Synergetic effect and structure-activity relationship of 3-hydroxy-3-methylglutaryl coenzyme a reductase inhibitors from crataegus pinnatifida bge

Article abstract of DOI:10.1021/jf903337f

The 3-hydroxy-3-methylglutaryl coenzyme A reductase (HMGR) inhibitors from hawthorn fruit (Crataegus pinnatifida Bge.) were isolated and evaluated for their antihyperlipidemic effect induced by high-fat diet in mice. After being further purified with silica and polyamide column chromatography from the fractions (fractions A, F, H, and G) with a high inhibitory rate (IR) to HMGR, 24 chromatographic fractions were obtained, including 8 active fractions with a high IR to HMGR. However, the total inhibitory activity of 24 fractions was decreased by about 70%. From eight active fractions, four compounds were obtained by recrystallization and identified as quercetin (a), hyperoside (b), rutin (c), and chlorogenic acid (d), the contents of which in hawthorn EtOH extract were 0.16, 0.32, 1.45, and 0.95%, respectively. The IR values of compounds ad to HMGR were 6.28, 9.64, 23.53, and 10.56% at the corresponding concentrations of 0.16, 0.32, 1.45, and 0.95 mg/mL, respectively. It was discovered that the IR of a mixture (2.85 mg/mL) matching the original percentage of compounds a-d in hawthorn EtOH extract was up to 79.5%, much higher than that of the single compound and the total IR of these four compounds (50.01%). The in vivo results also revealed that the mixture had a more significant lipid-lowering efficacy than the monomers. Structure-activity relationship revealed the inhibitory activity and lowering-lipid ability of compounds a-c decreased with increasing glycoside numbers. It was concluded that there were synergetic effects on inhibiting HMGR and lowering lipid among compounds a-d, and the weak hydrophilic ability benefits the inhibition to HMGR and lowering-lipid efficacy.

Full text of DOI:10.1021/jf903337f

3132 J. Agric. Food Chem. 2010, 58, 3132–3138
DOI:10.1021/jf903337f
Synergetic Effect and Structure-Activity Relationship of
3-Hydroxy-3-methylglutaryl Coenzyme A Reductase Inhibitors
from Crataegus pinnatifida Bge.
§
X
IAO-L
I
Y
E
,† WEN
-
WEN
ANG,§ YING
H
UANG,§ ZHU
C
HEN,§ XUE-GANG
§
L
I
,*,§ PING
L
I
,# PING
§
L
AN
,
L
IANG
W
G
AO,§ ZHONG-Q
I
Z
HAO
,
AND
X
IN
C
HEN
†School of Life Science, Southwest University, Chongqing 400715, China, §Chemistry Institute of
Pharmaceutical Resources, School of Pharmaceutical Science, Southwest University, Chongqing 400716,
China, and #Beibei Tranditional Chinese Medicie Hospital, Chongqing 400700, China
The 3-hydroxy-3-methylglutaryl coenzyme A reductase (HMGR) inhibitors from hawthorn fruit
(Crataegus pinnatifida Bge.) were isolated and evaluated for their antihyperlipidemic effect induced
by high-fat diet in mice. After being further purified with silica and polyamide column chromatog-
raphy from the fractions (fractions A, F, H, and G) with a high inhibitory rate (IR) to HMGR, 24
chromatographic fractions were obtained, including 8 active fractions with a high IR to HMGR.
However, the total inhibitory activity of 24 fractions was decreased by about 70%. From eight active
fractions, four compounds were obtained by recrystallization and identified as quercetin (a),
hyperoside (b), rutin (c), and chlorogenic acid (d), the contents of which in hawthorn EtOH extract
were 0.16, 0.32, 1.45, and 0.95%, respectively. The IR values of compounds a-d to HMGR
were 6.28, 9.64, 23.53, and 10.56% at the corresponding concentrations of 0.16, 0.32, 1.45, and
0.95 mg/mL, respectively. It was discovered that the IR of a mixture (2.85 mg/mL) matching the
original percentage of compounds a-d in hawthorn EtOH extract was up to 79.5%, much higher
than that of the single compound and the total IR of these four compounds (50.01%). The in vivo
results also revealed that the mixture had a more significant lipid-lowering efficacy than the
monomers. Structure-activity relationship revealed the inhibitory activity and lowering-lipid ability
of compounds a-c decreased with increasing glycoside numbers. It was concluded that there were
synergetic effects on inhibiting HMGR and lowering lipid among compounds a-d, and the weak
hydrophilic ability benefits the inhibition to HMGR and lowering-lipid efficacy.
KEYWORDS: Hawthorn fruit; 3-hydroxy-3-methylglutaryl coenzyme A reductase (HMGR); antihyperlipi-
demic effect; synergetic effect; structure-activity relationship (SAR)
1. INTRODUCTION
clinics, reduce TC, LDL-C, apolipoprotein B, and TG levels (5),
but they are associated with undesirable side effects such as severe
myopathy and statin-associated memory loss (6). Myopathy is a
serious side reaction, with the potential for rhabdomyolysis (the
pathological breakdown of skeletal muscle), leading to acute
renal failure (7).
Recently, we conducted a rationalized screening to search for
HMGR inhibitors from traditional Chinese medicine (TCM).
The EtOH extract of hawthorn fruit (Crataegus pimmatifida Bge.)
showed a very significant inhibition to HMGR by HPLC (8)
among 73 TCMs. C. pimmatifida Bge. (Rosaceae), named
Shanzha in Chinese, is widely used as a food and traditional
medicine in treating chronic heart failure, high blood pressure,
various digestive ailments(9), and arrhythmia(10), as well as
geriatric and antiarteriosclerosis remedies (11) by reducing serum
cholesterol. It has been documented that hawthorn fruit de-
creased the serum total cholesterol, LDL-C, and triglycerides in
hyperlipidemic humans (12).
Hyperlipidemia is a dangerous factor of cardio-cerebrovascu-
lar diseases, inducing essential hypertension, coronary heart
disease, and atherosclerosis (1). It is widely acknowledged that
lowering the level of total cholesterol (TC), low-density lipopro-
tein cholesterol (LDL-C), and triglycerides (TG) can interfere
with the progression of atherosclerosis and reduce cardiovascular
events inpatients with atherosclerosis and cardio-cerebrovascular
diseases (2, 3). The most widely used lipid-lowering drugs are a
class of 3-hydroxy-3-methylglutaryl-coenzyme
A reductase
(HMGR) inhibitors. HMGR is an endoplasmic reticulum (ER)
bound enzyme that catalyzes the conversion of hydroxymethyl-
glutaryl-CoA to mevalonate (4), an early rate-limiting step in
cholesterol biosynthesis, leading to cholesterol-lowering effects.
