S.M. Daskalova, et al.
Bioorganic&MedicinalChemistry28(2020)115642
targeting the DNA binding affinity of the enzyme, disrupting Pol β-
dependent DNA repair complexes, affecting post-translational mod-
ifications critical for enzyme function or directly inhibiting Pol β lyase
or polymerase activities. Lyase inhibitors might well be thought to hold
the greatest promise for specificity, given both the uniqueness of this
domain, and the obligatory intermediacy of a normally transient lyase-
DNA covalent binary complex.
of distinguishing the nature of binding interactions at the lyase active
site, we prepared and characterized three analogues of Lys72, namely
homolysine (III), p-aminophenylalanine (IV) and p-aminomethylphe-
nylalanine (V), that would reposition the side chain amine by about
0.1–1.3 Å yet retain similar nucleophilicity (Fig. 3A, Table 1). The ly-
catalytic pocket other than repositioning of the lysine side chain amine,
providing modified enzymes that retained significant lyase activity.
Additionally, we included N-methyl-lysine (II) as a positive control in
addition to wild type. The structural similarity of II to lysine (I), with
identical Cβ-Nε distance and slightly altered nucleophilicity and steric
bulk of the ε-amino group, was anticipated to have little impact on the
lyase activity of the modified enzyme. In comparison, substitution of
Lys72 with valine (VI) should afford an enzyme with no lyase activity,
thereby fulfilling the role of a negative control.
While numerous inhibitors of the lyase activity have been reported,
only limited work has been done to characterize the nature of the
binding mechanism(s) which conduce to effective and selective in-
hibition. Perhaps more surprisingly, little has been done to identify
inhibitors that engage the catalytic nucleophile in the lyase active site,
i.e. Lys72. Such inhibitors might be expected to affect the formation or
stability of the normally transient covalent protein-DNA binary com-
plex. Accordingly, a key goal of this study was to identify Pol β lyase
inhibitors that engage the primary catalytic nucleophile of Lys72 as part
of the inhibitor Pol β complex. To achieve that outcome, we developed
a strategy predicated on comparing the effect of Pol β inhibitors on the
wild-type enzyme vs modified enzymes containing lysine analogues at
position 72. An inhibitor having the desired characteristic of interacting
with the Lys72 side chain amine to form contacts critical for Pol β
binding would be expected to decrease the lyase activity of the wild-
type enzyme, but have a diminished (or no) effect on the activity of the
modified enzyme. This is illustrated in Fig. 1 in which inhibitor I1
Pol β enzymes, while inhibitor I2 affects only the wild-type enzyme,
putatively by interaction with the side chain amino group of Lys72.
Presently, we describe our study of the interaction of five Pol β lyase
inhibitors (Fig. 2) with wild-type and modified enzymes having Nε-
pounds 2–5 are newly reported inhibitors of Pol β. One of these in-
hibitors (cis-9,10-epoxyoctadecanoic acid (4)) exhibited binding to the
enzyme critically dependent on the presence of the primary catalytic
nucleophile of Lys72. A less pronounced, but qualitatively analogous
observation was also made for 3,4-dihydroxybenzoic acid (3).
The actual preparation of each of the polymerases was carried out as
described in Fig. 4. As illustrated in the Fig. for homolysine, N-pente-
transcript lacking the 3′-terminal pCpAOH moiety by the use of T4 RNA
by brief treatment with aqueous iodine.34,35 Each of the misacylated
tRNAs was used in a coupled in vitro transcription/translation of rat Pol
β from a DNA template having an amber stop codon corresponding to
position 72 of Pol β. The enzymes so derived were purified by Ni-NTA
affinity chromatography, then analyzed by sodium dodecyl-poly-
acrylamide gel electrophoresis (SDS-PAGE). The detailed procedures
have been described previously.36
The six amino acid analogues shown in Fig. 3A were incorporated
(Fig. 3B). As expected, the N-methyllysine-modified enzyme demon-
strated lyase activity similar to the wild-type enzyme, about 86%.
Among the other enzymes prepared, only the homolysine-substituted
Pol β (III) showed moderate lyase activity, ~35% of the wild-type Pol β
levels. The enzymes having IV and V at position 72 had only minimal
lyase activity at best (Fig. 3C), and were not investigated further. Based
on these results, we employed wild-type enzyme expressed in vivo as a
control and, in addition, in vitro synthesized enzymes having II or III at
position 72 to study the effects of compounds 1–5. The enzyme mod-
ified with II allowed the steric effects of a small substituent to be as-
sessed, and also verified that the method of enzyme production did not
affect the results. Homolysine-substituted Pol β was the only suitable
choice for a modified enzyme capable of distinguishing the nature of
binding interactions at the lyase active site, as a consequence of having
the NH2 group of its side chain repositioned, relative to that in wild
type.
The predominant mode of binding of the cis-epoxide of oleic acid
(4), was studied further by in silico docking, and the results obtained
provided additional support for the experimental findings. While none
of the compounds studied here is known to be selective for inhibition of
Pol β, nor is any potent enough to be employed for the inhibition of the
enzyme under physiological conditions, the compounds do provide
proof of principle for a new strategy that can be used to identify
structural motifs that may be developed further as useful inhibitors.
2. Results and discussion
To measure the activity of wild type and modified Pol β enzymes,
we employed an assay described previously in which a synthetic DNA
In order to identify a Pol β analogue modified at position 72 capable
Fig. 1. Inhibition of wild-type Pol β and modified Pol β containing homolysine at position 72 in lieu of lysine by two hypothetical inhibitors. Inhibitor I1 affects the
conversion of substrate to product by both enzymes; inhibitor I2 affects only wild type.
2