Hypochlorous acid scavenging activities of thioallyl compounds from garlic

Article abstract of DOI:10.1021/jf102423w

The hypochlorous acid (HOCl) scavenging capacities of 10 garlic compounds containing modifications in the thioallyl group (-S-CH₂CH=CH ₂) were determined by a catalase protection assay, and the corresponding structure-activity relationships using molecular descriptors were calculated. This scavenging activity was enhanced by increasing the number of S atoms or by the alanyl group (-CH₂CH-NH₂-COOH) and decreased in the absence of the C=C bond or in the presence of a sulfoxide group in the thioallyl group. Interestingly, S-allylcysteine and its corresponding sulfoxide (alliin) showed the highest and lowest HOCl-scavenging capacities, respectively. Quantitative modeling by multiple regression analysis and partial least-squares projections showed that the topological descriptor polar surface area and two electronic properties, namely, highest occupied molecular orbital and total energy, contributed mainly to variations in the HOCl scavenging activity of thioallyl compounds. These observations provide new insights on the antioxidant mechanism of garlic derivatives in processes involving HOCl production.

Full text of DOI:10.1021/jf102423w

11226 J. Agric. Food Chem. 2010, 58, 11226–11233
DOI:10.1021/jf102423w
Hypochlorous Acid Scavenging Activities of Thioallyl
Compounds from Garlic
†
#
§
´
.
´
´
´
R
AUL
A
RGUELLO-GARCIA, OMAR N. MEDINA-CAMPOS, NURY
P
EREZ-HERNANDEZ
,†
,
#
´
´
J
OSE
P
EDRAZA-CHAVERRI
,
AND
GUADALUPE
ORTEGA-PIERRES*
†
Departamento de Gene
7360 Mexico City, Mexico, Departamento de Biologı
Universidad Nacional Autonoma de Mexico, 04510 Mexico City, Mexico, and Programa Institucional de
Biomedicina Molecular, Escuela Nacional de Medicina y Homeopatıa, Instituto Politecnico Nacional,
Mexico City 07320, Mexico
´
tica y Biologı
´
a Molecular, Centro de Investigacio
´
n y de Estudios Avanzados-IPN,
§
0
´
a, Facultad de Quı
´
mica, Edificio “F”, Lab. 209,
#
´
´
´
´
The hypochlorous acid (HOCl) scavenging capacities of 10 garlic compounds containing modifica-
tions in the thioallyl group (-S;CH CHdCH ) were determined by a catalase protection assay, and
2
2
the corresponding structure-activity relationships using molecular descriptors were calculated. This
scavenging activity was enhanced by increasing the number of S atoms or by the alanyl group
(
-CH
sulfoxide group in the thioallyl group. Interestingly, S-allylcysteine and its corresponding sulfoxide
alliin) showed the highest and lowest HOCl-scavenging capacities, respectively. Quantitative
2 2
CH;NH ;COOH) and decreased in the absence of the CdC bond or in the presence of a
(
modeling by multiple regression analysis and partial least-squares projections showed that the
topological descriptor polar surface area and two electronic properties, namely, highest occupied
molecular orbital and total energy, contributed mainly to variations in the HOCl scavenging activity of
thioallyl compounds. These observations provide new insights on the antioxidant mechanism of
garlic derivatives in processes involving HOCl production.
KEYWORDS: Garlic; thioallyl compounds; hypochlorous acid; ROS scavenging; catalase protection
assay; molecular descriptors; structure-activity relationship
INTRODUCTION
antioxidants, membrane lipids, and DNA are also affected (4).
These processes are related to host tissue damage and to the
establishment of inflammatory diseases such as arthritis, cystic
fibrosis, asthma, heart disease, and some types of cancer (5-7).
The use of exogenous antioxidants from synthetic and natural
sources is an ongoing strategy for treating these oxidative stress-
mediated abnormalities.
Garlic (Allium sativum L.) is a perennial crop used worldwide
since ancient times as a food spice with a variety of medicinal
qualities. These include antimicrobial as well as health care prop-
erties such as anticancer, antihypertensive, antithrombotic,
hepato- and cardioprotective, and hypolipidemic activities
Oxidative damage is an important trait in the pathophysiology
of several conditions such as aging, cancer, and diseases involving
neurodegenerative, cardiovascular, or inflammatory processes (1).
•
-
Reactive oxygen species (ROS) such as superoxide (O ),
2
•
•
-
hydroxyl (HO ), peroxyl (HOO ), peroxynitrite (ONOO ), or
hypohalous acids (HOX, X = Cl, Br) are widely recognized
effectors in the oxidative transformation of cellular membranes
and DNA. ROS may be exogenously derived from the environ-
ment or endogenously produced in metabolic pathways during the
transformation of normal or xenobiotic substrates (2). Among
these strong oxidants, HOX have potent antibacterial activities
under normal function of the mammalian immune system. These
are produced in vivo by the heme-peroxidase enzyme myelo-
peroxidase from either activated neutrophils, monocytes, and
macrophages (MPO, HOCl-forming) or eosinophils (EPO,
HOBr-forming), and the predominant species corresponds to
HOCl (3). These HOX species may be controlled by endogenous
antioxidant systems such as superoxide dismutase, catalase, uric
acid, glutathione peroxidase, and glutatione S-transferase among
others (1). However, if HOX levels are excessive or misplaced
exceeding these systems, HOX react directly with amino acids and
proteins as major targets, but carbohydrates, thiol-containing
(
reviewed in ref 8). Substantial evidence suggests that these
biological activities are due to the unique organosulfur com-
pounds (OSCs) of garlic cloves, which have diverse molecular
effects including enzyme modulation and potent antioxidant
activities mediated by radical scavenging (9). OSCs commonly
found in garlic include four types of thioallyl compounds (TAC):
S-allylcysteine sulfoxide (alliin, the most abundant sulfur com-
pound contained in mesophyll storage cells), S-allylcysteine
(
obtained by long-time organic fermentation of whole bulbs in
ethanolic solutions), allyl thiosulfinates (e.g., allicin, obtained by
the transformation of alliin by the vascular sheath cell enzyme
allinase when bulbs are cut, crushed, or chopped in water), and
some products of allicin decomposition (allyl sulfides and ajoenes)
(10). Interestingly, aqueous extracts from raw or powdered garlic
*
Address correspondence to this author (phone þ52 55 5747-3331;
fax þ52 55 5747-3392; e-mail gortega@cinvestav.mx).
