Analytica Chimica Acta 687 (2011) 159–167
Contents lists available at ScienceDirect
Analytica Chimica Acta
journal homepage: www.elsevier.com/locate/aca
Determination of the nucleosides and nucleobases in Tuber samples
by dispersive solid-phase extraction combined with liquid
chromatography–mass spectrometry
Ping Liu a , Yuan-Yuan Li a,d , Hong-Mei Li a , Duan-Ji Wan a , Ya-Jie Tang a,b,c,∗
a
Key Laboratory of Fermentation Engineering (Ministry of Education), Hubei Provincial Key Laboratory of Industrial Microbiology,
College of Bioengineering, Hubei University of Technology, Wuhan 430068, China
b
Lab of Bioreﬁnery Engineering, Shanghai Advanced Research Institute, Chinese Academy of Sciences, Shanghai 201203, China
c
National Key Laboratory of Biochemical Engineering, Institute of Process Engineering, Chinese Academy of Sciences, Beijing 100080, China
d
Research Group for Bioactive Products, Department of Biology & Chemistry, City University of Hong Kong, Kowloon, Hong Kong SAR, China
a r t i c l e
i n f o
Article history:
Received 6 September 2010
Received in revised form
12 November 2010
Accepted 16 December 2010
Available online 23 December 2010
Keywords:
Trufﬂe
Nucleosides
Nucleobases
Protein removal
Dispersive solid-phase extraction
Liquid chromatography–mass
spectrometry (LC–MS)
a b s t r a c t
A simple, fast and inexpensive method based on dispersive solid phase extraction (DSPE) combined
with LC–MS was developed for simultaneous determination of 7 nucleosides and nucleobases (i.e., adenine, hypoxanthine, uridine, adenosine, guanine, guanosine, and inosine) in Tuber fruiting-bodies and
fermentation mycelia. The DSPE procedure was ﬁrstly introduced to remove the protein interference
from sample solutions, and D3520 macroporous resin was chosen as the DSPE sorbent because of its high
removal capability on protein interferences, but low adsorption rate on analytes. Besides, key parameters
on DSPE procedure (i.e., macroporous resin type, macroporous resin amount, methanol concentration,
and vortex time) were optimized, and the protein removal efﬁcacy could achieve about 95% after the
process optimization. Though the method validation test, the DSPE-LC–MS method was conﬁrmed to
be precise, accurate and sensitive, and the column blinding problem was solved successfully. By using
this established method, the total amount of nucleosides and nucleobases in the fermentation mycelia
was determined to range from 4881.5 to 12,592.9 ␮g g−1 , which was about 2–25 times higher than the
fruiting-bodies (from 498.1 to 2274.1 ␮g g−1 ). The formulation of nucleosides and nucleobases in the fermentation mycelia maintained relatively constant, while the formulation in Tuber fruiting-bodies varied
signiﬁcantly with their species. Hierarchical cluster analysis (HCA) showed the formulation similarity
of nucleosides and nucleobases between Tuber fermentation mycelia and the fruiting-bodies of Tuber
indicum and Tuber himalayense. From the viewpoint of nucleosides and nucleobases, this work conﬁrms
the potentiality of Tuber fermentation mycelia as the alternative resource for its fruiting-bodies.
© 2010 Elsevier B.V. All rights reserved.
1. Introduction
Trufﬂe, the hypogeous fungus in Tuber species with characteristic aroma and delicious taste, is precious and expensive delicacies,
and widely used in the famous French and Italian cuisines. Because
of the natural production decrease and the worldwide demand
increase, submerged fermentation to produce Tuber mycelia as an
alternative resource for its fruiting-body is a potential way to solve
this problem. Our lab successfully developed trufﬂe fermentation
system and systematically optimized the medium components,
including carbon source [1], nitrogen source [2], metal ion [3],
and plant oil [4], and developed a novel high-cell density fed-
∗ Corresponding author at: Key Laboratory of Fermentation Engineering (Ministry
of Education), Hubei Provincial Key Laboratory of Industrial Microbiology, College of
Bioengineering, Hubei University of Technology, Room 310, Wuhan 430068, China.
Tel.: +86 27 88015108; fax: +86 27 88015108.
E-mail address: [email protected] (Y.-J. Tang).
0003-2670/$ – see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.aca.2010.12.025
batch fermentation process [5]. Besides these, we identiﬁed the
existence of the bioactive androstenol in Tuber fermentation system [6], compared the volatile organic compound (VOC) between
Tuber fruiting-bodies and fermentation mycelia [7], and studied the
variation of VOC composition with culture condition [8]. All these
results demonstrated the chemical composition similarity between
Tuber fermentation mycelia and fruiting-bodies, and it is possible
to adjust the aroma of trufﬂe fermentation mycelia similar with the
natural fruiting-body through the control of fermentation process
[6–8].
In order to reveal the secret of trufﬂe special aroma, scientists
were more interested in the VOC composition in Tuber fruitingbodies, and more than 200 violates were identiﬁed in the recent
20 years [9,10]. Simultaneously, lipid soluble constituent, including
ceramide [11,12], fatty acids [13,14], and sterol [15], have also been
identiﬁed. However, little attention was paid to hydrosoluble components. As a serial of very important hydrosoluble constituents,
nucleosides and nucleobases have a lot of physiological activities
160
P. Liu et al. / Analytica Chimica Acta 687 (2011) 159–167
Table 1
Physical and chemical properties of the dispersive sorbent of macroporous resins.
Trade name
Functional group
Surface area (m2 g−1 )
Average pore diameter (nm)
D3520
D4006
D101
D312
AB-8
Polystyrene
Polystyrene
Polystyrene
Polystyrene
Polystyrene
480–520
400–440
500–550
500–650
480–520
85–90
65–75
90–100
50–60
130–140
including anticonvulsant activity [16], stimulating axon growth
in vitro and in the adult central nervous system [17], inﬂuencing
the growth and differentiation of the gastro intestinal tract [18],
and maintaining the immune response [19]. Therefore, nucleosides
and nucleobases were selected as the quality control marker in the
medicinal higher fungi Ganoderma and Cordyceps genus [20,21].
Due to the remarkable activities of nucleosides and nucleobases,
and their research blank in Tuber genus, it was necessary to qualitatively and quantitatively compare the formulation of nucleosides
and nucleobases between Tuber fruiting-bodies and fermentation
mycelia.
