Quantitation of monosaccharide isotopic enrichment in physiologic fluids by electron ionization or negative chemical ionization GC/MS using Di-O-isopropylidene derivatives

Article abstract of DOI:10.1021/ac990724x

The aldonitrile pentaacetate and other derivatives lack ions in the electron ionization (EI) spectra possessing an intact hexose structure and thus must be analyzed by chemical ionization GC/MS in order to study multiple isotopomers. We report methods for quantitation of hexose di-O-isopropylidene acetate (IPAc) or pentafluorobenzoyl (PFBz) esters. These were prepared in a two-step procedure using inexpensive reagents that do not adversely impact the isotopomer structure of the sugar. The acetate derivative possesses an abundant [M - CH₃] ion in the EI spectrum which is suitable for quantitative analysis of isotopomers. The negative chemical ionization (NCI) spectrum of the corresponding pentafluorobenzoyl derivative has a dominant molecular anion. Moreover, the PFBz derivative is about 100-fold more sensitive than the acetate, which offers some advantages for analysis of minor hexoses found in plasma. Isotopic calibration curves of [U-¹³C]glucose are linear over the 0.1-60% tracer/tracee range tested. The useful range for isotopic tracer studies is 25-2500 pmol for EI analysis of the acetate derivative and 0.1-55 pmol for NCI analysis of PFBz derivative (sample amount injected). For most studies where sample size is not limited, EI-GC/MS analysis of the IPAc derivative is preferred. NCI-GC/MS analysis is reserved when sample size is limiting or when studies involve hexoses other than glucose that are normally present at low concentration.

Full text of DOI:10.1021/ac990724x

Anal. Chem. 1999, 71, 4734-4739
Quantitation of Monosaccharide Isotopic
Enrichment in Physiologic Fluids by Electron
Ionization or Negative Chemical Ionization GC/MS
Using Di-O-isopropylidene Derivatives
David L. Hachey,* W. Reed Parsons, Siripoom McKay, and Morey W. Haymond
USDA/ARS Children’s Nutrition Research Center, Department of Pediatrics, Baylor College of Medicine,
Houston, Texas 77030
recently, investigators have focused their attention on quantifying
the enrichment of multiple isotopomers of glucose derived from
isotopic infusions of [2-13C]alanine, [2-13C]glycerol, and [U-13C]-
glucose to estimate the relative contribution of gluconeogenesis
to glucose production.3,6,12-18
The aldonitrile pentaacetate and other derivatives lack
ions in the electron ionization (EI) spectra possessing an
intact hexose structure and thus must be analyzed by
chemical ionization GC/ MS in order to study multiple
isotopomers. We report methods for quantitation of hex-
ose di-O-isopropylidene acetate (IP Ac) or pentafluoroben-
zoyl (P FBz) esters. These were prepared in a two-step
procedure using inexpensive reagents that do not ad-
versely impact the isotopomer structure of the sugar. The
acetate derivative possesses an abundant [M - CH3 ] ion
in the EI spectrum which is suitable for quantitative
analysis of isotopomers. The negative chemical ionization
(NCI) spectrum of the corresponding pentafluorobenzoyl
derivative has a dominant molecular anion. Moreover, the
P FBz derivative is about 1 0 0 -fold more sensitive than the
acetate, which offers some advantages for analysis of
minor hexoses found in plasma. Isotopic calibration
curves of [U-1 3 C]glucose are linear over the 0 .1 -6 0 %
tracer/ tracee range tested. The useful range for isotopic
tracer studies is 2 5 -2 5 0 0 pmol for EI analysis of the
acetate derivative and 0 .1 -5 5 pmol for NCI analysis of
P FBz derivative (sample amount injected). For most
studies where sample size is not limited, EI-GC/ MS
analysis of the IP Ac derivative is preferred. NCI-GC/ MS
analysis is reserved when sample size is limiting or when
studies involve hexoses other than glucose that are
normally present at low concentration.
Several techniques and derivatives have been described to
quantitate glucose isotopic enrichment. The more common
methods use the pentaacetate,6,19,20 aldonitrile pentaacetate,12,19,21-23
methyloxime pentatrimethylsilyl,1,12,24 bisbutylboronate acetate,12,25,26
and permethyl derivatives.12,27 Although each derivative has its
merits, those containing silicon or boron introduce significant
isotopic complexity in the mass spectrum, which limits their
usefulness for quantitation of isotopomers present at low enrich-
ment. The aldonitrile pentaacetate, simple hexose pentaacetates,
and permethylated hexoses generally lack ions in the electron
ionization (EI) spectrum possessing an intact hexose structure
and thus must be analyzed by chemical ionization (CI)-GC/ MS
in order to study the complete set of isotopomers.
