Homolog-Selective Degradation as a Strategy to Probe the Function of CDK6 in AML

Article abstract of DOI:10.1016/j.chembiol.2018.11.006

The design of selective small molecules is often stymied by similar ligand binding pockets. Here, we report BSJ-03-123, a phthalimide-based degrader that exploits protein-interface determinants to achieve proteome-wide selectivity for the degradation of cyclin-dependent kinase 6 (CDK6). Pharmacologic CDK6 degradation targets a selective dependency of acute myeloid leukemia cells, and transcriptomics and phosphoproteomics profiling of acute degradation of CDK6 enabled dynamic mapping of its immediate role in coordinating signaling and transcription. Brand et al. describe BSJ-03-123, a degrader with proteome-wide selectivity for CDK6. Specificity emerges from differential ternary complex formation with the E3 ligase CRL4^CRBN. BSJ-03-123 exploits a dependency of AML cells on CDK6, and rapid degradation allowed assaying the role of CDK6 in coordinating signaling and gene control in AML.

Full text of DOI:10.1016/j.chembiol.2018.11.006

Brief Communication
Homolog-Selective Degradation as a Strategy to
Probe the Function of CDK6 in AML
Graphical Abstract
Authors
Matthias Brand, Baishan Jiang,
Sophie Bauer, ..., Eric S. Fischer,
Nathanael S. Gray, Georg E. Winter
Selective CDK6 degradation through
differential ternary complex formation
CDK6
O
O
N
O
O
O
H N
O
BSJ-03-123
N
N
O
O
N
O
O
N
N
H N
N
O
N H
O
Correspondence
NH
O
O
N
HN
N
N
O
O
N
CRBN
E3
O
N
H
N
N
nathanael_gray@dfci.harvard.edu
(N.S.G.),
gwinter@cemm.at (G.E.W.)
O
N
Ub
O
O
O
O
CRBN
O
O
O
N
O
O
O
H N
O
N
N
O
CDK6
CDK4
N
N
H N
N
O
N
O
N H
O
CDK4
time
In Brief
Characterize CDK6
signaling network
Map CDK6-dependent
gene regulation
Exploit genetic
Brand et al. describe BSJ-03-123, a
degrader with proteome-wide selectivity
for CDK6. Specificity emerges from
CDK6 dependencies
Phosphoproteomics
RNA-seq
Proliferation assays
P
P
P
differential ternary complex formation
CRBN
P
P
P
P
with the E3 ligase CRL4
. BSJ-03-123
exploits a dependency of AML cells on
CDK6, and rapid degradation allowed
assaying the role of CDK6 in coordinating
signaling and gene control in AML.
fold change
time
Highlights
d
Development of a CDK6-selective small-molecule degrader
d
Mechanistically understood selectivity via differential ternary
complex formation
d
d
Profiling of consequences of CDK6 degradation on signaling
and gene regulation
Precise exploitation of genetic dependencies through
homolog-selective degradation
Brand et al., 2019, Cell Chemical Biology 26, 1–7
February 21, 2019 ª 2018 Elsevier Ltd.
https://doi.org/10.1016/j.chembiol.2018.11.006

Please cite this article in press as: Brand et al., Homolog-Selective Degradation as a Strategy to Probe the Function of CDK6 in AML, Cell Chemical
Biology (2018), https://doi.org/10.1016/j.chembiol.2018.11.006
Cell Chemical Biology
Brief Communication
Homolog-Selective Degradation as a Strategy
to Probe the Function of CDK6 in AML
Matthias Brand,1,4 Baishan Jiang,
2,3,4
Sophie Bauer, Katherine A. Donovan, Yanke Liang, Eric S. Wang,
1
2,3
2,3
2,3
2
,3
2
2,3
2,3
1
2,3
Rados1aw P. Nowak, Jingting C. Yuan, Tinghu Zhang, Nicholas Kwiatkowski, Andr e´ C. M u€ ller, Eric S. Fischer,
Nathanael S. Gray, * and Georg E. Winter
CeMM Research Center for Molecular Medicine of the Austrian Academy of Science, Vienna, Austria
2Department of Cancer Biology, Dana-Farber Cancer Institute, Boston, USA
3Department of Biological Chemistry and Molecular Pharmacology, Harvard Medical, Boston, USA
4These authors contributed equally
2,3,
1,5,
*
1
5Lead Contact
*Correspondence: nathanael_gray@dfci.harvard.edu (N.S.G.), gwinter@cemm.at (G.E.W.)
https://doi.org/10.1016/j.chembiol.2018.11.006
SUMMARY
induced targeted protein degradation strategy based on hetero-
bifunctional molecules that induce molecular proximity between
CRBN
The design of selective small molecules is often sty- a protein of interest (POI) and the CRL4
E3 ligase complex
mied by similar ligand binding pockets. Here, we (Lu et al., 2015; Winter et al., 2015; Zengerle et al., 2015). Inter-
estingly, studies of degraders derived from multi-targeted war-
heads revealed that only a subset of targets were efficiently
report BSJ-03-123, a phthalimide-based degrader
that exploits protein-interface determinants to
achieve proteome-wide selectivity for the degra-
dation of cyclin-dependent kinase 6 (CDK6). Phar-
macologic CDK6 degradation targets a selective
dependency of acute myeloid leukemia cells, and
transcriptomics and phosphoproteomics profiling
of acute degradation of CDK6 enabled dynamic
mapping of its immediate role in coordinating
signaling and transcription.
degraded, indicating the possibility of engineering selectivity
into poly-pharmacologic scaffolds (Bondeson et al., 2018;
Gadd et al., 2017; Huang et al., 2018; Olson et al., 2018; Winter
et al., 2015). Along those lines, it was demonstrated that selec-
tive degradation can be achieved by exploiting amino acid differ-
ences in the POI-E3 interface and thus differential ternary com-
plex formation (Gadd et al., 2017; Nowak et al., 2018).
Here, we explored the feasibility of homolog-selective degra-
dation of CDK6 versus CDK4 using ligands that bind the ATP
sites of both enzymes. We describe the discovery and character-
ization of BSJ-03-123 (BSJ), a selective degrader of CDK6 that
uniquely enables rapid pharmacological interrogation of CDK6-
dependent functions. Coupling acute CDK6 degradation with
phosphoproteomic and transcriptomic profiling allowed us to
establish a systems-level understanding of the sensitivity of
acute myeloid leukemia (AML) cell lines to selective CDK6 abla-
tion and to chart a network of CDK6-dependent signaling events.
INTRODUCTION
Cyclin-dependent kinases (CDKs) orchestrate fundamental
cellular processes such as cell cycle and transcription. CDKs
are serine/threonine kinases and their activity is dependent on
association with cyclins, enabling temporal control over enzy-
matic function. CDK4 and CDK6 control G1-S transition by asso-
ciating with D-type cyclins to phosphorylate the tumor suppres- RESULTS
sor retinoblastoma (Rb). This relieves Rb-mediated repression
of E2F transcription factors (TFs) to trigger gene expression Development of a Homolog-Selective CDK6 Degrader
required for S-phase entry. While CDK4 and CDK6 were long Toward the development of CDK4/6 degraders, we chose palbo-
perceived as redundant, recent studies have identified homo- ciclib (palbo) as a starting point given its high selectivity and po-
log-specific functions. In particular, CDK6 emerged as a regula- tency (Fry et al., 2004). Available crystal structures of palbo with
tory hub connecting cell cycle with metabolism (Wang et al., CDK6 (PDB: 5L2I) highlighted the solvent-exposed piperazine
2
017) and transcription (Tigan et al., 2016). In several contexts, moiety as an ideal site for linker conjugation (Chen et al.,
the transcriptional role of CDK6 is independent of its kinase func- 2016). We installed a 3-polyethylene glycol linker conjugated to
tion, but relies on molecular scaffolding (Kollmann et al., 2013; Ti- pomalidomide, resulting in the compound YKL-06-102 (YKL)
gan et al., 2016).
(Figure 1A). Biochemical characterization indicated that phthali-
ATP-competitive CDK4/6 inhibitors have shown significant mide conjugation results in a comparable, albeit weakly reduced
clinical activity, leading to approval of three drugs for breast binding affinity to both CDK4 and CDK6 (Figure S1A; Table S1).
cancer (O’Leary et al., 2016). However, none of these inhibitors Moreover, YKL bound to recombinant CRBN-DDB1 with compa-
can discriminate between CDK4 and CDK6, as they share 94% rable affinity as lenalidomide (Table S1). Unexpectedly, cellular
amino acid sequence identity in their ATP binding pockets. In treatment with YKL led to a selective destabilization of CDK6
addition, current inhibitors cannot disrupt scaffolding functions (Figures S1B–S1D). To further survey degrader selectivity, we
(
Kollmann et al., 2013). To address this, we turned to a ligand- employed global proteomics following cellular treatment. While
Cell Chemical Biology 26, 1–7, February 21, 2019 ª 2018 Elsevier Ltd.
1

Please cite this article in press as: Brand et al., Homolog-Selective Degradation as a Strategy to Probe the Function of CDK6 in AML, Cell Chemical
Biology (2018), https://doi.org/10.1016/j.chembiol.2018.11.006
A
D
BSJ-03-123 (nM)
E
BSJ (200 nM)
F
O
H N
H
N
O
N
N
O
O
N
O
YKL-06-102
BSJ
O
N
N
O
N
CDK6
CDK4
H3
CDK6
CDK4
H3
(200 nM)
-
+
-
+
O
O
N
N
H
O
N
H
CDK6
H3
B
G
1
00
100
80
60
40
20
0
YKL-06-102
BSJ-03-123
8
0
0
0
0
CDK4
6
IKZF1
CDK6
4
2
CDK6
IKZF3
0
adj.
adj.
BSJ-bump
C
H
I
O
N
O
O
NH
O
H
H
N
O
N
N
N
N
O
O
CDK6
CDK4
H3
O
BSJ-03-123
N
N
N
O
N
BSJ-bump
N
O
N
O
N
N
N
H
H
N
O
N
N
O
N
O
O
O
O
O
O
O
O
O
Figure 1. Development of a Homolog-Selective Degrader of CDK6
(
(
A) Structure of YKL-06-102.
B) Quantification of 5,945 proteins following treatment with YKL-06-102 (500 nM, 5 hr). False discovery rate (FDR)-adjusted p values. Kinases inhibited by
palbociclib in vitro are color-coded by magnitude of inhibition.
(
(
(
(
(
(
(
C) Structure of BSJ-03-123 (BSJ).
D) Immunoblot for CDK4, CDK6, and histone 3 after 4 hr BSJ treatment.
E) As in (D), but time-resolved at 200 nM BSJ.
F) Immunoblot for CDK6 and histone 3. 200 nM BSJ, 4 hr, in CRBN-deficient and wild-type MV4-11.
G) Quantification of 5,995 proteins following BSJ treatment (100 nM, 2 hr). FDR-adjusted p values. Kinase labeling as in (B).
H) Structure of BSJ-bump.
I) As in (D), but for BSJ-bump.
See also Figure S1 and Tables S1 and S2.
recapitulating the selective degradation of CDK6, we detected able to degrade CDK4 or CDK6, thus serving as excellent nega-
ꢀ
6
significant destabilization of IKZF1 and IKZF3 (p = 6.19 3 10
Figure 1B), suggesting warhead-independent CRBN substrate
modulation (An et al., 2017; Gandhi et al., 2014; Kronke et al., BSJ Triggers Homolog-Selective Degradation via
014; Lu et al., 2014). To remove IKZF1/3 activity of YKL, we de- Differential Ternary Complex Formation
)
tive control with matched physiochemical properties (Figure 1I).
(
2
signed BSJ by employing a phenoxyacetamide linker without Next, we sought to understand the mechanism of CDK6 selec-
altering the initial linker length (Figure 1C). Measurements of tivity. In vitro kinase assays confirmed comparable affinity to
in vitro binding to recombinant kinase domains (KINOMEScan) both kinases (Figure S1A; Table S1). Similar cellular thermal sta-
confirmed that phthalimide conjugation preserves the narrow bilization of CDK4/6 further suggested that selectivity does not
selectivity profile (Figure S1G; Table S2). BSJ induced CRBN- emerge from differential cellular target engagement (Figure 2A).
dependent, potent, fast, and homolog-selective degradation of We thus hypothesized that selectivity might stem from differen-
CDK6 (Figures 1D–1F and S1E). In proteome-wide selectivity tial ternary complex formation. To monitor tripartite assembly
studies, CDK6 was the only destabilized protein (Figures 1G in real time in intact cells, we designed a luciferase complemen-
and S1F).
Based on BSJ, we furthermore synthesized an N-methylated expressed C-terminal CDK4/6 LgBit fusions along with an N-ter-
tation assay based on NanoBiT technology (Figure 2B). We
CRBNꢀ/ꢀ
glutarimide analog (BSJ-bump) incapable of CRBN binding minal SmBit-CRBN fusion in 293T cells. BSJ induced
in vitro (Figure 1H; Table S1). Accordingly, BSJ-bump was un- rapid, dose-dependent ternary complex formation with CDK6
2
Cell Chemical Biology 26, 1–7, February 21, 2019

