1. Collaborative science to identify novel inhibitors for the pseudokinase TRIB2
Morgan Focas
Biology Faculty Sponsor: Dr. Mark Peifer
Principal Investigator: Dr. Tim Willson
Introduction
The Structural Genomics Consortium (SGC) is a global open science network that
focuses on the lesser-studied areas of the human genome. By focusing on these less well-studied
areas and making the reagents and information publically available, they hope to catalyze
additional research in both industry and academia. At UNC, the SGC is focused on the
historically understudied kinome. Their interest in this area originated with a publication that
analyzed the scientific literature and found that both academia and industry are working on less
than 50% of the kinome despite RNAi and driver mutation data showing that many other kinases
that are not being studied have direct disease relevance1.
Figure 1. The targeted kinome in 2010. (a) The kinases are ordered by number of citations in
PubMed. In the same graph kinases are also highlighted that were identified by RNAi as
potential targets that are required for growth and survival of the cancer type. Those with a
defined genetic background are indicated in the left column and those with statistically relevant
driver mutations are indicated in the right column. (b) Number of inhibitor patents per kinase
target. In each case, a kinase has only been included as a target of a patented compound if
activity data has been reported in the body of the patent.
Kinases are enzymes that catalyze the transfer of a phosphate group from ATP onto a
substrate protein. Kinases mediate many diverse cellular processes including proliferation,
transcription, and apoptosis. As a result of the important role they play in cellular signaling,

2. kinases have become an important area of research and this family has yielded over 30 approved
drugs in the US. They are promising targets for the treatment of diseases such as various cancers,
inflammatory diseases, central nervous system disorders, and diabetic complications. By
competing with ATP at the binding pockets of kinases, small molecule inhibitors have the
potential to remedy many diseases. The currently available and characterized kinase inhibitors
only potently inhibit a small fraction of the human kinome, creating a need to further investigate
understudied kinases that have the potential to be therapeutically relevant in disease2.
The therapeutic importance and lack of good tools fueled the creation of the Published
Kinase Inhibitor Set (PKIS), a set of 367 kinase inhibitors that have been widely distributed
under an open access philosophy in order to facilitate productive collaborations. This set has
been distributed to over 300 collaborators around the world covering a diversity of research
ranging from oncology to developmental biology to tropical diseases. This initial set of
compounds was chosen from kinase inhibitors that had been published by chemists from
GlaxoSmithKline (GSK). The criteria employed when selecting the constituents of PKIS
included physical availability of the compound, reduction of overrepresented chemotypes, and
diversity of original targets1. Based on the cross reactivity of ATP-competitive kinase inhibitors,
the PKIS set was screened against 220 kinases and mutants and previously unknown activities
emerged. These compounds were then screened through numerous phenotypic screens by
collaborators to reveal kinases that drive specific phenotypes. When screening and profiling
PKIS, a challenge is finding compounds with potency as well as selectivity. Based upon these
screening results the SGC-UNC looks at the compounds and the activity in the assay to
determine on structure activity relationships (SAR), and then develops a chemistry plan to
improve both selectivity and potency against a specific kinase.
One collaborator that we have worked with this semester is interested in a pseudokinase
named Tribbles2 (TRIB2). TRIB2 is a member of a family of kinase-like proteins that control an
array of biological functions such as mitosis, cell differentiation, transcription factor
stabilization, modulation of gene expression, and apoptosis. They differ from enzymatically
competent kinases in that they lack amino acid motifs required for catalysis and have a
conserved C-terminal patch3. The Tribbles kinase-like domain has been highly conserved
throughout evolution, suggesting that it has an important biological function4. Tribbles have been
linked to cellular signaling involved in metabolic and proliferative diseases like atherosclerosis
and cancer, which is why they have become a novel class of drug target. TRIB2 is a Tribbles
pseudokinase that is specifically associated with proliferation as well as other cancer sustaining
transcriptional programs5.
The binding of small molecule ligands can increase the thermal stability of a protein. The
stability is related to the Gibbs free energy of unfolding; if a compound binds to the protein then
the free energy contribution results in an increase in melting temperature of the protein. This can
be measured by an increase in the fluorescence of a dye with affinity for hydrophobic parts of the
protein. As the protein unfolds, those parts are exposed and fluorescence increases6. This
technique, a differential scanning fluorimetry (DSF) assay, assesses the ligand’s ability to bind to
TRIB2, which then can identify potent TRIB2 inhibitors. These hits can then be further
evaluated to determine the downstream effects and to see if they offer potential treatments for a
variety of diseases7.

