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

FR901464 (Fig. 1) is a natural product isolated from the culture broth of the bacterium Pseudomonas sp. 2663 (1). FR901464 was discovered as an activator of a stably transfected reporter gene driven by an SV40 DNA tumor virus promoter in human breast adenocarcinoma MCF-7 cells (1). FR901464 exhibited antiproliferative activity against MCF-7, human lung adenocarcinoma A549, colon cancer HCT116 and SW480, and murine leukemia P388 cell lines with IC₅₀ values of 1.8, 1.3, 0.61, 1.0, and 3.3 nmol/L, respectively. The IC₅₀ against mouse bone marrow cells was 9.9 nmol/L, indicating specificity toward tumor cells relative to bone marrow cells (1). FR901464 was effective against P388, A549, murine colon 38, and Meth A cell xenograft models in mice (2). To date, three total chemical syntheses of FR901464 have been reported (3–10). In addition to FR901464, the Jacobsen group synthesized 2, which was slightly more potent than FR901464 against Jurkat cells (5). The Kitahara group reported 3 to be more active than FR901464, although numerical data were not presented (11). We synthesized meayamycin (1) and showed this analogue to be two orders of magnitude more potent than FR901464 against MCF-7 cells (9). More recently, the Yoshida group reported that a target of FR901464 was the splicing factor 3b (SF3b) complex (12). Together with pladienolides (13), these two classes of natural products represent the first inhibitors of pre-mRNA splicing via SF3b binding and represent promise for the development of new chemotherapeutic approaches through splicing inhibition (14). A pladienolide analogue, E7107, is now in a phase I clinical trial for various cancers (15). In contrast, the development of a compound derived from FR901464 lags behind, although the Webb group recently reported analogues of FR901464 and showed their biological activity (16). Herein, we wish to report further studies of meayamycin in a cell-free system and in living cells.

For the stability of meayamycin in buffers, a Corning tube (15 mL) was filled with phosphate buffer (10 mL), PhCO₂H (10 mmol/L in DMSO, 20 μL), and DMSO (70 μL) under an open atmosphere and warmed to 37°C. Meayamycin (10mmol/L in DMSO, 10 μL) was added to the reaction mixture and the resulting mixture was sealed with a cap, vortexed for 15 s, and then placed in a 37°C incubator. The decomposition was monitored by high-performance liquid chromatography (HPLC) at the indicated times. The resulting data were normalized by dividing the ratios of meayamycin/PhCO₂H by the ratio of the first data point.

For the stability of meayamycin in cell culture media, a dram vial was filled with 10% fetal bovine serum (FBS) in RPMI 1640 (1 mL), rhodamine (10 mmol/L in DMSO, 1 μL), and DMSO (7 μL) under an open atmosphere and warmed to 37°C. Meayamycin (10 mmol/L in DMSO, 1 μL) was added to the reaction mixture, and the resulting mixture was sealed with a cap, vortexed for 15 s, and then placed in a 37°C incubator. The decomposition was monitored by HPLC at the indicated times. The resulting data were normalized by dividing the ratios of meayamycin/rhodamine by the ratio of the first data point.

The HPLC analysis for the stability of meayamycin in buffers was done on a Varian Pursuit XDR C18 column, 75 × 4.6 mm. The elution conditions—flow, 1.0 mL/min; eluant, 10% MeCN/H₂O (containing 1% HCO₂H) to 100% MeCN linear gradient elution from 0.5 to 9.5 min. Meayamycin and benzoic acid were monitored at 230 nm. Retention time was 6.7 min for meayamycin, and 4.3 min for benzoic acid. The HPLC analysis for stability of meayamycin in cell culture medium was done on a Varian Pursuit XRs 5 C18 column, 250 × 10.0 mm. The elution conditions—flow, 2.5 mL/min; eluant, 30% methanol/H₂O (containing 0.1% HCO₂H) to 100% methanol linear gradient elution from 0.5 to 30.5 min. Meayamycin was monitored at 232 nm. Rhodamine was monitored at 550 nm. Retention time was 4.3min for meayamycin and 18.4 min for rhodamine. The data were analyzed with GraphPad Prism 5.0. The calculation of meayamycin degradation in 10% FBS in RPMI 1640 was done based on a one-phase exponential decay equation in the software.

