Received: 31 July 2020 Revised: 25 January 2021 Accepted: 3 February 2021

DOI: 10.1111/1753-0407.13163

ORIGINAL ARTICLE

Inhibition of α-glucosidase and α-amylase by herbal

compounds for the treatment of type 2 diabetes: A
validation of in silico reverse docking with in vitro
enzyme assays

Morné Tolmie1 | Megan Jean Bester2 | Zeno Apostolides1

1
Department of Biochemistry, Genetics,
and Microbiology, University of Pretoria, Abstract
Pretoria, South Africa Background: α-Amylase and α-glucosidase are important therapeutic targets
2
Department of Anatomy, University of for the management of type 2 diabetes mellitus. The inhibition of these
Pretoria, Pretoria, South Africa
enzymes decreases postprandial hyperglycemia. In the present study, com-
Correspondence pounds found in commercially available herbs and spices were tested for their
Morné Tolmie, Department of ability to inhibit α-amylase and α-glucosidase. These compounds were
Biochemistry, Genetics and Microbiology,
acetyleugenol, apigenin, cinnamic acid, eriodictyol, myrcene, piperine, and
University of Pretoria, Private Bag X20,
Hatfield, Pretoria, 0028, South Africa. rosmarinic acid.
Email: [email protected] Methods: The enzyme inhibitory nature of the compounds was evaluated
Funding information
using in silico docking analysis with Maestro software and was further con-
University of Pretoria firmed by in vitro α-amylase and α-glucosidase biochemical assays.
Results: The relationships between the in silico and in vitro results were well
correlated; a more negative docking score was associated with a higher in vitro
inhibitory activity. There was no significant (P > .05) difference between the
inhibition constant (Ki) value of acarbose, a widely prescribed α-glucosidase
and α-amylase inhibitor, and those of apigenin, eriodictyol, and piperine. For
α-amylase, there was no significant (P > .05) difference between the Ki value
of acarbose and those of apigenin, cinnamic acid, and rosmarinic acid. The
effect of the herbal compounds on cell viability was assessed with the sul-
forhodamine B (SRB) assay in C2C12 and HepG2 cells. Acetyleugenol,
cinnamic acid, myrcene, piperine, and rosmarinic acid had similar (P > .05)
IC50 values to acarbose.
Conclusions: Several of the herbal compounds studied could regulate post-
prandial hyperglycemia. Using herbal plants has several advantages including
low cost, natural origin, and easy cultivation. These compounds can easily be
consumed as teas or as herbs and spices to flavor food.

KEYWORDS
α-amylase, α-glucosidase, herbal compounds, in vitro cytotoxicity, reverse molecular docking,
type 2 diabetes

© 2021 Ruijin Hospital, Shanghai Jiaotong University School of Medicine and John Wiley & Sons Australia, Ltd

Journal of Diabetes. 2021;1–13. wileyonlinelibrary.com/journal/jdb 1

2 TOLMIE ET AL.

Highlights
• There is a positive relationship between in silico docking and in vitro assays.
• Several of the herbal compounds studied regulate postprandial hyperglycemia.
• Apigenin is a monotherapeutic agent, inhibiting both α-amylase and
α-glucosidase.
• Most herbal compounds are not more toxic than acarbose.

1 | INTRODUCTION compounds found in commercially available herbs and

spices. The compounds were found in herbs that had
Diabetes mellitus (DM) is one of the major health chal- three characteristics according to the literature,20
lenges of the 21st century. It is a chronic metabolic disorder α-amylase inhibitory activity,20 α-glucosidase inhibitory
characterized by hyperglycemia, caused by insufficient insu- activity,20 and insulin mimicking effects.20 Acetyleugenol
lin secretion and/or insulin resistance.1,2 According to the is a phenol ester21 found abundantly in Syzygium
International Diabetes Federation (IDF), nearly half a bil- aromaticum (cloves).22 The flavone23 apigenin is found in
lion people have DM worldwide.3 Type 2 DM (T2DM) Petroselinum crispum (parsley).24 Cinnamic acid, a
accounts for about 90% of all diabetic cases4,5 and is primar- phenylpropanoid acid,25 is found in Cinnamomum
ily caused by various genetic, environmental, and behav- loureiroi (Vietnamese cinnamon).26 The flavanone27
ioral factors.4 Prolonged hyperglycemia causes various eriodictyol is found in Lippia graveolens (Mexican oreg-
complications, including stroke, cardiovascular diseases, ano).27 Myrcene is a monoterpene28 found in Myristica
neuropathy, retinopathy, and nephropathy.5-7 One of the fragrans (nutmeg).29 Piperine is an alkaloid30 found
best treatment strategies for T2DM involves hyperglycemic abundantly in Piper nigrum (black pepper),29 and ros-
control8-10 through the inhibition of carbohydrate hydrolyz- marinic acid is a phenolic acid31 found abundantly in
ing enzymes, such α-amylase and α-glucosidase.9 Mentha piperita (peppermint).31 Previous studies have
Pancreatic α-amylase is secreted once carbohydrates shown that herbs and spices such as cinnamon, cloves,
reach the intestinal lumen and is responsible for the hydroly- mint, nutmeg, oregano, parsley, and pepper alleviate
sis of carbohydrates into smaller oligosaccharides, such as abdominal pain, diarrhea, and flatulence,33 counteracting
maltotriose, maltose, and glucose.11 These oligosaccharides the side effects commonly caused by α-amylase and
are further broken down into glucose by α-glucosidase, found α-glucosidase inhibitors.
in the brush border of the intestine.8,9,11 The released glucose The aim of this study was to find new inhibitors
molecules are absorbed into the bloodstream, resulting in (Figure 1) of α-amylase and α-glucosidase in commer-
hyperglycemia.9,12,13 Therefore, inhibition of α-amylase and cially available herbs and spices, using in silico and
α-glucosidase decreases the rate of starch hydrolysis and pre- in vitro relationship studies. To the best of our
vents a sudden surge in glucose, resulting in lower postpran- knowledge, this is the first study documenting the
dial hyperglycemia.8,12,14,15 The oral drug, acarbose is a inhibition constants (K i) and mode of inhibition
widely prescribed α-amylase and α-glucosidase inhibitor, of these herbal compounds on α-amylase and
despite causing various gastrointestinal adverse effects such α-glucosidase.
as flatulence, diarrhea, and abdominal pain.11,16-18 There is
an increasing effort to discover effective α-amylase and
α-glucosidase inhibitors from natural sources.8,16-18 2 | MATERIALS AND METHODS
Herbs and spices have been used for centuries to fla-
vor and preserve food.19 They possess a wide variety of 2.1 | Chemicals
bioactivities, including anticancer, antidiabetic, anti-
inflammatory, antimicrobial, and antioxidant effects.19,20 The following analytical grade reagents were purchased
Using herbal compounds as an alternative treatment from Sigma-Aldrich Co (St Louis, Missouri):
strategy for T2DM is beneficial in many developing coun- 3,5-dinitrosalicylic acid (DNSA), acarbose, acetyleugenol,
tries where conventional medicine is logistically α-glucosidase (EC 3.2.1.20) from Saccharomyces
unavailable and/or cannot be afforded. Herbs and spices cerevisiae, apigenin, cinnamic acid, Dulbecco's Modified
are widely available and fairly inexpensive. Eagle Media (DMEM), eriodictyol, maltose monohydrate,
In the present study, we examined the α-amylase and myrcene, piperine, p-nitrophenol, p-nitrophenyl-α-D-
α-glucosidase inhibitory activity of seven herbal glucopyranoside (pNPG), porcine pancreatic α-amylase
TOLMIE ET AL. 3

