Journal of Food Composition and Analysis 114 (2022) 104744

Contents lists available at ScienceDirect

Journal of Food Composition and Analysis

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

HPLC-ESI-HRMS and chemometric analysis of carobs polyphenols –

Technological and geographical parameters affecting their
phenolic composition
Atalanti Christou a, Ana B. Martinez-Piernas b, Ioannis J. Stavrou c, Juan F. Garcia-Reyes b,
Constantina P. Kapnissi-Christodoulou a, *
a
Department of Chemistry, University of Cyprus, 1678 Nicosia, Cyprus
b
Analytical Chemistry Research Group (FQM-323), Department of Physical and Analytical Chemistry, University of Jaen, 23071 Jaen, Spain
c
Department of Life Sciences, European University Cyprus, 2404 Nicosia, Cyprus

A R T I C L E I N F O A B S T R A C T

Keywords: In the present study, the effects of the ripening stage, processing method, and geographical origin were inves­
Ceratonia siliqua L. tigated through the analysis of carobs and derived products by use of high-performance liquid chromatography-
HPLC-ESI-HRMS electrospray ionization high-resolution mass spectrometry (HPLC-ESI-HRMS), in combination with chemometric
Polyphenols
analysis. The distribution of secondary metabolites across the fruit was also established through the character­
Carob extract
ization of the phenolic pattern of different carob parts. It was observed that carob seeds represent a better po­
Carob pulp
Carob seeds tential source of polyphenols since significantly higher levels were demonstrated. The content of polyphenols of
Carob products both pulp and seeds was diminished upon maturation. Therefore, unripe pods were considerably more enriched
Chemometric analysis in polyphenols. Noticeable changes were also observed during carob processing, especially over the thermal
Principal component analysis (PCA) treatment of carob products. In regard to the effect of geographical origin, a multivariate statistical approach was
Partial least squares discriminant analysis (PLS- employed to study the relationship between the phenolic composition of carobs and their growing regions.
DA)

1. Introduction obtained after the removal of seeds (kibbles), has a lower commercial
value, and it is considered an agri-food waste material (Cavdarova and
Carob (Ceratonia siliqua L., Leguminosae family), a native tree of the Makris, 2014; Van Rijs and Fogliano, 2020). Although it has been
Mediterranean basin, is of considerable economic importance from both neglected for the last decades and survived mainly as a livestock feed,
industrial and nutritional points of view (Vekiari et al., 2011). Although the carob pulp has recently returned to the global marketplace as a
it is a typical tree of the Mediterranean area, with Spain, Italy, Morocco, potential functional and nutraceutical food ingredient (Antoniou et al.,
Portugal, Greece, Turkey, and Cyprus being the principal producing 2020; Kokkinofta et al., 2020). Nowadays, carob pulp is used as a raw
countries, recently, it has spread to other Mediterranean-like regions, material for the production of a variety of added-value biological
such as California, Arizona, Mexico, Chile, Argentina, Australia, South products, including carob sirups or molasses and roasted or unroasted
Africa, and India (Stavrou et al., 2018; Vekiari et al., 2011). It is a carob powders (a cocoa substitute, free of caffeine and theobromine).
well-known tree for its nutritional and health-promoting edible fruits, Due to its unique chemical composition, beneficial health effects,
also referred to as carob pods or carobs. They are comprised of two and potential application in the development of food-derived products,
major parts, the pulp and the seeds in a roughly 90:10 w/w ratio (Chait carob has gained considerable attention over the last few years. Carob
et al., 2020; Kyratzis et al., 2021). Currently, the seeds are considered pod is a rich source of different valuable components such as carbohy­
the most valuable part of the pods exploited industrially for the pro­ drates, dietary fibers, minerals, amino acids, and polyphenols (Christou
duction of locust bean gum (LBG, E410), a widely used natural food et al., 2021a). Most of the biological properties and health-promoting
thickening and stabilizing agent extracted from the seed endosperm effects of carobs and carob’s derived products have been attributed
(Antoniou et al., 2020; Kyratzis et al., 2021). The carob pulp, which is mainly to the presence of polyphenols (Chait et al., 2020; Ydjedd et al.,

* Corresponding author.
E-mail address: [email protected] (C.P. Kapnissi-Christodoulou).

https://doi.org/10.1016/j.jfca.2022.104744
Received 6 May 2022; Received in revised form 6 July 2022; Accepted 7 July 2022
Available online 9 July 2022
0889-1575/© 2022 Elsevier Inc. All rights reserved.
A. Christou et al. Journal of Food Composition and Analysis 114 (2022) 104744