The cells used for Table 1 and Fig. 2 were grown at 37°C in an atmosphere containing 5% CO₂ in Corning cell culture flasks (25 cm²) in RPMI 1640 cell culture medium containing 10% FBS, 1% penicillin/streptomycin, and 1% l-glutamine. Cells used for Fig. 3 were HEK-293 cells with a stably integrated splicing reporter consisting of a human triose phosphate isomerase gene exon 6-intron 6-exon 7 cassette upstream of luc2 (17). A549 human lung cancer cells and IMR-90 normal diploid lung fibroblasts were from the American Type Culture Collection. DC3F transformed Chinese hamster cells and their multidrug-resistant (MDR) VCRd5L counterparts were a gift from Prof. John Lazo (University of Pittsburgh, Pittsburgh, PA). These cells were grown at 37°C in an atmosphere containing 5% CO₂ in DMEM cell culture medium containing 10% FBS. Neurons used for Fig. 4 were dissociated from E18 rat cortical tissue and grown at 37°C with 6% CO₂ in Neurobasal medium with a B27 supplement.

The compounds were dissolved in DMSO as 10 mmol/L stocks and stored at −20°C. For the experiments, aliquots were thawed at room temperature and diluted solutions were prepared in RPMI 1640 containing 2% DMSO at twice the desired concentration and were warmed to 37°C prior to addition to the cells. Control cultures received prewarmed medium containing 2% DMSO.

Two independent methods were used to assess growth-inhibitory activity. For Table 1, cells were plated in 96-well plates at an initial density of 2,000 cells per well in 100 μL of medium, and were incubated for 24 h prior to compound addition. Serial 2-fold dilutions were used from 100nmol/L to 0.191 pmol/L. The compound was added to the cells at twice the desired concentration in 100 μL cell culture medium. The compound solutions were prewarmed to 37°C. The cells were then incubated for 3 to 5d. Cell proliferation was measured using a commercial MTS solution (20 μL/well). The absorbance (at 490 nm minus that at 630 nm) was measured by a Spectromax M2 plate reader (Molecular Devices). Test agent concentration points within each experiment were done in quadruplicate, and experiments were repeated at least three times to ensure reproducibility. The final GI₅₀ values were averaged.

For DC3F and VCRd-5L Chinese hamster cells, we used our previously described image-based analysis procedure (18). Briefly, human cancer cell lines were plated at a low density (500–2,000 cells/well depending on the cell line) in 384-well microplates. Cells were allowed to attach overnight and treated with compounds. After 3 d of continuous exposure to compounds, cells were labeled with Hoechst 33342 and nuclei enumerated on the ArrayScan II. Acquisition variables were set such that the instrument acquired a minimum of 5,000 cells in each well. Cell densities were calculated as objects per imaging field and normalized to vehicle control density at the end of the study.

Cells were plated in 96-well plates at an initial density of 2,000 cells per well in 100 μL of medium and were incubated for 24 h prior to the addition of the compound. One concentration was used in each experiment for all times examined. The compound was added to the cells at twice the desired concentration in 100 μL cell culture medium. At the desired time intervals, the medium containing the drug was removed, the wells were washed five times with new medium and 200 μL of new medium containing 1% DMSO was added. At the last time interval, after washing and replacing the medium, cell proliferation was measured using a commercial MTS solution (20 μL/well). The absorbance (at 490 nm minus that at 630 nm) was measured by a Spectromax M2 plate reader (Molecular Devices). Each concentration treatment was done in quadruplicate and the final numbers were averaged.

Uniformly, ³²P-labeled AdML pre-mRNA splicing substrate was generated by T7 runoff transcription methods and gel-purified. HeLa cell nuclear extracts were prepared as previously described (19–21). Splicing reactions containing 1 or 5, ∼20 nmol/L pre-mRNA and 30% nuclear extract, in 40 mmol/L Tris-HCl (pH 8), containing 20 mmol/L KCl, 2.5 mmol/L MgCl₂, 10 mmol/L creatine phosphate, 0.5mmol/L DTT, 0.4 units/μL RNasin, 0.5 mmol/L ATP, and 1% (v/v) DMSO were incubated for 60 min at 30°C. For analysis of splicing, reactions were stopped with 9 volumes of stop buffer [20 mmol/L Tris (pH 7.5), containing 300 mmol/L sodium acetate, 0.5% SDS, 100 mmol/L NaCl, 5 mmol/L EDTA, and 25 ng/μL tRNA], extracted twice with phenol and precipitated with ethanol. RNAs were separated on 15% denaturing polyacrylamide gels and visualizedwith a phosphorimager. Native splicing complexes were analyzed as described previously (21, 22). Briefly, heparin was added to a final concentration of 0.5mg/mL, reactions were incubated for 5 min at 30°C and placed on ice. Reactions were centrifuged at 14,000rpm in an Eppendorf microcentrifuge and immediately loaded onto a 4% acrylamide (80:1 acrylamide/bis), and 50mmol/L Tris-glycine native gel. Gels were visualized with a phosphorimager.

