Search

Selective methane oxidation is difficult chemistry. Here we describe a strategy for the electrocatalysis of selective methane oxidation by immobilizing tricopper catalysts on the cathodic surface. In the presence of dioxygen and methane, the activation of these catalysts above a threshold cathodic potential can initiate the dioxygen chemistry for O atom transfer to methane. The catalytic turnover is completed by facile electron injections into the tricopper catalysts from the electrode. This technology leads to dramatic enhancements in performance of the catalysts toward methane oxidation. Unprecedented turnover frequencies (>40 min -1 ) and high product throughputs (turnover numbers >30 000 in 12 h) are achieved for this challenging chemical transformation in water under ambient conditions. The technology is green and suitable for on-site direct conversion of methane into methanol.

The burgeoning field of twistronics, which concerns how changing the relative twist angles between two materials creates new optoelectronic properties, offers a novel platform for studying twist-angle dependent excitonic physics. Herein, by surveying a range of hexagonal phase transition metal dichalcogenides (TMD) twisted homobilayers, we find that 21.8 ± 1.0°-twisted ( 7 a × 7 a ) and 27.8 ± 1.0°-twisted ( 13 a × 13 a ) bilayers account for nearly 20% of the total population of twisted bilayers in solution-phase restacked bilayers and can be found also in chemical vapor deposition (CVD) samples. Examining the optical properties associated with these twisted angles, we found that 21.8 ± 1.0° twisted MoS 2 bilayers exhibit an intense moiré exciton peak in the photoluminescence (PL) spectra, originating from the refolded Brillouin zones. Our work suggests that commensurately twisted TMD homobilayers with short commensurate wavelengths can have interesting optoelectronic properties that are different from the small twist angle counterparts.

The particulate methane monooxygenase (pMMO) is the first enzyme in the C1 metabolic pathway in methanotrophic bacteria. As this enzyme converts methane into methanol efficiently near room temperature, it has become the paradigm for developing an understanding of this difficult C1 chemistry. pMMO is a membrane-bound protein with three subunits (PmoB, PmoA, and PmoC) and 12-14 coppers distributed among different sites. X-ray crystal structures that have revealed only three mononuclear coppers at three sites have neither disclosed the location of the active site nor the catalytic mechanism of the enzyme. Here we report a cyro-EM structure of holo -pMMO from Methylococcus capsulatus (Bath) at 2.5 Å, and develop quantitative electrostatic-potential profiling to scrutinize the nonprotein densities for signatures of the copper cofactors. Our results confirm a mononuclear Cu I at the A site, resolve two Cu I s at the B site, and uncover additional Cu I clusters at the PmoA/PmoC interface within the membrane ( D site) and in the water-exposed C -terminal subdomain of the PmoB ( E clusters). These findings complete the minimal set of copper factors required for catalytic turnover of pMMO, offering a glimpse of the catalytic machinery for methane oxidation according to the chemical principles underlying the mechanism proposed earlier.

Twisting the angle between van der Waals stacked 2D layers has recently sparked great interest as a new strategy to tune the physical properties of the materials. The twist angle and associated strain profiles govern the electrical and optical properties of the twisted 2D materials, but their detailed atomic structures remain elusive. Herein, using combined atomic-resolution electron microscopy and density functional theory (DFT) calculations, we identified five unique types of moiré features in commensurately twisted 7 a × 7 a transition metal dichalcogenide (TMD) bilayers. These stacking variants are distinguishable only when the moiré wavelength is short. Periodic lattice strain is observed in various commensurately twisted TMD bilayers. Assisted by Zernike polynomial as a hierarchical active-learning framework, a hexagon-shaped strain soliton network has been atomically unveiled in nearly commensurate twisted TMD bilayers. Unlike stacking-polytype-dependent properties in untwisted structures, the stacking variants have the same electronic structures that suggest twisted bilayer systems are invariant against interlayer gliding.

The active site of methanol dehydrogenase (MDH) contains a rare disulfide bridge between adjacent cysteine residues. As a vicinal disulfide, the structure is highly strained, suggesting it might work together with the pyrroloquinoline quinone (PQQ) prosthetic group and the Ca 2+ ion in the catalytic turnover during methanol (CH 3 OH) oxidation. We purify MDH from Methylococcus capsulatus (Bath) with the disulfide bridge broken into two thiols. Spectroscopic and high-resolution X-ray crystallographic studies of this form of MDH indicate that the disulfide bridge is redox active. We observe an internal redox process within the holo -MDH that produces a disulfide radical anion concomitant with a companion PQQ radical, as evidenced by an optical absorption at 408 nm and a magnetically dipolar-coupled biradical in the EPR spectrum. These observations are corroborated by electron-density changes between the two cysteine sulfurs of the disulfide bridge as well as between the bound Ca 2+ ion and the O5-C5 bond of the PQQ in the high-resolution X-ray structure. On the basis of these findings, we propose a mechanism for the controlled redistribution of the two electrons during hydride transfer from the CH 3 OH in the alcohol oxidation without formation of the reduced PQQ ethenediol, a biradical mechanism that allows for possible recovery of the hydride for transfer to an external NAD + oxidant in the regeneration of the PQQ cofactor for multiple catalytic turnovers. In support of this mechanism, a steady-state level of the disulfide radical anion is observed during turnover of the MDH in the presence of CH 3 OH and NAD + .

Methane monooxygenases (MMOs) mediate the facile conversion of methane into methanol in methanotrophic bacteria with high efficiency under ambient conditions. Because the selective oxidation of methane is extremely challenging, there is considerable interest in understanding how these enzymes carry out this difficult chemistry. The impetus of these efforts is to learn from the microbes to develop a biomimetic catalyst to accomplish the same chemical transformation. Here, we review the progress made over the past two to three decades toward delineating the structures and functions of the catalytic sites in two MMOs: soluble methane monooxygenase (sMMO) and particulate methane monooxygenase (pMMO). sMMO is a water-soluble three-component protein complex consisting of a hydroxylase with a nonheme diiron catalytic site; pMMO is a membrane-bound metalloenzyme with a unique tricopper cluster as the site of hydroxylation. The metal cluster in each of these MMOs harnesses O 2 to functionalize the C-H bond using different chemistry. We highlight some of the common basic principles that they share. Finally, the development of functional models of the catalytic sites of MMOs is described. These efforts have culminated in the first successful biomimetic catalyst capable of efficient methane oxidation without overoxidation at room temperature.

