Capillary electrophoresis and liquid chromatography for determining steroids in concentrates of purified water from Päijänne Lake

Journal of Chromatography A 1649 (2021) 462233

Contents lists available at ScienceDirect

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

Capillary electrophoresis and liquid chromatography for determining
steroids in concentrates of purified water from Päijänne Lake
Heli Sirén∗, Tuomas Tavaststjerna, Marja-Liisa Riekkola
Department of Chemistry, University of Helsinki, P.O. Box 55, FI-00014, Helsinki 00560, Finland

a r t i c l e

i n f o

Article history:
Received 13 February 2021
Revised 11 April 2021
Accepted 30 April 2021
Available online 7 May 2021
Keywords:
Steroid hormones
Tap water
Partial filling micellar electrokinetic
chromatography
Microemulsion electrokinetic capillary
chromatography
Liquid chromatography-mass spectrometry

a b s t r a c t

The research was done with partial filling micellar electrokinetic chromatography, microemulsion electrokinetic chromatography, and ultra-high performance liquid chromatography. The study focuses on determination of male and female steroids from cold and hot tap water of households in Helsinki City. The
districts´ raw water is made run from Päijänne Lake through a water tunnel to the purification plants
in Helsinki area. The effluents delivered from the plants to households as tap water were sampled and
used for the study. They were concentrated with solid phase extraction to exceed the detection limits of the three methods. With partial filling method the limits were 0.50, 0.48, 0.33, and 0.50 mg/L
for androsterone, testosterone, progesterone, and testosterone-glucuronide, respectively. In microemulsion method the limit values were 1.33, 1.11, and 0.40 mg/L for androsterone, testosterone, and progesterone, respectively, and 0.83, 0.45, and 0.50 mg/L for hydrocortisone, 17-α -hydroxyprogesterone, and
17-α -methyltestosterone, respectively.
In the tap water samples, progesterone concentrations represented the highest values being 0.22 and 1.18
ng/L in cold and hot water, respectively. They also contained testosterone (in all samples), its glucuronide
metabolite (in 25% of the samples), and androstenedione (in 75% of the samples). The ultra-high liquid
chromatographic method with mass spectrometric detection was used for identification of the steroids at
μg/L level.
© 2021 The Authors. Published by Elsevier B.V.
This is an open access article under the CC BY license ( />
1. Introduction
Humans and animals generate numerous natural and synthetic
steroids that are the main hormone sources e.g., in surface waters
[1–3]. Concerning to steroid contamination, nearly 80% of effluent
water from wastewater treatment plants (WWTPs) contain female
hormones [4,5].
Generally, purification of drinking water is processed with efficient pretreatment methodologies, which are needed before water
can be delivered to humans’ households [6]. Mostly, drinking water is made of water from environment, i.e., from ground water
or surface water from lake and river systems, or from sea [7] that
is purified in water purifying plants. In the present research, the
raw water is made run from Päijänne Lake through the Päijänne
Water Tunnel to the purification plants in Helsinki. Raw water is
purified by precipitation of organic humus material with iron sulphate. Furthermore, removals of the faults in smell and taste are
done with filtration through sand beds, removal of microbes with

∗

Corresponding author.
E-mail address: heli.m.siren@helsinki.fi (H. Sirén).

ozone gasification, organic material with carbon membrane filtration, and disinfection with UV light. Chlorine is added to prevent
´
microbesgrowth
in the water networks. Finally, calcinated water
and carbon dioxide are added [8].
When water components have not significant health effects,
they are not designed to be removed totally from the effluents.
Therefore, several steroids are found at extremely low concentrations in water of various pre-treatment plants [9]. Recently, it was
also observed that biological processes removed steroids, but they
also activated steroids to be transformed back to their steroidal
precursor androstenedione. Thus, the knowledge about the existence of steroids in drinking water is strengthen, when raw water
is used from lake and river systems surrounded by city, agricultural, and industrial areas [10,11].
According to literature, at many plants androgen steroids are
removed from intake water using biological purification and membrane filtration [12–14]. However, despite that in effluent water androgen steroids were detected [15]. Recently, the female hormones
17β -estradiol (E2) and 17α -ethinylestradiol (EE2) were also identified in tap and drinking water by GC-MS [16]. Another paper
showed that millimetric-size polymer ultrafilters (UF-PBSAC) used
as packed-layer membranes, removed estradiol at higher than 99%

/>0021-9673/© 2021 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license ( />

H. Sirén, T. Tavaststjerna and M.-L. Riekkola

Journal of Chromatography A 1649 (2021) 462233

recoveries. Thus, their performance fulfilled the European Union
proposed minimum detection limit of 1 ng/L in drinking water
[17] by assuring the filter-packages´ quality for estradiol removal

from water [18]. Anyhow, there are very few research papers on
determination of male steroids in local tap water supply systems of
the cities in the world [19–21]. The reasons may be the published
risk assessments that the low concentrations of steroids were very
unlikely to pose risks to human health [22]. Despite that, highquality water processing directives and regulations would be helpful to evaluate transparently the water quality in general [23].
In many countries, network systems for water services are out
of the coverage with regulations [24,25]. However, in Finland the
authorities have published results of the water quality in Helsinki
area using the purified effluent water from Päijänne Lake being
the source of drinking water. The results of the laboratory data
of the year 2020 dealt the public information about exceptionally
low amounts of total organic contaminant (TOC) [26]. However,
the concentrations of individual inorganic ions were more detailed
measured. They were individually studied to convince the suitability of the purified water for human use.
As to steroid hormones, they are slightly water-soluble but not
volatile, which enable the preservation of their features in surface
and ground waters [15, 27-29]. Since those water types are used in
preparation of drinking water, the processes may lead to meaningful concentrations of steroids in tap water. There are few recently
published papers about the studies on androgens, estrogens, and
progestogens in environmental water and in surface water, which
were used for processing of drinking water [30–32]. Then, it was
assumed that the purification processes omitted the removal of
pharmaceuticals due to problems in water pre-treatment. Furthermore, some groundwater plants virtually have reported that the
processes do not have any other treatment steps in the process
than aeration when producing drinking water [33–35]. Anyhow,
the knowledge about the individual steroid hormones in drinking
water is important, since they have hormonal effects, such as generation of femininity among animals (estrogens), occurrence of deformity, and increase of birth defects [10,22,31,35,36]. Furthermore,
clean water is the base solvent in manufacturing of commercially
products.
Mainly, steroid hormones in drinking water have been determined with liquid chromatography–tandem mass spectrometry

(LC-MS/MS) and gas chromatography-mass spectrometry (GC/MS)
[37–39]. Despite the strength of chromatographic techniques, capillary electrophoresis (CE) with UV detection has shown its usability in separation of parent steroids from their metabolites due to
the possible utilization of CE with different kinds of methodologies and by modifying the composition of electrolyte solutions. In
steroid analyses, a partial filling micellar electrokinetic capillary
chromatography (PF-MEKC) was used to enhance separation and
UV sensitivity of neutral steroid compounds in water samples from
WWTPs [10,15]. Then, water samples needed matrix removal and
steroid concentration before analyses.
Solid phase extraction (SPE) is an excellent technique to purify high volumes of water samples and to concentrate nonpolar analytes [10]. Various novel configurations of SPE materials,
like magnetic solid–phase extraction [40] before GC-MS studies of
estrane, androstenedione, and progesterone hormones was used.
Lately, SPE was filled with magnetite nanoparticles which were
palmitate-coated to enrich steroids and increase capacity with 100fold. Then, limit of detection for steroids were 4 - 8 ng/L via SPE
when measured with GC/MS [41]. The concentrates needed manipulation since steroids were derivatives to improve their volatility.
The concentrations of androstenedione, testosterone, and progesterone have shown to be 10 0 0-fold lower in drinking water
than in surface and ground water [28,29,42–46]. According to literature, individual steroid concentrations exist at ng/L level in en-

vironmental samples [10,15,24,28,29,47]. The increase of toxicity of
water to organisms is due to progesterone, which is permanently
measured at ng/L to higher concentrations [48,49]. Progesterone
was determined in surface water, lake and river water, tap water, and in influent and effluent water of WWTPs in various countries [10,24,50–54]. Its concentrations varied from 0.031 ng/L to
128.3 ng/L, being the lowest in tap water and the highest in influent water to WWTPs (Supplementary data Table S1). Recently
also, interlaboratory research to compare determination of ECDs in
drinking water, surface water and treated wastewater in six countries was organized and the results were reported [55]. Solid phase
extraction (SPE) was used for purification and concentration of
the ECDs and finalized with situation fitting extraction techniques
[55,56, Supplementary data Table S2]. The steroid concentrates are
favorable made from small sample volumes with micro techniques
[57,58].
The present study was made to determine steroid hormones

