Immunoprecipitation on magnetic beads and liquid chromatography-tandem mass spectrometry for carbonic anhydrase II quantification in human serum

Article abstract of DOI:10.1016/j.ab.2010.01.039

In this study, a magnetic bead-based platform amenable to high-throughput protein carbonic anhydrase II (CA II) capture is presented. The key steps in this approach involved immunoaffinity purification of the target protein from serum followed by on-bead digestion with trypsin to release a surrogate peptide. This tryptic peptide was quantified by liquid chromatography-electrospray ionization-tandem mass spectrometry (LC-ESI-MS/MS) operating in multiple reaction monitoring acquisition mode. Using a synthetic peptide standard and a structural analogue free-labeled internal standard, the resulting concentration was stoichiometrically converted to CA II serum concentration. The analytical steps, such as preparation of immunobeads, protein capture, proteolysis, and calibration, were optimized. The method was validated in terms of recovery (77%), reproducibility (relative standard deviation [RSD]<12%), and method detection limit (0.5pmolml^-1). The developed method was applied to determining the CA II in eight healthy subjects, and the concentration measured was 27.3pmolml^-1 (RSD=65%).

Full text of DOI:10.1016/j.ab.2010.01.039

Analytical Biochemistry 400 (2010) 195–202
Contents lists available at ScienceDirect
Analytical Biochemistry
journal homepage: www.elsevier.com/locate/yabio
Immunoprecipitation on magnetic beads and liquid chromatography–tandem
mass spectrometry for carbonic anhydrase II quantification in human serum
*
Luciano Callipo, Giuseppe Caruso, Patrizia Foglia, Riccardo Gubbiotti, Roberto Samperi, Aldo Laganà
Department of Chemistry, Sapienza University of Rome, 00185 Roma, Italy
a r t i c l e i n f o
a b s t r a c t
Article history:
In this study, a magnetic bead-based platform amenable to high-throughput protein carbonic anhydrase
II (CA II) capture is presented. The key steps in this approach involved immunoaffinity purification of the
target protein from serum followed by on-bead digestion with trypsin to release a surrogate peptide. This
tryptic peptide was quantified by liquid chromatography–electrospray ionization–tandem mass spec-
trometry (LC–ESI–MS/MS) operating in multiple reaction monitoring acquisition mode. Using a synthetic
peptide standard and a structural analogue free-labeled internal standard, the resulting concentration
was stoichiometrically converted to CA II serum concentration. The analytical steps, such as preparation
of immunobeads, protein capture, proteolysis, and calibration, were optimized. The method was vali-
dated in terms of recovery (77%), reproducibility (relative standard deviation [RSD] < 12%), and method
detection limit (0.5 pmol mlꢀ1). The developed method was applied to determining the CA II in eight
healthy subjects, and the concentration measured was 27.3 pmol mlꢀ1 (RSD = 65%).
Received 4 November 2009
Received in revised form 14 January 2010
Accepted 28 January 2010
Available online 1 February 2010
Keywords:
Biomarkers
Serum proteins
Immunobeads
Tandem mass spectrometry
Carbonic anhydrase
Ó 2010 Elsevier Inc. All rights reserved.
Most of the studies dealing with biomarker discovery on blood
are focused on the cell leakage products potentially present in that
matrix; indeed, increased serum levels of various proteins are rou-
tinely used for diagnostic purposes [1–4]. The analysis of serum is
challenging because of the large dynamic range of protein concen-
trations, spanning more than 10 orders of magnitude [1,5]. In par-
ticular, the presence of high-abundance proteins such as albumin
and immunoglobulins, whose concentrations represent 60–90% of
the total serum protein content, enhances the difficulty of detect-
ing the low-abundance proteins of interest [5]. Recently, a multisite
assessment of the precision and reproducibility of multiple reac-
tion monitoring (MRM)1-based measurements of target proteins in
plasma demonstrated that this platform may be very reproducible
for proteins present at moderate to high abundance (>2 mg/ml) in
nondepleted, nonfractionated plasma [6].
teins, and it is often achieved by high-abundance protein depletion
[7]. A common approach is the affinity removal method using anti-
body-based resins (both monoclonal and polyclonal) or affinity
dye-based resins [7–9]. There are several affinity removal columns
or kits commercially available for depleting up to 20 major
abundant serum proteins [9,10]. Although this strategy is highly
specific, even after depletion the remaining proteins are still suffi-
ciently abundant to hamper the low-abundance protein determi-
nation. Moreover, some authors have recently pointed out the
risk of losing low-abundance proteins and biomarkers when using
affinity-based depletion due to the association of the targeted pro-
teins with the abundant ones [5,9,11,12], but different results have
also been published [12–15].
Another method of reducing serum sample complexity is the fil-
tration through molecular weight cutoff membranes [9,11] for
removing the high-molecular-weight proteins, including albumin.
Solvent precipitation has also been employed, in particular for
albumin depletion [9]; the method is nonspecific but presents
the advantage of being rapid and cheap compared with the immu-
noaffinity-based methods. For isolation of the fraction containing
the target proteins, liquid chromatographic and electrophoretic
fractionation techniques have been used [9].
On the contrary, the reduction of sample complexity is an
essential first step in the analysis of the serum low-abundance pro-
* Corresponding author. Fax: +39 06 490631.
E-mail address: aldo.lagana@uniroma1.it (A. Laganà).
