Highly sensitive fluorescent method for the detection of cholesterol aldehydes formed by ozone and singlet molecular oxygen

Article abstract of DOI:10.1021/ac1006427

Cholesterol oxidation gives rise to a mixture of oxidized products. Different types of products are generated according to the reactive species being involved. Recently, attention has been focused on two cholesterol aldehydes, 3β-hydroxy-5β-hydroxy-B-norcholestane-6β- carboxyaldehyde (1a) and 3β-hydroxy-5-oxo-5,6-secocholestan-6-al (1b). These aldehydes can be generated by ozone-, as well as by singlet molecular oxygen-mediated cholesterol oxidation. It has been suggested that 1b is preferentially formed by ozone and 1a is preferentially formed by singlet molecular oxygen. In this study we describe the use of 1-pyrenebutyric hydrazine (PBH) as a fluorescent probe for the detection of cholesterol aldehydes. The formation of the fluorescent adduct between 1a with PBH was confirmed by HPLC-MS/MS. The fluorescence spectra of PBH did not change upon binding to the aldehyde. Moreover, the derivatization was also effective in the absence of an acidified medium, which is critical to avoid the formation of cholesterol aldehydes through Hock cleavage of 5α-hydroperoxycholesterol. In conclusion, PBH can be used as an efficient fluorescent probe for the detection/quantification of cholesterol aldehydes in biological samples. Its analysis by HPLC coupled to a fluorescent detector provides a sensitive and specific way to quantify cholesterol aldehydes in the low femtomol range.

Full text of DOI:10.1021/ac1006427

Anal. Chem. 2010, 82, 6775–6781
Highly Sensitive Fluorescent Method for the
Detection of Cholesterol Aldehydes Formed by
Ozone and Singlet Molecular Oxygen
Fernando V. Mansano, Rafaella M. A. Kazaoka, Graziella E. Ronsein, Fernanda M. Prado,
Thiago C. Genaro-Mattos, Miriam Uemi, Paolo Di Mascio, and Sayuri Miyamoto*
Departamento de Bioqu´ımica, Instituto de Qu´ımica, Universidade de Sa˜o Paulo,
CP26077, CEP 05513-970, Sa˜o Paulo, SP, Brazil
Cholesterol oxidation gives rise to a mixture of oxidized
products. Different types of products are generated ac-
cording to the reactive species being involved. Recently,
attention has been focused on two cholesterol aldehydes,
3ꢀ-hydroxy-5ꢀ-hydroxy-B-norcholestane-6ꢀ-carboxyalde-
hyde (1a) and 3ꢀ-hydroxy-5-oxo-5,6-secocholestan-6-al
(1b). These aldehydes can be generated by ozone-, as well
as by singlet molecular oxygen-mediated cholesterol oxi-
dation. It has been suggested that 1b is preferentially
formed by ozone and 1a is preferentially formed by singlet
molecular oxygen. In this study we describe the use of
1-pyrenebutyric hydrazine (PBH) as a fluorescent probe
for the detection of cholesterol aldehydes. The formation
of the fluorescent adduct between 1a with PBH was
confirmed by HPLC-MS/MS. The fluorescence spectra of
PBH did not change upon binding to the aldehyde.
Moreover, the derivatization was also effective in the
absence of an acidified medium, which is critical to avoid
the formation of cholesterol aldehydes through Hock
cleavage of 5r-hydroperoxycholesterol. In conclusion,
PBH can be used as an efficient fluorescent probe for the
detection/quantification of cholesterol aldehydes in bio-
logical samples. Its analysis by HPLC coupled to a
fluorescent detector provides a sensitive and specific way
to quantify cholesterol aldehydes in the low femtomol
range.
