Certification of a new selenized yeast reference material (SELM-1) for methionine, selenomethinone and total selenium content and its use in an intercomparison exercise for quantifying these analytes.

Abstract

A new selenized yeast reference material (SELM-1) produced by the Institute for National Measurement Standards, National Research Council of Canada (INMS, NRC) certified for total selenium (2,059+/-64 mg kg(-1)), methionine (Met, 5,758+/-277 mg kg(-1)) and selenomethionine (SeMet, 3,431+/-157 mg kg(-1)) content is described. The +/-value represents an expanded uncertainty with a coverage factor of 2. SeMet and Met amount contents were established following a methanesulfonic acid digestion of the yeast using GC-MS and LC-MS quantitation. Isotope dilution (ID) calibration was used for both compounds, using 13C-labelled SeMet and Met. Total Se was determined after complete microwave acid digestion based on ID ICP-MS using a 82Se spike or ICP-OES spectrometry using external calibration. An international intercomparison exercise was piloted by NRC to assess the state-of-the-art of measurement of selenomethione in SELM-1. Determination of total Se and methionine was also attempted. Seven laboratories submitted results (2 National Metrology Institutes (NMIs) and 5 university/government laboratories). For SeMet, ten independent mean values were generated. Various acid digestion and enzymatic procedures followed by LC ICP-MS, LC AFS or GC-MS quantitation were used. Four values were based on species-specific ID calibration, one on non-species-specific ID with the remainder using standard addition (SA) or external calibration (EC). For total selenium, laboratories employed various acid digestion procedures followed by ICP-MS, AFS or GC-MS quantitation. Four laboratories employed ID calibration, the remaining used SA or EC. A total of seven independent results were submitted. Results for methionine were reported by only three laboratories, all of which used various acid digestion protocols combined with determination by GC-MS and LC UV. The majority of participants submitted values within the certified range for SeMet and total Se, whereas the intercomparison was judged unsuccessful for Met because only two external laboratories provided values, both of which were outside the certified range.

Full text

Page 1

Anal Bioanal Chem (2006) 385: 168–180
DOI 10.1007/s00216-006-0338-0
ORIGINAL PAPER
Zoltán Mester.Scott Willie.Lu Yang.Ralph Sturgeon.Joseph A. Caruso.
María Luisa Fernández.Peter Fodor.Robert J. Goldschmidt.
Heidi Goenaga-Infante.Ryszard Lobinski.Paulette Maxwell.Shona McSheehy.
Aleksandra Polatajko.Baki B. M. Sadi.Alfredo Sanz-Medel.Christine Scriver.
Joanna Szpunar.Raimund Wahlen.Wayne Wolf
Certification of a new selenized yeast reference material
(SELM-1) for methionine, selenomethinone and total selenium
content and its use in an intercomparison exercise for quantifying
these analytes
Received: 5 December 2005 / Revised: 24 January 2006 / Accepted: 26 January 2006 / Published online: 5 April 2006
# Springer-Verlag 2006
Abstract Anewselenizedyeastreferencematerial(SELM-1)
produced by the Institute for National Measurement
Standards, National Research Council of Canada (INMS,
NRC) certified for total selenium (2,059±64 mg kg−1),
methionine (Met, 5,758±277 mg kg−1) and selenomethio-
nine (SeMet, 3,431±157 mg kg−1) content is described.
The ±value represents an expanded uncertainty with a
coverage factor of 2. SeMet and Met amount contents were
established following a methanesulfonic acid digestion of
the yeast using GC-MS and LC-MS quantitation. Isotope
dilution (ID) calibration was used for both compounds,
using
determined after complete microwave acid digestion
based on ID ICP-MS using a
spectrometry using external calibration. An international
intercomparison exercise was piloted by NRC to assess the
state-of-the-art of measurement of selenomethione in
SELM-1. Determination of total Se and methionine was
also attempted. Seven laboratories submitted results (2
National Metrology Institutes (NMIs) and 5 university/
government laboratories). For SeMet, ten independent
mean values were generated. Various acid digestion and
enzymatic procedures followed by LC ICP-MS, LC AFS or
GC-MS quantitation were used. Four values were based on
species-specific ID calibration, one on non-species-specific
ID with the remainder using standard addition (SA) or
external calibration (EC). For total selenium, laboratories
employed various acid digestion procedures followed by
ICP-MS, AFS or GC-MS quantitation. Four laboratories
employed ID calibration, the remaining used SA or EC. A
total of seven independent results were submitted. Results
for methionine were reported by only three laboratories, all
of which used various acid digestion protocols combined
with determination by GC-MS and LC UV. The majority of
participants submitted values within the certified range for
13C-labelled SeMet and Met. Total Se was
82Se spike or ICP-OES
Z. Mester (*).S. Willie.L. Yang.R. Sturgeon.P. Maxwell.
S. McSheehy.C. Scriver
Institute for National Measurement Standards,
National Research Council of Canada,
Ottawa, ON K1A 0R6, Canada
e-mail: Zoltan.mester@nrc.ca
Fax: +1-613-9932451
J. A. Caruso.B. B. M. Sadi
University of Cincinnati,
Department of Chemistry,
Cincinnati, OH 45221-0172, USA
M. L. Fernández.A. Sanz-Medel
Department of Physical and Analytical Chemistry,
University of Oviedo,
33006 Oviedo, Spain
P. Fodor
Department of Applied Chemistry,
Corvinus University of Budapest,
Villányi út 29-33, 1118 Budapest, Hungary
R. J. Goldschmidt.W. Wolf
Food Composition Laboratory,
BHNRC, ARS, USDA,
Beltsville, MD 20705, USA
H. Goenaga-Infante.R. Wahlen
Laboratory of the Government Chemist,
Teddington, Middlesex, TW11 0LY, UK
R. Lobinski.A. Polatajko.J. Szpunar
Group of Bio-Inorganic Analytical Chemistry,
CNRS UMR 5034,
Hélioparc, 2, Avenue Angot,
64053 Pau, France
R. Lobinski.A. Polatajko
Department of Chemistry,
Warsaw University of Technology,
Noakowskiego 3,
00-664 Warsaw, Poland

