Yuccalechins A–C from the Yucca schidigera Roezl ex Ortgies Bark: Elucidation of the Relative and Absolute Configurations of Three New Spirobiflavonoids and Their Cholinesterase Inhibitory Activities

The ethyl acetate fraction of the methanolic extract of Yucca schidigera Roezl ex Ortgies bark exhibited moderate acetylcholinesterase (AChE) and butyrylcholinesterase (BChE) inhibitory activity (IC50 47.44 and 47.40 µg mL−1, respectively). Gel filtration on Sephadex LH-20 and further RP-C18 preparative HPLC of EtOAc fraction afforded 15 known and 3 new compounds, stereoisomers of larixinol. The structures of the isolated spirobiflavonoids 15, 26, and 29 were elucidated using 1D and 2D NMR and MS spectroscopic techniques. The relative configuration of isolated compounds was assigned based on coupling constants and ROESY (rotating-frame Overhauser spectroscopy) correlations along with applying the DP4+ probability method in case of ambiguous chiral centers. Determination of absolute configuration was performed by comparing calculated electronic circular dichroism (ECD) spectra with experimental ones. Compounds 26 and 29, obtained in sufficient amounts, were evaluated for activities against AChE and BChE, and they showed a weak inhibition only towards AChE (IC50 294.18 µM for 26, and 655.18 µM for 29). Furthermore, molecular docking simulations were performed to investigate the possible binding modes of 26 and 29 with AChE.


Introduction
Yucca schidigera Roezl ex Ortgies (syn. Yucca mohavensis Sargent) belonging to the Asparagaceae family, is a small evergreen tree (up to 5 m tall), distributed from Southern Nevada to Mexico (Baja California) [1]. Native Americans used the plant for food and fiber, and its extracts have been used for centuries in folk medicine to treat a wide variety of inflammatory disorders, headaches, A great number of phenolic compounds have been identified as candidates for AD treatment [25,31,32]. They constitute of one of the widest chemical classes amongst plant secondary metabolites. To date, phenolic substances have been identified with many pharmacological effects including antioxidant, anti-inflammatory, antimutagenic, chemopreventive, anticancer, and antiviral activities. Some plant phenolics have been demonstrated to inhibit both AChE and BChE to varying extents. Most of these studies focused on in vitro tests, and only few studies were performed on insects, tissue, and animal models, with rarely any clinical studies [33]. Phenolics, besides their AChE and BChE inhibitory activities, also have very important antioxidant activity, which may enhance their protective effects. It has been proven that oxidative stress caused by reactive oxygen species (ROS) is involved in the aging processes. It has been suggested that free radicals damage mitochondria in certain areas of the brain that are particularly important for memory and cognitive processes and are associated with the pathogenesis of AD [34][35][36]. Hence, supplementation of the diet with antioxidants in people may reduce the risk of AD [34]. This was a major point of a number of studies performed on plants with high antioxidant potential [37]. Moreover, numerous reports indicated multitarget effects of resveratrol on AD [25]. Resveratrol oligomers showed a significant AChE/BChE inhibitory activity [38], and it was suggested to be used as a starting compound in the design of multitargeted drugs for the treatment of AD [39]. The diverse biological effects of the constituents of Y. schidigera bark encouraged us to investigate further the structurally related compounds using modern chromatographic and spectroscopic techniques.
Furthermore, as nature possesses the ability to create innumerable complex chemical structures, very often with chiral properties [40,41], the need to properly assign stereochemistry of natural products has emerged [42,43]. The use of quantum chemical calculations and computer-assisted structure elucidation (CASE) methods in solving structural validation problems simplified this task and reduced the risk of misinterpretations [44,45]. DP4-based nuclear magnetic resonance (NMR) chemical shift calculation is one of the most advanced approaches for stereochemical assignments of organic molecules when only one set of experimental data is available [46]. This method implements gauge-independent atomic orbital (GIAO) NMR for chemical shift calculations of geometries obtained by Merck molecular force fields (MMFFs) [47][48][49][50]. As it failed in assignment of some challenging molecules, Grimblat et al. [51] developed a DP4+ probability-based chemical shift analysis, where using B3LYP/6-31+G** geometries and adding the unscaled shift values significantly increased the performance of the method, which led to more accurate and confident results in establishing the stereochemistry of challenging isomeric compounds. However, relying solely on this approach often is not sufficient in determining the absolute configuration (AC) of natural compounds. One of the approaches, when the chiral compound possesses an appropriate chromophore, is the use of electronic circular dichroism (ECD) by comparing the experimental spectrum with the one calculated by time-dependent density functional theory (TDDFT) [52][53][54]. This, along with the careful study of nuclear Overhauser effects (NOEs) observed in the NMR spectra and using H-H or C-H coupling constants, provides an unambiguous tool for assignment of the AC.
