Accelerated Publication
THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 277, No. 24, Issue of June 14, pp. 21115–21118, 2002 © 2002 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A.
Two-state Conformational Changes in Inositol 1,4,5Trisphosphate Receptor Regulated by Calcium* Received for publication, April 23, 2002 Published, JBC Papers in Press, April 29, 2002, DOI 10.1074/jbc.C200244200 Kozo Hamada‡§, Tomoko Miyata¶, Kouta Mayanagi¶, Junji Hirota!**, and Katsuhiko Mikoshiba‡!‡‡ From the ‡Laboratory for Developmental Neurobiology, Brain Science Institute, RIKEN (The Institute of Physical and Chemical Research), 2-1 Hirosawa, Wako, Saitama, 351-0198, Japan, the ¶Biomolecular Engineering Research Institute (BERI), 6-2-3 Furuedai, Suita, Osaka, 565-0874, Japan, the !Division of Molecular Neurobiology, Department of Basic Medical Sciences, Institute of Medical Science, University of Tokyo, 4-6-1, Shirokanedai, Minato-ku, Tokyo 108-8639, Japan, and the ‡‡Calcium Oscillation Project, International Cooperative Research Project (ICORP), Japan Science and Technology Corporation (JST), 3-144, Shirokanedai, Minato-ku, Tokyo 108-0071, Japan
Inositol 1,4,5-trisphosphate receptor (IP3R) is a highly controlled calcium (Ca2!) channel gated by inositol 1,4,5-trisphosphate (IP3). Multiple regulators modulate IP3-triggered pore opening by binding to discrete allosteric sites within IP3R. Accordingly we have postulated that these regulators structurally control ligand gating behavior; however, no structural evidence has been available. Here we show that Ca2!, the most pivotal regulator, induced marked structural changes in the tetrameric IP3R purified from mouse cerebella. Electron microscopy of the IP3R particles revealed two distinct structures with 4-fold symmetry: a windmill structure and a square structure. Ca2! reversibly promoted a transition from the square to the windmill with relocations of four peripheral IP3-binding domains, assigned by binding to heparin-gold. Ca2!-dependent susceptibilities to limited digestion strongly support the notion that these alterations exist. Thus, Ca2! appeared to regulate IP3 gating activity through the rearrangement of functional domains.
messenger generated by various extracellular stimuli, neurotransmitters, neuromodulators, hormones, and lights (1, 2). The IP3R is widely distributed in living systems and plays pivotal roles in fundamental processes including fertilization, cellular proliferation and differentiation, cellular signaling, and vesicle secretion (2). Molecular cloning studies have revealed that there are three isoforms of IP3R and that alternative splicing results in several variants of the IP3R (2). These divergent primary structures of the IP3R and their differential distributions have been assumed to award the functional diversity of IP3R by nature. The most characterized type 1 IP3R (IP3R1), a predominant type in rodent cerebellar endoplasmic reticulum (ER) and spine apparatus, plays an integral role in Ca2! signaling (3–5) and neural plasticity (6, 7). The protomer of IP3R1, a 2749-amino acid polypeptide (Mr 313,000), contains the IP3-binding core (residues 226 –578), membrane-spanning domains (residues 2276 –2589), and widespread allosteric sites for intracellular effector molecules (Ca2!, calmodulin, and ATP) and for phosphorylation by protein kinases (cAMP-dependent protein kinase, protein kinase C, cGMP-dependent protein kinase, Ca2!/ calmodulin-dependent protein kinase II, and tyrosine kinase) (2). These cumulative allosteric regulations imply a structural paradigm for global conformational changes within the higher ordered structure of IP3R1. Because Ca2! rigorously determines the channel activity of IP3R and Ca2!