The classic work by Biemann firmly established the utility of
MS for structural analysis of pentose and hexose acetates20 and
di-O-isopropylidene28 derivatives. Despite the fact that the O-
isopropylidene derivatives have been known for more than 60
(9) Searle, G. L. Clin. Endocrinol. Metab. 1 9 7 6 , 5, 783-804.
(10) Bier, D. M.; et al. Diabetes 1 9 7 7 , 26, 1016-23.
(11) Horber, F. F.; Marsh, H. M.; Haymond, M. W. Diabetes 1 9 9 1 , 40, 141-9.
(12) Beylot, M.; et al. Anal. Biochem. 1 9 9 3 , 212, 526-31.
(13) Yudkoff, M.; Nissim, I. Clin. Perinatol. 1 9 9 5 , 22, 97-109.
(14) Lee, W. N.; et al. Anal. Biochem. 1 9 9 5 , 226, 100-12.
(15) Robert, J. J.; et al. Diabetes 1 9 8 5 , 34, 67-73.
Analysis of the stable isotopic enrichment in plasma glucose
is one of the more enduring procedures in physiology and
metabolism. These measurements are used to determine rates of
hepatic glucose output in a variety of clinical conditions.1-11 More
(16) Fernandez, C. A.; et al. J. Mass Spectrom. 1 9 9 6 , 31, 255-62.
(17) Argoud, G. M.; Schade, D. S.; Eaton, R. P. Am. J. Physiol. 1 9 8 7 , 252, (5, Pt
1), E606-15.
(18) Martineau, A.; et al. Anal. Biochem. 1 9 8 5 , 151, 495-503.
(19) Guo, Z. K.; et al. Anal. Biochem. 1 9 9 2 , 204, 273-82.
(20) Biemann, K.; DeJongh, D. C.; Schnoes, H. K. J. Am. Chem. Soc. 1 9 6 3 , 85,
1763-70.
* Corresponding author: Department of Pharmacology 812 MRBI Vanderbilt
University Nashville, TN 37232-6400; (615)343-8382 (voice); (615)343-1268
(fax); david.l.hachey@vanderbilt.edu (e-mail).
(21) Szafranek, J.; Pfaffenberger, C. D.; Horning, E. C. Carbohydr. Res. 1 9 7 4 ,
38, 97-105.
(1) Laine, R. A.; Sweeley, C. C. Anal. Biochem. 1 9 7 1 , 43, 533-8.
(2) Bier, D. M. Baillieres Clin. Endocrinol. Metab. 1 9 8 7 , 1, 817-36.
(3) Bier, D. M.; et al. Diabetes 1 9 7 7 , 26, 1005-15.
(22) Pfaffenberger, C. D.; et al. Anal. Biochem. 1 9 7 5 , 63, 501-12.
(23) Previs, S. F.; et al. Anal. Biochem. 1 9 9 4 , 218, 192-6.
(24) Desage, M.; et al. Biomed. Environ. Mass Spectrom. 1 9 8 9 , 18, 1010-5.
(25) Wiecko, J.; Sherman, W. R. J. Am. Chem. Soc. 1 9 7 6 , 98, 7631-7.
(26) Lepetit, N.; Rocchiccioli, F. Rapid Commun. Mass Spectrom. 1 9 8 9 , 3, 153-
5.
(4) Delarue, J.; et al. Diabetologia 1 9 9 3 , 36, 338-45.
(5) Shaw, J. H.; Klein, S.; Wolfe, R. R. Surgery 1 9 8 5 , 97, 557-68.
(6) Tserng, K. Y.; Kalhan, S. C. Am. J. Physiol. 1 9 8 3 , 245 (t, Pt 1), E476-82.
(7) Royle, G. T.; Wolfe, R. R.; Burke, J. F. J. Surg. Res. 1 9 8 3 , 34, 187-93.
(8) Royle, G. T.; et al. Clin. Sci. 1 9 8 2 , 62, 553-6.