Thus, the aim of this work was to study numerous phenolics of the plant, evaluate the AChE/BChE inhibitory activity of newly isolated compounds, and to assign their stereochemistry and absolute configurations using DP4+ probability-based chemical shift analysis and quantum chemical calculation methods.

Structural Characterization of the New Phenolic Compounds
Yuccalechin A (15) was isolated as an off-white, amorphous solid exhibiting a UV absorption maximum at 215 nm, with a specific rotation of [α] D 20 = +95. The negative HRESIMS spectra of 15 showed deprotonated molecule at m/z 541.1131, and its molecular formula was determined as C 30 H 21 O 10 (calcd. 541.1140). The 13 C-NMR spectra of 15 showed 25 signals, sorted by the distortionless enhancement by polarization transfer with retention of quaternaries (DEPTQ) and heteronuclear single quantum coherence (HSQC) experiments into 1 CH 2 , 9 CH, and 15 quaternary carbons with a characteristic signal at δ C 177.1, assignable to the γ-lactone carbonyl in C-1". The difference in the observed number of carbon atoms compared to HRESIMS was explained by the presence of two similar sets of aromatic protons, corresponding to the para-substituted aromatic groups at δ  Analysis of long-range correlations visible in the heteronuclear multiple bond coherence (HMBC) spectrum gave characteristic cross-peaks ( Figure 3) between H-3" and spiro-center C-2" (δ 61.5), γ-lactone C-1" (δ 177.1), p-hydroxyphenyl C-10" (δ 128.3), and dihydrobenzopyran C-8 (δ 106.7), while the H-2 correlated with C-8a (δ 153.1) and C-1 (δ 130.6) of the B-ring. These data were in close agreement with those of larixinol [16,67]. Analysis of long-range correlations visible in the heteronuclear multiple bond coherence (HMBC) spectrum gave characteristic cross-peaks ( Figure 3) between H-3′′ and spiro-center C-2′′ (δ 61.5), γ-lactone C-1′′ (δ 177.1), p-hydroxyphenyl C-10′′ (δ 128.3), and dihydrobenzopyran C-8 (δ 106.7), while the H-2 correlated with C-8a (δ 153.1) and C-1′ (δ 130.6) of the B-ring. These data were in close agreement with those of larixinol [16,67].   [68,69]. The 2,3-trans-flavan-3-ols showed the preference of the B-ring to be in an equatorial position, while for the 2,3-cis-compounds the distorted equatorial position, or a significant population of the axial C-2 aryl conformations was present [70]. The flavan-3-ol compounds, like (+)-(2R,3S)-catechin and (-)-(2R,3R)-epicatechin, consist of two benzene rings (A and B) and a pyran C-ring. They have two stereocenters and, therefore, four possible diastereomers, 2,3-trans-(2R,3S)/(2S,3R) and 2,3-cis-(2R,3R)/(2S,3S), with the C-ring being conformationally flexible. The rapid flexing within the C-ring can bring the B-ring into a pseudoequatorial (E-conformer) or a pseudoaxial (A-conformer) position ( Figure 4). The equilibrium between different forms depends on the solvent used, hence H-H and C-H coupling constants should be considered as time-averaged values (e.g., A:E ratios of about 41:59 (DMSO), 30:70 (dioxane), and 33:67 (water) for (+)-catechin) [68,69]. The 2,3-trans-flavan-3-ols showed the preference of the B-ring to be in an equatorial position, while for the 2,3-cis-compounds the distorted equatorial position, or a significant population of the axial C-2 aryl conformations was present [70]. According to the small coupling constant (J ≈ 2 Hz), protons H-2 and H-3 in 15 were deduced as having cis-relative configurations, similarly to epicatechin [71] and epiafzelechin [66]. The orientation According to the small coupling constant (J ≈ 2 Hz), protons H-2 and H-3 in 15 were deduced as having cis-relative configurations, similarly to epicatechin [71] and epiafzelechin [66]. The orientation of protons H-2 and H-3 can be deduced also from the direct H-C coupling constants, which depend on the stereochemistry of the heteroatom-substituted cyclohexane C-ring, presenting smaller values for axial H-C than for the equatorial atoms (typically ∆J = 4 Hz)-the so-called normal Perlin