Statins, a competitive inhibitor of HMGR most widely used in
*Address correspondence to this author at the Chemical Institute of
Pharmaceutical Resources, School of Pharmaceutical Science, South-
west University, Chongqing 400715, China (telephone þ86 23 68250728;
fax þ86 23 68250728; e-mail xuegangli2000@yahoo.com.cn).
The major components include flavonoids, proanthocyanidin,
triterpenes, organic acids, tannin, flavane and its polymers, and so
pubs.acs.org/JAFC
Published on Web 02/04/2010
© 2010 American Chemical Society

Article
J. Agric. Food Chem., Vol. 58, No. 5, 2010 3133
on(13). Among them, flavonoids and triterpenes were reported as
the main active hypolipidemic constituents (14, 15). However,
there had been no reports on the active components of HMGR
inhibitors in hawthorn fruit and their lipid-lowering efficacy.
In the present study, we separated and purified the EtOH
extract of hawthorn fruit by systematic chromatographic meth-
ods using HMGR as the target to research the inhibition to
HMGR and the lipid-lowering efficacy of active components and
further investigate their synergetic effect and structure-activity
relationship.
Table 1. Inhibition Rate of Constituents to HMG-CoA Reductase and of
Counts
weight
(g)
weight ratio
(%)
total inhibitory
activityb
IRa (%)
crude extract
680
100
5.1
18.4
0.01
0.2
34680
6476.80
44.80
precipitation (fraction A)
water fraction (fraction B)
10% EtOH (fraction C)
20% EtOH (fraction D)
30% EtOH (fraction E)
40% EtOH (fraction F)
50% EtOH (fraction G)
60% EtOH (fraction H)
70% EtOH (fraction I)
80% EtOH (fraction J)
90% EtOH (fraction K)
100% EtOH (fraction L)
35.2
448.06
24.92
11.92
16.63
43.6
31.5
18.5
5.23
2.09
4.5
5.18
65.89
3.66
1.75
2.45
6.41
4.63
2.72
0.77
0.31
0.66
0.86
49.84
71.52
0.6
5.6
931.28
12338.8
9544.5
4865.5
78.45
28.3
30.3
26.3
1.5
2. MATERIALS AND METHODS
2.1. Chemicals. Standard quercitrin, hyperoside, rutin, and chloro-
genic acid were purchased from the National Institute for the Control of
Pharmaceutical Biological Products (Beijing, China). Nicotinamide ade-
nine dinucleotide phosphate (NADPH), 1,4-dithiothreitol (DTT), ethyle-
nediaminetetraacetic acid (EDTA), and 3-hydroxy-3-methylglutaryl-
coenzyme A (HMG-CoA) were obtained from Sigma Chemical Co. (St.
Louis, MO). Other chemicals were purchased from Pharmaceutical and
Chemical Reagent Inc. (Chongqing, China). Silica gel (200-300 mesh),
polyamide (100-200 mesh), and D-101 macroporuos resin were pur-
chased from Qingdao Marine Chemical Co., Ltd. (Qingdao, China).
2.2. Plant Materials. Hawthorn fruit (C. pimmatifida Bge.) was
commercially available in Chongqing, China. It was dried at 75 °C in a
vibrating blast drier (model VFB-A, Beijing, China) and ground to pass
through a 60 mesh screen; the powder was stored in a desiccator at the
Chemistry Institute of Pharmacological Resource, Southwest University,
China.
2.5
1.3
52.25
58.5
5.86
2.2
128.92
total
648.01
95
34641.16
a IR indicates the inhibitory rate to HMGR at the concentration of 1 mg/mL. b Total
inhibitory activity = weight (mg) ꢀ IR (%).
column (8 cm ꢀ 100 cm), then eluted with different concentrations of
EtOH (0, 10, 20, 30, 40, 50, 60, 70, 80, 90, and 100%) to collect various
fractions, and concentrated under reduced pressure to obtain condensed
fractions designated fractions B, C, D, E, F, H, I, J, K, and L, respectively.
Then the IR of fractions A-L to the activity of HMGR was measured by
HPLC (Table 1).
2.5. Purification of Active Fractions with High Inhibition to
HMGR. According to results of section 2.4, four fractions (A, F, G, and
H) had a high inhibitory activity to HMGR (Table 1) and were further
purified.
One gram of fraction A was suspended in MeOH and loaded onto 10 g
of silicagel. After MeOH had been evaporated, silica gel with fraction A
was put into column (1.5 cm ꢀ 70 cm) containing 90 g of silica gel, and then
eluted with different ratios of CHCl3/MeOH to produce six fractions
(fractions 1-6) (shown in Table 2).
Fraction F (1 g) was loaded onto column chromatography on 100 g of
polyamide (1.5 cm ꢀ 70 cm) and eluted with a gradient of H2O/MeOH
eluent to yield six fractions (fractions 7-12).
Fraction G (1 g) was loaded onto 100 g of polyamide chromatography
(1.5 cm ꢀ 70 cm) and eluted with a gradient of H2O/EtOH to yield five
fractions (fractions 13-17).
2.3. Preparation and Assay of HMGR from Mouse Liver (8).
Briefly, liver was homogenized in potassium phosphate buffer A (PPB, pH
7.2), containing 50 mmol/L KCl, 40 mmol/L K2PO4, 30 mmol/L EDTA,
and 10 mmol/L sucrose. The homogenate was centrifuged at 16000g for 15
min, and the supernatant was centrifuged at 100000g for 60 min. The
resulting microsomal fraction was suspended in 3 mL of potassium
phosphate buffer A (pH 7.2), containing 250 mmol/L sucrose, 20 mmol/
L EDTA, and 50 μmol/L leupeptin, immediately frozen in liquid nitrogen,
and stored at -80 °C until analysis.
