Allosteric Network of camp-dependent Protein Kinase Revealed by Mutation of Tyr204 in the PC1 Loop

Transcription

1 doi: /j.jmb J. Mol. Biol. (2005) 346, Allosteric Network of camp-dependent Protein Kinase Revealed by Mutation of Tyr204 in the PC1 Loop Jie Yang 1, Siv M. Garrod 2, Michael S. Deal 1, Ganesh S. Anand 1 Virgil L. Woods Jr 3 and Susan Taylor 1,2 * 1 Howard Hughes Medical Institute, University of California San Diego, 9500 Gilman Drive, La Jolla, CA 92093, USA 2 Department of Chemistry and Biochemistry, University of California San Diego, 9500 Gilman Drive, La Jolla, CA 92093, USA 3 Department of Medicine University of California San Diego, 9500 Gilman Drive La Jolla, CA 92093, USA *Corresponding author Previous studies on the catalytic subunit of camp-dependent protein kinase (PKA) identified a conserved interaction pair comprised of Tyr204 from the PC1 loop and Glu230 at the end of the af-helix. Single-point mutations of Tyr204 to Ala (Y204A) and Glu230 to Gln (E230Q) both resulted in alterations in enzymatic kinetics. To understand further the molecular basis for the altered kinetics and the structural role of each residue, we analyzed the Y204A and the E230Q mutants using hydrogen/deuterium (H/D) exchange coupled with mass spectrometry and other biophysical techniques. The fact that the mutants exhibit distinct molecular properties, supports previous hypotheses that these two residues, although in the same interaction node, contribute to the same enzymatic functions through different molecular pathways. The Tyr204 mutation appears to affect the dynamic properties, while the Glu230 mutation affects the surface electrostatic profile of the enzyme. Furthermore, H/D exchange analysis defines the dynamic allosteric range of Tyr204 to include the catalytic loop and three additional distant surface regions, which exhibit increased deuterium exchange in the Y204A but not the E230Q mutant. Interestingly, these are the exact regions that previously showed decreased deuterium exchange upon binding of the RIa regulatory subunit of PKA. We propose that these sites, coupled with the PC1 loop through Tyr204, represent one of the major allosteric networks in the kinase. This coupling provides a coordinated response for substrate binding and enzyme catalysis. H/D exchange analysis also further defines the stable core of the catalytic subunit to include the ae, af and ah-helix. All these observations lead to an interesting new way to view the structural architecture and allosteric conformational regulation of the protein kinase molecule. q 2004 Elsevier Ltd. All rights reserved. Keywords: hydrogen/deuterium exchange; camp-dependent protein kinase; protein dynamics; long-range interaction; catalytic mechanism Introduction Abbreviations used: capk, camp-dependent protein kinase; C-subunit, the catalytic subunit of campdependent protein kinase; Y204A, the Tyr204 to Ala mutant of C-subunit; E230Q, the Glu230 to Gln mutant of C-subunit; H/D, hydrogen/deuterium; IP20, a 20-mer inhibitory peptide of the C-subunit; SP20, a 20-mer substrate peptide for the C-subunit; Kemptide, a commonly used peptide substrate for C-subunit; Ala- Kemptide, the phosphor-acceptor Ser in Kemptide is replaced by Ala; RIa(91 244), a truncated form of the regulatory RIa subunit of camp-dependent protein kinase that comprises residues ; DXMS, deuterium exchange coupled with mass spectrometry; PKA, protein kinase; SPR, surface plasma resonance. address of the corresponding author: staylor@ucsd.edu Protein kinases constitute one of the largest gene families and serve as critical switches for regulation of a myriad of biological networks. They share a conserved kinase fold consisting of a small and a large lobe that are hinged together to create a wedge-shaped active site cleft, as represented by the first such structure of the catalytic subunit (C) of the camp-dependent protein kinase (capk, PKA). 1 These enzymes represent a molecule that is comprised of an intricate network of interactions that make it poised for transferring the g-phosphate group from ATP to a protein substrate. In addition to the exquisite orchestration of the highly conserved active site cleft, the enzymes have evolved to /$ - see front matter q 2004 Elsevier Ltd. All rights reserved.

