Article Cite This: J. Org. Chem. 2018, 83, 5035−5043
pubs.acs.org/joc
DiCE: Diastereomeric in Silico Chiral Elucidation, Expanded DP4 Probability Theory Method for Diastereomer and Structural Assignment Dongyue Xin,† Paul-James Jones,‡ and Nina C. Gonnella*,† †
Material and Analytical Sciences, Boehringer Ingelheim Pharmaceuticals, Inc., Ridgefield, Connecticut 06877, United States Information Technology, Boehringer Ingelheim Pharmaceuticals, Inc., Ridgefield, Connecticut 06877, United States
‡
S Supporting Information *
ABSTRACT: NMR chemical shift prediction at the B3LYP/ cc-pVDZ level of theory was used to develop a highly accurate probability theory algorithm for the determination of the stereochemistry of diastereomers as well as the regiochemistry. DFT-GIAO calculations were performed for each conformer using geometry optimization and a CPCM solvent model. Boltzmann averaged shielding constants were converted to chemical shifts for 1H and 13C, using the generalized linear scaling terms determined in four different solvents for 1H and 13 C and extended to 15N in DMSO. The probability theory algorithm, DiCE, was based on the DP4 method and developed for 1H, 13C, and 15N NMR using individual and combined probability data. The chemical shift calculation errors were fitted to a Student’s t-distribution for 1H and 13C and a normal distribution for 15N. The application yielded a high accuracy for structural assignment with a low computational cost.
■
INTRODUCTION Stereochemistry is a critical structural element with profound consequences in a compound’s physical and biological properties. There are numerous pharmaceutical examples where stereochemistry significantly affected a molecule’s activity. Compounds such as Zoloft, an antidepressant for major depressive disorders, obsessive-compulsive disorder, panic disorder, and social anxiety disorder has two chiral centers, yet only the (S,S)-diastereomer, which has significant activity, is marketed.1 Likewise the antibiotic penicillin has three stereocenters yielding eight possible stereoisomers, yet only the (2S,5R,6R)-diastereomer has significant antibacterial activity, and chloramphenicol, an antibiotic for the treatment of meningitis, cholera, and typhoid fever, shows only the (R,R)diastereomer is active (Figure 1).2 In addition to compound activity, determining the correct structure, including the correct stereochemistry, is essential in protecting intellectual property and avoiding costly patent disputes.3,4 Two powerful technologies for determining the stereochemistry of diastereomers include single crystal X-ray and NMR spectroscopy. While X-ray crystallography can unequivocally establish the stereochemistry of diastereomers, the technology requires sufficient quantity, purity, and solubility of the material as well as the time and resources to grow a diffraction quality crystal. NMR spectroscopy has an advantage over X-ray in that generating a diffraction quality crystal is not needed. This © 2018 American Chemical Society
Figure 1. Chemical structures of Zoloft, penicillin, and chloramphenicol with the active stereochemistry shown.
