Discovery of a B-Cell Lymphoma 6 Protein-Protein Interaction Inhibitor by a Biophysics-Driven Fragment-Based Approach

Article abstract of DOI:10.1021/acs.jmedchem.7b00313

B-cell lymphoma 6 (BCL6) is a transcriptional factor that expresses in lymphocytes and regulates the differentiation and proliferation of lymphocytes. Therefore, BCL6 is a therapeutic target for autoimmune diseases and cancer treatment. This report presen

Full text of DOI:10.1021/acs.jmedchem.7b00313

Subscriber access provided by HKU Libraries
Article
Discovery of a B-cell lymphoma 6 Protein–Protein Interaction
Inhibitor by a Biophysics-driven Fragment-based Approach
Yusuke Kamada, Nozomu Sakai, Satoshi Sogabe, Koh Ida, Hideyuki Oki, Kotaro Sakamoto,
Weston Lane, Gyorgy Snell, Motoo Iida, Yasuhiro Imaeda, Junichi Sakamoto, and Junji Matsui
J. Med. Chem., Just Accepted Manuscript • Publication Date (Web): 04 May 2017
Downloaded from http://pubs.acs.org on May 4, 2017
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of Medicinal Chemistry is published by the American Chemical Society. 1155
Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 40
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Discovery of a Bꢀcell lymphoma 6 Protein–Protein
Interaction Inhibitor by a Biophysicsꢀdriven
Fragmentꢀbased Approach
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Yusuke Kamada*,†, Nozomu Sakai*,†, Satoshi Sogabe†, Koh Ida†, Hideyuki Oki†, Kotaro
Sakamoto†, Weston Lane‡, Gyorgy Snell‡, Motoo Iida†, Yasuhiro Imaeda†, Junichi Sakamoto†
and Junji Matsui†
†Takeda Pharmaceutical Company Limited, 26ꢀ1 MuraokaꢀHigashi 2ꢀchome, Fujisawa,
Kanagawa 251ꢀ8555, Japan
‡Takeda California Inc. 10410 Science Center Dr. San Diego CA 92121, USA
ABSTRACT: Bꢀcell lymphoma 6 (BCL6) is a transcriptional factor that expresses in
lymphocytes and regulates the differentiation and proliferation of lymphocytes. Therefore, BCL6
is a therapeutic target for autoꢀimmune diseases and cancer treatment. This report presents the
discovery of BCL6ꢀcorepressor interaction inhibitors by using a biophysicsꢀdriven fragmentꢀ
based approach. Using the surface plasmon resonance (SPR)ꢀbased fragment screening, we
successfully identified fragment 1 (SPR KD = 1200 µM, ligand efficiency (LE) = 0.28), a
competitive binder to the natural ligand BCoR peptide. Moreover, we elaborated 1 into the more
potent compound 7 (SPR KD = 0.078 µM, LE = 0.37, cellꢀfree proteinꢀprotein interaction (PPI)
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 40
1
2
3
4
5
6
IC50 = 0.48 µM (ELISA), cellular PPI IC50 = 8.6 µM (M2H)) by a structureꢀbased design and
structural integration with a second highꢀthroughput screening hit.
7
8
9
KEYWORDS: Bꢀcell lymphoma 6, surface plasmon resonance, fragmentꢀbased drug discovery,
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
fragmentꢀassisted drug discovery, proteinꢀprotein interaction
INTRODUCTION
Bꢀcell lymphoma 6 (BCL6) is a transcriptional factor that belongs to the bricꢀaꢀbrac, tramtrack,
broad complex/poxvirus zinc finger (BTB/POZ) family of proteins. It possesses BTB, RD2, and
zinc finger domains, and interacts with three corepressors, i.e., BCoR, SMRT, and NCoR.1 It
drives germinal center Bꢀcell formation and differentiation of T lymphocytes.2ꢀ5 It is also
involved in the differentiation and proliferation of diffuse large Bꢀcell lymphomas.6,
7
Consequently, it is thought to be an effective therapeutic target for the treatment of autoꢀimmune
diseases and cancer.
The crystal structure of the BCL6 BTB domain (BCL6BTB) complexed with the SMRT
peptide has already been reported.8 The crystal structure of the coꢀcomplex with the BCL6
inhibitor 79ꢀ6 has also been solved.9 These ligands bind in the lateral groove of the BCL6BTB
homodimer. Parekh et al. have suggested that the lateral groove of the BCL6BTB homodimer
could be an excellent therapeutic target to develop effective smallꢀmolecule inhibitors.6 Several
compounds and peptides, other than 79ꢀ6, have been reported as BCL6BTB inhibitors, including
RIꢀBPI with a KD of ~10 µM, Rifamycin SV with a KD of ~1 mM, and FX1 with a KD of 7 µM,
all of which have binding affinities weaker than 1 µM (Figure 1).10ꢀ12 Recently, we reported the
peptide F1324, which inhibits the BCL6ꢀBCoR interaction with a KD of 0.57 nM (Figure 1).13
Peptide F1324 also binds to the lateral groove of the BCL6BTB homodimer. These data suggest
ACS Paragon Plus Environment

Page 3 of 40
Journal of Medicinal Chemistry
1
2
that the lateral groove of the BCL6BTB homodimer could be a potent site for inhibitors to bind
with high affinity.
3
4
5
6
7
8
9
Fragmentꢀbased drug discovery (FBDD) has been widely practiced in both industry and
academia to identify compounds with good physiological profiles, new scaffolds, and new
binding sites.14, 15 The starting point of this approach is to identify molecule fragments with low
affinity using highꢀconcentration screening, and then to improve their affinity by linking or
growing them.16, 17 Biophysical assays (surface plasmon resonance (SPR), nuclear magnetic
resonance (NMR), thermal shift assays, and so on) have been used for the FBDD approach
because they are more sensitive than biochemical assays.18, 19 Of these, SPR biosensing is the
most costꢀeffective (low protein consumption and quick assay development) and provides highꢀ
content information on binding affinity and kinetics, and identification of panꢀassay interference
compounds (PAINS).20, 21 Thus, SPR biosensing assays are often used for the FBDD approach.22ꢀ
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
25
Proteinꢀprotein interactions (PPIs) represent a challenging target for the biophysicsꢀdriven
FBDD approach because the targets, like BCL6, do not possess a deep binding pocket.26ꢀ29
In this study, SPRꢀbased fragment screening was performed against BCL6 to discover BCL6ꢀ
cofactor interaction inhibitors. As a result, we successfully identified triazine fragment 1, which
binds to the lateral groove of the BCL6BTB homodimer where BCoR also binds. Furthermore,
fragment 1 was optimized into the more potent inhibitor 7, which exhibits cellular activity and
twoꢀdigit nanomolar binding affinity, by structureꢀbased drug design using Xꢀray analysis and
structural integration with a second highꢀthroughput screening (HTS) hit. This process represents
a highly successful fragmentꢀassisted drug discovery approach (FADD).30, 31
RESULTS AND DISCUSION
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 40
1
2
3
4
5
6
7
8
9
Protein preparation. To identify BCL6 inhibitors that disrupt the interaction between BCL6
and cofactors, we attempted to prepare BCL6BTB cofactors that bind to the lateral groove of the
BCL6BTB homodimer. BCL6BTB is approximately 15 kDa in size, but despite its small size, its
domain contains five cysteine residues. We speculated that the high proportion of Cys residues
would cause instability, which was reinforced by the fact that the coꢀcrystal structure of
BCL6BTB has been obtained using the Cysꢀmutant (Ala5ꢀGlu129, C8Q, C67R, and C84N) to
prevent aggregation.8 In addition, we needed to consider that the high concentration (mM)
required for SPR screening can cause false positives. In an attempt to mitigate this tendency for
false positives, we decided to use the Cysꢀmutation protein for SPRꢀbased screening.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
SPR assay development. For our SPRꢀbased screening, we used a Biacore 4000 highꢀ
throughput instrument with five detection areas for protein immobilization and four flow cells
for compound injection. This parallel immobilization enables the evaluation of selectivity and
nonspecific binding simultaneously.32ꢀ34 We also screened against both wild type (wt) and
mutant (mt) BCL6BTB by SPR. Figures 2A and 2B show the layout of the sensor chip used in this
screening. Aviꢀtagged wt BCL6BTB and aviꢀtagged mt BCL6BTB were captured onto spot 1 and
spot 2, respectively, by NeutrAvidin. In addition, wt BCL6BTB was covalently immobilized onto
spot 5 by an amine coupling reaction (Figure S1). This configuration enabled us to evaluate
binding to captured wt BCL6BTB, captured mt BCL6BTB, coupled wt BCL6BTB, and coupled
NeutrAvidin.
To verify the activity of each immobilized BCL6BTB, we evaluated the binding of the BCoR
peptide as a positive control (Figures 2C to 2F). The BCoR peptide successfully binds with each
BCL6BTB in a doseꢀdependent manner. The KD values for the BCoR peptide with each BCL6BTB
are in the twoꢀdigit micromolar range. Since there are no significant differences in the binding
ACS Paragon Plus Environment