We have developed a novel streamlined sample preparation procedure for mass spectrometric (MS) analysis of membrane proteins using surface-oxidized nanodiamond particles. The platform consists of solid-phase extraction and elution of the membrane proteins on nanodiamonds, concentrating the membrane proteins on the nanodiamonds and separating out detergents, chaotropic agents, and salts, and other impurities that are often present at high concentrations in solubilized membrane preparations. In this manner, membrane-protein extracts are transformed into MS-ready samples in minutes. The protocol is not only fast, but also widely adaptable and highly effective for preparing generic membrane protein samples for both MALDI-MS studies of membrane-protein complexes and shotgun membrane proteomics studies. As proof of concept, we have demonstrated substantial improvements in the MALDI-MS analysis of the particulate methane monooxygenase (pMMO) complex, a three-subunit transmembrane protein solubilized in various detergent buffers. Enzymatic digestions of membrane proteins are also greatly facilitated since the proteins extracted on to the nanodiamonds are exposed on the surface of the nanoparticles rather than in SDS gels or in detergent solutions. We illustrate the effectiveness of nanodiamonds for SDS removal in the preparation of membrane proteins for MS analysis on the proteome level by examining the quality of the tryptic peptides prepared by on-surface nanodiamond digestion of an E. coli membrane fraction for shotgun proteomics.

By means of time-resolved photoluminescence and photothermal techniques, after-effects from excited-state dynamics, energy migration, and conformational rearrangement of poly(9,9-di-n-octyl-2,7-fluorene) (PFO) and its homologues has been examined and interpreted with rotational potential maps from quantum mechanical calculations. Steady-state photoluminescence spectral changes and time-resolved photoluminescence measurements of oligofluorenes and PFO diluted in toluene suggest excited state ring torsion occurring within 30 ps of photoexitation. With all effects from internal conversion/intersystem crossing processes properly accounted for, we show that the conformational changes associated with this twisting motion can be quantitatively probed by means of photothermal methods. Results suggest mean torsion between neighboring fluorene units by ca. 40 degrees upon excitation, in agreement with the shift of rotational potential minimum from +/-40 degrees (and +/-140 degrees) in the ground state to +/-20 degrees (and +/-160 degrees) in the first excited singlet state according to results of quantum mechanical calculations.

[Figure: see text]. The growing need for inexpensive methods to convert methane to methanol has sparked considerable interest in methods that catalyze this process. The integral membrane protein particulate methane monooxygenase (pMMO) mediates the facile conversion of methane to methanol in methanotrophic bacteria. Most evidence indicates that pMMO is a multicopper enzyme, and these copper ions support redox, dioxygen, and oxo-transfer chemistry. However, the exact identity of the copper species that mediates the oxo-transfer chemistry remains an area of intense debate. This highly complex enzyme is notoriously difficult to purify because of its instability outside the lipid bilayer and tendency to lose its essential metal cofactors. For this reason, pMMO has resisted both initial identification and subsequent isolation and purification for biochemical and biophysical characterization. In this Account, we describe evidence that pMMO is a multicopper protein. Its unique trinuclear copper cluster mediates dioxygen chemistry and O-atom transfer during alkane hydroxylation. Although a recent crystal structure did not show this tricopper cluster, we provide compelling evidence for such a cluster through redox potentiometry and EPR experiments on the "holo" enzyme in pMMO-enriched membranes. We also identify a site in the structure of pMMO that could accommodate this cluster. A hydrophobic pocket capable of harboring pentane, the enzyme's largest known substrate, lies adjacent to this site. In addition, we have designed and synthesized model tricopper clusters to provide further chemical evidence that a tricopper cluster mediates the enzyme's oxo-transfer chemistry. These biomimetic models exhibit similar spectroscopic properties and chemical reactivity to the putative tricopper cluster in pMMO. Based on computational analysis using density functional theory (DFT), triangular tricopper clusters are capable of harnessing a "singlet oxene" upon activation by dioxygen. An oxygen atom is then inserted via a concerted process into the C-H bond of an alkane in the transition state during hydroxylation. The turnover frequency and kinetic isotope effect predicted by DFT show excellent agreement with experimental data.

The crystal structure of the particulate methane monooxygenase (pMMO) from Methylococcus capsulatus (Bath) has been reported recently [Lieberman, R. L., and Rosenzweig, A. C. (2005) Crystal structure of a membrane-bound metalloenzyme that catalyses the biological oxidation of methane, Nature 434, 177-182]. Subsequent work has shown that the preparation on which the X-ray analysis is based might be missing many of the important metal cofactors, including the putative trinuclear copper cluster at the active site as well as ca. 10 copper ions (E-clusters) that have been proposed to serve as a buffer of reducing equivalents to re-reduce the copper atoms at the active site following the catalytic chemistry [Chan, S. I., Wang, V. C.-C., Lai, J. C.-H., Yu, S. S.-F., Chen, P. P.-Y., Chen, K. H.-C., Chen, C.-L., and Chan, M. K. (2007) Redox potentiometry studies of particulate methane monooxygenase: Support for a trinuclear copper cluster active site, Angew. Chem., Int. Ed. 46, 1992-1994]. Since the aqueous-exposed domains of the 45 kDa subunit (PmoB) have been suggested to be the putative binding domains for the E-cluster copper ions, we have cloned and overexpressed in Escherichia coli the two aqueous-exposed subdomains toward the N- and C-termini of the subunit: the N-terminal subdomain (residues 54-178) and the C-terminal subdomain (residues 257-394 and 282-414). The recombinant C-terminal water-exposed subdomain is shown to behave like a Cu(I) sponge, taking up to ca. 10 Cu(I) ions cooperatively when cupric ions are added to the protein fragment in the presence of dithiothreitol or ascorbate. In addition, circular dichroism measurements reveal that the C-terminal subdomain folds into a beta-sheet structure in the presence of Cu(I). The propensity for the C-terminal subdomain to bind Cu(I) is consistent with the high redox potential(s) determined for the E-cluster copper ions in the pMMO. These properties of the E-clusters are in accordance with the function proposed for these copper ions in the turnover cycle of the enzyme.

As part of our continuing study of the effects of the turn sequence on the conformational stability as well as the mechanism of folding of a beta-sheet structure, we have undertaken a parallel investigation of the solution structure, conformational stability, and kinetics of refolding of the beta-sheet VFIVDGOTYTEV(D)PGOKILQ. The latter peptide is an analogue of the original Gellman beta-sheet VFITS(D)PGKTYTEV(D)PGOKILQ, wherein the TS(D)PGK turn sequence in the first hairpin has been replaced by VDGO. Thermodynamics studies revealed comparable conformational stability of the two peptides. However, unlike the Gellman peptide, which showed extremely rapid refolding of the first hairpin, early kinetic events associated with the refolding of the corresponding hairpin in the VDGO mutant were found to be significantly slower. A detailed study of the conformation of the modified peptide suggested that hydrophobic interactions might be contributing to its stability. Accordingly, we surmise that the early kinetic events are sensitive to whether the formation of the hairpin is nucleated at the turn or by sequestering of the hydrophobic residues across the strand, before structural rearrangements to produce the nativelike topology. Nucleation of the hairpin at the turn is expected to be intrinsically rapid for a strong turn. However, if the process must involve collapse of hydrophobic side chains, the nucleation should be slower as solvent molecules must be displaced to sequester the hydrophobic residues. These findings reflect the contribution of different forces toward nucleation of hairpins in the mechanism of folding of beta-sheets.

