Article pubs.acs.org/JPCB
Structural Basis of Tonic Inhibition by Dimers of Dimers in Hyperpolarization-Activated Cyclic-Nucleotide-Modulated (HCN) Ion Channels Bryan VanSchouwen† and Giuseppe Melacini*,†,‡ †
Department of Chemistry and Chemical Biology, McMaster University, 1280 Main Street West, Hamilton, Ontario L8S 4M1, Canada ‡ Department of Biochemistry and Biomedical Sciences, McMaster University, Hamilton, Ontario L8S 4M1, Canada S Supporting Information *
ABSTRACT: The hyperpolarization-activated cyclic-nucleotide-modulated (HCN) ion channels control rhythmicity in neurons and cardiomyocytes. Cyclic AMP (cAMP) modulates HCN activity through the cAMP-dependent formation of a tetrameric gating ring spanning the intracellular region (IR) of HCN. In the absence of cAMP, the IR cAMP-binding domain (CBD) mainly samples its inactive conformation, resulting in steric clashes that destabilize the IR tetramer. Although these clashes with the inactive CBD are released through tetramer dissociation into monomers, functional mutagenesis suggests that the apo IR is not fully monomeric. To investigate the inhibitory nonmonomeric IR species, we performed molecular dynamics simulations starting from “hybrid” structures that are tetrameric but contain inactive apo-state CBD conformations. The ensemble of simulated trajectories reveals that full dissociation of the tetramer into monomers is not necessary to release the steric hindrance with the inactive CBD. Specifically, we found that partial dissociation of the tetramer into dimers is sufficient to accommodate four inactive CBDs, while reduction of the quaternary symmetry of the nondissociated tetramer from 4- to 2-fold permits accommodation of two inactive CBDs. Our findings not only rationalize available electrophysiological, fluorometry, and sedimentation equilibrium data, but also provide unprecedented structural insight into previously elusive nonmonomeric autoinhibitory HCN species.
■
(Figure 1c).18−37 Similar apo-state structures were subsequently reported for the HCN2 CBD, highlighting the relevance of such structural rearrangement for allostery of other HCN isoforms.38,39 On the basis of these structures, it was proposed that the apo-state CBD conformation destabilizes the tetramer through steric clashes with the TD, which arise as the result of an incompatibility of the apo-state CBD conformation with the tight steric packing imposed by tetramer assembly (Figure 1d).17 Stabilization of a tetramerization-competent CBD conformation upon cAMP binding removes these steric clashes, thus providing an explanation for how cAMP controls HCN channel gating via TD tetramerization.17,40 Alternatively, in the absence of cAMP, the CBD/TD steric clashes are removed through dissociation of the tetramer into monomers, which results in increased TD flexibility and provides an explanation for the ion channel autoinhibition imposed by the apo-state CBD.40 While the aforementioned model17,40 provides a viable explanation for cAMP-dependent modulation of HCN IR tetramerization, it is currently unclear whether CBD conformational shift or full dissociation into monomers are the only
INTRODUCTION The hyperpolarization-activated cyclic-nucleotide-modulated (HCN) proteins are cAMP-regulated ion channels that play a key role in nerve impulse transmission and heart rate modulation in neuronal and cardiac cells, respectively.1−10 These proteins contain an N-terminal transmembrane region (TM), which forms a tetrameric assembly harboring the ion pore, and a C-terminal intracellular region (IR) that confers regulation by cAMP (Figure 1a,b).1,2,5,7,9−13 It has been suggested that tetramerization of the HCN IR in response to cAMP binding is closely linked to the cAMP-dependent upregulation of ion channel opening.3,14−16 Specifically, the HCN IR tetramer adopts an “elbow−shoulder” topology3 involving intermonomer interactions of two α-helical hairpins (αA′−αB′ and αC′−αD′) at the IR N-terminus, which are referred to as the tetramerization domain (TD; Figure 1a,b). Meanwhile, cAMP-binding domains (CBD) C-terminal to the TDs of each monomer (Figure 1a,b) allosterically control TD selfassociation, whereby cAMP binding to the CBD promotes TD tetramerization.3 In a recent study performed by Akimoto et al. (2014),17 the apo-state structure of the HCN4 CBD was solved by NMR, and was found to differ from the structure of the cAMP-bound CBD5 by a rearrangement of the CBD α-helical subdomain similar to that observed for CBDs of other cAMP receptors © XXXX American Chemical Society
Received: August 1, 2016 Revised: September 30, 2016
A
DOI: 10.1021/acs.jpcb.6b07735 J. Phys. Chem. B XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry B
Figure 1. Overview of HCN4 tetramer architecture and the “hybrid” tetramer structures simulated in the current work. (a) Domain organization of the intracellular region (IR) of HCN4, highlighting the major structural components of the IR: the IR tetramerization domain (“TD”) spanning the N-terminal αA′-αD′ helices, and the IR cAMP-binding domain (“CBD”), the latter of which includes the N-terminal α-helix bundle (“N3A”), the βsubdomain (“β-core”), and the C-terminal αB and αC helices. (b) Ribbon-structure representation of the active tetrameric structure of the HCN4 IR bound to four molecules of cAMP (PDB code “3OTF”). The tetramer is assembled via interprotomer interactions among the N-terminal αA′−αD′ helices (“TD”; green ribbons), while the C-terminal CBD in each protomer (red ribbons) allosterically controls tetramer assembly in response to cAMP (shown as blue sticks). (c) Overlay of the active CBD structure (red ribbons) and a representative structure in the ensemble of inactive-state CBD conformations solved previously for apo HCN4 (PDB code “2MNG”; black ribbons),17 illustrating the conformational changes that occur within the CBD α-helical subdomain during cAMP-dependent activation. Upon cAMP binding, the αB−αC region shifts from an “out” position to an “in” position relative to the β-subdomain, while the N3A shifts from an “in” position to an “out” position. For clarity, the β-cores of both structures are shown in gray, and the bound cAMP and αA′-αD′ helices are omitted. (d) Ribbon-structure illustration of one of the “hybrid” tetramer structures utilized as starting conformations in the current simulation work. The structure shown consists of the TD region from the active tetrameric structure (green ribbons) and the CBD conformation derived from a representative structure of the inactive-state CBD conformational ensemble (black ribbons). The TD/CBD steric clashes within this structure17 are indicated by red dotted ellipses, and this hybrid structure is shown with the same TD-region orientation as the structure in panel b to facilitate comparison of the tetramers. (e) Log-scale distribution of αA′−αB′/βcore intraprotomer, αC′−αD′/β-core interprotomer, and αB−αC/β-core interprotomer steric-clash potential energies for the hybrid tetramer structures derived from the inactive-state CBD conformational ensemble.17 The points that correspond to structures selected for simulations (i.e., the 2nd, 6th, 7th, 9th, and 10th structures from the ensemble) are indicated by red circles. (f) Total van der Waals (VDW) potential energy for the hybrid tetramer derived from a representative ensemble structure, as observed over the course of the respective simulation. The initial potential energy is off-scale, as indicated by the arrowhead in the plot.
