Structural and chemical insights into the covalent-allosteric inhibition of the protein kinase Akt

Article abstract of DOI:10.1039/c8sc05212c

The Ser/Thr kinase Akt (Protein Kinase B/PKB) is a master switch in cellular signal transduction pathways. Its downstream signaling influences cell proliferation, cell growth, and apoptosis, rendering Akt a prominent drug target. The unique activation mechanism of Akt involves a change of the relative orientation of its N-terminal pleckstrin homology (PH) and the kinase domain and makes this kinase suitable for highly specific allosteric modulation. Here we present a unique set of crystal structures of covalent-allosteric interdomain inhibitors in complex with full-length Akt and report the structure-based design, synthesis, biological and pharmacological evaluation of a focused library of these innovative inhibitors.

Full text of DOI:10.1039/c8sc05212c

Chemical
Science
View Article Online
View Journal
EDGE ARTICLE
Structural and chemical insights into the covalent-
allosteric inhibition of the protein kinase Akt†
Cite this: DOI: 10.1039/c8sc05212c
a
a
Niklas Uhlenbrock, ‡a Steven Smith,‡a Jorn Weisner,
Ina Landel,
All publication charges for this article
have been paid for by the Royal Society
of Chemistry
¨
Marius Lindemann,a Thien Anh Le,b Julia Hardick,
Rajesh Gontla,a
a
Rebekka Scheinpflug,a Paul Czodrowski,
Petra Janning,c Laura Depta,a
a
a
b
a
Lena Quambusch,a Matthias P. Muller,
¨
Bernd Engels
and Daniel Rauh
*
The Ser/Thr kinase Akt (Protein Kinase B/PKB) is a master switch in cellular signal transduction pathways. Its
downstream signaling influences cell proliferation, cell growth, and apoptosis, rendering Akt a prominent
drug target. The unique activation mechanism of Akt involves a change of the relative orientation of its
N-terminal pleckstrin homology (PH) and the kinase domain and makes this kinase suitable for highly
specific allosteric modulation. Here we present a unique set of crystal structures of covalent-allosteric
interdomain inhibitors in complex with full-length Akt and report the structure-based design, synthesis,
biological and pharmacological evaluation of a focused library of these innovative inhibitors.
Received 22nd November 2018
Accepted 12th February 2019
DOI: 10.1039/c8sc05212c
rsc.li/chemical-science
decades.8 A well-established approach in addressing protein
kinases has been the development of orthosteric inhibitors that
Introduction
bind to the active site of the kinase domain in an ATP-
competitive manner.9 A multitude of potent inhibitors such as
ipatasertib, based on a cyclopentapyrimidine-scaffold, and the
thiophenecarboxamide-derivative afuresertib have been identi-
ed and have entered phase I/II studies for mono- or combi-
nation therapy for a variety of indications.10–14 However, the
ATP-binding pocket of Akt is highly conserved among all three
isoforms of Akt and among several other kinases of the AGC
kinase superfamily, making selectivity an issue for this type of
inhibitors.15 In contrast to orthosteric inhibitors, allosteric
kinase inhibitors that bind at remote sites of the protein are
capable of inhibiting the kinase by stabilization of inactive
conformations, and can lead to great benets with respect to
selectivity.16,17
Due to the pleckstrin homology (PH) domain-mediated
regulation mechanism of Akt, targeting the interdomain
region between the kinase and the PH domain enables the
stabilization of the inactive “PH-in” conformation by allosteric
inhibitors.18 Initially identied by serendipity, a small number
of potent PH domain-dependent inhibitors have been devel-
oped to target this interdomain region and have resulted in the
clinical lead candidates MK-2206 19–21 and miransertib.22–24
Besides their benets in selectivity, it was shown recently that
the conformation-dependent, but kinase-independent, func-
tions of Akt are linked to cancer cell survival.23 Hence, stabi-
lizers of distinct kinase conformations could contribute not
only to a better understanding of this function of Akt beyond
catalysis, but also pave the way for allosteric Akt inhibitors in
a clinical setting.25,26 In view of this, we recently combined the
characteristics of allosteric Akt modulators with the benecial
The modulation of protein kinases with small molecule inhib-
itors accounts for a large share of the targeted therapy of a wide
range of diseases including cancer, inammatory diseases,
diabetes, and neurodegenerative disorders.1,2 The crucial role of
kinases with respect to protein phosphorylation in cellular
signaling makes this class of proteins important and attractive
targets in drug discovery.3 As a key player in the PI3K/Akt/
mTOR-pathway, the Ser-/Thr-specic kinase Akt, also known
as protein kinase B or PKB, reects the signicance of protein
kinases in cellular processes.4 Alterations and dysregulation in
the PI3K/Akt signaling pathway are related to different types of
solid tumors such as lung, prostate, endometrial, cervical
cancer, and melanoma.5 Furthermore, activating mutations of
Akt as well as overexpression have been identied as disease
drivers in certain metastatic breast cancers and are oen related
to resistance against chemo- and radiotherapy.6,7 These features
suggest a promising potential of the targeted modulation of Akt
with small molecule inhibitors in disease treatment and have
motivated the development of selective Akt inhibitors in recent
aFaculty of Chemistry and Chemical Biology, TU Dortmund University, Drug Discovery
¨
Hub Dortmund (DDHD) am Zentrum fur integrierte Wirkstoffforschung (ZIW),
Otto-Hahn-Strasse 4a, 44227 Dortmund, Germany. E-mail: daniel.rauh@
tu-dortmund.de; Web: http://www.ddhdortmund.de; www.twitter.com/DDHDortmund
b
¨
Faculty for Chemistry and Pharmacy, Institut fur Physikalische und Theoretische
¨
¨
¨
Chemie, Universitat Wurzburg, Emil-Fischer-Strasse 42, 97074 Wurzburg, Germany
c
¨
Max-Planck-Institut fur Molekulare Physiologie, Abteilung Chemische Biologie, Otto-
Hahn-Strasse 11, 44227 Dortmund, Germany
† Electronic supplementary information (ESI) available. See DOI:
10.1039/c8sc05212c
‡ These authors contributed equally to this work.
This journal is © The Royal Society of Chemistry 2019
Chem. Sci.

