Target discovery in small-molecule cell-based screens by
in situ proteome reactivity profiling
Michael J Evans1, Alan Saghatelian1, Erik J Sorensen2& Benjamin F Cravatt1
Chemical genomics aims to discover small molecules that
affect biological processes through the perturbation of
protein function1,2. However, determining the protein targets
of bioactive compounds remains a formidable challenge3.
We address this problem here through the creation of a
natural product–inspired small-molecule library bearing
protein-reactive elements. Cell-based screening identified
a compound, MJE3, that inhibits breast cancer cell
proliferation. In situ proteome reactivity profiling revealed
that MJE3, but not other library members, covalently labeled
the glycolytic enzyme phosphoglycerate mutase 1 (PGAM1),
resulting in enzyme inhibition. Interestingly, MJE3 labeling
and inhibition of PGAM1 were observed exclusively in intact
cells. These results support the hypothesis that cancer cells
depend on glycolysis for viability and promote PGAM1 as a
potential therapeutic target. More generally, the incorporation
of protein-reactive compounds into chemical genomics
screens offers a means to discover targets of bioactive small
molecules in living systems, thereby enabling downstream
Parallel advances in technologies for cell-based screening4and organic
synthesis5have generated sophisticated experimental platforms for the
rapid pharmacological analysis of structurally diverse small-molecule
libraries. These chemical genomics screens have identified compounds
that affect a wide range of cellular processes1,2,6,7. In most cases,
however, the identity of the protein target(s) of bioactive small
molecules remains unknown. Without this knowledge, subsequent
efforts to elucidate the mechanism of action of small-molecule probes,
as well as to refine their structures for pharmaceutical applications, are
Target identification presents a formidable problem in chemical
genomics experiments for multiple reasons. First, small molecules
originating from cell-based screens tend to produce their cellular
effects at high nM to mM concentrations, meaning that interactions
with their putative protein target(s) are of only moderate potency3.
Such relatively weak reversible interactions are difficult to characterize
using classical biochemical approaches like affinity purification
with immobilized small-molecule resins. Additionally, certain small
molecule–protein interactions may require factors or conditions in
the living cell that are not preserved in cell extracts. Some of these
issues can be circumvented using simple model organisms like yeast,
where, for example, genome-wide suppression screens can assist the
target discovery process8. However, such genetic tools are not straight-
forward to apply in mammalian systems. We reasoned that many of
the aforementioned technical hurdles might be overcome by using
chemical libraries that are designed to produce biological effects
through the covalent modification of proteins. After cell-based screen-
ing, the in situ proteome reactivity profiles of library members could
then be compared to identify proteins uniquely labeled by bioactive
compounds, even in cases where relevant interactions are of modest
affinity and only occur in the context of living cells.
There are numerous examples of small molecules that produce their
biological effects through the covalent modification of proteins,
including both natural products and commercial drugs9,10. Many of
these compounds possess moderately reactive electrophilic groups
that, when incorporated into the appropriate molecular scaffold,
lead to selective labeling and inactivation of specific proteins. Guided
by these principles, we designed and synthesized a library of protein-
reactive small molecules possessing the following elements: (i) a
spiroepoxide reactive group akin to those found in bioactive natural
products like fumagillin11, luminacin D12and FR901464 (ref. 13);
(ii) a variable amine-derived binding group intended to direct
probe reactivity to distinct subsets of proteins in the proteome; and
(iii) an alkyne handle for downstream conjugation to reporter tags via
the bio-orthogonal click-chemistry reaction14,15(Fig. 1). The epoxide
is a versatile electrophile that modifies a range of amino acid residues
in the active sites of mechanistically distinct enzymes11,16–18,
suggesting that a probe library incorporating this reactive group
should target a variety of enzymes in the proteome. Additionally,
the inclusion of a sterically inconspicuous alkyne group in the probe
scaffold permits the click chemistry–based analysis of proteins mod-
ified by chemical probes in living cells14,15without requiring the
presence of a bulky reporter tag on each probe during the initial step
of cell-based screening.
An B50-member probe library was synthesized following the
general route outlined in Supplementary Figure 1 online (see
Supplementary Fig. 2 online for structures of all compounds in the
probe library) and screened for antiproliferation activity against the
invasive human breast cancer line MDA-MB-231 using the XTTassay.
