Available online at www.sciencedirect.com
ScienceDirect Molecular simulations of metal-coupled protein folding Wenfei Li1,2, Jun Wang1,2, Jian Zhang1,2 and Wei Wang1,2 Many proteins require help from metal cofactors to function properly. Due to the involvement of metal binding, folding of these metalloproteins can be much more complicated. In recent years, several computational methods have been developed to reveal the essential features of metal-coupled protein folding, ranging from quantum mechanics (QM) to atomistic and coarse-grained (CG) simulations. These theoretical tools have achieved great successes in solving the multiscale difficulties arising from metal binding, and provided new insights into the mechanisms of metalloprotein folding. In this review, we first discuss the interaction features of metalcoordination and then introduce several computational models and their applications in metal-coupled folding. Finally we discuss the effects of metal-binding on the protein energy landscape, which is followed by some perspectives. Addresses 1 National Laboratory of Solid State Microstructure, Department of Physics, Nanjing University, Nanjing 210093, China 2 Collaborative Innovation Center of Advanced Microstructures, Nanjing University, Nanjing 210093, China Corresponding authors: Wang, Jun ([email protected]) and Wang, Wei ([email protected])
Current Opinion in Structural Biology 2015, 30:25–31 This review comes from a themed issue on Folding and binding Edited by Annalisa Pastore and Guanghong Wei http://dx.doi.org/10.1016/j.sbi.2014.11.006 0959-440X/# 2014 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3.0/).
bind to proteins and how the binding couples with protein folding and conformational motions have become the central topics of many discussions. In particular, people are curious about whether the new folding features arising from metal binding conform to the same principle governing the folding of metal-less proteins [3]. Answers to these questions are not only fundamental to understand the general principle of protein folding but also helpful for developing better strategies to design new metalloproteins with certain functions, which holds great promise for treating related diseases [2–8]. Previous investigations into metalloprotein folding have been mostly experimental. Representative examples include, but are not limited to, the folding of myoglobin/ hemoglobin, cytochrome c, blue copper proteins, iron– sulfur proteins, calcium binding proteins, and zinc fingers [5–14,15,16]. These experimental studies contributed significantly to our knowledge of the thermodynamics and kinetics of metalloprotein folding and are well documented in several reviews [5–9]. In comparison, the computational works on metalloprotein folding are very rare, mostly due to the difficulties in realistically modelling the key interactions involving metal ions. Presently, there are just a few computational studies focusing on the metal-coupled protein folding [17,18–22]. However, these computational studies have already begun to provide new insights into the metal-coupled protein folding mechanisms. In this review, we will discuss these progresses.
Key features of interactions arising from metal binding Metal induced protonation/deprotonation
Introduction Inorganic cofactors are indispensable participants of life processes in molecular level and take on vital roles in the biological world build up with organic biomolecules, for example, proteins, nucleic acids, lipids, etc. It has been found that more than 30% of the proteins require the help of cofactors for their proper functioning [1,2]. Multivalent metal ions, for example, zinc, calcium, iron, copper, nickel, magnesium, etc., are typical cofactors. Life systems have evolved to take advantage of these metal ions to perform a number of key functions, such as electron transfer, oxygen storage, catalysis, and structure modulations. Although the eventual roles of these metal ions in biological systems can be very diverse, modulating protein stability and/or folding is the common effect of metal binding. Consequently, questions of how these metal ions www.sciencedirect.com
Typical metal ligands include thiolates of cysteines, imidazoles of histidines, carboxylates of glutamic and aspartic acids, backbone groups, and water molecules. Many of these ligands are titratable and are therefore sensitive to acid/base environment. Consequently, metal binding is often accompanied with proton release [12]. Such metal binding induced deprotonation contributes significantly to the folding thermodynamics and, in some cases, can be a key driving force for successful folding [12,22]. Accordingly, interactions and folding mechanism involving metal ions are easier to regulate via pH [19,23], adding complexity to the modelling of metalloprotein folding [19]. Charge transfer and induced polarization
Metal ions are good electron acceptors and therefore can accept negative charge from the ligands. The extent of charge transfer can affect the strength of coordination Current Opinion in Structural Biology 2015, 30:25–31
26 Folding and binding
bonds and modulate the coordination geometry and metal selectivity [24]. Besides, the highly charged metal ions often lead to polarization of the surrounding protein groups, which not only contributes to the metal binding but also modulates the global electrostatic interactions and residue contact networks. Particularly, for the groups with delocalized electrons, such as the imidazole ring of histidine, the polarization induced by metal binding can be crucial for the binding affinity. It is essential to include the effects of polarization and charge transfer in order to accurately model the metal coordination, which is actually the focus of several studies on building theoretical models of metalloproteins [24–28,29].
information, one may deduce the key steps involved in the metal-coupled folding. One example is the all-electron QM calculations of the nucleation structure in the zinc-induced folding of zinc-finger peptide [21]. Zincfingers represent a wide class of proteins whose folding depend on metal binding. By combining DFT and continuum dielectric solvation model, the structures of zinc finger mimetics with low free energies were identified and considered as the key nucleation structures during metal-coupled folding. The QM calculations suggested a successive metal binding process and emphasized a crucial role of the backbone groups in mediating the zinc coordination [21].
Multiscale coupling
Similar DFT calculations have been used to investigate the most probable metal binding modes of amyloid-b peptide and octarepeat domain of prion protein [33,38,39]. For better efficiency, hybrid dynamics simulations combining QM and classical MD have also been applied to some metalloproteins [40]. Although the nonadditive effects of metal coordination can be accurately described by such QM calculations, a common challenge for these methods is to obtain converged sampling of protein conformational space, which requires long time simulations and is a prerequisite for accurately estimating the conformational entropy and energy landscape.
The multivalent metal ions often bind to proteins with specific coordination geometries, which are largely determined by their electron structures. Although the metal binding sites are local in three-dimensional space, the metal ligands can cover a wide range of sequence. Consequently, the local interactions of metal binding may result in global changes of protein structure and dynamics, demonstrating multiscale coupling. Additionally, the timescales of the metal-induced protonation/ deprotonation, charge transfer, and polarization are much shorter than those of large-scale conformational motions of polypeptide chain. Such multiscale coupling has been exploited by many metalloproteins to accomplish signal transduction and other biological functions [30–32]. At the same time, it poses challenges to the computational modelling of metalloprotein folding.
Simulations of metal-coupled protein folding with multiple models Due to the above-mentioned features arising from metal binding, modelling metalloprotein folding is much more challenging compared to that of metal-less proteins. Particularly, the multiscale challenge of the metal induced interactions makes it impractical to fully describe all aspects of metal-coupled folding with any a single model. Consequently, very diverse methods, ranging from QM [21,33] to all-atom molecular dynamics (MD) [22,34–37] and CG models [17,19,20,23], have been developed to characterize certain aspects of the protein folding coupled with metal binding (Figure 1). QM calculations
Charge transfer, induced polarization, and protonation/ deprotonation essentially involve the motions of electrons. To fully describe such electron degrees of freedom requires QM methods. In principle, the density functional theory (DFT) can be used to tackle all aspects of the metal-coupled folding if not considering its computational difficulty. In practice, however, such delicate methods are mostly used for structure optimization and energy estimation for small and static biomolecules. On the basis of the structural and energy Current Opinion in Structural Biology 2015, 30:25–31
All-atom simulations
One solution to the timescale difficulties of QM methods is to empirically model the quantum chemical details of metal coordination within classical all-atom level models which is successful for the folding of small metal-less proteins [41]. In the literature, several models have been proposed to describe the coordination bonds within the framework of all-atom force field, including bonded model [42,43], nonbonded model [22,24–27,44,45], and dummy cation model [46]. Among these models, the nonbonded model is more suitable for the metal-coupled folding problem. In the nonbonded model, the coordination bonds are treated by the electrostatic and van der Waals (vdW) potentials. Different nonbonded models are featured by their distinguished strategies to treat the electrostatic interactions in order to correctly describe the coordination geometry. One interesting effort is the coordination model by Sakharov and Lim for zinc proteins, in which a distance dependent charge transfer model and local polarization around zinc binding sites were suggested [24]. With this model, the coordination number and geometry in the native state of the zinc-fingers can be well reproduced. Along the same line, Li et al. extended the nonbonded coordination model to metal-coupled protein folding problem by further considering the metal-induced www.sciencedirect.com
Simulations of metal ion coupled protein folding Li et al. 27
Figure 1
Time (s) larger proteins, e.g., calmodulin, azurin, and cytochrome c 100
small proteins, e.g., zinc finger and Ab
10–3
implicit or explicit metal cofactors 1-6 beads per residue
10–6 metal binding sites
coarse-grained model 10–9
nonbonded model NH
HC
S
Zn
e .19 –0
C
all-atom model
C H
–0.1
10–12
C
5e N .14 –0
45e
S –0 .1 9e
N
CH
HC
ncy
icie
Eff
cy
ura
NH
Acc
C
10–15 quantum mechanics 10
102
103
104
Protein size(atom #) Current Opinion in Structural Biology
Application scopes of different metal-binding models. The QM are often used for structure optimization and energy estimation of metal binding sites or small protein mimetics; the all-atom nonbonded metal model can be used for metal-coupled folding of small metalloproteins; the CG model with explicit/implicit metal ions can be used for metal-coupled folding of larger metalloproteins.