View Article Online
Chemical Science
Edge Article
Fig. 1 Covalent-allosteric Akt inhibitors irreversibly stabilize the inactive PH-in conformation. (A) Crystal structure of full-length Akt in complex
with borussertib (1, highlighted in blue, PDB: 6HHF) shows covalent-allosteric mode of action while binding at the interdomain region between
the kinase-domain (white) and the PH-domain (green). Detailed view of the binding mode of borussertib (right) reveals covalent bond formation
to Cys296, H-bond interactions are illustrated with dotted lines. (B) Schematic representation of the key interactions of borussertib to the target
protein illustrate crucial p–p-stacking with Trp80 and water-mediated H-bond-interactions. (C) Structural analysis inspired the design of novel
derivatives to probe Akt functions.
properties of irreversible inhibitors to result in covalent- inhibitors (Fig. 1C). The characterization of the inhibitory and
allosteric inhibitors (CAIs).27 The rst-in-class inhibitor bor- kinetic properties as well as a series of complex crystal struc-
ussertib (1) is based on the 1,6-naphthyridinone-scaffold and tures resulted in the rst structure-activity relationship (SAR) of
features a warhead to allow for the formation of a covalent bond this innovative class of inhibitors. Furthermore, we demon-
to Cys296 via Michael addition, resulting in an increased strate the potent inhibition of cell proliferation in a series of
potency and selectivity by maximization of the drug-target cellular models. By in vitro ADME proling, we identied novel
residence time.28 The evaluation of borussertib in meaningful predestined candidates for further in vivo studies.
cellular and xenogra models emphasized the inhibitory
potency and in vivo efficacy of this novel class of inhibitors.29
The crystal structure in complex with full-length Akt provided
Results
Structure-based design
crucial information about the binding characteristics (Fig. 1A–
C). Based on these insights, we now report the structure-based
design and synthesis of a focused library of covalent-allosteric
Structural information about the binding characteristics of
covalent-allosteric inhibitors provides not only a detailed
Chem. Sci.
This journal is © The Royal Society of Chemistry 2019

View Article Online
Edge Article
Chemical Science
understanding of Akt modulation with this novel type of importance of the water-mediated hydrogen bond was investi-
inhibitor, but also denes starting points for further optimiza- gated by modications at the N1-position of the benzo[d]imi-
tion. The recently published co-crystal structure of full-length dazolone moiety. Furthermore, we investigated how altering the
Akt with borussertib (PDB: 6HHF)29 showed that the exibility and linker length between the 1,6-naphthyridinone
compound was indeed bound between the PH- and the kinase core and the acrylamide warhead impact the preorganization of
domain to adopt the proposed allosteric binding mode (Fig. 1A the inhibitors towards the targeted cysteine residues. We
and B). Most importantly, the crystal structure clearly revealed addressed this issue by replacing the benzo[d]imidazolone by
that the acrylamide warhead of borussertib formed a covalent scaffolds of different length, exibility and substitution pattern
bond with Cys296 of the kinase domain. The electrophilic group (ESI Fig. S18†). In line with our previous study, the retro-
was further stabilized by an H-bond to the amide backbone of synthetic analysis revealed the convergent division in 1,6-
Glu85. The 1,6-naphthyridinone core scaffold resided at the naphthyridinone building block 9 and the benzo[d]imidazolone
interface between the activation loop of the kinase domain and 13 (Scheme 1).27,32 Synthesis of 9 was dissected into two
the regulatory PH-domain, forming a favorable p–p-stacking synthetic pathways starting from pyridine-derivative 2 and p-
interaction with the side chain of Trp80 of the PH-domain, as cyanobenzaldehyde (5). Compound 2 was Boc-protected and
previously seen for different allosteric inhibitors in complex formylated in order to obtain intermediate 4.33 In the second
with Akt.23,30,31 The phenyl rings in the 3- and the 4-position of synthetic path, the aldehyde of 5 was initially protected with
the naphthyridinone core interact with the kinase domain by ethylene glycol and then converted by means of a Grignard
¨
sandwiching the side chain of Tyr272 and are located in reaction to the acetophenone-based intermediate 7. Friedlander
hydrophobic regions anked by Leu210, Leu264, and Ile290, quinoline synthesis, coupled with an in situ Boc-deprotection,
thus adding to favorable protein–ligand interactions. Interest- allowed the subsequent combination of 4 and 7 to the 1,6-
ingly, the unsubstituted amide nitrogen of the imidazolone naphthyridinone scaffold 8, which could then be deprotected to
pointed into a hydrophilic pocket which was occupied by obtain the desired building block 9.34
a conserved water molecule which mediates H-bond interac-
The synthesis of the benzo[d]imidazolone scaffold started
tions to the amino acid side chains of Glu17 of the PH- and from a selective nitration of 10 to afford intermediate 11
Tyr326 and Arg273 of the kinase domain. Structure–activity (Scheme 2A).35 Pd-catalyzed hydration of the nitro-group into
relationship (SAR) studies regarding this rather narrow pocket the amine and reductive dehalogenation succeeded in one
will be discussed below. Also, the 4- and 5-positions of the step.36 The desired product 13 was obtained aer selective Boc-
benzo[d]imidazolone core were solvent exposed. Thus, the protection of the anilinic amine.37 Great efforts were made to
crystal structure of borussertib inspired the structure-based optimize the reductive amination, combining the two major
design of novel probe molecules by directed derivatization. building blocks 9 and 13. Ultimately, we found a Leuckart–
Toward this end, we developed a robust synthetic route that Wallach amination using formic acid as the reducing agent gave
enabled the synthetic access to a series of highly diverse 1,6- the best results.38 Deprotection and attachment of the acryl-
naphthyridinone-based compounds.
amide yielded the desired product (Scheme 3, 24a–c). Besides
dehalogenation and simple alkylation reactions at the N1-
position (synthesis shown in ESI†), the benzo[d]imidazolone
core gave a very restricted scope regarding chemical space of
possible derivatives. Consequently, we decided to change the
scaffold in order to enable access to a greater variety of potential
inhibitors. One option was the ring-opening of the imidazolone
moiety leading to phenylurea-based derivatives (Scheme 2B).
Aer initial Cbz-protection of 3-nitro-4-chloroaniline (14), urea
Structure-based design-driven synthesis
To acquire insights into the structure–activity relationship of
1,6-naphthyridinone-based covalent-allosteric inhibitors, we
developed a robust synthetic route to generate a focused library
of Akt inhibitors. Novel inhibitors were designed based on the
insights of the structural analysis of the crystal structure of
borussertib in complex with Akt (Fig. 1C). More precisely, the
Scheme 1 Synthesis of 1,6-naphthyridinone core scaffold by a divergent synthetic route. (i) Boc2O, DMAP, TEA, DCM, rt, 2 h. (ii) tBuLi, DMF, THF,
À76 ꢀC – rt, ovn. (iii) Ethylene glycol, p-TsOH, toluene, reflux, ovn. (iv) Benzylmagnesium chloride, THF, 0 ꢀC – rt. (v) Sodium methoxide, MeOH,
65 ꢀC, 4 h. (vi) HClaq, reflux, 3 h.
This journal is © The Royal Society of Chemistry 2019
Chem. Sci.

