Discovery of New Boswellic Acid Hybrid 1H-1,2,3-Triazoles for Diabetic Management: In Vitro and In Silico Studies

A series of 24 new 1H-1,2,3-triazole hybrids of 3-O-acetyl-11-keto-β-boswellic acid (β-AKBA (1)) and 11-keto-β-boswellic acid (β-KBA (2)) was designed and synthesized by employing “click” chemistry in a highly efficient manner. The 1,3-dipolar cycloaddition reaction between β-AKBA-propargyl ester intermediate 3 or β-KBA-propargyl ester intermediate 4 with substituted aromatic azides 5a–5k in the presence of copper iodide (CuI) and Hünig’s base furnished the desired products—1H-1,2,3-triazole hybrids of β-AKBA (6a–6k) and β-KBA (7a–7k)—in high yields. All new synthesized compounds were characterized by 1H-, 13C-NMR spectroscopy, and HR-ESI-MS spectrometry. Furthermore, their α-glucosidase-inhibitory activity was evaluated in vitro. Interestingly, the results obtained from the α-glucosidase-inhibitory assay revealed that all the synthesized derivatives are highly potent inhibitors, with IC50 values ranging from 0.22 to 5.32 µM. Among all the compounds, 6f, 7h, 6j, 6h, 6g, 6c, 6k, 7g, and 7k exhibited exceptional inhibitory potency and were found to be several times more potent than the parent compounds 1 and 2, as well as standard acarbose. Kinetic studies of compounds 6g and 7h exhibited competitive and mixed types of inhibition, with ki values of 0.84 ± 0.007 and 1.18 ± 0.0012 µM, respectively. Molecular docking was carried out to investigate the binding modes of these compounds with α-glucosidase. The molecular docking interactions indicated that that all compounds are well fitted in the active site of α-glucosidase, where His280, Gln279, Asp215, His351, Arg442, and Arg315 mainly stabilize the binding of these compounds. The current study demonstrates the usefulness of incorporating a 1H-1,2,3-triazole moiety into the medicinally fascinating boswellic acids skeleton.


Introduction
Frankincense (oleo-gum exudates of Boswellia ssp.) is the most precious resin religiously, economically, socially, and medicinally. The distribution of the Boswellia trees ranges from South Arabia to India to the horn of Africa [1]. Frankincense exists in traditional Chinese medicine to treat ulcers, rheumatoid arthritis, dysmenorrhea, swelling, osteoarthritis, pain from injuries, and amenorrhea [2][3][4]. B. sacra is endemic to Oman, and the resin (olibanum or frankincense) of B. serrata is known for its superb anti-inflammatory and pro-apoptotic activity [5,6]. Frankincense and triterpenoids from B. sacra and other species of the same genus have been used in traditional medicine since ancient times to treat obesity and lipid disorders [4,7].
In the drug discovery process, the incorporation of an active pharmacophoric moiety into a core bioactive natural product is a key step in exploring a wide range of new chemical entities with enhanced biological activities. With this objective in our mind, and in continuation of our research interest involving the chemical transformation of compounds 1 and 2 through the molecular hybridization approach, we herein describe the synthesis of novel β-AKBA (1)-and β-KBA (2)-based 1H-1,2,3-triazole hybrids and molecular docking studies with respect to their α-glucosidase activities.
Hyperglycemia is one of the most serious diabetes outcomes, leading to long-term health issues such as cardiovascular diseases, metastatic cancer, retinopathy, and nephropathy [29]. According to a worldwide projection, factors such as population growth, aging, urbanization, and lifestyle may cause the global prevalence rate of diabetes to rise to 54 percent by 2030 [30]. α-glucosidase is a hydrolase enzyme which is present on the brush border of the small intestine and catalyzes carbohydrates digestion into glucose molecules, which are then absorbed into blood. The use of α-glucosidase inhibitors has been shown to be a reliable approach for drug development and to overcome hyperglycemia. As a result, α-glucosidase inhibitors (AGIs) have received a lot of interest due to their therapeutic usage in the treatment of nocturnal hypoglycemia [31]. A In addition, the importance of 1H-1,2,3-triazole molecules in pharmaceuticals and agrochemicals has increased [20]. Triazole compounds play a key role in organic chemistry due to their multiple and broad range of applications in bio-medicinal, biochemical, and material sciences [21]. The interest of the compounds containing the triazole moiety underwent a substantial growth over the past few decades. Furthermore, heterocycles containing 1,2,3-triazole scaffolds are known to have biological activities including antimicrobial [22], anti-bacterial [23], anti-viral [24], anti-HIV [25], anti-inflammatory [26], and anti-cancer [27] activities. Our group have recently reported the synthetic modifications of the β-AKBA and β-ABA hybrid triazole moieties as potent carbonic anhydrase-II [28].
