15N-Detected TROSY NMR experiments to study large disordered proteins in high-field magnets

Abstract

Intrinsically disordered regions (IDRs) of proteins are critical in the regulation of biological processes but difficult to study structurally. Nuclear magnetic resonance (NMR) is uniquely equipped to provide structural information on IDRs at atomic resolution; however, existing NMR methods often pose a challenge for large molecular weight IDRs. Resonance assignment of IDRs using ¹⁵N^D-detection was previously demonstrated and shown to overcome some of these limitations. Here, we improve the methodology by overcoming the need for deuterated buffers and provide better sensitivity and resolution at higher magnetic fields and physiological salt concentrations using transverse relaxation optimized spectroscopy (TROSY). Finally, large disordered regions with low sequence complexity can be assigned efficiently using these new methods as demonstrated by achieving near complete assignment of the 398-residue N-terminal IDR of the transcription factor NFAT1 harboring 18% prolines.

Fig. 1. Transverse relaxation rates (R₂) of ¹⁵N and ¹³CO spins.

a Comparison of R₂ for spins tumbling at different rotational correlation times. The magnetic field is fixed at 28.2 T. b Comparison of R₂ at different magnetic fields. The rotational correlation time is fixed at 20 ns.

Fig. 2. Schematic representation of magnetization transfer in the suite of ¹⁵N^H-TROSY direct detection experiments.

Magnetization transfer in the four experiments for backbone resonance assignments is shown with red arrows. Chemical shifts are evolved for the atoms shown in blue and atoms in black are used to transfer magnetization in pulse scheme without chemical shift encoding. The green shaded region is i − 1 amino acid residue and orange shaded region is i amino acid.

Fig. 3. IPAP scheme for detecting ¹⁵N^H-TROSY spin state.

a The pulse module appended towards the end to each of the four pulse schemes for evolving in-phase and anti-phase ¹⁵N coherence by differential placement of the 180° ¹H pulse. The delays Δ₁ and Δ₂ are $\frac{1}{4J_{NH}}$ and $\frac{1}{4J_{N{C}^{\prime }}}$, respectively. b ¹⁵N doublets are acquired in both in-phase and anti-phase spectra. c The in-phase and anti-phase spectra are added and subtracted to select the ¹⁵N^H anti-TROSY and TROSY spin state, respectively.

Fig. 4. Comparing ¹⁵N^H-TROSY and ¹⁵N^D detection.

a Box plot distribution of signal to noise ratio. b Overlay of 2D ¹³C′-¹⁵N projection of ¹⁵N^D detect (blue), ¹⁵N^H-TROSY (red) haCACON and ¹H detect (green) HNCO. c One dimensional slices in ¹⁵N dimension of marked peaks in b. The blue and red traces are of ¹⁵N^D detect and ¹⁵N^H-TROSY, respectively.

Fig. 5. CA-N dipeptide fingerprint.

A ${}^{13}C{}_{i-1}^{\alpha }-{}^{15}N{}_i$ projection from haCACON of NFAT_1–398. The fingerprint regions of dipeptides are marked with red boxes. ’X’ stands for any amino acid residue.