vascular diseases.2-5 Recently, attention has been focused on
cholesterol aldehydes that can be formed by the oxidation of
cholesterol by ozone8 and singlet molecular oxygen.9,10
The ozonation of cholesterol produces several oxidized prod-
ucts, in special two aldehydes (Figure 1), the cholesterol 5,6-
secosterols, 3ꢀ-hydroxy-5ꢀ-hydroxy-B-norcholestane-6ꢀ-carboxy-
aldehyde (1a) and 3ꢀ-hydroxy-5-oxo-5,6-secocholestan-6-al (1b).11,12
Wentworth and co-workers showed the presence of 1a and 1b
in atherosclerotic plaques8 and LDL oxidized with several oxi-
dants.13 These aldehydes have been also detected in neurode-
generative diseases, like Lewy body dementia14 and Alzheimer
disease.15 The role of 1a and 1b in the pathogenesis of
cardiovascular and neurodegenerative diseases has been inves-
tigated. In vitro studies have shown that cholesterol aldehydes
can covalently modify proteins, as well as, accelerate their
aggregation, as in the case of amyloid ꢀ-peptide formation15-17
and R-synuclein14. Further studies have shown that covalent
modification of apo-B by 1b causes this protein to misfold,
rendering the LDL particle more susceptible to macrophage
uptake.18 Moreover, some reports have also shown that
cholesterol aldehydes can induce apoptosis in macrophages
,20
and cardiomyoblasts.19 The induction of apoptosis in cardi-
(8) Wentworth, P., Jr.; Nieva, J.; Takeuchi, C.; Galve, R.; Wentworth, A. D.;
Dilley, R. B.; DeLaria, G. A.; Saven, A.; Babior, B. M.; Janda, K. D.;
Eschenmoser, A.; Lerner, R. A. Science 2003, 302, 1053–1056
(9) Brinkhorst, J.; Nara, S. J.; Pratt, D. A. J. Am. Chem. Soc. 2008, 130, 12224–
12225
(10) Uemi, M.; Ronsein, G. E.; Miyamoto, S.; Medeiros, M. H. G.; Di Mascio,
P. Chem. Res. Toxicol. 2009, 22, 875–884
(11) Gumulka, J.; Smith, L. L. J. Am. Chem. Soc. 1983, 105, 1972–1979
(12) Jaworski, K.; Smith, L. L. J. Org. Chem. 1988, 53, 545–554
(13) Wentworth, A. D.; Song, B. D.; Nieva, J.; Shafton, A.; Tripurenani, S.;
Wentworth, P. Chem. Commun. 2009, 3098–3100
(14) Bosco, D. A.; Fowler, D. M.; Zhang, Q.; Nieva, J.; Powers, E. T.; Wentworth,
P.; Lerner, R. A.; Kelly, J. W. Nat. Chem. Biol. 2006, 2, 249–253
.
.
Cholesterol (cholest-5-en-3ꢀ-ol) is a neutral lipid found in the
cellular membranes of mammals.1 Cholesterol is susceptible to
oxidation mediated by enzymatic and nonenzymatic mechanisms.2-5
The nonenzymatic oxidation can be mediated by reactive oxygen
species.2 Several oxidized products of cholesterol have been char-
acterized including hydroperoxides, epoxides, and aldehydes.2,6,7
These oxysterols have been detected in biological tissues and their
formation has been associated to neurodegenerative and cardio-
.
.
.
.
.
(15) Zhang, Q.; Powers, E. T.; Nieva, J.; Huff, M. E.; Dendle, M. A.; Bieschke,
J.; Glabe, C. G.; Eschenmoser, A.; Wentworth, P., Jr.; Lerner, R. A.; Kelly,
J. W. Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 4752–4757
(16) Scheinost, J. C.; Wang, H.; Boldt, G. E.; Offer, J.; Wentworth, P. Alzheimer
Dis. 2008, 47, 3919–3922
(17) Usui, K.; Hulleman, J. D.; Paulsson, J. F.; Siegel, S. J.; Powers, E. T.; Kelly,
J. W. Proc. Natl. Acad. Sci. U.S.A. 2009, 106, 18563–18568
(18) Takeuchi, C.; Galve, R.; Nieva, J.; Witter, D. P.; Wentworth, A. D.; Troseth,
R. P.; Lerner, R. A.; Wentworth, P. Biochemistry 2006, 45, 7162–7170
(19) Sathishkumar, K.; Gao, X.; Raghavamenon, A. C.; Parinandi, N.; Pryor, W. A.;
Uppu, R. M. Free Radical Biol. Med. 2009, 47, 548–558
(20) Gao, X.; Raghavamenon, A. C.; D’Auvergne, O.; Uppu, R. M. Biochem.
Biophys. Res. Commun. 2009, 389, 382–387
.
* To whom correspondence should be addressed. Phone: (55) (11) 3091-
3810 (x261). Fax: (55) (11) 3815-5579. E-mail: miyamoto@iq.usp.br.
.
(1) Maxfield, F. R.; Tabas, I. Nature 2005, 438, 612–621
(2) Brown, A. J.; Jessup, W. Atherosclerosis 1998, 142, 1–28
(3) Bjo¨rkhem, I.; Diczfalusy, U. Arterioscler. Thromb. Vasc. Biol. 2002, 22,
734–742
(4) Garenc, C.; Julien, P.; Levy, E. Free Radical Res. 2010, 44, 47–73
(5) Schroepfer, G. J., Jr. Physiol. Rev. 2000, 80, 361–554
(6) Smith, L. L. Lipids 1996, 31, 453–487
.
.
.
.
.
.
.
.
.
(7) Girotti, A. W. J. Photochem. Photobiol. B 1992, 13, 105–118
.
.