2017a). Phenolic compounds are a very heterogeneous group of plant substances might be proposed as a source of analytical data to tackle
metabolites that includes simple phenolic acids, flavonoids, stilbenes, characterization, classification, and authentication issues, often with the
and lignans (Stavrou et al., 2018). In general, these substances occur in assistance of chemometric methods (Barbosa et al., 2020b;
plant cells, mainly as glycosides and/or associated with various organic Izquierdo-Llopart and Saurina, 2019). The application of sophisticated
acids and/or polymerized molecules of high molecular weights, like data analysis procedures, the so-called chemometrics, has been proven a
tannins (Daglia, 2012). Polyphenols may protect cell constituents from versatile and valuable tool for the discrimination and classification of a
oxidative damage and, thus, limit the risk of several degenerative dis­ variety of fruit-based products according to their region and variety
eases that are believed to be associated with oxidative stress (Scalbert (Makris et al., 2006). Chemometric methods, such as non-supervised
et al., 2005). Carob and its derived products have demonstrated principal component analysis (PCA), a powerful method for the reduc­
anti-carcinogenic, anti-cardiovascular, and neuroprotective properties tion of data dimensionality, and supervised partial least squares
apparently related to their antioxidant constituents (Manach et al., discriminant analysis (PLS-DA), useful in maximizing the separability of
2004; Vekiari et al., 2011). known categories, are the most commonly used data analysis techniques
Several analytical methodologies have, so far, been proposed for the for establishing the origin of different food products.
characterization of the phenolic profile of carobs and carob-based In the present study, the effect of geographical origin, ripeness, and
products. These include (i) spectrophotometric methods for the esti­ processing method on the phytochemical composition of carobs and
mation of total phenolic content and (ii) chromatographic and electro­ carob-based products was assessed using HPLC-ESI-HRMS. The impact
phoretic methods used for the identification and quantification of of traditional processing technologies, such as thermal and mechanical
individual phenolic components (Christou et al., 2021b). Among them, treatment, on polyphenolic compounds was evaluated through the
high-performance liquid chromatography (HPLC) with either analysis of various carob-derived products. Precise data available on the
Ultraviolet-Visible (UV) detection or coupled to mass spectrometry (MS) phenolic pattern of carob-based products are at present insufficient.
is by far the method of choice for the qualitative and quantitative Thus, the phenolic composition of roasted/unroasted carob powder,
analysis of carob’s phenolics (Christou et al., 2021b; Stavrou et al., roasted carob pulp, and carob sirup treated under different temperatures
2018). Nevertheless, the diverse chemical nature of phenols, which was determined and compared. At the same time, the changes in the
differ in polarity and size (from simple phenolic acids to tannins), along phenolic composition of carob pods during the fruit ripening process
with the low concentration levels of the substances found in food were investigated. Although the great influence of the ripening process
matrices, make LC coupled to MS or tandem MS (LC-MS(/MS)) the most on carob’s phenolic composition has already been highlighted in pre­
appropriate technique for the structural characterization and determi­ vious investigations (Benchikh et al., 2016; Christou et al., 2021b; Farag
nation of both low- and high-molecular-weight polyphenols (Barbosa et al., 2019; Kyriacou et al., 2021; Ydjedd et al., 2017b), there is no data
et al., 2020a). In the last few years, high-resolution MS (HRMS), with available that compares the phenolic profile of both mature and
time-of-flight (TOF) and Orbitrap analyzers and electrospray ionization immature carob pulp and seeds. Through the characterization of the
(ESI) sources, has also gained wide acceptance as a highly sensitive and phenolic pattern of different carob parts (pulp and seeds) of both ripe
selective technique for the characterization and determination of phe­ and unripe stages, the distribution of secondary metabolites across the
nolics in a variety of food matrices (Barbosa et al., 2018; Lucci et al., fruit was also established.
2017). LC-HRMS has proven its excellent analytical performance by A multivariate statistical approach was successfully employed to
providing more comprehensive information in regard to the exact mo­ study the effect of geographical origin, and particularly the relationship
lecular mass, elemental composition, and detailed molecular structure between the phenolic composition of carobs, as given by HPLC-ESI-
of a given compound, and, at the same time, by allowing both targeted HRMS analysis, and their growing region. A data set consisting of 33
and non-targeted modes of analysis (Díaz-De-Cerio et al., 2018; Lucci carob samples (of 7 different geographical origins) and 10 measured
et al., 2017). variables (polyphenols markers) was evaluated using pattern recogni­
A lot of research has, so far, been performed in regard to the phenolic tion techniques, including PCA and PLS-DA. To the best of our knowl­
profile, and particularly on the identification and quantification of edge, no work has, so far, been performed to differentiate carob species
polyphenols in carob pulp, seeds, leaves, bark, and derived products according to their geographical origin employing phenolic profiles as
(Christou et al., 2021b). Different compositions were observed in the characteristic markers. Morphological characteristics (pods length/
different carob parts and, among them, carob pulp and leaves demon­ width/thickness, pulp weight, seeds number/yield/length/width/
strated the highest phenolic contents (Stavrou et al., 2018). Phenolic thickness/weight), nutritional compositions (moisture, fat, proteins,
analysis revealed high contents of phenolic acids (mainly gallic acid, sugars, dietary fibers, minerals), odor profile (volatile organic com­
including the free form and its derivatives such as methyl gallate), fla­ pounds), and IR spectra data have previously been utilized as potential
vonoids (particularly rich in flavonols such as quercetin, myricetin, markers for carob’s geographical origin discrimination (Christou et al.,
kaempferol, and their glycosidic derivatives), and tannins (with 2018; Kokkinofta et al., 2020; Krokou et al., 2020; Naghmouchi et al.,
condensed tannins composed of flavan-3-ol and flavan-3,4-diol units 2009; Sidina et al., 2009). Only Farag et al. (2019) have reported the use
being the most abundant) (Almanasrah et al., 2015; Loullis and Pina­ of secondary metabolites as chemical markers for carobs differentiation.
koulaki, 2018; Nasar-Abbas et al., 2016; Owen et al., 2003). Neverthe­ In their study though, the PCA approach appeared to be more effective
less, the literature data on the phenolic profile and content of carobs in regard to the carobs’ metabolites changes based on their ripening
demonstrated a great variation depending, not only on carob variety, stages rather than their region of origin.
gender, geographical origin, conditions of cultivation, and degree of
maturation, but also on technological practices followed such as 2. Material and methods
extraction and analytical techniques. In addition, different polyphenolic
compositions can be detected in carob products that have been treated 2.1. Chemicals and standards
under different conditions, and, as a consequence, the qualitative and
quantitative comparison of results from different studies is considered For the HPLC-ESI-MS analysis, LC-MS grade methanol (MeOH) and
extremely difficult (Christou et al., 2021b). formic acid were purchased from Sigma-Aldrich (St. Louis, MO, USA)
Taking into consideration that the polyphenolic profile of carobs and Fluka (Bucks, Switzerland), respectively. A Milli-Q-Plus water pu­
and, consequently, carob-based products is affected by several rification system (Millipore, Milford, MA) was utilized throughout the
geographical parameters, such as climatic (rainfall, temperature) and study to obtain ultra-pure water of analytical quality. The analytical
environmental (soil, altitude, sun exposure) conditions, phenolics can be standards of (+)-catechin, ( ± )-naringenin, pyrocatechol, trans-
exploited as reliable geographical origin indicators. Hence, these cinnamic acid, chlorogenic acid, 2,5-dihydroxybenzoic acid,