FIGURE 1 Chemical structures of the herbal compounds

(EC 3.2.1.1), rosmarinic acid, starch from potato, and sul- 2.2 | In silico docking analysis
forhodamine B (SRB). C2C12 myotubes (American Type
Culture Collection [ATCC] CRL-1772) and HepG2 The docking studies were performed using Schrödinger's
hepatocarcinoma cells (ATCC HB-8065) were obtained Maestro (Maestro v 11.5; Schrödinger LLC, New York,
from the ATCC. New York) program.
4 TOLMIE ET AL.

2.2.1 | Ligand preparation determine the bioavailability and toxicity of the eight
compounds. The SMILES of each compound was
In this study, the isomeric simplified molecular-input imported into Canvas, after which the physiochemical
line-entry system (SMILES) of roughly a thousand com- properties were calculated. Each compound-structure
pounds, identified from 30 commercially available herbs was minimized to obtain 3D structures before the
and spices,20 were imported onto Maestro. Acarbose was Qikprop descriptors were computed. Once the calcula-
selected as a standard drug reference molecule. The 3D tions were completed, the percentage human oral
structures34 of the ligands were prepared using the absorption, QPlogHERG, and #stars for each com-
LigPrep function, which generates several poses from pound were noted.
each input structure.35 The default parameters, including
“Retain specified chiralities,” were kept.
2.3 | In vitro enzyme analysis

2.2.2 | Protein preparation 2.3.1 | Kinetics of α-glucosidase

inhibition
The crystal structure of the two enzymes (Protein Data
Bank [PDB] entries: 3L4Y and 4GQR for α-glucosidase The in vitro α-glucosidase activity was measured using
and α-amylase, respectively) were downloaded from the previously described methods.15,39 For the enzymatic
PDB (www.rscb.org) and imported into Maestro. The pro- assay, the compounds were dissolved in 100 mM phos-
tein preparation wizard was used to prepare the proteins phate buffer (pH 6.8). Briefly, in a 96-well plate (Greiner,
for in silico experimentation. The imported 3D structures clear F-bottom) 25 μL enzyme (0.2 U/mL) diluted in
were made fit to study the docking.34 Therefore, all the phosphate buffer (100 mM, pH 6.8) was preincubated
cofactor and water molecules were removed from the with 100 μL inhibitor (250, 500, and 1000 μM) or the pos-
proteins, and the hydrogen bonding was optimized, itive control acarbose (250, 500, and 1000 μM) for
followed by an energy minimization step.36 10 minutes at 37 C. Thereafter, 25 μL pNPG (0, 0.15, 0.3,
0.6, 0.9, 1.15, and 1.5 mM) was added, and the reaction
mixtures were incubated at 37 C for 30 minutes. The
2.2.3 | Molecular docking reaction was stopped by adding 50 μL NaOH (0.1 M).
Enzyme activity was quantified by measuring the absor-
Grid representations of the active sites of both proteins bance at 405 nm using an ultraviolet-visible (UV-VIS)
were created using the receptor grid generation tool.37 spectrophotometer (Spectramax paradigm; Molecular
The default parameters were kept.38 Protein docking was Devices Inc, San Jose, California).
carried out using the Glide high-throughput virtual
screening (HTVS) protein docking module of the virtual
screening workflow function of Maestro. The interactions 2.3.2 | Kinetics of α-amylase inhibition
between the ligands and protein were quantified with the
GlideScore.34 The best docked pose with the lowest The α-amylase inhibitory activity of the compounds
GlideScore value was recorded for each ligand. was evaluated using a colorimetric assay described in
Research done by Pereira et al20 identified herbs and previous literature18 with slight modifications. In
spices with known α-amylase and α-glucosidase inhibition. short, 100 μL porcine pancreatic amylase enzyme
We validated their findings using Schrödinger's Maestro (2 U/mL) dissolved in 20 mM phosphate buffer
program and identified seven bioactive compounds found (pH 6.9) with 6.7 mM NaCl was preincubated with
in these herbs and spices to test for their inhibitory activity 100 μL herbal compounds (2.5, 5, and 10 μM) or the
in vitro. Acarbose was used as a positive control. We chose positive control, acarbose (2.5, 5, and 10 μM) for
four compounds with stronger docking than acarbose and 10 minutes at 25 C. One hundred microliter of the
three compounds with weaker docking as negative controls. substrate (0, 1.3, 2.6, 4.0, 5.3, and 6.6 mg/mL) starch
was added before the reaction was stopped after
10 minutes by adding 100 μL DNSA (96 mM). The
2.2.4 | Calculation of physiochemical reaction mixtures were heated at 85  C for 10 minutes.
properties After the final reaction mixtures were diluted with
1.1 mL double-distilled H 2O, 200 μL was pipetted into
Schrödinger's Canvas (Canvas v 3.5.011; Schrödinger 96-well plates to read the absorbance at 540 nm using
LLC, New York, New York) program was used to the UV-VIS spectrophotometer.
TOLMIE ET AL. 5

2.3.3 | In vitro cytotoxicity (compound) concentration on GraphPad Prism. The

kinetic parameters and IC50 values were subjected to a
The cytotoxicity of the herbal compounds was evaluated
using the method described by Vichai and Kirtikara.40
The cells were seeded (100 μL) at 1 × 105 per well and left
overnight at 37 C to allow attachment. A volume of T A B L E 1 Glide HTVS docking scores of the herbal compounds
100 μL herbal compounds (0.005-500 μM) was added, docked to α-amylase
and the plates were further incubated for 72 hours at Compound Glide score (Kcal/mol)
37 C. Saponin was used as a positive control. Wells con-
Rosmarinic acid −7.9
taining media only served as blanks, while wells con-
taining cells and media served as the negative control. Eriodictyol −7.4
The cells were fixed by adding 50 μL 50% (w/v) trichlor- Apigenin −6.6
oacetic acid and incubating the plates at 4 C for 24 hours. Piperine −5.8
The plates were washed with water and dried overnight Acarbose (positive control) −5.2
before 100 μL SRB dye (0.057% w/v) was added. After a Acetyleugenol −4.3
30-minute incubation period, the plates were washed
Cinnamic acid −4.2
with 1% acetic acid. The plates were dried overnight
Myrcene −1.6
before 200 μL tris (10 mM) was added. The plates were
shaken gently (550 rpm) for 1 hour, and the absorbance Abbreviation: HTVS, high-throughput virtual screening.
was read at 540 nm.

