Development of a selective ESI-MS derivatization reagent: Synthesis and optimization for the analysis of aldehydes in biological mixtures

Article abstract of DOI:10.1021/ac801429w

In LC-MS, derivatization is primarily used to improve ionization characteristics, especially for analytes that are not (efficiently) ionized by ESI or APCI such as aldehydes, sugars, and steroids. Derivatization strategies are then directed at the incorporation of a group with a permanent charge. A compound class that typically requires derivatization prior to LC-MS is the group of small aliphatic aldehydes that are, for instance, analyzed as the key biomarkers for lipid peroxidation in organisms. Here we report the development of a new tailor-made, highly sensitive, and selective derivatization agent 4-(2-(trimethylammonio)ethoxy)benzenaminium halide (4-APC) for the quantification of aldehydes in biological matrixes with positive ESI-MS/MS without additional extraction procedures. 4-APC possesses an aniline moiety for a fast selective reaction with aliphatic aldehydes as well as a quaternary ammonium group for improved MS sensitivity. The derivatization reaction is a convenient one-pot reaction at a mild pH (5.7) and temperature (10°C). As a result, an in-vial derivatization can be performed before analysis with an LC-MS/MS system. All aldehydes are derivatized within 30 min to a plateau, except malondialdehyde, which requires 300 min to reach a plateau. All derivatized aldehydes are stable for at least 35 h. Linearity was established between 10 and 500 nM and the limits of detection were in the 3-33 nM range for the aldehyde derivatives. Furthermore, the chosen design of these structures allows tandem MS to be used to monitor the typical losses of 59 and 87 from aldehyde derivatives, thereby enabling screening for aldehydes. Finally, of all aldehydes, pentanal and hexanal were detected at elevated levels in pooled healthy human urine samples.

Full text of DOI:10.1021/ac801429w

Anal. Chem. 2008, 80, 9042–9051
Development of a Selective ESI-MS Derivatization
Reagent: Synthesis and Optimization for the
Analysis of Aldehydes in Biological Mixtures
Mark Eggink,† Maikel Wijtmans,‡ Reggy Ekkebus,† Henk Lingeman,† Iwan J. P. de Esch,‡
Jeroen Kool,† Wilfried M. A. Niessen,† and Hubertus Irth*,†
Leiden/Amsterdam Center for Drug Research, Faculty of Sciences, Department of Chemistry, Section of Analytical
Chemistry & Applied Spectroscopy, and Department of Pharmacochemistry, Section of Medicinal Chemistry, VU
University Amsterdam, De Boelelaan 1083, 1081 HV Amsterdam, The Netherlands
In LC-MS, derivatization is primarily used to improve
ionization characteristics, especially for analytes that
are not (efficiently) ionized by ESI or APCI such as
aldehydes, sugars, and steroids. Derivatization strate-
gies are then directed at the incorporation of a group
with a permanent charge. A compound class that
typically requires derivatization prior to LC-MS is the
group of small aliphatic aldehydes that are, for instance,
analyzed as the key biomarkers for lipid peroxidation
in organisms. Here we report the development of a new
tailor-made, highly sensitive, and selective derivatiza-
volatility of the analyte, to alter its ionization characteristics, or
to influence its fragmentation behavior. In combined liquid
chromatography-MS (LC-MS), however, where soft ionization
techniques like electrospray (ESI) and atmospheric pressure
chemical ionization (APCI) are applied, derivatization is generally
not needed and avoided as much as possible. In LC-MS,
derivatization is primarily used to improve ionization character-
istics, especially for analytes that are not (efficiently) ionized by
ESI or APCI such as aldehydes, sugars, and steroids.1-4 Deriva-
tization strategies are then directed at the incorporation of a group
with a permanent charge (cationic groups for positive-ion mode
and strong acidic functionalities for negative-ion mode) or other
groups that enhance ionization (secondary or tertiary amine for
positive-ion mode or aromatic nitro groups in negative-ion mode).
In addition, derivatization may be directed at improving the
fragmentation characteristics in tandem MS (MS/MS). One of
the drawbacks of derivatization is that the reaction may need harsh
conditions, which may not only derivatize the target analyte but
can also affect other constituents of the complex samples to be
analyzed. Such side products may effect the derivatization reaction
or may require additional sample pretreatment steps or advanced
chromatographic separation to be removed.
The most efficient derivatization agents directed at achieving
higher ionization efficiency in ESI contain a group with a
permanent charge in addition to a reactive functional group. A
variety of such derivatization agents are available for different
classes of analytes, such as the commercially available Girard T
and P reagents, containing a quaternary ammonium and a
pyridinium moiety, respectively.5-8 Girard T and P reagents have
been reported as reagents for steroids, peptides, and nucleotides.
New synthesized derivatization agents for these molecules featur-
tion
agent
4-(2-(trimethylammonio)ethoxy)benze-
naminium halide (4-APC) for the quantification of
aldehydes in biological matrixes with positive ESI-MS/
MS without additional extraction procedures. 4-APC
possesses an aniline moiety for a fast selective reaction
with aliphatic aldehydes as well as a quaternary am-
monium group for improved MS sensitivity. The deriva-
tization reaction is a convenient one-pot reaction at a
mild pH (5.7) and temperature (10 °C). As a result,
an in-vial derivatization can be performed before analy-
sis with an LC-MS/MS system. All aldehydes are
derivatized within 30 min to a plateau, except malon-
dialdehyde, which requires 300 min to reach a plateau.
All derivatized aldehydes are stable for at least 35 h.
Linearity was established between 10 and 500 nM and
the limits of detection were in the 3-33 nM range for
the aldehyde derivatives. Furthermore, the chosen
design of these structures allows tandem MS to be used
to monitor the typical losses of 59 and 87 from aldehyde
derivatives, thereby enabling screening for aldehydes.