2
A. Christou et al. Journal of Food Composition and Analysis 114 (2022) 104744

kaempferol, ferulic acid, quercetin, myricetin, caffeic acid, and gallic 2.2.3. Solid-phase purification
acid were obtained from Sigma-Aldrich (St. Louis, MO, USA). Stock In an attempt to obtain a cleaner extract, without significant inter­
standard solutions, 1 mg/mL, were prepared by dissolving in MeOH (LC- ference from other components in the sample, such as sugars and pro­
MS grade) the appropriate amount of substance and stored at − 20 ◦ C for teins, a further purification step, which involves solid-phase extraction
not more than 3 months. Intermediate working solutions were prepared (SPE), was performed. Based on preliminary results of SPE sorbent
daily from these stock standard solutions by appropriate dilution with testing, a DISCOVERY® DPA-6S sorbent material was employed for
LC-MS grade MeOH. Materials for extracts preparation, including MeOH further purification of carob extracts (Christou et al., 2021a). Prior to
and acetone of HPLC-grade and L-ascorbic acid were supplied from extraction, the sorbent material was preconditioned with MeOH (5 mL)
Sigma-Aldrich (St. Louis, MO, USA). Extracts acidification was per­ and equilibrated with an acidified methanolic solution (5 mL), 20:80 v/v
formed using hydrochloric acid (HCl) of analytical grade obtained from MeOH:H2O acidified to pH 2.0 with HCl. The sample mixture, which was
Scharlau Chemie (Barcelona, Spain). prepared by dissolving the obtained freeze-dried extract in acidified
methanolic solution (20 mL), was passed through the cartridge, which
2.2. Extracts preparation was then washed with 5 mL of acidified water (pH 2.0, HCl), in order to
remove the coextracted substances. The bound phenolics were then
2.2.1. Sample material eluted with an aqueous acetone mixture (2 × 5 mL, 80:20 v/v acetone:
A total of 33 carob pod samples were collected during August and water), and the obtained purified extracts were evaporated to dryness.
September 2020 (at peak maturity) from 7 different cultivation regions, The residues were redissolved in a known volume of MeOH (2 mL),
represented by Cyprus (Larnaca, 1–3; Limassol, 4–6; Famagusta, 7–9; filtered through a 0.45-μm pore size membrane filter, properly diluted
Paphos, 10–12), Greece (Heraklion Crete, 13–15; Rethymno Crete, (1:100 or when necessary 1:1000 in 90:10 v/v H2O:MeOH), and injected
16–18), Italy (Sicily, 19–21), Spain (Mallorca, 22–24), Morocco (Mar­ into the HPLC-ESI-HRMS system.
rakesh, 25–27), Palestine (Gaza, 28–30) and USA (Winston-Salem North
Carolina, 31–33). The effect of the ripening process on polyphenolic 2.3. HPLC-ESI-HRMS analysis
composition was assessed in pods harvested from Paphos (Cyprus).
Thus, both mature and immature carobs were collected in 2020 from the An Agilent 1290 Infinity series HPLC system (Agilent Technologies,
same tree cultivated in the region of Archimandrita (Paphos, Cyprus), at Santa Clara, USA) equipped with a vacuum degasser, a binary pump and
different dates according to their ripening stages (June: unripe pods and an autosampler was utilized for the chromatographic determination.
September: ripe pods). The collected samples were washed and crushed The reversed-phase separation was achieved on an Agilent ZORBAX
manually; seeds and pulps were separated, lyophilized (LyoDry Compact Eclipse XDB-C18, Rapid Resolution HT, 4.6 × 100 mm, 1.8 µm column,
Benchtop Freeze Dryer, Mechateck Systems), ground to a fine powder thermostated at 25 ◦ C, under gradient elution condition. The HPLC-ESI-
(Thermomix® TM5, Vorwerk), and passed through a 250-μm sieve HRMS analysis was performed according to a previously described
(Endecotts) to obtain uniformly sized particles. The samples were method with slight modifications (Abrankó et al., 2012). Briefly, 0.1 %
vacuum-packed (4,100,050 Sealcom-V, J. P. SELECTA) and stored at − (v/v) formic acid in H2O (mobile phase A) and MeOH (mobile phase B)
20 ◦ C until extraction and analysis. were used as solvents at a flow rate of 0.5 mL/min. The gradient pro­
Furthermore, the impact of the processing method was investigated gram started at 5 % B (at 0 min), and after 5 min of an isocratic run,
through the analysis of various carob-derived products. Roasted carob solvent B was increased linearly and reached 55 % at 35 min and then
powder, raw carob flour (unroasted carob powder, treated under 40 ◦ C), 100 % at 40 min. Finally, 100 % B was kept constant for 5 min (until 45
carob sirup (processed over 40 ◦ C), raw carob sirup (processed under min). A 10-min re-equilibration time was used after each analysis.
40 ◦ C), as well as ripe carobs and roasted pods were purchased from a Before use, all solvents were filtered through 0.45-μm pore size PVDF
local company (Creta Carob, Rethymno, Crete, Greece). For the whole membrane filters, and before HPLC analysis, all samples were filtered
pods purchased (roasted and unroasted), only the pulp was investigated, through 0.45-μm single-use PTFE syringe filters. An injection volume of
which was obtained after crushing the fruit and removing the seeds. The 10 μL was used in all experiments.
pulp obtained from the roasted and unroasted pods was then treated as The HPLC system was coupled to a time-of-flight mass spectrometer
previously described (converted into a dry powder), while the remaining Agilent 6220 TOF (Agilent Technologies, Santa Clara, CA) equipped
derived products were subjected directly to the following extraction with an electrospray interface operating in negative ion polarity mode.
procedure. The ionization source operating conditions were as follows: capillary
voltage, 2500 V; nebulizer gas pressure, 40 psi; drying gas flow rate, 9 L/
2.2.2. Ultrasound-assisted extraction min; drying gas temperature, 325 ◦ C; and fragmentor voltage, set at 140
Ultrasound-assisted extraction (UAE) of phenolics was previously V. Nitrogen was used as both nebulizing and drying gas. HPLC-MS ac­
optimized using response surface methodology (RSM) (Christou et al., curate mass spectra were recorded across the range of 50–1000 m/z and
2021a). In brief, UAE of phenolics was performed by use of a 500-W data acquisition and processing were performed on the Agilent Mass
power and 20-kHz frequency ultrasonic probe system (CY-500, Optic Hunter software (version B.06.00).
Ivymen System). For the extraction, 2 g of carob sample (pulp/seed
powder or carob-based product) was mixed with 50 mL of aqueous 2.4. Statistical analysis
acetone solution (57 %, v/v, acetone) containing 0.5 % (w/v) ascorbic
acid, as an antioxidant to prevent polyphenols’ oxidation, and the ob­ All experimental assays were performed in triplicate. The results
tained suspension was exposed to acoustic cavitation for 14 min using 50 obtained were expressed as mean values ± standard deviation (SD). The
% of ultrasonic amplitude. The tip of the probe was submerged means were compared and statistical differences were obtained through
approximately 2 cm into the extraction solution and the sonication was one-way analysis of variance (ANOVA) followed by Duncan’s multiple
conducted in the pulsed mode (pulse duration:pulse interval 5 s:5 s). range test at a 95 % of confidence level. The differences between indi­
After the ultrasound treatment, the obtained mixture was centrifuged at vidual means were considered significant at p < 0.05. Then, the mean
4400 rpm for 20 min, and filtered through Whatman no.1 filter paper. values of the obtained data set were subjected to pattern recognition
The treated sample was then concentrated under vacuum using a rotary analysis. The data set to be treated consisted of a 33 × 10 matrix, in
evaporator (RE300, stuart®), lyophilized, and vacuum stored at − 20 ◦ C which rows represented the carob pulp samples analyzed (33 objects)
until further analysis. All extracts were prepared in triplicate. and the columns the concentrations of the individual phenolic com­
pounds detected by HPLC-ESI-HRMS analysis (10 variables). Prior to
multivariate analysis, the entire data matrix was mean-centered and

3
A. Christou et al. Journal of Food Composition and Analysis 114 (2022) 104744

scaled to unit variance to standardize the statistical importance of all respectively, are considered to be proper for the determination of
responses. PCA, as an unsupervised pattern recognition technique, was polyphenols in real samples. The intra- and interday precision were
initially applied to reduce data dimensionality and to identify any evaluated using the matrix-matched standard solutions at the concen­
existing clustering of carob samples based on geographical origin. tration level of 10 μg/g. For the intraday variability test, 5 replicate
Subsequently, PLS-DA, as a supervised method, was performed to analyses were performed within one single day, whereas for the interday
construct classification models and to extract information on discrimi­ assay, measurements were conducted once a day for five consecutive
nant features. The all-for-mentioned statistical analyses were performed days. All of the precision measurements are expressed in Table 1 as
using RStudio statistical software (version 1.3.1073). relative standard deviations (RSDs). The precision of the method,
expressed as % RSD of the peak areas, met acceptance criteria, since RSD
3. Results and discussion values were lower than 12 % and 20 % for intra- and interday precision,
respectively. The impact of the matrix on the ionization, in terms of ion
3.1. Method development and validation suppression or ion enhancement, was evaluated by comparing the slopes
obtained in the calibration with matrix-matched standards with those
The target analytes were selected based on a previous work that obtained with solvent-based standards, according to the following
studied the phenolic composition of carob and its derived products, as equation: Matrix Effect % = ((slope of calibration curve in matrix/slope
well as the availability of reference standards (Christou et al., 2021a). of calibration curve in pure solvent)− 1) × 100. Negative matrix effect
The selected analytes were separated within 39 min by use of the opti­ values indicate signal suppression due to the matrix, while positive re­
mized chromatographic conditions. Fig. S1 (Supplementary Material) sults demonstrate enhancing effects. No significant matrix effect was
demonstrates the typical extracted ion chromatograms (EICs) of the observed for most of the studied compounds (matrix effect values lower
standard analytes and the EICs of polyphenolic compounds present in a than 20 %). Due to the negligible matrix effect observed, solvent-based
real sample (carob pulp from Limassol). The identification of the tar­ calibration curves were utilized for the quantification process. In gen­
geted species was performed by comparing both retention time and eral, the proposed analytical method proved to be a very efficient tool
accurate mass spectra obtained from carob samples and standards of the for the determination of the selected compounds, revealing excellent
target analytes. The main analytical parameters for each compound, performance characteristics.
including retention time, chemical formula, characteristic fragmenta­
tion, theoretical and experimental accurate mass, are reported in
Table S1 (Supplementary Material). In order to fulfill the identification 3.2. Analysis of real samples – Factors affecting the phenolic composition
criteria for HRMS analysis, at least two ions (including the molecular ion
and a fragment ion) with a mass error of less than 5 ppm were selected to 3.2.1. Effect of ripening process and distribution of polyphenols across the
confirm the presence of each analyte. In some cases, a third ion was fruit
considered in the identification process, due to its high abundance. In Both ripe and unripe carob pods were harvested from the same tree,
any case, the quantification was carried out by monitoring the most thus ensuring similar and controlled environmental and cultivation
intense ion, [M-H]-. Based on the data obtained, it can be concluded that conditions. The analytical data obtained from the chromatographic
the method offers a high degree of confirmation because of its high mass determination of carob pulp and seeds phenolics in both ripening stages
accuracy, enabling accurate mass measurements of target ions within are demonstrated in Table 2. In agreement with previous studies, unripe
5-ppm error in most cases. pods (both pulp and seeds) were found to be more enriched in poly­
The proposed analytical method was validated in terms of linearity, phenols compared to the ripe ones. According to Ydjedd et al. (2017b),
limit of detection (LOD) and quantification (LOQ), intra- and interday the high content of phenolics recorded at the unripe stage is a result of
precision, and matrix effect (Table 1). Linearity was assessed by con­ the effort of the plant to protect itself against biotic and/or abiotic
structing matrix-matched standard calibration curves. In particular, environmental hazards during its growth. The measurable levels of
carob extracts (sample no. 16) were spiked with known amounts of almost all polyphenols detected in both pulp and seeds were observed to
working solutions in order to obtain the desired concentration range. In fall significantly upon increasing the degree of maturation. In particular,
all cases, regression coefficients (R2) were higher than 0.994, demon­ the contents of gallic acid, 2,5-dihydroxybenzoid acid, catechin, ferulic
strating acceptable linear relation between the range of concentrations acid, myricetin, and naringenin decreased significantly during carob’s
assayed and the detector response. LODs and LOQs were obtained ac­ pulp ripening, while the contents of quercetin and kaempferol remained
cording to the signal-to-noise ratio (S/N) of 3 and 10 criteria. Both of almost unchanged. As far as the phenolic composition of carob seeds
them were determined experimentally by injecting a series of dilute during the fruit ripening process is concerned, it was observed that gallic
matrix-matched standard solutions. As shown in Table 1, the LOD and acid, catechin, and ferulic acid concentrations demonstrated a signifi­
LOQ values, which ranged from 0.03 to 3 μg/g and 0.1–10 μg/g, cant decline, while the concentrations of caffeic acid, myricetin, and
naringenin remained constant and the concentrations of quercetin and