View Article Online
Chemical Science
Edge Article
Scheme 2 Synthesis of diverse building blocks for the derivatization of covalent-allosteric Akt inhibitors. (A) Synthetic scheme for synthesis of
benzo[d]imidazolone building block. (i) HNO3, o-xylene, 60 ꢀC, 2 h. (ii) Pd/C, NH4HCOO, MeOH, reflux, ovn. (iii) Boc2O, 10% AcOH/H2O, 1,4-
dioxane, rt, ovn. (B) Synthetic scheme for synthesis of urea-based building block. (i) Cbz-Cl, NaHCO3, THF, rt, 5 h. (ii) 18, K2CO3, DMF, 130 ꢀC, mW.
(iii) 4 N HCl/1,4-dioxane, rt, 2 h. (C) Synthetic scheme for synthesis of phenylamide-based building blocks. (i) HATU, 2,6-lutidine, DMF, rt, ovn. (ii)
4 N HCl/1,4-dioxane, rt, 2 h.
formation (16) occurred derived from literature procedure with with 4-amino-Boc-piperidine (18) and subsequent deprotection
4-amino-Boc-piperidine under microwave conditions (18).39 Acidic yielded novel building block that could be combined with the
hydrolysis of the Boc-protecting group led to another fragment core scaffold by Leuckart–Wallach amination (Schemes 2C
that was capable of attaching to the 1,6-naphthyridinone-scaffold and 3). By reduction and acrylamide attachment, the
by Leuckart–Wallach reaction.
phenylamide-based probe molecules (30a–g) were obtained. In
Hence, the desired compound was obtained by reduction of addition, the variation of the acrylamide warhead provided
25 and attachment of the Michael acceptor. Furthermore, another opportunity for modications. Instead of the undeco-
phenylamide-derivatives were considered to be an attractive rated acrylamide, the introduction of Michael acceptors
alternative scaffold for optimization. Based on a variety of carrying a dimethylamino solubilizing group were expected to
benzoic acids (19a–g), an amide coupling of these compounds mediate an improved solubility and additionally serve as an
Scheme 3 Synthetic scheme for the synthesis of covalent-allosteric inhibitor derivatives by assembling building blocks with Leuckart–Wallach-
reaction. (i) Amine, formic acid, MeCN, 80 ꢀC, ovn. (ii) 4 N HCl/1,4-dioxane, rt, 2 h. (iii) Acryloyl chloride, DIPEA, THF, rt, 3 h. (iv) Fe, NH4Cl, MeOH/
H2O, 80 ꢀC, 2 h.
Chem. Sci.
This journal is © The Royal Society of Chemistry 2019

View Article Online
Edge Article
Chemical Science
intramolecular base to catalyze the Michael addition.40 In order linker region of phenylamide-based inhibitors, compound 32,
to analyze the inuence of the piperidine-based linker which omitted the carbonyl function, showed a lack of activity.
region, piperazine-based molecule 31 and benzylamine-based We hypothesize that this is a conformation-driven effect. This
compound 32 were synthesized by related synthetic routes suggests that leaving out the amide function led to a loss of
(synthesis shown in ESI†). In total, 13 novel inhibitors with structural integrity of the inhibitor. Furthermore, as observed
different substitution patterns were synthesized. The for molecule 31, these phenyl amides were characterized by
biochemical and cellular evaluations of these molecules a superior potency in comparison to phenylpiperazine-based
should lead to insights into the SARs of covalent-allosteric Akt probes that were distinguished by a shortened molecule
inhibitors.
length. It is worth noting that the replacement of the phenyl
amide by the bioisosteric thiophene amide 30f was well toler-
ated.41 The kinetic evaluation of these inhibitors already evinced
the covalent binding character of these type of molecules. In
Biochemical characterization and SAR analysis
For activity-based characterization of the small focused 1,6- order to provide an orthogonal proof for covalent bond forma-
naphthyridinon-based library, we measured the inhibitory tion, we co-incubated the covalent-allosteric inhibitors with the
potency (IC50) of all synthesized compounds and important protein Akt1 and analyzed the samples by mass spectrometry
intermediates towards full-length Akt (Table 1). In addition, (Fig. 2). Although the resolution of the deconvoluted mass
a representative selection of these molecules was further spectra was low due to differently phosphorylated forms of Akt1,
investigated with respect to their kinetic proles. In large, the all spectra showed a mass shi in comparison to the apo
covalent allosteric inhibitors showed superior potency with protein with mass differences according to a mono-labeling of
half-maximal inhibitory activity in the nanomolar to sub- Akt1 with the corresponding inhibitor. While molecules with
nanomolar range when compared to the clinical candidate MK- a high inhibitory activity as shown for 1, 24b (Fig. 2B) and 27
2206. Furthermore, an evaluation of the ratio of kinact/Ki, (Fig. 2C) showed complete labeling of the protein, inhibitors
a parameter which combines both affinity and covalent bond with a lower potency such as 31 showed incomplete labeling
formation, indicated that these inhibitors had an outstanding within 1 h (Fig. 2E). In concert with the biochemical analysis,
kinetic prole. Small substitutions at the 5-position of the these experiments gave further qualitative evidence for selective
benzo[d]imidazolone-core as shown for the 5-chlorinated covalent modication of the target protein.
derivative 24a were tolerated, but showed 2-fold lower IC50
To further investigate inhibitor selectivity over other protein
values, respectively. Interestingly, the N-methylation of the kinases, we evaluated the benzo[d]imidazolone 24b, the phe-
benzo[d]imidazolone in 24b causes a 3-fold loss in inhibitory nylurea 27 and the phenylamide 30b in a proling study
activity, suggesting that also methylation at this position is well encompassing 100 selected kinases with similar characteristics
tolerated. However, such a mild drop in potency let us assume such as AGC kinase family members, PH-domain containing
that this water molecule is not energetically unfavorable. In kinases and kinases containing analogous cysteines on the
addition, the elongated electrophile of 24c showed a 20-fold activation loop (SelectScreen Kinase Proling Service from
decrease in potency indicating no positive inuence of the ThermoFisher Scientic, Tables S2–S4 and Fig. S19 in the ESI†).
dimethylamino-group on inhibitor binding. In order to extend At a concentration of 1 mM, all compounds show a potent
access to more diverse compounds occupying a broader chem- inhibition of all three Akt isoforms with a notable preference of
ical space, we varied the benzo[d]imidazolone scaffold by the Akt1 and Akt2 isoforms. For 24b and 30b, we observed
opening the imidazolone-ring to obtain phenylurea-derivatives. a selectivity prole similar to borussertib with poor inhibition of
Although a 10-fold loss in activity was observed compared to the MELK kinase (48% by 24b and 52% by 30b, respectively). For
1, an evaluation of compound 27 revealed that phenylurea- 30b, we detected additional moderate inhibitory activity against
based probes represent covalent-allosteric inhibitors with the AMP-activated protein kinases A1/B1/G1 (46%) and A2/B1/
a novel scaffold that inhibited Akt in the nanomolar range. On G1 (40%) and non-receptor tyrosine kinase BMX (44%). Inter-
the other hand, phenylamide-based inhibitors illustrated an estingly, we noticed subtle differences in the selectivity prole
alternative scaffold that, in contrast to benzo[d]imidazolones, of phenylurea-based compound 27, since we observed a more
allowed access to other diverse derivatives (30a–f, Table 1). In potent inhibition of MELK kinase (83%), AMPK A1/B1/G1 (77%)
general, our compounds exhibited half-maximal inhibition of and A2/B1/G1 (77%) and besides a moderate inhibition PASK
Akt at nanomolar concentrations, indicating that these phe- (61%) and ROCK1 (54%). Of particular note is that all these
nylamide derivatives were suitable for covalent allosteric inhi- kinases contain at least one analogous cysteine on the activa-
bition. The orientation of the Michael acceptor was varied by tion loop. All other kinases showed less than 40% inhibition
decoration of the acrylamide function at the ortho-, meta- and emphasizing the fact that covalent-allosteric inhibitors are
para-positions. However, the ortho-substituted compound 30a characterized by a promising selectivity prole. The structure-
and the meta-substituted compound 30b were favored over 30c based design and biochemical evaluation of the small-focused
carrying the acrylamide in the para-position, which is consistent library of covalent allosteric inhibitors revealed phenylurea-
with the idea that the orientation of the warhead towards the and phenylamide-derived molecules as Akt probes with novel
targeted cysteines plays an important role in irreversible scaffolds that opened chemical space for optimization. The
inhibitor binding and occurs preferentially in the ortho- and binding mode of these compound classes was further investi-
meta-positions. On closer consideration of the amide-bridged gated by X-ray crystallography.
This journal is © The Royal Society of Chemistry 2019
Chem. Sci.

