Protein methylation as a marker of aspartate damage
in glucose-6-phosphate dehydrogenase-deficient erythrocytes
Role of oxidative stress
Diego Ingrosso
1,2
, Amelia Cimmino
1
, Stefania D’Angelo
1
, Fiorella Alfinito
3
, Vincenzo Zappia
1,2
and Patrizia Galletti
1,2
1
Department of Biochemistry and Biophysics, School of Medicine, Second University of Naples, Italy;
2
Cardiovascular Research
Centre, School of Medicine, Second University of Naples, Italy;
3
Department of Hematology, School of Medicine,
University of Naples Federico II, Italy
The ÔMediterraneanÕ variant of glucose-6-phosphate dehy-
drogenase (G6PD) deficiency is due to the C563CT point
mutation, leading to replacement of Ser with Phe at position
188, resulting in acute haemolysis triggered by oxidants.
Previous work has shown increased formation of altered
aspartate residues in membrane proteins during cell ageing
and in response to oxidative stress in normal erythrocytes.
These abnormal residues are specifically recognized by

deamidation/isomerization in the mechanisms of cell injury
and haemolysis.
Keywords: erythrocyte membrane; glucose-6-phosphate
(G6PD) deficiency;
L
-isoaspartate residues; oxidative stress;
protein methylation.
Several biochemical variants of glucose-6-phosphate dehy-
drogenase (G6PD), corresponding to about 100 different
point mutations of the gene encoding this protein, have been
described [1]. Many of these are associated with chronic or
acute haemolysis. The ÔMediterraneanÕ clinical variant,
resulting from the C563CT change in the gene sequence,
results in the replacement of serine with phenylalanine at
position 188. Clinical outcome is characterized by neonatal
jaundice and acute haemolysis and haemoglobinuria.
Haemolysis is triggered by exposure to oxidants, e.g. fava
beans(thediseaseisoftenreferredtoasÔfavismÕ, and acute
haemolysis is called Ôfavic crisisÕ) or administration of drugs
such as primaquine, nitrofurantoin, and sulfamethoxazole,
or infection [1]. G6PD activity is almost undetectable in
most patients [2,3]. However, despite the vast amount of
data on the characterization of different G6PD variants, the
pathophysiological link between the enzyme defect and
haemolysis has not been unequivocally elucidated. Among
G6PD-deficient erythrocytes, aged cells are the most
sensitive to haemolysis, because of age-dependent decay of
the activity of several enzymes, including G6PD, and
antioxidant systems. There is further evidence that altera-
tions in the plasma membrane are central to the mechanism

Eur. J. Biochem. 269, 2032–2039 (2002) Ó FEBS 2002 doi:10.1046/j.1432-1033.2002.02838.x
isomerization of aspartate residues [6]. Major targets of
these alterations, also called Ôprotein molecular fatigue
damageÕ [7], are cytoskeletal components, such as ankyrin
and bands 4.1 and 4.2, as well as the integral membrane
protein band 3 (AE1; the anion transporter) [6]. In this
respect, asparaginyl deamidation has been shown to play a
role in the shift of band 4.1b to 4.1a [8] during erythrocyte
ageing. Increased molecular fatigue damage of several
cytoskeletal proteins has been found associated not only
with erythrocyte ageing, but also with intrinsic defects of
erythrocytes [7]. In fact,
L
-isoaspartate, a major protein
fatigue degradation product, is increased in hereditary
spherocytosis, and its increase correlates positively with the
degree of spectrin deficiency [9]. Erythrocyte passage
through the spleen microcirculation has also been found
to be a key determinant of this type of protein alteration in
spherocytosis [10]. Therefore, deamidation and isomeriza-
tion of membrane proteins have been proposed to play a
role in spleen conditioning [10].
Major byproducts of protein fatigue at the Asn/Asp
(Asx) level are
L
-isoaspartate residues. These residues are
selectively recognized by a specific S-adenosylmethionine
(AdoMet)-dependent enzyme,
L
-isoaspartate (