2 192 H/D Exchange Reveals PKAc Coupling Sites be sensitive to the binding of a variety of substrates, polypeptide inhibitors, and scaffold proteins. Crystal structures of the C-subunit alone or in complexes with various ligands have been solved. 2 7 The collection of structures has displayed a series of pictures with detailed interactions between the enzyme and the substrates along the reaction pathway. 8 While crystal structures provide us with snapshots of different conformational states, they do not provide much information regarding the dynamic features of catalysis and protein protein recognition. This requires solution-based methods. One way for monitoring protein dynamics in solution is to use hydrogen/deuterium (H/D) exchange coupled with mass spectrometry. 9,10 Under the conditions normally used for protein studies, only the exchange of the backbone amide protons is measured. The rate of deuterium exchange reflects the solvent accessibility to and the microenvironment of amide protons. Recent improvements in mass spectrometry and data analysis have facilitated increased sophistication in understanding protein dynamics in solution. These include the application of the matrix-assisted laser desorption/ ionization (MALDI) H/D exchange technique, 11 and the comprehensive deuterium exchange mass spectrometry system (DXMS). 12 We have used both methods to map interacting surfaces and conformational changes in both the catalytic and the regulatory (R) subunits of PKA We have also used the DXMS system to map the domain boundaries of the dual-specific PKA anchoring protein, DAKAP2. 17 Many functionally important residues can be identified through sequence and structural analysis. However, site-directed mutagenesis has also demonstrated the importance of a number of residues for enzyme function even though their precise roles are hard to predict based solely on structure. For example, we previously characterized two mutant forms of the C-subunit. In one case, Glu230 was substituted with Gln (E230Q), and in the other case, Tyr204 was substituted with Ala (Y204A). Although they are not located at the active site where phosphoryl transfer takes place, both mutants showed kinetic defects. Comparing with the wild-type enzyme, the k cat for Y204A and E230Q was 30-fold and twofold lower, respectively, and the K m for peptide substrate (Kemptide) was tenfold, and 45-fold higher, respectively. The K i for the inhibitory Ala-substituted Kemptide was 45- fold and 20-fold higher for Y204A and the E230Q, respectively. 18,19 Detailed analysis showed that the phosphoryl transfer rate in the E230Q mutant decreased 25-fold. As seen in Figure 1, Glu230 is located at the C terminus of the af-helix (Ala218-Ala232), while Tyr204 is in the PC1 loop (Leu198 Leu205). Coming from different subdomains, they converge at the recognition site for the P-2 (second residue N- terminal to the phosphorylation P-site) Arg in the substrate. From the crystal structure of the C-subunit with substrate/inhibitor peptide bound, Glu230 is ion-paired with the P-2 Arg. In addition, one of the carboxylate oxygen atoms of Glu230 is within hydrogen-bond distance from the hydroxyl group of Tyr204, and the other oxygen atom is within hydrogen-bond distance from the side-chain of Arg133 from the ad-helix. Thus, there appears to be a balance of interactions among these residues (Figure 1). Crystal structures of either E230Q (J. Wu, J.Y., N.-H. Xuong & S.T., unpublished results) or Y204A 7 did not provide an obvious explanation why these mutations would so dramatically change the kinetics of the enzyme reaction, especially the rate of phosphoryl transfer. Kinetic analysis coupled with a comprehensive structural analysis suggested that Tyr204 is involved in packing of the hydrophobic core of the large lobe. We predicted that mutation to Ala might alter the dynamics of the core and hence affect the stability of the catalytic and PC1 loops that are docked onto the core. 7,19 These disturbances could provide an explanation for the altered kinetics of Y204A. In the case of the E230Q mutant, the change in surface electrostatics most likely accounts for the altered kinetic properties. 18,20 In this study, we utilized biophysical methods to test these hypotheses and to better understand the mechanism of the conformational allostery within the protein molecule. We report the characterization of biophysical properties of the mutants and the observation of long-range dynamic consequences caused by the Tyr204 to Ala mutation. In addition, DXMS analysis further defines the essential stable core of the large lobe that has been recognized previously from structural analysis. Along with previous structural and biochemical data, these biophysical methods reveal a more detailed structure-function profile of the C-subunit. This approach of assessing dynamic consequences of single-point mutations to probe long-range interaction networks should be widely applicable to other molecules. Results Thermostability of the mutants Purified wild-type and mutant proteins were diluted into the same buffer to the same concentration and their CD spectra were measured (see Experimental Procedures). The wavelength scan profiles of Y204A and E230Q were very similar to that of the wild-type protein (data not shown). This suggested that there were no detectable changes in the secondary structure caused by the mutation, which is consistent with the fact that no significant differences were observed in the crystal structures of the mutants compared to the wild type protein (J. Wu, J.Y., N.-H. Xuong & S.S.T., unpublished results). 7 However, the temperature scan profiles showed a 6 deg. C decrease in melting temperature (T m ) for

