Synthesis and Antioxidant Activity of Rosmariquinone and Several Analogues

Article abstract of DOI:10.1021/jf970742k

Rosmariquinone (1) and six analogues were chemically synthesized using an ultrasound-promoted Diels - Alder cycloaddition in yields of 35-90%. The analogues included substitution of the isopropyl at carbon 13 (C-13) with a hydrogen (5), methyl (6), or tert-butyl (4) substituent. The hydrogen-substituted analogue had the lowest yield at 35%, due in part to the instability of the compound to air, while the highest yields were achieved for the methyl (85%) and tert-butyl (90%) analogues. The 60% yield obtained for the C-14 methyl analogue (7; no C-13 isopropyl) may have been caused by the meta-substituted catechol inhibiting the cycloaddition. The final two analogues were ring A modifications and included the removal of one C-4 methyl (3; 80% yield) or both C-4 methyl (2; 85% yield) groups. The analogues were tested against rosmariquinone in light-sensitized oxidation of stripped soybean oil. Analogues 5 and 6 were significantly (P < 0.05) better antioxidants than rosmariquinone and all other analogues. The antioxidant properties of compounds 2-7 were not significantly different (P < 0.05) from each other while compounds 2 and 4 had significantly (P < 0.05) lower antioxidant activity than rosmariquinone. This study demonstrated the importance of structural characteristics of antioxidants and that natural antioxidants, such as rosmariquinone, can be improved through chemical modification.

Full text of DOI:10.1021/jf970742k

J. Agric. Food Chem. 1998, 46, 1303−1310
1303
Syn th esis a n d An tioxid a n t Activity of Rosm a r iqu in on e a n d Sever a l
An a logu es
Clifford A. Hall III,*,† Susan L. Cuppett,‡ and Pat Dussault§
Department of Cereal Science, Harris Hall, North Dakota State University, Fargo, North Dakota 58105,
and Departments of Food Science and Technology and Chemistry, University of Nebraska-Lincoln,
Lincoln, Nebraska 68583
Rosmariquinone (1) and six analogues were chemically synthesized using an ultrasound-promoted
Diels-Alder cycloaddition in yields of 35-90%. The analogues included substitution of the isopropyl
at carbon 13 (C-13) with a hydrogen (5), methyl (6), or tert-butyl (4) substituent. The hydrogen-
substituted analogue had the lowest yield at 35%, due in part to the instability of the compound to
air, while the highest yields were achieved for the methyl (85%) and tert-butyl (90%) analogues.
The 60% yield obtained for the C-14 methyl analogue (7; no C-13 isopropyl) may have been caused
by the meta-substituted catechol inhibiting the cycloaddition. The final two analogues were ring A
modifications and included the removal of one C-4 methyl (3; 80% yield) or both C-4 methyl (2; 85%
yield) groups. The analogues were tested against rosmariquinone in light-sensitized oxidation of
stripped soybean oil. Analogues 5 and 6 were significantly (P < 0.05) better antioxidants than
rosmariquinone and all other analogues. The antioxidant properties of compounds 2-7 were not
significantly different (P < 0.05) from each other while compounds 2 and 4 had significantly (P <
0.05) lower antioxidant activity than rosmariquinone. This study demonstrated the importance of
structural characteristics of antioxidants and that natural antioxidants, such as rosmariquinone,
can be improved through chemical modification.
Keyw or d s: Antioxidant activity; Diels-Alder cycloaddition; rosmariquinone; ultrasound
INTRODUCTION
been given a great deal of attention because of their
strong antioxidant activity, but only carnosol, carnosic
acid, and rosmariquinone (RQ) have been produced
synthetically.
Many natural product syntheses have been directed
toward the production of medicinal compounds, and
little attention has been given to the production of
compounds for fields other than medicine. The produc-
tion of natural food antioxidants through synthetic
methods increases yields and reduces their costs of
supply to the food industry. For example, the synthetic
antioxidant butylated hydroxytoluene (BHT) costs $1.80/
lb while the natural counterpart, tocopherols, cost
$14.64/lb (Haumann, 1991), thus demonstrating the
substantial cost savings to the users of synthetic anti-
oxidants. A second advantage of synthetic organic
chemistry is in the production of analogues of a natu-
rally occurring compound. Miller and Quackenbush
(1957a,b) and Dugan et al. (1951) identified structural
characteristics of phenolic antioxidants and found that
simple variation in side groups (e.g., tert-butyl versus
methyl) improved the activity of an antioxidant. The
same idea can be applied to a naturally occurring
compound where the substitution of functional groups
at different regions of the molecule may bring about
increased activity of the compound. This type of im-
provement would be difficult if not impossible to control
in cell cultures and/or in the plant system. Antioxidants
isolated from rosemary (Rosmarinus officinalis L.) have
Several synthetic methods were developed for the
total synthesis of carnosol, carnosic acid, and their
dimethyl ethers (Meyer and Schindler, 1966; Meyer et
al., 1968; Shew and Meyer, 1968). The synthetic routes
to carnosol and carnosic acid have been given little
attention over the last 25-30 years due to low synthetic
yields and the large quantities (2.7-26%) of these
antioxidants that can be found in rosemary and sage
extracts (Cuvelier et al., 1994). Rosmariquinone (Hou-
lihan et al., 1985) or miltirone (Hayashi et al., 1970)
isolated from rosemary and Chinese sage, respectively,
was found to be a minor component of the plant
materials. Hall et al. (1994) isolated between 16 and
230 ppm (0.0016-0.023%) RQ from a commercial rose-
mary oleoresin which represents a concentration that
is 100 times lower than the amount of carnosol or
carnosic acid in the commercial rosemary oleoresin. For
this reason, little attention has been given to the
separation of RQ from the commercial rosemary oleo-
resin or as an antioxidant. Unlike carnosol or carnosic
acid, RQ can be synthetically produced in yields as high
as 60% (total synthesis yield) and large quantities can
be easily produced. Several synthetic methods have
been developed over the last 20 years following various
synthetic routes.
* Author to whom correspondence should be addressed.
† North Dakota State University.
There are three general approaches for the synthesis
of RQ. The first synthetic method (Nasipuri and Mitra,
1973; Chang et al., 1990, 1991; He et al., 1990) follows
the approach in which ring C was constructed first
‡ Department of Food Science and Technology, University
of Nebraska-Lincoln.
