Available via license: CC BY-NC-ND 4.0
Content may be subject to copyright.
Journal Pre-proof
Extraction of Artemisinin with Hydroxypropyl-β-Cyclodextrin Aqueous Solution for
Fabrication of Drinkable Extract
Yongqiang Zhang, Yingying Cao, Xiangzhan Meng, Phonphat Prawang, Hui Wang
PII: S2666-9528(20)30007-8
DOI: https://doi.org/10.1016/j.gce.2020.09.007
Reference: GCE 7
To appear in: Green Chemical Engineering
Please cite this article as: , https://doi.org/10.1016/j.gce.2020.09.007.
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition
of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of
record. This version will undergo additional copyediting, typesetting and review before it is published
in its final form, but we are providing this version to give early visibility of the article. Please note that,
during the production process, errors may be discovered which could affect the content, and all legal
disclaimers that apply to the journal pertain.
© 2020 Institute of Process Engineering, Chinese Academy of Sciences. Publishing services by Elsevier
B.V. on behalf of KeAi Communication Co. Ltd.
1
A green and sustainable process for extraction of artemisinin from Artemisia annua L.
using hydroxypropyl-β-cyclodextrin aqueous solution was developed.
Journal Pre-proof
1
Extraction of Artemisinin with Hydroxypropyl-
β-Cyclodextrin
Aqueous Solution for Fabrication of Drinkable Extract
Yongqiang Zhang
a,b
, Yingying Cao
a
, Xiangzhan Meng
a
, Phonphat Prawang
a
,
Hui Wang
a,b,∗
a
Beijing Key Laboratory of Ionic Liquids Clean Process, CAS Key Laboratory of Green Process and
Engineering, State Key Laboratory of Multiphase Complex Systems, Institute of Process Engineering,
Innovation Academy for Green Manufacture, Chinese Academy of Sciences, No. 1 Beierjie, Zhongguancun,
Haidian District, Beijing, 100190 P.R. China;
b
Sino-Danish College, University of Chinese Academy of Sciences, Yanqihu Campus, No. 380
Huaibeizhuang, Huairou District, Beijing, 101408 P.R. China
Abstract
A process for extraction of artemisinin (Ars) from Artemisia annua L. based on
aqueous solution of cyclodextrin derivative under ultrasonic irradiation was proposed.
Among the test additives, the system containing edible hydroxypropyl-β-cyclodextrin
(HPBCD) was shown to be efficient for this extraction. The optimal extraction conditions
were predicted by artificial neural network, under which the extraction amount of Ars
could reach 8.66 mg per gram of the leaves, much higher than the value in water (1.70
mg/g). Kinetics study indicated that HPBCD enhanced the equilibrium
amount of Ars in
water, and ultrasonic irradiation accelerated extraction rate and reduced the activation
energy. Mechanism study showed that the formation of Ars-HPBCD complex during
extraction enhanced the water solubility of Ars and facilitated the extraction, and the
configuration of Ars-HPBCD complex was modeled by molecular docking, which was
further verified by characterizations. This work provides a convenient way to fabricate
the traditional tea-based medicament of natural drugs to make production of this famous
antimalarial medicine cheaper.
1. Introduction
Artemisinin (Ars) is a famous drug mainly used for treatment of the life-threaten
∗ Corresponding author: E-mail address: huiwang@ipe.ac.cn (H. Wang)
Journal Pre-proof
2
disease malaria. Compared with other antimalarial drug like quinine, Ars has remarkable
advantages such as no toxicity and strong ability against multidrug-resistant malaria
parasites
[1]
. Ars has already saved millions of people’s lives, and the Chinese scientist
who discovered Ars to treat parasitic disease, Youyou Tu, won the 2015 Nobel Prize in
Physiology or Medicine
[2]
.
Ars is usually produced by treating Artemisia annua L. (Qinghao) with volatile and
explosive organic solvents like petroleum ether, followed by purification with
crystallization or chromatographic column separation
[3]
. The use of large amounts of
organic solvents to produce Ars would cause environment pollution and have the risk of
explosion. Moreover, Ars is a poorly water soluble drug (0.84 mg/mL, 25
o
C
[4]
), and in
order to achieve a better utilization and easier absorption by human body, it is generally
transformed into derivatives to produce higher water soluble products, such as
artemether-lumefantrine (AL), artesunate-amodiaquine (ASAQ),
dihydroartemisinin-piperaquine (DHA-PPQ) by semi-synthesis for clinical use
[5]
. The
derivatization steps would definitely increase the product cost, making Ars-based drugs
not accessible for the majority of malaria victims, especially in the rare and undeveloped
Africa. Therefore, development of a sustainable, simple, and cheap way to get Ars is
important to make it affordable to more people.
In China, a method of using Qinghao was documented in an ancient medical book
titled “Zhou Hou Bei Ji Fang”
[6]
, which was written approximately 1700 years ago for
treatment of fever by a famous alchemist of Jin Dynasty, Ge Hong (284-363 A.D.). In the
book, it was mentioned that “take one bunch of fresh Qinghao, soak it with 0.4 L water,
wring out the juice and drink it”. This method provides an alternative choice to extract
Ars from the plant leaves using water for the fabrication of the traditional “tea” based
medicament. However, Ars is a poorly water-soluble drug with an octanol-water partition
coefficient (logP) higher than 2, and the extraction amount was very low when using
water as the extractant
[7]
. Besides, Ars is a thermal sensitive compound and relatively
high temperatures (e.g., 70 ºC) could cause its decomposition
[8]
, and thus Artemisia
annua L. could not be boiled in hot water to obtain Ars like the treatment of other
traditional Chinese medicine plants.
Cyclodextrins (CDs) is a kind of well-known supramolecular compounds consisting
Journal Pre-proof
3
of α-(1,4)-linked glucose units. CDs and its hydrophilic derivatives (such as α-CDs,
β-CDs, γ-CDs, hydroxypropyl-, carboxymethyl-, sulfobutyl ether-α, β-, γ-CDs, etc.) are
widely used in food, cosmetic, and pharmaceutical industries due to their special
“internal-external” affinity, which could be contributed to the hydrophilic external groups
and hydrophobic internal cavity of these molecules
[9]
. CD derivatives are considered as
promising solubilizers and pharmaceutical excipients because they can combine with
some poorly soluble drug molecules to form water-soluble inclusion complexes to
increase the solubility of these drugs
[10-12]
. They can also be used to improve the
extraction efficiency of some Chinese herbal medicinal components, such as
camphorquinone
[13]
, mulberry anthocyanin
[14]
and flavonoids
[15]
. In addition, CDs are
starch derivatives, so they are edible, safe and not toxic, and thus it is more appropriate to
be used as the solubilizer than traditional organic compounds.
