Characterization of post-translational modifications common to three primary mesenchyme cell-specific glycoproteins involved in sea urchin embryonic skeleton formation

DEVELOPMENTAL

BIOLOGY

150,29&305

(19%)

Characterization of Post-translational Modifications Common to Three Primary Mesenchyme Cell-Specific Glycoproteins Involved in Sea Urchin Embryonic Skeleton Formation BRUCE KABAKOFF, SHENG-PING Department

of Biochemistry

L. HWANG, AND WILLIAM

J. LENNARZ~

and Cell Biology, State University of New York at Stony Brook, Stony Brook, New York 1179,&5215 Accepted December 11, 1991

Previous studies have established the importance of a complex, N-linked oligosaccharide chain, recognized by a monoclonal antibody (mAb 1223), in the formation of sea urchin embryonic skeletal components known as spicules. To further investigate the function of this epitope, mAb 1223 was added to primary mesenchyme (PM) cell cultures prior to spiculogenesis. The antibody did not inhibit cell migration, cell attachment, or synthesis of the filapodial networks upon which the spicules are deposited. However, it did block deposition of mineralized CaCOs along these filapodia, strongly supporting the previously proposed role for the 1223 epitope in calcium accumulation and/or deposition. Previously the 1223 epitope has been most extensively studied in association with a mesenchyme-specific protein of 130 kDa (msp 130). It has now been established, by Western blot analysis of whole embryo and PM cell extracts using mAb 1223, that two other proteins of 205 and 250 kDa contain the 1223 epitope. A study of the developmental profiles of expression of these glycoproteins revealed that all three were first expressed just prior to spiculogenesis, consistent with a role for any or all of these proteins in this process. Additionally all three proteins incorporated ethanolamine, myristate, and palmitate, the precursors of the glycosylphosphatidylinositol (GPI) anchor. Further labeling studies revealed differences in the metabolic lability of the GPI anchor in the three proteins; pulse-chase studies demonstrated that the ethanolamine moiety was stable in msp 130, but was rapidly chased from the 205-kDa protein (T,,z = 14 hr). Phosphatidylinositol-specific phospholipase C partially released (50%) msp 130 from the PM cell surface, whereas it had no effect on release of the 205- and 250-kDa proteins. Studies with %SO, labeling and PNGase F treatment directly established that all three proteins are sulfated, and that most of the sulfate is attached to the N-linked oligosaccharide chains. Thus, the three major mAb 1223-reactive glycoproteins in PM cells are also the three major proteins containing both sulfated N-linked oligosaccharide chains and GPI anchors. Further investigation of this intriguing correlation may help to define the o 1992 Academic PWSS, IIN. precise function of the 1223 epitope in the process of spicule formation. INTRODUCTION

During the course of the studies in our laboratory on the genesis of skeletal components (spicules) of the sea urchin embryo, it was found that a marked increase in the level of N-linked glycoprotein synthesis occurs just prior to the onset of gastrulation, an event that precedes spiculogenesis (Lennarz, 1983). Furthermore, it was demonstrated by using either an inhibitor of assembly of N-linked oligosaccharide chains (tunicamycin) or an inhibitor of dolichol synthesis (mevinolin) that synthesis of such N-linked glycoproteins was required for both gastrulation and subsequent spicule formation (Schneider et al, 1978; Carson and Lennarz, 1979,198l). In contrast to these observations, indicating a common requirement for de novo synthesis of N-linked glycoproteins in both of these morphogenetic events, in recent studies we found that the processing of high-mannosetype glycoproteins to complex-type glycoproteins is nec1 To whom correspondence should be addressed. 0012-X06/92

$3.00

Copyright 0 1992 by Academic Press, Inc. All rights of reproduction in any form reserved.

294

essary for the process of spiculogenesis, but not for gastrulation (Kabakoff and Lennarz, 1990). In the process of studying the possible role of specific glycoproteins in skeletogenesis, our laboratory (Carson et aZ., 1985) and others (McClay et al, 1985; Anstrom et al, 1987) generated monoclonal antibodies directed toward epitopes found to be localized to the primary mesenchyme (PM)’ cells, the cell type that is involved in embryonic spicule formation. All three antibodies recognized epitopes on a highly immunogenic 130-kDa protein designated as mesenchyme-specific protein 130 (msp 130). Further study in our laboratory with the monoclonal antibody (mAb) 1223 established that its epitope on the protein was a complex N-linked oligosaccharide chain (Farach-Carson et G!., 1989). More recent

2 Abbreviations used: CFSW, calcium-free seawater; GPI, glycosylphosphatidylinositol; HBS, Hepes-buffered saline; mAb, monoclonal antibody; msp 130, mesenchyme-specific protein 130, PIPLC, phosphatidyl inositol-specific phospholipase C, PM, primary mesenchyme; PNGaseF, peptide:N glycosidase F.

KABAKOFF,

HWANG,

AND LENNARZ

Common Modijkatims

studies with mAb 1223 demonstrated that it was found not only in msp 130, but also on at least one other PM cell glycoprotein of 205 kDa (Kabakoff and Lennarz, 1990) in embryo cultures. When the monoclonal antibody 1223 or a Fab fragment derived from it was added to PM cell cultures, spicule elongation was inhibited (Carson et al, 1985), suggesting that one or more of these proteins containing the mAb 1223-reactive complex oligosaccharide chains played an essential role in spicule formation. There is now considerable information about the structure of msp 130. Anstrom et al. (1987) provided indirect evidence to suggest that it is a sulfated glycoprotein. Recently, Parr et al. (1990) reported the cloning and deduced amino acid sequence of msp 130. They concluded that the protein has six potential N-linked glycosylation sites. In addition, based on the presence of a hydrophobic C-terminus and the sensitivity of the protein to phosphatidylinositol-specific phospholipase C (PIPLC) they concluded that msp 130 contained a glycosylphosphatidylinositol (GPI) anchor. Because the glycoprotein lacks a hydrophobic transmembrane domain it was suggested that the GPI moiety served to anchor it to the plasma membrane. Because of the finding that binding of antibody to the 1223 epitope blocked spiculogenesis, and because only the msp 130 has been characterized to any extent, we have extended the studies to examine the developmental profiles of two other PM cell-specific glycoproteins that we have found to contain the 1223 epitope, a 205- and a 250-kDa glycoprotein. We have shown that all three immunoreactive glycoproteins are first detectable just prior to spicule formation, which is consistent with the possibility that any or all of these proteins participate in some step in skeletogenesis. Additionally, because recent reports in other biological systems have shown the GPI anchor to be developmentally regulated (Hortsch and Goodman, 1990; Field et aZ., 1991), we further studied the GPI anchor associated with msp 130, and investigated the possibility that such anchors are also present in the 205- and 250-kDa glycoproteins that were found to contain the 1223 epitope. Embryos and PM cell cultures were incubated with various labeled precursors of the GPI anchor and incorporation into glycoproteins was determined. Complementary to these radiolabeling studies, experiments were carried out utilizing enzymatic treatment with PIPLC or peptide:N-glycosidase F (PNGase F). All of the results show that all three glycoproteins bearing the 1223 epitope contain GPI-like groups; they can be radiolabeled with palmitate, myristate, and ethanolamine. However, only in the case of msp 130 is the GPI-like anchor sensitive to PIPLC in intact PM cells. Additionally, the results with inorganic sulfate radiolabeling, immunoaffinity chromatography,