§
Department of Chemistry, University of Nebraska-
Lincoln.
S0021-8561(97)00742-5 CCC: $15.00 © 1998 American Chemical Society
Published on Web 03/25/1998

1304 J. Agric. Food Chem., Vol. 46, No. 4, 1998
Hall III et al.
F igu r e 1. Reaction sequence of the ultrasound-promoted Diels-Alder cycloaddition as postulated by Lee et al. (1990) for the
synthesis of miltirone (i.e., rosmariquinone; 1).
as parts per million (δ) and referenced to the internal residual
followed by ring B and finally ring A. The resulting
overall yields of 3-3.5% limits the cost-effectiveness of
this synthetic approach. The second method follows a
semisynthetic route using carnosol as the starting
material (Luis et al., 1996). This method resulted in a
yield of 42% and, provided carnosol is readily available,
represents a more efficient method to obtain RQ com-
pared to the earlier synthetic methods. The third and
most productive method involves the Diels-Alder cy-
cloaddition reaction between 3-isopropyl-O-benzoquino-
ne and 6,6-dimethyl-1-vinylcyclohexene (Knapp and
Sharma, 1985; Lee et al., 1990). The Diels-Alder
cycloaddition method of Lee et al. (1990) resulted in
significant improvement in the overall yield of RQ (i.e.,
miltirone) when compared to the method of Knapp and
Sharma (1985). Lee et al. (1990) reported a cycloaddi-
tion yield of 93% (60% overall yield), while Knapp and
Sharma (1985) were only able to achieve a 30% (20%
overall yield). The significant improvement in the
Diels-Alder approach between Lee et al. (1990) and
Knapp and Sharma (1985) was the elimination of an
external oxidation step prior to the cycloaddition step.
A second advantage to the ultrasound-promoted cy-
cloaddition was that less catechol (i.e., o-benzoquinone),
the more expensive reactant, was needed to produce
higher yields. A variety of abietanoid o-quinones have
been produced using ultrasound-promoted cycloaddi-
tions (Lee and Snyder, 1990). Because these authors
have demonstrated the application of ultrasound for the
production of a wide variety of o-quinones, the use of
ultrasound represents a potential synthetic method for
the production of RQ analogues.
The antioxidant activity of RQ in various lipid sys-
tems has been demonstrated (Houlihan et al., 1985;
Weng and Gordon, 1992; Hall et al. 1994) while ana-
logues of RQ (excluding tanshinones) have not been
tested. Due to the limited research on the structural
characterization of RQ, the objectives of this study were
to synthesize several RQ analogues, including the
appropriate starting materials, and to compare the
antioxidant activity of the RQ analogues to that of RQ.
The general approach will be through the use of ultra-
sound-promoted Diels-Alder reaction. To determine
the importance of the isopropyl group (ortho position),
substitution will be made using tert-butyl, methyl, or
hydrogen in place of isopropyl (Figure 1, Table 1). To
determine the importance an ortho substituent, an
analogue bearing a meta methyl in place of the ortho
isopropyl will be synthesized. Last, the significance of
ring A dimethyl substitution will be investigated with
analogues bearing no methyl or single methyl on carbon
4 (Table 1).
1
peak of deuteriochloroform (7.25 ppm) for H NMR and to the
center peak of the triplet (77.0 ppm) for 13C NMR. 1H shifts
and couplings were assigned using 1H-1H chemical shift
correlated spectroscopy (i.e., COSY) while 13C-1H hetero-
nuclear chemical shift correlated spectroscopy (i.e., HETCOR)
was used to define the 13C resonances. Infrared (IR) absorp-
tion spectra were recorded on an Analect model RFX-30 FT-
IR spectrophotometer using a potassium bromide cell. Ultra-
violet data were collected using a Milton Roy Genesys 5 UV-
vis spectrophotometer. Low-resolution mass spectra were
obtained using a Hewlett-Packard 5790 GC-MS system and a
Chiral-val column (Alltech, State College, PA) using both
chemical (methane) and electron impact ionization. Uncor-
rected melting point determinations were completed using
open Pyrex capillary tubes and a Mel-Temp apparatus. All
flash column chromatography was completed on E. Merck
silica gel 60 and thin-layer chromatography on commercial
silica gel plates (Analtech Silica HLF 250 m; Fisher Scientific,
St. Louis, MO). Solvents for chromatography and for reactions
were purchased from Fisher, and other chemical reactants
were purchased from Aldrich Chemical Co. (Milwaukee, WI).
When required, the solvents were dried over an appropriate
drying agent under nitrogen atmosphere and used immedi-
ately. Reactions involving moisture-sensitive reagents were
performed under dry nitrogen in flame-dried flasks. Evapora-
tion of solvents was performed first on a Bu¨chi rotary evapora-
tor and then at 0.10 mmHg.
Syn th etic P r oced u r es. The production of the vinylcyclo-
hexene compounds (8-10; Table 1) was completed using the
method of Tanis and Abdallah (1986). The 1-vinylcyclohexene
(8) and 6-methyl-1-vinylcyclohex-1,2-ene (9) were synthesized
from the available cyclohexanone and 1-methylcyclohexanone,
respectively. However, the reactant 6,6-dimethyl-1-vinylcy-
clohexene (10) synthesis required that 2,2-dimethylcyclohex-
anone be synthesized using the method of Boatman et al.
(1965) prior to Grignard addition and subsequent dehydration.
3-Isopropyl catechol (11) was synthesized following the
methods of Edwards and Cashaw (1955) and Tsuruta and
Mukai (1968) while 3-tert-butyl catechol (12) was produced
using the method of Casiraghi et al. (1980). All other catechols
(13-15; Table 1) were purchased from Aldrich.
The Diels-Alder cycloaddition was completed using the
ultrasound method of Lee et al. (1990) using the appropriate
catechol and vinylcyclohexene (Table 1). A general synthetic
method for RQ is listed below, and all other analogues were
prepared and purified in the same manner following the
conditions in Table 1. Physical data are listed in Table 2.