Page 2

SeMet and total Se, whereas the intercomparison was
judged unsuccessful for Met because only two external
laboratories provided values, both of which were outside
the certified range.
Keywords Selenomethionine.Yeast.Reference
material.Speciation.Intercomparison
Introduction
Mineral supplementation represents a sizeable portion of
the multibillion dollar nutritional supplement/functional
food/animal feed business.
Selenium is one of the most commonly used mineral
supplements. In most such dietary supplements the
predominant chemical form of selenium is selenomethio-
nine (SeMet), either as synthetic L-SeMet or yeast-based
SeMet.
Many of the supplemented metals including Se are
added as metal-enriched yeast (i.e., Zn, Cr) to the final
formulation. However, there are currently no CRMs
available for quality control of measurements characteriz-
ing these supplements. In the case of Se-enriched yeast, the
question of the quality of the product is even more difficult
because most of the Se present is covalently bound to
carbon, forming the non-canonical amino acid selenome-
thionine. Most of the functionality claims for Se-enriched
yeast are related to its SeMet content.
Accurate assessment of Se, and especially the SeMet
content, presents a significant analytical challenge. There
are no CRMs currently available with which to evaluate
measurement performance for SeMet. In a recent publica-
tion, several commonly used extraction methods were
compared for determination of SeMet; results were very
poor [1]. In response to this need, the Institute for National
Measurement Standards (INMS) of the National Research
Council Canada (NRC) recently completed certification of
a new selenized yeast material, certified for SeMet, Met
and total Se amount content.
As part of this process, an intercomparison exercise
piloted by INMS was organized to evaluate the state-of-
the-art of SeMet measurement capabilities. A call for
participation was sent out in August 2004 to laboratories
identified by INMS as having expert capabilities in the
field. By the end of August 2004, all respondents were sent
one bottle of sample containing 8 g of selenized yeast
material. Instructions for the storage of the samples were
provided to each participant. Results of five replicate
measurement were requested to be submitted to NRC no
later than 31 October 2004. Seven laboratories submitted
results (2 National Metrology Institutes (NMIs) and 5
university/government laboratories). A summary of partic-
ipants and their affiliations is presented in Table 1. This
report details the preparation and certification of the new
CRM and presents the results of the intercomparison
exercise.
Experimental
Certification
Instrumentation
Determination of SeMet and Met AThermo Finnigan TSQ
quantum AM triple quadrupole instrument (San Jose, CA,
USA) was used for ES-MS analysis. ES-MS conditions
(i.e., capillary voltage, lens voltage, multipole offset and
entrance voltage) were optimized for response from
selenomethionine using a standard tune procedure.
Anion exchange HPLC separations were achieved using
a Hamilton PRPx-100 (250 × 4.6 mm × 5 μm) column
(Hamilton, Reno, NV, USA) with a PRPx-100 guard
column (Hamilton). A Dionex BioLC, model LCM
(Dionex Corp., Sunnyvale, CA, USA) fitted with a
100-μL injection loop was employed for the anion
exchange HPLC separations. Reversed-phase separations
were undertaken using a Prevail C18 (150 × 2.1 mm ×
5 μm) column (Alltech, Deerfield, IL). A Hewlett–Packard
HP 1100 pump with autosampler was used for the
reversed-phase separations.
A Hewlett–Packard HP 6890 GC (Agilent Technologies
Canada Inc., Mississauga, ON, Canada) fitted with a DB-
5MS column (Iso-Mass Scientific Inc., Calgary AB,
Canada) was used for the separation of the Met and
SeMet in preparations of yeast extract. Detection was
achieved with an HP model 5973 mass-selective detector
(MS).
Table 1 Intercomparison participants
Acronym OrganizationCountry Contact
BCU
LGC
PAU
NRC
USDA
Budapest Corvinus University, Dept. of Applied Chemistry
Laboratory of the Government Chemist
Group of Bio-Inorganic Analytical Chemistry, CNRS UMR5034
Institute for National Measurement Standards, National Research Council Canada
United States Department of Agriculture, Beltsville Human Nutrition Research Center, Food
Composition Laboratory
University of Cincinnati, Department of Chemistry
University of Oviedo, Dept. of Physical and Analytical Chemistry
Hungary P. Fodor
UK
France
Canada
US
R. Wahlen
R. Lobinski
Z. Mester
R.
Goldschmidt
J. Caruso
A. Sanz-
Medel
UC
OVIE
US
Spain
169

Page 3

Total Se AThermoFinnigan Element2 sector field ICP-MS
(SF-ICP-MS; Bremen, Germany) in combination with a
Scott-type double-pass glass spray chamber fitted with a
PFA self-aspiratingnebuliser
Omaha, NE) was used along with a plug-in quartz torch
and a sapphire injector for measurement of total Se
content.
A Perkin–Elmer Optima 3000 radial view ICP-OES was
used for the determination of Se in yeast samples. A gem
tip cross-flow nebuliser and an alumina injector were used.
Background-corrected measurements were made at the
196.026-nm line.
A CEM MDS-2100 microwave digester (Matthews,
NC) equipped with Teflon vessels was used for closed
vessel high-pressure decomposition of yeast to enable
determination of total Se.
(ElementalScientific,
Reagents and solutions
Nitric acid was purified in-house prior to use by sub-
boiling distillation of reagent-grade feedstock in a quartz
still. Environmental-grade ammonium hydroxide was
purchased from Anachemia Science (Montreal, PQ, Can-
ada). OmniSolv methanol (glass-distilled) and chloroform
were purchased from EM Science (Gibbstown, NJ). High-
purity deionized water (DIW) was obtained from a
NanoPure mixed-bed ion-exchange system fed with re-
verse osmosis domestic feedwater (Barnstead/Thermolyne
Corp.). Certified-grade chloroform was sourced from
Fisher Scientific (Ottawa, Canada). Methanesulfonic acid
(98% purity) and methyl chloroformate (99% purity) were
obtained from Sigma Aldrich Canada (Oakville, ON,
Canada).
Enriched isotope82Se (elemental), purchased from the
Oak Ridge National Laboratory (Oak Ridge, TN), was used
for the determination of total Se. A stock solution of82Se
(≈90.5 μg mL−1) was prepared by dissolution of this
material in a few mL of HNO3and diluted with DIW. The
concentration of this82Se spike was verified by reverse
spike ID using natural-abundance Se standards prepared
from high-purity elemental Se.
Natural-abundance high-purity Met, SeMet,
riched Met and
purchased from Sigma Aldrich Canada. Individual stock
solutions of 1,000–2,500 g mL−1were gravimetrically
prepared in 1% HCl solution and kept refrigerated until
used. Standard purity was assessed using ICP MS, LC-ICP
MS and LC-MS methods.
A
obtained from W. Wolf (Food Composition Laboratory,
USDA, Beltsville, MD) and used to prepare a stock
solution of ≈450 μg mL−1in 1% HCl. The concentration of
74SeMet in the spike was verified by reverse spike ID
against the natural-abundance SeMet standards.
13C-en-
13C-enriched SeMet compounds were
74Se-enriched SeMet (74SeMet) compound was
Sample preparation for determination of total Se
in yeast by ICP-MS
Four 0.25-g subsamples of yeast were accurately weighed
into individual precleaned Teflon digestion vessels. A
suitable amount of the enriched inorganic82Se spike was
then added to each vessel. Three process blanks (spiked
with 10% of the amount of enriched isotope solution used
for the samples) were processed along with the samples.
After 5 mL of nitric acid and 0.2 mL of H2O2were added,
the vessels were capped and the sample digested in a CEM
MDS-2100 microwave oven. The heating conditions were
as follows: 10 min at a pressure of 20 psi and 40% power,
10 min at 40 psi and 50% power, 10 min at 80 psi and 50%
power, 20 min at 100 psi and 60% power, and 30 min at
120 psi and 70% power. After cooling, 0.25-mL volumes
of the digested solutions were transferred to precleaned
polyethylene screw-capped bottles and diluted to 25 mL
with 1% HNO3.
Calibration of the 90.5 μg mL−182Se-enriched spike was
achieved by reverse spike ID. This measurement was
performed at the same time as the measurement of total Se
in the yeast digests. Six replicate samples were prepared
by accurately pipetting 0.1003-mL volumes of 90.5 μg
mL−182Se spike solution into precleaned polyethylene
screw-capped bottles. Aliquots of 0.0700 mL of the first
1,000 μg mL−1natural-abundance Se stock solution were
added to the first three bottles; this was repeated for the
other three bottles using the second 1,000 μg mL−1Se
stock solution. The contents of each bottle were then
diluted with 15 mL of 1% HNO3for SF-ICP-MS analysis.
The digested yeast samples and the six reverse spike ID
calibration samples were analysed by SF-ICP-MS on the
same day. Mass bias correction was implemented based on
the theoretical natural-abundance ratio of an isotope pair
divided by the mean value of the isotope pair measured in a
natural-abundance Se standard. The Element2 SF-ICP-MS
was optimized daily following recommendations by the
manufacturer. Detector dead time was determined accord-
ing to the procedure described by Nelms et al. [2] using
three different concentrations of U.
Sample preparation for determination of total Se
in yeast by ICP-OES
Digestion of samples was conducted in a class-10 or class-
100 clean room environment. A minimum test mass of
250 mg of yeast material was weighed into pre-cleaned
Teflon CEM ACV digestion vessels and 7 mL of high-
purity nitric acid and 0.2 mL hydrogen peroxide (ASC
grade) were added. Three procedural blanks were prepared.
The samples and blanks were digested in the CEM MDS-
2100 microwave oven using a program noted earlier.
After cooling, the caps were removed from the vessels
and rinsed with DIW, with the rinse solution being added to
that in the vessel. The vessels were then placed on a hot
plate in a class-10 fume hood and the contents were
evaporated to near dryness. The residues were dissolved in
170

