ã

Oncogene (2001) 20, 2068 ± 2079 2001 Nature Publishing Group All rights reserved 0950 ± 9232/01 $15.00 www.nature.com/onc

UCS15A, a non-kinase inhibitor of Src signal transduction Sreenath V Sharma*,1, Chitose Oneyama1, Yoshinori Yamashita2, Hirofumi Nakano1, Katsura Sugawara3, Masako Hamada1, Nobuo Kosaka2 and Tatsuya Tamaoki2 1

Tokyo Research Laboratories, Kyowa Hakko Kogyo Co., Ltd. 3-6-6 Asahi-cho, Machida-shi, Tokyo 194, Japan; 2Pharmaceutical Research Laboratories, Kyowa Hakko Kogyo Co., Ltd 1188 Shimotogari, Nagaizumi-cho, Sunto-gun, Shizuoka 411, Japan; 3 Okazaki National Research Institute, National Institute for Basic Biology, Division of Morphogenesis, Nishigonaka 38, Myodaji, Okazaki 444-8585 Aichi, Japan

Src tyrosine kinase plays key roles in signal transduction following growth factor stimulation and integrinmediated cell-substrate adhesion. Since src-signal transduction defects are implicated in a multitude of human diseases, we have sought to develop new ways to identify small molecule inhibitors using a yeast-based, activatedsrc over-expression system. In the present study, we describe the identi®cation of a unique src-signal transduction inhibitor, UCS15A. UCS15A was found to inhibit the src speci®c tyrosine phosphorylation of numerous proteins in v-src-transformed cells. Two of these phosphoproteins were identi®ed as bona-®de src substrates, cortactin and Sam68. UCS15A di€ered from conventional src-inhibitors in that it did not inhibit the tyrosine kinase activity of src. In addition, UCS15A appeared to di€er from src-destabilizing agents such as herbimycin and radicicol that destabilize src by interfering with Hsp90. Our studies suggest that UCS15A exerted its src-inhibitory e€ects by a novel mechanism that involved disruption of protein-protein interactions mediated by src. One of the biological consequences of src-inhibition by UCS15A was its ability to inhibit the bone resorption activity of osteoclasts in vitro. These data suggest that UCS15A may inhibit the bone resorption activity of osteoclasts, not by inhibiting src tyrosine kinase activity, but by disrupting the interaction of proteins associated with src, thereby modulating downstream events in the src signal transduction pathway. Oncogene (2001) 20, 2068 ± 2079. Keywords: Src; UCS15A; bone resorption inhibitor; protein-protein interaction Introduction Members of the Src-family of non-receptor tyrosine kinase (Src, Fyn, Yes, Fgr, Lyn, Hck, Lck, Blk and Yrk) are ubiquitously expressed and play diverse roles in biological systems (reviewed in Brown and Cooper, 1996). The prototypic member of this family, src, has

*Correspondence: SV Sharma Received 11 September 2000; revised 24 January 2001; accepted 25 January 2001

the distinction of being not only the ®rst member to highlight the connection between oncogenes and their cellular counterparts, proto-oncogenes (Stehlin et al., 1976), but was also the ®rst tyrosine kinase identi®ed (Hunter and Sefton, 1980). While the tyrosine kinase activity of src has received much attention in growth factor-mediated signal transduction and transformation, more recent studies highlighting the kinaseindependent, sca€olding function of src represent a major paradigm shift in our understanding of this important protein (reviewed in Schwartzberg, 1998). This functional duality is re¯ected in the primary structure of the protein which contains, in addition to the catalytic kinase domain, two protein-protein interaction modules termed SH2 and SH3 domains (reviewed in Kay et al., 2000). The presence of these proteinprotein interaction domains underscores the adapter functions of src that play important roles not only in signal transduction by intermolecular interactions, but also in the regulation of src, itself, by intramolecular interactions. The importance of the latter interactions in regulating src function was highlighted by X-ray crystallographic studies (Xu et al., 1997). Src kinase plays essential roles in growth factormediated signal transduction, integrin-mediated signal transduction and, probably also, in mitotic progression of cells through its interaction with Sam68 (reviewed in Thomas and Brugge, 1997). In addition, activation of c-src and src kinase activity is a recurrent theme in Gprotein coupled receptor-induced ras-dependent signaling (reviewed in Luttrell et al., 1999). In growth factormediated signaling, critical substrates downstream of src include, among others, RasGAP, SHPTP2/Syp, p85 subunit of PI3kinase, SHC and PLCg (reviewed in Brown and Cooper, 1996). In integrin-mediated signaling, downstream substrates include FAK, vinculin, paxillin, tensin, cortactin, pp130cas and p110 AFAP (reviewed in Brown and Cooper, 1996; Schwartzberg, 1998). Less well characterized are the mitotic substrates of src, which include Sam68 and hnRNPK (Fumagalli et al., 1994; Taylor and Shalloway, 1994; Weng et al., 1994). While growth factormediated signaling requires the kinase activity of src, integrin-mediated signaling is kinase-independent and probably involves the adapter functions of src (reviewed in Schwartzberg, 1998).

UCS15A inhibits osteoclastic bone resorption SV Sharma et al

All small-molecule src kinase inhibitors, described to date, fall into three broad categories. The ®rst category, exempli®ed by Genistein, 67B-83-A, Staurosporin, WIN 61651, Tyrphostins, PP1, PP2 and CGP77675, consist of molecules that interfere with the kinase activity of src itself (Hanke et al., 1996 and references therein; Missbach et al., 1999). The second category includes agents, such as radicicol/monorden, geldanamycin and herbimycin, that interfere with src stability by blocking the function of its associated molecular chaperone, hsp90 (Whitesell et al., 1994; Stebbins et al., 1997; Schulte et al., 1998, 1999; Sharma et al., 1998; Roe et al., 1999). The third category consists of small molecule inhibitors that disrupt protein-protein interactions involving the src protein. This new category is currently represented by a single compound, AP22161, that disrupts SH2 domainmediated interactions of src with its associated proteins (Violette et al., 1999). Although the Src protein shows widespread cellular distribution, three cell-types, namely, platelets, neurons and osteoclasts, express the highest levels of src in the body (reviewed in Thomas and Brugge, 1997). Osteoclasts are multinucleated giant cells that are thought to originate from the same progenitors as the monocyte-macrophage lineage (Kurihara et al., 1990; reviewed in Roodman, 1999). Osteoclasts are involved in bone turnover and play important roles in bone resorption (reviewed in Roodman, 1999). Targeted disruption of src in mice leads to only one major phenotype, namely, osteopetrosis (Soriano et al., 1991), a condition that is characterized by a de®ciency in the bone resorption activity of osteoclasts. Further analysis reveals that the src7/7 mice have normal numbers of osteoclasts but that their bone resorption activity is de®cient (Lowe et al., 1993). Expression of src is essential for osteoclasts to form ru‚ed borders and resorb bone (Boyce et al., 1992). Surprisingly, however, osteoclast-speci®c expression of a kinase-inactive mutant of src was able to partially rescue the osteopetrotic phenotype of src7/7 mice, suggesting that kinase-independent functions of src are more important in osteoclast-mediated bone resorption (Schwartzberg et al., 1997). Kinase-inactive mutants of src are also able to rescue the cell spreading defect of ®broblasts from src7/7 mice (Kaplan et al., 1995), suggesting a possible link between integrin signaling and the osteopetrotic phenotype of src7/7 mice. Taken together, these studies underscore the importance of the src signal transduction pathway (especially its kinaseindependent functions) in bone resorption by osteoclasts and suggest that this pathway could be a useful target for bone resorption inhibitors. Osteoporosis is generally thought to result from an imbalance between bone formation and bone resorption. This imbalance may result from either a decrease in bone formation with age and/or an uncoupling of bone resorption and bone formation (reviewed in Manolagas, 2000). Regardless of the exact mechanism of osteoporosis, it is generally believed that controlling the bone resorption activity of osteoclasts might be an important

