Dissociation of proton-bound complexes reveals geometry and arrangement of double bonds in unsaturated lipids

Article abstract of DOI:10.1016/j.ijms.2015.07.006

Double bond position and stereochemistry in unsaturated lipids can have profound impact on biological properties and activities but the assignment of these features by mass spectrometry is frequently challenging. Conventional techniques for lipid identification rely on collision-induced dissociation (CID) and are most often unable to differentiate between lipid isomers, particularly those involving double bond position and geometry (i.e., cis and trans). In this study, CID performed on proton-bound complexes of fatty acid methyl esters and iodoaniline (and related reagents) reveals unusual fragmentation patterns. CID products are shown to result from proton transfer and are associated with specific structures of the unsaturated lipids. Notably, CID of these complexes can not only distinguish cis- and trans-fatty acid methyl esters, but also differentiate conjugated double bond arrangements from non-conjugated analogs. Herein, the mechanisms underpinning this unique CID behavior are investigated by stable isotope labeling and are proposed to involve both carbene and free radical intermediates.

Full text of DOI:10.1016/j.ijms.2015.07.006

International Journal of Mass Spectrometry 390 (2015) 170–177
Contents lists available at ScienceDirect
International Journal of Mass Spectrometry
journal homepage: www.elsevier.com/locate/ijms
Dissociation of proton-bound complexes reveals geometry and
arrangement of double bonds in unsaturated lipids
Huong T. Pham , Matthew B. Prendergast , Christopher W. Dunstan , Adam J. Trevittb,e,
a
b
b
d
a,∗
b,c,e,∗∗
Todd W. Mitchell , Ryan R. Julian , Stephen J. Blanksby
Department of Chemistry, University of California, Riverside, USA
School of Chemistry, University of Wollongong, Australia
Central Analytical Research Facility, Queensland University of Technology, Australia
School of Medicine, University of Wollongong, Australia
ARC Centre of Excellence for Free Radical Chemistry and Biotechnology, Australia
a
b
c
d
e
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 2 May 2015
Received in revised form 13 July 2015
Accepted 15 July 2015
Available online 23 July 2015
Double bond position and stereochemistry in unsaturated lipids can have profound impact on biological
properties and activities but the assignment of these features by mass spectrometry is frequently chal-
lenging. Conventional techniques for lipid identification rely on collision-induced dissociation (CID) and
are most often unable to differentiate between lipid isomers, particularly those involving double bond
position and geometry (i.e., cis and trans). In this study, CID performed on proton-bound complexes of
fatty acid methyl esters and iodoaniline (and related reagents) reveals unusual fragmentation patterns.
CID products are shown to result from proton transfer and are associated with specific structures of the
unsaturated lipids. Notably, CID of these complexes can not only distinguish cis- and trans-fatty acid
methyl esters, but also differentiate conjugated double bond arrangements from non-conjugated analogs.
Herein, the mechanisms underpinning this unique CID behavior are investigated by stable isotope labeling
and are proposed to involve both carbene and free radical intermediates.
Keywords:
Lipids
Isomers
Non-covalent complex
Collision-induced dissociation
Radicals
Carbenes
© 2015 Elsevier B.V. All rights reserved.
1
. Introduction
Double bond stereochemistry and positional arrangement in
undergo ionization in either positive or negative ion mode electro-
spray ionization (ESI). Subsequent CID yields diagnostic products
that allow the determination of double bond position(s) [5,6]. In
the same manner, Moe et al. pre-treated phospholipids and free
unsaturated lipids can have profound impact on biological prop-
erties and activities [1–3]. However, determination of these critical
features by mass spectrometry is challenging. Conventional low
energy collision-induced dissociation (CID) is the mainstay of mod-
ern lipid analysis but for isomeric lipids CID mass spectra are often
similar or identical, particularly where isomers differ only in dou-
ble bond position and geometry (i.e., cis and trans isomers) [4].
As a result, many studies have attempted to enhance CID with
either chemical derivatization of carbon-carbon double bonds or
the introduction of targeted radical dissociation processes in order
to obtain greater structural information. One example of the former
is to convert olefinic bonds to secondary ozonides, which can then
fatty acids with osmium tetroxide (OsO ) prior to electrospray
4
ionization. This method derivatized double bonds to vicinal diols,
which are readily cleaved upon CID, enabling identification of initial
double bond location [7]. Another method employs ion-molecule
reactions between unsaturated lipids and ozone in the gas phase,
namely ozone-induced dissociation (OzID) for online assignment
of the double bond position(s) [8,9]. Moreover, the differences
in reaction rates for gas-phase ozonolysis (and thus product ion
abundance) have been exploited to discriminate between lipid geo-
metric isomers (i.e., cis and trans isomers) [10]. Dramatic differences
in gas phase reaction rates have also been noted between conju-
gated and non-conjugated linoleic acid methyl esters [11]. Although
capable of detecting subtle structural differences, OzID technology
is currently not widely available.
