2
K. Macegoniuk et al. / Bioorganic & Medicinal Chemistry Letters xxx (2017) xxx–xxx
tioned pathogens stimulated research on urease inhibitors21–23 and
design assumption concerning the reaction with the cysteine resi-
due present at the entrance to the active site. Most probably, the
presence of neighboring histidine residue (His323), which could
act as a base, makes the covalent modifications of cysteine residue
reversible.
There are several interesting structure-activity relationships
visible in the study. First, the geometry of the central unsaturated
bond has a significant influence on the activity. E isomers of substi-
tuted double bond- and triple bond-containing compounds are
capable of urease inhibition, whereas Z isomers are completely
inactive against the target enzyme (e.g., 3a and 8a versus 4b or
3h and 8c versus 4a).
2
4–28
led to the discovery compounds such as phosphoramidates,
2
9–32
33–36
hydroxamates,
compounds.
phosphinates,
and
heterocyclic
3
7,38
Currently, only one urease inhibitor available on
39
the market, namely, acetohydroxamic acid, shows significant
side effects, including teratogenicity.40
The active site of urease contains two nickel ions coordinated by
four histidine residues (His137, His139, His249, His275), car-
bamoylated lysine (Kcx220) and aspartate (Asp363).41–43 The
active site is of relatively small volume (related to the size of urea)
and is covered by a movable flap (Fig. S8). This flap contains a cys-
teine residue that could be targeted by inhibitors.44 Currently, only
a limited number of urease inhibitors possess this mode of action,
including benzoquinone, ebselen derivatives, cyclohexanone and
Second, carboxylic acids showed similar, or in some cases even
higher, activity than analogous alkyl esters (e.g., 1a versus 1b, 3a
versus 3h, 8a versus 8c). This finding is of primary importance
because carboxylic acid-containing Michael acceptors are gener-
ally much less chemically reactive towards –SH groups than anal-
ogous esters. Therefore, the high activity of acids is related to their
specific interactions with the enzyme (see the following section).
Third, in most cases, the substituents of phenyl groups or
changes in ester functional groups have minor influence on the
activity (e.g., 1b versus 1d versus 1e or 3a-3f). This observation sug-
gests that these groups do not influence inhibitor binding.
Any compound showing high chemical reactivity with func-
tional groups available in biological systems, in particular –SH
groups, has to be avoided in medicinal chemistry due to unpre-
dictable influence on other biomolecules. Unspecific toxicity is cor-
4
5–50
cyclopentenone.
The product of reaction of urease and benzo-
51
quinone has been recently characterized structurally.
In this paper, we aimed to establish a general methodology to
evaluate the activity-reactivity profile of reversible Michael accep-
tor-type inhibitors based on studies of urease inhibitors. The vari-
ous types of Michael acceptors were extensively screened for their
inhibitory properties against bacterial urease.
The measured inhibitory activities were correlated with the
chemical reactivity (towards a model compound, glutathione) of
the studied compounds to find molecules that exhibit favorable
profiles related to specific interactions with the enzyme. Moreover,
the cytotoxicity (towards mouse fibroblasts BALB/3T3) of the cho-
sen compounds was analyzed. The most important findings were
also rationalized by modelling the structures of the inhibitor-
enzyme complexes.