Ultrasonic assisted extraction (UAE) has been widely used in natural product
extraction to intensify the process
[16-19]
because the ultrasonic cavitation effect can
significantly enhance the mass transfer and decrease the extraction temperature to protect
some thermal sensitive compounds like Ars
[20]
. The above information promoted us to
develop a process for Ars extraction using aqueous solutions of CD derivatives under
ultrasonic irradiation. Extraction efficiency of the aqueous systems containing various
CD derivatives was compared, and the conditions were optimized by single factor
experiments, orthogonal experiments, and artificial neural network (ANN). Molecular
simulation and characterizations including
1
H NMR, FTIR, DSC, and ESI-MS were used
to reveal the detailed mechanism of this extraction process. Kinetic study was also
investigated to further disclose the apparently improved efficiency of the proposed
system for Ars extraction
.
2. Materials and Methods
2.1 Materials and chemicals
Artemisia annua L. (Qinghao) dry herbal plant was kindly provided by Tianyuan
Biotechnology Co. Ltd. (Henan, China) and used as pulverized powder sieved with a
40-mesh sieve. Ars standard with a purity of 98% was purchased from Aladdin Industrial
Corporation (Shanghai, China). All kinds of CD derivatives (99%), i.e., β-CD, γ-CD,
Journal Pre-proof
4
hydroxypropyl-β-CD (HPBCD), sulfobutyl ether-β-CD (SEBCD),
methyl-β-CD (MBCD),
hydroxypropyl-γ-CD (HPGCD), were gained from Aladdin and used as received. Water
used in the experiments was prepared by Milli-Q reverse osmosis system (Bedford,
USA).
The Ars-HPBCD complex was prepared by evaporation of water in the Ars saturated
HPBCD aqueous solution (saturated concentration of Ars in 8 wt% HPBCD aqueous
solution was determined to be 12.96 mg/mL at 25
o
C). The physical mixture of Ars and
HPBCD was fabricated by mixing these two compounds at the same ratio. Briefly, 129.6
mg Ars and 0.8 g HPBCD were ground in a mortar to get homogeneously dispersed
mixture.
2.2 Ultrasound Apparatus
Ultrasonic assisted extraction experiments were carried out in an ultrasonic water
bath (KQ-250DB, Kunshan, China), the temperature of which could be maintained at a
certain value with an error range of ± 0.4 ºC by a thermal couple. The apparatus was a
rectangular container with a size of 23.5×13.3×10.2 cm, and a 20 kHz, 0-250 W
ultrasonic generator was annealed at the bottom. 2 L water was used for water bath
heating.
2.3 Ultrasound assisted extraction (UAE) of Ars
The UAE of Ars was carried out by putting 1.0 g dried leaves powder into a 50 mL
round bottom flask with CD derivative aqueous solution, and the flask was then sealed
with a glass stopper. The mixture was sonicated at the center of the water bath under a
certain power and temperature for a given time. After extraction, the mixture was filtrated
and the liquid phase was collected and measured by HPLC.
2.4 Quantification of Ars
Quantification of Ars in the solutions was performed on HPLC (LC-20AT, Shimadzu,
Japan) with a C18 column and UV detector. Acetonitrile with a flow rate of 0.6 mL/min
and water with a flow rate of 0.4 mL/min were used as the mobile phase. The analysis
process lasted for 20 min and chromatographs were collected under the UV wavelength
of 213 nm.
2.5 Optimization by orthogonal experiments
Journal Pre-proof
5
Orthogonal experiment is a typical method for multi-factor optimization and widely
applied in extraction processes
[21-23]
. For the optimization with 5 factors, a L
16
(4
5
)
orthogonal experiment design was established by Minitab 17.1 software based on the
boundaries of each parameter obtained in the single-factor optimization experiments. The
orthogonal experiment was shown in Table 1, and each experiment was repeated for 3
times.
2.6 Artificial neural network (ANN) model for predicting extraction results
ANN is a model of algorithm based on deep neural network, which is similar to a
nervous system of human brain and can be used to solve many complicate problems like
clustering, pattern classification, multi-factor modelling and prediction
[24]
.
Back-propagation (BP) as a typical ANN is a computing model for nonlinear multivariate
calculation and modelling, capable of predicting the results at certain conditions based on
the training of experimental data
[25-26]
. BP-ANN consists of input signal (experimental
conditions), hidden layer, output layer and output signal (results) as shown in Figure 1.
BP-ANN model can provide precise predicted results after constant training. The data
processing procedure of BP-ANN is illustrated in Figure 2.
Figure 1. A typical model of BP-ANN.
Input signal
Input layer
Hidden layer
Output layer
Output signal
Input original data
Construct a BP-
ANN model
Initialize the BP-ANN
Train the BP-ANN
Meet the accuracy
End the training
Y
Input new parameters
Calculate the results
Predict the results
Construction of BP-ANN
Training of BP-ANN
Prediction of results
N
Journal Pre-proof
6
Figure 2. Flowsheet of BP-ANN data processing.
In this study, BP-ANN model was established by IBM SPSS modeler 14.1, an ANN
module of this software used to construct a BP-ANN, and the results obtained from the
orthogonal experiments were used as the original and training data. Multi-layer
perceptron was set as the method of this model, and when both of the accuracy of total
model and prediction were higher than 99%, the modelling would be stopped. After
successful training, a predicted model with certain accuracy was generated. In order to
get more new parameters and improve the accuracy of the predicted results, a total factor
experiment with 1024 (4
5
) sets of conditions were designed. These conditions were put
into the BP-ANN model and the results were predicted. These results were then selected
as the optimal extraction conditions after sorting with Microsoft Excel (Microsoft 2016)
according to the extraction amounts.
2.7 Kinetics
In order to further investigate the influence of CD derivative and ultrasound on the
extraction, the kinetics study was performed under the following conditions: (1)
ultrasonic assisted extraction with CD derivative (8.0 wt% HPBCD aqueous solution, 20
mL/g for liquid to solid ratio, 40 min for time and 100 W for ultrasonic power); (2)
ultrasonic assisted extraction with pure water (20 mL/g for liquid to solid ratio, 40 min
for time and 100 W for ultrasonic power); (3) extraction with CD derivative without
ultrasound (8.0 wt% HPBCD aqueous solution, 20 mL/g for liquid to solid ratio, and 40
min for extraction time). The experiments were carried out at 30
o
C and 40
o
C under the
conditions mentioned above, and time gradient was set as 5 min. As solid-liquid
extraction can be regarded as the reverse of liquid-solid adsorption
[27-28]
, these two
processes share the same principle and kinetics. Thus, the liquid-solid adsorption kinetic
equations are applicable for solid-liquid extraction, and the second order kinetic model
can also be used to fit the extraction of natural products. The second order equation was
listed as below:
2
( )
te t
dC
k C C
dt
= −
(1)
where C
t
(mg/mL) is the concentration at time t (min), C
e
(mg/mL) is the equilibrium
Journal Pre-proof
7
concentration, and k is the second order rate constant (mg/(g·min)).