Table 1
Performance characteristics of the proposed analytical method.
Compound Linearity range (µg/g) Linearity (R2) LOD (μg/g) LOQ (μg/g) Precision (RSD %) Matrix effect (%)

Intraday Interday

Gallic acid 1–50 0.999 0.3 1 6 7 -13

Pyrocatechol 1–100 0.999 0.3 1 9 16 5
2,5-Dihydroxybenzoic acid 0.4–50 0.997 0.12 0.4 6 18 1
Catechin 1–100 0.994 0.3 1 6 20 7
Chlorogenic acid 1–50 0.999 0.3 1 4 14 3
Caffeic acid 0.5–50 0.999 0.15 0.5 7 14 11
Ferulic acid 0.5–100 0.999 0.15 0.5 6 15 -5
Myricetin 1–50 0.999 0.3 1 8 11 -15
trans-Cinnamic acid 10–1000 0.998 3 10 4 18 -1
Quercetin 1–50 0.995 0.3 1 12 19 -21
Naringenin 0.1–10 0.998 0.03 0.1 3 15 -2
Kaempferol 0.1–10 0.995 0.03 0.1 5 15 -18

4
A. Christou et al. Journal of Food Composition and Analysis 114 (2022) 104744

Table 2 2017b).
Phenolic composition of ripe and unripe carob pulp and seeds. Carob seeds are predominantly composed of catechin, gallic acid,
Compound Concentration (μg/g) and quercetin in both ripening stages. Ripe seeds were more enriched in
flavonols (quercetin and kaempferol), while flavanols (catechin) were
Unripe stage Ripe stage
the dominant compounds identified in seeds obtained from unripe pods.
Pulp Seeds Pulp Seeds Among phenolic acids, gallic acid was the major component detected in
Gallic acid 1510 ± 667 ± 46b 374 ± 9c 334 ± 10c both ripening stages, with unripe pods demonstrating considerably
70a higher amounts. In agreement with the present findings, Santonocito
Pyrocatechol nd nd nd nd
et al. (2020) highlighted the predominance of flavonols (mainly
2,5-Dihydroxybenzoic 1.49 ± ndb nqb ndb
acid 0.24a quercetin-deoxyhexoside) and flavanols (mainly catechin) in unripe and
Catechin 720 ± 10c 4270 ± nqd 1500 ± ripe seeds extracts, respectively. Gallic acid and its derivatives were also
155a 65b more abundant in seeds extracts from unripe pods compared to the ripe
Chlorogenic acid nd nd nd nd ones (Santonocito et al., 2020). Similar results were also obtained in
Caffeic acid nqb 1.31 ± nqb 1.88 ±
0.09a 0.47a
previous studies during the analysis of carob seed peel and germ ex­
Ferulic acid 2.40 ± 5.31 ± nqc 2.24 ± tracts. In particular, high levels of catechin, quercetin, epicatechin,
0.12b 0.80a 0.24b apigenin, gallic acid, and their derivatives were determined (Albertos
Myricetin 46.2 ± 13.0 ± 9.73 ± 6.68 ± et al., 2015; Ben Ayache et al., 2021; Rico et al., 2019).
7.2a 1.0b 2.18b 1.91b
trans-Cinnamic acid nq nd nd nd
Quercetin 25.5 ± 436 ± 10b 6.95 ± 1640 ± 3.2.2. Effect of processing
1.8c 1.59c 22a The main polyphenolic patterns of carob-derived products were also
Naringenin 6.77 ± 1.73 ± 5.23 ± 0.69 ± identified and quantified in order to investigate the effect of processing
0.99a 0.57c 0.24b 0.19c on carobs’ polyphenolic composition. Therefore, the phenolic profiles of
Kaempferol 1.49 ± 27.9 ± ndc 75.6 ±
0.39c 5.5b 11.8a
roasted and unroasted carob powders, roasted carob pulp, and carob
Total 2310 ± 5420 ± 396 ± 9d 3560 ± sirup treated under different temperatures were determined and
69c 173a 50b compared. At the same time, for comparison purposes, the phenolic
Different lower-case letters (a, b, c and d) within the same row indicate signif­ composition of raw carob pulp, based on which these products were
icant differences (p < 0.05) according to Duncan’s multiple range test; nd, non- derived, was also investigated. Table 3 demonstrates the qualitative and
detected; nq, non-quantified. quantitative results obtained from the analysis of both raw material and
carob-derived products.
kaempferol presented an unexpected increase (almost 3–4 times higher As observed, the processing of carob pods has a great effect on the
concentrations at the ripe stage). All these variations may be attributed polyphenolic pattern obtained and the extractable quantities of poly­
to the different activities of enzymes involved in the biosynthetic phenols. Gallic acid was the major component of carob sirups, while
pathways of these phytochemicals during fruit maturation (Liu et al., other polyphenols were detected in minor amounts. This suggests that
2015). gallic acid was co-extracted with carbohydrates during carob sirup
The distribution of the secondary metabolites across the fruit was preparation. Treatment with water may have facilitated the release of
also established by comparing the phenolic patterns of carob pulp and gallic acid from the carob matrix, which, in turn, resulted in significantly
seeds of both ripening stages. The results obtained (Table 2) demon­ higher levels of gallic acid in carob sirup samples compared to the raw
strated significant differences in the phenolic profiles of carob pulp and carob meal. This is in agreement with the results obtained by Papa­
seeds. In particular, the seeds exhibited a significantly higher amount of giannopoulos et al. (2004), who also reported a high amount of gallic
polyphenols in comparison to the seedless part of the pod. This is not in acid along with hydrolyzable tannins in carob sirup samples. Similar
agreement with the literature, which states that the majority of poly­ findings were also observed by Dhaouadi et al. (2014). In their study,
phenols in carobs are concentrated in the seedless part of the pod (Ben gallic acid was found to be the main compound of carob sirups, followed
Ayache et al., 2021; Benković et al., 2017; Papagiannopoulos et al., by quercetin glycoside and syringic acid.
2004). It is worth here to mention though, that the polyphenolic It was also observed that thermal treatment had a negative impact on
composition of carob pulp has not, so far, been compared with the sirups’ polyphenolic composition. Sirups that were processed above
phenolic pattern of whole carob seeds. It was only compared with spe­ 40 ◦ C had a markedly lower gallic acid content. An increase in the
cific parts of them, including seeds peel, endosperm, and germ. Ac­ processing temperatures resulted in several changes in the chemical
cording to the literature, seeds polyphenols are restricted to the outer, composition of carob sirup, including thermal degradation of poly­
dark brown seed coat, which generally has a small overall amount phenols due to decarboxylation reactions. A comparison with previous
(Papagiannopoulos et al., 2004). Therefore, the view that the carob pulp studies could not be made, since this is the first time that the phenolic
is characterized by a higher content of phenolic components should be content of carob sirup processed under mild temperatures and condi­
revised. tions was investigated.
Gallic acid was found to be the most abundant phenolic compound in Thermal degradation was also observed during the roasting of carob
both ripe and unripe carob pulp extracts. Unripe pulp exhibited powder. As observed in Table 3, the levels of gallic acid, catechin, and
comparatively high proportions of more flavonoids, including catechin, quercetin decreased with the roasting treatment, while, at the same
myricetin, and quercetin. The phenolic profile of the ripe pulp extract time, cinnamic acid, myricetin, and even 2,5-dihydroxybenzoic acid
was almost dominated by gallic acid. The decreased flavonoid content contents demonstrated a significant increase. According to the litera­
observed in the present study can be attributed to the polymerization ture, roasting treatment can induce degradation of hydrolyzed tannins,
and/or glycosylation of flavonoid substances during the ripening pro­ increasing the amount of gallic acid and lowering the content of tannins
cess and the subsequent reduction of the aglycone moieties (Christou in the final product (Papagiannopoulos et al., 2004). However, such a
et al., 2021b). In comparison with other investigations, remarkable trend was not observed in the present study. Previous reports demon­
differences were observed in terms of identified phenolics and respective strated higher phenolic levels for the roasted powder compared to the
quantification. Nonetheless, most of the studies performed highlighted not thermally treated one. This was attributed to the difference in the
the predominance of gallic acid in both ripe and unripe carob pulp ex­ hydrolysis of gallotannins and ellagitannins and the subsequent release
tracts (Benchikh et al., 2016; Goulas and Georgiou, 2020; Ydjedd et al., of gallic acid units (Eldeeb and Mosilhey, 2022; Şahin et al., 2009; Van
Rijs and Fogliano, 2020). It is worth here to mention that most of these

