Converting a weaker ATP-binding site inhibitor into a potent hetero-bivalent ligand by tethering to a unique peptide sequence derived from the same kinase

Article abstract of DOI:10.1039/c8ob01406j

Attaching an additional binding site directed moiety or a ligand to an ATP-binding site inhibitor has been used as a strategy to increase kinase binding affinity and specificity. The moieties typically used here as the second binding partner are varied from simple organic groups to ligands such as peptides derived from substrate binding site sequences. So far these hetero-bivalent ligands were developed targeting additional binding sites closer to the ATP-binding pocket. Here we report a unique expansion of this hetero-bivalent idea by: (I) targeting a new binding site much farther away from ATP-binding site, (II) using a peptide uniquely derived from a portion of the same kinase sequence that has been reported to turn and bind to the above distance binding pocket (used as the second binding ligand), and (III) optimizing a much longer and flexible linker (to connect ATP-binding site inhibitor and above mentioned second peptide sequence) with multistep, yet complete on-bead synthesis approach. We converted a very weak EphA3-kinase ATP-binding site inhibitor-PP2 into a potent hetero-bivalent ligand by tethering to a unique 5-mer peptide sequence that derived from the linker region of EphA3 that connects kinase and sterile alpha motif (SAM) domains. Our design highlight the use of distance binding pockets to ATP-binding site as the second targeted site, while introducing the idea of extracting natural peptide sequences that already exist within the same kinase sequence, by a careful screening of available crystal structures.

Full text of DOI:10.1039/c8ob01406j

View Article Online
View Journal
Organic &
Biomolecular
Chemistry
Accepted Manuscript
This article can be cited before page numbers have been issued, to do this please use: S. R. Kedika and
D. G. Udugamasooriya, Org. Biomol. Chem., 2018, DOI: 10.1039/C8OB01406J.
This is an Accepted Manuscript, which has been through the
Volume 14 Number
1 7 January 2016 Pages 1–372
Royal Society of Chemistry peer review process and has been
accepted for publication.
Organic &
Biomolecular
Chemistry
www.rsc.org/obc
Accepted Manuscripts are published online shortly after
acceptance, before technical editing, formatting and proof reading.
Using this free service, authors can make their results available
to the community, in citable form, before we publish the edited
article. We will replace this Accepted Manuscript with the edited
and formatted Advance Article as soon as it is available.
You can find more information about Accepted Manuscripts in the
author guidelines.
Please note that technical editing may introduce minor changes
to the text and/or graphics, which may alter content. The journal’s
standard Terms & Conditions and the ethical guidelines, outlined
in our author and reviewer resource centre, still apply. In no
event shall the Royal Society of Chemistry be held responsible
for any errors or omissions in this Accepted Manuscript or any
consequences arising from the use of any information it contains.
ISSN 1477-0520
COMMUNICATION
Takeharu Haino et al.
Solvent-induced emission of organogels based on tris(phenylisoxazolyl)
benzene
rsc.li/obc

Page 1 of 20
Organic & Biomolecular Chemistry
View Article Online
DOI: 10.1039/C8OB01406J
ABSTRACT
Attaching an additional binding site directed moiety or a ligand to an ATP-binding site
inhibitor has been used as a strategy to increase kinase binding affinity and specificity. The
moieties typically used here as the second binding partner are varied from simple organic groups
to ligands such as peptides derived from substrate binding site sequences. So far these hetero-
bivalent ligands were developed targeting additional binding sites closer to the ATP-binding
pocket. Here we report a unique expansion of this hetero-bivalent idea by: (I) targeting a new
binding site much farther away from ATP-binding site, (II) using a peptide uniquely derived from
a portion of the same kinase sequence that has been reported to turn and bind to the above
distance binding pocket (used as the second binding ligand), and (III) optimizing a much longer
and flexible linker (to connect ATP-binding site inhibitor and above mentioned second peptide
sequence) with multistep, yet complete on-bead synthesis approach. We converted a very weak
EphA3-kinase ATP-binding site inhibitor-PP2 into a potent hetero-bivalent ligand by tethering to
a unique 5-mer peptide sequence that derived from the linker region of EphA3 that connects
kinase and sterile alpha motif (SAM) domains. Our design highlight the use of distance binding
pockets to ATP-binding site as the second targeted site, while introducing the idea of extracting
natural peptide sequences that already exist within the same kinase sequence, by a careful
screening of available crystal structures.