T A B L E 2 Glide HTVS docking scores of the herbal compounds

2.4 | Data analysis docked to α-glucosidase

Compound Glide score (Kcal/mol)

All of the results represent at least three independent
Eriodictyol −5.5
experiments and are expressed as mean ± SD. The
kinetic parameters of the compounds and the type of Rosmarinic acid −5.4
inhibition they exert were studied on GraphPad Prism Apigenin −5.3
(version 8.3.0; GraphPad Software, San Diego, Califor- Piperine −4.2
nia) using Michaelis-Menten kinetics and the Acarbose (positive control) −4.1
corresponding Lineweaver-Burk double reciprocal plots. Cinnamic acid −3.4
The Ki values were obtained with secondary plots and
Acetyleugenol −3.4
were analyzed using a one-sided unpaired Student's
Myrcene −1.4
t test. The IC50 of the compounds was calculated by plot-
ting the percentage cell viability against the log drug Abbreviation: HTVS, high-throughput virtual screening.

T A B L E 3 Selected molecular
Compound Human oral absorptiona (%) QPlogHERGb #Starsc
properties of herbal compounds
Acarbose (positive control) 0 −5.6 13
Eriodictyol 63 −4.9 0
Piperine 100 −4.8 1
Acetyleugenol 100 −4.6 0
Myrcene 100 −3.8 5
Apigenin 71 −3.8 3
Rosmarinic acid 35 −3.7 2
Cinnamic acid 45 −3.7 4
a
Predicted human oral absorption on a 0% to 100% scale. The prediction is based on a quantitative multiple
linear regression model. A value of >80% is considered high, and <25% is considered poor.46
b
Predicted IC50 value for blockage of HERG K+ channels. A value below −5 is a concern.46
c
Number of property or descriptor values that fall outside the 95% range of similar values for known drugs.
A large number of stars suggests that a molecule is less drug-like than molecules with few stars. The
recommended range is 0 to 5, where 0 indicates no violation or best candidate.46
6 TOLMIE ET AL.

two-sided, unpaired Student's t test. Differences were and α-glucosidase, respectively. The compounds with bet-
considered significant at P < .05. ter docking scores than acarbose to both enzymes were
apigenin, eriodictyol, piperine, and rosmarinic acid.
Acetyleugenol, cinnamic acid, and myrcene had weaker
3 | R E SUL T S docking scores than acarbose.

3.1 | Molecular docking studies

3.2 | In silico physiochemical properties
We investigated roughly a thousand herbal com-
pounds in silico for α-amylase and α-glucosidase inhi- Canvas was used to evaluate the toxicity, bioavailability,
bition through docking analysis. To reduce the and druggability of the chosen herbal compounds in silico.
number of compounds for in vitro analysis, we used a These parameters were tested independently, without the
literature review 20 to identify compounds in our dock- enzymes. The results are purely based on the structure of
ing study with confirmed inhibitory activity of both the compound itself. Acetyleugenol, myrcene, and piperine
α-amylase and α-glucosidase. Identifying compounds had the highest percentage human oral absorption (100%),
with both α-amylase and α-glucosidase inhibitory while acarbose had the lowest at 0% (Table 3). Cinnamic
activity increased our chances of finding mon- acid and rosmarinic acid had the lowest HERG toxicity,
otherapeutic targets. while acarbose had the highest. Acarbose had the lowest
The results in Tables 1 and 2 show the docking scores druggability score at 13 stars. The herbal compounds had
of these seven herbal compounds docked to α-amylase no more than five stars, substantially less than acarbose.

T A B L E 4 Inhibitory activity of
Compound Type of inhibition Ki (μM) P value
herbal compounds against porcine
Acarbose (positive control) Competitive 3.8 ± 1.9 pancreatic α-amylase
Rosmarinic acid Noncompetitive 4.5 ± 2.9 .364
Apigenin Mixed 7.8 ± 2.7 .054
Cinnamic acid Noncompetitive 8.0 ± 4.5 .094
a
Eriodictyol Noncompetitive 10.5 ± 3.6 .023
Piperine Competitive 10.9 ± 5.5a .042
a
Acetyleugenol Mixed 12.1 ± 5.8 .039
a
Myrcene None 49.0 ± 33.7 .038

Note: Data are represented as mean ± SD (n = 3).

Abbreviation: Ki, inhibition constant.
a
Values significantly different (P < .05) from acarbose as determined by a one-sided Student's t test.

T A B L E 5 Inhibitory activity of
Compound Type of inhibition Ki (μM) P value
herbal compounds against yeast
Eriodictyol Mixed 130 ± 70 .284 α-glucosidase
Apigenin Mixed 160 ± 50 .452
Acarbose (positive control) Mixed 170 ± 80
Piperine Mixed 280 ± 120 .115
Cinnamic acid Noncompetitive 620 ± 380a .043
Acetyleugenol Noncompetitive 950 ± 240a .002
a
Myrcene None 1580 ± 650 .011
a
Rosmarinic acid Uncompetitive 2580 ± 550 .008

Note: Data are represented as mean ± SD (n = 3).

Abbreviation: Ki, inhibition constant.
a
Values significantly different (P < .05) from acarbose as determined by a one-sided Student's t test.
TOLMIE ET AL. 7

TABLE 6 IC50 values of herbal compounds on C2C12 and

HepG2 cells

C2C12 HepG2
Compound IC50 (μM) IC50 (μM)
Eriodictyol 11 ± 2a 41 ± 2
Apigenin 31 ± 4a 210 ± 41
Acarbose (positive 60 ± 15 >500
control)
Piperine 79 ± 8 >500
Rosmarinic acid 83 ± 2 >500
Myrcene 84 ± 3 >500
Cinnamic acid 87 ± 2 >500
Acetyleugenol 96 ± 24 >500

Note: Data are represented as mean ± SD (n = 3).

