Elucidating Connectivity and Metal-Binding Structures of Unlabeled

9693

2007, 111, 9693-9696 Published on Web 07/28/2007

Elucidating Connectivity and Metal-Binding Structures of Unlabeled Paramagnetic Complexes by 13C and 1H Solid-State NMR under Fast Magic Angle Spinning Nalinda P. Wickramasinghe, Medhat A. Shaibat, and Yoshitaka Ishii* Department of Chemistry, UniVersity of Illinois at Chicago, 845 West Taylor Street, Chicago, Illinois 60607 ReceiVed: April 6, 2007; In Final Form: July 2, 2007

Characterizing paramagnetic complexes in solids is an essential step toward understanding their molecular functions. However, methodologies to characterize chemical and electronic structures of paramagnetic systems at the molecular level have been notably limited, particularly for noncrystalline solids. We present an approach to obtain connectivities of chemical groups and metal-binding structures for unlabeled paramagnetic complexes by 13C and 1H high-resolution solid-state NMR (SSNMR) using very fast magic angle spinning (VFMAS, spinning speed g20 kHz). It is experimentally shown for unlabeled Cu(II)(Ala-Thr) that 2D 13C/1H correlation SSNMR under VFMAS provides the connectivity of chemical groups and assignments for the characterization of unlabeled paramagnetic systems in solids. We demonstrate that on the basis of the assignments provided by the VFMAS approach multiple 13C-metal distances can be simultaneously elucidated by a combination of measurements of 13C anisotropic hyperfine shifts and 13C T1 relaxation due to hyperfine interactions for this peptide-Cu(II) complex. It is also shown that an analysis of 1H anisotropic hyperfine shifts allows for the determination of electron-spin states in Fe(III)-chloroprotoporphyin-IX in solid states.

Introduction In nature, diverse paramagnetic metal ions are utilized as catalytic centers in complexes with organic ligands.1,2 Various small paramagnetic complexes, including peptide-metal complexes, have been developed as drugs,3,4 which are often offered as solid substances. The molecular structures in solids differentiate the characteristics of advanced paramagnetic materials5 and the effectiveness of these drugs.6 Thus, characterizing paramagnetic complexes in solids is an essential step toward understanding their molecular functions. However, methodologies to characterize chemical and electronic structures of paramagnetic systems at the molecular level are limited, particularly for noncrystalline solids. 13C and 1H high-resolution solid-state NMR (SSNMR) using magic angle spinning (MAS) has been widely used for the chemical and structural analysis of solid materials, including biomolecules.7 For unlabeled samples such as polymers and bioactive small compounds, 2D 13C/1H correlation SSNMR has provided resolution and assignment with useful structural information.7 For many interesting small paramagnetic complexes such as drugs and other advanced materials, preparing 13C or other isotopically labeled samples is generally difficult because suitable labeled compounds required for synthesis are often not readily available. Although SSNMR of 31P, 7Li/6Li, and other relatively abundant spins has been shown to be effective for the analysis of paramagnetic inorganic systems,8,9 these options are often not available for a variety of paramagnetic complexes. Thus, multidimensional SSNMR correlating 1H and dilute nuclei such as 13C is potentially indispensable * Corresponding author. E-mail: [email protected]. Phone: (+1) 312-4130076. Fax: (+1) 312-996-0431.