One compound, MJE3, was found to block MDA-MB-231 prolifera-
tion by 60%, a pharmacological effect that greatly exceeded those of
Received 11 July; accepted 17 August; published online 2 October 2005; 10.1038/nbt1149
1The Skaggs Institute for Chemical Biology and Departments of Cell Biology and Chemistry, The Scripps Research Institute, La Jolla, California 92037, USA.
2Department of Chemistry, Princeton University, Princeton, New Jersey 08544-1009, USA. Correspondence should be addressed to B.F.C. (email@example.com).
VOLUME 23NUMBER 10 OCTOBER 20051303
© 2005 Nature Publishing Group http://www.nature.com/naturebiotechnology
the other library members, none of which reduced proliferation by
more than 30% (Fig. 2a). The superior antiproliferative activity of
MJE3 compared to structurally related probes was also observed using
a 5-bromodeoxyuridine (BrdU) incorporation assay (Supplementary
Fig. 3a online), in which this compound exhibited a half-maximal
concentration (IC50value) for inhibition of cell growth of 19 mM
(single treatment; Fig. 2b). Finally, an analog of MJE3 in which the
spiroepoxide was replaced by an exocyclic alkene did not exhibit
antiproliferative effects (Fig. 2b), indicating that the reactivity of this
probe was essential for its biological activity.
We next compared the in situ proteome reactivity profiles of library
members to identify proteins that were selectively labeled by MJE3.
Probe-treated MDA-MB-231 cells were homogenized, incubated with
an azide-derivatized rhodamine tag (Fig. 1) under click-chemistry
conditions, and the reactions analyzed by in-gel fluorescence scanning.
Several protein targets of MJE3 were detected, most of which were also
labeled by other compounds that lacked antiproliferative activity
(Fig. 2c). Interestingly, however, a 26-kDa protein was observed
that reacted specifically with MJE3, but not with other probes in the
spiroepoxide library (double arrowhead, Fig. 2c and Supplementary
Fig. 3b online) or with the MJE3-alkene variant (Supplementary
Fig. 3b online). This target was enriched using a ‘trifunctional’ click-
chemistry tag bearing azide, rhodamine and biotin groups (Fig. 1),
and avidin chromatography, separated by SDS-PAGE and subjected
to in-gel trypticdigestionand
spectrometry (LC-MS) analysis. Database searches of the MS results
identified the protein as brain-type phosphoglycerate mutase 1
(PGAM1), a key enzyme in glycolysis that converts 3-phosphoglyce-
rate to 2-phosphoglycerate19.
Recombinant expression of PGAM1 in the SV40-transformed
monkey kidney cell line COS7 confirmed that this enzyme is a specific
target of MJE3, but not other structurally related probe library
members (Fig. 3a). Interestingly, MJE3 labeling of PGAM1 was only
observed in living cells, as neither the native (MDA-MB-231-derived)
nor recombinantly (COS7-derived) expressed enzyme reacted with
this probe in cell extracts (Fig. 3b,c, respectively). Initially, we
suspected that this finding might be explained by the presence of a
labile ester group in the MJE3 structure, which could be selectively
cleaved in situ to release an active product that targeted PGAM1.
Although the general tenets of this ‘pro-drug’ model appear to be
correct (see below), it did not explain the selective labeling of PGAM1
in situ, as this enzyme did not show significant reactivity with the
carboxylate analog of MJE3 (MJE51) in cell extracts (Fig. 3b,c).
Collectively, these results indicate that PGAM1 is a unique target of
the antiproliferative probe MJE3, and that this small molecule–protein
interaction depends on the native environment of the living cell.
To test whether MJE3 also inhibited PGAM1, we monitored the
activity of this enzyme in extracts of MDA-MB-231 cells treated in situ
with MJE3 (0–200 mM, 2 h) using a coupled-enzyme assay19. MJE3
was found to inhibit PGAM1 activity in MDA-MB-231 cells with an
IC50value of 33 mM (Fig. 4a). In contrast, the structurally related
probe MJE4 did not inhibit PGAM1 (Fig. 4b), consistent with its
inability to label this enzyme in situ (Fig. 2c). Treatment of MDA-MB-
231 proteomes in vitro with MJE3 did not affect PGAM1 activity
(Supplementary Fig. 4 online), indicating that covalent labeling of the
enzyme was required for inhibition.