Page 4

1.0 mL high-purity concentrated HNO3and then accurately
diluted gravimetrically to 25 g with DIW. External
calibration approach was used for quantitation.
Sample preparation for determination of SeMet
and Met in yeast following refluxing
with methanesulfonic acid
The yeast extraction procedure used in this study followed
that described by Wrobel et al. [3] with minor modifica-
tions. Three sample blanks and six subsamples of yeast
were prepared at the sametime. In brief, 0.25 gof yeast was
spiked with 0.250 mL of 2,193.4 μg mL−113C-enriched
SeMet and 1.00 mL of 1,093.1 μg mL−113C-enriched Met.
After addition of 16.75 mL of DIW and 6 mL of
methanesulfonic acid (resulting in a concentration of 4 M
for methanesulfonic acid and 24 mL volume in total), the
contents were refluxed on a hot plate for 16 h.
The derivatisation procedure followed that reported by
Haberhauer–Troyer et al. [4] based on use of a 1-mL
volume of extract. After addition of 0.48 mL of ammonium
hydroxide and 0.75 mL of methanol/pyridine (3:1 v/v) to a
10-mL glass vial containing 1 mL of extract, 0.250 mL of
methylchloroformate was slowly added. The vial was
shaken manually for 1 min with venting. A 1-mL aliquot of
chloroform was then added and the vial was shaken
manually for 1 min. The chloroform layer was then
transferred to a 1-mL glass vial for subsequent GC-MS
analysis.
Determination of SeMet and Met
GC-MS As shown in Fig. 1, good separation and peak
profiles for Met and SeMet were obtained under optimized
conditions. Thederivatized
(C8H15O4NS+) and the derivatized SeMet molecular ion
Metmolecular ion
4.204.40 4.604.805.005.20 5.40 5.605.80 6.006.20 6.40
3000000
4000000
5000000
6000000
7000000
Time, minutes
Met
SeMet
1000000
2000000
a
Abundance
218219220 221222223 224225
8000
16000
24000
32000
40000
48000
m/z
221
222
223
224
Abundance
Abundance
b
273
c
263 264 265 266 267 268 269 270 271 272 273
5000
10000
20000
25000
30000
35000
269 270
267
268
266
271 265272
263 264
m/z
Fig. 1 a Total ion GC-MS
chromatogram (m/z 50–300)
of a spiked (13C-enriched Met
and SeMet) yeast extract deriv-
atized with methyl chlorofor-
mate; b Met ion isotope pattern
following derivatisation with
methyl chloroformate; and
c SeMet ion isotope pattern
following derivatisation with
methyl chloroformate
171

