Application of a Substrate-Mediated Selection with c-Src Tyrosine Kinase to a DNA-Encoded Chemical Library

Abstract

As aberrant activity of protein kinases is observed in many disease states, these enzymes are common targets for therapeutics and detection of activity levels. The development of non-natural protein kinase substrates offers an approach to protein substrate competitive inhibitors, a class of kinase inhibitors with promise for improved specificity. Also, kinase activity detection approaches would benefit from substrates with improved activity and specificity. Here, we apply a substrate-mediated selection to a peptidomimetic DNA-encoded chemical library for enrichment of molecules that can be phosphorylated by the protein tyrosine kinase, c-Src. Several substrates were identified and characterized for activity. A lead compound (SrcDEL10) showed both the ability to serve as a substrate and to promote ATP hydrolysis by the kinase. In inhibition assays, compounds displayed IC_50's ranging from of 8-100 µM. NMR analysis of SrcDEL10 bound to the c-Src:ATP complex was conducted to characterize the binding mode. An ester derivative of the lead compound demonstrated cellular activity with inhibition of Src-dependent signaling in cell culture. Together, the results show the potential for substrate-mediated selections of DNA-encoded libraries to discover molecules with functions other than simple protein binding and offer a new discovery method for development of synthetic tyrosine kinase substrates.

Keywords: DNA-encoded chemical library; c-Src; protein tyrosine kinases; substrate-mediated selection.

Scheme 1

Construction of DNA encoded library with 5.5 × 10⁵ unique members.

Figure 1

Src kinase substrate selection scheme.

Figure 2

Cubic plot analysis of DNA-encoded chemical libraries (DEL) selection. (a) The highest 200 enriching members of the library encoded by the initial 48 codons (A library) are plotted indicating the synthon numbers of the first (x-axis), second (y-axis), and fourth (z-axis) chemical steps. (b) The highest 300 enriching members of the library encoded by the initial 48 codons (A library) that contained the phenol cap in the fourth step are plotted indicating the synthon numbers of the first (x-axis), second (y-axis), and third (z-axis) chemical steps. Color of the points corresponds to the log₁₀ of the enrichment relative to the unselected library as indicated in the color bar scale.

Figure 3

Validation of SrcDEL10 as a hit molecule from the substrate-mediated selection. (a) on-DNA binding affinity to c-Src measured by fluorescence polarization in the presence of ATP (100 µM). A SrcDEL10 modified oligo was hybridized to a fluorescein amide (FAM) modified complementary oligo for the binding assay. (b) Off-DNA HPLC analysis of SrcDEL10 (200 µM) with treatment of ATP (1 mM) and c-Src at high levels (1 µM, yellow, and 10 µM, red), which gave small amounts of the phosphorylated molecule. Peak identity was verified by LC/MS analysis (Figure S9).

Figure 4

Synthesized derivatives of SrcDEL10.

Figure 5

Structure activity relationship (SAR) studies of SrcDEL10 derivatives. (a) ADP-Glo assays of the SrcDEL10 derivative, SrcDEL10–6, lacking a phenolic oxygen show comparable activity to SrcDEL10. Derivative SrcDEL10–8, which contains the phenol but lacks the cyclopropyl group, shows no activity. (b) Analysis of enrichment in the substrate selection (shown in Scheme 1) shows the phenol is required for enrichment. (10-NE, SrcDEL10 with no enzyme treatment).

Figure 5

Structure activity relationship (SAR) studies of SrcDEL10 derivatives. (a) ADP-Glo assays of the SrcDEL10 derivative, SrcDEL10–6, lacking a phenolic oxygen show comparable activity to SrcDEL10. Derivative SrcDEL10–8, which contains the phenol but lacks the cyclopropyl group, shows no activity. (b) Analysis of enrichment in the substrate selection (shown in Scheme 1) shows the phenol is required for enrichment. (10-NE, SrcDEL10 with no enzyme treatment).

Figure 6

Inhibition of hit molecules in a radiolabel assay with ³²P-ATP (50 µM) and biotinylated-LYNtide (50 µM) as substrates. (a) Half maximal inhibitory concentration (IC₅₀) for six hit structures (structures shown in Table 1). (b) Representative data for SrcDEL10 and SrcDEL21.

Figure 7

NMR analysis of SrcDEL10 binding. (a) Overlay of ¹⁵N-HSQC spectra of the phosphorylated Src catalytic domain in the presence of ATP (green) or ATP and SrcDEL10 (red). Ligand concentrations are near saturation based on estimated K_m values. A portion of the spectrum showing significant changes is highlighted. See Figure SX for full spectrum. (b) A portion of the spectrum shows significant chemical shift perturbation of K321 in the presence of SrcDEL10. (c) Chemical shifts with large changes upon addition of SrcDEL10 are mapped (red) to the ribbon drawing of the Src kinase domain (PDB ID: 1Y57) [58].

Figure 8

Inhibition of Src-dependent kinase signaling in EGFR-transformed mammary epithelial cells (NME) cells by SrcDEL10. NME cells were serum deprived for 24 h in the presence of vehicle control (DMSO), Src family inhibitor PP2 (10 µM), or SrcDEL10-ester (100 µM) for the indicated amounts of time. Where indicated, these cells were subsequently stimulated with epidermal growth factor (EGF) (50 ng/mL) for an additional 30 min. Equal protein aliquots were analyzed by immunoblot with the indicated antibodies. Data are representative of at least 3 independent experiments.