as those of placental cells, were digested by endoglycosidase H indicating N-linked sugar chain(s) to be the high-mannose type. The mature (Y- and fi-subunits, secreted by BeWo cells as well as subunits of urinary hCG which are usually used as a standard hCG secreted by normal placental cells, were sensitive to neuraminidase treatment indicating that these subunits have terminal sialic acid(s). Contrary to placental cells, mature subunits of BeWo hCG could not be found in any subcellular fraction indicating a rapid secretion rate or supporting the hypothesis that BeWo cells secrete hCG subunits without the formation of secretory granules. The ol-subunit synthesized in BeWo cells had a slightly lower molecular weight than that of placental cells; however, the ar-subunit secreted by BeWo cells had a slightly higher molecular weight than the a-subunit of urinary hCG. The /I-subunits synthesized and secreted by BeWo cells had slightly higher molecular weights than /3-subunits of both placental cells and urinary hCG. Even after digestion by N-glycanase as well as endoglycosidase H, molecular weights were still different between the respective subunits of BeWo and placental cells indicating that the apoprotein structures of BeWo hCG subunits may differ from those of placental cells. Moreover, urinary P-subunit was sensitive to endo-a-N-acetylgalactosaminidase treatment but the P-subunit secreted by BeWo cells was not, indicating that the structure of O-linked sugar chain(s), if present, may be unusual. Analysis of assembled and free forms of subunits of BeWo cell cultures by sodium dodecyl sulfate-polyacrylamide gel electrophoresis under nonreducing conditions followed by immunobinding methods revealed that subunits are associated intracellularly and then secreted to the media as hCG. Moreover, only free P-subunits, but not cY-subunits, of BeWo hCG were found intra- and extracellularly.

Introduction Address for correspondence: Ryuzo Sakakibara, Ph.D., Department of Biochemistry, School of CIinicaI Pharmaceutical Sciences, Nagasaki University, 1-14 Bunkyo-machi, Nagasaki 852, Japan. 0303-7207/90/$03.50

0 1990 Elsevier Scientific

Publishers

Ireland,

Human coprotein Ltd.

chorionic hormone

gonadotropin (hCG) is a glyconsisting of noncovalently

146

bonded subunits, (Y and ,B. Both subunits contain two asparagine-linked (N-linked) sugar chains (Bellisario et al., 1973; Carlsen et al.. 1973; Endo et al., 1979; Kessler et al., 1979a). In addition, the P-subunit contains four mucin-type (O-linked) sugar chains (Kessler et al., 1979b). The hormone as well as its uncombined subunits (free (Y- and P-subunits) are secreted by placental trophoblasts (Ashitaka et al., 1974) and several trophoblastic and nontrophoblastic cancer cells. Differences in structure of N-linked sugar chains between hCGs purified from urine of healthy pregnant women and patients with choriocarcinoma have been elucidated (Mizuochi et al., 1983; Endo et al., 1987). In addition, similar comparative studies of O-linked sugar chains have been performed (Cole, 1987; Amano et al., 1988). It has also been suggested that altered glycosylation is induced in both (Y- and /?-subunits of hCG produced by choriocarcinoma (Endo et al., 1988). These observations permitted us to determine the nature of the structural differences of hormones synthesized in the cells (immature form) and secreted (mature form), as well as the uncombined subunits in normal trophoblasts and a choriocarcinoma cell line, BeWo, established by Patillo and Gey (1968). BeWo as well as other choriocarcinoma cell lines, JAR and JEG, and also other nontrophoblastic cell lines such as FOCUS human hepatoma have been shown to synthesize and secrete hormones and their free subunits (Pattilo and Gey, 1968; Ruddon et al., 1980; Ozturk et al., 1987; Endo et al., 1988). Molecular weights of hCG and its subunits in BeWo cells and culture fluid have been compared with standard hCG and its subunits by molecular sieve chromatography (Hussa, 1977, 1980). Recently, free P-subunit was purified from the culture fluid of BeWo cells using a specific monoclonal antibody (Thotakura and Bahl, 1986). However, comparative studies of synthesized and secreted hCGs by BeWo and normal placental cells, i.e., differences in molecular weight, structural differences of sugar chains, association characteristics of subunits and intracellular localization of hCG and its subunits, have not yet been performed. In order to clarify the mechanism of malignant alteration, it may be worthwhile to determine the structural characteristics of hCG and its subunits in choriocarcinoma and compare these

