The ratio of cholesterol 5,6-secosterols formed from ozone and singlet oxygen offers insight into the oxidation of cholesterol in vivo

Article abstract of DOI:10.1039/b821584g

Ongoing efforts to unravel the origins of the cholesterol 5,6-secosterols (1a and 1b) in biological systems have revealed that the two known chemical routes to these oxysterols, ozonolysis of cholesterol (3) and Hock-cleavage of 5-α-hydroperoxycholesterol (4a), are distinguishable based upon the ratio of the hydrazone derivatives (2a and 2b) formed in each case and this ratio offers an insight into the chemical origin of the secosterols in vivo. The Royal Society of Chemistry 2009.

Full text of DOI:10.1039/b821584g

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
The ratio of cholesterol 5,6-secosterols formed from ozone and singlet
oxygen offers insight into the oxidation of cholesterol in vivow
Anita D. Wentworth,a Byeong-Doo Song,a Jorge Nieva,b Asher Shafton,a
Sangeetha Tripurenania and Paul Wentworth Jr.*ac
Received (in Cambridge, UK) 2nd December 2008, Accepted 11th March 2009
First published as an Advance Article on the web 16th April 2009
DOI: 10.1039/b821584g
Ongoing efforts to unravel the origins of the cholesterol
5,6-secosterols (1a and 1b) in biological systems have revealed
that the two known chemical routes to these oxysterols,
ozonolysis of cholesterol (3) and Hock-cleavage of 5-a-hydro-
peroxycholesterol (4a), are distinguishable based upon the ratio
of the hydrazone derivatives (2a and 2b) formed in each case and
this ratio offers an insight into the chemical origin of the
secosterols in vivo.
1b, analyzed as their 2,4-DNP hydrazones (2a and 2b), during
the oxidation of cholesterol (3) in biologically-relevant
formats—human low-density lipoprotein (hLDL); human
high-density lipoprotein (hHDL); and liposomes using
cellular, biological and chemical oxidant systems in aqueous
buffer (Table 1 and ESIw).
The 2,4-DNP hydrazone of atheronal-B (2b) is observed at
levels above untreated hLDL (B0.01% of 3), after oxidation
by superoxide anion (Table 1, entries 1 and 2), by singlet
oxygen generating systems (Table 1, entries 6 and 7), and by
activated human polymorphonuclear leukocytes (hPMNs,
Table 1, entries 8 and 9). In contrast, no elevation in 2b above
untreated samples is observed after hLDL treatment with
H2O2, HOCl or hydroxyl radical generating systems
(Table 1, entries 3–5). The generation of 2b during hLDL
oxidation by phorbol myristate acetate (PMA)-activated
hPMNs, is proportional to cell density, is consistently higher
in lipid extracts of hLDL rather than the PMN cell membranes
and is elevated in D2O relative to H2O-containing buffer.
These combined data suggest that the oxidant responsible
for the generation of 2b is both sufficiently long-lived to be
released into the medium and its life-time is prolongꢀed in
deuterated buffers, hinting at roles for both ꢀO2À/HO2 and
1O2 (Table 1, entries 8 and 9 and ESI,w Tables S1–S3).
The photosensitized and hPMN-mediated oxidation of
cholesterol in hHDL in aqueous buffer (PBS, pH 7.4) parallels
the observations with hLDL in that the 2,4-DNP hydrazone of
cholesterol secosterol 2b is elevated above untreated levels and
In a recent report, Pratt and co-workers1 have shown that
Hock-cleavage of 5a-hydroperoxycholesterol (4a), which can
arise from the singlet oxygen ene reaction with cholesterol (3),
occurs facilely under acidic conditions in organic solvents
leading to the formation of primarily cholesterol 5,6-secosterol
atheronal-B (1b). Atheronal-A (1a) is either not formed at all
or is a minor component in participating solvents such as
ethanol (Fig. 1).
