The cytotoxicity of garlic-related disulphides and thiosulfonates in WHCO1 oesophageal cancer cells is dependent on S-thiolation and not production of ROS

Article abstract of DOI:10.1016/j.bbagen.2016.03.032

Background Garlic has been used for centuries in folk medicine for its health promoting and cancer preventative properties. The bioactive principles in crushed garlic are allyl sulphur compounds which are proposed to chemically react through (i) protein S-thiolation and (ii) production of ROS. Methods A collection of R-propyl disulphide and R-thiosulfonate compounds were synthesised to probe the importance of thiolysis and ROS generation in the cytotoxicity of garlic-related compounds in WHCO1 oesophageal cancer cells. Results A significant correlation (R² = 0.78, Fcrit (7,1) α = 0.005) was found between the cytotoxicity IC₅₀ and the leaving group pK_a of the R-propyl disulphides and thiosulfonates, supporting a mechanism that relies on the thermodynamics of a mixed disulphide exchange reaction. Disulphide (1) and thiosulfonate (11) were further evaluated mechanistically and found to induce G₂/M cell-cycle arrest and apoptosis, inhibit cell proliferation, and generate ROS. When the ROS produced by 1 and 11 were quenched with Trolox, ascorbic acid or N-acetyl cysteine (NAC), only NAC was found to counter the cytotoxicity of both compounds. However, NAC was found to chemically react with 11 through mixed disulphide formation, providing an explanation for this apparent inhibitory result. Conclusion Cellular S-thiolation by garlic related disulphides appears to be the cause of cytotoxicity in WHCO1 cells. Generation of ROS appears to only play a secondary role. General significance Our findings do not support ROS production causing the cytotoxicity of garlic-related disulphides in WHCO1 cells. Importantly, it was found that the popular ROS inhibitor NAC interferes with the assay.

Full text of DOI:10.1016/j.bbagen.2016.03.032

Biochimica et Biophysica Acta 1860 (2016) 1439–1449
Contents lists available at ScienceDirect
Biochimica et Biophysica Acta
journal homepage: www.elsevier.com/locate/bbagen
The cytotoxicity of garlic-related disulphides and thiosulfonates in
WHCO1 oesophageal cancer cells is dependent on S-thiolation
and not production of ROS
Muneerah Smith a, Roger Hunter b, Nashia Stellenboom b, Daniel A. Kusza b, M. Iqbal Parker a,c
,
Ahmed N.H. Hammouda b, Graham Jackson b, Catherine H. Kaschula b,
⁎
a
Department of Integrative Biomedical Sciences, Division of Medical Biochemistry, University of Cape Town, Anzio Road, Observatory,7925 Cape Town, South Africa
b
c
Department of Chemistry, University of Cape Town, Rondebosch, 7701 Cape Town, South Africa
International Centre for Genetic Engineering and Biotechnology, Anzio Road, Observatory, 7925 Cape Town, South Africa
a r t i c l e i n f o
a b s t r a c t
Article history:
Background: Garlic has been used for centuries in folk medicine for its health promoting and cancer preventative
properties. The bioactive principles in crushed garlic are allyl sulphur compounds which are proposed to chemically
react through (i) protein S-thiolation and (ii) production of ROS.
Methods: A collection of R-propyl disulphide and R-thiosulfonate compounds were synthesised to probe the
importance of thiolysis and ROS generation in the cytotoxicity of garlic-related compounds in WHCO1 oesophageal
cancer cells.
Received 15 December 2015
Received in revised form 19 March 2016
Accepted 23 March 2016
Available online 4 April 2016
Results: A significant correlation (R2 = 0.78, Fcrit (7,1) α = 0.005) was found between the cytotoxicity IC50 and the
leaving group pKa of the R-propyl disulphides and thiosulfonates, supporting a mechanism that relies on the ther-
modynamics of a mixed disulphide exchange reaction. Disulphide (1) and thiosulfonate (11) were further evaluated
mechanistically and found to induce G2/M cell-cycle arrest and apoptosis, inhibit cell proliferation, and generate
ROS. When the ROS produced by 1 and 11 were quenched with Trolox, ascorbic acid or N-acetyl cysteine (NAC),
only NAC was found to counter the cytotoxicity of both compounds. However, NAC was found to chemically react
with 11 through mixed disulphide formation, providing an explanation for this apparent inhibitory result.
Conclusion: Cellular S-thiolation by garlic related disulphides appears to be the cause of cytotoxicity in WHCO1 cells.
Generation of ROS appears to only play a secondary role.
General significance: Our findings do not support ROS production causing the cytotoxicity of garlic-related
disulphides in WHCO1 cells. Importantly, it was found that the popular ROS inhibitor NAC interferes with the assay.
© 2016 Elsevier B.V. All rights reserved.
1. Introduction
in whole garlic is alliin which is converted into the pungent allicin by
the enzyme alliinase when the clove is crushed (see Fig. 1). Allicin is un-
Exogenous factors are known to contribute to the development of
many cancers, with nutritional intervention playing a major role in its
prevention. Garlic (Allium sativum) is a medicinal plant used throughout
the ages for its beneficial health effects, which include protection
against cancer. Garlic consumption has been the subject of numerous
clinical trials designed to establish a link to reduced cancer risk. Critical
reviews of the epidemiological literature [1–5] are quite contradictory
depending on the tumour examined, the specific garlic preparation
and the analysis used, although there appears to be a consensus that
while the available evidence supports this link, further research is
needed. The complexity of the problem is compounded by the fact
that crude garlic extracts contain numerous pharmacologically active
compounds with varying activities and stability. The main compound
stable in concentrated form and readily self-reacts via S-thioallylation,
or decomposes via an αH Cope-type elimination to give a plethora of
second-generation allyl sulphur products [6], with some of the major
degradation products displayed in Fig. 1. These compounds share a
sulphide or polysulphide functional group flanked by allyl side
groups such as allyl disulphide (DAS), diallyl disulphide (DADS),
diallyl trisulphide (DATS), diallyl tetrasulphide (DAS4) and S-
allylmercaptocysteine (SAMC). Owing to the reactivity of allicin
and related garlic allyl sulphides in S-thioallylation, following ingestion,
these compounds spontaneously react with free intracellular thiols
such as GSH, or susceptible cysteine residues of redox-sensitive proteins
to form mixed disulphides and allyl mercaptan as the leaving group [7].
Allicin and related garlic allyl sulphur compounds, have been shown
to exert cytotoxic effects in cancer cells by inhibiting their cell growth
through G2/M cell cycle arrest, and inducing apoptosis both in vitro
and in vivo [8,9]. In addition, these compounds appear to exhibit low
⁎
Corresponding author.
http://dx.doi.org/10.1016/j.bbagen.2016.03.032
0304-4165/© 2016 Elsevier B.V. All rights reserved.

1440
M. Smith et al. / Biochimica et Biophysica Acta 1860 (2016) 1439–1449
Fig. 1. The predominant secondary products of allicin degradation in crushed garlic. Upon crushing of the garlic clove the enzyme alliinase comes into contact with its substrate alliin and an
elimination reaction ensues to generate allicin. Allicin degrades and rearranges to form secondary garlic products which include (i) the polysulfanes: diallyl disulphide (DADS) and diallyl
trisulphide (DATS), (ii) allyl methyl disulphide (AMDS), (iii) allyl allylmercaptan, (iv) allyl methyl sulphide (AMS), (v) S-allyl cysteine (SAC), (vi) S-allylmercaptocysteine (SAMC), (vii) E-
ajoene, and (viii) Z-ajoene.
cytotoxicity in animals [10], as well as a level of selectivity for tumour
over normal cells in culture [11–13]. The underlying chemical biology
behind their cytotoxicity in cancer cells is still unclear, although protein
S-thiolation and the production of ROS have both been proposed to play
a role [14]. There are a collection of studies which propose that the
specific target of DADS [15], DATS [16], DAS4 [17], Z-ajoene [13] and
SAMC [18] is β-tubulin, which has been shown by proteomics to be S-
thioallylated by DATS in vitro [16]. This modification is postulated to
inhibit the ability of β-tubulin to form microtubules during mitosis, an
important step in cell division. Other protein targets that are S-
thiolated by garlic allyl sulphides include human glutathione reductase
[19] and sulphur transferases [20]. Although a handful of proteins have
been identified as the specific targets of garlic allyl sulphur compounds,
we have recently found that ajoene accumulates in the endoplasmic
reticulum of MDA-MB-231 breast cancer cells where it S-thiolates a
multitude of protein targets [21]. Bulk S-thiolation of ER proteins was
found to interfere with the proper folding of newly synthesised
proteins, and to activate the unfolded protein response [21]. If this
mechanism is central to the cytotoxicity of garlic allyl sulphides in can-
cer cells then the ability of the compound to drive protein thiolysis in
the forward direction should be a determinant in its activity.