The particulate methane monooxygenase (pMMO) of Methylococcus capsulatus (Bath) is an integral membrane protein that catalyzes the conversion of methane to methanol. To gain some insight into the structure-reactivity pattern of this protein, we have applied attenuated total reflection Fourier transform infrared (ATR-FTIR) spectroscopy to investigate the secondary structure of the pMMO. The results showed that ca. 60% of the amino acid residues were structured as alpha-helices. About 80% of the peptide residues were estimated to be protected from the amide (1)H/(2)H exchange during a 21 h exposure to (2)H(2)O. In addition, a significant portion of the protein was shown to be sequestered within the bilayer membrane, protected from trypsin proteolysis. The ATR-FTIR difference spectrum between the intact and the proteolyzed pMMO-enriched membranes revealed absorption peaks only in the spectral regions characteristic for unordered and beta-structures. These observations were corroborated by amino acid sequence analysis of the pMMO subunits using the program TransMembrane topology with a Hidden Markov Model: 15 putative transmembrane alpha-helices were predicted. Finally, an attempt was also made to model the three-dimensional folding of the protein subunits from the sequence using the Protein Fold Recognition Server based on the 3D Position Specific Scoring Matrix Method. The C-terminal solvent-exposed sequence (N255-M414) of the pMMO 45 kDa subunit was shown to match the beta-sheet structure of the multidomain cupredoxins. We conclude on the basis of this ATR-FTIR study that pMMO is an alpha-helical bundle with ca. 15 transmembrane alpha-helices embedded in the bilayer membrane, together with a water-exposed domain comprised mostly of beta-sheet structures similar to the cupredoxins.

The particulate methane monooxygenase (pMMO) is a complex membrane protein complex that has been difficult to isolate and purify for biochemical and biophysical characterization because of its instability in detergents used to solubilize the enzyme. In this perspective, we summarize the progress recently made toward obtaining a purified pMMO-detergent complex and characterizing the enzyme in pMMO-enriched membranes. The purified pMMO is a multi-copper protein, with ca. 15 copper ions sequestered into five trinuclear copper clusters: two for dioxygen chemistry and alkane hydroxylation (catalytic or C-clusters) and three to provide a buffer of reducing equivalents to re-reduce the C-clusters following turnover (electron transfer or E-clusters). The enzyme is functional when all the copper ions are reduced. When the protein is purified under ambient aerobic conditions in the absence of a hydrocarbon substrate, only the C-clusters are oxidized; there is an apparent kinetic barrier for electron transfer from the E-cluster copper ions to the C-clusters under these conditions. Evidence is provided in support of both C-clusters participating in the dioxygen chemistry, but only one C-cluster supporting alkane hydroxylation. Acetylene modification of the latter C-cluster in the hydrophobic pocket of the active site lowers or removes the kinetic barrier for electron transfer from the E-clusters to the C-clusters so that all the copper ions could be fully oxidized by dioxygen. A model for the hydroxylation chemistry when a hydrocarbon substrate is bound to the active site of the hydroxylation C-cluster is presented. Unlike soluble methane monooxygenase (sMMO), pMMO exhibits limited substrate specificity, but the hydroxylation chemistry is highly regioselective and stereoselective. In addition, the hydroxylation occurs with total retention of configuration of the carbon center that is oxidized. These results are consistent with a concerted mechanism involving direct side-on insertion of an active singlet "oxene" from the activated copper cluster across the "C-H" bond in the active site. Finally, in our hands, both the purified pMMO-detergent complex and pMMO-enriched membranes exhibit high NADH-sensitive as well as duroquinol-sensitive specific activity. A possible role for the two reductants in the turnover of the enzyme is proposed.

The reaction of the quinol oxidase cytochrome bo3 from Escherichia coli with ubiquinol-2 (UQ2H2) was carried out using substoichiometric (0.5 equiv) amounts of substrate. Reactions were monitored through the use of freeze-quench EPR spectroscopy. Under 1 atm of argon, semiquinone was formed at the QB site of the enzyme with a formation rate constant of 140 s-1; the QB semiquinone EPR signal decayed with a rate constant of about 5 s-1. Heme b and CuB were reduced within the 10-ms dead time of the freeze-quench experiment and remained at a constant level of reduction over the 1-s time course of the experiment. Quantitation of the reduction levels of QB and heme b during this reaction yielded a reduction potential of 30-60 mV for heme b. Under a dioxygen atmosphere, the rates of semiquinone formation and its subsequent decay were not altered significantly. However, accurate quantitation of the EPR signals for heme b and heme o3 could not be made, due to interference from dioxygen. In the reaction between the QB-depleted enzyme and UQ2H2 under substoichiometric conditions, there was no observable change in the EPR spectra of the enzyme over the time course of the reaction, suggesting an electron transfer from heme b to the binuclear site in the absence of QB which occurs within the dead time of the freeze-quench apparatus. Analysis of the thermodynamics and kinetics of electron transfers in this enzyme suggests that a Q-cycle mechanism for proton translocation is more likely than a cytochrome c oxidase-type ion-pump mechanism.

As part of our studies on the structural and dynamic properties of hyperthermostable proteins, we have investigated the unfolding pathways of the small iron-sulfur protein rubredoxin from Pyrococcus furiosus (RdPf) at pH 2. Unfolding has been initiated by temperature jump, triggered by manual mixing of a concentrated protein solution into a thermally preequilibrated buffer. The process has been followed in real time by absorption, tryptophan fluorescence emission, and far-UV circular dichroism. Unlike the case of the mesophilic rubredoxin from Clostridium pasteurianum (RdCp), RdPf displays a complex unfolding kinetics, pointing to the formation of at least three intermediates. All of the steps, including the one involving metal ion release, are extremely slow. However, hydrophobic core relaxation--not Fe3+ loss--is rate-determining for RdPf unfolding. This clearly rules out the fact that Fe3+ is solely responsible for the kinetic stability of RdPf. Results have been discussed in terms of sequential vs parallel pathways, and the possible role of irreversible phenomena has been taken into consideration. Aggregation does not appear to play a significant role in the observed kinetic complexities. According to a proposed sequential mechanism, partial release of secondary structure elements precedes iron loss, which is then followed by further loss of beta-sheet content and, finally, by hydrophobic relaxation. Although the main features of the RdPf unfolding mechanism remain substantially unchanged over the experimentally accessible temperature range, final hydrophobic relaxation gets faster, relative to the other events, as the temperature is decreased. A qualitative assessment of the unfolding activation parameters suggests that this arises from the very low activation energies (Ea) that characterize this step.