A group of forty compounds of non-extended, Michael acceptor-
type compounds was selected for screening. These compounds
include molecules containing unsaturated functional groups of
various geometries: E and Z isomers of substituted double bonds
and linear triple bonds or allenes. All groups controlling the chem-
ical reactivity of double/triple bonds contained carbonyl groups
but showed significant differences in activating potency. Ketone
and ester functional groups are considered to activate strongly,
whereas carboxylates are much less reactive. Moreover, analogues
of known inhibitors, cyclopentenone (10a) and cyclohexanone
1
0
related with high reactivity of Michael acceptors. Therefore, the
preferred profile of the Michael acceptor-based active compound
is based on a combination of high inhibitory activity and low reac-
tivity towards thiols. The reactivity of representative compounds
was assayed using glutathione (GSH) as a model compound with
reactive cysteine residues. The modified Ellman method, where
the quantification of unreacted GSH was performed using a spec-
trophotometric-based
concentration-response
assay,
was
5
2
applied. The reactions of the extremely reactive Michael acceptor
compounds (3a, 8a) with GSH are completed within less than
15 min, whereas compound 1b demonstrates no reactivity, even
after 36 h of reaction (examples of progress curves and 1H NMR
spectra of reaction mixtures are shown in Figs. S3–S5, respec-
tively). The tested compounds spanned nearly six orders of magni-
tude of reactivity, with logkGSH values between ꢁ3.138 ± 0.077 and
2.588 ± 0.088 for ethyl cinnamonate (1a) and dimethyl
acetylenedicarboxylate (8a), respectively (Table S4). The depen-
(9a), were also included in the study for comparison.
All compounds were tested against model bacterial urease puri-
TM
fied from Sporosarcina pasteurii CCM 2056 with specific activity
of 2451 U/mg. The values of the kinetic parameters
ꢁ
1
(K
M
¼ 4:92 ꢂ 0:31 mM and vmax ¼ 4:224 ꢂ 0:060 mM s ) of highly
purified urease in an enzymatic reaction were determined by fit-
ting the initial reaction velocities measured over a range of urea
concentrations to the Michaelis-Menten equation by nonlinear
regression. The majority of the assayed compounds exhibited inhi-
bitory activity in the micromolar or nanomolar range (Table 1). The
nonlinear progress curves indicated that Michael acceptors pro-
duced time-dependent inhibition of urease activity, in which
dence between pK and logkGSH, shown graphically in Fig. 1, allows
i
for classification of the studied compounds into three groups. Com-
pounds from the first group show roughly linear dependence
between activity and reactivity (1a, 8a, 9a, 9c, 10a, 10b), which
suggests their nonspecific interaction with the enzyme. The second
group contains structures with high chemical reactivity and mod-
erate inhibitory activity (2a, 2b, 3a). In this case, the interactions of
the compound with the enzyme are specific but unfavorable due to
steric impairment. The third group is of the highest importance
because it combines compounds with a preferable profile: high
inhibitory potential and low –SH reactivity (1b, 3h, 8c). In these
cases, inhibition of the enzyme is driven not only by –SH reactivity
but also specific interactions between the ligand and protein. The
specificity of compounds 1b, 3h, 8c was additionally confirmed
by complete lack of inhibition of model cysteine-dependent
enzyme – papain. On the basis of the discussed graph, compounds
of interest are easily indicated; in this particular case, acetylenedi-
steady-state velocity (
) and steady-state velocity (
inhibitor concentration over the examined ranges. The linear
replots of 1= and 1= showed a slow binding mode of action
v
s
) was attained slowly with both the initial
(v
i
v
s
) decreasing with an increase in
v
i
v
s
according to mechanism B, in which the initial EI complex under-
ꢀ
goes conformational change into the final EI complex. The steady
À
Á
ꢀ
state inhibition constants
K
ranged from 0.00977 to 64.2 lM
for compounds 8a and 3e, respectively. Importantly, the potency
of some of the assayed compounds (5, 6a, 8a, 8b, 8c, 9a, 9b, 9c,
iLB
9
f, 10a and 10b) exhibited higher activities than the reference inhi-
ꢀ
bitor, acetohydroxamic acid (AHA). All inhibitors demonstrated
competitive and reversible type inhibition, as confirmed by Line-
weaver-Burk plots and fast dilution assay (data not shown). This
mode of action of the studied inhibitors is consistent with the
carboxylic acid (8c) with KiLB ¼ 42:5 nM and logkGSH ¼ ꢁ2:14 has
the most beneficial activity-reactivity profile.
The cytotoxicity towards normal mouse fibroblasts (BALB/3T3)
for
a group of selected Michael acceptor compounds was