The Meisenheimer Complex as a Paradigm in Drug Discovery: Reversible Covalent Inhibition through C67 of the ATP Binding Site of PLK1

Article abstract of DOI:10.1016/j.chembiol.2018.06.001

The polo kinase family are important oncology targets that act in regulating entry into and progression through mitosis. Structure-guided discovery of a new class of inhibitors of Polo-like kinase 1 (PLK1) catalytic activity that interact with Cys67 of th

Full text of DOI:10.1016/j.chembiol.2018.06.001

Article
The Meisenheimer Complex as a Paradigm in Drug
Discovery: Reversible Covalent Inhibition through
C67 of the ATP Binding Site of PLK1
Graphical Abstract
Authors
Russell J. Pearson, David G. Blake,
Mokdad Mezna, Peter M. Fischer,
Nicholas J. Westwood,
Campbell McInnes
Correspondence
mcinnes@cop.sc.edu
In Brief
Pearson et al. describe the structure-
guided discovery of benzothiazole
N-oxide PLK1 inhibitors that bind
67
covalently with C through
Meisenheimer complexes. These
represent prototypes to exploit this
unique mechanism of inhibition in drug
discovery to achieve high specificity and
still bind reversibly.
Highlights
d
d
d
d
Identification of potent and selective PLK1 inhibitors via a
structure-guided approach
Covalent binding with Cys67 of PLK1 occurs through a
Meisenheimer complex (MC)
Lack of activity against PLK1 C67S confirmed the role of this
residue in inhibition
The MC allows covalent but reversible inhibition as alternative
for irreversible drugs
Pearson et al., 2018, Cell Chemical Biology 25, 1–10
September 20, 2018 ª 2018 Elsevier Ltd.
https://doi.org/10.1016/j.chembiol.2018.06.001

Please cite this article in press as: Pearson et al., The Meisenheimer Complex as a Paradigm in Drug Discovery: Reversible Covalent Inhibition through
C67 of the ATP Binding Site of PLK1, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.06.001
Cell Chemical Biology
Article
The Meisenheimer Complex as a Paradigm
in Drug Discovery: Reversible Covalent Inhibition
through C67 of the ATP Binding Site of PLK1
Russell J. Pearson,1,4 David G. Blake, Mokdad Mezna, Peter M. Fischer, Nicholas J. Westwood,
2
2
3
4
2,5,6,
and Campbell McInnes
1
*
School of Pharmacy, Keele University, Staffordshire ST5 5BG, UK
2Cyclacel Ltd., James Lindsay Place, Dundee DD1 5JJ, UK
3School of Pharmacy and Centre for Biomolecular Sciences, University of Nottingham, NG7 2RD, UK
4Department of Chemistry, University of St Andrews, Fife KY16 9ST, UK
5Drug Discovery and Biomedical Sciences, University of South Carolina, Columbia, SC 29208, USA
6Lead Contact
*Correspondence: mcinnes@cop.sc.edu
https://doi.org/10.1016/j.chembiol.2018.06.001
SUMMARY
ated by activation of the APC (Golan et al., 2002; Kotani et al.,
1998), and also through phosphorylation of the SCC1 subunit
The polo kinase family are important oncology tar- of cohesin (Alexandru et al., 2001; Sumara et al., 2004). Mitotic
gets that act in regulating entry into and progression exit occurs after PLK1 interacts with and phosphorylates the ki-
nesin-like protein, CHO1/MKLP-1 and promotes cytokinesis
through mitosis. Structure-guided discovery of a new
(
(
Lee et al., 1995). Four closely related PLKs exist in humans
Glover et al., 1998) and structurally contain two separate, highly
class of inhibitors of Polo-like kinase 1 (PLK1) cata-
lytic activity that interact with Cys67 of the ATP bind-
ing site is described. Compounds containing the
benzothiazole N-oxide scaffold not only bind cova-
lently to this residue, but are reversible inhibitors
through the formation of Meisenheimer complexes.
This mechanism of kinase inhibition results in com-
homologous domains: an amino-terminal region that contains
the catalytic kinase domain and a C-terminal domain that has
two conserved Polo-box regions (Elia et al., 2003a, 2003b)
(one in PLK4 [Leung et al., 2002]). A fifth PLK has been identified
but shown to lack a kinase domain (de Carcer et al., 2011). The
polo boxes function in recruitment for subcellular localization
pounds that can target PLK1 with high selectivity, and compartmentalization (Lee et al., 1998). Furthermore there
while avoiding issues with irreversible covalent bind- is evidence suggesting that the Polo-box domain (PBD) partici-
ing and interaction with other thiol-containing mole- pates in regulating interactions with substrates, e.g., CDC25C
phosphatase (Elia et al., 2003a, 2003b), as an autoregulatory
cules in the cell. Due to renewed interest in covalent
drugs and the plethora of potential drug targets,
these represent prototypes for the design of kinase
inhibitory compounds that achieve high specificity
through covalent interaction and yet still bind revers-
ibly to the ATP cleft, a strategy that could be applied
to avoid issues with conventional covalent binders.
domain (Jang et al., 2002). Furthermore, PBD-dependent PLK1
activity is required for proper metaphase/anaphase transition
and cytokinesis (Seong et al., 2002; Yuan et al., 2011).
Overexpression of PLK1 is a common occurrence in cancer
and is of prognostic value for several tumor types (Knecht
et al., 1999; Takahashi et al., 2003; Tokumitsu et al., 1999; Wolf
et al., 1997; Yuan et al., 1997). In addition, constitutive expres-
sion of PLK1 in mammalian cells can lead to malignant transfor-
mation (Smith et al., 1997). PLK1 has been extensively validated
as an anti-tumor drug target using a variety of cellular and in vivo
INTRODUCTION
studies. Downregulation of PLK1 activity, through the adminis-
The Polo-like kinases (PLKs) are serine/threonine protein kinases tration of antisense oligonucleotides and small interfering RNA
that are key regulators of the cell cycle, and in mammals have molecules, has been demonstrated to be highly effective in inhi-
been reported to play numerous roles in regulating entry into bition of proliferation of cancer cells both in vitro and in vivo (Elez
and progression through mitosis (Craig et al., 2014; McInnes et al., 2000, 2003; Spankuch-Schmitt et al., 2002a, 2002b).
and Wyatt, 2011). PLK1 is expressed throughout the cell cycle, Several drug candidates have been investigated in human trials
although its activity peaks late in G2 and is sustained through including BI2536 and GSK461364A, with the most advanced
mitosis (Hamanaka et al., 1995; Lee et al., 1995). At the end of compound being volasertib (BI-6727), shown to have clinical ef-
G2, through phosphorylation and promotion of nuclear translo- ficacy in combination with cytarabine in acute myeloid leukemia
cation (Toyoshima-Morimoto et al., 2002), PLK1 acts on (Dohner et al., 2014; Gjertsen and Schoffski, 2014). Although
CDC25C (Roshak et al., 2000), resulting in activation of CDK1/ promising activity has been observed, there is continued scope
cyclin B by removal of inhibitory phosphate groups (Nurse, to improve clinical outcomes through optimization of mechanism
1
990). PLK1 also promotes sister chromatid separation medi- of inhibition and pharmacokinetic and dynamic properties.
Cell Chemical Biology 25, 1–10, September 20, 2018 ª 2018 Elsevier Ltd.
1

Please cite this article in press as: Pearson et al., The Meisenheimer Complex as a Paradigm in Drug Discovery: Reversible Covalent Inhibition through
C67 of the ATP Binding Site of PLK1, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.06.001
Figure 1. Chemical Structures of PLK1
Inhibitory Compounds (2TA, 1; 5TA, 2; and
adenosine 3) Used in Validation of the Ho-
mology Model and for Probing the Covalent
Modification of the ATP Binding Site
833
Here a structure-guided design approach to the discovery and adjacent to the C
and was shown in this to form a disulfide
development of benzothiazole N-oxide PLK1 inhibitors is bond and to inhibit the kinase activity. As the active site cysteine
67
described. These highly selective molecules inhibit the enzyme (C ) is on the opposite face of the ATP cleft in the PLK1 context,
0
through a covalent but reversible mechanism involving a Meisen- inhibition would necessitate the synthesis of 5 -thioadenosine
heimer complex (MC), which is without precedent in the protein (5TA), 2, (Figure 1) so as to position the ribose thiol proximal to
kinase literature. In recent years, covalent inhibition of drug tar- the reactive side chain. Flexible docking of 5TA 2, with PLK1
gets, and especially of protein kinases, has been recognized was completed and the binding mode was verified through com-
as an effective approach to improving selectivity and for over- parison with ATP bound crystal structures. Further to this, the di-
67
coming drug resistance; however, this is not without significant sulfide link between C and 5TA 2, was formed in silico and re-
issues including potential toxicities (Baillie, 2016). Inhibition sulted in a conformation (Figure 2) similar to the experimental
through MC formation is a promising strategy to combine the binding mode observed in other kinases. Subsequent to this, 2
benefits of a covalent mechanism while retaining reversible bind- was synthesized and shown to inhibit PLK1 kinase activity with
ing to the drug target of interest. A recent report described the a half maximal inhibitory concentration (IC50) of 39 mM. 2TA 1, ex-
observation of a MC in the context of nitrobenzothiazinone hibited markedly lower inhibition (120 mM) and adenosine 3 (Fig-
anti-tuberculosis drugs; however, it did not elaborate these as ure 1), itself had no observed activity against PLK1 (>200 mM).
contributing to potency and selectivity but merely from the
standpoint of the metabolic implications for continuing develop- Discovery of Potent and Selective Inhibitors of
ment of this important class of drugs (Kloss et al., 2017). The PLK1 Using Lidaeus
compounds described herein therefore represent prototypes The homology model of the kinase domain of PLK1 was vali-
and proof of concept for exploiting the MC in kinase inhibition dated through the observed correlation of intermolecular con-
and potentially more broadly in drug discovery.
tacts and binding energies of known inhibitors and through the
assay results obtained from the use of thiol-modifying agents.
Since it was considered to be of sufficient computational reli-
ability, the model structure was further used in high-throughput
docking calculations to identify novel PLK1 starting points. To
carry out this task, one of the previously docked thiazole ani-
RESULTS
PLK1 Is Irreversibly and Selectively Inhibited by Thiol
Modifying Agents
A structure-guided approach for the discovery of the cyclapolin lino-pyrimidine compounds in complex with PLK1 was used as
benzothiazole N-oxide class of inhibitors has been described a template structure for Lidaeus calculations. Lidaeus (Taylor
(
McInnes et al., 2006). This approach preceded the determina- et al., 2008; Wu et al., 2003), a tool for rapid flexible docking of
tion of the crystal structure of PLK1 kinase domain. Upon con- ligands into protein binding sites, generates site points from en-
struction of a homology structure and examination of the ATP ergies of probe atoms within a specified radius of the inhibitor.
binding site residues, it was observed that PLK1 had some un- Interatomic ligand vectors are then matched with the site points
usual features compared with other protein kinases. In particular, from fragments of the molecule followed by positioning of subse-
the residue at the end of the glycine-rich loop that commonly in- quent fragments and ranking by overall intermolecular energy
teracts with the ribose of ATP, and is frequently a valine, occurs score. As previously described, using the complex of a pyrimi-
6
7
as a cysteine in PLK1 (C ). This residue in other protein kinase- dine derivative with PLK1, Lidaeus was employed to dock
ligand crystal structures generally makes significant contacts approximately 200,000 commercially available small molecules.
1
with ATP competitive small molecules (e.g., V in CDK2). The From these, 350 in-silico-ranked compounds were assayed for
8
6
7
unique characteristic of C in PLK1 thus could allow exploitation PLK1 inhibition in a radiometric kinase assay quantifying the
of the nucleophilicity of the thiol side chain to enable covalent amount of radiolabeled ATP incorporated into the CDC25C sub-
binding of PLK1 inhibitors.
strate. A number of in vitro PLK1 inhibitors emerged from this
The development of irreversible kinase inhibitors has prece- strategy with an overall hit rate of 1% and a potency range of
dent in the literature in the context of the epidermal growth factor 0.5–20 mM IC50. While several confirmed hits were obtained,
receptor tyrosine kinases, where non-covalently binding ATP one series had more potent activity in the kinase assay and con-
competitive compounds were converted by addition of a sisted of a benzothiazole N-oxide core structure. The obtained
geometrically appropriate reactive functionality (Singh et al., hit, 4 (Table 1), was promising not only from its micromolar po-
1997). These compounds were designed to form a covalent tency but also from its low molecular weight and potential for
bond with a cysteine in the kinase active site and thus to take synthesizing a number of derivatives that could optimize potency
advantage of the thiol attack on the inhibitor. In the erbB1 kinase and drug-like properties. The phenotypic activity of compound 4
0
domain, 2 -thioadenosine (2TA), 1 (Figure 1), places the thiol has previously been described where novel roles of PLK1 in
2
Cell Chemical Biology 25, 1–10, September 20, 2018

