A quantitative analysis of spontaneous isoaspartate formation from N-terminal asparaginyl and aspartyl residues

Article abstract of DOI:10.1007/s00726-012-1454-0

The formation of isoaspartate (isoAsp) from asparaginyl or aspartyl residues is a spontaneous post-translational modification of peptides and proteins. Due to isopeptide bond formation, the structure and possibly function of peptides and proteins is altered. IsoAsp modifications within the peptide chain have been reported for many cytosolic proteins. Amyloid peptides (Aβ) deposited in Alzheimer's disease may carry an N-terminal isoAsp-modification. Here, we describe a quantitative investigation of isoAsp-formation from N-terminal Asn and Asp using model peptides similar to the Aβ N-terminus. The study is based on a newly developed separation of peptides using capillary electrophoresis (CE). ¹H NMR was employed to validate the basic finding of N-terminal isoAsp-formation from Asp and Asn. Thereby, the isomerization of Asn at neutral pH (0.6 day^-1, peptide NGEF) is approximately six times faster than that within the peptide chain (AANGEF). The difference in velocity between Asn and Asp isomerization is approximately 50-fold. In contrast to N-terminal Asn, Asp isomerization is significantly accelerated at acidic pH. The kinetic solvent isotope (k _D2O/k _H2O) effect of 2.46 suggests a rate-limiting proton transfer in isoAsp-formation. The proton inventory is consistent with transfer of one proton in the transition state, supporting the previous notion of rate-limiting deprotonation of the peptide backbone amide during succinimide-intermediate formation. The study provides evidence for a spontaneous N-terminal isoAsp-formation within peptides and might explain the accumulation of N-terminal isoAsp in amyloid deposits.

Full text of DOI:10.1007/s00726-012-1454-0

Amino Acids (2013) 44:1205–1214
DOI 10.1007/s00726-012-1454-0
ORIGINAL ARTICLE
A quantitative analysis of spontaneous isoaspartate formation
from N-terminal asparaginyl and aspartyl residues
•
•
Bert H.-O. Gu¨ttler Holger Cynis
•
Franziska Seifert Hans-Henning Ludwig
•
•
Andrea Porzel Stephan Schilling
Received: 18 September 2012 / Accepted: 24 December 2012 / Published online: 24 January 2013
ꢀ Springer-Verlag Wien 2013
Abstract The formation of isoaspartate (isoAsp) from
asparaginyl or aspartyl residues is a spontaneous post-
translational modification of peptides and proteins. Due to
isopeptide bond formation, the structure and possibly
function of peptides and proteins is altered. IsoAsp modi-
fications within the peptide chain have been reported for
many cytosolic proteins. Amyloid peptides (Ab) deposited
in Alzheimer’s disease may carry an N-terminal isoAsp-
modification. Here, we describe a quantitative investigation
of isoAsp-formation from N-terminal Asn and Asp using
model peptides similar to the Ab N-terminus. The study is
based on a newly developed separation of peptides using
times faster than that within the peptide chain (AANGEF).
The difference in velocity between Asn and Asp isomeri-
zation is approximately 50-fold. In contrast to N-terminal
Asn, Asp isomerization is significantly accelerated at acidic
pH. The kinetic solvent isotope (kD2O/kH2O) effect of 2.46
suggests a rate-limiting proton transfer in isoAsp-forma-
tion. The proton inventory is consistent with transfer of one
proton in the transition state, supporting the previous
notion of rate-limiting deprotonation of the peptide back-
bone amide during succinimide-intermediate formation.
The study provides evidence for a spontaneous N-terminal
isoAsp-formation within peptides and might explain the
accumulation of N-terminal isoAsp in amyloid deposits.
1
capillary electrophoresis (CE). H NMR was employed to
validate the basic finding of N-terminal isoAsp-formation
from Asp and Asn. Thereby, the isomerization of Asn at
neutral pH (0.6 day-1, peptide NGEF) is approximately six
Keywords Alzheimer’s disease ꢀ PIMT ꢀ Capillary
1
electrophoresis ꢀ H NMR ꢀ Isoaspartate
Introduction
B. H.-O. Gu¨ttler ꢀ H. Cynis ꢀ F. Seifert ꢀ H.-H. Ludwig ꢀ
S. Schilling (&)
Probiodrug AG, Weinbergweg 22, Biozentrum,
06120 Halle (Saale), Germany
e-mail: stephan.schilling@probiodrug.de
The formation of isoaspartate (isoAsp) from asparagine
(L-Asn) or aspartate (L-Asp) represents one common post-
translational modification, which is considered to deter-
mine the half-life of proteins (Robinson and Rudd 1974;
Robinson and Robinson 2001; Robinson et al. 1970).
Under mild chemical or physiological conditions, L-Asn
and L-Asp peptide bonds form a succinimide under con-
secutive release of ammonia (Robinson et al. 1970; Fig. 1)
or water, respectively. The intermediate is prone to spon-
taneous hydrolysis resulting in the formation of isoAsp or
Asp. Usually, L-asparaginyl residues decompose into
L-isoAsp and L-Asp in a ratio of approximately 3:1 to 5:1,
depending on the buffer and pH conditions (Meinwald
et al. 1986; Geiger and Clarke 1987; Patel and Borchardt
1990). Epimerization of the succinimide-intermediate may
Present Address:
H. Cynis
Brigham and Women’s Hospital, Harvard Medical School,
NRB 363, 77 Avenue Louis Pasteur, Boston, MA 02115, USA
Present Address:
F. Seifert
Institute of Pharmacy, Martin-Luther-University
Halle-Wittenberg, Weinbergweg 22,
06120 Halle (Saale), Germany
A. Porzel
Department of Bioorganic Chemistry, Leibniz Institute of Plant
Biochemistry, Weinberg 3, 06120 Halle (Saale), Germany
123

1206
B. H.-O. Gu¨ttler et al.
(Aswad et al. 2000; Geiger and Clarke 1987), conse-
quently altering its structure. Therefore, the isoAsp-
modification may cause also different protein function
(George-Nascimento et al. 1990; Velazquez et al. 1997;
Noguchi 2010; Corti and Curnis 2011). It was suggested
that Asn—because of a rapid isomerization as compared
with Asp functions as a molecular timer for protein
ageing, development, protein turnover and other biolog-
ical events (Robinson et al. 1970; Robinson 1974;
Robinson and Robinson 2001). The enzyme protein
L-isoaspartyl O-methyltransferase (PIMT, EC 2.1.1.77)
represents a cellular repair mechanism to recycle isoAsp-
to Asp-residues (Clarke 1985). PIMT utilizes the cofactor
S-adenosylmethionine (SAM) to transfer a methyl group
onto the a-carboxyl group of isoAsp leading to succin-
imidyl intermediate formation followed by subsequent
hydrolysis to L-Asp and L-isoAsp (compare to Fig. 1).