(27) Kochetkov, N. K.; Chizov, O. S. Tetrahedron 1 9 6 5 , 21, 2029-2047.
(28) DeJongh, D. C.; Biemann, K. J. Am. Chem. Soc. 1 9 6 4 , 86, 67-86.
4734 Analytical Chemistry, Vol. 71, No. 20, October 15, 1999
10.1021/ac990724x CCC: $18.00 © 1999 American Chemical Society
Published on Web 09/16/1999

Figure 1. Structures of the major hexose derivatives: 1,2:5,6-di-O-isopropylidene-R-D-glucofuranose (Ia,b), 2,3:5,6-di-O-isopropylidene-D-
mannofuranose (IIa,b), 1,2:3,4-di-O-isopropylidene-R-D-galactopyranose (IIIa,b), 2,3:4,5-di-O-isopropylidene-â-D-fructopyranose (IVa,b), 1,2:
4,5-di-O-isopropylidene-R-D-fructopyranose (Va,b).
years, possess favorable EI spectra and good chromatographic
properties, and are easily prepared in high yield using inexpensive
reagents, they have received scant attention from the biomedical
community.29,30 We report here a simple procedure for isola-
tion of glucose from plasma, preparation of di-O-isopropyl-
idene derivatives, and analysis of the acetylated derivatives
by electron ionization GC/ MS. Using a simple modification of
the final derivatization step, the procedure can be adapted to
microscale analysis of serum monosaccharides as the di-O-
isopropylidene pentafluorobenzoyl ester with isotopic analysis by
negative chemical ionization (NCI)-GC/ MS.
After informed consent was obtained, a glucose kinetic study
was conducted in a 56-year-old, 103.4-kg male subject diagnosed
with noninsulin-dependent diabetes mellitus. The subject received
a priming bolus dose of 20.0 µmol‚kg-1 and a constant infusion
of 20.0 µmol‚kg-1‚h-1 of [U-13C6]glucose for 6 h. Eight blood
samples were collected over the duration of the study. The
Institutional Review Board of Baylor College of Medicine approved
this protocol.
P reparation of Di-O-isopropylidene Acetate (IP Ac) De-
rivatives of Hexoses. The general derivatization chemistry
is illustrated in Figure 1 for glucose. Approximately 1-5 µmol
(180-1000 µg) of each carbohydrate was transferred to a 10-mL
screw-cap culture tube and evaporated to dryness in vacuo.
Ten microliters of water was added to dissolve the sugars,
followed by addition of 1 mL of 0.38 M sulfuric acid in acetone.
The samples were incubated at room temperature for 60 min
and then carefully neutralized with 2 mL of 0.44 M sodium
carbonate, followed by 2 mL of saturated sodium chloride.
O-Isopropylidene derivatives were extracted twice with 3 mL of
ethyl acetate, centrifuged, transferred to a 2-mL, vial and evapo-
rated to dryness under nitrogen on a cold manifold. The O-
isopropylidene derivatives were acetylated with 200 µL of ethyl
acetate/ acetic anhydride (1:1) at 60 °C for 30 min. Samples were
diluted as necessary with ethyl acetate to give a working
concentration of 125 µM (∼25 ng/ µL). The small amount of acetic
anhydride present is harmless to nonpolar polysiloxane capillary
GC columns.
MATERIALS AND METHODS
13
2
[U- C]Glucose and [6,6- H2]glucose (99 atom %) were obtained
from Cambridge Isotope Laboratories (Woburn, MA).
-galactose, -fructose, -mannose, and -xylose and derivatization
reagents were obtained from Sigma-Aldrich (St. Louis, MO).
Reference standards of 1,2:5,6-di-O-isopropylidene-R- -gluco-
furanose and 1,2:3,4-di-O-isopropylidene-R- -galactopyranose
D-Glucose,
D
D
D
D
D
D
were obtained from Aldrich Chemical Co. (Milwaukee, WI).
[14C]Glucose was obtained from CEA (Gif-sur-Yvette, France).
Capillary GC columns were obtained from Hewlett-Packard (Palo
Alto, CA) and Supelco (Bellefonte, PA). Common inorganic
reagents and solvents were obtained from Fisher Scientific
(Houston, TX). Stock solutions of labeled compounds and refer-
ence monosaccharides were prepared to a nominal concentration
of 7.5 mM in 20% acetone/ water to prevent microbial growth.
