Identification of Thermostabilizing Mutations for Membrane Proteins
Supporting Information Identification of Thermostabilizing Mutations for Membrane Proteins: Rapid Method Based on Statistical Thermodynamics Satoshi Yasuda,a,b,c Yuta Kajiwara,d Yuuki Takamuku,a Nanao Suzuki,a Takeshi Murata,*,a,b,e and Masahiro Kinoshita*,c a
Graduate School of Science, Chiba University, 1-33 Yayoi-cho, Inage, Chiba 263-8522, Japan
b
Molecular Chirality Research Center, Chiba University, 1-33 Yayoi-cho, Inage, Chiba 263-8522, Japan
c
Institute of Advanced Energy, Kyoto University, Uji, Kyoto 611-0011, Japan
d
Graduate School of Energy Science, Kyoto University, Uji, Kyoto 611-0011, Japan
e
JST, PRESTO, 1-33 Yayoi-cho, Inage, Chiba 263-8522, Japan
1. Integral equation theory The integral equation theory1 (IET) is based on classical statistical mechanics. In this theory, from the system partition function, various correlation functions are defined, and the basic equations satisfied by these functions are derived. The many-body correlations are approximately taken into account. The average value of a physical quantity is calculated for an infinitely large system and an infinitely large number of system configurations. In the case of bulk solvent of a single component, for example, the temperature, number density, and interaction potential form the input data. Once the basic equations are numerically solved and the correlation functions are obtained, we can calculate the microscopic structure and thermodynamic quantities. The solvent structure near a solute molecule and thermodynamic quantities of solvation, that is, changes in thermodynamic quantities upon solute insertion into the solvent, can also be calculated. A thermodynamic quantity of solvation is calculated for a fixed solute structure inserted into a fixed position within the solvent. When the solvent is a simple fluid with radial-symmetric potential, a complex solute molecule with a polyatomic structure (e.g., a protein) can directly be handled by the three-dimensional integral equation theory2−4 (3D-IET). However, the use of the 3D-IET is quite time consuming, which requires a special approach which we have developed (i.e., the morphometric approach). 2. Morphometric approach (MA) The solvation entropy should be calculated using the radial-symmetric integral equation theory (RSIET) in which the solvent effect as molecular assembly is fully taken into account. The computational burden, however, is too heavy to apply for calculation of a lot of mutant structures. This problem can be overcome by combing it with the morphometric approach5,6 (MA). The idea of the MA is to express S by the linear combination of only four geometric measures of a solute molecule: (S1)
SVH/kB = C1Vex + C2A + C3X + C4Y. S1
Here, Eq. S1 is referred to as the morphometric form, Vex is the excluded volume (EV) generated by the solute molecule, A is the solvent-accessible surface area, X and Y are the integrated mean and Gaussian curvatures of the solvent-accessible surface, respectively, and kB is the Boltzmann constant. The solvent-accessible surface is the surface that is accessible to the centers of solvent molecules. The EV is the volume that is enclosed by this surface. We note that C1 is independent of the solute-water interaction potential. Though SVH is influenced by all the four terms, C1Vex is the principal term at normal temperature and pressure. This is the reason for the fair insensitivity of SVH to the solute-water interaction potential. The contribution from the solvent molecules near the solute molecule is represented by the other three terms. In the MA, the solute shape enters SVH only via the four geometric measures. Therefore, the four coefficients (C1−C4) can be determined in simple geometries: They are calculated from the values of SVH of hard-sphere solutes immersed in hard spheres. The RSIET is employed in the calculation. The four coefficients are determined by the least square fitting applied to the following equation (i.e., Eq. S1 applied to hard-sphere solutes); SVH/kB = C1(4πR3/3) + C2(4πR2) + C3(4πR) + C4(4π), (S2) R = (dU + dS)/2. Here, dU denotes the diameter of a hard-sphere solute and sufficiently many different values of dU are considered (0.6dS ≤ dU ≤ 10dS). T is set at 298 K, and the number density of bulk water ρS is taken to be that of real water on the saturation curve, ρS = 0.0333 Å−3. Once the four coefficients are determined, SVH of the solute molecule with a prescribed structure is obtained from Eq. S1 only if the four geometric measures are calculated. 