Dark, Ultra-Dark and Ultra-Bright Nanodiscs for membrane protein investigations

Abstract

We describe the construction, expression and purification of three new membrane scaffold proteins (MSP) for use in assembling Nanodiscs. These new MSPs have a variety of luminescent properties for use in combination with several analytical methods. “Dark” MSP has no tryptophan residues, “Ultra-Dark” replaces both tryptophan and tyrosine with non-fluorescent side chains, and “Ultra-Bright” adds additional tryptophans to the parent membrane scaffold protein to provide a dramatic increase in native tryptophan fluorescence. All MSPs were used to successfully assemble Nanodiscs nominally 10 nm in diameter, and the resultant bilayer structure was characterized. An example of the usefulness of these new scaffold proteins is provided.

Graphical Abstract

1. Introduction

The challenges in dealing with integral and peripheral membrane proteins using the usual biochemical and analytical methods are well known. Several developments have helped alleviate these challenges and one of the most successful has been the Nanodisc platform. Developed almost 20 years ago [1–3], a genetically engineered amphipathic helical protein, termed membrane scaffold protein (MSP), was used to self assemble phospholipids into a soluble membrane bilayer structure. Amazingly, integral membrane proteins can be self-assembled into this bilayer, providing full functionality of the incorporated target [4]. In addition, various sizes of soluble discoidal bilayers can be generated (~6 to 16 nm) providing a membrane surface of defined area and composition for studies of peripheral protein binding events. The Nanodisc system has enjoyed wide, and increasingly growing, popularity allowing structural and functional investigations of membranes and membrane protein complexes. Several recent reviews further define the Nanodisc system and report successful examples.[3–8] The membrane scaffold protein (MSP) most commonly used, MSP1D1, generates discoidal bilayers ~10 nm in diameter and was engineered from the human Apo-AI protein sequence as represented in high density lipoprotein particles.[2] The sequence, shown in Figure 1, codes for a class A amphipathic helix motif [9] with short segments broken by proline residues. The short helices allow the MSP to assemble around the Nanodisc acting as a belt to stabilize the discoidal structure. In recent years several other novel membrane mimetic systems have been developed that also generate soluble discoidal particles that harbor desired lipid content and target proteins for subsequent biophysical characterization. These include the peptide based discs that use small polypeptides to stabilize the disc [10–12] and discs that are stabilized by polymers such as Styrene/Maleic Acid (SMA) and Diisobutylene/Maleic Acid (DIMBA) [13–15]. In this report we add to the ever-expanding repertoire of available membrane mimetic tools by introducing several MSP1D1 variants.

Figure 1.

MSP1D1 sequence: Helices are overlaid with blue cylinders, The 7X Histidine tag and TEV protease site are orange, tyrosines (*, yellow), tryptophans (†, green), and new tryptophan sites (‡,red). Tyrosine 236 is highlighted in red as a new tryptophan site.

Many applications of Nanodiscs to study membrane proteins and lipid recognition make use of fluorescence methods, including lifetime, resonance energy transfer, and anisotropy.[16–20] Typically, this requires labeling of target proteins with fluorophores that have convenient excitation and emission properties. One of the first examples applying fluorescence to the Nanodiscs system was the use of FRET to measure the binding of arrestin to light activated rhodopsin assembled into the Nanodisc bilayer. [16] In this case the MSP was labeled with a donor fluorophore, Texas Red, and arrestin was labeled with the dark quencher, QXL 610, allowing the measurement of arrestin binding to rhodopsin incorporated into labeled Nanodiscs via quenching of the Texas Red emission. More recently, fluorescence techniques have been expanded to a “label free” technique where binding of proteins to Nanodisc bilayers can be monitored by fluorescence anisotropy. To achieve this the Nanodisc protein belts are labeled with a long lifetime dye (~1 µs) which allow the measurement of the change in rotational correlation time of the Nanodisc upon binding of an unlabeled membrane protein. This technique was used to characterize the interaction of the oncogenic KRas4b to anionic lipids. [21,22]

