Ultrasonic-based protein quantitation by18O-labeling: Optimization and comparison between different procedures

Article abstract of DOI:10.1002/rcm.4816

Herein we report results regarding the optimization and comparison between different ultrasonic-based procedures for protein quantitation by the direct ¹⁸O-labeling approach. The labeling procedure was evaluated using different proteins, differ

Full text of DOI:10.1002/rcm.4816

Research Article
Received: 27 July 2010
Revised: 6 October 2010
Accepted: 8 October 2010
Published online in Wiley Online Library: 00 Month 2010
Rapid Commun. Mass Spectrom. 2011, 25, 75–87
(wileyonlinelibrary.com) DOI: 10.1002/rcm.4816
18
Ultrasonic-based protein quantitation by O-labeling:
optimization and comparison between different procedures
R. J. Carreira1, M. S. Diniz1 and J. L. Capelo1,2
*
1
´
ˆ
REQUIMTE, Departamento de Quımica, Faculdade de Ciencias e Tecnologia, Universidade Nova de Lisboa, 2829-516
Caparica, Portugal
2University of Vigo, Department of Physical Chemistry, Faculty of Sciences at Ourense Campus, c/As Lagoas s/n, E-32004
Ourense, Spain
Herein we report results regarding the optimization and comparison between different ultrasonic-based procedures
for protein quantitation by the direct 18O-labeling approach. The labeling procedure was evaluated using different
proteins, different ultrasonic devices and different reaction times: from 30 s to 10 min with the ultrasonic probe and
from 30 s to 30 min with the sonoreactor. Variables such as the enzyme-to-protein ratio and protein concentration were
also assessed. The results show that it is possible to accelerate the labeling reaction from 12 h to only 15 min with the
sonoreactor without compromising the labeling efficiency. A larger variation in the double labeling yield was
obtained among the different peptides, but the values for the smaller peptides are similar to the ones achieved with
the classic methodology. These findings were further confirmed by labeling a complex protein mixture from human
plasma. It was also found that the labeling reaction is affected by the sample concentration, even when performed
with the classic overnight procedure. Copyright ß 2010 John Wiley & Sons, Ltd.
The proteome of a living organism is the result of gene
expression but unlike the genome, it is highly dynamic
and influenced by cellular conditions and physiological
states. In order to study the protein components of biological
systems we need to obtain not only information about
the presence or absence of proteins (qualitative information),
but we also need to infer about protein expression level
(quantitative information).[1]
Mass spectrometry (MS) is nowadays an essential tech-
nique in the proteomics field and when coupled with stable
isotopic labeling (SIL) methods, MS can provide important
quantitative information.[2–5] There are several SIL method-
ologies, e.g.: stable isotope labeling by amino acids in cell
culture (SILAC),[6] isotope-coded affinity tags (ICAT),[7] and
isobaric tag for relative and absolute quantitation (iTRAQ),[8]
but the 18O-enzymatic labeling of proteins is one of the most
commonly used methods because it is a relatively cheap
technique, easy to perform and versatile.[9–13]
In the normal 18O-labeling workflow one sample is labeled
in 18O-enriched water while the other is labeled in natural
abundance 16O water. Then, the two samples are mixed and
analyzed by MS. Finally, the relative abundance of each
sample is calculated based on the relative intensities of the
‘light’ and ‘heavy’ labeled peptides provided by the mass
spectrum.[14] The labeling reaction occurs during the
hydrolysis of the peptide bond and, depending on the
enzyme and on the reaction conditions, one or two 18O-atoms
from the H218O-enriched medium are incorporated at the
C-terminal carboxyl group of the peptide.[9] Trypsin can
catalyze the incorporation of two 18O-atoms, resulting in the
ideal mass shift of þ4 Da for the labeled peptide fragment,
which is the minimum mass gap required to avoid naturally
occurring isotopic interferences (e.g. 13C, 15N, 34S) or isotopic
overlapping between labeled and unlabeled species in the
mass spectrum. There are two main approaches for the
18O-isotopic labeling of proteins: (i) the isotopic labeling
occurs during the enzymatic digestion in H218O buffer
medium, the direct labeling procedure; or (ii) the post-
digestion approach, the decoupling procedure, where the
proteins are first digested in H216O buffer medium, dried and
then labeled in H218O in the presence of trypsin.[15] The
post-digestion approach has the advantage of consuming less
H218O, an expensive reagent, and it provides better 18O
conversion at the C-terminus of the peptide, thus increasing
the labeling efficiency. However, the procedures used with
this approach are generally longer, more elaborate and
labor-intensive than the direct labeling protocols.[16] Due to
these disadvantages, and because the direct labeling is done
in one single step, minimizing technical variations and
facilitating on-line approaches for MS protein quantitation,
we chose this strategy for protein labeling in this study.
Despite the advantages of 18O-labeling over other labeling
techniques, there are also some drawbacks that can affect
the labeling efficiency, such as: (i) variable 18O-incorporation,
i.e. single or double 18O-incorporation at the peptide’s
C-terminus; and (ii) back-exchange reaction, i.e. the post-
labeling exchange of 18O from the peptide’s C-terminal
carboxyl group with 16O from medium contamination
with residual H216O. In order to improve the double
18O-incorporation and diminish the effect of the back-
exchange reaction in the labeling efficiency, several reaction
parameters such as (i) the pH of the labeling reaction, (ii) the
H216O/H218O ratio, and (iii) the residual enzymatic activity
need to be optimized and controlled.[12,13,17–19] The time-
* Correspondence to: J. L. Capelo, REQUIMTE, Departamento de
´
ˆ
Quımica, Faculdade de Ciencias e Tecnologia, Universidade
Nova de Lisboa, 2829-516 Caparica, Portugal.
E-mail: jlcapelom@uvigo.es
Rapid Commun. Mass Spectrom. 2011, 25, 75–87
Copyright ß 2010 John Wiley & Sons, Ltd.