The temperature dependence of the unfolding kinetics of rubredoxins from the hyperthermophile Pyrococcus furiosus (RdPf) and the mesophile Clostridium pasteurianum (RdCp) has been studied. Results show that RdPf unfolds much more slowly, under all experimentally accessible temperature regimes, than RdCp and other typical mesophilic proteins. Rates of RdCp and RdPf unfolding decrease upon increasing the pH above 2 and diverge dramatically at pH 7. As shown by detailed electrostatic energy calculations, this is the result of a differential degree of protonation of the negatively charged amino acids, which causes distinct electrostatic configurations as a function of pH. We propose that ion pairs, particularly those that are placed in key surface positions, may play a kinetic role by mildly clamping the protein and thereby influencing the nature and the number of the vibrational normal modes that are thermally accessible upon unfolding. More generally, these modes are also likely to be affected by the favorable electrostatic configurations, which we have shown to be directly linked to the extremely slow unfolding rates of RdPf at neutral pH. Even at pH 2, in the absence of any salt bridges, the unfolding rates of RdPf are much smaller than those of RdCp. This is ascribed to presently unidentified structural elements of nonelectrostatic nature. Since electrostatic effects influence the unfolding kinetics of both mesophilic and thermophilic rubredoxins, these findings may be of general significance for proteins.

The single-turnover kinetics of the oxidation of cytochrome aa3 by O2 have been studied using a new approach. Up to 1000 whole spectra covering both the Soret and alpha regions were sequentially collected at room temperature from single samples with a time resolution of 10 microns. All of the spectral and time information were used in analyses based on singular value decomposition. Four spectral transitions (i.e., intermediates) were distinguished with time constants near 0.01, 0.1, 1.1, and 30 ms. Two different kinds of sequential models were evaluated, one linear and the other branched. Although past kinetic analyses have emphasized the linear sequential model, the complexity of the intramolecular electron transfer in this enzyme suggests that a branched model be considered. This is especially true in a single-turnover experiment where earlier optical and EPR studies have pointed unequivocally to a branched model [Clore et al. (1980) Biochem. J. 185, 139-154; Blair et al. (1985) J. Am. Chem. Soc. 107, 7389-7399]. In the present study, analysis of spectral data in terms of the linear model did not reveal the formation and decay of the expected oxyferryl intermediate, whereas analysis of the branched model did. The results obtained using the branched model are consistent with all of the available evidence from a broad range of physical techniques that have been applied to examine the single-turnover kinetics of the oxidation of reduced cytochrome aa3 by O2.

The cytochrome bo3 ubiquinol oxidase complex from Escherichia coli contains two binding sites for ubiquinone(ol) (UQ(H2)). One of these binding sites, the ubiquinol oxidation site, is clearly in dynamic equilibrium with the UQ(H2) pool in the membrane. The second site has a high affinity for ubiquinone (UQ), stabilizes a semiquinone species, and is located physically close to the low-spin heme b component of the enzyme. The UQ molecule in this site has been proposed to remain strongly bound to the enzyme during enzyme turnover and to act as a cofactor facilitating the transfer of electrons from the substrate ubiquinol to heme b [Sato-Watanabe et al. (1994) J. Biol. Chem. 269, 28908-28912]. In this paper, the steady-state turnover of the enzyme is examined in the presence and absence of inhibitors (UHDBT and NQNO) that appear to be recognized as ubisemiquinone analogs. It is found that the kinetics are accounted for best by a noncompetitive inhibitor binding model. Furthermore, at high concentrations, the substrates ubiquinol-1 and ubiquinol-2 inhibit turnover in an uncompetitive fashion. Together, these observations strongly suggest that there must be at least two UQ(H2) binding sites that are in rapid equilibrium with the UQ(H2) pool under turnover conditions. Although these data do not rule out the possibility that a strongly bound UQ molecule functions to facilitate electron transfer to heme b, they are more consistent with the behavior expected if the two UQ(H2) binding sites were to function in a Q(H2)-loop mechanism (similar to that of the cytochrome bc1 complex) as originally proposed by Musser and co-workers [(1993) FEBS Lett. 327, 131-136]. In this model, ubiquinol is oxidized at one site and ubiquinone is reduced at the second site. While the structural similarities of the heme-copper ubiquinol and cytochrome c oxidase complexes suggest the possibility that these two families of enzymes translocate protons by similar mechanisms, the current observations indicate that the Q(H2)-loop proton translocation mechanism for the heme-copper ubiquinol oxidase complexes should be further investigated and experimentally tested.

The bases of the hyperthermostability of rubredoxin from Pyrococcus furiosus (RdPf) have been probed by structural perturbations induced by solution pH and ionic strength changes. Comparison of the solution behavior at pH 7 and pH 2, as probed by far- and near-UV circular dichroism, Trp fluorescence emission, 1-anilinonaphthalene-8-sulfonate (ANS) binding, and NMR spectroscopy, reveals the presence of only minimal structural variations at room temperature. At pH 2, the protein displays a surprising nearly native-like behavior at high ionic strength while, at low ionic strength, it is capable of strongly binding the hydrophobic probe ANS. All the secondary and tertiary structural features, including the environment of the hydrophobic core, appear to be intact regardless of pH and ionic strength. The apparent "melting" or denaturation temperature at pH 2, however, is 42 degrees C lower than at pH 7. This is attributed to the perturbation of many electrostatic interactions, including the disruption of all the ion pairs, which is complete at pH 2, as indicated by electrometric pH titrations. The implications of these findings for the origins of the hyperthermostability of rubredoxin are discussed.

