Food Analytical Methods

https://doi.org/10.1007/s12161-020-01756-w

Development and Validation of an Analytical Method

for the Determination of Aloin Using Liquid Chromatography
Tandem Mass Spectrometry
Chris Anagnostopoulos 1 & George Ampadogiannis 1

Received: 30 August 2019 / Accepted: 12 April 2020

# Springer Science+Business Media, LLC, part of Springer Nature 2020

Abstract
A new simple analytical method was developed for the identification and quantification of aloin in aloe gel. The extraction was
performed using acetonitrile acidified with formic acid (1%, v/v) with no additional cleanup step. The compound was determined
using high-performance liquid chromatography coupled with triple quadrupole mass spectrometry (QqQ). The recovery rates
obtained were between 80 and 110% with a relative standard deviation < 26%, at four different concentration levels. The
calculated limit of quantification (LOQ) was at 0.05 mg kg−1.

Keywords Aloin . Method development . LC-MS/MS . Uncertainty . Validation

Introduction the future (Li Wan et al. 2017; Ming-Chin Lee et al. 2014; Qin
Pan et al. 2013; Kumar et al. 2010). Aloin, aloesin, polysac-
The genus Aloe vera are plants which belong to the charides, aloe-emodin, and aloe essential oil are the most bi-
Xanthorrhoeaceae family (subfamily: Asphodelaceae) with ologically active ingredients of these species (Maryam
more than 250 species that are cultivated around the world. Akaberia et al. 2016). Various pathogenic microbes are sensi-
Aloe vera is a perennial, drought-resisting, succulent plant tive to different Aloe vera extracts. Inhibitory action has been
belonging to the Asphodelaceae family. The name, aloe, is demonstrated against S. aureus, E. coli (Nejatzadeh-
derived from the Arabic “alloeh” or Hebrew “halal” meaning Barandozi 2013), and Helicobacter pylori (L. Celini et al.
bitter shiny substance (Karkala and Bidya 2014). Aloe vera 2014). Aloe vera extracts exhibit anti-inflammatory effects.
gel derived from the leaf pulp has many applications in the The gel from aloe has anti-hyperglycemic and anti-
food and medical industry. In the food industry, Aloe vera hypercholesterolemia effects in patients with type 2 diabetes
products include concentrates, gel fillets, juices, and powders (Ramesh Pothuraju et al. 2016). The anti-inflammatory effect
that can be used for the production of jam, jellies, candies, of aloe-emodin was comparable with that of kaempferol and
gum, ready-to-drink or soft juices, smoothies yoghurt, curd, quercetin, indicating aloe-emodin as a possible key constitu-
and others (Ahlawat and Khatkar 2011). In veterinary medi- ent responsible for the anti-inflammatory activity of aloe (Mi-
cine, aloe is used as a laxative and human medicine including Young Park and Sung 2009).
uses as bitter tonic, eupeptic, and cholagogic agents when The leaves, which are the part of the plant used, can be
given in small doses and as laxatives at higher doses separated into two products: aloe latex and aloe gel. Aloe
(European Medicine Agency (EMA) 2000). gel is the colorless gel contained in the inner part of the leaves
There are several references for the anticancer activity of and consists primarily of water (> 98%) and polysaccharides
aloin showing promise for application in cancer treatment in (Reynolds and Dweck 1999). Aloe latex (or aloe juice) is the
bitter yellow exudate from the pericyclic tubules in the outer
skin of the leaves. Aloe latex contains a series of glycosides
* Chris Anagnostopoulos known as anthraquinones, which are the phenolic compounds
[email protected] present in the sap or yellow exudates with the most prominent
being Aloin.
1 Aloin is an anthraquinone glycosyl, meaning that its an-
Department of Pesticides Control and Phytopharmacy, Laboratory of
Pesticides’ Residues, Benaki Phytopathological Institute, 8 St. Delta thraquinone skeleton has been modified by the addition of a
Street, Kifissia, 14561 Athens, Greece sugar molecule. Aloin extracted from natural sources is a
 Food Anal. Methods