R. J. Carreira, M. S. Diniz and J. L. Capelo
consuming labeling reaction, 12 to 48 h, also creates a hurdle
for the application of this methodology to a wider range of
protein quantitation experiments.
a-lactalbumin (100 pmol/mL) were prepared in AmBic
(100 mM) using natural abundance water, and the protein
samples were reduced with DTT (10 mM) during 1 h at 378C,
and alkylated with IAA (50 mM) in the dark at room
temperature during 45 min. Aliquots of 10 mL were diluted to
100 mL with AmBic (100 mM) prepared in natural abundance
water, or in 95% 18O-enriched water. Trypsin (2 mL) was
added to the protein samples to a final concentration of
0.47 pmol/mL and the enzymatic digestion/labeling was
accelerated with: (i) ultrasonic probe operating at 50%
amplitude with a 0.5 mm probe, during 30, 60, 120, 300 s
and 10 min; (ii) sonoreactor (UTR) operating at 50%
amplitude during 30, 60, 120, 300 s, and 10, 15 and 30 min.
The enzymatic reaction was stopped after the addition of 5 mL
of formic acid (50%, v/v). All the experiments were done in
replicates of three (n ¼ 3).
Ultrasonic energy has been used in proteomic method-
ologies to enhance protein enzymatic digestion from over-
night to minutes.[20,21] More recently, we reported the
application of ultrasonic energy to accelerate the 18O-labeling
procedure with promising results. Several proteins were
labeled in only 30 min in an ultrasonic bath (USB) and similar
16O/18O ratios to the classical approach (12 h labeling) were
obtained. However, acceptable 18O1/18O2 ratios were only
obtained for bovine serum albumin (BSA).[22]
In this work, the application of direct ultrasonication with
the ultrasonic probe and indirect ultrasonication with the
sonoreactor (UTR) to enhance the enzymatic 18O-labeling
reaction is compared and reported. The influence of the type
of ultrasonic device, the ultrasonication time, the protein
concentration and the enzyme concentration in the labeling
efficiency and in the 18O-incorporation degree were assessed,
and compared with the results previously obtained with
the classical approach (overnight labeling) and with the
ultrasonic bath.[22]
Enzyme-to-protein ratio effect on the 18O-labeling reaction
The sample treatment before protein digestion/labeling was
performed as described in the previous section. Following
protein reduction and alkylation, aliquots of 10 mL of
BSA (60 mg) were diluted to 100 mL with AmBic (100 mM)
prepared in natural abundance water, or in 95% 18O-enriched
water. After the addition of 2 mL of trypsin the reaction was
performed during 15 min in the sonoreactor (50% amplitude).
Different enzyme-to-protein (E:P) ratios were used for
protein digestion/labeling: (i) 1:120 w/w (trypsin – 0.5 mg);
(ii) 1:80 w/w (trypsin – 0.75 mg); (iii) 1:60 w/w (trypsin –
1.0 mg); (iv) 1:40 w/w (trypsin – 1.5 mg); 1:30 w/w (trypsin –
2.0 mg). Formic acid 50% (v/v) (5 mL) was added to stop the
enzymatic digestion.
EXPERIMENTAL
Apparatus
Protein digestion/labeling was performed in 0.5 mL safe-lock
tubes (Eppendorf, Hamburg, Germany). A minicentrifuge-
vortex model Sky Line (ELMI, Riga, Latvia) and a model
Spectrafuge-mini minicentrifuge (Labnet, Madrid, Spain)
were used during the sample treatment. Milli-Q natural
abundance (H216O) water was obtained from a SimplicityTM
185 model (Millipore, Milan, Italy). A UTR200 sonoreactor
(200 W, 24 kHz) and a UP100H ultrasonic probe (100 W,
30 kHz, 0.5 mm diameter probe tip; Hielscher Ultrasonics,
Teltow, Germany) were used to accelerate enzymatic protein
digestion/labeling.
18O-labeling in low concentration protein samples
Protein reduction and alkylation was performed as described
above. BSA samples of 2.5; 5; 15; 30 and 60 mg were digested/
labeled in 100 mL of AmBic (100 mM) prepared in natural
abundance water, or in 95% 18O-enriched water. The E:P ratio
used throughout these experiments was always 1:40 (w/w)
except for the 2.5 mg samples, which were also digested/
labeled with different E:P ratios: 1:40 w/w (trypsin –
0.0625 mg); 1:20 w/w (trypsin – 0.125 mg); 1:6.7 w/w (trypsin
– 0.375 mg); and 1:3.3 w/w (trypsin – 0.75 mg). The enzymatic
reaction was performed in 15 min with the sonoreactor
(50% amplitude), and stopped after the addition of 5 mL of
formic acid (50% v/v).
Standards and reagents
Bovine serum albumin (BSA; 66 kDa, >97%), a-lactalbumin
(14.4 kDa, ꢀ85%), ovalbumin (45 kDa), human plasma
(lyophilized powder) and trypsin (proteomics grade) used
throughout the experiment were from Sigma (Steinheim,
Germany) as well as the DL-dithiothreitol (DTT, 99%) and
iodoacetamide (IAA) used for protein reduction and
alkylation, respectively. Ammonium bicarbonate buffer
(AmBic, pH 8.5, ꢀ99.5%) and formic acid (FA, ꢁ98%) were
from Fluka (Buchs, Switzerland), and the H218O (95 atom %)
used for protein isotopic labeling was from ISOTECTM
(Miamisburg, OH, USA). a-Cyano-4-hydroxycinnamic acid
(a-CHCA, ꢀ99.0%), acetonitrile (99.9%) and trifluoroacetic
acid (TFA, 99%) were from Fluka, Sigma-Aldrich and Riedel-de
18O-labeling of proteins from human plasma
Lyophilized human plasma was dissolved in 1 mL of
phosphate-buffered saline (PBS) buffer (pH 7.2). Aliquots
of 100 mL were precipitated overnight at ꢂ208C with
5 volumes of cold acetone. The samples were then centrifuged
at 10 000 g during 30 min (48C); the supernatant was
discarded and the pellet resuspended in 50 mL of AmBic
(100 mM). Protein reduction and alkylation was performed
as described above, and then aliquots of 10 mL were diluted
to 100 mL with AmBic (100 mM) prepared in natural
abundance water, or in 95% 18O-enriched water. Trypsin
(1 mg) was added and the enzymatic digestion/labeling
Haen (Seelze, Germany), respectively. The ProteoMassTM
¨
Peptide MALDI-MS calibration kit (MSCAL2) from Sigma
was used as mass calibration standard for MALDI-TOF-MS.
Sample treatment
Protein digestion/labeling
Protein digestion/labeling was performed as previously
described.[22] Briefly, stock solutions of BSA, ovalbumin and
wileyonlinelibrary.com/journal/rcm
Copyright ß 2010 John Wiley & Sons, Ltd.
Rapid Commun. Mass Spectrom. 2011, 25, 75–87