5
A. Christou et al. Journal of Food Composition and Analysis 114 (2022) 104744

Table 3
Phenolic composition of carob-derived products and raw material.
Compound Concentration (μg/g)

Sirup Powder Carob pulp

< 40 oC > 40 oC Unroasted Roasted Unroasted Roasted

Gallic acid 1920 ± 192a 1280 ± 63b 1100 ± 95c 668 ± 22e 323 ± 8 f 855 ± 81d
Pyrocatechol nd nd nd nd nd nd
2,5-Dihydroxybenzoic acid 4.97 ± 0.34a 4.17 ± 0.56b 1.08 ± 0.07d 1.86 ± 0.63c 0.89 ± 0.32d 1.18 ± 0.35 cd
Catechin nd nd 19.9 ± 0.7a 2.01 ± 0.88c 3.51 ± 0.92c 12.2 ± 3.9b
Chlorogenic acid nd nd nd nd nd nd
Caffeic acid 0.66 ± 0.10b 1.08 ± 0.10a nq nq nq nq
Ferulic acid 8.45 ± 1.69b 18.4 ± 4.0a 5.73 ± 1.37bc 4.37 ± 0.51bc 7.74 ± 1.30bc 6.39 ± 1.20bc
Myricetin 1.78 ± 0.07e 6.95 ± 0.39d 24.1 ± 4.4b 39.1 ± 2.9a 24.1 ± 1.1b 12.1 ± 2.2c
trans-Cinnamic acid nq nq nq 35.8 ± 1.1b nq 45.3 ± 7.8a
Quercetin 6.65 ± 0.16c 10.0 ± 0.5c 34.1 ± 3.4a 20.5 ± 2.9b 19.8 ± 2.0b 18.4 ± 6.9b
Naringenin 0.61 ± 0.21c 1.02 ± 0.26c 3.67 ± 0.57b 3.39 ± 0.54b 5.41 ± 1.44a 5.02 ± 0.15a
Kaempferol 0.65 ± 0.26c nd 1.99 ± 0.29a 1.64 ± 0.31ab 1.36 ± 0.29b 1.13 ± 0.48bc
Total 1940 ± 190a 1320 ± 67b 1190 ± 102b 777 ± 20c 386 ± 9d 957 ± 97c

Different lower-case letters (a-f) within the same row indicate significant differences (p < 0.05) according to Duncan’s multiple range test; nd, non-detected; nq, non-
quantified.

previous studies focused on the determination of the total phenolic content during the roasting process. During carob pulp roasting,
content (TPC) of carob powder samples. The Folin-Ciocalteu colori­ degradation of myricetin and an increase in the solubility of catechin
metric reagent, which is used to determine the TPC, detects all phenolic and cinnamic acid were also observed.
groups present in the sample (naturally occurring polyphenols and
newly formed compounds). In particular, during the roasting process, 3.2.3. Effect of geographical origin
Maillard reaction products with phenolic type structure can be deter­ To assess the effect of geographical origin on carobs polyphenolic
mined by use of this method of analysis and this results in higher levels composition, the phenolic profiles of carobs collected from different
of TPC in the final roasted product. Furthermore, the breakdown of regions were also investigated. Table S2 (Supplementary Material)
cellular structures during roasting could influence the solubility of demonstrates the concentration ranges, mean values, and standard de­
previously insoluble or bound phenolics, leading to the detection of viations of the polyphenols detected, sorted by sampling area. As
higher amounts of polyphenols. This applies to cinnamic acid, myricetin, observed, the polyphenols exhibited notable variations among the
and 2,5-dihydroxybenzoic acid results. samples analyzed, probably due to the influence of geographical origin.
The phenolic profile of unroasted carob powder was similar to the These variations among the carob pulp extracts are graphically pre­
one obtained with the initial raw material. The main difference involved sented in Fig. 1.
the considerably higher amounts of phenolics, possibly due to the The total phenolic contents ranged from 144 μg/g in pulp samples
technological processes used for obtaining carob flour. During the pro­ from Morocco to 2520 μg/g in extracts from USA. In all cases, the
cessing steps the matrix is thought to be ruptured and bound phenolics highest contents corresponded to gallic acid, which is in line with the
to be released. As a result, a greater amount of phenolics is detected in previously mentioned data (Section 3.2.1). No detectable amounts of
the final product. On the other hand, the roasting of carob pulp resulted chlorogenic acid and pyrocatechol were observed in the examined
in an increase in gallic acid content. In this case, the hydrolysis of hy­ samples; so, they were not included in the multivariate statistical
drolyzable tannins and the subsequent release of gallic acid moieties is analysis.
the only possible explanation for the observed increase of gallic acid Given the large number of polyphenols, as variables, PCA was first

Fig. 1. Graphical representation of major polyphenols detected in carob pulp samples from different geographical origins.

