H-Atom Abstraction vs Addition: Accounting for the Diverse Product Distribution in the Autoxidation of Cholesterol and Its Esters

Article abstract of DOI:10.1021/jacs.8b11524

We recently communicated that the free-radical-mediated oxidation (autoxidation) of cholesterol yields a more complex mixture of hydroperoxide products than previously appreciated. In addition to the epimers of the major product, cholesterol 7-hydroperoxide, the epimers of each of the regioisomeric 4- and 6-hydroperoxides are formed as is the 5α-hydroperoxide in the presence of a good H-atom donor. Herein, we complete the story by reporting the products resulting from competing peroxyl radical addition to cholesterol, the stereoisomeric cholesterol-5,6-epoxides, which account for 12% of the oxidation products, as well as electrophilic dehydration products of the cholesterol hydroperoxides, 4-, 6-, and 7-ketocholesterol. Moreover, we interrogate how their distribution - and abundance relative to the H-atom abstraction products - changes in the presence of good H-atom donors, which has serious implications for how these oxysterols are used as biomarkers. The resolution and quantification of all autoxidation products by LC-MS/MS was greatly enabled by the synthesis of a new isotopically labeled cholesterol standard and corresponding selected autoxidation products. The autoxidation of cholesteryl acetate was also investigated as a model for the cholesterol esters which abound in vivo. Although esterification of cholesterol imparts measurable stereoelectronic effects, most importantly reflected in the fact that it autoxidizes at 4 times the rate of unesterified cholesterol, the product distribution is largely similar to that of cholesterol. Deuteration of the allylic positions in cholesterol suppresses autoxidation by H-atom transfer (HAT) in favor of addition, such that the epoxides are the major products. The corresponding kinetic isotope effect (k_H/k_D ～ 20) indicates that tunneling underlies the preference for the HAT pathway.

Full text of DOI:10.1021/jacs.8b11524

Subscriber access provided by UNIV OF LOUISIANA
Article
H-Atom Abstraction vs. Addition: Accounting for the Diverse
Product Distribution in the Autoxidation of Cholesterol & its Esters
Zosia A. M. Zielinski, and Derek A. Pratt
J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.8b11524 • Publication Date (Web): 09 Jan 2019
Downloaded from http://pubs.acs.org on January 10, 2019
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 16
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
H-Atom Abstraction vs. Addition: Accounting for the Diverse
Product Distribution in the Autoxidation of Cholesterol & its Esters
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Zosia A. M. Zielinski & Derek A. Pratt*
Department of Chemistry and Biomolecular Sciences, University of Ottawa, Ottawa, Ontario, Canada K1N 6N5
ABSTRACT: We recently communicated that the free radical-mediated oxidation (autoxidation) of cholesterol yields a more com-
plex mixture of hydroperoxide products than previously appreciated. In addition to the epimers of the major product, cholesterol 7-
hydroperoxide, the epimers of each of the regioisomeric 4- and 6-hydroperoxides are formed, as is the 5a-hydroperoxide in the
presence of a good H-atom donor. Herein we determine the products resulting from competing peroxyl radical addition to cholesterol,
the stereoisomeric cholesterol-5,6-epoxides, which account for 12% of the oxidation products, as well as electrophilic dehydration
products of the cholesterol hydroperoxides, 4-, 6- and 7-ketocholesterol. Moreover, we interrogate how their distribution – and abun-
dance relative to the H-atom abstraction products – changes in the presence of good H-atom donors, which has serious implications
for how these oxysterols are used as biomarkers. The resolution and quantification of all autoxidation products by LC-MS/MS was
greatly enabled by the synthesis of a new isotopically-labeled cholesterol standard and corresponding selected autoxidation products.
The autoxidation of cholesteryl acetate was also investigated as a model for the cholesterol esters which abound in vivo. While
esterification of cholesterol imparts measurable stereoelectronic effects – most importantly reflected in the fact that it autoxidizes at
4 times the rate of unesterified cholesterol – the product distribution is largely similar to cholesterol. Deuteration of the allylic posi-
tions in cholesterol suppresses autoxidation by H-atom transfer (HAT) in favour of addition, such that the epoxides become the major
products. The corresponding kinetic isotope effect (kH/kD ~ 20) indicates that tunneling underlies the preference for the HAT pathway.
and essentially all of the brain’s cholesterol is synthesized lo-
INTRODUCTION
cally – of which an estimated 70% is contained in myelin.6 Chol
autoxidation products have been reported to be in the vicinity
of 1 µg/mL in human serum and LDL,7-9 and elevated levels
have been associated with many degenerative diseases,10
though it remains unclear whether their role is causative or con-
sequential – a debate that continues to garner interest.10-13 The
exception to this has been the secosterols 1 and 2, which have
been implicated in cardiovascular disease,11 cancer,12 and neu-
rodegeneration (including Alzheimer’s13 and Parkinson’s14).
These compounds have been proposed to arise in vivo from ozo-
nolysis of chol,15 or more likely, Hock fragmentation of chol 5-
hydroperoxide (chol 5-OOH) formed by oxidation with 1O2.16
Lipid autoxidation (peroxidation) has been linked to a myriad
of degenerative diseases and conditions. The initial products of
the free radical chain mechanism are usually hydroperoxides,
and downstream products include electrophilic species contain-
ing ketones or aldehydes, including the pathogenic compounds
malondialdehyde, derived from arachidonate, and 4-hy-
droxynonenal, derived from omega-6 polyunsaturated fatty ac-
ids (PUFAs), in general.1 Sterol autoxidation has received com-
paratively less attention than PUFA autoxidation, presumably
because sterols are typically less reactive, reflected in their
lower propagation rate constants (kp) (cf. Eq. 1 and Table 1).
The exception to this is 7-dehydrocholesterol (7-DHC), the
highly autoxidizable precursor to vitamin D3 and cholesterol
(chol). It has been shown that 7-DHC builds up in patients with
Smith-Lemli-Opitz syndrome, and its oxidation is believed to
be involved in the pathogenesis of the disease.2,3
H
H
H
H
H
H
ꢀ
ꢁ
O
ROO• + RH → ROOH + R•
(1)
O
HO
O
HO
OH
1
2
Table 1: Propagation rate constants for selected lipids.4
We recently showed17 that chol autoxidation yields a more
complex mixture of hydroperoxide products than previously
understood (cf. Figure 1B). In addition to chol 7-hydroperoxide
(7-OOH), both epimers of 4-OOH and 6-OOH were formed in
chol autoxidations carried out in aerated chlorobenzene due to
the C4-H abstraction pathway previously deemed uncompeti-
tive with C7-H abstraction (cf. Figure 1A). Furthermore, 5a-
OOH was formed in the presence of phenolic antioxidants,
which are significantly more reactive H-atom donors than chol
itself and are able to trap the 5a-OOl intermediate faster than
PUFA
kp (M-1s-1)
62
sterol
kp (M-1s-1)
11
linoleate
arachidonate
cholesterol
7-DHC
201
2260
While chol is less reactive to autoxidation than any of the
aforementioned lipids, it is considerably more abundant, consti-
tuting up to 50 mol% of the lipids in cell membranes.5 In addi-
tion, the brain contains more cholesterol than any other organ,
1
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 2 of 16
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 1: The competing pathways in cholesterol (chol) autoxidation: H-atom abstraction resulting in regioisomeric hydroperoxide
formation (A) and peroxyl radical addition yielding epoxides and peroxides (B). The associated product distribution from MeO-
AMVN-initiated autoxidations of chol in chlorobenzene at 37°C in the presence of 4-tert-butyl-2,6-dimethylphenol (BDMP), where
the alcohols are detected due to pre-treatment with PPh3 (C).17 Authentic chol epoxides have the same retention times as chol 4-OHs
(D). Data in panels C and D were obtained by HPLC (95:5 hexane:i-PrOH, 1.5 mL/min) with APCI+-MS/MS detection for the m/z
à m/z transitions indicated.
pathway preference, might also be affected. Furthermore, we
surmised that O2 could (reversibly) add to the alkyl radical
formed upon peroxyl radical addition, yielding yet another
class of compounds, and that this distribution could also be
sensitive to H-atom donors (cf. Figure 1C). Inclusion of this
manifold of reactivity provides a comprehensive view of cho-
lesterol autoxidation chemistry which provides crucial in-
sights on their use as biomarkers and implications in the dis-
ease pathogenesis.
the b-fragmentation of O2 (the reverse of oxygen addition) oc-
curs. These observations provide an alternative explanation
for the origin of the secosterols. Moreover, we found that chol
6-OOH also underwent Hock fragmentation to produce 1 and
2, suggesting chol autoxidation may have a larger role in dis-
ease pathogenesis than previously believed.
While these developments contributed to a more thorough
understanding of chol autoxidation, the complete mechanism
is even more nuanced. Herein, we report on our investigations
of the interplay between H-atom abstraction by peroxyl radi-
cals and their addition as propagation steps in cholesterol au-
toxidation (cf. Figure 1A and 1B, respectively). The interme-
diate radical formed upon addition can undergo intramolecu-
lar homolytic substitution (SHi) to yield 5a,6a-epoxycholes-
terol (chol a-epoxide) or 5b,6b-epoxycholesterol (chol b-
epoxide) and an alkoxyl radical. The alkoxyl radical can then
abstract an H-atom from another molecule of chol to propa-
gate the radical chain reaction. Although chol epoxides are
known products of chol autoxidation, their quantification in
relation to chol hydroperoxides has been complicated by the
fact that they are also believed to be formed by enzymes
and/or other oxidants (e.g. O3, inhaled in polluted air).9,18,19
The a:b ratio is considered to be an indication of the origin of
the epoxides; for example, the autoxidation ratio is reported
to be ca. 1:3, whereas epoxides from enzymatic origin are al-
most exclusively the a-isomer.18 We were therefore interested
in studying the interplay between the H-atom abstraction and
peroxyl radical addition pathways. Since the distribution of
hydroperoxide products was affected by the addition of H-
atom donors, we wondered if the ratio of epoxides, or the
RESULTS
I. Cholesterol Autoxidation Revisited. Standards of the cho-
lesterol-derived a- and b-epoxides were prepared by MCPBA
oxidation as a mixture, and analyzed using the same HPLC-
APCI+-MS/MS technique used in our previous work (see Ex-
perimental Section for details).17 To our surprise, not only did
the epoxides ionize efficiently with the same MS/MS transi-
tion (m/z 385.4 à m/z 367.4), but they also eluted at the same
retention time as chol 4a-OH and 4b-OH (cf. Figure 1D). At-
tempts to identify conditions that afforded clean separation of
chol b-epoxide and chol 4a-OH were unsuccessful. The epox-
ides did, however, have a unique MS/MS transition (m/z 403.4
à m/z 385.4, cf. Supporting Information page S3), enabling
their quantification and deduction of the amount of chol 4a-
OH (from MS/MS chromatograms corresponding to the
shared transition of both of the products, i.e. m/z 385.4 à m/z
367.4).
Isotopically-labeled standards were required for precise
quantification of chol 4a-OH and the epoxides, which were
prepared from the newly synthesized 17,20,21,21,21,22-d6-
2
ACS Paragon Plus Environment

Page 3 of 16
Journal of the American Chemical Society
chol (8). We had previously used 26,26,26,27,27,27-d6-chol vivo and has been frequently used as a biomarker of lipid ox-
idation.28-30 In theory, each of chol 4-OOH, 6-OOH and 7-
OOH can dehydrate to give the corresponding ketones. Au-
thentic standards for 4-, 6- and 7-ketochol were synthesized
(see Supporting Information), and although they ionized well
by APCI+, the parent M+H+ ion (m/z 401.4) did not fragment
to a significant extent, precluding MS/MS analysis. However,
given their α,β-unsaturation, the ketones have absorption
maxima at 234 nm, and could be quantified by UV detection.
Indeed, measurable amounts of 6- and 7-ketochol were
formed in the autoxidations. 4-Ketochol was surprisingly non-
polar – presumably due to the C3-OH participating in a hy-
drogen bond with the ketone – and could not be separated
from the solvent front. Since the ketochols could also be de-
tected by APCI-MS, select autoxidation samples were ana-
lyzed this way, but 4-ketochol was not detected.
1
2
3
4
5
6
7
8
to prepare our standards, which was synthesized using the
method of Holm and Crossland,20 with a (optimal) 7% yield
over 9 steps. To our knowledge, the only other synthesis of
chol-26,26,26,27,27,27-d6 uses an expensive desmosterol
,
starting material.21 22 We therefore opted to develop a synthe-
sis of 17,20,21,21,21,22-d6-chol (8), which was achieved in 6
steps from pregnenolone with an overall yield of 34%
(Scheme 1).
Scheme 1. Synthesis of 17,20,21,21,21,22-d6-cholesterol (8).
9
O
O
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
CD3
D
D2O, NaOD
Dioxane
TIPS-Cl, Im
DMF, DCM
RT, 4 h
4
110oC, 24 h
quant.
97%
TIPSO
HO
Br
5
OH
D
D3C
Chol autoxidations were also carried out in the presence of
the radical-trapping antioxidant BDMP to explore the kinetic
product distribution, where b-fragmentation of intermediate
peroxyl radicals is minimized. Expectedly, the total amount of
oxidation products decreased with increasing [BDMP] (cf.