10.1021/ac1006427  2010 American Chemical Society
Published on Web 07/21/2010
Analytical Chemistry, Vol. 82, No. 16, August 15, 2010 6775

Figure 1. Structures of cholesterol carboxyaldehyde (1a), cholesterol secocholestanal (1b) and their corresponding fluorescent adducts formed
upon derivatization with 1-pyrenebutyric hydrazide (2a and 2b).
omyoblasts was associated with the induction of ROS generation
and the activation of extrinsic and intrinsic pathway.19
This manuscript describes a new fluorescence-based method
to detect and quantify cholesterol aldehydes using the probe
1-pyrenebytiric hydrazine (PBH) (Figure 1). This method proved
to be highly sensitive and has the advantage of minimizing the
effect of interfering compounds and not requiring acidic media.
The use of acid conditions showed to cause an overestimation of
1a in a sample containing cholesterol 5R-OOH. Using PBH as
the fluorescent probe we could detect 1a in the low femtomolar
range by HPLC coupled to fluorescence detector. Moreover, this
methodology allowed detecting and calculating the ratio of 1a and
1b formed by the oxidation of cholesterol exposed to singlet
molecular oxygen and ozone, respectively.
Besides ozone, recent studies have evidenced the generation
of cholesterol aldehydes in the reaction of cholesterol with singlet
molecular oxygen. The proposed mechanisms involve Hock-
cleavage of cholesterol 5R-hydroperoxide (5R-OOH) in the pres-
ence of acids9 or cholesterol dioxetane decomposition.10 In both
cases, 1a was detected as the major product. Supporting these
studies, Tomono et al.21 have also detected 1a as the major
product in the incubation of cholesterol with human myeloper-
oxidase in the presence H2O2 and chloride ions, a system that
is suggested to generate singlet molecular oxygen in activated
neutrophils.22
EXPERIMENTAL SECTION
Several methods have been used for the determination of
cholesterol aldehydes, including mass spectrometry,10 UV-
visible,8,13,23 and fluorescence detection,14,21 most of them coupled
to liquid chromatography. Although mass spectrometry based
methods have shown a good sensitivity, this technique requires
specialized people and expensive equipment. UV-visible detection
methods using 2,4-dinitrophenylhydrazine (DNPH) have been
widely used to detect and quantify aldehydes derived from lipid
peroxidation. However, this method is usually carried out in a
strong acidic media, which is able to induce the cleavage of
cholesterol 5R-OOH to form 1a and 1b,9 therefore leading to an
overestimation of the real aldehyde concentration in biological
samples. An alternative method used to detect aldehydes involves
their reaction with fluorescent probes and the detection of the
fluorescent adducts formed. The fluorescent methods have been
used to increase detection limits and exclude interferences. The
fluorescent probe described in the literature for cholesterol 5,6-
secosterols detection is dansyl hydrazine. Bosco and co-workers
used this probe in acidic media to detect cholesterol aldehydes
in brain.14
Reagents. Cholesterol, silica gel (200-400 mesh, 60 Å),
1-pyrenebutyric hydrazide (PBH), deuterated chloroform (CDCl3),
sodium phosphate monobasic and dibasic were purchased from
Sigma (St. Louis, MO). Methylene blue was purchased from
Merck (Rio de Janeiro, Brazil). All the other solvents were of
HPLC grade and were acquired from Mallinckrodt Baker
(Phillipsburg, NJ). The water used in the experiments was
treated with the Nanopure Water System (Barnstead, Dubuque,
IA).
Synthesis of 3ꢀ-Hydroxy-5ꢀ-hydroxy-B-norcholestane-6ꢀ-
carboxyaldehyde (1a). 1a was synthesized by photooxidation
of cholesterol in the presence of methylene blue as described
previously.10 Briefly, 200 mg of cholesterol dissolved in 20 mL of
chloroform was mixed with 250 µL of methylene blue (10 mM in
methanol). This solution was cooled at 4 °C and irradiated under
continuous agitation using two tungsten lamps (500 W) during
2.5 h. The formation of cholesterol oxidation products were
checked by thin-layer chromatography using ethyl acetate and
isooctane (1:1, v/v) as the eluent. 1a was purified from the
reaction mixture by flash column chromatography, using silica
gel and a gradient of hexane and ethyl eter. The purified 1a was
analyzed and quantified in CDCl3 by NMR spectroscopy using
DRX500 instrument, AVANCE series (Bruker-Biospin, Rhein-
stetten, Germany) as described previously.10 For the quantifica-
(21) Tomono, S.; Miyoshi, N.; Sato, K.; Ohba, Y.; Ohshima, H. Biochem. Biophys.
Res. Commun. 2009, 383, 222–227
(22) Kiryu, C.; Makiuchi, M.; Miyazaki, J.; Fujinaga, T.; Kakinuma, K. FEBS
Lett. 1999, 443, 154–158
.
.
(23) Wang, K. Y.; Bermudez, E.; Pryor, W. A. Steroids 1993, 58, 225–229
.
