NMR Structure Elucidation of Small Organic Molecules and Natural

Aug 6, 2014 - Strychnine (1) and cervinomycin A2 (2) were employed as model compounds to represent proton-rich and proton-deficient classes of molecul...
0 downloads 12 Views 1MB Size
Download PDF
Article pubs.acs.org/jnp

NMR Structure Elucidation of Small Organic Molecules and Natural Products: Choosing ADEQUATE vs HMBC Alexei V. Buevich,*,† R. Thomas Williamson,‡ and Gary E. Martin‡ †

Discovery and Preclinical Sciences, Process and Analytical Chemistry, NMR Structure Elucidation, Merck Research Laboratories, Kenilworth, New Jersey 07033, United States ‡ Discovery and Preclinical Sciences, Process and Analytical Chemistry, NMR Structure Elucidation, Merck Research Laboratories, Rahway, New Jersey 07065, United States S Supporting Information *

ABSTRACT: Long-range heteronuclear shift correlation methods have served as the cornerstone of modern structure elucidation protocols for several decades. The 1H−13C HMBC experiment provides a versatile and relatively sensitive means of establishing predominantly 3JCH connectivity with the occasional 2JCH or 4JCH correlation being observed. The two-bond and fourbond outliers must be identified specifically to avoid spectral and/or structural misassignment. Despite the versatility and extensive applications of the HMBC experiment, it can still fail to elucidate structures of molecules that are highly proton-deficient, e.g., those that fall under the socalled “Crews rule”. In such cases, recourse to the ADEQUATE experiments should be considered. Thus, a study was undertaken to facilitate better investigator understanding of situations where it might be beneficial to apply 1,1- or 1,n-ADEQUATE to proton-rich or proton-deficient molecules. Equipped with a better understanding of when a given experiment might be more likely to provide the necessary correlation data, investigators can make better decisions on when it might be advisible to employ one experiment over the other. Strychnine (1) and cervinomycin A2 (2) were employed as model compounds to represent proton-rich and proton-deficient classes of molecules, respectively. DFT methods were employed to calculate the relevant nJCH heteronuclear proton−carbon and nJCC homonuclear carbon−carbon coupling constants for this study.

N

experiments are not nearly as sensitive as HMBC when they are utilized in conjunction with small-volume cryogenic probes (1.7 or 3 mm), several reports have appeared demonstrating the utilization of the technique with submilligram samples.6a−d The only disadvantage of 1,1-ADEQUATE relative to INADEQUATE is that the former is incapable of detecting 13 C−13C correlations between adjacent quaternary carbon resonances, but this is a relatively small price to pay for the dramatic gain in sensitivity over the INADEQUATE experiment. Scheme 1 shows the most prevalent correlation pathways in the HMBC,1,2 long-range INADEQUATE,7 and 1,n-ADEQUATE experiments.4c The focus of the present study was to make a systematic comparison of HMBC and 1,nADEQUATE experiments, defining their advantages and disadvantages and outlining possible means of improving the utilization of these when dealing with challenging protondeficient molecular structures. For this study, strychnine (1) and cervinomycin A2 (2) were used as model compounds. Strychnine has been extensively characterized both experimentally and theoretically by many groups and is readily available to investigators worldwide.8 Strychnine also represents what can be considered in the present context a “proton-rich” model compound that is readily amenable to application of the

MR structure elucidation strategies involving small organic molecules and natural products rely on several homonuclear NMR correlation techniques (COSY, TOCSY, and either NOESY or ROESY) and proton−carbon direct (HSQC) and long-range (HMBC)1 heteronuclear shift correlation techniques. These experiments afford a useful protocol for the characterization of typical proton-rich natural product structures.2 Over the past decade, magnet technology has evolved, making higher field instruments more readily accessible to many investigators. In parallel, the sensitivity of NMR probes has also dramatically improved with the development of small volume and cryogenic NMR probes.3 In concert with instrumental improvements, NMR-based structure elucidation strategies have expanded to include proton-detected 13C−13C direct and long-range homonuclear correlations as a complement to 1H−13C heteronuclear protocols.4 Although 13C−13C homonuclear correlation experiments have been available since 1980 in the form of the INADEQUATE experiment,5 the prodigious sample requirements to successfully perform the heteronucleus-detected INADEQUATE experiment have severely limited utilization of the experiment in natural product structure elucidation strategies. In contrast, the proton-detected 1,1-ADEQUATE experiment, first reported in 1996, which relies on the relatively large (30−80 Hz) 1JCC coupling constant, is theoretically 32 times more sensitive than its direct 13C-detected INADEQUATE counterpart.4,5 While the 1,1- and 1,n-ADEQUATE © XXXX American Chemical Society and American Society of Pharmacognosy

