Two-dimensional infrared spectroscopy has been used for the first time to study the conformation of proteins by hydrogen–deuterium exchange. In order to generate the two-dimensional synchronous and asynchronous maps, hydrogen–deuterium exchange of the amide protons of proteins deposited on attenuated total reflection crystals has been used as an external perturbation. Owing to the fact that the amide protons associated with each conformation are not exchanged at the same rate, the different conformational contributions of the amide bands could be separated. The use of different sampling time domains turned out to be very helpful in order to separate more efficiently the fast kinetics from the slower ones. The results obtained on myoglobin show that this method is particularly useful to unravel the different components under the poorly resolved amide I, II, and II' bands of proteins. The analysis of the synchronous and asynchronous maps of myoglobin demonstrates that the amide I band of this protein is composed of at least four components that could be assigned to α-helical, intermolecular β-sheet, β-turn, and random coil conformations.

The molecular orientation is generally expressed by an "order parameter," (P,), which depends on both the angular position and the shape of the orientation distribution. This parameter is an average made over all orientations of the structural units studied in a sample and, consequently, a given (P,) value can correspond to different orientation distributions. In this article, model distributions are used to show the relationship between the shape, width, and angular position of the center of the orientation distribution on the (P,) coefficient, for the case where the distribution of the molecular chains exhibits cylindrical symmetry with respect to the reference direction. A significant difference is observed between the order parameters calculated for distributions of Gaussian and Lorentzian shapes with similar width at half-height. The variation of the (P,) coefficient as a function of the width at half-height, W,,,, and of the position of the center of the distribution, 0,, is analyzed. Figures showing the range of W,,,-0, coordinates that can correspond to a given (P,) value are presented. As an example, the influence on the order parameter of the disorder between the different domains of phospholipid samples (mosaic spread) and of the conformational disorder in the acyl chains of these molecules is also studied. This example permits the evaluation of the magnitude of the errors that can be introduced in calculations of the tilt angle of the molecular chains in the case of distributions of finite widths or of bimodal character.Key words: orientation, orientation function, phospholipid bilayers, conformational disorder, mosaic spread.Resume : L'orientation molCculaire est gCnCralement exprimCe par un parambtre d'ordre, (P,), dont la valeur dCpend B la fois de la position angulaire et de la forme de la distribution d'orientation. Ce parambtre d'ordre constitue une moyenne faite sur toutes les orientations des unites structurales CtudiCes dans un Cchantillon et, en constquence, une valeur de (P,) donnCe peut correspondre 2 diffkrentes distributions d'orientation. Des distributions modkles sont utilisCes afin de montrer l'influence de la forme, de la largeur et de la position angulaire du centre de la distribution d'orientation sur le coefficient (P,), pour le cas oh les chaines molCculaires posskdent une symCtrie cylindrique par rapport B la direction de rCfCrence. Une difference significative est observCe entre les parambtres d'ordre calculCs pour des distributions de formes Gaussienne et Lorentzienne de mCme largeur i mi-hauteur. La variation du coefficient (P,) en fonction de la largeur 2 mi-hauteur, WIl2, et de la position du centre de la distribution, €Ic, est analysCe. Des figures montrant la plage de coordonnCes W,,,-0, pouvant correspondre B une valeur de parambtre d'ordre donnCe sont prCsentCes. A titre d'exemple, la variation du coefficient (P,) en fonction du dCsordre entre les diffkrents domaines dlCchantillons de phospholipides et en fonction du dCsordre conformationnel dans les chaines ...

The temperature-induced gel-to-liquid crystalline phase transition of dipalmitoylphosphatidylcholine (DPPC) is characterized by a shift towards high frequencies and an increase of the width of the bands due to the methylene stretching vibrations. These spectral modifications are frequently used to measure the conformational order of lipid acyl chains. However, it is not clear whether these bands contain two spectral components due to trans and gauche conformers or whether they gradually shift with temperature. The temperature-induced gel-to-liquid crystalline phase transition of DPPC has been investigated in the present study by two-dimensional infrared (2D-IR) correlation spectroscopy. Our results show that each methylene stretching band in both the synchronous and the asynchronous maps is characterized by two peaks. The same pattern is also observed when the temperature range is restricted to the gel phase. These results were compared to those obtained by spectral simulations using either a single band that shifts in frequency and gets broader with the increase of temperature (shifting-band model), simulating a continuously evolving one-phase system, or a band made of two components (two-band model), simulating the trans and gauche spectral contributions of a two-phase system. The results obtained for the asynchronous maps of the simulated spectra indicate clearly that the experimental results cannot be modeled by a pure two-phase system and are best simulated by the shifting-band model.