in cold and hot tap water from randomly selected households in
Helsinki City. The analyses were done with partial filling micellar
electrokinetic chromatography technique (PF-MEKC). Off-line concentration with SPE enrichment and analyte focusing by on-line
stacking were used to concentrate the steroids from ng/L (raw water level) to μg/mL concentrations (method level). The concentration was mostly focused to the off-line enrichment and validation
of the capillary electrophoresis methods by the properties of the
electrolytes (ionic strength, pH) and the composition of the micelle
phases to obtain regionally suitable conductivity zones for focusing
purposes in the capillary.
Often the raw water contains humus [8,24,59,60]. Thus, some
markers of the mold Botrytis cinerea were detected with PFMEKC using UV detection and identified with UHPLC-HRMS/MS. All
steroids in the samples were also identified with UHPLC-HRMS/MS.
The observations about the mold are in correlation with those
made on syntheses, with behaviour of fungal pathogen B. cinerea,
and with creation of the pathways written in literature [61,62].
2. Materials and methods
2.1. Chemicals
2.1.1. PF-MEKC
Steroids selected for tap water studies were progesterone
(≥98%, Sigma-Aldrich Co., Germany), androstenedione (≥ 98%, 6305-8, Sigma-Aldrich Co., Germany), testosterone (≥ 98%, SigmaAldrich Co., France), and testosterone-glucuronide (TLC: 1, STERALOIDS INC., USA).
2.1.2. MEEKC
The chemicals (progesterone, androstenedione, testosterone,
and testosterone-glucuronide) mentioned above were used along
with pregnelone (≥98%, Sigma-Aldrich), hydrocortisone (≥98%,
HPLC, Sigma-Aldrich), 17-α -hydroxyprogesterone (≥95%, SigmaAldrich), fluoxymesterone (TLC: 1, STERALOIDS), 5-androsten3β -ol-17-one (DHEA, 1 mg/1 mL in methanol, certified solution, Sigma-Aldrich), 5α -androstan-17β -ol-3-one (TLC: 1, STERALOIDS), 11β -hydroxytestosterone (TLC: 1, STERALOIDS), and 17α methyltestosterone (TLC: 1, STERALOIDS). They were studied with
MEEKC as the reference for sensitivity when optimizing the CE
analysis.
2.1.3. LC-MS
The chemicals (progesterone, androstenedione, testosterone,
and testosterone-glucuronide) mentioned above were used along
with 5β -androstan-3α -ol-17-one glucosiduronate (TLC: 1, STERALOIDS), 1,3,5(10)-estratrien-3,17β -diol 3-glucosiduronate (TLC: 1,

STERALOIDS), 4-androsten-17β -ol-3-one glucosiduronate (TLC:
2


H. Sirén, T. Tavaststjerna and M.-L. Riekkola

Journal of Chromatography A 1649 (2021) 462233

1,
STERALOIDS),
aldosterone,
11β ,17β -dihydroxy-9α -fluoro17α -methyl-4-androsten-3-one
(9α -fluoro-11β -hydroxy-17α methyltestosterone, fluoxy-mesterone), (TLC: 1, STERALOIDS),
4-androsten-3,11,17-trione, β -estradiol (TLC: 1, STERALOIDS),
estrone (TLC: 1, STERALOIDS), 17α -methyltestosterone (≥98%,
Sigma-Aldrich), and 17α -hydroxyprogesterone (≥98%, SigmaAldrich). They were used for validation of the column material and
the steroid identification with liquid chromatographic methods.
Methanol was used in gradient elution, since it endured better
separation between steroids, matrix, and steroid conjugates.

infusion (10 μl/min) into the mass spectrometer with the micro syringe pump and scan range in MS-measurements of m/z 50–10 0 0
Da were used. Parameters in the negative ionization mode were
-3.2 kV, 40–48 V depending on the metabolite, and 70°C for capillary voltage, cone voltage, and source temperature, respectively.
Parameters in the positive ionization mode were 3.5 kV, 35–50 V
depending on the metabolite, and 70°C, respectively. The temperature of turbo ion source was set to 450°C. The optimised instrument settings were ion spray voltage 4500 V, nebulizer gas flow
10 L/min, curtain gas 12 L/min, collision gas 5 L/min, focusing potential 185 V, entrance potential 8 V, and cell exit potential 13 V.
The MS2 measurements with [M – H]− (negative ionization
mode, nESI) and [M + H]+ (positive ionization mode, pESI) were
done with 15, 25, and 35 eV collision energies. Standard mixtures at 10-26 ppm concentrations were used. They were prepared
in acetonitrile-water mixture (1:1, v/v) containing 0.1 % NH4 OH

(v/v), when measurements were done in nESI. The solution was
acetonitrile-water (1:1, v/v) containing 0.1 % HCOOH (v/v), when
identification was done with positive ionization mode. For the MS2
analysis of the compounds the most intense peak in the mass spectrum was chosen providing the mass spectra of the product ions.

2.1.4. Other chemicals
Ammonium acetate (AA) (98%, Sigma-Aldrich CO., Germany) for
CE electrolyte solution, ammonium hydroxide (25%, VWR International S.A.S, France) for pH adjustment, sodium dodecyl sulphate
(SDS) for micelle formation (99%, Oy FF-Chemicals Ab, Germany),
sodium taurocholate (STC) for SDS-STC micelles (BioXtra ≥95%
(TLC), Sigma-Aldrich Co., Germany), methanol (HPLC-MS grade,
Fisher Scientific, UK), sodium hydroxide (NaOH) for condition of
capillary and pH adjustment of the solutions [Sigma-Aldrich Co.,
Finland), and MQ-water (pH 5.8) for electrolytes, eluents, and standard solutions (Direct-Q UV Millipore, Millipore S.A., Molshheim,
France). Methanol was used as the solvent in standards and as the
marker of electroosmosis.
The steroids were used as received and stored in a dark and
cold room (+ 4°C and -20 °C). Botrydial (C17 H26 O5 , 310.38534
g/mol) standard was not available. Therefore, it was identified with
the accurate mass method with non-target UHPLC-HRMS.

2.2.2. UHPLC coupled with a mass spectrometer for identification
The accurate masses of the compounds were measured with
the Thermo Ultimate 30 0 0 UHPLC coupled with an Orbitrap Fusion TMS (Tribrid mass spectrometer) using a Kinetex C18 column
˚ Phenomenex, Denmark). A filter unit cou(100 × 2.1 mm, 100 A,
pled with a precolumn was used before the analytical column. Accurate identification was made with an orbitrap-electrospray mass
spectrometer. The mass spectrometer was operated in the positive
ion electrospray mode (pESI) with at m/z 86.0 0–470.0 0 Da and
at EICs of the specific ion fragments. The 10 μL volumes of the
preconcentrated water samples were injected to the eluent at the

flow rate of 0.6 mL/min (column temperature 40°C). The parameters used for the mass spectrometer were as follows: spray voltage + 30 0 0 V in pESI mode, sweep gas flow rate 0 respective arbitrary units (AU), sheath gas flow rate 40 AU, aux gas flow rate
12 AU, ion transfer tube temperature 350°C, vaporizer temperature
300°C, maximum injection time 100 ms, automated gain control
(AGC) target 250 0 0 0, and S-lens RF level 60%. The orbitrap resolution used in this work was 120 0 0 0.

2.2. Instruments and methods
2.2.1. LC Liquid chromatography and mass spectrometry methods in
method development
2.2.1.1. LC-UV and UHPLC-UV-ESI-MS. Liquid chromatographic analyses were performed with a Hewlett–Packard Series 1050 liquid
chromatograph –MSD (Palo Alto, USA). The separation was investigated on different columns (Gemini C18 column (30 × 3.00 mm,
˚
5μm) and Gemini-NX C18 column (250 × 4.60 mm, 5μm, 110A)
from Phenomenex (Denmark). The best separation was obtained
with Gemini-NX C18 column using methanol-water (60:40, v/v)
modified by 25% ammonia-ammonium hydroxide (0.1%, v-%) as the
eluent. The program was isocratic with 1 mL/min flow rate and 30
min analysis time and the detection was with UV at 247 nm. Sample volume was 2 μL. The MS spectra were detected at mass range
of 50-600 Da and analysed at positive mode.

2.2.3. Capillary electrophoresis
2.2.3.1. PF-MEKC. A Hewlett-Packard 3D capillary electrophoresis
instrument (Agilent, Waldbronn, Germany) equipped with a photodiode array detector (190-600 nm) was used with ChemStation
programmes (Agilent) for instrument running and data handling.
Analyses were done with PF-MEKC at +25 kV voltage, 25°C temperature for 20 min time. The method provided 17 μA current,
which was monitored to store information about the possible instability of the analysis a function of time. It also informs in long
sequences about bubble formation (detected as current decrease to
zero). Current was a good marker to detect, when the discontinuous solution system was stabilized to continuous one (detected as
stabile current).
Fused silica capillaries (i.d. 50 μm, o.d. 362 μm) of 80 cm (Ltot
80 cm, Leff 71.5 cm for PF-MEKC) and 60 cm (Ltot 60 cm, Leff

51.5 cm for MEEKC) were from Polymicro Technologies (Phoenix,
AZ, USA). They were conditioned by sequentially flushing with 0.1
M NaOH, MQ-water, and the electrolyte solution for 20 min each
at 13.634 psi (940 mbar). When the sample was changed to another, the capillary was washed with water and the electrolyte solution for 5 min and 2 min, respectively. The compounds were detected with UV-wavelengths of 206 nm (testosterone-glucuronide
conjugate) and 247 nm (androstenedione, testosterone, and progesterone).

