Cucurbitacin covalent bonding to cysteine thiols: the filamentous-actin severing protein Cofilin1 as an exemplary target
© Gabrielsen et al.; licensee BioMed Central Ltd. 2013
Received: 29 May 2013
Accepted: 6 August 2013
Published: 14 August 2013
Cucurbitacins are a class of triterpenoid natural compounds with potent bioactivities that led to their use as traditional remedies, and which continue to attract considerable attention as chemical biology tools and potential therapeutics. One obvious target is the actin-cytoskeleton; treatment with cucurbitacins results in cytoskeletal rearrangements that impact upon motility and cell morphology.
Cucurbitacin reacted with protein cysteine thiols as well as dithiothreitol, and we propose that the cucurbitacin mechanism of action is through broad protein thiol modifications that could result in inhibition of numerous protein targets. An example of such a target protein is Cofilin1, whose filamentous actin severing activity is inhibited by cucurbitacin conjugation.
The implications of these results are that cucurbitacins are unlikely to be improved for selectivity by medicinal chemistry and that their use as chemical biology probes to analyse the role of specific signalling pathways should be undertaken with caution.
KeywordsActin Cofilin Cytoskeleton Cucurbitacin
The natural compound cucurbitacin triterpenes are a group of structurally related compounds that give a variety of plants and fungi their bitter taste as a defense against being eaten . Historically, cucurbitacin-producing plants or extracts have been used as traditional remedies for diseases such as cancer, inflammation and infection . More recently, cucurbitacins have attracted attention because of several notable properties. Their potent cytotoxicity has led to numerous investigations on their potential utility as anti-cancer therapeutics [3–5]. In addition, they have been used as chemical biology probes to explore the biological roles of signalling pathways including Jak/STAT3 [6, 7], NF-κB , MAPK/ERK [9–11] and PI3-kinase . Cucurbitacin analogues also have marked effects on the actin cytoskeleton, which in turn affects processes such as cell motility, tumour cell invasion and metastasis [12–15]. In fact, the ability of cucurbitacin E to inhibit filamentous actin (F-actin) depolymerisation led to the suggestion that it would be useful as a tool to study actin dynamics and actin-based processes in live cells .
With so many apparent biological activities, an important issue is how cucurbitacin compounds interact with and consequently inhibit their protein targets. This question is particularly important if a decision were made to use medicinal chemistry to optimize cucurbitacin compounds as anti-cancer therapeutics by improving their on-target selectivity and potency while minimizing their reported toxicities . The mode of cucurbitacin binding to protein targets is also an important issue if they are to be used as chemical biology probes with confidence. One attempt to address this question used in silico docking of cucurbitacin B and E into the hydrophobic ligand-binding pocket of B-Raf . However, no direct physical measurements were made to validate this hypothetical mechanism of action.
A previous attempt to characterize important cellular targets of cucurbitacin used biotinylated cucurbitacin E to purify interacting proteins from lysates of U937 human leukaemia cells and mass spectrometry (MS) for protein identification, which led to the discovery of Cofilin1 as a major interacting protein . In this study, we aimed to examine how cucurbitacin analogues affect Cofilin1 activity and to identify the mechanism of action. Using Cofilin1 as an exemplar, we sought to determine if there might be specific or general modes used by cucurbitacin compounds to interact with target proteins, and by inference whether these compounds could potentially be optimized for potency and selectivity.
Results and discussion
The actin severing assay was used to determine whether cucurbitacin E inhibited Cofilin1 F-actin severing activity (Figure 3E). While DMSO vehicle control did not change the S/P ratio, cucurbitacin E notably increased the pelleted F-actin, consistent with a direct effect on inhibiting depolymerisation . The effect of Cofilin1 was to shift actin to the monomeric S fraction, which was unaffected by DMSO vehicle control. In contrast, treatment with cucurbitacin E inhibited the actin severing activity of Cofilin1, since the P fraction was significantly (p = 0.05) increased relative to DMSO-treated Cofilin1 samples (Figure 3F).
It is worth noting that the cucurbitacin-conjugated Cofilin1 became extremely hydrophobic with a marked propensity for non-specific binding to dialysis membranes and columns, which resulted in complete loss of detectable cucurbitacin-conjugated Cofilin1 in some procedures. Only by treating proteins directly in solution or following extraction from polyacrylamide gels could any material be obtained for analysis. Although Cofilin1 was identified as a biotinylated-cucurbitacin E interacting protein in cell lysates by mass spectrometry , no mass shift was detected after incubation of purified Cofilin1 or Gelsolin with cucurbitacin I , likely due to the loss of cucurbitacin-conjugated protein because of their hydrophobicity. This property has also likely hampered efforts to identify additional cucurbitacin binding proteins.
