Single-molecule FRET reveals sugar-induced conformational dynamics in LacY

Majumdar et al. 10.1073/pnas.0700969104.

Supporting Information

Files in this Data Supplement:

SI Materials and Methods
SI Figure 6
SI Figure 7




SI Figure 6

Fig. 6. (A) E* distribution of C154G LacY double-labeled on the cytoplasmic side, no NPG added (gray bins) compared to E* distribution derived from dsDNA labeled with the same fluorophores (blue line). Two-state (green lines and tabs) and Gaussian distribution (red lines and tabs) model fitting of E* distribution (gray bins) of cytoplasmically labeled C154G LacY mutant in the absence (B) and presence (C) of NPG. The same modeling was applied to the E* distribution of wild-type LacY double-labeled on the periplasmic side in the absence (D) and presence (E) of NPG. Vertical tabs correspond to the mean value of indicated states and horizontal red bars indicate the E* values corresponding to DR.

SI Figure 7

Fig. 7. E* histogram in the absence of NPG (gray bins) modeled with a two interconverting states E* = 0.383 and E* = 0.639, with interconversion rates of 3 s^-1 (magenta), 150 s^-1 (green), and 3,000 s^-1 (cyan).

SI Materials and Methods

Materials. All galactoside derivatives were purchased from Sigma (St. Louis, MO). Oligonucleotides were synthesized by Integrated DNA Technologies (Coralville, IA). Restriction enzymes were purchased from New England Biolabs (Beverly, MA). The QuikChange II kit was from Stratagene (La Jolla, CA). Talon superflow resin was obtained from BD Clontech (Palo Alto, CA). All other materials were of reagent grade obtained from commercial sources.

Construction of Mutants. LacY mutants were constructed by site-specific mutagenesis using the QuikChange II site-directed mutagenesis kit and plasmid pT7-5/LacY as a template. Generally, 30- to 40-bp direct and reverse primers bearing the mutated triplet in the middle of the primer were designed using the Vector NTI 9.1 program (Invitrogen, Carlsbad, CA). All mutagenesis procedures were done according to the QuikChange II manual except that the temperature of the extension reaction was lowered to 68°C during PCR-directed mutagenesis to reduce possible primer duplication. Mutagenic constructs were sequenced over the entire gene to confirm mutations introduced and to discard unwanted mutations. All constructs were engineered with a C-terminal His-tag to enable purification by affinity chromatography.

LacY Purification. LacY mutants were purified from E. coli XL1-Blue cells (Stratagene) by using Co(II) affinity chromatography as described (1). All protein preparations were at least 95% pure as judged by silver staining after SDS polyacrylamide gel electrophoresis.

Ensemble Fluorescence Measurements. Steady-state fluorescence was measured at room temperature in an SLM-Aminco 8100 spectrofluorometer (Urbana, IL) modified by OLIS (Bogart, GA). Windows-based Olis GlobalWorks software was used for instrument control. All measurements were performed in degassed 50 mM NaP_i (pH 7.5)/0.02% DDM in 1 ´ 1 cm cuvettes (2 ml) with constant stirring. Emission spectra were recorded with slits at 4 and 8 mm for excitation and emission, respectively.

Molecular Modeling. Structures of Alexa Fluor 488 C₅-maleimide and Alexa Fluor 647 C₂-maleimide were taken from the Invitrogen catalog and from ref. 2, respectively. Products of the reaction of both Alexa dyes with L-cysteine were constructed in the ChemDraw Ultra 9.0 program, transferred to the Chem3D Ultra program (ChemOffice Ultra 9.0 Suite; CambridgeSoft, Cambridge, MA), and minimized using the Mopac module in the same program. The resulting PDB files were transferred to PyMOL 0.99 (DeLano Scientific), and atoms originated from cysteine (Ca, Cb, N, and O) of Cys-Alexa 488 or 647 were aligned with corresponding atoms of the amino acid at a given position in the LacY structure (PDB ID code 1PV7) where an additional Cys residue was introduced by site-directed mutagenesis.