means of controlling the disease. Indeed, several src kinase inhibitors such as Genistein, herbimycin A, radicicol, geldanamycin, AP22161 and CGP77675 have been shown to inhibit osteoclast-mediated bone resorption (Yoneda et al., 1993; Hall et al., 1994; Blair et al., 1996; Missbach et al., 1999; Violette et al., 1999). More recently, induction of the csk gene, an upstream inhibitor of src-kinase activity, was also shown to inhibit osteoclastic functions (Miyazaki et al., 2000), a result that is consistent with the indispensable nature of src in osteoclast-mediated bone resorption. In addition, src kinase activity is signi®cantly up regulated in human cancers, particularly colon and breast cancers (Jacobs and Rubsamen, 1983; Bolen et al., 1987) indicating the widespread role of src in human diseases. In our laboratory, src signal transduction inhibitors have been screened using activated src-over expressing yeast strains as a test organism. Given the large numbers of currently available src-kinase inhibitors, we were particularly interested in identifying nonkinase inhibitors of src signal transduction. Using this system, UCS15A, an antibiotic produced by Streptomyces sp., was identi®ed as a non-kinase inhibitor of src signal transduction. As a test of the physiological relevance of UCS15A in src-signaling inhibition, we examined its e€ects on the bone resorption activity of osteoclasts. UCS15A showed inhibitory e€ects on the bone resorption activity of multinucleated osteoclastic cells in vitro, as well as in a mouse calvaria organ culture system.

2069

Results Discovery of UCS15A, a new src signal transduction inhibitor, from a screen for src-signaling antagonists, using yeast-based high-throughput assays Over-expression of v-src, under the control of the Gal1 promoter in S. cerevisiae, has been shown to cause growth inhibition (Brugge et al., 1987; Kornbluth et al., 1987). Based on this observation, putative inhibitors of src signal transduction were screened for their ability to rescue the growth arrest of v-src over-expressing yeast strains (strategy outlined in Materials and methods). This method was used to screen microorganisms isolated from the soil for their ability to produce src inhibitors. Of 20 000 natural compounds screened, ®ve active compounds produced by Actinomycetes and fungi were isolated. These included well-known src inhibitors, such as herbimycin A and radicicol. An additional src inhibitor, designated UCS15A, produced by Streptomyces sp. was identi®ed as an active compound in these assays. Structural determination studies by Nuclear Magnetic Resonance and Mass Spectrometry revealed that the structure of UCS15A (shown in Figure 1) was identical to a previously identi®ed compound, SI4228A (Suzuki et al., 1983). It is interesting to note that even though the natural compound library contained several broadly speci®c, tyrosine kinase Oncogene

UCS15A inhibits osteoclastic bone resorption SV Sharma et al

2070

Figure 1 Chemical structure of UCS15A

inhibitors, such as Genistein and Staurosporin, these compounds were not identi®ed as positives in the assay system used. This indicated that the yeast-based assay system had a built-in bias for non-kinase inhibitors of src signal transduction (category 2 and 3 type of inhibitors). In addition, since the compounds were selected for their ability to induce growth of arrested yeast cells, it was likely that they were not detrimental to growth and hence, were not toxic. Since the mode of action of UCS15A was unknown, we focused our attention on this compound. Inhibition of tyrosine phosphorylation of proteins in v-src transformed NIH3T3 cells by UCS15A Since UCS15A appeared to inhibit src signal transduction in the yeast based assay system, we examined the e€ect of this compound on mammalian cells expressing the activated v-src oncogene. Previous studies have shown that cellular transformation by activated src resulted in the increased tyrosine phosphorylation of 15 to 30 cellular proteins (Kamps and Sefton, 1988; Kanner et al., 1990). Consistent with previous studies, NIH3T3 cells expressing the v-src oncogene (v-Src 3T3) speci®cally expressed several tyrosine phosphorylated proteins as compared with their untransformed parent line, NIH3T3 (Figure 2a left hand panel, compare lanes 6 ± 10 with lanes 1 ± 5). To ensure that appropriate amounts of protein were loaded in each lane, the nitrocellulose ®lter was stripped and stained with Coomassie Blue (Figure 2a, right hand panel). To examine the e€ect of UCS15A on the pro®le of tyrosine phosphorylated proteins in v-Src 3T3 cells, cells were treated with di€erent amounts of UCS15A and the tyrosine phosphorylation of proteins was examined by immunoblotting analysis, using an antiphosphotyrosine speci®c antibody (Figure 2b). Untreated v-src 3T3 cells contained a large number of tyrosine phosphorylated proteins that were tentatively designated I ± VI (Figure 2a,b, lane 1). Upon UCS15A treatment, the tyrosine phosphorylation of most of these proteins decreased in a dose-dependent manner (Figure 2b, lanes 2 ± 6). To determine the identity of some of these proteins, the immunoblot from 2b was stripped and re-probed with antibodies to several candidate proteins. The choice of candidates was Oncogene

dictated by known src substrates (reviewed in Brown and Cooper, 1996). Suggestive evidence that the proteins designated V and VI in Figure 2b might be cortactin and Sam68, respectively, was obtained by probing the ®lter from 2b with anti-cortactin and antiSam68 antibodies (Figure 2c,d, respectively). In addition, these results suggested that the levels of cortactin and Sam68 were unchanged following UCS15A treatment (Figure 2c,d, respectively), suggesting that, at the concentrations used, the drug did not have signi®cant cytotoxic e€ects. Even though the identity of the other phosphorylated proteins was not pursued in detail, based on the molecular weights it was possible that band III consisted of pp120FAK, pp130cas or both. Taken together these results suggested that UCS15A treatment resulted in the reduced tyrosine phosphorylation of several proteins, some of which may be src substrates. Inhibition of tyrosine phosphorylation of bona-fide src substrates, cortactin and Sam68 by UCS15A Previous studies have demonstrated that cortactin and Sam68 are bona-®de substrates of src (Wu et al., 1991; Fumagalli et al., 1994; Taylor and Shalloway, 1994). Given the tentative identi®cation of bands V and VI as cortactin and Sam68, respectively (Figure 2) more speci®c analyses were undertaken. To this end, cortactin and Sam68 proteins were immunoprecipitated from UCS15A treated v-src 3T3 cells and their tyrosine phosphorylation status was determined by immunoblotting with anti-phosphotyrosine antibodies (Figure 3). The tyrosine phosphorylation of both Sam68 and cortactin was inhibited by UCS15A in a dosedependent manner (Figure 3b,c, left-hand side panels). The levels of the two proteins were relatively unchanged by UCS15A treatment (Figure 3b,c, righthand side panels), as were the levels of src protein (Figure 3a). Together with results shown in Figure 2, these results strongly suggested that UCS15A inhibited the tyrosine phosphorylation of at least two src substrates, namely, cortactin and Sam68, without appreciable e€ects on their intracellular levels. UCS15A does not inhibit the tyrosine kinase activity of v-Src or its activation state Since tyrosine phosphorylation of src substrates might be dependent on the kinase activity of src, we examined src kinase activity in response to UCS15A by an in vitro kinase assay, using a synthetic peptide derived from cdc2, as a src family speci®c substrate (Cheng et al., 1992). The kinase activity of immunoprecipitated v-src protein was assayed in the presence of increasing concentrations of either UCS15A or a bona-®de src kinase inhibitor, PP2 (Hanke et al., 1996), which served as a positive control (Figure 4a). As described previously (Hanke et al., 1996), PP2 very e€ectively inhibited the in vitro kinase activity of src to about 20% of untreated controls (Figure 4a). Surprisingly, however, UCS15A failed to show any inhibition

