Extraction and preconcentration of compounds from the l-tyrosine metabolic pathway prior to their micellar electrokinetic chromatography separation

Journal of Chromatography A 1620 (2020) 461032

Contents lists available at ScienceDirect

Journal of Chromatography A
journal homepage: www.elsevier.com/locate/chroma

Extraction and preconcentration of compounds from the l-tyrosine
metabolic pathway prior to their micellar electrokinetic
chromatography separation
∗
Natalia Miekus
˛
, Alina Plenis, Marta Rudnicka, Natalia Kossakowska, Ilona Oledzka,
˛
Piotr Kowalski, Tomasz Baczek
˛
´ sk, Hallera, 107, 80-416, Gdan
´ sk, Poland
Department of Pharmaceutical Chemistry, Medical University of Gdan

a r t i c l e

i n f o

Article history:
Received 15 October 2019
Revised 9 March 2020
Accepted 10 March 2020
Available online 12 March 2020
Keywords:

Biogenic amines
Capillary electrophoresis
Hierarchical cluster analysis
Solid-phase microextraction
Solid-Phase Extraction
Dispersive Liquid-Liquid Microextraction

a b s t r a c t
The prominent biological effects of adrenaline (A), noradrenaline (NA) and dopamine (DA) as well as
the clinical importance of their metabolites (such as dihydroxyphenylacetic acid (DOPAC), methoxy–4hydroxyphenyl glycol (MHPG), dihydroxyphenylglycol (DHPG), metanephrine (M), normetanephrine (NM),
vanillylmandelic acid (VMA), homovanillic acid (HVA)) have forced researchers to evaluate new analytical methodologies for their isolation and preconcentration from biological samples. For this reason, the
three most popular extraction techniques (dispersive liquid-liquid microextraction (DLLME), solid-phase
extraction (SPE), solid-phase microextraction (SPME)) were tested. Micellar electrokinetic chromatography (MEKC) – a mode of capillary electrophoresis – with a diode array detector (DAD) was applied to
assess the extraction efficiency. Next, the enrichment factor (EF) of each applied method was calculated
in respect to standard mixtures of the analytes at the same concentration levels. The EF results of seven
selected metabolites of biogenic amines (BAs) from urine after sample preparation procedures based on
twenty-five different protocols (one DLLME, thirteen SPE and eleven SPME) were calculated and compared using hierarchical cluster analysis (HCA). The SPE as well as SPME procedures were proved to be
the most effective approaches for the simultaneous extraction of the chosen compounds. Moreover, an
ionic liquid (IL) – 1-ethyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide – added to methanol in
SPME additionally could successfully improve the extraction efficiency. It was also confirmed that the
HCA approach could be considered a supportive tool in the selection of a suitable sample preparation
procedure for that group of endogenous substances.
© 2020 The Authors. Published by Elsevier B.V.
This is an open access article under the CC BY-NC-ND license.
( />
1. Introduction
The secretory cells of the adrenal medulla mainly produce catecholamines: adrenaline (A), noradrenaline (NA) and dopamine
(DA). The metabolism of these relevant compounds takes place
mainly in the gastrointestinal tract (GI), intraneurally and in the
adrenal medulla owing to two enzymes: catechol-O-methyl transferase (COMT) and monoamine oxidase (MAO) (Fig. 1).

The main metabolites of DA are: dihydroxyphenylacetic acid
(DOPAC) and homovanillic acid (HVA), whereas A is converted into
metanephrine (M) and further into 3–methoxy–4-hydroxyphenyl

∗

Corresponding author.
E-mail addresses: , (N.
Miekus).
˛

glycol (MHPG). By the actions of alcohol dehydrogenase in the
liver, MHPG is metabolized to vanillylmandelic acid (VMA). The
NA metabolic pathway end products are also MHPG and VMA,
but NA is also metabolized intraneurally to dihydroxyphenylglycol (DHPG) and in the adrenal medulla to normetanephrine (NM)
[1]. The physiological metabolism of those monoamine neurotransmitters (NTs) could be interrupted (or their ratios of concentration
visibly changed) in pathophysiological stages of the human organism, which include neuroendocrine tumors (NETs) – pheochromocytoma (PHE) and neuroblastoma (NBL) [2,3]. As such, the determination of the concentration of HVA and VMA in urine samples
remains the gold standard for the biochemical diagnosis of NETs
[4]. Furthermore, the determination of DHPG and MHPG in plasma
samples could provide reliable information regarding the effects of
COMT and MAO on NA [5]. Nevertheless, the determination of Omethylated metabolites (M and NM) in plasma or urine samples

/>0021-9673/© 2020 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license. ( />

2

N. Miekus,
˛
A. Plenis and M. Rudnicka et al. / Journal of Chromatography A 1620 (2020) 461032

Fig. 1. Scheme of the main neuronal and non-neuronal pathways of dopamine, noradrenaline and adrenaline metabolism with pKa values for corresponding metabolites.
Legend: DA, dopamine; NA, noradrenaline; A adrenaline; DOPAC, 3,4-dihydroxyphenylacetic acid; DHPG, 3,4-dihydroxyglycol; NM, normetanephrine; M, metanephrine; HVA,
homovanillic acid; MHPG, 3–methoxy–4-hydroxyphenylglycol; VMA, vanillylmandelic acid; MAO, monoamine oxidase; COMT, catechol O-methyltransferase; ADH, alcohol
dehydrogenase.