Known substrates of PIMT comprise the histone proteins
H2A and H2B or synapsin I and II (Young et al. 2001,
2005; Carter and Aswad 2008). Recent results suggest
that PIMT might also act as marker for protein degra-
dation as shown for p53 (Lee et al. 2012), initiator for
chromatin remodeling as shown for histone H2A and
H2B or function as transcription activator (Park et al.
2012).
Possibly caused by lower protein turnover during ageing
and pathogenesis, the formation of isoAsp has been impli-
cated in the pathogenesis of autoimmune and neurodegen-
erative disorders. A detailed characterization of Ab-peptides
and neurofibrillary tangles deposited in Alzheimer’s disease
revealed presence of isoAsp in substantial quantity (Saido
et al. 1996; Shimizu et al. 2000). Isomerization of Asp1 and
Asp7 within the Ab sequence (D1AEFRHD7SGYE…40/42)
has been reported, which has been shown to enhance its
fibril formation propensity (Roher et al. 1993; Saido et al.
1996). Previous studies revealed that isoAsp influences also
the susceptibility of the amyloid precursor protein to
cleavage and thereby the molecular pathways of Ab pro-
duction: Formation of isoAsp at position 1 of Ab enhanced
cleavage by Cathepsin B, an alternative b-secretase,
whereby cleavage by the b-secretase BACE 1 was not
Fig. 1 Spontaneous deamidation of L-asparaginyl residues. Deam-
idation leads to formation of a succinimidyl intermediate that is
hydrolyzed to aspartyl or isoaspartyl residues. The formation and
resolution of the tetrahedral intermediate to form the succinimidyl
intermediate involves proton transition(s). L-aspartyl residues are also
undergoing such reactions (not shown) but at slower rate. On the state
of the succinimidyl intermediate, epimerization at the Ca carbon
might occur (not shown) leading to minute amounts of D-aspartyl and
D-isoaspartyl residues. L-isoaspartyl and D-aspartyl residues are prone
to methylation by PIMT, building up protein repair mechanism of the
cells (Lowenson and Clarke 1992)
¨
detected (Bohme et al. 2008a, b). Thus, occurrence of iso-
Asp within the peptide chain of APP might occur prior to Ab
release and, therefore, account for isoAsp-modified Ab
within deposits.
To further investigate the formation of isoAsp in pep-
tides, we aimed at an analysis of N-terminal isoAsp-for-
mation. Because previous studies did not focus on such
analysis, we developed a new assay based on the separation
of model peptides using capillary electrophoresis (CE-UV).
The results should provide further insights into the role and
formation of isoAsp at the N-terminus, which might have
also implications for, e.g., stability of protein drugs.
lead to generation of minute amounts of D-isoAsp and
D-Asp.
The isoAsp-formation introduces an additional meth-
ylene group into the backbone of the protein or peptide
123

Quantitative analysis of spontaneous isoaspartate formation
Materials and methods
1207
1H nuclear magnetic resonance
1H NMR spectra were recorded at 25 ꢁC on an Agilent
VNMRS 600-MHz spectrometer equipped with an inverse
detection cryoprobe. NMR samples contained 1.73 mM of
peptide in buffer without addition of D2O. 40 transients
were acquired. The water signal was suppressed using a 2 s
pre-saturation pulse. A 90ꢁ excitation pulse (6.1 ls) was
used with a pulse repetition rate of 15 s.
Synthesis and purification of model peptides
Peptides were generated as C-terminal amides in a
0.5 mmol scale on an automated peptide synthesizer
(Endeavour 90, AAPPTec) using Fmoc strategy. The pep-
tides were synthesized on a Rink Amide MBHA resin
(MERCK Biosciences). A twofold excess of Fmoc-AA-OH
and 2-(1H-benzotriazol-1-yl)-1,1,3,3-tetramethyluronium
tetrafluoroborate (TBTU) in dimethylformamide (DMF)
and 4 eq. N,N-Diisopropylethylamine for the coupling
process were used. The reaction time was approximately
25 min and each coupling step was repeated for one time.
For de-protection of the Fmoc group, a 20 % piperidine
solution in DMF was applied. The peptides were cleaved
from the resin with a mixture of 95 % trifluoroacetic acid
(TFA), 2.5 % water and 2.5 % triisopropylsilane (TIS) for
2 h. The product was evaporated and re-crystallized from
cold methanol/ether. The crude products were purified by
Determination of pKa values
Titrations have been conducted using a Mettler Toledo
DL50 Graphix equipped with a DG111-SC glass electrode.
Standard solutions of 0.1 N NaOH and 0.1 N HCl (Roth,
Karlsruhe, Germany) were applied. The peptides were
dissolved in distilled water and pre-incubated at 37 ꢁC. The
titration was performed at 37 ꢁC and the curve was eval-
uated using DL50 Graphix software. Each determination
has been performed in triplicate at least.
˚
RP-HPLC on a Luna 10 lm C18 column (100 A,
250 9 21.20 mm, Phenomenex) applying an acetonitrile/
Solvent isotope effect and proton inventory
water gradient (0.04 % TFA).