2.2.1.2. UHPLC-ESI-MS. The liquid chromatographic analyses were
also performed with a Hewlett–Packard Series 1100 liquid chromatograph (Palo Alto, USA) coupled with an Esquire 30 0 0 plus ion
trap mass spectrometer (Bruker Daltonics, USA), using a GeminiNX C18 (30 × 3.00 mm, 5μm, Phenomenex, Denmark) column
and methanol-water (60:40, v/v) eluent containing 0.1% (v-%)
ammonia-ammonium hydroxide (25%). Electrospray ionization was
used with the following parameters: capillary voltage ±4500 V,
end plate offset ±500 V, nebulizer pressure 50 psi (nitrogen), 9
L/min of drying gas (nitrogen), and drying temperature 350°C,
mass range 50–500 amu. Androstenedione (m/z 287 Da), progesterone (m/z 315 Da), and testosterone (m/z 289 Da) were analysed
in positive mode with 1 μL injection volume.
2.2.1.3. Tandem-MS and direct infusion. A Micromass Quattro II
triple quadrupole mass spectrometer (electrospray ionization, positive and negative ESI-MS/MS) was used to find the experimental
conditions to optimize the liquid chromatographic method. Direct
3


H. Sirén, T. Tavaststjerna and M.-L. Riekkola

Journal of Chromatography A 1649 (2021) 462233

2.2.3.2. MEEKC. Microemulsion electrokinetic chromatography
(MEEKC) was used as the comparison method for PF-MEKC. The
instrumentation and the capillary dimensions were the same as
informed above for PF-MEKC. Samples were injected by pressure

of 50 mbar for 5.0 s. Separation was made at +23 kV for 30 min.
The compounds were detected with UV-wavelengths of 238 nm
and 247 nm.

2.3.1.2. MEEKC. The microemulsion solution was made by weighting of 82.7 g disodium tetraborate decahydrate solution (10 mM
stock solution made in water and pH adjusted to 9.2), 0.5 g hexanol, 15 g acetonitrile, 1.2 g 1-butanol, and 0.6 g SDS. The solution
order was important made by weighting the components. The final
solution weight was 100 g. It contained disodium tetraborate decahydrate, hexanol, acetonitrile, 1-butanol, and SDS at 82.7, 0.5, 15,
1.2, and 0.6 w-%. The MEEKC method was used for determination
of progesterone and androstenedione, but since it was not as selective as PF-MEKC, other steroids, corticosteroids, and phthalates
could also be measured.

2.2.4. Solid-phase extraction
In many cases, these steroids cannot be detected due to too
small representative sample volumes used in the analyses. Therefore, sample concentrates are needed in analyses. Solid phase extraction (SPE, STRATA-X C18 columns) was used for treatment of
2 L cold and hot tap water samples from the locations informed
in Ch. 2.4. Before use, the sorbents for each water sample were
washed with methanol (6 mL) and MQ-water (conductivity 18 M ,
6 mL). After sorbent conditioning and introducing the sample (2
columns for 2 L), the SPE materials were dried for 30 min under vacuum. The adsorbed steroids were eluted with methanol
(6 mL). The eluates from the two C18 columns were combined and
evaporated to dryness under gentle nitrogen stream at 40°C. The
precipitate was reconstituted in 1 mL of methanol. Methanol was
used since it was investigated to be a more quantitative eluent for
steroids than acetonitrile. The control sample made of MQ-water
had an enrichment factor of 80 0 0 (2.0 0 L to 250 μL). The extract
was divided into 250 μL aliquots for the PF-MEKC and MEEKC
analyses followed by addition of 20 μL of 0.1 M NaOH into the
sample. For LC-UV-MS analyses the studied sample volumes were
200 μL (in methanol-MQ-water mixture, 1:1, v/v). The enrichment

factor is 10,0 0 0. Analyses of each steroid standard and water concentrates were performed with five replicate injections and with
eight sequential analyses.

2.3.2. Eluents in liquid chromatography
The eluents used in LC-UV-MS/MS analyses contained MQwater and methanol. The isocratic systems (Gemini NX C18) were
made of methanol-water (60:40, v/v) and 0.1% (v-%) ammonia. The
analysis time was 30 min. The eluent flow rate was 1 mL/min.
In the gradient system (UHPLC-orbitrap-MS/MS) the eluents
were A) 0.1% formic acid in MQ-water and B) 0.1% formic acid in
methanol. Gradient elution was used from 95:5 (v/v, A/B) by increasing B to 100% in 15 min and thereafter returning to the initial
composition within 1 min, and lastly keeping the composition for
4 min to equilibrate the system. The analysis time was 16 min. The
eluent flow rate was 0.6 mL/min.
2.4. Samples, sampling, and sample storage
The household water samples were taken by the resident of the
premises in the district area of Helsinki city. Their locations were
selected by site from the old and new city areas, but also from a
city coast area and a hill. The water samples were sampled on August to November in 2017 according to ISO/TC 147/SC 6 protocol
[64]. Within the framework of the current study, randomly chosen
tap water samples from the areas mentioned below were analysed
for the content of selected steroids and botrydis fungus (B. cinerea).
Cold and hot tap water (both 1 × 20 0 0 mL, sampled in Etu-Töölö,
Kumpula (4 households and Helsinki University), Munkkiniemi, Latokartano, Viikki, Pikku-Huopalahti, Vesala, Etelä-Haaga (2 households), Kalasatama, Pasila, and Ullanlinna). They were water from
households in apartment houses that were connected to water networks with pipelines that were constructed within 15 years. The
exceptions were tap water from Etu-Töölö, Munkkiniemi, EteläHaaga, and Pasila, where the refurbishment has started or will
start. In addition, the exception is Kalasatama area that has renewed pipelines. The reference samples were blank water made
of MQ-water - methanol (95:5, v/v) spiked with a steroid mixture
to reach 4 μg/mL steroid concentrations for the analyses. In the
artificial water sample the concentration was 0.5 ng/L. They were
pre-treated as the real samples (Ch. 2.2.3).

2.3. Solutions
2.3.1. Electrolyte solutions
2.3.1.1. PF-MEKC. The final micelle mixture was prepared by
adding 10 0 0 μL of 20 mM ammonium acetate (AA, pH 9.68) in
MQ-water, 440 μL of 100 mM sodium dodecyl sulphate (SDS) in
20 mM AA solution (pH 9.68) followed by addition of 50 μL of
100 mM sodium taurocholate in MQ-water, in this specific order.
The concentrations of AA, SDS, and STC were 19.33 mM, 29.50 mM,
and 3.356 mM, respectively. The total volume of the micelle solution was 1490 μL.
In the beginning of each analysis, all solutions were sequentially
introduced into the capillary from the inlet to the outlet in the
following order: The electrolyte (ammonium acetate, AA), micelle
solution (mixture of SDS in AA and STC in MQ water), a sample
solution (a standard or a sample concentrate), and the electrolyte
(AA). Then, the micelle plug was placed after the electrolyte solution used for separation and before the sample solution. Before
the electric field (+25 kV) is switch on, the AA plug was 94.88 %,
followed by the micelle plug of 3.55% (55.7 nl), the sample plug
of 0.415% (6.46 nl), and the final electrolyte plug of 1.18 % (18.30
nl). The percentages are calculated from the total volume of the
capillary (1570 nL) [63]. Furthermore, before voltage was switched
on, the analysis was forced to wait for 0.5 min to stabilize instability due to the pressure differences ( p) between the ends of the
capillary.
Originally, the PF-MEKC method was optimised for separation of pregnelone, progesterone, androstenedione, testosterone,
hydrocortisone, 17-α -hydroxyprogesterone, fluoxymesterone, 5androsten-3beta-ol-17-one (DHEA), 5α -androstan-17β -ol-3-one,
11β -hydroxytestosterone, and 17α -methyl-testosterone, but in this
study only for testosterone, androstenedione, progesterone, and
testosterone-glucuronide.