In addition to protein S-thiolation, there is a growing body of
evidence supporting the generation of ROS by garlic allyl sulphides
in cancer cells. Within cells cytosolic glutathione is predominantly
in the reduced state (GSH), in which a GSH:GSSG ratio in excess of
100:1 abounds [22]. Garlic allyl sulphides are reported to affect the
GSH:GSSG ratio through spontaneous reaction with GSH to form S-
allylmercaptoglutathione (GSS-allyl), with the expulsion of a thiol
leaving group [23]. Some researchers believe that these metabolites
are in fact the active principles in garlic. For example, allicin has been
shown to readily react with cysteine to form S-allylmercaptocysteine
(SAMC) [24], which may be the active metabolite of allicin in vivo,
although SAMC itself is also a constituent of crushed garlic. The garlic
compound 2-propenyl thiosulfate has been shown to spontaneously
react with GSH at physiological pH to form GSS-allyl [20], and such a
cross reaction would be expected to lower the GSH:GSSG ratio, thereby
allowing a build-up of ROS to occur. Indeed, Schaferberg et al. [25] found
that a fast uptake of ajoene accompanied an immediate reduction in the
GSH:GSSG ratio in cultured BJA-B lymphoma cells. This was also report-
ed following DAS4 treatment of U937 human histiocytic lymphoma
cells [17]. In parallel though, ajoene has also been shown to act as an in-
hibitor of the SH metabolising enzyme glutathione reductase (GR),
involved in the catalysis of GSSG to GSH, which could also play a role
in decreasing the GSH:GSSG ratio [19]. Overall, the garlic compounds
ajoene [26], DADS [27,28], DATS [29,30] and DAS4 [17] have all
been shown to induce a time-dependent production of ROS in
promyeloleukemic, glioblastoma, neuroblastoma, prostate and lympho-
ma cancer cell lines respectively.
In the current study, we aimed at probing whether thiolysis and
ROS-production by a small library of garlic-related disulphides and
thiosulfonates correlate with their cytotoxicity in WHCO1 oesophageal
cancer cells. We utilised a mixed library of R-substituted propyl
disulphides and R-substituted thiosulfonates with varying electronic ef-
fects feeding into the disulphide bond, and the cytotoxicity of these
compounds in WHCO1 oesophageal cancer cells was evaluated. Using
inhibitors of ROS, we then probed the relationship between ROS pro-
duction and the cytotoxicity of disulphide and thiosulfonate compounds
in the WHCO1 cells.
2. Materials and methods
2.1. Synthesis
Tetrahydrofuran was distilled under argon from lithium aluminium
hydride and then sodium wire with benzophenone. Column chroma-
tography was performed using silica-gel 60 (Merck 7734). Thin layer
chromatography (TLC) was carried out on aluminium-backed Merck
silica-gel 60 F254. Compounds were visualised on TLC using a UV lamp.
Infra-red (IR) absorptions were measured on a Perkin Elmer Spectrum
one FT-IR Spectrometer and measured in chloroform. Nuclear Magnetic
Resonance spectra were recorded on a Varian Unity 400 (100.6 MHz for
13C) NMR instrument in chloroform-d. Chemical shifts (δ) were

M. Smith et al. / Biochimica et Biophysica Acta 1860 (2016) 1439–1449
1441
recorded relative to residual chloroform (δ = 7.26 ppm in 1H NMR) and
2.4. Cell proliferation assay
(δ = 77.16 ppm in 13C NMR). All chemical shifts are reported in ppm,
and resonances are assigned according to IUPAC numbering, viz H-
1 = H on C-1. High-resolution mass spectra were recorded on a
Waters API Q-TOF Ultima machine at the Central Analytical Facility, Uni-
versity of Stellenbosch. Elemental analyses were performed using a
Fisons EA 1108 CHNS elemental analyser. Reagents were purchased
from Sigma-Aldrich or Merck.
Cell proliferation was quantitated using the Cell Proliferation ELISA,
BrdU kit (Roche) according to the manufacturer's instruction. Briefly,
cells were plated and treated with compounds according to the cell vi-
ability protocol above. After 48 h, BrdU labelling solution (10 μL) was
added for 4 h, followed by removal and addition of the fixative
(100 μL) and treatment with the anti-BrdU-POD (100 μL) working solu-
tion for 90 min at RT. After removal of the anti-BrdU-POD, the cells were
incubated with the peroxidase (POD) substrate and chemiluminescence
was measured using a Veritas microplate luminometer (Promega).
2.1.1. Synthesis of thiosulfonates and disulphides
The disulphides (1–8) used in the study were synthesised from the
two respective thiols using 1-chlorobenzotriazide (BtCl) as an oxidant
[31], while the thiosulfonates (9–12) were synthesised from potassium
p-toluenethiosulfonate and the respective R-substituted halide, RX
(X = Cl or Br) in DMF at 60 °C for 3 h [32].
2.5. Apoptosis quantification assay
Apoptosis was quantified using the Cell Death Detection ELISA kit
(Roche) according to the manufacturer's instruction. Briefly, cells were
plated and treated with compounds according to the cell viability
protocol above. After the indicated incubation time, cell lysates were
prepared and the cytosolic fraction (20 μL) was transferred to a
streptavidin coated plate for analysis. Briefly, the freshly prepared
immunoreagent (80 μL) containing anti-histone biotin and anti-DNA-
POD was added to each well and the plate was incubated on a shaker
at room temperature for 2 h. The solution was then removed by gentle
tapping, washed (3×) and then incubated with ABTS (100 μL) for
15 min. The reaction was quenched by adding the ABTS stop solution
(100 μL) after which the absorbance at 405 nm and 495 nm was read
using a Multiscan FC plate reader (Thermo Scientific).
2.1.2. (S)-N-acetyl-S-((4-methoxybenzyl)thio)cysteine (13)
To a solution of 11 (150.0 mg, 0.486 mmol) in THF (5 mL) was added
N-acetyl cysteine (40.0 mg, 0.243 mmol) dissolved in THF (0.5 mL) at
25 °C under nitrogen. The reaction was allowed to proceed for 16 h.
The solvent was then removed under reduced pressure and the residue
purified by silica gel chromatography using ethyl acetate/methanol to
afford the mixed disulphide product 13 (53.6 mg, 0.188 mmol) in a
77% yield.
M.p. 112–114 °C; (found: C, 49.83; H, 5.37; N, 2.79; S, 18.03%.
C
13H17NO4S2 requires C, 49.51; H, 5.43; N, 4.44; S, 20.33%), found
[M+H]+ 316.0668, C13H17NO4S2 requires 316.0633; νmax/cm−1 3326,
1701, 1223, 1178, 1607, 1557, 1511, 543 (S–S); δH (CDCl3, 300 MHz)
2.05 (3H, s, NAc), 2.88 (1H, dd, J = 6.0, 14.1 Hz, H-3), 2.94 (1H, dd,
J = 4.7, 14.1 Hz, H-3), 3.78 (3H, s, OCH3), 3.87 (2H, s, Bn), 4.75 (1H,
m, H-2), 6.49 (1H, d, J = 7.4 Hz, NH), 6.74 (1H, s, OH), 6.85 (2H, d,
J = 8.8 Hz, Ar), 7.22 (2H, d, J = 8.8 Hz, Ar); δC (CDCl3, 101 MHz) 23.1,
39.5, 43.1, 52.3, 55.5, 114.3, 128.8, 130.7, 159.4, 171.5, and 173.2.