6776 Analytical Chemistry, Vol. 82, No. 16, August 15, 2010

tion, 2 µL of proprionaldehyde were added as the internal
standard into 750 µL of the 1a solution in CDCl3. The 1H signal
corresponding to the aldehyde group of the internal standard
and 1a appeared at 9.73 and 9.62 ppm, respectively. The relative
signals of 1H of aldehydes were integrated by Mestre C
software (Figure S1 in the Supporting Information (SI)).
Synthesis of 3ꢀ-Hydroxy-5-oxo-5,6-secocholestan-6-al (1b).
1b was prepared by ozonation of cholesterol as described by
Wentworth et al.8 and Wang et al.24 Ozone was produced at a
rate of 100 mg/h by passing pure oxygen through AquaZone
PLUS 200 instrument (Red Sea Fish Pharm. Ltd., Houston, TX).
Oxygen flow rate was set to 10 mL/min. A solution of 10 mg/mL
of cholesterol in chloroform was cooled at dry ice temperature
and oxidized by bubbling ozone for 5 min. After oxidation,
chloroform was evaporated with nitrogen gas and the residue was
stirred for 2 h at room temperature with Zn powder (19.4 mg) in
water-acetic acid (1:19 v/v; 3.1 mL). Dichloromethane (15 mL)
was added and the mixture was washed five times with deionized
water (5 × 15 mL). The organic phase was evaporated to dryness
in vacuo. The residue was dissolved in isopropyl alcohol and kept
at -80 °C for further analysis. The characterization of 1b was
performed by NMR spectroscopy using a DRX500 instrument,
AVANCE series (Bruker-Biospin, Rheinstetten, Germany) operat-
ing at 11.7 T. Cholesterol ozonation followed by Zn reduction in
acetic acid gave 1b as a major aldehyde, as confirmed by 1H NMR
analysis (δ, ppm, CDCl3): δ 9.607 (s, J ) 0.5 Hz, 1H, CHO),
4.466 (s, 1H, H-3), 3.097 (dd, J ) 13.5, 4.0 Hz, 1H, H-4e), 1.010
(s, 3H, CH3-19), 0.671 (s, 3H, CH3-18) (Figure S2 in the SI).
Derivatization of Cholesterol Aldehydes with PBH. The
optimal conditions for the derivatization of cholesterol aldehydes
were established using the purified 1a. Time-course of PBH
adduct formation with 1a was monitored by incubating 100 µM
1a with 2 mM PBH in isopropanol at 37 °C under continuous
agitation for up to 8 h. Aliquots of 1 µL of the reaction mixture
were taken at specified times and analyzed by HPLC coupled to
fluorescence detector. The ideal concentration of the probe was
determined by incubating 1 µM of 1a in the presence of 1-1000
µM of PBH in isopropanol at 37 °C under continuous agitation
for 6 h. The effect of the pH in the formation of 2a was analyzed
by incubating 1 µM of 1a with 50 µM PBH in a reaction system
consisted of isopropanol:10 mM phosphate buffer at pH 5.7, 7.4,
and 8.0 (90:10, vol/vol). Quantitative analysis of 1a was done by
incubating an aliquot of sample with 600 µM of the probe in
isopropanol containing 1 mM phosphate buffer at pH 7.4 (neutral
condition) or 0.1 mM HCl (acid condition) at 37 °C for 6 h.
Analysis of the Fluorescent Adduct by HPLC Coupled
with Fluorescence Detector. HPLC analysis was carried out on
a Shimadzu Prominence system (Tokyo, Japan), consisted of LC-
20AT pumps, SIL-20AV autosampler, RF-10Axl fluorescence detec-
tor, and a CBM-20A controller. One microliter of the sample was
injected into a reversed-phase column Synergi C18 (50 × 4.6 mm,
2.5 µm particle size, Phenomenex, Torrance, CA). The HPLC
mobile phase consisted of water (A) and methanol (B), and the
flow rate was 1 mL/min. The separation of the fluorescent adducts
was done using the following condition: 86% B for 5 min, 86-92%
in 1 min, 92% B for 15 min, and 92-86% B in 1 min. The excitation
and emission wavelengths were fixed at 339 and 380 nm,
respectively.25 For the analysis 1 µL was injected through the
autosampler. Data were acquired at high sensitivity and gain of
1× in the fluorescence detector. Fluorescence spectra were
acquired by the fluorescence detector RF-551 (Shimadzu, Japan).
HPLC data were processed using the LC solution software
(Shimadzu, Japan).
Analysis of the Fluorescent Adducts by HPLC-MS/MS.
The PBH fluorescent adducts formed with 1a and 1b were
analyzed by a HPLC system connected to a UV-visible detector
(SPD 10 AVVP, Shimadzu, Kyoto, Japan) and a Quattro II triple
quadrupole tandem mass spectrometer (Micromass, Manchester,
UK) (HPLC-MS/MS). HPLC separation was carried out as
described above. The eluent from the column was monitored at
342 nm and 10% of the HPLC flow rate was directed into the mass
spectrometer. The fluorescent adducts were detected using
electrospray ionization (ESI) in the positive ion mode. The source
and desolvation temperature of the mass spectrometer were set
at 100 and 200 °C, respectively. The cone voltage was set to 50 V,
the extractor cone voltage was set to 5 V and the capillary and
the electrode potentials were set at 4.5 and 0.5 kV, respectively.