To determine the kinetic solvent isotope effect (KIE) of the
isomerization process, all buffers were prepared in water or
deuterium oxide. The pH and pD values have been adjusted
according to the formula: pD = (pH reading) ? 0.4 (Cook
1991). The solvent KIE (k(D2O)/k(H2O)) of isoAsp-formation
was determined with NGEF and D(OMe)GEF in three-com-
ponent buffer (12.5 mM MES, 12.5 mM acetic acid, 25 mM
Tris) at pH/pD 8.0and additionally for NGEF in tricine buffer
at pH/pD 8.0 to rule out potential influence of buffer ions.
To assess the proton inventory, the sample buffer was pre-
pared using H2O and D2O (at pH = pD) as solvent and mixed.
Thereafter, 1 mM of the peptides (NGEF or D(OMe)GEF) were
dissolved in H2O- and D2O-containing sample buffer at
pH/pD 8.0. Both peptide solutions were mixed in ratios of
80 % (v/v) H2O to 20 % (v/v) D2O; 60 % (v/v) H2O to 40 %
(v/v) D2O; 40 % (v/v) H2O to 60 % (v/v) D2O and 20 % (v/v)
H2O to 80 % (v/v) D2O. All peptide solutions were incubated
at 37 ꢁC and subsequently analyzed by CE-UV. For analysis,
the isomerization velocity was plotted against the mole frac-
tion of D2O. The resulting graph was analyzed using the pro-
gram ‘‘GraphPad Prism 4’’ from GraphPad Software Inc.
Analysis of isoAsp-formation using capillary
electrophoresis (CE-UV)
Peptides were incubated at 37 ꢁC using a sample buffer
containing 12.5 mM sodium acetate, 12.5 mM 2-(N-Mor-
pholino)ethanesulfonic acid (MES) and 25.0 mM
Tris(hydroxymethyl)aminomethane (Tris) ensuring a con-
stant ionic strength over a broad pH range (Ellis and Mor-
rison 1982). In addition, 0.02 % (w/v) sodium azide was
added to preserve sterile conditions. The pH value was
adjusted by hydrochloric acid or sodium hydroxide. The
incubation time ranged from minutes to months depending
on the peptide and the pH value.
The peptides have been separated using a P/ACETM
MDQ Glycoprotein System (Beckman CoulterTM, USA,
Reorder No.: 338472) equipped with a capillary tubing of
30 cm in length, 50 lm inner diameter. The separation was
accomplished in 50 mM citric acid, pH 2.0 at 20 ꢁC. The
absorption was detected at a wavelength of 200 5 nm.
The peptides NGEF, DGEF, DisoGEF, NAEF, DAEF,
D
isoAEF and D(OMe)GEF have been separated at 20 kV and
0.2 psi. Acetyl-NGEF, Acetyl-DGEF and Acetyl-DisoGEF
have been separated at 20 kV without applying additional
pressure. AANGEF, AADGEF and AADisoGEF have been
analyzed at 5 kV without pressure. After every run, the
capillary was flushed with 0.2 M NaOH, 1.0 N HCl fol-
lowed by running buffer. The data were evaluated by
comparing the peak areas of the model peptides and the
internal standard peptide HY.
Results
Capillary electrophoresis for analysis of N-terminal
isoAsp-formation
In a first attempt to chromatographically separate the model
peptides NGEF, DGEF and DisoGEF, HPLC using a
123

1208
B. H.-O. Gu¨ttler et al.
standard C-18 column was performed. A baseline separa-
tion could not be achieved (data not shown).
To standardize separation conditions in different approa-
ches, the relative migration time was calculated on the basis
of the internal standard. Accordingly, HY possessed a rel-
ative migration time of 1 (actual migration time 2.6 min),
followed by NGEF at 1.4 min, DGEF at 1.55 min and Dis-
oGEF at a migration time of 2.0 min (Fig. 2a). A similar
separation was observed with the other model peptides in
this study (see below, data not shown). In either case, the
order of migration was as shown in Fig. 2a, suggesting that
the differences in the pKa values caused by isoAsp-forma-
tion always result in changes of the net charge, making CE a
valuable tool to analyze peptide conversion in a quantitative
manner.
However, based on different pKa values of Asp and
isoAsp with regard to the carboxylic group [i.e., Asp:
pKa, a-COOH = 1.99, pKa, b-COOH = 3.90 (Wagenknecht and
Knapp 2009)], the formation of isoAsp was expected to
result in a switch of the net charge of the peptides at acidic
pH. As a consequence, the electrophoretic movement should
change, with isoAsp-modified peptides migrating slower
than the corresponding Asp and Asn containing peptides.
Indeed, choosing an acidic pH of 2.0 as electrolyte solution,
a separation of NGEF, DGEF and DisoGEF was observed
(Fig. 2a). In addition, an internal standard (the dipeptide
HY) was baseline-separated from the standard peptides.
For evaluation of the isoAsp content, the internal stan-
dard peptide HY was applied at a constant concentration,
and the relative UV signal areas at 200 nm calculated.
A linear relationship of the ratio was observed within a
broad concentration range (Fig. 2b).
Quantitative analysis of N-terminal isoAsp-formation
To investigate the N-terminal isoAsp-formation, a model
peptide (NGEF), which was designed to show a rapid
Fig. 2 a Separation of the peptides NGEF, DGEF and DisoGEF and
the standard peptide HY using capillary electrophoresis. The peptides
were detected by UV absorption at 200 nm. To standardize the
separation, the relative migration time is shown, applying HY as
standard. Accordingly, HY is detected at a relative migration time of
1, followed by NGEF at 1.4, DGEF at 1.6 and DisoGEF at 2.