Please cite this article in press as: Brand et al., Homolog-Selective Degradation as a Strategy to Probe the Function of CDK6 in AML, Cell Chemical
Biology (2018), https://doi.org/10.1016/j.chembiol.2018.11.006
A
C
6
5
6
5
CDK4:CRBN
CDK6:CRBN
4
3
2
4
2
0 µM
CDK4
3
2
1
CDK6
1
Temperature (°C)
4
6
49
52
55
0
2
4
6
8
10
0
2
4
6
8
10
time (min)
time (min)
DMSO
BSJ-bump
BSJ (500 nM)
BSJ (2 uM)
D
B
CDK6:CRBN
4
4
.5
.0
DMSO
SmBit
CRBN
palbo
len
SmBit
LgBit
3.5
3.0
LgBit
+
BSJ-03-123
BSJ+DMSO competition
BSJ+palbo competition
BSJ+len competition
2
.5
CRBN
HN
N
O
O
2.0
O
N
N
N
O
N
O
NH
O
N
N
HN
O
O
1
1
.5
.0
CDK4/6
CDK4/6
0
2
4
6
8
10
time (min)
Figure 2. BSJ Selectivity Is Explained by Differential Ternary Complex Formation
–
A) Immunoblot for CDK4 and CDK6 after thermal shift assay in MV4-11 CRBN cells after cellular treatment with DMSO, palbociclib, or BSJ (both at 20 mM).
/–
(
(
(
B) Schematic representation of the luciferase complementation assay.
C) Measurement of CDK4:CRBN and CDK6:CRBN binding via NanoBiT assay. Averages of five replicates are plotted, shaded areas represent 95% confidence
intervals.
D) Measurement of CDK6:CRBN complex formation after 1 hr pre-treatment with 20 mM palbociclib or lenalidomide. Statistics as in (C).
(
and CRBN, but not with CDK4 and CRBN (Figure 2C). Treatment pected, dual CDK4/6 inhibition severely impaired proliferation of
with BSJ-bump failed to induce CDK6:CRBN interactions (Fig- CDK4-dependent cell lines, whereas selective degradation of
ure 2C). Ternary complex formation was prevented by blocking CDK6 did not trigger a comparable effect (Figures 3B and
binding sites on CDK6 or CRBN via pretreatment with palbo or S2E). In contrast, BSJ caused a pronounced anti-proliferative ef-
lenalidomide (Figure 2D). We thus concluded that BSJ exploits fect in CDK6-dependent AML cell lines by inducing a G1 cell-cy-
structural differences between CDK4 and CDK6 to achieve ho- cle arrest without a measurable increase in apoptosis (Figures
molog-selective degradation of CDK6 via differential ternary 3C–3E). Of note, this effect exceeded the anti-proliferative con-
complex formation.
sequences of cellular treatment with BSJ-bump (Figures 3C and
S2F). The observed benefit of CDK6 degradation over CDK6 in-
hibition could be due to the disruption of additional, kinase activ-
ity-independent molecular mechanisms (Kollmann et al., 2013),
BSJ Can Exploit Homolog-Selective Dependency
on CDK6
To identify a cellular system that is disproportionally dependent or due to pharmacologic advantages of degraders, such as cat-
on CDK6 over CDK4, we turned to genome-scale CRISPR/Cas9 alytic target turnover. To address this, we compared BSJ with
screens in 342 cancer cell lines (Meyers et al., 2017). We identi- the clinically approved palbo in CDK4-deficient MV4-11 cells
fied a pronounced enrichment of AML cell lines among the most to enable a CDK6-centric readout. In this experimental setup,
CDK6-addicted models (Figure 3A). None of these cell lines selective degradation of CDK6 was not superior to enzymatic in-
showed a comparable dependency on CDK4 (Figure S2A). Inter- hibition (Figure 3F). While interpretation of this comparison is
section with gene expression data did not unveil a correlation be- non-trivial given the physicochemical differences between the
tween essentiality and mRNA transcript levels (Figures 3A and molecules, it suggests that the dependency of AML cells on
S2A) (Klijn et al., 2015). Accordingly, CRISPR/Cas9-mediated CDK6 is mostly dependent on its kinase function.
genetic depletion of CDK4 in the AML cell line MV4-11 revealed
largely preserved growth kinetics and cell-cycle distribution, Impairment of Selective Regulatory Networks after
supporting a dispensable role of CDK4 in the proliferation of CDK6 Degradation
AML cell lines (Figures S2B–S2D).
We next investigated drug impact on the known CDK4/6 Rb
Given their lack of homolog-selectivity, CDK4/6 inhibitors are phosphorylation site S780. Both BSJ and palbo treatment
incapable of exploiting such fine-tuned genetic dependencies. reduced levels of p-S780 (Figure 4A). Notably, BSJ-bump treat-
Concomitant inhibition of both homologs is thus expected to limit ment failed to affect phosphorylation of Rb, indicating that, at the
the achievable therapeutic window. Therefore, we wanted to test assayed concentration, enzymatic CDK4/6 inhibition is insignifi-
if BSJ could target the genetic dependency of AML cells on cant (Figure 4A). Measurable impact of BSJ-bump was only
CDK6, while sparing CDK4-dependent cancer cells lines. As ex- observed upon dose escalation (Figure S2G). In line with this,
Cell Chemical Biology 26, 1–7, February 21, 2019
3

Please cite this article in press as: Brand et al., Homolog-Selective Degradation as a Strategy to Probe the Function of CDK6 in AML, Cell Chemical
Biology (2018), https://doi.org/10.1016/j.chembiol.2018.11.006
A
B
C
E
D
F
Figure 3. BSJ Exploits Homolog-Selective Dependencies
A) Bottom: waterfall plot of 391 cell lines ranked by CDK6 dependency as determined in genome-wide CRISPR/Cas9 screens. CERES essentiality score is
(
normalized for copy-number variation and scaled by setting the median of pan-essential genes to ꢀ1. Top: mRNA levels of CDK4, CDK6, and D-type cyclins.
(
(
(
(
B) Colony-formation assays (12 days, refreshing the treatment every 2 days).
C) Growth curves of AML cell lines treated with 200 nM BSJ, BSJ-bump, or DMSO. Cells were counted and treatment refreshed every 2 days.
D) Cell cycle after treatment with 200 nM BSJ or BSJ-bump for 24 hr.
E) Percentage of apoptotic cells after 24 hr treatment with 200 nM palbociclib, BSJ, BSJ-bump, or DMSO (Caspase-Glo 3/7 assay). BET protein degrader dBET6
served as positive control.
F) Growth curves of CDK4-deficient MV4-11 treated with 200 nM BSJ, palbociclib, or DMSO.
See also Figure S2. Data presented in (C–F): mean ±SD, triplicate analysis.
(
ꢀ5
the effect of BSJ treatment on Rb phosphorylation is rescued 10 ) overlap with previously reported CDK4/6 substrates based
upon CRBN knockout (Figure S2H). on in vitro assays (Anders et al., 2011), but also uncovered unde-
Given the plethora of functions assigned to CDK6, we globally scribed substrates such as the splicing factor SRRM2 and the
extended our analysis to assay how acute and selective CDK6 RSF chromatin-remodeling component RSF1 (Figures S3D
disruption affects cellular signaling and gene activity in AML. and S3E).
Given the limited activity of BSJ-bump, we compared selective
In parallel, we compared the transcriptional response elicited
degradation of CDK6 via BSJ with catalytic CDK4/6 inhibition by palbo and BSJ. This allowed us to determine if CDK6 orches-
by palbo. We coupled acute drug treatment (2 hr, 200 nM trates transcription independent of its kinase domain. As ex-
each) to quantitative (Ser/Thr) phosphoproteomics to focus on pected, transcriptional impact of YKL significantly differed from
direct drug effects. We quantified a total of 22,804 phosphopep- BSJ/palbo owing to off-target effects on IKZF1 and IKZF3. In
tides in at least one technical replicate. In general, drug impact contrast, enzymatic CDK4/6 inhibition and CDK6 degradation
2
was modest, with 197 deregulated phosphopeptides after treat- yielded a very similar and highly correlated (R = 0.9375) tran-
ment with BSJ (44 up, 153 down), and 191 deregulated (57 up, scriptional response, arguing that the transcriptional role of
1
34 down) phosphopeptides following palbo treatment (Figures CDK6 in AML is mostly limited to its kinase function (Figures
B and S3A). Changed sites were enriched for CDK motifs based S4A–S4D). To further explore the functional implications of im-
4
on kinase-substrate enrichment analysis (Figures 4B and S3B), mediate CDK6 degradation, we performed gene ontology-term
validating that short treatments biased for upstream events analysis of transcriptional changes elicited by CDK6 degradation
(Casado et al., 2013). Proteins with altered phosphosites were and CDK4/6 inhibition (Figures 4D and S4E). Again, transcrip-
enriched for known regulators of cell-cycle progression (Fig- tional consequences were similar, indicating that both converge
ure S3C). Moreover, our data indicate a significant (p = 4.49 3 on known pathways such as cell-cycle regulation, but also
4
Cell Chemical Biology 26, 1–7, February 21, 2019

Please cite this article in press as: Brand et al., Homolog-Selective Degradation as a Strategy to Probe the Function of CDK6 in AML, Cell Chemical
Biology (2018), https://doi.org/10.1016/j.chembiol.2018.11.006
A
B
C
BSJ PAL bump
8h
24h
log2
fold change
YKL BSJ PAL
-2
0
2
p-RB S780
total RB
CDK6
BSJ-03-123
Palbociclib
CDK4
BSJ-03-123
Palbociclib
H3
TP/SP motif
log2 fold change
-4.5
0
3
E
D
PDAP1
TRA2B
DNA
repair
DNA replication
DNA damage
response
AHCTF1
MYL12B
EIF5B
RPS6
CCNT2
TAF7
chromosome
organization
GTF2B
CEBPZ
RNF2
RFX7
SIRT6 GTF2I
KAT2A
GTF3C2
NFYB
NFYA
ABCF3
phospho-
FLI1
KAT2B
RELA
SRRM 2T MPO
IRF1
CEBPD ETS1
MKL1
no
SP/TP SP/TP
CDC20
CHD4 KDM1A
NELFE
BCL11A
STAT1
SP4
-
log10 q-value
proteomics hit
PML
E2F1
NRF1
chromatin
organization
RB1
0
>20
ENCODE TR
PAX5
enrichment
FOXM1
RBL2
identified
SMARCC1
ATAD2
IRF3
cell cycle
regulation
phospho-peptides
SMARCB1
NCOR1
0
8
MYBL2
2
10
18
percentage of
dysregulated targets
RBL1
LDB1
RSF1
1
%
3% 8%
STAT2
Figure 4. An Integrated View of the Effects of Acute CDK6 Degradation on Cellular Signaling and Transcription
(
(
A) Immunoblot for p-RB S780, RB, CDK4, CDK6, and histone 3 after treatment with 200 nM BSJ, palbociclib, or BSJ-bump.
B) Global phosphoproteomics. Heatmap depicting fold changes in peptide phosphorylation after BSJ or palbociclib treatment (200 nM, 2 hr) compared with
DMSO for 305 differentially phosphorylated peptides. Hits (log2 fold change [FC] > 0.5 or < ꢀ0.5, adjusted p value < 0.05) and peptides phosphorylated at
canonical SP/TP CDK phosphorylation motifs are annotated.
(
(
(
C) Heatmap of DMSO-normalized fold changes in gene expression after BSJ, YKL, or palbociclib treatment (200 nM, 6 hr) for 993 significantly deregulated genes
log2 FC > 1.5 or < ꢀ1.5, adjusted p value < 0.05).
D) Functional network of BSJ treatment. Nodes represent gene ontology (GO) terms enriched among genes that are differentially expressed upon treatment,
scaled by magnitude, and color coded by significance of enrichment. Edges represent parent-child relationships of GO terms.
E) Molecular network of BSJ treatment. Hits identified via global phosphoproteomics were mapped on a protein-protein interaction network and expanded to
(
include first-order neighbors limited to ENCODE transcriptional regulators. Node shape distinguishes transcriptional regulators (TR) (diamonds) from phos-
phoproteomics hits (round). Node color represents the number of quantified phosphopeptides. Diamonds are scaled proportional to percentage of dysregulated
TR target genes upon treatment. Proteins phosphorylated at CDK consensus motif (SP/TP) are annotated by edge color. Edges represent physical interaction
between proteins.
See also Figures S3 and S4.
influence processes such as DNA replication, the DNA damage scaled based on the percentage of dynamically regulated target
response, or chromatin organization.
genes (Figure 4E). This integrated network allowed identification
To derive a systems-level understanding of the function of of known, consensus downstream signaling axes such as Rb-
CDK6, we condensed data from transcriptional and phosphopro- E2F1, but also nodes such as BCL11A and NCOR2, previously
teomic profiling into a network centered around proteins with not linked to CDK6. Network topologies are largely overlapping
altered phosphopeptides after immediate CDK6 degradation between BSJ or palbo treatment (Figure S3F), again supporting
(
Figure 4E). Proteins were differentiated based on presence of a the notion that, in AML, CDK6 coordinates signaling and gene
consensus CDK phosphorylation motif (SP/TP) in the detected control predominantly via its kinase function.
phosphopeptides. Next, we expanded this network by adding
ENCODE transcriptional regulators with established protein-pro- DISCUSSION
tein interactions to members from the experimentally derived
network. Network expansion was limited to factors with available The development of highly selective small molecules is a long-
genome-wide binding data. Dynamic transcriptional changes af- standing challenge in ligand discovery given the structural simi-
ter CDK6 depletion were calculated and altered target genes larities of substrate or cofactor binding sites. This is a particular
were assigned to the network-resident TFs, which were visually concern with ATP-competitive kinase inhibitors where a lack of
Cell Chemical Biology 26, 1–7, February 21, 2019
5