Microsomal suspensions of 10 μL (80 mg/mL protein) and 10 μL
inhibitors were preincubated for 5 min at 37 °C with 80 μL of PPB (pH
6.8), containing 300 mmol/L KCl, 240 mmol/L K2HPO4, 6 mmol/L
EDTA, and 15 mmol/L DTT. The assay was initiated by adding 10 μL
of 50 μmol/L NADPH and 10 μL of HMG-CoA. The incubation was
performed at 37 °C for 30 min and terminated by the addition of 500 μL of
0.5 mol/L of NaOH. Then the mixture was centrifuged at 10000g for 10
min, and 20 μL of supernatant was taken to measure the content of
NADPH by HPLC (8). In this reaction, the NADPH would be oxidized to
NADPþ, causing the decrease in absorbance of NADPH at 340 nm. After
adding inhibitors, the reaction velocity was decreased and the depletion of
NADPH was correspondingly declined. The changes in the content of
NADPH were analyzed, and the inhibitory rate (IR) was calculated
according to the following equation:
Fraction H (1 g) was loaded onto 100 g of silica gel chromatography
(1.5 cm ꢀ 70 cm) and eluted with a gradient of CHCl3/MeOH to yield
seven fractions (fractions 18-24).
The IR of fractions 1-24 to HMGR was measured by HPLC.
Those fractions with a high inhibition to HMGR were purified with thin
layer chromatography (TLC) and recrystallized to obtain eight compounds,
a-h, respectively. The separation process is shown in Scheme 1.
2.6. Structural Identification of Active Compounds. Melting
1
points, UV, IR, H NMR, and TLC were used to identify structures of
new compounds. Melting points were determined on an RD-2C electro-
thermal melting point apparatus and are uncorrected. The UV spectra
were recorded on a Hitachi U-1800 spectrophotometer. The IR spectra
Si -Sc
Sb -Sc
IR ¼
ꢀ 100%
ð1Þ
were carried out ona Perkin-Elmer IR spectrophotometer. The 1H and 13
C
Here, IR indicates the inhibitory rate and Si, Sb, and Sc indicate the peak
area (mAU*s) of NADPH in inhibitor, blank, and control group,
respectively.
HPLC conditions of analyzing NADPH were as follows: chromato-
graphic column, Hypersil C18-ODS (4.6 mm ꢀ 200 mm, 5 μm); mobile
phase, phosphate buffer/MeOH (78:22, pH 7.2); flow rate, 1 mL/min;
injection volume, 20 μL; wavelength of detection, 340 nm; temperature of
column, 25 °C.
NMR spectra were recorded on a Bruker model Avance DMX300
spectrometer (300 MHz for 1H and 75 MHz for 13C) using TMS as an
internal standard and DMSO-d6 as solvent. Macroporous resin was used
for crude separations. Analytical TLC was performed on precoated silica
gel G plates (Qingdao Marine Chemical Co. Ltd.) and polyamide thin film
(Yuanda Chemical Co. Ltd.), visualized with UV light, and then sprayed
with 1% vanillin/H2SO4 and heated or with 1%AlCl3/EtOH.
2.7. Quantitative Analysis of Active Components in Hawthorn
Fruit Extract. For quantification of active components in hawthorn fruit
extract, a Waters HPLC system equipped with a 510 pump, a UV detector,
and a Symmetry Hypersil C18-BDS (4.6 ꢀ 250 mm, 5 μm) was used.
Quercetin, hyperoside, rutin, and chlorogenic acid in methanol were
injected into the column as external standard compounds. After injection,
the column was eluted with a linear gradient of acetonitrile in water from
2.4. Preparation of Hawthorn Fruit Extracts. The dried powder
of hawthorn fruit (1000 g) was extracted three times with 15 L of 95%
ethanol under ultrasonic vibration. The supernatant was evaporated at 60
°C ina rotary evaporator under reduced pressure to produce 680 g of crude
extracts. Then the extract was dissolved in distilled water to obtain
supernatant and precipitate (35.2 g, fraction A) by centrifugation. The
supernatant was loaded onto the D-101 macroporous adsorption resins

3134 J. Agric. Food Chem., Vol. 58, No. 5, 2010
Ye et al.
Table 2. Contents and IR of Chromatographic Fractions to HMGR
IRa (%)
Six hours after the last treatment, the mice were sacrificed, their blood
was collected, and the serum for the measurement of cholesterol levels was
obtained by centrifugation. Then total cholesterol (TC), triglyceride (TG),
high-density lipoprotein-cholesterol (HDL-C), and low-density lipopro-
tein-cholesterol (LDL-C) were assayed using commercially available kits.
2.9. Statistics. The data were analyzed by SPSS11.5 software and
expressed as mean ( standard deviation (SD), and one-way analysis of
variance (ANOVA) was used for statistical evaluation. Differences were
accepted as statistically significant at P values of <0.05.
fraction
weight (mg)
weight ratio (%)
total activityb
A
1
1000
99
100.0
9.9
18.4
2.5
184.0
2.5
2
78
7.8
39.6
2.4
30.9
6.1
3
255
149
50
25.5
14.9
5.0
4
4.6
0.2
6.9
0.1
5
6
sum
154
785
15.4
78.5
10.3
15.9
62.3
3. RESULTS AND CONCLUSION
3.1. IR of Fractions A-L to the Activity of HMGR. Table 1
showedthe IR ofhawthornfruit extracttothe activity ofHMGR.
From 1000 g of dried powder of hawthorn fruit, 680 g of crude
EtOH extract was obtained. Its IR to HMGR was only 5.1%, and
the total inhibitory activity was about 34680.
F
1000
61
100.0
6.1
28.3
5.9
283.0
3.6
7
8
80
8.0
30.4
16.5
1.9
24.3
46.7
1.7
9
283
92
28.3
9.2
10
11
12
sum
80
8.0
11.1
7.9
8.9
Then the crude extract was dissolved in water to get the
precipitate and supernatant. The IR of precipitate (35.2 g,
fraction A) was up to 18.4%, and the total inhibitory activity
was about 6476.8. From the supernatant, 11 fractions (fractions
B-L) were obtained. Among them, fractions F, G, and H had
high inhibitory activities to HMGR, their IRs being up to 28.3,
30.3, and 26.3%, respectively. According to the data in Table 1,
the total activity of fractions A-L was 34641.16. This indicated
that there had been no loss of total inhibitory activity. The
inhibitory activities of fractions A, F, G, and H accounted for
95.9%, but their weights were only 18.9% of total weight.
3.2. IR of Fractions 1-24 to the Activity of HMGR. Fractions
A, F, G, and H were further purified with different column
chromatographies to obtain 24 fractions (fractions 1-24)
(Table 2).