Organic & Biomolecular Chemistry
Page 2 of 20
View Article Online
DOI: 10.1039/C8OB01406J
INTRODUCTION
Protein kinases are tightly regulated to maintain normal physiology and play a critical role
in cell signaling. Dysregulation of the activity of kinases causes a number of diseases, making
kinases a main therapeutic target. However, the high level of sequence and structural homology
shared by the more than 500 human protein kinases, in particular at the main adenosine
triphosphate (ATP) binding pocket, make it extremely difficult to develop high-binding and
selective inhibitors. Development of hetero-bivalent kinase inhibitors, wherein an ATP-binding
moiety is connected to another ligand that targets a location outside of the ATP-binding cleft, has
been explored for some time as an attractive strategy to improve both binding and specificity.1-4
Hetero-bivalent inhibitors of protein kinases can be separated into two main groups
depending on the two main binding pockets that they contact on the targeted kinase surface .2
The first group comprises the bisubstrate kinase inhibitors, which are composed of an ATP-
binding site compound that is covalently attached through a linker to a pseudosubstrate peptide
sequence derived from the substrate protein that interacts with a particular kinase. Ricouart et al.
first reported the bisubstrate inhibitor of the protein kinase C (PKC) family, which consists of more
than ten physiologically relevant isozymes.5 They designed the inhibitor by connecting a PKC
alpha-specific pseudosubstrate (SRRRRRR) with a 5-isoquinolinyl-sulfonyl moiety as the ATP
binder using a 16-Å linker based on the predicted distance from crystal structure. The resultant
bisubstrate inhibitor exhibited a 67-fold increase in potency (IC50=0.3 µM) when compared with
the 5-isoquinolinyl-sulfonyl moiety alone (IC50=20 µM). Over time, there have been reports of a
significant number of bisubstrate hetero-bivalents ligands.6-9
The second main type of hetero-bivalent inhibitors of protein kinases is composed of those
that have an ATP-binding site-directed compound attached to a second ligand that targets a
region outside of the active site, such as dimerization sites or other domain-interacting surfaces.
For example, Stebbins et al. reported a hetero-bivalent inhibitor for mitogen-activated protein

Page 3 of 20
Organic & Biomolecular Chemistry
View Article Online
DOI: 10.1039/C8OB01406J
kinase (MAPK) c-Jun N-terminal kinase 1 (JNK1) that is created by connecting a D-domain
peptide (IC50>50 µM) from a JNK1 scaffolding protein with SP600125 (an ATP competitive
inhibitor targeting the JNK1 family) using a glycine-glycine-diamino propyl linker.10 This bivalent
inhibitor showed a greater than 20,000-fold increase in inhibition (IC50=0.7 nM) compared with
that of the ATP competitive inhibitor SP600125 (IC50>14 µM). Multiple examples of hetero-bivalent
ligands that fall into this second category have also been reported.11-14
RESULTS AND DISCUSSION
Here we describe another interesting approach to hetero-bivalent ligand design that
utilizes a peptide sequence derived from a portion of the same kinase peptide sequence that acts
as a linker between two domains as one of the targeting moieties of the bivalent molecule (the
other targeting moiety is the ATP-binding site-targeted molecule). When we examined the EphA3
kinase domain crystal structure,15 we noted that a portion of the linker peptide region (amino acids
900–904) that connects the kinase domain to the sterile alpha motif (SAM) domain at the C-
terminus seems to be turning back and interacting with the kinase domain at the bottom (Fig. 1).
This 5-mer peptide sequence, NLLLD [also referred to as the EphA3 SAM Linker Peptide (ESL
peptide)], appears to be a true binding moiety on a unique binding pocket situated much further
away from the ATP-binding site (Figs. 1B and 1C). The side chains of Leu901 and Asp904 are
accommodated in complimentary pockets formed by G and H of the EphA3 kinase domain,
and interactions include three backbone-backbone hydrogen bonds with the GH loop (residues
Tyr841–Pro846) (Fig. 1C). Therefore, we envisioned developing a unique hetero-bivalent ligand
using this peptide sequence as the non-ATP binding moiety that could be connected to an ATP-
binding compound through an optimized linker (Fig. 1A). It is important to mention that our intent
here is not really developing a drug-lead for future use because the contribution of EphA3 kinase
inhibition for blocking cancer is still an ambiguous area of research.16-18 We selected EphA3

