Peptide Torsion Angle Measurements: Effects of Nondilute Spin Pairs on Carbon-Observed, Deuterium-Dephased PM5-REDOR Ingolf Sack,* Yael S. Balazs,* Shai Rahimipour,² and Shimon Vega* ,1 *Department of Chemical Physics and ² Department of Organic Chemistry, Weizmann Institute of Science, Rehovot 76100, Israel Received June 12, 2000; revised September 6, 2000 Reintroducing dipolar coupling between spin-1/2 nuclei (e.g., 13 C, 15 N) and spin-1 2 H, using phase-modulated deuterium dephasing pulses, provides a simple and efficient basis for obtain- ing peptide backbone torsion angles (, ) in specific stable- isotope enriched samples. Multiple homonuclear spin-1/2 interac- tions due to isotopic enrichment can arise between neighboring molecules or within a multiply labeled protein after folding. The consequences of 13 C homonuclear interactions present during 13 C- observed, 2 H-dephased REDOR measurements are explored and the theoretical basis of the experimentally observed effects is investigated. Two tripeptides are taken to represent both the general case of 2 H  -alanine (in the tripeptide LAF) and the special case of 2 H 2  -glycine (in the tripeptide LGF). The lyophilized tri- peptides exhibit narrowed spectral linewidths over time due to reduced conformational dispersion. This is due to a hydration process whereby a small fraction of peptides is reorienting and the bulk peptide fraction undergoes a conformational change. The new molecular packing arrangement lacks homonuclear 13 C spin interactions, allowing determination of (, ) backbone torsion angles. © 2001 Academic Press Key Words: solid-state NMR; CPMAS; 2 H REDOR; rotational resonance; dipolar coupling. INTRODUCTION Solid-state NMR is a unique tool for obtaining molecular level structural details of a wide variety of compounds without inherent restrictions to size or molecular ordering. Specific stable-isotope labeling is an effective strategy for obtaining local structural information in biological systems (reviews: (1–4)). An advantage of magic-angle spinning (MAS) solid- state NMR among sophisticated high-resolution structural methods is the flexibility in preparation state of the sample: crystalline, lyophilized, precipitated, membrane-bound, flash- frozen, or frozen solutions are all viable options. One of the most powerful and widely used MAS methods of modern solid-state NMR is rotational echo double resonance (REDOR) (5), which determines heteronuclear distances be- tween spin-1/2 nuclei. The experiments are easy to perform and in most cases the REDOR decay signals can be analyzed with the help of a simple master curve. Internuclear distance measurements between spin-1/2 and spin-1 deuterium ( 2 H) nuclei can be accurately measured using variations of standard spin-1/2 REDOR (6 –12). The method of PM5-REDOR (12) uses optimized phase-modulated pulses to obtain efficient 2 H dephasing. PM5-REDOR was recently used with specific stable-isotope labeling to determine  and  peptide torsion angles (13). Three peptide torsion angles define the backbone conformation of a protein, (, , ). A number of solid-state NMR ap- proaches exist to measure peptide backbone torsion angles and provide definitive information of local protein structure (14– 22). PM5-REDOR determines peptide backbone conforma- tional angles (, ) by reintroducing 13 C– 2 H or 15 N– 2 H dipolar couplings during MAS using the labeling scheme shown in Fig. 1. Site-specific stable-isotope labeling can be obtained by starting with commercially available enriched amino acids for solid-phase peptide synthesis or for growth medium supple- ments in bacterial biosynthesis (23). The torsion angle,  i , can be measured by reintroducing the dipolar coupling between a nonexchangeable deuterium at the -proton position ( 2 H i  ) and a 13 C-enriched i - 1 carbonyl carbon. The torsion angle  i is determined in a separate 15 N– 2 H i  PM5-REDOR experiment between 2 H i  and a 15 N-enriched i + 1 amide nitrogen. As in REDOR, every resolved peak in the observed spectrum can be analyzed from one experimental run. This scheme works well for both two-spin (e.g., 2 H  -ala- nine) and three-spin ( 2 H 2  -glycine) cases. The two-spin 15 N– 2 H i  PM5-REDOR experiment to obtain  i is sensitive to -sheet vs -helical conformations even when signal-to-noise is relatively low. Comparing experimental data and theoretical calculations results in two possible values for L-amino acid torsion angles or four degenerate values of glycine torsion angles. Readily available constraints from Ramachandran plots of 13 C– 15 N REDOR measurements on the same sample can result in unique torsion angles (13). The labeling scheme shown in Fig. 1 is advantageous for extracting multiple structural constraints from a single sample. Four different nuclei can be observed, providing complemen- tary spectroscopic information. The presence of multiple spin 1 To whom correspondence should be addressed. Fax: +972-8-934-4123. E-mail: civega@wis.weizmann.ac.il. Journal of Magnetic Resonance 148, 104 –114 (2001) doi:10.1006/jmre.2000.2214, available online at http://www.idealibrary.com on 104 1090-7807/01 $35.00 Copyright © 2001 by Academic Press All rights of reproduction in any form reserved.