2.5. Standard series in calibration

2.5.1. PF-MEKC and MEEKC
The stock solutions of steroids were 1100-2470 μg/mL in
methanol. They were diluted to 50 μg/mL solutions to prepare the
working solutions. Their final concentrations were 50 μg/mL for
optimizing the migration, verifying electroosmosis, and adjustment
of lamp energy (sensitivity). Concentration calibration was made
with solution mixtures of 0.5-6.0 μg/mL. Peak areas of each steroid
in the electropherograms were calculated as a function of concentrations.
In MEEKC the concentrations in calibration were 2.0, 4.0, 6.0,
8.0, and 10.0 μg/mL.
2.5.2. LC methods
The stock solutions were prepared to10 0 0 μg/mL and 200
μg/mL. They were diluted to 0.05, 0.1, 0.5, 1.0, 2.0, 3.0, 4.0, 5.0, 7.0,
4


H. Sirén, T. Tavaststjerna and M.-L. Riekkola

Journal of Chromatography A 1649 (2021) 462233

and 10.0 μg/mL with methanol to make the concentration calibration and to determinate the limits of detection (LOD).

separation efficiency by means of Gemini C18 and Gemini-NX C18
columns with isocratic elution (Supplementary, Fig. S1). The column phases behaved differently: The latter one was not as good as
the first one in respect of separation efficiency them all. However,
the system was not perfect because the steroid peaks were overly
broad with low sensitivity, although at higher concentrations the
better sorbent also lost resolution. Despite that particularly good
linearity was obtained for androstenedione, testosterone, and progesterone (R2 between 0.9986-0.9996, Supplementary, Ch. S3), even
if their movement to detection took long times, i.e., 9.800, 12.136,

and 24.243 min, respectively. Tandem-MS and direct infusion MS
were performed to optimise ionisation and the fragment intensities
(Supplementary, Table S1) with nESI and pESI (Supplementary, Figs.
S2–S5). The MS detection was promising, although the sensitivities
of the characteristic masses for androstenedione, androsteroneglucuronide, testosterone-glucuronide, and testosterone were too
low for detecting them in real water samples. The method limits of detection (MLD) for androstenedione, testosterone, and progesterone were 0.181, 0.661, and 0.149 μg/mL, respectively with
UV detection. Anyhow, LC-MS was only used for identification of
the steroids in pre-concentrated water samples. Finally, confirmed
analyses of the samples were done by means of UHPLC-MS/MS to
identify the steroids with accurate masses.
Determination of the studied steroids in the concentrated water samples were determined with PF-MEKC, which was optimised with standards (Experimental Ch. 2.1, 2.5, and 2.6). Due to
the structural and small polarity differences of the steroids they
did not interact interdependently with micelles in the binary micellar liquid (sodium dodecyl sulphate (SDS) and sodium taurocholate (STC)) zone before the separation electrolyte solution. Androstenedione is the lightest of the steroids, why it migrated as
the first compound to detection. Testosterone is a slightly heavier than androstenedione, but the linear geometry due to the OHgroup allows interaction with the micelles and migration as the
second to detection. On the contrary, progesterone has the highest
mass of the studied steroids due to its side-chain CO-CH3 , which
branches the structure, with the ketone group out of the nonpolar part of the molecule. Thus, progesterone can intensively interact with the binary micelles and migrate as the last steroid before
the micelles. Only testosterone-glucuronide (the anionic glucoside
conjugate with pKa 3.30, migrated near electroosmosis showing
the least interaction with micelles compared to testosterone (pKa
19.04–19.09) and progesterone (pKa 18.92). The method optimization showed that all steroid glucuronide conjugates migrated before the parent steroid. The final PF-MEKC method for tap water
measurements was validated with androstenedione, testosterone,
and progesterone standard mixtures. Table 1 informs that the system was efficient (N, calculated with N = 5.54 (tm /wh )2 for each of
the studied steroids. In addition, their resolution (R between peaks
1 and 2; R = 2(t2 −t1 )/((w1 +w2 )) was quantitative, with no overlapping with each other, matrix compounds, or the micelles.
The advantage of the PF-MEKC method was also that it was
repeatable regarding to the separation of the endogenous steroid
hormones and testosterone-glucuronide (Table 2). The correctness
of the obtained results was verified by the relative standard deviations of the absolute migration times of individual steroids,
their electrophoretic mobilities, and the mobility of electroosmosis,

which were 4.0–6.3% (RSD 1.4–3.6% in a mixture), 2.8–6.7% (RSD
4.7–7.5% in a mixture), and 3.2–4.1% (RSD 1.6–10% in a mixture),
respectively (Table 2). At applied voltage of +25 kV the relative
standard deviation of the current from run-to-run was 5.4%.
In general, the correlation coefficients for all steroid compounds
were higher than 0.99 (R2 ) (Table 3). The method limits of detection (MLD, S/N 3) and method limits of quantification (MLQ, 3 x
MLD) were from 0.33 to 0.61 μg/mL and from 0.99 to 1.84 μg/mL,
respectively. This meant that the steroid concentrations in the

2.5.3. Mass spectrometric detection
The 10, 12, 15, 25, and 50 μg/mL concentrations were used for
validation of the MS spectra of the steroids.
2.6. Method validation
Under the above-mentioned optimised conditions in the capillary electrophoresis and liquid chromatography, the methods were
validated with steroid standards and with concentrated water samples.
2.6.1. PF-MEKC and MEEKC
The electrophoretic mobilities (μep ) were measured from the
equation μep = (Ldet Ltot ) / (tR x V), where Ldet is the length of the
capillary to the detector, Ltot is its total length, tR absolute migration of the analyte to detection, and V the applied voltage during
the analyses.
The limit of detection (LOD) was measured for each steroid
mixture in methanol by using the equation LOD = (height (peak))
/ (height (noise)), which should be 3 x the noise value. The limit
of quantification (LOQ) was calculated as 3 x LOD.
Parameters (retention factor k, efficiency N, and resolution R) of
the steroids in PF-MEKC were calculated with the Eqs. (1)–(3).

k=
N=
R=

ta − tEOF
tEOF 1 −

μe V l
2DI
1√
N
4

=

ta
tmi

μe El
2D

μ
μ¯

(1)

(2)

(3)

where ta , tEOF and tmi are migration times of an analyte, electroosmosis, and micelle in Eq. (1); l is the effective length of capillary,
V is the applied voltage and E the electric field in the separation,
μe is the electrophoretic mobility and D the diffusion coefficient of

a steroid in Eq. (2); μ = μ2 − μ1 is the difference of the electrophoretic mobilities of two steroids and μ
¯ is the average of the
electrophoretic mobilities of the two steroids in Eq. (3). The micelle marker was sodium taurocholate (STC), which was the other
micelle chemical in the mixture. STC is a bile salts and has affinity
to SDS and it absorbs at the wavelengths sensitive for steroids.
The mobility of electroosmosis is calculated from each of the
analyses by using methanol as the neutral marker. Calculations
made with the equation μep = μtot − μeo , μtot = (Ldet Ltot )/(Ut
m ), and μ eo = (L det L tot )/(Ut eo ), where μ ep and μ eo are the
electrophoretic mobilities of the analyte and electroosmosis, L det
is the length of the capillary to the detector, L tot is the length of
the total capillary, U is the applied voltage during the analysis, and
tm and teo are the migration times of the analyte and electroosmosis, respectively.
2.6.2. UHPLC-UV-MS
The method optimization was made with choosing the best
separation efficiency and the highest sensitivity for the steroids in
UV and the most selective m/z value for each of the steroids.
3. Results
3.1. Optimization of the methods
Detectability of the steroid hormones was first studied extensively using both UV and ESI-MS detection with liquid chromatography. Then, twelve steroids in mixtures were used to measure
5


H. Sirén, T. Tavaststjerna and M.-L. Riekkola

Journal of Chromatography A 1649 (2021) 462233

Table 1
Validation data of selectivity parameters in PF-MEKC.
Steroid

Retention factor (k)∗

Efficiency (N)∗ [x 105 ]

Resolution (R)∗

Androstenedione (And)
Testosterone (Testo)
Progesterone (Prog)

1.57
3.01
3.49

9.37
9.76
9.84

114 (And-Testo)
26 (Testo-Prog)
not measured

∗
Average values∗∗ Testosterone-glucuronide could not been used in the study, since it could not
be detected at 247 nm.

Table 2
Validation data of sensitivity parameters in PF-MEKC and MEEKC.
Steroid

PF-MEKC Migration∗ time [min]

RSD [%]

Mobility∗ [m2 V−1 s−1 ]

RSD [%]

Repeated analyses series∗∗

Androstenedione
Testosterone
Progesterone
Testosterone-glucuronide
Electroosmosis

9.422
10.909
11.276
7.231
5.648

6.3
5.2
5.8
4.0
3.2

-2.705 × 10−8

-3.256 × 10−8
-3.370 × 10−8
-1.478 × 10−8
6.752 × 10−8

2.8
6.1
6.7
5.3
4.1

6
6
6
6
24

Steroid

MEEKC Migration∗ time [min]

RSD [%]

Mobility∗ [m2 V−1 s−1 ]

RSD [%]

Repeated analyses series∗∗

Hydrocortisone

Androstenedione
17-α -hydroxy progesterone
Testosterone
17-α -methyl testosterone
Progesterone
Electroosmosis

5.271
6.453
6.887
7.014
7.203
8.347
4.178

6.0
7.4
7.3
9.7
11.9
10.8
4.1

-4.248 × 10−8
-3.470 × 10−8
-3.251 × 10−8
-3.192 × 10−8
-3.109 × 10−8
-2.683 × 10−8
3.216 × 10−8

2.9
3.5
4.0
7.9
8.5
9.5
3.6

6
6
6
6
6
6
20

∗∗

The 6 series were done with a standard mixture randomly chosen days during one month.

Table 3
Calibration data of the steroids with PF-MEKC and MEEKC. Optimised parameters of the steroids in PF-MEKC and in MEEKC under the conditions explained in Experimental.