Organic & Biomolecular Chemistry
Page 4 of 20
View Article Online
DOI: 10.1039/C8OB01406J
kinase solely as a model system to use in a proof-of-principle study for our novel hetero-bivalent
ligand design hypothesis after observing this very unique feature of its crystal structures solved
so far.19-22
Figure 1: EphA3 kinase domain crystal structure and unique hetero-bivalent ligand design. (A)
Components of hetero-bivalent ligand that target binding sites. (B) ATP-binding site and unique
EphA3 SAM linker (ESL) peptide binding site located at the bottom of the kinase domain. (C)
Crystal structure showing ESL peptide segment binding to a unique binding site on the bottom of
the kinase domain.
The first task was to determine the approximate size of the linker that could effectively
connect two moieties (ATP-binding compound and ESL peptide). When considering the much
longer distance that this linker would be required to cover, we planned a two-step approach, while
giving more attention to the types of moieties we should use to build the linker. We reasoned that
the linker would have to be flexible to wrap around major portion of the kinase domain as freely

Page 5 of 20
Organic & Biomolecular Chemistry
View Article Online
DOI: 10.1039/C8OB01406J
as possible. Even though the high degree of entropy associated with increased flexibility may
negatively contribute to the binding process (i.e. higher entropic penalty at the binding event), we
still chose the most flexible linker residues to develop our model system. The goal was to provide
full freedom for the linker to fall onto the kinase surface and wrap around the most appropriate
path when the two binding moieties on the two ends of the ligand interacted at their respective
binding hot spots. Subsequently we would be able to rigidify this linker as much as possible to get
rid of the expected entropic penalty of the binding event.
We used both combinatorial and structure-based approaches to build this linker region in
a two-step process. The first step was to determine the approximate distance using the
combinatorial approach. The second step was to fine-tune the distance measurement by focusing
more on structural features. We chose a purine analog as the ATP-targeting moiety considering
the previously published work indicating the ease of synthesizing this analog .23 We utilized two
types of building blocks to develop the linker region, namely (I) AEEAc [polyethylene glycol (PEG)-
based amino-ethoxy-ethoxy-acetyl moiety) and (II) glycine residues (Scheme 1 and Fig. 2A). The
AEEAc would contribute a more water-soluble characteristic, whereas glycine would provide
some hydrophobicity. The AEEAc moiety, being nine atoms longer, would help build the
necessary extended lengths, while glycines would be useful in fine-tuning the lengths of the
linkers. Both moieties would provide the much needed flexibility. We first built 11 linkers that
covered a wide range of 54 -106 Å (Fig. 2A). Again, the goal was to allow this linker region to
wrap around the kinase in any conformation, including situations in which the optimized linker
may be not the shortest possible, but an extended linker which could find additional interactions
by falling onto various areas of the kinase surface (e.g., H-bonds through multiple amide bonds
within the linker). The linker lengths were varied by combining different numbers of the
aforementioned building blocks (Scheme 1).

Organic & Biomolecular Chemistry
Page 6 of 20
View Article Online
DOI: 10.1039/C8OB01406J
Scheme 1: General procedure for two rounds of hetero-bivalent derivative synthesis. R1 and R2
represent two different ATP-binding site directed moieties used (dark blue). ESL peptide
sequence is show in red and the linker region is shown in light blue.
We modified our previously published complete on-bead synthesis procedure to
synthesize the hetero-bivalent ligands in this study .24 We selected NovaSyn TGR resin
(MilliporeSigma, Darmstadt, Germany with low loading (0.20–0.30 mmol/g), which provides
enough space to grow complex high molecular weight molecules without much steric hindrance.
The first residue attached was Fmoc-Glu(biotinyl-PEG)-OH , which is essential for the subsequent