Often it is desirable to have a label free detection modality for monitoring binding and function which avoids the challenges of affixing a fluorescent label to a protein of interest, thus avoiding the risk of altering its inherent structure or activity. Nearly all proteins have intrinsic fluorescence resulting from the presence of tryptophan and tyrosine which can be leveraged as a detection modality. Examining the sequence of MSP1D1 (Figure 1), one notes there exist two tryptophan residues, five tyrosines and three phenylalanines. Trp fluorescence can be used to detect binding by monitoring the Stokes shift that reports on changes in the Trp environment, or anisotropy, which reports on the mobility of the Trp residues. In addition, the local environment of tyrosine can be assessed using second-derivative optical spectroscopy, as has been used to provide insight into the conformational changes upon ligand binding or interactions with other proteins or peptides. [23,24] To improve on these detection modalities, we constructed the two MSP variants described in this manuscript that generate the ~10 nm Nanodiscs: Dark MSP1D1 (D-MSP1D1), where all tryptophans († in Figure 1) have been replaced with a phenylalanine sidechain and Ultra Dark MSP1D1 (UD-MSP1D1), where, in addition, all tyrosines (* in Figure 1) are replaced by phenylalanine. We characterized the MSPs by mass spectrometry and assembled Nanodiscs with each of these new MSPs, documenting their biophysical properties, including absorption spectra, excitation and emission spectra, fluorescence lifetime, and anisotropy decay. We also provide a representative experiment to demonstrate the usefulness of UD-MSP’s by measuring the binding of the small peptide, indolicidin, to DMPC/DMPS containing Nanodiscs in a completely label free manner. Indolicidin is a 13 amino acid antimicrobial peptide isolated from bovine neutrophils and where the mechanism of action is believed to be dependent on the anionic continent of the membrane bilayer. [25–27]

In addition to the dark MSP’s we also characterize another MSP, Ultra Bright (UB-MSP1D1), which adds three new tryptophans to the MSP sequence (‡ in Figure 1). This MSP variant can be useful in monitoring the properties of Nanodiscs as well as acting as an intrinsic FRET donor for labeled proteins interacting with the bilayer surface. Techniques such as Analytical Ultra Centrifugation (AUC) can benefit from UB-MSP1D1, which can be used to measure binding of target proteins to the Nanodisc lipid bilayer and can be performed with the intrinsic fluorescence of UB-MSP1D1 without the need for extrinsic fluorophores.[28]

2. Materials and Methods

2.1. Construction of MSP variants.

The Generation of “dark” MSP (W71F, W108F) in our laboratory was described previously.[29] The plasmids with “ultra dark” MSP (W72F, Y100F, W108F, Y115F, Y166F, Y192F, Y236F) and “ultrabright” MSP (Q138W, L181W, Y236W) were generated using standard protocols. Briefly, the EcoR1-Sac1 gene fragment of MSP1D1 was replaced by the corresponding synthetic G-block fragments, synthesized by IDT DNA, Inc., which contained ~15-bp overlapping sequences. MSP1D1 plasmid was digested with Eco R1 and Sac1 restriction endonucleases. After separation on agarose gel, the 5.7 kbp fragment of the MSP1D1 plasmid was ligated using Gibson Assembly (New England Biolabs, Inc). After initial identification of the mutant clones by digestion mapping, the presence of the desired mutations were confirmed by DNA sequencing. Figure 1 shows a schematic representation of the location of the mutants with the helical structure of MSP1D1 overlaid as blue cylinders.

2.2. Protein Purification

MSPs were purified as previously described. [6] For UD-MSP1D1 the 7X histidine tag was removed by incubation with 6X His tagged-TEV protease in 20 mM Tris HCl 100mM NaCl, pH 7.5, 1 mM DTT at 30 C for 4 hours. TEV protease and unclipped MSP are removed from the reaction by first dialyzing the sample against buffer without DTT, then passing the sample over a Ni NTA column equilibrated in 20 mM Tris-HCL 100 mM NaCl, 20 mM imidazole and collecting the flow through fraction. The resulting protein, denoted as UD-MSP1D1(−), is completely devoid of Trp and Tyr. Concentrations of MSP’s were estimated using the computed extinction coefficients at 280 nm obtained using the ProtParam tool [30] with the exception of UD-MSP1D1(−). UD-MSP1D1(−) concentrations were estimated using Coomassie Plus protein assay (Pierce) based on total protein mass by employing unclipped UD-MSP1D1 as a standard. Electrospray ionization mass spectrometry was performed by the University of Illinois mass spectrometry laboratory.