Reverse transcription-PCR (RT-PCR) was carried out using the SuperScript III One-step RT-PCR kit (Invitrogen). Total RNA (60 ng) that was DNase-treated was used per 25 μL reaction, run on 1% agarose gels, and stained with ethidium bromide. Primers used for amplification for the experiment of Fig. 3 were 5′-TAA ACT TAA GCT TCA GCG CCT CGG-3′ and 5′-TAG CGC TTC ATG GCT TTG TGC AG-3′. For the experiments shown in Fig. 4, 1 μg of total RNA was reverse transcribed with Moloney murine leukemia virus reverse transcriptase (Invitrogen) with random primers. PCR reactions (10 μL) were done for 21 to 24 cycles with 1 μL of the reverse-transcribed sample. DNA products were separated on 1.2% agarose gels and visualized with ethidium bromide. Forward (F) and reverse (R) primer sequences used for the experiments in Fig. 4 are indicated as follows (5′ to 3′). GRIN1 C1 cassette: (F) ATGCCCGTAGGAAGCAGATGC, (R) CGTCGCGGCAGCACTGTGTC; actin E2, (F) AGCTTCTTTGCAGCTCCTTC, (R) TCTTCTCCATGTCGTCCCAG; KCNQ2 exon 8, (F) TCTATGCTACTAACCTCTCACGCAC, (R) GTGAGTCCAGATTTGCTCTTGAGATTC; HNRPH3 exon 3, (F) GGGACAGTGCGACTTCGTGG, (R) CTGTCCCAGCAATCTTCTTGGTGG; HNRPH3 exon 8, (F) GCTATGGAGGCTATGATG, (R) ACCAGTTACTCTGCCATC; CAV1.3 exon 2-exon 3, (F) TCCTATCTTGGCAAGCTGCAATCG, (R) AGCTAAGGCCACACAATTGGCA; H1F0 (intronless), (F) ATGACCGAGAACTCCACCTCC, (R) TTCTTCTTGCTGGCTCTCTTGG; GRIN1 exon 2-exon N1, (F) AGATGATGCGCGTCTACAACTGGAAC, (R) CATAGGACAGTTGGTCGAGGT; and GRIN1 intron2-exon N1, (F) GGTGTTTGCGAGCTCCAGG, (R) CATAGGACAGTTGGTCGAGGT.

Multiparametric cellular profiling studies followed procedures described previously (23, 24). A549 or IMR-90 cells (10,000/well) were plated in collagen 1–coated 384-well microplates and treated the next day with vehicle (DMSO) or 10 2-fold concentration gradients of test agents. Cells were incubated for 24 h at 37°C, fixed with formaldehyde, labeled with 10 μg/mL of Hoechst 33342, and incubated with a primary antibody cocktail containing rabbit monoclonal anti-cleaved caspase-3 (1:500) and sheep polyclonal anti-p53 (1:500), followed by a mixture of Cy3-labeled donkey anti-rabbit IgG (1:250) and Cy5-labeled donkey anti-sheep IgG (1:250, both from Jackson ImmunoResearch). Microplates were analyzed with an ArrayScan II (Cellomics) using the Compartmental Analysis Bioapplication. One thousand individual cells in each well were imaged at excitation/emission wavelengths of 350/461 nm (Hoechst), 556/573 nm (Cy3), and 649/670 nm (Cy5), respectively. A nuclear mask was generated from Hoechst 33342–stained nuclei. Fluorescence intensities in the Cy3 and Cy5 channels were then measured in an area defined by the nuclear mask. The following variables were extracted for data analysis: average nuclear intensity (a measure of nuclear condensation), nuclei per field (cell density), object area (nuclear size), and average Cy3 and Cy5 intensities (to determine levels of cleaved caspase and p53, respectively). To calculate the percentages of cells with condensed nuclei, p53 nuclear accumulation, and cleaved caspase, thresholds for Hoechst, Cy3, and Cy5 intensities were defined as the average intensity plus 1 SD from 28 vehicle-treated wells placed in the center of the microplate. Cells were classified as positive if their average Hoechst, Cy3, or Cy5 intensity exceeded this threshold.