F I G U R E 2 Graph of negative delta G vs the Ki value of the a
Values significantly different (P < .05) from acarbose determined with a
herbal compounds against α-amylase
two-sided Student's t test.

cinnamic acid and rosmarinic acid do not clearly differentiate

between mixed and noncompetitive inhibition, there is no
statistically significant difference between their Km values
(Table S1) and the Km value of the uninhibited reaction,
corresponding to noncompetitive inhibition. Acetyleugenol
and apigenin showed mixed inhibition. Acetyleugenol and
cinnamic acid inhibited α-glucosidase noncompetitively,
while acarbose, apigenin, eriodictyol, and piperine showed
mixed inhibition (Figure S2 and Table S2). Rosmarinic acid
showed uncompetitive inhibition. Myrcene showed no statis-
tically significant inhibition of α-amylase or α-glucosidase.
The inhibition potential of each compound was evalu-
ated and compared based on their Ki values. The Ki was
F I G U R E 3 Graph of negative delta G vs the Ki value of the
calculated with secondary graphs by plotting the recipro-
herbal compounds against α-glucosidase. Rosmarinic acid (shown cal of the Lineweaver-Burk plot slope against the inhibi-
with a square symbol) was identified as an outlier with undue tor concentration. There was no significant difference
influence on the slope between the Ki value of acarbose and those of apigenin,
cinnamic acid, and rosmarinic acid when inhibiting
α-amylase. Acetyleugenol eriodictyol, myrcene, and pip-
3.3 | Kinetics of the in vitro α-amylase erine had significantly higher Ki values than acarbose.
and α-glucosidase inhibition For α-glucosidase inhibition, there was no significant dif-
ference between the Ki value of acarbose and those of
The ability of the herbal compounds to inhibit α-amylase apigenin, eriodictyol, and piperine, while acetyleugenol,
and α-glucosidase was investigated in vitro. The inhibitory cinnamic acid, myrcene, and rosmarinic acid had a statis-
characteristics of the herbal compounds were explored tically significantly higher Ki than acarbose.
by performing kinetic assays, with double reciprocal The relationship between the docking scores and Ki were
Lineweaver-Burk plots to calculate the kinetic parameters. visualized with graphs by plotting negative delta G against
The type of inhibition exerted by the herbal compounds was the Ki of each compound. There is a positive relationship
deduced from the calculation of the Michaelis constant (Km) between the negative delta G score and the Ki values of the
and maximum enzyme velocity (Vmax). Regarding the inhibi- herbal compounds when inhibiting α-glucosidase and
tion of α-amylase, it can be concluded that acarbose and pip- α-amylase. The Ki value of rosmarinic acid, inhibiting α-glu-
erine were competitive inhibitors, whereas cinnamic acid, cosidase, was excluded from the calculation of the slope in
eriodictyol, and rosmarinic acid were noncompetitive inhibi- Figure 3 because it had undue influence on the line but is
tors. Although the Lineweaver-Burk graphs (Figure S1) of included in the graph with a different symbol.
8 TOLMIE ET AL.

T A B L E 7 Herbal dosage required

Amount in Amount (g) needed to
herb/spice relate to daily dose of to relate to a daily dose of acarbose
Compound Main source (mg/100 g) acarbose
Acetyleugenol Syzygium aromaticum 2075 7.2
(cloves)
Apigenin Petroselinum crispum 1263 11.9
(parsley, dried)
Cinnamic Cinnamomum loureiroi 1697 8.8
acid (Vietnamese
cinnamon)
Eriodictyol Lippia graveolens 85 176
(Mexican oregano,
dried)
Myrcene Myristica fragrans 333 45
(nutmeg)
Piperine Piper nigrum (black 5350 2.8
pepper)
Rosmarinic Mentha piperita 1734 8.6
acid (peppermint, dried)

Note: Data for acetyleugenol, apigenin, eriodictyol, myrcene, and piperine from foodb.ca.31
Data for acetyleugenol, eriodictyol, and rosmarinic acid from phenol-explorer.eu.
Data for cinnamic acid from Lee et al.26
All based on a 150-mg acarbose dose per day.

3.4 | In vitro cytotoxicity 4 | DISCUSSION

The cytotoxic effects of the herbal compounds on the cell The control of postprandial hyperglycemia is important
lines were quantified with IC50 values (Table 6). All herbal in the treatment of T2DM and the prevention of short-
compounds, except for apigenin and eriodictyol, displayed and long-term complications.41 Inhibition of enzymes,
limited cytotoxicity in the HepG2 cell line, where concen- such as α-amylase and α-glucosidase, involved in the
trations up to 500 μM did not induce 50% cell death metabolism of carbohydrates is an important therapeutic
(Figure S4); their IC50 values could not be calculated accu- approach for reducing postprandial hyperglycemia.41
rately. Eriodictyol displayed significant toxicity, while Interest in using herbal plants has grown recently due to
apigenin displayed milder toxicity against this cell line. their low cost, natural origin, and easy cultivation.42 This
In C2C12 cells, acetyleugenol, cinnamic acid, study highlights the inhibitory effect of acetyleugenol,
myrcene, piperine, and rosmarinic acid had similar IC50 apigenin, cinnamic acid, eriodictyol, myrcene, piperine,
values to acarbose. Thus, the toxicity of these com- and rosmarinic acid on the activity of yeast α-glucosidase
pounds was not statistically more significant (P > .05) and porcine pancreatic amylase.
than the toxicity of acarbose at the same concentration The predicted binding of the herbal compounds to
(500 μM). Eriodictyol and apigenin were significantly α-amylase and α-glucosidase were studied in silico before
(P < .05) more toxic than acarbose. any studies were performed in the lab. Herbal compound
docking scores for α-amylase and α-glucosidase were gen-
erated through Maestro. Maestro uses Schrödinger's
3.5 | Herbal dosage related to daily GlideScore function; this algorithm recognizes favorable
acarbose dose hydrophobic, hydrogen-bonding, and metal-ligation
interactions and penalizes steric clashes between the
The FooDB32 and Phenol Explorer31 databases were used ligand and the protein.43 The herbal compounds were
to search for natural sources of each herbal compound. docked into the catalytic site of both enzymes where they
The databases give the amount (mg/100 g) of each com- interacted with the amino acids through a negative bind-
pound found in many herbs and spices. This was used to ing energy, indicating a spontaneous binding and poten-
calculate the amount (g) of each herb required to relate tial inhibition. With α-amylase, the decreasing order of
to the average daily dose (150 mg) of acarbose (Table 7). the positive binding and potential inhibition was
TOLMIE ET AL. 9