∗ Corresponding author at: Department of Chemistry, University of California,
Riverside, USA. Tel.: +1 951 827 3958.
∗
versity of Technology, Australia. Tel.: +61 731383343.
E-mail addresses: ryan.julian@ucr.edu (R.R. Julian),
stephen.blanksby@qut.edu.au (S.J. Blanksby).
Radical-based dissociation techniques (e.g., EID, ETD, ECD and
RDD) have been applied to lipid analysis with the aim to induce
carbon-carbon bond cleavages and thus to reveal more of the intra-
chain bonding motifs (i.e., double bond position, chain branching,
∗ Corresponding author at: Central Analytical Research Facility, Queensland Uni-
http://dx.doi.org/10.1016/j.ijms.2015.07.006
1
387-3806/© 2015 Elsevier B.V. All rights reserved.

H.T. Pham et al. / International Journal of Mass Spectrometry 390 (2015) 170–177
171
etc.). For example, Yoo and Håkansson generated radical precur-
2.2. Sample preparation
sor ions at ∼10 eV electron energy and performed electron-induced
dissociation (EID) on unsaturated fatty acids with Mn2+ attached at
the carboxylic acid moiety. The resulting fragmentation produced
patterns associated with the double bond positions within the
acyl chain [12]. Electron-capture dissociation (ECD) and electron-
transfer dissociation (ETD) have also been applied to lipid structural
analysis but in general do not yield greater insight than con-
ventional CID [13,14]. Specifically, ECD and ETD examination of
phospholipids yielded product ions identifying the lipid headgroup
and the chain length and degree of unsaturation of the acyl chains
but did not reveal information that is specific to the site of unsatu-
ration nor stereochemistry of the double bond(s).
We have previously demonstrated another radical approach,
namely radical directed dissociation (RDD), to be a sensitive struc-
tural tool for both proteins [15,16] and lipids [17–20]. In the RDD
approach, bifunctional molecules containing a photo-activated rad-
ical initiator are adducted (either covalently or non-covalently)
to a lipid, ionized and then subjected to laser irradiation in an
ion-trap mass spectrometer. Subsequent, collisional activation of
the resulting radicals yields rich and structurally informative frag-
mentation, especially for elucidating carbon-bonding motifs within
lipid acyl chain substituents. For example, RDD was shown to
discriminate between branched and straight-chain phospholipid
isomers, and was effectively deployed to assign carbon-carbon
double bond position(s) in unsaturated triacylglycerols [17]. Inter-
estingly however, once the radical has been introduced into
the carbon chain, it can scramble the geometry of the unsat-
urated center, making it insensitive for distinguishing cis and
trans structures [18]. Serendipitously however, CID of some of the
same non-covalent complexes between lipids with iodo-containing
reagents—which had been initially designed for RDD—are found
to undergo surprising even-electron fragmentation that is strongly
associated with the geometry of the carbon-carbon double bond.
This observation is described herein for fatty acid methyl ester
(1)
The procedure to convert fatty acids (FA) to fatty acid methyl
esters (FAME) was described previously and is summarized in Eq.
1) [20]. Briefly, the FA (∼1 mg) was treated with BF3 in a 10%
(
methanol solution at room temperature (RT) and then extracted
into a non-polar solvent (e.g., n-pentane). Fresh samples were pre-
pared for mass spectrometry by collecting the upper n-pentane
layer (∼3 mM) and diluting in methanol to yield of final solution
of 10–20 ␮M in the resulting FAME. Then, each iodine-containing
reagent was added to a sample of this solution to a final con-
centration of 5–10 ␮M before adding 0.05% formic acid to aid the
formation of protonated 4-iodoaniline (pIA) for charge adducting.
(2)
Deuterium-exchanged experiments were undertaken as sum-
marized in Eq. (2), whereby a mixture of D -methanol and D O
1
2
(2:1) was used to dilute 1 ␮L of the n-pentane layer to exchange all
three protons in protonated amine group (−ND3+).
(3)
Preparation of the D -labeled FAME 1,1,1-trideuteromethyl (Z)-
3
(
FAME) lipids with cis/trans and conjugated/non-conjugated struc-
octadec-9-enoate is summarized in Eq. (3). In this procedure, 9Z-
ocatadecanoic acid (1 mg, 4 ␮mol) placed in D -methanol (CD OH,
tures. The fragmentation mechanism is investigated by stable
isotope labeling experiments and provides further understanding
that may support future applications of this approach to the field
of lipidomics.
