- Email: [email protected]
Inhibitors of metalloendoproteases block spiculogenesis in sea urchin primary mesenchyme cells
Experimental
Inhibitors
JUDITH
Cell Research 181 (1989) 542-550
of Metalloendoproteases Block Spiculogenesis Urchin Primary Mesenchyme Cells L. ROE,
Department of Biochemistry
HELEN R. PARK, WARREN and WILLIAM J. LENNARZ*
in Sea
J. STRITTMATTER,’
and Molecular Biology, UT M.D. Anderson Holcombe Boulevard, Houston, Texas 77030
Cancer Center, 1515
Metalloendoproteases have been implicated in a variety of fusion processes including plasma membrane fusion and exocytosis. As a prerequisite to skeleton formation in the sea urchin embryo, primary mesenchyme cells undergo fusion via tilopodia to form syncytia. The spicule is formed within the syncytial cable by matrix and mineral deposition. To investigate the potential involvement of a metahoendoprotease in spiculogenesis, the effect of inhibitors of this enzyme on skeleton formation was studied. Experiments with primary mesenchyme cells in vitro and in normal embryos revealed that skeleton formation was blocked by these inhibitors. These findings implicate a metalloendoprotease in spiculogenesis; such an enzyme has been demonstrated in homogenates of primary mesenchyme cells. The most likely site of action of the metalloendoprotease is at the cell membrane fusion stage and/or at subsequent events requiring membrane fusion. @ 1989 Academic FWSS, Inc.
Previous studies have implicated metalloendoproteases in a variety of cellular processes involving membrane fusion, such as exocytosis in mast cells and adrenal chromaffin cells [l] and plasma membrane fusion in myoblasts [2, 31. In addition, studies with sea urchin sperm have revealed the presence of a metalloendoprotease, inhibitors of which block the acrosome reaction [4]. Moreover, studies have implicated metalloendoproteases in sea urchin gamete fusion [5]. Another potential site of action of metalloendoproteases in a fusion event is syncytial formation of primary mesenchyme cells of the sea urchin embryo. These cells, which function in assembly of the calcite skeleton of the prism stage embryo, fuse with each other via filopodia to form a syncytium within which the CaC03-containing spicule is formed (for review see [6]). In view of our interests in skeletogenesis and membrane fusion, we have studied the effects of metalloendoprotease inhibitors and dipeptide substrates on cultures of isolated primary mesenchyme cells which form spicules in vitro in the presence of horse serum 17, 81. The results show that both an inhibitor and a dipeptide substrate of metalloendoproteases block the formation of spicules in vitro. The inhibitors also arrest spicule formation in the embryo. Direct evidence was obtained for a metalloendoprotease in homogenates prepared from primary mesenchyme cells isolated from mesenchyme blastula stage embryos or cultured from micromeres. From these studies we infer that a metalloendoprotease may be involved in the process of
I Department
of Neurology,
Baylor
College
of Medicine,
2 To whom reprint requests should be addressed. Copyright @ 1989 by Academic Press, Inc. All rights of reproduction in any form reserved 0014-4827189 $03.00
542
Houston,
TX 77030.
Metalloendoproteases
spiculogenesis, possibly by inhibiting mesenchyme cells. MATERIALS
and spiculogenesis
fusion of the cell membranes
AND
543
of the primary
METHODS
Materials. Strongylocentrotus purpuratus were obtained from Pacific Biomarine Supply Company (Venice, CA) and Marinus (Long Beach, CA). Falcon tissue culture dishes were obtained from Becton Dickinson and Co. (Lincoln Park, NJ). CBZ’-Gly-Phe-NH* and l,lO-phenanthroline were obtained from Sigma (St. Louis, MO) and CBZ-Gly-Gly-NH2 was obtained from Vega Biotechnologies, Inc. (Tucson, AZ). All other chemicals were reagent grade. Embryo culture andprimary mesenchyme cell culture. Gametes from S. purpuratus were obtained and embryos were cultured as described [9]. Primary mesenchyme cells were cultured essentially as described [8] with the addition of 1% penicillin-streptomycin (GIBCO Labs, Grand Island, NY), 0.2 gm/liter Ticarcillin disodium (Beecham Laboratories, TN), and 10 mg&ter erythromycin (Abbott Labs, North Chicago, IL) to the culture medium. Micromeres were isolated from 16-cell stage embryos by the method of Okazaki [18] and cultured in millipore-filtered artificial seawater (MFSW) supplemented with 3 % horse serum and 5 @ml gentamycin (GIBCO Labs). Metalloendoprotease assays. Metalloendoprotease activity was assayed using the fluorogenic protease substrate succinyl-Ala-Ala-Phe-aminomethylcoumarin (Sigma) (Succ-Ala-Ala-Phe-AMC). Primary mesenchyme cells, isolated from mesenchyme blastula stage embryos or from cells cultured from 16-cell stage micromeres, were assayed for metalloendoprotease activity in a crude, cell-free homogenate. Cells were scraped off the culture dish, washed three times in MFSW, and homogenized by three-freeze-thaw cycles in 50 n&f Hepes, pH 7.0, 150 mM KCl. Twenty-five microliters of homogenate, containing 1 to 5 ug protein, were incubated in 0.4 mM Succ-Ala-Ala-Phe-AMC in 5 % DMSO with 5ug chymostatin, in a total volume of 100 pl for 24 h at 37°C with or without the addition of metalloendoprotease inhibitors EGTA (5 mM), CBZ-Gly-Phe-NH2 (2 mM), or active site inhibitor 2-(iV-hydroxycarboxamido)-4-methylpentanoyl-~-alanyl glycine amide (Zincov, Calbiochem) (20 l&f). After incubation, free unquenched AMC was liberated from the hydrolyzed substrate by incubation for 30 min at 37°C with 20 @ml aminopeptidase M (Boehringer-Mannheim) with 1 x 10-r M phosphoramidon to inhibit a contaminating metalloendoprotease. Fluorescence (excitation 380 nm, emission 460 nm) of free AMC was measured in an Amino Bowman fluorometer. The amount of AMC produced was then determined by comparison with a standard curve.