295

of Three Glywproteins

and PNGase F treatment show directly that all three glycoproteins are sulfated and that most of the sulfate residues are located on the N-linked oligosaccharide chains. MATERIALS

AND

METHODS

Materials Strongylocentrotus purpuratus sea urchins were purchased from Marinus (Long Beach, CA). Artificial seawater (Instant Ocean) was obtained from Aquarium Systems (Mentor, OH). [l-3H]Ethanolamine hydrochloride (19-30 Ci/mmole), Nal%I (carrier free), and m&2‘Hlinositol(l7.4 Ci/mmole) were purchased from Amersham Corp. (Arlington Heights, IL). En3hance, N+?SO, (carrier free), [9,10-‘Hlpalmitic acid (60 CVmmole), and [9,10-3H]myristic acid (40 CVmmole) were from New England Nuclear Corp. (Boston, MA). Preimmune mouse IgG and goat anti-mouse IgG were supplied by Cappel Laboratories (Malvern, PA). Goat anti-mouse IgG was iodinated by the chloramine T method (Review 18, Amersham Corp.). PIPLC was purchased from Boehringer-Mannheim (Indianapolis, IN) and recombinant PNGase F from Genzyme Corp. (Boston, MA). Horse serum, penicillin-streptomycin, and gentamycin sulfate were obtained from GIBCO Laboratories (Grand Island, NY). Ticarcillin disodium, erythromycin, 3amino-1,2,4 triazole, fatty acid-free bovine serum albumin, and ethanolamine were from Sigma (St. Louis, MO). All other chemicals were of reagent grade. Analysis of Sea Urchin Embryo

Proteins

For labeling experiments with embryos, eggs were collected and fertilized as described previously (Heifetz and Lennarz, 1979). The embryos were cultured as 2% suspensions in 5 ml of seawater at 14°C in Costar dishes containing six 9.5-cm2 wells (Cambridge, MA). The labeled compounds were added at a concentration of 100 &i/ml for myristate and palmitate. The fatty acids were first dissolved in seawater with 10% BSA and sonicated in a water bath for 15 min before addition of 1110th vol to the culture media. This procedure ensured uniform dispersion of the lipids. For ethanolamine and inositol, the radiolabeled components were added to a final concentration of 50 and 30 &i/ml, respectively. Aliquots were removed at different stages of development and the embryos washed three times with seawater. The embryos were extracted and the proteins subjected to SDS-PAGE (8.75% acrylamide) as described previously (Kabakoff and Lennarz, 1990). The gels were subsequently treated with En3hance, rinsed, dried, and exposed to AR film (Kodak, Rochester, NY) at -70°C.

296

DEVELOPMENTALBIOLOGY

Micromeres were isolated and cultured as described by Kabakoff and Lennarz (1990). Media containing 100 &i/ml of palmitate or myristate were prepared and used as described above. Inositol was added to a final concentration of 50 &Yrnl and ethanolamine to 25 &i/ ml in normal culture media. Sulfate was added (100 &i/ ml) to culture media containing l/lOth the normal concentration of sulfate. The cells were harvested by incubation in calcium-free seawater (CFSW) with 1 mM EGTA to detach the cells and washed three times with CFSW prior to protein extraction, electrophoresis, and fluorography as described above. For pulse-chase experiments with ethanolamine, the label was added at 48 hr after fertilization and incubated for 12 hr. At this time the culture medium was removed and replaced with fresh medium containing unlabeled ethanolamine (1 mM). The cells were then harvested at the times after chase indicated in the figure legends. Densitometric analysis of fluorograms was conducted with a LKB (Bromma, Sweden) Model XL Laser Densitometer. For Western blot analysis of embryo and PM cell protein extracts, the protocols for protein extraction, electrophoresis, transfer, and immunoblotting were as described by Kabakoff and Lennarz (1990). Immunoafinity

Chromatography

MAb 1223 or preimmune mouse IgG were coupled to Affigel-10 (Bio-Rad, Richmond, CA) in a 0.1 1M Hepes, pH 7.5, buffer containing 80 mlM CaCI,. The coupling was allowed to proceed overnight at 4°C. As judged by the Bio-Rad protein assay prior and subsequent to coupling, both couplings occurred at greater than 95% efficiency and resulted in gels with about 4 mg IgG per milliliter of packed gel. The gels were washed exhaustively in coupling buffer and then unreacted active groups blocked by treatment with 1 Mglycine ethyl ester, pH 8.0, for 1 hr. The gels were then washed exhaustively with buffer containing 10 milf Hepes, pH 7.2,140 m&f NaCI, and 0.1% Triton X-100 (HBS). Aliquots of protein extracts were incubated with 100 ~1 of packed immunoaffinity gel overnight at 4°C. The samples were then centrifuged gently and the supernatant was removed. The pellet was washed twice with 200 ~1 of the HBS. The combined supernatant fractions were treated with 5 vol of acetone to precipitate the protein, and the resulting pellet was resuspended in sample buffer for electrophoresis. It is referred to as the unbound fraction in the text. The gel was washed three times with the same Hepes buffer containing 0.5 MNaCl to remove nonspecifically adsorbed protein and three times with icecold Hz0 to remove the excess salt. The gel was then resuspended in two times sample buffer, heated at 100°C for 10 min, and centrifuged and the supernatant