b Dependence of the relative absorption of NGEF, D(OMe)GEF, DGEF
and DisoGEF on the peptide concentration after separation and
detection. HY served as the internal standard peptide. The concen-
tration of the model peptide has been plotted against the quotient of
peak area of the model peptide divided by the peak area of the internal
standard peptide. Virtually identical slopes suggest equivalent
concentrations and absorption of the peptide standards, enabling
analysis of deamidation/isoAsp-formation
Fig. 3 a Spontaneous formation of DisoGEF and DGEF from NGEF
at 37 ꢁC and pH 7.0. The model peptide NGEF (1 mM) was incubated
for the indicated times, samples removed and the peptide composition
analyzed using CE-UV. Upon incubation, the signal for NGEF
decreases accompanied by increased signals of DisoGEF. b Spontane-
ous deamidation of NGEF, assessed using 1H NMR. The model
peptide NGEF was incubated at pH 7.0 and 37 ꢁC. Samples were
removed after the indicated times and analyzed by 1H NMR. The
integral of the superimposed E H-a signals of both NGEF and
DisoGEF (approx. 4.16 ppm) was used as an internal standard. NGEF
and DisoGEF are characterized by the well-separated H-a signals of N
(4.23 ppm) and Diso (4.05 ppm), respectively
123

Quantitative analysis of spontaneous isoaspartate formation
1209
Table 1 Quantitative analysis of isoAsp-formation determined by
1
CE-UV and H NMR
Incubation time (h) of
NGEF (1 mg/mL)
1H NMR
[DisoGEF] (%)
CE-UV
[DisoGEF] (%)
at 37 ꢁC and pH 7.0
24
48
72
15
24
37
18
26
40
Illustrated is the comparison of generated amount of IsoAsp from
1 mg/mL NGEF after incubation for 24, 48 and 72 h at 37 ꢁC and pH
7. The quantification was carried out by H NMR and by CE-UV
1
deamidation due to the adjacent Gly residue, was incubated
at 37 ꢁC and pH 7. At different time points, samples were
taken and the peptides separated using CE-UV. As shown
in Fig. 3a, during the incubation period of 118 h a signif-
icant formation of DisoGEF could be observed. Moreover,
minute amounts were already detectable after 3 h of
incubation. As expected, also DGEF originating from
deamidation of NGEF occurred in the mixture leading to a
final ratio of DisoGEF:DGEF of approximately 9:1.
To validate these results, in a following approach the
formation of isoAsp at the N-terminus was assessed using
1H NMR (Fig. 3b). Proton spectra of the model peptides
NGEF, DGEF and DisoGEF were recorded for signal
assignment. For quantification, the integral intensities of
the well-separated H-a signals of N and Diso were used,
normalized to the integral of the superimposed E H-a
1
signals of both, NGEF and DisoGEF. The H NMR-based
quantification shows good conformity with the results of
the capillary electrophoresis analysis (Table 1). On the
basis of this validation, the further analysis of spontaneous
isoAsp-formation in DGEF, NAEF, DAEF, (Ac)NGEF,
AANGEF and D(OMe)GEF was evaluated using CE-UV.
Fig. 4 Analysis of spontaneous deamidation/isoAsp-formation from
NGEF and DGEF. a Formation of DisoGEF and DGEF from NGEF at
pH 9.0 and 37 ꢁC; b Decomposition of NGEF at pH 3, 5, 7 and 9;
c Formation of DisoGEF from DGEF at pH 3, 5, 7 and 9. The peptides
were incubated at 37 ꢁC and samples were removed consecutively.
The concentration of the peptides has been determined using CE-UV
as described before. Data have been analyzed according to a model of
first order, as indicated by the linearization. A0 refers to the initial
concentration of the Asx peptide, X denotes the concentration of
product at a given time
Rate of spontaneous isoAsp-formation in different
model peptides
In a first approach to quantify the spontaneous deamidation
of NGEF, we investigated the progress of DisoGEF and
DGEF formation at pH 9.0 (Fig. 4a). The rapid decompo-
sition of NGEF was accompanied by formation of
D
isoGEF. Also DGEF was formed, however, the concen-
tration was always below 10 %, indicating that N-terminal
deamidation to form Asp is virtually negligible. The
decomposition rate fits best to a first-order kinetic model,
as suggested by linearization of the data (Fig. 4b).
(Fig. 4b). The rate of spontaneous deamidation and isoAsp
formation was fastest at basic pH and strictly declined at
acidic pH. The deamidation of Asn thereby yields Asp and
isoAsp residues at a pH-independent ratio of *1:9, effi-
ciently ruling out deamidation leading directly to Asp,
which accords to previous observations (Capasso et al.
1993). Also N-terminal Asp was found to isomerize over
time (Fig. 4c), although at significantly lower rate. In
To evaluate the rate of spontaneous isoAsp-formation,
several model peptides with Asx at the N-terminus or
within the peptide chain have been investigated at different
pH conditions. The rate of spontaneous isoAsp-formation
from N-terminal Asn in NGEF was clearly pH-dependent
123

1210
B. H.-O. Gu¨ttler et al.
contrast to deamidation of Asn, the rate of isomerization
was low at neutral pH. The highest rate of isoAsp-forma-
tion was observed at acidic pH.
The kinetic data of spontaneous deamidation and Asp
isomerization of all peptides investigated are summarized
in Table 2. In general, the isoAsp-formation from Asn
was found to be fastest at basic pH, whereas N-terminal
Asp-residues show fastest spontaneous isomerization
under basic and acidic conditions and a corresponding
minimum between pH 5 and 7. At neutral pH-conditions,
deamidation of Asn in NGEF is about 60 times faster than
isoAsp-formation from DGEF. The difference was sig-
nificantly lower with NAEF and DAEF, suggesting a
differential influence of the second amino acid on velocity
of isoAsp-formation. At pH 3.0, isoAsp is even faster
generated from N-terminal Asp as compared with N-ter-
minal Asn.
To investigate the role of the free a-amino group for
spontaneous isoAsp-formation, we assessed acetylated
NGEF [(Ac)NGEF]. As depicted in Table 2, the pH-
dependency of the isomerization of (Ac)NGEF was similar
to NGEF, however, the rate of isomerization was eightfold
slower. This was further substantiated by analysis of the
peptide AANGEF, which showed a similar rate of isoAsp-
formation as compared with (Ac)NGEF (Table 2). The
results point to an accelerated isomerization of peptides
carrying a free a-amino group without effect on the relative
pH-dependence of the reaction.