rosmarinic acid > eriodictyol > apigenin > piperine > indicating that most of the 24 pharmaceutically relevant
acarbose > acetyleugenol > cinnamic acid > myrcene descriptors lie within the recommended range of known
(Table 1). With α-glucosidase it was eriodictyol > ros- drugs.48 All of the compounds have a lower #stars than
marinic acid > apigenin > piperine > acarbose > acarbose, which means the herbal compounds are more
cinnamic acid > acetyleugenol > myrcene (Table 2). The drug-like than acarbose.
docking scores of these herbal compounds were com- The kinetic parameters of the herbal compounds
pared with the interactions of acarbose with α-amylase were calculated with Lineweaver-Burk double recipro-
and α-glucosidase, whose binding energies were −5.2 and cal plots. Acarbose was verified as a competitive inhibi-
−4.1 Kcal/mol, respectively. Evidently, apigenin, tor of α-amylase17,49,50 and a mixed inhibitor of
eriodictyol, piperine, and rosmarinic acid had better α-glucosidase.17,51 Acetyleugenol, apigenin, cinnamic
docking scores to α-amylase and α-glucosidase than aca- acid, eriodictyol, piperine, and rosmarinic acid dis-
rbose. Acetyleugenol, cinnamic acid, and myrcene played dose-dependent inhibition of α-amylase and
showed weaker docking scores than acarbose. The results α-glucosidase, using acarbose as positive control.
of the docking analysis encouraged us to further investi- Cinnamic acid, eriodictyol, and rosmarinic acid
gate the enzyme inhibitory activity using enzyme assays. inhibited α-amylase noncompetitively, the Km value
Selected physiochemical parameters of the herbal com- remained constant, and the Vmax value decreased
pounds were evaluated in silico. The cardiotoxicity of each (Table S1 and Figure S1). Acetyleugenol and apigenin
herbal compound was assessed using the QPlogHERG inhibited α-amylase in a mixed fashion, the Vmax was
function, which is the projected log IC50 value for the decreased, and the Km increased. Piperine inhibited
blockage of HERG potassium (K+) channels.44,45 The Can- α-amylase competitively, leading to an increased Km,
vas software calculates the distances and angles between while the Vmax of the reaction stayed the same.
the carbon atoms of each herbal compound and predicts Apigenin, eriodictyol, and piperine were mixed inhibi-
the pIC50.45 A value between 0 and −5 is desired.44,45 A tors of α-glucosidase, while acetyleugenol and cinnamic
more negative value can lead to a disorder called long Q-T acid were noncompetitive inhibitors. Rosmarinic acid
syndrome.45 All the herbal compounds had more positive inhibited α-glucosidase uncompetitively, decreasing
QPlogHERG values than acarbose, indicating that these both the Km and Vmax (Table S2 and Figure S2).
compounds have a lower probability than acarbose of Myrcene presented no inhibitory activity against
causing long Q-T syndrome. The percentage human oral α-amylase and α-glucosidase.
absorption is calculated by studying the number of metab- Ki values can be a useful tool to compare the inhibitory
olites, logP, rotatable bonds, solubility, and cell permeabil- activity of the herbal compounds, being an indicator of the
ity of each compound. binding affinity of the inhibitor. A lower Ki value suggests
Acarbose is an orally administered drug, thus a low oral a higher binding affinity. The Ki values of the tested com-
absorption is expected. In this case acarbose had a human pounds against α-amylase were: acarbose < rosmarinic
oral absorption of 0%. No oral absorption is required since acid < apigenin < cinnamic acid < eriodictyol < piperine
the target site for acarbose is the lumen of the gastrointesti- < acetyleugenol < myrcene. There was no statistically sig-
nal tract (GIT). Rosmarinic acid had a human oral absorp- nificant difference (Table 4, P > .05) between the Ki values
tion of 35%, which is the lowest of all the herbal of acarbose and those of rosmarinic acid, apigenin, and
compounds, while cinnamic acid and eriodictyol had an cinnamic acid. None of the herbal compounds had a Ki
oral absorption of 45% and 63%, respectively. Apigenin had value lower than that of acarbose, which might be thera-
an oral absorption of 71%. These values are substantially peutically desired. Mild inhibition of α-amylase is often
higher than the oral absorption of acarbose. Acetyleugenol, preferred to avoid excessive bacterial fermentation, leading
myrcene, and piperine had an oral bioavailability of 100%, to gastrointestinal side effects.11,52 Regarding α-glucosi-
which means most of the administered drug gets absorbed dase, the decreasing order of the Ki values were:
into the systemic circulation. Thus, little drug remains in eriodictyol < apigenin < acarbose < piperine < cinnamic
the GIT, where the drug's action is required. acid < acetyleugenol < myrcene < rosmarinic acid. There
The druggability of the compounds was evaluated was no statistically significant difference (Table 5, P > .05)
using the #stars function. The QikProp function of Can- in the inhibition effect of apigenin, eriodictyol, and piper-
vas includes 24 descriptors in the calculation of the ine when compared with that of acarbose. This indicates
#stars.46 The #stars of each compound indicates the num- that these herbal compounds inhibit α-glucosidase with
ber of descriptor values that fall outside of 95% of similar the same efficacy as acarbose.
values for known drugs.47 Therefore, a lower #stars indi- One aim of this study was to determine the relation-
cates a better drug-like molecule.48 It is clear from Table 3 ship between the in silico docking results and in vitro
that all herbal compounds had no more than five stars, inhibitory strength. Figures 2 and 3 show the relationship
10 TOLMIE ET AL.