Bovine heart cytochrome c oxidase (CcO) has been modified by 8-azido-adenosine 5'-triphosphate (8-azido-ATP), and the electron-transfer activity from ferrocytochrome c to the modified CcO under physiological ionic strengths has been studied by the laser flash photolysis technique with 5-deazariboflavin and EDTA as the electron donor. The kinetics of intermolecular electron transfer between the redox protein partners was shown to be reduced significantly. In addition, there is significant decrease in the binding affinity of the cytochrome c to the oxidase upon 8-azido-ATP modification. The 8-azido-ATP-modified CcO exhibited 50% of the intracomplex electron-transfer rate (ket) and 56% of the association constant (Ka) normally observed between cytochrome c and native CcO under otherwise identical conditions. Since the effective electron transfer rate constant is the product of ket and Ka under nonsaturation conditions, the overall electron-transfer rate has been curtailed by over a factor of 2. Similar observations have been noted with the native CcO in the presence of 3 mM ATP. In contrast, the redox potential of neither CuA nor cytochrome a was altered upon 8-azido-ATP modification or in the presence of 3 mM ATP. Also, no gross structural changes at either the CuA or the cytochrome a site were noted, as evidenced by a lack of any spectral perturbations in the EPR signals from both of these centers. We conclude that ATP modulates the electron transfer from cytochrome c to CcO by interacting with the CcO and altering allosterically the docking. In this manner, ATP can affect the branching of the electron input from ferrocytochrome c to cytochrome a and CuA.

Horse heart cytochrome c has been modified by 8-azido-ATP and the electron-transfer activity of the modified cytochrome c's to bovine heart cytochrome c oxidase (CcO) under physiological ionic strengths has been studied by the laser flash photolysis technique with 5-deazariboflavin and EDTA as the electron donor. The intermolecular electron transfer between the redox protein partners was shown to be extremely slow. The 8-azido-ATP-modified system exhibited less than 5% of the intracomplex electron-transfer rate observed between native cytochrome c and CcO under otherwise identical conditions. The binding affinity of the modified cytochrome c was greatly reduced (3 orders of magnitude) at low ionic strengths; however, it was only slightly reduced (by a factor of 2) relative to the native protein at physiological ionic strengths. Thus, the binding affinity of the ATP-cytochrome c adducts is relatively insensitive to the ionic strength compared to the native enzyme, suggesting that a different docking conformation is assumed by the ATP-cytochrome c adducts in their interaction with the oxidase. Since the redox potential of the modified cytochrome c is close to the value of its native form, we conclude that there has been a change in the docking of the cytochrome c to CcO and the electronic coupling between heme c and CuA upon 8-azido-ATP modification.

The cytochrome o complex is a bo-type ubiquinol oxidase in the aerobic respiratory chain of Escherichia coli. This complex has a close structural and functional relationship with the eukaryotic and prokaryotic aa3-type cytochrome c oxidases. The specific activity, subunit composition, and metal content of the purified cytochrome o complex are not consistent for different preparative protocols reported in the literature. This paper presents a relatively simple preparation of the enzyme starting with a strain of Escherichia coli which overproduces the oxidase. The pure enzyme contains four subunits by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). Partial amino acid sequence data confirm the identities of subunit I, II, and III from the SDS-PAGE analysis as the cyoB, cyoA, and cyoC gene products, respectively. A slight modification of the purification protocol yields an oxidase preparation that contains a possible fifth subunit which may be the cyoE gene product. The pure four-subunit enzyme contains 2 equivs of iron but only 1 equiv of copper. There is no electron paramagnetic resonance detectable copper in the purified enzyme. Hence, the equivalent of CuA of the aa3-type cytochrome c oxidases is absent in this quinol oxidase. There is also no zinc in the purified quinol oxidase. Finally, monoclonal antibodies are reported that interact with subunit II. One of these monoclonals inhibits the quinol oxidase activity of the detergent-solubilized, purified oxidase. Hence, although subunit II does not contain CuA and does not interact with cytochrome c, it still must have an important function in the bo-type ubiquinol oxidase.

Resonance Raman spectroscopy was employed to investigate the heme structures of catalytic intermediates of cytochrome c oxidase at room temperature. The high-frequency resonance Raman spectra were obtained for compound C (the two-electron-reduced dioxygen intermediate), ferryl (the three-electron-reduced dioxygen intermediate), and the fully oxidized enzyme. Compound C was formed by photolyzing CO mixed-valence enzyme in the presence of O2. The ferryl intermediate was formed by reoxidation of the fully reduced enzyme by an excess of H2O2. Two forms of the oxidized enzyme were prepared by reoxidizing the fully reduced enzyme with O2. Our data indicate that, in compound C, cyt a3 is either intermediate or low spin and is nonphotolabile and its oxidation state marker band, v4, appears a higher frequency than that of the resting form of the enzyme. The ferryl intermediate also displays a low-spin cyt a3, which is nonphotolabile, and an even higher frequency for the oxidation state marker band, v4. The reoxidized form of cytochrome c oxidase with a Soret absorption maximum at 420 nm has an oxidation state marker band (v4) in a position similar to that of the resting form, while the spin-state region resembles that of compound C. This species subsequently decays to a second oxidized from of the enzyme, which displays a high-frequency resonance Raman spectrum identical with that of the original resting enzyme.

Gramicidin A' (GA') has been added to three lipid systems of varying hydrophobic thicknesses: dimyristoyllecithin (DML), dipalmitoyllecithin (DPL), and distearoyllecithin (DSL). The similarity in length between the hydrophobic portion of GA' and the hydrocarbon chains of the lipid bilayers has been studied by using 31P and 2H NMR. Hydrophobic mismatch has been found to be most severe in the DML bilayer system and minimal in the case of DSL. In addition, the effects of hydrophobic mismatch on the cooperative properties of the bilayer have been obtained from 2H NMR relaxation measurements. The results indicate that incorporation of the peptide into the bilayer disrupts the cooperative director fluctuations characteristic of pure multilamellar lipid dispersions. Finally, the GA'/lecithin ratio at which the well-known transformation from bilayer to reverse hexagonal (HII) phase occurs (Van Echteld et al., 1982; Chupin et al., 1987) is shown to depend on the acyl chain length of the phospholipid. A rationale is proposed for this chain length dependence.

The reduction potential of the cytochrome a site in the carbon monoxide derivative of beef heart cytochrome c oxidase has been studied under a variety of conditions by thin-layer spectroelectrochemistry. The reduction potential exhibits no ionic strength dependence and only a 9 mV/pH unit dependence between pH 6.5 and 8.5. The weak pH dependence indicates that protonation of the protein is not stoichiometrically linked to oxidoreduction over the pH range examined. The temperature dependence of the reduction potential implies a relatively large standard entropy of reduction of cytochrome a. The measured thermodynamic parameters for reduction of cyctochrome a are (all relative to the normal hydrogen electrode) delta Go'(25 degrees C) = -6.37 kcal mol-1, delta Ho' = -21.5 kcal mol-1, and delta So' = -50.8 eu. When cytochrome c is bound to the oxidase, the reduction potential of cytochrome a and its temperature dependence are not measurably affected. Under all conditions studied, the cytochrome a site did not exhibit simple Nernstian n = 1 behavior. The titration behavior of the site is consistent with a moderately strong anticooperative interaction between cytochrome a and CuA [Wang, H., Blair, D. F., Ellis, W. R., Jr., Gray, H. B., & Chan, S. I. (1985) Biochemistry (following paper in this issue)].