Equation (1) can be integrated from 0 to t, and in this case C
t
could be expressed as
equation (2), which can be transformed to a linearized form as equation (3):
2
1+
e
t
e
k t C
C
k t C
⋅ ⋅
=
⋅ ⋅
(2)
2
1
t e e
t t
C C k C
= +
⋅
(3)
Plotting t/C
t
as a function of t could give the values of C
e
and k by calculating the
slope and intercept of the fitted line.
Arrhenius equation (4) is widely used to describe the relationship between
temperature (T) and k, and the extraction activation energy (E
a
) can then be calculated
from equation (4)
[29-30]
.
0
exp( )
Ea
k k
RT
= −
(4)
Equation (4) can be converted into a linearized equation (5) by taking logarithm of
both sides of the equation:
0
Ea
Lnk Lnk
RT
= −
(5)
By plotting Lnk as a function of 1/T, the values of k
0
and E
a
can be obtained, where
k
0
(mg/(g·min)) is a temperature independent factor.
2.8 Characterization
2.8.1 FTIR
FTIR spectra of pure HPBCD, the physical mixture of HPBCD and Ars, and
complex of Ars-HPBCD were collected on a FTIR Spectrometer (Nicolet560, USA) by
using KBr disc method. The sample was each compressed into a transparent tablet and
scanned from 400 cm
-1
to 4000 cm
-1
with an average of 32 scans.
2.8.2
1
H NMR
1
H NMR spectra were recorded on a Bruker Avance III HD 600 spectrometer
(Germany) at 600 MHz. Neat Ars was dissolved in D
2
O directly, while the Ars-HPBCD
complex sample for analysis was fabricated by dissolving an excess amount of Ars into
Journal Pre-proof
8
HPBCD-D
2
O solution, which was then filtered through a 0.22 µm
filter membrane to get
the liquid for test.
2.8.3 ESI-MS
The molecular weight of the Ars-HPBCD complex was recorded on Bruker QTOF
Electrospray Ionization Mass Spectrometer (Germany) with ultraclean water as the
solvent.
2.8.4 DSC
Melting points of the physical mixture of Ars and HPBCD as well as the
Ars-HPBCD complex were obtained by differential scanning calorimetry (Mettler-Toledo
DSC 1, Switzerland). The powder of each sample was sealed in an aluminum hermetic
pan, which was then heated from 25
o
C to 200
o
C with a heating rate of 10
o
C/min.
2.8.5 SEM
The morphologies of the leaves particles treated under different conditions were
observed by Hitachi SU8020 scanning electron microscope (Japan). The particles were
firstly sputtered on the conductive tape and then were coated with platinum layer under
the coating conditions of 25 mA and 60 s.
2.9 Molecular docking simulation
AutoDock is an open source molecular simulation software developed by Olson’s
laboratory, Scripps Research Institute (USA)
[31]
, which is mainly used for ligand-protein
docking
[32]
. This software is also widely used in the investigation of complexes formed
by CD derivative and guest drug molecules
[33-34]
. In this study, the structure of HPBCD
with four hydroxypropyl groups was downloaded from PubChem (U.S. national library
of medicine, CID of the compound: 56972821) and applied for molecular docking. Ars
molecule was generated through ChemDraw (Chemoffice Professional 15.1) and
optimized by MM2 method. AutoDock 4.2 with AutoDockTools (version 1.5.6) was
applied for molecular docking, and 100 independent dock runs were performed and the
genetic algorithm method was used for calculation. The other parameters were adopted
from this software’s default settings. The selection of clusters was based on their total
energy, and the configuration with the lowest energy was considered to be the most
probable configuration of the Ars-HPBCD complex.
2.10 The maximum extraction cycles
Journal Pre-proof
9
In order to determine how much Ars could be further extracted from the residual
leaves in the second, third, fourth cycles after the first extraction under the optimized
conditions, the residual leaves were mixed with 30 mL HPBCD aqueous solution.
Extraction followed procedures described in Section 2.4, and Ars concentration in the
solution was analyzed using HPLC. The same batch of leaves was extracted for 5 times.
3. Results and Discussion
3.1 Selection of CD derivatives
In this study, six kinds of CD derivatives, including β-CD, γ-CD, HPBCD, SEBCD,
MBCD, HPGCD (the molecular structures and information are shown in Figure S1 and
Table S1, Supporting Information (SI)), were used as the additive (1.0 wt%) to facilitate
the extraction of Ars from the leaves using water under ultrasonic irradiation, and the
efficiencies of these different systems are compared in Figure 3. The results
demonstrated that addition of each of the six CD derivatives could enhance the extraction
amount compared with neat water, and β-CD derivatives (i.e., β-CD, MBCD, HPBCD)
generally possessed higher capability to enhance the extraction efficiency than γ-CD
derivatives. This is because the solubilization effect of β-CD was higher than γ-CD as
reported by Usuda et al.
[35]
. HPBCD is a highly water-soluble compound (solubility ≥ 600
g/L at 25
o
C) and has no haemolytic side effects after being absorbed by human body.
Therefore, HPBCD, which exhibited desirable ability to enhance Ars extraction in water,
was selected as the additive of this extraction for further study.
Figure 3. Comparison of the extraction amounts of Ars by aqueous solutions with
different CD derivatives (extraction conditions: 1.0 wt% for concentration of CD
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
Neat water
Η
Η
ΗΗ
PGCD /H
2
O
γ
γ
γγ
-CD/ H
2
O
HPBCD/ H
2
O
MBCD/ H
2
O
SEBCD/ H
2
O
Extraction Amount (mg/g)
β
β
ββ
-CD/ H
2
O
Journal Pre-proof
10
derivative, 20 mL/g for liquid to solid ratio, 30
o
C for temperature, 20 min for time and
200 W for ultrasonic power).
3.2 Optimization by single-factor experiments
In this section, the effects of extraction parameters, including concentration of
HPBCD, liquid to solid ratio, temperature, time, and ultrasonic power, on the efficiency
were optimized by single-factor experiments and the results were shown in Figure 4. It
can be seen that the extraction amount of Ars increased dramatically as the HPBCD
concentration raised from 1.0 wt% to 7.0 wt% and seemed to remain constant when
further increasing the HPBCD concentration (Figure 4a), so the range of the additive
concentration for further optimization was chosen as 6.0 wt% to 9.0 wt%.