Figure S22). However, upon comparing the relative amounts
of each oxidation product, it is clear that the reversibility of
O2 addition is a contributing factor to the overall product dis-
tribution. The products derived from C4-H abstraction (chol
4-OOH and 6-OOH) and peroxyl radical addition (chol α- and
β-epoxide) decreased relative to chol 7-OOH with increasing
BDMP concentration, albeit after an initial surge in epoxide
concentration (Figure 2C). Conversely, the proportion of chol
5α-OOH increased with added BDMP. This dependence was
used to determine the β-fragmentation rate constant for 5a-
1) Mg, THF
70oC, 2h
2) 5, THF
70oC 16 h
84%
POCl3, Py
0oC to RT, 16h
48%
TIPSO
6
D
D
D3C
D3C
1) D2, PtO2
Dioxane, AcOD
RT, 16 h
D
D
2) TBAF, THF
RT, 16 h
84%
TIPSO
HO
7
8
Briefly, the 3-OH of pregnenolone was silylated with
TIPS-Cl and the product deuterated based on the method of
De Miranda et. al.,23 resulting in the quantitative formation of
5. While attempts to couple 5 and 1-bromo-4-methylpentane
by Wittig chemistry resulted in partial D/H exchange, the
analogous Grignard reaction proceeded without loss of D. The
subsequent elimination resulted in a mixture of three alkenes
(purported to be 55% ∆20(22), 25% ∆20(21) and 20% ∆17(20) when
carried out on a similar substrate).24 Carrying forward the
mixture would have undoubtedly increased the overall yield,
but would have led to a mixture of d5-and d6-chol in the final
product. Thus, the desired ∆20(22) alkene (7) was isolated from
the mixture by silver-impregnated silica gel chromatography.
Catalytic hydrogenation of 7 with D2 installed the final two
deuterium atoms, and fluoro-desilylation afforded
17,20,21,21,21,22-d6-chol. From this novel d6-chol, the de-
sired internal standards, d6-chol 5a-OH and the d6-chol epox-
ides (a mixture of a- and b-isomers) were prepared by known
methods (see Experimental Section for details).
Ÿ
OO kβ = (3.8 ± 0.9) × 105 s−1 (cf. Figure S23). 31 The initial
increase in epoxide was intriguing, and in addition to influ-
encing the amount of epoxides formed compared to hydrop-
eroxides, the ratio of a- to b-epoxide was also influenced by
added BDMP (cf. Figure 2D).
Chol autoxidations were also carried out in the presence of
PMC, an analogue of a-tocopherol, the most bioactive form
of Vitamin E. As a more potent H-atom donor than BDMP
(compare k inh = 3.8 × 106 M-1s-1 for PMC32 with 1.3 × 105 M-
1 s-1 for BDMP17) less autoxidation takes place in the presence
of comparable amounts of PMC. Nevertheless, the trends ob-
served in the PMC-loaded autoxidations are fully consistent
with those observed in the presence of BDMP. It is notewor-
thy that in the presence of PMC, chol 5a-OH increased to a
maximum of 15 mol% of the product distribution. Analogous
to the BDMP-loaded autoxidations, there was an initial surge
in epoxide concentration as the PMC concentration was in-
creased, reaching a maximum at roughly the same amount (ca.
30 mol%) before decreasing.33 The ratio of a-epoxide to b-
epoxide, on the other hand, increased roughly two-fold past
the BDMP maximum at higher [PMC] (cf. Figure S28).
Cholesterol autoxidations were then carried out as de-
scribed in the Experimental Section. In each instance the sam-
ples were treated with PPh3 to convert the hydroperoxide
products to their corresponding alcohols, which are more sta-
ble toward chromatography and MS analysis.25 Control exper-
iments showed the chol epoxides to be stable under these con-
ditions. In addition to the six primary products observed in our
previous work (chol 4α-OH, 4β-OH, 6α-OH, 6β-OH, 7α-OH,
and 7β-OH), the α- and β-epoxides were observed (Figure
2A), and found to constitute roughly 12% of the product dis-
tribution (Figure 2B).26
II. Computational Insights on the Origins of Chemo- and
Regioselectivity in Cholesterol Autoxidation. To help ration-
alize the chemo- and regioselectivity observed in the choles-
terol autoxidations, electronic structure calculations were car-
ried out to determine the transition state structures for both H-
atom abstraction from, and addition to, chol by a model per-
oxyl radical (methylperoxyl). For computational expediency,
the calculations employed a simplified model of chol which
included only the A and B rings at which the relevant chem-
istry occurs. In our preliminary report,17 we reported that the
C4-H and C7-H bond dissociation enthalpies – determined by
In addition to the primary hydroperoxide and epoxide
products of autoxidation, we sought to determine whether sig-
nificant amounts of secondary products were formed compet-
itively in the earlystages of the autoxidations.27 7-Ketocholes-
terol (7-ketochol), the dehydration product of chol 7-OOH or
termination product of two chol-7OO•, is known to form in
3
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 4 of 16
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 2: Product distribution from MeOAMVN-initiated autoxidation of chol in chlorobenzene at 37oC. (A) Chromatograms (con-
ditions in experimental section) from HPLC/APCI+-MS/MS analysis with m/z 403.4 à m/z 385.4 (top) and m/z 385.4 à m/z 367.4
(bottom), where the hydroperoxide-derived alcohols are detected following pretreatment of samples with PPh3. (B) Distribution of α-
and β-epimers, as well as dehydration products (keto). (C) Product distribution as a function of added BDMP. (D) Ratio of chol
epoxides to chol OH (<), and ratio of a-epoxide to b-epoxide (=) as a function of added [BDMP].
the high-accuracy CBS-QB3 compound method – were 89.0
and 83.2 kcal/mol, respectively. The origin of the difference
is clear when looking at the spin density distributions in the
allylic radicals formed from C4-H and C7-H abstraction (cf.
Figure 3); the latter benefits from having its spin density lo-
calized on both a tertiary and a secondary position, compared
to the two secondary positions in the former. Moreover, the
radical derived from C4-H abstraction is delocalized across
both A and B rings, and the concomitant planarity causes each
ring to distort from their optimal chair conformations. Never-
theless, the difference in BDEs does not alone govern the ratio
of C4-H to C7-H abstraction. Given that H-atom abstraction
is followed by the near diffusion-controlled addition of O2 to
the so-formed alkyl radicals, this kinetically controlled step is
instead dictated by the barrier of H-atom transfer. We had pre-
viously calculated the barriers for the key transition states
(TSs) for the C7-H and C4-H abstraction pathways with den-
sity functional theory, but herein the results of corresponding
CBS-QB3 calculations are presented.
(LDL) and in vivo. Indeed, H-atom abstraction from the C7α-
H has a relatively low enthalpic barrier (A1 in Figure 3A),
benefitting from a secondary orbital interaction between the
contribution of the internal peroxyl oxygen atom to the nomi-
nally singly occupied p* and the electron-rich C5-C6 p bond.
This (syn) TS structure is 1.4 kcal/mol lower in free energy
than the corresponding anti TS structure (A2 in Figure 3A).
H-atom abstraction from the C7β-H is slightly higher in en-
ergy, presumably due to the steric hindrance imposed by the
C10-methyl group and the equatorial β-H requiring more re-
organization of the B-ring to achieve the transition state.34
However, the syn geometry that benefits from the secondary
orbital interaction is still preferred over the anti geometry, de-
spite the steric interaction (A3 and A4, respectively, in Figure
3A).
The proximity of the 3-OH group introduces a competing
interaction in the C4-H abstraction TS. In fact, the preferred
geometry for HAT from C4β-H involves a H-bonding inter-
action between the internal peroxyl oxygen atom and the C3-
OH group (B1 in Figure 3B). In contrast, the syn geometry for
C4β-H (B2 in Figure 3B) abstraction is less favourable by 1.1
kcal/mol. The H-bonding interaction lowers the barrier to
within 2 kcal/mol of the most optimal C7-H abstraction TSs.
Conventional wisdom has been that the C7-H oxidation
pathway operates more or less exclusively, as chol 7-OOH,
chol 7-OH and 7-ketochol have been reported to be the vastly
dominant products in liposomes, low density lipoprotein
4
ACS Paragon Plus Environment

Page 5 of 16
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
Figure 3: (A) H-atom abstraction from chol by a methylperoxyl radical at C7-H with transition state (TS) structures: C7α-H with a
syn geometry and secondary orbital interaction (A1) or with an anti geometry (A2); or from C7β-H with a syn geometry and secondary
orbital interaction (A3) or with an anti geometry (A4); as well as spin density distribution for the allylic radical formed therefrom
(right). (B) H-atom abstraction from chol by a methylperoxyl radical at C4-H with TS structures: C4β-H with a H-bonding interaction
with the C3-OH (B1) or a syn geometry and secondary orbital interaction (B2); or from C4α-H with a syn geometry and secondary
orbital interaction (B3) or no secondary interaction (B4); as well as spin density distribution for the allylic radical formed therefrom
(right). Highest occupied molecular orbitals (HOMOs) are shown for TS structures in A and B that benefit from a secondary orbital
interaction. (C) Reaction coordinate for the epoxidation pathway (left) and TS structures for addition to chol by a methylperoxyl
radical at: C6 α-face (C1), C6 β-face (C2), C5 α-face (C3), or C5 β-face (C4), on the right. All structures were computed with CBS-
QB3, energies are reported in kcal/mol, and TS structures are accompanied by the associated barriers.
48
49
50
51
52
53
54
55
56
57
58
59
60
5
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 6 of 16
suppress C4-H abstraction. On the other hand, the trends in-
The computed TSs for C4a-H abstraction, which are much
1
2
3
4
5
6
7
8
volving the epoxides were different – there was no initial
surge in epoxide concentration relative to the other products,
and the ratio of a- to b-epoxide was unaffected by [BDMP].
Otherwise, the outcomes of chol and chol OAc autoxidations
were largely the same. The kb of the 5a-peroxyl derived from
chol OAc was 2-fold higher than with chol (vide supra), as it
was determined to be (6.7 ± 0.5) × 105 s−1 (cf. Figure S30).
Similarly, the amounts of ketochol products were consistent
with chol autoxidations, where 7-ketochol was ca. 10% of the
mixture, 6-ketochol was a minor component and no 4-keto-
chol was detected.
higher in energy, are devoid of H-bonding interactions be-
tween the peroxyl radical and the C3-OH, presumably due to
its equatorial position in the A-ring chair structure. Further-
more, C4α-H abstraction benefits from a secondary orbital in-
teraction only if the A-ring is flipped to a boat geometry (B3
in Figure 3B), which lowers the barrier slightly relative to H-
atom abstraction perpendicular to the framework (B4 in Fig-
ure 3B).
To explore the peroxyl radical addition pathway, we cal-
culated the barriers to peroxyl radical addition at C6 from the
a- or β-face of the sterol framework (C1 or C2, respectively,
in Figure 3C), or at C5 from the a- or β-face (C3 or C4, re-
spectively, in Figure 3C). It appears that two factors affect ad-
ditions at these four positions. First, C5 addition yields a sec-
ondary alkyl radical, whereas the C6 addition yields a more
stable tertiary alkyl radical. However, despite any steric hin-
drance imposed by the C10 methyl group, C6 addition is pre-
ferred on the β-face, and C5 addition on the a-face. This is
due to the resultant alkyl radicals adopting chair confor-
mations, while C6 addition on the a-face and C5 addition on
the β-face force the B-ring into a twist boat conformation. The
interplay of these two factors are likely why both C5 addition
pathways have higher barriers, but the intermediate formed
from C5a addition is more stable than the one formed from
C6a addition (cf. Figure 3C).
Since peroxyl radical addition is a reversible process, we
also calculated the subsequent SHi reaction to form an epoxide
and an alkoxyl radical, which had enthalpic barriers of 7.3 and
9.9 kcal/mol for the 6a- and 6β-peroxides, respectively, cor-
responding to unimolecular rate constants of 3.8×107 and 6.6
×105 s-1, respectively. Taken together, the overall rate of
epoxide formation is fastest for initial C6b addition, with a
computed kepox of 19 M-1s-1 (cf. Figure 3C). Interestingly, this
is predicted to be almost as fast as H-atom abstraction from
C7-H (kHAT = 24 M-1s-1), and significantly faster than abstrac-
tion of C4-H (kHAT = 0.96 M-1s-1), in good agreement with the
experimental observations.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Once again, computations were carried out to enable ra-
tionalization of our observations. The CBS-QB3-predicted C-
H BDEs are not significantly affected when chol is esterified
– they were computed to be 88.6 and 83.1 kcal/mol for C4-H
and C7-H, respectively, using CBS-QB3. Thus, the C4-H
bond is weakened by ca. 0.4 kcal/mol relative to chol, but is
still 5 kcal/mol stronger than the C7-H. TSs were calculated
for H-atom abstraction at the C4-H and C7-H positions. Anal-
ogous to chol, the preferred C7a-H abstraction geometry ben-
efits from a secondary orbital interaction (cf. Figure 4D). This
TS again was lower in energy than the C7b-H abstraction TS,
which is subject to steric hindrance from the C10-methyl. For
C4-H abstraction, no H-bonding interaction was possible be-
tween the peroxyl radical and the acetate group; however, the
enthalpic barrier to C4b-H abstraction with a secondary or-
bital interaction was only 20.0 kcal/mol – 0.5 kcal lower than
the same geometry in chol. This, again, was favoured relative
to C4a-H abstraction, which required an A-ring flip to a boat
conformation to benefit from the secondary orbital interac-
tion.
