Competition-based, quantitative chemical proteomics in breast cancer cells identifies new target profiles for sulforaphane

Article abstract of DOI:10.1039/c6cc08797c

Sulforaphane is a small molecule isothiocyanate which exhibits anticancer potential, yet its biological targets remain poorly understood. Here we employ a competition-based chemical proteomics strategy to profile sulforaphane's targets and identify over 500 targets along with their relative affinities. These targets provide a new set of mediators for sulforaphane's bioactivity, and aid understanding of its complex mode of action.

Full text of DOI:10.1039/c6cc08797c

ChemComm
View Article Online
View Journal
COMMUNICATION
Competition-based, quantitative chemical
proteomics in breast cancer cells identifies
Cite this:DOI: 10.1039/c6cc08797c
new target profiles for sulforaphane†
Received 2nd November 2016,
Accepted 13th April 2017
James A. Clulow,‡a Elisabeth M. Storck,‡a Thomas Lanyon-Hogg,
‡
a
a
b
Karunakaran A. Kalesh,
Lyn H. Jones
and Edward W. Tate*a
DOI: 10.1039/c6cc08797c
rsc.li/chemcomm
Sulforaphane is a small molecule isothiocyanate which exhibits previously been analysed by proteomic strategies;7–9 however,
anticancer potential, yet its biological targets remain poorly understood. investigations into alterations in protein expression and/or
Here we employ a competition-based chemical proteomics strategy to degradation do not provide insight into direct target binding.
profile sulforaphane’s targets and identify over 500 targets along with Attempts to profile the covalent target spectrum of sulforaphane
their relative affinities. These targets provide a new set of mediators have employed sulforaphane probes labelled with radioisotopes
for sulforaphane’s bioactivity, and aid understanding of its complex or bioorthogonal alkyne reporters.8,10 Analysis of radiolabelled
mode of action.
targets by 2D PAGE followed by proteomic identification has a
strong bias towards highly abundant targets. In contrast,
Isothiocyanates are a class of bioactive, electrophilic compounds advances in the use of ‘click’ chemistry functionalisation of
that are produced from metabolism of glucosinolates found in bioorthogonal reporters allow for a range of analysis, including
many cruciferous vegetables. Promising in vitro cell culture and global proteomic profiling.11 To date, however, a comprehensive
in vivo animal studies for this compound class have led to intense and unbiased screen of sulforaphane’s targets in a relevant
research interest.1 One such isothiocyanate, (À)-1-isothiocyanato- cancer system is still lacking. We therefore sought to employ a
(4R)-(methylsulfinyl)-butane (sulforaphane), has been particularly quantitative, competition-based chemical proteomics strategy to
well studied owing to its potential anti-cancer activity. Sulforaphane profile sulforaphane’s targets in two breast cancer cell lines, and
has been suggested to prevent or suppress the development of many to identify the relative affinities of these targets for sulforaphane.
types of cancer, including breast cancer;2,3 its properties have been
Ahn et al. previously reported the synthesis and application
explored in over 1500 publications and many registered clinical of cell-permeable alkyne-tagged sulforaphane probe 1 (Fig. 1a);10
trials, including ongoing trials of a stabilised sulforaphane–cyclo- the electrophilic isothiocyanate group is replaced with sulfoxy-
dextrin complex in breast cancer. Sulforaphane can covalently thiocarbamate, as the product of thiol addition to the latter is
modify proteins, as well as other biomolecules, and is widely more stable. Sulforaphane’s sulfoxide moiety can be replaced with
appreciated to be a polypharmacological agent that likely affects a ketone without affecting activity.10 1 was reported to identify over
multiple targets and signalling cascades.4,5 A small number 100 protein targets in HEK293 cells by mass spectrometry (MS);
of well-characterised protein targets have been identified for however, only two targets of this probe have been validated as
sulforaphane that provide an initial basis for explaining its genuine sulforaphane targets (macrophage migration inhibitory
cellular activities;6 however, despite growing interest in sulforaphane factor (MIF) and AKAP149), and profiling was not performed in
and its potential therapeutic application, its molecular targets and cancer cell lines.10 We therefore synthesised 1 alongside a novel
underlying mode of action remain poorly characterised.
probe 2 (Fig. 1a) with greatly reduced steric encumbrance
Understanding the full protein target spectrum of sulfora- around the electrophilic warhead. 2 was synthesised using a
phane could provide greater insight into its mode of action. The modified procedure to introduce the alkyne handle in a short
effect of sulforaphane treatment on cellular protein levels has synthetic sequence from propargylamine (Schemes S1 and S2,
ESI†). Sulforaphane and 2 display similar lipophilicity (clog P =
0.15 and À0.62, respectively), which is significantly lower than
a Department of Chemistry, Imperial College London, London, UK SW7 2AZ.
