View Article Online / Journal Homepage / Table of Contents for this issue

Molecular Dynamic Article Links

BioSystems
Cite this: Mol. BioSyst., 2012, 8, 2637–2644

www.rsc.org/molecularbiosystems PAPER
A chemical proteomics approach reveals Hsp27 as a target for
proapoptotic clerodane diterpenesw
Laura Faiella,a Fabrizio Dal Piaz,*b Angela Bisio,c Alessandra Tosco*b and
Nunziatina De Tommasib
Received 4th May 2012, Accepted 3rd July 2012
DOI: 10.1039/c2mb25171j
Published on 04 July 2012. Downloaded on 23/01/2018 20:12:17.

Clerodane diterpenoids are a class of naturally occurring molecules widely distributed in the
Lamiaceae family. Neo-clerodane diterpenoids from Salvia ssp were recently described as compounds
inhibiting the proliferation of human cancer cell lines. To gain new insights into molecular
mechanism(s) underlying the antitumor potential of this class of compounds, we used a chemical
proteomics approach to analyse the cellular interactome of hardwickiic acid (HAA) selected as a
representative molecule. HAA was linked to an opportune 1,1 0 -carbonyldiimidazole modified by
1,12-dodecanediamine and then immobilized on a matrix support. The modified beads were then
used as bait for fishing the potential partners of HAA in a U937 cell lysate. We identified heat shock
protein 27 (Hsp27), an ATP-independent antiapoptotic chaperone characterized for its tumorigenic
and metastatic properties and now referenced as a major therapeutic target in many types of cancer,
as a major HAA partner. Here, we also report the study of HAA–Hsp27 interaction by means of a
panel of chemical and biological approaches, including surface plasmon resonance measurements
limited proteolysis, and biochemical assays. Our data suggest that HAA could provide a potential
tool to develop strategies for the discovery of Hsp27 chemical inhibitors.

Introduction target(s) of hardwickiic acid (HAA) chosen as representative of

Compounds isolated from members of the Lamiaceae family are
traditionally used as a medicine against many diseases. Within
the Lamiaceae, the genus Salvia is worthwhile to be studied, as it
comprehends more than 900 species distributed throughout the
world.1 The genus comprises herbaceous, suffructicous or shrub-
by perennials, rarely biennial or annual species, many of them
are of ornamental interest also because of their tolerance to
dehydration;2 the genus is endowed with bioactive diterpenes,
commonly found in the complex secretion product of these
plants, whose secretory structures are widely distributed on their
aerial parts.3 Several clerodane-type diterpenes have been iso-
lated from these surface exudates, and many of them have
demonstrated antimicrobial, anti-inflammatory, anti-ulcer and
cytotoxic activity.1,4–6 Despite all these well described actions,
the molecular mechanisms by which this class of compound
exerts its effects are still unclear. In this context, we started an
unbiased screening aimed at the identification of intracellular
a
Dipartimento di Scienze Farmaceutiche, Università di Pisa,
via Bonanno 33, 56126 Pisa, Italy
b
Dipartimento di Scienze Farmaceutiche e Biomediche, Università
degli Studi di Salerno, via Ponte don Melillo 1, 84084 Fisciano,
Italy. E-mail: tosco@unisa.it, fdalpiaz@unisa.it;
Fax: +39 089969602; Tel: +39 089969797
c
Dipartimento di Chimica e Tecnologie Farmaceutiche e Alimentari,
Università di Genova, via Brigata Salerno 13, 16147 Genova, Italy
w Electronic supplementary information (ESI) available. See DOI:
10.1039/c2mb25171j
the clerodane diterpenes isolated from Salvia ssp. One of the
most versatile methods to profile cellular targets of selected drug
candidates is compound-immobilized affinity chromatography.
The procedure involves immobilization of the compound on a
solid support through a spacer arm and the use of this matrix as
a bait to fish for interacting proteins in a cellular lysate or tissue
extract.7,8 The power of affinity chromatography combined with
advances in protein identification by sensitive and high-throughput
mass spectrometric analysis offers huge potential to find out
previously unrecognized activities and potential therapeutic
applications. Using this approach we identified heat shock
protein 27 (Hsp27), an antiapoptotic protein characterized for
its tumorigenic and metastatic properties and now referenced as
a major therapeutic target in many types of cancer,9 as an HAA
target. Here, we report the validation of HAA–Hsp27 inter-
action by means of a panel of chemical and biological
approaches, including surface plasmon resonance measure-
ments, limited proteolysis, and biochemical assays. Our data
suggest that HAA could provide a potential tool to develop
strategies for the discovery of Hsp27 chemical inhibitors.
Results and discussion
Chemical proteomics
To identify clerodane diterpenes molecular target(s) responsible
for their numerous biological activities, chemical proteomics

This journal is c The Royal Society of Chemistry 2012 Mol. BioSyst., 2012, 8, 2637–2644 2637
Published on 04 July 2012. Downloaded on 23/01/2018 20:12:17.