effect-which reflects the greater length of the axial C-H bonds. The bond lengthening in the axial position occurs as a result of hyper-conjugative σ C-H → σ* C-H interactions of anti-periplanar C-H bonds in the axial position [72,73]. On the other hand, one has to remember that axial protons at the β-carbons (i.e., C-3 of C-ring) of hetero-substituted cyclohexanes could also present coupling constants larger than those of the equatorial protons (the reverse Perlin effect) [74,75], but no evidence was found to substantiate it in the case of 15. Furthermore, the 1 J CH coupling displays an angular dependence for C-H bonds adjacent to π-bonds. In many cases, σ C-H → π* interactions lengthen C-H bonds, with maximized overlap of interacting orbitals when the C-H bond is aligned with the π-bond, thus giving the minimum 1 J CH value [76]. In order to explore these previously undescribed properties of flavan-3-ols, we measured a series of C-H couplings ( 1 J CH , 2 J CH and 3 J CH ) using the F2-coupled HSQC [77] and HSQC-HECADE experiments [78] for (+)-catechin and (-)-epicatechin (close analogs of (+)-afzelechin and (-)-epiafzelechin) and compounds 15, 26, and 29 (Table 3). The accurate values of 1 J CH for C-2 (144.0 Hz) and C-3 (146.2 Hz) for compound 15 suggested that H-2 adopted a pseudoaxial, and H-3 a pseudoequatorial orientation.  Additionally, 2 J CH coupling provides conformational information that can identify the position of an oxygen functionality-when it is gauche to its geminal proton, the coupling becomes large (4)(5)(6)(7), and when it is anti, 2 J CH becomes small (0-2 Hz). On the other hand, 3 J CH coupling follows a Karplus-type dependence, thus showing smaller values for gauche (1-3 Hz) than for anti-periplanar conformation (5)(6)(7)(8) [79,80]. The data gathered in Table 3 for compound 15 agreed with the relative 2,3-cis-configuration.
The relative configuration between the C-2"/C-3" of the benzofuran F-ring was deduced primarily on the basis of the chemical shift of C-1"(177.1), as reported by Nakashima et al. [15]. The article reported that 13 C NMR spectra of yuccaols A-E and gloriosaols A-E in CD 3 OD showed the γ-lactone carbonyl atoms appearing within range δ C 174.6-176.9 for compounds with cis-configurations between C-2" and C-3", while for those with trans-configuration between δ C 178.6-181.1. Interestingly, we observed that yuccaols A-E, yuccalide A and gloriosaol A, and gloriosaols C-E isolated in this study had 1 J CH couplings measured for the C-3" appearing within range 154.2-155.0 Hz for compounds with the cis-configuration between C-2" and C-3" (2"R,3"S or 2"S,3"R), and 152.9-153.3 Hz for compounds adopting the trans-configuration (2"R,3"R or 2"S,3"S), and they were inversely correlated with the chemical shifts of corresponding C-1" carbon atoms. The 1 J CH for C-3" in compounds 15, 26, and 29 were 157.2 (cis-), 156.3 (cis-), and 150.8 Hz (trans-), respectively, while the chemical shifts of C-1" were 177.1, 177.1, and 180.6, respectively. Additionally, the DP4+ probability calculation was implemented as confirmation for establishing the relative configuration of 15. Based on the aforementioned observation and 2D-ROESY correlations from H-3" to H-2 and to H-2 /H-6 ( Figure 5), the relative configuration of 15 was determined as 2"R*,3"S*,2R*,3R*. However, because of the flexibility of the C-ring, to confirm the stereochemistry of chiral centers C-2 and C-3, all the possible isomers had to be generated and optimized using the DFT/B3LYP/6-31G(d,p)/IEFPCM/methanol (B3LYP = Becke, 3-parameter, Lee-Yang-Parr; IEFPCM = integral equation formalism for the polarizable continuum model) level of theory. Among them, only 2"R*,3"S*,2R*,3R* and 2"R*,3"S*,2S*,3R* were consistent with observed NOEs and coupling constants. Subsequently, geometrical optimization and further NMR calculations of the two isomers were performed using the gauge-independent atomic orbital (GIAO) method by implementing mpw1pw91/6-311G+ (d,p)/IEFPCM/methanol. The DP4+ probability calculation indicated that the probability of all atoms (sum of scaled and unscaled probabilities of all H and C atoms) for isomer 2′′R*,3′′S*,2R*,3R* was 100%, which was in accordance with the observed coupling constants ( Figure  S11). Furthermore, the mean average error (MAE) and corrected mean average error (CMAE) values along with correlation coefficients were calculated to evaluate correct/incorrect assignments. As shown in Table 4, MAE and CMAE values showed higher errors for incorrect isomers, confirming the relative configuration of 15 as for 2"R*,3"S*,2R*,3R*.  