Finally, of all aldehydes, pentanal and hexanal were
detected at elevated levels in pooled healthy human
urine samples.
(1) Naven, T. J. P.; Harvey, D. J. Rapid Commun. Mass Spectrom. 1996, 10,
829–834
(2) Roth, K. D. W.; Huang, Z. H.; Sadagopan, N.; Watson, J. T. Mass Spectrom.
Rev. 1998, 17, 255–274
(3) Fernandes-Whaley, M.; Muhlberger, F.; Whaley, A.; Adam, T.; Zimmer-
mann, R.; Rohwer, E.; Walte, A. Anal. Chem. 2005, 77, 1–10
(4) Barry, S. J.; Carr, R. M.; Lane, S. J.; Leavens, W. J.; Monte, S.; Waterhouse,
I. Rapid Commun. Mass Spectrom. 2003, 17, 603–620
(5) Griffiths, W. J.; Liu, S.; Alvelius, G.; Sjovall, J. Rapid Commun. Mass
Spectrom. 2003, 17, 924–935
(6) Habrioux, G. Ann. Biol. Clin. (Paris) 1978, 36, 433–442
(7) Hong, H. H.; Wang, Y. S. Anal. Chem. 2007, 79, 322–326
(8) Quirke, J. M. E.; Van Berkel, G. J. J. Mass Spectrom. 2001, 36, 1294–
1300
.
.
Analyte derivatization has played an important role in analysis
using combined gas chromatography-mass spectrometry (GC/
MS). In GC/MS, derivatization is performed to enhance the
.
.
.
* To whom correspondence should be addressed. E-mail: Irth@few.vu.nl.
† Department of Chemistry, Section of Analytical Chemistry & Applied
Spectroscopy.
.
.
‡ Department of Pharmacochemistry, Section of Medicinal Chemistry.
.
9042 Analytical Chemistry, Vol. 80, No. 23, December 1, 2008
10.1021/ac801429w CCC: $40.75 2008 American Chemical Society
Published on Web 10/15/2008

Scheme 1. Reaction Sequence for the Three-Step Synthesis of 4-APC
ing a permanent charge have been reported, providing selectivity
toward specific analytes.4,9-11 A compound class that requires
derivatization prior to LC-MS analysis is the group of small
aliphatic aldehydes that are, for instance, analyzed as the key
biomarkers for lipid peroxidation in organisms.12 Lipid peroxida-
tion plays an important role in a wide variety of diseases such as
heart disease, diabetes mellitus, and atherosclerosis.13-15 As
biomarkers, these aldehydes indicate the extent of free radical
damage to different polyunsaturated fatty acids. They are excreted
and thus analyzed in biological matrixes, for example, plasma,
urine, or exhaled breath condensate. The most widely used
derivatization agent for aldehydes is 2,4-dinitrophenylhydrazine
(DNPH).16-20 Alternatively, pentafluorophenyl-, 2-chloro-, and 2,4-
dichlorophenylhydrazine and cyclohexanedione21,22 are applied
as derivatization reagents. These reagents were typically designed
for UV-vis and fluorescence detection but are currently also used
in combination with MS detection, for example, for the detection
of aldehyde-DNPH derivatives in the nanomolar range by means
of negative-ion APCI.18,19 Derivatization of aldehydes is mostly
based on the fast reaction of hydrazines under relatively harsh
conditions, e.g., at high temperatures and in acidic medium. An
additional liquid-liquid extraction step, frequently with n-hexane,
is implemented prior to reversed-phase LC separation and MS
detection. DNPH derivatization is fast, and stable derivatives are
formed under acidic conditions. A drawback is that stereoisomeric
hydrazone derivatives are formed,23 which may be hydrolyzed
under nonacidic conditions. Furthermore, because hydrazines not
only react with aldehydes but also with ketones and small
carboxylic acids, this derivatization lacks selectivity.
In this paper, we report on a novel derivatization agent, 4-(2-
(trimethylammonio)ethoxy)benzenaminium halide (4-APC), that
allows highly sensitive and selective MS detection of aldehydes.
The design of the compound was based on a set of strict criteria
and relies on a unique pH profile of anilines in NaBH3CN-based
reductive aminations. A convenient gram-scale synthesis route is
given, allowing easy access to the material. Excellent quantification
of aldehydes in biological matrixes with positive ESI-MS/MS can
be achieved. Using 4-APC as derivatization agent, aldehydes can
be detected in the low-nanomolar range. One of the key features
of 4-APC, a permanent positive charge, will significantly reduce
ion suppression, provide gain in ionization efficiency, and is ideally
suited for structure-specific MS-MS screening strategies in
biological samples.
EXPERIMENTAL SECTION
Chemicals. Pentanal, 2-pentanone, trans-2-pentenal, hexanal,
heptanal, octanal, nonanal, decanal, cyclohexylcarboxaldehyde, 1,2-
dibromoethane, trimethylamine in ethanol (4.2 M), phenethy-
lamine, 1,1,3,3-tetramethoxypropane (TMP), hydrochloric acid,
p-hydroxyacetanilide, tetrabutylammonium perchlorate, sodium
cyanoborohydride (NaBH3CN), benzoylcholine chloride, chloro-
form, pyridoxamine, biotin hydrazide, aniline, sodium carbonate,
ammonium acetate, and methylethylketone were all purchased
from Sigma-Aldrich (Zwijndrecht, The Netherlands).
Methanol (MeOH), formic acid were purchased from Biosolve
(Valkenswaard, The Netherlands), ethanol was purchased from
Mallinckrodt-Baker (Deventer, The Netherlands), and amine PEO2
biotin was purchased from Thermo Fisher Scientific Inc. Pooled
urine was obtained from five healthy human volunteers.
Synthesis of 4-(2-(trimethylammonio)ethoxy)benzenami-
nium Chloride Bromide (1). General Note about the Synthesis.