Abbreviations used: MRM, multiple reaction monitoring; MS, mass spectrometry;
1
LC, liquid chromatography; CA II, carbonic anhydrase II; ESI, electrospray ionization;
MS/MS, tandem mass spectrometry; IS, internal standard; DMP, dimethyl pimelim-
idate dihydrochloride; PBS, phosphate-buffered saline; TFA, trifluoroacetic acid; DTT,
1,4-dithiothreitol; IAA, iodoacetamide; RT, room temperature; SPE, solid phase
extraction; TISP, TurboIonSpray; MARS, Multiple Affinity Removal System; RSD,
relative standard deviation; IgG, immunoglobulin G; MDL, method detection limit;
MQL, method quantification limit; RP–HPLC, reverse phase high-performance liquid
chromatography; LOD, limit of detection; MIL, method identification limit; MQL,
method quantification limit; S/N, signal/noise.
A promising approach is the employment of affinity reagents for
specific enrichment or isolation of target proteins [5,16]. Recently,
increasing attention has been given to the development and applica-
tion of separation techniques employing small magnetic particles.
Magnetic carriers bearing an immobilized affinity or hydrophobic
ligand or ion exchange groups, or magnetic biopolymer particles
0003-2697/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved.
doi:10.1016/j.ab.2010.01.039

196
CA II quantification in human serum / L. Callipo et al. / Anal. Biochem. 400 (2010) 195–202
having affinity to the isolated structure, are mixed with the sample
containing the target compound [17]. Following an incubation per-
iod in which the target compound binds to the magnetic particles,
the whole magnetic complex is easily and rapidly removed from
the sample using an appropriate magnetic separator. The isolated
target compound can then be eluted and used for downstream appli-
cations and detection methods.
Several types of mass spectrometry (MS) techniques in conjunc-
tion with various liquid chromatography (LC) separation methods
are adopted for proteomic measurements. In particular, the MRM
acquisition mode is the most employed one in targeted proteomics
[1,5,18–20]. For absolute quantitative analysis, labeling and label-
free strategies [21] are currently carried out, both aiming to corre-
late the mass spectrometric signal of proteotypic peptide with the
relative or absolute protein quantity directly [22].
Carbonic anhydrase II (CA II), a single polypeptide chain of
29 kDa molecular weight, is present in the cytosol of most tissues
even if the highest concentration is found in erythrocytes. It cata-
lyzes the hydration of CO2 and the hydrolysis of esters, and its defi-
ciency has been associated with pathological consequences such as
mental retardation and cerebral calcification, osteoporosis, and re-
nal tubular acidosis [23,24] as well as Down syndrome [25,26] and
Alzheimer’s disease [26]. Although serum is not the natural site of
CA II, in consequence of cell release, this protein could be present
at low concentration in serum. Furthermore, because CA II is in-
volved in serious diseases, it represents an interesting model to
investigate the separation, identification, and quantification of pro-
teins present in low concentration in complex matrices.
In a recent study by our laboratory [27], CA II quantification in
serum was performed by automated LC chip technology coupled
with an electrospray ionization (ESI) source and a triple quadru-
pole mass spectrometer, operating in MRM acquisition mode, after
purification by reversed phase LC, enzymatic digestion, and surro-
gate peptide selection. The highly sensitive nanoelectrospray chip
able from Carlo Erba Reagents (Milan, Italy) and were used without
any further purification. Ultrapure water was produced from dis-
tilled water by a Milli-Q system (Millipore, Billerica, MA, USA).
Modified porcine trypsin, sequencing grade, was commercialized
by Promega (Madison, WI, USA).
Peptide stock solutions were prepared in 0.1% TFA at 1 g Lꢀ1 and
stored at ꢀ20 °C. The human CA II standard was reconstituted with
0.1% TFA at 1 g Lꢀ1 and stored at ꢀ80 °C. Its actual title was tested
by the Bradford assay and was found to be 85 5%.
Samples
Serum samples were obtained from the Department of Experi-
mental Medicine at Sapienza University of Rome (Italy) by veni-
puncture of healthy volunteers (20–40 years of age), with
collection done in a BD P100 Blood Collection System (Becton Dick-
inson, Franklin Lakes, NJ, USA) with K2EDTA anticoagulant and pro-
tease inhibitors mix (GE Healthcare). After clot formation, the
sample was centrifuged at 1000g for 15 min. The serum was re-
moved, and aliquots were stored at ꢀ80 °C. All serum samples
were checked to verify the absence of hemolysis.
Immobilization of antibody on magnetic beads
The antibody solution (400
into an Eppendorf microcentrifuge tube containing 200
l
l of 1
l
g
l ꢀ1
l
anti-CA II) was added
l (6 mg)
l
of prewashed Dynabeads Protein G and incubated by gentle mixing
at room temperature (RT) for 2 h to allow time to attach onto the
surface of the beads. Following incubation, residual unbound anti-
body was removed by washing two times with 0.5 ml of citrate
phosphate buffer (pH 5.0) with 0.05% Tween 20. Approximately
10 lg of antibody was bound per milligram of beads. For cross-
linking of antibody to beads, the Dynabeads Protein G with immo-
bilized polyclonal antibody were washed with 1 ml of 0.2 mol Lꢀ1
triethanolamine (pH 8.2) and resuspended in 1 ml of freshly pre-
pared cross-linking buffer (20 mmol Lꢀ1 DMP in 0.2 mol Lꢀ1 trieth-
anolamine, pH 8.2). The mixture was incubated under gentle
rotation for 30 min at RT and placed on a magnet, and then the
supernatant was discarded. To stop the reaction, the beads were
resuspended in 1 ml of 50 mmol Lꢀ1 Tris (pH 7.5) and incubated
for 15 min, and then the beads were rinsed three times with 1 ml
of 100 mmol Lꢀ1 PBS with 0.05% Tween 20 before use.
technology was necessary because only 1 ll of serum could be sub-
mitted to the LC isolation step without column overloading.