The current techniques for assaying nucleosides and nucleobases mainly include liquid chromatography (LC) [22,23] or
capillary electrophoresis (CE) [24,25] combined with ultraviolet
detector (UV) or mass spectra (MS). It is worthy noting that most
of the injection samples were always extracted by water or low
concentration methanol aqueous solutions without any clear up
step. However, in our preliminary experiment, HPLC system pressure was raised by the assaying samples prepared by the described
method [23], and the peaks were distorted when multiple analyses
were undertaken. This phenomenon was probably due to the accumulation of dissolved sample protein in the column head, which
were co-extracted by high polarity solvent. Therefore, the clean-up
step for the removal of protein in solution prior to chromatographic
analysis is necessary for assaying nucleosides and nucleobases in
Tuber samples.
The usual clean-up procedure for removing protein from sample
solutions include organic solvent precipitation [26,27], solid-phase
extraction (SPE) [28,29], and precolumn-switching technique [30].
As to the organic solvent precipitation method, high proportion of
non-polar organic solvent was needed to remove protein. While,
high concentration of organic solvent would reduce the solubility of water-soluble compounds, and result in poor recovery. So,
organic solvent precipitation was not suitable for the water-soluble
analytes. The shortcomings of the SPE and precolumn-switching
technique were time-consuming, cumbersome, expensive, and the
extra equipment needed. The dispersive solid-phase extraction
(DSPE) procedure, a fairly rapid and simple technique with high
recovery and accuracy, is widely used in agricultural chemicals
residue [31,32] and other analytes [33] determination, which might
be also useful to remove the protein interference in the complex biological matrix. Therefore, with the purpose to remove
the protein from Tuber sample solution by a more simple way,
the new DSPE procedure was tried to solve to column blinding
problem. Generally speaking, the sorbents widely used in DSPE
are graphitized carbon black (GCB), primary secondary amine
(PSA), octadecylsilane (C18 ), aminopropyl (–NH2 ), and aluminaN [34]. No attention was paid to the kinds of advantageous
sorbents—macroporous resins. Because of the outstanding virtues,
including unique adsorption properties ideal pore structure and
various surface functional groups available, low operation expense,
less solvent consumption and easy regeneration, macroporous
resins have been gained a growing interest in the ﬁeld of bioactive
compounds separation [35,36]. Therefore, it might be a good try to
use a suitable macroporous resins as the DSPE cleanup sorbent to
remove the interference protein in the nucleosides and nucleobases
assay for Tuber samples.
Particle diameter (mm)
0.3–1.25
0.3–1.25
0.3–1.25
0.3–1.25
0.3–1.25
Polarity
Non-polar
Non-polar
Non-polar
Non-polar
Low polar
By developing the assay method coupling DSPE with LC–MS in
this work, we qualitatively and quantitatively assayed the nucleosides and nucleobases in Tuber fruiting-bodies and fermentation
mycelia for the ﬁrst time. The DSPE procedure was ﬁrstly introduced to remove the protein interference. More precisely, this work
include (1) identiﬁed the types of nucleosides and nucleobases
in Tuber samples; (2) developed the DSPE procedure for protein
removing; (3) assayed the contents of the target nucleosides and
nucleobases in various Tuber samples and analyzed the relationship of Tuber fruiting-bodies and fermentation mycelia through
hierarchical clustering analysis. Furthermore, the developed DSPE
method for the removal of protein could also be used in other biological samples with protein interference in matrix.
2. Experimental
2.1. Reagents and materials
Methanol (HPLC grade) was purchased from Merck (Darmstadt,
Germany). Ammonium acetate and formic acid were HPLC grade.
Water was prepared using a Millipore Milli Q-Plus system (Millipore, Bedford, MA).
Uridine, hypoxanthine, thymine, guanine, guanosine, adenine,
adenosine, and inosine were purchased from Sigma (St. Louis, MO,
USA). The mobile phase was used as the solvent for stock solution preparation, and the concentrations for each standard were
about 0.2 mg mL−1 except the guanine of 0.09 mg mL−1 . A certain
volume of stock solution was transferred to 10 mL volumetric ﬂask
and diluted with mobile phase to the desired concentration. All the
standard solutions were stored at 4 ◦ C in the dark.
Macroporous resins including D-3520, D-4006, D-101, and AB-8
were purchased from Nankai Hecheng S & T (Tianjin, China), and
macroporous resin D312 was purchased from Shandong Lukang
Pharmaceutical Group Co., Ltd. (Shandong, China). Their physical
and chemical properties were summarized in Table 1. Macroporous
resins were soaked in 95% ethanol, shaken for 24 h and thoroughly
washed by deionized water [35,37]. Moisture content of macroporous resin was determined from mass difference after drying at
60 ◦ C until the mass maintained constant.
2.2. Trufﬂe fruiting-body collection and mycelia culture
All natural trufﬂe fruiting-bodies were collected in China. Tuber
indicum, Tuber aestivum were obtained from Liangwangshan Nature
Reserve (Yunnan, China); Tuber borchii var. was obtained from West
Hills Forest Park of Kunming (Yunnan, China); Tuber himalayense
was obtained from Nujiang River (Yunnan, China); and Tuber
sinense was purchased from Mianyang Institute of Edible Fungi
(Sichuan, China). All fruiting-bodies were stored in the refrigerator
at −20 ◦ C.
Trufﬂe mycelia were cultured in our laboratory. The strains of
Tuber melanosporum, T. sinense and T. indicum were purchased from
Mianyang Institute of Edible Fungi (Sichuan, China). The strain of T.
aestivum was kindly provided by Huazhong Agricultural University
(Hubei, China). The details of the culture conditions and procedure
have been previously described [1].
P. Liu et al. / Analytica Chimica Acta 687 (2011) 159–167
2.3. Sample preparation by DSPE procedure
The freeze-dried samples were pulverized and then subjected
to pass through a 250-␮m stainless sieve. 100.0 mg sample powder was accurately weighed, and then extracted with 5 mL water by
ultrasonic cell disruption apparatus (J92-II, Ningbo Scientz Biotechnology Instrument, China) for 120 s (400 W power, 2 s ultrasound,
2 s intermittent, and 60 times). After centrifugation (3K15, Sigma,
Germany, 13,000 rpm and 10 min), 1 mL supernatant was diluted by
60% methanol aqueous to obtain the ﬁnal volume at 2 mL and ﬁnal
methanol concentration at 30%. Then 1 mL mixture was transferred
to a centrifuge tube, in which 0.05 g (dry weight) macroporous resin
D3520 was added as the DSPE sorbent. After vortexed for 60 s (XW80A, Shanghai Huxi Analysis Instrument, China) and centrifuged
at 13,000 rpm for 10 min, 0.5 mL supernatant was transferred to
another test tube and then evaporated to dry under a gentle nitrogen stream, and ﬁnally reconstituted with 0.5 mL of mobile phase.