VOLUME

150,19%

removed for electrophoresis. as the bound fraction. PIPLC

This fraction

is referred to

Digestion

PIPLC digestions of PM cell suspensions were carried out as described by Parr et al. (1990) for whole embryo cultures. The PM cells were grown as described and harvested at ‘72 hr after fertilization. The cells were rinsed twice with calcium- and magnesium-free seawater, twice with buffer containing 0.1 M Hepes, pH 7.8, and 0.35 M glycine, and then resuspended in the Hepes/glytine buffer. The incubation was initiated by addition of 0.5 units of PIPLC and was allowed to proceed for 2 hr at 37°C. The cells were pelleted by centrifugation at 16,OOOg for 5 min in a microfuge. The supernatant was removed and the proteins precipitated with acetone. Both the cell pellet and acetone pellet were redissolved by heating at 100°C for 5 min in SDS-PAGE sample buffer. PNGase

F Digestion

PNGase F digestions were performed as described by Roux et al. (1988) on extracts from either unlabeled or sulfate-labeled PM cell cultures. Protein samples were acetone precipitated and resuspended by heating at 100°C for 5 min in 100 ~1 of lysis buffer consisting of 50 mlM Tris-HCl, pH 7.5,0.1 M fl-mercaptoethanol, and 1% SDS. To this mixture was added 50 ~1 of 0.5 MEDTA, 300 ~1 of H,O, and 50 ~1 of concentrated PNGase F buffer consisting of 0.2 M Tris-HCl, pH 7.5, 0.2 M P-mercaptoethanol, and 10% NP-40 detergent. The reaction was begun by addition of 0.3 units (1.2 ~1) of recombinant PNGase F and the incubation was allowed to proceed overnight at 3’7°C. The samples were then heated at 100°C for 5 min and cooled to 37’C, and an additional 0.3 unit aliquot of enzyme was added. For Western blot analysis of unlabeled proteins, aliquots were removed at the stated times after initial enzyme addition for SDSPAGE. For sulfate labeled material, the incubation was allowed to proceed for 24 hr and then an aliquot was removed for fluorography as described. RESULTS

Although most studies on the mAb 1223~reactive oligosaccharide chains on glycoproteins of the micromere lineage have focused on the 130~kDa glycoprotein, msp 130, Western blot analysis has revealed that two other glycoproteins bear this epitope. Thus, a Western blot with mAb 1223 as the probe revealed that in gastrula stage (52 hr) embryos, in addition to msp 130, one other major immunoreactive glycoprotein at 205 kDa was detectable (Fig. 1A). Upon analysis of the proteins from

KABAKOFF,

A

HWANG,

AND LJSNNARZ

Common Mod$icutim

B

205+ 130-9 kDa

FIG. 1. Western immunoblot analysis of gastrula stage embryos and PM cells. Whole embryo cultures (A) and PM cell cultures (B) were harvested at 54 and ‘72 hr, respectively, after fertilization. Proteins were extracted and immunoblots performed as described under Materials and Methods.

micromere-derived PM cells cultured in vitro for ‘72 hr after fertilization, where one might expect a 20-fold enrichment for mesenchyme proteins relative to total embryonic proteins, an additional immunoreactive glycoprotein of 250 kDa was detected (Fig. 1B). Although the relative level of the 250-kDa protein is quite high in this particular preparation of cells, we have observed that the relative level varies considerably depending on the culture. Consistent with this observation we occasionally can detect the 250-kDa protein in immunoblots from whole embryo protein extracts (data not shown). Given that immunofluorescence microscopy of embryos earlier showed that the 1223 epitope is unique to the primary mesenchyme cells (Carson et a& 1985), it appears that all three of these glycoproteins are localized to this cell type. A key observation implicating the glycoproteins bearing the 1223 epitope in spiculogenesis was that addition of mAb 1223 to PM cell cultures containing spicules prevented their further elongation (Carson et a& 1985). This inhibitory effect was observed using either intact antibody or Fab fragments. We have now studied this effect in more detail, using cells that have not yet initiated formation of spicules. The results in Fig. 2 reveal that addition of mAb 1223 at 36 hr completely blocks spicule formation. As shown in Figs. 2A and 2B, addition of either 10 pg/ml of preimmune IgG or 1 pg/ml mAb 1223 had no effect on subsequent spicule formation. In contrast, cultures which received either 5 pg/ml (data not shown) or 10 pg/ml (Fig. 2C) of mAb 1223 formed no spicules. Examination of the mAb 1223treated (10 pg/ml) cultures at higher magnification

of Three Glycoproteins

297

(Fig. 2E) revealed that the cells did not detach or aggregate, which would have been suggestive of a role for the epitope in cell adhesion or cell migration. Instead, they formed the extensive filopodial networks that normally appear prior to spiculogenesis (Decker and Lennarz, 1988). But unlike the preimmune IgG-treated cell cultures (Fig. 2D), they did not proceed beyond this stage and deposit pseudocrystalline CaCO,. These findings are strongly supportive of the hypothesis that the glycoproteins associated with the 1223 epitope are involved in calcium ion accumulation and/or CaCO, deposition during the process of skeletogenesis. Because we found that (1) the 1223 epitope is present on three glycoproteins and (2) this epitope is important in CaCOs deposition, we measured the level of these glycoproteins with respect to the time course of embryonic development. The relative level of expression of both the msp 130 and 205-kDa glycoproteins averaged from results obtained with six individual embryo cultures is shown in Fig. 3. It is obvious that there is a considerable difference in the appearance and turnover of the two glycoproteins, at least with respect to the expression of the 1223 epitope. Although both immunoreactive glycoproteins are detected at late gastrula stage (50-55 hr) prior to the appearance of nascent spicules they subsequently differ in expression. The msp 130 protein first appears at mesenchyme blastula stage, peaks at late gastrula, and then declines to barely detectable levels by the larval stage (96 hr); the 205-kDa protein is absent at mesenchyme blastula stage, only at roughly 60% of its maximal amount at late gastrula, and continues to rise throughout the course of development. The turnover of the 1223 epitope in msp 130 parallels the turnover of the msp 130 protein itself as determined by Western blots utilizing a monoclonal antibody (mAb A,; kindly provided by Dr. R. Raff) to the polypeptide chain (data not shown). It is not possible at this time to make such an assessment with respect to the 205-kDa protein because an antibody to the polypeptide backbone is not yet available. The results of limited studies using PM cell culture (data not shown) revealed that the initial pattern of appearance of the msp 130 and the 205-kDa glycoproteins was quite similar to that observed in whole embryos. Both were detected prior to spicule formation and, as observed in embryo cultures, the 205-kDa protein continued to increase until the end of the time of culture. However, the pattern of appearance for the msp 130 in PM cell cultures diverged from that observed in whole embryos. In both systems msp 130 reached a maximum at the onset of spiculogenesis. In cultured cells it then remained at a constant level, whereas in embryos it rapidly declined. The time course of appearance of the 250kDa glycoprotein, which could only be consistently de-