polarization of the histidine imidazole ring and the deprotonation of metal ligands [22]. Applications to the folding of a classical zinc finger showed great success and revealed tight coupling between the peptide folding and zinc binding. The folding of zinc finger follows a metal binding induced folding mechanism, which is mediated by the packing of the conserved hydrophobic residues. They also found that the packing of the hydrophobic residues and the coordination of the native ligands are highly cooperative. Additionally, the simulations directly showed the misligation and ligand exchange due to the involvement of zinc ions. Other efforts to improve the modelling of metal binding include the SIBFA polarizable force field [29], the polarizable multipole based electrostatic model [27], the quantum calibrated force field [25], the multiscale polarizable model [26], the short-long effective functions www.sciencedirect.com
[28], and the 12-6-4 LJ-type nonbonded model [45], although the focuses of these works are reproducing the coordination geometry and modelling the conformational motions around the native structure instead of investigating the metal-coupled folding problem. Coarse-grained simulations
Even with the above empirical approximation to the metal coordination, the description of folding with allatom MD was still limited to relatively small proteins. For larger proteins, CG models, typically with one to six beads per amino acid, are more appropriate [47]. In particular, the structure-based models [48], often with additional considerations of physiochemical properties [17,49–51], non-native interactions [52,53], realistic chain flexibility [54], and denaturant effects [55], have been successfully used in the folding studies of large and topologically complex proteins. Current Opinion in Structural Biology 2015, 30:25–31
28 Folding and binding
When applied to metal-coupled protein folding, these models have to be modified to include the effects of metal ions. Two strategies have been used to model the metal binding. In one strategy, the metal ions are modelled explicitly [18–20,56]. The strength of coordination between the central metal ion and the possible ligands are calibrated to experiments. In addition, the protonation/ deprotonation of titratable ligands and its dependence on pH values can be modelled by applying a grand canonical formalism [3,18,19]. Such model can reproduce the sequential assembly mechanism of cytochrome c observed in hydrogen exchange experiment [18]. Especially, this model can well describe the chemical frustration arising from deprotonation and ligation of nonnative ligands at alkaline pH condition [18,19], which gives rise to several intermediates as observed experimentally [23]. More importantly, these studies showed that the mechanisms of metalloprotein folding conform to the same energy landscape principles governing the folding of metal-less proteins [18,57]. A similar model was also employed to investigate the zinc binding coupled folding of azurin [20]. By computing the free energy profiles based on a variational free-energy functional, Zong et al. showed that the binding of zinc ion can lead to a shifting transition state, which well explained the curved chevron plot observed in experiment [20]. Another strategy is the implicit metal binding model [17]. In this model, the metal ions are considered implicitly, and the energetic effect of metal binding is modelled by strengthening the contacting interactions of metal ligands. The binding/unbinding of the metal ions were modelled by Monte Carlo simulations. The binding rate depends on the concentration of metal ions assuming a diffusion-limited binding process, whereas the metal dissociation rate depends on the binding energies. Such implicit ligand binding model was originally used to model the ligand binding coupled allosteric motions of proteins [58–60]. Recently, it has been successfully used to study the folding of a typical allosteric protein, calmodulin [17]. The calmodulin domain has two functional states, open and closed, depending on the binding of calcium ions. By integrating with a double-basin energy function with improved modelling of physicochemical feature and chain flexibility, computer simulations with the implicit metal binding model revealed the multiroute folding pathways, and the shifts of folding mechanisms with increasing calcium concentrations [17]. In addition, the simulations demonstrated the folding frustrations introduced by the allosteric feature of the energy landscape. Such a simplified metal binding model successfully describes the global changes of energy landscape arising from metal binding, though it cannot be used to describe the details of the coordination reaction due to the lack of essential degrees of freedom. Current Opinion in Structural Biology 2015, 30:25–31
Interplays between the ‘coordination’ landscape and folding landscape Through the coordination interactions, the multivalent metal ions could introduce effective interactions between the metal ligands, shaping up a coordination landscape. Considering whether there is a specific metal binding mode, the coordination landscape could be funnel-like (with specific binding sites) or rugged (with multiple competitive binding patterns). Through the interplay between the coordination landscape and the folding landscape (the intrinsic energy landscape), several kinds of behaviors would emerge. Firstly, for the nonspecific coordination, the inclusion of the metal ion would increase the ruggedness of the composite landscape (Figure 2a). There would be multiple trapped conformations depending on the environment conditions (with the octarepeat domain of prion protein as the example) [39]. Differently, the funnel-shaped coordination landscape could stabilize a certain region of the composite energy landscape. Particularly, when there are no stable structures for the isolated peptide chain, the binding with the metal ion would shape up a stable folder (Figure 2b), and the metal ion is obligatory for the successful folding (with the zinc-finger protein as the example) [22]. When the folding funnel matches the coordination funnel, the folded state would be stabilized in a deeper funnel (Figure 2c), and the folding pathways may be modulated by the competitions of two kinds of funnels (with the azurin as the example) [19,20]. On the other hand, when two kinds of funnels are located at different regions of the landscape, two-basin funnel structure of the landscape would be favourite (Figure 2d). In this case, the binding with a metal ion would induce a large change of the landscape, and the metal ion acts as a key to switch such kind of changes (with the calmodulin as the example) [17]. A similar kind of cooperation/competition of binding and folding landscapes could be also observed in dimerization of proteins [61]. In summary, the metal ions introduce another degree of freedom to tailor the landscape and produce rich behaviors of proteins, which offers more opportunities to design proteins with new dynamics and functions.
Conclusions The recent development of computer resources and theoretical tools has greatly extended the scope and depth of the computational studies of protein folding. Consequently, metal-coupled protein folding, which has long been a challenging and extensively studied topic in experimental fields, has become the subjects of computational studies. In this review, we discussed several recent works that have developed computational tools to explore metal-coupled protein folding. These computational works have already shown great successes in describing many aspects of metal-coupled protein folding. However, relative to its abundance and biological importance, the computational works on metalloprotein www.sciencedirect.com
Simulations of metal ion coupled protein folding Li et al. 29
Figure 2
(a)
(b)
(c)
(d)
energy
folding landscape
composite landscape
nonspecific binding
specific binding Current Opinion in Structural Biology
Schematic diagram of the effects of metal binding on the energy landscapes of proteins. (a) Nonspecific binding leads to rugged composite landscape due to multiple competitive coordination modes; (b) binding creates a funnel on the composite landscape when the coordination landscape dominates the overall interactions; (c) binding induces a deeper composite landscape when the coordination landscape is consistent with the folding landscape; and (d) binding reshapes the folding landscape when the coordination landscape matches a metastable state on the folding funnel.
folding are still rare. In the future, developing multiscale methodologies, which can simultaneously model the metal-involved local interactions and the global motions of the polypeptide chain with high accuracy and efficiency, can further extend the computational studies of metalloprotein folding. In addition, similar methods can be used to model metal-coupled protein aggregations and nucleic acid folding [62–65].