has been shown to have higher sensitivity towards the diagnosis of
both NETs than the estimation of catecholamines or the concentration of VMA and HVA [6]. The levels of M and NM were evaluated
not only to give more reliable data during the diagnosis of PHE, but
also their concentration was positively correlated with the size and
adrenal or extra-adrenal location of a tumor. Even though the analysis of the concentration of M and NM in plasma is considered as
more appropriate than urine samples for the diagnosis of PHE, the
concentration of those two compounds in tumor tissues is usually
3 orders of magnitude higher than in plasma samples [5].
The precise diagnosis and description of the localization and
size of a tumor of a neuroendocrine origin (PHE or NBL) requires
the simultaneous analysis of the main catecholamine metabolites
from the biological specimens of patients. Modern, high throughput analytical methods could provide a great tool for the fast
and sensitive profiling of metabolites. However, the extraction of
trace amounts of metabolites needs to be evaluated since the
level of each metabolite is extremely low. Therefore, applied isolation methods should also have the advantage of preconcentrating analytes in the biological sample. To address this issue, the
three most popular and efficient extraction and preconcentration
techniques: dispersive liquid-liquid microextraction (DLLME), solidphase extraction (SPE) and solid-phase microextraction (SPME)
were tested. The DLLME procedure is simple and fast in terms of
sample preparation and is rarely used for complex biological matrices, which require the removal of ballast substances, such as
proteins. SPE is widely employed to concentrate and purify bio-

logical samples before analysis. SPME holds some advantages over
traditional sample preparation methods, such as little consumption of toxic and hazardous organic solvents and the relative ease
of online coupling to chromatographic systems. However, for each
extraction technique, the key extraction parameters affecting the

extraction efficiency should be optimized. For each of these approaches, the extraction efficiency was evaluated by the calculation of the enrichment factor (EF). This was done through a comparison of the signal intensity of the analytes in respect to signals
obtained for standard mixtures of the compounds of interest at
the same concentration levels. The signal intensities (peak heights)
were determined using an optimized micellar electrokinetic chromatography (MEKC) method coupled with a diode array detector
(DAD) for the determination of DHPG, VMA, MHPG, HVA, NM, M
and DOPAC in human urine samples. Additionally, three ionic liquids (ILs), namely: 1–butyl–3-methylimidazolium tetrafluoroborate
(IL1), 1-ethyl-3-methylimidazolium tetrafluoroborate (IL2) and 1ethyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide (IL3)
were tested as the desorbent additives for SPME to evaluate their
utility for a new class of compounds different to those tested in
our previous studies [7]. The decision to apply ILs was driven by
the fact that the use of traditional organic solvents could lead to
pollution of the natural environment and danger to human health.
ILs are believed to constitute alternative solvents, characterized
by low volatility, and chemical and physical stability [7–9]. Some
modern analytical methods have been optimized with the use of
distinct ILs during the analysis of biogenic amines (BAs) [7,8,10,11].


N. Miekus,
˛
A. Plenis and M. Rudnicka et al. / Journal of Chromatography A 1620 (2020) 461032

3

In our earlier studies, experimental data demonstrated the usefulness of IL2 at a concentration of 20 ng/mL in SPME, followed by
MEKC for the determination of DA, A, NA, l-tyrosine (L-Tyr) and ltryptophan (L-Tryp) from human urine samples [7]. Owing to the
optimized method, the extraction yields of BA precursors together
with some basic BAs increased from nearly 9 times for l-Tryp up
to 21 times for A [7].
The main objective of the presented work was to develop and

optimize an effective extraction protocol for seven analyzed BAs
(DHPG, VMA, MHPG, HVA, NM, M and DOPAC) from human urine
samples. In this study, different modifications of three sample
preparation procedures based on SPME, SPE and DLLME were investigated and discussed, in order to select the most efficient procedure providing the largest recovery of analytes and the most effective purification of the sample matrix. It was decided to use hierarchical cluster analysis (HCA) in order to check whether it could
be considered as a useful tool facilitating the selection of the most
effective sample preparation procedure for specific analytes. This
powerful analytical platform, consisting of the most appropriate
sample preconcentration, enrichment and analysis methods, was
evaluated for each of the studied compounds.

MEKC method. The reference urine samples were prepared daily,
just before use, by diluting the stock solution as appropriate with
2 mM of sodium tetraborate decahydrate (to a final concentration
of each analyte of 10 μg/mL). The stock standard solutions were
kept in a freezer (−20 °C), in closed containers and new solutions
were prepared once every two weeks. The working solutions were
stored at 4 °C in closed containers for a maximum of 7 h.

2. Materials and methods

2.5. DLLME conditions

2.1. Chemicals and reagents

1 mL of the human urine sample was spiked with the working
solution of analytes at a concentration of 10 μg/mL. Subsequently,
130 μL of cold acetone was added and the samples were shaken for
5 min (laboratory shaker – Elpin 358S, Lubawa, Poland) and centrifuged for 5 min (12 0 0 0 g) (laboratory centrifuge – MPW-211
or MPW- 350R, Warsaw, Poland). Next, 1 mL of supernatant was
separated and placed in a clean glass tube and then 1 mL of EtOH

and 500 μL of DCM were added. The samples were shaken mechanically for 10 min and centrifuged for 4 min (4 0 0 0 g). In order
to separate the organic phase, 400 μL of the solution was taken
from the bottom of the glass tube and transferred to a clean Eppendorf tube and then evaporated to dryness at 45 °C (Labconco®,
Kansas City, Missouri, USA). The residue was dissolved in 150 μL of
2 mM sodium tetraborate using a long vortex time (2 min for each
sample). Then the samples were injected into the capillary and analyzed by the elaborated MEKC method.

Methanol (MeOH), hexane, ethanol 96% (EtOH) and acetone were supplied by POCH (Gliwice, Poland). Reagents, such
as sodium dodecyl sulfate (SDS), DHPG, VMA, MHPG, HVA,
M, NM, DOPAC, acetonitrile (ACN), dichloromethane (DCM),
1–butyl–3-methylimidazolium tetrafluoroborate (IL1), 1-ethyl3-methyl-imidazolium tetrafluoroborate (IL2) and 1-ethyl-3methylimidazolium bis(trifluoromethylsulfonyl)imide (IL3) were
supplied by Sigma-Aldrich (Darmstadt, Germany). Sodium tetraborate decahydrate (borax), boric acid and sodium hydroxide (NaOH)
were obtained from Merck (Darmstadt, Germany). Capillary Regenerator Basic Wash Solution was purchased from Beckman
Coulter (CA, USA). All chemicals were of analytical grade and were
applied without further purification. The purified water used in
all experiments was obtained from Milli-Q equipment (Millipore,
Bedford, MA, USA).

