and 
Shifman et al. 10.1073/pnas.0606433103. |
Fig. 6. Changes in secondary structure of WT and mutant CaMs caused by Ca2+ binding. Secondary structures of WT and mutant CaMs were assessed by circular dichroism (CD) spectroscopy on an Aviv 62A DS CD spectrophotometer. Protein samples (10 mM) were diluted into a buffer containing 50 mM Tris, 100 mM NaCl, 1 mM MgCl2, and 2 mM CaCl2 (pH 7.2) in a 2-mm quartz cuvette at 25°C to obtain spectra of the Ca2+-bound species. CaCl2 was replaced with 5 mM EGTA to obtain spectra of the Ca2+-free species. The solid line corresponds to the spectra of the Ca2+-free conformation; the dotted line corresponds to the spectra of the Ca2+-bound conformation. (A) WT CaM. (B) CaM-CWT. (C) CaM-NWT.
Fig. 7. Measurement of binding affinity of the mutant CaMs for CaMKII-cbp. Binding of CaMKII-cbp to WT CaM and mutant CaMs was assessed by titrating the CaMKII-cbp into a solution of CaM and monitoring the circular dichroism signal at 222 nm, as described in ref. 1. A change in this signal is produced when CaM binding induces formation of helical structure in the peptide. Aliquots of CaMKII-cbp were titrated into a 5 mM deoxygenated solution of WT or mutant CaM with the use of an autotitrator with a mixing time of 1.5 min and an averaging time of 0.5 min. The binding experiments were conducted in a buffer containing 50 mM Tris, 100 mM NaCl, 1 mM MgCl2, and 2 mM CaCl2 (pH 7.2 at 25°C). q222 was recorded on an Aviv 62A DS CD spectrophotometer. The data were fit to a 1:1 binding model (2) to calculate the KD values.
1. Shifman JM, Mayo SL (2002) J Mol Biol 323:417-423.
2. Wyman J, Gill SJ (1990) Binding and Linkage: Functional Chemistry of Biological Molecules (University Science Books, Mill Valley, CA).
Fig. 8. Binding sites for the N and C halves of CaM in the a-subunit holoenzyme of CaMKII. (A and B) Two views of the x-ray structure of Ca2+/CaM bound to a peptide with the sequence of the CaM-binding domain in CaMKII (PDB ID code 1CDM) (1, 2). (A) Oriented to show the long axis of the CaM-binding peptide. (B) Rotated clockwise 90° around the y axis. CaM is shown as a backbone ribbon and Ca2+ ions are shown as spheres. The N terminus of CaM is blue, and the C terminus is yellow. The CaM-binding peptide is shown in space-filling mode. Residues associating with the N terminus of CaM are colored green; those associating with the C terminus of CaM are colored red. Two important residues in the interaction with the C terminus, Arg-296 and Met-307, are labeled. (C and D) Two views of the x-ray structure of the autoinhibited kinase domain of the a-subunit of CaMKII (PDB ID code 2BDW) (3). (C) The catalytic units form a dimer held together by a coiled-coil interaction between the regulatory regions of the catalytic units. The CaM-binding domains are shown in space-filling mode. The residues of the CaM-binding domain are colored as in A and B except that residues whose side chains participate in the coiled-coil interaction are colored yellow. The rest of the structure is shown as a backbone ribbon. One kinase domain (K-domain 1) is colored silver, and the other (K-domain 2) is colored black. Arg-296 and Met-307 from K-domain 1 are labeled for comparison to A and B. Arg-297, which is important for interaction with the N terminus of CaM, also is labeled in K-domain 1. (D) Rotated clockwise 90° along the x axis to reveal the exposed face of the CaM-binding domain. Residues that interact with the C terminus of CaM are accessible on the surface (see labeled Met-307's from both K domains). a-helices that extend to link with the association domains are colored violet. Black arrows indicate the directions of the extensions. PDB ID code 2BDW. (E) Orientation of the kinase domains with respect to the association domains. The structures of the kinase domain pairs and the oligomerized association domains were determined from separate crystals of peptides containing residues 1-318 and 340-468 of the a-subunit, respectively (3, 4). No structure has been determined for the linking region containing residues 319-339. Here, the pair of kinase domains depicted in D is rotated clockwise 90° around the x axis. Three of the six arms formed by the 12 association domains in the CaMKII holoenzyme (3) are depicted below the kinase domain pair. One arm is oriented toward the kinase domain pair as proposed in refs. 4 and 5. The a-helices in the association domains that link to the kinase domains are colored violet. The violet arrows indicate the approximate positions that linking residues 319-339 would occupy in the proposed structure. Acidic residues at the tip of the arm of the association domains are colored red, and basic residues are colored blue. The PDB file containing the dodecameric structure of the association domains is available as supporting online material for reference 4.