1,2,3,4-Tetr a h yd r o-4,4-d im eth yl-13-(1-m eth yleth yl)-11,-
12-p h en a n t h r en ed ion e (R Q, 1; F igu r e 1). A mixture of
3-isopropylcatechol (11; 105 mg, 0.69 mmol), 6,6-dimethyl-1-
vinylcyclohexene (10, 95.2 mg, 0.69 mmol), and silver oxide
(788 mg, 3.40 mmol) in ethanol (1.5 mL) was subjected to
ultrasonication for 3 h at 24 °C. The mixture was filtered,
and the solvent was removed in vacuo. The o-quinone was
separated from the crude material using flash chromatography
on silica gel (hexane, 100%; hexane:ethyl acetate 90:10 v/v) to
give a mixture of regioisomers. The final purification was done
using flash chromatography on silica gel eluting with hexane:
ethyl acetate (100:0; 98:2; 95:5; 90:10). The solvent was
removed in vacuo to afford a red oil. The oil was recrystallized
from hexane to give RQ as red crystals (178 mg, 0.63 mmol,
90%): mp 100-101 °C; IR (KBr) 2959, 2932, 2870, 1660, 1630,
EXPERIMENTAL PROCEDURES
Gen er a l Meth od s. 1H NMR and 13C NMR spectra were
recorded on a 300 MHz General Electric NMR instrument (300
1
MHz for H, 75 MHz for 13C). Chemical shift data are reported

Antioxidant Activity of Rosmariquinone
J. Agric. Food Chem., Vol. 46, No. 4, 1998 1305
Ta ble 1. Diels-Ald er Cycloa d d ition Con d ition s for th e P r od u ction of RQ Der iva tives
a
b
Catechol is converted to o-quinone prior to cycloaddition. All reactions run under ultrasound and anhydrous ethanol. c Catechol
purchased from Aldrich.
1581, 1563, 1460, 1387, 1294, 1258, 1227, 1141, 934 cm-1; 1H
NMR δ 7.56 (d, J ) 7.97 Hz, 1H), 7.09 (d, J ) 7.97 Hz, 1 H),
7.06 (s, 1 H), 3.14 (t, J ) 6.38 Hz, 2 H), 2.99 (sept, J ) 6.82,
1 H), 1.76 (m, 2 H), 1.62 (m, 2 H), 1.28 (s, 6 H), 1.14 (d, J )
6.89, 6 H); 13C NMR δ 182.3, 181.4, 149.6, 145.0, 144.8, 144.3,
139.8, 134.2, 133.7, 128.0, 127.8, 37.6, 34.3, 31.5 (2 C), 29.8,
26.7, 21.3 (2 C), 18.9; LRMS (EI, 70 eV) m/z 254 (M+ - 28);
(CI, methane) m/z 283 (M + 1); UV-vis (methanol) λmax (ꢀ)
209 (14 222), 257 (17 730), 359 (1803), 437 (2184).
1,2,3,4-Tetr a h yd r o-4-m eth yl-13-(1-m eth yleth yl)-11,12-
p h en a n th r en ed ion e (3; Ta ble 1). Isolated as red crystals
(351 mg, 1.31 mmol, 80%): mp 93-94 °C; IR (KBr) 2960, 2934,
2872, 1661, 1574, 1464, 1388, 1254, 1110, 1110, 940 cm-1; 1H
NMR δ 7.41 (d, J ) 7.5 Hz, 1H), 7.07 (d, J ) 7.5 Hz, 1 H), 7.06
(s, 1 H), 3.21 (t, J ) 5.7 Hz, 1 H), 3.01-3.10 (m, 2 H), 2.95
(sept, J ) 5.7 Hz, 1 H), 1.79-1.84 (m, 2 H), 1.73 (m, 1 H), 1.53
(m, 1 H), 1.27 (d, J ) 7.5 Hz, 3 H), 1.15 (d, J ) 7.5 Hz, 6 H);
13C NMR δ 181.9, 181.0, 145.9, 144.9, 144.8, 139.9, 135.2,
134.5, 128.0, 127.7, 32.9, 29.5, 28.8, 26.6, 22.6, 21.2 (2 C), 19.2;
LRMS (EI, 70 eV) m/z 267 (M-1); (CI, methane) m/z 269 (M +
1); UV-vis (methanol) λmax (ꢀ) 209 (13 475), 260 (13 741), 359
(1492), 437 (2134).
1,2,3,4-Tet r a h yd r o-13-(1-m et h ylet h yl)-11,12-p h en a n -
th r en ed ion e or Nor m iltir on e (2; Ta ble 1). Isolated as red
crystals (300 mg, 1.181 mmol, 85%): mp 101-102 °C; IR (KBr)
2931, 2869, 1652, 1563, 1458, 1418, 1382, 1255, 917 cm-1; 1H
NMR δ 7.30 (d, J ) 7.5 Hz, 1H), 7.09 (s, 1 H), 7.06 (d, J ) 7.5
Hz, 1 H), 3.20 (t, J ) 6.4 Hz, 2 H), 3.04 (sept, J ) 6.8 Hz, 1
H), 2.80 (t, J ) 6.4 Hz, 2 H), 1.80 (m, 4 H), 1.18 (d, J ) 6.9 Hz,
6 H); 13C NMR δ 182.2, 181.3, 145.0, 144.7, 140.7, 139.9, 135.9,
134.6, 128.1, 127.5, 30.5, 28.7, 26.9, 22.8, 21.9, 21.4 (2 C);
LRMS (EI, 70 eV) m/z 253 (M - 1); (CI, methane) m/z 255 (M
+ 1); UV-vis (methanol) λmax (ꢀ) 209 (10 630), 257 (13 487),
359 (1194), 437 (1695).
1,2,3,4-Tetr ah ydr o-4,4-dim eth yl-13-(1,1-dim eth yleth yl)-
11,12-p h en a n th r en ed ion e (4; Ta ble 1). Isolated as red-
orange crystals (322 mg, 1.09 mmol, 90%): mp 124-126 °C;
IR (KBr) 2958, 2869, 1661, 1582, 1562, 1460, 1366, 1221, 1142,
1
911 cm-1; H NMR δ 7.58 (d, J ) 7.5 Hz, 1H), 7.14 (s, 1 H),
7.10 (d, J ) 7.5 Hz, 1 H), 3.15 (t, J ) 6.6 Hz, 2 H), 1.76 (m, 2
H), 1.63 (m, 2 H), 1.28 (s, 15 H); 13C NMR δ 182.7, 181.8, 149.4,
146.5, 144.0, 140.4, 134.4, 133.6, 128.2, 37.8, 34.6, 34.3, 31.6

1306 J. Agric. Food Chem., Vol. 46, No. 4, 1998
Hall III et al.

