Optimization of activity-based probes for proteomic profiling of histone deacetylase complexes

Article abstract of DOI:10.1021/ja074138u

Histone deacetylases (HDACs) are key enzymatic regulators of the epigenome and serve as promising targets for anticancer therapeutics. Recently, we developed a photoreactive "clickable" probe, SAHA-BPyne, to report on HDAC activity and complex formation i

Full text of DOI:10.1021/ja074138u

Published on Web 01/25/2008
Optimization of Activity-Based Probes for Proteomic Profiling
of Histone Deacetylase Complexes
Cleo M. Salisbury and Benjamin F. Cravatt*
The Skaggs Institute for Chemical Biology and Department of Chemical Physiology, The Scripps
Research Institute, 10550 North Torrey Pines Road, La Jolla, California 92037
Received June 6, 2007; E-mail: cravatt@scripps.edu
Abstract: Histone deacetylases (HDACs) are key enzymatic regulators of the epigenome and serve as
promising targets for anticancer therapeutics. Recently, we developed a photoreactive “clickable“ probe,
SAHA-BPyne, to report on HDAC activity and complex formation in native biological systems. Here, we
investigate the selectivity, sensitivity, and inhibitory properties of SAHA-BPyne and related potential activity-
based probes for HDACs. While we identified several probes that are potent HDAC inhibitors and label
HDAC complex components in native proteomic preparations, SAHA-BPyne was markedly superior for
profiling HDAC activities in live cells. Interestingly, the enhanced performance of SAHA-BPyne as an in
situ activity-based probe could not be solely ascribed to potency in HDAC binding, implying that other
features of the molecule were key to efficient active site-directed labeling in living systems. Finally, we
demonstrate the value of in situ profiling of HDACs by comparing the activity and expression of HDAC1 in
cancer cells treated with the cytotoxic agent parthenolide. These results underscore the utility of activity-
based protein profiling for studying HDAC function and may provide insight for the future development of
click chemistry-based photoreactive probes for the in situ anaylsis of additional enzyme activities.
substrate specificity.4-6 Studies are further confounded by
difficulties in reconstituting HDAC activities in purified,
Introduction
The human genome encodes ∼20 histone deacetylases
(HDACs), a family of hydrolytic enzymes that catalyze the
removal of acetyl groups from the side chain of lysines found
in histone (and other) proteins. Since the downstream effects
of histone acetylation include changes to chromatin conforma-
tion, an increase in apoptosis and differentiation, and a decrease
in cell growth, there is a growing interest in the development
of HDAC inhibitors as anticancer agents.1-3 A number of small
molecule HDAC inhibitors have been reported (Figure 1), some
of which induce differentiation in cancer cell lines and reduce
tumor volume in ViVo.1-3 Recently, suberoylanilide hydroxamic
acid (SAHA, Vorinostat, 1, Figure 1) was approved by the FDA
for the treatment of cutaneous T cell lymphoma, and several
other HDAC inhibitors are in clinical development.1-3
recombinant systems.3,7 New methods are thus needed to
monitor the functional state of HDAC complexes in native
biological systems. Recently, we developed an active-site
directed chemical probe, SAHA-BPyne (5, Figure 2), for
profiling HDAC activities in proteomes and live cells.8
The previously reported SAHA-BPyne was designed based
on the scaffold of SAHA, a reversible HDAC inhibitor with a
hydroxamic acid moiety that chelates to the metal cation in
HDAC active sites to provide high nanomolar to low micromolar
inhibition.2 To convert SAHA into an irreversible probe
compatible with activity-based protein profiling (ABPP)9,10
experiments, we introduced a benzophenone group into SAHA-
BPyne, which can be photoactivated to promote covalent
labeling of proximal proteins. We also appended an alkyne
moiety onto SAHA-BPyne to facilitate click chemistry-mediated
conjugation of reporter tags for the rapid and sensitive detection
and affinity enrichment of probe-labeled proteins.11,12 Here, we
investigate in more detail the proteome labeling profiles and
Despite the intense pursuit of HDAC-targeted therapeutics,
we do not yet fully understand the connection between HDACs
and cancer on a molecular level. Many of the HDAC inhibitors
in clinical development, including SAHA, target several of the
11 known metal-dependent (formally Zn2+-dependent) HDACs
in humans. Efforts to rigorously understand which HDACs are
critical for cancer pathogenesis are complicated by the fact that
most of these enzymes are found in a variety of large
multisubunit complexes that regulate both HDAC activity and
(4) You, A.; Tong, J. K.; Grozinger, C. M.; Schreiber, S. L. Proc. Natl. Acad.
Sci. U.S.A. 2001, 98, 1454-8.
(5) Zhang, Y.; Ng, H. H.; Erdjument-Bromage, H.; Tempst, P.; Bird, A.;
Reinberg, D. Genes DeV. 1999, 13, 1924-35.
(6) Pflum, M. K.; Tong, J. K.; Lane, W. S.; Schreiber, S. L. J. Biol. Chem.
2001, 276, 47733-41.
(7) Curtin, M.; Glaser, K. Curr. Med. Chem. 2003, 10, 2373-92.
(8) Salisbury, C. M.; Cravatt, B. F. Proc. Natl. Acad. Sci. U.S.A. 2007, 104,
1171-6.
(1) Marks, P. A.; Breslow, R. Nat. Biotechnol. 2007, 25, 84-90.
(9) Evans, M. J.; Cravatt, B. F. Chem. ReV. 2006, 106, 3279-301.
(10) Jessani, N.; Cravatt, B. F. Curr. Opin. Chem. Biol. 2004, 8, 54-9.
(11) Speers, A. E.; Adam, G. C.; Cravatt, B. F. J. Am. Chem. Soc. 2003, 125,
4686-7.
(2) Bolden, J. E.; Peart, M. J.; Johnstone, R. W. Nat. ReV. Drug DiscoVery
2006, 5, 769-84.
(3) Dokmanovic, M.; Clarke, C.; Marks, P. A. Mol. Cancer Res. 2007, 5, 981-
89.
(12) Speers, A. E.; Cravatt, B. F. Chem. Biol. 2004, 11, 535-46.
9
2184
J. AM. CHEM. SOC. 2008, 130, 2184-2194
10.1021/ja074138u CCC: $40.75 © 2008 American Chemical Society

Activity Probes for Profiling HDAC Complexes
A R T I C L E S
Figure 1. Structures of representative HDAC inhibitors.
Figure 2. Structures of candidate ABPP probes for HDACs. Stereochemistry of benzophenone amino acid in compounds 9-18 (asterisks) is noted next to
the compound numbers.
sonicating and Dounce homogenizing cell pellets in PBS, followed by
centrifugation at 100 000 × g (45 min) to provide soluble cellular
proteome fractions (supernatant) and a membrane pellet. Nuclear
fractions were prepared by first adding cold 10 mM HEPES pH 7.7,
10 mM KCl, 0.1 mM EDTA, and 0.4% IGEPAL CA-630 to plates of
washed cells. After 10 min at rt, the lysed cells were centrifuged at
15 000 × g for 3 min and the soluble fractions removed. The pellet
was resuspended in 20 mM HEPES pH 7.7, 0.4 M NaCl, 10% glycerol
by vortexing for 2 h at 4 °C, followed by centrifugation at 15 000 ×
g (5 min, 4 °C) to provide the nuclear fractions (supernatant) and a
membrane pellet. All fractions were stored at -80 °C until use.
inhibitory properties of SAHA-BPyne and demonstrate that this
probe has a rare ability, relative to closely related analogues, to
label HDACs in living cells. Furthermore, we demonstrate that
SAHA-BPyne, and more generally, the ABPP technology,
provides a unique perspective for understanding the interplay
between enzyme abundance and enzyme activity in living cells.
Experimental Section
Chemicals. Tris(2-carboxyethyl) phosphine (TCEP) was purchased
from Fluka (St. Louis, MO). The click-chemistry ligand, tris[(1-benzyl-
1H-1,2,3-triazol-4-yl)methyl]amine (“ligand”), was purchased from
Aldrich (St. Louis, MO). Parthenolide was purchased from Tocris
(Ellisville, MO). Resins, amino acids, and reagents for solid-phase
peptide synthesis were purchased from EMD Biosciences (San Diego,
CA) and Bachem (Torrance CA). SAHA (1)8, SAHA-BPyne (5)8, and
rhodamine-azide11 were prepared as described.
