Sterols and triterpenoids as potential anti-inflammatories: Molecular docking studies for binding to some enzymes involved in inflammatory pathways

Sterols and triterpenoids as potential anti-inflammatories: Molecular docking studies for binding to some enzymes involved in inflammatory pathways

Accepted Manuscript Title: Sterols and triterpenoids as potential anti-inflammatories: Molecular docking studies for binding to some enzymes involved ...

572KB Sizes 0 Downloads 2 Views

Accepted Manuscript Title: Sterols and triterpenoids as potential anti-inflammatories: Molecular docking studies for binding to some enzymes involved in inflammatory pathways Author: Marco A. Loza-Mej´ıa Juan Rodrigo Salazar PII: DOI: Reference:

S1093-3263(15)30043-7 http://dx.doi.org/doi:10.1016/j.jmgm.2015.08.010 JMG 6592

To appear in:

Journal of Molecular Graphics and Modelling

Received date: Revised date: Accepted date:

19-1-2015 24-8-2015 26-8-2015

Please cite this article as: Marco A.Loza-Mej´ia, Juan Rodrigo Salazar, Sterols and triterpenoids as potential anti-inflammatories: Molecular docking studies for binding to some enzymes involved in inflammatory pathways, Journal of Molecular Graphics and Modelling http://dx.doi.org/10.1016/j.jmgm.2015.08.010 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Title. Sterols and triterpenoids as potential anti-inflammatories: Molecular docking studies for binding to some enzymes involved in inflammatory pathways

Author names and affiliations. Marco A. Loza-Mejía1, Juan Rodrigo Salazar1 1

Facultad de Ciencias Químicas, Universidad La Salle. Benjamín Franklin 47, 06140, México City,

Mexico. E-mail: [email protected], [email protected] +52-(55)-5278-9500

• Corresponding author. Both authors should be considered as corresponding authors.

Abstract Triterpenes and sterols are good candidates for the development of anti-inflammatory drugs and use in chemoprevention or chemotherapy of cancer via the interaction with therapeutic targets related to inflammation, such as COX-1 and -2; LOX-5; MPO, PLA2 and i-NOS. In this study, we use molecular docking to evaluate the potential binding of a database of selected sterol and triterpenoid compounds with several skeletons against enzymes related to inflammation to propose structural requirements beneficial for anti-inflammatory activity that can be used for the design of more potent and selective anti-inflammatory and antitumor drugs. Our results suggest that the substitution pattern is important and that there is an important relationship between the class of sterol or triterpenoid skeleton and enzyme binding.

Highlights •

We performed docking studies of key triterpenoids to enzymes involved in inflammation.



We determined the relationship between the skeleton type and the inhibited enzyme.



Both skeleton and substitution pattern are important for enzyme binding.



Our results support the use of triterpenoids as potential leads for drug design.

Graphical abstract.

Keywords Molecular docking; triterpenoids; inflammation; structure-activity relationship; natural products 1. Introduction The relationship between chronic inflammation and cancer was proposed in 1860 when Virchow and colleagues observed that cancer initiated in areas or places subject to chronic inflammation. There is growing evidence that chronic and persistent inflammation in damaged tissues contributes to the promotion, progression and metastasis of tumors.1–4 Therefore, pro-inflammatory immune cells, cytokines, chemokines, enzymes, such as cyclooxygenase-2 (COX-2) or inducible nitric oxide synthase (i-NOS), their products, prostaglandins (e.g., PGE2) and the gaseous free radical nitric oxide (NO), and transcription factors, such as nuclear factor kappa B (NF-kB), increase in the tumor microenvironment.5 This non-balanced pro-inflammatory microenvironment enhances cell proliferation and increases the risk of developing tumors, which partially due to the release of proinflammatory mediators and growth factors as well as the presence of inflammatory cells and agents that can damage genetic material. It is well known that the immune system acts as a "double-edged sword" due to its ability to act as a tumor suppressor along with its ability to act as an initiator or a promoter in tumors.6 Several clinical observations support the strong association between chronic inflammation and cancer.6–8

Therefore, it has been hypothesized that an alternative to non-specific and highly toxic traditional treatments of cancer is the use of some of the specific components of the chronic inflammatory response as potential therapeutic targets to achieve chemoprevention or chemotherapy of cancer.9–15

Our experience16–19 as well as that of other groups20–25 indicates that some triterpenes and sterols are good candidates for producing active compounds that interfere with the inflammatory process. Therefore, in this study, we select a set of natural and semisynthetic metabolites to prepare candidate compounds for the development of drugs for use in chemoprevention or chemotherapy of cancer via interaction with therapeutic targets related to inflammation, such as COX-1 and -2; LOX5; MPO, PLA2 and i-NOS. These enzymes have been extensively studied, and the relationship between their participation in chronic inflammation and cancer development has been suggested.26– 28

