Threshold protective effect of deuterated polyunsaturated fatty acids on peroxidation of lipid bilayers

Article abstract of DOI:10.1111/febs.14807

Autoxidation of polyunsaturated fatty acids (PUFAs) damages lipid membranes and generates numerous toxic by-products implicated in neurodegeneration, aging, and other pathologies. Abstraction of bis-allylic hydrogen atoms is the rate-limiting step of PUFA autoxidation, which is inhibited by replacing bis-allylic hydrogens with deuterium atoms (D-PUFAs). In cells, the presence of a relatively small fraction of D-PUFAs among natural PUFAs is sufficient to effectively inhibit lipid peroxidation (LPO). Here, we investigate the effect of various D-PUFAs on the stability of liposomes under oxidative stress conditions. The permeability of vesicle membranes to fluorescent dyes was measured as a proxy for bilayer integrity, and the formation of conjugated dienes was monitored as a proxy for LPO. Remarkably, both approaches reveal a similar threshold for the protective effect of D-PUFAs in liposomes. We show that protection rendered by D-PUFAs depends on the structure of the deuterated fatty acid. Our findings suggest that protection of PUFAs against autoxidation depends on the total level of deuterated bi-sallylic (CD₂) groups present in the lipid bilayer. However, the phospholipid containing 6,6,9,9,12,12,15,15,18,18-d₁₀-docosahexaenoic acid exerts a stronger protective effect than should be expected from its deuteration level. These findings further support the application of D-PUFAs as preventive/therapeutic agents in numerous pathologies that involve LPO.

Full text of DOI:10.1111/febs.14807

Received Date : 31-Dec-2018
Accepted Date : 06-Mar-2019
Threshold Protective Effect of Deuterated Polyunsaturated Fatty Acids
on Peroxidation of Lipid Bilayers
Alexander M. Firsova, Maksim A. Fomichb, Andrei V. Bekishb, Olga L. Sharkob,
Elena A. Kotovaa, Harry J. Saalc, Dragoslav Vidovicd, Vadim V. Shmanaib, Derek A. Pratte
Yuri N. Antonenkoa*, Mikhail S. Shchepinovc*
aBelozersky Institute of Physico-Chemical Biology, Lomonosov Moscow State University, Moscow
119991, Russia.
bInstitute of Physical Organic Chemistry, National Academy of Science, Minsk 220072, Belarus.
cRetrotope, Inc., Los Altos, CA 94022, USA.
dMonash University, Wellington Rd, Clayton VIC 3800, Melbourne, Australia.
eDepartment of Chemistry and Biomolecular Science, University of Ottawa, Ottawa, K1N 6N5 Canada
Corresponding authors are:
Mikhail S. Shchepinov, Retrotope, Inc., Los Altos, CA 94022, USA, email: shchepa65@yahoo.com
and Yuri N. Antonenko, Belozersky Institute of Physico-Chemical Biology, Lomonosov Moscow State
University, Moscow 119991, Russia, email: antonen@genebee.msu.ru
This article has been accepted for publication and undergone full peer review but has not
been through the copyediting, typesetting, pagination and proofreading process, which may
lead to differences between this version and the Version of Record. Please cite this article as
doi: 10.1111/febs.14807
This article is protected by copyright. All rights reserved.

Abstract
Autoxidation of polyunsaturated fatty acids (PUFAs) damages lipid membranes and
generates numerous toxic by-products implicated in neurodegeneration, aging and other
pathologies. Abstraction of bis-allylic hydrogen atoms is the rate-limiting step of PUFA
autoxidation, which is inhibited by replacing bis-allylic hydrogens with deuterium atoms
(D-PUFAs). In cells, the presence of a relatively small fraction of D-PUFAs among natural
PUFAs is sufficient to effectively inhibit lipid peroxidation. Here, we investigate the effect
of various D-PUFAs on the stability of liposomes under oxidative stress conditions. The
permeability of vesicle membranes to fluorescent dyes was measured as a proxy for
bilayer integrity, and the formation of conjugated dienes was monitored as a proxy for
lipid peroxidation. Remarkably, both approaches reveal a similar threshold for the
protective effect of D-PUFAs in liposomes. We show that protection rendered by D-PUFAs
depends on the structure of the deuterated fatty acid. Our findings suggest that
protection of PUFAs against autoxidation depends on the total level of deuterated
bisallylic (CD2) groups present in the lipid bilayer. However, the phospholipid containing
6,6,9,9,12,12,15,15,18,18-d10-docosahexaenoic acid (D10-DHA) exerts
a
stronger
protective effect than should be expected from its deuteration level. These findings
further support the application of D-PUFAs as preventive/therapeutic agents in numerous
pathologies that involve lipid peroxidation.
Running title: Deuterated bisallylic groups of phospholipids stabilize liposomes
Article type : Original Articles
Keywords: polyunsaturated fatty acids, lipid peroxidation, deuterium, reactive oxygen
species, liposome leakage, conjugated dienes, bisallylic groups.
This article is protected by copyright. All rights reserved.

Abbreviations: LPO, lipid peroxidation; ROS, reactive oxygen species; PUFA,
polyunsaturated fatty acid; D-PUFA, PUFA having bis-allylic hydrogens replaced with
deuterium atoms; KIE, kinetic isotope effect; OCR, oxygen consumption rate; ALS,
amyotrophic lateral sclerosis; FCS, fluorescence correlation spectroscopy; Asc, ascorbate;
Ara, arachidonic acid, DHA, docosahexaenoic acid; EPA, eicosapentaenoic acid; Lin, linoleic
acid; Lnn, linolenic acid; Ole, oleic acid; EtOAc, ethyl acetate; MsCl, methanesulfonyl
chloride; H-LPC, hydrogenated lysophosphatidylcholine; H-Lin-PC, 1-acyl-2-linoleyl-sn-
glycero-3-phosphatidylcholine;
D2-Lin-PC,
1-acyl-2-(11,11-D2-linoleyl)-sn-glycero-3-
phosphatidylcholine; H-Ara-PC, 1-acyl-2-arachidonoyl-sn-glycero-3-phosphatidylcholine; D6-
Ara-PC, 1-acyl-2-(7,7,10,10,13,13-D6-arachidonoyl)-sn-glycero-3-phosphatidylcholine; H-Lnn-
PC, 1-acyl-2-linolenyl-sn-glycero-3-phosphatidylcholine; D4-Lnn-PC, 1-acyl-2-(11,11,14,14-
D4-linolenyl)-sn-glycero-3-phosphatidylcholine; D2-Lnn-PC, 1-acyl-2-(14,14-D2-linolenyl)-sn-
glycero-3-phosphatidylcholine;
eicosapentaenoyl)-sn-glycero-3-phosphatidylcholine;
(6,6,9,9,12,12,15,15,18,18-D10-docosahexaenoyl)-sn-glycero-3-phosphatidylcholine;
D8-EPA-PC,
1-acyl-2-(7,7,10,10,13,13,16,16-D8-
D10-DHA-PC,
1-acyl-2-
SRB,
sulforhodamine B; D2-Lin, D4-Lnn, D6-Ara, D8-EPA and D10-DHA – linoleic acid, linolenic
acid, arachidonic acid, eicosapentaenoic acid and docosahexaenoic acid, respectively, with
full bis-allylic deuteration; HNE, 4-hydroxy-2-nonenal; 4-HHE, 4-hydroxy-2-hexenal; MDA,
malondialdehyde.
Introduction
Polyunsaturated fatty acids (PUFAs) are essential components of lipid membranes, providing
them with necessary fluidity. However, the very motif that makes PUFAs “fluid” – the
“skipped” 1,4-diene moiety – also makes PUFAs susceptible to lipid peroxidation (LPO). LPO
is increasingly recognized as a key contributing factor in numerous pathological events –
neuronal, ocular, vascular and age related. Two features make LPO particularly pernicious:
the autocatalytic radical chain reaction mechanism, and its non-enzymatic nature. The
former leads to extensive damage from a single initiating event, while the latter prevents
cells from evolving defences against direct control of LPO. The LPO-induced damage to living
This article is protected by copyright. All rights reserved.

systems is multifaceted (Scheme 1). Compromised membrane fluidity and barrier function
(Scheme 1E) occur in tandem with the formation of various carbonyl compounds (Scheme
1G) and other highly reactive species which irreversibly cross-link important biomolecules
and form mutagenic DNA conjugates [1-6]. Lipid bilayer integrity is so vital to neuronal
function that 5% of the total energy produced by the body is expended repairing damaged
lipids in the brain [7].
To form a membrane, PUFAs (as part of phospholipids) assume a dense, uninterrupted,
water-repelling regular formation, which other lipid-soluble molecules can disturb. This may
contribute to the inefficiency of antioxidants in inhibiting the LPO in vivo, established in
numerous clinical trials [3,8]. The level of antioxidants, such as tocopherols, ascorbate and
reduced glutathione, in the oxidative stress-exposed PUFA-rich parts of an organism can be
up-regulated by up to 40% [9]. Under stress, membranes rich in docosahexaenoic acid (DHA)
could lose up to 70% of their DHA [10], which is likely due to both down regulation of the
fraction of the most oxidizable PUFAs to reduce LPO, as well as loss of DHA to oxidation.
Similar decreases are observed for PUFAs, in general, in animals deficient in vitamin E (vit E)
[11]. Moreover, it has been suggested that certain PUFA-rich domains may have higher than
average levels of vit E at normal conditions [12]. Regardless, tightly controlled delivery
mechanisms cannot exceed a certain level of antioxidants in membranes, lest the structural
integrity or optimal parameters of the latter be compromised. Indeed, increasing the level
of vit E in model bilayers leads to decreased fluidity [13]. The reported physiological levels of
tocopherols relative to fatty acid residues in lipid membranes vary depending on cellular
and subcellular membrane types. The ratio of vit E to fatty acid residues can be as high as
1:130 in Golgi and lysosomal membranes [14], but more generally is around 1 tocopherol
molecule per 2000 fatty acid residues [3]. However even at this high level of antioxidants,
given the stochastic, random nature of ROS generation within the PUFA membranes, and
the two-three orders of magnitude difference in the molar ratio of antioxidants and PUFAs,
LPO cannot be completely suppressed, particularly in bilayers rich in long chain PUFAs which
are easier to oxidize. Indeed, the propagation rate constants (kp, M-1 s-1) for autoxidation of
PUFAs increases almost linearly with the number of bis-allylic methylene groups in a given
PUFA, i.e., linoleic acid kp = 62: arachidonic acid (Ara), 197 ± 13; eicosapentaenoic acid (EPA),
This article is protected by copyright. All rights reserved.