3. Discrimination of the native fold of GpA from a number of non-native folds using our free-energy function (FEF) In our earlier work,7 we succeeded in discriminating the native fold of transmembrane dimer of glycophorin A, GpA (PDB code: 1afo),8 from ~15000 non-native folds generated by Kokubo and Okamoto using a Monte Carlo simulation based on the replica-exchange method.9 This protein has two transmembrane α-helical segments: Figure S1 shows ribbon representations for the arrangement of the two segments in the native structure (NS) and in two representative non-native structures. The number of residues in each segment is 18. The structures stabilized in the second stage of the two-stage model were considered and their stabilities were evaluated using our free-energy function F. Since the method for calculating the energetic component Λ is changed as described in the main text, here we demonstrate that the discrimination is successful even when the modified version is employed. (F – FNS)/(kBT), X, and Y are plotted against the RMSD from the NS in Fig. S2. Here, X and Y are defined as X = Λ/(kBT0) − ΛNS/(kBT0) and Y = −S/kB − (−SNS/kB); F/(kBT) = X + Y (the subscript “NS” denotes the value for the NS). It is observed in Figure S2(A) that the NS is the most stable in terms of F, which shows the validity of the modified version. In the NS, no hydrogen bonds are formed between the two α-helical segments. For this reason, as observed in Figure S2(B), the NS is less stable than significantly many non-native structures in terms of Λ or X. Interestingly, Figure S2(C) indicates that the NS is the most stable in terms of −S or Y. Thus, the arrangement of α-helices is performed primarily through the entropic effect originating from the translational displacement of CH2 , CH3, and CH groups. It is verified that the entropic term is critically important factor in the second stage of the two-stage model for membrane protein folding. 4. Adenosine A2a receptor Human adenosine A2a receptor (A2aR) is an important GPCR involved in the control of various physiological activities, including regulation of glutamine and dopamine release in the brain.10 Specific inhibitors of the receptor are in advanced clinical trials for the treatment of Parkinson’s disease.11 Purified A2aR S2
in detergents is not very stable at room temperature. Half of the sample is denatured in 30 min incubation at 30°C .12 Several crystal structures of thermostabilizing A2aR mutants (T4 lysozyme [or b562RIL]-fusion mutants in intercellular loop 3 [ICL3] and alanine-scanning mutants) have been obtained by breakthroughs in protein engineering.12−15 We obtained the crystal structure (PDBID: 3VG9)16 of A2aR in complex with a mouse monoclonal-antibody Fab-fragment, which is used as a model structure in this work. 5. Thermal stabilities examined in terms of unfolding curves at a constant temperature The thermal stabilities of the wild type and its mutants can be evaluated from a different viewpoint. Figure S3 shows unfolding curves measured at 35°C for 30 min. It is observed that T88R is more stable than the wild type. 6. High conservation of S91 in G protein-coupled receptors We choose ten representative GPCRs from among those belonging to Class A whose wild-type structures are experimentally available and check their sequence alignment. We then find that S91 is conserved in many of them. See Figure S4.
Figure S1. Ribbon representations of transmembrane α-helical segments of GpA in the native structure viewed from two different angles (A) and for those in two representative non-native structures (B). They are drawn using PyMOL.17
S3
Figure S2. Relation between F (A), X (B), or Y (C) and RMSD from the native structure for GpA. Black asterisks: data points for the non-native structures. Red square: data point for the native structure.
S4
Figure S3. Thermal stabilities of purified A2aR mutants. This figure shows unfolding curves of the wild type and its mutants measured at 35°C for 30 min. The percentage of folded protein was calculated by the quotient of raw fluorescence measured at each time point divided by the maximal fluorescence as described elsewhere.18 The inset shows the half-life time at which the percentage of folded protein decreased to 50%. The experiments were independently carried out three times. A representative unfolding curve is shown here and the average value is adopted for each numerical data in the inset. The standard error for the half-life time is also given.