Received: June 2, 2014

A

dx.doi.org/10.1021/np500445s | J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products

Article

Scheme 1. Comparison of the Correlation Pathways of the HMBC, Long-Range INADEQUATE,7 and 1,n-ADEQUATE Experiments4

accuracy of DFT-calculated J-couplings has already been verified on a large set of experimentally measured 2JCH, 3JCH, and 1JCC couplings of strychnine,8,10 which supported the belief that this accuracy would be generally transferrable to other long-range couplings as well. For the long-range nJCC couplings that it has been possible to measure thus far, this is indeed the case.8 In total, for 22 protons and 21 carbons of strychnine, 462 n JCH couplings were calculated and associated with them an additional 200 n−1JCC couplings were obtained (see Supporting Information). The reduced number of carbon−carbon couplings in comparison with proton−carbon couplings is due to the fact that each of the nJCC couplings between any pair of protonated carbons corresponds to two different n+1JCH couplings and each of the methylene groups in strychnine has double the number of JCH couplings compared with JCC. In Figure 1 the correlation of the n-bond proton−carbon Jcouplings (nJCH) and (n − 1)-bond carbon−carbon Jcouplings11 (n−1JCC) associated with the same pathway for strychnine (1) is shown. All points in Figure 1 can be divided into two subgroups along the y-axis: one with n−1JCC > 30 Hz and the other with n−1 JCC < 10 Hz. The points in the top region are exclusively composed of 2JCH vs 1JCC correlations (2JCH/1JCC). Out of 40

HMBC experiment. Cervinomycin A2, in contrast, is a complex molecule with a high degree of aromaticity and large expanses of molecular space with sparse numbers of protons (proton deficient) that can be used to provide long-range connectivity information. Solving the structure of a molecule such as cervinomycin A2, if limited to only HMBC experiments for long-range correlations, would be extremely challenging, if not impossible.9 Detailed DFT studies have also been reported for this compound.10

■

RESULTS AND DISCUSSION With the development of the proton-detected 1,1- and 1,nADEQUATE pulse sequences,4a,b the overall appearance of HMBC and 1,n-ADEQUATE spectra have become nearly identical. In both experiments, spectral data are presented in a two-dimensional proton−carbon frequency plane, thereby facilitating their direct comparison and/or spectral overlay if desired.4a In the most general case, cross-peak intensity in the HMBC experiment is proportional to sin(πnJCHΔ2); in the 1,nADEQUATE experiment, cross-peak intensity is defined by sin2(2π1JCHΔ1)sin2(πn−1JCCΔ2). The delay, Δ1, can be optimized to closely match the term 1/(41JCH), thus leading to simplification of the 1,n-ADEQUATE transfer function to sin2(πn−1JCCΔ2). Traditionally, the Δ2 delay (see Supporting Information) in both experiments is chosen in the range 60− 100 ms, aiming to minimize losses due to relaxation and to maximize the magnetization transfer for J-couplings in the range 5−8 Hz.4b Provided that the Δ2 delay in both types of experiments is identical, which is usually the case, the same response between a given proton and a remote carbon is roughly proportional to the value of nJCH coupling in HMBC and to the n−1JCC coupling in the case of the 1,n-ADEQUATE experiment.11 For the purposes of the present discussion, the differences in absolute intensities of the peaks related to the natural abundance of coupled nuclei will be disregarded and the focus will be on J-couplings only. To make a comparison between the two types of experiments, a complete set of proton−carbon and carbon− carbon J-couplings was needed initially. Unfortunately, a complete set of experimental data was not available even for a molecule as well-studied as strychnine.8 Instead, DFT calculations were used to predict these couplings. The high