probe 1 (clog P = 2.7). This improvement in probe lipophilicity
E-mail: e.tate@imperial.ac.uk
b Medicine Design, Pfizer, 610 Main Street, Cambridge, MA, 02139, USA
(Dclog P = 3.3) was expected to impart reduced non-specific
binding to 2. Probe 1 has already been shown to exert similar
† Electronic supplementary information (ESI) available: ProteomeXchange with
biological effects to sulforaphane,10 suggesting 2 would also
identifier PXD006279. See DOI: 10.1039/c6cc08797c
‡ Authors share equal contribution.
mimic this biological activity. MCF7 and MDA-MB-231 breast
This journal is ©The Royal Society of Chemistry 2017
Chem. Commun.

View Article Online
Communication
ChemComm
(5 mM) alone, or 2 (5 mM) plus sulforaphane (5, 25 or 100 mM) for
30 min. In parallel, the cell line of interest was labelled with
DMEM media containing 15N413C6-arginine and 15N213C6-lysine
(R10K8) over multiple passages until incorporation of R10K8
was 498% (Tables S1 and S2, ESI†). R10K8-labelled cells were
treated with 2 (20 mM) for 30 min, and lysed to generate a ‘spike-in’
SILAC lysate. A fixed amount of this ‘spike-in’ lysate was added to
each competition lysate and the mixed lysates functionalised by
CuAAC with an azido-biotin capture reagent.16 Labelled proteins
were affinity enriched on neutravidin-sepharose, reduced, alkylated
and trypsin digested, and the resulting peptides analysed by LC-MS/
MS to identify protein targets and quantify enrichment levels across
the sulforaphane concentration gradient, using the ‘spike-in’ SILAC
peptide as an internal standard to generate heavy/light (H/L) ratios
(Fig. 2a). Medium-confidence targets were defined as those giving a
statistically significant (t-test, S0 = 1, FDR = 0.01) increase in H/L
ratio compared to samples treated with 2 only (Fig. S1–S3, ESI†).
High-confidence targets were defined as medium-confidence targets
present at 5, 25 and 100 mM, or 25 and 100 mM sulforaphane
competition (Fig. S4, ESI†). This identified 121 and 129 high-
confidence targets in MDA-MB-231 and MCF7 cells, respectively
(Fig. 2b), with 56 conserved targets (Fig. 2c and Tables S3 and S4,
ESI†). Analysis of published target abundances indicates that
the difference in target profile between cell lines may relate to
differences in expression levels of target proteins (Fig. S5, ESI†).
The most potent sulforaphane binders were identified through
analysis of relative amounts of target captured by 2 over three
sulforaphane concentrations (Fig. 2b). Targets exhibiting strong
Fig. 1 (a) Structures of sulforaphane and probes 1 and 2. The electrophilic
moiety of each molecule is highlighted in blue, and the bioorthogonal
alkyne handle in red. (b) MCF7 and (c) MDA-MB-231 lysate in-gel fluorescence competition for probe labelling at the lowest sulforaphane
and coomassie staining after in-cell labelling with 1 or 2, competed
against a concentration gradient of sulforaphane, iodoacetamide (IAA),
concentration were hypothesised to be the most potent binders
(Fig. 2b). Across both cell lines these included Kelch-like ECH-
associated protein 1 (KEAP1) and MIF. KEAP1 is a repressor of
or N-ethylmaleimide (NEM).
the nuclear factor (erythroid-derived 2)-like 2 (Nrf2) transcription
cancer cell lines were selected as biologically relevant systems factor controlling antioxidant response element (ARE)-driven gene
for treatment with 1 and 2. Following in-cell labelling with each expression,17 whilst MIF is a pro-inflammatory cytokine, whose
probe, cells were lysed and probe-labelled proteomes functionalised expression in cancer correlates with tumour aggressiveness and
with azido-TAMRA capture reagent12 to specifically visualise covalent metastatic potential.18 Both proteins have been previously identified
targets, using well-established copper-catalysed alkyne–azide as sulforaphane targets;17,18 however, this is the first time their
cycloaddition (CuAAC) methodology.13 Proteomes were resolved interaction with sulforaphane has been observed in a cellular
by SDS-PAGE and analysed by in-gel fluorescence (Fig. 1). To environment at endogenous expression levels. Identification of
determine the extent to which sulforaphane shares the same known sulforaphane targets demonstrates the validity of our
target profile as each probe, competition assays were performed approach, and provides the first assessment of the binding of
whereby cells were incubated with sulforaphane prior to probe sulforaphane towards these two targets relative to other proteins
treatment. Sulforaphane concentration-dependent reduction in in the cell, in the absence of artificial overexpression. Sulforaphane
labelling by 1 and 2 was observed for the majority of bands, is often used as a chemical probe inhibitor of KEAP1; however,
suggesting a high degree of target overlap (Fig. 1). Competition our results indicate very substantial target promiscuity at higher
assays against other electrophilic alkylating agents (iodoacetamide concentrations.
