Apr 3, 1972 ... Walton, A. G., Deverry, M. J. & Koenig, J. L. (1970). Calcified Tissue Res.6, 162-
167. 5. Peticolas, W. L., Hibler, G. W., Lippert, J. L., Peterlin,.
Proc. Nat. Acad. Sci. USA Vol. 69, No. 6, pp. 1467-1469, June 1972
Conformationally Dependent Low-Frequency Motions of Proteins by Laser Raman Spectroscopy (a-chymotrypsin/protein conformation) K. G. BROWN, S. C. ERFURTH, E. W. SMALL, AND W. L. PETICOLAS Department of Chemistry, University of Oregon, Eugene, Oreg. 97403
Communicated by Irving M. Klotz, April 3, 1972 Low-frequency Raman bands (lower than ABSTRACT 50 cm-') exist in certain proteins. They are dependent upon the conformation of the protein molecule, but are relatively independent of the form of the sample, i.e., whether it is a film or a crystal. Low-frequency Raman spectra were obtained from samples of a-chymotrypsin that had been prepared in several ways. A peak at about 29 cm-' was found for all samples except the one that had been denatured with sodium dodecyl sulfate. Such low frequency motions must arise from vibrations that involve all, or very large portions, of the protein molecule.
In the past few years the technique of laser Raman spectroscopy has been used with considerable success to obtain the Raman active vibrations of several proteins (1-4). However, if one examines the published spectra, it is apparent that it has been impossible in the past to obtain the Raman bands in proteins that lie below 150 cm-'. The reason for this is that the scattering due to the Rayleigh component is too large for a double-grating monochromator to discriminate against. Recently, in this laboratory, we have shown how to obtain Raman spectra only a few wave numbers from the exciting line on synthetic polymers, such as polyethylene and poly-i-alanine, by the iodine filter technique (5-7) and a Spex double-grating monochromator. More recently, we have also found it possible to obtain these low frequency bands using a Cary triple-grating monochromator. For the work reported here, we have used both of these instruments and have obtained equivalent results on each. This has been of the greatest help in elimination of the possibility of experimental artifacts. From our work with these instruments it is possible to show, for the first time, that definite low-frequency motions exist in many common proteins and that these vibrations appear to be sensitive to the conformation of the protein. As we will discuss below, such low-frequency motions must arise from vibrations that involve either all or very large portions of the protein molecule. Thus, it appears from our measurements that large portions of the protein molecule are constantly undergoing a coherent periodic vibration. The existence of such vibrations is of considerable interest even though the exact assignment of the motion is not possible at present. Figs. la and b show the low-frequency Raman spectra of samples of a-chymotrypsin that were prepared in several ways. The purpose of preparing the protein in so many different ways is to show that the frequency is dependent on Abbreviation: SDS, sodium dodecyl sulfate. 1467
the protein molecule and not on the method of sample preparation. Furthermore, we wanted to observe the effect of slight modifications of the protein, such as deuteration and acylation, as well as the effect of more drastic modification, such as denaturation with sodium dodecyl sulfate (SDS). The samples studied included: (i) lyophilyzed powder as supplied by the manufacturer (Worthington), (ii) a sample that had been acylated with indole-acrolyl-imidazole by the method of Rossi and Bernhard (8) and then cast as a clear film, (iii) a film cast from pure protein in a concentrated solution, (iv) a film cast from D20 to see the effect of deuteration, (v) a film cast from a solution of the protein that had been denatured with SDS by the method of Rossi and Bernhard (8) and then separated from the SDS by a Biogel-P30 column, and finally, (vi) a single crystal. of a-chymotrypsin that was grown by the method of Siegler et al. (9). The protein films were always cast from a protein solution of sufficiently low pH (< 5) that self digestion did not occur. In every case, except the sample that had been denatured with SDS, a pronounced peak at about 29 cm-' is found. It is apparent that there is some splitting of the peak in the single crystal and also some slight change in the shape of the peak with deuteration and acylation. However, upon denaturation with SDS the peak at 29 cm-' vanishes. Rather intense Raman scattering throughout the region of 20-150 cm-' is observed on the denatured material, but it is broad and structureless-a fact that probably reflects the decrease in the order of the protein conformation. The fact that this low-frequency band is dependent on conformation and disappears upon denaturation of the protein is to be expected from the nature of such vibrations that are lower in frequency than 50 cm-'. Calculations on model polypeptides show that vibrations in this low-frequency range must of necessity involve large portions of the polymer chain, because localized vibrations inevitably lie at a higher frequency (10-12). If we consider a macromolecule such as a protein with one or more linear backbone peptide chains, and if we take the linear backbone chains to be informationally ordered one-dimensional crystals, then these low-frequency modes might be considered to be the macromolecular analogues of the low-frequency acoustical modes that are well known in polymers (10-12). These are vibrations in which the motion between adjacent residues is strongly coupled so that a simultaneous coherent displacement of the atomic nuclei occurs throughout all or many of the residues of the molecule. Thus, since these vibrations must involve the coherent vibrational motion of all, or at least a large part,
1468
Chexi*try:
Proc. Nat. Acad. Sci. USA 69 (1972)
Brown et al.