The supernatant was ﬁltrated through 0.45 ␮m ﬁlter, 20.0 ␮L was
directly injected into the LC-UV or LC–MS for analysis.
161
based on the protein removal ratio (Pr). Vortex time was ﬁxed at
180 s during the experiment of optimizing the addition amount of
macroporous rein, and the addition amount of macroporous rein
ranged from 0.025 to 0.2 g. The addition amount of macroporous
resin was ﬁxed at 0.05 g during the experiment of optimizing vortex
time, and vortex time varied from 30 s to 180 s.
2.4.2. HPLC conditions for DSPE optimization
The analysis of un-adsorption ratio for nucleosides and nucleobases were preformed on Waters 600E system (Waters, USA),
equipped with an on-line degasser, a Waters 2487 UV detector. The column used for separation was Agela Venusil ASB C18
(250 mm × 4.6 mm i.d., 5 ␮m) ﬁtted with a C18 guard column
(Agela, Beijing, China). The optimized mobile phase was consisted
of methanol (A) and 5 mM ammonium acetate aqueous solution (B),
whose pH was adjusted to 2.0 by formic acid, and the separation
was carried on by the isocratic elution (A:B = 1:99, v/v). The column
oven temperature was maintained at 30 ◦ C and the ﬂow rate was
1.0 mL min−1 . The detection wave was ﬁxed at 254 nm.
2.4. DSPE procedure optimization
2.5. LC–MS analysis
2.4.1. Parameter optimization
A series of parameters, including macroporous resin type,
macroporous resin amount, methanol concentration, and vortex
time were optimized in the DSPE method development process. All
the experiments were performed in triplicate.
The selection of macroporous resin type (i.e., D-3520, D-4006,
D-101, AB-8, and D-312) and methanol concentration (e.g., 20–50%)
was based on the integrated consideration for protein removal
ratio and un-adsorption ratio of nucleosides and nucleobases.
During the experiment for investigating macroporous resin type
and methanol concentration, the other two parameters, macroporous resin amount and vortex time were ﬁxed at 0.2 g and
180 s, respectively. The experiments were performed as follows: 4
copies of 1 mL crude extract solution (concentration at 20 mg mL−1 :
100 mg sample extracted by 5 mL aqueous solution, and containing 20 ␮g mL−1 thymine as an internal standard) were diluted by
methanol aqueous with different concentrations (40%, 60%, 80% and
100%, respectively) to make the ﬁnal volume at 2 mL and the ﬁnal
methanol concentrations at 20%, 30%, 40%, and 50%, respectively.
After the process of vortex (180 s) and centrifugation (13,000 rpm
for 10 min), all the aforementioned sample solutions were treated
by the following procedure in parallel: ﬁrstly, 0.3 mL supernatant
was taken and treated according to Bradford method [38], and the
relative amount of protein was deﬁned as the value of “Absorbance”
at 595 nm (recorded as “Ac ”); and then another 0.3 mL supernatant
was directly injected to HPLC for analyzing the amount of nucleoside and nucleobases (conditions in detail were seen in Section
2.4.2), and the peak area for each analyte was recorded as “PAc ”.
Secondly, the left 1 mL supernatant was transferred into a new
centrifuge tube, in which 0.2 g (dry weight) macroporous resin
was added. After vortexed 180 s and centrifuged (13,000 rpm for
10 min), the relative amount of protein, as well as nucleosides and
nucleobases in the supernatant, were determined again according
to the aforementioned process and recorded as “As ”, and “PAs ”,
respectively. Thus the protein removal ratio (Pr) for a macroporous resin at a speciﬁc methanol concentration was calculated
by: Pr = 100% × (Ac − As )/Ac , and the un-adsorption ratio for each
analyte (UAr) was calculated by: UAr = 100% × PAs /PAc . The macroporous reins and methanol concentration in solutions, which lead
to high protein removal ratio and un-adsorption ratio of nucleosides and nucleobases, was preferred.
After the type of macroporous resin and concentration of
methanol were ﬁxed, the other two signiﬁcant factors of the addition amount of macroporous resin and vortex time were optimized
Analysis for nucleosides and nucleobases were preformed on
LCMS-2010EV system (Shimadzu, Tokyo, Japan), equipped with a
DGU-20A3 on-line degasser, two LC-20AD solvent delivery pumps,
a CTO-20AC column oven, a SIL-20A autosampler, a SPD-M20A photodiode array detector, a single quadrupole mass spectrometer with
an electrospray ionization interface, and a LCMS solution workstation. The column used for separation was Shim-pack VP-ODS
(Shimadzu, 250 mm × 2.0 mm i.d., 5 ␮m) ﬁtted with a C18 guard
column (Shimadzu). The optimized mobile phase was consisted of
methanol (A) and 5 mM ammonium acetate aqueous solution (B),
whose pH was adjusted to 2.0 by formic acid, and the separation
was carried on by the isocratic elution (A:B = 1:99, v/v). The column oven temperature was maintained at 30 ◦ C and the ﬂow rate
was 0.2 mL min−1 . Peaks were detected and scanned with the wavelength range from 200 to 600 nm by photodiode array detector, and
the positive mode was adopted in MS detection.
For qualitative analysis, mass spectrometry was carried out
in the scan mode from m/z 50 to 350u. However, for quantitative analysis, the selected ion monitoring (SIM) mode was used,
and the [M+H]+ at m/z 136, 137, 245, 268, 284, 152, 269 and
127 was selected as the characteristic ion fragment for adenine,
hypoxanthine, uridine, adenosine, guanosine, guanine, inosine, and
thymine. Mass spectrometric detection conditions for both scan
and SIM mode were as follows: ion source temperature was 250 ◦ C.
Curved desolvation line (CDL) and heat block temperatures for the
analysis were set at 250 and 200 ◦ C, respectively. Probe voltage was
+4.5 kV. Detector voltage was 1.5 kV. CDL voltage was −20 V. Drying and nebulizer gases of nitrogen were set at 1.5 L min−1 with a
pressure of 0.02 MPa.
2.6. Method validation
In order to determine the linearity of investigated compounds,
a series of standard solutions at 7 different concentrations, including the internal standard (I.S., 10 ␮g mL−1 ), were treated by the
proposed method in Section 2.3, and then analyzed by LC–ESIMS under SIM mode to establish calibration curve. The calibration
curve was constructed by plotting the peak area ratio of individual
standard to I.S. versus the ratio of their corresponding concentrations. The limits of detection (LOD) and quantiﬁcation (LOQ) for
each analyte were determined at a signal-to-noise ratio (S/N) of
about 3 and 10, respectively.