Ultrasonic-based protein quantitation by 18O-labeling
reaction allowed to proceed for 15 min in the sonoreactor
at 50% amplitude, or overnight at 378C. The enzymatic
reaction was stopped after the addition of 5 mL of formic acid
(50%, v/v). All the experiments were done using three
replicates (n ¼ 3).
RESULTS AND DISCUSSION
The 18O-labeling reaction can incorporate one or two
18O-atoms at the terminal carboxylic group of the peptide,
shifting the mass value of the naturally occurring isotopic
distribution by þ2 or þ4 Da. Figure 1 presents two theoretical
cases which can occur when performing protein 18O-labeling
quantitation. Spectrum (a) represents the theoretical result
obtained when the labeling reaction is complete and all the
peptides are double labeled. Here, when mixing the labeled
sample with the unlabeled control sample, there is no isotopic
overlapping in the mass spectrum between the two peptide
forms. Therefore, protein relative quantitation can simply
be done by measuring the relative intensities of the
monoisotopic peaks of each peptide. In case (b) the labeling
reaction is incomplete and a mixture of single and double
labeled peptides is generated, producing an isotopic overlap
between the non-labeled control sample and labeled sample.
This variable 18O-incorporation affects the measuring of the
relative abundances of the peptide and increases the error in
the calculation of the correct 16O/18O peptide ratios.[3,11,13]
Therefore, when performing 18O-labeling it is important to
consider not only the labeling efficiency (18Ototal %), which
measures the percentage of both single and double labeled
peptides, but also the labeling degree (18O2 %) which
measures the percentage of double labeled peptides.
MALDI-TOF-MS analysis
Before MS analysis, samples were mixed in a 1:1 ratio with
the matrix solution of a-CHCA (10 mg/mL) prepared in 50%
acetonitrile/0.1% TFA. Each sample (1 mL) was hand-spotted
onto a MALDI-TOF-MS stainless steel 96-well plate and
allowed to dry. The mass spectra were obtained with a
Voyager DE-PRO Biospectrometry Workstation (Applied
Biosystems, Foster City, USA), equipped with a nitrogen laser
radiating at 337 nm. Measurements were carried out in the
reflector positive ion mode, with an accelerating voltage of
20 kV, 75.1% of grid voltage, 0.002% of guide wire and a delay
time of 100 ns. The monoisotopic peaks of bradykinin,
angiotensin II, P14R and ACTH peptide fragments (m/z
[M þ H]þ: 757.3997, 1046.5423, 1533.8582 and 2465.1989,
respectively) were used for the external calibration of the
mass spectra. A total of 250 laser shots were summed per
spectrum.
18O-labeling with direct ultrasonication
Deconvolution of MALDI mass spectra
The ultrasonic probe provides ultrasonic energy directly into
the reaction media and the ultrasound intensity is at least
1500 times higher than that provided by the ultrasonic
bath.[25] Previous studies reported the use of direct ultra-
sonication to accelerate the 18O-labeling reaction in a
post-digestion approach with similar results to the ones
obtained with the conventional protocols.[26] However,
despite accelerating the enzymatic reactions up to only a
few minutes, this workflow still remains tedious and long
due to the several drying steps required, which increase
sample losses and preclude on-line labeling approaches.
Therefore, to assess the ultrasonic probe effect on peptide
18O-labeling using the direct labeling approach, aliquots of
BSA (60 mg) were digested in the presence of H218O and
The mass spectra deconvolution was performed with the
Data ExplorerTM software (version 4.0) from Applied
Biosystems. This software has an advanced peak filtering
method, the deisotope function, that uses a deisotoping
algorithm to determine the relative abundance of multiple
components with overlapping isotope distributions.[23] The
deisotope function deconvolutes the mass spectra and
reduces the isotopic cluster to a centroided plot composed
of the monoisotopic peaks from the peak list. Additionally,
for comparative purposes in order to test the correct
applicability of this function, the mathematical algorithm
for deconvolution described by Yao and coworkers was also
used in the first steps of this work.[24]
Figure 1. Complete 18O-labeling vs. variable 18O-labeling. (a) Complete 18O-labeling – theoretical MALDI-TOF mass spectrum
of a mixture (1:1) of the unlabeled (927 m/z) and the double labeled (931 m/z) peptide (YLYEIAR)Hþ from BSA. (b) Variable
18O-labeling – theoretical MALDI-TOF mass spectrum of a mixture (1:1) of the unlabeled (927 m/z) and the single (929 m/z) and
double labeled (931 m/z) peptide (YLYEIAR)Hþ from BSA.
Rapid Commun. Mass Spectrom. 2011, 25, 75–87
Copyright ß 2010 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/rcm