UCS15A inhibits osteoclastic bone resorption SV Sharma et al

2071

Figure 2 Inhibition of tyrosine phosphorylation of proteins in v-src 3T3 cells by UCS15A. Di€erent amounts of cell lysate (amounts indicated at the top of each lane in panel (a) from NIH3T3 (lanes 1 ± 5) and V-Src 3T3 cells (lanes 6 ± 10) were separated by polyacrylamide gel electrophoresis and analysed by immunoblotting with the HRP-conjugated anti-phosphotyrosine antibody (left panel). The position of migration of six proteins of interest (I ± VI) is indicated to the left of the autoradiogram. The same nitrocellulose ®lter used for the anti-phosphotyrosine Western blot, was stripped and stained with Coomassie blue (right panel). To examine the e€ects of UCS15A on the pro®le of Tyrosine phosphorylated proteins in src-transformed cells, v-src 3T3 cells were treated for 24 h with increasing concentrations of UCS15A as indicated at the top of each lane (b, c and d). Following the treatment, cells were lysed and equivalent amounts of total cellular proteins were separated by SDS ± PAGE, transferred to nitrocellulose membranes, and immunoblotted with the antibodies indicated at the bottom of each panel. Approximate molecular weights of standards are indicated in kilodaltons in the right-hand margin of (d). (b) The membrane was probed with an anti-pTyr antibody PY20, conjugated to horseradish peroxidase and detected by chemiluminescence. (c) The membrane from (b) was stripped of the antibody probe, and reprobed with anti-cortactin antibody, and horseradish peroxidase-conjugated secondary antibody, and detected as described above. Position of migration of cortactin is indicated by * in the right-hand margin of (c). (d) The membrane from (b) was stripped of the antibody probe, and reprobed with anti-sam68 antibody. Position of migration of sam68 is indicated by * in the right-hand margin of panel (d)

of src-kinase activity, even at concentrations as high as 20 mM (Figure 4a). Previous studies have shown that the phosphorylation of tyrosine 416 in the src protein is essential for its kinase activity (reviewed in Brown and Cooper, 1996). Given the inability of UCS15A to inhibit the tyrosine kinase activity of v-src in vitro (Figure 4a) the e€ect of the drug was tested on the in vivo activation of v-src. To this end, the phosphorylation status of tyrosine 416 of src was assessed using a phosphospeci®c antibody for src 416 (Figure 4b).

Tyrosine 416, the major autophosphorylation site of src, was phosphorylated in v-src 3T3 cells (Figure 4b, lane 1) and UCS15A treatment had no e€ect on the in vivo activity of v-src since tyrosine 416 remained phosphorylated in the presence of the drug (Figure 4b, lanes 2 ± 6). Taken together, these data indicated that UCS15A did not inhibit the kinase activity of src, thus, its mechanism of action was distinct from that of other pharmacological src-kinase inhibitors, such as PP1, PP2, Genistein, Staurosporin and Tyrphostins. Oncogene

UCS15A inhibits osteoclastic bone resorption SV Sharma et al

2072

Figure 3 Inhibition of tyrosine phosphorylation of Sam68 and Cortactin by UCS15A-treatment. V-src 3T3 cells were treated for 24 h with the concentrations of UCS15A indicated at the top of each lane. Following the treatment, cells were lysed and equivalent amounts of total cellular proteins were either used directly (a) or immunoprecipitated (IP; b and c) with the antibodies indicated at the bottom of each panel. Proteins were separated by SDS ± PAGE, transferred to nitrocellulose membranes and immunoblotted (WB) with the antibodies indicated at the bottom of each panel. Approximate molecular weights of standards are indicated in kilodaltons on the left-hand margin of each panel. Proteins of interest are indicated in the right-hand margin of each panel. (a) Nitrocellulose membrane containing total cell lysates was probed with an anti-src antibody. (b) Anti-sam68 immunoprecipitates from UCS15A-treated v-src 3T3 cells were probed with either an anti-phosphotyrosine antibody (left-hand panel) or with antisam68 antibody (right-hand panel). Positions of migration of sam68 and immunoglobulin heavy chain (Ig hc) are indicated in the right-hand margin of (b). (c) Anti-cortactin immunoprecipitates from UCS15A-treated v-src 3T3 cells were probed with either an anti-phosphotyrosine antibody (left-hand panel) or with anti-cortactin antibody (right-hand panel). Positions of migration of cortactin and immunoglobulin heavy chain (Ig hc) are indicated in the right-hand margin of (c). Lanes 1 and 10 contains whole cell lysate as a reference

UCS15A is distinct from Src destabilizing agents such as radicicol and herbimycin Previous studies have identi®ed a distinct group of pharmacological agents such as herbimycin, geldanamycin and radicicol that were originally thought to be src-kinase inhibitors (Uehara et al., 1988; Fukazawa et al., 1991; Kwon et al., 1992a). Subsequent studies demonstrated that these agents destabilized src protein by interfering with the molecular chaperone hsp90 (Whitesell et al., 1994; Schulte et al., 1998, 1999; Sharma et al., 1998). In addition, these agents have been shown to reverse the transformed morphology of cells expressing not only the src oncogene, but ras and mos, as well (Uehara et al., 1985; Kwon et al., 1992b, 1995; Whitesell et al., 1994; Zhao et al., 1995). Despite the fact that UCS15A did not a€ect the intracellular levels of src protein (Figure 3a) the possibility existed that UCS15A could have some features in common with this group of agents. To investigate this possibility, the e€ect of UCS15A was compared with that of herbimycin and radicicol (UCS1006) on the Oncogene

morphology of v-src 3T3 cells (Figure 5). Untreated vsrc 3T3 cells exhibited a very characteristic, refractile morphology that was consistent with their transformed nature (Figure 5a). Treatment of these cells with either UCS1006 (Figure 5b) or herbimycin (Figure 5d) resulted in a very characteristic reversal of the transformed phenotype. The cells were much ¯atter and more adherent than in the untreated controls. On the other hand, v-src 3T3 cells treated with UCS15A appeared even more refractile than the untreated control cells with the almost complete disappearance of ®lopodia (Figure 5c). Despite this extreme morphology change, UCS15A treated cells were attached to the substratum and did not undergo apoptosis (data not shown). Furthermore, these morphological changes were reversible upon removal of the drug (data not shown). Although the basis for this morphological change is at present unknown, these data clearly indicated that UCS15A was distinct in its mode of action from conventional src destabilizing agents such as radicicol, herbimycin and geldanamycin.