Page 5 of 40
Journal of Medicinal Chemistry
1
2
affinities among the different BCL6BTBs, the mutation of the Cys residues is not thought to affect
the interaction between BCL6 and BCoR, and we concluded that each BCL6BTB is immobilized
while retaining its activity.
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For performing highꢀquality SPRꢀbased screening, maintaining the stability of the
immobilized protein is important.35ꢀ37 Therefore, selection of a suitable buffer is critical to retain
stability. Compared with that used in biochemical assays, the concentration of dimethyl sulfoxide
(DMSO) required for SPRꢀbased screening is generally higher to solubilize the fragments at high
concentration. Accordingly, the influence of DMSO on the activities of the immobilized
BCL6BTBs was evaluated (Table 1 and Figure S2). The KD and Rmax values for the BCoR peptide
are unaffected by the concentration of DMSO. Moreover both wt BCL6BTB and mt BCL6BTB are
stable for 36 h (500 injections) under buffer conditions of 1% or 5% DMSO. Therefore, we
decided to perform screening with a 5% DMSO buffer to ensure a 1 mM solubility for the
fragments.
Primary fragment screening. In a primary SPRꢀbased screening of our inꢀhouse fragment
library, we were able to obtain the screening data for captured wt BCL6BTB, captured mt
BCL6BTB, coupled wt BCL6BTB, and NeutrAvidin simultaneously using the Biacore 4000
apparatus. In the screening with the captured mt BCL6BTB, the sensorgrams against most of the
fragments show a reasonable boxꢀshape, indicating fast binding (Figure 3B). Conversely,
captured wt BCL6BTB presents many plots that indicate slow binding (Figure 3A). These data
suggest that many fragments reversibly reacted with the Cys residues on wt BCL6 due to a
concentration as high as 1 mM. Regarding the immobilization method, the binding responses of
the positive control (BCoR) to coupled wt BCL6BTB decrease depending on the injection number,
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 40
1
2
while the binding responses of the positive control to each captured BCL6BTB are stable (Figures
3A to 3C). These results indicate that captured mt BCL6BTB is optimal for stable screening.
3
4
5
6
7
8
9
In the primary screening, we identified 266 fragments for captured wt BCL6BTB, 256
fragments for captured mt BCL6BTB, 273 fragments for coupled wt BCL6BTB, and 50 fragments
for NeutrAvidin. The Venn diagram in Figure 2E summarizes the results of this fragment
screening. A total of 64 compounds bind to all the BCL6BTBs, hence these fragments were
selected as the first priority for the followꢀup doseꢀresponse tests.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Discovery and characterization of fragment 1. To confirm the reproducibility and doseꢀ
responsivity of the 64 fragments, we performed doseꢀtitration experiments. All the fragments
show binding responses in a doseꢀdependent and reversible manner (data not shown). Next, we
performed saturation transfer differenceꢀnuclear magnetic resonance (STDꢀNMR) experiments to
prioritize the 64 fragments. Seven out of the 64 fragments exhibit binding to BCL6 (data not
shown). Of seven binding fragments, we focused on fragment 1 (Figure 4A), which shows
binding to BCL6 by STDꢀNMR (Figure 4B) and furthermore competition with the BCoR peptide
by SPR competition experiments (Figure 4C). The KD value for 1 was determined to be 1.2 mM
by SPR, and its ligand efficiency (LE) was calculated to be 0.28 from its KD value. An LE of at
least 0.29 is necessary to develop an orally available candidate.38 Therefore, 1 is a good starting
point for elaboration into potent BCL6 inhibitors.
For further optimization of fragment 1, we attempted to determine its binding mode by
determining the coꢀcrystal structure of 1 with wt BCL6BTB. Coꢀcrystals of 1 with BCL6BTB were
obtained by soaking into unliganded BCL6BTB crystals in which the biological homodimer is
related by crystallographic 2ꢀfold symmetry. Electron density indicated that 1 binds to the lateral
ACS Paragon Plus Environment

Page 7 of 40
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
groove at the dimer interface and exhibits alternate conformations of the aminotriazine moiety
(Figures 4 and S4). The linker nitrogen forms a hydrogen bond with the mainꢀchain oxygen of
Met51. Comparison with the binding mode of 79ꢀ6 shows that 1 occupies the same site with a
similar interaction pattern (data not shown). Therefore, we evaluated the BCL6 binding affinities
of several pyrimidine derivatives that are structurally related to 1.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Structure-activity relationships of compounds related to fragment 1. Table 2 summarizes
the structureꢀactivity relationships of the pyrimidine derivatives along with the original fragment
1. Pyrimidine derivative 2 exhibits a weak binding potency (KD = 3000 µM), while the
introduction of fluorine (3) or chlorine (4) atoms onto the 5ꢀposition of compound 2 results in
significantly improved binding affinities (KD = 180 and 68 µM, respectively). Chloropyrimidine
4 exhibits good LE (0.38), and may be a promising compound for further optimization with the
aim to develop more potent BCL6 inhibitors.
Structural integration with a second HTS hit. Simultaneously, an independent HTS
campaign for BCL6 inhibitors identified the weak BCL6 inhibitor 5 (KD = 88 ꢁM, LE = 0.16),
which also has a pyrimidine core structure, as shown in Figure 5; compound 5 was discovered by
the HTS campaign using the ELISA assay monitoring the interaction between the BCL6BTB and
the BCoR peptide. The structural resemblance of 5 to 4 prompted us to integrate these two
structures, forming the “hybrid compounds” shown in Figure 5. Coꢀcrystal structure
determination of compounds 4 and 5 with BCL6BTB confirmed that these compounds exhibit the
same binding modes as that of the initial triazine fragment 1 (Figures 6 and S4). As expected, the
two aromatic rings connected by the amine linker in the two structures significantly overlap. The
coꢀcrystal structure of 4 reveals that the chlorine atom at the 5ꢀposition of the pyrimidine core
occupies a small lipophilic region of BCL6BTB. The binding mode information of the HTSꢀ
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 40
1
2
3
4
5
6
7
8
9
derived pyrimidine derivative 5 shows that the carbonyl oxygen of the rightꢀhandꢀside cyclic
amide moiety interacts with Glu115 of BCL6BTB. Therefore, based on the structural
superposition of 5 with 4, the hybrid compounds 6 and 7 were designed and synthesized (Table
3). We selected a simple unsubstituted pyridinylmethyl substituent for compound 7 while taking
synthetic tractability into account.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The pyrimidine derivatives 6 and 7 exhibit significantly improved BCL6 binding affinities
(KD = 9.3 and 0.078 ꢁM, respectively) compared to those of the fragment derivative 4 and HTS
hit 5. To confirm the binding modes, the coꢀcrystal structures of compounds 6 and 7 were
determined, as well as those of the other compounds (Figures 7A and S4). The electron density
map shows that the pyridine moiety of compound 7 is exposed to solvent and mostly disordered.
Both compounds bind to BCL6BTB in the same binding mode as those for compounds 1, 4, and 5.
The linker nitrogen forms a hydrogen bond with the mainꢀchain oxygen of Met51, and the
carbonyl oxygen of the cyclic amide moiety interacts with Glu115. The reasons for the
significant improvement in binding affinity by introduction of the pyrimidine ring are unclear
because no specific structural features are observed.
As compound 7 exhibits twoꢀdigit nanomolar binding affinity along with good LE (0.37), we
also evaluated the BCL6ꢀBCoR PPI inhibitory activity of 7 in cellꢀfree and cellular assays
(Figures 7B to 7E). To evaluate the PPI inhibitory activities in cells, previously reported M2H
assay was used.13 Briefly, HEK293T cells were transfected with the GAL4ꢀresponsive reporter
plasmid, the bait expression plasmid (GAL4 DNAꢀbinding domain fused to wt BCL6, Ala5ꢀ
Glu129), and the VB expression plasmid (VP16 activation domain fused to BCoR, Leu112ꢀ
Ala753). Then, the PPI activities were evaluated by monitoring luciferase activities in the
transfected HEK293 cell lysates. Compound 7 exhibits potent cellꢀfree PPI inhibitory activity
ACS Paragon Plus Environment