R. J. Carreira, M. S. Diniz and J. L. Capelo
trypsin (1 mg) with a 0.5 mm sonotrode. The ultrasonication
time was between 30 s and 10 min.
suggest that a complex relation involving ultrasonic energy
and the first and second 18O-incorporations take place. It
seems that this reaction can be enhanced with ultrasonic
energy of low intensity, such as the one provided by the
ultrasonic bath and the sonoreactor. Yet, a system like the
ultrasonic probe, 30 times more intense than the sonoreactor
and 1500 times more intense than the ultrasonic bath, seems to
produce uncontrolled effects which compromise an effective
18O-incorporation, especially the double 18O-incorporation.
The nature of these effects remains unclear.
Table 1 presents the labeling degree results (18O2 %)
obtained for several BSA peptides after ultrasonication with
the ultrasonic probe. In general, the labeling degree increased
with the ultrasonication time reaching a maximum value
for t ¼ 120 s, and no improvements were achieved when
ultrasonication was performed during 5 and 10 min. In fact,
the ultrasonication of liquid media with the ultrasonic
probe during long periods of time (>2 min) leads to
sample overheating and aerosol formation, diminishing the
efficiency of the ultrasonic treatment.[27] In addition, in a
previous study developed by our group, we demonstrated
that 30 s of ultrasonication with the ultrasonic probe does not
affect enzyme activity, but after 60 s the activity decreases ca.
20% (casein hydrolysis with protease XIV). An ultrasonica-
tion time of 120 s led to the complete inactivation of the
enzyme.[28] The results obtained also show that the double
18O-labeling yield decreased with the increasing mass of the
peptide fragment. The smallest peptide fragment considered
in this study (YLYEIAR)Hþ – 927 m/z was double labeled with
an efficiency of 63.9% (t ¼ 120 s) whilst the larger peptide
(KVPQVSTPTLVEVSR)Hþ – 1639 m/z presented a double
labeling efficiency of only 12.6% for the same ultrasonication
time. Considering that the oxygen exchange rate is dependent
on the peptide size, sequence and type of amino acid,[29] it is
possible that the ultrasonic energy provided by the ultrasonic
probe was not sufficient to enhance the different reaction
rates during the short reaction times tested here, since when
the labeling reaction was performed during 12 h at 378C the
double labeling yield was ca. 76% for all the peptides.[22]
Concerning the isotopic labeling efficiency (Table 2), i.e. the
percentage of peptides labeled with at least one 18O (18Ototal
%), the results show that the best performance was obtained
for t ¼ 120 and t ¼ 300 s, and no significant improvement
was achieved when ultrasonication was performed during
10 min. The percentage of single or double labeled peptides
was higher than 90% for all peptides, except fragment
(LGEYGFQNALIVR)Hþ – 1479 m/z with a labeling efficiency
below 85% for all the ultrasonication times tested. These
results are closer to the values obtained with the USB, with a
labeling efficiency higher than 92% for all peptides, but they
are still below the values obtained by the classic methodology
(overnight labeling), with labeling efficiencies higher than
96%.[22] This is probably due to the fact that the labeling
reaction is a two-step reaction. In the first step the enzyme
forms an ester intermediate with the peptide and during
the hydrolysis of the amide bond the first 18O from the
medium is incorporated at the terminal carboxyl group.
During the second step of the reaction, also known as the
carboxyl exchange reaction, the enzyme forms another ester
intermediate with the peptide terminal carboxyl group and
after a series of esterification and hydrolysis cycles the
peptide will be double labeled.[29] Unfortunately, the rates of
the carboxyl oxygen exchange reaction are much lower than
the peptide bond hydrolysis, which lead to the variable
18O-incorporation when short reaction times are used. It is
also important to stress that the carboxyl exchange reaction is
a reversible reaction; therefore the ultrasonic energy not only
accelerates the 16O exchange from the peptide’s carboxyl
group with 18O from the medium, but also enhances the
reverse reaction. The results presented in Tables 1 and 2 also
18O-labeling with the UTR
The sonoreactor (UTR) can be defined as a high-intensity
ultrasonic bath.[25] Even though the ultrasonic energy
generated by the sonoreactor is 30 times less intense than
the one provided by the ultrasonic probe, several advantages
make this device suitable for proteomic workflows, such as:
(i) sample ultrasonication in sealed vials, which prevents
cross-contamination between samples; (ii) no aerosol
formation, which improves biosafety when working with
hazardous samples from pathogenic bacteria and viruses;
(iii) lower sample volume is needed for ultrasonication;
(iv) high throughput, since many samples can be treated at
once unlike common ultrasonic probes.[21,25]
The efficiency of the sonoreactor was first evaluated using
BSA as the model protein. The results in Table 1 show that the
percentage of double labeled peptides (18O2) increased with
the ultrasonication time. For the time range between 30 and
120 s the percentage of double labeled peptides obtained with
the sonoreactor was similar to the one achieved with the
ultrasonic probe. However, better results were obtained with
the sonoreactor when BSA was labeled during 5 min
and 10 min, especially for the largest peptides. This is in
agreement with the data presented before for the ultrasonic
probe, where it was said that the lack of improvement in the
labeling degree efficiency for larger ultrasonication times
could be related to aerosol formation, sample overheating
and unexpected reactions caused by the high ultrasound
intensity provided by the probe. When ultrasonication is
performed with the sonoreactor the aerosol formation is
insignificant and the temperature of the water bath can be
controlled. In addition the ultrasound intensity provided is 30
times lower than the one obtained with the ultrasonic probe.
Therefore, no sample overheating or spreading through the
walls of the container occurs, and the ultrasonic efficiency is
maintained during the process. Regarding the larger ultra-
sonication times tested, 15 and 30 min, the 18O-labeling
degree obtained with the sonoreactor was higher than when
the ultrasonic bath was used, and for the smallest peptide
fragments, (YLYEIAR)Hþ – 927 m/z; (ALKAWSVAR)Hþ
–
1001 m/z; and (RHPEYAVSVLLR)Hþ – 1439 m/z, the percen-
tages of double 18O-incorporation were higher and close to
the values obtained with the classic overnight protocol (76%).
The efficiency of the 18O-labeling reaction (single and
double labeling) also increased with the ultrasonication
time (Table 2). Like before, the results obtained with the
sonoreactor between 30 s and 5 min were similar to the ones
obtained with the ultrasonic probe, but with 10 min of
ultrasonication the labeling efficiency was higher for the
sonoreactor. As previously referred, this result suggests that
the labeling reaction is more effective when ultrasonic energy
wileyonlinelibrary.com/journal/rcm
Copyright ß 2010 John Wiley & Sons, Ltd.
Rapid Commun. Mass Spectrom. 2011, 25, 75–87

Ultrasonic-based protein quantitation by 18O-labeling
Rapid Commun. Mass Spectrom. 2011, 25, 75–87
Copyright ß 2010 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/rcm

R. J. Carreira, M. S. Diniz and J. L. Capelo
wileyonlinelibrary.com/journal/rcm
Copyright ß 2010 John Wiley & Sons, Ltd.
Rapid Commun. Mass Spectrom. 2011, 25, 75–87

Ultrasonic-based protein quantitation by 18O-labeling
of low intensity is used. This conclusion is confirmed by the
data obtained with the sonoreactor and the ultrasonic bath.
These devices, which provide low intensity ultrasonic energy,
can accelerate the labeling reaction to the same levels as the
overnight process (12 h) in just 15 min.
Additional proteins were 18O-labeled to assess the
efficiency of the UTR technology. Therefore, aliquots of
ovalbumin and a-lactalbumin were labeled in the presence of
H218O and trypsin during 15 and 30 min. The ultrasonication
times were chosen based on the best labeling degree and
efficiency obtained for BSA. As may be seen in Table 3, for the
most intense peptides of ovalbumin, the labeling efficiency
(18Ototal) was between 96 and 99% for both 15 and 30 min of
ultrasonication. In addition, the labeling efficiency obtained
in only 15 min with the sonoreactor was higher than that
obtained during 30 min with the ultrasonic bath, and it was
equal or higher than that obtained with the classic overnight
reaction, which was between 93 and 97%. In terms of the yield
of double labeled peptides (18O2), the results were mostly the
same, regardless of the ultrasonication time applied: for
peptide (VYLPR)Hþ – 647.39 m/z, the double labeling yield
was between 65 and 70% with 15 and 30 min, respectively;
and for the largest peptide (GGLEPINFQTAADQAR)Hþ
–
1687.84 m/z the double labeling yield was 71% with both
reaction times, which is higher than the yield obtained with
the classic method, 66%.
Concerning the labeling efficiency (18Ototal) obtained for
a-lactalbumin, the results show that at least 90% of the
peptides were labeled with one or two 18O-atoms for both
15 and 30 min of ultrasonication, whilst with the ultrasonic
bath a labeling efficiency of 87% was obtained after 30 min.
However, in this case the results were below the classic
overnight procedure with a labeling efficiency of ca. 97%.
As far as the double 18O-incorporation is concerned, the
sonoreactor performed much better than the ultrasonic
bath by promoting the double 18O-incorporation in 30% of
a-lactalbumin’s (CEVFR)Hþ (710.33 m/z) peptide, whereas
the results obtained with the ultrasonic bath for the same
peptide were 6 times lower: 5% of double 18O-incorporation
during 30 min of ultrasonication. However, the sonoreactor
results were still lower than the ones obtained with the classic
overnight methodology for which 71 to 76% of the peptides
were double 18O-labeled.
Influence of E:P ratio on the 18O-labeling reaction
The double 18O-incorporation at the terminal carboxylic
group of the peptide is essential to obtain mass spectra free
from isotopic overlap between unlabeled and labeled species,
thus improving the precision and accuracy of the protein
quantitation method. In order to achieve a complete double
18O-labeling it is important to accelerate the carboxyl
oxygen exchange reaction. This can be done in several
different ways, such as: (i) decreasing the pH of the enzyme-
catalyzed carboxyl oxygen exchange reaction from 8.5, the
optimum pH for trypsin proteolytic activity, to pH between
5 and 6, where trypsin presents the best catalytic activity
regarding the carboxyl oxygen exchange reaction;[15,30] (ii) or
by performing the carboxyl oxygen exchange in aqueous
solutions with organic solvents.[31] Despite providing prom-
ising results, these techniques rely on the post-digestion
labeling approach, where the sample is first digested in
Rapid Commun. Mass Spectrom. 2011, 25, 75–87
Copyright ß 2010 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/rcm