Figure 1. Correlation of DFT-calculated nJCH and n−1JCC couplings in strychnine (1). Note that in the case of nJCH, where n = 2, the equivalent 13C−13C homonuclear coupling would be n−1JCC or 1JCC. Hence, each point represents the same coupling pathway in strychnine via the heteronuclear nJCH coupling in the HMBC experiment (x-axis) and via the n−1JCC 13C−13C homonuclear coupling pathway in the 1,nADEQUATE experiment (y-axis). B

dx.doi.org/10.1021/np500445s | J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products

Article

points in this region, 17 correspond to 2JCH couplings smaller than 2 Hz and five to 2JCH smaller than 1 Hz. Such small values of 17 2JCH couplings would create a very significant detection problem for responses corresponding to these couplings by a conventional HMBC experiment. In contrast, significantly larger 1JCC coupling values would be easily detected in a standard or an accordion-type 1,1-ADEQUATE experiment12 optimized for a range of 1JCC couplings. This corresponds to a distinct advantage of 1,1-ADEQUATE over HMBC for the observation of key correlations despite the higher sensitivity demands of the former experiment. In fact, this has been successfully utilized advantageously in a recent study of the coniothyrione structure6c and in the characterization of an unprecedented degradant of the antifungal agent posaconazole.6b An expansion of the lower portion of Figure 1 is shown in Figure 2. The L-shaped distribution of points in this plot clearly

Thus, the majority of points in Figure 2 belong to the fourth region of unobservable points. These points correspond to n JCH/n−1JCC correlations with n = 5 or higher. Understandably, such very long-range couplings would not be expected to afford detectable correlations in either of the two experiments except in relatively rare cases.13a,b The 1,n-ADEQUATE experiment yields predominantly correlations via 3JCC homonuclear coupling pathways (4JCH) aside from the 1JCC (2JCH) correlations that unavoidably “leak” into 1,n-ADEQUATE spectra.4a,14 Unlike the HMBC experiment, which employs a low-pass J-filter to eliminate 1JCH correlations, there is no comparable filter in the ADEQUATE pulse sequences. This observation is consistent with the fact that on average 3JCC couplings are expected to be larger than the average four-bond JCH or JCC coupling.15 However, what is somewhat surprising is the sheer number of such correlations (i.e., 66). The predominance of ADEQUATE-only observed correlations over HMBC-only (i.e., 22) underscores the high degree of complementarity of long-range correlation data from 1,n-ADEQUATE and HMBC spectra. This observation is also a key determinant of when an investigator may wish to consider the acquisition of a 1,n-ADEQUATE spectrum to solve a challenging, proton-deficient structure. The 22 cross-peaks expected to be observed only in the HMBC spectrum, which are located in the lower right quadrant in Figure 2, are all with one exception related to the 3JCH/2JCC correlations, which is consistent with the fact that the threebond proton−carbon J-couplings are likely to be large, whereas 21 (out of 71) of the corresponding two-bond carbon−carbon J-couplings fall below the 1 Hz threshold. The only exception in this region is an unusual 4JC13,H23a/3JC13,C23 correlation with coupling constants of 2.15 and −0.47 Hz, respectively. Figure 2 contains 49 correlations for commonly observed cross-peaks, which are a mixture of expected 37 3JCH/2JCC and 12 4JCH/3JCC correlations. The latter group can be considered as an exception, in that their 4JCH couplings atypically exceeded the 1 Hz cutoff value. A distribution of all possible C−H couplings in strychnine (1), together with those that would result in observable HMBC and ADEQUATE correlations (i.e., nJCH > 1 Hz and n−1JCC > 1 Hz, respectively), is shown in Figure 3. As is clear from Figure 3, among the two-bond C−H correlations, 88% would be observable by HMBC and 100% by ADEQUATE; among the three-bond C−H correlations 80% would be observable by HMBC and 59% by ADEQUATE; and among the four-bond