However, a growing number of isolated and elucidated compounds from nature highlights the need for the use of computer-assisted techniques, which have emerged as promising tools to manage the enormous quantity of data related to the macromolecular targets and compounds and to extract new information from these data for the prediction or explanation of biological activities.29 Among these techniques, molecular docking predicts the preferred orientation of the ligand relative to the protein, which can be used to calculate the theoretical binding energy of the ligand to the protein. Recently, molecular docking has been used as a tool to explain the biological activity of some natural products (e.g., flavonoids as thrombin inhibitors,30 alkaloids and phenolic compounds with antileishmanial activity,31,32 preparations from traditional medicine as antitumor agents, and in the search of multi-target anti-inflammatories33). In this study, we use molecular docking to evaluate the potential binding of a database of selected triterpenoid compounds with several skeletons20 against enzymes related to inflammation to elucidate the structural requirements for anti-inflammatory activity that can be used for the design of more potent and selective anti-inflammatory and antitumor drugs. 2. Methodology

2.1. Ligand construction

All of the ligands were chosen from a review published by Akihisa and Yasukawa, which includes several triterpene compounds from different families with potential anti-inflammatory activity.20 All of the structures were constructed using Spartan ’10 for Windows34, and these geometries were optimized using the MMFF force field. A list of the studied molecules and their structures is available in the Supporting information.

2.2. Molecular docking studies The protein-ligand docking studies were carried out using Molegro Virtual Docker v. 6.0.135 based on the crystal structures of key enzymes that are important for the inflammatory process including COX-1 (PDB:1EQG36), COX-2 (PDB:3NT137), i-NOS (PDB:3E7G38), LOX-5 (PDB:3V9939), myeloperoxidase (PDB:3ZS040) and PDE4 (PDB:3SL441). All structures were retrieved from the Protein Data Bank.42 Co-crystallized inhibitors or substrates were present in all of the structures, and these sites were chosen as the center of the search area, to analyse the potential of sterols and triterpenoids as competitive inhibitors of the selected enzymes. Docking search area was sphere of 15 Å radius with a grid resolution of 0.30 Å. Protonation states and assignment of the charges in the proteins and ligands based on neutral pH were based on standard templates as part of the Molegro Virtual Docker program. As a test of the docking accuracy, co-crystallized ligands were re-docked into the protein structures. MolDock Optimizer was selected as docking protocol. Searching parameters were set to 5000 maximum iterations with a simplex evolution population size of 50 and a minimum of 25 runs for each ligand. The RMSD threshold for multiple cluster poses was set to <1.00 Å. After docking, a number of further scores were calculated including the binding affinity (MolDock Score) and re-ranking score (Rerank Score). The re-ranking score utilizes a more advanced scoring scheme than that used during docking and is often more useful for accurate ranking of the poses.35 The poses with the lowest Rerank score were selected for further analysis. To assess the efficacy of this procedure, crystal ligands were also docked to their respective enzymes, the top ranking score recorded, and the RMSD of that pose from the corresponding crystal co-ordinates computed. In all of the cases, the RMSD was lower than 2Å. Additionaly, to analyze the reproducibility of our results, docking studies with diferent starting crystallized enzyme-ligand complexes were carried out for the ligands with lower Rerank scores and for all ligands using a different starting COX-1:ligand complex. Results of these additional studies were consistent with the original docking studies. (See Supplemental information)

2.3. Analysis of the strongest docking poses

For each enzyme, the average Rerank Score for each family of triterpene compounds was calculated to determine the selectivity between a triterpenic nucleus and a particular enzyme. In order to analyze if there was selectivity of the sterol/triterpenoid skeleton for a particular enzyme, a onefactor ANOVA was carried out using Minitab 16.43 Interval plots showing a 95% confidence interval for the mean of each group can be consulted in the Supplemental Information. Then, the compounds of the family or families with the lowest average Rerank Score were analyzed to determine the additional structural requirements for enzyme binding. 3.

Results and discussion

Table 1 shows the average Rerank Score calculated for the triterpene families described in the Akihisa report. Docking scores of some of the studied compounds are close to those of well-known inhibitors like ibuprofen or naproxen. Also, these results indicate that the same type of triterpenoid skeleton may have an affinity to several enzymes involved in the inflammatory process. For example, cholestane and stigmastane skeletons have good Rerank scores for all of the enzymes analyzed even though their scores are not the best among all of the skeletons. Therefore, this class of compounds may bind to many enzymes even with low affinity causing a synergistic effect leading to reduction or inhibition of the inflammatory process. This result is consistent with a trend to reevaluate natural products as privileged structures and starting points for developing antiinflammatory multi-target drugs44 because interference with multiple targets is superior to targeting a single key factor for complex diseases, such as inflammation45, and triterpenoids are among the most studied class of compounds for this purpose.46–49 In addition, these results strongly suggest that the skeleton of the sterol/triterpenoid compound is associated with its potential binding to a specific enzyme. For example, the lanostane skeleton exhibits poor theoretical binding against COX-1, COX-2 and i-INOS. However, the lanostane skeleton has better scores in calculations performed with MPO, PDE4 and, especially, LOX5. Statistical analysis shows that average Rerank score of lanostane-LOX5 complexes are statiscally different from other enzyme complexes. On the other hand, the cholestane skeleton possess a good score averages against COX-1 and COX-2 but poor results in the other set of enzymes tested. Although there are many reports on the structure-activity relationship of the anti-inflammatory and antitumor50 activities of triterpenoids, most of these reports focus on describing the most favorable substitution patterns on a single skeleton or a few skeleton types. The presence of some particular