In the drug discovery process, the incorporation of an active pharmacophoric moiety into a core bioactive natural product is a key step in exploring a wide range of new chemical entities with enhanced biological activities. With this objective in our mind, and in continuation of our research interest involving the chemical transformation of compounds 1 and 2 through the molecular hybridization approach, we herein describe the synthesis of novel β-AKBA (1)-and β-KBA (2)-based 1H-1,2,3-triazole hybrids and molecular docking studies with respect to their α-glucosidase activities.
Hyperglycemia is one of the most serious diabetes outcomes, leading to long-term health issues such as cardiovascular diseases, metastatic cancer, retinopathy, and nephropathy [29]. According to a worldwide projection, factors such as population growth, aging, urbanization, and lifestyle may cause the global prevalence rate of diabetes to rise to 54 percent by 2030 [30]. α-glucosidase is a hydrolase enzyme which is present on the brush border of the small intestine and catalyzes carbohydrates digestion into glucose molecules, which are then absorbed into blood. The use of α-glucosidase inhibitors has been shown to be a reliable approach for drug development and to overcome hyperglycemia. As a result, α-glucosidase inhibitors (AGIs) have received a lot of interest due to their therapeutic usage in the treatment of nocturnal hypoglycemia [31]. A fundamental treatment method used to overcome hyperglycemia is to reduce carbohydrate digestion and absorption, which provides an insight into the important role of α-glucosidase. As a result, novel AGIs with fewer side effects are urgently needed, in contrast to the currently existing clinical AGIs, such as acarbose, voglibose, and miglitol [32,33]. Recently, it was reported in the literature that both 1H-1,2,3-triazole and BAs are active pharmacophores of α-glucosidase inhibitors [34][35][36]. In the current study, we conjugated the 1H-1,2,3-triazole with BAs to discover their hybrid behavior against α-glucosidase.
Keeping in mind the importance of boswellic acids, triazole moieties, and AGIs, we planned to synthesize their hybrids between BAs and triazoles, and screened them against the α-glucosidase enzyme. Furthermore, this is the first report, to the best of authors' knowledge, on the synthetic hybrids of 1H-1,2,3-triazole with β-AKBA (1) and β-KBA (2). In addition, here, we report the crystal structure of compound 4 for the first time. Additionally, some structure-activity relationships, molecular docking, and kinetic studies of the active analogs are also discussed.
The structure of compound 4 was unambiguously confirmed using X-ray crystallography. The molecular structure of compound 4 is illustrated in Figure 2a, while the crystal data are depicted in Table S2 (Supporting Information). Compound 4 crystallizes in the orthorhombic space group P2 1 P2 1 P2 1 . All cyclohexane rings of the pentacyclic backbone adopt chair conformations, with the exception of a ring containing sp 2 -hybdridized Catoms (C12-C13). C9 and C11-C14 form an approximately planar region, with C8 out of the plane, resulting in a sofa or a half chair conformation of the ring. The doublebond character of C12-C13 was confirmed by their distance of 1.343 (4) Å. The bond distance between C3-O3 was 1.431 Å, which confirms their single-bond character. The bond distance between C11-O4 was 1.226 (3) Å, which suggests a double-bond character. The C32-C33 bond distance in the propargilate fragment was 1.164 (3), which confirms the triple-bond character. Compound 4 is stabilized by intramolecular hydrogen bonding. The O-H group forms a hydrogen bond with the oxygen (O4) of the neighboring molecule (d O3-H1· · · O4 = 2.08 (3) Å, θ O3-H1· · · O4 = 172 (4) o ), resulting in a hydrogen-bonded one-dimensional chain along c-axis ( Figure 2b). hibiting signals at δ 174.7, 76.8, 75.0, and 51.8, corresponding to the ester (-COOR), acetylene (C≡CH) and oxymethelene (-OCH2-) groups. Further confirmation of compound 3 was conducted by HRMS (ESI), which showed protonated molecular ion at (m/z) 