2.6. Cell-cycle analysis
WHCO1 cells were seeded for triplicate experiments (1 × 106 cells,
8 mL) in 100 mm cell culture dishes and allowed to recover overnight.
To the cells were added DMSO solutions of each compound in media.
Control cells received 0.1% DMSO alone. After the specified incubation
time (0, 6, 12, or 24 h), cells were lifted with trypsin, re-suspended in
2 mL cold PBS, and fixed with cold 70% ethanol (8 mL). Cells were
pelleted, washed with cold PBS, and re-suspended in PBS (100 μL) con-
taining RNase A (50 μg/mL, Roche). Twenty minutes prior to analysis,
propidium iodide (10 μg/mL) was added and incubated with the sample
at 4 °C. Individual samples were run on a Becton Dickinson FACSCalibur
flow cytometer using a 488 nm coherent laser and the acquisition
software Cellquest Pro version 5.2.1. The cell population was identified
and gated (R1) on a forward scatter (FSC) vs. side scatter (SSC) dot
plot in acquisition mode. Fluorescent Channel 2 (FL2) at 575 nm was
used for propidium Iodide detection. A dot plot of FL2A (area) vs
FL2W (width) was used to identify single cells (R2) and to thus
eliminate doublets. A histogram plot of FL2A was used to enumerate
G1/G0, S-phase and G2 populations and a threshold of 52 on the FSC
channel was set to remove sample debris. Nile Red Fluorescent particles
were used for instrument standardisation, stability and reproducibility.
Data was analysed using MODFIT LT version 2.0, software.
2.2. Cell lines and culture conditions
The oesophageal cancer cell-line WHCO1 was derived from a biopsy
of primary oesophageal squamous cell carcinoma of South African
origin [33]; the Het-1A cell line is an oesophageal epithelial cell-line de-
rived from a 25-year-old black male, immortalised with the SV40 large T
antigen (ATCC CRL-2692). Cells were all incubated at 37 °C under 5%
CO2 and cultured with antibiotics in DMEM (Dulbecco's Modified
Eagle Medium) containing 10% FBS (Gibco). For experiments, cells
were plated at the specified cell density and allowed to recover
overnight before treatment with compounds (dissolved in DMSO).
2.3. Cellular viability assay
Cytotoxicity of compounds was evaluated using the standard 3-(4,5-
dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide cellular viabil-
ity assay (MTT assay). Briefly, WHCO1 or Het-1A cells were seeded at
a density of 2.5 and 5.0 × 103 cells per well respectively (90 μL) in 96
well plates and allowed to recover overnight. Compounds in DMSO
(0.1% v/v) were added to the cells (10 μL) at relevant concentrations
(2-fold dilutions) and incubated for 24 or 48 h. Control cells received
0.1% DMSO or media alone. To confirm that the compound itself does
not interfere with the assay, compound in the absence of cells was
included as a control. Thereafter 5 mg/mL MTT (10 μL) was added to
each well and incubated for 4 h at 37 °C, followed by addition of
100 μL of 10% SLS in 0.01 M HCl to solubilise the formazan crystals.
The absorbance was read at 595 nm on a Multiscan FC plate reader
(Thermo Scientific), and data was analysed using GraphPad Prism 4
software using sigmoidal dose–response variable slope curve fitting.
2.7. Reactive oxygen species assay
Reactive oxygen species were measured using the cell-permeable
fluorogenic probe 2′, 7′-dichlorodihydrofluorescin diacetate (DCFH-
DA, Sigma). WHCO1 cells were seeded in quadruplicate in sterile
white 96-well plates at 10 × 103 cells per well in 100 μL and allowed
to recover overnight. The media were then removed and the cells
were rinsed with PBS, followed by re-suspension in Krebs' Ringer Buffer
(KRB, Gibco). If inhibitors were used, they were added and
incubated with the cells for 30 min prior to adding the sulphur
compound (10 μL); 7.5 mM N-acetyl-L-cysteine (NAC) (Sigma), 50 μM
L-ascorbic acid (ascorbate) (Sigma) or 12.5 μM 6-Hydroxy-2,5,7,8-
tetramethylchroman-2-carboxylic Acid (Trolox) (Calbiochem). DCFH-

1442
M. Smith et al. / Biochimica et Biophysica Acta 1860 (2016) 1439–1449
Fig. 2. Synthesis of R-propyl disulphides and R-thiosulfonates. (A) Disulphides were synthesised from the two respective thiols using 1-chlorobenzotriazide as an oxidant [31].
(B) Thiosulfonates were synthesised from potassium p-toluenethiosulfonate and the respective R-substituted halide [32].
DA (10 μM) was then added, followed by treatment of the cells with
either 1 or 11 at the specified concentrations for the indicated time.
Control cells were treated with 0.1% DMSO or media alone. To account
for background DCFH-DA and/or compound fluorescence, wells without
cells were treated with DCFH-DA or compound (1 or 11) alone and fluo-
rescence was subtracted from the corresponding wells containing cells.
direction (Fig. 3A). Reducing the thiolate leaving ability in the form of
entry 4 bearing a p-tolyl group (still resonance-stabilised, but lacking
an electron-withdrawing group on the ring), resulted in a threefold re-
duction in activity (IC50 = 61 μM) in keeping with the relative electronic
character and anion stabilising ability of the group. By comparison,
disulphides 5, 6, 7 and 8 bearing different-length alkyl groups (8 with
an N-Boc cysteine Me ester group), all as poor leaving groups due to a
lack of anion stabilisation, grouped together with an activity reduction
around 1000-fold, pushing the activity into the millimolar range (8–
50 mM). In keeping with this analysis, the leaving group stability, as
reflected by the pKa of its conjugate acid (shown in Table 2 and taken
from literature data based on measurements carried out in water),
Fluorescence was quantified on
a Cary Eclipse fluorescence
spectrophotometer (Varian), Eex = 480 nm; Eem = 530 nm.
3. Results
3.1. Relationship between leaving group pKa and cytotoxicity of disulphides
and thiosulfonates in WHCO1 cells
Table 1
Previous studies in our laboratory have demonstrated that S-
thiolation of N-Boc-cysteine by ajoene [32] is regioselective, with nucle-
ophilic attack occurring selectively at the allyl-S rather than the vinyl-S
of the ajoene disulphide during thiolysis exchange. Furthermore, we
established that a dansyl group, strategically inserted on the allyl-S
end of ajoene (to form a derivative called DP) [21], is transferred to nu-
merous thiol groups of proteins following the addition of DP to MDA-
MB-231 breast cancer cells. These findings were rationalised on the
basis of the superior leaving ability of the vinylthio group, by virtue of
delocalisation of incipient charge into the double bond, and this aspect
was used as a crucial design element in the current library of disulphides
chosen. Hence, for the disulphide library propyl-SSR, R was chosen as an
anion stabilising group to favour regioselective attack involving propyl-
S as the transferred group in the disulphide exchange and RS as the leav-
ing group (see Fig. 3A). Specifically, R was varied as an aryl or heteroaryl
group that could stabilise the thiolate leaving group in the exchange by
resonance stabilisation, although some non-stabilising alkyl groups
were also included as controls in which presumably regioselectivity
was not secured. The disulphides were synthesised according to our
previously published protocol [31] for unsymmetrical disulphide syn-
thesis using 1-chlorobenzotriazole as the oxidant in the coupling of
two different thiols (see Fig. 2A), and the compounds have been previ-
ously characterised.
Cytotoxicity of disulphide and thiosulfonate compounds against WHCO1 cancer cell prolif-
eration. Disulphide (A) and thiosulfonate (B) compounds were quantitated for their anti-
proliferative activity using the MTT assay. Extent of inhibition is reported as the IC50 de-
fined as the concentration found to inhibit 50% WHCO1 cell growth after 48 h. Each IC50
value is an average of three independent determinations.
A
No.
−R
IC50 SD (μM)
WHCO1
1
9.9 1.9
2
3
22 3.0
23 1.1
4
5
61 6.3
1.01 0.57 × 104
6
7
8
4.93 1.24 × 104
9.59 1.08 × 103
8.51 7.87 × 103
All disulphides were tested for their ability to inhibit WHCO1
cancer-cell proliferation by an MTT cytotoxicity assay. This assay quan-
titates the enzymatic reduction of a yellow tetrazolium salt to purple
formazan crystals by NAD(P)H-dependent cellular oxidoreductases.