Page 9 of 40
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
(ELISA IC50 = 0.48 µM) along with moderate cellular potency (M2H IC50 = 8.6 µM). Taken
together, compound 7 is a promising BCL6 inhibitor candidate for further exploration. The total
process of the fragmentꢀbased approach was summarized in Figure 8. Scaffold hopping of the
initial triazine fragment yielded the pyrimidine fragment 2, which enabled an introduction of a
substituent occupying the small lipophilic region of BCL6BTB. Structureꢀbased modification of
the resulting compound 4 along with the consolidation of HTS hit 5 led to the discovery of the
cellꢀactive compound 7. The HTS hit 5 had a poor ligand efficiency (LE = 0.16), and its complex
chemical structure made its modification rather difficult. Conversely, the fragmentꢀderived
compound 4 was a very good step for further modification due to its good ligand efficiency (LE
= 0.38) and small molecular size.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
CHEMISTRY
Synthetic methods for pyrimidine derivatives are shown in Scheme 1. The reaction of the 4ꢀ
chloropyrimidines 8–11 with anilines 12 and 13 give the corresponding 4ꢀanilinopyrimidines 2, 6,
14, 15, and 16. Amination of 14–16 with amines 17 and 18 at the 2ꢀposition of the pyrimidine
derivatives gives the diꢀ or triꢀsubstituted pyrimidine derivatives 3, 4, and 7.
CONCLUSIONS
To identify novel and promising starting points for the discovery of potent BCL6 inhibitors,
we adopted an FBDD approach, which represents a challenge given that the target is a PPI. We
identified 64 novel binding fragments using SPRꢀbased fragment screening, and among them,
fragment 1 showed moderate LE (0.28) and competition with the BCoR peptide. Moreover,
BCL6 binding affinities of several structurally related pyrimidine derivatives were also measured,
and we identified the more potent fragmentꢀlike pyrimidine derivative 4 (KD = 68 ꢁM).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 40
1
2
3
4
5
6
7
8
9
Consolidation of the structural information of compound 4 with the independently obtained HTS
hit 5 led to the identification of the potent BCL6 inhibitor 7 (KD = 0.078 ꢁM, LE = 0.37), which
is >15,000ꢀfold more potent than the initial fragment 1. Compound 7 also exhibited an efficacy
in cellꢀfree and cellular PPI assays (ELISA IC50 = 480 nM, M2H IC50 = 8.6 ꢁM). These findings
suggest that the combination of biophysicsꢀdriven FBDD, SBDD, and FADD is a promising
strategy for hit identification and lead generation against challenging targets such as PPIs.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
EXPERIMENTAL SECTION
Preparation of BCL6. The fragments of His, avi, SUMO (LifeSensors), Flag, Tobacco Etch
Virus (TEV) protease recognition sequence, and BCL6BTB (human BCL6 from 5 to 129 aa) were
amplified by the polymerase chain reaction (PCR) and ligated into pET21 vector (Merck). The
mt BCL6BTB (C8Q, C67R, C84N) was constructed by the overlap PCR method. The proteins
were expressed in E. coli BL21 (DE3) (NIPPON GENE) and partially biotinylated on aviꢀtag by
endogenous BirA. The proteins were purified using NiꢀNTA (QIAGEN) and Superdex200 (GE
healthcare) columns. For Flagꢀtagged wt BCL6BTB, SUMO was digested by ULP1 (LifeSensors).
To remove ULP1, HisꢀaviꢀSUMO, and uncleaved protein, the digested solution was passed
through an NiꢀNTA column in 50 mM TrisꢀHCl pH 8.0, 300 mM NaCl, 5% glycerol, and 1 mM
DTT. To increase the purity, ionꢀexchange chromatography (monoQ, GE Healthcare) was
conducted. The purified proteins were concentrated to 2 mg/ml for storage at −80°C.
Fragment library. The library contained 1494 fragments. All fragments in the library had a
clogP <3.5, the number of Hꢀbond acceptors and donors were <6 and <3 respectively, and
rotatable bonds were <3. Heavy atom count ranged between 8 and 15. The molecular weight of
ACS Paragon Plus Environment

Page 11 of 40
Journal of Medicinal Chemistry
1
2
3
4
5
6
the fragments were <350 Da (average 180 Da), and the average of the aromatic rings was 1.5.
Fragments were dissolved at 200 mM in 100% DMSO and stored at −30 °C.
7
8
9
SPR. SPR biosensing experiments were performed at 22°C on Biacore 4000 and Biacore
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
S200 instruments equipped with SeriesS CM5 and SA sensor chips (GE healthcare).
HBSꢀP+ (10 mM Hepes pH 7.4, 150 mM NaCl, 0.05% Surfactant P20, GE healthcare)
supplemented with 1 mM DTT was used as the running buffer for immobilization. NeutrAvidin
(Thermo Fisher Scientific Inc.) was covalently coupled onto spots 1, 2, and 3 of a CM5 sensor
chip following the standard amine coupling procedure according to the manufacturer's
instructions. Typical immobilization levels of NeutrAvidin ranged from 10,000 to 13,000
resonance units (RUs). Subsequently, aviꢀtagged wt BCL6BTB and aviꢀtagged mt BCL6BTB were
injected to spots 1 and 2, respectively. The surfaces were blocked by injecting biocytin
(Thermofisher) to spots 1, 2, and 3. Approximately 13,000 RUs of each BCL6BTB were captured
by NeutrAvidin. Furthermore, wt BCL6BTB was immobilized on spot 5 via the standard amine
coupling procedure, resulting in immobilization levels of around 4,000 RUs.
For the interaction studies, binding experiments were performed in 20 mM Tris, pH 8.0, 150
mM NaCl, 0.01% Surfactant P20, 1 mM DTT, and 5% DMSO. Different concentration sample
solutions were injected for 30 or 60 s at a flow rate of 30 or 50 µL/min and the dissociation was
thereafter followed for up to 30, 60, or 150 s. Data processing and analysis were performed using
the Biacore 4000 and Biacore S200 evaluation software (GE healthcare). Solvent correction was
included as described in the Biacore software handbook. Sensorgrams were double referenced
prior to fitting the concentration series to a steadyꢀstate affinity model. The dissociation constant
KD was calculated using the equation below.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 40
1
2
3
4
KD = Rmax × C/R ꢀ C
5
6
7
8
9
where Rmax, R, and C correspond to the sample binding capacity of the surface (RU), the
normalized response of test sample (RU), and the concentration of the test solution (M),
respectively. LE was calculated using the equation below.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
LE = ∆G/HA = (−2.303 × R × T) × log (KD)/HA
where ∆G, HA, R, and T correspond to the Gibbs free energy (kcal/mol), heavy atom count, the
ideal gas constant (1.987 × 10−3 kcal/K/mol) and the temperature in Kelvin (K).39
For the primary screen, Takeda's fragment library were diluted to 1 mM in the running buffer
and injected for 30 s. Then, the dissociation was followed for 30 s. Every 33rd cycle, a positive
control
containing
100
µM
BCoR
peptide
((Arg498ꢀPro514)ꢀLys,
Acetylꢀ
RSEIISTAPSSWVVPPKꢀOH, chemically synthesized by Toray Research Center) and running
buffer as a blank were injected. Data processing and analysis were performed using the Biacore
4000 evaluation software. Solvent correction was included as described in the Biacore 4000
software handbook. The binding responses were normalized by the positive control and the
molecular weight of each sample. Fragments with a normalized binding response that exceeded
median + median absolute deviation were selected as primary hits. As for NeutrAvidin,
fragments with a normalized response that was more than 90% against their theoretical Rmax
were selected.
For competition studies, data were acquired for test compounds binding to BCL6 in the
presence of 100 µM BCoR peptide in the running buffer and samples. When a compound binds
to the same site as the BCoR peptide, the response for the compound decreases.40
ACS Paragon Plus Environment