-dependent behavior of IP3R is considered to be crucial for spatiotemporal organizations of Ca2! signaling (1, 4), the most important regulator for IP3R is Ca2! per se. Previous functional analysis indicates that a low Ca2! level acts as an essential coagonist for IP3-gated Ca2! release and a high Ca2! level inversely acts as a feedback repressor (4, 8, 9) via Ca2!/calmodulin in part (10). Thereby we assumed that Ca2! could induce alterations in conformational states of IP3R1 underlying such dynamic regulations. An investigation of this hypothesis requires information about structural rearrangements that has heretofore been unclear because of the structural polymorphism within IP3R particles, which is partially due to their fragile architectures, presented by previous electron microscopic studies (11–14). To address this issue, we improved rapid purification of the IP3R1 channel so that we could use electron microscopic study to visualize the domain arrangement and to investigate its structural change by Ca2!. EXPERIMENTAL PROCEDURES
Inositol 1,4,5-trisphosphate receptor (IP3R)1 is a tetrameric ion channel that release Ca2! from intracellular stores in response to the binding of 1,4,5-trisphosphate (IP3), a second * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. § To whom correspondence should be addressed. Tel.: 81-3-54495320; Fax: 81-3-5449-5420; E-mail:
[email protected]. ** Present address: Laboratory of Vertebrate Developmental Neurogenetics, The Rockefeller University, 1230 York Ave., New York, NY 10021. 1 The abbreviations used are: IP3R, inositol 1,4,5-trisphosphate receptor; IP3, inositol 1,4,5-trisphosphate; IP3R1, type 1 IP3R; ER, endoplasmic reticulum; CHAPS, 3-[(3-cholamidopropyl)dimethylammonio]1-propanesulfonate; Lys-C, lysyl endopeptidase. This paper is available on line at http://www.jbc.org
Purification of IP3R1—Immunoaffinity purification of IP3R1 was performed as described previously (15) with the following modifications. Microsomal membrane (3 mg/ml), prepared from mouse cerebella, was solubilized in 50 mM Tris buffer (pH 7.5) containing 1% (w/v) CHAPS, 150 mM KCl, 2 mM dithiothreitol, 200 !M phenylmethylsulfonyl fluoride, 10 !M pepstatin A, 10 !M leupeptin, 10 !M E-64, and 0.2 mM CaCl2 or 1 mM EDTA. The solubilized IP3R1 was mixed with pep6Ab-immobilized beads, incubated at 4 °C, washed, and eluted with 20 !M pep 6. The [3H]-IP3 binding assay was performed as described previously (11). Electron Microscopy—The purified IP3R1 (0.5 !l) was injected into 9.5 !l of 50 mM Tris buffer (pH 7.5) containing 1 mM CaCl2, 1 mM EDTA, or 1 mM EGTA and incubated for 30 min on ice. For heparin-gold labeling, purified IP3R1 was mixed with a 5–10-fold molar excess of heparin-gold (Sigma) and incubated for 10 min on ice. An aliquot of the mixture (4 !l) was applied onto carbon-coated copper grids. Excess solution was removed by filter paper, and the IP3R1 particles were stained with 2 !l of 1% (w/v) uranyl acetate solution. Dried grids were examined on a JEM 1200 EX transmission electron microscope (JEOL)
21115
21116
Conformational Changes in IP3 Receptor
