Solvent caging of internal motions in myoglobin at low temperatures

Abstract

Experimental and simulation studies have reported the presence of a transition in the internal dynamics of proteins at 220 K. This transition has been correlated with the onset of activity in several proteins. The role of the solvent in the dynamical transition has been the subject of increased attention. Here simulation techniques are used to distinguish dynamical features inherent to the protein energy landscape from those induced by the surrounding solvent. The present results indicate that the protein dynamical transition primarily affects the side-chains on the outer layers of the protein. Moreover, the results indicate that the solvent restrains protein motions at low temperatures.

---

Introduction

Various experimental techniques such as neutron scattering, Mössbauer spectroscopy, and X-ray scattering have shown the presence of a temperature-dependent transition in protein dynamics at around 180–220 K.^1–16 This dynamical transition has also been reproduced using Molecular Dynamics (MD) simulation techniques.^17–20 Experiments have shown that in several proteins biological function ceases below the dynamical transition.^5,14,22 The transition is thought to be associated with a transition in the dynamics of protein from harmonic to anharmonic, and is often referred to as the protein glass transition.^16,23 A number of experiments have indicated that when a protein is solvated the dynamical transition is strongly coupled to the surrounding solvent.^{1,5,7,24–27}

The observed dependence of the dynamical transition behavior on the solvent composition leads to the question of the role of solvent in the dynamical transition.^7,18 Simulations have demonstrated that proteins in vacuo undergo the dynamical transition.^17,18,21 However, whether solvent drives the dynamical transition in a hydrated protein is still open to question. Here we probe solvent effects by using dual heatbath methods to set the protein and its solvent at different temperatures. This approach enables the distinction of features inherent to the protein energy landscape from features due to properties of the solvent. Dual heatbath methods differ from standard simulation techniques in which the system has a constant and uniform temperature, in that different parts are set at different temperatures. Fig. 1 presents a schematic diagram of how such simulations are performed. The method involves using Nosé–Hoover–Chain (NHC) thermostats to set the temperatures of the different parts.^28,29 A similar approach, using the Nosé–Hoover thermostat, has previously been used by Vitkup et al. and confirmed the importance of solvent effects on internal fluctuations.³⁰ The present results extend this work and examine in greater detail the effect of the solvent on the dynamical transition in proteins.

Fig. 1 Schematic of dual-heatbath methods.

Our results indicate the protein side-chain atoms to be more mobile than the backbone atoms and that the protein layers closest to the solvent are the most affected by the dynamical transition, whereas the protein core is seen not to show any transition in fast dynamics along the observed temperature range. The results of the dual heatbath simulations show that cold solvent effectively cages protein motions.

Methods

The NHC algorithm was implemented and added in the CHARMM package, version 27b2.^31,32 Our system consisted of myoglobin surrounded by one shell of solvent (492 water molecules) as previously used in ref. 30. The details of the simulation were the same as in ref. 30, except for the use of the NHC thermostat. The system was prepared by a series of minimizations followed by dynamics runs at 180 K using harmonic constraints to ensure that the protein structure is not affected by the solvent equilibration. It was then brought to 180 K and equilibrated for 20 ps ready for the simulation runs. For each simulation run, protein and solvent were brought to their respective temperatures and equilibrated for 40 ps. A 200 ps production run was then performed and the trajectories recorded for subsequent analysis. Sets of simulations were performed where the protein (or solvent) temperature was held at a given temperature and the temperature of the other component varied between 80 and 300 K or between 180 and 300 K. These temperature ranges were chosen so as to cover the expected transition regime.

Dual heatbath molecular dynamics involves running a simulation with part of the system (here, the protein) set at one temperature and the other part (here, the solvent) at another. In the NHC method, used here, the different parts of the system are each regulated not by one but by two heatbaths, the first one regulating the system and the second regulating the first heatbath. NHC has the advantage over the Nosé–Hoover algorithm in that it reproduces exact canonical behavior and is more stable. The characteristic time adopted for the thermostat motion was 0.2 ps, a value commonly used for condensed phase molecular systems.