Compound

Concentra-tion
range in PF-MEKC
[μg/mL]

Androstene-dione

0.5-6.0

∗

0.996

y=0.341x + 0.077

0.997

y=0.570x + 0.080

0.993

y=0.394x + 0.005

0.991
Correlation (R2 )

0.884
0.880
0.931

Concentra-tion
range in MEEKC
[μg/mL]
2.0-10.0
2.0-10.0

2.0-10.0

0.50 (min 0.48
-max 0.70)
0.48 (min
0.42-max 0.53)
0.33 (min
0.27-max 0.39)
0.50 (min
0.43-max 0.81)
MLD [μg/mL]∗ ∗

1.84 (min 1.59
-max 2.10)
1.48 (min 1.26
-max 1.71)
0.99 (min 0.81
-max 1.17)
1.72 (min 1.29
-max 2.43)
MLQ [μg/mL] ∗ ∗ )

0.83
1.33
0.45

2.50
4.00
1.35

Concentration
calibration
equation∗∗ )
y=0.621x + 0.707
y=1.493x - 0.471
y=1.355x - 0.310

2.0-10.0
2.0-10.0

1.11
0.50

3.32
1.51

y=0.935x + 0.937
y=1.161x + 1.464

0.837
0.517

2.0-10-0

0.40

1.19

y=1.646x - 1.043

0.906

Progesterone

0.5-6.0

Testosteroneglucuronide
Compound

0.5-6.0

∗∗

y=0.293x + 0.053

MLQ [μg/mL]

0.5-6.0

∗

Correlation (R2 )

MLD [μg/mL]

Botrydial
Testosterone

Hydro-cortisone
Androstene-dione

17-α -hydroxyprogesterone
Testosterone
17-α -methyltestosterone
Progesterone

Concentration
calibration
equation∗∗

No standard available; identification was made with UHPLC-ESI-orbitrap-MS and tandem-MS with accurate mass method (
Average values

water samples (at ng/L level) needed enhancement to fulfil the
method-related quantification ranges in PF-MEKC-UV, because the
lowest amount in concentration calibration was 0.5 mg/L with UV
at 254 nm. The performance of the technique was studied by measuring six standard series by detecting the accuracy of the calibration range during a day and during each day in two months. The
results showed stable system throughout the work.
The usability of PF-MEKC method was compared with data
obtained microemulsion electrokinetic capillary chromatography
(MEEKC). MEEKC was chosen as the reference method since the
buffer itself does not absorb as much as the fully micellar MEKC
buffer modified with SDS. However, the measurements showed
that the MEEKC buffer could not give as low MDL values as those

less than 2; five decimals)

in PF-MEKC. As seen in Table 3 they were 1.33, 1.11, and 0.40 mg/L
for androsterone, testosterone, and progesterone, respectively. The
advantage of MEEKC was the fast analyses, since even progesterone
(the last compound) could migrate within 8.5 min (Fig. 1). Thus,

the PF method could be applied for sensitive, selective, and accurate determination of steroids in tap water.
3.2. Screening of steroid hormones with capillary electrophoresis
Steroid concentrations are extremely low in tap water [21,60].
Therefore, the water samples needed concentration enhancement
with SPE for steroid detection at the limits of the PF-MEKC method
[10,15,24,65,66].
6


H. Sirén, T. Tavaststjerna and M.-L. Riekkola

Journal of Chromatography A 1649 (2021) 462233

17α -hydroxyprogesterone. However, by using a longer column
(250 mm) hydrocortisone, 11-β -hydroxytestosterone, fluoxymesterone, androstenedione, testosterone, 17α -hydroxyprogesterone,
17α -methyl-testosterone, and progesterone were separated. Again,
the MEEKC method was used to receive reference to the estrones.
However, then high concentrations of the steroids were needed,
which lead to overlapping of the glucuronide conjugates of estrone,
estradiol, and estratriol. Thus, more LC studies were done, but with
direct infusion of the standards to MS and with UHPLC-orbitrapMS. The UHPLC studies showed that the concentrated samples enabled detection of testosterone-glucuronide, androstenedione, androsterone, testosterone, estrogens, and progesterone, although the
method was not used for quantification. However, the traditional
LC-ESI-MS was practical for detecting of the steroid fragmentation of steroids and their identification using direct infusion with
nESI and pESI modes. Mostly, the steroids were identified with
pESI mode, except detection of the glucuronide conjugates with
both nESI and pESI modes. Optimisation with high sensitivity was
needed in tandem-MS by modifying the selective collision voltage
(Table 4).
3.3. Determination of the concentrates of purified water samples

Fig. 1. Capillary electrophoresis separation of steroid standards in methanol.
(A) A PF-MEKC-UV profile. Peaks (1) testosterone glucuronide, (2) fluoxymesterone, (3) androstenedione, (4) testosterone, (5) 17α -hydroxyprogesterone, (6)
17α -methyltestosterone, and (7) progesterone. Concentration of all is 3 μg/mL.
(B) A MEEKC-UV profile. Peaks: (1) hydrocortisone, (2) androstenedione, (3) 17α -hydroxyprogesterone, (4) testosterone, (5) 17-α -methyltestosterone, (6) progesterone. Detection at 247 nm. Conditions as explained in Experimental. Current 13
μA. Concentration 20.0 μg/mL.

Steroids in the tap water samples were prepared as concentrates before the PF-MEKC. Any of the concentrates of purified water samples was steroid-free. The PF-MEKC electropherograms of the concentrates of purified water samples showed
peaks corresponding to testosterone glucuronide, androstenedione,
testosterone, and progesterone. The analytes were identified using
UHPLC-ESI-orbitrap-MS/MS to prove their existence in the concentrates. To convince the steroids, the concentrated samples were
spiked with 3 μg/mL steroid standards. They were detected by the
increase of their peak areas compared to those detected for the
concentrate samples. The steroids were also identified with electrophoretic mobility using methanol as the electroosmosis marker.
Similar procedure was performed with the MEEKC method, but the
SPE concentrated samples spiked concentrations were 8.6 and 20
μg/mL, since the MLD values were higher with MEEKC-UV method.
The PF-MEKC method allowed excellent relation between absorption and concentration compared with the equal correlation in
MEEKC was only satisfactory (Table 3).
The electropherograms of concentrated water samples justified
that the steroids in cold and hot water samples were androstenedione, testosterone, and progesterone (Fig. 2). All the samples contained progesterone. Its concentration was higher than those of
androstenedione and testosterone. Progesterone concentrations in
cold and hot tap water from Helsinki households ranged between
0.012-0.330 ng/L and 0.054-0.765 ng/L, respectively.
Fig. 3 shows the variation of progesterone concentration in the
16 locations studied in Helsinki City. According to literature on
LC studies, progesterone was found to be 0.003–0.5 ng/L in water samples [50,54]. Earlier, similar quantities at the range of 0.01–
0.33 ng/L were also measured with PF-MEKC [15]. In the present
study, the samples contained progesterone, testosterone (in 100%
of samples), its glucuronide metabolite (in 25% of samples), and
androstenedione (in 75% of the samples). The MEEKC method gave

similar information about the steroids in cold and hot water samples. In addition to the analytes discovered, in MEEK the additional
steroids were 17-α - hydroxyprogesterone and hydrocortisone, but
also diethyl phthalate. Their concentrations were 0.006–0.056 ng/L.
The results showed that the concentrations of progesterone were
the highest in the western and the central districts of Helsinki
City (Table 5). The reasons for that could not be discovered, but
the ages of the water pipelines influence the steroid concentration.
The regional differences may also be the reason, since materials of
the pipelines (steel, polymer coated, other), flow rate of water, and

The off-line concentration factor (enrichment factor, F) made
with SPE was 80 0 0 (2.0 0 L water enriched to 250 μL concentrated
sample.). Thus, by using SPE the ng/L concentrations could be concentrated to μg/mL (mg/L) level., which was the method detection level in CE with UV detection. The enrichment factor in online concentration was not measured, but it was detected from
very narrow speak shapes of the non-ionic and non-polar steroids.
Therefore, the recoveries of the analytes at 5.0 μg/mL depended on
their detectability. According to the present studies, the yields from
water were 33.8% and 64.4% for the parent steroids and testosterone glucuronide, respectively. However, when the solvent was
water-methanol mixture (1:1, v/v), the yields for all compounds
were 80–90%.
The matrix effect calculated from the analyte peak area in presence of matrix (after SPE) was divided by the analyte peak area
(standard) without matrix [67,68] gave the average values between
1.3 and 2.9 for androstenedione, testosterone, and progesterone
in PF-MEKC. Non-spiked MQ-water was also extracted and background area at the migration range of the target analytes were
subtracted from the concentrated sample profiles of the target analytes when calculating recovery. In the present work, the enrichment factor was 80 0 0 (meaning 2.0 0 L water sample concentrated
to 250 μL), which was chosen, since 250 μL volumes for four times
made repetitions were needed. The MEEKC used as the reference
was an important screening method to verify the PF-MEKC profiles due to hydrocortisone, 17-α -hydroxyprogesterone, and 17-α methyltestosterone, since in MEEKC the interaction of 17-α - functionalised steroids had considerable different mobility than their
parent steroids. Because the sensitivity was better in PF-MEKC-UV
than in MEEKC-UV, the partially filled system was selected for the
capillary electrophoresis study.