Collision energy was set at 20 eV for 2a and 30 eV for 2b. Full-
scan data was acquired over a mass range of 100-900 m/z. Data
was processed by means of the MassLynx NT software.
Cholesterol Oxidation Induced by Photooxidation, Ozone
and HOCl/H2O2. Cholesterol (10 mg/mL in chloroform) was
photooxidized in the presence of methylene blue for 2.5 h. An
aliquot of this sample was taken for cholesterol aldehyde deter-
mination. The ozonation of cholesterol (10 mg/mL in chloroform)
was conducted essentially as described for the synthesis of 1b.
After 5 min ozonation, chloroform was evaporated by nitrogen
gas and the residue was dissolved in isopropanol. An aliquot of
this sample was diluted and used for cholesterol aldehyde
determination. The oxidation of cholesterol by HOCl/H2O2
system was conducted by incubating a 10 mM cholesterol
solution containing 1% ethanol in 10 mM phosphate buffer at
pH 7.4 with 1 mM HOCl and H2O2, under continuous agitation
at 37 °C for 5 min.
Method Validation. The limit of quantification (LOQ) was
established as the amount of adduct formed that generated a signal
6-fold higher than baseline. The limit of detection (LOD) was
established as the amount of adduct formed that generated a signal
3-fold higher than baseline. The reproducibility was checked by
intra- and interday analysis. Interday analyses were conducted on
three consecutive days.
RESULTS
Synthesis and Characterization of 1a. Cholesterol carboxy-
aldehyde (1a) was synthesized by photooxidation of cholesterol.10
After the reaction, the oxidized products were purified by flash
column chromatography and 1a was isolated. The identity of 1a
was confirmed by NMR spectroscopy as described by Uemi et
al.10 The 1H NMR analysis showed a doublet peak at 9.62 ppm
consistent with the presence of the aldehyde group in 1a
structure (Figure S1 in the SI). 1a was quantified by 1H NMR
analysis using proprionaldehyde as the internal standard as
described in the experimental section.
Characterization of 2a Fluorescent Adduct. Using the
purified 1a sample, several experiments were conducted to
(24) Wang, K. Y.; Bermudez, E.; Pryor, W. A. Steroids 1993, 58, 225–229
.
(25) Morsel, J. T.; Schmiedl, D. Fresenius´ J. Anal. Chem. 1994, 349, 538–541.
Analytical Chemistry, Vol. 82, No. 16, August 15, 2010 6777

Figure 3. Fluorescence excitation and emission spectra of PBH and
2a. Both compounds showed the same excitation and emission
maximum wavelengths at 339 and 380 nm, respectively.
Figure 2. Analysis of 2a fluorescent adduct, formed in the reaction
of 1a with PBH by HPLC coupled to fluorescence detector. HPLC/
fluorescence chromatogram obtained for the analysis of 2a using
excitation at 339 nm and emission at 380 nm. For the derivatization,
5 pmol of 1a was incubated with PBH 100 µM in a final volume of
100 µL isopropanol at 37 °C for 6 h under agitation. Analysis was
done by injecting 1 µL of the reaction mixture into the HPLC. The
inset shows the time dependent formation of 2a during incubation of
100 µM of 1a in the presence of 2 mM of PBH for up to 8 h.
([M+H-PB]+, m/z 431), PBH and two protons ([M+H-PBH-
2H]+, m/z 398) and PBH and two water molecules
([M+H-PBH-2H2O]+, m/z 365), respectively. The fragment
ion at m/z 303 corresponds to the positively charged PBH ion.
Reproducibility and Stability. The intra- and interday repro-
ducibility of the assay procedure was determined by evaluating
three individual reactions containing 1a (1 pmol; 1 µM final
concentration) and PBH (50 µM). The relative standard deviation
for the intra- and interday reproducibility at this concentration was
1.7% and 7.6%, respectively.
The stability of 2a adduct was evaluated in triplicate reactions
of 1a (1 pmol; 1 µM final concentration) with PBH (50 µM). After
derivatization, these samples were kept in vials inside the
autosampler at 4 °C and the 2a peak was monitored up to 6 h.
There was no significant change in the peak area of the
fluorescence signal during this period.
Standard Curve, Limit of Detection and Limit of Quanti-
fication. A calibration curve was constructed using 5-100 fmol
(50-1000 nM) of 1a in triplicate reaction in a final volume of 100
µL. Peak area of 2a increased linearly over this concentration
range with R2 value of 0.9962 and Y ) 6.64 × 103 X + 1.65 ×
104. The limit of detection (LOD) and quantification (LOQ) of
1a was 10 fmol (10 nM) and 20 fmol (20 nM), respectively.