Page 5

(C8H15O4NSe+) were selected for quantitation in this
study [1, 5]. The increased abundance at m/z 222 and 270
in Fig. 1b,c reflects the contribution from added
enriched spikes. Ions at m/z 221 and 222 were selected as
reference and spike ions for ID analysis using
enriched Met to calculate the final concentration of Met in
the yeast. Relative abundances of 85.793 and 8.706% for
ions at m/z 221 and 222 in the sample and 8.670 and
85.932% in the spike were used for the quantitation.
Similarly, ions at m/z 269 and 270 with relative
abundances of 45.102 and 4.210% in the sample and
2.272 and 45.193% in the spike were selected for ID
analysis using a13C-enriched SeMet spike. All four ions
were monitored under selective ion monitoring (SIM)
mode. Peak areas were used to calculate the reference/
spike ion ratios, from which the analyte concentrations
were calculated. The following equation was used for the
quantitation of Met and SeMet in yeast:
13C-
13C-
C ¼ Cy?
my
w ? mx?Ay? By? Rn
where Cxis the analyte concentration (μg; g−1), Cyis the
concentration of enriched spike (μg g−1), my(g) is the
mass of spike used to prepare the blend solution of sample
and spike, mxis the mass (g) of yeast sample used, Ayis
the abundance of reference ion in the spike, By is the
Bx? Rn? Ax?AWx
AWy? Cb
(1)
abundance of spike ion in the spike, Axis the abundance of
reference ion in the sample, Bxis the abundance of spike
ion in the sample, Rnis the measured reference/spike ion
ratio (mass bias corrected) in the blend solution of sample
and spike, AWxis the atomic mass of analyte in the sample
and AWyis the atomic mass of analyte in the spike. As is
evident from this equation, only the reference/spike ion
ratios in the spiked samples need to be measured to derive
the final analyte concentrations. The mass bias correction
factor was calculated from the expected to measured ratio
using a natural abundance Met and SeMet standard
solution.
LC-MS Sample digests were diluted to 100 mL with DIW
and filtered (0.45 μm) prior to injection.
For ID calibration, reference and spike protonated
molecular ions at m/z 150 and 151 for Met and 198 and
199 for SeMet, respectively, were monitored in SIM mode
for integration. Peak areas were used to calculate the
reference/spike ion ratios of Met (150/151) and SeMet
(198/199) for use in Eq. 1. Detailed description of the LC-
MS protocol can found elsewhere [6].
Determination of moisture content
The water content of dried and non-dried selenium-
enriched yeast samples was determined using the Karl
Table 2 Analytical methods and instrumental techniques used by the participants for determination of total selenium
ParticipantCalibration MethodInstrumentation
BCU
LGC
PAU
NRC
USDA
UC
OVIE
EC
ID (77Se spike)
SA, IS
ID (82Se)
ID (82Se spike)
SA, IS (Yt)
ID (77Se spike)
Acid digest, hotplate
Acid digest, microwave
Acid digest, microwave
Acid digest, microwave
Acid digest/ chelation with TFMPD
Acid digest, microwave
Acid digest, microwave
AFS
ICP-MS (quadrupole, ORS)
ICP-MS
ICP-MS (SF)
GC-MS
ICP-MS (quadrupole, ORS)
ICP-MS (quadrupole, ORS)
ID-MS: isotope dilution mass spectrometry; SA: standard additions; IS: internal standard; EC: external calibration
Table 3 Analytical methods and instrumental techniques used by participants for determination of selenomethionine
Participant Calibration Extraction /derivatisationInstrumentation
BCU (1) EC
(2) EC
SA and EC
SA, IS (Rh)
(1) ID (13C,74Se)
(2) ID (13C,74Se)
(3) ID (13C,74Se)
Protease XIV, 24 h at 37 °C + shaking
3M p-toluol-sulfonic acid+0.2% triptamine
Protease+lipase, 20 h at 37 °C+shaking
Protease+lipase, 17 h at 37 °C+stirring × 3 (repeated three times)
Methanesulfonic acid, 16 h, reflux, Met-chloformate derivatisation GC-MS
Methanesulfonic acid, 16 h, reflux, Met-chloformate derivatisation GC-MS
Methanesulfonic acid, 16 h, reflux methanesulfonic acid, 16 h, reflux,
no derivatisation
Methanesulfonic acid, 16 h, reflux, CNBr derivatisation
Methanesulfonic acid, 16 h, reflux
Post column ID (77Se) Protease+lipase, 16 h at 37 °C
LC-UV-AFS
Amino acid analyzer
LC-ICP-MS (quadrupole)
LC-ICP-MS
LGC
PAU
NRC
LC-MS
USDA
UC
OVIE
ID (74Se spike)
SA, IS
GC-MS
HPLC ICP-MS (quadrupole)
HPLC ICP-MS (quadrupole,
DRC)
ID-MS: isotope dilution mass spectrometry; SA: standard additions; IS: internal standard; EC: external calibration
172

Page 6

Fischer titration technique. Formamide was used as the
extraction solvent. A weighed quantity of the sample was
placed in either a 30-mL hypo vial (dried samples) or a
25-mL screw-top vial and 20 mL of formamide was added
using a dispensing pipette. The containers were sealed,
shaken, briefly sonicated and allowed to stand for several
days prior to analysis. All containers had a septum-type cap
through which weighed sample aliquots were taken by a
syringe and injected into the titration apparatus. Blank
solvent samples were carried through the same procedure.
Results were corrected for the water content of the solvent.
Preparation of CRM
A 25-kg lot of selenized yeast material with a nominal total
Se concentration of 2,000 ppm was obtained from a
commercial source (Lallemand, Montreal, PQ). Homoge-
neity of this material was assessed by measurement of total
Se following representative sampling of the bulk. The
results of the testing provided a relative standard deviation
of the mean of analysis of seven subsamples of 0.9% (i.e.
2,008.7±18.6 μg g−1). Repeat analysis of the same
subsample yielded a mean of 2,003.5±18.6 μg g−1. The
material was thus deemed homogeneous for the purpose of
this exercise at the 250-mg subsample size and no attempt
to further homogenize the material was undertaken.
The sample was partitioned into 2,400 pre-cleaned glass
bottles, each containing a nominal 8 g of sample. All
bottles were gamma-irradiated (minimum 2.5 Mrad) to
ensure long-term stability, labelled and packaged in
trilaminate foil pouches and stored at −25 °C. Analysis
of randomly selected bottles of the yeast samples before
and after gamma-irradiation revealed no change in SeMet
composition, in contrast to the situation with other metal
organic compounds such as butyltins [7].
Intercomparison exercise
As noted earlier, the choice of analytical method was left to
the discretion of the participant. This includes the sample
extraction procedure, instrument method, calibration ap-
proach and procedure for moisture determination. The
calibration approaches varied from ID or calibration with a
natural abundance standard in combination with selected
internal standards (SA calibration with an accompanying
internal standard (SA-IS) andEC, with and without internal
standard). Tables 2 and 3 summarize the analytical methods
and instrumental techniques used by the participants for
determination of total selenium and selenomethionine.
Determination of total Se
Each laboratory employed various acid digestion protocols
combined with ICP-MS, AFS or GC-MS detection
schemes. The main focus of the Se intercomparison
exercise was the determination of SeMet and Met in the
new CRM. Table 2 summarizes the extraction, detection
and calibration methods employed.
Determination of SeMet and Met
Table 3 summarizes the extraction, separation, detection
and calibration methods reported by the participants. Each
of these is briefly summarized below.
BCU1 One-step enzymatic digestion with the help of
protease XIV was applied. The sample (100 mg) was
placed in a 15-mL polyethylene vial followed by the
addition of 20 mg enzyme dissolved in 4.5 mL 0.1 M
TRIS/HCl (pH 7.5). The samples were stirred at 200 rpm
for 24 h at 37 °C. After proteolysis, the samples were
centrifuged at 4,100 g for 25 min at 15 °C. The
supernatants were recovered, and the residue suspended
by a vortex device in 1.0 mL DIW and once again
centrifuged. After this step, the matching supernatants
were combined, filtered through a 0.45-μm cellulose
nitrate syringe filter (Millipore, Tullagreen, Ireland) and
the final volume made up to 25.0 mL using DIW. All
procedures and measurements were carried out in triplicate
using an HPLC-UV photochemical digestion–hydride
generation –atomic fluorescence detection (HPLC-UV-
HG-AFS) system. External calibration and standard addi-
tion were used for quantitation based on a stock solution of
seleno-DL-methionine (Sigma Chemicals, St. Luis, MO,
USA). The separation was carried out on an Hamilton PRP
X-100 anion exchange column using a mobile phase of
200 mmol L−1ammonium dihydrogen phosphate at pH
6.0 with isocratic elution [8].
BCU2 A 60-mg subsample of the yeast was mixed with
10 mL 3M p-toluol-sulfonic acid + 0.2% tryptamine in a
12-mL hydrolysing vessel and digested in a block
thermostat at 110 °C for 24 h. After hydrolysis, the
Table 4 Uncertainty components for SELM-1
Se (mg kg−1)SeMet (mg kg−1)Met (mg kg−1)
uchar
uhom
ustab
ucombined
uCRM(k=2)
28
15
29
20
70
78
157
54
31
124
138
277
0
32
64
Table 5 Certified values in SELM-1 [14]
Concentration (mg kg−1)±expanded
uncertainty (k=2)
Total selenium
Selenomethionine
Methionine
2,059±64
3,431±157
5,758±277
173