Method development with LC was studied with twelve different steroids and glucuronide metabolites using UV and MS
detectors. The efficiency of isocratic elution with the 30 mm
column used was quite good, although androstenedione could
not be separated from estrones, 17α -methyltestosterone, and
7


H. Sirén, T. Tavaststjerna and M.-L. Riekkola

Journal of Chromatography A 1649 (2021) 462233

Table 4
Mass fragments used for identification of some steroids in LC-ESI-MS.

Compound

Concentration [μg/mL]

Testosterone glucuronide

10

Androsterone glucuronide

Androstenedione

Testosterone

Confirmation ions m/z [Da]
(% = normalised intensities

of the fragments)
463 (100%)
464 (20%)
465 (5%)
463 (100%)
Ions 75, 85, 113, 403 (low
abundance)
465 (100%)
466 (30%)

15

Ionization mode negative
(-) positive (+)

Solvent
Acetonitrile-water (1:1,
v/v) cont. 0.1% ammonium
hydroxide

MS2 (-)
collision 25 eV
MS(-)
collision 25 eV
MS2 (-)
collision 35 eV

465 (20%)
Ions 75 (98%),
85 (100%),

113 (50%)
287 (100%)
288 (30%)
328 (40%)
329 (5%)
97 (100%)
109 (70%)
123 (10%)
289 (100%)
290 (20%)
330 (20%)
289 (100%)
271 (3%)
253 (4%)
97 (85%)
109 (50%)

12

12

MS(-)
collision 25 eV

MS(+)
collision 20 eV

MS2 (+)
collision 25 eV
MS(+)

collision 20 eV
MS2 (+)
collision 15 eV

Table 5
Concentrations of progesterone, testosterone, and androstenedione in cold and hot tap water extracts determined with the PF-MEKC. The water samples were from Latokartano, Pikku-Huopalahti,
Vesala, Etelä-Haaga, Kalasatama, Pasila, Kumpula (1-5), Ullanlinna, Etu-Töölö, Munkkiniemi, and Viikki. Conditions are as explained in Experimental. Testosterone-glucuronide was detected, but it was
not quantified, since the amount was near its MLQ value.
Progesterone

Testosterone

Androstenedione

Water sources

Cold
[ng/L]

Hot

Cold
[ng/L]

Hot

Cold
[ng/L]

Hot

Cold
[ng/L]

Steroids
Hot

Latokartano
Pikku-Huopalahti
Vesala
Etelä-Haaga 1
Kalasatama
Pasila
Etelä-Haaga 2
Kumpula 1
Kumpula 2
Kumpula 3
Kumpula 4
Kumpula 5
Ullanlinna
Etu-Töölö
Munkkiniemi
Viikki
16 extracts

0.013
0.033
nd
0.033
0.006

0.018
0.006
0.022
0.027
0.018
0.020
0.029
0.016
0.029
0.015
0.027

0.036
0.056
0.015
0.077
0.018
0.026
0.011
0.028
0.037
0.023
0.022
0.036
0.060
0.043
0.019
0.027

0.001

0.001
0.001
0.007
0.001
0.002
0.007
nd
nd
nd
nd
nd
0.003
nd
nd
nd

0.002
0.003
0.002
0.008
0.009
0.008
0.008
nd
nd
nd
nd
nd
0.005
nd

0.007
0.008

nd
nd
nd
nd
0.001
0.001
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd

nd
0.008
nd
0.006
0.008
0.005
0.005
nd
nd
nd

nd
nd
nd
nd
nd
nd

0.014
0.034
0.001
0.040
0.007
0.021
0.013
0.022
0.027
0.018
0.020
0.029
0.019
0.029
0.015
0.027

0.038
0.059
0.017
0.091
0.035
0.039

0.024
0.028
0.037
0.023
0.022
0.036
0.065
0.043
0.026
0.035

Average values from 4 repetitions. Three replicate samples (á 250 μL) from the concentrates of 1
mL. (nd = not determined)

network of households’ connections with the water pipeline have
disparity.
The steroid results showed that the concentration of progesterone in tap water from Ullanlinna was moderately higher in
hot water than in cold water (0.134 and 0.563 ng/L, respectively)
(Table 5). Similar effect was noticed in water samples of Latokartano, Pikku-Huopalahti, and Etelä-Haaga. In addition, elevated
amounts of other steroids were found in Ullanlinna compared to
water in Kumpula (Table 5). Testosterone and androstenedione
amounts were 30% lower in Kumpula (hill area) than in Ullanlinna

(seaside). Overall, in seaside area the steroid concentrations were
higher than in the central area and new constructed suburban areas.
Visual perception about the hot water samples from Ullanlinna
and Etelä-Haaga showed before SPE treatment that they were reddish in colour. The filtrated precipitation and the soluble mold
were identified with UHPLC-ESI-orbitrap-MS(MS2 ). The phytotoxic
sesquiterpene metabolite botrydial from Botrytis cinerea (Supplementary reference S11) mold was identified in all concentrates of
purified water. Identification of botrydial made with SCAN(+) MS

8


H. Sirén, T. Tavaststjerna and M.-L. Riekkola

Journal of Chromatography A 1649 (2021) 462233

Fig. 3. Progesterone concentrations in
PF-MEKC technique. The concentrated
Pikku-Huopalahti, Vesala, Etelä-Haaga,
Munkkiniemi, and Viikki. Conditions are

tap water samples determined with the
water samples were from Latokartano,
Kalasatama, Pasila, Kumpula, Ullanlinna,
as explained in Experimental.

erature that describes about discoveries of filamentous fungi identified from 14 drinking water systems in Norway [59]. The paper
informs that mold is detected in cold water. However, in our study,
the concentrations of botrydial metabolite of B. cinerea were higher
in hot water than in cold water samples. That was supported the
knowledge that only a small number of mold can grow at temperatures of 4°C or below (Supplementary reference S12). Steroids
belonging to oil compounds can resist B. cinerea, since mold cause
biodegradation of natural materials and may grow on dead organic
matter everywhere in nature. Thus, tap water may have an odd
odor and flavor problems.
5. Conclusions
Capillary electrophoresis showed capability in studies of male
and female steroids in concentrates of water which was made
run from Päijänne Lake and delivered from purification plants to

Helsinki City. Partial filled micellar electrokinetic chromatography
with UV detection applied for SPE treated concentrates of purified
water was sensitive and selective for infinite small steroid concentrations. The methods used detected also microbial pollutants in
drinking water. Our results prove that by applying UV detection at
specific wavelength for steroids, the steroid identification can be
conducted.

Fig. 2. Electropherograms as examples of hot tap water samples analysed with PFMECK-UV. The concentrate samples are from (A) Vesala and (B) Pasila, and (C)
the reference water. Compounds: (1) androstenedione, (2) testosterone, (3) progesterone, and (4) botrydial. The compounds were identified by spiking. Detection at
the UV wavelength 247 nm. Separation conditions are explained in Experimental.

and MS2 modes gave accurate molar mass of 311.18533 Da for
C17 H27 O5 ( 0.008250 ppm) with the fragment of 170.15390 Da
(80%). The MS2 results to confirm botrydial from the isolation
ion 311.18454 Da (C17 H27 O5 ,
-2.44430 ppm) gave the fragments
293.17419 Da (100%), 265.17931 Da (50%), 255.12222 Da (20%), and
237.11170 Da (10%).
Both cold and hot tap water contained botrydial as high as 861–
3900% of the biologically grown mold with respect to a drilled well
water. Except for the purified water all samples from Vesala area
very rich in mold growth being at 2320–2460% level. The lowest
botrydial amounts (9% in cold water and 23% in hot water) were
detected in the new built suburban area Viikki.

Statement of human and animal rights
In this project, humans or animals have not been used to test
the water samples. We have neither had any sensory evaluator
boards. The persons in households gave the permission to sample the tap water when they wanted. The households also have
received results of their own water samples analysed by the instrumentation informed in methods.

Declaration of Competing Interest
The authors declare that they do not have competing financial
interest concerning the project. They do not have any conflicts either.

4. Discussion
The results showed that all water samples pretreated for concentrates of purified water from Päijänne Lake contained progesterone. Its high concentration in tap water might be caused by
the endogenic progesterone but also contraceptive pharmaceuticals
since steroids need specific removal from effluents of purification
plants. The water samples can be contaminated with mold which
may grow in the water pipes. The observation is supported by lit-

CRediT authorship contribution statement
Heli Sirén: Methodology, Investigation, Data curation, Writing review & editing. Tuomas Tavaststjerna: Writing - review & editing, Investigation, Data curation. Marja-Liisa Riekkola: Supervision.
9


H. Sirén, T. Tavaststjerna and M.-L. Riekkola

Journal of Chromatography A 1649 (2021) 462233

Acknowledgments
[23]

The Helsinki City Foundation (Grant 2017) is acknowledged for
the financial support. The authors acknowledge I. Rekola, A. Puolakka, M. Tilli, and L. Guricza for their assistance in laboratory.

[24]

Supplementary materials
Supplementary material associated with this article can be

found, in the online version, at doi:10.1016/j.chroma.2021.462233.