Our results clearly show that cucurbitacin compounds bind covalently to the cysteines present in Cofilin1, and that this binding inhibits the ability of Cofilin1 to sever F-actin. By modelling the addition of cucurbitacin E to Cys139 on our recently solved crystal structure of human Cofilin1 (depicted with space filling; PDB ID 4BEX ) that was modelled with associated actin (ribbon) based on the C-terminal Cofilin-like domain of mouse twinfilin (Twf-C) in complex with a single unit of actin (PDB ID 3DAW) , it becomes apparent that the conjugated cucurbitacin would disrupt the interface between Cofilin1 and actin (Figure 5B). A similar conjugation to Cys147 would also likely lead to a clash with F-actin that would inhibit actin-severing. The increased hydrophobicity of the surface of cucurbitacin-modified Cofilin1 would also likely interfere with binding to F-actin. Although Cys39 and Cys80 are positioned away from the Cofilin1-actin interface, their relatively buried positions could result in disruption of protein folding if conjugated to the bulky cucurbitacin molecules, similar to the cucurbitacin-induced structural rearrangements observed for human serum albumin , which could also inhibit F-actin severing activity.
In addition to cucurbitacin E binding to thiols on F-actin , Cofilin1 and DTT, cucurbitacin B was reported to form adducts with thiol-containing N-acetylcysteine (NAC) and glutathione, but not thiol-free vitamin C or ascorbic acid . Although NAC blocked the cytotoxicity of cucurbitacin B, vitamin C and ascorbic acid did not . These results suggest that rather than inhibiting the effect of cucurbitacin B through anti-oxidant properties, which NAC, vitamin C and ascorbic acid all share, the inhibitory effect of NAC is more likely to be a consequence of its direct conjugation with cucurbitacin B. In fact, the ability of the reactive oxygen scavenging NAC to reverse the effects of cucurbitacin has widely been interpreted as evidence that generation of reactive oxygen species is part of its mechanism of action . Our alternative interpretation is that the neutralization of the biological effects of cucurbitacin compounds is via direct conjugation with NAC in tissue culture medium or in cells, meaning that all experiments in which NAC and cucurbitacins have been combined should be interpreted with caution.
Although cucurbitacin compounds have been proposed to be potential anti-cancer drugs and are used to inhibit specific signal transduction pathways, the results of this study and others [16, 23] support the conclusion that cucurbitacins non-specifically bind protein targets by forming thioether bonds via a Michaels-type addition. This would allow cucurbitacins to be conjugated with a broad array of potential protein targets, many of which would be inhibited or disrupted as a consequence. As a result, their value as chemical biology probes is limited and must be confirmed by independent means. For example, although cucurbitacin I was reported to be a selective inhibitor of Jak/STAT3 signalling  and has been used to test the involvement of this pathway in various processes, the ability of cucurbitacin I to activate Rac1 was not replicated by Jak2 or Stat3 knockdown by siRNA . The findings in this study also indicate that the binding mode of cucurbitacin compounds to protein targets means that optimization for selectivity would be unlikely to work, which would make it very difficult to minimize toxicities or improve the therapeutic window for future clinical development. Optimistically, alternative ways that their potential therapeutic utility could be improved in the future would be through targeted delivery to tumour cells, for example through antibody-conjugation or incorporation in liposome microparticles .
MCF-7 human breast cancer cells were grown in Dulbecco’s modified eagle’s medium (Gibco) with 10% fetal bovine serum (PAA) at 37°C and 5% CO2. Transfections with non-targeting control siRNA D001810-01-05 and human Cofilin1 siRNA J012707-05 (Abgene) were performed as described in .
Cytotoxicity and actin-phalloidin fluorescence assays
MCF-7 cells were trypsinized and to each well of Greiner black-walled, clear bottomed 96-well plates, 100 μl of cell suspension (2,000 cells for cytotoxicity, 5,000 cells for phalloidin staining) was added. Plates were left to stand in the hood for 10 minutes to aid uniform cell attachment prior to placing in CO2 incubator. After 24 hours attachment, 100 μl of media containing DMSO vehicle or cucurbitacin compounds was added to the wells. For cytotoxicity, cells were cultured for a further 3 days and then after fixation with 4% paraformaldehyde for 15 minutes at room temperature and DAPI staining of cell nuclei, cell number was assessed using an Operetta high content imaging system (Perkin Elmer). For actin-phalloidin fluorescence measurements, cells were cultured for 4 hours after treatment. After fixation with 4% paraformaldehyde for 15 minutes at room temperature, permeabilization with 0.25% (v/v) Triton X-100 and staining of actin structures with Texas Red-conjugated Phalloidin (2 μg/mL), fluorescence intensity was measured using an Operetta high content imaging system. Imaging of actin structures was done using a Zeiss 710 confocal microscope.