Please cite this article in press as: Brand et al., Homolog-Selective Degradation as a Strategy to Probe the Function of CDK6 in AML, Cell Chemical
Biology (2018), https://doi.org/10.1016/j.chembiol.2018.11.006
selectivity can limit the achievable therapeutic window. This
hinders pharmacologic exploitation of genetically defined depen-
dencies for therapeutic indications. Here, we present BSJ, a ho-
molog-selective CDK6 degrader. BSJ features proteome-wide
selectivity for CDK6 via differential ternary complex formation.
Selectivity of BSJ enables exploitation of genetic dependencies
beyond a resolution achievable with dual CDK4/6 inhibitors. In
particular, we show that BSJ is capable of exploiting a homo-
log-selective dependency of AML cells on CDK6, thus outlining
the potential for selective CDK6 degraders for further transla-
tional investigation, conceivably at reduced overall toxicity.
Degradation of CDK6 is fast and potent, allowing us to map global
consequences of acute CDK6 disruption on downstream
signaling networks and transcriptional programs. While we
cannot rule out that, at saturating concentrations, BSJ also cata-
lytically inhibits CDK4, no inhibition was measured at the assayed
concentration. Comparative profiling of CDK6 degradation and
CDK4/6 inhibition suggest that, in AML, CDK6 integrates
signaling and gene activity predominantly via its kinase activity.
Our analysis uncovered several signaling nodes and transcrip-
tional hubs previously not linked to CDK6, such as BCL11A and
NCOR2, and future research will be necessary to explore the
functional relevance of these pathways in AML and beyond.
CDK6 features a prominent role in other malignancies as well
as hematopoietic and leukemic stem cells. Selective degradation
of CDK6 will facilitate differentiating relevant molecular mecha-
nisms and further untangle kinase-dependent and -independent
functions. Future medicinal chemistry efforts will be necessary
to expand on the presented concept and to develop a toolbox
of selective degraders to investigate the role of protein kinases
at unprecedented precision and kinetic resolution.
d METHOD DETAILS
B Knockout Generation
B Proliferation Assays
B Cell Cycle and Apoptosis
B In Vitro CRBN Binding Assay
B Immunoblotting
B Cellular Thermal Shift Assay
B NanoBiTꢀ Assay
B Expression Proteomics BSJ-03-123
B Expression Proteomics YKL-06-102
B Phosphoproteomics
B RNA Sequencing
B Chemical Synthesis of BSJ-03-123, BSJ-03-190/BSJ-
bump, YKL-06-102
d QUANTIFICATION AND STATISTICAL ANALYSIS
B Expression Proteomics BSJ-03-123
B Expression Proteomics YKL-06-102
B Phosphoproteomics
B RNA Sequencing
d DATA AND SOFTWARE AVAILABILITY
SUPPLEMENTAL INFORMATION
Supplemental Information includes four figures, two tables, and one data
file and can be found with this article online at https://doi.org/10.1016/j.
chembiol.2018.11.006.
ACKNOWLEDGMENTS
We thank all members of the Gray and Winter labs. CeMM and the Winter lab
are supported by the Austrian Academy of Sciences. This work was supported
by an FWF grant to G.E.W. (project number FWF30271), and a Damon Runyon
Cancer Research Fellowship DRG-2270-16 (to E.S.W.). The Gray lab is sup-
ported by NIH grant CA154303-06, and the Fischer lab by NIH NCI R01
CA214608-02. E.S.F. is a Damon Runyon-Rachleff Innovator supported in
part by the Damon Runyon Cancer Research Foundation (DRR-50-18).
SIGNIFICANCE
Traditionally, drug discovery efforts focus on the develop-
ment of small-molecule inhibitors that block accessible hy-
drophobic pockets such as substrate or cofactor binding
sites. However, due to the often high sequence conservation
of these pockets, the design of selective inhibitors remains a
continuous challenge. Here, we report a degrader that ex-
ploits structural differences outside of the ligand binding
pocket to induce differential E3 ligase recruitment and
ensuing degradation of CDK6, but not CDK4. We employed
this exquisitely selective probe to dissect CDK6-dependent
signaling and gene expression in AML, and showed that it
allows unprecedented exploitation of genetically defined
dependencies. This work showcases a framework for engi-
neering selectivity into small-molecule probes and paves
the way for the development of tools to functionally under-
stand proteins at high kinetic resolution.
AUTHOR CONTRIBUTIONS
M.B. performed immunoblots, nluc-complementation assays, and prolifera-
tion assays. He also prepared samples for RNA sequencing and phosphopro-
teomics and expression proteomics, and performed the associated bio-
informatic analysis. B.J. synthesized BSJ-03-123 and BSJ-bump. S.B.
performed immunoblots and proliferation assays. K.A.D. performed, analyzed,
and compared expression proteomics. Y.L. synthesized YKL-06-102. E.S.W.
performed immunoblots and gave experimental advice. R.P.N. and J.C.Y.
conducted recombinant CRBN binding assays. T.Z. and N.K. gave experi-
mental and analytic advice. A.C.M. performed proteomics experiments.
E.S.F. analyzed data and supervised aspects of the project. N.S.G. and
G.E.W. wrote the manuscript, analyzed data, and have the overall project
responsibility.
DECLARATION OF INTERESTS
E.S.F. is a member of the scientific advisory board of C4 Therapeutics and is a
consultant to Novartis Pharmaceuticals. E.S.F. receives research funding from
Novartis Pharmaceuticals and Astellas Pharma not related to this work. N.S.G.
is equity holder and scientific advisor for Syros, Gatekeeper, Soltego, C4,
Petra, and Aduro companies. N.S.G., B.J., T.Z., N.K., B.N., and E.S.F. are in-
ventors on a patent covering CDK6 degraders owned by Dana-Farber.
STAR+METHODS
Detailed methods are provided in the online version of this paper
and include the following:
Received: July 23, 2018
d KEY RESOURCES TABLE
Revised: October 11, 2018
Accepted: November 6, 2018
Published: December 27, 2018
d CONTACT FOR REAGENT AND RESOURCE SHARING
d EXPERIMENTAL MODEL AND SUBJECT DETAILS
6
Cell Chemical Biology 26, 1–7, February 21, 2019