advantage further extends to the ability of NMR technology to analyze compound mixtures where only microgram quantities of material are available. NMR spectroscopy has been applied to solve the stereochemistry of diastereomers of rigid systems with well-defined conformational orientation but faces severe challenges in chiral elucidation with compounds having flexible, rotatable bonds. Chemical synthesis in combination with NMR spectroscopy offers an alternate approach to solving diastereomeric structures; however, a significant amount of time and resources need to be employed in synthesizing multiple stereoisomers. This approach is not practical for structures having multiple stereocenters and many possibilities. The Received: February 5, 2018 Published: April 6, 2018 5035
DOI: 10.1021/acs.joc.8b00338 J. Org. Chem. 2018, 83, 5035−5043
Article
The Journal of Organic Chemistry nuclear Overhauser effect (NOE), which is dependent on the through space distance (95%), and 9 were assigned with >75% probability. This corresponds to a success rate of 87% for the examples shown in Table 3, while the rate for DP4 is only 43%. In practice, the success rate could be even better since the examples in this study were already proven to be challenging by DP4. Both DiCE and DP4+ have shown improved performance over DP4, for 1H and 13C nuclei, in solving difficult stereochemistry problems.16,17 For this study, both DiCE and DP4+ show an overall comparable performance. Specifically, DiCE and DP4+ are largely in agreement, but DiCE performs better in cases 5h, 7a, 8a, 8d, 8f, 11c and 15b, while DP4+ is superior for entries 2b, 3c, 5c, 6b, 9a, 11b, and 11d. Relative to DP4 and DP4+, DiCE has the advantage of using a generalized linear scaling reference, which improves the accuracy, especially for compounds having a few atoms. DiCE does not require the use of unscaled data and the method used in the chemical shift prediction enables high accuracy at low computational cost.9 Case Study 1. Compound 2b was selected to illustrate the differences between the generalized linear scaling and linear scaling within each molecule. The experimental data of isomer 2b was compared with the calculated chemical shifts of 2a and 2b converted from the shielding constants via the two linear
a
DP4 and DP4+ data were obtained from the original papers8,16,17 or calculated as originally described. The probability for the correct isomer: +, P ≥ 95%; ?, 95% > P ≥ 50%; and -, P < 50%. bRelevant experimental data are not available for compounds 4a, 5g, 7c, 7d, 10b, 5038
DOI: 10.1021/acs.joc.8b00338 J. Org. Chem. 2018, 83, 5035−5043
Article
The Journal of Organic Chemistry
the deviation from the experimental data. With generalized linear scaling, the calculated chemical shifts favor the correct diasteromer 2b (Figure 5c, MAE = 1.75 ppm) over 2a (Figure 5a, MAE = 2.17 ppm), which is consistent with a 96% probability for the correct isomer 2b in the 13C DiCE calculation. However, although the same sets of shielding constants were analyzed, in the case of linear scaling within the molecule, the incorrect diastereomer 2a (Figure 5b, MAE = 1.15 ppm) shows a better match with the experimental data than 2b (Figure 5d, MAE = 1.92 ppm), yielding a false high probability for the incorrect structure 2a. A comparison of Figure 5a,c shows that the application of the generalized linear scaling accurately identifies 2b as the correct stereoisomer over 2a. Figure 5b,d, however, illustrates how the linear scaling within the molecule significantly reduced the MAE for structure 2a, while slightly increasing the MAE for 2b, as slopes and intercepts are different for the two structures, leading to an incorrect prediction of 2a. Therefore, real errors caused by the incorrectness of the structure might be suppressed or even completely concealed by linear scaling within the molecule. This may consequently skew the probability calculations and lead to high probabilities for incorrect structures in DP4 calculations. Given that the error distribution varies between the structures, the performance of the linear scaling within the molecule is likely to be highly dependent on the target molecule. Unlike the linear scaling within the molecule, the generalized linear scaling has the advantage of not being biased by specific molecules, especially when the number of atoms for the specific nucleus is limited. In addition, in the process of generating the generalized linear scaling terms, systematic errors for certain functional groups could be readily identified, such as heavy halogens Cl and Br, in cases of 13 C calculations9,18 and NH2 groups in 15N calculations.10 Since those systematic errors do not follow the same statistical distribution as the other atoms, they have been excluded from the probability calculations. Case Study 2. In general, DiCE probabilities have yielded impressive results. An example is compound 16 (Figure 6),
Table 3. continued and 14b. c13C atoms with known systematic errors9 were excluded from the 13C DiCE calculation. Details can be found in the Supporting Information. dCombined probabilities using 1H and 13C data.
Figure 4. Percentage of entries within each category for five methods. Categories: green, successful, P ≥ 95%; yellow, successful but questionable confidence, 95% > P ≥ 50%; red, unsuccessful, P < 50%.
scaling methods. Figure 5 shows the calculation errors, e = δexp − δcalc, of the generalized linear scaling conversion (Figure 5a,c) and linear scaling within each molecule (Figure 5b,d) for 2a and 2b, respectively. In each graph, the atoms are ordered according to the chemical shifts, with the more downfield atom on the right side of the bar graph (Figure 5a). The mean absolute error (MAE) was calculated in each case to quantify
Figure 6. Chemical structure of compound 16 (tricholomalide B) with chiral centers where the stereochemistry was varied marked with an asterisk.