mixture of two diastereomers, aloin A (or barbaloin) and aloin methodologies is presented in Table 1 where the lowest LOQ
B (or isobarbaloin), which have similar chemical properties. of aloin was achieved by using mass spectrometry. Due to the
At the EU level, the maximum permitted concentration is selectivity of the systems in MRM mode, in a single run con-
0.1 mg kg−1 in foodstuffs for human consumption with the firmation and quantification can be achieved; however, matrix
exception of alcoholic beverages where the limit is set at effect is still a significant problem with this technology but not
50 mg kg−1(European Council 1988b). The International in the form of interfering peaks in the chromatogram, like in
Aloe Science Council ((IASC) 2019) recommends the aloin other nonselective detectors, but in the form of ion
in products for oral consumption to be less than 10 mg kg−1 suppression.
(((IASC) 2019)). From surveys conducted in several aloe As an analytical matrix, aloe gel contains more than 98%
products (liquid samples), concentrations of aloin A or/and water, the rest is mostly occupied by polysaccharides and
aloe emodin were found below the LOQ of the analytical active compounds that may interfere the target analytes. In
method (0.005 mg L−1) or up to 263 mg L−1 (ElSohly et al. this study, a simple, easy, and robust analytical method was
2004; Kanama et al. 2015; Nobuyuki Okamura et al. 1996; developed and validated for the determination of aloin in aloe
Wang et al. 2012). gel using a simple and quick sample preparation procedure
Aloin A and B, with a molecular formula of C21H22O9, and determination by LC-MSMS. The proposed methodology
have a molecular weight of 418.398 g mol−1. Their structures can be used in routine analysis for the detection and quantifi-
are presented in Fig. 1 below. The water solubility can be cation of aloin in raw gel, processed gel, and final aloin prod-
estimated at 2.92 mg mL−1 showing a high solubility in water ucts (such as juice or beverages) ready for consumption.
with an estimated logP of 0.41. Due to this, the exactions of
the analyte from different matrices (leaves, gel, food, pharma-
ceutical products) were mostly preformed with polar solvents Material and Methods
or mixtures of an organic solvent with water like acetonitrile/
water (40/60, v/v). Reagents and Materials
Different analytical techniques were employed for the de-
termination of aloin A/B and aloe emodin. Liquid chromatog- Aloin reference standard (purity 94.9%) was purchased from
raphy is the technique used for the separation, using reverse- LGC. All the solvents, namely, acetonitrile, methanol, and
phase C18 columns in combination with UV/PDA (Azaroual water, were HPLC grade. For cleaning of the sample, PTFE
et al. 2012; Bozzi et al. 2007; Brown et al. 2014; Hsiu-Mei filters 13 mm diameter and 0.22 μm particle size were used
Chiang et al. 2012; Logaranjan et al. 2013; Tan et al. 2011; (Restek Cat 26147, 100pk.).
Zahn et al. 2008), DAD (Sánchez-Machado et al. 2017), or
fluorescence (Coran Silvia and Bambagiotti-Alberti Massimo Instrumentation
2011). The disadvantage of the methods using these detection
systems is that they involve more complex sample preparation For the determination of the analytes, liquid chromatography
procedures and have higher LOQ, which however are below (LC) in combination with triple quadrupole tandem mass
10 mg L−1. In the last years LC-MS/MS systems (Edder spectrometry (QqQ) with electron spay ionization was
Patrick et al. 2007) are more frequently used in the analysis employed. Two different LCMS-MS systems (Agilent and
of aloin A, B or/and emodin with low LOQs and short run Varian) were used. The methodology and analytical condi-
time (Wang et al. 2012). A summary of indicative analytical tions are presented below in detail.

Fig. 1 Chemical structures of aloin A (left) and B (right). Structure were drawn with ACD/ChemSketch (Freeware) 2018.1.1
 Food Anal. Methods

Table 1 Overview of analytical methodologies used for the analysis of aloin in various matrices

Matrix Scope Extraction principle Column Analytical system LOQ (mg/L) Reference

Aloe gel powders Aloin A Methanol sonicated extraction Cation-exchange UV – (Bozzi et al. 2007)
Aloin B HPLC column
Aloinoside A
Aloinoside B
5-Hydroxyaloin
Capsule, soft gel and Aloin A Acetonitrile/water (40:60, v/v) C18UPLC ESI(-)-MS/MS 0.005 (Wang et al. 2012)
liquid products Aloe emodin extraction with vortex
Aloe crude Aloin A Water extraction and vacuum drying C18HPLC UV-VIS 1.3 (aloin) (Hsiu-Mei Chiang et al. 2012)
Drugs Aloe emodin 0.5 (aloe-emodin)
Aloe juices, pills, Aloin A Shaking with ammonium acetate 2D-LC ESI(-)-MS/MS 0.0009 (aloin A) (Edder Patrick et al. 2007)
dairy products Aloin B 100 mM 0.0004 (aloin B)
Aloe ferox Miller and Aloin A Water or water methanol 50:50 C18HPLC UV 0.1 (Zahn et al. 2008)
aloe-related product
Aloe ferox (A) and Aloe Aloin Derivatization with H3BO3 in Silica gel 60 F254S, HPTLC densitometry 110 ng (Coran Silvia and
ferox (B) dried extracts, CH3OH as reagent 20 cm × 10 cm in fluorescence mode Bambagiotti-Alberti
pharmaceutical products Massimo 2011)
Fresh latex, fresh gel Aloin PBS buffered saline (PBS) for C18 HPLC DAD 0.012 (Sánchez-Machado et al. 2017)
Dry gel, dry latex extraction
Aloe crude Aloin, aloe-emodin Hydrolysis with HCl, ethyl acetate C18 HPLC UV-VIS 1.3 (aloin) (Hsiu-Mei Chiang et al. 2012)
Drugs 0.5 (aloe-emodin)
Raw materials and Aloin A & B Sonication with an acidified solvent C18 HPLC UV 0.23 (aloin A) 0.21 (Brown et al. 2014)
finished products (aloin B)
Leaves from Aloe vera Aloin A Sonication methanol C18 HPLC & Photodiode array 9.9 (Azaroual et al. 2012)
monolithic
C18 HPLC
Crude plant Materials Aloin A Ethanol and an aqueous two phase C8 HPLC Photodiode array 0.0038 Aloin A (LOD) (Logaranjan et al. 2013;
Aloin B system of polyethylene glycol/salt Tan et al. 2011)
Aloe emodin
 Food Anal. Methods

