Nuclear Magnetic Resonance Investigation of the Cytochrome Oxidase-Phospholipid Interaction: A New Model for Boundary Lipid

Deuterium nuclear magnetic resonance spectra of 1 -myristoyl-2- [formula ommited] myristoylphosphatidylcholine [formula ommited] dipalmitoylphosphatidylcholine (DPPC-d6), l-[16‘,16‘,16‘-2H3]palmitoyl-2-paImito-leylphosphatidylcholine (PPPC-d3), and 1-myristoyl-2-[6‘,-6‘-2H2]myristoylphosphatidylcholine [DMPC-2-(6‘, 6‘-d2)] have been obtained in the presence of cytochrome oxidase (ferrocytochrome c:O2 oxidoreductase, EC 1.9.3.1) as a function of temperature and composition and for the pure lipid as a function of temperature by using the quadrupole-echo Fourier transform method at 34.1 MHz. Above Tc, the temperature of the gel-to-liquid crystal phase transition of the pure phospholipid, all of the spectra have a single quadrupole splitting with no evidence of a second component. For DMPC-d3 the splitting at ~30 °C decreases monotonically with cytochrome oxidase concentration from ~3.7 kHz in the pure lipid to ~2.6 kHz at 90 wt % protein. Plots of quadrupole splitting vs. protein-lipid ratio are linear, indicating that a simple two-site exchange model may be sufficient to account for the results observed. This shows that the methyl groups are “less ordered” in the “boundary lipid” than in the free lipid bilayer. Moreover, the exchange between the two sites is fast enough (>103 s-1) to average out their 1-kHz difference in splitting. Similar, less extensive results were obtained for the other methyl-labeled lipids. However, DMPC-2-(6‘,6‘-d2) with 67 wt % cytochrome oxidase has a spectrum with somewhat broadened features but with about the same splitting (28 kHz) as the pure lipid. Thus, the 6‘-CH2 group not only is more ordered in the pure lipid than the terminal methyl group but also its order parameter is less affected by the protein. In lipid samples containing cytochrome oxidase, the exchange-averaging and increased disorder of the terminal methyl groups persist immediately below Tc. But as the temperature is lowered further, a broad gellike component with a splitting of ~ 15 kHz develops in addition to the narrow component (~3-kHz splitting) from the boundary lipid associated with the protein. So the exchange has become too slow to average out the 10-kHz difference in splitting. The greater disorder of the terminal methyl groups in the boundary lipid both above and below Tc is compared with spin-label data interpretations which have suggested that the boundary lipid is more ordered. It is suggested that the motions of the hydrocarbon chains in boundary lipid may be more random and of larger amplitude, though slower, than in pure lipid bilayers. Many of the spectra have features less sharp than those of the theoretical powder pattern. They were analyzed by using simulation techniques with several possible mechanisms for broadening the transitions. Two were found to be important under various circumstances―a Lorentzian broadening derived from the decay rate T2e of the quadrupole echo and a Gaussian distribution of order parameters.