Isotope dilution curves covering a tracer/ tracee ratio of 0.001:1-
0.60:1 were prepared by serial volumetric dilutions from these
standards. Stock solutions were stored in the freezer at -20 °C
until needed.
P reparation of Di-O-isopropylidene P entafluorobenzoyl
(P FBz) Derivatives of Hexoses. Approximately 10-50 nmol (2-
10 µg) of each carbohydrate was derivatized as just described,
except for the final treatment with acetic ethyl acetate/ anhydride.
To form the PFBz derivative, the dried di-O-isopropylidene
derivative was dissolved in 250 µL of anhydrous pyridine, followed
(29) Gagnaire, D. Y.; Taravel, F. R. Eur. J. Biochem. 1 9 8 0 , 103, 133-43.
(30) Morgenlie, S. Carbohydr. Res. 1 9 7 5 , 41, 285-9.
Analytical Chemistry, Vol. 71, No. 20, October 15, 1999 4735

by addition of 25 µL of pentafluorobenzoyl chloride. The sample
was vortexed and then heated at 60 °C for 10 min. The PFBz
derivatives were extracted with 3 mL of hexane and washed twice
with 2 mL of water to remove pyridine and excess pentafluo-
robenzoyl chloride. The hexane layer was transferred to a 2-mL
vial and evaporated to dryness under nitrogen. The samples were
diluted as necessary with hexane to give a working concentration
of 1 µM (∼180 pg‚µL-1).
Statistical and Numerical Methods. Statistical analysis of
the tracer data was performed using Minitab Statistical Software
Release 11.1 (Minitab, Inc., State College, PA). One-way analysis
of variance (ANOVA) was done to determine mean, standard
deviation, and pooled standard deviation. The measurement
systems analysis subroutines in Minitab were used to evaluate
within-sample and between-sample variance. Theoretical isotope
ratios were computed using a copy of ISOPRO version 3.0 that
was kindly supplied by Dr. Michael Senko of Finnigan MAT Corp.
Numerical deconvolution of the tracer data was done as described
by Hachey et al.31 The correction matrix for glucose was computed
from the mean tracer ratios of the two unenriched baseline
samples using the CORMAT computer program described by
P reparation of Di-O-isopropylidene Derivatives of P lasma
Glucose. Plasma (50-100 µL) was pipetted into a 1.5-mL
microcentrifuge tube, followed by 1 mL of cold (0 °C) acetone.
The tube was capped, vortexed, and centrifuged at 2000g for 5
min to remove plasma proteins. The acetone was decanted into a
10-mL Teflon-capped culture tube and evaporated to dryness under
nitrogen. One milliliter of freshly prepared 0.38 M sulfuric acid
in acetone was added, and the sample was incubated at room
temperature for 60 min. At this point, either the acetate or the
pentafluorobenzoyl esters can be prepared as described above.
Note that for routine measurement of glucose isotopomer enrich-
ment, the acetate derivative with analysis by EI-GC/ MS is
preferred. The pentafluorobenzoyl esters may require dilution with
large volumes of solvent, so this method is reserved for studies
of minor hexoses present at low concentration in plasma or when
plasma sample size is limited.
Fernandez et al.16 However, the [U- C]glucose entry in the matrix
13
was obtained from a pure tracer standard. Dr. Henri Brunengraber
of Case Western University was kind enough to provide a copy
of CORMAT.
RESULTS AND DISCUSSION
Derivatization P rocedures. Figure 1 shows the structures
of di-O-isopropylidene esters formed by acetonation with 0.38 M
sulfuric acid, followed by acylation with either acetic anhydride
or pentafluorobenzoyl chloride. The mass spectra and fragmenta-
tion patterns of the acetate esters were identified by comparison
to the hydroxy compounds published by Biemann28 and Morgen-
lie30 and the acetylated glucose derivative Ia reported by Gag-
naire.29 Acetylation improves and maintains the capillary GC
column performance and peak shape, which tends to degrade with
repeated analysis of large numbers of physiologic glucose samples.