However, because of the flexibility of the C-ring, to confirm the stereochemistry of chiral centers C-2 and C-3, all the possible isomers had to be generated and optimized using the DFT/B3LYP/6-31G(d,p)/IEFPCM/methanol (B3LYP = Becke, 3-parameter, Lee-Yang-Parr; IEFPCM = integral equation formalism for the polarizable continuum model) level of theory. Among them, only 2"R*,3"S*,2R*,3R* and 2"R*,3"S*,2S*,3R* were consistent with observed NOEs and coupling constants. Subsequently, geometrical optimization and further NMR calculations of the two isomers were performed using the gauge-independent atomic orbital (GIAO) method by implementing mpw1pw91/6-311G+ (d,p)/IEFPCM/methanol. The DP4+ probability calculation indicated that the probability of all atoms (sum of scaled and unscaled probabilities of all H and C atoms) for isomer 2"R*,3"S*,2R*,3R* was 100%, which was in accordance with the observed coupling constants ( Figure S11). Furthermore, the mean average error (MAE) and corrected mean average error (CMAE) values along with correlation coefficients were calculated to evaluate correct/incorrect assignments. As shown in Table 4, MAE and CMAE values showed higher errors for incorrect isomers, confirming the relative configuration of 15 as for 2"R*,3"S*,2R*,3R*. In order to determine the absolute configuration of 15, the 3D structure was optimized by conformational analysis (refer to method section). All conformers occurring in the energy window of 5 kcal mol −1 were subjected to geometrical optimization and minimization ( Figure S10). Further calculations of excitation states, rotatory strength, and, hence, simulated ECD spectrum were performed by the TD-DFT/B3LYP/6-31G(d,p)/IEFPCM/methanol level of theory. All spectra obtained were Boltzmann averaged and, after applying UV correction of +10 nm and half band of 0.2 eV, compared to the experimentally obtained spectrum of 15, recorded in methanol. The ECD spectrum calculated was in good agreement with the experimental one ( Figure 6), while its enantiomer showed opposite Cotton effects. Therefore, the absolute configuration of 15 was assigned as 2"R,3"S,2R,3R.   Table 2). The analysis of 2D-NMR spectra provided sufficient data to ascribe the planar structure of 26 as identical with the one of 15, but the difference in the observed retention time (Figure 1, Table 1) and different optical rotation suggested that compound 26 was a stereoisomer of 15. The 1 H-1 H vicinal coupling constants observed for the C-ring suggested that it adopted 2,3-trans-configuration (afzelechin-type), with a di-pseudo-equatorial orientation of H-2/H-3, as suggested by the large 1 JCH coupling constants, large 2 JC3-H2, and small 2 JC2-H3, which were reasonably similar to that of (+)-catechin (Table 3) and NOEs observed between H-2′/H-6′ and both H-3 and H-4 ( Figure 5). The relative configuration between C-2′′/C-3′′ of the benzofuran ring was deduced from the chemical shift of C-1′′ (177.1) and 1 JCH for C-3′′ (156.3 Hz); therefore, it was a cisisomer. In order to establish and prove the relative configuration of C-2/C-3, possible isomers of the  Table 2). The analysis of 2D-NMR spectra provided sufficient data to ascribe the planar structure of 26 as identical with the one of 15, but the difference in the observed retention time (Figure 1, Table 1) and different optical rotation suggested that compound 26 was a stereoisomer of 15. The 1 H-1 