Use of L-[15N] glutamic acid and homoglutathione to determine both glutathione synthesis and concentration by gas chromatography-mass spectrometry (GCMS)

Article abstract of DOI:10.1002/jms.185

A method for simultaneous measurement of both glutathione enrichment and concentration in a biological sample using gas chromatography mass spectrometry is described. The method is based on the preparation of N,S-ethoxycarbonylmethyl ester derivatives of glutathione, and the use of homoglutathione (glutamyl-cysteinyl-alanine) as an internal standard. A procedure for determination of glutamate concentration and enrichment is also reported. Both methods have within-day and day-to-day inter-assay coefficients of variation less than 5%, and recoveries of known added amounts of glutathione and glutamate are close to 100%. Taken together, these methods allowed determination of glutathione concentration and fractional synthesis rate in red blood cells using L-[¹⁵N] glutamic acid infusion. This approach was applied in vivo to investigate the effects of a 72 h fast, compared with a control overnight fast, on erythrocyte glutathione in a single dog. The 72 h fast was associated with a 39% decline in erythrocyte glutathione level, (2.9 ± 0.4 versus 4.7 ± 0.5 mmol 1^-1, fasting versus control) with no change in glutathione fractional synthesis (67.4 versus 71.3%d^-1, fasting versus control). Copyright

Full text of DOI:10.1002/jms.185

JOURNAL OF MASS SPECTROMETRY
J. Mass Spectrom. 2001; 36: 726–735
Use of L-[15N] glutamic acid and homoglutathione to
determine both glutathione synthesis and concentration
by gas chromatography-mass spectrometry (GCMS)
Bernard Humbert,1,2 Patrick Nguyen,2 Christiane Obled,3 Christine Bobin,1 Anne Vaslin,4
Shawn Sweeten4 and Dominique Darmaun1,4∗
1
INSERM U. 539, Centre de Recherche en Nutrition Humaine, Nantes, France
Unite´ de Nutrition et Alimentation, Ecole Nationale Ve´ te´ rinaire, Nantes, France
Institut National de la Recherche Agronomique, Theix, France
Nemours Children’s Clinic, Jacksonville, FL, USA
2
3
4
Received 12 December 2000; Accepted 26 April 2001
A method for simultaneous measurement of both glutathione enrichment and concentration in a
biological sample using gas chromatography mass spectrometry is described. The method is based on the
preparation of N,S-ethoxycarbonylmethyl ester derivatives of glutathione, and the use of homoglutathione
(glutamyl–cysteinyl–alanine) as an internal standard. A procedure for determination of glutamate
concentration and enrichment is also reported. Both methods have within-day and day-to-day inter-
assay coefficients of variation less than 5%, and recoveries of known added amounts of glutathione
and glutamate are close to 100%. Taken together, these methods allowed determination of glutathione
concentration and fractional synthesis rate in red blood cells using L-[15N] glutamic acid infusion. This
approach was applied in vivo to investigate the effects of a 72 h fast, compared with a control overnight
fast, on erythrocyte glutathione in a single dog. The 72 h fast was associated with a 39% decline in
erythrocyte glutathione level, (2.9 0.4 versus 4.7 0.5 mmol l−1, fasting versus control) with no change
in glutathione fractional synthesis (67.4 versus 71.3%d−1, fasting versus control). Copyright  2001 John
Wiley & Sons, Ltd.
KEYWORDS: stable isotopes; glutathione; glutamate; erythrocyte; dog
have established methods to investigate the kinetics of
GSH synthesis in vivo using stable isotopes over the course
of intravenous infusions of stable isotope-labeled glycine,4
cysteine,5,6 or glutamate.7
INTRODUCTION
The tripeptide glutathione (ꢀ-glutamyl–cysteinyl–glycine,
GSH) is present at millimolar concentrations in the intracel-
lular milieu in most tissues, and plays a prominent role in
the defense against oxidative stress, as well as in the detoxi-
fication of many xenobiotics.1 It arises from de novo synthesis
from its constituent amino acids, glutamate, cysteine, and
glycine, and has a very high turnover rate in most tissues.
GSH depletion has been described in various disease
states,2–4 and could arise from either a decreased rate of
synthesis, increased rate of utilization, or a combination
of both. To delineate the mechanisms responsible for the
alterations of GSH concentrations, determination of its rate
of synthesis is of physiological interest, and several groups
Jahoor et al.8 were first to use preparative high-
performance liquid chromatography (HPLC) to separate
blood GSH, then followed by hydrolysis of the tripeptide,
to determine isotopic enrichments in GSH-bound glycine
or glutamate by gas chromatography–mass spectrometry
(GCMS). This approach was successful in determining the
synthesis rate of GSH over the course of labeled glycine
or glutamate infusion, and applied to the investigation of
important aspects of the regulation of GSH homeostasis.8,9
The method is, however, somewhat tedious as it involves the
use of two separate analytical techniques, i.e. HPLC followed
by GCMS. In addition, when labeled glutamate is used as a
tracer to quantify GSH synthesis, accurate determination of
the isotope enrichment in free glutamate is required to assess
the isotope enrichment in the precursor pool used for GSH
synthesis. As biological samples contain both glutamine and
glutamate, an accurate determination of labeled glutamate
enrichment is often hampered by the spontaneous hydrolysis
ŁCorrespondence to: D. Darmaun, INSERM U. 539, Hoˆtel-Dieu,
HNB 3e e´tage nord, 44093 Nantes cedex 1, France.
E-mail: ddarmaun@nantes.inserm.fr
Contract/grant sponsor: Nemours Research Programs.
Contract/grant sponsor: European Society for Parenteral and
Enteral Nutrition (ESPEN).
Contract/grant sponsor: Socie´te´ Francophone de Nutrition Ente´rale
et Parente´rale (SFNEP).
Contract/grant sponsor: Conseil Re´gional des Pays-de-la-Loire.
DOI: 10.1002/jms.185
Copyright  2001 John Wiley & Sons, Ltd.