Conflict of interest statement Nothing declared.
Acknowledgements This work was supported by National Basic Research Program of China (973 Program) (no. 2013CB834100), Natural Science Foundation of China (Grants 11334004, 11174134, and 11274157) and Jiangsu Province (BK2011546).
References and recommended reading Papers of particular interest, published within the period of review, have been highlighted as: of special interest of outstanding interest 1.
Gray HB: Biological inorganic chemistry at the beginning of the 21st century. Proc Natl Acad Sci USA 2003, 100:3563-3568.
2.
Wittung-Stafshede P: Role of cofactors in protein folding. Acc Chem Res 2002, 35:201-208.
3.
Weinkam P, Zimmermann J, Romesberg FE, Wolynes PG: The folding energy landscape and free energy excitations of cytochrome c. Acc Chem Res 2010, 43:652-660.
4.
Klug A: Towards therapeutic applications of engineered zinc finger proteins. FEBS Lett 2005, 579:892-894.
www.sciencedirect.com
5.
Cox EH, McLendon GL: Zinc-dependent protein folding. Curr Opin Chem Biol 2000, 4:162-165.
6.
Gomes GM, Wittung-Stafshede P: Metal ions, protein folding, and conformational states: an introduction. In Protein Folding and Metal Ions. Edited by Gomes GM, Wittung-Stafshede P. Boca Raton: CRC Press; 2011:3-11.
7.
Wilson CJ, Apiyo D, Wittung-Stafshede P: Role of cofactors in metalloprotein folding. Q Rev Biophys 2004, 37:285-314.
8.
Bren KL, Pecoraro VL, Gray HB: Metalloprotein folding. Inorg Chem 2004, 43:7894-7896.
9.
Palm-Espling ME, Niemiec MS, Wittung-Stafshede P: Role of metal in folding and stability of copper proteins in vitro. Biochim Biophys Acta 2012, 1823:1594-1603.
10. Palmieri M, Malgieri G, Russo L, Baglivo I, Esposito S, Netti F, Del Gatto A, de Paola I, Zaccaro L, Pedone PV et al.: Structural Zn(II) implies a switch from fully cooperative to partly downhill folding in highly homologous proteins. J Am Chem Soc 2013, 135:5220-5228. 11. Morleo A, Bonomi F, Iametti S, Huang VW, Kurtz DM Jr: Ironnucleated folding of a metalloprotein in high urea: resolution of metal binding and protein folding events. Biochemistry 2010, 49:6627-6634. 12. Blasie CA, Berg JM: Structure-based thermodynamic analysis of a coupled metal binding-protein folding reaction involving a zinc finger peptide. Biochemistry 2002, 41:15068-15073. 13. Miura T, Satoh T, Takeuchi H: Role of metal–ligand coordination in the folding pathway of zinc finger peptides. Biochim Biophys Acta 1998, 1384:171-179. 14. Lv C, Gao X, Li W, Xue B, Qin M, Burtnick LD, Zhou H, Cao Y, Robinson RC, Wang W: Single-molecule force spectroscopy reveals force-enhanced binding of calcium ions by gelsolin. Nat Commun 2014, 5:4623. 15. Stigler J, Rief M: Calcium-dependent folding of single calmodulin molecules. Proc Natl Acad Sci USA 2012, 109: 17814-17819. A single molecule optical tweezer experiment demonstrated the modulations of folding mechanisms of calmodulin by calcium concentrations. The kinetic measurements showed that the folding pathway with calcium Current Opinion in Structural Biology 2015, 30:25–31
30 Folding and binding
bound to the denatured state is the dominant pathway at high calcium concentrations. 16. Tang J, Kang SG, Saven JG, Gai F: Characterization of the cofactor-induced folding mechanism of a zinc-binding peptide using computationally designed mutants. J Mol Biol 2009, 389:90-102. 17. Li WF, Wang W, Takada S: Energy landscape views for interplays among folding, binding, and allostery of calmodulin domains. Proc Natl Acad Sci USA 2014, 111:10550-10555. A coarse-grained model provided an integrated description of folding, metal binding and allostery. The simulations demonstrated the shift of folding mechanisms of calmodulin domains with increasing calcium concentrations. 18. Weinkam P, Wolynes PG: The folding landscape of metalloproteins. In Protein Folding and Metal Ions. Edited by 247-273. Gomes GM, Wittung-Stafshede P. Boca Raton: CRC Press; 2011: 19. Weinkam P, Romesberg FE, Wolynes PG: Chemical frustration in the protein folding landscape: grand canonical ensemble simulations of cytochrome c. Biochemistry 2009, 48:2394-2402.
34. Dasmeh P, Kepp KP: Unfolding simulations of holomyoglobin from four mammals: identification of intermediates and betasheet formation from partially unfolded states. PLoS One 2013, 8:e80308. 35. Evoli S, Guzzi R, Rizzuti B: Dynamics and unfolding pathway of chimeric azurin variants: insights from molecular dynamics simulation. J Biol Inorg Chem 2013, 18:739-749. 36. Chen CJ, Huang YZ, Jiang XW, Xiao Y: Binding free-energy calculation of an ion–peptide complex by constrained dynamics. Phys Rev E 2013, 87:062705. 37. Li WF, Zhang J, Su Y, Wang J, Qin M, Wang W: Effects of zinc binding on the conformational distribution of the amyloid-beta peptide based on molecular dynamics simulations. J Phys Chem B 2007, 111:13814-13821. 38. Furlan S, Hureau C, Faller P, La Penna G: Modeling copper binding to the amyloid-beta peptide at different pH: toward a molecular mechanism for Cu reduction. J Phys Chem B 2012, 116:11899-11910.
20. Zong C, Wilson CJ, Shen T, Wittung-Stafshede P, Mayo SL, Wolynes PG: Establishing the entatic state in folding metallated Pseudomonas aeruginosa azurin. Proc Natl Acad Sci USA 2007, 104:3159-3164.
39. Furlana S, La Pennab G: Metal ions and protons compete for ligand atoms in disordered peptides: examples from computer simulations of copper binding to the prion tandem repeat. Coord Chem Rev 2012, 256:2234-2244.
21. Dudev T, Lim C: All-electron calculations of the nucleation structures in metal-induced zinc-finger folding: role of the peptide backbone. J Am Chem Soc 2007, 129:12497-12504. 22. Li W, Zhang J, Wang J, Wang W: Metal-coupled folding of Cys2His2 zinc-finger. J Am Chem Soc 2008, 130:892-900.
40. Sparta M, Shirvanyants D, Ding F, Dokholyan NV, Alexandrova AN: Hybrid dynamics simulation engine for metalloproteins. Biophys J 2012, 103:767-776. By combining the quantum mechanics with an all-atom discrete molecular dynamics method, they can conduct simulations for metalloproteins with high accuracy for a longer time duration.
23. Weinkam P, Zimmermann J, Sagle LB, Matsuda S, Dawson PE, Wolynes PG, Romesberg FE: Characterization of alkaline transitions in ferricytochrome c using carbon-deuterium infrared probes. Biochemistry 2008, 47:13470-13480.
41. Piana S, Klepeis JL, Shaw DE: Assessing the accuracy of physical models used in protein-folding simulations: quantitative evidence from long molecular dynamics simulations. Curr Opin Struct Biol 2014, 24:98-105.