The site of MJE3 labeling in PGAM1 was characterized using a
recently described tandem orthogonal proteolysis (TOP) strategy for
chemical proteomics20. TOP analysis identified a single probe-labeled
peptide in PGAM1, which corresponded to residues 91–107 modified
by one molecule of ester-hydrolyzed MJE3 (Supplementary Fig. 5 and
Supplementary Methods online). This peptide contains several resi-
dues that line the substrate-binding pocket of PGAM1 (e.g., Y92,
K100)21. We were unable to further refine the site of labeling by
tandem MS analysis, owing to dominant probe fragmentation peaks
(Supplementary Methods online). Because the conditions used to
prepare the labeled protein sample for TOP analysis are mild enough
to preserve ester linkages20, it appears that MJE51, the acid variant of
MJE3, is the active form of the probe. Curiously, however, MJE51 was
not found to react with PGAM1 in living cells (Fig. 3b) nor to inhibit
proliferation (Supplementary Fig. 6a online), possibly owing to the
poor cellular uptake of this probe, which labeled very few targets
in situ (Supplementary Fig. 6b online). These results suggest that the
antiproliferative probe MJE3 is hydrolyzed in situ to produce an acid
product (MJE51) that covalently modifies the PGAM1 active site,
resulting in enzyme inhibition. Also consistent with this model, a
R = N MJE50
R = O MJE3
Figure 1 A natural product-inspired library of protein-reactive chemical genomics probes. Several bioactive natural products, including fumagillin11,
luminacin D12and FR901464 (ref. 13), possess an electrophilic 1-oxa-spiro[2.5]octane substructure capable of covalently modifying the active sites of
enzymes. Based on this scaffold, a library of structurally diverse, protein-reactive compounds was synthesized, that contains a spiroepoxide electrophile,
a variable binding group (BG), and an alkyne (for coupling to azide-modified reporter tags by click chemistry14,15). Representative BG elements are shown
(for a complete list of BGs, see Supplementary Fig. 2 online). The synthesis of the probe library was accomplished via the route shown in Supplementary
Figure 1 online.
1304VOLUME 23NUMBER 10OCTOBER 2005NATURE BIOTECHNOLOGY
L E T T E R S
© 2005 Nature Publishing Group http://www.nature.com/naturebiotechnology
methyl ester analog of MJE3 (MJE52) was
found to covalently label PGAM1 and inhibit
proliferation of MDA-MB-231 cells, whereas
an amide analog (MJE50) was not active in
either assay (Supplementary Fig. 6 online).
Many natural products and drugs produced
their biological effects through the covalent
modification of proteins9,10. Despite their wide-
spread use by nature and humankind, protein-
reactive small molecules have not yet been
purposely incorporated into cell-based screens, possibly out of concern
for their potential lack of target selectivity22. Here, we have shown that
libraries of protein-reactive small molecules can be used, in combination
with functional proteomic technologies, to identify a compound with
unique pharmacological and biochemical effects.
The discovery of a small-molecule inhibitor of PGAM1 that reduces
cell proliferation is provocative for multiple reasons. First, it has been
recognized for decades that cancer cells depend more greatly on
glycolysis for energy production than normal cells (the Warburg
effect), which suggests that inhibitors of this metabolic pathway
could be valuable for cancer treatment23. Second, few, if any small-
molecule inhibitors of PGAM1 have been identified that are functional
in living cells. The only additional inhibitors that we could identify in
the literature are highly polyanionic structures, such as benzene
Figure 3 Identification and characterization of
PGAM1 as the 26-kDa MJE3 target. (a) In situ
labeling of recombinantly expressed PGAM1
by MJE3, but not structurally related probes.
PGAM1 was expressed in COS7 cells by transient
transfection and probes were added to cells for
12 h before cell homogenization and in-gel
fluorescence analysis. Equivalent expression
levels of PGAM1 among different samples were confirmed by western blotting analysis (lower panel). Mock cells were transfected with empty vector.
(b,c) MJE3 covalently labeled PGAM1 in living cells, but not in vitro. Treatment of extracts of MDA-MB-231 (b) or transfected COS7 (c) cells with increasing
concentrations of MJE3 for 1 h failed to result in labeling of PGAM1 (in vitro). In situ labeling with MJE3 for 1 h resulted in strong labeling of PGAM1
(b and c, upper right panels). The ester-hydrolyzed product of MJE3 (MJE51) did not significantly label PGAM1 in vitro (b and c, middle panels) or in situ
(b, right middle panel). For (b), the double-asterisked band above PGAM1 corresponds to a distinct protein in MDA-MB-231 cells that was labeled by both
the MJE3 and MJE51 probes in situ and in vitro.
