Béatrice de Foresta

Fluorescence anisotropy decays of n-(9-anthroyloxy) fatty acids in dipalmitoyl phosphatidylcholine vesicles. Localization of the effects of cholesterol addition

Abstract The localization of the effects of cholesterol addition on the dynamic structure of the fatty acyl chains of dipalmitoyl phosphatidylcholine vesicles has been investigated by the time-resolved fluorescence anisotropy technique with a set of n-(9-anthroyloxy) fatty acids probes. The major effect of cholesterol is observed in the 7–9 carbon region where both parameters of the anisotropy decays, the residual anisotropy (r∞) and the correlation time, are greatly enhanced whatever the temperature (21, 37 and 47°C). In the 12–16 carbon region, the r∞ values are lowered upon addition of cholesterol in the gel phase, in agreement with the effect monitored by the 1,6-diphenyl-1,3,5-hexatriene probe. Only slight perturbations on the r∞ values are observed in the 2-carbon region whatever the temperature.

Tryptophan Octyl Ester in Detergent Micelles of Dodecylmaltoside: Fluorescence Properties and Quenching by Brominated Detergent Analogs

The fluorescence properties of tryptophan octyl ester (TOE), a hydrophobic model of Trp in proteins, were investigated in various mixed micelles of dodecylmaltoside (DM) and 7,8-dibromododecyl beta-maltoside (BrDM) or 10,11-dibromoundecanoyl beta-maltoside (BrUM). This study focuses on the mechanism via which these brominated detergents quench the fluorescence of TOE in a micellar system. The experiments were performed at a pH at which TOE is uncharged and almost completely bound to detergent micelles. TOE binding was monitored by its enhanced fluorescence in pure DM micelles or its quenched fluorescence in pure BrUM or BrDM micelles. In DM/BrUM and DM/BrDM mixed micelles, the fluorescence intensity of TOE decreased, as a nonlinear function of the molar fraction of brominated detergent, to almost zero in pure brominated detergent. The indole moiety of TOE is therefore highly accessible to the bromine atoms located on the detergent alkyl chain because quenching by bromines occurs by direct contact with the fluorophore. TOE is simultaneously poorly accessible to iodide (I(-)), a water-soluble collisional quencher. TOE time-resolved fluorescence intensity decay is heterogeneous in pure DM micelles, with four lifetimes (from 0.2 to 4.4 ns) at the maximum emission wavelength. Such heterogeneity may arise from dipolar relaxation processes in a motionally restricted medium, as suggested by the time-dependent (nanoseconds) red shift (11 nm) of the TOE emission spectrum, and from the existence of various TOE conformations. Time-resolved quenching experiments for TOE in mixed micelles showed that the excited-state lifetime values decreased only slightly with increases in the proportion of BrDM or BrUM. In contrast, the relative amplitude of the component with the longest lifetime decreased significantly relative to that of the short-lived species. This is consistent with a mainly static mechanism for the quenching of TOE by brominated detergents. Molecular modeling of TOE (in vacuum and in water) suggested that the indole ring was stabilized by folding back upon the octyl chain, forming a hairpin conformation. Within micelles, the presence of such folded conformations, making it possible for the entire molecule to be located in the hydrophobic part of the micelle, is consistent with the results of fluorescence quenching experiments. TOE rotational correlation time values, in the nanosecond range, were consistent with a hindered rotation of the indole moiety and a rotation of the complete TOE molecule in the pure DM or mixed detergent micelles. These results, obtained with a simple micellar model system, provide a basis for the interpretation of fluorescence quenching by brominated detergents in more complex systems such as protein- or peptide-detergent complexes.

Single-spanning membrane protein insertion in membrane mimetic systems: role and localization of aromatic residues