24. Sakharov DV, Lim C: Zn protein simulations including charge transfer and local polarization effects. J Am Chem Soc 2005, 127:4921-4929.
42. Lin F, Wang RX: Systematic derivation of AMBER force field parameters applicable to zinc-containing systems. J Chem Theory Comput 2010, 6:1852-1870.
25. Zhu T, Xiao X, Ji C, Zhang JZH: A new quantum calibrated force field for zinc–protein complex. J Chem Theory Comput 2013, 9:1788-1798.
43. Burger SK, Lacasse M, Verstraelen T, Drewry J, Gunning P, Ayers PW: Automated parametrization of AMBER force field terms from vibrational analysis with a focus on functionalizing dinuclear zinc(II) scaffolds. J Chem Theory Comput 2012, 8: 554-562.
26. Huang YD, Shuai JW: Induced dipoles incorporated into allatom Zn protein simulations with multiscale modeling. J Phys Chem B 2013, 117:6138-6148. 27. Zhang J, Yang W, Piquemal JP, Ren P: Modeling structural coordination and ligand binding in zinc proteins with a polarizable potential. J Chem Theory Comput 2012, 8:13141324. 28. Wu R, Lu Z, Cao Z, Zhang Y: A transferable nonbonded pairwise force field to model zinc interactions in metalloproteins. J Chem Theory Comput 2011, 7:433-443. 29. Devereux M, Gresh N, Piquemal JP, Meuwly M: A supervised fitting approach to force field parametrization with application to the SIBFA polarizable force field. J Comput Chem 2014, 35:1577-1591. This paper provided a promising approach to the parameter fitting of the SIBFA polarizable force field which was one of the first potentials to model the metal-protein interactions considering explicit charge transfer and polarization effects. 30. Chakravorty DK, Parker TM, Guerra AJ, Sherrill CD, Giedroc DP, Merz KM Jr: Energetics of zinc-mediated interactions in the allosteric pathways of metal sensor proteins. J Am Chem Soc 2013, 135:30-33. 31. Chakravorty DK, Wang B, Lee CW, Giedroc DP, Merz KM Jr: Simulations of allosteric motions in the zinc sensor CzrA. J Am Chem Soc 2012, 134:3367-3376. 32. Reyes-Caballero H, Campanello GC, Giedroc DP: Metalloregulatory proteins: metal selectivity and allosteric switching. Biophys Chem 2011, 156:103-114. 33. Marino T, Russo N, Toscano M, Pavelka M: On the metal ion (Zn2+, Cu2+) coordination with beta-amyloid peptide: DFT computational study. Interdiscip Sci 2010, 2:57-69. Current Opinion in Structural Biology 2015, 30:25–31
44. Stote RH, Karplus M: Zinc binding in proteins and solution: a simple but accurate nonbonded representation. Proteins 1995, 23:12-31. 45. Li PF, Merz KM: Taking into account the ion-induced dipole interaction in the nonbonded model of ions. J Chem Theory Comput 2014, 10:289-297. A nonbonded model of metal ions with a 12-6-4 LJ type potential was proposed to describe the metal ion induced dipole interaction. This model can reproduce the hydration free energy and the coordination geometry for a number of metal ions and is readily incorporated into standard allatom force fields. 46. Pang YP, Xu K, Yazal JE, Prendergas FG: Successful molecular dynamics simulation of the zinc-bound farnesyltransferase using the cationic dummy atom approach. Protein Sci 2000, 9:1857-1865. 47. Takada S: Coarse-grained molecular simulations of large biomolecules. Curr Opin Struct Biol 2012, 22:130-137. An excellent survey of the recently developed and widely used coarsegrained models for the simulations of large biomolecules. The pros and cons as well as the application scope of the coarse-grained models were discussed. 48. Clementi C, Nymeyer H, Onuchic JN: Topological and energetic factors: what determines the structural details of the transition state ensemble and ‘‘en-route’’ intermediates for protein folding? An investigation for small globular proteins. J Mol Biol 2000, 298:937-953. 49. Li W, Terakawa T, Wang W, Takada S: Energy landscape and multiroute folding of topologically complex proteins adenylate kinase and 2ouf-knot. Proc Natl Acad Sci USA 2012, 109: 17789-17794. www.sciencedirect.com
Simulations of metal ion coupled protein folding Li et al. 31
50. Li W, Wolynes PG, Takada S: Frustration, specific sequence dependence, and nonlinearity in large-amplitude fluctuations of allosteric proteins. Proc Natl Acad Sci USA 2011, 108: 3504-3509.
59. Wang Y, Tang C, Wang E, Wang J: Exploration of multi-state conformational dynamics and underlying global functional landscape of maltose binding protein. PLoS Comput Biol 2012, 8:e1002471.
51. Li W, Yoshii H, Hori N, Kameda T, Takada S: Multiscale methods for protein folding simulations. Methods 2010, 52:106-114.
60. Kenzaki H, Koga N, Hori N, Kanada R, Li W, Okazaki K-i, Yao X-Q, Takada S: CafeMol: a coarse-grained biomolecular simulator for simulating proteins at work. J Chem Theory Comput 2011, 7:1979-1989.
52. Chan HS, Zhang Z, Wallin S, Liu Z: Cooperativity, local-nonlocal coupling, and nonnative interactions: principles of protein folding from coarse-grained models. Annu Rev Phys Chem 2011, 62:301-326. 53. Shental-Bechor D, Smith MT, Mackenzie D, Broom A, Marcovitz A, Ghashut F, Go C, Bralha F, Meiering EM, Levy Y: Nonnative interactions regulate folding and switching of myristoylated protein. Proc Natl Acad Sci USA 2012, 109:17839-17844. 54. Terakawa T, Higo J, Takada S: Multi-scale ensemble modeling of modular proteins with intrinsically disordered linker regions: application to p53. Biophys J 2014, 107:721-729. 55. Thirumalai D, Liu Z, O’Brien EP, Reddy G: Protein folding: from theory to practice. Curr Opin Struct Biol 2013, 23:22-29. 56. Wang Q, Liang KC, Czader A, Waxham MN, Cheung MS: The effect of macromolecular crowding, ionic strength and calcium binding on calmodulin dynamics. PLoS Comput Biol 2011, 7:e1002114.
61. Chu X, Gan L, Wang E, Wang J: Quantifying the topography of the intrinsic energy landscape of flexible biomolecular recognition. Proc Natl Acad Sci USA 2013, 110:E2342-E2351. Structure-based model was used to quantitatively characterize the energy landscapes of interfacial binding and monomeric folding for fifteen homodimers, and the interplay between the two kinds of energy landscapes can be well related to the association mechanism of the monomers. 62. Miller Y, Ma BY, Nussinov R: Metal binding sites in amyloid oligomers: complexes and mechanisms. Coord Chem Rev 2012, 256:2245-2252. 63. Bian Y, Tan C, Wang J, Sheng Y, Zhang J, Wang W: Atomistic picture for the folding pathway of a hybrid-1 type human telomeric DNA G-quadruplex. PLoS Comput Biol 2014, 10:e1003562.
57. Onuchic JN, Luthey-Schulten Z, Wolynes PG: Theory of protein folding: the energy landscape perspective. Annu Rev Phys Chem 1997, 48:545-600.
64. Li W, Hou XM, Wang PY, Xi XG, Li M: Direct measurement of sequential folding pathway and energy landscape of human telomeric G-quadruplex structures. J Am Chem Soc 2013, 135:6423-6426.
58. Okazaki K, Takada S: Dynamic energy landscape view of coupled binding and protein conformational change: inducedfit versus population-shift mechanisms. Proc Natl Acad Sci USA 2008, 105:11182-11187.