RESULTS Inhibitors of Metalloendoproteases Block Spicule Formation in Primary Mesenchyme Cell Cultures
The chelator, l,lO-phenanthroline (a heavy metal chelator with a high affinity for Zn2+) has been shown to inhibit myoblast fusion [2] and mast and adrenal chromaffin cell exocytosis [l]. These findings, coupled with the demonstration that a Zn2+-dependent metalloendoprotease activity is present in these cells, have led to the hypothesis that this enzyme may be required for membrane fusion [l-3]. Because the formation of spicules is preceded by fusion of primary mesenthyme cells, we tested the effect of this chelator on spicule formation by these cells. As previously described [8], primary mesenchyme cells were cultured in vitro by plating dissociated cells from the mesenchyme blastula stage (2 h posthatching). The primary mesenchyme cells were allowed to adhere to the culture dish and the unattached cells were washed out 18 h after plating. As shown in Fig. 1, ’ CBZ, carbobenzoxy; EGTA, 35-898334
ethylene
glycol
DMSO, dimethyl sulfoxide; EDDA, ethylenediamine-N’,N’-diacetic bis(j3-aminoethyl
ether)-N’,N’,N’,N’-tetraacetic
acid.
acid;
544
Roe et al.
Fig. I. Effect of l,lO-phenanthrohne on cultures of isolated primary mesenchyme cells. Primary mesenchyme cells were isolated as described under Materials and Methods. Nonadherent cells were removed at 18 h and cultures were incubated at 14°C for 72 h. Control cultures (A) and cultures treated with 0.05% DMSO (B) showed extensive spicule formation. Alternatively, cultures were incubated with 10 (C), 50 (D), 100 (,?!I),or 250 @4 (fl l,lO-phenanthroline immediately after removal of nonadherent cells and cultured for an additional 72 h. (Magnification 47x .)
untreated cultures showed extensive spicule formation 72 h after removal of nonadherent cells (Fig. 1 A), as did cultures treated with DMSO alone (Fig. 1 B) or with 10 @4 (Fig. 1 C) or 50 @4 (Fig. 1 D) l,lO-phenanthroline. l,lO-Phenanthroline was added to such cultures after the nonadherent cells had been removed. However, inhibition of spicule formation after 72 h was seen in cultures treated with 100 u.A4 1, lo-phenanthroline (Fig. 1 E) and spicule formation was completely inhibited with 250 @4 1, lo-phenanthroline (Fig. 1fl. The inhibitory effect was not the result of inhibition of primary mesenchyme cell attachment because cells did attach in the presence of the chelator (data not shown). Filopodial extensions were not seen between cells in 1, lo-phenanthroline-treated cultures. The effect of
Metalloendoproteases
and spiculogenesis
545
Fig. 2. Effect of dipeptides on cultures of primary mesenchyme cells. Cultures of primary mesenthyme cells (see Materials and Methods) were cultured in normal culture medium (A) or treated with 2 rnIU CBZ-Gly-Gly-NH2 (B) or 2 mA4 CBZ-Gly-Phe-NH2 (C) after removal of nonadherent cells and cultured an additional 72 h. (D), cultures were incubated for 72 h in 2 m&I CBZ-Gly-Phe-NH2 and washed and then the cells were replated in fresh media and incubated an additional 120 h. (Magnification (A-C) 47x and (0) 94x .)
1, IO-phenanthroline on spiculogenesis was irreversible, consistent with previous reports on its effect on myoblast fusion [2] and metalloendoprotease activity [ 101. The effects of 1, IO-phenanthroline are not likely to be due to Ca” chelation, as Ca*+ ions are in great excess in seawater. However, because this chelator may have pleiotropic effects, we tested the effect of more specific inhibitors of metalloendoprotease activity. Studies on the specificity of metalloendoproteases using various peptide substrates have shown that these proteases cleave the peptide bond on the amino side of uncharged aromatic and large aliphatic amino acids. Dipeptides of the general type Gly-Phe or Gly-Leu containing their N- and C-termini blocked have been shown to be substrates of these proteases [ll]; when they are added in excess they would be expected to inhibit the action of the metalloendoprotease on endogenous substrates. Indeed, as shown in Fig. 2, the metalloendoprotease substrate CBZ-Gly-Phe-NH2 was found to completely inhibit spicule formation in the primary mesenchyme cell cultures (Fig. 2 C). Cells incubated with the control dipeptide CBZ-Gly-Gly-NH2 (Fig. 2B), which is not a substrate for metalloendoprotease, formed spicules as well as control cultures formed (Fig. 2A). The cells cultured in the presence of CBZ-Gly-Phe-NH2 filopodia to a variable extent as seen in the early stages of control cultures. It was found that the effect of the inhibitory dipeptide was reversible (Fig. 20); when
546
Roe et al.
3. Effect of 1, IO-phenanthroline on spiculogenesis in vivo. Early gastrula stage embryos (36 h) were cultured for an additional 30 h in the absence (control) (II) or presence of 1 m&I l,lOphenanthroline in 0.2% DMSO (CT),in 2.5 m&f l,lO-phenanthroline in 0.5% DMSO (D), or in 0.5% DMSO alone (I?). Embryos were viewed using Normarski optics. (Magnification 196x .) Fig.