2.3. Nanodisc assembly

The process of assembling lipids and MSPs into Nanodiscs has been extensively described.[4] Lipids, either DMPC (1,2-dimyristoyl-sn-glycero-3-phosphocholine) or a 70:30 mixture of DMPC:DMPS (1,2-dimyristoyl-sn-glycero-3-phosphoserine), were dissolved in chloroform and (importantly) concentration was measured by determination of total phosphorous.[31,32] Chloroform stocks were dried as a thin film in a glass test tube under a steady stream of nitrogen. The film was further dried in a vacuum desiccator for 4 – 12 hours at room temperature. The lipids were then solubilized with sodium cholate (200 mM) and mixed with the MSP in 20 mM Tris-HCl, 100 mM NaCl, pH 7.5 at a ratio of total lipid:protein of 85:1 and incubated at room temperature (~23 C) for 20 minutes. Cholate was removed from the mixture by the addition 0.5 volumes of Amberlite XAD beads (Sigma) and incubated for 4 hours at room temperature with gentle agitation. The resulting self-assembled Nanodiscs were filtered through 0.22 micron spin filters (Ultrafree-MC, Millipore Sigma) and purified using size exclusion chromatography on a Superdex 200 column (GE Life Sciences, Piscataway, NJ).

2.4. Spectroscopy

UV-VIS absorption measurements were made on a Varian 300 spectrophotometer from 250 nm to 350 nm with a 2 nm band pass. Steady state fluorescence spectra and lifetimes used a K2 fluorometer equipped with excitation and emission polarizers and a digital frequency domain module (DFD) for lifetime and anisotropy decay measurements (ISS, Urbana, IL, USA). Steady state emission spectra were collected using a Xenon arc lamp equipped with a monochromator as an excitation source. Cross correlation phase fluorometry was used to measure fluorescence lifetimes and anisotropy decays [33,34]. A 286 nm LED was used as an excitation source and modulated from 5 to 150 MHz for phase modulated lifetime and anisotropy decay measurements. Intensities in the frequency domain were measured with a PMT equipped with a 325 nm cutoff filter (Edge Basic 325, Semrock). 1,4-diphenylbenzene (p-terphenyl) (1.05 ns) was used as a lifetime standard for all lifetime measurements. [35,36] Fluorescence lifetimes and anisotropy decays were analyzed by the methods previously reported using the Vinci software package (ISS, Urbana, IL USA) and in house Matlab routines. [33,34,37,38]

2.5. Peptide-Nanodisc interactions

Indolicidin was purchased from Anaspec and dissolved in water at 1 mg/ml. Concentrations were measured using a computed extinction coefficient of 27500 mM⁻¹ at 280 nm. Indolicidin was diluted to 110 nM and titrated with UD-MSP1D1 Nanodiscs containing either 70% DMPC / 30% DMPC or 100% DMPC. The samples were excited with 280 nm light and the polarization was measured through 325 nm cut off filters. In some experiments, full emission spectra were collected at each titration point to monitor the change in Stokes shift upon binding of Nanodiscs.

3. Results and Discussion

3.1. Characterization of MSP Varaints

UD-MSP1D1, D-MSP1D1 and UB-MSP1D1 were purified as described in the methods. Computed extinction coefficients and expected molecular weights are reported in Table 1. To confirm the production of the MSP variants, each protein sample was subjected to electrospray ionization mass spectrometry (University of Illinois Mass Spectrometry Lab). In all cases the measured mass confirms the production of the expected protein variants, indicating the MSP’s are stable during bacterial expression.

Table 1.

Physical Properties of MSP1D1 Variants

	Extinction Coefficient 280 nm (cm−1 M−1)	MW (Da)	Measured Mass (Da)
UB-MSP1D1	36440	24816	24815
MSP1D1	21430	24662	24664
D-MSP1D1	10430	24583	24586
UD-MSP1D1	2980	24503	24583