65. Shen X, Ding KK, Zhang FS: Ion effects of Sr2+, Cs+ and IS on DNA in aqueous solutions. Chem Phys Lett 2013, 574: 100-105.
www.sciencedirect.com
Current Opinion in Structural Biology 2015, 30:25–31
ScienceDirect Molecular simulations of metal-coupled protein folding Wenfei Li1,2, Jun Wang1,2, Jian Zhang1,2 and Wei Wang1,2 Many proteins require help from metal cofactors to function properly. Due to the involvement of metal binding, folding of these metalloproteins can be much more complicated. In recent years, several computational methods have been developed to reveal the essential features of metal-coupled protein folding, ranging from quantum mechanics (QM) to atomistic and coarse-grained (CG) simulations. These theoretical tools have achieved great successes in solving the multiscale difficulties arising from metal binding, and provided new insights into the mechanisms of metalloprotein folding. In this review, we first discuss the interaction features of metalcoordination and then introduce several computational models and their applications in metal-coupled folding. Finally we discuss the effects of metal-binding on the protein energy landscape, which is followed by some perspectives. Addresses 1 National Laboratory of Solid State Microstructure, Department of Physics, Nanjing University, Nanjing 210093, China 2 Collaborative Innovation Center of Advanced Microstructures, Nanjing University, Nanjing 210093, China Corresponding authors: Wang, Jun ([email protected]) and Wang, Wei ([email protected])
Current Opinion in Structural Biology 2015, 30:25–31 This review comes from a themed issue on Folding and binding Edited by Annalisa Pastore and Guanghong Wei http://dx.doi.org/10.1016/j.sbi.2014.11.006 0959-440X/# 2014 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3.0/).
bind to proteins and how the binding couples with protein folding and conformational motions have become the central topics of many discussions. In particular, people are curious about whether the new folding features arising from metal binding conform to the same principle governing the folding of metal-less proteins [3]. Answers to these questions are not only fundamental to understand the general principle of protein folding but also helpful for developing better strategies to design new metalloproteins with certain functions, which holds great promise for treating related diseases [2–8]. Previous investigations into metalloprotein folding have been mostly experimental. Representative examples include, but are not limited to, the folding of myoglobin/ hemoglobin, cytochrome c, blue copper proteins, iron– sulfur proteins, calcium binding proteins, and zinc fingers [5–14,15,16]. These experimental studies contributed significantly to our knowledge of the thermodynamics and kinetics of metalloprotein folding and are well documented in several reviews [5–9]. In comparison, the computational works on metalloprotein folding are very rare, mostly due to the difficulties in realistically modelling the key interactions involving metal ions. Presently, there are just a few computational studies focusing on the metal-coupled protein folding [17,18–22]. However, these computational studies have already begun to provide new insights into the metal-coupled protein folding mechanisms. In this review, we will discuss these progresses.
Key features of interactions arising from metal binding Metal induced protonation/deprotonation
Introduction Inorganic cofactors are indispensable participants of life processes in molecular level and take on vital roles in the biological world build up with organic biomolecules, for example, proteins, nucleic acids, lipids, etc. It has been found that more than 30% of the proteins require the help of cofactors for their proper functioning [1,2]. Multivalent metal ions, for example, zinc, calcium, iron, copper, nickel, magnesium, etc., are typical cofactors. Life systems have evolved to take advantage of these metal ions to perform a number of key functions, such as electron transfer, oxygen storage, catalysis, and structure modulations. Although the eventual roles of these metal ions in biological systems can be very diverse, modulating protein stability and/or folding is the common effect of metal binding. Consequently, questions of how these metal ions www.sciencedirect.com
Typical metal ligands include thiolates of cysteines, imidazoles of histidines, carboxylates of glutamic and aspartic acids, backbone groups, and water molecules. Many of these ligands are titratable and are therefore sensitive to acid/base environment. Consequently, metal binding is often accompanied with proton release [12]. Such metal binding induced deprotonation contributes significantly to the folding thermodynamics and, in some cases, can be a key driving force for successful folding [12,22]. Accordingly, interactions and folding mechanism involving metal ions are easier to regulate via pH [19,23], adding complexity to the modelling of metalloprotein folding [19]. Charge transfer and induced polarization
Metal ions are good electron acceptors and therefore can accept negative charge from the ligands. The extent of charge transfer can affect the strength of coordination Current Opinion in Structural Biology 2015, 30:25–31
26 Folding and binding
bonds and modulate the coordination geometry and metal selectivity [24]. Besides, the highly charged metal ions often lead to polarization of the surrounding protein groups, which not only contributes to the metal binding but also modulates the global electrostatic interactions and residue contact networks. Particularly, for the groups with delocalized electrons, such as the imidazole ring of histidine, the polarization induced by metal binding can be crucial for the binding affinity. It is essential to include the effects of polarization and charge transfer in order to accurately model the metal coordination, which is actually the focus of several studies on building theoretical models of metalloproteins [24–28,29].
information, one may deduce the key steps involved in the metal-coupled folding. One example is the all-electron QM calculations of the nucleation structure in the zinc-induced folding of zinc-finger peptide [21]. Zincfingers represent a wide class of proteins whose folding depend on metal binding. By combining DFT and continuum dielectric solvation model, the structures of zinc finger mimetics with low free energies were identified and considered as the key nucleation structures during metal-coupled folding. The QM calculations suggested a successive metal binding process and emphasized a crucial role of the backbone groups in mediating the zinc coordination [21].
Multiscale coupling
Similar DFT calculations have been used to investigate the most probable metal binding modes of amyloid-b peptide and octarepeat domain of prion protein [33,38,39]. For better efficiency, hybrid dynamics simulations combining QM and classical MD have also been applied to some metalloproteins [40]. Although the nonadditive effects of metal coordination can be accurately described by such QM calculations, a common challenge for these methods is to obtain converged sampling of protein conformational space, which requires long time simulations and is a prerequisite for accurately estimating the conformational entropy and energy landscape.
The multivalent metal ions often bind to proteins with specific coordination geometries, which are largely determined by their electron structures. Although the metal binding sites are local in three-dimensional space, the metal ligands can cover a wide range of sequence. Consequently, the local interactions of metal binding may result in global changes of protein structure and dynamics, demonstrating multiscale coupling. Additionally, the timescales of the metal-induced protonation/ deprotonation, charge transfer, and polarization are much shorter than those of large-scale conformational motions of polypeptide chain. Such multiscale coupling has been exploited by many metalloproteins to accomplish signal transduction and other biological functions [30–32]. At the same time, it poses challenges to the computational modelling of metalloprotein folding.
Simulations of metal-coupled protein folding with multiple models Due to the above-mentioned features arising from metal binding, modelling metalloprotein folding is much more challenging compared to that of metal-less proteins. Particularly, the multiscale challenge of the metal induced interactions makes it impractical to fully describe all aspects of metal-coupled folding with any a single model. Consequently, very diverse methods, ranging from QM [21,33] to all-atom molecular dynamics (MD) [22,34–37] and CG models [17,19,20,23], have been developed to characterize certain aspects of the protein folding coupled with metal binding (Figure 1). QM calculations
Charge transfer, induced polarization, and protonation/ deprotonation essentially involve the motions of electrons. To fully describe such electron degrees of freedom requires QM methods. In principle, the density functional theory (DFT) can be used to tackle all aspects of the metal-coupled folding if not considering its computational difficulty. In practice, however, such delicate methods are mostly used for structure optimization and energy estimation for small and static biomolecules. On the basis of the structural and energy Current Opinion in Structural Biology 2015, 30:25–31
All-atom simulations
One solution to the timescale difficulties of QM methods is to empirically model the quantum chemical details of metal coordination within classical all-atom level models which is successful for the folding of small metal-less proteins [41]. In the literature, several models have been proposed to describe the coordination bonds within the framework of all-atom force field, including bonded model [42,43], nonbonded model [22,24–27,44,45], and dummy cation model [46]. Among these models, the nonbonded model is more suitable for the metal-coupled folding problem. In the nonbonded model, the coordination bonds are treated by the electrostatic and van der Waals (vdW) potentials. Different nonbonded models are featured by their distinguished strategies to treat the electrostatic interactions in order to correctly describe the coordination geometry. One interesting effort is the coordination model by Sakharov and Lim for zinc proteins, in which a distance dependent charge transfer model and local polarization around zinc binding sites were suggested [24]. With this model, the coordination number and geometry in the native state of the zinc-fingers can be well reproduced. Along the same line, Li et al. extended the nonbonded coordination model to metal-coupled protein folding problem by further considering the metal-induced www.sciencedirect.com
Simulations of metal ion coupled protein folding Li et al. 27
Figure 1
Time (s) larger proteins, e.g., calmodulin, azurin, and cytochrome c 100
small proteins, e.g., zinc finger and Ab
10–3
implicit or explicit metal cofactors 1-6 beads per residue
10–6 metal binding sites
coarse-grained model 10–9
nonbonded model NH
HC
S
Zn
e .19 –0
C
all-atom model
C H
–0.1
10–12
C
5e N .14 –0
45e
S –0 .1 9e
N
CH
HC
ncy
icie
Eff
cy
ura
NH
Acc
C
10–15 quantum mechanics 10
102
103
104
Protein size(atom #) Current Opinion in Structural Biology
Application scopes of different metal-binding models. The QM are often used for structure optimization and energy estimation of metal binding sites or small protein mimetics; the all-atom nonbonded metal model can be used for metal-coupled folding of small metalloproteins; the CG model with explicit/implicit metal ions can be used for metal-coupled folding of larger metalloproteins.