Page 7

solution was neutralized with 10 mL of 2 N NaOH and
diluted to 20 mL with DIW. The final solution was passed
through a 0.45-μm membrane filter and stored at −20 °C
until analysed. The analysis was conducted using an LC-
based amino acid analyser (Biotronik, LC3000, Germany)
which employed post-column derivatisation (ninhydrin)
photometric detection.
LGC Enzymatic hydrolysis was achieved by addition of
40 mg of protease and 20 mg of lipase in 5 mL of a 30 mM
Tris-HCl buffer solution (pH 7.5) to approximately 0.2 g
of yeast. Incubation at 37 °C was then carried out in the
dark for 20 h [9, 10]. During enzymolysis, the sample
slurries were constantly and gently homogenized using a
rotary shaker set at 60 rpm. Hydrolysed samples were
centrifuged at 3,000 rpm for 30 min and the supernatants
filtered, degassed with He for 5 min and stored at 4 °C in
the dark (for no longer than 2 h) until analysis. A 500-fold
dilution of these samples was performed prior to their
analysis for speciation of Se.
Quantitation of SeMet in the digests was based on a
50-μL subsample analysed by RP-IP HPLC-ICP-MS at a
flow rate of 1 mL min−1using a 98:2 water/methanol
mixture containing 0.1% (v/v) trifluoroacetic acid (TFA, as
ion pairing agent) as the mobile phase. Results obtained by
external calibration and by the standard addition technique
at three concentration levels (using peak area measure-
ments of the chromatographic signals) by monitoring the
82Se signal agreed within the respective uncertainties of
the measurements.
PAU Extraction of selenomethionine was based on use of a
0.2-g subsample of selenized yeast hydrolysed with 5 mL
of 30 mM TRIS-HCl solution (pH 7.5) containing 20 mg
of protease and 10 mg of lipase. The sample was incubated
for 17 h at 37 °C while being magnetically stirred and
finally centrifuged for 10 min at 2,500 rpm. The super-
natant was stored in the fridge after addition of 5 μL of β-
mercaptoethanol (0.1%) to avoid oxidation. The residue
was then subjected to a further proteolytic digestion with
two fresh enzymatic solutions. Finally, the three superna-
tants were pooled and the total amount of extracted
selenium was quantified by ICP-MS using the method of
standard additions in the presence of a 10 ppb Rh internal
standard.
Chromatographic separation of the Se species from
sample extracts was performed by HPLC ICP-MS using
an Agilent Technologies 1100 HPLC system (Palo Alto,
CA, USA) fitted with a Hamilton PRP X-100 anion
exchange column coupled to an Agilent 7500C ICP-MS
for element-specific detection. Gradient elution was used
for species separation from a 100-μL injection volume. A
flow rate of 1.5 mL min−1of the following was used:
0–5 min, 100% A; 5–30 min, 0–100% B; 30–40 min,
100% B where A is 20 mM acetic acid containing 10 mM
triethylamine at pH 4.7 and B is a mixture of 200 mM
acetic acid and 100 mM triethylamine at pH 4.7.
Table 6 Participant’s results for
determination of SeMet in
SELM-1
Participant Mean SeMet
(mg kg−1)
Standard deviation
(mg kg−1)
95% confidence interval
(mg kg−1)
Number
of replicates
BCU(1) 2,576
(2) 2,073
2,991
3,437
(1) 3,461
(2) 3,407
(3) 3,481
3,279
1,970
3,587
139
50
115
65
27
14
73
18
59
48
386
215
319
180
61
33
203
50
164
133
5
3
5
5
7
8
8
5
5
5
LGC
PAU
NRC
USDA
UC
OVIE
Table 7 Participant’s results for
determination of total Se in
SELM-1
ParticipantMean Se
(mg kg−1)
Standard deviation
(mg kg−1)
95% confidence interval
(mg kg−1)
Number
of replicates
BCU
LGC
PAU
NRC
USDA
UC
OVIE
1,836
1,983
2,044
2,103
2,012
2,083
2,145
13
16
35
14
30
87
23
36
44
97
32
381
241
64
5
5
6
9
2
5
5
174

Page 8

1500
2000
2500
3000
3500
4000
UCBCU-2 BCU-1LGC USDANRC-2PAUNRC-1 NRC-3OVIECERT
1500
1700
1900
2100
2300
2500
BCULGC USDA PAUUCNRCOVIE CERT
5000
6000
7000
8000
9000
10000
11000
USDA NRC-3 NRC-2 NRC-1BCULGCPAU UCOVIE CERT
a
b
c
Concentration, mg/kg
Fig. 2 Results for a total
SeMet, b Se and c Met in
SELM-1 CRM. Note that error
bars represent 95% confidence
interval. The last data point
(CERT) along with the dotted
horizontal lines represent CRM
SELM-1 certified value and
combined expanded uncertainty
(k=2) interval. Solid line indi-
cates the median value of the
intercomparison; dashed line is
the deviation from the median
Table 8 Participant’s results
for determination of Met in
SELM-1
ParticipantMeanMet
(mg kg−1)
Standard deviation (±)95% confidence interval (±)Number
of replicates
BCU
NRC
10,050
(1) 5,886
(2) 5,772
(3) 5,698
5,381
132
25
17
77
97
568
69
40
182
269
3
8
8
8
5 USDA
175