Please cite this article in press as: Pearson et al., The Meisenheimer Complex as a Paradigm in Drug Discovery: Reversible Covalent Inhibition through
C67 of the ATP Binding Site of PLK1, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.06.001
Table 1. Structure-Activity Relationships for the Benzothiazole
N-Oxide PLK1 Inhibitors
Structure
Kinase Inhibition (mM)
CDK2/
Compound
R
1
R
2
R
3
R
4
PLK1
Cyclin E
4
5
6
7
8
CF
F
3
NO
NO
NO
NO
NO
2
2
2
2
2
CONH
CONH
CONH
CONH
CONH
2
2
2
2
2
O
O
O
O
O
2.47 ± 1.23 >100
18.1 ± 2.69 >100
0.36 ± 0.01 >100
7.11 ± 0.86 >100
8.4 ± 0.86 44.2 ±
I
COOCH
3
CN
3.18
9
CH
SCF
COOH
CF
CF3
CF
NO
CF
3
NO
NO
NO
NO
2
2
2
2
CONH
CONH
CONH
CN
2
2
2
O
O
O
O
O
–
>100
>100
1
1
1
0
1
2
3
0.39 ± 0.07 >100
>100
Figure 2. Complex of 5TA, 2, with the ATP Binding Site of PLK1
Complex was generated through flexible docking of the free thiol compound
into the PLK1 homology model with selected conformations that correspond
to the known binding mode of ATP derivatives. The irreversible inhibition of this
derivative is supported by the formation of the disulfide linkage in silico and
minimization of the covalently linked inhibitor.
3
17.0 ± 4.24
31.0 ± 15.6 >100
>100
13
NO2 CONHOH
14
15
16
3
NO
2
2
CONH
2
2
2
CF
3
CONH
O
O
>100
3
NO
CONHNH
2
>100
spindle pole maintenance were demonstrated through use of
this molecule as a chemical biology probe (McInnes et al.,
2006). The structure-activity relationship of 4 and its related an-
alogs are described for the first time in the following results.
increased in potency from 2.5 to approximately 0.4 mM, suggest-
ing that the spacing functionality resulted in more complemen-
Structure-Activity Relationship Determination of
Benzothiazole Inhibitors
3
tary interactions of the CF group in the ATP binding site. Gener-
Subsequent to the identification of 4 as a potent PLK1 inhibitor, a ation of the free carboxylate group from hydrolysis of the methyl
set of commercial compounds expanding on the benzothiazole ester 7, to generate compound 11, interestingly resulted in an
N-oxide scaffold were obtained and tested for inhibition of inactive molecule.
PLK1. This included a significant number of secondary amide
Further synthetic modification of the substituents on the thia-
and ester derivatives at the R position of the thiazole ring. Sur- zole ring were carried out and revealed that the incorporation
3
prisingly none of these compounds exhibited any significant of other small polar groups replacing the amide resulted in active
level of in vitro inhibition of PLK1, thus suggesting an important compounds, although none as potent as the parent hit (Table 1).
contribution of the primary amide (see structure in Table 1, inac- Compounds tested included those with nitrile 12, hydroxamate
tive esters not shown) to binding.
To further explore the structure-activity relationship (SAR) of
13, and hydrazide 16 functions.
An additional observation made was that closely related com-
this series, a synthetic route for the benzothiazole N-oxide was pounds lacking the N-oxide on the thiazole ring were completely
developed, and a number of analogs related to the in silico hit devoid of activity as exemplified by the reduction of the N-oxide
compound were prepared. Firstly, the substituents on the aryl in 4 to generate 14 and that of the reduction of 10 to its counter-
component of the pharmacophore were modified, and testing part (data not shown). These results were not in line with the
of these compounds in the PLK1 assay yielded information on predicted small contribution of a single atom to intermolecular
the SAR. A number of derivatives were synthesized (Table 1) binding but pointed to the importance of the electrostatic nature
replacing the CF
iodo 6, methyl ester 7, nitrile 8, and methyl 9. Surprisingly, only Further SAR for the benzothiazole N-oxide series was estab-
compounds possessing electron-withdrawing substituents in lished through the synthesis of compound 15 with the CF and
this position exhibited activity against PLK1. In particular, the nitro groups reversed. This compound was surprisingly inactive
methyl group derivative 9, which is isosteric with CF , was pointing to essential interactions of the nitro group with the ATP
completely inactive. When the CF group in the in-silico-discov- binding site of PLK1.
ered compound was replaced with a trifluoromethyl-thioether To gain some insight into the binding of the benzothiazole
0, an increase in potency was observed. This compound N-oxide lead compounds in the PLK1 ATP cleft, and to assess
3
with other groups and included fluoro 5, and electron-withdrawing potential of the N-oxide functionality.
3
3
3
1
Cell Chemical Biology 25, 1–10, September 20, 2018
3

Please cite this article in press as: Pearson et al., The Meisenheimer Complex as a Paradigm in Drug Discovery: Reversible Covalent Inhibition through
C67 of the ATP Binding Site of PLK1, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.06.001
covalently with the cysteine. Further investigation into the chem-
istry of the benzothiazole N-oxide series suggested a literature
precedent for reactivity and susceptibility to nucleophilic attack
on the aromatic ring (Bunting, 1979). This reaction has been
shown to occur ortho or para to a nitro-substituted phenyl ring,
where a covalent intermediate is formed as per the mechanism
shown above Table 2, and has been named the ‘‘Meisenheimer
complex.’’
Examination of the docked structure of 4 with the PLK1 model
67
structure indicates a close proximity of the thiol of C and the C4
aromatic carbon (para to the nitro group and highlighted in Fig-
ure 3), and hence strongly supports this hypothesis. Confirma-
1
tion of the MC was indicated through a set of model H nuclear
magnetic resonance (NMR) experiments with the active and
inactive compounds from the benzothiazole series. Addition of
n-butylamine as a surrogate nucleophile to the active com-
pounds including 4, 7, 8, and 10, in each case, resulted in a dra-
matic color change of the compound in solution. For solubility
purposes, all of these experiments relied on the use of the ethyl
Figure 3. Molecular Docking of the In-Silico-Discovered Benzothia-
zole N-Oxide, 4, with the ATP Binding Site of PLK1
The binding mode was predicted based on most favorable interaction energies
and the SAR of the primary amide group is thus consistent with hinge
H-bonding interactions. The displayed pose is consistent with the Mei-
senheimer hypothesis requiring the proximity of the benzothiazole C4 to the
3 2
ester derivatives of each benzothiazole (R = CO Et) dissolved in
deuterated chloroform, and therefore the majority of model reac-
tions were carried out using n-butylamine. Experiments were
also undertaken with sodium 1-butanethiolate and showed
67
attacking thiol nucleophile from C
.
the contributions of various substituents for binding, molecular similar dramatic changes in the color of the compound solution
docking experiments were carried out using the PLK1 model and in the UV spectrum (Figures S1 and S2). The requirement
structure and subsequently with the available crystal structures. for polar aprotic NMR solvents (potentially reactive) to solubilize
Using this approach, numerous docking solutions were evalu- the 1-butanethiolate precluded NMR experiments with this more
ated in terms of interaction energy, docking score, and, with appropriate model nucleophile. On the other hand, non-polar
respect to consistency, with known kinase inhibitory interac- solvents did not allow NMR study of the MC due to a lack of com-
tions. Particular attention was paid to the observed requirement pound solubility.
for a primary amide group on the thiazole. A binding mode that
Further experiments utilizing UV-vis spectrophotometry were
fulfilled each of these criteria was observed (Figure 3). Interest- carried out to support this hypothesis using ethanol as the sol-
6
7
ingly, C is in close proximity to the C4 (aromatic ring) of the vent of choice. During initial experiments to identify the presence
benzothiazole system, although it does not make the comple- of the MC, it was observed that highly colored solutions were
mentarity expected from the valine residue that is commonly obtained upon addition of the model nucleophile (n-butylamine)
found in this position. In addition, the interactions observed in to the benzothiazole N-oxide inhibitor, an observation consistent
the docked structure provided a clear rationale for CF
3
S group. with the literature, where a red shift in the UV-vis spectrum is
These calculations indicate that the CF
3
S projects into a sub- seen for lmax (Taylor, 1970). An extensive UV-vis spectrophoto-
pocket and shows a high degree of complementarity with this re- metric characterization of the series of active and inactive benzo-
gion. A good correlation was observed between the non-bonded thiazoles was subsequently carried out to examine if a relation-
docking score and the potency of the inhibitors.
ship existed between the induced color change and potency of
the pharmacophoric series (Table 2). It was determined that,
for each of the active PLK1 inhibitors from this series, a substan-
tial shift in UV absorbance was identified, accompanied by a dra-
Benzothiazole N-Oxide Inhibitors Interact Covalently
with the PLK1 Active Site
To confirm that the benzothiazoles were binding to the ATP cleft matic increase in extinction coefficient for the new signal. Out of
of PLK1 and were not targeting substrate binding or possibly eight compounds tested, all six analogs with measurable IC50
another site resulting in allosteric inhibition of ATP binding, the values against PLK1 exhibited a significant UV shift and an
K
m,ATP was determined at various inhibitor concentrations. This enhancement of extent of absorbance. Conversely, neither of
measurement was repeated for two of the most potent PLK1 the inactive molecules in the series displayed either a UV shift
inhibitory compounds, 4 and 10. While compound 4 (2.5 mM or an increase in extinction coefficient (9a and 11a). In addition,
IC50) was shown to be ATP competitive, as demonstrated by a rough correlation of the increase in absorbance and potency of
the variation in the K
analog, 10 was found to have similar K
m
for ATP, the significantly more potent the compounds can be extracted from the measured data.
for ATP at every ligand Furthermore two additional inactive compounds in this series
m
concentration studied (Figure 4). These data indicate that the were tested and no UV shift or absorbance increase was
equilibrium of binding for compound 10 shifts toward being observed (data not shown).
1
Further characterization of the H NMR spectra of the inhibi-
non-ATP competitive in contrast to inhibitor 4. In light of these
6
7
results, observation of the C side chain in the PLK1 model tors before and after addition of the nucleophile (n-butylamine)
structure and the outcomes of using thiol modifying agents, it resulted in a disappearance of both the aromatic protons in the
was hypothesized that these compounds may be interacting spectrum (Table 2) and a shift of the signals further upfield, which
4
Cell Chemical Biology 25, 1–10, September 20, 2018