R. J. Carreira, M. S. Diniz and J. L. Capelo
natural abundance water media, then dried and finally
labeled in an appropriate buffer enriched with 18O-water.
Due to the extra drying steps introduced, this method
requires more time, is not suitable for on-line approaches and
the results might suffer from a higher technical variation.
Another way to increase the yield of the double labeling is to
use a higher E:P ratio, although this might be a problem due
to enzyme autolysis, because of the ion suppression effect
caused by enzyme peptides over the protein peptides in the
mass spectra. The use of calcium salts to improve trypsin’s
activity and prevent autolysis has been reported,[32,33] but this
implicates an extra step to remove the salts before MS
analysis to avoid interferences with peptide ionization.
Immobilized trypsin can also be used as an alternative to
overcome the problem of back-exchange and enzyme
autolysis,[16,34] but the cost of this reagent may be discoura-
ging to some laboratories.
To further evaluate the combined effect of ultrasound
and different enzyme-to-protein (E:P) ratios, aliquots of BSA
(60 mg) were labeled during 15 min with the sonoreactor in
the presence of H218O and different amounts of trypsin. The
E:P ratio varied between 1:120 and 1:30 (w/w). As may be
seen in Table 4, the lowest labeling efficiency (18Ototal %) was
obtained when only 0.5 mg (1:120 w/w E:P ratio) of trypsin
was used. In this case the labeling efficiency varied from 80%,
for the two larger peptides (1479 and 1639 m/z), to 93% for
the smallest peptide (927 m/z). When other E:P ratios were
used the labeling efficiencies obtained with each ratio were
similar among them: between 96 and 98% for the smaller
peptides – 927, 1001 and 1439 m/z; and between 93 and 96%
for the largest peptides, 1479 and 1639 m/z. Regarding the
double 18O-incorporation (18O2) yield, the worst performance
was obtained with the lowest amount of trypsin (0.5 mg): 50%
of the (YLYEIAR)Hþ – 927 m/z peptides were double labeled
with 15 min of ultrasonication, while the classical overnight
labeling methodology achieved a double labeling yield of
75% for the same peptide. Focusing on peptide fragments
(YLYEIAR)Hþ – 927 m/z and (LGEYGFQNALIVR)Hþ –
1479 m/z, results show that the double labeling degree
increased with the amount of enzyme, reaching a maximum
when 1:40 (w/w) E:P ratio was used: 75 and 65%, respectively.
For these peptides no improvement in the double incorporation
yield was obtained when the E:P ratio was raised to 1:30 (w/w).
Regarding peptide fragments (RHPEYAVSVLLR)Hþ
–
1439 m/z and (KVPQVSTPTLVEVSR)Hþ –1639 m/z, the label-
ing degree obtained also increased with the amount of
trypsin, but no significant differences were found when the
E:P ratio varied from 1:60 to 1:30 (w/w). In addition, no
significant interference from trypsin autolysis peptides was
found in the mass spectra of the samples labeled with a higher
quantity of trypsin (Fig. 2). Overall, the best results were
obtained when the E:P ratio was 1:40 (w/w), which is in
the range of the recommended E:P ratios by Sigma-Aldrich1
for the in-solution protein digestion: between 1:100 and 1:20
(w/w).[35]
Influence of sample concentration on the 18O-labeling of
proteins
When the sample concentration decreases the probability of
having protein and enzyme molecules to collide with
each other and establish bonds also decreases, which may
compromise protein digestion/18O-labeling. Thus, the ultra-
sound effect on the isotopic labeling reaction of low
concentration protein samples was evaluated: BSA samples
ranging from 2.5 to 60 mg were 18O-labeled during 15 min
with the sonoreactor in the presence of trypsin, using an E:P
ratio of 1:40 (w/w) which was previously found to be the best
Table 4. Effect of the enzyme-to-protein ratio (E:P) on the labeling efficiency (18Ototal %) and labeling degree (18O2 %).
Aliquots of BSA (60 mg) were labeled during 15 min with the sonoreactor (50% amplitude) in the presence of H218O and
trypsin. Different E:P ratios were used: (i) 1:120 w/w (trypsin – 0.5 mg); (ii) 1:80 w/w (trypsin – 0.75 mg); (iii) 1:60 w/w
(trypsin – 1.0 mg); (iv) 1:40 w/w (trypsin – 1.5 mg); (v) 1:30 w/w (trypsin – 2.0 mg). The five most intense mass peaks were
considered: 927.49 m/z – (YLYEIAR)Hþ; 1001.59 m/z – (ALKAWSVAR)Hþ; 1439.81 m/z – (RHPEYAVSVLLR)Hþ; 1479.80 m/z
– (LGEYGFQNALIVR)Hþ; 1639.94 m/z – (KVPQVSTPTLVEVSR)Hþ (n ¼ 3)
E:P ratio [Trypsin (mg)]
18O2 %
[M þ H]þ
(m/z)
1:120 [0.5]
1:80 [0.75]
1:60 [1.0]
1:40 [1.5]
1:30 [2.0]
927.49
50.46 ꢃ 1.86
58.98 ꢃ 6.70
24.19 ꢃ 3.52
23.15 ꢃ 1.69
6.74 ꢃ 2.22
70.63 ꢃ 2.08
76.03 ꢃ 2.13
52.22 ꢃ 0.22
45.44 ꢃ 3.40
19.02 ꢃ 1.45
74.75 ꢃ 0.01
71.72 ꢃ 10.03
69.86 ꢃ 1.97
58.63 ꢃ 5.79
75.43 ꢃ 0.51
72.63 ꢃ 1.28
*
*
1001.59
1439.81
1479.80
1639.94
69.05 ꢃ 1.96
64.51 ꢃ 0.18
40.38 ꢃ 7.54
68.90 ꢃ 1.26
58.61 ꢃ 1.62
42.71 ꢃ 3.02
41.12 ꢃ 2.48
18
O
%
Total
927.49
93.56 ꢃ 0.08
93.54 ꢃ 4.30
90.64 ꢃ 1.20
78.35 ꢃ 7.88
83.79 ꢃ 3.40
97.59 ꢃ 2.49
98.66 ꢃ 0.01
97.12 ꢃ 4.08
97.45 ꢃ 0.64
93.17 ꢃ 1.29
93.25 ꢃ 0.63
95.90 ꢃ 2.52
97.39 ꢃ 0.17
*
*
1001.59
1439.81
1479.80
1639.94
100
95.93 ꢃ 1.43
94.31 ꢃ 0.51
92.51 ꢃ 0.38
97.61 ꢃ 1.32
96.78 ꢃ 0.85
95.53 ꢃ 1.22
95.63 ꢃ 1.02
92.38 ꢃ 0.83
92.23 ꢃ 2.32
*
Peptide not present in the spectra.
wileyonlinelibrary.com/journal/rcm
Copyright ß 2010 John Wiley & Sons, Ltd.
Rapid Commun. Mass Spectrom. 2011, 25, 75–87

