J. McNulty, et al.
Bioorganic&MedicinalChemistryLetters30(2020)127559
Fig. 1. Structure of current approved anti-herpesvirus drugs acyclovir 1, vala-
cyclovir 2 and novel non-nucleoside hits 3–5, anti-Dengue 6 and antimalarial 7
acylhydrazones.
Fig. 2. Anti-viral assessment of the initial set of analogues 21–31 in HSV-1
infected iPSC neurons.
compounds, such as the indole derivative 25 and the highly substituted
trimethoxy-derivative 27 proved more active. Nonetheless, the con-
clusion derived from the SAR on the initial array was that no derivative
exhibited potency close to 4F17, indicating the requirement of a small,
electron-rich heterocycle on the hydrazone portion of the molecule.
These results with the array of aryl derivatives in comparison to the
furano-substituted parent compound 5, focussed our attention on the
preparation of a second generation of analogues containing small,
electron-rich heterocyclic rings. This second-generation compound
array was readily accessed from the penultimate hydrazide 13 con-
densing with a range of commercially available heterocyclic aldehydes,
Scheme 1. Synthesis of library target 4F17 (5) proceeding via homogeneous
copper (II) catalysed cyclopropanation with 20, hydrazinolysis and hydrazone
condensation with furfural.
furfural 14 under acid catalysis to give synthetic 4F17, compound 5.
Analysis of synthetic 5 by NMR showed the pure molecule to exist as a
3:1 mixture of rotamers 5:8, identical to that observed for the com-
mercial library product 4F17, thus confirming the structure and stabi-
lity and validating the initial anti-viral activity described for this
compound. Synthetic 5 also exhibited anti-HSV1 activity indis-
tinguishable from the library compound 4F17.
in the same assay as described above and the overall results presented
in Fig. 3. The 2-thiofurano analogue 32 showed substantial activity,
similar to the positive control acyclovir. To our delight, the corre-
sponding 2-pyrrolo analogue 33 proved highly potent. Interestingly, the
two constitutionally isomeric imidazoles 34 and 35 proved significantly
less active, indicating that a basic nitrogen atom is detrimental, perhaps
due to solubility and/or basicity of the side chain. In order to probe
basicity further, we prepared the corresponding 2-pyridyl analogue 36
which proved highly potent, and the isomeric 3-pyridyl derivative 37
which was less active. The bromofurano analogue of 5, compound 38,
proved less active than 5. Overall, the second-generation structure-an-
tiviral assessment allowed identification of the 2-pyrrolo- and 2-pyr-
idinyl- substituted compounds 33 and 36 as being significantly and
slightly more potent that acyclovir and the parent 2-furano derivative 5
itself.
The synthetic route developed toward 5 outlined in Scheme 1 was
antiviral activity investigations following our late-stage (last step) di-
versity-oriented approach that has proven successful elsewhere.11 The
first array of hydrazone derivatives was readily accessed from the pe-
nultimate intermediate 13 through condensation with a range of alde-
hydes and ketones to generate the acyl hydrazones 21–31 shown in
iPSC-neurons following the previously developed protocol.5 The assay
read-out is based on a recombinant HSV-1 strain derived from the KOS
virus expressing the enhanced green fluorescent protein (EGFP) from
the early ICP0 promoter and a monomeric red fluorescent protein (RFP)
from the glycoprotein C promoter.6 Uninfected cells as well as those
treated with the recombinant HSV-1 in the presence of the positive
control acyclovir (ACV) or synthetic 4F17 exhibited a low proportion of
fluorescent cells, while those infected with HSV-1 alone or any of the
initial compound screening set 21–31, demonstrated relatively low
inhibition of viral infection, with compounds 25, 27 and 29–31
showing slight inhibition.6
These results set the stage for the preparation of the third generation
of compounds that would allow probing the contribution of the 1,1-
diarylmethano fragment to the anti-viral activity. Strategically, the
synthetic approach followed the original synthetic protocol, allowing
access to arrays through late-stage incorporation of 2-furanyl, 2-pyridyl
and 2-pyrrolyl substituents. The synthesis of the required substituted
1,1-diarylethylenes required is shown in Scheme 2. The reaction of a 4-
substituted benzaldehyde 39 or 40 with the corresponding 4-
Overall, the structure–activity correlation of the variously sub-
stituted benzaldehyde derived hydrazones disclosed no linear electronic
effect, however, on the whole, electron poor analogues were seen to be
somewhat less active (e.g. 23, 28). Conversely, electron rich
2