A Toolbox for the Generation of Chemical Probes for Baculovirus IAP Repeat Containing Proteins

Abstract

E3 ligases constitute a large and diverse family of proteins that play a central role in regulating protein homeostasis by recruiting substrate proteins via recruitment domains to the proteasomal degradation machinery. Small molecules can either inhibit, modulate or hijack E3 function. The latter class of small molecules led to the development of selective protein degraders, such as PROTACs (PROteolysis TArgeting Chimeras), that recruit protein targets to the ubiquitin system leading to a new class of pharmacologically active drugs and to new therapeutic options. Recent efforts have focused on the E3 family of Baculovirus IAP Repeat (BIR) domains that comprise a structurally conserved but diverse 70 amino acid long protein interaction domain. In the human proteome, 16 BIR domains have been identified, among them promising drug targets such as the Inhibitors of Apoptosis (IAP) family, that typically contain three BIR domains (BIR1, BIR2, and BIR3). To date, this target area lacks assay tools that would allow comprehensive evaluation of inhibitor selectivity. As a consequence, the selectivity of current BIR domain targeting inhibitors is unknown. To this end, we developed assays that allow determination of inhibitor selectivity in vitro as well as in cellulo. Using this toolbox, we have characterized available BIR domain inhibitors. The characterized chemical starting points and selectivity data will be the basis for the generation of new chemical probes for IAP proteins with well-characterized mode of action and provide the basis for future drug discovery efforts and the development of PROTACs and molecular glues.

Keywords: E3 Ligase; IAP; NanoBRET; PROTAC; Ubiquitin.

FIGURE 1

(A) domain structure of the BIRC family E3 ligases displaying conservation of the BIR3 domain (yellow) and the RING domain (grey). (B) domain structure of the residual BIRC family members showing no conservation. (C) phylogenetic analysis of the BIR domains from all BIRC proteins showing clustering for the BIR1 (green), BIR2 (orange) and BIR3 (yellow) domains together with some ungrouped domains (blue). BIRC, Baculovirus IAP Repeat containing; BIR, Baculovirus IAP Repeat; UBA, Ubiquitin-associated; CARD, Caspase recruitment domain; RING, really interesting new gene; CC, Coiled coil; NACHT, NAIP, CIITA, HET-E und TEP1; WH, Winged helix; HD, Helical domain; UBC, Ubiquitin-conjugating; LLR, Leucine-rich repeat.

FIGURE 2

Exemplary results of the biophysical assays performed on the BIR3 domain of BIRC2. (A) and (B) FP assay curves of the compound titrations Data were expressed as mean ± SD (n = 3). (A) titration curve of A 410099.1 with an estimated IC₅₀ of 5.4 nM and the calculated K _I of 2.0 nM. (B) titration curve of Birinapant with an estimated IC₅₀ of 14.2 nM and the calculated K _I of 4.8 nM. (C) and (D) ΔRFU/ΔT plotted against the temperature obtained from DSF (n = 3). (C) melting point determination for A 410099.1 (blue dotted line) in comparison to the DMSO control (red dotted line). (D) melting point determination for Birinapant treated protein in comparison to the DMSO control. (E,F) ITC results of A 410099.1 and Birinapant with µcal/s plotted against the time in the upper frame and kcal/mol of injection plotted against the molar ratio in the bottom frame with a K _D, n, ΔH, TΔS and ΔG given (n = 2).

FIGURE 3

NanoBRET tracer titrations of the different BIRC constructs. (A) Determination of the binding affinities of the tracers to the different BIRC constructs. Data were expressed as mean ± SEM using two independent experiments performed in duplicates (n = 4). (B) Tracer potency across the BIRC family. Here, BIRC6 was not tested and the BIRC1-BIR1 domain did not show a significant luciferase signal and therefore, both constructs were excluded (italic). (C) Tracer titration curves for BIRC2 full-length proteins and its corresponding BIR domains (n = 4). (D) Structure of the NanoBRET IAP tracer.

FIGURE 4

NanoBRET cellular target engagement assay using the full-length and single domain proteins. Full-length BIRC2, BIRC3, BIRC4, BIRC7, BRIC8, and BIRC1-BIR2, BIRC2-BIR2, BIRC2-BIR3, BIRC3-BIR2, BIRC3-BIR3, BIRC7-BIR and BIRC8-BIR were investigated. Data were expressed as mean ± SEM using two independent experiments performed in duplicates (n = 4). (A) Potency of CUDC-427 against tested BIRC domains. The excluded assays for BIRC6-BIR and BIRC1-BIR1 are shown in italic. (B,C) Normalized NanoBRET ratio [%] against the tested concentration of CUDC-427 and A 410099.1, respectively. (D) Potency of most potent compound A 410099.1 against tested BIRC domains. The excluded assays for BIRC6-BIR and BIRC1-BIR1 are expressed in italic. (E,F) chemical structures of CUDC-427 and AZD5582 which mark starting points towards BIRC7 selective inhibitors. (G) Heat map for compound potencies towards the corresponding protein constructs. Bivalent compounds are marked (*).