Ultrasonic-based protein quantitation by 18O-labeling
Figure 2. Effect of the enzyme-to-protein (E:P) ratio on the MALDI-TOF mass spectra obtained after protein 18O-labeling with
different amounts of trypsin: (a) 1:120 w/w (trypsin – 0.5 mg); (b) 1:80 w/w (trypsin – 0.75 mg); (c) 1:60 w/w (trypsin – 1.0 mg);
(d) 1:40 w/w (trypsin – 1.5 mg); and (e) 1:30 w/w (trypsin – 2.0 mg).
Table 5. Effect of the sample concentration in the labeling efficiency (18Ototal %) and labeling degree (18O2 %). Aliquots of BSA:
(i) 2.5 mg; (ii) 5mg; (iii) 15 mg; (iv) 30 mg; and (v) 60 mg were labeled during 15 min with the sonoreactor (50% amplitude) in
the presence of H218O and trypsin. A constant enzyme-to-protein ratio was used in this experiment: 1:40 w/w. The five
most intense mass peaks were considered: 927.49m/z – (YLYEIAR)Hþ; 1001.59 m/z – (ALKAWSVAR)Hþ; 1439.81 m/z –
(RHPEYAVSVLLR)Hþ; 1479.80 m/z – (LGEYGFQNALIVR)Hþ; 1639.94 m/z – (KVPQVSTPTLVEVSR)Hþ (n¼ 3)
BSA (mg)
18O2 %
[M þ H]þ
(m/z)
2.5
5
15
30
60
927.49
25.23 ꢃ 3.57
25.60 ꢃ 2.15
13.00 ꢃ 0.44
32.55 ꢃ 0.88
27.06 ꢃ 3.69
8.26 ꢃ 2.54
61.48 ꢃ 2.79
67.98 ꢃ 0.56
38.89 ꢃ 0.02
34.72 ꢃ 2.03
72.69 ꢃ 1.67
86.48 ꢃ 6.60
54.19 ꢃ 4.51
48.67 ꢃ 2.55
18.17 ꢃ 10.28
75.43 ꢃ 0.51
*
1001.59
1439.81
1479.80
1639.94
69.05 ꢃ 1.96
64.51 ꢃ 0.18
40.38 ꢃ 7.54
15.35 ꢃ 12.92
8.80 ꢃ 1.22
*
*
5.77 ꢃ 1.16
18
O
%
Total
927.49
94.15 ꢃ 0.11
89.29 ꢃ 3.45
90.05 ꢃ 7.44
92.39 ꢃ 0.03
91.48 ꢃ 0.48
92.29 ꢃ 1.80
96.77 ꢃ 0.55
97.51 ꢃ 1.07
92.31 ꢃ 0.08
87.50 ꢃ 2.63
90.74 ꢃ 2.52
98.35 ꢃ 0.35
95.90 ꢃ 2.52
*
1001.59
1439.81
1479.80
1639.94
100.00
95.25 ꢃ 0.65
95.48 ꢃ 1.10
91.89 ꢃ 0.80
97.61 ꢃ 1.32
96.78 ꢃ 0.85
95.53 ꢃ 1.22
86.04 ꢃ 4.48
83.17 ꢃ 1.77
*
*
*
Peptide not present in the spectra.
Rapid Commun. Mass Spectrom. 2011, 25, 75–87
Copyright ß 2010 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/rcm

R. J. Carreira, M. S. Diniz and J. L. Capelo
working E:P ratio. The results in Table 5 show that the
labeling efficiency (18Ototal) and the yield of double labeled
peptides (18O2) were higher when the protein concentration
used was 0.6 mg/mL (60 mg of BSA). Regarding the smallest
BSA peptide, (YLYEIAR)Hþ – 927 m/z, the variation in the
labeling efficiency was between 94 and 96% for the lowest
(0.025 mg/mL) and highest (0.6 mg/mL) protein concentration
samples, respectively. Yet, the variation obtained in the
percentage of double labeled peptides between the two
protein samples was much higher: 25% of the peptides were
double labeled in the 0.025 mg/mL BSA samples, in contrast
with the 75% yield obtained when 0.6 mg/mL of BSA was
used. The same pattern was observed for the other peptides:
the labeling efficiency and the labeling degree increased with
the increasing sample concentration.
These results suggest that the ultrasonic energy provided
by the sonoreactor is suitable for the enhancement of the
peptide bond hydrolysis, but not for the acceleration of the
carboxyl oxygen exchange reaction in samples of low protein
concentration. Thus, in order to achieve a better double
labeling yield at the low concentration range using only
15 min of ultrasonication, we increased the E:P ratios
from 1:40 w/w (trypsin – 0.0625 mg) to 1:3.33 w/w (trypsin
– 0.75 mg) in the isotopic labeling of 2.5 mg of BSA. The
overnight labeling reaction was also performed for com-
parative purposes. The results in Fig. 3 show that the
percentage of double 18O-incorporation (18O2) obtained with the
sonoreactor (15 min) increased with the E:P ratio. Yet, a labeling
yield higher than 70% was only achieved for the smallest
peptides, 927 m/z and 1001 m/z, when the E:P ratio was 1:3.3
(w/w). Interestingly, the results achieved with the 12 h labeling
method also presented some variation, especially for the
larger peptides. Considering peptides (LGEYGFQNALIVR)Hþ
– 1479 m/z and (KVPQVSTPTLVEVSR)Hþ –1639 m/z, the
labeling efficiency (18Ototal %) was between 92 and 98%,
but the double labeling yield (18O2 %) was only superior to
70% for the E:P ratios higher than 1:6.7 (w/w). It is important
to note that these E:P ratios are much higher than the
recommended ones.[35] Therefore, when working with
samples of low protein concentration, the peptides chosen
for protein quantitation are of special importance, as well
as the total time for labeling. In order to achieve the maximum
double 18O-incorporation yield the reaction time must
probably be increased beyond the 12 h, if no ultrasonication
is used, or smaller peptides should be chosen for protein
quantitation. Protein concentration strategies, like protein
precipitation or ultrafiltration methods, might also be
adopted in complex protein samples in order to increase
sample concentration.
Ultrasound-based 18O-labeling of proteins from human
plasma
The labeling procedure reported in this manuscript was
further tested in a complex protein sample from human
plasma. As we were only interested in studying the labeling
efficiency, protein identification was not performed.
Thus, after precipitation with cold acetone, protein aliquots
of 10 mL in ammonium bicarbonate (100 mM, pH 7.5–8.5)
were reduced, alkylated and finally labeled with trypsin
in 18O- or 16O-enriched buffer by two different methods:
(i) overnight (378C) and (ii) in the sonoreactor during 15 min
(50% amplitude), which was previously found to be the best
ultrasonic enhanced 18O-labeling method. Regarding the
Figure 3. Effect of the sample concentration on the labeling efficiency (18Ototal %) and labeling degree (18O2 %). Aliquots of BSA
(2.5 mg) were labeled with increasing amounts of trypsin during 12 h (overnight) at 378C, and 15 min with the sonoreactor (50%
amplitude): (A) overnight labeling with 0.0625 mg trypsin; (B) sonoreactor labeling with 0.0625 mg trypsin; (C) overnight labeling
with 0.125 mg trypsin; (D) sonoreactor labeling with 0.125 mg trypsin; (E) overnight labeling with 0.375 mg trypsin;
(F) sonoreactor labeling with 0.375 mg trypsin; (G) overnight labeling with 0.75 mg trypsin; and (H) sonoreactor labeling with
0.75 mg trypsin (n ¼ 3).
wileyonlinelibrary.com/journal/rcm
Copyright ß 2010 John Wiley & Sons, Ltd.
Rapid Commun. Mass Spectrom. 2011, 25, 75–87