and N-ethylmaleimide) indicated common target reactivity with
the probes. The sulforaphane-mimetic probe 2 showed the most informatic platform Ingenuitys Pathway Analysis. The major
robust competition by sulforaphane (Fig. 1). canonical pathway upregulated in a dose-dependent manner in
Sulforaphane’s targets were globally assessed using the bio-
To identify the protein targets of sulforaphane we employed both cell lines was apoptosis signalling (Fig. S6, ESI†), consistent
‘spike-in’ stable isotopic labelling of amino acids in cell culture with sulforaphane’s ability to induce apoptosis in numerous
(SILAC) methodology coupled with competition-based chemical cancer cell lines at elevated concentrations.19 In both cell lines a
proteomics.14 Sulforaphane exerts biological activity at low mM common apoptosis signalling target was NF-kB subunits (Fig. S7,
concentrations in MCF7 and MDA-MB-231 cell lines,15 therefore ESI†). NF-kB is a ubiquitous transcription factor controlling
cells were cultured under normal conditions and treated with 2 gene expression of a variety of pro-inflammatory mediators,
Chem. Commun.
This journal is ©The Royal Society of Chemistry 2017

View Article Online
ChemComm
Communication
Fig. 3 (a) Scatter plot of log2(QS) for selected targets across three
concentrations of sulforaphane competition (5, 25, and 100 mM). Targets
were given a value of 0 when absent from a cell line. (b) MCF7 and
(c) MDA-MB-231 western blot analysis showing STAT1 and STAT3 are targets
of 2 and sulforaphane labelling competition and pull-down. L = lysate,
SN = supernatant, PD = pull-down.
Major downregulated canonical pathways included growth
hormone and ERK/MAP kinase signalling pathways (Fig. S6, ESI†),
which have been identified as targeted by sulforaphane.24,25
Several downregulated pathways are mediated via the signal
transducer and activator of transcription (STAT) 1 and 3 proteins
(Fig. S8, ESI†). STAT3 was a high-confidence target in both lines,
and STAT1 a high-confidence and medium-confidence target
in MCF7 and MDA-MB-231, respectively. STAT proteins are
latent cytosolic transcription factors; STAT3 in particular is an
oncogenic transcription factor constitutively expressed in a
variety of cancers, resulting in expression of various genes
involved in cell proliferation.26 STAT1 and STAT3 were therefore
validated as sulforaphane targets via labelling with 2 in competition
Fig. 2 (a) ‘Spike-in’ SILAC workflow. 2 in competition with a sulforaphane
(SULF) concentration gradient was incubated with cells grown in normal
media, and R10K8 labelled cells were treated with 2 only. Cells were lysed,
and ‘spike-in’ R10K8 lysate labelled with 2 added to light lysates. Probe-bound
proteins were functionalised by CuAAC, affinity enriched and digested for
LC-MS/MS. Heavy/light ratios (R(0)–R(100)) were calculated and quantification
scores (QS) generated as a ratio of ratios to the probe only sample. (b) Heat
maps showing log2(QS) over four sulforaphane concentrations (0, 5, 25 and
100 mM). Blue = no competition, red = high competition. (c) Distribution of
unique and common high-confidence targets.
and activation is linked to cancer cell survival.20 Proposed with sulforaphane, followed by CuAAC functionalisation, pull-down
mechanisms for sulforaphane inhibition of NF-kB include and western blot analysis (Fig. 3b, c and Fig. S9, ESI†). Sulforaphane
interference with an activating kinase IkBa,21 reduced NF-kB inhibition of STAT3 signalling has been reported previously, with
upstream signalling,22 or direct interaction with NF-kB p50 the proposed mechanism of action through reduction in protein
subunit.23 Our data shows sulforaphane binds two other level and phosphorylation status of STAT3’s activator kinase, JAK2.27
NF-kB subunits p65 (RELA, high-confidence in both lines) Our data show sulforaphane also directly binds STAT1 and STAT3.