Binding of cucurbitacin to Cofilin1
Cofilin1 was expressed and purified as described previously . Binding of cucurbitacin to Cofilin1 was achieved by incubating 5 μM Cofilin1 with a dose range from 5 μM to 500 μM cucurbitacin (solubilised in 100% DMSO) or an equivalent volume of DMSO at 4°C overnight. Cucurbitacin binding to Cofilin1 was analysed on 12% Bis-Tris SDS-PAGE gels and SimplyBlue (Invitrogen) stained gels scanned on a LiCor Odyssey.
5 μmol Cofilin1 was incubated with 0.5 mM Cucurbitacin E, separated by SDS-PAGE, stained with colloidal Coomassie Blue and digested with trypsin. The digests were analysed by LC-MS using an AB-Sciex 5600 TripleTOF mass spectrometer coupled to an Eksigent 2D Ultra HPLC system fitted with a 150 × 0.075 mm C18 packed emitter (New Objective). Digests were loaded in 2% acetonitrile/0.1% formic acid (Buffer A) and then eluted with a linear gradient of buffer B (100% acetonitrile/0.1% formic acid) at 300 nL/min. LC-MS data was searched using Mascot 2.4 run on a local server against SwissProt allowing for trypsin cleavages and cysteine modifications of oxidation and cucurbitacin isoforms. To determine the intact mass of cucurbitacin modified Cofilin1, complexes were diluted into 25% acetonitrile/0.5% formic acid and separated on an Eskigent 150 × 0.3 mm ChromXP C18CL column. The mass spectra were acquired on the 5600 TripleTOF with the intact protein script activated (AB-Sciex) from m/z 600–2000 and spectra were deconvoluted using BioAnalyst 1.5 to calculate the exact mass of the protein complex.
100 μM Cucurbitacin E was incubated plus or minus 10 mM DTT in 50 mM TEAB pH 8 for 3 h at room temperature and diluted 1:10 with 50% acetonitrile/0.1% formic acid. Samples were analysed by MS on 5600 TripleTOF by infusing at 1 μl/min and data was acquired in negative ion mode (TOF-MS and MS-MS).
F-actin depolymerization assay
The Actin Binding Protein Biochem Kit with Non-Muscle Actin (Cytoskeleton) was used according to manufacturer’s instructions. Aliquots of actin polymerization buffer and ATP were rapidly defrosted in a room temperature water bath and kept on ice. Cofilin1 samples were centrifuged for 1 h at 100,000 X g at 4°C, supernatants taken and kept on ice as test protein stocks. General actin buffer was supplemented with 0.2 mM final ATP. One 250 μg aliquot of actin was diluted to 1 mg/ml with 225 μl of ATP-supplemented general actin buffer, mixed to ensure complete re-suspension and left on ice for 30 min. 25 μl of actin polymerization buffer was pipetted into the actin protein and mixed well. The actin was incubated at room temperature for 1 h and kept as F-actin stock at 21 μM actin. All tubes were incubated at room temperature for 30 min. Afterwards the tubes were centrifuged at 150,000 X g for 1.5 h at 24°C. Supernatants were removed and kept on ice. To each tube 10 μl of 5× Laemmli reducing sample buffer was added. The pellets were re-suspended in 30 μl water by mixing for 2 min, leaving on ice for 10 min and repeated mixing. The samples were transferred into tubes and supplemented with 30 μl 2× Laemmli reducing-sample buffer. The samples were frozen at −20°C until further analysis by SDS-PAGE. Quantification was performed with a Licor Odyssey by scanning the fluorescence intensity of SimplyBlue (Invitrogen) stained bands at 700 nm.
Half maximal effective concentration
Monomeric globular actinMS mass spectrometry
Sodium dodecyl sulphate polyacrylamide gel electrophoresis
Time of flight
C-terminal Cofilin-like domain of mouse twinfilin.
Funded by Breast Cancer Campaign (2012MaySP008) and CRUK.
