Journal of Advanced Research 24 (2020) 545–555

Contents lists available at ScienceDirect

Journal of Advanced Research

journal homepage: www.elsevier.com/locate/jare

Molecular networking aided metabolomic profiling of beet leaves using

three extraction solvents and in relation to its anti-obesity effects
Nesrine M. Hegazi a,⇑, Rasha A. Radwan b, Sherein M. Bakry a, Hamada H. Saad a,c
a
Phytochemistry and Plant Systematics Department, Division of Pharmaceutical Industries, National Research Centre, PO Box 12622, Cairo, Egypt
b
Biochemistry Department, Faculty of Pharmacy, Sinai University-Kantara Branch, El Ismailia, 41611, Egypt
c
Department of Pharmaceutical Biology, Pharmaceutical Institute, Eberhard Karls University of Tübingen, PO Box 72074, Tübingen, Germany

g r a p h i c a l a b s t r a c t

a r t i c l e i n f o a b s t r a c t

Article history: In the present study, the efficiency of three different solvents (H2O, acidified H2O, and 70% Methanol) for
Received 12 May 2020 metabolites extraction from the leaves of sugar beet (Beta vulgaris subsp. vulgaris var. rubra) was inves-
Accepted 1 June 2020 tigated along with their inhibitory activity on pancreatic a-amylase and lipase for obesity management.
Available online 3 June 2020
The metabolic profile of the three extracts was analyzed by ultra-performance liquid chromatography
(UPLC) coupled with electrospray ionization high-resolution mass spectrometric (ESI-HRMS-MS). Mass
Keywords: spectrometry-based molecular networking was employed to aid in metabolites annotation and for the
Beta vulgaris
visual investigation of the known metabolites and their analogues. The study led to the tentative identi-
UPLC-HRMS-MS
Molecular networking
fication of 45 metabolites including amino acids, purine derivatives, phenolic acids, flavonoids, fatty
a- amylase acids, and an alkaloid, articulating 24 compounds as a first time report from beet leaves along with 2
Pancreatic lipase new putatively identified compounds: a flavone feruloyl conjugate (39) and a malonylated acacetin digly-
coside (40). The three extracting systems exhibited comparable efficiency for pulling out the secondary
metabolites from the beet leaves. The in vitro study supported this finding and demonstrated that the
three extracts inhibited the activity of both pancreatic a-amylase and lipase enzymes with no significant

Peer review under responsibility of Cairo University.

⇑ Corresponding author at: Phytochemistry and Plant Systematics Department, National Research Centre, Dokki, Cairo, Egypt.
E-mail address: [email protected] (N.M. Hegazi).

https://doi.org/10.1016/j.jare.2020.06.001
2090-1232/Ó 2020 The Authors. Published by Elsevier B.V. on behalf of Cairo University.
This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
546 N.M. Hegazi et al. / Journal of Advanced Research 24 (2020) 545–555

difference observed regarding the percentage of the inhibition of the enzymes. Conclusively, the extrac-
tion protocol has a minimal effect on the anti-obesity properties of beet leaves.
Ó 2020 The Authors. Published by Elsevier B.V. on behalf of Cairo University. This is an open access article
under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Introduction beet) leaves using UPLC-PDA-(±) ESI-HRMS/MS . Mass spectral

similarity networking via the global natural products social
Obesity is a growing universal issue and regarded as one of molecular networking platform (GNPS) was employed for the
the utmost intimidations to global health in this era [1]. Statisti- visualization and exploration of the tandem mass spectrometry
cally, more than two billion adults worldwide are overweight data and to aid in dereplication of known metabolites and their
among which 650 million of them are clinically obese [2]. possible analogues. Simultaneously, the diminishing effect of the
Currently, there are two distinct classes of drugs for obesity three extracts of B. vulgaris leaves on pancreatic a-amylase and
management [3]. The first class acts by the inhibition of pancre- lipase enzymes were inspected for their potential use for obesity
atic lipase (orlistat), and hence reducing intestinal fat absorption management.
[4]. The second one counts on suppressing the appetiteor anorec-
tics represented by sibutramine [5]. However, both medications Materials and methods
have several drawbacks such as hypertension, xerostomia,
decreased bowel movements, headache, and sleeplessness Plant material
[6–8]. Accordingly, there is always a persistent demand for an
alternative approach to effective and safe obesity control [9]. Sugar beet leaves (Beta vulgaris, subsp. vulgaris var. rubra) were
Even though the widespread daily practice of consuming various obtained from local merchandise in Giza, Egypt in March 2018. A
dietary supplements for obesity surveillance, their effectiveness plant sample was verified by Prof. Dr. Salwa Khawatshi, professor
and safety are poorly investigated and have not yet persuasively of Taxonomy, National Research Centre, Cairo, Egypt. A voucher
proved in such context [1]. specimen (M132) was deposited in the herbarium of the National
Digestive enzymes such as pancreatic a-amylase and lipase are Research Centre (CAIRC), Phytochemistry and Plant Systematics
known to be responsible for the oligosaccharides and triacylglyc- Department, National Research Centre, Dokki, Cairo, Egypt. The
erols hydrolysis into simple molecules. Naturally occurring leaves were cleaned thoroughly from the fine dust and debris with
polyphenols are firmly ascertained to inhibit such digestive bi-distilled H2O, ground to paste, and frozen at
20 °C for further
enzymes which can govern the food caloric content by reducing analysis.
its absorption and prolonging the digestive process. By this means,
a bodyweight reduction could be an achievable task and conse- Chemicals
quently yield significant health improvement [10].
A substantial number of reports has accentuated the nutritional Methanol (HPLC grade) was obtained from Fisher Chemical, UK.
value and health benefits of functional foods. Special attention was Sodium formate (MS grade) was provided by Honeywell Fluka,
given to vegetables and fruits for their overall wellbeing effects Germany. All other chemicals for phytochemical analysis and bio-
including reduced incidence of metabolic, cardiovascular disorders, logical assays were purchased from Sigma-Aldrich (Merck, USA).
and cancers [11–13].
Beta vulgaris L. species belong to the Amaranthaceae family and Preparation of the extracts
is widely distributed throughout Asian Turkey, the Mediterranean,
and Europe [14]. The species is recognized in traditional medicines To extract the beet leaves metabolites, three different solvents
as an immunostimulant and as adjuvant therapy in cancer treat- were used: bi-distilled H2O (a), acidified bi-distilled H2O with 1%
ment [15]. The Red beetroot ‘Beta vulgaris, subsp. vulgaris var. (w:v) ascorbic acid (b) and 70:30 MeOH: bi-distilled H2O (c). For
rubra’ is widely used in folk medicines, besides its widespread each extracting solvent, 500 ml were used to sonicate 250 g of
use in the food industry as food colouring [16]. Red beets represent leaves for 30 min at 50 °C, each in triplicates. Extracts were then
a rich source of natural nitrates which proved helpful in diseases filtered, evaporated under reduced pressure, and finally lyophilized
associated with low bioavailability of NO including hypertension and kept frozen at
20 °C.
and endothelial function [17]. Several in vivo and in vitro reports
demonstrated innumerable biological activities for beetroots Sample preparation for HPLC profiling and MS analyses
including antioxidant and anti-inflammatory [18], antimicrobial
[19], stimulant of the hematopoietic and immune systems; renal The lyophilised samples of the three extracts (50 mg each) were
and hepatic protective, antioxidant, anti-inflammatory and antitu- dissolved in 70% MeOH (HPLC-grade) with sonication (10 mins),
mor properties [20]. The antioxidant and antidiabetic potential of then centrifuged. Aliquots were then evaporated under reduced
beetroots advocates its possible use in obesity management pressure followed by freeze-drying for 48 hrs. For HPLC profiling,
[21,22]. 10 mg of the dried extracts were dissolved in 500 ml MeOH
Beetroot phytochemicals were previously extracted with vari- (HPLC-grade), each in triplicates, and 10 ml were injected. Mean-
ous solvents including distilled water [23], 80% alcohol [24,25], while, triplicates of 1 mg in 250 ml MeOH (MS-grade) were pre-
and acidified ethanol with citric acid [26,27]. Nevertheless, signif- pared for MS analysis consuming 5 ml as an injection volume in
icant differences were observed in the content of the individual the UPLC-MS analysis.
compounds as affected by the solvent choice [25].
In this context and motivated by the public usage of beets HPLC profiling
and its documented biological properties, this study fundamen-
tally focused on the exploration of the phytochemical con- Profiling of the obtained extracts triplicates was achieved using
stituents of three different extracts (aqueous, acidified aqueous an HPLC system composed of Waters 1525 Binary Pump with a
with 1% ascorbic acid and 70% methanol) of B. vulgaris (sugar 7725i Rheodyne injection port; a Kromega Solvent Degasser:
 N.M. Hegazi et al. / Journal of Advanced Research 24 (2020) 545–555 547