2.4. MEKC conditions
For MEKC separation, the following parameters were applied:
uncoated fused silica capillary with an effective/total length of
50/60.2 cm and 75 μm i.d.; wavelength of UV detection 200 nm;
hydrodynamic injection (15 s, at 0.5 psi); total analysis time
20 min; applied voltage 22 kV; temperature 25 (± 0.1) °C. Between
each run, the capillary was rinsed with 0.1 M NaOH for 1 min under a pressure of 50 psi and, subsequently, with the Milli-Q water for 1 min under a pressure of 50 psi. The background electrolyte (BGE) consisted of 5 mM of sodium tetraborate decahydrate,
150 mM of boric acid, 50 mM of SDS and 15% (v/v) of MeOH. The
apparent pH value of the BGE equalled 7.3.

2.6. SPE conditions
2.2. Apparatus

All separation studies were carried out using a capillary electrophoresis (CE) apparatus (P/ACE MDQ Capillary Electrophoresis
System, Beckman Coulter, Fullerton, CA, USA). The appliance was
equipped with an automatic sample dispenser and a DAD detector. Analysis of the obtained data was made using 32 Karat 8.0
software (Beckmann, Fullerton, CA, USA). The device was additionally equipped with a capillary thermostat system by means of a
coolant, which allowed the temperature to remain constant during
the analysis.
2.3. Preparation of stock and working solutions
Stock solutions were prepared by accurately weighing 1.0 mg
of each analyte on an electronic scale (Ohaus, PA, USA). Then
the weighed analytes, namely: DHPG, VMA, MHPG, HVA, M, NM,
DOPAC were dissolved separately in 1 mL of MeOH. Subsequently,
they were shaken on an MS 3 Basic, IKA® shaker (USA). Standard
human urine samples were enriched with each of the analytes
to a final concentration of 10 μg/mL and then put aside for one
of the extraction methods. After the isolation procedure, the dry
residue containing the extracted analytes was dissolved with 50
μL of 2 mM sodium tetraborate decahydrate and separated by the

SPE (Agilent Vac Elut SPS 24 Manifold, Santa Clara, United
States) was carried out on hydrophilic-lipophilic balanced (HLB)
(SupelTM -select HLB, Sigma Aldrich, Germany), octadecyl sorbent (C18) (Discovery® DSC-18, Sigma Darmstadt, Germany) and
´
cyanopropyl (CN) (Chromabond®-CN, VWR, Gdansk,
Poland) cartridges, previously activated with 1 mL of MeOH and 1 mL of MilliQ water. The 1 mL of human urine was spiked with the working
solution of analytes at a concentration of 10 μg/mL. Next, the samples were applied to the SPE columns which were next washed
with 1 mL of Milli-Q water and dried in a vacuum for 5 min.
The analytes were desorbed from the SPE cartridges to clean glass
tubes with 1 mL of one of the tested eluents: MeOH, DCM, hexane, acetone and ACN:MeOH (1:1, v/v). The solvent was evaporated
to dryness at 45 °C. The dry residue was dissolved in 150 μL of
2 mM sodium tetraborate and analyzed by the elaborated MEKC

method.
2.7. SPME conditions
1 mL of the human urine sample was spiked with the working
solution of analytes at a concentration of 10 μg/mL. In the meantime, 96-well SPME brushes with polystyrene-divinylbenzene (PSDVB) resin, which was used as complementary to SPE HLB-type


4

N. Miekus,
˛
A. Plenis and M. Rudnicka et al. / Journal of Chromatography A 1620 (2020) 461032

resin or C18 resin, were conditioned with 1 mL of MeOH/H2 O (1:1;
v/v) for 30 min and washed with 1 mL of deionized water for
10 s. Then, 1 mL of each sample was applied to the wells of a 96well plate and the BAs were extracted for 60 min (shaking speed
850 rpm). Then the brush fibers were washed again with deionized
water for 10 s to remove impurities and a 60 min desorption step
with one of the four tested desorbents: MeOH, acetone, ACN:MeOH
(1:1; v/v) or DCM was carried out. Afterwards, the samples were
evaporated to dryness on a centrivap (45 °C, 1.5 h) and the residue
was dissolved in 150 μL of 2 mM sodium tetraborate and analyzed
by the elaborated MEKC method.
2.8. Data analysis
In order to calculate the EF of each tested method, the MEKC
separation of the sample undergoing the sample extraction procedure was carried out, as well as the control sample containing all
the analytes at a concentration of 10 μg/mL in 2 mM borax, which
was not undergoing the sample preparation procedure. The value
of the EF was calculated according to Eq. (1):

EF = H/H0

(1)

where: H – the peak height of the analyte determined by the MEKC
method in the human urine sample undergoing the extraction procedure, H0 – the peak height of the analyte determined by MEKC
in the control sample without sample pretreatment.
Due to the fact that the sample undergoing the extraction procedure was evaporated and next, the residue was dissolved in 150
μL of 2 mM sodium tetraborate, the effect was the concentration
of the analyte in the sample. Therefore, the calculated height of
the peak could be higher than that of the control sample without
sample pretreatment. In consequence, the value of the EF calculated according to Eq. (1) could be above 1.
The comparative study of the EF results of seven BAs from urine
samples obtained under the 25 tested sample preparation procedures was conducted under HCA using the Euclidean distance
method and the single linkage method. Statistica 13.3 software
(StatSoft, Tulsa, USA) was used to achieve the task. The numbering
of the extraction procedures, as presented in Table 1, was retained
unchanged in the chemometric calculation.
3. Results and discussion
Analyte pre-concentration procedures are essential, particularly
when less sensitive detection methods (spectrophotometric, e.g.
DAD) are employed. The most common purification techniques that
allow a differentiated degree of isolation of analytes are liquidliquid extraction (LLE), DLLME, SPE, SPME and their variants and
combinations. Each of them provides a different degree of analyte
concentration (initial off-line concentration) and each is differently
useful for the isolation of specific analytes from the matrix [12]. In
the case of BAs, this is particularly important because their concentrations in biological matrices are extremely low (ng, pg or less),
and BAs have a hydrophilic nature and are characterized by photoand thermo-lability [13].
A detailed description of the aforementioned sample preparation protocols has recently been described in our previous papers
[7,10,11]. For this reason, in the presented study, the advantages
and disadvantages of each of the isolation methods were omitted, while the focus was on the isolation of seven BA metabolites with three different analytical approaches based on DLLME,

SPE or SPME (fully described in Sections 2.5–2.7). To the best of
our knowledge, there have been no other studies to date for the
simultaneous determination of such a large group of compounds
from the l-tyrosine metabolic pathway. The extraction efficiency

for each of the method was confirmed trough the analysis of electropherograms obtained by the MEKC-DAD method supported by
the HCA chemometric analysis. The signal recorded on the electropherogram from each isolated analyte was compared with the signal from the reference sample with the same concentration of the
test compound. At the beginning, the sample buffer and the BGE
composition were optimized to ensure the best separation conditions for multiple biomolecules.