Results & discussion

In order to validate our simulations the dynamical transition was first reproduced on a set of control runs with the protein and solvent at the same temperature. The results are shown in Fig. 2. As can be seen, the experimentally observed transition in mean square deviation, 〈u²〉, at ∼220 K is well reproduced. 〈u²〉 is seen to increase relatively slowly up to ∼220 K, whereas beyond this temperature it increases more sharply with temperature, giving rise to the characteristic dynamical transition feature.

Fig. 2 Mean-square fluctuations, 〈u²〉 of the protein non-hydrogen atoms for different sets of simulations.

The data from this initial set of runs was analysed to investigate which parts of the proteins are subject to the dynamical transition. To do this the 〈u²〉 of individual sets of heavy atoms were examined. The 〈u²〉 of sidechain atoms was found to be 6 times greater than the backbone 〈u²〉 at 80 K, and twice as large at 300 K (data not shown). Fig. 3 shows the side-chain 〈u²〉 in the control simulations as a function of temperature for shells of atoms at different distances from the protein centre of mass. The inner shells do not show any dynamical transition feature, their 〈u²〉 increasing linearly with temperature. In contrast, the outer shells show a marked increase at 220 K, the outermost shell being the most affected. Thus, the atoms found to be most influenced by the dynamical transition are the side-chain atoms on the outer layers of the protein i.e., the protein atoms interacting with the solvent shell.

Fig. 3 Mean-square fluctuations 〈u²〉 of the protein side-chain heavy atoms for 5 different shells, each 4 Å thick (except for the inner shell (8 Å) and outer shell (6 Å)). The inset shows the difference in slopes of lines fitted below and above 220 K as a function of distance from the protein centre of mass.

Fig. 2 also presents the protein fluctuations calculated from the dual heatbath simulations, performed fixing the temperature of one component below the dynamical transition while varying the temperature of the other component. Fixing the solvent temperature at 80 or 180 K suppresses the dynamical transition, the protein 〈u²〉 increasing linearly with temperature up to 300 K. Therefore, low temperature solvent cages the protein dynamics. Fig. 2 also shows that holding the protein temperature constant at 80 or 180 K and varying the solvent temperature also abolishes the dynamical transition behavior in the protein. The solvent does increase the protein fluctuations to some extent even when the latter is held constant at 180 K. However, this effect is of much lesser extent than reported in ref. 30. In summary then, holding either component at a low temperature suppresses the protein dynamical transition.

Conclusions

Previous experimental, as well as, computer simulation studies have shown the presence of a transition in the dynamics of hydrated proteins at 220 K. The present results show that the sidechains on the outer layers of the protein are most affected by the dynamical transition. These atoms are in contact with the solvent. The possible role of the solvent in the dynamical transition has been mentioned in previous work.^1,19,25,30 The dual heatbath technique used in the present approach enables the further investigation of the role of the solvent. The present data indicates that the cold (80 and 180 K) solvent effectively cages protein dynamics over the whole range of protein temperatures examined (from 80 up to 180 K). This indicates the important role of solvent in influencing protein dynamics. Work is in progress to further investigate the role of the solvent so as to find an answer to the burning question: ‘Is the solvent driving the dynamical transition in a protein?’.