In the current study, a Protein G magnetic bead-based antibody
platform amenable to high throughput is proposed for the selective
enrichment of CA II in human serum, followed by its accurate and
reliable quantification by a label-free procedure performed on a
less technologically sophisticated LC–ESI–tandem mass spectrom-
etry (MS/MS) system. Magnetic beads bound to polyclonal anti-CA
II antibodies as affinity probes to isolate the specific target protein,
and a conventional LC–ESI–MS/MS apparatus amenable to small
molecule quantitation, were used for the determination of the CA
II proteotypic peptide produced through proteolysis.
CA II capture and digestion
To capture CA II, the Dynabeads Protein G with immobilized
antibody were incubated at 4 °C for 1 h, under gentle mixing, with
20 ll of serum diluted with 100 mM PBS (1:49, v/v), and then the
Materials and methods
beads were recovered from the sample and washed three times
with 1 ml of 100 mmol Lꢀ1 PBS. Immunocomplex was resuspended
Materials
in 40
bicarbonate and 2
for 1 h, under slight agitation, to denature captured antigen. Then
l of 10 mmol Lꢀ1 IAA was added, and the immunocomplex
l
l of 6 mol Lꢀ1 urea solution in 25 mmol Lꢀ1 ammonium
l
l of 10 mmol Lꢀ1 DTT and incubated at 37 °C
Anti-CA II polyclonal antibody was purchased from U.S. Biolog-
ical (Swampscott, MA, USA), and Dynabeads Protein G were ob-
tained from Invitrogen (Carlsbad, CA, USA). Synthetic peptide
standards (5 mg each, certified title P95%), corresponding to
GGPLDGTYR (CA II proteotypic peptide) and GGPLEGTYR (internal
standard [IS]), were purchased from CRIBI Center (Padova Univer-
sity, Italy). The human CA II standard (<80%), dimethyl pimelimi-
date dihydrochloride (DMP), triethanolamine, Tween 20, Bradford
reagent, ammonium bicarbonate, phosphate-buffered saline (PBS,
pH 7.2), and trifluoroacetic acid (TFA) were obtained from Sig-
ma–Aldrich (St. Louis, MO, USA). Tris (hydroxymethyl) amino-
methane, 1,4-dithiothreitol (DTT), iodoacetamide (IAA), urea, and
protease inhibitor mix were purchased from GE Healthcare (Upp-
sala, Sweden). All organic solvents were of the highest grade avail-
8
l
was incubated at RT for 1 h in the dark. Subsequently, 8 ll of
10 mmol Lꢀ1 DTT was added and incubated at 37 °C for 1 h, under
slight agitation, to consume any leftover alkylating agent and to
avoid trypsin alkylation. The mixture was then diluted with
25 mmol Lꢀ1 ammonium bicarbonate to obtain a final urea concen-
tration of 1 mol Lꢀ1. Reconstituted trypsin solution (20 g mlꢀ1 in
l
25 mmol Lꢀ1 ammonium bicarbonate) was added to a final concen-
tration of 5.8
g mlꢀ1. The samples were allowed to digest under
gentle mixing overnight at 37 °C, the digestion was quenched by
adding 4 l of formic acid, and then 50
l of 3 pg mlꢀ1 IS solution
was added. The supernatant (ꢁ400 l) containing the peptides
coming from both antigen and antibody digestion was recovered
l
l
l
l

CA II quantification in human serum / L. Callipo et al. / Anal. Biochem. 400 (2010) 195–202
197
using a magnet, and an aliquot was subjected to LC–MS/MS
analysis.
The scheme of the on-line system is shown in Fig. 1. The two
alternate positions of the software-controlled eight-port valve al-
lowed flow switching in the trap column. The LC binary pump (pump
1) was used to deliver a 300-
umn for loading and washing the injected sample, whereas the LC
binary micropump (pump 2) was used to deliver a 70- l/min flow
ll/min flow rate through the trap col-
Chromatographic and mass spectrometric conditions
l
The LC apparatus consisted of a series 200 binary LC micro-
pump, a series 200 binary LC pump, two vacuum degassers, an
autosampler (PerkinElmer, Norwalk, CT, USA), and an eight-port
rate for eluting the analyte from the trap column in back flushing
mode, for carrying out the chromatographic run, and subsequently
for flushing and equilibrating the column. The mobile phases used
for sample loading and washing (pump 1) were water (A) and aceto-
nitrile (B), with both containing 1% (v/v) formic acid. Mobile phases
for analyte elution (pump 2) consisted of water (C) and acetonitrile/
methanol (60:40, v/v) (D), with both containing 0.1% (v/v) formic
acid.
The sample was loaded onto the trap column and desalted by
washing with 2% B for 2 min. Elution was performed by switching
the eight-port valve; after an isocratic step at 10% D for 1 min, D
was linearly increased to 40% within 7 min and then the eight-port
valve was resettled to the starting position and B and D were
brought to 95% within 1 min and held constant for 4 min to rinse
the column and trap column. Finally, the D content was lowered
Valco valve equipped with a 100-
tional, Schenkon, Switzerland). The LC separation was carried out
on a Jupiter Proteo 4- m aver-
ll peek loop (VICI AG Interna-
l
m C12 column (150 ꢂ 1 mm i.d., 4
l
age particle size, 90 Å pore size) equipped with a 2.1-mm i.d. guard
column from Phenomenex (Torrance, CA, USA). A C18 trap column
(4.0 ꢂ 2.0 mm i.d.) supplied by Phenomenex was used for on-line
solid phase extraction (SPE) with 100 ll of samples being injected.
