SEA-TROSY (Solvent Exposed Amides with TROSY): A Method to Resolve the Problem of Spectral Overlap in Very Large Proteins Maurizio Pellecchia,* ,† David Meininger, † Anna L. Shen, ‡ Rick Jack, † Charles B. Kasper, ‡ and Daniel S. Sem † Triad Therapeutics Inc., 5820 Nancy Ridge Dr. San Diego, California 92121 McArdle Laboratory for Cancer Research, Medical School UniVersity of Wisconsin, Madison, Wisconsin 53706 ReceiVed December 5, 2000 ReVised Manuscript ReceiVed April 3, 2001 The recent advances of 2 H labeling in combination with the revolutionary spectral improvement provided by the development of TROSY-type experiments 1 largely reduce the problems of nuclear spin relaxation particularly in 15 N, 1 H experiments and allowed the study of complexes of molecular mass >50 kDa where only one of the components (with ∼200 amino acid residues) was 15 N/ 2 H labeled. 2 However, for even larger proteins with a large number of amino acid residues (>300) the problem of resonance overlap still represents a major obstacle for resonance assignment and chemical shift perturbation studies. Recently, several research groups proposed the use of segmental labeling to conduct studies on selectively labeled domains in multidomain proteins. 3 Although these methods are very promising for studying interdomain interactions or performing structural characterization in multidomain proteins, low expression yields still represent a major hurdle. Recently, Sattler and Fesik 4 proposed the use of lanthanide-induced shifts to increase the chemical shift dispersion. However, the increase in dispersion is rather limited (∼1 ppm at most) and a small detrimental relaxation effect has also been observed. In this communication we present a modification of 15 N, 1 H correlation experiments that is designed to reduce the problem of resonance overlap in very large proteins (>70 kDa) with a large number of amino acid residues (>300). Our idea is based on the concept that for binding studies only amides that are exposed to the solvent are of interest whereas those buried in the interior of the protein are not likely to be involved in intermo- lecular interactions. The selection of solvent exposed amide protons in a perdeuterated 15 N-labeled sample dissolved in H 2 O is obtained with the pulse scheme of Figure 1a. The sequence starts with a 15 N double filter 5 that serves to eliminate all the magnetization generated from amide protons. Water magnetization is not affected by the 15 N filter and subsequently is returned along the z-axis by the last 90° 1 H pulse prior to τ m (Figure 1a). At this time point, water z magnetization is allowed to exchange with amide protons during a variable mixing time, τ m (Figure 1a). Backbone amides that are exposed to the solvent will acquire magnetization from the solvent that can be subsequently detected with a TROSY-type experiment. 1,6 A water flip-back version of TROSY 1,6 has been adopted as it is a requisite for the successful implementation of the SEA selection. The appearance of the resulting spectrum is the same as a 15 N, 1 H TROSY spectrum, but containing only backbone amides that are solvent exposed and therefore with much fewer resonances. The signal intensity (I) of a given amide proton in the resulting 15 N, 1 H TROSY spectrum at a given mixing time, τ m , is related to the exchange rate, k ex , according to: 7 where I ∞ is the intensity at infinite mixing