Figure 2. Correlation of DFT-calculated nJCH and n−1JCC couplings in strychnine (bottom part of the plot in Figure 1): 3JCH/2JCC (red), 4 JCH/3JCC (blue), nJCH/n−1JCC (n > 4) (yellow). Each point represents the same correlation pathway in strychnine (1) and contributes to the same cross-peak in HMBC and 1,n-ADEQUATE experiments.

underscores the complementary effect that 1,n-ADEQUATE data can have on the information gleaned from the HMBC spectrum. Thus, all points along the y-axis and close to zero along the x-axis would constitute additional correlations observable by 1,n-ADEQUATE that would be expected to be absent in an HMBC spectrum. More specifically, the J-coupling correlation ensemble in Figure 2 can be subdivided into four regions of cross-peaks detectable by (a) 1,n-ADEQUATE only; (b) HMBC only; (c) by both methods; and (d) by neither method. A 1 Hz threshold was chosen as a dividing criterion of the aforementioned regions. It was assumed that any J-coupling below 1 Hz (in absolute value) would not produce a cross-peak and any J-coupling larger than 1 Hz (in absolute value) would produce one. In accordance with this assumption in the case of strychnine (1), the region of 1,n-ADEQUATE-only observed correlations contains 66 points; correlations observed only by HMBC contain 22 points; correlations observed by both methods, 49 points; and those correlations that are unobservable by either of the two methods, 263 points.

Figure 3. Distribution of possible nJCH and associated n−1JCC couplings in strychnine (in blue). Distribution of HMBC-observable (nJCH > 1 Hz) and ADEQUATE-observable (n−1JCC > 1 Hz) couplings in strychnine (1) is shown in red and green, respectively. C

dx.doi.org/10.1021/np500445s | J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products

Article

C−H correlations only 14% would be observable by HMBC, compared to 76% by 1,n-ADEQUATE. These distribution plots once again indicate the preference of ADEQUATE experiments over HMBC in detecting two and four C−H bond correlations. It is also interesting to note that although the total number of n JCH couplings (for n > 1) in strychnine is larger than n−1JCC (440 vs 200), the number of potentially observable cross-peaks is significantly (by 44%) greater in the n−1 J CC -based ADEQUATE experiment (40 + 66 + 49 = 155) compared with the nJCH-based HMBC (35 + 22 + 49 = 106). The second example chosen to illustrate the differences between HMBC and 1,n-ADEQUATE experiments is the severely proton-deficient xanthone antibiotic cervinomycin A2 (2).9 Cervinomycin A2 is a predominantly aromatic molecule that represents an entirely different type of structural framework when compared with the the rigid alicyclic skeleton of strychnine (1). As in the case of strychnine, DFT methods were used to calculate 406 proton−carbon J-couplings and associate with them 253 carbon−carbon J-couplings in cervinomycin A2 at the B3LYP/6-311+G(d,p)//B3LYP/631G(d) level of theory (see Supporting Information). Correlations via nJCH and those associated with the same pathway via n−1JCC for cervinomycin A2 are shown in Figure 4.

Figure 5. Correlation of DFT-calculated nJCH and n−1JCC couplings in cervinomycin A2 (bottom part of the plot in Figure 3): 3JCH/2JCC (red), 4JCH/3JCC (blue), nJCH/n−1JCC (n > 4) (yellow). Each point represents the same pathway in cervinomycin A2 and contributes to the same cross-peak in HMBC and ADEQUATE experiments.