Application of PBH for Cholesterol Aldehyde Detection
in Oxidized Cholesterol Samples. The newly developed fluo-
rescence-based method was applied for the detection and quan-
tification of 1a in cholesterol samples oxidized by photooxidation
or ozone. Cholesterol photooxidation gave rise to a major peak at
12.5 min (Figure 5A). On the other hand, cholesterol ozonation
yielded an intense peak at 10.3 min as well as some other smaller
peaks, including the same peak at 12.5 min observed for the
photooxidation (Figure 5B). Peak assignment was done by
comparison of the retention times of the peaks with those obtained
for the standard samples of 1a (Figure 5C) and 1b (Figure 5D).
In this way, the peaks observed at 10.3 and 12.5 min were assigned
to the fluorescent adducts 2b and 2a, respectively. Additionally,
the identity of the peaks was confirmed by HPLC-MS/MS analysis
(Figure S4 in the SI). It should be noted that the overall retention
times for the analysis by HPLC-MS/MS were increased by almost
2 min compared to the analysis by HPLC/fluorescence detection.
This retention time delay was due to the differences in room
establish the optimal conditions for the reaction with PBH.
Incubations of 1a with PBH were conducted in isopropanol at 37
°C under continuous agitation. HPLC analysis using fluorescence
detection showed the appearance of an intense peak at 12.5 min,
corresponding to the fluorescent adduct 2a (Figure 2). The pH
effect on 2a adduct formation was analyzed. Incubations con-
ducted at pH 5.7, 7.4, and 8.5 showed that this adduct was
preferentially formed at pH 5.7 (Figure S3 in the SI), which is in
accordance with favorable formation of Schiff base adduct at
slightly acidic conditions.
A time-course analysis of 2a showed a time-dependent increase
of the peak area up to 4 h, reaching a plateau after this time (inset,
Figure 2). Based on this, the incubation time for the derivatization
was fixed to 6 h. The ideal concentration of PBH for the reaction
was established by incubating 1a (1 µM) in the presence of
1-1000 µM of PBH. The fluorescent adduct formation reached
its maximum with PBH concentration higher than 50 µM.
The fluorescence of 2a was characterized to establish the
optimal excitation and emission wavelengths. Fluorescence analy-
sis of 2a showed a spectrum similar to the original probe with
excitation and emission wavelengths maximum at 339 and 380
nm, respectively (Figure 3). Thus, indicating that adduct formation
does not alter the fluorescence spectra of the probe.
The formation of 2a was also confirmed by HPLC-MS/MS.
The mass spectrum of 2a acquired by ESI in the positive ion mode
showed peaks corresponding to 2a molecular ion ([M+H]+) and
its sodium adducts ([M+Na]+) at m/z 703 and 725, respectively
(Figure 4A and 4B). Two other peaks at m/z 685 and 667 were
also detected, corresponding to the loss of one ([M+H-H2O]+)
and two water molecules ([M+H-2H2O]+), respectively. These
ions were also observed in the MS/MS spectrum of m/z 703
(Figure 4C). The MS/MS spectrum also showed other fragment
ions at m/z 431, 398, 365, and 303. The first three, correspond to
the ions formed by the loss of pyrene butyric (PB) group
6778 Analytical Chemistry, Vol. 82, No. 16, August 15, 2010

Figure 4. Analysis of 2a by HPLC-MS/MS using ESI in the positive ion mode. The cone voltage and the collision energy were set to 50 V and
20 eV, respectively. Selected ion chromatogram of the ion at m/z 703 (A). Spectrum of the peak corresponding to 2a at 14 min (B). Fragment
ion spectrum of m/z 703 (C).
and the sodium adduct ([M+Na]+) of 2b (Figure 6B). Collision
induced dissociation of the ion at m/z 703, showed the same
fragment ions observed for 2a, suggesting that the peak corre-
sponds to its isomer, 2b (Figure 6C). Fragments ions observed
at m/z 685 and 667 are formed by the loss of one and two water
molecules from 2b, respectively. The ions at m/z 431, 398, 383,
and 365 are formed by the loss of PB, PBH and two protons, PBH
and one water, and PBH and two water molecules from 2b,
respectively (Figure 6C).
Aiming to get a quantitative data on cholesterol aldehyde
formation, a calibration curve constructed for 1a was used to
determine its concentration. For the quantification, oxidized
cholesterol samples were diluted to get a final cholesterol
concentration of 219 µM. The amounts of 1a detected in the
samples by PBH method at neutral pH were 19.1 4.0 µM and
0.7 0.3 µM for the photooxidation and ozonation, respectively.