H]methionine (55 CiÆmmol
)1
) were from Amer-
sham International. Ready Gel liquid-scintillation cocktail
was from Beckman Inc. (Cuppertino, CA, USA). Percoll
was purchased from Pharmacia (Uppsala, Sweden). Selec-
tographin was obtained from Schering (Berlin, Bergkamen,
Germany). Thiobarbituric acid and t-butylhydroperoxide
(t-BHP) (70% aqueous solution) were from Sigma Co. (St
Louis, MO, USA).
Patient enrolment and sample processing
All subjects to be enrolled were assessed by a standard
screening panel, evaluating blood G6PD activity and the
presence of the C563T point mutation. A group of patients
and age/sex-matched normal controls were selected.
Patients gave informed consent and were made aware of
the outcomes of the study. All procedures and manipula-
tions, including blood sampling and genetic diagnostics,
were subject to authorization by the patients. Experimental
design was subject to approval by the bioethics committee,
as required. At the time of the study, patients were free of
haemolysis and in good clinical condition. Routine bio-
chemical blood tests (Hitachi 911 Automatic Analyzer) and
a standard haematological screening test for anaemia were
performed. For all erythrocyte testing, blood samples were
withdrawninEDTA(1mgÆmL
)1
blood) and further
processed for determination of G6PD activity [16]. Control
G6PD activity was 8.34 ± 1.59 UÆg

formed by incubating the cells at 37 °C in a shaking water
bath, in the presence of t-BHP at the indicated concentra-
tion, in 25-mL flasks (final haematocrit 10%). After
incubation, supernatants were used to determine levels
of thiobarbituric acid-reactive substances (TBARS) as
described below. Erythrocytes were washed seven times
with isotonic buffer to remove t-BHP.
Evaluation of oxidation markers
Determination of methaemoglobin and oxyhaemoglo-
bin. Methaemoglobin and oxyhaemoglobin contents were
determined by a spectrophotometric method [20]. Briefly,
5 lL packed oxidized erythrocytes were mixed with 995 lL
stabilizing solution (2.7 m
M
EDTA, pH 7.0, and 0.7 m
M
2-mercaptoethanol). After shaking, oxyhaemoglobin and
methaemoglobin concentrations were measured spectro-
photometrically.
Ó FEBS 2002 Protein methylation in G6PD-deficient erythrocytes (Eur. J. Biochem. 269) 2033
Evaluation of lipid peroxidation. Lipid peroxidation was
evaluated by detecting the amount of TBARS, mainly
malondialdehyde, as described previously [21]. Briefly, 2 mL
of the supernatant of oxidized erythrocyte pellet was mixed
with 1 mL 30% (w/v) trichloroacetic acid and centrifuged at
5000 g for 15 min. A 2-mL aliquot of supernatant was added
to 0.5 mL 1% (w/v) thiobarbituric acid in 0.05
M
NaOH and
heated in a boiling-water bath for 10 min. The absorbance of

M
sodium phosphate,
pH 8.0, containing 25 m
M
phenylmethanesulfonyl fluoride).
Membranes were then washed twice with the same hypotonic
solution at decreasing pH (7.2 and 6.2) to preserve methyl
ester stability. Radioactivity incorporated as protein methyl
esters was determined after solubilizationof10 lLmembrane
preparation in 125 lL10 m
M
acetic acid/2.5% SDS. Protein
concentration was determined as described previously [6].
Electrophoretic analysis of membrane proteins
SDS/PAGE of membrane erythrocytes was performed by
method of Fairbanks et al.[22] with modifications [9]. The
gels were 1.5 mm thick and contained acrylamide mix 5.6%
(mass/vol), in the presence of 1% SDS, at pH 7.4. All
samples were run in duplicate so that one control and one
treated (and/or patient) samples were analysed in parallel on
each gel half of the same gel. At the end of the run, gels were
cut into half, and one half was stained with Coomassie
Brilliant Blue to visualize protein bands and densitometri-
cally scanned for area quantification [9]. The other half was
used for methyl ester quantification. For this, lanes were
sliced into 2-mm fractions and the incorporated radioacti-
vity was determined after elution of proteins from each slice
[6,9]. Radioactivity was expressed as d.p.m./band area.
Determination of AdoMet and
S