3 H/D Exchange Reveals PKAc Coupling Sites 193 Figure 1. Glu230 and Tyr204 pair in the C-subunit of PKA. Overall structure of the C-subunit bound with the substrate peptide SP20 is shown at the left, with SP20 colored in blue, the PC1 loop in tan, the catalytic loop in red, and the ad and af-helix in yellow. Tyr204, Glu230 and Arg133 are rendered as yellow sticks, and P-2 Arg as blue sticks. The Tyr204 Glu230 interaction environment is shown in detail at the right. The golden and red spheres locate the C a atoms of Thr197 and P-site Ser, respectively. Dotted lines indicate that the two atoms are in hydrogen-bond distance. The ternary C- subunit structure with SP20 and transition-state analogue aluminum fluoride (PDB ID 1L3R) was used to generate this Figure. Y204A compared with the wild-type protein, indicating a reduction in its thermostability. On the other hand, the T m for the E230Q mutant was similar to that of the wild-type protein (Table 1). Further experiments showed that the presence of MgATP and IP20 increase the thermostability of Y204A, similar to the wild-type protein, 21 although the T m was always slightly lower under the same conditions (Table 1). IP20 by itself did not exert significant effect on the T m of wild-type or the mutant proteins, whereas MgATP increased the T m of the wild-type or the mutants by deg. C. Significantly, MgATP and IP20 together brought about 12.3 deg. C and 9.2 deg. C increase in T m for the wild-type and the Y204A mutant protein, respectively, but had no synergistic effect on the T m of the E230Q mutant. The synergistic effect of MgATP and IP20 is probably correlated with the shift in conformation equilibrium to favor the closed state upon ligand binding. It appeared that at the concentrations of protein and ligands used in the experiment, there was no ternary complex formation, or the complex is not stable for E230Q. Surprisingly, and in agreement with these results, the crystal structure of E230Q revealed an open conformation without any bound ligands even though the crystals were grown in the presence of higher concentrations of MgATP and IP20 that produced the ternary complex for Y204A and the wild-type protein (J. Wu, J.Y., N.-H. Xuong & S.S.T., unpublished results). Somehow E230Q seems to have very low actual affinity for MgATP and IP20. This is intriguing, because previous data have shown that K m of ATP and K i of the inhibitory Ala-Kemptide were both twofold higher for Y204A than for E230Q, 18,19 yet Y204A crystallized as a ternary complex in the presence of MgATP and IP20, but the E230Q did not. 7 It appears that the pathway for the synergistic effect of MgATP and IP20 is destroyed in E230Q, but retained in the Y204A mutant. Since no significant structural differences were observed for the crystal structure of Y204A complexed with MgATP and IP20, 7 we speculated that the difference in thermostability of Y204A and the wild-type protein is probably a reflection of differences in protein dynamics. DXMS thus seemed to be the most appropriate tool to address this question. Deuterium exchange analysis at the catalytic loop of the C-subunit The deuterium exchange profiles of the apo form of Y204A and E230Q mutants were analyzed and compared with that of the wild-type protein. For all the samples, after pepsin digestion, recovered peptides covered more than 90% of the entire sequence. The pepsin digestion map of the Y204A mutant is shown in Figure 2, maps of other samples Table 1. Thermostability (T m (8C)) as an indicator of protein stability and ligand binding affinity apo CIP20 CMgATP CMgATP/IP20 Wt-C 48.90G G G G0.51 Y204A 42.29G a 51.57G0.27 E230Q 48.92G G G G0.96 a Representative of duplicated results.

4 194 H/D Exchange Reveals PKAc Coupling Sites Figure 2. Pepsin digestion peptide coverage map of the Y204A mutant. Peptides identified and analyzed are shown as bars. Those used in this study are shown in dark grey. are provided as Supplementary Material. Timecourses of deuterium exchange profiles of all common peptides among the mutants and the wild-type protein were compared and are provided as Supplementary Material. E230Q was essentially identical with the wild-type protein: no significant difference in exchange rates of recovered peptides could be detected. Most regions of Y204A also exhibited exchange profiles very similar to those of the wild-type protein. However, peptides covering the catalytic loop (Asp165 Asn171) in Y204A showed enhanced exchange rates (Figure 3B). Three common peptides (recovered in all three protein samples) in this region ( , , ) all showed very similar exchange profiles: starting from about two minutes to 50 minutes, each of the peptides from Y204A incorporated two more deuterons, whereas the exchange profiles of E230Q peptides were similar to those of the wild-type protein (Figure 3B, showing peptide ). Exchange profiles of peptides flanking this region are also shown (Figure 3A and C). On one side, peptide showed no differences in exchange rates for both mutants and wild-type proteins

5 H/D Exchange Reveals PKAc Coupling Sites 195 Figure 3. Increased deuteron exchange in the catalytic loop (Arg165 Asn171) of the Y204A mutant. The C-subunit is shown as a ribbon diagram. Glu230 and Tyr204 are shown as red and orange sticks, respectively. T197 is shown as an orange sphere, to locate the activation segment and the PC1 loop. The catalytic loop (C-loop) is highlighted in red and its flanking region in tan. A C, The time-course of the deuteron exchange profiles of peptides from the catalytic loop B and its flanking regions A and C. The color scheme in A C is red for the Y204A, gray for the E230Q and black for the wild type protein. D F, Compare the exchange profiles of Y204A (red) and wild-type protein (black) in the absence (circle) and presence (triangle) of MgATP/IP20 for the same peptides shown in A C. The X-axis shows exchange time in logarithm scale. The Y-axis indicates the number of deuterons added to the peptide. (Figure 3A); while peptide on the other side already started to show some differences (Figure 3A). The crystal structure of Y204A complexed with MgATP and IP20 showed that the conformation of the catalytic loop and even temperature factors of each residue therein were almost indistinguishably low compared with those of the wild-type ternary complex. Binding of MgATP and IP20 most likely reverts the catalytic loop of the mutant to the wildtype conformation and stability. 7 We thus performed the DXMS experiments on Y204A in the presence of MgATP and IP20. Indeed, the exchange rate for the catalytic loop peptides in the ternary complex of MgATP/IP20/Y204A decreased dramatically (Figure 3E). Almost no deuteron exchange was observed for those same peptides that had increased exchange rates in the apo-form (Figure 3E and F, red symbols). The absence of deuteron exchange was similar to what was observed for the wild-type ternary complex (Figure 3E and F, black symbols). 22 The reduction in deuterium exchange may be resulted from reduced solvent access to, or stabilization of the catalytic loop, upon binding of MgATP/IP20 to the protein. Taking into account the lower temperature factor profile of the catalytic loop in the ternary complex than that in the apo form, as represented in the wild-type structure, it is likely that the stabilization of the catalytic loop makes an important contribution to the extremely low exchange. Deuterium exchange profiles at other regions of the C-subunit Peptides from three additional regions in Y204A,