where four of the five stereocenters are varied. Experimental data were available for 16a and 16b.19,20 For 16a, 1H DiCE (>99.9%) performs better than 13C DiCE (84.42%); however, the combined analysis is >99.9%. The results for 16b have both 13 C DiCE (99.78%) and 1H (99.18%) predicting the correct stereochemistry with high confidence. This example illustrates the power of DiCE in predicting the correct stereochemistry from a field of 16 complex disastereomers with multiple chiral centers (Table 4). Case Study 3. DiCE could be readily expanded to nuclei other than 1H and 13C. On the basis of the fact that combined
Figure 5. Deviation of calculated chemical shifts from experimental chemical shifts of 2b. Calculated chemical shifts from (a) 2a with generalized linear scaling (gls), MAE = 2.17 ppm; (b) 2a with linear scaling within the molecule (ls), MAE = 1.15 ppm; (c) 2b with generalized linear scaling (gls), MAE = 1.75 ppm; and (d) 2b with linear scaling within the molecule (ls), MAE = 1.92 ppm. 5039
DOI: 10.1021/acs.joc.8b00338 J. Org. Chem. 2018, 83, 5035−5043
Article
The Journal of Organic Chemistry Table 4. 1H and 13C DiCE Calculations for Compound 16a
for dibromopalau’amine. In fact, only 13C DiCE yielded the highest probability for the correct isomer, while 1H and 15N DiCE ranked the correct isomer as the second highest probability structure. Combining 13C and 1H DiCE could lead to a higher tendency for the correct isomer 18a (12R, 17R, 20R), but even so, the probability value is only 76%, indicating ambiguity in the prediction. However, after the inclusion of 15N DiCE, the probability for the correct diastereomer was >98% (Table 5). This improved performance may be attributed to the Table 5. Summary of DiCE Probabilities for Dibromopalau’amine
a13
C atoms with known systematic errors9 were excluded from13C DiCE calculation. Details can be found in the Supporting Information. a
Atoms with known systematic errors9,10 were excluded from DiCE calculations. Details can be found in the Supporting Information.
1
H and 13C DiCE probabilities are improved over those of a single nucleus, we explored the inclusion of other nuclei besides 1 H and 13C to further improve the accuracy of the DiCE probability calculation. In previous work, we systematically investigated 15N NMR chemical shift prediction as a useful tool in the structure elucidation of nitrogen-rich molecules, such as pharmaceutically relevant compounds and natural products.10 These statistical parameters from the previous study can be readily applied in 15N DiCE calculations for combined analysis. As a proof-of-concept study, palau’amine21 17 (Figure 7), a nitrogen-rich polycyclic dimeric pyrrole-imidazole alkaloid
independence of the DiCE probabilities for different nuclei. The probabilities for the correct and incorrect diastereomers are comparable to signal and noise. Ideally, noise, if completely random, could be canceled out by averaging more independent measurements, while signal would become more prominent in this process. Therefore, the incorporation of the 15N DiCE calculation can be highly beneficial for the structure elucidation of the nitrogen containing compounds. Case Study 4. The use of the 15N chemical shift DiCE probabilities has also been effective in defining and corroborating regiochemistry. This has been notably illustrated with oxazoles (Figure 8, 19a, 19b), isoxazoles (Figure 8, 20a, 20b), and oxadiazoles (Figure 8, 21a, 21b). As previously reported,10 the 15N chemical shift prediction can be a powerful tool in structure determination. However, such results may be
Figure 7. Structures of palau’amine 17 and dibromopalau’amine 18.