H vicinal coupling constants observed for the C-ring suggested that it adopted 2,3-trans-configuration (afzelechin-type), with a di-pseudo-equatorial orientation of H-2/H-3, as suggested by the large 1 J CH coupling constants, large 2 J C3-H2 , and small 2 J C2-H3 , which were reasonably similar to that of (+)-catechin (Table 3) and NOEs observed between H-2 /H-6 and both H-3 and H-4 ( Figure 5). The relative configuration between C-2"/C-3" of the benzofuran ring was deduced from the chemical shift of C-1" (177.1) and 1 J CH for C-3" (156.3 Hz); therefore, it was a cis-isomer. In order to establish and prove the relative configuration of C-2/C-3, possible isomers of the C-ring-2"R,3"S,2R,3S, 2"R,3"S,2S,3R, and 2"S,3"S,2S,3S-were generated and imported to conformational analysis followed by geometrical optimization ( Figure S24) and DP4+ probability analysis of calculated chemical shifts. However, after the first optimization step, isomer 2"R,3"S,2S,3R was excluded from further calculation. There were two reasons for this: a) the lowest energy and most stable conformers showed over space correlations between H-3" and H-2 /6 , while this is the connection only observed in the case of compound 15, and b) low-energy conformers of 2"R,3"S,2S,3R showed ax-eq configuration between H-2 and H-3, which were not in agreement with observed coupling constants. The DP4+ calculations showed that 2"R,3"S,2R,3S with a probability of 99.19% was the correct isomer ( Figure S25). The calculated MAE and CMAE values were lower for the correct isomer and further confirmed by the higher correlation coefficient value for that one. Overall, the results obtained were compatible with the result of DP4+ probability calculations (Table 4). Therefore, the relative configuration of 26 was established as 2"R*,3"S*,2R*,3S*. Furthermore, to determine the absolute configuration of 26, optimized conformers were subjected to ECD spectra simulations using TD-DFT/cam-B3LYP/6-31G(d,p)/IEFPCM/methanol. A comparison of the Boltzmann-averaged spectrum with one obtained experimentally, after applying UV correction of +23 nm and half band of 0.25 eV, resulted in the absolute configuration of 26 determined as 2"R,3"S,2R,3S (Figure 7). This outcome coincided with the structure of 3,2 -epi-larixinol, a compound isolated from the aerial parts of Abies chensiensis [19], but differing in both NMR chemical shifts and optical rotation compared to 26. However, the authors did not imply the AC of the aforementioned compound and relied solely on NOEs in determining the relative configuration; thus, our findings still hold true. . This outcome coincided with the structure of 3,2′-epi-larixinol, a compound isolated from the aerial parts of Abies chensiensis [19], but differing in both NMR chemical shifts and optical rotation compared to 26. However, the authors did not imply the AC of the aforementioned compound and relied solely on NOEs in determining the relative configuration; thus, our findings still hold true. In this way, the structure of yuccalechin B was established as (2S,2′R,3R,3′S)-3′,4,5′,6tetrahydroxy-2,2′-bis(4-hydroxyphenyl)-3′,4′-dihydro-2H,2′H,8′H-spiro[benzofuran-3,9′-furo. [2,3h]chromen]-8′-one.  Yuccalechin C (29) was obtained as an off-white, amorphous solid exhibiting UV absorption maximum at 211 nm, with a specific rotation of [α] D 20 = +148. The negative HRESIMS spectra of 29 showed a deprotonated molecule at m/z 541.1121, and its molecular formula was determined as C 30 H 21 O 10 (calcd 541.1140), the same as 15 and 26. The planar structure of 29 was assigned to be identical to compounds 15 and 26 based on the analysis of 1D and 2D-NMR spectra. However, the 1 H and 13 C-NMR spectra presented certain differences when compared to yuccalechin B (26), namely signals of upfield-shifted H-2 (δ 4.19, 1H, d, J = 9.5 Hz), H-3 (δ 3.85, 1H, ddd, J = 9.8, 9.5, 5.8 Hz), and H-3" (δ 5.72, 1H, s) and downfield-shifted C-3 (δ 30.5), C-1" (δ 180.6), and C-3" (δ 94.5). These changes, along with a