FIG. 2. Effects of mAb 1223 on spicule formation in PM cell cultures. PM cells were isolated from embryos and cultured as described under Materials and Methods. At 36 hr after fertilization, cultures were supplemented with 10 pg/ml of preimmune IgG (A,D), or 1 fig/ml (B) or 10 pg/ml (C,E) of mAb 1223. The cultures were continued until 96 hr after fertilization when the above micrographs were recorded. Bar (A-D), 60 pm. Bar (E,F), 5 pm.

tected in these cell cultures, closely resembled that of msp 130. Clearly, the developmental profiles observed for the three glycoproteins bearing the 1223 epitope are consistent with any or all of them being involved in spicule formation. In view of the above findings of these glycoproteins bearing the 1223 epitope and the presence of a GPI anchor in msp 130 (Parr et al, 1990), it became of interest to determine whether the GPI anchor was common to all three of these glycoproteins. To investigate this possibility, we determined if all three glycoproteins could be radiolabeled with precursors to the GPI anchor. In initial studies with embryos and 3H-labeled inositol, ethanolamine, palmitate, or myristate, we were unable to detect any labeled proteins comigrating with the two major proteins (130 and 205 kDa) that bear the 1223 epitope (data not shown). Given that the glycoproteins of interest are mesenchyme cell-specific, and that this cell type comprises only 5% of the cells in the embryo, we turned to the micromere-derived PM cells cultures for labeling studies. The results of the fatty acid labeling experiments with these cells are shown in Fig. 4. Although the spe-

cific activity of the palmitate was l.&fold higher than that of the myristate, it was still clear that the cells incorporated palmitate more readily (lfold) than myristate (Fig. 4A). As the arrows denote, labeled proteins that comigrated with the 130-, 205-, and 250-kDa glycoproteins bearing the 1223 epitope proteins were detectable. Consistent with the observations of Bolanowski et al. (1984) in whole embryos, many other proteins were also labeled with the fatty acids. Given this fact it was important to confirm that the three major fatty acid-labeled proteins did indeed contain the 1223 epitope. Therefore, the palmitate-labeled extracts were subjected to immunoaffinity chromatography using either control preimmune mouse IgG or mAb 1223 coupled to Affigel-10. The results shown in Fig. 4B establish that in the preimmune control all three fatty acid-labeled proteins were found in the unbound (Ub) fraction, whereas with the mAb 1223 gel the three labeled proteins of interest were retained in the bound (Bd) fraction. Identical results were obtained with myristate-labeled proteins (data not shown). Because ethanolamine has been shown to be a precursor of GPI anchors (Cross, 1990), labeling of the PM cell

KABAKOFF, HWANG, AND LENNARZ 1.1

Cmmm

other precursors, presumably as a result of extensive isotopic dilution (personal communication, G. Hart). To confirm that the ethanolamine was linked to the mAb 1223-reactive glycoproteins of interest via a GPI anchor, the sensitivity of these ethanolamine-labeled glycoproteins to phosphatidylinositol-specific phospholipase C (PIPLC) was tested. A representative Western blot of untreated and PIPLC-treated PM cell suspensions after a 2-hr incubation with enzyme is shown in Fig. ‘7. Relative to the control, only in the case of the 130-kDa glycoprotein was a significant portion (50%) of the immunoreactive glycoprotein released from the cell surface into the supernatant by enzymatic action. Further incubation up to 5 hr or doubling of the amount of PIPLC added did not alter the pattern (data not shown). Consistent with this observation, when ethanolamine-

1

1.0 -

0.9 0.8 0.7

-

0.6

-

0.5

-

0.4

-

0.3

-

0.2

-

0.1

7

0.0 0

24

I

I

48

72

Time

299

Mod$ficutions of Three Glgwproteins

96

(h)

FIG. 3. Developmental expression of msp 130- and 205kDa glycoproteins in embryo cultures. Aliquots from embryo cultures were harvested at the given times after fertilization, proteins were extracted, immunoblots were performed, and film was exposed for autoradiograms as described under Materials and Methods. Densitometric scanning of the resultant autoradiograms was used to quantitate the levels of msp 130 (0 0) and the 205-kDa (m - - - n ) glycoproteins. Values were normalized as a ratio of the intensity of the protein band at the time point to the intensity at the maximum level to facilitate comparison of different embryo cultures. Each value represents the average of the ratios from six individual cultures + SEM.

A

B -- PI 1223 UBBdUBBd

200kDa

116. 9?66s

culture was carried out with this compound. As shown in Fig. 5A, three of the major ethanolamine-labeled proteins corresponded to the three glycoproteins of interest (Fig. 5A). Immunoaffinity chromatography showed that all three ethanolamine-labeled glycoproteins (130, 205, and 250 kDa) were specifically bound to mAb 1223 (Fig. 5B). In an experiment where PM cell cultures were pulsed with labeled ethanolamine for 12 hr and then chased for 12, 24, and 36 hr, the label in the 130-kDa glycoprotein remained stable, whereas that in the 205kDa and probably the 250-kDa glycoprotein were lost (Fig. 5C). Densitometric quantification of the extent of turnover of the ethanolamine moiety in the 130- and 205-kDa glycoproteins is shown in Fig. 6. Within the limits of error there is no turnover of the ethanolamine moiety of msp 130, whereas that in the 205-kDa glycoprotein turned over with a T,,z of 14 hr. Attempts to label the cell cultures with inositol to further demonstrate the presence of the components typical of a GPI anchor were unsuccessful (data not shown). This finding is not unique to sea urchin cells; apparently, in other biological systems radiolabeled inositol is not incorporated as readily into the anchor as the

43.