Cell Culture. The human uveal melanoma cell line MUM-2B
(provided by Mary Hendrix at Northwestern University) was grown in
RPMI medium 1640 containing 10% FCS. HCT-15 colon cancer cells
and MDA-MB-231 breast cancer cells (both from the American Type
Culture Collection) were grown in Advanced RPMI and DMEM media,
respectively, with 10% FCS. Whole cell lysate was prepared by
sonicating cell pellets in PBS. Soluble fractions were prepared by
In Vitro Proteome Labeling and Analysis. Proteome samples were
diluted in PBS to 50 µL at a final concentration of 1 mg/mL protein (4
mg/mL for whole cell lysate). SAHA-BPyne was added to samples at
a final concentration of 50 nM-5 µM in the presence or absence of
excess SAHA competitor (final concentration of 1 µM-50 µM). All
samples were then placed into the wells of an ice-cooled 96-well plate
and irradiated at 365 nm using a Spectroline ENF 260C UV lamp for
1 h on ice. After UV cross-linking, rhodamine-azide (12.5-100 µM
final concentration) was added, followed by TCEP (0.5 mM final
concentration) and ligand (100 µM final concentration). Samples were
gently vortexed, and the cycloaddition was initiated by the addition of
CuSO4 (1 mM final concentration). The click chemistry reactions were
9
J. AM. CHEM. SOC. VOL. 130, NO. 7, 2008 2185

A R T I C L E S
Salisbury and Cravatt
incubated at rt for 1 h. An equal volume of 2× standard SDS loading
buffer was added, and the samples were separated by 1D SDS-PAGE
and visualized by in-gel fluorescent scanning using a Hitachi FMBio
Ile flatbed scanner (MiraiBio, Alameda CA).
purified using preparatory C18 reversed-phase HPLC (tR: 20.7 min)
to afford 7.7 mg (58%) of probe 6 as a yellow solid. tR (analytical):
13.2 min. 1H NMR (500 MHz) δ 1.62-1.65 (m, 4H), 1.84 (quintet, J
) 7.1, 2H), 2.04 (t, J ) 6.7, 2H), 2.30 (td, J ) 7.0, 2.5, 2H), 2.42 (t,
J ) 6.7, 2H), 2.53-2.56 (m, 2H), 2.89 (t, J ) 2.6, 1H), 7.75-7.83
(m, 8H), 10.20-10.42 (m, 3H). 13C NMR (125 MHz) δ 18.21, 24.65,
25.58, 27.70, 33.02, 36.07, 37.19, 72.57, 84.83, 119.05, 119.08, 131.75,
131.92, 132.55, 132.59, 143.88, 143.94, 172.09, 172.32, 172.58, 194.23.
HR-MS m/z calcd for C25H27N3O5 (M + H): 450.2023. Found:
450.2011.
In Situ Labeling of Cancer Cells. MUM2B cells, grown to 90%
confluency in 6 cm dishes, were washed with PBS (3 × 3 mL) and
treated with 1 mL of PBS containing probe (50 nM-25 µM), with or
without excess SAHA (10×). The samples were irradiated at 365 nm
at 37 °C for 6 min (13 min for concentration gradient study) by placing
a Spectroline ENF 260C UV approximately 1′′ above the covered dish
of cells. After irradiation, the cells were washed with PBS (3 × 3 mL),
scraped, and pelleted by centrifugation. The cell pellet was resuspended
in PBS (50 µL), homogenized by sonication, and diluted to 2.5 mg/
mL with PBS (4 mg/mL for concentration gradient). To initiate the
click chemistry reaction, rhodamine azide (100 µM final concentration),
TCEP (0.5 mM final concentration), ligand (100 µM final concentra-
tion), and CuSO4 (1 mM final concentration) were added. The reactions
were incubated at rt for 1 h. An equal volume of 2× loading buffer
was added, and the samples were separated by 1D SDS-PAGE and
visualized and by in-gel fluorescent scanning.
Parthenolide-treated cells were grown to ∼70% confluency, and the
media was replaced with media containing 15 µM parthenolide. After
incubating the cells for 3 h at 37 °C, the cells were washed in cold
PBS and treated with SAHA-BPyne (2 µM) with or without excess
SAHA (40 µM) using the above conditions. After click chemistry, SDS-
PAGE separation, and scanning as described above, labeled proteins
were quantified by measuring the integrated band intensities.
Probe 7 (CSIV.109). 4-(4-Formyl-3-methoxyphenoxy)butyryl (FMPB)
AM resin (50 mg, 0.045 mmol) was added to a fritted syringe and
swollen with dichloroethane (DCE, 1 mL). Monotrityl-protected 1,6-
diaminohexane (188 mg, 0.450 mmol) in 2:1 DCE/trimethylorthofor-
mate (TMOF) (0.5 mL) was added to the resin, followed by NaBH-
(OAc)3 (85 mg, 0.45 mmol). The resin was agitated overnight and then
successively washed with DMF, 10% i-Pr2EtN in DMF, MeOH, THF,
and DMF (1 × 1 mL each). A solution of Fmoc-4-benzoyl-L-
phenylalanine (Fmoc-L-Bpa-OH, 108 mg, 0.220 mmol), O-(7-azaben-
zotriazol-1-yl)-1,1,3,3-tetramethyluroniumhexafluorophosphate (HATU,
84 mg, 0.22 mmol), and i-Pr2EtN (74 µL, 0.45 mmol) in DMF (0.5
mL) was added to the resin and agitated for 4 h. The resin was then
filtered and washed with DMF (3 × 1 mL), the Fmoc group was
removed (subjected to 20% piperidine in DMF for 30 min), and the
resin was washed again (3 × 1 mL DMF). A solution of hexynoic
acid (24 µL, 0.22 mmol), HATU (84 mg, 0.22 mmol), and i-Pr2EtN
(74 µL, 0.45 mmol) in DMF (0.5 mL) was added to the resin and
agitated for 4 h. The resin was then filtered and washed successively
with DMF, THF, MeOH, THF, and CH2Cl2 (3 × 1 mL). The trityl
group was removed by treatment with 2% TFA in CH2Cl2 (10 × 2 mL
× 10 min), and the resin was washed with CH2Cl2 and DMF (3 × 1
mL). A solution of Fmoc-PABA-OH (79 mg, 0.22 mmol), HATU (84
mg, 0.22 mmol), and i-Pr2EtN (74 µL, 0.45 mmol) in DMF (0.5 mL)
was added to the resin and agitated for 4 h. The resin was then filtered
and washed with DMF (3 × 1 mL), the Fmoc group was removed,
and the resin was washed successively with DMF, THF, MeOH, THF,
and CH2Cl2 (3 × 1 mL). A solution of methyl-8-chloro-8-oxooctanoate
(51 µL, 0.36 mmol) and i-Pr2EtN (60 µL, 0.36 mmol) in CH2Cl2 (0.5
mL) was added to the resin and agitated overnight. The acylation was
repeated, and the resin was washed successively with DMF, THF,
MeOH, THF, and CH2Cl2 (3 × 1 mL). The methyl ester analogue of
the probe was cleaved from support by treatment with 97.5% TFA/
2.5% H2O (1 mL), and the eluant was concentrated by rotary
evaporation. DMF (300 µL), THF (300 µL), and 50% H2NOH in water
(300 µL, 5 mmol) were added to the residue, and the reaction stirred
for 48 h. The THF was removed in Vacuo, and the remainder of the
sample was purified using preparatory C18 reversed-phase HPLC to
Western Blotting. All samples were separated on SDS-PAGE gels
and transferred to nitrocellulose. The blots were blocked with 3% milk
in TBST, probed with polyclonal R-HDAC1 (1:800 in TBST, 06-720,
Upstate, Lake Placid NY) for 2 h at rt, washed with TBST, incubated
with goat-R-rabbit IgG (H + L) HRP conjugate (1:20000, Biorad,
Hercules CA) for 2 h at rt, washed again in TBST, and visualized with
a SuperSignal West Pico kit (Pierce, Rockford IL).
Alternate Probe Synthesis. Unless otherwise noted, chemicals were
obtained from commercial suppliers and used without further purifica-
tion. Dry THF and CH2Cl2 were obtained by passing commercially
available predried, oxygen-free formulations through activated alumina
columns. NMR spectra were obtained in DMSO-d6 and were recorded
on a Bruker DRX-500 instrument unless otherwise noted. NMR
chemical shifts are reported in ppm downfield relative to the internal
solvent peak, and J values are in Hz. High-resolution mass spectrometry
data (HRMS) were performed at The Scripps Research Institute Mass
Spectrometry Core and recorded on an Agilent mass spectrometer using
ESI-TOF (electrospray ionization-time-of-flight). Preparatory HPLC
was performed with a C18 column (Higgins Analytical, 5 µm, 150
mm × 10 mm) gradient of CH3CN/H2O-0.1% TFA, 5-66% over 50
min, 5 mL/min, with detection at 280 nm for 70 min. Analytical HPLC
was performed with a C18 column (Higgins Analytical, 5 µm, 150
mm × 4.6 mm) gradient of CH3CN/H2O-0.1% TFA, 5-95% over 17
min, 0.5 mL/min, with detection at 220/254/280 nm for 22 min.