3
3
◦
100 ␮L), with the addition of sulfuric acid (2%), was heated at 70 C
for 15 min in an oil bath and then allowed to cool for 15 min. Milli-Q
water (1 mL) and n-pentane (1 mL) were added to the solution. After
vigorous shaking, to ensure thorough mixing, the mixture was left
◦
for 30 min at 4 C to allow the separation of the aqueous and organic
layers. The n-pentane layer was collected with a Pasteur pipette and
was prepared for mass spectrometric analysis as described above.
2
. Experimental methods
2.1. Materials
The following fatty acid standards: oleic acid FA 18:1(9Z), elaidic
acid FA 18:1(9E), cis-vaccenic acid FA 18:1(11Z), trans-vaccenic acid
FA 18:1(11E), palmitoleic acid FA 16:1(9Z), palmitelaidic acid FA
(
4)
Methyl (Z)-2,2-dideutero-octadec-9-enoate was prepared by
adapting a procedure previously used to synthesize methyl
2,2-dideuteropentanoate and is summarized in Eq. (4) [21].
D -Methanol (CH OD, 8.13 g, 246 mmol) was cooled in an ice-
bath under a nitrogen atmosphere and sodium metal (75.4 mg,
3.28 mmol) was added. After the metal had dissolved, 3.25 mL of
the solution was added to methyl (Z)-octadec-9-enoate (1.00 g,
3.71 mmol) and the resulting mixture heated at reflux under nitro-
gen for 48 h. The solvent was removed in vacuo to yield the product
as brown oil. The sample was then prepared for mass spectrometric
analysis as described above.
1
6:1(9E) were purchased from Nu-Chek Prep (Elysian, Minnesota)
and were all ∼99% purity. Three linoleic acid methyl ester isomers,
FAME 18:2(9Z,12Z), FAME 18:2(10E,12Z) and FAME 18:2(9Z,11E),
1
3
were also purchased from Nu-Chek Prep (Elysian, Minnesota). D
-
7
1
ꢀ
ꢀ
Oleic acid (containing 17 deuterium atoms at the 11, 11 , 12, 12 ,
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀꢀ
1
∼
3, 13 , 14, 14 , 15, 15 , 16, 16 , 17, 17 , 18, 18 , and 18 positions,
99% deuterium incorporation) was obtained from Cayman Chem-
ical (Ann Arbor, MI).
Boron trifluoride-methanol 10% solution, D -methanol
1
(
CH OD), D -methanol (CD OH), deuterium oxide (D O) and
3 3 3 2
all iodo-containing reagents: 2-iodoaniline (oIA), 3-iodoaniline
mIA), 4-iodoaniline (pIA), 4-iodobenzylamine (IBzA) and 3-iodo-4-
(
2.3. Mass spectrometry
methylaniline (IMeA) were obtained from Sigma–Aldrich (St. Louis,
MO). Other HPLC grade solvents, such as methanol, chloroform
and n-pentane, were purchased from Thermo-Fisher Scientific
Experiments were performed using a linear ion-trap mass
spectrometer, Thermo Fisher Scientific LTQ (San Jose, California).
Sample solutions were introduced into the electrospray ionization
(
Waltham, MA) and were used without further purification.

1
72
H.T. Pham et al. / International Journal of Mass Spectrometry 390 (2015) 170–177
by direct infusion to generate the gaseous non-covalent complexes.
Typical source parameters included: spray voltage +4.0–4.5 kV;
◦
capillary temperature 250 C, capillary voltage 5–11 V, and tube
lens voltage 25–40 V. Nitrogen served as sheath, sweep and aux-
iliary gases and all were set to between 5 and 20 (arbitrary
units) and helium was used as buffer/collision gas. Ions were
mass-selected with an isolation width of 1–3 Da and activated by
collision-induced dissociation (CID) using default instrument MSn
parameters. Mass spectra were typically averaged over 50–100
scans to obtain the desired signal-to-noise.
3
. Results and discussions
3.1. Differentiation of cis and trans isomers
Neutral glycerolipids readily form non-covalent complexes with
ammonium ions and CID of the resulting complexes is widely used
for structural analysis [22]. Such complexation is similarly effi-
cient for aniline-derived compounds [17,18]. Fig. 1(a-e) shows the
CID spectra obtained for complexes of FAME 18:1(9Z) with (a) 4-
iodobenzylamine and (b-e) four different iodo-aniline variants. The
most abundant product ion in each case corresponds to the ammo-
nium cation (specific structures shown in red) occurring at either
m/z 234 or m/z 220, depending on the mass of the adducting amine.
These ions are formed by simple dissociation of the non-covalent
bond and neutral loss of the lipid. Fig. 1(a) shows an additional
product ion at m/z 217, which is a secondary loss of ammonia
Fig. 1. CID of non-covalent complexes between (a-e) oleic acid methyl ester FAME
18:1(9Z) and five different iodine-containing reagents with structures noted in red
(
f) elaidic acid methyl ester FAME 18:1(9E) and pIA. The 20× magnification indicated
at the top is applied to all spectra.