FR901464 was not very stable in pH 7.4 phosphate buffer (t_1/2 = 4 hours at 37°C; ref. 9). Although we previously reported the half-life of the right fragment of meayamycin (9), it was not obvious whether the result could be applied to the entire molecule. Thus, meayamycin was incubated at 10 μmol/L in pH 7.4 phosphate buffer at 37°C, and its decomposition was monitored by HPLC analysis. Under these conditions, the half-life was 80 hours (Fig. 2A). When the pH of the solution in which stability was gauged was decreased to 7.4, 7, and 6, meayamycin exhibited progressively longer half-lives. At pH 5, the stability of meayamycin was similar to that at pH 7. A DMSO solution of meayamycin kept at 24°C for 4 weeks then examined by HPLC showed no evidence of the decomposition (data not shown).

More relevant data in terms of a drug's biological effect on cells would be their stability in tissue culture medium. Yoshida and coworkers showed that the half-life of FR901464 in RPMI culture medium containing 10% FBS was only 45 minutes (12). We found the half-life of meayamycin in RPMI containing 10% FBS to be 37 hours (Fig. 2A). Therefore, meayamycin is 50 times more stable than its parent compound and during cell culture experiments, >50% of the initial dose of meayamycin would be presumed to be preserved during the first 24 hours of compound treatment.

Because we have not yet determined the mechanisms by which meayamycin is degraded, the initial-rate kinetics of meayamycin was analyzed under pseudo–first order conditions [−d (meayamycin) / dt = k (meayamycin)]. The rate constants were found to be 1.1 × 10⁻⁶, 0.44 × 10⁻⁶, 1.0 × 10⁻⁶, 1.8 × 10⁻⁶, and 5.2 × 10⁻⁶ s⁻¹ in pH 5 buffer, pH 6 buffer, pH 7 buffer, pH 7.4 buffer, and culture medium, respectively.

We examined antiproliferative properties of meayamycin against various tumor cell lines. The average 50% growth-inhibitory concentrations (GI₅₀) are reported in Table 1. Human breast cancer MCF-7 (a gift from Professor Marc Lippman, Georgetown University, Washington, D.C.) and MDA-MB231 cells were particularly responsive to the inhibitory effects of meayamycin, against which low picomolar GI₅₀s were observed. Human colon cancer HCT-116 and human prostate cancer PC-3 cells were also quite responsive. Human non–small cell lung cancer H1299 and human lung cancer A549 cells showed very respectable mid-picomolar responses toward meayamycin. Human prostate cancer DU-145 cells were slightly less responsive.

The reversibility of growth inhibition by meayamycin was examined to determine whether the epoxide might form a covalent bond with target biomolecule(s). A549 cells were treated with meayamycin at 2 nmol/L, and after the indicated times (Fig. 2B), the compound was removed from some of the wells. The assay was continued with and without meayamycin for total of 4 days, at which point, cell growth was assessed by the MTS assay. An 8-hour exposure of cells to 2 nmol/L of meayamycin was sufficient to irreversibly inhibit growth even after drug removal. We observed similar results with MCF-7 cells and HeLa cells (data not shown)

The data in Fig. 2B suggests that meayamycin exerted its effect on cells within 8 hours. We therefore examined drug concentration–dependent cell proliferation after 8 hours of exposure to meayamycin followed by its removal from the culture medium. If meayamycin covalently binds to intracellular target protein(s) within 8 hours, the growth inhibition curve of the cells treated with meayamycin for 8 hours should be similar to that of cells treated with meayamycin for the entire assay period (25). As Fig. 2C shows, when A549 cells were exposed to meayamycin for 8 hours and incubated for an additional 88 hours in its absence, the concentration-dependent growth inhibition was similar to that observed with meayamycin for 96 hours.