Waters 996 Photodiode Array Detector; and an Aeris peptide XB- Sirius + CSI: FingerID 4.0.1 for the prediction of elemental compo-
C18 (5 mm, 250  4.6 mm, Phenomenx). ACN (solvent A) and sition (C, H, N, O, S, P) and molecular structure database search
H2O + 0.1% TFA (solvent B) were used for the gradient elution of with m/z tolerance set to 20 ppm using online Pubchem database
the analytes at a steady flow rate of 0.7 ml/min, with an injection [30,31] and DEREP-NP database which was manually integrated
volume of 5 ll. A non-uniform gradient was employed: for the ini- into the software [32].
tial 8 min, 10% A, then 30% A for 12 min, 35% A for additional
15 min, 50% A for 10 min, followed by 100% A for 10 min, 20% A Alpha-amylase inhibition
for 1 min, and lastly 10% A for the last 4 min. The detection wave-
lengths were 238, 280, and 336 nm. The a-amylase bioassay was performed as described by Miller
with some modifications [33,34]. In brief, potato starch was
UPLC-HRMS-MS analysis mixed with 20 mmol L
1 sodium phosphate buffer with
6.7 mmol L
1 sodium chloride to obtain a 0.5% w/v starch solu-
MaXis-4G instrument (Bruker Daltonics, Bremen, Germany) tion. For the preparation of the enzyme solution, 25.3 mg of
attached to an Ultimate 3000 HPLC (Thermo Fisher Scientific) were a-amylase (10 U mg -1) was stirred in 100 ml of cold distilled
used for HR-MS analysis. The HPLC-method was (0.1% FA in H2O as water. Extracts solutions were made by dissolving them in a buf-
solvent A and MeOH as solvent B), an isocratic gradient of 10% B for fer to afford a final concentration ranging from 1000 mg mL
1 to
10 min, 10% to 100% B in 30 min, 100% B for an additional 15 min, 31.25 mg mL
1. Sodium potassium tartrate solution (12.0 g of
using a flow rate of 0.3 ml/min; 5 ll injection volume and UV sodium potassium tartrate tetrahydrate in 8.0 ml of 2 M NaOH)
detector (UV/VIS) wavelength monitoring at 336, 280 and and 96 mmol L
1 of 3, 5 dinitro salicylic acid solution were
238 nm. The separation was performed on a Nucleoshell 2.7 mm mixed for the colourimetric reagent. The control (acarbose as a
150  2 mm column (Macherey-Nagel), and the range for MS positive control) and the extracts solutions with concentrations
acquisition was m/z 50–1800. A capillary voltage of 4500 V, nebu- (75–600 mg/mL) were mixed individually with the starch solu-
lizer gas pressure (nitrogen) of 2 (1.6) bar, ion source temperature tion and allowed to react with a-amylase in alkaline conditions
of 200 °C, the dry gas flow of 9 L/min source temperature, and at 25◦C. Maltose production was quantified by the reduction of
spectral rates of 3 Hz for MS1 and 10 Hz for MS2 were used. For 3,5-dinitrosalicylic acid to 3-amino5-nitrosalicylic acid. The
MS/MS fragmentation, the 10 most intense ions per MS1 were cho- reaction was detected at 540 nm using ELX 808 (Bio-Tek
sen for subsequent CID with stepped CID energy applied. The Instrumental, Italy). The following equation was employed for
employed parameters for tandem MS were applied following the calculation of the percentages of inhibition: 100- [{A sample/
[28]. Sodium formate was used as a calibrant, and with acquired A control  100].
data calibrated using a Bruker-developed script.

Data analysis and preprocessing Pancreatic lipase inhibition

Data visualization was performed using Bruker Daltonics Data Following Conforti protocol [35], the inhibition of pancreatic
Analysis 4.4, while Metaboscape 3.0 (Bruker Daltonics) was used lipase was evaluated. Pancreatic lipase aqueous solution (1
for molecular features selection. Raw data files were imported into mgmL
1) was made from type II crude porcine enzyme. The solu-
MetaboScape 3.0 for the entire data treatment and pre-processing. tion of 4-nitrophenyl octanoate (NPC) (5 mmol L
1) in dimethyl
T-ReX 3D (Time aligned Region Complete eXtraction) algorithm sulfoxide was prepared as the substrate. The reaction mixture
was used for retention time alignment. It automatically detects was executed as follows: 100 ml of 5 mmol L
1 NPC, 4 ml of Tris-
and combines isotopes, adducts, and fragments intrinsic to the HCl buffer (pH = 8.5), 100 ml of extract solutions in concentrations
same compound into one feature. All detected features were dis- from (12–100 mg/ mL), and 100 ml of enzyme solution. Next, the
played as a bucket table with their Rt, measured m/z, molecular prepared solution was incubated at 37 °C for 25 min before adding
weight, detected ions, and their intensity in each sample [29]. the substrate. For the negative control, the same volume of
The Bucket table was created with intensity threshold 10e3 and dimethyl sulfoxide was added instead of the extract solution. The
10e4 for negative and positive ionization modes, respectively. The absorbance was measured in cuvettes at 412 nm using ELX 808
retention time range was set from 1 to 35 min and the mass range (Bio-Tek Instrumental, Italy). A blank sample without the enzyme
from 140 to 1800 m/z. was measured for each extract. For comparison, orlistat was used
as a positive control at a final concentration of 20 mgmL
1.The fol-
Molecular networking and metabolites annotation lowing formula was used to calculate the percentage of inhibition:
100- [{A sample/ A control}  100].
The features list of the three extraction solvents was exported
from Metaboscape as two single MGF files for both of the positive Statistical analysis
and negative measurements. Both MGF files were uploaded sepa-
rately to the GNPS online platform where two molecular networks Both enzymatic experiments were carried out in triplicates. The
were generated with the online workflow (GNPS 2.0). A molecular results were given as mean values and standard deviations (SDs).
network was created with a cosine score above 0.65 and 0.7 for The differences among the various extracts were analyzed using
positive and negative modes, respectively. The minimum number a one-way analysis of variance (ANOVA) followed by Tukey’s hon-
of matched fragment ions was adjusted to 6. Further edges est significant difference post hoc test at p values < 0.05 [36]. A lin-
between two nodes were kept in the network if (and only if) each ear regression curve was employed to calculate the IC50 (extract
of the nodes appeared in each other’s respective top 10 most sim- concentration that inhibits 50% of the enzyme activity). The hori-
ilar nodes. The network spectra were searched against GNPS’ spec- zontal axis displayed the concentrations of the extracts, which
tral libraries using a minimum of 6 matched fragments for spectral were (from 75 to 600 lg/ml) for a-amylase and (from 60 to
matching. Cytoscape 3.5.1 was used for molecular network 600 lg/ml) and pancreatic lipase. While the vertical axis presented
visualization. the percentage of inhibition [37]. All analyses were calculated
Manual putative structures identification was achieved by sub- using SPSS v. 22.0 (IBM, Chicago, USA). Microsoft Excel 2010 was
mitting the preprocessed MS2.mgf output file from Metaboscape to used for graph construction.
548 N.M. Hegazi et al. / Journal of Advanced Research 24 (2020) 545–555