MJE43 MJE42MJE41 MJE40MJE39MJE38 MJE35MJE32MJE30
MJE10MJE4 MJE29MJE28 MJE26MJE25 MJE24MJE23 MJE22 MJE21MJE14MJE11 MJE9 MJE3
log [I] (µM)
(% of control)
IC50 = 19 µM
Proliferation (% of control)
Probe (20 µM)
MJE50 MJE49 MJE48MJE47 MJE46MJE45 MJE44MJE43MJE42 MJE41MJE40 MJE39MJE38 MJE37MJE36 MJE35MJE34 MJE33MJE32 MJE31 MJE30MJE29MJE28 MJE27 MJE26MJE25 MJE24MJE23 MJE22 MJE21MJE20 MJE19MJE18 MJE17MJE16 MJE15MJE14 MJE13MJE12 MJE11 MJE10
Transfected COS7 MDA-MB-231
Figure 2 Evaluation of the antiproliferation
effects and in situ proteome reactivity profiles
of spiroepoxide probes. (a) A comparison of the
effects of the 50-member spiroepoxide library
on the proliferation of MDA-MB-231 human
breast cancer cells identifies a probe MJE3 with
significantly greater antiproliferative activity
compared to other library members (red bar).
Probes were applied twice to cells in 6-h
intervals before measurement of proliferation
using the XTT assay (final probe concentrations:
first 6 h, 20 mM; last 6 h, 27 mM). Results
represent the average of four independent trials,
**, P o 0.01, MJE3-treated versus vehicle
(DMSO)-treated cells. (b) MJE3 inhibits cancer
cell proliferation with an IC50value of 19 mM
(determined 12 h after a single treatment of
probe using a colorimetric immunoassay for
BrdU incorporation (n ¼ 4)). An MJE3 analog
in which the spiroepoxide was replaced with an
exocyclic alkene (MJE3-alkene) failed to inhibit
proliferation at any of the tested concentrations.
(c) In situ proteome reactivity profiles of
representative members of the spiroepoxide
library reveal a 26-kDa protein uniquely targeted
by MJE3 in MDA-MB-231 cells (double
arrowhead). Probe-labeled proteins were
detected by click chemistry–mediated coupling
to an azide-rhodamine reporter tag and in-gel
fluorescence scanning (fluorescent images shown
in gray scale). The in situ proteome reactivity
profiles of other probe library members are shown
in Supplementary Figure 3b online.
VOLUME 23NUMBER 10 OCTOBER 20051305
L E T T E R S
© 2005 Nature Publishing Group http://www.nature.com/naturebiotechnology
hexacarboxylate, which inhibit yeast PGAM with low mM affinity
in vitro24, but would likely display poor cell permeability.
Recent findings from other labs25,26also support a role for PGAM1
in cancer cell proliferation. For example, a peptide inhibitor of
PGAM1 phosphorylation and activity promotes growth arrest in
tumor cell lines25. Conversely, overexpression of PGAM1 leads to
the immortalization and indefinite proliferation of mouse embryonic
fibroblasts26. Reduction in PGAM1 expression through RNA inter-
ference (RNAi) produced a senescent phenotype in these cells26.
Collectively, these findings suggest that the inhibition of PGAM1
alone may be sufficient to produce an antiproliferative effect in cancer
cells. Our initial attempts to address this question by treating MDA-
MB-231 cells with RNAi probes to PGAM1 have failed to substantially
alter the levels of this enzyme (data not shown).
An alternative possibility is that MJE3 produces its antiproliferative
effect through a combination of inhibiting PGAM1 and other proteins
in MDA-MB-231cells. Indeed, many drugs appear to work by targeting
multiple proteins27,28, and MJE3 was found to label several additional
proteins in MDA-MB-231 cells. Still, because these additional targets of
MJE3 were shared by other probes lacking antiproliferative effects, their
covalent modification was not sufficient to affect cell proliferation in
the absence of PGAM1 inhibition. Finally, it is also possible that MJE3’s
antiproliferative effects are mediated by a distinct target not detected by
our SDS-PAGE analyses of probe-labeled MDA-MB-231 proteomes.
However, given the amount of complementary data implicating
PGAM1 as a regulator of cancer cell viability25,26, we favor this enzyme
as a likely target for the observed antiproliferative effects of MJE3.