Page 7 of 20
Organic & Biomolecular Chemistry
View Article Online
DOI: 10.1039/C8OB01406J
binding assay to detect the binding event (Scheme 1, step 1). We next removed
fluorenylmethyloxycarbonyl (Fmoc) using 20% piperidine and coupled Fmoc-Lys(ivDde)-OH (N-
α-Fmoc-N-ε-1-(4,4-dimethyl-2,6-dioxocyclohex-1-ylidene)-3-methylbutyl-L-lysine)
to
obtain
compound A. Lysine acts as a central linker in compound A, with two orthogonal protecting groups
(i.e., Fmoc and ivDde) on two branched arms to facilitate the synthesis of hetero-bivalent ligands
in a controlled fashion (Scheme 1, step 1). The ESL peptide-containing arm was built first by
selectively deprotecting the Fmoc protecting group using 20% piperidine in dimethylformamide
(DMF) and then elongating the linker portion by coupling Fmoc-AEEAc-OH moieties (y = 2, 3, 4,
and 5) to obtain intermediate compound B (Scheme 1, step 2). We then synthesized the NLLLD
ESL sequence as an elongation to compound B by coupling Fmoc-Asp-OH, Fmoc-Leu-OH (3
times), and Fmoc-Asn(Trt)-OH (MilliporeSigma, Darmstadt, Germany. Subsequently, the N-
terminal was capped using Boc-Gly-OH to obtain compound C (Scheme 1, step 3). Next, the ATP-
binding moiety-containing arm was built by deprotecting the ivDde protecting group using 5%
hydrazine in DMF and growing the linker using Fmoc-AEEAc-OH (x = 2, 3, and 4) and Fmoc-Gly-
OH moieties (n = 0, 1, and 2) to yield compound D (Scheme 1, step 4). After deprotecting Fmoc
using 20% piperidine in DMF, 6-amino purine hexanoic acid was coupled at the N-terminal to
obtain compound F (Scheme 1, step 5). Finally, the compounds were cleaved from the resin using
cleaving cocktail [95% trifluoroacetic acid (TFA), 2.5% triisopropylsilane (TIS), and 2.5% H2O] that
deprotects all protecting groups in the hetero-bivalent ligand (Scheme 1, step 6). The biotinylated
purine analogue was synthesized in large scale following the published protocol (Supplementary
FigS1). The biotinylated ESL peptide was synthesized using standard Fmoc chemistry
(Supplementary FigS2).

Organic & Biomolecular Chemistry
Page 8 of 20
View Article Online
DOI: 10.1039/C8OB01406J
Figure 2: First round of hetero-bivalent ligand design. (A) Outline of the hetero-bivalent ligand
design. (B) ELISA-like binding assay results of 11 hetero-bivalent ligands binding to EphA3 kinase
domain at 50 µM. EPHB1.4 showed the best binding.
We next performed an ELISA-like binding assay to identify the derivative(s) that
preferentially bound to the EphA3 kinase domain. We coated the white bottoms of 96-well nickel-
coated plates (Pierce™) with 50 µL of EphA3 recombinant human protein (Thermo Fisher
Scientific, Waltham, MA) (1 µg/mL) for 1 h in Tris-buffered saline (TBS). Wells were washed three
times with 200 µL TBS and blocked with StartingBlock Blocking Buffer (Thermo Fisher Scientific).
Fifty microliters of biotinylated compounds [(I) EphA3-binding hetero-bivalent (EPHB) derivatives,
(II) purine analogue, and (III) ESL peptide] in blocking buffer were added to wells and allowed to
incubate for 1 h. Solutions were removed, the wells were washed as above, and incubated for 1
h with 50 µL of streptavidin-horse radish peroxidase (HRP) from Novus Biologicals (Littleton, CO)

Page 9 of 20
Organic & Biomolecular Chemistry
View Article Online
DOI: 10.1039/C8OB01406J
at a 1:800 dilution in blocking buffer. Wells were similarly washed. Next, 50 µL of SuperSignal
ELISA Pico Chemiluminescent Substrate (Thermo Fisher Scientific) was added to each well. The
resultant luminescence signal was detected at all wavelengths using a Spectramax i3 microplate
reader (Molecular Devices, San Jose, CA). As shown in Figure 2B, derivative EPHB1.4 stands
out, displaying the highest binding, which has a length of approximately 64 Å. The derivatives that
were longer or shorter than EPHB1.4 did not show significant binding (with the exception of
EPHB1.2, which seems to have a slightly better binding than the others). As expected, the two
monomeric moieties (purine derivative and ESL peptide) displayed the lowest binding
characteristics.
Our next goal was to fine-tune the rough linker length we determined in the first round. At
the same time we wanted to replace the purine analog with a known molecule that targets the
ATP-binding site. We chose the PP2 analogue that recognizes the ATP-binding site. The PP2
molecule was initially identified to be a selective Src kinase inhibitor. However, later reports
suggest that it can inhibit multiple kinases (including EphA3) by competing for the ATP binding
site.25, 26 Therefore, we chose an analogue of PP2 for which the synthesis route is available. More
importantly, this analogue has been shown to attach to other moieties via click chemistry ,26 which
enabled us to use it for our purpose without extensive initial investigations (Fig. 3A). We decided
to explore a much narrower region of linker lengths, which is also more towards shorter linker
lengths in our previous range (51- 65 Å) for the second round of linker optimizations, because
derivative EPHB1.1 also showed somewhat higher binding in the preliminary evaluation (Fig. 2B)
in addition to the best binding derivative EPHB1.4. We designed five derivatives, again combining
different numbers of AEEAc,and glycine moieties (Fig. 3A).