Six fractions were obtained from precipitate (fraction A).
Among them, fraction 2 showed the highest inhibitory activity
to HMGR, up to 39.6%, 2-fold higher than that of the pre-
cipitate. However, after purification by column chromatography,
the loss of weight was only 21.5%, and total activity was up to
66.2%, compared with that of fraction A (Table 2). Fraction 2
was further purified with TLC and recrystallized in MeOH, and a
pure compound (a) with yellow needle-like crystals was obtained
and identified as quercetin.
Similarly, six fractions (fractions 7-12) were obtained from
fraction F. The loss of total weight and total inhibitory activity
were 23.4 and 65.2%, respectively. Among six fractions, fractions
8, 9, and 11 had strong inhibitory activities, their IR values being
30.4, 16.5, and 11.1%, respectively. After purification by TLC
and recsytallization in MeOH, compounds b, c, and d were
obtained from fractions 8, 9, and 11, respectively, and identified
as hyperoside (b), rutin (c), and chlorogenic acid (d).
After purification with polyamide, five fractions were obtained
from fraction G. The loss of total weight and total inhibitory
activity were 23.3 and 56.9%, respectively. Two fractions (14 and
15) with high IR to HMGR values were obtained from fraction
G. After purification by TLC and recrystallization in MeOH,
compounds e and f were produced from fractions 14 and 15,
respectively, and identified as hyperoside (b) and rutin (c).
After purification on a silica gel column, seven fractions were
produced from fraction H. According to the data in Table 2, the
loss of total weight and inhibitory activity of these fractions were
25.7 and 64.3%, respectively. From these fractions, two fractions
(19 and 21) with high IR to HMGR values were obtained. After
purification by TLC and recrystallization in MeOH, compounds
g and h were produced from fractions 19 and 21, respectively, and
identified as quercetin (a) and hyperoside (b).
170
766
17.0
76.6
13.4
98.7
G
1000
72
100.0
7.2
30.3
3.4
303.0
2.4
13
14
15
16
17
sum
308
164
80
30.8
16.4
8.0
30.4
16.5
0.3
93.6
27.1
0.2
153
777
15.3
77.7
4.8
7.3
130.7
H
1000
120
15
100.0
12.0
0.8
26.3
6.9
263.0
8.3
18
19
20
21
22
23
24
sum
39.6
2.4
5.9
1.6
65
6.5
128
73
12.8
7.3
30.4
1.9
38.9
1.4
30
3.0
2.6
11.9
0.8
312
713
31.2
74.3
37.1
93.9
a IR indicates the inhibitory rate to HMGR at the concentration of 1 mg/mL. b Total
inhibitory activity = weight (mg) ꢀ IR (%).
10 to 50% (containing 0.3% KH2PO4) over 30 min before returning to
initial conditions for 10 min. The flow rate was 1 mL/min, and effluent was
monitored at 360 nm.
Before being analyzed, the EtOH extract of hawthorn fruit was
dissolved in methanol, and after a brief centrifugation (1000g, 10 min), a
5 mL aliquot of the supernatant was passed through a 2 mL cleanup
column and eluted with methanol. Twenty microliters of collected eluant
was used to analyze the content of active compounds by HPLC.
2.8. Animal Experiments. Kunming mice (20 ( 2 g) of both
genders were purchased from Animal Breeding Center of the Third
Military Medical University (Chongqing, China). Animals were cared
for according to the institutional guidelines of Chongqing City Laboratory
Animal Administration Committee of China. Mice were housed in an air-
conditioned room (24 ( 2 °C, 55 ( 10% relative humidity, 15 air changes
in 1 h, and 12 h light cycle) with 5 mice per cage. The regular and the high-
fat and high cholesterol (HFHC) diets contained 2% cholesterol, 10%
lard, 10% yolk powder, and 0.5% sodium deoxycholate, purchased from
Animal Breeding Center of the Third Military Medical University
(Chongqing, China).
After 1 week of accommodation period, the mice were divided into
normal control (HFHC free), model (HFHC diet), compound a (2.85 mg/
kg/day), compound b (2.85 mg/kg/day), compound c (2.85 mg/kg/day),
compound d (2.85 mg/kg/day), and mixture of compounds a-d (2.85 mg/
kg/day), and 10 mice were included in each group. The mixture of
compounds a-d was matched according to their ratio in hawthorn fruit
extract (0.16:0.32:1.42:0.95). Mice in the normal control were fed regular
mouse diet. Other animals were fed HFHC diet for 2 months to get the
hyperlipidemic mice. Then hyperlipidemic mice were untreated or treated
with above extracts using a stomach tube for 6 weeks, respectively.
During the period of separating active compounds with the
different column chromatographies, we observed that the total

Article
J. Agric. Food Chem., Vol. 58, No. 5, 2010 3135
Scheme 1. Scheme of Separating Active Compounds with High Inhibition to HMGR from Crataegus pinnatifida Bge.
weight of those fractions (A, F, G, and H) with a high IR was
reduced by 20-30%, and the total activity (fractions 1-24) was
reduced by 50-70%. This indicated that further purification to
crude extract weakened the inhibitory activity of active compo-
nents to HMGR, and it was speculated that there was a synergetic
effect among the active components.
quercetin, indicating the aglycone is quercetin. Another result of
the TLC chromatogram confirmed they contained galactose.
Their structural properties were as follows: yellow amorphous
powder (MeOH), mp 234-236 °C; ESI-MS (m/z) 465 [ M þ 1];
Anal. Cacld C21H20O12; UV λmax (MeOH) 359 nm; IR v (cm-1
)
3436.6 (OH stretching vibration), 2923.2 (saturated C;H
stretching vibration), 1654.5 (CdO stretching vibration),
1607.1, 1505.1, 1452.4 (aromatic ring skelton vibration), 1384.4,
1363.9, 1305.6, 1261.7, 1205.6, 1172.7, 116.4, 1134.8, 996.2, 819.7,
3.3. Structural Identification of Active Compounds from Haw-
thorn Fruit. Compounds a and g were negative in Molish’s
reaction and positive in HCl-Mg reaction. Both of them had
the same structural properties: yellow needle-like crystal
(MeOH), mp 310-312 °C; EI-MS (m/z) 302 [M þ 1]; Anal.