Probe 6 (CSIV.87). To a flame-dried 2 dram vial fitted with a stir
bar were added N-(4-(4-aminobenzoyl)phenyl)hex-5-ynamide8 (35 mg,
0.11 mmol), i-Pr2EtN (38 µL, 0.22 mmol), and THF (700 µL). The
flask was cooled to 0 °C in an ice bath, and methyl adipoyl chloride
(24 µL, 0.14 mmol) was added. The mixture was allowed to warm to
rt and stirred overnight. CH2Cl2 (5 mL) was added to the reaction, and
the organic layer was washed with NaHCO3 (sat., 5 mL) and 10% citric
acid (5 mL). The organic layer was dried (Na2SO4) and filtered, and
the solvent was removed by rotary evaporation to provide the methyl
ester analogue of the probe as a yellow solid (32 mg, 65% yield, HR-
MS m/z calcd for C26H28N2O5 (M + H): 449.2071. Found: 449.2052).
To a 2 dram vial fitted with a stir bar were added the methyl ester
intermediate (13 mg, 0.03 mmol), DMF (300 µL), THF (300 µL), and
50% H2NOH in water (300 µL, 5 mmol). After stirring overnight, the
THF was removed in Vacuo, and the remainder of the sample was
1
afford 2.1 mg (6%) of probe 7. tR (analytical): 14.2 min. H NMR
(500 MHz) δ 1.28-1.65 (m, 18H), 1.98-2.06 (m, 4H), 2.20 (td, J )
7.4, 1.7, 2H), 2.37 (t, J ) 7.4, 2H), 2.60-2.62 (m, 1H), 2.78-2.80
(m, 1H), 2.89-2.95 (m, 2H), 3.06-3.14 (m, 3H), 3.27 (q, J ) 6.7,
2H), 4.58-4.63 (m, 1H), 7.46 (d, J ) 8.2, 2H), 7.61 (t, J ) 7.7, 2H),
7.69-7.76 (m, 6H), 7.83 (d, J ) 7.8, 2H), 8.02 (t, J ) 5.5, 1H), 8.19
(d, J ) 8.6, 1H), 8.33 (t, J ) 5.6, 1H), 10.12 (s,1H), 10.38 (s, 1H).
HR-MS m/z calcd for C43H53N5O7 (M + H): 752.4018. Found:
752.4005.
General Protocol for Preparation of Probes 8, 17, and 18. FMPB
AM resin (50 mg, 0.045 mmol) was added to a fritted syringe and
swollen with DCE (1 mL). 6-Aminohexyne hydrochloride13 (61 mg,
0.45 mmol) in 2:1 DCE/TMOF (0.5 mL) was added to the resin,
followed by NaBH(OAc)3 (85 mg, 0.45 mmol). The resin was agitated
overnight and then successively washed with DMF, 10% i-Pr2EtN in
DMF, MeOH, THF, and DMF (1 × 1 mL each). A solution of Fmoc-
Bpa-OH (L for probes 8 and 18, D for probe 17, 108 mg, 0.220 mmol),
(13) Muller, T. E.; Lercher, J. A.; Van Nhu, N. AIChE J. 2003, 49, 214-24.
9
2186 J. AM. CHEM. SOC. VOL. 130, NO. 7, 2008

Activity Probes for Profiling HDAC Complexes
A R T I C L E S
HATU (84 mg, 0.22 mmol), and i-Pr2EtN (74 µL, 0.45 mmol) in DMF
(0.5 mL) was added to the resin and agitated for 4 h. The resin was
then filtered and washed with DMF (3 × 1 mL), the Fmoc group was
removed (subjected to 20% piperidine in DMF for 30 min), and the
resin was washed again (3 × 1 mL DMF). [For probe 8, a solution of
Fmoc-PABA-OH (79 mg, 0.22 mmol), HATU (84 mg, 0.22 mmol),
and i-Pr2EtN (74 µL, 0.45 mmol) in DMF (0.5 mL) was added to the
resin and agitated for 4 h. The resin was then filtered and washed with
DMF (3 × 1 mL), the Fmoc group was removed, and the resin was
washed successively with DMF, THF, MeOH, THF, and CH2Cl2 (3 ×
1 mL).] A solution of methyl-8-chloro-8-oxooctanoate (51 µL, 0.36
mmol) and i-Pr2EtN (60 µL, 0.36 mmol) in CH2Cl2 (0.5 mL) was added
to the resin and agitated overnight. The acylation was repeated, and
the resin was washed successively with DMF, THF, MeOH, THF, and
CH2Cl2 (3 × 1 mL). The methyl ester analogue of the probe was cleaved
from support by treatment with 95% TFA/5% CH2Cl2 (1 mL), and the
eluant was concentrated by rotary evaporation. DMF (300 µL), THF
(300 µL), and 50% H2NOH in water (300 µL, 5 mmol) were added to
the residue, and the reaction stirred for 48 h. The THF was removed
in Vacuo, and the remainder of the sample was purified using
preparatory C18 reversed-phase HPLC.
0.52 mmol) in DMF (0.5 mL) was added to the resin and agitated
overnight. The resin was then filtered and washed with DMF (3 × 1
mL), the Fmoc group was removed, and the resin was washed again
(3 × 1 mL DMF). A solution of 6-hexynoic acid (36 µL, 0.33 mmol),
HATU (125 mg, 0.33 mmol), and i-Pr2EtN (109 µL, 0.66 mmol) in
DMF (0.5 mL) was added to the resin and agitated for 4 h. The acylation
was repeated overnight, and the resin was then filtered and washed
successively with DMF, THF, MeOH, THF, and CH2Cl2 (3 × 1 mL).
The probes were cleaved from support by treatment with 97.5% TFA/
2.5% H2O, and the eluant was concentrated and then purified using
preparatory C18 reversed-phase HPLC.
Probe 9 (CSIV.102A). The general procedure provided 4.1 mg
1
(15%) of probe 9. tR(analytical): 13.8 min. H NMR (500 MHz) δ
1.60-1.66 (m, 2H), 2.06 (td, J ) 7.2, 2.6, 2H), 2.23 (td, J ) 7.4, 3.3,
2H), 2.79 (t, J ) 2.6, 1H), 2.94-2.98 (m, 1H), 3.19 (dd, J ) 13.7, 5.2,
1H), 4.38 (d, J ) 5.9, 2H), 4.67-4.72 (m, 1H), 7.31 (d, J ) 8.2, 2H),
7.50 (d, J ) 8.2, 2H), 7.61-7.64 (m, 2H), 7.71-7.77 (m, 7H), 8.29
(d, J ) 8.4, 1H), 8.64 (t, J ) 5.9, 1H). 13C NMR (125 MHz) δ 18.10,
25.13, 34.93, 42.732, 54.58, 72.25, 84.95, 100.01, 127.71, 127.76,
128.56, 129.40, 129.41, 130.27, 130.30, 130.32, 130.47, 133.37, 135.95,
138.17, 171.98, 172.39, 196.36. HR-MS m/z calcd for C30H29N3O5 (M
+ H): 512.218. Found: 512.2172.
Probe 8 (CSIV104). The general procedure provided 4.7 mg (37%)
1
of probe 8. tR(analytical): 14.8 min. H NMR (500 MHz) δ 1.33-
Probe 10 (CSIV.102B). The general procedure provided 4.2 mg
1
1.64 (m, 14H), 1.99 (t, J ) 7.0, 2H), 2.20 (m, 2H), 2.37 (t, J ) 7.4,
2H), 2.79 (m, 1H), 2.95 (m, 1H), 4.78 (m, 2H), 7.55-7.84 (m, 13H),
8.16 (br s,1H), 8.52 (br s,1H), 8.70 (br s,1H), 10.15 (s, 1H), 10.38 (s,
1H). HR-MS m/z calcd for C37H42N2O6 (M + H): 639.3177. Found:
639.3159.