ESI–MS was carried out on an API 3000 triple quadrupole
instrument equipped with a TurboIonSpray (TISP) interface and
with a built-in software-controlled eight-port valve (Applied Bio-
systems/MDS Sciex, Concord, ON, Canada). The LC–MS system, data
acquisition, and processing were managed by Analyst software
(version 1.4.2, Applied Biosystems/MDS Sciex).
Fig. 1. On-line LC–MS/MS CA II proteotypic peptide analysis after off-line immunoextraction and tryptic digestion. Analytical system scheme: 1, position used for sample
loading, extraction, analytical column equilibration, and cartridge regeneration; 0, position used for peptide elution from the extraction/enrichment trap column and peptide
separation by the analytical column. Mobile phases: A, water; B, acetonitrile (containing 1% [v/v] HCOOH); C, water; D, acetonitrile/methanol (60:40, v/v, both containing 0.1%
[v/v] HCOOH). a, pump 1; b, pump 2; c, autosampler; d, 100-ll loop; e, eight-port valve; f, extraction/concentration trap column; g, analytical column; h, mass spectrometer;
w, waste.

198
CA II quantification in human serum / L. Callipo et al. / Anal. Biochem. 400 (2010) 195–202
to 10% and the B content was lowered to 2% over 1 min, and both
the column and trap column were reequilibrated for 15 min.
The mass spectrometer was calibrated using polypropylene
glycol as standard (Applied Biosystems). Ionization and mass spec-
trometric conditions were optimized for both of the peptides (CA
sition pairs for the two peptides were extracted from the LC MRM
dataset, the resulting peak areas were measured, and the plot of
the ratio between the peak areas of the CA II synthetic peptide
and the IS peptide versus concentration was obtained. The ma-
trix-matched calibration line was also constructed by using a ser-
um pool obtained by mixing serum aliquots from the eight
II-specific peptide and IS peptide) by infusing at a 5-
l
l minꢀ1 flow
rate a 0.1-ng l ꢀ1
l
solution prepared in water/acetonitrile/formic
healthy subjects. Nine 20-ll aliquots were processed as reported
acid (70:30:0.1, v/v/v). TISP interface was operated in the positive
ionization mode by applying to the capillary a voltage of 5500 V.
Nitrogen was used as curtain, nebulizing, and turbo spray gases
(heated at 300 °C), and the gas pressures were set at 20, 30, and
40 psi, respectively. Nitrogen, kept at medium pressure (arbitrary
units), also served as collision gas.
above. Next, to obtain the same concentrations as for the external
calibration, the samples were transferred into vials spiked with the
suitable volume of standard and IS solutions after solvent evapora-
tion. All samples were run in triplicate interleaving three blank
samples before the next series of injections, and results were
averaged.
From the initial set of candidates, two suitable transition pairs
were chosen for acquisition in MRM mode for both peptides. The
declustering, entrance, and focusing potentials were maintained
at the optimal values of 28, 7, and 370 V, respectively, whereas col-
lision energy was optimized for each ion. The LC–ESI–MS/MS
parameters are summarized in Table 1.
Diprotonated molecules ([M + 2H]2+) were mass selected by the
first quadrupole and fragmented. In the product ion scan mode, the
range of m/z 200–1000 was monitored for both CA II proteotypic
peptide and IS.
A flow chart of the analytical procedure is shown in Fig. 2.
Recovery studies were conducted by spiking human serum
samples with different amounts of CA II standard at different pro-
cedure steps while maintaining the IS amount constant.
Calibration and recovery studies
Standard solutions for calibration were prepared by drawing
the appropriate volume of the CA II-specific peptide GGPLDGTYR
working standard solution so as to obtain nine concentration levels
Results and discussion
in the range of 0.05–100 pg l ꢀ1 l of a 3-pg l ꢀ1
and 50 l l IS peptide
l
GGPLEGTYR working standard solution. The solvent was evapo-
Abundant protein depletion
rated at 37 °C under a gentle N2 stream, and then the solution
was reconstituted with 400
l
l of 25 mmol Lꢀ1 ammonium bicar-
To avoid significant interference due to nonspecific interaction
of the most abundant serum proteins with the antibody capture
of CA II, in a first attempt, immunoaffinity depletion of serum sam-
ples was carried out as the first analytical step before immunopre-
cipitation. Depletion was performed on the Agilent Human-14
Multiple Affinity Removal System (MARS) immunoaffinity spin
cartridge according to the recommendations of the manufacturer.
After depletion, the sample was handled as described in Materials
and methods. Although recoveries from spiked samples were 80–
90%, quantitative determinations on unspiked individual serum
samples gave both lower concentrations and lower between-sam-
ple standard deviations (16.3 pmol mlꢀ1 and relative standard
deviation [RSD] = 18%, respectively, n = 8) compared with the pre-
vious results obtained in our laboratory [27]. This result suggested
that serum depletion may lead to loss of CA II to an unpredictable
extent. Probably, the CA II is concomitantly removed during deple-
tion due to the nonspecific binding to depleted proteins [28]. Thus,
immunoprecipitation of CA II was optimized on whole serum.
bonate acidified with formic acid (25 mmol Lꢀ1). Volumes of
100
ll were injected. The ion current profiles of the selected tran-
Development of magnetic separation technique
Among the different commercially available magnetic bead sys-
tems tested, characterized by a variety of surface chemistries, the
Dynabeads Protein G were chosen for the CA II antigen capture
procedure. Indeed, this magnetic bead system showed high anti-
body coupling efficiency, good functional orientation of antibody
for antigen capture, low background binding of serum peptides
and proteins, and appropriate specificity for the succeeding MS
analysis.