References

1. J. Fitter, R. E. Lechner and N. A. Dencher, Biophys. J., 1997, 73, 2126.
2. W. Doster, S. Cusack and W. Petry, Nature, 1989, 337, 754.
3. W. Doster, S. Cusack and W. Petry, Phys. Rev. Lett., 1990, 65, 1080.
4. R. V. Dunn, V. Reat, J. Finney, M. Ferrand, J. C. Smith and R. M. Daniel, Biochem. J., 2000, 346 Pt 2, 355.
5. M. Ferrand, A. J. Dianoux, W. Petry and G. Zaccai, Proc. Natl. Acad. Sci. U. S. A., 1993, 90, 9668.
6. R. M. Daniel, J. C. Smith, M. Ferrand, S. Hery, R. Dunn and J. L. Finney, Biophys. J., 1998, 75, 2504.
7. V. Reat, R. Dunn, M. Ferrand, J. L. Finney, R. M. Daniel and J. C. Smith, Proc. Natl. Acad. Sci. U. S. A., 2000, 97, 9961.
8. D. J. Bicout and G. Zaccai, Biophys. J., 2001, 80, 1115.
9. M. Diehl, W. Doster, W. Petry and H. Schober, Biophys. J., 1997, 73, 2726.
10. F. Parak, E. N. Frolov, R. L. Mossbauer and V. I. Goldanskii, J. Mol. Biol., 1981, 145, 825.
11. S. G. Cohen, E. R. Bauminger, I. Nowik, S. Ofer and J. Yariv, Phys. Rev. Lett., 1981, 46, 1244.
12. E. W. Knapp, S. F. Fischer and F. Parak, J. Am. Chem. Soc., 1982, 86, 5042.
13. R. F. Tilton, Jr., J. C. Dewan and G. A. Petsko, Biochemistry, 1992, 31, 2469.
14. B. F. Rasmussen, A. M. Stock, D. Ringe and G. A. Petsko, Nature, 1992, 357, 423.
15. A. L. Lee and A. J. Wand, Nature, 2001, 411, 501.
16. J. L. Green, J. Fan and C. A. Angell, J. Phys. Chem., 1994, 98, 13780.
17. J. Smith, K. Kuczera and M. Karplus, Proc. Natl. Acad. Sci. U. S. A., 1990, 87, 1601.
18. J. A. Hayward and J. C. Smith, Biophys. J., 2002, 82, 1216.
19. A. R. Bizzarri, A. Paciaroni and S. Cannistraro, Phys. Rev. E. Stat. Phys., 2000, 62, 3991.
20. P. J. Steinbach and B. R. Brooks, Proc. Natl. Acad. Sci. U. S. A., 1993, 90, 9135.
21. P. J. Steinbach and B. R. Brooks, Proc. Natl. Acad. Sci. U. S. A., 1996, 93, 55.
22. F. Parak, E. N. Frolov, A. A. Kononenko, R. L. Mossbauer, V. I. Goldanskii and A. B. Rubin, FEBS Lett., 1980, 117, 368.
23. C. A. Angell, Science, 1995, 267, 1924.
24. A. Paciaroni, S. Cinelli and G. Onori, Biophys. J., 2002, 83, 1157.
25. J. Fitter, Biophys. J., 1999, 76, 1034.
26. M. M. Teeter, A. Yamano, B. Stec and U. Mohanty, Proc. Natl. Acad. Sci. U. S. A., 2001, 98, 11242.
27. L. Cordone, M. Ferrand, E. Vitrano and G. Zaccai, Biophys. J., 1999, 76, 1043.
28. G. J. Martyna, M. E. Tuckerman, D. J. Tobias and M. L. Klein, Mol. Phys., 1996, 87, 1117.
29. M. E. Tuckerman and G. J. Martyna, J. Phys. Chem., 2000, 194, 159.
30. D. Vitkup, D. Ringe, G. A. Petsko and M. Karplus, Nat. Struct. Biol., 2000, 7, 34.
31. B. R. Brooks, R. E. Bruccoleri, B. D. Olafson, D. J. States, S. Swaminathan and M. Karplus, J. Comput. Biol., 1983, 4, 187.
32. A. D. MacKerell, Jr., D. Bashford, M. Bellott, R. L. Dunbrack, Jr., J. D. Evanseck, M. J. Field, S. Fischer, J. Gao, H. Guo, S. Ha, D. Joseph-McCarthy, L. Kuchnir, K. Kuczera, F. T. K. Lau, C. Mattos, S. Michnick, T. Ngo, D. T. Nguyen, B. Prodhom, W. E. Reiher, III, B. Roux, M. Schlenkrich, J. C. Smith, R. Stote, J. Straub, M. Watanabe, J. Wiorkiewicz-Kuczera, D. Yin and M. Karplus, J. Phys. Chem., 1998, 102, 3586.

---

Footnote

† This paper was originally presented as a poster at the Faraday Discussion 122 meeting.

---