|Home | About | Journals | Submit | Contact Us | Français|
Current distance measurements between spin-labels on multimeric protonated proteins using double electron-electron resonance (DEER) EPR spectroscopy are generally limited to the 15–60 Å range. Here we show how DEER experiments can be extended to dipolar evolution times of ~80 µs, permitting distances up to 170 Å to be accessed in multimeric proteins. The method relies on sparse spin-labeling, supplemented by deuteration of protein and solvent, to minimize the deleterious impact of multispin effects and substantially increase the apparent spin-label phase memory relaxation time, complemented by high sensitivity afforded by measurements at Q-band. We demonstrate the approach using the tetradecameric molecular machine GroEL as an example. Two engineered surface-exposed mutants, R268C and E315C, are used to measure pairwise distance distributions with mean values ranging from 20 to 100 Å and from 30 to 160 Å, respectively, both within and between the two heptameric rings of GroEL. The measured distance distributions are consistent with the known crystal structure of apo GroEL. The methodology presented here should significantly expand the use of DEER for the structural characterization of conformational changes in higher order oligomers.
Sparse nitroxide spin-labeling of symmetric, fully deuterated multimeric proteins, extends DEER distance measurements between nitroxide spin labels up to 170 Å. This is achieved by increasing the length and amplitude of the slow component of phase memory relaxation, thereby permitting DEER data to be collected out to long dipolar evolution times (~80 µs). The approach is demonstrated on the molecular machine GroEL comprising 14 identical subunits arranged in two heptameric rings.
Double electron-electron resonance (DEER) EPR spectroscopy offers a powerful method of probing conformational changes in biological macromolecules by measuring quantitative distances between pairs of spin labels. Typically, distances within the range 15–60 Å for protonated proteins and up to 90 Å for deuterated proteins are accessible by DEER.[2c,3] Symmetric multimeric proteins, however, present a particular challenge for the quantitative interpretation of DEER data, owing to the negative impact of multispin effects which arise when three or more spin labels are located in close proximity.[2c,4] The phase memory relaxation time, Tm, which dictates the length of the dipolar coupling evolution time in a DEER experiment and hence both the accessible distance range and signal-to-noise, decreases as the number of spins within a particular molecular assembly increases.[4d,5] Further, the presence of more than two spins results in so-called “ghost peaks” in the DEER-derived distance distributions owing to higher-order sum and difference dipolar frequency contributions to the DEER echo curve. In this paper we present a simple approach employing sparse spin-labeling that can circumvent the deleterious impact of multispin effects. We demonstrate distance measurements of 160 Å using the molecular machine GroEL comprising 14 identical subunits arranged in two stacked heptameric rings (Fig. 1).
In these studies, two engineered, surface-exposed cysteine mutants of GroEL were employed: R268C and E315C. For any given spin label, there are 6 intra-ring and 7 inter-ring spin pairs (Figs. 1 and and2A).2A). The distances between spin-labels, calculated from the crystal structure (PDB 1XCK) using the program MMMv2013.2, range from 15 to 80 Å within a heptameric ring, and from 90 to 170 Å between rings. The latter encompass a broad range of distributions centred about 100 Å and 160 Å for the R268C and E315C constructs of GroEL, respectively. The room temperature X-band CW spectrum for nitroxide spin-labeled GroEL(E315C) is characterized by relatively narrow linewidths indicative of mobile spin labels (SI Fig. S1A); that for GroEL(R268C), however, appears broad (SI Fig. S1A) despite the fact that a wide range of rotamers is predicted for both spin-labels by MMMv2013.2 (Fig. 1A). This is due to the fact that for the R268C construct (but not the E315C construct) there are pairs of spin labels within a heptameric ring separated by less than 15 Å. These short distances result in strong dipolar interactions that broaden the EPR spectrum, as confirmed by the Q-band echo-detected EPR spectrum (SI Fig. S1B).