The liquid to solid (i.e., the leaves) ratio is closely related to the contact area
between the solvent and the leaf particles, and thus is also an important factor to affect
the extraction amount. The results in Figure 4b indicated an obvious increasing trend for
extraction amount when the liquid to solid ratio was from 5 mL/g to 25 mL/g, and it
remained almost unchanged when the ratio was higher than 25 mL/g. Increasing the
liquid to solid ratio can make the solvent and leaves contact more effectively and could
also enhance the concentration difference between solid particles and the solution, thus
the extraction amount increased. After the liquid to solid ratio reached a certain value, the
leaves and solvent could contact effectively, resulting in no obvious change in extraction
amount of Ars when further raising the ratio. The boundary of liquid to solid ratio was
determined to be 20 mL/g to 35 mL/g.
Temperature is a major factor of an extraction process and the extraction amount of
Ars kept increasing with temperature from 20
o
C to 60
o
C, and there was no significant
difference in the extraction efficiencies obtained at 60
o
C and 70
o
C, as indicated in
Figure 4c. The temperature range for subsequent optimization was determined to be 40
o
C to 70
o
C. It usually takes some time for an extraction process to reach equilibrium, and
thus time is also a vital parameter to affect the extraction amount of Ars from the leaves.
In Figure 4d, the amount increased continually with time extending from 5 min to 30
min, with increasing tendency being somewhat retarded from 30 min to 40 min. The time
range was chosen as 25 min to 40 min for further optimization.
The effect of ultrasonic power on extraction efficiency is shown in Figure 4e. It can
Journal Pre-proof
11
be seen that the extraction amount of Ars increased from 100 to 150 W, and it slowly
decreased with the power further increasing. Ultrasonic irradiation could enhance the
transition of Ars from the solid particle to the liquid extractant because of the cavitation
effect, but higher ultrasonic power (e.g., ≥ 150 W) caused the extraction amount
decreased. This is because the ultrasound can produce a local high temperature
[36]
, which
can make the thermal sensitive artemisinin decomposed
[37]
. The boundary of ultrasonic
power was selected as 100 W to 175 W.
Figure 4. Effects of HPBCD concentration (a), liquid to solid ratio (b), temperature (c),
time (d), ultrasonic power (e) on extraction amounts of Ars.
3.3 Orthogonal experiments
Conditions for Ars extraction using aqueous solution of HPBCD were also
optimized by orthogonal experiments. The boundary of each factor was obtained by
single factor experiment and the levels were set according to L
16
(4
5
) in the orthogonal
table. Table 1 showed the conditions and the results of orthogonal experiments. These
data was further used for construction and training of ANN in Section 3.4.
Table 1. The results of orthogonal experiments.
No. Concen.
(wt%)
Ratio
(mL/g)
Time
(min)
Tempera
ture (
o
C)
Ultrasonic
power (W)
Extraction
amount
Predicted
values by
0 5 10 15 20 25 30 35 40 45
1.6
1.8
2.0
2.2
2.4
2.6
2.8
3.0
3.2
Extraction amount (mg/g)
Ratio (mL/g)
20 30 40 50 60 70
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
Extration amount (mg/ g)
Temperature (
o
C)
0 5 10 15 20 25 30 35 40 45
2.6
2.8
3.0
3.2
3.4
3.6
3.8
Extraction amount (mg/g)
Time (min)
100 120 140 160 180 200 220 240 260
1.0
1.5
2.0
2.5
3.0
3.5
Extraction amount (mg/g)
Ultrasonic power (W)
bc
de
a
0 2 4 6 8 10 12 14 16
2
3
4
5
6
7
8
Extraction amount (mg/g)
HPBCD concentrations
Journal Pre-proof
12
(mg/g)
ANN
1
6
20
25
40
100
4.53
4.54
2
6
25
30
50
125
5.95
5.88
3
6
30
35
60
150
6.70
6.72
4
6
35
40
70
175
8.28
8.33
5
7
20
30
60
175
6.77
6.78
6
7
25
25
70
150
6.74
6.73
7
7
30
40
40
125
5.28
5.31
8
7
35
35
50
100
7.30
7.32
9
8
20
35
70
125
5.41
5.42
10
8
25
25
60
100
6.02
6.07
11
8
30
40
50
175
5.91
5.93
12
8
35
30
40
150
4.05
4.04
13
9
20
40
50
120
6.58
6.55
14
9
25
35
40
175
5.93
5.95
15
9
30
30
70
100
5.43
5.42
16
9
35
25
60
125
4.50
4.49
3.4 ANN
ANN was used in this study to develop a model for results prediction of Ars
extraction under different conditions based on back-propagation and genetic algorithm.
Figure 5 (a) showed the relationship between training times and their accuracy, which
indicated that the accuracy increased dramatically during the first 30 trainings and it
reached 99.91% after training for 100 times with the data in Table 1. After successful
establishment and training of the ANN model, the data from experiments and prediction
was compared in Table 1 and Figure 5 (b), and R
2
of the fitted line was 0.99925, which
Journal Pre-proof
13
demonstrated that our established model had a very high prediction precision. The model
could also give the weight of these five factors in the ANN model, indicating their
importance. As can be seen from Figure 6, the factor importance decreased in the order
of: temperature > time > HPBCD concentration > ultrasonic power > liquid to solid ratio.
a b
Figure 5. The relationship between training times and accuracy (a), and the linear
relationship between experimental and predicted data (b).
Figure 6. The importance of the five factors in the ANN model.
If an experimental design with more conditions were put into this model, more
results could be predicted to find the best conditions. After obtaining the ANN model
with high accuracy, we designed experiments taking all the factors into consideration
with 1024 sets of conditions and the extracted amounts were predicted. Then these
0 10 20 30 40 50 60 70 80 90 100
30
40
50
60
70
80
90
100
Accuracy (%)
Training times
4 5 6 7 8 9
4
5
6
7
8
9
Predicted data
Experimental data
0.0
0.2
0.4
0.6
0.8
1.0
Ratio
Ultrasonic powerConcentration
Time
Factors
Temperature
Most unimportant
Most important
Journal Pre-proof
14
obtained results were sorted by Excel and the best conditions were shown to be: 8 wt%
for HPBCD concentration, 30 mL/g for liquid to solid ratio, 40 min for extraction time,
40
o
C for temperature, and 150 W for ultrasonic power with a predicted extraction
amount of 8.34 mg/g. Ars extraction was performed for five times under these optimal
conditions to check accuracy of the model and the experimental results were listed in
Table S2, SI. Under these conditions, the extraction amount was 8.66 mg/g, which was
quite close to the predicted value of 8.34 mg/g, but higher than the value (8.07 mg/g)
obtained in hydrothermal extraction with petroleum ether at 50
o
C. The standard
deviation (SD) of 0.55% illustrates that repetition of the experiments was good.