FIG. 4. Fatty acid labeling and immunoaffinity binding of protein in PM cell cultures. PM cell cultures were prepared and rJH]myristate (Myr) or [8Hjpalmitate (Palm) was added as described at 48 hr after fertilization. The cells were incubated for an additional 24 hr before the cells were harvested, proteins were extracted, and an aliquot was placed on SDS gels for fluorography (A) as described. The remainder of the sample labeled with palmitate was incubated with preimmune IgG Affigel-10 (PI) or mAb 1223 Affigel-10 (1223) as described. The unbound (UB) and bound (Bd) fractions were analyzed by SDS-PAGE (B). The arrows denote the position of migration of the 1223 epitopeassociated proteins.

300

DEVELOPMENTALBIOLOGY

VOLUME150,1092

C

6 Hours 36 48 200kDa

after 54

0

Fertilization 60

66

Hours

of Chase I2

24

36

72 PI 1223 UB Bd UBBd

Il69766-

200kDa Il69766-

43-

FIG. 5. Ethanolamine labeling and immunoaffinity binding of proteins in PM cell cultures. (A) PM cell cultures were prepared as described in the text and [aH]ethanolamine (25 &i/ml) was added after fertilization at the hour denoted at the top of each lane of the fluorogram. The incubation was stopped after 12 hr and the cells were harvested and processed as described. (B) [*H]Ethanolamine (25 &i/ml) was added to PM cell cultures at 48 hr after fertilization and the incubation was allowed to proceed for 24 hr before harvesting and extraction of the cells. Equal portions of the protein were incubated with Affigel-10 coupled to preimmune IgG (PI) or mAb 1223 (1223). The unbound (UB) and bound (Bd) fraction were electrophoresed and fluorograms developed as described. (C) PM cell cultures were grown for 48 hr after fertilization and then PHjethanolamine (25 &i/ml) was added. After a 12-hr incubation, the radioactive medium was removed and medium containing unlabeled ethanolamine (1 mM) was added for the times noted above each lane of the fluorogram before the cells were harvested and processed. For all fluorograms the arrows denote where the immunoreactive glycoprotein migrate.

labeled PM cell suspensions were treated with PIPLC, the only labeled protein that was significantly released into the supernatant was the 130-kDa protein (Fig. 8, left panel). As also shown in Fig. 8 (right panel), the labeled protein released into the supernatant contained the 1223 epitope, since it was specifically retained on an immunoaffinity column. Thus, while these data show that all three glycoproteins incorporate palmitate and ethanolamine, only in the case of msp 130 is release by PIPLC observed. Previous efforts in our laboratory to directly determine by %O, labeling experiments if the 130-kDa glycoprotein is sulfated have been unsuccessful in whole embryos. In light of our finding that metabolic labeling of components of the mAb 1223 reactive glycoproteins is greatly enhanced in the PM cell culture system, the question of sulfation of this glycoprotein and, more specifically, the sulfation of its oligosaccharide chains(s) was re-examined using this cell culture system. Carrier-free %O, was added to culture media containing a tenth of the normal sulfate concentration to minimize isotopic dilution. As shown in Fig. 9A, many labeled sulfated proteins were synthesized in PM cells. Moreover, major labeled bands corresponding to the 130-, 205-, and 250kDa glycoproteins were readily detectable and these

three Sod-labeled glycoproteins specifically bound to the 1223 immunoaffinity column (Fig. 9B). These findings established that these glycoproteins are indeed sulfated, but they did not define the site of linkage of the sulfate. In order to obtain more information pertaining to this, the labeled glycoproteins were digested with PNGase F, which cleaves both complex and simple Nlinked chains. In Fig. 10A is shown a Western blot of the time course of PNGase F action on the glycoproteins containing the 1223 epitope. In confirmation of previous work by Farach-Carson et al. (1989), enzyme treatment resulted in a loss of the immunoreactivity of the 130kDa glycoprotein. Additionally, with both the 205- and 250-kDa glycoproteins there appeared to be a great diminishment of immunoreactivity and a distinct shift to lower molecular weights of roughly 175 and 230 kDa, respectively. As shown in Fig. lOB, when sulfate-labeled glycoproteins from PM cell cultures were treated with PNGase F for 24 hr, a similar pattern of change was observed and the sulfate label was released as a consequence of cleavage of the oligosaccharide chains from the protein. The above experiments demonstrate that the three glycoproteins bearing the 1223 epitope are sulfated and that most of the sulfate is incorporated into N-linked

KABAKOFF, HWANG, AND LENNARZ 2.0 .

Common Mod@xtims of ThreeGlycoproteins

I

130

SUP PI-i23 -Q%i~"BBdUBBd

-PIPLC + PIPLC

kDa

Q..%*! 1::

301

/\/'I

0.5 -

2.0

1.5

1.0

0.5

0.0

0

12

24

36

HOURS OF CHASE FIG. 6. Kinetics of turnover of PHjethanolamine incorporated into the 130- and 205-kDa glycoproteins. The fluorograms in Fig. 5C were densitometrically scanned and the relative levels of radiolabel associated with the 130-kDa protein (top) and 205-kDa (bottom) were measured as a function of time of chase.

oligosaccharide chains. However, because 130 could contain up to six oligosaccharide et al, 1990) and not all of these necessarily with the 1223 epitope, the above results swered the question of whether some or all is on the oligosaccharide chains that bear

-PIPLC

at least msp chains (Parr are modified leave unanof the sulfate the 1223 epi-

+PIPLC

2503 2054 1304 kDa FIG. 7. Western immunoblot of PM cell protein extracts treated with PIPLC. PM cell cultures were prepared as described in the text. The cultures were harvested at 72 hr after fertilization and the cells resuspended in the Hepes/glycine buffer described under Materials and Methods. The cells were incubated in the absence (-PIPLC) or presence of 0.5 units of lipase (+PIPLC) as described for 2 hr at 37°C. The cells were pelleted (Pel) and the proteins in the supernatant (Sup) precipitated with acetone prior to SDS-PAGE followed by Western immunoblot analysis.