We discovered the cholesterol 5,6-secosterols 1a and 1b
within human atherosclerotic plaque material in vivo2 and
surmised that, based upon the wealth of literature in the field
of chemical-, biological- and auto-oxidation of cholesterol at
the time,3,4 that only ozone was capable of generating 1a from
cholesterol (3).2 In fact, we considered the presence of 1a
[measured as 2a after extraction and derivatization with
2,4-dinitrophenyl (2,4-DNP) hydrazine] within inflammatory
arteries as indirect evidence that an oxidant with the chemical
signature of ozone may be generated during atherosclerosis
progression. We also showed that 1a undergoes an almost
instantaneous aldolization process to form 1b in whole blood
and therefore the cholesterol 5,6-secosterols 1a and 1b were
both potential signature molecules for cholesterol ozonolysis
in vivo.2
The oxysterols 1a and 1b are proatherogenic5 and induce
protein misfolding and amyloidogenesis6 in a number of
biologically-relevant proteins.2,7 Given the pathological effects
of 1a and 1b, we are striving to understand their chemical
origin in vivo. Herein we have studied the formation of 1a and
a Department of Chemistry and the Skaggs Institute for Chemical
Biology, The Scripps Research Institute, 10550 N. Torrey Pines Rd.,
La Jolla, CA 92037, USA. E-mail: paulw@scripps.edu;
Fax: +1 858 784 7385; Tel: +1 858 784 2576
b Department of Molecular and Experimental Medicine, The Scripps
Research Institute, John Jay Hopkins Dr., La Jolla, CA 92037, USA
c Department of Biochemistry, The University of Oxford,
South Parks Rd., Oxford, UK OX1 3QU
w Electronic supplementary information (ESI) available: All experi-
mental methods for chemical synthesis, 1H NMR analyses and
biological assays. See DOI: 10.1039/b821584g
Fig. 1 Chemical routes from cholesterol 3 into the cholesterol
secosterols 1a and 1b are being sought.
ꢁc
This journal is The Royal Society of Chemistry 2009
3098 | Chem. Commun., 2009, 3098–3100

View Article Online
Table 1 Oxidation of hLDL-bound cholesterol in aqueous buffera
Productsb (% of 3)
directly into 1b in aqua (followed by 2,4-DNP hydrazine
derivatization after extraction); or (3) conversion of 4a or 4b
into 2b during the 2,4-DNP hydrazine derivatization process.
A number of experiments have been performed that help to
deconvolute these alternatives. When authentic 4a is incubated
in aqueous buffers for 20 h (pH 5.8–7.5), the allylic hydro-
peroxide undergoes a low percentage isomerization into 4b,
but there is no formation of 1a or 1b (ESI,w Table S7 and Fig.
S3). In addition, we have shown previously that authentic 1a
does not undergo measurable aldolization to 1b under the
conditions of these studies.7 Finally, Pratt and co-workers1
have recently observed that Hock-cleavage of 4a occurs in the
presence of 2,4-DNP hydrazine to give 2b, with a trace of 2a
being formed. Using authentic 4a, we find that this hydro-
peroxide undergoes Hock-cleavage during the hydrazine
derivatization process utilized in this study (2 mM 2,4-DNP
hydrazine, 0.08% HCl in isopropanol–ethanol (9 : 1)) to yield
quantitative amounts of 2b (no 2a is detected, ESI,w Fig. S4).
In addition, 7a-hydroperoxide 4b does not undergo conversion
into either 2a or 2b during the 2,4-DNP hydrazine derivatization
process vide supra. Therefore, the 2b that arises after ꢀO2À, 1O2
and PMA-activated hPMN-mediated oxidation of hLDL,
hHDL and liposomes (Table 1) is most likely a result of
cholesterol peroxidation to generate 4a, followed by Hock-
cleavage of 4a during the 2,4-DNP hydrazine derivatization
process.
Entry Oxidant system
Reactive species 2a
2b
1
2
3
4
5
6
7
8
9
10
Xan/XOc
ꢀO2À/HO2
ꢀO2À/HO2
H2O2
n.d.