natural product, was investigated with multinucleus DiCE calculations. Historically, the structure of palau’amine has been revised several times, and particularly the relative configurations of the the eight stereogenic centers have puzzled the scientific community for almost 15 years.21 The stereochemistry at C12, C17, and C20 appeared to be especially ambiguous and had the most disagreements in the literature, until the final structural revision22 and total synthesis of this natural product.23 To test the performance of DiCE, 1H, 13C, and 15N DiCE calculations were performed for eight possible isomers generated by varying the stereochemistry at positions 12, 17, and 20 in dibromopalau’amine 18 (Figure 7), as experimental 15N data are available in the literature for this palau’amine derivative.22 Of the eight diastereomers, none of the single nucleus DiCE calculations could reliably assign the relative stereochemistry
Figure 8. Numbered chemical structures showing the regiochemistry of oxazoles, isoxazoles, and oxadiazoles. 5040
DOI: 10.1021/acs.joc.8b00338 J. Org. Chem. 2018, 83, 5035−5043
Article
The Journal of Organic Chemistry
was notably demonstrated in the ability to expand DiCE to include probability calculations with 15N data, yielding a significant benefit for the structure elucidation of nitrogencontaining compounds. In addition, the generalized linear scaling terms enable the widespread application for 1H and 13C in four different solvents and for 15N in DMSO. Hence because the process is generalized and the reference term is not linked to a specific molecule, the procedure may be broadly applied to different compounds. This facilitates the process and allows an element of automation to be incorporated in the method. The general performance of DiCE appears to be better than that of DP4 and similar to DP4+ for 1H and 13C, based on the molecules studied in Table 3. In addition, DiCE was specifically developed for the B3LYP/cc-pVDZ method, which we previously reported afforded us accurate predictions at low computational cost.9 Because we use the same method for geometry optimization and shielding constant calculation, this economy in computer time is naturally extended to the DiCE calculations.
augmented with the application of probabilities, especially when the chemical shift differences are within the accepted statistical distribution range. Examples given in Table 6 show that even Table 6. 15N DiCE Probabilities for Compounds 19−21
■
CONCLUSIONS We developed a probability theory algorithm for 1H, 13C, and 15 N NMR data that improved the accuracy in the prediction of the stereochemistry and regiochemistry in cases where only one set of experimental data is available. The algorithm was generated using NMR chemical shift calculations at the B3LYP/cc-pVDZ level of theory using geometry optimization and the CPCM solvent model. The generalized linear scaling terms calculated for four different solvents had a major impact on the predicted accuracy. The improved chemical shift predictions enabled a higher level of precision in calculating the statistical distribution, standard deviation, and degrees of freedom. DiCE was found to be highly successful at making accurate predictions in very challenging diastereomer and regioisomer cases, where the chemical shift prediction comparisons were ambiguous or inconclusive.
a
Calculated 15N chemical shifts for the correct isomer. bCalculated values for the regioisomer with swapped R1- and R2-substituents c Systematic error correction of −16.5 ppm was applied.10 This atom was excluded from the DiCE calculation.
