Characterization of ATP-independent ERK inhibitors identified through in silico analysis of the active ERK2 structure

Abstract

The extracellular signal-regulated kinases (ERK1 and ERK2) are important mediators of cell proliferation. Constitutive activation of the ERK proteins plays a critical role in the proliferation of many human cancers. Taking advantage of recently identified substrate docking domains on ERK2, we have used computer-aided drug design (CADD) to identify novel low molecular weight compounds that interact with ERK2 in an ATP-independent manner and disrupt substrate-specific interactions. In the current study, a CADD screen of the 3D structure of active phosphorylated ERK2 protein was used to identify inhibitory compounds. We tested 13 compounds identified by the CADD screen in ERK-specific phosphorylation, cell proliferation, and binding assays. Of the 13 compounds tested, 4 compounds strongly inhibited ERK-mediated phosphorylation of ribosomal S6 kinase-1 (Rsk-1) and/or the transcription factor Elk-1 and inhibited the proliferation of HeLa cervical carcinoma cells with IC(50) values in the 2-10 microM range. These studies demonstrate that CADD can be used to identify lead compounds for development of novel non-ATP-dependent inhibitors selective for active ERK and its interactions with substrates involved in cancer cell proliferation.

Figure 1

Superimposed structures of the unphosphorylated (green) and phosphorylated (purple) forms of ERK2. A. Superimposed ribbon image showing the location and conformational changes associated with the ATP binding domain, activation site, and the ED and CD domains. B. Superimposed ribbon image in the vicinity of the CD (Asp 316 and 319) and ED (Thr 157 and 158) domains.

Figure 2

Effects of test compounds on ERK substrate phosphorylation. HeLa cells were pre-treated with or without 100 μM of the test compounds listed for 15 minutes prior to the addition of epidermal growth factor (EGF or E) to stimulate the ERK pathway or anisomycin (A) to stimulate p38 MAP kinase. Cells were harvested and the extracted proteins were subjected to immunoblot analysis. A. Immunoblots of Elk-1 phosphorylated on S383 (pElk-1), α-tubulin, and active ERK1/2 (ppERK1/2). The lower graph shows the quantification of the ratio of pElk-1 to α-tubulin as measured by densitometry. The pElk-1 phosphorylation in the presence of the test compounds was compared to the EGF only treatment, which was set at 100%. B. Immunoblots of Rsk-1 phosphorylated on T573 (pRsk-1) and α-tubulin as a loading control. The lower graph shows the quantification of the ratio of pRsk-1 to α-tubulin as measured by densitometry. The pRsk-1 phosphorylation in the presence of the test compounds was compared to the EGF only treatment, which was set at 100%. C. Immunoblot analysis of phosphorylated ATF2 (pATF2) and α-tubulin as a loading control. The lower graph shows the quantification of the ratio pATF2 to α-tubulin as measured by densitometry. The pATF2 phosphorylation in the presence of the test compounds was compared to the anisomycin (A) only treatment, which was set at 100%. The data were reproduced in at least 3 independent experiments.

Figure 2

Effects of test compounds on ERK substrate phosphorylation. HeLa cells were pre-treated with or without 100 μM of the test compounds listed for 15 minutes prior to the addition of epidermal growth factor (EGF or E) to stimulate the ERK pathway or anisomycin (A) to stimulate p38 MAP kinase. Cells were harvested and the extracted proteins were subjected to immunoblot analysis. A. Immunoblots of Elk-1 phosphorylated on S383 (pElk-1), α-tubulin, and active ERK1/2 (ppERK1/2). The lower graph shows the quantification of the ratio of pElk-1 to α-tubulin as measured by densitometry. The pElk-1 phosphorylation in the presence of the test compounds was compared to the EGF only treatment, which was set at 100%. B. Immunoblots of Rsk-1 phosphorylated on T573 (pRsk-1) and α-tubulin as a loading control. The lower graph shows the quantification of the ratio of pRsk-1 to α-tubulin as measured by densitometry. The pRsk-1 phosphorylation in the presence of the test compounds was compared to the EGF only treatment, which was set at 100%. C. Immunoblot analysis of phosphorylated ATF2 (pATF2) and α-tubulin as a loading control. The lower graph shows the quantification of the ratio pATF2 to α-tubulin as measured by densitometry. The pATF2 phosphorylation in the presence of the test compounds was compared to the anisomycin (A) only treatment, which was set at 100%. The data were reproduced in at least 3 independent experiments.

Figure 2

Effects of test compounds on ERK substrate phosphorylation. HeLa cells were pre-treated with or without 100 μM of the test compounds listed for 15 minutes prior to the addition of epidermal growth factor (EGF or E) to stimulate the ERK pathway or anisomycin (A) to stimulate p38 MAP kinase. Cells were harvested and the extracted proteins were subjected to immunoblot analysis. A. Immunoblots of Elk-1 phosphorylated on S383 (pElk-1), α-tubulin, and active ERK1/2 (ppERK1/2). The lower graph shows the quantification of the ratio of pElk-1 to α-tubulin as measured by densitometry. The pElk-1 phosphorylation in the presence of the test compounds was compared to the EGF only treatment, which was set at 100%. B. Immunoblots of Rsk-1 phosphorylated on T573 (pRsk-1) and α-tubulin as a loading control. The lower graph shows the quantification of the ratio of pRsk-1 to α-tubulin as measured by densitometry. The pRsk-1 phosphorylation in the presence of the test compounds was compared to the EGF only treatment, which was set at 100%. C. Immunoblot analysis of phosphorylated ATF2 (pATF2) and α-tubulin as a loading control. The lower graph shows the quantification of the ratio pATF2 to α-tubulin as measured by densitometry. The pATF2 phosphorylation in the presence of the test compounds was compared to the anisomycin (A) only treatment, which was set at 100%. The data were reproduced in at least 3 independent experiments.

Figure 3

Chemical structures of the thirteen compounds tested in this study.

Figure 4

Effects of test compounds on ERK2 fluorescence. A. Fluorescence titrations were done with test compounds 92 (open circles), 93 (closed squares), 94 (open squares), and 95 (closed circles). B. Fluorescence titrations with test compounds 86 (closed circles), 89 (closed squares) and 98 (open circles). The percentage of ERK2 fluorescence (F) was plotted against the log concentration in moles/liter (Log [M]) of each test compound using the peak fluorescence in the absence of test compound set at 100% as the reference.