and p52 (NFKB2, medium- and high-confidence in MCF7 and
Disease and biofunction analysis of sulforaphane’s targets
MDA-MB-231, respectively). Other conserved apoptosis signalling indicated upregulation of organismal death (40 targets in MCF7
targets include DFFA, PLCG1, and RPS6KA1; apoptosis signalling and 34 in MDA-MB-231) and downregulation of cell proliferation
targets such as BID and ROCK1 were only identified in one cell (58 targets in MCF7) and cell viability (24 targets in MDA-MB-231)
line (Fig. 3a). Sulforaphane mediates apoptosis in a cell line- with sulforaphane treatment (Fig. S10–S12, ESI†). Common in
specific manner through different apoptotic pathways;19 these these functions were high-affinity, high-confidence conserved
differences in target profile between MCF7 and MDA-MB-231 targets serine/threonine kinase (STK) 3 and 4, which function in
may explain such effects.
the Hippo pathway to regulate cell proliferation and apoptosis.28
This journal is ©The Royal Society of Chemistry 2017
Chem. Commun.

View Article Online
Communication
ChemComm
3 U. Telang, D. A. Brazeau and M. E. Morris, Exp. Biol. Med., 2009, 234,
287–295.
4 Z. Wang, J. Fan, M. Liu, S. Yeung, A. Chang, M. S. S. Chow, D. Pon
and Y. Huang, Expert Opin. Invest. Drugs, 2013, 22, 1613–1626.
5 S. M. Tortorella, S. G. Royce, P. V. Licciardi and T. C. Karagiannis,
Antioxid. Redox Signaling, 2015, 22, 1382–1424.
6 L. Mi, A. J. Di Pasqua and F.-L. Chung, Carcinogenesis, 2011, 32,
1405–1413.
7 L. Mastrangelo, A. Cassidy, F. Mulholland, W. Wang and Y. Bao,
Cancer Res., 2008, 68, 5487–5491.
8 A. S. Agyeman, R. Chaerkady, P. G. Shaw, N. E. Davidson, K. Visvanathan,
A. Pandey and T. W. Kensler, Breast Cancer Res. Treat., 2012, 132,
175–187.
9 C. Angeloni, S. Turroni, L. Bianchi, D. Fabbri, E. Motori, M. Malaguti,
E. Leoncini, T. Maraldi, L. Bini, P. Brigidi and S. Hrelia, PLoS One,
2013, 8, e83283.
10 Y.-H. Ahn, Y. Hwang, H. Liu, X. J. Wang, Y. Zhang, K. K. Stephenson,
T. N. Boronina, R. N. Cole, A. T. Dinkova-Kostova, P. Talalay and
P. A. Cole, Proc. Natl. Acad. Sci. U. S. A., 2010, 107, 9590–9595.
11 E. W. Tate, K. A. Kalesh, T. Lanyon-Hogg, E. M. Storck and
E. Thinon, Curr. Opin. Chem. Biol., 2015, 24, 48–57.
12 E. Thinon, R. A. Serwa, M. Broncel, J. A. Brannigan, U. Brassat,
M. H. Wright, W. P. Heal, A. J. Wilkinson, D. J. Mann and E. W. Tate,
Nat. Commun., 2014, 5, 4919.
13 M. H. Wright, B. Clough, M. D. Rackham, K. Rangachari,
J. A. Brannigan, M. Grainger, D. K. Moss, A. R. Bottrill, W. P. Heal,
M. Broncel, R. A. Serwa, D. Brady, D. J. Mann, R. J. Leatherbarrow,
R. Tewari, A. J. Wilkinson, A. A. Holder and E. W. Tate, Nat. Chem.,
2014, 6, 112–121.
14 K. A. Kalesh, J. A. Clulow and E. W. Tate, Chem. Commun., 2015, 51,
5497–5500.
15 S. Kanematsu, N. Uehara, H. Miki, K. Yoshizawa, A. Kawanaka,
T. Yuri and A. Tsubura, Anticancer Res., 2010, 30, 3381–3390.
16 M. Broncel, R. A. Serwa, P. Ciepla, E. Krause, M. J. Dallman,
A. I. Magee and E. W. Tate, Angew. Chem., Int. Ed. Engl., 2015, 54,
5948–5951.
17 A. T. Dinkova-Kostova, W. D. Holtzclaw, R. N. Cole, K. Itoh,
N. Wakabayashi, Y. Katoh, M. Yamamoto and P. Talalay, Proc. Natl.
Acad. Sci. U. S. A., 2002, 99, 11908–11913.
