Brian J. Albert

Meayamycin inhibits pre–messenger RNA splicing and exhibits picomolar activity against multidrug-resistant cells

FR901464 is a potent antitumor natural product that binds to the splicing factor 3b complex and inhibits pre-mRNA splicing. Its analogue, meayamycin, is two orders of magnitude more potent as an antiproliferative agent against human breast cancer MCF-7 cells. Here, we report the picomolar antiproliferative activity of meayamycin against various cancer cell lines and multidrug-resistant cells. Time-dependence studies implied that meayamycin may form a covalent bond with its target protein(s). Meayamycin inhibited pre-mRNA splicing in HEK-293 cells but not alternative splicing in a neuronal system. Meayamycin exhibited specificity toward human lung cancer cells compared with nontumorigenic human lung fibroblasts and retained picomolar growth-inhibitory activity against multidrug-resistant cells. These data suggest that meayamycin is a useful chemical probe to study pre-mRNA splicing in live cells and is a promising lead as an anticancer agent. [Mol Cancer Ther 2009;8(8):2308–18]

Zirconium-promoted epoxide rearrangement-alkynylation sequence.

Additions of terminal alkynes to electrophiles are important transformations in organic chemistry. Generally, activated terminal alkynes react with epoxides in an S(N)2 fashion to form homopropargylic alcohols. We have developed a new synthetic method to form propargylic alcohols from epoxides and terminal alkynes via 1,2-shifts. This method involves cationic zirconium acetylides as both the activator of epoxides and nucleophiles. Due to the mild conditions to pre-activate alkynes with silver nitrate, this synthetic method is useful for both electron-rich and electron-deficient alkynes with other acid- and base-sensitive functional groups.

How rapidly do epoxides nonspecifically form covalent bonds with thiols in water

Epoxides are part of many biologically active natural products. For example, trapoxin B (Scheme 1) binds to histone deacetylACHTUNGTRENNUNGases (HDAC), and its analogues are currently in clinical trials for cancer treatment. Epoxomicin is a specific ligand for 20S proteasome. Fumagillin binds to methionine aminopeptidACHTUNGTRENNUNGase 2, and its analogue TNP-470 is now in phase III trials for cancer. All of these small molecules form a covalent bond with their target proteins through epoxide openings. This feature of such natural products inspired the Cravatt group to use MJE3 as an activity-based probe in chemical proteomics. However, epoxides in natural products do not necessarily cross-link to their target proteins; for example, epothilones and triptolide, an ingredient in traditional Chinese herbal therapy lei gong teng to treat inflammatory disorders, do not react with their target proteins. Despite the importance of these biologically active compounds, there is concern about nonspecific reactions with abundant intracellular nucleophiles, particularly with thiols such as glutathione and thiolate-containing proteins. We recently became interested in FR901464 because of its promising potency against various cancers in vivo. Meayamycin is an FR901464 analogue with the remarkable GI50 value of 10 pm in MCF-7 cells. In our ongoing investigations, we, like many others, became concerned about the presence of a spiroepoxide. In theory, epoxides have the potential to react with intraand extracellular thiols, presumably the most abundant and powerful nucleophiles in biological environments, before they reach their desired targets. Despite its significance, quantitative analysis of the opening of epoxides with thiols has been limited to arene oxides and ethylene oxide at 30 and 20 8C, respectively. Subsequently, we questioned whether the nonspecific covalent-bond formation of epoxides with endogenous thiols would be relevant in biological experiments. If so, this would significantly reduce the effective concentration of epoxide-containing bioactive compound in cells. Moreover, it is possible that both the steric and electronic local environments of the epoxide could drastically alter its reactivity towards nucleophiles. Since we are interested in biologically active natural products and many contain epoxides, we desired thiol reactivity data towards these epoxide motifs that until now have remained unstudied. In designing experiments to address this question, one must consider that the local concentrations of thiols and epoxides can vary in vivo, thiols can be activated (deprotonated) under certain biological conditions like in the presence of glutathione transferases, and epoxide hydrolases can cause enzymatic digestion of epoxides. In order to address the unanswered question of whether nonenzymatic covalent bonding of epoxides and thiols was of major biological concern, we chose the reaction between epoxide 1 and N-acetylcysteamine (2) to give thioether 3 (Scheme 2) as a model system, because these substrates are water soluble and readily accessible, and the mercapto group of this thiol exhibits a similar pKa value (9.92) [13] to that of the mercapto group of glutathione (9.42 0.17). Glutathione was avoided as a model thiol because the structurally complex products would be difficult to characterize and because an undesirable kinetic resolution could occur between racemic epoxides and glutathione. To obtain kinetic data, we successfully monitored the reaction between 1 and 2 at 37 8C in buffered D2O in an NMR tube; it only gave 3 (verified by preparation in organic solvent, see the Experimental Section) as the major, expected SN2 product (method A). However, we later found that this was not a general method due to overlap of signals in the NMR spectra when using other epoxides. Consequently, we decided to extract each of the epoxides from an aqueous medium to CDCl3 and determine their consumption based on an internal standard (method B). The accuracy of this method was validated by comparing the data obtained by methods A and B (not shown). Figure 1 displays the data for the consumption of 1. We first examined the pH dependence of the consumption of 1 and found the half-life to be 71 h at pH 6 when [1]=20 mm and [2]=100 mm. At pH 7, the half-life decreased dramatically to 7.0 h. We lowered the concentration of 2 from 100 to 40 mm and observed a proportional increase in the half-life of 1, thereby confirming that the reaction is bimolecular. At pH 8, the SN2 reaction was significantly faster with a half-life of 1.7 h. Since we had previously determined the half-life of a related spiroepoxide to be approximately two days in pH 5–8 phosphate buffers at 37 8C in the absence of thiol, the background hydrolysis reaction should be minor at and above pH 7 at these concentrations in the presence of a thiol. We proceeded to examine the reactivities of other common epoxide motifs found in natural products. Spiroepoxide 4 had half-lives in pH 7 and pH 8 buffers of 3.7 and 0.85 h, respectively (Figures 2 and 3). Monosubstituted epoxide 5 displayed halflives of 2.0 and 0.59 h at pH 7 and 8. 1,2-Disubstituted epoxide 6 gave half-lives of 9.3 and 2.9 h in pH 7 and 8 buffers. The reactivity of a,b-epoxyketone 7 was far greater than that of the other epoxides we had used thus far ; the half-lives at pH 7 and 8 were 1.8 h and 0.22 h, respectively, at much lower concentrations (6.5 s and 0.79 s when put into the same concentrations employed with epoxides 1 and 4–6). To verify that the rapid reactions of 7 were due to the nucleophilic attack of thiols, this epoxide’s half-lives were measured in pH 7 or 8 buffers in the absence of thiol and found to be greater than one day. Thus, these reactions were again bimolecular between the epoxide and the thiol. [a] B. J. Albert, Prof. K. Koide Department of Chemistry, University of Pittsburgh 219 Parkman Avenue, Pittsburgh, PA 15260 (USA) Fax: (+1)412-624-8611 E-mail : koide@pitt.edu Supporting information for this article is available on the WWW under http://www.chembiochem.org or from the author.