The reactions of nitric oxide (NO) with the oxidized and reduced forms of fungal and tree laccase, as well as with tree laccase depleted in type 2 copper, are reported. The products of the reactions were determined by NMR and mass spectroscopy, whereas the oxidation states of the enzymes were monitored by EPR and optical spectroscopy. All three copper sites in fungal laccase are reduced by NO. In addition, NO forms a specific complex with the reduced type 2 copper. NO similarly reduces all of the copper sites in tree laccase, but it also oxidizes the reduced sites produced by ascorbate or NO reduction. A catalytic cycle is set up in which N2O, NO2-, and various forms of the enzyme are produced. On freezing of fully reduced tree laccase in the presence of NO, the type 1 copper becomes reoxidized. This reaction does not occur with the enzyme depleted in type 2 copper, suggesting that it involves intramolecular electron transfer from the type 1 copper to NO bound to the type 2 copper. When the half-oxidized tree laccase is formed in the presence of NO, a population of molecules exists which exhibits a type 3 EPR signal. A triplet EPR signal is also seen in the same preparation and is attributed to a population of the enzyme molecules in which NO is bound to the reduced copper of a half-oxidized type 3 copper site.

Band 3 is proposed to contain substrate channels that lead from the aqueous medium to a transport site buried within the membrane, and which can be blocked by inhibitors. The inhibitors 1,2-cyclohexanedione (CHD) and dipyridamole (DP) each inhibit the transport site 35Cl NMR line broadening, but neither competes with Cl- for binding. Thus these inhibitors do not occupy the transport site; instead they slow the migration of Cl- between the transport site and the medium. The simplest explanation for this behavior is that CHD and DP block one or more substrate channels. CHD is an arginine-specific covalent modification reagent, and its effectiveness as a channel blocker indicates that the channel contains arginine positive charges to facilitate the migration of anions through the channel. DP is a noncovalent channel blocker that binds with a stoichiometry of 1 molecule per band 3 dimer. DP binding is unaffected by CHD but is prevented by phenylglyoxal (PG), 4,4'-dinitrostilbene-2,2'-disulfonate (DNDS), or niflumic acid. Thus the DP and CHD binding sites are distinct, with DP binding sufficiently close to the transport site to interact with PG and DNDS. It is proposed that substrate channels may be a general feature of transport proteins.

Modification of the CuA site in mammalian cytochrome c oxidase has been used to elucidate the functional role of this center in the catalytic cycle of the enzyme. Both heat treatment in detergents and chemical modification by p-(hydroxymercuri)benzoate (pHMB) convert CuA to a lower potential type II center and effectively remove the site from the electron-transfer pathway during turnover. In this study, resonance Raman spectroscopy has been employed to investigate the effects of these CuA modifications on the heme active sites. The Raman data indicate some environmental perturbation of the heme a3 chromophore in the modified derivatives. Only pHMB modification and SB-12 heat treatment produced significant effects in the Raman spectra of the fully reduced enzyme. These perturbations are much less evident in the spectra obtained within 10 ns of CO photolysis from the fully reduced species of the modified enzymes. Transient Raman studies further indicate that the half-time for CO religation in the modified enzymes is quite similar to that of the native protein.

The temperature dependence of the reduction potential of the CuA site in carbon monoxide inhibited cytochrome c oxidase has been measured with a spectroelectrochemical method adapted to the relatively weak near-infrared absorption of this copper ion. These measurements, together with parallel measurements on the 604-nm absorption due to Fea, indicate that an interaction between CuA and Fea causes the reduction potential for one of these sites to be decreased by approximately 40 mV upon reduction of the other. The temperature dependence of the CuA reduction potential indicates a relatively large and negative standard entropy of reduction of CuA (delta So' = -48.7 +/- 2.3 eu). Possible implications of the intersite redox interaction and the large standard entropy of reduction of the CuA site are discussed.

A model for the molecular interaction between cholesterol and phospholipid in bilayer membranes is presented. We propose that cholesterol forms associations with phospholipids with stoichiometries of both 1:1 and 1:2. A hydrogen bond between the beta-OH of cholesterol and the glycerol ester oxygen of a phospholipid is suggested as a likely mechanism for tight binding in a 1:1 complex. A second phospholipid molecule is loosely associated with the complex to form domains of 1:2 stoichiometry, which may coexist with pure phospholipid domains. Interfacial boundary phospholipid separates these two domains. Under conditions in which interfacial phospholipid is maximal, the perturbed phospholipid assumes a composition of 20 mol % cholesterol. To account for the phase behavior and surface properties of cholesterol-lipid membranes, we propose a molecular packing model for linear arrays within the cholesterol-rich domains. In this arrangement, two rows of 1:1 complex run antiparallel with loosely associated phospholipid intercalated between them. The loosely associated phospholipid can pack in the nearly hexagonal manner in which pure crystalline phospholipid is known to pack. The model provides maximal van der Waals contact in the hydrocarbon region of the bilayer and can maintain phospholipids as cholesterol's nearest neighbors at all concentrations up to 50 mol % cholesterol. The model is compatible with the diverse experimental observations compiled by many investigators over the past decade.

During the translocation of the band 3 transport site between the inward- and outward-facing orientations, the Cl- transport site complex passes through a transition state lying on the reaction pathway between the two extreme orientations. Niflumic acid, 2-[(7-nitrobenzofurazan-4-yl)amino]ethanesulfonate, and 2,4,6-trichlorobenzenesulfonate each are translocation blockers that can bind to both the inward- and outward-facing conformations of band 3. The principal mechanism of these inhibitors is a reduction in the translocation rate, since they have essentially no effect on the apparent KD for Cl- binding to the transport site and the migration of Cl- between the transport site and solution. Instead, these inhibitors raise the free energy of formation of the transition state during translocation and thereby can lock the transport site into either the inward- or outward-facing orientation. In contrast, 2,4-dinitrofluorobenzene (DNFB) appears to restrict the accessibility of the transport site to solution Cl-; also, the DNFB reaction rate is increased by Cl-, suggesting that DNFB modification may occur during translocation. Thus DNFB is proposed to trap the Cl--transport site complex site during translocation to yield a conformation intermediate to the inward- and outward-facing orientations. A model is presented for the molecular mechanism of transport across biological membranes. The transport machinery is proposed to contain greater than or equal to 6 transmembrane helices that surround a central channel containing a sliding hydrophobic barrier. The transport site lies between two of the channel-forming helices and remains stationary while the hydrophobic barrier slides from one end of the channel to the other, thereby exposing the transport site to the opposite solution compartment.