Fig. 1 Chemical structure of compounds 1–16.
experiments were performed. The diterpene HAA10
(compound 1 in Fig. 1) was selected as a testing compound
because of its 4-carboxyl group, allowing us to covalently
bind the compound to an agarose matrix activated with
1,1 0 -carbonyldiimidazole and modified by 1,12-dodecanediamine
as a spacer. The reaction between the spacer-modified resin
and HAA was performed by activating the carboxyl group
with N-(3-dimethylaminopropyl)-N 0 -ethylcarbodiimide (EDC)
and N-hydroxysuccinimide (NHS). The experimental condi-
tions (buffer, pH, incubation time, and temperature) were
optimized to prepare beads with a final HAA concentration
of 15 mmol mL1 resin; the coupling process was monitored
by RP-HPLC analysis and was completed by acetylating the
residual free amino groups of the spacer. In characterizing
binding partners for a small molecule by MS, the major
challenge is to identify bona fide interacting partners because
very high sensitivity of MS analysis can permit identification
of almost all proteins, even contaminants present at very low
levels in the sample, thus the use of a negative control is crucial.
Total protein extracts were derived from U937 cell cultures.
2638 Mol. BioSyst., 2012, 8, 2637–2644
View Article Online
An equal amount of cell lysate was loaded onto the covalently
coupled beads and onto unmodified beads in parallel to
discriminate between proteins specific interaction with hard-
wickiic acid and the unspecific background.11 After extensive
washing, the anchored proteins were released from the resin by
elution with Laemmli buffer and then resolved by SDS-PAGE;
each gel line was cut in 10 pieces, in order to highlight even not
clearly visible differences, digested with trypsin, and analyzed
by mass spectrometry through nanoflow LC/MS/MS. A com-
parison between peptides eluted from HAA modified and
unmodified beads led to the identification of the proteins
specifically captured by the clerodane. As reported in the Venn
diagram (Fig. 2), 87 proteins were identified from the control
gel and 88 in the samples from HAA modified beads. Only 2
proteins were exclusively identified in the binding experiment,
and were considered as potential hardwickiic acid targets
(Table 1). One of the identified proteins is chloride intra-
cellular channel protein1 (CLIC1) that belongs to a particular
class of ion channel proteins that can exist both as soluble and
as integral membrane protein.12 The other protein identified is
This journal is c The Royal Society of Chemistry 2012

Fig. 2 Venn diagram of the chemical proteomics results.
the small heat shock protein Hsp27; the involvement of this
chaperone in many cellular processes related to apoptosis and
cell proliferation13–15 led us to further investigate the interaction
of clerodane diterpenes with this target. Chemical proteomic
Published on 04 July 2012. Downloaded on 23/01/2018 20:12:17.
experiments were performed in triplicate to ensure data repro-
ducibility leading to the identification of about 85% of identical
proteins in the three experiments; moreover, a gel free approach
was also used, leading to similar results. However, CLIC1 and
Hsp27 were always detected in the samples from modified beads
and never revealed in those from control gels.
Surface plasmon resonance experiments
To confirm the binding between Hsp27 and HAA and to
measure the parameters of the interaction, a surface plasmon
resonance (SPR) based binding assay was used. The same
analysis was also performed using 15 terpenoids showing
structural analogies with HAA (compounds 2–16 in Fig. 1)
in order to perform some structure–activity relationship
evaluations. Before immobilizing Hsp27 on a sensor chip, we
evaluated the aggregation level of the protein; indeed, this
chaperone can exist in different oligomerization states,
depending on its concentration and on the experimental
conditions;16 however, to obtain reproducible and consistent
SPR results it is crucial the immobilization of homogeneous
protein population. Therefore, we performed a gel electro-
phoresis analysis under non-denaturing conditions by loading
the protein at different concentrations. Achieved results
demonstrated that when loaded at concentrations up to 2 mM,
Hsp27 migrates as a dimer, in agreement with previously
reported results indicating that the dimers are the minimal
structural units of large Hsp27 oligomers.17 On the basis of
these results, we performed protein immobilization on the
sensor chip using a 0.5 mM protein solution. Some of the
sensorgrams obtained in these experiments are shown in
Fig. 3. Compounds 1–8 efficiently interacted with Hsp27,
and the good quality of the acquired sensorgrams allowed
fitting the curves to a single-site bimolecular interaction model
(A + B = AB) leading to the KD measurements reported in
Table 2. Compounds 1–4 showed the higher affinity towards
the chaperon suggesting that the interaction with Hsp27 was
not affected by the dimension of the small molecules; as
expected, very similar KD were measured for compounds 1
and 2 only differing for the presence of an hydroxyl group at
C-6. Some affinity towards Hsp27 was also demonstrated for
compounds 5–8, even if a 2- to 30-fold increase in the respec-
tive KD compared to that measured for HAA was observed.
Remaining compounds (9–16) showed poor or no interaction
This journal is c The Royal Society of Chemistry 2012
View Article Online
with the chaperone. These data confirmed Hsp27 as a possible
target of clerodane, demonstrating a remarkable affinity
towards the protein for at least 4 out of the 16 investigated
compounds. Moreover, the critical role played by the furane
ring in the Hsp27/diterpene interaction was inferred by the
results obtained for compounds 10 and 15.