FIG. 1. Dual discrete structures of IP3R1. A, purified IP3R (1.2 !g) on SDS-PAGE (5% gel) stained with Coomassie Brilliant Blue. Electron microscopic images of negatively stained IP3R1 show a typical top view of a windmill-shaped particle (B) and a typical top view of a squareshaped particle (C). An arrow in B indicates a globular domain at the tip of a radial wing. The scale bar represents 20 nm. operated at 80-kV acceleration voltage. Micrographs were taken at magnifications ranging from "25,000 to "50,000. Partial Proteolysis—Partial proteolysis experiments were carried out in the solution containing IP3R1 and lysyl endopeptidase (Lys-C). Reaction mixtures were incubated for 30 min at 37 °C. The proteolysis was stopped by heat treatments at 55 °C for 15 min in the sample buffer including SDS. Proteolytic fragments were analyzed by discontinuous PAGE (5 or 10% gel) and immunoblotting using monoclonal antibodies 18A10 and 4C11 (11, 15). RESULTS AND DISCUSSION
We purified IP3R1 to apparent homogeneity as judged by gel electrophoresis (Fig. 1A) from mouse cerebella, which was functionally active as estimated by the specific activity of maximum binding to IP3 (3 nmol/mg of protein). The purified IP3R1 was negatively stained and imaged in a transmission electron microscope. Electron microscopy explicitly showed two distinct structures in negatively stained samples. One was a windmilllike structure (Fig. 1B), and another was a square-shaped structure (Fig. 1C). The windmill structure contained four segregated radial wings and a central core. Each wing structure appeared to be composed of two domains: a globular domain, which often exhibited a central spot that was densely stained, and a constricted segment forming a bridge between the globular domain and the central core domain. The dimension of the windmill structure was 31 # 2 nm (n $ 76) from the tip of one wing to the tip of the opposite wing. The globular domain in the wing was 8.1 # 0.8 nm (n $ 30) in diameter, and the central core was 9.8 # 0.6 nm (n $ 15) in diameter. The shape and dimensions of the windmill structure are consistent with those previously reported for IP3R isolated from smooth muscle (12). The dimensions of the square structure were 19 # 1 nm (n $ 54) on a side and 24 # 1 nm (n $ 54) on a diagonal line (Fig. 1C), similar to that of small dense projections in the smooth ER of rat Purkinje cells (14). Comparison with the dimensions of ryanodine receptor, another intracellular Ca2! channel (16), provides support that the projected size of the square structure in this study is reasonable because of the ratio in the molecular mass of the protomer. We also found that the IP3R1 particles appeared as other forms, suggesting their variances of orientation or intrinsic flexibilities. Our microscopic data revealed two distinct states of the IP3R1 molecule, leading us to the hypothesis that the conformation of IP3R1 alters. We tried to capture a Ca2!-dependent transition between the dual structures by imaging IP3R1 particles injected into 1 mM CaCl2, 1 mM EDTA, or 1 mM EGTA. Significantly the windmill structures were abundantly observed in the presence of Ca2! (Fig. 2A). In contrast, the relative abundance of windmill structures appeared to decrease in specimens prepared in solution containing 1 mM EDTA (Fig.
FIG. 2. Structural changes in the IP3R1 particle. Electron micrographs of IP3R1 injected into the solution containing 1 mM CaCl2 (A) or 1 mM EDTA (B). Marked projections respectively show top or tilted views of windmill particles (circled) and square particles (boxed). Bars, 100 nm. C, relative abundance of windmill and square particles. IP3R1 was purified in 0.2 mM CaCl2 (a), 1 mM EDTA (b), or neither CaCl2 nor EDTA (c), and then it was injected into 1 mM CaCl2, 1 mM EGTA, or 1 mM EDTA. Ratios were calculated from the identifiable windmill and square structures counted on electron micrographs. For statistical evaluations on each condition, 5–11 fields were evaluated resulting in a total of 169 – 685 molecules indicated at the top of each bar. The standard deviation was calculated by comparing different micrographs.