Results and discussion and edge attributes were used, so that the colour of the node corre-
sponds to the origin of the sample (used extraction solvent). Nodes
HPLC and UPLC-HRMS/MS metabolites profiling of the extraction were displayed as a pie chart to reflect the semi-relative abundance
solvents of each ion in the three extracts. The edges were employed using
either their thickness as a caliber to the similarity between the con-
This study aimed to chart the comparable efficiency of three dif- nected nodes or as a measure of the mass shifts between the associ-
ferent extracting solvents in terms of as many as attainable ated variants during the dereplication process.
metabolites from beet leaves. The HPLC profiling of the three dif- Two networks were separately generated for the positive and
ferent extracts including the triplicates exhibited insignificant dif- negative ionization modes using the GNPS 2 platform (Figs. 2 &
ferences as seen in Suppl. Fig. S1 suggesting that the three solvent 3). The positive network contained 107 nodes compromising 8
(s) mixtures could be similarly and efficiently used for extracting clusters and 71 self-looped nodes whereas the negative afforded
the beet leaves metabolites under the studied conditions. a total of 88 nodes arranged into 10 clusters and 49 self-looped
Driven by the high similarity in the HPLC readings of the differ- nodes. The created networks allowed the visual inspection of the
ent extracts, a holistic comparative secondary metabolic analysis of different compound families, analogues and aided in isomers dis-
the red beet leaves for 3 different extraction solvents was necessi- crimination. Clusters A in positive network, A & B in the negative
tated via reversed-phase (RP-C18) UPLC-PDA-ESI-HRMS/MS in network comprised the C-glycosidic flavones and their acylated
both ionization modes. Overlaid base peak chromatogram (BPC) derivatives (Figs. 2 & 3). Nevertheless, cluster A (Fig. 3) in the
for the three solvents is shown in Fig. 1. positive network was the one mostly used for the exploration of
C-glycosidic flavones complemented with their fragmentation
Molecular networking-based categorization of beet leaves metabolites pattern in the negative ionization mode. While in the negative
network (Fig. 2), clusters C, D, and E represented the flavonol
Molecular networking depends on the fact that structurally sim- O-glycosides, phenolic acids, and fatty acids, respectively. Other
ilar metabolites have similar MS/MS fragmentation patterns allow- identified metabolites appeared as self-looped nodes within both
ing for the instantaneous visual investigation of identical of the networks.
molecules, analogues, or compound families [38]. In other words,
it groups mass spectrometric data by mining spectral similarity UPLC-HRMS/MS metabolites annotation
between the MS/MS fragmentation patterns of different but struc-
turally related metabolites. Within the network, each node corre- Putative metabolites annotation was counted based on their
sponds to one consensus MS/MS spectrum, and is typically labeled retention times, molecular formula, UV absorption maxima, and
with the neutral precursor mass. Nodes having common fragmenta- their fragmentation pattern in comparison to previously reported
tion patterns are connected with edges (lines). In this study, node data aided with the molecular networking investigation and GNPS

Fig. 1. Overlaid base peak chromatograms of the three extracts of beet leaves in the negative & positive ionization modes. Red colour: H2O extract, green: acidified H2O
extract and blue 70% MeOH extract.
 N.M. Hegazi et al. / Journal of Advanced Research 24 (2020) 545–555 549

Fig. 2. Full molecular networking created using MS/MS data in negative mode from extracts of the leaves of B. vulgaris subsp. vulgaris var rubra. Nodes are labelled with
parent mass. The network is displayed as pie chart with blue, red and orange colours representing distribution of theprecursor ion intensity in the acidified H2O, H2O, 70%
MeOH extracts correspondingly. While, nodes with bold edges are nodes which matched with GNPS spectral libraries. * Substitution may differ.

spectral library search combined with the suggested fragmentation Purine derivatives: Guanosine 2 and methylbutenyl isoguanine 5
trees by Sirius. In total, 45 compounds belonging to different were also annotated as stated earlier [41,42].
metabolite classes were tentatively identified including amino Alkaloids: A self-looped node in the positive spectral network
acids, purine derivatives, phenolic acids, flavonoids, fatty acids, was found for the first time abundantly in the H2O extract of beet
and an alkaloid (Table 1). More than 50% of the annotated features leaves as a pyrrolizidine alkaloid, trachelanthine 6. It exhibited a
(24 compounds, chiefly phenolic acids, and C-glycosidic flavones) molecular ion at m/z 302.1956 [M + H]+ with a formula of
are reported for the first time from the beet leaves along with 2 C15H27NO5 and a characteristic product ion at m/z 184.17
new putatively identified metabolites as a flavone feruloyl conju- [M + H-C6H14O2] in accordance with Hama & Strobel, 2019 [43].
gate (39) and malonylated acacetin glycoside (40).
Amino acids and derivatives: Glutamine 1 and tryptophan 9 [39] Phenolic acids and their derivatives:
were detected only in the positive mode whilst tryptophan N-
glycoside 8 [40] was observed in the negative mode and appeared The negative network using MS/MS data managed to gather the
in the spectral network as self-looped nodes. Their identifications closely related phenolic acid glycosides together as cluster D
were based on their retention times, UV absorbance, molecular for- encompassing compounds 4, 7, 13, and 14 (Fig. 2). Dihydroxy ben-
mula, and their fragmentation patterns. zoic acid methyl ester-O-hexoside, 4 had a parent ion at m/z
550 N.M. Hegazi et al. / Journal of Advanced Research 24 (2020) 545–555

Fig. 3. Full molecular networking created using MS/MS data in the positive mode from extracts of the leaves of B. vulgaris subsp. vulgaris var rubra. Nodes are labelled with
parent mass. The network is displayed as pie chart with blue, red and orange colours representing distribution of theprecursor ion intensity in the acidified H2O, H2O, 70%
MeOH extracts correspondingly. While, nodes with bold edges are nodes which matched with GNPS spectral libraries. * Substitution may differ.

329.087 [M-H]-, C14H18O9 and a base peak at 167 [M-H-162]-, due ment ions at m/z 267 [M-H-176]- suggesting glucoronic acid moi-
to the loss of the hexose moiety [44]. Meanwhile, compound 7, m/z ety loss and m/z 193 [M-H-176-73]- for the glucoronosyl glycerol
359.0986 [M-H]- as C15H20O10, was connected to 4 with a mass dif- allowed the characterization of compound 16 as feruloyl glu-
ference of 30 Da as a possible extra OCH2. The fragments observa- curonosyl glycerol, formerly found in B. vulgaris cell cultures but
tion at m/z 197 [M-H-162]- and m/z 153 [(M-H-162)-44]- referring not from beet leaves [50]. An additional N-containing phenyl-
to the elimination of hexose moiety and CO2, respectively eased its propanoid was found to be a N-feruloyl-methoxytyramine 41 with
assignment as syringic acid hexoside 7 [45]. Ferulic acid hexoside a molecular formula of C19H21NO5, m/z 342.1352 [M
H]-, along-
13 and sinapic acid hexoside 14 were also allied in the same cluster side two characteristic peaks at m/z 178, and 148 as previously
exhibiting in their negative MS2 distinctive fragment ions at m/z described in the literature [51].
193 and m/z 223 owing to identical losses of O-hexosyl moiety,
respectively [22,46].
While other phenolic features 3 & 11 were marked as self- C- glycosidic flavones and their acylated derivatives:
looped nodes in the negative network due to their different frag-
mentation behavior which was reflected in their putative identities Flavone C-glycosides were the most abundant metabolite class
either as a non– or di-glycosylated species in contrast to 4, 7, 13 in the beet leaves. They compromised cluster A in the positive
and 14 assumed to be mono-glycosylated variants. Thus, com- and A & B in the negative spectral networks (Figs. 2 & 3). MS/MS
pound 3 was assigned to be 2-hydroxyl-cinnamaldehyde, with a could nicely differentiate between O-glycosyl and C-glycosyl
molecular ion at m/z 149.0600 [M + H]+, and a molecular formula derivatives. Neutral losses of 162 / 132 or 146 corresponds to O-
C9H8O2 with fragment ions formed by losses of H2O (-18 Da) and hexoside / pentoside or deoxyhexoside, respectively. Whereas,
CO (-28 Da) [47]. However, Compound 11 was annotated as hydro- the fragmentation of C-glycosylated derivatives exhibits loss of
xyl benzoyl hexosyl deoxyhexoside with m/z 445.1351[M-H]-, H2O [M-18], and cross ring cleavages of the sugar moieties
C19H26O12 supported by the characteristic product ions at m/z observed as [M-120/90] +/- for C-hexosides, [M-90/60] +/- for C-
283 [M-H-162]- for the loss of a hexose motif and m/z 137 [M-H- pentosides and [M-104/47] +/- for C- deoxyhexosides [52]. Addi-
162-146]- for the consecutive loss of a deoxyhexoside moiety tionally, neutral losses of 42 and 86 accounts for acetylation and
[48]. Assisted by spectral similarity networking, the occurrence malonylation, respectively [46]. Further, the dereplication of the
of 2-hydroxyl-cinnamaldehyde 3, dihydroxybenzoic acid methyl C- glycosidic flavones was much facilitated with the created net-
ester-O-hexoside 4, syringic acid hexoside 7 and hydroxy benzoyl works that allowed the visualization of analogues and the quick
hexosyl deoxyhexoside 11 was possible for the first time in beet discrimination of isomers, if any. The mass differences between
leaves. the connected nodes accentuated the underlying structure modifi-
Ferulic acid conjugates appeared as scattered four nodes in the cations. For instances, mass differences of 14 Da or 30 Da were
negative network implying their peculiar fragmentation behavior. often due to a sugar replacement (pentoside to hexoside and vice
Compound 12 was assigned as feruloyl pentosyl-hexoside based versa), while 42 Da and 86 Da were characteristic of acetylated
on its molecular ion at m/z 487.1453 [M-H]-, C21H28O13 and its and malonylated descendants, respectively. While the acetylated
characteristic base peak at m/z 193 after losing pentosyl and hex- and malonylated derivatives were connected together with 44 Da
osyl moieties [46]. Compound 15, m/z 459.1506 [M-H]- with a for- mass difference.
mula of C20H28O12 as non-previously reported feature in B. vulgaris, The first eluted flavone glycoside was (iso)vitexin 60 ’-O-
was dereplicated as butane-tetraol (O-feruloyl) -O- hexoside (pae- pentoside 10, which had a parent ion at m/z 565.1570 [M + H]+,
derol B) evidenced by a daughter ion at m/z 193 [(M-H-162)-104]- C26H28O14 and an abundant ion at m/z 433 [M + H-132]+ for the loss
thanks to the losses of a hexose and butan-tetraol moieties [49]. of a pentose unit and m/z 313 [(M + H-132)-120]+ characteristic for
Similarly, a parent ion [M-H]- at m/z 443.1199 with two key frag- the cross ring cleavage of C-glycosyl derivatives [52].
 N.M. Hegazi et al. / Journal of Advanced Research 24 (2020) 545–555 551

