10.1002/cbic.201800587
ChemBioChem
COMMUNICATION
dihydroxyisocoumarins 25a and 25b commences with a similar strategy.
3,5-Dimethoxybromo benzene 1 is herein converted into 2-bromo-4,6-
dimethoxybenzaldehyde under standard Vilsmeier-Haack conditions.[10]
KMnO4 oxidation[11] and methyl iodide treatment in the presence of
K2CO3 furnishes the desired methyl ester 18. Palladium-catalyzed Stille
coupling reaction with allytributylstannane affords the allylic
intermediate 19 (92% yield) which is subjected to the Lemieux-Johnson
oxidative cleavage[12] to provide smoothly the aldehyde 20. Subsequent
enantioselective Brown allylation[13] delivers the chiral homoallylic
alcohol 21 which partially undergoes spontaneous intramolecular
esterification with the neighboring ester functionality. Final cyclization
of the remaining acyclic material is promoted by Amberlyst-15
catalyst,[14] leading to the allyl-modified isochromanone 22 in a 60%
yield over 2 steps. This compound is finally exposed to a catalytic
amount of Grubbs II complex (5%) in benzene at 45°C to perform a
cross-coupling metathesis reaction with the homoallylic alcohol (S)-(+)-
5-hexen-2-ol, leading to a remarkably clean formation of the E-olefin 23
in 64% yield. At this stage, despite the fact that we were seemingly only
two routine operations away from the final lactone 25, we could not
force the complete stereoselectivity of the ring closure step. Moreover,
given our interest in accessing both diastereoisomers, no particular effort
was further dedicated to guide specifically the stereoselectivity for the
ring closure. Hence, the strategy continues with an iodine-mediated
cyclization allowing the rapid generation of the C5-membered ring
system. The so-formed iodo-intermediate is engaged directly in a
halogen-hydride exchange reaction employing Bu3SnH/AIBN to yield
to a mixture of two diastereoisomers 24a and 24b, highly inseparable.
Only after proceeding to their separation with a reversed-phase chiral
HPLC 24a and 24b were isolated with 99% d.e.[15] Ultimate deprotection
under Maier’s conditions (AlI3, TBAI, and phloroglucinol)[16] leads to
the lactones 25a and 25b, which could be unambiguously characterized
with a fully assigned configuration (See supporting information).
Sonogashira coupling reaction with prop-2-yn-1-ylcyclohexane 26 in
the presence of a catalytic amount of copper iodide and trimethylamine,
leading to the alkynylaryl ester 27 in 80% yield. FeCl3-mediated
cyclization of 27 selectively generated the expected six-membered ring
product which was subjected to BBr3 treatment, to afford the analogue
28 ready for biological evaluation (Scheme 3). [18]
Scheme 3. Third Route – A simplified approach to cladosporin analogue 28.
Reagents and conditions: a) Pd(PPh3)4, CuI, NEt3, DMF, 80%; b) FeCl3, DCM, rt, 18h,
40%; c) BBr3, DCM, rt, 8h, 55%.
Evaluation of biological potency and metabolic stability of the
compound library- With all these analogues in hand, we initiated a
profiling campaign to assess their respective biological
activity/selectivity for different t-RNA synthetases : lysyl, threonine and
phenylalanine as well as their metabolic stability. Follow up analysis,
using the haploinsufficiency profiling (HIP) method and genetic
mutational profiling, [19] confirmed that compounds 9, 10, and 28 were
still on target and the overall lysyl-tRNA binding mode was unchanged
(Supporting information). [20] This key assay allowed not only to assess
activity and selectivity in a cell-based fashion but also to verify that the
tested derivatives also have the potential to penetrate even thick
biological cell walls and membranes, an important requirement for
potential anti-infectives.
Critically, while presenting a relatively close structural similarity
with cladosporin, compounds 9 and 10 (IC50, p.f.LysRS = 0.3; 0.2 µM) were
found several-fold less potent and less selective than the parent
compound cladosporin (Table 1). Compound 25a and 25b did not
present any activity (IC50, p.f.LysRS >10 µM). HIP analysis of compounds
17 and 29 (obtained from commercial source) gave a most surprising
outcome as their respective activity revealed that the observed cell-based
activity was not elicited by inhibition of lysyl-tRNA synthetase, but
rather threonyl- and phenylalanyl-tRNA synthetase, respectively
(Supporting information).
Compound
IC50
IC50
HIP target
S.cerevisiae
with human
Lys.R.S.
(µM)
S.cerevisiae
with P.
falciparum
Lys.R.S.
hypothesis
Scheme 2. Second route proposed, yielding to the isolation of 25a and 25b.
Reagents and conditions: a) POCl3, DMF, 100°C, 4h, 91%, b) KMnO4, H2O, 75°C, 4h,
64%, c) MeI, K2CO3, DMF, rt, 3h, 95%. d) Pd(PPh3)4, LiCl, allyltributylstanane, DMF,
100 °C, 24h, 92%; e) OsO4, NaIO4, dioxan/water (3:1), rt, 5h, 62%; f) (+)-
Ipc2B(allyl)borane, Et2O, –78°C to rt, 2h; g) Amberlyst 15, DCM, rt, 2 days, 60%
(yield over two steps); h) (S)-(+)-5-hexen-2-ol, Grubbs II (5 mol %), benzene, 45 °C,
6h, 64%; i) I2, NaHCO3, toluene, –78 °C rt, 18h; then Bu3SnH, AIBN, benzene, 80 °C,
24a 21%, 24b 21% ( 57% yield over two steps and 42% after chiral separation);
j) Al, I2, TBAI, phloroglucinol, benzene, 30 min, 5 °C, 25a 58%, 25b 42%.
(µM)
cladosporin
100
0.05
Lysyl-tRNA synth.
9
30
15
0.3
0.2
40
Lysyl-tRNA synth.
Lysyl-tRNA synth.
10
17
180
Phenylalanyl-tRNA
synth.
Our third strategy resembles a small pilot synthesis, on a
simplified set of analogues. A representative case is outlined in Scheme
3. In fact, the presence of the isocoumarin ring was known to be vital
but some structural modifications were allowed by the “π- cage” formed
with the aromatic residues. Inspired by previous findings from a set of
compounds generated by biotransformation, we proposed to introduce
a strategic unsaturation within the isocoumarin core as illustrated by
compound 28.[17a]
25a
25b
28
>100
>100
10
>100
>100
0.02
n.d.
n.d.
Lysyl-tRNA synth.
As shown in Scheme 3, our pivotal starting material, methyl 2-
bromo-4,6-dimethoxybenzoate 18 is subsequently engaged in
a
This article is protected by copyright. All rights reserved.