2B). For statistical evaluation, we counted the windmill particles with more than three identifiable wings and the square forms with a homologous dimension, which had no wing, on electron micrographs. Quantitative analysis clearly indicates a significant difference in the ratio of the two structures in a Ca2!-dependent manner (Fig. 2C). Readdition of CaCl2 into the IP3R1 purified with EDTA restored the windmill configuration, and the number of windmill structures showed a marked reduction upon readdition of EDTA into the IP3R1 purified in the presence of Ca2!, indicating that the structural rearrangements are reversible (Fig. 2C). IP3 did not induce significant changes in each state at this resolution, thus the binding of IP3 may cause a fine structural change to open the channel. Our findings provide the first evidence of structural alterations in IP3R1 molecules. The structural alteration could account for the polymorphism in IP3R particles presented by previous electron microscopic studies (11–14). The unique architectures and conspicuous conformational changes within the IP3R1 particle differ remarkably from those in the ryanodine receptor particle (17, 18). These differences might result from intrinsic properties of the gating machinery of IP3R1. To correlate the structural changes in IP3R1 observed by electron microscopy with changes in solution, we monitored its sensitivity to limited protease digestion. The patterns of degradative intermediates were clearly Ca2!-dependent (Fig. 3). In particular, a 38-kDa fragment detected with 4C11 within the IP3-binding domain was markedly generated by cleavage in CaCl2 solution; however, a 48-kDa intermediate and a 38 – 60kDa ladder of bands detected with 4C11 were dominantly observed in EDTA solution (Fig. 3B). Furthermore a C-terminal 130-kDa fragment detected with 18A10 was abundantly detected by cleavage of purified IP3R1 in CaCl2 solution compared with 85- and 120-kDa fragments (Fig. 3B). These results suggest structural changes in purified IP3R1 rather than a simple acceleration or regulation of proteolysis by Ca2!. The degradation by contaminant protease, such as Ca2!-activated calpains, was insignificant because of negligible production of digested proteins without Lys-C (Fig. 3). In addition, we also investigated the dynamic property of IP3R1 in crude microsomal membrane. The 38- and 30-kDa fragments detected with 4C11 were evenly precipitated in both CaCl2 and EDTA solutions; however, these fragments were more releasable to supernatant fractions in the presence of CaCl2 than in EDTA (Fig. 3C). We also confirmed the reproducibility of these proteolysis experi-
Conformational Changes in IP3 Receptor
2!
FIG. 3. Effects of Ca on partial proteolysis. A, fragments (arrowheads) of purified IP3R digested by Lys-C in the presence of 1 mM CaCl2. The intact IP3R is marked by an asterisk. The products were probed with 18A10. B, fragments of purified IP3R1 digested in the presence of 2.5 mM EDTA (left four lanes) or 1 mM CaCl2 (right four lanes). The products were probed with 4C11 (upper) and 18A10 (lower). An asterisk indicates a ladder of bands (38 – 60 kDa). C, release of digested fragments from crude microsome membrane. The gels show fragments in 15,000 " g supernatants (sup) and precipitates (ppt). The products were detected by 4C11. Microsome fractions were pretreated with 1 mM EDTA (left) or 1 mM CaCl2 (right) and digested by Lys-C (0 –10 !g/ml) at 4 °C (open bars) or 37 °C (solid bars). D, schematic representation of Ca2!-dependent fragments (solid bars). The middle row indicates boundaries of partially digested fragments shown in A. Binding sites for antibodies and heparin, indicated by dotted lines, and IP3-binding and channel regions (bottom) are based on previous studies (2, 20, 21).
ments by using 0.5 mM EGTA and 0.2 mM CaCl2 solution. This heightened ability to release may be related to the structural changes within IP3R1 embedded in the microsomal membrane. Taken together, these biochemical data strongly support the presence of structural alterations in IP3R1. To determine how functional domains are arranged, we used colloidal gold conjugated with heparin, which is a competitive inhibitor of IP3 binding (19) and specifically binds to the Nterminal IP3-binding region (20). Heparin-gold particles bound not only to the windmill structure but also to the square form (Fig. 4, A and B). In the windmill structure, the gold particles bound to the globular domain of the radial wings (Fig. 4A). In the square form, the gold particles attached at the sites close to the corners (Fig. 4B). We assigned these heparin-binding domains to the N-terminal IP3-binding domain. Our results provide the first evidence of domain arrangement in quaternary configurations of the IP3R1 particle. In both states, the distribution of peripheral IP3-binding domains occurred away from the center, a plausible Ca2! gateway, by over 5 nm, suggesting that a long range allosteric transmission took place underlying the IP3-gated Ca2! release. The comparison between dual structures and mapping of heparin-binding sites indicates that the structural transition from square to windmill is caused by the relocation of functional domains. Therefore, we propose a “flapping model” for the large scale rearrangements in IP3R1 (Fig. 4C). The windmill structure may be a consequence of the IP3-binding domain splitting from the channel domain. The dynamic flapping may
21117
FIG. 4. Mapping of heparin-binding sites and proposed flapping model for structural rearrangements within IP3R1. A and B, typical top view images (upper rows) of windmill (A) and square (B) particles and those bound to heparin-gold (middle rows). The contours are shown below each particle to delineate the heparin-gold (closed circles) and the IP3R (dotted lines). The scale bar represents 20 nm. C, IP3-binding domain (light gray), channel domain (gray), and bridge domain (white). Black arrows indicate plausible domain movements.
