Bianca Sperl

Discovery of chromone-based inhibitors of the transcription factor STAT5.

Molecular signals originating at the cell surface are conveyed by a complex system of interconnected signaling pathways to the nucleus. They converge at transcription factors, which in turn regulate the transcription of sets of genes which ultimately determine the cellular phenotype. Whereas enzymes involved in signaling pathways, that is, intracellular kinases and phosphatases and receptor tyrosine kinases, have been recognized and exploited as intervention points for modulating cellular properties with small organic molecules, transcription factors are often considered “nondruggable” because of their lack of enzymatic activities. However, as many transcription factors require interactions with themselves or other proteins, cell-permeable inhibitors of protein–protein interactions could provide an approach towards the inhibition of this important class of proteins, and would thereby allow for the analysis of transcription factor functions and for therapeutic intervention of diseased states. 3] Initially regarded as unfeasible, a growing body of evidence indicates that the inhibition of protein–protein interactions can be potently and selectively achieved by drug-like molecules, some of which are even undergoing clinical trials. STATs (signal transducers and activators and transcription) are a family of transcription factors, which require their Src-homology 2 (SH2) domain at two steps of the signaling process to be active. Firstly, STATs need to bind via their SH2 domain to activated receptors and nonreceptor tyrosine kinases (NRTKs), and can subsequently be phosphorylated at a conserved tyrosine residue C-terminal of their SH2 domain (Scheme 1). Secondly, upon tyrosine phosphorylation, STATs dissociate from the respective receptor or NRTK, and form dimers via reciprocal interactions between their SH2 domains and the sequences surrounding the phosphorylated tyrosine residue. Therefore, a small molecule which inhibits the protein–protein interactions mediated by the SH2 domain of STATs could inhibit STAT functions efficiently (Scheme 1). Direct inhibition of STATs is less likely to result in unintentional inhibition of additional signaling pathways than the targeting of upstream kinases. Two STAT family members, STAT3 and STAT5, have been recognized as therapeutic targets for many human tumors. 16] We have recently identified a small-molecule inhibitor of STAT3, which acts by selectively inhibiting the function of the STAT3 SH2 domain, thus validating the outlined approach toward STAT inhibition. Two isoforms of STAT5 exist, dubbed STAT5a and STAT5b, which are 93% identical at the amino acid level. STAT5 is overactive in several kinds of leukemias, and also in breast cancer, uterine cancer, prostate cancer, and squamous cell carcinoma of the head and neck (SCCHN). As the inhibition of signaling by STAT5 has been shown to inhibit tumor growth and to induce apoptosis of tumor cells, direct inhibition of the STAT5 protein would be desirable to help dissect and counteract the role of STAT5 in cancer. Smallmolecule inhibitors of STAT5 could furthermore be useful tools to clarify the relevance of STAT5 for various cellular processes in genetically unmodified systems. Despite the significant interest in small-molecule inhibitors of STAT5, to the best of our knowledge, nonpeptidic molecules which inhibit the function of the STAT5 SH2 domain have not been published to date. To identify organic molecules which can inhibit the function of the SH2 domain of STAT5, we used a homogeneous assay based on fluorescence polarization which monitors binding of the peptide 5-carboxyfluorescein-GYACHTUNGTRENNUNG(PO3H2)LVLDKW, which is derived from the erythropoietin (EPO) receptor, to the SH2 domain of STAT5b. Screening of diverse chemical libraries consisting of a total of 17298 molecules for compounds which disrupt the interaction between STAT5b and its binding peptide led to the identification of the chromone-derived acyl hydrazone 1 (Table 1, apparent IC50=47 17 mm). The functions of the SH2 domains of STAT3, STAT1, and of the tyrosine kinase Lck were inhibited to a lesser extent (Table 1, Figure 1). Scheme 1. Simplified model of STAT signaling induced by activated cytokine receptors. The signaling steps indicated by the dashed arrows (phosphorylation and dimerization) could be inhibited by an inhibitor of the SH2 domain of STAT family members.

