UNC1215, the UNC2533-L3MBTL3 co-crystal structure revealed a 2:2 dimer
complex with UNC2533 bridging the L3MBTL3 dimer interface and interacting
with the MBT1 domain of one monomer and the MBT2 domain of the second
(PDB: 4 L59) [182]. When compared to the UNC1215-L3MBTL3 complex, the
two monomers rotated at the dimer interface to accommodate for the smaller size
of UNC2533 rather than altering the conformation of the protein backbone in the
ligand-binding site [182].
A series of UNC2533-based compounds with modifications to both amine
moieties and the two carbon linker between the phenyl and pyrrolidinyl rings were
synthesized and evaluated [182]. Increasing the size of the (pyrrolidinyl)piperidine
moiety or replacing the piperidine with an aliphatic C-2 chain was not well-tolerated
as evidenced by significant decreases in activity against L3MBTL3. Reducing the
ethyl linker to a methylene reduced potency approximately sixfold (IC 50 ¼ 0.35 μM),
while increasing the tether to three carbons maintained potency against L3MBTL3
(IC 50 ¼ 0.07 μM), but did not demonstrate the same binding affinity in ITC studies
(K D ¼ 1.8 μM). The requirement for a basic amine on both sides of the phenyl
ring was highlighted by the 40- (analog #8; numbering as in the original paper)
and 200-fold (analog #11) loss in potencies seen when these moieties were
either modified or removed. Interestingly, without the second basic amine, a
decrease in L3MBTL1 activity was also observed, indicating the significance of
the pyrrolidine ring in binding to this reader. Replacing the pyrrolidine with
an azetidine (analog #14) resulted in a compound with equipotent activity toward
L3MBTL3 (IC 50 ¼ 0.048 μM) and increased selectivity with respect to L3MBTL1
(IC 50 ¼ 86 μM) when compared to UNC2533 [182].
From a SAR perspective, it can be concluded that two basic moieties at each
end of the molecule are necessary to improve potency and selectivity toward
L3MBTL3 due to its unique dimer formation. Furthermore, smaller inhibitors
like UNC2533 highlight the rotational mechanism by which L3MBTL3 may use
to bind specific ligands or methylated histones without altering Kme-binding
pockets. The ability to “tune” molecules for L3MBTL3 inhibition brings promise
to the development of potent inhibitors not only targeting L3MBTL3 but also
other MBT domain-containing effector proteins.
4.5 The PWWP Domain
4.5.1 Structure and Functions
The PWWP domain was named for the Pro-Trp-Trp-Pro motif initially identified
in the Wolf-Hirschhorn syndrome candidate 1 (WHSC1) protein [183]. This motif
is loosely conserved in several other proteins, including over 20 human proteins
[184]. The first (Pro) and second (Trp) residues can vary, while the third (Trp)
and fourth (Pro) are well preserved. As revealed by the first PWWP domain structure
to be determined, from mouse DNA methyltransferase DNMT3B, the PWWP
Methyl-Readers and Inhibitors
371
complex with UNC2533 bridging the L3MBTL3 dimer interface and interacting
with the MBT1 domain of one monomer and the MBT2 domain of the second
(PDB: 4 L59) [182]. When compared to the UNC1215-L3MBTL3 complex, the
two monomers rotated at the dimer interface to accommodate for the smaller size
of UNC2533 rather than altering the conformation of the protein backbone in the
ligand-binding site [182].
A series of UNC2533-based compounds with modifications to both amine
moieties and the two carbon linker between the phenyl and pyrrolidinyl rings were
synthesized and evaluated [182]. Increasing the size of the (pyrrolidinyl)piperidine
moiety or replacing the piperidine with an aliphatic C-2 chain was not well-tolerated
as evidenced by significant decreases in activity against L3MBTL3. Reducing the
ethyl linker to a methylene reduced potency approximately sixfold (IC 50 ¼ 0.35 μM),
while increasing the tether to three carbons maintained potency against L3MBTL3
(IC 50 ¼ 0.07 μM), but did not demonstrate the same binding affinity in ITC studies
(K D ¼ 1.8 μM). The requirement for a basic amine on both sides of the phenyl
ring was highlighted by the 40- (analog #8; numbering as in the original paper)
and 200-fold (analog #11) loss in potencies seen when these moieties were
either modified or removed. Interestingly, without the second basic amine, a
decrease in L3MBTL1 activity was also observed, indicating the significance of
the pyrrolidine ring in binding to this reader. Replacing the pyrrolidine with
an azetidine (analog #14) resulted in a compound with equipotent activity toward
L3MBTL3 (IC 50 ¼ 0.048 μM) and increased selectivity with respect to L3MBTL1
(IC 50 ¼ 86 μM) when compared to UNC2533 [182].
From a SAR perspective, it can be concluded that two basic moieties at each
end of the molecule are necessary to improve potency and selectivity toward
L3MBTL3 due to its unique dimer formation. Furthermore, smaller inhibitors
like UNC2533 highlight the rotational mechanism by which L3MBTL3 may use
to bind specific ligands or methylated histones without altering Kme-binding
pockets. The ability to “tune” molecules for L3MBTL3 inhibition brings promise
to the development of potent inhibitors not only targeting L3MBTL3 but also
other MBT domain-containing effector proteins.
4.5 The PWWP Domain
4.5.1 Structure and Functions
The PWWP domain was named for the Pro-Trp-Trp-Pro motif initially identified
in the Wolf-Hirschhorn syndrome candidate 1 (WHSC1) protein [183]. This motif
is loosely conserved in several other proteins, including over 20 human proteins
[184]. The first (Pro) and second (Trp) residues can vary, while the third (Trp)
and fourth (Pro) are well preserved. As revealed by the first PWWP domain structure
to be determined, from mouse DNA methyltransferase DNMT3B, the PWWP
Methyl-Readers and Inhibitors
371