Targeting of the Protein Chaperone, HSP90, by the Transformation Suppressing Agent, Radicicol
Total Page:16
File Type:pdf, Size:1020Kb
Oncogene (1998) 16, 2639 ± 2645 1998 Stockton Press All rights reserved 0950 ± 9232/98 $12.00 http://www.stockton-press.co.uk/onc Targeting of the protein chaperone, HSP90, by the transformation suppressing agent, radicicol Sreenath V Sharma1, Tsutomu Agatsuma2 and Hirofumi Nakano2 1Department of Microbiology and Immunology, University of Tennessee, Memphis, 858 Madison Avenue, Memphis, Tennessee, USA; 2Kyowa Hakko Kogyo Co., Ltd., Tokyo Research Laboratories, 3-6-6 Asahimachi, Machidashi, Tokyo 194, Japan Radicicol, a macrocyclic anti-fungal antibiotic, has the oncogenic eects of Ras is well documented (Leevers et ability to suppress transformation by diverse oncogenes al., 1994; Moodie et al., 1993; Vanaelst et al., 1993; such as Src, Ras and Mos. Despite this useful property, Vojtek et al., 1993; Warne et al., 1993; reviewed in the mechanism by which radicicol exerts its anti- Marshall, 1995b), although other studies suggest that transformation eects is currently unknown. To under- additional, Raf-independent events are also important stand the transformation-suppressing eects of radicicol, in Ras-mediated transformation (Khosravi-Far et al., a biotinylated derivative of radicicol was chemically 1996; reviewed in Katz and McCormick, 1997). Besides synthesized and used as a probe in a Western-blot format Ras, other proteins that interact with Raf, and perhaps to visualize cellular proteins that interact with radicicol. regulate it, include Mitogen- or extracellular-regulated In transformed and untransformed mouse ®broblasts, the kinase (MEK1/2); proteins of the 14-3-3 family; most prominent cellular protein that bound to radicicol molecular chaperones such as HSP90 and Cdc37/p50; had a molecular weight of approximately 90 kDa. and the protein Kinase Suppressor of Ras (KSR) Further analysis revealed that this protein was the mouse (reviewed in Morrison and Cutler, 1997). Interaction of homologue of the 90 kDa heat shock protein (HSP90). HSP90 and Cdc37/p50 appears to be important for the This was con®rmed by demonstrating the ability of proper intracellular localization and stability of Raf radicicol to speci®cally bind puri®ed human HSP90. (Schulte et al., 1995, 1996; Stancato et al., 1997; Speci®city of binding was demonstrated by the inhibition reviewed in Hunter and Poon, 1997). of binding of biotinylated radicicol by the native drug. Members of the Ras family, as well as proteins that Taken together with other studies the present observa- regulate them, are particularly important with respect to tions suggest that the anti-transformation eects of human cancer, since they are frequently deregulated in radicicol may be mediated, at least in part, by the human tumors (reviewed in Hunter, 1997). Given the association of radicicol with HSP90 and the consequent central role of Ras oncogenes in human cancer, much dissociation of the Raf/HSP90 complex leading to the eort has been devoted to the identi®cation of agents attenuation of the Ras/MAP kinase signal transduction that are capable of suppressing its oncogenic capabilities pathway. (reviewed in Oli et al., 1996). One fruitful approach has utilized the detailed information available regarding Keywords: radicicol; molecular chaperones; transforma- the Ras oncoprotein and exploited the requirement for tion suppression; HSP90; Raf membrane association of Ras for its activity (reviewed in Gibbs et al., 1996; Koblan et al., 1996). Other approaches target the Ras signal transduction pathway as a means to control its activity (reviewed in Marshall, Introduction 1995a). An example of an agent that suppresses Ras- transformation by the latter approach, is radicicol The mitogen-activated protein (MAP) kinase pathway (Kwon et al., 1995; Soga et al., 1998; Zhao et al., serves as one of the principal conduits for transmission 1995). In addition to Ras-transformed cells, radicicol of signals generated by extracellular mitogenic growth also suppresses the transformed phenotype of cells factors (reviewed in Marshall, 1994; Robinson and expressing activated Src (Kwon et al., 1992b) and Mos Cobb, 1997). Oncogenes, or cancer causing genes, oncogenes (Zhao et al., 1995). In both Ras- and Mos- frequently subvert signal transduction pathways transformed cells radicicol treatment led to the general- utilized by growth factors (reviewed in Hunter, 1997). ized attenuation of the Ras-Raf-MAP kinase pathway In particular, oncogenes such as Raf, Mos and the in vivo (Zhao et al., 1995), but had no eects on the recently discovered oncogene Tpl2 (Salmeron et al., kinase activity of MAP kinase kinase (MAPKK) or 1996) have been implicated in directly hyper-activating MAP kinase in vitro (Kwon et al., 1995). This apparent the MAP kinase pathway and this seems to be, in part, paradox adds to the conundrum of the mechanism of responsible for their transforming capabilities (re- transformation suppression by radicicol. Attempts to viewed in Hunter, 1997; Marais and Marshall, 1996). unravel the mechanism of action of radicicol