FIG. 8. Release of ethanolamine-labeled 130 kDa from the cell surface by PIPLC treatment and its specific binding to an immunoaffinity column. PM cell cultures were prepared as described and allowed to grow for 48 hr after fertilization. PH]Ethanolamine was added (25 &i/ml) at this point and the cultures were incubated for an additional 24 hr, after which time the cells were harvested and cell suspensions were prepared in Hepes/glycine buffer. Equal aliquots were incubated in the absence (-PIPLC) or presence (+PIPLC) of lipase for 2 hr at 37°C. The cells were pelleted (Pel) and the supernatant (Sup) was precipitated with acetone. Both pellets were redissolved in 100 ~1 of buffer containing 2% cholate and 10 mMTris-HCl, pH 8.0. Approximately 20% of each of these suspensions were used for the fluorograms represented as -PIPLC and +PIPLC. The remaining 80% of the supernatant (Sup) of the lipase-treated sample was divided equally and incubated with either preimmune IgG (PI) or mAb 1223 (1223) affinity columns, The unbound (UB) and bound (Bd) fractions were processed as described. The arrows denote the position of migration of the immunoreactive glycoproteins.

tope. Unfortunately, because the free oligosaccharide chains released by PNGase F no longer react well with mAb 1223 it was not possible to determine if the epitope is sulfated by immunoaffinity purification of the radiolabeled, free oligosaccharide. DISCUSSION

The results of this study establish the existence of at least three glycoproteins in PM cells that contain an epitope that reacts with mAb 1223. A study of the profiles of expression of these glycoproteins revealed that all three first appear prior to the onset of spiculogenesis, as expected if they play a role in spicule formation.

302

DEVELOPME~AL

A 35S04

200

BIOLOGY

B PI

1223

UBBd

UBBd

hDa

-200 kDa

116

- 116

97

-

97

66

43

VOLUME 150,1992

130 is related to the two other proteins with respect to a common carbohydrate epitope, previous work has clearly shown them to be unrelated with respect to the polypeptide backbone, because neither the 205 nor 250kDa glycoproteins react with an antibody directed against the polypeptide chain of msp 130 (Leaf et al, 1987; Kabakoff and Lennarz, 1990). Given the finding of Parr et al. (1990) that the polypeptide backbone of msp 130 does not contain transmembrane sequences, it seems unlikely that this glycoprotein acts as Ca2+ channels. However, further work to elucidate the primary structure of the other two glycoproteins may reveal other common features that will be clues to their function. Another reason to learn more about the primary structure of these other glycoproteins is our finding

B

FIG. 9. Sulfate labeling and immunoaffinity binding of protein extract from PM cell cultures. Cultures were prepared as described and incubated at 14°C. At 48 hr after fertilization, the culture medium was removed and 1 ml of medium containing l/lOth the normal concentration of sulfate and 100 pCi/ml of carrier-free NaaS6S04was added back to the cells. The cultures were incubated for an additional 24 hr and then harvested and processed as described. An aliquot (10%) of the protein extract was analyzed on SDS-PAGE followed by fluorography (A). The remainder (B) was divided into two equal portions and incubated with preimmune IgG (PI) and mAb 1223 (1223) coupled to Affigel-lo. The unbound (UB) and bound (Bd) fractions were analyzed by SDS-PAGE followed by fluorography.

We have extended the observations of Carson et al. (1985) to demonstrate that the antibody which recognizes these glycoproteins, mAb 1223, not only prevents elongation of preformed spicules, but also blocks spicule formation entirely when it is present in PM cell cultures at a concentration of 5 pg/ml or greater. Because we also found that the treated cultures still contain the filopodial networks that form prior to skeletogenesis (Decker and Lennarz, 1988), it is likely that the glycoprotein(s) containing the 1223 epitope is involved in Ca2’ accumulation or the mineralization process per se, rather than the earlier events involving cell migration or development of cell interactions between the PM cells. Given the earlier observation (Farach-Carson et a& 1989) that the mAb 1223 immunoreacts with a specific subset of N-linked oligosaccharide chains, we conclude that these chains are important in some step in the biomineralization process. Although this work has shown that msp

FIG. 10. PNGase F digestion of protein extracts from PM cell cultures. Unlabeled PM cell cultures were prepared as described. In cultures labeled with “SO,, the radiolabel (100 &i/ml) was added at 48 hr after fertilization. Both cultures were harvested at ‘72 hr after fertilization and protein extracts were prepared by resuspending the cells in lysis buffer and heating for 10 min at 1OO’C. The other buffer components were then added as described and the PNGase F incubation begun at 37’C by addition of enzyme. For Western immunoblot analysis (A), aliquots were removed for SDS-PAGE at the times after initial enzyme addition as indicated above each lane. For fluorograms (B), proteins were incubated with (+PNGase F) or without (-PNGase F) enzyme for 24 hr before the reaction was stopped and aliquots prepared for SDS-PAGE and fluorography.