Glutathione isotope enrichment and level by GCMS 727
of glutamine to glutamate in vitro during sample storage and
processing. Although the method described by Reeds et al.9
was aimed at minimizing the degradation of glutamine by
using a ‘‘soft’’ derivatization procedure to reduce the ‘‘con-
tamination’’ of glutamate by degraded glutamine, the assay
did not provide a correcting mechanism in the case when
glutamine degradation was not negligible.
More recently, Capitan et al.5 used the direct deriva-
tization of GSH extracted from blood and tissue to its
ethoxylcarboxymethyl ester to measure the incorporation
of labeled cysteine into blood and tissue GSH with a simple
GCMS assay. The method, however, did not allow for a
simultaneous determination of the stable isotope enrichment
and concentration of blood GSH in the same sample.
The current report describes an improved approach, in
as much as: (a) the GCMS assay makes use of an internal
standard to allow for the simultaneous determination of
the stable isotope enrichment and concentration of blood
GSH; (b) the preparation of labeled GSH in vitro allows
for the use of a standard curve in the determination of
labeled GSH enrichments; and (c) the method can be used to
measure accurately the ‘‘true’’ enrichment in the precursor
free glutamate pool despite the potential ‘‘contamination’’
of glutamate by glutamine hydrolysis in vitro when L-[15N]
glutamic acid is used as a tracer to assess in vivo GSH kinetics.
the supernatant was removed and transferred into a 3 ml
glass tube, and adjusted to pH 7.5 with 0.8 mol lꢀ1 NaOH
before derivatization.
Preparation of standard solutions
For GSH concentration determination, the tripeptide hGSH
(glutamyl–cysteinyl–alanine)—an analogue of GSH that is
exclusively synthesized in plants, and is not found in native
human blood—was chosen as an internal standard because
of its structure closely resembling that of GSH. To prepare a
standard curve for the assay of GSH concentrations, graded
amounts of natural GSH were added to 1 ml of phosphate
buffer containing 1 µmol of hGSH and 80 µmol of DTT,
to obtain GSH/hGSH molar ratios ranging between zero
and two.
For GSH enrichment determination, a 15N-labeled GSH
was produced in vitro in the laboratory, by incubating 5 ml
°
of hemolyzed human red blood cells, for 2 h at 37 C with
300 µmol L-[15N] glutamic acid, 200 µmol L-cysteine, 200 µmol
L-glycine, 200 µmol ATP, 200 µmol D-glucose, 1.5 mmol
Tris–HCl, 300 µmol magnesium chloride, in 10 ml distilled
water at pH 7.5.10 After incubation, protein was precipitated
by adding 1.5 ml of 50% (w/v) SSA. The mixture was
°
then centrifuged for 15 min at 3000 g and at 10 C, the
supernatant removed and filtered, and its pH adjusted up
to 7.5 with 10 mol lꢀ1 NaOH before the solution could be
°
stored at ꢀ80 C. The concentration of this final solution
EXPERIMENTAL
was assayed before each use by adding hGSH as internal
standard as described above. Its GSH concentration was
623 š 25 µmol lꢀ1, and, assuming a slope of unity for the
determination of the 364/363 ion ratio on GCMS, the 15N-
GSH/natural GSH mole ratio was 0.365 š 0.007.
Reagents
Homoglutathione (hGSH) was obtained from BACHEM
Biochimie (Voisins le Bretonneux, France). Reduced glu-
tathione (GSH), ethyl chloroformate, homoglutamate (hglu),
sulfosalicylic acid (SSA) and dithiothreitol (DTT) were
obtained from Sigma (Sigma–Aldrich, Steinheim, Germany).
L-[15N] glutamic acid (15N 95–99%) and L-[U-13C] glutamine
(13C 99%) were obtained from Cambridge Isotope Laborato-
ries (Andover, MA, USA). AG50 cation-resin was obtained
from Aldrich (Sigma–Aldrich). Heptafluorobutyric anhy-
dride (HFBA) was obtained from Fluka (Sigma–Aldrich).
GSH derivatization
The derivatization procedure was adapted from Kataoka
et al.11 and Capitan et al.,5 and converted GSH and hGSH to
their respective N,S-ethoxycarbonyl methyl (NSECM) esters.
After the pH was adjusted to 7.5, 200 µl of ethyl chloroformate
were added, and the mixture was shaken for 10 min at
room temperature. The pH was then adjusted to 1.5 with
1 mol lꢀ1 HCl, the mixture transferred into a 10 ml glass tube
containing 0.5 g of NaCl, and extracted twice with 4 ml of
peroxide-free diethyl ether. The organic phase extract was
GSH analysis
Preparation of red blood cells
Venous blood was collected in 5 ml (ethylenediaminete-
traacetic acid) (EDTA) tubes, immediately put on ice, and
centrifuged at 5000g for 10 min to separate the red cell pel-
let from plasma. After centrifugation of whole blood, a fine
indelible ink line was drawn on the tube wall to mark the
plasma level. Plasma was then removed, and replaced with
cold distilled water drop by drop up to the mark, in order
to achieve hemolysis of the red blood cells. After this step,
°
evaporated to dryness under nitrogen flux at 50 C. 250 µl
of 1 mol lꢀ1 HCl in methanol (prepared freshly by mixing
250 µl of 36% HCl into 7.5 ml methanol) were added, and the
mixture was incubated for 10 min at 80 C. After cooling, the
mixture was evaporated under nitrogen flux and the residue
subsequently dissolved in 400 µl of ethyl acetate for injection
into the gas chromatography–mass spectrometer.