To illustrate the role of the b-carboxyl group for
isoGEF formation from DGEF, the isomerization velocity
D
of an activated form of DGEF (D(OMe)GEF) was deter-
mined. Methylation of the side chain led to significant
increase of the reaction velocity by factors of 14 and 740 as
compared with NGEF and DGEF, respectively (Table 2).
The acceleration factor was 3,000-fold as compared with
DGEF at pH 9.0.
Solvent isotope effect on isoAsp-formation
To analyze, whether proton transfer is rate limiting for
N-terminal isoAsp-formation, we incubated peptides in
deuterium oxide (Fig. 5a). The isomerization of NGEF and
D
(OMe)GEF at pH/pD 8.0 was significantly lower in D2O.
The quantitative analysis of the isomerization velocities
revealed a KIE of 2.46 for NGEF and 2.42 for D(OMe)GEF,
suggesting that the transfer of at least one proton is rate-
limiting in the isomerization reaction.
In a separate investigation, we assessed the dependence
of the velocity on pD (Table 3). The rate dependence
mirrored that observed in H2O, suggesting slower isoAsp-
formation at low pD. The relative drop in the rate was
stronger in D2O, resulting in an increasing solvent isotope
effect with decreasing pH/pD.
123

Quantitative analysis of spontaneous isoaspartate formation
1211
Table 3 Kinetic isotope Effect in dependence of pH/pD value for
D(OMe)GEF
pD/pH
D
(OMe)GEF
KIE
[kH2O/kD2O]
[kD2O (day-1)]
8
7
5
3
5.42 0.04
2.79 0.03
2.42
2.56
3.36
3.75
0.38 0.01
0.0218 0.0004
Displayed is the decomposition velocity of D(OMe)GEF at pD 3, 5, 7
and 8 and the resulting kinetic isotope effect (KIE) at each pD value.
The KIE is increasing with decreasing pH/pD value
using HPLC (Stephenson and Clarke 1989; Patel and
Borchardt 1990; Capasso et al. 1993). Novel approaches
also relied on degradation of peptides using endoprotease
Asp-N and electron transfer dissociation (ETD) mass
spectrometry (Sargaeva et al. 2011; Ni et al. 2010).
Because the previous HPLC methods are not capable of
separating peptides with an N-terminal isoAsp from the
precursors, we here employed CE-UV for analysis. The
method takes advantage of the different pKa values of the
carboxyl groups in Asp and isoAsp at acidic pH. Because
the basic principle of different net charge of Asn, Asp and
isoAsp at pH of 2–3 applies generally, we used CE-UV
also for assessment of isoAsp-formation within the peptide
chain enabling a comparison. Under the applied conditions,
capillary electrophoresis allowed the analysis of the sample
within less than 8 min, requiring a sample volume of 5 ll.
Thus, the newly introduced method provides an opportu-
nity to rapidly analyze Asx isomerization, which might
have also implications for other experimental approaches
such as analysis of isoAsp-peptide conversion by protein
isoaspartyl methyltransferase (PIMT).
Fig. 5 a Spontaneous deamidation of NGEF and formation of
DisoGEF from D(OMe)GEF in deuterium oxide (D2O). Both peptides
(1 mM) have been incubated at 37 ꢁC and pH/pD 8. The data were
analyzed as before and the kinetic isotope effect (KIE) calculated
from kH2O and kD2O. The resulting KIE was 2.46 for NGEF and 2.42
for D(OMe)GEF. b Dependence of the rate constant of DisoGEF
formation on the mole fraction of deuterium in the sample (proton
inventory). NGEF (squares) and D(OMe)GEF (triangles, 1 mM both)
were incubated in mixtures of buffers (pH/pD 8.0), which have been
prepared with H2O or D2O. The resulting rate constants were plotted
depending on the mole fraction of deuterium in solution. The data are
consistent with transfer of one proton in the transition state. A0 refers
to the initial concentration of the Asx peptide, X denotes the
concentration of product at a given time
To further study the reaction, we determined the rate
constants in mixtures of H2O and D2O and plotted the rate
against the mole fractions of D2O (Fig. 5b). The proton
inventory is consistent with transfer of a single proton in
the transition state.
The formation of isoAsp at the N-terminus proceeds
readily at neutral pH. The calculated half-lives range from
79 days for the Ab-sequence DAEF to 140 min for the
isomerization of D(OMe)GEF. Although we observed an
accelerated peptide bond degradation with NGEF com-
pared to the modified N-termini in (Ac)NGEF and AAN-
GEF, the half–lives determined are within the range of
those reported previously for different proteins and pep-
tides (Stephenson & Clarke 1989; Geiger & Clarke 1987),
which are 1-100 days for Asn and 40-300 days for Asp,
depending on the peptide sequence. In accordance with our
analysis, the Asx-Gly linkages are rapidly degrading pep-
tide bonds. Thus, the general requirements and mechanism
of the reactions appear to follow the same principle. Such a
conclusion is finally also supported by our initial mecha-
nistic analysis. The data clearly suggest a differential pH-
dependence of isomerization between Asn and Asp resi-
dues. In contrast, modification of Asp in D(OMe)GEF results
in a similar pH-rate profile as observed with Asn (Table 2).
Discussion
The N-terminus of peptides and proteins might be prone to
different post-translational modifications such as acylation
or formation of pyroglutamic acid from glutaminyl and
glutamyl precursors. In contrast to these modifications,
which are catalyzed by different transferases, the N-termi-
nal formation of isoAsp is considered a purely spontaneous
reaction in vitro and in vivo. Here, we investigated the rate
and mechanism of N-terminal isomerization of Asn- and
Asp-residues, which may constitute the N-Terminus of
amyloid peptides deposited in Alzheimer’s disease.
Previous studies to analyze the isoAsp-formation within
the peptide chain were based on separation of products
123

1212
B. H.-O. Gu¨ttler et al.
Thus, the rate of isomerization is generally determined by
the electrophilicity of the b-carbonyl carbon. Likewise,
because the side chain of Asp is protonated at pH \ 3.0,
the rate of nucleophilic attack by the nitrogen of the
Asx-AA linkage and the Asp b-carbonyl carbon increases.