UCS15A inhibits osteoclastic bone resorption SV Sharma et al

Figure 4 E€ect of UCS15A and PP2A on the tyrosine kinase activity of v-src. (a) To assess the e€ect of UCS15A on the in vitro tyrosine kinase activity of v-src, anti src immunoprecipitates from BRK v-src cells were subjected to in vitro kinase assays using a cdc2 peptide as substrate (see Materials and methods) in the presence of either UCS15A (*) or PP2 (&), at the concentrations indicated along the x-axis. Points shown are averages of data obtained in three independent experiments and the error bars represent the standard error of the mean. (b) To examine the e€ect of UCS15A on the in vivo activation state of v-src, v-src 3T3 cells were treated for 15 h with the concentrations of drug indicated at the top of each lane. Following the treatment, cells were lysed and 20 mg of total protein was analysed by SDS ± PAGE on 8.5% gels. Proteins were transferred to nitrocellulose membranes and immunoblotted with a phosphospeci®c anti-src antibody (anti-Src [pY416]). Position of migration of phosphorylated Src is indicated to the right of the antoradiogram. Approximate molecular weights of standards are indicated in kilodaltons to the left of the autoradiogram

UCS15A disrupts the association of v-Src with Sam68 and another unidentified protein The ®nding that UCS15A potently inhibited the tyrosine phosphorylation of src substrates (Figures 2 and 3), but failed to destabilize src (Figure 3a) or inhibit src-kinase activity (Figure 4), suggested the possibility that the drug may exert its e€ects by some other means. To clarify the mechanism of action of the

drug, the e€ects of UCS15A on src protein-protein complexes was examined in v-src 3T3 cells (Figure 6). Src protein was immunoprecipitated from UCS15A treated v-src 3T3 cells and the phosphotyrosine content of src and its associated proteins was examined by immunoblot analyses with anti-phosphotyrosine antibody (Figure 6a). Under the conditions used, only two prominent tyrosine phosphorylated proteins were found associated with src in untreated cells (Figure 6a, lane 1) with molecular weights corresponding to 64 and 68 kDa. This was somewhat surprising since src has been reported to associate with a large number of tyrosine phosphorylated proteins in v-src transformed cells. While the reason for this discrepancy is unclear, it may relate to the speci®c extraction conditions used in the present study (see Materials and methods). Indeed, other faint bands corresponding to higher molecular weight proteins were also evident but these were very minor (Figure 6a, lane 1). Upon UCS15A treatment, the tyrosine phosphorylation of both the 64 (pp64) and 68 kDa proteins diminished in a dose-dependent manner (Figure 6a, lanes 2 ± 6), albeit with di€erent eciencies. The tyrosine phosphorylation of pp64 was almost completely abolished by nanomolar concentrations of UCS15A (400 nM), while the 68 kDa protein required micromolar concentrations of the drug. Under similar conditions, there was little or no change in the tyrosine phosphorylation of src itself (Figure 6a). The 68 kDa protein was identi®ed as Sam68 by stripping the ®lter from 6a and re-probing with an anti-Sam68 antibody (Figure 6b). Such an analysis revealed that the decrease in tyrosine phosphorylation of Sam68, upon UCS15A treatment (Figure 6a), mirrored the decrease in src-associated Sam68 (Figure 6b). Taken together, these results suggested that the UCS15Amediated inhibition of tyrosine phosphorylation of Sam68 might be due to the ability of the drug to disrupt the src-Sam68 complex. The identity of pp64 is currently unknown, but it is possible that the inhibition of its tyrosine phosphorylation, as well as the tyrosine phosphorylation of other src-substrates, in response to UCS15A, may occur by disrupting protein-protein interactions, in a manner analogous to that which occurred with Sam68.

2073

Effects of UCS15A on bone resorption activity of MNCs Given the ability of UCS15A to inhibit the tyrosine phosphorylation of potential src substrates and/or disrupt src complexes in src-transformed cells, we investigated its activities in a biologically relevant system. To this end, we examined the e€ects of UCS15A on the bone resorption activity of osteoclast-like multinucleated cells (MNCs). The choice of this system was predicated on the fact that a large body of work has implicated src in the bone resorption activity of these cells (reviewed in Schwartzberg, 1998; Roodman, 1999). In addition, previous studies have shown that inhibition of src-kinase inhibited osteoclastmediated bone resorption (Yoneda et al., 1993; Hall et al., 1994; Blair et al., 1996; Missbach et al., 1999; Oncogene

UCS15A inhibits osteoclastic bone resorption SV Sharma et al

2074

Figure 5 E€ect of UCS15A, radicicol and herbimycin on the morphology of v-src 3T3 cells. V-src 3T3 cells were treated for 20 h with the indicated amounts of the di€erent drugs, as shown at the bottom of each panel. (a) untreated; (b) radicicol treated; (c) UCS15A treated; and (d) herbimycin treated. Cells were photographed under the microscope at a magni®cation of 2006

Figure 6 E€ect of UCS15A on the association of proteins with src. V-src 3T3 cells were treated for 24 h with various concentrations of UCS15A indicated at the top of each lane. Anti-src immunoprecipitates (IP) from UCS15A-treated cells were probed (WB) with either an anti-phosphotyrosine antibody (a) or with anti-sam68 antibody (b). Positions of migration of sam68, pp64 and src are indicated in the left-hand margin of each panel. Approximate molecular weights of standards are indicated in kilodaltons on the right-hand margin of each panel Oncogene

UCS15A inhibits osteoclastic bone resorption SV Sharma et al

Violette et al., 1999; Miyazaki et al., 2000). To assess the potential use of UCS15A as an anti-osteoporosis drug, the e€ects of UCS15A on the bone resorption activity of MNCs was examined (Figure 7). UCS15A inhibited bone resorption in a dose-dependent manner with an IC50 of 3 mM and almost completely inhibited bone resorption at 30 mM (Figure 7). Cytotoxicity of UCS15A was also assessed against primary mouse osteoblastic cells. Osteoblasts were used in the cytotoxicity assays, instead of MNCs, due to the diculty of obtaining pure populations of di€erentiated MNCs (see Materials and methods). As shown in Figure 7, UCS15A, at concentrations up to 30 mM, had minimal cytotoxic e€ects on osteoblastic cells. These data suggested that the inhibitory e€ect of UCS15A on bone resorption was probably not due to cytotoxicity. Previous studies showed that herbimycin A, a well-known src inhibitor, inhibited bone resorption in this assay with an IC50 of 0.6 mM but was cytotoxic with an IC50 of 3 mM (Sugawara et al., 1998). Similarly, radicicol, another src inhibitor, inhibited bone resorption with an IC50 of 1 mM but was cytotoxic with an IC50 of 10 mM. In contrast to these src inhibitors, UCS15A had minimal cytotoxicity over a wide range of concentrations at which bone resorption was signi®cantly inhibited. Effects of UCS15A on bone resorption in mouse calvaria organ culture systems Given the e€ects of UCS15A on the bone resorption activity of MNCs in vitro (Figure 7), the e€ect of the

compound was evaluated in a bone organ culture system. The mouse calvaria organ culture system, which is a closer approximation to the in vivo situation than isolated MNCs, has been a useful bioassay in the identi®cation of new agents that modulate bone resorption (reviewed in Roodman, 1999). UCS15A e€ectively inhibited the human parathyroid hormone peptide (PTH residues 1-34)-stimulated bone resorption activity of isolated mouse calvaria (Table 1). Stimulation of bone resorption activity by PTH was con®rmed by measuring the calcium concentration in the culture medium, which was signi®cantly higher in the PTHstimulated samples than in the control. The inhibitory e€ect of UCS15A on bone resorption in the organ culture system was dose-dependent, with an IC50 of 6.6 mM, which was similar to the IC50 determined using MNCs (3 mM). Taken together, these results suggested the potential usefulness of UCS15A as a therapeutic agent in the treatment of osteoporosis.