with molecules having only one nitrogen atom, probabilities may be accurately predicted using DiCE. Notably, compounds 19a and 21b show chemical shift differences that fall within an accepted statistical distribution range for 15N. The DiCE analysis of 19a and 21b supports the correct isomer with ≥95% confidence in the structure. Comparison of DiCE, DP4, and DP4+. Although the basic protocols for DiCE, DP4, and DP4+ all involve DFT calculations and probability theory calculations, the implementation details and, therefore, the performance vary significantly. Table 7 lists the differences between DiCE, DP4, and DP4+. In terms of the shielding constant calculation, both DP4+ and DiCE applied a solvent model and geometry optimization at the DFT level, which led to improved accuracy in the shielding constant calculations relative to DP4. Another critical step in the process is the conversion of the shielding constants to calculated chemical shifts. DiCE applied generalized linear scaling, which could readily identify systematic errors for certain functional groups and reduce bias in specific molecules compared with linear scaling within the molecule, a method employed both by DP4 and DP4+. As noted previously, generalized linear scaling has an advantage in improved accuracy over linear scaling within the molecule since it may be applied to molecules where nuclei of interest are sparse. This
■
EXPERIMENTAL SECTION
Sample Preparation. All compounds shown were purchased from Sigma-Aldrich, Matrix Scientific, Enamine, SynChem Inc., Maybridge Inc., or Combi-blocks Inc. NMR samples were dissolved in approximately 600 μL of deuterated solvent and used without further manipulation unless otherwise indicated. NMR Spectroscopy. NMR spectra were acquired using a Bruker Avance III 600 MHz NMR spectrometer (Bruker-Biospin) operating at 600.04 MHz for 1H, 150.88 MHz for 13C, and 60.8 MHz for 15N and using a 5 mm Cryo-BBFO Z gradient probe. Spectra were acquired at 300 K. One-dimensional spectra were obtained using a sweep width of 12 kHz, a relaxation delay of 2.0 s, an acquisition time of 1.32 s, and 32 K data points for 1H NMR spectra and a sweep width
Table 7. Comparison of DiCE, DP4, and DP4+
solvent model geometry optimization at DFT level shielding constant calculation (GIAO) use of unscaled shift linear scaling with molecule generalized linear scaling systematic error identified and excluded nuclei performance for molecules studied
DP4
DP4+
DiCE
none none B3LYP/6-31G** no yes no no 1 H, 13C fair
PCM B3LYP/6-31G* mPW1PW91/6-31+G** yes yes no no 1 H, 13C good
CPCM B3LYP/cc-pVDZ B3LYP/cc-pVDZ no no yes yes 1 H, 13C, 15N good
5041
DOI: 10.1021/acs.joc.8b00338 J. Org. Chem. 2018, 83, 5035−5043
Article
The Journal of Organic Chemistry
(2) Hutt, A. G.; O’Grady, J. Drug chirality: a consideration of the significance of the stereochemistry of antimicrobial agents. J. Antimicrob. Chemother. 1996, 37, 7−32. (3) Young, D. W.; Morecombe, D. J.; Sen, P. K. The stereochemistry of beta-lactam formation in penicillin biosynthesis. Eur. J. Biochem. 1977, 75, 133−147. (4) Bycroft, B. W.; Wels, C. M.; Corbett, K.; Maloney, A. P.; Lowe, D. A. Biosynthesis of penicillin G from D- and L-[14C]- and [α-3H]valine. J. Chem. Soc., Chem. Commun. 1975, 0, 923−924. (5) Rychnovsky, S. D.; Rogers, B. N.; Richardson, T. I. Configurational Assignment of Polyene Macrolide Antibiotics Using the [13C]Acetonide Analysis. Acc. Chem. Res. 1998, 31, 9−17. (6) Peng, J.; Place, A. R.; Yoshida, W.; Anklin, C.; Hamann, M. T. Structure and absolute configuration of karlotoxin-2, an ichthyotoxin from the marine dinoflagellate Karlodinium veneficum. J. Am. Chem. Soc. 2010, 132, 3277−3279. (7) Neuhaus, D.; Williamson, M. P. The Nuclear Overauser Effect in Structural and Conformational Analysis, 2nd ed.; Wiley-VCH: New York, 2000. (8) Smith, S. G.; Goodman, J. M. Assigning stereochemistry to single diastereoisomers by GIAO NMR calculation: the DP4 probability. J. Am. Chem. Soc. 2010, 132, 12946−12959. (9) Xin, D.; Sader, C. A.; Chaudhary, O.; Jones, P. J.; Wagner, K.; Tautermann, C. S.; Yang, Z.; Busacca, C. A.; Saraceno, R. A.; Fandrick, K. R.; Gonnella, N. C.; Horspool, K.; Hansen, G.; Senanayake, C. H. Development of a 13C NMR Chemical Shift Prediction Procedure Using B3LYP/cc-pVDZ and Empirically Derived Systematic Error Correction Terms: A Computational Small Molecule Structure Elucidation Method. J. Org. Chem. 2017, 82, 5135−5145. (10) Xin, D.; Sader, C. A.; Fischer, U.; Wagner, K.; Jones, P. J.; Xing, M.; Fandrick, K. R.; Gonnella, N. C. Systematic investigation of DFTGIAO 15N NMR chemical shift prediction using B3LYP/cc-pVDZ: application to studies of regioisomers, tautomers, protonation states and N-oxides. Org. Biomol. Chem. 2017, 15, 928−936. (11) Molecular Operating Environment (MOE), version 2014.09; Chemical Computing Group Inc., 2014. (12) Lodewyk, M. W.; Siebert, M. R.; Tantillo, D. J. Computational prediction of 1H and 13C chemical shifts: a useful tool for natural product, mechanistic, and synthetic organic chemistry. Chem. Rev. 2012, 112, 1839−1862. (13) Pierens, G. K. 1H and 13C NMR scaling factors for the calculation of chemical shifts in commonly used solvents using density functional theory. J. Comput. Chem. 2014, 35, 1388−1394. (14) Ermanis, K.; Parkes, K. E.; Agback, T.; Goodman, J. M. Expanding DP4: application to drug compounds and automation. Org. Biomol. Chem. 2016, 14, 3943−3949. (15) Jones, E.; Oliphant, E.; Peterson, P. Scipy: Open Source Scientific Tools for Python. http://www.scipy.org/. (16) Grimblat, N.; Zanardi, M. M.; Sarotti, A. M. Beyond DP4: an Improved Probability for the Stereochemical Assignment of Isomeric Compounds using Quantum Chemical Calculations of NMR Shifts. J. Org. Chem. 2015, 80, 12526−12534. (17) Zanardi, M. M.; Suarez, A. G.; Sarotti, A. M. Determination of the Relative Configuration of Terminal and Spiroepoxides by Computational Methods. J. Org. Chem. 2017, 82, 1873−1879. (18) Kaupp, M.; Malkin, O. L.; Malkin, V. G. Interpretation of 13C NMR chemical shifts in halomethyl cations on the importance of spinorbit coupling and electron correlation. Chem. Phys. Lett. 1997, 265, 55−59. (19) Tsukamoto, S.; Macabalang, A. D.; Nakatani, K.; Obara, Y.; Nakahata, N.; Ohta, T. Tricholomalides A-C, new neurotrophic diterpenes from the mushroom Tricholoma sp. J. Nat. Prod. 2003, 66, 1578−1581. (20) Wang, Z.; Min, S. J.; Danishefsky, S. J. Total synthesis and structural revision of (±)-tricholomalides A and B. J. Am. Chem. Soc. 2009, 131, 10848−10849. (21) Kock, M.; Grube, A.; Seiple, I. B.; Baran, P. S. The pursuit of palau’amine. Angew. Chem., Int. Ed. 2007, 46, 6586−6594.
of 36 kHz, a relaxation delay of 0.5 s, an acquisition time of 0.904 s, and 65 K data points for 13C NMR spectra. 1H and 13C chemical shift assignments were referenced to the residual solvent peak: chloroform (7.24 ppm, 77.0 ppm), DMSO (2.50 ppm, 39.5 ppm), methanol (3.30 ppm, 49.0 ppm), and acetonitrile (1.93 ppm, 1.3 ppm). Chemical shift assignments were made using two-dimensional gradient selected correlated spectroscopy (1H,1H-gCOSY), rotating frame nuclear Overhauser effect correlation spectroscopy (1H,1H-ROESY), gradient selected heteronuclear single-quantum correlation (1H,13C-gHSQC), and gradient selected heteronuclear multiple bond correlation (1H,13CgHMBC) spectra.24−27 ROESY data were collected with a 300−500 ms mixing time. Computational Methods. The procedure used to compute NMR chemical shifts began with a minimized chemical structure. A conformational search was conducted in the gas phase using the LowModeMD algorithm28 in the molecular operating environment (MOE 2014.09) software.11 An upper energy limit of 7.0 kcal/mol and an upper conformation limit of 30 produced a computationally tractable number of conformers and an accurate solution. Input files for the optimization with the quantum mechanical methods were generated, and geometry optimizations, calculation of analytical frequencies, and calculation of NMR isotropic shielding constants were performed at the density functional theory level in Gaussian 09, revision D.01.29 The effects of the solvation were achieved using the conductor-like polarizable continuum model (CPCM)30 with four different solvents. The isotropic shielding constants were computed for all of the optimized conformers within 7 kcal/mol of the lowest energy conformer and with no imaginary frequencies. Boltzmann weighting factors for each conformer were determined using the relative free energies obtained from the frequency calculation. The gauge-independent atomic orbital approach31,32 was used for the prediction of the NMR isotropic shielding constants, which were converted to chemical shifts by the generalized linear scaling after Boltzmann averaging. The functional used in the geometry optimizations and GIAO calculations was B3LYP,33,34 with the correlation consistent cc-pVDZ basis set of Dunning.35 DiCE probabilities were calculated with an in-house developed program.