Antioxidant Activity of Rosmariquinone
J. Agric. Food Chem., Vol. 46, No. 4, 1998 1307
(2 C), 29.6, 29.0 (3 C), 22.5, 18.9; LRMS (EI, 70 eV) m/z 294
(M - 2); (CI, methane) m/z 297 (M + 1); UV-vis (methanol)
λmax (ꢀ) 218 (11 058), 251 (11 646), 356 (1616), 428 (2346).
1,2,3,4-Tetr a h yd r o-4,4-d im eth yl-11,12-p h en a n th r en e-
d ion e (5; Ta ble 1). Isolated as red-brown crystals (102 mg,
0.42 mmol, 35%): mp 90-92 °C; IR (KBr) 2957, 2933, 2865,
dants. TBHQ served as the positive control while an untreated
sample served as the negative control.
The jars were randomly placed under two, 15 W cool
fluorescent lamps at a level sufficient to illuminate 4200 lux
of fluorescent radiation at 25 ( 1 °C. Aluminum foil was
placed in the open areas between the side of the jars and the
bottom of the lamps to create uniform lighting. Duplicate
peroxide values (PV) were determined every 24 h during the
light exposure using the American Oil Chemists Society
(AOCS; 1989 official method Cd-8-53) until a PV of 20 meq/kg
was reached for each oil sample. When necessary, PV were
taken after 12 h rather than 24 h.
Statistical Analysis. The entire oxidation evaluation was
completed three times and the data were analyzed by analysis
of variance (ANOVA) using Statistical Analysis System (SAS)
software (1985). The least significant differences (Steel and
Torrie, 1980) were used to determine a 95% confidence level
(P < 0.05) between the mean number of hours required to
reach a PV of 20 meq/kg for the treatments. The synthesis
was completed three times, and the analytical data were
reported as an average.
1
1661, 1558, 1457, 1246, 1138, 852 cm-1; H NMR δ 7.61 (d, J
) 8.7 Hz, 1H), 7.35 (d, J ) 8.7 Hz, 1 H), 7.15 (d, J ) 8.7 Hz,
1 H), 6.35 (d, J ) 8.7 Hz, 1 H), 3.17 (t, J ) 6.0 Hz, 2 H), 1.78
(m, 2 H), 1.63 (m, 2 H), 1.28 (s, 6 H); 13C NMR δ 181.5, 181.4,
151.0, 146.9, 145.0, 133.9, 133.7, 129.3, 128.4, 126.4, 37.6, 31.6
(2 C), 31.3, 29.9, 18.9; LRMS (EI, 70 eV) m/z 240 (M+); (CI,
methane) m/z 241 (M + 1); UV-vis (methanol) λmax (ꢀ) 209
(10 703),251 (13 172), 356 (1201), 419 (1717).
1,2,3,4-T e t r a h y d r o -4,4-d im e t h y l-13-m e t h y l-11,12-
p h en a n th r en ed ion e (6; Ta ble 1). Isolated as red-orange
crystals (175 mg, 0.69 mmol, 85%): mp 106-107 °C; IR (KBr)
2961, 2929, 2861, 1692, 1570, 1462, 1264, 1143 cm-1; 1H NMR
δ 7.57 (d, J ) 7.5 Hz, 1H), 7.12 (s, 1 H), 7.05 (d, J ) 7.5 Hz,
1 H), 3.16 (t, J ) 6.6 Hz, 2 H), 2.0 (s, 3 H), 1.76-1.80 (m, 2 H),
1.61-1.65 (m, 2 H), 1.28 (s, 6 H); 13C NMR δ 181.9, 181.6,
149.5, 144.3, 142.9, 134.8, 134.2, 133.7, 128.2, 127.5, 37.5, 34.3,
31.5 (2 C), 29.7, 18.8, 14.9; LRMS (EI, 70 eV) m/z 253 (M+
1); (CI, methane) m/z 255 (M + 1); UV-vis (methanol) λmax (ꢀ)
209 (11 299), 257 (12 649), 359 (1444), 434 (1927).
-
RESULTS AND DISCUSSION
Syn th esis. To elucidate the structure of the RQ
analogues, spectral data of the analogues were com-
pared to RQ spectral data. The infrared spectrum of
RQ exhibited an aromatic C-H stretching at 3059 cm-1
and, more importantly, a conjugated ketone system at
1659 cm-1. The UV-vis scan indicated four primary
absorption signals at 209, 257, 359, and 437 nm. The
absorbance maximum at 257 nm indicated the presence
of an aromatic ring system while 359 and 437 nm
indicated further substitution and conjugation within
the molecular structure. All infrared and UV-vis data
were in agreement with those reported by Houlihan et
al. (1985), Knapp and Sharma (1985), and Lee et al.
(1990).
1,2,3,4-T e t r a h y d r o -4,4-d im e t h y l-14-m e t h y l-11,12-
p h en a n th r en ed ion e (7; Ta ble 1). Isolated as red-orange
crystals (123 mg, 0.49 mmol, 60%): mp 119-120 °C; IR (KBr)
2946, 2866, 1654, 1619, 1566, 1446, 1377, 1285, 1248, 859
cm-1; 1H NMR δ 7.60 (d, J ) 8.4 Hz, 1H), 7.32 (d, J ) 8.4 Hz,
1 H), 6.23 (s, 1 H), 3.09 (t, J ) 6.6 Hz, 2 H), 2.29 (s, 3 H), 1.69
(m, 2 H), 1.57 (m, 2 H), 1.24 (s, 6 H); 13C NMR δ 182.8, 181.3,
154.7, 150.8, 144.3, 134.4, 133.4, 129.2, 126.3, 124.6, 37.6, 34.4,
31.7 (2 C), 30.1, 21.0, 19.1; LRMS (EI, 70 eV) m/z 253 (M+
1); (CI, methane) m/z 255 (M + 1); UV-vis (methanol) λmax (ꢀ)
209 (10 987), 254 (12 453), 356 (1201), 413 (1422).