2075

Discussion The present studies have led to the following conclusions: ®rst, UCS15A is a unique src inhibitor with a potentially novel mode of action that does not involve src tyrosine kinase inhibition per se, or destabilization of the src protein itself. Second, UCS15A appears to inhibit the tyrosine phosphorylation of several src substrates in vivo, including cortactin and sam68. Third, UCS15A appears to exert its inhibitory e€ects on src by destabilizing the interactions of src with its associated substrates. Fourth, UCS15A inhibits the bone resorption activity of osteoclast-like multinucleated cells both in vitro and in organ culture systems. Taken together, these results point to the unique mode of action of UCS15A and suggest the potential usefulness of this agent in the treatment of osteoporosis. Why the yeast-based assay system used is biased against catalytic src-kinase inhibitors is presently unclear. It is possible that broadly speci®c kinase inhibitors such as genistein and staurosporin, inhibit one or more essential pathways and thus are toxic to Table 1 E€ects of UCS15A on bone resorption in mouse calvariae

Figure 7 E€ect of UCS15A on bone resorption activity of MNCs. MNCs and osteoblastic cells were treated with the concentrations of UCS15A indicated along the x-axis. The percentages of bone resorption (*) or cell viability (*) are indicated along the left- and right-hand sides of the y-axis, respectively. Points shown are averages of data obtained in three independent experiments and the error bars represent the standard error of the mean

Treatment

Calcium concentration (mg/dl)

Inhibition of bone resorption (%)

Control (vehicle) PTH 10 nM +UCS15A 0.1 mM 0.3 mM 1 mM 3 mM 10 mM 30 mM

9.05+0.11 13.10+0.14a 13.00+0.15 13.20+0.16 12.90+0.10 12.60+0.26b 11.00+0.11c 8.83+0.06c

± ± 2+7 2+5 5+3 13+7 52+5 106+4

a

P50.01 vs control; bP50.05 vs PTH 10 nM; cP50.01 vs PTH 10 nM. Half-calvaria from 5-day-old newborn mice were treated with 10 nM of human parathyroid hormone, 1 ± 34 fragment (PTH), and cultured at 378C for 96 h, in the presence or absence of UCS15A Oncogene

UCS15A inhibits osteoclastic bone resorption SV Sharma et al

2076

Oncogene

yeast cells. Such compounds would not be identi®ed in this screen that relies on the restoration of the growth of arrested yeast cells. Catalytic tyrosine-kinase inhibitors that exclusively inhibit src tyrosine kinase may not be present in the compound library, since none were identi®ed in our screen. Therefore, only agents that otherwise a€ect the src-signaling pathway, are identi®ed by this system. In addition, such compounds are not detrimental to growth because the assay relies on their ability to restore growth of arrested yeast cells. The ability of UCS15A to inhibit the tyrosine phosphorylation of src substrates without directly a€ecting the tyrosine kinase activity of src appears, on the surface, to be paradoxical. However, these results can be reconciled by the fact that UCS15A appears to disrupt src-mediated protein-protein interactions. It is intuitively obvious that before a substrate can be phosphorylated it must make contact with the kinase. Protein-protein interactions and their roles in the diversity of src functions are still poorly understood. The protein binding modules of src, namely its SH2 and SH3 domains have been implicated in the association of src with a plethora of substrates (reviewed in Brown and Cooper, 1996). These modules presumably bind to the substrates, resulting in an opening up of the closed (inactive) src molecule, and place the substrates in an optimal con®guration for action by the activated kinase domain (Yamaguchi and Hendrickson, 1996; Sicheri et al., 1997; Xu et al., 1997; reviewed in Schwartzberg, 1998). Thus, despite the fact that UCS15A fails to inhibit the tyrosine kinase activity of src in vitro (Figure 3) its ability to disrupt the association of proteins with src is consistent with the prevention of tyrosine phosphorylation of src substrates (Figures 1 and 2). A similar inhibition of tyrosine phosphorylation of p58gag by src was reported by peptides corresponding to the SH3 domain of src (Huang et al., 1999). Using peptide-library scans, it was determined that the consensus sequence for src phosphorylation is similar to the consensus sequence for SH2 binding, suggesting that some targets of src kinase physically bind to the SH2 domain of src (Songyang et al., 1994). Similarly, other studies have shown that the src SH3 domain is also important for the tyrosine phosphorylation of src substrates (Weng et al., 1994; reviewed in Dalgarno et al., 1998). Consistent with these results, over-expressed SH2 and SH3 domains of src resulted in dominant-negative e€ects on src signaling (Twamley-Stein et al., 1993; Broome and Hunter, 1996; Erpel et al., 1996). Morphological changes induced by UCS15A are unique and distinct from other known src destabilizing agents such as herbimycin, radicicol (Figure 5) and geldanamycin (Whitesell et al., 1994). In addition, previous studies have demonstrated that src-kinase inhibitors, such as PP2, also induced morphological reversion of src-transformed cells (Violette et al., 1999), which is not the case with UCS15A (Figure 5). UCS15A treated, src-transformed cells appear to be less adherent and less well-spread out than their untreated counterparts (Figure 5). This appearance is reminiscent of the

cell-spreading defects in src7/7 ®broblasts that could be rescued by the introduction of kinase de®cient src molecules (Kaplan et al., 1995). Taken together with other studies, our results suggest that UCS15A treatment may interfere with the non-kinase, adapter function of src that serves to recruit signaling complexes in response to integrin engagement (Schaller et al., 1994; Schlaepfer et al., 1997; reviewed in Schwartzberg, 1998). In addition, given the fact that src activity is intimately linked to the dynamics of the cytoskeleton (reviewed in Brown and Cooper, 1996), it is not surprising that a src inhibitor like UCS15A would have profound e€ects on cyto-architecture and cell morphology (Figure 5). In this regard, it is interesting that one of the proteins, whose tyrosine phosphorylation is altered by UCS15A, is the actin-binding protein, cortactin (Figures 2 and 3). Previous reports have demonstrated that the srcsubstrate cortactin, which is localized to the cell periphery in normal cells, becomes hyperphosphorylated on tyrosine and concentrated in podosomes in srctransformed cells (Wu et al., 1991). The reduced tyrosine-phosphorylation of cortactin, upon UCS15A treatment (Figures 2 and 3), correlates with the disappearance of ®lopodia, and possibly podosomes, in src-transformed cells (Figure 5). However, the fact that this decreased phosphorylation of cortactin does not result in the restoration of stress-®bers (Figure 5) suggests that the relationship between cortactin, cell morphology and UCS15A is complex and warrants further investigation. The extreme rounding of srctransformed cells, upon UCS15A treatment, was reversible and was not due to the induction of apoptosis, as assayed by the caspase-mediated cleavage of Poly ADP Ribose Polymerase (data not shown). In addition to their roles in src signaling, proteinprotein interactions also play a key role in the intracellular localization of src. Activation of the PDGF receptor leads to the membrane localization of src, whereas src localizes to focal adhesions following integrin engagement (reviewed in Schwartzberg, 1998). Localization of src to the membrane cytoskeleton, especially at focal adhesions, has been shown to be kinase-independent, but shows an absolute requirement for the SH3 domain of src (Kaplan et al., 1994). However, other studies point to the importance of the SH2 and catalytic domains, and not the SH3 domain, in these interactions (Fukui et al., 1991; Okamura and Resh, 1994). In either case, it is clear that proteinprotein interactions play important roles in the subcellular localization of src and its involvement in signaling complexes. In our experiments, the most prominent tyrosine phosphorylated protein associated with src and whose association is disrupted by UCS15A treatment, is Sam68 (Figure 6). Previous studies have suggested that the increase in src kinase activity, speci®cally in mitosis (Chackalaparampil and Shalloway, 1988), is accompanied by its association with Sam68 in this phase of the cell cycle (Fumagalli et al., 1994; Taylor and Shalloway, 1994). These ®ndings suggest the possibility that src and its mitotic substrate Sam68 may play a