Limited proteolysis
Surface plasmon results encouraged us to deepen the study of
interaction between clerodane diterpenes and Hsp27. Even if
compounds 3 and 4 showed an affinity towards immobilized
Hsp27 slightly higher than that measured for HAA, their low
solubility and the small amount available led us to keep on
using HAA as the testing compound.
As a first step to evaluate the effect of the interaction
between HAA and Hsp27 on protein structure and activity,
we used a limited proteolysis-mass spectrometry based strategy.
This approach is based on the evidence that exposed, weakly
structured, and flexible regions of a protein can be recognized
by a proteolytic enzyme. The differences in the proteolytic
patterns observed in the presence or in the absence of putative
protein ligands can be analyzed to identify the protein regions
involved in the molecular interactions.18 Again, compound 16
was used as a negative control. As a first step of this study, we
performed Hsp27 enzymatic digestion under strictly controlled
conditions (buffer, temperature, incubation time and enzyme/
substrate ratio), optimized to preserve the dimeric form of
Hsp27 (protein concentration 1 mM) and the correct protein
folding. Trypsin, chymotrypsin and GluC were used as
proteolytic agents. Digestion products were analyzed by mass
spectrometry and the observed fragments were identified on
the basis of their molecular weight, the selectivity of the used
enzymes and their complementarities. Only the fragments
released by a single hydrolytic event have been taken into
account, in order to evaluate the proteolytic accessibility of the
intact protein. Under the experimental conditions used, the
protein remained largely undigested. Only a limited number of
fragments were released from the protein molecule, showing
that its native conformation is susceptible to proteolysis at a
few very specific, and probably flexible, sites (Fig. 4A): most of
them are located in the N-terminal domain of the protein, in
agreement with previous observations depicting this region as
poorly structured.19,20 Interestingly, a good agreement
between the results achieved with the three enzymes was
observed. When the same experiments were performed by
pre-incubating Hsp27 with HAA, a different proteolytic pattern
was observed (Fig. 4B). First of all, a higher amount of
undigested protein was observed, regardless of the proteolytic
enzyme used; besides, some of the sites prone to enzymatic
hydrolysis in the N-terminal domain of the protein were
protected following the Hsp27 interaction with HAA. In
particular all the cleavage sites detected in the portion 26–36
escaped from hydrolysis in the complex, thus suggesting this
region to be involved in the Hsp27 interaction with the
diterpene. Experiments carried out on pre-incubating Hsp27
with compound 16 produced a proteolytic pattern super-
imposable to that observed for the protein. All the limited
proteolysis detailed data are provided in ESI.w
Mol. BioSyst., 2012, 8, 2637–2644 2639

Table 1 Proteins identified as possible HAA targets by chemical proteomics
Swiss Prot code Protein
HSPB1_HUMAN Heat shock protein 27 kDa
CLIC1_HUMAN Chloride intracellular channel protein 1
Published on 04 July 2012. Downloaded on 23/01/2018 20:12:17.
Fig. 3 Sensorgrams obtained by injecting different concentrations
(50 nM, 100 nM, 250 nM, 500 nM, 1 mM) of compound 1 (A), 3 (B)
or 10 (C) on immobilized Hsp27.
Hsp27 is composed of 205 aminoacids, its structure consists
of an N-terminal domain that includes the so-called WDPF-
motif, the a-crystalline domain and a short C-terminal sequence
(Fig. 4C).13 Both the WDPF-motif and the a-crystalline domain
seem to be important for Hsp27 oligomerization, this small heat
shock protein can form variable size oligomers of up to
800 kDa16 and the state of oligomerization depends on many
2640 Mol. BioSyst., 2012, 8, 2637–2644
View Article Online
Mascot score Sequence coverage (%) Peptides
376 63 16
255 48 10
Table 2 Thermodynamic constants measured by SPR for the inter-
action between tested compounds and immobilized Hsp27
Compound KD (nM)
1 269  35
2 202  18
3 158  21
4 130  26
5 513  42
6 802  38
7 1310  107
8 8987  527
9 No binding
10 No binding
11 No binding
12 No binding
13 No binding
14 No binding
15 No binding
16 No binding
Fig. 4 Preferential cleavage sites observed submitting to limited
proteolysis Hsp27 (A) or Hsp27/1 complex (B). Schematic representa-
tion of structural organization of Hsp27. (C): a-crystalline domain is
underlined in gray and WFPD motif is in black. Reported and
putative phosphorylation sites are indicated by large or small black
circles, respectively.
different factors such as post-translational modifications
(phosphorylation and thiolation) and physical parameters
(pH and temperature).21,22 In particular, phosphorylation leads
to formation of small oligomers, while increase in the tempera-
ture or decrease in the pH leads to the formation of large
oligomers. Ser26 is a putative phosphorylation site that stands
just at the beginning of the region of interaction between the
protein and the HAA, suggesting that the small molecule could
influence the posttranslational modification of this residue. But,
more probably, the binding with the diterpene clerodane could
provoke a conformational change of Hsp27 influencing the
interaction with its protein targets.