Selective Inhibition of c-Myc/Max Dimerization by a Pyrazolo[1,5-a]pyrimidine

The c-Myc proto-oncogene is involved in many human tumors, and needs to bind to its activation partner Max for all of its known biological activities. Inhibition of the protein–protein interactions between c-Myc and Max by cell-permeable molecules is therefore an attractive goal. Dimerization between c-Myc and Max occurs via a-helical domains comprising leucine zipper motifs, which display no obvious binding sites for inhibitory ligands. We recently identified two pyrazolo ACHTUNGTRENNUNG[1,5a]pyrimidines which inhibit c-Myc/Max dimerization from a diverse collection of chemicals (Figure 1a). These compounds, dubbed Mycro1 and Mycro2, inhibited c-Myc/Max dimerization and DNA binding with preference over other structurally related transcription factors in vitro, and exhibited c-Myc dependent effects in cellular assays. To explore the chemical space around the pyrazoloACHTUNGTRENNUNG[1,5-a]pyrimidine core structure for substitution patterns which are associated with activity against cMyc/Max dimerization, and to possibly identify a Myc/Max dimerization inhibitor with improved properties, we screened a 1438-membered pyrazoloACHTUNGTRENNUNG[1,5-a]pyrimidine library based on the structures of Mycro1 and Mycro2 (Figure 1b and Figure S1 in the Supporting Information). As c-Myc can bind DNA only as dimer with Max, we tested the compounds for their abilities to inhibit DNA binding of cMyc in a fluorescence polarization assay. Five test compounds (1–5) inhibited DNA binding of c-Myc/Max with preference over Max/Max DNA binding by more than 50% at a concentration of 100 mm (Table 1). DNA binding of Max/Max dimers is the most stringent specificity control possible, as cMyc and Max are 59% similar at the protein level in the dimerization domains, and the overall structure of the DNA-bound dimers are very similar to each other. As an additional specificity control, we analyzed the effect of compounds 1–5 on the function of the Src-homology 2 (SH2) domain of the structurally unrelated transcription factor STAT3. None of the compounds affected the interactions between STAT3 and a phosphotyrosine-containing peptide comprising the STAT3 binding motif to a major extent (Table 1). Confirmation of compound activities in well-controlled cellular systems would provide a strong argument against the notion that any in vitro data could be influenced by factors which are irrelevant under cellular conditions. In addition, cellular assays allow the analysis of compound specificities against all relevant proteins; therefore, their scope is incomparably larger than any in vitro analysis. We chose a cell proliferation assay to further analyze the effects of inhibitor candidates 1–5. Cell cycle progression and proliferation of almost all cell types, including U-2OS osteosarcoma cells, requires c-Myc function, and hence c-Myc’s ability to bind to its activation partner Max. However, for reasons not yet fully understood, PC-12 pheochromocytoma cells proliferate independent of cMyc/Max dimerization, as they express a truncated Max protein which is unable to interact with c-Myc. Therefore, a selective inhibitor of c-Myc/Max dimerization can be expected to inhibit proliferation of the c-Myc/Max-dependent U-2OS cells, without inhibiting the growth of the c-Myc/Max-independent PC-12 cells, provided it is cell permeable and stable in the cellular environment. In contrast, compounds which act unspecifically in the cellular context will either cause a reduction in the proliferative rate of both U-2OS and PC-12 cells, or display toxic effects. Whereas compound 5 appeared to be toxic in both cell lines, compound 4 inhibited proliferation of the cell lines only to a minor extent at 20 mm (Figure S2 in the Supporting Information), possibly for reasons related to cellular uptake or intracellular stability. In contrast, compounds 2 and 3 inhibited proliferation of U-2OS cells with good selectivity over proliferation of PC-12 cells at 10 mm (Figure S2 in the Supporting Information), even though compound 3 was toxic at 20 mm. The best selectivity was observed with compound 1, which strongly inhibited proliferation of U-2OS cells (70% reduction of cell number after 5 days, Figure 2a), but had no significant effect on the proliferation of PC-12 cells (Figure 2b). These data sugFigure 1. a) Structures of the c-Myc/Max dimerization inhibitors Mycro1 and Mycro2. b) Structural diversity elements of the pyrazolopyrimidine library, and classification of substitution patterns present in the library. All compounds carrying furane or thiophene as R contain CF3 or CF2Cl as substituent X.

Natural product inhibitors of protein–protein interactions mediated by Src-family SH2 domains

In this Letter, we report the natural products salvianolic acid A, salvianolic acid B, and caftaric acid as inhibitors of the protein-protein interactions mediated by the SH2 domains of the Src-family kinases Src and Lck, two established disease targets. Moreover, we propose a binding mode for the inhibitors based on molecular modeling, which will facilitate chemical optimization efforts of these important lead structures for drug discovery.