Organic & Biomolecular Chemistry
Page 10 of 20
View Article Online
DOI: 10.1039/C8OB01406J
Figure 3: Second round of fine-tuned hetero-bivalent ligand design. (A) Outline of the fine-tuned
hetero-bivalent ligand design. (B) ELISA-like binding assay results of five focused hetero-bivalent
ligands binding to EphA3 kinase domain at 50, 200, and 500 nM. EPHB2.3 showed the best
binding.
The synthesis was completed following the same synthesis route except that
azidopropylamine was incorporated in place of R1 after the N-terminal Fmoc deprotection in step
5. We employed a standard microwave-assisted peptoid protocol using bromoacetic acid and
diisopropylcarbodiimide (DIC) followed by addition of 3-azido-1-propylamine.27, 28 Then, the PP2
analogue (R2) was incorporated using Cu(I)-catalyzed click chemistry with N-methyl-2-

Page 11 of 20
Organic & Biomolecular Chemistry
View Article Online
DOI: 10.1039/C8OB01406J
pyrrolidone (NMP) as solvent. Next, the resin was treated with cleaving cocktail (95% TFA, 2.5%
TIS, and 2.5% H2O) to cleave compounds from the resin and simultaneously remove the
protecting groups to yield hetero-bivalent ligands (Scheme 1, step 6). The PP2 was synthesized
in large scale following the published protocol (Supplementary FigS3), and biotinylation was
completed using click chemistry (Supplementary FigS4).
We then performed the same ELISA-like binding assay on three different concentrations
of ligand (50, 200, and 500 nM). The derivative EPHB2.3 that had a linker length of approximately
57 Å displayed the strongest binding (Fig. 3B). The derivatives with next longer linker length
(EPHB2.4 with 61 Å) and the immediate shorter linker length (EPHB2.2 with 54Å) displayed lesser
binding characteristics, indicating EPHB2.3 is the optimized linker-containing derivative. Two of
the remaining derivatives (EPHB2.5 and EPHB2.1) that are furthest from EPHB2.3 in terms of
linker lengths (longest and shortest) displayed the lowest binding, further confirming that EPHB2.3
is the optimized derivative.
We next selected the best derivative (EPHB2.3), a weaker derivative (EPHB2.5) that had
a somewhat closer linker length to EPHB2.3, and two monomeric components (PP2 and ESL
peptide) to elucidate the full binding curves. The same ELISA-like binding assay (Fig. 4) provided
estimated dissociation constant (KD) values for EPHB2.3 and EPH2.5 in the range of
approximately 250 - 300 nM. The two monomeric units we utilized displayed very weak binding
characteristics. We could not obtain accurate binding curves due to their low solubility at higher
concentrations. Considering the previously published information about the PP2 interaction with
EphA3 kinase,26 we predicted the KD would be approximately 200 - 300 µM. Also, we could not
detect binding of ESL peptide in the range of concentrations we examined, but there are reports
of multiple crystal structures that display the binding of this sequence on a unique binding pocket
situated on the bottom of EphA3 kinase.19-22 Therefore, we believe ESL peptide is indeed a ligand
for EphA3, but with very low affinity (may be even KD at millimolar range). Taken together, this