Please cite this article in press as: Brand et al., Homolog-Selective Degradation as a Strategy to Probe the Function of CDK6 in AML, Cell Chemical
Biology (2018), https://doi.org/10.1016/j.chembiol.2018.11.006
REFERENCES
Kollmann, K., Heller, G., Schneckenleithner, C., Warsch, W., Scheicher, R.,
Ott, R.G., Schafer, M., Fajmann, S., Schlederer, M., Schiefer, A.I., et al.
(2013). A kinase-independent function of CDK6 links the cell cycle to tumor
angiogenesis. Cancer Cell 24, 167–181.
Alanis-Lobato, G., Andrade-Navarro, M.A., and Schaefer, M.H. (2017). HIPPIE
v2.0: enhancing meaningfulness and reliability of protein-protein interaction
networks. Nucleic Acids Res. 45, D408–D414.
Kronke, J., Udeshi, N.D., Narla, A., Grauman, P., Hurst, S.N., McConkey, M.,
Svinkina, T., Heckl, D., Comer, E., Li, X., et al. (2014). Lenalidomide causes se-
lective degradation of IKZF1 and IKZF3 in multiple myeloma cells. Science
343, 301–305.
An, J., Ponthier, C.M., Sack, R., Seebacher, J., Stadler, M.B., Donovan, K.A.,
and Fischer, E.S. (2017). pSILAC mass spectrometry reveals ZFP91 as IMiD-
dependent substrate of the CRL4(CRBN) ubiquitin ligase. Nat. Commun.
8
, 15398.
Lu, G., Middleton, R.E., Sun, H., Naniong, M., Ott, C.J., Mitsiades, C.S., Wong,
K.K., Bradner, J.E., and Kaelin, W.G., Jr. (2014). The myeloma drug lenalido-
mide promotes the cereblon-dependent destruction of Ikaros proteins.
Science 343, 305–309.
Anders, L., Ke, N., Hydbring, P., Choi, Y.J., Widlund, H.R., Chick, J.M., Zhai,
H., Vidal, M., Gygi, S.P., Braun, P., et al. (2011). A systematic screen for
CDK4/6 substrates links FOXM1 phosphorylation to senescence suppression
in cancer cells. Cancer Cell 20, 620–634.
Lu, J., Qian, Y., Altieri, M., Dong, H., Wang, J., Raina, K., Hines, J., Winkler,
J.D., Crew, A.P., Coleman, K., et al. (2015). Hijacking the E3 ubiquitin ligase
cereblon to efficiently target BRD4. Chem. Biol. 22, 755–763.
Bondeson, D.P., Smith, B.E., Burslem, G.M., Buhimschi, A.D., Hines, J.,
Jaime-Figueroa, S., Wang, J., Hamman, B.D., Ishchenko, A., and Crews,
C.M. (2018). Lessons in PROTAC design from selective degradation with a
promiscuous warhead. Cell Chem. Biol. 25, 78–87.e5.
Meyers, R.M., Bryan, J.G., McFarland, J.M., Weir, B.A., Sizemore, A.E., Xu, H.,
Dharia, N.V., Montgomery, P.G., Cowley, G.S., Pantel, S., et al. (2017).
Computational correction of copy number effect improves specificity of
CRISPR-Cas9 essentiality screens in cancer cells. Nat. Genet. 49, 1779–1784.
Casado, P., Rodriguez-Prados, J.C., Cosulich, S.C., Guichard, S.,
Vanhaesebroeck, B., Joel, S., and Cutillas, P.R. (2013). Kinase-substrate
enrichment analysis provides insights into the heterogeneity of signaling
pathway activation in leukemia cells. Sci. Signal. 6, rs6.
Nowak, R.P., DeAngelo, S.L., Buckley, D., He, Z., Donovan, K.A., An, J.,
Safaee, N., Jedrychowski, M.P., Ponthier, C.M., Ishoey, M., et al. (2018).
Plasticity in binding confers selectivity in ligand-induced protein degradation.
Nat. Chem. Biol. 14, 706–714.
Chen, P., Lee, N.V., Hu, W., Xu, M., Ferre, R.A., Lam, H., Bergqvist, S.,
Solowiej, J., Diehl, W., He, Y.A., et al. (2016). Spectrum and degree of CDK
drug interactions predicts clinical performance. Mol. Cancer Ther. 15,
O’Leary, B., Finn, R.S., and Turner, N.C. (2016). Treating cancer with selective
2273–2281.
CDK4/6 inhibitors. Nat. Rev. Clin. Oncol. 13, 417–430.
Fry, D.W., Harvey, P.J., Keller, P.R., Elliott, W.L., Meade, M., Trachet, E.,
Albassam, M., Zheng, X., Leopold, W.R., Pryer, N.K., et al. (2004). Specific in-
hibition of cyclin-dependent kinase 4/6 by PD 0332991 and associated anti-
tumor activity in human tumor xenografts. Mol. Cancer Ther. 3, 1427–1438.
Olson, C.M., Jiang, B., Erb, M.A., Liang, Y., Doctor, Z.M., Zhang, Z., Zhang, T.,
Kwiatkowski, N., Boukhali, M., Green, J.L., et al. (2018). Pharmacological
perturbation of CDK9 using selective CDK9 inhibition or degradation. Nat.
Chem. Biol. 14, 163–170.
Gadd, M.S., Testa, A., Lucas, X., Chan, K.H., Chen, W., Lamont, D.J.,
Zengerle, M., and Ciulli, A. (2017). Structural basis of PROTAC cooperative
recognition for selective protein degradation. Nat. Chem. Biol. 13, 514–521.
Remillard, D., Buckley, D.L., Paulk, J., Brien, G.L., Sonnett, M., Seo, H.S.,
Dastjerdi, S., Wuhr, M., Dhe-Paganon, S., Armstrong, S.A., et al. (2017).
Degradation of the BAF complex factor BRD9 by heterobifunctional ligands.
Angew. Chem. Int. Ed. 56, 5738–5743.
Gandhi, A.K., Kang, J., Havens, C.G., Conklin, T., Ning, Y., Wu, L., Ito, T.,
Ando, H., Waldman, M.F., Thakurta, A., et al. (2014). Immunomodulatory
agents lenalidomide and pomalidomide co-stimulate T cells by inducing
degradation of T cell repressors Ikaros and Aiolos via modulation of the E3
ubiquitin ligase complex CRL4(CRBN.). Br. J. Haematol. 164, 811–821.
Tigan, A.S., Bellutti, F., Kollmann, K., Tebb, G., and Sexl, V. (2016). CDK6 - a
review of the past and a glimpse into the future: from cell-cycle control to tran-
scriptional regulation. Oncogene 35, 3083–3091.
Wang, H., Nicolay, B.N., Chick, J.M., Gao, X., Geng, Y., Ren, H., Gao, H., Yang,
G., Williams, J.A., Suski, J.M., et al. (2017). The metabolic function of cyclin
D3-CDK6 kinase in cancer cell survival. Nature 546, 426–430.
Huang, H.T., Dobrovolsky, D., Paulk, J., Yang, G., Weisberg, E.L., Doctor,
Z.M., Buckley, D.L., Cho, J.H., Ko, E., Jang, J., et al. (2018). A chemoproteomic
approach to query the degradable kinome using a multi-kinase degrader. Cell
Chem. Biol. 25, 88–99.e6.
Winter, G.E., Buckley, D.L., Paulk, J., Roberts, J.M., Souza, A., Dhe-Paganon,
S., and Bradner, J.E. (2015). Drug development. Phthalimide conjugation as a
strategy for in vivo target protein degradation. Science 348, 1376–1381.
Klijn, C., Durinck, S., Stawiski, E.W., Haverty, P.M., Jiang, Z., Liu, H.,
Degenhardt, J., Mayba, O., Gnad, F., Liu, J., et al. (2015). A comprehensive
transcriptional portrait of human cancer cell lines. Nat. Biotechnol. 33,
Zengerle, M., Chan, K.H., and Ciulli, A. (2015). Selective small molecule
induced degradation of the BET bromodomain protein BRD4. ACS Chem.
Biol. 10, 1770–1777.
306–312.
Cell Chemical Biology 26, 1–7, February 21, 2019
7

Please cite this article in press as: Brand et al., Homolog-Selective Degradation as a Strategy to Probe the Function of CDK6 in AML, Cell Chemical
Biology (2018), https://doi.org/10.1016/j.chembiol.2018.11.006
STAR+METHODS
KEY RESOURCES TABLE
REAGENT or RESOURCE
SOURCE
IDENTIFIER
Antibodies
CDK4
Cell Signaling
Santa Cruz
Cat#12790; RRID: AB_2631166
Cat#sc-7961; RRID: AB_627242
Cat#ab1791; RRID: AB_302613
Cat#A5441; RRID: AB_476744
Cat#8180; RRID: AB_10950972
Cat#9309; RRID: AB_823629
CDK6
Histone 3
b-actin
p-Rb S780
Rb
Abcam
Sigma-Aldrich
Cell Signaling
Cell Signaling
Jackson Immuno Research
Peroxidase-conjugated AffiniPure Goat Anti-
Mouse IgG
Cat#115-035-003; RRID: AB_10015289
Peroxidase-conjugated AffiniPure Goat Anti-
Rabbit IgG
Jackson Immuno Research
Cat#111-035-003; RRID: AB_2313567
Chemicals, Peptides, and Recombinant Proteins
BSJ-03-123
This study
This study
This study
Selleckchem
N/A
BSJ-bump
N/A
YKL-06-102
N/A
Palbociclib
Cat#S1116
Critical Commercial Assays
Caspase-Glo 3/7
Promega
Promega
Cat#G8090
Cat#N2011
NanoBiT
Deposited Data
Project Achilles: CRISPR-Cas9 Avana
Meyers et al., 2017
Klijn et al., 2015
https://portals.broadinstitute.org/achilles/
datasets/18/download
Gene expression data of 675 cell lines
ENCODE transcription factor targets
https://www.nature.com/articles/nbt.3080#
supplementary-information
ENCODE Consortium
Alanis-Lobato et al., 2017
http://amp.pharm.mssm.edu/Harmonizome/
dataset/ENCODE+Transcription+Factor+Targets
Human protein-protein interaction network
Human reference genome NCBI build 38, GRCh38
http://cbdm-01.zdv.uni-mainz.de/ꢁmschaefer/
hippie/
http://hgdownload.cse.ucsc.edu/goldenPath/
hg38/bigZips/
RNA sequencing upon CDK6 degradation
This study
GEO: GSE116187
CDK4/6 targets
Experimental Models: Cell Lines
MV4-11
Anders et al., 2011
https://doi.org/10.1016/j.ccr.2011.10.001
ATCC
RRID: CVCL_0064
RRID: CVCL_0006
RRID: CVCL_2119
RRID: CVCL_1632
RRID: CVCL_0013
RRID: CVCL_1559
RRID: CVCL_0320
RRID: CVCL_0063
RRID: CVCL_0332
THP-1
ATCC
MOLM13
DMSZ
P31/FUJ
CCLE/Broad Institute
ATCC
MOLT4
NCI-H358
ATCC
HT29
ATCC
HEK293T
ATCC
Hs578T
ATCC
Oligonucleotides
sgCDK4 sense (5’- CACCGAGCCACTGGCTCAT
ATCGAG-3’)
Sigma-Aldrich
Sigma-Aldrich
N/A
N/A
sgCDK4 antisense (5’- AAACCTCGATATGAGCC
AGTGGCTC-3’)
(Continued on next page)
e1 Cell Chemical Biology 26, 1–7.e1–e9, February 21, 2019

Please cite this article in press as: Brand et al., Homolog-Selective Degradation as a Strategy to Probe the Function of CDK6 in AML, Cell Chemical
Biology (2018), https://doi.org/10.1016/j.chembiol.2018.11.006
Continued
REAGENT or RESOURCE
SOURCE
IDENTIFIER
N/A
sgCRBN sense (5’- CACCGTAAACAGAC
ATGGCCGGCGA-3’)
Sigma-Aldrich
sgCRBN antisense (5’- AAACTCGCCGGCCATG
TCTGTTTAC-3’)
Sigma-Aldrich
N/A
Recombinant DNA
Lenti_sgRNA_EFS_GFP
pBit1.1-C [TK/LgBiT]
pBit2.1-N [TK/SmBiT]
Software and Algorithms
Cytoscape
Addgene
Promega
Promega
https://www.addgene.org/65656
Cat#N2014
Cat#N2014
Institute for Systems Biology
Thermo Fisher Scientific
http://cytoscape.org; RRID: SCR_003032
GraphPad Prism
https://www.graphpad.com/scientific-
software/prism/
Proteome Discoverer
https://www.thermofisher.com/order/catalog/
product/OPTON-30795
KSEA App
STAR Aligner
limma
https://casecpb.shinyapps.io/ksea
https://github.com/alexdobin/STAR
https://bioconductor.org/packages/release/bioc/
html/limma.html
GOrilla
http://cbl-gorilla.cs.technion.ac.il/
https://networkx.github.io/
networkx
CONTACT FOR REAGENT AND RESOURCE SHARING
Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Georg
Winter (gwinter@cemm.at).
EXPERIMENTAL MODEL AND SUBJECT DETAILS
MV4-11 (RRID:CVCL_0064), THP-1 (RRID:CVCL_0006), MOLM13 (RRID:CVCL_2119), P31/FUJ (RRID:CVCL_1632), MOLT4
(
(
RRID:CVCL_0013), NCI-H358 (RRID:CVCL_1559) and HT29 (RRID:CVCL_0320) cells were cultured in RPMI-1640 growth medium
Gibco) supplemented with 10% fetal bovine serum (Gibco) and 1% Penicillin/Streptomycin solution (ThermoFisher Scientific).
HEK293T (RRID:CVCL_0063) and Hs578T (RRID:CVCL_0332) cells were cultured in DMEM (Gibco) supplemented with 10% fetal
ꢂ
bovine serum (Gibco) and 1% Penicillin/Streptomycin solution (ThermoFisher Scientific). All cell lines were cultured at 37 C with
5
2
% CO and tested for mycoplasma contamination and cell identity.
Sex of human cell lines: MV4-11 (Male, 10 years old), THP-1 (Male, 1 year old), MOLM13 (Male, 20 years old), P31/FUJ (Male,
years old), MOLT4 (Male, 19 years old), NCI-H358 (Male, age unknown), HT29 (Female, 44 years old), HEK293T (Female, fetus),
Hs578T (Female, 74 years old).
7
METHOD DETAILS
Knockout Generation
0
0
0
sgRNAs (CDK4: 5 -AGCCACTGGCTCATATCGAG-3 , CRBN: 5 - TAAACAGACATGGCCGGCGA-3’) were cloned into Lenti_
sgRNA_EFS_GFP (addgene #65656). Cells were transduced with lentiviral particles, selected by sorting for GFP positive cells and
successful knockout verified by immunoblot and Sanger sequencing.
Proliferation Assays
For growth over time experiments of suspension cell lines, 300.000 cells per well were seeded in 24-well plates in triplicates. Cells
were counted and treatments renewed every 2 days. For growth over time experiments of adherent cell lines, 100.000 cells per well
were seeded in 12-well plates in triplicates. Cells were trypsinized and counted and treatments renewed every 3 days. For colony
formation assays, 1000 cells per well were seeded in 6-well plates in duplicates. Every 2 days, culture medium was exchanged
and treatments were renewed. After 12 days, cells were fixed in 1% PFA in PBS for 15 min at room temperature, washed 3 times
with PBS and stained in Crystal Violet solution (0.1% in 10% EtOH) for 15 min at room temperature. The cells were again washed
3
times with PBS and left to dry overnight.
Cell Chemical Biology 26, 1–7.e1–e9, February 21, 2019 e2