The significant enhancement of isoAsp-formation from
described (He and Barrow 1999; Schilling et al. 2006;
Kummer et al. 2011). Among those, N-terminally truncated
and pyroglutamate (pE) modified Ab (Ab3pE-42) gained
considerable interest due to high aggregation propensity
and oligomer toxicity (Schlenzig et al. 2009; Nussbaum
et al. 2012). Compared with spontaneous formation of pE
at the N-terminus of Ab, which exhibits a half-life of
*10 years at neutral pH (Seifert et al. 2009), the N-ter-
minal isoAsp-formation in the model peptide DAEF (rep-
resenting the N-terminus of Ab 1–4) is rather fast
(78 days). In contrast to isoAsp, however, pE-formation
has been shown to be matter of enzymatic catalysis in vivo
(Schilling et al. 2004, 2008). Hence, the results of the
present study can be reconciled with a spontaneous for-
mation of N-terminal isoAsp after generation and deposi-
tion of Ab in AD pathology. However, it should be
considered that previous studies have shown that isoAsp-
modifications in the Ab precursor protein APP alter the
D
(OMe)GEF and the pH-dependence are all consistent with
a rate-determining formation of the succinimidyl-inter-
mediate (which is identical formed from NGEF, DGEF and
D
(OMe)GEF). It has been suggested previously that a de-
protonation of the attacking nitrogen might be determining
for the rate of the overall reaction at pH \ 3 (Capasso et al.
1993) (see also Fig. 1). The solvent kinetic isotope effect
and the proton inventories of our study clearly support such
a conclusion. The equal size of the isotope effect and the
slope of unity obtained with NGEF and D(OMe)GEF support
transfer of one single proton in the transition state. The
general finding of a rate-limiting proton transfer is, how-
ever, not surprising because numerous acyl transfer reac-
tions have been described for which such limiting factors
are concluded (Satterthwait and Jencks 1974; Cox and
Jencks 1978; Barnett and Jencks 1969).
¨
processing and the liberation of Ab (Bohme et al. 2008b).
These reports are accompanied by several epidemiological
studies that point to a crucial influence of the sequence
surrounding the b-secretase cleavage site for generation of
the N-terminus of Ab (Haass et al. 1995; Di Fede et al.
2009; Jonsson et al. 2012). Therefore, it cannot be excluded
that an isoAsp-formation—at least partially—precedes the
APP cleavage, which, in turn, modifies the cellular
processing.
It remains unclear, though, whether a formation of isoAsp
at the N-terminus affects the function of certain proteins in a
similar manner as described for isoAsp within the peptide
chain (George-Nascimento et al. 1990; Velazquez et al. 1997;
Weber and McFadden 1997; Corti and Curnis 2011). Con-
sidering the half-lives of N-terminal Asx of days to months to
form isoAsp, the turnover of most intracellular proteins is
probably too high (Varshavsky 1996) to yield accumulation
of the N-terminally modified proteins in considerable
amounts. In contrast, amyloid peptides deposited in extra-
cellular space such as Ab have been described to contain
isoAsp attheN-terminus and within thechain, possibly dueto
low if any turnover (Roher et al. 1993; Lowenson et al. 1994;
Saido et al. 1996). These observations are also in line with the
enzymatic properties and distribution of PIMT. Previous
studies on the substrate specificity of PIMT suggested that
conversion of isoAsp strongly depends on the size and amino
acid constitution of the substrate (Lowenson and Clarke
1990; Lowenson and Clarke 1991). Already these studies
pointed towards an exceptionally inefficient conversion of a
dipeptide containing isoAsp at the N-terminus (Lowenson
and Clarke 1991). Likewise, in initial experiments we
observed only poor methylation of isoDAEF by PIMT (not
shown). Moreover, previous studies revealed that PIMT is
primarily distributed to intracellular compartments or may be
retained within the endoplasmatic reticulum (MacLaren et al.
1992; Potter et al. 1992; Lowenson et al. 1994). Thus,
extracellular Ab might escape the cellular isoAsp repair
system (Lowenson et al. 1994).
In summary, we performed the first quantitative analysis
of isoAsp-formation at the N-terminus of model peptides.
The study has been enabled by introducing CE-UV for
separation and quantification of substrates and products.
Compared to in-chain Asx residues, the half-lives of the
N-terminal sites appear shorter; however, the rate constants
are generally within the range of those described previ-
ously. The results provide some clues for the appearance of
isoAsp at the N-terminus of deposited amyloid peptides,
and might be also important for other proteins which are
stored for prolonged time, e.g., protein drugs.
Acknowledgments We are grateful to Prof. Dr. H.-U. Demuth for
his helpful comments and suggestions on the manuscript. We thank
Dr. T. Hoffmann for his technical support. This work was financially
supported by the Investitionsbank Sachsen-Anhalt, Grant#
1004/00082 to Probiodrug AG (StS).
Conflict of interest BG, HC, FS, H-HL and StS are former or
present employees of Probiodrug AG.
References
Aswad DW, Paranandi MV, Schurter BT (2000) Isoaspartate in
peptides and proteins: formation, significance, and analysis.
J Pharm Biomed Anal 21:1129–1136
In addition to isoAsp formation, also other post-trans-
lational modifications of amyloid peptides have been
123

Quantitative analysis of spontaneous isoaspartate formation
1213
Barnett RE, Jencks WP (1969) Diffusion-controlled proton transfer in
intramolecular thiol ester aminolysis and thiazoline hydrolysis.
J Am Chem Soc 91:2358–2369
Lee JC, Kang SU, Jeon Y, Park JW, You JS, Ha SW, Bae N, Lubec G,
Kwon SH, Lee JS, Cho EJ, Han JW (2012) Protein L-isoaspartyl
methyltransferase regulates p53 activity. Nat. Commun. 3:927
Lowenson JD, Clarke S (1990) Identification of isoaspartyl-contain-
ing sequences in peptides and proteins that are usually poor
substrates for the class II protein carboxyl methyltransferase.