Calcd C15H10O7; UV λmax (MeOH) 359 nm; IR v (cm-1) 3416.82
(OH stretching vibration), 1662.1 (CdO stretching vibration),
1612.2, 1560.4, 1522.6, 1450.7 (aromatic ring skelton vibration),
1383.8, 1320.8, 1320.1, 1264.2, 1200.7, 1170.0, 1132.3, 1093.1,
1015.2, 942.1, 864.4, 842.4, 824.9; 1H NMR (300 MHz, DMSO-
d6) δ 12.49 (s, 1H, 5-Ar-OH), 10.78 (s, 1H, 7-Ar-OH), 9.60
(s, 1H, 40-Ar-OH), 9.37 (s, 1H, 3-Ar-OH), 9.31 (s, 1H, 30-Ar-
OH), 7.67 (s, 1H, 20-Ar-H), 7.54 (d, 1H, J = 8.4 Hz, 60-Ar-H),
6.88 (d, 1H, J = 8.4 Hz, 50-Ar-H), 6.41 (s, 1H, 8-Ar-H), 6.19 (s,
1H, 6-Ar-H); 13C NMR (75 MHz, DMSO-d6) δ 175.87 (C-4),
163.91 (C-7), 160.75 (C-5), 156.16 (C-9), 147.73 (C-40), 146.82 (C-
2), 145.09 (C-30), 135.77 (C-3), 121.98 (C-10), 120.00 (C-60), 115.63
(C-20), 115.09 (C-50), 103.04 (C-10), 98.21 (C-6), 93.38 (C-8).
According to the results of structural identification and color
reaction, compounds a and g were confirmed as quercetin, 4H-1-
benzopyran-4-one, 2-(3,4-dihydroxyphenyl)-3,5,7-trihydroxy,
assigned as compound a.
1
796.7; H NMR (300 MHz, DMSO-d6) δ 12.73 (s, 1H, 5-Ar-
OH), 10.96 (s, 1H, 7-Ar-OH), 9.83 (s, 1H, 40-Ar-OH), 9.25 (s,
1H, 30-Ar-OH), 7.78-7.75 (dd, 1H, J = 5.9 Hz, 60-Ar-H), 7.62
(s, 1H, 20-Ar-H), 6.91 (d, 1H, J = 8.5 Hz, 50-Ar-H), 6.49 (d, 1H,
J = 1.4, 8-Ar-H), 6.29 (s, 1H, 6-Ar-H), 5.47 (d, 1H, J = 7.6,
H-Ar-100), 5.23 (br s, 1H, galactosyl OH), 4.95 (br s, 1H,
galactosyl OH), 4.53 (br s, 2H, galactosyl OH), 3.74 (s, 1H,
galactosyl H), 3.66-3.26 (m, 6H, H-200-600); 13C NMR (75 MHz,
DMSO-d6) δ 177.48 (C-4), 164.10 (C-7), 161.22 (C-5), 156.28 (C-
2, C-9), 148.44 (C-40), 144.81 (C-30), 133.49 (C-3), 121.98 (C-60),
121.10 (C-10), 115.95 (C-50), 115.17 (C-20), 103.92 (C-10), 101.81
(C-100), 98.65 (C-6), 93.48 (C-8), 75.83 (C-500), 73.19 (C-300), 71.20
(C-200), 67.92 (C-400), 60.13 (C-600). According to the results of
structural identification and color reaction, it was confirmed that
compounds b, e, and h are hyperoside, 4H-1-benzopyran-4-one,
2-(3,4-dihydroxyphenyl)-3-(β-D-galactopyranosyloxy)-5,7-dihy-
droxy, assigned as compound b.
Compounds c and f were all positive in HCl-Mg reaction,
aluminum chloride (AlCl3) reaction, and Molish’s reaction. Both
of them had the same structural properties: yellow needle-like
crystal (MeOH), mp 195-198 °C; ESI-MS (m/z) 610 [M þ 1];
Compounds b, e, and h were all positive in Molish’s reaction
and HCl-Mg reaction, indicating they were a kind of flavonoid
glycoside. Those in absolute methanol were hydrolyzed by 2%
H2SO4 at 60 °C and then neutralized with NaOH to remove
H2SO4. After being concentrated under reduced pressure, the
residue was recrystallized in ethyl acetate to get the yellow crystal.
Its mp was 310-312 °C. The Rf value was the same as for
Anal. Calcd C27H30O16; UV λmax (MeOH) 357 nm; IR v (cm-1
)
3429.3 (OH stretching vibration), 2983.5, 2937.7, 2901.1 (saturated
C-H stretching vibration), 1654.1 (CdO stretching vibration),
1599.8, 1505.9, 1454.5 (aromatic ring skelton vibration), 1362.4,
1295.8, 1233.6, 1203.3, 1168.8, 1124.2, 1091.9, 1061.4, 1042.1,

3136 J. Agric. Food Chem., Vol. 58, No. 5, 2010
Ye et al.
1
1014.0, 969.7, 944.0, 911.9, 879.7, 827.7, 807.9; H NMR (300
MHz, DMSO-d6) δ 12.69 (s, 1H, 5-Ar-OH), 10.93 (s, 1H, 7-
Ar-OH), 9.77 (s, 1H, 40-Ar-OH), 9.28 (s, 1H, 30-Ar-OH), 7.63
(d, 2H, J = 7.4, 60-Ar-H), 6.92 (d, 1H, J = 8.6 Hz, 50-Ar-H),
6.47 (s, 1H, 8-Ar-H), 6.28 (d, 1H, 6-Ar-H), 5.43 (d, 1H, J = 6.3
Hz, 100-H), 4.47 (d, 1H, J = 9.3 Hz, 1000-H), 1.07 (d, 3H, J = 5.9
Hz, 6000-H); 13C NMR (75 MHz, DMSO-d6) δ 177.37 (C-4),
164.05 (C-7), 161.22 (C-5),156.42 (C-2), 156.60 (C-9), 148.40 (C-
40), 144.74 (C-30), 133.31 (C-3), 121.58 (C-10), 121.18 (C-60),
116.27 (C-50), 115.22 (C-20), 103.97 (C-10), 101.19 (C-100),
100.73 (C-1000), 98.66 (C-6), 93.57 (C-8), 76.46 (C-300), 75.91(C-
500), 74.07 (C-200), 71.85 (C-4000), 70.56(C-400), 70.37(C-2000),
70.01(C-3000), 68.23 (C-5000), 66.99 (C-600), 17.71 (C-6000). It was
confirmed that compounds c and f were rutin, 4H-1-benzopyran-
4-one, 3-[(6-O-(6-deoxy-R-
L
-mannopyranosyl)-β-D-glucopyranosyl]-
oxy]-2-(3,4-dihydroxyphenyl)-5,7-dihydroxy, assigned as com-
pound c.