3.1. The BGE and sample buffer for the simultaneous separation of
analytes
Because of the diversity of pKa values (data in Fig. 1) and the
amphoteric nature of the selected panel of BAs, their simultaneous
determination by conventional capillary zone electrophoresis (CZE)
is relatively difficult. Moreover, their chemical structures contain a
few functional groups which could be ionized in a wide pH range,
hence finding the optimal ingredients of the BGE is a real challenge. In our study, a borate buffer was selected because catechol
compounds and some substituted catechols (like BAs) can become
charged in a weakly alkaline electrolyte. In effect, the analytes contain vicinal hydroxyl groups which after becoming charged are able
to form complexes with borate ions [14,15], which allows adequate
electrophoretic mobility to be obtained. However, due to the lack
of a satisfactory separation of all analytes, the addition of surfactants and an organic modifier was examined. For this purpose, the
influence of an anionic surfactant such as SDS in the range of 0
– 50 mM was tested. A borate buffer without SDS and one at a
concentration below 50 mM SDS did not allow satisfactory separation; however, a concentration at 50 mM SDS gave full separation
of seven compounds of interest (Fig. 2). Moreover, ACN and MeOH
(in different volume proportions in the range of 5 – 20%, v/v) were
tested as organic components of the running buffer in order to increase sensitivity and improve the resolution for BAs. The experimental results indicated that the effective separation of the peaks
of interest could be observed when 15% of MeOH (v/v) was added

to the BGE. Therefore, a mixture of sodium tetraborate (5 mM),
boric acid (150 mM), SDS (50 mM) and MeOH (15%, v/v) (apparent
pH 7.3) was selected to separate all analytes at the highest resolution without any interferences (Fig. 3). In these MEKC conditions,
DHPG, VMA, MHPG and HVA were cations and therefore did not
interact with the hydrophobic interior of SDS micelles, and their
migration times (MTs) were shorter than NM, M and DOPAC, which
were anions.
To obtain an increase in the analyte signal, it was also necessary
for the injection parameters and the sample buffer composition to
be optimized. The ionic strength of the sample plays a significant
role in CE and can positively or negatively affect the separation, the
detectability of analytes and the migration time. The viscosity and
pH of the sample buffer as well as the potential amount of organic
component in the sample are of great importance. The appropriate
selection of the sample buffer allows for significant narrowing of
the analyte band, which promotes the simultaneous separation of
many components of the analyzed mixture.
In our research, we simultaneously developed two online preconcentration techniques: the first was stacking – accomplished by
placing the sample in a solution the ionic strength of which is significantly less than that of the separation buffer, and the second
– sweeping, based on the interaction between analytes in the matrix free of the SDS and the surfactant molecule-formed pseudostationary phase in the BGE. In this study, for the selection of the
sample buffer (injection medium), different borax concentrations
(in the range of 2 – 10 mM) were tested. Our research showed the
best sharpness and symmetry of peaks for a sample containing a
2 mM borax solution. Ultimately, experimental data revealed that


Table 1
Mean EF-values for seven BAs extracted from urine samples with twenty five different sample preparation procedures (n = 3).
SPE_2

SPE_3

SPE_4

SPE_5

SPE_6

SPE_7

SPE_8

SPE_9

SPE_10

SPE_11

SPE_12

—

HLB

HLB

HLB

HLB

C18

C18

C18

C18

CN

CN

DCM

MeOH

Acetone

Hexane

Acetone

MeOH

DCM

Hexane

C18
MeOH/ACN (1:1,

v/v)

MeOH

Acetone

0.06 ± 0.005
0.1 ± 0.002
0.03 ± 0.002
0.4 ± 0.003
0.0 ± 0.0
0.0 ± 0.0
0.0 ± 0.0

0.0
0.0
0.0
0.0
0.0
0.0
0.0

0.02 ± 0.001
0.03 ± 0.002
0.02 ± 0.002
0.0 ± 0.0
0.1 ± 0.01
0.03 ± 0.002
0.01 ± 0.002

0.2 ± 0.01
0.2 ± 0.02
0.1 ± 0.01
0.0 ± 0.0
0.07 ± 0.005
0.2 ± 0.01
0.0 ± 0.0

3.4
2.5
0.6
7.4
0.3
1.7
4.1

Eluting/desorbing agent —
Analytes

Mean EF-values (n = 3)

DHPG
VMA
MHPG
HVA
NM
M
DOPAC

0.4

0.2
0.1
0.8
0.2
0.6
0.2

Type of solid phase in
SPE/SPME

±
±
±
±
±
±
±

0.03
0.02
0.01
0.07
0.02
0.05
0.01

SPE_13
CN

Eluting/desorbing agent Hexane

0.1 ± 0.01
0.2 ± 0.02
0.05 ± 0.004
0.3 ± 0.02
0.1 ± 0.01
0.02 ± 0.002
0.04 ± 0.003

1.2
1.8
0.2
3.1
0.4
1.7
0.5

SPE_14
CN

SPME_15
C18

SPME_16
C18

SPME_17
C18

SPME_18

C18

SPME_19
PS-DVB

SPME_20
PS-DVB

SPME_21
PS-DVB

SPME_22
PS-DVB

SPME_23
PS-DVB

DCM

DCM

Acetone

MeOH

MeOH/ACN
(1:1, v/v)

MeOH/ACN
(1:1, v/v)