Table 1
Identified metabolites in the different extracts of the leaves of B. vulgaris subsp. vulgaris var. rubra.

Compound Rt Proposed structure [M + H]+ [M
H]- MS2 Molecular formula Extraction References
number (min) (error in ppm) solvent
a b C
p
1 1.76 Glutamine 175.1197 129 C6H14N4O2, (3.9) [39]
p p
2 2.85 Guanosine 284.0996 152 C10H13N5O5, (2.2) GNPS
libraries,
[41]
p p p
3 4.1 Hydroxy-cinnamaldehyde* 149.0600 121, 130 C9H8O2, (1.9) [47]
p p p
4 4.13 Dihydroxybenzoic acid methyl ester-O- 329.087 167, 152, 123 C14H18O9, (0.92) [44]
hexoside *
p
5 4.2 Methylbutenyl isoguanine* 220.1185 130 C10H13N5O5, (3.6) [42]
p p p
6 4.57 Trachelanthine* 302.1956 184, 123 C15H27NO5, (1.71) [43]
p p p
7 4.6 Syringic acid O- hexoside* 359.0986 197, 153 C15H20O10, (0.63) [45]
p p
8 5.03 Tryptophan N-hexoside 365.1356 203 C17H22N2O7, (1.59) [40]
p p
9 5.9 Tryptophan 205.0977 188, 146, 118 C11H12N2O2, (2.5) GNPS
libraries,
[39]
p p p
10 5.8 60 ’-O-pentosyl (iso)vitexin 565.157 433, 313 C26H28O14, (3.2) GNPS
libraries,
[52]
p p p
11 6.11 Hydroxybenzoyl O-hexosyl-O- 445.1351 283, 137 C19H26O12, (0.32) [48]
deoxyhexoside*
p p p
12 6.91 Ferulic acid O-pentosyl-hexoside 487.1453 193 C21H28O13, (0.84) [46]
p p p
13 7.04 Ferulic acid O-hexoside 355.1035 193, 178, C16H20O9, (0.2) [22,46]
p p p
14 7.48 Sinapic acid -O- hexoside 385.1145 223, 208 C17H22O10, (1.2) [46]
p p p
15 7.63 Butane-tetraol-O-feruloyl-O- 459.1506 193, 175 C20H28O12, (0.13) [49]
glucopyranoside. paederol B*
p p p
16 8.62 Ferulic acid-O-glucuronosyl glycerol 443.1199 267, 249, 193, C19H24O12, (1.37) [50]
175, 134
p p p
17 9.04 Apigenin di- C-hexoside 595.1513 593.1515 473, 413, 313 C27H30O15, (0.34) [53]
p p p
18 9.12 Apigenin di- C-hexoside isomer 595.1595 593.1515 473, 413, 313 C27H30O15, (1.77) [53]
p p p
19 9.3 Apigenin -C- -deoxyhexoside O – hexoside* 579.1717 577.1567 415, 311 C27H30O14, (0.72) GNPS
libraries,
[52]
p p p
20 9.41 Apigenin C -deoxyhexoside O – hexoside 579.1702 577.1567 415, 311 C27H30O14, (0.72) GNPS
isomer* libraries,
[52]
p p p
21 9.48 20 ’ O-pentosyl (iso)vitexin 563.1410 413, 293 C26H28O14, (0.6) GNPS
libraries,
[54]
p p p
22 9.50 (iso)vitexin-60 ’-malonyl-20 ’-O- pentoside 651.1561 649.1403 563, 455, 311 C29H30O17, (0.725) [46]
p p p
23 9.6 Acetyl 20 ’ O-pentosyl (iso)vitexin* 605.1523 563, 413, 293 C28H30O15 (0.1) [55]
p p p
24 9.92 Apigenin -C- hexoside 433.1136 431.1054 313, 283 C21H20O10, (2.04) [22]
p p p
25 9.99 Isorhamnetin di- O-hexoside 641.172 639.1565 317, 315 C28H32O17, (1.35) [46]
p p p
26 10.15 malonyl (iso)vitexin 200 -O-hexoside 681.1599 679.152 575, 455, 311, C30H32O18, (0.32) [46]
293
p p p
27 10.23 Acetyl apigenin- C-pentoside- C-hexoside* 605.1521 563, 545, 455, C28H30O15, (1.61) [52]
353, 311
p p p
28 10.25 (iso)vitexin 60 ’-acetyl 20 ’-O- hexoside* 635.1626 455, 413, 293 C29H32O16 (0.2) [52]
p p p
29 10.26 (iso)vitexin-60 ’-malonyl-20 ’-O- pentoside 651.1556 649.1415 563, 455, 311 C29H30O17, (1.38) [46]
p p p
30 10.39 Isorhamnetin -O- pentoside O-hexoside* 609.1465 315 C27H30O16 (0.2) GNPS
libraries,
[58]
p p p
31 10.45 Isorhamnetin -O- pentoside O- hexoside 609.1461 315 C27H30O16 (0.2) GNPS
isomer* libraries,
[58]
p p p
32 10.48 malonyl 600 -O-deoxyhexosyl -C-hexosyl 663.1633 545, 311 C30H32O17, (0.9) [52]
apigenin *
p p p
33 10.62 Acetyl apigenin- C-pentoside- C-hexoside* 605.1515 563, 545, 455, C28H30O15, (0.44) [52]
353, 311
p p p
34 11.02 Malonyl (iso)vitexin * 519.114 517.0986 413, 341, 311 C24H22O13, (0.07) [57]
p p p
35 11.2 Acacetin di- C-hexoside* 607.1669 487, 293 C28H32O15 (2.1) [56]
p p p
36 11.3 malonyl 600 -O-deoxyhexosyl -C-hexosyl 663.1561 545, 311 C30H32O17, (0.85) [52]
apigenin *
p p p
37 11.8 Acacetin-C-hexoside
200 -O-pentoside* 579.1716 577.1565 427, 307 C27H30O14, (1.9) [39]
p p p
38 12.06 Dihydroxy dimethoxy flavone -C-hexoside 607.1636 457, 337 C28H32O15 (2.3) [52]
200 - O pentoside*
p p p
39 12.61 Tetrahydroxy monomethoxy flavone -O- 815.2035 639, 315 C38H40O20, (0.15)
feruloyl hexosyl hexoside**
p p
40 12.8 Acacetin-C-malonyl hexoside-O-pentoside** 665.1723 533, 447, 327 C30H32O17, (1.75)
p p p
41 14.27 N -feruloyl-methoxytyramine 342.1352 178, 327, 148 C19H21NO5, (1.44) [51]
p p
42 17.3 Trihydroxy-octadecadienoic acid 327.2178 211 C18H32O5, (0.4) [59]