Experimental Table 2 Overview of the LCMS-MS parameters for aloin A and B

Analyte MRM LCMSMS01 LCMSMS02

Preparation of Standard Solutions
Capillary voltage CE Fragmentor CE
The stock solution of aloin was prepared by accurately
Aloin A 417 > 297 92 17 70 28
weighing 13.17 mg of the reference standard in volumetric
Aloin B 417 > 268 92 34 70 22
flasks (certified “A” class) and dissolving in 25 mL acetonitrile
reaching a final concentration of 500 mg/L−1. The stock stan- 417 > 251 92 46.5 70 56

dard solution was stored at − 20 °C. A working standard solu-

tion of 5 mg L−1 was prepared by diluting an aliquot of each System LC-MS/MS01
stock solution in acetonitrile. The working standard solution
was also stored at − 20 °C and before each use was left to reach Chromatographic separation was achieved by using a ProStar
room temperature. From this working standard solution, 2 se- 420 AutoSampler, an online degassing system and two Varian
ries of calibration standards were prepared within the range of
ProStar 210 pumps (Varian, Palo Alto, CA, USA) equipped
0.025–5 mg L−1 by serial dilution in acetonitrile or aloe extract. with a reverse-phase Zorbax SB C18 3.5 μm particle size,
150 mm × 2 mm analytical column (Agilent), at a flow rate
Laboratory Sample Pre-treatment of 250 μL/min with a mobile phase consisting of H2O/ACN
(90:10 v/v)—1 mM ammonium formate, 0.1 HCOOH (solvent
As to obtain a representative analytical sample, the laboratory A) and ACN/H2O (90:10 v/v)—1 mM ammonium formate,
sample treatment differs depending on the nature of the
0.1% HCOOH (solvent B). A gradient program was used
product. consisting of 80% of solvent A and 20% of solvent B, ramped
In the case of the whole leaf sample, the lower part of the linearly over the course of 8 min to 95% of solvent B. This
leaf in which a reddish decolorization is observed, is cut at
composition was held for a further 7 min before returning to
about 3 cm and removed. This was done since after the leaf is the initial condition. The column was re-equilibrated for 5 min
removed from the plant in the cutting point, the concentration at the initial mobile phase composition. The total run-time was
might be higher and thus not representative to the whole con-
15 min. The injection volume was 5 μL. In order to avoid
centration in the rest of the leaf. carryover, the autosampler syringe was purged with a mixture
In the case of aloe gel, the sample is stirred as to be ho- of methanol/water (50:50 v/v) before sample injection.
mogenized and an aliquot is taken with a spoon. Similar in the
Detection was achieved using a triple quadrupole mass
case of the liquid product of aloe, the sample is shaken or spectrometer (Varian model 1200 L) equipped with an
vortexed as to achieve homogenization and an analytical sam-
electrospray ionization interface operating in negative mode.
ple is taken with a pipette. Typical source parameters were as follows: capillary voltage
and collision energy varied depending on the precursor or
Analytical Sample Preparation product ion as shown in Table 2, source temperature was set
at 250 °C and drying gas temperature at 250 °C. Drying and
An aliquot of 2 g (2 ± 0.02) were weighed in a 50-mL poly- nebulizing gas was nitrogen generated from a high-purity gen-
propylene centrifuge tube and extracted with 4 mL acetoni-
erator, and their pressures were set at 18 and 55 psi, respec-
trile/H2O (80:20), 1% HCOOH for 1 min by shaking vigor- tively. For the operation in MS/MS mode, Argon 99.999%
ously by hand. The sample was then centrifuged (4000 rpm) was used as collision gas with a pressure of 1.8 mTorr. The
for 5 min, and the final extract was transferred into a screw cup
multiple-reaction monitoring experiments were conducted
storage vial and stored in the freezer until analysis. Before the with a dwell time of 100 ms. For instrument control, data
injection in the chromatographic system, the final solution acquisition, and processing, Varian MS Workstation software
was filtered through a 0.22-μm disposable PTFE syringe filter.
version 6.8 was used.
Following this extraction procedure, the concentration
C mg kg−1 of the analytes in the sample corresponds to 2*C
μg/mL of the analytes in the extract. System LC-MS/MS02

LC-ESI[-]-MS/MS Analysis LC-MS/MS analysis was performed by an Agilent Series

1200 liquid chromatograph with degasser (G1379B),
Validation experiments were performed in two different LC- autosampler (Hip/ALS G1367A) with thermostat (FC/ALS
MS/MS systems as to ensure the robustness and reproducibil- Therm G1330B), binary pump (G1312A), and a thermostatic
ity of the method. The conditions used for each system are column department (TCCG1316A) equipped with a reverse-
described below. phase Zorbax Eclipse XDBC 18 3.5-μm particle size,
Food Anal. Methods