functional groups has been noted. For example, functional groups that increase the permeability or the presence of glycosidic residues have been reported.51 Our calculations suggest that a key relationship may exist between the skeleton type and enzyme binding. Nevertheless, this relationship does not exclude the necessary presence of some particular functional groups for biological activity. This observation is supported by statistical analyses which show a wide interval for almost all sterol/triterpenoid families (see Supplemental Information).

Bold italics indicate the best docking score average An asterisk indicates there is statistical difference (p > 0.05) with the skeleton-enzyme complex with the lowest average Rerank score Co-crystallized ligands: ibuprofen (COX-1); naproxen (COX-2); AT2_1906 (i-NOS); arachidonic acid (LOX5); Z20_1579 (MPO); JN4_23 (PDE4) Ligands with lowest docking score: cholestane 9 (COX-1); cholestane 16 (COX-2); cardiac steroid 294 (iNOS); lanostane 57 (LOX5); lupane 269 (MPO); dammarane 85 (PDE4). The number after the sterol/triterpenoid skeleton is the identification number of the ligand in the Akihisa report. Structures could be reviewed in the supporting information.

The results for the calculations from the docking studies performed on COX-1 indicate that the cholestane skeleton has the best theoretical binding among all of the skeletons analyzed. Apparently, this result is related to the “width” of the active site in COX-1 because it is too narrow to allow access of compounds with pentacyclic skeletons. In fact, compounds with the latter type of skeletons have the highest Rerank Scores.This is in agreement with some experimental results that indicate that anti-inflammatory effect of some pentacyclic triterpenoids like amyrin or boswellic acid cannot be explained by cyclooxygenase inhibition.52–56 Among the ten compounds with lower docking scores, eight of the compounds belong to the cholestane type of compounds while the other compounds with similar Rerank Score values belong to the stigmastane family (see Supplementary Information). The binding poses of the ten compounds with the lower Rerank Score is shown in Figure 1, and the conformation adopted by most of the compounds involves the tetracyclic template interacting with the amino acids closer to the cofactor while the aliphatic chain lays “outside” the enzyme adopting a “hook-like” conformation. This conformation appears to be important because compounds with unsaturation in the aliphatic chain, cannot adopt this conformation and exhibited lower binding energies.

Experimentally,

5α-stigmastane-3,6-dione exhibits higher anti-

inflammatory activity than 5α-stigmastane-23-ene-3,6-dione57 where the only difference between the two compounds is the presence of unsaturation in the aliphatic chain in the latter compound.

In addition to several steric interactions, most compounds bearing a 3β-hydroxyl or 3-keto group form a hydrogen bond with Met 522. The binding pose of cholestane 16 (identification of compounds is the same of review by Akihisa20 and could be consulted in the Supporting Information) is shown in Figure 2. This compound is interesting because the keto group in position 7 forms an additional hydrogen bond with the Ser 530 residue, which suggests that cholestane or steroid compounds with hydrogen bond donors or acceptors in position 6 or 7 may exhibit good anti-inflammatory activity (e.g., the higher anti-inflammatory activity of 5α-stigmastane-3,6-dione and 5α-stigmastane-23-ene-3,6-dione compared to that of 3β-hydroxyl-5α-stigmastane-24-ene).19,57

Figure 1. Binding poses of the ten compounds with lower Rerank Score in the COX-1 docking study

Figure 2. 2D diagram of binding pose of cholestane 16 in the complex with COX-1 where the keto group in position 7 forms a hydrogen bond with Ser 530 (see text for details, 2D diagram generated with Maestro Version 10.058

The results from the docking studies performed on COX-2 are very similar to those for COX-1 because compounds with pentacyclic skeletons exhibited low theoretical enzyme binding while sterols with tetracyclic skeletons exhibited better binding. Among the ten compounds with lower COX-2 docking scores, five of these compounds belong to the cholestane class, and five of these compounds below to the stigmastane class. It is interesting to note that in general, stigmastane compounds have slightly lower Rerank Scores for COX-2 than for COX-1. This result is consistent with some experiments performed with β-sitosterol and some of its analogues that exhibit greater inhibition against COX-2 than COX-1.59 All of the compounds with good Rerank Scores adopt the same “hook-like” conformation, as observed in the COX-1 docking studies. To conclude, we propose some key structural requirements for good COX binding including a) a tetracyclic skeleton, such as cholestane or stigmastane, b) incorporation of an aliphatic chain in C-17 and c) the presence of hydroxyl groups at C-3 or C-6.