polarization of the histidine imidazole ring and the deprotonation of metal ligands [22]. Applications to the folding of a classical zinc finger showed great success and revealed tight coupling between the peptide folding and zinc binding. The folding of zinc finger follows a metal binding induced folding mechanism, which is mediated by the packing of the conserved hydrophobic residues. They also found that the packing of the hydrophobic residues and the coordination of the native ligands are highly cooperative. Additionally, the simulations directly showed the misligation and ligand exchange due to the involvement of zinc ions. Other efforts to improve the modelling of metal binding include the SIBFA polarizable force field [29], the polarizable multipole based electrostatic model [27], the quantum calibrated force field [25], the multiscale polarizable model [26], the short-long effective functions www.sciencedirect.com
[28], and the 12-6-4 LJ-type nonbonded model [45], although the focuses of these works are reproducing the coordination geometry and modelling the conformational motions around the native structure instead of investigating the metal-coupled folding problem. Coarse-grained simulations
Even with the above empirical approximation to the metal coordination, the description of folding with allatom MD was still limited to relatively small proteins. For larger proteins, CG models, typically with one to six beads per amino acid, are more appropriate [47]. In particular, the structure-based models [48], often with additional considerations of physiochemical properties [17,49–51], non-native interactions [52,53], realistic chain flexibility [54], and denaturant effects [55], have been successfully used in the folding studies of large and topologically complex proteins. Current Opinion in Structural Biology 2015, 30:25–31
28 Folding and binding
When applied to metal-coupled protein folding, these models have to be modified to include the effects of metal ions. Two strategies have been used to model the metal binding. In one strategy, the metal ions are modelled explicitly [18–20,56]. The strength of coordination between the central metal ion and the possible ligands are calibrated to experiments. In addition, the protonation/ deprotonation of titratable ligands and its dependence on pH values can be modelled by applying a grand canonical formalism [3,18,19]. Such model can reproduce the sequential assembly mechanism of cytochrome c observed in hydrogen exchange experiment [18]. Especially, this model can well describe the chemical frustration arising from deprotonation and ligation of nonnative ligands at alkaline pH condition [18,19], which gives rise to several intermediates as observed experimentally [23]. More importantly, these studies showed that the mechanisms of metalloprotein folding conform to the same energy landscape principles governing the folding of metal-less proteins [18,57]. A similar model was also employed to investigate the zinc binding coupled folding of azurin [20]. By computing the free energy profiles based on a variational free-energy functional, Zong et al. showed that the binding of zinc ion can lead to a shifting transition state, which well explained the curved chevron plot observed in experiment [20]. Another strategy is the implicit metal binding model [17]. In this model, the metal ions are considered implicitly, and the energetic effect of metal binding is modelled by strengthening the contacting interactions of metal ligands. The binding/unbinding of the metal ions were modelled by Monte Carlo simulations. The binding rate depends on the concentration of metal ions assuming a diffusion-limited binding process, whereas the metal dissociation rate depends on the binding energies. Such implicit ligand binding model was originally used to model the ligand binding coupled allosteric motions of proteins [58–60]. Recently, it has been successfully used to study the folding of a typical allosteric protein, calmodulin [17]. The calmodulin domain has two functional states, open and closed, depending on the binding of calcium ions. By integrating with a double-basin energy function with improved modelling of physicochemical feature and chain flexibility, computer simulations with the implicit metal binding model revealed the multiroute folding pathways, and the shifts of folding mechanisms with increasing calcium concentrations [17]. In addition, the simulations demonstrated the folding frustrations introduced by the allosteric feature of the energy landscape. Such a simplified metal binding model successfully describes the global changes of energy landscape arising from metal binding, though it cannot be used to describe the details of the coordination reaction due to the lack of essential degrees of freedom. Current Opinion in Structural Biology 2015, 30:25–31
Interplays between the ‘coordination’ landscape and folding landscape Through the coordination interactions, the multivalent metal ions could introduce effective interactions between the metal ligands, shaping up a coordination landscape. Considering whether there is a specific metal binding mode, the coordination landscape could be funnel-like (with specific binding sites) or rugged (with multiple competitive binding patterns). Through the interplay between the coordination landscape and the folding landscape (the intrinsic energy landscape), several kinds of behaviors would emerge. Firstly, for the nonspecific coordination, the inclusion of the metal ion would increase the ruggedness of the composite landscape (Figure 2a). There would be multiple trapped conformations depending on the environment conditions (with the octarepeat domain of prion protein as the example) [39]. Differently, the funnel-shaped coordination landscape could stabilize a certain region of the composite energy landscape. Particularly, when there are no stable structures for the isolated peptide chain, the binding with the metal ion would shape up a stable folder (Figure 2b), and the metal ion is obligatory for the successful folding (with the zinc-finger protein as the example) [22]. When the folding funnel matches the coordination funnel, the folded state would be stabilized in a deeper funnel (Figure 2c), and the folding pathways may be modulated by the competitions of two kinds of funnels (with the azurin as the example) [19,20]. On the other hand, when two kinds of funnels are located at different regions of the landscape, two-basin funnel structure of the landscape would be favourite (Figure 2d). In this case, the binding with a metal ion would induce a large change of the landscape, and the metal ion acts as a key to switch such kind of changes (with the calmodulin as the example) [17]. A similar kind of cooperation/competition of binding and folding landscapes could be also observed in dimerization of proteins [61]. In summary, the metal ions introduce another degree of freedom to tailor the landscape and produce rich behaviors of proteins, which offers more opportunities to design proteins with new dynamics and functions.
Conclusions The recent development of computer resources and theoretical tools has greatly extended the scope and depth of the computational studies of protein folding. Consequently, metal-coupled protein folding, which has long been a challenging and extensively studied topic in experimental fields, has become the subjects of computational studies. In this review, we discussed several recent works that have developed computational tools to explore metal-coupled protein folding. These computational works have already shown great successes in describing many aspects of metal-coupled protein folding. However, relative to its abundance and biological importance, the computational works on metalloprotein www.sciencedirect.com
Simulations of metal ion coupled protein folding Li et al. 29
Figure 2
(a)
(b)
(c)
(d)
energy
folding landscape
composite landscape
nonspecific binding
specific binding Current Opinion in Structural Biology
Schematic diagram of the effects of metal binding on the energy landscapes of proteins. (a) Nonspecific binding leads to rugged composite landscape due to multiple competitive coordination modes; (b) binding creates a funnel on the composite landscape when the coordination landscape dominates the overall interactions; (c) binding induces a deeper composite landscape when the coordination landscape is consistent with the folding landscape; and (d) binding reshapes the folding landscape when the coordination landscape matches a metastable state on the folding funnel.
folding are still rare. In the future, developing multiscale methodologies, which can simultaneously model the metal-involved local interactions and the global motions of the polypeptide chain with high accuracy and efficiency, can further extend the computational studies of metalloprotein folding. In addition, similar methods can be used to model metal-coupled protein aggregations and nucleic acid folding [62–65].