(A)
the dipeptide was removed the cells were able to resume differentiation and form spicules. More cells were observed to cluster around the spicules than in control cultures. Metalloendoprotease Inhibitors Block Spiculogenesis in the Embryo
Because the metalloendoprotease inhibitors were found to block spiculogenesis in cultures of isolated primary mesenchyme cells, we tested the effects of the inhibitors on spiculogenesis in normal embryos. Inhibitors were added to cultures of S. purpuratus embryos at the gastrula stage (Fig. 3A), when the archenteron has formed and the primary mesenchyme cells are migrating within the blastocoel. As shown in Fig. 3, embryos were treated with l,lO-phenanthroline at 36 h and cultured for an additional 30 h. No spicules were formed in embryos cultured
Metalloendoproteases
and spiculogenesis
547
Fig. 4. Effect of dipeptides on spiculogenesis in vim Early gastrula stage embryos (36 h) (Fig. 3A) were cultured for an additional 30 h in the absence (control) (A) or presence of 2 mA4 CBZ-Gly-GlyNH2 in 0.5% DMSO (B) or 2 mA4 CBZ-Gly-Phe-NH2 in 0.5% DMSO (C). Embryos were viewed using Nomarski optics. (Magnification 196x .)
in 1 mM (Fig. 3 C) or 2.5 mM (Fig. 3 0) l,lO-phenanthroline. By this time (66 h) untreated cultures had~ developed complete triradial skeletons (Fig. 3B). The inhibition was not reversible, although the embryos continued to swim and appeared viable. The concentration of 1 ,lO-phenanthroline required to inhibit spiculogenesis in the embryo was 4- to IO-fold higher than observed in the primary mesenchyme cell cultures. In other experiments it was found that if the chelator was added after the formation of a small triradial skeleton (40 h), further elongation of the spicule was inhibited (data not shown), suggesting that I ,lOphenanthroline can block spicule deposition. The dipeptide metalloendoprotease substrate, CBZ-Gly-Phe-NHz, also inhibited spicule formation in the embryo. When CBZ-Gly-Phe-NH2 was added to embryos at 36 h and the embryos were cultured for 30 h, as in the experiment described in Fig. 3, no spicules were observed (Fig. 4C), while untreated embryos formed complete skeletons (Fig. 4A). As shown, addition of the dipeptide CBZ-Gly-Gly-NH*, which is not a metalloendoprotease substrate, had no effect on the embryos (Fig. 4B). These data show that inhibitors of a metalloendoprotease are able to inhibit spiculogenesis in the normal embryo as well as in isolated cultures of primary mesenchyme cells.
548
Roe et al.
Primary
Mesenchyme
Cells Contain Metalloendoprotease
Activity
Metalloendoprotease activity was identified in primary mesenchyme cells using the synthetic fluorogenic endoprotease substrate succinyl-Ala-Ala-Phe-AMC, used by us previously to characterize this protease in other cells [ 1, 41. Homogenates prepared from primary mesenchyme cells isolated from blastula stage embryos or cultured from 16-cell stage micromeres were found to contain endoprotease activity that hydrolyzed succinyl-Ala-Ala-Phe-AMC at a linear rate for 24 h. The metal chelator EGTA (5 mM) inhibited substrate hydrolysis by 85 %; this metal-dependent endoprotease activity ranged from 4.0 to 34~ 10-l’ mol substrate hydrolyzed/24h/ug protein in five different cell preparations. The metal chelator 1, lo-phenanthroline cannot be used in this assay because it inhibits the aminopeptidase M used in this coupled assay (see Materials and Methods and [l]). The dipeptide CBZ-Gly-Phe-NH* (2 mM) inhibited hydrolysis of the synthetic substrate by 50% and the active site metalloendoprotease inhibitor, 2-W hydroxycarboxamido)-4-methylpentanoyl-L-alanylglycine amide (20 uM) (Zincov), inhibited substrate hydrolysis by 60%. These data demonstrate the presence of a metalloendoprotease activity in the spicule forming cells that is inhibited by the same peptide substrate shown to block spicule formation. DISCUSSION The mechanism of membrane fusion has been studied using a variety of model systems including cell-cell fusion, vesicle-vesicle fusion, vesicle-plasma membrane fusion, and virus-host cell fusion. Many of these events are Ca*+-dependent and may require the activity of a Zn*+ -dependent metalloendoprotease [ 121. Although in some systems the generation of a fusogenic protein by proteolysis is required for fusion, a general hypothesis of the mechanism of involvement of metalloendoproteases has not yet been formed. In the sea urchin sperm, we have recently described a Zn*+-dependent metalloendoprotease that may be involved in the acrosome reaction, an event that requires vesiculation of the membrane overlying the acrosomal granule [4]. Sea urchin gamete fusion also appears to require the activity of a metalloendoprotease [5]. In the current report we have studied the effects of metalloendoprotease inhibitors on the differentiation of primary mesenchyme cells of the sea urchin embryo. During this differentiation, a syncytial structure within which the skeleton is formed is generated by fusion of the filopodia of the primary mesenchyme cells. The nature of the syncytia formed by primary mesenchyme cells in vivo has been studied at both the light [13-151 and electron microscope levels [6, 16-191. the cells are connected via a syncytial cable which was formed by the fusion of tilopodia. The cell bodies are often connected to the syncytial cable to short stalks. The spicule is formed within the cable by mineral and matrix deposition, and the syncytial cable is connected by tilopodial extensions to the blastocoel wall. The primary mesenchyme cells appear to be highly fusogenic and Okazaki [13] has observed that the filopodia of cells can fuse, release, and re-fuse with another cell in vivo during the syncytial formation. The tilopodia appear to fuse