3.5 Kinetics
In order to get the extraction rate under different conditions, the extraction amounts
were plotted as a function of time. As shown in Figure 7, the extraction amount increased
over time and it didn’t change too much after a certain time. For instance, the UAE
extraction carried out in HPBCD/H
2
O at 40
o
C reached a balance at around 20 min, and
the equilibrium time for the water system at 40
o
C or 30
o
C was 10 min. The equilibrium
time for the processes without ultrasound irradiation was around 20-25 min. The
extraction amounts in the processes under 40
o
C were higher than the corresponding
values at 30
o
C, and the addition of HPBCD to the extraction system was demonstrated to
be able to enhance the extraction amount (A vs. C; B vs. D in Figure 7). UAE is shown to
be able to accelerate this extraction process and enhance the equilibrium amount of
extraction by comparing A with E (or B with F) in Figure 7.
Figure 7. Extraction amount as a function of time under different conditions: (A: UAE
with HPBCD/H
2
O at 40
o
C; B: UAE with HPBCD/H
2
O at 30
o
C; C: UAE with water at
0 5 10 15 20 25 30 35 40
0
1
2
3
4
5
6
7
8
Extraction amount (mg/g)
Time (min)
A
B
C
D
E
F
Journal Pre-proof
15
40
o
C; D: UAE with water at 30
o
C; E: extraction with HPBCD/H
2
O at 40
o
C with
hydrothermal heating; F: extraction with HPBCD/H
2
O at 30
o
C with hydrothermal
heating).
The second order kinetics model was applied to describe the Ars extraction, which
could provide more insights into the UAE and HPBCD assisted process. The fitting of
second order kinetics is shown in Figure 8, and their parameters are presented in Table 2.
Figure 8 indicated that the second order kinetics model was reliable for fitting the
kinetics in this study as all the corresponding correlation coefficients are higher than
0.986. C
e
and k values in Table 2 increased as the temperature raised from 30
o
C to 40
o
C.
The much higher value of C
e
in ultrasonic assisted Ars extraction by HPBCD/H
2
O versus
that with neat water indicated that the presence of the additive increased the equilibrium
amount of Ars in water. UAE with water as the extractant had the highest extraction rate
(k), which might be attributed to lower equilibrium amount (C
e
), rapid mass transfer, and
cavitation effects of ultrasound. Moreover, extraction rate in water was faster than the
HPBCD/H
2
O process under UAE, which could be attributed to the inclusion process for
the formation of Ars-HPBCD complex (discussed later), resulting in a postponed
equilibrium point as some time was needed to form the complex. The energy that
ultrasound produced can make the Ars and HPBCD molecules well mixed and increase
the probability of collision of these two molecules to form the complex, thus C
e
in
HPBCD/H
2
O was significantly increased from 5.42 mg/g in hydrothermal heating to 8.88
mg/g under ultrasound. The HPBCD/H
2
O extraction process with hydrothermal heating
processed the highest E
a
(115.50 kJ/mol), while those for the extractions with
HPBCD/H
2
O and water under UAE were 32.97 kJ/mol and 24.58 kJ/mol, respectively.
Therefore, it was obvious that ultrasound can significantly decrease the apparent
activation energy. In addition, because of the inclusion process for the formation of
Ars-HPBCD complex, E
a
of the extraction with HPBCD was a little higher than that of
the water process under UAE.
Journal Pre-proof
16
Figure 8. Fitting of second order extraction kinetics under different conditions (A: UAE
with HPBCD/H
2
O at 40
o
C; B: UAE with HPBCD/H
2
O at 30
o
C; C: UAE with water at
40
o
C; D: UAE with water at 30
o
C; E: extraction with HPBCD/H
2
O at 40
o
C under
hydrothermal heating; F: extraction with HPBCD/H
2
O at 30
o
C under hydrothermal
heating).
Table 2. Parameters of extraction kinetics and activation energies under different
conditions.
* C
e
is the equilibrium extraction amount.
3.6 Molecular docking
AutoDock software was utilized for molecular docking, and a grid approximation
was used to position the atoms to make the energy calculation fast. Because of the grid
setup, the binding energy value was a close approximation of the real case. Here 50
Systems Temperature (
o
C) k (g/(mg·min)) C
e
(mg/g)* E
a
(kJ/mol)
UAE,
HPBCD/H
2
O
40 0.031 8.88 32.97
30 0.021 5.98
Hydrothermal
heating,
HPBCD/H
2
O
40 0.028 5.42 115.50
30 0.018 4.96
UAE,
Water
40 0.337 1.70 24.58
30 0.187 1.16
Journal Pre-proof
17
configurations were achieved and their estimated free energies of binding were shown in
Figure 9. The results suggested that configuration of 17 had the lowest binding energy
(-5.15 kcal/mol), which indicated that this configuration was a relatively stable
Ars-HPBCD complex.
Figure 9. The configurations and their estimated free energies of binding.
The most probable configuration was shown in Figure 10. The results suggested that
the hydrophobic part of Ars was oriented inside the cavity of HPBCD, whereas the more
polar part tended to be exposed to the bulk solvent (Figure 10b). The groups (lactone,
peroxide bond in the dashed blue circle) that positioned outside the wider opening of the
HPBCD cavity could form hydrogen bonds easily with the outer hydroxyls and
hydroxypropyl groups of HPBCD.
Figure 10. The most probable configuration of Ars-HPBCD complex (a) and group
positions of Ars in Ars-HPBCD complex (b).
3.7 Verification of Ars-HPBCD complex
The Ars-HPBCD complex was characterized by a series of spectroscopic techniques
0 3 6 9 12 15 18 21 24 27 30 33 36 39 42 45 48 51
-5.2
-5.0
-4.8
-4.6
-4.4
-4.2
Estimated free energ y of binding (kcal/mol)
Configuration numbers
Journal Pre-proof
18
including FTIR,
1
H NMR, ESI-MS, and DSC to verify the configuration. FTIR spectra of
neat Ars, HPBCD, their physical mixture and the complex were shown in Figure 11. The
characteristic peaks for symmetric and asymmetric vibrations of -CH
3
, -CH
2
- of Ars can
be observed in the spectra of Ars and the physical mixture of Ars and HPBCD at around
2800~2900 cm
-1
, while these peaks disappeared in the Ars-HPBCD complex because of
the inclusion by HPBCD. The peak for -C=O-O group (lactone bond, see Figure 10b) of
Ars can be clearly seen around 1740 cm
-1
in the spectra of both the physical mixture and
the Ars-HPBCD complex, which indicated that the lactone bond positioned outside of
HPBCD in the Ars-HPBCD complex. The wavenumber of hydroxyl and hydroxypropyl
of HPBCD shifted from 3420.41 cm
-1
in neat HPBCD to 3381.63 cm
-1
in the complex,
illustrating that hydrogen bonds were formed in the Ars-HPBCD complex
[20]
. The FTIR
results confirmed the configuration obtained in molecular docking.