A major challenge in cancer therapy is the development of small molecules that are active against cells that have acquired multidrug resistance (MDR), a phenotype that is typically mediated by the expression of xenobiotic efflux pumps such as P-glycoprotein (26). We used the DC-3F (drug-sensitive)/VCRd-5L (MDR) cell line pair that has been extensively used and is well-described in the literature (27). VCRd-5L cells were rendered resistant to vincristine and show cross-resistance to a variety of agents due to expression of efflux pumps (27). Figure 5 shows that VCRd5L cells were resistant to vincristine, paclitaxel, and to a lesser extent, FR901464, but retained complete sensitivity to picomolar concentrations of meayamycin.

Although FR901464 inhibits pre-mRNA splicing in live cells (12), it was not obvious whether meayamycin, a far more stable analogue of FR901464 (9), would exhibit similar biological activity. As such, HeLa nuclear extracts were treated with meayamycin and 5 at various concentrations and with isoginkgetin, a known inhibitor of pre-mRNA splicing (17) as a comparison (Fig. 3). Meayamycin inhibited pre-mRNA splicing of the AdML pre-mRNA substrate in a concentration-dependent manner and 50 nmol/L was sufficient for complete inhibition (Fig. 3, left). Incubation of the same nuclear extract with up to 10 μmol/L of the nonepoxide analogue 5 had no effect upon splicing. Furthermore, compound 5 showed no antiproliferative activity even at 10μmol/L (data not shown). These data imply that the epoxide functionality in both FR901464 and meayamycin is necessary for both antiproliferative activity and pre-mRNA splicing inhibition. Analysis of the RNA-containing complex on a native gel showed that meayamycin blocked assembly of the spliceosomal A complex (Fig. 3, middle), which is consistent with SF3b inhibition (12).

We then treated HEK-293 cells with meayamycin (10nmol/L) and monitored splicing inhibition over time (Fig. 3, right). In 2 hours, unspliced RNA was noticeable, and in 4 to 7 hours, meayamycin exhibited its full inhibitory activity. This time-dependence is similar to the antiproliferative activity of meayamycin against human cancer cells (Fig. 2B and C), supporting the notion that the antiproliferative activity of meayamycin is mediated by pre-mRNA splicing inhibition.

To extend this analysis, we measured the effects of meayamycin on alternative splicing. Cortical neurons were chosen as the experimental system because at high frequency, alternative splicing produces modular changes in the expression of ion channels and other protein factors as these cells differentiate in culture (28, 29). Neurons were treated with meayamycin for 24 hours in culture, and total RNA was harvested for analysis of endogenous splicing patterns by RT-PCR. Primers were chosen for the RT-PCR analysis to amplify three-exon regions of these transcripts corresponding to the exons of interest. As additional controls, we tested a region of the intronless histone H1F0 transcript, as well as intron-containing regions of known pre-mRNAs.

A group of three alternatively spliced cassette exons from the glutamate NMDA R1 receptor, GRIN1 (exon C1), the potassium channel KCNQ2 (exon 8), and the splicing regulatory factor, HNRPH3 (exon 3) showed no change in their splicing patterns upon treatment with meayamycin (Fig. 4A). That is, the sizes of the observed exon-included and exon-skipped products were as expected (30), and the relative ratios of these products did not vary with the presence or absence of meayamycin. Constitutive exons from actin (exon 2) and HNRPH3 (exon 8) also showed no variations with the expected (single) mRNA product. We noticed that the mRNA product levels decreased for some transcripts at a final concentration of 20 pmol/L of meayamycin. A similar decrease was also observed for the intronless H1F0 transcript and for the pre-mRNA fragments containing introns. Nonetheless, the neurons remained viable to the same extent in the presence or absence of meayamycin as measured by trypan blue staining (Fig. 4B). Thus, it is likely that meayamycin induces a decrease in transcription and/or mRNA stability, but does not affect alternative splicing regulation in these cells.

Because the intracellular functions of SF3b are poorly understood, it was not clear what phenotype should be observable when cells are treated with FR901464 class compounds, nor if inhibitors of SF3b would show selectivity for the transformed phenotype. We therefore probed cancer-relevant activities of meayamycin by high-content multiparametric cellular profiling in A549 human lung cancer cells and IMR-90 nontumorigenic human lung fibroblasts. We chose the IMR-90 line because it is well characterized and known to be resistant to a variety of cytotoxic and antisignaling agents (31–34). The IMR-90 cells have a finite life span of approximately 60 population doublings (35) before entering crisis, which we confirmed in our laboratory (data not shown).