pubs.acs.org/JAFC
Published on Web 10/13/2010
© 2010 American Chemical Society

Article
J. Agric. Food Chem., Vol. 58, No. 21, 2010 11227
•
-
•
•
-
are able to scavenge O2 , HO , HOO , ONOO , and HOCl
species (11-13). However, these abilities may be attributed to
multiple TAC from garlic as these exhibit different patterns of
radical scavenging as previously reported, that is, alliin scavenged
acetone diluted 1:2 with water. AA was disolved in 50 mM phosphate
buffer, pH 7.4.
HOCl Scavenging Assay. The elimination of the catalase peak as a
consequence of the destruction of the heme prosthetic group by HOCl was
determined spectrophotometrically at 404 nm (13). The HOCl scavenging
capacity of garlic TAC or the reference compounds (AA and NAC) was
made evident by the inability of HOCl to eliminate/decrease the peak in a
concentration-dependent way. Experiments were carried out essentially as
described (19), and data were registered in a DU-640 series Beckman
spectrophotometer (Beckman Coulter, Fullerton, CA). A solution of
49.8 μM bovine liver catalase (16.6 μM, final concentration) was mixed
with 18 mM HOCl (6 mM, final concentration) with increasing concen-
trations of organosulfur compounds (Table 1). AA and NAC were used as
reference compounds (13).
Determination of HOCl Scavenging Capacity. Scavenging capa-
city was expressed as 50% of the inhibitory concentration (IC50) value,
which represents the concentration of organosulfur test or reference
compounds required to give a 50% reduction in catalase destruction
relative to the one obtained in the tube that did not contain organosulfur
compound or reference compounds. The IC50 of each test compound was
calculated by the least-squares method. The lower the IC50 value, the
higher the scavenging capacity of the compound (13).
Structure-Activity Relationship Analyses. The three-dimensional
structures of the studied compounds were geometrically optimized using
the SPARTAN ’04 package from Wave function (Irvine, CA). The dipole
moment (μ) and total energy (ETOTAL) and the corresponding energies for
highest occupied molecular orbitals (EHOMO) and lowest unoccupied
molecular orbitals (ELUMO) were calculated by the semiempirical quan-
tum-chemical method AM1. For each molecule, the conformer with the
major contribution to Boltzman population as found by Monte Carlo
search was considered. In addition, two properties related to molecular
transport (absorption prediction) were included. One was the octanol/
water partition coefficient (log Poct), the values of which were taken as
reported in previous studies for AMS, DPS, DPDS, and ALC (20), for
DAS, DADS, and DATS (21), and for SAC and ALI (22), whereas for
AM this value was obtained from the Good Scents Co. database. The
other molecular transport property analyzed was the molecular polar
surface area (PSA). This was calculated for all compounds following the
fragment-based contribution procedure (23). The relationship between the
values of each of these molecular descriptors and their HOCl scavenging
activities (expressed as IC50 values) was determined by multiple regression
analysis (MLR), and the relative contributions of each descriptor to
simplified models were cross-validated by the partial least-squares (PLS)
procedure, both tools contained in the 16.0.18 version of Statgraphics
Centurion XVI software.
•-
•
•
•-
O2 and HO , allyl disulfide scavenged HO but not O , and
allicin was unable to scavenge O2 (14) and its HO and HOO
2
•
-
•
•
scavenging abilities are controversial (14-16). S-Allylcysteine is
•
-
•
•
-
able to scavenge O2 ,HO ,HOO ,ONOO ,andHOCl(14,17,18).
With regard to the HOCl scavenging activity of garlic aqueous
extracts, it is eliminated if intact cloves are microwave-treated but
not if these are heated before or after cutting, suggesting that
allinase activity (i.e., allicin formation) may not be critical (13).
In the present work we have analyzed the ability of eight TAC
from garlic and two propyl sulfide analogues from onion to
scavenge HOCl using the absorption spectrum of the antioxidant,
heme-containing enzyme catalase as a model that tests the
disrupting action of hypochlorite over the prosthetic group
(
13, 19). In addition, the structures of the 10 compounds under
study were characterized by calculating several nonempirical
quantum descriptors, especially those related to electronic and
molecular transport properties, and a quantitative modeling
study was carried out using multiple regression analysis (MLR)
and partial least-squares (PLS) procedures. Thus, the main
purpose of these analyses was to obtain structure-activity para-
meters from TAC of garlic to understand their different HOCl
scavenging capacities, which can provide insights for further
molecular design of ROS-specific scavengers.