150 mm × 2.1 mm analytical column (Varian, Palo Alto, CA, reproducibility for both systems (LCMSMS01 and
USA). Injection was performed using an injection program LCMSMS02) and expressing as relative standard deviation
(Anagnostopoulos et al. 2013) in which the samples is diluted (RSD %). The RSD was calculated for each spiking level.
with water minimizing the matrix effect. As mentioned, the Matrix effects are generally recognized as a suppression or
separation was achieved with a C18 column at a flow rate of enhancement of the analytical signal due to co-eluting matrix
285 μL/min with a mobile phase consisting of water with components. The magnitude of the matrix effect (ME)
5 mM ammonium formate, 0.1% formic acid and 0.02% ace- expressed in percentage peak enhancement or suppression
tonitrile as solvent A, and methanol with 5 mM ammonium compared with the peak of the analyte in pure solvent
formate and 0.1% formic acid as solvent B. The LC gradient (acetonitrile) was estimated as (Trufelli et al. n.d..)
program initially started with 20% solvent B until 8 min, in-
ΜΕ ð%Þ ¼ ðΜF–1Þ  100
creased to 100% at 9 min and remained until 15 min. The
column was re-equilibrated for 5 min at the initial mobile were MF (matrix factor) is calculated from the equation
phase composition. The total run-time was 20 min.
Detection was achieved using a triple quadrupole mass bmatrix
MF ¼
spectrometer (Agilent Triple Quad 6410) equipped with an bsolvent
electrospray ionization interface operating at the negative
were bmatrix and bsolvent are the slopes of the calibration curves
mode. Typical source parameters were as follows: capillary
of the analyte in matrix and solvent, respectively.
voltage and collision energy varied depending on the precursor
A value of ME = 0% means that no matrix effect occurred.
or product ion as shown in Table 2, and drying gas temperature
Negative values represent suppressions of the analyte signal,
was set at 300 °C. Drying and nebulizing gas was nitrogen-
and positive values stand for enhancements induced by matrix.
generated from a high-purity nitrogen generator (NitroFlow
Matrix effects are known to be both compound-dependent and
Basic Mobile, Parker Filtration & Separation B.V.) both at
matrix-dependent. In our cases, the matrix effect is considered
11 L/min and 30 psi. For the operation in MS/MS mode, ni-
to be important; this is the reason why the samples are not
trogen was used as collision gas with a pressure of 1.5 mTorr.
quantitated from the standard solutions in a solvent but in the
The multiple reaction monitoring experiments were conducted
matrix. The matrix effect can be classified in one of three
with a dwell time of 50 ms. For instrument control, Agilent
categories: soft (− 20 to 20%), medium (± 20–50%), and strong
Mass Hunter data acquisition Triple Quad B.01.04 was used,
(> 50% and ≤ 50%). Soft matrix effect is considered as ME (%)
and for data processing Agilent MassHunter Workstation
values within repeatability range values (− 20 to 20%) accept-
Quantitative Analysis B.01.04 was used.
ed in pesticide analysis (Ferrer et al. 2011).
In order to have an additional statistical confirmation for
the significance or not of the matrix effect, on the peak area
Validation Design
and therefore the sensitivity of the analytes, the slopes of the
linear calibration curves obtained for the solvent and the ma-
Due to lack of a specific European legislation regarding the
trix (aloe) were compared using a Student t test. tcal is defined
validation of analytical methods for aloin or related com-
in the following equation:
pounds and taking into consideration the background of the
laboratory (pesticide residue analysis), the European guide- jb1 −b2 j
lines (European Commission 2017) that apply to pesticide t cal ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
S b1 2 þ S b2 2
residue analysis were used to perform validation experiments.
Thus, the following parameters were assessed during valida- with b the slopes of the calibration curves and Sb the standard
tion of the analytical method: linearity, accuracy, precision, deviation of the slopes.
limit of quantification, (LOQ), limit of detection (LOD), ma- If the theoretical value (ttheo) of 2306 (df = 8 + 8 − 4 = 12,
trix effects (ME), and uncertainty (U). two-sided critical region, probability 95%) exceeds the calcu-
Linearity was studied using pure solvent (acetonitrile) and lated value tcal, the null hypothesis (that there is no significant
matrix extract (aloe gel extract) in eight concentrations (0.025, difference between the two calibration lines) is accepted.
0.05, 0.1, 0.25, 0.5, 1, 2.5, and 5 mg L−1).
The accuracy of the method was estimated by spiking
blank homogenized aloe gel samples with aliquot of working
standard solution at four concentration levels: 0.05, 0.1, 1, and Setting the LOQ and LOD
10 mg kg−1. The spiked samples were set aside for 30 min and
then processed according to the procedure described above. The limit of quantification (LOQ) of the method was set as the
Precision was estimated by accessing (a) the repeatability of lowest concentration of an analyte that can be determined with
the recovery experiments (n = 6) for each system (b) the acceptable precision (repeatability) and accuracy under the
 Food Anal. Methods

stated conditions (Eurachem 1998); however, a minimum The combined standard uncertainty (uc(X)) was finally cal-
signal-to-noise ratio (S/N) of ten spiked samples was desirable culated as follows:
and obligatory as for a chromatographic peak to be considered pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
sufficient. ucðX Þ ¼ U vol 2 þ U mass 2 þ U RefMas þ U bias 2 ð1Þ

Methods suitable for official control must produce results

with standard measurement uncertainties less than the maxi-
Uncertainty Estimation mum standard measurement uncertainty calculated using the
formula below:
The uncertainty of the method was calculated based on the U ¼ ucððX ÞÞ  k ð2Þ
Eurachem/Citac Guidelines (Ellison and Williams 2012a). For
most purposes in analytical chemistry, a relative uncertainty U The formulas for the calculation of the independent uncer-
should be used. The relative uncertainty provides an interval tainties are presented below.
within, which the value of the analyte concentration is be-
lieved to lie within a higher level of confidence. U is obtained
by multiplying the combined standard uncertainty uc(y), by a Estimation of Type B Uncertainty
coverage factor k. The combined uncertainty uc(y) was calcu-
lated from the summary squared of several independent pa- Uncertainty Associated with Volume Measurement (Uvol)
rameters as
Since a rectangular distribution is assumed, the uncertainty
(a) the volume uncertainty Uvol from the addition of the extraction solvent is calculated as
(b) the mass uncertainty Umass follows:
(c) the reference material URefMat
S
(d) the stock standard solutions Ustock uvol ¼ pffiffiffi ð3Þ
(e) the calibration uncertainty UCalCurv 3
(f) the bias uncertainty Ubias where S is the standard deviation from the calibration certifi-
cate of the mechanical pipette. The relative uncertainty Uvol is
In the cases were parameters are estimated though other calculated from the following formula:
sources like certificates, the type of estimation will be referred
uvol
as type A; in cases where a series of experimental values is U vol ¼ ð4Þ
used, type B. V sample