Figure S4. Sequence alignment of ten representative GPCRs belonging to class A; adenosine A2a receptor (A2aR), serotonin 5-HT2B receptor (HTR2B), β2-adrenergic receptor (ADRB2), histamine H1 receptor (HRH1), dopamine D3 receptor (DRD3), P2Y12 receptor (P2Y12), rat neurotensin receptor 1 (NTSR1), kappa opioid receptor (OPRK1), C-X-C chemokine receptor 4 (CXCR4), and C-C chemokine receptor 5 (CCR5). The residue numbers and secondary structures of A2aR are indicated on its sequence. The residues corresponding to T88 and S91 of A2aR are surrounded by dark green and light green squares, respectively. S91 is conserved except in NTSR1 and CCR5. By contrast, the conservation of T88 is much lower. S5
References (1) Hansen, J. -P.; McDonald, I. R. Theory of Simple Liquids, 3rd ed. (Academic, London) 2006. (2) Beglov, D.; Roux, B. J. Chem. Phys. 1995, 103, 360. (3) Ikeguchi, M.; Doi, J. J. Chem. Phys. 1995, 103, 5011. (4) Kinoshita, M. J. Chem. Phys. 2002, 116, 3493. (5) König, P. M.; Roth, R.; Mecke, K. R. Phys. Rev. Lett. 2004, 93, 160601. (6) Roth, R.; Harano, Y.; Kinoshita, M. Phys. Rev. Lett. 2006, 97, 078101. (7) Yasuda, S.; Oshima, H.; Kinoshita, M. J. Chem. Phys. 2012, 137, 135103. (8) MacKenzie, K. R.; Prestegard, J. H.; Engelman, D. M. Science 1997, 276, 131. (9) Kokubo, H.; Okamato, Y. Chem. Phys. Lett. 2004, 383, 397. (10) Fredholm, B. B.; Chen, J. F.; Masino, S. A.; Vaugeois, J. M. Annu. Rev. Pharmacol. Toxicol. 2005, 45, 385. (11) Müller, C. E.; Jacobson, K. A. Biochim. Biophys. Acta 2011, 1808, 1290. (12) Xu, F.; Wu, H.; Katritch, V.; Han, G. W.; Jacobson, K .A.; Gao, Z. G.; Cherezov, V.; Stevens, R. C. Science 2011, 332. 322. (13) Doré, A. S.; Robertson, N.; Errey, J. C.; Ng, I.; Hollenstein, K.; Tehan, B.; Hurrell, E.; Bennett, K.; Congreve, M.; Magnani, F.; Tate, C. G.; Weir, M.; Marshall, F. H. Structure 2011, 19, 1283. (14) Lebon, G.; Warne, T.; Edwards, P. C.; Bennett, K.; Langmead, C. J.; Leslie, A. G. W.; Tate, C. G. Nature 2011, 474. 521. (15) Liu, W.; Chun, E.; Thompson, A. A.; Chubukov, P.; Xu, F.; Katritch, V.; Han, G. W.; Roth, C. B.; Heitman, L. H.; Ijzerman, A. P.; Cherezov, V.; Stevens, R.C. Science, 2012, 337, 232. (16) Hino, T.; Arakawa, T.; Iwanari, H.; Yurugi-Kobayashi, T.; Ikeda-Suno, C.; Nakada-Nakura, Y.; Kusano-Arai, O.; Weyand, S.; Shimamura, T.; Nomura, N.; Cameron, A. D.; Kobayashi, T.; Hamakubo, T.; Iwata, S.; Murata, T. Nature 2012, 482, 237. (17) The PyMol Molecular Graphics System, Version 1.3, Schrödinger, LLC. (18) Sonoda, Y.; Newstead, S.; Hu, N. J.; Alguel, Y.; Nji, E.; Beis, K.; Yashiro, S.; Lee, C.; Leung, J.; Cameron, A. D.; Byrne, B.; Iwata, S.; Drew, D. Structure 2011, 19, 17.