551.3713 [M+H] + . The 1 H NMR spectrum of compound 4 showed prominent signals at δ 2.48 and 4.66, corresponding to acetylenic (-C≡CH) and oxymethelene (-OCH2-) protons, respectively. The 13 C NMR spectrum of compound 4 confirmed the above observations by exhibiting signals at δ 167.1, 77.4, 74.7, and 51.5, corresponding to the ester (-COOR), acetylene (-C≡CH), and oxymethelene (-OCH2-) functionalities. Further confirmation of compound 4 was achieved by HRMS (ESI), which showed protonated molecular ion at (m/z) 509.3609 The structure of compound 4 was unambiguously confirmed using X-ray crystallography. The molecular structure of compound 4 is illustrated in Figure 2a, while the crystal data are depicted in Table S2 (Supporting Information). Compound 4 crystallizes in the orthorhombic space group P21P21P21. All cyclohexane rings of the pentacyclic backbone adopt chair conformations, with the exception of a ring containing sp 2 -hybdridized Catoms (C12-C13). C9 and C11-C14 form an approximately planar region, with C8 out of the plane, resulting in a sofa or a half chair conformation of the ring. The double-bond character of C12-C13 was confirmed by their distance of 1.343 (4) Å. The bond distance between C3-O3 was 1.431 Å, which confirms their single-bond character. The bond distance between C11-O4 was 1.226 (3) Å, which suggests a double-bond character. The C32-C33 bond distance in the propargilate fragment was 1.164 (3), which confirms the triplebond character.  The structures of all the synthesized compounds (6a-6k and 7a-7k) were characterized by spectral data analysis ( 1 H and 13 C NMR, and HRMS). 19 F NMR spectroscopy was used for the compounds containing fluorine. The formation of 1H-1,2,3-triazole hybrids of β-AKBA (6a-6k) and β-KBA (7a-7k) derivatives was confirmed by the presence of a characteristic singlet observed in their corresponding 1 H NMR spectra at δ 7.60-8.30, which can be attributed to the H-5 proton of the triazole ring. Further, the 13 C NMR spectra of the synthesized hybrids (6a-6k and 7a-7k) determined the diagnostic triazole ring carbon signals between δ 123~125 and 142~144. The structures of the synthesized 1H-1,2,3-triazole hybrids (6a-6k and 7a-7k) were further supported by HRMS.

α-Glucosidase Activity
The synthesized derivatives of boswellic acids were evaluated for their α-glucosidaseinhibitory potential. Consequently, all the compounds exhibited significant inhibitory activities in the range of 0.22-5.32 µM, at a level even superior to the standard inhibitor acarbose (IC 50 = 942 ± 0.74 µM). Compounds 3 and 4, containing the same propargylic groups, displayed slight variations in their α-glucosidase-inhibitory activity after the addition of the triazole moiety with different substituents. Compound 3 with the propargylic group displayed a potent inhibitory activity, with an IC 50 value of 4.72 ± 0.52 µM, while the addition of the triazole moiety in 6a was favorable in order to enhance the inhibitory activity (IC 50 = 2.02 ± 0.031 µM). However, the substitution of the ortho-methyl group in 6b slightly decreased the inhibitory activity of 6b (IC 50 = 2.25 ± 0.10 µM), as compared to 6a with no substitution. However, the substitution of the methoxy group at the orthoposition in 6c (IC 50 = 1.23 µM) and 6e (IC 50 = 2.01 µM) showed only a modest impact on the inhibition, as compared to substitution at the para-position. The substitution of the tri-fluoro methyl group at the meta-position of 6f (IC 50 = 0.22 µM) remarkably increased the inhibitory activity, more so than substitutions at the ortho-and para-positions of compounds 6d (IC 50 = 2.37 µM) and 6k (IC 50 = 1.26 µM). Moreover, a minute variation was observed in the inhibitory potential by the substitution of the bromo phenyl group in compounds 6g and 6h, with IC 50 values of 0.78 and 0.69 µM, respectively. In compound 6j, the addition of a highly electronegative fluorine atom at the meta-position displayed significant inhibitory potential, with an IC 50 value of 0.58 µM, as compared to the replacement of the chlorine atom at the same position in 6i (IC 50 = 3.06 µM) ( Table 2).
