Identification and quantitation of N-(carboxymethyl)valine adduct in hemoglobin by gas chromatography/mass spectrometry

Article abstract of DOI:10.1002/(SICI)1096-9888(199905)34:5<537::AID-JMS806>3.0.CO;2-H

A sensitive, specific and reproducible method was developed for the quantitation of the hemoglobin (Hb) adduct N-(carboxymethyl)valine (CMV). This adduct is one of various products from the Maillard reaction, involving reducing sugars and amino acids, proteins or other molecules with a free amino group. Such adducts, including N'-(carboxymethyl)lysine (CML), are called advanced glycation end products (AGE) and have been correlated with aging and severity of diabetes in human tissues. This method was developed to examine the CMV-Hb adduct as a possible AGE formed by reaction of Hb with glucose or other oxidation products. CMV was cleaved selectively from isolated globin using pentafluorophenyl isothiocyanate (PFPITC) in a modified Edman degradation at pH 9.5. The carboxyl group of the adduct was derivatized to its methyl ester with diazomethane. The resulting derivative, 5-isopropyl-1-(methyl acetate)-3-pentafluorophenyl-2-thiohydantoin, was detected by gas chromatography/mass spectrometry with selected ion monitoring (GC/SIM/MS). Quantitation was based on the response factor of the derivative molecular ion (m/z 396) from synthesized CMV and N-(2-carboxyethyl)valine (molecular ion m/z 410) as internal standard. This method exhibits reproducibility and linearity in the range 0.2-100 ng CMV. The limit of quantitation (0.2 ng CMV) gave a signal-to-noise ratio greater than 5:1 using a 1:30 sample aliquot. The GC/SIM/MS method can detect CMV adduct in 5 mg globin samples with relative standard deviations less than 5%. This approach avoids tedious acid hydrolysis and interference from other amino acids. The molecular ion and other CMV derivative ion assignments from samples were confirmed by accurate mass determinations using GC/high resolution SIM/MS. Measurements from random mouse, rat and human globin samples gave mean CMV levels of about 6, 5 and 14 nmol g^-1 Hb in these species, respectively.

Full text of DOI:10.1002/(SICI)1096-9888(199905)34:5<537::AID-JMS806>3.0.CO;2-H

JOURNAL OF MASS SPECTROMETRY
J. Mass Spectrom. 34, 537È543 (1999)
IdentiÐcation and Quantitation of
N-(Carboxymethyl)valine Adduct in Hemoglobin
by Gas Chromatography/Mass Spectrometry¤
Jian Cai and Harrell E. Hurst*
Department of Pharmacology and Toxicology, School of Medicine, University of Louisville, Louisville, Kentucky 40292, USA
A sensitive, speciÐc and reproducible method was developed for the quantitation of the hemoglobin (Hb) adduct
N-(carboxymethyl)valine (CMV). This adduct is one of various products from the Maillard reaction, involving
reducing sugars and amino acids, proteins or other molecules with a free amino group. Such adducts, including
N e-(carboxymethyl)lysine (CML), are called advanced glycation end products (AGE) and have been correlated
with aging and severity of diabetes in human tissues. This method was developed to examine the CMV–Hb adduct
as a possible AGE formed by reaction of Hb with glucose or other oxidation products. CMV was cleaved selec-
tively from isolated globin using pentaÑuorophenyl isothiocyanate (PFPITC) in a modiÐed Edman degradation at
pH 9.5. The carboxyl group of the adduct was derivatized to its methyl ester with diazomethane. The resulting
derivative, 5-isopropyl-1-(methyl acetate)-3-pentaÑuorophenyl-2-thiohydantoin, was detected by gas chromatog-
raphy/mass spectrometry with selected ion monitoring (GC/SIM/MS). Quantitation was based on the response
factor of the derivative molecular ion (m/z 396) from synthesized CMV and N-(2-carboxyethyl)valine (molecular
ion m/z 410) as internal standard. This method exhibits reproducibility and linearity in the range 0.2–100 ng CMV.
The limit of quantitation (0.2 ng CMV) gave a signal-to-noise ratio greater than 5 : 1 using a 1 : 30 sample aliquot.
The GC/SIM/MS method can detect CMV adduct in 5 mg globin samples with relative standard deviations less
than 5%. This approach avoids tedious acid hydrolysis and interference from other amino acids. The molecular ion
and other CMV derivative ion assignments from samples were conÐrmed by accurate mass determinations using
GC/high resolution SIM/MS. Measurements from random mouse, rat and human globin samples gave mean CMV
levels of about 6, 5 and 14 nmol g—1 Hb in these species, respectively. Copyright ( 1999 John Wiley & Sons, Ltd.