data to those of normal placental cells. In this paper, we describe differences in molecular weight, sugar chain structure and association characteristics of synthesized (immature form) and secreted (mature form) hCG subunits comparing them to those of normal placental cells. Materials and methods Materials The BeWo choriocarcinoma cell line was kindly provided by Dr. S. Imamura (Nagasaki University, School of Medicine). Urinary hCG, used as secreted mature hCG from placenta, was purified (12,000 IU/mg) in our laboratory from crude hCG (5000 IU/mg; Mochida Pharm. Co.) by the published method (Sakakibara et al., 1986). Antibodies against hCG (anti-hCG) and its subunits (anti-a and anti-p) were raised as described in Sakakibara et al. (1987a). Glycosidases were obtained as follows: endo-P-acetylglucosaminidase (from Streptomyces griseus, endoglycosidase H) and endo-cr-N-acetylgalactosaminidase (from AIcarigenes sphacerotilus) from Seikagaku Kogyo Co.; neuraminidase (from Clostridium perfringens) from Sigma Chemical Co. and N-glycanase (from Flauobacterium meningosepticum, peptide-N4[ Nacetyl-j3-glucosaminyl]asparagine amidase) from Genzyme Corp. Reagents were obtained as follows: N 6-2’-O-dibutyryladenosine 3’S’-cyclic monophosphate (db-CAMP) and theophylline from Sigma Co.; goat anti-rabbit IgG antibody from ICN; a low molecular weight calibration kit for sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) from Pharmacia; donkey anti-rabbit IgG ‘251-labeled whole antibody (9 pCi/pg) and Na[12’I] (17 Ci/mg) from New England Nuclear; phenylmethylsulfonyl fluoride (PMSF) from Boehringer Mannheim and Durapore (GVHP) filter from Millipore. All other chemicals were reagent grade compounds obtained from Techno. Suzuta Co. Placental cells First-trimester human placental sues, kindly provided by physicians elective abortion, were processed to tal cells and their cell lysates as Sakakibara et al. (1986).

choriomc tisat the time of obtain placendescribed by

147

Cell culture BeWo cells were cultured in Waymouth’s medium supplemented with 40% Gey’s balanced salt solution and 10% fetal bovine serum (WGF medium) at 37 o C in a 95% air, 5% carbon dioxide humified atmosphere in 25-cm* Corning plastic cell culture flasks. Routine subculturing was carried out every 7 days using 0.25% trypsin treatment for 5 min at 37” C to detach the cells. For experiments, the cells detached with trypsin treatment were washed once with WGF medium, then plated in 25-cm* plastic cell culture flasks (1.0 X lo6 cells per flask). The cells were subsequently allowed to attach to the flasks and grow to a monolayer sheet in WGF medium for 6-8 days. Thereafter, the cells were washed once with Dulbecco’s phosphate-buffered saline and were recultured in the same medium (5 ml) in the presence or absence of 1 mM db-CAMP and 1 mM theophylline for the indicated periods. The media were collected and the cells were washed twice with Hanks’ balanced salt solution and lysed by 20 mM Tris-HCl, pH 7.4 containing 1.0% Triton X-100 (0.4 ml) followed by sonication and centrifugation (at 15,000 rpm for 5 min, Tomy 15A). The resultant supernatant solution was used as the cell lysates. Samples for analyses by radioimmunoassay (RIA) and SDS-PAGE of synthesized (in the cells) and secreted (to the media) hCG and its subunits were prepared as follows. Aliquots of media and cell lysates were diluted to the appropriate concentrations for RIA. For SDS-PAGE, aliquots (10 ~1) of media, which had been concentrated 3-fold by ultrafiltration, and unconcentrated cell lysates were used as samples for electrophoresis. Radioimmunoassay of hCG [‘251]hCG was prepared by iodination with Na[‘*‘I] using the chloramine T method (Greenwood et al., 1963). The ‘251-labeled hCG had a specific activity of 11.6 pCi/pg and had about 60% of total counts precipitable with excess anti-hCG. Urinary hCG (12,000 IU/mg) served as a reference preparation for hCG RIA. For determination of the amounts of hCG in samples, tubes (0.4 ml total vol.) containing samples, [‘251]hCG (50,000 cpm/ml) and anti-hCG (0.1 pg/ml) in phosphate-buffered saline containing

0.1% bovine serum albumin (RIA buffer) were incubated for 3 days at 4’C. Then 10 ~1 of nonimmune rabbit IgG (1.2 mg/ml) and 0.3 ml of goat anti-rabbit IgG (0.6 mg/ml) were added and further incubated for 24 h at 4°C. The precipitates were washed twice with RIA buffer and radioactivities were measured. Glycosidase treatments Culture media and cell lysates obtained from BeWo cell cultures after 48 h, placental cell lysates and urinary hCG were digested with various exoand endoglycosidases as follows. Proteins in each sample were equivalent to 0.8 pg of hCG by RIA determination. For endoglycosidase H digestion, samples were pretreated with 2 mM PMSF and 100 U/ml of aprotinin for 30 min at 0°C in 0.15 M sodium citrate (pH 5.0), then 2 mU of endoglycosidase H was added and incubated for O-20 h at 37’C. Treatment of samples with neuraminidase (40 mu) was carried out at 37 ‘C for O-4 h in 0.15 M acetate buffer (pH 4.5) containing 0.45 M NaCl, 0.027 M CaC12, 1 mM PMSF and 100 U/ml aprotinin. For N-glycanase treatment, samples were denatured by boiling for 3 min in 0.5% SDS and 0.1 M 2-mercaptoethanol then incubated with enzyme (3 U) in 0.3 M sodium phosphate buffer (pH 8.6). Treatment with endo-cY-Nacetylgalactosaminidase (140 mu) was carried out at 37°C for 16 h in 0.2 M sodium citrate buffer (pH 4.5) after terminal sialic acids of the samples were removed by neuraminidase treatment under the same conditions as described above. SDS-PAGE and immunobinding analysis of a- and P-subunits and the combined form (hCG@) Proteins in samples, being equivalent to 0.2 pg hCG by RIA determination, were separated with SDS-PAGE according to the method of Laemmli (1970) using 14% polyacrylamide gel under either reducing or nonreducing conditions as described in a previous paper (Tominaga et al., 1989) with some modification. For determination of each subunit, culture medium and cell lysate were prepared under reducing condition by mixing with an equal volume of SDS-PAGE sample buffer (32 mM Tris-HCl, pH 6.8 containing 1% SDS, 2.2% 2-mercaptoethanol, 5% glycerol and 0.001% bromphenol blue) and boiling for 5 min. For