A small amount of water (∼10 µL) was used to dissolve crystalline
standards before acetonation due to the poor solubility of hexoses
in anhydrous acetone. High concentrations of water generally
resulted in poor yields of isopropylidene derivatives. To form
stable acetone adducts a cis 1,2- or cis 1,3-diol conformer is
required. Thus, various di-O-isopropylidene structures are possible
depending on the initial orientation of the hydroxyl groups in the
hexose. The composition of the final product is determined by
thermodynamic factors caused by intramolecular steric interac-
tions with the bulky gem dimethyl groups. Glucose, for example,
Derivatization Efficiency. Laboratory control plasma and
aqueous glucose standards were spiked with about 50 000
dpm‚mL-1 [14C]glucose. A 100-µL aliquot of each sample was
subjected to the derivatization procedure just described. Aliquots
were taken after the protein precipitation step and at the final
derivatization step for radioactivity determination. Results were
expressed as percent of the initial spike normalized to the volume
of plasma analyzed.
GC/ MS Analysis of Di-O-isopropylidene Acetate or P en-
tafluorobenzoyl Derivatives. Capillary GC/ MS analysis was
performed using a nonpolar HP-5ms 5% diphenyldimethylpolysi-
loxane column (30 m × 0.25 mm i.d., 0.25-µm phase thickness),
a medium-polarity HP-1701 (cyanopropylphenyl)dimethylpoly-
siloxane column (20 m × 0.18 mm i.d., 0.25-µm phase thickness),
or a SPB-17 50% (diphenyldimethyl)polysiloxane column (30 m
× 0.25 mm i.d.). Samples were chromatographed using injector
and detector temperatures of 250 °C, a helium inlet pressure of 6
psi, and splitless injection. The typical column temperature
program was from 80 to 260 °C at 10 °C‚min-1, after a 1-min
injection hold. Electron ionization (70 eV) mass spectra were
acquired on either a Hewlett-Packard HP5970 mass-selective
detector or a HP5989 MS Engine. NCI mass spectra were acquired
on the HP5989A instrument. The ion source temperature was 200
°C, methane reagent gas pressure was typically 0.9 Torr, and
electron energy was 150 eV. Full-scan spectra were acquired from
m/ z 20 to 550 at about 2 scans/ s. Selected ion monitoring data
were acquired using a 25-ms ion integration time. Instruments
were tuned and calibrated in EI mode using the instrument
“autotune” procedure. NCI tuning was done manually to achieve
the best sensitivity, resolution, and ion peak shape. Calibration
curves and quantitative standards were generally analyzed in
triplicate. Biological samples were usually analyzed twice, depend-
ing on the level of enrichment.
forms predominantly the 1,2:5,6-di-O-isopropylidene-R-D-glucofura-
nose (Figure 1, Ia,b), with a trace amount of two other diacetone
adducts, as noted also by Morgenlie.30 Although Ia always
dominates the mixture, the relative composition depends on the
amount of water and the time allowed for derivatization. Galactose
also forms a minor acetone adduct, but mannose forms exclusively
the 2,3:5,6-adduct (Figure 1, IIa,b). Fructose forms a 2:1 mixture
of 2,3:4,5- (Figure 1, IVa,b) and the 1,2:4,5-adducts (Figure 1,
Va,b). Figure 2 shows the chromatographic separation of a
standard mixture of pentose and hexose di-O-isopropylidene
derivatives. Although the minor glucose derivatives are present
at only a few percent of the total, they are generally resolved from
other hexoses (Figure 2). However, they may interfere with
quantitation of the minor hexoses due to the much greater
concentration of glucose in plasma (∼5.5 mM) than that of other
hexoses. A chromatogram of the PFBz derivatives is shown in
(31) Hachey, D. L.; Blais, J. C.; Klein, P. D. Anal. Chem. 1 9 8 0 , 52, 1131-5.
4736 Analytical Chemistry, Vol. 71, No. 20, October 15, 1999

Figure 3. Total ion chromatogram of the PFBz derivatives of xylose
and hexose standards. Peak identitification is described in Figure 3.
Note the absence of a xylose peak. Di-O-isopropylidene xylofuranose,
and other pentoses, lack a free hydroxyl and do not form PFBz
derivatives. See Figure 1 for a description of the structures.
Figure 2. Total ion chromatogram of a mixture of IPAc derivatives
(A) of xylose (1), fructose (2, IVa,b), glucose (3, Ia,b), a glucose
byproduct (4), mannose (5, IIa,b), galactose (6, IIIa,b), fructose (7,
Va,b), and minor glucose (8) and galactose (9) byproducts. Chro-
matogram B illustrates the typical composition of plasma hexoses in
the fed state. Note the presence of a small amount of mannose at
16.9 min. See Figure 1 for a description of the structures.
di-O-isopropylidene derivatives. No molecular ions were observed.
