Quantitative Analysis of Protein-Lipid Interactions Using Tryptophan Fluorescence

See allHide authors and affiliations

Science Signaling  01 Dec 2009:
Vol. 2, Issue 99, pp. pl4
DOI: 10.1126/scisignal.299pl4


The fluorescent properties of the amino acid tryptophan make it a useful tool for fluorometric assays. Because tryptophan fluorescence is remarkably sensitive to the polarity of the environment, it can be used to determine the affinity of tryptophan-containing peptides for phospholipid vesicles of varying compositions. Here, we describe a method for using tryptophan fluorescence to determine the binding affinities of peptides derived from the proteins Raf-1 and KSR-1 to small unilamellar vesicles containing phosphatidic acid. The method can be extrapolated to measure the binding of other tryptophan-containing peptides or proteins to lipid vesicles.


The fluorescent properties of the aromatic amino acids tryptophan, phenylalanine, and tyrosine were initially described 50 years ago (13). Thereafter, Teale demonstrated that the fluorescence quantum efficiency per aromatic residue (that is, the normalized intensity of the emitted fluorescence at a given excitation intensity) was not the same for all proteins (4). Furthermore, the excitation and emission properties of the fluorescent residues within the same protein show substantial variation (5). These observations are linked to the remarkable sensitivity of the fluorescent properties of aromatic amino acids to changes in the polarity of the local environment.

Because of its high quantum yield, tryptophan is the dominating intrinsic fluorophore in peptides and proteins. Tryptophan fluorescence is highly sensitive to the local environment: Changes in the intrinsic tryptophan fluorescence of proteins often occur upon conformational changes or binding of ligands. In general, the quantum yield of tryptophan fluorescence increases in intensity when tryptophan is exposed to a hydrophobic environment and decreases when it is exposed to an aqueous medium (5). A blue shift of the emitted light often accompanies quantum yield changes (6, 7). Tryptophan fluorescence has been used to determine the kinetics of protein-lipid and protein-protein interactions (8, 9), the depth of membrane insertion of proteins (1012), and the topology of peptide-membrane complexes (11, 13). The degree of blue shift has been correlated with the depth of tryptophan insertion into a membrane (10).

The importance of protein-lipid interactions in cell signaling is widely acknowledged. Many signaling proteins contain specific lipid-binding domains that govern their translocation to specific membrane surfaces. The affinities of lipids for signaling proteins and their specific lipid-interacting domains have been an area of active research. Most of these interactions have been measured with comparative semiquantitative procedures based on the use of immobilized lipid matrices (1416). Other procedures include plasmon resonance studies using lipid vesicles immobilized onto lipophilic chips (17) and sedimentation equilibrium data using lipid vesicles of known compositions (18).

The issue with most semiquantitative methods is that the information obtained is reliable only for comparative purposes. Thus, it is possible to estimate the relative affinities of various lipids to their protein-interaction domains with membrane-immobilized lipid substrates, but reliable quantitative estimates of the actual affinity binding constants cannot be obtained by this method. Plasmon resonance, on the other hand, affords quantification of the on and off rate constants of the formation of the protein lipid constants, but its use is hindered by technical limitations, such as unstirred layers and rebinding issues.

The protocol described here takes advantage of the properties of tryptophan fluorescence to determine the affinity of the phosphatidic acid (PA)–binding regions (PABRs) of the serine-threonine kinase Raf-1 and the closely related kinase suppressor of Ras 1 (KSR-1). The binding of Raf-1 and KSR-1 to PA is essential for the activation of the extracellular signal-regulated kinase (ERK) cascade (15, 19, 20). Deletion experiments with the Raf-1 model placed the PABR between residues 389 and 423 (residue numbering is based on mouse Raf-1) (15, 21). This 35–amino acid sequence is characterized by hydrophobic residues flanking a polybasic motif, which contains two arginine residues. Substitution of these arginines for alanine results in a mutant unable to interact with PA (15, 22). An equivalent mutation of the PABR of KSR-1 results in dominant-negative activity (19).