To the best of our knowledge there are no reports that es-
terified chol autoxidizes more rapidly than chol, as our com-
putations would suggest. We therefore determined the propa-
gation rate constants (kp) for chol and chol OAc using the per-
oxyl radical clock approach (cf. Figure 4C).4 The kp of chol
OAc was almost four-fold higher than chol, at 31.4 ± 0.9 and
8.4 ± 0.6 M-1 s-1, respectively.35
III. Cholesteryl Acetate Autoxidation. Given that the H-
bonding interaction between the C3-OH of chol and the per-
oxyl radical in the C4-H abstraction TS seemed to be a rea-
sonable explanation for the observation of chol 4-OOH and 6-
OOH in autoxidations, we sought to further probe this exper-
imentally. We therefore carried out cholesteryl acetate (chol
OAc) autoxidations, as described in the Experimental Section.
In each instance, the samples were treated with PPh3 to con-
vert the hydroperoxide products to their corresponding alco-
hols. The mixtures were further treated with 1M KOH in
MeOH to convert the chol OAc-derived oxidation products to
the corresponding chol-derived products, as the chol OAc
products were not resolvable by normal phase HPLC. Control
experiments were conducted to ensure quantitative deprotec-
tion of chol OAc and recovery of the oxidation products under
these conditions.
IV. Isotopic Substitution Inverts Product Distribution in
Cholesterol Autoxidation. Given that the computed rates of
H-atom abstraction from, and peroxyl radical addition to, chol
were similar, we wondered whether tunneling in the HAT
pathway is responsible for the predominance of hydroperox-
ides over epoxides in cholesterol autoxidations. We therefore
synthesized 2,2,4,4,7,7-d6-cholesterol (9, 2,2,4,4,7,7-d6-chol)
by the method of Fujimoto et. al.22 and carried out analogous
autoxidations to those described above. Amazingly, the prod-
uct distribution observed in the 2,2,4,4,7,7-d6-chol autoxida-
tions was completely inverted compared to that observed from
chol, such that the epoxides accounted for over 75% of the
product mixture. The ratio of hydroperoxide and epoxide
products can be used to estimate a kinetic isotope effect (KIE)
of kH/kD = 20 ± 1 on H-atom abstraction from chol by chain-
propagating peroxyl radicals (cf. Figure 5). This value is sig-
nificantly larger than the classical limit, supporting the contri-
bution of tunneling in the HAT pathway(s). It is noteworthy
that the rate of autoxidation of 2,2,4,4,7,7-d6-chol was signif-
icantly slower than of chol, yielding ca. 4-fold less oxidation
products in total, but roughly the same amount of epoxide
products. This suggests that deuteration affected only the H-
Somewhat surprisingly, the amount of 4-OOH and 6-OOH
only decreased by about half, and the product distribution was
otherwise similar to chol autoxidations (cf. Figure 4A). The
presence of chol 4-OOH and 6-OOH, albeit in a slightly lesser
amount than in the chol autoxidations, indicated that the ab-
sence of the H-bonding interaction was not enough to fully
6
ACS Paragon Plus Environment

Page 7 of 16
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 4: (A) Product distribution from MeOAMVN-initiated autoxidation of cholesteryl acetate (chol OAc) in chlorobenzene at
37oC as a function of added BDMP. The hydroperoxide-derived alcohols were quantified following pretreatment of samples with
PPh3. (B) Ratio of epoxides derived from chol OAc to alcohols (hydroperoxides) derived from chol OAc (<), and ratio of a-epoxide
to b-epoxide (=) as a function of added [BDMP] (empty data points are the corresponding data for chol autoxidations, presented for
comparison). (C) Propagation rate constants (kp) for chol (<) and chol OAc (=) as determined from the c,t/t,t ratio of conjugated
hydroperoxides arising from autoxidation of methyl linoleate in the presence of chol and chol OAc (see Experimental Section). (D)
CBS-QB3 computed lowest energy transition states for H-atom abstraction from a model of chol OAc by a model peroxyl radical at
C7α-H (top) or C4β-H (bottom) and the associated barriers in kcal/mol.
atom abstraction pathway(s), and had a negligible effect on
the rate of peroxyl radical addition.
initial reports suggested chol epoxides may be alkylating
agents, it was later shown that they have surprisingly weak
electophilicity compared to other epoxides.36 Their interaction
with several enzymes (including topoisomerase-II, liver-X-re-
ceptors, cholesterol acyl transferases, and estrogen receptors)
is more likely to be the reason for their pathogenesis. The
epoxides are known to be metabolized to chol 3,5,6-triols by
chol epoxide hydrolase enzymes (present in the liver, and to a
lesser extent, the lungs), though the a-epoxide is significantly
more susceptible to hydrolysis.9,18 We therefore sought to
quantify the product distribution of the H-atom abstraction
and addition pathways simultaneously, as well as probe
whether the addition of H-atom donors affects the distribution
of products derived from the latter.
DISCUSSION
Our preliminary study on chol autoxidation focused on the
formation of the regioisomeric hydroperoxide products and
the changes in their distribution upon addition of H-atom do-
nors to the system (cf. Figure 1A). However, the mechanism
of chol autoxidation is more complex, as peroxyl radicals also
add to the C5-C6 unsaturation in competition with H-atom ab-
straction. The products of this addition pathway may, theoret-
ically, include chol epoxides and peroxides (cf. Figure 1B).
Indeed, the epoxides are known autoxidation products of chol,
but are also produced by other oxidants, which has compli-
cated their quantification in the past. There has been signifi-
cant interest in the bioactivity of chol epoxides, as they are
reported to influence sterol synthesis and esterification, and
have links to a number of diseases, including cancer.9,18 While
Using the modified HPLC-MS/MS protocol, the chol
epoxides were quantified and found to constitute a significant
amount of the product distribution in chol autoxidation, at ca.
15 mol% in the absence of H-atom donor. Meanwhile, the
7
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 8 of 16
either propagate the radical chain, or it could be trapped as the
1
2
3
4
5
6
7
8
corresponding hydroperoxide by a H-atom donor – a pathway
that is more competitive at higher [BDMP] (or [PMC]). Since
subsequent analysis reduced the hydroperoxide products to
their corresponding alcohols, one could envision a displace-
ment of the added peroxide fragment to yield an epoxide.
Thus, an increase in the peroxide product (3) would lead to an
artificial increase in the chol epoxide measured. The hydrop-
eroxide in 3 could also be reduced to its corresponding alcohol
by other means in vivo (i.e. glutathione peroxidase 4), sug-
gesting that chol epoxides may not simply arise from SHi
chemistry, as previously thought. Indeed, upon direct infusion
of select chol autoxidation samples, we observed the m/z for
the corresponding alcohol of 3 wherein R is initiator-derived.
Upon increasing [BDMP], this signal weakened, and the m/z
for a corresponding BDMP-phenoxyl radical addition product
was observed (cf. Figure S41).
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 5. Product distributions in autoxidations of chol and
2,2,4,4,7,7-d6-chol.
Scheme 2: Reversible addition of O2 to lipid-derived (bis)al-
lylic radicals is key to the product distributions observed in
autoxidations. β-Fragmentation rate constants are shown for
key classes of lipid-derived peroxyl radicals.
amount of chol 4-OOH accounted for ca. 5 mol% of the oxi-
dation products – roughly equivalent to the amount of 6-OOH
– and the remainder of the distribution belonged to 7-OOH.
Upon addition of BDMP, a good H-atom donor, chol 5α-OOH
was also observed. In our preliminary report, we proposed that
this is due to the reversibility of the addition of O2 to the allylic
radical derived from C7-H abstraction. The addition of O2 to
allylic radicals is known to be reversible – the forward reac-
tion is fast (approaching the diffusion limit),37 while the rate
of the reverse reaction, β-fragmentation, varies depending on
the substrate. This has been demonstrated in other lipids, such
as oleate38 and linoleate39 (cf. Scheme 2), whereby kβ is in-
versely proportional to the C-OO• bond strength (i.e. the rela-
tive stability of the peroxyl radical, compared to the alkyl rad-
ical it derives from). In the context of autoxidation, peroxyl
radicals with large kβ may only be observed in the presence of
an antioxidant (AH) that can compete with β-fragmentation
(i.e., kinh [AH] > kβ).
OO
Previous work (Porter):
R2
0.5 s-1
R1
R2
R2
O2
3 s-1
R1
R1
R2
OO
e.g. oleate
R1
OO
R2
30 s-1
O2
R1
R2
R1
1.9 x 106 s-1
R2
R1
e.g. linoleate
400 s-1
O2
O2
R2
R1
R2
R1
OO
This work:
+
Ÿ
HO
slow
HO
HO
The β-fragmentation rate constant, kβ, for 5a-OO was de-
termined to be (3.8 ± 0.9) × 105 s−1 (and ca. 2-fold higher for
cholesterol
3.8 x 105 s-1
slow
slow
O2
O2
the chol OAc-derived peroxyl). We had previously estimated
Ÿ
a value of kβ for 4β-OO to be (8.6 ± 0.4) × 103 s−1, however,
chol 5α-OO
chol 7-OO chol 4-OO
chol 6-OO
due to co-elution with the chol epoxides, the a/b ratio only
appeared to be varying. It is now clear that 4β-OOH is the
dominant product, regardless of [BDMP]. Furthermore, the
ratio of [4-OH]/[6-OH] is unaffected by changing [BDMP],
The fact that the a:b epoxide ratio varied with [BDMP] is
also in line with the mechanism outlined in Scheme 3. Calcu-
lations predict that peroxyl radical addition to chol occurs at
the C6 position on either the a- or b-face (with slight prefer-
ence for the b-face), suggesting subsequent O2 addition would
yield two isomers of 3 (a,a or a,b). Upon reduction to the
corresponding alcohols, the a,b-isomer is poised for displace-
ment of the peroxide moiety, whereas the a,a-isomer is likely
to be relatively stable. Therefore, the suggestion that increas-
ing [BDMP] leads to a larger relative amount of 3 is also con-
sistent with an increase in the ratio of a:b epoxide, though this
also implies the a,a-isomer was unaccounted for in this anal-
ysis.
Ÿ
Ÿ
suggesting the 4-OO and 6-OO do not undergo measurable
β-fragmentation under the reaction conditions. These results
are consistent with computed C-O bond strengths in the 8 per-
Ÿ
Ÿ
oxyl radicals, since the C4-OO and C6-OO are both similar
(26.0 and 24.1 kcal/mol, respectively, for the a-epimers, and
26.0 and 25.2 kcal/mol, respectively, for the β-epimers) and
Ÿ
Ÿ
much stronger than for C5-OO and C7-OO (20.4 and 21.0
kcal/mol, respectively, for the a-epimers, and 17.1 and 21.2
kcal/mol, respectively, for the β-epimers).
Increasing [BDMP] also impacted the peroxyl radical ad-
dition pathway - initially increasing the amount of chol epox-
ides relative to the other products. An explanation for the in-
crease involves a competition between SHi and O2 addition to
the alkyl radical initially formed from peroxyl radical addition
to chol (cf. Scheme 3). The peroxyl radical formed could
Interestingly, the effect of increasing [BDMP] on the rela-
tive amount of epoxide and its stereochemistry differed in
chol OAc autoxidations (cf. Figure 4B). However, since the
workup required treatment with KOH to hydrolyze the ace-
tate, any 3 formed would likely be cleaved by hydroxide to
8
ACS Paragon Plus Environment

Page 9 of 16
Journal of the American Chemical Society
yield chol 3,5,6-triols (chol triol) rather than intramolecular to form exclusively hydroperoxide products through the initial
1
2
3
4
5
6
7
8
displacement to yield chol epoxides. Indeed, we were able to
detect an increased amount of chol triol in select chol OAc
autoxidations relative to the analogous chol experiments.27
Additionally, when chol autoxidations were treated with
KOH, a similar amount of chol triol was present (cf. Figure
S37), and the ratio of a:b epoxide (ca. 0.3) was unaffected by
increasing [BDMP], supporting the fact that KOH could dis-
place the peroxide in 3. Unfortunately, attempts to cleave the
peroxidic bond in 3 by other means (e.g. Zn or Mg) also lead
to the hydrolysis of chol epoxides, complicating the product
analysis.
allylic H-atom abstraction. On the other hand, peroxyl radical
addition to 7-DHC is known to occur, not at C6 as in choles-
terol (to yield the stabilized tertiary alkyl radical), but instead
at C5 – exclusively on the α-face – to yield a stabilized allylic
radical.3 The 5α,6α-epoxide so formed is known to be con-
verted to 3β,5α-dihydroxycholest-7-en-6-one (DHCEO) in
vivo, which has been used as a biomarker in Smith-Lemli-
Opitz syndrome (SLOS). Interestingly, Porter and co-workers
observed a decrease in DHCEO in a mouse model for SLOS
upon administration of antioxidant (H-atom donor).41 Surely
the antioxidant would slow 7-DHC autoxidation in general;
however, should the foregoing results for chol autoxidation
translate to 7-DHC, one might assume the H-atom abstraction
and peroxyl radical addition pathways could be differentially
impacted by the addition of H-atom donor.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 3: Simplified mechanism of cholesterol autoxidation
(black) highlighting the competitive O2 addition after initial
peroxyl radical addition to chol (red).