Cardiac steroids and phytoecdysone skeletons have the lowest Rerank score averages in the i-NOS docking studies, suggesting that compounds with cis-fused rings could be important leads for

designing inhibitors for this enzyme. Compounds with this type of skeleton bind closer to the heme group. Particularly, in the case of cardiac steroids, the cis conformation allows the compounds to “interiorize” towards the enzyme catalytic site and interact via hydrogen bonds with Gln 263, Try 347, Tyr 373, Gln 377, Arg 382 and Arg 388, with the butenolide ring interacting with heme group (Figure 3). To our best knowledge, there are no examples describing this potential activity for this class of compounds.

Figure 3. Binding poses of studied cardiac steroids with i-NOS

Despite these type of compounds exhibited the highest theoretical enzyme binding, it was not possible to find a specific skeleton type that exhibited greater selectivity for this enzyme, as observed in the docking studies carried out in COX-1 and COX-2. Some substitution patterns and types of functional groups that improve binding with the enzyme can be proposed, because voluminous groups conferred triterpenoids with better theoretical binding to i-NOS. In particular, the compounds with feruloyl or stearoyl groups in position 3 were among those with lower Rerank Scores. These residues help compounds bind to residues close to the heme group. However, the compounds with more polar groups and aromatic residues (i.e., acetyl, feruloyl) tend to bind away from the heme groups and apparently could remove access to the enzyme catalytic center through interaction with residues located in the entrance to catalytic site, while less polar atoms of the ligands bind near the catalytic site. This observation is well exemplified by oleanane 155 and ergostane 20 (Figure 4), particularly the former compound has higher theoretical binding than the latter, this could be attributed to an additional hydrogen bond with Arg 381 which is located closer to heme group. Taking these observations in consideration, some structural requirements could be proposed for binding to i-NOS: a) a non-polar group or aromatic group that could interact with heme group via π−π interactions, b) next to this non-polar group, hydrogen bond forming groups

are required for interaction with polar residues like Gln 263, Tyr 347, Arg 381 and Asp 382 and c) aromatic polar groups that interact with residues located in the entrance to catalytic site.

Many studies have reported a decrease in NO induced by the administration of triterpenoids and sterols but this effect is due to their inhibition of the induction of iNOS,60,61 whose expression is stimulated by various inflammatory cytokines, such as interferon gamma (IFN-γ), tumor necrosis factor-alpha (TNF-α) and interleukin-1 (IL-1), blocking the ability of inflammatory cytokines to induce transcription of the iNOS gene and not to a direct enzyme inhibition.62 However, our results suggest that some triterpenoids and sterols could also act as direct iNOS inhibitors to some extent.

Figure 4. Binding poses of a ergostane 20 bearing a feruloyl group in position 3 (yellow) and an oleanane 155 (white, CPK rendering) in the i-NOS catalytic site. The polar atoms in the feruloyl or acetyl group are located away from the catalytic site but could improve enzyme binding.

In general, most of the triterpenoid families exhibit low Rerank Score averages in the calculations performed for LOX5. However, the lanostane compounds were among those with lower Rerank Scores. In fact, within the ten compounds with lower Rerank Scores, five of the compounds belong to the lanostane type. The presence of a carboxylic group appears to improve LOX5 binding, as lanostane-type compounds bearing a carboxylic group in position 21 have good binding. In fact, lanostanes with carboxylic group in position 21 have better binding than those having it in position 26. Apparently, carboxylic groups are important for enzyme binding because they interact with basic histidine residues near to catalytic site (Figure 5). This is supported by the fact other compounds possessing a carboxylic group in the aliphatic chain, regardless of the skeleton type also exhibit low Rerank Scores (Figure 6). The inhibition of lipoxygenase products induced by triterpene compounds, especially boswellic acid derivatives, has been extensively studied, and examples of the structure-activity relationship have been described,52,63 including structural features such as the incorporation of hydrogen bond forming groups. The docking results of the oleanane and ursane derivatives bearing a carboxylic group partially explains these findings because it appears that the incorporation of such functional group is crucial for activity.

Figure 5. Docking poses for lanostane-.type compounds bearing a carboxylic group in position 21. Interaction with basic residues around the catalytic site of LOX5 with carboxylate group in position 21 is displayed.