Fig. 2 Fragmentation pathway proposed for the [M–H]− ion of aloin

Food Anal. Methods

Table 3 Summary of calibration line parameters at concentration levels 0.025–10 mg L−1 for both analytes in solvent, and aloe gel extract

System Matrix n r r2 b Sb a Sa UCalCurv Intersect at zero

LCMSMS01 Solvent 12 0.999 0.998 8.71E + 07 1.50E + 06 9.03E + 06 3.02E + 06 0.080 Yes
LCMSMS02 12 0.998 0.997 1.67E + 06 2.57E + 04 − 1.13E + 05 5.18E + 04 0.101 Yes
LCMSMS01 Aloe gel 12 0.992 0.984 9.05E + 07 4.70E + 06 1.02E + 07 9.48E + 06 0.241 Yes
LCMSMS02 12 0.995 0.990 1.32E + 06 3.63E + 04 − 1.42E + 05 7.32E + 04 0.180 Yes

n number of levels, b slope of the regression line, Sb mean standard deviation of the slope of the regression line, Sa mean standard deviation of the mean of
the population that corresponds to x = 0, r correlation coefficient, r2 correlation coefficient squared, UCalCurv the uncertainty in the predicted value C
(mg/kg) due to the variability in the area

Uncertainty Associated with Mass Measurement (Umass) Uncertainty Associated with Bias (Ubias)

The uncertainty from the measurement of the mass is estimat- The uncertainty related with the bias was estimated though a
ed though the calibration certificate of the balance used Ubal. series of measurements of spiked samples at different levels
Since the distribution is the same with the one of the volume and in different analytical systems and thus is calculated
measurement, the same formula is applied. The relative uncer- though the following formula:
tainty Umass is calculated from the following formula: qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
U bias ¼ RMSbias 2 þ uC recovery 2 ð7Þ
U bal
U mass ¼ ð5Þ qffiffiffiffiffiffiffiffiffiffiffiffi2ffi
msample
where RMSbias ¼ Σ ðbias n
Þ
is the average (root mean square)
where msample is the mass of the analytical sample. bias and uC recovery is the uncertainty related to the
spiked concentration and calculated as uC recovery ¼
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Uncertainty Associated with the Standard Reference Material uðRefMatÞ2 þ uðvolÞ2 .
(URefMat)

The relative uncertainty of the reference material of aloin used

Results and Discussion
for quantification (and identification) is calculated from the
same formula (3) as above. In this case, S is the standard
Optimization of LC-MS/MS Conditions
deviation from the certificate of the purchased material.
The use of MRM mode based on QqQ mass spectrometry
Estimation of Type A Uncertainty allows low analyte detectability and is, at present, one of the

Uncertainty Associated with Calibration Curve (UCalCurv)

Table 4 Mean recoveries and RSD (%) for aloin in aloe gel in 2
different LCMS-MS systems
The uncertainty associated with the linear least square fitting
procedure is calculates as follows: Level (mg kg−1)
vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
u  2 0.05 0.1 1 10
u
Sy u 1 1 yo −y n 6 6 6 6
U CalCurv ¼ u þ þ  2 ð6Þ
b tN n LCMSMS01
b2 þ ∑Ni¼1 xi −x MeanRecovery (%) 93.5 96.7 98.5 88.1
SDr (%) 26.6 3.0 9.7 9.9
where Sy is the residual standard deviation, b is the slope of LCMSMS02
the calibration curve, N is the number of calibration levels (x), MeanRecovery (%) 70.8 76 120.2 107.7
n is the number of values per level, yo is the experimental SDr (%) 11.5 9.7 2.3 5.2
value of y for which the concentration xo, y is the average of Combined
the yi values and ∑Ni¼1 ðxi −xÞ2 i = 1 is the sum of the obtained n 12 12 12 12
concentration (given by the calibration curve) minus the aver- MeanRecovery (%) 82.1 86.4 97.9 109.4
age concentration of the standards used in the calibration SDr (%) 25.8 14.0 12.7 12.1
curve.
 Food Anal. Methods

Fig. 3 a Multiple Ion transition chromatograms of aloin at 0.016 mg kg−1 in aloe matrix in LC-MS/MS01 system. b Multiple Ion transition
chromatograms of aloin at 0.01 mg kg−1 in the aloe matrix in LC-MS/MS02 system

most selective approaches for trace analysis. However, as to maximize the signal for each pesticide. The first step was to
achieve the maximum sensitivity, in terms of S/N, a precise select the parent ion (precursor ion). Since electron spray ion-
optimization of MS/MS parameters is needed in order to ization (ESI) is a “soft” ionization mode compared with other
Food Anal. Methods

techniques (e.g., electron impact in GC-MS or MS/MS sys-

tems), the deprotonated ion [M-H]− (417 m/z) was selected
and the highest abundance was optimized at capillary or
fragmentor voltage of 70 V depending on the system. The
source optimization was achieved by introducing the analyte
into the mass spectrometer through direct infusion via a sy-
ringe pump Model “11” Plus (Harvard apparatus, MA, USA)
at a flow rate of 250 mL/min. Breakdown curves for each ion
were automatically constructed by the software. The break-
down curve demonstrates the ion intensity as function of ex-
citation amplitude (volts). The optimum value as given by the
breakdown curves of the cone voltage and collision energy
that gave maximum response for each precursor ion and its
products was selected. Aloin showed 3 abundant product ions
297, 268, and 251. The first one can be explained by the loss
of C4H8O4 (417 m/z → 297 m/z), the second by C5H9O5
(417 m/z → 268 m/z) and the last by the loss of C6H11O5
(417 m/z → 251 m/z). The fragmentation pathway proposed
for the [M–H]− ion of aloin is presented in Fig. 2.