0.02
0.05
40.01
40.01
40.01
2.30
2.61
0.30
0.35
0.10
ꢀ
Acet/XOd
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
0.40
ꢀ
Glc/GlcOxe
Glc/GlcOx/MPOf
Glc/GlcOx/FeCl2
H2O2/HOCl
H2O2/HOꢀ
1O2
g
HPIX/hnh
Glc/GlcOx/HPIX/hn H2O2/1O2
hPMNsi
hPMNs/D2O
O3/O2
Multiple
Multiple
O3
j
a
For full details see the ESI;w lipid extracts were derivatized with
2,4-DNP hydrazine (2 mM) in isopropanol–ethanol (9 : 1) containing
b
HCl (0.08%) Products were analyzed as their 2,4-DNP hydrazones 2a
or 2b and are reported as the amount of secosterol formed
c
as a percentage of the cholesterol in hLDL at t = 0. Xanthine
d
(Xan, 0.5 mM)/xanthine oxidase (XO, 0.4 mg mlÀ1). Acetaldehyde
e
(Acet, 10 mM) incubated for 30 min. Glucose (Glc, 2 mM)/glucose
f
oxidase (GlcOx, 0.1 mg mlÀ1) incubated for 14 h at 37 1C. Same as e
g
with myeloperoxidase (MPO, 0.2 mM). Same as
e with FeCl2
h
(20 mM). Hematoporphyrin IX (HPIX, 40 mM), hn (400–700 nm),
14 h incubation. hPMNs, (5.4 Â 107 cells per ml) activated with PMA
i
(0.1 mg mlÀ1) in PB or PB in D2O. A stream of O3/O2 gas (5 ml minÀ1
)
j
was passed over a solution of hLDL in PBS (pH 7.4) for 40 s.
no 2a is observed (ESI,w Tables S4 and S5). Similarly, the
photosensitized oxidation of 3 in liposomes (DOPC 80 mol%;
We have studied the singlet oxygen route to the cholesterol
secosterols further; photochemical oxidation of 3 in Al2O3-
treated CDCl3 with in situ acid (TFA, 1% v/v) leads to a good
yield of 1b (469%), and surprisingly, albeit in low yield,
of 1a (B3.5%). Inclusion of participating solvents, such
as methanol, or increasing acid concentration does not
significantly increase the yield of isolated secosterol 1a
(ESI,w Table S8). When the concentration of TFA is lowered
sufficiently (0.02% v/v), the progress of Hock-cleavage of 4a to
1
3 5 mol%; PIP-2 15 mol%) in PBS (pH 7.4) implicates a O2
pathway to 2b, because the measured levels of 2b are elevated
in PBS containing D2O, relative to H2O (ESI,w Table S6).
A clear finding of the hLDL oxidation studies detailed in
Table 1 is that the 2,4-DNP hydrazone of atheronal-A (2a) is
only generated after ozonolysis (Table 1). Thus, passage of an
ozone–oxygen mixture over the surface of an hLDL solution
containing catalase in aqueous buffer (pH 7.4), followed by
2,4-DNP hydrazine derivatization of the lipid extract, yields 2a
and 2b in a ratio of B4 : 1 (Table 1, entry 10, ESI,w Fig. S1).
To validate this result, cholesterol 3, which had been exposed
to ozone in aluminium oxide (Al2O3)-treated CDCl3, was
shaken in aqueous buffers at a range of physiological pHs
(5.8, 6.5, 7.0 and 7.4) in the presence or absence of a reductant
(Zn dust), and the resultant products were assessed by
1H NMR (ESI,w Fig. S2). This analysis reveals that, in all
cases tested, only cholesterol secosterol 1a is generated in this
aqueous treatment of ozonolyzed 3. Secosterol 1b is not
formed by this route as confirmed by 2,4-DNP hydrazine
1
form 1b can be followed by H NMR, reinforcing that 1a is
not formed even as a discrete intermediate, during this process
on a time-scale where allylic isomerization of 4a to 4b
competes measurably (ESI,w Table S9 and Fig. S5). Further-
more, when authentic 4a is incubated in isopropanol–ethanol
1
(9 : 1) with 0.08% HCl for 3 h, H NMR reveals a complex
mixture of 1b and 1b-derived dehydration and acetal products
but no 1a or 1a-derived products. Derivatization of this
4a-derived mixture with 2,4-DNP hydrazine then yields 2b
and hydrazones of 1b-derived dehydration products with no
measurable 2a or 2a-derived dehydration or acetal products in
either case (ESI,w Fig. S6). Thus, the summary of this singlet
oxygen/Hock-cleavage route to the cholesterol secosterols in
organic solvents is that in all the systems we have studied, 1b is
formed in vast excess over 1a.
1
derivatization of the H NMR samples.