performed here started from “hybrid” tetramer structures that contained CBD conformations derived from the apo-state structural ensemble,17 in order to deliberately introduce the CBD/TD steric clashes that arise in the presence of these conformations, and probe the structural responses to these clashes. The simulations revealed that full dissociation of the tetramers into monomers is not necessary to release the CBD/TD steric clashes. Rather, inactive-to-active CBD conformational transitions in selected protomers or partial tetramer dissociation events, which were captured by our MD trajectories, provide alternative mechanisms sufficient to
possible mechanisms for releasing the steric clashes between the TD and the inactive CBD conformation of the apo state. Indeed, electrophysiology of tandem HCN mutants suggested that tonic channel inhibition in the absence of cAMP arises from HCN IR dimers.41 However, no structural insight is currently available on such self-associated forms of the apo HCN IR. In order to gain structural and dynamical insight into the apo HCN IR, we assessed through molecular dynamics (MD) simulations28,42−57 how the HCN4 IR tetramer responds to the presence of inactive CBD conformations (Figure 1d). Unlike previous MD analyses,40 the majority of the simulations B
DOI: 10.1021/acs.jpcb.6b07735 J. Phys. Chem. B XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry B
different regions of the intratetramer steric-clash potential energy distribution spanned by the structure ensemble (Figure 1e).17 In addition, an active-state apo/dimer structure was obtained from the active-state tetramer structure by removing all cAMP molecules and two of the four HCN4 protomers. MD Simulation Protocol. All MD simulations were performed using the NAMD 2.9 software on the Shared Hierarchical Academic Research Computing Network (SHARCNET), using a protocol similar to that implemented previously.40 As before, addition of hydrogen atoms was implemented in a manner consistent with an experimental pH of 7.0, such that all His side chains were in their un-ionized τstate and the N-/C-termini and Asp, Glu, Arg, and Lys side chains were in their ionized states. The protein was then placed in a cubic box of TIP3P water molecules with a minimum distance of 12 Å from the edge of the solvent box, and Na+ and Cl− ions were added to the solvent box to a NaCl concentration of 50 mM to achieve charge neutrality. After initial energy minimization, heating, and 1.0 ns NVEensemble and NPT-ensemble equilibrations were performed as described previously,40 the NPT-ensemble production-run simulations were executed for 25 ns for all tetramers, and for 120 ns in triplicate for the active-state apo/dimer, saving structures every 10 000 time steps (i.e., every 10.0 ps) for subsequent analysis. It should be noted that MD trajectories in the tens of nanoseconds, while often too brief to capture transitions initiating from equilibrium structures, are still valuable in capturing transitions starting from nonequilibrium conformations, as is the case for the tetramer simulations presented here. Indeed, the “hybrid” tetramer structures model an HCN4 structural state that is inherently far from equilibrium (due to its intrinsic CBD/TD steric clashes, as discussed above), and we were interested primarily in capturing the tetramers’ structural responses to the steric clashes. Probing such responses requires our simulations to relax the steric clashes in the hybrid tetramers, and such relaxation occurs already within the first nanosecond of our MD trajectories as shown in Figures 1f and S1. Furthermore, backbone N−H order parameters (S2) computed from 25 ns MD simulations of the cAMP-bound IR monomer40 capture the trends observed in the experimental S2 values (Figure S2),40 suggesting that key dynamic features can be successfully reproduced by relatively brief simulations, and thus lending credibility to the use of shorter simulations for the current work. Last but not least, running simulations in the 10−100 ns range allowed us to record multiple trajectories with different initial conditions, an approach known to enhance conformational sampling.58−62 Analysis of Structural Dynamics. Tetramer Backbone Root-Mean-Square Deviations. As an initial assessment of structural propensities within the simulations, root-meansquare deviations (RMSDs) from the initial “hybrid” tetramer structures were computed for the TD and CBD regions of each tetramer, overlaying all four protomers during the calculation. In addition, RMSDs from the active-state tetramer structure were computed for the CBD region of each tetramer (again overlaying all four protomers) in order to assess the extent of tendency toward the active-state structure that was exhibited by this region during the simulations. In all calculations, the Cterminal αC helices (which were truncated in the “2MNG”derived CBD structures) were ignored in order to avoid calculation bias resulting from artificial flexibility in the truncated C-termini, and because this region was previously
alleviate the CBD/TD steric clashes. In particular, the MD simulations reveal the structural and dynamical basis for how a subset of two inactive CBDs is compatible with a tetrameric assembly in which the TD symmetry is reduced from 4- to 2fold, and for how four inactive CBDs are accommodated not only by fully dissociated IR monomers but also by IR dimers of dimers. In addition, the stability of dimers with two inactive CBDs was independently confirmed by control simulations starting from an initial structure different from any of the “hybrid” tetramers, i.e., an apo dimer structure with two active CBDs. These >100 ns long trajectories run in triplicate support the notion that both CBDs are able to simultaneously transition to an inactive conformation without requiring dissociation into monomers. Overall, the MD simulations provide a more comprehensive picture of autoinhibited HCN than was previously available, and offer a structural rationalization of electrophysiological, fluorometry, and sedimentation equilibrium data on HCN autoinhibition and cAMP-dependent activation.
■
METHODS Overview. Molecular dynamics (MD) simulations in explicit solvent were performed for five representative “hybrid” HCN4 IR tetramers,17,40 which were built starting from the active-state tetramer structure (Figure 1b) by substituting the active CBD regions with inactive CBD conformations (Figure 1c) obtained from selected frames in the previously solved apo HCN4 CBD structural ensemble (Figure 1d).17 These “hybrid” HCN4 IR constructs were designed to simulate a situation in which inactive-state CBD conformations are placed into the context of an HCN4 IR tetramer, thus giving rise to CBD/TD steric clashes.17 The resulting simulations were analyzed through RMSDs, RMSD-based similarity measures, RMSFs, and inter-center-of-mass distances in order to assess the potential structural consequences of the steric clashes. Details of the setup, execution, and analysis of the MD simulations are described below. Molecular Dynamics Simulation Protocol. Initial Structure Preparation. An HCN4 construct composed of the TD and CBD regions of human HCN4 was used for the MD simulations. The active-state tetramer structure was obtained from the X-ray crystal structure of the cAMP-bound intracellular region of human HCN4 (PDB code “3OTF”; see Figure 1b) as described previously.40 Initial structures for the “hybrid” apo/tetramers (Figure 1d) were then obtained by grafting the CBD-region structure from each selected frame in the previously solved apo HCN4 CBD structure ensemble (PDB code “2MNG”)17 onto the C-linker of the “3OTF”derived protomer structure via a backbone superimposition at the αA helix, saving the “2MNG”-derived αA−αC atomic coordinates and “3OTF”-derived αA′−αF′ atomic coordinates as a new structure. While structural grafts via a full-N3A backbone superimposition were also attempted, joining the “2MNG”- and “3OTF”-derived structures together at the αE′ or αF′ helix (rather than at the αA helix) resulted in covalent backbone distortion due to a lesser degree of backbone alignment at the αE′ and αF′ helices, and in either superimposition, the greatest degree of backbone alignment still occurred at the αA helix. The four αA-grafted protomers were then assembled into a tetramer by superimposing the TDregion backbone onto the active-state tetramer structure. This process was performed for five frames from the apo HCN4 CBD structure ensemble, which were selected to represent C
DOI: 10.1021/acs.jpcb.6b07735 J. Phys. Chem. B XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry B found to be controlled predominantly by cAMP binding40 which was absent here. Intraprotomer Backbone Root-Mean-Square Deviations. As an assessment of protomer structural propensities within the simulations, RMSDs from the initial protomer structures were computed for the TD and N3A regions of the constituent protomers (Figure 1a,c) in the last 20 ns of each simulation. The RMSDs were computed via a protomer overlay onto the reference structure at the β-core (Figure 1a,c) in order to permit examination of the movement of these regions relative to the central scaffold of the CBD (i.e., the β-core; Figure 1a,c). In addition, TD and N3A RMSDs from the “3OTF”-derived protomer structure were computed in order to assess the tendency toward the active-state structure that was exhibited by these regions during the simulations. RMSD-Based Structural Distributions To Probe Inactive/ Active Transition Propensities in the CBDs. To further assess CBD structural propensities within the simulations, RMSDbased active-versus-inactive structure similarity measures were computed for the constituent protomers within each simulation. First, the intraprotomer RMSDs of the N3A and αB−αC regions (Figure 1a,c) from both the “3OTF”-derived protomer structure (representing the active conformation) and the average structure derived from the “2MNG” NMR structural ensemble (representing the inactive conformation) were computed as described above. Then, the RMSD-based similarity measures for the N3A and αB−αC regions in any given structure (i.e., “SMN3A” and “SMαB−αC”) were computed from the RMSD values for each constituent protomer as described previously:40 SM =