6
A. Christou et al. Journal of Food Composition and Analysis 114 (2022) 104744

introduced in order to investigate the structure of the data set and to describe the Italian carobs phenolic composition. The USA samples are
identify variables responsible for the similarities and differences of the characterized by high contents of all of the aforementioned polyphenols,
examined samples. In particular, PCA was employed on the autoscaled along with high levels of kaempferol, quercetin, and catechin. On the
data matrix, which consisted of 33 samples and 10 variables (all poly­ other hand, caffeic acid is alone at the bottom of the plot, being char­
phenols detected), to locate any existing clustering of carob pulp sam­ acteristic of the Greek samples. These samples are dispersed along the
ples based on their geographical origin. The first three principal PC2, which indicates a great variability of the contents of caffeic acid as
components (PCs) accounted for 65.47 % of the total variability of the well as naringenin. The third group is negatively correlated to caffeic
system. When scores of the samples were presented on the two- (Fig. 2) acid and consists of naringenin. As observed, samples from Greece,
and three-dimensional spaces (data not shown) defined by PC1 Cyprus, and Palestine are best described by their naringenin contents,
(explaining 32.03 % of the variance), PC2 (explaining 19.34 % of the since these samples are found on the top side of the plot.
variance), and PC3 (explaining 14.10 % of the variance), respectively, a However, the PCA failed to achieve any statistically significant
distinct separation of species collected from USA, Spain, and Italy was segregation of specimens collected from Cyprus, Greece, Morocco, and
observed. However, samples from Cyprus, Greece, Morocco, and Palestine, probably due to the similar climatic and environmental con­
Palestine could not be separated due to the obvious overlaps of their ditions prevailing in these areas. On the contrary, the samples collected
clusters in both bidimensional and tridimensional plots. from the USA, outside the Mediterranean region, demonstrated a clear
In addition, in Fig. 2, it is clearly demonstrated that the majority of separation. These samples were located on the right part of the plot, far
the polyphenols are concentrated and, thereby, correlated, in the right away from the Mediterranean specimens. The PCA reflects the signifi­
part of the plot. These compounds are characteristic of the separated cant difference between the Mediterranean and American carob sam­
groups of samples (USA, Italy, and Spain) since all of them are pooled in ples, with the latter demonstrating significantly higher levels of
the right part of the plot. In particular, the phenolic content of the polyphenols. Therefore, even though the complete segregation of sam­
Spanish samples is determined by myricetin, gallic acid, and trans- ples collected from the Mediterranean region was not achieved, with the
cinnamic acid, while 2,5-dihydroxybenzoic acid and ferulic acid exception of the Italian and Spanish samples, PCA enabled the clear

Fig. 2. Principal component score projection of carob pulp samples collected from seven different geographical regions.

7
A. Christou et al. Journal of Food Composition and Analysis 114 (2022) 104744

differentiation of samples collected from the Mediterranean and non- 4. Conclusions

Mediterranean regions, probably due to the different climatic condi­
tions of the areas. It is worth here to mention again that the phenolic In recent years, carobs phenolics have gained wide attention due to
composition of carobs is affected by several other parameters, including the growing evidence that their consumption is associated with a
carob variety, gender, conditions of cultivation, and degree of matura­ plethora of health-promoting properties. Several factors, from intrinsic
tion, which were impossible to control. As a result, the similarities or to agronomic, affect carobs phenolic composition. Among them, the
differences observed between the examined samples may be due to effects of ripening stage, processing method, and geographical origin
factors other than their different geographical origins. were well investigated in the present study, through the analysis of both
PLS-DA was then applied to the autoscaled data matrix in order to carob pods (pulp and seeds) and derived products. Besides that, the
develop discrimination models for the efficient classification of carob distribution of the secondary metabolites across the fruit was also
samples according to their geographical origins. Four latent variables established by comparing the phenolic patterns of carob pulp and seeds
(LVs) were selected to construct the PLS-DA model, which account for of both ripening stages. The qualitative and quantitative analysis of the
99.99% of the parameters variance (X) and 44.17% of the class variance main polyphenolic substances of the examined samples was performed
(Y), and demonstrates the lowest root mean square error of leave-one- by the use of an HPLC-ESI-HRMS analytical method, which was previ­
out cross-validation (LOOCV). The USA samples were again well sepa­ ously validated in terms of linearity, LODs, LOQs, intra- and interday
rated. However, complete visual discrimination between the remaining precision, and matrix effect.
six regions, of the Mediterranean basin, was not achieved. Fig. S2 (A) Unripe pods, both pulp and seeds, were found to be more enriched in
demonstrates the scores plot of LV1 and LV2, where the USA samples polyphenols compared to the ripe ones (pulp and seeds). In particular,
located on the bottom left of the graph, displaying negative LV1 and LV2 the levels of almost all polyphenols detected in both pulp and seeds were
values, were well discriminated. The scatter plot of loadings of LV1 observed to fall significantly upon an increase in the degree of matu­
versus LV2 describing the behavior of the polyphenolic variables is ration. At the same time, seeds were found to exhibit a significantly
demonstrated in Fig. S2 (B). higher amount of polyphenols compared to the seedless part of the pod.
Since the single PLS-DA model proved inadequate to discriminate the Gallic acid was found to be the most abundant phenolic component in
examined samples according to their geographical region, a classifica­ carob pulp extracts, while carob seeds were predominantly composed of
tion decision tree, consisting of consecutive smaller two-input class PLS- catechin, quercetin, and gallic acid, as well.
DA models, was proposed. The flow chart of the classification decision Furthermore, the processing of carob pods had a great impact on the
tree is demonstrated in Fig. S3. The corresponding scores plots of the phenolic pattern of carob-derived products. As observed, gallic acid was
consecutive PLS-DA models (A. USA vs others, B. Cyprus vs others, C. the major component of carob sirups, which was probably co-extracted
Italy vs others, D. Morocco vs others, E. Spain vs others, E. Greece vs with carbohydrates during carob sirup preparation. Thermal treatment
Palestine), which are presented in Fig. 3, provided successively a satis­ had a negative impact on the phenolic content of both carob sirup and
factory classification for the studied classes. Even though, in some cases, powder, resulting in thermal degradation of polyphenols. Unroasted
a slight overlapping was observed, due to the great variability of the carob powder demonstrated a similar phenolic profile with that of raw
respective groups, a satisfactory classification of the examined samples carob material, while a considerably higher amount of phenolics was
was achieved. Table 4 demonstrates the LOOCV results of the proposed observed. Matrix disrupture and subsequent release of bound phenolics
models, ensuring their satisfactory performance. Through LOOCV, are suggested to be the factors that may have contributed to the high
models were evaluated by accuracy (the ratio of the sum of true positives contents of polyphenols detected in carob powder. Unlike carob powder,
and true negatives), sensitivity (capability to detect true positives) and the roasting of carob pulp resulted in an increase in phenolic content,
specificity (capability to detect true negatives). All sample classes probably due to the hydrolysis of tannins and the subsequent release of
demonstrated satisfactory specificity (≥ 0.944), sensitivity (≥ 0.889) gallic acid moieties.
and accuracy (≥ 0.917) values, except from Italy (Italy vs Others model) Noticeable differences were also observed during the analysis of
that presented an accuracy of 0.889 and a poor sensitivity (0.333). Based carob pulp samples from different geographical regions. PCA enabled
on the above, polyphenols could be used as markers to trace the origin of the classification of the USA and Mediterranean samples. The subse­
carob pulp samples effectively. Probably a larger number of examined quent classification decision tree PLS-DA enabled the satisfactory
samples would give more representative results. discrimination of all the examined samples according to their
geographical origin. In conclusion, polyphenols, when combined with

Fig. 3. PLS-DA scores plots of LV1 and LV2 using the smaller two-input class models consisting of: (A) USA vs others, (B) Cyprus vs others, (C) Italy vs others, (D)
Morocco vs others, (E) Spain vs others and (F) Greece vs Palestine.