View Article Online
Chemical Science
Edge Article
Table 1 Biochemical evaluation of covalent-allosteric Akt inhibitors
Aktwt
kinact [minÀ1
Cpd
R
X
C
IC50 [nM]
Ki [nM]
]
kinact/Ki [mMÀ1 sÀ1
]
1
0.8 Æ 0.3
2.2 Æ 0.3
0.111 Æ 0.020
0.110 Æ 0.023
0.853 Æ 0.038
24a
24b
C
C
1.2 Æ 0.3
3.0 Æ 0.3
4.1 Æ 0.7
0.447 Æ 0.074
0.190 Æ 0.025
10.7 Æ 0.5
0.121 Æ 0.016
0.050 Æ 0.009
24c
C
18.1 Æ 4.9
33.0 Æ 2.4
0.025 Æ 0.005
27
C
C
9.1 Æ 1.5
17.5 Æ 3.6
27.2 Æ 4.2
0.081 Æ 0.019
0.138 Æ 0.031
0.080 Æ 0.008
0.085 Æ 0.018
30a
10.8 Æ 2.5
30b
C
3.6 Æ 0.8
6.8 Æ 0.3
0.083 Æ 0.016
0.202 Æ 0.035
30c
30d
C
C
285 Æ 65
362 Æ 74
0.042 Æ 0.007
0.085 Æ 0.008
0.002 Æ 0.001
0.014 Æ 0.002
52.8 Æ 16.3
98.6 Æ 10.0
30e
C
3.7 Æ 0.7
7.6 Æ 1.1
0.078 Æ 0.022
0.168 Æ 0.025
30f
31
C
N
14.0 Æ 6.1
126 Æ 31
39.3 Æ 6.6
67.1 Æ 5.5
0.088 Æ 0.010
0.054 Æ 0.006
0.038 Æ 0.007
0.013 Æ 0.001
32
C
192 Æ 45
327 Æ 27
0.071 Æ 0.004
0.004 Æ 0.001
Capivasertib
Ipatasertib
MK-2206
0.9 Æ 0.1
3.5 Æ 0.6
10.0 Æ 2.1
—
—
—
—
—
—
—
—
—
Chem. Sci.
This journal is © The Royal Society of Chemistry 2019

View Article Online
Edge Article
Chemical Science
Fig. 2 Deconvoluted mass spectra of Akt1wt after incubation with DMSO (Apo, (A)) and a small focused library of covalent-allosteric Akt inhibitors
showed a mass shift (displayed as D values) for all tested inhibitors with mass differences according to a labeling of the protein. (B) Mass spectra of
benzo[d]imidazolone-based inhibitors. (C) Mass spectrum of urea-based inhibitor. (D) Mass spectra of phenylamide-based inhibitors. (E) Mass
spectra of inhibitors 31 and 32. Mass spectra were recorded using denaturing conditions.
and PH-domain. Furthermore, the naphthyridinone-core of
these inhibitors adopted similar conformations to the previ-
ously described binding mode of borussertib by p–p-stacking
interaction with Trp80. Surprisingly, the crystal structures of
24b and 27 showed that the electrophile warhead did not form
a covalent bond to Cys296 as it was initially shown in the crystal
structure of borussertib, but covalently labeled the nearby
Cys310. This alternative binding mode also caused a different
interaction prole of these ligands within the binding pocket
Structural analysis
For a deeper understanding of the binding mode of the novel
Akt inhibitors, we co-crystallized full-length Akt with benzo[d]
imidazolone-derivative 24b, phenylurea-derivative 27, the
phenylamide-derivative 30b and the phenylpiperazine-based
ligand 31 (Fig. 3A–D). A more detailed close-up of the protein–
ligand interactions is given in ESI Fig. S20.† All four complex
structures conrmed the expected covalent-allosteric mode of
action while binding to the interdomain region between kinase
This journal is © The Royal Society of Chemistry 2019
Chem. Sci.

View Article Online
Chemical Science
Edge Article
Fig. 3 Co-crystal structures of full-length Akt in complex with different covalent-allosteric inhibitors, 2Fo–Fc maps contoured at 0.8s. (A) Co-
crystal structure of Akt with 24b (PDB: 6HHJ) and (B) with 27 (PDB: 6HHG) revealed novel binding mode while labeling Cys310. (C) Co-crystal
structure of Akt with 30b (PDB: 6HHI) and (D) with 31 (PDB: 6HHH) exhibited similar binding mode to borussertib while labeling Cys296.
(ESI Fig. S20A and B†). For 24b, the additional N-methylation molecular dynamic (MD) simulations. The pKa values of
impeded the polar contact of this amine to the conserved cysteine residues strongly correlate with their reactivity because
water molecule in the narrow polar pocket as shown in the deprotonated cysteine residues are considerably more reac-
crystal structure of borussertib. Consequently, the piperidine tive.42,43 Based on the similarities of their enzyme environment
ring adopted an altered conformation inducing a shi of the Cys296 and Cys310 are expected to possess equal pKa values and
benzo[d]imidazolone-core and thus enabled the covalent bond reactivities.44 In such cases the reaction probability of the sulfur
formation to Cys310 and additional H-bond interactions to the of a cysteine moiety with the electrophilic center of the inhibitor
amide backbone of Lys297 and water-mediated to the side chain is mainly determined by their distance in the non-covalent
of Gln79. In case of 27, a small shi of ligand conformation enzyme inhibitor complex, i.e. before the reaction.45–48 Against
caused labeling of Cys310 as well as H-bond formation of the this background, we used theoretically prepared non-covalent
acrylamide amine to the protein backbone of Glu17. The shi enzyme–inhibitor complexes and conducted MD simulation in
seemed to be induced by the extended molecule structure of the which we evaluated the distances of the acrylamide electrophile
urea-based compound in comparison to the benzo[d]imidazo- to the sidechain sulfur of Cys296 and Cys310 over the simula-
lone- and phenylamide-based inhibitors. This assumption was tion time (ESI Fig. S22–S24†). For borussertib (ESI Fig. S19†),
˚
further conrmed by the co-crystal structure of Akt with 30b and our MD simulations predict distances of 3–4 A between the
31, given that both molecules with shortened molecule length electrophilic center of the inhibitor and the nucleophilic sulfur
revealed a covalent modication of Cys296 and a similar of the Cys296 residue which is sufficiently short that a reaction
binding mode compared to borussertib. The interaction of the can take place. Additionally, the simulations indicated only
latter inhibitors was additionally stabilized by formation of an small variations in the distances between both centers which is
H-bond to the amide nitrogen of Glu85 (ESI Fig. S19C and D†). also favorable for a reaction.45,46 The computed distances
Hence, the crystal structure analysis of these covalent-allosteric between the electrophilic center of the inhibitor and the
inhibitors showed different prioritization in labelling the adja- nucleophilic sulfur of Cys310 are larger for most of the simu-
cent residues Cys296 and Cys310. Interestingly, previously lation time, although distances which enable a reaction are
described MS/MS analysis of borussertib showed a covalent rarely adopted. Summarizing, our simulations predict reactions
modication of both cysteines.27 We did additional MS/MS to Cys296 and Cys310, but the reaction to Cys296 is clearly
studies for inhibitor 27 and obtained similar results (ESI preferred. This nicely explains the apparent contradiction
Fig. S21†). Therefore, we further investigated the labeling of the between crystallographic and MS/MS experiments. For
residues of Cys296 and Cys310 by different inhibitors by compound 24b (ESI Fig. S20†) the computed distances between
Chem. Sci.
This journal is © The Royal Society of Chemistry 2019