KABAKOFF,

HWANG,

AND LENNARZ

Cmwum ModQicatims

that components of the GPI anchor, previously identified on msp 130 (Parr et c& 1990), were also found in the two other glycoproteins (205 and 250 kDa). All three of the 1223 antigen-bearing proteins in the PM cell culture readily labeled with ethanolamine and palmitate, two components of the GPI anchor. The 130-kDa protein, in confirmation of previous observations from Raff’s laboratory (Parr et al, 1990), was found to be partially sensitive to PIPLC treatment. However, both the 205 and 250-kDa proteins were completely resistant to this enzyme. The most likely explanation for this observation is that these two glycoproteins incorporate the labeled compounds into a type of anchor which is resistant to PIPLC treatment. For example, a recent report suggests that the contact site A glycoprotein in Dictvostelium discoideum is attached to the membrane via a novel ceramide-based lipid glycan (Stadler et a& 1989) and therefore is not susceptible to PIPLC treatment. Alternatively, these two glycoproteins may also have GPI anchors, but these anchors may be resistant to enzyme action due to either chemical modification or physical inaccessibility. There is precedence for the former since it has been shown that certain GPI anchors containing fatty acyl groups on the inositol ring are not cleaved by PIPLC (Mayor et aL, 1990). Preliminary experiments with crude phosphatidyl inositol-specific phospholipase D from rat serum, which has been shown to act on such modified inositol rings, were inconclusive. We have investigated the latter possibility of physical inaccessibility in two ways: PM cell suspensions were sonicated to form vesicles and then treated with PIPLC. Such treatment did not alter the results obtained with intact cells. A second approach was to study the partitioning of the three glycoproteins into the detergent phase of Triton X-114 before or after lipase treatment. Unfortunately, even in the absence of lipase treatment, all three proteins partitioned into the aqueous phase, thus precluding the use of this technique. There are precedents in the literature to support the possibility of physical constraints preventing release of a GPI anchored protein from the cell surface. Recently Stahl et al. (1990) demonstrated that in cells the scrapie prion protein is PIPLC resistant even though it has been clearly demonstrated that this protein, when isolated, does have a GPI anchor sensitive to PIPLC. The authors speculate that this resistance is due to tight packing of the protein on the cell surface, which is thought to be the reason for the resistance of VSG to PIPLC-catalyzed release from the surface of the trypanosome (Cross, 1975, 1984; Tetley et aZ., 1981). A second possibility is that not all three GPI containing proteins are located on the cell surface. Yet another possibility offered by Stahl et al. (1990), and earlier by Low (1987) with reference to Thy-l, is that the anchor is actually cleaved, but that the

of Three Glyccproteins

303

protein is not released from the cell surface because of interactions with another cell membrane protein. This would be a particularly appealing possibility in that we have proposed that the protein(s) containing the 1223 epitope may be involved in calcium transport not as a transporter, but rather as a means of concentrating calcium to the cell surface in the vicinity of a transporter with which it interacts (Farach-Carson et al, 1989). Another, less likely, explanation for the lipase resistance of two of the three glycoproteins may be that both palmitate and ethanolamine are coincidentally incorporated in these two proteins at sites other than a GPI anchor. In the case of palmitate there is a large body of literature on the fatty acylation of proteins via a thioester linkage or via an amide to the N-terminal amino acid residue (Towler et ab, 1988). In fact, the former linkage has been shown to be present in sea urchin embryos (Bolanowski et al, 1984). With respect to ethanolamine modifications of proteins, it has been shown that murine elongation factor-la contains ethanolamine linked via an amide linkage to internal glutamic acid residues (Whiteheart et al., 1989). However, other than those containing GPI anchors we are not aware of instances where both of these moieties have been found in the same protein. It is important to note that the congruence of the occurrence of the 1223 epitope and the GPI anchor in all three proteins does not mean that the reactivity to mAb 1223 resides in the anchor. In fact, the finding that the epitope is sensitive to recombinant PNGase F treatment, which hydrolyzes N-linked oligosaccharides but has no demonstrated effect on GPI anchors, strongly argues against such a possibility. Moreover, although many proteins in PM cells incorporate sulfate and palmitate, it is clear that the three glycoproteins containing the 1223 epitope are by far the major proteins which incorporate ethanolamine and presumably are the major PM cell proteins with GPI anchors. It will be interesting to determine the nature of the signal(s) which specifically allows for the assembly of both mAb 1223reactive N-linked oligosaccharide chains and GPI anchors in all three of these proteins. Despite these common features the 130- and 205-kDa glycoproteins differ in the stability of the labeled ethanolamine moiety incorporated into them. When cells were pulsed with labeled ethanolamine and then chased, the radioactivity remained stably associated with the 130-kDa glycoprotein, whereas it was rapidly chased out of the 205-kDa glycoprotein. However, the preliminary studies in PM cell cultures revealed that the msp 130 did not decline after gastrulation as was observed in the embryos. Therefore, it is possible that some signal is required from an ectodermal or endodermal source to

304

DEVEL~PMENTALBIOLOGY

effect turnover of msp 130 and subsequently turnover of the GPI anchor. Regardless, both possibilities have intriguing implications with respect to the developmental biology and biochemistry of spicule formation. The latter scenario suggests that a paracrine signal of unknown nature from ectodermal or endodermal sources is necessary for proper turnover of one mesenchyme cell-surface protein, msp 130, and its associated GPI anchor, but not for the 205-kDa protein even though they share several common post-translational modifications. The other possibility suggests that the relative stabilities of the label in these proteins and their different developmental profiles reflect differential functions of the two proteins during development of the sea urchin embryo. In fact, little is known of the functional significance of membrane attachment via a GPI anchor. Recently it has been reported that an extracellular matrix protein of neural cell origin in Drosophila, fasciclin, has a GPI anchor which is thought to be developmentally controlled in such a manner as to regulate its adhesive function (Hortsch and Goodman, 1990). It also has been suggested that this type of membrane anchor imparts a greater degree of fluidity for the protein within the membrane. If, as postulated, the proteins with the 1223 antigen are involved in Ca2’ uptake during spiculogenesis, such added fluidity may be important in this process. Clearly, the PM cells of the sea urchin embryo should provide a good system for further study of the developmental regulation and functional role of the GPI membrane anchor. Earlier, Anstrom et al. (1987) presented indirect evidence to suggest that msp 130 is a sulfated glycoprotein; in this study using cultured PM cells we have provided direct evidence to support this idea. Moreover, %04 labeling experiments revealed that not only msp 130, but also the 205- and 250-kDa glycoproteins are sulfated. Thus, all three proteins bearing mAb 1223-reactive Nlinked oligosaccharide chains also contain GPI-like anchors and also are sulfated. The majority of the sulfate label appears to be associated with the N-linked oligosaccharide chains, but as of yet we have been unable to determine if it is exclusively attached to the mAb 1223reactive N-linked oligosaccharide chains or to other Nlinked oligosaccharide chains that may be present. It is hoped that structural studies now in progress on the mAb 1223-reactive oligosaccharide chains will answer this question and will be useful in determining the precise role of these novel glycoproteins in spiculogenesis. We express our appreciation to Andrea Bonner for her considerable help in preparing primary mesenchyme cell cultures and to Lorraine Conroy for her assistance with the preparation of the manuscript.

VOLUME150,1992 This work was supported by a grant from the National Institutes of Health (HD21483) to W.J.L.