Page 13 of 40
Journal of Medicinal Chemistry
1
2
X-ray Crystallography. Crystals of unliganded BCL6BTB for soaking experiments were
obtained as described previously.13 Crystallization was by vaporꢀdiffusion using the sittingꢀdrop
method from 0.1 M BisꢀTris pH 6.5, 0.7 M potassium/sodium tartrate at 20°C. To generate
proteinꢀligand complexes, crystals were typically soaked for 2 h in a reservoir solution
containing 1 mM ligand(s). Prior to data collection, crystals were immersed in the reservoir
solution with the addition of 30% glycerol as a cryoprotectant, and were flashꢀfrozen in liquid
nitrogen. Diffraction data were collected from a single crystal using the CCD detector Quantum
315 (ADSC) at beamlines 5.0.2 and 5.0.3 of the Advanced Light Source (Berkeley) under a 100
K nitrogen cryostream. The data were reduced and scaled with HKL2000.41 The structures were
solved by the molecular replacement method with Molrep42 of the CCP4 software suite43 using
the BCL6BTB structure (PDB code: 1R28) as a search model. The structures were refined through
an iterative procedure utilizing REFMAC44 followed by model building in COOT.45 The final
models were validated using Molprobity.46 Crystallographic processing and refinement statistics
are summarized in Table S1. All structural figures were generated using PyMOL (Schrödinger).
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
STD-NMR. 1H NMR spectra were recorded using a 3 mm NMR tube on a 600 MHz Avance
spectrometer (Bruker Biospin) equipped with a 5 mm TCI cryoprobe. All NMR samples were
prepared in 50 mM phosphate D2O buffer at pH 7.4 containing 150 mM NaCl, 1 mM DTTꢀd10,
and 0.2% DMSOꢀd6 in D2O solvent. The STD spectrum was acquired from a sample containing
400 µM fragment 1 and 20 µM BCL6BTB at 15°C with 32 scans with onꢀ and offꢀresonance
saturation frequencies of 0.4 ppm and 30 ppm, respectively, and a total saturation time of 2.0 s.
The decrease in signal intensity for onꢀresonance irradiation results from the transfer of
saturation from BCL6BTB to fragment 1. The STD spectrum of fragment 1 was obtained by
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 40
1
2
3
4
subtracting the onꢀresonance spectrum from the offꢀresonance spectrum. A positive signal in the
5
6
STD spectrum indicates binding of fragment 1 to BCL6BTB
.
7
8
9
ELISA. Biotinylation of the εꢀamino group of the Cꢀterminus Lys on the BCoR peptide was
carried out with the Biotinꢀ(AC5)2 SulfoꢀOSu (DOJINDO) according to the protocol
recommended by the manufacturer. The wells of a Nunc Maxisorp microplate (460ꢀ518) were
coated with streptavidin (SA) (Wako) and were blocked with phosphateꢀbuffered saline (PBS)
that contained 1.0% Casein. The biotinylated BCoR peptide was captured by the SA, and a wt
BCL6BTB (0.5 nM) solution in PBS that contained 0.05% Tween20 and 1 mM DTT (PBST) was
added to the wells. After washing with PBST, bound BCL6BTB was detected using horseradish
peroxidase (HRP)ꢀconjugated antiꢀFLAG antibody (Sigma). The amounts of HRP in the wells
were measured using a chemical luminescent regent (Thermo Fisher Scientific Inc.).13 Percent
inhibition was calculated based on wells without BCL6BTB as a high control and without
compound as a low control.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
M2H. The assay was performed according to a procedure described previously.1 The vectors
pGL4.35, pBind, and pACT were obtained from Promega Corp. As template DNA, human BCL6
cDNA was isolated by PCR from a human skeletal muscle cDNA library (TAKARA Bio) and
human BCoR cDNA was purchased from GeneCopoeia Inc. Each cDNA fragment was granted a
restriction site by PCR and digested with restriction enzymes to insert into pBIND or pACT,
respectively. M2H was performed in HEK293T cells that were transfected with the reporter
constructs pGL4.35 containing GAL4 special response element of firefly luciferase
(9×GAL4UAS); pBIND/GAL4ꢀBCL6 (Ala5–Glu129); and pACT/VP16ꢀBCoR (Leu112–
Ala753) by Fugene HD (Promega). The transfected cells were seeded at 1 × 104 cells/15 ꢁL/well
on 384ꢀwell plates (CORNNING) in Dulbecco’s modified Eagle’s medium that contained 10%
ACS Paragon Plus Environment