In the four-pulse DEER experiment,[1d] the reliability of the P(r) distance distribution is governed by the maximum dipolar evolution time, tmax.[2c] For tmax = 2 µs, reliable distances can be obtained up to ~50 Å, and this limit scales as tmax1/3.[2c] Thus, tmax values of ~20 and ~80 µs would be required to accurately determine distances of ~100 and ~170 Å, respectively. In the literature tmax is usually only extended out to ~1.5Tm to obtain adequate sensitivity.[3a] As a result, such long tmax values are generally precluded since Tm values are typically of a few microseconds or less in protonated protein samples. Deuteration of both protein and solvent can significantly extend the Tm by removing electron-nuclear dipole interactions between spin labels and nearby protons.[3b, 12] This is illustrated by the spin-echo decay curve for a model system comprising doubly spin-labeled, deuterated, monomeric protein A shown in SI Fig. S2 (Tm ~ 64 µs) which clearly shows that DEER data can easily be acquired to tmax = 80 µs in a simple two spin system. For a multispin system, however, deuteration alone is entirely insufficient, as shown by the spin-echo decay curve for deuterated GroEL(E315C) in which every subunit is spin-labeled (i.e. a total of 14 spin labels): a bi-exponential decay is observed with fast and slow ) components arising from spin-spin relaxation between dipole-coupled spin labels and various forms of spin diffusion (black trace in Fig. 2A). Further, the amplitude of the fast component is very high (~70%; cf. Table 1) precluding the use of tmax values much larger than about 2.5 to 3 µs in a DEER experiment, corresponding to an upper distance of ~60 Å, in accordance with previous observations on protein assemblies.[4a–c]
The simple solution to the above problem is to make use of sparse spin-labeling by diamagnetic dilution of the spin-label reagent with its diamagnetic analog (Fig. 1C), in our case MTSL (1-oxyl-2,2,5,5-tetramethyl-D3-pyrroline-3-methyl methanethio-sulfonate) and MTS (1-acetoxy-2,2,5,5-tetramethyl-d-3-pyrroline-3-methyl methanethiosulfonate), respectively. Alternatively one can reduce the ratio of spin-label reagent to free cysteines (SI Fig. S2). In both instances, as the fractional MTSL labeling is reduced, the decay rate and amplitude of the slow component of the phase memory relaxation are increased, while the amplitude of the fast component is concomitantly reduced even though its decay rate is unaffected (Fig. 2A and Table 1, and SI Fig. S2 and Table S1).
To observe dipolar coupling at least 2 sites (out of 14 in GroEL) must be labelled, but labeling of more than 2 sites decreases the Tm. Thus the ratio of MTSL to MTS needs to be optimized to minimize the fraction of GroEL labeled at only a single site or at more than 2 sites. A 1:7 ratio of MTSL:MTS would result in the largest fraction of GroEL molecules occupied by two spin labels but would also result in a significant fraction of GroEL molecules with only a single spin label which would not contribute to the DEER dipolar evolution curve. Thus, we decided to use a ratio of MTSL to MTS (~1:5 in this instance) that results in the largest fraction of molecules with 2 to 3 spin labels (Fig. 2B), reducing the one-spin label species while still retaining a large amplitude for (Fig. 2A and Table 1), and a modulation depth in excess of 50% (SI Fig. S3). The presence of some species containing three spin labels will be associated with “ghost” peak contributions to the P(r) distance distributions which can be minimized by reducing the normalized modulation depth (Δ/Δ180) to a value of ~0.6 by decreasing the ELDOR pump pulse flip angle from the usual 180° to ~60–70°.[6b]
Q-band DEER echo curves recorded on fully deuterated GroEL(R268C) and GroEL(E315C), spin-labeled using a ratio of 1:5 MTSL to MTS, with different tmax values (20 and 35 µs for the R268C sample, and 47 and 80 µs for the E315C sample) are shown in Fig. 3A (left and right panels, respectively). For comparison a DEER echo curve recorded with a tmax of 18 µs is included for fully spin-labeled and deuterated GroEL(E315C), highlighting the resulting very rapid signal decay and the huge gains afforded by sparse spin-labeling. P(r) distributions derived from the DEER echo curves using the programs DD and DeerAnalysis are shown in Fig. 3B. The former makes use of a mathematical model (in this instance a sum of Gaussians) to directly fit the DEER data (including automated background subtraction with a best-fit exponential decay), while the latter is model free and uses Tikhonov regularization. The predicted P(r) distributions derived from the apo GroEL crystal structure using MMMv2013.2 are shown by the grey envelopes. For GroEL(R268C) the predicted intra-ring distances between spin labels fall into two classes comprising two distances around 15 Å, and four distances that coalescence into a broad distribution centered at 37 Å; the seven predicted intra-ring distances merge into a single distribution centred around 100 Å. For GroEL(E315C), the six intra-ring distances fall into three resolved distributions centred at 35, 60 and 78 Å, while the inter-ring distances which range from 146–170 Å, fall into two classes centred at 152 and 162 Å. Overall, for both GroEL(R268C) and GroEL(E315C), the P(r) distributions derived from the DEER data recorded with the longer tmax (red curves; 35 µs for R268C and 80 µs for E315C), using either DD or DeerAnalysis (Fig. 3B, top and bottom panels, respectively), match closely with the predicted distributions from the crystal structure, reflecting the increased accuracy afforded by extending the measurement to longer dipolar evolution times.