18 J. V. Cross, J. M. Rady, F. W. Foss, C. E. Lyons, T. L. Macdonald and
D. J. Templeton, Biochem. J., 2009, 423, 315–321.
The conserved target profile between cell lines similarly
affected apoptotic and growth signalling, with corresponding
upregulation of cell death biofunctions (Fig. S13, ESI†). Analysis
of the interaction network of conserved targets highlighted a
high degree of connectivity around Akt protein kinases and
caspases (Fig. S14, ESI†). Although neither are identified as a
direct target of sulforaphane in our study, previous studies have
shown sulforaphane can downregulate total and phosphorylated Akt
levels leading to antiproliferative effects,29 and promote activation of
various caspases in apoptosis.19 Our target profile therefore identifies
several potential mediators of these effects.
The presented dataset represents the most comprehensive
direct target profile of sulforaphane to date, and highlights the
wide range of targets that sulforaphane covalently binds. It
should, however, be noted that use of sulfoxythiocarbamate
warheads in electrophilic probes may affect target reactivity or
target labelling dependent on transient metabolic modification,
for example by glutathionylation. Sulforaphane induces varied
biological effects at different concentrations,30 and our quantitative
data provides insight into how such effects could be mediated
by demonstrating targets that engage sulforaphane at different
therapeutically relevant concentrations. These target profiles are
orthogonal datasets that complement general proteomic and
transcriptomic studies of sulforaphane treatment,8 and will
provide a valuable starting point for future studies.
Target profile differences between the two cell lines high-
light the potential for sulforaphane to differentiate between breast
cancers based on molecular mechanism. The use of cleavable
capture reagents may also allow identification of sulforaphane
binding sites in future. Increased understanding of sulforaphane’s
targets and underlying mode of action may ultimately provide
better insight into how best to apply this agent or related electro-
philic compounds in the clinic.
19 A. Pledgie-Tracy, M. D. Sobolewski and N. E. Davidson, Mol. Cancer
Ther., 2007, 6, 1013–1021.
20 Y. Ben-Neriah and M. Karin, Nat. Immunol., 2011, 12, 715–723.
´
21 C. Xu, G. Shen, C. Chen, C. Gelinas and A.-N. T. Kong, Oncogene,
This work was supported by the UK Engineering and Physical
Sciences Research Council and Pfizer (EPSRC Industrial CASE
Studentship award to J. A. C.). E. M. S. acknowledges the award
of a PhD studentship from the British Heart Foundation. T. L.-H.
was supported by Cancer Research UK (C29637/A20781). K. A. K.
was funded by the European Commission’s Research Executive
Agency (Marie Curie International Incoming Fellowship,
FP7-PEOPLE-2011-IIF). L. H. J. is an employee and shareholder
of Pfizer.
2005, 24, 4486–4495.
22 H. S. Youn, Y. S. Kim, Z. Y. Park, S. Y. Kim, N. Y. Choi, S. M. Joung,
J. A. Seo, K.-M. Lim, M.-K. Kwak, D. H. Hwang and J. Y. Lee,
J. Immunol., 2010, 184, 411–419.
¨
23 E. Heiss, C. Herhaus, K. Klimo, H. Bartsch and C. Gerhauser, J. Biol.
Chem., 2001, 276, 32008–32015.
24 S. K. Roy, R. K. Srivastava and S. Shankar, J. Mol. Signaling, 2010,
5, 10.
25 S. Pinz, S. Unser and A. Rascle, PLoS One, 2014, 9, e99391.
26 B. B. Aggarwal, G. Sethi, K. S. Ahn, S. K. Sandur, M. K. Pandey,
A. B. Kunnumakkara, B. Sung and H. Ichikawa, Ann. N. Y. Acad. Sci.,
2006, 1091, 151–169.
27 E.-R. Hahm and S. V. Singh, Cancer Prev. Res., 2010, 3, 484–494.
28 S. Wu, J. Huang, J. Dong and D. Pan, Cell, 2003, 114, 445–456.
29 D. Chaudhuri, S. Orsulic and B. T. Ashok, Mol. Cancer Ther., 2007, 6,
334–345.
30 T. A. Shapiro, J. W. Fahey, A. T. Dinkova-Kostova, W. D. Holtzclaw,
K. K. Stephenson, K. L. Wade, L. Ye and P. Talalay, Nutr. Cancer,
2006, 55, 53–62.
Notes and references
1 T.-F. Yin, M. Wang, Y. Qing, Y.-M. Lin and D. Wu, World
J. Gastroenterol., 2016, 22, 7058–7068.
2 C. B. Ambrosone, S. E. McCann, J. L. Freudenheim, J. R. Marshall,
Y. Zhang and P. G. Shields, J. Nutr., 2004, 134, 1134–1138.
Chem. Commun.
This journal is ©The Royal Society of Chemistry 2017