Ultrasonic-based protein quantitation by 18O-labeling
Figure 4. 18O-labeling of complex protein samples from human plasma. Spectra (a) and (b) correspond to the overnight protein
digestion at 378C with trypsin in 16O- and 18O-enriched buffer, respectively. Spectra (c) and (d) correspond to protein digestion
with the sonoreactor (15 min; 50% amplitude) in 16O- and 18O-enriched buffer (for details, see section 18O-labeling of proteins from
human plasma).
number of peptides obtained with the different procedures,
181 peptides were labeled with the overnight method,
177 peptides were labeled with the sonoreactor and, from
these peptides, 122 were common to both methods. The
remaining peptides, which were characteristic to each
method, had a relative intensity below 15%. Furthermore,
as can be seen in Fig. 4, the background noise and baseline in
the mass spectra obtained with the different labeling methods
were similar.
Regarding the labeling efficiency for the most intense
peptides, the results obtained were similar between the two
methodologies tested and showed that all the peptides were
labeled with at least one 18O-atom in a percentage higher than
90% when the sonoreactor methodology was used (Fig. 5(a)).
These results are also very close to the best results obtained
previously for the standard proteins. However, as far as the
labeling degree is concerned (Fig. 5(b)), the results presented
a larger variation between different peptides, as obtained
for a-lactalbumin when the accelerated procedure was used.
In fact, only the peptide corresponding to 1623 m/z was
double labeled with a similar percentage to the overnight
procedure: ca. 88%. Peptides corresponding to 927, 960 and
1467 m/z present double labeling percentages higher than
50%, but lower than the double labeling yield of 85%
obtained with the overnight methodology. This is probably
related to the presence of multiple proteins with different
characteristics, some of them more efficiently digested with
trypsin than others. It must be also noted that peptides
corresponding to 1160, 1226 and 1342 m/z were present in the
mass spectra of the ultrasonicated samples with a lower
relative intensity when compared to the mass spectra
corresponding to the overnight labeled samples. Actually,
if we exclude these peptides, it is possible to confirm the trend
observed for the standard protein samples: the higher
the peptide mass, the lower the percentage of double
18O-incorporation.
Figure 5. 18O-labeling of complex protein samples from
human plasma. Comparison between: (a) 18O-labeling effi-
ciency (18Ototal %) and (b) 18O-labeling degree (18O2 %)
obtained with the overnight (12 h; 378C) and the sonoreactor
(15 min; 50% amplitude) methodologies (n ¼ 3).
Rapid Commun. Mass Spectrom. 2011, 25, 75–87
Copyright ß 2010 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/rcm