Please cite this article in press as: Pearson et al., The Meisenheimer Complex as a Paradigm in Drug Discovery: Reversible Covalent Inhibition through
C67 of the ATP Binding Site of PLK1, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.06.001
Figure 4. Enzyme Kinetics of Benzothiazole N-Oxide PLK1 Inhibitors
(
A and B) Lineweaver Burk Plot analysis of compounds 4 and 10 suggest that they act via competitive (A) and non-competitive (B) binding, respectively.
C) The ATP dependence of these two inhibitors further confirms this analysis (compound 4 with red data points and 10 in magenta).
(
is characteristic of the cyclohexa-2,5-dien-1-ylidene azinate of cysteine 67 for serine, was generated. As serine is consider-
structure formed after nucleophilic attack (see above Table 2). ably less nucleophilic than the native cysteine residue, it would
Subsequent examination of the inactive compounds in a similar not be expected to form the MC to a significant degree and
experimental protocol determined that for these benzothiazoles, therefore the activity of the benzothiazole N-oxide series should
1
no change in the aromatic H NMR signals occurred. This was be dramatically lower in the context of this mutant. Testing of vol-
shown to be true for the previous incongruously inactive com- asertib in the first instance showed that while its activity against
3 3
pounds, including the CF to CH replacement, 9, and the PLK1 C67S was reduced compared with the wild-type, an IC50 of
carboxylate derivative, 11. Analysis of the compounds (those 0.151 mM was obtained (Figure 5). Staurosporine was also
active against PLK1) recovered after removal of the n-butylamine shown to potently inhibit PLK1 C67S (Figure S3). This contrasted
showed that there was no evidence of the Meisenheimer adduct to compound 4, which was found to be completely inactive at
therefore confirming that this is a fully reversible reaction.
concentrations of >100 mM, thereby providing confirmation
that Cys67 forms a covalent MC with the benzothiazole N-oxide
series as strongly suggested by the previously described data.
Activity of Compound 4 and volasertib (BI6727) against
PLK1 WT and PLK1 C67S
6
7
To confirm whether C plays a role in the activity of the benzo- Kinase Selectivity and Cellular Activity of Lead
thiazole N-oxide series, as strongly suggested by the SAR data, Inhibitors
along with the enzyme kinetics and spectroscopy results, com- Since specificity can be problematic in the development of
pound 4 was investigated along with volasertib (BI6727), a ATP competitive inhibitors of any protein kinase, the benzothia-
potent PLK1 ATP inhibitor currently approved by the US Food zole N-oxide series was tested on a panel of in-house enzymes
and Drug Administration for the treatment of acute myeloid leu- (Table S1). Interestingly none of these were inhibited by the
kemia. Due to requirements for commercial assay screening, in-silico-identified lead compound 4. To further probe the selec-
these compounds were tested against the wild-type PLK1 with tivity of this compound, a further 20 kinases available from an
a lower ATP concentration than previously used and shown to out-sourced screen were examined for inhibition in the presence
have IC50 values of 0.463 and 0.016 mM, respectively. Subse- of 100 mM of 4, and included three kinases with a cysteine in the
quent to this, a mutant PLK1 enzyme possessing the exchange position equivalent to PLK1 (Table S2). None of this additional
Cell Chemical Biology 25, 1–10, September 20, 2018
5

Please cite this article in press as: Pearson et al., The Meisenheimer Complex as a Paradigm in Drug Discovery: Reversible Covalent Inhibition through
C67 of the ATP Binding Site of PLK1, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.06.001
2
Table 2. Spectroscopic Results of Benzothiazole N-Oxide Meisenheimer Formation (in the Presence of n-BuNH )
l
max (nm),
b
2
b
2
ꢀ1
ꢀ1 c
PLK1 IC50 (mM)
When R = CONH
d
Ar1 (ppm)
+
d
Ar2 (ppm)
+
l
max (nm),
3
(L mol cm ) +
a
b
b
ꢀ1
ꢀ1 c
Compound
R1
3
2
d
Ar1 (ppm)
d
Ar2 (ppm)
n-BuNH
n-BuNH
3
(L mol cm
)
n-BuNH
2
4
5
6
7
8
9
a
a
a
a
a
a
CF
3
2.47 ± 1.23
18.1 ± 2.69
0.36 ± 0.01
7.11 ± 0.86
8.4 ± 0.86
>100
9.00–8.99
–
8.85–8.84
–
6.78–6.77
6.61
405, 5,369
380, 3,143
410, 3,937
400, 2,443
410, 4,399
400, 5,085
405, 4,412
395, 4,063
440, 24,832
450, 7,429
465, 12,992
420, 19,870
450, 14,076
395, 5,367
445, 18,382
390, 4,375
F
I
–
–
–
–
8.94
8.96
–
8.81
8.81
–
COOCH
CN
3
7.10
–
7.00
–
CH
SCF
COOH
3
8.65
8.93
8.96
8.42
8.78
8.80
8.70–8.40
6.78
9.00
8.70–8.40
6.64
1
0a
1a
3
0.39 ± 0.07
>100
1
8.82
Position C4 of the benzothiazole N-oxide is less hindered, favored by the SNAr mechanism and in close proximity to C67 in the docked models.
a
Primary amide of R3 is replaced by COOEt.
1
b
3
All H NMR performed in CDCl .
c
All UV-vis in EtOH.
panel exhibited a significant level of inhibition, with the exception cysteine residue. In addition, the observation that adenosine
of CSK, whose activity was slightly blocked by the action of this does not inhibit PLK1 gives weight to the hypothesis that selec-
compound (<50% inhibition at 100 mM compound). Due to the tive covalently interacting inhibitors can be designed and synthe-
level of potency obtained for this fragment like series, and the sized based on the structural information obtained from this
high degree of selectivity observed, it is apparent that the spec- approach. These results, taken as a whole, strongly suggest
ificity arises from the demonstrated mechanism of inhibition that the cysteine residue present in the PLK1 ATP cleft can
through the MC. The selectivity of the most potent inhibitor, form covalent interactions with appropriately positioned electro-
1
panel and confirmed its selectivity toward PLK1.
0, was also examined by testing against an in-house kinase philic groups in an inhibitor. This novel feature of PLK1 should
thus enable the discovery and design of inhibitors that potently
and selectively inhibit the activity of PLKs. To this end, as the
homology structure was extensively validated, it was confidently
utilized for database searching for new PLK1 inhibitors and
DISCUSSION
The results presented here demonstrate that an approach to resulted in the identification of the benzothiazole N-oxide phar-
generating PLK1 inhibitors involving homology modeling, molec- macophore. Results show that potency and selectivity for
ular docking of known ligands, and in vitro testing has deter- PLK1 over many other kinases can be achieved through this
mined novel features of the PLK1 ATP binding site that can be chemical series, which not only binds avidly to the active site
67
exploited in the structure-guided design of compounds specific but also interacts covalently with C . While the homology struc-
for this kinase. The ability to selectively inhibit PLK1 via covalent ture was used in initial studies to propose the covalent mecha-
6
7
interaction with C of the ATP binding site was confirmed nism of inhibition and for discovery of the benzothiazole N-oxide
through the synthesis of thio-adenosine derivatives (1 and 2). inhibitors, the publication of multiple crystal structures of the
67
The cysteine is located on the glycine-rich loop in the N-lobe PLK1 kinase domain confirmed the position of C and further
region in the position occupied by a valine in the majority of validated the results obtained with the homology structure.
protein kinases and is quite unique to the PLKs. The preferential Furthermore, compounds were redocked with PLK1 crystal
inhibition of PLK1 by 5TA, 2, which, according to known kinase- structures and, as expected, also found to have binding modes
ATP structures and subsequent modeling (Figure 2) in the PLK1 compatible with formation of the MC.
context, results from proximity of the thiol group to the reactive It is apparent from these results that the necessity for an elec-
0
cysteine (as opposed to the 2 -thio derivative, where the thiol tron-withdrawing group on the phenyl substituent of the benzo-
group projects toward the C-lobe and therefore would be less thiazole, and the presence of the N-oxide for PLK1 inhibition,
6
7
able to react with C ). Although 2TA, 1, inhibits PLK1 to a certain relate to the ability of these compounds to form MCs. This was
degree, its observed activity probably results from the flexibility determined by demonstrating a correlation between inhibitors
of the ribose in the ATP binding site, enabling reaction with the of the benzothiazole series that show activity against PLK1 and
6
Cell Chemical Biology 25, 1–10, September 20, 2018