(continued on next page)

552 N.M. Hegazi et al. / Journal of Advanced Research 24 (2020) 545–555

Table 1 (continued)

Compound Rt Proposed structure [M + H]+ [M
H]- MS2 Molecular formula Extraction References
number (min) (error in ppm) solvent
a b C
p p
43 18.6 Pinellic acid, trihydroxyoctadecenoic acid 329.2336 171 C18H34O5, (0.9) [59]
p p p
44 18.74 trihydroxyoctadecenoic acid isomer* 329.233 229, 171, 127 C18H34O5, (0.17) [68]
p p p
45 19.26 trihydroxy-octadecadienoic acid 327.2187 239, 171, 229 C18H32O5, (0.71) [59]

Solvent a: H2O extract, b: acidified H2O extract, c: 70% MeOH: H2O.

*Compounds reported for the first time in beet leaves.
**Compounds not previously described in nature.

With the aid of feature-based molecular networking (GNPS 2) 86 Da to 35. Consequently, it was assigned as apigenin C-
platform, isomers could be discriminated by their retention time hexoside possessing a parent ion at m/z at 433.1139 [M + H]+
difference within the same cluster reflecting its advantage over and 431.1056 [M-H]- and a molecular formula of C21H20O10. The
the classical one. Apigenin 6,8-di-C-hexoside isomers, compounds observation of m/z 343 [M + H-90]+ and 313 [M + H-120]+ as a sup-
17 & 18 were observed as two separate connected nodes with portive product ions in the positive mode clarified the cross-ring
the same parent ion within cluster A in both networks (Figs. 2 & fissions in the C-hexoside moiety [22].
3) indicating a possible isomeric structural features. The extracted Whereas, compound 26 with [M-H]- at 679.1518, C30H32O18
ion chromatograms furthermore confirmed the presence of two correlated to apigenin di-C-hexoside (17 & 18) with a mass differ-
isomers rendered as two discrete peaks eluted at 9.04 and ence of 86 Da signifying a probable malonylated derivative (Fig. 2).
9.12 min. with a molecular ions [M-H]- at m/z 593.1515 and However, its negative MS2 spectrum afforded a product ion at 455
595.1595 [M + H]+ corresponding to a formula of C27H30O15. The [M-H-(162)-18]– which is characteristic of 200 -O-glycosyl-C-
direct attachment of 17 & 18 to compound 10 with a mass shift glycosyl derivatives rather than the C-glycosyl ones. Other
of 30 Da as probable OCH2 extension with characteristic fragments observed ions were m/z 635 [M-H-44]- for CO2 elimination, and
at m/z 473 [M-H-120]- for the cross ring cleavage of C-hexoside 575 [M-H-86-18]- for the loss of malonyl group. Hence, it was iden-
moiety and m/z 413 (loss of an additional 60 Da) perceived the tified as malonyl (iso)vitexin-200 -O-hexoside [52].
replacement of O-pentoside into C-hexoside. Thus, both com- The negative EIC of 605 Da displayed 3 distinct peaks sharing
pounds were assigned as apigenin 6,8-di-C-glucoside, previously the same elemental composition C28H30O15. The first one eluted
identified in B. vulgaris leaves [53]. Equivalently, two additional 23 at 9.6 min. followed by two isomers 27 & 33 at 10.31 and
isomers 19 and 20 eluted at 9.30 and 9.41 min. could be recognized 10.64 min. However, their MS2 spectra were quite different such
as two nodes linked by a thick edge with a null mass difference in that compound 23, 605.1523[M-H]- showed the fragmentation
the ions constellation (Fig. 2). Their linkage by delta mass of 14 pattern of acetylated 20 ’-O-pentosyl C-glycosides with fragments
(+CH2), and 16 (-O) Da to 10, and 17 & 18, respectively propagated at m/z 563 [M-H-42]-, m/z 413 for an extra loss of 150 Da signifi-
their assignment as apigenin C-deoxyhexoside-O-hexoside isomers cant of the cleavage of 20 ’-O-pentosyl moiety followed by the loss
boosted by MS2 fragments at m/z 415 [M-H-162]- for the loss of O- of 120 Da indicating a C-glycoside flavone [55]. And hence, it was
hexose moiety and a base peak at m/z 311 [M-162-104-H]- indica- supposed to be an acetyl 20 ’-O-pentosyl (iso)vitexin which was
tive of C-deoxyhexoside moiety [52]. additionally certified by its presence in C-glycosidic flavones clus-
Within the same cluster (Fig. 2), compounds 19 & 20 were clo- ter of the spectral negative network being connected to the
sely associated to an ion feature 21 with a mass difference of 14 Da malonylated (iso)vitexin pentosides (22 & 29) with a 42 Da mass
having a parent ion at m/z 563.1409 [M-H]
and a formula of difference (Fig. 2). Inversely, the full mass spectra of 27 and 33
C26H28O14. Considering the suggested MFs difference, 14 Da could revealed di-C-glycosidic derivatives with identical fragment ions
point to CH2 that was traced up in the MSMS comparatively against at m/z 563 [M-H-42]- corresponding to the loss of acetyl group,
19 & 20 to allocate the structural variation. A dual modification m/z 455 [(M-42-18)-90-H]- from the cross ring cleavage of C-
was suggested in the attached sugars as O-pentoside rather than glycosidic moieties and m/z 353, 311 characteristics for apigenin
O-hexoside concerning the fragment observed at m/z 413 [M-H- di-C-glycosides. Consequently, they were annotated as acetylated
150]- and C- hexoside unit instead of C-deoxyhexoside correspond- apigenin C-hexoside C-pentoside isomers [52].
ing to m/z 293 for an extra loss of 120 amu. Accordingly, it was Compound 28 was found with a molecular ion of 635.1626 [M-
annotated as (iso)vitexin 20 ’-O-pentoside as previously reported H]- and a formula of C29H32O16. Its direct association to compound
in beet leaves [54]. 23 in the negative spectral network (Fig. 2) with an extra 30 Da
In parallel, compounds 22 and 29, tentatively identified as (iso) (OCH2) along with its fragmentation pattern proposed a hexoside
vitexin-60 ’-malonyl-20 ’-O- pentoside, were directly recognized as derivative rather than a pentoside one. This assumption was even
likely isomeric structures since they were visualized as two bound confirmed by its fragmentation sequence with a fragment at 455
ions sharing the same precursor ion in the same cluster within the [M-H-180]- for the loss of 20 ’-O-hexosyl moiety followed by the
positive network (Fig. 3). They exhibited molecular ion peaks at m/ loss of 42 Da at m/z 413 significant of its acetylation and finally a
z 649.1415 [M-H]-, 651.1561[M + H]+ suggesting a molecular for- fragment at m/z 293 characteristic of apigenin C-hexosides.
mula of C29H30O17. Their annotation was based on their observed Accordingly, it was identified as (iso)vitexin 60 ’-acetyl 20 ’-O-
negative MS2 spectra having fragment ions at m/z 563 [M-H-86]- hexoside [52].
accounting for the loss of a malonyl group, m/z 455 [M-H-44- Similarly, cluster A in the negative mode (Fig. 2) showed addi-
150]- for an extra loss of 20 ’-O-pentoside moiety and the character- tional distinct isomers 32 and 36 with m/z 663.15611 [M-H]- cor-
istic fragment at m/z 311 of apigenin C-hexosides [46]. related to the malonylated (iso)vitexin pentosides (22 & 29) with a
While compound 24 appeared as a significant ion, in cluster A in mass shift of 14 Da which was followed in their MS2 spectra to sug-
the positive network (Fig. 3) due to its linkage to most of the con- gest the presence of a O-deoxyhexosyl moiety instead of pentosyl.
stituting nodes. It was connected to compounds 10 with a mass This was again supported with the observed key fragment ions at
difference of 132 Da, 162 Da to 17 & 18, 146 Da to 19 & 20 and m/z 545 [M-H-86-18]- due to loss of H2O along with the malonyl
 N.M. Hegazi et al. / Journal of Advanced Research 24 (2020) 545–555 553