The synthetic pathway is shown in Scheme 1. It should be
mentioned that, in our latest efforts to further optimize the
synthesis of the derivatization reagent, we found that the last step
(i.e., from C to 1) can be more conveniently carried out with
aqueous 5 N HBr instead of aqueous 5 N HCl in an otherwise
identical synthetic procedure. This protocol is higher yielding and
gives 4-(2-(trimethylammonio)ethoxy)benzenaminium dibromide.
As expected, in the derivatization protocols, this dibromide salt
behaves identically in all respects to the current Cl,Br salt.
N-(4-(2-Bromoethoxy)phenyl)acetamide (B). A procedure adapted
from Murakami et al. with minor modifications was used.24
p-Hydroxyacetanilide A (10.0 g, 66 mmol) and 1,2-dibromoethane
(57.2 mL, 124.7 g, 664 mmol) were added to saturated sodium
carbonate solution (60 mL). Tetrabutylammonium perchlorate
(9) Johnson, D. W. Rapid Commun. Mass Spectrom. 2007, 21, 2926–2932
(10) Spikmans, V.; Lane, S. J.; Leavens, B.; Manz, A.; Smith, N. W. Rapid
Commun. Mass Spectrom. 2002, 16, 1377–1388
(11) Yang, W. C.; Mirzaei, H.; Liu, X. P.; Regnier, F. E. Anal. Chem. 2006, 78,
4702–4708
(12) De Zwart, L. L.; Meerman, J. H. N.; Commandeur, J. N. M.; Vermeulen,
N. P. E. Free Radical Biol. Med. 1999, 26, 202–226
(13) Bhor, V. M.; Raghuram, N.; Sivakami, S. Int. J. Biochem. Cell Biol. 2004,
36, 89–97
(14) Shacter, E. Drug Metab. Rev. 2000, 32, 307–326
(15) Halliwell, B.; Whiteman, M. Br. J. Pharmacol. 2004, 142, 231–255
(16) Nagy, K.; Pollreisz, F.; Takats, Z.; Vekey, K. Rapid Commun. Mass Spectrom.
2004, 18, 2473–2478
(17) Kolliker, S.; Oehme, M.; Dye, C. Anal. Chem. 1998, 70, 1979–1985
(18) Andreoli, R.; Manini, P.; Corradi, M.; Mutti, A.; Niessen, W. M. A. Rapid
Commun. Mass Spectrom. 2003, 17, 637–645
(19) Korchazhkina, O.; Exley, C.; Spencer, S. A. J. Chromatogr., B 2003, 794,
353–362
(20) Stafiej, A.; Pyrzynska, K.; Ranz, A.; Lankmayr, E. J. Biochem. Biophys.
Methods 2006, 69, 15–24
(21) O’Brien-Coker, I. C.; Perkins, G.; Mallet, A. I. Rapid Commun. Mass
Spectrom. 2001, 15, 920–928
(22) Luo, X. P.; Yazdanpanah, M.; Bhooi, N.; Lehotay, D. C. Anal. Biochem.
1995, 228, 294–298
(23) Uchiyama, S.; o, M.; Aoyagi, S. J. Chromatogr., A 2003, 996, 95–102
.
.
.
.
.
.
.
.
.
.
.
.
.
.
(24) Murakami, H.; Kawabuchi, A.; Matsumoto, R.; Ido, T.; Nakashima, N. J. Am.
.
Chem. Soc. 2005, 127, 15891–15899.
Analytical Chemistry, Vol. 80, No. 23, December 1, 2008 9043

(0.45 g, 1.3 mmol) was added as a catalyst. The mixture was
vigorously stirred for 28 h at 80 °C after which water and the
excess of 1,2-dibromoethane were evaporated under high vacuum.
The residue was extracted with chloroform (250 mL) to remove
inorganic salts and amounts of disubstituted products. The
chloroform extract was washed with a saturated sodium carbonate
solution (2×). The chloroform layer was dried over MgSO4,
filtered, and concentrated. Recrystallization from ethanol/water
yielded a white solid (5.3 g, 20.6 mmol, 31%). Spectral data were
in agreement with the literature.24 1H NMR (CDCl3/drop DMSO-
d6, 200 MHz) δ 7.5 (d, 2H, CHCNHAc), 6.8 (d, 2H, CHCOCH2),
4.2 (t, 2H, OCH2), 3.6 (t, 2H, CH2Br), 2.0 (s, 3H, COCH3).
(2-(4-Acetamidophenoxy)ethyl)trimethylammonium Bromide
(C). A procedure adapted from Marlow et al. with major modifica-
tions was used.25 To prevent working with gaseous trimethy-
lamine, a commercial solution of Me3N in EtOH was implemented
in the existing procedure. Thus, compound B (5.2 g, 20.2 mmol)
was mixed with a solution of Me3N in EtOH (4.2 M, 47.6 mL,
199.9 mmol). Cosolvent methylethylketone (40 mL) was added,
and the mixture was stirred at room temperature for 4 days, during
which time the product precipitated. It was filtered and recrystal-
lized from hot EtOH/water to afford C as pinkish/beige transpar-
ent crystals (4.1 g, 12.9 mmol, 64%). 1H NMR (D2O, 200 MHz) δ
7.3 (d, 2H, CHCNHAc), 7.0 (d, 2H, CHCOCH2), 4.4 (m, 2H,
OCH2), 3.7 (m, 2H, CH2NMe3), 3.2 (s, 9H, NMe3), 2.1 (s, 3H,
COCH3). Extended splitting of some 1H-signals is well-known for
choline-type structures.26
TMP to MDA (final concentration ∼10 mM). The concentration
of MDA was determined by measuring its absorbance at 245 nm
(ꢀ ) 31 800 mol L-1 cm-1) according to Esterbauer et al.27 Before
analysis, MDA stock solution was diluted with H2O to working
solution. For each aldehyde, a stock solution of 10 mM in MeOH
was prepared and stored in the freezer at -20 °C. The stock
solution was diluted with H2O to the final working mixed
aldehydes solution.