J Biol Chem 265:3106–3110
¨
¨
Bohme L, Bar JW, Hoffmann T, Manhart S, Ludwig HH, Rosche F,
Demuth HU (2008a) Isoaspartate residues dramatically influence
substrate recognition and turnover by proteases. Biol Chem
389:1043–1053
¨
Bohme L, Hoffmann T, Manhart S, Wolf R, Demuth HU (2008b)
Lowenson JD, Clarke S (1991) Structural elements affecting the
recognition of L-isoaspartyl residues by the L-isoaspartyl/D-
aspartyl protein methyltransferase. Implications for the repair
hypothesis. J Biol Chem 266:19396–19406
Isoaspartate containing amyloid precursor protein derived pep-
tides alter efficacy and specificity of potential beta-secretases.
Biol. Chem. 389:1055–1066
Capasso S, Mazzarella L, Sica F, Zagari A, Salvadori S (1993)
Kinetics and mechanism of succinimide ring formation in the
deamidation process of asparagine residues. J Chem Soc Perkin
Trans 2:679–682
Lowenson JD, Clarke S (1992) Recognition of D-aspartyl residues in
polypeptides by the erythrocyte L-isoaspartyl/D-aspartyl protein
methyltransferase. Implications for the repair hypothesis. J Biol
Chem 267:5985–5995
Carter WG, Aswad DW (2008) Formation, localization, and repair of
L-isoaspartyl sites in histones H2A and H2B in nucleosomes
from rat liver and chicken erythrocytes. Biochemistry 47:
10757–10764
Clarke S (1985) Protein carboxyl methyltransferases: two distinct
classes of enzymes. Annu Rev Biochem 54:479–506
Cook PF (1991) Enzyme mechanism from isotope effects. CRC press,
Boca Raton
Corti A, Curnis F (2011) Isoaspartate-dependent molecular switches
for integrin-ligand recognition. J Cell Sci 124:515–522
Cox MM, Jencks WP (1978) General acid catalysis of the aminolysis
of phenyl acetate by a preassociation mechanism. J Am Chem
Soc 100:5956–5957
Di Fede G, Catania M, Morbin M, Rossi G, Suardi S, Mazzoleni G,
Merlin M, Giovagnoli AR, Prioni S, Erbetta A, Falcone C, Gobbi
M, Colombo L, Bastone A, Beeg M, Manzoni C, Francescucci B,
Lowenson JD, Roher AE, Clarke S (1994) Protein aging extracellular
amyloid formation and intracellular repair. Trends Cardiovasc
Med 4:3–8
MacLaren DC, Kagan RM, Clarke S (1992) Alternative splicing of
the human isoaspartyl protein carboxyl methyltransferase RNA
leads to the generation of a C-terminal -RDEL sequence in
isozyme II. Biochem Biophys Res Commun 185:277–283
Meinwald YC, Stimson ER, Scheraga HA (1986) Deamidation of the
asparaginyl-glycyl sequence. Int J Pept Protein Res 28:79–84
Ni W, Dai S, Karger BL, Zhou ZS (2010) Analysis of isoaspartic acid
by selective proteolysis with Asp-N and electron transfer
dissociation mass spectrometry. Anal Chem 82:7485–7491
Noguchi S (2010) Conformational variation revealed by the crystal
structure of RNase U2A complexed with Ca ion and 20-adenylic
˚
acid at 1.03 A resolution. Protein Pept Lett 17:1559–1561
Nussbaum JM, Schilling S, Cynis H, Silva A, Swanson E, Wangsanut T,
Tayler K, Wiltgen B, Hatami A, Ronicke R, Reymann K, Hutter-
Paier B, Alexandru A, Jagla W, Graubner S, Glabe CG, Demuth HU,
Bloom GS (2012) Prion-like behaviour and tau-dependent cytotox-
icity of pyroglutamylated amyloid-beta. Nature 485:651–655
Park JW, Lee JC, Ha SW, Bang SY, Park EK, Yi SA, Lee MG, Kim
DS, Nam KH, Yoo JH, Kwon SH, Han JW (2012) Requirement
of protein l-isoaspartyl O-methyltransferase for transcriptional
activation of trefoil factor 1 (TFF1) gene by estrogen receptor
alpha. Biochem Biophys Res Commun 420:223–229
Patel K, Borchardt RT (1990) Chemical pathways of peptide
degradation. II. Kinetics of deamidation of an asparaginyl
residue in a model hexapeptide. Pharm Res 7:703–711
Potter SM, Johnson BA, Henschen A, Aswad DW, Guzzetta AW
(1992) The type II isoform of bovine brain protein L-isoaspartyl
methyltransferase has an endoplasmic reticulum retention signal
(…RDEL) at its C-terminus. Biochemistry 31:6339–6347
Robinson AB (1974) Evolution and the distribution of glutaminyl and
asparaginyl residues in proteins. Proc Natl Acad Sci USA
71:885–888
Robinson NE, Robinson AB (2001) Molecular clocks. Proc Natl Acad
Sci USA 98:944–949
Robinson AB, Rudd CJ (1974) Deamidation of glutaminyl and
asparaginyl residues in peptides and proteins. Curr Top Cell
Regul 8:247–295
Robinson AB, McKerrow JH, Cary P (1970) Controlled deamidation
of peptides and proteins: an experimental hazard and a possible
biological timer. Proc Natl Acad Sci USA 66:753–757
Roher AE, Lowenson JD, Clarke S, Wolkow C, Wang R, Cotter RJ,
Reardon IM, Zurcher-Neely HA, Heinrikson RL, Ball MJ (1993)
Structural alterations in the peptide backbone of beta-amyloid
core protein may account for its deposition and stability in
Alzheimer’s disease. J Biol Chem 268:3072–3083
`
Spagnoli A, Cantu L, Del Favero E, Levy E, Salmona M,
Tagliavini F (2009) A recessive mutation in the APP gene with
dominant-negative effect on amyloidogenesis. Science 323:
1473–1477
Ellis KJ, Morrison JF (1982) Buffers of constant ionic strength for
studying pH-dependent processes. Methods Enzymol 87:405–426
Geiger T, Clarke S (1987) Deamidation, isomerization, and racemi-
zation at asparaginyl and aspartyl residues in peptides. Succin-
imide-linked reactions that contribute to protein degradation.