Selective Targeting of Disease-Relevant Protein Binding Domains by O-Phosphorylated Natural Product Derivatives

Phosphorylation-dependent protein binding domains are crucially important for intracellular signaling pathways and thus highly relevant targets in chemical biology. By screening of chemical libraries against 12 structurally diverse phosphorylation-dependent protein binding domains, we have identified fosfosal and dexamethasone-21-phosphate as selective inhibitors of two antitumor targets: the SH2 domain of the transcription factor STAT5b and the substrate-binding domain of the peptidyl-prolyl isomerase Pin1, respectively. Both compounds are phosphate prodrugs with documented clinical use as anti-inflammatory agents in humans and were discovered with a high hit rate from a small subgroup within the screening library. Our study indicates O-phosphorylation of appropriately preselected natural products or natural product derivatives as a generally applicable strategy for the identification of non-reactive and non-peptidic ligands of phosphorylation-dependent protein binding domains. Moreover, our data indicate that it would be advisable to monitor the bioactivities of clinically used prodrugs in their uncleaved state against phosphorylation-dependent protein binding domains.

Germline variant FGFR4 p.G388R exposes a membrane-proximal STAT3 binding site

Variant rs351855-G/A is a commonly occurring single-nucleotide polymorphism of coding regions in exon 9 of the fibroblast growth factor receptor FGFR4 (CD334) gene (c.1162G>A). It results in an amino-acid change at codon 388 from glycine to arginine (p.Gly388Arg) in the transmembrane domain of the receptor. Despite compelling genetic evidence for the association of this common variant with cancers of the bone, breast, colon, prostate, skin, lung, head and neck, as well as soft-tissue sarcomas and non-Hodgkin lymphoma, the underlying biological mechanism has remained elusive. Here we show that substitution of the conserved glycine 388 residue to a charged arginine residue alters the transmembrane spanning segment and exposes a membrane-proximal cytoplasmic signal transducer and activator of transcription 3 (STAT3) binding site Y390-(P)XXQ393. We demonstrate that such membrane-proximal STAT3 binding motifs in the germline of type I membrane receptors enhance STAT3 tyrosine phosphorylation by recruiting STAT3 proteins to the inner cell membrane. Remarkably, such germline variants frequently co-localize with somatic mutations in the Catalogue of Somatic Mutations in Cancer (COSMIC) database. Using Fgfr4 single nucleotide polymorphism knock-in mice and transgenic mouse models for breast and lung cancers, we validate the enhanced STAT3 signalling induced by the FGFR4 Arg388-variant in vivo. Thus, our findings elucidate the molecular mechanism behind the genetic association of rs351855 with accelerated cancer progression and suggest that germline variants of cell-surface molecules that recruit STAT3 to the inner cell membrane are a significant risk for cancer prognosis and disease progression.