Please cite this article in press as: Brand et al., Homolog-Selective Degradation as a Strategy to Probe the Function of CDK6 in AML, Cell Chemical
Biology (2018), https://doi.org/10.1016/j.chembiol.2018.11.006
Cell Cycle and Apoptosis
For cell cycle measurement, 10 cells per well were seeded in 24-well plates in triplicates. After 24 h of drug treatment, cells were
6
harvested by centrifugation at 500g for 5 min and washed in 1 ml cold PBS. Cells were spun down again, resuspended in 200 ml
ꢂ
PBS and fixed by addition of 800 ml 70% EtOH and incubation for 20 min at -20 C. Cells were washed with 1 ml PBS and stained
with 500 ml propidium iodide solution (50 mg/ml propidium iodide (Sigma Aldrich), 200 mg/ml RNase A (ThermoFisher Scientific), in
PBS). Cellular DNA content was measured by flow cytometry using FACSCalibur (BD Biosciences) and analyzed using the FlowJo
v10 software.
Analysis of apoptotic cells was performed using Caspase-Glo 3/7 (Promega) according to manufacturer’s instructions. 20 000 cells
per well were seeded in triplicate in a white 96-well plate in a total volume of 50 ml. After 24 h incubation with the treatment, 45 ml
Caspase-Glo substrate were added per well. The plate was incubated at room temperature in the dark for 1 h and luminescence
measured on a Victor X3 (Perkin Elmer) microplate reader. Luminescent signal was normalized to vehicle-treated cells.
In Vitro CRBN Binding Assay
Compounds in Atto565-Lenalidomide displacement assay were dispensed in a 384-well microplate (Corning, 4514) using D300e Dig-
ital Dispenser (HP) and normalized to 1% DMSO into 10 nM Atto565-Lenalidomide, 100 nM DDB1DB-CRBN, 50 mM Tris pH 7.5,
2
00 mM NaCl, 0.1% Pluronic F-68 solution (Sigma). The change in fluorescence polarization was monitored using a PHERAstar
FS microplate reader (BMG Labtech) for 30 cycles of 187s each. Data from three independent measurements (n = 3) was plotted
and IC50 values estimated using variable slope equation in GraphPad Prism 7.
Immunoblotting
6
x10 cells were lysed in RIPA buffer (150 nM NaCl, 0.1% Triton X-100, 0.5% sodium deoxycholate, 0.1% SDS, 50 mM Tris-HCl
2
(
pH 8.0), protease inhibitors (Thermo Scientific), benzonase (Novagen)). Cell lysates containing 40 mg of protein were loaded on
Boltꢁ 4-12% Bis-Tris Plus Gels (Thermo Scientific) and ran for SDS-PAGE before transfer onto 0.45 mM nitrocellulose membrane
(
GE Healthcare). Following antibodies were used for detection:
Name
Origin
Cat. No.
#12790
sc-7961
ab1791
A5441
CDK4
Cell Signaling
Santa Cruz
CDK6
Histone 3
b-actin
p-Rb S780
Rb
Abcam
Sigma-Aldrich
Cell Signaling
Cell Signaling
Jackson Immuno Research
#8180
#9309
Peroxidase-conjugated AffiniPure Goat Anti-
Mouse IgG
115-035-003
Peroxidase-conjugated AffiniPure Goat Anti-
Rabbit IgG
Jackson Immuno Research
111-035-003
Cellular Thermal Shift Assay
6
x10 MV4-11 CRBN-/- cells were treated for 3 h with 20 mM palbociclib, BSJ-03-123 or vehicle. 1x10 cells were spun down at
6
4
ꢂ
500 x g for 5 min, and the supernatant removed. Pellets were incubated at 46, 49, 52 or 55 C for 3 min followed by 3 min incubation
at room temperature. 30 ml of lysis buffer (20 mM Tris-HCl pH 8.0, 120 mM NaCl, 0.5% NP-40, protease inhibitors) were added and
ꢂ
cells lysed by 3 rounds of snap freezing and thawing. Denatured proteins were removed by 20 min centrifugation at 14.000 x g at 4 C
and supernatants analyzed by Western Blotting.
NanoBiTꢀ Assay
The NanoBiTꢀ assay (Promega) was performed following manufacturer’s instructions. Full-length CDK4 and CDK6 were cloned
into pBit1.1-C [TK/LgBiT] vector, full-length CRBN was cloned into pBit2.1-N [TK/SmBiT]. 400.000 CRBN-deficient HEK293T
cells were seeded in a 6-well plate and transfected the day after with 750 mg each of CDK4/6-LgBit and CRBN-SmBit using
Lipofectamine (Thermo Fisher Scientific) according to manufacturer’s instructions. After 24 h, cells were detached and
5
0.000 cells per well seeded in a 96-well plate and allowed to attach overnight. Before the assay, the medium was exchanged
with 100 ml of ambient temperature fresh medium supplemented with 25 mM HEPES and the plate incubated for 10 min at room
temperature to equilibrate. 25 ml Nano-Gloꢀ Live Cell Reagent was added per well, incubated for 5 min and baseline lumines-
cence measured on a Victor X3 (Perkin Elmer) with 2 s integration time. 10 ml of 13.5X drug stock were added per well and
emission of luminescent signal monitored by continuous measurement every 2 min. Signals were normalized to baseline and
vehicle treated cells.
e3 Cell Chemical Biology 26, 1–7.e1–e9, February 21, 2019

Please cite this article in press as: Brand et al., Homolog-Selective Degradation as a Strategy to Probe the Function of CDK6 in AML, Cell Chemical
Biology (2018), https://doi.org/10.1016/j.chembiol.2018.11.006
Expression Proteomics BSJ-03-123
Sample Preparation
6
5x10 MOLT4 cells were treated with 100 nM BSJ-03-123 for 2 h, in triplicate. Cells were washed three times with PBS, the super-
1
2
natant aspirated and pellets snap-frozen in liquid N . Each washed cell pellet was lysed separately in 40 mL of freshly prepared lysis
buffer containing 50 mM HEPES (pH 8.0), 2% SDS, 0.1 M DTT, 1 mM PMSF, and protease inhibitor cocktail (Sigma-Aldrich). Samples
ꢂ
rested at RT for 20 minutes before heating to 99 C for 5 min. After cooling down to RT, DNA was sheared by sonication using a Co-
ꢂ
varis S2 high performance ultrasonicator. Cell debris were removed by centrifugation at 20.000 3 g for 15 min at 20 C. Supernatant
was transferred to fresh eppendorf tubes and protein concentration determined using the BCA protein assay kit (Pierce Biotech-
nology, Rockford, IL). FASP was performed using a 30 kDa molecular weight cutoff filter (VIVACON 500). Fifty microliters of each
cleared protein extract were mixed with 200 mL of freshly prepared 8 M urea in 100 mM Tris-HCl (pH 8.5) (UA-solution) in the filter
ꢂ
unit and centrifuged at 14.000 3 g for 15 min at 20 C to remove SDS. Any residual SDS was washed out by a second washing
step with 200 mL of UA. The proteins were alkylated with 100 mL of 50 mM iodoacetamide in the dark for 30 min at RT. Afterward,
three washing steps with 100 mL of UA solution were performed, followed by three washing steps with 100mL of 50 mM TEAB buffer
ꢂ
(Sigma-Aldrich). Proteins were digested with trypsin overnight at 37 C. Peptides were recovered using 40 mL of 50 mM TEAB buffer
followed by 50 mL of 0.5 M NaCl (Sigma-Aldrich). Peptides were desalted using C18 solid phase extraction spin columns (The Nest
Group, Southborough, MA). After desalting, peptides were labeled with TMT 10plexꢁ reagents according to the manufacturer
(
Pierce, Rockford, IL). After quenching of the labeling reaction, labeled peptides were pooled, organic solvent removed in a vacuum
ꢂ
concentrator at 45 C and reconstituted in 5% acetonitrile containing 20mM ammonia formate buffer, pH 10 for offline fractionation
using high pH reversed phase liquid chromatography (2D-RP/RP-HPLC).
2D-RP/RP Liquid Chromatography Mass Spectrometry
Two-dimensional liquid chromatography was performed by reverse-phase chromatography at high and low pH. In the first dimen-
˚
sion, peptides were separated on a Gemini-NX C18 (150 3 2 mm, 3 mm, 110 A, Phenomenex, Torrance, USA) in 20 mM ammonia
formate buffer, pH 10, and eluted over 45 min by a 5-70% acetonitrile gradient at 100 mL/min using an Agilent 1200 HPLC system
(
Agilent Biotechnologies, Palo Alto, CA, USA). Ninty-six time-based fractions were collected and consolidated into 40 fractions. After
solvent removal in a vacuum concentrator, samples were reconstituted in 5% formic acid for LC-MS/MS analysis. Mass spectrom-
etry was performed on an Orbitrap Fusion Lumos mass spectrometer (ThermoFisher Scientific, San Jose, CA) coupled to an Agilent
1
200 HPLC nanoflow system (Agilent Biotechnologies, Palo Alto, CA) via nanoflex source interface. Tryptic peptides were loaded
onto a trap column (Zorbax 300SB-C18 5 mm, 5 3 0.3 mm, Agilent Biotechnologies) at a flow rate of 45 mL/min using 0.1% TFA
as loading buffer. After loading, the trap column was switched in-line with a 25 cm, 75 mm inner diameter analytical column (packed
in-house with ReproSil-Pur 120 C18-AQ, 3 mm, Dr. Maisch, Ammerbuch-Entringen, Germany). Mobile-phase A consisted of 0.4%
formic acid in water and mobile-phase B of 0.4% formic acid in a mix of 90% acetonitrile and 10% water. The flow rate was set
to 230 nL/min and a 90 min gradient used (6 to 30% solvent B within 81 min, 30 to 65% solvent B within 8 min and, 65 to 100% solvent
B within 1 min, 100% solvent B for 6 min before equilibrating at 6% solvent B for 18 min). Analysis was performed in a data-dependent
acquisition mode using the multi-notch MS3-based TMT method (termed SPS-MS3 on the Lumos instrument). MS1 spectra were
acquired within a mass range of 350 - 1550 m/z in the orbitrap at a resolution of 120,000 (at 200Th). Automatic gain control (AGC)
5
was set to a target of 2 3 10 and a maximum injection time of 50 ms. Precursor ions for MS2/MS3 analysis were selected using
a Top10 dependent scan approach. MS2 spectra were acquired in the linear quadrupole ion trap (IT) using a quadrupole isolation
window of 0.8Da; collision induced dissociation (CID) for fragmentation; a normalized collision energy (NCE) of 35%; an AGC target
3
of 4 3 10 ; and a maximum injection time of 150 ms. For TMT quantitation, each MS2 scan followed a SPS-MS3 scan of the same
precursor ion using the multiple frequency notches approach 3. The quadrupole isolation window for MS3 scans was set to 2 Da and
the top 5 most intense MS2 fragment ions were isolated by SPS for fragmentation by higher energy collision-induced dissociation
(
HCD) using a NCE of 65%. Resulting fragment ions were analyzed in the Orbitrap at a resolution of 50,000 (at 200 Th). AGC settings
4
were 5 3 10 and a maximum injection time of 150 ms. Dynamic exclusion for selected ions was 60 s. A single lock mass at m/z
45.120024 was employed. Xcalibur version 4.0.0 and Tune 2.1 were used to operate the instrument.
4
Expression Proteomics YKL-06-102
MOLT4 cells were treated with DMSO or 500 nM of compound YKL-06-102 in biological triplicates for 5 hours and cells harvested by
centrifugation. Lysis buffer (8 M Urea, 50 mM NaCl, 50 mM 4-(2hydroxyethyl)-1-piperazineethanesulfonic acid (EPPS) pH 8.5, Pro-
tease and Phosphatase inhibitors from Roche) was added to the cell pellets and homogenized by 20 passes through a 21 gauge
(
1.25 in. long) needle to achieve a cell lysate with a protein concentration between 1 – 4 mg mL-1. A micro-BCA assay (Pierce)
was used to determine the final protein concentration in the cell lysate. 200 mg of protein for each sample were reduced and alkylated.
Proteins were precipitated using methanol/chloroform. In brief, four volumes of methanol were added to the cell lysate, followed by
one volume of chloroform, and finally three volumes of water. The mixture was vortexed and centrifuged to separate the chloroform
phase from the aqueous phase. The precipitated protein was washed with three volumes of methanol, centrifuged and the resulting
washed precipitated protein was allowed to air dry. Precipitated protein was resuspended in 4 M Urea, 50 mM HEPES pH 7.4, fol-
lowed by dilution to 1 M urea with the addition of 200 mM EPPS, pH 8. Proteins were first digested with LysC (1:50; enzyme:protein)
for 12 hours at room temperature. The LysC digestion was diluted to 0.5 M Urea with 200 mM EPPS pH 8 followed by digestion with
ꢂ
trypsin (1:50; enzyme:protein) for 6 hours at 37 C. Tandem mass tag (TMT) reagents (Thermo Fisher Scientific) were dissolved in
anhydrous acetonitrile (ACN) according to manufacturer’s instructions. Anhydrous ACN was added to each peptide sample to a final
Cell Chemical Biology 26, 1–7.e1–e9, February 21, 2019 e4