A summary of the intra- and inter-ring mean distances between spin labels derived from the long tmax DEER data using DD and DeerAnalysis is provided in Table 2. The long inter-ring mean distances, ~103 and ~160 Å for GroEL(R268C) and GroEL(E315C), respectively, obtained by the two methods of analysis are in excellent agreement with one another. The same is approximately true for the three shorter intra-ring mean distances obtained for GroEL(E315C). The intra-ring distances for GroEL(R268C) are well reproduced from the DD analysis, but appear to be shifted to slightly higher values with DeerAnalysis. This may be due the fact that DD describes the DEER echo curve for GroEL(R268C) as a function of three Gaussians, while no assumption on the number of distances is made by the model free DeerAnalysis approach. The unconstrained nature of the latter, particularly when the data at short dipolar evolution times are undersampled (since we were basically interesting in measuring the longer distances beyond 100 Å), may result in potential overfitting of the data, accounting for both the shifts in peak positions and possibly the appearance of a longer distance at around 50 Å.
The total integrated intensities for the long inter- and short intra-ring P(r) distance distributions should in principle be roughly comparable given that there are 6 intra- and 7 inter-ring distances. This is approximately true for GroEL(R268C), but not for GroEL(E315C) where the integrated intensity of the long inter-ring distances centered around 160 Å is clearly underestimated (Fig. 3B). This is due to the fact that the value of tmax (in µs) required to measure an accurate mean distance (rm in Å) is given by 2(rm/50)3 µs, but the value required to accurately determine both mean distance and distribution width (and hence integrated intensity) is given by 2(rm/40)3 µs.[2c] The latter condition is satisfied with tmax = 35 µs for the 103 Å mean distance in GroEL(R268C), but not with tmax = 80 µs for the 160 Å mean distance in GroEL(E315C) where a tmax value of ≥130 µs would be required.
In conclusion, distances between nitroxide spin labels in the 100 to 170 Å range are accessible to DEER in symmetric multimeric proteins using sparse spin-labeling to increase both the length (~1.5-fold) and amplitude (3 to 3.5 fold) of the slow component of phase memory relaxation, over and above what is already achievable by full deuteration of protein and solvent. As a result DEER echo curves can be acquired for long dipolar evolution times up to a tmax of ~ 80 µs which corresponds to an upper mean distance limit of ~170 Å.[2c] This approach, which is easy to implement by simply controlling the ratio of nitroxide spin-label to diamagnetic analog, should find wide applicability for the study of conformational changes involving many multimeric systems, including an array of molecular machines, ion transporters and transmembrane assemblies.
This work was supported by funds from the Intramural Research Program of NIDDK, NIH, and from the Intramural AIDS Targeted Antiviral Program of the Office oft he Director of the NIH (to G.M.C.).
Supporting information for this article is given via a link at the end of the document.
Thomas Schmidt, Laboratory of Chemical Physics, National Institutes of Diabetes and Digestive and Kidney Diseases, National Institutes of Health Bethesda, MD 20892-0520 (U.S.A.)
Marielle A. Wälti, Laboratory of Chemical Physics, National Institutes of Diabetes and Digestive and Kidney Diseases, National Institutes of Health Bethesda, MD 20892-0520 (U.S.A.)
James L. Baber, Laboratory of Chemical Physics, National Institutes of Diabetes and Digestive and Kidney Diseases, National Institutes of Health Bethesda, MD 20892-0520 (U.S.A.)
Eric J. Hustedt, Department of Molecular Physiology and Biophysics, Vanderbilt University, Nashville, TN 37232 (U.S.A.)
G. Marius Clore, Laboratory of Chemical Physics, National Institutes of Diabetes and Digestive and Kidney Diseases, National Institutes of Health Bethesda, MD 20892-0520 (U.S.A.)