Site-Directed Spin Labeling of NADPH-Cytochrome P450 Reductase
Jimmy B. Feix, Professor of Biophysics
The overall goals of this project are to define the conformational changes involved in ligand binding and electron transfer by NADPH-cytochrome P450 reductase (CPR), and to define the structure of an N-terminal polypeptide domain that anchors CPR to the lipid bilayer. CPR is an integral membrane protein that is a key component of microsomal electron transport. It is one of only two mammalian proteins known to contain both FMN and FAD and a NADPH binding site, the other being nitric oxide synthase. The major function of CPR is to transfer electrons provided by NADPH to the heme iron of a cytochrome P450 acceptor. CPR is composed of at least two structural domains and is generally considered to contain four functional domains. A hydrophobic N-terminal domain anchors the protein in the endoplasmic reticulum and ensures proper spatial orientation for electron transfer. Removal of this 56-amino acid N-terminal domain by digestion with trypsin releases a 70-kDa soluble domain that retains the ability to transfer electrons to artificial acceptors but is incapable of reducing cytochrome P450. All previous structural studies have been on the truncated, water-soluble form of CPR, and we now turn our attention to the full-length, membrane-bound form.
To determine the structural organization of the 56-residue N-terminal domain, a series of single-cysteine-containing mutants were prepared in a background in which the seven native cysteine residues of CPR had been replaced by either serine or alanine. These mutants were purified and spin-labeled with a sulfhydryl-specific nitroxide spin label, providing a site-specific molecular probe. The rotational mobility and accessibility to paramagnetic relaxation agents was determined for the spin label at each of these mutational sites and indicated that this region of the protein forms a single, transmembrane á-helix. These results establish the manner in which CPR is oriented on the membrane surface and will serve as a basis set for further studies aimed at characterizing domain movements during substrate binding and electron transfer.
Formation of Cholesterol-Glycosphingolipid Raft-Domains in Unsaturated Phosphatidylcholine Membranes and Their Characterization
W. Karol Subczynski, Associate Professor of Biophysics
We tested the application of different spin labels for discrimination of liquid-ordered domains. The DOT (discrimination by oxygen transport) method can be applied only if spin labels are distributed between liquid-disordered (ld) and liquid-ordered (lo) domains. We showed that in the DMPC/Chol membranes, 5-PC, 16-PC, and 5-SASL are sensitive probes for distinguishing domains. The cholesterol analog ASL is also a sensitive molecular probe that in the presence of oxygen shows distinct double-exponential SR signals for conditions where two domains are expected in a DMPC/Chol mixture. However, another cholesterol analog, CSL, shows only single-exponential SR signals under conditions where other spin labels indicate double-exponential SR signals. These data for temperatures above and below the phase transition temperature of DMPC are presented in Figure 1 below. This figure is important to understand both the advantages and the limitations of the DOT method.
Below the phase transition temperature, both ASL and CSL indicate only one environment, while data obtained with 5-PC and 16- PC (data not shown) and from the literature (Almeida PFF et al.  Biochemistry 31:6739) show that at these conditions (DMPC containing 15 mol% cholesterol) two phases, solid-ordered (so) and lo, should be present. The so phase should consist of a pure DMPC bilayer, and the lo phase should be composed of a mixture of DMPC with 20 mol% cholesterol. Thus, cholesterol molecules and cholesterol analogs should be located in only one domain (lo domain), which is indicated in Figure 1 above. It is worth noting that the values of the oxygen transport parameter measured with ASL (with the nitroxide moiety close to the membrane center) lie between those measured with 5-PC and 16-PC for lo phase membranes in the same temperature region. The value of the oxygen transport parameter for CSL is extremely low, indicating extremely low oxygen transport within the polar head group region of the lo phase. Above the phase transition cholesterol molecules (and cholesterol analogs) should be distributed between ld and lo domains. ASL clearly shows the presence of these domains, while CSL shows only a single environment in terms of the collision rate with oxygen. This result clearly indicates that the oxygen collision rate with CSL cannot differentiate the ld and lo domains. This example shows that spin labels should be carefully selected for DOT experiments.
We tested discrimination of liquid-ordered domains with the use of CuKTSM2. Collisions with CuKTSM2 are sensitive to a much larger area around the nitroxide moiety than collisions with molecular oxygen because of the size of the copper complex. This is why CuKTSM2 transport can sometimes discriminate the existence of liquid-ordered domains for conditions at which oxygen transport cannot.
This case is illustrated in Figure 1, where collisions of CuKTSM2 with CSL indicate the existence of two domains in DMPC/Chol membranes, while collisions with oxygen do not (see left Figure 1). The two-domain pattern is observed only for temperatures above the main phase transition of DMPC. From this figure, it is also clear that the CuKTSM2 transport parameter is not a sensitive indicator of membrane domains below the main phase transition because the movement of the copper complex is too slow. Thus, the ability of pulse EPR to discriminate membrane domains is increased by the alternative use of these two probe molecules, with certain limitations.
Solution Structure of Visual Arrestin
Candice S. Klug, Associate Professor of Biophysics
The aim of this research is to characterize the solution structure of visual arrestin, as several aspects of its structure in solution remain unknown. Progress on the first aim, which was to determine the solution structure of the "plastic regions" of arrestin found in the crystal structure, has been accomplished through the nitroxide scanning of residues 158-163. The crystal structure indicates two differing conformations, one a helix and the other a loop, for this section of arrestin. It was determined by site-directed spin labeling (SDSL) to be mainly helical at room temperature, based on power saturation methods. With these results, the stage is set to now be able to move on to more advanced study of these residues by MQ and SR techniques being established at the Center. In order to distinguish between populations of conformations in equilibrium, it will be of interest to study this region by multiquantum and saturation recovery at both Q- and W-bands.
Progress on the second aim has also been achieved. In order to determine the location of the dimer interface in solution, spectra of combinations of spin-labeled F85C, F197C, and K267C with WT protein have been recorded and analyzed. It was found that an additional motional component arises in the spectra of 85 and 197 upon the formation of dimers, allowing for the analysis and quantitation of a dimerization constant for arrestin of approximately 25mM. Preliminary results indicate that the predominant dimer configuration is head to tail, with the N-terminal half of one monomer interacting with the C-terminal half of the opposite monomer. These results are the first to identify this interface in solution.
In addition, initial MQ at Q-band and SR at Q-band experiments were carried out on spin-labeled K267C at the Center. 3Q and 5Q spectra have been recorded on 30nL of a 300mL sample (~10pmol) at various powers in the presence of oxygen and NiEDDA. These spectra showed an enhanced sensitivity to both T1 and T2 (Fig. 1). Spectra have also been successfully recorded on <1pmol of protein. A 30nL sample size is enabled by the development of a three-loop-two-gap Q-band resonator. Also, T1 values were obtained using SR at both X- and Q-bands for this site. Single exponentials were recorded for each frequency, and the T1 values in the presence of nitrogen were found to increase from X-band (T1 = 2.57ms) to Q-band (T1 = 3.35ms), as anticipated. T1 values also increased when recorded in the presence of increasing amounts of oxygen. Thus, the foundation has been laid for the feasibility and success of this proposed study using the new technology developed.