Optimization of Reaction Conditions. To determine the deriva-
tization speed and stability, the formed derivatives were measured
at 1, 10, 30, 60, 120, and 600 min. Alternatively, derivatizations
were left for 35 hours and measurements taken every hour.
Repeatability of the derivatization was determined by performing
the derivatization in triplicate and analyzing the formed derivatives.
Detection limits and linearity of the selected aldehydes were
determined by a calibration series ranging from 0 to 500 nM
aldehydes. The calibration series were measured in buffer and in
urine.
Sample Pretreatment and Derivatization. For the optimization
reactions, 200 µL of 2 mg/mL biotin hydrazide, amine PEO2 biotin,
phenethylamine, pyridoxamine, aniline, or 4-APC in 50 mM
ammonium acetate buffer pH 5.7, 50 µL of NaBH3CN (4 mg/mL
in MeOH), 50 µL of 2 µM benzoylcholine chloride in H2O as
internal standard (IS) and 250 µL of mixed aldehyde standard
consisting of trans-2-pentenal, pentanal, hexanal, cyclohexylcar-
boxaldehyde, heptanal, octanal, nonanal, decanal, and MDA with
a final concentration of 500 nM were mixed and vortexed for 60 s.
The urine samples were centrifuged at 13 600 rpm for 15 min at
10 °C before the derivatization reaction. The derivatization was
carried out at 10 °C in the cooled autosampler of HP1100. After
3 h, the first sample was injected in the LC-MS/MS for analysis.
Pooled Urine Samples. Pooled urine sample of five healthy
volunteers were collected and stored at -20 °C. The 250 µL of
urine was mixed with 200 µL of 4-APC (2 mg/mL) in buffer pH
5.7, 50 µL of NaBH3CN (4 mg/mL) in MeOH, and 50 µL of
aldehyde spike consisting of pentanal, hexanal, octanal, and
decanal (final concentration 10 nM). The derivatization was
performed as mentioned above.
4-(2-(Trimethylammonio)ethoxy)benzenaminium Chloride Bro-
mide (1, 4-APC). A procedure adapted from Marlow et al. with
major modifications was used.25 Salt C (4.0 g, 12.6 mmol) was
heated at reflux with hydrochloric acid (5 N, 40 mL) for 30 min.
After cooling, the solvent was evaporated from the solution leaving
a solid residue, which was recrystallized from hot EtOH/water.
Filtering, washing, and drying afforded shiny beige crystals with
excellent purity (1.9 g). Elemental analysis revealed that this
particular crystal lattice consisted of the desired dication with 0.43
chloride counterion and 1.57 bromide counterion. Thus, for the
yield calculation and throughout our work, a mass of 337.0 was
used for the final product. Yield 45% (5.64 mmol, 1.9 g). The
product is stable for at least several weeks (as judged from 1H
NMR) in aerated D2O solution and for months as a solid at room
temperature in the dark. λmax (H2O) 277 nm. 1H NMR (D2O, 200
MHz) δ 7.4 (d, 2H, CHCNH3), 7.1 (d, 2H, CHCOCH2), 4.5 (m,
2H, OCH2), 3.8 (m, 2H, CH2NMe3), 3.2 (s, 9H, NMe3). 13C NMR
(D2O, 50 MHz) δ 158.8, 125.9, 124.9, 117.4, 66.5, 63.6, 55.5.
HPLC. All HPLC separations were performed on an Agilent
1100 HPLC system (Agilent Technologies, Amstelveen, The
Netherlands) controlled by Chemstation Rev B.01.09. A Waters
XTerra MS reversed-phase column (C18 100 × 2.1 mm, 3 µM) at
45 °C and with a flow rate of 150 µL/min was used for the
separation of the derivatized aldehydes standards and urine
samples. Samples were injected (10 µL) from a thermostatic
autosampler kept at 10 °C.
1
Extended splitting of some H-signals is well-known for choline-
type structures.26 HR-MS (C11H19N2O+): calcd 195.1492, found
195.1497. Elem. Anal. (C11H20Br1.57Cl0.43N2O): calcd (%) C 39.21,
H 5.98, Br 37.23, Cl 4.52, N 8.31, O 4.75; found (%, n ) 2) C 39.22,
H 6.27, Br 37.25, Cl 4.43, N 8.38, O 4.82. LC-MS purity: >98%
(MS, BPI).
Preparation of Malondialdehyde (MDA) Standard and Aldehyde
Solution. TMP was used to prepare a malondialdehyde stock
solution. A volume of 17 µL was dissolved in HCl 0.1 M (10 mL),
and this solution was incubated at 40 °C for 60 min to hydrolyze
The gradient elution was programmed as follows: After
injection, 100% mobile phase A (95% H2O + 5% MeOH + 0.1%
formic acid) was maintained for 5 min and then solvent B (5%
H2O + 95% MeOH + 0.1% formic acid) was increased from 0 to
90% in 10 min with a 5-min hold at 90% B. After this, the column
was reconditioned for 10 min at 100% mobile phase A. The effluent
from the LC column was directed to the mass spectrometer.
Mass Spectrometer. A Micromass (Wythenshawe, Manchester,
U.K.) Q-TOF2 mass spectrometer equipped with a Micromass
Z-spray ESI source was used for detection. MassLynx software
(version 3.5) running under Windows NT was used for control of
(25) Marlow, H. F.; Metcalfe, J. C.; Burgen, A. S. V. Mol. Pharmacol. 1969, 5,
156–165
(26) Sega, E. M.; Tormena, C. F.; de Oliveira, P. R.; Rittner, R.; Tinoco, L. W.;
Figueroa-Villar, J. D.; Hoehr, N. F. J. Mol. Struct. 2006, 797, 44–48
.