be mediated by the bridge domain, which may act as a hinge structure. Additionally the digested IP3R retains the assembly of domains under Ca2!-free conditions (21, 22); thus interdomain coupling may also stabilize the more compact square structure. The Ca2!-dependent cleavage sites and releasable regions presented here are candidates for the apparent hinge and interface structures linking between functional domains. Our findings show that the functional regulator altered the relative locations of IP3-binding and channel domains even if there is no IP3, further emphasizing that the domain arrangement is crucial for the transmission of IP3 binding cues to Ca2! pores. Two discrete configurations of IP3R1 imply dual relay modes controlling the allosteric transmission. Which of two structures is more active? As it appears to have an advantage for direct transferring of conformational changes by IP3 binding toward the central channel, we cannot exclude the possibility that the compact square form is an active state. Based on single channel analysis, however, high Ca2! only acts as an activator toward purified IP3R1 (10). Hence it is favorable that the windmill structure is more active. In this case, it is conceivable that the bridge domain presents an effective relay of ligand binding signals. This notion fascinates us as a new type of mechanical control on ligand gating behavior. The threedimensional structure at much higher resolution will answer our central question on the precise pathway of allosteric transmission from the IP3-binding core to the Ca2! gateway underlying IP3-gated channel opening. Ca2! dependence of IP3R is known to interplay with the allosteric regulation by ATP (23) and the cooperative gating by IP3 (24). Thereby the Ca2!-dependent global conformational changes may concern primary states for other allosteric ligands. Since IP3R is known to interact with phosphatidylinositol 4,5-bisphosphate incorporated in the plasma membrane (25), with the transient receptor potential protein, which is a calcium channel assumed to be involved in capacitative Ca2! entry (26), and with the Homer protein linking with metabotropic glutamate receptor involved in neural plasticity (27), it is
21118
Conformational Changes in IP3 Receptor
interesting to study how the dramatic conformational change of IP3R1 alters the association with these signaling molecules. Our novel model for structural rearrangements and the methodology presented here should be useful for understanding of the further biological significance of structural plasticity within IP3R1 in neural systems. Acknowledgments—We thank A. Terauchi (JST) for excellent technical assistance; Y. Kimura (BERI) and H. Sagara (Tokyo University) for discussion about electron microscopy; A. Mizutani (RIKEN) for support in protein purification and for biochemical advice; S. Ohmi (Tokyo University) for peptide synthesis; and A. Takahashi (RIKEN) for 18A10 and 4C11. REFERENCES 1. Berrige, M. J., and Irvine, R. F. (1989) Nature 341, 197–205 2. Furuichi, T., Michikawa, T., and