1. Meador WE, Means AR, Quiocho FA (1992) Science 257:1251-1255.
2. Wall ME, Clarage JB, Phillips GN (1997) Structure (London) 5:1599-1612.
3. Rosenberg OS, Deindl S, Sung RJ, Nairn AC, Kuriyan J (2005) Cell 123:849-860.
4. Rosenberg OS, Deindl S, Comolli LR, Hoelz A, Downing KH, Nairn AC, Kuriyan J (2006) FEBS J 273:682-694.
Supporting Text
Methods
Detailed Computational Methods.
The NMR structure of Ca2+-free calmodulin (CaM) was used as a starting structure for all calculations (1). CaM residues 20, 22, 24, 27, 31, 56, 58, 60, 62, and 67 were selected for optimization to obtain the mutant CaM that binds Ca2+ only at C-terminal sites (CaM-CWT). CaM residues 93, 95, 97, 100, 104, 129, 131, 133, 135, and 140 were selected for optimization to obtain the mutant CaM that binds Ca2+ only at N-terminal sites (CaM-NWT). Residues selected for design were divided into core, boundary, and surface classes and their amino acid identity was restricted according to this classification as described in ref. 2. The rotamer library used for the optimizations was based on the backbone-dependent library of Dunbrack and Karplus (3). A potential energy function that included terms for van der Waals, electrostatic, and hydrogen bonding interactions and surface-area based solvation was used to calculate side chain/side chain and side chain/backbone pairwise interactions as described in refs. 2 and 4. The calculated energies served as input to a side-chain selection procedure that used the Dead-End Elimination theorem (5, 6). All standard parameters were used in the optimization procedure except that the distant-dependent dielectric constant was lowered from the standard value of 40r to 10r. This change was done in accordance with our previous results (7) that revealed the benefits of emphasizing electrostatic interactions when optimizing CaM-target interfaces. Calculations were performed with 195 MHz SGI R10000 processors.Fluorescence Measurements of Calcium Binding.
The measured fluorescence signal F was used to calculate the free Ca2+ concentration [Ca2+]free and the concentration of Ca2+ bound to the dye [Ca2+·Dye] according to:and
Here, Fmax and Fmin are the fluorescence signals when Ca2+ is in the saturated and the depleted conditions, respectively. Fmin was obtained by measuring the fluorescence signal of the sample in the presence of 2 mM EGTA. [Dye]tot is the total dye concentration. Ca2+ concentrations were calibrated with the use of the Ca2+ concentration buffer kit (Molecular Probes, Carlsbad, CA), and data were fit with the CaLigator program (www.bpc.lu.se/research/caligator).
We repeated the measurements after addition of WT or mutant CaMs (10 mM). The concentration of Ca2+ bound to CaM [Ca2+·CaM] was obtained by subtracting [Ca2+]free and [Ca2+·Dye] from the total Ca2+ concentration [Ca2+]tot. The ratio of [Ca2+·CaM] and [CaM]tot was used to determine the number of bound Ca2+ ions per CaM molecule. Ca2+ binding to CaM in the presence of either purified rat brain Ca2+/calmodulin-dependent protein kinase II (CaMKII) or peptide having the sequence of the CaM-binding domain of CaMKII (CaMKII-cbp) was measured similarly in a solution of 5 mM CaM/5 mM Fluo4FF/5 mM of either CaMKII or CaMKII-cbp. BSA (0.9 mg/ml) and 5 mM dithioerythritol (DTE) were added to the buffer described above to stabilize the CaMKII enzyme.
Note on Quench-Flow Measurement of Turnover Number for Autophosphorylation of CaMKII
. We found that the decrease in pH to 2.9 was necessary to stop the reaction within »100 ms. When stopped with acidic stop solution, the curve of autophosphorylated CaMKII crossed the x axis at -100 ms when extrapolated to zero autophosphorylation. In contrast, when reactions were stopped by addition of SDS at neutral pH, the curve crossed the x axis at »-700 ms after a similar extrapolation. We presume that this difference reflects the fact that SDS denatures CaMKII more slowly than acidic pH suppresses catalysis of phosphorylation.Discussion
The higher affinity for CaMKII displayed by CaM-CWT (KD = 5 mM, Table 2) compared with CaM-NWT (KD = 20 mM, Table 2) reflects, in part, the intrinsically higher affinity of CaM-CWT for the peptide sequence of the CaM binding domain in CaMKII (70 nM vs. 6 mM for CaM-NWT, Fig. 7). However, recently published structures of the catalytic and association domains of CaMKII (8, 9) suggest that CaM-CWT also may be less sterically hindered than CaM-NWT in its initial interaction with the appropriate site in the CaMKII holoenzyme (Fig. 8). The atomic structure of CaM bound to the a-helical CaMKII-cbp reveals that the C- and N-terminal lobes of CaM interact with side chains on opposite faces of the helix (Fig. 8 A and B) (10). Residues Arg-296, Leu-299, Ile-303, Thr-306, Met-307 and Thr-310 (colored red in Fig. 8 A and B) make contact with the C-terminal lobe of CaM. Residues Arg-297, Lys-298, Gly-301, Leu-304, Thr-305, and Leu-308 (colored green in Fig. 8 A and B) contact the N-terminal lobe. The x-ray structure of the catalytic domain of CaMKII reveals that it forms a dimer held together by a coiled-coil interaction between helices formed by residues 273-317, which includes the CaM-binding domain (Figs. 8 C and D) (9). Rosenberg et al. (9) propose that interaction of a CaM molecule with a CaM-binding domain in a dimer disrupts the coiled-coil interaction, freeing both catalytic subunits and increasing the affinity of the second subunit for CaM. This model explains the high cooperativity (Hill coefficient »2) for binding of CaM to the CaMKII holoenzyme. In the dimer structure, the side chains that interact with the N terminus of CaM (green) are buried in a pocket formed by the catalytic domains (Fig. 8C). In contrast, many of the side chains that interact with the C terminus of CaM (red) are exposed on one face of the dimer (Fig. 8D).