Single-Molecule Fluorescence Data Analysis. The number of photons in a burst detected from donor (D) and acceptor (A) fluorophores, as correlated with D and A excitation periods, reports on the apparent FRET efficiency (E*) and stoichiometry of dyes in a molecule (S). E* is calculated by dividing the number of acceptor photons by the sum of D and A photons (detected during the D excitation periods) as shown by (3)

 (1),

where

is the background-corrected D-excitation-based D-emission, and is the D-excitation-based A-emission, corrected for background and direct excitation of A. The Ratio S is defined as

 (2),

where

is the sum of all D-excitation-based emission and is the sum of all A-excitation based emission. The inherent heterogeneity of statistical labeling of LacY is addressed by this ratio: D-only molecules (S »1) and A-only molecules (S » 0) are effectively "sorted" from molecules labeled with both D and A (S » 0.5). Thus, each burst, as defined by the burst search and characterized in terms of E* and S, is binned in a 2D S-E* histogram with 100 bins on each axis. Relevant FRET events were selected from S-E* histograms using the dual-channel burst search (4). The resulting subpopulation of events was projected onto a 1D E* histogram for further analysis (presented with 50 bins and smoothed with a running average of 3 bins window size). Gaussian fits to the 1D E* histogram were used to estimate E* and DE*. Each experiment was repeated 3-5 times, and the results are reported as <E*> and <DE*> of these data sets.

The apparent FRET efficiency, E*, is a FRET-dependent, distance-dependent ratio, related to the FRET efficiency by the following relationship:

 (3),

where g is a correction factor dependent on quantum efficiency and detection efficiency. Measurement of g was performed as described previously (3), using two different D-A-labeled double-stranded DNA molecules (same fluorophores as those used to label LacY). Because g might be sensitive to local environment of D and A (through changes in quantum yields), our measured g for DNA is not necessarily the same as the true g for the protein. The calculated "accurate" E values are therefore not fully corrected.

Distance (R) is derived from

 (4)

assuming k² = 2/3.

is the measured Förster radius (5). The cited distances are therefore not fully corrected and carry errors due to the assumptions regarding g and k². Nonetheless, calculated distances agree well with modeling.

Wild-type and C154G LacY were mutated bearing a single additional Cys replacement at each labeling position (R73, 401, 164, and 375). Each single-Cys mutant was individually labeled with D and A for quantum yield measurements of D-labeled molecules and absorption spectra of A-labeled molecules. The spectral overlap of D and A were used to calculate R₀ in each construct reported herein, and we assume an average value of 2/3 for k², the orientation factor (6). Consistency and reproducibility of R₀ values is important for validation of statistical labeling. For example, R₀ values are 50.7 Å and 51.7 Å (donor at 73 and 401, respectively) for C154G LacY and 52.7 Å and 48.2 Å (same positions) for wild-type LacY doubly-labeled on the cytoplasmic side. R₀ values calculated for wild-type labeled on the periplasmic side are 48.4 Å and 53.4 Å (donor at 164 and 375, respectively) and 47.1 Å and 51.8 Å (same positions) for C154G. Calculations of interdye distances for doubly labeled LacY are based on the average of each corresponding R₀ pair.

The fractional change in the NPG titrations was measured using the following expression:

 (5),

where C_X measures the fractional change at X [mM] NPG, and where N₀, N₁₀₀₀, and N_x refer to the measured value of E* or DE* at concentrations of 0, 1,000, or X (mM).

Analysis of Distribution of Conformers. None of the E* histograms obtained for wild-type and C154G mutant LacY could be explained by a single conformer. As an example, Fig. 6A compares the E* histogram measured for the C154G mutant on the cytoplasmic side with no sugar (gray bins) with a histogram derived from double-stranded DNA (a relatively rigid molecule representing a single conformer) labeled with the same fluorophores that are separated by a similar distance (blue line). Evidently, the C154G mutant distribution is wider. The broader distributions could be explained either by (i) dynamic interconversion between two (or more) states or by (ii) a static (or slowly interconverting) distribution of closely-spaced states (that could not be resolved due to shot-noise broadening). In the following, we examine the data presented in Figs. 2C and 4B with respect to these two very simplistic models (the data in Figs. 2B and 4C do not exhibit much of a change). Although simplistic, these models do provide some limits and bounds on distribution of conformers and interconversion rates.