The base peak at m/ z 43 is due to intramolecular rearrangement
of the 1,3-dioxolane ring on the acetonide and ejection of a stable
CH3CO+ cation. Although there are significant differences in
fragmentation patterns between the pyranose and furanose
structures, the ion at m/ z 287, [M - CH3]+, is a common feature
in all the hexoses. Because this ion is relatively abundant, is free
of interfering fragments, and contains the intact hexose carbon
backbone, it is useful for isotopic quantitation. Other analytically
useful ions are observed at m/ z 169 [M - 133], which retains C1
- C6; m/ z 143 [M - 159], which retains C1 - C4; and m/ z 101
[M - 201] caused by scission between C4 and C5, which retains
C5 - C6.
Negative Chemical Ionization Mass Spectra. The methane
NCI spectra contain a dominant molecular anion at m/ z 454 [M]-
caused by resonance capture of a thermal electron, as illustrated
for the galactose derivative IIIb (Figure 5). Minor ions are present
at m/ z 396 [M - CH3COCH3]- and 380 [M - acetone - O]-. In
addition, there are ions at m/ z 148, C6F4-; m/ z 167, C6F5-; and
m/ z 195 C6HF5CO- from the PFBz group. The NCI spectrum of
the other hexoses is virtually identical to that of IIIb, differing
slightly in the relative abundance of the minor ions. The ion at
m/ z 454 is used for isotopic ratio measurements.
Figure 3. Separation of various hexoses is similar to that of the
acetate, but glucose and fructose are not resolved on a 30-m 5%
diphenyldimethylsiloxane column. Fructose and galactose con-
centration in plasma of healthy, fasting subjects is normally below
detection limits.
Isolation and derivatization efficiency was determined using
[14C]glucose. Protein precipitation using acetone resulted in 60
( 5% recovery in the supernatant, with the remainder bound to
the protein pellet. Extraction of the final derivatized acetonide with
hexane gave 22 ( 11% recovery, whereas extraction with ethyl
acetate gave 44 ( 21% overall recovery. Traces of water tend
to quench acetonide formation, so the acetone supernatant
obtained in the protein precipitation should be dried carefully.
Because the emphasis in this work is on isotope ratio determi-
nation and the concentration of glucose in plasma is sufficiently
high, no further attempts to improve the derivatization were
attempted. For quantitative analysis, the addition of a stable isotope
labeled internal standard would overcome the somewhat low
recovery.
Electron Ionization Mass Spectra. The 70-eV electron
ionization mass spectra of the four major hexoses are shown in
Figure 4. The corresponding acetates generally followed the
fragmentation pathways described by Biemann28 for the hydroxy
Isotope Calibration Curves. Isotope calibration curves were
constructed using [U- C6]glucose covering a tracer/ tracee ratio
of 0.001:1-0.60:1. Isotope dilution standards were analyzed
in triplicate. The amount of sample injected on the column
13
Analytical Chemistry, Vol. 71, No. 20, October 15, 1999 4737

significant analytical bias when different amounts of analyte
were injected. The slopes of the lines were 0.994-1.049 (R )
0.999) for the EI data and 0.952-0.987 (R ) 0.999) for the NCI
data.