Acetone

MeOH

MeOH with IL1

MeOH with MeOH with
IL2
IL3

0.0 ± 0.0
0.02 ± 0.002
0.0 ± 0.0
0.3 ± 0.002
0.0 ± 0.0
0.06 ± 0.004
0.02 ± 0.002

0.1 ± 0.01
0.2 ± 0.02
0.08 ± 0.01
0.8 ± 0.05
0.0 ± 0.0
0.1 ± 0.01
0.4 ± 0.03

0.2 ± 0.01
0.2 ± 0.01
0.1 ± 0.01

0.8 ± 0.05
0.4 ± 0.03
0.7 ± 0.02
0.04 ± 0.003

0.2 ± 0.01
0.2 ± 0.01
0.06 ± 0.005
0.8 ± 0.04
0.6 ± 0.04
2.8 ± 0.2
0.3 ± 0.02

0.0 ± 0.0
0.02 ± 0.001
0.0 ± 0.0
0.03 ± 0.002
0.3 ± 0.02
0.05 ± 0.004
0.5 ± 0.02

0.7
1.4
0.1
2.0
0.4
1.2
1.2

Analytes

Mean EF-values (n = 3)

DHPG
VMA
MHPG
HVA
NM
M
DOPAC

0.1 ± 0.03
0.1 ± 0.02
0.1 ± 0.01
0.06 ± 0.005
0.0 ± 0.0
0.0 ± 0.0
0.0 ± 0.0

0.03 ± 0.002
0.02 ± 0.001
0.01 ± 0.002
0.01 ± 0.001
0.0 ± 0.0
0.0 ± 0.0
0.0 ± 0.0

±
±
±

±
±
±
±

0.1
0.2
0.01
0.4
0.03
0.2
0.03

±
±
±
±
±
±
±

0.0
0.0
0.0
0.0
0.0
0.0
0.0

1.3

1.7
0.5
4.0
2.8
1.1
2.5

±
±
±
±
±
±
±

0.1
0.2
0.03
0.3
0.2
0.1
0.2

3.5
4.9
1.8
7.8
4.5
0.6
5.6

±
±
±
±
±
±
±

0.3
0.3
0.2
0.6
0.4
0.05
0.4

±
±
±
±
±
±
±

0.05
0.2
0.01
0.2
0.03

0.2
0.1

1.9
1.8
1.9
3.9
4.1
5.2
1.9

±
±
±
±
±
±
±

0.2
0.1
0.1
0.3
0.3
0.4
0.1

0.8
0.7
1.0

1.8
4.1
5.6
1.1

±
±
±
±
±
±
±

±
±
±
±
±
±
±

0.03
0.2
0.005
0.6
0.02
0.2
0.3

0.06

0.05
0.1
0.2
0.3
0.4
0.1

1.5
0.6
0.6
0.6
0.0
0.0
0.8

0.6
0.7
1.0
1.8
3.3
5.7
1.0

±
±
±
±
±
±
±

±
±
±
±
±
±
±

1.6 ± 0.1
0.9 ± 0.05
0.6 ± 0.04
0.8 ± 0.03
0.5 ± 0.02
0.1 ± 0.01
0.6 ± 0.04

0.2
0.04
0.05
0.03
0.0
0.0
0.06

0.04
0.05
0.1
0.1
0.3

0.4
0.1

SPME_24
PS-DVB

1.8
0.9
1.1
2.4
4.2
6.2
1.1

±
±
±
±
±
±
±

0.1
0.1
0.1
0.2
0.3
0.4
0.1

SPME_25
PS-DVB
DCM

0.0
0.0
0.0
0.0
0.0
0.0
0.0

±
±
±
±
±
±
±

0.0
0.0
0.0
0.0
0.0
0.0
0.0

Legend: EF – enrichment factor; BAs – biogenic amines; DLLME – dispersive liquid-liquid microextraction; SPE – solid phase extraction; SPME – solid phase microextraction; HLB – hydrophilic-lipophilic balanced sorbent;
DVB – divinylbenzene resin; C18 – octadecyl sorbent; CN – cyanopropyl sorbent; MeOH – methanol; ACN – acetonitrile; DCM – dichloromethane; IL1 – 1–butyl–3-methylimidazolium tetrafluoroborate; IL2 – 1-ethyl-3-methylimidazolium tetrafluoroborate; IL3 – 1-ethyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide; DHPG – dihydroxyphenylglycol; VMA – vanillylmandelic acid; MHPG – methoxy–4-hydroxyphenyl glycol; HVA – homovanillic

acid; NM – normetanephrine; M – metanephrine; DOPAC – dihydroxyphenylacetic acid.

N. Miekus,
˛
A. Plenis and M. Rudnicka et al. / Journal of Chromatography A 1620 (2020) 461032

Type of solid phase in
SPE/SPME

DLLME_1

5


6

N. Miekus,
˛
A. Plenis and M. Rudnicka et al. / Journal of Chromatography A 1620 (2020) 461032

Fig. 2. Effect of SDS concentration in the electrolyte on the MEKC separation of BAs metabolites. Separation parameters: applied voltage 22 kV, an effective/total capillary
length 50/60.2 cm and 75 μm i.d., λ = 200 nm, hydrodynamic injection 15 s at 0.5 psi, temp. 25 (± 0.1) °C; BGE: 5 mM sodium tetraborate decahydrate, 150 mM boric acid,
50 mM SDS and 15% (v/v) MeOH, pH = 7.3.
Legend: 1 – DHPG, 2 – VMA, 3 – MHPG, 4 – HVA, 5 – NM, 6 – M, 7 – DOPAC.

Fig. 3. Electropherogram obtained for the standard sample of seven analytes (each at the concentration of 10 μg/mL) dissolved in 2 mM sodium tetraborate (water solution)
under the optimized MEKC conditions. Separation parameters and legend as in Fig. 2.
Legend: 1 – DHPG, 2 – VMA, 3 – MHPG, 4 – HVA, 5 – NM, 6 – M, 7 – DOPAC.

the 2 mM sodium tetraborate (pH 8.7) solution was optimal as a

sample buffer for the studied compounds.
The developed MEKC-DAD method with the optimized sample
buffer, BGE, injection time, pressure, current and capillary length
allowed limits of detection (LODs) to be obtained of less than
0.1 μg / ml for all BAs.