KEYWORDS: N-(Carboxymethyl)valine; Edman degradation; hemoglobin; glycation end product; biomarker; gas
chromatography/mass spectrometry
be accumulated with aging, in diabetic patients and
INTRODUCTION
uremia.6,11 It was also found to be a major AGE
antigen in tissue proteins.12,13 Formation of CML
requires oxidative conditions,14h16 as the appearance of
this adduct in vitro can be inhibited by removing
oxygen or by using scavengers of reactive oxygen
species. Such formation of this adduct can also be accel-
erated by introducing reactive oxygen species such as
hydroxyl radical.17 These results indicate that CML can
be used as a biomarker of oxidative stress.
Glycation of hemoglobin (Hb) is well recognized.18,19
The major glycation sites in Hb are the amino groups at
side-chains of lysine residues and the amino groups of
N-terminal Hb valine residues.20 Oxidation of glycated
valine residues should generate an analogous valine
adduct, N-(carboxymethyl)valine (CMV), which might
be used as a more accessible biomarker of Maillard
reaction and/or oxidative stress.
The Maillard reaction is the reaction between reducing
sugars and amino acids, proteins or other molecules
that have a free animo group.1 The Ðnal products of the
Maillard reaction are called advanced glycation end
products (AGE), which cause browning, Ñuorescence
and cross-linking in proteins. The accumulation of AGE
has been correlated with aging and severity of dia-
betes.2,3 The Ne-carboxymethyllysine (CML) adduct in
proteins is one of a few AGE that have been established
in vivo in human tissues.4,5 This adduct was shown to
* Correspondence to: H. E. Hurst, Department of Pharmacology
and Toxicology, School of Medicine, University of Louisville, Louis-
ville, Kentucky 40292, USA. E-mail: h.hurst=louisville.edu
¤ Portions of this paper were presented at the 1997 Annual Meeting
of the Society of Toxicology held 9È13 March 1997 in Cincinnati,
Ohio, and included as Abstract No. 1853, Fundam. Appl. T oxicol., 36
(No. 1, Part 2, T he T oxicologist), 365 (1997).
This study established a method for the quantitation
of CMV by gas chromatography/mass spectrometry
with selected ion monitoring (GC/SIM/MS). The
method used a modiÐed Edman degradation21h23 to
cleave the CMV adduct from Hb. Further reaction of
Contract/grant sponsor: Office of Research and Development,
Medical Research Service, Department of Veterans A†airs.
CCC 1076È5174/99/050537È07 $17.50
Copyright ( 1999 John Wiley & Sons, Ltd.
Received 29 December 1998
Accepted 9 February 1999

538
J. CAI AND H. E. HURST
the Edman product with diazomethane produced the
Ðnal derivative, 5-isopropyl-1-(methyl acetate)-3-
pentaÑuorophenyl-2-thiohydantoin, which is suitable
for the sensitive, reproducible analysis of adduct levels
by GC/SIM/MS. This study conÐrms the existence of
this adduct in human, mouse and rat hemoglobin.
dissolved in 10 ml of water, then 25 mmol of tri-
ethylamine were added to the solution as an acid scav-
enger. The mixture was allowed to react at 50 ¡C for 3
days, then was washed four times with 20 ml of
dichloromethane before it was condensed under nitro-
gen to about 10 ml. The washed mixture was loaded on
an ion-exchange column and eluted with water and 0.8
M formic aid. TLC indicated products in fractions from
the ion-exchange column. Those fractions containing
valine adduct were collected and the solvent were
evaporated under nitrogen. The valine adducts were
washed Ðnally with acetone (2 ] 50 ml) and dried under
nitrogen. The [M ] H]` ions in the fast atom bom-
bardment mass spectra of CMV and CEV were m/z 176
and 190, respectively. Major ion fragments (not shown)
could be rationalized from the structures of CMV and
CEV.
EXPERIMENTAL
Chemicals
Diethylenetriaminepentaacetic acid (DTPA), bromo-
acetic acid, 3-bromopropionic acid and Diazald were
purchased from Aldrich Chemical (Milwaukee, WI,
USA), valine and phosphate-bu†ered saline (pH 7.4)
from Sigma Chemical (St Louis, MO, USA), pentaÑuo-
rophenyl isothiocyanate (PFPITC) from Fluka Chemi-
cal (Milwaukee, WI, USA), constant-boiling (6 M
hydrochloric acid from Pierce Chemical (Rockford, IL,
USA) and perÑuorokerosene-755 (PFK-755) from Lan-
caster Synthesis (Windham, NH, USA). All reagents
were of the highest grade commercially available and
were used without additional puriÐcation.