.
(27) Esterbauer, H.; Cheeseman, K. H. Methods Enzymol. 1990, 186, 407–421.
9044 Analytical Chemistry, Vol. 80, No. 23, December 1, 2008

Figure 1. Time curves with derivatization speed and stability of nonanal derivatized with biotin hydrazide, aniline, pyridoxamine, biotin PEO2
amine, 4-APC, and phenethylamine.
the system and data acquisition. The time-of-flight analyzer was
operated at a 20-kHz frequency with a spectrum integration time
of 1 s in “full spectrum” MS in the positive-ion mode in the range
m/z 150-450 (“interscan” time, 0.1 s). The ESI source conditions
for the HPLC analysis were as follows: source temperature 100
°C, desolvation temperature 250 °C, and capillary voltage 2.5 kV.
The sampling cone voltage was 20 V. For the MS/MS experi-
ments, the collision energy was set on 20 V. Nitrogen (99.999%
purity; Praxair, Oevel, Belgium) was used with flow rates of 20
L/h for nebulization, 50 L/h for cone gas, and 350 L/h for
desolvation. Argon (99.9995% purity; Praxair) was used in the
collision cell.
derivatives also stable in acidic environment. The biotin derivatives
can be detected with high sensitivity in positive-ion ESI-MS, the
derivatization reaction is fast, and the reaction conditions are mild.
Unfortunately, ketones are also derivatized under these conditions,
whereas biotin derivatives of larger aliphatic aldehydes such as
nonanal and decanal show solubility problems, as can be con-
cluded from Figure 1 where the biotin hydrazone derivative of
nonanal is declining over time. Smaller aliphatic aldehydes were
also measured and showed same derivatization rates as nonanal
but stable derivatives. These findings illustrate that derivatization
with neither DNPH nor biotin hydrazide was suitable for our
purposes. This prompted us to develop a tailor-made derivatization
reagent that would meet the criteria listed below.
RESULTS AND DISCUSSION
Selection Criteria for a Tailor-Made Derivatization Agent.
The development of our new derivatization agent for aldehydes
was based on a number of important criteria: The derivatization
agent should contain a group reactive to aldehydes, but it should
not react with other carbonyl-containing compounds such as
ketones and carboxylic acids. The reaction time for derivatization
should be in the same order as the benchmark DNPH derivati-
zation method, but the reaction should take place under less harsh
conditions, i.e., mild pH and temperature. The fragmentation
pattern should give additional conformation to identify possible
derivatized aldehydes. It should involve a convenient one-pot
reaction, preferably not requiring additional extraction procedures.
Furthermore, the formed derivatives should be water-soluble and
be stable for a significant amount of time to allow large sample
sets to be derivatized simultaneously and analyzed subsequently.
Finally, with the aim of maximizing MS sensitivity in positive-ion
ESI-MS analysis, the compound has to be equipped with a
permanent positive charge.
Initial Screen for Candidate Derivatization Agents. At the
onset of our project, we screened several existing derivatization
agents for selective aldehyde derivatization (Figure 1). Because
other biomarkers will be determined in the same reaction mixture
in the near future, we selected positive-ion ESI-MS as our method
of choice. Therefore, since benchmark DNPH derivatives are
typically analyzed in negative-ion APCI-MS, another hydrazide had
to be chosen. A possible agent is biotin hydrazide, which is
typically used for biotinylation of carbohydrates and carbonyl
compounds in combination with positive-ion ESI-MS.28,29 Biotin
hydrazide rapidly reacts with aldehydes and ketones at pH 4-6
and attaches the biotin group through stable hydrazone bonds.
However, our experiments show that these hydrazones in an
aqueous environment are only stable in the presence of biotin
hydrazide or at alkaline pH (data not shown). Furthermore, the
LC separation of the aldehyde derivatives at basic pH results in
doublet peaks for a single aldehyde adduct. The identical m/z
for the two peaks indicate that stereoisomers are formed (cis/
trans). Actually, Uchiyma et al.23 reported a similar effect with
the derivatization of aldehydes with DNPH. Reduction of the
hydrazone bond with sodium cyanoborohydride (NaBH3CN) is
required to avoid stereoisomer formation and to obtain biotin
Selection of the Nucleophilic Group. We decided to explore
reversible imine formation from an aldehyde and a nucleophilic
NH2 group followed by an in situ irreversible reduction by
NaBH3CN to yield stable secondary amines. This is attractive for
selectivity reasons as it is known that many ketones tend to react
more slowly than aldehydes in this sequence.30 A key obstacle is
(28) Masso, F.; Varela, E.; Paez, A.; Zenteno, E.; Montano, L. F. Glycoconjugate
J. 1998, 15, 843–846
(29) van Bommel, M. R.; de Jong, A. P. J. M.; Tjaden, U. R.; Irth, H.; van der
Greef, J. J. Chromatogr., A 1999, 855, 397–409
.
(30) Borch, R. F.; Bernstein, M. D.; Durstlb, H. D. J. Am. Chem. Soc. 1971, 93,
.
2897–2904.
Analytical Chemistry, Vol. 80, No. 23, December 1, 2008 9045

Figure 2. Pathways for imine formation and reduction using aldehydes/ketones/malondialdehyde and 4-APC.
that efficient reduction by NaBH3CN only takes place at proto-
nated imines, i.e., at pH <7.30 We considered various nucleophilic
NH2 groups for derivatization at such subneutral pH values. The
protonation requirement did not bode well for use of an alkylamine
(pKa ≈ 10-11) in our reagent, because such a group will be fully
protonated at subneutral pH and thereby lose its nucleophilic
character. To confirm these expectations, we selected three
different primary alkylamines, amine PEO2 biotin, phenethylamine,
and pyridoxamine, as derivatization agents and evaluated the
reaction time. As shown in the Figure 1, the reaction times of all
the primary alkylamines were longer than 8 h for a complete
derivatization of nonanal.