Several advantages of performing chemical genomics experiments
with protein-reactive compounds are evident from this study. First,
small molecules that produce their biological effects through the
covalent modification of proteins provide a straightforward way to
couple cellular phenotype to underlying biochemical mechanism. By
comparing the in situ proteome reactivity profiles of a series of
bioactive and inactive small molecules, proteins uniquely targeted by
the former compounds can be rapidly identified. Covalent small
molecule–protein interactions are also easier to characterize across a
wide range of affinities than reversible binding events. This feature is
particularly relevant for chemical genomics experiments, which often
generate small molecule leads with biological effects of only moderate
potency (high nM to low mM)3. Thirdly, and perhaps most provoca-
tively, covalent small molecules permit the identification of protein
targets, like PGAM1, for which chemical inhibition only occurs in the
context of the living cell. Although it remains unclear why labeling of
PGAM1 by MJE3 is exclusively observed in situ, this enzyme engages
in protein-protein complexes29and is post-translationally modified
(e.g., by phosphorylation25), which may prime it for chemical inhibi-
tion in situ. These findings are instructive for the design and execution
of chemical genomics projects, as they suggest that certain small
molecule–protein interactions that occur in native cellular environ-
ments may be difficult to maintain in vitro.
One potential drawback of performing pharmacological screens
with protein-reactive small molecules is that reversible inhibitors, due
to their presumed superior target selectivity, are generally preferred
over covalent agents as lead compounds for drug discovery. It should
be noted, however, that covalent probes like MJE3 can also be used in
competitive proteomic screens30to identify reversible compounds that
target enzymes in living cells. In this way, protein-reactive small
molecules originating from chemical genomics experiments should
empower future cell biological and medicinal chemistry investigations
aimed at translating basic research discoveries into new modes for
Synthesis of the spiroepoxide probe library and reporter tags. A general
scheme for the synthesis of the spiroepoxide probe library is shown in
Supplementary Figure 1 online. With the exception of MJE50, all probes were
synthesized with commercially available amines (the amine substituent for
MJE50 was synthesized as shown in Supplementary Methods online). The
MJE3-alkene probe was synthesized as shown in Supplementary Methods
online. A complete description of the synthesis and characterization of the
probe library is available upon request and will be disclosed in a separate
publication in due course. The structures of probe library members were
confirmed by1H-NMR analysis and high-resolution exact mass measurements
(ESI-TOF). The rhodamine-azide and the trifunctional rhodamine-biotin azide
reporter tags were prepared as previously described15. All compounds were
stored as 10 mM DMSO stocks at ?20 1C.
Cell proliferation screens. MDA-MB-231 cells (ATCC) were grown to 80%
confluency in RPMI-1640 medium (Invitrogen) containing 10% FCS and
supplemented with 200 mM L-glutamine in 10-cm dishes at 37 1C in a 5%
CO2atmosphere before seeding in 96-well plates (Corning) at a density of
10,000 cells/well in 50 ml medium (20–30% confluency). Cells were incubated
for 10 h (37 1C, 5% CO2) and then treated with 50 ml of medium containing
40 mM probe in DMSO or DMSO alone (20 mM final probe concentration;
1% final DMSO concentration). The cells were incubated with probe for 6 h, at
which time an additional 50 ml aliquot of medium containing 40 mM probe was
added (27 mM final probe concentration). After a total treatment time of 12 h,
proliferation was assayed using the XTT colorimetric cell proliferation kit
(Roche) following manufacturer’s guidelines (read at 450 nm; reference at
650 nm). Data represent the average ± s.d. for four trials. The concentration-
dependence of probe effects on proliferation was determined using the BrdU
incorporation assay (Roche) as described in Supplementary Methods online.
In situ proteome reactivity profiling. MDA-MB-231 cells were grown to
20–30% confluency in 10-cm plates under the conditions described
above. Medium was exchanged, and 20 ml of probe was added directly to
10 ml medium in the plate (20 mM final probe concentration). After incubating
for 12 h, the medium was aspirated, and the cells were washed 3? in PBS and
harvested by scraping. Soluble proteome extracts were prepared by homogeni-
zation of cell pellets in PBS and centrifugation at 100,000g (supernatant) for 1 h
and were normalized for protein concentration (Dc protein assay kit; Bio-Rad).