UCS15A inhibits osteoclastic bone resorption SV Sharma et al

role in mediating certain mitotic events. Consistent with these observations, UCS15A is able to arrest cells in mitosis (Nakai et al., 2001, in preparation). The ability of UCS15A to inhibit the bone resorption of MNCs and mouse calvariae with minimal cytotoxicity (Figure 6 and Table 1) suggests a potential therapeutic use for this compound. Our studies indicate that UCS15A does not directly inhibit the kinase activity of src (Figure 4) neither does it in¯uence the stability of the src protein (Figure 3a). In this regard it is di€erent from most currently known src antagonists. The mode of action of UCS15A most closely resembles the compound AP22161, which selectively binds to the SH2 domain of src and inhibits osteoclast mediated bone resorption (Violette et al., 1999). Using mammalian-cell based two-hybrid assays, AP22161 was shown to inhibit the binding of peptide ligands to the src SH2 domain (Violette et al., 1999). However, while AP22161 causes morphological reversion of rat ®broblasts transformed by activated src (Violette et al., 1999), UCS15A did not cause morphological reversion of src-transformed NIH3T3 cells (Figure 5). Thus, while both AP22161 and UCS15A may inhibit src activity by disrupting protein-protein interactions, the speci®city of the two agents may be di€erent. The observation that UCS15A disrupts the association between src and sam68 (Figure 6b) may be important in this regard. Previous studies have shown that the association between src and sam68 is primarily mediated by the SH3 domain of src (Shen et al., 1999). Therefore, it is possible that UCS15A speci®cally disrupts SH3 mediated interactions of src with its substrates, unlike AP22161, which disrupts src SH2 interactions. We are currently investigating this possibility in an e€ort to better understand the mechanism of action of UCS15A in the disruption of src-mediated protein-protein interactions.

Materials and methods Reagents, cell-lines and animals UCS15A was produced by Streptomyces sp. and puri®ed in our laboratories. 1,25-dihydroxyvitamin D3 was obtained from Wako Pure Chemicals Ind. Ltd., (Osaka, Japan) and parathyroid hormone (PTH, human synthetic 1-34 fragment) was purchased from Sigma (St. Louis, MO, USA). Dulbecco's modi®cation of Eagle's medium (DMEM) was obtained from Nissui Pharmaceutical Co. Ltd. (Tokyo, Japan). Anti phosphotyrosine antibody was purchased as an HRP-conjugate (ICN), anti-cortactin antibody (mouse monoclonal; UBI), anti-Sam68 (rabbit polyclonal; SantaCruz Biotechnology), Anti-Src (mouse monoclonal GD11; UBI), Anti-src [pY418] (rabbit polyclonal; Biosource International). Secondary antibodies consisted of either anti-mouse or anti-rabbit antibodies conjugated to HRP (Amersham). Immunoprecipitates were collected on Protein A/G beads (Santa Cruz Biotechnology). Src kinase activity was estimated using a commercially available kit (UBI). Cell-lines used were Src-transformed NIH3T3 cells, v-src 3T3, and adenovirus E1A immortalized Baby Rat Kidney cells transformed by the v-src oncogene, BRK v-src (kindly provided by Dr Margaret

Quinlan). Their culture conditions were as described previously (Fischer and Quinlan, 2000). Newborn mice (ddY) and 8-week old male mice (ddY) were purchased from SLC (Shizuoka, Japan).

2077

Screening for Src inhibitors Saccharomyces cerevisiae, RAY 3A-D (ura3, his3, trp1, leu2) (kindly provided by Dr Kunihiro Matsumoto, Nagoya Univ.), carrying the v-src gene controlled by the Gal-1 promoter was grown to stationary phase in glucose medium. These mycelia were then transferred to agar plates containing galactose. In the presence of galactose, v-src was overexpressed and, as described previously, this resulted in the growth-arrest of the cells (Brugge et al., 1987; Kornbluth et al., 1987). Paper disks, soaked in the samples to be tested, were placed on the galactose agar plates, which were then incubated at 308C for 2 days. Active compounds (src inhibitors) were identi®ed as those that could relieve the growth-inhibition imposed by the over-expressed v-src, thus allowing a halo of growth around the paper disks. A natural product library consisting of about 20 000 natural compounds was screened using this system. Immunoblotting and immunoprecipitation analyses The e€ect of UCS15A on the tyrosine phosphorylation of proteins in v-src 3T3 cells was examined as follows. Brie¯y, a sub-con¯uent 60 mm dish of cells was trypsinized and plated into a six well multi-plate. After the cells had adhered and spread on the dish (approximately 8 h post-plating), each well was treated with a di€erent concentration of UCS15A (0, 0.4, 0.8, 2, 4 and 8 mM respectively) for 24 h. At the end of the treatment period, cells from each well were scraped and collected in their own medium. Cell pellets were lysed by the addition of 75 ml of ice-cold RIPA lysis bu€er (50 mM TRIS/ HCl pH 8.0; 2 mM EDTA; 150 mM NaCl; 0.1% SDS; 1% sodium deoxycholate; 1% Triton X-100; 10 mg/ml aprotinin; 10 mg/ml leupeptin; 1 mM sodium orthovanadate; 1 mM PMSF). Lysates were clari®ed by microcentrifugation (15 000 r.p.m. for 30 min at 48C). A third of each clari®ed lysate (25 ml) was mixed with an appropriate volume of 46Laemmli's sample bu€er, boiled for 10 min and proteins were separated on a 7.5% polyacrylamide gel. Proteins were transferred to 0.45 mm nitrocellulose membranes (Protran; Schleicher and Schuell) and blocked overnight at 48C in blocking bu€er (PBS; 0.25% gelatin; 0.2% Tween 20). Membranes were blotted with HRP-conjugated anti-phosphotyrosine antibody and subjected to chemiluminescent detection. When the activation state of src in vivo was examined, the nitrocellulose ®lter was blocked in TBS-T bu€er (Trisbu€ered saline containing 0.1% Tween) containing 3% bovine serum albumin (TBS-TB) for 1 h followed by incubation with the primary antibody (rabbit polyclonal anti-src [pY418] at a 1 : 1000 dilution) in TBS-TB for 4 h. After brie¯y washing with TBS-T, the ®lter was incubated with the secondary antibody (HRP-conjugated Goat antiRabbit IgG at a 1 : 4000 dilution) in TBS-TB for 1 h followed by chemiluminescent detection. To compare the pro®le of tyrosine phosphorylated proteins in src-transformed and untransformed NIH3T3 cells, cells were used 2 days after plating. Cells were lysed as described above and the protein content of each lysate was estimated (Bio-Rad DC Protein Assay kit). Di€erent amounts of the cell lysates were separated by polyacrylamide gel electrophoresis Oncogene