3.2. Verification of the isolation and preconcentration methods
The DLLME approach for DA, A, NA, l-Tryp, l-Tyr, 5-HT, l-DOPA
[10,16] as well as HVA and VMA [11] was previously evaluated by
our team [10,11,16]. Here, it was applied for a new group of analytes. In the case of SPE, knowing the physicochemical properties
of DHPG, VMA, MHPG, HVA, M, NM and DOPAC, SPE with HLB,


N. Miekus,
˛
A. Plenis and M. Rudnicka et al. / Journal of Chromatography A 1620 (2020) 461032

CN or C18 cartridges was applied for the isolation of these analytes from the standard urine samples. In turn, the SPME-based
methods were carried out with the application of similar solvents
for desorption as those used in the SPE-based methods described
here. Also, based on our previous research – ionic liquids (IL)
were tested as new SPME desorbent additives which could improve
the isolation and preconcentration of analytes, e.g. DA, A, NA, lTryp and l-Tyr [7]. The elaborated DLLME, SPE and SPME extraction methods were compared by the calculation of their EF-values
on the basis of the peak heights of the analytes obtained by the
MEKC-DAD analysis. The obtained results, summarized in Table 1,
indicate that the final extractions of BAs from urine depend on the
physicochemical nature of the analytes, the type of sorbents used,
and the eluting/desorbing agents. The extraction results indicate
that in the case of SPE procedures based on MeOH as the eluting agent, the most effective solid phase to retain the analytes was
the non-polar sorbent (C18), followed by HLB, while CN offered the

worst retention. However, the efficiency of the tested organic solvents for the simultaneous elution of the analytes was varied. Thus,
the most effective simultaneous elution of the analytes, except for
M, was obtained using SPE-C18 with MeOH as the eluting agent.
For HLB cartridges, the use of MeOH gave also the best elution results, although these values were significantly lower than for C18.
In the case of CN, acetone was the most effective eluting agent. In
the case of SPME procedures, the properties of PS-DVB sorbent in
combination with MeOH as the desorbent allowed very good extraction efficiency to be obtained with EF-values of 5.2 for M, 4.1
for NM, 3.9 for HVA, 1.9 for DHPG, MHPG and DOPAC, and 1.8 for
VMA. On the other hand, the EF value for M increased when ILs
were added to MeOH, and obtained the highest value for MeOH
with IL3 (6.2). For SPME procedures, the possibilities for the effective desorption of the analytes by the tested organic solvents were
also different, and they were dependent on the type of SPME fiber
used. However, among the tested organic solvents, DCM gave the
worst results.
Next, these data were evaluated by the HCA multivariate analysis in order to find the most effective sample preparation procedure for the simultaneous isolation of the analytes. HCA offers
graphic data visualization of the relationships between the variables and/or objects without losing any significant information.
This statistical approach is based on an algorithm that groups similar objects into groups called clusters. The endpoint is a set of clusters, where each cluster is distinct from any other cluster, and the
objects within each cluster are broadly similar to each other. The
main output of HCA is a dendrogram which shows the hierarchical
relationship between the clusters [17]. In this study, the Euclidean
distance was used for measuring the dissimilarity between each
pair of observations, while average linkage clustering was applied
to determine which clusters should be joined at each stage. Dendrograms calculated on the basis of the established EF-values are
illustrated in Fig. 4A (variables) and 4B (objects), respectively.
3.2.1. Relationships between the tested sample preparation protocols
established by HCA
According to the HCA results for the variables (Table 1, Fig. 4A),
the tested extraction procedures were located in clusters I, II and
III. Taking into account the distance between the variables observed on the dendrogram, the biggest differences can be noticed
for the SPE-C18 procedures included in cluster I. Among them,

SPE_7 with MeOH as the eluting agent was found as an outlier.
This procedure offered the highest EFs for DHPG (3.5), VMA (4.9),
NM (4.5), DOPAC (5.6) and HVA (7.8). Slightly lower values were
measured for MHPG in comparison to those calculated after using
SPME_21 (1.8 vs 1.9). Only one analyte – M – was poorly extracted
from urine samples (EF = 0.6). The application of the mixture of
MeOH/ACN (1:1, v/v) as the eluting agent (SPE_10) provided a more

7

effective extraction of M (EF = 1.7) and a comparable extraction of
DHPG (EF = 3.4) and HVA (EF = 7.4). On the other hand, a less effective isolation of DOPAC (EF = 4.1) and VMA (EF = 2.5) was obtained, while the lowest isolation was found for MHPG (EF = 0.6)
and NM (EF = 0.3). The EF-values of the analytes were more comparable after using SPE_6 with acetone, although this solute modification caused a further decrease in efficiency for DHPG, VMA,
MHPG, HVA, M and DOPAC. However, in the case of NM, this effect
was contrary to using SPE_10. The EF-value for NM after SPE_10
was 0.3, while after the procedure with SPE_6, it was 2.8. Summarizing, among the procedures positioned in cluster I, the SPE_7
protocol located on the left of the dendrogram offered the best EF
results for six of the tested analytes. Only M was poorly extracted
using this protocol.
Taking into account the sample preparation protocols located in
cluster II, it can be noticed that each of them was based on SPME
with PS-DVB coatings. The methods using MeOH with the addition of three different ILs (IL1, IL2 and IL3) as desorbing solvents
(SPME_22–24, respectively) were located together, while the protocol with pure MeOH (SPME_21) was found as an outlier of cluster
II. In fact, SPME_22 and 23 gave comparable EF results for all tested
analytes, but they were lower than after using SPE_7. The only EF
parameter achieved for M was almost 10 times higher (5.2). On
the other hand, this value was slightly lower than that calculated
for SPME_24 with IL3 (EF = 6.2). This protocol was also more effective than SPME_22 and 23 for other tested analytes. Thus, the
position of SPME_24 at a small distance to the above-mentioned
procedures is fully justified. As it was mentioned above, SPME_21