Compound 4 (IC 50 value of 5.32 ± 0.15 µM) with the propargylic group displayed the least α-glucosidase-inhibitory potential among all the compounds. However, the addition of triazole group, along with various substitutions at different positions, improved the inhibitory potential of compound 4 derivatives. The phenyl-substituted derivative of 4 (7a) exhibited a higher level of in vitro inhibitory activity (IC 50 = 2.31 µM) than compound 4. Similarly, 7b with the methyl substitution at the meta-position of the phenyl ring also displayed a potent inhibitory activity (IC 50 = 2.11 µM), while the substitution of the tri-fluoro methyl group at the meta-and ortho-positions of 7d and 7f, respectively, produced only a slight difference in the inhibitory potential, with IC 50 values of 2.62 and 2.52 µM, respectively, in contrast to the addition of the same group at the para-position in 7k, which displayed an increase in the inhibitory activity (IC 50 = 1.39 µM). Compound 7c with a meta-methoxysubstituent displayed a slight increase in inhibitory activity (IC 50 = 2.18 µM), as compared to 7e (IC 50 = 2.37 µM) with a similar substituent at the para-position. Compound 7h with para-bromo substitution displayed more potent inhibitory activity (IC 50 = 0.43 µM) than 7g (IC 50 = 1.34 µM) with the same substitution at the ortho-position. In contrast, compound 7j with para-fluoro substitution exhibited more potent inhibitory activity (IC 50 = 2.64 µM) than 7i with a chloro-substituent at the same para-position (IC 50 = 3.18 µM). Comparing both series (6a-6k and 7a-7k), four compounds of AKBA derivatives (6c, 6f, 6g, and 6j) displayed a higher level of inhibition than KBA derivatives (7c, 7f, 7g, and 7j), compared to the standard, which showed the importance of the acetate group over hydroxyl, while the rest of the compounds determined an almost similar inhibition. Only one compound, 7h (KBA derivative), showed a higher level of inhibin than 6j (AKBA derivative), which could possibly be due to the substituted group attached to the triazoles moiety. compared to the standard, which showed the importance of the acetate group over hydroxyl, while the rest of the compounds determined an almost similar inhibition. Only one compound, 7h (KBA derivative), showed a higher level of inhibin than 6j (AKBA derivative), which could possibly be due to the substituted group attached to the triazoles moiety. compared to the standard, which showed the importance of the acetate group over hydroxyl, while the rest of the compounds determined an almost similar inhibition. Only one compound, 7h (KBA derivative), showed a higher level of inhibin than 6j (AKBA derivative), which could possibly be due to the substituted group attached to the triazoles moiety. compared to the standard, which showed the importance of the acetate group over hydroxyl, while the rest of the compounds determined an almost similar inhibition. Only one compound, 7h (KBA derivative), showed a higher level of inhibin than 6j (AKBA derivative), which could possibly be due to the substituted group attached to the triazoles moiety. compared to the standard, which showed the importance of the acetate group over hydroxyl, while the rest of the compounds determined an almost similar inhibition. Only one compound, 7h (KBA derivative), showed a higher level of inhibin than 6j (AKBA derivative), which could possibly be due to the substituted group attached to the triazoles moiety. compared to the standard, which showed the importance of the acetate group over hydroxyl, while the rest of the compounds determined an almost similar inhibition. Only one compound, 7h (KBA derivative), showed a higher level of inhibin than 6j (AKBA derivative), which could possibly be due to the substituted group attached to the triazoles moiety. compared to the standard, which showed the importance of the acetate group over hydroxyl, while the rest of the compounds determined an almost similar inhibition. Only one compound, 7h (KBA derivative), showed a higher level of inhibin than 6j (AKBA derivative), which could possibly be due to the substituted group attached to the triazoles moiety. compared to the standard, which showed the importance of the acetate group over hydroxyl, while the rest of the compounds determined an almost similar inhibition. Only one compound, 7h (KBA derivative), showed a higher level of inhibin than 6j (AKBA derivative), which could possibly be due to the substituted group attached to the triazoles moiety. compared to the standard, which showed the importance of the acetate group over hydroxyl, while the rest of the compounds determined an almost similar inhibition. Only one compound, 7h (KBA derivative), showed a higher level of inhibin