RMSDfrom inactive − RMSDfrom active RMSDinactive vs active
Distances. To further assess interprotomer movements of the TD and CBD regions during the simulations, time profiles of inter-center-of-mass (inter-CM) distances between the constituent protomers were computed for the TD and CBD regions over the course of each simulation. In this analysis, the Cartesian coordinates of the centers of mass for the TD or CBD of each protomer were calculated at each time point as follows: N
rCOM =
∑i = 1 (mi)(ri) N
∑i = 1 mi
(2)
where “rCOM” denotes the Cartesian coordinates of the center of mass, “mi” denotes the mass of atom “i”, “ri” denotes the Cartesian coordinates of atom “i”, and the summation is calculated over all atoms within the protomer TD or CBD (denoted as “N”). From these Cartesian coordinates, the interprotomer distances between the calculated centers of mass were then computed. As in the RMSD calculations, the Cterminal αC helix (which is truncated in the “2MNG”-derived CBD structures) was ignored in order to avoid calculation bias resulting from artificial floppiness in the truncated C-terminus, and because this C-terminal region is controlled primarily by cAMP which was absent here.40 Assessment of Protomer Structural Propensities in the Active-State Apo/Dimer Simulations. To confirm that inactive-state-like CBD conformations were being explored within the active-state apo/dimer simulations, conformations of each constituent protomer exhibiting minimal N3A RMSDs (i.e., 2 Å (i.e., larger than the TD RMSDs observed for hybrid tetramers #2, #9, and #10; Figure 2b−d). Therefore, the CBD/TD steric clashes within hybrid tetramers #2, #9, and #10 (collectively referred to herein as tetramer “group I”) appear to drive their overall structural arrangements toward the active-state tetrameric structure, while the steric clashes within hybrid tetramers #6 and #7 (collectively referred to herein as tetramer “group II”) push their overall structural arrangements further away from the active-state structure. The differential behavior of groups I and II warranted a closer examination of the TD and CBD structural propensities within each of the two groups, as discussed in the following sections. F
DOI: 10.1021/acs.jpcb.6b07735 J. Phys. Chem. B XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry B
Figure 4. Inter-center-of-mass distance trajectories for the CBD portion of the tetramer, as observed for each of the simulated hybrid tetramers. (a− f) Distance trajectories computed for the inter-center-of-mass distances between the constituent protomers of each tetramer: (a) cAMP-bound active-state tetramer,40 used here as a control; (b−f) hybrid tetramers #2, #9, #10, #6, and #7, respectively. As in Figure 2, groups of hybrid tetramers that exhibit similar types of structural propensities (i.e., groups I and II) are indicated. An outline of the computed inter-center-of-mass distances is shown in panel g, where the CBD of each protomer is represented as a numbered circle, and the distances between protomers are shown as colorcoded arrows. (h) Schematic outline of notable interprotomer movements detected for the group I tetramers, as exemplified by hybrid tetramers #9 and #10. (i) Schematic outline of notable interprotomer movements observed for hybrid tetramer #7.
for hybrid tetramer #7 was not detected for hybrid tetramer #6 though, as TD-region dissociation for just a single protomer pair (Figure 3e) suggested only a partial disassembly of this tetramer. Notably, the extent of tetramer-to-dimers shift observed for hybrid tetramer #7 relative to hybrid tetramer #6 appears to qualitatively correlate with the tetramer’s high degree of overall shift away from the active-state tetramer structure (as quantified by tetramer RMSDs; Figure 2e,f) and with the large CBDregion inter-CM distances for the dissociating protomer pairs (Figure 4e,f). In addition, the CBD-region RMSDs for the dimers of hybrid tetramer #7 (i.e., the protomer 1/protomer 4 and protomer 2/protomer 3 dimers outlined in Figure 4i) revealed that the individual dimers equilibrated to RMSD values comparable to those of the group I tetramers (blue and orange plots in Figure 2f vs black plots in Figure 2b−d), even as the tetramer RMSDs increase toward larger values (black and red plots in Figure 2f), further suggesting the occurrence of interdimer movement. Therefore, the structural shift of hybrid tetramer #6 may represent a dissociation intermediate exhibiting partial dimer dissociation, while the structural shift of hybrid tetramer #7 is more illustrative of full dimer dissociation. Overall, the group II MD trajectories point to partial or full tetramer dissociation into dimers as being an
represent a potential means of relieving CBD/TD steric clashes within the tetramers. Interestingly, the propensity for diagonal distortion to lower symmetry was also observed for the activestate tetramer (Figures 3a and 4a), suggesting that such diagonal distortions are an intrinsic property of the tetrameric state. In marked contrast to group I, the tetramers of group II exhibited a propensity for disruption of the tetrameric state of the protein, as suggested by a visible deviation of the TD interCM distances between adjacent protomers from those exhibited by the active-state tetramer (Figure 3a,e,f). For example, for hybrid tetramer #7, the TD inter-CM distances for two adjacent pairs of protomers demonstrate greater values than those observed for the active-state tetramer (black and blue plots in Figure 3a,e,f,i), suggesting a tendency for dissociation of the tetramers into dimers (Figure 3i). A propensity for protomer-pair dissociation was also observed for the CBD regions of the tetramers (red and purple plots in Figure 4e,f,i), although the CBD dissociation appeared to occur in a direction orthogonal to that of the TD dissociation: i.e., the CBDs of protomers 2 and 3 drifted away from those of protomers 1 and 4 (Figure 4i), while the TD regions of protomers 1 and 3 moved away from those of protomers 2 and 4 (Figure 3i). The full tetramer-to-dimer dissociation observed G
DOI: 10.1021/acs.jpcb.6b07735 J. Phys. Chem. B XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry B
Figure 5. Assessment of conformational propensities within the CBD regions of the constituent protomers from the simulated hybrid tetramers. (a) Box plots of backbone root-mean-square deviations (RMSDs) of the N3A region from its position in the active-state structure (PDB code 3OTF), as observed for the constituent protomers of each tetramer. The box plots were constructed using Origin 9.1 graphing software (OriginLab Corporation) based on the RMSD data for all four protomers of each tetramer, and all RMSDs were computed with the structures overlaid at the βcore. The RMSDs between the respective initial hybrid structures and the active-state structure are indicated by dotted horizontal lines, and as in Figure 2, groups of hybrid tetramers that exhibit similar types of structural propensities (i.e., groups I and II) are labeled. (b−e) Two-dimensional plots of the computed active vs inactive structure similarity measure (SM) distributions for the N3A and αB−αC regions, as observed for the constituent protomers of all hybrid tetramers belonging to each of the aforementioned groups: (b, c) group I tetramers; (d, e) group II tetramers. The data for all tetramers in each group are plotted together, using the following color code for the constituent protomers of each tetramer: red points = protomer 1 (“H P1”); green points = protomer 2 (“H P2”); blue points = protomer 3 (“H P3”); orange points = protomer 4 (“H P4”). For comparison, the previously computed SM distribution for the active-state apo/tetramer40 is also plotted (brown points), and the quadrants that represent inactive (“In”) and active (“Act”) CBD conformations are indicated.40 The left and right plots display the same data, but with a different order of front vs back layers, to highlight the extent of overlap (or lack thereof) among the plotted SM distributions. (f) Outline of the relative positions of the four protomer CBDs (represented as numbered, color-coded circles) within the active tetramer. (g, h) Schematic outline of the prevailing CBD conformational propensities observed for the (g) group I tetramers and (h) group II tetramers, as identified based on N3A structural propensities. Inactive CBD conformations as gauged based on the N3A (“In”) are indicated by bolded outlines, while active CBD conformations as gauged based on the N3A (“Act”) are denoted by color-filled shapes.