3. Figure 2– GW881A – PKIS thienopyrimidine hit from TRIB2 DSF assay.
PKIS compounds were evaluated in this assay with one potent hit (GW881A) identified
with a ΔTm = 5.5°C in the DSF assay, suggesting that it could be a potent inhibitor. Based on this
data, we were tasked with synthesizing more of the compound for further evaluation in TRIB2
assays.
Experimental Procedures
Synthesis of 6-bromo-N-(3-chloro-4-((3-fluorobenzyl)oxy)phenyl)thieno[3,2-
d]pyrimidin-4-amine 3.
To a solution of 1 (149mg, 0.59mmol) in ethanol (7.3 mL) at room temperature was
added 2 (150mg, 0.59mmol), and then this mixture was stirred at 60°C overnight. The reaction
had gone to completion by TLC (30/70 solvent?). To the reaction mixture was added 100mL of
water. The precipitated solid was collected and washed with water. The yellow solid was then
dried to give 3 (140mg, 50% yield). 1H NMR (400 MHz, DMSO-d6) δ= 9.66 (s,1H), 8.48 (s,
1H), 7.85 (d, 1H), 7.63 (s, 1H), 7.53 (dd, 1H), 7.41 (q, 1H), 7.27 (dd, 2H), 7.20 (d, 1H), 7.12 (t,
1H), 5.20 (s, 2H).
Synthesis of N-{3-chloro-4-[(3-fluorophenyl)methoxy]phenyl}-6-(1H-pyrrol-2-
yl)thieno[3,2-d]pyrimidin-4-amine 5.
To a microwave vial was added 3 (80.9 mg, 0.174 mmol), 4 (56.1 mg, 0.191 mmol),
DMF (1.13 mL), toluene (1.13 mL), Na2CO3 (0.5 mL, 2 M solution), and finally Pd(dppf)Cl2 (10
%, 12.7 mg, 0.0174 mmol). The mixture was heated in the microwave at 160°C for 15 minutes.
The solution was then run on a TLC plate (50/50 ethyl acetate/heptanes) Rf = 0.4. To the solution
was added ethyl acetate, washed with brine (3x 50 mL), dried over Na2SO4, and concentrated in
vacuo. The residue was subjected to column chromatography on silica gel, with heptanes/ethyl
acetate in a 50:50 (v/v) ratio as eluent to furnish BOC-protected 5 as 80% pure yellow solid by
LC/MS. The semi-pure 5 was dissolved in DCM (5 mL) and to the solution was added TFA (0.5
mL). Reaction was complete in five minutes as indicated by TLC. The solvent was removed in
vacuo. The residue was dissolved in DCM and subjected to column chromatography on silica
gel, with heptanes/ethyl acetate in a 50:50 (v/v) ratio as eluent to furnish N-{3-chloro-4-[(3-
fluorophenyl)methoxy]phenyl}-6-(1H-pyrrol-2-yl)thieno[3,2-d]pyrimidin-4-amine 5 (4.9mg, 6%
yield).1H NMR (400 MHz, MeOD-d6) δ= 8.45 (s,1H), 7.74 (s, 1H), 7.45 (dd, 1H), 7.37 (q, 1H),
7.28 (s, 2H), 7.22 (d, 1H), 7.15 (d, 1H), 7.02 (t, 2H), 6.95 (s, 1H), 6.62 (d, 1H), 6.20 (d, 1H),
5.44 (s, 2H).
Results and Discussion
Chemistry
We needed to synthesis additional GW881A to progress into further experiments for TRIB2. We
outlined the synthetic route in scheme 1. It is a three-step synthesis starting with an aromatic

4. substitution reaction with an aniline followed by a Suzuki reaction and lastly a removal of the
BOC protecting group.
Scheme 1- Proposed synthetic route to GW881A
The thienopyrimidine 1 was reacted with the aniline 2 under basic conditions to furnish
compound 3. The next step (b) was a Suzuki cross-coupling reaction with the commercially
available BOC-pyrrole boronic acid pinacol borane 4. Purification at this step yielded semi-pure
product. Instead of repurifying the intermediate, we proceeded directly to the BOC reaction. The
semi-pure sample was treated with TFA and was further purified to yield the desired final
product 5. The combined yield over two steps (b, c) was low due to multiple purifications
performed in addition to poor resolution of the product and the impurities.
Biology
A small focus set of kinase inhibitors were screened through the TRIB2 DSF assay. This
assay identified several hits from the thienopyrimidine series. Within the thienopyrimidine
series, there were other compounds that were synthesized with slight variations at the R position
(Table 1) while maintaining the same aniline head group. All of those compounds were screened
through the TRIB2 DSF assay, which gave us some insight into the SAR.