Cytoplasmic side (C154G mutant). In the absence of NPG, fitting the histogram with a two-state dynamic model (i above) yields

= 0.383 and = 0.639 (Fig. 6B, green line and two green tabs; distances of 57.5 Å and 48.3 Å, respectively; DR = 9.2 Å) with an apparent equilibrium rate constant K_eq = 1. In this model, the underlying interconversion rate must be slower than diffusion (<2,000 s^-1); fitting yields a rate k_12/21 < 150 s^-1 (see also Fig. 7). For NPG-bound LacY mutant, fitting with this model yields = 0.420 and = 0.670 (Fig. 6C, green line and two green tabs; distances of 56.1 Å and 47.2 Å, respectively; DR = 8.9 Å), with an apparent equilibrium rate constant of K_eq = 2.55, k₁₂ < 280 s^-1, and k₂₁ < 110 s^-1 (indeed well below the diffusion rate). According to this model, the reduction in DR due to sugar binding is DDR = 0.3 Å. The same data could be fitted with a multiple-state static model (ii above), i.e., a Gaussian distribution of static conformers with a mean <R> and a root-mean-square width DR, which is a measure of the static conformational heterogeneity. In absence of NPG, fitting the histogram with this model yields R = 53.0 Å and DR = 5.1 Å (Fig. 6B, red line and tab). With NPG, the fitting yields R = 49.7 Å and DR = 3.9 Å (Fig. 6C, red line and tabs). According to this model, the reduction in DR due to sugar binding is DDR = 1.2 Å.

Periplasmic side (wild type). In the absence of NPG, fitting the histogram with a two-state dynamic model yields

= 0.337 and = 0.506 (Fig. 6D, green line and two green tabs; distances of 60.7 Å and 53.2 Å, respectively; DR = 7.5 Å) with an apparent equilibrium rate constant K_eq = 1. In this model the underlying interconversion rate must be slower than diffusion; fitting yields a rate k_12/21 < 10 s^-1. For NPG-bound wild-type LacY, fitting with this model yields = 0.270 and = 0.440 (Fig. 6E, green line and two green tabs; distances of 64.9 Å and 55.8 Å, respectively; DR = 9.1 Å), with an apparent equilibrium rate constant of K_eq = 1, and rate k_12/21 < 10 s^-1 (below the diffusion rate). The increase in DR due to sugar binding according to this model is DDR = 1.6 Å. Fitted with the Gaussian distribution model, the same data in the absence of NPG yields R = 56.8 Å and DR = 3.4 Å (Fig. 6D, red line and red tab). With NPG, the fitting yields R = 59.9 Å and DR = 3.7 Å (Fig. 6E, red line and red tabs). According to this model, the increase in DR due to sugar binding is DDR = 0.3 Å.

1. Ermolova NV, Smirnova IN, Kasho VN, Kaback HR (2005) Biochemistry 44:7669-7677.

2. White SS, Li H, Marsh RJ, Piper JD, Leonczek ND, Nicolaou N, Bain AJ, Ying L, Klenerman D (2006) J Am Chem Soc 128:11423-11432.

3. Lee NK, Kapanidis AN, Wang Y, Michalet X, Mukhopadhyay J, Ebright RH, Weiss S (2005) Biophys J 88:2939-2953.

4. Nir E, Michalet X, Hamadani KM, Laurence TA, Neuhauser D, Kovchegov Y, Weiss S (2006) J Phys Chem B 110:22103-22124.

5. Kapanidis AN, Lee NK, Laurence TA, Doose S, Margeat E, Weiss S (2004) Proc Natl Acad Sci USA 101:8936-8941.

6. Ha T, Enderle T, Ogletree DF, Chemla DS, Selvin PR, Weiss S (1996) Proc Natl Acad Sci USA 93:6264-6268.

• Early Edition
• Archives
• Online Submission

• Feature Articles

  (Expanded research articles of exceptional breadth)

• Commentaries

  (Expert commentary on leading research)

• Letters

  (Online-only letters to the editor)

• Inaugural Articles

  (Articles by recently elected NAS members)

• PNAS Profiles

  (Biographical profiles of Academy members)

• Sustainability Science

  (Interactions of human and environmental systems)

• Collected Papers

  (Editorials, Perspectives, Reviews, Colloquia, etc.)

• Special Features

  (Solicited articles on exceptional research topics)

• Sackler Colloquia

  (Scientific reports held under NAS auspices)