than 6j (AKBA derivative), which could possibly be due to the substituted group attached to the triazoles moiety. compared to the standard, which showed the importance of the acetate group over hydroxyl, while the rest of the compounds determined an almost similar inhibition. Only one compound, 7h (KBA derivative), showed a higher level of inhibin than 6j (AKBA derivative), which could possibly be due to the substituted group attached to the triazoles moiety. compared to the standard, which showed the importance of the acetate group over hydroxyl, while the rest of the compounds determined an almost similar inhibition. Only one compound, 7h (KBA derivative), showed a higher level of inhibin than 6j (AKBA derivative), which could possibly be due to the substituted group attached to the triazoles moiety. compared to the standard, which showed the importance of the acetate group over hydroxyl, while the rest of the compounds determined an almost similar inhibition. Only one compound, 7h (KBA derivative), showed a higher level of inhibin than 6j (AKBA derivative), which could possibly be due to the substituted group attached to the triazoles moiety. compared to the standard, which showed the importance of the acetate group over hydroxyl, while the rest of the compounds determined an almost similar inhibition. Only one compound, 7h (KBA derivative), showed a higher level of inhibin than 6j (AKBA derivative), which could possibly be due to the substituted group attached to the triazoles moiety. Kinetic Studies To investigate their inhibitory mechanisms, the most potent compounds 6f and 7h were subjected to kinetic studies. Compound 6f exhibited a competitive type of inhibition, with a ki value of 0.84 ± 0.007 µM (Figure 3). In this type of inhibition, the inhibitor binds with the active site residue of enzyme and, therefore, Vmax remains the same, with an increase in the Km. The kinetic studies of compound 7h revealed a mixed type of inhibition, with a ki value of 1.18 ± 0.0012 µM (Figure 4). In this type of inhibition, both inhibitors-Vmax and Km-are affected by Vmax decreases and Km increases. To investigate their inhibitory mechanisms, the most potent compounds 6f and 7h were subjected to kinetic studies. Compound 6f exhibited a competitive type of inhibition, with a ki value of 0.84 ± 0.007 μM (Figure 3). In this type of inhibition, the inhibitor binds with the active site residue of enzyme and, therefore, Vmax remains the same, with an increase in the Km. The kinetic studies of compound 7h revealed a mixed type of inhibition, with a ki value of 1.18 ± 0.0012 μM (Figure 4). In this type of inhibition, both inhibitors-Vmax and Km-are affected by Vmax decreases and Km increases.

Molecular Docking Studies
Initially, the docking protocol was validated through the re-docking of the competitive inhibitor: maltose in the active site of the enzyme (PDB code: 3A4A). The re-docking results showed that the maltose was docked at its binding site with RMSD = 1.31Å ( Figure  S1, Supplementary Materials) and a docking score of −5.94 kcal/mol, which indicated that the docking method was reliable to use to predict the binding modes of our compounds. To investigate their inhibitory mechanisms, the most potent compounds 6f and 7h were subjected to kinetic studies. Compound 6f exhibited a competitive type of inhibition, with a ki value of 0.84 ± 0.007 μM (Figure 3). In this type of inhibition, the inhibitor binds with the active site residue of enzyme and, therefore, Vmax remains the same, with an increase in the Km. The kinetic studies of compound 7h revealed a mixed type of inhibition, with a ki value of 1.18 ± 0.0012 μM (Figure 4). In this type of inhibition, both inhibitors-Vmax and Km-are affected by Vmax decreases and Km increases.

Molecular Docking Studies
Initially, the docking protocol was validated through the re-docking of the competitive inhibitor: maltose in the active site of the enzyme (PDB code: 3A4A). The re-docking results showed that the maltose was docked at its binding site with RMSD = 1.31Å ( Figure  S1, Supplementary Materials) and a docking score of −5.94 kcal/mol, which indicated that the docking method was reliable to use to predict the binding modes of our compounds.

Molecular Docking Studies
Initially, the docking protocol was validated through the re-docking of the competitive inhibitor: maltose in the active site of the enzyme (PDB code: 3A4A). The redocking results showed that the maltose was docked at its binding site with RMSD = 1.31Å ( Figure S1, Supplementary Materials) and a docking score of −5.94 kcal/mol, which in-Pharmaceuticals 2023, 16, 229 9 of 20 dicated that the docking method was reliable to use to predict the binding modes of our compounds.