EDTA, pH 7.0, and 0.7 m
M
2-mercaptoethanol) [16].
Wild type
AB
Mediterranean
417 120
120100317
1a 1b 2
417
1 2 3 4 5 6 7 8 9 10 11 St
317
120
100
Fig. 1. Diagnostic assessment of molecular defect in G6P-deficient patients. Patient selection, sampling and DNA extraction were as described in
Materials and Methods. The C563T mutation, associated with the Mediterranean variant was identified after PCR amplification of exon 5 and 6,
followed by digestion with MboII restriction enzyme. (A) Schematic representation of the expected restriction fragment length polymorphism
(RFLP) in wild-type and Mediterranean mutants, where the latter show an additional MboII site. (B) RFLP analysis of some patients and controls.
1, 3, 10, Mediterranean variant male patients; 4, 5, heterozygous females; 2, 7, 9, 11, normal controls. Mutants are characterized by sensitivity to
MboII digestion of PCR amplified fragments, yielding additional bands of 317 and 100 bp, respectively.
2034 D. Ingrosso et al.(Eur. J. Biochem. 269) Ó FEBS 2002
Membranes were removed by centrifugation at 10 000 g for
20 min. The assay mixture contained, in a final volume of
40 lL, 1.6 mg ovalbumin as the methyl acceptor, 2.8 mg
cytosolic proteins, 0.1
M
sodium citrate buffer, pH 6.0, and
30 l
M
(final concentration) S-adenosyl-

number of abnormal aspartate residues, spontaneously
arising from
L
-asparaginyl deamidation and/or
L
-aspartyl
isomerization reactions [7]. In addition it has been reported
that isoaspartate residues, detected by the PCMT in situ
assay, increase in membrane proteins of normal erythro-
cytes subjected to oxidative stress [14], suggesting that
susceptibility to oxidative damage may render these mem-
brane protein components more prone to deamidation/
isomerization.
We measured methyl esterification of membrane proteins
in G6PD-deficient erythrocytes, to establish if abnormal
isoaspartate residues occur, in this condition, at a higher
rate than normal, while in the circulation. To this end, cells
were fractionated according to density, the two most
abundant, intermediate fractions being used in the subse-
quent procedure. Cell recovery in these fractions, with
respect to the total amount of cells loaded on to the
gradient, was 70.3 ± 5.1% (control) vs. 80.9 ± 3.8%
(G6PD). Cell percentages were 46.6 ± 4.1% (buoyant
fraction) vs. 23.7 ± 1.9 (dense fraction) for the control
samples, and 61.0 ± 3.2% (buoyant fraction) vs.
19.9 ± 2.3 (dense fraction) for the G6PD samples. An
equal number of cells from each fraction was incubated with
methyl-labelled methionine, the in vivo AdoMet precursor.
PK activity was measured in parallel, as a cell age marker, in
cytosolic extracts of the same erythrocyte fractions. PK

treatment with t-BHP, leads to significant membrane
alterations, including the occurrence of deamidated/isomer-
ized Asx residues of membrane-cytoskeletal proteins [15].
Fig. 2. Membrane protein methylation levels and PK activity of density-
fractionated G6PD-deficient erythrocytes. (A) PK activity, as a cell-age
marker, was determined in erythrocyte cytosol, as detailed in Materials
and methods. (B) Membrane protein methylation levels were deter-
mined in two different erythrocyte age/density fractions obtained by
isopycnic centrifugation on a Percoll gradient. Methyl esterification in
intact erythrocytes was assayed by incubating them in the presence of
[
3
H]methionine according to the in situ procedure (see Materials and
methods).
Ó FEBS 2002 Protein methylation in G6PD-deficient erythrocytes (Eur. J. Biochem. 269) 2035
We were therefore intrigued to investigate whether the
abnormal susceptibility of G6PD-deficient erythrocytes to
oxidative stress could be responsible for their increased
tendency to isoaspartate formation in membrane proteins.
Therefore we monitored development of this alteration to
evaluate its pathophysiological meaning in the mechanism
of cell damage, using the in situ methylation assay, in
isolated G6PD-deficient erythrocytes subjected to oxidative
stress.
Erythrocytes from both normal control and G6PD-
deficient patients were subjected to oxidant treatment, with
t-BHP, before the in situ methylation assay. To limit
possible interference of cell manipulation with the oxidative
stress conditions, cells were not fractionated according to
density. The effects of reactive oxygen species on erythrocyte