Metalloendoproteases
and spiculogenesis
549
only at the tips, suggesting localized fusogenic sites. The syncytia appears to be a requirement for spicule formation because single cells have been observed only to form small calcareous deposits [18]. We carried out studies with two types of inhibitors of metalloendoproteases, a Zn*+ chelator and dipeptide substrates of the enzyme that have been shown to inhibit fusion events in other systems [l-5]. The effect of these agents was examined in normal embryos and cultures of isolated primary mesenchyme cells. l,lO-Phenanthroline (a heavy metal chelator and metalloendoprotease inhibitor) blocked the formation of spicules by primary mesenchyme cells both in isolated primary mesenchyme cell cultures and in the normal embryo. Addition of excess Zn*+ ions is known to arrest development in the sea urchin embryo at the blastula stage, and this “animalizing” effects has been shown to be reversible by chelation of excess Zn*+ by EDDA and EGTA after Zn*+ treatment [2&22]. The addition of chelator alone reportedly had no effect on development. In contrast, the results of the present study show a significant effect of removal of Zn*+ ions (inhibition of spiculogenesis), which may be due to better penetration of the chelator, 1, lo-phenanthroline. Alternatively, the effect of 1, lo-phenanthroline may be due to its ability to chelate Fe*+ ions as well, and may be mimicking effects seen with the Fe*+ chelator, a-a’-dipyridyl. This compound blocks the collagen-processing enzymes, prolyl and lysyl hydroxylase, and inhibits spiculogenesis both in mixed cultures containing primary mesenchyme cells [23] and in micromere cultures [24]. Other inhibitors of collagen processing have been shown to arrest normal embryos at the mesenchyme blastula stage as well [25, 261. Spiculogenesis was also inhibited by the metalloendoprotease dipeptide substrate, CBZ-Gly-Phe-NH*, both in cultures of primary mesenchyme cells and in the normal embryo. A control dipeptide, CBZ-Gly-Gly-NH2, that is not a substrate for metalloendoproteases had no effect on the cells. Thus, terminal differentiation of primary mesenchyme cells, the formation of CaC03-containing spicules within a syncytial cavity, may require the activity of a metalloendoprotease. Using assays with chelators, dipeptide substrate, or an active site inhibitor, the presence of such an enzyme was demonstrated in homogenates of primary mesenchyme cells isolated from mesenchyme blastula stage embryos or cultured from 16-cell micromeres. Possibly this metalloendoprotease is involved in the fusion of the filopodia of the cells to form syncytia and/or in subsequent membrane fusion events. Experiments are being designed to more precisely define the locus of action of the enzyme. We thank Diana Welch for manuscript preparation, Oscar Fong for help with sea urchin cultures, Dr. Fred Wilt, University of California, Berkeley, for advice on micromere culture, Dr. Bruce Kabakoff for micromeres, and members of the laboratory for critical review of this manuscript. This work was supported by NIH Grants HD 18590 and HD 21483 to W. J. Lennarz. Dr. Lennarz, who is a Robert A. Welch Professor of Chemistry, gratefully acknowledges the Robert A. Welch Foundation. Judith L. Roe acknowledges the support of a Rosalie B. Hite Fellowship.
REFERENCES 1. Mundy, D. I., and Strittmatter, 2. Couch, C. B., and Strittmatter,
W. J. (1985) Cell 40, 645-656. W. J. (1983) Cell 32, 257-265.
550
Roe et al.
3. Couch, C. B., and Strittmatter, W. .I. (1984) J. Biol. Chem. 259, 5396-5399. 4. Farach, H. A., Mundy, D. I., Strittmatter, W. J., and Lennarz, W. J. (1987) J. Biol. Chem.
262,
5483-5487.
5. Roe, J. L., Farach, H. A., Strittmatter,
W. J., and Lennarz,
W. J. (1988) J. Cell Biol.
107,
539-544.
Decker, G. L., and Lennarz, W. J. (1988) Dezwlopment 103, 231-247. Venkatasubramanian, K., and Solursh, M. (1984) Exp. Cell Res. 154, 421-431. 8. Carson, D. D., Farach, M. C., Earles, D. S., Decker, G. L., and Lennarz, W. J. (1985) Cell 41,
6. 7.
639-648. 9.
10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20.
Heifetz, A., and Lennarz, W. J. (1979) J. Biol. Chem. 254, 6119-6127. Barrett, A. J. (1977) in Proteinases in Mammalian Cells and Tissues (Barrett, A. J., Ed.), pp. l-55, Elsevier, Amsterdam. Morihara, K. (1974) Adv. Enzymol. 41, 179-243. Strittmatter, W. J., Couch, C. B., and Mundy, D. I. (1985) in Membrane Fluidity in Biology (Aloia, R. C., and Boggs, J., Eds.), Vol. 4, pp. 259-291, Academic Press, New York. Okazaki, K. (1965) Exp. Cell Res. 40, 585-596. Gustafson, T., and Wolpert, L. (1961) Exp. Cell Res. 24, 64-79. Okazaki, K. (1960) Embryologia 5, 283-320. Gibbins, J. R., Tilney, L. G., and Porter, K. R. (1969) J. Cell Biol. 41, 201-226. Hagstrom, B. E., and Lonning, S. (1969) Protoplasma 68, 271-288. Okazaki, K. (1975) Amer Zool. 15, 567-581. Decker, G. L., Morrill, J. B., and Lennarz, W. J. (1987) Development 101, 297-312. Nemer, M., Wilkinson, D. G., and Travaglini, E. C. (1985) Dev. Biol. 109, 418-427.
21. Nemer, M. (1986) Dev. Biol. 114, 214-224. 22. Mitsunaga, K., Fujiwara, A., Yoshima, T., and Yasomaso, I. (1983) Dev. Growth Differ. 25, 249-260. 23. Mintz, G. R., DeFrancesco, S., and Lennarz, W. J. (1981) J. Biol. Chem. 256, 13,105-13,111. 24. Blankenship, J., and Benson, S. (1984) Exp. Cell Res. 152, 98-104. 25. Wessel, G. M., and McClay, D. R. (1987) Dev. Biol. 121, 149-165. 26. Butler, E., Hardin, J., and Benson, S. (1987) Exp. Cell Res. 173, 174-182.