Scheme 4: Peroxyl radical addition in select other hydrocar-
bons.
ROO
+
RO
OOR
R2
O
ROO
HO
HO
HO
+
RO
O
R1
R2
R1
R1
R2
OOR
R(O)O
O2
e.g. oleate
R(O)OH
ArOH
chol
ROO
HO
3
+
RO
HO
HO
O2
R−H
OO
HOO
HO
7-dehydrocholesterol
HO
OOR
OOR
OOR
O
R
PPh3
OOR
chol-OOH
HO
HO
+
OOR
OH
OH
HO
HO
DHCEO
O
While we were able to identify chol 4-OOH, 6-OOH 7-
OOH and 5,6-epoxides as the primary products of chol autox-
idation (in addition to chol 5a-OOH in the presence of a good
H-atom donor), much of the pathogenesis of lipid autoxida-
tion is attributed to secondary (electrophilic) products. In our
previous work we showed that secosterols 1 and 2 are formed
upon Hock fragmentation of either chol 5a-OOH16 or 6-
OOH.17 While they are likely less electrophilic, chol-derived
ketones (ketochols) may also have pathogenic potential. In the
course of the autoxidation, ketochols may arise from peroxyl-
peroxyl termination, homolysis of hydroperoxides, or H-atom
abstraction from the carbon of a hydroperoxide.42 Interest-
ingly, it appears the three secondary hydroperoxides form ke-
tochols to differing extents – 4-ketochol was not observed un-
der any conditions, suggesting 4-OOH is less prone to the
aforementioned reactions. This could be due to H-bonding be-
tween the C3-OH and the hydroperoxide hindering ketone for-
mation. Indeed, our computations suggested this type of inter-
chol-OOH
+
HO
O
In chol and chol OAc autoxidations alike, the relative
amount of chol epoxide in the product mixture decreased with
increasing [BDMP] – after its initial surge in the former case.
This is likely due to the increasing contribution of the antiox-
idant-mediated peroxidation mechanism, where at high
[BDMP] the predominant chain-carrying radical is no longer
Ÿ
the substrate-derived chol-OO , but instead the antioxidant-
derived phenoxyl radical (which was also suggested to be the
reason for the decrease in 4- and 6-OOH, as previously dis-
cussed).17 Phenoxyl radicals are unlikely to add to a double
bond, as this would be both endothermic and highly reversible
since SHi would not be possible. Therefore, at higher concen-
trations of BDMP the peroxyl radical addition pathway would
give way to an increasing contribution from the C7-H abstrac-
tion pathway. Taken together, the foregoing discussion reiter-
ates that mechanistic generalizations based on the amount or
ratio of the stereoisomeric epoxides is tenuous at best, as they
depend greatly on conditions.
Ÿ
action strengthens the C-OO bond in the chol 4-peroxyl rad-
ical. Conversely, the 6-ketochol was measurable only at low
[BDMP] where the amount formed was above the threshold
of UV detection. Without additive, the amount of 6-ketochol
was ca. 1:1 relative to the amount of 6-OH. This was signifi-
cantly more than the 7-ketochol, which was only ca. 1:5 rela-
tive to 7-OH in the timeframe of our autoxidations. It has been
a great surprise that we detect measurable quantities of chol
4-OOH and 6-OOH, while the former had only been tenuously
tied to chol autoxidation in the past,43 and the latter has not
been detected in vivo, to our knowledge. While analytical
challenges surely contribute, the susceptibility of chol 6-OOH
to Hock fragmentation and significant 6-ketochol formation
A discussion of the interplay between the H-atom abstrac-
tion and peroxyl radical addition pathways in chol would not
be complete without a comparison to other lipids. In particu-
lar, other monounsaturated lipids, such as oleate, are not re-
ported to form epoxides through autoxidation at all.40 Indeed,
the secondary alkyl radical formed upon addition of the per-
oxyl radical is less stable than the tertiary one in chol, presum-
ably so much so that its β-fragmentation is significantly faster
than the SHi reaction (cf. Scheme 4). Instead, oleate is reported
9
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 10 of 16
could mean the concentration of chol 6-O(O)H in vivo is fleet- as opposed to a linoleate-derived peroxyl (right). C7-H ab-
1
2
3
4
5
6
7
8
ing. However, the products so formed are likely significantly
more insidious, suggesting the contribution of the C4-H ab-
straction pathway may be significant contributors to the
pathological potential of chol oxidation products.
straction is shown for example.
We were intrigued by the fact that the computed rate con-
stants for the C7-H-atom abstraction and C6-addition path-
ways were similar (e.g. 24 and 19 M-1s-1, respectively, for
chol). Based on these results, one might expect the experi-
mental chol hydroperoxide/epoxide ratio to be roughly 1:1.
However, the experimentally determined ratio (in the absence
of H-atom donor) was ca. 6:1. Recent reports of tunneling in
hydrocarbon autoxidation prompted us to consider whether
this might underlie the observed preference for H-atom ab-
straction from chol. For example, the propagation rate con-
stants in the autoxidation of linoleic acid44 and tetralin45 are
reported to have a kH/kD of 13 and 16, respectively. Further-
more, the α-tocopherol-loaded autoxidation of 7-dehydrocho-
lesterol has an isotope effect of 21 for H or D transfer to the
chain-carrying tocopheroxyl radical.46 The same experiment
yielded values of 23 and 32 for linoleic acid, α-linolenic acid,
respectively.47 We were able to use the distribution of H-atom
transfer vs. peroxyl radical addition products in chol and
2,2,4,4,7,7-d6-chol autoxidations to estimate the kH/kD for
HAT to a peroxyl radical to be 20 ± 1. Inclusion of corre-
sponding deuterium atoms in our computational model, and
application of Skodje-Truhlar tunneling corrections, the kH/kD
were predicted to be 29.9, 45.7, and 0.992, for C7 H-atom ab-
straction, C4-H atom abstraction and addition to the C5-C6
double bond, respectively. This is fully consistent with our ex-
perimental results, including the fact that the 4-OOH and 6-
OOH products were more abundant than their (classical) re-
action barriers had predicted; the greater contribution from
tunneling evidently enables products derived from C4-H ab-
straction to contribute.
The presence of chol 4-OOH and 6-OOH in chol autoxida-
tions was initially rationalized on the basis of a H-bonding in-
teraction in the H-atom abstraction transition state (cf. Figure
3B). We therefore expected the C4-H abstraction pathway to
be insignificant in chol OAc autoxidations, as this interaction
would no longer exist; however, measurable amounts of 4-
OOH and 6-OOH were nonetheless present. This led us to
consider if another interaction was affecting the analogous
transition state with chol OAc. Interestingly, the two key TSs
for H-atom abstraction from chol OAc were abstraction of the
C7a-H or C4β-H with syn TSs that benefit from secondary
orbital interactions (cf. Figure 4D), and both were lowered by
ca. 1 kcal/mol relative to the analogous TSs in chol. While
this was consistent with the slightly weakened C4-H BDE in
the chol OAc computations, we were perplexed by this sub-
stituent effect. One might expect the C3-OAc to be more in-
ductively withdrawing than the C3-OH, which would both
strengthen the C4-H bond and lower the C5=C6 HOMO, thus
slowing HAT. To the best of our knowledge there are no re-
ports of differing reactivity between chol and chol OAc. We
were therefore surprised to find that the computational predic-
tion was consistent with experimental evidence – the meas-
ured kp was ca. 4-fold higher for chol OAc than for chol. This
implies areas rich in esterified chol, such as low density lipo-
protein (LDL) particles, may be more susceptible to autoxida-
tion than those enriched in free chol, such as the cell mem-
brane. It also suggests that when esterified, chol is about half
as reactive as linoleate!
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
In light of the kp values determined in the peroxyl radical
clock experiments, it is interesting that the total amount of ox-
idation products detected in our chol and chol OAc autoxida-
tions were similar. We believe this could be due to steric hin-
drance in the propagation between two sterol molecules in
chol and chol OAc autoxidations (cf. Figure 6). For example,
the chain length in our chol and chol OAc autoxidations was
~1 in the absence of BDMP, and significantly lower (<0.1) in
its presence (cf. Supporting Information page S38). This sug-
gests propagation is indeed inefficient in our system, and that
the peroxyl radical species abstracting (or adding) is instead
an (unhindered) initiator-derived radical. On the contrary, a
chol-doped methyl linoleate autoxidation involves propaga-
tion between a sterol and a significantly less-hindered PUFA.
We believe the latter to be more biologically relevant, indicat-
ing the reactivity difference between chol and cholOAc is
likely to exist in vivo.
The aforementioned large kH/kD values in PUFA autoxida-
tion are the basis of the surprising protection afforded by the
treatment of cells with deuterated-PUFAs. For example,
Shchepinov and co-workers have demonstrated that the incor-
portation of 11,11-d2-linoleate can protect yeast from lipid au-
toxidation-induced cell death, and surprisingly only 20% deu-
terated linoleate incorporation is required to achieve it.44 This
protective effect was also observed upon administration of
deuterated-PUFAs in a mouse model for Parkinson’s dis-
ease.48 To the best of our knowledge, deuterated sterols have
not been the subject of analogous investigations. Such an in-
vestigation would be interesting, as it may also shed light on
the overall importance of chol oxidation and whether HAT or
addition pathways are more or less damaging to cell mem-
branes.
CONCLUSIONS
We have demonstrated that the products of chol autoxidation
include chol 4-OOH, chol 6-OOH, chol 7-OOH, chol 5,6-
epoxides, 6-ketochol and 7-ketochhol whether cholesterol is
esterified or not. Furthermore, in the presence of a good H-
atom donor, chol 5α-OOH is formed. This product distribu-
tion is far more complex than previously reported, and is in-
fluenced by factors including: the presence of good H-atom
donors, the esterification of the C3-OH, and the contribution
of quantum mechanical tunneling in the H-atom abstraction
pathways. Furthermore, the pathogenesis of chol-derived
oxysterols has, to date, focused on secosterols 1 and 2, prod-
ucts purported to be derived from high energy oxidants like
1O2 and O3. While we have already demonstrated that chol 5-
H
H
H
H
H
H
HO
H
HO
H
O
H
O
vs.
H
H
O
O
H
H
HO
H
O
O
Figure 6: The steric hindrance is suggested to be greater in
the H-atom abstraction of chol by a chol-derived peroxyl (left)
10
ACS Paragon Plus Environment

Page 11 of 16
Journal of the American Chemical Society
OOH and 6-OOH undergo Hock fragmentation to yield 1 and 50.2, 44.2, 43.2, 39.0, 37.6, 36.7, 32.5, 32.05, 31.98, 31.7,
1
2
3
4
5
6
7
8
2, the complex product distribution of chol autoxidation pre-
sented herein begs the question whether there are other hith-
erto unidentified pathogenic compounds that may derive
from, for example, chol 4-OOH or 7-OOH. While the epox-
ides were previously known to be autoxidation products, their
quantification has been complicated by the fact that they are
not unique to autoxidation. Others have advocated for the use
of the a:b ratio of epoxides as an indicator of the oxidant they
were derived from; however, we have shown that the presence
of H-atom donors in the autoxidation medium can affect this
ratio, suggested that this practice should be avoided, or at the
very least, that interpretations of these data be made with cau-
tion.
24.7, 23.0, 21.2, 19.6, 18.3, 17.9, 13.4, 12.5; HRMS (EI): cal-
culated for (loss of isopropyl from TIPS) C30H52O2Si
429.31833, found 429.32169.
3β-((Triisopropylsilyl)oxy)-pregn-5-en-20-one-17,21,21,21-
d4 (5). To 4 (2.0 g, 4.65 mmol) in 100 mL dry dioxane was
added 7.5 mL D2O and 2.5 mL sodium deuteroxide (40% so-
lution in D2O) in a flask fitted with a reflux condenser. The
reaction was heated to 110oC for 24 hours under nitrogen, then
cooled to room temperature. The solvent was removed in
vacuo, and the residue taken up in ether (100 mL). The solu-
tion was washed successively with 100 mL each of H2O, 1M
HCl, NaHCO3 (aqueous, saturated), and then brine. The or-
ganic layer was dried over MgSO4 and concentrated in vacuo.