14x

determination of hCG@, culture media and cell lysates were prepared at 4 o C under nonreducing conditions by mixing with an equal volume of SDS-PAGE sample buffer without 2-mercaptoethanol. The separated proteins on the gel under reducing conditions were blotted directly onto a Durapore filter. On the other hand, separated proteins in the gel under nonreducing conditions were blotted onto a Durapore filter after the gel was incubated with a solution of 25 mM Tris-HCI, pH 8.3, containing 192 mM glycine, 0.1% SDS and 100 mM 2-mercaptoethanol for revealing antigenicities of (Y- and ,B-subunits composing hCGc@ on the gel against both anti-a and anti-p. Immunobinding analysis was performed as described in Sakakibara et al. (1986) using anti-a and anti-p followed by donkey anti-rabbit IgG ‘251-labeled whole antibodies. Subcellular fractionation of Be Wo cells BeWo cells were cultured for 48 h in the presence of 1 mM db-CAMP and 1 mM theophylline. The cells were detached by trypsin treatment and washed twice with ice-cold 0.25 M sucrose containing 20 mM Tris-HCl (pH 7.6) 50 mM KC1 and 2 mM MgCl 2. The cells were homogenized and fractionated subcellularly (fractions PA to PE and PP, see Fig. 1 in Sakakibara et al., 1987b) according to the method developed in placental cells. Fraction PB corresponds to the Golgi apparatus/smooth endoplasmic reticulum. Fractions PC and PD correspond to rough endoplasmic reticulum. Each fraction (PA to PE and PP) was analyzed by SDS-PAGE and an immunobinding method. Protein determination Protein determinations were carried out by the method of Lowry et al. (1951) using bovine serum albumin as the standard. Results hCG production of Be Wo cell BeWo cells were cultured ence or absence of db-CAMP the indicated period, amounts hCG in the cell lysates and

cultures for 72 h in the presand theophylline. At of immunoreactive t,he media were de-

(B)

,rlO 24

46 72 Incubation

8

24

Time

48

72

O

(h)

Fig. 1. Effects of db-CAMP on the production of hCG from BeWo cell cultures. The levels of hCC in cell lysates (A) and culture media (B) of control (0) and db-CAMP-treated cells (0) were determined by RIA. Each point represents the mean f SE of four separate experiments.

termined by RIA using ‘251-labeled hCG and antihCG. As shown in Fig. 1, db-CAMP with combination of theophylline stimulated hCG production time dependently until 48 h. After 48 h of culture, the amounts of hCG in the cell lysates and media increased 40-fold (9.7 f 1.0 IU/lOh cells) and 25fold (64.5 + 5.5 IU/lOh cells), respectively, over the control. To ascertain whether hCG subunits can be detected by immunobinding analysis, cell lysates and media prepared at each period were subjected to SDS-PAGE followed by immunobinding analysis. Immunoreactive (Y- and &subunits in the cell lysates and media could be detected as single protein bands after a culture period of 3 h, and the widths of those protein bands increased by stimulation with db-CAMP in a time-dependent manner until 48 h (data not shown), consistent with RIA data. hCG subunits of Be Wo cell cultures As shown in Fig. 2 and Table 1, apparent molecular weights of the hCG subunits of BeWo cell cultures were compared to those of placental cell lysates and urinary hCG. As previously reported (Sakakibara et al.. 1986), not only immature a-subunits (a part of 21 kDa) and immature P-subunits (23 and 19 kDa) but also their mature subunits, demonstrating the same molecular weights as urinary hCG subunits (21 and 31 kDa for (Y- and P-subunits, respectively), were detected

149

in lane 3 of Fig. 2 B protein band, since other bands by not petition experiments

in placental cell lysates. An immunoreactive band against anti-q having a lower molecular weight (19 kDa) than that of placental cells and urinary hCG (21 kDa), was detected in the cell lysates of BeWo cell cultures, named BeWo C-hCGa as a cellular form of the hCG a-subunit in BeWo cells (lane 2 in Fig. 2A). An i~unoreactive band against anti-a, having a slightly higher molecular weight (23 kDa) than the urinary hCG a-subunit (21 kDa), was detected in the media of BeWo cell cultures, named BeWo S-hCGLwas a secreted form of the hCG a-subunit from BeWo cells (lane 3 in Fig. 2A). Only a single immunoreactive band against anti-p, differing from placental cells by having a molecular weight (24 kDa) higher than both immature forms (23 and 19 kDa) but lower than the mature form (31 kDa) of the &subunits in placental cells was detected in cell lysates of BeWo cell cultures, named BeWo C-hCGj3 as a cellular form of hCG &subunit in BeWo cells (lane 2 in Fig. 2B). In addition, an immunoreactive band against anti-j3 having a slightly higher molecular weight (33 kDa) than the urinary hCG &subunit (31 kDa), was detected in the media of BeWo cell cultures, named BeWo S-hCG/? as a secreted form of hCG P-subunit from BeWo cells (lane 3 in Fig. 2B). A minor immunoreactive band

TABLE

may be a nonspecific reacted only this band differed from disappearing in immunocom(data not shown).