When cytochrome c oxidase is incubated at 43 degrees C for approximately 75 min in a solution containing the zwitterionic detergent sulfobetaine 12, the CuA site is converted into a type II copper as judged by changes in the 830-nm absorption band and the EPR spectrum of the enzyme. SDS-PAGE and sucrose gradient ultracentrifugation indicate concomitant loss of subunit III and monomerization of the enzyme during the heat treatment. Comparison of the optical and resonance Raman spectra of the heat-treated and native protein shows that the heme chromophores are not significantly perturbed; the resonance Raman data indicate that the small heme perturbations observed are limited to the cytochrome a3 site. Proton pumping measurements, conducted on the modified enzyme reconstituted into phospholipid vesicles, indicate that these vesicles are unusually permeable toward protons during turnover, as previously reported for the p-(hydroxymercuri)benzoate-modified oxidase and the modified enzyme obtained by heat treatment in lauryl maltoside. The sulfobetaine 12 modified enzyme is no longer capable of undergoing the recently reported conformational transition in which the tryptophan fluorescence changes upon reduction of the low-potential metal centers. Control studies on the monomeric and subunit III dissociated enzymes suggest that the disruption of this conformational change in the heat-treated oxidase is most likely associated with perturbation of the CuA site. These results lend support to the suggestion that the fluorescence-monitored conformational change of the native enzyme is initiated by reduction of the CuA site [Copeland et al. (1987) Biochemistry 26, 7311].

The resonance Raman (RR) spectra of oxidized, reduced, and oxidized cyanide-bound cytochrome c oxidase with excitation at several wavelengths in the 600-nm region are presented. No evidence is found for laser-induced photoreduction of the oxidized protein with irradiation at lambda approximately 600 nm at 195 K, in contrast to the predominance of this process upon irradiation in the Soret region at this temperature. The Raman spectra of all three protein species are very similar, and there are no Raman bands which are readily assignable to either cytochrome a or cytochrome a3 exclusively. The Raman spectra of the three protein species do, however, exhibit a number of bands not observed in the RR spectra of other hemoproteins upon exicitation in their visible absorption bands. In particular, strong Raman bands are observed in the low-frequency region of the RR spectra (less than 500 cm-1). The frequencies of these bands are similar to those of the copper-ligand vibrations observed in the RR spectra of type 1 copper proteins upon excitation in the 600-nm absorption band characteristic of these proteins. In cytochrome c oxidase, these bands do not disappear upon reduction of the protein and, therefore, cannot be attributed to copper-ligand vibrations. Thus, all the observed RR bands are associated with the two heme A moieties in the enzyme.

The interaction of alamethicin with lecithin bilayer vesicles in the presence of a transmembrane potential difference has been studied by proton magnetic resonance spectroscopy. Asymmetrical vesicles with alamethicin trapped within the intravesicular compartment were prepared and the formation of ion channels, were then triggered by imposing a suitable potassium ion gradient across the bilayer membrane in the presence of valinomycin. These alamethicin channels were found to permit the outward passage of europium ions into the extravesicular medium when the transmembrane potential is more positive for the inner compartment.

In order to elucidate the role of lysolecithin in membranes, we have examined the effect of lysolecithin on the structure and permeability of lecithin bilayer membranes. Small L-alpha-dimyristoyllecithin (DML) vesicles with myristoyllysolecithin (MLL) incorporated as well as small L-alpha-dipalmitoyllecithin (DPL) vesicles with palmitoyllysolecithin (PLL) were studied by nuclear magnetic resonance (NMR) methods at temperatures both above and below the alpha-gel in equilibrium liquid crystalline phase transition temperature (Tc) and as a function of the concentration of the incorporated lysolecithin. Europium (III) ion was used as a probe to measure the permeability of the vesicular bilayer membrane. At temperatures below Tc, these vesicles were found to be extremely permeable to europium (III) ions. The ion translocation was found to be too fast to be measured by the NMR method under these conditions. However, above the phase transition temperature the ionic permeability decreases to a rate which could be conveniently monitored, and the permeability was shown to increase with temperature and lysolecithin concentration. Analysis of the lysolecithin concentration dependence suggests the formation of ion channels within the lipid bilayer involving four lysolecithin molecules. The data below Tc suggest a phase separation below the phase transition temperature of the host lipid, leading to the formation of patches of lysolecithin molecules within the lecithin matrix. These lysolecithin clusters are presumably long-lived under these conditions and are sufficiently structurally perturbed or disordered to serve as channels for rapid ion permeation.

Two histone IV fragments obtained by cleavage at Met-84 by cyanogen bromide have been examined by proton magnetic resonance (PMR) spectroscopy as a function of temperature, peptide concentration, ionic strength, and pD. Sedimentation and gel electrophoresis studies on these peptides have also been carried out. The 220-MHz PMR spectrum of the N-peptide in both the high- and low-field regions was shown to be almost identical with that calculated for an extended coil N-peptide monomer. The calculated random coil and experimental C-peptide spectra, on the other hand, differ in many respects. Evidence was obtained for the presence of rigid secondary structure in the C-peptide. In addition, the Val, Leu, Ile CH3 resonance displays a prominent high-field satellite band which shifts downfield with increasing temperature. Sedimentation studies on the N-peptide reveal the formation of extremely large, remarkably homogeneous aggregates at ionic strengths larger than or equal to 0.01. The C-peptide, on the other hand, does not appear to form aggregates of sufficient size to be detectable in velocity sedimentation studies of a few hours duration. The relative area changes which have previously been noted in the PMR spectrum of histone IV with increasing ionic strength were also observed for the N-peptide but not the C-peptide. Interpretation of these relative area changes has been made in terms of the amino acid sequence of histone IV, and an effort was made to identify that segment of the polypeptide which undergoes secondary structural change with increasing ionic strength.