Figure 2A shows the UV-VIS absorbance spectra of the variant MSP’s. The positions of absorbance maximums are consistent with those expected based on computed extinction coefficients. For UD-MSP1D1 in the absence of tryptophan and tyrosine we observe several bands between 250 nm and 270 nm due to the vibrational structure of the phenyalanine benzene ring.[39] Figure 2B shows the concentration normalized emission spectra of the MSP1D1 variants upon excitation at 280 nm. UB-MSP1D1 displays a fluorescence intensity nearly 2-fold higher than MSP1D1 and almost 10-fold higher than D-MSP1D1 and, as expected, the peak emission of D-MSP1D1 appears at 302 nm from the tyrosine residues. UD-MSP1D1 emission drops to zero between 320 and 350 nm. Comparison of the emission spectra of MSP1D1 and UB-MSP1D1 (Figure 2C) shows a slight red shift in the emission normalized emission spectra, suggesting that a larger fraction of the Trp residues are solvent exposed. This is also evident in the change in the spectral center of gravity, shifting from 342.2 to 345.5 nm (Table 1). As expected, upon assembly of Nanodiscs the emission spectra of MSP1D1 and UB-MSP1D1 blue shifts, suggesting that the Trp residues are in an environment less accessible to solvent (Figure 2D). To further probe the tryptophan environments we measured the fluorescence lifetimes of the MSP1D1 an UB-MSP1D1 using cross correlation phase fluorometry (Figure 3) [34,38]. In both lipid free samples and assembled Nanodiscs, the fluorescence intensity decay is best described by two exponentials as shown in Table 2. Although the lifetimes of MSP1D1 and UB-MSP1D1 are quite similar, the fraction of the slow lifetime decreases slightly compared to MSP1D1 in both lipid free and DMPC Nanodiscs. This is not unexpected as the Trp fluorescence in proteins can vary wildly due to subtle changes in tryptophan conformation or environment resulting in the observed differences [40].

Figure 2.

UV-Vis and Normalized Emission Spectra. A. UV-VIS spectra of MSP1 variants: UB-MSP1D1 (solid red), MSP1D1 (dotted blue), D-MSP1D1 (dash-dot, yellow), UD-MSP1D1 (dashed blue), clipped UD-MSP1D1 (bold red) UDMSP spectra are magnified 3X. B. Emission spectra of MSP1D1 variants normalized to protein concentration: UB-MSP1D1 (red solid) MSP1D1 (blue dotted), D-MSP1D1 (green dashed ), UD-MSP1D1 (black bold). C. Intensity normalized emission spectra of MSP1D1 (blue dashed) and UB-MSP1D1 (red solid) showing a red shifted emission spectrum. D. Normalized emission spectra of Lipid Free MSP1D1 (blue dashed) and MSP1D1 DMPC Nanodiscs (red solid).

Figure 3.

Cross correlation phase fluorometry of MSP1D1 (red circles) Phase Delay, (blue squares) Modulation Ratio. Solid lines represent a fit of the luminescence decay.

Table 2.

Fluorescence properties of MSP1D1 variants

	Emission Spectral COG (nm)	F1	Tau1 (ns)	F2	Tau2 (ns)
MSP1D1 Lipid Free	342.6	0.75 ± 0.1	5.8 ± 0.6	0.25 ± 0.1	1.4 ± 1.0
MSP1D1 DMPC	339.4	0.74 ± 0.1	6.4 ± 0.6	0.26 ± 0.1	1.4 ± 0.8
UB MSP1D1 Lipid Free	345.5	0.62 ± 0.1	6.1 ± 0.9	0.38 ± 0.1	1.9 ± 0.6
UB MSP1DI DMPC	342.1	0.65 ± 0.08	6.2 ± 0.5	0.35 ± 0.08	1.5 ± 0.8