Insuline and citrate synthase aggregation assays
The ability of HAA to modulate the in vitro chaperone activity
of Hsp27 was evaluated by monitoring the insulin DTT
This journal is c The Royal Society of Chemistry 2012
Published on 04 July 2012. Downloaded on 23/01/2018 20:12:17.
Fig. 5 Aggregation kinetics of reduced insulin (A) or CS at 43 1C (B)
determined by light scattering. The spontaneous aggregation of the
proteins (E) and the aggregation of the proteins in the presence of
0.150 mM Hsp27 (m), of 0.150 mM Hsp27 and 0.3 mM compound 1
(’), or of 0.150 mM Hsp27 and 0.3 mM compound 16 (K).
induced21 and the citrate synthase (CS) thermal induced22
aggregation in the presence of Hsp27, with or without HAA
or compound 16, used as a negative control. The molecular
chaperone activity of small heat shock proteins is often
evaluated by DTT reduction of insulin. Reduction breaks
the disulfide bonds holding together insulin chains A and B
whereupon chain B aggregates but this aggregation can be
prevented by adding stoichiometric amounts of Hsp27
(Fig. 5A). When a 4-fold molar excess of HAA was added
to this mixture, the slope of the curve became comparable to
that observed without the chaperone, suggesting the inhibition
of Hsp27 chaperone activity by this compound. Conversely,
the presence of compound 16 did not perturb the Hsp27 anti-
aggregation effect. Similar results were achieved by evaluating
the effects of Hsp27 on CS thermal induced aggregation.
Upon incubation at elevated temperatures, CS underwent
quantitative protein aggregation but the presence of stoichio-
metric amounts of Hsp27 changed the aggregation kinetics
significantly (Fig. 5B). Again, adding a 4-fold molar excess of
HAA the anti-aggregation effect of Hsp27 was substantially
reverted, whereas the same amount of compound 16 did not
produce any sensible effect. These results demonstrate that
when HAA binds to the small heat shock protein it influences
the ability of Hsp27 to interact with denatured proteins
in vitro.
Caspase 3 activation
In the effort to give a cellular significance to Hsp27–HAA
interaction, we decided to explore the pro-apoptotic activity of
the small molecule in human monocytes. Hsp27 has recently
been shown to be highly and constitutively expressed in human
monocytic leukemia cells.15 High levels of Hsp27 seem to
block apoptosis induced by chemotherapeutic drugs through
This journal is c The Royal Society of Chemistry 2012
View Article Online
Fig. 6 Apoptosis evaluation by analysis of hypodiploid nuclei per-
centage (A) and of caspase-3 activity measurements by fluorimetric
assays (B) in U937 cells incubated with different concentrations of
HAA and compound 16.
the binding of caspase-3 prodomain, which provides a protec-
tion against the proteolytic cleavage necessary for caspase-3
activation. U937 cells were incubated with different concen-
trations of hardwickiic acid (50, 100 and 200 mM) and 200 mM
of compound 16, after 24 hours cells were harvested and
analyzed for apoptosis and caspase-3 activation. Fig. 6A
shows an increasing ability of HAA to induce apoptosis,
measured as the percentage of hypodiploid nuclei, while
Fig. 6B shows the results of the caspase-3 activity measure-
ments in cell lysates, demonstrating a concurring increase in
the executioner caspase activation. Compound 16 did not
show any apoptotic effect.
Experimental
Plant materials and compounds isolation
Fresh aerial parts of Salvia jamensis J. Compton, S. wagneriana,
S. miniata Fernald, and S. blepharophylla Brandegee ex Epling
were obtained from Centro Regionale di Sperimentazione ed
Assistenza Agricola (Albenga, Italy). The species have been
identified by Dr. Gemma Bramley and a voucher specimen is
deposited in Kew Herbarium (K). The isolation and structural
characterization of compounds 1–16 were performed as
reported in our previous papers.10,23–28 The purity of
Mol. BioSyst., 2012, 8, 2637–2644 2641

compounds 1–16 was assessed by MS and NMR experiments
and resulted to be higher than 95% for all the tested compounds.
Chemical proteomics
The solid support was a 1-carboimidazole activated agarose
beaded resin with a capacity of 50 mmol mL1 and a particle
size range of 45–165 mm (Reacti gel 6x – Support Thermo
Scientific). The resin was derivatized by a bifunctional spacer,
1,2-diaminododecyldecane (Sigma Aldrich, St. Louis, USA).