Method Validation

Good linearity with satisfactory coefficients, were obtained in

both solvent and aloe matrix, ranging from 0.992 to 1. The
basic calibration line parameters are presented in Table 3.
Thus, a reliable quantification can be performed in the whole
calibration range, using either the calibration line, single point
calibration or bracketing.
The matrix effect, as estimated from the comparison of the
calibration slopes, was 21% and (−)3.8% for LCMSMS02 and
LCMSMS01, respectively, indicating that the significance of
the matrix effect is not straightforward and varied depending
on the system used. In the case of LCMSMS02, the matrix
effect is medium but significant (tcal = 7.9 > ttheo) but in the
case of LCMSMS01 is considered soft and not significant (t-
cal = 0.67 < ttheo). It should be noted that for both systems the
ionization and MS parameters were optimized accordingly;
therefore, the difference can be attributed to the different ion
kinetics in the source or the ion formation due to different
buffer concentrations and pH in the mobile phases.
However, as to ensure adequate and reliable quantification,
the measurements were conducted with matrix-matched cali-
bration standards. This procedure is also common practice in
analytical methods with a large scope in which no internal
standards are used. Even if the method is “targeted” for aloin,
due to the medium matrix effect in one of the 2 systems, the
calibration standard should be in the matrix as to achieve
reliable quantification in all systems. However, an in-house
validation for the implementation and duplication on the
method in a routine laboratory that can prove the non-
significance of the matrix effect could waive the need for
Fig. 3 (continued) matrix-matched calibration standards for the specifically
laboratory.
 Food Anal. Methods

Fig. 4 Ishikawa diagram of the main uncertainties evaluated in aloin analysis by the developed analytical method

Recoveries obtained were satisfactory in both systems Uncertainty Analysis

and ranged from 88.1 to 120.2% with RSD < 26.6%. At the
lowest fortification level (0.05 mg kg−1), recoveries were The main uncertainty sources were identified and quantified,
82.1% and RSD < 25.8% (average of both systems). The followed by the determination of the combined uncertainty
mean recoveries and RSD values per system and level are (uc(x)) using a Gauss propagation model. Relative uncertainty
presented in Table 4. (U) was calculated using a coverage factor k = 2 (95% of con-
LOQ is an important validation parameter and was used fidence level): U = uc(X) × k (European Commission 2017;
to evaluate the sensitivity of the method. The LOQ was set Ellison and Williams 2012b).
at 0.05 mg kg−1 which is the lowest fortification level. The identification of all relevant uncertainty sources are
According to Council Directive 88/388/EEC of 22 presented graphically in a case and effect (Ishikawa) diagram
June 1988 (European Council 1988a), the maximum con- (Fig. 4). A detailed description of the estimation is presented
tent of aloin (synonym: barbaloin) in foodstuffs for human below.
consumption is set to 0.1 mg kg−1 in both foodstuffs and
beverages, with the exception of alcoholic beverages,
which may contain up to 50 mg kg−1 aloin. In both cases, Estimation of Type B Uncertainty
the estimated LOQ of 0.05 mg kg−1 covers the EU legisla-
tion requirements. For the uncertainty associated with volume measurement
The estimation of the LOD was done using the equation (Uvol) of the extraction solvent, the standard deviation from
LOQ = 3.3 × LOD. Since the LOQ was set at 0.05 mg kg−1, an the calibration certificated of the micropipette (max volume
LOD of 0.015 mg kg−1 can be calculated. 5 mL) is 0.05% and used in the calculations. Since a volume
The selectivity of the discussed LC–MS/MS method was of 4 mL is used for extraction, the relative uncertainty is cal-
investigated by analyzing blank matrix samples. The extracted culated as
chromatograms did not show any signals for neither aloin A or 0:0025
B nor any interfering compounds at the corresponding reten- pffiffiffi
3 0:00144 mL
tion time. U vol ¼ ¼ ¼ 0:00036
4 4 mL
In Fig. 3 a and b, chromatograms of matrix-matched
calibration standards at 0.05 mg/L−1 in the aloe matrix is For the uncertainty associated with mass measurement
presented for both transitions and in both LC-MS/MS (Umass) of the analytical sample, the standard deviation from
systems used. the calibration certificate of the balance was used in the
Food Anal. Methods