REFERENCES Anstrom, J. A., Chin, J. E., Leaf, D. S., Parks, A. L., and Raff, R. A. (1987). Localization and expression of msp 130, a primary mesenthyme lineage-specific surface protein of the sea urchin embryo. Development 101,255-265. Bolanowski, M., Earles, B. J., and Lennarz, W. J. (1984). Fatty acylation of proteins during development of sea urchin embryos. J. Biol Chenh 2.59,4934-4940. Carson, D. D., and Lennarz, W. J. (1979). Inhibition of polyisoprenoid and glycoprotein synthesis causes abnormal embryonic development. Proc. Natl. Ad Sci. USA 76,5709-5713. Carson, D. D., and Lennarz, W. J. (1981). Relationship of dolichol synthesis to glycoprotein synthesis during embryonic development. J. Biol Chem 256,4679-4686. Carson, D. D., Farach, M. C., Earles, D. S., Decker, G. L., and Lennarz, W. J. (1985). A monoclonal antibody inhibits calcium accumulation and skeleton formation in cultured embryonic cells of sea urchin. Cell 41, 639-648. Cross, G. A. M. (1975). Identification, purification and properties of clone-specific glycoprotein antigens constituting the surface coat of Trypanosoma brucei. Parasitologzl 71,393-417. Cross, G. A. M. (1984). Release and purification of Trypanosoma brucei variant surface glycoprotein. J. CeElBiochem 24,79-90. Cross, G. A. M. (1990). Glycolipid anchoring of plasma membrane proteins. Annu. Rev. Cell. BioL 6, l-40. Decker, G. L., and Lennarz, W. J. (1988). Skeletogenesis in the sea urchin embryo. Development 103,231-247. Farach-Carson, M. C., Carson, D. D., Collier, L., Park, H. R., and Wright, G. C. (1989). A calcium-binding asparagine-linked oligosaccharide is involved in skeleton formation in sea urchin embryo. J. Cell Biol 109,1289-1299. Field, M. D., Menon, A. K., and Cross, G. A. M. (1991). Developmental variation of glycosylphosphatidylinositol membrane anchors in Tryparwsomn ~vY.Ix~~: Identification of a candidate biosynthetic precursor of the glycosylphosphatidylinositol anchor of the major procyclic stage surface glycoprotein. J. Biol Chem 266,8392-8400. Heifetz, A., and Lennarz, W. J. (1979). Biosynthesis of N-glycosidically linked glycoproteins during gastrulation of sea urchin embryos. J. Biol Ch,ewz.254,6119-6127. Hortsch, M., and Goodman, C. S. (1990). Drosophila fasciclin I, a neural cell adhesion molecule, has a phosphatidylinositol lipid membrane anchor that is developmentally regulated. J. BioL Chmz. 265, 15,104-15,109. Kabakoff, B., and Lennarz, W. J. (1990). Inhibition of glycoprotein processing blocks assembly of spicules during development of the sea urchin embryo. J. Cell Biol 111,391-400. Leaf, D. S., Anstrom, J. A., Chin, J. E., Harkay, M. A., Showman,R. M., and Raff, R. A. (1987). Antibodies to a fusion protein identify a cDNA clone encoding msp 130, a primary mesenchyme-specific cell surface protein of the sea urchin embryo. De-v. Bid 121,29-40. Lennarz, W. J. (1983). Glycoprotein synthesis and embryonic development. Cri’t. Rev. BiocFwm. 14,257-272. Low, M. G. (1987). Biochemistry of the glycosyl-phosphatidylinositol membrane protein anchors. Biochem. J. 244,1-13. Mayor, S., Memon, A. K., and Cross, G. A. M. (1990). Glycolipid precursors for the membrane anchor of Trypamwma Bruce’ variant surface glycoproteins. II. Lipid structures of phosphatidylinositol-specific phospholipase C sensitive and resistant glycolipids. J. Bid Chem 265,6174-6181.

KABAKOFF,

HWANG,

AND LENNARZ

Common MocQfications of Three Glycoproteins

McClay, D. R., Matranga, V., and Wessel, G. (1985). Expression and appearance of germ-layer specific antigens on the surface of embryonic sea urchin cells. In “The Cellular and Molecular Biology of Invertebrate Development” (R. H. Sawyer and P. M. Showman, Eds.), pp. 171-186. University of South Carolina Press, Columbia, SC. Parr, B. A., Parks, A. L., and Raff, R. A. (1990). Promoter structure and protein sequence of msp 130, a lipid-anchored sea urchin glycoprotein. J. BioL Chem 265.14081413. Roux, L., Holojda, S., Sundblad, G., Freeze, H. H., and Varki, A. (1988). Sulfated N-linked oligosaccharides in mammalian cells. I. Complextype chains with sialic acids and O-sulfate esters. J. BioL Chem. 263, 88’794889. Schneider, E. G., Nguyen, H. T., and Lennarz, W. J. (1978). The effect of tunicamycin, an inhibitor of protein glycosylation, on embryonic development in the sea urchin. J. BioL Chem. 253,2348-2355. Stadler, J. T., Keenan, W., Bauer, G., and Gerisch, G. (1989). The con-

305

tact site A glycoprotein carries a phospholipid anchor of a novel type. EMBO. J. 8,371-377. Stahl, N., Borchelt, D. R., and Prusiner, S. B. (1990). Differential release of cellular and scrapie prion proteins from cellular membranes by phosphatidylinositol-specific phospholipase C. Biochxmistq/ 29,5405-5412. Tetley, L., Vickerman, K., and Moloo, S. K. (1981). Absence of a surface coat from metacyclic Twpanosoma tiva~ Possible implications for vaccination against vivax trypanosomiasis. Trans. R. Sot Trap. Med Hyg. 75,409-414. Towler, D. A., Gordon, J. I., Adams, S. P., and Glaser, L. (1988). The biology and enzymology of eukaryotic protein acylation. Annu. Rev. Biochem. 57,69-99. Whiteheart, S. W., Shenbagamurthi, P., Chen, L., Cotter, R. J., and Hart, G. W. (1989). Murine elongation factor la is posttranslationally modified by novel amide-linked ethanolamine-phosphoglycerol moieties. Addition of ethanolamine-phosphoglycerol to specific glutamic acid residues on EF-la. J. BioL Chem 264,14,334-14,341.