(
2
−17 Da) originating from iodobenzyl ammonium cation at m/z
34.
In addition to the abundant ammonium ions observed in all
and, in the case of the para-iodoaniline variant, to the stereochem-
istry of the double bond. Substitution of bromo- for iodo-aniline
within the proton-bound complex results in loss of the m/z 295
feature for all isomers (one such example is shown in Supporting
Information, Figure S-1). Furthermore, these product ions are not
observed if the complex is subjected to UV irradiation, which pri-
marily results in homolytic cleavage of the carbon-iodine bond and
subsequent radical directed dissociation products.
To examine the generality of these observations, several addi-
tional lipids were examined with the pIA reagent. Fig. 2 shows
the CID mass spectra for two pairs of geometrical isomers, namely
FAME 18:1(11Z) and FAME 18:1(11E) (Fig. 2a and b, respectively)
and FAME 16:1(9Z) and FAME 16:1(9E) (Fig. 2c and d, respec-
tively). Again the ratio of the product ion abundances of m/z 297
and m/z 295 for the 18:1 isomers and the equivalent m/z 269 and
spectra shown in Fig. 1, for the mIA and pIA complexes (Fig. 1d and
e, respectively) there are also two interesting product ions of appre-
ciable abundance at m/z 297 and m/z 295. Of these peaks, only the
m/z 297 peak is observed for mIA (Fig. 1d), while both m/z 295 and
2
97 are of comparable abundance for the pIA (Fig. 1e). Neither of
these product ions are observed for the IBzA complex (Fig. 1a), and
they are only present at very low abundance for the IMeA and oIA
complexes (Fig. 1b and c, respectively). Focusing on the pIA com-
plexing reagent which produces both the m/z 295 and 297 ions,
the ratio of these two peaks was found to invert if the stereochem-
istry of the double bond is cis or trans as illustrated with the FAME
1
8:1(9E) isomer in Fig. 1(f). Therefore, the products at m/z 295 and
2
97 are sensitive to both the structure of the complexation reagent
Fig. 2. CID of non-covalent complexes of [FAME + pIA]+ precursor ion from two pairs of stereoisomers (a) FAME 18:1(11Z), (b) FAME 18:1(11E), (c) FAME 16:1(9Z) and (d)
FAME 16:1(9E).

H.T. Pham et al. / International Journal of Mass Spectrometry 390 (2015) 170–177
173
to yield m/z 311 with the former process being favored. The data in
Fig. 3(e) suggest that hydrogens from either half of the lipid can be
lost in formation of the marker ion at m/z 295. Given the previously
demonstrated stereochemical sensitivity of this pathway, the most
likely positions are the two sites allylic to the carbon-carbon double
bond.
Based on the results presented in Figs. 1–3, we propose the
mechanism shown in Scheme 1 to explain the diagnostic prod-
uct ions at m/z 295 and 297. As the first step toward formation
of the product at m/z 295, a proton is transferred from the amine
to the ortho-position on the aniline ring. The energetics of the ring-
protonated tautomers of the anilinium cation have previously been
calculated, and the ortho-site is found to have a proton affinity only
−
1
1
1 kJ mol lower than the nitrogen while the para-site is slightly
−
1
favored over nitrogen for protonation by 8 kJ mol [23]. While the
effect of iodine on the relative energetics of these tautomers has
not been reported, it is reasonable to suppose that protonation
of the ortho-, meta- and para-positions of iodoanilines is simi-
larly thermoneutral. While protonation of the ring is energetically
reasonable, barriers to intra-molecular proton transfer have been
calculated to be substantial, e.g., proton transfer from nitrogen to
−
1
the ortho-carbon is calculated to be 238 kJ mol for the anilinium
cation. In this case however, the analogous proton migration can
be catalyzed by the adducted lipid. The methyl ester moiety in par-
ticular could play a crucial role in shuttling the proton between the
nitrogen and the ring and even between positions on the ring at rel-
atively modest energetic cost [24]. Once the proton has migrated
to the para-position, the activated system is configured for loss
of hydrogen iodide producing a resonance-stabilized carbene and
placing the formal charge back on the nitrogen. This intermediate
carbene could insert into an allylic C H bond on either side of the
double bond, temporarily fusing the two molecules. At this point,
the lipid-ring bond is cleaved in an electron transfer process driven
by formation of aniline and the lipid is released as a carbocation at
m/z 295. The meta-iodoaniline cannot form a resonance-stabilized
carbene intermediate, which may explain the absence of a product
ion at m/z 295 in Fig. 1(b) or (d). Although the ortho-iodoaniline
could form a resonance-stabilized carbene, it is likely that proxim-
ity to the amine prevents loss of HI by facilitating the reverse proton
transfer. The combination of these resonance and steric effects can
account for the spectral differences between the ortho-, meta- and
para-isomers observed in Fig. 1. Furthermore in this mechanism, a
hydrogen is abstracted from the lipid without transfer of a proton to
the lipid, which is consistent with the absence of deuterium scram-
bling observed in Fig. 3(d) and the loss of deuterium observed in
+
Fig. 3. CID spectra acquired for (a) the unlabeled standard [FAME 18:1(9Z) + pIA]
+
at m/z 516, (b) the D2-labeled [D2-FAME 18:1(9Z) + pIA] at m/z 518, (c) the D3-
labeled [D3-FAME 18:1(9Z) + pIA] at m/z 519, (d) the D3-exchanged complexes
FAME 18:1(9Z) + D3-pIA] at m/z 519 and (e) the D17-labeled lipid standard [D17-
FAME 18:1(9Z)-d17 + pIA] at m/z 533. Deuterium incorporation is illustrated in
+
+
[
+
Supporting Information (Figure-S2).