°
°
samples could be stored at ꢀ80 C for several weeks, until
analysis. For analysis, 1.5 ml of hemolyzed sample was trans-
ferred into a 5 ml glass tube containing an 850 µl aliquot of
0.2 mol lꢀ1 phosphate buffer with 1.5 µmol of hGSH as inter-
nal standard and 40 µmol of DTT; the pH was adjusted to
8.5 with 0.8 mol lꢀ1 NaOH, and the mixture was left 15 min
at room temperature. Samples were then deproteinized by
adding 375 µl of 50% (w/v) SSA and 375 µl of 0.2 mol lꢀ1
GCMS parameters
Analysis was carried out on an electron impact ionization gas
chromatograph–mass selective detector (5890 series II gas
chromatograph coupled with a 5971 mass selective detector,
Hewlett-Packard, Palo Alto, CA, USA) using an HP-5MS
capillary column (10 m ð 0.25 mm internal diameter (id),
0.1 µm film thickness, Hewlett-Packard) operated in the split
mode with a 20 : 1 split. Injector and detector temperatures
°
phosphate buffer. After centrifugation (3000 g, 15 min, 10 C)
Copyright  2001 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2001; 36: 726–735

728 B. Humbert et al.
°
°
were 270 C and 280 C respectively; the column head
Preparation of standard solutions
pressure was 3 psi, and the helium flux was set at 1 ml minꢀ1
.
Three sets of standards are required with this method.
The first set of standards was designed to assess glu-
tamine concentration and its degradation to glutamate, and
was prepared by mixing graded amounts of [U-13C] glu-
tamine and natural glutamine in order to obtain [U-13C]
glutamine/glutamine mole ratios ranging between 0 and
15%. The second set of standards was designed to mea-
sure the sum of ‘‘true’’ glutamate plus glutamate derived
from glutamine degradation, and was obtained by adding
graded amounts of natural glutamate to homoglutamate,
in order to obtain mole ratios in the 0–6% range. The
third set of standards was designed to assess 15N-glutamate
enrichment and was obtained by preparing L-[15N] glu-
tamic acid/natural glutamic acid mixtures with mole ratios
ranging between 0 and 20%. A 200 µl aliquot of each stan-
dard was evaporated to dryness under vacuum before
derivatization.
The volume injected was 1–2 µl. The oven temperature was
°
°
initially set to 150 C, and maintained at 150 C for 30 s
ꢀ1
°
°
after injection, then raised to 275 C at a rate of 15 C min
,
ꢀ1
°
°
and finally ramped up at the rate of 40 C min to 300 C,
where it was maintained for 3 min. The acquisition time
was 12 min. These settings allow a sufficient separation of
the GSH and hGSH. Using the selective ion monitoring
(SIM) mode of the mass detector, with a dwell time set
at 100 ms, ions at mass-to-charge ratio m/z D 363 and
364, representing natural and 15N-GSH respectively, were
selectively monitored for the GSH peak, and the ion at
m/z D 363 for the hGSH peak.
Glutamate and glutamine analysis
To determine red blood cell GSH fractional synthesis rate
over the course of an infusion of L-[15N] glutamic acid, the
15N-enrichment in intracellular red blood cell glutamate,
the precursor pool for erythrocyte GSH enrichment, must be
measured accurately. Most biological samples, however, con-
tain both glutamine and glutamate. As glutamine’s amide
nitrogen is spontaneously lost during sample storage and
processing in vitro, a fraction of the glutamine present in
the original sample can degrade, and result in the ‘‘contam-
ination’’ of ‘‘true’’ glutamate with glutamate derived from
degraded glutamine. Although our previously published
method12 quantified the ‘‘contamination’’ of glutamate by
deamidated glutamine, the separation of glutamine from
glutamate was tedious, and the glutamine and glutamate
fractions extracted from biological samples had to be treated
and analyzed separately. We therefore developed a method
that allowed for the concomitant assessment of glutamate
and glutamine enrichments and concentrations, while still
allowing for quantitation of glutamine degradation in the
same sample. Homoglutamate, a six-carbon analog of glu-
tamate, was used as internal standard for glutamate, and
[U-13C] glutamine as an internal standard for glutamine. The
derivatization procedure used was previously described by
Reeds et al.9
Derivatization
The derivation procedure converted glutamine and gluta-
mate to their heptafluorobutyramide propyl (PHFBA) ester
derivatives. The procedure we used produces the same
derivative for glutamate as described by Knapp13 and
applied by Matthews and Campbell,14 but the use of a lower
temperature prevents complete glutamine degradation to
glutamate as described by Reeds et al.9
A 600 µl amount of 5/1 (v/v) propan-1-ol to acetylchlo-
ride mixture was added to dry samples or standard solutions.
The mixture was incubated 2 h at room temperature,9 then
evaporated to dryness under a nitrogen stream. A 50 µl
amount of heptafluorobutyric anhydride (HFBA) was added,
°
tubes were shaken, capped, incubated 20 min at 60 C in
heating block, and evaporated under nitrogen stream. The
residue was then dissolved in either 2 ml or 10 ml of ethyl
acetate for samples or standard solutions respectively.
GCMS parameters
Analyses were carried out using electron impact ionization
GCMS with a DB-1 capillary column (30 m ð 0.25 mm id,
0.25 µm film thickness, J&W Scientific, Folsom, CA, USA)
in the splitless mode. Injector and detector temperatures
Preparation of red blood cells
°
were 250 and 280 C respectively; the column head pressure
Red blood cells were hemolyzed as described above for
GSH analysis. 500 µl of hemolyzed cells were spiked with
20 nmol [U-13C] glutamine as an internal standard to
quantify glutamine degradation to glutamate, and 75 nmol
homoglutamate to quantify the sum (glx) of ‘‘true’’ glutamate
plus glutamine-derived glutamate. Protein was precipitated
with 300 µl of 25% (w/v) SSA, and samples were centrifuged
was 8 psi, and the helium flux was set at 1 ml minꢀ1. The
°
oven temperature was initially set at 80 C, and then ramped
to 250 C at the rate of 15 C minꢀ1. The acquisition time
was 13 min. The volume injected was 1 µl. These settings
allowed a sufficient separation of glutamine, glutamate and
homoglutamate (at 7.3, 8.9 and 9.7 min, respectively). Ions at
m/z D 252 (natural glutamate, glutamine-derived glutamate
and homoglutamate), 279 (natural glutamine), 280 (natural
glutamate and 15N-glutamine), 281 (15N-glutamate), 283 ([U-
13C] glutamine) and 284 ([U-13C] glutamate derived from
[U-13C] glutamine degradation) were monitored using the
SIM mode, with a dwell time of 30 ms.