MATERIALS AND METHODS
Reagents. Bovine liver catalase, ascorbic acid (AA), and N-acetyl-
cysteine (NAC) were from Sigma Chemical Co. (St. Louis, MO), and
sodium hypochlorite (NaOCl), sulfuric acid, and acetone were from
J. T. Baker (Xalostoc, Edo. Mexico, Mexico). The following TAC were
´
used: alliin (S-allylcysteine sulfoxide, ALI), diallyl sulfide (DAS), and
dipropyl sulfide (DPS) from Fluka Chemika AG (Buchs, Switzerland);
diallyl disulfide (DADS), dipropyl disulfide (DPDS), allyl methyl sulfide
AMS), and allyl mercaptan (AM) from Aldrich Chemical Co. (Milwaukee,
WI); S-allyl- -cysteine (SAC) from TCI (Tokyo Kasei Kogyo Co., Tokyo,
Japan); and diallyl trisulfide (DATS) from MP Biomedicals Inc. (Solon,
OH). Allicin (diallyl thiosulfinate, ALC) was produced by oxidation of
DADS using the protocol reported by Lawson (17). In this, 1 g of DADS
was dissolved in 5 mL of acetic acid under stirring in an ice bath. Hydrogen
peroxide (1.5 mL, 30% v/v) was added stepwise, and the reaction was
allowed to proceed for 30 min. After this step, the reaction was kept at
(
L
RESULTS
1
3 °C for 20 min, and it was placed on an ice bath again for 5 h to minimize
HOCl Scavenging Activities of Garlic TAC. The 10 TAC tested
at different concentrations were able to re-establish at different
degrees the absorption peak of catalase at 404 nm that was
abolished by HOCl attack when TAC were not present (compare
curves 1 and 2 in each spectrophotometric pattern shown in
Figure 1). This protective effect was performed in a dose-dependent
manner and was compound-specific because none of the vehicles
used (acetone and water) affected the catalase peak. The two
reference compounds (AA and NAC) also protected catalase
the content of remaining DADS. Reaction was stopped by adding 15 mL
of distilled water at pH 6.5, and it was extracted with dichloromethane (3 ꢀ
30 mL). The combined organic extracts were washed with 5% of sodium
bicarbonate (3 ꢀ 20 mL) and distilled water (2 ꢀ 20 mL) and dried over
Na SO . The solvent was left to evaporate until yellowish oil (ALC, yield
2
4
of approximately 47% per assay) was obtained. For stabilization and
storage of ALC, the oily product was resuspended in water at 1.5% (w/v)
and kept in 1 mL aliquots at -70 °C until analyzed and used. Under these
conditions, ALC retained its full activity for up to 6 months. The
identification of ALC was performed by nuclear magnetic resonance
(
not shown), although the latter was more efficient; indeed,
1
(
3
1
C and H spectra) using a JEOL Eclipse spectrometer with CDCl
3
as
these compounds served as reference scavengers of low and
high efficiency (IC50 = 3.24 ( 0.31 and 0.99 ( 0.07 mM,
respectively).
solvent and tetramethylsilane as reference. The purity of ALC was
determined by HPLC (Waters model 996 coupled to a Waters model
6
00 pump) using 4.6 ꢀ 250 mm Spherisorb columns (10 μM particle size),
On the basis of the concentration-response data, the IC50
values were calculated for all test compounds (Table 1). The activ-
ities ranged from 0.56 to 3.57mM. The simplest TAC tested (AM)
had an intermediate efficacy (1.45 mM) that was decreased by
substituting the H atom by a methyl group as in AMS (1.88 mM).
If the H atom was substituted by another allyl group as in DAS,
the IC50 value increased to a lower extent (1.68 mM). When the
two CdC double bonds of DAS were saturated, as in the onion
derivative DPS, the scavenging capacity dramatically decreased
mobile phase 30% methanol/70% water, flow rate = 0.4 mL/min, and UV
detection at 254 nm, rendering an average time retention of 3.61 min and a
purity of 90-92% for synthetic allicin.
Preparation of HOCl and TAC. HOCl was obtained immediately
before use by mixing a 1% (v/v) solution of NaOCl with 0.6 M sulfuric
acid at pH 6.2. The HOCl concentration was further determined spectro-
photometrically at 235 nm using the molar extinction coefficient of
-1
-1
1
5
00 M cm . Solutions of ALC, SAC, and ALI were prepared in
0 mM phosphate buffer, pH 7.4; all other compounds were dissolved in

11228 J. Agric. Food Chem., Vol. 58, No. 21, 2010
Arg €u ello-Garc ´ı a et al.
a
Table 1. Chemical Structures and HOCl Scavenging Activity of Eight Thioallyl Compounds from Garlic and Two Related Propyl Sulfides from Onion
a
IC50 = millimolar concentration of compound required to achieve 50% of HOCl scavenging capacity. Data are the mean ( standard deviation (S.D.) of three separate
experiments.
to 3.16 mM. In fact, DPS was one of the two compounds (along
with ALI) exhibiting the lowest HOCl scavenging activity. Other-
wise, a significant enhancing effect of scavenging capacity was
observed when one additional S atom was present as in DADS
Quantitative Estimations of the Structure-HOCl Scavenging
Activity Relationship of TAC. The values of 4 electronic and 2
molecular transport descriptors were retrieved or calculated for
the 10 TAC tested and these are listed in Table 2. With regard
to the electronic descriptors, some trends were preliminarily
observed. The ETOTAL parameter ranged from ≈ þ30 kcal/mol
(DAS, DADS, and DATS, good HOCl scavengers) to negative
values of ≈-111 kcal/mol (ALI, the lowest HOCl scavenger),
albeit SAC, the best scavenger tested, had a rather negative value
(-87 kcal/mol). The EHOMO values were always negative in a
narrow range where the lowest scavenger (ALI) exhibited the
highest negative value (-230 kcal/mol) and SAC the lowest
negative one (-191 kcal/mol). Otherwise, the ELUMO descriptor
exhibited a range from -51 kcal/mol (DATS) to þ20 kcal/mol
(DPS) without a suggestive trend. This was also the case for the
absolute values of the dipole moment, which ranged from 1.40
(DATS) to 3.49 D (ALI), the latter being the unique compound
with μ higher than SAC (2.30 D). With rgard to the molecular
transport descriptors, neither the lipophilicity (as depicted by log
Poct) nor the topological PSA seemed to keep a significant
relationship to HOCl scavenging capacity because the more
hydrophilic and polar compounds (SAC and ALI) exhibited
opposite scavenging activities. From these observations, it was
very likely that more than a single property should explain the
differential scavenging activities of TAC tested.
(1.02 mM), whereas the corresponding onion derivative with satu-
rated double bounds, DPDS, displayed lower activity (1.37 mM).