Membrane protein insertion in the lipid bilayer is determining for their activity and is governed by various factors such as specific sequence motifs or key amino-acids. A detailed fluorescence study of such factors is exemplified with PMP1, a small (38 residues) single-membrane span protein that regulates the plasma membrane H+-ATPase in yeast and specifically interacts with phosphatidylserines. Such interactions may stabilize raft domains that have been shown to contain H+-ATPase. Previous NMR studies of various fragments have focused on the critical role of interfacial residues in the PMP1 structure and intermolecular interactions. The C-terminal domain contains a terminal Phe (F38), a single Trp (W28) and a single Tyr (Y25) that may act together to anchor the protein in the membrane. In order to describe the location and dynamics of W28 and the influence of Y25 on protein insertion within membrane, we carried out a detailed steady-state and time-resolved fluorescence study of the synthetic G13-F38 fragment and its Tyr-less mutant, Y25L in various membrane mimetic systems. Detergent micelles are conveniently used for this purpose. We used dodecylphosphocholine (DPC) in order to compare with and complement previous NMR results. In addition, dodecylmaltoside (DM) was used so that we could apply our recently described new quenching method by two brominated analogs of DM (de Foresta et al. 2002, Eur. Biophys. J. 31:185–97). In both systems, and in the presence and absence of Y25, W28 was shown to be located below but close to the polar headgroup region, as shown by its maximum emission wavelengths (λmax), curves for the quenching of Trp by the brominated analogs of DM and bimolecular constants for quenching (kq) by acrylamide. Results were interpreted by comparison with calibration data obtained with fluorescent model peptides. Time-resolved anisotropy measurements were consistent with PMP1 fragment immobilization within peptide-detergent complexes. We tentatively assigned the two major Trp lifetimes to the Trp (χ1=60° and 180°) rotamers, based on the recent lifetime–rotamer correlation proposed for model cyclic peptides (Pan and Barkley 2004, Biophys J 86:3828–35). We also analyzed the role of the hydrophobic anchor, by comparing the micelle binding of fragments of various lengths including the synthesized full-length protein and detected peculiar differences for protein interaction with the polar headgroups of DM or DPC.

Structural and dynamic properties of juxta-membrane segments of caveolin-1 and caveolin-2 at the membrane interface

Caveolins (cav1–3) are essential membrane proteins found in caveolae. The caveolin scaffolding domain of cav-1 includes a short sequence containing a CRAC motif (V94TKYWFYR101) at its C-terminal end. To investigate the role of this motif in the caveolin–membrane interaction at the atomic level, we performed a detailed structural and dynamics characterization of a cav-1(V94-L102) nonapeptide encompassing this motif and including the first residue of cav-1 hydrophobic domain (L102), in dodecylmaltoside (DM) or dodecylphosphocholine (DPC) micelles, as membrane mimics. Cav-1(V94-L102) partitioned better in DPC and in DM/anionic lipid micelles than in DM micelles, as shown by fluorescence titration and CD. NMR data revealed that this peptide folded as an amphipathic helix located in the polar head group region of DPC micelles. The two tyrosine side-chains, flanked by arginine and lysine residues, are situated on one face of this helix, whereas the phenylalanine and tryptophan side-chains are located on the opposite face. Fluorescence studies showed significant Trp subnanosecond rotations, the presence of several rotamers, and a heterogeneous location within the water/micelle interface. NMR studies of the shorter cav-1(V94-R101) peptide and of the homologous sequence of cav-2(I79SKYVMYKF87) allowed the description of the effect of L102 and of the amino acid variations occurring in cav-2 on the structure and localization in DPC micelles. Based on the topological model of caveolins, our results suggest that the cav-1 and cav-2 nonapeptides studied form interfacial α-helix membrane anchors in which the K/RhhhYK/Rh motif, also found in cav-3, may play a significant role.

The polar headgroup of the detergent governs the accessibility to water of tryptophan octyl ester in host micelles

Many attempts have been made to rationalize the use of detergents for membrane protein studies [J. Biol. Chem. 264 (1989) 4907]. The barrier properties of the detergent headgroup may be one parameter critically involved in protein protection. In this paper, we analyzed these properties using a model system, by comparing the accessibility of tryptophan octyl ester (TOE) to water-soluble collisional quenchers (iodide and acrylamide) in three detergent micelles. The detergents used differed only in the chemical nature of their polar headgroups, zwitterionic for dodecylphosphocholine (DPC) and nonionic for octa(ethylene glycol) dodecyl monoether (C(12)E(8)) and dodecylmaltoside (DM). In all cases, in phosphate buffer at pH 7.5, the binding of 5 microM TOE was complete in the presence of a slight excess of detergent micelles over TOE molecules, resulting in a significant blue shift and greater intensity of TOE fluorescence emission. The resulting quantum yield of bound TOE was between 0.08 (in DPC) and 0.12 (in DM) with an emission maximum (lambda(max)) of approximately 335 nm whatever the detergent micelle. Time-resolved fluorescence intensity decays of TOE at lambda(max) were heterogeneous in all micelles (3-4 lifetime populations), with mean lifetimes of 1.7 ns in DPC, and 2 ns in both C(12)E(8) and DM. TOE fluorescence quenching by iodide, in detergent micelles, yielded linear Stern-Volmer plots characteristic of a dynamic quenching process. The accessibility of TOE to this ion was the greatest with C(12)E(8), followed by DPC and finally DM (Stern-Volmer quenching constants K(sv) of 2 to 5.5 M(-1)). In contrast, the accessibility of TOE to acrylamide was greatest with DPC, followed by C(12)E(8) and finally DM (K(sv)=2.7-7.1 M(-1)). TOE also presents less rotational mobility in DM than in the other two detergents, as shown from anisotropy decay measurements. These results, together with previous TOE quenching measurements with brominated detergents [Biophys. J. 77 (1999) 3071] provide reference data for analyzing Trp characteristics in peptide (and more indirectly protein)-detergent complexes. The main finding of this study was that TOE was less accessible (to soluble quenchers) in DM than in DPC and C(12)E(8), the cohesion of DM headgroup region being suggested to play a role in the ability of this detergent to protect function and stability of solubilized membrane proteins.