Detection of probe-labeled proteins was carried out using previously described
click-chemistry methods14,15(also see Supplementary Methods online).
Affinity isolation and identification of the 26-kDa MJE3-labeled protein.
Large-scale MDA-MB-231 cultures (8–12 ? 150 mm plates) were grown to
20–30% confluency and treated with 20 mM MJE3 for 12 h, after which cells
were washed, harvested and processed as described above. To remove excess
probe, the MJE3-labeled proteomic extract (2 ml at 2 mg/ml) was applied to
a PD-10 (Amersham-Pharmacia) size-exclusion column pre-equilibrated in
Probe (100 µM)
% PGAM1 activity% PGAM1 activity
Log [MJE3] (µM)
IC50 = 33 µM
Figure 4 MJE3 inhibits PGAM1 activity in MDA-MB-231 cells.
(a) Treatment of MDA-MB-231 cells with increasing concentrations of
MJE3 for 2 h resulted in the inhibition of PGAM1 activity with a half-
maximal (IC50value) of 33 mM. (b) PGAM1 activity in MDA-MB-231 cells
was selectively inhibited by MJE3, but not by the structurally related probe
MJE4, which also did not label PGAM1 (Fig. 2c). Probes were tested at
100 mM. Data represent the average of three independent experiments;
**, P o 0.01 compared to vehicle controls.
1306 VOLUME 23NUMBER 10OCTOBER 2005 NATURE BIOTECHNOLOGY
L E T T E R S
© 2005 Nature Publishing Group http://www.nature.com/naturebiotechnology
10 mM Na+/K+phosphate buffer, pH 8.0 (PB) and eluted with 3.5 ml PB.
Samples were fractionated by Q-sepharose chromatography (0–2 M NaCl
gradient in PB) using an AKTA fast protein (FP)LC (Amersham-Pharmacia).
Fractions containing the 26-kDa MJE3-labeled protein were identified by
performing diagnostic click-chemistry reactions with a rhodamine-azide repor-
ter tag and 20 ml of each 500-ml Q fraction. These fractions were pooled and
reacted with the trifunctional biotin/rhodamine-azide reporter tag15(20 mM,
Fig. 1) and the probe-labeled proteins affinity purified using avidin agarose
beads (Sigma) as described in the Supplementary Methods online. Proteins
were separated and visualized with SDS-PAGE and in-gel fluorescence scanning,
respectively. The 26-kDa MJE3 target was excised from the gel, and the gel band
was washed with methanol:water:acetic acid (5:4:1) for 1 h, then methanol:-
water (1:1, 1 h, then 12 h with fresh solution) and subjected to in-gel trypsin
digestion. The resulting peptide mixture was then analyzed by nano LC-tandem
MS (LC-MS/MS) using a quaternary Agilent 1100 series high-performance
(HP)LC pump directly coupled to a Finnigan LTQ ion trap electrospray mass
spectrometer running the Xcalibur data system (Thermo Electron) as previously
described20. The MS data were used to search public databases, which identified
the 26-kDa protein as brain-type phosphoglycerate mutase 1 (PGAM1).
Recombinant expression and characterization of PGAM1. A PGAM1 cDNA
was purchased as the full-length human clone (Invitrogen) and subcloned via
PCR into the eukaryotic expressionvector pcDNA3.1-myc/His B. The construct
was confirmed by DNA sequencing and transiently transfected into COS7 cells
with Lipofectamine (Invitrogen). After 40 h, probe (20 mM, unless otherwise
stated) was added to the medium and incubated for 12 h, at which point cells
were harvested and processed as described above. Mock (empty vector)-
transfected COS7 cells were also labeled and processed. Probe-labeled proteins
were resolved with SDS-PAGE (13 mg protein/lane) and in-gel fluorescence
scanning. Equivalent levels of PGAM1 expression across samples were con-
firmed by western blotting analysis with either anti-myc (Invitrogen) or anti-
PGAM1 (Abcam) antibodies. Large-scale preparations of PGAM1-transfected
COS-7 cells were used for the characterization of the site of MJE3 labeling by
the TOP method20as described in Supplementary Methods online.