-
Oxid a tion Eva lu a tion . Analysis of Soybean Oil Compo-
nents. Tocopherols (R, δ), â-carotene, and chlorophyll stan-
dards were obtained from Sigma Chemical Co. (St. Louis, MO).
γ-Tocopherol was obtained from Eastman Chemical Products,
Inc. (Kingsport, TN). The tocopherols were determined by the
high-performance liquid chromatography method of Carpenter
(1979). Chlorophyll and carotenoids were spectrophotometri-
cally analyzed by modified Association of Official Analytical
Chemists (AOAC) methods (1984) (Hall and Cuppett, 1993).
Antioxidants. tert-Butylhydroquinone (TBHQ) was obtained
from Eastman. RQ and analogues were synthesized as
discussed above (see the synthetic procedures section).
Stripping of Soybean Oil. Commercial soybean oil (SBO)
was purchased from a local supermarket and stored in the
dark at -18 °C until needed. The oil was stripped using the
modified method of Kiritisakis and Dugan (1985) which
included a batch process rather than a column stripping (Hall,
1996). Stripping was conducted under a nitrogen atmosphere
until tocopherol was no longer detected by HPLC (2-4 h). The
bleaching material was filtered under a stream of nitrogen,
and the solvent was then removed in vacuo from the stripped
SBO at 30 °C.
Final purification was completed by re-suspending the SBO
into solvent (2.0 times the SBO content) and passing it through
a column containing the purification material [silicic acid
(36.4%); absorptive magnesia (27.2%; magnesium oxide); Hyflo
SuperCel (18.2%) and Florisil (18.2%) packed over Actisil (0.1
parts to the amount of PM)]. The SBO/solvent mixture was
passed through the column under nitrogen and vacuum.
Purification was monitored by spectrophotometry at 436 nm
(â-carotene) and 452 nm (Lutein) and was re-purified if
pigments were observed. Solvent was evaporated under
vacuum at 30 °C and the stripped SBO was stored at -18 °C
for no longer than 4 weeks.
The 1HNMR spectrum of RQ indicates several regions
which correspond to functional groups. RQ has a
doublet at δ 1.14 ppm which corresponds to the protons
of the methyl groups of the isopropyl region (Figure 2A;
protons at carbon 16 [C-16] and 17 [C-17]). The singlet
at δ 1.28 ppm indicated the six protons of the two
methyl groups of the aliphatic ring (Figure 2A; C-18 and
C-19). The methine protons (δ 2.99 ppm) correspond
to the protons at C-15, located in the isopropyl region,
while the triplet at δ 3.15 ppm indicated aliphatic
protons at C-1. Other aliphatic protons were found
between δ 1.6 and 1.8 ppm. The final region of concern
is that of the aromatic protons. RQ has three aromatic
signals at δ 7.06, 7.09, and 7.56 ppm (Figure 2A). As a
general rule, the proton at C-14 will give a singlet signal
and those at C-6 and C-7 doublet signals. The proton
at C-6 is influenced by the methyl protons (C-18 and
C-19) of the aliphatic ring and shifts the signal upfield
(δ 7.09 ppm) from the aromatic protons at C-7.
The 13CNMR spectrum of RQ confirmed the presence
of the aromatic, aliphatic, and alkyl carbons and sup-
ported the presence of the carbonyl functional groups
at δ 181 and 182 ppm (Figure 2B; C-11 and C-12). The
aromatic region (δ 125-160 ppm) represents all carbons
containing double bonds. The most intense signals (C-
6, C-7, and C-14) are attached to protons while the weak
signals indicate quaternary carbons (Figure 2B; C-5,
C-8, C-9, C-10, and C-13). C-9 and C-13 are influenced
by the magnetic field of the oxygen atoms and, as a
result, shift furthest downfield in comparison to the
other aromatic carbons. Carbons C-5 and C-10 are
influenced by the methyl protons and aliphatic protons,
respectively, and shift upfield when compared to C-8
Oxidation of Soybean Oils. Stripped soybean oil (100 g) was
weighed into 110 mL glass jars. RQ, RQ analogues, and TBHQ
were added to separate containers of stripped soybean oil at a
concentration of 200 ppm (0.02%). Each sample was thor-
oughly mixed to ensure complete dispersion of the antioxi-

1308 J. Agric. Food Chem., Vol. 46, No. 4, 1998
Hall III et al.
multiplets for C-2 and C-3 becomes one multiplet signal.
The 13CNMR spectrum of compounds 2 and 3 are
similar to RQ and all spectral data for compounds 2 and
3 are in agreement with Lee et al. (1990).
The replacement of the C-13 isopropyl group with a
tertiary butyl group (4, Table 1) results in the loss of
1
the methine H signal (δ 2.99 ppm) and doublet signal
(δ 1.14 ppm). Because the methyl groups give the
equivalent chemical shift, the singlet represents the
methyl protons at C-16, C-17, C-18, C-19, and C-20. The
carbon spectrum is similar to that of RQ with the
exception that C-16, C-17, and C-18 are represented by
a single peak at δ 29 ppm while the methyl carbons (C-
16 and C-17) are found near δ 22 ppm.
The removal of ring C substituents (5; Table 1) results
in a change in the aromatic region of the 1HNMR
spectrum. The protons at C-6 are strongly influenced
by the C-16 protons and, as a result, shift upfield from
δ 7.09 ppm in RQ to δ 6.35 ppm in compound 5. The
proton at C-13 and C-7 are shifted downfield due to the
deshielding properties of the aromatic protons. The
o-benzoquinones have similar downfield signals as with
C-13 and C-14 and give support to the labeling of these
protons. The intensity of the C-13 and C-14 signal in
the 13CNMR spectrum indicated the C-H bonding and
can be supported by the HETCOR techniques.