Effects of temperature and benzyl alcohol on the structure and adenylate cyclase activity of plasma membranes from bovine adrenal cortex.

Adenylate cyclase activation by corticotropin (ACTH), fluoride and forskolin was studied as a function of membrane structure in plasma membranes from bovine adrenal cortex. The composition of these membranes was characterized by a very low cholesterol and sphingomyelin content and a high protein content. The fluorescent probes 1,6-diphenylhexa-1,3,5-triene (DPH) and a cationic analogue 1-[4-(trimethylamino)phenyl]-6-phenylhexa-1,3,5-triene (TMA-DPH) were, respectively, used to probe the hydrophobic and polar head regions of the bilayer. When both probes were embedded either in the plasma membranes or in liposomes obtained from their lipid extracts, they exhibited lifetime heterogeneity, and in terms of the order parameter S, hindered motion. Under all the experimental conditions tested, S was higher for TMA-DPH than for DPH but both S values decreased linearly with temperature within the range of 10 to 40 degrees C, in the plasma membranes and the liposomes. This indicated the absence of lipid phase transition and phase separation. Addition to the membranes of up to 100 mM benzyl alcohol at 20 degrees C also resulted in a linear decrease in S values. Membrane perturbations by temperature changes or benzyl alcohol treatment made it possible to distinguish between the characteristics of adenylate cyclase activation with each of the three effectors used. Linear Arrhenius plots showed that when adenylate cyclase activity was stimulated by forskolin or NaF, the activation energy was similar (70 kJ.mol-1). Fluidification of the membrane with benzyl alcohol concentrations of up to 100 mM at 12 or 24 degrees C produced a linear decrease in the forskolin-stimulated activity, that led to its inhibition by 50%. By contrast, NaF stabilized adenylate cyclase activity against the perturbations induced by benzyl alcohol at both temperatures. In the presence of ACTH, biphasic Arrhenius plots were characterized by a well-defined break at 18 degrees C, which shifted at 12.5 degrees C in the presence of 40 mM benzyl alcohol. These plots suggested that ACTH-sensitive adenylate cyclase exists in two different states. This hypothesis was supported by the striking difference in the effects of benzyl alcohol perturbation when experiments were performed below and above the break temperature. The present results are consistent with the possibility that clusters of ACTH receptors form in the membrane as a function of temperature and/or lipid phase fluidity.(ABSTRACT TRUNCATED AT 400 WORDS)

ANALYSIS OF EFFECTS OF CORTICOTROPIN, FORSKOLIN AND FLUORIDE ON ACTIVITY OF ADENYLATE CYCLASE OF BOVINE ADRENAL CORTEX

A mathematical model relating the activity of adenylate cyclase (AC) with concentrations of stimulators, equilibrium dissociation constants, specific activity and efficacies of AC depending on the states of its binding sites has been developed and used for analysis of the data on activation of AC of bovine adrenal cortex plasma membranes presented in (De Foresta et al. (1987) FEBS Lett. 216, 107-112). Equilibrium dissociation constants. chi h and chi l, corresponding to high- and low-affinity forskolin-binding sites were estimated to be 0.37 and 17 microM: these constants characterize forskolins potency more adequately than does ED50, the concentration eliciting half-asymptotic activity of AC. Corticotropin does not affect the affinity of AC for forskolin whereas fluoride increases this affinity, thus augmenting forskolins potency. Hormone receptor of adenylate cyclase of bovine adrenal cortex has been suggested to have two or more binding sites for corticotropin. Some unidentified factor(s) may be responsible for the differences found in adenylate cyclase activity in different experiments carried out under similar conditions. The model applied for the analysis may be thought to be the best means for the moment to relate dose-response dependencies with what is known or can be hypothesized about the mechanisms underlying activation of adenylate cyclase.