Preparation of diazomethane
The preparation of diazomethane and derivatization
with diazomethane were carried out in a hood with
good ventilation. Diazomethane were prepared using a
mini Diazald apparatus (Aldrich Chemical). Ethanol (5
ml) and potassium hydroxide solution (2.5 g in 4 ml of
water) were added to the reaction vessel, which was
heated to 65 ¡C in a water-bath. A separating funnel
(with Clear-Seal joint, see Caution note below) was
attached over the reaction vessel and charged with a
solution of Diazald (N-methyl-N-nitroso-p-tolu-
enesulfonamide) in diethyl ether (2.5 g in 25 ml). The
Diazald solution was slowly added over about 20 min.
The diazomethane generated was condensed with
diethyl ether on a cold Ðnger Ðlled with a dry-iceÈ
acetone slurry and collected in a 50 ml receiving Ñask
(with Clear-Seal joint), which was cooled in an ice-bath.
Use of a diethyl ether trap at the side-arm decreased the
release of diazomethane into air, and the distillation
was stopped when the distillate turned colorless. The
diazomethane solution was used immediately or kept at
[20 ¡C before use within D2 h. Caution: diazomethane
is a very toxic, potentially carcinogenic compound that
can cause allergic sensitization. Diazomethane is poten-
tially explosive in contact with ground-glass joints or
sharp edges. Use extreme caution when handling this
reagent. Follow the instructions provided by Aldrich
Chemical.
Thin-layer chromatography
Thin-layer chromatography (TLC) on silica gel K6
plates (5 ] 10 cm) from Alltech (DeerÐeld, IL, USA)
was used to separate and visualize valine and valine
adducts. Plates were developed using waterÈmethanolÈ
acetonitrile (1 : 1 : 1). Amino acid spots were visualized
with ninhydrin (0.5 g in 100 ml of 1-butanol with 5 ml
of acetic acid). The R values (ninhydrin product colors)
f
for triethylamine (brown), valine (purple), CMV (purple)
and N-(2-carboxylethyl)-valine (CEV) (purple) were 0,
0.64, 0.74 and 0.79, respectively.
Ion-exchange chromatography
Synthetic standards were puriÐed by ion-exchange chro-
matography on AG 1-X8 resin (100È200 mesh, acetate
form) from Bio-Rad Laboratories (Hercules, CA, USA).
This resin was converted to the formate form by
washing with 2 M formic acid and then used to pack an
ion-exchange column (23 cm ] 1.8 cm i.d.). The column
was washed with 200 ml of 2 M formic acid, followed by
washing with doubly distilled water until the pH of the
elute became near neutral. Aliquots of 2È4 ml of the
mixture were loaded on the column. The column was
eluted with 20 ml of water, followed by 0.8 M formic
acid until all the components had passed through the
column. The column was washed with 2 M formic acid
and water before reloading additional material.
Blood sample collection
Human blood samples, drawn into vacuum tubes con-
taining EDTA as an anticoagulant, were collected ran-
domly from waste hospital specimens. Mouse blood
samples were obtained from male strains A/J and
C57BL/6 mice weighing 20È25 g (Jackson Laboratories,
Bar Harbor, ME, USA). Rat blood samples were col-
lected from male SpragueÈDawley rats weighing about
250 g from Harlan (Indianapolis, IN, USA). The
animals were anesthetized brieÑy with diethyl ether
before blood was drawn via cardiac puncture into a
heparinized syringe. Blood samples were stored under
Synthesis of CMV and CEV
Starting materials, including 10 mmol valine and 12
mmol bromoacetic acid or 3-bromopropionic acid, were
Copyright ( 1999 John Wiley & Sons, Ltd.
J. Mass Spectrom. 34, 537È543 (1999)

N-(CARBOXYMETHYL)VALINE ADDUCT IN HEMOGLOBIN
539
refrigeration brieÑy before preparation of globin
samples.
0.5 ml of constant-boiling HCl. The solution was trans-
ferred to a 1 ml vacuum hydrolysis tube and hydrolyzed
at 110 ¡C for 48 h under vacuum. After hydrolysis, the
acid was evaporated at 60 ¡C under nitrogen. The
residue was derivatized using the procedures described
above.
Preparation of globin samples
Whole blood (0.5È1.0 ml) was centrifuged for 5 min in a
benchtop centrifuge to pellet the erythrocytes. The cells
were washed twice with 3 ml of phosphate-bu†ered
saline and lysed with 2 ml of distilled water. The
resulting cellular debris was pelleted by centrifugation
(2000 rpm for 10 min). The supernatant Hb solution
was added dropwise to 20 ml of acidiÐed acetone (3 ml
of concentrated HCl per liter of acetone) chilled with
ice. Precipitated globin was obtained by centrifugation,
washed twice with 20 ml of acetone and evaporated to
dryness under nitrogen. Dried globin samples were
stored at [20 ¡C before analysis.