[25]

References

[27]

[26]

[1] A.C. Naldi, P.B. Fayad, S. Sauve, M. Prevost, Analysis of steroid hormones and
their conjugated forms in water and urine by on-line solid-phase extraction
coupled to liquid chromatography tandem mass spectrometry, Chem. Central
J. 1 (2016) 1–17.
[2] J.K. Leet, S. Sassman, J.J. Amberg, A.W. Olmstead, L.S. Lee, G.T. Ankley,
M.S. Sepulveda, Environmental hormones and their impacts on sex differentiation in fathead minnows, Aquat. Toxicol. 158 (2015) 98–107.
[3] C.P. Silva, M. Otero, V. Esteves, Processes for the elimination of estrogenic
steroid hormones from water: a review, Env. Pollut. 165 (2012) 38–58.
[4] A.C. Johnson, H.-R. Aerni, A. Gerritsen, M. Gibert, W. Giger, K. Hylland, M. Jürgens, T. Nakari, A. Pickering, M.J.F. Suter, A. Svenson, F.E. Wettstein, Comparing steroid estrogen, and nonylphenol content across a range of European
sewage plants with different treatment and management practices, Water Res.
39 (2005) 47–58.
[5] Z.H. Liu, Y. Kanjo, S. Mizutani, Removal mechanisms for endocrine disrupting
compounds (EDCs) in wastewater treatment — physical means, biodegradation,
and chemical advanced oxidation: a review, Sci. Total Env. 407 (2009) 731–748.
[6] S.Y. Wee, A.Z. Aris, Endocrine disrupting compounds in drinking water supply
system and human health risk implication, Env. Inter. 106 (2017) 207–233.
[7] />[8] https://yle.fi/uutiset/3-8731424
[9] M. Ferrey, Wastewater Treatment Plant Endocrine Disrupting Chemical Monitoring Study, Report, Minnesota Pollution Control Agency, 2011.
[10] H. Sirén, S. El Fellah, Androgens, oestrogens, and progesterone concentrations
in wastewater purification processes measured with capillary electrophoresis,

Env. Sci. Poll. Res. 24 (2017) 16765–16785.
[11] M.J. Benotti, R.A. Trenholm, B.J. Vanderford, J.C. Holady, B.D. Stanford, S.A. Snyder, Pharmaceuticals and endocrine disrupting compounds in U.S. drinking water, Environ. Sci. Technol. 43 (2009) 597–603.
[12] M.O. Barbosa, N.F.F. Moreira, A.R. Ribeiro, M.F.R. Pereira, A.M.T. Silva, Occurrence and removal of organic micropollutants: an overview of the watch list
of EU Decision 2015/495, Water Res. 94 (2016) 257–279.
[13] D. Balabanicˇ , D. Hermosilla, A. Blanco, N. Merayo, A. Krivograd Klemencˇ icˇ , The
possibility of removal of endocrine disrupters from paper mill waste waters
using anaerobic and aerobic biological treatment, membrane bioreactor, ultrafiltration, reverse osmosis and advanced oxidation processes, Environmental
Toxicology III, www.witpress.com, ISSN 1743-3541 (on-line); WIT Transactions
on Ecology and the Environment, Vol 132, © 2010 WIT Press; Environmental
Toxicology III, pp. 33-44. doi:10.2495/ETOX10 0 041
[14] P. Rajasulochana, V. Preethy, Comparison on efficiency of various techniques in
treatment of waste and sewage water – a comprehensive review, Resource-Effici. Technol. 2 (2016) 175–184.
[15] H. Sirén, S. El Fellah, Steroids contents in waters of wastewater purification
plants: determination with partial-filling micellar electrokinetic capillary chromatography and UV detection, Int. J. Env. Anal. Chem. 96 (2016) 1003–1021.
[16] R. Vallejo-Rodríguez, P.B. Sánchez-Torres, A. López-López, E. León-Becerril,
M. Murillo-Tovar, Detection of steroids in tap and drinking water using an optimized analytical method by gas chromatography–mass spectrometry, Expo
Health 10 (2018) 189–199.
[17] Proposal for a Directive of the European Parliament and of the Council on
the quality of water intended for human consumption (recast). COM (2017)
753 final 2017/0332 (COD). p. 18. />water-drink/pdf/revised_drinking_water_directive.pdf
[18] M. Tagliavini, P.G. Weidler, C. Njel, J. Pohl, D. Richter, B. Böhringer, A.I. Schäfer,
Polymer-based spherical activated carbon – ultrafiltration (UF-PBSAC) for the
adsorption of steroid hormones from water: material characteristics and process configuration, Water Res. 185 (2020) 116249.
[19] C. Chia-Yang, W. Tzu-Yao, W. Gen-Shuh, C. Hui-Wen, L. Ying-Hsuan,
L. Guang-Wen, Determining estrogenic steroids in Taipei waters and removal
in drinking water treatment using high-flow solid-phase extraction and liquid chromatography/tandem mass spectrometry, Sci. Total Environ. 378 (2007)
352–365.
[20] E. Magi, C. Scapolla, M Di Carro, C. Liscio, Determination of endocrine-disrupting compounds in drinking waters by fast liquid chromatography-tandem mass
spectrometry, J. Mass Spectrom. 45 (2010) 1003–1011.
[21] Y. Yu, L. Wu, A.C. Chang, Seasonal variation of endocrine disrupting compounds, pharmaceuticals and personal care products in wastewater treatment

plants, Sci. Total Env. 442 (2013) 310–316.
[22] C.J. Houtman, J. Kroesbergen, K. Lekkerkerker-Teunissen, J.P. van der Hoek, Human health risk assessment of the mixture of pharmaceuticals in Dutch drink-

[28]

[29]

[30]

[31]

[32]

[33]

[34]

[35]

[36]

[37]

[38]

[39]

[40]
[41]

[42]

[43]

[44]

[45]

[46]

[47]

10

ing water and its sources based on frequent monitoring data, Sci. Total Env.
496 (2014) 54–62.
H.Fr. Schröder, W. Gebhardt, M. Thevis, Anabolic, doping, and lifestyle drugs,
and selected metabolites in wastewater—detection, quantification, and behaviour monitored by high-resolution MS and MS n before and after sewage
treatment, Anal. Bioanal. Chem. 398 (2010) 1207–1229.
S. El Fellah, G. Duporte, H. Siren, Steroid hormones, inorganic ions and botrydial in drinking water. Determination with capillary electrophoresis and liquid chromatography-orbitrap high resolution mass spectrometry, Microchem.
J. 133 (2017) 126–136.
R. Mäkinen, Drinking water quality and network materials in Finland. Summary Report (2008).
Finnish accreditation service, T225, (EN/ISO/IEC 17025, Quality of water 1.1.31.12.2020
A. Arditsoglou, D. Voutsa, Partitioning of endocrine disrupting compounds in
inland waters and wastewaters discharged into the coastal area of Thessaloniki, Northern Greece, Environ. Sci. Pollut. Res. 17 (2010) 529–538.
E. Vulliet, C. Cren-Olivé, Screening of pharmaceuticals and hormones at the
regional scale, in surface and groundwaters intended to human consumption,
Env. Poll. 159 (2011) 2929–2934.
S.T. Glassmeyer, E.T. Furlong, D.W. Kolpin, A.L. Batta, R. Benson, J.S. Boone,
O. Conerly, M.J. Donohue, D.N. King, M.S. Kostich, H.E. Masha, S.L. Pfaller,

K.M. Schenck, J.E. Simmons, E.A. Varughese, S.J. Vesper, E.N. Villegas, V.S. Wilson, Nationwide reconnaissance of contaminants of emerging concern in
source and treated drinking waters of the United States, Pharmaceuticals. Sci.
Total Environ. 581-582 (2017) 909–922.
V.L. Cunningham, C. Perino, V.J. D’Aco, A. Hartmann, R. Bechter, Human health
risk assessment of carbamazepine in surface waters of North America and Europe, Regul. Toxicol. Pharmacol. 56 (2010) 343–351.
A.C. Johnson, M.D. Juergens, R.J. Williams, K. Kuemmerer, A. Kortenkamp,
J.P. Sumpter, Do cytotoxic chemotherapy drugs discharged into rivers pose a
risk to the environment and human health? An overview and UK case study,
J. Hydrol. 348 (2008) 167–175.
E. Vulliet, C. Cren-Olivé, Screening of pharmaceuticals and hormones at the
regional scale, in surface and groundwaters intended to human consumption,
Environ. Pollut. 159 (2011) 2929–2934.
M.J. Benotti, B.D. Stanford, E.C. Wert, S.A. Snyder, Evaluation of a photocatalytic reactor membrane pilot system for the removal of pharmaceuticals and
endocrine disrupting compounds from water, Water Res. 43 (2009) 1513–1522.
K. Lekkerkerker-Teunissen, E.T. Chekol, S.K. Maeng, K. Ghebremichael, C.J. Houtman, A.R.D. Verliefde, J.Q.J.C. Verberk, G. Amy, R. van Egmond, Pharmaceutical
removal during managed aquifer recharge with pretreatment by advanced oxidation, Water Sci. Technol. Water Supply 12 (2012) 755–767.
C.J. Houtman, J. Kroesbergen, K. Lekkerkerker-Teunissen, J.P. van der Hoek, Human health risk assessment of the mixture of pharmaceuticals in Dutch drinking water and its sources based on frequent monitoring data, Sci. Total Env.
496 (2014) 54–62.
E. Magi, C.M. Scapolla, M. Di Carro, C. Liscio, Determination of endocrine-disrupting compounds in drinking waters by fast liquid chromatography–tandem
mass spectrometry, J. Mass Spectr. 45 (2010) 1003–1011.
L. Barreiros, J.F. Queiroz, L.M. Magalhaes, A.M.T. Silva, M.A. Segundo, Analysis of 17 β -estradiol and 17-α -ethinylestradiol in biological and environmental
matrices-a review, Microchem. J. 126 (2016) 243–262.
R. Vallejo-Rodríguez, P.B. Sánchez-Torres, A. López-López, E. León-Becerril,
M. Murillo-Tovar, Detection of steroids in tap and drinking water using an optimized analytical method by gas chromatography–mass spectrometry, Exposure Health (2017) 1–11.
H.M. Kuch, K. Ballschmiter, Determination of endocrine-disrupting phenolic
compounds and estrogens in surface and drinking water by HRGC−(NCI)−MS
in the picogram per liter range, Environ. Sci. Technol. 35 (2001) 3201–3206.
I. Vasconcelos, C. Fernandes, Magnetic solid phase extraction for determination
of drugs in biological matrices, TrAC Trends Anal. Chem. 89 (2017) 41–52.
R.A. Pérez, B. Albero, J.L. Tadeo, E. Molero, C. Sánchez-Brunete, Analysis of