Compound d was blue in ferric chloride (FeCl3) reaction:
yellow powder (MeOH/H2O), mp 207-209 °C; ESI-MS (m/z)
354 [M þ 1]; Anal. Calcd16H18O9; UV λmax (MeOH) 328.5 nm; IR
v (cm-1) 3389.5 (OH stretching vibration), 2956.0, 2928.8, 2857.1
(C;H stretching vibration), 2626.4, 1685.6 (CdO stretching
vibration), 1638.9, 1614.3, 1528.7, 1517.7, 1442.8, 1384.1 (aro-
matic ring skelton vibration), 1289.4, 1252.4, 1190.4, 1158.8,
1133.6, 1114.1, 1086.2, 1038.1, 970.2, 818.4; 1H NMR (300
MHz, H2O-d6) δ 7.47 (d, 1H, J = 15.9 Hz, R-H), 7.03 (s, 1H,
20-H), 6.96 (d, 1H, J = 8.3 Hz, 60-H), 6.83 (d, 1H, J = 8.3 Hz, 50-
H), 6.17 (d, 1H, J = 15.9 Hz, β-H), 5.25-5.18 (m, 1H, 3-H), 4.19
(d, J = 3.0 Hz, 5-H), 3.80 (dd, 1H, J = 8.5 Hz, 4-H), 2.22-2.00
(m, 4H, 2-H, 6-H); 13C NMR (75 MHz, H2O-d6) δ 1176.57 (C-70),
168.13 (C-9), 146.60 (C-4), 145.71 (C-3,), 143.72 (C-7), 126.39 (C-
1), 122.24 (C-6), 115.65 (C-5), 114.65 (C-2), 113.78 (C-8), 74.40
(C-10), 70.93 (C-40), 70.13 (C-30), 68.70 (C-50), 36.08 (C-60), 35.95
(C-20). It was confirmed that compound d was chlorogenic
acid, (1S,3R,4R,5R)-3-[[3-(3,4-dihydroxyphenyl)-1-oxo-2-pro-
penyl]oxy]-1,4,5-trihydroxycyclohexanecarboxylic acid.
The structures of compounds a-d are shown in Figure 1.
3.4. Quantification of Four Active Components in Hawthorn
Fruit Extract. The HPLC chromatogram of the ethanol extract of
hawthorn fruit is shown in Figure2. There are more than 10 peaks
in the extract, which accounted for about 4%. In 100 g of
dried extract, chlorogenic acid (d) accounted for 0.95 g; rutin
(c), 1.42 g; hyperoside (b), 0.32 g; and quercetin (a), 0.16 g. In
addition, ursolic acid accounted for 0.32 g; oleaniolic, 0.08 g; and
maslinic acid, 0.08 g. The total content of compounds a-d was
about 2.85 g in 100 g of dried extract of hawthorn fruit.
3.5. IR of Four Monomers and Their Mixture to HMGR.
According to the ratio of the four monomers in crude extract of
hawthorn fruit, quercetin (a), hyperoside (b), rutin (c), and
chlorogenic acid (d) were adjusted to concentrations of 0.16,
0.32, 1.42, and 0.95 mg/mL with PPB (pH 6.8), respectively. Then
their IR to the activity of HMGR was analyzed by HPLC. The IR
values of compounds a-d at 1 mg/mL were 6.28, 9.64, 23.53, and
10.56%, respectively. Their total IR was about 50.01% (Table3).
According to their ratio in hawthorn fruit extract, the mixture of
compounds a-d was matched with the concentration of 2.85 mg/
mL. The IR of the matched mixture was up to 79.5%, much
higher than that of the sum of four monomers.
Figure 1. Structures of compounds a-d: (a) quercetin; (b) hyperoside;
(c) rutin; (d) chlorogenic acid.
period. There were no significant differences in the body weights
among different groups during treatment, suggesting the mixture
or compounds a-d was safe and well tolerated in the mice.
The mice were fed a high-fat and -cholesterol (HFHC) diet for
2 months. Compared to the normal diet group, the levels of TC,
TG, and LDL-C increased (p < 0.01) and that of HDL-C
decreased (p < 0.05) in the hyperlipidemic group. Then, hyperli-
pidemic mice were treated with compounds a-d or matched
mixture from hawthorn fruit orally for 6 weeks or left untreated.
The levels of TC, TG, and LDL-C in therapy groups decreased in
different degrees and that of HDL-C increased in the blood of the
hyperlipidemic mice (Table 4). Among these monomers, querce-
tin (a) showed the highest lipid-lowering effect (p < 0.05). The IR
values of hyperoside, rutin, and chlorogenic acid decreased in
order, which was similar to the result of their inhibitory activity to
HMGR. In general, the lipid-lowering efficacy of compounds
a-d was not perfect; the content of TC, TG, and LDL-C showed
no difference from that of the hyperlipidemic control. However,
the mixture of compounds a-d could significantly reduce the
levels of TC, TG, and LDL-C by 46.5, 49.6, and 58.1%,
respectively (p < 0.01). This confirmed that there was a positive
synergetic effect on lowering lipid among these four compounds,
similar to that of HMGR.
3.7. Structure-Activity Relationship of Compounds a-c to
HMGR and Lipid-Lowering Efficacy. As seen in the structures of
compounds a-c in Figure 1, they contained a quercetin glycone.
Hyperoside (b) is a monoglycoside of compound a, linked with a
β-
glycoside of compound a, linked with a disaccharide of 6-O-
rhamnosyl- -glucose at the 3-OH of quercetin. In general, the
D-galatose on the 3-OH of quercetin. Rutin (c) is a disaccharide
L-
D
introduction of the glycosyl group increased the hydrophilic
ability of a compound. The greater the number of the glycosyl
group is, the stronger is the hydrophilic ability. From the data of
Table 5, the IR values of quercetin, hyperoside, and rutin to
HMGR were 39.6, 30.4, and 16.5%, respectively, which gradually
decreased with the increase of glycosyl group number. Similarly,
the efficacy of quercetin, hyperoside, and rutin in lowering TC,
TG, and LDL-C was decreased with the increase of glycosyl
group number (Table 4). Quercetin (a), without a glycosyl group,
showed the highest inhibitory activity and lipid-lowering effect.