Figure 6. Docking poses for cycloartane 70 (yellow), oleanane 126 (red) and multiflorane 214 (white). All compounds have a carboxylate group that interact with basic Hys 367 or Hys 372 residues

Myeloperoxidase (MPO) has been studied as a potential target for triterpenoids and may be responsible for the anti-inflammatory effect of some compounds.64 Based on the results from the MPO docking studies, the quassine compounds exhibit the lowest Rerank Score averages. However, other skeletons also exhibit low docking scores, suggesting that the substitution pattern is more important for MPO binding than the type of skeleton. In addition, the incorporation of voluminous aromatic groups in position 3 and a carboxylic group in position 21, 28 or 30 are among the structural features common to compounds with good theoretical binding. Voluminous aromatic

residues tend to interact with residues close to the heme group, as shown in Figure 7, and the polar groups in the aliphatic chain interact with polar residues that exist in the catalytic site.

Figure 7. Binding pose for lupane 269 in the MPO docking study. Aromatic ring of feroloyl residue is close to heme close

In the PDE4 docking results, cardiac steroids and phytoecdysone skeletons have the lowest Rerank Score averages. In general, tetracyclic compounds exhibit better Rerank Scores than pentacyclic compounds, which suggests that the size of the triterpenoid skeleton is important for enzyme binding. The dammarane type compounds also exhibit good theoretical binding against this enzyme, and some semisynthetic dammarane derivatives have been evaluated and exhibit good antiinflammatory activity.65 The binding pose of dammarane 85 which was the compound with the lowest Rerank Scores, is shown in figure 8. In this type of compound, the non-polar aliphatic chain binds in the interior of the enzyme, and the tetracyclic skeleton bearing a polar hydroxy group in position 3 is accommodated at the entrance to the catalytic site. This trend was observed among other skeletons, with non-polar groups binding inside the enzyme, which confirms that the combination of an appropriate substitution pattern and proper skeleton type are necessary for good enzyme binding.

Figure 8. Binding pose and 2D representation of interaction of dammarane 85 with PDE4

4. Conclusions Based on all of these results, the substitution pattern seems to be very important for the formation of enzyme-sterol or triterpenoid complexes. However, the skeleton-activity relationship should be considered when using these templates as a base for semisynthesis projects directed toward drug design. In particular, tetracyclic skeletons theoretically bind better to COX-1, COX-2 and PDE4, while pentacyclic skeletons exhibit low theortical binding to these enzymes, which is consistent with available experimental data. These results also support the use of triterpenoids and sterols as a starting point for the design of multi-target strategies because they may exhibit an affinity for many targets, resulting in a synergistic effect that may be useful in the treatment or chemoprevention of some diseases. Acknowledgments The authors wish to thank the Universidad La Salle for funding through project number Q-099/14 and Fernando Parra, M.Sc. for his guidance in the statistical analysis. Supplementary Information Docking poses, figures comparing docking poses of some compounds with those of known inhibitors/ligands, table of analyzed structures and their calculated Rerank Score can be consulted in the following archives. MVD workspace files can be opened with Molegro Molecular Viewer.

References 1.

Balkwill, F. & Mantovani, A. Inflammation and cancer: back to Virchow? Lancet 357, 539– 45 (2001).

2.

Schwartsburd, P. M. Chronic inflammation as inductor of pro-cancer microenvironment: pathogenesis of dysregulated feedback control. Cancer Metastasis Rev. 22, 95–102 (2003).

3.

Mantovani, A., Allavena, P., Sica, A. & Balkwill, F. Cancer-related inflammation. Nature 454, 436–444 (2008).

4.

Porta, C. et al. Cellular and molecular pathways linking inflammation and cancer. Immunobiology 214, 761–77 (2009).

5.

Kundu, J. K. & Surh, Y.-J. Inflammation: gearing the journey to cancer. Mutat. Res. 659, 15–30

6.

Balkwill, F., Charles, K. A. & Mantovani, A. Smoldering and polarized inflammation in the initiation and promotion of malignant disease. Cancer Cell 7, 211–7 (2005).

7.

Coussens, L. M. & Werb, Z. Inflammation and cancer. Nature 420, 860–7

8.

Roder, D. M. The epidemiology of gastric cancer. Gastric Cancer 5 Suppl 1, 5–11 (2002).

9.

De Souza Pereira, R. Selective cyclooxygenase-2 (COX-2) inhibitors used for preventing or regressing cancer. Recent Pat. Anticancer. Drug Discov. 4, 157–63 (2009).

10.

Dolcet, X., Llobet, D., Pallares, J. & Matias-Guiu, X. NF-kB in development and progression of human cancer. Virchows Arch. 446, 475–82 (2005).

11.

Fitzpatrick, B., Mehibel, M., Cowen, R. L. & Stratford, I. J. iNOS as a therapeutic target for treatment of human tumors. Nitric Oxide 19, 217–24 (2008).

12.

Laye, J. P. & Gill, J. H. Phospholipase A2 expression in tumours: a target for therapeutic intervention? Drug Discov. Today 8, 710–716 (2003).

13.

Turini, M. E. & DuBois, R. N. Cyclooxygenase-2: a therapeutic target. Annu. Rev. Med. 53, 35–57 (2002).

14.

Wang, X. & Lin, Y. Tumor necrosis factor and cancer, buddies or foes? Acta Pharmacol. Sin. 29, 1275–88 (2008).

15.

Wink, D. A., Ridnour, L. A., Hussain, S. P. & Harris, C. C. The reemergence of nitric oxide and cancer. Nitric Oxide 19, 65–7 (2008).

16.

Flores-Rosete, G. & Martínez-Vázquez, M. Anti-inflammatory and Cytotoxic Cycloartanes from Guayule (Parthenium argentatum). Nat. Prod. Commun. 3, 413–422

17.

Oviedo-Chávez, I., Ramírez-Apan, T., Soto-Hernández, M. & Martínez-Vázquez, M. Principles of the bark of Amphipterygium adstringens (Julianaceae) with anti-inflammatory activity. Phytomedicine 11, 436–45 (2004).

18.

Chávez, I. O., Apan, T. R. & Martínez-Vázquez, M. Cytotoxic activity and effect on nitric oxide production of tirucallane-type triterpenes. J. Pharm. Pharmacol. 57, 1087–91 (2005).

19.

Salazar, J. R. et al. Anti-inflammatory and cytotoxic activities of chichipegenin, peniocerol, and macdougallin isolated from Myrtillocactus geometrizans (Mart. ex Pfeiff.) Con. Zeitschrift fur Naturforsch. C-Journal Biosci. 66, 24 (2011).

20.

Akihisa, T. & Yasukawa, K. in Bioact. Nat. Prod. (Part F) (Atta-ur-Rahman) 25, Part F, 43–87 (Elsevier, 2001).

21.

Akihisa, T., Yasukawa, K. & Tokuda, H. Bioactive Natural Products (Part J). Stud. Nat. Prod. Chem. 29, 73–126 (Elsevier, 2003).

22.

Dzubak, P. et al. Pharmacological activities of natural triterpenoids and their therapeutic implications. Nat. Prod. Rep. 23, 394–411 (2006).

23.

Kinoshita, K. et al. Antinociceptive Effect of Triterpenes from Cacti. Pharm. Biol. 36, 50–57 (1998).

24.

Medeiros, R. et al. Effect of two active compounds obtained from the essential oil of Cordia verbenacea on the acute inflammatory responses elicited by LPS in the rat paw. Br. J. Pharmacol. 151, 618–27 (2007).

25.

Recio, M. C., Giner, R. M., Máñez, S. & Ríos, J. L. Structural requirements for the antiinflammatory activity of natural triterpenoids. Planta Med. 61, 182–5 (1995).

26.

Aggarwal, B. B., Shishodia, S., Sandur, S. K., Pandey, M. K. & Sethi, G. Inflammation and cancer: how hot is the link? Biochem. Pharmacol. 72, 1605–21 (2006).

27.

Hussain, S. P. & Harris, C. C. Inflammation and cancer: an ancient link with novel potentials. Int. J. Cancer 121, 2373–80 (2007).

28.

Reuter, S. et al. Thiocolchicoside exhibits anticancer effects through downregulation of NFκB pathway and its regulated gene products linked to inflammation and cancer. Cancer Prev. Res. (Phila). 3, 1462–72 (2010).

29.

Rollinger, J. M. Accessing target information by virtual parallel screening—The impact on natural product research. Phytochem. Lett. 2, 53–58 (2009).

30.

Liu, L. et al. A series of natural flavonoids as thrombin inhibitors: structure-activity relationships. Thromb. Res. 126, e365–78 (2010).

31.

Ogungbe, I. V., Ng, J. D. & Setzer, W. N. Interactions of antiparasitic alkaloids with Leishmania protein targets: a molecular docking analysis. Future Med. Chem. 5, 1777–99 (2013).

32.

Ogungbe, I. V., Erwin, W. R. & Setzer, W. N. Antileishmanial phytochemical phenolics: molecular docking to potential protein targets. J. Mol. Graph. Model. 48, 105–17 (2014).

33.

Ehrman, T. M., Barlow, D. J. & Hylands, P. J. In silico search for multi-target antiinflammatories in Chinese herbs and formulas. Bioorg. Med. Chem. 18, 2204–18 (2010).

34.

Spartan 10. (2010). at

35.

Thomsen, R. & Christensen, M. H. MolDock:  A New Technique for High-Accuracy Molecular Docking. J. Med. Chem. 49, 3315–3321 (2006).

36.

Selinsky, B. S. Structural analysis of NSAID binding by prostaglandin H2 synthase: timedependent and time-independent inhibitors elicit identical enzyme conformations. Biochemistry 40, 5172–5180 (2001).

37.