To characterize the mobility of the tryptophan residues the anisotropy decays of MSP1D and UB-MSP1D1 were measured in both lipid free form and DMPC Nanodiscs. Figure 4 shows the phase modulated anisotropy spectrum of MSP1D1. The line represents a fit using the measured fluorescence life times and two rotational correlation times (θ). Results for lipid free and DMPC Nanodiscs are summarized in Table 3. In all cases there is a fast anisotropy decay < 8 ns and a slower decay > 30 ns with contributions to the total anisotropy indicated by r_fast and r_slow. In a survey of the anisotropy decays of several tryptophan containing proteins, many displayed a complex decay model with a fast decay (< 1ns) that is attributed to the local motions of Trp [41]. In the case of Nanodiscs, we attribute the slow decay to the tumbling of the protein in solution. Consistent with this is the significant increase in the slow rotation when the MSP is incorporated into Nanodiscs. The diameters of the resulting Nanodiscs are approximately 10 nm which results in a rotational correlation time greater than 50 ns [42] which is difficult to resolve due to the short lifetime of the tryptophan fluorescence. The fast rotations on the other hand show a distinct difference in the lipid free vs. assembled Nanodiscs. The 4-fold slowing of the fast component can result from a decrease in the probability of homo-transfer between Trp residues. In the lipid free state, in the absence of detergent, the MSP most likely adopts a wide range of structures including oligomers, thus increasing the likelihood that Trp residues approach distances to promote depolarization through energy transfer, whereas in native MSP1D1 Nanodiscs the tryptophans become spatially spread out over the MSP belt, reducing the extent of homo-transfer. A second hypothesis, and the more likely scenario, is that the fast anisotropy decay is reporting on the segmented motions of the tryptophan monitored by the fluorescence. Inspection of a solution NMR structure of a shortened MSP missing helix-5 shows that, in the ten reported conformations of the scaffold protein, the tryptophan residues either lie in the protein-lipid interface or at the MSP-MSP protein-protein interface.[43] Although these NMR structures show a relatively planar lipid bilayer, Nanodiscs are clearly dynamic, particularly if less than ideal numbers of lipids are incorporated, and molecular dynamics simulations have revealed several complex small motions.[44–47] Therefore, when incorporated in the structure of the Nanodisc, the segmented motion slows significantly, whereas in the lipid free from the tryptophans are significantly more mobile.

Figure 4.

Anisotropy decay of MSDP1D1 monitored by Cross Correlation Phase Fluorometry. (red circles) Delta Phase, (blue squares) Amplitude Ratio

Table 3.

Anisotropy decay of MSP1D1 variants

	rslow	θslow (ns)	rfast	θfast (ns)
Lipid Free MSP1D1	0.14	44	0.08	1.4
MSP1D1 DMPC Nanodiscs	0.08	260	0.03	5.8
Lipid Free UB	0.09	105	0.03	1.9
UB DMPC Nanodiscs	0.10	131	0.03	7.6

3.2. Indolicidin Binding

To demonstrate the utility of UD-MSP1D1, we measured the binding of indolicidn to Nanodiscs containing 70%:30% DMPC:DMPS and 100% DMPC. Indolicidin is a small (13 amino acid) cationic antimicrobial peptide (AMP) isolated from bovine neutrophils. AMPs have been under intense study as a therapeutic against gram-positive bacteria due to the significant rise in antibiotic resistance of many pathogens, and more recently as a potential cancer therapeutic.[25,48,49] Cationic AMPs interactions are driven by negatively charged membrane surfaces owing to the increased prevalence of negatively charged lipids expressed on the surface of bacterial pathogens. Due to the small molecular mass of indolicidin, interaction of the peptide can be easily measured by monitoring the polarization of the tryptophan fluorescence. Figure 5 shows the tryptophan emission polarization as a function of UD-MSP1D1 containing 100% DMPC (blue) and 70% DMPS/30% DMPC (red). The solid line represents a simple Langmuir fit to a single binding site. As expected, the affinity is much tighter on Nanodiscs containing anionic lipids having a dissociation constant of 30 +/− 10 nM in the presence of 30% DMPC while pure DMPC displays a much weaker binding at ~ 650 nM.

Figure 5.

Indolicidn - Nanodisc binding isothemerms. Titration of 110 nM indolicidn with Nanodiscs. (red circles) 30% DMPS UD-MSD1D1 (blue squares) 100% DMPC UD-MSP1D1 Nanodiscs. Solid line represents a fit to a single binding site.

In many cases, proteins that interact with the membrane surface possess a rotational correlation time much too long to be monitored by tryptophan fluorescence polarization. The tryptophan lifetimes in proteins are short (~3 –5 ns) thus limiting the ability to measure changes in polarization of proteins ~ 50 kDa using the intrinsic fluorescence. In the absence of a detectable polarization signal, changes in the tryptophan environment (i.e. embedding into the lipid bilayer) can be detected in the Stokes shift of the Trp emission. In the case of indolicidin binding, the Trp environment is perturbed significantly, shifting the emission spectral center of gravity from 355 to 347 nm (Figure 6A). We use this shift to reproducibly measure the binding of indolicidin to 70% DMPC / 30% DMPS Nanodiscs (Figure 6B). The use of UD-MSP1D1 is thus extended to the label free detection of large proteins binding to a controlled bilayer surface with one caveat: the tryptophan emission of the large target protein must be sensitive to an environment change upon binding the Nanodisc. This approach is not limited to protein lipid interactions, but also protein-ligand or protein-lipid interactions that occur at the bilayers surface.