The assay described here was designed to obtain a quantitative description of the lipid-binding properties of the PABRs of Raf-1 and KSR-1. Neither Raf-1 nor KSR-1 contains a tryptophan in their PABR; however, they both contain a phenylalanine residue within the hydrophobic region downstream of the central polybasic motif. Therefore, we designed PABR-based peptides in which this endogenous phenylalanine was replaced with tryptophan. We also designed peptides in which both arginines of the polybasic core were replaced with alanines. The sequences of these peptides are shown in Fig. 1. Although this method was designed and tested primarily with synthetic peptides, it may be adapted to the study of proteins and protein-lipid binding domains. It should be noted that the analysis of the lipid-binding properties of proteins and peptides containing multiple tryptophan residues may be substantially more complex than the simple case presented here.

Fig. 1

Sequences of the Raf-1– and KSR-1–derived peptides used in the development of this protocol. WT denotes the wild-type sequence, and DM indicates double mutants in which the two arginines of the central polybasic motif were replaced by alanines (27). The W denotes the position where tryptophan was substituted.


Lipids, in chloroform solution (Avanti Polar Lipids)

Tryptophan-containing peptides, preferably above 95% purity by high-performance liquid chromatography (HPLC) (Molecular Medicine Institute Peptide Synthesis Facility of the University of Pittsburgh)

CHAPS (thin-layer chromatography grade)

Note: CHAPS is the detergent of choice for several reasons. The most important one is that CHAPS has no detectable fluorescence at the wavelengths of interest. In addition, very pure detergent is commercially available. Triton X-100 should be avoided, because of its very substantial fluorescence in the near-ultraviolet (UV) range.





Bath sonicator

Tip sonicator


Note: We use a Fluoromax-3 single-beam spectrofluorometer equipped with grating excitation and emission monochromators. However, double-beam spectrofluorometers are preferred, because these instruments allow for automatic baseline corrections.

Cuvettes, with matching stir bars

Stir plate

Note: This is only necessary if the spectrofluorometer is not equipped with a stirring mechanism.

Graphing and curve-fitting computer software (such as Microsoft Excel, SigmaPlot, or GraphPad Prism)


Recipe 1: Assay Buffer

Tris-HCl, pH 7.550 mM
NaCl100 mM
EGTA100 nM

Prepare 250 ml. Solution may be stored for several weeks at 4°C.

Recipe 2: Micelle Buffer

Add CHAPS to Assay Buffer (Recipe 1) to a final concentration of 18 mM. Prepare 40 ml. Solution may be stored for weeks at 4°C.


Preparation of Peptide Samples

We typically assay the peptides at final concentrations ranging from 15 to 150 nM. The optimal concentration depends on the number of tryptophans contained in the sequence and peptide solubility. Hydrophobic peptides tend to aggregate in aqueous solution, and this aggregation is usually accompanied by increased tryptophan fluorescence. Optimal concentrations may be determined empirically by measuring the fluorescence of the peptide solution without lipid and after the addition of 2 μM lipid and subtracting the appropriate baselines (see the detailed procedure described below). Each experiment requires 2 ml of peptide solution.

  • 1. Dissolve peptides in Assay Buffer (Recipe 1) to a final concentration between 0.6 and 6 μM to prepare stock solutions.

Note: If HPLC-purified synthetic peptides (purity >95%) are used, prepare solutions by weight. The concentration of purified recombinant peptides can be determined with a BCA protein concentration assay.

  • 2. After dissolving the peptides, sonicate the solution thoroughly with a bath sonicator.

  • 3. Dilute 10 to 50 μl of the stock peptide solution in 2 ml of Assay Buffer (Recipe 1) directly in the fluorometer cuvette. The final concentration of the peptide in the cuvette should be less than 150 nM.

Note: The peptide solutions must be clear—large clusters of oligomerized peptides will introduce noise into the data because of light scattering.

Preparation of Lipid Stock Solutions

Lipids are shipped dissolved in chloroform, which must be completely removed by drying the lipids under nitrogen before preparing lipid stock solutions. We typically prepare 1 ml of lipid stock solutions ranging from 2 to 5 mM. Each experiment requires <1 μl of primary lipid stock.

  • 1. Dry the lipid under nitrogen, creating a film on the bottom of the tube.

Note: Thorough drying is very important. We routinely leave the lipids under vacuum for at least 24 hours after the elimination of the chloroform.

  • 2. Resuspend the dried lipid in an appropriate volume of Micelle Buffer (Recipe 2) to make primary stock lipid solutions at concentrations between 2 and 5 mM.

  • 3. Prepare additional lipid solutions by dilution of the primary stock solutions in Micelle Buffer (Recipe 2). We recommend preparing 1-ml stocks of the following dilutions: 1 μM, 10 μM, 100 μM, and 1 mM.