The method precision and accuracy were evaluated by analyzing the mixed standards in three replicates for short-term (1 day)
162
P. Liu et al. / Analytica Chimica Acta 687 (2011) 159–167
Table 2
The molecular spectra information of the nucleoside and nucleobases in Tuber samples.
Analyte
Adenine
Hypoxanthine
Uridine
Adenosine
Guanine
Guanosine
Inosine
Thyminec
a
b
c
tR (min)
4.16
5.75
7.41
8.72
12.48
12.48
13.48
10.08
UV␭max
(nm)
[M+H]+
(m/z)
261
248
257
264
254
254
248
262
136
137
245
268
152
284
269
127
[M+Na]+
(m/z)
[2M+H]+
(m/z)
[2M+Na]+
(m/z)
159
267
273
295
Fruiting-body
a
T. sin.
√b
√
√
√
√
√
√
T. ind.
√
√
√
√
√
√
√
Fermentation mycelia
T. him.
√
√
√
√
√
√
√
T. aest.
√
√
√
√
√
√
√
T. borc.
√
√
√
√
√
T. mel.
√
√
√
√
√
√
√
T. sin.
√
√
√
√
√
√
√
T. ind.
√
√
√
√
√
√
√
T. aest.
√
√
√
√
√
√
√
T. sin. = T. sinense; T. ind. = T. indicum; T. him. = T. himalayense; T. aest. = T. aestivum; T. borc. = T. borchii var.; T. mel. = T. melanosporum.
√
means the compound was identiﬁed in the corresponding samples.
Thymine was not found in any Tuber samples, and was selected as the internal standard substance for the following experiments.
and long-term (3 days). The intra-day precision was deﬁned as relative standard deviation (RSD) calculated from three independent
assay in the same day, and the inter-day precision was the RSD
calculated from three independent assays in the separate days.
The accuracy was evaluated by the mean deviation between the
measured concentration and its spiked concentration.
The sample recovery was performed by adding known amount
of individual standards into an accurately weighed mixed sample mentioned above. The mixed sample was extracted and
analyzed using the method mentioned above. For each concentration, three replicate experiments with the whole analysis
process were performed. Recovery was calculated with the
following equation: recovery (%) = 100 × amount found/(original
amount + amount spiked).
2.7. Hierarchical clustering analysis
Hierarchical cluster analysis is a multivariate analysis technique,
which is used to sort samples into groups. In our study, the hierarchical clustering analysis (HCA) of samples was performed using
SPSS 15.0 software (Chicago, USA). The Between-groups linkage
cluster method, the squared Euclidean distance measure and zscores standardization [39] were used to establish clusters.
3. Results and discussion
3.1. Qualitative analysis for the nucleosides and nucleobases in
Tuber fermentation mycelia and fruiting-bodies
In order to identify the nucleosides and nucleobases in Tuber
samples, the standard substances (i.e., adenine, hypoxanthine, uridine, adenosine, guanine, guanosine, inosine, and thymine) and
Tuber samples extracting solution were analyzed by LC-UV/ESI-MS
in parallel. The mass and UV spectrum data for individual standard
substance was compared in Table 2. As the UV spectrum of each
standard substance was too similar with each other, they could
not be used to identify the nucleosides and nucleobases. Comparatively, the MS spectra obtained by scan mode provided much more
useful information for identiﬁcation, because the signals of [M+H]+ ,
as well as [2M+H]+ , [M+Na]+ , and [2M+Na]+ were speciﬁc to the
individual nucleoside and nucleobases. Therefore, the nucleosides
and nucleobases in Tuber samples were identiﬁed by comparing
the retention time and the on-line MS spectra information with
the standard substance.
Although the guanosine and guanine were co-eluted at the same
time (tR = 12.48 min) (Table 2), they could also be well distinguished
with each other by MS ion fragment [M+H]+ , because the m/z of
152 was speciﬁc to guanine, and the m/z of 284 was to guanosine.
Among the investigated nucleoside and nucleobases, thymine was
not found in any Tuber samples. Adenine, uridine, hypoxanthine,
Fig. 1. Effects of methanol concentration and macroporous resin type on the protein removal ratio for Tuber sample water extraction solution. Conditions: sample
concentration, 20 mg mL−1 crude extract; sample volume, 0.5 mL; the weight of
macroporous resin, 0.2 g; and vortex time, 180 s.
adenosine, guanosine, guanine and inosine were conﬁrmed to be
presented in all Tuber samples, except that adenine and uridine
were absent in the fruiting-body of T. borchii var. Therefore, these
seven compounds (i.e., adenine, uridine, hypoxanthine, adenosine,
guanosine, guanine and inosine) were selected as the target nucleosides and nucleobases, and thymine was selected as the internal
standard substance for the following experiments.
3.2. DSPE procedure optimization
With the purpose to remove the protein from the sample, the
DSPE procedure was adopted to selectively adsorb the protein,
and keep most of the investigated compounds remain in the sample. The signiﬁcances of key factors (e.g., the type and amount of
macroporous resin, methanol concentration, and vortex time) on
the performance of DSPE procedure were studied in detail.
3.2.1. Inﬂuences of macroporous resin and methanol
concentration on the protein removal ratio
Fig. 1 shows that macroporous resin D3520 has the best performance on protein removing, whose protein removal ratio was
around 100% regardless of methanol concentration. For the other
four kinds of macroporous resins, their protein removal ratios were
increased with the increase of methanol concentration within the
range of investigated. Therefore, macroporous resin D3520 was
the best choice from the viewpoint of protein removal ratio. The
purpose of DSPE was to selectively adsorb protein, and to reserve
analytes in solution as more as possible, so the un-adsorption effect
of macroporous resins on analytes should be further studied.
P. Liu et al. / Analytica Chimica Acta 687 (2011) 159–167
163
Fig. 2. Effects of methanol concentration and macroporous resin type on the un-adsorption ratio for nucleosides and nucleobases. (A) At the methanol concentration of 20%,
(B) at the methanol concentration of 30%, (C) at the methanol concentration of 40%, (D) at the methanol concentration of 50%. Conditions: sample concentration, 20 mg mL−1
crude extract; sample volume, 0.5 mL; the weight of macroporous resin, 0.2 g; and vortex time, 180 s.