Twenty-four hours after treatment, meayamycin consistently caused a more pronounced cell loss in the A549 lung cancer cells compared with normal cells (Fig. 6A and B). Nuclear morphology measurements indicated that meayamycin did not induce chromatin condensation (Fig. 6A and B, open squares), suggesting a nonapoptotic mode of cell death and the absence of mitotic arrest. Consistent with nonapoptotic death, no caspase cleavage was observed (Fig. 6A and B, closed circles). We occasionally observed a low level of caspase cleavage that seemed to be higher in the A549 cells, which accounts for the slightly higher response seen with A549 cell in Fig. 6D, but the effect was not seen consistently and its magnitude was low. Similar results were obtained after 48 hours of treatment (data not shown). Both cell lines responded to meayamycin with p53 induction, although the magnitude of response was smaller in the normal cells (Fig. 6D). The meayamycin-induced increase in p53 was, however, not responsible for antiproliferative activity because wild-type Tp53^+/+ and Tp53^−/− HCT116 cell lines (36) were equally sensitive to meayamycin, with GI₅₀ values of 0.66 nmol/L for wild-type HCT116 and 0.95 nmol/L for p53^−/− HCT116 cells.

Meayamycin consistently caused enlarged nuclei (Fig. 6C), which may be related to splicing inhibition-induced RNA accumulation that results in speckle enlargement (37, 38). Similar results were obtained after a 48-hour exposure to meayamycin (data not shown). Data were repeatable and consistent over the course of several independent experiments (Fig. 6D), and suggested that meayamycin preferentially affected the survival of lung cancer cells compared with nontumorigenic lung fibroblasts but did not cause apoptosis.

Meayamycin was very stable in pH 7.4 phosphate buffer and relatively stable in cell culture medium. The enhanced chemical stability relative to FR901464 is meritorious because unlike FR901464, solutions of meayamycin could be stored even at ambient temperature for months without significant decomposition. The half-life of meayamycin in tissue culture media is useful information to design cell-based experiments and interpret data.

The growth of various tumor cell lines was inhibited by picomolar concentrations of meayamycin. The activity profile of meayamycin is noticeably different from that of pladienolide; for example, although the MCF-7 and HCT-116 cell lines are very sensitive to both pladienolide and meayamycin, the MDA-MB231 cell line is far more affected by meayamycin than by pladienolide (13). Additionally, the A549 and DU-145 cell lines are very sensitive to pladienolide but less so to meayamycin.

Some epoxide-containing anticancer natural products such as fumagillin form a covalent bond with specific proteins in vivo, whereas others such as the epothilones do not (39), which is consistent with the low intrinsic electrophilicity of epoxides under biological conditions (39, 40). Therefore, we asked if meayamycin would act as an electrophile to form a covalent bond with target biomacromolecule(s). This is a testable question because covalent bond formation between drugs and biomacromolecules typically results in irreversible growth inhibition within the experimental time frame (3–5 days), which manifests itself as continuous growth inhibition after removing the drug from the culture medium (25). Noncovalent binding usually results in a return to cell proliferation after drug removal (17).

Biochemical studies by the Yoshida group showed that FR901464-SF3b protein binding may be reversible (12). If FR901464 reversibly binds to a splicing factor and this binding is responsible for the antiproliferative activity, removal of meayamycin at an early stage should have resulted in a return to the typical rate of cell growth. However, transient treatment of A549 (Fig. 2B and C), HeLa, and MCF-7 cells (data not shown) with meayamycin resulted in growth inhibition profiles similar to those seen with cells treated with meayamycin for 4 days. These data suggest that, at least in these cell lines, the binding mode of meayamycin may be irreversible. Alternatively, reversible binding of meayamycin may induce irreversible cellular events that lead to cell growth arrest. Further studies are needed to precisely determine the nature of the interaction of meayamycin with its proposed cellular targets.

Therefore, together with the lack of activity of 4 (5) and 5, there is a line of evidence suggesting that the antiproliferative activity of meayamycin is mediated by biomolecule(s) that attack the epoxide to form a covalent bond.