Figure 11. FTIR spectra of Ars, HPBCD, their physical mixture and the Ars-HPBCD
complex.
DSC was another characterization technique to provide the evidence of Ars-HPBCD
complex formation by detecting the melting behavior
[38]
. In Figure 12, the peak around
100
o
C in the curve of the physical mixture of Ars and HPBCD was due to evaporation of
water and that around 156
o
C belonged to the melting of Ars. Whereas the second peak
disappeared in the DSC curve of the Ars-HPBCD complex, indicating that a new
compound was formed between Ars and HPBCD, which was not a simple physical
mixture of them.
4000 3500 3000 2500 2000 1500 1000 500
Ars-HPBCD complex
HPBCD
Physical mixing
Ars
Absorptance
Wavenumber (cm
-1
)
Journal Pre-proof
19
Figure 12. DSC curves of the physical mixture and Ars-HPBCD complex.
NMR is also an important tool to verify the structure of Ars-HPBCD complex
[39]
.
The formation of the complex generally results in proton downfield shifts because of the
anisotropic ring current effect
[40]
. The
1
H NMR spectra of neat Ars and Ars-HPBCD
complex were shown in Figure S2, SI and the chemical shifts of the typical protons were
recorded in Figure 10b and Table 3 (other protons were demonstrated to have no obvious
shifts). The downfield shifts of H-a, H-b, H-c after forming the complex indicated that
this part was inserted in the cavity of HPBCD, and chemical shift of H-d appeared in a
higher field because of the formation of hydrogen bonds of lactone in Ars with the
outside hydroxyl and hydroxypropyl groups in HPBCD. These results provided further
evidence that the configuration obtained in molecular docking was reasonable.
Table 3.
1
H NMR chemical shifts of typical protons in Ars and Ars-HPBCD complex.
ESI-MS was used to analyze the molecular weights of the Ars-HPBCD complex (the
molecular weight of Ars is 282.3, and HPBCD is a mixture of β-CDs with different
numbers of hydroxypropyl group). The spectra were shown in Figure S3, SI, and the
molecular weights were recorded in Table 4. Peaks of 6-, 7-, 8-, 9-[HPBCD+Na]
+
(each
Recorded H Chemical shift (ppm) ∆δ (ppm)
Neat Ars Ars-HPBCD complex
H-a 0.934 1.031 + 0.097
H-b 1.137 1.236 + 0.099
H-c 1.731 1.811 + 0.080
H-d 6.171 6.092 - 0.079
Journal Pre-proof
20
number stands for the number of hydroxypropyl group in β-CD) were detected. The
difference of molecular weights between HPBCD and their complex was around 282.3
(one Ars molecule), which indicated that the Ars-HPBCD complex was a combination of
Ars and HPBCD with a ratio of 1:1. HPBCD and Ars-HPBCD complex with more
hydroxypropyl groups showed signals with high intensity because more hydroxypropyl
groups can enhance the water solubility of Ars-HPBCD complex since water was used as
the solvent in ESI-MS analysis.
Table 4. The molecular weights of [HPBCD+Na]
+
and their complexes.
Cations Numbers of
hydroxypropyl HPBCD
(Da) Complex
(Da) Difference
3-[HPBCD+Na]
+
3 1331.5 -
*
-
4-[HPBCD+Na]
+
4 1389.5 - -
5-[HPBCD+Na]
+
5 1447.5 - -
6-[HPBCD+Na]
+
6 1505.6 1787.9 282.3
7-[HPBCD+Na]
+
7 1563.6 1846.4 282.8
8-[HPBCD+Na]
+
8 1621.7 1903.7 282.0
9-[HPBCD+Na]
+
9 1679.7 1961.8 282.1
*: - indicates that the signal was very low and could be negligible.
SEM was used to investigate the effect of ultrasound on the morphology of the solid
leaf particles. SEM images of the leaves treated with thermal heating and UAE process
under the optimal conditions were shown in Figure 13. The surface of leaves in image (a)
was relatively smooth and intact after extraction under hydrothermal conditions, while
significant change occurred after extraction with UAE as the surface was broken into
debris, indicated by image (b). This phenomenon was also reported by other works
[41-43]
in UAE process, and generally disruption of the solid matrix was attributed to the
cavitation of the ultrasound, which can reduce the resistance of solid-liquid mass transfer,
thus decreasing the apparent activation energy as the kinetics study indicated.
a
b
Journal Pre-proof
21
Figure 13. SEM micrographs of solid leaf particle surface: (a) treated under
hydrothermal heating; (b) treated under UAE.
3.8
Application of drinkable Ars extract
There have been plenty of reports on the Ars extract based “tea”
[7, 44-45]
. The results
showed that the absorption rate of artemisinin in tea was faster than that in oral formula,
and the bioavailability is almost the same. Moreover, it has been reported that complexes
of Ars-CD derivatives such as hydroxypropyl-β-cyclodextrin, sulfobutyl ether-β-CD, etc.,
exhibited a higher Ars absorption rate and extent of bioavailability than the commercial
product Artemisinin 250
®[46]
. Therefore, this study provided a simple and efficient way to
obtain drinkable extract with high concentration of Ars.
4
Conclusions
An efficient method for extraction of Ars from
Artemisia annua L.
by using HPBCD
aqueous solution under ultrasound to fabricate a “tea” based medicament was developed.
The extraction conditions were optimized by single-factor optimization, orthogonal
experiment and ANN, and the extraction amount (8.66 mg/g) was much higher than the
corresponding value (1.70 mg/g) obtained in water. The kinetic study demonstrated that
this extraction process fitted well with the second order kinetic model. The presence of
HPBCD increased the equilibrium amount of Ars in water, and UAE accelerated the
extraction rate and reduced the activation energy. Mechanism study indicated that an
Ars-HPBCD complex with a 1:1 molar ratio was formed during the extraction process,
resulting in enhancement of the water solubility of Ars. The configuration of Ars-HPBCD
complex was calculated and modeled by molecular docking, which was confirmed by
FTIR,
1
H
NMR, ESI-MS and DSC. The efficient and simple method for artemisinin
extraction developed in this study led to the formation of “artemisinin tea”, which would
Journal Pre-proof
22
decrease the production cost of this drug and make this efficient antimalarial drug more
accessible to malarial victims from undeveloped area.