6 196 H/D Exchange Reveals PKAc Coupling Sites Figure 4. Increased deuteron exchange in RI-binding interface in Y204A. Three regions (labeled as 1 3) exhibit increased exchange rates in apo-y204a are highlighted in red. The catalytic loop remains lightly colored for reference. The surface model on the top right shows the surface location of these three regions. The color scheme for the deuteron exchange profiles is the same as that in Figure 3. A C, The exchange profiles in three peptides (as indicated) for the apo Y204A mutant, the E230Q mutant and the wild-type protein. D F, Compare the exchange profiles of Y204A (red) and the wild-type protein (black) in these regions in the absence (circle) and presence (triangle) of MgATP/IP20. Peptide 1, ; peptide 2, ; and peptide 3, Axis definition is the same as in Figure 3. peptide , peptide and peptide , also exhibited increased exchange rates. They were labeled as regions 1, 2 and 3, respectively, in Figure 4. Region 1 is located after the PC1 loop, region 2 included the C-terminal part of the ag-helix and the following loop region, and region 3 included the ai-helix and the preceding loop region. Only one common peptide was recovered in region 1. After 50 minutes, this region in the Y204A mutant had exchanged three more deuterons than the same region in the wild-type protein (Figure 4A). For region 2, three common peptides were recovered. In the Y204A mutant, it appeared that the exchange had already reached a maximum after five minutes, whereas in the wild-type protein, it took more than 50 minutes to achieve the same extent of deuteration (Figure 4B). Three common peptides were also recovered for region 3. Again, in the Y204A mutant, this region exchanged two or three more deuterons in the time-frame examined (Figure 4C). In E230Q, the exchange rates for these regions were very similar to the wild-type protein (Figure 4A C). Interestingly, these regions have been implicated in the involvement of RIa-subunit binding, since they showed decreased exchange rates in the C-RIa complex. 13 Crystal structures of C-RIa complex indicated that residues from two of the peptides, peptide and , formed part of the interface of the complex. 23 Interestingly, the presence of MgATP and IP20 slowed the exchange rates in these regions to a level similar to that of the wild-type protein (Figure 4D F),even though IP20 does not directly interact with these regions of the protein. In the case of wild-type protein, the presence of MgATP/IP20 did not bring significant further reductionintheexchangeratesintheseregions (Figure 4D F, black symbols).

7 H/D Exchange Reveals PKAc Coupling Sites 197 Altered binding kinetics of Y204A: RIa interaction One of the functions that has been proposed for Tyr204 is to secure the PC1 loop onto the stable hydrophobic core in the large lobe. 7 Increased flexibility of the PC1 loopiny204acanbeexpected. However, we were surprised to detect increased exchange rates for exactly those three regions that were protected upon RIa binding. Based on our observations, we therefore predicted that Y204A might exhibit altered kinetics in its interaction with the RIa-subunit. To answer this question, we measured the binding kinetics of the C-subunit and RIa by surface plasma resonance (SPR) on a BiaCore Full-length RIa contains the dimerization domain and an inhibitory region at the N terminus and two camp-binding domains (CBDA and CBDB) at the C-terminal part of the molecule. H/D exchange data have established that a short form of RIa containing the inhibitory region and CBDA, RIa (91 244), is sufficient for interacting with the C- subunit, since the C-RIa holoenzyme and C-RIa(91 244) revealed the same interaction interfaces. 13 To quantify the interactions with RIa(91 244), either the Y204A mutant or the wild-type protein was immobilized on the chip surface and the RIa was flowed through the cell. As seen in Table 2, the difference in the binding kinetics was clearly reflected more on the association rate constant (k a ) than the dissociation rate constant (k d ). The mutant showed a slight threefold increase in k d, but a tenfold decrease in k a.these resulted in a 30-fold increase in the equilibrium dissociation constant K D for the mutant (Table 2). This is in line with the hypothesis that the Y204 to Ala mutation affects the dynamics of the PC1 loop, which affects more its initial association with substrate. Preliminary studies also suggested that the k d for RIa increased significantly for E230Q mutant (data not shown), although its dynamic properties monitored by H/D were similar to those of the wild-type protein. The reason for the altered k d is likely due to the change in the surface electrostatic profile, since the charge attraction is a key factor in enzyme substrate recognition in PKA. Thus, both Y204A and E230Q affect RIa-binding but through different mechanisms. Y204A affects the dynamics of the RIa-binding sites, while E230Q affects the surface electrostatic field. Regions in the C-subunit resistant to deuteron exchange The above DXMS experiments were carried out in Table 2. Binding kinetics of the Y204A or the wild-type protein to RIa(91 244) k a (1/M s) k d (1/s) K D (M) Wt 7.2! !10 K10 Y204A 7.2! !10 K9 Data represent duplicated binding experiments. a time-frame of seconds. As summarized in Figure 5, peptides and from the ae-helix (His142 His158), and peptide from the ah-helix (Ser263 Leu272) in all three proteins showed almost no deuteron exchange in this time-frame. Peptides and from the af-helix (Ala218 Ala233) were not recovered in the E230Q mutant, but