group, m/z 311 [M-H-146-120]- proving the loss of O-deoxyhexosyl cleavage of an O-pentosyl moieties [M + H-132]+ at m/z 447 and
moiety and cross cleavages of a C-hexosyl residue characteristic for 533 for 37 & 40, respectively were found (Fragmentation scheme
600 -O-deoxyhexosyl C-hexosyl derivatives [52]. Accordingly, they and MS2 spectrum for compound 40 are shown as Suppl. Fig. S4,
were tentatively assigned as malonyl 600 -O-deoxyhexosyl C- S5). Whereas, m/z 447 [M + H-132–89]+ was spotted later for
hexosyl apigenin congeners. compound 40 from the sequential loss of a malonyl group. Other
The EIC of m/z 607.1669 [M-H]- with a molecular formula of shared characteristic fragments at m/z 327 representing the loss
C28H32O15, unexpectedly revealed two peaks as 35 and 38 at of 120 Da as a cross ring cleavage of a C-hexoside moiety
11.20 and 12.06 min., respectively with quite different MSMS conclusively facilitated its tentative assignment as acacetin
which was further validated by the spectral network as two dis- O-pentosyl malonyl C-hexoside that was not previously reported
crete nodes. Compound 35 was found to group with the C-glyco- in nature.
sidic flavones within negative cluster B (Fig. 2). Its MS2 spectrum Among the identified known C-glycosidic flavones: apigenin C-
showed the presence of a fragment ion at m/z 487 for [M-H- deoxyhexoside-O-hexoside 19 & 20; (iso)vitexin 200 -O-(600 -malonyl)
120]- and the absence of the aglycone ion, which is typical of di- hexoside 22 & 33, acetyl apigenin C-pentoside C-hexoside 27 & 32,
C-glycosyl flavones. This evidence, along with the UV–vis spectrum malonyl 600 O-deoxyhexosyl-C-hexosyl apigenin 31 & 36, malonyl
led to its assignment as dihydroxy monomethoxy flavone di-C- (iso)vitexin 34, acacetin di-C-hexoside 35, acacetin C-hexoside -
hexoside; acacetin di-C-hexoside [56]. Differently, compound 38 200 -O-pentoside 37 and dihydroxy dimethoxy flavone C-hexoside
which was observed as a self-looped node and displayed a different 200 -O-pentoside 38 are reported for the first time to occur in beet
fragmentation pattern in which daughter ions at m/z 457 [M-H- leaves along with the putatively annotated new compounds
132-18]- characteristic for the cleavage of 200 -O-pentosyl moiety tetrahydroxy monomethoxy flavone-O-feruloyl hexosyl hexoside
and m/z 337 [(M-H-150)-120]- typical for C-hexosyl derivatives 39, and acacetin O-pentosyl C-malonyl hexoside 40.
were monitored. As a result, it was annotated as dihydroxy Flavonol O-glycosides as a distinct compound family were laid
dimethoxy flavone C-hexoside 200 -O-pentoside [52]. out as cluster C having 3 closely related nodes in the negative spec-
Other flavone C-glycosides like 34 and 37 were also pictured as tral network, exclusively (Fig. 2). The first of which was isorham-
self-looped nodes in the negative mode while in the positive mode netin di-O- hexoside 25, exhibited a molecular ion at m/z
they were assembled within cluster A (Figs. 2 & 3). Such ion scat- 641.172 [M + H]+ and 639.1565 [M
H]- corresponding to a chem-
tering in the negative mode could be explained in the light of the ical formula of C28H32O17 and fragment ions at m/z 317 [M + H-
fewer generated fragments which are the key parameter for the 324]+, 315 [M-H-324]- attributed to the loss of two hexose moi-
ions assembling. Such ions assemblage in the positive mode prop- eties [46]. Two positional diglycosylated isorhamnetin isomers 30
agated the annotation of these ions through the observed mass & 31, m/z 609.1465 [M-H]-, were in a direct connection to 25 with
shifts of the connecting edges. For example, compound 34, m/z a 30 Da difference (OCH2) proposing the replacement of one of the
517.0986 [M-H]- C24H22O13, was found to be in a direct connectiv- hexoses with a pentose moiety. This was further affirmed with the
ity with the formerly annotated apigenin C-hexoside 24 with a detected formula of C27H30O16 and their MS2 spectra which
mass difference equals to 86 Da which could be possibly envi- showed an abundant ion at m/z 315 [M-H-132-162]- from the loss
sioned as an extra malonylation. The existence of confirmatory of O-pentosyl and O-hexosyl moieties. Hence, they were tenta-
fragment ions as m/z 413 [M-H-86-18]- revealing the loss of a mal- tively assigned as not previously reported isorhamnetin O-
onyl unit together with m/z 311 with 341 characteristics for the C- pentoside O-hexoside isomers in beet leaves [58].
hexoside moiety breakdown of the flavone led to its assignment as Fatty acids: additionally, the ESI-MSMS spectra showed the
malonyl(iso)vitexin [57]. presence of polyhydroxylated fatty acids in the negative spectral
Likewise, compound 37 was depicted as structurally related network as an individual family (Fig. 2). The observed four nodes
congener to 21 with an additive CH2 taking into account the sug- cluster represented two trihydroxy-octadecadienoic acid isomers
gested molecular formulas. Manual inspection of the comparative 42 & 45 along with two trihydroxy-octadecenoic acid isomers 43
MS2 spectra delivered a typical fragmentation behavior to 21 with & 44. They exhibited molecular ions at m/z 327.2178 [M
H]-
a constant mass shift of 14 Da, m/z 427 & 413 [M-H-150]
and m/z (C18H32O5) and m/z 329.2336 (C18H34O5), respectively [59] with
307 & 293 [(M-H-150)-120]- permitting the allocation of such MS2 spectra, in which sequential elimination of H2O [M
H
18]-
extra 14 Da on the aglycone moiety which led to its identification followed by multiple losses of CH2 [M
H
14]- were spotted.
as acacetin C-hexoside-200 -O-pentoside [56].
A unique feature 39 appeared as a self-looped node in the neg-
ative network (Fig. 2) at m/z 815.2035 [M-H]- having a chemical
formula C38H40O20. The MS2 spectrum showed fragment ions at
m/z 639 [M-H-177]- pointing to a potential loss of a feruloyl moiety
and m/z 315 [(M-H-177)-324]- for the loss of di-O-hexoside units
(Fragmentation scheme and MS2 spectrum are shown as Suppl.
Fig. S2, S3). Thus, it was tentatively identified as tetrahydroxy
monomethoxy flavone-O-feruloyl hexosyl hexoside as a possible
new natural product.
While the positive spectral network (Fig. 3) revealed an ion fea-
ture 40 with a parent ion at m/z 665.1723 [M + H]+ . It was directly
linked to a previously annotated acacetin O-pentosyl C-hexoside
37 and with close relatedness to compound 22 in the positive
molecular network, (iso) vitexin 20 ’-O-(60 ’-malonyl) hexoside with
a mass difference of 14 Da (CH2) proving the methylation of the
aglycone moiety. When considering the generated molecular for-
mula of C30H32O17 along with the recognized mass difference of
86 Da, a probable malonylation of 37 was assumed. Comparing Fig. 4. IC50 values for the inhibition of pancreatic a- amylase and lipase enzymes
the fragmentation patterns of the two compounds additionally with beet leaves extracts. Data are presented as (Mean ± S.D.) *: indicates significant
validated this assumption where shared fragment ions from the difference at p < 0.05.
554 N.M. Hegazi et al. / Journal of Advanced Research 24 (2020) 545–555

metabolites belonging to diverse classes among which 24 were

not formerly described in beet leaves including a putative descrip-
tion of 2 new metabolites, namely a flavone feruloyl derivative (39)
and a malonylated acacetin di-glycoside (40). Full structural
elucidation of these novel metabolites using other spectroscopic
techniques i.e., NMR post isolation should now follows.
Simultaneously, the inhibitory activities of the three extracts were
tested against pancreatic a-amylase and lipase enzymes for its
possible use for obesity management. All three extracting solvents
were proved statistically efficient and maintained the desired bio-
logical activity. These results may help to expand the potential
health benefits of beet leaves for obese people as well as being a
rich source of nutraceuticals. However, additional studies are
essential to determine its biological activity in more complex sys-
tems and various food matrices both in in vitro and in vivo diges-
Fig. 5. Pancreatic a- amylase and lipase inhibition of beet leaves extracts at 300 ug/
ml and 100 ug/ml, respectively. Data are presented as (Mean ± S.D.) *: indicates
tion processes.
significant difference at p < 0.05.
Compliance with ethics requirements