Study of fluorescent tryptophyl residues and extrinsic probes for the characterization of molecular domains of Folch-Pi apoprotein.

The highly hydrophobic myelin Folch-Pi apoprotein can be solubilized in organic as well as in aqueous media. In order to understand the molecular organization changes consecutive to changes in the solvent medium, the environment of intrinsic probes and extrinsic labels has been studied by fluorescence and accessibility to some reagents. In acqueous solution, only two tryptophan residues per protein molecule of 23,500 molecular weight have been shown to fluoresce, and their fluorescence characterisitics indicate an hydrophobic and/or constrained environment. Two ANS binding sites have also been observed having a high quenching effect on the intrinsic chromophore fluorescence. A large accessibility has been evidenced for the protein sulfhydryl groups in chloroform-methanol 2:1 (v/v), both by kinetic study of the protein reaction with a specific reagent, N-(1-anilino-naphtyl-4) maleimide, and by the fluorescence characteristics of this probe once linked to the protein. The free sulfhydryl groups were still reactive in acqueous solution, but extrinsic fluorescence of the labelled apoprotein transferred from chloroform-methanol 2:1 (v/v) into water gave evidence of constraints on the probe or on its environment. Such constraints may contribute to the solubilization in acqueous solution of this highly hydrophobic protein.

Interaction with membrane mimics of transmembrane fragments 16 and 17 from the human multidrug resistance ABC transporter 1 (hMRP1/ABCC1) and two of their tryptophan variants

The human multidrug resistance-associated protein 1 (hMRP1/ABCC1) belongs to the ATP-binding cassette transporter superfamily. Together with P-glycoprotein (ABCB1) and the breast cancer resistance protein (BCRP/ABCG2), hMRP1 confers resistance to a large number of structurally diverse drugs. The current topological model of hMRP1 includes two cytosolic nucleotide-binding domains and 17 putative transmembrane (TM) helices forming three membrane-spanning domains. Mutagenesis and labeling studies have shown TM16 and TM17 to be important for function. We characterized the insertion of the TM16 fragment into dodecylphosphocholine (DPC) or n-dodecyl-beta-d-maltoside (DM) micelles as membrane mimics and extended our previous work on TM17 (Vincent et al., 2007, Biochim. Biophys. Acta 1768, 538). We synthesized TM16 and TM17, with the Trp residues, W1198 in TM16 and W1246 in TM17, acting as an intrinsic fluorescent probe, and TM16 and TM17 Trp variants, to probe different positions in the peptide sequence. We assessed the interaction of peptides with membrane mimics by evaluating the increase in fluorescence intensity resulting from such interactions. In all micelle-bound peptides, the tryptophan residue appeared to be located, on average, in the head group micelle region, as shown by its fluorescence spectrum. Each tryptophan residue was partially accessible to both acrylamide and the brominated acyl chains of two DM analogs, as shown by fluorescence quenching. Tryptophan fluorescence lifetimes were found to depend on the position of the tryptophan residue in the various peptides, probably reflecting differences in local structures. Far UV CD spectra showed that TM16 contained significant beta-strand structures. Together with the high Trp correlation times, the presence of these structures suggests that TM16 self-association may occur at the interface. In conclusion, this experimental study suggests an interfacial location for both TM16 and TM17 in membrane mimics. In terms of overall hMRP1 structure, the experimentally demonstrated amphipathic properties of these TM are consistent with a role in the lining of an at least partly hydrophilic transport pore, as suggested by the currently accepted structural model, the final structure being modified by interaction with other TM helices.

Adenylate cyclase of bovine adrenal cortex plasma membranes: divergence between corticotropin and fluoride combined effects with forskolin

The diterpene forskolin maximally stimulated bovine adrenal cortex adenylate cyclase activity 9‐fold with a concentration producing half‐maximum effect (ED50) of about 4 μM. The effects of forskolin and the fully active corticotropin fragment ACTH (1–24) were additive over nearly the whole range of concentration of both effectors, indicating separate and independent mechanisms of action. By contrast, 10 mM NaF blocked forskolin action in the nanomolar range of the diterpene concentration, while it allowed a partial stimulation by forskolin in the micromolar range. NaF thus reveals a heterogeneity of forskolin action in the adrenal cortex plasma membranes. Moreover, our data suggest that ACTH and NaF activation effects, both mediated by the stimulatory regulatory protein Gs, proceed through different mechanisms.