Note: This should result in a clear suspension of lipid micelles.

Preparation of Lipid Vesicles

To prepare small unilamellar vesicles, it is important to dilute the detergent (CHAPS) below the critical micelle concentration (CMC), which, for CHAPS, is 6.2 mM in 100 mM NaCl (23). The concentration of CHAPS in the diluted samples must be at least 1/10th that of the CHAPS CMC in order to produce vesicles instead of micelles (24).

  1. Dilute the concentrated lipid stocks into Assay Buffer (Recipe 1) by a factor of at least 100 to yield a final CHAPS concentration less than 200 μM and lipids at concentrations between 10 nM and 2 μM.

  2. With a microtip sonicator, sonicate the diluted lipid solution thoroughly, for 1 min at 80 to 90% power in an ice bath (to prevent overheating).

Note: The lipid vesicle solution should be clear, consisting primarily of small unilamellar vesicles, not large multilamellar vesicles.

Generation of Emission Spectra

We collect data at room temperature with a Fluoromax-3 single-beam spectrofluorometer equipped with grating excitation and emission monochromators and with the instrument configured to excite the samples at 280 nm. Tryptophan spectra should be collected in the 300- to 400-nm range. To minimize errors, we recommend configuring the instrument such that each spectra data point is integrated over several seconds of measurement. A sample data trace obtained with a single-beam instrument is shown in Fig. 2.

Fig. 2

Fluorescence emission spectra of the WT-KSR1 PABR peptide. Data were collected using 280 nm as the excitation wavelength. The figure compares the fluorescence of the peptide in solution with the fluorescence of the peptide in the presence of lipid vesicles containing dioleoyl phosphatidic acid (DOPA).

When a single-beam spectrofluorometer is used, two control spectra must be collected. The first one is obtained from the peptide solution dissolved in Assay Buffer. The second set of control spectra is obtained from lipid vesicles at each lipid concentration used to obtain the titration curve. This second set of control spectra is not necessary when using a dual-beam spectrofluorometer in which the reference cuvette contains lipids at exactly the same concentration as the sample cuvette.

Because measurements are initiated at the time of addition of the peptide solution to the sample cuvette followed by the addition of the lipid vesicles, it is convenient to use a fluorometer equipped with a built-in stirring mechanism. If the instrument does not have this feature, the solution must be properly stirred by other methods, such as carefully pipetting the solution in and out of the cuvette several times. Care must be taken to avoid trapping air bubbles within the cuvette.

  • 1. Agitate 2 ml of peptide solution for 30 s either in the spectrofluorometer (if equipped with a stirring mechanism) or by pipetting the sample repeatedly. Avoid trapping air bubbles in the cuvette, because these will interfere with the assay.

  • 2. Collect the baseline emission spectrum for the peptide alone by exciting the sample at 280 nm and collecting the emission at wavelengths from 300 to 400 nm.

  • 3. With vesicles prepared with the lowest concentration of lipid, dilute the vesicles directly into the peptide-containing sample cuvette (and into the reference cuvette if using a dual-beam instrument). Stir continuously, or mix thoroughly by pipetting the solution repeatedly.

  • 4. Collect the spectrum by exciting the sample at 280 nm and collecting the emission at wavelengths from 300 to 400 nm following the instructions for the spectrofluorometer.

  • 5. For each successively higher lipid concentration, continue adding the lipid vesicle solution to the peptide sample, thus conserving peptide. Stir the sample after each addition.

Note: It is very important that the CHAPS concentration in the sample cuvette remains no less than1/10th that of the CMC. Table 1 shows an example of a titration experiment using in succession four lipid stock solutions of different concentrations.

  • 6. Collect an emission spectrum from the vesicle solutions at each concentration of lipids to which the peptides were exposed.

Note: This step is unnecessary if a dual-beam instrument is used and the lipid vesicles are added to reference cuvette for each concentration tested.

Table 1

Design of a typical titration experiment. This experiment is designed to collect fluorescence data within a lipid concentration range of 10 nM to 1.33 μM. The experiment is initiated by diluting the peptide stock solution and measuring its spectrum. Then, for each point, the indicated volume of lipid stock is added, and the spectrum is collected after careful mixing. Each lipid addition changes the volume of the sample by less than 0.05% (0.14% for the complete titration curve). The final CHAPS concentration in this example is 144 μM, about 1/50th that of the CMC of CHAPS.