The product was used without further purification. White
waxy solid; 99% yield. 1H-NMR (400 MHz; CDCl3): δ 5.30
(d, J = 5.2 Hz, 1H), 3.55 (tdd, J = 10.3, 6.1, 4.1 Hz, 1H), 2.28-
2.25 (m, 2H), 2.19-2.13 (m, 1H), 2.04-1.97 (m, 2H), 1.83-1.79
(m, 2H), 1.69-1.43 (m, 9H), 1.27-1.12 (m, 2H), 1.05-1.04 (m,
21H), 1.02 (s, 3H), 0.98-0.89 (m, 1H), 0.62 (s, 3H); 13C-NMR
(100 MHz; CDCl3): δ 209.9, 141.8, 120.9, 72.5, 57.1, 50.2,
44.1, 43.2, 39.0, 37.6, 36.7, 32.5, 32.21, 32.02, 31.96, 24.7,
22.8, 21.2, 19.6, 18.2, 17.8, 13.4, 12.5; HRMS (EI): calculated
for (loss of isopropyl from TIPS) C27H41D4O2Si+ 433.34344,
found 433.34469.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
EXPERIMENTAL SECTION
All chemicals and solvents were purchased from Sigma Al-
drich Co. LLC and used as received, unless otherwise stated.
Cholesterol (92%) was purified to >99% (NMR purity) by
Fieser’s bromination method to remove other sterol impuri-
ties.49 The >99% cholesterol was then freshly recrystallized
from methanol prior to each experiment to remove trace oxi-
dation product impurities. 2,2'-Azobis (4-methoxy-2,4-dime-
thylvaleronitrile) (MeOAMVN)50 was purchased from Wako
Pure Chemical Industries, Ltd. and used as received. 2,6-Di-
tert-butyl-4-methylphenol (BHT) was recrystallized from
hexane prior to use. 3-Chloroperbenzoic acid (mCPBA,
≤77%) was recrystallized from CH2Cl2 prior to use.
3β-((Triisopropylsilyl)oxy)-cholest-5-en-20-ol-17,21,21,21-
d4 (6). Magnesium powder (306 mg, 12.6 mmol) was taken up
in 5 mL dry THF in a round bottom flask fitted with a reflux
condenser. A few drops of anhydrous HCl (4.0 M in dioxane)
were added and the mixture was heated to 70oC. The magne-
sium was activated by adding a solution (~10 mg/mL) of io-
dine in dry THF in three 300µL portions. 1-bromo-4-
methylpentane (920 µL, 6.3 mmol) was then added dropwise
in 5 mL dry THF at room temperature, and the mixture was
refluxed for 2 hours then cooled to room temperature. 5 (500
mg, 1.05 mmol) was taken up in 5 mL dry THF and the Gri-
gnard reagent was added dropwise at room temperature. The
reaction was refluxed overnight, then cooled to room temper-
ature. 15 mL of NH4Cl (aqueous, saturated) was added and
the mixture was stirred for 30 minutes, then extracted with 3
´ 15 mL EtOAc. The combined organic extracts were washed
with brine, then dried over MgSO4 and concentrated in vacuo.
The resultant oil was purified by flash column on silica gel
(9:1 hexanes:ethyl acetate) to yield a clear oil; 84% yield. 1H-
NMR (400 MHz; CDCl3): δ 5.31 (d, J = 5.0 Hz, 1H), 3.56
(tdd, J = 10.3, 6.1, 4.1 Hz, 1H), 2.31-2.21 (m, 2H), 2.08 (dt, J
= 12.4, 3.4 Hz, 1H), 2.00-1.93 (m, 1H), 1.83-1.42 (m, 13H),
1.33-1.11 (m, 9H), 1.05 (s, 21H), 1.01 (s, 3H), 0.89-0.86 (m,
9H); 13C-NMR (100 MHz; CDCl3): δ 141.8, 121.1, 75.2, 72.6,
57.1, 50.3, 44.2, 43.2, 42.7, 40.2, 39.8, 37.6, 36.7, 32.5, 32.0,
31.5, 28.20, 28.09, 24.0, 22.88, 22.72, 22.4, 22.1, 21.1, 19.6,
18.3, 13.7, 12.5; HRMS (EI): calculated for (loss of isopropyl
from TIPS) C33H55D4O2Si+ 519.45299, found 519.45353.
Synthesis. 4α-Hydroxycholesterol (chol 4α-OH), 4β-hy-
droxycholesterol (chol 4β-OH), 5α-hydroxycholesterol (chol
5α-OH), 6α-hydroxycholesterol (chol 6α-OH), 6β-hydroxy-
cholesterol (chol 6β-OH), and 7α-hydroxycholesterol (chol
7α-OH), 7β-hydroxycholesterol (chol 7β-OH), and 4-(tert-bu-
tyl)-2,6-dimethylphenol (BDMP) were prepared as previously
reported.17
Cholesterol-5α,6α-epoxide (chol α-epoxide) and cholesterol
5β,6β-epoxide (chol β-epoxide). mCPBA (485 mg, 2.8 mmol)
was added to 1.0 g cholesterol (2.6 mmol) in 25 mL dry
CH2Cl2 and stirred for 6 hours under nitrogen. Aqueous, satu-
rated NaHCO3 (25 mL) was added and the mixture was stirred
for 10 minutes. The organic layer was washed with water then
brine, dried overMgSO4 and concentrated in vacuo. The crude
solid was purified by flash column on silica gel (7:3 hex-
anes:EtOAc) to give a mixture of chol α- and β-epoxide, with
a ratio of 1:0.3, as a white solid; 94% yield. Spectroscopic
characterization was consistent with literature precedent.51
3β-((Triisopropylsilyl)oxy)-pregn-5-en-20-one (4). Pregn-5-
en-3β-ol-20-one (pregnenolone, 1.5 g, 4.7 mmol) was dis-
solved in 40 mL dimethylformamide and 10 mL dichloro-
methane, and imidazole (1.2 g, 17.6 mmol) and chlorotriiso-
propylsilane (1.7 mL, 7.8 mmol) were added. The reaction
was stirred at room temperature under nitrogen for 4 hours.
The reaction was taken up in ethyl acetate (50 mL) and
washed with NaHCO3 (aqueous, saturated, 2 x 50 mL) then
brine (50 mL), dried over MgSO4 and concentrated in vacuo.
The resultant oil was purified by flash column on silica gel
(9:1 hexanes:ethyl acetate) to yield a white waxy solid; 97%
yield. 1H-NMR (400 MHz; CDCl3): δ 5.31 (d, J = 5.2 Hz, 1H),
3.56 (tdd, J = 10.3, 6.1, 4.2 Hz, 1H), 2.53 (t, J = 9.0 Hz, 1H),
2.28-2.16 (m, 3H), 2.12 (s, 3H), 2.06-1.96 (m, 2H), 1.84-1.79
(m, 2H), 1.72-1.44 (m, 9H), 1.28-1.13 (m, 2H), 1.06-1.05 (m,
21H), 1.00 (s, 3H), 0.98-0.89 (m, 1H), 0.63 (s, 3H); 13C-NMR
(100 MHz; CDCl3): δ 209.7, 141.8, 120.9, 72.6, 63.9, 57.1,
3β-((Triisopropylsilyl)oxy)-cholesta-5,20-diene-17,21,21,21-
d4 (7). Compound 6 (400 mg, 0.72 mmol) was dissolved in 20
mL dry pyridine and cooled to 0oC. POCl3 (110 µL, 1.2 mmol)
was added dropwise, and the reaction was allowed to come to
room temperature and stirred for 2 hours. Additional portions
of POCl3 (110 µL) were added at 0oC until no starting material
remained by TLC (19:1 hexanes:ethyl acetate, where the Rf of
6 and 7 were 0.2 and 0.9, respectively). The reaction was
quenched at 0oC by dropwise addition of methanol (2 mL),
then water (20 mL), then extracted twice with 50 mL ether.
11
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 12 of 16
The combined ether extracts were washed twice with 0.1 M 28.2, 23.95, 23.86, 23.0, 22.90, 22.7, 21.2, 14.8, 12.3; HRMS
1
2
3
4
5
6
7
8
HCl (50 mL), then with 50 mL NaHCO3 (aqueous, saturated),
then 50 mL brine, dried over MgSO4 and concentrated in
vacuo. This reaction is known to give a mixture of three elim-
ination products (purported to be 55% ∆20(22), 25% ∆20 and
20% ∆17(20) when carried out on a similar substrate),24 there-
fore, the resultant oil was purified by flash column on silica
gel impregnated with 10% silver nitrate (10:1 hexanes:tolu-
ene) to yield the ∆20(22) product 7 as a clear oil; 48% yield. 1H-
NMR (400 MHz; CDCl3): δ 5.32 (d, J = 5.1 Hz, 1H), 5.17 (t,
J = 7.0 Hz, 1H), 3.56 (tdd, J = 10.4, 6.0, 4.3 Hz, 1H), 2.32-
2.22 (m, 2H), 2.05-1.95 (m, 3H), 1.84-1.77 (m, 4H), 1.70-1.38
(m, 9H), 1.25-1.12 (m, 5H), 1.06 (s, 21H), 1.01 (s, 3H), 0.88
(d, J = 6.6 Hz, 6H), 0.54 (s, 3H); 13C-NMR (100 MHz;
CDCl3): δ 141.9, 134.0, 125.9, 121.2, 77.2, 72.6, 56.5, 50.6,
43.5, 43.3, 39.4, 38.8, 37.6, 36.8, 32.53, 32.37, 32.1, 27.9,
26.1, 24.7, 24.5, 22.78, 22.73, 21.2, 19.7, 18.3, 13.0, 12.5;
HRMS (EI): calculated for (loss of isopropyl from TIPS)
C33H53D4OSi+ 501.44243, found 501.44429.
(EI): calculated for C27H40D6O2 408.38744, found 408.38674.
Cholesterol-5,6-epoxide-17,20,21,21,21,22-d6. The title com-
pound was prepared from 8 (50 mg, 0.13 mmol scale) using
the same procedure as detailed above for the synthesis of the
chol epoxides to give a mixture of d6-chol α- and β-epoxide,
in a ratio of 4:1, as a white solid; 89% yield. Characterization
was consistent with that previously reported51 for the protiated
analogues (chol epoxides). Cholesterol-5α,6α-epoxide-
17,20,21,21,21,22-d6: 1H-NMR (400 MHz; CDCl3): δ 3.91 (tt,
J = 10.9, 5.3 Hz, 1H), 2.90 (d, J = 4.7 Hz, 1H), 2.07 (t, J =
12.0 Hz, 1H), 1.95-1.87 (m, 3H), 1.82-1.10 (m, 23H), 1.06 (s,
3H), 1.02-0.92 (m, 3H), 0.86 (d, J = 6.6 Hz, 6H), 0.60 (s, 3H);
13C-NMR (75 MHz; CDCl3): δ 68.9, 65.8, 59.4, 57.0, 42.7,
42.4, 40.0, 39.48, 39.42, 35.0, 32.5, 31.3, 30.0, 29.0, 28.2,
28.1, 24.2, 24.1, 23.0, 22.7, 20.8, 16.1, 12.0; HRMS (EI): cal-
culated for C27H40D6O2 408.38744, found 408.38852. Choles-
terol-5β,6β-epoxide-17,20,21,21,21,22-d6: 3.70 (tt, J = 10.7,
5.3 Hz, 1H), 3.06 (d, J = 2.5 Hz, 1H), 2.07 (t, J = 12.0 Hz,
1H), 1.95-1.87 (m, 3H), 1.82-1.10 (m, 23H), 1.06 (s, 3H),
1.02-0.92 (m, 3H), 0.86 (d, J = 6.6 Hz, 6H), 0.63 (s, 3H); 13C-
NMR (75 MHz; CDCl3): δ 69.6, 63.9, 63.1, 56.4, 51.5, 42.7,
42.4, 39.6, 39.5, 37.4, 35.0, 32.8, 31.2, 29.9, 28.2, 27.9, 24.0,
23.9, 22.9, 22.8, 22.2, 17.2, 11.9; HRMS (EI): calculated for
C27H40D6O2 408.38744, found 408.38852.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Cholesterol-17,20,21,21,21,22-d6 (8). Compound 7 (300 mg,
0.6 mmol) was dissolved in 15 mL dry dioxane and 0.3 mL
AcOD, and platinum oxide (25 mg, 0.09 mmol) was added.
The flask was briefly placed under vacuum, then affixed with
a balloon containing approximately 30 mL D2. The reaction
was stirred at room temperature overnight, passed through a
1
pad of celite, then concentrated in vacuo. Crude H NMR
4-Ketocholesterol. To a solution of 4-ketocholesteryl acetate
(150 mg, 0.34 mmol, prepared as previously reported)17,53 in
6 mL MeOH was added KOH (340 mg, 6 mmol), and the mix-
ture was stirred for 3 hours. CH2Cl2 (6 mL) was added, and
the organic layer separated and washed with 1M HCl, then
NaHCO3 (aq. sat.), then brine, dried over MgSO4 and concen-
trated in vacuo. The crude solid was purified by silica gel flash
column (hexanes), and portions were further purified by pre-
parative TLC (hexanes); white solid, 78% yield, based on the
isolated inseparable mixture of 4-ketocholesterol and 4-hy-
droxycholest-4-en-3-one (characterization of the latter has
showed disappearance of the methylene proton at 5.17 ppm.