This indicated that the activities of inhibition to HMGR and
lowering lipid were related to its hydrophilic ability. The weak
3.6. Effects of Four Monomers and Their Mixture on Lipid-
Lowering Efficacy in Mice. As mixture of compounds a-d
showed an improved IR to HMGR in vitro, the effects of four
monomers and their mixture on lipid-lowering efficacy in vivo
were investigated.
After a 6 week treatment, all mice showed good health status,
and no mortality was recorded during the whole experimental

Article
J. Agric. Food Chem., Vol. 58, No. 5, 2010 3137
Figure 2. HPLC chromatogram of ethanol extract of hawtorn fruit: (A) chromatogram of standard compounds a-d; (B) chromatogram of sample (1,
chlorogenic acid; 2, rutin; 3, hyperoside; 4, quercetin). Symmetry Hypersil C18-BDS (4.6 ꢀ 250 mm, 5 μm) was used and was eluted with a linear gradient of
acetonitrile in water from 10 to 50% (containing 0.3% KH2PO4) over 30 min before returning to initial conditions for 10 min. The flow rate was 1 mL/min, and
effluent was monitored at 360 nm.
Table 3. Inhibition Rate of Compounds from Hawthorn Fruit to HMG-CoA
Reductase
EtOH extract of hawthorn fruit by systematic chromatographic
methods using HMGR as the target to obtain and identify the
active compounds with high inhibitory activity to HMGR. The
lipid-lowering efficacy of these components was verified by
animal experiments.
chlorogenic
quercetin hyperoside rutin
acid
1.42 0.95
23.53 10.56
totala mixtureb
2.85 2.85
concentration (mg/ 0.16
mL)
0.32
From Table 1, the IR and the total inhibitory activity of crude
extract with EtOH were about 5.1% and 34680, respectively.
After being purified with D-101 macroporous adsorption resin,
four main active fractions (fractions A, F, G, and H) were obtained.
Their total weight was 128.8 g, and the total inhibitory activity was
33225.6, which was near that of crude extract. After further
purification with different column chromatography to these four
active fractions, we observed that the loss of weight was about 30%,
but the loss of inhibitory activity was up to 70% (Table 2). This
indicated that further purification to crude extract weakened the
inhibitory activity of active components to HMGR, and it was
speculated that there was a synergetic effect among the active
components. From the chromatographic fractions, four com-
pounds were obtained by recrystallization in MeOH and identified
as quercetin (a), hyperoside (b), rutin (c), and chlorogenic acid (d),
which indicated that flavonoids were the active ingredients with a
high inhibitory activity to HMGR in hawthorn fruit.
To further investigate the synergetic effect on the HMGR
activity, we compared the IR values of a, b, c, and d and their
mixture matched according to the ratio in crude extract. It was
discovered that the IR of the mixture to HMGR was up to 79.5%,
much higher than that of any of the single compounds and the
total IR of their sum (50.01%). In addition, ursolic acid, oleanolic
acid, and maslinic did not inhibit the activity of HMGR (data not
shown). Although triterpenes, especially ursolic acid, were re-
ported as the main active hypolipidemic constituents, the present
results indicate triterpenes could lower lipid (15) not by the
mechanism of inhibiting HMGR activity.
IR (%)
6.28
9.64
50.01 79.5
a Total indicated the sum of compounds a-d. b Mixture indicated compounds
a-d were matched according to the ratio of compounds a-d in hawthorn fruit
extract, which was 0.16:0.32:1.42:0.95, respectively. IR indicated the inhibitory rate
at the corresponding concentration in Table 3.
hydrophilic ability benefited the inhibition to HMGR and lipid-
lowering efficacy.
4. DISCUSSION
C. pimmatifida Bge. (Rosaceae) is widely used as a food and
traditional medicine in the treatment of chronic heart failure, high
blood pressure, arrhythmia, and various digestive ailments (9), as
well as geriatric and antiarteriosclerosis remedies(11). It has
been documented that it decreased the serum total cholesterol,
LDL-C, and triglyceride in hyperlipidemic humans (12). In
hawthorn fruit, the active ingredients mainly contained flavo-
nols (quercetin, hyperoside, rutin, and vetixin), flavanols
(epicatechin), phenylpropanoids (chlorogenic acid), triterpenes
(ursolic acid, oleanolic acid, and maslinic acid) and other in-
gredients. It was recorded that flavonoids and triterpenes were the
active ingredients of hypolipidemic (15), but it is not clear whether
these ingredients can inhibit the HMGR activity and lower the
lipid level in blood of mice.
To confirm the hypolipidemic activity and mechanism of
action, the inhibitory activity of various fractions from hawthorn
fruit to HMGR was traced. We separated and purified the

3138 J. Agric. Food Chem., Vol. 58, No. 5, 2010
Ye et al.
Table 4. Lipid-Lowering Efficacy of Compounds a-d and Mixture (n = 10, Mean ( SD)
group
dose (mg kg-1
)
TCa (mmol L-1
)
TGa (mmol L-1
)
LDL-Ca (mmol L-1
)
HDL-Ca (mmol L-1
)
normal control
hyperlipidemic control
compound a
compound b
compound c
compound d
mixtureb
2.40 ( 0.25
6.58 ( 0.53##
4.12 ( 0.65*
5.13 ( 0.67
5.88 ( 0.64
6.15 ( 0.60
3.52 ( 0.18**
0.79 ( 0.11
3.04 ( 0.41##
1.66 ( 0.33*
1.83 ( 0.57
2.54 ( 0.48
2.87 ( 0.60
1.53 ( 0.32**
1.49 ( 0.08
4.59 ( 0.74##
2.17 ( 0.30
2.56 ( 0.56
3.28 ( 0.34
3.77 ( 0.44
1.92 ( 0.25**
1.56 ( 0.10
1.34 ( 0.13#
1.47 ( 0.23
1.38 ( 0.26
1.40 ( 0.34
1.39 ( 0.14
1.51 ( 0.12
2.85
2.85
2.85
2.85
2.85
a #, p < 0.05, ##, p < 0.01 versus normal control group; *, p < 0.05, **, p < 0.01 versus hyperlipidemic control group. b The ratio of compounds a-d in the mixture was
0.16:0.32:1.42:0.95.