Please cite this article in press as: Brand et al., Homolog-Selective Degradation as a Strategy to Probe the Function of CDK6 in AML, Cell Chemical
Biology (2018), https://doi.org/10.1016/j.chembiol.2018.11.006
concentration of 30% v/v, and labeling was induced with the addition of TMT reagent to each sample at a ratio of 1:4 peptide:TMT
label. The 10-plex labeling reactions were performed for 1.5 hours at room temperature and the reaction quenched by the addition of
hydroxylamine to a final concentration of 0.3% for 15 minutes at room temperature. The sample channels were combined at a
1
:1:1:1:1:1:1:1:1:1 ratio, desalted using C18 solid phase extraction cartridges (Waters) and analyzed by LC-MS for channel ratio com-
parison. Samples were then combined using the adjusted volumes determined in the channel ratio analysis and dried down in a
speed vacuum. The combined sample was then resuspended in 1% formic acid, and acidified (pH 2ꢀ3) before being subjected
to desalting with C18 SPE (Sep-Pak, Waters). Samples were then offline fractionated into 96 fractions by high pH reverse-phase
HPLC (Agilent LC1260) through an aeris peptide xb-c18 column (phenomenex) with mobile phase A containing 5% acetonitrile
and 10 mM NH4HCO3 in LC-MS grade H2O, and mobile phase B containing 90% acetonitrile and 10 mM NH4HCO3 in LC-MS grade
H2O (both pH 8.0). The 96 resulting fractions were then pooled in a non-continuous manner into 24 fractions and these fractions were
used for subsequent mass spectrometry analysis.
Data were collected using an Orbitrap Fusion Lumos mass spectrometer (Thermo Fisher Scientific, San Jose, CA, USA) coupled
with a Proxeon EASY-nLC 1200 LC pump (Thermo Fisher Scientific). Peptides were separated on a 75 mM inner diameter microca-
˚
pillary column packed with ꢁ40 cm of Accucore C18 resin (1.6 mM, 100 A, Thermo Fisher Scientific). Peptides were separated using a
1
50 min gradient of 6–27% acetonitrile in 1.0% formic acid with a flow rate of 400 nL/min.
Each analysis used an MS3-based TMT method. The data were acquired using a mass range of m/z 340 – 1350, resolution
20,000, AGC target 1 x 106, maximum injection time 100 ms, dynamic exclusion of 120 seconds for the peptide measurements
1
in the Orbitrap. Data dependent MS2 spectra were acquired in the ion trap with a normalized collision energy (NCE) set at 35%,
AGC target set to 1.8 x 104 and a maximum injection time of 120 ms. MS3 scans were acquired in the Orbitrap with a HCD collision
energy set to 55%, AGC target set to 1.5 x 105, maximum injection time of 150 ms, resolution at 50,000 and with a maximum syn-
chronous precursor selection (SPS) precursors set to 10.
Phosphoproteomics
6
5x10 MV4-11 cells were treated with 200 nM BSJ-03-123 or Palbociclib for 2 h, in triplicate. Cells were washed three times with
1
2
PBS, the supernatant aspirated and pellets snap-frozen in liquid N .
Each washed cell pellet was lysed separately in 40 mL of freshly prepared lysis buffer containing 50 mM HEPES (pH 8.0), 2% SDS,
.1 M DTT, 1 mM PMSF, phosSTOP and protease inhibitor cocktail (Sigma-Aldrich). Samples rested at RT for 20 minutes before heat-
0
ing to 99 C for 5 min. After cooling down to RT, DNA was sheared by sonication using a Covaris S2 high performance ultrasonicator.
ꢂ
ꢂ
Cell debris was removed by centrifugation at 20.000 X g for 15 min at 20 C. Supernatant was transferred to fresh eppendorf tubes and
protein concentration determined using the BCA protein assay kit (Pierce Biotechnology, Rockford, IL). FASP was performed using a
30 kDa molecular weight cutoff filter (VIVACON 500). In brief, 100 mg total protein per sample were reduced by adding DTT at a final
ꢂ
concentration of 83.3 mM followed by incubation at 99 C for 5 min. After cooling to room temperature, samples were mixed with
2
00 mL of freshly prepared 8 M urea in 100 mM Tris-HCl (pH 8.5) (UA-solution) in the filter unit and centrifuged at 14.000 3 g for
ꢂ
1
5 min at 20 C to remove SDS. Any residual SDS was washed out by a second washing step with 200 mL of UA. The proteins
were alkylated with 100 mL of 50 mM iodoacetamide in the dark for 30 min at RT. Afterward, three washing steps with 100 mL of
UA solution were performed, followed by three washing steps with 100mL of 50 mM TEAB buffer (Sigma-Aldrich). Proteins were
ꢂ
digested with trypsin at a ratio of 1:35 overnight at 37 C. Peptides were recovered using 40 mL of 50 mM TEAB buffer followed by
5
0 mL of 0.5 M NaCl (Sigma-Aldrich). Peptides were desalted using C18 solid phase extraction spin columns (The Nest Group, South-
borough, MA). After desalting, peptides were labeled with TMT 10plexꢁ reagents (label 128N was omitted) according to the manu-
facturer (Pierce, Rockford, IL). After quenching of the labeling reaction, labeled peptides were pooled, organic solvent removed in
vacuum concentrator and labelled peptides loaded onto a SPE column. Peptides were eluted with 300mL 80% acetonitrile containing
0
immobilized metal affinity chromatography (IMAC). Briefly, two times 100 mL of Ni-NTA superflow slurry (QIAGEN Inc., Valencia,
.1% trifluoroacetic to achieve a final peptide concentration of ꢁ1mg/ml. Eluate was then used for phosphopeptide enrichment via
2
USA) were washed with LCMS-grade water and Ni stripped off the beads by incubation with 100 mM of EDTA, pH 8 solution
+
3+
for 1 hr at room temperature. Stripped NTA resin was recharged with Fe -ions by incubation with a fresh solution of Fe(III)Cl
3
and 2 x 110 mL of charged resin slurry used for the enrichment of a total of ꢁ900 mg TMT-labelled peptide. The unbound fraction
was transferred to a fresh glass vial and used for offline fractionation for the analysis of the whole chromatome proteome. After
washing the slurry with 0.1% TFA, phosphopeptides were eluted with a freshly prepared ammonia solution containing 3mM
EDTA, pH 8 and all used for offline fractionation for the analysis of the phophoproteome.
Offline Fractionation via RP-HPLC at High pH
Peptides were re-buffered in 20 mM ammonium formiate buffer shortly before separation by reversed phase liquid chromatography
at pH 10. Phosphopeptides was separated into 20 fractions on a Dionex column (500 mm 3 50 mm PepSwift RP, monolithic, Dionex
Corporation, Sunnyvale, CA, USA) using an Agilent 1200 series nanopump delivering solvent at 4 mL/min. Peptides were separated by
applying a gradient of 90% aceonitrile containing 20 mM ammonium formiate, pH 10. After solvent removal in a vacuum concentrator,
ꢂ
samples were reconstituted in 0.1% trifluroacetic acid (TFA) for LC-MS/MS analysis. Prepared samples were kept at -80 C until the
analysis.
2
D-RP/RP Liquid Chromatography Mass Spectrometry
Mass spectrometry was performed on an Orbitrap Fusion Lumos mass spectrometer (ThermoFisher Scientific, San Jose, CA)
coupled to an Dionex Ultimate 3000RSLC nano system (ThermoFisher Scientific, San Jose, CA) via nanoflex source interface. Tryptic
e5 Cell Chemical Biology 26, 1–7.e1–e9, February 21, 2019