■
ASSOCIATED CONTENT
* Supporting Information S
The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.joc.8b00338. Generalized linear scaling plots, statistical parameters in various solvents, derivation of eq 2, DiCE probabilities for compounds 1−15, and experimental and calculated chemical shifts for compounds 1−16 and 18 (PDF)
■
AUTHOR INFORMATION
Corresponding Author
*E-mail:
[email protected] ORCID
Nina C. Gonnella: 0000-0001-8944-7180 Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS We gratefully acknowledge Mr. David Craska for support in the setup and management of our High Performance Computing resources and Mr. Scot Campbell for assistance with NMR data collection.
■
REFERENCES
(1) McConathy, J.; Owens, M. J. Stereochemistry in Drug Action. Prim Care Companion J. Clin Psychiatry 2003, 05, 70−73. 5042
DOI: 10.1021/acs.joc.8b00338 J. Org. Chem. 2018, 83, 5035−5043
Article
The Journal of Organic Chemistry (22) Grube, A.; Kock, M. Structural assignment of tetrabromostyloguanidine: does the relative configuration of the palau’amines need revision? Angew. Chem., Int. Ed. 2007, 46, 2320−2324. (23) Seiple, I. B.; Su, S.; Young, I. S.; Lewis, C. A.; Yamaguchi, J.; Baran, P. S. Total synthesis of palau’amine. Angew. Chem., Int. Ed. 2010, 49, 1095−1098. (24) Aue, W. P.; Bartholdi, E.; Ernst, R. R. Two-dimensional spectroscopy: Application to nuclear magnetic resonance. J. Chem. Phys. 1976, 64, 2229−2246. (25) Jeener, J.; Meier, B. H.; Bachmann, P.; Ernst, R. R. Investigation of exchange processes by two-dimensional NMR spectroscopy. J. Chem. Phys. 1979, 71, 4546−4553. (26) Bodenhausen, G.; Ruben, D. J. Natural abundance nitrogen-15 NMR by enhanced heteronuclear spectroscopy. Chem. Phys. Lett. 1980, 69, 185−189. (27) Bax, A.; Summers, M. F. Proton and carbon-13 assignments from sensitivity-enhanced detection of heteronuclear multiple-bond connectivity by 2D multiple quantum NMR. J. Am. Chem. Soc. 1986, 108, 2093−2094. (28) Labute, P. LowModeMD−implicit low-mode velocity filtering applied to conformational search of macrocycles and protein loops. J. Chem. Inf. Model. 2010, 50, 792−800. (29) 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, J. A., Jr.; 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, Ö .; Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J. Gaussian 09, Revision D.01; Gaussian, Inc.: Wallingford, CT, 2009. (30) Barone, V.; Cossi, M. Quantum Calculation of Molecular Energies and Energy Gradients in Solution by a Conductor Solvent Model. J. Phys. Chem. A 1998, 102, 1995−2001. (31) Kupka, T.; Ruscic, B.; Botto, R. E. Toward Hartree−Fock- and Density Functional Complete Basis-Set-Predicted NMR Parameters. J. Phys. Chem. A 2002, 106, 10396−10407. (32) Kupka, T.; Stachow, M.; Nieradka, M.; Kaminsky, J.; Pluta, T. Convergence of Nuclear Magnetic Shieldings in the Kohn-Sham Limit for Several Small Molecules. J. Chem. Theory Comput. 2010, 6, 1580− 1589. (33) Lee, C.; Yang, W.; Parr, R. G. Development of the Colle-Salvetti correlation-energy formula into a functional of the electron density. Phys. Rev. B: Condens. Matter Mater. Phys. 1988, 37, 785−789. (34) Becke, A. D. Density-Functional Thermochemistry. III. J. Chem. Phys. 1993, 98, 5648−5652. (35) Dunning, J. T. H. Gaussian Basis Sets for use in Correlated Molecular Calculations. I. The Atoms Boron Through Neon and Hydrogen. J. Chem. Phys. 1989, 90, 1007−1023.
5043
DOI: 10.1021/acs.joc.8b00338 J. Org. Chem. 2018, 83, 5035−5043