Comparison of in situ and in vitro labeling of PGAM1. For in vitro labeling
experiments, soluble MDA-MB-231 or transfected-COS7 proteome extracts
were adjusted to 1 mg protein/ml in PB (pH 8.0) and treated with 0, 5, 10, 20,
50 and 100 mM probe (MJE3 or MJE51) at 37 1C for 1 h. For comparison,
cultures of MDA-MB-231 or transfected COS7 were incubated with 20 mM
MJE3 for 1 h at 37 1C, after which the cells were washed, scraped and the
soluble fraction isolated by centrifugation and adjusted to 1 mg protein/ml in
PB (pH 8.0). Probe labeling was visualized by in-gel fluorescence scanning and
equivalent levels of PGAM1 expression across samples was confirmed by
western blotting analysis with an anti-PGAM1 antibody.
Inhibition studies of PGAM1. MDA-MB-231 cells were grown to 25%
confluency and treated in situ with 0, 10, 20, 50, 100 and 200 mM probe
(MJE3 or MJE4 in DMSO (final concentration of 1%)) for 2 h. Cells were
harvested as described above and soluble proteomes were prepared in 30 mM
Tris-HCl (pH 7.0) and tested for PGAM1 activity by monitoring the conversion
of 3-phosphoglyceric acid (3-PGA) to 2-phosphoglyceric acid at 20–251 C
using an assay coupled to the oxidation of NADH19. Specifically, the reactions
contained: 20 mg protein, 20 mM KCl, 5 mM MgSO4, 0.2 mM ADP (potassium
salt), 0.2 mM NADH (disodium
glycerate (Tris salt), 10 mM 3-PGA (trisodium salt), 0.1 U enolase (Sigma),
0.6 U lactic dehydrogenase (Sigma) and 0.5 U pyruvate kinase (Sigma) buffered
in 30 mM Tris-HCl (pH 7.0). We added 3-PGA last to initiate the reaction,
which was monitored by a decrease in fluorescence at 340 nm in 20-s intervals
for 10 min using an HP 845E UV/Vis spectrophotometer (Hewlett-Packard).
IC50values were determined from concentration-response curves derived with
data averaged from three independent trials using PRISM software.
Note: Supplementary information is available on the Nature Biotechnology website.
We thank J. Tamiya for valuable contributions to the initial design of the probe
library and A. Speers for assistance with the identification of the probe labeling
site on PGAM1. This work was supported by the National Institutes of Health
grant CA087660 (to B.F.C.), the California Breast Cancer Research Foundation
(B.F.C.), a Merck Fellowship of the Life Sciences Research Foundation (A.S.)
and the Skaggs Institute for Chemical Biology.
COMPETING INTERESTS STATEMENT
The authors declare they have no competing financial interests.
Published online at http://www.nature.com/naturebiotechnology/
Reprints and permissions information is available online at http://npg.nature.com/
1. Strausberg, R.L. & Schreiber, S.L. From knowing to controlling: a path from genomics
to drugs using small molecule probes. Science 300, 294–295 (2003).
2. Stockwell, B.R. Exploring biology with small organic molecules. Nature 432, 846–854
3. Burdine, L. & Kodadek, T. Target identification in chemical genetics: the (often) missing
link. Chem. Biol. 11, 593–597 (2004).
4. Clemons, P.A. Complex phenotypic assays in high-throughput screening. Curr. Opin.
Chem. Biol. 8, 334–338 (2004).
5. Schreiber, S.L. Target-oriented and diversity-oriented organic synthesis in drug dis-
covery. Science 287, 1964–1969 (2000).
6. Dolma, S., Lessnick, S.L., Hahn, W.C. & Stockwell, B.R. Identification of genotype-
selective antitumor agents using synthetic lethal chemical screening in engineered
human tumor cells. Cancer Cell 3, 285–296 (2003).
7. Wu, X., Ding, S., Ding, Q., Gray, N.S. & Schultz, P.G. Small molecules that induce car-
diomyogenesis in embryonic stem cells. J. Am. Chem. Soc. 126, 1590–1591 (2004).
8. Luesch, H. et al. A genome-wide overexpression screen in yeast for small-molecule
target identification. Chem. Biol. 12, 55–63 (2005).
9. Robertson, J.G. Mechanistic basis of enzyme-targeted drugs. Biochemistry 44, 5561–
10.Drahl, C., Cravatt, B.F. & Sorensen, E.J. Protein-reactive natural products. Angew.
Chem. Int. Edn. Engl. 44, 5788–5809 (2005).
11.Liu, S., Widom, J., Kemp, C.W., Crews, C.M. & Clardy, J. Structure of human methion-
ine aminopeptidase-2 complexed with fumagillin. Science 282, 1324–1327 (1998).