Please cite this article in press as: Brand et al., Homolog-Selective Degradation as a Strategy to Probe the Function of CDK6 in AML, Cell Chemical
Biology (2018), https://doi.org/10.1016/j.chembiol.2018.11.006
peptides were loaded onto a trap column (Pepmap 100 5mm, 5 3 0.3 mm, ThermoFisher Scientific, San Jose, CA) at a flow rate of
1
0 mL/min using 2% ACN and 0.1% TFA as loading buffer. After loading, the trap column was switched in-line with a 30 cm, 75 mm
inner diameter analytical column (packed in-house with ReproSil-Pur 120 C18-AQ, 3 mm, Dr. Maisch, Ammerbuch-Entringen, Ger-
many). Mobile-phase A consisted of 0.4% formic acid in water and mobile-phase B of 0.4% formic acid in a mix of 90% acetonitrile
and 10% water. The flow rate was set to 230 nL/min and a 90 min gradient used (6 to 30% solvent B within 81 min, 30 to 65% solvent
B within 8 min and, 65 to 100% solvent B within 1 min, 100% solvent B for 6 min before equilibrating at 6% solvent B for 18 min).
Analysis was performed in a data-dependent acquisition mode. Full MS scans were acquired with a mass range of 375 -
5
650 m/z in the orbitrap at a resolution of 120,000 (at 200Th). Automatic gain control (AGC) was set to a target of 2 3 10 and a
1
maximum injection time of 50 ms. Precursor ions for MS analysis were selected using a TopN dependant scan approach with a cycle
2
2
time of 3 seconds. MS spectra were acquired in the orbitrap (FT) using a quadrupole isolation window of 1 Da and higher energy
4
collision induced dissociation (HCD) at a normalized collision energy (NCE) of 38%. AGC target was set to 5 3 10 with a maximum
2
injection time of 150 ms and MS scans acquired at a resolution of 15,000 (at 200 Th). Dynamic exclusion for selected ions was 60 s. A
single lock mass at m/z 445.120024 was employed. Xcalibur version 4.0.0 and Tune 2.1 were used to operate the instrument.
RNA Sequencing
6
x10 MV4-11 cells per well were seeded in triplicate in a 12-well plate and treated with 200 nM BSJ-03-123, YKL-06-102, Palbociclib
3
or vehicle for 6 h. Cells were lysed using QIAshredder columns (QIAGEN) and RNA isolated using RNeasy kit (QIAGEN) according to
manufacturer’s instructions.
The amount of total RNA was quantified using the Qubit Fluorometric Quantitation system (Life Technologies) and the RNA integrity
number (RIN) was determined using the Experion Auto-mated Electrophoresis System (Bio-Rad). RNA-seq libraries were prepared
with the TruSeq Stranded mRNA LT sample preparation kit (Illumina) using both, Sciclone and Zephyr liquid handling robotics
(
PerkinElmer). Library concentrations were quantified with the Qubit Fluorometric Quantitation system (Life Technologies) and the
size distribution was assessed using the Experion Automated Electrophoresis System (Bio-Rad). For sequencing, samples were
diluted and pooled into NGS libraries in equimolar amounts. Expression profiling libraries were sequenced on Illumina HiSeq
3
000/4000 instruments in 50-base-pair-single-end mode and base calls provided by the Illumina Real-Time Analysis (RTA) soft-
ware were subsequently converted into BAM format (Illumina2bam) before de-multiplexing (BamIndexDecoder) into individual, sam-
ple-specific BAM files via Illumina2bam tools (1.17.3).
Chemical Synthesis of BSJ-03-123, BSJ-03-190/BSJ-bump, YKL-06-102
Synthesis of 6a and 6b
2
-((2-(2,6-dioxopiperidin-3-yl)-1,3-dioxoisoindolin-4-yl)oxy)acetic acid (6a). 6a was prepared according to the literature (Remillard
et al., 2017). 3-Hydroxyphthalic anhydride (1) (1.64 g, 10 mmol) and 3-aminopiperidine-2,6-dione hydrochloride (2a) (1.65 g, 10 mmol)
ꢂ
were dissolved in pyridine (40 mL, 0.25 M) and heated to 110 C. After 14h, the mixture was cooled to room temperature and
concentrated under reduced pressure. The residue was purification by column chromatography on silica gel (0-10% MeOH/DCM)
to give 2-(2,6-dioxopiperidin-3-yl)-4-hydroxyisoindoline-1,3-dione (3a) as a grey solid (2.41 g, 88%). LC-MS: m/z 275 [M+1]. To a
solution of 3a (2.19 g, 8 mmol) in 8 mL of DMF was added K
2 3
CO (1.66 g, 12 mmol) and t-butyl bromoacetate (4) (1.18 mL,
8
mmol) respectively. The mixture was stirred at room temperature for 2h, then diluted with EtOAc and washed once with water
then twice with brine. The organic layer was dried over anhydrous sodium sulfate, filtered and concentrated under reduced pressure.
The residue was purification by column chromatography on silica gel (5-100%EtOAc/hexanes) to give tert-butyl 2-((2-(2,6-dioxopi-
peridin-3-yl)-1,3-dioxoisoindolin-4-yl)oxy)acetate (5a) as a cream colored solid (2.70 g, 87%). LC-MS: m/z 389 [M+1]. 5a (2.06g,
5
.3 mmol) was then dissolved in TFA (53 mL, 0.1M) at room temperature. After 4 hours, the solution was diluted with DCM and
concentrated under reduced pressure to give 6a as a cream colored solid (1.5 g, 85%) was deemed sufficiently pure and carried
1
onto the next step without further purification. LC-MS: m/z 333 [M+1]. H NMR (500 MHz, DMSO-d6) d 11.09 (s, 1H), 7.79 (dd,
J = 8.4, 7.4 Hz, 1H), 7.48 (d, J =7.4 Hz, 1H), 7.39 (d, J = 8.6 Hz, 1H), 5.10 (dd, J = 12.8, 5.4 Hz, 1H), 4.99 (s, 2H), 2.93-2.89
m, 1H), 2.63-2.51 (m, 2H), 2.11-2.03 (m, 1H).
-((2-(1-methyl-2,6-dioxopiperidin-3-yl)-1,3-dioxoisoindolin-4-yl)oxy)acetic acid (6b). 6b was synthesized with similar procedures
(
2
as 6a from 3-Hydroxyphthalic anhydride (1) (328 mg, 2 mmol), 3-amino-1-methylpiperidine-2,6-dione (2b) (357 mg, 2 mmol) and
t-butyl bromoacetate (4) (0.295 mL, 2 mmol). 6b was obtained as an off-white solid (451 mg, 65% yield in 3 steps). LC-MS: m/z
1
47 [M+1]. H NMR (500 MHz, DMSO-d
3
6
) d 13.24 (s, 1H), 7.80 (dd, J = 8.5, 7.3 Hz, 1H), 7.48 (d, J = 7.3 Hz, 1H), 7.40
(
(
d, J = 8.5 Hz, 1H), 5.17 (dd, J = 13.0, 5.4 Hz, 1H), 4.99 (s, 2H), 3.02 (s, 3H), 2.99-2.91 (m, 1H), 2.80-2.73 (m, 1H), 2.59-2.52
1
3
m, 1H), 2.09-2.02 (m, 1H). C NMR (125 MHz, DMSO) d 172.24, 170.14, 169.95, 167.17, 165.62, 155.61, 137.26, 133.70, 120.36,
16.76, 116.24, 65.46, 49.82, 31.56, 27.07, 21.65.
1
Synthesis of BSJ-03-123 and BSJ-03-190/BSJ-Bump
tert-butyl (2-(2-(2-(2-(4-(6-((6-acetyl-8-cyclopentyl-5-methyl-7-oxo-7,8-dihydropyrido[2,3-d]pyrimidin-2-yl)amino)pyridin-3-yl)piper-
azin-1-yl)ethoxy)ethoxy)ethoxy)ethyl)carbamate (9). To a suspension of Palbociclib (7) (100 mg, 0.22 mmol) in DMSO (5 mL) was
added tert-butyl (2-(2-(2-(2-bromoethoxy)ethoxy)ethoxy)ethyl)carbamate (8) (156 mg, 0.44 mmol) and DIPEA (0.115 mL,
ꢂ
0
.66 mmol). The mixture was heated to 80 C and kept stirring for 48h. The mixture was then cooled down to room temperature, ex-
tracted, dried, filtered and concentrated. The residue was purified by reverse phase HPLC (5-95% MeOH in H
1
as a yellow solid (103mg, 65%). LC-MS: m/z 723 [M+1]. H NMR (500 MHz, DMSO-d
2
O) to give 9 (TFA salt)
6
) d 10.34 (s, 1H), 8.97 (s, 1H), 8.12 (d, J = 3.0 Hz,
Cell Chemical Biology 26, 1–7.e1–e9, February 21, 2019 e6

Please cite this article in press as: Brand et al., Homolog-Selective Degradation as a Strategy to Probe the Function of CDK6 in AML, Cell Chemical
Biology (2018), https://doi.org/10.1016/j.chembiol.2018.11.006
1
H), 7.90 (d, J = 9.1 Hz, 1H), 7.64-7.58 (m, 1H), 6.81-6.74 (m, 1H), 5.89-5.78 (m, 1H), 3.93-3.75 (m, 4H), 3.67-3.60 (m, 4H), 3.59-3.55
m, 2H), 3.55-3.46 (m, 4H), 3.44-3.35 (m, 4H), 3.27 (br, 2H), 3.15-3.01 (m, 4H), 2.42 (s, 3H), 2.32 (s, 3H), 2.28-2.19 (m, 2H), 1.95-1.84
(
(
13
m, 2H), 1.83-1.71 (m, 2H), 1.64-1.52 (m, 2H), 1.36 (s, 9H). C NMR (126 MHz, DMSO) d 202.38, 160.69, 158.32, 158.08, 157.98,
1
5
55.59, 154.77, 145.07, 141.96, 141.83, 134.97, 129.58, 126.24, 115.21, 106.93, 77.61, 69.67, 69.60, 69.52, 69.47, 69.17, 64.24,
4.85, 52.96, 50.95, 45.48, 31.29, 28.21, 27.57, 25.14, 13.65.
N-(2-(2-(2-(2-(4-(6-((6-acetyl-8-cyclopentyl-5-methyl-7-oxo-7,8-dihydropyrido[2,3-d]pyrimidin-2-yl)amino)pyridin-3-yl)piperazin-1-
yl)ethoxy)ethoxy)ethoxy)ethyl)-2-((2-(2,6-dioxopiperidin-3-yl)-1,3-dioxoisoindolin-4-yl)oxy)acetamide (10a). To a solution of the 9
(
30.5 mg, 0.0422 mmol) in DCM (2 mL) was added TFA (1 mL) and the resulting solution was stirred at room temperature for 1h.
The mixture was concentrated and the residue was then dissolved in DMF (1 mL) followed by adding 6a (14mg, 0.0422 mmol),
HATU (33 mg, 0.0844 mmol) and DIPEA (37 mL, 0.211 mmol). The resulting mixture was stirred for 1h at room temperature, then evap-
orated the solvent and purified by reverse phase HPLC (5-95% MeOH in H
2
O) to give 10a (TFA salt) as a yellow solid (34.4 mg, 87%).
) d 11.2 (s, 1H, NH), 9.03 (s, 1H, AR-H), 8.09 (dt, J = 10.6, 3.1 Hz, 2H, AR-H), 7.87
t, J = 8.5 Hz, 1H, AR-H), 7.79 (dd, J = 8.5, 7.3 Hz, 1H, AR-H), 7.47 (d, J = 7.2 Hz, 1H, AR-H), 7.40 (d, J = 8.5 Hz, 1H, AR-H), 5.89-5.80
m, 1H, N-CH), 5.11 (dd, J = 12.8, 5.5 Hz, 1H, N-CH), 4.80 (s, 2H, CO-CH -O), 3.95-3.82 (m, 4H, 2xO-CH ), 3.75-3.55 (m, 8H, 4xO-
), 3.54-3.52 (m, 2H, N-CH ), 3.51-3.45 (m, 2H, N-CH ), 3.42-3.19 (m, 8H, 4xN-CH ), 3.02-2.98 (br, 1H, H’N-C-CH2), 2.65-2.52 (m,
H, CO-CH ), 2.44 (s, 3H, CO-CH ), 2.28-2.12 (m, 2H, C-CH ), 2.05 (ddt, J = 15.2, 7.8, 2.8 Hz, 1H, H’N-C-CH2),
.98-1.85 (m, 2H, C-CH ), 1.82-1.74 (m, 2H, C-CH ). C NMR (126 MHz, DMSO) d 202.43, 172.82, 169.90,
1
LC-MS: m/z 937 [M+1]. H NMR (500 MHz, DMSO-d
6
(
(
2
2
CH
2
2
2
2
2
1
1
1
3
2
3
), 2.35 (s, 3H, AR-CH
3
2
13
2
2
), 1.64-1.53 (m, 2H, C-CH
2
66.97, 166.72, 165.51, 160.70, 158.33, 158.23, 157.98, 154.92, 154.82, 144.95, 141.95, 141.83, 136.98, 133.03, 129.77, 126.77,
20.40, 116.78, 116.14, 115.35, 107.13, 69.71, 69.65, 69.63, 69.52, 68.83, 67.57, 64.19, 54.90, 53.00, 50.96, 48.85, 45.48, 38.40,
1.32, 30.98, 27.61, 25.19, 22.02, 13.70.
N-(2-(2-(2-(2-(4-(6-((6-acetyl-8-cyclopentyl-5-methyl-7-oxo-7,8-dihydropyrido[2,3-d]pyrimidin-2-yl)amino)pyridin-3-yl)piperazin-1-
yl)ethoxy)ethoxy)ethoxy)ethyl)-2-((2-(1-methyl-2,6-dioxopiperidin-3-yl)-1,3-dioxoisoindolin-4-yl)oxy)acetamide (10b). 10b was
synthesized with similar procedures as 10a from 9 (30.5 mg, 0.0422 mmol) and 6b (14.6mg, 0.0422 mmol). 10b was obtained
1
as a yellow solid (33.7 mg, 84%). LC-MS: m/z 951 [M+1]. H NMR (500 MHz, DMSO-d
6
) d 9.75 (br, 1H, NH), 8.97 (s, 1H, AR-H),
8
.10 (d, J = 3.1 Hz, 1H, AR-H), 8.03-7.93 (m, 1H, AR-H), 7.87 (d, J = 9.1 Hz, 1H, AR-H), 7.85-7.76 (m, 1H, AR-H), 7.64 (dd,
J = 9.2, 3.2 Hz, 1H, AR-H), 7.50 (d, J = 7.3 Hz, 1H, AR-H), 7.40 (d, J = 8.5 Hz, 1H, HAR), 5.87-5.76 (m, 1H, N-CH), 5.21-5.10
m, 1H, N-CH), 4.79 (s, 2H, CO-CH -O), 3.91-3.82 (m, 2H, O-CH ), 3.81-3.77 (m, 2H, O-CH ), 3.65-3.59 (m, 2H, O-CH ), 3.58-
.55 (m, 2H, O-CH ), 3.54 (br, 4H, 2xO-CH ), 3.47 (t, J = 5.8 Hz, 2H, N-CH ), 3.43-3.36 (m, 2H, N-CH ), 3.35-3.30 (m, 2H, N-
CH ), 3.30-3.18 (m, 2H, N-CH ), 3.13-3.04 (m, 2H, N-CH ), 3.02 (s, 3H, N-CH ), 2.99-2.91 (m, 2H, N-CH ), 2.80-2.74 (m, 2H,
CO-CH ), 2.74-2.70 (m, 1H, H’N-C-CH2), 2.43 (s, 3H, CO-CH ), 2.32 (s, 3H, Ar-CH ), 2.29-2.16 (m, 2H, C-CH ), 2.06 (dtd,
J = 13.0, 5.5, 2.4 Hz, 1H, H’N-C-CH2), 1.95-1.84 (m, 2H, C-CH ), 1.83-1.72 (m, 2H, C-CH ), 1.65-1.51 (m, 2H, C-CH
126 MHz, DMSO) d 202.37, 171.75, 169.62, 166.93, 166.67, 165.47, 160.67, 158.32, 158.20, 158.03, 157.95, 154.95, 154.79,
(
2
2
2
2
3
2
2
2
2
2
2
2
3
2
2
3
3
2
1
3
2
2
2
). C NMR
(
1
6
44.90, 141.92, 141.79, 136.98, 133.00, 129.75, 126.78, 120.40, 116.73, 116.12, 115.34, 107.12, 69.68, 69.61, 69.49, 68.81,
7.55, 64.17, 54.86, 52.97, 50.92, 49.38, 45.45, 38.37, 31.30, 31.09, 27.59, 26.61, 25.16, 21.21, 13.68.
Synthesis of YKL-06-102
tert-butyl 2-(4-(6-((6-acetyl-8-cyclopentyl-5-methyl-7-oxo-7,8-dihydropyrido[2,3-d]pyrimidin-2-yl)amino)pyridin-3-yl)piperazin-1-yl)
acetate (11). To a suspension of Palbociclib (7) (100 mg, 0.22 mmol) in DMSO (5 mL) were added t-butyl bromoacetate (4)
ꢂ
(
then cooled down to room temperature, extracted, dried, filtered and concentrated. The residue was purified by reverse phase
65 mL, 0.44 mmol) and DIPEA (0.115 mL, 0.66 mmol). The mixture was heated to 80 C and kept stirring for 24h. The mixture was
1
HPLC (5-95% MeOH in H
DMSO-d6) d 10.09 (s, 1H), 8.95 (s, 1H), 8.05 (d, J = 3.0 Hz, 1H), 7.85 (d, J = 9.0 Hz, 1H), 7.49-7.43 (m, 1H), 5.88-5.75 (m, 1H),
.20-3.13 (m, 6H), 2.67 (t, J = 4.9 Hz, 4H), 2.42 (s, 3H), 2.31 (s, 3H), 2.27-2.18 (m, 2H), 1.89 (br, 2H), 1.82-1.73 (m, 2H), 1.65-1.52
2
O) to give 11 (TFA salt) as a yellow solid (90 mg, 73%). LC-MS: m/z 562 [M+1]. H NMR (500 MHz,
3
1
3
(
m, 2H), 1.43 (s, 9H). C NMR (126 MHz, DMSO) d 202.91, 169.67, 161.22, 159.04, 158.74, 155.22, 144.77, 143.88, 142.55,
35.89, 125.21, 115.60, 107.02, 100.00, 80.72, 59.66, 53.36, 52.12, 48.78, 31.78, 28.29, 28.02, 25.58, 14.10.
-(4-(6-((6-acetyl-8-cyclopentyl-5-methyl-7-oxo-7,8-dihydropyrido[2,3-d]pyrimidin-2-yl)amino)pyridin-3-yl)piperazin-1-yl)-N-(2-(2-
1
2
(
2-(2-((2-(2,6-dioxopiperidin-3-yl)-1,3-dioxoisoindolin-4-yl)amino)ethoxy)ethoxy)ethoxy)ethyl)acetamide (14). To a solution of the
1 (23.7 mg, 0.0422 mmol) in DCM (2 mL) was added TFA (1 mL) and the resulting solution was stirred for 1h at room temperature.
The mixture was concentrated to give 12 which was used directly for the next step without further purification. 13 (23.1 mg,
.0422 mmol), which was synthesized as previously described (Olson et al., 2018), was also dissolved in DCM (2 mL), TFA
1 mL) was added and the mixture was kept stirring for 1h at room temperature. Then the solvent was evaporated and the residue
1
0
(
was re-dissolved in DMF (2mL). To the DMF solution were added 12, HATU (33 mg, 0.0844 mmol) and DIPEA (37 mL, 0.211 mmol).
The resulting mixture was stirred for 1h at room temperature, then evaporated the solvent and purified by reverse phase HPLC
1
(
5-95% MeOH in H
DMSO-d
.93 (d, J = 9.4 Hz, 1H, AR-H), 7.57 (dd, J = 8.6, 7.1 Hz, 1H, AR-H), 7.13 (d, J = 8.6 Hz, 1H, AR-H), 7.02 (d, J = 7.0 Hz, 1H,
AR-H), 5.90-5.75 (m, 1H, N-CH), 5.05 (dd, J = 12.8, 5.4 Hz, 1H, N-CH), 4.05 (s, 2H, CO-CH -N), 3.73-3.59 (m, 4H, 2xO-CH ),
.59-3.54 (m, 4H, 2xO-CH ), 3.53-3.50 (m, 4H, 2xO-CH ), 3.49-3.43 (m, 6H, 3xN-CH ), 3.43-3.25 (m, 6H, 3xN-CH ), 2.94-2.84
m, 1H, H’N-C-CH2), 2.63-2.51 (m, 2H, CO-CH ), 2.44 (s, 3H, CO-CH ), 2.35 (s, 3H, AR-CH ), 2.28-2.15 (m, 2H, C-CH ), 2.08–
.00 (m, 1H, H’N-C-CH2), 1.99-1.91 (m, 2H, C-CH ), 1.86-1.75 (m, 2H, C-CH ), 1.65-1.53 (m, 2H, C-CH ). C NMR (126 MHz,
2
O) to give 14 (TFA salt) as a yellow solid (34.4 mg, 77%). LC-MS: m/z 936 [M+1]. H NMR (500 MHz,
6
) d 9.05 (s, 1H, AR-H), 8.90 (t, J = 5.6 Hz, 1H, NH), 8.19 (dd, J = 9.6, 2.8 Hz, 1H, AR-H), 8.12 (d, J = 2.9 Hz, 1H, AR-H),
7
2
2
3
2
2
2
2
(
2
3
3
2
1
3
2
2
2
2
e7 Cell Chemical Biology 26, 1–7.e1–e9, February 21, 2019