This reaction is only possible in metabolically active cells and therefore
gives a quantitative measure of cell viability. The extent of inhibition is
reported as a half maximal inhibitory concentration (IC50) after a 48 h
incubation period. All compounds were tested a minimum of three
times, and the IC50 value is reported as an average thereof. Within this
series of disulphides a wide variation in the cytotoxicity was observed,
with values ranging from 10 μM to 50 mM, some 10,000 fold (see
Table 1A).
B
No.
−R
IC50 SD (μM)
WHCO1
9
13 2.3
16 3.0
10
The three most active compounds in the disulphide series were en-
tries 1–3 with electron-withdrawing pyridyl, p-nitophenyl and
benzothiazolyl groups respectively, with IC50's of 10–20 μM. Notably,
these R groups all provide resonance stabilisation of the thiolate leaving
group that is expected to drive the thiolysis exchange in the forward
11
12
15 1.9
13 1.2

M. Smith et al. / Biochimica et Biophysica Acta 1860 (2016) 1439–1449
1443
Fig. 3. Proposed thiolysis mechanism between reduced glutathione or protein thiol and a disulphide or thiosulfonate compound. (A) S-thiolation involving R-propyl disulphides with the
expulsion of a thiolate leaving group; (B) S-thiolation involving R-thiosulfonates with expulsion of p-toluenesulfinic acid as a leaving group. GSH = reduced glutathione, PSH = protein
thiol.
could be correlated with the observed cytotoxicity. This was done by
converting the IC50's to Logs and plotting against the reported pKa for
the corresponding leaving group. The p-tolylsulfonyl leaving group of
the thiosulfonates (see next section) also followed the observed trends
and was included in the analysis. A significant correlation was found be-
tween the LogIC50 and the pKa of the leaving group with an R2 = 0.78,
Fcrit (7,1) α = 0.005. (see Fig. 4).
from the p-tolylthio (disulphide) case in entry 4 (IC50 = 61 μM) about
fourfold to be on a par with the most potent derivatives in the di-
sulphide series (entries
1 and 2) bearing strongly electron-
withdrawing groups in the ring. Once again, these results further vali-
date our hypothesis that thiolate stability (leaving ability) closely aligns
with cytotoxicity, and in which thiolysis exchange is operating as the
driver in the cytotoxic behaviour.
In the thiosulfonate library the leaving group (LG) was kept constant
as p-tolylsulfonyl, while the transferring group (T) was now varied from
propyl (entry 12) to a benzyl group (entry 9), which in two cases was
substituted with a p-substituent as electron-withdrawing (F) in entry
10 and electron-releasing (OMe) in entry 11. Similar IC50's were found
for all of the four thiosulfonates at around 15 μM (see Table 1B). These
results indicate that the thiosulfonate moiety is solely responsible for
the activity, in which the level of cytotoxicity once again closely aligns
with the group's stability (leaving ability) as indicated through pKa
(2.8 for p-tolysulfinic acid — see Fig. 3B). The presence of two
electron-withdrawing oxygens on the sulphur improves the activity
3.2. Cytotoxicity of 1 and 11 in WHCO1 cells
Evidence from the literature on the cytotoxicity of garlic allyl
sulphides in cancer cells shows that although ajoene, allicin, DATS,
DADS and DAS have different sulphur-containing functional groups,
they share common mechanistic features. These include: inhibition of
cell growth, G2/M cell cycle arrest, increased levels of ROS and induction
of apoptosis via the intrinsic pathway [34–36]. In this current report, our
structure–activity data for both the disulphides and thiosulfonates
support a common mechanism in which thiolysis exchange is a key
Table 2
Structures of disulphides and thiosulfonates and their respective leaving groups in a thiolysis exchange reaction.
No.
Disulphide
LogIC50
1.00
Leaving group (LG)
LG pKa
1
−1.07 [48,49]
2
1.30
5.1 [49,50]
3
4
1.36
1.79
7.09 [51]
6.82 [52]
5
6
7
8
4.01
4.69
3.93
3.98
10.0 [53]
12.2 [49,53]
10.2 [49,53]
10.4 [54]
12
1.11
2.8 [55]

1444
M. Smith et al. / Biochimica et Biophysica Acta 1860 (2016) 1439–1449
or thiosulfonate pharmacophore that is responsible for the cytotoxic
effects in WHCO1 cells. The latter include the low micromolar inhibition
of proliferation and induction of apoptosis, as well as G2/M cell cycle
arrest.
3.3. Inter-relationship between ROS production and the cytotoxicity of 1
and 11 in WHCO1 cells
Garlic allyl sulphides are reported to induce ROS in cancer cells,
which may play a role in their cytotoxicity. To investigate whether 1
and 11 produce ROS in WHCO1 cells, the 2,7-dichlorodihydroflourescin
diacetate (DCFH-DA) reagent was used. DCFH-DA is able to permeate
cells and only fluoresces upon ester cleavage and oxidation to
dichlorofluorescein. The DCFH-DA reagent is oxidised by a broad
spectrum of ROS and NOS which include hydrogen peroxide, hydroxyl,
peroxy radicals, nitric oxide and peroxynitrite radicals making it a
popular reagent for quantifying oxidative stress in cells. [37,38].
WHCO1 cells were treated with 1 or 11 at 1/2 IC50, 24 h, IC50, 24 h or
2 × IC50, 24 h concentrations, and subsequent fluorescence was recorded
up to 6 h. As can be seen in Fig. 6A WHCO1 cells treated with 11 resulted
in a time- and concentration-dependent increase in the production of
ROS compared to the control. The levels of ROS in the untreated sample
also increased with time due to cell proliferation and endogenous ROS
being produced. By comparison, compound 1 failed to produce signifi-
cant levels of ROS beyond the control up to 6 h at both 1/2 IC50, 24 h
Fig. 4. Plot of pKa vs LogIC50. Correlation between the pKa of the leaving group and the Log
of the cytotoxicity IC50. Compounds included in the plot are propyl derivatives 1, 2, 3, 4, 5,
6, 7, 8, and 12. Equation: LogIC50 = 0.3168 pKa + 0.3365. R2 = 0.78, Fcrit (7,1) α = 0.005.
reaction. Therefore, one disulphide and one thiosulfonate were selected
for more detailed mechanistic evaluation. Compound 1 was selected
from the disulphide group based on it displaying the lowest IC50. As all
the thiosulfonates displayed very similar IC50s, 11 were chosen
arbitrarily for further analysis.
We first set out to confirm that 1 and 11 do indeed inhibit cell prolif-
eration, as the MTT assay measures viability, which is only an indirect
measure of proliferation. For this purpose the BrdU cell-proliferation
assay was used, which directly measures cell proliferation through
quantification of incorporated tagged uracil into newly synthesised
DNA. Compounds 1 and 11 were tested at 2 × IC50 concentration for
48 h, with control cells treated with media alone or 0.1% DMSO in
media alone. In agreement with the MTT results, complete inhibition
of cell proliferation was observed for both compounds 1 (p b 0.001)
and 11 (p b 0.001) (see Fig. 5A).
In order to assess whether 1 and 11 induce apoptosis in WHCO1
cancer cells, cells were treated with IC50 concentrations of each
compound for 24 h. Morphology analysis of the treated cells was
found to support apoptosis, with cells appearing condensed and round-
ed up following treatment, a characteristic feature of apoptosis. These
cells were then harvested, and apoptosis was quantitated using an
ELISA-based assay that detects histone-associated DNA fragments in
the cytoplasm after treatment. Histone-associated DNA fragments
were detected photometrically using the ABTS substrate. From Fig. 5B
it is statistically evident that both 1 (p b 0.005) and 11 (p b 0.001)
induce apoptosis in WHCO1 cancer cells.