Other potential NH2 groups of interest are primary aromatic
amines (the class of anilines) because these have lower pKa values.
Two recent papers by Dirksen et al. describe the elegant use of
aniline as a very effective catalyst for the condensation of
aldehydes with oximes or hydrazides.31,32 Key to the acceleration
is the in situ formation of protonated, highly reactive aniline
imines.32 While trying to use this aniline methodology in catalyzing
our early-stage derivatizations with biotin hydrazide as mentioned
above, we were surprised to see that in situ treatment with
NaBH3CN gave no reduced biotin hydrazones at all. Rather, only
clean and rapid formation of the reduced imine resulting from
the aldehyde and the catalyst aniline was observed. This is fully
in line with the observations by Dirksen et al., who noticed the
exact same behavior when aniline was applied as a catalyst in the
reaction of glyoxylyls and hydrazides.31 This unexpected outcome
is a direct result of the pKa of aniline (pKa ) 4.6), which is low
enough to retain nucleophilic character yet high enough to warrant
fast reduction of the resulting imine by NaBH3CN, all at pH 4.5
or, in our case, at pH 5.7.31 This dual reactivity clearly renders
aniline a superior candidate for our purposes. Indeed, when
nonanal was mixed with just aniline and NaBH3CN, a very rapid
formation of the stable reduced aniline adduct was observed
(Figure 1). It was therefore decided to take advantage of this
unique pH profile of aniline by incorporating it into the new
derivatization reagent.
Additional Structural Features. To meet our criteria, aniline
needed to be subjected to other structural demands, e.g., inclusion
of a permanent positive charge. For this, we turned to p-MeO-
aniline (anisidine, pKa ) 5.3), which was also reported by Dirksen
et al.32 Replacing the MeO group by a substituted alkoxy group
will have little effect on the pKa and thus NH2 reactivity but
provides a point for convenient synthetic elaboration. The unit
providing the permanent charge is a quaternary tetraalkylammo-
nium group, which will also increase the water solubility of
adducts. The ammonium group was connected to the oxyaniline-
head through an ethyl spacer to minimize any potential inductive
effects from the positive charge on the nucleophility of the NH2.
In the end, this led to 4-APC (1) shown in Scheme 1 as our novel
derivatization agent. Although 4-APC has previously been studied
in research on (acetyl)choline25,33 and multilamellar vesicles,34
to our knowledge it has never been reported as a derivatization
agent nor has any aniline in a reductive-amination based protocol
for aldehyde biomarker research.
An added bonus of this type of structure is the known and
predictable fragmentation pattern of the loss of the quaternary
amine, which is known from e.g. tandem MS on acetylcholine
and choline.35,36 We introduced strategic changes into an existing
synthetic procedure25 to give gram amounts of 4-APC without a
need for chromatographic purification.
Derivatization Reactions of Aldehydes, Ketones, and
MDA. The 4-APC derivatization reagent in combination with
NaBH3CN (both in excess) was subjected to a diverse set of
carbonyl compounds at pH 5.7. Figure 2 shows the reaction
pathways involved. Striking differences in reaction outcome with
(31) Dirksen, A.; Dirksen, S.; Hackeng, T. M.; Dawson, P. E. J. Am. Chem. Soc.
2006, 128, 15602–15603
(32) Dirksen, A.; Hackeng, T. M.; Dawson, P. E. Angew. Chem., Int. Ed. 2006,
45, 7581–7584
(33) Batzold, F.; Dehavent, R.; Kuhart, M. I.; Birdsall, N. Biochem. Pharmacol.
1980, 29, 2413–2416
(34) Takakura, K.; Sugawara, T. Langmuir 2004, 20, 3832–3834
(35) Uutela, P.; Reinila, R.; Piepponen, P.; Ketola, R. A.; Kostiainen, R. Rapid
Commun. Mass Spectrom. 2005, 19, 2950–2956
(36) Dunphy, R.; Burinsky, D. J. J. Pharm. Biomed. Anal. 2003, 31, 905–915
.
.
.
.
.
.
9046 Analytical Chemistry, Vol. 80, No. 23, December 1, 2008

Figure 3. Reaction kinetics of derivatization of 4-APC and trans-2-pentenal, pentanal, hexanal, heptanal, nonanal, decanal, malondialdehyde,
and benzoylcholine as IS.
Figure 4. Adduct intensity with (A) 500 nM 2-pentanone after 1 h, (B) 500 nM pentanal after 1 h of derivatization, (C) 500 nM 2-pentanone
after 20 h, and (D) 500 nM pentanal after 20 h of derivatization.
aldehydes, ketones, and MDA were observed (see Figure 3). As
anticipated, aldehydes are rapidly and irreversibly reduced by
NaBH3CN to amines 3 through intermediacy of imines 2.
Ketones, on the other hand, were only poorly derivatized and no
amines 6 or the precursor imines were observed by MS at
nanomolar range. Figure 4 shows that, at a concentration of 500
nM of 2-pentanone, no peak was detected after 1 h whereas
pentanal gave a large peak within the same time. After 20 h, a
small peak of 2-pentanone adduct was detected where pentanal
gave the same amount as after 1 h. Based on a previous report,
it is anticipated that this low reactivity of ketones is due to
difficulties in the initial imine formation between the ketone and
the aniline moiety in aqueous environments.30 Intriguingly, MDA
did react with 4-APC, although slower than for saturated alde-
hydes, but the resulting imine/enamine 4 was left untouched by
NaBH3CN. In general, enamines conjugated with a carbonyl are
Analytical Chemistry, Vol. 80, No. 23, December 1, 2008 9047

Figure 5. Typical LC-MS (combined) XIC chromatogram of nine aldehyde standards (50 nM) and benzoylcholine as IS (80 nM).