The resultant clear oil was taken up in 15 mL THF and TBAF
(1 M in THF) was added in 1 mL portions until the starting
material was consumed by TLC. The mixture was taken up in
ether (50 mL), washed with 50 mL water, then 50 mL brine,
dried over MgSO4, and concentrated in vacuo. The crude solid
was purified by flash column on silica gel (9:1 hexanes:ethyl
acetate) to yield a white solid; 84% yield. Note that in the 13C
NMR, D-bearing carbons (55.4, 35.9, 35.1, and 17.8 ppm) are
barely visible due to D-coupling, and are therefore not re-
1
ported. H-NMR (400 MHz; CDCl3): δ 5.34 (d, J = 5.2 Hz,
1
been reported54). 4-Ketocholesterol: H-NMR (600 MHz;
1H), 3.51 (tt, J = 10.6, 5.1 Hz, 1H), 2.31-2.20 (m, 2H), 2.02-
1.93 (m, 2H), 1.85-1.77 (m, 3H), 1.70 (s, 1H) 1.55-1.43 (m,
7H), 1.25-1.11 (m, 8H), 1.00 (s, 3H), 0.95-0.89 (m, 2H), 0.86
(dd, J = 6.6, 1.9 Hz, 6H), 0.57 (s, 3H); 13C-NMR (100 MHz;
CDCl3): δ 140.9, 121.8, 71.9, 56.9, 50.3, 42.4, 42.3, 39.8,
39.6, 37.4, 36.6, 32.05, 32.03, 31.8, 28.2, 28.0, 24.4, 23.9,
23.0, 22.7, 21.2, 19.5, 12.0; HRMS (EI): calculated for
C27H40D6O 392.39253, found 392.39493.
CDCl3): δ 6.91 (dd, J = 5.2, 2.7 Hz, 1H), 6.07 (br, 1H), 6.03
(dd, J = 6.8, 2.8 Hz, 1H), 2.55-2.41 (m, 1H), 2.33-2.26 (m,
1H), 2.07-1.96 (m, 2H), 1.90-1.78 (m, 2H), 1.68-1.59 (m, 2H),
1.58-1.46 (m, 4H), 1.42-1.21 (m, 8H), 1.12 (s, 3H), 1.09-0.93
(m, 6H), 0.91 (dd, J = 10.5, 6.6 Hz, 3H), 0.86 (dt, J = 4.3, 2.3
Hz, 6H), 0.69 (s, 3H); 13C-NMR (150 MHz; CDCl3): δ 184.5,
146.4, 137.2, 115.4, 56.6, 56.2, 49.3, 42.5, 39.64, 39.64,
39.61, 37.9, 36.3, 35.9, 32.2, 31.2, 28.3, 28.2, 28.2, 24.3, 24.0,
23.0, 22.7, 21.9, 20.9, 18.9, 12.0; HRMS (EI): calculated for
C27H44O2 400.33413, found 400.33523.
5α-Hydroxycholesterol-17,20,21,21,21,22-d6. The title com-
pound was prepared from 8 in a manner analogous to the pre-
viously reported52 synthesis of chol 5α-OH from chol: A so-
lution of 8 (30 mg, 0.08 mmol) and Rose Bengal (1 mg, 0.001
mmol) in 2 mL pyridine was irradiated with a 400 W Na lamp
at a distance of 5 cm at 0°C with constant O2 bubbling for 6
hours. The pyridine was removed in vacuo and the residue
taken up in 2 mL EtOH. PPh3 (22 mg, 0.08 mmol) was added
and the solution was stirred for 2 hours, then concentrated in
vacuo. The residue was purified by flash column on silica gel
(7:3 hexanes:ethyl acetate) to yield a white solid; 73% yield.
1H-NMR (600 MHz; CDCl3): δ 5.63 (dd, J = 9.83, 1.83 Hz,
1H), 5.57 (dd, J = 9.82, 2.70 Hz, 1H), 4.12 (m, 1H), 2.02 (m,
1H), 1.94-1.78 (m, 4H), 1.65-1.56 (m, 3H), 1.51-0.97 (m, 17
H), 0.92 (s, 3H), 0.86 (dt, J = 4.30, 2.34 Hz, 6H), 0.70 (s, 3H);
13C-NMR (150 MHz; CDCl3): δ 133.34, 133.32, 74.1, 67.3,
54.0, 45.2, 43.7, 41.0, 40.1, 39.6, 38.6, 38.2, 30.7, 28.7, 28.3,
6-Ketocholesterol. Cholest-4-en-3,6-dione (100 mg, 0.25
mmol, prepared as previously reported)17,55 was taken up in 2
mL dry THF and 1 mL dry EtOH, and CeCl3x7H2O (100 mg,
0.27 mmol) was added. The mixture was cooled to 0°C and
NaBH4 (3.2 mg, 0.085 mmol) was added. The reaction was
allowed to reach room temperature, then stirred for 1 hour and
quenched with 2 mL water. The mixture was extracted with
ether, washed with brine, dried over MgSO4, and concentrated
in vacuo. The crude solid was purified by flash column on sil-
ica gel (7:3 hexanes:ethyl acetate); white solid, 45% yield.
The compound matched previous characterization,56,57 though
we report 13C NMR and a more complete 1H NMR here: 1H-
NMR (400 MHz; CDCl3): δ 6.17 (d, J = 1.24 Hz, 1H), 4.32
(dt, J = 10.8, 5.2 Hz, 1H), 2.46-2.30 (m, 2H), 2.17 (ddd, J =
11.9, 5.6, 3.4 Hz, 1H), 2.06-1.96 (m, 3H), 1.89-1.46 (m, 7H),
1.43-1.24 (m, 7H), 1.18 (s, 3H), 1.13-1.00 (m, 7H), 0.91 (d, J
12
ACS Paragon Plus Environment

Page 13 of 16
Journal of the American Chemical Society
40.0, 39.6, 37.2, 36.3, 35.9, 34.9, 29.7, 28.3, 28.1, 24.3, 24.0,
= 6.6 Hz, 3H), 0.86 (dd, J = 6.6, 1.9 Hz, 6H), 0.70 (s, 3H);
13C-NMR (100 MHz; CDCl3): δ 199.7, 171.9, 119.8, 68.8,
56.2, 55.7, 53.9, 42.6, 41.6, 39.62, 39.61, 39.2, 36.4, 36.2,
35.9, 34.3, 33.9, 28.3, 28.2, 24.3, 23.9, 23.0, 22.7, 21.2, 18.8,
18.4, 12.1.
23.0, 22.7, 22.1, 18.8, 17.2, 11.9; HRMS (EI): calculated for
C27H40D6O2 408.38744, found 408.38675.
1
2
3
4
Autoxidations. Cholesterol autoxidations. To 19 mg choles-
terol (0.05 mmol) in a HPLC vial was added H-atom donor
(BDMP or PMC) in a stock solution, and enough chloroben-
zene to make up the volume to 90 µL. MeOAMVN (10 µL of
a 200 mM solution in benzene) was added. The vial was
capped with a preslit HPLC cap, and incubated at 37°C for 16
hours. The solution was quenched by adding 760 µL of a so-
lution of 6.5 mM BHT and 6.5 mM PPh3 in HPLC grade hex-
ane, and the following internal standards were added as 1 mM
solutions in chlorobenzene: 20 µL of d6-5α-hydroxycholes-
terol, 20 µL of d6-cholesterol-5,6-epoxide, and 100 µL of 4-
nitrophenol.
7-Ketocholesterol. Cholesterol (500 mg, 1.3 mmol) and N-hy-
droxyphthalimide (30 mg, 0.2 mmol) were dissolved in PhCl
(25 mL) and azobisisobutyronitrile (20 mg, 0.1 mmol) was
added. The reaction was stirred open to air at 60°C for 72
hours, then PPh3 (250 mg, 1 mmol) was added. The crude
mixture was concentrated in vacuo and purified by silica gel
flash column (gradient from 9:1 to 7:3 hexanes:ethyl actate)
to yield a white solid; 13% yield. Spectroscopic characteriza-
tion was consistent with literature precedent.58
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Cholestane-3β,5a,6β-triol (chol triol). Cholesterol epoxide
(200 mg, 0.5 mmol) was dissolved in 7 mL of acetone:water
(10:1) in an ice bath and 60 µL of 70% perchloric acid was
added. The reaction was warmed to room temperature and
stirred for 2 hours. NaOH (50%, aqueous) was added until the
pH was neutral, and the acetone was removed in vacuo. The
crude mixture was taken up in 10 mL ether, washed with brine
(3 x 10 mL), dried over MgSO4 and concentrated in vacuo.
The crude solid was purified by flash column on silica gel
(gradient from 1:1 EtOAc:hexanes to EtOAc) to give a white
solid; 86% yield. Spectroscopic characterization was con-
sistent with literature precedent.59
Cholesteryl acetate autoxidations. To 22 mg cholesteryl ace-
tate (0.05 mmol) in a vial was added BDMP in a stock solu-
tion, and enough chlorobenzene to make up the volume to 90
µL. MeOAMVN (10 µL of a 200 mM solution in benzene)
was added. The vial was capped loosely, and incubated at
37°C for 16 hours. The solution was quenched by adding 100
µL of 0.2 M BHT and PPh3 in PhCl, then 2 mL of 1.0 M KOH
in MeOH was added and the vial was shaken occasionally
over 3 hours. 6 mL CH2Cl2 was added and the mixture was
washed twice with 6 mL H2O, dried over MgSO4, concen-
trated in vacuo. The residue was taken up in 860 µL of HPLC
grade hexane, and the following internal standards were added
as 1 mM solutions in chlorobenzene: 20 µL of d6-5α-hydroxy-
cholesterol, 20 µL of d6-cholesterol-5,6-epoxide, and 100 µL
of 4-nitrophenol.
Cholestane-3β,5a,6β-triol-17,20,21,21,21,22-d6. The title
compound was prepared from cholesterol-5,6-epoxide-
17,20,21,21,21,22-d6 (10 mg, 0.025 mmol scale) using the
same procedure as above for chol triol; white solid, 91% yield.
1H-NMR (600 MHz; CDCl3): δ 4.10 (tt, J = 10.9, 5.5 Hz, 1H),
3.54 (t, J = 2.9 Hz, 1H), 2.08 (dd, J = 12.9, 11.2 Hz, 1H), 2.00-
1.94 (m, 1H), 1.87-1.69 (m, 3H), 1.63-1.48 (m, 7H), 1.43-1.22
(m, 8H), 1.18 (s, 3H), 1.15-0.95 (m, 6H), 0.86 (dd, J = 6.6, 2.6
Hz, 6H), 0.67 (s, 3H) ; 13C-NMR (150 MHz; CDCl3): δ 76.3,
76.2, 67.8, 56.1, 46.1, 42.8, 40.9, 40.0, 39.6, 38.5, 34.7, 32.5,
31.0, 30.3, 28.2, 28.0, 24.3, 23.9, 23.0, 22.7, 21.3, 17.1, 12.3;
Cholesterol-2,2,4,4,7,7-d6 autoxidations. Cholesterol-d6 (20
mg) was treated as in the cholesterol autoxidation above, with-
out additive, and 5α-hydroxycholesterol and cholesterol-5,6-
epoxide were used as internal standards rather than the d6-an-
alogues.
HPLC-APCI-MS/MS Analyses. From each autoxidation
(unless otherwise noted), 10 µL of the sample was injected
onto a normal phase HPLC column (SunFire 5 µm silica; 4.6
mm × 250 mm) and eluted with the following gradient of hex-
ane:iso-propanol (i-PrOH) at a flow rate of 1.5 mL/min: 99:1
to 98:2 over 30 minutes, 98:2 from 30 to 40 minutes, 98:2 to
95:5 from 40 to 41 minutes, 95:5 from 41 to 49 minutes, 95:5
to 99:1 from 49 to 50 minutes, and 99:1 from 50 to 60 minutes.
APCI-MS/MS detection in positive ion mode was used for the
MS/MS transitions: 403.4 m/z à 385.4 m/z and 385.4 m/z à
367.4 m/z, for chol epoxides and alcohols, respectively; as
well as 409.4 m/z à 391.4 m/z and 391.4 m/z à 373.4 m/z,
for d6-chol epoxides and alcohols, respectively. The settings
for mass spectral analysis were as follows: corona current 2
µA; cone voltage 25 V; source temperature 150°C; APCI
probe temperature 500°C; desolvation gas flow 600 L/hr;
cone gas flow 20 L/hr; collision gas flow 0.2 mL/min; colli-
sion energy 15 eV. Simultaneous UV detection was carried
out by photodiode array, and the 6- and 7-ketochols were de-
tected at 234 nm relative to 4-nitrophenol as internal standard.
+
HRMS (EI): calculated for (M+-H2O) C27H41D6O2
409.39472, found 408.39340.