between the negative docking scores and Ki values. The herbal compound tested in this study is currently found
slopes of both graphs were positive, indicating a positive in commercially available herbs and spices, safety at culi-
relationship between the docking scores and the Ki nary relevant levels is apparent.
values of the herbal compounds. A high negative docking The results have shown that various herbal com-
score corresponds to a low Ki value, both indicating pounds are effective α-amylase and α-glucosidase inhibi-
higher inhibition efficacy. The compounds with better tors. We wanted to estimate the daily dose of each herbal
docking scores than acarbose to both α-amylase and plant that would be equivalent to the average daily dose
α-glucosidase were apigenin, eriodictyol, piperine, and of acarbose, which is 150 mg daily for a 60 kg individ-
rosmarinic acid. With regard to α-amylase, it can be seen ual.59,60,61 Since there is no statistically significant differ-
that the in vitro results of rosmarinic acid and apigenin ence (Tables 4 and 5, P > .05) between the Ki of acarbose
correspond well to the in silico results. For both enzymes, and those of apigenin, cinnamic acid, eriodictyol, piper-
the herbal compounds that had weaker docking scores ine, and rosmarinic acid, we made the assumption that
than acarbose showed a significantly weaker inhibition 150 mg of each compound will have the same effect as
effect than acarbose in vitro, with the exception of acarbose, without taking bioavailability into account.
cinnamic acid when inhibiting α-amylase. Concerning However, this assumption needs validation with animal
α-glucosidase, the docking scores of eriodictyol, apigenin, studies in the future.
and piperine correlate well with the in vitro results, indi- There are various simple ways to include herbs and
cating a good overall correlation between the in silico spices in one's daily life. The most common way would
docking results and in vitro inhibition results. be to add the herbs or spices to food. Another way would
The cytotoxic effects of the herbal compounds on the be to brew a herbal tea. Since a tea bag usually contains
cell lines were quantified with IC50 values. Cytotoxicity about 2.5 g dried leaves or herbs, it would be an easy way
was only observed at concentrations larger than 50 μM to incorporate a large amount of phytochemicals into
(see Figures S3 and S4); this provides evidence of the pre- one's diet. The amount (Table 7) of acetyleugenol,
clinical safety of these compounds at culinary relevant apigenin, piperine, and rosmarinic acid that needs to be
concentrations. The HepG2 cells were overall more ingested to relate to a daily dose of acarbose can thus be
cytotoxic resistant than the C2C12 cells, likely due to realistically achieved. Pepper is insoluble in water and
the liver's detoxification capabilities. Acetyleugenol, should thus be used as a spice to be sprinkled over food;
cinnamic acid, myrcene, piperine, and rosmarinic acid however, cloves, mint, and parsley can be brewed in a
did not induce a 50% decrease in cell viability at the tea.62-64
highest tested concentration (500 μM) in HepG2 cells,
confirming their low toxicity. Only two compounds,
apigenin and eriodictyol, had toxic effects on both cell 5 | CONCLUSION
lines. These compounds have significantly lower IC50
values than acarbose. Apigenin and eriodictyol are both Due to the increasing prevalence of T2DM, there has been
flavonoids and therefore have very similar structures. an ongoing effort to find natural compounds that can con-
Various studies have reported the toxicity of flavonoids, trol hyperglycemia. Here we report that several herbal
including apigenin and eriodictyol, at high concentra- compounds possess potential antidiabetic activities due to
tions.53 In C2C12 cells, the IC50 values of acetyleugenol, their ability to inhibit both α-amylase and α-glucosidase.
cinnamic acid, myrcene, piperine, rosmarinic acid, and The strength of the enzyme inhibition was calculated in
acarbose were similar (P > .05), indicating that these silico using docking analysis and compared with in vitro
compounds are not more toxic than acarbose, a widely inhibition efficacy using enzymatic assays. The relation-
prescribed drug. All of these compounds had higher IC50 ships between the in silico and in vitro results were well
values than acarbose, implying that they are less toxic correlated, a more negative docking score translated to a
than acarbose at the same dose. Neiro and Machado- higher in vitro inhibitory activity. Our results have shown
Santelli54 (2013) confirmed the low in vitro and in vivo that apigenin, cinnamic acid, and rosmarinic acid are
cytotoxicity of cinnamic acid. Şahin, et al55 (2017) effective α-amylase inhibitors, while apigenin, eriodictyol,
reported no cytotoxic activity when five different cell and piperine inhibited α-glucosidase effectively. Apigenin
lines were treated with rosmarinic acid. The cytotoxicity was identified as a monotherapeutic agent, inhibiting both
of myrcene56 and piperine57,58 has been reported to be α-amylase and α-glucosidase. The present study provides
low in HepG2 and other cell lines, although the use of in vitro evidence of the safety of each herbal compound
myrcene was recommended with caution. To the extent tested at concentrations below 50 μM in C2C12 myotubes
of our knowledge, our work is the first report of the cyto- and HepG2 cells. These compounds are found in a wide
toxicity of acetyleugenol in these cell lines. Since each variety of herbs and spices and can easily be incorporated
TOLMIE ET AL. 11