Gas chromatography/selected ion monitoring/mass
spectrometry
The Ðnal CMV derivative, 5-isopropyl-1-(methyl
acetate)-3-pentaÑuorophenyl-2-thiohydantoin,
analyzed under 70 eV electron ionization (EI) condi-
tions using an HP5890/HP5971A GC/MS system
(Hewlett-Packard, Palo Alto, CA, USA). Compound
was
separation was achieved using
a
bonded-phase
(polydimethylsiloxane) capillary column (DB-1, 10
m ] 0.2 mm i.d. ] 0.33 lm Ðlm thickness) from J&W
ScientiÐc (Folsom, CA). The GC injection port and
interface temperatures were set to 280 ¡C, with helium
carrier gas maintained at 6 psig. Injections were made
in the splitless mode with the inlet port purged for 1
min following injection. The GC oven temperature was
held initially at 100 ¡C for 1 min, then increased at a
rate of 20 ¡C min~1 to 290 ¡C, which was held for 1.5
min. Under these conditions the retention times for
CMV and CEV derivatives were 7.83 and 8.35 min,
respectively. Molecular (m/z 396) and major fragment
(m/z 354) ions of CMV derivative were monitored as
quantifying and qualifying ions, respectively. The
molecular ion (m/z 410) of the CEV derivative was mon-
itored as an internal standard. A standard sample
(prepared from CMV standard solution) was run with
each batch of samples. The response factor, determined
by peak area from weighed, synthesized CMV standard
vs. that of the CEV internal standard, was used to cal-
culate the CMV adduct concentrations in globin
samples.
Preparation of CMV-modiÐed globin
To produce model CMVÈHb adduct material for
method evaluation, human hemoglobin solution (about
2 mM) was allowed to react for 3 days with an approx-
imately 50-fold molar excess of bromoacetic acid and
triethylamine at 50 ¡C. The mixture was dialyzed
against water twice (1 l for 12 h each) using molecular
porous membrane tubing with a molecular mass cut-o†
of 6000È8000 Da (Spectra/Por 1, Spectrum Medicial
Industries, Los Angles, CA, USA). The product was pre-
cipitated with acidiÐed acetone, washed with acetone
and dried under nitrogen.
Derivatization of CMV adduct
For analysis each weighed globin sample (3.5È7 mg) was
dissolved in a mixture of 0.5 ml of 1-propanol and 1.0
ml of 0.5 M KHCO containing 1 mM DTPA. Aliquots
3
containing 20 ng of aqueous CEV (internal standard)
High-resolution selected ion monitoring
and 7 ll of PFPITC were added. The mixture was
vortex mixed to dissolve PFPITC completely, allowed
to react at room temperature overnight and heated at
50 ¡C for 2 h. The sample was then washed twice with 2
ml of hexane, acidiÐed with 0.2 ml of 6 M HCl and
extracted with 2 ] 2 ml of hexane. The hexane extracts
were combined and a 2 ml aliquot of ethereal diazo-
methane was added. The mixture was allowed to sit at
room temperature for 1 h, then the solvent was evapo-
rated under nitrogen. The residue was dissolved in
toluene (1 ml) and washed with 2 ml of distilled water.
The toluene phase was collected and evaporated to
dryness under nitrogen. The residue was dissolved in 30
ll of toluene and an aliquot (1 ll) was analyzed by GC/
SIM/MS.
(GC/HR/SIM/MS)
A VG 70SE high-resolution magnetic sector instrument
(VG Analytical, Manchester, UK) with an EI source
and HP5890 gas chromatograph was used in the experi-
ment. The electron ionization energy was 35 eV, the
trap current was set to 500 lA and the source tem-
perature was 250 ¡C. The instrument was tuned to
10 000 resolving power (10% baseline deÐnition). A cap-
illary column (DB-1, 15 m ] 0.25 mm i.d. ] 0.25 lm
Ðlm thickness) was directly coupled to the ion source.
Injections were made in the splitless mode with the
injection port purged 1.0 min following injection.