Please cite this article in press as: Pearson et al., The Meisenheimer Complex as a Paradigm in Drug Discovery: Reversible Covalent Inhibition through
C67 of the ATP Binding Site of PLK1, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.06.001
Figure 5. Comparison of the Activities of
Compound 4 (left) and volasertib (right)
against PLK1 C67S, a Mutant Designed to
67
Test the Contributions to C to the Activity
of the Benzothiazole N-Oxide Series (error
bars represent standard deviations calcu-
lated for duplicate experiments).
The data obtained for the replacement of the
cysteine with a serine, an isosteric residue, and
which is considerably less nucleophilic, confirms
the formation of the Meisenheimer complex with 4.
their ability to induce spectroscopic changes upon addition of a ported by the observation that this compound was refractory
model nucleophile. This was further supported by the require- to formation of the MC, as demonstrated by the described spec-
ment of the R
drawing substituent. Small changes in the chemical structure generate compound 15 resulted in complete loss of activity on
of this series (i.e., replacement with isosteric groups, e.g., CF PLK1, an effect probably due to the critical nature of the nitro
to CH or removal of the N-oxide functionality), which should function to direct binding, and supported by electrostatic inter-
1 3 2
position (see Table 1) to be an electron-with- troscopic methods. Reversal of the CF and NO groups of 4 to
3
3
82
not be so detrimental to kinase activity, resulted in complete actions with K in the docked structures (Figure 3). Since the
lack of PLK1 inhibition. These changes do not significantly alter nitro group is acting as an electron sink it will be more negatively
the steric nature of the inhibitor, but would dramatically change charged in the MC and therefore contributes significantly to the
the electronic nature of the structures. Increasing the electron- enthalpy of binding.
withdrawing character at R
1
would therefore make the C4 and
Confirmation for the MC formation and the resulting covalent
67
C6 positions of the nitro-substituted benzothiazole ring less inhibition of PLK1 through C was obtained through comparison
electron-rich and more susceptible to nucleophilic attack, result- of the activities of compound 4 and volasertib (BI6727) in inhibit-
ing in formation of the MC (see above Table 2). This was corrob- ing the wild-type PLK1 versus the C67S mutant. The retention of
orated by the observed correlations that exist between the potent activity of volasertib for the mutant PLK1 and the com-
electronegativity of the R
potency, and also between the UV absorbance of the Meisen- evidence that the benzothiazole N-oxide series act through MC
heimer adduct and the PLK1 activity. formation. The decreased activity of volasertib with PLK1
1 2
and R groups and the inhibitor plete loss of inhibition of 4 in this context provides strong
The higher potency of compound 10 and the differing kinetics C67S is most likely due to the loss of van der Waals contacts
versus 4 (Figure 4.) can be explained by the greater affinity and in the context of the serine as the oxygen has a significantly
complementarity of the SCF
site and the structural basis for this potency increase has been
determined computationally. The SCF group has almost twice zothiazole ring and the sulfur nucleophile leads to high-affinity
the lipophilicity of the CF group and therefore has greater poten- binding of this series and is likely responsible for the selectivity
tial to interact through van der Waals interactions in a binding for PLK1 versus all of the other enzymes tested. This method
pocket (Feng et al., 2016). Calculations show that the SCF of of kinase inhibition is currently without precedent in the literature,
0 has a significantly higher degree of complementarity with a and has been demonstrated to be a paradigm through which se-
3
group with the PLK1 ATP binding smaller atomic radius than the sulfur.
The formation of the Meisenheimer adduct between the ben-
3
3
3
1
3
sub-pocket of the ATP binding site compared with CF (4), occu- lective kinase inhibitors can be obtained. Despite the observa-
pying it almost completely and therefore explaining the potency tion that these act through covalent bond formation with the
increase in these terms alone (the significantly increased interac- reactive cysteine in the PLK1 active site, this series should not
3
tions of the SCF group are shown comparatively; Figure S4). As suffer from reactivity with general nucleophilic groups in biolog-
a result and consistent with enzyme inhibition kinetics, it is ical milieu since the MC is not a stable intermediate and is under
known that, if the affinity is tight enough, compounds can display normal conditions fully reversible. This was demonstrated in the
non-competitive behavior. The kinetic differences of 4 versus 10 context of the benzothiazole N-oxide series in that no stable in-
can be explained by the increased potency of 10, which results in termediate MC could be isolated after addition of model nucleo-
stabilization of the binding mode and therefore, by extension, the philes even though the presence of this species was confirmed
covalent MC (Table 1). Other explanations such as displacement using two spectroscopic methods.
of the SCF
pounds are very unlikely due to the similar electronegativities of confirmed the mechanism of inhibition of the benzothiazole
the CF and SCF groups, and also the restrictions on the nucle- series through the MC, the molecular docking calculations
ophilic aromatic substitution reaction (SNAr) in this system. that were carried out strongly support the proposed covalent
3
group or differences in the oxidation of the two com-
In addition to the experimental methods used and which
3
3
67
The lack of inhibition of the carboxylate derivative 11, despite interaction with C . The binding modes and interaction ener-
possessing an electron-withdrawing substituent most probably getics generated using a rigorous molecular mechanics docking
results from delocalization of the negative charge into the aro- approach (allowing flexibility of both the ligand and ATP binding
matic ring thus disfavoring nucleophilic attack. This was sup- site residues) were predicted for a set of the active analogs. As
Cell Chemical Biology 25, 1–10, September 20, 2018
7

Please cite this article in press as: Pearson et al., The Meisenheimer Complex as a Paradigm in Drug Discovery: Reversible Covalent Inhibition through
C67 of the ATP Binding Site of PLK1, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.06.001
the SAR data pointed to the critical nature of the primary amide anism of kinase inhibition, which yields high selectivity while
group, and since a frequent binding determinant of kinase inhib- avoiding off target effects that occur through covalent bind-
itors includes an H-bond donor and acceptor pair to the inter- ing to other thiol-containing molecules in the cell. Due to
domain connecting loop (‘‘hinge’’), it is highly probable that recent renewed interest in covalently binding inhibitors as
the amide is responsible for these interactions (no other a class of drugs, and the plethora of targets that can be
H-bond donors in the parent benzothiazole molecule, 4). Dock- inhibited using this approach, these compounds represent
ing of other active isosteres at this position including the nitrile prototypes for the design of kinase inhibitors and, in fact,
1
2, hydroxamate 13, and the hydrazide 16 derivatives confirm inhibitors in general that exploit the Meisenheimer complex
the contributions of the hinge contacts, and therefore the overall to generate potent and selective compounds that potentially
binding mode. Overall, an excellent correlation between the also avoid toxicity issues arising for irreversibly bind-
interaction energy of this series of benzothiazole N-oxide deriv- ing drugs.
atives and the IC50 values was observed and therefore confirms
6
7
the binding mode that positions C for nucleophilic attack to
form the MC.
STAR+METHODS
A study has demonstrated that potent and selective inhibitors
of protein kinases could be obtained using a structural bioinfor-
matics approach that targeted the corresponding cysteine
described in this work as resulting in specific PLK1 inhibition
Detailed methods are provided in the online version of this paper
and include the following:
(
Cohen et al., 2005). Sequence alignment analysis revealed
d KEY RESOURCES TABLE
that eight other protein kinases have the equivalent cysteine res-
idue representing four different kinase families. In this example,
however, in addition to cysteine, a threonine ‘‘gatekeeper’’ was
required in order to provide additional space for a substituent
on the designed inhibitor. While three of these kinases (MSK2,
NEK2, and RSK1) were probed for inhibition by compounds 4
and 10, and no significant activity observed, each could poten-
tially be inhibited by appropriate tuning of the inhibitor containing
a ‘‘Meisenheimer Complex Scaffold.’’ Furthermore, recent work
described the analysis of cysteine residues across the human
kinome and which would be available for covalent inhibitor
development (Zhao et al., 2017). This study concluded that there
are 44 kinases containing cysteine residues that have been
successfully exploited by covalent kinase inhibitors. There is,
therefore, a plethora of potential kinase drug targets that the
MC approach could be applied to for potent and selective inhibi-
tion. While resistance through mutation of the cysteine residue is
always possible, it remains so regardless of the mode of covalent
inhibition.
d CONTACT FOR REAGENT AND RESOURCE SHARING
d METHOD DETAILS
B General Method for the Synthesis of 2-alkoxycarbonyl-
N-oxide Derivatives
B Synthesis of 2-Carbamoyl-7-nitro-5-(trifluoromethyl)
benzothiazole 3-oxide, Compound 4
B Synthesis of 4-Fluoro-2,6-dinitrophenyl 4-methylben-
zenesulfonate, Precursor for Compound 5
B Synthesis of 2-Carbamoyl-5-fluoro-7-nitrobenzothia-
zole 3-oxide, Compound 5
B Synthesis of 2-Carbamoyl-5-iodo-7-nitrobenzothia-
zole 3-oxide, Compound 6
B Synthesis of 2-carbamoyl-5-(methoxycarbonyl)-7-ni-
trobenzothiazole 3-oxide, Compound 7
B Synthesis of 2-Carbamoyl-5-cyano-7-nitrobenzothia-
zole 3-oxide, Compound 8
B Synthesis of 2-Carbamoyl-5-methyl-7-nitrobenzothia-
zole 3-oxide , Compound 9
B Synthesis of 2-Carbamoyl-7-nitro-5-((trifluoromethyl)
thio)benzothiazole 3-oxide, Compound 10
B Synthesis of 2-Carbamoyl-5-carboxy-7-nitrobenzo-
thiazole 3-oxide, Compound 11
In summary, through study of the PLK1 kinase domain,
unique features of its ATP binding site have been exploited in
the discovery and design of inhibitors. A class of PLK1 inhibi-
tors based on the benzothiazole N-oxide have been described
and shown to selectively inhibit the enzyme through an MC, a
mechanism of action not previously described for protein ki-
nase inhibitors. This pharmacophore represents a prototype
for a new approach to generate PLK1 inhibitors and, in fact, in-
hibitors in general that possess high selectivity through revers-
ible covalent interaction. Such inhibitors have the potential
added benefit of minimizing off target effects that occur with
irreversible inhibitors.
B Synthesis of 2-Cyano-7-nitro-5-(trifluoromethyl)ben-
zothiazole 3-oxide , Compound 12
B Synthesis of 2-(Hydroxycarbamoyl)-7-nitro-5-(trifluor-
omethyl)benzothiazole 3-oxide , Compound 13
B Synthesis of 2-Carbamoyl-7-nitro-5-(trifluoromethyl)
benzothiazole , Compound 14
B Synthesis of 2-Carbamoyl-5-nitro-7-(trifluoromethyl)
benzothiazole 3-oxide , Compound 15
B Synthesis of 2-(Hydrazinecarbonyl)-7-nitro-5-(trifluor-
omethyl)benzothiazole 3-oxide , Compound 16
B Expression and Purification of Plk1 for Kinase Assays
B Construction, Expression and Purification of Cdc25C
B Procedure for PLK1 Assay
SIGNIFICANCE
The Polo family of mitotic kinases are clinically validated
oncology targets. Structure-based discovery and design
has led to a class of inhibitors of PLK1 catalytic activity
B Cloning and Expression of PLK1 C67S
B Procedure for PLK1 and PLK1 C67S Kinase Assay
B Procedures for Molecular Modeling
6
7
that covalently interact with C of the ATP binding site.
Benzothiazole N-oxide derivatives reversibly inhibit PLK1
through the formation of Meisenheimer complexes, a mech-
d QUANTIFICATION AND STATISTICAL ANALYSIS
d DATA AND SOFTWARE AVAILABILITY
8
Cell Chemical Biology 25, 1–10, September 20, 2018