3.2.2. Inﬂuences of macroporous resin and methanol
concentration on the un-adsorption ratio for the target
compounds
In order to make sure most of the nucleosides and nucleobases were remained in solvent instead of being adsorbed by
sorbent, effects of macroporous resin and methanol concentration
on the un-adsorption ratio of analytes were studied. As shown
in Fig. 2, all ﬁve macroporous resins had adsorption effect on
nucleosides and nucleobases, as well as the internal standard substance (I.S., thymine). The un-adsorption ratios for analytes were
rising accompanying with the increase of methanol concentration. Among the ﬁve macroporous resins, D4006 and D312 had
the weakest adsorption capabilities on nucleosides and nucleobases, thus their un-adsorption ratios for analytes and I.S. were
higher than the others’ at the same concentration of methanol.
However, macroporous resin D4006 and D312 were not satisﬁed
to our purpose for protein remove because of their poor protein adsorption capabilities (Fig. 1). Although the analytes and
I.S. were partly adsorbed by macroporous resins D101, AB-8, and
D3520, the un-adsorption ratios of at least 60% could still achieve
when the methanol concentration was not below 30%. Such a
small faction of analytes and I.S. adsorbed by macroporous resins
would not affect the performance of the quantitive assay. Comprehensive consideration of protein removing ratio and analytes
un-adsorption ratio, macroporous resin D3520 is the optimal DSPE
sorbent.
Although the un-adsorption ratios for analytes and I.S. could
achieve at least 60% when the methanol concentration was not
below 30%, the un-adsorption ratio for individual analyte and I.S.
was more equal with the methanol concentration at 30%, which
was more beneﬁt to good accuracy for the applied inter-standard
method. Furthermore, the DSPE procedure with 30% methanol was
much clearer and colorless than any other solutions with higher
methanol concentrations. This phenomenon demonstrated the relatively high adsorption capability of macroporous resin in low
methanol concentration, not only for protein, but also the pigments
and some other nonpolarity impurity. So, 30% of methanol was
selected for DSPE procedure.
3.2.3. Inﬂuences of macroporous resin amount and vortex time
on the protein removal ratio
The macroporous resin amount and the vortex time were also
the important factors for DSPE procedure which signiﬁcantly affecting the protein removal ratio. As shown in Fig. 3A, the protein
removal ratio for macroporous resin D3520 increased with the
increase of the addition macroporous resins weight from 0 to 0.20 g,
and then achieved a plateau when the weight exceeded 0.05 g.
This result indicated that 0.05 g of macroporous resin D3520 was
enough to remove about 95% of protein from the 0.5 mL 20 mg mL−1
crude extract. The inﬂuence of vortex time on the protein removal
ratio was shown in Fig. 3B. The protein removal ratio increased
from 0 to about 95% when vortex time increased from 0 to 60 s, and
then maintained constant when the vortex time exceeded 60 s. This
result indicated that it takes 60 s for macroporous resin D3520 to
achieve the protein adsorption balance. So, the optimal dispersive
sorbent amount and vortex time were selected as 0.05 g and 60 s
for the 0.5 mL 20 mg mL−1 crude extract, respectively.
To conclude, for 0.5 mL 20 mg mL−1 Tuber crude extract, 0.05 g
macroporous resin D3520 was chosen as the dispersive sorbent; the
methanol concentration in solution was adjusted to 30%; and vortex
time was ﬁxed at 60 s. Within this DSPE procedure, protein removal
ratio was about 95%, the un-adsorption ratio for individual analyte
and I.S. were keep at the same level (above 70%). And, the phenomenon, such as HPLC system pressure rising, and the distorted
peaks, did not happen again.
3.3. Method validation
Internal standard method was adopted to counteract the partial
loss of analytes during DSPE procedure. Thymine was selected as
the internal standard substance because it was not found in any
Tuber samples.
As shown in Table 3, the correlation coefﬁcients (R2 ) ranged from
0.9993 to 0.9999. The high correlation coefﬁcient values indicated
good correlations between investigated compounds concentrations and their peak area ratios. The different ratios of LOQ and LOD
may be derived from the difference of MS response to the analytes,
164
P. Liu et al. / Analytica Chimica Acta 687 (2011) 159–167
Fig. 3. Effect of macroporous resin D3520 weight (A) and voxter time (B) on the protein removal ratio. Conditions for (A): sample concentration, 20 mg mL−1 crude extract;
sample volume, 0.5 mL; methanol concentration, 30%; and vortex time, 180 s. Conditions for (B): sample concentration, 20 mg mL−1 crude extract; sample volume, 0.5 mL;
methanol concentration, 30%; and the weight of macroporous resin, 0.05 g.
Table 3
Linear regression data and validation of the developed determination method for nucleosides and nucleobases (n = 3).
Analyte
Linear regression data
Linear range (␮g mL−1 )
Adenine
Hypoxanthine
Uridine
Adenosine
Guanine
Guanosine
Inosine
a
0.5–100.1
0.5–109.3
0.5–98.9
0.6–111.2
0.4–84.5
0.5–98.8
0.5–90.1
Regression equation
R2 a
y = 0.1772 × −0.0482
y = 0.1591 × −0.0723
y = 0.0759 × −0.0212
y = 0.1164 × −0.0507
y = 0.0710 × −0.0211
y = 0.0868 × −0.0352
y = 0.0943 × −0.0374
0.9998
0.9993
0.9998
0.9995
0.9997
0.9996
0.9999
LOD (␮g mL−1 )
LOQ (␮g mL−1 )
0.4
0.4
0.5
0.1
0.3
0.2
0.3
1.5
1.0
1.6
0.3
1.0
0.7
0.8
R2 , squares of correlation coefﬁcients for the standard curves.
and the LODs were between 0.1 and 0.5 ␮g mL−1 , and the LOQs
were between 0.3 and 1.6 ␮g mL−1 , respectively. These results indicated that LC–MS was sensitive for the qualitative and quantitative
determination of the analytes.
As shown in Table 4, the relative standard deviation (RSD) was
not higher than 8.8%, and the highest analytical error was 6.5% and
9.4% for short-term and long-term, respectively.
Table 5 shows that the recoveries of the 7 investigated compounds ranged from 95.0% to 111.0% with RSD between 0.2 and
10.8%, respectively, which indicated good performance of the DSPE
method for quantitative assay.
3.4. Determination and discrimination of the target nucleosides
and nucleobases in Tuber fruiting-bodies and fermentation
mycelia
The contents of the target 7 nucleosides and nucleobases in
Tuber fruiting-bodies and fermentation mycelia were determined
by LC–ESI-MS. The SIM technique in mass spectrum was adopted
because it can distinguish the co-eluted compounds, which could
not be identiﬁed by UV chromatogram. The SIM chromatograms
of various Tuber samples were shown in Fig. 4, and the determination results were summarized in Table 6. Except adenine and
Table 4
Intra- and inter-day variability for the assay of nucleosides and nucleobases (n = 3).