A2 there were 298 unobservable correlations, 41 correlations that could be observed by ADEQUATE only, six observed by HMBC only, and 47 observed by both methods. These numbers were well aligned with expectations for strychnine, thus suggesting that even in the mostly aromatic and protondeficient molecules, such as cervinomycin A2, the ADEQUATE-only correlations are more prevalent than HMBConly correlations (41 vs 6) and the total number of potentially observable cross-peaks in the ADEQUATE experiments prevails over HMBC (88 vs 53). A distribution of all possible C−H couplings in cervinomycin A2 (2) together with those that would result in observable HMBC and ADEQUATE correlations is shown in Figure 6. Similar to strychnine (Figure 3), in cervinomycin A2 among the two-bond C−H correlations 84% would be observable by HMBC and 100% by ADEQUATE; among the three-bond C− H correlations 90% would be observable by HMBC and 87% by ADEQUATE; and among the four-bond C−H correlations 15% would be observable by HMBC and 76% by ADEQUATE. In contrast to strychnine (1), the highly conjugated aromatic frame of cervinomycin A2 also allows the observation of some five-bond C−H correlations (3% by HMBC and 12% by ADEQUATE) and even two six-bond correlations, one by HMBC (6JC24,H5b/5JC5,C24 = −1.09 Hz/0.25 Hz) and the other one by 1,n-ADEQUATE (6JC27,H10/5JC10,C27 = −0.46 Hz/1.29 Hz). The advantages of HMBC over ADEQUATE are primarily associated with the inherently greater sensitivity of HMBC and the hence greater probability of detecting three-bond proton− carbon correlations than the corresponding two-bond carbon− carbon correlations identified by the 1,n-ADEQUATE experiment. However, in those situations when the amount of sample is not critical, the 1,1-ADEQUATE experiment has a considerable advantage over HMBC data in the aspect of correlations through 1JCC vs 2JCH couplings. Often small and somewhat unpredictable values of 2JCH couplings pose serious problems of interpretation relative to highly predictable, large,

Figure 4. Correlation of DFT-calculated nJCH and n−1JCC couplings in cervinomycin A2 (2). Each point represents the same pathway in cervinomycin A2 and contributes to the same cross-peak in HMBC and ADEQUATE experiments in a manner analogous to that explained in the Figure 1 caption for strychnine (1).

Analogous to strychnine, the 2JCH/1JCC correlations were well separated from the rest of the ensemble in the top portion of the plot (n−1JCC > 30 Hz). The range of 2JCH couplings between 0 and 8 Hz (in absolute value terms) was similar to that of the 3 JCH couplings. This observation once again highlights the potential utility of large, 30−80 Hz, 1JCC couplings that can be accessed with a 1,1-ADEQUATE spectrum to resolve ambiguities between 2JCH and 3JCH coupling correlations and magnify those that are associated with small values of 2JCH.3 The region in Figure 4 that encompasses the long-range correlations is expanded and shown in Figure 5 with 3JCH/2JCC (in red), 4JCH/3JCC (in blue) and nJCH/n−1JCC (n > 4) (in yellow). Applying the same 1 Hz detection limit criterion for cross-peak observation in ADEQUATE and HMBC experiments, it has been determined that in the case of cervinomycin D

dx.doi.org/10.1021/np500445s | J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products

Article

Figure 6. Distribution of possible nJCH and associated n−1JCC couplings in cervinomycin A2 (in blue). Distribution of HMBC-observable (nJCH > 1 Hz) and ADEQUATE-observable (n−1JCC > 1 Hz) couplings in cervinomycin A2 is shown in red and green, respectively.