249 ± 16; DHA, 334 ± 37 [15]. For comparison, a monounsaturated lipid, such as oleic acid,
which has no bis-allylic –CH2- sites, has kp = 0.9 [1]. In unilamellar liposomes, the values,
obtained using the same method, were: Ara, 115 ± 7; EPA, 145 ± 8; DHA, 172 ± 13, thus
following a pattern similar to the solution values.
Hydrogen atom transfer from carbon to oxygen is a relatively slow process, because of
strong triplet repulsion or repulsion between the electrons of the CH bond and the electron
of the oxygen radical. The dichotomy between LPO-associated toxicity and inefficiency of
antioxidants has recently been addressed by utilizing a kinetic isotope effect (a slowing of
reaction kinetics as a result of isotopic substitution) to slow down the propagation step of
the LPO chain reaction [16,17]. By incorporating deuterium atoms at the bis-allylic sites of
PUFAs (D-PUFAs), they become resistant to LPO, without a change in their chemical
structure (Scheme 1). A useful “non-linear” feature of this approach is that LPO in cells is
inhibited even when the D-PUFAs are present in membranes at relatively low (around 20
molar %) levels [18], making the approach more practical [19-25]. The exact nature of the
latter effect is not fully understood, but a possibility that various pathways in a living cell
might play a role cannot be ruled out.
Previous studies of this non-linear effect have involved determination of the kinetics of
D-PUFA oxidation in organic solvents by monitoring oxygen consumption [26] and
peroxidation by-products [18,27], thus likely bearing little relevance to processes in lipid
bilayers. Studies have also been carried out in living cells [18,28], using cell survival and
oxygen consumption rate (OCR) end points, but mechanistic interpretations are complicated
by complex biological interactions following the initial LPO event. Such interactions may be
particularly unpredictable in PUFA-rich mitochondria, where besides biochemical
interactions, other confounding factors include an LPO-dependent membrane breakdown
due to high membrane potential [29] and a recently reported high physiological operating
temperature of mitochondria, which at 50oC may further increase the rate of LPO [30,31].
Here, we use unilamellar liposomes to test whether the LPO in simple non-living lipid
membrane systems is indeed inhibited by small fractions of D-PUFAs, similar to what is
observed in living cells. We find this to be the case and provide mechanistic insights into the
process. The study further justifies the use of D-PUFAs as novel therapeutic agents for
treatment of many retinal and neurological diseases [32-34], such as Alzheimer’s [35,36],
This article is protected by copyright. All rights reserved.

Parkinson’s [37], ALS [38], in addition to many other major and orphan diseases where LPO
is known to play a key detrimental role.
Results
Synthesis of phospholipids and preparation of liposomes
D2-Lin, D4-Lnn [17] and D6-Ara [43] were prepared as described previously by assembling
corresponding polyyne chains and subjecting them to catalytic hydrogenation. Although a
similar convergent synthetic strategy, based on a Wittig olefination, was developed to
obtain selectively deuterated DHA species [44], D10-DHA could be prepared using the less
laborious approach of catalytic H/D exchange [45]. D8-EPA was synthesized according to a
newly developed procedure (Scheme 2A) based on the polyacetylene approach followed by
partial hydrogenation. The efficiency of the latter was improved by using a novel catalyst,
Ni-P2 poisoned with Pb2+ (Ni-PB2), optimized as described in the Supporting Information.
Phospholipids used in this work were synthesized from lyso-phospholipid and corresponding
D- and H-PUFAs using standard protocols as shown in Scheme 2B [46,47]. Liposomes were
prepared by the standard method of extrusion of lipid mixtures through a membrane filter
of defined pore size [48,49].
Fluorescence Correlation Spectroscopy study of Fe/Asc-induced liposome leakage
In several recent publications [50-54], the permeability of vesicle membranes to fluorescent
dyes has been studied by FCS. This approach does not require loading of liposomes with
dyes at very high, self-quenching concentrations. Previously, we have shown that the
addition of Fe2+ (5 µM) together with ascorbate (100 µM) brings about permeabilization of
liposomes formed from the mixture of soybean phosphatidylcholine with bovine heart
This article is protected by copyright. All rights reserved.

cardiolipin, as monitored by changes in the fluorescence intensity autocorrelation function
(G(τ)) of the water-soluble dye SRB encapsulated in liposomes [54]. Here we studied the
effect of the combination of Fe2+ and ascorbate on G(τ) of SRB-loaded liposomes prepared
from lipid mixtures containing linoleoyl stearoyl phosphatidylcholine (H-Lin-PC) as the bulk
lipid (unless otherwise stated) with various contents of a series of 1-stearoyl
phosphatidylcholines, bearing deuterated Lin, Lnn, Ara, EPA or DHA fatty acids at position 2
of glycerol. Importantly, SRB has been shown to be resistant to oxidative damage [49]. The
amplitude of G(τ) at the limit τ ꢀ 0 is determined by the reciprocal of a mean number (N) of
fluorescent particles in the observation volume [40,41]. In dye-leakage experiments, N
comprises dye-loaded liposomes and free dye molecules released from liposomes. To
measure G(τ ꢀ 0) more precisely, we performed FCS experiments under stirring conditions
[39]. Fig. 1A shows a time dependence of the G(τ) functions of SRB for liposomes prepared
from H-Lin-PC lipids with 10 % D2-Lin-PC measured without Fe2+/ascorbate (red, dark
yellow, etc. lines) and after the addition of Fe/ascorbate (green, blue, etc. lines). Incubation
with Fe/ascorbate led to a decrease in the G(τ ꢀ 0) amplitude (compare, for example, green
and dashed blue lines in Fig. 1A), although the decrease in the G(τ ꢀ 0) amplitude took
place even without Fe2+/ascorbate. Disruption of liposomes by the addition of Triton X-100
resulted in a drop of G(τ ꢀ 0) to nearly zero (dark green (dotted) line in Fig. 1A). Fig. 1B
shows the same type of experiment carried out with liposomes having 25 % D2-Lin-PC in H-
Lin-PC. In this case, the incubation with Fe2+/ascorbate did not reduce the amplitude of G(τ
ꢀ 0) or reduced it to a very low extent (for example, dark yellow and blue long dashed
lines). The reduction of the G(τ ꢀ 0) amplitude apparently reflected an increase in the
number of fluorescent particles due to SRB release from liposomes. It should be noted that
the contribution of different fluorescent species (dye-loaded liposomes and free dye
molecules released from liposomes) to G(τ ꢀ 0) is proportional to the square of their
brightness [40,41].
This article is protected by copyright. All rights reserved.

The G(τ ꢀ 0) values at various periods of incubation with Fe2+/ascorbate (t) can be
converted into the extent of liposome leakage α using the equation [54]:
Gt (τ → 0)
(2)
α(t) =1−
G0 (τ → 0)
where G0(τꢀ0) and Gt(τꢀ0) represent G(τ) at the limit τꢀ0 at the moment just before
Fe2+/ascorbate addition (zero time) and t min after the addition, respectively. Fig. 2A
displays time courses of the extent of liposome leakage α (plus/minus Fe2+/ascorbate) for
liposomes having different % of D2-Lin-PC (namely, 0 %, 10 %, 25 %, and 100 %). The
difference between α with and without Fe2+/ascorbate was found to be high for 0 % D2-Lin-
PC (red circles and red diamonds) and was very low for 100 % of the deuterated lipid (pink
circles and pink diamonds).
Fe/ascorbate-induced liposome leakage and accumulation of conjugated dienes
To compare the liposome leakage induced by Fe2+/ascorbate with the process of lipid
peroxidation, we measured the formation of conjugated diene in the samples at 234 nm
(ΔA234) [42]. Fig. 2B shows the time courses of ΔA234 after the addition of Fe2+/ascorbate to
liposomes having various % of D2-Lin-PC. This parameter increased in time considerably for
lipid mixtures with low percentage of deuterated D2-Lin-PC and did not increase at high
content of D2-Lin-PC (Fig. 2B). Importantly, ΔA234 values were stable and did not increase in
time without the addition of Fe2+/ascorbate (data not shown).
The dependence of ΔA234 at t = 10 min (panel B) and of the difference in α in the
presence/absence of Fe2+/ascorbate (panel A) on the % of D2-Lin-PC are shown in Fig. 3. The
curves are close to each other exhibiting a sharp decrease at low % of D2-Lin-PC leading to a
very low level of ΔA234 and a very low value of the difference in α at about 20 % of D2-Lin-PC
and higher. Fig. 4 shows the results of similar experiments with liposomes made from lipids
containing arachidonic acid (H-Ara-PC) with various contents of D6-Ara-PC. Interestingly the
protecting effects of D2-Lin-PC (Fig. 3) and D6-Ara-PC (Fig. 4) exerted on the corresponding
This article is protected by copyright. All rights reserved.

H-Lin-PC and H-Ara-PC liposomes were observed in the same concentration range of about
10 to 20 % of the deuterated lipids. These data are in good agreement with the results on
the protective action of D2-Lin against LPO in membranes of yeast [18]. Surprisingly, the
protective effect of 20% D2-Lin-PC in the H-Ara-PC matrix was somewhat stronger than
expected, and quite close to that of D6-Ara-PC in H-Ara-PC (blue line, Fig 4B).
The extinction coefficient ε for conjugated dienes has been estimated to be 28000 mol-1 cm-
1
at 234 nm [42]. Therefore, for the maximal ΔA of 0.05 measured after 30 minute
incubation with Fe2+/ascorbate (Fig. 2B), the concentration of dienes formed is around 2
μM. For 10 μg/ml (or 10 μM) total lipids used, the level of conjugated dienes could
therefore be estimated to be around 20%.
Comparison of the protective effect of D2-Lin-PC in the H-Lin-PC matrix with that of
D6-Ara-PC in the H-Lin-PC matrix reveals (Fig. 5) that the protective effect of D6-Ara-PC
against H-Lin-PC liposome leakage required the same concentration of D6-Ara-PC as in Fig. 4
(10-20 %), while the effect on formation of conjugated diene required a smaller amount of
D6-Ara-PC (10 %).
We measured the accumulation of conjugated diene at various loadings of D10-DHA-
PC in the undeuterated H-Lin-PC matrix and compared them to those measured for D8-EPA-
PC, D6-Ara-PC, D4-Lnn-PC, D2-Lnn-PC and D2-Lin-PC (Fig. 6). D10-DHA-PC exhibited
substantially enhanced protection from LPO compared to D6-Ara-PC and other deuterated
lipids. The IC50 concentration was about 0.2 % for D10-DHA-PC, while IC50 for D8-EPA-PC
was about 1.5 %, about 2 % for D6-Ara-PC, about 5 % for D4-Lnn-PC, about 22 % for D2-Lnn-
PC, and about 15 % for D2-Lin-PC. Generally, increasing the number of bis-allylic CD2 groups
in a PUFA led to a stronger protective effect, more or less in a linear way. However, D10-
DHA was found to be substantially more protective than should be expected from its degree
of deuteration.
This article is protected by copyright. All rights reserved.

To assess the degree of protection elicited by D-PUFAs over a long time scale (Fig.7),
we compared the accumulation of diene conjugates (ΔA234) in liposomes consisting entirely
of D2-Lin-PC (black curve) and ΔA234 in liposomes consisting of a mixture of 20 % D2-Lin-PC
and 80 % H-Lin-PC (green curve) with ΔA234 in H-Lin-PC liposomes (red curve) as a control.
While oxidation of the 100% H-PUFA liposomes occurred on the timescale of tens of
minutes (red curve), the 100% D-PUFA liposomes were substantially more stable (black
curve), with the oxidation timescale extending to multiple days. In case of the 20% D2-Lin-
PC containing liposomes, ΔA234 reached saturation in about 30 hours (green curve).
Monte-Carlo simulation models threshold effect on chain length
Adding D-PUFAs to a non-deuterated membrane reduces the number of oxidatively
damaged molecules, but the mechanism of the steep decline and the resulting almost
complete reduction with only 20% concentration is not fully elucidated. Building a simple
model of a 2D surface populated with a selected density of D-PUFAs (which would “stop”
the chain propagation with varying probability) could shed light on the observations,
without relying heavily on the physio-chemical effects themselves. A Monte-Carlo
simulation was developed, and run while varying two different percentages, firstly the
fraction of cells which were D-PUFAs that potentially terminate the chain, and secondly, the
probability of the chain reaction stopping once a particular cell in the grid (a D-PUFA in the
H-PUFA membrane) had been encountered. The simulation results, shown in Fig. 8, indeed
reproduce the general shape and behavior of the experimental data, starting from first
principles in a highly simplistic model. Interestingly, varying the % chance that the chain is
stopped (i.e., 10% chance would mean that in 90% cases, the chain continues onwards) from
10 to 75% does not significantly shift the “knee” of the curves, somewhat compressing them
vertically instead.
This article is protected by copyright. All rights reserved.

Discussion
The key propagation step of the radical chain reaction in LPO is inhibited if the abstraction-
prone H-atoms at the bis-allylic positions are replaced with deuterium atoms [16]. We used
phospholipids with various H- and D-PUFAs at the 2-position of the sn-glycero backbone to
build liposomes with variable ratios of H- to D-PUFAs in lipid bilayers, and determined their
stability/integrity to oxidative insult. We have previously reported that a relatively small
fraction of D-PUFAs is sufficient to inhibit LPO in living cells [18,28]. The exact nature of this
“20% effect” is not fully understood. The effect has been observed by monitoring oxygen
consumption over the course of LPO in organic solutions of PUFAs [26]; while direct
measurements of LPO end-products revealed only a 12-fold isotope effect along with some
degree of LPO in the presence of 20% D-PUFAs [18,26]. In co-oxidation experiments in
organic solvents, a linear relationship was observed between the total amounts of oxidation
products formed and the percentage of D-PUFA, suggesting that D-PUFA did not act as an
antioxidant in the co-oxidation reactions and only acted as a less reactive co-substrate in the
autoxidations carried out in solution [18,27].
Studies of the non-linear protective effect in living cells using the cell survival end
point [17,18] or the oxygen consumption rate (OCR) [28], while informative, are likely
complicated by other factors linking LPO to cell death, so interpreting the data
mechanistically is not straightforward. It is far from clear what particular stage, or
product(s), of the LPO process (Scheme 1) are the most harmful to the cell and impact its
survival. It was long assumed that lipid peroxides would be the most detrimental for bilayer
integrity [55], while various carbonyl compounds (Scheme 1G), though toxic elsewhere,
were less damaging to the membrane function. However, recent data questions this
assumption, indicating that aldehydes, rather than lipid peroxides, are the group of
compounds most damaging to the membrane integrity [56].
The ability to stop LPO at as early a stage as possible is highly desirable. To further
elucidate the non-linear protective effect in the lipid bilayer format, but without
complications elicited by biochemical pathways, we now report how D-PUFAs, in the form of
synthetic phospholipids incorporated into lipid bilayers of liposomes, increase the resistance
of liposomes to oxidative stress and LPO. Various methods exist for monitoring LPO. While
This article is protected by copyright. All rights reserved.

monitoring carbonyls (HNE, HHE, MDA, etc) or isoprostanes only covers some aspects of the
LPO process, which typically proceeds through multiple pathways (Scheme 1), fluorescent
LPO markers such as BODIPY dyes have other limitations [18]. Accordingly, we chose the
liposome leakage measurement [57], i.e. membrane permeabilization, as a way of
monitoring the membrane damage process as a whole; and conjugated diene formation as
an upstream indicator of LPO progress. The liposome leakage method [58] is a convenient
way to monitor lipid bilayer integrity. However, the relatively large size of SRB may require a
rather extensive degree of membrane destruction for it to leak out, which may be too
substantial compared to the damage which occurs to membranes in vivo, even at
pathological conditions. We therefore also measured the formation of conjugated dienes
(Fig. 1E) [59] to more directly follow LPO.
To compare the relative inhibitory efficiency of various D-PUFAs, we tested 1-
stearoyl-phosphatidylcholines containing 11,11-d2-linoleic (D2-Lin-PC), 11,11,14,14-d4-
linolenic (D4-Lnn-PC), 14,14-d2-linolenic (D2-Lnn-PC), 7,7,10,10,13,13-d6-arachidonic (D6-
Ara-PC),
7,7,10,10,13,13,16,16-d8-eicosapentaenoic
(D8-EPA-PC)
and
6,6,9,9,12,12,15,15,18,18-d10-docosahexaenoic (D10-DHA-PC) acids at position 2 of glycerol.
Formation of liposomes and their subsequent processing are typically performed in air. To
minimize the non-specific oxidation of H-PUFA-PCs during liposome preparation, we used H-
Lin-PC as a non-deuterated bulk component of bilayers in all studies, except in Fig. 4, where
H-Ara-PC was used. In the series of deuterated lipids studied here, the relative efficacy of
the various D-PUFA-PCs to protect H-Lin-PC from oxidation increased with increasing
number of bis-allylic CD2 groups in the D-PUFA (assuming the LPO inhibition results in a 90%
decrease in the ΔA234 measured) as is obvious from comparing the D2-Lin-PC (at ~ 20-25% in
bilayer), D2-Lnn-PC (~ 40%), D4-Lnn-PC (~ 15-20%), D6-Ara-PC (~ 10-15%) and D8-EPA-PC (~
7-10%) series (Fig. 6). D10-DHA-PC also followed this trend. However, the protection
rendered by D10-DHA-PC, suppressing LPO at as low as 1% level of incorporation in the H-
Lin-PC matrix (Fig.6), was substantially stronger (10-fold stronger than for D6-Ara as
measured by the conjugated diene assay) than should be expected simply from the total
number of its CD2 groups. It must be acknowledged that incorporation of DHA may lead to
some physical changes in the structure of the bilayer arising from the difference in chain
length between the two PUFAs (Lin: 18 carbons; DHA: 22 carbons), although the increased
This article is protected by copyright. All rights reserved.

number of double bonds would compensate somewhat for the length difference, by
effectively “shortening” the chain. Moreover, it is known that DHA occupies a larger
effective volume within the bilayer due to its curved configuration and lateral rotation, as
well as other anomalous properties [60] such as its high melting point (Ara, -49.5oC; EPA, -
54oC; DHA, -44oC). However, it is unclear why such physical changes would make the bilayer
more resistant to oxidation.
The importance of the ratio of CD2 to CH2 at the bis-allylic sites in the lipid bilayer is
further supported by the observation that the percentage of D6-Ara-PC, needed for
protection of H-Ara-PC matrix, is similar to that found for the case of D2-Lin-PC in the H-Lin-
PC matrix (Fig. 4A,B; Fig. 6) However, a surprisingly low level of D2-Lin-PC prevented the
damage in H-Ara-PC matrix (Fig. 4B, blue). Furthermore, the percentage of D2-Lnn-PC
needed for protection of the H-Lin-PC is about two-fold higher than that of D4-Lnn-PC (dark
yellow and black curves in Fig.6). Increasing the total number of such “CD2 equivalents”
seems to “buffer” the chain propagating radicals, slowing them down and inhibiting LPO.
This is consistent with two important earlier observations. First, mono-deuterated 11,11-
D,H-Lin, bearing one H-atom and one D-atom at the same bis-allylic site protects against
LPO, even though a hydrogen atom is available for abstraction. This seems to suggest a
decreased reactivity of the intermediate pentadienyl and/or dienylperoxyl radicals that
contain a deuterium atom (a possible role in this case of the secondary KIE was analysed in
[18]). Second, a mixture of non-oxidizable oleic acid (Ole) with H-Lin (1:1) is toxic to cells
while a mixture of D2-Lin with H-Lin (1:1) is not. This suggests that D2-Lin itself is not just a
stable, non-oxidizable component of the mixture (like Ole is), but is somehow actively
participating in stopping Lin from undergoing LPO (these, and other observations are
discussed in detail in [18]). Hence it seems that the protective effect of D-PUFAs may have
two parts: (a) a kinetic isotope effect on abstraction from the bis-allylic site, and (b) a
difference in the properties of the resultant D-containing radicals as compared to H-
containing radicals. The isotope effect on the atom abstraction step may be an aggregate of
several effects, including both a primary and a secondary KIE. Indeed, abstraction of
hydrogen from H-Lin by tocopheryl radical is 23 times faster than abstraction of D from D2-
Lin, while the corresponding intramolecular KIE for reaction of tocopheryl radical with
monodeuterated 11,11-D,H-Lin is only 8.9 [27]. However, it is more challenging to
This article is protected by copyright. All rights reserved.

rationalize different reactivities of the D-containing radicals compared to H-containing
radicals in terms of our current understanding of the mechanism of LPO (see [62] for a
comprehensive recent review). Moreover, it is difficult to explain how isotopic substitution
may impact the rate of formation of products which derive from these intermediates, such
as carbonyl compounds which may derive from Hock fragmentation [61], peroxyl radical
termination and/or alkoxyl radical β-scission [56,63]. At this stage, the possibility of a
secondary isotope effect on the conjugated pentadienyl system, or even some form of
hyperconjugation [64] between the unpaired electron and deuterium in the intermediate
peroxyl, although unlikely, cannot be fully ruled out.
The threshold mechanism of protection also emerges from the Monte-Carlo
simulation, performed for different “deuteration” levels, showing a similarity with the
oxygen consumption rate curve for D2-Lin:H-Lin mixtures oxidation [26], particularly the
“knee” in the chain length curve (Fig.7 and SUPPORTING INFO). Oxidation of liposomes
containing pure, 100% D2-Lin-PC, was found to proceed at substantially longer (multiple
days) timescale (data not shown).
A different kind of a threshold occurs when the integrity of a bilayer under oxidative
stress is monitored over a longer period of time (Fig 7). Here, the 20% fraction of D2-Lin,
protective over a shorter time scale (Fig 6), holds up (green curve) in a way similar to that
for a fully deuterated bilayer (black curve), but after approx. 20-24 h, succumbs to oxidative
damage over a fairly narrow time window (approx. 6 h), so that by 30 h the 20% D2-Lin
bilayer is as damaged as the non-deuterated control (red curve), implying that by approx. 24
h the D-PUFA has been mostly consumed. Confirming this by mass-spectrometry may be
required, while more detailed elucidation of the mechanism is likely to be complicated by a
possible variable rate of initiation over time, and further difficulties arising from the fact
that accumulating LPO products are often more oxidizable than the starting PUFAs.
The present study showed similarity between concentrations of deuterated
phospholipids that were effective in suppressing LPO [17,18,28] and those providing
protection of membrane integrity (Figs. 3 and 4). We did not quantify oxidized
phospholipids or other LPO products, relying instead on the well-known correlations
between membrane leakage and conjugated diene formation, “bird’s eye view” proxies of
This article is protected by copyright. All rights reserved.

the membrane integrity and LPO product formation, respectively [56,57,65]. At present, it
remains unclear what degree of LPO is required to impair bilayer function, and also, which
particular LPO products are responsible for the induction of membrane leakiness and loss of
integrity. Some have suggested that lipid hydroperoxides, the primary LPO products, are
responsible [55], while others have suggested that fragmentation products, such as lipid
aldehydes with truncated fatty acid tails are key [56]. Therefore, the mechanism of
peroxidative permeabilization of lipid membranes may vary depending on the kind of LPO
induction and the presence of different protectors. Of relevance to this point could be a
noticeable difference between the contents of D6-Ara-PC in the H-Lin-PC matrix that were
effective in protecting against LPO (IC50 about 10 %) and SRB leakage (IC50 about 2.5 %), as
seen in Fig.5, in contrast to Figs. 3 and 4.
In conclusion, the foregoing clearly demonstrates a strong protective effect of small
amounts of deuterated PUFA in liposomal lipid bilayers, the phenomenon previously
described in homogenous solution [26], cells [66,67,69] and living systems [18,28]. For each
D-PUFA studied here, there exists a threshold percentage (e.g., about 20-25% for D2-Lin-PC
and much less for D10-DHA-PC) in the H-PUFA matrix, inhibiting the LPO rate, which may be
attributed to the ability of a D-PUFA-derived radical to interrupt LPO chain reactions in a
lipid bilayer. The degree of protection of non-deuterated PUFAs thus correlates with the
degree of deuteration, or the total number of -CD2- equivalents, present in the lipid bilayer.
This useful property of D-PUFAs justifies their evaluation as potential preventive and/or
therapeutic agents for various diseases in which LPO has been implicated, as attested by
ongoing and completed [23] trials in neurological diseases as well as several positive studies
in animal models [19-22,24,25]. Lastly, the current study suggests applications of D-PUFAs in
liposome-based drug delivery approaches [68], as both the drug vehicles, and the drugs to
be delivered.
This article is protected by copyright. All rights reserved.

Experimental procedures
Materials
Soybean lysophosphatidylcholine (Lipoid) and other commercial reagents and solvents were
used as received. Chloroform was washed with water, dried with CaCl2 and distilled over
1
P2O5. H (500 MHz) and 13C (126 MHz) NMR spectra were obtained with Bruker DRX-500
1
spectrometer and referenced to residual solvent signals (CDCl3: 7.26 ppm for H; CD3OD:
4.87 ppm for 1H and 49.00 ppm for 13C). Analytical thin-layer chromatography was
performed on TLC Silica gel 60 F254 (Merck). Silica gel column chromatography was
performed using Merck Kieselgel 60 0.063–0.200 nm. Other reagents were from Sigma-
Aldrich.
Synthesis of Ethyl 7,7,10,10,13,13,16,16-D8-Eicosapentaenoate
1,1,4,4-D4-Octa-2,5-diyn-1-ol (2). 250 mL round-bottom flask was charged with DMF (30
mL), CuI (10.0 g, 52 mmol), KI (9.0 g, 54 mmol) and K2CO3 (11.0 g, 80 mmol), then bromide 1
(10.0 g, 67 mmol) and 1,1-D2-propargyl alcohol (89 %, 4.90 g, 75 mmol) were added. The
mixture was stirred overnight under argon at room temperature and quenched with sat.
NH4Cl (80 mL), sat. NaCl (60 mL) and hexanes/EtOAc (60 mL, 3:1 v/v). After 15 min, the
residue was filtered on the fine glass filter and washed with hexanes/EtOAc (3:1 v/v). The
organic layer was separated, and the aqueous layer was extracted with hexanes/EtOAc (3:1
v/v, 6×50 mL). Combined organic layers were concentrated, filtered through a silica plug in
hexanes/Et2O (2:1 v/v) to give compound 2 as a colorless liquid. Yield: 6.88 g (82 %). 1Н NMR
(CDCl3, 500 MHz) δ 2.15 (q, J = 7.5 Hz, 2H), 2.02 (s, 1H), 1.10 (t, J = 7.5 Hz, 3H). 13С NMR
(CDCl3, 126 MHz) δ 82.56, 80.73, 78.48, 72.77, 13.90, 12.42.
This article is protected by copyright. All rights reserved.

1,1,4,4-D4-1-Bromoocta-2,5-diyne (3). The solution of 2 (6.85 g, 54 mmol) in Et2O (45 mL)
was cooled to –10 °С, then Et3N (8.7 mL, 63 mmol) and MsCl (4.80 mL, 62 mmol) were
added dropwise. The reaction was stirred under –10 °С for 10 min and at rt for 30 min. To
convert the obtained mesylate to bromide 3, MgBr2 in Et2O was prepared by adding 1,2-
dibromoethane (25.5 g, 136 mmol) in Et2O (60 mL) to magnesium turnings (3.35 g, 140
mmol) in Et2O (30 mL) and added to the reaction mixture. The mixture was refluxed for 30
min. The reaction was quenched with ice-cold water (30 mL), and the residue was dissolved
(ether boiling!). The organic layer was separated, and the aqueous layer was extracted with
Et2O (3×50 mL). Combined organic layers were concentrated, filtered through a silica plug in
hexanes to give compound 3 as a pale-yellow liquid. Yield: 9.86 g (96 %). 1Н NMR (CDCl3, 500
13
MHz) δ 2.16 (q, J = 7.5 Hz, 2H), 1.11 (t, J = 7.5 Hz, 3H). С NMR (CDCl3, 126 MHz) δ 82.80,
82.15, 75.30, 72.24, 13.92, 12.46.
1,1,4,4,7,7-D6-Undeca-2,5,8-triyn-1-ol (4). The compound was synthesized according to the
procedure for 2, using bromide 3 (9.83 g, 52 mmol), 1,1-D2-propargyl alcohol (89 %, 3.91 g,
60 mmol), CuI (7.6 g, 40 mmol), KI (7.0 g, 42 mmol) and K2CO3 (8.6 g, 62 mmol) in DMF (23
mL). The extraction was carried out using hexanes/EtOAc (3:1 v/v, 10×50 mL). Hexanes/Et2O
(1:1 v/v) was used as an eluent for chromatography on silica. Crude 4 as pale-yellow oil was
mixed with hexanes and cooled at 0 °С overnight. The crystalline solid was filtered, washed
1
with cold hexanes and dried in vacuo. Yield: 6.47 g (75 %). Н NMR (CDCl3, 500 MHz) δ 2.15
13
(q, J = 7.5 Hz, 2H), 1.91 (s, 1Н), 1.10 (t, J = 7.5 Hz, 3H). С NMR (CDCl3, 126 MHz) δ 82.41,
80.06, 78.79, 75.58, 73.78, 72.98, 13.93, 12.45.
1,1,4,4,7,7-D6-1-Bromoundeca-2,5,8-triyne (5). The compound was synthesized according
to the procedure for 3, using compound 4 (6.39 g, 39 mmol), MsCl (4.98 g, 43.5 mmol), Et3N
(6.1 mL, 43.9 mmol) in Et2O (40 mL), Mg turnings (2.38 g, 100 mmol) in Et2O (20 mL) and 1,2-
dibromoethane (18.2 g, 97 mmol) in Et2O (40 mL). When filtering through a silica plug, 5 %
1
Et2O/hexanes was used as an eluent. Yellow oil. Yield: 8,45 g (96 %). Н NMR (CDCl3, 500
13
MHz) δ 2.16 (q, J = 7.5 Hz, 2H), 1.11 (t, J = 7.5 Hz, 3H). С NMR (CDCl3, 126 MHz) δ 82.46,
81.41, 75.84, 75.63, 73.29, 72.92, 13.97, 12.49.
This article is protected by copyright. All rights reserved.

1,1,4,4,7,7,10,10-D8-Tetradeca-2,5,8,11-tetrayn-1-ol (6). The compound was synthesized
according to the procedure for 2, using bromide 5 (8.42 g, 37 mmol), 1,1-D2-propargyl
alcohol (89 %, 2.76 g, 42 mmol), CuI (5.4 g, 28 mmol), KI (5.0 g, 30 mmol) and K2CO3 (6.1 g,
44 mmol) in DMF (15 mL). The extraction was carried out using hexanes/EtOAc (3:1 v/v,
7×50 mL). Crude 6 was mixed with hexanes and cooled at 0 °С overnight. The crystalline
solid was filtered, washed with cold hexanes and dried in vacuo to give white crystals (mp
1
63-65 °С). Yield: 3.25 g (43 %). Н NMR (CDCl3, 500 MHz) δ 2.15 (q, J = 7.5 Hz, 2H), 1.81 (s,
13
1Н), 1.10 (t, J = 7.5 Hz, 3H). С NMR (CDCl3, 126 MHz) δ 82.36, 79.97, 78.85, 75.39, 74.90,
74.11, 74.00, 73.04, 13.93, 12.45.
1,1,4,4,7,7,10,10-D8-1-Bromotetradeca-2,5,8,11-tetrayne (7). The compound was
synthesized according to the procedure for 3, using compound 6 (3.20 g, 16 mmol), MsCl
(2.00 g, 17.5 mmol), Et3N (2.1 mL, 17.7 mmol) in Et2O (20 mL), Mg turnings (0.95 g, 40
mmol) in Et2O (8 mL) and 1,2-dibromoethane (7.33 g, 39 mmol) in Et2O (16 mL). When
filtering through a silica plug, Et2O was used as an eluent. White crystals (mp 52-56 °С).
1
Yield: 3,74 g (91 %). Н NMR (CDCl3, 500 MHz) δ 2.16 (q, J = 7.5 Hz, 2H), 1.10 (t, J = 7.5 Hz,
3H). 13С NMR (CDCl3, 126 MHz) δ 82.36, 81.28, 75.67, 75.43, 75.13, 73.94, 73.61, 73.03,
13.96, 12.47.
Ethyl 7,7,10,10,13,13,16,16-D8-eicosa-5,8,11,14,17-pentaynoate (8). Round-bottom flask
was charged with DMF (25 mL), CuI (4.9 g, 25 mmol), KI (7.3 g, 44 mmol) and K2CO3 (5.2 g, 38
mmol), then bromide 7 (3.69 g, 13.7 mmol) and ethyl-5-hexynoate (1.92 g, 13.7 mmol) were
added. The mixture was stirred overnight under argon at room temperature and quenched
with sat. NH4Cl (55 mL), sat. NaCl (40 mL) and hexanes/EtOAc (40 mL, 4:1 v/v). After 15 min,
the residue was filtered on the fine glass filter and washed with hexanes/EtOAc (4:1 v/v).
The organic layer was separated, and the aqueous layer was extracted with hexanes/EtOAc
(4:1 v/v, 3×30 mL). Combined organic layers were concentrated, filtered through a silica plug
1
in 5 % EtOAc/hexanes to give compound 8 as a yellow liquid. Yield: 3.41 g (76 %). Н NMR
(CDCl3, 500 MHz) δ 4.11 (q, J = 7.2 Hz, 2H), 2.39 (t, J = 7.5 Hz, 2H), 2.21 (t, J = 7.0 Hz, 2H),
2.14 (q, J = 7.5 Hz, 2H), 1.78 (p, J = 7.2 Hz, 2H), 1.24 (t, J = 7.2 Hz, 3H), 1.09 (t, J = 7.5 Hz, 3H).
This article is protected by copyright. All rights reserved.

13С NMR (CDCl3, 126 MHz) δ 82.29, 79.67, 75.29, 75.10, 74.67, 74.56, 74.49, 74.22, 74.09,
73.05, 60.43, 33.23, 23.96, 18.27, 14.33, 13.93, 12.44.
Ethyl 7,7,10,10,13,13,16,16-D8-(all-Z)-eicosa-5,8,11,14,17-pentaenoate (9). Nickel acetate
tetrahydrate (380 mg) was dissolved in 96 % EtOH (6 mL) at 50-60 °C. The flask was filled
with hydrogen, and a solution of NaBH4 in EtOH (1.57 mL), obtained from NaBH4 (87 mg)
and EtOH (2 mL) after stirring for 15 min and filtering, was added dropwise during 15 min.
After 15 min, Pb(NO3)2 (5.8 mg) in water (40 μL), 1,2-ethylenediamine (0.47 mL,) and ester 8
(3.38 g, 10 mmol) in EtOAc (2.5 mL) were added. The mixture was stirred under hydrogen (1
atm) at 0 °C until the absorption of hydrogen stopped (3.5 h), diluted with hexanes (15 mL),
glacial AcOH (0.66 mL), water (0.18 mL). Organic layer was separated; aqueous layer was
extracted with hexanes/EtOAc (4:1, 4×40 mL). Combined organic layers were washed with
1M H2SO4, sat. NaHCO3 and sat. NaCl (15 mL each) and dried over Na2SO4. Solvents were
evaporated to provide a crude ester 9 (826 mg, 80 %), which was further chromatographed
on AgNO3-silica using Et2O/hexanes mixtures (1:9, 2:8, 3:7, 4:6, 5:5, 8:2, 10:0) as an eluent
to wash out the impurities. Then EtOAc was used to elute pure 9. The sorbent was prepared
by drying the mixture of SiO2 (15 g) and AgNO3 (4 g) in acetonitrile (10 mL) on rotary
evaporator and evacuating it at 15 mmHg and 80 °C for 3 h. Pure fractions (as analyzed by
TLC) were collected, evaporated and filtered through a silica plug in 10 % EtOAc in hexanes
1
to give 9 as a colorless liquid. Yield: 330 mg (32 %). Н NMR (CDCl3, 500 MHz) δ 5.37 (m,
10H), 4.12 (q, J = 7.1 Hz, 2H), 2.30 (t, J = 7.6 Hz, 2H), 2.09 (m, 4H), 1.70 (p, J = 7.5 Hz, 2H),
13
1.25 (t, J = 7.2 Hz, 3H), 0.97 (t, J = 7.6 Hz, 3H). С NMR (CDCl3, 126 MHz) δ 173.71, 132.17,
129.16, 128.81, 128.60, 128.29, 128.16, 127.93, 127.03, 60.34, 33.84, 26.70, 24.94, 20.69,
14.38.
This article is protected by copyright. All rights reserved.

Synthesis of D-PUFA-containing phospholipids
Hydrogenated lysophosphatidylcholine (H-LPC, 11). Soybean lysophosphatidylcholine (3.5
g) was dissolved in MeOH (50 mL) and 5 % Pd/C (350 mg) was added to the solution. The
mixture was stirred under H2 (1 atm) for 1 h, filtered through a silica plug and cooled to 0 °C
overnight. The resulting crystalline product was filtered off, washed with cold MeOH and
dried in vacuo. The filtrate was evaporated to give another portion of product. Yield 3.39 g
(95 %). 1H NMR (CD3OD) δ 4.28 (m, 2H), 4.17 (dd, J = 11.3, 4.6 Hz, 2H), 4.10 (dd, J = 11.3, 6.4
Hz, 2H), 4.00-3.85 (m, 3H), 3,64 (m, 2H), 3.34 (s, 2H), 3.30 (p, J = 1.7 Hz, 3H), 3.22 (s, 9H),
2.35 (t, J = 7.5 Hz, 2H), 1.61 (p, J = 7.2 Hz, 2H), 1.28 (m, 28H), 0.89 (t, J = 6.9 Hz, 3H). 13C NMR
(CD3OD) δ 175.34, 69.79, 67.84, 67.47, 66.22, 60.45, 54.66, 34.91, 33.09, 30.79, 30.64,
30.49, 30.45, 30.25, 26.00, 23.75, 14.46.
12a: RCOOH = linoleic acid
12b: RCOOH = 11,11-d2-linoleic acid
12c: RCOOH = arachidonic acid
12d: RCOOH = 7,7,10,10,13,13-d6-arachidonic acid
12e: RCOOH = linolenic acid
12f: RCOOH = 11,11,14,14-d4-linolenic acid
12g: RCOOH = 14,14-d2-linolenic acid
12h: RCOOH = 7,7,10,10,13,13,16,16-d8-eicosapentaenoic acid
12i: RCOOH = 6,6,9,9,12,12,15,15,18,18-d10-docosahexaenoic acid
Acylated H-LPCs (12a-i). In a typical experiment, 11 (300 mg, 0.57 mmol) was dissolved in
CHCl3 (10 mL), then fatty acid (1.15 mmol), 4-DMAP (140 mg, 1.15 mmol) and DCC (236 mg,
1.15 mmol) were added. The reaction mixture was stirred at 20 °C for 5 days, the solids
were filtered off and washed with CHCl3. The solvent was evaporated and the residue was
chromatographed on silica in CHCl3, CHCl3 — MeOH (2:1, 1:1, 1:2) and MeOH. The pure
fractions were evaporated to give the product as a colourless waxy substance. Rf 0.40
(CHCl3, MeOH, conc. NH3 13:5:1).
This article is protected by copyright. All rights reserved.

1
1-Acyl-2-linoleyl-sn-glycero-3-phosphatidylcholine (H-Lin-PC, 12a). Yield 344 mg (77 %). H
NMR (CDCl3) δ 5.36 (m,4H), 4.41(m,1H), 4.30(m,2H), 4.11 (dd, J = 12.1, 7.4 Hz, 1H), 4.00-3.85
(m, 2H), 3,76 (m,2H), 3.33 (m,9H), 2.80 (m, 2H), 2.28 (m, 4H), 2.10 (m,4H), 1.58 (m, 4H), 1.24
(m, 40H), 0.97 (m, 6H). MS (ESI) m/z calcd. (M+HCOO–) 830.61, found 830.50.
1-Acyl-2-(11,11-d2-linoleyl)-sn-glycero-3-phosphatidylcholine (D2-Lin-PC, 12b). Yield 364
mg (81 %). 1H NMR (CDCl3) δ 5.36 (m, 4H), 4.41 (m, 1H), 4.30 (m, 2H), 4.11 (dd, J = 12.1, 7.4
Hz, 1H), 4.00-3.85 (m, 2H), 3,76 (m, 2H), 3.33 (m, 9H), 2.28 (m, 4H), 2.10 (m, 4H), 1.58 (m,
4H), 1.24 (m, 40H), 0.97 (m, 6H). MS (ESI) m/z calcd. (M+HCOO–) 832.62, found 832.42.
1-Acyl-2-arachidonoyl-sn-glycero-3-phosphatidylcholine (H-Ara-PC, 12c). Yield 381 mg (82
1
%). H NMR (CDCl3) δ 5.36 (m, 8H), 5.19 (m, 1H), 4.41 (m, 1H), 4.30 (m, 2H), 4.11 (dd, J =
12.1, 7.4 Hz, 1H), 4.00-3.85 (m, 2H), 3,76 (m, 2H), 3.33 (m, 9H), 2.80 (m, 6H), 2.28 (m, 4H),
2.10 (m, 4H), 1.58 (m, 4H), 1.24 (m, 38H), 0.97 (m, 6H). MS (ESI) m/z calcd. (M+HCOO–) 854.60,
found 854.42.
1-Acyl-2-(7,7,10,10,13,13-d6-arachidonoyl)-sn-glycero-3-phosphatidylcholine (D6-Ara-PC,
1
12d). Yield 350 mg (75 %). H NMR (CDCl3) δ 5.36 (m, 8H), 5.19 (m, 1H), 4.41 (m, 1H), 4.30
(m, 2H), 4.11 (dd, J = 12.1, 7.4 Hz, 1H), 4.00-3.85 (m, 2H), 3,76 (m, 2H), 3.33 (m, 9H), 2.28
(m, 4H), 2.10 (m, 4H), 1.58 (m, 4H), 1.24 (m, 38H), 0.97 (m, 6H). MS (ESI) m/z calcd. (M+HCOO–)
860.64, found 860.58.
1-Acyl-2-linolenyl-sn-glycero-3-phosphatidylcholine (H-Lnn-PC, 12e). Yield 330 mg (74 %).
1H NMR (CDCl3) δ 5.34 (m, 6H), 5.19 (m, 1H), 4.38 (m, 1H), 4.28 (m, 2H), 4.11 (dd, J = 12.1,
7.4 Hz, 1H), 4.00-3.85 (m, 2H), 3,76 (m, 2H), 3.33 (m, 9H), 2.80 (m, 4H), 2.28 (m, 4H), 2.10
(m, 4H), 1.58 (m, 4H), 1.24 (m, 38H), 0.97 (t, J = 7.5 Hz, 3H), 0.89 (t, J = 6.9 Hz, 3H). MS (ESI)
m/z calcd. (M+HCOO–) 828.59, found 828.42.
This article is protected by copyright. All rights reserved.

1-Acyl-2-(11,11,14,14-D4-linolenyl)-sn-glycero-3-phosphatidylcholine (d4-Lnn-PC, 12f). Yield
1
410 mg (90 %). H NMR (CDCl3) δ 5.34 (m, 6H), 5.19 (m, 1H), 4.38 (m, 1H), 4.28 (m, 2H), 4.11
(dd, J = 12.1, 7.4 Hz, 1H), 4.00-3.85 (m, 2H), 3,76 (m, 2H), 3.33 (m, 9H), 2.28 (m, 4H), 2.05
(m, 4H), 1.58 (m, 4H), 1.24 (m, 38H), 0.97 (t, J = 7.5 Hz, 3H), 0.89 (t, J = 6.9 Hz, 3H). MS (ESI)
m/z calcd. (M+HCOO–) 832.61, found 832.42.
1-Acyl-2-(14,14-D2-linolenyl)-sn-glycero-3-phosphatidylcholine (d2-Lnn-PC, 12g). Yield 366
mg (81 %). 1H NMR (CDCl3) δ 5.34 (m, 6H), 5.19 (m, 1H), 4.38 (m, 1H), 4.28 (m, 2H), 4.11 (dd,
J = 12.1, 7.4 Hz, 1H), 4.00-3.85 (m, 2H), 3,76 (m, 2H), 3.39 (m, 9H), 2.80 (m, 2H), 2.28 (m,
4H), 2.10 (m, 4H), 1.58 (m, 4H), 1.24 (m, 38H), 0.97 (t, J = 7.5 Hz, 3H), 0.89 (t, J = 6.9 Hz, 3H).
MS (ESI) m/z calcd. (M+HCOO–) 830.60, found 830.42.
1-Acyl-2-(7,7,10,10,13,13,16,16-D8-eicosapentaenoyl)-sn-glycero-3-phosphatidylcholine
1
(d8-EPA-PC, 12h). Yield 371 mg (80 %). H NMR (CDCl3) δ 5.36 (m, 10H), 5.19 (m, 1H), 4.39
(dd, J = 12.0, 2.9 Hz, 1H), 4.28 (m, 2H), 4.11 (dd, J = 12.1, 7.3 Hz, 1H), 4.00-3.85 (m, 2H), 3,76
(m, 2H), 3.33 (m, 9H), 2.31 (t, J = 7.6 Hz, 2H), 2.26 (t, J = 7.6 Hz, 2H), 2.10 (m, 2H), 1.66 (m,
2H), 1.58 (m, 2H), 1.24 (m, 28H), 0.97 (t, J = 7.5 Hz, 3H), 0.89 (t, J = 6.9 Hz, 3H). MS (ESI) m/z
calcd. (M+HCOO–) 860.64, found 862.50.
1-Acyl-2-(6,6,9,9,12,12,15,15,18,18-d10-docosahexaenoyl)-sn-glycero-3-
1
phosphatidylcholine (D10-DHA-PC, 12i). Yield 390 mg (81%). H NMR (CDCl3) δ 5.36
(m,12H), 5.19 (m,1H), 4.38 (m, 1H), 4.28 (m, 2H), 4.12 (dd, J = 12.1, 7.4 Hz, 1H), 4.00-3.85
(m,2H), 3,76 (m, 2H), 3.67 (m, 4H), 3.33 (m, 9H), 2.36 (m, 4H), 2.28 (m, 2H), 1.58 (m, 2H),
1.24 (m,28H), 0.97 (t, J = 7.5 Hz, 3H), 0.89 (t, J = 6.9 Hz, 3H). MS (ESI) m/z calcd. (M+HCOO–)
888.67, found 889.58.
This article is protected by copyright. All rights reserved.

Preparation of liposomes
Plain and dye-loaded unilamellar liposomes were prepared by evaporation under a
stream of nitrogen of a 2 % solution of a mixture of lipids in chloroform followed by
hydration with a buffer solution containing a fluorescent marker for dye-loaded liposomes.
We used 5 mg of appropriate phosphatidylcholine mixture and added 0.5 ml of 1 mM
sulforhodamine B (SRB) in 100 mM KCl, 10 mM Tris, 10 mM MES, pH 7.4. The mixture was
vortexed, passed through several cycles of freezing and thawing, and extruded through 0.1-
μm pore size Nucleopore polycarbonate membranes using an Avanti Mini-Extruder. The
unbound fluorescent marker was then removed by passage through a Sephadex G-50 coarse
column with a buffer solution containing 100 mM KCl, 10 mM Tris, 10 mM MES, pH 7.4.
Fluorescence correlation spectroscopy
The custom-made setup was described previously [39]. Briefly, fluorescence
excitation and detection utilized a Nd:YAG solid state laser with a 532-nm beam attached to
an Olympus IMT-2 epifluorescent inverted microscope equipped with a 40x, NA 1.2 water
immersion objective (Carl Zeiss, Jena, Germany). The fluorescence light passed through an
appropriate dichroic beam splitter and a long-pass filter and was imaged onto a 50-μm core
fiber coupled to an avalanche photodiode (SPCM-AQR-13-FC, PerkinElmer Optoelectronics,
Vaudreuil, Quebec, Canada). The output signal was sent to a PC using a fast interface card
(Flex02-01D/C, Correlator.com, Bridgewater, NJ). Data acquisition time was 30 s.
Fluorescence was recorded from the confocal volume located at about 50 μm above the
coverslip surface with 50 μl of the buffer solution added. Most of the data were collected
under the conditions of stirring liposome dispersions with a paddle-shaped 3-mm plastic bar
rotated at 600 rpm. To calibrate the setup, we recorded the fluorescence autocorrelation
function of Rhodamine 6G solution. Assuming the diffusion coefficient of the dye to be 2.5 x
10-6 cm2/s, the value of the confocal radius ω=0.42 μm was obtained. The correlated
This article is protected by copyright. All rights reserved.

fluorescence emission signals were fitted to the three-dimensional autocorrelation function
[40,41].


















1
1
1




G(τ ) =
(1)
ω2τ
z02τD
τ
N
1+




1+
τD
with τD being the characteristic correlation time during which a molecule resides in the
observation volume of radius ω and length z0, given by τD = ω2/4D, where D is the diffusion
coefficient, N is the mean number of fluorescent particles in the confocal volume. The
amplitude of the autocorrelation function is inversely proportional to the number of
fluorescent particles (N=1/G(τꢀ0)), but is independent of the fluorescence intensity of a
single particle (in a system of identical particles) and therefore does not depend on the
number of fluorophores per vesicle. Particles can be any fluorescent “point objects” in
comparison to the dimension of the observation volume (i.e. about 1 micrometer).
Therefore, particles can be single molecules of dye (i.e. SRB), as well as liposomes carrying
different numbers of dye molecules. Initially (before the leakage induction) the system has a
limited number of particles per observation volume comprising predominantly several
liposomes loaded with the dye. After the leakage, the number of particles increases
substantially, because every liposomal particle produces thousands of particles of free dye
leading to a significant decrease of the parameter G(τꢀ0).
Conjugated diene assay
The level of conjugated dienes was determined from the absorption at 234 nm [42].
The UV spectrum of the liposome suspension was measured with a CM 2203
spectrophotometer (SOLAR, Belarus) at various time intervals after the addition of 5 µM
FeSO4 and 100 µM ascorbate, subtracting the initial spectrum at time zero. The peak of
conjugated dienes was seen as a shoulder in the broad band of ascorbate. As a measure of
the level of diene conjugates, the absorbance value at 234 nm was taken from which the
This article is protected by copyright. All rights reserved.

averaged absorbance at 228 nm and 240 nm was subtracted, namely ΔA234=A234
-
(A228+A240)/2. The data points represented Mean±S.D. of at least three independent
experiments.
Monte Carlo simulation
A simulation program that roughly modeled the effect of D-PUFAs on the lipid chain reaction
was written using Processing 3.0. A rectangular grid with 400 x 400 cells was populated
randomly with given concentration of D-PUFA. The chain reaction was initiated at the central
cell, thereafter a 2D random walk simulation proceeded to one of the eight neighboring cells
with equiprobability, damaging encountered H-PUFAs with 100% probability until a D-PUFA
was hit. The chain was then randomly stopped based on a parameter (“strength”, a
parameter that represents the probability that the chain terminates) of the model, which
embodies a combination of those properties that quench the propagating peroxyl radical. An
average of 100 runs was performed for each set of parameters.
Acknowledgements
We are very grateful to Prof Cathy F. Clarke (UCLA, USA) and Prof Derek Pratt (Ottawa
University, Canada) for providing their insights on the manuscript. This work was financially
supported by the Russian Science Foundation Grant 16-14-10025.
Author Contributions:
MSS initiated the project; YNA and MSS designed the experiments, and YNA, MSS and DAP
interpreted the data; AMF performed biophysical and biochemical experiments; VVS and DV
designed schemes of synthesis of D-PUFA-containing lipids; MAF, AVB and OLS performed
synthesis of the deuterated lipids; HJS performed the Monte-Carlo simulation; MSS, YNA,
EAK and DAP wrote the manuscript.
This article is protected by copyright. All rights reserved.

Disclaimers
H.J. Saal and M.S. Shchepinov are employed by, and hold stocks in Retrotope, Inc.
References
1. Yin H, Xu L, Porter NA (2011) Free radical lipid peroxidation: mechanisms and
analysis. Chem Rev 111, 5944–5972.
2. Esterbauer H, Schaur RJR, Zollner H (1991) Chemistry and biochemistry of 4-
hydroxynonenal, malonaldehyde and related aldehydes. Free Radic Biol Med 11, 81–
128.
3. Halliwell B, Gutteridge JMC (2007) Free Radicals in Biology and Medicine, Oxford
University Press, New York.
4. Roberts LJ, Fessel JP (2004) The biochemistry of the isoprostane, neuroprostane, and
isofuran pathways of lipid peroxidation. Chem Phys Lipids 128, 173–186, doi:
10.1016/j.chemphyslip.2003.09.016.
5. Negre-Salvayre A, Coatrieux C, Inguenau C, Salvayre R (2008) Advanced lipid
peroxidation end products in oxidative damage to proteins: potential role in diseases
and therapeutic prospects for the inhibitors. Br J Pharmacol 153, 6–20.
6. Blair IA (2008) DNA adducts with lipid peroxidation products. J Biol Chem 283,
15545–15549.
7. Brenna JT, Carlson SE (2014) Docosahexaenoic acid and human brain development:
evidence that a dietary supply is needed for optimal development. J Human Evol 77,
99–106, doi: 10.1016/j.jhevol.2014.02.017.
8. Kamat CD, Gadal S, Mhatre M, Williamson KS, Pye QN, Hensley K (2008) Antioxidants
in central nervous system disease: preclinical promise and translational challenges. J
Alzheimers Dis 15, 473–493.
9. Penn JS, Naash MI, Anderson RE (1987) Effect of light history on retinal antioxidants
and light damage susceptibility in the rat. Exp Eye Res 44, 779–788.
10. Penn JS, Anderson RE (1987) Effect of light history on rod outer-segment membrane
composition in the rat. Exp Eye Res 44, 767–778.
This article is protected by copyright. All rights reserved.

11. Zhang Y, Turunen M, Appelkvist EL (1996) Restricted uptake of dietary coenzyme Q is
in contrast to the unrestricted uptake of α-tocopherol into rat organs and cells. J
Nutr 126, 2089–2097, doi: 10.1093/jn/126.9.2089.
12. Atkinson J, Harroun T, Wassall SR, Stillwell W, Katsaras J (2010) The location and
behaviour of α-tocopherol in membranes. Mol Nutr Food Res 54, 641–665, doi:
10.1002/mnfr.200900439.
13. Urano S, Yano K, Matsuo M (1988) Membrane-stabilizing effect of vitamin E: effect of
alpha-tocopherol and its model compounds on fluidity of lecithin liposomes.
Biochem Biophys Res Commun 150, 469–475.
14. Wang X, Quinn PJ (2000) The location and function of vitamin E in membranes. Mol
Membr Biol 17, 143–156.
15. Xu L, Davis TA, Porter NA (2009) Rate constants for peroxidation of polyunsaturated
fatty acids and sterols in solution and in liposomes. J Am Chem Soc 131, 13037–
13044, doi: 10.1021/ja9029076.
16. Shchepinov MS (2007) Reactive oxygen species, isotope effect, essential nutrients,
and enhanced longevity. Rejuvenation Res 10, 47–59, doi: 10.1089/rej.2006.0506.
17. Hill S, Hirano K, Shmanai VV, Marbois BN, Vidovic D, Bekish AV, Kay B, Tse V, Fine J,
Clarke CF, Shchepinov MS (2011) Isotope-reinforced polyunsaturated fatty acids
protect yeast cells from oxidative stress. Free Radic Biol Med 50, 130–138, doi:
10.1016/j.freeradbiomed.2010.10.690.
18. Hill S, Lamberson CR, Xu L, To R, Tsui HS, Shmanai VV, Bekish AV, Awad AM, Marbois
BN, Cantor CR, Porter NA, Clarke CF, Shchepinov MS (2012) Small amounts of
isotope-reinforced polyunsaturated fatty acids suppress lipid autoxidation. Free
Radic Biol Med 53, 893–906, doi: 10.1016/j.freeradbiomed.2012.06.004.
19. Shchepinov MS, Chou VP, Pollock E, Langston JW, Cantor CR, Molinari RJ, Manning-
Bog AB (2011) Isotopic reinforcement of essential polyunsaturated fatty acids
diminishes nigrostriatal degeneration in a mouse model of Parkinson’s disease.
Toxicol Lett 207, 97–103, doi: 10.1016/j.toxlet.2011.07.020.
20. Berbee JFP, Mol IM, Milne GL, Pollock E, Hoeke G, Lutjohann D, Monaco C, Rensen
PCN, van der Ploeg LHT, Shchepinov MS (2017) Deuterium-reinforced
polyunsaturated fatty acids protect against atherosclerosis by lowering lipid
peroxidation and hypercholesterolemia. Atherosclerosis 264, 100–107, doi:
10.1016/j.atherosclerosis.2017.06.916.
21. Elharram A, Czegledy NM, Golog M, Milne GL, Pollock E, Bennett BM, Shchepinov MS
(2017) Deuterium-reinforced polyunsaturated fatty acids improve cognition in a
This article is protected by copyright. All rights reserved.

mouse model of sporadic Alzheimer’s disease. FEBS J 284, 4083–4095, doi:
10.1111/febs.14291.
22. Raefsky SM, Furman R, Milne G, Pollock E, Axelsen P, Mattson MP, Shchepinov MS
(2018) Deuterated polyunsaturated fatty acids reduce brain lipid peroxidation and
hippocampal amyloid beta-peptide levels, without discernible behavioural effects in
an APP/PS1 mutant transgenic mouse model of Alzheimer’s disease. Neurobiol Aging
66, 165–176, doi: 10.1016/j.neurobiolaging.2018.02.024.
23. Zesiewicz T, Heerinckx F, De Jager R, Omidvar O, Kilpatrick M, Shaw J,. Shchepinov
MS (2018) Randomized, clinical trial of RT001: early signal of efficacy in Friedreich’s
ataxia. Mov Disord 33, 1000–1005, doi: 10.1002/mds.27353.
24. Hatami A, Zhu C, Relaño-Gines A, Elias C, Galstyan A, Jun M, Milne G, Cantor CR,
Chesselet MF, Shchepinov MS (2018) Deuterium-reinforced linoleic acid lowers lipid
peroxidation and mitigates cognitive impairment in the Q140 knock in mouse model
of Huntington's disease. FEBS J, doi: 10.1111/febs.14590.
25. Angelova PR, Horrocks MH, Klenerman D, Gandhi S, Abramov AY, Shchepinov MS
(2015) Lipid peroxidation is essential for α-synuclein-induced cell death. J
Neurochem 133, 582–589, doi: 10.1111/jnc.13024.
26. Shchepinov MS, Roginsky VA, Brenna JT, Molinari RJ, To R, Tsui H, Clarke CF,
Manning-Bog AB (2014) Deuterium protection of polyunsaturated fatty acids against
lipid peroxidation: a novel approach to mitigating mitochondrial neurological
diseases. In Omega 3 Fatty Acids in Brain and Neurological Health (R.S. Watson & F.
De Meester, eds.), pp. 373–383, Elsevier Academic Press, Amsterdam.
27. Lamberson CR, Xu L, Muchalski H, Montenegro-Burke JR, Shmanai VV, Bekish AV,
McLean JA, Clarke CF, Shchepinov MS, Porter NA (2014) Unusual kinetic isotope
effects of deuterium reinforced polyunsaturated fatty acids in tocopherol-mediated
free radical chain oxidations. J Am Chem Soc 136, 838–841, doi: 10.1021/ja410569g.
28. Andreyev AY, Tsui HS, Milne GL, Shmanai VV, Bekish AV, Fomich MA, Pham
MN, Nong Y, Murphy AN, Clarke CF, Shchepinov MS (2015) Isotope-reinforced
polyunsaturated fatty acids protect mitochondria from oxidative stress. Free Radic
Biol Med 82, 63–72, doi: 10.1016/j.freeradbiomed.2014.12.023.
29. Vladimirov UA, Parnev OM, Annaberdyeva EM, Putvinskij AB, Putchkova TV (1984)
Electric Stability of Mitochondrial Membranes. Biol Membr (Rus) 4, 428-434.
30. Chretien D, Benit P, Ha HH, Keipert S, El-Khoury R, Chang YT, Jastroch M, Jacobs HT,
Rustin P, Rak M (2018) Mitochondria are physiologically maintained at close to 50°C.
PLoS Biol 16(1), e2003992, doi: 10.1371/journal.pbio.2003992.
This article is protected by copyright. All rights reserved.

31. Lane N (2018) Hot mitochondria? PLoS Biol 16(1), e2005113, doi:
10.1371/journal.pbio.2005113.
32. Reed TT (2011) Lipid peroxidation and neurodegenerative disease. Free Radic Biol
Med 51, 1302–1319, doi: 10.1016/j.freeradbiomed.2011.06.027.
33. Yagi K (1987) Lipid peroxides and human diseases. Chem Phys Lipids 45, 337–351.
34. Anderson RE, Rapp LM, Wiegand RD (1984) Lipid peroxidation and retinal
degeneration. Curr Eye Res 3, 223–227.
35. Sayre LM, Zelasko DA, Harris PLR, Perry G, Salomon RG, Smith MA (1997) 4-
Hydroxynonenal-derived advanced lipid peroxidation end products are increased in
Alzheimer’s disease. J. Neurochem 68, 2092–2097.
36. Butterfield DA, Lauderback CM (2002) Lipid peroxidation and protein oxidation in
Alzheimer’s disease brain: potential causes and consequences involving amyloid
beta-peptide-associated free radical oxidative stress. Free Radic Biol Med 32, 1050–
1060.
37. Dexter DT, Carter CJ, Wells FR, Javoy-Agid F, Agid Y, Lees A, Jenner P, Marsden CD
(1989) Basal lipid peroxidation in Substantia Nigra is increased in Parkinson’s disease.
J Neurochem 52, 381–389.
38. Pedersen WA, Fu W, Keller JN, Markesbery WR, Appel S, Smith RG, Kasarskis E,
Mattson MP (1998) Protein modification by the lipid peroxidation product 4-HNE in
the spinal cords of amyotrophic lateral sclerosis patients. Annals Neurol 44, 819–824.
39. Perevoshchikova IV, Zorov DB, Antonenko YN (2008) Peak intensity analysis as a
method for estimation of fluorescent probe binding to artificial and natural
nanoparticles: tetramethylrhodamine uptake by isolated mitochondria. Biochim
Biophys Acta 1778, 2182–2190. doi: 10.1016/j.bbamem.2008.04.008.
40. Hess ST, Huang S, Heikal AA, Webb WW (2002) Biological and chemical applications
of fluorescence correlation spectroscopy: a review. Biochemistry 41, 697–705.
41. Krichevsky O, Bonnet G (2002) Fluorescence correlation spectroscopy: the technique
and its applications. Rep Prog Phys 65, 251–297, doi: 10.1088/0034-4885/65/2/203.
42. Yamamoto Y, Niki E, Kamiya Y, Shimasaki H (1984) Oxidation of lipids. 7. Oxidation of
phosphatidylcholines in homogeneous solution and in water dispersion. Biochim
Biophys Acta 795, 332–340.
43. Fomich MA, Bekish AV, Vidovic D, Lamberson CR, Lysenko IL, Lawrence P, Brenna JT,
Sharko OL, Shmanai VV, Shchepinov MS (2016) Full library of (bis-allyl)-deuterated
This article is protected by copyright. All rights reserved.

arachidonic acids: synthesis and analytical verification. Chemistry Select 1, 4758–
4764, doi: 10.1002/slct.201600955.
44. Rosell M, Villa M, Durand T, Galano J‐M, Vercauteren J, Crauste C (2017) Total
syntheses of two bis‐allylic‐deuterated DHA analogues. Asian J Org Chem 6(3), 322–
334, doi: 10.1002/ajoc.201600565.
45. Smarun AV, Petkovic M, Shchepinov MS, Vidovic D (2017) Site-specific deuteration of
polyunsaturated alkenes. J Org Chem 82, 13115–13120. doi:
10.1021/acs.joc.7b02169.
46. Neises B, Steglich W (1978) Simple method for esterification of carboxylic acids.
Angew Chem Internat Ed 17, 522–524.
47. Nishimura S (2001) Hydrogenation of Alkenes. In Handbook of Heterogeneous
Catalytic Hydrogenation for Organic Synthesis, pp. 64-147, John Wiley &Sons, Inc.,
NY.
48. Olson F, Hunt CA, Szoka FC, Vail WJ, Papahadjopoulos D (1979) Preparation of
liposomes of defined size distribution by extrusion through polycarbonate
membranes. Biochim Biophys Acta 557, 9–23.
49. Pashkovskaya A, Kotova E, Zorlu Y, Dumoulin F, Ahsen V, Agapov I, Antonenko Y
(2010) Light-triggered liposomal release: membrane permeabilization by
photodynamic action. Langmuir 26, 5726–5733, doi: 10.1021/la903867a.
50. Yu L, Ding JL, Ho B, Feng SS, Wohland T (2008) Investigation of the mechanisms of
antimicrobial peptides interacting with membranes by fluorescence correlation
spectroscopy. The Open Chemical Physics Journal 1, 62–80, doi:
10.2174/1874412500801010062.
51. Blicher A, Wodzinska K, Fidorra M, Winterhalter M, Heimburg T (2009) The
temperature dependence of lipid membrane permeability, its quantized nature, and
the influence of anesthetics. Biophys J 96, 4581–4591, doi:
10.1016/j.bpj.2009.01.062.
52. Sorochkina AI, Kovalchuk SI, Omarova EO, Sobko AA, Kotova EA, Antonenko YN
(2013) Peptide-induced membrane leakage by lysine derivatives of gramicidin A in
liposomes, planar bilayers, and erythrocytes. Biochim Biophys Acta 1828, 2428–
2435. doi: 10.1016/j.bbamem.2013.06.018.
53. Kristensen K, Henriksen JR, Andresen TL (2014) Quantification of leakage from large
unilamellar lipid vesicles by fluorescence correlation spectroscopy. Biochim Biophys
Acta 1838, 2994–3002, doi: 10.1016/j.bbamem.2014.08.007.
This article is protected by copyright. All rights reserved.