During method development, each step was optimized to max-
imize the final recovery of CA II. An increasing amount of anti-CA II
polyclonal antibody (25, 50, 100, and 150
fixed amount of Dynabeads Protein G (50
1.5 mg of beads) in a fixed sample volume of 100
l
g) was incubated with a
l, corresponding to
l, keeping the
l
l
incubation time constant (5 h). An estimate of the antibody
amount bound to Protein G magnetic beads was performed by
comparing protein concentrations measured by Bradford assay
by using an immunoglobulin G (IgG) standard. The supernatants
Fig. 2. Flow chart of the analytical procedure.

CA II quantification in human serum / L. Callipo et al. / Anal. Biochem. 400 (2010) 195–202
199
Table 1
Instrumental parameter settings under MRM conditions for the two peptides: CA II proteotypic peptide and IS peptide.
Peptide
Retention time (min)
4.2
Precursor ion
[M + 2H]2+
MRM transitions (m/z)
Ion type
Relative collision energy (%)a
Relative abundance (%)
GGPLDGTYR
CA II peptide
GGPLEGTYR
IS peptide
468.3 ? 411.2
468.3 ? 496.3
475.3 ? 418.2
475.3 ? 625.3
y2+
y
20
17
21
16
30
100
90
4.0
[M + 2H]2+
y2+
y
100
a
Relative collision energy is expressed as percentage with respect to the maximum voltage difference value between the high-pressure entrance quadrupole and the
collisional cell quadrupole ( 130 V) permitted by the instrument.
Table 2
Hydrolysis yield.
containing the unbound antibody were discarded, and the complex
was washed twice with the washing buffer to remove unbound
antibody. Then the antibody was recovered after complex dissoci-
ation with citrate buffer (pH 2.5), and its concentration was deter-
mined spectrophotometrically. The bonded amount of anti-CA II
Trypsin (mg)
Peptide (%)
RSD (%)
0.5
1.0
2.0
3.0
48
83
97
95
18
8
3
polyclonal antibody increased up to 14.9 0.5
solution using 100 g of antibody. This ratio (6.7
beads) is very similar to that reported by Berna and coworkers
[29] for a different antibody (8 g antibody/mg beads). Next, keep-
ing constant the amount of anti-CA II polyclonal antibody (100 g)
and the amount of Dynabeads Protein G (50 l), the efficiency of
lg for 50
ll of bead
5
l
l
g antibody/mg
Note. CA II standard added: 2 nmol, total volume = 400 ll.
l
l
l
Method validation
the Dynabeads Protein G/anti-CA II polyclonal antibody binding
reaction was tested by increasing the incubation time (1 h, 2 h,
4 h, 8 h, and overnight). No significant increase in the binding reac-
tion was observed from 2 h to overnight incubation times. The ef-
fect of cross-linking between Dynabeads Protein G and anti-CA II
polyclonal antibody was also investigated. In all cases in which
the anti-CA II polyclonal antibody was covalently bound to the
magnetic beads by a cross-linking reaction, the LC–MS/MS analysis
of the peptide at m/z 468.2 ([M + 2H]2+) exhibited a significant
peak area increase (>30%). Probably, when cross-linking was not
performed, a fraction of the immunocomplex was lost during
washing. After the best conditions for obtaining the complex anti-
body beads had been determined, we optimized the incubation
volume and the incubation time for serum samples. The same
To determine the performance characteristics of the magnetic
beads-based capture system coupled to LC–ESI–MS/MS analysis,
we assessed linear range, method detection limit (MDL), method
quantification limit (MQL), recovery, accuracy, and precision.
In addition to ionization efficiency and reproducibility, the se-
lected proteotypic peptides also need to meet the proteotypic
selection criteria described in literature [30,31]. In our previous
work based on protein isolation by reverse phase high-perfor-
mance liquid chromatography (RP–HPLC) before protein digestion,
the peptide GGPLDGTYR was selected as an appropriate peptide of
CA II and the peptide GGPLEGTYR, having very similar chemical
and physical properties, was selected as the relative IS. The benefits
and limits of a nonlabeled peptide as IS have been discussed previ-
ously [32]. For this reason, they were also the first candidates for
this method. From the product ion MS/MS spectrum of diprotonat-
ed GGPLDGTYR peptide at m/z 468.3 (shown in Fig. 3) and diprot-
onated GGPLEGTYR at m/z 475.3, the two most intense transitions
were chosen for MRM acquisition. The supernatant containing the
tryptic cleaved peptide mixture was separated from magnetic
beads, and an aliquot was injected into the LC–ESI–MS/MS system.
The mass chromatograms relevant to the two selected transitions
for CA II-specific peptide are shown in Fig. 4A and B, and we see
that background compounds present in the sample give, for the
transition 468.3 ? 212.3 (Fig. 4A), a signal poorly resolved from
that relevant to the proteotypic peptide. This problem was solved
by changing the selected transition 468.3 ? 212.3 to 468.3 ?
496.3 (Fig. 4C) that did not show background peaks. The IS was
eluted approximately 0.2 min earlier and did not show any back-
ground interference.
amount of magnetic beads with immobilized antibody (200
was used for extracting 20 l of serum spiked with 0.5 pmol of
CA II standard and diluted in different volumes (200, 500, 750,
and 1000 l) of PBS. The samples were incubated overnight at
4 °C. The best LC–MS/MS result in terms of recovery was obtained
using 20 l of serum diluted to 500 l with PBS buffer. Moreover,
the effect of sample incubation time was investigated (30 min,
1 h, 2 h, 4 h, 8 h, and overnight) using 20 l of diluted serum.
ll)
l
l
l
l
l
Recovery did not further increase after 1 h of incubation, whereas
for higher incubation times an increase of background signal in
the MRM mass chromatogram was noted, probably due to the
extraction of other proteins by means of nonspecific interactions.