Page 15 of 40
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
fetal bovine serum. After incubation for 20 h at 37°C under 5% CO2, cells were lysed to measure
luciferase activity using the BrightꢀGlo luciferase assay system (Promega).13 Percent inhibition
was calculated based on the well without transfection of pBIND/GALꢀBCL6 (Ala5–Glu129) as
the high control and the well without the compound as a low control.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
Chemistry. H NMR spectra were obtained on a Bruker AVANCE II (300 MHz)
1
spectrometer. Chemical shifts for H NMR are reported in parts per million (ppm) downfield
from tetramethylsilane (δ) as an internal standard in deuterated solvent, and the coupling
constants (J) are given in hertz (Hz). The following abbreviations are used for spin multiplicity: s
= singlet, d = doublet, t = triplet, q = quartet, dd = doublet of doublets, dt = doublet of triplets, m
= multiplet, and brs = broad singlet. Reaction progress was monitored by thinꢀlayer
chromatography (TLC) analysis on silica gel 60 F254 plates (Merck) or an NH TLC plate (Fuji
Silysia Chemical Ltd.). Column chromatography was carried out on a silica gel column
(Chromatorex® NHꢀDM1020, 100–200 mesh; Fuji Silysia Chemical, Ltd.) or on PurifꢀPack (SI ϕ
60 ꢁM or NH ϕ 60 ꢁM; Fuji Silysia Chemical, Ltd.). Preparative HPLC (prepꢀHPLC)
purification was performed on a Waters 2525 equipped with a Lꢀcolumn2 ODS (20 × 150 mm
i.d., 5 ꢁm particle size, CERI, Japan), eluting with 0.1% TFA in distilled water (Mobile phase A)
and 0.1% TFA in acetonitrile (Mobile phase B) using the following elution gradient: 20% B to
55% B over 7 min followed by 100% B isocratic over 4 min at a flow rate of 20 mL/min. The
products were detected by monitoring the mass spectra recorded using a Waters ZQ2000 with
electrospray ionization. The purity of all compounds used in biological studies was determined to
be ≥95% by elemental analysis or HPLCꢀMS analysis. Lowꢀresolution mass spectra were
acquired using a Shimadzu UFLC/MS (Shimazu LCꢀ20AD/LCMSꢀ2020) in electron spray
ionization mode (ESI+). The column used was an Lꢀcolumn 2 ODS (3.0 × 50 mm i.d., 3 ꢁm;
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 40
1
2
3
4
5
6
7
8
9
CERI, Japan) at a temperature of 40°C and a flow rate of 1.2 or 1.5 mL/min. Condition 1:
Mobile phases A and B under acidic conditions were 0.05% TFA in water and 0.05% TFA in
MeCN, respectively. The proportion of mobile phase B was increased linearly from 5% to 90%
over 0.9 min, and maintained at 90% over the next 1.1 min. Condition 2: Mobile phases A and B
under neutral conditions were a mixture of 5 mmol/L AcONH4 and MeCN (9:1, v/v) and a
mixture of 5 mmol/L AcONH4 and MeCN (1:9, v/v), respectively. The proportion of mobile
phase B was increased linearly from 5% to 90% over 0.9 min, and maintained at 90% over the
next 1.1 min. Elemental analyzes were carried out by Sumika Chemical Analysis Service, Ltd.
All commercially available solvents and reagents were used without further purification. Yields
were not optimized. Compounds 1 and 5 are commercially available.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
N4-Phenylpyrimidine-2,4-diamine (2). A mixture of 4ꢀchloropyrimidinꢀ2ꢀamine (100 mg,
0.77 mmol), aniline (108 mg, 1.16 mmol), and conc. HCl (1 drop) in i-PrOH (1.0 mL) was
heated at 80°C for 2 h. The reaction mixture was diluted with saturated aqueous NaHCO3
solution and extracted with EtOAc. The organic layer was concentrated, and the residue was
purified by column chromatography (NH silica gel, eluted with 50% to 100% EtOAc in hexane).
The desired fractions were concentrated, and the residue was triturated with hexaneꢀEtOAc to
1
give compound 2 (98 mg, 69%) as an offꢀwhite solid. H NMR (DMSOꢀd6) δ 6.00 (1H, brs),
6.20 (2H, brs), 6.94 (1H, brs), 7.27 (2H, brs), 7.53–8.00 (3H, m), 9.08 (1H, brs). MS m/z 187.1
(M+H)+. Anal. Calcd for C10H10N4: C, 64.50; H, 5.41; N, 30.09. Found: C, 64.30; H, 5.56; N,
30.21.
5-Fluoro-N4-phenylpyrimidine-2,4-diamine (3). A mixture of compound 14 (63 mg, 0.28
mmol) and 28% aqueous ammonia solution (2.5 mL, 37.2 mmol) in i-PrOH (2.5 mL) was heated
at 150°C under microwave irradiation for 4 h. The reaction mixture was concentrated, and the
ACS Paragon Plus Environment

Page 17 of 40
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
residue was purified by prepꢀHPLC. The desired fraction was concentrated and neutralized with
saturated aqueous NaHCO3 solution. The precipitated solid was collected, washed with H2O and
IPE, airꢀdried, and dried in vacuo to give 3 (9.1 mg, 16%) as a colorless solid. 1H NMR (DMSOꢀ
d6) δ 6.18 (2H, s), 7.01 (1H, t, J = 7.1 Hz), 7.28 (2H, t, J = 7.7 Hz), 7.80–7.89 (3H, m), 9.04 (1H,
s). MS m/z 205.1 (M+H)+. Anal. Calcd for C10H9FN4·0.1 H2O: C, 58.30; H, 4.50; N, 27.20.
Found: C, 58.33; H, 4.67; N, 27.32.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
5-Chloro-N4-phenylpyrimidine-2,4-diamine (4). A mixture of 15 (100 mg, 0.42 mmol) and
2 M ammonia i-PrOH solution (5.0 mL, 10 mmol) was heated at 150°C under microwave
irradiation for 4 h. The reaction mixture was concentrated, and the residue was purified by prepꢀ
HPLC. The desired fraction was concentrated and neutralized with saturated aqueous NaHCO3
solution. The precipitated solids were collected, washed with H2O and IPE, and dried in vacuo to
1
give 4 (23 mg, 25%) as an offꢀwhite solid. H NMR (DMSOꢀd6) δ 6.41 (2H, s), 7.01–7.08 (1H,
m), 7.26–7.33 (2H, m), 7.74 (2H, d, J = 8.1 Hz), 7.91 (1H, s), 8.45 (1H, s); MS m/z 221.1
(M+H)+. HPLC analysis conditions 1: Rt = 0.65 min (100% purity); 2: Rt = 0.89 min (100%
purity).
5-((5-Chloropyrimidin-4-yl)amino)-1,3-dihydro-2H-indol-2-one (6). A mixture of 4,5ꢀ
dichloropyrimidine (119 mg, 0.80 mmol), 5ꢀaminoindolinꢀ2ꢀone (119 mg, 0.80 mmol), and TEA
(0.133 ml, 0.96 mmol) in DMSO (1.0 mL) was stirred at room temperature for 14 h. The reaction
mixture was diluted with a 9:1 mixture of H2O and EtOH (10 mL). The precipitated solid was
collected, washed with H2O, and dried in vacuo to give 6 (95 mg, 46%) as a pale brown solid. 1H
NMR (DMSOꢀd6) δ 3.49 (2H, s), 6.80 (1H, d, J = 8.2 Hz), 7.25–7.36 (1H, m), 7.40–7.49 (1H,
m), 8.39 (1H, s), 8.42 (1H, s), 9.02 (1H, s), 10.36 (1H, s). MS m/z 261.2 (M+H)+. Anal. Calcd for
C12H9ClN4O·0.8 H2O: C, 52.39; H, 3.88; N, 20.37. Found: C, 52.32; H, 4.12; N, 20.30.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 40
1
2
3
4
5
6
7
8
9
5-((5-Chloro-2-((pyridin-3-ylmethyl)amino)pyrimidin-4-yl)amino)-1,3-dihydro-2H-
indol-2-one (7). A mixture of 16 (120 mg, 0.41 mmol), pyridinꢀ3ꢀylmethanamine (132 mg, 1.22
mmol), and TEA (49 mg, 0.49 mmol) in DMSO (1.0 mL) was heated at 95 °C for 3 h. The
reaction mixture was directly purified by column chromatography (silica gel, eluted with 50% to
100% EtOAc in hexane, followed by 0% to 20% MeOH in EtOAc) to give 7 (34 mg, 22%) as a
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
yellow solid. H NMR (DMSOꢀd6) δ 3.40 (2H, brs), 4.37 (2H, brs), 6.62–6.77 (1H, m), 7.13–
7.71 (5H, m), 7.92 (1H, s), 8.37–8.52 (3H, m), 10.30 (1H, s). MS m/z 367.0 (M+H)+. Anal. Calcd
for C18H15ClN6O·0.1–H2O·0.1–EtOAc: C, 58.56; H, 4.27; N, 22.27. Found: C, 58.40; H, 4.52;
N, 22.12. HPLC analysis conditions 1: Rt = 0.54 min (100% purity); conditions 2: Rt = 0.80 min
(97.3% purity).
2-Chloro-5-fluoro-N-phenylpyrimidin-4-amine (14). A mixture of 2,4ꢀdichloroꢀ5ꢀ
fluoropyrimidine (500 mg, 2.99 mmol), aniline (0.287 ml, 3.14 mmol), and TEA (333 mg, 3.29
mmol) in i-PrOH (10 mL) was stirred at room temperature for 3 h. The reaction mixture was
diluted with hexane, and the insoluble materials were removed by filtration. The filtrate was
concentrated, and the resulting solid was washed with water and dried in vacuo to give 14 (530
mg, 79%) as a colorless solid. 1H NMR (DMSOꢀd6) δ 7.15 (1H, t, J = 7.2 Hz), 7.69 (2H, d, J =
8.2 Hz), 8.32 (1H, d, J = 3.5 Hz), 9.99 (1H, s). MS m/z 224.0 (M+H)+.
2,5-Dichloro-N-phenylpyrimidin-4-amine (15). A mixture of 2,4,5ꢀtrichloropyrimidine
(1025 mg, 5.59 mmol), aniline (520 mg, 5.59 mmol), and TEA (565 mg, 5.59 mmol) in i-PrOH
(20 mL) was heated at reflux for 14 h. The mixture was poured into H2O and extracted with
EtOAc. The organic layer was washed with H2O and brine, dried over MgSO4, and concentrated
in vacuo. The residue was purified by column chromatography (silica gel, eluted with 5% to 50%
ACS Paragon Plus Environment