Helium (5 psig) was used as the carrier gas. The injec-
tion port and GC/MS interface temperatures were 280
and 270 ¡C, respectively. The oven temperature, held at
100 ¡C for 1 min initially, increased at 20 ¡C min~1 to
290 ¡C and held at 290 ¡C for 2.5 min. Under these con-
ditions, the retention times for the CMV and CEV
derivatives were 7.22 and 7.72 min, respectively. Ions
monitored included m/z 396.0567, 354.0098 and
Acid hydrolysis of globin samples
An aliquot of CMV-modiÐed globin containing about
70 ng of CMV adduct and 100 ng of CEV was added to
Copyright ( 1999 John Wiley & Sons, Ltd.
J. Mass Spectrom. 34, 537È543 (1999)

540
J. CAI AND H. E. HURST
321.9835 from CMV. The ion of m/z 392.9760 from
PFK-755 was used as the lock mass. Peak ratios
obtained at 7.22 min from samples were calculated and
compared with the corresponding peak ratios obtained
from CMV standard.
observed with N-(carboxymethyl)valine amino acid
standard. In the case of Hb, the CMV adduct is not
cleaved from the remaining globin protein and the
derivative cannot be detected with GC/MS.
Two peaks were detected when CMV standard
reacted with PFPITC under neutral pH (see Fig. 2). The
early peak with a retention time of 7.84 min was deriv-
ative B, while the later peak (7.92 min) was derivative A.
Initial assignment of identity was based on the fact that
only one peak was detected in CMV-modiÐed globin.
With the same retention time and mass spectrum, the
single derivative from CMVÈglobin had to be derivative
A. Di†erences in the M [ 42 fragments in the mass
spectra of these peaks shown in Fig. 2 also support this
assignment of structures to the eluted peaks. Loss of 42
Da gives one of the major fragments (m/z 354) in the
spectrum of the latter peak. This fragmentation can be
explained by McLa†erty rearrangement in derivative A.
This fragmentation was conÐrmed using mass spectra of
isotopically labeled CMV standards and GC/high
resolution (HR) MS (data not shown). Such a McLaf-
ferty rearrangement might be expected in derivative B,
but the fragment is not likely to be as stable as with
derivative A. This fragment was almost invisible in the
mass spectrum of the earlier peak, corresponding to the
structure of derivative B.
RESULTS
Edman degradation and derivatization
In modiÐed Edman degradation, PFPITC is used to
cleave the N-terminal CMV from globin. As this adduct
exhibits a terminal carboxyl group, there are two pos-
sible ways that PFPITC can react with CMV to form
di†erent derivatives, shown in Fig. 1. In the desired
reaction, PFPITC can react with the amine group and
peptide bond. This reaction cleaves the CMV adduct
from the remaining globin protein and reacts with the
former peptide-bond carboxyl group to form an inter-
mediate derivative. This derivative can be esteriÐed with
diazomethane to formed derivative A (Fig. 1), which is
detected by GC/MS. Alternatively, PFPITC can also
react with the amine group and the adduct carboxyl
group to generate another derivative. The derivative B
pathway (Fig. 1) illustrates this competing reaction
Figure 1. Derivatives formed from CMV using PFPITC and diazo-
methane. Only derivative A is formed from the CMV–Hb adduct.
Derivative B of this adduct does not form, as the corresponding
intermediate derivative remains linked to globin through a peptide
bond.
Figure 2. TIC and 70 eV mass spectra of CMV derivatives A and
B from modified Edman degradation using PFPITC reacting at
neutral pH.
Copyright ( 1999 John Wiley & Sons, Ltd.
J. Mass Spectrom. 34, 537È543 (1999)

N-(CARBOXYMETHYL)VALINE ADDUCT IN HEMOGLOBIN
541
For reproducibility and sensitivity, ideal reaction
conditions should favor the formation of derivative A
and avoid the formation of derivative B. Conditions
favoring these reactions were examined with synthesized
CMV standard. A single product is produced when
reaction with PFPITC occurs under basic conditions
(pH 9.5). Derivative A is essentially the only peak
detected with standards, which produce results similar
to the sample analyses shown in Fig. 3.
CMV quantitation and detection limit
For routine analyses, CMV adducts in Hb were
detected and quantiÐed by selected ion monitoring at
nominal mass resolution. Typical ion chromatograms
following the analysis of the CMV adduct in a sample
are shown in Fig. 3 and a calibration curve for CMV
quantitation is shown in Fig. 4. As the calibration curve
was linear through the working concentration range
(0.2È100 ng) and passed essentially through the origin,
the response factor from 20 ng of CMV relative to 10 ng
of CEV (internal standard) run with each batch provid-
ed the means for CMV quantitation. The detection limit
for this method was determined to be 0.2 ng per 5 mg
globin sample or 7 pg per injection (signal-to-noise ratio
P5).