View Article Online
Edge Article
Chemical Science
the electrophilic center and Cys310 become smaller and the lines exhibit genetic alterations in the PI3K/Akt and Ras/MAPK
corresponding uctuations decrease in comparison to the pathways, including in-frame deletions in PIK3R1, loss-of-
situation of borussertib. In contrast, the distances and the function or downregulating genetic lesions in PTEN or TP53 or
uctuation of the distances to Cys296 increase. This implies activating point mutations in PI3KCA, Akt1 and NRas. Evidently,
that for 24b the reaction with Cys310 becomes more probable the tested cell lines AN3CA, T47D, ZR-75-1, MCF-7, BT-474 and
hence is in line with the experimental crystal data which KU-19-19 are dependent on Akt signaling considering that both
indicate a covalent bonding to Cys310. In case of 27, an even reference compounds and the majority of the CAI-library were
higher preference to Cys310 was ascertained which was re- able to inhibit cell proliferation. As gured out in recent studies,
ected by the fact that during simulation time the distance the breast cancer cell line ZR-75-1 showed the highest sensitivity
between the electrophile of 27 and Cys296 never adopted against Akt inhibition and, more particularly, the rst-in-class
distances that enable covalent bond formation. Although MS/ covalent-allosteric inhibitor borussertib exhibited excellent
MS analysis showed a Cys296-labeled species which was not cellular activity in the nanomolar range with superior prole
detectable during the MD simulations, these results are in against clinical candidate Akt inhibitors as well as the cytostatic
good correlation with the crystallography studies and indicate drug doxorubicin. Interestingly, the chlorinated derivative of
that the covalent bond formation of 27 to Cys310 is clearly borussertib 24a and the methylated derivative 24b evoked a 2-fold
favored.
higher inhibitory activity with low nanomolar EC50-values for ZR-
75-1 inhibition and similar tendencies for the endometrium
cancer cell line AN3CA and the breast cancer cell lines T47D,
MCF-7 and BT-474. The other benzo[d]imidazolone derivative 24c
showed lower cellular activity with a 200- and 100-fold reduction
of EC50, respectively. For the novel urea-based inhibitor 27,
a good cellular inhibition could also be determined, in particular
with 39 nM inhibition of ZR-75-1, however, a 20-fold decrease in
activity was observed in comparison to 24a. At a glance, the series
of phenylamide-based inhibitors showed an outstanding inhibi-
tory prole on all different cell lines, whereas 30b was identied
to be the most active covalent-allosteric inhibitor in this series.
Cellular evaluation
The biochemical and structural analysis of the small synthesized
library revealed probe molecules acting as covalent-allosteric
inhibitors of Akt. In order to proof their activity in cellular
systems, we tested the inhibitors and we determined the half-
maximal effective concentration (EC50)-values (Table 2). The
tested cell lines AN3CA (endometrium), T47D (breast), ZR-75-1
(breast), MCF-7 (breast), BT-474 (breast) and KU-19-19 (bladder)
are described to be sensitive against Akt inhibition.23,49 All cell
Table 2 Cellular evaluation of covalent-allosteric Akt inhibitors
Cell viability assay EC50 [nM]
AN3CA
R557_K561>Q
R130 fs
T47D
ZR-75-1
MCF-7
BT-474
K111N
KU-19-19
R1023Q
PIK3R1a
PIK3CAa
PTENa
AKT1a
H1047R
E545K
L108R
E17K/E49K
Q61R
NRASa
TP53a
L194F
E285K
Capivasertib
Ipatasertib
MK-2206
869 Æ 278
925 Æ 457
972 Æ 322
30 Æ 9
475 Æ 92
443 Æ 6
191 Æ 68
219 Æ 83
63 Æ 21
555
1605 Æ 450
2371 Æ 745
1682 Æ 316
623 Æ 134
20 866 Æ 7603
24 037 Æ 9491
7054 Æ 421
72 Æ 8
5036 Æ 1830
571 Æ 111
153 Æ 9
583 Æ 245
127 Æ 40
Doxorubicin
100 Æ 26
Borussertib
24a
24b
24c
191 Æ 90
159 Æ 70
48 Æ 15
25 Æ 7
5 Æ 1
2 Æ 0
277 Æ 90
138 Æ 81
373 Æ 54
259 Æ 68
7770 Æ 641
2313 Æ 908
7157 Æ 2112
> 30 000
230 Æ 75
11 Æ 2
2 Æ 0
32 Æ 8
109 Æ 16
15 333 Æ 3264
2063 Æ 438
1328 Æ 60
382 Æ 31
2114 Æ 49
336 Æ 25
41 Æ 19
427 Æ 23
39 Æ 3
20 267 Æ 7039
1111 Æ 117
1024 Æ 621
336 Æ 113
2841 Æ 685
2437 Æ 1017
224 Æ 88
11 971 Æ 1636
1785 Æ 458
2268 Æ 700
464 Æ 31
27
1660 Æ 46
30a
30b
30c
30d
30e
30f
31
32
22 Æ 7
5543 Æ 1506
14 942 Æ 5541
11 332 Æ 4428
6067 Æ 1104
6442 Æ 1238
3055 Æ 848
17 119 Æ 6379
6755 Æ 1462
95 Æ 20
11 Æ 3
3467 Æ 1693
1311 Æ 158
568 Æ 106
879 Æ 223
3614 Æ 1292
3288 Æ 298
1359 Æ 454
546 Æ 87
179 Æ 30
183 Æ 52
832 Æ 81
1897 Æ 578
151 Æ 47
73 Æ 25
10 Æ 3
5211 Æ 1892
1956 Æ 391
621 Æ 85
33 Æ 9
1190 Æ 518
1218 Æ 58
2196 Æ 125
1104 Æ 212
4617 Æ 1059
4019 Æ 912
116 Æ 16
384 Æ 62
a
Genetic alterations of the evaluated cancer cell lines according to COSMIC database.
This journal is © The Royal Society of Chemistry 2019
Chem. Sci.