known to resist NaBH3CN reductions at mild pH,30 and in our
case, a set of papers by Tamura et al. on the reaction of MDA
with 4-MeO-aniline (anisidine) can be invoked to rationalize our
findings with MDA in greater detail.37,38
Optimization of Reaction Conditions. The data in Figure 3
clearly demonstrate that the aldehydes pentanal, hexanal, heptanal,
nonanal, decanal, and trans-2-pentenal were derivatized within 30
min to stable products which last for at least 10 h. Further stability
studies have shown (see Supporting Information Figure S5) that
the derivatives are stable for at least 35 h. The participation of
trans-2-pentenal is noteworthy as the double bond does not
interfere in the derivatization. As explained above, MDA reacts
significantly slower than the other aldehydes: a plateau is reached
after 300 min. A 250-fold molar excess of 4-APC and 125-fold molar
excess of NaBH3CN is used in the derivatization for three reasons.
Lower amounts give the same results but in the end biological
samples with unknown amounts of aldehydes will be used.
Furthermore, low concentrations of endogenous primary amines
can be present in these biological samples that could also react
with the aldehydes. Last, employing an excess of 4-APC will
effectively block a potential second reductive amination with 3
as the nucleophile.
MDA, trans-2-pentenal, pentanal, hexanal, heptanal, octanal,
nonanal, and decanal are potential lipid peroxidation biomarkers
in biological samples. Therefore, cyclohexylcarboxaldehyde was
added to biological samples prior to derivatization as a nonen-
dogenous aldehyde for control. Benzoylcholine was used as
internal standard to correct for any signal fluctuations during
long sequence runs in the LC-MS. Because trace amounts of
formaldehyde, acetaldehyde, propionaldehyde, and butyralde-
Tamura et al. showed that the equilibrium constant Keq for acid-
catalyzed hydrolysis of ꢀ-(anisidino)acrolein to MDA and anisidine
is 1.31 × 10-6 (1% EtOH/H2O, 25 °C), meaning that the reverse
reaction (imine formation from anisidine and MDA) is thermo-
dynamically highly favored.37 They also showed that the pKa value
for protonated ꢀ-(anisidino)acrolein is 1.26.38 Given the high
structural similarity between anisidine and 4-APC, these numbers
are directly applicable to our case. The estimated pKa of an
alkylimine such as 2 is 3.3,32 meaning that at pH 5.7 sufficient
amounts of protonated 2 are present to warrant a fast reduction
by the excess NaBH3CN to adduct 3.30 In contrast, imine 4 is
more stable and likely has a pKa ≈ 1.3. As a result, no appreciable
reduction of 4 to 5 takes place at pH 5.7.30 These combined data
clearly indicate that reaching equilibrium in our reaction of MDA
with 4-APC means that virtually all MDA will be in the stable imine
form (4) and that this does not appreciably react further in a
reduction, which is supported by the experiments shown in Figure
3.
(37) Ono, M.; Ohtsuka, S.; Tamura, S. Chem. Pharm. Bull. 1983, 31, 3129–
3138
(38) Tamura, S.; Ono, M.; Furuyama, K. Chem. Pharm. Bull. 1980, 28, 2356–
2366
.
.
9048 Analytical Chemistry, Vol. 80, No. 23, December 1, 2008

Table 1. Linear Dynamic Range, Correlation Coefficients (r2), Limit of Detection (LODs), and Characteristic Ions of
LC-MS/MS of 4-APC Derivatives of Aldehydes
aldehydes
linear range (nM)
r2
LOD (nM)b
M+
[M+-59.07]
tR RSD (%)
area RSD (%)a
malondialdehyde
trans-2-pentenal
pentanal
10-500
100-500
10-500
10-500
10-500
10-500
10-500
10-500
10-500
80
0.9998
0.9957
0.9998
0.9968
0.9983
0.9972
0.9928
0.9999
0.9994
3
33
3
249.2
263.2
265.2
279.3
293.3
291.3
307.3
321.3
335.4
208.1
190.1
204.1
206.1
220.2
234.2
232.2
248.2
262.2
276.3
149
0.71
1.04
0.67
0.39
0.29
0.40
0.18
0.11
0.08
0.70
5.23
18.72
8.72
5.32
2.51
2.12
5.48
4.96
4.68
4.43
hexanal
3
heptanal
cylcohexylcarboxaldehyde
octanal
nonanal
decanal
benzoylcholine (IS)
3
3
3
3
3
1.5
a RSDs are average of the calibration series. b Defined as 3 times the signal-to-noise ratio.
Figure 6. Tandem MS spectra of 50 nM cyclohexylcarboxaldehyde, heptanal, decanal, and malondialdehyde derivative standard.
hyde are present in methanol and formic acid used as reagents,
these aldehydes were excluded in this study.
and malondialdehyde derivatives are also presented in Figure 5.
The derivatives show ions with m/z 208, 293, 321, and 249,
respectively. The intense ion with m/z 195 is due to the 4-APC
derivatization reagent. Table 1 shows the linearity data of the nine
derivatized aldehyde standards in buffer. All aldehydes show good
linearity over almost 3 orders of magnitude. Limit of detection,
defined as 3 times the signal-to-noise ratio, of these derivatives
were determined at 3 nM, except for trans-2-pentenal, which was
33 nM. The RSD of the retention times were all within 1.05%,
which indicates a robust separation system. The average RSD of
Determination of Aldehyde Derivatives. From the in-vial
derivatization reaction mixture, a 10-µL injection is performed onto
the C18 column of the LC system. Figure 5 shows a combined
extracted ion chromatogram of the nine aldehyde derivatives of
trans-2-pentenal, pentanal, hexanal, heptanal, cyclohexylcarbox-
aldehyde, octanal, nonanal, decanal, malondialdehyde and IS,
benzoylcholine. All derivatized aldehydes examined showed an
M+ ion. The MS spectra of benzoylcholine, heptanal, nonanal,
Analytical Chemistry, Vol. 80, No. 23, December 1, 2008 9049

Figure 7. (A) Pooled urine spiked with 10 nM of aldehydes, (B) 10 nM aldehyde standard. 1, Pentanal; 2, hexanal; 3, heptanal; 4, octanal; and
5, decanal.