The derivatization reaction was carried out for 12 hours
at room temperature under stirring, incubating 2 mL of
1,2-diaminododecyldecane solution (10 mg mL1) in sodium
bicarbonate (50 mM, pH 8.5) and acetonitrile (80%) with
1 mL of support, previously equilibrated with sodium bicarbonate
(50 mM, pH 8.5) and acetonitrile (80%). The kinetic reaction
Published on 04 July 2012. Downloaded on 23/01/2018 20:12:17.
was monitored by Kaiser’s test kit (Sigma Aldrich). To bind
hardwickiic acid (HAA) to the derivatized resin, the carboxylic
group of the diterpene was activated by 1-ethyl-3-(3-dimethyl-
aminopropyl)carbodiimide (EDC) and N-hydroxysuccinimide
(NHS): 15 mL of EDC (0.48 M in H2O) and 70 mL of NHS
(0.1 M in H2O) were added to 0.6 mg of hardwickiic acid
(6 mg mL1 in CH3OH). The volume was adjusted to 300 mL with
a mixture of CH3OH and NaHCO3, 50 mM, pH 8.5 (1 : 1 v/v).
The reaction mixture was incubated over night at 4 1C under
stirring with 30 mL of derivatized resin, while 30 mL of resin
with 85 mL of H2O, 57.5 mL of NaHCO3 (50 mM, pH 8.5) and
157.5 mL of CH3OH was the unloaded resin. To verify the
percentage of the bound hardwickiic acid, LC-MS analysis of
the reaction mixture was performed by comparing the HAA
peak areas before and after the incubation with the resin. An
LCQ instrument with an electrospray source and an ion trap
detector, coupled with an HPLC chromatographic system
(Thermo-Fisher Scientific, Waltham, USA), was used: 10 mL
of the reaction mixture were added to 1 mL of a solution of
CH3OH : H2O (1 : 1), the eluents were H2O, 95%, CH3CN,
5%, and HCOOH, 0.1%, using a Luna (2  50 mm) C4
column (Phenomenex, Tarrance, USA).
U937 (human leukemic monocyte lymphoma cell line)
were cultured in RPMI medium (Euroclone, Milano, Italy)
supplemented with 10% (v/v) fetal bovine serum (Euroclone),
100 U mL1 penicillin, and 100 mg mL1 streptomycin (Sigma,
St. Louis, MO) at 37 1C in a 5% CO2 atmosphere. For protein
extracts, cells were harvested and washed three times with PBS
(Sigma Aldrich), cell pellets were suspended in lysis buffer
(PBS 1X, 0.1% Igepal, 10% glycerol) supplemented with
protease inhibitor cocktail (Sigma), incubated for 30 min on
ice and clarified by centrifugation for 15 min at 12 000 g at 4 1C.
Unloaded and hardwickiic acid-loaded beads (100 mL) were
washed extensively with lysis buffer and then incubated with 1 mg
of U937 protein extract for 16 h at 4 1C with continuous shaking
on a rotator tube holder. The modified and unmodified beads were
washed three times with lysis buffer and then three times with PBS
1X, 0.1% Igepal. The elution of interacting proteins was performed
with 50 mL of Laemmli buffer (60 mM Tris–HCl pH 6.8, 2%
sodium dodecylsulfate, 10% glycerol, 0.01% blue bromophenol,
5% b-mercaptoethanol). Eluted samples were loaded on a mono-
dimensional 12% SDS-PAGE, and separated proteins were
stained with Brilliant Blue G-Colloidal (Sigma Aldrich).
2642 Mol. BioSyst., 2012, 8, 2637–2644
View Article Online
In gel trypsin digestions were performed as described.29
Briefly, Coomassie-stained protein bands were excised from
the polyacrylamide gel, reduced, alkylated using iodoacetamide,
and digested by trypsin. The resulting fragments were extracted
and analyzed by LC/MS/MS using a Q-TOF premier instrument
(Waters, Milford, USA) equipped with a nano-ESI source
coupled with a nano-Acquity capillary UPLC (Waters):
peptide separation was performed on a capillary BEH C18
column (0.075 mm  100 mm, 1.7 mm, Waters) using aqueous
0.1% formic acid (A) and CH3CN containing 0.1% formic
acid (B) as mobile phases. Peptides were eluted by means of
linear gradient from 5% to 50% of B in 45 min and a
300 nL min1 flow rate. Capillary ion source voltage was set
at 2.5 kV, cone voltage at 35 V, and extractor voltage at 3 V.
Peptide fragmentation was achieved using argon as collision
gas and a collision cell energy of 25 eV. Mass spectra were
acquired in an m/z range from 400 to 1800, and MS/MS
spectra in a 25–2000 range. Mass and MS/MS spectra calibra-
tion was performed using a mixture of angiotensin and insulin
as an external standard and [Glu]-Fibrinopeptide B human as
a lock mass standard. A gel-free approach was also used:
proteins were eluted from the resin using guanidine hydro-
chloride 6 M, dithiothreitol 20 mM, Tris–HCl 50 mM solution,
pH 7.8. After a 30 min incubation with iodoacetamide, the
protein mixtures underwent trypsin digestion. Resulting
peptides were analyzed by the previously described LC/MS/
MS apparatus, using for the chromatographic separation
a linear gradient from 3% to 55% of B in 90 min. MS and
MS/MS conditions were identical to those described above.
MS and MS/MS data were used by Mascot (Matrix Science)
to interrogate the Swiss Prot non-redundant protein database.