R. J. Carreira, M. S. Diniz and J. L. Capelo
CONCLUSIONS
Acknowledgements
Ricardo J. Carreira acknowledges the doctoral grant SRFH/
BD/28563/2006 from FCT (Science and Technological
Foundation) of Portugal. The University of Vigo is
acknowledged for financial support under projects INOU/
UVIGO/K919/2009 and INOU/UVIGO/K914/2009. Xunta
de Galicia, Spain, is acknowledged by the Isidro Parga Pondal
Program (J. L. Capelo) and for financial support given
The results obtained show that the ultrasonic probe is
capable of accelerating the labeling reaction from 12 h,
the classic overnight methodology, to only 120 s without
compromising the labeling efficiency. Yet, the labeling
degree, i.e. the percentage of double 18O-labeled peptides,
was lower than that obtained with the classic methodology,
especially for larger peptides. It was also found that the use
of an ultrasonic probe is not recommended for the
acceleration of the labeling reaction when the ultrasonica-
tion time is higher than 120 s, at least with the conditions
reported here, because the aerosol formation, sample
overheating and uncontrolled secondary reactions, that
occur during ultrasonication at high intensities, compro-
mise the double 18O-incorporation at the carboxyl group of
the peptide.
´
under project 09CSA043383PR. Mario S. Diniz acknowledges
ˆ
program Ciencia 2008 from FCT/MCTES.
REFERENCES
[1] M. J. MacCoss, D. L. Matthews. Anal. Chem. 2005, 77, 295A.
[2] A. P. Deleenheer, L. M. Thienpont. Mass Spectrom. Rev. 1992,
11, 249.
[3] S. E. Ong, M. Mann. Nat. Chem. Biol. 2005, 1, 252.
[4] S. Sechi, Y. Oda. Curr. Opin. Chem. Biol. 2003, 7, 70.
[5] M. Bantscheff, M. Schirle, G. Sweetman, J. Rick, B. Kuster.
Anal. Bioanal. Chem. 2007, 389, 1017.
[6] S. E. Ong, B. Blagoev, I. Kratchmarova, D. B. Kristensen, H.
Steen, A. Pandey, M. Mann. Mol. Cell. Proteomics 2002, 1, 376.
[7] S. P. Gygi, B. Rist, S. A. Gerber, F. Turecek, M. H. Gelb, R.
Aebersold. Nat. Biotechnol. 1999, 17, 994.
[8] P. L. Ross, Y. L. N. Huang, J. N. Marchese, B. Williamson, K.
Parker, S. Hattan, N. Khainovski, S. Pillai, S. Dey, S. Daniels,
S. Purkayastha, P. Juhasz, S. Martin, M. Bartlet-Jones, F. He,
A. Jacobson, D. J. Pappin. Mol. Cell. Proteomics 2004, 3, 1154.
[9] M. Schnolzer, P. Jedrzejewski, W. D. Lehmann. Electropho-
resis 1996, 17, 945.
[10] X. D. Yao, A. Freas, J. Ramirez, P. A. Demirev, C. Fenselau.
Anal. Chem. 2001, 73, 2836.
[11] M. Miyagi, K. C. S. Rao. Mass Spectrom. Rev. 2007, 26, 121.
[12] C. Fenselau, X. D. Yao. J. Proteome Res. 2009, 8, 2140.
[13] X. Ye, B. Luke, T. Andresson, J. Blonder. Brief. Funct. Geno-
mics Proteomics 2009, 8, 136.
[14] A. J. Patwardhan, E. F. Strittmatter, G. C. David, R. D. Smith,
M. G. Pallavicini. Proteomics 2006, 6, 2903.
[15] L. Zang, D. P. Toy, W. S. Hancock, D. C. Sgroi, B. L. Karger. J.
Proteome Res. 2004, 3, 604.
[16] J. L. Capelo, R. J. Carreira, L. Fernandes, C. Lodeiro, H. M.
Santos, J. Simal-Gandara. Talanta 2010, 80, 1476.
[17] I. I. Stewart, T. Thomson, D. Figeys. Rapid Commun. Mass
Spectrom. 2001, 15, 2456.
[18] H. F. Storms, R. van der Heijden, U. R. Tjaden, J. van der
Greef. Rapid Commun. Mass Spectrom. 2006, 20, 3491.
[19] P. M. Angel, R. Orlando. Anal. Biochem. 2006, 359, 26.
[20] D. Lopez-Ferrer, J. L. Capelo, J. Vazquez. J. Proteome Res.
2005, 4, 1569.
[21] R. Rial-Otero, R. J. Carreira, F. M. Cordeiro, A. J. Moro, L.
Fernandes, I. Moura, J. L. Capelo. J. Proteome Res. 2007, 6,
909.
[22] R. J. Carreira, R. Rial-Otero, D. Lopez-Ferrer, C. Lodeiro, J. L.
Capelo. Talanta 2008, 76, 400.
[23] Data Explorer Software User’s Guide, version 4.0, section 3,
Applied Biosystems.
[24] X. D. Yao, A. Freas, J. Ramirez, P. A. Demirev, C. Fenselau.
Anal. Chem. 2004, 76, 2675.
[25] H. M. Santos, J. L. Capelo. Talanta 2007, 73, 795.
[26] D. Lopez-Ferrer, T. H. Heibeck, K. Petritis, K. K. Hixson, W.
Qian, M. E. Monroe, A. Mayampurath, R. J. Moore, M. E.
Belov, D. G. Camp, R. D. Smith. J. Proteome Res. 2008, 7, 3860.
[27] R. J. Carreira, F. M. Cordeiro, A. J. Moro, M. G. Rivas, R.
Rial-Otero, E. M. Gaspar, I. Moura, J. L. Capelo. J. Chroma-
togr. A 2007, 1153, 291.
Regarding the sonoreactor, the results obtained from
30 to 120 s were similar to the ones obtained with the
direct ultrasonication method, but in contrast to the
ultrasonic probe, as the ultrasonication time is increased,
higher labeling efficiencies and higher double labeling
yields are obtained. Furthermore, the sonoreactor techno-
logy has some advantages over the ultrasonic probe: (i) it
provides indirect and less intense ultrasonic energy,
preventing aerosol formation; (ii) no sample overheating
occurs, because the temperature of the water bath where
ultrasonication takes place can be controlled; and last
but not least (iii), the ultrasonication is performed in
sealed vials, preventing sample contamination. The results
achieved for the labeling degree (18O2 %) in just 15 min
of ultrasonication were similar to the ones obtained
previously with the overnight methodology. This was
further confirmed with the labeling results obtained for
ovalbumin.
When the ultrasonication was performed with different
enzyme-to-protein (E:P) ratios the results showed that the
labeling efficiency and the labeling degree were best with an
E:P ratio of 1:40 (w/w). However, for low concentration
protein samples, higher E:P ratios were required in order to
achieve an acceptable double labeling yield, even when the
classic methodology was performed.
Our results demonstrate that the isotopic labeling
reaction can be performed in simple protein samples in
only 15 min in a direct labeling approach using indirect
ultrasonication provided by the sonoreactor. No inter-
mediate drying steps are required in this workflow,
which facilitate on-line approaches for protein quantitation.
In addition, the sonoreactor has a higher sample through-
put than the ultrasonic probe, which minimizes the
sample treatment time and simplifies the overall workflow.
However, when applied to a complex protein sample
like human plasma, this technology was not capable
of promoting efficient double 18O-incorporation, thus
compromising protein quantitation. Therefore, in the
presence of this type of samples, two approaches can be
used: (i) the decoupled labeling procedure in which
peptides are double labeled in percentages higher than
95%;[36] or (ii) mathematical algorithms that measure the
effective 18O-incorporation rate due to variable enzyme
substrate specificity during the labeling reaction and correct
for the 18O-abundance.[37,38]
wileyonlinelibrary.com/journal/rcm
Copyright ß 2010 John Wiley & Sons, Ltd.
Rapid Commun. Mass Spectrom. 2011, 25, 75–87

Ultrasonic-based protein quantitation by 18O-labeling
[28] G. Vale, S. Pereira, A. Mota, L. Fonseca, J. L. Capelo. Talanta
2007, 74, 198.
[34] J. R. Sevinsky, K. J. Brown, B. J. Cargile, J. L. Bundy, J. L.
Stephenson. Anal. Chem. 2007, 79, 2158.
[29] S. Julka, F. Regnier. J. Proteome Res. 2004, 3, 350.
[30] A. Staes, H. Demol, J. Van Damme, L. Martens, J. Vande-
kerckhove, K. Gevaert. J. Proteome Res. 2004, 3, 786.
[31] J. Blonder, M. L. Hale, K. C. Chan, L. R. Yu, D. A. Lucas, T. P.
Conrads, M. Zhou, M. R. Popoff, H. J. Issaq, B. G. Stiles, T. D.
Veenstra. J. Proteome Res. 2005, 4: 523.
[35] Technical Bulletin – Product Code T6567, Sigma-Aldrich.
[36] R. J. Carreira, C. Lodeiro, M. S. Diniz, I. Moura, J. L. Capelo.
Proteomics 2009, 9, 4974.
[37] J. E. Eckel-Passow, A. L. Oberg, T. M. Therneau, C. J. Mason,
D. W. Mahoney, K. L. Johnson, J. E. Olson, H. R. Bergen.
Bioinformatics 2006, 22, 2739.
[32] T. Sipos, J. R. Merkel. Biochemistry 1970, 9, 2766.
[33] E. Papaleo, P. Fantucci, L. De Gioia. J. Chem. Theory Com-
putation 2005, 1, 1286.
[38] C. J. Mason, T. M. Therneau, J. E. Eckel-Passow, K. L. Johnson,
A. L. Oberg, J. E. Olson, K. S. Nair, D. C. Muddiman, H. R.
Bergen. Mol. Cell. Proteomics 2007, 6: 305.
Rapid Commun. Mass Spectrom. 2011, 25, 75–87
Copyright ß 2010 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/rcm