Organic & Biomolecular Chemistry
Page 12 of 20
View Article Online
DOI: 10.1039/C8OB01406J
information indicates that the hetero-bivalent ligand EPHB2.3 we developed might have achieved
approximately a 1000-fold or more improvement.
Figure 4: Concentration-dependent binding curves of EPHB2.3, EPHB2.5, PP2, and ESL peptide
evaluated using ELISA-like binding assay.
Also, we performed ELISA-like binding assay of EPHB2.3 in the presence of Dasatinib,
which is a known potent EphA3 kinase inhibitor.29 We performed this assay by first coating the
nickel coated white round bottom plates with EPHA3 protein as described above. Then wells were
washed with 200uL TBS for three times and blocked with Starting block buffer. Fifty microliters of
Dasatinib (5 uM) in blocking buffer were added to each well to saturate ATP binding pockets of
EPHA3 kinase. Dasatinib untreated wells were used as controls. Then, solutions were removed
and wells were washed with TBS as discussed above. Next 50 uL solutions of biotinylated
EPHB2.3 (0.5 and 1 uM, separately) were added to each well and removed after 1 hr incubation.
These concentrations were chosen to provide saturation levels of EPHB2.3 (see Fig. 4). Similar
washing were performed in each well and then incubated with 50 uL of streptavidin-horse radish
peroxidase (HRP) in blocking buffer. Again wells were similarly washed. Then we added 50 uL of
SuperSignal ELISA Pico Chemiluminescent Substrate to each well and recorded luminescence
readings at all wavelengths. As shown in Fig. 5, Dasatinib clearly reduce the binding of EPHB2.3,

Page 13 of 20
Organic & Biomolecular Chemistry
View Article Online
DOI: 10.1039/C8OB01406J
indicating EPHB2.3 indeed targeting EphA3 kinase pocket (via PP2 moiety) where Dasatinib also
binds. This experiment also ruled out any concerns of non-specific binding derived readings in
our ELISA-like binding assay (which typically exists in such plate based ELISA-like assays), as
the only way the fluorescence signal from EPHB2.3 can be reduced is only when EPHB2.3 binds
to the EphA3 where Dasatinib also binds.
EPHB2.3
500000
EPHB2.3 + Dasatinib
400000
300000
200000
100000
0
Concentration of EPHB2.3 (nM)
Figure 5: Binding competition of biotinylated EPHB2.3 with Dasatinib against EPHA3 kinase using
ELISA like assay. Dasatinib readily competed with EPHB2.3 to bind to EphA3 kinase domain.
In addition, our PP2 analogue docking results (not shown) are consistent with previously
published reports in which PP2 as a type – I inhibitor, which inhibited the DFG in conformation of
kinases.30 In EphA3, this active conformation is majorly contributed by phosphorylation of two
tyrosines in juxtamembrane region (Tyr 596 and Tyr 602) and minorly by Tyr 779 in activation
segment.15, 31 In type – 1 inhibitor binding mode, the alkyne moiety of PP2 analogue (that connects
PP2 to the linker) faces the solvent exposed region, which directs the linker of EPHB2.3 towards

Organic & Biomolecular Chemistry
Page 14 of 20
View Article Online
DOI: 10.1039/C8OB01406J
the N terminus of activation loop segment. But, considering the flexibility of our linker, it has the
possibility of interacting with the activation loop segment or even fall outside of it, where in both
ways without affecting the overall conformation. Also, as per our knowledge, the ESL peptide
binding site was not previously reported to directly involve in binding of adaptor proteins. Most of
the reported adapter proteins (Nck, Sh2) for EPHA3 known to interact with two
phophorylated tyrosines at the juxtamembrane and the one in the activation loop.32-34 However, a
recent study reports that SAM domain linker region can induce the phosphorylation of
juxtamembrane region by disconnecting its interaction with the bottom of the EphA3 kinase.35 This
is the exact interaction we also studied here, and very interestingly, our ESL peptide sequence
also binds to the same binding site. Their study reports that disruption of the bonding between
Leu901 of the SAM domain linker region (i.e. ESL peptide of our study) from its αG-αH binding
pocket allosterically affect the auto phosphorylation of the juxtamembrane region. So it may be
possible to use our compound to reduce the auto phosphorylation of the juxtamembrane region,
which can possibly affect the binding of adapter proteins and downstream signaling.
CONCLUSIONS
Our approach provides a new way of designing hetero-bivalent ligands for kinases that
utilize unique short peptide sequences from the very same kinase that display unique interactions
within that protein. Because we derived the short peptide sequence from a linker region that
connects kinase domain to another domain, it is worth searching through available crystal
structures of kinases and even other proteins and their complexes to identify such short peptide
regions that may serve as ‘naturally provided’ binding ligands that can be utilized to build hetero-
bivalent molecules. Although the ESL sequence we utilized seems to have very low affinity by
itself, we have shown that when it is connected to another hotspot binding unit (e.g., ATP-binding
pocket), a significant increment of affinity can be achieved. The biggest advantage of our