while the tetramers of group II did not exhibit this overall structural shift (Figure 5a). This group I vs group II difference in structural behavior was further highlighted by RMSD-based active-vs-inactive structure similarity measures (SMs; Figure 5b−e).40 Specifically, the group I tetramers were subject to inactive-to-active CBD conformational shifts, as evidenced by SM values that significantly overlap the previously computed40 distribution of values spanned by the active-state apo tetramer (Figure 5b,c). Interestingly, the group I tetramers did not consistently exhibit a shift of all four protomer CBDs toward the active-state structure, but rather also explored structural states in which two of the four protomer CBDs retained an inactive conformation (Figure 5b,c,g; red and orange plots versus green and blue plots). Therefore, the tetramer distortion exhibited by the group I tetramers (Figures 3b−d,h and 4b−d,h) appears to permit a tolerance of inactive CBD conformations in one pair
alternative means of relieving CBD/TD steric clashes within the tetramers. However, the interprotomer movements examined so far do not report on possible changes in structure and dynamics within the constituent protomers. Thus, in the following sections, the intraprotomer CBD and TD structural dynamics are assessed. Intraprotomer CBD Structural Dynamics. As an initial assessment of intraprotomer CBD structural dynamics within the two hybrid tetramer groups, RMSDs of the N-terminal αhelical structural elements of the CBD (i.e., the N3A region; Figure 1c) from their active-state positions relative to the CBD β-subdomain were analyzed as a proxy for inactive/active conformational transitions (Figure 1c).40 From this analysis (Figure 5a), it was found that the tetramers of group I exhibited a shift of their protomer CBDs closer to the active-state CBD structure, as evidenced by an overall drop in the N3A-region RMSDs compared to the respective initial hybrid structures, H
DOI: 10.1021/acs.jpcb.6b07735 J. Phys. Chem. B XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry B
Figure 6. Box plots of backbone root-mean-square deviations (RMSDs) of the TD region from its position in the active-state structure, as observed for each of the simulated hybrid tetramers. (a) Overall RMSD distributions observed for each hybrid tetramer. As in Figure 2, groups of hybrid tetramers that exhibit similar types of structural propensities (i.e., groups I and II) are indicated. (b−f) RMSDs observed for the individual protomers of each hybrid tetramer (color-coded as in Figure 5b−e). All box plots were constructed using Origin 9.1 graphing software (OriginLab Corporation), and all RMSDs were computed with the structures overlaid at the β-core. For reference, the RMSDs between the respective initial hybrid structures and the active-state structure are indicated by dotted horizontal lines in all plots. (b, Inset) Schematic outline of the “elbow− shoulder” topology3 of the four protomer TD regions (represented as numbered, color-coded αA′−αD′ segments) within the active tetramer.
protomers can explore inactive CBD structures comparable to the apo/monomer40 (Figures S3a−d and S4a−d) and can do so simultaneously (Figures S3e and S4e). Furthermore, dimer structure visualization (Figure S4e) and TD inter-CM distances comparable to those exhibited by the active-state tetramer (Figures S4f and 3a) confirmed maintenance of a dimeric state. These observations further support the notion that HCN4 IR dimers are compatible with the inactive CBD conformations, and suggest that the inactive-CBD retention in group II is made possible by CBD/TD steric clash relief due to the tetramer-todimers shift exhibited by the group II tetramers (Figure 5h). Intraprotomer TD Structural Dynamics. As an initial assessment of intraprotomer TD structural dynamics within the two hybrid tetramer groups, RMSDs of the TD region from its active-state position relative to the CBD β-subdomain (Figure 1b) were examined (Figure 6). From this analysis, it was found that the group I tetramers exhibited a shift of the protomer TD regions closer to their active-state positions, as evidenced by an overall drop in the TD-region RMSDs compared to the respective initial hybrid structures (Figure 6a−d). This structural shift of the group I tetramers is in agreement with the inter-CM distance results (Figures 3b−d,h and 4b−d,h) and confirms the maintenance of the tetrameric state of the
of adjacent protomers without requiring tetramer dissociation (Figure 5g). Unlike group I, the group II tetramers retained predominantly inactive CBD conformations, as evidenced by protomer SM values primarily within the SM plot quadrant corresponding to inactive CBD conformations,40 and by little overlap with the active-state distribution (Figure 5d,e). The confinement of the group II tetramers to inactive CBD conformations is not likely due to the short simulation time frame though, since this time frame was sufficient to capture inactive-to-active CBD conformational transitions for the group I tetramers. In addition, as a further control, we implemented longer MD simulations (120 ns in length) acquired in triplicate starting from a completely different initial structure, i.e., an apo/dimer with both CBDs in the active conformation (derived from the active tetramer structure solved by X-ray crystallography, as explained in the Methods section).3 The 120 ns simulations confirmed that both protomer CBDs within the dimer are able to transition to an inactive conformation without requiring dimer dissociation into monomers. Specifically, similarity measures (Figure S3) and visualization of inactive-like CBD structures identified based on N3A-region RMSDs40 (Figure S4a−e) suggested that both I
DOI: 10.1021/acs.jpcb.6b07735 J. Phys. Chem. B XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry B
Figure 7. Residue-specific root-mean-square fluctuations (RMSFs) for the tetramerization domain (TD), computed with structure overlay at the βcore, for each of the simulated hybrid tetramers. (a−e) Results for the constituent protomers of hybrid tetramers (a) #2, (b) #9, (c) #10, (d) #6, and (e) #7, color-coded as in Figures 5b−e and 6b−f. The constituent α-helical secondary structure elements are indicated along the horizontal dimension of each graph (black bars), and for comparison, the previously computed protomer-averaged RMSFs for the active-state apo/monomer (brown plots) and apo/tetramer (black plots)40 are also shown. (f) Schematic outline of the relative positions of the four protomer TD regions (represented as numbered, color-coded boxes) within the active tetramer, indicating the α-helical secondary structure elements involved in the interactions between protomers (i.e., the αA′−αB′ and αC′−αD′ elements; see Figure 1a).