Received October 19, 1988 Revised version received November 22, 1988
Printed
in Sweden
Inhibitors
JUDITH
Cell Research 181 (1989) 542-550
of Metalloendoproteases Block Spiculogenesis Urchin Primary Mesenchyme Cells L. ROE,
Department of Biochemistry
HELEN R. PARK, WARREN and WILLIAM J. LENNARZ*
in Sea
J. STRITTMATTER,’
and Molecular Biology, UT M.D. Anderson Holcombe Boulevard, Houston, Texas 77030
Cancer Center, 1515
Metalloendoproteases have been implicated in a variety of fusion processes including plasma membrane fusion and exocytosis. As a prerequisite to skeleton formation in the sea urchin embryo, primary mesenchyme cells undergo fusion via tilopodia to form syncytia. The spicule is formed within the syncytial cable by matrix and mineral deposition. To investigate the potential involvement of a metahoendoprotease in spiculogenesis, the effect of inhibitors of this enzyme on skeleton formation was studied. Experiments with primary mesenchyme cells in vitro and in normal embryos revealed that skeleton formation was blocked by these inhibitors. These findings implicate a metalloendoprotease in spiculogenesis; such an enzyme has been demonstrated in homogenates of primary mesenchyme cells. The most likely site of action of the metalloendoprotease is at the cell membrane fusion stage and/or at subsequent events requiring membrane fusion. @ 1989 Academic FWSS, Inc.
Previous studies have implicated metalloendoproteases in a variety of cellular processes involving membrane fusion, such as exocytosis in mast cells and adrenal chromaffin cells [l] and plasma membrane fusion in myoblasts [2, 31. In addition, studies with sea urchin sperm have revealed the presence of a metalloendoprotease, inhibitors of which block the acrosome reaction [4]. Moreover, studies have implicated metalloendoproteases in sea urchin gamete fusion [5]. Another potential site of action of metalloendoproteases in a fusion event is syncytial formation of primary mesenchyme cells of the sea urchin embryo. These cells, which function in assembly of the calcite skeleton of the prism stage embryo, fuse with each other via filopodia to form a syncytium within which the CaC03-containing spicule is formed (for review see [6]). In view of our interests in skeletogenesis and membrane fusion, we have studied the effects of metalloendoprotease inhibitors and dipeptide substrates on cultures of isolated primary mesenchyme cells which form spicules in vitro in the presence of horse serum 17, 81. The results show that both an inhibitor and a dipeptide substrate of metalloendoproteases block the formation of spicules in vitro. The inhibitors also arrest spicule formation in the embryo. Direct evidence was obtained for a metalloendoprotease in homogenates prepared from primary mesenchyme cells isolated from mesenchyme blastula stage embryos or cultured from micromeres. From these studies we infer that a metalloendoprotease may be involved in the process of
I Department
of Neurology,
Baylor
College
of Medicine,
2 To whom reprint requests should be addressed. Copyright @ 1989 by Academic Press, Inc. All rights of reproduction in any form reserved 0014-4827189 $03.00
542
Houston,
TX 77030.
Metalloendoproteases
spiculogenesis, possibly by inhibiting mesenchyme cells. MATERIALS
and spiculogenesis
fusion of the cell membranes
AND
543
of the primary
METHODS
Materials. Strongylocentrotus purpuratus were obtained from Pacific Biomarine Supply Company (Venice, CA) and Marinus (Long Beach, CA). Falcon tissue culture dishes were obtained from Becton Dickinson and Co. (Lincoln Park, NJ). CBZ’-Gly-Phe-NH* and l,lO-phenanthroline were obtained from Sigma (St. Louis, MO) and CBZ-Gly-Gly-NH2 was obtained from Vega Biotechnologies, Inc. (Tucson, AZ). All other chemicals were reagent grade. Embryo culture andprimary mesenchyme cell culture. Gametes from S. purpuratus were obtained and embryos were cultured as described [9]. Primary mesenchyme cells were cultured essentially as described [8] with the addition of 1% penicillin-streptomycin (GIBCO Labs, Grand Island, NY), 0.2 gm/liter Ticarcillin disodium (Beecham Laboratories, TN), and 10 mg&ter erythromycin (Abbott Labs, North Chicago, IL) to the culture medium. Micromeres were isolated from 16-cell stage embryos by the method of Okazaki [18] and cultured in millipore-filtered artificial seawater (MFSW) supplemented with 3 % horse serum and 5 @ml gentamycin (GIBCO Labs). Metalloendoprotease assays. Metalloendoprotease activity was assayed using the fluorogenic protease substrate succinyl-Ala-Ala-Phe-aminomethylcoumarin (Sigma) (Succ-Ala-Ala-Phe-AMC). Primary mesenchyme cells, isolated from mesenchyme blastula stage embryos or from cells cultured from 16-cell stage micromeres, were assayed for metalloendoprotease activity in a crude, cell-free homogenate. Cells were scraped off the culture dish, washed three times in MFSW, and homogenized by three-freeze-thaw cycles in 50 n&f Hepes, pH 7.0, 150 mM KCl. Twenty-five microliters of homogenate, containing 1 to 5 ug protein, were incubated in 0.4 mM Succ-Ala-Ala-Phe-AMC in 5 % DMSO with 5ug chymostatin, in a total volume of 100 pl for 24 h at 37°C with or without the addition of metalloendoprotease inhibitors EGTA (5 mM), CBZ-Gly-Phe-NH2 (2 mM), or active site inhibitor 2-(iV-hydroxycarboxamido)-4-methylpentanoyl-~-alanyl glycine amide (Zincov, Calbiochem) (20 l&f). After incubation, free unquenched AMC was liberated from the hydrolyzed substrate by incubation for 30 min at 37°C with 20 @ml aminopeptidase M (Boehringer-Mannheim) with 1 x 10-r M phosphoramidon to inhibit a contaminating metalloendoprotease. Fluorescence (excitation 380 nm, emission 460 nm) of free AMC was measured in an Amino Bowman fluorometer. The amount of AMC produced was then determined by comparison with a standard curve.