Acknowledgements
This work was financially supported by
Innovation Academy for Green Manufacture,
Chinese Academy of Sciences (No. IAGM-2019-A12), National Natural Science
Foundation of China (No. 21878311), CAS Pioneer Hundred Talents Program, and K. C.
Wong Education Foundation (No. GJTD-2018-04).
References
[1] Klayman, D. L., Qinghaosu (Artemisinin) - An antimalarial drug from China. Science
1985, 228 (4703), 1049-1055.
[2] The Noble Prize in Physiology or Medicine 2015; Nobel Foundation, Retrieved 7
October 2015
[3] Lapkin, A. A.; Plucinski, P. K.; Cutler, M., Comparative assessment of technologies for
extraction of artemisinin. J. Nat. Prod. 2006, 69 (11), 1653-1664.
[4] Xiao, D.; Yang, B.; Yang, X. M.; Yi, D.; Liao, X. L.; Yang, J.; Gao, C. Z., Synthesis of
Water Soluble Chitosan-Artemisinin Conjugate. Asian J. Chem. 2013, 25 (8), 4637-4639.
[5] WHO, Guidelines for the treatment of malaria. WHO: Geneva, Switzerland, 2015.
[6] Ge, H., Zhou Hou Bei Ji Fang. Shangwuyinshuguan: Taibei, China, 1983.
[7] Wright, C. W.; Linley, P. A.; Brun, R.; Wittlin, S.; Hsu, E., Ancient Chinese methods are
remarkably effective for the preparation of artemisinin-rich extracts of Qing Hao with potent
antimalarialactivity. Molecules 2010, 15 (2), 804-812.
[8] Sixt, M.; Strube, J., Systematic and Model-assisted evaluation of solvent based- or
pressurized hot water extraction for the extraction of artemisinin from Artemisia annua L.
Processes 2017, 5 (4), 15.
[9] Bai, C. C.; Tian, B. R.; Zhao, T.; Huang, Q.; Wang, Z. Z., Cyclodextrin-catalyzed organic
synthesis: reactions, mechanisms, and applications. Molecules 2017, 22 (9), 17.
[10] Hamada, H.; Ishihara, K.; Masuoka, N.; Mikuni, K.; Nakajima, N., Enhancement of
water-solubility and bioactivity of paclitaxel using modified cyclodextrins. J. Biosci. Bioeng.
2006, 102 (4), 369-371.
[11] Ol'khovich, M. V.; Sharapova, A. V.; Perlovich, G. L.; Skachilova, S. Y.; Zheltukhin, N.
K., Inclusion complex of antiasthmatic compound with 2-hydroxypropyl-beta-cyclodextrin:
Journal Pre-proof
23
Preparation and physicochemical properties. J. Mol. Liq. 2017, 237, 185-192.
[12] Petrovic, G.; Stojanovic, G.; Palic, R., Modified beta-cyclodextrins as prospective agents
for improving water solubility of organic pesticides. Environ. Chem. Lett. 2011, 9 (3),
423-429.
[13] Zhang, J.; Xiao, P.; Shi, S. Q.; Nie, J., Preparation and characterization of a water soluble
methylated beta-cyclodextrin/camphorquinone complex. Polym. Adv. Technol. 2009, 20 (9),
723-728.
[14] Xie, J.; Xu, Y.; Shishir, M. R. I.; Zheng, X.; Chen, W., Green extraction of mulberry
anthocyanin with improved stability using beta-cyclodextrin. J. Sci. Food Agric. 2019, 99 (5),
2494-2503.
[15] Feng, T.; Liu, F.; Sun, L.; Huo, H.; Ren, X.; Wang, M., Associated-Extraction Efficiency
of Six Cyclodextrins on Various Flavonoids in Puerariae Lobatae Radix. Molecules 2019, 24
(1).
[16] Teh, S. S.; Birch, E. J., Effect of ultrasonic treatment on the polyphenol content and
antioxidant capacity of extract from defatted hemp, flax and canola seed cakes. Ultrason.
Sonochem. 2014, 21 (1), 346-353.
[17] Fang, X. S.; Wang, J. H.; Wang, Y. Z.; Li, X. K.; Zhou, H. Y.; Zhu, L. X., Optimization
of ultrasonic-assisted extraction of wedelolactone and antioxidant polyphenols from Eclipta
prostrate L using response surface methodology. Sep. Purif. Technol. 2014, 138, 55-64.
[18] Dranca, F.; Oroian, M., Optimization of ultrasound-assisted extraction of total
monomeric anthocyanin (TMA) and total phenolic content (TPC) from eggplant (Solanum
melongena L.) peel. Ultrason. Sonochem. 2016, 31, 637-646.
[19] Chen, Y.; Mu, T., Application of deep eutectic solvents in biomass pretreatment and
conversion. Green Energy & Environ. 2019, 4 (2), 95-115.
[20] Zhang, Y.; Prawang, P.; Li, C.; Meng, X.; Zhao, Y.; Wang, H.; Zhang, S., Ultrasonic
assisted extraction of artemisinin from Artemisia Annua L. using monoether-based solvents.
Green Chem. 2018, 20 (3), 713-723.
[21] Zhou, S. K.; Bi, T. N.; Xu, Y. F.; Zhang, R. L.; Yang, M. J., Extraction optimization of
carbohydrate compound from Huangqi using orthogonal design. Int. J. Biol. Macromol. 2013,
58, 13-17.
[22] RenJie, L., Orthogonal test design for optimization of the extraction of polysaccharides
from Phascolosoma esulenta and evaluation of its immunity activity. Carbohydr. Polym. 2008,
73 (4), 558-563.
Journal Pre-proof
24
[23] Zhang, L. J.; Wang, M. S., Optimization of deep eutectic solvent-based
ultrasound-assisted extraction of polysaccharides from Dioscorea opposita Thunb. Int. J. Biol.
Macromol. 2017, 95, 675-681.
[24] Ilamathi, P.; Selladurai, V.; Balamurugan, K.; Sathyanathan, V. T., ANN-GA approach
for predictive modeling and optimization of NOx emission in a tangentially fired boiler.
Clean Technol. Environ. Policy 2013, 15 (1), 125-131.
[25] Bourquin, J.; Schmidli, H.; van Hoogevest, P.; Leuenberger, H., Advantages of Artificial
Neural Networks (ANNs) as alternative modelling technique for data sets showing non-linear
relationships using data from a galenical study on a solid dosage form. Eur. J. Pharm. Sci.
1998, 7 (1), 5-16.