30−80 Hz, 1JCC couplings utilized in the 1,1-ADEQUATE experiment. The 1,n-ADEQUATE experiment also has an advantage in finding carbon−carbon long-range three-bond correlations (3JCC) that otherwise would be observed as four-bond correlations (4JCH) in HMBC experiments and could be very difficult to detect by the latter method. This example with cervinomycin A2 underscores the conclusion that the 1,nADEQUATE experiment also has a four times higher probability of finding “extra-long” four-bond carbon−carbon correlations compared with the five-bond proton−carbon correlations in the HMBC experiment. Overall, the 1,n-ADEQUATE experiment is shown to be a valuable complement to the HMBC experiment. The primary advantage of the 1,n-ADEQUATE experiment is that it is capable of routinely reaching one “extra bond” (3JCC vs 4JCH) compared to HMBC and can therefore resolve ambiguities related to occasional similarities in the magnitude of 2JCH, 3JCH, and 4JCH couplings. It is noteworthy that the dual-optimized inverted 1JCC 1,n-ADEQUATE experiment14b combines readily identified 1JCC correlations and 3JCC correlation data in a single experiment with improved sensitivity relative to the conventional 1,n-ADEQUATE experiment. In the recently reported application of this experiment to the complex, proton-deficient alkaloid staurosporine,6b a number of 1JCC correlations not observed in a standard 1,1-ADEQUATE experiment were observed in the dual-optimized inverted 1JCC 1,n-ADEQUATE spectrum.14b As more and more laboratories gain access to high-field NMR spectrometers equipped with smaller diameter (3 or 1.7 mm) cryogenic probe capabilities, the utility and applicability of ADEQUATE carbon−carbon correlation experiments to solve difficult natural products structural problems can be expected to continually increase.4 We have also shown that the utilization of DFT calculations can help to facilitate an understanding of the complementarity of HMBC and ADEQUATE experiments and also provides a further understanding of when long-range correlations are likely to be observed in these useful experiments.10

■

and d orbital functions and then spin−dipolar, diamagnetic spin− orbital, and paramagnetic spin−orbital contributions were calculated on a regular contracted basis set. Prior to J-coupling calculations, the molecular geometries of strychnine (1) and cervinomycin A2 (2) were fully optimized at the B3LYP/6-31G(d) level. All calculations were done using the Gaussian 09 software package.17

■

ASSOCIATED CONTENT

* Supporting Information S

HMBC and ADEQUATE pulse sequence schemes; DFTcalculated J CH and J CC couplings for strychnine and cervinomycin A2. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*Tel: 908-740-3990. Fax: 908-740-4042. E-mail: alexei. [email protected]. Notes

The authors declare no competing financial interest.

■

REFERENCES

(1) Bax, A.; Summers, M. F. J. Am. Chem. Soc. 1986, 108, 2093−2094. (2) (a) Breton, R. C.; Reynolds, W. F. Nat. Prod. Rep. 2013, 30, 501− 524. (b) Reynolds, W. F.; Enriquez, R. G. J. Nat. Prod. 2002, 65, 221− 244. (c) Martin, G. E.; Hadden, C. E.; Russell, D. J. In Handbook of Spectroscopy; Gauglitz, G.; Vo-Dinh, T., Eds.; Wiley-VCH: New York, 2003; pp 209−268. (d) Bross-Walch, N.; Kühn, T.; Moskau, D.; Zerbe, O. Chem. Biodiversity 2005, 2, 147−177. (e) Lohr, L. L.; Marquez, B. L.; Martin, G. E. In Comprehensive Analytical Chemistry; Ahuja, S.; Jespersen, N., Eds.; Elsevier: Amsterdam, 2007; pp 259− 318. (3) (a) Martin, G. E. In Annual Reports on NMR Spectroscopy; Webb, G. A., Ed.; Elsevier: London, 2005; Vol. 56, Chapter 1, pp 1−99. (b) Hilton, B. D.; Martin, G. E. J. Nat. Prod. 2010, 73, 1465−1469. (4) (a) Martin, G. E. eMagRes; 2014, DOI: 10.1002/ 9780470034590.emrstm1370. (b) Martin, G. E.; Reibarkh, M.; Buevich, A. V.; Blinov, K. A.; Williamson, R. T. In Encyclopedia of NMR Spectroscopy; Wiley: New York, 2014; in press. (5) Uhrín, D. In Annual Reports on NMR Spectroscopy; Webb, G. E., Ed.; Elsevier: Amsterdam, 2010; Vol. 70, Chapter 1, pp 1−34. (6) (a) Martin, G. E.; Hilton, B. D.; Willcott, M. R., III; Blinov, K. A. Magn. Reson. Chem. 2011, 49, 350−358. (b) Zhong, W.; Yang, C.; Tong, W.; Martin, G. E. J. Pharm. Biomed. Anal. 2012, 66, 40−49. (c) Martin, G. E.; Buevich, A. V.; Reibarkh, M.; Singh, S. B.; Ondeyka, J. G.; Williamson, R. T. Magn. Reson. Chem. 2013, 51, 383−389. (d) Senior, M. M.; Williamson, R. T.; Martin, G. E. J. Nat. Prod. 2013, 76, 2088−2093.