(15%) of probe 10. tR(analytical): 13.8 min. H NMR (400 MHz) δ
1.52-1.60 (m, 2H), 1.99 (td, J ) 7.2, 2.4, 2H), 2.16 (td, J ) 7.3, 2.0,
2H), 2.74 (t, J ) 2.6, 1H), 2.89 (dd, 13.0, 10.2, 1H), 3.13 (m, 13.4,
4.8, 1H), 4.32 (d, J ) 5.9, 2H), 4.60-4.66 (m, 1H), 7.24 (d, J ) 8.1,
2H), 7.43 (d, J ) 8.2, 2H), 7.56 (t, J ) 7.5, 2H), 7.65-7.71 (m, 7H),
8.25 (d, J ) 8.5, 1H), 8.64 (t, J ) 5.7, 1H). 13C NMR (100 MHz) δ
17.25, 24.28, 34.07, 41.87, 53.74, 71.44, 84.10, 100.01, 126.87, 126.88,
128.56, 129.43, 129.49, 129.64, 131.28, 132.54, 135.09, 137.30, 142.53,
143.47, 171.15, 171.54, 195.52. HR-MS m/z calcd for C30H29N3O5 (M
+ H): 512.218. Found: 512.2171.
Probe 17 (CSIV.108B). The general procedure provided 8.2 mg
1
(36%) of probe 17. tR(analytical): 14.2 min. H NMR (500 MHz) δ
1.15-1.24 (m, 4H), 1.40-1.52 (m, 8H), 1.95 (t, J ) 7.4, 2H), 2.10 (t,
J ) 7.4, 2H), 2.19 (td, J ) 6.9, 2.6, 2H), 2.78 (t, J ) 2.7, 1H), 2.92
(dd, J ) 9.4, 4.1, 1H), 3.07-3.15 (m, 3H), 4.58-4.61 (m, 1H), 7.47
(d, J ) 8.2, 2H), 7.62 (t, J ) 7.7, 2H), 7.70-7.76 (m, 5H), 8.04 (t, J
) 5.6, 1H), 8.12 (d, J ) 8.6, 1H), 10.36 (s, 1H). 13C NMR (125 MHz)
δ 18.24, 25.83, 25.95, 26.13, 28.99, 29.11, 29.20, 33.10, 35.99, 38.77,
38.85, 54.42, 72.10, 85.22, 129.40, 130.25, 130.30, 130.41, 133.39,
135.88, 138.16, 144.30, 169.94, 171.63, 172.85, 196.35. HR-MS m/z
calcd for C30H37N3O5 (M + H): 520.2806. Found: 520.2800.
Probe 11 (CSIV.93). The general procedure provided 6.6 mg (24%)
1
of probe 11. tR(analytical): 14.3 min. H NMR (500 MHz) δ 1.62-
1.68 (m, 2H), 2.09 (td, J ) 7.3, 2.6, 2H), 2.25 (td, J ) 7.5, 2.1, 2H),
2.80 (q, J ) 2.6, 1H), 3.04 (dd, J ) 13.4, 9.9, 1H), 3.20-3.23 (dd, J
) 13.5, 4.9, 1H), 4.80-4.85 (m, 1H), 7.54 (d, J ) 8.2, 2H), 7.61 (t, J
) 7.7, 2H), 7.70-7.80 (m, 1H), 8.44 (d, J ) 8.2, 1H), 10.42 (s, 1H).
13C NMR (125 MHz) δ 18.09, 25.12, 34.79, 38.49, 55.40, 72.26, 84.92,
119.51, 128.55, 129.39, 130.28, 130.31, 130.33, 130.43, 133.40, 136.07,
138.11, 142.13, 143.82, 171.34, 172.58, 196.37. HR-MS m/z calcd for
C29H27N3O5 (M + H): 498.2023. Found: 498.2014.
Probe 12 (CSIV.94). The general procedure provided 4.8 mg (18%)
of probe 12. tR(analytical): 14.3 min. 1H NMR (400 MHz) δ 1.28
(quintet, J ) 7.2, 2H), 2.02 (td, J ) 7.2, 2.6, 2H), 2.25 (dd, J ) 7.2,
6.0, 2H), 2.75 (t, J ) 2.6, 1H), 2.96 (dd, J ) 13.4, 10.0, 1H), 3.14 (dd,
J ) 13.8, 5.0, 1H), 4.73-4.78(m, 1H), 7.47-7.73 (m, 13H), 8.40 (d,
J ) 8.2, 1H), 10.38 (s, 1H). 13C NMR (125 MHz) δ 18.10, 25.14,
34.81, 38.49, 55.41, 72.28, 84.94, 119.51, 128.58, 129.40, 130.29,
130.29,130.34, 130.44, 133.41, 136.08, 138.12, 142.16, 143.82, 171.34,
172.58, 196.38. HR-MS m/z calcd for C29H27N3O5 (M + H): 498.2023.
Found: 498.2007.
General Protocol for Preparation of Probes 13-16. A solution
of Fmoc-Amb-OH (31 mg, 0.084 mmol) (for probes 13 and 14) or
Fmoc-PABA-OH (30 mg, 0.08 mmol) (for probes 15 and 16) and i-Pr2-
EtN (46 µL, 0.28 mmol) in dry CH2Cl2 (0.5 mL) was added to a fritted
syringe containing 2-chlorotrityl resin (50 mg, 0.07 mmol). After 4 h,
the resin was filtered, quenched by the addition of i-Pr2EtN/MeOH/
CH2Cl2 (1:2:17), and washed with CH2Cl2 and DMF (3 × 1 mL). The
Fmoc group was removed (subjected to 20% piperidine in DMF for
30 min), and the resin was washed again (3 × 1 mL DMF). A solution
of Fmoc-Bpa-OH (172 mg, 0.350 mmol) (L for probes 13 and 15, D
for probes 14 and 16), HATU (133 mg, 0.350 mmol), and i-Pr2EtN
(116 µL, 0.700 mmol) in DMF (0.5 mL) was added to the resin and
agitated overnight. The resin was then filtered and washed with DMF
Probe 18 (CSIV.108A). The general procedure provided 5.9 mg
1
(26%) of probe 18. tR(analytical): 14.1 min. H NMR (500 MHz) δ
1.16-1.26 (m, 4H), 1.41-1.52 (m, 8H), 1.95 (t, J ) 7.3, 2H), 2.10 (t,
J ) 7.3, 2H), 2.19 (td, J ) 6.8, 2.6, 2H), 2.78 (t, J ) 2.1, 1H), 2.92
(dd, J ) 9.5, 3.9, 1H), 3.05-3.17 (m, 3H), 4.57-4.62 (m, 1H), 7.47
(d, J ) 8.2, 2H), 7.62 (t, J ) 7.7, 2H), 7.70-7.76 (m, 5H), 8.04 (t, J
) 5.7, 1H), 8.12 (d, J ) 8.6, 1H), 10.36 (s, 1H). 13C NMR (125 MHz)
δ 18.22, 25.81, 25.93, 26.10, 28.97, 29.09, 29.18, 33.08, 35.99, 38.75,
38.84, 54.40, 72.10, 85.20, 129.39, 130.24, 130.29, 130.40, 133.38,
135.87, 138.15, 144.30, 169.93, 171.62, 172.85, 196.34. HR-MS m/z
calcd for C30H37N3O5 (M + H): 520.2806. Found: 520.2799.
General Protocol for Preparation of Probes 9-12. Wang resin
(200 mg, 0.220 mmol) was modified to display a resin bound
O-hydroxylamine prepared as previously described.14 A portion of the
resin (60 mg) was added to a fritted syringe and swollen with CH2Cl2
(1 mL) for 1 h. A solution of Fmoc-Amb-OH (97 mg, 0.26 mmol) (for
probes 9 and 10) or Fmoc-PABA-OH (79 mg, 0.22 mmol) (for probes
11 and 12), HATU (99 mg, 0.26 mmol), and i-Pr2EtN (86 µL, 0.52
mmol) in DMF (0.5 mL) was added to the resin and agitated for 4 h.
The resin was then filtered and washed with DMF (3 × 1 mL), the
Fmoc group was removed (subjected to 20% piperidine in DMF for
30 min), and the resin was washed again (3 × 1 mL DMF). A solution
of Fmoc-Bpa-OH (162 mg, 0.33 mmol) (L for probes 9 and 11, D for
probes 10 and 12), HATU (99 mg, 0.26 mmol), and i-Pr2EtN (86 µL,
(14) Floyd, C. D.; Lewis, C. N.; Patel, S. R.; Whittaker, M. Tet. Lett. 1996, 37,
8045-8.
9
J. AM. CHEM. SOC. VOL. 130, NO. 7, 2008 2187

A R T I C L E S
Salisbury and Cravatt
(3 × 1 mL), the Fmoc group was removed, and the resin was washed
again (3 × 1 mL DMF). A solution of hexynoic acid (39 µL, 0.35
mmol), HATU (133 mg, 0.350 mmol), and i-Pr2EtN (116 µL, 0.700
mmol) in DMF (0.5 mL) was added to the resin and agitated for 4 h.