[26] Aineer, K.; Chun, B. S.; Kwon, J. H., Optimization of supercritical fluid extraction of
steviol glycosides and total phenolic content from Stevia rebaudiana (Bertoni) leaves using
response surface methodology and artificial neural network modeling. Ind. Crop. Prod. 2017,
109, 672-685.
[27] Xie, P. J.; Huang, L. X.; Zhang, C. H.; Deng, Y. J.; Wang, X. J.; Cheng, J., Enhanced
extraction of hydroxytyrosol, maslinic acid and oleanolic acid from olive pomace: Process
parameters, kinetics and thermodynamics, and greenness assessment. Food Chem. 2019, 276,
662-674.
[28] Bucic-Kojic, A.; Planinic, M.; Tomas, S.; Bilic, M.; Velic, D., Study of solid-liquid
extraction kinetics of total polyphenols from grape seeds. J. Food Eng. 2007, 81 (1),
236-242.
[29] Karmakar, A. K.; Biswas, R. K., A study on the kinetics of extraction of Ti(IV) from
sulphate medium by Cyanex 302. Sep. Purif. Technol. 2019, 221, 331-337.
[30] Fang, S.; Tao, T. Y.; Cao, H. B.; Zheng, X. H.; Hu, Y. Y.; Zhang, Y.; Sun, Z., Selective
recovery of Gallium (Indium) from metal organic chemical vapor deposition dust-A
sustainable process. ACS Sustainable Chem. Eng.2019, 7 (10), 9646-9654.
[31] Norgan, A. P.; Coffman, P. K.; Kocher, J. P. A.; Katzmann, D. J.; Sosa, C. P., Multilevel
Parallelization of AutoDock 4.2. J. Cheminformatics 2011, 3, 9.
[32] Park, H.; Lee, J.; Lee, S., Critical assessment of the automated AutoDock as a new
docking tool for virtual screening. Proteins 2006, 65 (3), 549-554.
[33] Li, J. T.; Yu, T.; Xu, G. F.; Du, Y.; Liu, Z. R.; Feng, Z. J.; Yang, X.; Xi, Y.; Liu, J.,
Synthesis and application of ionic liquid functionalized beta-cyclodextrin,
mono-6-deoxy-6-(4-amino-1,2,4-triazolium)-beta-cyclodextrin chloride, as chiral selector in
Journal Pre-proof
25
capillary electrophoresis. J. Chromatogr. A 2018, 1559, 178-185.
[34] Illapakurthy, A. C.; Sabnis, Y. A.; Avery, B. A.; Avery, M. A.; Wyandt, C. M., Interaction
of artemisinin and its related compounds with hydroxypropyl-beta-cyclodextrin in solution
state: Experimental and molecular-modeling studies. J. Pharm. Sci. 2003, 92 (3), 649-655.
[35] Usuda, M.; Endo, T.; Nagase, H.; Tomono, K.; Ueda, H., Interaction of antimalarial
agent artemisinin with cyclodextrins. Drug Dev. Ind. Pharm. 2000, 26 (6), 613-619.
[36] Abulizi, A.; Okitsu, K.; Zhu, J. J., Ultrasound assisted reduction of graphene oxide to
graphene in L-ascorbic acid aqueous solutions: Kinetics and effects of various factors on the
rate of graphene formation. Ultrason. Sonochem. 2014, 21 (3), 1174-1181.
[37] Lin, A. J.; Klayman, D. L.; Hoch, J. M.; Silverton, J. V.; George, C. F., Thermal
rearrangement and decomposition products of artemisinin (Qinghaosu). J. Org. Chem. 1985,
50 (23), 4504-4508.
[38] Rodriguez-Tenreiro, C.; Alvarez-Lorenzo, C.; Concheiro, A.; Torres-Labandeira, J. J.,
Characterization of cyclodextrin-carbopol interactions by DSC and FTIR. J. Therm. Anal.
Calorim. 2004, 77 (2), 403-411.
[39] Schneider, H. J.; Hacket, F.; Rudiger, V.; Ikeda, H., NMR studies of cyclodextrins and
cyclodextrin complexes. Chem. Rev. 1998, 98 (5), 1755-1785.
[40] Ali, S. M.; Maheshwari, A., A proton magnetic resonance study of
fexofenadine/beta-cyclodextrin inclusion complexes in aqueous solution. Bull. Korean Chem.
Soc. 2005, 26 (12), 2061-2064.
[41] Hu, B.; Li, C.; Qin, W.; Zhang, Z. Q.; Liu, Y. T.; Zhang, Q.; Liu, A. P.; Jia, R. Y.; Yin, Z.
Q.; Han, X. F.; Zhu, Y. D.; Luo, Q. Y.; Liu, S. X., A method for extracting oil from tea
(Camelia sinensis) seed by microwave in combination with ultrasonic and evaluation of its
quality. Ind. Crop. Prod. 2019, 131, 234-242.
[42] Both, S.; Chemat, F.; Strube, J., Extraction of polyphenols from black tea - Conventional
and ultrasound assisted extraction. Ultrason. Sonochem. 2014, 21 (3), 1030-1034.
[43] Wang, Y. G.; Zhang, X.; Ma, X. Q.; Zhang, K.; Li, S. W.; Wang, X. J.; Liu, X. F.; Liu, J.
C.; Fan, W. G.; Li, Y. L.; Li, Q. Y.; Zhu, X. Q., Study on the kinetic model, thermodynamic
and physicochemical properties of Glycyrrhiza polysaccharide by ultrasonic assisted
extraction. Ultrason. Sonochem. 2019, 51, 249-257.
[44] Mouton, J.; Jansen, O.; Frederich, M.; van der Kooy, F., Is Artemisinin the Only
Antiplasmodial Compound in the Artemisia annua Tea Infusion? An in Vitro Study. Planta
Med. 2013, 79 (6), 468-470.
Journal Pre-proof
26
[45] Rath, K.; Taxis, K.; Walz, G.; Gleiter, C. H.; Li, S. M.; Heide, L., Pharmacokinetic study
of artemisinin after oral intake of a traditional preparation of Artemisia annua L. (annual
wormwood). Am. J. Trop. Med. Hyg. 2004, 70 (2), 128-132.
[46] Wong, J. W.; Yuen, K. H., Improved oral bioavailability of artemisinin through inclusion
complexation with beta- and gamma-cyclodextrins. Int. J. Pharm. 2001, 227 (1-2), 177-185.
Journal Pre-proof
•
A novel process for artemisinin extraction with aqueous solution was developed.
•
Hydroxypropyl-β-cyclodextrin was a good solutizer for this extraction process.
•
More artemisinin could be extracted by this method compared to conventional
process.
•
The mechanism was unveiled by molecular docking simulation and
characterization.
Journal Pre-proof