Sensitivities of hCG subunits of Be Wo cell cultures to gly~osidases

To obtain information about structures of sugar chains in hCG subunits of BeWo cell cultures, cell lysates and media were treated with various glycosidases and analyzed by SDS-PAGE followed by an i~unobin~ng method. Fig. 3 shows the results from endoglycosidase H treatment, in which the enzyme usually cleaves the bond of 4GlcNAcfil-4GlcNAcfll consisting of the core portion of a high-mannose type N-linked sugar chain. As previously reported (Sakakibara et al., 1986) and shown in lanes 1 and 2, immature subunits of placental cells (part of 21 kDa of the a-subunit and 23 and 19 kDa of the P-subunit) were digested by endoglycosidase H treatment, decreasing their molecular weights, but mature forms of the P-subunit (31 kDa) and a-subunit (a part of 21 kDa; it is very hard to detect as the remaining band in this picture, but it was clearly detected in the original film) were not digested. BeWo ChCGP, similarly to immature forms of placental

1

APPARENT MOLECULAR WITH GLYCOSIDASES

WEIGHTS

OF BeWo AND

PLACENTAL

hCG

Molecular weights were calculated from the data in Figs. 2-5 using a molecular marker proteins under the same conditions. Subunit

Intracelhtlar (-) a

immature

form (kDa)

Endo H b

Secreted mature (-) a

n-Subunit Placenta

21

14

21

BeWo

19

12

23

23,19 24

15 16

31 33

SUBUNITS

AND

weight calibration

THEIR

DIGESTED

curve obtained

by SDS-PAGE

form (kDa) N-gly .. ’

f

Neu d

Endo-a _

19 21

~-Subunit Placenta BeWo

f

22 23

Not treated with glycosidase. Endoglycosidase H digestion. N-glycanase digestion. Neura~nidase digestion. End~a-~-acetyl~~actosaminidase digestion. ’ Molecular weight of urinary hCG subunits which are the same as those of mature N.D.. not digested by glycosidase.

24 N.D.

26 29

a b ’ a ’

forms in placental

cells.

FORMS

e

of

150

(A)

12

3

1

123

I

2

3

z

Fig. 2. Detection of hCG subunits in BeWo cell cultures, placental cells and urinary hCG. The hCG subunits in the cell lysates (lane 2) and media (lane 3) prepared from BeWo cell cultures (48 h). cell lysates of placental cells (lane 1) and urinary hCG (lane 4) were analyzed by SDS-PAGE under reducing conditions followed by an immunobinding method for a-( A) and P-subunits (B). The molecular weight values indicate the migration positions of hCG subunits which are described in Sakakibara et al. (1986) as follows: 21 kDa, placental and urinary a-subunits; 31 kDa, placental mature and urinary /3-subunits; 23 and 19 kDa, placental immature P-subunits. BeWo C-hCGcu and C-hCG/3. intracellular forms of hCG subunits of BeWo cells; BeWo S-hCGcr and S-hCG/3. secreted forms of hCG subunits from BeWo cells.

cells, was completely digested by endoglycosidase H, decreasing its molecular weight as indicated by the open arrow on the right-hand side of lane 4 in Fig. 3B. BeWo C-hCGa was digested incompletely under these conditions (lane 4 in Fig. 3A). However, BeWo C-hCGa was completely converted to a new band, as indicated by the open arrow on the right-hand side of Fig. 3A, with further enzyme digestion. On the other hand, secreted hCG subunits from BeWo cells, BeWo S-hCGa and BeWo S-hCGP, as well as mature urinary hCG subunits were not digested by endoglycosidase H {data not shown). As shown in Fig. 4, mature forms of hCG subunits in the placental cells (part of 21 kDa of the a-subunit and 31 kDa of the P-subunit) as

4

Fig. 3. Endoglycostdaae H sensitivrty of hCG subunits in placental and BeWo cell lysates. The cell lysates were Incubated with (lanes 2 and 4) or without (lanes 1 and 3) endoglycosidase H for 20 h followed by SDS-PAGE under reducing conditions and rmmunobmding analysis for a- (A) and &subunits (B). Lanes 1 and 2. placental cell lysatea; lanes 3 and 4, BeWo cell lysates. Open arrows indrcate the positions of digested bands obtamed by enzyme digestion.

well as urinary hCG subunits (data not shown in this figure) were digested by the neuraminidase, which cleaves off the terminal sialic acid of both N-linked and O-linked sugar chains, decreasing their molecular weights as indicated by the open

(A)

BeWo S-hCGa

1

2

3

4

4eWo

;Q

1

2

3

S-hCG6

4

Fig. 4. Neuraminidase sensitivity of hCG subunits in placental cell lysates and the media of BeWo cell cultures. Placental cell lysates (lanes 1 and 2) and media of BeWo cell cultures (lanes 3 and 4) were incubated with (lanes 2 and 4) or without (lanes 1 and 3) neuraminidase for 4 h followed by SDS-PAGE under reducing conditions and immunobinding analysis for a- (A) and P-subunits (B). Open arrows indicate the positions of digested bands obtained by enzyme digestion.