View Article Online
Chemical Science
Edge Article
Furthermore, a good correlation between effect on cell viability in analysis of the metabolic hot spots of this inhibitor is required
different cancer cells that were previously shown to be sensitive to enable the chemical optimization of its pharmacokinetic
towards Akt inhibition and the enzymatic inhibition of Akt was properties. The phenylamide-based probe molecules also
detected. For instance, the meta-derivative 30b exhibited superior showed good microsomal stability, whereas 30b stuck out with
inhibitory activity compared to its ortho-derivative 30a and its a 4-fold longer half-life in murine microsomes. Consequently,
para-derivative 30c.
with 30b another covalent-allosteric inhibitor of protein kinase
In addition, small substitutions revealed by 30e or bio- Akt could be identied with optimized stability for in vivo
isosteric scaffolds as shown for 30f are well tolerated, and mouse studies without a decrease of activity as validated in
correspond well to their biochemical prole. In summary, out of biochemical and cellular evaluation.
a small focused library, a set of inhibitors was identied having
an excellent biochemical and cellular activity. In order to
ascertain the best candidates for in vivo characterization and
Discussion
lead development, we further investigated a selection of prom-
By combining the benecial properties of allosteric and irre-
ising inhibitors concerning pharmacokinetic properties.
versible inhibitors, we developed covalent-allosteric Akt inhib-
itors which represent a novel class of probe molecules with
outstanding selectivity and potency proles. The structural
information emanating from the recently published rst crystal
structure of a covalent-allosteric Akt inhibitor borussertib
inspired a series of structure-based designed inhibitors that
acquire deeper insights into the structure–activity relationships
(Fig. 4). The established synthetic route enabled us to access
diverse derivatives that could be analyzed concerning their
biochemical and cellular characteristics. Through biological
and pharmacokinetic evaluations, we were able to evaluate
these compounds regarding their inhibitory efficacy and ability
to act as eligible candidates for further preclinical investiga-
tions. The structural evaluation revealed that p–p-interactions
of the 1,6-naphthyridinone scaffold with Trp80 were of great
importance for binding affinity, and thus the core scaffold was
maintained in the design of the novel compounds. On the other
Evaluation of phase I metabolic stability
The cellular evaluation demonstrated high inhibitory activity of
a series of covalent-allosteric Akt inhibitors on different cancer
cell lines, thus making these molecules potential candidates for
further preclinical investigation. To understand subtle differ-
ences between the best inhibitors and to assess their ability to
examine them in subsequent in vivo studies, we further
analyzed a representative selection of inhibitors with respect to
their pharmacokinetic properties (Table 3). For this purpose, we
used a microsomal stability assay that is a good predictor for
phase I metabolic stability.50 As discussed in previous studies,
borussertib revealed a good pharmacokinetic prole with
reasonable microsomal stability in human and murine micro-
somes.29 Inhibitor 24a, which exhibited an equivalent activity
compared to borussertib in cellular evaluations, showed similar
stability properties. Combining the superior cellular activity
and comparable stability, 24a is a promising candidate with
optimized potency for further in vivo investigation. In contrast,
4-fold lower stability in microsomes occurred for 24b, thus the
methyl group seems to be a metabolically labile position of the
inhibitor.
A similar loss in stability was observed for
phenylurea-derived inhibitor 27. Therefore, a more detailed
Table 3 Evaluation of covalent-allosteric Akt inhibitors in a micro-
somal stability assay
Microsomal stability phase I
t1/2 [min]
Clint [mL minÀ1 mgÀ1
]
(Ha/Mb)
(Ha/Mb)
Capivasertib
Ipatasertib
MK-2206
Borussertib
24a
24b
27
30b
30e
33/87
39/231
347/693
99/46
87/41
25/23
19/20
58/173
39/69
14/16
12/6
1/2
5/30
5/34
19/60
24/68
8/8
Fig. 4 Overview illustrates the synthesized and evaluated library of
covalent-allosteric Akt inhibitors. Phenylamide-, benzo[d]imidazo-
lone- and urea-based derivates gain insights into the structure–
activity-relationships. The diversification of scaffolds and resulted in
novel binding modes with different preference in labeling Cys296 and
Cys310. Phenylamide-based inhibitors broadened chemical space and
showed improved metabolic stability for 30b. Small substitutions of
the benzo[d]imidazolone-core allowed a gain of cellular potency.
Urea-based inhibitor 27 revealed a novel binding mode by promoting
Cys310-modification.
12/20
12/24
30f
39/58
a
Human liver microsomes. b Murine liver microsomes.
Chem. Sci.
This journal is © The Royal Society of Chemistry 2019

View Article Online
Edge Article
Chemical Science
side of the molecule, the benzo[d]imidazolone moiety allows an additionally an improved pharmacokinetic prole in compar-
optimal orientation of the acrylamide warhead towards the ison to borussertib due to better microsomal stability.
covalently targeted cysteine residues. Consistent with the
complex structure, small substituents in the 5-position of this
scaffold were well tolerated, and, subsequently, the chlorine
Conclusions
substitution in molecule 24a had a positive inuence on the
cellular efficacy with a reasonable pharmacokinetic prole.
Furthermore, the unoccupied polar narrow pocket in the
direction of the free amide of the imidazolone ring was of high
interest in the design of novel probe molecules. N-Methylation
at this position led to 24b with a methyl group pointing toward
this area with increased cellular potency in comparison to
borussertib. Although the metabolic stability assay indicated
a metabolic hot spot at this position, further investigations
concerning a presumable cleavage of the methyl group in phase
I metabolism should specify a potential role of 24b as a prodrug
concept for borussertib. The complex crystal structure of 24b in
complex with Akt showed a novel alternative binding mode with
labeling of the adjacent Cys310 instead of Cys296, while the
absence of polar interactions towards this conserved water
molecule induced an altered conformation of the piperidine
ring. The inuence of more polar substituents at this position
on the distinct binding mode should contribute to deeper
understanding and further optimization. However, the ability to
label both cysteine residues could have a benecial effect
regarding resistance mutations of these amino acids that are
known to occur for other cysteine-modifying covalent kinase
inhibitors in cancer treatment.51 In order to allow access to
more substituents occupying higher chemical space, the benzo
[d]imidazolone core was substituted by ring opening to afford
phenylurea-based compounds. The closely related structure of
both scaffolds suggested that this replacement was well toler-
ated. The slight increase in half-maximal inhibitory concen-
tration for 27 could be attributed to the higher entropic degrees
of freedom induced by the ring opening, leading to a higher
entropic penalty upon binding to the kinase. By solving the
complex crystal structure of 27, we also revealed the altered
binding mode by labeling Cys310, which is promoted by the
elongated structure of 27. Given the lower microsomal stability,
the urea-based inhibitor still has a need for further optimiza-
tion. However, they represent an interesting novel class of probe
molecules. In addition, phenylamide-based derivatives consti-
tuted another novel type of covalent allosteric Akt inhibitors. A
series of derivatives veried the possibility to introduce
different small substituents as shown for the halide-derivatives
30d and 30e, and bioisosteric variants as shown for 30f are well
tolerated. The biochemical evaluation of these compounds
showed a good correlation to the inhibitory effect on different
cancer cell lines. In particular, this aspect became apparent in
comparing the ortho-, meta- and para-derivatives 30a–c. These
probe molecules additionally conrmed the importance of an
optimal preorganization of the acrylamide warhead toward the
targeted cysteine for covalent bond formation, also manifested
by the co-crystal structure of Akt in complex with 30b. The
shortened molecule length promoted labeling of Cys296, also
illustrated by the complex crystal structure of 31. Probe mole-
cule 30b stood out having high nanomolar cellular activity and
In summary, we have validated the concept of covalent-
allosteric inhibitors as highly potent modulators of the
protein kinase Akt by a structure-guided design of a small
focused library of novel probe molecules. Phenylurea- and
phenylamide-based inhibitors opened a broader chemical
space for further derivatization and optimization and led to
a synthesis route that allows access to a variety of novel deriv-
atives. Biochemical characterization, kinetic parameters, and
cellular evaluation provided a detailed view of the structure–
activity relationships of covalent-allosteric inhibitors. By further
analyzing the microsomal stability, the inhibitors 24a and 30b
were identied as potent and stable novel candidates for in vivo
studies and lead development.
Statement of contributions
N. U. and S. S. contributed equally and designed and carried out
the synthesis. J. W. performed the activity-based assay and
kinetic experiments. I. L., M. P. M. and R. S. solved the crystal
structures. M. L., R. G., L. D. and L. Q. contributed to synthesis.
J. W. and I. L. conducted the cell viability assay. T. A. L. and B. E.
performed the molecular dynamic simulations. J. H. deter-
mined the microsomal stability. P. C. carried out the molecular
modeling studies. P. J. performed MS/MS studies. N. U., S. S.
and M. L. wrote the manuscript and all authors have given
approval to the nal version. D. R. conceived and designed all
experiments.
Conflicts of interest
There are no conicts to declare.
Acknowledgements
This work was co-funded by the German Federal Ministry
for Education and Research (NGFNPlus and e:Med) (Grant
No. BMBF 01GS08104, 01ZX1303C), the Deutsche For-
schungsgemeinscha (DFG), the German federal state North
Rhine-Westphalia (NRW) and the European Union (European
Regional Development Fund: Investing In Your Future) (EFRE-
800400), NEGECA (PerMed NRW), EMODI and Drug Discovery
Hub Dortmund (DDHD).
Notes and references
1 M. Rask-Andersen, J. Zhang, D. Fabbro and H. B. Schioth,
Trends Pharmacol. Sci., 2014, 35, 604–620.
2 J. Lategahn, M. Keul and D. Rauh, Angew. Chem., 2018, 57,
2307–2313.
3 P. Wu, T. E. Nielsen and M. H. Clausen, Drug Discovery Today,
2016, 21, 5–10.
This journal is © The Royal Society of Chemistry 2019
Chem. Sci.