In order to assess whether 1 and 11 display any selectivity for
tumour over normal cells, the anti-proliferative IC50 was determined
at the 24 h timepoint for both WHCO1 and Het-1A cells using the MTT
assay. The WHCO1 cell line is derived from a biopsy of oesophageal
squamous cell carcinoma, whereas the Het-1A cell line is a non-cancer
cell line of oesophageal epithelial origin, immortalised with a SV40
large T antigen. As shown in Fig. 5D, the 24 h IC50 of 1 and 11 in the
Het-1A cells was found to be 206 and 65 μM, which reveals a favourable
10- and 2.5-fold selectivity for tumour over normal cells respectively.
Garlic allyl sulphur compounds are reported to inhibit cancer cell
growth by inducing G2/M cell cycle arrest. We therefore tested whether
compounds 1 and 11 affect cell cycle progression in WHCO1 cells. Cells
were treated with 2 × IC50, 24 h concentration of 1 or 11 for 24 h; control
cells were treated with 0.1 % DMSO in media alone. Cells were then
harvested and propidium iodide was used to stain the DNA and
analysed by FACS. For WHCO1 cells treated with either 1 or 11, a
time-dependent increase in the population of cells in the G2/M phase
was observed, with a concurrent decrease in the S and G1 populations
(see Fig. 5C). An increase in the amount of debris formed after 24 h
was also observed in both cases supportive of apoptosis induction at
24 h. Therefore, although analogues 1 and 11 are structurally different
to natural garlic compounds, they contain an exchangeable disulphide
and IC50,
concentrations, although low (yet significant) levels
24
h
were detected at the 2 × IC50, 24 h concentration at 6 h (Fig. 6A).
To probe the influence of ROS production on the cytotoxicity of 1 and
11 in WHCO1 cells, we evaluated cytotoxicity in the presence of ROS
inhibitors. Three ROS inhibitors were chosen for this experiment,
namely: N-acetyl cysteine (NAC), L-ascorbic acid (ascorbate or vitamin
C) and Trolox (a soluble form of vitamin E). Vitamin E is the primary
lipid soluble, small molecule antioxidant in cells, while vitamin C is
the corresponding water soluble antioxidant that protects lipids and
lipid structures against peroxidation [39]. NAC on the other hand is a
synthetic antioxidant that stimulates GSH synthesis, and its free radical
scavenging property arises either directly via the redox potential of
thiols, or via increasing glutathione levels in the cells [40]. The results
for compounds 1 and 11 will be discussed separately and then
compared.
First, we first established that the chosen ROS inhibitors were
capable of scavenging ROS. Here, pre-incubation of the cells with any
of the three inhibitors resulted in a reduction of ROS to levels below
that of endogenous (Fig. 6B(i)), although ascorbate was less effective
than the other two. Treatment of the cells with thiosulfonate 11 (3 h,
IC50, 24 h concentration) in the absence of any inhibitor resulted in a
1.5 fold increase in the levels of ROS compared to the control
(Fig. 6B(i), lane 2) which was completely suppressed following pre-
incubation with any of the ROS inhibitors.
To determine how ROS production relates to the cytotoxicity of 11,
cells were pre-treated with the ROS scavengers NAC, ascorbate or Trolox
and then treated with 11 for 24 h, followed by cytotoxicity
quantification by the MTT assay. It is immediately evident from
Fig. 6B(ii) that those cells pre-treated with inhibitors ascorbate (lane
6) and Trolox (lane 8) were still susceptible to the cytotoxic effects of
11. However, in cells pre-treated with NAC (lane 4), cytoprotection
was observed.
The results for disulphide 1 were somewhat different. Firstly, unlike
thiosulfonate 11, compound 1 failed to increase ROS above the back-
ground, suggesting a crucial difference between 1 and 11 in their rela-
tionship to ROS production. As with 11, though, the inhibitors alone or
a combination of inhibitor with 1 also resulted in a reduction of ROS
to levels below the background, again with ascorbate proving the less
effective of the three (see Fig. 6C(i)). A striking difference though is in
the cytotoxicity study, where a similar profile to 11 was generated,
with only NAC able to restore cell viability (lane 4, Fig. 6C(ii)), while

M. Smith et al. / Biochimica et Biophysica Acta 1860 (2016) 1439–1449
1445
Fig. 5. Cytotoxic effects of compounds 1 and 11 in WHCO1 cells. (A) Anti-proliferative activity of 1 and 11 in WHCO1 cells using the BrdU assay. Cells were treated with the compounds at
twice IC50, 48 h for 48 h. (B) Apoptosis-inducing activity of 1 and 11 in WHCO1 cells through quantification of histone-associated DNA fragments in the cytoplasm after treatment with IC50,
24 h concentration for 24 h. (C) WHCO1 cells were treated with 1 or 11 for 24 h and cells were stained with propidium iodide and analysed by FACS. Compounds 1 and 11 were found to
induce complete G2/M cell cycle arrest in WHCO1 cells when treated at twice IC50, 24 h concentration for 24 h compared to untreated DMSO controls. (D) Fold selectivity of compounds 1
and 11 for WHCO1 cells over the normal epithelial control (Het1A) measured by the MTT cell proliferation assay for 24 h. All experiments were performed independently in triplicate and
values are reported as the average standard deviation.
pre-treatment with ascorbate (lane 6) or Trolox (lane 8) failed to re-
verse the cytotoxicity of 1. This important contrast indicates that NAC
and 1 together restore cell viability via a mechanism other than scav-
enging ROS. Upon closer inspection of the chemical structure of NAC,
we decided to test whether 11 and NAC may self-react through thiolysis
(Fig. 7).
yield. It was characterised by 1H and 13C NMR spectroscopy and HRMS
to be N-acetyl cysteine methoxybenzyl disulphide, demonstrating the
feasibility of this exchange. Therefore, the reversal of 11's cytotoxicity
by NAC, in view of the results with the other ROS inhibitors cannot be
due to a quenching of ROS but is likely due to the removal of 11 from
the system via direct reaction with NAC.
3.4. Evidence for S-thiolation between 11 and NAC
4. Discussion
In a model reaction, 11 and NAC were incubated in THF at room tem-
perature overnight. A product formed that was less polar on TLC than
NAC and it was isolated by silica gel chromatography as an oil in 77%
There is much evidence to support a link between garlic consump-
tion and reduced cancer risk, especially cancers of the gastrointestinal
tract [2–4]. This link is attributed to the allyl sulphur compounds that

1446
M. Smith et al. / Biochimica et Biophysica Acta 1860 (2016) 1439–1449
Fig. 6. ROS production by compounds 1 and 11 in WHCO1 cells. (A) Timecourse for ROS production by compounds 1 and 11 in WHCO1 cells treated at 1/2 IC50, 24 h, IC50, 24 h and 2 × IC50, 24 h
concentrations or DMSO alone (control) measured from 1–6 h using the DCFH-DA assay. (B) Effect of ROS inhibitors NAC (10 mM), ascorbate (50 μM) and Trolox (12.5 μM) by 11on
(i) production of ROS after 3 h measured by DCFH-DA assay and (ii) cytotoxicity after 24 h measured by MTT assay. (C) Effect of ROS inhibitors NAC (10 mM), ascorbate (50 μM) and
Trolox (12.5 μM) by 1on (i) production of ROS after 3 h measured by DCFH-DA assay and (ii) cytotoxicity after 24 h measured by MTT assay. All experiments were performed
independently in triplicate and values are reported as the average standard deviation.
are found in crushed cloves sharing a sulphide/disulphide/polysulphide
or thiosulfinate functional group as a common structural motif. In turn,
this has prompted the notion that the underlying chemical biology
behind the anti-cancer effects of these compounds involves a
combination of glutathione and protein S-thiolation via the disulphide
pharmacophore, and generation of reactive oxygen species, which
together gives rise to a cascade of cytotoxic events leading to growth
inhibition, G2/M cell cycle arrest and induction of apoptosis.