10.1021/jp0727289 CCC: $37.00

for the analysis of paramagnetic systems. In contrast, applications of 2D SSNMR involving dilute nuclei for paramagnetic systems has been limited to isotopically labeled systems10,11 until recent studies using very fast MAS (VFMAS, spinning speed g20 kHz) by our group and other.12-15 Furthermore, in spite of previous efforts,16-19 it has been challenging to elucidate metal-binding structures in unlabeled paramagnetic systems by SSNMR because of limited resolution and the difficulty of signal assignment. Recent studies by McDermott et al.20 and Emsley and Bertini et al.21 demonstrated that high-resolution 2D 13C/ 13C or 13C/15N correlation SSNMR can be obtained for 13Cand 15N-labeled paramagnetic proteins such as cytochrome P450 BM-3 heme protein and Cu(II)-Zn(II) superoxide dismutase. In these studies, however, SSNMR signals near the metal center (R e 5 Å) have not been observed, presumably because the traditional CPMAS approach was used. SSNMR methods for small unlabeled paramagnetic compounds using VFMAS can serve as the basis of methodologies for examining metal-binding structures in paramagnetic metalloproteins. In this study, we demonstrate that 2D 13C/1H correlation SSNMR under VFMAS provides the connectivity of chemical groups and assignments for the characterization of unlabeled paramagnetic systems. Also, we examine the possibility of elucidating multiple 13C-metal distance constraints for paramagnetic systems. Furthermore, we characterize electron-spin states of Fe(III) for unlabeled hemin chloride, for which the metal-binding structure is relatively well defined. Materials and Methods Cu(II)(Ala-Thr)‚1/2(H2O) was prepared following ref 22. All materials required for the synthesis and hemin chloride were purchased from Sigma-Aldrich Co. (St. Louis, MO). All © 2007 American Chemical Society

9694 J. Phys. Chem. B, Vol. 111, No. 33, 2007

Letters

Figure 1. (a) 2D 13C/1H correlation NMR spectrum of Cu(Ala-Thr)‚ 1 /2H2O obtained with adiabatic CP transfer, together with (b) an X-ray structure of this complex. The sidebands in part (a) are denoted by *. The spectrum in part (a) was obtained with a contact time of 0.3 ms at a 25 kHz spinning speed for 16 mg of the sample at 35.5 °C. In (a), the connectivity-based assignments are presented by color-coded lines for Ala (red dotted line) and Thr (blue solid line). Green circles denote cross peaks between 13C-1H pairs that are not directly bonded. The total experimental time was 35 h. See Supporting Information for details. In (b), the atoms are color coded as follows: yellow (C), blue (N), red (Cu), and pink (O), where hydrogens are omitted. The distances between Cu and 13C from the present SSNMR experiments and those from the X-ray structure22 (in parentheses) are compared.

TABLE 1: Experimental 13C Anisotropic Hyperfine Shift for 13CO in Ala of Cu(Ala-Thr) Together with Cu(II)-13C Distances Determined from the Data (RCO) and the X-ray Crystal Structure (RX-ray) 13

C δiso (ppm) -338

a

∆aniso (ppm)

RCO (Å)a

RX-ray (Å)

-251.1 ( 9.8

2.62 ( 0.03

2.78

T ) 5 °C (278 K) was used to calculate RCO.

SSNMR spectra were acquired at 9.4 T (400.2 MHz for 1H NMR) with a Varian InfinityPlus 400 NMR spectrometer using a home-built 2.5-mm MAS double-resonance NMR probe. Except for the data in Table 1, VFMAS conditions were used without 1H or 1H-1H rf decoupling.14,23 All data were processed with Varian Spinsight software. Further details including the assignment of the CH3 group in hemin chloride are available in Supporting Information (SI). Results and Discussion To examine the applicability of the methods to unlabeled paramagnetic peptides and other moderately complicated paramagnetic complexes, we applied 2D 13C/1H correlation SSNMR to Cu(II)(Ala-Thr)‚1/2H2O (Cu(Ala-Thr)) as a benchmark molecule. Cu(II)-peptide complexes have been important as drugs4 and model systems that mimic the structures and functions of natural copper-bound proteins.22 Cu(Ala-Thr) was chosen because of the moderate size, availability of its X-ray structure, and superoxide-dismutase-like activity.22 Figure 1a shows 2D 13C/1H chemical-shift correlation NMR spectrum of Cu(AlaThr) obtained with adiabatic CP transfer12,24 at a spinning speed of 25 kHz. The signal assignments to chemical groups in Figure 1a were obtained by a 2D 13C/1H correlation experiment acquired with the dipolar INEPT sequence,14,25 1D 13C dipolar INEPT experiments,14,25 and 1D 13C-1H REDOR experiments12 (details in SI). In addition to signals due to directly bonded 13C-1H pairs, some cross peaks due to remote 13C and 1H pairs separated by more than one bond were observed in the spectrum, as indicated by green circles in Figure 1a. Tracing the longrange cross peaks allows us to identify the connectivity of the