The acylation was repeated overnight, and the resin was then filtered
and washed successively with DMF, THF, MeOH, THF, and CH2Cl2
(3 × 1 mL). The probes were cleaved from support by treatment with
5% TFA/95% CH2Cl2 (1 mL), and the eluant was concentrated and
then purified using preparatory C18 reversed-phase HPLC. 1-Hydroxy-
benzotriazole (HOBt, 15 mg, 0.080 mmol) and DMF/DCE (1:6, 700
µL) were added to the residue. After the residue dissolved, 1-ethyl-3-
[3-dimethylaminopropyl]carbodiimide hydrochloride (EDC, 15 mg,
0.080 mmol) was added and the reaction stirred for 5 min. 1,2-
Phenylenediamine (76 mg, 0.70 mmol) was added to the reaction vessel,
and the reaction stirred overnight. Additional portions of HOBt, EDC,
and diamine were added to the reaction, and again the reaction stirred
overnight. EtOAc (10 mL) was added to the reaction, and the solution
was washed with NaHCO3 (sat aq., 2 × 5 mL) and 10% citric acid
(aq., 2 × 5 mL). The organic layer was dried (Na2SO4), filtered, and
concentrated by rotary evaporation. The crude probes were then purified
using preparatory C18 reversed-phase HPLC.
Probe 19 (CSIV.101). Wang resin (200 mg, 0.220 mmol) was
modified to display a resin bound O-hydroxylamine prepared as
previously described.14 A portion of the resin (60 mg) was added to a
fritted syringe and swollen with CH2Cl2 (1 mL) for 1 h. A solution of
Fmoc-aminohexanoic acid-OH (71 mg, 0.20 mmol), HATU (76 mg,
0.20 mmol), and i-Pr2EtN (66 µL, 0.40 mmol) in DMF (0.5 mL) was
added to the resin and agitated overnight. The resin was then filtered
and washed with DMF (3 × 1 mL), the Fmoc group was removed
(subjected to 20% piperidine in DMF for 30 min), and the resin was
washed again (3 × 1 mL DMF). A solution of Fmoc-PABA-OH (72
mg, 0.20 mmol), HATU (76 mg, 0.20 mmol), and i-Pr2EtN (66 µL,
0.40 mmol) in DMF (0.5 mL) was added to the resin and agitated for
4 h. The resin was then filtered and washed with DMF (3 × 1 mL),
the Fmoc group was removed, and the resin was washed again (3 × 1
mL DMF). A solution of Fmoc-L-Bpa-OH (98 mg, 0.20 mmol), HATU
(76 mg, 0.20 mmol), and i-Pr2EtN (66 µL, 0.40 mmol) in DMF (0.5
mL) was added to the resin and agitated overnight. The resin was then
filtered and washed with DMF (3 × 1 mL), the Fmoc group was
removed, and the resin was washed again (3 × 1 mL DMF). A solution
of 6-hexynoic acid (22 µL, 0.20 mmol), HATU (76 mg, 0.20 mmol),
and i-Pr2EtN (66 µL, 0.40 mmol) in DMF (0.5 mL) was added to the
resin and agitated overnight. The acylation was repeated overnight, and
the resin was then filtered and washed successively with DMF, THF,
MeOH, THF, and CH2Cl2 (3 × 1 mL). The probe was cleaved from
support by treatment with 97.5% TFA/2.5% H2O, and the eluant was
concentrated and then purified using preparatory C18 reversed-phase
Probe 13 (CSIV.100A). The general procedure provided 5.7 mg
1
(13%) of probe 13. tR(analytical): 15.1 min. H NMR (500 MHz) δ
1.61-1.67 (m, 2H), 2.06 (td, J ) 7.2, 2.6, 2H), 2.24 (td, J ) 7.4, 2.6,
2H), 2.79-2.80 (m, 1H), 2.95-3.00 (m, 1H), 3.20 (dd, J ) 13.7, 5.0,
1H), 4.43 (d, J ) 6.1, 2H), 4.69-4.73 (m, 1H), 6.95 (br s, 1H), 7.07
(d, J ) 7.8, 1H), 7.17 (t, J ) 7.3, 1H), 7.07 (t, J ) 8.2, 2H), 7.39 (d,
J ) 8.2, 2H), 7.51 (t, J ) 7.5, 2H), 7.62 (t, J ) 7.6, 2H), 7.72-7.77
(m, 5H), 7.91 (d, J ) 8.3, 1H), 7.99 (d, J ) 8.3, 1H), 8.32 (d, J ) 8.5,
1H), 8.71 (t, J ) 5.8, 1H), 9.97 (s,1H). 13C NMR (125 MHz) δ 18.09,
25.12, 34.93, 38.53, 42.72, 54.62, 72.24, 84.94, 127.44, 127.58, 127.67,
127.68, 127.90, 128.70, 129.36, 129.39, 130.26, 130.30, 130.32, 130.47,
133.36, 133.56, 135.95, 138.14, 143.99, 144.298, 172.02, 172.41,
182.09, 196.35. HR-MS m/z calcd for C36H34N4O4 (M + H): 587.2653.
Found: 587.2655.
1
HPLC (tR 14.2 min) to provide 5.6 mg (16%) of probe 19. H NMR
(500 MHz) δ 1.32-1.35 (m, 2H), 1.55-1.67 (m, 6H), 2.00-2.11 (m,
4H), 2.25 (td, J ) 7.3, 1.6, 2H), 2.80 (t, J ) 2.6, 1H), 3.01-3.06 (m,
1H), 3.19-3.30 (m, 3H), 4.80-4.84 (m, 1H), 7.54-7.62 (m, 4H), 7.70-
7.73 (m, 8H), 7.86 (d, J ) 8.7, 1H), 8.37-8.43 (m,2H), 10.39 (s, 1H).
13C NMR (125 MHz) δ 18.11, 25.14, 25.81, 27.02, 29.84, 33.13, 34.81,
38.51, 41.01, 55.43, 72.28, 84.93, 119.35, 128.84, 129.39, 130.28,
130.32, 130.34, 130.43, 133.40, 136.07, 138.11, 142.00, 143.83, 166.35,
169.94, 171.30, 172.58, 196.37. HR-MS m/z calcd for C35H38N4O6 (M
+ H): 611.2864. Found: 611.2856.
Probe 14 (CSIV.100B). The general procedure provided 4.7 mg
1
(11%) of probe 14. tR: 15.1 min. H NMR (500 MHz) δ 1.62-1.66
HDAC Activity Assay. HDAC activity was determined using a
fluorogenic substrate assay kit (Fluor de Lys, Biomol, Plymouth
Meeting, PA). In vitro assays were performed per manufacturer’s
instructions. HeLa nuclear lysate was diluted 1:30 with the provided
assay buffer (50 mM Tris-HCl, pH 8.0, 137 mM NaCl, 2.7 mM KCl,
1 mM MgCl2), and 15 µL were added to the well of a white 96 well
plate. Inhibitors diluted in assay buffer (10 µL of 0-900 µM inhibitor,
less than 10% DMSO) were added to nuclear lysate and preincubated
for 15 min at rt. The assay was initiated by the addition of substrate
(25 µL of 1.25 µM substrate in assay buffer). After 25 min at rt, the
assay was quenched with the addition of 50 µL of developer (1:40
developer solution, 2 µM TSA in assay buffer). The fluorescence of
the assay wells was detected at λex/λem 360/480 nm using a SPEC-
TRAmax plate reader (Molecular Devices, Sunnyvale CA). IC50 values
were determined by fitting the data from replicate trials (n ) 2 or 3)
to a sigmoidal dose-response curve using GraphPad Prism (GraphPad
Software, San Diego, CA).
(m, 2H), 2.06 (td, J ) 7.2, 2.6, 2H), 2.24 (td, J ) 7.4, 2.6, 2H), 2.95-
3.00 (m, 1H), 3.20 (dd, J ) 13.3, 5.2, 1H), 4.43 (d, J ) 5.7, 2H),
4.69-4.73 (m, 1H), 6.86 (br s, 1H), 7.01 (d, J ) 1.1, 1H), 7.11-7.15
(m, 1H), 7.29-7.39 (m, 3H), 7.50 (t, J ) 8.1, 2H), 7.62 (t, J ) 7.6,
2H), 7.72-7.77 (m, 5H), 7.91 (d, J ) 8.3, 1H), 7.99 (d, J ) 8.1, 1H),
8.31 (d, J ) 8.5, 1H), 8.70 (t, J ) 6.1, 1H), 9.88 (s, 1H). HR-MS m/z
calcd for C36H34N4O4 (M + H): 587.2653. Found: 587.2653.