Page 19 of 40
Journal of Medicinal Chemistry
1
2
EtOAc in hexane) to give 15 (900 mg, 67%) as a colorless oil. 1H NMR (DMSOꢀd6) δ 7.20 (1H,
s), 7.40 (2H, s), 7.58 (2H, d, J = 8.2 Hz), 8.38 (1H, s), 9.52 (1H, s). MS m/z 240.0 (M+H)+.
3
4
5
6
7
8
9
5-((2,5-Dichloropyrimidin-4-yl)amino)indolin-2-one (16).
A
mixture of 2,4,5ꢀ
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
trichloropyrimidine (0.50 g, 2.73 mmol) and 5ꢀaminoindolinꢀ2ꢀone (0.40 g, 2.73 mmol) in
DMSO (1.0 mL) was added with TEA (0.379 ml, 2.73 mmol) at room temperature, and the
resultant mixture was stirred at room temperature for 2 h. The resulting solution was diluted with
H2O/EtOH (5:1, 10 mL), and the precipitated solid was collected by filtration. The obtained solid
1
was washed with H2O and dried in vacuo to give 16 (634 mg, 79%) as a pale brown solid. H
NMR (DMSOꢀd6) δ 3.51 (2H, s), 6.82 (1H, d, J = 8.2 Hz), 7.24–7.30 (1H, m), 7.34 (1H, s), 8.31
(1H, s), 9.37–9.49 (1H, m), 10.33–10.51 (1H, m). MS m/z 295.1 (M+H)+.
ASSOCIATED CONTENT
Supporting Information.
The Supporting Information is available free of charge on the ACS Publications website at DOI:
SPR sensorgrams for parallel immobilization, evaluation of stability and DMSO effect, SPR
sensorgrams and equilibrium plots for compounds 2ꢀ6, electron density OMIT map, data
collection and refinement statistics (PDF), molecular formula strings (CSV)
Accession Codes
Atom coordinates and structure factors have been deposited in the Protein Data Bank with
accession codes: 5X4M (BCL6/1), 5X4N (BCL6/4), 5X4O (BCL6/5), 5X4P (BCL6/6), and
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 40
1
2
3
4
5X4Q (BCL6/7). Authors will release the atomic coordinates and experimental data upon article
5
6
publication.
7
8
9
AUTHOR INFORMATION
Corresponding Authors
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
*yusuke.kamada1@takeda.com. tel: +81ꢀ0466ꢀ32ꢀ2782; nozomu.sakai@takeda.com. tel: +81ꢀ
0466ꢀ32ꢀ1273
ORCID
Y.K.: 0000ꢀ0003ꢀ3701ꢀ2075
Author Contributions
YK and NS contributed equally to this work and wrote most of this paper. YK discovered and
analyzed the fragments and their relatives by SPR. NS designed and synthesized the compounds.
SS, WL, and GS conducted crystal structure analysis and wrote a part of this paper. KI prepared
the recombinant proteins. HO and MI conducted NMR experiments. KS conducted ELISA and
M2H assays. YI, JS, and JM supervised and supported this work.
Notes
The authors declared no potential conflicts of interest with respect to the research, authorship,
and publication of this article.
ACKNOWLEDGMENT
ACS Paragon Plus Environment

Page 21 of 40
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
The authors would like to thank Takamichi Muraki for biochemical experiments, Haihong Wu,
Yuuka Hirao, Yoshifumi Nishimura, and Hideaki Shimojo for STDꢀNMR experiments, Shugo
Ohno and Derek Cole for fragment library and compounds supply, and Kengo Okada and
Takashi Ito for suggestions, help, and support to this work. We also acknowledge Takeshi
Yamamoto, Takeshi Yasui, Takafumi Takai, Yoshihisa Nakada, Takashi Ichikawa, Takayuki
Tatamiya, and Masashi Iwata for their kind suggestions. We also thank the staff of the Berkeley
Center for Structural Biology who operated the ALS beamlines 5.0.2 and 5.0.3. Berkeley Center
for Structural Biology is supported in part by the U.S. National Institutes of Health and U.S.
National Institute of General Medical Sciences. The Advanced Light Source is supported by the
Director, Office of Science, Office of Basic Energy Sciences of the U.S. Department of Energy
under Contract No. DEꢀAC02ꢀ05CH11231.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ABBREVIATIONS
BCL6, Bꢀcell lymphoma 6; BCoR, BCL6 corepressor; SMRT, silencing mediator of retinoic acid
and thyroid hormone receptor; NCoR, nuclear receptor corepressor; M2H, mammalian two
hybrid assay; Et3N, triethylamine; EtOAc, ethyl acetate; EtOH, ethanol; IPE, diisopropyl ether;
MeCN, acetonitrile; MeOH, methanol; iꢀPrOH, 2ꢀpropanol; TEA, triethylamine; TEV, tobacco
etch virus; RU, resonance unit.
REFERENCES
(1)
Huynh, K. D.; Fischle, W.; Verdin, E.; Bardwell, V. J. BCoR, a novel corepressor
involved in BCLꢀ6 repression. Genes Dev. 2000, 14, 1810ꢀ1823.
(2)
Dent, A. L.; Shaffer, A. L.; Yu, X.; Allman, D.; Staudt, L. M. Control of inflammation,
cytokine expression, and germinal center formation by BCLꢀ6. Science 1997, 276, 589ꢀ592.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 40
1
2
3
4
(3)
Linterman, M. A.; Vinuesa, C. G. Signals that influence T follicular helper cell
5
6
differentiation and function. Semin. Immunopathol. 2010, 32, 183ꢀ196.
7
8
9
(4)
Nurieva, R. I.; Chung, Y.; Martinez, G. J.; Yang, X. O.; Tanaka, S.; Matskevitch, T. D.;
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Wang, Y. H.; Dong, C. Bcl6 mediates the development of T follicular helper cells. Science 2009,
325, 1001ꢀ1005.
(5)
Xin, N.; Fu, L.; Shao, Z.; Guo, M.; Zhang, X.; Zhang, Y.; Dou, C.; Zheng, S.; Shen, X.;
Yao, Y.; Wang, J.; Wang, J.; Cui, G.; Liu, Y.; Geng, D.; Xiao, C.; Zhang, Z.; Dong, R. RNA
interference targeting Bclꢀ6 ameliorates experimental autoimmune myasthenia gravis in mice.
Mol. Cell. Neurosci. 2014, 58, 85ꢀ94.
(6)
Parekh, S.; Prive, G.; Melnick, A. Therapeutic targeting of the BCL6 oncogene for
diffuse large Bꢀcell lymphomas. Leuk. Lymphoma 2008, 49, 874ꢀ882.
(7)
Duy, C.; Hurtz, C.; Shojaee, S.; Cerchietti, L.; Geng, H.; Swaminathan, S.; Klemm, L.;
Kweon, S. M.; Nahar, R.; Braig, M.; Park, E.; Kim, Y. M.; Hofmann, W. K.; Herzog, S.; Jumaa,
H.; Koeffler, H. P.; Yu, J. J.; Heisterkamp, N.; Graeber, T. G.; Wu, H.; Ye, B. H.; Melnick, A.;
Muschen, M. BCL6 enables Ph+ acute lymphoblastic leukaemia cells to survive BCRꢀABL1
kinase inhibition. Nature 2011, 473, 384ꢀ388.
(8)
Ahmad, K. F.; Melnick, A.; Lax, S.; Bouchard, D.; Liu, J.; Kiang, C. L.; Mayer, S.;
Takahashi, S.; Licht, J. D.; Prive, G. G. Mechanism of SMRT corepressor recruitment by the
BCL6 BTB domain. Mol. Cell 2003, 12, 1551ꢀ1564.
(9)
Cerchietti, L. C.; Ghetu, A. F.; Zhu, X.; Da Silva, G. F.; Zhong, S.; Matthews, M.;
Bunting, K. L.; Polo, J. M.; Fares, C.; Arrowsmith, C. H.; Yang, S. N.; Garcia, M.; Coop, A.;
ACS Paragon Plus Environment