Shanmugam, M. K., Nguyen, A. H., Kumar, A. P., Tan, B. K. H. & Sethi, G. Targeted inhibition of tumor proliferation, survival, and metastasis by pentacyclic triterpenoids: Potential role in prevention and therapy of cancer. Cancer Lett. 320, 158–170 (2012).

38.

Garcin, E. D. et al. Anchored plasticity opens doors for selective inhibitor design in nitric oxide synthase. Nat. Chem. Biol. 4, 700–7 (2008).

39.

Gilbert, N. C. Conversion of human 5-lipoxygenase to a 15-lipoxygenase by a point mutation to mimic phosphorylation at Serine-663. Faseb J. 26, 3222–3229 (2012).

40.

Tidén, A.-K. et al. 2-thioxanthines are mechanism-based inactivators of myeloperoxidase that block oxidative stress during inflammation. J. Biol. Chem. 286, 37578–89 (2011).

41.

Nankervis, J. L. et al. Thiophene inhibitors of PDE4: crystal structures show a second binding mode at the catalytic domain of PDE4D2. Bioorg. Med. Chem. Lett. 21, 7089–93 (2011).

42.

Berman, H. M. et al. The Protein Data Bank. Nucleic Acids Res. 28 , 235–242 (2000).

43.

Minitab 16 Statistical Software. at

44.

Koeberle, A. & Werz, O. Multi-target approach for natural products in inflammation. Drug Discov. Today 19, 1871–1882 (2014).

45.

Morphy, R. & Rankovic, Z. Designed multiple ligands. An emerging drug discovery paradigm. J. Med. Chem. 48, 6523–43 (2005).

46.

Ríos, J. L., Recio, M. C., Máñez, S. & Giner, R. M. in Bioact. Nat. Prod. (Part C) (Chemistry, A.-R. B. T.-S. in N. P.) Volume 22, 93–143 (Elsevier, 2000).

47.

Shanmugam, M. K. et al. Ursolic acid in cancer prevention and treatment: Molecular targets, pharmacokinetics and clinical studies. Biochem. Pharmacol. 85, 1579–1587 (2013).

48.

Shanmugam, M. K. et al. Oleanolic acid and its synthetic derivatives for the prevention and therapy of cancer: Preclinical and clinical evidence. Cancer Lett. 346, 206–216 (2014).

49.

Yadav, V. R., Prasad, S., Sung, B., Kannappan, R. & Aggarwal, B. B. Targeting inflammatory pathways by triterpenoids for prevention and treatment of cancer. Toxins (Basel). 2, 2428–66 (2010).

50.

Xu, K. et al. Cytotoxic activity of Pulsatilla chinensis saponins and their structure-activity relationship. J. Asian Nat. Prod. Res. 15, 680–6 (2013).

51.

Du, J.-R., Long, F.-Y. & Chen, C. in Nat. Prod. Cancer Signal. Isoprenoids, Polyphenols Flavonoids (Enzymes, S. Z. B. and F. T. B. T.-T.) Volume 36, 95–130 (Academic Press, 2014).

52.

Sailer, E.-R. et al. Acetyl-11-keto-β-boswellic acid (AKBA): structure requirements for binding and 5-lipoxygenase inhibitory activity. Br. J. Pharmacol. 117, 615–618 (1996).

53.

Ammon, H. P., Safayhi, H., Mack, T. & Sabieraj, J. Mechanism of antiinflammatory actions of curcumine and boswellic acids. J. Ethnopharmacol. 38, 113–9 (1993).

54.

Safayhi, H. et al. Boswellic acids: novel, specific, nonredox inhibitors of 5-lipoxygenase. J. Pharmacol. Exp. Ther. 261, 1143–1146 (1992).

55.

Medeiros, R., Otuki, M. F., Avellar, M. C. W. & Calixto, J. B. Mechanisms underlying the inhibitory actions of the pentacyclic triterpene alpha-amyrin in the mouse skin inflammation induced by phorbol ester 12-O-tetradecanoylphorbol-13-acetate. Eur. J. Pharmacol. 559, 227–35 (2007).

56.

Verhoff, M. et al. Tetra- and pentacyclic triterpene acids from the ancient anti-inflammatory remedy frankincense as inhibitors of microsomal prostaglandin E(2) synthase-1. J. Nat. Prod. 77, 1445–51 (2014).

57.

Okoye, F. B. C. et al. Anti-inflammatory and membrane-stabilizing stigmastane steroids from Alchornea floribunda leaves. Planta Med. 76, 172–7 (2010).

58.

Schrödinger Release 2014-4: Maestro. (2014). at

59.

Nair, J. J., Aremu, A. O. & Van Staden, J. Anti-inflammatory effects of Terminalia phanerophlebia (Combretaceae) and identification of the active constituent principles. South African J. Bot. 81, 79–80 (2012).

60.