°
°
°
for 20 min at 3000g at 4 C. The supernatant was transferred,
adjusted to pH 3, poured on top of a disposable plastic
column containing 1.5 ml of AG50 ion exchange resin,
and washed with 10 ml distilled water. After applying
the sample, the column was washed a second time with
10 ml cold distilled water, and the eluate discarded. 2 ml
of 6 mol lꢀ1 ammonium hydroxide were added to each
column to elute both glutamate and glutamine, and the
eluate collected, and evaporated to dryness under vacuum
before derivatization.
Calculations
GSH concentration ([GSH]/µmol lꢀ1) was determined from
the peak area ratio of the ion at m/z D 363 in the GSH peak
Copyright  2001 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2001; 36: 726–735

Glutathione isotope enrichment and level by GCMS 729
to the ion at the same m/z D 363 in the hGSH peak:
The fraction of glx that is glutamate derived from
glutamine degradation (fdegr), was calculated using:
[GSH] D nhGSH ð [ꢁ363GSH/363hGSHꢂ
ꢀ b1]/ꢁa1 ð Hcrit/100ꢂ
fdegr D ꢁ284glx/280glxꢂ/a2
where 284glx/280glx is the ratio of ꢁm C 4ꢂ/m peak area
assessed by GCMS in the glx fraction.
Glutamate concentration ([glu]/µmol lꢀ1) was calculated
using:
where Hcrit is the hematocrit, defined as the fraction of
whole blood volume accounted for by red blood cells (and
expressed as a percentage); nhGSH the number of nanomoles
of hGSH added as internal standard per milliliter of sample;
363GSH/363hGSH is the peak area ratio of GSH/hGSH at
m/z 363; a1 and b1 are the slope and the intercept respectively
of the standard curve obtained by adding graded amounts
of natural GSH in hGSH aqueous solutions; and 100 converts
the hematocrit into a fraction of unity.
[glu] D [glx] ð [1 ꢀ fdegr]
The 15N-enrichment in the glutamate peak (glx, a
mixture of true glutamate plus degraded glutamine), E
15Nglx, expressed in MPE, was calculated as follows.
The 15Nglx/natural glx mole ratio (MR 15Nglx) was first
calculated as:
Glutamine concentration ([gln]/µmol lꢀ1) was calculated
using:
[gln] D a2 ð [U-13C-gln]/[ꢁ283gln/279gln ꢀ b2ꢂ ð Hcrit/100]
MR 15Nglx D ꢁ281glx/280glx ꢀ b3ꢂ/a3
where a2 and b2 are the slope and the intercept respectively of
the standard curve obtained by adding increasing amounts
of [U-13C]glutamine in natural glutamine, [U-13C-gln] is
the concentration of [U-13C] glutamine (added internal
standard) in the sample, expressed in micromoles per
liter, and 283gln/279gln is the ratio of ꢁm C 4ꢂ/m in the
glutamine peak.
The 15N-enrichment in erythrocyte glutamine (E 15Ngln,
expressed in mole percent excess; MPE) was calculated as
follows. First, the 15N-glutamine/natural glutamine mole
ratio at time t (MR 15Ngln) was calculated as:
where 281glx/280glx is the peak area ratio of ion 281 to 280
in the glutamate plus glutamine-derived glutamate fraction
assessed by GCMS, and b3 and a3 are the intercept and slope
of the corresponding standard curve respectively.
Erythrocyte 15N-glx enrichment (E 15Nglx) was then
calculated as:
MR 15Nglxt ꢀ MR 15Nglxt0
E
15Nglx D
ð 100
MR 15Nglxt ꢀ MR 15Nglxt0 C 1
where MR 15Nglxt and MR 15Nglxt0 are the 15Nglx/natural
glx mole ratios measured before the tracer infusion (at
baseline) and at isotopic plateau respectively.
15N-glutamate enrichment (E 15Nglu/MPE) was calculated
using:
MR 15Ngln D ꢁ280gln/279gln ꢀ b3ꢂ/a3
t
where 280gln/279glnt is the 280/279 peak area ratio
measured by GCMS in erythrocyte glutamine peak at time t,
and a3 and b3 are the slope and the intercept respectively of
the standard curve obtained by adding increasing amounts
of 15N-glutamate in natural glutamate. The 15N-enrichment
in erythrocyte glutamine at time t was then calculated using:
[glx] ð E 15Nglx ꢀ fdegr ð [glx] ð E 15Ngln
E
15Nglu D
[glu]
where E 15Nglx and E 15Ngln were the 15N enrichments in the
glutamate plus glutamine-derived glutamate fraction and in
the glutamine fraction, respectively, expressed in MPE.
The GSH fractional synthesis rate (FSR/%dꢀ1), is calculated
as follows. The 15N-GSH/natural GSH mole ratio (MR
15Ngsh) was calculated using:
E
15Ngln D ꢁMR 15Nglnt ꢀ MR 15Nglnt0ꢂ/ꢁMR 15Nglnt
ꢀ MR 15Nglnt0 C 1ꢂ ð 100
where MR 15Nglnt and MR 15Nglnt0 are the molar ratios
of 15N in glutamine at isotopic plateau and before L-[15N]
glutamic acid infusion (i.e. at baseline) respectively.
The concentration of the sum of ‘‘true’’ glutamate plus
glutamate arising from glutamine degradation to glutamate
([glx]/µmol lꢀ1) was calculated thus.