chemical groups for the unlabeled metal peptide, as indicated by a blue line (Thr) and a red line (Ala). Except for the 13CO2group in Thr, all chemical groups including the OH group in Thr are correlated through the long-range cross peaks. Although we cannot generally exclude the possibility of observing “intercomplex” cross peaks, the “intracomplex” cross peaks between 13C and 1H separated by two bonds should yield stronger cross peaks for the short 13C-1H distances (∼2 Å). Thus, we did not include the effects in the analysis. Because many organic compounds, including peptides, often consist of the same set of chemical or functional groups, the assignment approach based on long-range couplings provides a useful means for identifying the connectivity of chemical groups. With the improved resolution and signal assignment protocol established here, we examine the applicability of SSNMR in structure measurements for paramagnetic systems. For this purpose, we propose combining measurements of 13C T1 values and anisotropic hyperfine (pseudocontact) shifts to determine the distance between 13C and the metal ion (R). In the analysis of the anisotropic hyperfine shifts, we assumed an isotropic g tensor and a point electron dipole at the metal center, which yield an axially symmetric shift tensor. For this simplified spin interaction, the span of the anisotropic hyperfine shift or the unique principal value of the pseudocontact shift tensor, ∆aniso ≡ |δzz - δiso|, is given by16,18

∆aniso ) 2c

( ) µ0

(1)

4πR3

where µ0 is the permeability of vacuum, c ) g2µB2S(S + 1)/ 3kBT, T is the temperature, kB is the Boltzmann constant, µB is the Bohr magneton, g is the isotropic value of the electron g tensor, δiso is the isotropic shift, and δzz is the principal value of the pseudocontact shift tensor that gives the largest |δkk δiso| (k ) 1, 2, 3).26 S is the electron spin quantum number (S ) 1/2 for Cu(II)). See the SI for details about the equation. By substituting known constants into eq 1, we obtain

∆aniso )

{1.661 × 106S(S + 1)} (R3T)

(ppm)

(2)

where R and T are in units of Å and K, respectively.18,27 Equation 2 is valid for any nucleus. Thus, an analysis of |∆aniso| provides distance information on metal-binding structures. Table 1 shows R for 13CO in Ala (RCO) calculated from the experimental pseudocontact shift obtained by analyzing a spinning sideband pattern under MAS at 12.5 kHz at 5 °C (details in SI). Although the analysis presumes the pseudocontact shift within a complex to be a dominant anisotropic shift interaction, the resulting distance of 2.62 Å shows reasonable agreement with the corresponding X-ray distance (RX-ray) of 2.78 Å. For other 13C spins in Ala-Thr, the analysis of sidebands was difficult for signal overlapping or for weak sideband intensities, which do not allow for the determination of ∆aniso with reasonable accuracy. Thus, we attempted to determine R through measurements of 13C T1 as a result of paramagnetic relaxation (T1p), which is known to be proportional to R6. Although the R6 dependence of T1p has long been known, because of the lack of reliable assignment methods this relationship has been utilized to assign signals rather than to elucidate structures.28 Here, we simultaneously determined relative distances (Rj/RCO) for different carbons (i.e., 13CJ from the ratios of 13C T1p, (T1pj/T1pCO)1/6, where T1pCO and T1pj denote T1p for 13CO in Ala and 13Cj,

Letters

J. Phys. Chem. B, Vol. 111, No. 33, 2007 9695

Figure 2. (a) 2D 13C/1H correlation NMR spectrum of hemin chloride obtained with dipolar INEPT transfer14 (τ ) 12 µs) at a spinning speed of 26.3 kHz at 27 °C, together with (b) a slice along the 1H dimension of the CH3 peak marked with a red box. The experimental time was 19 h. Nine center bands in part a are indicated by blue arrows. (c-h) Corresponding simulated 1H spinning sideband patterns for electron spin states of S ) (c, f) 1/2, (d, g) 3/2, and (e, h) 5/2 from the Fe-H distances for (c-e) the lower limit of 5.74 Å and (f-h) the higher limit of 6.38 Å.38 (i) X-ray structure of hemin chloride,38 where CH3 groups are highlighted by yellow, other carbons are denoted by gray, and hydrogens are omitted. The color coding for other atoms is the same as that in Figure 1b except for Cl (green) and Fe (orange).