J Biol Chem 262:785–794
George-Nascimento C, Lowenson J, Borissenko M, Calderon M,
Medina-Selby A, Kuo J, Clarke S, Randolph A (1990) Replace-
ment of a labile aspartyl residue increases the stability of human
epidermal growth factor. Biochemistry 29:9584–9591
Haass C, Lemere CA, Capell A, Citron M, Seubert P, Schenk D,
Lannfelt L, Selkoe DJ (1995) The Swedish mutation causes
early-onset Alzheimer’s disease by beta-secretase cleavage
within the secretory pathway. Nat Med 1:1291–1296
He W, Barrow CJ (1999) The A beta 3-pyroglutamyl and 11-pyro-
glutamyl peptides found in senile plaque have greater beta-sheet
forming and aggregation propensities in vitro than full-length A
beta. Biochemistry 38:10871–10877
Jonsson T, Atwal JK, Steinberg S, Snaedal J, Jonsson PV, Bjornsson
S, Stefansson H, Sulem P, Gudbjartsson D, Maloney J, Hoyte K,
Gustafson A, Liu Y, Lu Y, Bhangale T, Graham RR, Huttenl-
ocher J, Bjornsdottir G, Andreassen OA, Jonsson EG, Palotie A,
Behrens TW, Magnusson OT, Kong A, Thorsteinsdottir U, Watts
RJ, Stefansson K (2012) A mutation in APP protects against
Alzheimer’s disease and age-related cognitive decline. Nature
488:96–99
Kummer MP, Hermes M, Delekarte A, Hammerschmidt T, Kumar S,
¨
Terwel D, Walter J, Pape HC, Konig S, Roeber S, Jessen F,
Klockgether T, Korte M, Heneka MT (2011) Nitration of
tyrosine 10 critically enhances amyloid b aggregation and plaque
formation. Neuron 71:833–844
Saido TC, Yamao-Harigaya W, Iwatsubo T, Kawashima S (1996)
Amino- and carboxyl-terminal heterogeneity of beta-amyloid
peptides deposited in human brain. Neurosci Lett 215:173–176
123

1214
B. H.-O. Gu¨ttler et al.
Sargaeva NP, Lin C, O’Connor PB (2011) Differentiating N-terminal
aspartic and isoaspartic acid residues in peptides. Anal Chem
83:6675–6682
Shimizu T, Watanabe A, Ogawara M, Mori H, Shirasawa T (2000)
Isoaspartate formation and neurodegeneration in Alzheimer’s
disease. Arch Biochem Biophys 381:225–234
Satterthwait AC, Jencks WP (1974) The mechanism of the aminolysis
of acetate esters. J Am Chem Soc 96:7018–7031
Schilling S, Hoffmann T, Manhart S, Hoffmann M, Demuth HU
(2004) Glutaminyl cyclases unfold glutamyl cyclase activity
under mild acid conditions. FEBS Lett 563:191–196
Stephenson RC, Clarke S (1989) Succinimide formation from aspartyl
and asparaginyl peptides as a model for the spontaneous
degradation of proteins. J Biol Chem 264:6164–6170
Varshavsky A (1996) The N-end rule: functions, mysteries, uses. Proc
Natl Acad Sci USA 93:12142–12149
Schilling S, Lauber T, Schaupp M, Manhart S, Scheel E, Bohm G,
Demuth HU (2006) On the seeding and oligomerization of pGlu-
amyloid peptides (in vitro). Biochemistry 45:12393–12399
Schilling S, Zeitschel U, Hoffmann T, Heiser U, Francke M, Kehlen
A, Holzer M, Hutter-Paier B, Prokesch M, Windisch M, Jagla W,
Schlenzig D, Lindner C, Rudolph T, Reuter G, Cynis H, Montag
D, Demuth HU, Rossner S (2008) Glutaminyl cyclase inhibition
attenuates pyroglutamate Abeta and Alzheimer’s disease-like
pathology. Nat Med 14:1106–1111
Schlenzig D, Manhart S, Cinar Y, Kleinschmidt M, Hause G,
Willbold D, Funke SA, Schilling S, Demuth HU (2009)
Pyroglutamate formation influences solubility and amyloidoge-
nicity of amyloid peptides. Biochemistry 48:7072–7078
Seifert F, Schulz K, Koch B, Manhart S, Demuth HU, Schilling S
(2009) Glutaminyl cyclases display significant catalytic profi-
ciency for glutamyl substrates. Biochemistry 48:11831–11833
Velazquez P, Cribbs DH, Poulos TL, Tenner AJ (1997) Aspartate
residue 7 in amyloid beta-protein is critical for classical
complement pathway activation: implications for Alzheimer’s
disease pathogenesis. Nat Med 3:77–79
Wagenknecht H-A, Knapp H (2009) Roempp Online. Version 3.5
Weber DJ, McFadden PN (1997) Injury-induced enzymatic methyl-
ation of aging collagen in the extracellular matrix of blood
vessels. J Protein Chem 16:269–281
Young AL, Carter WG, Doyle HA, Mamula MJ, Aswad DW (2001)
Structural integrity of histone H2B in vivo requires the activity
of protein L-isoaspartate O-methyltransferase, a putative protein
repair enzyme. J Biol Chem 276:37161–37165
Young GW, Hoofring SA, Mamula MJ, Doyle HA, Bunick GJ, Hu Y,
Aswad DW (2005) Protein L-isoaspartyl methyltransferase
catalyzes in vivo racemization of aspartate-25 in mammalian
histone H2B. J Biol Chem 280:26094–26098
123