the peak areas (n ) 3) were within 5.5% for all compounds except
for pentanal and trans-2-pentenal, which were 8.72 and 18.72%,
respectively. The larger variation can be explained by the lower
concentration range. With these detection limits, aldehydes can
be determined in biological samples.12,39
ionization suppression of the signal but the linearity (R2 >0.99)
for all the tested aldehydes was not affected. Benzoylcholine
concentration was kept constant in all samples; its response did
not change in time and showed a signal loss of 10% by the
ionization suppression in urine. Because the matrix effects were
not constant for all the aldehydes and different from the IS,
standard addition for each aldehydes was used to determine the
concentration in urine. The extracted ion chromatograms in
Figure 7 represent pentanal, hexanal, heptanal, octanal, and
decanal with a concentration of 10 nM spiked in urine (A) and 10
nM in buffer (B). All selected aldehydes are detected in both urine
and buffer sample. Pentanal and hexanal were detected at elevated
concentration in the pooled urine samples, i.e., 61 ± 10 nM for
pentanal and 26 ± 2 nM for hexanal. Konidari et al.40 also reported
pentanal as biomarker of oxidative stress in human urine. The
peak heights for pentanal and hexanal are 80.2 and 30.1,
respectively, in spiked urine, and 25.2 and 22.1 in the aldehyde
standard.
An important additional feature of tandem MS is the ability to
provide analyte identification or to confirm its identity. Tandem
MS experiments on aldehyde-4-APC derivatives show that typical
neutral losses of 59 and 87 Da, consistent with losses of C3H9N
and C5H13N, respectively, are observed for all derivatives. All
aldehyde derivatives show the two abundant fragment ions at a
collision voltage of 20 V. Figure 6 show the MS/MS spectra of
the 50 nM derivatized cyclohexylcarboxaldehyde, heptanal, de-
canal, and malondialdehyde, providing the abundant fragment ions
with m/z 204.1 and 232.2 for cyclohexylcarboxaldehyde (selected
precursor ion with m/z 291.2), with m/z 206.1 and 234.2 for
heptanal (m/z 293.2), with m/z 248.2 and 276.2 for decanal (m/z
335.3), and with m/z 162.1 and 190.1 for malondialdehyde (m/z
249.2). Although the losses of [M+ - 59] and [M+ - 87] are
observed for all 4-APC-derivatized aldehydes, the abundance ratio
of the fragment ions changes with increasing aliphatic chain
length. Increasing collision energies could be used, but in this
case, we have chosen for a fixed collision energy. Also, the [M+
- 59] gave excellent linearity of R2 > 0.99 for all the tested
derivatized aldehydes.
Furthermore, malondialdehyde was detected in the derivatized
urine and was conformed by the retention time of derivatized
malondialdehyde standard (data shown in Supporting Information
Figure S6).
Again, for identification, the specific losses of [M+ - 59] and
[M+ - 87] in tandem MS can be used to confirm identity of the
aldehyde derivatives in urine, where a pooled urine sample was
spiked with 100 nM hexanal, cyclohexylcarboxaldehyde, heptanal,
octanal, and decanal, and all MS/MS spectra (Supporting Informa-
tion Figure S7) showed these typical losses.
Detection of Aldehyde Derivatives in Biological Samples.
To determine the aldehyde concentration in biological samples,
standard addition was applied. Comparing the calibration curves
in buffer and in urine, the derivatives show matrix-induced
(39) Orhan, H.; Van Holland, B.; Krab, B.; Moeken, J.; Vermeulen, N. P. E.;
(40) Konidari, C. N.; Giannopoulos, T. S.; Nanos, C. G.; Stalikas, C. D. Anal.
Hollander, P.; Meerman, J. H. N. Free Radical Res. 2004, 38, 1269–1279
.
Biochem. 2005, 338, 62–70.
9050 Analytical Chemistry, Vol. 80, No. 23, December 1, 2008

CONCLUSION
sufficient to determine aldehydes in urine. In future studies, the
possibilities of isotopically labeled internal standards or fast and
simple SPE-based sample pretreatment will be explored to correct
for the ionization suppression effects. Future efforts will also be
directed at structural refinement of 4-APC as well as at the use of
this type of reagent in the quantification or screening of aldehydes
in large sets of biological samples for lipid peroxidation-related
biomarkers.
We successfully designed and prepared a novel derivatization
reagent, 4-APC. The key advantage of 4-APC is that it allows highly
sensitive and selective MS detection of aldehydes. The derivati-
zation reaction is a fast and convenient one-pot reaction at a mild
pH of 5.7. As a result, no additional extraction steps are necessary
before analysis with an LC-MS/MS system for the quantification
of aldehydes as biomarkers for lipid peroxidation in urine at low-
nanomolar range. As a result of the chosen design of the
derivatization agent, tandem MS can be used to confirm identity
of the aldehydes based on the typical neutral losses of 59 and 87
Da from the aldehyde derivative and to screen for “unknown”
aldehydes from lipid peroxidation in urine samples. Although
some of the aldehydes show a 2-4 times response loss due to
matrix effect, quantification was still possible with the use of
standard addition. The quantification levels of this method are
SUPPORTING INFORMATION AVAILABLE
Additional information as noted in text. This material is
available free of charge via the Internet at http://pubs.acs.org.
Received for review July 10, 2008. Accepted August 25,
2008.
AC801429W
Analytical Chemistry, Vol. 80, No. 23, December 1, 2008 9051