The effect of bovine myelin basic protein (MBP) on dimyristoylphosphatidic acid (DMPA) and phosphatidic acid prepared from egg yolk phosphatidylcholine (EPA) has been investigated by transmission and attenuated total reflectance (ATR) Fourier transform infrared spectroscopy. Interaction of MBP with DMPA and EPA dispersions decreases the lipid acyl chain conformational disorder as a consequence of hydrophobic interactions of the protein with the lipids. Since these effects are more important for EPA dispersions than for DMPA, MBP is believed to penetrate more into EPA bilayers. This could be due to the fact that the hydrogen bond network formed by the charged polar headgroups of EPA is weaker than that of DMPA. This is supported by the spectra of the phosphate region showing that the phosphate groups of EPA are less hydrogen bonded than DMPA. In the presence of MBP, the hydrogen bond network is replaced by electrostatic interactions of the protein with the polar headgroups of the lipid. Infrared spectra of the polar headgroup region also show evidence that MBP enhanced the second ionization state of the phosphate group at neutral pH, this effect being more important for EPA than for DMPA bilayers. Also, infrared spectra of the lipid carbonyl stretching region show evidence that MBP limits the accessibility of water molecules to the interfacial part of the lipid bilayer. Finally, ATR measurements on oriented films of lipid/protein complexes indicate that the penetration of the protein into the lipid bilayer is followed by a reorientation of the lipid acyl chains toward the normal to the bilayer in the case of EPA.(ABSTRACT TRUNCATED AT 250 WORDS)

The insertion mode of the long fatty acid chain of the asymmetric glycosphingolipid C26:0-cerebroside sulfate (C26-CBS) in symmetric matrices of phosphatidylcholines of different acyl chain length has been investigated by transmission and attenuated total reflectance (ATR) infrared spectroscopy. The concentration of C26-CBS in myelin is increased in the demyelinating disease adrenoleukodystrophy. The conformational order and the orientation of the chains of the asymmetric glycosphingolipid have been evaluated for C26-CBS incorporated at 8 mol % in perdeuterated dimyristoylphosphatidylcholine (DMPC-d54) and perdeuterated dipalmitoylphosphatidylcholine (DPPC-d62). The results, for the gel phase, are consistent with interdigitation of the C26-CBS long acyl chain across the bilayer center of an all-trans-DMPC bilayer in which DMPC is less tilted than in the absence of CBS. In contrast, in DPPC the results suggest that although the CBS long chain interdigitates across the center of the bilayer, it does not change the tilt angle of the DPPC molecules in the gel phase. Furthermore, in DPPC, C26-CBS is less well oriented than the host DPPC molecules and it increases the gauche content of the DPPC acyl chains. The observation of the amide spectral region indicates that exposure of the sphingosine amide moiety to buffer is greater in the longer chain length DPPC bilayer than in the shorter chain length DMPC bilayer. The thermotropic behavior of the lipid mixtures of C26-CBS at 8 mol % in DMPC or DPPC shows that the glycosphingolipid stabilizes the gel phase of the short chain length bilayer while it destabilizes the long chain length one. Our results further demonstrate that, at this concentration, C26-CBS is completely miscible in DMPC and DPPC in the gel and the liquid crystalline phases. The difference in behavior of C26-CBS in DMPC and DPPC is a consequence of the greater mismatch between the C26 chain length and the bilayer thickness of DPPC relative to DMPC. They may help to understand the deleterious effects of glycosphingolipids with very long chain fatty acids in adrenoleukodystrophy.