When the isopropyl group at C-13 is simply replaced
with a methyl group (6; Table 1), several noticeable
1
changes in the HNMR spectrum can be observed. The
loss of the methine signal (septet, δ 2.99 ppm), normally
found in the RQ spectrum, can be explained by the loss
of the two methyl groups at C-15. Because of the loss
of C-15 methyl groups, no hydrogens are present on
neighboring carbons (C-16 and C-17 as in RQ for
example; Figure 2a) to give a septet signal, and for this
same reason, the C-15 methyl gives rise to a singlet
signal. The chemical shift of the C-15 protons also
moves upfield. The loss of the C-15 methyl groups also
explains the disappearance of the doublet signal, found
in RQ, at δ 1.28 ppm. The final difference in the proton
spectrum is the downfield shift of the C-14 proton
beyond that of the C-6 proton signal. The reason for
this is that the C-15 methyl group has less of an
influence on the magnetic field of the C-14 proton, in
comparison to the isopropyl influence, thus allowing the
C-14 proton to shift downfield. The 13CNMR spectrum
remains similar to RQ except that the signal for C-15
can be found at δ 15 ppm instead of δ 26.7 ppm as in
RQ.
F igu r e 2. 1HNMR (A) and 13CNMR (B) spectra of ros-
mariquinone.
and C-9. The signal for C-5 is the weakest due to the
strong shielding of the methyl protons. The alkyl and
aliphatic region (δ 14-45 ppm) can be easily identified
and correlated to the appropriate protons using the
HETCOR technique. The methyl carbons can be found
at δ 30 ppm (aliphatic ring) and δ 21 ppm (isopropyl
group) and each corresponds to two carbons. The
HETCOR also established the location of the aliphatic
carbons (C-1, C-2, and C-3) at δ 29, 19, and 38 ppm,
respectively. C-4 and C-15 give the weakest signals due
to the neighboring methyl groups. The final analytical
data includes mass spectrometry and is used primarily
to support the proposed structure. In the case of RQ,
electron impact ionization mass spectrometry resulted
in a molecular ion of 254 m/z (M - 28) while the
chemical ionization resulted in molecular ion peak of
283 m/z (M + 1).
In all analogues, the infrared and UV-vis data were
similar to that of RQ, as would be expected. Ring A
alterations included the removal of one C-4 methyl (3;
Table 1) and both C-4 methyl groups (2; Table 1). The
removal of one methyl group resulted in a change in
the the δ 2.9-3.3 ppm region of the spectrum. The
biggest change was the presence of the triplet signal at
δ 3.2 and a quartet at 3.0 ppm and indicates the proton
at C-4 position. The removal of both C-4 methyl groups
resulted in the formation of a triplet signal at δ 2.9 ppm
and indicated the presence of the C-4 protons. In the
absence of C-4 methyl groups, the characteristic set of
1,2,3,4-Tetrahydro-4,4-dimethyl-14-methyl-11,12-
1
phenanthrenedione (7, Table 1) has a HNMR spectrum
similar to compound 6 (Table 1). The 1HNMR spectrum
for the C-14 methyl substituted compound has singlets
at δ 2.29 and 6.23 ppm. The singlet at δ 2.29 ppm and
the loss the methine signal (δ 2.99 ppm) support the
presence of the methyl substitution based on the discus-
sion above for the C-13 methyl substituted compound.
The reduced shielding of the C-15 methyl protons by
the C-12 oxygen atom results in a downfield shift of the
C-15 proton’s signal. The opposite holds true for the
C-13 proton, where the shielding effects of C-15 and
C-12 result in an upfield shift (δ 134 ppm) compared
with the δ 144 ppm shift observed for C-14 of analogue
6. All other carbons remain at similar chemical shifts.
The ultrasound-promoted Diels-Alder reaction re-
sulted in yields between 35 and 90% for the RQ
analogues. The poorest yield (35%) occurred in the
nonsubstituted ring C compound 1,2,3,4-tetrahydro-4,4-

Antioxidant Activity of Rosmariquinone
J. Agric. Food Chem., Vol. 46, No. 4, 1998 1309
electron density of the hydroxyl (OH) group, which in
turn decreases the bond dissociation energy of the AO-
H, resulting in an antioxidant that will readily donate
a hydrogen to the lipid radical. Also, phenoxy radical
can couple through the para position enhancing the
stability of the radical; preventing the antioxidant
radical from participating in free radical propagation
reactions (Miller and Quackenbush, 1957a,b; Gordon,
1990). Since the hydrogen (5) and methyl (6) are less
bulky than the C-13 isopropyl of RQ, the AOA was
expected to be worse than that of RQ.
For analogue 4, tert-butyl substitution at C-13 had
an AOA that was significantly (P < 0.05) worse than
that of RQ. This was an unexpected result since the
bulkier C-13 tert-butyl was expected to give a better
AOA for reasons mentioned above. The addition of a
methyl at the C-14 position (7) resulted in an AOA that
was not significantly different from RQ or analogue 3.
Both compounds 3 and 7 had slightly lower protective
times of 157 and 168 h, respectively, when compared to
RQ. These same analogues had nonsignificant AOA
differences when compared to analogues 2 and 4. All
analogues were found to control the oxidation relative
to the SBO control thus demonstrating the AO potential
for the analogues.
F igu r e 3. Number of hours RQ (1), RQ analogues (2-7), and
TBHQ delayed the oxidation of stripped SBO to a level of 20
meq/kg. Different letters represent significant differences (P
< 0.05) between treatments.
dimethyl-11,12-phenanthrenedione (5). There are two
primary reasons for the low yield of this cycloaddition:
(1) Several regioisomers may be produced during the
cycloaddition due in part to the catechol which lacks
ortho substitution and (2) Chang et al. (1990) reported
the synthesis of compound 5 but found that the com-
pound was unstable upon storage. The second lowest
yield (60%) was found in the 14-methyl-substituted
compound 1,2,3,4-tetrahydro-4,4-dimethyl-14-methyl-
11,12-phenanthrenedione (7). The presence of the
methyl at the meta position may have inhibited the
cycloaddition. The C-13 methyl (6), C-13 tert-butyl (4),
and C-13 isopropyl (2 and 3) analogues were formed in
yields between 80 and 90%, indicating the importance
of the C-13 (ortho) substitution in the orthoquinone. The
starting diene appears to have little effect on the
reduced yields of the cycloaddition products. From this
synthetic study, the ultrasound-promoted Diels-Alder
was a valuable method for the synthesis of structurally
similar phenanthrenediones.