Analyte
Conc. (␮g mL−1 )
Intra-day
Found (␮g mL
Inter-day
−1
)
R.S.D.(%)
Error (%)
Found (␮g mL−1 )
R.S.D. (%)a
Error (%)b
a
b
Adenine
20.4
4.4
19.9
4.3
3.1
3.9
2.9
3.6
19.4
4.1
4.0
4.8
4.9
5.7
Hypoxanthine
21.6
4.2
21.0
4.1
2.7
3.3
2.7
3.0
20.6
4.2
1.6
2.2
4.6
1.5
Uridine
20.8
3.8
19.9
3.8
1.1
6.5
4.4
4.7
19.3
4.0
1.5
6.1
7.4
5.0
Adenosine
20.6
4.1
20.2
4.1
3.0
3.5
2.6
2.9
20.1
4.0
4.3
2.2
3.0
1.9
Guanine
20.2
3.9
20.0
4.0
1.3
2.0
1.2
3.5
20.8
3.8
1.7
5.2
3.1
4.8
Guanosine
21.0
3.8
21.6
4.0
8.8
2.2
5.9
6.5
20.7
4.2
6.0
2.6
4.6
9.4
Inosine
19.8
4.1
20.2
4.2
3.9
2.8
2.6
2.7
20.4
4.3
8.2
2.7
6.9
3.9
a
b
R.S.D. (%) =
100 × S.D./mean.
Error (%) =
ABS(spiked concentration − measured concentration)/3 spiked concentration × 100.
P. Liu et al. / Analytica Chimica Acta 687 (2011) 159–167
165
Table 5
Recoveries for the assay of nucleosides and nucleobases in Tuber samples (n = 3).
R.S.D. (%)b
Original (␮g)
Spiked (␮g)
Adenine
11.99
22.54
28.18
33.81
36.82
42.11
48.54
106.6
104.8
106.0
1.3
1.0
1.2
1.96
1.98
2.47
2.96
4.12
4.21
4.68
104.7
95.0
95.0
6.3
1.6
1.3
Uridine
29.98
11.02
13.77
16.52
40.62
45.14
50.46
99.1
103.2
108.5
1.0
5.8
2.3
Adenosine
45.69
42.31
52.89
63.46
86.46
98.07
104.77
98.3
99.5
96.0
0.7
0.6
1.6
Guanine
61.21
44.46
55.57
66.69
112.15
122.58
134.30
106.1
105.0
105.0
1.4
0.6
0.6
Guanosine
65.86
41.81
52.27
62.72
115.59
128.49
136.17
107.3
108.8
105.9
1.6
3.9
0.2
1.17
3.54
4.43
5.31
5.10
6.21
6.71
108.1
111.0
103.5
10.8
8.4
4.5
Hypoxanthine
Inosine
a
b
Found (␮g)
Recovery (%)a
Analyte
Recovery (%) = 100 × amount found/(original amount + amount spiked).
R.S.D. (%) = 100 × S.D./mean.
uridine were not detected in the fruiting-bodies of T. borchii var.,
the 7 target nucleosides and nucleobases were detected in all Tuber
samples. Despite their contents varied with the species, adenosine, guanine, and guanosine were the predominant nucleosides
in all fermentation mycelia. However, the nucleosides formulation for each species was quite different, so it is difﬁcult to ﬁnd
the common major nucleosides in the fruiting-bodies. Typically,
the highest contents of adenine and uridine were determined in
T. sinense fruiting-bodies, while they were not found in T. borchii
var. fruiting-bodies. The highest concentration of adenosine was
detected in the fruiting-bodies of T. indicum and T. himalayense,
while its concentration was extreme low in the fruiting-bodies of
T. aestivum and T. borchii var. More interestingly, total amount of
nucleosides and nucleobases existed signiﬁcant difference between
the fruiting-bodies and fermentation mycelia. As shown in Table 6,
the total amount of nucleosides and nucleobases in fermentation mycelia ranged from 4881.5 to 12,592.9 ␮g g−1 , which was
about 2–25 times higher than those in the fruiting-bodies (i.e.,
498.1–2274.1 ␮g g−1 ). The contents of the three key nucleosides
(i.e., adenosine, guanine, and guanosine) in fermentation mycelia
Fig. 4. The typical chromatograms in SIM mode. (A) The mixed standards, (B) the fruiting-bodies of T. sinense, (C) the fruiting-bodies of T. himalayense, (D) the fruiting-bodies
of T. aestivum, (E) the fruiting-bodies of T. borchii var., and (F) the fermentation mycelia of T. melanosporum. Symbols for the analytes: U, uridine; H, hypoxanthine; Ad,
adenosine; G, guanine; Gu, guanosine; I, inosine; and IS, thymine. IS concentration was 10 ␮g mL−1 in both mixed standards and samples.
166
P. Liu et al. / Analytica Chimica Acta 687 (2011) 159–167
Table 6
The contents (␮g g−1 ) of nucleosides and nucleobases in Tuber fruiting-bodies and fermentation mycelia (n = 3).
Analyte
Fruiting-bodies
a
T. sin.
Adenine
Hypoxanthine
Uridine
Adenosine
Guanine
Guanosine
Inosine
Total
a
b
c
Fermentation mycelia
T. ind.
318.3 ± 17.9b 172.3
161.2 ± 2.1
37.0
274.5 ± 8.4
309.2
241.8 ± 10.9
415.6
185.8 ± 9.8
376.0
238.6 ± 7.2
404.8
85.9 ± 3.4
77.2
1506.1 ± 64.2 1792.1
T. him.
±
±
±
±
±
±
±
±
3.3
2.2
5.7
10.0
12.7
11.7
5.3
79.2
346.3
31.8
260.1
542.1
481.2
544.5
68.1
2274.1
T. aest.
±
±
±
±
±
±
±
±
7.1
1.6
6.2
9.6
9.5
2.6
2.1
106.5
265.0
126.2
137.3
50.1
200.0
226.2
215.5
1220.3
±
±
±
±
±
±
±
±
12.1
6.7
2.5
2.4
7.1
3.6
10.1
71.9
T. borc.
T. mel.
n.d.c
175.9 ± 2.3
n.d.