TABLE 2: Cu(II)-13C Distances Estimated from 13C T1 Values and Anisotropic Hyperfine Shift of 13CO in Ala for Cu(Ala-Thr) residue

chemical group

13C

δiso (ppm)

T1p (ms)a

(R/RCO)T1b

RSSNMR (Å)

Ala

CO CH CH3

-361 -304 181

2.9 ( 0.3 1.00 2.62 ( 0.03c 9.7 ( 0.3 1.23 ( 0.02 3.21 ( 0.07 45.2 ( 3.0 1.58 ( 0.03 4.15 ( 0.10

Thr

CO2CH CHOH CH3

-265 167 80 46

2.1 ( 0.3 6.2 ( 0.3 31.0 ( 0.9 56.1 ( 1.3

0.95 ( 0.03 1.14 ( 0.02 1.49 ( 0.03 1.64 ( 0.03

2.49 ( 0.08 2.97 ( 0.07 3.89 ( 0.09 4.30 ( 0.10

a 13C T values for CH were corrected for the effect of diamagnetic 1 3 T1 as discussed in the text. No corrections were made for other 13C species. b (R/RCO)T1 was calculated from {T1/(T1 for CO in Ala)}1/6.c The distance for 13CO in Ala (RCO) was obtained from ∆aniso in Table 1.

respectively). 13C T1 values were measured by a simple 1D inversion recovery experiment under VFMAS, and we assumed that T1p ) T1 except for CH3, which typically has a short T1 in diamagnetic systems.29 13C T1p values for CH3 were corrected for the effect of diamagnetic T1, T1dia using 1/T1p ) 1/T1 1/T1dia, where we adopted T1dia ) 75 ( 2 ms for L-Ala, which was measured by the inversion recovery method at a spinning of 20 kHz in our group. In Table 2, we list 13C T1p and (Rj/RCO) values. We also calculated absolute distances as RSSNMR using the relative distances and RCO obtained from the pseudocontact shift. The analysis was performed without prior knowledge of the corresponding distances in the X-ray structure.22 In Figure 1b, we compared the SSNMR distances with the X-ray distances. The SSNMR distances agreed with the corresponding X-ray distances within 12%. The results also confirm the validity of the assignment based on the 2D 13C/1H correlation SSNMR. Although previous SSNMR studies for quadrupolar metal nuclei showed excellent examples of distance measurements from metal to 13C or other spins for diamagnetic systems,30-33 a few previous studies demonstrated practical strategies for measurements of metal-binding structure for unlabeled paramagnetic systems.14 This study presents a promising general strategy for obtaining multiple 13C-metal distances for moderately large unlabeled paramagnetic systems by SSNMR using VFMAS. Fe(III)-protoporphyrin-IX is utilized as a prosthetic group of various heme proteins. Its aggregated forms have attracted

renewed attention as potential targets for anti-malaria drugs.34,35 In contrast to Cu(II), which takes only the value of S ) 1/2, Fe(III) is known to exhibit multiple electron spin states.1,2 It has been reported that Fe(III)-chloroprotoporphyin-IX or hemin chloride undergoes thermal equilibrium between low-spin (S ) 1/2) and high-spin (S ) 5/2) states in pyridine-chloroform solution.36,37 Because the spin state modulates the chemical nature, including the binding affinity, and the Fe(III) binding structures to porphyrin rings are generally well defined, it is rather interesting to examine the possibility of detecting the electron spin state and its dynamic nature for the hemin in solids by SSNMR. Figure 2a shows the 2D 13C/1H correlation SSNMR spectrum of hemin chloride at a spinning speed of 26.3 kHz. Although the limited sensitivity and spectral dispersion made signal assignments difficult, we could assign a 13C peak at 84 ppm to one of four 13CH3 groups (yellow in Figure 2i), as indicated in Figure 2a (assignment in SI). A slice along the 1H dimension for the CH3 peak in Figure 2b shows numerous sidebands due to hyperfine interactions. Figure 2c-h shows simulated 1H spinning sideband patterns for the indicated electron spin states for the lower and upper limits of the known Fe-H distances (r) of (c-e) 5.74 Å and (f-h) 6.38 Å.38 The simulated sideband patterns were calculated from a pseudocontact shift estimated from r and S (see SI). Although contact shifts may produce nonnegligible anisotropies,16 the extent of the thermally averaged contact coupling predicted from the isotropic 1H shift for CH is small (∼40 ppm), as demonstrated for other 3 1H spins in heme groups.27 Thus, we analyzed the data assuming that the pseudocontact interaction dominates the anisotropic shift. The experimental 1H spinning sideband pattern best fits the simulation at S ) 5/2 for both distances. Therefore, the system is most likely to be in the high-spin state (S ) 5/2) in solids. Conclusions We demonstrated strategies to obtain connectivities, multiple distances, and electron spin states for paramagnetic complexes using 1H and 13C SSNMR under VFMAS. It is likely that the presented methods are useful for a variety of unlabeled paramagnetic complexes, including drugs, catalysts, and nanoassembled materials. Because solution NMR is often not effective for these small paramagnetic complexes,13 SSNMR