MR 15Ngsh D ꢁ364gsh/363gsh ꢀ b5ꢂ/a5
where 364gsh/363gsh is the peak ratio of ion 364 to 363 in
the GSH peak, and a5 and b5 are the slope and intercept
respectively of the standard curve obtained by adding
increasing amounts of 15N-GSH in natural GSH. The 15N-
enrichment in plasma GSH at time t (E 15Ngsh) was then
calculated using:
[glx] D 252glx/252hglu ð nhglu/ꢁa4 ð Hcrit/100ꢂ
where 252glx/252hglu is the peak area ratio of ion 252 in
the glutamate plus glutamine-derived glutamate peak (‘‘glx
peak’’) to ion 252 in the homoglutamate peak, nhglu is the
number of nanomoles of homoglutamate added as internal
standard per milliliter of sample, and a4 is the slope of the
standard curve obtained by adding increasing amounts of
natural glutamate in homoglutamate.
MR 15Ngsht ꢀ MR 15Ngsht0
E
15Ngsh D
ð 100
MR 15Ngsht ꢀ MR 15Ngsht0 C 1
where MR 15Ngsht and MR 15Ngsht0 are the molar ratios
of 15N in glutathione at time t and before 15N-glutamate
infusion respectively.
Copyright  2001 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2001; 36: 726–735

730
B. Humbert et al.
The GSH FSR was then defined as:
the recovery of added GSH reached 103.3 š 5.4%, which did
not differ significantly from 100% (Table 3).
FSR D 100 ð 24 ð E 15Ngsh/ꢁt ð E 15Ngluꢂ
GSH enrichment
where E 15Ngsh (MPE hꢀ1) is the rise in erythrocyte 15N-gsh
enrichment during the time interval t (expressed in hours),
when erythrocyte-free 15N-glutamate enrichment (E 15Nglu/
MPE) was at steady state, e.g. between the end of the second
and fifth hours of isotope infusion; 24 converts FSR from hꢀ1
to dayꢀ1, and 100 converts a fraction of unity to a percentage.
The 15N-enrichment in erythrocyte GSH from both in vivo
tracer studies and in vitro synthesized labeled GSH was
measured with within-day and day-to-day coefficients of
variation of less than 3% (Tables 1 and 2). A linear standard
curve was obtained when increasing amounts of 15N-
glutathione were added to natural glutathione (Table 4).
The intercept value represents the m C 1/m peak area ratio
in unlabeled GSH. This value can be predicted and cannot
In vivo experiment
2
be expected to be zero, due to the presence of 13C, H, 17O,
A single dog, from the kennel of the National Veterinary
School of Nantes, was studied after approval of the
experimental procedure by the ethical committee of the
National Veterinary School.
15N and 33S at their natural abundance in natural GSH. For
instance, for the ion at m/z D 363 in the GSH derivative,
the chemical formula is: C13H23O7N2S. We calculated the
predicted percent natural abundance of ion at m/z D 364 in
GSH as follows:
The synthesis rate of erythrocyte GSH was measured
in vivo in a single Beagle dog on two separate occasions:
(i) after an overnight physiological fast; and (ii) after a 72 h
fast. On the day of isotope infusion at 8 : 30 am the dog was
weighed, and two short intravenous catheters (Vasocan
20 gauge, BBraun Medical, Emmenbru¨cke, Germany) were
placed using an aseptic technique: one in the cephalic vein
of the forelimb for isotope infusion, and the other in the con-
tralateral forelimb for blood sampling. At 8 : 45 am, a baseline
blood sample was obtained to determine background isotope
enrichment and concentration of erythrocyte GSH and free
glutamate. Starting at 9 : 00 am, the dog received a priming
dose of L-[15N] glutamic acid ꢁ45 µmol kgꢀ1ꢂ, immediately
followed by a continuous 5.5 h intravenous infusion of
L-[15N] glutamic acid ꢁ45 µmol kgꢀ1 hꢀ1ꢂ delivered by means
of a calibrated Bioblock (Fisher Scientific, Illkirch, France)
syringe-pump. 5 ml venous blood samples were obtained
in EDTA tubes at 90, 120, 150, 180, 210, 240, 270, 300, and
330 min after the start of isotope infusion. Tubes were kept
13 ð 1.11 C 23 ð 0.015 C 7 ð 0.037
C2 ð 0.37 C 1 ð 0.76 D 17.6%
i.e. 0.176 when expressed as a fraction, where 13, 23, 7,
2 and 1 are the numbers of atoms of carbon, hydrogen,
oxygen, nitrogen, and sulfur respectively in the compound,
and 1.11, 0.015, 0.037, 0.37, and 0.76 are the percentage
natural abundances of 13C, 2H, 17O, 15N and 33S respectively.
Therefore, the experimental value we measured (0.171) is not
different from this predicted value.
Glutamate concentration
Figure 3 shows the typical chromatogram and spectrum of
erythrocyte glutamate and homoglutamate. Linear standard
curves were obtained when homoglutamate was used to
assay glutamate concentration (Table 4). We verified that the
degradation of glutamine to glutamate during storage and
in vitro sample processing was not negligible, as degraded
glutamine accounted for as much as 22% of the erythrocyte
‘‘glx’’ peak (i.e. ‘‘true glutamate’’ C glutamate arising from
degraded glutamine) in some instances.
With the appropriate correction for glutamine degra-
dation—allowed by the use of [U-13C] glutamine internal
standard—however, the precision of the ‘‘true’’ glutamate
assay, as defined by its within-day coefficient of variation,
was 0.4% (Table 1). When hemolyzed red blood cell samples
were spiked with known amounts of natural glutamate, the
recovery of added glutamate was not different from 100%
(Table 3).
°
on ice until centrifugation at 4 C at 5000g for 10 min. Plasma
was immediately removed and an equivalent volume of cold
distilled water was added. The mixture was shaken, and
°
immediately frozen at ꢀ80 C. Hematocrit was determined
in triplicate in baseline samples for further calculations, by
centrifuging blood aliquots in capillary tubes at 5000 g, for
5 min. Hematocrit (expressed in percent) was calculated as
the length of the red blood cell fraction in the capillary tube,
divided by the total length ꢁplasma C red blood cellsꢂ, as
measured after the end of the centrifugation.