Pancreatic a- amylase and lipase inhibition by different This article does not contain any studies with human or animal
extracts of beet leaves subjects

Lately, numerous studies have proved the potentiality of plant Declaration of Competing Interest
extracts to inhibit the digestive enzymes a-amylase and pancreatic
lipase [1,34,60]. Inhibition of carbohydrate-hydrolyzing enzymes, The authors declare that they have no known competing finan-
such as a-amylase may aid in obesity management by delaying cial interests or personal relationships that could have appeared
glucose absorption and overall carbohydrate digestion time. Lipase to influence the work reported in this paper.
inhibition is another important tactic for obesity prevention
through the suppression of triglycerides absorption.
Appendix A. Supplementary data
As shown in Figs. 4 & 5, the three extracts inhibited the
a-amylase activity with the lowest IC50 for the aqueous extract
Supplementary data to this article can be found online at
150.35 ± 3.7 mg/ml while the other two extracts exhibited quite
https://doi.org/10.1016/j.jare.2020.06.001.
similar IC50 values. Likewise, the three extracts inhibited the pan-
creatic lipase equivalently as proved with their IC50 values with
References
an average of 24.47 mg/ml. No significant difference was observed
between the three extraction solvents regarding the percentage [1] Birari RB, Bhutani KK. Pancreatic lipase inhibitors from natural sources:
of the enzyme inhibition. This can be capitalized that all three unexplored potential. Drug Discovery Today 2007;12(19–20):879–89.
designed extraction schemes are efficient enough for grabbing [2] Organistaion, W.H. https://www.who.int/news-room/fact-sheets/detail/
obesity-and-overweight 2016.
the active compounds to exert the desirable biological potency. [3] Kakkar AK, Dahiya N. Drug treatment of obesity: current status and future
(Dose-response curves for the enzymes’ activity are displayed prospects. European Journal of Internal Medicine 2015;26(2):89–94.
as Suppl. Fig. S6 & S7. While the dose–response curve of the [4] Drew BS, Dixon AF, Dixon JB. Obesity management: update on orlistat.
Vascular health and risk management 2007;3(6):817.
enzymes’ inhibition with the three extracts are provided as Suppl.
[5] Tziomalos K, Krassas GE, Tzotzas T. The use of sibutramine in the management
Fig. S8-S13). of obesity and related disorders: an update. Vascular health and risk
The observed results could be ascribed to the phenolic metabo- management 2009;5:441.
lite(s) present in the beet leaves. Naturally occurring phenolics are [6] de Simone G, D’Addeo G. Sibutramine: balancing weight loss benefit and
possible cardiovascular risk. Nutrition, Metabolism and Cardiovascular
reported to regulate carbohydrate and lipid metabolism, by modi- Diseases 2008;18(5):337–41.
fying the activity of digestive enzymes [61,62]. More specifically, [7] Karamadoukis L et al. An unusual complication of treatment with orlistat. Clin
several investigations proved the potential health benefits of natu- Nephrol 2009;71(4):430–2.
[8] Slovacek L, Pavlik V, Slovackova B. The effect of sibutramine therapy on
ral flavonoids in obesity management [63] especially the C- glyco- occurrence of depression symptoms among obese patients. Nutrition,
sidic flavones which exerted potent inhibitory action on pancreatic Metabolism and Cardiovascular Diseases 2008;18(8):e43–4.
lipase as previously reported by [64]. More particular, vitexin or [9] Yun JW. Possible anti-obesity therapeutics from nature–A review.
phytochemistry, 2010,;71(14–15):1625–41.
extracts containing vitexin were proved to exert anti-obesity activ- [10] Worsztynowicz P et al. Pancreatic a-amylase and lipase inhibitory activity of
ity through different mechanisms of action [65,66]. polyphenolic compounds present in the extract of black chokeberry (Aronia
To the best of our knowledge, this is the first report for the ben- melanocarpa L.). Process Biochem 2014;49(9):1457–63.
[11] Andrew C. The encyclopedia of medicinal plants. A Dorling Kindersley Book;
eficial effect of beet leaves for the prevention and/or treatment of 1996.
obesity. Although the in vivo antidiabetic and lipid-lowering poten- [12] Tardío J et al. Ethnobotanical and food composition monographs of selected
tial of beet leaves extracts were previously proved by [22,67]. Mediterranean wild edible plants, in Mediterranean Wild Edible. Plants. 2016;
Springer:273–470.
[13] Septembre-Malaterre A, Remize F, Poucheret P. Fruits and vegetables, as a
source of nutritional compounds and phytochemicals: Changes in bioactive
Conclusion compounds during lactic fermentation. Food Res Int 2018;104:86–99.
[14] Nikan M, Manayi A. Beta vulgaris L, in Nonvitamin and Nonmineral Nutritional.
Supplements. 2019;Elsevier:153–8.
The effectiveness of three extracting systems (H2O, 1% acidified [15] Sacan O, Yanardag R. Antioxidant and antiacetylcholinesterase activities of chard
H2O with ascorbic acid and 70%MeOH) of beet leaves for pulling (Beta vulgaris L. var. cicla). Food and chemical toxicology 2010;48(5):1275–80.
out the secondary metabolites from the beet leaves were compared [16] Ninfali P, Angelino D. Nutritional and functional potential of Beta vulgaris cicla
and rubra. Fitoterapia 2013;89:188–99.
via holistic UPLC-HRMS/MS analysis. Molecular networking [17] Clifford T et al. The potential benefits of red beetroot supplementation in
propagated metabolite profiling allowed the dereplication of 45 health and disease. Nutrients 2015;7(4):2801–22.
 N.M. Hegazi et al. / Journal of Advanced Research 24 (2020) 545–555 555