was positioned on the right of these procedures, which offered the
more effective isolation of DHPG, VMA, MHPG, HVA and DOPAC.
This confirms that the modification of the desorbing solvent by the
addition of ILs, especially IL1 and IL2, should be avoided for the analytes containing acid groups (VMA, HVA and DOPAC) or more than
two hydroxide groups in the side chain of the molecule (DHPG,
MHPG).
It can also be noticed that most of the tested sample preparation procedures with more complicated structures were located in
cluster III. Therefore, SPME_20 based on PS-DVB and SPE_3 based
on HLB coatings were positioned in subcluster IIIA located closely
to cluster II. For them, acetone (SPME_20) and pure MeOH (SPE_3)
were used for the desorption/elution of the compounds of interest.
The EF parameters were higher for the analytes, except for DOPAC,
after SPE_3. SPME_C18 with the mixture of MeOH:ACN as the desorbing solvent was located on the right of subcluster IIIA as an
outlier. This protocol offered a significantly less effective extraction
of the analytes than that calculated for the protocols included in
clusters I and II, except for M. For this analyte, EF parameters almost 5 times higher were calculated with respect to SPE_7. Two
SPE procedures based on CN coatings were placed in subcluster
III B1. These protocols gave low EF results for all analytes except
DHPG.
The SPME_19 protocol based on the PS-DVB cartridge and the
mixture of MeOH/ACN as a desorbing agent was positioned on the
right as an outlier of subcluster III B.2. This method offered very
low efficiency of the analytes (EFs from 0.0 to 0.5). Five protocols with various solid phases were located in subcluster III B 2.1,
whereas SPE_5 using hexane was positioned on the right of this
group as an outlier. These procedures were also described by very
low EF parameters in comparison to other tested protocols, especially SPME_25 and SPE_5, which were not able to isolate any compound of interest. Slightly more effective protocols with respect
to those located in subcluster III B.1 were located in subcluster III
B2.2, where DCM in combination with C18 (SPME_15) and the HLB
cartridge (SPE_2) as well as HLB and acetone as the eluting agent
(SPE_4) were applied. Unfortunately, these methods also offered

poor extraction of the analytes. On the left of the dendrogram, sub-


8

N. Miekus,
˛
A. Plenis and M. Rudnicka et al. / Journal of Chromatography A 1620 (2020) 461032

Fig. 4. HCA results obtained for the variables (A) and the objects (B) on the basis of EF-values obtained for the analytes after DLLME, SPME and SPE protocols following the
MEKC method.

cluster III B3 was distinguished, which contained two procedures
based on C18 (SPME_16 and 17) and the DLLME_1 method. Their
positions with respect to cluster III B2.2 suggest that slightly different EF results were calculated. In fact, these protocols offered
more effective isolation of all tested analytes, especially for HVA.
Unfortunately, each of them offered significantly worse results than
those determined for other protocols included in clusters I and II.
3.2.2. Relationships between the tested analytes established by HCA
Taking into account the HCA results for the objects, it can be
observed that seven tested analytes were located in clusters I and

II (Fig. 4B). Their positions were clearly correlated with their chemical structures, which defined the different physicochemical characteristics of these molecules. Therefore, M and NM which possess
the same chemical structure of the main molecule, but with a different type of amino group in the side chain (-NHCH3 and -NH2 ,
respectively, Fig. 1) were included in cluster I. On the other hand,
the relatively high distance between them indicates that relatively
different EF results were calculated from them. This confirms that
the type of amino group can decide about the final interaction between the analyte and the molecules of the solvents used as extraction/desorption agents or mobile-/solid-phase components de-

N. Miekus,
˛
A. Plenis and M. Rudnicka et al. / Journal of Chromatography A 1620 (2020) 461032

pending on specific experimental conditions. The analytes included
in cluster II possess carboxyl groups (HVA, DOPAC, VMA) or two
hydroxide groups in the side chain of the molecule (MHPG, DHPG)
(Fig. 1). Among them, HVA was located as an outlier of cluster II,
while MHPG was positioned at a small distance to DOPAC, VMA
and DHPG. This can be correlated with the fact that twenty extraction protocols were able to isolate HVA more effectively than
other analytes, especially SPE_3, 6, 7 and 10, and SPME_16 and
20. Moreover, pure MeOH and the mixture of MeOH/ACN were the
best eluting solvents for HVA, whereas the addition of ILs caused
a decrease of this parameter. In contrast to HVA, MHPG was significantly less effectively isolated from urine than other analytes,
e.g. after the application of DLLME_1, SPE_3 and 6, and SPME_18–
20. The most effective protocols for this compound were SPME_21
and SPE_7. DOPAC and VMA possessing carboxyl groups were comparably isolated using most tested protocols. SPE_7 offered the
best conditions for the extraction of these BAs. On the other hand,
SPE_10, especially for DOPAC, can be considered as an interesting
alternative.
Summarizing, the obtained HCA results indicated that SPE-C18
with MeOH as the eluting agent offered the most effective isolation of DHPG, VMA, HVA, NM and DOPAC, whereas SPME-PS-DVB
with the same solvent for the desorption was the best choice for M
and MHPG. This SPME approach also guaranteed a relatively high
extraction for other BAs. Moreover, the addition of IL3 (1-ethyl-3methylimidazolium bis(trifluoromethylsulfonyl)imide) to MeOH in
SPME increased the efficiency for M, did not change the extraction parameters for DHPG and NM, but should be avoided for VMA,
MHPG, HVA and DOPAC. It was also confirmed that the success or
failure of the tested extraction procedures was dependent on the
specific chemical structures of the BAs.
4. Conclusions

The present research is the first example of the comparison of
three different extraction approaches based on DLLME, SPME and
SPE for the isolation of compounds from the l-tyrosine metabolic
pathway from human urine samples. In this study, one DLLME,
thirteen SPE and eleven different SPME protocols were performed,
then the separation of the analytes based on the developed MEKC
method was performed, and next, the EF values for each analyte in
specific extraction conditions were calculated. Finally, the EF values
were compared by HCA. The obtained results confirmed that the
use of HCA increases the probability of the selection of the most
appropriate sample preparation procedure for the specific analysis, including the simultaneous or specific determination of the selected BAs. Similarly, the HCA results showed that SPE-C18 with
MeOH as the eluting agent and SPME-PS-DVB with the same solvent for the desorption should be considered as alternative tools
for the extraction of the seven tested BAs. The addition of 1-ethyl3-methylimidazolium bis(trifluoromethylsulfonyl)imide to MeOH in
SPME offered a more effective extraction of M but can decrease efficiency for VMA, MHPG, HVA and DOPAC. Thus, the application of
multivariate data processing, i.e. HCA can be considered as a valuable starting point for improving the reliable evaluation of sample preparation protocols in pharmaceutical practice. Moreover, it
was confirmed that the developed MEKC method, supported by
SPE or SPME, can be used as an off-line preconcentration technique
for the simultaneous isolation and determination of seven catechol
compounds in urine samples for diagnostic purposes.
Author’s contribution
N.M. coordinated the manuscript writing and submission, N.M.,
I.O. planned all the experiments, wrote the experimental part of