difference in chromatographic behavior (RT = 8.93 min vs. 8.39 min for 26) and optical rotation, indicated that compound 29 was a stereoisomer of 15 and 26. Similarly to 26, protons H-2 and H-3 of the C-ring were in 2,3-trans-configuration (afzelechin-type), but with an energetically favorable di-axial orientation, as indicated by a large 3 J H-H coupling constant, small 1 J CH coupling constants, relatively large 2 J C3-H2 and 2 J C2-H3 (Table 3) , and NOEs observed between H-2 /H-6 and H-3 and between H-2 and H-4β ( Figure 5). The relative configuration of the spiro-center was apparent from the chemical shift of C-1" (δ 180.6) and 1 J CH for C-3" (150.8 Hz), and as a result, the relative configuration between C-2"/C-3" of the F-ring was deduced as trans. Subsequently, to prove the relative configuration, two possible isomers 2"R,3"R,2S,3R and 2"R,3"R,2R,3S were subjected to an NMR calculation followed by a DP4+ probability calculation. Results implied that the 2"R,3"R,2S,3R isomer was the correct isomer with a probability value of 100% ( Figure S35). Although the MAE and CMAE values were lower and in favor of the incorrect isomer, the correlation coefficient was higher for correct assignment (∆r = 0.00795). Additionally, the ECD calculation of optimized structures ( Figure S34) was performed using TD-DFT/cam-B3LYP/6-31G(d,p)/CPCM/methanol (CPCM = conductor-like polarizable continuum model) and resulted in establishing the absolute configuration of 29 as 2"S,3"S,2R,3S (Figure 8).

Molecular Docking Simulations of Yuccalechins B and C
To explore possible binding interactions of 26 and 29, molecular docking simulations were carried out on hAChE (PDB: 4EY7) using a Glide module implemented in the Schrödinger Small-Molecule Drug Discovery Suite. The results revealed that 26 and 29 exhibited binding energies of -8.43 and -7.76 kcal mol −1 , respectively, against AChE, which were modulated by hydrogen bonds and π-π stacking contacts inside the active site.
According to molecular docking results, the orientation of 26 was driven by the interactions with the peripheral anionic site (PAS) comprising residues. Compound 26 was located in the bottleneck of the active site gorge via π-π stacking contacts with Tyr341 and forming a hydrogen bond with Tyr72 at the entrance to the gorge through the 5-positioned hydroxyl group of the A-ring. The oxygen atom of the F-ring pointed toward the acyl binding site and was found to interact with the Phe295 backbone via hydrogen bonding. Moreover, the π-π stacking contact between the E-ring and the oxyanion hole residue Phe338 stabilized the occupation of 26 in the binding site ( Figure 9). The observation for 26 was found to be in good agreement with the data reported, indicating the aromatic residues at the PAS as a major site of interaction for polycyclic aromatic compounds [84]. via hydrogen bonding. Moreover, the π-π stacking contact between the E-ring and the oxyanion hole residue Phe338 stabilized the occupation of 26 in the binding site ( Figure 9). The observation for 26 was found to be in good agreement with the data reported, indicating the aromatic residues at the PAS as a major site of interaction for polycyclic aromatic compounds [84]. As depicted in Figure 10, 29 was positioned deeply in the enzyme gorge occupying the region between the oxyanion hole and the acyl binding pocket, which was close to catalytic triad residues. The E-ring displayed interactions with His447 and Phe338 through hydrogen bonding and π-π stacking contacts, respectively, which were found to be responsible for stabilization of 29 at the binding site. As depicted in Figure 10, 29 was positioned deeply in the enzyme gorge occupying the region between the oxyanion hole and the acyl binding pocket, which was close to catalytic triad residues. The E-ring displayed interactions with His447 and Phe338 through hydrogen bonding and π-π stacking contacts, respectively, which were found to be responsible for stabilization of 29 at the binding site.