in the Y204A and the wildtype protein, they also showed no exchange. To examine the degree of stability in these regions, Y204A was subjected to a longer duration of deuteron exchange. As shown in Figure 5, all these regions showed negligible deuteron exchange even after 48 hours at room temperature. Before carrying out DXMS experiments, CD spectra were measured to confirm that the protein did not undergo significant conformational changes over this long time-frame. The T m remained unchanged after incubation at room temperature for 48 hours (data not shown). Notably, peptides cover the ajhelix (Trp302 Tyr306), peptide from the E230Q and peptide from the wild-type protein also showed no deuterium exchange after 3000 seconds but, since a similar peptide in this region was not recovered from the Y204A protein, it is not known whether there were any differences in this mutant. Discussion In this study, we utilized several solution-based methods to examine the consequences of singlepoint mutations of an interaction pair, Tyr204 Glu230 in the catalytic subunit of PKA. It provides in-depth insights into the extended molecular pathways that lead from specific residues to global functions. By probing the conformational dynamics of the mutants, we identify the catalytic loop as well as three distant sites on the surface of the molecule to be coupled with the PC1 loop. We propose that this allosteric coupling of distant sites for coordinated function is a common mechanism in protein molecules. Lockless et al. have proposed pathways of energetic connectivity in protein molecules through theoretical energy calculations. 24 In the present study, we demonstrate by experimental approaches the organization of such pathways and the propagation of conformational changes from one site to other sites. In the crystal structures of various C-subunits, be it the apo form or the liganded forms, Glu230 and Tyr204 maintain a hydrogen-bond interaction through their side-chains (Figure 1). This pair of interaction also appears to be highly conserved in Ser/Thr kinases. A Trp is often seen at the position of Tyr204 in most of the tyrosine kinases, and the hydrogen-bond interaction with the paired Glu is retained through the N 3 atom in the indole ring of Trp. 7 In PKA, this interaction node provides a platform to dock the P-2 Arg from the substrate. Previous studies showed that both E230Q and Y204A exhibited a decreased phosphoryl transfer

8 198 H/D Exchange Reveals PKAc Coupling Sites Figure 5. Stable core in the large lobe revealed by DXMS. In the ribbon diagram of the C-subunit, three regions that exhibited no deuteron exchange up to 48 hours in Y204A are colored in dark blue. The aj-helix is colored in light blue. The catalytic loop is colored in red. The orange sphere locates the C a atoms of Thr197. The charts show the deuterium exchange profiles of each one of the peptides from the ae, af and ah-helix. Axis definition is the same as in Figure 3. rate as well as an increased K m for substrate peptide. 18,19 Furthermore, K i of the inhibitory Ala- Kemptide for both mutants showed a fold increase. Although both residues locate to the same interaction node, the Y204A mutant shows decreased thermostability and increased deuterium exchange rates in the catalytic loop and three additional regions on the surface, while the thermostability or the deuterium exchange profiles of the E230Q mutant are very similar to those of the wildtype protein. However, while MgATP and IP20 can synergistically increase the thermostability of Y204A as they do for the wild-type protein, they failed to increase the thermostability of the E230Q mutant (Table 1). This is also reflected in the crystal structure of the E230Q, where it assumes an open conformation with no ligands bound, even though the crystal was grown in the presence of MgATP and IP20 that produced the ternary complex for Y204A and the wild-type protein (J. Wu, J.Y., N.-H. Xuong & S.S.T., unpublished results). The E230Q mutant appears to lose the ability to convey the synergistic response of MgATP and IP20 on its thermostability. Using fluorescence anisotropy, Li et al. have shown that the C-subunit assumes a closed conformation in the presence of MgATP and IP The disruption of the synergistic effect of MgATP and IP20 in the E230Q mutant may play a critical role by favoring the enzyme in an open-state conformation as seen in the crystal structure. This would contribute greatly to its deficiency in catalysis, since the ability of the enzyme to toggle between open and closed conformation states in response to ligands is essential for the catalytic cycle. DXMS analysis reveals that Tyr204 to Ala mutation caused not only increased deuterium exchange of the catalytic loop but also selectively increased the exchange at three additional sites on the surface of the molecule. These surface regions correspond strikingly to the surfaces of the wildtype C-subunit that showed protection upon RIa interaction. Interestingly, all these regions in the Y204A mutant are protected in the presence of MgATP and IP20, even though these ligands do not have direct contact with any of these regions. From the crystal structure of the C-RIa(91 244) complex, only region 1 and 2 in the C-subunit (Figure 4) are involved in direct contacts with the RIa-subunit. 23 Based on these observations, we propose that all these regions are allosterically coupled with the PC 1 loop (Figure 6). Perturbation (mutation in this case, or binding of RIa) at any one of the sites will be transmitted to other coupling sites through Figure 6. Long-range coupling of the PC1 loop with the catalytic loop and three additional regions on the surface of the kinase. Part of the large lobe of the catalytic subunit is shown as a in ribbon diagram. Regions resistant to deuterium exchange are shown in dark green. Regions exhibiting increased deuterium exchange in the Y204A mutant are highlighted in red. The PC1 loop is colored in tan. Yellow arrows illustrate hydrophobic interactions. Dotted lines illustrate hydrogen-bond interactions. The bottom cap refers to the segment from the ag-helix to the aj-helix. This part is connected to the hydrophobic core through hydrophobic interactions and the ionic interaction of the Glu208 Arg280 pair.