m/z 267 ion pair for the 16:1 isomers are shown to be sensitive
to the stereochemistry of the carbon-carbon double bond. Fur-
thermore, these data demonstrate that the dissociation pathway
yielding these fragments can occur at different double bond loca-
tions (i.e., 9- and 11-positions) and with varying chain lengths (i.e.,
1
6- and 18-carbon variants).
3.2. Deuterium labeling
In order to probe the mechanisms yielding the diagnostic
products ions identified above, a series of deuterium-labeling
experiments were carried out and the results are summarized in
Fig. 3. Where the lipid carries the deuterium label, direct dissocia-
tion of the complex yields unlabeled anilinium ions at m/z 220 (see
Fig. 3b, c and e) while exchange of the anilinium protons for deu-
terium results in a quantitative shift of this ion to m/z 223 (Fig. 3d).
These observations suggest that there is no hydrogen scrambling
within the complex prior to dissociation. The 2 Da spacing between
fragments of interest observed at m/z 295 and 297 for the unlabeled
archetype (Fig. 3a) is conserved when the lipid is deuterium labeled
at the labile ␣-carbon (Fig. 3b) or the ester methyl group (Fig. 3c).
This suggests that neither site plays a significant mechanistic role
in the fragmentation processes. Exchange of the anilinium protons
Fig. 3(e). Finally, the insertion of the carbene into the allylic C H
bond is central to the proposed mechanism. The energetics of these
bonds will be subtly different within the cis and trans isomers pro-
viding a rationale for the differing relative abundance of the m/z
295 and 297 signals for each of the lipids (see examples in Fig. 2).
Collision energy was found to alter the abundance ratio of the m/z
295 and 297 product ions but the m/z 295 channel is consistently
favored for the trans isomer while the m/z 297 channel is preferred
for the cis isomer at all energies explored (Supporting Information,
Figure S-5). The observed variation of the m/z 295 and 297 prod-
uct ion abundances with collision energy (see breakdown curves
shown in Supporting Information, Figure S-6) is consistent with the
different kinetic constraints of the two competing processes shown
in Scheme 1 (i.e., proton transfer compared to rearrangement).
The ion at m/z 297 arises from simple proton transfer from
the aniline to the lipid. The near exclusive transfer of a deuteron
(Fig. 3d) increases the spacing between the ions of interest to 3 Da.
Importantly, the peak observed at m/z 298 results from transfer
of a deuteron to the lipid, increasing the mass by 1 Da while the
peak at m/z 295 is not shifted, suggesting no transfer of a deuteron
or scrambling with the lipid occurs in the formation of this ion. In
Fig. 3(e), labeling of the final 17-positions of the lipid alkyl chain
is shown to produce loss of both two hydrogen atoms yielding a
product ion at m/z 312 and one hydrogen and one deuterium atom
from D -iodoanilinium cation to the lipid (Fig. 3d) suggests that
3
proton transfer from the ring protonated states does not occur to
a significant extent. If this were the case, significant scrambling
would be expected since the deuteron transfer to ring creates the
potential for both deuteron and proton transfer. The overall low

1
74
H.T. Pham et al. / International Journal of Mass Spectrometry 390 (2015) 170–177
Scheme 1. Proposed mechanism for the formation of m/z 297 and 295 product ions shown in Fig. 3(a) from CID of the non-covalent complex [FAME 18:1(9Z) + pIA]+ m/z 516
precursor ion.
abundance of the protonated lipid at m/z 297 compared to the
anilinium and benzylammonium ions at m/z 220 and 234 (see
Fig. 1) is consistent with the relative expected proton affinities
−
1
of the lipid (e.g., PA[CH CO CH ] = 822 kJ mol ) and aniline (e.g.,
3
2
1
3
−
PA[C H5NH ] = 883 kJ mol ) [25].
