Letter
one above except that uPAATF was placed on the microtiter plate.
Compound 5 was incubated with surface-immobilized uPAATF
for 30 min, and after several washing steps uPAR was added to
the wells followed by detection to measure the extent of uPAR
binding to immobilized uPAATF. The compound did not inhibit
the protein−protein interaction, suggesting that 5 was unlikely
inhibiting the protein−protein interaction by binding to uPA
directly (Figure 3B).
ATF
Figure 1. Chemical structure of 1 (IPR-2992).
Although our compounds do not seem to possess any
moieties that are likely to react with nucleophiles on uPAR, we
nevertheless used the fluorescence-based thiol-reactive assay
substituent of 1 with other larger groups (Table 1). These
compounds were tested using a fluorescence polarization (FP)
assay that uses a previously described fluorescently labeled α-
helical peptide AE147 that binds to the uPAATF binding site
(
MSTI) to rule out covalent bond formation with cysteine. As
expected, 5 and its derivatives did not react toward the activated
nucleophile (Figure 3C). Next, we explored the possibility that
the compound may inhibit through an aggregation mechanism.
It is worth pointing out that the FP, ELISA or BLI curves do not
suggest aggregation as an inhibition mechanism. The inhibition
curves span two-log units of concentration. To rule out
aggregation, we repeated the inhibition studies with increasing
levels of detergent Triton X-100. Increased levels of detergent
are known to disrupt aggregates. If the compounds inhibit
through aggregation, then the activity of the compounds should
be eliminated at higher detergent levels. However, increasing the
levels of Triton X-100 in the ELISA assay from 0.01% to 0.1%
did not affect the inhibition profile of 5 (Figure 3D). We also
resorted to dynamic light scattering (DLS), another common
approach to detect the presence of large particles. At higher
concentrations of 5, we did observe formation of large particles
in assay buffer containing PBS and 0.01% Triton X-100. The
aggregation was detectable at concentrations of 12.5 μM or
higher (Figure 3E), with heterogeneous particle sizes being
detected with diameters in the range of 150−2000 nm and high
polydispersity (17−30%). The diameter and the polydispersity
increased with increasing concentration of compound. The
count rate also increases with increasing concentration of
compound (Figure 3F). However, the concentration at which
large particles are observed, 12.5 μM, is 25-fold higher than the
IC50 of the compound in the ELISA assay. This along with the
fact that a 10-fold increase in Triton X-100 levels showed no
effect on compound activity suggests that aggregation is unlikely
the mechanism by which the compound inhibits uPAR binding
to uPAATF. To further establish the specificity of compound 5,
we tested whether the compound inhibited the interaction
(Figure 2A). We also tested the compounds using a microtiter-
based enzyme-linked immunosorbent assay (ELISA), which,
unlike the FP assay, includes the entire uPAR·uPA protein
interface (Figure 2B). For the FP assay, two compounds, 5 and 6
(
IPR-3243), revealed concentration-dependent inhibition, but
the inhibition curves plateaued at about 40% (Figure 2A and
Table 1). For the ELISA, two compounds exhibited complete
inhibition of the uPAR·uPA interaction within the concentration
ranges considered. Remarkably, compound 5 inhibited with
submicromolar IC of 0.5 ± 0.0 μM (Table 1). The IC of 5 in
50
50
this assay improved by 280-fold over the parent 1, which was
25
reported at 140.6 ± 19.0 μM. Compound 5 did not interfere
with the fluorescence of the AE147-FAM peptide (SI; Figure
S1). Interestingly, replacing the thiophene of 5 with a phenyl
moiety as in compound 2 (IPR-3238) resulted in substantially
higher IC of 13.9 ± 1.5 μM. A phenylethyl group in 4 (IPR-
5
0
3
241) did not inhibit the interaction, and the phenylethynyl and
isoindolinedione substituents in 3 (IPR-3239) and 6, led to high
double-digit IC s, 50.0 ± 6.0 μM and 73.8 ± 9.7 μM,
5
0
respectively (Table 1). To confirm the ELISA results, we
resorted to biolayer interferometry (BLI) in competition mode.
Compound 5 inhibited the uPAR·uPAATF interaction in a
concentration-dependent manner with an IC of 8.6 ± 0.2 μM
50
(
Figure 2C and D).
Biochemical Studies to Explore Potential Nonspecific
Inhibition. The unusually high potency of 5, especially
considering its relatively low molecular weight and flat structure,
prompted us to explore the possibility that the compound may
be inhibiting through a nonspecific mechanism. First, the
compound was tested for inhibition of other protein−protein
interactions. Compound 5 was tested against two unrelated
protein−protein interactions, namely, TEA Domain Tran-
scription Factor 4 (TEAD4) and Yes-Associated protein 1
28
the vitronectin binding site of uPAR (SI; Figure S2). The
antibody binds tightly to uPAR with a K of 20.6 ± 1.1 nM. BLI
d
sensors with immobilized 8B12 were dipped into wells
containing uPAR preincubated with increasing concentration
of compounds 5 or 8. Neither 5 nor 8 (IPR-3430) inhibited the
interaction between uPAR and 8B12. If the compounds were
aggregating the protein or nonspecifically inhibiting uPAR, we
would expect inhibition in this assay. We used microscale
pounds 5, 8, and 9 (IPR-3432) to uPAR (SI; Figure S3).
(
Yap1), and voltage-gated calcium channel 2.2 beta subunit
(Cav2.2 ) and the autoinhibitory domain (AID), for which we
β3
26,27
have established FP-based assays.
Both interactions are
high-affinity protein−protein interactions with dissociation
equilibrium constants in the nanomolar range, and they both
occur over a large interface. The TEAD4·Yap1 interaction is
devoid of a pocket and is considered a tertiary interaction as the
interface involves multiple secondary structures of Yap1 binding
Compound 9 showed good binding to uPAR with a K of 13.0 ±
d
to the surface of TEAD4. The Cav2.2 ·AID interaction is
3.0 μM. We did not detect binding of 5 and 8 to apo uPAR,
possibly because these compounds may bind with higher affinity
to uPAR when it is in complex with uPA. The binding of 9
provides evidence of direct engagement of uPAR by these
compounds.
Design and Synthesis of Derivatives of 5. Another 11
derivatives (7−17) were prepared based on the scaffold of 5,
with the thiophene group on the quinoline core (Table 1). As
β3
considered a secondary interaction as the interface consists of an
α-helix (AID) binding to a pocket on the Cav2.2 subunit. The
β3
compounds did not inhibit either the binding of TEAD4 to Yap1
or Cav2.2 to the AID peptide (Figure 3A).
β3
Next, we wanted to eliminate the possibility that the
compound binds to uPA instead of uPAR to inhibit the
protein−protein interaction. We used an ELISA similar to the
6
1
ACS Med. Chem. Lett. 2021, 12, 60−66