Please cite this article in press as: Pearson et al., The Meisenheimer Complex as a Paradigm in Drug Discovery: Reversible Covalent Inhibition through
C67 of the ATP Binding Site of PLK1, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.06.001
SUPPLEMENTAL INFORMATION
Elez, R., Piiper, A., Kronenberger, B., Kock, M., Brendel, M., Hermann, E.,
Pliquett, U., Neumann, E., and Zeuzem, S. (2003). Tumor regression by com-
bination antisense therapy against Plk1 and Bcl-2. Oncogene 22, 69–80.
Supplemental Information includes six figures and two tables and can be
found with this article online at https://doi.org/10.1016/j.chembiol.2018.
Elia, A.E., Cantley, L.C., and Yaffe, M.B. (2003a). Proteomic screen finds pSer/
06.001.
pThr-binding domain localizing Plk1 to mitotic substrates. Science 299,
1
228–1231.
ACKNOWLEDGMENTS
Elia, A.E., Rellos, P., Haire, L.F., Chao, J.W., Ivins, F.J., Hoepker, K.,
Mohammad, D., Cantley, L.C., Smerdon, S.J., and Yaffe, M.B. (2003b). The
molecular basis for phosphodependent substrate targeting and regulation of
Plks by the Polo-box domain. Cell 115, 83–95.
We would like to thank many at Cyclacel who contributed to this project and
also especially acknowledge the Scottish Executive for provision of funding
through a SCORE award. We also acknowledge Drs Christopher Meades, Jan-
ice McLachlan, Andrew Osnowski, Andy Plater, Joshua Bolger, and James
Bozard for their assistance with compound synthesis and BPS Biosciences
for the cloning, expression of PLK1 C67S.
Feng, P., Lee, K.N., Lee, J.W., Zhan, C., and Ngai, M.Y. (2016). Access to a
new class of synthetic building blocks via trifluoromethoxylation of pyridines
and pyrimidines. Chem. Sci. 7, 424–429.
Gjertsen, B.T., and Schoffski, P. (2014). Discovery and development of the
Polo-like kinase inhibitor volasertib in cancer therapy. Leukemia 29, 11.
AUTHOR CONTRIBUTIONS
Glover, D.M., Hagan, I.M., and Tavares, A.A. (1998). Polo-like kinases: a team
that plays throughout mitosis. Genes Dev. 12, 3777–3787.
R.J.P. synthesized and characterized the molecules used in this study and
contributed to the writing of the paper. D.B. was involved in project manage-
ment and scientific direction. M.M. performed the in vitro kinase assays.
P.M.F. was involved in project management and scientific direction. N.J.W.
supervised the synthetic chemistry aspects of the study. C.McI. performed
the molecular modeling studies, was involved in project management and
also contributed extensively to the writing of the paper.
Golan, A., Yudkovsky, Y., and Hershko, A. (2002). The cyclin-ubiquitin ligase
activity of cyclosome/APC is jointly activated by protein kinases Cdk1-cyclin
B and Plk. J. Biol. Chem. 277, 15552–15557.
Hamanaka, R., Smith, M.R., O’Connor, P.M., Maloid, S., Mihalic, K., Spivak,
J.L., Longo, D.L., and Ferris, D.K. (1995). Polo-like kinase is a cell cycle-regu-
lated kinase activated during mitosis. J. Biol. Chem. 270, 21086–21091.
Jang, Y.-J., Lin, C.-Y., Ma, S., and Erikson, R.L. (2002). Functional studies on
the role of the C-terminal domain of mammalian polo-like kinase. Proc. Natl.
Acad. Sci. USA 99, 1984–1989.
DECLARATION OF INTERESTS
Dr. D. Blake is an employee of Cyclacel, and a shareholder of Cyclacel Phar-
maceuticals. Dr. McInnes, as well as an employee of the University of South
Carolina, is Founder, President and Chief Scientific Officer of PPI Pharmaceu-
ticals; however, this company was not involved with this published study. Drs.
McInnes, Mezna, and Fischer are co-inventors of a patent covering the benzo-
thiazole N-oxides described in this study (WO, 2004/05,700 A1).
Kloss, F., Krchnak, V., Krchnakova, A., Schieferdecker, S., Dreisbach, J.,
Krone, V., Mollmann, U., Hoelscher, M., and Miller, M.J. (2017). In vivo dearo-
matization of the potent antituberculosis agent BTZ043 via Meisenheimer
complex formation. Angew. Chem. Int. Ed. 56, 2187–2191.
Knecht, R., Elez, R., Oechler, M., Solbach, C., Von Ilberg, C., and Strebhardt,
K. (1999). Prognostic significance of polo-like kinase (PLK) expression in squa-
mous cell carcinomas of the head and neck. Cancer Res. 59, 2794–2797.
Received: July 19, 2017
Revised: September 25, 2017
Accepted: May 31, 2018
Published: July 12, 2018
Kotani, S., Tugendreich, S., Fujii, M., Jorgensen, P.M., Watanabe, N., Hoog,
C., Hieter, P., and Todokoro, K. (1998). PKA and MPF-activated polo-like
kinase regulate anaphase-promoting complex activity and mitosis progres-
sion. Mol. Cell 1, 371–380.
REFERENCES
Lee, K.S., Grenfell, T.Z., Yarm, F.R., and Erikson, R.L. (1998). Mutation of the
polo-box disrupts localization and mitotic functions of the mammalian polo
kinase Plk. Proc. Natl. Acad. Sci. USA 95, 9301–9306.
Alexandru, G., Uhlmann, F., Mechtler, K., Poupart, M.-A., and Nasmyth, K.
(
2001). Phosphorylation of the cohesin subunit Scc1 by Polo/Cdc5 kinase reg-
Lee, K.S., Yuan, Y.-L.O., Kuriyama, R., and Erikson, R.L. (1995). Plk is an
M-phase-specific protein kinase and interacts with a kinesin-like protein,
CHO1/MKLP-1. Mol. Cell. Biol. 15, 7143–7151.
ulates sister chromatid separation in yeast. Cell 105, 459–472.
Baillie, T.A. (2016). Targeted covalent inhibitors for drug design. Angew. Chem.
Int. Ed. 55, 13408–13421.
Leung, G.C., Hudson, J.W., Kozarova, A., Davidson, A., Dennis, J.W., and
Sicheri, F. (2002). The Sak polo-box comprises a structural domain sufficient
for mitotic subcellular localization. Nat. Struct. Biol. 9, 719–724.
Bunting, J.W. (1979). Heterocyclic Meisenheimer complexes. Adv. Heterocycl.
Chem. 25, 67–74.
Cohen, M.S., Zhang, C., Shokat, K.M., and Taunton, J. (2005). Structural bio-
informatics-based design of selective, irreversible kinase inhibitors. Science
McInnes, C., Mazumdar, A., Mezna, M., Meades, C., Midgley, C., Scaerou, F.,
Carpenter, L., Mackenzie, M., Taylor, P., Walkinshaw, M., et al. (2006).
Inhibitors of Polo-like kinase reveal roles in spindle-pole maintenance. Nat.
Chem. Biol. 2, 608–617.
308, 1318–1321.
Craig, S.N., Wyatt, M.D., and McInnes, C. (2014). Current assessment of
polo-like kinases as anti-tumor drug targets. Expert Opin. Drug Discov. 9,
McInnes, C., and Wyatt, M.D. (2011). PLK1 as an oncology target: current sta-
773–789.
tus and future potential. Drug Discov. Today 16, 619–625.
de Carcer, G., Escobar, B., Higuero, A.M., Garcia, L., Anson, A., Perez, G.,
Mollejo, M., Manning, G., Melendez, B., Abad-Rodriguez, J., et al. (2011).
Plk5, a polo box domain-only protein with specific roles in neuron differentia-
tion and glioblastoma suppression. Mol. Cell. Biol. 31, 1225–1239.
Nurse, P. (1990). Universal control mechanism regulating onset of M-phase.
Nature 344, 503–508.
Roshak, A.K., Capper, E.A., Imburgia, C., Fornwald, J., Scott, G., and
Marshall, L.A. (2000). The human polo-like kinase, PLK, regulates cdc2/cyclin
B through phosphorylation and activation of the cdc25C phosphatase. Cell.
Signal. 12, 405–411.
Dohner, H., Lubbert, M., Fiedler, W., Fouillard, L., Haaland, A., Brandwein,
J.M., Lepretre, S., Reman, O., Turlure, P., Ottmann, O.G., et al. (2014).
Randomized, phase 2 trial comparing low-dose cytarabine with or without vol-
asertib in AML patients not suitable for intensive induction therapy. Blood 124,
Seong, Y.-S., Kamijo, K., Lee, J.-S., Fernandez, E., Kuriyama, R., Miki, T., and
Lee, K.S. (2002). A spindle checkpoint arrest and a cytokinesis failure by the
dominant-negative polo-box domain of Plk1 in U-2 OS cells. J. Biol. Chem.
1426–1433.
2
77, 32282–32293.
Elez, R., Piiper, A., Giannini, C.D., Brendel, M., and Zeuzem, S. (2000). Polo-
like kinase 1, a new target for antisense tumor therapy. Biochem. Biophys.
Res. Commun. 269, 352–356.
Singh, J., Dobrusin, E.M., Fry, D.W., Haske, T., Whitty, A., and McNamara, D.J.
(1997). Structure-based design of a potent, selective, and irreversible inhibitor
Cell Chemical Biology 25, 1–10, September 20, 2018
9