Digestion efficiency
The enzymatic digestion protocol was optimized to obtain the
best peptide formation yield. For this purpose, the digestion effi-
ciency was checked by spiking immunocomplex sample after
immunoprecipitation with a known amount of CA II standard
Peptide adsorption is an important factor that may affect accu-
racy and reproducibility. To reduce peptide adsorptive processes,
protein low-retention tubes were used. Surface adsorption phe-
nomena are ruled by the adsorption constant, so the adsorbed
compound amount decreases as the solvent volume increases. In
addition, salt removal by means of C18 mini-columns is a time-
consuming critical step in the peptide analysis protocols. Bearing
in mind these considerations, we devised an on-line injection sys-
tem that permitted us to inject by an autosampler a relatively large
(2 nmol) while maintaining a constant volume (400 ll), and the
recovery of surrogate peptide by increasing the trypsin concentra-
tion was evaluated by LC–MS/MS. Experiments were done in qua-
druplicate, and the results are reported in Table 2. A recovery of
97 3% was obtained by adding to approximately 250
ll of sample
100 ), suggesting that in these
l
l of a trypsin solution (20 ng l ꢀ1
l
conditions the proteotypic peptide released from CA II digestion
could stoichiometrically represent the absolute amount of its par-
ent protein in serum. This concentration of enzyme is higher than
that usually adopted in other protocols, including those adopted by
our laboratory in a previous study [27,29].
volume (100
ll) into a small bore (1 mm i.d.) column. Our system
is more complicated than that previously used for a similar study
[29], but the current system has some advantages: (i) the injection
and trap washing with water can be made during analytical col-
umn equilibration, (ii) highly retained compounds did not enter

200
CA II quantification in human serum / L. Callipo et al. / Anal. Biochem. 400 (2010) 195–202
Fig. 3. Product ion spectrum of the diprotonated molecule of CA II-specific peptide GGPLDGTYR (m/z 468.3) acquired at 24% relative collision energy, obtained by analyzing a
10-pg l ꢀ1 l minꢀ1 flow rate.
standard solution in infusion mode, at a 5-
l
l
the analytical column, and (iii) the trap washing with strong sol-
vent can be made during sample separation.
results being averaged. MIL (signal/noise [S/N] = 3 for the second
most intense transition in MRM) and MQL (S/N = 10 for the sum
of the two selected transitions), expressed as pmol mlꢀ1 CA II in
serum, were 0.3 and 0.5, respectively.
Although the performances of the method based on immuno-
precipitation by magnetic beads were not superior to those based
on reversed phase LC fractionation, it present two advantages: (i)
many samples can be treated contemporaneously and (ii) due to
the higher selectivity and loadability, a less sophisticated MS/MS
platform can be used.
Calibration graphs were constructed as described in Materials
and methods. Two different regression lines were considered:
external calibration and matrix-matched calibration. When the
areas of the standard peaks were not normalized for the IS re-
sponse, the ratio between the slopes (bmm/bext) was 0.73 0.6,
whereas after normalization it was 1.04 0.05. These data showed
that ion suppression due to matrix compounds is moderate but not
negligible and that the unlabeled IS was adequate. The memory ef-
fect was nearly 100% when the most diluted standard solution
(0.05 pg l ꢀ1
l
)
was injected after the most concentrated one
CA II quantitation in human serum samples
(100 pg l ꢀ1
l
) and became negligible only after three blank injec-
tions. The response was found to be linear for the calibration range
used, with R2 = 0.9990 and 0.9863 for the external and matrix-
matched calibrations, respectively.
Recovery, accuracy, and reproducibility of the method were cal-
culated by analyzing six 20-ll aliquots of a serum pool unspiked
and spiked with CA II at three concentration levels during 3 weeks,
with three LC–MS/MS analyses being performed for each immuno-
precipitated sample. Results are shown in Table 3. If we consider
that the tryptic digestion regarding the representative peptide
recovery is quantitative, the roughly 23% analyte loss should be
due to an incomplete capture by the antibody. The RSD (<12%)
was comparable to the RSDs of similar studies [29,33,34].
Instrumental limit of detection (LOD), calculated as three times
the intercept of the external calibration regression line, was 2 pg of
the injected synthetic peptide (ꢁ2 fmol). The method identification
limit (MIL) and method quantification limit (MQL) were estimated
from the serum samples, with samples being run in triplicate and
Eight human serum samples from apparently healthy subjects
(seven males and one female, 20–40 years of age) were analyzed,
and for each sample three RP–HPLC runs were done. For CA II-spe-
cific peptide quantification, external calibration was employed and
recovery correction was done. The concentration of the endoge-
nous CA II in serum was found to be 27.3 pmol mlꢀ1 (RSD = 65%).
Both the mean concentration and RSD were very different from
those found in a previous study by our laboratory (56 pmol mlꢀ1
and 21%, respectively). Because both methods were validated in a
similar way and the sampled population was different but similar,
a possible explanation may be that the external addition of stan-
dard CA II was not similar to the endogenous CA II situation at
the isolation step in both cases. The chromatographic isolation
used in the previous work fractionated denaturated samples,
whereas the immunologic capture is performed in native samples.
If the protein were present in different forms (modified or associ-
ated in complex structures), the affinity for the antibody would be

CA II quantification in human serum / L. Callipo et al. / Anal. Biochem. 400 (2010) 195–202
201
preclinical validation of candidate biomarker by LC–MS/MS is
becoming a hot topic.