PCMT specific activity, after oxidative treatment, confirm-
ing our previous findings [15].
As a whole, the results indicate that the lack of
reducing power is a crucial element in conditioning
erythrocyte susceptibility to undergo membrane protein
damage in the form of Asx deamidation/isomerization.
Isoaspartate formation may also be one of the ultimate
events in cell destruction. Evidence shows that erythrocyte
removal during cell ageing or after oxidative damage is
mediated by binding of band 3 antibodies to band 3
antigenic sites [5]. Therefore, the occurrence of altered
aspartate residues in band 3 of normal and abnormal
erythrocytes during ageing [6,7] or oxidative stress [15]
may be relevant to the fact that the same protein becomes
a major site of new antigen generation under the same
conditions.
It should be pointed out, in this respect, that erythrocyte
ageing was initially believed to be the main determinant of
isoaspartate formation in membrane proteins [6]. Our
results are in line with a different interpretation, which
underscores the equally important role played by cell stress
in the occurrence of such protein damage. This may be
particularly relevant to pathological conditions, such as
Fig. 3. Evaluation of oxidation markers in G6PD-deficient erythrocytes
subjected to oxidative stress. Measurements were performed on both
G6PD-deficient and normal control erythrocytes after exposure to
oxidative stress with t-BHP. (A) TBARS evaluation of incubation
medium; (B) methaemoglobin content in erythrocyte cytosol.
Fig. 4. Membrane protein methylation levels and PK activity of G6PD-
deficient erythrocytes subjected to oxidative stress. (A) PK activity, as a

and integral membrane proteins glycophorin C and band 3
(AE1) [30].
The functional consequences of deamidation/isomeriza-
tion have often been investigated under near-pathological
conditions. Homozygous knockout mice for PCMT are
affected by growth retardation, and die prematurely with
tonic-clonic seizures [31,32]. In these animals, isomerized
proteins accumulate in all organs and tissues, indicating lack
of PCMT-driven repair activity [31,32]. However, the
functional outcome of such alterations on individual
proteins is still uncertain, although the biological activity
of different proteins appears to be compromised in vitro by
deamidation and isomerization. Previous experience with
several cell models has shown that the isoaspartate content
of intracellular proteins is increased as the result of heat
shock [33] as well as of UVA irradiation [34]. As far as the
Fig. 5. SDS/PAGE profile of membrane proteins from G6PD-deficient
and normal erythrocytes subjected to oxidative stress. Oxidative stress
was induced, where indicated, by t-BHP treatment. Lane 1, nonoxi-
dizednormalerythrocyte;lane2,oxidizednormalerythrocyte;lane3,
nonoxidized G6PD-deficient erythrocyte; lane 4, oxidized G6PD-
deficient erythrocyte.
Fig. 6. Schematic representation of the overall hypothesis on the relationships between oxidative stress and isoaspartate formation in G6PD deficiency.
G6PD-deficient erythrocytes are intrinsically less resistant to subliminal oxidant levels, so that protein deamidation/isomerization products (i.e.
isoaspartate residues) tend to accumulate despite the fact that the life span of these cells is, on average, shorter than normal. In other words, they
reach levels of aspartate damage that are typical of a much older normal erythrocyte population. Exposure to certain foods or drugs (fava beans,
nonsteroidal anti-inflammatory drugs, antimalaria drugs, chemotherapeutics, etc.) trigger the haemolytic crisis, which is also associated with a
further increase in the levels of deamidated/isomerized proteins. The mechanism linking oxidation to haemolysis involves membrane alterations.
Ó FEBS 2002 Protein methylation in G6PD-deficient erythrocytes (Eur. J. Biochem. 269) 2037
erythrocyte is concerned, there is evidence that deamida-

deficiency, which renders erythrocyte adaptation to an
oxidative microenvironment more difficult, makes mem-
brane proteins more prone to isoaspartate formation, both
during cell ageing and, even more so, under stress conditions
(see scheme in Fig. 6). Taken as a whole, the results support
the role of this post-biosynthetic protein modification in the
mechanism of haemolysis in G6PD deficiency.
ACKNOWLEDGEMENTS
Genetic testing of patients was accomplished at the International
Institute of Genetic and Biophysics (I.I.G.B.) of the National Research
Council, Naples, Italy, under the supervision of Dr Giuseppe Martini
and Stefania Filosa. The work was supported in part by research grants
from Ministero dell’Istruzione, dell’Universita
`
e della Ricerca, Progetti
di Rilevante Interesse Nazionale (M.I.U.R. P.R.I.N., 1999): ÔExtra and
intracellular nucleotide and nucleoside: chemical signals, metabolic
regulators and potential drugsÕ and ÔHyperhomocysteinemia as a
cardiovascular risk factor: biochemical mechanism(s)Õ.
REFERENCES
1. Luzzatto, L., Mehta, A. & Vulliamy, T. (2001) Glucose 6-phos-
phate dehydrogenase deficiency. In The Metabolic and Molecular
Bases of Inherited Disease (Scriver,C.R.,Beaudet,A.L.,Sly,W.S.
& Valle, D., eds), vol. 3, pp. 4517–4553. McGraw-Hill, New York.
2. Mason, J.P. (1996) New insights into G6PD deficiency. Br.
J. Haematol. 94, 585–591.
3. Beutler, E., Vulliamy, T. & Luzzatto, L. (1996) Hematological
important mutation: glucose-6-phosphate dehydrogenase. Blood
Cell Mol. Dis. 22, 49–56.
4. Pandolfi, P.P., Sonati, F., Rivi, R., Mason, P., Grosveld, F. &