151

Fig. 5. N-glycanase and endo-a-N-acetylgalactosaminidase sensitivities of urinary and BeWo hCG P-subunits. Urinary hCG (lanes l-4) and media of BeWo cell cultures (lanes 5-8) were incubated with (lanes 2-4.and 6-8) or without (lanes 1 and 5) glycosidases as described in Materials and Methods followed by SDS-PAGE under reducing conditions and immunobinding analysis for P-subunits. Lanes 2 and 6, incubation with N-glycanase; 3 and 7, incubation with neuraminidase; 4 and 8. successive incubations with neuraminidase and endo-a-N-acetylgalactosaminidase.

arrows on the left (lanes 1 and 2 in Fig. 4A and B). Similarly, secreted forms of hCG subunits from BeWo cells, BeWo S-hCGcy (lanes 3 and 4 in Fig. 4A) and BeWo S-hCGP (lanes 3 and 4 in Fig. 4B), but not intracellular forms of BeWo cells (data not shown) were digested by treatment of neuraminidase, slightly decreasing their molecular weights and increasing their immunogenicity as indicated by the open arrows on the right of Fig. 4. Glycosidase treatments with N-glycanase, which cleaves the bond of 4GlcNAcj31-Asn consisting of various types of N-linked sugar chains (high-mannose, hybrid and complex types), and endo-cr-N-acetylgalactosaminidase, which cleaves the bond of 3GalNAcal-Ser consisting of asialo

O-linked sugar chain, were performed as shown in Fig. 5. Urinary hCG P-subunit and BeWo S-hCG/? were digested with N-glycanase (lanes 1, 2, 5, and 6). The apparent molecular weight of a new protein band derived from a urinary hCG P-subunit as indicated by an open arrow on the left was 22 kDa. Digestion of BeWo S-hCG/? was incomplete under these conditions, since two new bands were detected at this time, but the upper protein band was converted to the lower protein band, having an apparent molecular weight of 23 kDa by further incubation with enzyme. Urinary hCG P-subunit could be digested by successive treatments with neuraminidase and endo-a-N-acetylgalactosaminidase (Fig. 5, lane 4) but BeWo S-hCG/? was not (lanes 4 and 8).

Fig. 6. Detection of assembled hCG subunits (hCGaS) and their free forms (free hCGa and free hCGB) in BeWo cell cultures, placental cells and urinary hCG. Urinary hCG (lanes 3 and 4). cell lysates of placental (lanes 5 and 6) and BeWo cells (lanes 7 and 8). and media from BeWo cell cultures (lanes 9 and 10) were subjected to SDS-PAGE under nonreducing conditions followed by immunobinding analysis using either anti-a (lanes 3, 5, 7 and 9) or anti-8 (lanes 4, 6, 8 and 10). As a control, urinary hCG was also subjected to SDS-PAGE under reducing conditions followed by immunobinding analysis using anti-a (lane 1) and anti-/3 (lane 2). The molecular weight values indicate the positions of urinary hCG& (SO kDa) and its dissociated subunits (31 and 21 kDa). a and a’, placental immature hCGe$s; b, free placental a-subunit; c and c’, free placental /&subunits; d, BeWo C-hCGn/3; e, free BeWo C-hCGP; f, BeWo S-hCGa/?; g, free BeWo S-hCGP.

1

2

I

2

Fig. 7. Assembly and secretion of BeWo hCG subunits. BeWo cells were cultured in the presence of db-CAMP for the indicated periods. Cell lysates (A and R) and media (C and D) were subjected to SDS-PAGE under nonreducing conditions followed by immunobinding analysis using anti-a (A and C) and anti-b (B and of_ Lane 1, 3 h: 2, 8 h; 3, 24 h; 4. 4X h: and 5. 72 h cultures.

Assembled and free forms of hCG subunits in Be Wo cell cultures As shown in lanes 3 and 4 of Fig. 6, urinary hCG& (assembled form of (Y- and P-subunits) was electrophoresed at the position of 50 kDa under nonreducing conditions and was detected as a common band by immunobinding analysis using both anti+ and anti-p. When placental cell lysates were subjected to SDS-PAGE under nonreducing conditions followed by immunobinding analysis, a minor band corresponding to 50 kDa and two major double protein bands (bands a and a’) were detected. 50 kDa bands and bands a and a’ must be mature hCGarp and immature hCG@s, respectively, since those bands reacted with both anti-a and anti-p. Free hCGa (band b) and free hCG/3 (bands c and c’) were also detected, since each band only reacted to anti-a or anti-p and their molecular weights were similar to those of their respective subunits. It is noteworthy that the apparent molecular weights of subunits, in particular the P-subunit, observed on the gel of SDSPAGE under nonreducing conditions usually appear to be slightly higher than those observed under reducing conditions. When cell lysates and media of BeWo cell cultures were analyzed, not only common protein bands (band d and band f) were recognized by both anti-a and anti-l) but