In the method described in this work, the protein of interest, CA
II, was isolated from serum by immunoprecipitation. A proteotypic
peptide, produced stoichiometrically through proteolysis with
trypsin, was ultimately quantified, using a synthetic peptide and
a structural analogue free-labeled synthetic peptide as IS, by LC–
ESI–MS/MS in MRM acquisition mode. An analytical column having
1 mm internal diameter and a triple quadrupole instrument were
employed. A column switching system was used for on-line SPE
sample cleanup. The assay was validated by recovery studies of
both intact proteins and proteotypic peptide. Good precision and
MDL as low as 0.5 pmol mlꢀ1 were obtained.
This strategy, based on isolation of target protein by immunoaf-
finity and its absolute quantification by quantifying one of its pro-
teotypic products, could represent a very specific and sensitive
analytical approach to reach the analytical goal of low-abundance
protein determination in clinical samples such as serum. Neverthe-
less, a comparison of physiological CA II concentrations in sera of
eight apparently healthy subjects obtained by using this approach
with those found in a previous study by using chromatographic
isolation of the fraction containing the target protein followed by
a more technologically advanced platform, such as chip LC–MS/
MS [27], showed very different values. If our data are not affected
by an unrecognized error, these results pose a question. A protein
may be present in a certain biological specimen in different forms
(e.g., free and involved in a complex with other molecules). Is the
analysis of spiked samples ever a correct way for making valida-
tion? When trying to isolate the target from the most abundant
proteins, what form is enriched?
References
[1] R. Schiess, B. Wollscheid, R. Aabersold, Targeted proteomic strategy for clinical
biomarker discovery, Mol. Oncol. 3 (2009) 33–44.
[2] D.J. Grossklaus, J.A. Smith, S.B. Shappell, C.S. Coffey, S.S. Chang, M.S. Cookson,
The free/total prostate-specific antigen ratio (%fPSA) is the best predictor of
tumor involvement in the radical prostatectomy specimen among men with an
elevated PSA, Urol. Oncol. 7 (2002) 195–198.
[3] C. Whitehouse, E. Solomon, Current status of the molecular characterization of
the ovarian cancer antigen CA125 and implications for its use in clinical
screening, Gynecol. Oncol. 88 (2003) S152–S157.
[4] A.S. Jaffe, Cardiovascular biomarkers: the state of the art in 2006, Clin. Chim.
Acta 381 (2007) 9–13.
[5] M. Pernemalm, R. Lewensohn, J. Lehtiö, Use of narrow-range peptide IEF to
improve detection of lung adenocarcinoma markers in plasma and pleural
effusion, Proteomics 9 (2009) 3414–3424.
[6] T.A. Addona, S.E. Abbatiello, B. Schilling, S.J. Skates, D.R. Mani, D.M. Bunk, C.H.
Spiegelman, L.J. Zimmerman, A.-J.L. Ham, H. Keshishian, et al., Multi-site
assessment of the precision and reproducibility of multiple reaction
monitoring-based measurements of proteins in plasma, Nat. Biotechnol. 27
(2009) 633–641.
[7] Y. Tanaka, H. Akiyama, T. Kuroda, G. Jung, K. Tanahashi, H. Sugaya, J. Utsumi, H.
Kawasaki, H. Hirano, A novel approach and protocol for discovering extremely
low-abundance proteins in serum, Proteomics 6 (2006) 4845–4855.
[8] L. Urbas, P. Brne, B. Gabor, M. Barut, M. Strliè, T. Èerk Petriè, A. Štrancar, Depletion
of high-abundance proteins from human plasma using a combination of an
affinity and pseudo-affinity column, J. Chromatogr. A 1216 (2009) 2689–2694.
[9] Y. Jmeian, Z. El Rassi, Liquid-phase-based separation systems for depletion,
prefractionation, and enrichment of proteins in biological fluids for in-depth
proteomics analysis, Electrophoresis 30 (2009) 249–261.
[10] C. Melm, M. Schuchard, D.E. Chen, H. Duewel, Examination of non-specifically
bound proteins and peptides following depletion of twenty high abundance
proteins from human plasma, Life Sci. Innov. 24 (2008) 17–20 (Sigma–Aldrich).
[11] M. Zhou, D.A. Lucas, K.C. Chan, H.J. Issaq, E.F. PetricoinIII, L.A. Liotta, T.D.
Veenstra, T.P. Conrads, An investigation into the human serum ‘‘interactome”,
Electrophoresis 25 (2004) 1289–1298.
[12] Y. Gong, X. Li, B. Yang, W. Ying, D. Li, Y. Zhang, S. Dai, Y. Cai, J. Wang, F. He, X.J.
Qian, Different immunoaffinity fractionation strategies to characterize the
human plasma proteome, J. Proteome Res. 5 (2006) 1379–1387.
[13] N. Zolotarjova, J. Martosella, G. Nicol, J. Bailey, B.E. Boyes, W.C. Barrett,
Differences among techniques for high-abundant protein depletion,
Proteomics 5 (2005) 3304–3313.
Fig. 4. LC–ESI–MS/MS extracted ion current in MRM mode deriving from the
analysis of an unspiked sample relative to 468.3 ? 212.3 transition (A),
468.3 ? 411.2 transition (B), and 468.3 ? 496.3 transition (C).
Table 3
Method recoveries and precision.
CAII added (pmol mlꢀ1
)
ꢀ
30
37.7
74
7
100
95,5
80
150
132.5
78
CAII found (pmol mlꢀ1
Recovery
)
15.5
–
RSD% (n = 6)
12
9
5
different, so the recovery of the standard may give erroneous
indications.
Conclusions
As demonstrated by the first attempts to conduct a multilabora-
tory assessment of the precision and reproducibility of MRM-based
measurements of proteins in plasma published recently [6,35], the
[14] N.J. Hannan, C.J. Stoikos, A.N. Stephens, L.A. Salamonsen, Depletion of high-
abundance serum proteins from human uterine lavages enhances detection of
lower-abundance proteins, J. Proteome Res. 8 (2009) 1099–1103.

