Effects of lipid fluidity on quenching characteristics of tryptophan fluorescence in yeast plasma membrane.

Fluorescence characteristics of tryptophan residues in yeast plasma membrane indicate that the residues are buried. The fluorescence is fully quenchable by iodide with similar quenching kinetics at temperatures from 8 to 37 degrees C in oleate-enriched membranes and from 25 to 37 degrees C in palmitelaidate-enriched membranes. Substantial increases in lipid microviscosity in palmitelaidate-enriched membranes reduce the fraction of quenchable tryptophan fluorescence by about 40% and increase the effective quenching constant 3-fold. These observations indicate that at above 25 degrees C, proteins in this membrane undergo transient conformational changes and that freedom of conformational changes of the proteins is regulated by lipid microviscosity.

Fluorescence characteristics of tryptophan residues in yeast plasma membrane indicate that the residues are buried. The fluorescence is fully quenchable by iodide with similar quenching kinetics at temperatures from 8 to 37 "C in oleate-enriched membranes and from 25 to 37 "C in palmitelaidate-enriched membranes. Substantial increases in lipid microviscosity in palmitelaidate-enriched membranes reduce the fraction of quenchable tryptophan fluorescence by about 40% and increase the effective quenching constant %fold. These observations indicate that at above 25 "C, proteins in this membrane undergo transient conformational changes and that freedom of conformationd changes of the proteins is regulated by lipid microviscosity.
Tryptophan fluorescence and phosphorescence of some native soluble proteins are quenched by a number of quenchers even though fluorescence spectra and, where available, x-ray data suggest that the fluorophores are buried and are out of direct contact with the solvent (1-5). Consequently, it has been suggested that, in solution, proteins undergo structural fluctuations on the nanosecond scale that permit penetration of the quencher to the fluorophore(s) (1-5). There is no evidence to suggest that this is also true of membrane proteins in situ, and, if so, whether the conformational fluctuations are affected by lipid fluidity. The latter point is particularly important since, for optimal activity of some membranebound enzymes, lipids must be in the liquid-crystalline state (6-11). We have evaluated the effects of fatty acid composition and lipid microviscosity on the kinetics of iodide quenching of tryptophan fluorescence in yeast plasma membrane. The results indicate that membrane proteins undergn transient conformational changes and are in a "fluid" state at physiolcnical temperatures. The results also indicate that freedom of conformational changes depends on lipid microviscosity.

MATERIALS AND METHODS
Fatty acids and DPH' were obtained from Sigma. Plasma membrane was isolated, as previously described (12), from a fatty hcid auxotroph of Saccharomyces cereuisiae grown at 30 "C in a medium containing 0.02% each of myristic and oleic or palmitelaidic acid (13). Tryptophan content of the membrane, determined according to * This work was supported by United States Public Health Service Grant GM-26452. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Iodide quenching experiments were carried out essentially as described by Lehrer (15). Each assay mixture contained membrane (0.1 mg of protein) in 2 ml of 0.05 M potassium phosphate buffer, pH 7.2, and unless otherwise indicated, 0 to 0.5 M KI. Ionic strength was adjusted to 1 by addition of KCl. Iodide solutions contained 1 X M S203-' to prevent 13-formation (15). Unless otherwise indicated, the reaction mixture was maintained for 15 min at 25 "C and tryptophan fluorescence emission was recorded at the same temperature.
DPH polarization was carried out, as described earlier (16), except that 0.1 mg of membrane protein was used. Spectra were recorded with a Perkin-Elmer MPF-44 fluorescence spectrophotometer. Temperature was maintained as described before (16). Excitation wavelengths were 290 and 358 nm for tryptophan and DPH, respectively.
Quantum yield was determined according to Kirby (17).

