i. PubChem Bioassay Name(s), AID(s), Assay-Type (Primary, DR, Counterscreen, Secondary)

ii. Assay Rationale & Description

This XIAP dose response assay was developed and performed to confirm hits originally identified in the XIAP HTS binding assay (AID-1018) and to study the structure-activity relationship on analogues of the confirmed hits. Compounds were acquired from commercial sources or synthesized in-house.

The rationale for the assay was based on the disruption of fluorescence polarization resulting from binding of a his-tagged-BIR1-BIR2 (bacoloviral IAP repeat, “Bir1/2”) domain protein derived from two of the three conserved caspase binding “BIR” domains of XIAP to a rhodamine tagged 7-mer N-terminal SMAC peptide.

iii. Summary of Results

We developed a binding assay based upon fluorescence polarization, using a short peptide representing residues from the N-terminus of activated SMAC with an attached fluorochrome. This fluorescence polarization assay (FPA) formed the basis for a high-throughput competitive displacement assay that was optimized for screening. We screened the NIH compound library using this FPA to identify chemical compounds that compete with the SMAC peptide for binding to IAPs. A secondary assay was also devised, so the hits would be independently confirmed. However, the single hit from the conventional HTS screen was very weak (> 25 μM IC₅₀), we therefore decided to take a structure-based rational design approach to develop novel chemical probes of XIAP. Chemistry has made significant progress towards this goal and the results are described in section 3c below.

i. Description of SAR & chemistry strategy that led to the probe

The design and synthesis of novel antagonists of XIAP that mimic the biological properties of the SMAC peptide has attracted considerable attention in the last few years. Short peptides (3–7 amino acids) typically demonstrate superior affinity for the BIR3 domain of XIAP to the detriment of its binding affinity for the BIR2 domain. With this in mind, most research to date has focused on the search for small molecules that disengage XIAP from caspase-9 by inhibiting the BIR3 interaction. Figure 1 represents the evolution of the SAR studies that we undertook, via chemical synthesis assisted with a structure-based approach, to develop new BIR2-selective SMAC peptidomimetics. X-ray crystal structures of the BIR2 (PDB code 1i3o) and BIR3 (PDB code 1g73) domains of XIAP were used to guide the structure-based molecular design. One key difference between the BIR2 and BIR3 protein structures at the SMAC binding site is that BIR3 has a sub-pocket on the right (Fig.2B) occupied by the isoleucine (Ile) residue of the SMAC N-terminal AVPI sequence, whereas in Bir2 this region is a shallow surface (Fig.2A). This difference in the SMAC binding site therefore forms a rational basis for designing selective Bir2 antagonists. As a result, various substitutions at R4 group in Figure 1 were designed to specifically target this non-conserved binding regions in BIR2 and BIR3. In addition, two complementary approaches drove our synthetic efforts: (1) studies to determine the effect of different substituents on the potency and selectivity of inhibitors of the BIR1/2 vs. BIR3 domains of XIAP; and (2) studies to determine the effect of the stereochemistry of amino acids in P1–3 on the potency and selectivity of inhibitors of the BIR1/2 vs. BIR3 domains.

Figure 1

SAR strategy for tripeptide 1.

Figure 2Protein surfaces of BIR2 (A) and BIR3 (B) at the SMAC binding site; docked poses of CID25241665 (MLS-0391005) in BIR2 (C) and CID25241639 (MLS-0390982) in BIR3 (D)

We synthesized a library of 47 compounds which were evaluated for binding to the BIR1/2 or BIR3 domains of XIAP by employing fluorescence polarization (FP) assays. The IC₅₀ values for selected inhibitors are summarized in Table 2. Although all the compounds shown here were synthesized employing natural L-amino acids, several derivatives were synthesized using unnatural D-amino acids. The use of these unnatural building blocks in general resulted in inactive compounds (IC₅₀>100 μM) for both the BIR1/2 and BIR3 domains. Compounds CID25241665 (MLS-0391005) and CID25241679 (MLS-0391014), represented by the sequence Ala1-Val2-Pro3, a methyl substituted N-terminus and a quinoline scaffold in the C-terminal capping group, showed high BIR1/2 affinity and selectivity (eight and nine-fold respectively relative to BIR3). This is consistent with our molecular modeling studies. Our docking results suggested that this quinoline group sat close to the side chain of the residue Lys206 in BIR2 and may form favorable H-bonding (Fig. 2C). The substitution of the quinoline fragment for a phenyl hydrazine (CID25241673 (MLS-0391011)) showed an almost three-fold improved binding affinity for BIR1/2, albeit losing most of the selectivity that characterizes the two aforementioned compounds. Interestingly, the (R)-tetrahydronaphthyl amide derivative CID25241639 (MLS-0390982) exhibited an extraordinary affinity for BIR3 with an IC₅₀= 0.28 μM. The substitution of the Val2 residue for the amino acids Ile2 or Abu2 afforded (R)-tetrahydronaphthyl analogues CID25241667 (MLS-0391006) and CID25241659 (MLS-0391002) respectively. Although this modification translated into a small erosion of their binding affinities, both compounds still exhibited very high affinity and selectivity for the BIR3 domain of XIAP. Our modeling results showed that the (R)-tetrahydronaphthyl group of MLS-0391002 was likely bound to the sub-pocket of Bir3 in a similar fashion as the ILE residue of the SMAC tetrapeptide AVPI (Fig.2B, D). Compound CID25241687 (MLS-0390979), which contains the sequence Ala1-Val2-Pro3, a non-substituted N-terminus and a diphenylmethane fragment in the C-terminal capping group, follows the same trend as the other three highly BIR3 selective analogues.

Table 2

Binding affinity of representative analogues for the XIAP BIR1/2 and BIR3 domains.

To help validate our synthetic approach, we decided to undertake the challenging synthesis of a previously reported bivalent Smac peptidomimetic, which represents the current state-of-the-art for inhibitors that bind to the BIR2 domain of XIAP. To this end we completed the synthesis of the bivalent inhibitor CID25181204 (MLS-0390866) and this was tested in the FP assays (Figure 3). As shown in Figure 3, the affinity values obtained were not as potent (or selective) as the results for the best new inhibitors summarized in Table 2, suggesting that our new inhibitors are an improvement over the state-of-the-art compound CID25181204 (MLS-0390866).

Figure 3

Bivalent Smac peptidomimetic CID-25181204 (MLS-0390866).

In summary, the SAR studies carried out around the tripeptide 1 have established that the presence of a quinoline substituent in the C-terminal capping group is important for selective binding of compounds to the BIR1/2 domain of XIAP. To our knowledge there is no precedent for this observation in the literature.