View Article Online
Chemical Science
Edge Article
4 I. Hers, E. E. Vincent and J. M. Tavare, Cell. Signalling, 2011,
23, 1515–1527.
5 P. S. Mundi, J. Sachdev, C. McCourt and K. Kalinsky, Br. J.
Clin. Pharmacol., 2016, 82, 943–956.
Y. Y. Feng, K. Krysiak, J. Hoog, Z. Guo, L. Nehring,
K. B. Wisinski, E. Mardis, I. S. Hagemann, K. Vij, S. Sanati,
H. Al-Kateb, O. L. Griffith, M. Griffith, L. Doyle,
C. Erlichman and M. J. Ellis, Clin. Cancer Res., 2017, 23,
6823–6832.
6 M. Rudolph, T. Anzeneder, A. Schulz, G. Beckmann,
¨
A. T. Byrne, M. Jeffers, C. Pena, O. Politz, K. Kochert, 22 M. J. Lindhurst, M. R. Yourick, Y. Yu, R. E. Savage, D. Ferrari
R. Vonk and J. Reischl, BMC Cancer, 2016, 16, 1–12. and L. G. Biesecker, Sci. Rep., 2015, 5, 17162.
7 A. S. Clark, K. West, S. Streicher and P. A. Dennis, Mol. Cancer 23 J. M. Lapierre, S. Eathiraj, D. Vensel, Y. Liu, C. O. Bull,
Ther., 2002, 1, 707–717.
S. Cornell-Kennon, S. Iimura, E. W. Kelleher, D. E. Kizer,
S. Koerner, S. Makhija, A. Matsuda, M. Moussa,
N. Namdev, R. E. Savage, J. Szwaya, E. Volckova,
N. Westlund, H. Wu and B. Schwartz, J. Med. Chem., 2016,
59, 6455–6469.
8 G. M. Nitulescu, D. Margina, P. Juzenas, Q. Peng, O. T. Olaru,
E. Saloustros, C. Fenga, D. Spandidos, M. Libra and
A. M. Tsatsakis, Internet J. Oncol., 2016, 48, 869–885.
9 J. Zhang, P. L. Yang and N. S. Gray, Nat. Rev. Cancer, 2009, 9,
28–39.
10 J. Lin, D. Sampath, M. A. Nannini, B. B. Lee, M. Degtyarev,
J. Oeh, H. Savage, Z. Guan, R. Hong, R. Kassees, L. B. Lee,
T. Risom, S. Gross, B. M. Liederer, H. Koeppen,
24 G. S. Roth, Z. Macek Jilkova, A. Zeybek Kuyucu, K. Kurma,
S. T. Ahmad Pour, G. Abbadessa, Y. Yu, B. Busser,
P. N. Marche, V. Leroy and T. Decaens, Mol. Cancer Ther.,
2017, 16, 2157–2165.
N. J. Skelton, J. J. Wallin, M. Belvin, E. Punnoose, 25 I. Vivanco, Z. C. Chen, B. Tanos, B. Oldrini, W. Y. Hsieh,
L. S. Friedman and K. Lin, Clin. Cancer Res., 2013, 19,
1760–1772.
N. Yannuzzi, C. Campos and I. K. Mellinghoff, eLife, 2014,
3, e03751.
11 R. L. B. Costa, H. S. Han and W. J. Gradishar, Breast Cancer 26 J. Rauch, N. Volinsky, D. Romano and W. Kolch, Cell
Res. Treat., 2018, 169, 397–406. Commun. Signaling, 2011, 9, 23.
12 M. Dumble, M.-C. Crouthamel, S.-Y. Zhang, M. Schaber, 27 J. Weisner, R. Gontla, L. van der Westhuizen, S. Oeck,
¨
D. Levy, K. Robell, Q. Liu, D. J. Figueroa, E. A. Minthorn,
M. A. Seefeld, M. B. Rouse, S. K. Rabindran,
D. A. Heerding and R. Kumar, PLoS One, 2014, 9, e100880.
J. Ketzer, P. Janning, A. Richters, T. Muhlenberg, Z. Fang,
A. Taher, V. Jendrossek, S. C. Pelly, S. Bauer, W. A. van
Otterlo and D. Rauh, Angew. Chem., 2015, 54, 10313–10316.
13 A. Spencer, S.-S. Yoon, S. J. Harrison, S. R. Morris, 28 J. Singh, R. C. Petter, T. A. Baillie and A. Whitty, Nat. Rev.
D. A. Smith, R. A. Brigandi, J. Gauvin, R. Kumar,
J. B. Opalinska and C. Chen, Blood, 2014, 124, 2190–2195.
14 A. W. Tolcher, A. Patnaik, K. P. Papadopoulos, D. W. Rasco,
C. R. Becerra, A. J. Allred, K. Orford, G. Aktan, G. Ferron-
Brady, N. Ibrahim, J. Gauvin, M. Motwani and
M. Cornfeld, Cancer Chemother. Pharmacol., 2015, 75, 183–
189.
Drug Discovery, 2011, 10, 307–317.
29 J. Weisner, I. Landel, C. Reintjes, N. Uhlenbrock,
M.
Trajkovic-Arsic,
N.
Dienstbier,
S.
Ladigan,
M. Lindemann, S. Smith, L. Quambusch, R. Scheinpug,
L. Depta, R. Gontla, A. Unger, H. Muller, M. Baumann,
C. Schultz-Fademrecht, G. Guenther, A. Maghnouj,
¨
¨
M. P. Muller, M. Pohl, C. Teschendorf, H. Wolters,
15 L. R. Pearce, D. Komander and D. R. Alessi, Nat. Rev. Mol.
Cell Biol., 2010, 11, 9–22.
R. Viebahn, A. Tannapfel, W. Uhl, J. Hengstler, S. Hahn,
J. Siveke and D. Rauh, 2018, revised.
¨
16 Z. Fang, C. Grutter and D. Rauh, ACS Chem. Biol., 2013, 8, 58– 30 W.-I. Wu, W. C. Voegtli, H. L. Sturgis, F. P. Dizon,
70.
G. P. A. Vigers and B. J. Brandhuber, PLoS One, 2010, 5,
e12913.
¨
17 J. R. Simard, M. Getlik, C. Grutter, R. Schneider, S. Wulfert
and D. Rauh, J. Am. Chem. Soc., 2010, 132, 4152–4160.
18 S. F. Barnett, D. Defeo-Jones, S. Fu, P. J. Hancock,
K. M. Haskell, R. E. Jones, J. A. Kahana, A. M. Kral,
K. Leander, L. L. Lee, J. Malinowski, E. M. McAvoy,
D. D. Nahas, R. G. Robinson and H. E. Huber, Biochem. J.,
2005, 385, 399–408.
31 M. A. Ashwell, J.-M. Lapierre, C. Brassard, K. Bresciano,
C. Bull, S. Cornell-Kennon, S. Eathiraj, D. S. France,
T. Hall, J. Hill, E. Kelleher, S. Khanapurkar, D. Kizer,
S. Koerner, J. Link, Y. Liu, S. Makhija, M. Moussa,
N. Namdev, K. Nguyen, R. Nicewonger, R. Palma,
J. Szwaya, M. Tandon, U. Uppalapati, D. Vensel,
L. P. Volak, E. Volckova, N. Westlund, H. Wu, R.-Y. Yang
and T. C. K. Chan, J. Med. Chem., 2012, 55, 5291–5310.
19 P. N. Lara Jr, J. Longmate, P. C. Mack, K. Kelly,
M. A. Socinski, R. Salgia, B. Gitlitz, T. Li, M. Koczywas,
K. L. Reckamp and D. R. Gandara, Clin. Cancer Res., 2015, 32 Z. Fang, J. R. Simard, D. Plenker, H. D. Nguyen, T. Phan,
21, 4321–4326.
P. Wolle, S. Baumeister and D. Rauh, ACS Chem. Biol.,
2015, 10, 279–288.
20 R. K. Ramanathan, S. L. McDonough, H. F. Kennecke,
S. Iqbal, J. C. Baranda, T. E. Seery, H. J. Lim, A. F. Hezel, 33 M. T. Bilodeau, A. E. Balitza, J. M. Hoffman, P. J. Manley,
G. M. Vaccaro and C. D. Blanke, Cancer, 2015, 121, 2193–
2197.
21 C. X. Ma, V. Suman, M. P. Goetz, D. Northfelt, M. E. Burkard,
F. Ademuyiwa, M. Naughton, J. Margenthaler, R. A, R. Gray,
S. F. Barnett, D. Defeo-Jones, K. Haskell, R. E. Jones,
K. Leander, R. G. Robinson, A. M. Smith, H. E. Huber and
G. D. Hartman, Bioorg. Med. Chem. Lett., 2008, 18, 3178–
3182.
A. Tevaarwerk, L. Wilke, T. Haddad, T. Moynihan, 34 C. Chen, B. W. Eastman and E. H. Hu, Inhibitors of Akt
C. Loprinzi, T. Hieken, E. K. Barnell, Z. L. Skidmore, activity, WO/2006/110638, 2006.
Chem. Sci.
This journal is © The Royal Society of Chemistry 2019