The oxidation of 3 by 1O2 is a well-understood process that
was first reported by Schenck.8 In brief, the 5a-hydroperoxide
4a is formed as the major product, which can rearrange to the
7a-hydroperoxide 4b under certain conditions.9 Therefore,
the observation that 1O2 and superoxide anion-mediated
biological- and chemical-oxidations of hLDL, hHDL and
liposome-derived 3 in aqueous buffers result in elevated levels
of 2b could be a result of the following possibilities: (1)
conversion of 4a or 4b into 1a and subsequent aldolization
of 1a into 1b in aqua (followed by 2,4-DNP hydrazine
derivatization after extraction); (2) conversion of 4a or 4b
1
The O2 route to secosterol hydrazone 2b has a number of
implications for our quantification method of atheronals in
biological systems. Analytical methods to quantify cholesterol
secosterols from biological extracts must be able to deconvolute
the 5a-hydroperoxycholesterol 4a from the cholesterol secosterol
1b (the 2,4-DNP hydrazine method does not). To this end, we
have developed an isotope dilution (ID) GC-MS assay,
which measures the levels of 1b in biological extracts after
treatment with bis(trimethylsilyl) trifluoroacetamide (BSTFA)
ꢁc
This journal is The Royal Society of Chemistry 2009
Chem. Commun., 2009, 3098–3100 | 3099

View Article Online
in pyridine, conditions where 4a is completely stable (ESI,w
Table S10, Fig. S7-S8)). A comparison of this IDGCMS assay
with the 2,4-DNP hydrazine method in a plasma analysis of
1b-derivatives in the apoEÀ/À mouse model of atherosclerosis
reveals that the measured levels are not significantly different
[GCMS method = 7.5 Æ 1.3 nM (n = 4); 2,4-DNP hydrazine
method = 9.3 Æ 2.4 nM (n = 6); p 4 0.05], suggesting that 4a
is not contributing significantly to the measured amounts of 2b
in this system (ESI,w Fig. S9).
Analysis of human arterial plaque extracts by IDGCMS
reveals a mean [1a(b)] of 32 Æ 15 pmol mgÀ1 of tissue which is
similar to levels we previously reported for 2b in plaque
extracts2 (ESI,wFig. S10–S12). Interestingly, treatment of these
arterial plaque extracts with Zn/acetic acid prior to lipid
extraction results in an increase in the measured 1a(1b) levels,
suggesting that either 4a or cholesterol ozonides are present in
these atheroma samples and are being converted into 1a and
1b during tissue processing. We are currently investigating this
finding (ESI,w Fig. S11 and S12).
We are also interested in understanding to what extent the
acid-catalyzed rearrangement of 4a or a chemical oxidant with
the chemical signature of ozone may be contributing to
cholesterol secosterol biogenesis in vivo. As described vide
supra authentic 4a is stable at physiological pH for 420 h
and 5a-hydroperoxide 4a has been shown to be isolable from
biological samples by a number of groups.4,10 However, one
cannot discount that the formation of 1b, which is ubiquitous
in all biological samples we have tested thus far, does not, at
least in part, arise from this route.
Pratt and co-workers1 have shown that Hock-cleavage of
authentic 4a in organic solvents leads to levels of 1bc1a and
we herein have shown that 4a, generated by chemical- and
biological-oxidations of cholesterol in aqueous buffers,
generates 2b (with no measurable 2a) upon 2,4-DNP hydrazine
derivatization. Ozonolysis of 3 in aqueous buffers generates
1ac1b (at all physiologically-relevant pHs either in the
presence or absence of added chemical reductant). We have
also studied the ozone-mediated oxidation of hLDL-bound
cholesterol in aqueous buffer over a range of pH values,
temperatures and durations of reaction and found that the
ratio 1a : 1b varies from B15 : 1 to B4 : 1, but in all cases, the
level of atheronal-A is greater than atheronal-B (ESI,w
Table S12). An analysis of our original report of the levels of
cholesterol secosterols (measured as their hydrazones 2a
and 2b) in human arterial plaque samples reveals that the
ratio of 2a : 2b varies greatly (61.3 : 1 to 0.06 : 1, n = 28)
(ESI,w Table S11). However, the ratio of 2a : 2b is 41 in 20 of
the 28 samples and 410 in 10 of the 28 samples. Therefore, the
dominant form of cholesterol 5,6-secosterols in arterial plaque
extracts, where aldolization is minimal, is 1a.2 Based upon our
new understanding of the ozone and singlet oxygen pathways
from 3 into 2a and 2b (after 2,4-DNP hydrazine derivatization),
this clinical data offers support for an oxidant with the
chemical signature of ozone playing a composite role in
arterial plaque cholesterol secosterol 1a biogenesis.