We have recently found that ajoene targets the endoplasmic
reticulum in MDA-MB-231 breast cancer cells where it S-thiolates a
multitude of protein targets [21]. There is some evidence pointing
towards the extent of S-thiolation being linked to the cytotoxicity of
the garlic allyl sulphur compounds. For example, ajoene is expected to
be a superior thiolating agent to diallyl disulphide due to the presence
of the vinyl group which would be expected to increase the electrophi-
licity of the disulphide as well as provide a relatively more stable leaving

M. Smith et al. / Biochimica et Biophysica Acta 1860 (2016) 1439–1449
1447
Fig. 7. Reaction between N-acetyl cysteine and thiosulfonate 11 to generate the mixed disulphide 13.
group, thus promoting the forward thiolysis reaction. This may be
reflected in its enhanced cytotoxicity in cancer cells (IC50–20 μM) [13,
32] relative to that reported for DADS (IC50–60 μM) [16,41]. The present
study thus sought to build on this idea by testing a small library of
disulphides and thiosulfonates with varying leaving group activities.
Indeed, the results of the cytotoxicity experiments as measured by the
MTT assay revealed a significant correlation (R2 = 0.78, p = 0.002)
between the cytotoxicity IC50 in WHCO1 cells and the reported pKa of
the leaving group, a thermodynamic driving factor in mixed disulphide
formation. In addition, a very wide rift (1000–10,000 fold) was
observed between disulphides having a resonance-stabilising group
attached to the thiolate leaving group and those containing non-
stabilising (alkyl) groups. These results support the importance of S-
thiolation in the cytotoxic mechanism of disulphides in WHCO1 cells.
In our previous study [21], we found that ajoene accumulates in the
ER of cancer cells by a trapping mechanism involving protein S-
thiolation. Disulphides entering the cytosol are expected to become
depleted by forming a mixed disulphide with glutathione (GSH),
whereas those entering the lumen of the ER form mixed disulphides
with free cysteine thiols of newly synthesised proteins (PSH). The
cytosolic targeting to GSH vs ER targeting to protein sulfhydryl groups
may be driven by the higher concentrations of GSH that are present in
the cytosol over the ER [42]. Garlic related disulphides interfere with
cellular processes by acting as GSSG mimics, and interfering with the
redox state of the cell and with protein folding [21]. However, based
on the superior leaving group ability of garlic disulphides and related
analogues in thiolysis exchange, these mimics are expected to be supe-
rior thiolating agents compared to GSSG. Our findings that cytotoxicity
correlates to S-thiolation, as measured by the thermodynamics of the
leaving group in thiolysis exchange, supports this hypothesis. In
addition to protein S-thiolation, there is much literature evidence
supporting ROS generation by garlic allyl sulphides as an important
factor in their cytotoxicity. More specifically, ajoene [25,26], DADS [27,
28], DATS [29,30] and DAS4 [16] have all been shown to induce a
time-dependent production of ROS in a variety of cancer cell lines.
Therefore, to shed some light on this in context, mechanistic aspects
of the cytotoxicity of thiosulfonate 11 and disulphide 1 were investigat-
ed. In agreement with reports for other related garlic allyl sulphur com-
pounds, both were found to be strong inducers of G2/M cell-cycle arrest
and apoptosis. By comparison and of great significance were the con-
trasting results for 1 and 11 in ROS production (Fig. 6B and C(i)). In
spite of almost identical cytotoxicity IC50's, thiosulfonate 11 produced
striking difference in final ROS levels for these two agents, in which
those for 11 are significantly higher than those for 1 (Fig. 6), deserves
comment. The assumption that these differences are not kinetic ones
due to substrate rate differences in the disulphide exchange (assume
complete conversion for each), implies that the observed ROS level
attenuation is due to the leaving groups produced in the exchange, i.e.
a 2-pyridylthiolate (or its neutral thione) for 1 and a thiosulfinate (or
its conjugate acid) for 11. In turn, the unlikelihood of these species
being able to substitute for GSH in any enzyme-catalysed ROS detoxifi-
cation process suggests that any ROS attenuation involving them is a
chemically-based process. Here, the reactivity of thiosulfinate ion in
solution may be a factor, but a clear and satisfactory explanation of
the said results at this stage is not forthcoming, and will require further
experimentation for clarification.
Importantly, our results demonstrate that ROS generation is a
secondary effect as far as cytotoxicity is concerned for both 1 and 11,
since pre-treatment with ROS inhibitors Trolox and ascorbic acid did
not negate the cytotoxic effects. Of importance was the significant
correlation observed between the leaving group pKa and the cytotoxic-
ity of both classes of compound in the absence of any ROS consider-
ations. By comparison, pre-treating in the individual case with the ROS
inhibitor NAC did result in shutting down the cytotoxicity of both 11
and 1, even though 1 failed to produce any significant increase in ROS.
Inspection of the chemical structure of NAC suggests that NAC may
chemically react with disulphides. This possibility was investigated by
performing a chemical reaction between NAC and the disulphide 11 in
THF overnight, which indeed resulted in a reaction product between
inhibitor and compound to form a mixed disulphide. We therefore ex-
plain the apparent reversal of the cytotoxicity by 11 through the ability
of NAC to reduce and inactivate the disulphide test compound.
A number of downstream events have been linked to ROS produc-
tion by garlic allyl sulphur compounds that include activation of the
cJun NH2 terminal kinase cJun signalling cascade [27,29,45],
hyperphosphorylation of Cdc25C [30] and activation of the mitochon-
drial membrane caspase cascade [46]. However quite a few of these
studies have drawn their conclusions based on inhibitor studies using
NAC [26,30,46,47]. Other ROS cytotoxicity studies, however, have also
observed positive effects which do not make use of NAC. For example
Filomeni et al. observed that overexpression of the antioxidant enzyme
Cu–Zn SOD offered some protection against lipid peroxidation, protein
carbonylation, G2/M cell cycle arrest and apoptosis in SH-SY5Y neuro-
blastoma cells and that use of the spin trap DMPO offered some protec-
tion against G2/M cell cycle arrest and apoptosis [27]. In another study,
addition of the SOD and catalase mimetic, EUK134, was found to offer
some protection against G2/M cell cycle arrest and Cdc25C phosphoryla-
tion [29]. In a study by Das et al., ascorbate pre-treatment was found to
reverse p38 MAPK phosphorylation and JNK1 activation by DAS and
DADS, thereby highlighting the importance of ROS production in this
signalling cascade in human glioblastoma T98G and U87MG cells. In a
study by Kelkel et al., the authors failed to detect any ROS production
by DAS4, and supportive of our results, found that NAC but not Trolox
pre-treatment resulted in reversal of the cytotoxic effects by DAS4 in
histiocytic lymphoma cells [17].
high levels of ROS after only 3 h and at concentrations below the IC50
.
Compound 1 however produced very low levels of ROS which were
only significant at twice the IC50. Since these two compounds have
similar IC50's due to a similar degree of thiolysis, the contrasting ROS
results may imply that the different leaving groups have a bearing on
ROS generation.
Both allicin [43,44] and 2-propenylthiosulfinate [20] have been
shown to spontaneously react with GSH in vitro. Therefore, it is expect-
ed that disulphides and thiosulfonates entering the cytosol will undergo
spontaneous thiolysis exchange with GSH. This reaction would lower
the levels of intracellular GSH and lead to increased ROS. However, the

1448
M. Smith et al. / Biochimica et Biophysica Acta 1860 (2016) 1439–1449
Fig. 8. Proposed cytotoxic mechanism of thiosulfonate 11 in WHCO1 cancer cells. Thiosulfonate (11) accumulates in the ER where it S-thiolates newly synthesised proteins and induces ER
stress. This reaction is thermodynamically driven by the stability of the leaving group. Thiosulfonate 11 entering the cytosol S-thiolates glutathione leading to lowered GSH:GSSG levels;
and subsequent increased levels of ROS.
In the current study the finding that NAC alone is able to nullify the
cytotoxicity of thiosulfonate 11 and disulphide 1 strongly suggests a
reaction between the inhibitor and the compound. If correct, this
important finding suggests that results obtained from inhibitor studies
between NAC and garlic-related disulphides should be interpreted
with caution. Needless to say, there are inhibitor studies that utilise
inhibitors other that NAC, which have found a positive correlation
between ROS production and G2/M cell cycle arrest and apoptosis in
SH-SY5Y neuroblastoma cells [27], and G2/M cell cycle arrest and
Cdc25C phosphorylation in DU145 and PC3 prostate cancer cells [29].