Two other modifications of the structure of DADS rendered a
differential trend in scavenger efficacy: when one additional S atom
was present as in DATS, the scavenging activity was even better
(0.89 mM); however, the addition of one oxo group (Od) to one of
the two S atoms, as was the case of diallyl thiosulfinate (ALC),
resulted in a lower scavenger efficacy (1.17 mM). By substituting
the thioallyl group in ALC by an alanyl group (-CH CH;
2
NH -COOH), which gave rise to alliin (ALI), a further dramatic
2
decrease in scavenging ability was observed (3.57 mM), and this
compound was the lowest HOCl scavenger tested. Interestingly,
the simple removal of the oxo group from ALI to became SAC
gave rise to the more dramatic enhancing effect on scavenging
capacity that indeed converted SAC to be the most efficient
scavenger tested (IC50 = 0.56 mM).
As a next step, it was of interest to compare the HOCl
scavenging activity of the TAC tested with that observed in two
reference scavengers and antioxidants of physiological impor-
tance as are AA and NAC. For statistical estimations, Dunnett’s
multiple-comparisons test was used. In this case, AA behaved as a
scavenger of low efficiency (IC50 = 3.49 ( 0.31 mM), and all
compounds tested except for DPS and ALI were more effective
HOCl scavengers than AA (P < 0.05; Figure 2). On the contrary,
As a statistical approach, the MLR method was used for the
10 TAC to assess possible relationships between their HOCl
scavenging activity and the 6 electronic/transport descriptors in a
quantitative manner. Some considerations were decisive for the
inclusion or rejection of descriptors (predictor variables) in the
screened and simplified models. In particular, pairs of variables
when NAC was used as a scavenger of high efficiency (IC50
.99 ( 0.07 mM), only DATS and SAC were more effective in
scavenging HOCl (P<0.05; Figure 2).
=
0

Article
J. Agric. Food Chem., Vol. 58, No. 21, 2010 11229
Figure 1. Spectrophotometric patterns of concentration-response data in the assay system for HOCl using eight TAC from garlic and two onion derivatives:
(
(
A) allicin; (B) alliin; (C) allyl mercaptan; (D) allyl methyl sulfide; (E) diallyl disulfide; (F) diallyl sulfide; (G) dipropyl disulfide; (H) dipropyl sulfide; (I) diallyl trisulfide;
J) S-allylcysteine. Arrows numbered 1 indicate the catalase peak without HOCl, and arrows numbered 2 indicate the abolished catalase peak by 6 mM HOCl.
equation could be obtained. In an initial approach using all six
descriptors, the maximal fit observed was as follows:
log IC50 ¼ 4:6994 ð1:557Þ þ 0:0303 ð0:008ÞEHOMO
ð0:056Þ
ð0:039Þ
þ 0:0004 ð0:010ÞELUMO þ 0:0053ð0:003Þ ETOTAL
þ 0:308ð0:302Þ μ þ 0:0135ð0:011Þ PSA þ 0:0515ð0:135Þ log Poct ð1Þ
ð0:965Þ
ð0:224Þ
ð0:383Þ
ð0:331Þ
ð0:729Þ
2
2
R
¼ 0:905;R ðcvÞ ¼ 0:715; s ¼ 0:131; n ¼ 10; DW ¼ 1:746
In this and the following equation, the numbers in parentheses
indicate the standard errors and P values (below) of the respective
coefficients, n is the number of compounds tested, s is the
2
2
standard error, R is the squared correlation coefficient, R (cv)
is the squared cross-validation coefficient, and DW is the Durbin-
Watson statistic for possible correlations among residuals.
Despite the high statistical significance determined, this equation
was not suitable for interpretations because there were three pairs
of variables strongly intercorrelated (r values of ETOTAL/μ,
Figure 2. HOCl scavenging capacity (IC50 values, in mM) of garlic TAC
and reference compounds (AA, ascorbic acid; ALC, allicin; ALI, alliin; AM,
allyl mercaptan; AMS, allyl methyl sulfide; DADS, diallyl disulfide; DAS,
diallyl sulfide; DATS, diallyl trisulfide; DPDS, dipropyl disulfide; DPS,
dipropyl sulfide; NAC, N-acetylcysteine; SAC, S-allylcysteine). /,
IC lower than AA (-- -) (P < 0.05); //, IC50 lower than NAC ( - -)
5
0
3 3
-
0.768; ETOTAL/log Poct, 0.816; and μ/log Poct, -0.779). Instead,
(
P < 0.05).
the equation that proved useful for statistically valid interpreta-
tions was based on three descriptors as follows:
withanr g 0.75wereconsidered asbeing intercorrelated, and that
with the lower R statistics versus HOCl scavenging ability was
2
log IC50 ¼ 3:4367ð0:915Þ þ 0:0203ð0:004Þ EHOMO
discarded in the simplified models. Also, only regression models
without serious multicollinearity (i.e., absolute values of coeffi-
cient correlation <0.5) and a P<0.05 in ANOVA among all
predictor variables were considered as statistically valid. Thus,
the possible 60 combinations of predicted versus observed activ-
ities were screened. From these, only one simplified and validated
ð0:009Þ
ð0:005Þ
þ 0:0041ð0:001Þ ETOTAL þ 0:0115ð0:002Þ PSA
ð2Þ
ð0:009Þ
ð0:003Þ
2
2
R
¼ 0:824; R ðcvÞ ¼ 0:736; s ¼ 0:126; n ¼ 10; DW ¼ 2:176

11230 J. Agric. Food Chem., Vol. 58, No. 21, 2010
Arg €u ello-Garc ´ı a et al.
Table 2. Values of Molecular Descriptors for TAC Shown in Table 1
a
2
polar surface area (A )
˚
compd
log Poct
Etotal (kcal/mol)
EHOMO (kcal/mol)
E
LUMO (kcal/mol)
dipole (D)
AM
1.51
1.76
13.71
8.62
-205.72
-195.63
-195.48
-196.45
-206.77
-206.96
-209.33
-212.71
-191.66
-230.07
9.74
11.54
1.78
1.57
1.60
1.54
2.16
2.14
1.40
2.25
2.30
3.68
39.80
25.30
25.30
25.30
50.60
50.60
75.90
61.58
88.62
105.69
AMS
DPS
DAS
DPDS
DADS
DATS
ALC
2.88
2.75
-35.56
26.38
20.50
8.12
3.84
2.91
-29.49
31.57
30.46
-33.94
-36.87
-51.57
-33.70
12.31
3.01
1.87
-0.78
-87.10
-111.47
SAC
ALI
-2.15
-3.87
-1.40
a
ALC, allicin; ALI, alliin; AM, allyl mercaptan; AMS, allyl methyl sulfide; DADS, diallyl disulfide; DAS, diallyl sulfide; DATS, diallyl trisulfide; DPDS, dipropyl disulfide; DPS,
dipropyl sulfide; SAC, S-allylcysteine.