Oxidized cytochrome c oxidase is shown to exist in three conformations in addition to the transient "g5" conformation previously reported [Shaw, R. W., Hansen, R. E., & Beinert, H. (1978) J. Biol. Chem. 253, 6637-6640]. The "resting" and "g12" conformations are distinguished by an NO-induced cytochrome a3 electron paramagnetic resonance (EPR) signal and an EPR signal at g' = 12, respectively. The "oxygenated" conformatin exhibits an unusual EPR signal in the presence of fluoride and is identical with the "oxygenated" state first discovered by Okunuki et al. [Okunuki, K., Hagihora, B., Sekuzu, I., & Horio, T. (1958) Proc. Int. Symp. Enzyme Chem., Tokyo, Kyoto, 264]. It is proposed that when the reduced enzyme is reoxidized by dioxygen, the oxidized enzyme first relaxes from the "g5" into the "oxygenated" conformation after which a percentage of the molecules slowly relax into the "g12" conformation. The "resting" conformation is not formed when the enzyme is reoxidized. On the basis of the EPR observations, it is proposed that these various conformations of the oxidized enzyme differ in the structure of the cytochrome a3--Cua3 site. Structures for the cytochrome a3--Cua3 site are proposed for each conformation, and a mechanism by which these conformations undergo interconversion among themselves is described.

Gramicidin S (GrS), an amphiphilic cyclosymmetric decapeptide produced by Bacillus brevis G-B and Nagano, binds nucleotides in water to yield a complex which partitions into organic solvents. The observed phase-transfer efficiencies at a given pH increase in the order AMP less than ADP less than ATP. The lipophilic complexes have well-defined stoichiometries, which were determined to be 1:1 for ADP-GrS at pH 7 and ATP-GrS at pH 3 and 1:2 for ATP-GrS at pH 7. The interaction is primarily ionic, involving coordination of the ornithine N delta H3+ groups of the peptide and the phosphoryl groups of the nucleotide, with little contribution from the nucleoside moiety. Exchange of organic and inorganic phosphates was also found to be mediated by GrS. The nucleotide complexes are sparingly soluble in water and self-associate extensively in CHCl3, most likely by cross-beta-aggregation, to yield large, ribbonlike aggregates which give rise to broad NMR resonances. Structures for the 1:1 and 1:2 complexes are proposed. In the latter, two GrS molecules envelop the nucleotide, orienting their apolar faces externally in opposite directions, while the lateral faces retain considerable polar character and direct aggregation in organic media. The 1:1 complex possesses a single apolar face and is less lipophilic. Binding constants were estimated by simulation of the extraction data. For the 1:1 complexes, K1:1 congruent to 4 X 10(4) M-1 for either ADP or ATP. Phase transfer of the ATP complex at pH 7 could be modeled either by stochastically independent binding to two noninteracting sites on the nucleotide with K1 approximately K2 approximately K1:1 or by a sequential process with K1 approximately K1:1 and K2/K1 less than 100. It is concluded that the apparent selectivity of GrS for ATP over ADP is a consequence of the greater lipophilicity and tendency to aggregate of the 1:2 complex, rather than an intrinsically higher binding affinity for triphosphates. GrS is, to our knowledge, the first peptide known to possess phase-transfer activity toward nucleotides; this is, in addition, the first molecular recognition process in which GrS is demonstrated to participate in vitro at physiologically active concentrations.

Proton and deuterium order parameters measured for the liquid crystalline phase of unsonicated lipid bilayer membranes are interpreted in terms of two motions: (i) chain reorientation and (ii) chain isomerization via kink diffusion. The observed order parameters are found to be compatible with angular deflections of the chain of about 50 degrees with respect to the bilayer normal, coupled with a probability of trans orientation of a methylene segment in the upper part of the chain of about 0.8-0.9. The motional model can be shown to account for the dynamic properties of the membrane system as measured by nuclear magnetic relaxation measurements, assuming that the chain isomerization occurs at a rate of approximately 10(10) s-1 and chain reorientation at a rate of approximately 10(7) s-1. Analysis of proton and deuterium line-width data in terms of this model shows that sonication has the effect of increasing the rate and amplitude of chain reorientation without substantially changing the isomerization motion along the acyl chain. These conclusions are briefly compared with similar observations recently reported in Raman spectroscopic studies.

Cytochrome c oxidase in which the CuA site has been perturbed by extensive modification of the enzyme with the thiol reagent p-(hydroxymercuri)benzoate has been reconstituted into phospholipid vesicles. The reconstituted vesicles lack respiratory control, and the orientation of the enzyme in the vesicles is similar to that of the native cytochrome c oxidase. In the proton translocation assay, the vesicles containing the modified enzyme behave as if they are unusually permeable to protons. When the modified and native proteins were coreconstituted, a substantial portion of the latter became uncoupled as revealed by low respiratory control and low overall proton pumping activity. These results suggest that the modified enzyme catalyzes a passive transport of protons across the membrane. When milder conditions were used for the chemical modification, a majority of the thiols reacted while the CuA site remained largely intact. Reconstitution of such a partially modified cytochrome c oxidase produced vesicles with respiratory control and proton translocating activity close to those of reconstituted native enzyme. It thus appears that the appearance of a proton leak is related to the perturbation of the CuA site. These observations suggest that the structure of CuA may be related to the role of this site in the proton pumping machinery of cytochrome c oxidase.

The interaction of alamethicin with artificial lecithin multilamellar dispersions was investigated by nuclear magnetic resonance (NMR) and Raman spectroscopies. 31P NMR studies revealed perturbation of the lipid head groups in the presence of the icosapeptide. Simulation of the 31P NMR spectra indicated that the observed spectral changes could be attributed to slight variations in the average tilt angle of the head groups. In contrast, no noticeable effect of the peptide on the segmental order of the hydrophobic acyl chains of the lipid molecules was detected by 2H NMR and Raman spectroscopic measurements. Taken together, these results indicated that, in the absence of a transmembrane electric potential, alamethicin interacts primarily at the water-lipid interface without significant insertion or incorporation into the bilayer leaflet.

The effect of cholesterol concentration on the thermotropic phase behavior of aqueous phospholipid multi-bilayers was monitored by means of electron spin resonance spectroscopy (ESR) of a cholestane spin-label (CSL). The spin-label itself induces an additional transition in several different phospholipids, which is attributed to local melting around the spin probe. In contrast, cholesterol prevents its neighboring phospholipids from undergoing fluidization. Small additions of cholesterol affect the position of the probe-induced lipid mobilization curve. The phospholipid main gel-liquid-crystal transition, which is also observed as a separate change in probe mobilization, is not affected by low concentrations of cholesterol. These observations indicate the presence of two phases, a cholesterol-rich phase and a pure phospholipid phase, and indicate that CSL preferentially enters the cholesterol-rich phase. Addition of more than 20 mol % cholesterol abolishes the bulk phospholipid phase. This is evidenced by the disappearance of the gel-liquid-crystal transition as observed by ESR. However, the CSL-induced transition is present at all concentrations of cholesterol and CSL. The behavioral differences between the two sterols caution against using this probe as a direct substitute for cholesterol. However, it remains a useful tool for monitoring the phase behavior of cholesterol-phospholipid bilayer systems.