Janakiram, N. B. et al. Chemoprevention of colon carcinogenesis by oleanolic acid and its analog in male F344 rats and modulation of COX-2 and apoptosis in human colon HT-29 cancer cells. Pharm. Res. 25, 2151–7 (2008).

61.

Wang, G.-J., Tsai, T.-H., Chang, T.-T., Chou, C.-J. & Lin, L.-C. Lanostanes from Phellinus igniarius and their iNOS inhibitory activities. Planta Med. 75, 1602–7 (2009).

62.

Sporn, M. B. et al. New synthetic triterpenoids: potent agents for prevention and treatment of tissue injury caused by inflammatory and oxidative stress. J. Nat. Prod. 74, 537–45 (2011).

63.

Najid, A. et al. Characterization of ursolic acid as a lipoxygenase and cyclooxygenase inhibitor using macrophages, platelets and differentiated HL60 leukemic cells. FEBS Lett. 299, 213–217 (1992).

64.

Meira, N. A. et al. Anti-inflammatory and anti-hypersensitive effects of the crude extract, fractions and triterpenes obtained from Chrysophyllum cainito leaves in mice. J. Ethnopharmacol. 151, 975–983 (2014).

65.

Scholz, D. et al. Synthesis of dammarane-type triterpenoids with anti-inflammatory activity in vivo. Bioorg. Med. Chem. Lett. 14, 2983–2986 (2004).

Table 1. Average Rerank Score for each family studied.

Arborinane Cardiac steroids Cholestane Cucurbitane Cycloartane Dammarane Ergostane Euphane Friedelane Glutinane Hopane Lanostane Lupane Moretenane Multiflorane Oleanane Phytoecdisone

COX1 115.84 12.22 -77.83 22.25(*) 11.10(*) 45.33(*) -11.19(*) 56.39(*) 206.15(*) 217.23(*) 124.74(*) 43.44(*) 247.95(*) 109.68(*) 188.80(*) 241.09(*) 45.64(*)

COX2 130.61 58.30(*) -28.54(*) 120.33(*) 44.67(*) 99.91(*) 28.28(*) 164.81(*) 240.97(*) 297.10(*) 141.47(*) 53.89(*) 269.26(*) 220.66(*) 225.30(*) 306.96(*) 44.45(*)

i-NOS -30.61 -97.31 -73.23 -44.85 -40.69(*) -8.44 -93.82 2.68 4.66 1.91 -16.00 -49.69(*) -50.55 39.72(*) -8.90 -19.75 -101.28

LOX5 -86.23 -62.07(*) -81.10 -97.96 -92.98 -84.71 -71.70 -83.02 -47.91 11.32 -69.06 -106.60 -85.27 -13.76 -46.22 -62.44 -82.32

MPO -54.93 -75.23(*) -19.80(*) -61.63 -79.94 -70.07 -13.14 -59.32 -36.93 -56.04 -63.13 -71.44(*) -72.52 -55.96 -58.81 -59.08 -56.47

PDE4 -61.63 -81.30(*) -51.94(*) -58.37 -67.19(*) -80.71 -57.09 -74.96 -38.85 -55.25 -75.20 -67.39(*) -47.25 -52.60 -20.85 -42.13 -80.79

Quassine Spirostane Stigmastane Taraxastane Taraxerane Tirucallane Ursane Co-crystallized ligands Lowest score found in the dataset

195.86(*) 55.63(*) -21.62(*) 196.22(*) 199.77(*) 53.61(*) 231.81(*) -73.88 -94.06

220.49(*) 109.48(*) -33.90(*) 182.79(*) 244.94(*) 91.86(*) 242.97(*) -85.30 -83.23

-78.41 -83.04 -77.69 0.53 -20.80 -66.01 2.64 -91.56 -124.96

-72.56 -53.41 -77.90 -40.38 -66.15 -71.23 -61.30 -24.39 -133.17

-86.63 -71.77(*) -45.24 -59.88 -5.15 -80.93 -57.13(*) -36.79 -114.65

-63.26 -29.72(*) -52.92 -41.67 -49.01 -52.06 -19.42(*) -104.5 -114.5

Bold italics indicate the best docking score average

An asterisk indicates there is statistical difference (p > 0.05) with the skeleton-enzyme complex with the lowest average Rerank score

Co-crystallized ligands: ibuprofen (COX-1); naproxen (COX-2); AT2_1906 (i-NOS); arachidonic acid (LOX5); Z20_1579 (MPO); JN4_23 (PDE4)

Ligands with lowest docking score: cholestane 9 (COX-1); cholestane 16 (COX-2); cardiac steroid 294 (iNOS); lanostane 57 (LOX5); lupane 269 (MPO); dammarane 85 (PDE4). The number after the sterol/triterpenoid skeleton is the identification number of the ligand in the Akihisa report. Structures could be reviewed in the supporting information.