- Chen JC, Chiu MH, Nie RL, Cordell GA, Qiu SX: Cucurbitacins and cucurbitane glycosides: structures and biological activities. Nat Prod Rep. 2005, 22: 386-399. 10.1039/b418841c.PubMedView ArticleGoogle Scholar
- Lee DH, Iwanski GB, Thoennissen NH: Cucurbitacin: ancient compound shedding new light on cancer treatment. Sci World J. 2010, 10: 413-418.View ArticleGoogle Scholar
- Chen X, Bao J, Guo J, Ding Q, Lu J, Huang M, Wang Y: Biological activities and potential molecular targets of cucurbitacins: a focus on cancer. Anticancer Drugs. 2012, 23: 777-787. 10.1097/CAD.0b013e3283541384.PubMedView ArticleGoogle Scholar
- Patlolla JM, Rao CV: Triterpenoids for cancer prevention and treatment: current status and future prospects. Curr Pharm Biotechnol. 2012, 13: 147-155. 10.2174/138920112798868719.PubMedView ArticleGoogle Scholar
- Alghasham AA: Cucurbitacins - a promising target for cancer therapy. Int J Health Sci. 2013, 7: 77-89.View ArticleGoogle Scholar
- Blaskovich MA, Sun J, Cantor A, Turkson J, Jove R, Sebti SM: Discovery of JSI-124 (cucurbitacin I), a selective Janus kinase/signal transducer and activator of transcription 3 signaling pathway inhibitor with potent antitumor activity against human and murine cancer cells in mice. Cancer Res. 2003, 63: 1270-1279.PubMedGoogle Scholar
- Sun J, Blaskovich MA, Jove R, Livingston SK, Coppola D, Sebti SM: Cucurbitacin Q: a selective STAT3 activation inhibitor with potent antitumor activity. Oncogene. 2005, 24: 3236-3245. 10.1038/sj.onc.1208470.PubMedView ArticleGoogle Scholar
- Jin HR, Jin X, Dat NT, Lee JJ: Cucurbitacin B suppresses the transactivation activity of RelA/p65. J Cell Biochem. 2011, 112: 1643-1650. 10.1002/jcb.23078.PubMedView ArticleGoogle Scholar
- Chan KT, Li K, Liu SL, Chu KH, Toh M, Xie WD: Cucurbitacin B inhibits STAT3 and the Raf/MEK/ERK pathway in leukemia cell line K562. Cancer Lett. 2010, 289: 46-52. 10.1016/j.canlet.2009.07.015.PubMedView ArticleGoogle Scholar
- Zhou X, Yang J, Wang Y, Li W, Li-Ling J, Deng Y, Zhang M: Cucurbitacin B inhibits 12-O-tetradecanoylphorbol 13-acetate-induced invasion and migration of human hepatoma cells through inactivating mitogen-activated protein kinase and PI3K/Akt signal transduction pathways. Hepatol Res. 2012, 42: 401-411. 10.1111/j.1872-034X.2011.00933.x.PubMedView ArticleGoogle Scholar
- Ahmed MS, Halaweish FT: Cucurbitacins: potential candidates targeting mitogen-activated protein kinase pathway for treatment of melanoma. J Enz Inhib Med Chem. 2013, 10.3109/14756366.2012.762646.Google Scholar
- Nakashima S, Matsuda H, Kurume A, Oda Y, Nakamura S, Yamashita M, Yoshikawa M: Cucurbitacin E as a new inhibitor of cofilin phosphorylation in human leukemia U937 cells. Bioorg Med Chem Lett. 2010, 20: 2994-2997. 10.1016/j.bmcl.2010.02.062.PubMedView ArticleGoogle Scholar
- Zhang T, Li J, Dong Y, Zhai D, Lai L, Dai F, Deng H, Chen Y, Liu M, Yi Z: Cucurbitacin E inhibits breast tumor metastasis by suppressing cell migration and invasion. Breast Cancer Res Treat. 2012, 135: 445-458. 10.1007/s10549-012-2175-5.PubMedView ArticleGoogle Scholar
- Khabbazi S, Jacques RO, Moyano Cardaba C, Mueller A: Janus kinase 2 and signal transducer and activator of transcription 3 activation is not essential for CCL3-, CCL5- or CCL8-induced chemotaxis. Cell Biochem Func. 2012, 10.1002/cbf.2901.Google Scholar