Please cite this article in press as: Pearson et al., The Meisenheimer Complex as a Paradigm in Drug Discovery: Reversible Covalent Inhibition through
C67 of the ATP Binding Site of PLK1, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.06.001
of the catalytic domain of the erbB receptor subfamily of protein tyrosine
Toyoshima-Morimoto, F., Taniguchi, E., and Nishida, E. (2002). Plk1 promotes
kinases. J. Med. Chem. 40, 1130–1135.
nuclear translocation of human Cdc25C during prophase. EMBO Rep. 3,
3
41–348.
Smith, M.R., Wilson, M.L., Hamanaka, R., Chase, D., Kung, H., Longo, D.L.,
and Ferris, D.K. (1997). Malignant transformation of mammalian cells initiated
by constitutive expression of the polo-like kinase. Biochem. Biophys. Res.
Commun. 234, 397–405.
Wagner, K., Heitzer, H., and Oehlmann, L. (1973). Benzthiazol-N-oxide, I.
Synthese und Reaktivitat von 2-Alkoxycarbonyl-benzthiazol-N-oxiden.
€
Chem. Ber. 106, 640–654.
Spankuch-Schmitt, B., Bereiter-Hahn, J., Kaufmann, M., and Strebhardt, K.
Wagner, K., and Oehlmann, L. (1976). Benzothiazol-N-oxide, IV1) Synthese und
Reaktivit a€ t von2-Cyanbenzothiazol-N-Oxiden und 2-Benzothiazolcarbonitrilen.
(
2002a). Effect of RNA silencing of polo-like kinase-1 (PLK1) on apoptosis
and spindle formation in human cancer cells. J. Natl. Cancer Inst. 94,
863–1877.
Chem. Ber. 109, 611–618.
1
Wolf, G., Elez, R., Doermer, A., Holtrich, U., Ackermann, H., Stutte, H.J.,
Altmannsberger, H.-M., R u€ bsamen-Waigmann, H., and Strebhardt, K.
Spankuch-Schmitt, B., Wolf, G., Solbach, C., Loibl, S., Knecht, R., Stegmuller,
M., von Minckwitz, G., Kaufmann, M., and Strebhardt, K. (2002b).
Downregulation of human polo-like kinase activity by antisense oligonucleo-
tides induces growth inhibition in cancer cells. Oncogene 21, 3162–3171.
(
1997). Prognostic significance of polo-like kinase (PLK) expression in non-
small cell lung cancer. Oncogene 14, 543–549.
Wu, S.Y., McNae, I., Kontopidis, G., McClue, S.J., McInnes, C., Stewart, K.J.,
Wang, S., Zheleva, D.I., Marriage, H., Lane, D.P., et al. (2003). Discovery of a
novel family of CDK inhibitors with the program LIDAEUS: structural basis for
ligand-induced disordering of the activation loop. Structure 11, 399–410.
Sumara, I., Gimenez-Abian, J.F., Gerlich, D., Hirota, T., Kraft, C., de la Torre,
C., Ellenberg, J., and Peters, J.M. (2004). Roles of polo-like kinase 1 in the
assembly of functional mitotic spindles. Curr. Biol. 14, 1712–1722.
Takahashi, T., Sano, B., Nagata, T., Kato, H., Sugiyama, Y., Kunieda, K.,
Kimura, M., Okano, Y., and Saji, S. (2003). Polo-like kinase 1 (PLK1) is overex-
pressed in primary colorectal cancers. Cancer Sci. 94, 148–152.
Yuan, J., Hoerlin, A., Hock, B., Stutte, H.J., Ruebsamen-Waigmann, H., and
Strebhardt, K. (1997). Polo-like kinase, a novel marker for cellular proliferation.
Am. J. Pathol. 150, 1165–1172.
Taylor, P., Blackburn, E., Sheng, Y.G., Harding, S., Hsin, K.Y., Kan, D., Shave,
S., and Walkinshaw, M.D. (2008). Ligand discovery and virtual screening using
the program LIDAEUS. Br. J. Pharmacol. 153 (Suppl 1), S55–S67.
Yuan, J., Sanhaji, M., Kramer, A., Reindl, W., Hofmann, M., Kreis, N.N.,
Zimmer, B., Berg, T., and Strebhardt, K. (2011). Polo-box domain inhibitor po-
loxin activates the spindle assembly checkpoint and inhibits tumor growth
in vivo. Am. J. Pathol. 179, 2091–2099.
Taylor, R. (1970). Novel Meisenheimer complexes, alkyl-2,4,6-trinitrocyclo-
hexadienate anions. J. Org. Chem. 35, 3578–3579.
Tokumitsu, Y., Mori, M., Tanaka, S., Akazawa, K., Nakano, S., and Niho, Y.
Zhao, Z., Liu, Q., Bliven, S., Xie, L., and Bourne, P.E. (2017). Determining cys-
teines available for covalent inhibition across the human kinome. J. Med.
Chem. 60, 2879–2889.
(1999). Prognostic significance of polo-like kinase expression in esophageal
carcinoma. Int. J. Oncol. 15, 687–692.
10 Cell Chemical Biology 25, 1–10, September 20, 2018

Please cite this article in press as: Pearson et al., The Meisenheimer Complex as a Paradigm in Drug Discovery: Reversible Covalent Inhibition through
C67 of the ATP Binding Site of PLK1, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.06.001
STAR+METHODS
KEY RESOURCES TABLE
REAGENT or RESOURCE
Chemicals, Peptides, and Recombinant Proteins
Triethylamine
SOURCE
IDENTIFIER
Sigma-Aldrich
Sigma-Aldrich
Sigma-Aldrich
TCI Eurpoe
T0886
ethyl thioglycolate
E34307
338818
ammonium hydroxide solution
2
4
2
2
2
2
2
-chloro-5-trifluoromethyl-1,3-dinitrobenzene
C0994
-fluoro-2,6-dinitrophenol
Fisher Scientific
ACES Pharma, Inc.
carbosynth Ltd
carbosynth Ltd
Sigma-Aldrich
NONE
11403557
112587
-chloro-5-iodo-1,3-dinitrobenzene
-chloro-5-methoxycarbonyl-1,3-dinitrobenzene
-chloro-5-cyano-1,3-dinitrobenzene
-chloro-5-methyl-1,3-dinitrobenzene
-chloro-5-(trifluoromethylthio)-1,3-dinitrobenzene
FM67720
FC70294
PH011654
Can be synthesized according to:
Journal of Fluorine Chemistry,
125(9), 1305-1316; 2004
sodium hydroxide
phosphorus oxychloride
Pyridine
Sigma-Aldrich
Sigma-Aldrich
Sigma-Aldrich
Sigma-Aldrich
Sigma-Aldrich
CambridgeChem
Sigma-Aldrich
ThermoFisher
Clontech
795429
262099
270970
hydroxylamine solution
triethylphosphine
467804
245275
2-chloro-1,5-dinitro-3-(trifluoromethyl)benzene
318514052
751855
Hydrazine solution
Sf9 cells in Sf-900ꢀ II SFM
Plk1 open reading frame (X75932)
BL21(DE3)pLysS Chemically Competent E. coli
Critical Commercial Assays
PLK1 Kinase assay
11496015
HA040548
C606003
ThermoFisher
BPS Biosciences LLC
PLK1
Software and Algorithms
Insight II
Biovia
Affinity
PyMol
Schrodinger
https://pymol.org/2/
CONTACT FOR REAGENT AND RESOURCE SHARING
Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Campbell
McInnes (mcinnes@cop.sc.edu).
METHOD DETAILS
General Method for the Synthesis of 2-alkoxycarbonyl-N-oxide Derivatives
See Figure S5 for synthetic scheme (McInnes et al., 2006). Step 1: Triethylamine (1.1eq) was added to a suspension of 5 substituted,
3
,3-dinitrobenzyl-2-chloride (1eq) and ethyl thioglycolate (1eq) in ethanol (3cm ) at 10-20 C (Supplementary Figure 5.). Onset of the
ꢁ
1
reaction was observed by the solvent beginning to reflux even though the mixture was being cooled in a water bath. After stirring for a
further 3h a precipitate formed which was collected by filtration, washed with ice-cold water followed by a further wash with ice-cold
methanol. Recrystallization from methanol gave the desired ethyl ester derivatives. Step 2: The 2-alkoxycarbonylbenzothiazol-3-ox-
3
ꢁ
ide (1eq) was suspended in ethanol (2 cm ) at 20-30 C and ammonium hydroxide (1.2eq) added. After stirring for 4h, the desired
-carbamoyl benzthiazol-3-oxide crystallised from solution, was collected by filtration and was washed with ice-cold water and
then methanol. Recrystallisation from methanol gave the desired amide derivatives. For further information see McInnes et al., 2006.
2
Cell Chemical Biology 25, 1–10.e1–e4, September 20, 2018 e1