Settings were as follows: mass accuracy window for parent ion,
50 ppm; mass accuracy window for fragment ions, 200 milli-
mass units; fixed modification, carbamidomethylation of
cysteines; variable modifications, oxidation of methionine.
Proteins with more than 2 peptides and program scores >75
were considered as reliable proteins. Protein Prospector soft-
ware was eventually used to confirm protein identification;
setting parameters included selection of trypsin with up to 2
missed cleavage sites, fixed modification, carbamidomethylation
of cysteines; variable modifications, oxidation of methionine.
Surface plasmon resonance
To perform SPR studies optical biosensor BIACORE 3000
(GE Healthcare) was used. Recombinant Hsp27 (Stressgen,
Brussels, Belgium; 5 mM) in sodium acetate (50 mM, pH 4)
was immobilized on a CM5 sensor chip. The carboxylic groups
of the chip were previously activated by EDC 0.2 M and NHS
0.05 M. The exceeding active groups were inactivated with
ethanolamine 1 M. Compounds 1–16 were dissolved in 100%
DMSO to obtain 4 mM solutions, and then diluted 1 : 1000
(v/v) in PBS (10 mM NaH2PO4, 150 mM NaCl, pH 7.4) to a
final DMSO concentration of 0.1%. A concentration series of
compounds was prepared as twofold dilutions into running
buffer: for each sample, the complete binding study was
performed using a six-point concentration series, typically
spanning 0.025–1 mM, and triplicate aliquots of each com-
pound’s concentration were dispensed into single-use vials.
This journal is c The Royal Society of Chemistry 2012

Multiple blank samples of running buffer alone were included
in each analysis. Binding experiments were performed at
25 1C, using a flow rate of 50 mL min1, with 60 s monitoring
of association and 200 s monitoring of dissociation. Simple
interactions were adequately fit to a single-site bimolecular
interaction model (A + B = AB), yielding a single KD.
Sensorgram elaborations were performed using the BIAevaluation
software provided by GE Healthcare.
Limited proteolysis
Limited proteolysis experiments were performed at 37 1C in
PBS, 0.1% DMSO, and 20 mM DTT, using trypsin, endo-
proteinase GluC (V8) or chymotrypsin as proteolytic agents;
30 mL of a 2 mM Hsp27 solution were used for each experi-
ment. Binary complex Hsp27/hardwickiic acid was formed by
Published on 04 July 2012. Downloaded on 23/01/2018 20:12:17.
incubating the protein with a 5 : 1 molar excess of the
compound at 37 1C for 15 min prior to proteolytic enzyme
addition. Both protein and the complex were digested using a
1 : 100 (w/w) enzyme to substrate ratio. The extent of the
reactions was monitored on a time-course basis by sampling
the incubation mixture after 5, 15, and 30 min of digestion.
Samples were desalted by ziptip C4 (Merk Millipore) and the
proteolytic fragments were analyzed by MALDITOF/MS
using a MALDI micro MX (Waters). In order to optimize
the sensitivity and accuracy of the mass measurements, two
different m/z ranges were explored for each sample: a first
range, from m/z 500 to 3500 was analyzed in reflector mode;
the other range, from m/z 3500 to 25 000 was analyzed in linear
mode. Each m/z range was calibrated using a suitable peptide
or protein mixture. Mass data were elaborated using the
Masslynx software (Waters). Preferential hydrolysis sites on
Hsp27 under different conditions were identified on the basis
of the fragments released during the enzymatic digestions.
When comparative experiments were carried out on Hsp27
in the presence or absence of HAA, differences in the suscepti-
bility of specific cleavage sites were detected, from which
protein regions involved in the conformational changes could
be inferred.
Insulin and citrate synthase (CS) aggregation assays
The ability of HAA to modulate the chaperone activity of
Hsp27 was evaluated by monitoring the insulin DTT induced
and the citrate synthase (CS) thermal induced aggregation in
the presence of stoichiometric amounts of Hsp27, and with or
without a 4-fold molar excess of HAA. The aggregation
process was monitored by right angle light scattering measure-
ments, using a Perkin Elmer LS 50B fluorimeter with a Perkin
Elmer PTP-1 controlled water bath in stirred and thermo-
stated quartz cells. Both the emission and excitation wave-
lengths were set at 500 nm, and the band pass was 5 nm.
Kinetics traces reported here are the averages of two measure-
ments. Insulin aggregation was initiated incubating 0.57 nmol
of the polypeptide with 0.1 M DTT in PBS, pH 7.4 at 37 1C.
CS aggregation was initiated by submitting the protein
(0.075 mM) to an incubation in 40 mM HEPES–KOH, pH
7.5 at 43 1C. To eliminate the exceeding citrate before the
analyses, CS was dialyzed against Tris–HCl, 50 mM, and
EDTA, 2 mM, at pH 8 using a 500 mL cut-off 10 000 Da filter.