(5) Gao, F.; Linhartova, L.; Johnston, A. McD.; Thickett, D. R. Statins
Table 5. IR of Compounds a-c to HMG-CoA Reductase
and sepsis. Br. J Anaesth. 2008, 100 (3), 288–298.
quercetin
hyperoside
rutin
(6) Endres, M.; Laufs, U.; Huang, Z.; Nakamura, T.; Huang, P.;
Moskowitz, M. A.; Liao, J. K. Stroke protection by 3-hydroxy-3-
methylglutaryl (HMG)-CoA reductase inhibitors mediated by en-
dothelial nitric oxide synthase. Proc. Natl. Acad. Sci. 1998, 95,
8880-8885.
concentration (mg/mL)
IRa (%)
1
39.6
1
30.4
1
27.1
a IR indicates the inhibitory rate to HMGR at the concentration of 1 mg/mL.
(7) Graham, D. J.; Staffa, J. A.; Shatin, D. Incidence of hospitalized
rhabdomyolysis in patients treated with lipid-lowering drugs.
JAMA-J. Am. Med. Assoc. 2004, 292 (21), 2585–2590.
(8) Yu, G.; Cao, X. G.; Ye, X. L.; Li, X. G.; Chen, Z.; Yuan, L. J.; Wang,
L. J. Determination of 3-hydroxy-3-methylglutaryl coenzyme A
reductase inhibitor activity by high performance liquid chromatog-
raphy. Chinese J. Anal. Chem. 2009, 37, 87–90.
(9) Chu, C. Y.; Lee, M. J.; Liao, C. L.; Lin, W. L.; Yin, Y. F.; Tseng,
T. H. Inhibitory effect of hot-water extract from dried fruit of
Crataegus pinnatifida on low-density lipoprotein (LDL) oxidation
in cell and cell-free systems. J. Agric. Food Chem. 2003, 51, 7583–
7588.
The result of animal tests in vivo further confirmed that the
lipid-lowering efficacy of the matched mixture was better than
that of every monomer and showed a significant difference
(p < 0.01). Therefore, compounds a-d had not only the
inhibitory activity to HMGR in vitro but also the lipid-lowering
efficacy in vivo. The better lipid-lowering activity of compounds
a-d was realized by their synergetic inhibition to HMGR.
Structure-activity relationship study revealed that the inhibi-
tory activities to HMGR and lipid-lowering efficacy were related
with the glycosyl numbers in compounds a-c. Quercetin without
a glycosyl group showed the highest activity, then hyperoside with
a monoglycoside. Therefore, the weak hydrophilic ability benefits
the inhibition to HMGR and lipid-lowering efficacy.
€
(10) Al Makdessi, S.; Sweidan, H.; Mullner, S.; Jocob, R. Myocardial
protection by pretreatment with Crataegus oxyacantha: an assess-
ment by means of the release of lactate dehydrogenase by the
ischemic and reperfused Langendorff heart. Arzneim.-Forsch. 1996,
46, 25–27.
ABBREVIATIONS USED
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
(11) Tadic, V. M.; Dobric, S.; Markovic, G. M.; Dordevic, S. M.; Arsic,
HMGR, hydroxy-3-methylglutaryl coenzyme A reductase; TC,
total cholesterol; LDL-C, low-density lipoprotein cholesterol;
TG, triglyceride; HDL-C, high-density lipoprotein-cholesterol;
HFHC, high fat and high cholesterol; NADPH, nicotin-
amide adenine dinucleotide phosphate; DTT, 1,4-dithiothreitol;
EDTA, ethylenediaminetetraacetic acid; SAR, structure-activity
relationship.
ꢀ
ꢀ
I. A.; Menkovic, N. R.; Stevic, T. Anti-inflammatory, gastroprotec-
tive, free-radical-scavenging, and antimicrobial activities of haw-
thorn berries ethanol extract. J. Agric. Food Chem. 2008, 56,
7700-7709.
(12) Chen, J. D.; Wu, Y. Z.; Tao, Z. L.; Chen, Z. M.; Liu, X. P. Hawthorn
(shan zha) drink and its lowering effect on blood lipid levels in
humans and rats. World Rev. Nutr. Diet. 1995, 77, 147–154.
(13) Zhang, Z. S.; Ho, W. K. K.; Huang, Y.; Chen, Z. Y. Hypocholester-
olemic activity of hawthorn fruit is mediated by regulation of
cholesterol-7a-hydroxylase and acyl CoA: cholesterol acyltransfer-
ase. Food Res. Int. 2002, 35, 885–891.
(14) Azuma, K.; Nakayama, M.; Ito, H.; Higashio, H.; Terao, J.;
Ippoushi, K. Absorption of chlorogenic acid and caffeic acid in
rats after oral administration. J. Agric. Food Chem. 2000, 48, 5496–
5500.
LITERATURE CITED
(1) Balbisi, E. A. Management of hyperlipidemia: new LDL-C targets
for persons at high-risk for cardiovascular events. Med. Sci. Monit.
2006, 12, RA34-RA39.
(2) Stamler, J.; Daviglus, M. L.; Garside, D. B.; Dyer, A. R.; Greenland,
P.; Neaton, J. D. Relationship of baseline serum cholesterol levels in
3 large cohorts of younger men to long-term coronary, cardiovas-
cular, and all-cause mortality and to longevity. JAMA-J. Am. Med.
Assoc. 2000, 284, 311–318.
(3) Grundy, S. M.; Cleeman, J. I.; Merz, C. N.; Brewer, H. B.; Clark, L.
T.; Hunninghake, D. B.; Pasternak, R. C.; Smith, S. C.; Stone, N. J.
Implications of recent clinical trials for the National Cholesterol
Education Program Adult Treatment Panel III guidelines. Circula-
tion 2004, 110, 227–239.
(15) Min, S. W.; Jung, S. H.; Cho, K. H.; Kim, D. H. Antihyperlipidemic
effects of red ginseng, Crataegii gructus and their main constituents
ginsenoside Rg3 and utrsolic scid in mice. Biomol. Ther. 2008, 16,
364–369.
Received for review August 19, 2009. Accepted January 22, 2010. This
work was supported by the Major Technologies R&D Program of CQ
CSTC (2008AA5021) and Chongqing Municipal Health Bureau (Yu
(2008)1-2).
(4) Goldstein, J. L.; Brown, M. S. Regulation of the mevalonate path-
way. Nature 1990, 343, 425–430.