Cholesterol-5,6-epoxide-2,2,4,4,7,7-d6. The title compound
was prepared from cholesterol-2,2,4,4,7,7-d6 (prepared as pre-
viously reported)22 (20 mg, 0.05 mmol scale) using the same
procedure as above for the chol epoxides, to give a mixture of
d6-chol α- and β-epoxide, with a ratio of 4:1, as a white solid;
86% yield. Spectroscopic characterization was consistent with
literature precedent for the protiated analogues (chol epox-
ides).51 Cholesterol-5α,6α-epoxide-2,2,4,4,7,7-d6: 1H-NMR
(600 MHz; CDCl3): δ 3.89 (s, 1H), 2.89 (s, 1H), 1.96-1.93 (m,
1H), 1.84-1.77 (m, 1H), 1.67 (d, J = 13.52, 1H), H (br, 1H),
1.57-1.48 (m, 3H), 1.38-1.28 (m, 6H), 1.25-1.20 (m, 3H),
1.14-1.06 (m, 4H), 1.05 (s, 3H), 1.0-0.93 (m, 3H), 0.88 (d, J
= 6.6 Hz, 3H), 0.85 (dd, J = 6.6, 3.0 Hz, 6H), 0.60 (s, 3H);
13C-NMR (150 MHz; CDCl3): δ 68.7, 65.7, 59.3, 56.9, 56.0,
42.6, 42.5, 39.63, 39.56, 36.3, 35.9, 34.9, 32.4, 29.8, 28.2,
28.1, 24.2, 24.0, 23.0, 22.7, 20.8, 18.8, 16.1, 12.0; HRMS
(EI): calculated for C27H40D6O2 408.38744, found 408.38675.
Cholesterol-5β,6β-epoxide-2,2,4,4,7,7-d6: 1H-NMR (600
MHz; CDCl3): δ 3.68 (s, 1H), 3.04 (s, 1H), 1.96-1.93 (m, 1H),
1.84-1.77 (m, 1H), 1.67 (d, J = 13.52, 1H), H (br, 1H), 1.57-
1.48 (m, 3H), 1.38-1.28 (m, 6H), 1.25-1.20 (m, 3H), 1.14-1.06
(m, 4H), 1.05 (s, 3H), 1.0-0.93 (m, 3H), 0.88 (d, J = 6.6 Hz,
3H), 0.85 (dd, J = 6.6, 3.0 Hz, 6H), 0.63 (s, 3H); 13C-NMR
(150 MHz; CDCl3): δ 69.4, 63.7, 63.0, 56.3, 56.3, 51.4, 42.4,
ASSOCIATED CONTENT
Supporting Information
Additional experimental details, NMR spectra, computational de-
tails, and energies and Cartesian coordinates of computed struc-
tures. The Supporting Information is available free of charge on
the ACS Publications website.
13
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 14 of 16
Annu. Rev. Pharmacol. Toxicol. 2016, 56, 447; (c) Guillemot-
AUTHOR INFORMATION
1
2
3
4
5
6
7
8
Legris, O.; Mutemberezi, V.; Muccioli, G. G. Oxysterols in Met-
abolic Syndrome: From Bystander Molecules to Bioactive Li-
pids. Trends Mol. Med. 2016, 22, 594; (d) Mutemberezi, V.; Guil-
lemot-Legris, O.; Muccioli, G. G. Oxysterols: From cholesterol
metabolites to key mediators. Prog. Lipid Res. 2016, 64, 152; (e)
Doria, M.; Maugest, L.; Moreau, T.; Lizard, G.; Vejux, A. Con-
tribution of cholesterol and oxysterols to the pathophysiology of
Parkinson's disease. Free Radic. Biol. Med. 2016, 101, 393; (f)
Kloudova, A.; Guengerich, F. P.; Soucek, P. The Role of Oxys-
terols in Human Cancer. Trends Endocrinol. Metab. 2017, 28,
485; (g) Gargiulo, S.; Testa, G.; Gamba, P.; Staurenghi, E.; Poli,
G.; Leonarduzzi, G. Oxysterols and 4-hydroxy-2-nonenal con-
tribute to atherosclerotic plaque destabilization. Free Radic. Biol.
Med. 2017, 111, 140; (h) Mutemberezi, V.; Buisseret, B.;
Masquelier, J.; Guillemot-Legris, O.; Alhouayek, M.; Muccioli,
G. G. Oxysterol levels and metabolism in the course of neuroin-
flammation: insights from in vitro and in vivo models. J. Neu-
roinflammation 2018, 15, 74; (i) Testa, G.; Rossin, D.; Poli, G.;
Biasi, F.; Leonarduzzi, G. Implication of oxysterols in chronic
inflammatory human diseases. Biochimie 2018, 153, 220; (j) Be-
zine, M.; Namsi, A.; Sghaier, R.; Ben Khalifa, R.; Hamdouni, H.;
Brahmi, F.; Badreddine, I.; Mihoubi, W.; Nury, T.; Vejux, A.;
Zarrouk, A.; de Sèze, J.; Moreau, T.; Nasser, B.; Lizard, G. The
effect of oxysterols on nerve impulses. Biochimie 2018, 153, 46;
and references cited therein.
(11) Takeuchi, C.; Galvé, R.; Nieva, J.; Witter, D. P.; Wentworth,
A. D.; Troseth, R. P.; Lerner, R. A.; Wentworth, P., Jr. Proather-
ogenic Effects of the Cholesterol Ozonolysis Products, Athero-
nal-A and Atheronal-B. Biochemistry 2006, 45, 7162.
(12) Nieva, J.; Song, B.-D.; Rogel, J. K.; Kujawara, D.; Altobel,
L., III; Izharrudin, A.; Boldt, G. E.; Grover, R. K.; Wentworth, A.
D.; Wentworth, P., Jr. Cholesterol Secosterol Aldehydes Induce
Amyloidogenesis and Dysfunction of Wild-Type Tumor Protein
p53. Chem. Biol. 2011, 18, 920.
(13) Usui, K.; Hulleman, J. D.; Paulsson, J. F.; Siegel, S. J.; Pow-
ers, E. T.; Kelly, J. W. Site-specific modification of Alzheimer’s
peptides by cholesterol oxidation products enhances aggregation
energetics and neurotoxicity. Proc. Natl. Acad. Sci. USA 2009,
106, 18563.
(14) Bosco, D. A.; Fowler, D. M.; Zhang, Q. H.; Nieva, J.; Pow-
ers, E. T.; Wentworth, P., Jr.; Lerner, R. A.; Kelly, J. W. Elevated
levels of oxidized cholesterol metabolites in Lewy body disease
brains accelerate a-synuclein fibrilization. Nat. Chem. Biol. 2006,
2, 249.
Corresponding Author
* dpratt@uottawa.ca
ACKNOWLEDGEMENT
This work was supported by grants from the Natural Sciences and
Engineering Research Council (NSERC) of Canada and the Can-
ada Foundation for Innovation. D.A.P. and Z.A.M.Z. also
acknowledge the support of the Canada Research Chairs and the
NSERC Post-Graduate Scholarships programs, respectively.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ABBREVIATIONS
7-DHC, 7-dehydrocholesterol; AH, antioxidant; BDMP, 4-tert-
butyl-2,6-dimethylphenol; chol, cholesterol; chol a-epoxide,
5a,6a-epoxycholesterol; chol b-epoxide, 5b,6b-epoxycholes-
terol; chol OAc, cholesteryl acetate; chol triol, 5,6-dihydroxycho-
lesterol; d6-chol, cholesterol-17,20,21,21,21,22-d6; DHCEO,
3β,5α-dihydroxycholest-7-en-6-one; HOMO, highest occupied
molecular orbital; kb, b-fragmentation rate constant; kHAT, H-
atom transfer rate constant; kinh, inhibition rate constant; kp, prop-
agation rate constant; KIE, kinetic isotope effect; LDL, low den-
sity lipoprotein; MeOAMVN, 2,2'-Azobis (4-methoxy-2,4-dime-
thylvaleronitrile); PMC, 2,2,5,7,8-Pentamethyl-6-chromanol;
PUFA, polyunsaturated fatty acid; SHi, intramolecular homolytic
substitution; SLOS, Smith-Lemli-Opitz syndrome; TS, transition
state.
REFERENCES
(1) Ullery, J. C.; Marnett, L. J. Protein modification by oxidized
phospholipids and hydrolytically released lipid electrophiles: In-
vestigating cellular responses. Biochim. Biophys. Acta,
Biomembr. 2012, 1818, 2424.
(2) Xu, L.; Porter, N. A. Reactivities and Products of Free Radical
Oxidation of Cholestadienols. J. Am. Chem. Soc. 2014, 136,
5443.
(3) Xu, L.; Korade, Z.; Porter, N. A. Oxysterols from Free Radi-
cal Chain Oxidation of 7-Dehydrocholesterol: Product and Mech-
anistic Studies. J. Am. Chem. Soc. 2010, 132, 2222.
(4) Xu, L.; Davis, T. A.; Porter, N. A. Rate Constants for Peroxi-
dation of Polyunsaturated Fatty Acids and Sterols in Solution and
in Liposomes. J. Am. Chem. Soc. 2009, 131, 13037.
(5) van Meer, G.; Voelker, D. R.; Feigenson, G. W. Membrane
lipids: where they are and how they behave. Nat. Rev. Mol. Cell
Biol. 2008, 9, 112.
(6) Björkhem, I.; Meaney, S. Brain Cholesterol: Long Secret Life
Behind a Barrier. Arterioscler. Thromb. Vasc. Biol. 2004, 24,
806.
(7) Stiles, A. R.; Kozlitina, J.; Thompson, B. M.; McDonald, J.
G.; King, K. S.; Russell, D. W. Genetic, anatomic, and clinical
determinants of human serum sterol and vitamin D levels. Proc.
Natl. Acad. Sci. USA 2014, 111, E4006.
(8) Dzeletovic, S.; Breuer, O.; Lund, E.; Diczfalusy, U. Determi-
nation of Cholesterol Oxidation Products in Human Plasma by
Isotope Dilution-Mass Spectrometry. Anal. Biochem. 1995, 225,
73.
(15) Wentworth, P.; Nieva, J.; Takeuchi, C.; Galve,́ R.; Went-
worth, A.; Dilley, R. B.; DeLaria, G. A.; Saven, A.; Babior, B.
M.; Janda, K. D.; Eschenmooser, A.; Lerner, R. A. Evidence for
Ozone Formation in Human Atherosclerotic Arteries. Science
2003, 302, 1053.
(16) Brinkhorst, J.; Nara, S. J.; Pratt, D. A. Hock Cleavage of
Cholesterol 5r-Hydroperoxide: An Ozone-Free Pathway to the
Cholesterol Ozonolysis Products Identified in Arterial Plaque and
Brain Tissue. J. Am. Chem. Soc. 2008, 130, 12224.
(17) Zielinski, Z. A. M.; Pratt, D. A. Cholesterol Autoxidation
Revisited: Debunking the Dogma Associated with the Most Vili-
fied of Lipids. J. Am. Chem. Soc. 2016, 138, 6932.
(18) Poirot, M.; Silvente-Poirot, S. Cholesterol-5,6-epoxides:
Chemistry, biochemistry, metabolic fate and cancer. Biochimie
2013, 95, 622.
(19) Speen, A. M.; Kim, H.-Y. H.; Bauer, R. N.; Meyer, M.;
Gowdy, K. M.; Fessler, M. B.; Duncan, K. E.; Liu, W.; Porter, N.
A.; Jaspers, I. Ozone-derived Oxysterols Affect Liver X Receptor
(LXR) Signaling. J. Biol. Chem. 2016, 291, 25192.
(20) Holm, T.; Crossland, I. A Simple Preparation of Hexadeuter-
iocholesterol. J. Label. Compd. Radiopharm. 1996, 38, 803.
(9) Schroepfer, G. J. Oxysterols: Modulators of Cholesterol Me-
tabolism and Other Processes. Physiol. Rev. 2000, 80, 361.
(10) For example see: (a) Porter, F. D.; Scherrer, D. E.; Lanier,
M. H.; Langmade, S. J.; Molugu, V.; Gale, S. E.; Olzeski, D.;
Sidhu, R.; Dietzen, D. J.; Fu, R.; Wassif, C. A.; Yanjanin, N. M.;
Marso, S. P.; House, J.; Vite, C.; Schaffer, J. E.; Ory, D. S. Cho-
lesterol Oxidation Products Are Sensitive and Specific Blood-
Based Biomarkers for Niemann-Pick C1 Disease. Sci. Transl.
Med. 2010, 2, 1; (b) Luu, W.; Sharpe, L. J.; Capell-Hattam, I.;
Gelissen, I. C.; Brown, A. J. Oxysterols: Old Tale, New Twists.
14
ACS Paragon Plus Environment

Page 15 of 16
Journal of the American Chemical Society
derived phenoxyl radical has not been reported, the reported dif-
(21) Kirk, D. N.; Varley, M. J.; Makin, H. L. J.; Trafford, D. J. H.
1
2
3
4
5
6
7
8
Synthesis of [26,27-2H6]Cholesterol and Derivatives substituted
in the Side chain. J. Chem. Soc., Perkin Trans. 1 1983, 2563.
(22) Chol A/B ring deuteration is precedented (e.g. 2,2,4,4,7,7-
d6-chol), however its conversion to the epoxide and 5a-OH stand-
ards would yield the former as d6-labelled and the latter as d5-
labelled – and still requires 8 steps. See: Fujimoto, Y.; Maeda, I.;
Ohyama, K.; Hikiba, J.; Kataoka, H. Biosynthesis of 20-hydrox-
yecdysone in plants: 3b-Hydroxy-5b-cholestan-6-one as an inter-
mediate immediately after cholesterol in Ajuga hairy roots. Phy-
tochemistry 2015, 111, 59.