This journal is c The Royal Society of Chemistry 2012
View Article Online
Apoptosis evaluation
U937 cells were seeded in 48-well plastic plates, 250 mL per
well at 2  105 cells mL1 and HAA or compound 16 was
added at different concentrations (50, 100 and 200 mM) for
24 h. At the end of the treatments 250 mL of a hypotonic buffer
(0.1% sodium citrate, 0.1% Triton X-100 and 50 mg mL1
propidium iodide) was added to cell suspensions. Apoptosis
was evaluated by analysing the percentage of hypodiploid
nuclei by propidium iodide incorporation. Samples were ana-
lysed by a FACS-Calibur flow cytometer using the Cell Quest
software (Becton Dickinson, North Ryde, NSW, Australia).
All the experiments were performed at least in triplicate.
Analysis of caspase-3 activity
Caspase-3 activity was measured by a fluorimetric assay based
on the proteolytic cleavage of the 7-amido-4-methylcoumarin
(AMC)-derived substrate (Ac-DEVD-AMC, Sigma-Aldrich),
which yields a fluorescent product. U937 cells were treated
with different concentrations of HAA for 24 h, then harvested
and washed three times with PBS 1X. Pellets were lysed with
caspase-3 reaction buffer (50 mM HEPES pH 7.5, 0.1 mM
EDTA, 0.1% Nonidet P-40, 0.1% CHAPS, 1 mM DTT)
incubated for 30 min on ice and clarified by centrifugation
for 15 min at 15 000 g at 4 1C. Assays were performed in 96-
well plates by incubating 10 mg of protein extracts with 20 mM
Ac-DEVD-AMC in 200 mL of assay buffer at 37 1C for 3 h.
Samples were analysed using a microplate reader (L55 Fluorescence
Spectrometer Perkin Elmer Instruments; excitation: 360 nm,
emission: 460 nm).
Conclusions
Our investigation aimed at the identification of the molecular
target(s) of clerodane diterpenes suggested that these com-
pounds can interact with Hsp27, modulating its biological
activity. Chemical proteomics approaches, used in this study
resulted efficient and reliable, allowing us to identify the
possible molecular target responsible for the observed effect
of clerodane diterpenes on certain cancer cells proliferation
and survey. The multidisciplinary approach used to validate
the chemical proteomics results confirmed the affinity of these
molecules towards the chaperone, also suggesting the protein
region involved in the interaction, and permitted to demon-
strate the inhibitory effects of such interaction on Hsp27
chaperon activity. At the cellular level, this inhibition resulted
in a concentration dependent activation of caspase-3 deter-
mining the observed proapoptotic effect.
Hsp27 has many diverse functions including chaperone
activity, mRNA stabilization, inhibition of apoptosis, cell
proliferation, and modulation of actin polymerization.
Hsp27 can interfere with a number of cell death pathways
induced by heat shock, oxidative stress or death receptor
agonist.30 Expression of Hsp27 is up-regulated in numerous
types of tumors and promotes unfavorable outcome.9 Over-
expression of this protein is frequently associated with
increased resistance to radiotherapy and to anti-cancer drugs,
such as cisplatin, doxorubicin and etoposide.31 Targeting
Hsp27 by an antisense strategy increases cancer cell death
Mol. BioSyst., 2012, 8, 2637–2644 2643

in vitro and in vivo.32 Despite the increasing interest towards
small heat shock proteins because of their emerging role in
different physiological and pathological processes, few com-
pounds capable of modulating their biological activity have
been actually described. However, such molecules could
represent a useful tool both to design new therapeutic agents
and as probes to investigate the structure and function of these
chaperones. In this light, HAA and related compounds could
be promising in the development of optimized compounds.
Notes and references
1 L. Rodriguez-Hahn, B. Esquivel and J. Cardenas, Trends Org.
Chem., 1992, 3, 99–111.
2 R. M. Augé, J. L. Moore, K. Cho, J. C. Stutz, D. M. Sylvia,
A. K. al-Agely and A. M. Saxton, J. Plant Physiol., 2003, 160,
1147–1156.
Published on 04 July 2012. Downloaded on 23/01/2018 20:12:17.
3 S. O. Duke and A. Oliva, in Allelopathy, chemistry and mode of
action of allelochemicals, ed. F. A. Macı́as, J. C. G. Galindo,
J. M. G. Molinillo and H. G. Cutler, CRC press, 2004,
pp. 201–216.
4 E. L. Whitson, C. L. Thomas, C. J. Henrich, T. J. Sayers,
J. B. McMahon and T. C. McKee, J. Nat. Prod., 2010, 73,
2013–2018.
5 G. M. Vieira-Júnior, L. A. Dutra, P. M. Ferreira, M. O. de
Moraes, L. V. Costa Lotufo, Ó. Pessoa Cdo, R. B. Torres,
N. Boralle, S. Bolzani Vda and A. J. Cavalheiro, J. Nat. Prod.,
2011, 74, 776–781.