We have recently found that ajoene targets the ER in MDA-MB-231
breast cancer cells [21]. Being lipophilic, we propose that ajoene and re-
lated disulfides may rapidly cross the membrane to enter the cytoplasm
via passive diffusion [43]. The cytoplasm is highly reducing due to high
levels of GSH which is expected to reduce the disulphides by thiolysis,
thereby lowering the GSH levels and allowing ROS to increase (see
Fig. 8). Thiolysis exchange in the current study on disulphides and
thiosulfonates has been shown to be thermodynamically driven by the
leaving group stability in a mixed disulphide reaction.
Town (UCT), the Cancer Society of South Africa (CANSA) and the Cancer
Research Trust.
Appendix A. Supplementary data
Supplementary data to this article can be found online at http://dx.
doi.org/10.1016/j.bbagen.2016.03.032.
References
[1] E. Dorant, P. Van Den Brandt, R. Goldbohm, R. Hermus, F. Sturmans, Garlic and its
significance for prevention of cancer in humans: a critical view, Br. J. Cancer 67
(1993) 424–429.
[2] A.T. Fleischauer, L. Arab, Garlic and cancer: a critical review of the epidemiologic
literature, J. Nutr. 3S (2001) 1032S–1040S.
[3] J.Y. Kim, O. Kwon, Garlic intake and cancer risk: an analysis using the Food and Drug
Administration's evidence-based review system for the scientific evaluation of
health claims, Am. J. Clin. Nutr. 89 (2009) 257–264.
[4] S.N. Ngo, D.B. Williams, L. Cobiac, R.J. Head, Does garlic reduce risk of colorectal
cancer? A systematic review, J. Nutr. 137 (2007) 2264–2269.
[5] Y. Zhou, W. Zhuang, W. Hu, G.J. Liu, T.X. Wu, X.T. Wu, Consumption of large amounts
of Allium vegetables reduces risk for gastric cancer in a meta-analysis, Gastroenter-
ology 141 (2011) 80–89.
[6] E. Block, S. Ahmad, J.L. Catalfamo, M.K. Jain, R. Apitz-Castro, The chemistry of alkyl
thiosulfinate esters. 9. Antithrombotic organosulfur compounds from garlic:
structural, mechanistic, and synthetic studies, J. Am. Chem. Soc. 108 (1986)
7045–7055.
[7] J.T. Pinto, B.F. Krasnikov, A.J. Cooper, Redox-sensitive proteins are potential targets
of garlic-derived mercaptocysteine derivatives, J. Nutr. 136 (2006) 835S–841S.
[8] A. Herman-Antosiewicz, A. Powolny, S.V. Singh, Molecular targets of cancer
chemoprevention by garlic-derived organosulfides, Acta Pharmacol. Sin. 9 (2007)
1355–1364.
5. Conclusion
The current study has delineated the effects of thiolysis exchange
and ROS production by disulphides and thiosulfonates in WHCO1
oesophageal cancer cells. The thermodynamics of thiolysis was found
to correlate to cytotoxicity as measured by the pKa of the leaving
group in thiolysis exchange. This supports our previous findings on
ajoene, which demonstrated ER targeting and an accumulation of
misfolded proteins. We conclude that protein S-thiolation by these
agents is the determining factor in their cytotoxicity, driving ER stress
and the unfolded protein response. Our results also suggest that ROS
is generated as a result of decreased GSH levels following thiolysis ex-
change and that the role of the leaving group on ROS attenuation
needs further investigation. The current study has relevance to the
design of effective cytotoxic agents against cancer cells.
[9] Y. Shukla, N. Kalra, Cancer chemoprevention with garlic and its constituents, Cancer
Lett. 247 (2007) 167–181.
[10] K. Agarwal, Therapeutic actions of garlic constituents, Med. Res. Rev. 16 (1996)
111–124.
[11] C.H. Kaschula, R. Hunter, H.T. Hassan, N. Stellenboom, J. Cotton, X.Q. Zhai, M.I. Parker,
Anti-proliferative activity of synthetic ajoene analogues on cancer cell lines, Anti
Cancer Agents Med. Chem. 11 (2011) 260–266.
[12] K. Scharfenberg, R. Wagner, K.G. Wagner, The cytotoxicity effect of ajoene, a natural
product from garlic, investigated with different cell lines, Cancer Lett. 53 (1990)
103–108.
[13] M. Li, J.-R. Ciu, Y. Ye, J.-M. Min, L.-H. Zhang, K. Wang, M. Gares, J. Cros, M. Wright, J.
Leung-Track, Antitumor activity of Z-ajoene, a natural compound purified from gar-
lic: antimitotic and microtubule-interaction properties, Carcinogenesis 23 (2002)
573–579.
Transparency document
[14] U. Münchberg, A. Anwar, S. Mecklenburg, C. Jacob, Polysulfides as biologically active
ingredients of garlic, Org. Biomol. Chem. 5 (2007) 1505–1518.
The Transparency document associated with this article can be
found, in online version.
[15] K. Aquilano, P. Vigilanza, G. Filomeni, G. Rotilio, M.R. Ciriolo, Tau dephosphorylation
and microfilaments disruption are upstream events of the anti-proliferative effects
of DADS in SH-SY5Y cells, J. Cell. Mol. Med. 14 (2010) 564–577.
[16] T. Hosono, T. Fukao, J. Ogihara, Y. Ito, H. Shiba, T. Seki, T. Ariga, Diallyl trisulfide sup-
presses the proliferation and induces apoptosis of human colon cancer cells through
oxidative modification of β-tubulin, J. Biol. Chem. 280 (2005) 41487–41493.
[17] M. Kelkel, C. Cerella, F. Mack, T. Schneider, C. Jacob, M. Schumacher, M. Dicato, M.
Diederich, ROS-independent JNK activation and multisite phosphorylation of Bcl-2
Acknowledgements
This work was supported by grants and fellowships from the Nation-
al Research Foundation of South Africa (NRF), the University of Cape

M. Smith et al. / Biochimica et Biophysica Acta 1860 (2016) 1439–1449
1449
link diallyl tetrasulfide-induced mitotic arrest to apoptosis, Carcinogenesis 33
(2012) 2162–2171.
[34] A.A. Powolny, S.V. Singh, Multitargeted prevention and therapy of cancer by diallyl
trisulfide and related Allium vegetable-derived organosulfur compounds, Cancer
Lett. 269 (2008) 305–314.
[18] D. Xiao, J.T. Pinto, J.-W. Soh, A. Deguchi, G.G. Gundersen, A. Palazzo, J.-T. Yoon, H.
Shirin, I.B. Weinstein, Induction of apoptosis by the garlic-derived compound S-
allylmercaptocysteine (SAMC) is associated with microtubule depolymerization
and c-Jun NH2-terminal kinase 1 activation, Cancer Res. 63 (2003) 6825–6837.
[19] H. Gallwitz, S. Bonse, A. Martinez-Cruz, I. Schlichting, K. Schumacher, R.L. Krauth-
Siegel, Ajoene is an inhibitor and subversive substrate of human glutathione reduc-
tase and Trypanosoma cruzi trypanothione reductase: crystallographic, kinetic, and
spectroscopic studies, J. Med. Chem. 42 (1999) 364–372.
[35] C.H. Kaschula, R. Hunter, M.I. Parker, Garlic-derived anti-cancer agents: structure
and biological activity of ajoene, Biofactors 36 (2010) 78–85.
[36] A. Herman-Antosiewicz, S.V. Singh, Signal transduction pathways leading to cell
cycle arrest and apoptosis induction in cancer cells by Allium vegetable derived
organosulfur componds: a review, Mutat. Res. 555 (2004) 121–131.
[37] B. Kalyanaraman, V. Darley-Usmar, K.J. Davies, P.A. Dennery, H.J. Forman, M.B.
Grisham, G.E. Mann, K. Moore, L.J. Roberts II, H. Ischiropoulos, Measuring reactive
oxygen and nitrogen species with fluorescent probes: challenges and limitations,
Free Radic. Biol. Med. 52 (2012) 1–6.
[20] R. Nepravishta, R. Sabelli, E. Iorio, L. Micheli, M. Paci, S. Melino, Oxidative species and
S-glutathionyl conjugates in the apoptosis induction by allyl thiosulfate, FEBS J. 279
(2012) 154–167.