6
2
3.3. Conjugated/non-conjugated lipids
Lipids commonly carry multiple double bonds with the rel-
ative arrangement and stereochemistry of these motifs giving
rise to many isomeric variants. Examination of the dissociation
behavior of pIA complexes with polyunsaturated lipids containing
both conjugated and non-conjugated double bonds are shown in
Fig. 4.
As seen in Fig. 4(a), the CID fragmentation for non-conjugated
FAME 18:2(9Z,12Z) is very similar to that observed for the mono-
unsaturated cases previously discussed (see for example, Fig. 1e
and f). This implies that non-conjugated double bonds react in
the same fashion as single double bonds. The spectra in Fig. 4(b)
and (c) are notably different from Fig. 4(a), revealing that con-
jugated isomers (9Z,11E) and (10E,12Z) exhibit behavior distinct
from the non-conjugated variant. There are several new or sig-
nificantly enhanced reaction channels for the conjugated isomers.
Peaks at m/z 387, 263, and 245 represent new products. The peak
at m/z 295 is analogous to the m/z 297 signal for the monounsat-
urated lipids discussed earlier but is much more abundant for the
conjugated system, rivaling the abundance of the iodoanilinium
cation at m/z 220. The fragments at m/z 263 and 245 are sec-
ondary products arising from the loss of methanol (−32 Da) and
methanol/water (−50 Da) from the protonated lipid m/z 295. The
relatively high abundance of m/z 295 ion in these spectra indi-
cates a significantly higher proton-affinity of conjugated FAME
compared to its non-conjugated isomer. Deuterium labeling indi-
cates that the proton is derived from the anilinium nitrogen (see
Fig. 4. CID spectra of non-covalent complexes [FAME + pIA]+ from three isomers:
a) non-conjugated FAME 18:2(9Z,12Z), (b) conjugated FAME 18:2(10E,12Z) and (c)
conjugated FAME 18:2(9Z,11E).
(
Supporting Information, Figure S-4) as previously observed for the
mono-unsaturated lipids. The observation of greater abundance
of the protonated FAME for the conjugated systems arises in part
from the greater proton affinity of the conjugated double bonds

H.T. Pham et al. / International Journal of Mass Spectrometry 390 (2015) 170–177
175
Scheme 2. Proposed mechanism with initial protonation of conjugated double bonds then inducing an electrophilic addition of the lipid to the aniline to form m/z 387 radical
ion and two major MS3 product ions at m/z 216 and 302 from [FAME 18:2(9Z,11E) + pIA] in Fig. 5b.
+
themselves due to the potential for resonance stabilization of the
equal to that of pIA. Preference for proton transfer to the conjugated
FAMEs interferes with the diagnostic dehydrogenation pathways
that report on double bond geometry; however, the abundance of
the proton transfer product itself clearly distinguishes the conju-
gated FAME from its non-conjugated isomers (Fig. 4).
−
1
resulting carbocation (e.g., PA[2E-butene] = 747 kJ mol compared
−1
to PA[1,3-butadiene] = 783 kJ mol ) [25]. When combined with
additional stabilization afforded by interactions with the oxygens
on the lipid ester moiety, the proton affinity appears to be nearly
Fig. 5. MS3 mass spectra performed on the radical product ions at m/z 387 generated in CID of the two non-covalent complexes between protonated pIA and conjugated
isomers: (a) FAME 18:2(10E,12Z) and (b) FAME 18:2(9Z,11E).

1
76
H.T. Pham et al. / International Journal of Mass Spectrometry 390 (2015) 170–177
The product ion observed at m/z 387 in Fig. 4(b) and (c) is
acknowledge funding from the National Science Foundation (CHE
1401737). The authors thank Hill Harman (UCR) for stimulating
discussion.
generated by neutral loss of atomic iodine and is unique to the
conjugated FAMEs. The observation of the m/z 387 ion in these
spectra is unusual as low energy CID does not commonly give rise
to radical ions, a phenomenon commonly referred to as the even-
electron rule [26]. Interestingly, this feature is not observed for
mIA as shown in Figure S-3 (Supporting Information). We propose
the mechanism shown in Scheme 2 to account for these obser-
vations and suggest that the same facile transfer of the proton
between the anilinium ion and the conjugated double bond motif,
as described above, can facilitate an electrophilic addition of the
lipid to the aniline. Addition of the intermediate carbocation at
the para-position on the aromatic ring will lower the bond dis-
sociation energy of C I bond and thus result in the homolytic
cleavage of this bond and loss of atomic iodine. The resonance
stabilization of the intermediate carbocation and the resulting rad-
ical by the aniline nitrogen is critical to these processes and thus
explains the absence of such pathways for mIA. Evidence for the
proposed covalent addition of the aniline to the alkyl chain is pro-
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.ijms.2015.07.006
References
[1] C.-h. Huang, H.-n. Lin, S. Li, G. Wang, Influence of the positions of cis dou-
ble bonds in the sn-2-acyl chain of phosphatidylethanolamine on the Bilayer’s
melting behavior, J. Biol. Chem. 272 (35) (1997) 21917–22192.