58.6 ± 1.3
49.1 ± 3.6
95.7 ± 5.5
118.8 ± 3.8
498.1 ± 24.7
357.5
85.5
1671.9
2178.4
2629.6
2884.4
42.7
9850.0
T. sin.
±
±
±
±
±
±
±
±
22.3
2.7
25.4
95.1
76.0
83.4
1.9
365.9
475.8
96.7
1285.2
2762.8
3821.1
4082.2
69.1
12592.9
T. ind.
±
±
±
±
±
±
±
±
17.7
5.4
49.2
50.6
105.9
195.5
1.7
239.3
532.2
162.0
1265.7
2182.7
2576.6
2812.1
119.8
9651.1
T. aest.
±
±
±
±
±
±
±
±
20.7
4.3
51.9
62.2
65.6
51.6
2.09
202.4
320.0
82.0
1110.7
1322.8
966.1
1036.7
43.2
4881.5
±
±
±
±
±
±
±
±
19.3
4.03
37.3
73.7
53.5
45.1
1.2
146.4
T. sin. = T. sinense; T. ind. = T. indicum; T. him. = T. himalayense; T. aest. = T. aestivum; T. borc. = T. borchii var.; T. mel. = T. melanosporum.
Each value is expressed as mean ± SD (n = 3).
n.d., not detected.
Fig. 5. Dendrogram of clustering of Tuber samples based on the contents of nucleosides and nucleobases. Options set were: method, Between-groups linkage; measure of
distance, squared Euclidean distance; standardization of variables, z-scores. (a FM = fermentation mycelia; b FB = fruiting-bodies).
were 2–52 times higher than those in the fruiting-bodies. All these
results indicated that Tuber fermentation mycelia contained much
more the target nucleosides than the fruiting-bodies. The formulation of the target 7 nucleosides and nucleobases almost kept
constant in fermentation mycelia, which were varied signiﬁcantly
with the species in Tuber fruiting-bodies.
Based on the analytes contents, the relationship of Tuber samples was further analyzed by hierarchical cluster analysis (HCA).
Fig. 5 clearly shows that total 9 Tuber samples were divided into
5 clusters. Cluster-I was formed by the fermentation mycelia of T.
sinense, T. indicum, T. melanosporum, which were cultured under the
same condition. We also observed that the fermentation mycelia of
T. aestivum closed to the other 3 fermentation mycelia although T.
aestivum fermentation mycelia did not belong to Cluster-I. From
the viewpoint of nucleosides and nucleobases, Tuber fermentation
mycelia were quite similar when mycelia were cultured under the
same condition. Similar phenomenon was also observed in our
previous studies about the volatile organic compounds in Tuber
mycelia [8]. Interestingly, the fruiting-bodies of T. indicum and T.
himalayense belonged to the same cluster, which closed to Tuber
fermentation mycelia. This result indicated that Tuber fermentation
mycelia were very similar with the fruiting-bodies of T. indicum and
T. himalayense from the viewpoint of nucleosides and nucleobases.
The fruiting-bodies of T. borchii var., T. aestivum and T. sinense did
not belong to one cluster, and shows pretty long distance with the
fruiting-bodies of T. indicum and T. himalayense. The signiﬁcant distinction mainly resulted from the heterogeneity of the investigated
fruiting-bodies (e.g., age, origin, soil type).
To conclude, the formulation similarity of nucleosides and
nucleobases existed between the fermentation mycelia of various Tuber species cultured under the same condition, and between
Tuber fermentation mycelia and the fruiting-bodies of T. indicum
and T. himalayense. From the viewpoint of nucleosides and
nucleobases, this work partly conﬁrmed the rationality of Tuber
fermentation mycelia as the alternative resource for trufﬂe fruitingbodies.
4. Conclusion
Based on the comparison of their retention time and the charactering ion fragment to those of authentic standards, 7 nucleosides
and nucleobases (i.e., adenine, hypoxanthine, uridine, adenosine,
guanine, guanosine, and inosine) in Tuber fruiting-bodies and
fermentation mycelia were identiﬁed for the ﬁrst time. As a
clean-up procedure, dispersive solid phase extraction (DSPE) was
ﬁrstly introduced to remove the protein from the extract solution,
and D3520 macroporous resin was chosen as the DSPE sorbent
because of its high removal capability on protein interferences,
but low adsorption rate on analytes. Additionally, key parameters on DSPE procedure (i.e., macroporous resin type, macroporous
resin amount, methanol concentration, and vortex time) were optimized, and the protein removal efﬁcacy could achieve about 95%
after the process optimization. Besides, the column blinding problems, such as HPLC system pressure rising, and peaks distorting,
were solved successfully. The mode of selective ion monitoring
(SIM) was applied in the quantitative assay method to distinguish
the co-eluting guanine and guanosine. In the method validation
test, the R.S.D. of precisions was not higher than 8.8%; he highest
analytical error was 6.5% and 9.4% for short-term and long-term,
respectively; the recoveries ranged from 95.0% to 111.0%; the
limits of detection and quantiﬁcation for analytes were in the
order of 0.1–0.5 ␮g mL−1 and 0.3–1.6 ␮g mL−1 , respectively; and
all calibration curves showed good linearity (R2 > 0.999) within
the range as investigated. All these results demonstrated the DSPE
combined with LC–MS method was accurate, precise, replicable,
and sensitive. The 7 target nucleosides and nucleobases were
P. Liu et al. / Analytica Chimica Acta 687 (2011) 159–167
detected in all Tuber samples except the fruiting-body of T. borchii
var. Adenosine, guanine, and guanosine were the predominant
nucleosides in all fermentation mycelia, while the major nucleosides and nucleobases in the fruiting-bodies were signiﬁcantly
varied among different species. The total amount of nucleosides
and nucleobases in the fermentation mycelia ranged from 4881.5
to 12,592.9 ␮g g−1 , which was about 2–25 times higher than
those (i.e., 498.1–2274.1 ␮g g−1 ) in the fruiting-bodies. Hierarchical cluster analysis (HCA) further demonstrated the similarity of
nucleosides and nucleobases formulation among various Tuber fermentation mycelia, as well as between the Tuber fermentation
mycelia and the fruiting-bodies of T. indicum and T. himalayense.
This work partly conﬁrms the rationality of Tuber fermentation
mycelia as the alternative resource for trufﬂe fruiting-bodies.
Acknowledgement
Financial support from the National Natural Science Foundation
of China (NSFC, Project Nos. 20706012 and 20976038), National
Basic Research Program of China (973 Program, 2007CB714306),
the Key Project of Chinese Ministry of Education (Project No.