8
A. Christou et al. Journal of Food Composition and Analysis 114 (2022) 104744

Table 4
Quality parameters of the calibration of the proposed PLS-DA models.
Calibration model Specificity Sensitivity Accuracy

USA vs Others USA Others USA Others USA Others

1 1 1 1 1 1
Cyprus vs Others Cyprus Others Cyprus Others Cyprus Others
0.944 0.917 0.917 0.944 0.933 0.933
Italy vs Others Italy Others Italy Others Italy Others
1 0.333 0.333 1 0.889 0.889
Morocco vs Others Morocco Others Morocco Others Morocco Others
1 1 1 1 1 1
Spain vs Others Spain Others Spain Others Spain Others
1 0.889 0.889 1 0.917 0.917
Greece vs Palestine Greece Palestine Greece Palestine Greece Palestine
1 1 1 1 1 1

chemometrics, can be exploited as reliable geographical origin in­ Barbosa, S., Pardo-Mates, N., Hidalgo-Serrano, M., Saurina, J., Puignou, L., Núnez, O.,
2018. Detection and quantitation of frauds in the authentication of Cranberry-based
dicators for carob pulp classification.
extracts by UHPLC-HRMS (Orbitrap) polyphenolic profiling and multivariate
calibration methods. J. Agric. Food Chem. 66, 9353–9365. https://doi.org/10.1021/
CRediT authorship contribution statement acs.jafc.8b02855.
Barbosa, S., Campmajó, G., Saurina, J., Puignou, L., Núñez, O., 2020a. Determination of
phenolic compounds in paprika by ultrahigh performance liquid chromatography-
Atalanti Christou: Methodology, Investigation, Writing – original tandem mass spectrometry: application to product designation of origin
draft. Ana B. Martinez-Piernas: Investigation, Resources, Writing – authentication by chemometrics. J. Agric. Food Chem. 68, 591–602. https://doi.org/
10.1021/acs.jafc.9b06054.
review & editing. Ioannis J. Stavrou: Writing – review & editing, Su­
Barbosa, S., Saurina, J., Oscar, N., 2020b. Capsaicinoid profiling for the chemometric
pervision. Juan F. Garcia-Reyes: Resources, Writing – review & edit­ characterization and classification of paprika with protected designation of origin
ing, Supervision, Project administration. Constantina P. Kapnissi- (PDO) attributes. Molecules 25, 1623. https://doi.org/10.3390/
Christodoulou: Writing – review & editing, Supervision, Project molecules25071623.
Ben Ayache, S., Reis, F.S., Inês Dias, M., Pereira, C., Glamočlija, J., Soković, M., Behija
administration, Funding acquisition. Saafi, E., Ferreira, C.F.R., Barros, I., Achour, L, L., 2021. Chemical characterization
of carob seeds (Ceratonia siliqua L.) and use of different extraction techniques to
promote its bioactivity. Food Chem. 351, 129263 https://doi.org/10.1016/j.
Funding
foodchem.2021.129263.
Benchikh, Y., Paris, C., Louaileche, H., Charbonnel, C.éline, Ghoul, M., Chebil, L., 2016.
This work was co-funded by the European Regional Development Comparative characterization of green and ripe carob (Ceratonia siliqua L.):
Fund and the Republic of Cyprus through the Research & Innovation physicochemical attributes and phenolic profile. SDRP J. Food Sci. Technol. 1,
85–91.
Foundation (Project: BlackGold INTEGRATED /0916/0019). ABMP is Benković, M., Belščak-Cvitanović, A., Bauman, I., Komes, D., Srečec, S., 2017. Flow
grateful for a postdoctoral contract to the Junta de Andalucia properties and chemical composition of carob (Ceratonia siliqua L.) flours as related
(DOC_01319). to particle size and seed presence. Food Res. Int. 100, 211–218. https://doi.org/
10.1016/j.foodres.2017.08.048.
Cavdarova, M., Makris, D.P., 2014. Extraction kinetics of phenolics from carob
(Ceratonia siliqua L.) kibbles using environmentally benign solvents. Waste Biomass
Declaration of Competing Interest Valoriz. 5, 773–779. https://doi.org/10.1007/s12649-014-9298-3.
Chait, Y.A., Gunenc, A., Bendali, F., Hosseinian, F., 2020. Simulated gastrointestinal
The authors declare that they have no known competing financial digestion and in vitro colonic fermentation of carob polyphenols: bioaccessibility
and bioactivity. LWT - Food Sci. Technol. 117, 108623 https://doi.org/10.1016/j.
interests or personal relationships that could have appeared to influence
lwt.2019.108623.
the work reported in this paper. Christou, A., Stavrou, I.J., Kapnissi-Christodoulou, C.P., 2021a. Continuous and pulsed
ultrasound-assisted extraction of carob’s antioxidants: processing parameters
Data availability optimization and identification of polyphenolic composition. Ultrason. Sonochem.
76, 105630 https://doi.org/10.1016/j.ultsonch.2021.105630.
Christou, A., Stavrou, I.J., Kapnissi-Christodoulou, C.P., 2021b. Combined use of
Data will be made available on request. β-cyclodextrin and ionic liquids as electrolyte additives in EKC for separation and
determination of carob’s phenolics – a study of the synergistic effect. Electrophoresis
42, 1945–1955. https://doi.org/10.1002/elps.202100085.
Appendix A. Supporting information Christou, C., Agapiou, A., Kokkinofta, R., 2018. Use of FTIR spectroscopy and
chemometrics for the classification of carobs origin. J. Adv. Res. 10, 1–8. https://doi.
org/10.1016/j.jare.2017.12.001.
Supplementary data associated with this article can be found in the
Daglia, M., 2012. Polyphenols as antimicrobial agents. Curr. Opin. Biotechnol. 23,
online version at doi:10.1016/j.jfca.2022.104744. 174–181. https://doi.org/10.1016/j.copbio.2011.08.007.
Dhaouadi, K., Belkhir, M., Akinocho, I., Raboudi, F., Pamies, D., Barrajón, E., Estevan, C.,
Fattouch, S., 2014. Sucrose supplementation during traditional carob syrup
References
processing affected its chemical characteristics and biological activities. LWT - Food
Sci. Technol. 57, 1–8. https://doi.org/10.1016/j.lwt.2014.01.025.
Abrankó, L., García-Reyes, J.F., Molina-Díaz, A., 2012. Systematic bottom-up approach Díaz-De-Cerio, E., Aguilera-Saez, L.M., María Gómez-Caravaca, A., Verardo, V.,
for flavonoid derivative screening in plant material using liquid chromatography Fernández-Gutiérrez, A., Fernández, I., Arráez-Román, D., Laganà, A., Capriotti, A.L.,
high-resolution mass spectrometry. Anal. Bioanal. Chem. 403, 995–1006. https:// Cavaliere, C., 2018. Characterization of bioactive compounds of Annona cherimola
doi.org/10.1007/s00216-012-5865-2. L. leaves using a combined approach based on HPLC-ESI-TOF-MS and NMR
Albertos, I., Jaime, I., María, A., Gonz, L., Rico, D., 2015. Carob seed peel as natural Published in the topical collection discovery of bioactive compounds with guest
antioxidant in minced and refrigerated (4C) Atlantic horse mackerel (Trachurus editors. Anal. Bioanal. Chem. 410, 3607–3619. https://doi.org/10.1007/s00216-
trachurus). LWT - Food Sci. Technol. 64, 650–656. https://doi.org/10.1016/j. 018-1051-5.
lwt.2015.06.037. Eldeeb, G.S.S., Mosilhey, S.H., 2022. Roasting temperature impact on bioactive
Almanasrah, M., Roseiro, L.B., Bogel-Lukasik, R., Carvalheiro, F., Brazinha, C., Crespo, J., compounds and PAHs in Carob powder (Ceratonia siliqua L.). J. Food Sci. Technol.
Kallioinen, M., Mänttäri, M., Duarte, L.C., 2015. Selective recovery of phenolic 59, 105–113. https://doi.org/10.1007/s13197-021-04989-7.
compounds and carbohydrates from carob kibbles using water-based extraction. Ind. Farag, M.A., El-Kersh, D.M., Ehrlich, A., Choucry, M.A., El-Seedi, H., Frolov, A.,
Crops Prod. 70, 443–450. https://doi.org/10.1016/j.indcrop.2015.02.051. Wessjohann, L.A., 2019. Variation in Ceratonia siliqua pod metabolome in context of
Antoniou, C., Kyratzis, A., Rouphael, Y., Stylianou, S., Kyriacou, M.C., 2020. Heat- and its different geographical origin, ripening stage and roasting process. Food Chem.
ultrasound-assisted aqueous extraction of soluble carbohydrates and phenolics from 283, 675–687. https://doi.org/10.1016/j.foodchem.2018.12.118.
carob kibbles of variable size and source material. Foods 9, 1364. https://doi.org/
10.3390/foods9101364.