S6
Graduate School of Science, Chiba University, 1-33 Yayoi-cho, Inage, Chiba 263-8522, Japan
b
Molecular Chirality Research Center, Chiba University, 1-33 Yayoi-cho, Inage, Chiba 263-8522, Japan
c
Institute of Advanced Energy, Kyoto University, Uji, Kyoto 611-0011, Japan
d
Graduate School of Energy Science, Kyoto University, Uji, Kyoto 611-0011, Japan
e
JST, PRESTO, 1-33 Yayoi-cho, Inage, Chiba 263-8522, Japan
1. Integral equation theory The integral equation theory1 (IET) is based on classical statistical mechanics. In this theory, from the system partition function, various correlation functions are defined, and the basic equations satisfied by these functions are derived. The many-body correlations are approximately taken into account. The average value of a physical quantity is calculated for an infinitely large system and an infinitely large number of system configurations. In the case of bulk solvent of a single component, for example, the temperature, number density, and interaction potential form the input data. Once the basic equations are numerically solved and the correlation functions are obtained, we can calculate the microscopic structure and thermodynamic quantities. The solvent structure near a solute molecule and thermodynamic quantities of solvation, that is, changes in thermodynamic quantities upon solute insertion into the solvent, can also be calculated. A thermodynamic quantity of solvation is calculated for a fixed solute structure inserted into a fixed position within the solvent. When the solvent is a simple fluid with radial-symmetric potential, a complex solute molecule with a polyatomic structure (e.g., a protein) can directly be handled by the three-dimensional integral equation theory2−4 (3D-IET). However, the use of the 3D-IET is quite time consuming, which requires a special approach which we have developed (i.e., the morphometric approach). 2. Morphometric approach (MA) The solvation entropy should be calculated using the radial-symmetric integral equation theory (RSIET) in which the solvent effect as molecular assembly is fully taken into account. The computational burden, however, is too heavy to apply for calculation of a lot of mutant structures. This problem can be overcome by combing it with the morphometric approach5,6 (MA). The idea of the MA is to express S by the linear combination of only four geometric measures of a solute molecule: (S1)
SVH/kB = C1Vex + C2A + C3X + C4Y. S1
Here, Eq. S1 is referred to as the morphometric form, Vex is the excluded volume (EV) generated by the solute molecule, A is the solvent-accessible surface area, X and Y are the integrated mean and Gaussian curvatures of the solvent-accessible surface, respectively, and kB is the Boltzmann constant. The solvent-accessible surface is the surface that is accessible to the centers of solvent molecules. The EV is the volume that is enclosed by this surface. We note that C1 is independent of the solute-water interaction potential. Though SVH is influenced by all the four terms, C1Vex is the principal term at normal temperature and pressure. This is the reason for the fair insensitivity of SVH to the solute-water interaction potential. The contribution from the solvent molecules near the solute molecule is represented by the other three terms. In the MA, the solute shape enters SVH only via the four geometric measures. Therefore, the four coefficients (C1−C4) can be determined in simple geometries: They are calculated from the values of SVH of hard-sphere solutes immersed in hard spheres. The RSIET is employed in the calculation. The four coefficients are determined by the least square fitting applied to the following equation (i.e., Eq. S1 applied to hard-sphere solutes); SVH/kB = C1(4πR3/3) + C2(4πR2) + C3(4πR) + C4(4π), (S2) R = (dU + dS)/2. Here, dU denotes the diameter of a hard-sphere solute and sufficiently many different values of dU are considered (0.6dS ≤ dU ≤ 10dS). T is set at 298 K, and the number density of bulk water ρS is taken to be that of real water on the saturation curve, ρS = 0.0333 Å−3. Once the four coefficients are determined, SVH of the solute molecule with a prescribed structure is obtained from Eq. S1 only if the four geometric measures are calculated. 