Page 15 of 20
Organic & Biomolecular Chemistry
View Article Online
DOI: 10.1039/C8OB01406J
approach with regard to designing drug-leads is that we do not have to derive these sequences
by structure-based or combinatorial means because the sequences are naturally available for the
utilization. We just need to explore available crystal structures carefully with a new mode of
curiosity to identify such existing sequences. In addition, because these random binding events
have a much higher chance of occurring very uniquely within a certain kinase, we may be able to
extract the much-needed specificity for that kinase through hetero-bivalent ligand design. Both
the PP2 and ESL peptide monomers were very weak binders, and the linker we designed had the
maximum flexibility, factors that in general are not the optimum contributions for ligand design.
Even with these weakest possible characteristics, we have shown a significant improvement of
binding once those weaker characteristics are combined in a proper way to build a single binding
unit. Therefore, if we were to begin with already stronger binding units, we could achieve very
strong binding hetero-bivalent ligands. The immediate future direction for our project is to test the
specificity of EPHB2.3 over the whole kinome and also validate it affects the allosteric regulation
of SAM domain linker using kinase functional assay. In addition, providing some level of rigidity
to the linker, while exploring additional binding hot spots through its path on the kinase surface,
are other goals we are currently pursuing on this project. We believe these designs will provide
further improvements in the binding affinity as well as the specificity.
ACKNOWLEDGMENTS
We thank University of Houston for funding this project.
CONFLICT OF INTEREST
Authors declare no conflict of interest.

Organic & Biomolecular Chemistry
Page 16 of 20
View Article Online
DOI: 10.1039/C8OB01406J
REFERENCES
1.
2.
3.
4.
5.
D. Lavogina, E. Enkvist and A. Uri, ChemMedChem, 2010, 5, 23-34.
C. M. Gower, M. E. Chang and D. J. Maly, Crit Rev Biochem Mol Biol, 2014, 49, 102-115.
K. Parang and P. A. Cole, Pharmacology & therapeutics, 2002, 93, 145-157.
V. Lamba and I. Ghosh, Curr Pharm Des, 2012, 18, 2936-2945.
A. Ricouart, J. C. Gesquiere, A. Tartar and C. Sergheraert, Journal of medicinal chemistry,
1991, 34, 73-78.
6.
T. L. Schneider, R. S. Mathew, K. P. Rice, K. Tamaki, J. L. Wood and A. Schepartz, Org
Lett, 2005, 7, 1695-1698.
7.
K. Shen and P. A. Cole, Journal of the American Chemical Society, 2003, 125, 16172-
16173.
8.
K. Parang, J. H. Till, A. J. Ablooglu, R. A. Kohanski, S. R. Hubbard and P. A. Cole, Nat
Struct Biol, 2001, 8, 37-41.
9.
A. C. Hines and P. A. Cole, Bioorganic & medicinal chemistry letters, 2004, 14, 2951-
2954.
10.
11.
J. L. Stebbins, S. K. De, P. Pavlickova, V. Chen, T. Machleidt, L. H. Chen, C. Kuntzen, S.
Kitada, M. Karin and M. Pellecchia, Journal of medicinal chemistry, 2011, 54, 6206-6214.
S. C. Meyer, C. D. Shomin, T. Gaj and I. Ghosh, Journal of the American Chemical
Society, 2007, 129, 13812-13813.
12.
13.
T. K. Johnson and M. B. Soellner, Bioconjug Chem, 2016, 27, 1745-1749.
B. C. Lechtenberg, P. D. Mace, E. H. Sessions, R. Williamson, R. Stalder, Y. Wallez, G.
P. Roth, S. J. Riedl and E. B. Pasquale, ACS Med Chem Lett, 2017, 8, 726-731.
A. A. Profit, T. R. Lee and D. S. Lawrence, Journal of the American Chemical Society,
1999, 121, 280-283.
14.
15.
T. L. Davis, J. R. Walker, P. Loppnau, C. Butler-Cole, A. Allali-Hassani and S. Dhe-
Paganon, Structure, 2008, 16, 873-884.