9
A. Christou et al. Journal of Food Composition and Analysis 114 (2022) 104744

Goulas, V., Georgiou, E., 2020. Utilization of carob fruit as sources of phenolic Owen, R.W., Haubner, R., Hull, W.E., Erben, G., Spiegelhalder, B., Bartsch, H., Haber, B.,
compounds with antioxidant potential: extraction optimization and application in 2003. Isolation and structure elucidation of the major individual polyphenols in
food models. Foods 9, 20. https://doi.org/10.3390/foods9010020. carob fibre. Food Chem. Toxicol. 41, 1727–1738. https://doi.org/10.1016/S0278-
Izquierdo-Llopart, A., Saurina, J., 2019. Characterization of sparkling wines according to 6915(03)00200-X.
polyphenolic profiles obtained by HPLC-UV/Vis and principal component analysis. Papagiannopoulos, M., Wollseifen, H.R., Mellenthin, A., Haber, B., Galensa, R., 2004.
Foods 8, 22. https://doi.org/10.3390/foods8010022. Identification and quantification of polyphenols in carob fruits (Ceratonia siliqua L.)
Kokkinofta, R., Yiannopoulos, S., Stylianou, M.A., Agapiou, A., 2020. Use of and derived products by HPLC-UV-ESI/MSn. J. Agric. Food Chem. 52, 3784–3791.
chemometrics for correlating carobs nutritional compositional values with https://doi.org/10.1021/jf030660y.
geographic origin. Metabolites 10, 62. https://doi.org/10.3390/metabo10020062. Rico, D., Martín-Diana, A.B., Martínez-Villaluenga, C., Aguirre, L., Silván, J.M.,
Krokou, A., Kokkinofta, R., Stylianou, M., Agapiou, A., 2020. Decoding carob flavor Dueñas, M., De Luis, D.A., Lasa, A., 2019. In vitro approach for evaluation of carob
aroma using HS–SPME–GC–MS and chemometrics. Eur. Food Res. Technol. 246, by-products as source bioactive ingredients with potential to attenuate metabolic
1419–1428. https://doi.org/10.1007/s00217-020-03500-4. syndrome (MetS). Heliyon 5, e01175. https://doi.org/10.1016/j.heliyon.2019.
Kyratzis, A.C., Antoniou, C., Papayiannis, L.C., Graziani, G., Rouphael, Y., Kyriacou, M. e01175.
C., 2021. Pod morphology, primary and secondary metabolite profiles in non-grafted Şahin, H., Topuz, A., Pischetsrieder, M., Özdemir, F., 2009. Effect of roasting process on
and grafted carob germplasm are configured by agro-environmental zone, genotype, phenolic, antioxidant and browning properties of carob powder. Eur. Food Res.
and growing season. Front. Plant Sci. 11, 1–19. https://doi.org/10.3389/ Technol. 230, 155–161. https://doi.org/10.1007/s00217-009-1152-7.
fpls.2020.612376. Santonocito, D., Granata, G., Geraci, C., Panico, A., Siciliano, E.A., Raciti, G., Puglia, C.,
Kyriacou, M.C., Antoniou, C., Rouphael, Y., Graziani, G., Kyratzis, A., 2021. Mapping the 2020. Carob seeds: food waste or source of bioactive compounds ? Pharmaceutics 12,
primary and secondary metabolomes of carob (Ceratonia siliqua l.) fruit and its 1090. https://doi.org/10.3390/pharmaceutics12111090.
postharvest antioxidant potential at critical stages of ripening. Antioxidants 10, Scalbert, A., Manach, C., Morand, C., Rémésy, C., Jiménez, L., 2005. Dietary polyphenols
1–21. https://doi.org/10.3390/antiox10010057. and the prevention of diseases. Crit. Rev. Food Sci. Nutr. 45, 287–306. https://doi.
Liu, Y., Ma, S.S., Ibrahim, S.A., Li, E.H., Yang, H., Huang, W., 2015. Identification and org/10.1080/1040869059096.
antioxidant properties of polyphenols in lotus seed epicarp at different ripening Sidina, M.M., El Hansali, M., Wahid, N., Ouatmane, A., Boulli, A., Haddioui, A., 2009.
stages. Food Chem. 185, 159–164. https://doi.org/10.1016/j. Fruit and seed diversity of domesticated carob (Ceratonia siliqua L.) in Morocco. Sci.
foodchem.2015.03.117. Hortic. 123, 110–116. https://doi.org/10.1016/j.scienta.2009.07.009.
Loullis, A., Pinakoulaki, E., 2018. Carob as cocoa substitute: a review on composition, Stavrou, I.J., Christou, A., Kapnissi-Christodoulou, C.P., 2018. Polyphenols in carobs: a
health benefits and food applications. Eur. Food Res. Technol. 244, 959–977. review on their composition, antioxidant capacity and cytotoxic effects, and health
https://doi.org/10.1007/s00217-017-3018-8. impact. Food Chem. 269, 355–374. https://doi.org/10.1016/j.
Lucci, P., Saurina, J., Núñez, O., 2017. Trends in LC-MS and LC-HRMS analysis and foodchem.2018.06.152.
characterization of polyphenols in food. TrAC - Trends Anal. Chem. 88, 1–24. Van Rijs, P., Fogliano, V., 2020. Roasting carob flour decreases the capacity to bind
https://doi.org/10.1016/j.trac.2016.12.006. glycoconjugates of bile acids. Food Funct. 11, 5924–5932. https://doi.org/10.1039/
Makris, D.P., Kallithraka, S., Mamalos, A., 2006. Differentiation of young red wines d0fo01158d.
based on cultivar and geographical origin with application of chemometrics of Vekiari, S.A., Ouzounidou, G., Ozturk, M., Görk, G., 2011. Variation of quality
principal polyphenolic constituents. Talanta 70, 1143–1152. https://doi.org/ characteristics in Greek and Turkish carob pods during fruit development. Procedia -
10.1016/j.talanta.2006.03.024. Soc. Behav. Sci. 19, 750–755. https://doi.org/10.1016/j.sbspro.2011.05.194.
Manach, C., Scalbert, A., Morand, C., Rémésy, C., Jiménez, L., 2004. Polyphenols: food Ydjedd, S., Chaalal, M., Richard, G., Kati, D.E., López-Nicolás, R., Fauconnier, M.L.,
sources and bioavailability. Am. J. Clin. Nutr. 79, 727–747. https://doi.org/ Louaileche, H., 2017. Assessment of antioxidant potential of phenolic compounds
10.1093/ajcn/79.5.727. fractions of Algerian Ceratonia siliqua L. pods during ripening stages. Int. Food Res.
Naghmouchi, S., Khouja, M.L., Romero, A., Tous, J., Boussaid, M., 2009. Tunisian carob J. 24, 2041–2049.
(Ceratonia siliqua L.) populations: morphological variability of pods and kernel. Sci. Ydjedd, Siham, Bouriche, S., López-Nicolás, R., Sánchez-Moya, T., Frontela-Saseta, C.,
Hortic. 121, 125–130. https://doi.org/10.1016/j.scienta.2009.02.026. Ros-Berruezo, G., Rezgui, F., Louaileche, H., Kati, D.E., 2017. Effect of in vitro
Nasar-Abbas, S.M., e-Huma, Z., Vu, T.H., Khan, M.K., Esbenshade, H., Jayasena, V., gastrointestinal digestion on encapsulated and nonencapsulated phenolic
2016. Carob Kibble: a bioactive-rich food ingredient. Compr. Rev. Food Sci. Food compounds of carob (Ceratonia siliqua L.) pulp extracts and their antioxidant
Saf. 15, 63–72. https://doi.org/10.1111/1541-4337.12177. capacity. J. Agric. Food Chem. 65, 827–835. https://doi.org/10.1021/acs.
jafc.6b05103.

10