Probe 15 (CSIV.97). The general procedure provided 3.9 mg (16%)
of probe 15. tR: 15.5 min. 1H NMR (500 MHz) δ 1.64-1.68 (m, 2H),
2.07-2.11 (m, 2H), 2.24-2.28 (m, 2H), 2.79-2.82 (m, 1H), 2.95 (s,
1H), 3.02-3.09 (m, 1H), 3.20-3.25 (m, 1H), 4.82-4.86 (m, 1H), 6.73-
6.76 (m, 1H), 6.90 (d, J ) 8.3, 1H), 7.07 (m, 1H), 7.19 (m, 2H), 7.25-
7.29 (m, 1H), 7.54-7.63 (m, 4H), 7.71-7.78 (m, 6H), 7.96 (d, J )
8.7, 1H), 8.01-8.04 (m, 1H), 8.45 (d, J ) 8.0, 1H), 9.72 (s,1H), 10.46-
10.51 (m, 1H). HR-MS m/z calcd for C35H32N4O4 (M + H): 573.2496.
Found: 573.2490.
Probe 16 (CSIV.99). The general procedure provided 6.3 mg (16%)
of probe 16. tR: 15.7 min. 1H NMR (500 MHz) δ 1.64-1.69 (m, 2H),
2.07-2.12 (m, 2H), 2.24-2.18 (m, 2H), 2.79-2.82 (m, 1H), 2.95 (s,
1H), 3.02-3.09 (m, 1H), 3.20-3.25 (m, 1H), 4.82-4.86 (m, 1H), 6.91-
6.93 (m, 1H), 7.05 (d, J ) 7.4, 1H), 7.16 (td, J ) 7.5, 1.0, 2H), 7.32
(d, J ) 7.3, 1H), 7.54-7.63 (m, 4H), 7.71-7.80 (m, 7H), 7.96 (t, J )
8.8, 1H), 8.04 (d, J ) 8.7, 1H), 8.46 (dd, J ) 8.1, 2.6, 1H), 9.90 (s,-
1H), 10.48-10.51 (m, 1H). 13C NMR (125 MHz) δ 18.11, 25.14, 34.81,
38.51, 55.48, 72.29, 84.94, 119.38, 119.488, 127.60, 129.38, 129.403,
129.68, 129.68, 130.30, 130.35, 130.35, 130.45, 130.45, 133.40, 133.43,
136.09, 138.11, 142.61, 143.80, 171.44, 172.62, 196.38. HR-MS m/z
calcd for C35H32N4O4 (M + H): 573.2496. Found: 573.2497.
Cytotoxicity Assay. Cell viability assays were performed by
distributing cells into the wells of a clear-bottomed 96-well plate (1000
cells/well). After 24 h, the media was replaced with media (100 µL/
well) containing 0-60 µM parthenolide (or vehicle alone) and incubated
for 4 h at 37 °C. After 4 h, WST-1 assay reagent (10 µL/well, Roche
Applied Science) was added. After an additional incubation of 2.5 h,
the absorbance of each well was detected at 420 nm using a
SPECTRAmax plate reader (Molecular Devices, Sunnyvale CA). The
background absorbance from 680 nm and from media + reagent alone
was subtracted from the absorbance of each well, and the average
corrected value from three independent trials was reported.
9
2188 J. AM. CHEM. SOC. VOL. 130, NO. 7, 2008

Activity Probes for Profiling HDAC Complexes
A R T I C L E S
Results and Discussion
Finally, we compared in Vitro proteome reactions with
labeling in live cells (in situ). SAHA-BPyne (50 nM-5 µM)
was added, either alone or in the presence of excess SAHA
(20×), to cultured preparations of MUM2B cells. The cells were
then irradiated with UV light, washed (to remove excess probe),
and homogenized. Treatment of the whole cell lysates with
rhodamine-azide under click chemistry conditions and analysis
by SDS-PAGE coupled with in-gel fluorescence scanning
revealed a number of SAHA-sensitive targets (Figure 3D). Here,
we observed clearly superior selectivity compared to in Vitro
labeling experiments, with low background even at the highest
concentrations of probe tested, and were able to detect a fifth
specific target of SAHA-BPyne at ∼125 kD. Similar results
were found with the breast cancer cell line MDA-MB-231, with
which we also observed improved sensitivity and selectivity of
HDAC labeling profiles in situ relative to in Vitro proteomic
preparations (data not shown). These data collectively indicate
that SAHA-BPyne is a versatile ABPP probe for HDAC
complexes capable of functioning in a variety of proteomic
fractions, including those as complicated as whole cell lysates,
and remarkably shows its most robust behavior in situ.
Optimization of Proteome Labeling Conditions for SAHA-
BPyne. In the original experiments with SAHA-BPyne, we used
a combination of gel- and mass-spectrometry-based approaches
to identify multiple HDAC enzymes, including HDACs 1, 2,
3, and 6, as specific targets of this probe.8 Furthermore, by virtue
of the benzophenone moiety, SAHA-BPyne was also found to
cross-link to and report on HDAC-associated proteins (e.g.,
CoREST, MBD3, and the metastasis associated proteins MTA1
and 2) bound proximally to the HDAC active site. While similar
levels of HDAC complex components were found in many of
the cell lines tested, some differences were observed. For
example, HDAC6 labeling was observed in higher levels in the
nonaggressive melanoma cell line, MUM2C, relative to its
invasive counterpart, MUM2B. Conversely, the levels of CoR-
EST in complex with HDACs were found to be higher in the
aggressive MUM2B cell line. These results were validated using
traditional western blotting methods, demonstrating that SAHA-
BPyne can accurately measure differences in both HDAC
content and complex assembly in native proteomes.
SAHA-BPyne was demonstrated to label HDAC complex
proteins both in proteomes at 100 nM and in live cells at 500
nM, but we wondered whether we could achieve improved
yields of protein labeling and a better depth of coverage by
varying the probe concentration. To this end, we studied the
effects of SAHA-BPyne concentration in the MUM2B cell line,
where specific, HDAC-related labeling events were defined as
those that could be blocked by the addition of excess SAHA
(Figure 3). SAHA-BPyne was added to the soluble fraction of
MUM2B proteome (1 mg/mL) at a concentration range of 50
nM-5 µM with or without excess SAHA (20×), crosslinked
at 365 nm on ice, and treated with click chemistry reagents to
append a rhodamine tag. The samples were separated by SDS-
PAGE analysis and analyzed by in-gel fluorescence scanning.
We observed increased signals for three specific targets, two at
∼60 kD (previously determined to be HDACs 1 and 2) and
one at ∼30 kD (previously determined to be the HDAC-
associated protein MBD3),8 with increasing probe concentration;
however, increased background labeling (not competed by
SAHA) presented a problem, particularly at higher concentra-
tions of probe (1-5 µM) (Figure 3A).
To determine if HDAC-enriched proteomic preparations
might display decreased background, we compared the soluble
proteome to the nuclear fraction of MUM2B cells (Figure 3B).
Nuclear proteomes demonstrated improved sensitivity (higher
signal intensities of SAHA-sensitive targets) and modest
improvements in selectivity (decreased intensity of SAHA-
insensitive bands, most notable around the molecular mass range
of MBD3). We also observed one additional target, at ∼45 kD,
that we had not observed in the soluble fraction. However, since
some HDACs are not found in the nucleus,1 we hoped that
SAHA-BPyne might also be amenable to profiling in whole
cell lysates, which encompass all proteomic fractions of the cell
(soluble, nuclear, and membrane). Surprisingly, we observed
improved selectivity for SAHA-sensitive targets in whole cell
proteomes relative to either the simplified soluble or nuclear
fractions, as well as maintenance of labeling of the 45 kD target
(Figure 3C), indicating that SAHA-BPyne performed particularly
well in proteomic samples of very high complexity.
Structure-Activity Relationship (SAR) for Labeling of
HDAC Complexes by SAHA-BPyne and Related Probes. We
next studied the SAR of SAHA-BPyne and examined if probes
based on different scaffolds might exhibit enhanced labeling
of specific targets. For these studies, we chose to maintain three
key features of SAHA-BPyne: (1) the polar headgroup (con-
served among all reported HDAC inhibitors and thought to
interact with the active site metal cation),15 (2) the benzophenone
crosslinker (to provide irreversible labeling), and (3) the alkyne
tag (to allow click chemistry-mediated conjugation to a rhodamine
reporter tag), which, together, would allow us to assay the probes
via the gel-based ABPP methods described above. We prepared
a series of potential probes (Figure 2) that varied in (1) linker
length (distance between polar headgroup and affinity group),
(2) benzophenone position, and (3) the identity of the polar
headgroup and other affinity elements.