13C-metal

9696 J. Phys. Chem. B, Vol. 111, No. 33, 2007 provides an excellent alternative for such systems. Although the distance and electron spin state analysis involve considerable simplifications, the present SSNMR analysis provided reasonably accurate structural parameters or precious information on electron spin states for systems involving Cu(II) and Fe(III), which form important classes of paramagnetic systems in both biological and material science. The data obtained here also indicate that signals for larger peptides/proteins that are selectively 13C labeled at metal-binding sites can be well resolved in 2D 13C/1H correlation SSNMR with sufficient sensitivity. Although applications of our approach to biomolecules require further studies, the VFMAS approach presented here may open an avenue toward the analysis of detailed metalbinding structures of paramagnetic proteins, in combination with modern biomolecular SSNMR methodologies, which were recently applied to paramagnetic metalloprotiens.20,21,39 General applicability, a detailed discussion on structural elucidation, and limitations of the present approach will be examined in our forthcoming studies. Acknowledgment. We thank Profs. Richard Kassner and Cynthia Jameson at UIC for helpful discussions. This work was supported in part by the NIH RO1 program (AG028490), the Dreyfus Foundation Teacher-Scholar Award program, and the NSF CAREER program (CHE 449952). Supporting Information Available: Further information including experimental details and the signal assignments for Figures 1 and 2. This material is available free of charge via the Internet at http://pubs.acs.org. References and Notes (1) Bertini, I.; Gray, H. B.; Lippard, S. J.; Valentine, J. S. Bioinorganic Chemistry; University Science Books: Sausalito, CA, 1994. (2) Kaim, W.; Schwederski, B. Bioinorganic Chemistry: Inorganic Elements in the Chemistry of Life: An Introduction and Guide; John Wiley & Sons: New York, 1994. (3) Farrell, N. Uses of Inorganic Chemistry in Medicine; Royal Society of Chemistry: Cambridge, U.K., 1999. (4) Ming, L. J. Med. Res. ReV. 2003, 23, 697. (5) Curry, R. J.; Gillin, W. P.; Clarkson, J.; Batchelder, D. N. J. Appl. Phys. 2002, 92, 1902. (6) Byrn, S. R.; Pfeiffer, R. R.; Stephenson, G.; Grant, D. J. W.; Gleason, W. B. Chem. Mater. 1994, 6, 1148. (7) Schmidt-Rohr, K.; Spiess, H. W. Multidimensional Solid-State NMR and Polymers; Academic Press: San Diego, CA, 1994. (8) Zhang, X. M.; Zhang, C.; Guo, H. Q.; Huang, W. L.; Polenova, T.; Francesconi, L. C.; Akins, D. L. J. Phys. Chem. B 2005, 109, 19156.