12.Wakabayashi, T., Kageyama-Kawase, R., Naruse, N., Funahashi, Y. & Yoshimatsu, K.
Luminacins: a family of capillary tube formation inhibitors from Streptomyces sp. II.
Biological activities. J. Antibiot. (Tokyo) 53, 591–596 (2000).
13.Nakajima, H. et al. New antitumor substances, FR901463, FR901464 and
FR901465. II. Activities against experimental tumors in mice and mechanism of
action. J. Antibiot. (Tokyo) 49, 1204–1211 (1996).
14.Speers, A.E., Adam, G.C. & Cravatt, B.F. Activity-based protein profiling in vivo using
a copper(I)-catalyzed azide-alkyne [3 + 2] cycloaddition. J. Am. Chem. Soc. 125,
15.Speers, A.E. & Cravatt, B.F. Profiling enzyme activities in vivo using click chemistry
methods. Chem. Biol. 11, 535–546 (2004).
16.Taunton, J. How to starve a tumor. Chem. Biol. 4, 493–496 (1997).
17.Bogyo, M., Verhelst, S., Bellingard-Dubouchaud, V., Toba, S. & Greenbaum, D.
Selective targeting of lysosomal cysteine proteases with radiolabeled electrophilic
substrate analogs. Chem. Biol. 7, 27–38 (2000).
18.Herrchen, M. & Legler, G. Identification of an essential carboxylate group at the active
site of lacZ beta-galactosidase from Escherichia coli. Eur. J. Biochem. 138, 527–531
19.White, M.F. & Fothergill-Gilmore, L.A. Development of a mutagenesis, expression and
purification system for yeast phosphoglycerate mutase. Investigation of the role of
active-site His181. Eur. J. Biochem. 207, 709–714 (1992).
20.Speers, A.E. & Cravatt, B.F. A tandem orthogonal proteolysis strategy for high-content
chemical proteomics. J. Am. Chem. Soc. 127, 10018–10019 (2005).
21.Wang, Y. et al. Crystal structure of human B-type phosphoglycerate mutase bound with
citrate. Biochem. Biophys. Res. Commun. 331, 1207–1215 (2005).
22.Sirois, S., Hatzakis, G., Wei, D., Du, Q. & Chou, K.C. Assessment of chemical libraries
for their druggability. Comput. Biol. Chem. 29, 55–67 (2005).
23.Gatenby, R.A. & Gillies, R.J. Why do cancers have high aerobic glycolysis? Nat. Rev.
Cancer 4, 891–899 (2004).
24.Rigden, D.J., Walter, R.A., Phillips, S.E. & Fothergill-Gilmore, L.A. Polyanionic
inhibitors of phosphoglycerate mutase: combined structural and biochemical analysis.
J. Mol. Biol. 289, 691–699 (1999).
25.Engel, M., Mazurek, S., Eigenbrodt, E. & Welter, C. Phosphoglycerate mutase-derived
polypeptide inhibits glycolytic flux and induces cell growth arrest in tumor cell lines.
J. Biol. Chem. 279, 35803–35812 (2004).
26.Kondoh, H. et al. Glycolytic enzymes can modulate cellular life span. Cancer Res. 65,
27.Kung, C. et al. Chemical genomic profiling to identify intracellular targets of a multiplex
kinase inhibitor. Proc. Natl. Acad. Sci. USA 102, 3587–3592 (2005).
28.Wong, S. et al. Sole BCR-ABL inhibition is insufficient to eliminate all myeloprolifera-
tive disorder cell populations. Proc. Natl. Acad. Sci. USA 101, 17456–17461 (2004).
29.Mazurek, S., Zwerschke, W., Jansen-Durr, P. & Eigenbrodt, E. Effects of the human
papilloma virus HPV-16 E7 oncoprotein on glycolysis and glutaminolysis: role of pyruvate
kinase type M2 and the glycolytic-enzyme complex. Biochem. J. 356, 247–256 (2001).
30.Leung, D., Hardouin, C., Boger, D.L. & Cravatt, B.F. Discovering potent and selective
reversible inhibitors of enzymes in complex proteomes. Nat. Biotechnol. 21, 687–691
VOLUME 23NUMBER 10OCTOBER 2005 1307
L E T T E R S
© 2005 Nature Publishing Group http://www.nature.com/naturebiotechnology