RESULTS Inhibitors of Metalloendoproteases Block Spicule Formation in Primary Mesenchyme Cell Cultures
The chelator, l,lO-phenanthroline (a heavy metal chelator with a high affinity for Zn2+) has been shown to inhibit myoblast fusion [2] and mast and adrenal chromaffin cell exocytosis [l]. These findings, coupled with the demonstration that a Zn2+-dependent metalloendoprotease activity is present in these cells, have led to the hypothesis that this enzyme may be required for membrane fusion [l-3]. Because the formation of spicules is preceded by fusion of primary mesenthyme cells, we tested the effect of this chelator on spicule formation by these cells. As previously described [8], primary mesenchyme cells were cultured in vitro by plating dissociated cells from the mesenchyme blastula stage (2 h posthatching). The primary mesenchyme cells were allowed to adhere to the culture dish and the unattached cells were washed out 18 h after plating. As shown in Fig. 1, ’ CBZ, carbobenzoxy; EGTA, 35-898334
ethylene
glycol
DMSO, dimethyl sulfoxide; EDDA, ethylenediamine-N’,N’-diacetic bis(j3-aminoethyl
ether)-N’,N’,N’,N’-tetraacetic
acid.
acid;
544
Roe et al.
Fig. I. Effect of l,lO-phenanthrohne on cultures of isolated primary mesenchyme cells. Primary mesenchyme cells were isolated as described under Materials and Methods. Nonadherent cells were removed at 18 h and cultures were incubated at 14°C for 72 h. Control cultures (A) and cultures treated with 0.05% DMSO (B) showed extensive spicule formation. Alternatively, cultures were incubated with 10 (C), 50 (D), 100 (,?!I),or 250 @4 (fl l,lO-phenanthroline immediately after removal of nonadherent cells and cultured for an additional 72 h. (Magnification 47x .)
untreated cultures showed extensive spicule formation 72 h after removal of nonadherent cells (Fig. 1 A), as did cultures treated with DMSO alone (Fig. 1 B) or with 10 @4 (Fig. 1 C) or 50 @4 (Fig. 1 D) l,lO-phenanthroline. l,lO-Phenanthroline was added to such cultures after the nonadherent cells had been removed. However, inhibition of spicule formation after 72 h was seen in cultures treated with 100 u.A4 1, lo-phenanthroline (Fig. 1 E) and spicule formation was completely inhibited with 250 @4 1, lo-phenanthroline (Fig. 1fl. The inhibitory effect was not the result of inhibition of primary mesenchyme cell attachment because cells did attach in the presence of the chelator (data not shown). Filopodial extensions were not seen between cells in 1, lo-phenanthroline-treated cultures. The effect of
Metalloendoproteases
and spiculogenesis
545
Fig. 2. Effect of dipeptides on cultures of primary mesenchyme cells. Cultures of primary mesenthyme cells (see Materials and Methods) were cultured in normal culture medium (A) or treated with 2 rnIU CBZ-Gly-Gly-NH2 (B) or 2 mA4 CBZ-Gly-Phe-NH2 (C) after removal of nonadherent cells and cultured an additional 72 h. (D), cultures were incubated for 72 h in 2 m&I CBZ-Gly-Phe-NH2 and washed and then the cells were replated in fresh media and incubated an additional 120 h. (Magnification (A-C) 47x and (0) 94x .)
1, IO-phenanthroline on spiculogenesis was irreversible, consistent with previous reports on its effect on myoblast fusion [2] and metalloendoprotease activity [ 101. The effects of 1, IO-phenanthroline are not likely to be due to Ca” chelation, as Ca*+ ions are in great excess in seawater. However, because this chelator may have pleiotropic effects, we tested the effect of more specific inhibitors of metalloendoprotease activity. Studies on the specificity of metalloendoproteases using various peptide substrates have shown that these proteases cleave the peptide bond on the amino side of uncharged aromatic and large aliphatic amino acids. Dipeptides of the general type Gly-Phe or Gly-Leu containing their N- and C-termini blocked have been shown to be substrates of these proteases [ll]; when they are added in excess they would be expected to inhibit the action of the metalloendoprotease on endogenous substrates. Indeed, as shown in Fig. 2, the metalloendoprotease substrate CBZ-Gly-Phe-NH2 was found to completely inhibit spicule formation in the primary mesenchyme cell cultures (Fig. 2 C). Cells incubated with the control dipeptide CBZ-Gly-Gly-NH2 (Fig. 2B), which is not a substrate for metalloendoprotease, formed spicules as well as control cultures formed (Fig. 2A). The cells cultured in the presence of CBZ-Gly-Phe-NH2 filopodia to a variable extent as seen in the early stages of control cultures. It was found that the effect of the inhibitory dipeptide was reversible (Fig. 20); when
546
Roe et al.
3. Effect of 1, IO-phenanthroline on spiculogenesis in vivo. Early gastrula stage embryos (36 h) were cultured for an additional 30 h in the absence (control) (II) or presence of 1 m&I l,lOphenanthroline in 0.2% DMSO (CT),in 2.5 m&f l,lO-phenanthroline in 0.5% DMSO (D), or in 0.5% DMSO alone (I?). Embryos were viewed using Normarski optics. (Magnification 196x .) Fig.