[18] Vulić JJ et al. In vivo and in vitro antioxidant effects of beetroot pomace using ultrafiltration and liquid chromatography–tandem mass spectrometry. J
extracts. J Funct Foods 2014;6:168–75. Pharm Biomed Anal 2012;59:96–101.
[19] Jha R, Gupta RK. Development of energy drink containing Aegle marmelos, [42] Lan MS et al. Chemical constituents of Phyllanthus reticulatus. Helv Chim Acta
Rubia cordifolia, Phyllanthus emblica and Beta vulgaris and its phytochemical, 2010;93(11):2276–80.
nutritive and antimicrobial analysis. Journal of Pharmacognosy and [43] Hama JR, Strobel BW. Pyrrolizidine alkaloids quantified in soil and water using
Phytochemistry 2016;5(1):186–93. UPLC-MS/MS. RSC Adv 2019;9(52):30350–7.
[20] Martinez RM et al. Anti-inflammatory activity of betalain-rich dye of Beta [44] Peng H et al. Major chemical constituents and antioxidant activities of
vulgaris: effect on edema, leukocyte recruitment, superoxide anion and different extracts from the peduncles of Hovenia acerba Lindl. Int J Food Prop
cytokine production. Arch Pharmacal Res 2015;38(4):494–504. 2018;21(1):2135–55.
[21] Zielińska-Przyjemska M et al. In vitro effects of beetroot juice and chips on [45] Pan Z et al. Non-targeted metabolomic analysis of orange (Citrus sinensis [L.]
oxidative metabolism and apoptosis in neutrophils from obese individuals. Osbeck) wild type and bud mutant fruits by direct analysis in real-time and
Phytotherapy Research: An International Journal Devoted to Pharmacological HPLC-electrospray mass spectrometry. Metabolomics 2014;10(3):508–23.
and Toxicological Evaluation of Natural Product Derivatives 2009;23 [46] Vissers A et al. Enzymatic browning in sugar beet leaves (Beta vulgaris L.):
(1):49–55. influence of caffeic acid derivatives, oxidative coupling, and coupled oxidation.
[22] El-Ghffar EAA et al. HPLC-ESI-MS/MS analysis of beet (Beta vulgaris) leaves J Agric Food Chem 2017;65(24):4911–20.
and its beneficial properties in type 1 diabetic rats. Biomed Pharmacother [47] Guan H et al. Identification of the Chemical Constituents of an Anti-Arthritic
2019;120:109541. Chinese Medicine Wen Luo Yin by Liquid Chromatography Coupled with Mass
[23] Ben Haj Koubaier, H., et al., Betalain and phenolic compositions, antioxidant Spectrometry. Molecules 2019;24(2):233.
activity of Tunisian red beet (Beta vulgaris L. conditiva) roots and stems [48] Garcez WS, Yoshida M, Gottlieb OR. Benzylisoquinoline alkaloids and flavonols
extracts. International journal of food properties, 2014. 17(9): p. 1934-1945. from Ocotea vellosiana. Phytochemistry 1995;39(4):815–6.
[24] Nowacki L et al. Betanin-enriched Red beetroot (Beta vulgaris L.) extract [49] Chin Y-W et al. Two new phenylpropanoid glycosides from the aerial parts of
induces apoptosis and autophagic cell death in MCF-7 cells. Phytother Res Paederia scandens. Bull Korean Chem Soc 2010;31(4):1070–2.
2015;29(12):1964–73. [50] Bokern M et al. Ferulic acid conjugates and betacyanins from cell cultures of
[25] Kujala T, Loponen J, Pihlaja K. Betalains and phenolics in red beetroot (Beta Beta vulgaris. Phytochemistry 1991;30(10):3261–5.
vulgaris) peel extracts: extraction and characterisation. Zeitschrift für [51] Song Q et al. Potential of hyphenated ultra-high performance liquid
Naturforschung C 2001;56(5–6):343–8. chromatography-scheduled multiple reaction monitoring algorithm for
[26] Attia GY, Moussa M, Sheashea E. Characterization of red pigments extracted from large-scale quantitative analysis of traditional Chinese medicines. RSC Adv
red beet (Beta vulgaris L.) and its potential uses as antioxidant and natural food 2015;5(71):57372–82.
colorants. Egyptian. Journal of Agricultural Research 2013;91(3):1095–110. [52] Farag MA et al. Comparative metabolite profiling and fingerprinting of genus
[27] Azizah AN et al. The Use of Natural Dyes from Beetroot Skin Extract (Beta Passiflora leaves using a multiplex approach of UPLC-MS and NMR analyzed by
Vulgaris) as Teaching Material on Cell Division for Senior High School chemometric tools. Anal Bioanal Chem 2016;408(12):3125–43.
Students. Indonesian Journal on Learning and Advanced Education (IJOLAE) [53] Simirgiotis M et al. The Passiflora tripartita (Banana Passion) fruit: A source of
2019;2(1):20–6. bioactive flavonoid C-glycosides isolated by HSCCC and characterized by
[28] Garg N et al. Mass spectral similarity for untargeted metabolomics data HPLC–DAD–ESI/MS/MS. Molecules 2013;18(2):1672–92.
analysis of complex mixtures. Int J Mass Spectrom 2015;377:719–27. [54] Ninfali P et al. Characterization and biological activity of the main flavonoids
[29] Olmo-García L et al. Exploring the Capability of LC-MS and GC-MS Multi-Class from Swiss Chard (Beta vulgaris subspecies cycla). Phytomedicine 2007;14(2–
Methods to Discriminate Virgin Olive Oils from Different Geographical 3):216–21.
Indications and to Identify Potential Origin Markers. Eur J Lipid Sci Technol [55] Zheleva-Dimitrova D et al. Chemical characterization with in vitro biological
2019;121(3):1800336. activities of Gypsophila species. J Pharm Biomed Anal 2018;155:56–69.
[30] Dührkop K et al. Searching molecular structure databases with tandem mass [56] Barreca D et al. Kumquat (Fortunella japonica Swingle) juice: Flavonoid
spectra using CSI: FingerID. Proc Natl Acad Sci 2015;112(41):12580–5. distribution and antioxidant properties. Food Res Int 2011;44(7):2190–7.
[31] Böcker S, Dührkop K. Fragmentation trees reloaded. Journal of cheminformatics [57] Alamgir KM et al. Acylflavonoid C-glycosides from Eleusine coracana.
2016;8(1):5. Phytochem Lett 2008;1(2):111–4.
[32] Zani CL, Carroll AR. Database for rapid dereplication of known natural products [58] Wang D-M et al. A new isorhamnetin glycoside and other phenolic compounds
using data from MS and fast NMR experiments. J Nat Prod 2017;80 from Callianthemum taipaicum. Molecules 2012;17(4):4595–603.
(6):1758–66. [59] Hamberg M, Olsson U. Efficient and Specific Conversion of 9-Lipoxygenase
[33] Miller GL. Use of dinitrosalicylic acid reagent for determination of reducing Hydroperoxides in the Beetroot. Formation of Pinellic Acid. Lipids 2011;46
sugar. Anal Chem 1959;31(3):426–8. (9):873–8.
[34] Marrelli M et al. Inhibition of key enzymes linked to obesity by preparations [60] Dastjerdi ZM, Namjoyan F, Azemi ME. Alpha amylase inhibition activity of
from Mediterranean dietary plants: effects on a-amylase and pancreatic lipase some plants extract of Teucrium species. European Journal of Biological
activities. Plant Foods Hum Nutr 2013;68(4):340–6. Sciences 2015;7(1):26–31.
[35] Conforti F et al. Wild Mediterranean dietary plants as inhibitors of pancreatic [61] Ikarashi N et al. The inhibition of lipase and glucosidase activities by acacia
lipase. Phytother Res 2012;26(4):600–4. polyphenol. Evidence-Based Complementary and Alternative Medicine
[36] Uysal S et al. A comparative in vitro and in silico study of the biological 2011;2011.
potential and chemical fingerprints of Dorcycinum pentapyllum subsp. [62] Tadera K et al. Inhibition of a-glucosidase and a-amylase by flavonoids. J Nutr
haussknechtii using three extraction procedures. New J Chem 2017;41 Sci Vitaminol 2006;52(2):149–53.
(22):13952–60. [63] Kawser Hossain M et al. Molecular mechanisms of the anti-obesity and anti-
[37] Chiocchio I et al. Screening of a hundred plant extracts as tyrosinase and diabetic properties of flavonoids. Int J Mol Sci 2016;17(4):569.
elastase inhibitors, two enzymatic targets of cosmetic interest. Ind Crops Prod [64] Lee EM et al. Pancreatic lipase inhibition by C-glycosidic flavones isolated from
2018;122:498–505. Eremochloa ophiuroides. Molecules 2010;15(11):8251–9.
[38] Wang M et al. Sharing and community curation of mass spectrometry data [65] Yang JH et al. Potent anti-inflammatory and antiadipogenic properties of
with Global Natural Products Social Molecular Networking. Nat Biotechnol bamboo (Sasa coreana Nakai) leaves extract and its major constituent
2016;34(8):828. flavonoids. J Agric Food Chem 2017;65(31):6665–73.
[39] Stintzing FC, Schieber A, Carle R. Identification of betalains from yellow beet [66] Wang F et al. Vitexin alleviates lipopolysaccharide-induced islet cell injury by
(Beta vulgaris L.) and cactus pear [Opuntia ficus-indica (L.) Mill.] by high- inhibiting HMGB1 release. Mol Med Rep 2017;15(3):1079–86.
performance liquid chromatography
electrospray ionization mass [67] Al-Dosari M et al. Effect of Beta vulgaris L. on cholesterol rich diet-induced
spectrometry. J Agric Food Chem 2002;50(8):2302–7. hypercholesterolemia in rats. Farmacia 2011;59:669–78.
[40] Diem S, Bergmann J, Herderich M. Tryptophan-N-glucoside in fruits and fruit [68] Tao Y et al. Ultrafiltration coupled with high-performance liquid
juices. J Agric Food Chem 2000;48(10):4913–7. chromatography and quadrupole-time-of-flight mass spectrometry for
[41] Liu S et al. Characterization of compounds and potential neuraminidase screening lipase binders from different extracts of Dendrobium officinale.
inhibitors from the n-butanol extract of Compound Indigowoad Root Granule Anal Bioanal Chem 2015;407(20):6081–93.