Page 9

NRC1 See above (GC-based ID analysis) operator N1
NRC2 See above (GC-based ID analysis) operator N2
NRC3 See above (LC-MS-based ID)
USDA Selenomethionine was determined by stable ID-MS
[2–4] using an HP 6890 GC with 5973 MSD. For sample
preparation
added to each of the 5 samples (between 0.2 g and 0.22 g).
The samples were boiled under reflux in 10 mL of 4 M
methanesulfonic acid for 14 h. Samples were then
centrifuged and 1-mL portions of the supernatants were
transferred to conical glass reaction vials. To the samples
were added 0.6 mL of 5 N NH4OH, 1 mL of 2% SnCl2in
0.3 M HCl, and then 200 μL of 3 M CNBr in methylene
chloride (with vigorous mixing between additions).
Samples were kept at 37 °C for 19 h with periodic
vortexing. The reaction product, CH3SeCN, was then
extracted with chloroform. For GC-MS analysis (negative
chemical ionisation with methane reagent gas, HP-5MS
column) SIM mode was used, with the ions at m/z 99.9
and m/z 105.9 being monitored (corresponding to SeCN).
74Se-enriched selenomethionine spikes were
UC An Agilent 1100 series HPLC system was used for
chromatographic separation of selenomethionine followed
by ICP-MS detection using an Agilent 7500ce (Agilent
Technologies, Palo Alto, CA, USA).
For the determination of selenomethionine, five repli-
cate 0.2-g subsamples of yeast were precisely weighed.
Methanesulfonic acid (20 mL, 4 M) and β-mercaptoeth-
anol (400 μL) were added to each of the samples. The
mixtures were heated under reflux for 16 h. The extracts
were filtered and transferred to 100-mL volumetric flasks
followed by dilution with DIW.
OVIE A digital control immersion thermostat model
Digiterm 100 (J.P. Selecta, Barcelona, Spain) was used
for incubation of samples during enzymatic hydrolysis.
Extracts were centrifuged in a Biofuge stratos centrifuge
(Heraeus, Hanau, Germany).
For extraction, 20 mg of protease, 10 mg of lipase and
5 mL of milli-Q water were added to 0.2 g of sample
which was then incubated for 16 h at 37 °C. Hydrolysed
samples were further centrifuged and passed through a
0.45-μm filter.
Quantitation of SeMet in the extract was carried out by
HPLC using post-column isotope dilution ICP-MS
(Agilent 7500c Palo Alto, CA, USA) detection. A flow
of 4 mL min−1of hydrogen was used in order to pressurize
the octapole chamber. The sample introduction system
consisted of a Meinhard nebuliser with a Scott double-pass
quartz spray chamber cooled to 2 °C.
For chromatographic separations, a Shimadzu LC-10 A
HPLC pump (Kyoto, Japan) was used and injections were
carried out using a Rheodyne 7725 injection valve
(Rheodyne, CA, USA) fitted with a 100-μL loop. Anion
exchange separation was performed on a Hamilton
PRP-X100 column (250×4.1 mm, 10 μm) (Hamilton
Company, Reno, Nevada, USA). A flow rate of
0.9 mL min−1and pH 5.6, 5mM ammonium acetate as
mobile phase was used along with an injection volume of
100 μL.
Due to the high concentration of selenomethionine in
these samples, yeast extracts were diluted approximately
1:300 with ultrapure water before injection into the HPLC
Table 9 Moisture content and methodologies
Participant Moisture content (%)Method
BCU
LGC
3.2
5.04
NR
0.5 g of sample heated to 105 °C for 3 h and cooled down to room temperature in a desiccator.
The procedure was repeated (1 h heating) until constant weight was reached
1 g of sample heated to 95 °C for 24 h. Repeated until constant weight was reached
0.3 g of sample, 360 h freeze-dried
0.25–0.2 g sample heat dried at 80 °C for 4 h
Vacuum drying at 0.5 mm Hg for 24 h of 0.5 g of samples
1.0 g sample heated to 80° for 30 min. Repeated until constant weight was reached
PAU
NRC
USDA
UC
OVIE
3.7
4.25
3.88
1.28
4.01
Table 10 Summary statistics
for SeMet intercomparison
SeMet (mg kg−1) Se (mg kg−1) Met (mg kg−1)
Mean
Median
Standard deviation
Deviation from the median
Range
Minimum
Maximum
Number of data sets
3,026
3,343
607
283
1,617
1,970
3,587
2,029
2,044
101
6,557
5,772
1,961
169
4,669
5,381
10,050
87
309
1,836
2,145
1075
176

Page 10

system. The determination of Se species post-column ID
was as follows: a77Se-enriched standard solution of the
appropriate concentration was continuously introduced (at
a flow rate of 0.1 mL min−1) through a T piece at the end
of the column using a peristaltic pump (model HP4,
Sharlau Sciences, Barcelona, Spain) [11]. The resultant
intensity chromatograms were converted, after adequate
mathematical treatment, into mass flow chromatograms.
The integration of the peak corresponding to selenome-
thionine (using an Origin software) provided the amount
of Se present in that peak.
Discussion
Method development
An exhaustive literature search indicated that the vast
majority of studies devoted to the speciation of selenium
are based on various soft extraction methods, such as
methanol/water or proteolytic enzymatic digestion. There
have been very few attempts to validate any of these
methods. The challenge in determining SeMet is that it is
basically an amino acid (AA) measurement. Quantitative
AA analysis based on acid digestion has been conducted
for decades. An acid digestion protocol based on use of
methanesulfonic acid [3] was therefore first evaluated.
Acid reflux of the yeast samples with methanesulfonic acid
provided consistent results in an easily controlled experi-
mental setup. In order to evaluate the potential loss of
liberated Met and SeMet during the extraction process,
spike recovery experiments were conducted. Quantitative
recovery of an added stable isotope-labelled spike was
achieved [3]. Consequently, use of any standard addition
protocol (including ID) was deemed to be acceptable.
However, even if the liberated SeMet and Met are stable
throughout the extraction process, the question of extrac-
tion efficiency remains to be addressed. To evaluate the
efficiency of liberation of the targeted amino acids from the
yeast proteome, a model system was created consisting of a
synthetic SeMet-and
(AHPDVLTVXLQMLDDGR, X denoting SeMet) which
was subjected to methanesulfonic acid digestion. During
the course of the digestion, the synthetic peptide
completely decomposed to free amino acids. Additionally,
the rate of liberation of SeMet and Met from the synthetic
peptide was identical to that for the selenized yeast,
suggesting that the employed model system behaved
similarly to the yeast [12]. It was thus concluded that the
digestion process was suitable for determination of SeMet
and Met. In order to confirm the suitability of this system,
several enzymatic digestion protocols were tested. If
sufficient digestion times and/or large quantities of
enzymes were used, the acid hydrolysis data were
reproducible, thereby providing an independent confirma-
tion of the validity of this protocol. It was also very clear
that by improving the efficiency of the enzymatic protocols
reported in the literature, results in conformance with data
generated by acid hydrolysis were obtained. At this point,
Met-containingpeptide
the extraction method was deemed suitable for the
quantitative analysis of these two amino acids.
Certification
Moisture content
The moisture content of the sample wasdetermined by Karl
Fisher titration following analysis of six subsamples
collected from six different bottles. Three 0.3-g test
samples were placed in a desiccator at room temperature
and three were placed in a freeze dryer. After 4 days, the
average weight (moisture) loss of material placed in the
desiccator was 2.72±0.17% (SD), while those placed in
the freeze dryer lost 4.20±0.02% (SD). After 350 h, 3.60±
0.35% and 4.25±0.03%weight losswas measured for those
samples stored in the desiccator and in the freeze dryer,
respectively.
Karl Fisher titration of total moisture content based on 4
subsamples was 4.87±0.08% (SD). Analysis of previously
dried samples revealed 1.26% water content in those dried
in a desiccator and 0.73% for those dried in a freeze dryer.
Total moisture content should correspond to the sum of
those determined by water loss during the drying
procedures and those from the Karl Fisher titration data
reflecting the remaining moisture content. With desicca-
tion, this amounted to 3.60%+1.26%=4.86±0.35% (SD),
whereas for the freeze drying this amounted to 4.25%+
0.73%=4.98±0.03% (SD). These values are in agreement
with the total moisture content of 4.87±0.08% (SD)
obtained by a direct Karl Fisher titration, indicating that
the only (or at least the main) component lost during either
drying procedure is moisture and not a significant quantity
of other volatile species. As the freeze drying procedure
achieves faster equilibration, the dry-weight correction
factor used for all measurements was 4.25±0.03%.
Uncertainty
Included in the overall uncertainty estimate are contribu-
tions from the batch characterization (uchar), uncertainties
related to possible between-bottle variation (uhom) and
instability derived from effects relating to long-term
storage and transport (ustab). Expressed as standard
uncertainties, these components can be combined as [14]:
u2
cðCRMÞ¼ u2
charþ u2
homþ u2
stab
(2)
The characterization uncertainties (uchar) were calculated in
accordance with Eq. 3,
uchar¼
sffiffiffip
p
(3)
wherein s is the standard deviation of the means and p is
the number of mean results included in the calculation. The
177

