A comprehensive description is given of instrumental and theoretical methods employed to make accurate measurements of rotational correlation times using passage saturation transfer electron paramagnetic resonance (ST–EPR). Saturation transfer methods extend by several orders of magnitude the sensitivity of EPR to very slow motion; for example, for nitroxide spin labels, correlation times as long as 10−3 sec become accessible to measurement. Two ST–EPR detection schemes are discussed in detail: dispersion, detected 90° out of phase with respect to the 100 kHz field modulation, and absorption, detected 90° out of phase with respect to the second harmonic of the 50 kHz field modulation. The sensitivities of these configurations are illustrated with experimental spectra obtained from a system obeying isotropic Brownian rotational diffusion; namely, maleimide spin labeled human oxyhemoglobin in aqueous glycerol solutions. Two theoretical approaches, one employing coupled Bloch equations and the other utilizing the stochastic Liouville equation for the density matrix with the orientation variables treated by transition rate matrix or orthogonal eigenfunction expansion methods, are in excellent agreement with each other and with model system spectra. Both experimental and theoretical spectra depend on a number of relaxation processes other than rotational diffusion; consequently, considerable care must be taken to ensure the accuracy of measured rotational correlation times. Although the absorption method is currently the more sensitive and convenient one to apply with most conventional (commercial) spectrometers, the dispersion ST–EPR method is potentially more powerful, providing strong motivation for future technological efforts to decrease noise levels in dispersion experiments.

We have used electron paramagnetic resonance (EPR) spectra to study spin labels selectively and rigidly attached to myosin heads in glycerinated rabbit psoas muscle fibers. Because the angle between the magnetic field and the principal axis of the probe determines the position of the EPR absorption line, spectra from labeled fibers oriented parallel to the magnetic field yielded directly the distribution of spin label orientations relative to the fiber axis. Two spin labels, having reactivities resembling iodoacetamide (IASL) and maleimide (MSL), were used. In rigor fibers with complete filament overlap, both labels displayed a narrow angular distribution, full width at half maximum approximately 15 degrees, centered at angles of 68 degrees (IASL) and 82 degrees (MSL). Myosin subfragments (heavy meromyosin and subfragment-1) were labeled and allowed to diffuse into fibers. The resulting spectra showed the same sharp angular distribution that was found for the labeled fibers. Thus is appears that virtually all myosin heads in a rigor fiber have the same orientation relative to the fiber axis, and this orientation is determined by the actomyosin bond. Experiments with stretched fibers indicated that the spin labels on the fraction of heads not interacting with actin filaments had a broad angular distribution. Addition of ATP to unstretched fibers under relaxing conditions produced orientational disorder, resulting in a spectrum almost indistinguishable from that of an isotropic distribution of probes. Addition of either an ATP analog (AMPPNP) or pyrophosphate produced partial disorder. That is a fraction of the probes remained sharply oriented as in rigor while a second fraction was in a disordered distribution similar to that of relaxed fibers.

Phospholamban is an integral membrane protein that regulates the contractility of cardiac muscle by maintaining cardiomyocyte calcium homeostasis. Abnormalities in association of protein kinase A with PLB have recently been linked to human heart failure, where a single mutation is responsible for dilated cardiomyopathy. To date, a high-resolution structure of phospholamban in a lipid environment has been elusive. Here, we describe the first structure of recombinant, monomeric, biologically active phospholamban in lipid-mimicking dodecylphosphocholine micelles as determined by multidimensional NMR experiments. The overall structure of phospholamban is "L-shaped" with the hydrophobic domain approximately perpendicular to the cytoplasmic portion. This is in agreement with our previously published solid-state NMR data. In addition, there are two striking discrepancies between our structure and those reported previously for synthetic phospholamban in organic solvents: a), in our structure, the orientation of the cytoplasmic helix is consistent with the amphipathic nature of these residues; and b), within the hydrophobic helix, residues are positioned on two discrete faces of the helix as consistent with their functional roles ascribed by mutagenesis. This topology renders the two phosphorylation sites, Ser-16 and Thr-17, more accessible to kinases.

Phospholamban (PLB), a 52-residue protein integral to the cardiac sarcoplasmic reticulum, is a key regulator of the Ca pump. PLB has been shown to form pentamers in the denaturing detergent sodium dodecyl sulfate (SDS), but its oligomeric state in the natural environment of the lipid membrane remains unknown. In order to address this issue, we performed electron paramagnetic resonance (EPR) experiments on two types of lipid-reconstituted, recombinant PLB: wild type (WT PLB) and a mutant substituted with alanine at leucine 37 (L37A PLB), whose propensity to oligomerize in SDS is greatly diminished. The lipid used in reconstitution was dioleoylphosphatidylcholine (DOPC) doped with a phospholipid spin-label that detects protein contact. EPR spectroscopy was used to determine the fraction of the total lipid molecules in contact with PLB. Our results show that, in phospholipid bilayers, WT PLB is oligomeric (effective oligomeric size of 3.52 +/- 0.71), while L37A PLB is monomeric (effective oligomeric size of 1.15 +/- 0.15). Thus, the oligomeric states of these proteins in the lipid membrane are remarkably similar to those in SDS solution. In particular, the point mutation in L37A PLB greatly destabilizes the PLB oligomer. Phosphorylation of PLB by protein kinase A, which has been shown to relieve inhibition of the cardiac Ca pump, changes the lipid-PLB interactions, decreasing the number of lipids restricted by contact with protein. The results are consistent with a phosphorylation-dependent increase of the effective oligomer size of WT PLB from 3.52 to 5.34 and of L37A PLB from 1.15 to 1.91. These phosphorylation effects were abolished in a medium with a high ionic strength. We conclude that the oligomeric states of PLB in lipid membranes are in a dynamic equilibrium that is perturbed by phosphorylation due to reduced electrostatic repulsion among PLB protomers.