into one's daily life. In this study, we identified potentially 12. Powers AC, Niswender KD, Evans-Molina C. In: Jameson J,
effective herbal compounds that can be used as alterna- Fauci AS, Kasper DL, Hauser SL, Longo DL, Loscalzo J, eds.
tives to the existing widely prescribed α-amylase and Diabetes Mellitus: Diagnosis, Classification, and Pathophysiol-
ogy. Harrison's Principles of Internal Medicine. 20th. New York:
α-glucosidase inhibitors. Using herbs and spices has sev-
McGraw-Hill; 2018. https://accesspharmacy.mhmedical.com/
eral advantages, including its widespread availability, content.aspx?bookid=2129&sectionid=19228 8322. Accessed:
affordability, and health benefits. April 17, 2020.
13. Brayer GD, Sidhu G, Maurus R, et al. Subsite mapping of the
A C K N O WL E D G E M E N T human pancreatic R-amylase active site through structural,
M.T. was partially funded by the University of Pretoria kinetic, and mutagenesis techniques. Biochemistry. 2000;39(16):
with a postgraduate scholarship. 4778-4791.
14. Israili ZH. Advances in the treatment of type 2 diabetes
mellitus. Am J Ther. 2011;18(2):117-152.
CONFLICT OF INTEREST
15. Ali MS, Jahangir M, ul Hussan SS, Choudhary MI. Inhibition
The authors declare no conflict of interest. of a-glucosidase by oleanolic acid and its synthetic derivatives.
Phytochemistry. 2002;60:295-299.
ORCID 16. Rouzbehan S, Moein S, Homaei A, Moein MR. Kinetics of
α-glucosidase inhibition by different fractions of three species
Morné Tolmie https://orcid.org/0000-0003-2025-1790
of Labiatae extracts: a new diabetes treatment model. Pharm
Biol. 2017;55(1):1483-1488.
R EF E RE N C E S 17. Stoilova I, Trifonova D, Marchev A, Stanchev V, Angelova G,
1. Goldenberg R. Definition, classification and diagnosis of diabe- Krastanov A. Phytochemical constituents and in vitro anti-
tes, prediabetes and metabolic syndrome. Can J Diabetes. 2013; diabetic properties of Ziziphus jujubа (Rhamnaceae) fruits. Int
37(Suppl 1):S8-S11. J Pharmacogn Phytochem Res. 2017;9(2):150-158.
2. American Diabetes Association. Diagnosis and classification of 18. Adisakwattana S, Chantarasinlapin P, Thammarat H, Yibchok-
diabetes mellitus. Diabetes Care. 2009;33(Supplement 1):S62-S69. Anun S. A series of cinnamic acid derivatives and their inhibi-
3. International Diabetes Federation. In: Karuranga S, tory activity on intestinal α-glucosidase. J Enzyme Inhib Med
Malanda B, Saeedi P, Salpea P, eds. IDF Diabetes Atlas. 9. Brus- Chem. 2009;24(5):1194-1200.
sels, Belgium: International Diabetes Federation. https://idf. 19. Asowata-Ayodele AM, Afolayan AJ, Otunola GA. Ethnobotani-
org/e-library/epidemiology-research/diabetes-atlas.html. cal survey of culinary herbs and spices used in the traditional
Accessed: March 23, 2020; 2019. medicinal system of Nkonkobe municipality, eastern cape,
4. Olokoba AB, Obateru OA, Olokoba LB. Type 2 diabetes South Africa. S Afr J Bot. 2016;104:69-75.
mellitus- a review of current trends. Oman Med J. 2012;27(4): 20. Pereira ASP, Banegas-Luna AJ, Peña-García J, Pérez-
269-273. Sánchez H, Apostolides Z. Evaluation of the anti-diabetic activ-
5. Wilke T, Boettger B, Berg B, et al. Epidemiology of urinary ity of some common herbs and spices: providing new insights
tract infections in type 2 diabetes mellitus patients: an analysis with inverse virtual screening. Molecules. 2019;24(22):4030.
based on a large sample of 456,586 German T2DM patients. 21. National Center for Biotechnology Information. PubChem Com-
J Diabetes Complications. 2015;29(8):1015-1023. pound Summary for CID 7136, Acetyleugenol. In. PubChem
6. Papatheodorou K, Banach M, Bekiari E, Rizzo M, Edmonds M. Compound Summary for CID 7136, Acetyleugenol; October
Complications of diabetes 2017. J Diabetes Res. 2018;2018: 2, 2020: National Library of Medicine; 2020. https://pubchem.
3086167. ncbi.nlm.nih.gov/compound/Acetyleugenol. Accessed: October
7. Harding JL, Pavkov ME, Magliano DJ, Shaw JE, Gregg EW. 2, 2020.
Global trends in diabetes complications: a review of current 22. Shan B, Cai YZ, Sun M, Corke H. Antioxidant capacity of
evidence. Diabetologia. 2019;62(1):3-16. 26 spice extracts and characterization of their phenolic constit-
8. Rivera-Chavez J, Gonzalez-Andrade M, Gonzalez Mdel C, uents. J Agric Food Chem. 2005;53(20):7749-7759.
Glenn AE, Mata R, Thielavins A. J and K: alpha-glucosidase 23. Shukla R, Pandey V. Role of flavonoids in management of
inhibitors from MEXU 27095, an endophytic fungus from Hint- inflammatory disorders. In: Vadnere GP, Lodhi S, eds. Bioactive
onia latiflora. Phytochemistry. 2013;94:198-205. Food as Dietary Interventions for Arthritis and Related Inflam-
9. Proenca C, Freitas M, Ribeiro D, et al. Alpha-glucosidase inhibi- matory. 2nd. Cambridge, Massachusetts: Academic Press; 2019.
tion by flavonoids: an in vitro and in silico structure-activity rela- https://www.sciencedirect.com/book/9780128138205/
tionship study. J Enzyme Inhib Med Chem. 2017;32(1):1216-1228. bioactive-food-as-dietary-interventions-for-arthritis-and-
10. He K, Shi JC, Mao XM. Safety and efficacy of acarbose in the related-inflammatory-diseases. Accessed October 2, 2020.
treatment of diabetes in Chinese patients. Ther Clin Risk 24. Shankar E, Goel A, Gupta K, Gupta S. Plant flavone apigenin: an
Manag. 2014;10:505-511. emerging anticancer agent. Curr Pharmacol Rep. 2017;3(6):423-446.
11. Proença C, Freitas M, Ribeiro D, et al. Evaluation of a flavo- 25. Das AB, Goud VV, Das C. Phenolic compounds as functional
noids library for inhibition of pancreatic α-amylase towards a ingredients in beverages. Value-Added Ingredients and Enrich-
structure–activity relationship. J Enzyme Inhib Med Chem. ments of Beverages. India: Elsevier; 2019. 285–323. https://
2019;34(1):577-588. www.sciencedirect.com/book/9780128166871/value-added-
12 TOLMIE ET AL.