Conflict of interest statement Nothing declared.
Acknowledgements This work was supported by National Basic Research Program of China (973 Program) (no. 2013CB834100), Natural Science Foundation of China (Grants 11334004, 11174134, and 11274157) and Jiangsu Province (BK2011546).
References and recommended reading Papers of particular interest, published within the period of review, have been highlighted as: of special interest of outstanding interest 1.
Gray HB: Biological inorganic chemistry at the beginning of the 21st century. Proc Natl Acad Sci USA 2003, 100:3563-3568.
2.
Wittung-Stafshede P: Role of cofactors in protein folding. Acc Chem Res 2002, 35:201-208.
3.
Weinkam P, Zimmermann J, Romesberg FE, Wolynes PG: The folding energy landscape and free energy excitations of cytochrome c. Acc Chem Res 2010, 43:652-660.
4.
Klug A: Towards therapeutic applications of engineered zinc finger proteins. FEBS Lett 2005, 579:892-894.
www.sciencedirect.com
5.
Cox EH, McLendon GL: Zinc-dependent protein folding. Curr Opin Chem Biol 2000, 4:162-165.
6.
Gomes GM, Wittung-Stafshede P: Metal ions, protein folding, and conformational states: an introduction. In Protein Folding and Metal Ions. Edited by Gomes GM, Wittung-Stafshede P. Boca Raton: CRC Press; 2011:3-11.
7.
Wilson CJ, Apiyo D, Wittung-Stafshede P: Role of cofactors in metalloprotein folding. Q Rev Biophys 2004, 37:285-314.
8.
Bren KL, Pecoraro VL, Gray HB: Metalloprotein folding. Inorg Chem 2004, 43:7894-7896.
9.
Palm-Espling ME, Niemiec MS, Wittung-Stafshede P: Role of metal in folding and stability of copper proteins in vitro. Biochim Biophys Acta 2012, 1823:1594-1603.
10. Palmieri M, Malgieri G, Russo L, Baglivo I, Esposito S, Netti F, Del Gatto A, de Paola I, Zaccaro L, Pedone PV et al.: Structural Zn(II) implies a switch from fully cooperative to partly downhill folding in highly homologous proteins. J Am Chem Soc 2013, 135:5220-5228. 11. Morleo A, Bonomi F, Iametti S, Huang VW, Kurtz DM Jr: Ironnucleated folding of a metalloprotein in high urea: resolution of metal binding and protein folding events. Biochemistry 2010, 49:6627-6634. 12. Blasie CA, Berg JM: Structure-based thermodynamic analysis of a coupled metal binding-protein folding reaction involving a zinc finger peptide. Biochemistry 2002, 41:15068-15073. 13. Miura T, Satoh T, Takeuchi H: Role of metal–ligand coordination in the folding pathway of zinc finger peptides. Biochim Biophys Acta 1998, 1384:171-179. 14. Lv C, Gao X, Li W, Xue B, Qin M, Burtnick LD, Zhou H, Cao Y, Robinson RC, Wang W: Single-molecule force spectroscopy reveals force-enhanced binding of calcium ions by gelsolin. Nat Commun 2014, 5:4623. 15. Stigler J, Rief M: Calcium-dependent folding of single calmodulin molecules. Proc Natl Acad Sci USA 2012, 109: 17814-17819. A single molecule optical tweezer experiment demonstrated the modulations of folding mechanisms of calmodulin by calcium concentrations. The kinetic measurements showed that the folding pathway with calcium Current Opinion in Structural Biology 2015, 30:25–31
30 Folding and binding
bound to the denatured state is the dominant pathway at high calcium concentrations. 16. Tang J, Kang SG, Saven JG, Gai F: Characterization of the cofactor-induced folding mechanism of a zinc-binding peptide using computationally designed mutants. J Mol Biol 2009, 389:90-102. 17. Li WF, Wang W, Takada S: Energy landscape views for interplays among folding, binding, and allostery of calmodulin domains. Proc Natl Acad Sci USA 2014, 111:10550-10555. A coarse-grained model provided an integrated description of folding, metal binding and allostery. The simulations demonstrated the shift of folding mechanisms of calmodulin domains with increasing calcium concentrations. 18. Weinkam P, Wolynes PG: The folding landscape of metalloproteins. In Protein Folding and Metal Ions. Edited by 247-273. Gomes GM, Wittung-Stafshede P. Boca Raton: CRC Press; 2011: 19. Weinkam P, Romesberg FE, Wolynes PG: Chemical frustration in the protein folding landscape: grand canonical ensemble simulations of cytochrome c. Biochemistry 2009, 48:2394-2402.
34. Dasmeh P, Kepp KP: Unfolding simulations of holomyoglobin from four mammals: identification of intermediates and betasheet formation from partially unfolded states. PLoS One 2013, 8:e80308. 35. Evoli S, Guzzi R, Rizzuti B: Dynamics and unfolding pathway of chimeric azurin variants: insights from molecular dynamics simulation. J Biol Inorg Chem 2013, 18:739-749. 36. Chen CJ, Huang YZ, Jiang XW, Xiao Y: Binding free-energy calculation of an ion–peptide complex by constrained dynamics. Phys Rev E 2013, 87:062705. 37. Li WF, Zhang J, Su Y, Wang J, Qin M, Wang W: Effects of zinc binding on the conformational distribution of the amyloid-beta peptide based on molecular dynamics simulations. J Phys Chem B 2007, 111:13814-13821. 38. Furlan S, Hureau C, Faller P, La Penna G: Modeling copper binding to the amyloid-beta peptide at different pH: toward a molecular mechanism for Cu reduction. J Phys Chem B 2012, 116:11899-11910.
20. Zong C, Wilson CJ, Shen T, Wittung-Stafshede P, Mayo SL, Wolynes PG: Establishing the entatic state in folding metallated Pseudomonas aeruginosa azurin. Proc Natl Acad Sci USA 2007, 104:3159-3164.
39. Furlana S, La Pennab G: Metal ions and protons compete for ligand atoms in disordered peptides: examples from computer simulations of copper binding to the prion tandem repeat. Coord Chem Rev 2012, 256:2234-2244.
21. Dudev T, Lim C: All-electron calculations of the nucleation structures in metal-induced zinc-finger folding: role of the peptide backbone. J Am Chem Soc 2007, 129:12497-12504. 22. Li W, Zhang J, Wang J, Wang W: Metal-coupled folding of Cys2His2 zinc-finger. J Am Chem Soc 2008, 130:892-900.
40. Sparta M, Shirvanyants D, Ding F, Dokholyan NV, Alexandrova AN: Hybrid dynamics simulation engine for metalloproteins. Biophys J 2012, 103:767-776. By combining the quantum mechanics with an all-atom discrete molecular dynamics method, they can conduct simulations for metalloproteins with high accuracy for a longer time duration.
23. Weinkam P, Zimmermann J, Sagle LB, Matsuda S, Dawson PE, Wolynes PG, Romesberg FE: Characterization of alkaline transitions in ferricytochrome c using carbon-deuterium infrared probes. Biochemistry 2008, 47:13470-13480.
41. Piana S, Klepeis JL, Shaw DE: Assessing the accuracy of physical models used in protein-folding simulations: quantitative evidence from long molecular dynamics simulations. Curr Opin Struct Biol 2014, 24:98-105.