While the debate as to the relevance of ozone generation in
biological systems continues,11 we continue to search for more
direct markers of this oxidant in biological extracts. The
terminal products from cholesterol ozonolysis, 1a and 1b,
may not discriminate the pathways to their formation under
all conditions, however, a ratio of 1a to 1b of 41 is strong
support for oxidation of 3 by an oxidant with the chemical
signature of ozone. Our current search is aimed at quantifying
more direct species that originate as a result of the primary
chemical reaction of ozone with biological molecules, such as
1,2,4-trioxolanes.
Financial support from the Skaggs Institute for Chemical
Biology and the NIH (AG028300).
Notes and references
1 J. Brinkhorst, S. J. Nara and D. A. Pratt, J. Am. Chem. Soc., 2008,
130, 12224.
2 P. Wentworth, Jr., J. Nieva, C. Takeuchi, R. Galve,
A. D. Wentworth, R. B. Dilley, G. A. DeLaria, A. Saven,
B. M. Babior, K. D. Janda, A. Eschenmoser and R. A. Lerner,
Science, 2003, 302, 1053.
3 L. L. Smith, Cholesterol Autoxidation, Plenum Press, New York,
1981; K. Osada and A. Sevanian, Methods Enzymol., 2000, 319,
188.
4 A. W. Girotti and W. Korytowski, Methods Enzymol., 2000, 319,
85.
5 C. Takeuchi, R. Galve, J. Nieva, D. P. Witter, A. D. Wentworth,
R. P. Troseth, R. A. Lerner and P. Wentworth, Jr., Biochemistry,
2006, 45, 7162.
6 J. Bieschke, Q. Zhang, D. A. Bosco, R. A. Lerner, E. T. Powers,
P. Wentworth and J. W. Kelly, Acc. Chem. Res., 2006, 39, 611.
7 Q. Zhang, E. T. Powers, J. Nieva, M. E. Huff, M. A. Dendle,
J. Bieschke, C. G. Glabe, A. Eschenmoser, P. Wentworth, Jr.,
R. A. Lerner and J. W. Kelly, Proc. Natl. Acad. Sci. U. S. A., 2004,
101, 4752; J. C. Scheinost, H. Wang, G. E. Boldt, J. Offer and
P. Wentworth, Jr., Angew. Chem., Int. Ed., 2008, 47, 3919;
J. Nieva, A. Shafton, L. J. Altobell, S. Tripuraneni, J. K. Rogel,
A. D. Wentworth, R. A. Lerner and P. Wentworth, Biochemistry,
2008, 47, 7695; D. A. Bosco, D. Fowler, Q. Zhang, J. Nieva,
E. T. Powers, P. Wentworth, Jr., R. A. Lerner and J. W. Kelly,
Nat. Chem. Biol., 2006, 2, 249.
8 G. O. Schenck, Angew. Chem., 1957, 69, 579.
9 G. O. Schenck, O.-A. Neumuller and W. Eisfeld, Justus Liebigs
¨
Ann. Chem., 1958, 618, 202; A. Nickon and J. F. Bagli, J. Am.
Chem. Soc., 1959, 81, 6330; A. Nickon and J. F. Bagli, J. Am.
Chem. Soc., 1961, 83, 1498.
10 S. Yamazaki, N. Ozawa, A. Hiratsuka and T. Watanabe, Free
Radical Biol. Med., 1999, 27, 110.
11 A. W. Kettle, B. M. Clark and C. C. Winterbourn, J. Biol. Chem.,
2004, 279, 18521; W. A. Pryor, K. N. Houk, C. S. Foote,
J. M. Fukoto, L. J. Ignarro, G. L. Squadrito and K. J.
A. Davies, Am. J. Physiol., 2006, 291, R491; H. Sies, Angew.
Chem., Int. Ed., 2004, 43, 3514.
ꢁc
This journal is The Royal Society of Chemistry 2009
3100 | Chem. Commun., 2009, 3098–3100