Please cite this article in press as: Brand et al., Homolog-Selective Degradation as a Strategy to Probe the Function of CDK6 in AML, Cell Chemical
Biology (2018), https://doi.org/10.1016/j.chembiol.2018.11.006
DMSO) d 202.22, 172.82, 170.07, 168.94, 167.27, 163.93, 160.54, 157.29, 157.19, 154.95, 146.39, 143.30, 141.74, 141.59, 136.25,
1
5
32.07, 130.81, 117.46, 116.43, 110.71, 109.23, 108.31, 72.16, 70.52, 69.81, 69.77, 69.60, 68.89, 68.71, 60.18, 55.90, 53.12,
0.89, 48.57, 44.97, 43.62, 41.71, 31.27, 30.99, 27.68, 25.27, 22.16, 13.79.
QUANTIFICATION AND STATISTICAL ANALYSIS
Statistical parameters are reported in the figure legends of the paper.
Expression Proteomics BSJ-03-123
Acquired raw data files were processed using the Proteome Discoverer 2.2.0. platform, utilizing the Sequest HT database search
engine and Percolator validation software node (V3.04) to remove false positives with a false discovery rate (FDR) of 1% on peptide
and protein level under strict conditions. Searches were performed with full tryptic digestion against the human SwissProt database
v2017.06 with up to two miscleavage sites. Oxidation (+15.9949 Da) of methionine was set as variable modification, whilst carbami-
domethylation (+57.0214 Da) of cysteine residues and TMT 6-plex labelling of peptide N-termini and lysine residues were set as fixed
modifications. Data was searched with mass tolerances of ±10 ppm and 0.6 Da on the precursor and fragment ions (CID), respec-
tively. Results were filtered to include peptide spectrum matches (PSMs) with Sequest HT cross-correlation factor (Xcorr) scores
of R1 and high peptide confidence. PSMs with precursor isolation interference values of R 50% and average TMT-reporter ion
signal-to-noise values (S/N) % 10 were excluded from quantitation. Isotopic impurity correction and TMT channel-normalization
based on total peptide amount were applied. For statistical analysis and p-value calculation, the integrated ANOVA hypothesis
test was used. TMT ratios with p-values below 0.01 were considered as significant.
Expression Proteomics YKL-06-102
Proteome Discoverer 2.1 (Thermo Fisher Scientific) was used for .RAW file processing and controlling peptide and protein level false
discovery rates, assembling proteins from peptides, and protein quantification from peptides. MS/MS spectra were searched against
a Uniprot human database (September 2016) with both the forward and reverse sequences. Database search criteria are as follows:
tryptic with two missed cleavages, a precursor mass tolerance of 20 ppm, fragment ion mass tolerance of 0.6 Da, static alkylation of
cysteine (57.02146 Da), static TMT labelling of lysine residues and N-termini of peptides (229.16293 Da), and variable oxidation of
methionine (15.99491 Da). TMT reporter ion intensities were measured using a 0.003 Da window around the theoretical m/z for
each reporter ion in the MS3 scan. Peptide spectral matches with poor quality MS3 spectra were excluded from quantitation
(
summed signal-to-noise across 10 channels < 200 and precursor isolation specificity < 0.5), and resulting data was filtered to
only include proteins that had a minimum of 3 unique peptides identified. Reporter ion intensities were normalised and scaled using
in-house scripts in the R framework. Statistical analysis was carried out using the limma package within the R framework.
Phosphoproteomics
Acquired raw data files were processed using the Proteome Discoverer 2.2.0. platform, utilising the Sequest HT database search
engine and Percolator validation software node (V3.04) to remove false positives with a false discovery rate (FDR) of 1% on peptide
and protein level under strict conditions. Searches were performed with full tryptic digestion against the human SwissProt database
v2017.12 with up to two miscleavage sites. Oxidation (+15.9949 Da) of methionine and phosphorylation (+79.9663 Da) of serine,
threonine and tyrosine was set as variable modification, whilst carbamidomethylation (+57.0214 Da) of cysteine residues and
TMT 10-plex labelling of peptide N-termini and lysine residues were set as fixed modifications. Data was searched with mass toler-
ances of ±10 ppm and 0.025 Da on the precursor and fragment ions (HCD), respectively. Results were filtered to include peptide
spectrum matches (PSMs) with Sequest HT cross-correlation factor (Xcorr) scores of R1 and high peptide confidence. The ptmRS
algorithm was additionally used to validate phospopeptides with a set score cutoff of 90. PSMs with precursor isolation interference
values of R 50% and average TMT-reporter ion signal-to-noise values (S/N) % 10 were excluded from quantitation. Isotopic impurity
correction and TMT channel-normalization based on total peptide amount were applied. For statistical analysis and p-value calcu-
lation, the integrated ANOVA hypothesis test was used. TMT ratios with p-values below 0.01 were considered as significant.
Hits identified via global phosphoproteomics (log2 fold change > 0.5 or < -0.5, FDR adj. p-value < 0.05) were mapped on a protein-
protein interaction network using the networkx python package. The protein-protein interaction network was queried from the Hu-
man Integrated Protein-Protein Interaction rEference (HIPPIE) (Alanis-Lobato et al., 2017). Interactions without PubMed IDs were
removed. The resulting subgraph was expanded to include first order neighbors limited to transcriptional regulators for which
ENCODE genome localization data was available and visualized using Cytoscape. Kinase substrate prediction was performed
with KSEA App (https://casecpb.shinyapps.io/ksea).
RNA Sequencing
NGS reads were trimmed based on quality and adapter sequence content with Trimmomatic in single-end (ILLUMINACLIP:TruSeq3-
SE.fa:2:30:10:1:true, SLIDINGWINDOW:4:15, MINLEN:20) mode. The resulting reads were aligned with the ‘‘Spliced Transcripts
Alignment to a Reference’’ (STAR) aligner to the hg38 reference genome assembly provided by the UCSC Genome Browser resem-
bling the Genome Reference Consortium GRCh38 assembly. Ensembl transcript annotation from version e87 (December 2016)
served as reference transcriptome. Reads were counted using htseq-count. Differential expression analysis was performed on
Cell Chemical Biology 26, 1–7.e1–e9, February 21, 2019 e8

Please cite this article in press as: Brand et al., Homolog-Selective Degradation as a Strategy to Probe the Function of CDK6 in AML, Cell Chemical
Biology (2018), https://doi.org/10.1016/j.chembiol.2018.11.006
quantile-normalized read counts using the voom-limma R package. Following thresholds were applied for hit calling: fold change > 1.5
or < -1.5, FDR adj. p-value < 0.05. GO-term enrichment analysis was performed using GOrilla. For the network visualization of en-
riched GO-terms, parent-child relationships of terms were extracted from the go-basic.obo ontology and the network assembled
using the python networkx package.
DATA AND SOFTWARE AVAILABILITY
The accession number for the RNA sequencing data reported in this paper is GEO: GSE116187.
e9 Cell Chemical Biology 26, 1–7.e1–e9, February 21, 2019