24. Sakharov DV, Lim C: Zn protein simulations including charge transfer and local polarization effects. J Am Chem Soc 2005, 127:4921-4929.
42. Lin F, Wang RX: Systematic derivation of AMBER force field parameters applicable to zinc-containing systems. J Chem Theory Comput 2010, 6:1852-1870.
25. Zhu T, Xiao X, Ji C, Zhang JZH: A new quantum calibrated force field for zinc–protein complex. J Chem Theory Comput 2013, 9:1788-1798.
43. Burger SK, Lacasse M, Verstraelen T, Drewry J, Gunning P, Ayers PW: Automated parametrization of AMBER force field terms from vibrational analysis with a focus on functionalizing dinuclear zinc(II) scaffolds. J Chem Theory Comput 2012, 8: 554-562.
26. Huang YD, Shuai JW: Induced dipoles incorporated into allatom Zn protein simulations with multiscale modeling. J Phys Chem B 2013, 117:6138-6148. 27. Zhang J, Yang W, Piquemal JP, Ren P: Modeling structural coordination and ligand binding in zinc proteins with a polarizable potential. J Chem Theory Comput 2012, 8:13141324. 28. Wu R, Lu Z, Cao Z, Zhang Y: A transferable nonbonded pairwise force field to model zinc interactions in metalloproteins. J Chem Theory Comput 2011, 7:433-443. 29. Devereux M, Gresh N, Piquemal JP, Meuwly M: A supervised fitting approach to force field parametrization with application to the SIBFA polarizable force field. J Comput Chem 2014, 35:1577-1591. This paper provided a promising approach to the parameter fitting of the SIBFA polarizable force field which was one of the first potentials to model the metal-protein interactions considering explicit charge transfer and polarization effects. 30. Chakravorty DK, Parker TM, Guerra AJ, Sherrill CD, Giedroc DP, Merz KM Jr: Energetics of zinc-mediated interactions in the allosteric pathways of metal sensor proteins. J Am Chem Soc 2013, 135:30-33. 31. Chakravorty DK, Wang B, Lee CW, Giedroc DP, Merz KM Jr: Simulations of allosteric motions in the zinc sensor CzrA. J Am Chem Soc 2012, 134:3367-3376. 32. Reyes-Caballero H, Campanello GC, Giedroc DP: Metalloregulatory proteins: metal selectivity and allosteric switching. Biophys Chem 2011, 156:103-114. 33. Marino T, Russo N, Toscano M, Pavelka M: On the metal ion (Zn2+, Cu2+) coordination with beta-amyloid peptide: DFT computational study. Interdiscip Sci 2010, 2:57-69. Current Opinion in Structural Biology 2015, 30:25–31
44. Stote RH, Karplus M: Zinc binding in proteins and solution: a simple but accurate nonbonded representation. Proteins 1995, 23:12-31. 45. Li PF, Merz KM: Taking into account the ion-induced dipole interaction in the nonbonded model of ions. J Chem Theory Comput 2014, 10:289-297. A nonbonded model of metal ions with a 12-6-4 LJ type potential was proposed to describe the metal ion induced dipole interaction. This model can reproduce the hydration free energy and the coordination geometry for a number of metal ions and is readily incorporated into standard allatom force fields. 46. Pang YP, Xu K, Yazal JE, Prendergas FG: Successful molecular dynamics simulation of the zinc-bound farnesyltransferase using the cationic dummy atom approach. Protein Sci 2000, 9:1857-1865. 47. Takada S: Coarse-grained molecular simulations of large biomolecules. Curr Opin Struct Biol 2012, 22:130-137. An excellent survey of the recently developed and widely used coarsegrained models for the simulations of large biomolecules. The pros and cons as well as the application scope of the coarse-grained models were discussed. 48. Clementi C, Nymeyer H, Onuchic JN: Topological and energetic factors: what determines the structural details of the transition state ensemble and ‘‘en-route’’ intermediates for protein folding? An investigation for small globular proteins. J Mol Biol 2000, 298:937-953. 49. Li W, Terakawa T, Wang W, Takada S: Energy landscape and multiroute folding of topologically complex proteins adenylate kinase and 2ouf-knot. Proc Natl Acad Sci USA 2012, 109: 17789-17794. www.sciencedirect.com
Simulations of metal ion coupled protein folding Li et al. 31
50. Li W, Wolynes PG, Takada S: Frustration, specific sequence dependence, and nonlinearity in large-amplitude fluctuations of allosteric proteins. Proc Natl Acad Sci USA 2011, 108: 3504-3509.
59. Wang Y, Tang C, Wang E, Wang J: Exploration of multi-state conformational dynamics and underlying global functional landscape of maltose binding protein. PLoS Comput Biol 2012, 8:e1002471.
51. Li W, Yoshii H, Hori N, Kameda T, Takada S: Multiscale methods for protein folding simulations. Methods 2010, 52:106-114.
60. Kenzaki H, Koga N, Hori N, Kanada R, Li W, Okazaki K-i, Yao X-Q, Takada S: CafeMol: a coarse-grained biomolecular simulator for simulating proteins at work. J Chem Theory Comput 2011, 7:1979-1989.
52. Chan HS, Zhang Z, Wallin S, Liu Z: Cooperativity, local-nonlocal coupling, and nonnative interactions: principles of protein folding from coarse-grained models. Annu Rev Phys Chem 2011, 62:301-326. 53. Shental-Bechor D, Smith MT, Mackenzie D, Broom A, Marcovitz A, Ghashut F, Go C, Bralha F, Meiering EM, Levy Y: Nonnative interactions regulate folding and switching of myristoylated protein. Proc Natl Acad Sci USA 2012, 109:17839-17844. 54. Terakawa T, Higo J, Takada S: Multi-scale ensemble modeling of modular proteins with intrinsically disordered linker regions: application to p53. Biophys J 2014, 107:721-729. 55. Thirumalai D, Liu Z, O’Brien EP, Reddy G: Protein folding: from theory to practice. Curr Opin Struct Biol 2013, 23:22-29. 56. Wang Q, Liang KC, Czader A, Waxham MN, Cheung MS: The effect of macromolecular crowding, ionic strength and calcium binding on calmodulin dynamics. PLoS Comput Biol 2011, 7:e1002114.
61. Chu X, Gan L, Wang E, Wang J: Quantifying the topography of the intrinsic energy landscape of flexible biomolecular recognition. Proc Natl Acad Sci USA 2013, 110:E2342-E2351. Structure-based model was used to quantitatively characterize the energy landscapes of interfacial binding and monomeric folding for fifteen homodimers, and the interplay between the two kinds of energy landscapes can be well related to the association mechanism of the monomers. 62. Miller Y, Ma BY, Nussinov R: Metal binding sites in amyloid oligomers: complexes and mechanisms. Coord Chem Rev 2012, 256:2245-2252. 63. Bian Y, Tan C, Wang J, Sheng Y, Zhang J, Wang W: Atomistic picture for the folding pathway of a hybrid-1 type human telomeric DNA G-quadruplex. PLoS Comput Biol 2014, 10:e1003562.
57. Onuchic JN, Luthey-Schulten Z, Wolynes PG: Theory of protein folding: the energy landscape perspective. Annu Rev Phys Chem 1997, 48:545-600.
64. Li W, Hou XM, Wang PY, Xi XG, Li M: Direct measurement of sequential folding pathway and energy landscape of human telomeric G-quadruplex structures. J Am Chem Soc 2013, 135:6423-6426.
58. Okazaki K, Takada S: Dynamic energy landscape view of coupled binding and protein conformational change: inducedfit versus population-shift mechanisms. Proc Natl Acad Sci USA 2008, 105:11182-11187.
65. Shen X, Ding KK, Zhang FS: Ion effects of Sr2+, Cs+ and IS on DNA in aqueous solutions. Chem Phys Lett 2013, 574: 100-105.
www.sciencedirect.com
Current Opinion in Structural Biology 2015, 30:25–31