12. Ryttersgaard, C., Griffith, S.C., Sawaya M.R., MacLaren D.C.,
Clarke, S. & Yeates, T.O. (2002) Crystal structure of human
L
-isoaspartyl methyltransferase. J. Biol. Chem., Manuscript
M200229200.
13. Ota, I.M. & Clarke, S. (1990) The function and enzymology of
protein
D
-aspartyl/
L
-isoaspartyl methyltransferase in eukaryotic
and prokaryotic cells. In Protein Methylation (Paik, W.K. & Kim,
S., eds), pp. 179–194. CRC Press, Boca Raton, FL.
14. Ingrosso, D., Cotticelli, M.G., D’Angelo, S., Buro, M.D., Zappia,
V. & Galletti, P. (1997) Influence of osmotic stress on protein
methylation in reasealed erythrocytes. Eur. J. Biochem. 244,918–
922.
15. Ingrosso, D., D’Angelo, S., di Carlo, E., Perna, A.F., Zappia, V. &
Galletti, P. (2000) Increased methyl esterification of altered
aspartyl residues in erythrocyte membrane proteins in response to
oxidative stress. Eur. J. Biochem. 267, 4397–4405.
16. Beutler, E., Blume, K.G., Lohr, G.W., Ramot, B. & Valentine,
W.N. (1977) International committee for standardization in hae-
matology: recommended methods for red-cell enzyme analysis.
Br. J. Haematol. 35, 331–341.
17. Vettore, L., De Matteis, M.C. & Zampini, P. (1980) A new density
gradient system for the separation of human red blood cells.
Am. J. Hematol. 8, 291–297.
18. Sykes, B.C. (1983) DNA in heritable disease. Lancet 2, 787–788.
19. Alfinito, F., Cimmino, A., Ferraro, F., Cubellis, M.V., Vitagliano,

throcytes. J. Biol. Chem. 263, 1386–1130.
27. Vlassara, H. (1990) Advanced non-enzymatic tissue glycosylation:
mechanism implicated in the complications associated with aging.
In Molecular Biology of Aging (Finch, C.E. & Johnson, T.E., eds),
pp. 171–185. Wiley-Liss, New York.
28. Ladino, C.A. & O’Connor, C.M. (1991) Identification of a site for
carboxyl methylation in human a-globin. Biochem. Biophys. Res.
Commun. 180, 742–747.
29. Bonaventura, J., Bonaventura, C., Sullivan, B., Ferruzzi, G.,
McCurdy, P.R., Fox, J. & Moo-Penn, W.F. (1976) Hemoglobin
providence. Functional consequences of two alterations of the
2,3,diphosphoglycerate binding site at position b82. J. Biol. Chem.
251, 7563–7571.
30. Cohen, C.M. & Gastard, P. (1992) Regulation and post-trasnla-
tional modofocation of erythrocyte membrane protein and
membrane skeletal proteins. Semin. Hematol. 29, 244–292.
31. Kim, E., Lowenson, J.D., MacLaren, D.C., Clarke, S. & Young,
S.G. (1997) Deficiency of a protein-repair enzyme results in
the accumulation of altered proteins, retardation of growth,
and fatal seizures in mice. Proc. Natl. Acad. Sci. USA 94,
6132–6137.
32. Kim, E., Lowenson, J., Clarke, S. & Young, S.G. (1999) Pheno-
typic analysis of seizure-prone mice lacking
L
-isoaspartate
(
D
-aspartate) O-methyltransferase. J. Biol. Chem. 274, 20671–
20678.
33. Ladino, C.A. & O’Connor, C.M. (1992) Methylation of atypical