RESULTS
GSH concentration
As shown in Figs 1 and 2, the GSH and hGSH peaks were
free of interfering peaks on the chromatogram. Under the
chromatographic conditions used, GSH eluted just before
hGSH, but their separation was sufficient to measure their
respective ion intensities separately without any cross-
contamination (Fig. 2). Standard curves were linear over
the range of GSH/hGSH molar ratios chosen (Fig. 2). The
within-day and day-to-day coefficients of variation were less
than 5% (Tables 1 and 2). When hemolyzed red blood cell
samples were spiked with known amounts of natural GSH,
Glutamate enrichment
The 15N-enrichment in erythrocyte glutamate was measured
with a within-day coefficient of variation of less than 4%
(Table 1).
Effects of fasting on GSH synthesis in red blood
cells in one dog
A 72 h fast was associated with a 39% decrease in red blood
cell GSH concentration without any alteration in glutamine
and glutamate concentrations (Table 5). Over the course of
Copyright  2001 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2001; 36: 726–735

Glutathione isotope enrichment and level by GCMS 731
Abundance
TIC
140000
130000
120000
110000
100000
90000
80000
70000
60000
50000
40000
30000
20000
10000
Glutathione
Homoglutathione
0
5.50
6.00
84
6.50
7.00
7.50
8.00
Scan
8.50
(7.606 min)
9.00
9.50
10.00 10.50 11.00 11.50
Time->
Abundance
13000
Glutathione:
148
12000
11000
10000
9000
8000
7000
6000
5000
4000
3000
2000
1000
m/z = 479
C − NH − CH − CO − NH − CH2 − CO − O − CH3
CH2
CH2
CH2
S − CO − O − C2H5
m/z = 363
m/z = 216
CH − NH − CO − O − C2H5
CO − O − CH3
216
56
110
74
317
331
363
98
174
128
159
160
0
m/z->
60
80
84
100
120
140
180
Scan
200
220
240
260
280
300
320
340
360
Abundance
(7.893 min)
Homoglutathione:
14000
m/z = 493
12000
10000
8000
6000
4000
2000
C − NH − CH − CO − NH − CH − CO − O − CH3
CH2
CH2
CH2
CH3
S − CO − O − C2H5
216
m/z = 363
m/z = 216
148
CH − NH − CO − O − C2H5
CO − O − CH3
55
104
77
188
363
259
128
116
317
94
159170
160
345
0
m/z->
60
80
100
120
140
180
200
220
240
260
280
300
320
340
360
Figure 1. Typical chromatogram (upper panel) and mass spectra of N,S-ethoxycarbonyl methyl ester of GSH (middle panel) and
hGSH (bottom panel) in red blood cell sample analyzed by electron impact ionization GCMS operating in the scan mode.
the L-[15N] glutamic acid infusion, erythrocyte 15N-glutamate
enrichment reached a plateau by the end of the second hour of
tracer infusion, and 15N-GSH enrichment increased linearly
after the 15N-glutamate enrichment plateau had been reached
(Fig. 4). The slopes of the lines describing the rise in 15N-GSH
enrichment did not differ between the two experimental days
Copyright  2001 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2001; 36: 726–735

732
B. Humbert et al.
ion 363,00 (362,70 to 363,70)
Abundance
2000
Table 2. Day-to-day inter-assay coefficient of variation (CV)
for GSH and glutamate concentration, and for GSH enrichment
in red blood cell hemolyzates. Data are expressed as
mean š SE and coefficients of variation between different days
of assessment of the same sample
1800
1600
Homoglutathione
Glutathione
1400
1200
1000
800
600
400
200
0
Number Number
CV
of
of
Mean š SE (%) assays injections
[GSH] (µmol lꢀ1
)
743 š 13
65.1 š 0.4
310 š 2
4.7
1.8
0.8
7
8
2
2
2
6
Time ->
5,50 6,00 6,50 7,00 7,50 8,00 8,50 9,00 9,50 10,00 10,50 11,00 11,50
Enr 15N-GSH (MPE)
[glutamate] (µmol lꢀ1
)
4
y = 1,7755x - 0,0988
R
2 = 0,9975
3
2
1
0
Table 3. Recovery of added GSH and glutamate in red blood
cell mixture. Recovery is defined as the number of micromoles
in excess in the sample divided by the number of micromoles
added to the original sample, and expressed as a percentage.
Values are means š SE. (n D 7 for each added amount of GSH
tested, and n D 2 for each added amount of glutamate tested)
0
0.5
1
1.5
2
Molar ratio (GSSH /hGSH)
Figure 2. Upper panel: typical chromatogram of
Glutathione
Glutamate
N,S-ethoxycarbonyl methyl esters of GSH and hGSH in red
blood cell sample analyzed by electron impact ionization
GCMS in the SIM mode. Bottom panel: typical standard curve
obtained by adding graded amounts of GSH in hGSH in a
molar ratio in the range of 0–2. The ion selected was m/z 363
in both GSH and hGSH peaks.
GSH added
ꢁµmolꢂ
Recovery
(%)
Glutamate added
Recovery
(%)
ꢁµmolꢂ
0.5
1
98.0 š 3.1
0.125
0.25
0.5
99.9
98.1
106.5
108.7 š 9.5
Mean
103.3 š 5.4
Mean
101.6 š 2.4
Table 1. Typical precision of the assays: the within-day
(intra-assay) coefficient of variation (CV) is given for the assay
of erythrocyte GSH and glutamate concentrations ([GSH] and
[glutamate] respectively), and for 15N-enrichment (Enr)
determination in GSH and in glutamate. Data are expressed as
mean š standard error
FSR (Table 5). For this calculation, we hypothesized that the
slope of the standard curve used for the measurement of
15N-GSH-to-natural GSH mole ratio was not different from
unity and that this slope did not differ between the different
analysis days. We verified the second hypothesis through
repeated measurements on aliquots of the 15N-enriched GSH
mixture synthesized in the laboratory (Tables 2 and 4).