steroid hormones in water using palmitate coated magnetite nanoparticles
solid-phase extraction and gas chromatography–tandem mass spectrometry,
Chromatographia 77 (2014) 837–843.
M. Velicu, R. Suri, Presence of steroid hormones and antibiotics in surface water of agricultural, suburban and mixed-use areas, Env. Monit. Ass. 154 (2009)
349–359.
E. Aydin, I. Talinli, Analysis, occurrence and fate of commonly used pharmaceuticals and hormones in the Buyukcekmece Watershed, Turkey, Chemosphere
90 (2013) 2004–2012.
H. Chang, S. Wu, J. Hu, M. Asami, S. Kunikane, Trace analysis of androgens and
progestogens in environmental waters by ultra-performance liquid chromatography-electrospray tandem mass spectrometry, J. Chrom. A 1195 (2008) 44–
51.
H.T.T. Thuy, T.D. Nguyen, The potential environmental risks of pharmaceuticals
in Vietnamese aquatic systems: case study of antibiotics and synthetic hormones, Env. Sci. Pollut. Res. 20 (2013) 8132–8140.
P.D. Scott, M. Bartkow, S.J. Blockwell, H.M. Coleman, S.J. Khan, R. Lim, J.A. McDonald, H. Nice, D. Nugegoda, V. Pettigrove, L.A. Tremblay, M.S.J. Warne,
F.D.L. Leusch, An assessment of endocrine activity in Australian rivers using
chemical and in vitro analyses, Env. Sci. Pollut. Res. 21 (2014) 12951–12967.
T.Y. Fang, S.M. Praveena, C. de Burbure, A.Z. Aris, S.N. Ismail, I. Rasdi, Analytical
techniques for steroid estrogens in water samples - a review, Chemosphere 165
(2016) 358–368.


H. Sirén, T. Tavaststjerna and M.-L. Riekkola

Journal of Chromatography A 1649 (2021) 462233

[48] H.R. Kasambala, M.J. Rwiza, R.H. Mdegela, Levels and distribution of progesterone in receiving waters and wastewaters of a growing urban area, Water
Sci. Technol. 80 (6) (2019) 1107–1117.
[49] K. Zhang, Y. Zhao, K. Fent, Occurrence and ecotoxicological effects of free,
conjugated, and halogenated steroids including 17α -hydroxypregnanolone and
pregnanediol in swiss wastewater and surface water, Environ. Sci. Technol. 51
(11) (2017) 6498–6506.

[50] F. Zhang, L. Yang, Y. Xie, D. Wang, X. Li, Z. Nie, Accumulation of steroid hormones in soil and its adjacent aquatic environment from a typical intensive
vegetable cultivation of North China, Sci. Total Environ. 538 (2015) 423–430.
[51] P. Avar, G. Maasz, P. Takacs, S. Lovas, Z. Zrinyi, R. Svigruha, A.G. Takatsy, L. Toth,
Z. Pirger, HPLC-MS/MS analysis of steroid hormones in environmental water
samples, Drug Test Anal. 8 (2016) 124–128.
[52] S. Comtois-Marotte, T. Chappuis, S. Vo Duy, N. Gilbert, A. Lajeunesse, S. Taktek, M. Desrosiers, E. Veilleux, S. Sauve, Analysis of emerging contaminants in
water and solid samples using high resolution mass spectrometry with a Q
Exactive orbital ion trap and estrogenic activity with YES-assay, Chemosphere
166 (2017) 400–411.
[53] M.A.K. Hashmi, B.I. Escher, M. Krauss, I. Teodorovic, W. Brack, Effect-directed
analysis (EDA) of Danube River water sample receiving untreated municipal wastewater from Novi Sad, Serbia, Sci. Total Environ. 624 (2018) 1072–
1081.
[54] R. Guedes-Alonso, Z. Sosa-Ferrera, J.J. Santana-Rodrıguez, An on-line solid
phase extraction method coupled with UHPLC-MS/MS for the determination
of steroid hormone compounds in treated water samples from waste water
treatment plants, Anal. Methods 7 (2015) 5996–6005.
[55] F.D.L. Leusch, P.A. Neale, C. Arnal, N.H. Aneck-Hahn, P. Balaguer, A. Bruchet,
B.I. Escher, M. Esperanza, M. Grimaldi, G. Leroy, M. Scheurer, R. Schlichting,
M. Schriks, A. Hebert, Analysis of endocrine activity in drinking water, surface
water and treated wastewater from six countries, Water Res. 139 (2018) 10–18.
[56] H. Tomšíková, J. Aufartová, P. Solich, Z. Sosa-Ferrera, J.J. Santana-Rodríguez,
L. Nováková, High-sensitivity analysis of female-steroid hormones in environmental samples, TrAC Trends Anal. Chem. 34 (2012) 35–58.
[57] M. Saraji, M.K. Boroujeni, Recent developments in dispersive liquid-liquid microextraction, Anal. Bioanal. Chem. 406 (2014) 2027–2066.

[58] X.E Zhao, P. Yan, R. Wang, S. Zhu, J. You, Y. Bai, H. Liu, Sensitive determination
of cholesterol and its metabolic steroid hormones by UHPLC-MS/MS via derivatization coupled with dual ultrasonic-assisted dispersive liquid-liquid microextraction, Rapid Commun. Mass Spectrom. 30 (2016) 147–154.
[59] M.N. Mons, M.B. Heringa, J. van Genderen, L.M. Puijker, W. Brand, C.J. van
Leeuwen, P. Stoks, J.P. van der Hoek, D. van der Kooij, Use of the threshold
of toxicological concern (TTC) approach for deriving target values for drinking
water contaminants, Water Res. 47 (2013) 1666–1678.

[60] G. Hageskal, P. Gaustad, B.T. Heier, I. Skaar, Occurrence of moulds in drinking
water, J. Appl. Microbiol. 102 (2007) 774–780.
[61] D.A. Schisler, W.J. Janisiewicz, Teu Boekhout, C.P. Kurtzman, The Yeasts (Fifth
Edition), Chapter 4 - Agriculturally important yeasts: biological control of field
and postharvest diseases using yeast antagonists, and yeasts as pathogens of
plants. (2011) 45-52. 0 0 04-5
[62] A. Grabke, G. Stammler, A Botrytis cinerea Population from a single strawberry
field in Germany has a complex fungicide resistance ptern, Plant Disease 99
(2015) 1078–1086.
[63] />[64] SFS Manual 147-1 - Water quality. Part 1: Methods for sampling. ISO/TC 147/SC
6.
[65] H. Noppe, K. Verheyden, W. Gillis, D. Courtheyn, P. Vanthemsche, H.F. De
Brabander, Multi-analyte approach for the determination of ng L−1 levels
of steroid hormones in unidentified aqueous samples, Anal. Chim. Acta 586
(2007) 22–29.
[66] L. Patrolecco, N. Ademollo, P. Grenni, A. Tolomei, A. Barra Caracciolo, S. Capri,
Simultaneous determination of human pharmaceuticals in water samples by
solid phase extraction and HPLC with UV-fluorescence detection, Microchem.
J. 107 (2013) 165–171.
[67] S. El Fellah, H. Sirén, Recent developments in capillary electrophoresis of
steroids and sterols, J. Biomed. Res. Pract. 2 (1) (2018) 1–21 Article ID: 10 0 0 06.
[68] M. Nic, J. Jirat, B. Kosata, IUPAC. Compendium of Chemical Terminology (The
"Gold Book"), 2nd ed., Compiled by A. D. McNaught and A. Wilkinson. Blackwell Scientific Publications, Oxford (1997). XML on-line corrected version:
(2006), updates compiled by A. Jenkins. ISBN 09678550-9-8.

11