Conformational Analysis of Uniformly 13C-Labeled Peptides by Rotationally Selected 13Cα-13CH3 Double-Quantum Solid-State NMR

Abstract

Peptides are an important class of biomolecules that perform many physiological functions and which occupy a significant and increasing share of the pharmaceutical market. Methods to determine the solid-state structures of peptides in different environments are important to help understand their biological functions and to aid the development of drug formulations. Here, a new magic-angle spinning (MAS) solid-state nuclear magnetic resonance (SSNMR) approach is described for the structural analysis of uniformly ¹³C-labeled solid peptides. Double-quantum (DQ) coherence between selective pairs of ¹³C nuclei in peptide backbone and side-chain CH₃ groups is excited to provide restraints on (i) ¹³C-¹³C internuclear distances and (ii) the relative orientations of C-H bonds. DQ coherence is selected by adjusting the MAS frequency to the difference in the resonance frequencies of selected nuclear pairs (the rotational resonance condition), which reintroduces the dipolar coupling between the nuclei. Interatomic distances are then measured using a constant time SSNMR experiment to eliminate uncertainties arising from relaxation effects. Further, the relative orientations of C-H bond vectors are determined using a DQ heteronuclear local field SSNMR experiment, employing ¹³C-¹H coupling amplification to increase sensitivity. These methods are applied to determine the molecular conformation of a uniformly ¹³C-labeled peptide, N-formyl-l-methionyl-l-leucyl-l-phenylalanine (fMLF). From just six distance and six angular restraints, two possible molecular conformations are determined, one of which is in excellent agreement with the crystal structure of a closely related peptide. The method is envisaged to a useful addition to the SSNMR repertoire for the solid-state structure determination of peptides in a variety of forms, including amyloid fibrils and pharmaceutical formulations.

Keywords: amyloid; biopharmaceuticals; magic-angle spinning; rotational resonance.

Figure 1

(a) Chemical structure of the fMLF peptide. Filled circles denote Cα (blue) and CH₃ (orange) sites that were ¹³C–¹³C dipolar recoupled at rotational resonance. (b) Pulse sequence of the 2D ¹³C–¹³C DQSQRR experiment. All filled rectangles represent π/2 pulses. Signal is acquired over an 8-step phase cycle, with pulse phases ϕ₁ = −y; ϕ₂ = +x; ϕ₃ = +x; ϕ₄ = +x; ϕ₅ = +x; ϕ₆ = +y; ϕ₇ = +x +y − x − y; ϕ₈ = +y − x − y + x; ϕ₉ = +x +x +x +x − x − x − x − x. The receiver phase ϕ_rec = +x +y − x − x − x − y +x + y. (c) One-dimensional ¹³C CP-MAS NMR spectrum of [U-¹³C,¹⁵N]fMLF showing Cα and aliphatic side-chain resonances only. (d) Two-dimensional ¹³C–¹³C DQSQRR spectra obtained using the pulse sequence in Figure 1b at MAS frequencies corresponding to the exact n = 1 rotational resonance condition with respect to Met Cα–Leu Cδ’ (5562 Hz), Met Cα–Met Cε (6642 Hz), Phe Cα–Met Cε (7078 Hz) and Leu Cα–Met Cα (7551 Hz). Note the shorthand: Lα, etc., refers to Leu Cα, etc. At ν_R = 6642 Hz, the chemical shift difference for Lα and Lδ’ is only 91 Hz away from n = 1 rotational resonance and therefore dipolar recoupling of these nuclear sites is observed. Arrows denote the carrier frequency. The ZQ excitation time, t_ex, is 4 ms for each spectrum. Green and blue contours represent positive and negative intensities, respectively.

Figure 2

The basis of the 1D DQRR experiments for selective measurement of ¹³C–¹³C distances and ¹³C–¹H bond orientations. (a) Basic pulse sequence. Phase cycling (ϕ₁–ϕ₉, ϕ_rec) is as described in Figure 1b. FSLG = frequency-switched Lee–Goldburg sequence for ¹H–¹H decoupling. (b) How the difference intensity is measured from the observed NMR spectrum. (c) At n = 1 RR, dipolar interactions are recoupled between pairs of ¹³C nuclei if the separation is less than ~7 Å. In the DQRR-CC experiment, the difference intensity is modulated by varying t_ex and t_rec. (d) In the DQRR-HLF experiment, t_ev is varied and the difference intensities are modulated according to the relative orientations of pairs of ¹³C–¹H bonds (defined by angles θ₁, θ₂ and θ₃).

Figure 3

Determination of the Leu Cα–Met CεH₃ relative orientations and ¹³Cα–¹³Cε distance in solid [U-¹³C,¹⁵N]fMLF. The MAS frequency was set to the n = 1 RR condition for Leu Cα and Met Cε (ν_R = 7551 Hz). (a) Series of spectra obtained in the CT-DQRR-CC experiment by varying t_ex and t_rec. (b) Difference intensities (filled circles) measured from the spectra in (a) and simulated curves for different internuclear distances, r_CC. Error bars represent the level of the noise. (c) Series of spectra obtained using the DQRR-HLF experiment, by varying t_ev up to the duration of one rotor cycle, t_R, and maintaining t_ex and t_rec at 4 ms. The top spectra were obtained without dipolar amplification using the pulse sequence in Figure 2a, and the bottom spectra were obtained with dipolar amplification using the pulse sequence in Figure S5. (d) DQ evolution over one rotor cycle. Filled circles denote the DQ-filtered difference intensities measured from the amplified and unamplified experiments. The line of best fit (dotted line) is shown on the amplified data set. The red shaded areas bounded by the solid lines represent the range of variability of the difference intensities to all possible C–H and CH₃ orientations.

Figure 4

Conformational analysis of fMLF from the DQRR distance and angle restraints. (a) An unrestrained random structure of fMLF, highlighting the 6 selective ¹³C–¹³C DQ coherences forming the basis for the structural restraints. (b) Combinations of peptide side-chain and main-chain torsional angles 1–8 that are consistent with the 6 distance restraints only. Connecting lines represent individual peptide conformations. The torsional angle numbers are defined in Figure S9 and explained in Table S3. Angles 9 and 10 are omitted for clarity. (c) Combinations of peptide torsional angles that are consistent with all restraints (see also Table 2). Two conformations are represented by triangles (conformer 1) and circles (conformer 2). (d) Peptide molecular conformations (conformer 1 and conformer 2) corresponding to the restrained torsional angles in (c) and Table 2. (e) Molecular conformations of fMLF determined previously from SSNMR restraints (PDB 1Q7O). (f) Molecular conformation of C-terminal methoxy fMLF [28].

Figure 5

Preliminary SSNMR analysis of the peptide [U-¹³C-FVA]Med43-50. (a) Negative-stain TEM image of the peptide fibrils. (b) Two-dimensional DQSQRR SSNMR spectrum of the peptide fibrils at a MAS frequency of 7250 Hz, which is close to, but does not correspond exactly to, the n = 1 RR condition for any pair of observed nuclei. The red arrow signifies the spectrometer carrier frequency. The doubling of resonances (Aα, A’α, Aβ, A’β, etc.) is attributed to the polymorphism of the fibrils.