Number Number of
CV
of
injections
Assay
Mean š SE (%) samples per sample
[GSH] (µmol lꢀ1
)
1982 š 18
64.5 š 0.7
351 š 1
1.8
2.2
0.4
4
4
4
3
3
2
3
3
Enr 15N-GSH (MPE)
[glutamate] (µmol lꢀ1
Enr 15N-glutamate
(MPE)
DISCUSSION
)
The current report describes a method to assess both GSH
isotope enrichment and concentration simultaneously in the
same biological sample, and in the same run, using GCMS.
In addition, we describe an approach to assess glutamate
enrichment and concentration. Finally, we describe a simple
procedure for the preparation of highly enriched stable-
isotope-labeled GSH to produce calibration standards for
8.68 š 0.17 3.4
(0.41 MPE hꢀ1 versus 0.37 MPE hꢀ1, fasting versus control
respectively), and fasting failed to affect red blood cell GSH
Table 4. Slope and intercept of the standard curves used for determination of GSH and glutamate
concentrations and enrichments. Results are expressed as mean š SE
Standard curve
Slope
Intercept
Correlation coefficient
Number of analyses
[GSH] ꢁa1, b1ꢂ
[U-13C] gln ꢁa2, b2ꢂ
15N-glx ꢁa3, b3ꢂ
[glx] ꢁa4ꢂ
1.724 š 0.055
0.933 š 0.038
1.069 š 0.020
0.639 š 0.088
0.977 š 0.019
ꢀ0.090 š 0.019
0.000 š 0.001
0.113 š 0.008
0.000 š 0.000
0.171 š 0.002
0.998 š 0.000
0.997 š 0.001
0.999 š 0.000
0.992 š 0.006
1.000 š 0.000
9
4
3
3
2
15N-GSH ꢁa5, b5ꢂ
Copyright  2001 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2001; 36: 726–735

Glutathione isotope enrichment and level by GCMS 733
TIC:
Abundance
700000
650000
600000
550000
500000
450000
400000
350000
300000
250000
200000
150000
100000
50000
Homoglutamate
Glutamate
0
0
12.50
4.50 5.00 5.50 6.00 6.50 7.00 7.50 8.00 8.50 9.00 9.50 10.00 10.50 11.00 11.50 12.00
Time->
Abundance
45000
Scan
(8.910 min)
252
H7C3 − O − ΟC − CH2 − CH2 − CH − CO − O − C3H7
40000
35000
30000
25000
20000
15000
10000
5000
N − CO − C3F7
H
m/z = 340
m/z = 280
m/z = 252
280
PHFBA-Glutamate:
m/z = 427
85
340
298
69
239
240
169
57
367
360
100
100
266
260
115
128
313 326
214
220
0
m/z->
80
120
140
160
180
200
280
300
320
340
60
Scan
(9.682 min):
280
Abundance
H7C3 − O − ΟC − CH2 − CH2 − CH2 − CH − CO − O − C3H7
140000
130000
120000
110000
100000
90000
80000
70000
60000
50000
40000
30000
20000
10000
0
N − CO − C3H7
H
252
m/z = 294
m/z = 252
m/z = 354
PHFBA-homoglutamate:
m/z = 441
354
169
67
226
295
98
382
82
128
313
300 320
520
493
197
146
440 460 480 500 520
420
340 360 380 400
280
220 240 260
60 80 100 120 140 160 180 200
m/z->
40
Figure 3. Typical chromatogram and mass spectra of glutamate (upper scan) and homoglutamate (bottom scan)
heptafluorobutyramide propyl ester derivatives, as analyzed by electron impact ionization GCMS operating in the scan mode.
the assay of labeled GSH enrichment. Taken together, these
methods allow determination of GSH concentration and
synthesis rate in red blood cells or other tissues, using a 5.5 h
L-[15N] glutamic acid tracer infusion in vivo.
When we applied the current assay to the measurement
of erythrocyte GSH in one dog, we found a concentration
of 4.7 mmol lꢀ1. Although this value is higher than the
2.6 mmol lꢀ1 values previously measured in Beagle dogs
Copyright  2001 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2001; 36: 726–735

734
B. Humbert et al.
Table 5. Concentrations and stable isotope enrichments of
GSH, glutamate (glu), glutamine (gln), and
glutamate C glutamine-derived fraction (glx), and GSH FSR in
red blood cells, during an infusion of L-[15N] glutamic acid in a
single dog after overnight fast (control), and after 72 h of fasting
3
2
1
0
Control
51.3
Fasting
44.0
L-[15N] glutamic acid infusion
rate ꢁµmol kgꢀ1 hꢀ1
ꢂ
[U-13C] gln/natural gln molar
ratio
0.132 š 0.003 0.125 š 0.003
0.024 š 0.001 0.035 š 0.001
0
2
4
6
[U-13C] glx/natural glx molar
ratio
control
fasting
15N-glx enrichment (MPE)
15N-gln enrichment (MPE)
10.4 š 0.3
1.85 š 0.33
194 š 7
11.0 š 0.3
1.55 š 0.30
192 š 14
138 š 9
20
15
10
5
[glx] ꢁµmol lꢀ1
[glu] ꢁµmol lꢀ1
ꢂ
ꢂ
159 š 7
15N-glu enrichment (MPE)
15N-GSH/natural GSH molar
ratio at time 0
12.2 š 0.3
0.0059
14.5 š 0.3
ꢀ0.0028
15N-GSH/natural GSH molar
ratio at time t
0.0252
0.0176
0
t (h)
5.5
4.5
GSH fractional synthesis rate
71.3
67.4
0
2
4
6
(% dayꢀ1
)
Time course (h)
[GSH] ꢁmmol lꢀ1
[glutamine] ꢁµmol lꢀ1
ꢂ
4.66 š 0.52
670 š 16
2.87 š 0.40
653 š 14
Figure 4. Time course of erythrocyte 15N-GSH (upper panel)
and 15N-glutamate (bottom panel) enrichments (enr) during an
infusion of L-[15N] glutamic acid after overnight physiological
fasting (control), and after a 72 h fasting period (fasting) in the
same dog.