Page 23 of 40
Journal of Medicinal Chemistry
1
2
3
4
Mackerell, A. D., Jr.; Prive, G. G.; Melnick, A. A smallꢀmolecule inhibitor of BCL6 kills
5
6
DLBCL cells in vitro and in vivo. Cancer Cell 2010, 17, 400ꢀ411.
7
8
9
(10) Cardenas, M. G.; Yu, W.; Beguelin, W.; Teater, M. R.; Geng, H.; Goldstein, R. L.;
Oswald, E.; Hatzi, K.; Yang, S. N.; Cohen, J.; Shaknovich, R.; Vanommeslaeghe, K.; Cheng, H.;
Liang, D.; Cho, H. J.; Abbott, J.; Tam, W.; Du, W.; Leonard, J. P.; Elemento, O.; Cerchietti, L.;
Cierpicki, T.; Xue, F.; MacKerell, A. D., Jr.; Melnick, A. M. Rationally designed BCL6
inhibitors target activated B cell diffuse large B cell lymphoma. J. Clin. Invest. 2016, 126, 3351ꢀ
3362.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(11) Cerchietti, L. C.; Yang, S. N.; Shaknovich, R.; Hatzi, K.; Polo, J. M.; Chadburn, A.;
Dowdy, S. F.; Melnick, A. A peptomimetic inhibitor of BCL6 with potent antilymphoma effects
in vitro and in vivo. Blood 2009, 113, 3397ꢀ3405.
(12) Evans, S. E.; Goult, B. T.; Fairall, L.; Jamieson, A. G.; Ko Ferrigno, P.; Ford, R.;
Schwabe, J. W.; Wagner, S. D. The ansamycin antibiotic, rifamycin SV, inhibits BCL6
transcriptional repression and forms a complex with the BCL6ꢀBTB/POZ domain. PLoS One
2014, 9, e90889.
(13) Sakamoto, K.; Sogabe, S.; Kamada, Y.; Sakai, N.; Asano, K.; Yoshimatsu, M.; Ida, K.;
Imaeda, Y.; Sakamoto, J. I. Discovery of highꢀaffinity BCL6ꢀbinding peptide and its structureꢀ
activity relationship. Biochem. Biophys. Res. Commun. 2017, 482, 310ꢀ316.
(14) Erlanson, D. A.; Fesik, S. W.; Hubbard, R. E.; Jahnke, W.; Jhoti, H. Twenty years on: the
impact of fragments on drug discovery. Nat. Rev. Drug Discovery 2016, 15, 605ꢀ619.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 40
1
2
3
4
5
6
7
8
9
(15) Ludlow, R. F.; Verdonk, M. L.; Saini, H. K.; Tickle, I. J.; Jhoti, H. Detection of
secondary binding sites in proteins using fragment screening. Proc. Natl. Acad. Sci. U. S. A.
2015, 112, 15910ꢀ15915.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(16) Jordan, J. B.; Whittington, D. A.; Bartberger, M. D.; Sickmier, E. A.; Chen, K.; Cheng,
Y.; Judd, T. Fragmentꢀlinking approach using (19)F NMR spectroscopy to obtain highly potent
and selective inhibitors of betaꢀsecretase. J. Med. Chem. 2016, 59, 3732ꢀ3749.
(17) Kohlmann, A.; Zech, S. G.; Li, F.; Zhou, T.; Squillace, R. M.; Commodore, L.;
Greenfield, M. T.; Lu, X.; Miller, D. P.; Huang, W. S.; Qi, J.; Thomas, R. M.; Wang, Y.; Zhang,
S.; Dodd, R.; Liu, S.; Xu, R.; Xu, Y.; Miret, J. J.; Rivera, V.; Clackson, T.; Shakespeare, W. C.;
Zhu, X.; Dalgarno, D. C. Fragment growing and linking lead to novel nanomolar lactate
dehydrogenase inhibitors. J. Med. Chem. 2013, 56, 1023ꢀ1040.
(18) Larsson, E. A.; Jansson, A.; Ng, F. M.; Then, S. W.; Panicker, R.; Liu, B.;
Sangthongpitag, K.; Pendharkar, V.; Tai, S. J.; Hill, J.; Dan, C.; Ho, S. Y.; Cheong, W. W.;
Poulsen, A.; Blanchard, S.; Lin, G. R.; Alam, J.; Keller, T. H.; Nordlund, P. Fragmentꢀbased
ligand design of novel potent inhibitors of tankyrases. J. Med. Chem. 2013, 56, 4497ꢀ4508.
(19) Wielens, J.; Headey, S. J.; Rhodes, D. I.; Mulder, R. J.; Dolezal, O.; Deadman, J. J.;
Newman, J.; Chalmers, D. K.; Parker, M. W.; Peat, T. S.; Scanlon, M. J. Parallel screening of
low molecular weight fragment libraries: do differences in methodology affect hit identification?
J. Biomol. Screening 2013, 18, 147ꢀ159.
ACS Paragon Plus Environment