To further examine structural fluctuations in the TD region of each protomer, backbone root-mean-square fluctuations (RMSFs) and N−H order parameters (S2) were computed on a residue-specific basis (Figures 7 and 8). In agreement with the TD RMSD results (Figure 6a−d), the protomer TD regions of the group I tetramers demonstrated minimal fluctuations in their positions relative to the β-core that were comparable to those of the active-state apo tetramer (Figure 7a−c). In contrast, the group II tetramers demonstrated a tendency for enhanced fluctuations in the αA′−αB′ region of the TD, whereby the most dynamic protomers in each tetramer are the same protomers that demonstrated the greatest TD-region RMSDs (blue and orange plots in Figures 6f and 7e; green plots in Figures 6e and 7d). However, localized backbone dynamics were consistently found to be low in the αA′−αB′ region of the TD, and greater in the αC′−αD′ region of the TD, as reflected by higher and lower S2 values, respectively (Figure 8). Therefore, the enhanced TD-region RMSDs and αA′−αB′ fluctuations in the group II tetramers (Figures 6 and 7) are due to movement of the αA′−αB′ regions relative to their respective CBDs (or vice versa), with the αC′−αD′
protein. In contrast, the group II tetramers exhibited a greater tendency for shift of the protomer TD regions away from their active-state positions, as evidenced by increases in TD-region RMSDs compared to the respective initial hybrid structures (Figure 6a,e,f). Notably, the two protomer TD regions of hybrid tetramer #7 that underwent increases in RMSDs (blue and orange plots in Figure 6f) are diagonally opposite one another within the tetramer (blue and orange protomers in Figure 6b inset), highlighting the 2-fold dimer symmetry that is also reflected by the inter-CM distance results (black, red, blue, and purple plots in Figures 3f,i and 4f,i). Furthermore, the TD region of the hybrid tetramer #6 protomer that underwent a larger RMSD increase than the other three protomers (green plot in Figure 6e) spans the interface between the two protomer TD regions that demonstrated propensity for dissociation from one another in hybrid tetramer #7 but not in hybrid tetramer #6 (blue arrows in Figure 3i; green and blue protomers in Figure 6b inset), suggesting that a hinge-like movement of hybrid tetramer #6 occurs across this interface as the other interprotomer interface dissociates (black arrows in Figure 3i; red and orange protomers in Figure 6b inset). J
DOI: 10.1021/acs.jpcb.6b07735 J. Phys. Chem. B XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry B
Figure 8. Residue-specific backbone N−H order parameters (S2) for the tetramerization domain (TD), as computed for each of the simulated hybrid tetramers. (a−e) Results for the constituent protomers of hybrid tetramers (a) #2, (b) #9, (c) #10, (d) #6, and (e) #7, color-coded as in Figures 5b− e and 6b−f. The constituent α-helical secondary structure elements are indicated as in Figure 7, and for comparison, the previously computed protomer-averaged order parameters for the active-state apo/monomer (brown plots) and apo/tetramer (black plots)40 are also shown. The region of the TD that is prone to greater localized backbone dynamics (as reflected by lower S2 values) is indicated by red highlights. (f) Outline of the position within the TD of the region prone to greater dynamics (highlighted in red), illustrated within the protomer structure from one of the hybrid tetramers. The CBD and the α-helical secondary structure elements of the TD (Figure 1a) are labeled, and the presence of movement of the αA′−αB′ region and CBD relative to one another, as observed for hybrid tetramers #6 and #7, is denoted by red arrows.
regions serving as intrinsically dynamic hinges across which the relative movements can occur (Figure 8f).
hindrance (Figure 9d,e) and provide a comprehensive picture of the autoinhibitory HCN IR ensemble. One group of hybrid tetramer simulations (i.e., group II) revealed that a full set of four inactive CBDs is accommodated without steric clashes when the tetramer dissociates into dimers (Figure 9e). A particularly prominent tetramer-to-dimers dissociation was observed for hybrid tetramer #7 (Figures 3f,i and 4f,i), demonstrating that as the protomer pairs separate, the dissociating αC′−αD′ regions become flexible hinge regions (blue and orange protomers in Figure 9e) that permit release of the steric clashes with the CBDs of the same protomers via movement of these CBDs away from the still-associated TDregion dimers (blue/red and orange/green protomer pairs in Figure 9e), thus explaining the separation of CBD pairs (Figure 4f,i) and the αA′−αB′/CBD relative movements (blue and orange plots in Figures 6f and 7e) observed for this tetramer. Furthermore, the dissociating αA′−αB′ regions (red and green protomers in Figure 9e) move relative to the respective CBDs (red and green plots in Figure 6f) due to increased conformational freedom, which together with dissociation of
■
DISCUSSION In our previous work, we determined that autoinhibitory CBD/ TD steric clashes arise within the HCN IR tetramer when the CBD is in an inactive conformation (Figure 9a). We also determined that such clashes are released either by a shift of the CBD to a tetramerization-competent active conformation (Figure 9c), which is stabilized upon cAMP binding, or by dissociation of the tetramer into dynamic monomers that retain the inactive CBD (Figure 9b), which prevails in the absence of cAMP.17,40 Here, our simulations of hybrid tetramers containing inactive CBD conformations show that full dissociation of tetramers into monomers is not the only viable means for releasing the steric clashes arising from the inactive CBD. In particular, the five hybrid tetramers analyzed clustered into two distinct consensus groups, which highlight possible nonmonomeric apo IR structural assemblies devoid of steric K
DOI: 10.1021/acs.jpcb.6b07735 J. Phys. Chem. B XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry B
Figure 9. Schematic summary of possible structural transitions to relieve the previously identified CBD/TD steric clashes that arise for inactive CBD conformations in the context of the HCN IR tetramer. (a) Outline of the initial hybrid tetrameric structure, with CBD/TD steric clashes indicated by gray starbursts. (b, c) Previously proposed structural transitions to relieve the steric clashes:17,40 (b) dissociation into monomers; (c) shift of the CBDs to their active conformational state, with preservation of a 4-fold symmetric tetramer. (d, e) Newly identified structural transitions that also relieve the steric clashes. (d) Possible transition identified from the group I hybrid tetramers, which involves a mix of CBD conformational transitions and/or tetramer distortion, and preserves the tetrameric state but with a symmetry reduction from 4- to 2-fold. A possible subsequent transition is also indicated (dashed arrow). (e) Possible structural transition identified from the group II hybrid tetramers, involving tetramer dissociation into dimers (which may subsequently dissociate into monomers; dashed arrow). All protomers are color-coded as in Figures 5−8, with the αA′−αB′ and αC′−αD′ segments of the protomer TD regions (see Figure 1a) indicated as long and short rectangles, respectively, and the CBDs as circles and squares. Less ordered αA′−αB′ and/or αC′−αD′ segments in panels b and e are shown as labeled, curved lines, and the central axes of symmetry for the tetrameric and dimeric structures (a, c-e) are indicated by “+” symbols. In all panels, inactive CBD conformations (“In”) are indicated by bolded outlines and active CBD conformations (“Act”) by color-filled shapes.
the adjoining αC′−αD′ regions permits release of all intra- and interprotomer steric clashes with these CBDs (Figure 9e). Meanwhile, hybrid tetramer #6 demonstrated a propensity for dissociation of only one of the two TD-region interfaces between protomer pairs (black plots in Figure 3e,i; red and orange protomers in Figure 9e), with a concurrent hinging movement of the αA′−αB′ region that spans the other TDregion interface between protomer pairs (green protomer in Figure 9e; green plots in Figures 6e and 7d). This partially dissociated tetramer illustrates a possible intermediate in the tetramer-to-dimer dissociation pathway, in which the tetramer begins to open without full dissociation into dimers. The notion emerging from the group II simulations that HCN IR dimers are able to accommodate a full set of inactive CBDs provides a structural rationalization as to why populations of dimers were previously observed based on sedimentation equilibrium data for the isolated apo HCN IR,3,16,67 and functional mutagenesis data for integral HCN ion channels in the absence of cAMP.41 Indeed, comparative analysis of electrophysiological data for tandem HCN dimers and tetramers demonstrated that tonic inhibition of gating in the absence of cAMP is best explained by an activation pathway
that proceeds via association of dimers.41 Therefore, the group II simulations provide an unprecedented atomistic glimpse into the nature of these functional inhibitory dimers. Furthermore, the notion of inactive HCN IR dimers was also independently confirmed by a set of three longer control MD simulations starting from an apo/dimer structure with both CBDs in the active conformation (Figures S3 and S4). These trajectories demonstrated that the CBDs of both protomers within the dimer are able to simultaneously transition to an inactive conformation without requiring further dissociation into monomers (Figures S3 and S4). The other group of simulations (i.e., group I) revealed that inactive CBD conformers are accommodated in a subset of protomers without dissociation (Figure 9d) by distorting the tetramer along its diagonal dimensions, thus reducing the tetramer symmetry from 4- to 2-fold (Figure 9d). Specifically, the symmetry-reducing diagonal distortions permit accommodation of structural states in which only a pair of adjacent protomer CBDs shifts toward the active-state structure, while the other two CBDs retain inactive conformations (Figure 5b,c,g). L
DOI: 10.1021/acs.jpcb.6b07735 J. Phys. Chem. B XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry B
symmetry, providing a structural rationalization of previous equilibrium sedimentation, fluorometry, and electrophysiological data, and suggesting that modulation of HCN activity by the IR proceeds with a dimeric organization.
Considering that the inactive and active CBD states are the main conformers sampled by the apo and cAMP-bound CBDs, respectively, the ability of low-symmetry tetramers to accommodate a mixture of active and inactive CBDs suggests that diagonally distorted tetramers (Figure 9d) may serve as putative intermediates populated in the presence of substoichiometric amounts of cAMP along the pathway of formation of the fully active tetramers (Figure 9c). In this context, the diagonal movement facilitates tetramer formation by permitting accommodation of inactive CBD conformations in a subset of protomers prior to final completion of the inactive-to-active CBD conformational shift for all four protomers (Figure 9a,c,d). Although the results outlined here were obtained through simulations that started from hybrid-tetramer structures of the HCN4 isoform, the generality of the proposed model (Figure 9d) is supported by the observation that similar diagonal distortions were also detected for the non-hybrid active-state tetramer of HCN4 (Figures 3a and 4a), as well as from MD simulations and normal-mode analyses of the HCN2 activestate tetramer.68,69 Thus, such symmetry-lowering distortions are an intrinsic property of the tetrameric IR that is shared across different HCN isoforms. The 2-fold symmetry of the partially active distorted tetramers (Figure 9d) provides a structural interpretation of previous experimental studies on intact HCN ion channels.41,70,71 These studies demonstrate that activation proceeds via a pair of allosterically coupled dimers, whereby pairs of protomers undergo inactive-to-active shifts in a highly cooperative manner. In particular, electrophysiology experiments for HCN2 binding-site mutants have indicated that as the number of cAMP-bound protomers is varied, the voltage required for half-maximal channel opening follows a trend that closely matches the pattern expected for activation via coupled dimers, but deviates from the trends expected for concerted changes in all four protomers, as posited by the Monod− Wyman−Changeux (MWC) model, or for four independent promoter transitions, as hypothesized by the Hodgkin−Huxley (HH) model.41 Additionally, functional fluorometry experiments have shown that HCN channels tend to be more thermodynamically stable with zero, two, or four bound cAMP ligands than with one or three bound ligands, further confirming the occurrence of pairwise cooperativity in cAMPassociated HCN activation.70 Therefore, concerted changes in all four protomers are a less likely pathway for activation/ inactivation than are the transitions via dimers and distorted tetramers summarized in Figure 9d,e,41,70 suggesting that these structural assemblies represent key intermediates in the activation/inactivation of intact HCN channels. In conclusion, the hybrid tetramer simulations presented here have provided unprecedented insight into the structural ensemble sampled by the apo HCN IR, in which the inactive CBD conformation prevails. We have shown that full dissociation of the tetramer into monomers is not the only means to release the steric clashes that arise when the inactive CBD is confined to a tetramer with 4-fold symmetry. Lowering the tetramer symmetry to 2-fold via diagonal distortions without dissociation or opening of the tetramer is sufficient to accommodate two adjacent inactive CBDs, while dissociation into dimers permits accommodation of a full set of four inactive CBDs (Figure 9d,e). Notably, the observed patterns of tetramer dissociation (Figure 9e) and inactive-to-active CBD conformational transition (Figure 9d) both exhibit a pair-of-dimers
■
ASSOCIATED CONTENT
S Supporting Information *
The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jpcb.6b07735. Figure S1 showing total van der Waals (VDW) potential energies observed for each of the simulated hybrid tetramers over the course of the respective simulations; Figure S2 showing comparison of calculated vs experimental order parameters; Figure S3 showing computed active vs inactive structure similarity measures (SM) as observed in the context of simulations of the active-state apo/dimer structures of the HCN4 IR; Figure S4 showing sampling of inactive-state-like CBD topologies within the active-state apo/dimer HCN4 IR simulations; Figure S5 showing root-mean-square fluctuations (RMSFs) for all amino acid residues, computed with structure overlay at the β-core, for each of the simulated hybrid tetramers; Figure S6 showing backbone N−H order parameters (S2) for all amino acid residues, as computed for each of the simulated hybrid tetramers (PDF)
■
AUTHOR INFORMATION
Corresponding Author
*Telephone: (905) 525-9140, extension 26959. Fax: (905) 522-2509. E-mail:
[email protected]. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS We thank Professors W. N. Zagotta (University of Washington), E. Accili (The University of British Columbia), A. Moroni (University of Milan), and S. S. Taylor (University of California, San Diego), M. Akimoto (Keio University), and S. Boulton and K. Moleschi (McMaster University) for helpful discussions. This study received funding from Canadian Institutes of Health Research (Grant MOP-68897) to G.M., Natural Sciences and Engineering Research Council of Canada (Grant RGPIN-2014−04514) to G.M.
■
REFERENCES
(1) Biel, M.; Wahl-Schott, C.; Michalakis, S.; Zong, X. Hyperpolarization-activated cation channels: from genes to function. Physiol. Rev. 2009, 89, 847−885. (2) Craven, K. B.; Zagotta, W. N. CNG and HCN channels: two peas, one pod. Annu. Rev. Physiol. 2006, 68, 375−401. (3) Zagotta, W. N.; Olivier, N. B.; Black, K. D.; Young, E. C.; Olson, R.; Gouaux, E. Structural basis for modulation and agonist specificity of HCN pacemaker channels. Nature 2003, 425, 200−205. (4) Kusch, J.; Biskup, C.; Thon, S.; Schulz, E.; Nache, V.; Zimmer, T.; Schwede, F.; Benndorf, K. Interdependence of receptor activation and ligand binding in HCN2 pacemaker channels. Neuron 2010, 67, 75− 85. (5) Xu, X.; Vysotskaya, Z. V.; Liu, Q.; Zhou, L. Structural basis for the cAMP-dependent gating in the human HCN4 channel. J. Biol. Chem. 2010, 285, 37082−37091. (6) Wang, J.; Chen, S.; Nolan, M. F.; Siegelbaum, S. A. Activitydependent regulation of HCN pacemaker channels by cyclic AMP:
M