Letters (9) Lee, Y. J.; Wang, F.; Grey, C. P. J. Am. Chem. Soc. 1998, 120, 12601. (10) Spaniol, T. P.; Kubo, A.; Terao, T. Mol. Phys. 1999, 96, 827. (11) Crozet, M.; Chaussade, M.; Bardet, M.; Emsley, L.; Lamotte, B.; Mouesca, J. M. J. Phys. Chem. A 2000, 104, 9990. (12) Ishii, Y.; Wickramasinghe, N. P.; Chimon, S. J. Am. Chem. Soc. 2003, 125, 3438. (13) Ishii, Y.; Wickramasinghe, N. P. 1H and 13C High-Resolution SolidState NMR of Paramagnetic Compounds under Very Fast Magic Angle Spinning. In Modern Magnetic Resonances; Webb, G., Ed.; Springer: Berlin, 2006. (14) Wickramasinghe, N. P.; Ishii, Y. J. Magn. Reson. 2006, 181, 233. (15) Kervern, G.; Pintacuda, G.; Zhang, Y.; Oldfield, E.; Roukoss, C.; Kuntz, E.; Herdtweck, E.; Basset, J. M.; Cadars, S.; Lesage, A.; Coperet, C.; Emsley, L. J. Am. Chem. Soc. 2006, 128, 13545. (16) Nayeem, A.; Yesinowski, J. P. J. Chem. Phys. 1988, 89, 4600. (17) Liu, K.; Ryan, D.; Nakanishi, K.; McDermott, A. J. Am. Chem. Soc. 1995, 117, 6897. (18) Brough, A. R.; Grey, C. P.; Dobson, C. M. J. Am. Chem. Soc. 1993, 115, 7318. (19) Heise, H.; Kohler, F. H.; Xie, X. L. J. Magn. Reson. 2001, 150, 198. (20) Jovanovic, T.; McDermott, A. E. J. Am. Chem. Soc. 2005, 127, 13816. (21) Pintacuda, G.; Giraud, N.; Picrattelli, R.; Bockmann, A.; Bertini, I.; Emsley, L. Angew. Chem., Int. Ed. 2007, 46, 1079. (22) Facchin, G.; Torre, M. H.; Kremer, E.; Piro, O. E.; Castellano, E. E.; Baran, E. J. J. Inorg. Biochem. 2002, 89, 174. (23) Wickramasinghe, N. P.; Shaibat, M.; Ishii, Y. J. Am. Chem. Soc. 2005, 127, 5796. (24) Hediger, S.; Meier, B. H.; Ernst, R. R. Chem. Phys. Lett. 1995, 240, 449. (25) Vita, E. D.; Frydman, L. J. Magn. Reson. 2001, 148, 327. (26) Antzutkin, O. N.; Lee, Y. K.; Levitt, M. H. J. Magn. Reson. 1998, 135, 144. (27) Bertini, I.; Luchinat, C.; Parigi, G. Solution NMR of Paramagnetic Molecules: Applications to Metallobiomolecules and Models, 1st ed.; Elsevier: Amsterdam; New York, 2001. (28) Aime, S.; Bertini, I.; Luchinat, C. Coord. Chem. ReV. 1996, 150, 221. (29) Naito, A.; Ganapathy, S.; Akasaka, K.; McDowell, C. A. J. Magn. Reson. 1983, 54, 226. (30) Gullion, T.; Vega, A. J. Prog. Nucl. Magn. Reson. Spectrosc. 2005, 47, 123. (31) Dybowski, C.; Bai, S. Anal. Chem. 2006, 78, 3853. (32) Grant, C. V.; Frydman, V.; Harwood, J. S.; Frydman, L. J. Am. Chem. Soc. 2002, 124, 4458. (33) Kao, H. M.; Grey, C. P. J. Phys. Chem. 1996, 100, 5105. (34) Egan, T. J. J. Inorg. Biochem. 2002, 91, 19. (35) Vangapandu, S.; Jain, M.; Kaur, K.; Patil, P.; Patel, S. R.; Jain, R. Med. Res. ReV. 2007, 27, 65. (36) Dugad, L. B.; Medhi, O. K.; Mitra, S. Inorg. Chem. 1987, 26, 1741. (37) Hill, H. A. O.; Morallee, K. G. J. Am. Chem. Soc. 1972, 94, 731. (38) Koenig, D. F. Acta. Crystallogr. 1965, 18, 663. (39) Balayssac, S.; Bertini, I.; Lelli, M.; Luchinat, C.; Maletta, M. J. Am. Chem. Soc. 2007, 129, 2218.