6 S. Kanokmedhakul, K. Kanokmedhakul and M. Buayairaksa,
J. Nat. Prod., 2007, 70, 1122–1126.
7 M. Raida, Curr. Opin. Chem. Biol., 2011, 15, 570–575.
8 U. Rix and G. Superti-Furga, Nat. Chem. Biol., 2009, 5, 616–624.
9 D. R. Ciocca and S. K. Calderwood, Cell Stress Chaperones, 2005,
10, 86–103.
10 A. Bisio, N. DeTommasi and G. Romussi, Pharmazie, 2004, 59,
309–311.
11 F. Dal Piaz, A. Vassallo, L. Lepore, A. Tosco, A. Bader and N. De
Tommasi, J. Med. Chem., 2009, 52, 3814–3828.
12 D. R. Littler, S. J. Harrop, S. C. Goodchild, J. M. Phang,
A. V. Mynott, L. Jiang, S. M. Valenzuela, M. Mazzanti,
L. J. Brown, S. N. Breit and P. M. G. Curmi, FEBS Lett., 2010,
584, 2093–2101.
13 M. C. Mymrikov, A. S. Seit-Nebi and N. B. Gusev, Physiol. Rev.,
2011, 91, 1123–1159.
2644 Mol. BioSyst., 2012, 8, 2637–2644
View Article Online
14 A. Parcellier, E. Schmitt, S. Gurbuxani, D. Seigneurin-Berny,
A. Pance, A. Chantôme, S. Plenchette, S. Khochbin, E. Solary
and C. Garrido, Mol. Cell. Biol., 2003, 23, 5790–5802.
15 O. H. Voss, S. Batra, S. J. Kolattukudy, M. E. Gonzalez-Mejia,
J. B. Smith and A. I. Doseff, J. Biol. Chem., 2007, 282,
25088–25099.
16 B. Lelj-Garolla and A. G. Mauk, J. Mol. Biol., 2005, 345, 631–642.
17 M. Ehrnsperger, H. Lilie, M. Gaestel and J. Buchner, J. Biol.
Chem., 1999, 274, 14867–14874.
18 A. Casbarra, L. Birolo, G. Infusini, F. Dal Piaz, M. Svensson,
P. Pucci, C. Svanborg and G. Marino, Protein Sci., 2004, 13,
1322–1330.
19 E. V. Baranova, S. D. Weeks, S. Beelen, O. V. Bukach,
N. B. Gusev and S. V. Strelkov, J. Mol. Biol., 2011, 411, 110–122.
20 E. V. Baranova, S. Beelen, N. B. Gusev and S. V. Strelkov, Acta
Crystallogr., Sect. F: Struct. Biol. Cryst. Commun., 2009, 65,
1277–1281.
21 B. Lelj-Garolla and A. G. Mauk, J. Biol. Chem., 2006, 281,
8169–8174.
22 T. Rogalla, M. Ehrnsperger, X. Preville, A. Kotlyarov, G. Lutsch,
C. Ducasse, C. Paul, M. Wieske, A. P. Arrigo, J. Buchner and
M. Gaestel, J. Biol. Chem., 1999, 274, 18947–18956.
23 A. Bisio, N. DeTommasi and G. Romussi, Planta Med., 2004, 70,
452–457.
24 A. Bisio, D. Fraternale, G. Damonte, E. Millo, A. P. Lanteri,
E. Russo, G. Romussi, B. Parodi, D. Ricci and N. De Tommasi,
Nat. Prod. Commun., 2009, 4, 1621–1630.
25 A. Bisio, G. Romussi, E. Russo, N. DeTommasi, N. Mascolo,
A. Alfieri, M. C. Bonito and C. Cicala, Nat. Prod. Commun., 2008,
6, 881–884.
26 A. Bisio, G. Damonte, D. Fraternale, E. Giacomelli, A. Salis,
G. Romussi, S. Cafaggi, D. Ricci and N. De Tommasi, Phyto-
chemistry, 2011, 72, 265–275.
27 A. Bisio, A. Corallo, P. Gastaldo, G. Romussi, G. Ciarallo,
N. Fontana, N. De Tommasi and P. Profumo, Ann. Bot. (London,
U. K.), 199, 83, 431–452.
28 F. Dal Piaz, N. Malafronte, A. Romano, D. Gallotta,
M. A. Belisario, G. Bifulco, M. J. Gualtieri, R. Sanogo, N. De
Tommasi and C. Pisano, Phytochemistry, 2012, 75, 78–89.
29 A. Shevchenko, H. Tomas, J. Havlis, J. V. Olsen and M. Mann,
Nat. Protoc., 2006, 1, 2856–2860.
30 C. Paul, S. Simon, B. Gibert, S. Virot, F. Manero and A. P. Arrigo,
Exp. Cell Res., 2010, 316, 1535–1552.
31 Y. Zhang and X. Shen, Clin. Cancer Res., 2007, 13, 2855–2864.
32 M. T. Aloy, E. Hadchity, C. Bionda, C. Diaz-Latoud, L. Claude,
R. Rousson, A. P. Arrigo and C. Rodriguez-Lafrasse, Int. J.
Radiat. Oncol., Biol., Phys., 2008, 70, 543–553.
This journal is c The Royal Society of Chemistry 2012