Page 25 of 40
Journal of Medicinal Chemistry
1
2
3
4
(20) Giannetti, A. M.; Koch, B. D.; Browner, M. F. Surface plasmon resonance based assay
5
6
for the detection and characterization of promiscuous inhibitors. J. Med. Chem. 2008, 51, 574ꢀ
7
8
9
580.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(21) Retra, K.; Irth, H.; van MuijlwijkꢀKoezen, J. E. Surface plasmon resonance biosensor
analysis as a useful tool in FBDD. Drug Discovery Today: Technol. 2010, 7, e181ꢀe187.
(22) Chimenti, M. S.; Bulfer, S. L.; Neitz, R. J.; Renslo, A. R.; Jacobson, M. P.; James, T. L.;
Arkin, M. R.; Kelly, M. J. A fragmentꢀbased ligand screen against part of a large protein
machine: the ND1 domains of the AAA+ ATPase p97/VCP. J. Biomol. Screening 2015, 20, 788ꢀ
800.
(23) Giannetti, A. M.; Zheng, X.; Skelton, N. J.; Wang, W.; Bravo, B. J.; Bair, K. W.;
Baumeister, T.; Cheng, E.; Crocker, L.; Feng, Y.; GunznerꢀToste, J.; Ho, Y. C.; Hua, R.;
Liederer, B. M.; Liu, Y.; Ma, X.; O'Brien, T.; Oeh, J.; Sampath, D.; Shen, Y.; Wang, C.; Wang,
L.; Wu, H.; Xiao, Y.; Yuen, P. W.; Zak, M.; Zhao, G.; Zhao, Q.; Dragovich, P. S. Fragmentꢀ
based identification of amides derived from transꢀ2ꢀ(pyridinꢀ3ꢀyl)cyclopropanecarboxylic acid as
potent inhibitors of human nicotinamide phosphoribosyltransferase (NAMPT). J. Med. Chem.
2014, 57, 770ꢀ792.
(24) Hendrick, A. G.; Muller, I.; Willems, H.; Leonard, P. M.; Irving, S.; Davenport, R.; Ito,
T.; Reeves, J.; Wright, S.; Allen, V.; Wilkinson, S.; Heffron, H.; Bazin, R.; Turney, J.; Mitchell,
P. J. Identification and investigation of novel binding fragments in the fatty acid binding protein
6 (FABP6). J. Med. Chem. 2016, 59, 8094ꢀ8102.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 40
1
2
3
4
5
6
7
8
9
(25) Lund, B. A.; Christopeit, T.; Guttormsen, Y.; Bayer, A.; Leiros, H. K. Screening and
design of inhibitor scaffolds for the antibiotic resistance oxacillinaseꢀ48 (OXAꢀ48) through
surface plasmon resonance screening. J. Med. Chem. 2016, 59, 5542ꢀ5554.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(26) Davies, T. G.; Wixted, W. E.; Coyle, J. E.; GriffithsꢀJones, C.; Hearn, K.; McMenamin,
R.; Norton, D.; Rich, S. J.; Richardson, C.; Saxty, G.; Willems, H. M.; Woolford, A. J.; Cottom,
J. E.; Kou, J. P.; Yonchuk, J. G.; Feldser, H. G.; Sanchez, Y.; Foley, J. P.; Bolognese, B. J.;
Logan, G.; Podolin, P. L.; Yan, H.; Callahan, J. F.; Heightman, T. D.; Kerns, J. K. Monoacidic
inhibitors of the kelchꢀlike ECHꢀassociated protein 1: nuclear factor erythroid 2ꢀrelated factor 2
(KEAP1:NRF2) proteinꢀprotein interaction with high cell potency identified by fragmentꢀbased
discovery. J. Med. Chem. 2016, 59, 3991ꢀ4006.
(27) F. Bower, J.; Pannifer, A. Using fragmentꢀbased technologies to target proteinꢀprotein
interactions. Curr. Pharm. Des. 2012, 18, 4685ꢀ4696.
(28) Wartchow, C. A.; Podlaski, F.; Li, S.; Rowan, K.; Zhang, X.; Mark, D.; Huang, K. S.
Biosensorꢀbased small molecule fragment screening with biolayer interferometry. J. Comput.-
Aided Mol. Des. 2011, 25, 669ꢀ676.
(29) Yoshimura, C.; Miyafusa, T.; Tsumoto, K. Identification of smallꢀmolecule inhibitors of
the human S100Bꢀp53 interaction and evaluation of their activity in human melanoma cells.
Bioorg. Med. Chem. 2013, 21, 1109ꢀ1115.
(30) Whittaker, M. Picking up the pieces with FBDD or FADD: invest early for future
success. Drug Discovery Today 2009, 14, 623ꢀ624.
ACS Paragon Plus Environment

Page 27 of 40
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
(31) Varnes, J. G.; Geschwindner, S.; Holmquist, C. R.; Forst, J.; Wang, X.; Dekker, N.;
Scott, C. W.; Tian, G.; Wood, M. W.; Albert, J. S. Fragmentꢀassisted hit investigation involving
integrated HTS and fragment screening: Application to the identification of phosphodiesterase
10A (PDE10A) inhibitors. Bioorg. Med. Chem. Lett. 2016, 26, 197ꢀ202.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(32) Elinder, M.; Geitmann, M.; Gossas, T.; Kallblad, P.; Winquist, J.; Nordstrom, H.;
Hamalainen, M.; Danielson, U. H. Experimental validation of a fragment library for lead
discovery using SPR biosensor technology. J. Biomol. Screening 2011, 16, 15ꢀ25.
(33) Hamalainen, M. D.; Zhukov, A.; Ivarsson, M.; Fex, T.; Gottfries, J.; Karlsson, R.;
Bjorsne, M. Labelꢀfree primary screening and affinity ranking of fragment libraries using parallel
analysis of protein panels. J. Biomol. Screening 2008, 13, 202ꢀ209.
(34) Nordstrom, H.; Gossas, T.; Hamalainen, M.; Kallblad, P.; Nystrom, S.; Wallberg, H.;
Danielson, U. H. Identification of MMPꢀ12 inhibitors by using biosensorꢀbased screening of a
fragment library. J. Med. Chem. 2008, 51, 3449ꢀ3459.
(35) Navratilova, I.; Macdonald, G.; Robinson, C.; Hughes, S.; Mathias, J.; Phillips, C.; Cook,
A. Biosensorꢀbased approach to the identification of protein kinase ligands with dualꢀsite modes
of action. J. Biomol. Screening 2012, 17, 183ꢀ193.
(36) Nordin, H.; Jungnelius, M.; Karlsson, R.; Karlsson, O. P. Kinetic studies of small
molecule interactions with protein kinases using biosensor technology. Anal. Biochem. 2005,
340, 359ꢀ368.
(37) Stenlund, P.; FrostellꢀKarlsson, A.; Karlsson, O. P. Studies of small molecule interactions
with protein phosphatases using biosensor technology. Anal. Biochem. 2006, 353, 217ꢀ225.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 40
1
2
3
4
(38) Hopkins, A. L.; Groom, C. R.; Alex, A. Ligand efficiency: a useful metric for lead
5
6
selection. Drug Discovery Today 2004, 9, 430ꢀ431.
7
8
9
(39) Hopkins, A. L.; Keseru, G. M.; Leeson, P. D.; Rees, D. C.; Reynolds, C. H. The role of
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ligand efficiency metrics in drug discovery. Nat. Rev. Drug Discovery 2014, 13, 105ꢀ121.
(40) Perspicace, S.; Banner, D.; Benz, J.; Muller, F.; Schlatter, D.; Huber, W. Fragmentꢀbased
screening using surface plasmon resonance technology. J. Biomol. Screening 2009, 14, 337ꢀ349.
(41) Otwinowski, Z.; Minor, W. Processing of Xꢀray diffraction data collected in oscillation
mode. In Macromolecular Crystallography, part A, Carter, C. W. J.; Sweet, R. M., Eds.
Academic Press: New York, 1997; Vol. 276, pp. 307ꢀ326.
(42) Vagin, A.; Teplyakov, A. MOLREP: An automated program for molecular replacement.
J. Appl. Crystallogr. 1997, 30, 1022ꢀ1025.
(43) Collaborative Computational Project, N. The CCP4 suite: programs for protein
crystallography. Acta Crystallogr., Sect. D: Biol. Crystallogr. 1994, 50, 760ꢀ763.
(44) Murshudov, G. N.; Skubak, P.; Lebedev, A. A.; Pannu, N. S.; Steiner, R. A.; Nicholls, R.
A.; Winn, M. D.; Long, F.; Vagin, A. A. REFMAC5 for the refinement of macromolecular
crystal structures. Acta Crystallogr., Sect. D: Biol. Crystallogr. 2011, 67, 355ꢀ367.
(45) Emsley, P.; Lohkamp, B.; Scott, W. G.; Cowtan, K. Features and development of Coot.
Acta Crystallogr., Sect. D: Biol. Crystallogr. 2010, 66, 486ꢀ501.
(46) Chen, V. B.; Arendall, W. B., 3rd; Headd, J. J.; Keedy, D. A.; Immormino, R. M.;
Kapral, G. J.; Murray, L. W.; Richardson, J. S.; Richardson, D. C. MolProbity: allꢀatom structure
ACS Paragon Plus Environment

Page 29 of 40
Journal of Medicinal Chemistry
1
2
3
4
5
6
validation for macromolecular crystallography. Acta Crystallogr., Sect. D: Biol. Crystallogr.
2010, 66, 12ꢀ21.
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment