Synthesis, discovery, and quantitation of dihomo-isofurans: Biomarkers for in vivo adrenic acid peroxidation

Article abstract of DOI:10.1002/anie.201402440

The growing importance of lipidomics, and the interest of non-enzymatic metabolites of polyunsaturated fatty acids (PUFAs) prompted us to initiate the synthesis of novel dihomo-IsoF compounds. Such metabolites of adrenic acid, the main PUFA in white matte

Full text of DOI:10.1002/anie.201402440

Angewandte
Chemie
DOI: 10.1002/anie.201402440
Total Synthesis
Synthesis, Discovery, and Quantitation of Dihomo-Isofurans:
Biomarkers for In Vivo Adrenic Acid Peroxidation**
Aurꢀlien de La Torre, Yiu Yiu Lee, Camille Oger, Per Torp Sangild, Thierry Durand,
Jetty Chung-Yung Lee,* and Jean-Marie Galano*
Abstract: The growing importance of lipidomics, and the
interest of non-enzymatic metabolites of polyunsaturated fatty
acids (PUFAs) prompted us to initiate the synthesis of novel
dihomo-IsoF compounds. Such metabolites of adrenic acid, the
main PUFA in white matter, were synthesized using a divergent
approach based on an orthoester cyclization. LC-MS/MS
investigation on pig brains showed the potential of this novel
biomarker for the first time, as a powerful new tool for brain
lipid peroxidation assessment.
Oxygenated metabolites of polyunsaturated fatty acids
(PUFAs) are key classes of natural products generated by
living organisms. The levels are regulated by the presence of
reactive oxygen species (ROS). Among them, isoprostanes
(IsoPs), liberated through lipid peroxidation of arachidonic
acid (AA), and more recently neuroprostanes (NeuroPs),
from docosahexaenoic acid (DHA), are widely studied.[1,2]
IsoPs are used as systemic oxidative stress biomarkers in vivo
whereas NeuroPs are specific for neurodegenerative dis-
eases.[3] A novel peroxidation pathway of PUFAs was
discovered and it generates tetrahydrofuran (THF) deriva-
tives, that is, isofurans (IsoFs) from AA[4] and neurofurans
(NeuroFs) from DHA.[5] These compounds could be highly
valuable biomarkers of oxidative stress because they appear
to be more abundantly produced than their isoprostanoid
counterparts, and because their formation is dependent upon
oxygen tension. We previously showed that F2-dihomo-IsoPs,
isoprostanoid metabolites of adrenic acid (AdA, being the
most important PUFA in white matter),[6] and 7- and 17-F2t-
dihomo-IsoPs are biomarkers for the early detection of Rett
syndrome (RTT), and subsequently, oxidative damage of the
Scheme 1. Representative structures of dihomo-IsoP and dihomo-IsoF
from AdA peroxidation. ROS=reactive oxygen species, R=phosphate
group.
myelin (Scheme 1).[7] It is therefore reasonable to assume that
IsoF metabolites of AdA, namely dihomo-IsoFs, do exist and
could be complementary biomarkers of RTT and other
neuronal damage. In pursuing this hypothesis, the first
synthesis of dihomo-IsoFs is described herein. For the first
time, concentrations of dihomo-IsoFs were determined by
LC-MS/MS to confirm their presence in vivo, particularly in
the brain, and to compare them with those of known
biomarkers of oxidative stress, mainly IsoPs, NeuroPs,[8] and
F2-dihomo-IsoPs.
The biosynthesis of dihomo-IsoFs can theoretically give
32 possible diastereoisomers, and because two biosynthesis
pathways coexist two classes of four families are present, that
is, the alkenyl-IsoFs and enediol-IsoFs.[9] Therefore, a total of
256 isomers are potentially generated in vivo.
[*] A. de La Torre, Dr. C. Oger, Dr. T. Durand, Dr. J.-M. Galano
Institut des Biomolꢀcules Max Mousseron (IBMM) UMR CNRS
5247—Universitꢀs de Montpellier I et II, ENSCM, Facultꢀ de
Pharmacie
15 Av. Charles Flahault, 34093 Montpellier cedex 05 (France)
E-mail: jean-marie.galano@univ-montp1.fr
Three different synthetic strategies have been reported
for IsoFs and NeuroFs over the past decade by the groups of
Falck, Taber, and Zanoni-Vidari.[10–14] To identify more
distinct and relevant metabolites for biological investigation,
we developed a versatile strategy based on a novel framework
for both alkenyl- and enediol-dihomo-IsoFs.
Dr. J. C. Y. Lee
School of Biological Sciences, The University of Hong Kong
Pokfulam Road, Hong Kong SAR (China)
E-mail: jettylee@hku.hk
P. T. Sangild
Retrosynthetic analysis of alkenyl-dihomo-IsoFs enabled
us to identify the THF precursors A (Scheme 2). To develop
a common scaffold for both alkenyl- and enediol-isofura-
noids, the polyalcohol derivative C would be a suitable target.
Two cyclization processes (5-exo-tet for alkenyl and 5-endo-
tet for enediol) could be foreseen with a suitably mono-
protected orthoester derivative (B for alkenyl-type) following
Department of Nutrition, Exercise and Sports, Faculty of Science
University of Copenhagen, Frederiksberg (Denmark)
[**] We thank the Universitꢀ de Montpellier I (BQR 2011), CNRS and the
Fondation pour la Recherche Mꢀdicale for financial support, and Dr.
Martin Christlieb and Prof. Ullrich Jahn for their technical advice.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201402440.
Angew. Chem. Int. Ed. 2014, 53, 1 – 5
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1
These are not the final page numbers!

.
Angewandte
Communications
prior to another Sharpless asymmetric dihydroxylation, thus
affording the diol 6 in good yield and good diastereoselectiv-
ity (d.r. = 88:12).
Regioselective monoprotection of 6 was achieved using
Kusumotoꢀs method through orthoester hydrolysis.[18] Treat-
ment of 6 with trimethyl orthoacetate and PTSA resulted in
placing the acyl group in the b-position (4:1). However,
subsequent acidic cleavage of the acetonide with PTSA
caused a migration of the acyl moiety to the liberated 1,2-diol.
Nevertheless, monoprotection was achieved by using tri-
methyl orthobenzoate to introduce the benzoyl group to the
b-position of 7, which was isolated in addition to 8 with a 4:1
regioselectivity (Scheme 4). Interestingly, the benzoyl group
Scheme 2. Retrosynthetic analysis of alkyl-dihomo-IsoF derivatives.
LA=Lewis acid.
the use of Borhanꢀs stereoselective orthoester cyclization of
1,2,n-triols.[15] The compound C should be available from
trans-b-muconic acid (2). Following this analysis we now
describe the first synthesis of 10-epi-17(RS)-SC-D15-11-
dihomo-IsoF (1).
The synthesis started with a four-step sequence to access
to the 1,6-diol 4 in 62% yield and in virtually optically pure
form (ee > 99%; Scheme 3).[16] Esterification of 2 with acetyl
chloride in an anhydrous iPrOH solution and subsequent
Scheme 4. Reagents and conditions: a) PhC(OMe)3, PTSA, CH2Cl2, RT;
then PTSA, H2O, RT, 88% as a (4:1) mixture of 7 and 8; b) PTSA,
MeOH/H2O (9:1), RT, then NaHCO3(s), 66%; c) MeC(OMe)3, PPTS,
CH2Cl2, RT; then BF3.Et2O, 158C, 60%; d) K2CO3, MeOH, RT, 72%;
e) TBSCl, imidazole, DMAP, CH2Cl2, RT, quant; f) LiBH4, MeOH, THF,
868C, 86%. DMAP=4-(N,N-dimethylamino)pyridine, PPTS=pyridi-
nium para-toluenesulfonate, TBS=tert-butyldimethylsilyl.
did not migrate from the b-position under acidic conditions,
but did migrate from the a- to the b-position under basic
conditions. Thus, the mixture of regioisomers 7 and 8 was
directly treated with PTSA in MeOH/H2O (9:1), followed by
a basic quench with solid NaHCO3 to afford the triol 9 as
a single regioisomer. The triol 9 was then treated under
reaction conditions developed by the group of Borhan,[15] that
is, orthoester formation with MeC(OMe)3 and PPTS, fol-
lowed by in situ addition of catalytic BF3·Et2O to conduct the
intramolecular attack at the orthoester intermediate. The
cyclization process was extremely temperature dependent.
Lewis acid addition at 308C led to consecutive cleavage of the
PMB group, whereas running it at 08C prevented cyclization
and deprotection. Best results were obtained at 158C, thus
yielding the THF derivative 10 in 60% yield. The relative
configuration of 10 was confirmed by NOESY NMR. At the
stage of the synthesis and after purification by flash chroma-
tography, 10 appears to be free of the minor diastereoisomer
observed when 6 was isolated. The Ac and Bz protecting
groups of 10 were subsequently transformed into TBS groups
by methanolysis with K2CO3 in MeOH followed by addition
of TBSCl. Reduction of the methyl ester group with LiBH4
Scheme 3. Reagents and conditions: a) AcCl, iPrOH, 1008C, 89%;
b) AD-mix b, MeSO2NH2, NaHCO3, tBuOH/H2O (1:1), 08C, 77%;
c) PTSA, 2,2-dimethoxypropane, RT, 99%; d) LiAlH4, THF, 08C!RT,
91%; e) MnO2, Ph3PCHCO2Et, 1,2-dichloroethane, 628C, 62%;
f) PMBTCA, La(OTf)3, toluene, RT, 93%; g) AD-mix a, MeSO2NH2,
tBuOH/H2O (1:1), 08C to RT, 94%, d.r.=88:12. PMBTCA=para-
methoxybenzyl trichloroacetimidate, PTSA=para-toluenesulfonic acid,
Tf =trifluoromethanesulfonyl.
Sharpless asymmetric dihydroxylation gave the correspond-
ing diol 3. Acetonide protection and ester reduction with
LiAlH4 gave 4 on large scale (15 g). This C2-symmetric
structure was then selectively mono-oxidized and elongated
by a Wittig reaction in the same pot,[17] thereby resulting in the
a,b-unsaturated ester 5 in 62% yield. Protection of the
remaining free alcohol yielded the corresponding PMB ether
2
www.angewandte.org
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 1 – 5
These are not the final page numbers!

Angewandte
Chemie
gave the compound 11 in 62% yield over three steps. This
compound is the key intermediate for the generation of all
alkenyl-IsoFs, alkenyl-dihomo-IsoFs, and alkenyl-NeuroFs.
The primary alcohol 11 was transformed into the corre-
sponding aldehyde with the Dess–Martin periodinane and
followed by a Horner–Wadsworth–Emmons reaction with the
commercially available phosphonate 12 (Scheme 5). The
compared to grey matter.[6,19] The difference in PUFA
localization and oxygen tension suggests the importance of
measuring the appropriate LPO metabolites in the different
parts of the brain. Lipids of the prefrontal cortex and medial
prefrontal cortex of preterm pig (116 days) brains were
extracted by Folch solution and subjected to alkaline hydrol-
ysis. Thereafter, AA, DHA, and AdA and the LPO metab-
olites were purified by anionic solid-phase extraction and
analyzed by LC-MS/MS.[20] It was found, that the levels of
DHA:AA:AdA amounted to approximately 1:5:0.1 in the
prefrontal cortex and 1:2:0.07 in the medial prefrontal cortex.
These levels represent the total PUFA concentration of the
brain tissue without differentiating grey and white matter.
The determination of LPO metabolites revealed signifi-
cantly high concentrations of 7(RS)-7-F2t-dihomo-IsoP and
4(RS)-4-F4t-NeuroP in the prefrontal cortex compared to 15-
F2t-IsoP, 10-F4t-NeuroP, or 17(RS)-17-F2t-dihomo-IsoP (Fig-
ure 1A). Similarly, the levels of 4(RS)-4-F4t-NeuroP and
7(RS)-7-F2t-dihomo-IsoP were significantly higher in the
medial prefrontal cortex compared to those of 5-F2t-IsoP
and 10-F4t-NeuroP (Figure 1B).
Interestingly, despite the relatively low concentration of
AdA, 10-epi-17(RS)-SC-D15-11-dihomo-IsoF levels in both
tissues were three- to fourfold higher than those of IsoFs and
NeuroFs (Figures 1C and D). This difference is significant,
since only up to 32 stereoisomers (25) of the particular
regioisomeric dihomo-IsoF series, represented by the stan-
dard 1, out of the 256 theoretical isomers were specifically
measured, whereas IsoFs and NeuroFs levels represent the
sum of all 256 and 512 potential metabolites, respectively.
The growth and development of the pig brain is similar to
that of the human brain, therefore the results have a signifi-
cant impact on the study of neuronal damage in humans.
Primarily, our study revealed that 10-epi-17(RS)-SC-D15-11-
dihomo-IsoF levels were extraordinarily high compared to
IsoFs and NeuroFs in the prefrontal cortex tissue of pig
brains. Secondly, the dihomo-IsoF levels are comparable to
those of 4(RS)-4-F4t-NeuroP (Figure 1A vs C), which is to
date the most prominent biomarker for oxidative stress
related neuronal damage.[2,5] These results show that the
whole spectrum of isoprostanoid and isofuranoid metabolites
represented by the novel 10-epi-17(RS)-SC-D15-11-dihomo-
IsoF from AdA and known 4(RS)-4-F4t-NeuroP from DHA
should be considered when evaluating neuronal damage and
disease. Moreover, the isofuranoid AdA metabolite 1 has
a much higher thermal and oxidative stability than 4(RS)-4-
F4t-NeuroP, and should lead to significantly more reliable and
reproducible determination of LPO in brain tissues.
Scheme 5. Reagents and conditions: a) DMP, CH2Cl2, RT; b) Ba(OH)2,
12, THF, RT, 60% over two steps; c) CeCl3.7H2O, NaBH4, MeOH, 08C;
d) TBSCl, imidazole, DMAP, CH2Cl2, RT, 85% over two steps; e) DDQ,
CH2Cl2/H2O (10:1), 08C to RT, 80%; f) DMP, CH2Cl2, RT; g) 15,
NaHMDS, THF, ꢀ788C to RT, 65% over two steps; h) TBAF, THF, RT;
i) LiOH, THF/H2O (1:1), RT, 76% over two steps. DMP=Dess–Martin
periodinane, DDQ=2,3-dichloro-5,6-dicyano-1,4-benzoquinone,
NaHMDS=sodium bis(trimethylsilyl)amide, TBAF=tetrabutylammo-
nium fluoride.
enone 13 was obtained in 60% yield and thus further reduced
under Lucheꢀs conditions and protected with a TBS group.
Cleavage of the PMB group provided the primary alcohol 14
in 68% yield over three steps. Another oxidation/Wittig
sequence with the phosphonium bromide 15 gave the ethyl
ester 16 in 65% yield. Finally, exhaustive silyl ether depro-
tection and saponification of the ester group yielded 10-epi-
17(RS)-SC-D15-11-dihomo-IsoF (1) in 76% yield.
Brain tissue consumes approximately 20% of the total
oxygen in the body. This consumption increases the likelihood
of brain lipid peroxidation when injured because of the high
abundance of PUFAs, especially AA, AdA, and DHA. The
localization of PUFA in the brain is specific: DHA is enriched
in grey matter and AdA is found in white matter, whereas AA
and EPA are evenly distributed. These brain PUFAs are
subject to lipid peroxidation (LPO) and oxidative damage
within the central nervous system. Furthermore, the products
released upon LPO may depend on the oxygen tension in the
brain; a higher tension was reported in brain white matter
Importantly, the pig brain samples of the present study
were at a homeostatic state. This state delineates the present
results from those of a previous report,[21] which found
elevated IsoF levels at greater than 21% oxygen tension and
elevated NeuroF at greater than 40% in the prefrontal cortex
of piglets after resuscitation from hypoxaemia compared to
nontreated controls.
A damaged prefrontal cortex is associated with cognitive
dysfunction, bipolar disorder, and schizophrenia patients.[22,23]
This association parallels the recent finding of a relationship
to oxidative stress, because the myelin fraction in neuronal
Angew. Chem. Int. Ed. 2014, 53, 1 – 5
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3
These are not the final page numbers!

.
Angewandte
Communications
Keywords: biomarkers · lipids ·
.
mass spectrometry · oxidation ·
total synthesis
[1] J. D. Morrow, K. E. Hill, R. F. Burk,
T. M. Nammour, K. F. Badr, L. J. Rob-
erts, Proc. Natl. Acad. Sci. USA 1990, 87,
9383 – 9387.
[2] L. J. Roberts II, T. J. Montine, W. R.
Markesbery, A. R. Tapper, P. Hardy, S.
Chemtob, W. D. Dettbarn, J. D. Morrow,
J. Biol. Chem. 1998, 273, 13605 – 13612.
[3] U. Jahn, J.-M. Galano, T. Durand,
Angew. Chem. 2008, 120, 5978 – 6041;
Angew. Chem. Int. Ed. 2008, 47, 5894 –
5955.
[4] J. P. Fessel, N. A. Porter, K. P. Moore,
J. R. Sheller, L. J. Roberts, Proc. Natl.
Acad. Sci. USA 2002, 99, 16713 – 16718.
[5] W.-L. Song, J. A. Lawson, D. Reilly, J.
Rokach, C.-T. Chang, B. Giasson, G. A.
FitzGerald, J. Biol. Chem. 2008, 283, 6 –
16.
[6] E. R. Skinner, C. Watt, J. a. O. Besson,
P. V. Best, Brain 1993, 116, 717 – 725.
[7] C. De Felice, C. Signorini, T. Durand, C.
Oger, A. Guy, V. Bultel-Ponce, J.-M.
Galano, L. Ciccoli, S. Leoncini, M.
DꢀEsposito, S. Filosa, A. Pecorelli, G.
Valacchi, J. Hayek, J. Lipid Res. 2011,
52, 2287 – 2297.
[8] C.-Y. J. Lee, R. C. S. Seet, S. H. Huang,
L. H. Long, B. Halliwell, Antioxid.
Redox Signaling 2009, 11, 407 – 420.
[9] D. F. Taber, J. P. Fessel, L. J. Roberts II,
Prostaglandins Other Lipid Mediators
2004, 73, 47 – 50.
Figure 1. Concentration of isoprostanoids and isofuranaoids in pig brain tissues. Values are
meanꢁSD, n=12. Concentration of each compound is expressed per g tissue. Significant
differences are annotated by *p<0.05, **p<0.01, ***p<0.001.
[10] J. R. Falck, L. Manmohan Reddy, K.
Byun, W. B. Campbell, X.-Y. Yi,
Bioorg. Med. Chem. Lett. 2007, 17, 2634 – 2638.
cells from the prefrontal cortex of bipolar disorder patients
had increased F2-IsoPs levels compared to controls.[23]
[11] D. F. Taber, P. Gu, R. Li, J. Org. Chem. 2009, 74, 5516 – 5522.