also uncommon protein bands (band e and band g) were recognized only by anti-& Band d and band e detected in the cell lysates must correspond to BeWo C-hCGa@ (assembled form of hCG subunits in BeWo cells) and free BeWo C-hCGP, respectively. In addition, a very weak band corresponding to band f was detected in the cell lysates. Similarly, band f and band g detected in the media must correspond to BeWo S-hCG&? (assembled form of hCG subunits secreted from BeWo cells) and free BeWo S-hCGP, respectively. Kinetics of assembly and secretion of BeWo hCG subunits As shown in Fig. 7, assembled and free forms of BeWo hCG subunits in cell lysates and media, obtained from BeWo cell cultures for 3.-72 h in the presence of db-CAMP, were analyzed by SDSPAGE under nonreducing conditions followed by immunobinding. The peak synthesis of total hCG was at 48 h (lane 4 of Fig. ?A and B) which was consistent with the results from RIA (Fig. 1A). The free BeWo C-hCG/3 and BeWo C-hCGa/3 were detected from 3 h (lane 1 in Fig. 78) and 24 h (lane 3 in Fig. 7A and B) cultures onwards, respectively. In the 48 h culture, a minor protein band having the same molecular weight as the secreted form of hCG@ (BeWo S-hCG&) was

153

Discussion

(A) BeUo C-N&t*

123456

111 (8) BeYo C-hCGII *

m

123436’ Fig. 8. Intracellular distribution of hCG subunits in BeWo cells. Subcellular fractionation was performed according to the previously reported method in Sakakibara et al. (1987b). Each subcellular fraction was subjected to SDS-PAGE under reducing conditions followed by immunobinding analysis for a- (A) and fi-subunits (E). Lanes l-5 and 6, fractions PA-PE and PP.

also detected in the cell lysates (lane 4 in Fig. 7A and B). In the 72 h culture, the amount of total hCG decreased, which was consistent with the results from RIA. In the media (Fig. 7C and D), total hCG increased time dependently until 72 h, which was also consistent with the results from RIA (Fig. 1B). BeWo S-hCGc@ and free BeWo S-hCGP were detected in media from the 24 h culture. On the other hand, no free form of the a-subunit was detected in either the cell lysates or in the media under these culture conditions. Intracellular localization of BeWo hCG subunits The BeWo cells, cultured for 48 h in the presence of db-CAMP, were fractionated subcellularly. As previously reported by Sakakibara et al. (1987b), immature placental hCG subunits were detected in fractions PC and PD, corresponding to rough endoplasmic reticulum. Mature placental hCG subunits were detected in fraction PC corresponding to the Golgi apparatus. As shown in Fig. 8, immature hCG subunits in BeWo cells (BeWo c-hCGa and C-hCGj3) were detected mainly in fraction PD. However, in fraction PB no protein bands corresponding to immature or mature BeWo hCG subunits were detected under these conditions.

The purpose of the present study was to clarify malignant alterations involved in hCG by comparing the molecular characteristics of hCG subunits between choriocarcinoma, BeWo cells, and normal placental cells using immunobinding analysis. For detection of hCG subunits in BeWo cell cultures by the immunobinding method, it was necessary to increase the synthesis and secretion of hCG subunits. Hussa et al. (1974) have reported that db-CAMP with or without combination of theophylline resulted in a marked stimulation of hCG subunits. When BeWo cells were cultured in the presence of 1 mM db-CAMP in combination with 1 mM theophylline, the amounts of synthesized and secreted hCG subunits increased time-dependently and were sufficient for hCG subunits to be detected by immunobinding analysis; however, the amount of intracellular hCG decreased after culture for 72 h. This may have resulted from degradation of hCG in the cells or acceleration of the secretion rate. Therefore, the cell lysates and media of BeWo cell cultures after 48 h were used in the various experiments. Both the (Y- and P-subunits of hCG are glycoproteins. Previous reports indicated that both subunits exist predominantly as immature intermediates having N-linked high-mannose type sugar chains in placental cells, and they appear to be secreted after maturation with alteration of their sugar chains to a complex type by the actions of various glycosyltransferases (Ruddon et al., 1981a; Sakakibara et al., 1987b). In addition, Olinked sugar chains are associated with maturation of the P-subunit. In this paper, we have described the molecular characterization of hCG and its subunits of BeWo cell cultures, comparing them to those of normal placental cells and urinary hCG. Apparent molecular weights of hCG subunits obtained in these experiments are summarized in Table 1. Results concerning BeWo hCG P-subunit are as follows: the BeWo cells differed from placental cells in that only one species of intracellular P-subunit (BeWo C-hCGP) was detected, the molecular weight (24 kDa) of which was higher than both immature forms of placental cells (23 and 19 kDa); BeWo C-hCGP as well as immature forms of placental cells were sensitive to endogly-