DENATURED
a I
I
1
b I,
0 100 soa2 FREQUENCY IN CM-, FIG. 1. (a) Low-frequency Raman spectra of a-chymotrypsin. The relative intensity of the Raman bands is of no significance as the spectra of protein samples have been separated for clarity. A band maximum at about 29 cm-' occurs in every sample except that which was denatured with SDS. (b) Low-frequency Raman spectra of a single crystal of a-chymotrypsin. 150
50 100 FREQUENCY IN CM
29
of the whole protein molecule, one would expect them to be sensitive to conformation and denaturation. For a perfectly ordered polymer, such as the a-helix or the all-trans structure of polyethylene in the single crystal, these vibrations are simply an accordion-like motion of the all-trans structure. In this latter case, the frequency of the vibration has been shown experimentally to be directly proportional to the inverse of the length of the all-trans chain (13). This is, of course, exactly what is observed for macroscopic Hookean springs. In structures of low symmetry, such as proteins, the exact description of these low-frequency vibrations is more difficult to assign. When we look at various, easily available proteins, we find that these low-frequency Raman bands may or may not be present. For example, we have found that the protein, very
0
150
pepsin, shows a strong Raman band at about 32 cm-, that disappears upon heat denaturation, and we have been unable to find even a weak broad band in the protein, carboxypeptidase. The purpose of this paper has been to show that lowfrequency Raman bands exist in certain proteins and that they are dependent on the conformation of the protein molecule but relatively independent of the form of the sample, i.e., whether the sample is a film or a crystal. Undoubtedly, more work needs to be done to elucidate the nature of these vibrations. In the past, attempts to calculate the normal modes of molecules as large as a-chymotrypsin would have been considered completely impossible. However, recently it has been possible to calculate the low-frequency motions of the polyalanine a-helix from a very simplified
Proc. Nat. Acad. Sci.- USA 69 (1972)
Low-Frequency Raman Spectra of Proteins
1469
model, which treats the peptide group as involving only a pair of masses (14). If this model can be extended to proteins, it should be possible to calculate the low-frequency motions that involve the whole protein chains using very simplified force fields.
6. Peticolas, W. L. (1971) Bull. Amer. Phys. Soc. Ser. II 16, 319-320. 7. Lippert, J. L., Hibler, G. W., Small, E. W. & Peticolas, W. L. (1971) in Light Scattering in Solids, ed. Balkanski, M. (University of Paris VI Flammarion Sciences, Paris). 8. Rossi, G. L. & Bernhard, S. A. (1971) J. Mol. Biol. 55,
The authors gratefully acknowledge the generosity of Prof. Joseph Nibler and the faculty at Oregon State University for allowing us to use their Cary triple grating Raman spectrograph. We also acknowledge the support of NIH Grant 5-R01GM 15547-05. K.G.B. is a PHS Post doctoral Fellow. 1. Tobin, M. C. (1968) Science 161, 68-70. 2. Lord, R. C. & Yu, N. T. (1970) J. Mol. Biol. 50, 509-524. 3. Lord, R. C. & Yu, N. T. (1970) J. Mol. Biol. 51, 203-213. 4. Walton, A. G., Deverry, M. J. & Koenig, J. L. (1970) Calcified Tissue Res. 6, 162-167. 5. Peticolas, W. L., Hibler, G. W., Lippert, J. L., Peterlin, A. & Olf, H. (1971) Appl. Phys. Lett. 18, 87-89.
215-230. 9. Siegler, P. B., Jefferey, B. A., Matthews, B. W. & Blow, D. M. (1966) J. Mol. Biol. 15, 175-192. 10. Small, E. W., Fanconi, B. & Peticolas, W. L. (1970) J. Chem. Phys. 52, 4369-4379. 11. Itoh, K. & Schimanouchi, T. (1970) Biopolymers 9, 383399. 12. Fanconi, B., Small, E. W. & Peticolas, W. L. (1971) Biopolymers 10, 1277-1298. 13. Peticolas, W. L., Hibler, G. W., Lippert, J. L., Peterlin, A. & Olf, H. (1971) Appl. Phys. Lett. 18, 87-89. 14. Fanconi, B. & Peticolas, W. L. (1971) Biopolymers 10, 2223-2229.