Figure 4. Calibration curve from synthesized CMV used as a
quantitative standard with 25 ng of CEV as internal standard.
samples. The relative standard deviation (RSD) from
di†erent batch analyses was usually less than 5% (Table
1), while the RSD from repeated injections of the same
sample was less than 3%.
Method reproducibility
The reproducibility of the method was examined by
repeated measurement of di†erent aliquots from globin
Validation of the method
In this method synthesized CMV was used as a stan-
dard to measure the response factor used in quantitat-
ion. A systematic error would be introduced if the
recovery following derivatization was di†erent between
CMV adduct in globin and the CMV amino acid stan-
dard. A method involving hydrolysis was used to check
this possibility. Since globin is hydrolyzed by acid to
amino acids, no systematic error should be introduced
using acid hydrolysis. Therefore, comparison of results
Table 1. Reproducibility of CMV adduct quantitation in
human globin samples
Globin
1a
2a
3a
4a
Nb
Meanc
SD
9612A-2
12.7
11.1
15.1
15.0
12.3
10.6
14.3
14.4
12.5
11.7
14.3
13.9
13.1
11.7
14.5
—
4
4
4
3
12.6
11.3
14.6
14.4
0.34
0.51
0.37
0.54
9612A-5
9612A-11d
9612A-11d
a Different preparations (3.5–5 mg) of three globin samples from
different patients. Values in the table are mean values of CMV
adduct (nmol gÉ1 globin) from 3–5 injections. RSDs from repeat-
ed injection were less than 3%.
b Number of preparations from same globin sample in a batch.
c Mean values of CMV adduct (nmol gÉ1 globin) from different
preparations.
d Sample 9612A-11 was run on different days. Note the agreement
in results indicating good reproducibility between batches run on
different days.
Figure 3. Typical ion chromatograms from CMV–Hb adduct
quantitation. Peaks for m/z 396 and 354 at 7.83 min are from
molecular and M ÉC H ions, used for quantification and confir-
mation of CMV derivative A, respectively. The peak for m/z 410 at
8.35 min is the molecular ion of the CEV derivative, used as an
internal standard.
3
6
Copyright ( 1999 John Wiley & Sons, Ltd.
J. Mass Spectrom. 34, 537È543 (1999)

542
J. CAI AND H. E. HURST
2. Good agreement for CMV quantitation between
Table 2. Results of CMV adduct quantitation in CMV-
hydrolytic and modiÐed Edman degradation indicates
that no signiÐcant error was introduced by using syn-
thesized CMV as standards rather than a CMVÈpeptide
standard.
modiÐed globina using di†erent procedures
Procedure
nd
Meanse
SD
Edmanb
4
4
4
20.9
20.7
20.1
1.23
0.38
1.21
Edmanb
Hydrolysisc
ConÐrmation of CMV adduct in Hb
a CMV-modified globin was obtained by reacting Hb with
bromoacetic acid.
Since the concentrations of CMV adduct in Hb are in
the pmol mg~1 Hb range, full-scan mass spectra could
not be obtained to identify this adduct. The existence of
the CMV adduct in Hb was conÐrmed in selected
samples by GC/HR/SIM/MS at 10 000 resolving power
to decrease the probability of misidentiÐcation. The
three major ions of the CMV derivative determined
with CMV standard by GC/HR/MS are at m/z
396.0567, 354.0098 and 321.9835, which are the molecu-
b Globin samples were analyzed by modified Edman degra-
dation.
c Globin samples were hydrolyzed with constant-boiling
HCl at 110 ¡C for 48 h.
d Number of preparations from CMV-modified globin.
e Mean values of CMV adduct (lmol gÉ1 globin) detected
in samples.
lar ion (C
15 13 2 3 3
[ C H [ CH O] ion, respectively. Peaks with same
H
N O F S), [M [ C H ] ion, and [M
3 6
with acid hydrolysis can be used to examine the
recovery of the modiÐed Edman degradation. Quanti-
tative results from analyses of a CMV-modiÐed globin
sample by the two di†erent methods are shown in Table
3 6
4
retention time as CMV standards were found in these
ion channels in all globin samples examined. The peak
ratios were calculated and compared with CMV stan-
dard (see Table 3). The good agreement of the peak
ratios with CMV standard conÐrmed the existence of
CMV adduct in human, mouse and rat Hb. Table 4 lists
mean CMV adduct levels found in di†erent globin
samples from these species.