To investigate the binding mode of the compounds in the active site of α-glucosidase, we carried out a molecular docking experiment. In the docking studies, it was observed that the propargyl-linked carbonyl moiety of 3 fit deep inside the active site and formed a hydrogen bond (H-bond) with the side chain of Arg442. In contrast, the carbonyl group of 4 did not interact with the surrounding residues; instead, its -OH group formed an H-bond with the side chain of Arg442. The boswellic acid skeleton of both of the compounds fit at the entrance of the active site and blocked the access to substrate in the active site. The binding modes of compounds 3 and 4 derivatives suggest that the addition of the triazole-substituted phenyl ring dragged the boswellic acid moieties of the compounds towards the entrance, while the substituted groups fit deep inside the active site.
The most active compound, 6f, mediated multiple H-bonds within the active site. The triazole-linked carbonyl oxygen of 6f formed an H-bond with His280, while the acetate moiety of 6f mediated a bidentate interaction with Arg315. Additionally, the side chain of Ser240 donated a H-bond to the carbonyl group in the boswellic acid. The binding modes of 7h and 6j reflect that the carbonyl and -OH moieties of 7h mediated H-bonding with Arg315 and Ser157, respectively, whereas the acetate and the carbonyl groups of 6j formed a bidentate interaction with Arg315 and Ser240, respectively. Similarly, the carbonyl moiety of 6h formed an H-bond with Arg315, while its triazole ring interacted with Arg442 through the H-bond. The carbonyl groups of 6g mediated H-bonds with His280 and Ser240. The side chains of Ser240 and Gln279 bound with the boswellic acid carbonyl oxygen and the phenyl-substituted methoxy group of 6c, respectively, while His280 bound with the triazole-substituted oxygen of 6c and carbonyl oxygen of 6k. Moreover, the -OH and the triazole-substituted oxygen of 7g formed H-bonds with the side chains of Glu411 and Gln279, respectively. The acetate moiety of 7k and 6e was linked with the side chain of Gln279.
Furthermore, the side chains of His351, Arg442, and Asp242 formed H-bonds with the triazole ring of 6a and the -OH of 7b and 7c, respectively. Meanwhile, 6b formed H-bonds with the side chain of Ser240 and His280 and -OH of 7a bind with Arg442. The carbonyl oxygen and fluorine atom of 6d formed an H-bond with the side chains of Gln279 and His351, respectively. Similarly, Ser240 provided an H-bond to the carbonyl oxygen of 7e. The binding modes of 7f and 7d showed that 7f formed H-bonds with the side chain of Gln279 and His112, while 7d only mediated an H-bond with the side chain of Arg213. The docked view of 7j suggested that the compound fit deep inside the active site and mediated multiple H-bonds with His351, Asp215, and Gln279. Similarly, the acetate, triazole, and carbonyl groups of 6i formed several H-bonds with Arg442, His351, Asn350, and Arg315. The docked conformation of 7i was found to be more surface-exposed, and formed only an H-bond with the Tyr158. Afterwards, compounds 2 and 1 exhibited the least inhibitory activities, compared to other compounds in this series.
The acetate and -OH groups of 2 mediated H-bonds with the side chains of Arg442 and Asp215, respectively. Meanwhile, the acetate moiety of 1 interacted with the side chains of Arg213 and His351. The binding modes of these compounds indicate that His280, Gln279, Asp215, His351, Arg442, and Arg315 are the important residues that provide H-bonds to these compounds, while Phe301 and Phe303 provide π-π interaction to most of the compound at the entrance of the active site. The binding mode of the most active hit is shown in Figure 5. The docking score of the compounds were in the range of −9.18 to −2.23 Kcal/mol. The atom-based protein-ligand interaction of each compound is tabulated in Table S1 (Supplementary Materials), along with their docking scores. The docking results support our experimental finding.