calculations. Since the analytical sample used is 2 g, the rela- The relative uncertainty U is calculated by multiplying the
tive uncertainty is calculated as combined standard uncertainty with a coverage factor of 2 to
give
0:02037
U mass ¼ ¼ 0:0102
2 U ¼ 2  18:9 ¼ 37:8%
For the uncertainty associated with the standard reference
material (URefMat), the standard deviation from the calibration
certificate of the analytical standard is 0.5% and used in the
calculations. The relative uncertainty is calculated as Conclusions
0:5 A new simple analytical method was developed for the iden-
U RefMat ¼ pffiffiffi ¼ 0:00289
100 3 tification and quantification of aloin in aloe gel. A “dilute and
shoot” method was applied for the extraction, without a clean-
up step and the determination was performed with LCMS-MS
systems in negative ionization, using MRM mode. The meth-
Estimation of Type A Uncertainty od was validated, obtaining good results in terms of trueness,
reproducibility, and repeatability, but with a significant matrix
The linearity was investigated in a range of 0.025–10 mg L−1. effect which however is estimated and does not affect the
The uncertainty associated with the calibration curve sensitivity or selectivity.
(UCalCurv) is 0.18 (18%) for a calibration line in aloe extract
and 0.101 (10.1%) in solvent.
Compliance with Ethical Standards
The uncertainty related with the bias (accuracy and predic-
tion) was estimated though a series of 48 measurements of Conflict of Interest Chris Anagnostopoulos declares that he has no con-
spiked samples at 4 different levels (0.05, 0.1, 1, 10 mg L−1) flict of interest. George Ampadogiannis declares that he has no conflict of
and in 2 different analytical systems (LCMSMS01 and interest.
LCMSMS02). The relative uncertainty is calculated as
Ethical Approval This article does not contain any studies with human
RMSbias ¼ 18:9 participants or animals performed by any of the authors.

As mentioned, the standard deviation from the calibration Informed Consent Not applicable
certificate of the micropipette is 0.05%; therefore,
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
uC recovery ¼ uðRefMatÞ2 þ uðvolÞ2 ¼ 0:002892 þ 0:052 ¼ 0:05
References
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Therefore, Ubias is estimated as 18:92 þ 0:052 = 18.89. (IASC) (2019). IASC Aloe vera Frequently Asked Questions. https://
The above value includes the uncertainty from the calibration www.iasc.org/Consumers/AloeVeraFAQ.aspx. Accessed 03/04/
curve. 2019
The combined standard uncertainty (uc(X)) associated Ahlawat KS, Khatkar BS (2011) Processing, food applications and safety
of aloe vera products: a review. J Food Sci Technol 48:525–533.
with the analytical results at the LOQ concentration lev- https://doi.org/10.1007/s13197-011-0229-z
el was calculated according to the procedure described Anagnostopoulos C, Bourmpopoulou A, Miliadis G (2013) Development
above in “Uncertainty estimation,” and results are and validation of a dispersive solid phase extraction liquid chroma-
shown below: tography mass spectrometry method with electrospray ionization for
the determination of multiclass pesticides and metabolites in meat
and milk. Anal Lett 46:2526–2541. https://doi.org/10.1080/
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 00032719.2013.803251
ucðX Þ ¼ U vol 2 þ U mass 2 þ U RefMas 2 þ U bias 2 Azaroual L, Liazid A, Barbero GF, Brigui J, Palma M, Barroso CG
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi (2012) Improved chromatographic methods for determination of
¼ 0; 000362 þ 0:01022 þ 0:002892 þ 18:92 bioactive compounds from Aloe vera leaves. Int Sch Res Netw
ISRN Chromatogr 609095:7. https://doi.org/10.5402/2012/609095
¼ 18:9% Bozzi A, Perrin C, Austin S, Arce Vera F (2007) Quality and authenticity
of commercial Aloe vera gel powders. Food Chem 103:22–30.
https://doi.org/10.1016/j.foodchem.2006.05.061
The contribution of the type B uncertainty is considered
Brown PNY, Kuan R, Hui C, Finley J, Mudge EM, Steven D (2014)
negligible compared with type A indicating that the perfor- Determination of aloin A and aloin B in Aloe vera raw materials
mance of the method in the laboratory is the most crucial and finished products by high-performance liquid chromatography:
parameter as to have a reliable measurement. single-laboratory validation. J AOAC Int 97:1323–1328. https://doi.
org/10.5740/jaoacint.13-028
 Food Anal. Methods