UCS15A inhibits osteoclastic bone resorption SV Sharma et al

2078

and analysed by immunoblotting with the HRP-conjugated anti-phosphotyrosine antibody as described above. For immunoprecipitation analyses, 2 day-old, sub-con¯uent cells (26100 mm dishes per treatment) were treated with 0, 0.4, 0.8, 2, 4 and 8 mM of UCS15A for 24 h. Cells were collected as described above and lysed by the addition of 1 ml of ice-cold RIPA lysis bu€er. Clari®ed lysates were immunoprecipitated overnight with the indicated antibodies (3 mg of anti-cortactin antibody or 1 mg of anti-Sam68 antibody). Immunoprecipitates were collected on Protein A/ G beads (30 ml per immunoprecipitation). Beads were washed three times in Triton/NP-40 lysis bu€er, re-suspended in 30 ml of Laemmli's sample bu€er and resolved on 7.5% SDSpolyacrylamide gels. Proteins were transferred, immunoblotted and subjected to chemiluminescent detection as described above. In some cases, the blots were stripped by incubation in stripping bu€er (62.5 mM Tris-HCl, pH 6.7, 2% SDS, 100 mM 2-mercaptoethanol) at 608C for 1 h. After thoroughly washing, the blots were re-probed with an appropriate primary and secondary antibody followed by chemiluminescent detection. To examine proteins associated with src, anti-src immunoprecipitations were performed as described above, except that cells were lysed in 1 ml of ice-cold Triton/NP-40 lysis bu€er (50 mM Tris/HCl pH 7.4, 150 mM NaCl, 1 mM EDTA, 1% Triton X-100, 0.5% NP-40, 10 mg/ml aprotinin, 10 mg/ml leupeptin, 1 mM sodium orthovanadate, 1 mM PMSF).

spotted on numbered P81 paper squares, which were then washed ®ve times (5 min each) with 0.75% phosphoric acid and once with acetone. Radioactivity associated with the ®lter squares was determined by liquid scintillation counting. Bone resorption assays

V-Src 3T3 cells were plated in six-well multi-plates as described above. Approximately 8 h after plating, cells were either left untreated or treated with di€erent amounts of radicicol (UCS1006), herbimycin or UCS15A for 20 h. At the end of the treatment period, cell morphology was visualized by light microscopy (2006 magni®cation) and recorded by photomicroscopy.

Osteoclast-like multinucleated cells (MNCs) were prepared in vitro and bone resorption activity of MNCs was measured as reported previously (Akatsu et al., 1992; Sugawara et al., 1998). Brie¯y, to induce osteoclast-like MNCs, osteoblastic cells obtained from newborn-mouse calvaria and bone marrow cells obtained from 8-week old mice were co-cultured in the presence of 1,25-dihydroxyvitamin D3 (10 nM) and dexamethasone (100 nM) at 378C for 6 days in a humidi®ed atmosphere of 5% CO2. Induced MNCs were treated with enzyme solution (0.1% collagenase and 0.2% dispase in PBS), and the MNCs were then placed on dentine slices. Bone resorption by MNCs was performed for 48 h with or without UCS15A, and calcium concentration in the culture medium was measured as a function of bone resorption activity, using a commercial test kit (Calcium-C Test, Wako Pure Chemicals, Osaka, Japan). Organ culture was performed as reported previously (Hall and Alexander, 1987). Half-calvaria from 5-day-old newborn mice were transferred to six-well dish culture plates containing 1.5 ml of DMEM. Bone resorption was stimulated by adding 10 nM of human PTH (1 ± 34 fragment). UCS15A was dissolved in 100% ethanol and 0.1% volume of culture medium was added to culture plates. Calvaria were cultured at 378C for 96 h in a humidi®ed atmosphere of 5% CO2. The culture medium was changed at 48 h. Calcium concentration in the culture medium was measured. Each series of experiments was repeated at least three times. The results of a typical experiment were expressed as the mean+ standard deviation (s.d.). Signi®cance determination was evaluated using Student's t-test or Dunnett's multiple test.

Src kinase assays

Cytotoxicity determination

Src kinase assays using cdc2 peptide as a phospho-acceptor were performed as described previously (Zhu et al., 1998), using a commercially available src-kinase assay kit (UBI). Subcon¯uent, (2-days old) BRK v-src cells from a 100 mm dish were lysed in Triton/NP-40 lysis bu€er and immunoprecipitated with anti-src antibody. Immunoprecipitates were collected on Protein A/G beads, washed twice in Triton/NP40 lysis bu€er, three times in kinase assay bu€er and resuspended in kinase assay bu€er containing di€erent concentrations of UCS15A or PP1 (positive control). To the reaction was added the cdc2 peptide, [g-32P]ATP in Mg/ ATP cocktail and incubated at 308C for 10 min. The reactions were stopped by the addition of 20 ml of 40% trichloroacetic acid. Twenty-®ve ml of the supernatant was

Cytotoxicity of UCS15A was evaluaetd by MTT assay (Mosmann, 1983; Sugawara et al., 1998) using mouse osteoblastic cells. Because it was dicult to isolate MNCs for cytotoxicity determination, osteoblastic cells were used instead. Osteoblastic cells were cultured with UCS15A for 48 h, identical to the bone resorption assay period.

Photomicroscopy

Acknowledgments We thank Dr Margaret Quinlan for the critical reading of this manuscript and for her helpful discussion. The technical assistance of Miss H Tsunoda is gratefully acknowledged.

References Akatsu T, Tamura T, Takahashi N, Udagawa N, Tanaka S, Sasaki T, Yamaguchi A, Nagata N and Suda T. (1992). J. Bone Miner. Res., 7, 1297 ± 1306. Blair HC, Jordan SE, Peterson TG and Barnes S. (1996). J. Cell. Biochem., 61, 629 ± 637. Bolen JB, Veillette A, Schwartz AM, DeSeau V and Rosen N. (1987). Proc. Natl. Acad. Sci. USA, 84, 2251 ± 2255. Boyce B, Yoneda T, Lowe C, Soriano P and Mundy G. (1992). J. Clin. Invest., 90, 1622 ± 1627. Oncogene

Broome MA and Hunter T. (1996). J. Biol. Chem., 271, 16798 ± 16806. Brown MT and Cooper JA. (1996). Biochim. Biophys. Acta., 1287, 121 ± 149. Brugge JS, Jarosik G, Andersen J, Queral-Lustig A, FedorChaiken M and Broach JR. (1987). Mol. Cell. Biol., 7, 2180 ± 2187. Chackalaparampil I and Shalloway D. (1988). Cell, 52, 801 ± 810.