>Figure 6.

Emission spectra of indolicidn Nanodisc complex. Panel A: Emission spectra of indolicidn (blue dashed) and indolicidin bound to 30% DMPS Nanodiscs (red solid). Panel B: Indolcidn binding to Nanodiscs by monitoring the change in emission spctrum center of gravity. Solid line represents a fit to a single binding site, Kd=100 nM.

In an effort to facilitate label free detection methods we have generated UB-MSP1D1 UD-MSP1D1. The key advantage of the MSP based system is the ability to tune the UV absorption and fluorescence properties of the resulting Nanodisc preparation in order to leverage them as a detection modality. In the case of UB-MSP1D1 we have increased the UV extinction coefficient and the resulting fluorescence emission intensity to allow for easy detection in the absence of exogenous labels which is especially useful when a target protein contains few tyrosine or tryptophan residues. In particular techniques such as AUC and microscale thermophoresis (MST) can greatly benefit from the UB-MSP1D1 Nanodisc system. Unfortunately if one is interested in the fluorescence properties of the proteins interacting with the membrane bilayer, the presence of tryptophan and tyrosine in the scaffold proteins will limit the ability to monitor the UV absorption or fluorescence of the protein of interest. UD-MSP1D1 provides a platform to monitor target membrane proteins in a completely label free modality through the measurement of intrinsic fluorescence properties. The Stokes shift reports on the environment of the Trp residues while anisotropy reports on the tryptophan mobility. While useful, these techniques can only be employed if the system is poised to provide a significant change in tryptophan fluorescence properties. For example, anisotropy is only effective in studying a small target proteins (< 50 kDa) where a large increase in tryptophan anisotropy observed due to the slowed rotational diffusion upon interaction with a Nanodsics membrane. For experiments on large target proteins the Stokes shift can be used only if the tryptophan environment changes significantly upon interaction with the bilayer surface.

When compared to other discoidal membrane mimetics, UD-MSP1D1 has the advantage of a much lower UV extinction coefficients. Systems that utilize unlinked short peptides, when assembled at ratios that produce similar size discs, can contain up to 20 tryptophan residues [10][11], while styrene maleic acid polymers (SMAPs) based discoidal bilayers have a significant UV absorbance due to the styrene functionality [13]. The high UV extinction will interfere with fluorescence measurements either thorough increased background intensity or though the inner filter effect, thus decreasing the sensitivity of the assays. More recently a polymer membrane mimetic based on DIMBA has been developed with much lower UV absorbance which are also useful in measuring UV spectral properties of incorporated targets [15].

4. Conclusions

This communication reports the generation of MSPs harboring altered tryptophan emission properties. UB-MSP1D1 is a variant containing five Trp residues per MSP belt, thus giving the assembled Nanodisc a total of ten tryptophan residues. This variant is nearly twice as luminescent as MSP1D1 presenting researchers the opportunity to use Trp fluorescence as a detection modality for techniques such as analytical ultracentrifugation and microscale thermophoresis (MST).[50] The generation of a membrane scaffold protein lacking both tryptophan and tyrosine, UD-MSP1D1, provides an easily accessible label free detection modality for protein-lipid and protein-protein interactions in the controlled Nanodisc lipid environment. The utility of UD-MSP1D1 is dependent on the system being studied. For polarization measurements, the target protein must be < 50 kDa in order to resolve changes in rotational correlation upon interacting with the Nanodisc bilayer. A secondary mode of detection that includes larger proteins is monitoring the Stokes shift of the tryptophan emission if the system has a Trp residue that experiences a change in environment upon association with the membrane bilayer or when involved in protein-protein interactions. The addition of UD-MSP1D1 and UB-MSP1D1 increased the utility of the Nanodisc membrane mimetic to allow a purely label free detection of protein-membrane and protein-protein interactions that occur at the bilayer interface.

Highlights.

Nanodiscs, Fluorescence spectroscopy, Membrane Scaffold Proteins, Fluorescence lifetime, Fluorescence anisotropy, Protein-membrane interactions