View Article Online
Edge Article
Chemical Science
35 A. T. James and E. E. Turner, J. Chem. Soc., 1950, 1515–1519, 45 M. Mladenovic, K. Junold, R. F. Fink, W. Thiel,
DOI: 10.1039/JR9500001515.
T. Schirmeister and B. Engels, J. Phys. Chem. B, 2008, 112,
5458–5469.
46 M. Mladenovic, T. Schirmeister, S. Thiel, W. Thiel and
B. Engels, ChemMedChem, 2007, 2, 120–128.
47 T. Schirmeister, J. Kesselring, S. Jung, T. H. Schneider,
A. Weickert, J. Becker, W. Lee, D. Bamberger, P. R. Wich
and U. Distler, J. Am. Chem. Soc., 2016, 138, 8332–8335.
36 P. Selvam, S. U. Sonavane, S. K. Mohapatra and
R. V. Jayaram, Tetrahedron Lett., 2004, 45, 3071–3075.
37 V. Perron, S. Abbott, N. Moreau, D. Lee, C. Penney and
B. Zacharie, Synthesis, 2009, 283–289.
38 M. O. Frederick and D. P. Kjell, Tetrahedron Lett., 2015, 56,
949–951.
39 H.-K. Kim and A. Lee, Org. Biomol. Chem., 2016, 14, 7345– 48 T. C. Schmidt, A. Welker, M. Rieger, P. K. Sahu,
7353.
C. A. Sotriffer, T. Schirmeister and B. Engels,
ChemPhysChem, 2014, 15, 3226–3235.
40 A. Wissner and T. S. Mansour, Arch. Pharm., 2008, 341, 465–
477.
49 O. Politz, F. Siegel, L. Barfacker, U. Bomer, A. Hagebarth,
W. J. Scott, M. Michels, S. Ince, R. Neuhaus, K. Meyer,
A. E. Fernandez-Montalvan, N. Liu, F. von Nussbaum,
D. Mumberg and K. Ziegelbauer, Int. J. Cancer, 2017, 140,
449–459.
41 L. M. Lima and E. J. Barreiro, Curr. Med. Chem., 2005, 12, 23–
49.
42 A. Paasche, M. Schiller, T. Schirmeister and B. Engels,
ChemMedChem, 2010, 5, 869–880.
43 M. Mladenovic, R. F. Fink, W. Thiel, T. Schirmeister and 50 T. H. Lave, S. Dupin, C. Schmitt, R. C. Chou, D. Jaeck and
B. Engels, J. Am. Chem. Soc., 2008, 130, 8696–8705. P. Coassolo, J. Pharm. Sci., 1997, 86, 584–590.
44 R. L. Lundblad, Chemical reagents for protein modication, 51 K. S. Thress, C. P. Paweletz, E. Felip, B. C. Cho, D. Stetson,
CRC press, 2014.
B. Dougherty, Z. Lai, A. Markovets, A. Vivancos, Y. Kuang,
D. Ercan, S. E. Matthews, M. Cantarini, J. C. Barrett,
P. A. Janne and G. R. Oxnard, Nat. Med., 2015, 21, 560–562.
This journal is © The Royal Society of Chemistry 2019
Chem. Sci.