Table 3. Area ratios of CMV derivative in globin
samples detected with GC/HR/SIM/MSa
Sampleb
Meanc ÀSD (322/396)
Meanc ÀSD (354/396)
DISCUSSION
CMV standard
Human globin
0.258 À0.021
0.284 À0.031
0.261 À0.027
0.301 À0.033
0.300 À0.049
0.261 À0.023
0.353 À0.040
0.286 À0.020
0.300 À0.036
0.294 À0.044
0.738 À0.064
0.715 À0.052
0.698 À0.039
0.762 À0.064
0.773 À0.060
0.639 À0.050
0.851 À0.074
0.740 À0.030
0.751 À0.055
0.730 À0.062
Since its discovery and identiÐcation in human tissues,
CML has received attention as a marker of oxidative
damage. CML was found to accumulate in collagen,
lens proteins and other tissues in diabetes and aging.
The formation of CML is related to reactive oxygen
species, and it has been used as a biomarker of oxida-
tive stress.8,10,11 One advantage of using CML as a bio-
marker of oxidative stress is that it can be used to
monitor localized oxidative stress exposure. A disadvan-
tage of this approach is that in many cases it is difficult
or impossible to collect tissue samples, as such requires
justiÐcation of biopsy. CML is most often detected in
long-lived proteins such as collagen and lens proteins.
In these cases CML was noted as a long-term bio-
marker of oxidative damage to tissues, but it may not
reÑect recent changes in oxidative stress.
Mouse globin
Rat globin
a Monitored ions of CMV derivatives were m/z 396.0567
(molecular ion), 354.0097 (M ÉC H ) and 321.9835 (M
ÉC H ÉCH O). Values were determined from peak areas
3
6
3
6
4
measured at 10 000 mass resolving power.
b Derivatives of three globin samples from different patients
or animals were analyzed.
c Mean area ratio and SD values were from 3–8 injections of
each sample.
If the formation of the CMV adduct in Hb occurs
through mechanisms similar to that of CML, the CMV
adduct in Hb might be used as a biomarker of oxidative
stress.24 There are several potential advantages of using
this CMVÈHb adduct as such an indicator. An advan-
tage for time course studies is that the life span of Hb in
human is about 4 months. The CMV adduct in Hb
might be used for detection of recent exposures or for
dosimetry of oxidative damage over time with serial
sampling. Such an intermediate-term biomarker should
be useful in pathophysiology investigations of both
diabetes- and age-related complications.25
Table 4. CMV adduct detected in globin samples
Globin
Nb
Meansc
SD
Humana
10
4
13.5
5.3
3.18
0.49
1.41
0.69
Mouse, strain A/J
Mouse, C57BL/6
Rat, Sprague–Dawley
5
6.8
4
5.4
a Human globin samples were obtained from 10 different
random waste blood samples.
b Number of samples from different patients or animals.
c Mean values for CMV adduct (nmol gÉ1 globin) detected
in samples.
Other advantages of this method stem from the con-
venient methodology. Blood sample collection and
globin sample preparation are both easy to accomplish.
Copyright ( 1999 John Wiley & Sons, Ltd.
J. Mass Spectrom. 34, 537È543 (1999)

N-(CARBOXYMETHYL)VALINE ADDUCT IN HEMOGLOBIN
543
Only a few drops of blood are required for this assay.
CMV can be detected with modiÐed Edman degrada-
tion, which is sensitive, speciÐc and reliable. As noted in
this procedure for modiÐed Edman degradation, the N-
terminal adduct can be cleaved from globin in one step
instead of two steps, which are required to cut N-
terminal valine that has a free (primary) amino group.
This di†erence essentially eliminated interference in
quantitation from all other residues in globin, including
Hb N-terminal valine that has no adduct.
For typical CML quantitation, proteins are usually
hydrolyzed to amino acids before analysis with HPLC
or GC/SIM/MS. Glycated lysine residues (Amadori
products) can be partially converted to CML during
acid hydrolysis,8 which may introduce errors in CML
quantitation. To avoid this problem, Amadori products
must be reduced before hydrolyzing protein. Addi-
tionally, large amounts of amino acids from protein
need to be removed with a clean-up procedure, other-
wise they can cause problems in quantitation with
HPLC or GC/SIM/MS. These requirements combine to
consuming procedure. With modiÐed Edman degrada-
tion, CMV derivatization is accomplished under mild
conditions, and remaining globin is easily removed
during liquid extraction. These di†erences made this
method much simpler to use.
This method exhibits very good sensitivity and repro-
ducibility. The limit of quantitation for the current
method is 0.2 ng per globin sample. If 5 mg of globin
are used in analysis, the sensitivity is sufficient to detect
CMV concentrations as low as 0.23 nmol g~1 globin,
which is much lower than the levels found in globin
samples (see Table 4). This laboratory will use the
method published here to investigate mechanisms of the
formation of CMV in Hb by reactive oxygen species.