Glucose Infusion in a Diabetic Subject. To illustrate the
suitability of these methods for analysis of tracer data from
metabolic kinetic studies, a primed-constant infusion protocol with
[U-13C6]glucose was conducted in a 56-year-old male subject with
noninsulin-dependent diabetes mellitus. The mean ion ratios from
triplicate determinations of the [M + 1] through [M + 6]
isotopomers are summarized in Table 1. From the ion ratios, a
corrected set of isotopomer data (Table 2) was computed by
solving a set of simultaneous linear equations using standard
matrix algebraic techniques.31 The isotopic enrichment of the [M
+ 6] isotopomer represents that of the parent unmetabolized
tracer, while that of the [M + 3] isotopomer reflects the primary
metabolic cleavage into two three-carbon metabolites and rein-
corporation into glucose by endogenous hepatic biosynthesis. The
[M +1] and [M + 2] isotopomers are derived from further
metabolism of the three-carbon fragments, followed by reincor-
poration into glucose, while the [M + 4] and [M + 5] isotopomers
are recombinant products of the [M + 3] isotopomer with either
the [M + 1] or the [M + 2] isotopomer. The isotopic enrichment
of the [M + 2] isotopomer is detectable, but with poor precision
due to the low absolute infusion rate of tracer in this subject. The
isotopic enrichment of the [M + 1] was not quantifiable due to
the inherent difficulty of measuring a small enrichment on top of
a large natural-abundance isotopomer. The [M + 4] and [M + 5]
isotopomers were not detectable due to the low probability of two
three-carbon unit recombination. Normally such studies are
conducted at 10-30-fold greater infusion rates in order to achieve
sufficient precision to be physiologically useful. In this example,
the plateau isotopic enrichment of the [M + 6] isotopomer
(0.011 29 ( 0.001 42) corresponds to a glucose flux of 1771
µmol‚kg-1‚h-1 in the fed state. The precision of the assay for the
[M + 6] isotopomer based on triplicate measurements was
typically better than 3%. Of the total variance in the measurement
system, the within-sample variance was 6.3%, which reflects the
inherent instrument precision. Most of the variance is explained
by sample-to-sample processing factors, even though the absolute
error is quite low. These data are consistent with a heteroscedastic
variance structure typically observed in stable isotope tracer
studies.
Figure 4. Electron ionization mass spectra of the hexose di-O-
isopropylidene derivatives of (A) fructose (IVa), (B) glucose (Ia), (C)
galactose (IIIa), and (D) mannose (IIa). See Figure 1 for a description
of the structures.
Measurement of the enrichment of glucose isotopomers is
most commonly done by methane positive chemical ionization
in order to preserve intact the carbon skeleton of the molecule.
Few glucose derivatives possess a favorable electron ionization
fragmentation pattern except the bis-n-butylboronate deriva-
tive.12,25,26 However, the unfavorable isotope distribution of boron
limits the utility of this derivative for quantitation of multiple
isotopomers. On the basis of the classical work by Biemann and
his students on the electron ionization mass spectra of carbohy-
drates,28 we expected the acetylated di-O-isopropylidene derivatives
to possess a favorable EI spectrum. This derivative can be
prepared rapidly and safely using inexpensive reagents; it has
excellent thermal and chemical stability; and it affords good
capillary GC performance. Moreover, the carbon burden of the
Figure 5. Methane NCI mass spectrum of 1,2:3,4-di-O-isopropy-
lidene-6-pentafluorobenzoyl-R-D-galactopyranose (IIIb).
ranged from 5 to 500 ng for EI measurements and 50 to 5000 pg
for NCI. Under these conditions, the curves were linear over
this range of isotopic enrichment, and they did not show any
4738 Analytical Chemistry, Vol. 71, No. 20, October 15, 1999

Table 1. Ion Abundance (×100) of Isotopomers in Plasma Glucose from a Subject Receiving a Constant Infusion of
[U-13C6]Glucose
sample
time (h)
[M + 1]
[M + 2]
[M + 3]
[M + 4]
[M + 5]
[M + 6]
1
2
3
4
5
6
7
8
-0.50
-0.25
3.00
4.00
4.50
5.00
5.50
6.00
15.541 ( 0.031
15.230 ( 0.113
15.477 ( 0.045
14.890 ( 0.165
15.318 ( 0.020
14.984 ( 0.016
15.267 ( 0.049
15.809 ( 0.031
2.642 ( 0.011
2.600 ( 0.035
2.727 ( 0.010
2.620 ( 0.061
2.695 ( 0.014
2.622 ( 0.067
2.655 ( 0.028
2.808 ( 0.022
0.297 ( 0.022
0.287 ( 0.024
0.383 ( 0.016
0.387 ( 0.045
0.368 ( 0.028
0.389 ( 0.040
0.385 ( 0.050
0.396 ( 0.006
0.046 ( 0.012
nd
0.052 ( 0.008
nd
0.050 ( 0.003
nd
0.054 ( 0.008
0.051 ( 0.007
nd
nd
nd
nd
0.077 ( 0.013
nd
0.077 ( 0.006
0.106 ( 0.029
0.072 ( 0.005
0.081 ( 0.018
1.200 ( 0.010
1.028 ( 0.019
1.092 ( 0.079
1.035 ( 0.036
1.030 ( 0.002
1.386 ( 0.010
pooled SD
pooled CV (%)
sample variance (%)
group variance (%)
isotope ratio (theory)
0.077
0.50
6.4
93.6
15.025
0.037
1.39
24.1
75.9
2.451
0.032
8.87
39.5
60.5
0.212
0.008
16.73
77.8
22.2
0.014
0.017
20.31
75.7
0.037
3.24
6.3
24.3
93.7
a
nd, not detected.