Figure 3. Relationships between the experimental and calculated HOCl scavenging activities of garlic TAC shown in Table 1 using MLR analyses
corresponding to eq 2 (a) and comparison of PLS regression coefficients assessing the contribution of each molecular descriptor listed in Table 2 for eq 2 (b).
This equation is highly significant, and there were no intercorrre-
lations among the variables (r values of EHOMO/ETOTAL, 0.283;
EHOMO/PSA, -0.229; ETOTAL/PSA, -0.644). Indeed, the inclu-
sion of two electronic descriptors (EHOMO and ETOTAL) and one
transport descriptor (PSA) yielded good-fitting and predictive capa-
bilities to the simplified model as assessed by a low s value and high
DISCUSSION
Scavenging of ROS is one of the mechanisms whereby garlic
derivatives may confer important health care benefits. HOCl is a
strong oxidant endogenously and excessively produced in inflam-
matory, degenerative, and neoplastic disorders; thus, effective
therapies and/or prophylaxes using exogenous scavengers of high
specificity are required. From our previous studies, it was observed
that aqueous extracts from raw garlic and unpeeled cloves treated
with distinct thermal processes were able to differentially scavenge
HOCl, and SAC was the only garlic compound previously studied
in this context (13, 18). In this work, eight derivatives of garlic
displaying systematic modifications of the basic thioallyl structure
2
2
R and R (cv) values that led to a very satisfactory agreement
between the observed and predicted IC50 values (Figure 3a).
To define the relative contribution of each descriptor included
in the simplified model shown in eq 2, this was cross-validated by
the PLS procedure using the leave-one-out method with unmodi-
fied data. The resulting model was statistically significant by
ANOVA test (P value = 0.011). By analyzing the relationship of
the selected descriptors and the corresponding PLS standardized
coefficients, it could be observed that in this three-component
model the topological descriptor PSA was the main contributor
to variations in scavenging capacity of TAC (PLS coefficient =
(TAC) and two propyl sulfide analogues from onion were tested
using the catalase protection assay to explore their ability to
scavenge HOCl. The thioallyl group was studied on the basis of
its natural abundance in garlic compounds and on its bona fide
characteristics as pharmacophore. This group has been identified
as the moiety having trophic and protective effects at organ and
cellular levels upon drug toxicity and carcinogenesis in experi-
mental models (24, 25). Synthetic, nongarlic TAC have been
further obtained and evaluated (26) providing a direct precedent
for rational drug design with this pharmacophore.
1.31, Figure 3b), whereas both electronic descriptors (ELUMO and
ETOTAL) had a lesser contribution to IC50 variations (PLS
coefficients = 0.93 and 0.85, respectively; Figure 3b). Therefore,
the electronic densities, primarily at the topological and secon-
darily at the molecular orbital levels, mostly explained the
variations in HOCl scavenging capacity of the compounds tested.

Article
As has been observed for other ROS, the HOCl scavenging
J. Agric. Food Chem., Vol. 58, No. 21, 2010 11231
molecules and the fourth reacts with the R-amino group (30),
which renders it a much more efficient scavenger. This is
consistent with the strong relationship of the acetylalanyl moiety
of NAC to the alanyl group of SAC and the thiol group of NAC
that was functionally improved by the increasing number of S
atoms present in DATS, two structural traits that combined to
HOCl dissipation by the HCl product could reinforce the high
scavenging efficacy of SAC and DATS. Otherwise, DPS and ALI
have structural traits (absence of CdC bond and presence of the
oxo group, respectively) that must hinder the reactivity of HOCl
on R;SH or R;S;S;R moieties of TAC. On the basis of
these observations, it is tempting to speculate that other TAC
exhibiting disulfide bonds and alanyl groups (e.g., S-allyl-
mercaptocysteine) might be predicted as very efficient HOCl
scavengers (even better than SAC), whereas others with disulfide
bonds and oxo groups (e.g., ajoenes) would have a more limited
ability to scavenge HOCl.
capacity was distinct among garlic TAC but relied on certain
structural features of the compounds. On the one hand, the HOCl
scavenging capacity of garlic extracts was not dependent on ALC
formation (13), and the present data support that although ALC
had a good scavenging activity, several other compounds con-
tained in fresh, heated/boiled garlic extracts (including ALI, a
scavenger of low capacity) are still able to trap HOCl. On the
other hand, the concentration-activity data for all TAC tested
showed that their HOCl trapping ability was enhanced by sub-
stituting the former H atom from the thiol group by an increasing
number of S atoms or by the alanyl group; however, scavenging
was decreased by the absence of the CdC bond or the presence of
sulfoxide groups in the thioallyl moiety. This latter observation
was supported by the higher HOCl trapping activity of DADS as
compared to ALC and that of SAC versus ALI.