[
[
[
2] Y. Lin, E. Schuurbiers, S. Van der Veen, E.A.M. De Deckere, Conjugated
linoleic acid isomers have differential effects on triglyceride secretion in
Hep G2 cells, Biochim. Biophys. Acta - Mol Cell Biol Lipids 1533 (1) (2001)
38–46.
3] H.J. Heipieper, F. Meinhardt, A. Segura, The cis-trans isomerase of unsaturated
fatty acids in Pseudomonas and Vibrio: biochemistry, molecular biology and
physiological function of a unique stress adaptive mechanism, FEMS Microbiol.
Lett. 229 (1) (2003) 1–7.
4] S.J. Blanksby, T.W. Mitchell, Advances in mass spectrometry for lipidomics,
Annu. Rev. Anal. Chem. 3 (1) (2010) 433–465.
3
vided by the MS spectra obtained from the radical anions of m/z
3
87 (Fig. 5). These spectra reveal abundant product ions at m/z 216
and m/z 302 that are consistent with cleavage of the alkyl chain
occurring at or near the site of the aniline attachment. Consistent
with this explanation, the abundance of the product ions arising
from the intermediate radical cations appear to be sensitive to
the arrangement of double bonds in the parent lipid and thus the
[5] K.A. Harrison, R.C. Murphy, Direct mass spectrometric analysis of ozonides:
application to unsaturated glycerophosphocholine lipids, Anal. Chem. 68
(
1996) 3224–3230.
[
[
6] S.R. Ellis, J.R. Hughes, T.W. Mitchell, M.i.h. Panhuis, S.J. Blanksby, Using ambient
ozone for assignment of double bond position in unsaturated lipids, Analyst 137
3
MS spectra could be used to distinguish between conjugated lipid
(
5) (2012) 1100–1110.
isomers.
7] M.K. Moe, T. Anderssen, M.B. Strom, E. Jensen, Total structure characterization
of unsaturated acidic phospholipids provided by vicinal di-hydroxylation of
fatty acid double bonds and negative electrospray ionization mass spectrome-
try, J. Am. Soc. Mass Spectrom. 16 (2005) 46–59.
4
. Conclusions
[
[
8] M.C. Thomas, T.W. Mitchell, D.G. Harman, J.M. Deeley, J.R. Nealon, S.J. Blanksby,
Ozone-induced dissociation elucidation of double bond position within mass-
selected lipid ions, Anal. Chem. 80 (1) (2008) 303–311.
9] T.W. Mitchell, H.T. Pham, M.C. Thomas, S.J. Blanksby, Identification of double
bond position in lipids: from GC to OzID, J. Chromatogr. B 877 (26) (2009)
2722–2735.
The photodissociation of non-covalent complexes of iodo-
anilines with neutral lipids has previously been demonstrated as
an analytically powerful approach to lipid structure elucidation.
This RDD approach is particularly useful for the assignment of
position(s) of carbon-carbon double bonds and chain branching
motifs in lipids [17]. Here we have shown, for the first time, that
CID of these same complexes can also give rise to diagnostic frag-
ments that report on complementary structural features, the most
important of which is the stereochemistry of the carbon-carbon
double bonds. While these diagnostic CID product ions were not
observed for all the unsaturated lipids investigated, the fact that
they can be derived from the same non-covalent complex used for
RDD suggests that in future workflows may be possible that switch
rapidly between RDD and CID analysis to more fully describe the
structure of unknown lipids. Importantly, the diagnostic trend in
ion abundance ratios is preserved across a wide range of collision
energies (Supporting Information Figure S-5) and is also observed
on instruments of different geometries (Supporting Information
Figure S-6).
[
[
[
10] B.L.J. Poad, H.T. Pham, M.C. Thomas, J.R. Nealon, J.L. Campbell, T.W. Mitchell,
S.J. Blanksby, Ozone-induced dissociation on a modified tandem linear ion-
trap: observations of different reactivity for isomeric lipids, J. Am. Soc. Mass
Spectrom. 21 (12) (2010) 1989–1999.
11] H. Pham, A. Maccarone, J.L. Campbell, T. Mitchell, S. Blanksby, Ozone-induced
dissociation of conjugated lipids reveals significant reaction rate enhancements
and characteristic odd-electron product ions, J. Am. Soc. Mass Spectrom. 24 (2)
(2013) 286–296.