View this table:

Titration Calculations

When choosing a wavelength for calculation purposes, it is important to take into account the blue shift of the tryptophan emission. Additionally, the wavelength selected should maximize the signal-to-noise ratio. For the Raf-1 PABR peptides, we calculated the tryptophan fluorescence at 325 nm (Fig. 3) because at 325 nm there was a large difference between the fluorescence intensity in the peptide alone compared with that in the presence of the lipid vesicles (Fig. 2).

Fig. 3

Typical titration curves. Data were collected using the following wavelengths: excitation, 280 nm; emission, 325 nm. Data were fitted to a sigmoidal binding curve with variable Hill coefficient using GraphPad Prism. The apparent Hill coefficient obtained from the fit was 1.15, indicating binding of a single phosphatidic acid molecule to the peptide.

Curve-fitting software, such as GraphPad Prism, is made to assume that the values entered as ligand concentrations correspond to the free ligand. This assumption is approximately correct at high lipid concentrations, but does not apply to the lowest concentration points, especially when the lipid binds the peptide with high affinity; therefore, instructions for calculating the free lipid concentration at each point are provided as steps 2 to 5 below. The calculations described can be easily programmed in an Excel worksheet.

  1. With a single emission wavelength chosen to maximize the signal-to-noise ratio and to minimize the effect of the blue shift, determine the fluorescence of the peptide exposed to all lipid concentrations from the emission traces.

  2. Using fixed excitation and emission bands (excitation, 280 nm; emission, 325 nm), for each concentration of peptide analyzed, determine the maximal change in fluorescence (ΔFmax) in the presence of saturating concentrations of lipid (1 μM or higher).

  3. Calculate the fractional binding (f) for each concentration of lipid by determining the ratio ΔF/ΔFmax (where ΔF is the change in fluorescence observed at a given lipid concentration).

  4. Calculate the concentration of bound lipid, which is equal to the product of the fractional binding times the total peptide concentration.

  5. Calculate the free lipid concentration by subtracting bound lipid from the total lipid concentration.

  6. With the ΔF, the free lipid concentration values, and curve-fitting software, calculate the dissociation equilibrium constants (KD).

Note: We routinely use GraphPad Prism, but any statistical package that provides curve-fitting analysis using least-squares or maximum-likelihood procedures is appropriate.


Light Scattering

Very hydrophobic peptides tend to aggregate, which introduces errors due to light-scattering caused by the aggregates. Likewise, small unilamellar vesicles tend to fuse and produce multilamellar vesicles, which also cause light scattering because of their larger size. These problems can usually be avoided by sonicating the lipid and peptide solutions immediately before the experiment.

No Detectable Change in Tryptophan Fluorescence

It is always possible that the binding of lipid to the protein or peptide does not alter significantly the fluorescence of the tryptophan residues, resulting in very small or undetectable changes in the fluorescence. In this case, mutation to tryptophan of another residue that is closer or farther away from the putative lipid-binding pocket may produce detectable changes in fluorescence upon lipid binding. Determining the best residue to mutate may need to be determined empirically and, thus, may become time-consuming and expensive.

Binding is a bimolecular reaction. Therefore, the rate of binding is a function of the concentration of the components. The fluorescent signal may take some time to develop when using very low lipid concentrations. It is advisable to determine appropriate incubation times by measuring the time course of peptide-lipid binding at low lipid concentrations by running a binding experiment using the lowest or second lowest lipid concentration as described above and recording the spectra at increasing intervals (1, 5, and 10 min). If the fluorescence does not change with time, the system can be assumed to be in equilibrium at the earliest time point.

Notes and Remarks

Using changes in tryptophan fluorescence to monitor protein-lipid interactions does have some limitations. The procedures described work well for peptides or proteins in which there is a single tryptophan. Proteins that contain more than one tryptophan residue will exhibit more complex spectra, and the fluorescence changes induced by lipid binding may be very different from those for proteins that contain a single tryptophan. However, as long as there is a change in the fluorescence that can be measured, this method can be adapted to this more complex situation.

Lipid vesicles introduce substantial light scattering under the best of circumstances. Thus, it is difficult to adapt the technique when high lipid concentrations (that is, 10 μM or greater) are required.


Funding was from NIH R01 DK54782 (to G.R.) and T32-GM-54812.

References and Notes

View Abstract

Stay Connected to Science Signaling

Navigate This Article