These data seem to be consistent with several types of reaction
mechanisms between TAC and HOCl where the -S;CH ;
In further analyses, molecular descriptors of TAC were used as
quantitative resources to model their interaction with HOCl as
the biological target. In this context, it is well-known that further
to their antioxidant properties, TAC from garlic are also able to
modulate a broad spectrum of enzymes involved in a plethora of
biological responses (31-33). The few previous studies that
considered molecular descriptors of organosulfur compounds
2
CHdCH moiety is likely delivered as a modified product from
2
primary reactions. One such reaction might be the oxidation of
sulfhydryl groups to form disulfides (27) as with AM:
H;S;CH
2
;CHdCH
2
þ HOCl
þ H
-
f CH
2
dCH;CH ;S;S;CH
2
2
;CHdCH
2
2
O þ Cl ð3Þ
(
OSCs) from garlic, including some TAC, assessed structure-
Another likely reaction may occur with the di- or trisulfides to
produce sulfinic acids (27) as follows:
enzyme inhibition (20, 21) and neuroprotective activity-ROS
scavenging relationships (22). Although the enhancing effect of
the thioallyl moiety on the inhibition of carcinogen (benzo-
R;S
f R;S
n
;S;CH
2
;CHdCH
2
þ HOCl
[
a]pyrene) activation mediated by cytochrome P450 1 (20) and
O
2
H þ HO ;S;CH
2
2
;CHdCH
2
þ HCl ð4Þ
the effect of the alanyl group conferring higher neuroprotective
activity (22) were determined, data concerning quantitative
structure-activity relationships (QSAR) for OSCs were reported
only in a model of inhibition of soybean 15-lipoxygenase by a
series of 28 OSCs from garlic essential oil comprising mono-, di-,
and trisulfides (21). Of interest, from a large number of descrip-
tors evaluated by MLR and PLS analyses, only descriptors
related to molecular size (solvent-accessible surface area and
average distance/distance degree) and electron-acceptor nature
n
In this case, R corresponds to an allyl (as in DADS and DATS),
propyl (as is DPDS), or allyl sulfinate group (as in ALC). In this
reaction, the formation of HCl contributes to further HOCl
dissipation. Third, if R-amino acids (such as SAC and ALI) are
exposed to HOCl, a chlorination reaction likely occurs at the
R-amino group (28):
(
ELUMO) of OSCs contributed significantly to variations in their
inhibitory effect on lipoxygenase activity (32). Nevertheless,
modeling TAC-enzyme interactions should be more difficult
than for TAC-HOCl interactions as enzymes are much more
complex molecules with intricate shapes and sizes within and
neighboring catalytic sites when compared to the very reactive
and small HOX molecule. In the present work this latter ob-
servation would contribute to retrieve statistically valid models
even using a smaller set of 10 compounds tested. Prompted by this
rationale, the present study systematically included electronic/
quantum-chemical (ETOTAL, EHOMO, ELUMO, and dipole) and
transport/permeability (log Poct and PSA) indices in MLR and
PLS analyses to identify the properties of 10 TAC by determining
their HOCl scavenging capacity and the relative contributions of
each descriptor. To our knowledge, this is the first time that
quantitative structure-antioxidant activity relationships for
garlic compounds are determined.
Because this group is attached to an amino acid, this
chloramine product readily decomposes to the correspond-
ing amide, releasing ammonia and HCl that in turn help to
dissipate HOCl. Nevertheless, the mechanism for the split of
the (allyl)S-C bond in AMS, DPS, DAS, ALC, SAC, and
ALI by HOCl seems to be much more complex and likely
needs multiple electron transfers and reaction intermediates
to take place. Furthermore, there is a possibility that some
other reaction products might also dissipate HOCl as afore-
mentioned. One such example was shown with 2-propene-
sulfenic acid, a Cope elimination product of ALC that was
•
shown to be the responsible for HOO scavenging instead of
ALC itself (16). The HOCl trapping by TAC regarding this
specific issue deserves future research.
Other implications from the structure-activity data obtained
in this study were derived from the comparison of the HOCl
scavenging activity of tested TAC with that of two scavengers of
low and high efficiency that in humans are derived from exogen-
ous and endogenous sources (AA and NAC, respectively). The
different ability to trap HOCl by these compounds is explained by
the underlying mechanism: AA has two hydroxyl groups that are
oxidized by two HOCl to form dehydroascorbic acid (29),
whereas NAC is able to scavenge up to four HOCl molecules as
its sulfhydryl group may be oxidized three times by three HOCl
Among electronic properties, EHOMO and ETOTAL, in that
order, contributed to variations in IC50 values (eq 2). This is in
good agreement with the high reactivity and electron-donor
nature of a TAC antioxidant that has to be oxidized upon
interaction with HOCl (eqs 3-5). In other terms, the less negative
(
i.e., less stable) is the electronic charge level on the more external
electron orbitals, the higher is the HOCl scavenging ability of
TAC. Furthermore, the improved model fitting TAC-HOCl
interactions (eq 2) included the topological PSA descriptor as a
primary contributor. First, it is well-known that this index, at

11232 J. Agric. Food Chem., Vol. 58, No. 21, 2010
Arg €u ello-Garc ´ı a et al.
2
˚
values of 50-90 A , predicts good cellular permeability, and it
was the case for some TAC exhibiting high HOCl scavenging
activity (e.g., ALC, DADS, DATS, and SAC). This would imply
an increased bioavailability of these compounds (by crossing the
blood-brain barrier) and helps to explain their neuroprotective
effect that in turn is related to ROS scavenging (22). Second, the
DADS, diallyl disulfide; DAS, diallyl sulfide; DATS, diallyl
trisulfide; DPDS, dipropyl disulfide; DPS, dipropyl sulfide;
NAC, N-acetylcysteine; SAC, S-allylcysteine; IC50, half-maximal
inhibitory concentration; MLR, multilinear regression analysis;
PLS, partial least-squares analysis.
ACKNOWLEDGMENT
importance of the broader size of the polar area on the R;NH
26.02 A ), R;SH (38.8 A ), R;S;S;R (50.6 A ), and R;
S;S;S;R (75.9 A ) moieties that interact with the HOCl
2
2
2
2
˚
˚
˚
(
We are grateful to Arturo Perez Taylor-Reyes for his help with
´
part of the artwork.
2
˚
oxidant (eqs 3-5) as compared to other groups with lower or
˚
LITERATURE CITED
negligible reactivity with HOCl (e.g., RdO, 17.07 A, and R;OH,
2
˚
2
0.23 A , both present in AA). Despite the fact that ALI had the
(1) Halliwell, B.; Gutteridge, J. M. C. Free Radicals in Biology and
Medicine, 4th ed.; Clarendon Press: Oxford, U.K., 2007; p 88.