Page 17 of 20
Organic & Biomolecular Chemistry
View Article Online
DOI: 10.1039/C8OB01406J
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
P. W. Janes, C. I. Slape, R. H. Farnsworth, L. Atapattu, A. M. Scott and M. E. Vail, Growth
Factors, 2014, 32, 176-189.
B. P. Fox, C. J. Tabone and R. P. Kandpal, Biochemical and biophysical research
communications, 2006, 342, 1263-1272.
N. I. Herath, M. D. Spanevello, J. D. Doecke, F. M. Smith, C. Pouponnot and A. W. Boyd,
Eur J Cancer, 2012, 48, 753-762.
H. Zhao, J. Dong, K. Lafleur, C. Nevado and A. Caflisch, ACS Med Chem Lett, 2012, 3,
834-838.
K. Lafleur, J. Dong, D. Huang, A. Caflisch and C. Nevado, Journal of medicinal chemistry,
2013, 56, 84-96.
A. Unzue, J. Dong, K. Lafleur, H. Zhao, E. Frugier, A. Caflisch and C. Nevado, Journal of
medicinal chemistry, 2014, 57, 6834-6844.
J. Dong, H. Zhao, T. Zhou, D. Spiliotopoulos, C. Rajendran, X. D. Li, D. Huang and A.
Caflisch, ACS Med Chem Lett, 2015, 6, 79-83.
H. S. Lim, C. T. Archer and T. Kodadek, Journal of the American Chemical Society, 2007,
129, 7750-7751.
J. C. Hooks, J. P. Matharage and D. G. Udugamasooriya, Biopolymers, 2011, 96, 567-
577.
K. R. Brandvold, M. E. Steffey, C. C. Fox and M. B. Soellner, ACS chemical biology, 2012,
7, 1393-1398.
K. R. Brandvold, S. M. Santos, M. E. Breen, E. J. Lachacz, M. E. Steffey and M. B.
Soellner, ACS chemical biology, 2015, 10, 1387-1391.
R. N. Zuckermann, J. M. Kerr, S. B. H. Kent and W. H. Moos, Journal of the American
Chemical Society, 1992, 114, 10646-10647.
H. J. Olivos, P. G. Alluri, M. M. Reddy, D. Salony and T. Kodadek, Org. Lett., 2002, 4,
4057‒4059.

Organic & Biomolecular Chemistry
Page 18 of 20
View Article Online
DOI: 10.1039/C8OB01406J
29.
M. W. Karaman, S. Herrgard, D. K. Treiber, P. Gallant, C. E. Atteridge, B. T. Campbell, K.
W. Chan, P. Ciceri, M. I. Davis, P. T. Edeen, R. Faraoni, M. Floyd, J. P. Hunt, D. J.
Lockhart, Z. V. Milanov, M. J. Morrison, G. Pallares, H. K. Patel, S. Pritchard, L. M.
Wodicka and P. P. Zarrinkar, Nat Biotechnol, 2008, 26, 127-132.
30.
31.
32.
33.
34.
35.
N. K. Williams, I. S. Lucet, S. P. Klinken, E. Ingley and J. Rossjohn, J Biol Chem, 2009,
284, 284-291.
N. Singla, H. Erdjument-Bromage, J. P. Himanen, T. W. Muir and D. B. Nikolov, Chemistry
& biology, 2011, 18, 361-371.
A. H. Zisch, C. Pazzagli, A. L. Freeman, M. Schneller, M. Hadman, J. W. Smith, E.
Ruoslahti and E. B. Pasquale, Oncogene, 2000, 19, 177-187.
I. D. Lawrenson, S. H. Wimmer-Kleikamp, P. Lock, S. M. Schoenwaelder, M. Down, A. W.
Boyd, P. F. Alewood and M. Lackmann, Journal of cell science, 2002, 115, 1059-1072.
S. J. Holland, N. W. Gale, G. D. Gish, R. A. Roth, Z. Songyang, L. C. Cantley, M.
Henkemeyer, G. D. Yancopoulos and T. Pawson, EMBO J, 1997, 16, 3877-3888.
A. Kwon, M. John, Z. Ruan and N. Kannan, J Biol Chem, 2018, 293, 5102-5116.

Page 19 of 20
Organic & Biomolecular Chemistry
View Article Online
DOI: 10.1039/C8OB01406J
80x38mm (300 x 300 DPI)