DOI: 10.1021/acs.jpcb.6b07735 J. Phys. Chem. B XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry B signaling through dynamic allosteric coupling. Neuron 2002, 36, 451− 461. (7) Bois, P.; Chatelier, A.; Bescond, J.; Faivre, J. F. Pharmacology of hyperpolarization-activated cyclic nucleotide-gated (HCN) channels. In Ion Channels and Their Inhibitors; Gupta, S. P., Ed.; Springer-Verlag: Berlin, 2011; pp 33−51. (8) Rosenbaum, T.; Gordon, S. E. Quickening the pace: looking into the heart of HCN channels. Neuron 2004, 42, 193−196. (9) Johnson, J. P.; Zagotta, W. N. The carboxyl-terminal region of cyclic nucleotide-modulated channels is a gating ring, not a permeation path. Proc. Natl. Acad. Sci. U. S. A. 2005, 102, 2742−2747. (10) Biel, M.; Schneider, A.; Wahl, C. Cardiac HCN channels: structure, function, and modulation. Trends Cardiovasc. Med. 2002, 12, 206−213. (11) Mika, D.; Richter, W.; Conti, M. A CaMKII/PDE4D negative feedback regulates cAMP signaling. Proc. Natl. Acad. Sci. U. S. A. 2015, 112, 2023−2028. (12) Deubel, V.; Schlesinger, J. J.; Digoutte, J. P.; Girard, M. Comparative immunochemical and biological analysis of African and South American yellow fever viruses. Arch. Virol. 1987, 94, 331−338. (13) Houslay, M. D.; Conti, M.; Fancis, S. H. Preface. In Phosphodiesterases as Drug Targets; Francis, S. H., Conti, M., Houslay, M. D., Eds.; Handbook of Experimental Pharmacology, Vol. 204; Springer: Berlin, 2011; pp v−vii, DOI: 10.1007/978-3-64217969-3 (14) Craven, K. B.; Zagotta, W. N. Salt bridges and gating in the COOH-terminal region of HCN2 and CNGA1 channels. J. Gen. Physiol. 2004, 124, 663−677. (15) Chow, S. S.; VanPetegem, F.; Accili, E. A. Energetics of cyclic AMP binding to HCN channel C terminus reveal negative cooperativity. J. Biol. Chem. 2012, 287, 600−606. (16) Lolicato, M.; Nardini, M.; Gazzarini, S.; Möller, S.; Bertinetti, D.; Herberg, F. W.; Bolognesi, M.; Martin, H.; Fasolini, M.; Bertrand, J. A.; et al. Tetramerization dynamics of C-terminal domain underlies isoform-specific cAMP gating in hyperpolarization-activated cyclic nucleotide-gated channels. J. Biol. Chem. 2011, 286, 44811−44820. (17) Akimoto, M.; Zhang, Z.; Boulton, S.; Selvaratnam, R.; VanSchouwen, B.; Gloyd, M.; Accili, E. A.; Lange, O. F.; Melacini, G. A mechanism for the auto-inhibition of hyperpolarization-activated cyclic nucleotide-gated (HCN) channel opening and its relief by cAMP. J. Biol. Chem. 2014, 289, 22205−22220. (18) Huang, G. Y.; Kim, J. J.; Reger, A. S.; Lorenz, R.; Moon, E. W.; Zhao, C.; Casteel, D. E.; Bertinetti, D.; VanSchouwen, B.; Selvaratnam, R.; et al. Structural basis for cyclic-nucleotide selectivity and cGMPselective activation of PKG I. Structure 2014, 22, 116−124. (19) Kim, C.; Cheng, C. Y.; Saldanha, S. A.; Taylor, S. S. PKA-I holoenzyme structure reveals a mechanism for cAMP-dependent activation. Cell 2007, 130, 1032−1043. (20) Su, Y.; Dostmann, W. R.; Herberg, F. W.; Durick, K.; Xuong, N. H.; Ten-Eyck, L.; Taylor, S. S.; Varughese, K. I. Regulatory subunit of protein kinase A: structure of deletion mutant with cAMP binding domains. Science 1995, 269, 807−813. (21) Das, R.; Esposito, V.; Abu-Abed, M.; Anand, G. S.; Taylor, S. S.; Melacini, G. cAMP activation of PKA defines an ancient signaling mechanism. Proc. Natl. Acad. Sci. U. S. A. 2007, 104, 93−98. (22) Das, R.; Melacini, G. A model for agonism and antagonism in an ancient and ubiquitous cAMP-binding domain. J. Biol. Chem. 2007, 282, 581−593. (23) Akimoto, M.; Selvaratnam, R.; McNicholl, E. T.; Verma, G.; Taylor, S. S.; Melacini, G. Signaling through dynamic linkers as revealed by PKA. Proc. Natl. Acad. Sci. U. S. A. 2013, 110, 14231− 14236. (24) Akimoto, M.; Moleschi, K.; Boulton, S.; VanSchouwen, B.; Selvaratnam, R.; Taylor, S. S.; Melacini, G. Allosteric linkers in cAMP signalling. Biochem. Soc. Trans. 2014, 42, 139−144. (25) Abu-Abed, M.; Das, R.; Wang, L.; Melacini, G. Definition of an electrostatic relay switch critical for the cAMP-dependent activation of protein kinase A as revealed by the D170A mutant of RI alpha. Proteins: Struct., Funct., Genet. 2007, 69, 112−124.
(26) McNicholl, E. T.; Das, R.; SilDas, S.; Taylor, S. S.; Melacini, G. Communication between tandem cAMP binding domains in the regulatory subunit of protein kinase A-I alpha as revealed by domainsilencing mutations. J. Biol. Chem. 2010, 285, 15523−15537. (27) VanSchouwen, B.; Selvaratnam, R.; Giri, R.; Lorenz, R.; Herberg, F. W.; Kim, C.; Melacini, G. Mechanism of cAMP partial agonism in protein kinase G (PKG). J. Biol. Chem. 2015, 290, 28631− 28641. (28) Boulton, S.; Melacini, G. Advances in NMR methods to map allosteric sites: from models to translation. Chem. Rev. 2016, 116, 6267−6304. (29) Moleschi, K. J.; Akimoto, M.; Melacini, G. Measurement of state-specific association constants in allosteric sensors through molecular stapling and NMR. J. Am. Chem. Soc. 2015, 137, 10777− 10785. (30) Boulton, S.; Akimoto, M.; Selvaratnam, R.; Bashiri, A.; Melacini, G. A tool set to map allosteric networks through the NMR chemical shift covariance analysis. Sci. Rep. 2014, 4, 7306. (31) Selvaratnam, R.; Mazhab-Jafari, M. T.; Das, R.; Melacini, G. The auto-inhibitory role of the EPAC hinge helix as mapped by NMR. PLoS One 2012, 7, e48707. (32) Selvaratnam, R.; VanSchouwen, B.; Fogolari, F.; Mazhab-Jafari, M. T.; Das, R.; Melacini, G. The projection analysis of NMR chemical shifts reveals extended EPAC autoinhibition determinants. Biophys. J. 2012, 102, 630−639. (33) Selvaratnam, R.; Chowdhury, S.; VanSchouwen, B.; Melacini, G. Mapping allostery through the covariance analysis of NMR chemical shifts. Proc. Natl. Acad. Sci. U. S. A. 2011, 108, 6133−6138. (34) Das, R.; Chowdhury, S.; Mazhab-Jafari, M. T.; Sildas, S.; Selvaratnam, R.; Melacini, G. Dynamically driven ligand selectivity in cyclic nucleotide binding domains. J. Biol. Chem. 2009, 284, 23682− 23696. (35) Das, R.; Mazhab-Jafari, M. T.; Chowdhury, S.; SilDas, S.; Selvaratnam, R.; Melacini, G. Entropy-driven cAMP-dependent allosteric control of inhibitory interactions in exchange proteins directly activated by cAMP. J. Biol. Chem. 2008, 283, 19691−19703. (36) Mazhab-Jafari, M. T.; Das, R.; Fotheringham, S. A.; SilDas, S.; Chowdhury, S.; Melacini, G. Understanding cAMP-dependent allostery by NMR spectroscopy: comparative analysis of the EPAC1 cAMP-binding domain in its apo and cAMP-bound states. J. Am. Chem. Soc. 2007, 129, 14482−14492. (37) Das, R.; Abu-Abed, M.; Melacini, G. Mapping allostery through equilibrium perturbation NMR spectroscopy. J. Am. Chem. Soc. 2006, 128, 8406−8407. (38) Saponaro, A.; Pauleta, S. R.; Cantini, F.; Matzapetakis, M.; Hammann, C.; Donadoni, C.; Hu, L.; Thiel, G.; Banci, L.; Santoro, B.; et al. Structural basis for the mutual antagonism of cAMP and TRIP8b in regulating HCN channel function. Proc. Natl. Acad. Sci. U. S. A. 2014, 111, 14577−14582. (39) Puljung, M. C.; DeBerg, H. A.; Zagotta, W. N.; Stoll, S. Double electron-electron resonance reveals cAMP-induced conformational change in HCN channels. Proc. Natl. Acad. Sci. U. S. A. 2014, 111, 9816−9821. (40) VanSchouwen, B.; Akimoto, M.; Sayadi, M.; Fogolari, F.; Melacini, G. Role of dynamics in the auto-inhibition and activation of the hyperpolarization-activated cyclic-nucleotide-modulated (HCN) ion channels. J. Biol. Chem. 2015, 290, 17642−17654. (41) Ulens, C.; Siegelbaum, S. A. Regulation of hyperpolarizationactivated HCN channels by cAMP through a gating switch in binding domain symmetry. Neuron 2003, 40, 959−970. (42) Shehu, A.; Nussinov, R. Computational methods for exploration and analysis of macromolecular structure and dynamics. PLoS Comput. Biol. 2015, 11, e1004585. (43) Tsai, H. H.; Tsai, C. J.; Ma, B.; Nussinov, R. In silico protein design by combinatorial assembly of protein building blocks. Protein Sci. 2004, 13, 2753−2765. (44) Nussinov, R.; Tsai, C. J. Allostery without a conformational change? Revisiting the paradigm. Curr. Opin. Struct. Biol. 2015, 30, 17− 24. N