This corresponds to a 1a yield of 8.7% for the photooxidation and
0.3% for the ozonation. For comparison, 1a quantification was also
done using dansyl hydrazine method in the presence of acid (0.1
mM HCl) (method in the SI). The concentrations of 1a deter-
mined by this method were 35.2 0.7 µM and 1.0 0.1 µM in the
photooxidation and ozonation, respectively. This provides 1a
yields of 16% for the photooxidation and 0.4% for the ozonation.
As can be clearly noticed, the 1a yield determined by the dansyl
hydrazine method was overestimated by almost 2-folds for the
photooxidation. The same result was obtained when PBH deriva-
tization was conducted in the presence of acid. Under this
Figure 5. Analysis of cholesterol aldehydes formed in the photo-
oxidation (A) and ozonation (B) of cholesterol using PBH. An aliquot
of the sample was reacted with PBH (600 µM) for 6 h and analyzed
by HPLC coupled to fluorescence detector. For peak assignment
standard samples containing 1a (C) or 1b (D) were also reacted with
PBH and analyzed.
temperatures, which was about 5 °C lower in the case of HPLC-
MS/MS analysis.
The major fluorescent adduct detected in cholesterol ozonation
was further characterized by HPLC-MS/MS analysis (Figure 6).
Selected ion mass chromatogram of the ion at m/z 703 showed a
single peak at 12 min. (Figure 6A). Mass spectrum for this peak
showed two major ions, one at m/z 703 and other at m/z 725
condition, the 1a concentration was 32.6
0.7 µM, which
corresponds to a 1a yield of 15% in the photooxidation. These
results are consistent with the fact that photooxidized cholesterol
(Figure 6B), which corresponds to the molecular ion ([M+H]+
)
Analytical Chemistry, Vol. 82, No. 16, August 15, 2010 6779

Figure 6. Analysis of 2b by HPLC-MS/MS using ESI in the positive ion mode. The cone voltage was set to 50 V and the collision energy was
set to 30 eV. Selected ion chromatogram for the ion at m/z 703 (A). Mass spectrum of the peak corresponding to 2b at 12 min (B). Fragment
ion spectrum of m/z 703 (C).
samples contain cholesterol 5R-OOH, which in the presence of
acids is easily converted to 1a.9
Moreover, the sensitivity of PBH and dansyl hydrazine
methods were also compared (Figure S5 in the SI). PBH method
conducted at neutral pH was almost 10 times more sensitive than
dansyl hydrazine method. This difference was even larger when
PBH reaction was carried out with acid, reaching a 90 times higher
sensitivity.
PBH was also used for the quantification of ChAld formed
during cholesterol oxidation promoted by the HOCl/H2O2 reac-
tion system. This is a biologically relevant reaction system that
is known to generate stoichiometric amounts of singlet mo-
lecular oxygen. The formation of 2a was analyzed by HPLC
coupled with fluorescence detector (Figure 7). The complete
reaction system containing cholesterol and HOCl/H2O2 showed
the appearance of an intense fluorescent peak corresponding
to 2a. The fluorescent adduct was quantified using the
calibration curve and a value of 0.2 µM of 1a was found. This
corresponds to a yield of 0.2%.
Figure 7. Analysis of cholesterol aldehydes generated by the
reaction of cholesterol with H2O2 and HOCl using PBH. HPLC/
fluorescence chromatograms obtained for cholesterol (A), and cho-
lesterol incubated with H2O2 (B), HOCl (C), and H2O2 + HOCl (D).
Reactions were carried out with 100 µM cholesterol in the presence
of 1 mM HOCl, 1 mM H2O2 and 100 µM PBH for 6 h at 37 °C under
continuous agitation.
DISCUSSION
Cholesterol is oxidized in the presence of reactive oxygen
species generating aldehydes, hydroperoxides and epoxides.6,7
Recently, two cholesterol aldehydes have attracted attention, the
cholesterol carboxyaldehyde (1a) and cholesterol secoaldehyde
(1b). The formation of 1a and 1b was first described in the
oxidation of cholesterol with ozone.11,12 Wentworth and co-workers
reported the presence of 1a and 1b in atherosclerotic plaques
and related the detection of these aldehydes as an evidence for
ozone generation in human tissues.8 On the other hand, two recent
studies identified the generation of 1a in the reaction of choles-
terol with singlet molecular oxygen.9,10 Indeed, Brinkhorst and
co-workers reported the formation of 1a from cholesterol 5R-
hydroperoxide by Hock cleavage in acid media9 and Uemi and
co-workers reported the formation of 1a in the reaction of
cholesterol with singlet molecular oxygen generated either by
photooxidation or by the thermodecomposition of endoperox-
ides.10
In this study, we have developed a new detection method for
cholesterol carboxyaldehyde, 1a, using the fluorescent probe
PBH, which contains the highly fluorescent pyrene group. The
6780 Analytical Chemistry, Vol. 82, No. 16, August 15, 2010

reaction of 1a with PBH can be conducted in the absence of acids,
which is important to avoid the formation of 1a from acid catalyzed
cholesterol 5R-OOH decomposition.9 Bosco and co-workers14 used
dansyl hydrazine as a fluorescent label to detected 1a and 1b in
a brain tissue sample from Lewy body disease. However, the
derivatization of cholesterol aldehydes with dansyl hydrazine is
conducted in the presence of strong acids, such as, 5 mM sulfuric
acid.14 In fact, we have done a comparative analysis of the two
fluorescent probes (PBH and dansyl hydrazine) to detect 1a in
vitro and observed that the method using dansyl hydrazine can
overestimate the basal level of 1a, especially in samples containing
cholesterol 5R-OOH. The 1a yields in photooxidized cholesterol
samples estimated by PBH (neutral pH) and dansyl hydrazine
were 8.7% and 16.0%, respectively. This data indicates that the
derivatization in the absence of acid is a critical point that should
be considered when estimating 1a basal level in samples,
especially when they were oxidized by singlet molecular oxygen.