The first analogue that we prepared possessed a shorter alkyl
chain linker between the hydroxamate headgroup and benzophe-
none crosslinker than that for SAHA-BPyne, envisioning that
such a change might allow for more optimal binding to HDACs
with shallower active site pockets. Probe 6, containing a glutaric
acid linker (two methylenes shorter than SAHA-BPyne’s suberic
acid linker), was prepared in three steps from commercially
available 4,4′-diaminobenzophenone (Figure 4A). Conversely,
we also examined the impact of increasing the distance between
the hydroxamate to the benzophenone, which we thought might
allow us to cross-link to more distally located HDAC-associated
proteins. Probes 7 and 8, which share an identical linker as
SAHA-BPyne, but place the crosslinking moiety at a farther
distance were prepared by loading a primary amine onto a BAL-
type linker via reductive amination, followed by acylation using
standard Fmoc-based solid-phase peptide synthesis (SPPS)
techniques (Figure 4B). Further elaboration of the structures
by SPPS, followed by cleavage from the support, yielded the
methyl ester analogues of the probes. The methyl esters were
treated with hydroxylamine in solution to provide the desired
hydroxamate probes. Additionally, probes 17 and 18, which have
(15) Grozinger, C. M.; Schreiber, S. L. Chem. Biol. 2002, 9, 3-16.
9
J. AM. CHEM. SOC. VOL. 130, NO. 7, 2008 2189

A R T I C L E S
Salisbury and Cravatt
Figure 3. Comparison of proteome labeling by SAHA-BPyne over a probe concentration range of 50-5000 nM in soluble fractions (A), nuclear fractions
(B), whole cell lysates (C), and in situ labeled cells (D). Gel lanes were loaded with 15 and 60 µg of protein for soluble/nuclear proteome and whole cell
proteome/in situ labeling experiments, respectively. Specifically labeled targets of SAHA-BPyne were defined as proteins whose labeling was completed by
excess SAHA (marked with arrowheads). Double arrowhead highlights additional specific targets of SAHA-BPyne that were detected with greater sensitivity
in nuclear fractions (45 kDa), whole cell lysate (45 kD), and live cells (in situ) (45 and 125 kDa) relative to soluble fractions.
the same suberic acid linker as SAHA-BPyne, but modified
affinity elements, were prepared by similar methods.
(Figure 4C). We also prepared probe 19, an alkyl hydroxamate
similar to SAHA-BPyne, but with a linker displaying a different
pattern of H-bond donors and acceptors, by this method.
The final group of probes, 13-16, was prepared with a
benzamide polar headgroup, as found in the clinical candidates
MS-275 (3) and CI-994 (4), instead of a hydroxamate. These
probes were synthesized via addition of the first amino acid
onto chlorotrityl resin, followed by standard Fmoc-based SPPS
and cleavage from support to provide the corresponding
carboxylic acid analogue of the probes. The benzamide was then
installed by coupling to 1,2-phenylenediamine in solution
(Figure 4D).
We next switched the linker from an alkyl chain to an aryl
ring, a variation found in some previously reported HDAC
inhibitors.16,17 Probes 9-12 were prepared by modifying Wang
resin to display a resin-bound O-hydroxylamine,14 followed by
standard Fmoc-based SPPS and cleavage from the solid support
(16) Bouchain, G.; Delorme, D. Curr. Med. Chem. 2003, 10, 2359-72.
(17) Vannini, A.; Volpari, C.; Filocamo, G.; Casavola, E. C.; Brunetti, M.;
Renzoni, D.; Chakravarty, P.; Paolini, C.; De Francesco, R.; Gallinari, P.;
Steinkuhler, C.; Di, Marco, S. Proc. Natl. Acad. Sci. U.S.A. 2004, 101,
15064-9.
9
2190 J. AM. CHEM. SOC. VOL. 130, NO. 7, 2008

Activity Probes for Profiling HDAC Complexes
A R T I C L E S
Figure 4. Synthesis of HDAC-directed ABPP probes. (A) Solution-phase synthesis of SAHA-BPyne and shorter linker length analogue 6. (B) Preparation
of 7. Probes 8, 17, and 18 were prepared via a similar method, starting with reductive amination of 1-amino-5-hexyne onto FMPB AM resin. (C) Preparation
of aryl hydroxamate probes. Probe 19 was prepared by coupling N-Fmoc-Ahx-OH onto O-hydroxylamine resin, followed by deprotection and elaboration
via the steps shown. (D) Preparation of benzamide probes.
We tested the probes for specific HDAC labeling using both
in Vitro and in situ assays. First, we compared the labeling of
protein targets in cytosolic and nuclear fractions of MUM2B
cells as described above. Almost all of the hydroxamate probes
performed adequately, showing labeling of HDACs 1 and 2 and
MBD3 in the nuclear fraction (Figure 5A). Of the four aryl
hydroxamate probes (9-12), the aminomethyl benzoic acid
(Amb) derivatives (9 and 10) produced the most sensitive
labeling of MBD3, whereas the highest signals for HDACs 1
and 2 were observed with para-aminobenzoic acid (PABA)
derivatives (11 and 12). These findings suggest that the longer
linker length of the Amb derivatives may allow the benzophe-
none moiety to extend further out of the HDAC pocket, thereby
improving interactions with HDAC-associated proteins, whereas
the benzophenone of the PABA derivatives interacts almost
exclusively with the HDACs themselves. Of the two epimeric
aryl hydroxamate PABA derivatives, higher intensity labeling
was observed when the benzophenone amino acid input,
4-benzoyl-phenylalanine (Bpa), displayed an L conformation
(compare compounds 17 versus 18). These differences may
reflect a lack of symmetry in the HDAC binding pockets, in
either structure or sequence (benzophenones preferentially cross-
link to Leu, Val, Lys, Arg, and Met residues18). In contrast to
the hydroxamate probes, none of the four benzamide probes
13-16 showed any SAHA-sensitive targets (data not shown).
The electronics of the benzophenone did not appear to dominate
(18) Dorman, G.; Prestwich, G. D. Biochemistry 1994, 33, 5661-73.
9
J. AM. CHEM. SOC. VOL. 130, NO. 7, 2008 2191

A R T I C L E S
Salisbury and Cravatt
Figure 5. Comparison of in Vitro (nuclear) versus in situ labeling profiles of HDAC-directed ABPP probes. Only relevant portions of the SDS-PAGE gels
that contain HDACs 1 and 2 and MDB3 are shown.
the ability of the probes to photolabel HDAC complexes, as
the probes with a weaker crosslinking potential18 (probes 5 and
6, containing electron donating substituents on the benzophe-
none) provided improved labeling over many of the other
probes. Despite the diversity of structures, no new significant
SAHA-sensitive targets were detected by gel analysis for any
of the probes tested when compared to the profile of the original
SAHA-BPyne probe.
We next tested all probes for in situ labeling of HDAC
complexes in MUM2B cells, using SAHA-BPyne as a positive
control. Surprisingly, none of the new sets of probes performed
as well as SAHA-BPyne in situ (Figure 5B) even at concentra-
tions up to 25 µM probe. At 5 µM, modest labeling of HDACs
1 and 2 and strong labeling of MBD3 were observed with probe
6, the shorter linker length analogue of SAHA-BPyne. Probes
7, 17, and 18 also showed weak in situ labeling of MBD3 at 25
µM but no significant labeling of HDACs 1 or 2. These results
demonstrate that in Vitro labeling is not necessarily predictive
of in situ labeling for ABPP probes.
Figure 6. IC50 values for inhibition of HDAC activities in HeLa cell nuclear
lysate by candidate ABPP probes. Results represent the average values (
standard errors for two-three independent experiments.
Comparison of the Inhibitory Activities of Candidate
HDAC-Directed ABPP Probes. To ascertain whether the
differences observed in proteome labeling among the various
probes were simply a reflection of their respective affinities for
HDACs, we assessed the potency of probe binding by measuring
inhibition of HDAC activity using a fluorogenic substrate assay
kit (Figure 6). Overall, the hydroxamates were significantly more
potent inhibitors of MUM2B HDAC activities than the benza-
mides, not a totally unexpected result, given that the benzamide
analogues of some hydroxamate HDAC inhibitors show dra-
matically higher IC50 values (e.g., 3.5 µM vs 5 nM for
trichostatin A).16 This result suggests that the poor labeling of
HDACs by benzamide-based probes may simply reflect weak
binding interactions. Among the hydroxamate probes, no
obvious correlation between IC50 values and HDAC labeling
in proteomes was observed. Several probes inhibited HDAC
activity more potently than SAHA-BPyne (e.g., probes 11 and
12) but exhibited much weaker labeling of HDAC complexes
in proteomes (Figure 5). This intriguing finding suggests that
other factors, such as the location and orientation of the
crosslinking moiety (and possibly the accessibility of the alkyne
tag), are critical for converting a tight-binding reversible
inhibitor into an optimal performing ABPP probe. On this note,
benzophenones are capable of crosslinking to proteins within a
range of ∼ 3.1 Å from the carbonyl bond,18 and therefore, the
location of the benzophenone (or other photocrosslinking
groups) is likely an important consideration for the successful
design of ABPP probes that operate by a photoreactive
mechanism.