[38] A. Gomes, E. Fernandes, J.L. Lima, Fluorescence probes used for detection of reactive
[21] C.H. Kaschula, R. Hunter, J. Cotton, R. Tuveri, E. Ngarande, K. Dzobo, G. Schäfer, V.
Siyo, D. Lang, D.A. Kusza, B. Davies, A.A. Katz, M.I. Parker, The garlic compound
ajoene targets protein folding in the endoplasmic reticulum of cancer cells, Mol.
Carcinog. (2015) (Ahead of Print).
oxygen species, J. Biochem. Biophys. Methods 65 (2005) 45–80.
[39] G.R. Buettner, The pecking order of free radicals and antioxidants: lipid peroxida-
tion, alpha-tocopherol, and ascorbate, Arch. Biochem. Biophys. 300 (1993) 535–543.
[40] M. Halasi, M. Wang, T.S. Chavan, V. Gaponenko, N. Hay, A.L. Gartel, ROS inhibitor N-
acetyl-L-cysteine antagonizes the activity of proteasome inhibitors, Biochem. J. 454
(2013) 201–208.
[22] H. Malhi, R.J. Kaufman, Endoplasmic reticulum stress in liver disease, J. Hepatol. 54
(2011) 795–809.
[23] A.I. Bhuiyan, V.T. Papajani, M. Paci, S. Melino, Glutathione–garlic sulfur conjugates:
slow hydrogen sulfide releasing agents for therapeutic applications, Molecules 20
(2015) 1731–1750.
[41] D. Xiao, J.T. Pinto, G.G. Gundersen, I.B. Weinstein, Effects of a series of organosulfur
compounds on mitotic arrest and induction of apoptosis in colon cancer cells, Mol.
Cancer Ther. 4 (2005) 1388–1398.
[24] C.J. Cavallito, J.S. Buck, C.M. Suter, Allicin, the antibacterial principle of Allium
sativum. II. Determination of the chemical structure, J. Am. Chem. Soc. 66 (1944)
1952–1954.
[25] K. Scharfenberg, T. Ryll, R. Wagner, K.G. Wagner, Injuries to cultivated BJA-B cells by
ajoene, a garlic-derived natural compound: cell viability, glutathione metabolism,
and pools of acidic amino acids, J. Cell. Physiol. 158 (1994) 55–60.
[26] V.M. Dirsch, A.L. Gerbes, A.M. Volmar, Ajoene, a compound of garlic, induces apopto-
sis in human promyleoleukemic cells, accompanied by generation of reactive oxy-
gen species and activation of nuclear factor kB, Mol. Pharmacol. 53 (1998) 402–407.
[27] G. Filomeni, K. Aquilano, G. Rotilio, M.R. Ciriolo, Reactive oxygen species-dependent
c-Jun NH2-terminal kinase/c-Jun signaling cascade mediates neuroblastoma cell
death induced by diallyl disulfide, Cancer Res. 63 (2003) 5940–5949.
[28] J.-S. Yang, G.-W. Chen, T.-C. Hsia, H.-C. Ho, C.-C. Ho, M.-W. Lin, S.-S. Lin, R.-D. Yeh, S.-
W. Ip, H.-F. Lu, J.-G. Chung, Diallyl disulfide induces apoptosis in human colon cancer
cell line (COLO 205) through the induction of reactive oxygen species, endoplasmic
reticulum stress, caspases casade and mitochondrial-dependent pathways, Food
Chem. Toxicol. 47 (2009) 171–179.
[42] C. Hwang, A.J. Sinskey, H.F. Lodish, Oxidized redox state of glutathione in the
endoplasmic reticulum, Science 257 (1992) 1496–1502.
[43] T. Miron, A. Rabinkov, D. Mirelman, M. Wilchek, L. Weiner, The mode of action of
allicin: its ready permeability through phospholipid membranes may contribute
to its biological activity, Biochim. Biophys. Acta 1463 (2000) 20–30.
[44] A. Rabinkov, T. Miron, D. Mirelman, M. Wilchek, S. Glozman, E. Yavin, L. Weiner, S-
Allylmercaptoglutathione: the reaction product of allicin with glutathione possesses
SH-modifying and antioxidant properties, Biochim. Biophys. Acta 1499 (2000)
144–153.
[45] A. Das, N.L. Banik, S.K. Ray, Garlic compounds generate reactive oxygen species lead-
ing to activation of stress kinases and cysteine proteases for apoptosis in human
glioblastoma T98G and U87MG cells, Cancer 110 (2007) 1083–1095.
[46] Y.A. Kim, D. Xiao, H. Xiao, A.A. Powolny, K.L. Lew, M.L. Reilly, Y. Zeng, Z. Wang, S.V.
Singh, Mitochondria-mediated apoptosis by diallyl trisulfide in human prostate can-
cer cells is associated with generation of reactive oxygen species and regulated by
Bax/Bak, Mol. Cancer Ther. 6 (2007) 1599–1609.
[47] B. Pagliei, K. Aquilano, S. Baldelli, M.R. Ciriolo, Garlic-derived diallyl disulfide modu-
lates peroxisome proliferator activated receptor gamma co-activator 1 alpha in neu-
roblastoma cells, Biochem. Pharmacol. 85 (2013) 335–344.
[29] J. Antosiewicz, A. Herman-Antosiewicz, S. Marynowski, S. Singh, c-Jun NH2-terminal
kinase signalling axis regulates diallyl trisulfide-induced generation of reacive oxy-
gen species and cell arrest in human prostrate cancer cells, Cancer Res. 66 (2006)
5379–5385.
[48] Kirk-Othmer Encyclopedia of Chemical Technology, fourth ed. John Wiley and Sons
Inc, New York, 1998.
[30] D. Xiao, A. Herman-Antosiewicz, J. Antosiewicz, H. Xiao, M. Brisson, J.S. Lazo, S.V.
Singh, Diallyl trisulfide-induced G2-M phase arrest in human prostate cancer cells
is caused by reactive oxygen species-dependent destruction and
hyperphosphorylation of Cdc25C, Oncogene 24 (2005) 6256–6268.
[31] S. Stellenboom, R. Hunter, M.R. Caira, One-pot synthesis of unsymmetrical disulfides
using 1-chlorobenzotriazole as an oxidant: interception of the sulfenyl chloride in-
termediate, Tetrahedron 66 (2010) 3228–3241.
[32] C.H. Kaschula, R. Hunter, N. Stellenboom, M.R. Caira, S. Winks, T. Ogunleye, P.
Richards, J. Cotton, K. Zilbeyaz, Y. Wang, V. Siyo, E. Ngarande, M.I. Parker,
Structure–activity studies into the anti-proliferation activity of ajoene derivatives
in WHCO1 oesophageal cancer cells, Eur. J. Med. Chem. 50 (2012) 236–254.
[33] R.B. Veale, A.L. Thornley, Increased single class low-affinity EGF receptors expressed by
human oesophageal squamous carcinoma cell lines, S. Afr. J. Sci. 85 (1989) 375–379.
[49] R. Williams, pKa data. Available from: http://www.research.che.edu.psu.edu/
brpgroup/pKa_compilation.pdf
[50] G.L. Ellman, A colorimetric method for determining low concentrations of mercap-
tans, Arch. Biochem. Biophys. 74 (1958) 443–450.
[51] L.D. Pettit, H.K.J. Powell, in: A. software (Ed.), Stability Constant Database, 1993
(Timble, Otley, Yorks LS21 2PW, UK).
[52] X. Wang, J. Cao, C. Zhao, Design of a ratiometric fluorescent probe for benzenethiols
based on a thiol-sulfoxide reaction, Org. Biomol. Chem. 10 (2012) 4689–4691.
[53] J.T. Edsall, J. Wymann, Biophysical Chemistry, Academic Press Inc, New York, 1958.
[54] C.F. Bernasconi, R.B. Killion, J. Am. Chem. Soc. 110 (1988) 7506–7512.
[55] D. De Filippo, F. Momicchioli, Benzenesulfinic and seleninic acids. Determination
and theoretical interpretation of pK, Tetrahedron 25 (1969) 5733–5744.