Discussion
The multistep purification of Yucca schidigera bark led to the isolation of three new spirobiflavonoids and confirmed the presence of numerous phenolic compounds, among which aromadendrin, naringenin, yuccalide A, and gloriosaols A and C-E are reported for the first time in this plant. Structures of isolated compounds were elucidated using various spectroscopic methods, including HRESIMS, UV, and ECD spectroscopy and optical rotations. For the new compounds, the relative configuration was established based on NMR chemical shifts, H-H and C-H coupling constants, DP4+ probability calculations, and NOE effects observed in the 2D-ROESY NMR spectra. Here, we report for the first time the usage of 1 JCH, 2 JCH, and 3 JCH coupling constants in the determination of relative stereochemistry of flavan-3-ols and spirobiflavonoids. Additionally, the absolute configuration of chiral spirobiflavonoids has been described for the first time using ab initio calculations of ECD spectra. The identification of stereochemistry of such compounds reported so far was based on a comparison of the ECD spectra with larixinol (abiesinol E), which possesses the 2′′R,3′′R,2R,3R absolute configuration established by the X-ray crystallographic analysis [19], or chemical methods [20]. Our work, to the best of our knowledge, is the first report on the cholinesterase inhibitory activity of spirobiflavonoids. Tested compounds, yuccalechins B (26) and C (29), turned out very weak, but they were selective inhibitors of AChE.

Chemicals and Reagents
Methanol and chloroform as well as acetic acid, n-hexane, and ethyl acetate, all of analytical reagent grade, were purchased from Fisher Chemical (Loughborough, UK) and Merck (Darmstadt, Based on the results of the molecular docking simulations, the reduced number of favorable π-π stacking and hydrogen bonding contacts with key residues, which tend to increase the binding affinity, might be the explanation for the lower inhibitory capacity of these compounds against AChE.

Discussion
The multistep purification of Yucca schidigera bark led to the isolation of three new spirobiflavonoids and confirmed the presence of numerous phenolic compounds, among which aromadendrin, naringenin, yuccalide A, and gloriosaols A and C-E are reported for the first time in this plant. Structures of isolated compounds were elucidated using various spectroscopic methods, including HRESIMS, UV, and ECD spectroscopy and optical rotations. For the new compounds, the relative configuration was established based on NMR chemical shifts, H-H and C-H coupling constants, DP4+ probability calculations, and NOE effects observed in the 2D-ROESY NMR spectra. Here, we report for the first time the usage of 1 J CH , 2 J CH , and 3 J CH coupling constants in the determination of relative stereochemistry of flavan-3-ols and spirobiflavonoids. Additionally, the absolute configuration of chiral spirobiflavonoids has been described for the first time using ab initio calculations of ECD spectra. The identification of stereochemistry of such compounds reported so far was based on a comparison of the ECD spectra with larixinol (abiesinol E), which possesses the 2"R,3"R,2R,3R absolute configuration established by the X-ray crystallographic analysis [19], or chemical methods [20]. Our work, to the best of our knowledge, is the first report on the cholinesterase inhibitory activity of spirobiflavonoids. Tested compounds, yuccalechins B (26) and C (29), turned out very weak, but they were selective inhibitors of AChE.