General
All experiments were carried out in dry reaction vessels under a dry nitrogen atmosphere. All reagents were purchased from Sigma-Aldrich, Germany. Solvents were purified and dried according to the standard procedures. The high-resolution electrospray ionisation mass spectra (HR-ESI-MS) were recorded on an Agilent 6530 LC Q-TOF LC/MS (Agilent, country of origin USA/EU, made in Singapore). The 1 H and 13 C NMR spectra were recorded on a nuclear magnetic resonance (NMR) spectrometer (Bruker, Zürich, Switzerland) operating at 600 MHz (150 MHz for 13 C) using the solvent peak as an internal reference (CDCl3, δ H: 7.26; δ C: 77.0). The 19 F NMR spectra were recorded at 564 MHz. Data were reported in the following order: chemical shift (δ) in ppm, multiplicities, and coupling constants (J) in Hertz (Hz). The column chromatography was carried out using silica gel of the selected particle size of 100-200 mesh. All reactions were monitored by thin layer chromatography (TLC) using silica gel F254 pre-coated plates (Merck, Darmstadt, Germany). Visualization was accomplished using UV-light and I2 stain. The solvents for column chromatography (EtOAc, n-hexane) were technical grade and distilled prior to use. The organic extracts were dried over anhydrous MgSO4. Single crystals of 4 were mounted on a MiTeGen loop with grease and examined on a Bruker D8 Venture APEX diffractometer equipped with a Photon 100 CCD area detector at 296 (2) K using graphitemonochromated Mo-Kα radiation (λ = 0.71073 Å). Data were collected using the APEX-II software [42], integrated using SAINT [43], and corrected for absorption using a multiscan approach (SADABS) [44]. The final cell constant was determined from full least squares refinement of all observed reflections. The structure was solved using intrinsic phasing (SHELXT) [45]. All non-H atoms were in subsequent difference maps and refined anisotropically. H-atoms were added at calculated positions and refined with a riding model. H atoms on O and N atoms were in a difference map and refined with constrained O-H and N-H distances. The structure of 4 was deposited with the CCDC (CCDC deposition number = 2,151,126).

Preparation of Boswellic Acids (BAs) Cluster
The air-dried ground material (100 g) of B. sacra resin was exhaustively extracted with 100% MeOH at room temperature (three times). The extract was evaporated to yield the yellowish residue (70 g). The MeOH extract was subjected to column chromatography

General
All experiments were carried out in dry reaction vessels under a dry nitrogen atmosphere. All reagents were purchased from Sigma-Aldrich, Germany. Solvents were purified and dried according to the standard procedures. The high-resolution electrospray ionisation mass spectra (HR-ESI-MS) were recorded on an Agilent 6530 LC Q-TOF LC/MS (Agilent, country of origin USA/EU, made in Singapore). The 1 H and 13 C NMR spectra were recorded on a nuclear magnetic resonance (NMR) spectrometer (Bruker, Zürich, Switzerland) operating at 600 MHz (150 MHz for 13 C) using the solvent peak as an internal reference (CDCl 3 , δ H: 7.26; δ C: 77.0). The 19 F NMR spectra were recorded at 564 MHz. Data were reported in the following order: chemical shift (δ) in ppm, multiplicities, and coupling constants (J) in Hertz (Hz). The column chromatography was carried out using silica gel of the selected particle size of 100-200 mesh. All reactions were monitored by thin layer chromatography (TLC) using silica gel F 254 pre-coated plates (Merck, Darmstadt, Germany). Visualization was accomplished using UV-light and I 2 stain. The solvents for column chromatography (EtOAc, n-hexane) were technical grade and distilled prior to use. The organic extracts were dried over anhydrous MgSO 4 . Single crystals of 4 were mounted on a MiTeGen loop with grease and examined on a Bruker D8 Venture APEX diffractometer equipped with a Photon 100 CCD area detector at 296 (2) K using graphite-monochromated Mo-K α radiation (λ = 0.71073 Å). Data were collected using the APEX-II software [42], integrated using SAINT [43], and corrected for absorption using a multi-scan approach (SADABS) [44]. The final cell constant was determined from full least squares refinement of all observed reflections. The structure was solved using intrinsic phasing (SHELXT) [45]. All non-H atoms were in subsequent difference maps and refined anisotropically. H-atoms were added at calculated positions and refined with a riding model. H atoms on O and N atoms were in a difference map and refined with constrained O-H and N-H distances. The structure of 4 was deposited with the CCDC (CCDC deposition number = 2,151,126).