Mikoshiba, K. (1999) in Calcium as a Cellular Regulator (Carafoli, E., and Klee, C., eds) pp. 200 –248, Oxford University Press, New York 3. Miyazaki, S., Yuzaki, M., Nakada, K., Shirakawa, H., Nakanishi, S., Nakade, S., and Mikoshiba, K. (1992) Science 257, 251–255 4. Miyakawa, T., Mizushima, A., Hirose, K., Yamazawa, T., Bezprozvanny, I., Kurosaki, T., and Iino, M. (2001) EMBO J. 20, 1674 –1680 5. Street, V. A., Bosma, M. M., Demas, V. P., Regan, M. R., Lin, D. D., Robinson, L. C., Agnew, W. S., and Tempel, B. L. (1997) J. Neurosci. 17, 635– 645 6. Inoue, T., Kato, K., Kohda, K., and Mikoshiba, K. (1998) J. Neurosci. 18, 5366 –5373 7. Miyata, M., Finch, E. A., Khiroug, L., Hashimoto, K., Hayasaka, S., Oda, S., Inouye, M., Takagishi, Y., Augustine, G. J., and Kano, M. (2000) Neuron 28, 233–244 8. Finch, E. A., Turner, T. J., and Goldin, S. M. (1991) Science 252, 443– 446
9. Bezprozvanny, I., Watras, J., and Ehrlich, B. E. (1991) Nature 351, 751–754 10. Michikawa, T., Hirota, J., Kawano, S., Hiraoka, M., Yamada, M., Furuichi, T., and Mikoshiba, K. (1999) Neuron 23, 799 – 808 11. Maeda, N., Niinobe, M., and Mikoshiba, K. (1990) EMBO J. 9, 61– 67 12. Chadwick, C. C., Saito, A., and Fleischer, S. (1990) Proc. Natl. Acad. Sci. U. S. A. 87, 2132–2136 13. Katayama, E., Funahashi, H., Michikawa, T., Shiraishi, T., Ikemoto, T., Iino, M., and Mikoshiba, K. (1996) EMBO J. 15, 4844 – 4851 14. Kanaseki, T., Ikeuch, Y., and Tashiro, Y. (1998) Cell Struct. Funct. 23, 373–387 15. Nakade, S., Rhee, S. K., Hamanaka, H., and Mikoshiba, K. (1994) J. Biol. Chem. 269, 6735– 6742 16. Lai, F. A., Erickson, H. P., Rousseau, E., and Liu, Q. (1988) Nature 331, 315–319 17. Orlova, E. V., Serysheva, I. I., van Heel, M., Hamilton S. L., and Chiu, W. (1996) Nat. Struct. Biol. 3, 547–552 18. Sharma, M. R., Jeyakumar, L. H., Fleischer, S., and Wagenknecht, T. (2000) J. Biol. Chem. 275, 9485–9491 19. Ghosh, T. K., Eis, P. S., Mullaney, J. M., Ebert, C. L., and Gill, D. L. (1988) J. Biol. Chem. 263, 11075–11079 20. Miyawaki, A., Furuichi, T., Ryou, Y., Yoshikawa, S., Nakagawa, T., Saitoh, T., and Mikoshiba, K. (1991) Proc. Natl. Acad. Sci. U. S. A. 88, 4911– 4915 21. Yoshikawa, F., Iwasaki, H., Michikawa, T., Furuichi, T., and Mikoshiba, K. (1999) J. Biol. Chem. 274, 316 –327 22. Boehning, D., and Joseph, S. K. (2000) EMBO J. 19, 5450 –5459 23. Thrower, E. C., Mobasheri, H., Dargan, S., Marius, P., Lea, E. J. A., and Dawson, A. P. (2000) J. Biol. Chem. 275, 36049 –36055 24. Missiaen, L., De Smedt, H., Parys, J. B., Sienaert, I., Vanlingen, S., and Casteels, R. (1996) J. Biol. Chem. 271, 12287–12293 25. Lupu, V. D., Kaznacheyeva, E., Krishna, U. M., Falck, J. R., and Bezprozvanny, I. (1998) J. Biol. Chem. 273, 14067–14070 26. Boulay, G., Brown, D. M., Qin, N., Jiang, M., Dietrich, A., Zhu, M. X., Chen, Z., Birnbaumer, M., Mikoshiba, K., and Birnbaumer, L. (1999) Proc. Natl. Acad. Sci. U. S. A. 96, 14955–14960 27. Tu, J. C., Xiao, B., Yuan, J. P., Lanahan, A. A., Leoffert, K., Li, M., Linden, D. J., and Worley, P. F. (1998) Neuron 21, 717–726