The Natural Product Betulinic Acid Inhibits C/EBP Family Transcription Factors

Transcription factors are convergence points in cellular signaling. They receive and integrate signals originating at the cell surface, which are conveyed by a complex system of interconnected signaling pathways to the nucleus. In response to these signals, transcription factors regulate the transcription of sets of genes that ultimately determine the cellular phenotype. Their central position makes them highly desirable targets for functional modulation by cell-permeable organic molecules. Unfortunately, transcription factors are challenging targets, which are frequently considered to be “non-druggable”, because their inhibition would require interfering with large protein–DNA or protein–protein interactions. Nevertheless, it has recently been demonstrated that direct inhibition of dimeric transcription factors (that is, transcription factors that need to form homodimers or heterodimers in order to bind DNA) by small organic molecules is indeed achievable. The most common approaches used in these studies involve the in vitro or in silico screening of diverse chemical libraries, mostly consisting of non-natural products without previous assignment of protein targets, against a single, preselected dimeric transcription factor. Stimulated by these success stories, we wondered whether known bioactive molecules—that is, natural products, molecular tools used in cell biology, and drugs/drug candidates for human usage—might also affect the activities of transcription factors. Analysis of the activity of a known bioactive molecule against various transcription factors could lead to a more complete appreciation of the compound’s target spectrum, and thus allow a more comprehensive understanding of the compound’s cellular mode of action. In addition, because natural products have been demonstrated to be excellent starting points for drug discovery projects, there exists the real potential that a known compound could be identified as a lead structure for a hitherto unrecognized target. In order to explore to what extent known bioactive small molecules target transcription factors, we assayed the 480member ICCB screening library of known bioactive compounds (BioMol) against a panel of seven transcription factors. The screening library was chosen because it contains important known bioactive small molecules whilst being readily manageable by academic laboratories due to its modest size. We assayed the screening library against seven disease-relevant dimeric transcription factors in fluorescence polarization assays. DNA binding assays were used for dimeric transcription factors that associate through extensive a-helical domains, for the two basic helix–loop–helix leucine zipper (bHLHZip) protein dimers c-Myc/Max and Max/Max, and also for the two basic leucine zipper protein dimers Jun/Jun and C/EBPa/C/EBPa. This assay type detects inhibitors of DNA binding of intact transcription factor dimers and also inhibitors of the protein–protein interactions between the two transcription factor subunits, because homodimerization or heterodimerization of the transcription factor subunits is a prerequisite for their DNA binding. The second group of assays employed binding between the SH2 domains of the transcription factors STAT1, STAT3, and STAT5b and their respective phosphotyrosine-containing peptide recognition motifs, and thus investigates protein–protein interactions, not protein–DNA binding. In order to reduce the proportion of false-positive hits frequently observed in high-throughput screens, we performed the initial screen at two concentrations (15 mm and 45 mm) and only followed up compounds that displayed dose-dependent inhibition. Test compounds that inhibited a single target by more than 35 % at 15 mm and by more than 70 % at 45 mm in a dose-dependent manner were defined as active. Cross examination of the active compounds’ activities against all targets investigated in this study and manual retesting of selected screening hits led to the identification of the pentacyclic triterpene betulinic acid (1; Scheme 1) as an inhibitor of C/EBPa. In view of the prominent role of C/EBP family proteins as small-molecule targets against adipogenesis and skin tumors, we set out to investigate the activity of betulinic acid in more detail. Betulinic acid (1) was found in the fluorescence polarization assay to inhibit DNA binding of the bZip protein C/EBPa with an IC50 value of 13.4 0.2 mm (Figure 1 A). This activity lies well within the wide range of IC50 values (1–30 mm) reported for betulinic acid activity in tumor cell lines. DNA binding of Jun homodimers from the same protein family was affected to a significantly lesser extent (ca. 40 % inhibition at the highest concentration tested (50 mm)). DNA binding of c-Myc/Max dimers was also inhibited to a lesser extent (IC50 = 39 12 mm). In contrast, DNA binding of Max/Max homodimers was not affected. Because the DNA sequences used for the c-Myc/Max and Max/Max assays are identical (the E-box motif CACGTG), this argues against the possibility that betulinic acid exerts its effect through DNA intercalation, consistently with a previous report. In addition, neither DNA binding of NF-kB p50 homodimers nor that of NF-kB p65 homodimers was significantly [a] Dr. M. Gr ber, Prof. Dr. T. Berg Institute of Organic Chemistry, University of Leipzig Johannisallee 29, 04103 Leipzig (Germany) E-mail : tberg@uni-leipzig.de [b] A. Hollis, B. Sperl, Prof. Dr. T. Berg Department of Molecular Biology, Max Planck Institute of Biochemistry Center for Integrated Protein Science Munich (CIPSM) Am Klopferspitz 18, 82152 Martinsried (Germany) Supporting information for this article is available on the WWW under http ://dx.doi.org/10.1002/cbic.201100652.

Binding of phosphorylated peptides and inhibition of their interaction with disease-relevant human proteins by synthetic metal-chelate receptors

The modulation of biological signal transduction pathways by masking phosphorylated amino acid residues represents a viable route toward pharmacologic protein regulation. Binding of phosphorylated amino acid residues has been achieved with synthetic metal‐chelate receptors. The affinity and selectivity of such receptors can be enhanced if combined with a second binding site. We demonstrate this principle with a series of synthetic ditopic metal‐chelate receptors, which were synthesized and investigated for their binding affinity to phosphorylated short peptides under conditions of physiological pH. The compounds showing highest affinity were subsequently used to inhibit the interaction of the human STAT1 protein to a peptide derived from the interferon‐γ receptor, and between the checkpoint kinase Chk2 and its preferred binding motif. Two of the investigated ditopic synthetic receptors show a significant increase in inhibition activity. The results show that regulation of protein function by binding to phosphorylated amino acids is possible. The introduction of additional binding sites into the synthetic receptors increases their affinity, but the flexibility of the structures investigated so far prohibited stringent amino acid sequence selectivity in peptide binding. Copyright