Page 11

calculated uncertainty components related to the character-
ization of SELM-1 are reported in Table 4.
Homogeneity
The material was tested for homogeneity and the data
subjected to ANOVA. Bottles from various points of the
bottling process were selected for analysis. Results from
different bottles, as determined by ID-GC-MS, resulted in
an uncertainty as reported in Table 4. The inhomogeneity
contribution to uncertainty, uhom, was equated to the
experimentally determined between-unit standard devia-
tion (sbetween) as the best estimate of the uncertainty due to
between-unit heterogeneity. However, the situation de-
picted by Eq. 4 occurred for Se:
S2
between<S2
means
n
(4)
wherein smeasis the repeatability standard deviation for the
method used in the homogeneity assessment and n is the
number of replicates per unit. For this case, uhomwas
calculated according to Eq. 5:
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
VMSwithin
uhom¼
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
MSwithin
n
r
4
2
s
(5)
where MSwithinrepresents the mean squares within groups
and vMSwithinis the number of degrees of freedom. The
homogeneity is warranted for samples of 250 mg and
above.
Stability
Long-term component Long-term stability was assessed by
monitoring the samples over a period of 6 months. A
0.34% decrease in the SeMet concentration was found
during the first period. The Met content decreased by
0.33%. The uncertainties related to long-term stability
(ustab-long) were calculated in accordance with Eq. 6,
u2
instab¼ u ct
ð Þ2þ
δffiffiffi
3
p
??2
(6)
where u(ct) represents the uncertainty of the concentration
correction applied in order to adjust the certified value to
the halfway point of the shelf life and 2δ is the predicted
total change over the expected shelf life [13].
Based on long-term stability data, a shelf life of
18 months was initially selected for this CRM. The
stability will be continuously monitored and updates will
be provided that will extend the expiry period. Certified
values for SeMet and Met were adjusted to the expected
halfway point of this time curve and the rate of change
incorporated as an uncertainty to the certified value.
Short-term component A short-term isochronous stability
study was performed on the material following storage of
two bottles at −25 °C, +4 °C, +22 °C and +44 °C for
31 days. These temperatures were selected to model
various storage conditions (i.e. freezer, fridge and room
temperature, respectively, with 44 °C selected to mimic
possible exposure of the CRM during transport). At the
end of the storage period the samples were equilibrated to
room temperature and analysed for SeMet and Met
content. No short-term stability tests were performed for
total Se because of available historical evidence of
stability of Se in other similar biological CRMs. No
significant change in the Met concentration was found, but
that for SeMet slightly decreased. The uncertainties related
to short-term stability (ustab-short) were calculated in
accordance with Eq. 6, except that the uncertainty
component related to the correction to the halfway point
(u(ct)) was removed from the equation. The uncertainties
associated with long- and short-term stabilities were
combined and are reported in Table 4.
Table 5 shows those analytes for which certified values
have been established for this selenium-enriched yeast
reference material (SELM-1)[14]. Certified values are
based on unweighted mean results from data generated by
three independent sets of measurements in the case of
selenomethionine (SeMet) and methionine and two inde-
pendent measurements for total selenium and are corrected
for stability. The methods used to obtain the three sets of
SeMet data are described earlier and all three methods are
based on isotope dilution MS. Met has been determined
using the same three methods as used for SeMet, and all
Met measurements were based on ID-MS. Total Se was
determined by SF ICP-MS using ID calibration and by
ICP-OES using external calibration. All data used for the
certification was generated by NRC personnel using
instrumentation located in the NRC laboratories.
Intercomparison exercise for determination of Se,
SeMet and Met
Two laboratories submitted more than one result, based on
an alternative separation/detection methodology or, as in
the case of NRC(1–2), the same extraction, calibration and
measurement method but different operators using inde-
pendently prepared standards. For the purpose of this
exercise, all such values were used and treated as being
independent and are delineated as the appropriate labora-
tory acronym accompanied by a numerical suffix e.g. BCU
(1).
No control sample was used to validate participants’ data
as no RMs are available for determination of SeMet. No
uncertainty budgets were requested from the participants
for either of the analytes and thus only data for the sample
standard deviation (1s) and 95% confidence interval are
reported by each participant and presented in Tables 6, 7
and 8 and illustrated in Fig. 2. All data reported here is
based on dry weight.
178