An tioxid a n t Com p a r ison . The light-sensitized oxi-
dation study involving RQ and analogues in stripped
soybean oil provided some unexpected results (Figure
3). As expected, TBHQ had the highest (P < 0.05)
antioxidant activity (AOA) and delayed oxidation to the
20 meq/Kg level for 500 h, while the nearest RQ
derivative (5) was significantly (P < 0.05) lower at 294
h. Compounds 5 and 6 had the best AOAs of the
analogues and were significantly (P < 0.05) better than
the remaining analogues, including RQ. Although
compounds 5 and 6 had comparable AOAs, the methyl
substitution (6) was slightly greater (294 h versus 268
h) but not statistically different. Hall and Cuppett
(1997) proposed that the antioxidant mechanism of RQ
relies on the ability of RQ to convert to the active 11,-
12-dihydroxy compound arucadiol which then partici-
pates in hydrogen donation to a free radical. The
formation of this 11,12-dihydroxy structure would then
be expected to occur in all RQ analogues and thus act
as a phenolic antioxidant. The AOA of the hydrogen
(5) or methyl (6) C-13-substituted analogues was sur-
prisingly better (P < 0.05) than that of RQ. The ortho
substitution of phenolic antioxidants with large or bulky
side groups usually improved the AOA of a phenolic
compound (Miller and Quackenbush, 1957a,b). Large
ortho alkyl (e.g., isopropyl) substitutions enhance the
SUMMARY AND CONCLUSIONS
The AOA of the RQ analogues was found to vary
depending on the substitution pattern. The unexpected
improvement in the AOA by C-13 methyl (6) and C-13
hydrogen (5) and the unexpected decrease in AOA by
the C-13 tert-butyl substitution (4) suggest that the AOA
is not completely dependent on the hydrogen donating
activity observed in phenolic antioxidants. The ob-
served AOA of the RQ analogues probably is due to the
ability of the compounds to undergo a rearrangement
to the 11,12-dihydroxy analogues (Hall and Cuppett,
1997) following a photooxidation mechanism similar to
that of tanshinone IIa proposed by Kusumi et al. (1976).
The bulky tert-butyl C-13 substitution may interfere
with the peroxide polymer formation thus eliminating
or reducing the concentration of the important 11,12-
dihydroxy intermediates that form during the break-
down of the peroxide polymer. A more likely explana-
tion may be that the C-13 tert-butyl substitution
undergoes a radical coupling reaction with the C-12
oxygen radical intermediate (i.e., C-11,12 semiquinone)
similar to the mechanism proposed by Kurechi and
Kunugi (1983) for intermediates formed from the pho-
tooxidation of TBHQ. Under this mechanism, the C-12
hydroxyl group would not form resulting in a compound
that had fewer hydrogen, which could be donated to the
lipid radical, than the 11,12-dihydroxy intermediates,
thus explaining the reduced AOA observed in analogue
4. The opposite may be true for the C-13 methyl and
hydrogen substitution in that the important 11,12-
dihydroxy intermediate form quickly, thus creating the
active analogues.
The difference between AOA of the C-13 methyl (6)
and C-14 methyl (7) compounds may be due to the
observations that meta-substituted phenols tend to be
less active due to the inability to stabilize the phenoxy
radicals (Miller and Quackenbush, 1957a,b). The sig-
nificant reduction of AOA observed in analogue 2 is not
completely understood at this time but may be due to
the polymerization processes which reduces the concen-

1310 J. Agric. Food Chem., Vol. 46, No. 4, 1998
Hall III et al.
He, Y.; Chang, H.; Lau, Y.; Cui, Y.; Wang, R.; Mak, T.; Wong,
H.; Lee, C. Synthesis of diazepam related analogues of
miltirone, an active central benzodiazepine receptor ligand
isolated from Salvia miltiorrhiza Bunge (Danshen). J .
Chem. Soc., Perkin Trans. 1 1990, 3359-3361.
Houlihan, C.; Ho, C.-T.; Chang, S. The structure of ros-
mariquinone-A new antioxidant isolated from Rosmarinus
officinalis L. J . Am. Oil Chem. Soc. 1985, 62, 96-98.
Kiritsakis, A.; Dugan, L. R. Studies in photooxidation of olive
oil. J . Am. Oil Chem. Soc. 1985, 62, 892-896.
Knapp, S.; Sharma, S. Synthesis of the naturally occurring
antioxidant rosmariquinone. J . Org. Chem. 1985, 50, 4996-
4998.
Kurechi, T.; Kunugi, A. Studies on the antioxidants XIX:
Photooxidation products of tertiary butyl hydroquinone
(TBHQ) (II). J . Am. Oil Chem. Soc. 1983, 60, 1882-1887.
Kusumi, T.; Kishi, T.; Kakisawa, H.; Kinoshita, T. Photooxi-
dation of tanshinone IIa (6,7,8,9-tetrahydro-1,6,6-trimethyl-
phenanthro[1,2-b]furan-10,11-dione). J . Chem. Soc., Perkin
Trans. 1 1976, 1716-1718.
Lee, J .; Snyder, J . Ultrasound-promoted cycloadditions in the
synthesis of Salvia miltiorrhiza abietanoid o-quinones. J .
Org. Chem. 1990, 55, 4995-5008.
Lee, J .; Mei, H.; Snyder, J . Synthesis of miltirone by an
ultrasound-promoted cycloaddition. J . Org. Chem. 1990a ,
55, 5013-5016.
Luis, J .; Andres, L.; Fletcher, W.; Lahlou, El H.; Perales, A.
Rearrangement of methyl 11,12-di-O-methyl-6,7-didehydro-
carnosate in basic medium. Easy hemisynthesis of miltirone.
J . Chem. Soc., Perkin Trans. 1 1996, 2207-2211.
Meyer, W.; Schindler, E. Diterpenoid total synthesis, an A B
C approach. III. Total synthesis of ethyl d1-carnosate
dimethyl ether. Tetrahedron Lett. 1966, 36, 4261-4263.
Meyer, W.; Huffman, R.; Schroeder, P. Diterpenoid total
synthesis, an A B C approach-VI: 10-cyano-12-hydroxy-7-
oxo-17-norpodocarpa-5,8,11,13-tetraene, a model for tricyclic
C-aromatic diterpenoids. Tetrahedron 1968, 24, 5959-5969.