UCS15A inhibits osteoclastic bone resorption SV Sharma et al

Cheng H-C, Nishio H, Hatase O, Ralph S and Wang JH. (1992). J. Biol. Chem., 267, 9248 ± 9256. Dalgarno DC, Bot®eld MC and Rickles RJ. (1998). Biopolymers., 43, 383 ± 400. Erpel T, Alonso G, Roche S and Courtneidge SA. (1996). J. Biol. Chem., 271, 16807 ± 16812. Fischer RS and Quinlan MP. (2000). Virol., 269, 404 ± 419. Fukazawa H, Li P-M, Yamamoto C, Murakami Y, Mizuno S and Uehara Y. (1991). Biochem. Pharmacol., 42, 1661 ± 1671. Fukui Y, O'Brien MC and Hanafusa H. (1991). Mol. Cell. Biol., 11, 1207 ± 1213. Fumagalli S, Totty NF, Hsuan JJ and Courtneidge S. (1994). Nature, 368, 871 ± 874. Hall GE and Alexander DK. (1987). Calcif. Tiss. Int., 40, 212 ± 218. Hall TJ, Schaeublin M and Missbach M. (1994). Biochem. Biophys. Res. Commun., 199, 1237 ± 1244. Hanke JH, Gardner JP, Dow RL, Changelian PS, Brisette WH, Weringer EJ, Pollok BA and Connelley PA. (1996). J. Biol. Chem., 271, 695 ± 701. Huang J, Zhang B-T, Li Y, Mayer B, Carraway KL and Carraway CAC. (1999). Oncogene, 18, 4099 ± 4107. Hunter T and Sefton BM. (1980). Proc. Natl. Acad. Sci. USA, 77, 1311 ± 1315. Jacobs C and Rubsamen H. (1983). Cancer Res., 43, 1696 ± 1702. Kamps MP and Sefton BM. (1988). Oncogene, 2, 305 ± 315. Kanner SB, Reynolds AB, Vines RR and Parsons JT. (1990). Proc. Natl. Acad. Sci. USA, 87, 3328 ± 3332. Kaplan KB, Bibbins KB, Swedlow JR, Arnaud M, Morgan DO and Varmus H. (1994). EMBO J., 13, 4745 ± 4756. Kaplan KB, Swedlow JR, Morgan DO and Varmus HE. (1995). Genes Dev., 9, 1505 ± 1517. Kay BK, Williamson MP and Sudol M. (2000). FASEB J., 14, 231 ± 241. Kornbluth S, Jove R and Hanafusa H. (1987). Proc. Natl. Acad. Sci. USA, 84, 4455 ± 4459. Kurihara N, Civin CL and Roodman GD. (1990). Endocrinol., 126, 2733 ± 2741. Kwon HJ, Yoshida M, Abe K, Horinouchi S and Beppu T. (1992b). Biosci. Biotech. Biochem., 56, 538 ± 539. Kwon HJ, Yoshida M, Fukui Y, Horinouchi S and Beppu T. (1992a). Cancer Res., 52, 6926 ± 6930. Kwon HJ, Yoshida M, Muroya K, Hattori S, Nishida E, Fukui Y, Beppu T and Horinouchi S. (1995). J. Biochem., 118, 221 ± 228. Lowe C, Yoneda T, Boyce BF, Chen H, Mundy GR and Soriano P. (1993). Proc. Natl. Acad. Sci. USA, 90, 4485 ± 4489. Luttrell LM, Daaka Y and Lefkowitz RJ. (1999). Curr. Opin. Cell Biol., 11, 177 ± 183. Manolagas SC. (2000). Endocr. Rev., 21, 115 ± 137. Missbach M, Jeschke M, Feyen J, Muller K, Glatt M and Susa M. (1999). Bone, 24, 437 ± 439. Miyazaki T, Takayanagi H, Isshiki M, Takahashi T, Okada M, Fukui Y, Oda H, Nakamura K, Hirai H, Kurokawa T and Tanaka S. (2000). J. Bone Miner. Res., 15, 41 ± 51. Mosmann T. (1983). J. Immunol. Meth., 65, 55 ± 63. Okamura H and Resh MD. (1994). Oncogene, 9, 2293 ± 2303. Roe MS, Prodromou C, O'Brien R, Ladbury JE, Piper PW and Pearl LH. (1999). J. Med. Chem., 42, 260 ± 266. Roodman GD. (1999). Exp. Hematol., 27, 1229 ± 1241. Schaller MD, Hildebrand JD, Shannon JD, Fox JW, Vines RR and Parsons JT. (1994). Mol. Cell. Biol., 14, 1680 ± 1688.

Schlaepfer DD, Broome MA and Hunter T. (1997). Mol. Cell. Biol., 17, 1702 ± 1713. Schulte TW, Akinaga S, Agatsuma T, Murakata T, Sugimoto S, Nakano H, Simen BB, Argon Y, Toft DO, Neckers LM and Sharma SV. (1999). Mol. Endocrinol., 13, 1435 ± 1448. Schulte TW, Akinaga S, Soga S, Sullivan W, Stensgard B, Toft D and Neckers LM. (1998). Cell Stress and Chaperones, 3, 100 ± 108. Schwartzberg PL. (1998). Oncogene, 17, 1463 ± 1468. Schwartzberg PL, Xing L, Ho€mann O, Lowell CA, Garrett L, Boyce BF and Varmus HE. (1997). Genes Dev., 11, 2835 ± 2844. Sharma S, Agatsuma T and Nakano H. (1998). Oncogene, 16, 2639 ± 2645. Shen Z, Batzer A, Koehler JA, Polakis PP, Schlessinger J, Lydon NB and Moran MF. (1999). Oncogene, 18, 4647 ± 4653. Sicheri F, Moare® I and Kuriyan J. (1997). Nature, 385, 602 ± 609. Songyang Z, Blechner S, Hoagland N, Hoekstra MF, Piwnica-Worms H and Cantley LC. (1994). Curr. Biol., 4, 973 ± 982. Soriano P, Montgomery C, Geske R and Bradley A. (1991). Cell, 64, 693 ± 702. Stebbins CE, Russo AA, Schneider C, Rosen N, Hartl FU and Pavletich NP. (1997). Cell, 89, 239 ± 250. Stehlin D, Varmus HE, Bishop JM and Vogt PK. (1976). Nature, 260, 170 ± 173. Sugawara K, Hamada M, Hosoi S and Tamaoki T. (1998). Anal. Biochem., 255, 204 ± 210. Suzuki G, Kobayashi I and Mitsutake. (1983). (Patent # JP58116686). Japan Kokai Tokkyo Koho. Taylor SJ and Shalloway D. (1994). Nature, 368, 867 ± 871. Thomas SM and Brugge JS. (1997). Annu. Rev. Cell. Dev. Biol., 13, 513 ± 609. Twamley-Stein GM, Pepperkok R, Ansorge W and Courtneidge SA. (1993). Proc. Natl. Acad. Sci. USA, 90, 7697 ± 7700. Uehara Y, Hori M, Takeuchi T and Umezawa H. (1985). Jpn. J. Cancer Res., 76, 672 ± 675. Uehara Y, Murakami Y, Suzukake-Tsuchiya K, Moriya Y, Sano H, Shibata K and Omura S. (1988). J. Antibiot., 41, 831 ± 834. Violette SM, Shakespeare WC, Barlette C, Guan W, Smith JA, Rickles RJ, Bohacek RS, Holt DA, Baron R and Sawyer TK. (1999). Chem. Biol., 7, 225 ± 235. Weng Z, Thomas SM, Rickles RJ, Taylor JA, Brauer AW, Seidel-Dugan C, Michael WM, Dreyfuss G and Brugge JS. (1994). Mol. Cell. Biol., 14, 4509 ± 4521. Whitesell L, Mimnaugh EG, De CB, Myers CE and Neckers LM. (1994). Proc. Natl. Acad. Sci. USA, 91, 8324 ± 8328. Wu H, Reynolds AB, Kanner SB, Vines RR and Parsons JT. (1991). Mol. Cell. Biol., 11, 5113 ± 5124. Xu W, Harrison SC and Eck MJ. (1997). Nature, 385, 595 ± 602. Yamaguchi H and Hendrickson WA. (1996). Nature, 384, 484 ± 489. Yoneda T, Lowe C, Lee C-H, Gutierrez G, Niewolna M, Williams PJ, Izbicka E, Uehara Y and Mundy GR. (1993). J. Clin. Invest., 91, 2791 ± 2795. Zhao JF, Nakano H and Sharma S. (1995). Oncogene, 11, 161 ± 173. Zhu T, Goh ELK, LeRoith D and Lobie PE. (1998). J. Biol. Chem., 273, 33864 ± 33875.

2079

Oncogene