Page 12

No drying method was prescribed for this intercompar-
ison exercise. Table 9 summarizes the moisture content
determined by the participants. There is an agreement
between the laboratories except for UC which reported
significantly lower moisture content (less than one third of
the average value). Additionally, average moisture content
based on measurements from 6 laboratories (except UC) is
4.01%, in reasonable agreement with the values determined
by NRC during the production of this material (4.25%).
As there were obvious extreme values, a median of the
means submitted by the individual laboratories was chosen
to represent the data (Fig. 2). The precision in the data
generated by an individual laboratory was generally quite
high. It is important to appreciate that the levels measured
in these samples were substantial, presenting no significant
challenge in terms of detection limits. This could result in a
situation wherein data from individual laboratories are
quite precise but not accurate.
Measurements performed using ID calibration (13C-
labelled Met and SeMet by NRC and74Se-labelled SeMet
by USDA) clearly provided better precision over other
calibration approaches.
Results for determination of SeMet in CRM SELM-1
were acceptable for the majority of the participants. This is
illustrated by the data in Table 6 and Fig. 2. Results
submitted by BCU (1–2) and UC were biased low with
respect to the accepted measurement window defined by
the certified value and its associated expanded uncertainty.
It is interesting to note that the data reported for SeMet
gradually approaches a maximum, likely because the
extraction/liberation efficiency of selenomethionine may
vary significantly depending on the digestion method used.
It appears that the acid digestion methods employed by
USDA and NRC (1–3) provided consistently higher
extraction efficiencies (UC employed the same acid
digestion protocol but LC-ICP-MS quantitation and SA
calibration and their results are over 30% lower than the
expected values). In earlier studies, it was shown that the
efficiency of the enzymatic protocol could vary signifi-
cantly, depending on the extraction conditions and the
activity of the enzyme used. It appears that oversizing of
the extraction protocol (i.e. the use of large quantities of
enzyme) or the use of a repeated/multi-step extraction
protocol, such as the one employed by PAU, generates
consistent results. The single-step enzymatic extraction and
the measurement protocols used by LGC and OVIE were
quite similar but their reported values differed by 20%,
suggesting a large potential variation in the efficiency of
digestion when enzymes with different activities are used.
An additional limitation of the enzymatic extraction
approach is that no canonic amino acids could be measured
because autolysis of the enzymes generates extremely high
and unpredictable background levels for these amino acids.
Only five data sets were submitted for Met (Table 8).
Data provided by BCU are almost double those for the
other four sets of numbers. Although the mean values
reported by the USDA are out of the certified range, the
95% confidence interval overlaps the certified range. It is
interesting to note that values reported by BCU were
generated using a conventional LC-based amino acid
analyser with post column derivatisation (ninhydrin) and
photometric detection.
Data for total selenium were acceptable for all
laboratories but BCU, which was biased approximately
10% lower than the certified value (Table 7). Data from
OVIE and LGC were both outside the certified range but
the range defined by the 95% confidence interval overlaps
with the certified range.
Table 10 summarizes the statistics arising from this
study.
It is clear from Fig. 2 and Table 10 that the
intercomparison median values for all three target analytes
are within the certified ranges for these species. However it
is important to note that for this median calculation, NRC’s
data are also included which significantly bias the overall
consensus data towards the certified values (e.g. the
methionine where three out of the five data points come
from NRC). By averaging the data sets produced by NRC
and treating them as a single data point for the purposes of
the intercomparison, the medians would still fall with in the
certified ranges for Met and total Se but just slightly below
that for SeMet. Because there is relatively little compara-
tive experience with measurement of SeMet, no data points
were eliminated from the calculation of the median even if
they clearly failed statistical outlier tests such as the
Cochrane and Grubb’s tests. They were retained in order to
preserve the integrity of the data and the overall robustness
of the study.
Challenges
Determination of SeMet requires significant sample prep-
aration to ensure exhaustive extraction of the analyte
without decomposition, as well as derivatisation of the
analyte prior to instrumental analysis if GC separation is
used. Precision and accuracy can be enhanced through
application of ID-MS techniques, but spike equilibration
must be assured if results are to be acceptable.
Acknowledgements
and V. Clancy for preparation of the yeast standard and G. Gardner
and J. Lam for the Karl Fischer measurements. The NRC is grateful
to Institute Rosell-Lallemand for supporting this research and
providing the yeast sample used in this study.
BCU thank the Department of Biochemistry and Food Technology
at BUTE for the amino acid analysis.
LGC acknowledges support under contract by the UK Department
of Trade and Industry as part of the National Measurement System
Valid Analytical Measurement (VAM) program.
OVIE acknowledges support from project BQU-2003-04671 and
thanks Dr. Vanessa Díaz Huerta for her instrumental work.
The NRC thanks V. Boyko, M. McCooeye
179

Page 13

References
1. Yang L, Sturgeon RE, McSheehy S, Mester Z (2004) J
Chromatogr A 1055:177–184
2. Nelms SM, Quetel CR, Prohaska T, Vogl J, Taylor PDP (2001)
J Anal At Spectrom 16:333–338
3. Wrobel K, Kannamkumarath SS, Caruso JA (2003) Anal
Bioanal Chem 375:133–138
4. Haberhauer Troyer C, Alvarez Llamas G, Zitting E, Rodriguez
Gonzalez P, Rosenberg E, Sanz Medel A (2003) J Chromatogr
A 1015:1–10
5. Yang L, Mester Z, Sturgeon RE (2004) Anal Chem 76:
5149–5156
6. McSheehy S, Yang L, Sturgeon R, Mester Z (2005) Anal Chem
77:344–349
7. Yang L, Bancon Montigny C, Mester Z, Sturgeon RE, Willie
SN, Boyko VJ (2003) Anal Bioanal Chem 376:85–91
8. Bodo ET, Stefanka Z, Ipolyi I, Soros C, Dernovics M, Fodor P
(2003) Anal Bioanal Chem 377:32–38
9. Infante HG, G O C, Rayman M, Wahlen R, Entwisle J, Norris
P, Hearn R, Catterick T (2004) J Anal At Spectrom 19:
1529–1538
10. Polatajko A, Sliwka Kaszynska M, Dernovics M, Ruzik R,
Encinar JR, Szpunar J (2004) J Anal At Spectrom 19:114–120
11. Reyes LH, Sanz FM, Espilez PH, Marchante Gayon JM,
Alonso JIG, Sanz Medel A (2004) J Anal At Spectrom
19:1230–1235
12. McSheehy S, Yang L, Mester Z (2006) Microchimica Acta,
http://www.springerlink.com/openurl.asp?genre=article&id=doi:
10.1007/s00604-006-0520-2
13. Ellison SLR, Burke S, Walker RF, Heydorn K, Mansson M,
Pauwels J, Wegscheider W, Nijenhuis BT (2001) Accred Qual
Assur 6:274–277
14. http://inms-ienm.nrc-cnrc.gc.ca/en/calserv/crm_files_e/
SELM1_certificate.pdf (2005)
180