3. Discrimination of the native fold of GpA from a number of non-native folds using our free-energy function (FEF) In our earlier work,7 we succeeded in discriminating the native fold of transmembrane dimer of glycophorin A, GpA (PDB code: 1afo),8 from ~15000 non-native folds generated by Kokubo and Okamoto using a Monte Carlo simulation based on the replica-exchange method.9 This protein has two transmembrane α-helical segments: Figure S1 shows ribbon representations for the arrangement of the two segments in the native structure (NS) and in two representative non-native structures. The number of residues in each segment is 18. The structures stabilized in the second stage of the two-stage model were considered and their stabilities were evaluated using our free-energy function F. Since the method for calculating the energetic component Λ is changed as described in the main text, here we demonstrate that the discrimination is successful even when the modified version is employed. (F – FNS)/(kBT), X, and Y are plotted against the RMSD from the NS in Fig. S2. Here, X and Y are defined as X = Λ/(kBT0) − ΛNS/(kBT0) and Y = −S/kB − (−SNS/kB); F/(kBT) = X + Y (the subscript “NS” denotes the value for the NS). It is observed in Figure S2(A) that the NS is the most stable in terms of F, which shows the validity of the modified version. In the NS, no hydrogen bonds are formed between the two α-helical segments. For this reason, as observed in Figure S2(B), the NS is less stable than significantly many non-native structures in terms of Λ or X. Interestingly, Figure S2(C) indicates that the NS is the most stable in terms of −S or Y. Thus, the arrangement of α-helices is performed primarily through the entropic effect originating from the translational displacement of CH2 , CH3, and CH groups. It is verified that the entropic term is critically important factor in the second stage of the two-stage model for membrane protein folding. 4. Adenosine A2a receptor Human adenosine A2a receptor (A2aR) is an important GPCR involved in the control of various physiological activities, including regulation of glutamine and dopamine release in the brain.10 Specific inhibitors of the receptor are in advanced clinical trials for the treatment of Parkinson’s disease.11 Purified A2aR S2
in detergents is not very stable at room temperature. Half of the sample is denatured in 30 min incubation at 30°C .12 Several crystal structures of thermostabilizing A2aR mutants (T4 lysozyme [or b562RIL]-fusion mutants in intercellular loop 3 [ICL3] and alanine-scanning mutants) have been obtained by breakthroughs in protein engineering.12−15 We obtained the crystal structure (PDBID: 3VG9)16 of A2aR in complex with a mouse monoclonal-antibody Fab-fragment, which is used as a model structure in this work. 5. Thermal stabilities examined in terms of unfolding curves at a constant temperature The thermal stabilities of the wild type and its mutants can be evaluated from a different viewpoint. Figure S3 shows unfolding curves measured at 35°C for 30 min. It is observed that T88R is more stable than the wild type. 6. High conservation of S91 in G protein-coupled receptors We choose ten representative GPCRs from among those belonging to Class A whose wild-type structures are experimentally available and check their sequence alignment. We then find that S91 is conserved in many of them. See Figure S4.
Figure S1. Ribbon representations of transmembrane α-helical segments of GpA in the native structure viewed from two different angles (A) and for those in two representative non-native structures (B). They are drawn using PyMOL.17
S3
Figure S2. Relation between F (A), X (B), or Y (C) and RMSD from the native structure for GpA. Black asterisks: data points for the non-native structures. Red square: data point for the native structure.
S4
Figure S3. Thermal stabilities of purified A2aR mutants. This figure shows unfolding curves of the wild type and its mutants measured at 35°C for 30 min. The percentage of folded protein was calculated by the quotient of raw fluorescence measured at each time point divided by the maximal fluorescence as described elsewhere.18 The inset shows the half-life time at which the percentage of folded protein decreased to 50%. The experiments were independently carried out three times. A representative unfolding curve is shown here and the average value is adopted for each numerical data in the inset. The standard error for the half-life time is also given.