Please cite this article in press as: Pearson et al., The Meisenheimer Complex as a Paradigm in Drug Discovery: Reversible Covalent Inhibition through
C67 of the ATP Binding Site of PLK1, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.06.001
Synthesis of 2-Carbamoyl-7-nitro-5-(trifluoromethyl)benzothiazole 3-oxide, Compound 4
The general method above was followed using 2-chloro-5-trifluoromethyl-1,3-dinitrobenzene. The title compound was obtained as a
1
yellow crystalline solid: H NMR (DMSO-d
+
8.76 (2H, s, Ar-H and NH), 8.96 (1H, s, Ar-H) and 9.36 (1H, s, NH). MS (ESI )
6
, 500 MHz): d
H
+
m/z 308.01 [M+H] (C
ꢁ
ꢁ
9
H
5
N
3
O
4
F
3
S requires 308.00). mp 222-223 C (lit 225-226 C, Wagner et al., 1973). Anal. RP-HPLC: t
R
13.75 min.
(
10-70% MeCN).
Synthesis of 4-Fluoro-2,6-dinitrophenyl 4-methylbenzenesulfonate, Precursor for Compound 5
3
-Fluoro-2,6-dinitrophenol (1eq) was dissolved in toluene (5cm ) and tosyl chloride (1.1eq) added followed by pyridine (1.1eq)
4
(
Figure S6). After stirring for 12h a white precipitate formed which was collected by filtration. The mother liquor was evaporated to
1
dryness and used without further purification. H NMR (DMSO-d
.67 (2H, d, J 8.0, 2 x Ar-H) and 8.51 (2H, d, J 8.0, 2 x Ar-H).
6 H 3
, 500 MHz): d 2.48 (3H, s, CH ), 7.50 (2H, d, J 8.0, 2 x Ar-H),
7
Synthesis of 2-Carbamoyl-5-fluoro-7-nitrobenzothiazole 3-oxide, Compound 5
The title compound was obtained as a yellow crystalline solid from 4-fluoro-2,6-dinitrophenyl 4-methylbenzenesulfonate using the
1
general method for the synthesis of 2-alkoxycarbonyl-N-oxides. H NMR (DMSO-d
J 7.5, Ar-H), 8.97 (1H, d, J 7.5, Ar-H) and 9.41 (1H, s, NH). C
Anal. RP-HPLC: t 9.83 min. (10-70% MeCN).
6
, 500 MHz): d
H
8.82 (1H, s, NH), 8.88 (1H, d,
ꢁ
8
H
5
N
3
O
4
FS requires M 257.99848, found 257.99861. mp 282-283 C.
R
Synthesis of 2-Carbamoyl-5-iodo-7-nitrobenzothiazole 3-oxide, Compound 6
The title compound was obtained as a yellow crystalline solid from 2-chloro-5-iodo-1,3-dinitrobenzene using the general method.
1
6 H 8 4 3 4
H NMR (DMSO-d , 500 MHz): d 8.81 (1H, s, Ar-H), 8.82 (1H, s, NH), 8.94 (1H, s, Ar-H), and 9.41 (1H, s, NH). C H N O IS requires
M 364.89673, found 364.89634.
Synthesis of 2-carbamoyl-5-(methoxycarbonyl)-7-nitrobenzothiazole 3-oxide, Compound 7
The title compound was obtained as a yellow crystalline solid from 2-chloro-5-methoxycarbonyl-1,3-dinitrobenzene using the
1
6 H 3
general method. H NMR (DMSO-d , 300 MHz): d 3.89 (3H, s, CH ), 8.78 (4H, s, 2 Ar-H and 2 NH).
Synthesis of 2-Carbamoyl-5-cyano-7-nitrobenzothiazole 3-oxide, Compound 8
The title compound was obtained as a yellow crystalline solid from 2-chloro-5-cyano-1,3-dinitrobenzene using the general method.
1
H NMR (DMSO-d
6
, 500 MHz): d
H
8.86 (1H, s, NH), 9.16 (1H, s, Ar-H), 9.22 (1H, s, Ar-H) and 9.33 (1H, s, NH). C
ꢁ
9
H
4
N
4
O
4
S requires
19.60 min.
ꢁ
M 263.9953, found 263.9951. mp 222-223 C (lit 225-226 C, Wagner and Oehlmann., 1976), Anal. RP-HPLC: t
10-70% MeCN).
R
(
Synthesis of 2-Carbamoyl-5-methyl-7-nitrobenzothiazole 3-oxide , Compound 9
The title compound was obtained as a yellow crystalline solid from 2-chloro-5-methyl-1,3-dinitrobenzene using the general method.
ꢁ
ꢁ
R
mp 271-272 C (lit 274-275 C, Wagner et al., 1973). Anal. RP-HPLC: t 11.32 min. (10-70% MeCN).
Synthesis of 2-Carbamoyl-7-nitro-5-((trifluoromethyl)thio)benzothiazole 3-oxide, Compound 10
The title compound was obtained as a yellow crystalline solid from 2-chloro-5-(trifluoromethylthio)-1,3-dinitrobenzene using the
1
general method. H NMR (DMSO-d
6
, 500 MHz): d
requires M 339.96736, found 339.96739. mp 255-256 C, Anal. RP-HPLC: t
H
8.82 (1H, s, Ar-H), 8.85 (1H, s, Ar-H), 8.90 (1H, s, NH) and 9.35 (1H, s, NH).
ꢁ
C
9
H
5
N
3
O
4
F
3
S
2
R
15.79 min. (10-70% MeCN).
Synthesis of 2-Carbamoyl-5-carboxy-7-nitrobenzothiazole 3-oxide, Compound 11
Compound 7 (26mg, 0.09mmol) was suspended in water (1.0ml) containing methanol (0.1ml). Sodium hydroxide (4mg, 0.1mmol) in
water (0.36ml) was added and the resultant precipitate collected by filtration and washed with water. The title compound was
1
obtained as a yellow crystalline solid, H NMR (DMSO-d
6
, 500 MHz): d
ꢁ
H
7.88 (1H, s, Ar-H) and 8.35 (1H, s, Ar-H), C
16.38 min. (10-70% MeCN).
9 5 3 6
H N O S requires
ꢁ
M 282.9899, found 282.9900. mp 181-183 C (lit 188-190 C, Wagner et al., 1973), t
R
Synthesis of 2-Cyano-7-nitro-5-(trifluoromethyl)benzothiazole 3-oxide , Compound 12
Phosphorus oxychloride (306mg, 0.186ml, 2.0mmol) was added to a suspension of 4 (672mg, 2.0 mmol) in dry pyridine (1ml) at
ꢁ
2
0-30 C and stirred for 12h (Wagner and Oehlmann., 1976). Ice-water was added and the resultant precipitate collected by filtration.
1
The title compound was obtained as a yellow crystalline solid. H NMR (DMSO-d
6
, 500 MHz): d
H
8.91 (1H, s, Ar-H) and 9.02 (1H, s,
ꢁ
ꢁ
Ar-H). mp 188-189 C (lit 190-191 C, Wagner et al., 1973), t
R
15.60 min. (10-70% MeCN).
Synthesis of 2-(Hydroxycarbamoyl)-7-nitro-5-(trifluoromethyl)benzothiazole 3-oxide , Compound 13
The title compound was obtained as a yellow crystalline solid from the reaction of compound 4 with hydroxylamine using routine
1
methodology (Wagner et al., 1973). H NMR (DMSO-d
6
, 500 MHz): d
H
8.86 (1H, s, Ar-H), 8.96 (1H, s, Ar-H), 10.17 and 12.13
ꢁ
+
+
+
2H, s, NH and OH). MS (ESI ) m/z 346.00 [M+Na] and 324.07 [M+H] (C
(
9
H
4
N
3
O
5
F
3
S requires 322.98). mp 230-231 C (lit 234-
ꢁ
2
35 C, Wagner et al., 1973). Anal. RP-HPLC: t
R
12.33 min. (10-70% MeCN).
e2 Cell Chemical Biology 25, 1–10.e1–e4, September 20, 2018