EXPERIMENTAL SECTION

General Experimental Procedures. DFT calculations of strychnine (1) and cervinomycin A2 (2) J-couplings were done at the B3LYP/6-311+G(d,p) level of theory, as previously described.8 Specifically, the two-step method of J-coupling calculations was employed.16 First, the most dominant Fermi contact contributions to J-couplings were calculated on an uncontracted basis set with tighter s E

dx.doi.org/10.1021/np500445s | J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products

Article

(7) (a) Weigelt, J.; Otting, G. J. Magn. Reson. A 1995, 113, 128−130. (b) Sparks, S. W.; Ellis, P. D. J. Magn. Reson. 1985, 62, 1−11. (8) Williamson, R. T.; Buevich, A. V.; Martin, G. E. Org. Lett. 2012, 14, 5098−5101. (9) O̅ mura, S.; Iwai, Y.; Hinotozawa, K.; Takahashi, K. Y.; Kato, J.; Nakagawa, A.; Hirano, A.; Shimizu, H.; Haneda, K. J. Antibiot. 1982, 35, 645−652. (10) Williamson, R. T.; Buevich, A. V.; Martin, G. E.; Parella, T. J. Org. Chem. 2014, 79, 3387−3394. (11) The nomenclature nJCH, where n = 3, refers to the familiar threebond heteronuclear coupling that is most prevalent in the HMBC experiment. The equivalent correlation in a 1,n-ADEQUATE expeirment would be 2JCC, which is equivalent to n−1JCC when we are using n = 3 from the HMBC experiment. (12) (a) Williamson, R. T.; Marquez, B. L.; Gerwick, W. H.; Koehn, F. E. Magn. Reson. Chem. 2001, 39, 544−548. (b) Williamson, R. T.; Boulanger, A.; Vulpanovici, A.; Roberts, M. A.; Gerwick, W. H. J. Org. Chem. 2002, 67, 7927−7936. (c) Williamson, R. T.; Mc Donald, L. A.; Barbieri, L. R.; Carter, G. T. Org. Lett. 2002, 4, 4659−4662. (d) He, H.; Janso, J. E.; Williamson, R. T.; Yang, H. Y.; Carter, G. T. J. Org. Chem. 2003, 68, 6079−6082. (13) (a) Araya-Maturana, R.; Delgado-Castro, T.; Cardona, W.; Weiss-López, B. E. Curr. Org. Chem. 2001, 5, 253−263. (b) ArayaMaturana, R.; Pessoa-Mahana, H.; Weiss-López, B. Nat. Prod. Commun. 2008, 3, 445−450. (14) (a) Martin, G. E.; Williamson, R. T.; Dormer, P. G.; Bermel, W. Magn. Reson. Chem. 2012, 50, 563−568. (b) Reibarkh, M.; Williamson, R. T.; Martin, G. E.; Bermel, W. J. Magn. Reson. 2013, 236, 126−133. (15) Krivdin, L. B.; Della, E. W. Prog. NMR Spectrosc. 1991, 23, 301− 610. (16) Deng, W.; Cheeseman, R. J.; Frisch, M. J. J. Chem. Theory Comput. 2006, 2, 1028−1037. (17) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H. P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Montgomery, Jr. J. A.; Peralta, J. E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin, K. N.; Staroverov, V. N.; Kobayashi, R.; Normand, J.; Raghavachari, K.; Rendell, A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega, N.; Millam, J. M.; Klene, M.; Knox, J. E.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.; Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, O.; Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J. Gaussian 09, Revision A.02; Gaussian, Inc.: Wallingford, CT, 2009.

F

dx.doi.org/10.1021/np500445s | J. Nat. Prod. XXXX, XXX, XXX−XXX