202
CA II quantification in human serum / L. Callipo et al. / Anal. Biochem. 400 (2010) 195–202
[15] R. Stempfer, M. Kubicek, I.M. Lang, N. Christa, C. Gerner, Quantitative
assessment of human serum high-abundance protein depletion,
Electrophoresis 29 (2008) 4316–4323.
[27] L. Callipo, P. Foglia, R. Gubbiotti, R. Samperi, A. Laganà, HPLC–CHIP coupled to a
triple quadrupole mass spectrometer for carbonic anhydrase II quantification
in human serum, Anal. Bioanal. Chem. 394 (2009) 811–820.
[16] J.R. Freije, R. Bischoff, The use of affinity sorbents in targeted proteomics, Drug
Discov. Today 3 (2006) 5–11.
[28] A.K. Yocum, K. Yu, T. Oe, I.A. Blair, Effect of immunoaffinity depletion of human
serum during proteomic investigations, J. Proteome Res. 4 (2005) 1722–1731.
[29] M.J. Berna, Y. Zhen, D.E. Watson, J.E. Hale, L.A. Bradley, Strategic use of
immunoprecipitation and LC/MS/MS for trace-level protein quantification:
myosin light chain 1, a biomarker of cardiac necrosis, Anal. Chem. 79 (2007)
4199–4205.
[30] J. Kamiie, S. Ohtsuki, R. Iwase, K. Ohmine, Y. Katsukura, K. Yanai, Y. Sekine, Y.
Uchida, S. Ito, T. Terasaki, Quantitative atlas of membrane transporter
proteins: development and application of a highly sensitive simultaneous
LC/MS/MS method combined with novel in-silico peptide selection criteria,
Pharm. Res. 25 (2008) 1469–1483.
[31] P. Mallick, M. Schirle, S.S. Chen, M.R. Flory, H. Lee, D. Martin, J. Ranish, B.
Raught, R. Schmitt, T. Werner, B. Kuster, R. Aebersold, Computational
prediction of proteotypic peptides for quantitative proteomics, Nat.
Biotechnol. 25 (2007) 125–131.
[32] D.R. Barnidge, E.A. Dratz, T. Martin, L.E. Bonilla, L.B. Moran, A. Lindall, Absolute
quantification of the G protein-coupled receptor rhodopsin by LC/MS/MS using
proteolysis product peptides and synthetic peptide standards, Anal. Chem. 75
(2003) 445–451.
[33] M. Careri, L. Elviri, J.B. Lagos, A. Mangia, F. Speroni, M. Terenghi, Selective and
rapid immunomagnetic bead-based sample treatment for the liquid
chromatography–electrospray ion-trap mass spectrometry detection of Ara
h3/4 peanut protein in foods, J. Chromatogr. A 1206 (2008) 89–94.
[34] J.R. Whiteaker, L. Zhao, H.Y. Zhang, L.C. Feng, B. DPiening, L. Anderson, A.G.
Paulovich, Antibody-based enrichment of peptides on magnetic beads for
mass-spectrometry-based quantification of serum biomarkers, Anal. Biochem.
362 (2007) 44–54.
[17] I. Safarik, M. Safarikova, Magnetic techniques for the isolation and purification
of proteins and peptides, Biomagn. Res. Technol. 2 (2004) 7–24.
[18] N.R. Kitteringham, R.E. Jenkins, C.S. Lane, V.L. Elliott, B.K. Park, Multiple
reaction monitoring for quantitative biomarker analysis in proteomics and
metabolomics, J. Chromatogr. B 877 (2009) 1229–1239.
[19] B. Domon, R. Aebersold, Mass spectrometry and protein analysis, Science 312
(2006) 212–217.
[20] L. Monaci, A. Visconti, Mass spectrometry-based proteomics methods for
analysis of food allergens, Trends Anal. Chem. 28 (2009) 581–591.
[21] S. Sechi, Quantitative Proteomics by Mass Spectrometry (Methods in
Molecular Biology), Humana Press, Totowa, NJ, 2007. p. 359.
[22] M. Bantscheff, M. Schirle, G. Sweetman, J. Rick, B. Kuster, Quantitative mass
spectrometry in proteomics: a critical review, Anal. Bioanal. Chem. 389 (2007)
1017–1031.
[23] W.S. Sly, P.Y. Hu, Human carbonic anhydrases and carbonic anhydrase
deficiencies, Annu. Rev. Biochem. 64 (1995) 375–401.
[24] R.E. Tashian, D. Hewett-Emmett, S.J. Dodgson, R.E. Forster, W.S. Sly, The
value of inherited deficiencies of human carbonic anhydrase isozymes to
our understanding of their cellular roles, Ann. NY Acad. Sci. 429 (1984)
284–286.
[25] S. Palminiello, E. Kida, K. Kaur, M. Walus, K.E. Wisniewski, T. Wierzba-
Bobrowicz, A. Rabe, G. Albertini, A.A. Golabek, Increased levels of carbonic
anhydrase II in the developing Down syndrome brain, Brain Res. 1190 (2008)
193–205.
[26] R. Vlkolinsky´, N. Cairns, M. Fountoulakis, G. Lubec, Decreased brain levels of
20,30-cyclic nucleotide-30-phosphodiesterase in Down syndrome and
Alzheimer’s disease, Neurobiol. Aging 22 (2001) 547–553.
[35] M. McIntosh, M. Fitzgibbon, Biomarker validation by targeted mass
spectrometry, Nat. Biotechnol. 27 (2009) 622–623.