A structure of the entire dodecameric holoenzyme of CaMKII has been proposed based on the x-ray structures of the catalytic and association domains, and on small angle x-ray scattering (SAXS) data from the holoenzyme in solution. In this proposed structure, each dimerized pair of catalytic subunits is held in place by the association domains that interact with other association domains to form a rosette (8, 9, 11). The data suggests that the face of each dimer that contains the binding site for CaM-CWT is pointed toward the outward end of an arm of the rosette. Because the structure of »20 of the residues that form the link between the kinase domain dimers and the arms of the rosette (Fig. 8E) is unknown, the size of the gap into which the C terminus of CaM must insert is not well defined and, in fact, may be flexible. For comparison, the violet a-helices in the association domains depicted in Fig. 8E are » 22 residues long.
In support of the notion that the linker region between the kinase and association domains influences affinity for the C terminus of CaM, we note that the affinity of CaM for b-subunits in holoenzymes of CaMKII is »2-fold higher than for a-subunits (12). The most prominent difference between these two subunits is the insertion into the b-subunit of 68 additional residues in the linker region between the catalytic and association domains (13). These additional residues may increase the size of the gap into which the C terminus of CaM must insert and, thus, decrease the on rate for association between the C terminus of CaM and the CaM-binding domain. Earlier work examining activation of isoforms of Drosophila CaMKII by a series of CaM point mutants, also suggested that splice variants with longer linker regions were activated by CaM at lower concentrations than those with shorter linker regions (14).
The studies by GuptaRoy et al. (14) also provided evidence for two phases of activation of CaMKII by CaM, a binding phase and an activation phase. In their studies, residue E31 in Ca2+-binding site 1 in the N terminus appeared critical for activation of CaMKII, as well as residue E140 in binding site 4 in the C terminus. Our studies are consistent with this finding and suggest further that binding of CaM to CaMKII after binding of Ca2+ to the C-terminal lobe may be sufficient to position the N-terminal lobe of CaM so that it can activate CaMKII even without Ca2+ bound to the N-terminal-binding sites.
1. Kuboniwa H, Tjandra N, Grzesiek S, Ren H, Klee CB, Bax A (1995) Nat Struct Biol 2: 768-776.
2. Dahiyat BI, Mayo SL (1997) Science 278:82-87.
3. Dunbrack RL, Karplus M (1993) J Mol Biol 230:543-574.
4. Street AG, Mayo SL (1999) Struct Fold Des 5:R105-R109.
5. Desmet J, De Maeyer M, Hazes B, Lasters I (1992) Nature 356:539-542.
6. Gordon DB, Hom GK, Mayo SL, Peirce NA (2003) J Comp Chem 24:232-243.
7. Shifman JM, Mayo SL (2003) Proc Natl Acad Sci USA 100:13274-13279.
8. Rosenberg OS. Deindl S, Comolli LR, Hoelz A, Downing KH, Nairn AC, Kuriyan J (2006) FEBS J 273:682-694.
9. Rosenberg OS, Deindl S, Sung RJ, Nairn AC, Kuriyan J (2005) Cell 123:849-860.
10. Meador WE, Means AR, Quiocho FA (1992) Science 257:1251-1255.
11. Hoelz A, Nairn AC, Kuriyan J (2003) Mol Cell 11:1241-1251.
12. Miller SG, Kennedy MB (1985) J Biol Chem 260:9039-9046.
13. Bulleit RF, Bennett MK, Molloy SS, Hurley JB, Kennedy MB (1988) Neuron 1:63-72.
14. GuptaRoy B, Beckingham K, Griffith LC (1996) J Biol Chem 271:19846-19851.