2
˚
highest PSA value (105.69 A ), the presence of the oxo group
surpasses the influence of PSA by hindering the reactivity of the
adjacent S atom that likewise displays an apparently favorable
(2) Sies, H., Ed. Oxidative Stress: Oxidants and Antioxidants; Academic
Press: London, U.K., 1991.
2
(3) Davies, M. J.; Hawkins, C. L.; Pattison, D. I.; Rees, M. D.
Mammalian heme peroxidases: from molecular mechanisms to
health implications. Antiox. Redox Signal 2008, 10, 1199-1234.
˚
PSA (32.09 A ). A likely explanation for this effect is that the
electron density of the S atom is enhanced by the adjacent allyl
group but drastically decreased by the O atom, which is less
polarizable, indeed impairing the nucleophilic nature of the
adjacent S atom.
(
4) Prutz, M. A. Hypochlorous acid interactions with thiols, nucleo-
tides, DNA, and other biological substrates. Arch. Biochem. Biophys.
1996, 332, 110-120.
The rational design of HOCl scavengers with higher efficacy
still deserves further analyses for several reasons. One relates to
the fact that some compounds structurally related to TAC, as are
the 4-mercaptoimidazoles (32), or structurally distinct, as are the
nonsteroid anti-inflammatory pyrazolones (33) or the β-lactam
antibiotics cephalosporins (34), are able to scavenge HOCl at the
micromolar range in vitro. However, these compounds are
actually intended for other therapeutic purposes and may have
severe secondary effects. Another relates to the fact that TAC
from garlic are able to scavenge other ROS, and particularly
SAC, the most efficient HOCl scavenger probed in the present
(5) Van der Vliet, A.; Nguyen, M. N.; Shigenaga, M. K.; Eiserich, J. P.;
Marelich, G. P.; Cross, C. E. Myeloperoxidase and protein oxidation
in cystic fibrosis. Am. J. Physiol. Lung Cell Biol. Physiol. 2000, 279,
L537-L546.
(
6) Hazell, L. J.; Arnold, L.; Flowers, D.; Waeg, G.; Malle, E.; Stocker,
R. Presence of hypochlorite-modified proteins in human athero-
sclerotic lesions. J. Clin. Invest. 1996, 97, 1535-1544.
7) Weitzmann, S. A.; Gordon, L. I. Inflammation and cancer: role of
phagocyte-generated oxidants in carcinogenesis. Blood 1990, 76,
655-663.
(
(8) Iciek, M.; Kwiecie nꢀ , I.; Wlodek, L. Biological properties of garlic
and garlic-derived organosulfur compounds. Environ. Mol. Mutagen.
2009, 50, 247-265.
•
-
•
•
-
work, is also an efficient O2 , HO , HOO , and ONOO
(9) Borek, K. Antioxidant health effects of aged garlic extract. J. Nutr.
scavenger at the submillimolar range in vitro (14, 17); moreover,
SAC itself has been reported as either an inefficient (20) or highly
active (32, 35) enzyme modulator. However, SAC exhibits other
important several advantages including easy gastrointestinal
absorption and bioavailability, low toxicity, and cost-effective
synthesis (22). Ultimately, garlic derivatives have important
advantages over most synthetic compounds as they have proven
benefits in chemoprevention besides their multiple therapeutic
applications (8, 9, 24).
2
001, 131, 1010S-1015S.
(
10) Lawson, L. D. The composition and chemistry of garlic cloves and
processed garlic. In Garlic: the Science and Therapeutic Applications
of Allium sativum L. and Related Species, 1st ed.; Koch, H. P., Lawson,
D., Eds.; Williams & Wilkins: Baltimore, MD, 1996; p 57.
ꢀ
(11) Pedraza-Chaverrı
Zu
scavenging capacity of different cooked garlic preparations. Life Sci.
006, 78, 761-770.
12) Pedraza-Chaverrı, J.; Medina-Campos, O. N.; Segoviano-Murillo,
´
, J.; Medina-Campos, O. N.; Avila-Lombardo, R.;
˜
´ niga-Bustos, A. B.; Orozco-Ibarra, M. Reactive oxygen species
2
(
´
Taken together, the observations derived from this work may
have important implications in modeling HOCl-mediated oxida-
tive stress because catalase is an antioxidant, protective enzyme
that may be located at inflammation sites and is able to be
inactivated by HOCl at the physiologically feasible concentra-
tions (29) mimicked in this study. Also, the presented data might
provide a platform for future design of HOCl scavengers with
improved action in processes associated with overproduction
and/or misplacing of this oxidant. In addition, more studies
concerning the structure-trapping relationship with other ROS
by the TAC tested herein will help to unravel the differential
requirements and mechanisms underlyingthe multiplehealth care
benefits associated with consumption of garlic nutraceuticals.
S. Effect of heating on peroxynitrite scavenging capacity of garlic.
Food Chem. Toxicol. 2007, 45, 622-627.
(
13) Pedraza-Chaverrı, J.; Arriaga-Noblecıa, G.; Medina-Campos, O. N.
´ ´
Hypochlorous acid scavenging capacity of garlic. Phytother. Res.
2007, 21, 884-888.
(14) Chung, L. Y. The antioxidant properties of garlic compounds: allyl
cysteine, alliin, allicin, and diallyl disulfide. J. Med. Food 2006, 9,
205-213.
(
15) Prasad, K.; Laxdal, V. A.; Yu, M.; Raney, B. L. Antioxidant activity of
allicin, an active principle in garlic. Mol. Cell. Biochem. 1995, 19, 183-189.
16) Galano, A.; Francisco-Marquez, M. Peroxyl-radical scavenging
activity of garlic: 2-propenesulfonic acid versus allicin. J. Phys.
Chem. B 2009, 113, 16077-16081.
(
(
17) Maldonado, P. D.; Barrera, D.; Rivero, I.; Mata, R.; Medina-
Campos, O. N.; Hernandez-Pando, R.; Pedraza-Chaverrı, J. Anti-
´ ´
ABBREVIATIONS USED
oxidant S-allyl cysteine prevents gentamicin-induced oxidative stress
and renal damage. Free Radical Biol. Med. 2003, 35, 317-324.