9 H/D Exchange Reveals PKAc Coupling Sites 199 allosteric molecular interaction networks. The communication of these regions could be through the hydrophobic core in the large lobe, especially for regions 2 and 3 that are located at the bottom cap of the molecule. Flanked by these two regions is the ah-helix that exhibits resistance to deuterium exchange over the 48 hour period. Figure 6 highlights some of the specific interactions connecting these coupling sites. As shown earlier, Leu167 from the catalytic loop and Tyr204 from the PC1 loop both participate in the packing of the hydrophobic core. 7 Glu208 from the PC1 loop and Arg280 from the end of the ah-helix form a highly conserved ion pair that will play an important role in connecting the PC1 loop and the loop between the ag and ah-helix (region 2). The hydrophobic methylene portion of the side-chains of this pair also packs loosely against Trp222 from the af-helix that forms the core of the large lobe. Although a protein molecule comprises a very complex network of interactions, these interactions are not connected randomly but are organized into clusters of coupled sites to achieve a precise coordination for enzymatic function. An analogy to this concept is to view the organization of a protein sequences as motifs and linkers. The coupling of the PC1 loop with the catalytic loop and substrate-binding sites provides a mechanism for coordinated multiple-site substrate binding to optimally position the substrate and the catalytic loop for phosphoryl transfer. We have shown in previous studies that, for example, binding of substrate to the PC1 loop promotes the interaction between Thr201 (from the PC1 loop) and Asp166 (the proposed catalytic base in the catalytic loop). Other clusters of coupled sites may be used for other functions, for example, in interaction with other partners or binding with a different substrate. This model provides a global picture for understanding the long-range interactions that have been observed for the C-subunit and other proteins. Similar long-range effects have been observed in other kinase molecules. Lee et al. have recently shown that in p38a MAP kinase and ERK2 kinases, binding of the DEJL motif-containing peptide at a region near the ad ae loop caused increased dynamics at the distal PC1 loop region. 26 Through recent structural studies of the apo form of the wild-type C-subunit 3 and the Y204A mutant, 7 the stable core of the large lobe with the ae and afhelix at its centre has been further appreciated. Using DXMS, we can now experimentally define this core more precisely to include also the ahhelix. Intertwined with these stable regions are the more flexible segments that are involved in substrate binding and catalysis. For example, between the most stable ae and af-helices are those catalytically important loops, including the catalytic loop, the activation loop and the PC1 loop. Between the af-helix and the stable ah-helix, is the less stable ag-helix, whose role in substrate binding is emerging; and the ah-helix may play an important role for anchoring the bottom cap (including region 2 and 3) of the molecule to the core (Figure 5, left). Alternatively, one can appreciate more of the structural architecture by viewing the large lobe as a stable core anchoring or presenting flexible regions that are poised for allosteric molecular interactions. By this design, the enzyme not only has a solid structural support but also has regions to achieve conformation flexibilities required for function. Communication through the common core may also provide the molecular basis for the long-range allosteric coupling of different subsites into a functional cluster. Experimental Procedures Protein purification E230Q and Y204A single mutations were introduced into the murine Ca subunit as described earlier. 18,19 Mutant proteins were expressed from prset vector in E. coli strain BL21 (DE3) cells. Protein expression and purification were essentially the same as that of the wildtype C-subunit with a slight modification of the buffer ph. 7 Thermostability determination by circular dichroism spectrometry Stability of wild-type and mutant proteins was assessed using circular dichroism spectrometry, by monitoring the ellipiticity change at 222 nm upon thermal denaturation. Purified proteins were diluted to 0.1 mg/ml (2.5 mm) in 10 mm Pipes (1,4-piperazinebis(ethansulfonic acid)), ph 7.0, plus 150 mm NaCl, in the presence or absence of 5mM MgCl 2, 1 mm ATP, and/or 50 mm inhibitory peptide IP20. Samples were incubated over a range of temperature from 25 8C to 65 8C at