9

the manuscript and introduction section, N.K., M.R., N.M. performed the experiments and collected the raw data, I.O., P.K. optimize the BGE for MEKC separation, wrote parts focused on the BGE
optimization in “Results and Discussion” section, A.P. performed
the HCA chemometric analysis of raw experimental data and described them in the manuscript, T.B. obtained financial support for
the experiments and manuscript publication, T.B., I.O. supervised
the experimental procedures, All authors prepared and approved

all the files related to Manuscript (figures, responses to Reviewers’
comments, tables).
Acknowledgement
The authors acknowledge the support of the MTB Korea V4 joint
project from the following sources: National Center for Research
and Development in Poland (DZP/V4-Korea- I/20/2018).
References
[1] H. Lehnert, Pheochromocytoma : Pathophysiology and Clinical management,
Frontiers of Hormone Research, Karger, Basel, Switzerland, 2004.
[2] James Mike, Anaesthesia For Patients With Endocrine Disease, Oxford University Press, NA, USA, 2010.
[3] K. Pacak, Phaeochromocytoma: a catecholamine and oxidative stress disorder,
Endocr. Regul 45 (2011) 65–90.
[4] N. Miekus,
˛
T. Baczek,
˛
Non-invasive screening for neuroendocrine tumors - Biogenic amines as neoplasm biomarkers and the potential improvement of "gold
standards, J. Pharm. Biomed. Anal. 130 (2016) 194–201, doi:10.1016/j.jpba.2016.
06.013.
[5] Catecholamines, Bridging Basic Science With Clinical Medicine, 1st Edition,
Academic Press, NY, USA, 1998.
[6] G. Eisenhofer, M. Peitzsch, B.C. McWhinney, Impact of LC-MS/MS on the laboratory diagnosis of catecholamine-producing tumors, Trends Anal. Chem. 84
(2016) 106–116, doi:10.1016/j.trac.2016.01.027.
[7] N. Miekus,
˛
I. Oledzka,
˛
N. Kossakowska, A. Plenis, P. Kowalski, A. Prahl,
T. Baczek,
˛

Ionic liquids as signal amplifiers for the simultaneous extraction of
several neurotransmitters determined by micellar electrokinetic chromatography, Talanta 186 (2018) 119–123, doi:10.1016/j.talanta.2018.04.041.
[8] N. Kossakowska, I. Oledzka,
˛
A. Kowalik, N. Miekus,
˛
P. Kowalski, A. Plenis, E.
´ A. Kaczorowska, M.A. Krawczyk, E. Adamkiewicz-Drozy
˙ nska,
´
Bien,
T. Baczek,
˛
Application of SPME supported by ionic liquids for the determination of biogenic amines by MEKC in clinical practice, J. Pharm. Biomed. Anal.. 173 (2019)
24–30. 10.1016/j.jpba.2019.05.021.
[9] N.F. Atta, E.H. El-Ads, Y.M. Ahmed, A. Galal, Determination of some neurotransmitters at cyclodextrin/ionic liquid crystal/graphene composite electrode, Electrochim. Acta 199 (2016) 319–331, doi:10.1016/j.electacta.2016.02.078.
´ A. Miekus,
[10] N. Miekus,
˛
I. Oledzka,
˛
A. Plenis, P. Kowalski, E. Bien,
˛
M.A. Krawczyk,
˙ nska,
´
E. Adamkiewicz-Drozy
T. Baczek,
˛
Determination of urinary biogenic

amines’ biomarker profile in neuroblastoma and pheochromocytoma patients
by MEKC method with preceding dispersive liquid–liquid microextraction, J.
Chromatogr. B 1036–1037 (2016) 114–123, doi:10.1016/j.jchromb.2016.10.007.
´ A. Miekus, M. Krawczyk,
[11] N. Miekus, P. Kowalski, I. Oledzka, A. Plenis, E. Bien,
´
E. Adamkiewicz-Drozynska,
T. Baczek, Cyclodextrin-modified MEKC method for
quantification of selected acidic metabolites of catecholamines in the presence
of various biogenic amines. application to diagnosis of neuroblastoma, Journal
of Chromatography B: Analytical Technologies in the Biomedical and Life Sciences 1003 (2015) 27–34, doi:10.1016/j.jchromb.2015.09.003.
[12] N. Drouin, S. Rudaz, J. Schappler, Sample preparation for polar metabolites in
bioanalysis, Analyst 143 (2018) 16–20, doi:10.1039/c7an01333g.
[13] A. Plenis, I. Oledzka,
˛
P. Kowalski, N. Miekus,
˛
T. Baczek,
˛
Recent trends in the
quantification of biogenic amines in biofluids as biomarkers of various disorders: a review, J. Clin. Med. 8 (2019) 640, doi:10.3390/jcm8050640.
[14] J. Cao, B. Li, Y.-.X. Chang, P. Li, Direct on-line analysis of neutral analytes
by dual sweeping via complexation and organic solvent field enhancement
in nonionic MEKC, Electrophoresis 30 (2009) 1372–1379, doi:10.1002/elps.
20 080 0523.
[15] M.J. Markuszewski, P. Britz-McKibbin, S. Terabe, K. Matsuda, T. Nishioka, Determination of pyridine and adenine nucleotide metabolites in bacillus subtilis
cell extract by sweeping borate complexation capillary electrophoresis, Journal of Chromatography. A. 989 (2003) 293–301, doi:10.1016/s0021-9673(03)
0 0 031-1.
[16] N. Miekus,
˛

I. Oledzka,
˛
D. Harshkova, I. Liakh, A. Plenis, P. Kowalski, T. Baczek,
˛
Comparison of three extraction approaches for the isolation of neurotransmitters from rat brain samples, Int. J. Mol. Sci. 19 (2018) 1–11, doi:10.3390/
ijms19061560.
[17] S. Lin, S. Van Poucke, Z. Zhang, P. Lan, F. Murtagh, Hierarchical cluster analysis in clinical research with heterogeneous study population: highlighting its
visualization with R, Ann. Transl. Med. 5 (2017) 75–75, doi:10.21037/atm.2017.
02.05.