SAHA-BPyne allows for detection of changes in HDAC
activity in live cells. The Van Dyke lab recently reported that
the natural product parthenolide specifically depletes HDAC1
in a number of cell lines.19 Using western blotting, the group
was able to demonstrate that HDAC1 protein abundance
decreases upon exposure to micromolar levels of parthenolide.
To assess whether HDAC actiVities would be affected by
parthenolide, we treated two human cancer cells lines, the colon
cancer line HCT-15 and the breast cancer line MDB-MB-231,
with 15 µM parthenolide for 3 h, followed by in situ analysis
with SAHA-BPyne. SAHA-BPyne labeling confirmed that
HDAC1 activity was decreased after parthenolide treatment in
both cancer lines (Figure 7A and B). Attempts to induce a
complete loss of HDAC1 signals by treating cancer cells with
higher concentrations of parthenolide were thwarted by the
marked toxicity displayed by this compound in HCT-15 cells
(see below). Interestingly, in both cancer lines, parthenolide-
treatment not only reduced HDAC1 activity but also produced
(19) Gopal, Y. N.; Arora, T. S.; Van, Dyke, M. W. Chem. Biol. 2007, 14, 813-
23.
9
2192 J. AM. CHEM. SOC. VOL. 130, NO. 7, 2008

Activity Probes for Profiling HDAC Complexes
A R T I C L E S
In the course of the aforementioned studies, we noted that
HCT-15 cells appeared to exhibit much greater sensitivity to
the cytotoxic effects of parthenolide compared to MDA-MB-
231 cells. This premise was confirmed by evaluating the
cytotoxic effects of parthenolide in these two cell lines. At all
concentrations of parthenolide tested, HCT-15 cells showed a
significantly greater cytotoxic response compared to MDA-MB-
231 cells (Figure 7C). Curiously, however, HDAC1 activity
signals were equal in intensity in these cell lines, as judged by
SAHA-BPyne labeling (Figure 7A and B), indicating that the
levels of HDAC1 in active histone-deacetylating complexes did
not account for the observed differences in cell viablity. On
the other hand, western blotting showed much greater total levels
of HDAC1 protein in HCT-15 cells compared to MDA-MB-
231 cells (Figure 7A, compare lanes 1 and 2 to lanes 4 and 5
in lower panels). These data thus suggest that, while the elevated
sensitivity of certain cancer cells to parthenolide may be due
to higher levels of HDAC1 expression, much of the excess
HDAC1 does not appear to be engaged in active histone
deacetylating complexes. The overexpression of HDAC1 in
certain cancer cells may, hence, create a “free” pool of HDAC1
that possesses functions relevant to cell viability that extend
beyond participation in protein complexes for histone deacety-
lation.
Conclusions
Here, we have investigated the proteome labeling profiles
and inhibitory properties of SAHA-BPyne and related ABPP
probes for HDACs. SAHA-BPyne, somewhat surprisingly,
showed superior performance in terms of sensitivity and
selectivity, when applied to the most complex biological
systems, including whole cell lysates and living cells. One
intriguing rationalization for this finding is that the maintenance
of HDAC complexes in an optimally active form may require
a variety of endogenous factors, some of which could be lost
upon cell homogenization and/or fractionation. Given the
diversity of post-translational mechanisms utilized by cells to
regulate enzymatic pathways,20 we expect that the optimal
performance of many ABPP probes may be positively correlated
with the integrity of biological samples under examination.
SAHA-BPyne further distinguished itself by effectively
labeling HDAC complexes in live cells, whereas other probes
tested targeted HDACs only in Vitro. Interestingly, SAHA-
BPyne does not appear to derive its value as an ABPP probe
simply from enhanced potency in binding HDACs, as several
probes showed stronger inhibitory activity in a substrate assay.
Instead, other features of the SAHA-BPyne scaffold, including
the location and accessibility of the benzophenone and alkyne
moieties, are likely critical to efficiently translate binding into
covalent labeling of HDAC complexes. Even these additional
properties may not be sufficient to explain the unique capacity
of SAHA-BPyne to label HDAC complexes in living cells. Here,
we suspect that potential advantages of SAHA-BPyne may
include enhanced cell permeability, distribution, and/or meta-
bolic stability. Although these parameters are difficult to
rationally design into small-molecule structures, they do suggest
that constructing moderately sized probe libraries and testing
these agents both in Vitro and in situ may provide the most
efficient route to optimized ABPP probes. Our results with
Figure 7. Parthenolide treatment of human cancer cell lines induces changes
in HDAC1 activity and expression. (A) Effects of parthenolide on HDAC1
activity and protein levels, as judged by labeling with SAHA-BPyne (upper
panels) and western blotting (lower panels), respectively. Note that
parthenolide treatment induces the formation of a lower molecule mass,
putatively “cleaved” form of HDAC1 that comigrates in the activity gel
with HDAC2 (note that the slight differences in relative migration distance
between the HDAC1 and HDAC1 “cleaved” signals on the activity gels
and western blots are a result of running the samples on long and short
SDS-PAGE gels, respectively). Gels and blots are representative examples
of two to three independent experiments. (B) Quantification of the SAHA-
BPyne labeling signals for HDAC1 in cancer cells treated with 0 or 15 µM
parthenolide (treatment time of 3 h). Results are the average values (
standard errors for two to three independent experiments. (C) Cytotoxic
effects of parthenolide in human cancer cell lines. Cancer lines were treated
with parthenolide for 4 h, after which cell viability was measured as
described in the Experimental Section. Results represent the average values
( standard errors for three independent experiments. **p < 0.01, planned
comparisons.
an increase in activity signals at a lower molecular mass that
overlapped with HDAC2 (Figure 7A, upper panel). Western
blotting revealed that, in addition to reducing HDAC1 protein
levels in HCT-15 and MDB-MB-231 cells, parthenolide also
induced the appearance of a lower molecular mass, putatively
“cleaved” form of this protein that exhibited a similar molecular
mass to HDAC2 (Figure 7A, lower panel; note that HDAC2
activity levels did not show any obvious changes following
parthenolide treatment, consistent with previous findings indi-
cating that this compound selectively reduces the levels of
HDAC119). To our knowledge, these results are the first
demonstration that parthenolide can induce the formation of a
lower molecular mass form of HDAC1 in cancer cells.
(20) Kobe, B.; Kemp, B. E. Nature 1999, 402, 373-6.
9
J. AM. CHEM. SOC. VOL. 130, NO. 7, 2008 2193

A R T I C L E S
Salisbury and Cravatt
HDACs further argue that these probes do not need to be based
on inhibitors of exceptional potency. Indeed, SAHA only
displays a high nanomolar to low micromolar affinity for
HDACs1-3 but was still a suitable scaffold upon which to
construct selective ABPP probes. Finally, by demonstrating that
the levels of HDAC1 protein, but not necessarily activity, in
cancer cells correlate with increased sensitivity to the cytotoxic
agent parthenolide, we show how ABPP complements expres-
sion-based proteomic methods for discerning the impact of
bioactive small-molecules on complex biological systems. We
anticipate that the findings reported here, in addition to
confirming the value of SAHA-BPyne as a reporter of HDAC
activity and complex assembly in biological systems, may serve
as a stimulus for the future development of additional classes
of photoreactive, “clickable” ABPP probes for chemical pro-
teomic endeavors.21-23
Acknowledgment. We acknowledge H. Hoover for technical
assistance with cell culture and the Cravatt lab for their helpful
discussions and suggestions. This work was supported by the
American Cancer Society (PF-06-009-01-CDD, to C.M.S.), the
National Institutes of Health (CA087660 and CA118696), and
the Skaggs Institute for Chemical Biology.
(21) Ballell, L.; Alink, K. J.; Slijper, M.; Versluis, C.; Liskamp, R. M.; Pieters,
R. J. ChemBioChem 2005, 6, 291-5.
(22) van Scherpenzeel, M.; van der Pot, M.; Arnusch, C. J.; Liskamp, R. M.;
Pieters, R. J. Bioorg. Med. Chem. Lett. 2007, 17, 376-8.
(23) Sieber, S. A.; Niessen, S.; Hoover, H. S.; Cravatt, B. F. Nat. Chem. Biol.
2006, 2, 274-81.
JA074138U
9
2194 J. AM. CHEM. SOC. VOL. 130, NO. 7, 2008