Chemicals and Reagents
Methanol and chloroform as well as acetic acid, n-hexane, and ethyl acetate, all of analytical reagent grade, were purchased from Fisher Chemical (Loughborough, UK) and Merck (Darmstadt, Germany), respectively. Acetonitrile and methanol (LC-MS grade) were purchased from Merck (Darmstadt, Germany), while MS-grade formic acid and other chemicals were purchased from Sigma-Aldrich (Steinheim, Germany). Ultrapure water was prepared using a Milli-Q water purification system (Millipore, Milford, MA, USA).

Extraction and Isolation
The yucca bark was powdered using an Ultra Centrifugal Mill ZM 200 (Retsch, Germany) with 0.5 mm sieves, and then 202.8 g of the powder was extracted with 100% MeOH (3 L × 3, 1 day each) using an ultrasonic bath (Polsonic 33, Warsaw, Poland) at room temperature. All extraction, isolation, and separation procedures were performed in the dark to avoid any isomerization of compounds. Combined filtered solutions were concentrated under reduced pressure at 35 • C followed by dilution with water and defatted with n-hexane in a separating funnel. The solution obtained was evaporated to eliminate MeOH and subsequently was extracted with ethyl acetate to yield 12.  Prior to the biological activity assays, the purities of compounds 13, 26, and 29 were checked by NMR spectroscopy (electronic reference to access in vivo concentrations 2 (ERETIC2) method) [86] using D-(-)-quinic acid (C 7 H 12 O 6 , 98%, Sigma-Aldrich, St. Louis, MO, USA) in deuterated methanol (30 mM) as the reference sample. The purity was over 60% for all compounds tested.

High-Resolution LC-MS
The EtOAc fraction was subjected to high-resolution LC-MS analysis. Chromatographic separation was performed on a Thermo Scientific Ultimate 3000RS chromatographic system on a Waters BEH C18 column (150 × 2.1 mm i.d.; 1.7 µm, Milford, USA). The effluent was analyzed using a photodiode array detector (200-600 nm, 10 Hz acquisition frequency), Q-TOF MS (Bruker Impact II HD, Bruker, Billerica, MA, USA), and a charged aerosol detector (CAD, Thermo Corona Veo RS) as described in detail in our previous publication [87].

Optical Rotation [α]
Optical rotations were determined on a P-2000 polarimeter (Jasco, Easton, PA, USA) in MeOH solutions at concentrations of 1 mg mL −1 .

Electronic Circular Dichroism (ECD) Spectroscopy
Circular dichroism spectra of all analyzed compounds were made at room temperature on a Jasco J-1500 magnetic circular dichroism spectrometer (1 mm or 5 mm pathlength was used) in MeOH. CD spectra were collected at a scan rate of 100 nm min −1 with a response time of 1 s. Measurements were taken in the 200-500 nm range. All spectra were baseline corrected, and the final plot was taken from five accumulated plots. The concentrations of compounds were 50 mM. protein preparation wizard tool. Water molecules were removed from the crystallographic structure followed by the addition of hydrogen atoms. All atom charges and atom types were assigned. Finally, energy minimization and refinement of the structure was performed up to 0.3 Å RMSD by applying an OPLS3e force field. The centroid of the active site residues was defined as a grid box. Van der Waals (vdW) radius scaling factor 1.00, partial charge cutoff 0.25, and an OPLS3e force field were used for receptor grid generation. The compounds prepared by LigPrep were docked into AChE using the IFD protocol [91], which considers flexibility of both the compounds and receptor. Residues Asp74, Trp86, Tyr124, Tyr133, Ser203, Trp286, Phe295, Phe297, Try337, Phe338, and His447 lining the binding site of AChE were kept as flexible. The initial docking protocol was set to employ a 0.50 vdW radius scaling factor, and the resulting top 20 poses of each compound were taken. An extra-precision (XP) algorithm was employed in redocking of the compounds, with the low energy refined structures generated by the Prime MM-GBSA (molecular mechanics -generalized Born surface area) method. The best conformation for each compound was chosen based on the lowest XP glide score.