[68]. In particular, incorporation of the hydroxamic acid Zn
2+ -binding group onto
the macrocyclic structure of trapoxin B (TpxB, 6.20) yielded a compound with
cross-class activity and enhanced half-life in blood. This compound was named
cyclic hydroxamic acid-containing peptide (CHAP) 1 (6.21). The optimal separation
between the macrocycle and the chelating group was found to be 5 carbon atoms,
which was further applied for the design of subsequent inhibitors inspired by
compound 6.18, chlamydocin (6.8), cyl-1 (6.14) and cyl-2 (6.16), WF3161 (6.12),
and HC-toxin (6.10) [68, 69]. The same strategy has been followed by others, to the
extent that several naturally occurring macrocyclic HDAC inhibitors have been
modified with the hydroxamic acid moiety, with various effects on in vitro activity
and selectivity (Fig. 6). Modified apicidin (6.1) had already been published by
Merck Research Laboratories few months before Yoshida’s study, showing slight
improvement in in vitro anti-deacetylase activity for the derivative with respect to
the natural compound [47]. Later, in 2007, Ganesan and coworkers synthesized both
the natural and hydroxamic acid-containing version of azumamide E (6.6), which
also exhibited the same trends in potency [48].
In 1993, when the natural product 6.18 was identified as an HDAC inhibitor,
in vitro experiments suggested “slow-binding” kinetics, with decreasing, nonlinear
rates measured over the course of incubation with the inhibitor, only reaching
equilibrium after 1 h incubation with the enzyme. In addition, it was argued that
the epoxide moiety acts as a covalent handle, since HDAC activity was not recovered by dialysis [22]. This experiment was also performed with analogue 6.21, and
restoration of the deacetylase activity indicated that the hydroxamic acid group
interacts in a non-covalent manner with HDACs as expected [68, 69]. However, it
was not investigated whether the binding kinetics of trapoxin-inspired CHAPs also
followed slow-binding profiles. Recent results from Olsen and coworkers demonstrate that hydroxamic acid-containing macrocycles may in fact be able to exhibit
in vitro slow, tight-binding behavior. In particular, compounds 6.4 and 6.21
displayed delayed equilibria in the inhibition of HDAC isotypes 1–3 and 6 [49]. It
is possible to calculate the equilibrium constants (K i ) from kinetic parameters. For
example, estimated K i values for HDAC3 inhibition were reported to be 40 pM and
20 pM for macrocycles 6.4 and 6.21, respectively. In the cases of HDAC1 and
HDAC2 in which the complex dissociation rate is close to 0 (tight binding),
estimation of K i was not possible however. HDAC8 was also potently inhibited by
both hydroxamic acid-containing macrocycles, but, in this case, the mechanism of
inhibition was observed to be fast-on/fast-off (6.4: K i ¼ 3.5 Æ 0.9 nM, 6.21:
K i ¼ 1.4 Æ 0.1 nM) [49]. Moreover, the recently reported demyristoylase activity
of HDAC11 was inhibited by compound 6.4 in a slow, tight-binding manner and by
compound 6.21 with a fast-on/fast-off profile (K i ¼ 24 Æ 1 nM) [17]. This opens the
door for the development of new inhibitors exhibiting isotype-selective inhibition
mechanisms. Further results, including slow, tight-binding inhibition data for macrocycles 6.10 and 6.18, are included in Fig. 6. It is important to mention that the
performance of slow, tight-binding inhibitors such as 6.4 and 6.21 in conventional
end-point experiments yielded apparent K i values up to 30 times higher than those
obtained from continuous assays. These differences result from the fact that
Hydroxamic Acid-Containing Peptides in the Study of Histone Deacetylases
39
2+ -binding group onto
the macrocyclic structure of trapoxin B (TpxB, 6.20) yielded a compound with
cross-class activity and enhanced half-life in blood. This compound was named
cyclic hydroxamic acid-containing peptide (CHAP) 1 (6.21). The optimal separation
between the macrocycle and the chelating group was found to be 5 carbon atoms,
which was further applied for the design of subsequent inhibitors inspired by
compound 6.18, chlamydocin (6.8), cyl-1 (6.14) and cyl-2 (6.16), WF3161 (6.12),
and HC-toxin (6.10) [68, 69]. The same strategy has been followed by others, to the
extent that several naturally occurring macrocyclic HDAC inhibitors have been
modified with the hydroxamic acid moiety, with various effects on in vitro activity
and selectivity (Fig. 6). Modified apicidin (6.1) had already been published by
Merck Research Laboratories few months before Yoshida’s study, showing slight
improvement in in vitro anti-deacetylase activity for the derivative with respect to
the natural compound [47]. Later, in 2007, Ganesan and coworkers synthesized both
the natural and hydroxamic acid-containing version of azumamide E (6.6), which
also exhibited the same trends in potency [48].
In 1993, when the natural product 6.18 was identified as an HDAC inhibitor,
in vitro experiments suggested “slow-binding” kinetics, with decreasing, nonlinear
rates measured over the course of incubation with the inhibitor, only reaching
equilibrium after 1 h incubation with the enzyme. In addition, it was argued that
the epoxide moiety acts as a covalent handle, since HDAC activity was not recovered by dialysis [22]. This experiment was also performed with analogue 6.21, and
restoration of the deacetylase activity indicated that the hydroxamic acid group
interacts in a non-covalent manner with HDACs as expected [68, 69]. However, it
was not investigated whether the binding kinetics of trapoxin-inspired CHAPs also
followed slow-binding profiles. Recent results from Olsen and coworkers demonstrate that hydroxamic acid-containing macrocycles may in fact be able to exhibit
in vitro slow, tight-binding behavior. In particular, compounds 6.4 and 6.21
displayed delayed equilibria in the inhibition of HDAC isotypes 1–3 and 6 [49]. It
is possible to calculate the equilibrium constants (K i ) from kinetic parameters. For
example, estimated K i values for HDAC3 inhibition were reported to be 40 pM and
20 pM for macrocycles 6.4 and 6.21, respectively. In the cases of HDAC1 and
HDAC2 in which the complex dissociation rate is close to 0 (tight binding),
estimation of K i was not possible however. HDAC8 was also potently inhibited by
both hydroxamic acid-containing macrocycles, but, in this case, the mechanism of
inhibition was observed to be fast-on/fast-off (6.4: K i ¼ 3.5 Æ 0.9 nM, 6.21:
K i ¼ 1.4 Æ 0.1 nM) [49]. Moreover, the recently reported demyristoylase activity
of HDAC11 was inhibited by compound 6.4 in a slow, tight-binding manner and by
compound 6.21 with a fast-on/fast-off profile (K i ¼ 24 Æ 1 nM) [17]. This opens the
door for the development of new inhibitors exhibiting isotype-selective inhibition
mechanisms. Further results, including slow, tight-binding inhibition data for macrocycles 6.10 and 6.18, are included in Fig. 6. It is important to mention that the
performance of slow, tight-binding inhibitors such as 6.4 and 6.21 in conventional
end-point experiments yielded apparent K i values up to 30 times higher than those
obtained from continuous assays. These differences result from the fact that
Hydroxamic Acid-Containing Peptides in the Study of Histone Deacetylases
39