- Ren S, Ouyang DY, Saltis M, Xu LH, Zha QB, Cai JY, He XH: Anti-proliferative effect of 23,24-dihydrocucurbitacin F on human prostate cancer cells through induction of actin aggregation and cofilin-actin rod formation. Cancer Chemother Pharmacol. 2012, 70: 415-424. 10.1007/s00280-012-1921-z.PubMedView ArticleGoogle Scholar
- Sorensen PM, Iacob RE, Fritzsche M, Engen JR, Brieher WM, Charras G, Eggert US: The natural product cucurbitacin E inhibits depolymerization of actin filaments. ACS Chemical Biol. 2012, 7: 1502-1508. 10.1021/cb300254s.View ArticleGoogle Scholar
- Raikhlin-Eisenkraft B, Bentur Y: Ecbalium elaterium (squirting cucumber)–remedy or poison?. Clin Toxicol. 2000, 38: 305-308. 10.1081/CLT-100100936.Google Scholar
- Knecht DA, LaFleur RA, Kahsai AW, Argueta CE, Beshir AB, Fenteany G: Cucurbitacin I inhibits cell motility by indirectly interfering with actin dynamics. PLoS One. 2010, 5: e14039-10.1371/journal.pone.0014039.PubMed CentralPubMedView ArticleGoogle Scholar
- Klejnot M, Gabrielsen M, Wood J, Mleczak A, Talapatra S, Kozielski F, Pannifer A, Olson MF: Analysis of the human Cofilin1 structure reveals conformational changes required for actin-binding. Acta Crystallogr D: Biol Crystallogr. 2013, 69: 1780-1788. 10.1107/S0907444913014418.View ArticleGoogle Scholar
- Paavilainen VO, Oksanen E, Goldman A, Lappalainen P: Structure of the actin-depolymerizing factor homology domain in complex with actin. J Cell Biol. 2008, 182: 51-59. 10.1083/jcb.200803100.PubMed CentralPubMedView ArticleGoogle Scholar
- Greige-Gerges H, Abou Khalil R, Chahine R, Haddad C, Harb W, Ouaini N: Effect of cucurbitacins on bilirubin-albumin binding in human plasma. Life Sci. 2007, 80: 579-585. 10.1016/j.lfs.2006.10.005.PubMedView ArticleGoogle Scholar
- Kausar H, Munagala R, Bansal SS, Aqil F, Vadhanam MV, Gupta RC: Cucurbitacin B potently suppresses non-small-cell lung cancer growth: identification of intracellular thiols as critical targets. Cancer Lett. 2013, 332: 35-45. 10.1016/j.canlet.2013.01.008.PubMedView ArticleGoogle Scholar
- Yasuda S, Yogosawa S, Izutani Y, Nakamura Y, Watanabe H, Sakai T: Cucurbitacin B induces G2 arrest and apoptosis via a reactive oxygen species-dependent mechanism in human colon adenocarcinoma SW480 cells. Mol Nutr Food Res. 2010, 54: 559-565. 10.1002/mnfr.200900165.PubMedView ArticleGoogle Scholar
- Lopez-Haber C, Kazanietz MG: Cucurbitacin I inhibits rac1 activation in breast cancer cells by a reactive oxygen species-mediated mechanism and independently of janus tyrosine kinase 2 and p-rex1. Mol Pharmacol. 2013, 83: 1141-1154. 10.1124/mol.112.084293.PubMed CentralPubMedView ArticleGoogle Scholar
- Habib L, Khreich N, Jraij A, Abbas S, Magdalou J, Charcosset C, Greige-Gerges H: Preparation and characterization of liposomes incorporating cucurbitacin E, a natural cytotoxic triterpene. Int J Pharm. 2013, 448: 313-319. 10.1016/j.ijpharm.2013.03.027.PubMedView ArticleGoogle Scholar
- Croft DR, Crighton D, Samuel MS, Lourenco FC, Munro J, Wood J, Bensaad K, Vousden KH, Sansom OJ, Ryan KM, Olson MF: p53-mediated transcriptional regulation and activation of the actin cytoskeleton regulatory RhoC to LIMK2 signaling pathway promotes cell survival. Cell Res. 2011, 21: 666-682. 10.1038/cr.2010.154.PubMed CentralPubMedView ArticleGoogle Scholar
- Scott RW, Hooper S, Crighton D, Li A, Konig I, Munro J, Trivier E, Wickman G, Morin P, Croft DR, et al: LIM kinases are required for invasive path generation by tumor and tumor-associated stromal cells. J Cell Biol. 2010, 191: 169-185. 10.1083/jcb.201002041.PubMed CentralPubMedView ArticleGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.