time (complete * Address correspondence to this author. E-mail: mpellecchia@triadt.com. † TRIAD Therapeutics, Inc. San Diego. ‡ University of Wisconsin. (1) (a) Pervushin, K.; Riek, R.; Wider, G.; Wu ¨thrich, K. Proc. Natl. Acad. Sci. U.S.A. 1997, 94, 12366-12371. (b) Wider, G.; Wu ¨thrich, K. Curr. Op. Struct. Biol. 1999, 9, 594-601. (c) Saltzmann, M.; Pervushin, K.; Wider, G.; Senn, H.; Wu ¨ thrich, K. Proc. Natl. Acad. Sci. U.S.A. 1998, 95, 13585-13590. (2) Pellecchia, M.; Sebbel, P.; Hermanns, U.; Wu ¨thrich, K.; Glockshuber, R. Nat. Struct. Biol. 1999, 6, 336-339. (3) (a) Yamazaki, T.; Otomo, T.; Oda, N.; Kyogoku, Y.; Uegaki, K.; Ito, N.; Ishino, Y.; Nakamura, H. J. Am. Chem. Soc. 1998, 120, 5591-5592. (b) Xu, R.; Ayers, B.; Cowburn, D.; Muir, T. W. Proc. Natl. Acad. Sci. U.S.A. 1999, 96, 388-393. (4) Sattler, M.; Fesik, S. W. J. Am. Chem. Soc. 1997, 119, 7885-7886. (5) (a) Otting, G.; Wu ¨thrich, K. Q. ReV. Biophys. 1990, 23, 39-96. (b) Breeze, A. L. Prog. NMR Spectrosc. 2000, 36, 323-372. (6) (a) Pervushin, K.; Wider, G.; Wu ¨thrich, K. J. Biomol. NMR 1998, 12, 345-348. (b) Zhu, G.; Kong, X. M.; Sze, K. H. J. Biomol. NMR 1999, 13, 77-81. (7) Gemmeker, G.; Jahnke, W.; Kessler, H. J. Am. Chem. Soc. 1993, 115, 11620-11621. Figure 1. (a) Pulse sequence to selectively observe solvent exposed amide protons with TROSY (SEA-TROSY). Narrow and thin bars represent 90° and 180° radio frequency pulses, respectively. Unless specified otherwise, pulse phases are along the x-axis. The pulsed field gradients are 500 μs duration with strengths of g1 ) 20 G/cm, g2 ) 30 G/cm, g3 ) 40 G/cm, g4 ) 15 G/cm, g5) 55 G/cm. The bipolar gradient gd is 0.5 G/cm and it is used to avoid radiation damping effects during t1. 15 The delay τ was set to 2.7 ms. The phase cycle was as follows: φ1 ) y, -y, -x, x; φ2 ) y; φ3 ) x; Ψrec ) x, -x, -y, y. A phase sensitive spectrum in the 15 N dimension is obtained by recording a second FID for each t2 value, with φ1 )-y, y, -x, x, φ2 )-y, and φ3 )-x, and the data were processed as described by Pervushin et al. 6 The SEA element is outlined by the dashed rectangle. (b) Pulse scheme for the 3D SEA-HNCA- TROSY. 13 C R 180° pulses are RE-BURP pulses 9 of 250 μs duration centered at 53 ppm and are designed to selectively excite the aliphatic region (excitation of ∼8000 Hz) without exciting the 13 CO region (∼177 ppm). This avoids losses of magnetization due to 15 N- 13 CO 1 J coupling constants (∼15 Hz). 13 CO decoupling pulses are off-resonance Gaussian shaped pulses of 120 μs duration shifted to 177 ppm. A phase sensitive spectrum in the 15 N dimension is obtained by recording a second FID for each t2 value, with φ1 ) 2(x), 2(-x) and φ2 ) x, -x, and the data were processed as described by Pervushin et al. 6 States-TPPI 10 quadrature detection in the 13 C R dimension was achieved by incrementing φ1. The pulsed field gradients are 500 μs duration and strengths of g1 ) 20 G/cm, g2 ) 30 G/cm, g3 ) 40 G/cm, g4 ) 25 G/cm, g5 ) 20 G/cm, g6 ) 15 G/cm, g7 ) 55 G/cm. 2 H decoupling during 13 C R evolution is achieved with a WALTZ-16 composite pulse 11 at a field strength of 2.5 kHz. For both schemes, suppression of residual water is achieved with a WATER- GATE sequence using a 3-9-19 composite pulse. 12 I ) I ∞ (1 - e - k ex τ m ) (1) 4633 J. Am. Chem. Soc. 2001, 123, 4633-4634 10.1021/ja005850t CCC: $20.00 © 2001 American Chemical Society Published on Web 04/20/2001