Please cite this article in press as: Pearson et al., The Meisenheimer Complex as a Paradigm in Drug Discovery: Reversible Covalent Inhibition through
C67 of the ATP Binding Site of PLK1, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.06.001
Synthesis of 2-Carbamoyl-7-nitro-5-(trifluoromethyl)benzothiazole , Compound 14
ꢁ
Compound 4 (30mg, 0.097 mmol) was dissolved in ethanol (2ml) and heated to 70 C. Triethyl phosphine (17mg, 1.1eq) was added
ꢁ
and the mixture remained at 70 C for a further 4h. On cooling the resultant precipitate was collected by filtration. The title compound
1
was obtained as a yellow crystalline solid: H NMR (DMSO-d
6
, 500 MHz): d
H
8.36 (1H, s, NH), 8.68 (1H, s, NH), 8.77 (1H, s, Ar-H) and
16.01 min. (10-70% MeCN).
ꢁ
ꢁ
8
.92 (1H, s, Ar-H). mp 222-224 C (lit 225-226 C, Wagner and Oehlmann., 1976). Anal. RP-HPLC: t
R
Synthesis of 2-Carbamoyl-5-nitro-7-(trifluoromethyl)benzothiazole 3-oxide , Compound 15
The title compound was obtained as a yellow crystalline solid from 2-chloro-1,5-dinitro-3-(trifluoromethyl)benzene using the general
1
6 H 9 4 3 3 4
method. H NMR (DMSO-d , 500 MHz): d 8.88 (1H, s, Ar-H), 8.90 (1H, s, NH), 8.97 (1H, s, Ar-H) and 9.33 (1H, s, NH), C H F N O S
ꢁ
ꢁ
requires M 306.9875, found 306.9880. mp 224-225 C (lit 225-226 C, Wagner et al., 1973).
Synthesis of 2-(Hydrazinecarbonyl)-7-nitro-5-(trifluoromethyl)benzothiazole 3-oxide , Compound 16
The title compound was obtained as a yellow crystalline solid from the reaction of compound 4 with hydrazine using standard
1
procedures (Wagner et al., 1973). H NMR (DMSO-d
6
, 500 MHz): d
H
5.20 (2H, s, NH
1.20 (1H, s, NH). mp 229-230 C (lit 232-233 C, Wagner et al., 1973). Anal. RP-HPLC: t
2
), 8.88 (1H, s, Ar-H), 8.95 (1H, s, Ar-H) and
11.88 min. (10-70% MeCN).
ꢁ
ꢁ
1
R
Expression and Purification of Plk1 for Kinase Assays
The Plk1 open reading frame (X75932) was amplified from a fetal lung complementary DNA library (Clontech) using primers incorpo-
0
0
rating restriction enzyme sites. The 5 primer (5 -GCCGCTAGCGACGATGACGATAAGATGAGTGCTGCAGTGACTGCAGGGA
0
0
0
0
AGC-3 ) had an NheI site upstream of the ATG start codon. The 3 primer (5 -GGAATTCTTAGGAGGCCTTGAGACGG-3 ) incorpo-
rated an EcoRI site downstream of the stop codon. The PCR product was subcloned into the NheI/EcoRI sites of a baculovirus
expression vector (pSSP1) derived from pFastBac Hta (Invitrogen). Cloning into this vector resulted in a hexahistidine-tag fusion
at the N terminus of the Plk1 construct. Sf9 strain cells of a passage number less than 20 were split back to give a 300-ml culture
volume, at a cell density of 1.5 3 106 cells mlꢀ1. Cells were only used for expression in logarithmic growth phase. Plk1 baculovirus
(
from P2 amplification) was added to give a multiplicity of infection of 3; this is equivalent to three virus particles for each insect cell.
ꢁ
The flasks were incubated at 27 C, with shaking at 100 r.p.m., for 48 h. On harvest, cell density and viability were determined, and the
ꢁ
cultures were spun down at 2,500 r.p.m. for 5 min and washed with ice-cold (0 C) phosphate-buffered saline. The wash was re-spun
at the same speed and the pellet was snap frozen. Plk1 protein was purified on a metal affinity column. The insect cell pellet was lysed
in a buffer (10 mM Tris-HCl, pH 8.0, 150 mM NaCl, 5 mM b-mercaptoethanol, 1 mM PMSF, 1 mM benzamidine, 20 mM imidazole and
protease inhibitor cocktail (Sigma)) and the precleared supernatant was loaded onto Ni-NTA-agarose (Qiagen). The affinity column
was washed with the lysis buffer and the bound protein was eluted with 250 mM imidazole in the same buffer. After overnight dialysis
against 25 mM Tris-HCl, pH 7.5, 100 mM NaCl, 1 mM DTT, 1 mM PMSF, 1 mM benzamidine, protease inhibitor cocktail (Sigma) and
ꢁ
1
0% glycerol, the purified protein was stored at ꢀ70 C until used.
Construction, Expression and Purification of Cdc25C
Using standard techniques a full-length Cdc25C clone was isolated by PCR from HeLa mRNA and inserted on a BamHI-HindIII
fragment into the plasmid pRsetA. The N-terminal Cdc25C fragment (encoding residues 1–300) was excised from this vector and
inserted into the plasmid pET28a (between the NcoI and BamHI sites). Expression was under the control of the T7 promoter, and
the encoded protein contains a hexahistidine tag at the C terminus. The vector was transformed into E. coli strain BRL(DE3) pLysS
ꢁ
for expression experiments. The protein was expressed in BL21(DE3) RIL bacteria cells that were grown in LB medium at 37 C until
optical density at 600 nm of 0.6 was reached. Expression was induced with 1 mM IPTG, and the bacterial culture was grown further
for 3 h. The bacteria were harvested by centrifugation and the cell pellet resuspended in 50 mM Tris, pH 7.5, and 10% sucrose, flash
ꢁ
frozen, and stored at ꢀ70 C until used. Purification of the protein was then carried out by lysing the bacterial pellet in 10 ml of lysis
buffer (10 mM Tris-HCl, pH 8.0, 150 mM NaCl, 5 mM b-mercaptoethanol and 20mM imidazole) supplemented with a cocktail of
protease inhibitors and sonicated six times at 20-s bursts. The lysate was then centrifuged for 15 min at 15,000 r.p.m. and filtered
through a 0.45-mm filter. The sample was then loaded onto a Ni-NTA agarose column and washed several times, and then the
Cdc25 protein was eluted with a buffer containing 10 mM Tris-HCl, pH 8.0, 100 mM NaCl, 5 mM b-mercaptoethanol, 0.02% Nonidet
P-40 (EMD Millipore) and 250 mM imidazole. The eluate was then dialyzed, concentrated, snap frozen in liquid nitrogen and stored
ꢁ
at ꢀ70 C until used.
Procedure for PLK1 Assay
PLK1 kinase activity (compounds in Tables 1 and 2) was assayed using either Casein or the N-terminal domain of the human Cdc25C
phosphatase (a natural substrate for PLK1) as substrates. The assays were carried out using a 96-well plate format by incubating
Cdc25C (2 mg/well) or Casein (50mg/well) with varying concentrations of the Ni-NTA purified PLK1 and varying concentrations of
the negative, non-transformed host Sf9 cell lysate in a total volume of 25ul of 20 mM Tris/HCl buffer pH 7.0, supplemented with
2
3
5 mM b-glycerophosphate, 5 mM EGTA, 1 mM DTT and 1 mM NaVO . Reaction was initiated by the addition of 100 mM ATP
3
2
ꢁ
and 0.5 mCi of [g- P]-ATP. The reaction mixture was incubated at 30 C for 1 h, then stopped with 75 mM aq orthophosphoric
acid, transferred onto a 96-well P81 filter plate (Whatman), dried, and the extent of Cdc25C phosphorylation was assessed by
scintillation counting using a Packard TopCount plate reader.
Cell Chemical Biology 25, 1–10.e1–e4, September 20, 2018 e3

Please cite this article in press as: Pearson et al., The Meisenheimer Complex as a Paradigm in Drug Discovery: Reversible Covalent Inhibition through
C67 of the ATP Binding Site of PLK1, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.06.001
Cloning and Expression of PLK1 C67S
This was undertaken at BPS Bioscience (San Diego). After C67S mutation and C-terminal His tag were introduced by PCR, the insert
was cloned into pFastBac vector (Thermo Fisher Scientific), followed by sequence analysis. The plasmid was transformed into
DH10Bacꢀ E. coli Competent Cells (Thermo Fisher Scientific) where it recombined with the parent bacmid to form an expression
bacmid. The purified bacmid DNA was tested for recombination by PCR. Then the bacmid DNA was transfected into insect cells
for production of recombinant baculovirus particles. P3 virus was used for protein production. Protein was expressed in Sf9 cells
using recombinant baculovirus at a MOI of 3 grown in SF900II media (Gibcoꢀ). Cell were harvest by centrifugation three days after
transfection and His tagged protein was purified using Ni-NTA affinity. Briefly, cells were lysed by sonication after resuspension in
TBS-T buffer. Protein was purified using Ni2+-NTA agarose (Sigma) eluting with TBS-T buffer plus imidazole. High purity elutions
as assessed by SDS PAGE were pooled with glycerol and DTT added to a final concentration of 20% and 3mM respectively. Protein
was quantified by Bradford reagent (Pierceꢀ) using BSA as the standard. Purity was determined to be >90% by SDS PAGE.
Procedure for PLK1 and PLK1 C67S Kinase Assay
For the data shown in Figures 5 and S3, the assay was performed by BPS Bioscience (San Diego) using Kinase-Glo Plus lumines-
cence kinase assay kit (obtained from Promega). It measures kinase activity by quantitating the amount of ATP remaining in solution
following a kinase reaction. The luminescent signal from the assay is correlated with the amount of ATP present and is inversely corre-
lated with the amount of kinase activity. The compounds were diluted in 10% DMSO and 5ml of the dilution was added to a 50ml
ꢁ
reaction so that the final concentration of DMSO is 1% in all of reactions. All of the enzymatic reactions were conducted at 30 C
for 45 minutes for the wild type PLK1 and 60 min for PLK1(C67S). The 50ml reaction mixture contains 40 mM Tris, pH 7.4, 10 mM
MgCl2, 0.1 mg/ml BSA, 1 mM DTT, 0.1 mg/ml PLK1 peptide substrate, 10 mM (or 5 mM) ATP and PLK1. After the enzymatic reaction,
50 ml of Kinase-Glo Plus Luminescence kinase assay solution (Promega) was added to each reaction and incubate the plate for 15 mi-
nutes at room temperature. The luminescence signal was measured using a BioTek Synergy 2 microplate reader
Procedures for Molecular Modeling
The homology model for PLK1 was generated using the program module Homology within the molecular modelling package InsightII
(
Accelrys, San Diego, CA) using PKA as a template structure. Compounds were docked with the PLK1 homology model using the
˚
Affinity program within InsightII implementing a molecular dynamics docking routine. The binding site was defined as an 8 A radius
from the centre of a ligand placed in the ATP site. The calculation was performed using the CVFF force field in a two-step process and
an implicitly derived solvation model and geometric H-bond restraints. Initially, the inhibitor was minimized into the ATP cleft, using a
simple non-bonded method where the Coulombic and Van der Waals terms are scaled to zero and 0.1, respectively. The refinement
phase involved molecular dynamics calculated over 5 ps in 100 fs stages, where the temperature was scaled from 500 K to 300 K
followed by a final minimisation over 1,000 steps using the Polak-Ribiere Conjugate Gradient method. The docked structures
were ranked energetically using the in-house developed programs Calsor and Calsorcont and the Ludi module of InsightII.
QUANTIFICATION AND STATISTICAL ANALYSIS
PLK1 activity assays were performed in duplicate at each concentration. The luminescence data were analyzed using the computer
software, Graphpad Prism. The difference between luminescence intensities in the absence of PLK1 (Lut) and in the presence of
PLK1 (Luc) was defined as 100 % activity (Lut – Luc). Using luminescence signal (Lu) in the presence of the compound, % activity
was calculated as:
% activity = {(Lut - Lu)/(Lut - Luc)}3100%, where Lu= the luminescence intensity in the presence of the compound (all percent
activities below zero were shown zero in the table).
The values of % activity versus a series of compound concentrations were then plotted using non-linear regression analysis of
Sigmoidal dose-response curve generated with the equation Y=B+(T-B)/1+10((LogEC50-X)3Hill Slope), where Y=percent activity,
B=minimum percent activity, T=maximum percent activity, X= logarithm of compound and Hill Slope=slope factor or Hill coefficient.
The IC50 value was determined by the concentration causing a half-maximal percent activity.
DATA AND SOFTWARE AVAILABILITY
No data or software beyond what is included in this manuscript and supplementary information are available.
e4 Cell Chemical Biology 25, 1–10.e1–e4, September 20, 2018