(A)
the dipeptide was removed the cells were able to resume differentiation and form spicules. More cells were observed to cluster around the spicules than in control cultures. Metalloendoprotease Inhibitors Block Spiculogenesis in the Embryo
Because the metalloendoprotease inhibitors were found to block spiculogenesis in cultures of isolated primary mesenchyme cells, we tested the effects of the inhibitors on spiculogenesis in normal embryos. Inhibitors were added to cultures of S. purpuratus embryos at the gastrula stage (Fig. 3A), when the archenteron has formed and the primary mesenchyme cells are migrating within the blastocoel. As shown in Fig. 3, embryos were treated with l,lO-phenanthroline at 36 h and cultured for an additional 30 h. No spicules were formed in embryos cultured
Metalloendoproteases
and spiculogenesis
547
Fig. 4. Effect of dipeptides on spiculogenesis in vim Early gastrula stage embryos (36 h) (Fig. 3A) were cultured for an additional 30 h in the absence (control) (A) or presence of 2 mA4 CBZ-Gly-GlyNH2 in 0.5% DMSO (B) or 2 mA4 CBZ-Gly-Phe-NH2 in 0.5% DMSO (C). Embryos were viewed using Nomarski optics. (Magnification 196x .)
in 1 mM (Fig. 3 C) or 2.5 mM (Fig. 3 0) l,lO-phenanthroline. By this time (66 h) untreated cultures had~ developed complete triradial skeletons (Fig. 3B). The inhibition was not reversible, although the embryos continued to swim and appeared viable. The concentration of 1 ,lO-phenanthroline required to inhibit spiculogenesis in the embryo was 4- to IO-fold higher than observed in the primary mesenchyme cell cultures. In other experiments it was found that if the chelator was added after the formation of a small triradial skeleton (40 h), further elongation of the spicule was inhibited (data not shown), suggesting that I ,lOphenanthroline can block spicule deposition. The dipeptide metalloendoprotease substrate, CBZ-Gly-Phe-NHz, also inhibited spicule formation in the embryo. When CBZ-Gly-Phe-NH2 was added to embryos at 36 h and the embryos were cultured for 30 h, as in the experiment described in Fig. 3, no spicules were observed (Fig. 4C), while untreated embryos formed complete skeletons (Fig. 4A). As shown, addition of the dipeptide CBZ-Gly-Gly-NH*, which is not a metalloendoprotease substrate, had no effect on the embryos (Fig. 4B). These data show that inhibitors of a metalloendoprotease are able to inhibit spiculogenesis in the normal embryo as well as in isolated cultures of primary mesenchyme cells.
548
Roe et al.
Primary
Mesenchyme
Cells Contain Metalloendoprotease
Activity
Metalloendoprotease activity was identified in primary mesenchyme cells using the synthetic fluorogenic endoprotease substrate succinyl-Ala-Ala-Phe-AMC, used by us previously to characterize this protease in other cells [ 1, 41. Homogenates prepared from primary mesenchyme cells isolated from blastula stage embryos or cultured from 16-cell stage micromeres were found to contain endoprotease activity that hydrolyzed succinyl-Ala-Ala-Phe-AMC at a linear rate for 24 h. The metal chelator EGTA (5 mM) inhibited substrate hydrolysis by 85 %; this metal-dependent endoprotease activity ranged from 4.0 to 34~ 10-l’ mol substrate hydrolyzed/24h/ug protein in five different cell preparations. The metal chelator 1, lo-phenanthroline cannot be used in this assay because it inhibits the aminopeptidase M used in this coupled assay (see Materials and Methods and [l]). The dipeptide CBZ-Gly-Phe-NH* (2 mM) inhibited hydrolysis of the synthetic substrate by 50% and the active site metalloendoprotease inhibitor, 2-W hydroxycarboxamido)-4-methylpentanoyl-L-alanylglycine amide (20 uM) (Zincov), inhibited substrate hydrolysis by 60%. These data demonstrate the presence of a metalloendoprotease activity in the spicule forming cells that is inhibited by the same peptide substrate shown to block spicule formation. DISCUSSION The mechanism of membrane fusion has been studied using a variety of model systems including cell-cell fusion, vesicle-vesicle fusion, vesicle-plasma membrane fusion, and virus-host cell fusion. Many of these events are Ca*+-dependent and may require the activity of a Zn*+ -dependent metalloendoprotease [ 121. Although in some systems the generation of a fusogenic protein by proteolysis is required for fusion, a general hypothesis of the mechanism of involvement of metalloendoproteases has not yet been formed. In the sea urchin sperm, we have recently described a Zn*+-dependent metalloendoprotease that may be involved in the acrosome reaction, an event that requires vesiculation of the membrane overlying the acrosomal granule [4]. Sea urchin gamete fusion also appears to require the activity of a metalloendoprotease [5]. In the current report we have studied the effects of metalloendoprotease inhibitors on the differentiation of primary mesenchyme cells of the sea urchin embryo. During this differentiation, a syncytial structure within which the skeleton is formed is generated by fusion of the filopodia of the primary mesenchyme cells. The nature of the syncytia formed by primary mesenchyme cells in vivo has been studied at both the light [13-151 and electron microscope levels [6, 16-191. the cells are connected via a syncytial cable which was formed by the fusion of tilopodia. The cell bodies are often connected to the syncytial cable to short stalks. The spicule is formed within the cable by mineral and matrix deposition, and the syncytial cable is connected by tilopodial extensions to the blastocoel wall. The primary mesenchyme cells appear to be highly fusogenic and Okazaki [13] has observed that the filopodia of cells can fuse, release, and re-fuse with another cell in vivo during the syncytial formation. The tilopodia appear to fuse