(23) De Miranda, M. G.; Albert, A. L. M.; Cardoso, J. N.; Lopes,
R. S. C.; Lopes, C. C. Straightforward Synthesis of 2,2,4,4,5,7,7-
d7-Cholestane: A New Deuterated Standard in Petroleum Analy-
sis. Quim. Nova 2013, 36, 1160.
(24) Giner, J.-L.; Faraldos, J. A. A Biomimetic Approach to the
Synthesis of an Antiviral Marine Steroidal Orthoester. J. Org.
Chem. 2002, 67, 2717.
(25) We note that these hydroxycholesterols and ketocholesterols
could also form from peroxyl radical termination reactions and
from further autoxidation or decomposition of the initially-
formed hydroperoxides.42 A control experiment in the absence of
PPh3 showed the contribution of those pathways is relatively mi-
nor (<20%) in chol autoxidations without additive (cf. page S40.
(26) The amount of chol 4-OH accounted for only 5% of the prod-
uct distribution, indicating the amount previously reported17 had
been overestimated due to co-elution with the chol epoxides.
(27) Both chol epoxides may be hydrolyzed to give 5a,6b-dihy-
droxycholesterol18 (chol triol), and the authentic chol triol was
found to have a MS/MS transitions corresponding to M+H+−H2O
à M+H+−2H2O (m/z 403.4 à m/z 385.4) as well as a second
transition of M+H+−2H2O à M+H+−3H2O (m/z 385.4 à m/z
367.4). The chol triol was formed only in trace amounts (< 0.5
mol% of the products) during the 16-hour autoxidation (cf. Figure
S21), which is not surprising given that harsh acidic or basic con-
ditions were required to hydrolyze the epoxides synthetically (see
Experimental Section for details).
ference of 86,000-fold33b in kt between 2,6-di-tert-butyl-4-
methylphenoxyl radical and the 2,4,6-tri-tert-butylphenoxyl rad-
ical enables an estimate from the value of kt for 2,4,6-trime-
thylphenol (4.5´107)33b of ~500 M-1s-1 – smaller than for PMC
(a-tocopherol), for which 3000 M-1s-1 has been determined.33b (b)
Burton, G. W.; Doba, T.; Gabe, E.; Hughes, L.; Lee, F. L.; Prasad,
L.; Ingold, K. U. Autoxidation of Biological Molecules. 4. Max-
imizing the Antioxidant Activity of Phenols. J. Am. Chem. Soc.
1985, 107, 7053.
(34) Xu, L.; Porter, N. A. Free radical oxidation of cholesterol
and its precursors: Implications in cholesterol biosynthesis disor-
ders. Free Radic. Res. 2015, 49, 835.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(35) The kp for chol was slightly lower in our hands than the lit-
erature value of 11 ± 2 M-1 s-1.4 In contrast to autoxidation, we
have found that 1O2 reacts more slowly with chol OAc than chol
(unpublished results). Furthermore, the substituent has the oppo-
site effect on peroxyl radical addition, where kepox is slowed
slightly to 7.8 and 13 M-1s-1 for addition to chol OAc at 6α and
6β, respectively. The origins of these differences are unclear and
remain under active investigation in our labortatories.
(36) Paillasse, M. R.; Saffon, N.; Gornitzka, H.; Silvente-Poirot,
S.; Poirot, M.; de Medina, P. Surprising unreactivity of choles-
terol-5,6-epoxides towards nucleophiles. J. Lipid Res. 2012, 53,
718.
(37) Maillard, B.; Ingold, K. U.; Scaiano, J. C. Rate Constants for
the Reactions of Free Radicals with Oxygen in Solution. J. Am.
Chem. Soc. 1983, 105, 5095.
(38) Porter, N. A.; Mills, K. A.; Carter, R. L. AMechanistic Study
of Oleate Autoxidation: Competing Peroxyl H-Atom Abstraction
and Rearrangement. J. Am. Chem. Soc. 1994, 116, 6690.
(39) Tallman, K. A.; Pratt, D. A.; Porter, N. A. Kinetic Products
of Linoleate Peroxidation: Rapid b-Fragmentation of Nonconju-
gated Peroxyls. J. Am. Chem. Soc. 2001, 123, 11827.
(40) However, a recent report has shown that peroxyl radical ad-
dition to the unsaturation in arachidonic acid yields epoxyeicosa-
trienoic acids via peroxyl radical addition, see: Aliwarga, T.; Rac-
cor, B. S.; Lemaitre, R. N.; Sotoodehnia, N.; Gharib, S. A.; Xu,
L.; Totah, R. A. Enzymatic and free radical formation of cis- and
trans- epoxyeicosatrienoic acids in vitro and in vivo. Free Radic
Biol Med 2017, 112, 131.
(41) Xu, L.; Korade, Z.; Rosado, D. A.; Liu, W.; Lamberson, C.
R.; Porter, N. A. An oxysterol biomarker for 7-dehydrocholes-
terol oxidation in cell/mouse models for Smith-Lemli-Opitz syn-
drome. J. Lipid Res. 2011, 52, 1222.
(42) Zielinski, Z. A. M.; Pratt, D. A. Lipid Peroxidation: Kinetics,
Mechanisms, and Products. J. Org. Chem. 2017, 82, 2817.
(43) Breuer, O.; Dzeletovic, S.; Lund, E.; Diczfalusy, U. The
oxysterols cholest-5-ene-3b,4a-diol, cholest-5-ene-3b,4b-diol
and cholestane-3b,5a,6a-triol are formed during in vitro oxida-
tion of low density lipoprotein, and are present in human athero-
sclerotic plaques. Biochim. Biophys. Acta, Lipids Lipid Metab.
1996, 1302, 145.
(44) Hill, S.; Lamberson, C. R.; Xu, L.; To, R.; Tsui, H. S.;
Shmanai, V. V.; Bekish, A. V.; Awad, A. M.; Marbois, B. N.;
Cantor, C. R.; Porter, N. A.; Clarke, C. F.; Shchepinov, M. S.
Small amounts of isotope-reinforced polyunsaturated fatty acids
suppress lipid autoxidation. Free Radic. Biol. Med. 2012, 53, 893.
(45) Muchalski, H.; Levonyak, A. J.; Xu, L.; Ingold, K. U.; Por-
ter, N. A. Competition H(D) Kinetic Isotope Effects in the Au-
toxidation of Hydrocarbons. J. Am. Chem. Soc. 2015, 137, 94.
(46) Muchalski, H.; Xu, L.; Porter, N. A. Tunneling in tocoph-
erol-mediated peroxidation of 7-dehydrocholesterol. Org. Bio-
mol. Chem. 2015, 13, 1249.
(28) Niki, E. Biomarkers of lipid peroxidation in clinical material.
Biochim. Biophys. Acta, Gen. Subj. 2014, 1840, 809.
(29) Lee, C.-Y. J.; Huang, S. H.; Jenner, A. M.; Halliwell, B.
Measurement of F2-isoprostanes, hydroxyeicosatetraenoic prod-
ucts, and oxysterols from a single plasma sample. Free Radic.
Biol. Med. 2008, 44, 1314.
(30) Iuliano, L.; Micheletta, F.; Natoli, S.; Ginanni Corradini, S.;
Iappelli, M.; Elisei, W.; Giovannelli, L.; Violi, F.; Diczfalusy, U.
Measurement of oxysterols and a-tocopherol in plasma and tissue
samples as indices of oxidant stress status. Anal. Biochem. 2003,
312, 217.
(31) This was a slight decrease from our previously reported
value.17
(32) Haidasz, E. A.; Van Kessel, A. T. M.; Pratt, D. A. A Contin-
uous Visible Light Spectrophotometric Approach To Accurately
Determine the Reactivity of Radical-Trapping Antioxidants. J.
Org. Chem. 2016, 81, 737.
(33) (a) Interestingly, the 30 mol% maximum in epoxide concen-
tration occurs at higher [PMC] than [BDMP] (cf. Figures S26 and
2C, respectively). We suggest that the interplay between the H-
atom donating ability of the two phenols (PMC being ~10-fold
more potent), and the concentration of antioxidant at which the
contribution of antioxidant-mediated peroxidation (AMP) be-
comes significant, dictates the maximum epoxide concentration.
AMP is expected to be more efficient for BDMP than for PMC
since the aryloxyl radical derived from BDMP is expected to be
both less stable (O-H BDE for the former should be roughly 4
kcal/mol higher than for the latter, and similar to that in choles-
terol) and more persistent owing to the t-butyl group on the (para)
position. While the self-termination rate constant for the BDMP-
(47) Lamberson, C. R.; Xu, L.; Muchalski, H.; Montenegro-
Burke, J. R.; Shmanai, V. V.; Bekish, A. V.; McLean, J. A.;
Clarke, C. F.; Shchepinov, M. S.; Porter, N. A. Unusual Kinetic
15
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 16 of 16
profile of oxysterols oxidized at C4 (4a- and 4b-hydroxycholes-
terol) and C7 (7-ketocholesterol, 7a- and 7b-hydroxycholesterol)
on cells of the central nervous system. Eur. J. Med. Chem. 2013,
70, 558.
(54) Thomas, M. G.; Suckling, C. J.; Pitt, A. R.; Suckling, K. E.
The synthesis of A- and B-ring fluorinated analogues of choles-
terol. J. Chem. Soc., Perkin Trans. 1 1999, 3191.
(55) Korde, S. S.; Udasi, R. A.; Trivedi, G. K. Oxidation of Ste-
roidal 5-En-3β-ol with Pyridinium Chlorochromate: Isolation of
Key Intermediate, Steroidal 6β-Hydroxy-4-en-3-one. Synth.
Commun. 1997, 27, 3419.
(56) Cui, J.; Huang, L.; Fan, L.; Zhou, A. A facile and efficient
synthesis of some (6E)-hydroximino-4-en-3-one steroids, steroi-
dal oximes from Cinachyrella spp. Sponges. Steroids 2008, 73,
252.
(57) Shamsuzzaman; Ahmad, S.; Khan, B. Z.; Shafiullah. A Con-
venient Method for the Synthesis of 3β-Hydroxy 4-En-6-one
Steroids. J. Org. Chem. 1991, 56, 1936.
(58) Li, S.; Pang, J.; Wilson, W. K.; Schroepfer, G. J. Sterol syn-
thesis. Preparation and characterization of fluorinated and deuter-
ated analogs of oxygenated derivatives of cholesterol. Chem.
Phys. Lipids 1999, 99, 33.
(59) Carvalho, J. F. S.; Silva, M. M. C.; Melo, M. L. S. E.
Efficient trans-diaxial hydroxylation of D5-steroids. Tetrahedron
2010, 66, 2455.
Isotope Effects of Deuterium Reinforced Polyunsaturated Fatty
Acids in Tocopherol-Mediated Free Radical Chain Oxidations. J.
Am. Chem. Soc. 2014, 136, 838.
(48) Shchepinov, M. S.; Chou, V. P.; Pollock, E.; Langston, J.
W.; Cantor, C. R.; Molinari, R. J.; Manning-Boğ, A. B. Isotopic
reinforcement of essential polyunsaturated fatty acids diminishes
nigrostriatal degeneration in a mouse model of Parkinson’s dis-
ease. Toxicol. Lett. 2011, 207, 97.
1
2
3
4
5
6
7
8
(49) Fieser, L. F. Cholesterol and Companions. VII. Steroid Di-
bromides. J. Am. Chem. Soc. 1953, 75, 5421.
(50) Noguchi, N.; Yamashita, H.; Gotoh, N.; Yamamoto, Y.; Nu-
mano, R.; Niki, E. 2,2`-Azobis (4-Methoxy-2,4-dimethylvale-
ronitrile), A New Lipid-Soluble Azo Initiator: Application to Ox-
idations of Lipids and Low-Density Lipoprotein in Solution and
in Aqueous Dispersions. Free Radic. Biol. Med. 1998, 24, 259.
(51) Carvalho, J. F. S.; Cruz Silva, M. M.; Moreira, J. N.; Simões,
S.; Sá E Melo, M. L. Efficient Chemoenzymatic Synthesis, Cyto-
toxic Evaluation, and SAR of Epoxysterols. J. Med. Chem. 2009,
52, 4007.
(52) Beckwith, A. L. J.; Davies, A. G.; Davison, I. G. E.; Maccoll,
A.; Mruzek, M. H. The Mechanisms of the Rearrangements of
Allylic Hydroperoxides: 5a-Hydroperoxy-3b-hydroxycholest-6-
ene and 7a-Hydroperoxy-3b-hydroxycholest-5-ene. J. Chem.
Soc., Perkin Trans. 2 1989, 815.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(53) Nury, T.; Samadi, M.; Zarrouk, A.; Riedinger, J. M.; Lizard,
G. Improved synthesis and in vitro evaluation of the cytotoxic
Insert Table of Contents artwork here
16
ACS Paragon Plus Environment