12] H.J. Yoo, K. Håkansson, Determination of double bond location in fatty acids by
manganese adduction and electron induced dissociation, Anal. Chem. 82 (16)
(
2010) 6940–6946.
[13] P.F. James, M.A. Perugini, R.A.J. O’Hair, Electron capture dissociation of com-
plexes of diacylglycerophosphocholine and divalent metal ions: competition
between charge reduction and radical induced phospholipid fragmentation, J.
Am. Soc. Mass Spectrom. 19 (7) (2008) 978–986.
[14] X. Liang, J. Liu, Y. LeBlanc, T. Covey, A. Ptak, J. Brenna, S. McLuckey, Electron
transfer dissociation of doubly sodiated glycerophosphocholine lipids, J. Am.
Soc. Mass Spectrom. 18 (10) (2007) 1783–1788.
[
15] T. Ly, R.R. Julian, Residue-specific radical-directed dissociation of whole pro-
teins in the gas phase, J. Am. Chem. Soc. 130 (1) (2008) 351–358.
The chemistries giving rise to the diagnostic CID product ions are
proposed to involve carbene and free radical intermediates formed
within the non-covalent complex (see Schemes 1 and 2). If found to
be general, such process could be exploited for structure elucidation
in other classes of compounds or indeed, to explore the bimolecular
gas phase chemistries of reactive intermediates.
[16] X. Zhang, R.R. Julian, Investigating the gas phase structure of KIX with radical
directed dissociation and molecular dynamics: retention of the native struc-
ture, Int. J. Mass Spectrom. 308 (2–3) (2011) 225–231.
[
[
[
[
17] H.T. Pham, T. Ly, A.J. Trevitt, T.W. Mitchell, S.J. Blanksby, Differentiation of com-
plex lipid isomers by radical-directed dissociation mass spectrometry, Anal.
Chem. 84 (17) (2012) 7525–7532.
18] H.T. Pham, A.J. Trevitt, T.W. Mitchell, S.J. Blanksby, Rapid differentiation of iso-
meric lipids by photodissociation mass spectrometry of fatty acid derivatives,
Rapid Commun. Mass Spectrom. 27 (7) (2013) 805–815.
19] H.T. Pham, R.R. Julian, Mass shifting radical delivery with crown ether attach-
ment for separation and analysis of phosphatidylethanolamine lipids, Anal.
Chem. 86 (6) (2014) 3020–3027.
20] H.T. Pham, R.R. Julian, Radical delivery and fragmentation for structural analysis
of glycerophospholipids, Int. J. Mass Spectrom. 370 (2014) 58–65.
Acknowledgements
S.J.B., A.J.T. and T.W.M. are grateful to the Australian Research
Council (DP150101715 and DP120102922) and the University of
Wollongong for funding. T.W.M. is an Australian Research Coun-
cil Future Fellow (FT110100249) and S.J.B. and A.J.T. are supported
by the Australian Research Council’s Centre of Excellence for Free
Radical Chemistry and Biotechnology (CE0561607). R.R.J and H.T.P
[21] R.D. Bowen, H.G.M. Edwards, D.W. Farwell, S.E. Morgan, Raman spectroscopy of
n-pentyl methyl ether and deuterium labelled analogues, J. Raman Spectrosc.
41 (12) (2010) 1725–1734.
[
22] A.M. McAnoy, C.C. Wu, R.C. Murphy, Direct qualitative analysis of triacylglyce-
rols by electrospray mass spectrometry using a linear ion trap, J. Am. Soc. Mass
Spectrom. 16 (9) (2005) 1498–1509.

H.T. Pham et al. / International Journal of Mass Spectrometry 390 (2015) 170–177
177
[
[
23] R. Flammang, N. Dechamps, L. Pascal, Y. Van Haverbeke, P. Gerbaux, P.C. Nam,
M.T. Nguyen, Ring versus nitrogen protonation of anilines, Lett. Org. Chem. 1
[25] P.J. Linstrom, W.G. Mallard (Eds.), NIST Chemistry WebBook, NIST Standard
Reference Database Number 69, National Institute of Standards and
Technology, Gaithersburg, MD 20899, 2015 http://webbook.nist.gov/
(retrieved April 30, 2015).
[26] K. Miriam, A. Mandelbaum, The ‘even-electron rule’, Org. Mass Spectrom. 15
(2) (1980) 53–64.
(
1) (2004) 23–30.
24] A.J. Chalk, L. Radom, Proton-transport catalysis: a systematic study of the
rearrangement of the isoformyl cation to the formyl cation, J. Am. Chem. Soc.
1
19 (32) (1997) 7573–7578.