P. Sivaprakasam, et al.
Bioorganic&MedicinalChemistryLetters30(2020)127531
of the bicyclic core of (S,S)-1 loosely interacts with Gln95 through two
water molecules in the HIV-1 integrase CCD, suggesting the potential
for the introduction of a substituent at this position (i.e. at the C-8
position of the IZP macrocyclic series) that would not negatively impact
the anti-HIV-1 activity. Unlike the HIV-1 integrase CCD structure, the 2-
phenyl moiety of (S,S)-1 is deeply positioned in a hydrophobic pocket
in the PXR LBD. The 2-phenyl ring of (S,S)-1 in the PXR LBD interacts
tightly with Leu239, Pro227, Tyr225, Leu213, Tyr306, and Val211. The
bicyclic core of (S,S)-1 sits above Met243 from the α3-helix and faces
the distal end of the α2-helix in the PXR LBD where there is no ob-
servable electron density for residues Cys207 to Leu209. Grafting this 3-
residue structural element from other PXR LBD structures into the
missing region of the (S,S)-1-bound structure revealed a tight junction
around the core N-4 position in (S,S)-1. This region is fairly solvent
exposed in the HIV-1 integrase CCD co-crystal structure and enabled the
hypothesis that suitably spaced bulky substituents projecting from this
region of 4 would sterically collapse with the distal end of the α2-helix
in the PXR LBD. Upon visualizing the IZP macrocycle 4 with a relatively
small group at this region which is close to the core heterocycle, this
hypothesis was tested by introducing extended, bulky, and polar sub-
stituents at the R position, as captured in Table 1. Taking advantage of
the aldehyde intermediate S19 used in the synthesis of 4a, analogues
4b-d, which incorporate basic amines, were quickly generated (Scheme
1; see Supplementary material for syntheses details). All three of these
compounds demonstrated EC50 values of > 50 µM in the PXR transac-
tivation assay while retaining single digit nM EC50 inhibitory values
toward HIV-1 replication in cell culture. The calculated pKa values
support the possibility that 4b-d exist in the protonated form under the
assay conditions.27 The basic amine moieties installed in 4b-d would
incur destabilizing interactions in the hydrophobic region of the PXR
LBD binding pocket. While binding and desolvation of a positively
charged side chain projecting from the 8-position is expected to be
unfavorable in the interior of the PXR LBD, it is predicted to bind just
fine at the solvent front in the HIV-1 integrase CCD. In fact the C-8
position in the IZP macrocycle chemotype offers an ideal site to install
such solubilizing groups without compromising potency, a vector not
available to the ALLINIs (S,S)-1 and 2. It is believed that the combi-
nation of high polarity and steric bulk at the C-8 position that sterically
destabilizes the α2-helix of the PXR LBD of the IZP macrocyclic core
contributes to attenuating the PXR transactivation liability.
Fig. 2. Overlay of X-ray co-crystal binding modes of (S,S)-1 in the HIV-1 in-
tegrase CCD dimer (orange) and the PXR LBD (green). The figures were pre-
pared using the program Pymol (The Pymol Molecular Graphics system, v.2.3.4
Schrödinger, LLC).
bicyclic core (Fig. 2). In the HIV-1 CCD co-crystal structure, the tert-
butoxy group of (S,S)-1 sits in a well-defined hydrophobic pocket lined
with Thr125, Gln95, Tyr99 and Thr174. Similarly, the tert-butoxy group
of (S,S)-1 occupies a hydrophobic sub-pocket created by Trp299,
Met323, His327, and Gln285 in the PXR LBD structure. These ob-
servations highlight a notable difference between the binding mode of
the carboxylic acid element of (S,S)-1 in the HIV-1 integrase and the
PXR LBD (Fig. 2). The PXR LBD takes advantage of the conformational
flexibility of the carboxylic acid moiety in (S,S)-1 and in the bound
form, this element is flipped upside down whereby the carboxylic acid
and the tert-butoxy groups are favorably positioned in the hydrophilic
and hydrophobic sub-pockets, respectively, relative to its conformation
in the HIV-1 integrase CCD. One could conceive of conformationally
locking the carboxylic acid and the tert-butoxy groups in their observed
states from the HIV-1 integrase CCD structure which would place the
two groups in the opposing sub-pockets in the PXR LBD. However,
synthetically arriving at such conformationally locked versions in this
system would be quite challenging.
In summary, a novel series of macrocyclic IZP-based ALLINIs with
potent HIV-1 inhibitory activity in cell culture is described. Relative to
the previously disclosed ALLINI 2, this new series demonstrates 14 to
23-fold better antiviral potency. However, 4a suffered from a sig-
nificant PXR transactivation liability that would compromise its use as a
therapeutic agent. Using the co-crystal structure complex of a structu-
rally similar ALLINI bound to the PXR LBD, it was hypothesized that the
PXR liability in this series could be abrogated by the judicious in-
troduction of substituents at C-8 of the core heterocycle. The validity of
the design hypothesis was confirmed with the synthesis of several de-
rivatives that presented greatly improved PXR activation profiles in a
The C7-chromane ring of (S,S)-1 occupies a deep hydrophobic
pocket formed by Leu102, Ala129, Thr174, Met178, Trp132 and
Ala169 in the HIV-1 integrase CCD. Similarly, the C7-chromane ring is
caged in a hydrophobic pocket created by Tyr306, Met246, Cys301,
Phe288, and Trp299 in the PXR LBD.
The bicyclic core of (S,S)-1 and the 2-phenyl substituent are well
positioned onto the hydrophobic surface of one of the two subunits of
the dimer in the HIV-1 integrase CCD. The 2-phenyl ring of (S,S)-1
makes van der Waals contact with Ala128 in the α3-helix(124–133) of
one of the subunits in the HIV-1 integrase CCD. The N-4 nitrogen atom
Scheme 1. Synthetic protocol for the synthesis of 4b-d from S19.
4