ꢂ
by Vajdovich et al.15 using an enzymatic assay, there are
many differences in experimental design between that study
and the current report. For instance, in that earlier study the
dogs received a diet containing 75% protein, versus 25% in
the current study. In addition, Vajdovich et al. documented
higher glutathione levels in female dogs, compared with
male dogs, and the single dog used in the current report was
female.
in red blood cell GSH content, without any alteration in
GSH fractional synthesis rate. This result is consistent with
the view that (a) fasting may have transiently increased
GSH breakdown, resulting in lower erythrocyte level, and
(b) the failure of a rise in GSH synthesis prevented the
repletion of the erythrocyte GSH pool, presumably due to
the discontinuation of dietary protein and energy intake.
Although circulating amino acid levels were not determined
in the current study, lack of protein intake may indeed limit
the availability of the precursor amino acids glycine, cysteine
and glutamate during the 72 h fast. As the current results
are, however, based on a single dog, they should only be
construed as very preliminary data, and the effect of fasting
on dog erythrocyte GSH synthesis would clearly warrant
further study.
To our knowledge, the current study is the first to report
on the fractional synthesis rate of erythrocyte GSH in dogs.
Although the 67–71% dayꢀ1 GSH FSR values obtained in
the current study in dogs differ from the value (23% dayꢀ1
)
reported in one study carried out in a single human using
[2H2] cysteine,5 this discrepancy is unlikely to result from
either interspecies or methodological differences. As a matter
of fact, two other studies carried out with [2H2] glycine,4 and
L-[1-13C] cysteine6 report erythrocyte GSH FSR of 71 or 65%
dayꢀ1 in humans, i.e. very close to the findings of the current
study.
Acknowledgments
This study was supported, in part, by grants from Nemours
Research Programs, Jacksonville, Florida (USA). Bernard Humbert
was supported, in part, by grants from the European Society for
Parenteral and Enteral Nutrition (ESPEN), the Socie´te´ Francophone
de Nutrition Ente´rale et Parente´rale (SFNEP), and the Conseil
Re´gional des Pays-de-la-Loire. We acknowledge the superb technical
help of Pascale Mauge´re.
The current report also describes a simple and rapid tech-
nique to synthesize in vitro an inexpensive 15N-GSH-enriched
mixture, useful for calculation of 15N-GSH enrichments and
normalization of the results independently of the day of the
measurement.
Although fasting has been shown to decrease GSH
levels in several tissues, such as intestine,16,17 liver, heart,
and muscle18 in rodents, the effect of food deprivation
had not, to our knowledge, been tested on GSH synthesis
in canine red blood cells. Using the current method, we
found that a 72 h fast was associated with a 39% drop
REFERENCES
1. Beutler E. Annu. Rev. Nutr. 1989; 9: 287.
2. Staal FT, Ela SW, Roederer M, Anderson MT, Herzenberg LA.
Lancet 1992; 339: 909.
Copyright  2001 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2001; 36: 726–735

Glutathione isotope enrichment and level by GCMS 735
3. Hammarqvist F, Luo JL, Cotgreave IA, Andersson K, Werner-
man J. Crit. Care Med. 1997; 25: 261.
11. Kataoka H, Takagi K, Makita M. Biomed. Chromatogr. 1995; 9:
85.
4. Jahoor F, Jackson A, Gazzard B, Philips G, Sharpstone D, Frazer
ME, Heird W. Am. J. Physiol. 1999; 276: E205.
5. Capitan P, Malmezat T, Breuille´ D, Obled C. J. Chromatogr. B.
1999; 732: 127.
6. Lyons J, Rauh-Pfeiffer A, Yu YM, Lu XM, Zurakowski D, Top-
kins RG, Ajami AM, Young VR, Cersosimo E. Proc. Natl. Acad.
Sci. U.S.A. 2000; 97: 5071.
12. Darmaun D, Manary MJ, Matthews DE. Anal. Biochem. 1985; 147:
92.
13. Knapp DR. In Handbook of Analytical Derivatization Reactions.
Wiley–Interscience: New York, 1979; 255.
14. Matthews DE, Campbell RG. Am. J. Clin. Nutr. 1992; 55:
963.
15. Vajdovich P, Gaal T, Szilagyi A, Harnos A. Vet. Res. Commun.
1997; 21: 463.
7. Reeds PJ, Burrin DG, Stoll B, Jahoor F, Wykes L, Henry J,
Frazer ME. Am. J. Physiol. 1997; 273: E408.
´
16. Jonas CR, Estivarz CF, Jones DP, Gu LH, Wallace TM, Diaz EE,
8. Jahoor F, Wykes L, Reeds PJ, Henry J, Del Rosario MP, Frazer
ME. J. Nutr. 1995; 125: 1462.
Pascal RR, Cotsonis GA, Ziegler TR. J. Nutr. 1999; 127:
1278.
9. Reeds PJ, Burrin DG, Jahoor F, Wykes L, Henry J, Frazer ME.
Am. J. Physiol. 1996; 270: E413.
10. Majerus PW, Brauner MJ, Smith MB, Minnich V. J. Clin. Invest.
1971; 50: 1637.
17. Boza JJ, Moe¨nnoz D, Vuichoud J, Jarret A, Gaudard-de-Weck D,
Fritsche´ R, Donnet A, Schiffrin E, Perruisseau G, Balle´vre O. J.
Nutr. 1999; 129: 1340.
18. Roberts JC, Francetic DJ. Anal. Biochem. 1993; 211: 183.
Copyright  2001 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2001; 36: 726–735