Coran Silvia A, Bambagiotti-Alberti Massimo BG (2011) Selective de- Maryam Akaberia ZS, Javadi B, Sahebkar A, Emami SA (2016)
termination of aloin in different matrices by HPTLC densitometry in Therapeutic effects of Aloe spp. in traditional and modern medicine:
fluorescence mode. J Pharm Biomed Anal 54:422–425. https://doi. a review. Biomed Pharmacother 84:759–772. https://doi.org/10.
org/10.1016/j.jpba.2010.09.019 1016/j.biopha.2016.09.096
Edder Patrick KM, Yves J, Olivie Z, (2007) Aloin determination by 2D- Mi-Young Park H-JK, Sung M-K (2009) Evaluation of aloin and aloe-
LC with MS/MS detection in foodstuffs containing Aloe vera. Paper emodin as anti-inflammatory agents in aloe by using murine macro-
presented at the 3rd International Symposioum on recent advances phages. Biosci Biotechnol Biochem 73:828–832
in food analysis, Prague, 7–9 November 2007 Ming-Chin Lee J-DL, Huang W-L, Fang-Yi Jiang Y-ZJ, Jin Y-Y, Tseng Y-
Ellison SLR, Williams A, (2012a) EURACHEM/CITAC guide: quanti- S (2014) Aloin-induced cell growth arrest, cell apoptosis, and au-
fying uncertainty in analytical measurement, 2nd edition tophagy in human non-small cell lung cancer cells. Biomarkers
ElSohly MA, Gul W, Murphy TP (2004) Analysis of the anthraquinones Genomic Med 6:144–149
aloe-emodin and aloin by gas chromatography/mass spectrometry. Nejatzadeh-Barandozi F (2013) Antibacterial activities and antioxidant
Int Immunopharmacol 4:1739–1744. https://doi.org/10.1016/j. capacity of Aloe vera. . Organic and Medicinal Chemistry Letters 3
intimp.2004.07.005 Nobuyuki Okamura MA, Hine N, Yagi A (1996) High-performance liq-
Eurachem (1998) The fitness for purpose of analytical methods: a labo- uid chromatographic determination of phenolic compounds in aloe
ratory guide to method validation and related topics vol species. J Chromatogr A 746:225–231
EURACHEM Guide Qin Pan HP, Lou H, Xu Y, Lu T (2013) Inhibition of the angiogenesis and
European Commission (2017) Guidance document on analytical quality growth of Aloin in human colorectal cancer in vitro and in vivo.
control and method validation procedures for pesticide residues and Cancer Cell Int 13:1–9
analysis in food and feed Ramesh Pothuraju RKSSKO, Singh S, Hussain SA (2016)
European Council (1988a) Council Directive (88/388/EEC) 29:1–10 Hypoglycemic and hypolipidemic effects of Aloe vera extract prep-
European Council (1988b) Directive 88/388/EEC of 22 June 1988 on the arations: a review. Phytother Res 30:200–207
approximation of the laws of the Member States relating to Reynolds T, Dweck AC (1999) Aloe vera leaf gel: a review update. J
flavourings for use in foodstuffs and to source materials for their Ethnopharmacol 68:3–37
production, Ellison SLR , Williams A (2012b) EURACHEM/CITAC Guide: quanti-
European Medicine Agency (EMA) (2000) Aloes fying uncertainty in analytical measurement
Ferrer C, Lozano A, Agüera A, Girón AJ, Fernández-Alba AR (2011) Kumar S, Sreeraj Gopi DPM, Sivakumar S, Narasimhan S (2010)
Overcoming matrix effects using the dilution approach in Synthesis of cytotoxic and antioxidant Schiff’s base analogs of
multiresidue methods for fruits and vegetables. J Chromatogr A aloin. J Asian Nat Prod Res 12:360–370
1218:7634–7639. https://doi.org/10.1016/j.chroma.2011.07.033 Sánchez-Machado DI, Jaime L-C, Mariscal-Domínguez MF, Cruz-Flores
Chiang H-M, Lin Y-T, Hsiao P-L, Yu-Han S, Tsao H-T, Wen K-C (2012) P, Campas-Baypoli ON, Cantú-Soto Ernesto U, Ana S-S (2017) An
Determination of marked components —aloin and aloe-emodin— HPLC procedure for the quantification of aloin in latex and gel from
in Aloe vera before and after hydrolysis. J Food Drug Anal 20:646– Aloe barbadensis leaves. J Chromatogr Sci 55:251–257
652. https://doi.org/10.6227/jfda.2012200311 Tan Z, Jian LF, Xing JM (2011) Separation and purification of aloe
Kanama SK, Viljoen AM, Kamatou GPP, Chen W, Sandasi M, Adhami anthraquinones using PEG/salt aqueous two-phase system separa-
H-R, Van Wyk B-E (2015) Simultaneous quantification of anthrones tion. Sci Technol 46:1503–1510
and chromones in Aloe ferox (“cape aloes”) using UHPLC–MS. Trufelli H, Palma P, Famiglini G, Cappiello A An overview of matrix
Phytochem Lett 13:85–90. https://doi.org/10.1016/j.phytol.2015. effects in liquid chromatography–mass spectrometry. Mass
04.025 Spectrom Rev 30:491–509. https://doi.org/10.1002/mas.20298
Karkala M, Bidya B (2014) Aloe vera: a wonder plant its history, culti- Wang PG, Zhou W, Wamer WG, Krynitsky AJ, Rader JI (2012)
vation and medicinal uses. J Pharm Phytochem 2:85–88 Simultaneous determination of aloin A and aloe emodin in products
L. Celini SDB, Di Campli E, Genovese S, Locatelli M, Di Giulio M containing Aloe vera by ultra-performance liquid chromatography
(2014) In vitro activity of Aloe vera inner gel against Helicobacter with tandem mass spectrometry. Anal Methods 4:3612–3619.
pylori strains. Lett Appl Microbiol 59:43–48. https://doi.org/10. https://doi.org/10.1039/c2ay25599e
1111/lam.12241 Zahn M, Trinh T, Jeong ML, Wang D, Abeysinghe P, Jia Q, Ma W (2008)
Li Wan LZ, Fan K, Wang J (2017) Aloin promotes A549 cell apoptosis A reversed-phase high-performance liquid chromatographic method
via the reactive oxygen species-mitogen activated protein kinase for the determination of aloesin, aloeresin a and anthraquinone in
signaling pathway and p53 phosphorylation. Mol Med Rep 16: Aloe ferox. Phytochem Anal 19:122–126. https://doi.org/10.1002/
5759–5768 pca.1024
Logaranjan K, Devasena T, Pandian K (2013) Quantitative detection of
aloin and related compounds present in herbal products and Aloe Publisher’s Note Springer Nature remains neutral with regard to jurisdic-
vera plant extract using HPLC method. Am J Anal Chem 4:600–605 tional claims in published maps and institutional affiliations.