18) Medina-Campos, O. N.; Barrera, D.; Segoviano-Murillo, S.; Rocha,
TAC, thioallyl compounds; PSA, topological polar surface
area; ETOTAL, total energy; EHOMO, energy of the higher occupied
molecular orbital; ELUMO, energy of the lower unoccupied
molecular orbital; μ, dipole moment; log Poct, logarithm of the
octanol/water partition coefficient; ROS, reactive oxygen species;
(
D.; Maldonado, P. D.; Mendoza-Patin
J. S-allyl cysteine scavenges singlet oxygen and hypochlorous acid
and protects LLC-PK cells of potassium dichromate-induced cyto-
toxicity. Food Chem. Toxicol. 2007, 45, 2030-2039.
˜
o, N.; Pedraza-Chaverrı,
´
1
•-
•
•
O2 , superoxide radical; HO , hydroxyl radical; HOO , peroxyl
(19) Aruoma, O. I.; Halliwell, B. Action of hypochlorous acid on the
antioxidant protective enzymes superoxide dismutase, catalase and
glutathione peroxidase. Biochem. J. 1987, 248, 973-976.
-
radical; ONOO , peroxynitrite; AA, ascorbic acid; ALC, allicin;
ALI, alliin; AM, allyl mercaptan; AMS, allyl methyl sulfide;

Article
J. Agric. Food Chem., Vol. 58, No. 21, 2010 11233
(
(
(
(
(
(
20) Jung Kim, H.; Sook Chun, H. Structure-activity relationships
(28) Block, S. S., Ed. Disinfection, Sterilization and Preservation; Lea &
Febiger, Philadelphia, PA, 1991.
(29) Carr, A. C.; Frei, B. Human neutrophils oxidize low-density lipo-
protein by a hypochlorous acid-dependent mechanism: the role of
vitamin C. Biol. Chem. 2002, 383, 627-636.
(30) Winterbourne, C. C. Comparative reactivities of various biological
compounds with myeloperoxidase-hydrogen peroxide-chloride, and
similarity to the oxidant to hypochlorite. Biochim. Biophys. Acta
1985, 840, 204-210.
(31) Li, C. K.; Tsai, C. W.; Wu, C. C. Garlic allyl sulfides display
differential modulation of rat P450 2B1 and the placental form
glutathione S-transferase in various organs. J. Agric. Food Chem.
2006, 54, 5191-5196.
of organosulfur compounds as inhibitors of cytochrome P450
1
-mediated activation of benzo[a]pyrene in a human cell model.
Biotechnol. Lett. 2002, 24, 2041-2046.
21) Camargo, A. B.; Marchevsly, E.; Luco, J. M. QSAR study for the
soybean 15-lipoxygenase inhibitory activity of organosulfur com-
pounds derived from the essential oil of garlic. J. Agric. Food Chem.
2
007, 55, 3096-3103.
22) Kim, J. M.; Joo Chang, H.; Kim, W. K.; Chang, N.; Sook Chun, H.
Structure-activity relationship of neuroprotective and reactive
oxygen species scavenging activities for Allium organosulfur com-
pounds. J. Agric. Food Chem. 2006, 54, 6547-6553.
23) Ertl, P.; Rohde, B.; Selzer, P. Fast calculation of molecular polar
surface area as a sum of fragment-based contributions and its
application to the prediction of drug transport. J. Med. Chem.
(
32) Gupta, N.; Porter, T. D. Garlic and garlic-derived compounds inhibit
human squalene monooxigenase. J. Nutr. 2001, 131, 1662-1667.
33) Costa, D.; Marques, A. P.; Reis, R. L.; Lima, J. L.; Fernandes, E.
Inhibition of human neutrophils oxidative burst by pyrazolone
derivatives. Free Radical Biol. Med. 2006, 40, 632-640.
2
000, 43, 3714-3717.
(
24) Wang, E. J.; Li, Y.; Lin, M.; Chen, L.; Stein, A. P.; Reuhl, K. R.;
Yang, C. S. Protective effects of garlic and related organosulfur
compounds on acetaminophen-induced hepatotoxicity in mice.
Toxicol. Appl. Pharmacol. 1996, 136, 146-154.
25) Moriguchi, T.; Matsuura, H.; Kodera, Y.; Itakura, Y.; Katsuki, H.;
Saito, H.; Nishiyama, N. Neurotrophic activity of organosulfur
compounds having a thioallyl group on cultured rat hippocampal
neurons. Neurochem. Res. 1997, 22, 1449-1452.
26) Kwon, S. K.; Moon, A. Synthesis of 3-alkylthio-6-allylthiopyrida-
zine derivatives and their antihepatocarcinoma activity. Arch. Pharm.
Res. 2005, 28, 391-394.
27) Pereira, W. E.; Hoyano, Y.; Summons, R. E.; Bacon, R. E.; Duffield,
A. M. Chlorination studies: II. The reaction of aqueous hypo-
chlorous acid with R-aminoacids and dipeptides. Biochim. Biophys.
Acta 1973, 313, 170-180.
(
34) Lapenna, D.; Cellini, L.; De Gioia, S.; Mezzeti, A.; Ciofni, G.; Festi,
D.; Cuccurullo, F. Cephalosporins are scavengers of hypochlorous
acid. Biochem. Pharmacol. 1995, 49, 1249-1254.
35) Pinto, J. T.; Krasnikov, B. F.; Cooper, A. Redox-sensitive proteins
are potential targets of garlic derived mercaptocysteine derivatives.
J. Nutr. 2006, 136, 835S-841S.
(
(
Received for review June 22, 2010. Revised manuscript received
September 10, 2010. Accepted September 11, 2010. This work was
supported in part by funds from SEP-CONACyT-ANUIES-ECOS
Grant M06-S03, DGAPA Grant 201910, and CONACYT Grants
49724-M and 129838.
(