154

cosidase H digestion and the decrease in those molecular weights were both 8 kDa (BeWo ChCG/3: from 24 to 16 kDa and placental immature P-subunit: from 23 to 15 kDa); the secreted BeWo hCG P-subunit (BeWo S-hCGP, 33 kDa) also showed a higher molecular weight than the urinary hCG /!&subunit (31 kDa), and the BeWo S-hCGP as well as the urinary hCG P-subunit were sensitive to neuraminidase but not endoglycosidase H digestion, although the decrease in their molecular weights by neuraminidase digestion was not similar. These results suggest that BeWo C-hCGp was an immature form having N-linked high-mannose sugar chain(s) and BeWo S-hC’G/I was a mature form having terminal sialic acid(s) similar to placental and urinary subunits, respectively. It is also suggested that the structures of N-linked sugar chains in the BeWo C-hCGP and immature forms of placental cells are very similar, since the decrease of those molecular weights by endogiycosidase H digestion was almost the same. Nowever, the molecular weights of original and endoglycosidase H-digested BeWo C-hCGP were higher by approximately 1 kDa than each respective subunit in placental cells. This may suggest that BeWo C-hCG/? has either an extra sugar chain or an additional polypeptide chain, although we did not determine the molecular weight of the polypeptide moiety of the P-subunit, i.e., by using a cell-free protein synthesis system. On the other hand, the difference in molecular weight between the mature form of BeWo and the urinary hCG ~-subunjt (BeWo S-hCG/I and urinary hCG P-subunit; a difference of 2 kDa) was larger than that between immature forms (1 kDa). However, the difference between mature forms after N-glycanase treatment (1 kDa) was close to that between immature forms after endoglycosidase H digestion. These results suggest that the structure of N-linked sugar chain(s) in l3eWo S-hCGj3 is different from that of the normal hCG P-subunit, i.e., it may have a triantennary structure as reported by Mizuochi et al. (1983). Moreover, it is suggested that BeWo S-hCGj3 has either an unusual structure of Q-linked sugar chain(s) or none at all, since BeWo S-hCGj3 was not digested by endo-a-N-acetylgalactosaminidase. Meanwhile, from the results regarding the BeWo a-subunit it is suggested that the intracellular form

of the BeWo hCG a-subunit (BeWo C-hCGa) is an immature form having a high-mannose sugar chain(s) like the intracellular n-subunit of placental cells, since both a-subunits were digested by endoglycosidase H and decreases in their molecular weights were very similar. The molecular weight of BeWo C-hCGa (19 kDa) was lower than the immature form of placental cells (21 kDa) by about 2 kDa and the difference in molecular weight between both a-subunits was not altered by endoglycosidase H digestion. This may suggest that the polypeptide moiety of BeWo C-hCGa is shorter than that of placental cells, although we did not compare molecular weights of both apoproteins on SDS-PAGE. On the other hand, contrary to BeWo C-hCGcu. the secreted a-subunit of BeWo cells (BeWo S-hCGcu, 23 kDa) showed a higher molecular weight than the mature form derived from placental cells (urinary hCG n-subunit. 21 kDa). This may suggest that processing of N-linked sugar chain(s) in BeWo cells is different from that of normal placental cells. Mizuochi et al. (1983) have reported that more than 97% of the sugar chains of hCG purified from urine of patients with choriocarcinoma was free from sialic acid, while the sugar chains of normal hCG were mostly sialylated. As cited above, BeWo S-hCG subunits have terminai sialic acid(s) indicating the sialylation in a choriocarcinema cell line, BeWo, is similar to normal placental cells rather than those from patients with choriocarcinoma. hCG consists of noncovalently bonded a- and P-subunits. hCGnP (assembled form of a- and ,&subunits) as well as free subunits are secreted in vivo and in vitro. To clarify whether c$3-dimer formation is occurring in BeWo cells and both assembled hCG and free subunits are secreting into the media, we analyzed assembled and free forms of mature and immature subunits by SDSPAGE under nonreducing conditions. It is suggested that intracellular as well as secreted BeWo hCG subunits are assembled in the cells and media and that free ~-subunits are present there, as well. Free a-subunits, if present, were very few, as we were unable to detect any. It is also suggested, from kinetic data, that synthesis and secretion of the assembled hCG and the free P-subunit occur independently. This observation may indicate a

155

biological action of free BeWo S-hCGP itself. Peters et al. (1984) reported that uncombined (Yand P-subunits were detected in both the cell lysates and the media of choriocarcinoma (JAR) cell cultures and the uncombined a-subunit rather than the P-subunit was secreted. However, it was reported that a much greater amount of free psubunits was detected in a choriocarcinoma patient’s sera (Ozturk et al., 1988). That is, the synthesis and secretion of hCG of BeWo cell cultures may be reflected in native choriocarcinema cells. In the BeWo cells, in contrast to placental cells, only trace amounts or none of mature hCG subunits were detected. As we reported previously, mature and immature subunits of placental cells were localized in Golgi apparatus-rich and rough endoplasmic reticulum-rich fractions, respectively (Sakakibara et al., 1987b). The immature subunits of BeWo cells were detected in the rough endoplasmic reticulum-rich fraction, similar to the immature subunits of placental cells, but no mature subunits of BeWo cells were detected in any subcellular fraction. These results are in agreement with those of Yorde et al. (1979), who suggested that hCG subunits appear to be secreted from BeWo cells without the formation of secretory granules, coinciding with the immunocytochemical observations, and also those of Ruddon et al. (1981b), who reported that no mature hCG subunits were stored in JAR cells. Moreover, Hanover et al. (1982) have indicated that completely glycosylated forms of hCG subunits cannot be detected intracellularly in BeWo cells during pulse-chase experiments. Although it is clear from the current experiments that the molecular characteristics of BeWo hCG and its subunits (molecular weights, presumable structures of sugar chains) are obviously different from those of normal placental cells, it is still unclear, in particular, why the variation of sugar chain structures occurs in choriocarcinoma cells. This must result from alterations of glycosyltransferases involved with processing the sugar chains. For instance, it is supposed that 0-GalNAc transferase activity may be deficient or very low in BeWo cells, since BeWo C-hCGP did not bind O-linked sugar chains judging from the sensitivity of it to endo-a-N-acetylgalactosaminidase. It may

be worthwhile to determine if such an alteration the enzyme level is present.