When comparing dansyl hydrazine and PBH methods, the latter
was approximately 10 times more sensitive for 1a detection
(Figure S5 in the SI). This difference was even larger when PBH
derivatization was carried out at acid condition, clearly evidencing
the superiority of PBH method compared to dansyl hydrazine in
terms of sensitivity.
A number of studies has detected the formation of cholesterol
aldehydes in vitro,9,10,21 as well as in vivo.8,14-16 In this study, we
have used PBH to analyze the formation of 1a during cholesterol
oxidation mediated by ozone and singlet molecular oxygen. The
reaction of cholesterol with ozone yielded 1a and 1b and also
other minor unidentified products. We could not quantify exactly
the amount of 1b due to the lack of an appropriate pure standard
for it. Nonetheless, the relative peak intensities suggest that
cholesterol ozonation yields 1b as a major product and a small
amount of its aldolization product (Figure 5). The estimated ratio
of 1a:1b in the ozonation was 1:10. Similar results were described
for the ozonation of human LDL, where a relative yield of 1:4 was
found.13 On the other hand, cholesterol oxidation promoted by
singlet molecular oxygen yielded 1a as the major product
consistent with the data described by Brinkhorst et al.9 and Uemi
et al.10 The estimated yield of 1a in the photooxidation of
cholesterol was approximately 8.7% and the ratio of 1a:1b was
165:1. These values are in agreement with those reported for the
photooxidation of human LDL using hematoporphyrin IX for 14 h
where only 1a was detected.13 The incubation of cholesterol in
the presence of H2O2 and HOCl also gave 1a as the major
product, consistent with the oxidation of cholesterol promoted
by singlet molecular oxygen. It is known that the reaction of
H2O2 with HOCl generate stoichiometric amounts of singlet
molecular oxygen26 and this reaction is suggested to play an
important role during inflammatory conditions.22 The predomi-
nant formation of 1a over 1b by this reaction was also reported
by Tomono and co-workers.21 They incubated cholesterol (100
µM) in the presence of H2O2 (100 µM) and NaOCl (100 µM)
and detected five times more 1a than 1b using dansyl
hydrazine as a fluorescent probe.21
CONCLUSIONS
We have developed a new sensitive method for the detection
of cholesterol aldehydes using PBH as a fluorescent probe. This
new methodology allows detecting and quantifying the relative
amounts of 1a and 1b with high sensitivity, which is important
to assess the relative contributions of ozone and singlet molecular
oxygen to the oxidation of cholesterol in biological systems.
Derivatization of cholesterol aldehydes using PBH can be con-
ducted without the addition of strong acids, which is important
for the determination of the basal level of cholesterol aldehydes
present in biological tissues from different sources. In conclusion,
the easy and reliable method developed in this study can help to
investigate the formation cholesterol aldehydes in biological
system, which is critical to clarify their relevance in disease
progression.
ACKNOWLEDGMENT
This work was supported by the Brazilian research funding
institutions FAPESP (Fundac¸˜ao de Amparo a` Pesquisa do Estado
de S˜ao Paulo), CNPq (Conselho Nacional para o Desenvolvimento
Cient´ıfico e Tecnolo´gico), INCT de Processos Redox em Biome-
dicina - Redoxoma, and Pro´-Reitoria de Pesquisa-USP.
SUPPORTING INFORMATION AVAILABLE
NMR analysis of 1a and 1b. HPLC-MS/MS detection of 2a
and 2b. HPLC-fluorescence data showing the pH effect on 2a
formation and the sensitivity of dansyl hydrazine and PBH
methods. Dansyl hydrazine derivatization method. This material
is available free of charge via the Internet at http://pubs.acs.org.
Received for review March 10, 2010. Accepted July 8,
2010.
(26) Khan, A. U.; Kasha, M. J. Am. Chem. Soc. 1970, 92, 3293–3300
.
AC1006427
Analytical Chemistry, Vol. 82, No. 16, August 15, 2010 6781