DOI: 10.1021/acs.jpcb.6b07735 J. Phys. Chem. B XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry B (45) Kim, J.; Masterson, L. R.; Cembran, A.; Verardi, R.; Shi, L.; Gao, J.; Taylor, S. S.; Veglia, G. Dysfunctional conformational dynamics of protein kinase A induced by a lethal mutant of phospholamban hinder phosphorylation. Proc. Natl. Acad. Sci. U. S. A. 2015, 112, 3716−3721. (46) Cembran, A.; Masterson, L. R.; McClendon, C. L.; Taylor, S. S.; Gao, J.; Veglia, G. Conformational equilibrium of N-myristoylated cAMP-dependent protein kinase A by molecular dynamics simulations. Biochemistry 2012, 51, 10186−10196. (47) Shi, L.; Cembran, A.; Gao, J.; Veglia, G. Tilt and azimuthal angles of a transmembrane peptide: a comparison between molecular dynamics calculations and solid-state NMR data of sarcolipin in lipid membranes. Biophys. J. 2009, 96, 3648−3662. (48) Zhang, L.; Borthakur, S.; Buck, M. Dissociation of a dynamic protein complex studied by all-atom molecular simulations. Biophys. J. 2016, 110, 877−886. (49) Zhang, L.; Polyansky, A.; Buck, M. Modeling transmembrane domain dimers/trimers of plexin receptors: implications for mechanisms of signal transmission across the membrane. PLoS One 2015, 10, e0121513. (50) Shoichet, B. K.; Walters, W. P.; Jiang, H.; Bajorath, J. Advances in computational medicinal chemistry: a reflection on the evolution of the field and perspective going forward. J. Med. Chem. 2016, 59, 4033− 4034. (51) Byun, J. A.; Melacini, G. Disordered regions flanking ordered domains modulate signaling transduction. Biophys. J. 2015, 109, 2447− 2448. (52) VanSchouwen, B.; Selvaratnam, R.; Fogolari, F.; Melacini, G. Role of dynamics in the autoinhibition and activation of the exchange protein directly activated by cyclic AMP (EPAC). J. Biol. Chem. 2011, 286, 42655−42669. (53) Fogolari, F.; Corazza, A.; Fortuna, S.; Soler, M. A.; VanSchouwen, B.; Brancolini, G.; Corni, S.; Melacini, G.; Esposito, G. Distance-based configurational entropy of proteins from molecular dynamics simulations. PLoS One 2015, 10, e0132356. (54) Malmstrom, R. D.; Kornev, A. P.; Taylor, S. S.; Amaro, R. E. Allostery through the computational microscope: cAMP activation of a canonical signalling domain. Nat. Commun. 2015, 6, 7588. (55) Boras, B. W.; Hirakis, S. P.; Votapka, L. W.; Malmstrom, R. D.; Amaro, R. E.; McCulloch, A. D. Bridging scales through multiscale modeling: a case study on protein kinase A Front. Physiol. 2015, 6, DOI: 10.3389/fphys.2015.00250. (56) Gullingsrud, J.; Kim, C.; Taylor, S. S.; McCammon, J. A. Dynamic binding of PKA regulatory subunit RI alpha. Structure 2006, 14, 141−149. (57) Huang, Y. M.; Huber, G.; McCammon, J. A. Electrostatic steering enhances the rate of cAMP binding to phosphodiesterase: Brownian dynamics modeling. Protein Sci. 2015, 24, 1884−1889. (58) Caves, L. S. D.; Evanseck, J. D.; Karplus, M. Locally accessible conformations of proteins: multiple molecular dynamics simulations of crambin. Protein Sci. 1998, 7, 649−666. (59) Lexa, K. W.; Carlson, H. A. Full protein flexibility is essential for proper hot-spot mapping. J. Am. Chem. Soc. 2011, 133, 200−202. (60) Chang, L.; Miyata, Y.; Ung, P. M. U.; Bertelsen, E. B.; McQuade, T. J.; Carlson, H. A.; Zuiderweg, E. R. P.; Gestwicki, J. E. Chemical screens against a reconstituted multiprotein complex: myricetin blocks DnaJ regulation of DnaK through an allosteric mechanism. Chem. Biol. 2011, 18, 210−221. (61) Monticelli, L.; Sorin, E. J.; Tieleman, D. P.; Pande, V. S.; Colombo, G. Molecular simulation of multistate peptide dynamics: a comparison between microsecond timescale sampling and multiple shorter trajectories. J. Comput. Chem. 2008, 29, 1740−1752. (62) Flöck, D.; Colacino, S.; Colombo, G.; Di Nola, A. Misfolding of the amyloid β-protein: a molecular dynamics study. Proteins: Struct., Funct., Genet. 2006, 62, 183−192. (63) Brooks, B. R.; Brooks, C. L., III; Mackerell, A. D., Jr.; Nilsson, L.; Petrella, R. J.; Roux, B.; Won, Y.; Archontis, G.; Bartels, C.; Boresch, S.; et al. CHARMM: the biomolecular simulation program. J. Comput. Chem. 2009, 30, 1545−1614.
(64) Pawley, N. H.; Gans, J. D.; Nicholson, L. K. Factors determining the reliable description of global tumbling parameters in solution NMR. J. Biomol. NMR 2002, 24, 215−229. (65) Wulf, J.; Pascuzzi, P. E.; Fahmy, A.; Martin, G. B.; Nicholson, L. K. The solution structure of type III effector protein AvrPto reveals conformational and dynamic features important for plant pathogenesis. Structure 2004, 12, 1257−1268. (66) García de la Torre, J.; Huertas, M. L.; Carrasco, B. HYDRONMR: prediction of NMR relaxation of globular proteins from atomic-level structures and hydrodynamic calculations. J. Magn. Reson. 2000, 147, 138−146. (67) Zhou, L.; Olivier, N. B.; Yao, H.; Young, E. C.; Siegelbaum, S. A. A conserved tripeptide in CNG and HCN channels regulates ligand gating by controlling C-terminal oligomerization. Neuron 2004, 44, 823−834. (68) Berrera, M.; Pantano, S.; Carloni, P. cAMP modulation of the cytoplasmic domain in the HCN2 channel investigated by molecular simulations. Biophys. J. 2006, 90, 3428−3433. (69) Zhou, L.; Siegelbaum, S. A. Gating of HCN channels by cyclic nucleotides: residue contacts that underlie ligand binding, selectivity, and efficacy. Structure 2007, 15, 655−670. (70) Kusch, J.; Thon, S.; Schulz, E.; Biskup, C.; Nache, V.; Zimmer, T.; Seifert, R.; Schwede, F.; Benndorf, K. How subunits cooperate in cAMP-induced activation of homotetrameric HCN2 channels. Nat. Chem. Biol. 2012, 8, 162−169. (71) Marni, F.; Wu, S.; Shah, G. M.; Xu, X. P.; Hackett, A. R.; Xie, C.; Shrestha, S.; Liu, L.; Liu, Q.; Zhou, L. Normal-mode-analysis-guided investigation of crucial intersubunit contacts in the cAMP-dependent gating in HCN channels. Biophys. J. 2012, 103, 19−28.
O
DOI: 10.1021/acs.jpcb.6b07735 J. Phys. Chem. B XXXX, XXX, XXX−XXX