Metalloendoproteases
and spiculogenesis
549
only at the tips, suggesting localized fusogenic sites. The syncytia appears to be a requirement for spicule formation because single cells have been observed only to form small calcareous deposits [18]. We carried out studies with two types of inhibitors of metalloendoproteases, a Zn*+ chelator and dipeptide substrates of the enzyme that have been shown to inhibit fusion events in other systems [l-5]. The effect of these agents was examined in normal embryos and cultures of isolated primary mesenchyme cells. l,lO-Phenanthroline (a heavy metal chelator and metalloendoprotease inhibitor) blocked the formation of spicules by primary mesenchyme cells both in isolated primary mesenchyme cell cultures and in the normal embryo. Addition of excess Zn*+ ions is known to arrest development in the sea urchin embryo at the blastula stage, and this “animalizing” effects has been shown to be reversible by chelation of excess Zn*+ by EDDA and EGTA after Zn*+ treatment [2&22]. The addition of chelator alone reportedly had no effect on development. In contrast, the results of the present study show a significant effect of removal of Zn*+ ions (inhibition of spiculogenesis), which may be due to better penetration of the chelator, 1, lo-phenanthroline. Alternatively, the effect of 1, lo-phenanthroline may be due to its ability to chelate Fe*+ ions as well, and may be mimicking effects seen with the Fe*+ chelator, a-a’-dipyridyl. This compound blocks the collagen-processing enzymes, prolyl and lysyl hydroxylase, and inhibits spiculogenesis both in mixed cultures containing primary mesenchyme cells [23] and in micromere cultures [24]. Other inhibitors of collagen processing have been shown to arrest normal embryos at the mesenchyme blastula stage as well [25, 261. Spiculogenesis was also inhibited by the metalloendoprotease dipeptide substrate, CBZ-Gly-Phe-NH*, both in cultures of primary mesenchyme cells and in the normal embryo. A control dipeptide, CBZ-Gly-Gly-NH2, that is not a substrate for metalloendoproteases had no effect on the cells. Thus, terminal differentiation of primary mesenchyme cells, the formation of CaC03-containing spicules within a syncytial cavity, may require the activity of a metalloendoprotease. Using assays with chelators, dipeptide substrate, or an active site inhibitor, the presence of such an enzyme was demonstrated in homogenates of primary mesenchyme cells isolated from mesenchyme blastula stage embryos or cultured from 16-cell micromeres. Possibly this metalloendoprotease is involved in the fusion of the filopodia of the cells to form syncytia and/or in subsequent membrane fusion events. Experiments are being designed to more precisely define the locus of action of the enzyme. We thank Diana Welch for manuscript preparation, Oscar Fong for help with sea urchin cultures, Dr. Fred Wilt, University of California, Berkeley, for advice on micromere culture, Dr. Bruce Kabakoff for micromeres, and members of the laboratory for critical review of this manuscript. This work was supported by NIH Grants HD 18590 and HD 21483 to W. J. Lennarz. Dr. Lennarz, who is a Robert A. Welch Professor of Chemistry, gratefully acknowledges the Robert A. Welch Foundation. Judith L. Roe acknowledges the support of a Rosalie B. Hite Fellowship.
REFERENCES 1. Mundy, D. I., and Strittmatter, 2. Couch, C. B., and Strittmatter,
W. J. (1985) Cell 40, 645-656. W. J. (1983) Cell 32, 257-265.
550
Roe et al.
3. Couch, C. B., and Strittmatter, W. .I. (1984) J. Biol. Chem. 259, 5396-5399. 4. Farach, H. A., Mundy, D. I., Strittmatter, W. J., and Lennarz, W. J. (1987) J. Biol. Chem.
262,
5483-5487.
5. Roe, J. L., Farach, H. A., Strittmatter,
W. J., and Lennarz,
W. J. (1988) J. Cell Biol.
107,
539-544.
Decker, G. L., and Lennarz, W. J. (1988) Dezwlopment 103, 231-247. Venkatasubramanian, K., and Solursh, M. (1984) Exp. Cell Res. 154, 421-431. 8. Carson, D. D., Farach, M. C., Earles, D. S., Decker, G. L., and Lennarz, W. J. (1985) Cell 41,
6. 7.
639-648. 9.
10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20.
Heifetz, A., and Lennarz, W. J. (1979) J. Biol. Chem. 254, 6119-6127. Barrett, A. J. (1977) in Proteinases in Mammalian Cells and Tissues (Barrett, A. J., Ed.), pp. l-55, Elsevier, Amsterdam. Morihara, K. (1974) Adv. Enzymol. 41, 179-243. Strittmatter, W. J., Couch, C. B., and Mundy, D. I. (1985) in Membrane Fluidity in Biology (Aloia, R. C., and Boggs, J., Eds.), Vol. 4, pp. 259-291, Academic Press, New York. Okazaki, K. (1965) Exp. Cell Res. 40, 585-596. Gustafson, T., and Wolpert, L. (1961) Exp. Cell Res. 24, 64-79. Okazaki, K. (1960) Embryologia 5, 283-320. Gibbins, J. R., Tilney, L. G., and Porter, K. R. (1969) J. Cell Biol. 41, 201-226. Hagstrom, B. E., and Lonning, S. (1969) Protoplasma 68, 271-288. Okazaki, K. (1975) Amer Zool. 15, 567-581. Decker, G. L., Morrill, J. B., and Lennarz, W. J. (1987) Development 101, 297-312. Nemer, M., Wilkinson, D. G., and Travaglini, E. C. (1985) Dev. Biol. 109, 418-427.
21. Nemer, M. (1986) Dev. Biol. 114, 214-224. 22. Mitsunaga, K., Fujiwara, A., Yoshima, T., and Yasomaso, I. (1983) Dev. Growth Differ. 25, 249-260. 23. Mintz, G. R., DeFrancesco, S., and Lennarz, W. J. (1981) J. Biol. Chem. 256, 13,105-13,111. 24. Blankenship, J., and Benson, S. (1984) Exp. Cell Res. 152, 98-104. 25. Wessel, G. M., and McClay, D. R. (1987) Dev. Biol. 121, 149-165. 26. Butler, E., Hardin, J., and Benson, S. (1987) Exp. Cell Res. 173, 174-182.
Received October 19, 1988 Revised version received November 22, 1988
Printed
in Sweden