Ligands with short aromatic linkers and large rigid cap groups were docked into the
target binding pocket, and a carbazole cap group was found to fit well. These
compounds were synthesized, tested, and found to be indeed highly selective
HDAC6 inhibitors [58]. More cases of successful structure-based design of HDAC
inhibitors are described in recent reviews [59, 60].
3.2 Sirtuins
Sirtuins, in contrast to the classical zinc-dependent HDACs, use nicotinamide
adenine dinucleotide (NAD
+
) as cofactor to carry out the deacetylation step. For
most of the seven human sirtuin (Sirt) isoforms, crystal structures have already been
solved. The reported crystal structures of human Sirt1 [61, 62], Sirt2 [63–66], Sirt3
[65, 67–75], Sirt5 [70, 74, 76–78], and Sirt6 [79], as well as sirtuins derived from
other species [80–88] shed light onto the overall structure of sirtuins. Sirt2 has been
extensively studied by crystallographers, who solved the 3D structure of this isoform
in the apo, cofactor-bound, substrate bound, as well as in inhibited form. So far, no
3D structure of human Sirt4 and Sirt7 is available. A recent review focused on the
structural details of sirtuins has been published, and the interested reader is referred
to [89]. Sirtuins contain a conserved 275 amino acid catalytic domain with variable
N- and C-termini. The structure of the catalytic domain consists of the so-called
Rossmann fold and a smaller zinc-binding domain. The sirtuin structure shows
several subpockets that accommodate adenine (A), ribose (B), and nicotinamide
(C) as well as the acetyllysine substrate peptide (Fig. 4a). Upon binding of the
substrate peptide, the zinc-binding domain rotates toward the Rossmann fold domain
and induces a “closure” of the active cleft. This closure of both domains helps to
correctly orientate the modified lysine group in the hydrophobic tunnel, allowing the
formation of a hydrogen bond between the Nε atom of lysine and the backbone
carbonyl of a conserved valine residue (Val233, Sirt2 numbering). A further conformation of sirtuins that has been recently reported is the “locked-open conformation” that is observed upon binding of the allosteric and highly selective Sirt2
inhibitor SirReal2 [90] (Fig. 4b). Upon binding, SirReal inhibitors induce a major
rearrangement of Sirt2 active site and open up a so far unexploited binding pocket.
This unique rearrangement is the basis for the high potency and isotype selectivity of
the SirReals. This pocket was therefore termed “selectivity pocket.”
Several inhibitors are available for Sirt2 that have been discovered by applying
computer-based methods [91–93]. One of the first potent Sirt2 inhibitors, the
vinylnitrile derivative AGK2 (13, Fig. 5), has been identified by focused-library
screening [94], and the interaction of the inhibitor with Sirt2 has been analyzed by
docking studies. The docking of the active inhibitors was carried out into different
conformations of Sirt2 that were generated by homology modeling. The high
flexibility of the active site loop made this strategy necessary [95]. The detailed
analysis of the docking results of AGK2 and inactive analogs resulted in the
Structure-Based Design of Epigenetic Inhibitors
463
target binding pocket, and a carbazole cap group was found to fit well. These
compounds were synthesized, tested, and found to be indeed highly selective
HDAC6 inhibitors [58]. More cases of successful structure-based design of HDAC
inhibitors are described in recent reviews [59, 60].
3.2 Sirtuins
Sirtuins, in contrast to the classical zinc-dependent HDACs, use nicotinamide
adenine dinucleotide (NAD
+
) as cofactor to carry out the deacetylation step. For
most of the seven human sirtuin (Sirt) isoforms, crystal structures have already been
solved. The reported crystal structures of human Sirt1 [61, 62], Sirt2 [63–66], Sirt3
[65, 67–75], Sirt5 [70, 74, 76–78], and Sirt6 [79], as well as sirtuins derived from
other species [80–88] shed light onto the overall structure of sirtuins. Sirt2 has been
extensively studied by crystallographers, who solved the 3D structure of this isoform
in the apo, cofactor-bound, substrate bound, as well as in inhibited form. So far, no
3D structure of human Sirt4 and Sirt7 is available. A recent review focused on the
structural details of sirtuins has been published, and the interested reader is referred
to [89]. Sirtuins contain a conserved 275 amino acid catalytic domain with variable
N- and C-termini. The structure of the catalytic domain consists of the so-called
Rossmann fold and a smaller zinc-binding domain. The sirtuin structure shows
several subpockets that accommodate adenine (A), ribose (B), and nicotinamide
(C) as well as the acetyllysine substrate peptide (Fig. 4a). Upon binding of the
substrate peptide, the zinc-binding domain rotates toward the Rossmann fold domain
and induces a “closure” of the active cleft. This closure of both domains helps to
correctly orientate the modified lysine group in the hydrophobic tunnel, allowing the
formation of a hydrogen bond between the Nε atom of lysine and the backbone
carbonyl of a conserved valine residue (Val233, Sirt2 numbering). A further conformation of sirtuins that has been recently reported is the “locked-open conformation” that is observed upon binding of the allosteric and highly selective Sirt2
inhibitor SirReal2 [90] (Fig. 4b). Upon binding, SirReal inhibitors induce a major
rearrangement of Sirt2 active site and open up a so far unexploited binding pocket.
This unique rearrangement is the basis for the high potency and isotype selectivity of
the SirReals. This pocket was therefore termed “selectivity pocket.”
Several inhibitors are available for Sirt2 that have been discovered by applying
computer-based methods [91–93]. One of the first potent Sirt2 inhibitors, the
vinylnitrile derivative AGK2 (13, Fig. 5), has been identified by focused-library
screening [94], and the interaction of the inhibitor with Sirt2 has been analyzed by
docking studies. The docking of the active inhibitors was carried out into different
conformations of Sirt2 that were generated by homology modeling. The high
flexibility of the active site loop made this strategy necessary [95]. The detailed
analysis of the docking results of AGK2 and inactive analogs resulted in the
Structure-Based Design of Epigenetic Inhibitors
463