a rate of 1 deg. C/minute, in a rectangular quartz cuvette with a light path of 0.1 cm. The equilibration time at each temperature was ten seconds. Data collection and averaging time was ten seconds. Spectra with ellipiticity at each temperature point (step size 1 deg. C) were recorded on an AVIV 202 spectropolarimeter and data analysis was carried out using software Origin. Each spectrum was subtracted from the blank (buffer alone) and fitted with a two-state unfolding sigmoidal model. The inflection point was taken as the melting temperature T m, which reflected the temperature that caused the unfolding of 50% of the total protein sample. Deuterium exchange analysis Preparation of deuterated samples and subsequent DXMS analysis was carried out as previously described. 17 Briefly, the deuterium exchange reaction was started by combining 5 ml of protein sample (5 mg/ml, or mm, mutant or wild-type PKAc) with 15 ml of 20 mm Mops (ph 7.4), 50 mm NaCl, 1 mm dithiothreitol in 2 H 2 O (deuteration buffer). After incubation at room temperature for various times of 10, 30, 100, 300, 1000, or 3000 seconds, the reaction was quenched with 30 ml of 0.8% (v/v) formic acid, 1.6 M guanidine-hcl in 2 H 2 O, ph at 0 8C. Samples were immediately passed through a solid-phase pepsin column (66 ml bed volume, Upchurch

10 200 H/D Exchange Reveals PKAc Coupling Sites Scientific) and eluted with 0.05% trifluoroacetic acid (TFA) at 0.2 ml/minute for two minutes. Proteolytic peptides were collected by a C18 column (Vydac), which was subsequently developed with a linear gradient of 10 ml of 8% 40% (v/v) acetonitrile in 0.05% TFA, at 0.2 ml/minute. Mass spectrometric analysis was carried out with a Finnigan LCQ mass spectrometer as previously described. 17 Fully deuterated samples were prepared by incubating the protein in 0.5% formic acid in 95% 2 H 2 O for 24 hours. Longer deuteration up to 48 hours was carried out for the Y204A protein to define regions most resistant to deuterium exchange. The time-points were 100, 300, 1000, 3000 seconds, 12, 24 and 48 hours at room temperature. Recovered peptide identification and analysis were carried out using software specialized in processing DXMS data. 17 Binding kinetic analysis Binding experiments were performed on Biacore 3000 (Biacore Inc, Piscataway, NJ). Purified wild-type or the Y204A protein was immobilized onto Sensor Chip CM (Biacore Inc) through amine-coupling to a ligand density response of w300 RU. The R max for the analyte was RU. RIa-A-domain protein was serially diluted by a factor of 2 from 128 nm to 0.06 nm in 50 mm Mops (ph 7.0), plus 50 mm NaCl, % (v/v) P-20 (Biacore, Inc), 1 mm DTT, 1 mm ATP and 5 mm MgCl 2 (binding buffer), and flowed through the cell at a flow rate of ml/minute. The surface was regenerated by 100 mm camp and 2 mm EDTA in 50 mm Mops (ph 7.0), plus 50 mm NaCl. The experiments were performed in duplicates. The kinetic parameters for binding were determined from the analysis of sensorgrams using software Biaevaluation version 4.1 (Biacore, Inc). Each sensorgram curve was fit to a 1:1 binding model. Acknowledgements We thank Celina Juliano for some of the protein preparation, Chris Gessner (Department of Pharmacology, University of California San Diego) for technical support and Elzbieta Radzio-Andzelm for some of the Figure preparation. We thank Jaume Canaves for his help with the CD experiments, and Ganesh Iyer and Francis Kinderman for insightful discussion. This work is supported by grant GM19301 to S.S.T. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi: /j.jmb References 1. Knighton, D. R., Zheng, J. H., ten Eyck, L. F., Ashford, V. A., Xuong, N. H., Taylor, S. S. & Sowadski, J. M. (1991). Crystal structure of the catalytic subunit of cyclic adenosine monophosphate-dependent protein kinase. Science, 253, Johnson, D. A., Akamine, P., Radzio-Andzelm, E., Madhusudan, M. & Taylor, S. S. (2001). Dynamics of camp-dependent protein kinase. Chem. Rev. 101, Akamine, P., Madhusudan, M., Wu, J., Xuong, N. H., ten Eyck, L. F. & Taylor, S. S. (2003). Dynamic features of camp-dependent protein kinase revealed by apoenzyme crystal structure. J. Mol. Biol. 327, Narayana, N., Cox, S., Nguyen-huu, X., ten Eyck, L. F. & Taylor, S. S. (1997). A binary complex of the catalytic subunit of camp-dependent protein kinase and adenosine further defines conformational flexibility. Structure, 5, Zheng, J., Knighton, D. R., ten Eyck, L. F., Karlsson, R., Xuong, N.