ingredients-and-enrichments-of-beverages. Accessed: Septem- 40. Vichai V, Kirtikara K. Sulforhodamine B colorimetric assay for
ber 29, 2020. cytotoxicity screening. Nat Protoc. 2006;1(3):1112-1116.
26. Lee J, Lee DG, Park JY, Chae S, Lee S. Analysis of the trans- 41. Ademiluyi AO, Oboh G. Soybean phenolic-rich extracts inhibit
Cinnamic acid content in Cinnamomum spp and commercial key-enzymes linked to type 2 diabetes (alpha-amylase and
cinnamon powder using HPLC. J Agr Chem Environ. 2015;04 alpha-glucosidase) and hypertension (angiotensin I converting
(04):102-108. enzyme) in vitro. Exp Toxicol Pathol. 2013;65(3):305-309.
27. Lin LZ, Mukhopadhyay S, Robbins RJ, Harnly JM. Identifica- 42. Luyen NT, Binh PT, Tham PT, et al. Wedtrilosides A and B,
tion and quantification of flavonoids of Mexican oregano two new diterpenoid glycosides from the leaves of Wedelia
(Lippia graveolens) by LC-DAD-ESI/MS analysis. J Food Com- trilobata (L.) Hitch. with α-amylase and α-glucosidase inhibi-
post Anal. 2007;20(5):361-369. tory activities. Bioorg Chem. 2019;85:319-324.
28. Clarke S. Families of compounds that occur in essential oils. 43. Friesner RA, Banks JL, Murphy RB, et al. Glide: a new
Essential Chemistry for Aromatherapy. London: Churchill Liv- approach for rapid, accurate docking and scoring. 1. Method
ingstone; 2008. https://www.sciencedirect.com/book/ and assessment of docking accuracy. J Med Chem. 2004;47(7):
9780443104039/essential-chemistry-for-aromatherapy. 1739-1749.
Accessed: April 27, 2020. 44. Sharma R, Kothapalli R, Van Dongen AM, Swaminathan K.
29. Duke JA. Handbook of Phytochemical Constituents of GRAS Chemoinformatic identification of novel inhibitors against
Herbs and Other Economic Plants. Boca Raton, FL: CRC Press; Mycobacterium tuberculosis L-aspartate alpha-decarboxylase.
1992. https://phytochem.nal.usda.gov/phytochem/search/list. PLoS One. 2012;7(3):1-7.
Accessed: October 2, 2020. 45. Ntie-Kang F. An in silico evaluation of the ADMET profile of
30. Izawa K, Amino Y, Masanori K. Human–environment interac- the StreptomeDB database. SpringerPlus. 2013;2(353):1–11.
tions – taste. Comprehensive Natural Products II. Vol 4. Japan: 46. QikProp 3. 5 User Manual [Schrodinger Press]. 2012. https://
Elsevier; 2010:631-671. https://www.sciencedirect.com/ www.schrodinger.com. Accessed: April 27, 2020.
referencework/9780080453828/comprehensive-natural- 47. Jain PP, Degani MS, Raju A, Ray M, Rajan MGR. Rational
products-ii. Accessed May 2, 2020. drug design based synthesis of novel arylquinolines as anti-
31. Neveu V, Perez-Jiménez J, Vos F, et al. Phenol-explorer: an tuberculosis agents. Bioorg Med Chem Lett. 2013;23:6097-
online comprehensive database on polyphenol contents in 6105.
foods. Database. 2010. https://doi.org/10.1093/database/ 48. Rohini K, Shanthi V. Discovery of potent Neuraminidae inhibi-
bap024. http://phenol-explorer.eu (free access). tors using a combination of pharmacophore-based virtual
32. The Metabolomics Innovation Centre: FooDB (Version 1.0). screening and molecular simulation approach. Appl Biochem
Database, Canada: The Metabolomics Innovation Centre. 2017. Biotechnol. 2018;184:1421-1440.
https://foodb.ca (free access), 49. Rahimzadeh M, Jahanshahi S, Moein S, Moein MR. Evaluation
33. Peter KV. Handbook of Herbs and Spices. 2nd ed. Cambridge: of alpha- amylase inhibition by Urtica dioica and Juglans regia
Woodhead Publishing; 2012. https://www.sciencedirect.com/ extracts. Iran J Basic Med Sci. 2014;17:465-469.
book/9780857090393/handbook-of-herbs-and-spices. Accessed: 50. Poovitha S, Parani M. In vitro and in vivo α-amylase and
April 12, 2020. α-glucosidase inhibiting activities of the protein extracts from
34. Balachandran P, Parthasarathy V, Kumar ATV. Isolation of two varieties of bitter gourd (Momordica charantia L.). BMC
compounds from Sargassum wightii by GCMS and the molecu- Complement Altern Med. 2016;16(S1):185.
lar docking against anti-inflammatory marker COX2. Int Lett 51. Son HU, Lee SH. Comparison of alpha-glucosidase inhibition
Chem Phys Astron. 2016;63:1-12. by Cudrania tricuspidata according to harvesting time. Biomed
35. Subramaniyan V, Palani M, Srinivasan P, Kumar Singh S. Rep. 2013;1(4):624-628.
Novel ligand-based docking; molecular dynamic simula- 52. Etxeberria U, de la Garza AL, Campión J, Martínez JA,
tions; and absorption, distribution, metabolism, and excre- Milagro FI. Antidiabetic effects of natural plant extracts via
tion approach to analyzing potential acetylcholinesterase inhibition of carbohydrate hydrolysis enzymes with emphasis
inhibitors for Alzheimer's disease. Jf Pharm Anal. 2017;8(6): on pancreatic alpha amylase. Expert Opin Ther Targets. 2012;16
413-420. (3):269-297.
36. Protein Preparation. Protein Preparation. Glide 6.7 User Man- 53. Galati G, O'Brien PJ. Potential toxicity of flavonoids and other
ual. Schrödinger Press; 2015. http://gohom.win/ManualHom/ dietary phenolics: significance for their chemopreventive and
Schrodinger/Schrodinger_2015-2_docs/glide/glide_user_ anticancer properties. Free Radic Biol Med. 2004;37(3):287-303.
manual.pdf. Accessed: October 27, 2019. 54. Neiro EL, Machado-Santelli M. Cinnamic acid induces apopto-
37. Dizdaroglu Y, Albay C, Arslan T, et al. Design, synthesis and tic cell death and cytoskeleton disruption in human melanoma
molecular modeling studies of some pyrazole derivatives as car- cells. J Exp Clin Cancer Res. 2013;32(1):31.
bonic anhydrase inhibitors. J Enzyme Inhib Med Chem. 2019;35 55. Şahin S, Tezcan G, Demir C, Tunca B, ÇeÇener G, Egeli Ü.
(1):289-297. Cytotoxic activity of Rosmarinic acid isolated from Prunella
38. Banerjee K, Gupta U, Gupta S, et al. Molecular docking of vulgaris L. and Prunella grandiflora L. in different tumor cells.
glucosamine-6-phosphate synthase in Rhizopus oryzae. Biomed Trak Univ J Nat Sci. 2017;18(1):9-13.
Inform. 2011;7(6):285-290. 56. Orlando JB, Silva BO, Pires-Cunha CL, Hiruma-Lima CA,
39. Oki T, Matsui T, Osajima Y. Inhibitory effect of α-glucosidase Gaivao IOM, Maistro EL. Genotoxic effects induced by beta-
inhibitors varies according to its origin. J Agric Food Chem. myrcene following metabolism by liver HepG2/C3A human
1999;47:550-553. cells. J Toxicol Environ Health A. 2019;82(3):176-185.
TOLMIE ET AL. 13

57. Han Y, Chin Tan TM, Lim LY. In vitro and in vivo evaluation 64. Pȩkal A, Dróżdż P, Biesaga M, Pyrzynska K. Screening of the anti-
of the effects of piperine on P-gp function and expression. oxidant properties and polyphenol composition of aromatised green
Toxicol Appl Pharmacol. 2008;230(3):283-289. tea infusions. J Sci Food Agric. 2012;92(11):2244-2249.
58. Paarakh PM, Sreeram DC, SS D, Ganapathy SP. In vitro cyto-
toxic and in silico activity of piperine isolated from Piper
nigrum fruits Linn. In Silico Pharmacol. 2015;3(1):9.
59. Mclver LA, Tripp J. Acarbose [StatPearls Publishing] 2020. SU PP O R TI N G I N F O RMA TI O N
https://www.ncbi.nlm.nih.gov/books/NBK493214/. Accessed: Additional supporting information may be found online
April 27, 2020. in the Supporting Information section at the end of this
60. Balfour JA, McTavish D. Acarbose: an update of its pharmacology article.
and therapeutic use in diabetes mellitus. Drugs. 1993;46(6):1025-1054.
61. Coniff RF, Shapiro JA, Robbins D, et al. Reduction of glyco-
sylated hemoglobin and postprandial hyperglycemia by aca- How to cite this article: Tolmie M, Bester MJ,
rbose in patients with NIDDM: a placebo-controlled dose- Apostolides Z. Inhibition of α-glucosidase and
comparison study. Diabetes Care. 1995;18(6):817-824.
α-amylase by herbal compounds for the treatment
62. Pirički AP, Moslavac T, Vugrinec M. Acceptability of parsley tea
of type 2 diabetes: A validation of in silico reverse
by adolescents. Glasnik Zastite Bilja. 2010;33(1):46-53.
63. Mani F, Braga CP, Novelli ELB, Sforcin JM. Influence of clove docking with in vitro enzyme assays. Journal of
tea (Syzygium aromaticum) on body weight and biochemical Diabetes. 2021;1–13. https://doi.org/10.
parameters of rats subjected to ethanol consumption and absti- 1111/1753-0407.13163
nence. Med Chem. 2012;2(4):81-85.