6. Table 1 – DSF data (ΔTm in °C with standard deviation) from PKIS on close structural analogs
of GW881A.
Based on the results from the DSF assay, we theorized that by synthesizing a small array
of compounds we could improve potency. Several analogs, for example GSK497A, added an
acetylene spacer to a pyrrolidine but these were all less potent than GW881A. When the pyrrole
present in GW881A was modified to be a furan, the potency dropped significantly. However
when the furan was linked to an N-methyl piperazine, the potency returned. This result indicated
that extending the linker past the pyrrole could be a way to increase potency. The pyrrole can
also be modified to a thiophene with a linker containing a pendant methylsulfonyl group to
maintain activity, further giving us confidence that a longer linker could improve potency as well
as other properties of the inhibitor.
One interesting result that emerged from the screening of commercially available kinase
inhibitors was the potency of Lapatinib against TRIB2. Lapatinib is published as a selective and
potent inhibitor of both Her2 and EGFR kinases8. Lapatinib maintains the same aniline head
group as GW881A but changes the core from the thienopyrimidine to a quinazoline (Figure 3).
Figure 3 – Structures of both GW881A and Lapatinib.
Lapatinib gave a ΔTm of 3.3°C, which is one of the most potent analogs identified. This
result could give another direction to make additional analogs for evaluation in the TRIB2 DSF
assay. A few analogs were also tested where the aniline was replaced (Table 2).
Table 2 - SAR around aniline modification
Although only two analogs were tested that had modifications to the aniline, both were
significantly less potent. These changes were perhaps too bulky or did not make favorable

7. interactions in the inhibitor pocket. We will further investigate the SAR with a small number of
additional analogs that have modifications at this position.
Conclusions and Next Steps
Based upon the thermal shift data collected around the thienopyrimidine series we have
planned a small array around each section of the molecule to interrogate how modifications made
to the structure can generate additional gains in potency (Figure 4). The molecule can be broken
down into three main portions: the core being the central thienopyrimidine, the aniline being the
aromatic head piece, and lastly the aryl group directly connected to the core.
Figure 4 - Three regions of GW881A where modifications will be made and evaluated
First, we will modify the aryl region based on what was observed with the existing SAR.
Figure 5 – Analogs modifying the aryl group
We will synthesize the analogs 6 and 7 to evaluate whether having a longer linker
improves potency. We will also replace the piperazine with a morpholine (8) to determine if
oxygen would make any favorable interactions. We also propose synthesizing the imidazole 9 to
S
N
N
HN
O
F
N
H
Cl
GW881Acore
aniline
aryl

8. evaluate the impact of additional nitrogen added to the ring. We will also synthesize the pyrrole
10 with a pendant ester at the 2-position.
For changes to the aniline position we propose making very modest modifications (Figure
6) based on the parent compound and commercially available anilines, since we have limited
knowledge of how changes affect activity at this position.
Figure 6 – Proposed analogs around modifications to the aniline
In all compounds we will remove the chloro and will introduce small changes such as
methylating the benzyl (14) or swapping the fluoro for an alcohol (13). These small changes will
give us some initial insight into whether this group can be modified while retaining activity.
Lastly we will synthesize an analog 15 changing the core to the quinazoline based upon
the screening results of Lapatinib (Figure 7).
Figure 7 – Compound designed to merge GW881A and Lapatinib
Based upon the screening results of the compounds we are proposing to synthesize, we
will then make a plan for further analogs.

9. References
1. Fedorov O., Muller S., Knapp S. The (un)targeted cancer kinome. Nature Chemical
Biology. 6, 166-169 (2010).
2. Wu P., Nielsen T., Clausen M. FDA- approved small-molecule kinase inhibitors. Trends
in Pharmacological Sciences. 36, 422-439 (2015).
3. Eyers, P. TRIBBLES: A Twist in the Pseudokinase Tail. Structure. 23, 1974-1975
(2015).
4. Hegedus Z., Czibula A., Kiss-Toth E. Tribbles: A family of kinase-like proteins with
potent signaling and regulatory function. Cellular Signaling. 19, 238-250 (2007).
5. Bailey F.P. et al. The Tribbles 2 (TRB2) pseudokinase binds to ATP and
autophosphorylates in a metal-independent manner. Biochem. J. 467, 47-62 (2015).
6. Niesen F. H., Berglund H., Vedadi M. The use of differential scanning fluorimetry to
detect ligand interactions that promote protein stability. Nature Protocols. 2, 2212-2221
(2007).
7. Foulkes D.M. et al. Tribbles pseudokinases: novel targets for chemical biology and drug
discovery? Biochem. Soc. Trans. 43, 1095-1103 (2015).
8. Konecny G. et al. Activity of the Dual Kinase Inhibitor Lapatinib (GW572016) against
HER-2-Overexpressing and Trastuzamab-Treated Breast Cancer Cells. Cancer Res. 66,
1630 (2006).
Acknowledgements
I kindly thank Patrick Eyers (University of Liverpool) for providing the DSF screening data
against TRIB2 and SGC-UNC (Tim Willson, Carrow Wells, Alison Axtman, David Drewry,
Dave Morris, Bill Zuercher, and Kim Swain) for allowing me to have the opportunity to work
and learn in their laboratory this semester.