Miller, G.; Quackenbush, F. A comparison of alkylated phenols
as antioxidants for lard. J . Am. Oil Chem. Soc. 1957a , 34,
249-250.
tration of the active intermediates which form during
the photoenolization as proposed by Kusumi et al.
(1976).
On the basis of the results of this study, the RQ
analogues exhibited various AOA in a photooxidation
environment and thus demonstrate the importance of
chemically modifying natural antioxidants to improve
antioxidant capabilities.
LITERATURE CITED
AOAC. Official Methods of Analysis, 14th ed.; Association of
Official Analytical Chemists: Washington, DC, 1984; Method
Section 3.140 and 43.017.
AOCS. Official Methods and Recommended Practices of the
American Oil Chemists’ Society, 4th ed.; American Oil
Chemists’ Society: Champaign, IL, 1989; Method Cd-8-53.
Boatman, S.; Harris, T.; Hauser, C. Alkylations of the R-me-
thinyl group of R-formyl cyclic ketones through their dicar-
banions. Angular Alkylations. J . Am. Chem. Soc. 1965, 87,
82-86.
Carpenter, A. P. Determination of tocopherols in vegetable oil.
J . Am. Oil Chem. Soc. 1979, 56, 668-671.
Casiraghi, G.; Casnati, G.; Puglis, G.; Sartori, G.; Terenghi,
G. Selective reactions between phenols and formaldehyde.
A Novel Route to Salicylaldehydes. J . Chem. Soc., Perkin
Trans. 1 1980, 1862-1865.
Chang, H.; Chui, K.; Tan, F.; Yang, Y.; Zhong, Z.; Lee, C.;
Sham, H.; Wong, H. Structure-activity relationship of
miltirone, an active central benzodiazepine receptor ligand
isolated from Salvia miltiorhiza Bunge (Danshen). J . Med.
Chem. 1991, 34, 1675-1692.
Chang, H.; Choang, T.; Chui, K.; Hon, P.; Lee, C.; Mak, T.;
Wong, H. Novel constituents of Chinese drug Danshen
(Salvia miltiorhiza Bunge): Isolation of (6,7,8,9-tetrahydro-
1,6,6-trimethylfuro[3,2-c]naphth[2,1-e]oxepine-10,12-di-
one) and its analogues, secondary metabolites produced by
photooxidation reactions. J . Chem. Res. (S) 1990, 114-115.
Cuvelier, M.-E.; Berset, C.; Richard, H. Antioxidant constitu-
ents in sage (Salvia officinalis). J . Agric. Food Chem. 1994,
42, 665-669.
Dugan, L. R., J r.; Hoffert, E.; Blumen, G.; Dagkiewicz, I.;
Kraybill, H. The antioxidant behavior of the isomers of
butylated hydroxyanisole. J . Am. Oil Chem. Soc. 1951, 28,
493-495.
Edwards, J . D., J r.; Cashaw, J . L. The synthesis of 1,2-
dimethoxy-3-isopropylbenzene. J . Org. Chem. 1955, 20,
847-849.
Miller, G.; Quackenbush, F. Antioxidant activity as related to
effect of substituent groups. J . Am. Oil Chem. Soc. 1957b,
34, 404-407.
Nasipuri, D.; Mitra, A. Synthetic studies in the diterpene
series. Part VIII. Synthesis of miltirone, a diterpenoid
quinone. J . Chem. Soc., Perkin Trans. 1 1973, 285-287.
Shew, D.; Meyer, W. Diterpenoid total synthesis, and A B C
approach. V. Total synthesis of d1-carnosol dimethyl ether
and d1-carnosic acid dimethyl ether. Tetrahedron Lett. 1968,
25, 2963-2965.
Statistical Analysis System (SAS); SAS Institute, Inc., Cary
NC, 1985.
Steel, R.; Torrie, J . Principles and Procedures of Statistics. A
Biometrical Approach; McGraw-Hill Book Company: New
York, 1980; pp 173-177.
Gordon, M. The mechanism of antioxidant action in vitro. In
Food Antioxidants; Hudson, B. J . F., Ed.; Elsevier: London,
U.K., 1990; Chapter 1.
Hall, C., III. In A Mechanistic Study to Determine and Define
the Antioxidant Activity of Rosmariquinone. Ph.D. Disserta-
tion, University of Nebraska-Lincoln, Lincoln NE, 1996;
pp 143-145.
Hall, C., III; Cuppett, S. Effects of a bleached and unbleached
rosemary oleoresin on light-sensitized oxidation of soybean
oil. J . Am. Oil Chem. Soc. 1993, 70, 477-482.
Hall, C., III,; Cuppett, S. The hydrogen donating mechanism
of rosmariquinone. In Session 37. General Lipid Oxidation
and Quality I; Presented at the American Oil Chemist’s
Society 88th Annual Meeting and Exposition, Seattle, WA,
May 11-14, 1997.
Hall, C., III; Cuppett, S.; Wheeler, D.; Fu, X. Effects of bleached
and unbleached rosemary oleoresin and rosmariquinone on
light-sensitized oxidation of soybean oil. J . Am. Oil Chem.
Soc. 1994, 71, 533-535.
Haumann, B. Antioxidants: Firms seeking products they can
label as natural. Int. News Fats Oils Relat. Mater. 1991, 1,
1002.
Hayashi, T.; Kakisawa, H.; Hsu, H.-Y.; Chen, Y. The structure
of miltirone, a new diterpenoid quinone. Chem. Commun.
1970, 1, 299.
Tanis, S.; Abdallah, Y. A study of the Diels-Alder regioselec-
tivity of substituted vinyl cyclohexenes. Synth. Commun.
1986, 16, 251-259.
Tsuruta, H.; Mukai, T. The Grignard reaction of 3-hal-
otropolone methyl ethers. Bull. Chem. Soc. J pn. 1968, 41,
2489-2494.
Weng, C.; Gordon, M. Antioxidant activity of quinones ex-
tracted from Tanshen (Salvia miltiorrhia Bunge). J . Agric.
Food Chem. 1992, 40, 1331-1336.
Received for review August 28, 1997. Revised manuscript
received February 4, 1998. Accepted February 9, 1998.
Published as J ournal Series No. 11983, Nebraska Agricultural
Research Division, Department of Food Science and Technol-
ogy, University of Nebraska, Lincoln, NE 68583-0724.
J F970742K