-h., Taylor, S. S. & Sowadski, J. M. (1993). Crystal structure of the catalytic subunit of campdependent protein kinase complexed with MgATP and peptide inhibitor. Biochemistry, 32, Madhusudan, Akamine, M., Xuong, N. H. & SS, T. (2002). Crystal structure of a transition state mimic of the catalytic subunit of camp-dependent protein kinase. Nature Struct. Biol. 9, Yang, J., ten Eyck, L. F., Xuong, N. H. & Taylor, S. S. (2004). Crystal structure of a camp-dependent protein kinase mutant at 1.26A: new insights into the catalytic mechanism. J. Mol. Biol. 336, Taylor, S. S., Yang, J., Wu, J., Haste, N. M., Radzio- Andzelm, E. & Anand, G. (2004). PKA: a portrait of protein kinase dynamics. Biochim. Biophys. Acta, 1697, Zhang, Z. & Smith, D. L. (1993). Determination of amide hydrogen exchange by mass spectrometry: a new tool for protein structure elucidation. Protein Sci. 2, Hoofnagle, A. N., Resing, K. A. & Ahn, N. G. (2003). Protein analysis by hydrogen exchange mass spectrometry. Annu. Rev. Biophys. Biomol. Struct. 32, Baerga-Ortiz, A., Hughes, C. A., Mandell, J. G. & Komives, E. A. (2002). Epitope mapping of a monoclonal antibody against human thrombin by H/D-exchange mass spectrometry reveals selection of a diverse sequence in a highly conserved protein. Protein Sci. 11, Woods, V. L., Jr & Hamuro, Y. (2001). High resolution, high-throughput amide deuterium exchange-mass spectrometry (DXMS) determination of protein binding site structure and dynamics: utility in pharmaceutical design. J. Cell. Biochem. Suppl. 37, Anand, G. S., Law, D., Mandell, J. G., Snead, A. N., Tsigelny, I., Taylor, S. S. et al. (2003). Identification of the protein kinase A regulatory RIalpha-catalytic subunit interface by amide H/2H exchange and protein docking. Proc. Natl Acad. Sci. USA, 100, Hamuro, Y., Anand, G. S., Kim, J. S., Juliano, C., Stranz, D. D., Taylor, S. S. & Woods, V. L., Jr (2004). Mapping intersubunit interactions of the regulatory subunit (RIalpha) in the type I holoenzyme of protein kinase a by amide hydrogen/deuterium exchange mass spectrometry (DXMS). J. Mol. Biol. 340, Anand, G. S., Hughes, C. A., Jones, J. M., Taylor, S. S. & Komives, E. A. (2002). Amide H/2H exchange reveals communication between the camp and catalytic subunit-binding sites in the R(I)alpha subunit of protein kinase A. J. Mol. Biol. 323, Hamuro, Y., Zawadzki, K. M., Kim, J. S., Stranz, D. D., Taylor, S. S. & Woods, V. L., Jr (2003). Dynamics of

11 H/D Exchange Reveals PKAc Coupling Sites 201 capk type IIbeta activation revealed by enhanced amide H/2H exchange mass spectrometry (DXMS). J. Mol. Biol. 327, Hamuro, Y., Burns, L., Canaves, J., Hoffman, R., Taylor, S. & Woods, V. (2002). Domain organization of D-AKAP2 revealed by enhanced deuterium exchange-mass spectrometry (DXMS). J. Mol. Biol. 321, Grant, B. D., Tsigelny, I., Adams, J. A. & Taylor, S. S. (1996). Examination of an active-site electrostatic node in the camp-dependent protein kinase catalytic subunit. Protein Sci. 5, Moore, M. J., Adams, J. A. & Taylor, S. S. (2003). Structural basis for peptide binding in protein kinase A. Role of glutamic acid 203 and tyrosine 204 in the peptide-positioning loop. J. Biol. Chem. 278, Tsigelny, I., Grant, B. D., Taylor, S. S. & ten Eyck, L. F. (1996). Catalytic subunit of camp-dependent protein kinase: electrostatic features and peptide recognition. Biopolymers, 39, Herberg, F. W., Doyle, M. L., Cox, S. & Taylor, S. S. (1999). Dissection of the nucleotide and metalphosphate binding sites in camp-dependent protein kinase. Biochemistry, 38, Andersen, M. D., Shaffer, J., Jennings, P. A. & Adams, J. A. (2001). Structural characterization of protein kinase A as a function of nucleotide binding. Hydrogen-deuterium exchange studies using matrix-assisted laser desorption ionization-time of flight mass spectrometry detection. J. Biol. Chem. 276, Kim, C., Xuong, N.-H. & Taylor, S. S. (2004). Crystal structure of a complex between the catalytic and regulatory (RI alpha) subunits of PKA. Science. In the press. 24. Lockless, S. W. & Ranganathan, R. (1999). Evolutionarily conserved pathways of energetic connectivity in protein families. Science, 286, Li, F., Gangal, M., Juliano, C., Gorfain, E., Taylor, S. S. & Johnson, D. A. (2002). Evidence for an internal entropy contribution to phosphoryl transfer: a study of domain closure, backbone flexibility, and the catalytic cycle of camp-dependent protein kinase. J. Mol. Biol. 315, Lee, T., Hoofnagle, A. N., Kabuyama, Y., Stroud, J., Min, X., Goldsmith, E. J. et al. (2004). Docking motif interactions in MAP kinases revealed by hydrogen exchange mass spectrometry. Mol. Cell. 14, Edited by P. Wright (Received 20 August 2004; received in revised form 11 November 2004; accepted 12 November 2004)