at

Acknowledgements We wish to thank Drs. S. Imamura and S. Okamoto for useful advice. We also thank Drs. M. Matsumoto and S. Imamichi for supplying the aborted placentas. We are grateful to Drs. T. Yamamoto and K. Tokuyasu, Seikagaku Kogyo Co., for providing the glycosidases. This work was supported by a Grant-in-Aid from the Ministry of Education and Culture of Japan. References Amano, J., Nishimura, R., Mochizuki, M. and Kobata, A. (1988) J. Biol. Chem. 263, 1157-1165. Ashitaka, Y., Nishimura, R., Futamura, K., Ohashi, M. and Tojo, S. (1974) Endocrinol. Jpn. 21, 547-550. Bellisario, R., Carlsen, R.B. and Bahl, 0m.P. (1973) J. Biol. Chem. 248, 6796-6809. Carlsen, R.B., Bahl, 0m.P. and Swaminathan, N. (1973) J. Biol. Chem. 248, 6810-6827. Cole, L.A. (1987) J. Clin. Endocrinol. Metab. 65, 811-813. Endo, Y., Yamashita, K., Tachibana, Y., Tojo, S. and Kobata, A. (1979) J. B&hem. 85, 669-679. Endo, T., Nishimura, R., Kawano, T., Mochizuki, M. and Kobata, A. (1987) Cancer Res. 47, 5242-5245. Endo, T., Nishimura, R., Mochizuki, M., Kochibe, N. and Kobata, A. (1988) J. Biochem. 103, 1035-1038. Greenwood, F.C., Hunter, W.M. and Glover, J.S. (1963) Biothem. J. 89, 114-123. Hanover, J.A., Elting, J., Mintz, G.R. and Lennarz, W.J. (1982) J. Biol. Chem. 257, 10172-10177. Hussa, R.O. (1977) J. Clin. Endocrinol. Metab. 44, 1154-1162. Hussa, R.O. (1980) J. Clin. Endocrinol. Metab. 50, 5-9. Hussa, R.O., Story, M.T. and Pattillo, R.A. (1974) J. Clin. Endocrinol. Metab. 38, 338-340. Kessler, M.J., Reddy, M.S., Shah, R.H. and Bahl, 0m.P. (1979a) J. Biol. Chem. 254, 790-7908. Kessler, M.J., Mise, T., Ghai, R.D. and Bahl, 0m.P. (1979b) J. Biol. Chem. 254, 7909-7914. Laemmli, U.K. (1970) Nature 227, 680-685. Lowry, O.H., Rosebrough, N.J., Farr, A.L. and Randall, R.J. (1951) J. Biol. Chem. 193, 265-275. Mizuochi, T., Nishimura, R., Derappe, C., Taniguchi, T., Hamamoto, T., Mochizuki, M. and Kobata, A. (1983) J. Biol. Chem. 258, 14126-14129. Ozturk, M., Bellet, D., Isselbacher, K. and Wands, J. (1987) Endocrinology 120, 559-566. Ozturk, M., Berkowitz, R., Goldstein, D., Bellet, D. and Wands, J.R. (1988) Obstet. Gynecol. 158, 193-198. Pattillo, R.A. and Gey, G.O. (1968) Cancer Res. 28,1231-1236.

156

Peters, B.P., Krzesicki, R.F., Hartle, R.J.. Perini, F. and Ruddon, R.W. (1984) J. Biol. Chem. 259. 15123-15130. Ruddon, R.W., Anderson, C. and Meade-Cobun. K.S. (1980) Cancer Res. 40. 4519-4523. Ruddon, R.W., Hartle, R.J., Peters, B.P., Anderson, C., Huot. RI. and Stromberg, K. (1981;) J. Biol. Chem. 256. 11389-11392. Ruddon, R.W., Bryan, A.H., Hanson, C.A., Perini, F.. Ceccorulli, L.M. and Peters, B.P. (1981b) J. Biol. Chem. 256, 5189-5196. Sakakibara, R.. Tominaga, N. and Ishiguro. M. (1986) Biothem. Biophys. Res. Commun. 137, 443-452.

Sakakibara, R., Tominaga, N.. Sakai, A. and Ishiguro. M. (1987a) Anal. Biochem. 162, 150-155. Sakakibara, R.. Yokoo, Y., Yoshikoshi, K., Tominaga, N.. Eida. K. and Ishiguro. M. (198%) J. Biochem. 102. 993-1001. Thotakura, N.R. and Bahl, 0m.P. (1986) Endocrinology 119. 1887-1894. Tominaga. N., Sakakibara, R.. Yokoo, Y. and Ishiguro, M. (1989) J. Biochem. 105, 992-997. Yorde. D.E.. Hussa. R.O., Garancis. J.C. and Patti&x R.A. (1979) Lab. Invest. 40, 391-398.