RESULTS AND DISCUSSION
Fluorescence Spectra-The fluorescence emission spectrum of the membrane excited at 290 nm exhibits a single broad maximum at 335 nm with a half band width of approximately 60 nm (Fig. 1). These are characteristics of tryptophan emission (18, 19) and remain the same when 1 M KC1 or 1 M KI are added. In 6 M guanidine HC1, the emission max is shifted to 350 nm. These indicate the fluorescing tryptophyl residues are shielded from water and in 6 M guanidine HCl they are transfered to the aqueous environment (1, 4, 5, 15, 20,21). It should be noted that iodide even at 2 M did not shift the X, , , .
Quenching by Iodide-Iodide is a general quencher of fluorophores (22). Quenching kinetic parameters can be calculated from a modified Stern-Volmer equation (15), where Fo is fluorescence intensity in the absence of the quencher ( X ) , A F is the change in fluorescence intensity due to addition of a given concentration of X , K,, is a constant which is the product of the collisional quenching rate constant and fluorescence life-time in the absence of the quencher, and fa is the effective fraction of the tryptophanyl fluorescence that is quenchable. The effects of 1 M iodide on fluorescence intensity of tryptophan residues and of DPH incorporated into the membrane are shown in Fig. 2. DPH is not quenched. Since DPH is located in the paraffinic region of the lipid-bilayer (23) and in 60% ethanol is quenched by I-, with a K,, of approximately 1 M", then the quencher cannot diffuse through the lipid-bilayer. The kinetic parameters of tryptophan quenching in oleate-enriched membranes were calculated from modified Stern-Volmer plots (Fig. 3A). fa was found to be 1 and K,, was 1.3 M-'. These values changed little at assay temperatures ranging from 8 to 37 "C. Since fa is 1, then either the fluorescence of each tryptophan residue is susceptible to quenching by iodide (15) or, alternatively, some may be accessible but all are connected by energy transfer networks. In either case, such assessibility is feasible if transient conformational changes, similar to those proposed for soluble proteins ( 1 4 , take place in membrane proteins. Effect quenching kinetics with plasma membranes enriched with oleate or palmitelaidate was conducted at different temperatures. The results are shown in Figs. 3-6, Fluorescence polarization values for DPH, determined in order to monitor microviscosity changes (23), were the same for the two types of membranes both a t 25 and 37 "C, but were substantially higher in palmitelaidate-enriched membranes a t 8 and 15 "C ( Fig. 4). Gas chromatography of methylated fatty acids of the membrane phospholipid fraction showed that the exogenous unsaturated fatty acid plus its derivative constituted 70% of the total in the palmitelaidate-enriched membrane and 45% in the oleate-enriched membrane. The relationship between DPH polarization values and quenching parameters, Ks, and fa, are shown in Fig. 5. In the oleate-enriched membrane, polarization increased from 0.16 a t 37 "C to 0.23 at 8 "C with no changes in either fa or K,, which were 1  The changes in fa or K,, cannot be simply due to the effect of temperature per se since both fa and K,, were constant from 8 to 37 "C in the oleate-enriched membrane. The quantum yield for both types of membrane was determined and found to be 0.04 at 25 "C, and 0.05 at 8 "C. Therefore, the increase in K,, must be due to an increase in collisional quenching rate constant and not due to increases in fluorescence life-time The effect of decreased fluidity on kinetics of iodide quenching is manifest in yet another fashion. All Stern-Volmer plots (24) were found to be linear at the assay temperature of 8 to 37 "C in the oleate-enriched membrane (not shown) and at 25 "C and above in the palmitelaidate-enriched membrane (Fig. 6). However, the plots deviated from linearity in the palmitelaidate-enriched membrane at 15 "C and below (Fig.   (25-27).

6). This indicates that at
higher lipid viscosity the tryptophyl residues are no longer subject to a similar degree of fluorescence quenching (15) and must consist of a heterogenous population.
The possibility of a transition from liquid-crystalline to the gel phase leading to the observed changes in quenching kinetics of the pahitelaidate-enriched membrane at 15 "C and below should be considered. Such a phase change could conceivably inhibit possible penetration of Ito the level of glyceryl backbone and prevent quenching of the tryptophanyl groups that might be located in this region. However, neither in a previous report (16) nor in studies of Fig. 4 did we observe such a phase transition in the palmitelaidate-membrane. For such a phase change to occur and to explain our findings, we should have observed a change in slope of DPH polarization of the palmitelaidate-enriched membrane between 15 and 25 "C (23). In addition, the Arrhenius plots of data in Fig. 4 revealed a constant activation energy between 8 and 25 "C. Therefore, the above possibility is ruled out.
The constancy of K,, in the oleate-enriched membrane (Fig.   5) is indicative of a very low energy of activation for quenching and a fluid protein matrix (5). Therefore, at the growth temperature of 30 "C both in this and in the palmitelaidateenriched membrane, where K,, was the same at 25 and 37 "C, there must be a substantial degree of mobility in the poly- peptide chains which may be necessary for optimal membrane function. It is then the restriction of this mobility, as a result of lipid phase transition, that could impair activity of membrane-bound enzymes (6-11).

Lipid Fluidity
The following interpretations are consistent with the observed changes in fa and K,,, respectively. An increase in lipid microviscosity restricts conformational changes of some proteins, preventing collision of the quencher with some of the tryptophan residues. Of the remaining, as a result of conformational changes induced in proteins by less fluid lipids, as has recently been documented for cytochrome oxidase (28), some are placed closer to the solvent. This shortens the path that the quencher must cover in order to collide with the fluorophore, resulting in an increase in the collisional quenching rate constant and a higher K,, value. This suggestion is further substantiated by the observation that at the higher lipid microviscosity (Le. palmitelaidate-enriched membrane at 8 or 15 "C) Stern-Volmer plots were no longer linear, indicating that the fluorescing residues had become heterogenous and subject to different degrees of quenching (15).
The present studies give evidence for transient conformational changes in membrane proteins in situ that are consistent with the existence of protein matrices in a fluid state at growth temperature. They further show that protein fluidity is modulated by alterations in lipid fluidity and give evidence for conformational changes in proteins induced by increased lipid microviscosity. These observations provide a plausible explanation for previously reported effects of membrane lipid composition on activity of some membrane-bound enzymes   Fig. 4). Bars show mean deviations.
FIG . 6 (right). Stern-Volmer plots of the quenching of tryptophan fluorescence by iodide in palmitelaidate-enriched membranes. The data of Fig. 3B were plotted according to the Stern-Volmer equation. Membranes assayed at 25 "C (0) and 8 "C