V. Zwick et al. / Bioorg. Med. Chem. Lett. xxx (2015) xxx–xxx
3
Scheme 2. Reagents and conditions: (i) Grubbs 1st generation catalyst continuous addition, CH2Cl2, reflux, 6 h; (ii) H2, Pd/C, AcOEt; (iii) TFA, CH2Cl2; (iv) Grubbs 2nd
generation catalyst continuous addition, CH2Cl2, reflux, 7 h; (v) pentenoylchloride, CH2Cl2, NEt3, rt, 12 h.
product. These first results confirmed the difficult metathesis with
such compounds. Optimized conditions were investigated with
compound 4, using 1.25% catalyst added each hour, with 4 to 6
loadings (Table 1, entries 1–3). As shown in Table 1, the more load-
ings realized, the higher yields being obtained. At that point a con-
tinuous catalyst injection was implemented corresponding to 7.5%
catalyst added over 6 h (Table 1, entry 4), leading to a dramatic
yield improvement. This was even better with an additional 1 h
reflux after the final catalyst addition (Table 1, entry 5), without
a noticeable change in the E/Z ratio and with an isolated yield of
70%. In these optimized conditions, compound 5 gave a better con-
version, but with Grubbs 2nd generation catalyst (Table 1, entries 6
and 7). Derivative 6 was not reacting in any of these conditions. In
the article describing metathesis of indoles we used as starting
point, the N–H group of the indole moiety was blocked. In its free
form it is probably interfering with the catalyst and a blocking
group seemed to be necessary. This possibility is currently
explored in our laboratory. The next steps were the reduction of
the double bond and deprotection, performed classically with
hydrogen and palladium on charcoal, followed by trifluoroacetic
acid treatment to remove the BOC groups, affording the final com-
pounds 9 and 10 as white powders.
At this stage we verified the activity of compounds 9 and 10 and
their respective intermediates with a BRET assay in living cells
(Fig. 2, Supplementary information Fig. S1).32 Compound 10, the
benzamide derivative of our SAHA analogues, retained low HDAC
inhibition in cells while its precursors 18 and 19 have no activity
(data not shown). Compound 9 showed good activity as expected
while its precursor 13 was inactive and 14 showed modest activity
(Fig. 2A). No major toxicity of the compounds was observed
demonstrating the reliability of the BRET measurements (Fig. 2B).
The hydroxamic acid function appeared essential to achieve good
activity, compared to the benzamide group. The comparison of
the activities for compounds 9, 13 and 14 indicated that the satu-
rated alkyl chain of compounds 9 and 14 would allow a high
molecular flexibility that contributes to the exposure of polar
groups which may enhance cell membrane permeability. On the
other hand, when blocked by the BOC groups the compound still
retains some activity, although modest. Due to their limited activ-
ities, the benzamide series was not studied further, whereas the
BRET results for 9 and 14 were compared to HDAC inhibition from
nuclear extracts (Table 2), showing some differences for 14 (prob-
ably due to lower cell permeability) and a coherent result for 9.
Because of our interest in identifying selective HDAC inhibitors,
compounds were tested against class I (HDACs 1–3) and class II
(HDAC6) HDACs. Results on the selected isoforms are reported in
Table 3 and compared to SAHA and CI-994. As expected, compound
9 is an effective HDAC inhibitor with a slight selectivity for HDAC6
compared to HDAC1 and 2 (3 and 4-fold respectively), and ten-fold
less active against HDAC3 compared to HDAC1. Interestingly, the
intermediate compound 14 showed micro molar inhibition on
HDAC1, 2 and 6 isoforms with an approximate ten-fold selectivity
for HDAC6. Interestingly, HDAC6 selectivity has been reported in
the past for a series of BOC protected chiral thiol derivatives 3,25
being the BOC group able to selectively interact with HDAC6 rim
channel. Our results are in agreement with these data.
These combined results highlight the difference in identifying
HDAC isoform activities on isolated proteins that are not always
correlated with effective activity in cells. The presence of the dou-
ble bond in 13 seems detrimental to the activity, possibly introduc-
ing geometric constraint and reducing cell permeability, as
hypothesized via MLP calculations33 (Supplementary information).
Compound 4, the precursor of 13, also has no activity suggesting
that to become active, compound 14 would need the two NO-
diBOC groups and to be reduced to an alkane to be more cell-per-
meable or stabilized in the active site.
As some HDAC6 inhibition preference was observed with
compounds 9, and 14, we investigated their effect on tubulin
acetylation. HDAC6 was found to be a microtubule deacetylase34
at the alpha subunit of tubulin in both the dimer and microtubule
polymer forms.35 The reversible post-translational acetylation of
the microtubule system, at residue Lys-40 of the tubulin subunit,
regulates the microtubule functions. We have elaborated a com-
plex test system to characterize the effect of the inhibition of
deacetylases on the acetylation level of the microtubule system
at molecular and cellular levels.36 Experiments were performed
with HeLa cells with deacetylated microtubule network. The addi-
tion of the well-established HDAC inhibitor, trichostatin A (TSA)
resulted in extensive acetylation of the network as detected by
immunofluorescence microscopy using specific anti-acetyl-tubulin
antibody (Fig. 3A). Thus the inhibitory potency of the compounds
were tested by cELISA, an assay of cell level established previ-
ously.37 The concentration-dependent inhibitory effect of the con-
trol compounds (TSA and SAHA), and compounds 9, 13 and 14
rendered it possible to obtain quantitative data for their IC50 val-
ues. The results as shown in Table 4 indicate that the inhibitory
potencies of 9 against the tubulin deacetylase activity of HDAC6
is the strongest, with micromolar activity, followed by the 5-fold
Table 1
Reactivity of compounds 4 and 5 at 50 °C, with 1.25% catalyst loaded each hour
Entry
Compd
Number of loading
Ratio A:AA (E/Z)
1
2
3
4
5
6
7
4
4
4
4
4
5
5
4
5
6
34:66
27:73 (54/46)
23:77 (54/46)
17:83
9:91 (58/42) 70%
91:9, Grubbs I
18:82, Grubbs II
Continuous, 6 h
Continuous, 6 h + 1 h reflux
Continuous, 6 h,+1 h reflux
Continuous, 6 h,+1 h reflux