210132), Hubei Provincial Natural Science Foundation for Innovative Research Team (Project No. 2008CDA002), Discipline Leader
Project of Wuhan Municipality (Project No. 200951830553), Scientiﬁc Research Key Project of Hubei Provincial Department of
Education (Project No. Z20101401), the Open Project Program for
Key Laboratory of Fermentation Engineering (Ministry of Education), and the Open Funding Project of the National Key Laboratory
of Biochemical Engineering (2010KF-06) are gratefully acknowledged. Ya-Jie Tang also thanks the Chutian Scholar Program (Hubei
Provincial Department of Education, China) (2006).
References
[1] Y.J. Tang, L.L. Zhu, D.S. Li, Z.Y. Mi, H.M. Li, Process Biochem. 43 (2008) 576–586.
[2] R.S. Liu, D.S. Li, H.M. Li, Y.J. Tang, Process Biochem. 43 (2008) 868–876.
[3] Y.J. Tang, L.L. Zhu, R.S. Liu, H.M. Li, D.S. Li, Z.Y. Mi, Bioresour. Technol. 99 (2008)
7606–7615.
[4] R.S. Liu, Y.J. Tang, Appl. Microbiol. Biotechnol. 87 (2010) 1689–1697.
[5] Q.N. Liu, R.S. Liu, Y.H. Wang, Z.Y. Mi, D.S. Li, J.J. Zhong, Y.J. Tang, Bioresour.
Technol. 100 (2009) 3644–3649.
[6] G. Wang, Y.Y. Li, D.S. Li, Y.J. Tang, J. Chromatogr. B 870 (2008) 209–215.
167
[7] Y.Y. Li, G. Wang, Y.J. Tang, Phytochemistry, revised for publication.
[8] Y.J. Tang, G. Wang, Y.Y. Li, J.J. Zhong, Anal. Chim. Acta 647 (2009) 40–45.
˜
˜
[9] P. Díaz, E. Ibánez,
F.J. Senoráns,
G. Reglero, J. Chromatogr. A 1017 (2003)
207–214.
[10] R. Splivallo, S. Bossi, M. Maffei, P. Bonfante, Phytochemistry 68 (2007)
2584–2598.
[11] J.M. Gao, A.L. Zhang, H. Chen, J.K. Liu, Chem. Phys. Lipids 131 (2004) 205–213.
[12] J.M. Gao, A.L. Zhang, C.Y. Wang, X.H. Wang, J.K. Liu, Chin. Chem. Lett. 13 (2002)
325–326.
[13] M.A. Murcia, M. Martínez-Tomé, A. Vera, A. Morte, A. Gutierrez, M. Honrubia,
A.M. Jiménez, J. Sci. Food Agric. 83 (2003) 535–541.
[14] E. Harki, D. Bouya, R. Dargent, Food Chem. 99 (2006) 394–400.
[15] J.M. Gao, L. Hu, J.K. Liu, Steroids 66 (2001) 771–775.
[16] M. Ballarin, M. Herrera-Marschitz, M. Cassa, U. Ungerstedt, Neurosci. Lett. 83
(1987) 338–344.
[17] L.I. Benowitz, D.E. Goldberg, N. Irwin, Prog. Brain Res. 137 (2002) 389–399.
[18] R. Uauy, G. Stringel, R. Thomas, R. Quan, J. Pediatr. Gastroenterol. Nutr. 10 (1990)
497–503.
[19] C. Morimoto, A.D. Steinberg, S.F. Schlossman, Y. Borel, J. Clin. Invest. 71 (1983)
1402–1409.
[20] H.Y. Cheung, C.W. Ng, D.J. Hood, J. Chromatogr. A 911 (2001) 119–126.
[21] Pharmacopoeia Commission of PRC (Ed.), Pharmacopoeia of the People’s
Republic of China, vol. I, Chemical Industry Press, Beijing, PR China, 2005, pp.
75.
[22] S.P. Li, P. Li, C.M. Lai, Y.X. Gong, K.K.W. Kan, T.T.X. Dong, K.W.K. Tsim, Y.T. Wang,
J. Chromatogr. A 1036 (2004) 239–243.
[23] J.L. Gao, K.S.Y. Leung, Y.T. Wang, C.M. Lai, S.P. Li, L.F. Hu, G.H. Lu, Z.H. Jiang, Z.L.
Yu, J. Pharm. Biomed. Anal. 44 (2007) 807–811.
[24] Y.X. Gong, S.P. Li, P. Li, J.J. Liu, Y.T. Wang, J. Chromatogr. A 1055 (2004) 215–221.
[25] X. Cahours, H. Dessans, P. Morin, M. Dreux, L. Agrofoglio, J. Chromatogr. A 895
(2000) 101–109.
[26] C. Polson, P. Sarkar, B. Incledon, V. Raguvaran, R. Grant, J. Chromatogr. B 785
(2003) 263–275.
[27] C.K. Fagerquist, A.R. Lightﬁeld, S.J. Lehotay, Anal. Chem. 77 (2005) 1473–1482.
[28] B. Streel, Ph. Hubert, A. Ceccato, J. Chromatogr. B 742 (2000) 391–400.
[29] F. Bruno, R. Curini, A. Di Corcia, M. Nazzari, R. Samperi, J. Agric. Food Chem. 49
(2001) 3463–3470.
[30] M. Walles, J. Borlak, K. Levsen, Anal. Bioanal. Chem. 374 (2002) 1179–1186.
ˇ
[31] M. Anastassiades, S.J. Lehotay, D. Stajnbaher,
F.J Schenck, J. AOAC Int. 86 (2003)
412–431.
[32] T. Dagnac, M. Garcia-Chao, P. Pulleiro, C. Garcia-Jares, M. Llompart, J. Chromatogr. A 1216 (2009) 3702–3709.
[33] M.C. Alcudia-León, R. Lucena, S. Cárdenas, M. Valcárcel, Anal. Chem. 81 (2009)
1184–1190.
[34] A. Hercegová, M. Dömötörová, E. Matisová, J. Chromatogr. A 1153 (2007) 54–73.
[35] G.T. Jia, X.Y. Lu, J. Chromatogr. A 1193 (2008) 136–141.
[36] Y.J. Fu, Y.G. Zu, S.M. Li, R. Sun, T. Efferth, W. Liu, S.G. Jiang, H. Luo, Y. Wang, J.
Chromatogr. A 1177 (2008) 77–86.
[37] Y.J. Fu, Y.G. Zu, W. Liu, C.L. Hou, L.Y. Chen, S.M. Li, X.G. Shi, M.H. Tong, J. Chromatogr. A 1139 (2007) 206–213.
[38] M.M. Bradford, Anal. Biochem. 72 (1976) 248–254.
[39] M. Beyer, R. Hahn, S. Peschel, M. Harz, M. Knoche, Sci. Hortic. 96 (2002) 139–150.