Figure S4. Sequence alignment of ten representative GPCRs belonging to class A; adenosine A2a receptor (A2aR), serotonin 5-HT2B receptor (HTR2B), β2-adrenergic receptor (ADRB2), histamine H1 receptor (HRH1), dopamine D3 receptor (DRD3), P2Y12 receptor (P2Y12), rat neurotensin receptor 1 (NTSR1), kappa opioid receptor (OPRK1), C-X-C chemokine receptor 4 (CXCR4), and C-C chemokine receptor 5 (CCR5). The residue numbers and secondary structures of A2aR are indicated on its sequence. The residues corresponding to T88 and S91 of A2aR are surrounded by dark green and light green squares, respectively. S91 is conserved except in NTSR1 and CCR5. By contrast, the conservation of T88 is much lower. S5
References (1) Hansen, J. -P.; McDonald, I. R. Theory of Simple Liquids, 3rd ed. (Academic, London) 2006. (2) Beglov, D.; Roux, B. J. Chem. Phys. 1995, 103, 360. (3) Ikeguchi, M.; Doi, J. J. Chem. Phys. 1995, 103, 5011. (4) Kinoshita, M. J. Chem. Phys. 2002, 116, 3493. (5) König, P. M.; Roth, R.; Mecke, K. R. Phys. Rev. Lett. 2004, 93, 160601. (6) Roth, R.; Harano, Y.; Kinoshita, M. Phys. Rev. Lett. 2006, 97, 078101. (7) Yasuda, S.; Oshima, H.; Kinoshita, M. J. Chem. Phys. 2012, 137, 135103. (8) MacKenzie, K. R.; Prestegard, J. H.; Engelman, D. M. Science 1997, 276, 131. (9) Kokubo, H.; Okamato, Y. Chem. Phys. Lett. 2004, 383, 397. (10) Fredholm, B. B.; Chen, J. F.; Masino, S. A.; Vaugeois, J. M. Annu. Rev. Pharmacol. Toxicol. 2005, 45, 385. (11) Müller, C. E.; Jacobson, K. A. Biochim. Biophys. Acta 2011, 1808, 1290. (12) Xu, F.; Wu, H.; Katritch, V.; Han, G. W.; Jacobson, K .A.; Gao, Z. G.; Cherezov, V.; Stevens, R. C. Science 2011, 332. 322. (13) Doré, A. S.; Robertson, N.; Errey, J. C.; Ng, I.; Hollenstein, K.; Tehan, B.; Hurrell, E.; Bennett, K.; Congreve, M.; Magnani, F.; Tate, C. G.; Weir, M.; Marshall, F. H. Structure 2011, 19, 1283. (14) Lebon, G.; Warne, T.; Edwards, P. C.; Bennett, K.; Langmead, C. J.; Leslie, A. G. W.; Tate, C. G. Nature 2011, 474. 521. (15) Liu, W.; Chun, E.; Thompson, A. A.; Chubukov, P.; Xu, F.; Katritch, V.; Han, G. W.; Roth, C. B.; Heitman, L. H.; Ijzerman, A. P.; Cherezov, V.; Stevens, R.C. Science, 2012, 337, 232. (16) Hino, T.; Arakawa, T.; Iwanari, H.; Yurugi-Kobayashi, T.; Ikeda-Suno, C.; Nakada-Nakura, Y.; Kusano-Arai, O.; Weyand, S.; Shimamura, T.; Nomura, N.; Cameron, A. D.; Kobayashi, T.; Hamakubo, T.; Iwata, S.; Murata, T. Nature 2012, 482, 237. (17) The PyMol Molecular Graphics System, Version 1.3, Schrödinger, LLC. (18) Sonoda, Y.; Newstead, S.; Hu, N. J.; Alguel, Y.; Nji, E.; Beis, K.; Yashiro, S.; Lee, C.; Leung, J.; Cameron, A. D.; Byrne, B.; Iwata, S.; Drew, D. Structure 2011, 19, 17.
S6