have The Ras oncogene also activates the MAP kinase revealed that radicicol treatment of Ras transformed pathway, albeit indirectly, via its down-stream eector, cells leads to a destabilization of the Raf-1 protein Raf (reviewed in Avruch et al., 1994; Morrison and (Soga et al., 1998) and consequently inhibition of Cutler, 1997). The importance of Raf in mediating the MAPK activity. Although this ®nding explains how radicicol can inhibit MAPK in vivo, but not in vitro,it still leaves unresolved the mechanism of Raf-1 destabilization by radicicol. Correspondence: SV Sharma The present studies demonstrate the ability of Received 23 October 1997; accepted 19 December 1997 radicicol to directly associate with murine and human Direct association of radicicol with HSP90 SV Sharma et al 2640 HSP90 in a highly selective and speci®c manner. These teins that radicicol binds speci®cally may point to key ®ndings suggest the possibility that radicicol, by intracellular targets, which may be valuable candidates binding to HSP90, may destabilize the Raf-1-HSP90 for chemotherapeutic intervention. With this goal in complex in a manner reminiscent of the benzoquinone mind, a biotinylated derivative of radicicol was ansamycin, geldanamycin (Schulte et al., 1996). In this synthesized. The structure of radicicol (shown in regard, radicicol is the ®rst non-ansamycin antibiotic to Figure 1b) contains an epoxide residue that is critical associate speci®cally with HSP90. for its antifungal properties (Ayer et al., 1980) and may play a role in its biological activity as a tyrosine kinase inhibitor (Kwon et al., 1992a). In addition, radicicol, like many other tyrosine kinase inhibitors, Results possesses phenolic ring structures that may be necessary for its function. With these constraints in Generation and use of a biotinylated derivative of mind, the biotin group was attached to radicicol, at radicicol to identify radicicol binding cellular proteins C9, via an oxime linkage, using a 12 carbon spacer The ability of radicicol to suppress transformation by arm (Figure 1a). Biotinylated radicicol (KT8529) was a diverse array of oncogenes suggested the possibility used in a Western-blot format, in which total cellular that intracellular targets of radicicol may lie at the proteins were immobilized on nitrocellulose mem- intersection of dierent pathways utilized by various branes and probed with KT8529, followed by oncogenes. Therefore, identi®cation of cellular pro- Streptavidin-Horse Radish Peroxidase (HRP). The a b c SRC-3T3 LYSATES (mI) 10 20 30 40 10 20 30 40 10 20 30 40 – – – – – – – – – – – – 115 115 89 89 52 52 – – – – – – – – – – – – – – 12345678910 11 12 13 14 UCS1006 KT8529 1¡-UCS1006 2¡-KT8529 Figure 1 Structure of the biotinylated derivative of radicicol, KT8529, and its use in the identi®cation of 90 kDa radicicol-binding protein(s). (a) Structure of KT8529, a biotinylated derivative of radicicol. (b) Structure of radicicol. (c) Dierent amounts of SRC- 3T3 cell lysates (indicated at the top of each lane), were electrophoresed on a 5% polyacrylamide gel, transferred to nitrocellulose membranes and probed with either 2.6 mg/ml of UCS1006 (lanes 1 ± 4), 1.3 mg/ml of KT8529 (lanes 6 ± 9) or ®rst with 2.6 mg/ml of UCS1006 followed by 1.3 mg/ml of KT8529 (lanes 11 ± 14) as indicated at the bottom of the autoradiograms. Biotinylated protein bands were detected by incubating the ®lters with streptavidin-HRP and visualized by chemiluminescence (for details see Experimental procedures). For reference pre-stained molecular weight markers (high range from BioRad) were used and their approximate positions are indicated on either side of the ®gure. They are in descending order of size: 115 000 daltons (b- galactosidase), 89 000 daltons (bovine serum albumin) and 52 000 daltons (ovalbumin). The position of the 90 kDa protein is indicated by an asterisk in lanes 5 and 10 Direct association of radicicol with HSP90 SV Sharma et al 2641 presence of proteins bound to radicicol were detected by chemiluminescence (Figure 1c). KT8529 was capable of binding a doublet of proteins of about I) I) 90 kDa in cell lysates prepared from SRC-3T3 cells m m (Figure 1c, lanes 6 ± 9; position demarcated by an ( ( g) g) m m asterisk). That these proteins were speci®cally ( ( SRC-3T3 LYSATES RAS-3T3 LYSATES recognized by radicicol was demonstrated by the fact huHSP90 huHSP90 that pre-treatment of the ®lters with native radicicol (UCS1006) eectively abolished the binding of KT8529 to these proteins (Figure 1c, lanes 11 ± 14). 551020401122 40 20 10 Figure 1c, lanes 1 ± 4 demonstrate that despite the 202 202 exquisite speci®city of Streptavidin for biotin, cell lysates contain proteins that bind to Streptavidin. These represent the background noise in this system and was the rationale for pre-clearing the lysates with 109 109 two rounds of Streptavidin-Agarose prior to Western- blot analysis (for details see Experimental procedures). 78 78 These results suggest that KT8529 can be used quite eectively to detect cellular proteins that bind radicicol and validate the Western-blot system as a convenient means to examine such proteins.