[12] D. F. Taber, Z. Zhang, J. Org. Chem. 2006, 71, 926 – 933.
[13] D. F. Taber, Y. Pan, X. Zhao, J. Org. Chem. 2004, 69, 7234 – 7240.
[14] M. Valli, P. Bruno, D. Sbarbada, A. Porta, G. Vidari, G. Zanoni, J.
Org. Chem. 2013, 78, 5556 – 5567.
[15] T. Zheng, R. S. Narayan, J. M. Schomaker, B. Borhan, J. Am.
Chem. Soc. 2005, 127, 6946 – 6947.
[16] B. H. Lipshutz, D. Buzard, C. Olsson, K. Noson, Tetrahedron
2004, 60, 4443 – 4449.
[17] D. J. Phillips, K. S. Pillinger, W. Li, A. E. Taylor, A. E. Graham,
Tetrahedron 2007, 63, 10528 – 10533.
[18] M. Oikawa, A. Wada, F. Okazaki, S. Kusumoto, J. Org. Chem.
1996, 61, 4469 – 4471.
[19] Z.-C. Feng, E. L. Roberts Jr., T. J. Sick, M. Rosenthal, Brain Res.
1988, 445, 280 – 288.
[20] M. L. S. Chung, K. Y. E. Lee, C.-Y. J. Lee, Food Chem. Toxicol.
2013, 53, 205 – 213.
In conclusion, a flexible strategy allowed the total syn-
thesis of both C17 epimers of 10-epi-SC-D15-11-dihomo-IsoF
in 22 steps with a 1.5% overall yield. 10-epi-SC-D15-11-
Dihomo-IsoF was discovered as a major metabolite of
nonenzymatic free-radical-mediated peroxidation of AdA in
the prefrontal and medial cortex of pig brain at homeostatic
state. Thus, the new isofuranoid metabolite 10-epi-17(RS)-
SC-D15-11-dihomo-IsoF from AdA and known 4(RS)-4-F4t-
NeuroP from DHA should be considered as the leading
complementary LPO biomarkers when evaluating neuronal
damage and disease, whereby the new isofuran metabolite
will have the advantage of considerably higher stability
compared to NeuroP. Their quantification in the brain will
likely provide beneficial insights into neurochemistry and
shed light on intervention studies of neurobiology.
[21] R. Solberg, M. Longini, F. Proietti, P. Vezzosi, O. D. Saugstad, G.
Buonocore, Free Radical Biol. Med. 2012, 53, 1061 – 1067.
[22] D. R. Euston, A. J. Gruber, B. L. McNaughton, Neuron 2012, 76,
1057 – 1070.
[23] A. C. Andreazza, J.-F. Wang, F. Salmasi, L. Shao, L. T. Young, J.
Neurochem. 2013, 127, 552 – 561.
Received: February 14, 2014
Revised: March 19, 2014
Published online: && &&, &&&&
4
www.angewandte.org
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 1 – 5
These are not the final page numbers!

Angewandte
Chemie
Communications
Total Synthesis
A. de La Torre, Y. Y. Lee, C. Oger,
P. T. Sangild, T. Durand, J. C. Y. Lee,*
J.-M. Galano*
&&&& — &&&&
Synthesis, Discovery, and Quantitation of
Dihomo-Isofurans: Biomarkers for In
Vivo Adrenic Acid Peroxidation
White matter matters: The enantioselec-
tive synthesis of dihomo-Isofurans
(dihomo-IsoFs), metabolites from in vivo
peroxidation of adrenic acid (found in
white matter of the brain), is described
and employs a C2-symmetric precursor in
a divergent strategy. Mass-spectrometry
analysis accurately validated its presence
in pig brain, and quantitation analysis
showed, for the first time, the importance
of such a novel biomarker in assessing
lipid peroxidation.
Angew. Chem. Int. Ed. 2014, 53, 1 – 5
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5
These are not the final page numbers!