Table 2. Corrected Enrichment (×100) of Glucose Isotopomers from a Subject Receiving a Constant Infusion of
[U-13C6]Glucose
sample
time (h)
[M + 1]
[M + 2]
[M + 3]
[M + 4]
[M + 5]
[M + 6]
1
2
3
4
5
6
7
8
-0.50
-0.25
3.00
4.00
4.50
5.00
5.50
6.00
0.185 ( 0.031
-0.122 ( 0.112
0.122 ( 0.044
-0.458 ( 0.164
-0.036 ( 0.020
-0.365 ( 0.016
-0.086 ( 0.049
0.450 ( 0.031
0.001 ( 0.014
0.003 ( 0.018
0.093 ( 0.013
0.070 ( 0.042
0.084 ( 0.013
0.059 ( 0.064
0.052 ( 0.022
0.126 ( 0.025
0.011 ( 0.022
0.008 ( 0.028
0.084 ( 0.017
0.105 ( 0.047
0.075 ( 0.025
0.106 ( 0.031
0.096 ( 0.046
0.085 ( 0.008
0.016 ( 0.011
-0.027 ( 0.003
0.008 ( 0.006
-0.042 ( 0.005
-0.008 ( 0.027
-0.042 ( 0.005
-0.008 ( 0.036
0.004 ( 0.007
-0.004 ( 0.001
0.001 ( 0.000
-0.002 ( 0.013
-0.059 ( 0.001
0.007 ( 0.012
0.041 ( 0.028
0.005 ( 0.003
-0.008 ( 0.017
0.000 ( 0.000
0.000 ( 0.000
1.164 ( 0.009
1.004 ( 0.019
1.058 ( 0.077
1.000 ( 0.035
0.998 ( 0.001
1.345 ( 0.011
ACKNOWLEDGMENT
IPAc derivative is less than that of glucose pentaacetate or silylated
derivatives, making it attractive for isotopic abundance measure-
ments using GC-combustion-isotope ratio mass spectrometry.
Although in this work we used the acetylated derivative to improve
capillary GC performance, the final acetylation step could be
eliminated to further minimize the exogenous isotope dilution
burden.
The authors acknowledge the efforts of Edward Jones and
Astride Altomare of the Nemours Children’s Clinic in Jacksonville,
FL, who did some of the preliminary work evaluating other
derivatives of glucose and galactose. We also acknowledge the
statistical advice of Dr. E. O’Brian Smith of the CNRC. This work
is a publication of the USDA/ ARS Children’s Nutrition Research
Center, Department of Pediatrics, Baylor College of Medicine and
Texas Children’s Hospital, Houston, TX. This project has been
funded in part with federal funds from the U.S. Department of
Agriculture, Agricultural Research Service under Cooperative
Agreement 58-6250-6-001. The contents of this publication do not
necessarily reflect the views or policies of the U.S. Department
of Agriculture, nor does mention of trade names, commercial
products, or organizations imply endorsement by the U.S. Govern-
ment.
Using a simple modification, the pentafluorobenzoyl ester
derivative can be made. The PFBz derivatives are about 100-fold
more sensitive when analyzed by NCI GC/ MS than the acetate
derivative, making them attractive for studies in infants and other
subjects from whom limited amounts of blood are available. In
addition, studies of galactose, mannose, and fructose metabolism
become practical. These carbohydrates are normally absent in the
fasting state, and their concentration in the fed state may be more
than 50-fold lower than that of glucose. The NCI spectrum has a
dominant molecular anion, unlike many PFBz esters that produce
mainly a pentafluorobenzoyl anion. Because glucose is present
at about 5-10 mmol L-1 in plasma, the EI methods are adequate
for most routine isotopomer analyses.
Received for review June 30, 1999. Accepted July 20,
1999.
AC990724X
Analytical Chemistry, Vol. 71, No. 20, October 15, 1999 4739