Acknowledgement
This material is based upon work supported in part by the Office of
Research and Development, Medical Research Service, Department of
Veterans A†airs.
make CML quantitation
a
laborious and time-
REFERENCES
1. J. W. Baynes and V. M. Monnier (Eds), The Maillard Reaction
in Aging, Diabetes and Nutrition. Alan R. Liss, New York
(1989).
14. K. J. Wells-Knecht, D. V. Zyzak, J. E. Litchfield, S. R. Thorpe
and J. W. Baynes, Biochemistry 34, 3702 (1995).
15. M. C. Wells-Knecht, S. R. Thorpe and J. W. Baynes, Biochem-
istry 34, 15134 (1995).
2. M. Brownlee, Diabetes 43, 836 (1994).
3. V. M. Monnier, D. R. Sell, R. H. Nagaraj, S. Miyata, S.
Grandhee, P. Odetti and S. A. Ibrahim, Diabetes 41 (Suppl.
2), 36 (1992).
4. M. U. Ahmed, S. R. Thorpe and J. W. Baynes, J. Biol. Chem.
26, 4889 (1986).
16. M. X. Fu, J. R. Requena, A. J. Jenkins, T. J. Lyons, J. W.
Baynes and S. R. Thorpe, J. Biol. Chem. 271, 9982 (1996).
17. R. Nagai, K. Ikeda, T. Higashi, H. Sano, Y. Jinnouchi, T. Araki
and S. Horiuchi, Biochem. Biophys. Res. Commun. 234, 167
(1997).
5. M. U. Ahmed, J. A. Dunn, M. D. Walla, S. R. Thorpe and J. W.
Baynes, J. Biol. Chem. 263, 8816 (1988).
18. P. J. Higgins and H. F. Bunn, J. Biol. Chem. 256, 5204
(1981).
6. J. A. Dunn, J. S. Patrick, S. R. Thorpe and J. W. Baynes, Bio-
chemistry 28, 9464 (1989).
19. D. E. Goldstein, R. R. Little, H. M. Wiedmeyer, J. D. England
and E. M. McKenzie, Clin. Chem. 32, B64 (1986).
20. R. Shapiro, M. J. McManus, C. Zalut and F. Bunn, J. Biol.
Chem. 255, 3120 (1980).
21. J. Mowrer, M. Tornqvist, S. Jensen and L. Ehrenberg, Toxicol.
Environ. Chem. 11, 215 (1986).
22. M. Tornqvist, J. Mowrer, S. Jensen and L. Ehrenberg. Anal.
Biochem. 154, 255 (1986).
23. E. Bailey, A. G. F. Brooks, C. T. Dollery, P. B. Farmer, B. J.
Passingham, M. A. Sleightholm and D. W. Yates, Arch.
Toxicol. 62, 247 (1988).
7. J. A. Dunn, D. R. McCance, S. R. Thorpe, T. J. Lyons and
J. W. Baynes, Biochemistry 30, 1205 (1991).
8. K. J. Knecht, J. A. Dunn, K. F. McFarland, D. R. McCance,
T. J. Lyons, S. R. Thorpe and J. W. Baynes, Diabetes 40, 190
(1991).
9. E. D. Schleicher, E. Wagner and A. G. Nerlich, J. Clin. Invest.
99, 457 (1997).
10. K. Horie, T. Miyata, K. Maeda, S. Miyata, S. Sugiyama, H.
Sakai, C. van Y. de Strihou, V. M. Monnier, J. L. Witztum and
K. Kurokawa, J. Clin. Invest. 100, 2995 (1997).
11. T. Miyata, Y. Wada, Z. Cai, Y. Iida, K. Horie, Y. Yasuda, K.
Maeda, K. Kurokawa and Y de Strihou, Kidney Int. 51, 1170
(1997).
24. T. J. Lyons, K. E. Bailie, D. G. Dyer, J. A. Dunn and J. W.
Baynes, J. Clin. Invest. 87, 1910 (1991).
25. Z. Makita, H. Vlassara, E. Rayfield, K. Cartwright, E. Friedman,
R. Rodby, A. Cerami and R. Bucala, Science 258, 651 (1992).
12. S. Reddy, J. Bichler, K. J. Wells-Knecht, S. R. Thorpe and
J. W. Baynes, Biochemistry 34, 10872 (1995).
13. K. Ikeda, T. Higashi, H. Sano, Y. Jinnouchi, M. Yoshida, T.
Araki, S. Ueda and S. Horiuchi, Biochemistry 35, 8075
(1996).
Copyright ( 1999 John Wiley & Sons, Ltd.
J. Mass Spectrom. 34, 537È543 (1999)