HomePage | Research Overview | Research Highlights | Publications | Members | Employment Opportunities

The Vibrational Structures of Proteins

Protein Structure from Vibrational Spectroscopy

Infrared and Raman spectroscopies measure the vibrational frequencies of a group of bonded atoms in a molecule. The frequencies are determined by masses of the atoms, the force constants and the geometry of the molecule (1-2). Thus, it is possible, at least in principle, to determine structural properties of a molecular group, such as bond orders, bond length/angle, ionization state of ionizable moieties, etc., from vibrational frequencies. Furthermore, the interactions that take place between molecular groups when a small molecule binds to a protein, such as hydrogen bonding and other bond polarizing electrostatic interactions, geometry distortion from steric clashes, and the formation of new bonds, affect vibrational frequencies and, hence, are reported by the changes in frequencies. The accuracy of determining these parameters from vibrational frequencies is very high, generally much better than probes of protein structure such as X-ray crystallographic studies (i.e., bond length accuracy of up to 0.001 Anstroms). On the other hand, while vibrational spectroscopy is a very fine scale probe, it is not (currently) feasible to determine the full structure of a protein from its vibrational spectrum. There is thus a high degree of synergism amongst these experimental probes in the information they provide.

Enzymes bind their reacting substrates and distort them with the goal of accelerating the reaction between them. Often, this means lowering the transition state energy found in the corresponding solution reaction, although a particular enzyme may find an entirely new reaction pathway. In any case, the bond distortions imposed on bound substrates are generally small compared to the bond strength itself because non-covalent interactions are typically involved. However, it is precisely these interactions that bring about the large rate enhancements associated with enzymatic catalysis. Hence, measurement of the vibrational spectra of bound substrates, with its high resolution, is generally very valuable in understanding enzymatic catalysis on a molecular level.

It is a daunting technical problem to determine the vibrational spectrum of a specific bond embedded in a protein due to spectral crowding; the protein background spectrum typically overwhelms the spectrum of any particular bond. Early studies were able to exploit resonance Raman spectroscopy in the study of chromophoric substrates. Recent developments in FT-IR and Raman difference techniques (1-3) make it possible to measure the vibrational spectra of any small molecule when it is bound to a protein. In such an experiment, a protein spectrum is measured as is that of a protein complexed with the molecule. Subtraction of the two yields the spectrum of the bound ligand. Alternatively, an atom within a bond of interest is labeled with a stable isotope (2H, 13C, 15N, 18O, etc), and this shifts the frequency of the modes that involve the motion of the labeled atom. Subtraction of labeled and unlabeled protein spectra yields an 'isotopically edited' difference spectrum whereby only those modes involving the labeled atom show up in the spectrum. These general approaches have stimulated rapid progress in studies of enzymes using vibrational spectroscopy.

Purine Nucleoside Phosphorylase Activation

As an example, we recently used isotope-edited vibrational spectroscopy (4) on a transition state analogue complex of human PNP, PNP/ImmH-PO₄, to investigate how PNP activates the phosphate nucleophile near the transition state (Scheme 1). Our difference Raman and FTIR results show that the dianionic phosphate in the complex is forced into a unique bonding arrangement in which one of the P=O bonds is greatly polarized by enzyme active site interactions, such that it resembles a P=O bond that is about one quarter of the way towards forming a bridging P-O-C single P-O bond as shown below.

The difference FTIR (left panel) and difference Raman (right panel) spectra between P^16O4 and P^18O4 (a) in aqueous solution at pH 9.5 (dianionic) and (b) in the PNP/ImmH-PO4 complex (4 mM:4 mM:4 mM, pH 7.2). The positive bands are from P^16O4 and the negative bands are from P^18O4. The resolutions for IR and Raman spectra were 2 cm^-1 and 8 cm^-1, respectively.

PNP forms a dissociative transition state with fully developed carbocation character at C1' of the ribose ring. Our results suggest that during catalysis, the active site hydrogen bonds hold PO₄ tightly in place while the combination of the positive charge on C1' of ribose and H-bonds to the nucleophilic P=O bond oxygen are responsible for the polarization of the P=O bond to form a highly reactive oxygen nucleophile. The reaction coordinate is completed by migration of the C1' carbon cation toward the activated and immobilized phosphate group.

	Bound phosphate in the PNP•ImmH•PO4 complex as determined by vibrational spectroscopy.

References:

(1) "Non-Resonance Raman Difference Spectroscopy: A General Probe Of Protein Structure, Ligand Binding, Enzymatic Catalysis, and the Structures Of Other Biomacromolecules", Robert Callender and Hua Deng, Ann. Rev. of Biophysics and Biomolecular Structure 23, 215 (1994).

(2) "Vibrational Studies of Enzymatic Catalysis", Hua Deng and Robert Callender, in Infrared and Raman Spectroscopy of Biological Materials, eds. Hans-Ulrich Gremlich and Bing Yan, (Marcel Dekker, Inc., New York, pp.1-581), pp. 477-515 (2001).

(3) “Spectroscopic probes of hydride-transfer activation by enzymes”, Handbook of Hydrogen Transfer, Vol. 2 (Biological Aspects of Hydrogen Transfer), R.L. Schowen and Judith Klinman, eds., Wiley VCH, 2005, in press.

(4) "Activating The Phosphate Nucleophile At The Catalytic Site Of Purine Nucleoside Phosphorylase: a vibrational spectroscopic study", Hua Deng, Andrzej Lewandowicz, Vern L. Schramm, and Robert Callender, J. Am. Chem. Soc. 126, 9516-9517 (2004).