A striking advantage of meayamycin, aside from its improved potency and stability compared with FR901464, was found to be its potency against an MDR cell line. The improved potency (100-fold) of meayamycin relative to that of FR901464 may be explained by its enhanced stability in cell culture media (56-fold); however, the explanation behind its improved potency against cells with the MDR phenotype is less obvious. A common mechanism of MDR is the overexpression of P-glycoprotein (ABCB1), which often exports aromatic ring–containing small molecules (41). However, other factors play a significant role in MDR mechanisms. For example, when the ester moiety of an epothilone was replaced by an amide bond, the resulting analogue lost cytotoxicity against an MDR cell line (42). Clearly, the minor chemical modifications represented by the change in FR901464 to meayamycin dramatically improved the potency against an MDR cell line. Currently, we do not know what causes the extraordinary ability of meayamycin to inhibit the growth of MDR cells. The interaction of meayamycin with cellular proteins that mediate multidrug resistance should be a focus of future investigations.

The activity of meayamycin in splicing assays in vitro and in live cells was found to be similar to that of 3 and pladienolide D. In in vitro assays, pre-mRNA substrate is initially coated with a heterogeneous mixture of RNA-binding proteins, referred to as the H complex. The first step in spliceosome assembly (the E complex) is the recognition of the 5′ splice site by the U1 small nuclear ribonucleoprotein and recognition of the branchpoint by U2 snRNA and its associated proteins (43). Spliceosomal A complex assembly is the first ATP-requiring step and results in the stabilization of the branchpoint-U2 small nuclear ribonucleoprotein interaction. Addition of the U4/U5/U6 tri-small nuclear ribonucleoprotein, forming the B complex, and subsequent structural rearrangements lead to the catalytically competent C complex spliceosome. Native gel electrophoresis of pre-mRNA transcripts from in vitro splicing reactions shows that a pre-mRNA substrate moves through a progression of complexes as defined by their electrophoretic mobilities. Consistent with interfering with binding and/or function of SF3b, in vitro spliceosome assembly was blocked by meayamycin at its earliest stages. As expected from the SF3b binding, the presence of meayamycin in in vitro splicing reactions prevented the transition from the H complex to the A complex.

The increased complexity in human proteome relative to its genomic counterpart is primarily due to alternative RNA splicing. As such, the regulation of alternative splicing plays pivotal roles in many biological processes and human disease (44, 45). We examined whether meayamycin perturbed alternative splicing in cortical neuron cells. None of the alternative splicing processes that we examined were affected by meayamycin at 20 pmol/L. We observed a decrease in the mRNA level of intronless H1F0, suggesting that meayamycin might induce a decrease in transcription and/or mRNA stability. This observation raises a question about whether the meayamycin-mediated splicing inhibition is less sensitive under in vivo conditions in which splicing is coupled to transcriptional regulation, for which further studies are needed (46).

Although meayamycin does not perturb alternative RNA splicing in the system we examined, this does not exclude the possibility that this compound interferes with other splicing regulatory systems. For example, SAP155, one of several possible targets for FR901464 (12), binds to ceramide-responsive RNA cis-element 1 and regulates the alternative splicing of Bcl-x (47). Therefore, meayamycin may regulate the alternative splicing of Bcl-x.

The toxicity of meayamycin in nontumorigenic cells was addressed; the result was encouraging in that the compound did not show significant toxicity against IMR-90 human lung fibroblasts. At this point, it is not clear whether such cell type specificity can be explained by splicing inhibition. Nonetheless, this result should warrant future drug development. Meayamycin had a unique profile of cellular activities. Exposure to meayamycin resulted in nonapoptotic toxicity that was selective for transformed cells as measured by cell loss, chromatin condensation, and caspase cleavage. Consistent with previous reports using FR901464 (1, 2), meayamycin did not induce mitotic arrest (Fig. 6; data not shown). We also examined alterations in the levels of the p53 protein, an indicator for DNA damage. As Fig. 6A and B illustrate, meayamycin concentration-dependently increased the level of p53 protein in A549, and to a lesser extent, inIMR-90 cells. This is interesting because FR901464 has been reported to lower the mRNA level of p53, potentially via splicing inhibition (2). It is possible that p53 protein degradation is inhibited because the p53 mRNA impairs Mdm2 through direct binding, resulting in increased levels of p53 protein (48). The meayamycin-induced increase in p53 was, however, not responsible for antiproliferative activity because wild-type and p53^−/− HCT116 cell lines were equally sensitive to meayamycin.

No potential conflicts of interest were disclosed.

We thank Professor Stephen Weber (University of Pittsburgh) for sharing his Spectromax M2 plate reader with us; we also thank Professors Bert Vogelstein (Johns Hopkins) and Lin Zhang (University of Pittsburgh) for sharing their p53^−/− cell line with us.